afab-vol.1-issue 2

Volume 1, Issue 2November 2011

Sooyoun Ahn Arkansas State University, USA

Walid Q. AlaliUniversity of Georgia, USA

Kenneth M. Bischoff NCAUR, USDA-ARS, USA

Claudia S. Dunkley University of Georgia, USA

Lawrence GoodridgeColorado State University, USA

Leluo GuanUniversity of Alberta, Canada

Joshua GurtlerERRC, USDA-ARS, USA

Yong D. HangCornell University, USA

Divya JaroniSouthern University, USA

Weihong Jiang Shanghai Institute for Biol. Sciences, P.R. China

Michael JohnsonUniversity of Arkansas, USA

Timothy KellyEast Carolina University, USA

William R. KenealyMascoma Corporation, USA

Hae-Yeong Kim Kyung Hee University, South Korea

W.K. KimUniversity of Manitoba, Canada

M.B. KirkhamKansas State University, USA

Todd KostmanUniversity of Wisconsin, Oshkosh, USA

Y.M. Kwon University of Arkansas, USA

Maria Luz Sanz MuriasInstituto de Quimica Organic General, Spain

Melanie R. MormileMissouri University of Science and Tech., USA

Rama NannapaneniMississippi State University, USA

Jack A. Neal, Jr.University of Houston, USA

Benedict OkekeAuburn University at Montgomery, USA

John PattersonPurdue University, USA

Toni Poole FFSRU, USDA-ARS, USA

Marcos RostagnoLBRU, USDA-ARS, USA

Roni ShapiraHebrew University of Jerusalem, Israel

Kalidas ShettyUniversity of Massachusetts, USA

EDITORIAL BOARD

EDITOR-IN-CHIEFSteven C. RickeUniversity of Arkansas, USA

EDITORSTodd R. CallawayFFSRU, USADA-ARS, USA

Cesar CompadreUniversity of Arkansas for Medical Sciences, USA

Philip G. CrandallUniversity of Arkansas, USA

EDITORIAL STAFFMANAGING EDITOR

Ellen J. Van LooGhent, Belgium

LAYOUT EDITORMelody Rust

Eureka Springs Arkansas, USA TECHNICAL EDITOR

Jessica C. ShabaturaFayetteville Arkansas, USA

ONLINE EDITION EDITORC.S. ShabaturaFayetteville Arkansas, USA

ABOUT THIS PUBLICATION

Agriculture, Food & Analytical Bacteriology (ISSN

2159-8967 is published quarterly, beginning with this

inaugural issue.

Instructions for Authors may be obtained at the

back of this issue, or online via our website at

www.afabjournal.com

Manuscripts: All correspondence regarding pend-

ing manuscripts should be addressed Ellen Van Loo,

Managing Editor, Agriculture, Food & Analytical

Bacteriology: [email protected]

Information for Potential Editors: If you are interested

in becoming a part of our editorial board, please con-

tact Editor-in-chef, Steven Ricke, Agriculture, Food &

Analytical Bacteriology: [email protected]

Advertising: If you are interested in advertising with

our journal, please contact us at advertising@afab-

journal.com for a media kit and current rates.

Reprint Permission: Correspondence regarding re-

prints should be addressed Ellen Van Loo, Managing

Editor, Agriculture, Food & Analytical Bacteriology

[email protected]

Ordering Print Copies: print editions of this journal

may be purchased and shipped internationally from

our website order form at www.afabjournal.com

Subscription Rates: Subscriptions are not available

at this time. To be advised when subscriptions plans

are made available, please join our newsletter at

www.afabjournal.com

Mailing Address: 637 S. Ray Ave. . Fayetteville, AR . 72701 Website: AFABjournal.com

TABLE OF CONTENTS

Production of Cyclosporin A by Static Fermentation Using Tolypocladium inflatum MTCC 557 S. A. Survase, U. S. Annapure and R. S. Singhal

105

A Transcriptomic Expression Array, PCR and Disk Diffusion Analysis of Antimicrobial Resistance Genes in Multidrug-Resistant Bacteria S. A. Khan, K. Sung and M. Nawaz

123

Antibiotic Resistance and Plasmid Profiles in Bacteria Isolated from Market-Fresh Vegetables S. Akter, Rafiq-Un-Nabi, F. Ahmed Rupa, Md. L. Bari and M. A. Hossain

140

Survival of Salmonella in Organic and Conventional Broiler Feed as Affected by Temperature and Water Activity A. Petkar, W. Q. Alali, M. A. Harrison and L. R. Beuchat

175

Influence of Winter and Summer Hutch Coverings on Fecal Shedding of Pathogenic Bacteria in Dairy Calves R. L. Farrow, T. S. Edrington, B. Carter, T. H. Friend, T. R. Callaway, R. C. Anderson and D. J. Nisbet

98

Comparative Studies on the Survival of Verocytotoxigenic Escherichia coli and Salmonella in Different Farm Environments C. J. O’Neill, D. J. Bolton and S. Fanning

116

Minimizing the Risk of Listeria monocytogenes in Retail Delis by Developing Employee Focused, Cost Effective Training P. G. Crandall, J. A. Neal Jr., C. A. O’Bryan, C. A. Murphy, B. P. Marks and S. C. Ricke

159

ARTICLES

REVIEWS

Instructions for Authors193

EXTRAS

The publishers do not warrant the accuracy of the articles in this journal, nor any views or opinions by their authors.

BRIEF COMMUNICATIONS

Impact of Calcium Chloride Dip and Temperature on Microbial Quality of Organically and Conventionally Grown Melons H. T. Aldrich, L. Goodridge, M. Bunning, C. Stushnoff and P. Kendall

150

Isolation and Initial Characterization of Plasmids in an Acetogenic Ruminal Isolate R. S. Pinder and J. A. Patterson

186

Agric. Food Anal. Bacteriol. • AFABjournal.com • Vol. 1, Issue 2 - November 2011 98

www.afabjournal.comCopyright © 2011

Agriculture, Food and Analytical Bacteriology

ABSTRACT

The effects of hutch coverings utilized during the summer and winter months to moderate extreme tem-

peratures were examined on fecal prevalence of E. coli O157:H7 and Salmonella in newborn dairy calves.

In the initial study, the effects of shade using screens in three treatment groups: no shade, partial shade,

and full shade were examined. Two additional studies were designed where individual calf hutches were

modified with a hutch blanket (treatment) or no hutch blanket (control) in the winter study and a ventilated

hutch design added as a third treatment in the summer study. During the summer experiment, prevalence

of E. coli O157:H7 and Salmonella was low; however, Salmonella was increased (P < 0.05) in the ventilated

hutch versus the control treatment. In the winter study, quantifiable results for both E. coli O157:H7 and

Salmonella were largely negative. Salmonella positive samples were numerically higher, however no treat-

ment differences were observed. In the shade cloth study all fecal samples were E. coli O157:H7 negative.

Salmonella was cultured from all treatment groups, however no differences were observed between treat-

ments. Summarily, there is no evidence that hutch treatments decreased the period prevalence of fecal

shedding of Salmonella, E. coli O157:H7 or Enterococcus.

Keywords: Dairy calves, Salmonella, E. coli, Hutch covering

INTRODUCTION

Dairy producers have long realized the need for

individualized care of newborn calves from both man-

agement and health perspectives. Dairy calves are

born without circulating antibodies that are critical

Correspondence: T. S. Edrington, [email protected]: +1 -979-260-3757 Fax: +1-979-260-9332

for immune function and current management prac-

tices strive to insure that newborns receive adequate

amounts of colostrum to develop immunity to a host

of potentially pathogenic organisms; however, these

calves have an increased susceptibility to disease (Roy,

1970). In 2006 the U. S. Department of Agriculture re-

ported a mortality rate of 11 percent for pre-weaned

dairy calves with the majority of those deaths resulting

from enteric and respiratory pathogens (USDA, 2006).

BRIEF COMMUNICATION Influence of Winter and Summer Hutch Coverings on Fecal Shedding

of Pathogenic Bacteria in Dairy CalvesR. L. Farrow1, T. S. Edrington1, B. Carter2, T. H. Friend2, T. R. Callaway1, R. C. Anderson1, and D. J. Nisbet1

1Food and Feed Safety Research Unit, ARS, USDA, College Station, TX 778452Texas A&M University, Department of Animal Science, College Station, TX 77843

‡Mention of trade name, proprietary product, or specific equipment does not constitute a guarantee or warranty by the USDA and

does not imply its approval to the exclusion of other products that may be suitable.

Agric. Food Anal. Bacteriol. 1: 98-104, 2011

99 Agric. Food Anal. Bacteriol. • AFABjournal.com • Vol. 1, Issue 2 - November 2011

Calf hutches have been used as a means to reduce

direct contact among newborn calves in an effort to

reduce transmission of disease. Dairy calves that

were comingled had an increased prevalence of E.

coli O157:H7 compared to calves that were housed

individually (Garber et al., 1995). Poos and Sordillo

(1982) reported that calves housed in hutches had

improved growth and performance as well as re-

duced mortality when compared to other methods.

In extreme cold weather environments, calves reared

in hutches had an overall better performance, ate

more starter feed and required fewer medical treat-

ments (McKnight, 1978).

In warm climates, supplemental shade used dur-

ing the summer months decreased severity of the

heat stress experienced by calves when compared

to those in hutches alone (Spain and Spears, 1996).

Scott et al. (1976) reported that calves exposed to

chronic heat loads had lower IgG than those that

were housed at thermo-neutrality. Given the afore-

mentioned research, it is hypothesized that reducing

temperature variations in periods of extreme heat

and cold will reduce the level of stress in calves and

thereby increase their resistance to pathogenic or-

ganisms. The objectives of this research are to de-

termine if the application of shade to dairy hutches

decreases fecal prevalence of E. coli O157:H7 and

Salmonella in newborn dairy calves.

MATERIALS AND METHODS

This research was conducted on a single, large

commercial Holstein dairy (> 2000 head) located in

the Texas Panhandle and managed as typical for dair-

ies in Eastern New Mexico and the Texas Panhandle.

During the periods at which the observations took

place for the summer and winter studies tempera-

tures averaged approximately 24°C (18.1 - 31.4°C)

and 4.9°C (-2.9 - 12.7°C), respectively. At the time of

these studies this dairy was a closed herd that reared

its own replacement females. Three studies were de-

signed to evaluate the effects of hutch modifications

on pathogen shedding. The first study (shade cloth)

evaluated the effects of a suspended shade cloth

over hutches on pathogen shedding. The second

(winter) and third (summer) studies evaluated the ef-

fects of reflective insulation applied to calf hutches;

during the summer study an additional treatment

designed to increase hutch ventilation was incorpo-

rated. Holstein heifer calves were placed alternately

in treatments immediately after birth and managed

as typical for dairy calves in this region. Heifers were

fed approximately 2.0 L of pasteurized waste milk

twice daily until they would drink from a bucket (1

to 3d), after which they were fed approximately 7.5L

pasteurized waste milk twice daily. As per protocol

of dairies typical of this region calves were provided

approximately 7.5L water upon consumption of milk,

and in addition calves were gradually introduced to

solid feeds between 10 to 14 d of age. All hutches

were commercially available polyethylene hutches

(Calf-Tel Pro®, Hampel Corp., Germantown, WI) af-

fixed with a 1 x 2 m outdoor pen made of welded

wire panels.

Shade Cloth Study

This study utilized 80% shade cloth (Sunblocker

Premium, Farmtek., Dyersville IA) to provide partial

(n = 4), full (n = 6) or no shade (n = 14). The cloth was

suspended 3 m above the hutches so that a portion of

the pen in front of the hutches was also shaded; par-

tially shaded hutches were considered as those that

received shade in the morning or evening because

they were located near the edge of the shade struc-

ture. Heifers were assigned to treatment as described

above. However, during the first collection calves had

only been placed in half of the hutches, resulting in

reduced sample sizes for partial (n = 2), full (n = 2) and

no shade (n = 9). Fecal samples were collected via

rectal palpation or from freshly voided, uncontami-

nated fecal pats within the hutch area on 20JUL2006

(day 1), 25AUG2006 (day 37), and 23SEP2006 (day 66).

Summer Study

As described below, hutch coverings were placed

on hutches upon placement of the calf in the hutch,

shortly following birth starting June 10, 2007 (day 1),


and were removed on September 1, 2007 (day 84). An

additional 10 hutches were modified by creating a 14

x 18 inch hole, 18 inches off the ground, in the back

wall of the hutch (ventilation treatment) to increase

air flow through the hutch. Heifers were assigned to

treatment as described above, calves were fed using

this opening on the inside the hutch, versus normal

feeding which occurred outside of the hutch in the

wire pen, thus allowing them to remain shaded while

eating. One calf in this type of hutch died early in

the experimental period and was not replaced result-

ing in 9 hutches of this type at each collection. Fecal

samples were collected as above on days 6, 19, 41,

65 and 84 of the experimental period as described

above.

Winter Study

During the winter and summer studies, treated

hutches were covered with a 2.2 x 2.5 m sheet of

Tempshield™ reflective insulation (Innovative Insula-

tion Inc., Arlington, TX). Grommets were place on

the edges of the “hutch blanket” to facilitate attach-

ment to the hutch in three places on each side of the

hutches long sides by elastic cord. Treated hutches

(n = 20) alternated with control hutches (n = 19) and

were all placed in a single row, contained within mul-

tiple rows of calf hutches. Heifers were assigned to

treatments as above. Hutch blankets were applied

on December 18, 2007 (day 1) and removed February

23, 2008 (day 68). Fecal samples were collected as

described above on days 29, 42, 56 and 68 of the ex-

perimental period. Fecal samples were collected us-

ing sterile palpation sleeves, placed on ice and trans-

ported to our laboratory in College Station, Texas for

bacterial culture described in the following section.

Bacterial Culture and Isolation

All fecal samples were processed the day follow-

ing collection for qualitative analysis of Salmonella

(Edrington et al., 2009) and E. coli O157:H7 as de-

scribed previously (Robinson et al., 2004) and modi-

fied (Brichta-Harhay et al., 2007). Enterococcus was

qualitatively cultured as previously described by

Edrington and co-workers (2009). Unless noted oth-

erwise, all reagents and antibiotics were obtained

from Sigma Chemical Co. (St. Louis, MO).

Antimicrobial susceptibility was determined on

isolates using the Sensititre automated antimicrobial

susceptibility system and the National Antibiotic Re-

sistance Monitoring System’s (NARMS) testing pan-

els (Trek Diagnostics Systems Inc., Cleveland, OH)

for isolates as described previously (Edrington et al.,

2009).

Statistical Analysis

Data were analyzed using SAS Version 9.2 (SAS

Inst. Inc., Cary, NC, USA). Quantitative enumeration

was infrequent and not sufficient to detect differenc-

es among treatments, therefore data is presented as

prevalence (% positive) and not actual counts. The

incidence of fecal pathogen shedding was subjected

to a Chi-square analysis using the PROC FREQ pro-

cedure. Additionally, the PROC MIXED procedure

was used to examine the main effects of treatment

and day, and treatment x day interaction. Differences

in means were considered significant at a 5% level of

significance.

RESULTS

Shade Cloth Study

When combined across sampling dates, partially

shaded calves shed Salmonella more frequently

(70%) when compared to no shade and full shade (54

and 43%, respectively), however no statistical differ-

ences between treatments were observed (Table 1).

Salmonella isolates (n = 29) from the first two collec-

tion times were analyzed for antimicrobial suscepti-

bility. Two isolates per positive sample from the first

collection period and a single isolate from collec-

tion two were subjected to antimicrobial screening.

Most isolates (86%) were pan susceptible; resistance

to sulphathiazole was observed in two isolates and

resistance to tetracycline was observed in a single

isolate. One isolate that displayed resistance to


chlortetracycline, oxytetracycline, and tetracycline

belonged to serogroup C2. No differences were

observed between treatments with respect to anti-

microbial resistance. All fecal samples were E. coli

O157:H7 negative throughout the study.

Summer Study

Fecal prevalence of E. coli O157:H7 was low

throughout the study, with 4 of 205 samples posi-

tive following enrichment and IMS (Table 2). Three

of those positive samples were cultured on d 84 in

the hutch covering treatment; however no significant

differences were observed between control and ven-

tilated hutch treatments (P = 0.06). No samples were

culture positive for E. coli O157:H7 using quantitative

methodology. Similarly, Salmonella was cultured in-

frequently following direct plating with positive sam-

ples detected only on d 84 in the control and ven-

tilated treatments (22% positive in each treatment).

Following enrichment however, Salmonella positive

fecal samples were detected on all collection days

except d 41. The highest incidence was observed

on days 19 and 84, although only on d 19 were sig-

nificant treatment differences observed. Salmonella

prevalence was increased (P < 0.05) in the ventilated

hutch compared to the control treatment, but was

similar for control and hutch blanket treatments

(17.2, 31.6 and 66.7% for control, hutch blanket and

ventilated treatments, respectively). When averaged

across collection days, only E. coli O157:H7 preva-

lence (as detected by IMS) was affected by treat-

ment, with a higher (P < 0.05) percentage of positive

samples in the hutch blanket treatment compared

to control and ventilated treatments. There were no

treatment x day interactions observed. Five different

Salmonella serogroups were identified (C1, C2, E1, K

and poly A-I, vi) with C2 accounting for 55% of the

isolates. Twenty-one Salmonella isolates were exam-

ined for antimicrobial susceptibility (7, 8 and 6 each

from the control, blanket and ventilated hutch treat-

ments, respectively) and all were susceptible to all of

the antibiotics examined (data not shown).

Enterococcus isolates (n = 14) cultured from hutch

calves on this farm during the shade study were ex-

amined for antimicrobial susceptibility. All 14 isolates

were resistant to quinupristin/dalfopristin and four of

these isolates were additionally resistant to erythro-

mycin, lincomycin, streptomycin, tetracycline, and

tylosin. No treatment differences were observed and

all Enterococcus isolates were resistant to vancomy-

cin (data not shown). Based on this data and similar

data generated in the Winter Study, Enterococcus

isolates were not examined in the summer hutch

covering study.

Winter Study

Results of the quantitative culture of E. coli

O157:H7 and Salmonella over the course of the win-

ter study were largely negative (Table 3). None of

the samples contained quantifiable populations of

Salmonella (0/136) while only two animals had quan-

tifiable concentrations of E. coli O157:H7 (d 56 and

Table 1. Prevalence (number and % positive) of Salmonella in fecal samples collected from dairy calves housed in control hutches or hutches with partial shade or full shade. Fecal samples were enriched for qualitative analysis prior to plating (ENR).

Control Partial Shade Full Shade

Day Method no. % no. % no. %

1 ENR 9/9 100 2/2 100 2/2 100

37 ENR 8/14 57 4/4 100 4/6 67

66 ENR 3/14 21 1/4 25 0/6 0.0

all days ENR 20/37 54 7/10 70 6/14 43

No significant differences between treatments were observed


68, control treatment). Enrichment of the samples

followed by IMS identified the same two positive

animals in the control treatment and an additional

six positive animals in the hutch blanket treatment

(2 on d 42, 4 on d 56). The number of Salmonella

positive samples increased slightly with enrichment,

although no treatment differences were observed.

The majority of serogrouped Salmonella belonged

to group K.

Antimicrobial susceptibility screening was con-

ducted on five E. coli O157:H7 isolates (data not

shown). All isolates were resistant to sulphisoxazole

and two of the five were also resistant to streptomy-

cin. Enterococcus isolates from d 29 (n = 32) and 42

(n = 11) collections were also examined (data not

shown). Most isolates however were susceptible to

all of the antibiotics examined with the exception of

quinupristin/dalfopristin, to which all but three were

resistant. Isolates (in both treatments) displayed re-

sistance to seven different antibiotics (chlorampheni-

col, erythromycin, lincomycin, quinupristin/dalfopris-

tin, streptomycin, tetracycline and tylosin) including

three isolates in the control treatment and two in the

hutch blanket treatment, accounting for most of the

observed resistance. Of the 11 isolates examined 13

d later on d 42, only one isolate in the hutch blanket

treatment was multi-resistant (6 antibiotics, exhibit-

ing the same pattern as the previous isolates). As

discussed above, the majority of the isolates were

susceptible to all antibiotics with the exception of

quinupristin/dalfopristin (data not shown). All En-

terococcus isolates were susceptible to vancomycin.

Due to the limited number of Salmonella positive

samples, isolates were not retained for antimicrobial

susceptibility screening.

DISCUSSION

Results of the current research highlight the spo-

radic nature of pathogen shedding in naturally-colo-

nized animals. Prevalence of E. coli O157:H7 and Sal-

monella was relatively low in this study compared to

previous research conducted by our laboratory ex-

amining pathogen prevalence in dairy cattle during

the summer months (Edrington et al., 2004). Even so,

Table 2. Prevalence (number and % positive) of E. coli O157:H7 (EC) and Salmonella (Salm) in fecal samples collected during the summer study, by day of collection (when at least on sample was cul-ture positive) and across all collection days, from dairy calves housed in control hutches or hutches modified with a blanket or ventilated to improve calf comfort. Fecal samples were either plated directly to quantify bacterial populations (DIR) or enriched for qualitative analysis prior to plating (ENR).

Control Hutch blanket Ventilated hutch

Bacteria Day Method no. % no. % no. %

EC 6 ENR 0/4 0 1/4 25 nsa .

84 ENR 0/18 0 3/15 20 0/9 0

All days ENR 0/100 0b 4/73 5.48c 0/32 0b

Salm 6 ENR 2/4 50 2/4 50 nsa .

19 ENR 5/29 17.2b 6/19 31.6b 6/9 66.7c

65 ENR 1/22 4.55 0/16 0 0/5 0

84 DIR 4/18 22.2 0/15 0 2/9 22.2

84 ENR 3/18 16.7 1/15 6.67 1/9 11.1

All days ENR 11/100 11 9/73 12.3 7/32 21.9

a no samples collected. bc Treatment means within a row with different superscripts differ (P < 0.05).


some significant treatment differences and trends

were noted in fecal shedding of E. coli O157:H7 and

Salmonella on various collection days, suggesting

hutch coverings increased pathogen prevalence. Vari-

ous scenarios including environmental changes within

the hutch due to hutch coverings may have contrib-

uted to this outcome, however identifying a single

cause is not possible. It is unclear to the authors what

contributed to a greater proportion of Salmonella

positive calves to have originated from the partially

shaded treatment relative to both no shade and full

shade within the shade study. It should be noted that

for a study of this type to achieve an α = 0.95 and a

ß = 0.20 with 20% prevalence expected in the control

versus 10% in the treated groups it would require a

sample of 219 animals per treatment group. Given the

resources available to the researchers sample sizes of

this magnitude were unobtainable. However, due to

the low number of sampling units used in this study

readers are cautioned not to over interpret these data.

Serogrouping of the Salmonella isolates identi-

fied the majority as belonging to groups C2 (summer

study), K (winter study) and C1 and C2 (shade cloth

study). This is not surprising as we have reported

seasonal differences in serogroup prevalence in

dairy cattle previously (Edrington et al., 2004). Sal-

monella Newport, a serotype frequently Multidrug-

Rresistant (MDR), belongs to serogroup C2. However,

antimicrobial susceptibility testing revealed no MDR

isolates; therefore it is likely that these isolates be-

longed to another common dairy serotype within the

C2 group, likely Kentucky. Serotyped isolates from

previous dairy research belonging to serogroup K,

were frequently identified as Cerro (Edrington et al.,

2004).

Antimicrobial susceptibility screening of the vari-

ous isolates yielded few MDR isolates (mostly En-

terococcus) and these were resistant to antimicro-

bials frequently utilized in veterinary medicine and

susceptible to antibiotics used in human medicine.

In summary, further research is necessary to pro-

vide additional evidence to support the use of hutch

coverings to moderate temperatures and conse-

quently increase calf comfort. These data indicate

that in an effort to increase animal level comfort

Table 3. Prevalence (number and % positive) of E. coli O157:H7 (EC) and Salmonella (Salm) in fecal samples collected during the winter study, by day of collection (when at least one sample was cul-ture positive) and across all collection days, from dairy calves housed in control hutches or hutches modified with a reflective blanket to improve calf comfort. Fecal samples were either plated direct-ly to quantify bacterial populations (DIR) or enriched for qualitative analysis prior to plating (ENR).

Control Hutch blanket

Bacteria Day Method no. % no. %

EC 42 ENR 0/17 0 2/17 11.8

56 DIR 1/18 5.6 2/15 13.3

56 ENR 1/18 5.6 4/15 26.7

68 DIR 1/18 5.6 0/16 0

68 ENR 1/18 5.6 0/16 0

All days DIR 2/71 2.8 2/65 3.1

All days ENR 2/71 2.8 6/65 9.2

Salm 29 ENR 1/18 5.6 0/17 0

42 ENR 0/17 0 1/17 5.9

56 ENR 2/18 11.1 1/15 6.7

All days ENR 3/71 4.2 2/65 3.1


dairy producers may inadvertently increase patho-

gen shedding. Additional studies should be focused

on locating herds that have high enough levels of

pathogen shedding and sufficient quantity of sub-

jects to discern measures of effect should they exist.

CONCLUSION

Efforts to mitigate extreme temperatures within

calf hutches via hutch coverings have been shown

to increase the pathogen burden of dairy calves

within this study. However, due to the small number

of observations further work is warranted to better

understand the effects of temperature modification

within hutches with various forms of shades and re-

flective material. Therefore, the authors recommend

that each dairy assess their needs and situation indi-

vidually to best determine how increase calf comfort

while ensuring that an increased pathogen burden

has not been imposed.

ACKNOWLEDGEMENT

The authors wish to thank the Food Animal Con-

cerns Trust for partial funding of this research and

Shannon Garey for assistance during sample collec-

tion.

REFERENCES

Brichta-Harhay D. M., T. M. Arthur, J. M. Bolsilevac,

M. N. Guerini, N. Kalchayanand, M. Koohmaraie.

2007. Enumeration if Salmonella and Escherichia

coli O157:H7 in ground beef, cattle carcass, hide

and faecal samples using direct plating methods.

J. Appl. Microbiol. 103:1657-1668.

Edrington, T. S., B. H. Carter, T. H. Friend, G. R. Ha-

gevoort, T. L. Poole, T. R. Callaway, R. C. Anderson,

D. J. Nisbet. 2009. Influence of sprinklers, used to

alleviate heat stress, on fecal shedding of E. coli

O157:H7 and Salmonella and antimicrobial suscep-

tibility of Salmonella and Enterococcus in lactating

dairy cattle. Lett. Appl. Microbiol. 48:738-743.

Edrington, T. S , M. E. Hume, M. L. Looper, C. L.

Schultz, A. C. Fitzgerald, T. R. Callaway, K. J. Geno-

vese, K. M. Bischoff, J. L. McReynolds, R. C. An-

derson, D. J. Nisbet. 2004. Variation in the faecal

shedding of Salmonella and E. coli O157:H7 in lac-

tating dairy cattle and examination of Salmonella

genotypes using pulsed-field gel electrophoresis.

Lett. Appl. Microbiol. 38:366-372.

Garber, L. P., S. J. Wells, D. D. Hancock, M. P. Doyle,

J. Tuttle, J. A. Shere, T. H. Zhao. 1995. Risk factors

for shedding Escherichia coli O157:H7 in dairy

calves. J. Am. Vet. Med. Assoc. 207:46-49.

McKnight, D. R. 1978. Performance of newborn dairy

calves in hutch housing. Can. J. Anim. Sci. 58:517-

520.

Poos, M. I. and L. Sordillo. 1982. The effect of type

of housing and supplementation on performance

of dairy calves from birth to weaning. J. Dairy Sci.

65:121.

Robinson, S. E, E. J. Wright, N. J. Williams, C. A. Hart,

and N. P. French. 2004. Development and applica-

tion of a spiral plating method for the enumera-

tion of Escherichia coli O157:H7 in bovine feces. J.

Appl. Microbiol. 97:581-589.

Roy, J. H. B. 1970. The calf: management and feed-

ing. Butterworths, London, UK. 224 p.

Scott, G. H., F. Wiersma, B. E. Menefee, F. R. Rad-

wanski. 1976. Influence of environment on passive

immunity in calves. J. Dairy Sci. 59:1306-1311.

Spain, J. N. and D. E. Spears. 1996. Effects of sup-

plemental shade on thermoregulatory response of

calves to heat challenge in a hutch environment. J.

Dairy Sci. 79:639-646.

USDA. 2006. Characterization and enhancement of

immune responses of calves. Annual Report FY

2006. United States Department of Agriculture,

Agricultural Research Services, National Animal

Disease Center, Ames, IA. p 1-5.




ABSTRACT

The effect of different nutrients (carbon and nitrogen sources) and fermentation factors (such as inocu-

lum size and volume of production media) was evaluated on the production of Cyclosporin A (CyA) in sta-

tionary culture of Tolypocladium inflatum MTCC 557 over a 21-day period at 25ºC. Glycerol was found to be

the best carbon source and gave a maximum CyA production of 410 mg/L. Further, response surface meth-

odology was used to optimize the concentrations of medium components which took the CyA production

to 452 mg/L. The exogenous supplementation of various amino acids individually and in combination was

also studied. The time of addition of the optimized combination of amino acids was also evaluated. A

maximum CyA production of 1241 mg/L was obtained when L-valine and L-leucine were added after 4th day

of the fermentation. The optimum concentrations of media components were (in g/L) glycerol 98.8, casein

peptone 28.5, malt extract 20, peptone 10 and α-amino butyric acid 5.7.

Keywords: cyclosporin A, Tolypocladium inflatum, stationary culture, response surface method, fermen-

tation, static, amino acids, production

INTRODUCTION

Cyclosporin A (CyA), a cyclic undecapeptide, is

one of the most commonly prescribed immunosup-

pressive drugs for the treatment of patients with or-

gan transplantation, autoimmune diseases, includ-

ing AIDS, owing to its superior T-cell specificity and

low levels of myelotoxicity (Kahan, 1984; Schindler,

1985). The organisms reported to produce CyA in-

Correspondence: Shrikant A. Survase, [email protected] Tel: 91-022-24145616 Fax: +91-022-24145614

clude Tolypocladium inflatum (Agathos et al., 1986),

Fusarium solani (Sawai et al., 1981), Neocosmospora

varinfecta (Nakajima et al., 1988) and Aspergillus

terreus (Sallam et al., 2003). CyA is reported to be

produced by submerged culture fermentation (Aga-

thos et al., 1986), static fermentation (Balaraman and

Mathew, 2006), solid state fermentation (Survase et

al., 2009a), and also synthesized enzymatically (Billich

and Zocher, 1987).

The production of CyA in submerged fermenta-

tion has been reported to vary with respect to the

production strain, fermentation conditions, and the

nutrient composition of the culture medium. The ef-

Production of Cyclosporin A by Static Fermentation Using Tolypocladium inflatum MTCC 557

S. A. Survase1, U. S. Annapure1 and R. S. Singhal1

1Food Engineering and Technology Department, Institute of Chemical Technology, Matunga, Mumbai 400 019, India



fect of media components such as carbon and nitro-

gen sources (Abdel-fattah et al., 2007; Agathos and

Parekh, 1990) and environmental factors such as pH

and inoculum density (Issac et al., 1990) on CyA pro-

duction have been studied.

Static culture conditions were used successfully

for production of glucansucrase using Leuconostoc

dextranicum NRRL B-1146 (Majumdar and Goyal,

2008), cordycepin by Cordyceps militaris CCRC

32219 (Shih et al., 2007) and exo-polysaccharide of

Agaricus brasiliensis (Fan et al., 2007). Balaraman and

Mathew, (2006) reported higher production of CyA

in static fermentation as compared to submerged

and solid state fermentation using Tolypocladium

sp (VCRC F21 NRRL No.18950). They studied me-

dium optimization with respect to composition and

reported maximum CyA production of 2.22 g/L me-

dium or 5.85 g/kg biomass. With this background we

have examined production of CyA using T. inflatum

MTCC 557.

Response surface methodology (RSM) is used

to evaluate the relative significance of variables in

the presence of complex interactions with limited

number of experiments. It has been successfully em-

ployed for optimization of medium constituents for

the production of metabolites such as cholesterol

oxidase (Chauhan et al., 2009) and cephamycin C

(Bussari et al., 2008).

To the best of our knowledge, there is just one

single report on optimization of the medium con-

stituents for the production of CyA by static fermen-

tation. In the present study, the effects of various

carbon sources, fermentation time, inoculum size

and production medium volume were investigated.

Thereafter, the concentration of media components

was optimized by using response surface methodol-

ogy. The effect of various amino acids was also stud-

ied for the first time in static fermentation.


Materials

Glucose, maltose, sucrose, glycerol, yeast extract,

casein peptone, bactopeptone, malt extract and

agar were procured from Himedia Ltd, Mumbai, In-

dia. Sodium chloride and calcium chloride; amino

acids L-valine, L-leucine, glycine, DL-amino butyric

acid, DL-methionine and solvents acetonitrile, n-bu-

tyl acetate, sodium hydroxide, concentrated hydro-

chloric acid and sulphuric acid were all purchased

from S. D. Fine Chemicals Ltd. Mumbai, India. All the

solvents used were of AR grade except acetonitrile,

which was of HPLC grade. Standard CyA (authentic

sample) was a gift sample through the kind courtesy

of RPG Life Sciences Ltd., Mumbai, India.

Microorganisms

Strains of T. inflatum MTCC 989, T. inflatum MTCC

557 (indicated as Beauveria nivea in the MTCC cata-

log), T. inflatum NCIM 1283, were procured from

MTCC, Chandigarh and NCIM, Pune, India. T. in-

flatum NRRL 18950 was a gift sample from the ARS

Culture Collection (NRRL), Peoria, Illinois, USA. The

cultures were maintained on agar slants containing

malt extract 2% and yeast extract 0.4% (MYA), pH 5.4

at 4°C after growing it for 12 days at 24°C.

Preparation of the seed inoculum and fermentation

The organism was subcultured onto a fresh MYA

slant and incubated at 25 ± 2°C for 12 days to a fully

grown slant. To this slant, 10 ml of sterile saline con-

taining 0.1% Tween 20 was added and well mixed. One

milliliter of this saline containing approximately 108 to

109 spores was added to 50 ml of medium composed

of malt extract 2%, yeast extract 0.4%, pH 5.4 taken in a

250 ml flask and incubated at 180 rpm for 72 h at 25 ±

2°C. This was used as the seed for static fermentation.

Seed inoculum (10 % v/v) was used to inoculate sterile

production medium. The fermentation was carried out

at 25 ± 2 ˚C, pH of 5.7 ± 0.2 for 21 days.

Screening of microorganisms and fer-mentation media

Four strains as indicated in the previous section

were screened for the maximum production of CyA


using medium reported by Balaraman and Mathew

(2006), and the strain which gave maximum produc-

tion was used for further optimization studies.

Three different production media were tested for

maximum production of CyA using T. inflatum MTCC

557. Various fermentation media screened were as

follows; medium I containing (g/L) glucose 80, casein

peptone 30, malt extract 20, bacto-peptone 10 and

DL-amino butyric acid 5 (Balaraman and Mathew,

2006); medium II containing (g/L) glucose 58.46, casein

peptone 8.66, KH2PO4 4.48 and KCl 3.23 (Survase et

al., 2009b); and medium III developed in our labora-

tory containing (gL) glucose 50, ammonium sulphate

15, K2HPO4 1.25. The production medium which sup-

ported maximum production of CyA was used as the

basal medium for optimization of CyA production.

Effect of fermentation parameters on production of CyA using T. inflatum MTCC 557

T. inflatum MTCC 557 was tested for the production

of CyA at different time intervals between 1 to 21 days

using medium I as the production medium. Fermenta-

tion medium was inoculated with different inoculum size

(5- to 20 % v/v), and its effect was observed on produc-

tion of CyA. Glucose in the basal medium was replaced

with various carbon sources such as sucrose, soluble

starch, maltodextrin and glycerol as carbon source for

the production of CyA using T. inflatum MTCC 557. All

carbon sources were evaluated at 8 % w/v in the culture

medium. Three nitrogen sources viz., casein peptone,

malt extract and bacto-peptone at different concen-

tration combinations (1:2:3; 1:3:2;2:3:1;2:1:3;3:2:1 and

3:1:2) were evaluated for maximum production of CyA.

The effect of volume of production medium was inves-

tigated by putting different volumes (25 to 150 ml) in

250 ml Erlenmeyer flasks.

Response surface methodology

A central composite rotatable design (CCRD) for

three independent variables was used to obtain the

combination of values that optimizes the response

within the region of three dimensional observation

spaces. The experiments were designed using the

software, Design Expert Version 6.0.10 trial version

(State Ease, Minneapolis, MN).

The medium components (independent variables)

selected for the optimization were glycerol, casein

peptone, and amino butyric acid. The experimental

design showing the coded as well as actual values of

Table 1. The central composite rotatable de-sign (CCRD) matrix of independent variables in coded as well as actual form with their corre-sponding response

RunGlycerol

(%)

Casein Peptone

(%)

Amino butyric acid (%)

CyAa (mg/l)

1 6 (-1)b 2 (-1) b 0.25 (-1) b 201± 14

2 10 (1) 2 (-1) 0.25 (-1) 221 ± 24

3 6 (-1) 4 (1) 0.25 (-1) 186 ± 15

4 10 (1) 4 (1) 0.25 (-1) 315 ± 17

5 6 (-1) 2 (-1) 0.75 (1) 187 ± 22

6 10 (1) 2 (-1) 0.75 (1) 274 ± 16

7 6 (-1) 4 (1) 0.75 (1) 145 ± 24

8 10 (1) 4 (1) 0.75 (1) 405 ± 25

9 4.63 (-1.68)

3 (0) 0.5 (0) 213 ± 12

10 11.36 (1.68)

3 (0) 0.5 (0) 431 ± 14

11 8 (0) 1.31 (-1.68)

0.5 (0) 155 ± 21

12 8 (0) 4.68 (1.68)

0.5 (0) 247 ± 12

13 8 (0) 3 (0) 0.07 (-1.68)

213 ± 33

14 8 (0) 3 (0) 0.92 (1.68) 274 ± 12

15 8 (0) 3 (0) 0.5 (0) 410 ± 27

16 8 (0) 3 (0) 0.5 (0) 412 ± 20

17 8 (0) 3 (0) 0.5 (0) 408 ± 25

18 8 (0) 3 (0) 0.5 (0) 411 ± 24

19 8 (0) 3 (0) 0.5 (0) 413 ± 26

20 8 (0) 3 (0) 0.5 (0) 411 ± 22

a values are mean ± SD of three determinationsb Values in the parenthesis are coded values of the independent variablesCyA is cyclosporin A


independent variables is shown in Table 1. Regres-

sion analysis was performed on the data obtained

from the designed experiments. The second order

polynomial coefficients were calculated to estimate

the responses of the dependent variable. Response

surface plots were also obtained using Design Ex-

pert Version 6.0.10.

Effect of amino acids

The effect of different amino acid members of

CyA molecule on drug production was evaluated

by replacing α-amino butyric acid in the fermenta-

tion media with different amino acids such as L-va-

line, L-leucine, DL-valine, L-methionine and glycine

at 6 g/L. They were screened individually as well as

in combination with others. The time of addition of

combination of amino acids (0 to 144 h) was also op-

timized to further increase the yield.

Analytical determinations

CyA extraction and estimation

The CyA extraction from the culture broth was car-

ried out according to the method of Agathos et al.

(1986). The fermentation broth was homogenized in a

blender. Ten milliliters of homogenized culture broth

was extracted with equal volume of n-butyl acetate.

Before extracting the sample, a concentrated solution

of NaOH was added to reach the concentration of 1N

and heated at 60°C for 30 min. The mixed sample was

kept on rotary shaker (180 rpm) for 24h. After centri-

fuging, the extract was filtered using Whatman filter

paper (No.1) and then through Pall 0.2 μm membrane

filter (Ultipor® N66® Nylon 6, 6 membranes) to give

clear extract. One milliliter of the extract was evapo-

rated under vacuum to dryness. The dried extract was

dissolved in equal volume (1 ml) of HPLC grade ace-

tonitrile. Twenty microliters of sample was analyzed

for CyA content using HPLC (Jasko system) fitted with

a reverse phase column Waters Sperisorb® ODS (C18

octadecyl silane, 250 X 4.6 mm ID) by the method de-

scribed by Survase et al. (2009a). The mobile phase

consisted of 70:30 ratio of acetonitrile and water with

a flow rate of 1 ml/min. The temperature of the col-

umn was maintained at 70°C and the HPLC profile

was monitored at 210 nm.

Estimation of biomass

A 10 ml sample of homogenized broth was centri-

fuged at 10,000 rpm for 20 min, washed twice with dis-

tilled water and taken on a preweighed Whatman fil-

ter paper. This was dried to a constant weight at 80°C.

RESULTS AND DISCUSSION

Screening of the available strains showed T. inflatum

MTCC 557 to produce maximum CyA (256 mg/L) under

static conditions followed by T. inflatum NRRL 18950

which gave 196 mg/L of CyA after 21 days. T. inflatum

MTCC 989 and T. inflatum NCIM 1283 gave lower titers

of CyA. Balaraman and Mathew (2006) reported use of

Tolypocladium sp. (VCRC F21 NRRL No.18950) for pro-

duction of CyA under static conditions. The reported

maximum CyA production was 2.22 g/L medium or

5.85 g/kg biomass. The difference in the yield could

be explained as a strain difference. They have used a

modified strain of T. inflatum NRRL 18950 whereas we

used a wild type strain for our studies.

Among the three media screened by using T. in-

flatum MTCC 557 for maximum production of CyA,

medium reported by Balaraman and Mathew (2006)

which contained glucose as carbon source and

combination of three different nitrogen sources was

found to be promising. Medium I produced 254 mg/L

of CyA followed by medium II which produced 145

mg/L. Biomass production was also higher 20.4 g/L

(measured as dry cell weight (DCW)) in medium I as

compared to other two media. Medium III produced

the lowest amount (97 mg/L) of CyA and biomass 10

g/L. Medium II reported by Survase et al. (2009b) pro-

duced maximum CyA yield of 134.5 mg/L from shake

flask cultures.

Three flasks after every two days from day 1 to 23

days were evaluated to study the effect of fermenta-

tion time on CyA production. It was observed that

production of CyA started after 3rd day of fermenta-

tion which reached maximum of 254 mg/L after 21

days. The CyA production was found to be decreased

to 238 mg/L on 23rd day which could be due to the


autolysis of fungal cells (Fig. 1). The biomass was also

found to be decreased after the 21st day.

It was observed that 15 % v/v of 72 h old seed

culture resulted in maximum CyA production of 312

mg/L as compared to 255 mg/l using 10 % v/v in-

oculum (data not shown). An increase in the seed

culture age and inoculum size did not increase the

yields significantly. Isaac et al. (1990) reported that

a higher spore density gave higher production of

CyA in submerged fermentation using T. inflatum

UAMH 2472. In our previous study (Survase et al.,

2009b), 10 % v/v inoculum was found to be opti-

mum under shaking conditions (180 rpm).

It was observed that variation in concentrations

of nitrogen sources viz. casein peptone, malt ex-

tract and bacto-peptone in the basal medium

changed the production of CyA (between 123 mg/L

to 252 mg/L) where DCW values ranging from 16.2

to 20.6 g/L were observed (Table 2). The optimum

concentrations included (in g/L) casein peptone 30,

malt extract 20 and bacto-peptone 10.

The effect of different carbon sources on produc-

tion of CyA using T. inflatum MTCC 557 is shown

in Fig. 2. It was observed that, glycerol supported

Figure 1. Effect of fermentation time on production of CyA using T. inflatum MTCC 557 (DCW is dry cell weight and CyA is cyclosporin A)

Table 2. Effect of various concentrations of ni-trogen sources on CyA production in the basal medium where, glucose was used as a carbon source at 80 g/l

Nitrogen Source (%)CyAa (mg/l)

DCWa (g/l)Casein

PeptoneMalt

extractBacto-

peptone

2 1 3197.4 ± 8.6

17.9 ± 0.58

3 1 2225.2 ± 7.5

19.7 ± 0.45

3 2 1252.5 ± 7.4

20.6 ± 0.64

1 3 2146.2 ± 8.2

16.2 ± 0.52

1 2 3188.2 ± 4.5

17.5 ± 0.55

2 3 1123.4 ± 6.8

17.4 ± 0.41

a values are mean ± SD of three determinationsDCW is dry cell weight and CyA is cyclosporin A

L


maximum production of 410 mg/L with biomass

production of 16.5 g/L DCW. This was followed

by sucrose which supported maximum production

of 287 mg/L. Glucose produced biomass of 19.65

g/L and CyA production was yielded 254 mg/L.

Survase et al. (2009b) reported glycerol as a car-

bon source in agitated culture system supported

biomass growth, but with a lower CyA production.

Survase et al. (2009a) also reported use of glycerol

as the best carbon source for supplementation in

solid state fermentation system. Abdel-fattah et

al. (2007) used three carbon sources as glucose,

sucrose and starch in combination yielded maxi-

mum CyA production using T. inflatum DSMZ 915

in agitated culture system. Agathos et al. (1986)

used sorbose as carbon source for production of

CyA using T. inflatum ATCC 34921.

The volume of production medium was varied

from 25 ml to 150 ml in 250 ml Erlenmeyer flasks

and its effect was observed on production of CyA

and biomass (data not shown). It was observed

that 50 ml of the production medium in 250 mL

flask produced maximum amount of CyA (249

mg/L) and biomass (19.25 g/L). With increasing

production medium volume, there was a decrease

in CyA production as well as biomass production.

Response Surface Methodology

The combined effect of three independent vari-

ables A: glycerol (g/L); B: casein peptone (g/L) and C:

α-aminobutyric acid (g/L) on production of CyA was

examined using CCRD with 20 experimental runs.

The experimental values of yields of CyA are given in

Table 1. The quadratic model was suggested for the

given set of experimental results.

The results were analyzed by using ANOVA i.e.

analysis of variance suitable for the experimental de-

sign used. The ANOVA of the quadratic model indi-

cated the model to be significant (Table 3). The Mod-

el F-value of 76.9 and Model P-value (Prob > F) of <

0.0001 implies the model to be significant.

The P values were used as a tool to check the sig-

nificance of each of the coefficients, which, in turn are

Figure 2. Effect of carbon source on CyA production using T. inflatum MTCC 557. All carbon sources were added at 8 % w/v (DCW is dry cell weight and CyA is cyclosporin A)


necessary to understand the pattern of the mutual

interactions between the test variables. The smaller

the magnitude of the P, the more significant is the

corresponding coefficient. Values of P less than 0.050

indicate the model terms to be significant. The coef-

ficient estimates and the corresponding P values sug-

gested that, the variables A, B, C, A2, B2 and C2, are

significant model terms whereas the interactions AB

and AC were significant. The second order response

model found after analysis for the regression was:

CyA (mg/L) = 535.42 - 113.26 (glyc-

erol) + 35.98 (casein peptone)

- 179.98 (amino butyric acid) + 4.20 (glycerol2) - 25.97

(casein peptone2) -175.18 (amino butyric acid2) + 17.62

(glycerol x casein peptone) + 49.50 (glycerol x amino

butyric acid) + 5.0 (casein peptone x amino butyric

acid) (1)

Eq. (1) represents the mathematical model relat-

ing the production of CyA with the independent

process variables, and the second order polynomial

coefficient for each term of the equation determined

through multiple regression analysis using the Design

Expert 6.0.10.

The fit of the model was also expressed by the

coefficient of regression (R2), which was found to be

0.99, indicating that 99 % of the confidence level of

the model to predict the response (CyA yield). The

“Pred R-Squared” of 0.98 is in reasonable agreement

with the “Adj R-Squared” of 0.99. “Adeq Precision”

measures the signal to noise ratio. A ratio greater

than 4 is desirable. Here, the ratio of 32 indicates an

adequate signal.

Accordingly, three-dimensional graphs (Fig. 3) were

Table 3. Analysis of variance (ANOVA) for the experimental resultsv of the central-composite design (Quadratic Model)

Source Coefficient estimate

Sum of Squares

F Value

Prob > F

Model 267.09 96322.82 76.9 < 0.0001

A 63.16 54487.8 391.51 < 0.0001

B 23.63 7626.31 54.8 < 0.0001

C 13.96 2659.79 19.11 0.0014

A2 16.8 4069.69 29.24 0.0003

B2 -25.98 9723.54 69.87 < 0.0001

C2 -10.95 1727.73 12.41 0.0055

AB 35.25 9940.5 71.43 < 0.0001

AC 24.75 4900.5 35.21 0.0001

BC 1.25 2.5 0.09 0.7705

Figure 3. 3D-surface plot for cyclosporin A pro-duction; A Effect of glycerol and casein pep-tone when other variables are held at zero lev-el; B Effect of glycerol and amino butyric acid when other variables are held at zero level

A

B


generated for the pair-wise combination of the three

variables, while keeping the other two at their center

point levels. The optimum value of the combination of

the three media constitutes was determined using the

special features of the RSM tool such as “contour plot

generation” and “point prediction” for the maximum

production of CyA. The optimal combination of me-

dium components obtained from the study included

(in g/L) glycerol 98.8, casein peptone 28.5, malt extract

20, peptone 10 and α-amino butyric acid 5.7. The opti-

mal composition was verified experimentally and com-

pared with the data calculated from the model. The

experimentally obtained CyA yield was 452.32 mg/L,

whereas the predicted value from the polynomial mod-

el was 455.05 mg/L, thereby confirming the high accu-

racy of the model under the investigated conditions.

Effect of amino acids

CyA consists of 11 amino acids in its structure and

its production is affected by the addition of exogenous

amino acids which are members of the CyA ring (Bal-

akrishnan and Pandey, 1996; Lee and Agathos, 1989;

Survase et al., 2009b). Hoppert et al. (2001) reported

that the supplemented amino acids modify the en-

dogenous amino acid pool of the fungus and directed

the biosynthesis of CyA as precursors. The effect of ex-

ogenous amino acid was studied in both submerged

as well as solid state fermentation (Balakrishnan and

Pandey, 1996; Lee and Agathos, 1989; Survase et al.,

2009b), but the effect of amino acids in submerged

fermentation under static conditions has not yet been

reported.

In the present study, α-amino butyric acid from the

basal medium was replaced with constituent amino

acids (Fig. 4) and their effect on CyA production and

biomass production was observed. Of all the amino

acids tested, L-valine produced the maximum CyA of

595 mg/L followed by L-leucine 556 mg/L. DL-valine,

on the other hand, did not increase the product titer

as that of L-valine. Addition of L-methionine in the pro-

duction medium reduced the CyA production. Zocher

et al. (1984) reported that methionine could not take

part in the biosynthesis, as methylated amino acids in-

Figure 4. Effect of different amino acids on production of CyA using T. inflatum MTCC 557 (DCW is dry cell weight and CyA is cyclosporin A)


terfere with the biosynthesis of CyA in vivo.

When added together, L-valine and L-leucine in-

creased drug production dramatically (897 mg/L) (Fig.

5). These two amino acids seem to act independently

and their mode of action is different. This effect was not

observed with other amino acid combinations. Supple-

mentation of L-methionine with L-valine resulted in loss

of the stimulatory effect of L-valine. Similar results were

encountered by Balakrishnan and Pandey (1996), Lee

and Agathos (1989) and Nisha et al. (2008) in CyA bio-

synthesis.

The optimal amount and time of addition of L-valine

was also investigated. It was observed that CyA pro-

duction increased with an increase in concentration of

L-valine up to 9 g/L, Maximum CyA production of 884

mg/L was observed at initial L-valine concentration of

9 g/L (data not shown). Balakrishnan and Pandey (1996)

and Lee and Agathos (1989) reported that CyA pro-

duction reached the saturation level at L-valine 4 g/L

whereas, Survase et al. (2009b) reported the saturation

level to be 6 g/L after which there was no further in-

crease in CyA production. The optimum time for L-va-

line (9 g/L) and L-leucine (6 g/L) addition for maximum

product titer was found to be 4 days (Data not shown).

When added after 4 days, CyA production of 1241

mg/L was obtained. Balakrishnan and Pandey (1996)

and Lee and Agathos (1989) reported that addition of

L-valine in the fermentation medium after 18 h and 20

h respectively, gave the maximum production of CyA

under shake flask conditions.

CONCLUSION

Static fermentation could successfully be used for

production of CyA. The exogenous supply of amino

acids along with time of addition played an impor-

tant role in maximizing the production of CyA.

ACKNOWLEDGEMENT

We are thankful to Department of Biotechnology,

Government of India for funding this project. The

gift of CyA standard from RPG Life Sciences Ltd,

Mumbai is gratefully acknowledged.

Figure 5. Effect of combination of amino acids on production of CyA using T. inflatum MTCC 557 (DCW is dry cell weight and CyA is cyclosporin A)


REFERENCES

Abdel-fattah, Y. R., H. E. Enshasy, M. Anwar, H. Omar,

E. Abolmagd, and R. A. Zahra. 2007. Application of

factorial experimental designs for optimization of

cyclosporin A production by Tolypocladium infla-

tum in submerged culture. J. Microbiol. Biotech-

nol. 17:1930–1936.

Agathos, S. N., and R. Parekh. 1990. Enhancement of

cyclosporin production in Tolyclapodium inflatum

strain after epichorhydrin treatment. J. Biotechnol.

13:73-81.

Agathos, S. N., J. W. Marshall, C. Moraiti, R. Parekh,

and C. Madhosingh. 1986. Physiological and ge-

netic factors for process development of cyclospo-

rin fermentations. J. Ind. Microbiol. 1:39-48.

Balakrishnan, K., and A. Pandey. 1996. Influence of

amino acids on the biosynthesis of cyclosporin A

by Tolypocladium inflatum. Appl. Microbiol. Bio-

technol. 45:800-803.

Balaraman, K., and N. Mathew. 2006. Optimization

of media composition for the production of cyclo-

sporin A by Tolypocladium species. Indian J. Med.

Res. 123:525-530.

Billich, A., and R. Zocher. 1987. Enzymatic synthesis

of cyclosporin-A. J. Biol. Chem. 262:17258–17259.

Bussari, B., P. S. Saudagar, N. S. Shaligram, S. A.

Survase, and R. S. Singhal. 2008. Production of

cephamycin C by Streptomyces clavuligerus NT4

using solid-state fermentation. J. Ind. Microbiol.

Biotechnol. 35:49–58.

Chauhan, A. K., S. A. Survase, J. K. Kumar, and U. S.

Annapure. 2009. Medium optimization by orthog-

onal array and response surface methodology for

cholesterol oxidase production by Streptomyces

lavendulae NCIM 2499. J. Gen. Appl. Microbiol.

55:171-180.

Fan, L., A. T. Soccol, A. Pandey, and C. R. Soccol.

2007. Effect of nutritional and environmental con-

ditions on the production of exo-polysaccharide of

Agaricus brasiliensis by submerged fermentation

and its antitumor activity. LWT - Food Sci. Technol.

40:30-35.

Hoppert M., C. Gentzch, and K. Schorgendrofor.

2001. Structure and localization of cyclosporine

synthetase. The key enzyme of cyclosporine bio-

synthesis in Tolypocladium inflatum. Arch. Micro-

biol. 176:285-293.

Isaac, C. C., A. Jones, and M. A. Pickard. 1990.

Production of cyclosporine A by Tolypocladium

niveum strains. Antimicrob. Agents Chemother.

34:121-127.

Kahan, B. D. 1984. Cyclosporin: Biological activity

and clinical applications. Crune & Straton Inc., Or-

lando, FL. 895 p.

Lee, J., and S. Agathos. 1989. Effect of amino acids

on the production of cyclosporin A by T. inflatum.

Biotechnol. Letts. 2:77–82.

Majumder A., and A. Goyal. 2008. Optimization of

culture conditions of a novel glucan producing

glucansucrase from Leuconostoc dextranicum

NRRL B-1146. Curr. Trends in Biotechnol. Pharm.

2:260-268.

Nakajima, H., T. Hamasaki, K. Nishimura, Y. Kimura,

S. Udagawa, S. Sato, and E. F. Smith. 1988. Isola-

tion of 2-acetylamino-3-hydroxy-4-methyl-oct-

6-enoic acid, a derivative of the ‘C9 amino acid’

residue of cyclosporins, produced by the fungus

Neocosmospora vasinfecta. Agric. Biol. Chem.

52:1621–1623.

Nisha, A. K., S. Meignanalakshmi, and K. Ramasamy.

2008. Comparative effect of amino acids in the pro-

duction of cyclosporin A by solid and submerged

fermentations. Biotechnol. 7: 205-208.

Sallam, L. A. R., A. H. El-Refai, A. A. Hamdi, A. H. El-

Minofi, and S. I. Abd-Elsalam. 2003. Role of some

fermentation parameters on cyclosporin A produc-

tion by a new isolate of A. terreus. J. Gen. Appl.

Microbiol. 49:321–328.

Sawai, K., T. Okuno, Y. Tereda, Y. Harada. K.Wawamura,

H. Sasaki, and S. Takao. 1981. Isolation and prop-

erties of two antifungal substances from Fusarium

solani. Agric. Biol. Chem. 45:1223–1228.

Schindler, R. 1985. Cyclosporin in autoimmune dis-

eases. Springer-Verlag, Berlin. 373 p.

Shih, I.-L., K. L. Tsai, and C. Y. Hsieh. 2007. Effects

of culture conditions on the mycelial growth and

bioactive metabolite production in submerged

culture of Cordyceps militaris. Biochem. Eng. J. 33:

193–201.


Survase, S. A., N. S. Shaligram, R. C. Pansuriya, U. S.

Annapure, and R. S. Singhal. 2009a. A novel me-

dium for the enhanced production of cyclospo-

rin A by Tolypocladium inflatum MTCC 557 using

solid state fermentation. J. Microbiol. Biotechnol.

19:462-467.

Survase, S. A., U. S. Annapure, and R. S. Singhal.

2009b. Statistical optimization of cyclosporin: A

production on a semi-synthetic medium using

Tolypocladium inflatum MTCC 557, Global J. Bio-

technol. Biochem. 4:184-192

Zocher, R., N. Madry, H. Peeters, and H. Kleinkauf.

1984. Biosynthesis of cyclosporin A. Phytochemis-

try 23:549-551.




ABSTRACT

The aim of this study was to investigate the survival characteristics of verocytotoxigenic Escherichia coli

(VTEC) and Salmonella in soil, slurry, farm water and bovine feces. Samples of each of the aforementioned

media were inoculated with separate cocktails of VTEC and Salmonella and stored at 4 and 14 ºC, repre-

senting average Winter and Summer temperatures respectively. Samples were withdrawn periodically and

surviving cells enumerated by directly plating onto selective media. Decimal reduction times (D-values)

were calculated from the inverse of the slope obtained by linear regression of a plot of time versus the log

of surviving cells. In the latter stages of the experiment, presence or absence was determined by enrich-

ment and selective plating. VTEC and Salmonella D-values ranged from 3.59 to 23.84 days. Temperature

significantly affected VTEC survival in water (P<0.05) and Salmonella survival in bovine feces (P<0.01) but

not in any of the other farm media tested. In general there was no significant difference (P>0.001) between

VTEC and Salmonella survival in a given medium under similar storage temperatures. However, Salmonella

D-values were significantly higher in slurry (4°C) and bovine feces (4°C and 14°C).

This study provides critical comparative data on VTEC and Salmonella death rates in a range of environ-

ments commonly encountered on farms to support the development of quantitative microbial risk assess-

ment (QMRA) and provide the scientific basis for an effective good agricultural practice (GAP) food safety

program.

Keywords: Salmonella Typhimurium, Salmonella Dublin, Verocytotoxigenic E. coli, persistence, slurry, soil, water, feces

Correspondence: D. J. Bolton, [email protected]: +353 (1) 805 9539 Fax: +353 (1) 805 9550

BRIEF COMMUNICATION Comparative Studies on the Survival of Verocytotoxigenic Escherichia coli and

Salmonella in Different Farm Environments

C. J. O’Neill1, 2, D. J. Bolton1 and S. Fanning2

1Teagasc-Ashtown Food Research Centre, Ashtown, Dublin 15, Ireland2 Centre for Food Safety, School of Public Health, Physiotherapy and Population Science, University College Dublin,

Belfield, Dublin 4, Ireland



INTRODUCTION

Cattle serve as a source of both Salmonella and

verocytotoxigenic E. coli especially E. coli O157:H7,

which are shed in the feces without any clinical symp-

toms (Losinger et al. 1995; Alice, 1997; Rasmussen et

al. 1993). Few studies have addressed the survival of

E. coli O157:H7 in soil, feces, slurry and water and

none has integrated survival in the different media

into the same study nor have provided comparative

data for Salmonella spp. Furthermore, data on Sal-

monella survival rates in the farming environment is

absent.

Under current EC regulations all food proces-

sors have a legal responsibility to provide safe food.

Hazard analysis and critical control point (HACCP)

is generally regarded to be the most effective ap-

proach to food safety, but is not a legal requirement

during primary production in Europe. Effective food

safety is therefore dependent on good agricultural

practice (GAP) (Horchner et al., 2006). Risk assess-

ment, which provides a quantifiable metric on the

effectiveness or otherwise of specific farming activi-

ties, is a prerequisite to an effective science based

GAP covering such issues as farm waste manage-

ment and biosecurity.

However, there is limited information available

for on-farm risk assessments (McGee et al., 2002;

Semenov et al., 2008; Williams et al., 2008,) and in-

sufficient data to fully characterize verocytotoxigenic

Escherichia coli (VTEC) and Salmonella enterica sur-

vival in the farming environment. Specific data on the

persistence of these two pathogens in slurry, water,

feces and in soil is therefore required. Mismanage-

ment of animal waste, in particular, will lead to direct

or indirect infection of humans (Vernozy-Rozand et

al., 2002; Guan and Holley, 2003). Survival data for

VTEC and Salmonella in the main environments en-

countered on the farm will aid in the development of

proper science-based management practices which

will in turn decrease farm to fork transmission. The

objective of this study was therefore to provide com-

parative data on VTEC and Salmonella survival in

slurry, soil, water and bovine feces.


Organisms

A VTEC cocktail and a Salmonella cocktail com-

posed of isolates of farm origin were chosen for this

study. The former included bovine and ovine E. coli

O157:H7 strains and an untypable canine strain. The

canine strain initially serotyped as O157:H7 (hence

its addition in this study) but was later discovered to

be untypable. The Salmonella organisms included

bovine Salmonella ser. Dublin and Salmonella Ty-

phimurium DT104 strains, along with a Salmonella

Typhimurium DT193 isolated from farm trough wa-

ter. All were obtained from the culture collection at

Teagasc Food Research Centre (Ashtown), Dublin 15,

Ireland.

Collection of samples

Composite samples of farm water, slurry, cattle

feces and soil were collected from a local farm in

County Meath, Ireland. Approximately 10 kg each of

slurry, feces and soil, in addition to 10 l of farmyard

water were collected and transferred to sterile hold-

ing containers. Samples were transported in a cool-

box at approximately 2°C to the laboratory.

Preparation of inoculum

A cryogenic bead of each isolate of Salmonella

and VTEC were aseptically transferred to 30 ml Brain

Heart Infusion broth (BHI, Oxoid, Basingstoke, UK)

and incubated for 18 h at 37ºC. One ml was then trans-

ferred to a 30 ml fresh BHI and incubated at 37ºC for

18 h to obtain a stationary phase culture. The cultures

were then centrifuged at 3000 g (Eppendorf, Davidson

and Hardy, Ireland). Cells were washed three times in

sterile Maximum Recovery Diluent (MRD, Oxoid). The

resulting pellets from the three VTEC isolates were

then mixed in 100 ml MRD to form a VTEC cocktail.

The same methods were used to prepare the Salmo-

nella cocktail. A serial dilution was performed on both

cocktails, followed by plating onto Plate Count Agar

(Oxoid) to determine the initial inoculum.


Inoculation of soil, fecal and slurry samples with VTEC strains

Soil, fecal and slurry samples (400 g) were inocu-

lated with 1 ml of approximately log10 8.0 cfu ml-1 of

the VTEC cocktail to yield a final concentration of

approximately log10 6.0 colony forming units (cfu)

g-1. Samples were mixed and stored in 500 g sterile

plastic containers, in triplicate, at both 4ºC and 14ºC

representing mean Winter and Summer tempera-

tures, respectively. Uninoculated slurry, fecal and soil

samples were used as controls. Sampling took place

on days 0, 3, 9, 13, 21, 34, 42, 64, 76, 88 and 102. Sur-

viving cells were enumerated as described below.

Inoculation of Water Samples with VTEC strains

Water samples (400 ml) were inoculated with 1 ml

of the VTEC cocktail to yield a final concentration of

approximately log10 6.0 cfu ml-1. Storage and sam-

pling was as previously described.

Inoculation of soil, slurry and fecal sam-ples with Salmonella species

Soil, slurry and fecal samples (400 g) were inoculat-

ed with log10 8.0 cfu ml-1 of the Salmonella cocktail to

yield a final concentration of approximately log10 6.0

cfu g-1. Storage was as per the VTEC above. Sampling

took on days 0, 3, 7, 12, 19, 32, 40, 60, 69, 81 and 102.

Inoculation of water samples with Sal-monella species

Water samples (400 ml) were inoculated with the

Salmonella cocktail to yield a final concentration of

approximately log10 6.0 cfu ml-1. Uninoculated wa-

ter samples were used as controls. Samples were

stored in 500 ml sterile Duran bottles. Storage and

sampling was as previously described.

Enumeration of VTEC

The inoculated water, slurry, soil and feces were

first mixed using a sterile spatula. Ten g/ ml samples

were then obtained to which 90 ml MRD was added

before stomaching for 90 seconds. Following serial

dilution, samples were plated directly onto CT-SMAC

(Oxoid) in 100 μl volumes in duplicate. Plates were

then incubated at 37ºC for 24 h. Enrichment cultures

were used to detect the presence of low numbers of

bacteria in the latter stages of the experiment. Ten

g/ ml of samples were enriched 1:10 in buffered pep-

tone water (BPW) (Oxoid) and incubated at 37ºC for

24 h. Samples were then plated onto CT-SMAC in

duplicate and incubated as before. The enrichment

procedure was carried out until the organism was

not detected for two consecutive sampling days.

Enumeration of Salmonella species

Samples were diluted and stomached as previ-

ously described. Following serial dilution, samples

were plated in 100 μl volumes onto XLD Agar (Ox-

oid) in duplicate. Plates were then incubated at 37ºC

for 24 h. Enrichment cultures were used to detect

the presence of low numbers of bacteria in the latter

stages of the experiment. Ten g/ ml of samples were

enriched 1:10 in BPW and incubated at 37ºC for 24

h. Samples were then plated onto XLD in duplicate

and incubated as before. The enrichment procedure

was carried out until the organism was not detected

for two consecutive sampling days.

Statistical analysis of the results

Each experiment was performed in triplicate. The

rate of decline over time was obtained by plotting

surviving cells against time. The line of best fit for

each set of points was found using linear regression

analysis (Genstat 5, Statistics Department, Rothamst-

ed Experimental Station, Hertfordshire, UK). The dec-

imal reduction values reported were calculated using

the average slope (D = -1/slope). The decimal reduc-

tion values were then statistically analyzed using the

t-test (Genstat 5, Statistics Department, Rothamsted

Experimental Station, Hertfordshire, UK). Differences

are reported at the 5%. 1% and 0.1% levels.


RESULTS

Neither Salmonella nor VTEC were detected

in the uninoculated control samples at any stage

throughout the experiment. At 4°C there was no sig-

nificant difference (P> 0.001) in the VTEC D-values

which ranged from 3.9 days (slurry) to 8.79 days

(soil) (Table 1). At 14°C the rate of decline in water

(D14-value: 3.59 days) was similar to that in slurry

(D14-value: 6.5 days) but significantly (P<0.001)

faster to that observed in soil (D14-value: 8.16 days)

and feces (D14-value: 9.79 days). A comparison of

the effect of storage temperature found that the

differences observed between the D-values ob-

tained at 4 and 14°C were only significant (P<0.05)

in water.

At 4°C there was no significant difference (P>

0.001) in the Salmonella D-values in water (4.89 days)

and soil (6.78 days) but the rate of decline in water

was significantly (P<0.001) faster than in feces (D4-

Table 1. VTEC and Salmonella D-values in soil, feces, slurry and water

Bac

teria

Sto

rag

e te

mp

erat

ure

(°C

)

Med

ium

D-v

alue

1(d

ays)

S.E

.

4°C

2 ver

sus

14°C

Bac

teria

Sto

rag

e te

mp

erat

ure

(°C

)

Med

ium

D-v

alue

1(d

ays)

S.E

.

4°C

ver

sus

14°C

VTE

C3 v

ersu

s Sa

lmo

nella

VTEC 4 Soil 8.79 a 0.0343 NSSalmo-nella

4 Soil 6.78ab 0.0236 NS NS

VTEC 4 Feces 7.56 a 0.0169 NSSalmo-nella

4 Feces 15.32 c 0.0155 P<0.001 P<0.05

VTEC 4 Slurry 3.90 a 0.070 NSSalmo-nella

4 Slurry 11.97bc 0.0215 NS P<0.05

VTEC 4 Water 8.63 a 0.0383 P<0.05Salmo-nella

4 Water 4.89 a 0.0189 NS NS

VTEC 14 Soil 8.16 b 0.0132 NSSalmo-nella

14 Soil 9.43 b 0.0079 NS NS

VTEC 14 Feces 9.79 b 0.0288 NSSalmo-nella

14 Feces 23.84 c 0.0090 P<0.001 P<0.05

VTEC 14 Slurry 6.50 ab 0.0311 NSSalmo-nella

14 Slurry 10.51 b 0.0094 NS NS

VTEC 14 Water 3.59 a 0.0564 P<0.05Salmo-nella

14 Water 5.25 a 0.0302 NS NS

1 D-values with the same superscript letter are not significantly different (P>0.001). Comparisons were made within a given storage condition (4°C or 14°C)

2NS=non significant; for significant differences the degree of significance is indicated. Comparisons were between storage conditions (4°C and 14°C)

3NS=non significant; for significant differences the degree of significance is indicated. Comparisons were made within a given storage condition (4°C or 14°C)

S.E= Standard error


value: 15.32 days) or slurry (D4-value: 11.97 days),

which were statistically similar (Table 1). At 14°C

the D-values in soil (9.43 days) and slurry (10.51

days) were statistically the same but significantly

(P<0.001) slower than in water (D14-value 5.25 days)

and significantly faster (P<0.001) than in feces (D14-

value: 23.84 days)

A comparison of the effect of storage tempera-

ture found that the differences observed between

the D-values obtained at 4 and 14°C were only sig-

nificant (P<0.001) in feces.

In general, the D-values obtained for VTEC and

Salmonella were similar with significant differences

only observed between the two bacterial cocktails

in feces at 4°C (P<0.05), slurry at 4°C (P<0.05) and

feces at 14°C (P<0.05). In each case the Salmonella

demonstrated better survival.

DISCUSSION

The D-values for VTEC in feces at 4 and 14°C were

7.56 and 9.79 days, respectively. The former com-

pares with the D4-value of 9.04 days reported by Hi-

mathongkham et al. (1999) for E. coli O157:H7 in the

upper layers of manure piles. It has been reported

that colonised cattle may shed up to 108 VTEC (Bess-

er et al., 2001; Fukushima and Seki, 2004) cells per

gram of feces, suggesting prolonged persistence

of these bacteria in feces for up to 60 days (Winter)

and 78 days (Summer) based on this study. Feces de-

posited onto grassland is washed by rainfall into the

soil (Bolton et al., 1999). During the warmer months,

when animals are grazing outdoors, cross-contami-

nation of soil is therefore inevitable.

The VTEC D-values in soil at 4°C (8.79 days) and

14°C (8.16 days) are similar to those previously re-

ported (Bolton et al., 1999, Besser et al., 2001; Islam

et al,. 2004; Fremaux et al., 2008) and provide further

evidence that once contaminated, soil will remain a

source of these pathogenic organisms for extended

periods, contaminating plants (Islam et al., 2004; Is-

lam et al., 2005), livestock (McGee et al., 2001) and

posing a risk for recreational use (Moore et al., 1993;

Chapman et al., 1997; Jackson et al., 1998).

During the Winter months, VTEC are more likely

to end up in the slurry tank. The D4-value and D14-

value for VTEC in slurry of 3.9 days and 6.5 days were

considerably lower than the 38.76 days reported by

Himathongkham et al. (1999) for E. coli O157:H7 in

slurry stored at 4°C and the 15.27 days reported by

McGee et al. (2001) for the same organism in the

same medium at 10 °C. This may be due to differ-

ences in slurry composition, in addition to bacterial

strain variation.

At 14°C, survival of VTEC in water was significantly

reduced. This was probably due to increased micro-

biological competition at the higher temperature

(Avery et al,. 2008). The D4 and D14-values of 8.63

and 3.59 days demonstrates a longevity that cor-

roborates the findings of other similar studies on

the survival of these organisms in aquatic environ-

ments (Wang and Doyle, 1998; McGee et al., 2002).

Farm water contaminated with this organism would

remain a potential reservoir for VTEC, facilitating dis-

semination and livestock re-infection (Bolton et al.,

1999; McGee et al., 2002).

The Salmonella D4 and D14-values in soil, slurry,

feces and water ranged from 4.89 to 23.84 days.

In general, the survival rates for the VTEC and Sal-

monella cocktails were statistically similar with the

exception of feces and slurry at 4 °C and feces at

14 °C, where the Salmonella survival rates were sig-

nificantly higher. These mixed findings are compa-

rable to those of Himathongkham et al., (1999) who

reported similar D-values for E. coli O157:H7 and

Salmonella Typhimurium, with the predominant or-

ganism varying depending on the temperature and

location within the manure pile. Furthermore, in a

study of the survival of enteric bacteria in inoculated

feces spread on grass pasture land, Hutchinson et

al,. (2005) also reported that there was no significant

difference between the reduction times for the dif-

ferent bacterial pathogens including E. coli O157:H7

and Salmonella spp.

CONCLUSIONS

In conclusion, the D-values provided in this study

may be used in risk assessments to estimate survival

in soil, feces, slurry and water under different envi-


ronmental conditions and to predict the duration

of slurry storage prior to grazing land application

or duration of the soil resting period prior to crop

sowing. If the target is a minimal 105-fold reduction

of VTEC or Salmonella spp., then slurry should be

stored at 60 days at 4°C or 53 days at 14°C before

spreading. Fields that may be contaminated with

these organisms, either as a result of grazing or fecal

waste spreading should be left for a minimum of 44

days during Winter or 47 days during Summer be-

fore crops are grown.

ACKNOWLEDGEMENT

This research was funded by the US-Ireland Co-

operative Program in Agricultural Science and Tech-

nology administered by the Department of Agricul-

ture, Fisheries and Food, Ireland.

REFERENCES

Alice, N. P. 1997. Public and microbes: public and

animal health problem. J. Dairy Sci. 80:2673-2681.

Avery, L. M., A. P. Williams, K. Killham, and D. L.

Jones. 2008. Survival of Escherichia coli O157:H7

in waters from lakes, rivers, puddles and animal-

drinking troughs. Sci. Total Environ. 389:378-385.

Besser, T. E., B. L. Richards, D. H. Rice, and D. D.

Hancock. 2001. Escherichia coli O157:H7 infection

of calves: infectious dose and direct contact trans-

mission. Epidemiol. Infect. 127: 555-560.

Bolton D. J., C. M. Byrne, J. J. Sheridan, D. A. Mc-

Dowell and I. S. Blair. 1999. The survival character-

istics of a non-toxigenic strain of Escherichia coli

O157:H7. J. Appl. Microbiol. 86:407-11.

Chapman, P. A., C. A. Siddons, J. Manning, and C.

Cheetham. 1997. An outbreak of infection due to

verocytotoxin-producing Escherichia coli O157 in

four families: the influence of laboratory methods

on the outcome of the investigation. Epidemiol.

Infect. 119:245-250.

Fukushima, H. and R. Seki. 2004. High numbers of

Shiga-toxin producing Escherichia coli found in

bovine faeces collected at slaughter in Japan.

FEMS Microbial. Lett. 238:189-197.

Fremaux, B., C. Prigent-Combaret, and C. Vernozy-

Rozand. 2008. Long-term survival of Shiga toxin-

producing Escherichia coli in cattle effluents and

environment: An updated review. Vet. Microbiol.

132:1-18.

Guan, T. Y. and R. A. Holley. 2003. Pathogen survival

in swine manure environments and transmission of

human enteric illness- A review. J. Environ. Qual.

32:383-392.

Himathongkham, S., S. Bahari, H. Riemann, and D.

Cliver. 1999. Survival of Escherichia coli O157:H7

and Salmonella Typhimurium in cow manure and

cow manure slurry. FEMS Microbial. Lett. 178:251-

257.

Horchner, P. M., D. Brett, B. Gormley, I. Jenson, and

A. M. Pointon. 2006. HACCP- based approach to

the derivation of an on-farm food safety program

for the Australian red meat industry. Food Control

17:497-510.

Hutchison, M. L., L. D. Walters, T. Moore, D. J. Thom-

as, and S. M. Avery. 2005. Fate of pathogens pres-

ent in livestock wastes spread onto farm plots.

Appl. Environ. Microbiol. 71:691-696.

Islam, M., M. P. Doyle, S. C. Phatak, P. Millner, and

X. Jiang. 2004. Persistence of enterohemorrhagic

Escherichia coli O157:H7 in soil and on leaf let-

tuce and parsley grown in fields treated with con-

taminated manure composts or irrigation water. J.

Food Prot. 67:1365-1370.

Islam, M., M. P. Doyle, S. C. Phatak, P. Millner, and X.

Jiang. 2005. Survival of Escherichia coli O157:H7

in soil and on carrots and onions grown in fields

treated with contaminated manure composts or ir-

rigation water. Food Microbiol. 22:63-70.

Jackson, S. G., R. B. Goodbrand, R. P. Johnson, V.

G. Odorico, D. Alves, K. Rahn, J. B. Wilson, M.

K. Welch, and R. Khakhria. 1998. Escherichia coli

O157:H7 diarrhoea associated with well water and

infected cattle on an Ontario farm. Epidemiol. In-

fect. 120:17-20.

Losinger W. C., S. J. Wells, L. P. Garber, H. S. Hurd,

and L. A. Thomas. 1995. Management factors re-

lated to Salmonella shedding by dairy heifers. J.

Dairy Sci. 78:2464-72.

McGee, P., D. J. Bolton, J. J. Sheridan, B. Earley and


N. Leonard. 2001. The survival of Escherichia coli

O157:H7 in slurry from cattle fed different diets.

Lett. Appl. Microbiol. 32:152-155.

McGee, P., D. J. Bolton, J. J. Sheridan, B. Earley, G.

Kelly and N. Leonard. 2002. Survival of Escherichia

coli O157:H7 in farm water: its role as a vector in

the transmission of the organism within herds. J.


Moore, A. C., B. L. Herwaldt, G. F. Craun, R. L. Calde-

ron, A. K. Highsmith and D. D. Jaunek. 1993. Sur-

veillance for waterborne disease outbreaks in the

United States 1991-1992. M M W R. 42:1-22.

Rasmussen, M. A., W. C. J. Cray, T. A. Casey and S. C.

Whipp. 1993. Rumen contents as a reservoir of En-

terohemorrhagic Escherichia coli. FEMS Microbiol.

Lett. 114:79-84.

Semenov, A.V., E. Franz, E., L. van Overbeek, A. J.

Termorshuizen and A. H. C. van Bruggen. 2008. Es-

timating the stability of Escherichia coli O157:H7

survival in manure-amended soils with differ-

ent management histories. Environ. Microbiol.

10:1450-1459.

Vernozy-Rozand, C., M. P. Montet, F. Lequerrec, E.

Serillon, B. Tilly, C. Bavai, S. Ray-Gueniot, J. Bou-

vet, C. Mazuy-Cruchauet and Y. Richard. 2002.

Prevalence of verotoxin-producing Escherichia coli

in slurry, farmyard manure and sewage sludge in

France. J Appl. Microbiol. 93:473-478.

Wang, G. and M. P. Doyle. 1998. Survival of entero-

hemorrhagic Escherichia coli O157:H7 in water. J.

Food. Prot. 61: 662–667.

Williams, A. P., K. A. McGregor, K. Killham and D. L.

Jones. 2008. Persistence and metabolic activity

of Escherichia coli O157:H7 in farm animal faeces.

FEMS Microbiol. Lett. 287:168.




ABSTRACT

We report the development of a microarray-based transcriptomic expression analysis method for the

detection of 131 antimicrobial resistance genes in multidrug-resistant clinical (Enterococcus faecium, E.

faecalis), aquaculture (Aeromonas veronii), poultry (Campylobacter jejuni, Staphylococcus aureus), and out-

break (Salmonella enterica serovar Typhimurium) strains. Unmodified oligonucleotide probes and 131 pair

of primers for the genes conferring resistance to 22 different antimicrobials were used for transcriptomic

array and PCR analysis. Detection of resistance genes by transcriptomic array and PCR methods correlated

well with the susceptibility profiles of the isolates used in the study. However, some of the genes conferring

resistance to ampicillin (amp), bleomycin (ble), lincomycin (linAn2, lmr, lmrA, lmrB, mgt), neomycin (neo,

nptII, himaR), oleandomycin (oleB, oleC-orf4, oleC-orf5), penicillin (mecA), and rifampin (arr2) could not be

detected by above methods. Co- or cross-resistance to these drugs and their extracellular transport due

to the presence of 2 to 5 efflux pump genes (oleB, marA, tetA, tetB, vraD) is believed to confer resistance

to the above antimicrobials. Moreover, the presence of resistance genes for bacitracin, chloramphenicol,

erythromycin, kanamycin, streptomycin, and tetracycline in phenotypically sensitive isolates indicated ei-

ther inactive versions of these genes or modulation of gene expression. Overall, the transcriptomic ar-

ray method provided a valuable insight into the mechanism(s) of resistance, status of gene expression at

transcription level, and detection of all the antimicrobial resistance genes among bacteria from different

ecological sources.

Keywords: Antimicrobial, resistance genes, microarray, hybridization, PCR, disk diffusion, oligonucleotide

Correspondence: S. A. Khan, [email protected]: +1 -870 543-7197 Fax: +1-870 543-7307

A Transcriptomic Expression Array, PCR and Disk Diffusion Analysis of Antimicrobial Resistance Genes in Multidrug-Resistant Bacteria

S. A. Khan1, K. Sung1, and M. Nawaz1

1U.S. Food and Drug Administration, National Center for Toxicological Research, Division of Microbiology, 3900 NCTR Road, Jefferson, AR 72079, USA.



INTRODUCTION

The most common techniques used to detect sus-

ceptibility of bacteria to different antimicrobials are

the disk diffusion (Clinical and Laboratory Standards

Institute, 2009a) and broth dilution assays (Clinical

and Laboratory Standards Institute, 2009b). While

these assays show susceptibility of the organisms to

different antimicrobial agents and their level of re-

sistance in most cases, they cannot tell which of the

multiple resistance genes for a given antimicrobial

are present. Polymerase chain reaction and other

genetic analysis tools that are used to detect them

are time-consuming, expensive, and labor-intensive

for detecting hundreds of genes in a single experi-

ment (Khan et al., 2000). The success of PCR tech-

nology, among various other factors, depends upon

published genomic DNA sequences. Any variation

of the PCR primer sequence, either due to a discrep-

ancy in the published DNA sequences or because of

point mutations, insertion and deletion events could

lead to false or unsuccessful amplifications requir-

ing other methods of verification. Lately, microarray

technology has been employed to identify and type

food-borne pathogens (Al-Khaldi et al., 2002; Gara-

izer et al., 2006; Majtan et al., 2007; Rasooly et al.,

2008). It has also been used successfully to detect

the presence of multiple antimicrobial resistance

genes among multidrug-resistant bacteria (Antwer-

pen et al., 2007; Brazas and Hancock, 2005; Call et

al., 2003; Cassone et al., 2008; Chen et al., 2005; Frye

et al., 2006; Perreten et al., 2005; Van Hoek et al.,

2005; Zhu et al., 2007) with enhanced sensitivity and

specificity. The cost of detection still remains prohib-

itive for routine monitoring of resistance genes in a

high-throughput manner. One of the factors that can

substantially increase the cost of custom arrays is the

use of long and modified oligonucleotide probes (Ly-

ons, 2003). Modification of oligonucleotide probes

requires the addition of amino, thiol, and biotin func-

tional groups to improve their binding efficiency on

the chip surface and has been shown to be unneces-

sary (Schüler et al., 2009).

The microarray-based methods described earlier

used labeled genomic DNA for hybridization with

the probes. Detection of the genes with labeled ge-

nomic DNA can pinpoint the presence of resistance

genes but it cannot tell whether the gene is tran-

scribed or expressed without using other genetic

testing methods (Cockerill III, 1999). The transcrip-

tomic array described in this study uses 40 nucleo-

tide long unmodified oligonucleotide probes that

were printed on epoxy and polylysine-coated glass

slides and used to detect transcriptional expression

of antimicrobial resistance genes from multidrug-re-

sistant bacteria. A comparison of the transcriptomic

array, PCR and disk diffusion data exhibited a very

good correlation and offers a powerful combination

for comprehensive analysis of antimicrobial resis-

tance genes and their transcriptomic expression in

multidrug-resistant bacteria. Our results also indicat-

ed that expensive modification of oligonucleotide

probes was not necessary for the detection of anti-

microbial resistance genes when printed on epoxy-

coated glass slides.


Bacterial strains used in the study

An in-house collection of multidrug-resistant bac-

teria from different ecological sources, including

Enterococcus faecium ATCC51559 and E. faecalis

ATCC 51299 (human clinical isolates), Aeromonas

veronii isolates Avt101 and Aet2002 (catfish isolates),

Campylobacter jejuni isolates TH200 and TH205,

Staphylococcus aureus isolates P20 and P34 (poultry

isolates) and S. enterica serovar Typhimurium phage

types DT16 and DT21 (outbreak isolates) were used

for the detection of antimicrobial resistance genes in

them. The number of genes targeted for detection

by microarray and PCR was 131. Most were related

to antibiotics used either in clinical practice or the

National Antibiotic Resistance Monitoring System

(NARMS) panel. Although bacteria were resistant to

multiple antimicrobials, they were grown in the pres-

ence of antibiotics used for their initial selection. For

example, E. faecium and E. faecalis were grown in

the presence of vancomycin (30 μg ml-1), A. veronii in

the presence of tetracycline (30 μg ml-1), C. jejuni and


S. enterica serovar Typhimurium phage types DT16

and DT21 in the presence of ampicillin (10 μg ml-1),

and S. aureus in the presence of erythromycin (15 μg

ml-1) in Luria Bertani (LB) broth.

Microarray supplies

Cy3 and/or Cy5-labeled antisense (complemen-

tary to the probes) oligonucleotides (40-mer), 2 x

hybridization buffer, epoxy- and polylysine-coated

slides were purchased from Eurofins MWG Operon

Biotech (Huntsville, AL, USA). An RT-PCR kit was

purchased from Invitrogen (Carlsbad, CA, USA) and

CyScribe first-strand cDNA labeling kit was pur-

chased from Amersham Biosciences (Piscataway, NJ,

USA). RNA isolation kit was purchased from Qiagen

(Valencia, CA, USA). For the printing of oligonucle-

otide probes on the slides, a microarray core facil-

ity at the University of Arkansas for Medical Sciences

was utilized. The slide scanner was purchased from

PerkinElmer (Boston, MA, USA). Hybridization cham-

bers were from Corning Life Sciences (Acton, MA,

USA).

Antimicrobial susceptibility testing by disk diffusion

All the isolates were tested for antimicrobial sus-

ceptibility (Clinical and Laboratory Standards In-

stitute, 2009a, b) by the criteria of the Clinical and

Laboratory Standards Institute (CLSI). Twenty two

different antibiotics, with their concentrations noted

after their names in μg ml-1, and three salts (tylosin,

tellurium, and benzalkonium chloride) that are used

in postharvest processing were tested to determine

the sensitivity of bacteria to these compounds: ampi-

cillin-10, apramycin-15, bacitracin-10, benzalkonium

chloride-100, bleomycin-512, chloramphenicol-30,

erythromycin-15, gentamicin-120, hygromycin-512,

kanamycin-30, lincomycin-2, methicillin-5, neomy-

cin-30, oleandomycin-512, penicillin-10, rifampin-5,

spectinomycin-100, streptomycin-300, sulfamethox-

azole-trimethoprim-1.25, tellurium-512, tetracy-

cline-30, tylosin-128, tobramycin-10, vancomycin-30,

and virginiamycin-512 (AB BIODISK, Piscataway, NJ,

USA). CLSI interpretive criteria for apramycin, bleo-

mycin, oleandomycin, QUADs, sulphonamide and

virginiamycin were not available for some of the bac-

terial species tested. In these cases, when there was

growth around the disks, the isolates were marked as

resistant but when a zone of clearance was present,

they were marked as question marks (Tables 1 and 2).

Probe designing and printing

The entire sequences for 131 genes, represent-

ing 22 different antimicrobials and three salts (ty-

losin, tellurium and QUADs), were copied from the

nucleotide sequence database and BLAST searched

for homologous sequences. The regions of a gene

that had the least homology to unrelated sequences

in the database were used to create a library of 40-

mer oligonucleotide probes and each of the probes

was BLAST searched again. Three probes for each

gene were selected based upon the minimum ho-

mology against the whole database. These were

then aligned against each other and screened for

the similarity index; finally, one probe per gene was

selected. The probes were suspended in water at a

concentration of 100 pmoles/μl. These were then di-

luted with 2 x printing buffer A (MWG Operon). Us-

ing a Gene Machine printer (San Carlos, CA, USA),

the probes were printed on polylysine- and epoxy-

coated slides in triplicate and each experiment was

done at least twice. The probes were printed in

twelve rows and twelve columns and numbered as

1 to 131. Row number 1 had the probes from 1 to

12, row 2 had the probes from 13 to 24, and row 3

had the probes from 25 to 36 and so on. Antisense

probes that were complementary to the 40-mer oli-

gonucleotide probes were labeled with either Cy3 or

Cy5 and obtained from MWG Operon.

Isolation of DNA and RNA

Chromosomal DNA was isolated from bacterial

cells grown overnight in MH broth at 35ºC. The cells

(1 ml) were centrifuged at 10,000 X g and the pel-

let was resuspended in 100 μl of sterile Milli Q (MQ)

water. The DNA was isolated by using QIAmp DNA


Table 1. Disk Diffusion, PCR and Microarray Analysis of Multiple Antibiotic Resistance Markers


Mini kit (Qiagen, Valencia, CA). RNA from a 1 ml bac-

terial culture was isolated using Qiagen’s RNeasy

Mini kit as per manufacturer’s instructions. DNA and

RNA concentrations were determined by measur-

ing the absorbance at 260 nm in a Nanodrop 2000

spectrophotometer (Thermo Fisher Scientific, Wilm-

ington, DE) and using a conversion ratio, A260 of 1 =

50 μg/ml of DNA and 40 μg/ml of RNA. The integ-

rity of an RNA sample was tested by measuring the

A260/280 ratio and analysis on a 1.2% agarose gel con-

taining 0.66 M formaldehyde. An RNA preparation

with an A260/280 ratio of 1.8 to 2.0 and the absence of

visible degradation on agarose gel was considered

good. The gels were run in MOPS electrophoresis

buffer (20 mM MOPS, 5 mM sodium acetate, 1 mM

EDTA, pH 7.0) for 3 h at 90 mV (Antwerpen et al.,

2007), stained with ethidium bromide (1 μg ml-1 in

electrophoresis buffer) and photographed using

a gel documentation system, GDS 8000 (UVP, Inc.

Upland, CA, USA). All the procedures starting from

RNA isolation to cDNA labeling, hybridization and

PCR analysis were replicated at least twice unless

indicated otherwise.

cDNA labeling by reverse transcription

To minimize the loss of RNA by degradation and

the loss of sensitivity, RNA isolated by the Qiagen

method was immediately used to make Cy3-la-

beled cDNA by using CyScribe first-strand cDNA

labeling kit (Amersham BioSciences). Total RNA

(10 μg) in a volume of 10 μl double-distilled water

(ddH2O) was mixed with 1 μl of random hexamer

primers (Amersham Biosciences, USA) and incu-

bated for 5 min at 70°C followed by incubation for

10 min at room temperature. After the incubation,

4 μl of 5 X CyScript buffer, 2 μl of 0.1 M dithioth-

reitol, 1 μl of 2 mM dNTP mix (1 mM dCTP + 1 μl

of 1 mM Cy3-labeled dCTP), and 1 μl of CyScript

reverse transcriptase were added to it. The syn-

thesis of cDNA was carried out in the dark at 42°C

for 1.5 h. RNA template was degraded by alkaline

hydrolysis with 2 μl of 2.5 M NaOH at 37°C for 15

min. The reaction was then neutralized by adding

10 μl of 2 M HEPES free acid. Labeled cDNA was

purified with CyScribe GFX purification kit (Amer-

sham Biosciences) and dried prior to hybridization

using DNA 120 SpeedVac (Thermo Savant, San

a Melting temperature of primers is in parentheses in the primer’s column. The primers within bold lines indi-cate 30-100% homology among the genes within same class of antibiotics

b Presence of a PCR product is indicated by “P” and the strength of the hybridization signal for various genes

in different bacteria is indicated by color in “S/N” column. ■ indicates S/N ≥ 3, less darker shade indicates S/N ≤ 3 but more than 2, and the least darker shade indicates S/N between 1 and 2.

Antimicrobial susceptibility data is not available for apramycin, bleomycin, oleandomycin, QACs, and virgin-iamicin. Therefore, the resistance profile of bacteria for these antibiotics is indicated by “?” mark; S = sensi-tive, R = resistant; I = intermediate resistant

Numbers within parenthesis in front of the gene names indicate their position on microarray slide.


Jose, CA, USA). It was stored at -20°C until use

and dissolved in 10 μl of nuclease-free water be-

fore hybridization.

Hybridization and specificity

A mock hybridization and washing procedure

of the printed slides followed by staining with

SYBR Green I (Molecular Probes, Inc., Eugene,

OR, USA) was carried out to determine whether

the probes remained immobilized or were washed

off from the slides. To determine the specificity of

the probes, Cy3- and/or Cy5-labeled antisense oli-

gonucleotides were hybridized with probes on the

slide. Hybridization was carried out by first using

a mixture of Cy3- and Cy5-labeled antisense oli-

gonucleotides to find out if they hybridized spe-

cifically, inhibited hybridization or cross-hybridized

with other probes. Antisense oligonucleotides or

Cy3-labeled cDNA (10 μl) were heated at 95°C for

2 min followed by a quick cooling in ice bath for

30 sec. A 2 X hybridization buffer (MWG) was pre-

heated to 55°C in a water bath and then combined

with of Cy3-labeled antisense oligonucleotides or

cDNA. The mixture was applied to the microarray

slide and the hybridization solution was spread

evenly. The slides were then covered with a cover

slip and sealed in humid hybridization chambers

that were then put in a 42°C water bath overnight.

The hybridized slides were washed with 1X SSC

and 0.2% (w/v) sodium dodecyl sulfate (SDS) at

55°C for 10 min. Subsequent two washes were car-

ried out with 0.1X SSC, 0.2% (w/v) SDS at 55°C for

10 min followed by rinsing in distilled water for 10

sec. The slides were then placed in a 50 ml capac-

ity plastic tube and centrifuged at 4,000 x g for

3 min to dry the slide surface. Hybridization with

Cy3-labeled cDNA was also carried out using the

above procedure.

PCR detection of antimicrobial resis-tance genes

Primers for the partial amplification of different

antimicrobial resistance genes were designed by

using a primer selection module of the Lasergene

program (DNASTAR, Inc., Madison, WI, USA) and

synthesized by MWG Biotech. PCR amplification

was carried out in a reaction volume of 25 μl by us-

ing a PCR Kit (Applied Biosystems, Foster City, CA,

USA). Each reaction tube contained 5 μl of bacte-

rial DNA (0.1-0.5 μg), 5 μl of a 10 μM mixture of the

forward and reverse primers (Table 1), and 15 μl of

PCR mix (200 μl of PCR mix contains: 33.3 μl of 10 X

XL buffer II, 27 μl of 25 mM magnesium acetate, 66

μl of 10 mM dNTP mix and 7 μl of Taq DNA poly-

merase and 66.7 μl of water). A total of 35 cycles of

amplification were carried out. Each cycle consisted

of 94°C denaturation for 1 min, annealing for 1 min

at 1°C below the lowest Tm of a given primer pair,

and 72°C extension for 5 min. The first denaturation

and the last extension steps were extended for 2

and 15 min, respectively. The PCR amplicons were

analyzed on 1.5% agarose gels.

Signal detection, quantification and val-idation

Two independent hybridization experiments were

performed for each bacterial strain using Cy3-la-

beled cDNA with three replicates of an array on each

slide. Thus, the microarray hybridization results were

from six subarrays for each bacterium tested. After

hybridization, the slides were immediately scanned

using PerkinElmer’s ScanArray Express Scanner (Per-

kin Elmer, Boston, MA, USA) and ScanArray Express

1.1 software at 5 μ resolution. Quantification of the

hybridized signal was carried out using QuantAr-

ray 3.0 software (Packard BioSciences, Billerica, MA,

USA). Signal intensities were measured by the adap-

tive circle method followed by local background

subtraction. The mean of intensities from six hybrid-

ization panels were calculated and signal-to-noise

ratios (S/N) were determined for all the genes. Arbi-

trary signal intensity units of 2000 or a S/N ratio of 3

and above are usually considered as positive for the

expression of the genes and a S/N ratio of less than

2 was considered as negative (Al-Khaldi et al., 2002;

Cassone et al., 2008). Genes with S/N ratios of ≥ 3 are

represented by dark shade squares and those with

■


S/N ratios of ≥ 2 but less than 3 are indicated by light

shade squares in Table 1. In cases where bacteria ex-

hibited resistance to an antimicrobial and the corre-

sponding gene(s) indicated the S/N ratio between 2

and 3, PCR and sequence analysis was carried out to

verify the presence of those genes. Genes were used

as positive controls when detected by PCR and mi-

croarray and their presence corresponded with the

antimicrobial resistance profile. Genes that were not

detected by PCR and microarray and correlated well

with the susceptibility profiles of bacteria were used

as negative controls.

Sequence Analysis

PCR products were purified by a QIAquick gel

extraction kit (Qiagen, Valencia, CA, USA), eluted in

nuclease-free water and directly sequenced with PCR

primers (Table 1) using Sanger’s sequencing method

(Sanger et al., 1977). Nucleotide sequences of the

PCR products were BLAST searched against the ex-

isting GenBank database to confirm their identity.

RESULTS

Slide chemistry and specificity of the probe

Staining of slides with SYBR Green I after post-

print processing indicated that the probes were

present on both the polylysine- and epoxy-coated

slides (Figs. 1A, B). However, after mock hybridiza-

tion at 55°°C, the probes were washed off of poly-

lysine (Fig. 1C) slides but not epoxy-coated slides

(Fig. 1D). Repetition of the experiment yielded sim-

ilar results. Hybridization with a mixture of Cy3- and

Cy5-labeled antisense oligonucleotides indicated

that they hybridized specifically with corresponding

probes on the slide. Cross hybridization with other

probes was also observed (Figs. 2A, B, C, D) but

since the S/N ratio of cross-hybridized probes was 1

or less, it was considered insignificant. These genes

that cross-hybridized were tetB and murX (Fig. 1A),

tetB, murX, and vanD (Fig. 1B), cat, bcrB, tetA,

lmrA, mgt, bla-TEM-1D, bla-VEB, and aacA4 (Fig.

Figure 1. Testing slide chemistry: Slides were stained with SYBR Green I before and after mock hybrid-ization and photographed. A and B. SYBR Green I stained polylysine (A) and epoxy-coated (B) slides before mock hybridization, C and D. SYBR Green I stained polylysine (C) and epoxy-coated (D) slides after mock hybridization

Figure 2. Testing probe specificity: After hybridization with a mixture of Cy3- and Cy5-labeled oligonucle-otidess, the slides were scanned under their respec-tive channels. A. 1-11, 23-33, Cy3; and 12-22, 34-44, Cy5-labeled antisense oligonucleotide hybridization profile. B. 1-11, 23-33, 45-60, 75-90, Cy3 and 12-22, 34-44, 61-74, Cy5-labeled antisense oligonucleotide hybridization profile. C. 45-60, 75-90, Cy3 and 61-74 Cy5-labeled antisense oligonucleotide hybridization profile. D. 91-100, 112-122, Cy3 and 101-111, 123-131, Cy5-labeled antisense oligonucleotide hybrid-ization profile. All the figures are representatives of the replicate experiments.


1C), vraG, lmrA, bla-OXA10, bla-TLA1, aacA1b,

aacC, aadA1, aadB, kamA and llm (Fig. 1D). Amino-

glycoside resistance genes aadA1 and aadA2 had

98.8%, hptII and hyg had 99.2%, and aph4 and nos

had 100% sequence homology. It did not pose a

major problem because only 1 to 3 of them were

present at a given time in a single bacterium and

were considered present after verification by PCR

and sequence analysis.

Microarray detection of antimicrobial resistance genes

Electrophoretic analysis of RNA samples on

agarose gel prior to using in microarray hybridiza-

tion experiments indicated that it was intact (Fig.

3A). Lanes 2 and 3 in figure 3A show the RNA pro-

files of E. faecium ATCC 51559 and E. faecalis ATCC

51229, respectively. Similar RNA profiles were ob-

tained from other bacterial strains used in the study

(data not shown). When total RNA was reverse tran-

scribed to prepare Cy3-labeled cDNA for hybridiza-

tion with the probes, hybridization signals with vary-

ing signal strengths were obtained (Figs. 3B, C, D,

E) for different bacteria tested. The data shown are

for E. faecium, ATCC 51559 (Fig. 3B), S. aureus iso-

late P34 (Fig. 3C), A. veronii isolate Aet2002 (Fig. 3D),

and S. Typhimurium phage type DT16 (Fig. 3E). Most

of the bacteria that exhibited resistance to multiple

antimicrobials had strong hybridization signals (S/N

= 3 or more) are shown by dark shades in Table 1.

Those, exhibiting S/N ratios between 2 and 3 are

shown by slightly lighter shading in Table 1. These

included kamC, hmrB, aad9, tlrB, tet34, tetA(E),

vanB2, and vanD in E. faecalis ATCC 51229; bacA,

cmlA5, ermF, nos, lmrB, hmrB, oleB, oleC-orf5, spa,

tlpD, tlrB, tetA, tetO, ddl2, murX, and vanB in E. fae-

cium ATCC 51559; kamB, kamC, ermF, and tlrB in

A. veronii isolate Avt101; amp, aac(3)IV, bacA, ermA,

aadD, kanR, oleB, oleC-orf5, mecA, aad9, spa, tlrB,

tet, ddl2, and vanB2 in A. veronii Aet2002; mecA,

sulI, tlrB, otrA, tet34, vanB2, and vgb in C. jejuni iso-

late TH200; aac(3)IV, hyg, kanR, oleB, mecA, aad9,

and tet34 in C. jejuni isolate TH205; kamC, himaR,

Figure 3. Hybridization of Cy3-labeled cDNA with antibiotic resistance gene probes. A. RNA analysis on a 1.2% agarose gel. Lanes 1 and 4, RNA ladder; lane 2, RNA from E. faecium ATCC 51559; lane 3, RNA from E. faecalis ATCC 51229. B. Hybridization profile of E. faecium ATCC 51559. C. Hybridization profile of S. aureus strain P34. D. Hybridization profile of A. veronii strain Aet2002. E. Hybridization profile of S. Typhimurium phage type DT16. All the figures are representatives of the replicate experiments.


Table 2. Correlation of microarray and PCR data in bacteria exhibiting resistant, sensitive, and/or inconclusive susceptibility patterns.

OrganismSensitivity

ProfileAntibiotics and their corresponding resistancemarkers used for detection by microarray and PCR

Enterococcus faecalis ATCC 51299 R NEO [neo, himaR, nptII]

S TET [otrA, tetO]

? SUL [spa, sulI]

Enterococcus faecium ATCC 51559 RAMP [amp]; BLE [ble]; NEO [neo, himaR, nptII]; PEN [mecA]; RIF [arr2]

S TET [otrA, tetO]

? OLE [oleB, oleC-orf5]

Aeromonas veronii Avt101 R PEN [mecA]; RIF [arr2]

S CHL [cmlA5]; STR/SPE [aad9]

? APR [aac(3)IV, kamB, kamC]; SUL [spa, sulI]; QAC [qacH]

Aeromonas veronii Aet2002 R RIF [arr2]

S CHL [cmlA5]? SUL [spa, sulI]; QAC [qacH]

Campylobacter jejuni TH200 R AMP [amp]; RIF [arr2]

S CHL [cmlA5]; ERY [ermA, msrSA]; STR/SPE [aad9]

?APR [aac(3)IV]; OLE [oleB]; QAC [qacH]; SUL [spa]; TEL [tlrB]; VIR [vgb]

SUL [spa]; TEL [tlrB]; VIR [vgb]

Campylobacter jejuni TH205 R AMP [amp]

S CHL [cmlA5]; ERY [ermA, msrSA]; STR/SPE [aad9]

? APR [aac(3)IV]; OLE [oleB]; QAC [qacH]; SUL [spa]

Staphylococcus aureus P20 RLIN [linAn2, lmr, lmrA, lmrB, mgt]; OLE [oleB, oleC-orf4, oleC-orf5]

S BAC [bacA, bcrA]; KAN [aadB, kanR]]

? SUL [spa]; VIR [vgb]

Staphylococcus aureus P34 R AMP [amp]

S BAC [bacA, bcrA, bcrB];

? APR [kamC]; SUL [spa]; VIR [vgb]

Salmonella Typhimurium DT16 R RIF [arr2]

S CHL [cmlA5]]; STR/SPE [aad9]; TET [tetA(E), tetO]

? APR [aac(3)IV, kamB]; QAC [qacH]

Salmonella Typhimurium DT21 R AMP [amp]; BLE [ble]; NEO [neo, himaR, nptII]; RIF [arr2]

? APR [aac(3)IV, kamB]; QAC [qacE∂1, qacH]

Names of the antimicrobial agents are indicated by three capital letter codes in bold and their corresponding resistance determinant genes within parentheses. The gene names that are in italics were not detected by either microarray or PCR but those in bold and italics were detected by microarray and PCR. AMP = Ampicillin; APR = Apramycin; BAC = Bacitracin; BLE = Bleomycin; CHL = Chloramphenicol; ERY = Erythromycin; HYG = Hygromycin; KAN = Kanamycin; LIN = Lincomycin; NEO = Neomycin; OLE = Oleandomycin; PEN = Penicillin; QAC = Quater-nary ammonium compounds; RIF = Rifampin; SUL = Sulphonamide; STR = Streptomycin; SPE = Spectinomycin; TEL = Tellurium; TET = Tetracycline; VAN = Vancomycin; VIR = Virginiamyciåån; S = Sensitive; ? = Antimicrobial agents for which either sensitivity criteria was not available or the results of the disk diffusion assay were inconclusive but their corresponding resistance genes were detected by microarray and PCR.


tetM, tet34, ddl2, sfiX, vanB, vanC2-c3, vanE and

vanG in S. aureus isolates P20; nos, lmrB, hmrB, neo,

oleB, oleC-orf5, mecA, sulI, tet, ddl2, ddl, and vanB

in S. aureus P34; ermF, lmr, femB, mecA, spa, tetA(E),

dacB, and vanB in S. Typhimurium phage type DT16;

lmr, femA, femB, oleC-orf5, spa, sulI, tlrB, dacB, sfiX,

vanA, vanB2, and vanE in S. Typhimurium phage

type DT21 (Table 1, Fig. 4A).

Apart from the specific antimicrobial resistance

genes, 1 to 6 additional aminoglycoside resistance

genes were also present in all the bacteria except in

C. jejuni isolate TH205 (Table1). The gene satG was

present in A. veronii isolates Avt101 and Aet2002,

and S. Typhimurium phage types DT16 and DT21.

E. faecalis ATCC 51299, S. aureus isolates P20 and

P34 contained the gene aacA1b. The genes aacA4

and aacC were present in E. faecalis ATCC 51299, A.

veronii isolate Avt101, C. jejuni isolate TH200, and S.

Typhimurium phage types DT16 and DT21. The gene

aadA was present in S. aureus isolate P20, aadA1 in

A. veronii isolate Avt101 and S. Typhimurium phage

type DT16, and aadA2 in E. faecalis ATCC 51299,

A. veronii isolate Avt101, and S. Typhimurium phage

type DT21. The gene speC was present in E. faecalis

ATCC 51299 and E. faecium ATCC 51559. E. faecalis

ATCC 51299, and S. aureus isolates P20 and P34 pos-

sessed the gene ant(3’’)-Ih-aac(6’)-Iid. Moreover, all

the bacteria used in the study also contained 2 to 5

efflux pump genes (Table 1).

In the case of bacteria that were sensitive to a

given antimicrobial agents, some of them were still

found to contain the corresponding antimicrobial re-

sistance genes. These included bacitracin resistance

genes bacA and bcrA in S. aureus isolate P20, and

bacA, bcrA, and bcrB in S. aureus isolate P34; chlor-

amphenicol resistance gene cmlA5 in A. veronii iso-

lates, Avt101 and Aet2002, C. jejuni isolates TH200

and TH205, and S. Typhimurium phage types DT16;

erythromycin resistance genes ermA and msrSA in

C. jejuni isolates TH200 and TH205; kanamycin re-

sistance genes aadB and kanR in S. aureus isolate

P20; streptomycin resistance gene aad9 in C. jejuni

isolates TH200 and TH205, A. veronii isolate Avt101,

and S. Typhimurium phage types DT16; and tetracy-

cline resistance genes otrA and tetO in E. faecium

ATCC 51559 and E. faecium ATCC 51229 (Table1,

Fig. 4B).

In the absence of CLSI interpretive criteria for

apramycin, bleomycin, oleandomycin, QUADs, sul-

phonamide and virginiamycin, susceptibility pat-

terns of bacteria could not be established for these

antimicrobials. Using the criterion of S/N ratio >2

and PCR detection, sulphonamide resistance genes

spa, and sulI were found in E. faecium ATCC 51559.

Oleandomycin resistance gene oleB, and oleC-orf5

were detected in E. faecium ATCC 51559. Apramycin

resistance genes aac(3)IV, kamB, and kamC, sulphon-

amide resistance genes spa, and sulI and QUADs

resistance gene qacH were detected in A. veronii

isolate Avt101. A. veronii isolate Aet2002 contained

sulphonamide resistance genes spa and sulI, and

QUADs resistance gene qacH. C. jejuni isolates

TH200 and TH205 possessed apramycin resistance

genes aac(3)IV, hygromycin resistance genes hyg and

nos, oleandomycin resistance gene oleB, QUADs re-

sistance gene qacH, and sulphonamide resistance

genes spa and sulI. C. jejuni isolate TH200 had ad-

ditional tellurium and virginiamycin resistance genes

tlrB and vgb, respectively (Table 1). S. aureus isolates

P20 and P34 contained sulphonamide resistance

gene spa and virginiamycin resistance gene vgb. S.

enterica serovar Typhimurium phage types DT16 and

DT21 contained QUADs resistance gene qacH.

PCR detection of antimicrobial resis-tance genes

Among resistant bacteria, most of the genes

with S/N ratios of 3 and above could be detected

by PCR. Some with S/N ratios > 2 but < 3 were also

detected (Table 1, Fig. 4A). Some genes associated

with the corresponding antimicrobial agents could

not be amplified from resistant isolates (Table 1, 2).

These included neomycin resistance genes neo, hi-

maR, and nptII in E. faecalis ATCC 51299, E. faecium

ATCC 51559 and S. Typhimurium phage type DT21.

Ampicillin resistance gene amp was missing in E.

faecium ATCC 51559, C. jejuni TH200, and TH205,

S. aureus P34 and S. Typhimurium phage type DT21.

Penicillin resistance gene mecA was absent in E.


faecium ATCC 51559 and A. veronii Avt101. Bacte-

rial strains E. faecium ATCC 51559, A. veronii Avt101,

Aet2002, C. jejuni TH200 and S. Typhimurium phage

types DT16 and DT21 were missing rifampin resis-

tance gene arr2. E. faecium ATCC 51559 and S. Ty-

phimurium phage type DT21 did not have bleomy-

cin resistance gene ble. S. aureus P20 was missing

lincomycin resistance genes linAn2, lmr, lmrA, lmrB

and mgt, and oleandomycin resistance genes oleB,

oleC-orf4 and oleC-orf5 (Table 1, 2).

Most of the antimicrobial resistance genes did not

amplify from sensitive isolates but some yielded am-

plicons corresponding to the antimicrobial agents

which they were sensitive to (Table 1, 2, Fig. 4B). The

antimicrobial agents for which no CLSI interpretive

criteria were available, the corresponding resistance

genes were also detected by PCR analysis (Table 1,

2, Fig. 4B).

DISCUSSION

This study highlights the development of an in-

expensive transcriptomic array method for the de-

tection of multiple antimicrobial resistance genes in

Figure 4. PCR amplification of antimicrobial resistance genes from bacteria that exhibited resistance to different antimi-crobials and had S/N ratio of ≤ 3. A. Lanes 1, 13, and 24, 100 bp ladder; lane 2, kamC; lane 3, aad9; lane 4, bacA; lane 5, ermF; lane 6, hmrB; lane 7, tlpD; lane 8, oleC-orf5; lane 9, ermA; lane 10, amp; lane 11, aadD; lane 12, tlrB; lane 14, mecA; lane 15, bcrB; lane 16, kanR; lane 17, mecA; lane 18, nos; lane 19, neo; lane 20, mecA; lane 21, femB; lane 22, sulI; lane 23, femB. B. Resis-tance genes from bacteria that exhibited sensitivity to different antimicrobials and those with inconclusive susceptibility pat-terns. Lanes 1, 13, and 24, 100 bp ladder; lane 2, tetO; lane 3, sulI; lane 4, otrA; lane 5, oleB; lane 6, cmlA5; lane 7, spa; lane 8, cmlA5; lane 9, ermA; lane 10, tlrB; lane 11, aad9; lane 12, mgt; lane 14, aadB; lane 15, spa; lane 16, bacA; lane 17, bcrA; lane 18, vgb; lane 19, tetA(E); lane 20, tetO; lane 21, qacH; lane 22, aac(3)IV; lane 23, kamB. The names of bacteria from which these genes were amplified are shown on top of the lanes. All the figures are representatives of the replicate experiments.


a variety of multidrug-resistant bacteria. Unlike the

previous methods that utilized expensive modified

oligonucleotide probes and the genomic DNA la-

beling for detection of multiple antimicrobial resis-

tance genes (Call et al., 2003; Frye et al., 2006; Per-

reten et al., 2005; Relógio et al., 2002; Van Hoek et

al., 2005; Volokhov et al., 2003), this method uses in-

expensive and unmodified oligonucleotide probes.

The data generated by this method was compared

and correlated with the results obtained by PCR

and disk diffusion assays. During the course of this

study, we used 40 nucleotide long unmodified oligo-

nucleotide probes to detect 131 genes that confer

resistance to multiple antimicrobial agents in a lim-

ited set of bacteria. Preliminary tests revealed that

the proper slide chemistry played an important role

when unmodified oligonucleotide probes were used

in a microarray experiment. Use of polylysine-coated

slides resulted in the loss of hybridization signals

due to the washing off of the probes during a mock

hybridization experiment. Epoxy-coated slides per-

formed better with unmodified probes as they re-

mained bound to the surface of the slides. Further, a

test of specificity of probes indicated that hybridiza-

tions with commercially synthesized and Cy3/Cy5-la-

beled antisense oligonucleotide sequences yielded

specific hybridization signals. An insignificant level of

cross-hybridization (S/N = 1 or less) was seen with

other genes. In actual tests, hybridization with Cy3-

labeled cDNA, prepared by the reverse transcription

of total RNA from ten different multidrug-resistant

bacteria yielded high intensity (IU 2000 and above)

and low intensity (IU < 2000) hybridization signals for

different antimicrobial resistance genes. Most of the

genes conferring resistance to multiple antimicrobial

agents in bacteria could be detected by PCR and mi-

croarray (S/N > 3) but in some cases where bacte-

ria exhibited resistance to antimicrobial agents, the

observed S/N ratios for the corresponding genes

were between 2 and 3. Using the S/N cut off limit of

3 (Al-Khaldi et al. 2002; Cassone et al. 2008), these

genes would be considered as being absent but

the fact that the susceptibility pattern of the bacte-

ria matched with PCR and sequence analysis data,

a lower level of gene expression in these cases was

deemed as good enough to confer resistance to

the above antimicrobials. Similar observations were

made earlier (Bernstein et al., 2002; Frye et al., 2006;

Wain and Kidgell, 2004, Woodford, 2001) and cor-

responded with our findings.

In most of the bacteria that exhibited resistance to

an antimicrobial agent, the corresponding resistance

genes were detected in them but the presence of re-

sistance genes corresponding to ampicillin, bleomy-

cin, erythromycin, lincomycin, neomycin, oleando-

mycin, penicillin, and rifampin could not be verified

by the above techniques in some of the isolates. The

absence of these genes suggested that the observed

resistance in these isolates was either due to cross

resistance or the involvement of efflux pump genes.

All the bacteria used in the study contained 2 to 5 ef-

flux pump genes (oleB, marA, tetA, tetB, vraD). The

presence of efflux pump genes causes the bacteria

to become resistant to multiple unrelated antimicro-

bial agents and was probably responsible for the ob-

served resistance in these organisms. For example,

the gene marA itself has been shown to produce

cross resistance to tetracycline, chloramphenicol,

ampicillin, nalidixic acid, and ciprofloxacin, as well

as to other toxic chemicals (QUADs) that are pres-

ent in household, industrial, and hospital disinfec-

tants (Alekshun and Levy, 1999). Cross resistance of

kanamycin-resistant mutants of Escherichia coli Q13

has been shown to confer various degrees of cross-

resistance to streptomycin, gentamicin, neomycin,

and dibekacin in vivo (Choi et al., 1980). Absence of

neomycin resistance genes (neo, himaR and nptII)

in E. faecalis ATCC51299, E. faecium ATCC 51559,

and S. Typhimurium phage type DT21 could be due

to the presence of kanamycin resistance observed

in these isolates. Extended spectrum ß-lactamases

such as bla-TEM1D and bla-OXA10 genes that have

been shown to confer resistance to ampicillin and

penicillin (Paterson and Bonomo, 2005) may be re-

sponsible for the observed resistance to these an-

tibiotics in the absence of ampicillin and/or penicil-

lin resistance genes amp and mecA, respectively, in

E. faecium ATCC 51559, A. veronii Avt101, C. jejuni

TH200 and TH205, S. aureus P34 and S. Typhimuri-

um phage type DT21. The absence of rifampin re-


sistance gene arr2, an ADP-ribosylating transferase,

in A. veronii Avt101, Aet2002, C. jejuni TH200, and

S. Typhimurium phage types DT16 and DT21, could

mean the involvement of multidrug resistance pro-

teins. Cloning and expression of a chromosomally-

located multidrug resistance gene cmr from Cory-

nebacterium glutamicum has been shown to confer

increased resistance to rifampin, bleomycin, tetracy-

cline and many other antibiotics in E. coli (Jäger et

al., 1997). Similar proteins may have been involved

in conferring resistance to rifampin in above isolates

and bleomycin resistance observed in E. faecium

ATCC 51559 and S. Typhimurium phage type DT21.

Absence of lincomycin and oleandomycin-specific re-

sistance genes in S. aureus isolate P20 is probably

due to the presence of ermB gene which is present

in this isolate and known to confer cross resistance

to all macrolides, lincosamides and stretogramins B

(Malbruny et al., 2011).

For most of the bacteria that exhibited sensitivity

to an antimicrobial, disk diffusion data corroborated

with microarray and PCR data but some of the bacte-

ria that exhibited sensitivity to an antimicrobial agent

possessed the corresponding antimicrobial resistance

genes as demonstrated by transcriptomic array, PCR

and partial sequence analysis. The question, there-

fore, arises that why did the bacteria show sensitiv-

ity to antibiotics when the corresponding resistance

genes were present? They were detected by PCR for

the simple fact that genomic DNA harboring these

genes was used as a template. Their detection by

transcriptomic array analysis indicated that the genes

were transcribed. After transcription of the resistance

genes, they have to be translated to exert their phe-

notypic effect but if the translation product is either

inhibited or silenced by any of the components of the

growth media and/or growth conditions, the organ-

isms will exhibit a sensitive phenotype for a given an-

timicrobial agent. In E. coli, silencing of the resistance

genes blaOXA-2, aadA1, sul1, and tetA has been re-

ported to be responsible for the observed sensitive

phenotype (Enne et al., 2006) and the results present-

ed in this study also point to a similar phenomenon.

On another note, a growth temperature increase from

37 to 42°C has been shown to up- or downregulate

approximately 20% of the C. jejuni genes (Stintzi,

2003). Change in DNA supercoiling induced by a

wide range of environmental stresses in coli has been

shown to affect the gene transcription (Cheung et al.,

2003). Environmental factors such as heat, acid, and

osmotic stress have been demonstrated to cause dif-

ferential expression of 13 to 18% of the genes (Xie,

2004). Earlier, we have also shown that susceptibility

profiles of the organisms could change from resis-

tant to sensitive in a disk diffusion assay depending

upon the media used for testing (Nayak et al., 2002).

In the light of current observations and the earlier re-

ports, it is possible for a resistance gene to be present

without exerting its phenotypic effect for the reasons

mentioned above. While susceptibility assays by disk

diffusion methods are sensitive to various experimen-

tal conditions, transcriptomic array and PCR analyses

can still provide information regarding the presence

or absence of a resistance gene. The latter, how-

ever, can not differentiate between a transcribed or

untranscribed gene unless cDNA derived from the

RNA transcripts is used as a template in PCR analysis.

Therefore, gene expression analysis by transcriptomic

or proteomic analysis methods would provide more

information compared to the genomic DNA-based

PCR or microarray technologies.

The transcriptomic array technique described in

this study proved more useful in the sense that it

was able to detect the presence of antimicrobial

resistance genes and also provided information

about the status of their transcriptomic expression.

PCR and genomic DNA labeling and detection

techniques, while able to detect the genes, do not

offer an explanation about the presence of resis-

tance genes in sensitive isolates or absence thereof

in resistant isolates and whether they expressed or

remained unexpressed. In genomic DNA labeling

and detection method, the entire genomic DNA

is labeled and used for the detection of antimicro-

bial resistance genes whether they are transcribed

or not. Since transcriptomic expression method

involves the labeling of cDNA during reverse tran-

scription of RNA, only the genes transcribed under

experimental conditions are detected. While each

of the methods described above have certain mer-


its, the transcriptomic array method, apart from

being able to detect the presence and absence

of the antimicrobial resistance genes, appeared to

provide more insight into the mechanism of antimi-

crobial resistance, status of gene expression, core-

sistance, and cross resistance compared to the ge-

nomic DNA labeling and PCR detection methods.

Furthermore, the technique could also be used for

research in food safety, probiotic product develop-

ment, epidemiological screening, and a thorough

and comprehensive screening of multiple antimi-

crobial resistance genes among bacteria from dif-

ferent ecological sources.

CONCLUSIONS

The transcriptomic array method utilizing un-

modified oligonucleotide probes was successfully

used to detect the presence of 131 antimicrobial re-

sistance genes conferring resistance to 22 different

antimicrobials described in this paper. It provides a

valuable insight into the mechanism(s) of resistance,

status of gene expression at transcription level, and

detection of the antimicrobial resistance genes

among bacteria from different ecological sources.

For most of the genes, the transcriptomic array

and PCR data correlated well with the susceptibility

profiles of the isolates used in the study. However,

some of the genes conferring resistance to certain

antibiotics could not be detected by above meth-

ods. Co- or cross-resistance to these drugs and their

extracellular transport due to the presence of 2 to 5

efflux pump genes is believed to confer resistance to

the above antimicrobials. Moreover, the detection of

resistance genes for few antibiotics in phenotypically

sensitive isolates indicated either inactive versions of

these genes or modulation of gene expression.

ACKNOWLEDGEMENT

The authors thank Dr. Doug Wagner and Dr. John

Sutherland of the Division of Microbiology, NCTR,

for their critical reading and evaluation of the manu-

script. We also thank Mr. William Branham and Dr.

James Fuscoe of the Division of Systems Toxicology

at NCTR for their help in printing the array slides.

The work was supported by intramural funding by

the US Food and Drug Administration. Views pre-

sented in this paper do not necessarily reflect those

of the FDA.

REFERENCES

Alekshun, M. and S. Levy. 1999. The mar regulon:

multiple resistance to antibiotics and other toxic

chemicals. Trends Microbiol. 7:410-413.

Al-Khaldi, S. F., S. A. Martin, A. Rasooly, J. D. Evans.

2002. DNA microarray technology used for study-

ing food-borne pathogens and microbial habitats.

JAOAC Int. 85:906-910.

Antwerpen, M. H, M. Schellhase, E. Ehrentreich-

Förster, F. Bier, W. Witte, U. Nübel. 2007. DNA mi-

croarray for detection of antibiotic resistance de-

terminants in Bacillus anthracis and closely related

Bacillus cereus. Mol. Cell. Probes. 21:152-160.

Bernstein, J. A., A. B. Khodursky, P. H. Lin, S. Lin-

Chao, S. Cohen. 2002. Global analysis of mRNA

decay and abundance in E. coli at a single-gene

resolution using two-color fluorescent DNA micro-

arrays. Proc. Nat. Acad. Sci. USA. 99:9697-9702.

Brazas, M.D, R.E. Hancock. 2005. Using microarray

gene signatures to elucidate mechanisms of an-

tibiotic action and resistance. Drug Discov. Today

10:1245-1252.

Call, D. R, M. K Bakko, M. J. Krug, M. C. Roberts.

2003. Identifying antimicrobial resistance genes

with DNA microarrays. Antimicrob. Agents Che-

mother. 47:3290-3295.

Cassone, M., M. Del Grosso, A. Pantosti, A. Giorda-

no, G. Pozzi. 2008. Detection of genetic elements

carrying glycopeptides resistance clusters in En-

terococcus by DNA microarrays. Mol. Cell. Probes.

22:162-167.

Chen, S., S. Zhao, P. F. McDermott, C. M. Schroeder,

D. G. White, J. A. Meng. 2005. DNA microarray for

identification of virulence and antimicrobial resis-

tance genes in Salmonella serovars and Escherich-

ia coli. Mol. Cell. Probes. 19:195-201.

Cheung, K. J., V. Badrinarayana, D. W. Selinger, D.

Janse, G. M. Church. 2003. A microarray-based


antibiotic screen identifies regulatory role for su-

percoiling in the osmotic stress response of Esch-

erichia coli. Genome Res. 13:206-15.

Choi, E. C., T. Nishimura, Y. Tanaka, N. Tanaka. 1980.

In vivo and in vitro cross-resistance of kanamycin-

resistant mutants of E. coli to other aminoglyco-

side antibiotics. J. Antibiot. 33:1527-1531.

Clinical and Laboratory Standards Institute. 2009a.

Performance standards for antimicrobial disk sus-

ceptibility tests; 10th ed. M02-A10. Wayne, PA:

CLSI.

Clinical and Laboratory Standards Institute. 2009b.

Methods for dilution antimicrobial susceptibility

tests for bacteria that grow aerobically; eighth ed.

M07-A8. CLSI , Wayne, PA.

Cockerill III, F.R. 1999. Genetic methods for assess-

ing antimicrobial resistance. Antimicrob. Agents

Chemother. 43:199-212.

Enne, V. I., A. A. Delsol, J. M. Roe, P. M. Bennett. 2006.

Evidence of antibiotic resistance gene silencing in

Escherichia coli. Antimicrob. Agents Chemother.

50:3003-3010.

Frye, J. G, T. Jesse, F. Long, G. Rondeau, S. Porwol-

lik, M. McClelland, C. R. Jackson, M. Englen, P. J.

Fedorka-Cray. 2006. DNA microarray detection of

antimicrobial resistance genes in diverse bacteria.

Int. J. Antimicrob. Agents. 27:138-151.

Garaizer, J., A. Rementeria, S. Porwollik. 2006. DNA

microarray technology: a new tool for the epidemi-

ological typing of bacterial pathogens? FEMS Im-

munol. Med. Microbiol. 47:178-189.

Jäger, W., J. Kalinowski, A. Pühler. 1997. A Coryne-

bacterium glutamicum gene conferring multidrug

resistance in the heterologous host Escherichia coli.

J. Bacteriol. 179:2449-2451.

Khan, S. A, M. S. Nawaz, A. A. Khan, C. E. Cerniglia.

2000. Transfer of erythromycin resistance from poul-

try to human clinical strain of Staphylococcus aure-

us. J. Clin. Microbiol. 38:1832-1838.

Lyons, P. 2003. Advances in spotted microarray re-

sources for expression profiling. Brief Funct. Ge-

nomics Proteomics. 2:21-30.

Majtan, T., L. Majtanova, J. Timko, V. Majtan. 2007. Oli-

gonucleotide microarray for molecular characteriza-

tion and genotyping of Salmonella spp. Strains. J.

Antimicrob. Chemother. 60:937-946.

Malbruny, B., A. M. Werno, D. R. Murdoch, R. Leclercq,

V. Cattoir. 2011. Cross-resistance to lincosamides,

streptogramins A, and pleuromutilins due to the

lsa(C) gene in Streptococcus agalactiae UCN70. An-

timicrob. Agents Chemother. 55:1470-1474.

Nayak, R., S. A. Khan, R. H. Watson, C. E. Cerniglia.

2002. Influence of growth media on vancomycin

resistance of Enterococcus isolates and correlation

with resistance gene determinants. FEMS Microbi-

ol. Lett. 214:159-163.

Paterson, D. L., R. A. Bonomo. 2005. Extended-spec-

trum ß-lactamases: a clinical update. Clin. Microbi-

ol. Rev. 18:657-686.

Perreten, V., L. Vorlet-Fawer, P. Slickers, R. Ehricht, P.

Kuhnert, J. Frey. 2005. Microarray-based detection

of 90 antibiotic resistance genes of Gram-positive

bacteria. J. Clin. Microbiol. 43:2291-2302.

Rasooly, A., K. E. Herold. 2008. Food microbial patho-

gen detection and analysis using DNA microarray

technologies. Foodborne Pathog. Dis. 5:531-550.

Relógio, A., C. Schwager, A. Richter, W. Ansorge, J.

Valcárcel. 2002. Optimization of oligonucleotide-

based DNA microarrays. Nucleic Acids Res. 30:1-30.

Sanger, F., S. Nicklen, A. R. Coulson. 1977. DNA se-

quencing with chain-terminating inhibitors. Proc.

Natl. Acad. Sci. U.S.A. 74:5463-5467.

Schüler, T., A. Nykytenko, A. Csaki, R. Möller, W. Frit-

zsche, J. Popp. 2009. UV cross-linking of unmodi-

fied DNA on glass surfaces. Anal. Bioanal. Chem.

395:1097-1105.

Stintzi, A. 2003. Gene expression profile of Campy-

lobacter jejuni in response to growth temperature

variation. J. Bacteriol. 185:2009-2016.

Van Hoek, A. H., I. M. Scholtens, A. Cloeckaert, H. J.

Aarts. 2005. Detection of antibiotic resistance genes

in different Salmonella serovars by oligonucleotide

microarray analysis. J. Microbiol. Methods. 62:13-23.

Volokhov, D., V. Chizhikov, K. Chumakov, A. Rasooly.

2003. Microarray analysis of erythromycin resistance

determinants. J. Appl. Microbiol. 95:787-798.

Wain, J., C. Kidgell. 2004. The emergence of multi-

drug resistance to antimicrobial agents for the

treatment of typhoid fever. Trans. R. Soc. Trop. Med.

Hyg. 98:423-430.


Woodford, N. 2001. Epidemiology of the genetic el-

ements responsible for acquired glycopeptides re-

sistance in enterococci. Microb. Drug Resist. 7:229-

236.

Xie, Y., L. S. Chou, A. Cutler, B. Weimer. 2004. DNA

microarray profiling of Lactococcus lactis subsp.

lactis IL1403 gene expression during environmental

stresses. Appl. Environ. Microbiol. 70:6738-6747.

Zhu, L. X., Z. W. Zhang, C. Wang, H. W. Yang, D. Jiang,

Q. Zhang, K. Mitchelson, J. Cheng. 2007. Use of a

DNA microarray for simultaneous detection of anti-

biotic resistance genes among staphylococcal clini-

cal isolates. J. Clin. Microbiol. 45:3514-3521.




ABSTRACT

This study was carried out to evaluate the total bacterial load and to isolate and identify antibiotic resis-

tant bacteria in fresh vegetables sold in the local markets of Dhaka city and also to determine its resistance

pattern to antibiotics. The highest Aerobic Plate Count (APC) found in tomato and carrot samples of local

markets were 5.17 log10 colony forming units (CFU/g) (Annanda Bazar Market) and 5.78 log10 CFU/g (Khil-

gaon Market), respectively. However, the lowest APC found in tomato and carrot samples of local markets

were 4.90 log10 CFU/g (Malibagh Bazar) and 5.50 log10 CFU/g (Malibagh Market), respectively. Antibiotic

sensitivity patterns of the isolates were determined and nearly all of them were resistant to commonly used

antibiotics. The percentage of resistant bacteria to the total load was also high. Amoxicillin resistant bac-

teria counts were 3.4%, followed by cefixime 2.15%, and ciprofloxacin count at 1.61%. There were 0.26% of

the bacteria found to be multi drug resistant. Therefore, the fresh vegetables samples collected from local

markets were heavily contaminated with resistant bacteria and are of special concern for human consump-

tion. Plasmid profile, curing and transformation study results demonstrated that resistance to amoxicillin

is plasmid mediated but for cefixime and ciprofloxacin were not. These study results demonstrated that

plasmids are one of the important ways to spread resistance but chromosomal mutation by environmental

selection might also responsible for resistance.

Keywords: Isolation, antibiotic resistant bacteria, fresh vegetables and plasmid profile

INTRODUCTION

Fresh vegetables are essential parts of the diet

of humans. An increased number of microbial in-

fections associated with consumption of fresh veg-

etables have been reported in recent years. Docu-

Correspondence: A. Hossain, [email protected]: +8802-9661920-59 Ext 7735 Fax: + 8802-8615583

mented illnesses have been caused by bacteria,

parasites, and viruses and are transmitted via many

types of fruits and vegetables (Beuchat, 1996; NAC-

MCF, 1999). Since antimicrobials are used in livestock

and crop production to control pathogens, there is

concern about antibiotic resistance development in

these pathogens and subsequent transfer to humans

through contaminated food. Antibiotic resistance

among bacteria associated with food animals has

Antibiotic Resistance and Plasmid Profiles in Bacteria Isolated from Market-Fresh Vegetables

S. Akter1, Rafiq-Un-Nabi2, F. Ahmed Rupa1, Md. L. Bari3 and M. A. Hossain1

1Department of Microbiology, University of Dhaka, Dhaka-1000, Bangladesh2Virology Lab, LSD, ICDDR, B, Mohakhali, Dhaka-1212, Bangladesh

3Food Analysis and Research Laboratory, Center for Advanced Research in Sciences, University of Dhaka, Dhaka-1000, Bangladesh



been well documented; research regarding resis-

tance profiles of bacteria isolated from raw produce

is very limited. New research shows that antibiotic

resistant bacteria also may be ingested with vegeta-

bles (Kilonzo Nthenge, A., 2009). For half a century,

meat producers have fed antibiotics to farm animals

to increase their growth and stave off infections.

Now scientists have discovered that those drugs are

appearing in unexpected places. Vegetables such as

corn, green onion and cabbage absorb antibiotics

when grown in soil fertilized with livestock antibiotics

contaminated manure (Kumar et al., 2005).

Antibiotic resistance is a food safety problem

for several reasons (CDC, 2005). First, antibiotic re-

sistance is increasing to some antibiotics, such as

fluoroquinolones and third-generation cephalospo-

rins. These antibiotics are commonly used to treat

serious infections caused by bacterial pathogens fre-

quently found in food, such as Salmonella and Cam-

pylobacter. Each year, several million people in the

United States are infected with Salmonella and Cam-

pylobacter, which usually cause diarrhea that lasts

about a week (CDC, 2006). Antibiotics are not rec-

ommended for treatment of most of these diarrheal

illnesses, but are used to prevent complications in

infants, persons with weakened immune systems,

and older persons. Antibiotics may be life-saving for

several thousand people each year who have seri-

ous invasive infections, such as bacteremia (infection

in the bloodstream) and meningitis (infection of the

lining of the brain and spinal cord). Salmonella infec-

tions are treated with ampicillin, trimethoprim-sulfa-

methoxazole, fluoroquinolones or third-generation

cephalosporins, but some Salmonella and Campy-

lobacter infections have become resistant to these

antibiotics (CDC, 2005).

A second reason for antibiotic resistance as a food

safety problem is that more people may become ill.

Ordinarily, healthy persons who consume few Salmo-

nella may carry them for a few weeks without hav-

ing any symptoms, because those few Salmonella

are held in check by the indigenous bacteria in their

intestines. However, even a few antibiotic-resistant

Salmonella in food can cause illness if the person

who consumes the contaminated food then takes an

antibiotic for another reason. The antibiotic can kill

normal bacteria in the gut, allowing a few Salmonella

that ordinarily would be unlikely to cause illness, take

over and cause illness (CDC, 2005).

A third possible reason for the resistance-related

problem is that the food supply may be a source of

antibiotic-resistant genes. Harmless bacteria present

in food-producing animals could be resistant, and

humans could acquire these bacteria when they eat

meat products from these animals. Once ingested,

resistant genes from these bacteria could be trans-

ferred to bacteria that cause disease (Woodford and

Ellington, 2007).

According to the Centers for Disease Control and

Prevention (CDC, 2005), more than 70% of the bac-

teria that cause infections acquired in hospitals are

resistant to at least one of the drugs most commonly

used to treat them. Out of the total antibiotics con-

sumed, about 25% are used in hospitals. In Germany,

estimated use of antibiotics in 1998 was approxi-

mately 412 tons (Kümmerer and Henninger, 2004)

and taking the excretion rate into consideration, the

amount of antibiotics released into municipal waste-

water was 305 tons. Furthermore it is predicted that

most of the resistant bacteria originate from hospi-

tals (Klaus Kümmerer, 2009) and which in turn end up

with municipality sewage wastewater. This wastewa-

ter pollution is more alarming in developing coun-

tries such as Bangladesh, because of poor wastewa-

ter management system. Food like vegetables can

be contaminated by resistant bacteria through this

water and resistant bacteria can subsequently spread

through the entire population very easily. Therefore,

the aims of this study were: 1) to isolate and identify

the resistant bacteria; 2) to examine their resistant

pattern; and 3) to examine their plasmid profile and

their effect on resistance.


Sample collection

A total of four kilograms of carrot and tomato

samples were collected from four different local mar-

kets (Khailgaon market, Anada market, Polashi mar-


ket and Malibag market) of Dhaka city. One kilogram

of each carrot and tomato samples were purchased

from each local market and brought aseptically to

the Laboratory of Microbiology, University of Dhaka

for microbial analysis. Special sterile Ziploc bags

were used to carrying all the samples to avoid fur-

ther contamination. Samples were processed and

used within 24 h of collection.

Processing of samples and microbial analyses

The samples were washed with sterile distilled

water and cut into small pieces in a biosafety cabi-

net using sterile knives. Twenty-five grams of each

tomato and carrot samples were measured carefully

and accurately by using a weighing machine and dis-

solved in 225 mL of sterile physiological saline (0.85%

NaCl solution) in a stomacher bag, separately. Each

sample was then stomached for 90 seconds and 0.1

mL of the diluted and undiluted stomacher treated

samples was subsequently spread plated onto Plate

Count Agar (PCA; Oxoid Ltd., Hampshire, England)

for the determination of total bacterial load and

onto DifcoTM Muller Hinton Agar (MHA; Becton,

Dickinson Company Ltd, Sparks, MD, USA) contain-

ing antibiotics for the isolation of antibiotic resistant

bacteria. The MHA plates were supplemented with

0.16 μg/mL ciprofloxacin, 5 μg/mL cefixime, 10 μg/

mL amoxicillin separately. For isolating multi drug

resistance (MDR) bacteria, MHA agar plates supple-

mented with all three antibiotics simultaneously at

the same concentrations mentioned were used. The

plates were then incubated at 37°C for 24 h and vi-

able bacteria were enumerated.

Biochemical tests

Biochemical tests were done according to the

manual for general bacteriology of the American

Society of Microbiology (1981). Biochemical tests

conducted in this study were as follows: Oxidase

Test, Catalase Test, Carbohydrate fermentation/

Utilization test, Kligler’s Iron Agar (KIA) test, Hydro-

gen sulfide production (H2S), Methyl Red (MR) test,

Voges-Proscauer (VP) test, Citrate Utilization test, Ni-

trate reduction test, Indole production, Urea (MIU)

tests, and Motility tests were performed to identify

the bacteria of interest (Cappuccino and Sherman,

1990).

Antibiotic susceptibility tests

Susceptibility of isolated bacteria to different an-

timicrobial agents was measured in vitro according

to Bauer-Kirby methods (1996). Briefly, commercially

available antimicrobial discs (Oxoid Ltd, Basing-

stoke, Hants, UK) of Streptomycin (S = 10 μg/mL),

Ciprofloxacin (Cip = 5 μg/mL), Ampicillin (A = 25

μg/mL), Rifampicin (R® = 5 μg/mL), Oxytetracyclin

(OT =30 μg/mL), Cephalosporin (Ce/Cef = 30 μg/

mL), Cefixime (CFM = 30 μg/mL) and Chlorampheni-

col (C© = 30 μg/mL) were placed on the inoculated

agar plates and incubated in an upright position

overnight at 37ºC. The results were expressed as the

zone of inhibition around the antibiotic disc.

Minimum inhibitory and bactericidal concentrations

The minimum inhibitory concentrations (MICs) of

all antimicrobials were determined by microdilution

techniques in Mueller-Hinton broth according to

Sanches et al. (2005). The inocula were prepared at

a density adjusted to a 0.5 McFarland turbidity stan-

dard 108 colony-forming units (CFU/mL) and diluted

1:10 for the broth microdilution procedure. Microti-

ter plates were incubated at 37°C and the MICs were

recorded after 24 h of incubation. Two susceptibility

endpoints were recorded for each isolates. The MIC

was defined as the lowest concentration of antimi-

crobials at which the microorganism tested did not

demonstrate visible growth. Minimum bactericidal

concentration (MBC) was defined as the lowest con-

centration yielding negative subcultures or only one

colony.

Statistical analysis

All trials were replicated three times. Reported


plate count data represent the mean values obtained

from three individual trials, with each of these values

being obtained from duplicated samples. Data were

subjected to analysis of variance using the Microsoft

Excel program (Redmond, Washington DC, USA).

Significant differences in plate count data were es-

tablished by the least-significant difference at the 5%

level of significance.

Isolation of Plasmid DNA

A single bacterial colony was transferred into

a conical flask containing Luria Broth (LB) medium

amended with the appropriate antibiotics and incu-

bated at 37°C overnight with mild shaking (180 x g).

After incubation, 1 mL of the culture was taken and

centrifuged (16,000 x g; 30 s) at 4°C. The superna-

tant was removed and the pellet was resuspended

in 150 μL of Tris- EDTA buffer 10 mM Tris chloride

(pH 8), 1mM EDTA (pH 8)] solution by vigorous vor-

texing. Two hundred microliter of NaOH-SDS (0.2 M

NaOH, 1% SDS) solution and 150 μL of 3M potas-

sium acetate (pH 4.8) were then added and vortexed

for 10 s. The content was centrifuged (16,000 x g; 5

min) again at 4°C and the supernatant was precipi-

tated with 600 μL of ice cold ethanol (Birnboim and

Doly, 1979). A portion (15 μL) of plasmid DNA was

loaded on to a 1.0% agarose gel containing 0.5 μg

m L-1 ethidium bromide and electrophoresed in TBE

(Tris -Boric acid- EDTA) buffer. The plasmid DNA

were visualized by placing the gel on a UV (300 nm)

transilluminator and recorded using the digital doc-

umentation imaging system (model universal Hood

II; BIO-RAD Laboratories, Hercules, CA , USA).

Plasmid curing

Curing of R-plasmid was done according to the

method of Tomeda et al. (1968). Test organisms were

selected from which plasmids had been isolated pre-

viously. An overnight culture of each test organism

in LB containing amoxicillin was diluted to 104 cfu/

mL using freshly prepared sterile LB by serial dilu-

tion technique. From this diluted culture, 0.5 mL was

added with 4.5 mL LB containing different concen-

trations of curing agents Ethidium Bromide: 75 μg/

mL, 100 μg/mL and 125 μg/mL. Thus the resulting

concentration was 103 cfu/mL and the cultures were

then incubated at 37ºC in an orbital shaker at 180 x g

for 48 hours. After incubation, the broth culture was

again diluted to 103 cfu/mL with sterile physiologi-

cal saline. A 10 μL aliquot was subsequently spread

Table 1. Total viable bacterial count & resistant bacterial count found in tomato & carrot samples from four different markets in Dhaka citya

Sample Tomato Carrot

ResistanceTotal count

AmoxR CefixR CiproR MDRTotal count

AmoxR CefixR CiproR MDR

Sampling site

log(CFU/g)

log(CFU/g)

log(CFU/g)

log(CFU/g)

log(CFU/g)

log(CFU/g)

log(CFU/g)

log(CFU/g)

log(CFU/g)

log(CFU/g)

Anondo Market

5.17A 4.17 A 3.71B 3.41 B 2.77B 5.71 A 3.85 B 3.74 A 3.74 A 2.00 B

Malibagh Market

4.98 B 4.04 B 3.90 A 3.64 A 2.60 B 5.56 B 3.94 A 3.75C 2.69 C 2.95 A

Khilgaon Market

5.11 A 4.00 B 3.93 A 3.49 B 2.84 A 5.78 A 3.97 A 3.86 A 3.77 A 3.04 C

Palashi Market

5.07 B 3.99 B 3.96 A 3.71 A 3.04 C 5.57B 3.95 A 3.89 A 3.81 A 3.11 C

aData represent the mean values (n=6) obtained from three individual trials, with each of these values being obtained from duplicated samples and are expressed in logarithmic colony forming unit per gram (CFU/g). Significant differences in plate count data were established by the least-significant difference at the 5% level of significance. Mean values with the same letter in the same column are not significantly different (P<0.05).


on Luria Agar medium. After 24 hour incubation at

37°C, plates were examined for growth. From this

plate culture, individual colonies were selected and

picked with a sterile toothpick on one Luria agar me-

dium without antibiotic and one Luria agar contain-

ing amoxicillin (25 μg/mL), with a numbered grid line

attached on the bottom of each plate. After 24-hour

incubation at 37°C, plates were observed for the

cured cells. The cured plasmid cells were detected

comparing the development of bacterial colonies on

antibiotic containing plate with that of the normal

(without antibiotic) plate. The samples that showed

colonies on normal LB agar but failed to grow on LB

agar supplemented with amoxicillin were the pos-

sible cured isolates.

Plasmid transformation

Producing Competent Cells

Plasmid DNA from different strains was extracted

by the method of Birniboirn & Doly (1979). A single

colony from a freshly grown plate was picked and

dispersed in 100 mL of LB media in a 1 L flask. The

culture was incubated at 37ºC with vigorous shaking

for approximately 3 hours. Cell density was moni-

tored by determining optical density (OD)600 and was

kept at less than 108 cfu/mL (log phase of growth). A

50 mL aliquot of this culture was transferred to a 50

mL conical tube and centrifuge at 4,000 x g for 10

min at 4°C. The supernatant was decanted and the

pellet was resuspended in 10 mL of ice cold 0.1 M

CaCl2. After resuspension, cells were centrifuged at

4,000 x g for 10 min. The supernatant was decanted

and the pellet was resuspended in 1.0 mL of ice cold

0.1 M CaCl2. Competent cells were stored by add-

ing ice-cold sterile glycerol to a final concentration

of 10% (v/v). Cells were mixed and left on ice for 30

min, then stored at -70ºC.

Plasmid Transformation

To transform, 0.1 mL of competent cells were

transferred into each of 3 chilled Eppendorf tubes

and labeled the tubes. A 1 ng amount of known plas-

mid (DH5α) to one tube and 10 ng was added to

the other and the tubes placed on ice for 30 min,

then held at 42ºC for 120 seconds and the cells were

returned to ice for 1 to 2 minutes. One hundred μL

of the diluted and undiluted transformation mixtures

were spread plated on to amoxicillin containing Lu-

ria agar plates and incubated at 37ºC overnight. To

calculate the competencies of cells, the number of

colonies on the plate were divided by the amount of

DNA (in ng) added to the transformation (Sambrook

et al. 1989).


Antibiotic resistance is a worldwide public health

problem that continues to grow. It occurs when

strains of bacteria harbored in the human body be-

come resistant to antibiotics due to improper use

and abuse of antibiotics. For this study, we collected

samples from four different local markets in Dhaka

city. The total bacterial load in tomato sample was

between 4.98 to 5.10 log CFU/g in 4 different local

markets in Dhaka city. However, the bacterial loads

in the carrot samples were 5.50 to 5.70 log CFU/g,

which is significantly (P<0.05) higher than that of to-

mato samples. (Table 1).

The percentage of resistant bacteria to the to-

tal viable bacterial load was also high. Amoxicil-

lin resistant bacterial count was 3.4%, followed by

cefixime 2.15%, and ciprofloxacin count at 1.61%.

There were 0.26% of bacteria that exhibited multi

drug resistance (MDR) (Table 2). A total of 2607 re-

sistant bacteria were found and of them, amoxicil-

lin resistant bacteria were 42%, cefixime resistant

bacteria were 37%, ciprofloxacin resistant bacteria

were 18% and multi drug resistant bacteria were

3% (Figure 1). The percentage of amoxicillin resis-

tant bacteria to all resistant bacteria was found to

Table 2. Percentage of resistant bacteria to total viable bacterial countTotal

Bacteria% amoxicillin

resistant bacteria% cefixime resistant

bacteria% ciprofloxacin

resistant bacteria% MDRbacteria

2.37×106 3.4 2.45 1.61 0.26


be at the highest frequency.

Based on morphology, colony characteristics,

single colonies were picked up randomly by sterile

loop and studied further. The colonies were then

plated on several differential and selective media

such as McConkey Agar media, Eosine Methylene

Blue (EMB) agar, Cetrimide, XLD (Xylose-lysine-

Deoxycholate) agar and TCBS media. Morphologi-

cal studies and biochemical tests of the isolates

revealed that both Gram-negative bacteria includ-

ing Escherichia coli, Shigella dysenteriae, Klebsilla

pneumonia, Salmonella Typhimurium and Gram-

positive bacteria including Pseudomonas vulgaris,

Micrococcus luteus, Corynebacterium xerosis, Strep-

tococcus lactis, Staphyloccus aureus were present in

the samples. The most common resistant bacteria

were found to be E. coli. Based on the morphology,

colony characteristics, gram staining and biochemi-

cal test, 22 colonies were picked for further study

(data not shown).

Minimum bactericidal concentration (MBC) and

MIC tests were performed using a microdilution

method. The MIC for amoxicillin ranged from 32 to

64 μg/mL and MBC ranged from 64 to 128 μg/mL.

MIC for cefixime ranged from 16 to 32 μg/mL and

MBC ranged from 32 to 64 μg/mL. MIC for ciproflox-

acin ranged from 1 to 2 μg/mL and MBC ranged from

4 to 8 μg/mL (Table 3). These results are in agree-

ment with the results of other researchers (Schwartz

et al., 2003; Willis 2000; Nemi et al., 1983; Klech and

Lee, 1978). These results also suggest that the bac-

teria can become resistant to higher concentrations

of antibiotics.

An antibiogram study of 22 isolates showed that

the isolates were also resistant to streptomycin,

ampicillin, rifampicin, oxytetracyclin, cephalospo-

rin and chloramphenicol (Table 4). Most of isolates

were resistant to ampicillin (18), followed by oxytet-

racyclin (14), and very low resistance to ciprofloxa-

cin (4) and cephalosporin (4). These experimental

results also suggested that multidrug resistance to

environmental bacteria was increasing and almost

every bacterium was resistant to more than one an-

tibiotic (Table 4).

Plasmid profile of the 22 isolates showed that

five isolates (4, 7, 11, 14 and 21) contained plasmid.

Isolate 4 had almost 8 kb plasmid and isolates 7, 11

and 21 had greater than 5 to 6 kb plasmid. Isolate

14 contained two plasmids and they were 8 and

greater than 4 to 5 kb plasmid (Figure 2). Kalantar

et al. (2011) studied the presence of plasmids of mo-

lecular sizes ranging from 1.4 kb to 4.5 kb among

the acute diarrhea causing E. coli isolates showing

resistance to ampicillin (A), chloramphenicol (C) and

tetracycline (T), and stated that these resistances are

plasmid mediated. However, the authors did not car-

ry out conjugation or plasmid curing experiments in

order to establish the involvement of the plasmid in

carrying ACT-resistance among the isolates studied.

Curing tests were performed to determine the

role of plasmid on resistance. Curing of amoxicillin

isolate 7, cefixime isolate 14 and ciprofloxacin iso-

late 21 were done. After curing they were subjected

to an antibiotic susceptibility assay and it was found

that only isolate 14 showed sensitive to amoxicillin

after curing. However the other two isolates (7 and

21) were not sensitive to their respective resistant an-

tibiotics (data not shown).

For the confirmation of curing, the plasmid of

cured bacteria was again isolated and after agarose

Figure 1. Total resistant bacteria and the per-centages of resistant bacteria to particular an-tibiotics


gel electrophoresis, the three cured isolate showed

that they had lost their plasmid. However, the growth

of isolates 7 in cefixime and isolates 21 in ciprofloxa-

cin indicate that their resistance is not dependent on

the presence of plasmid (data not shown). Therefore,

a transformation study was done to confirm that the

antibiotic resistance is plasmid mediated.

Transformation by plasmid on DH5α from isolate

14 was performed to confirm that plasmid was re-

sponsible for the resistance. After transformation,

the bacteria was placed on media containing amoxi-

cillin antibiotics and incubated at 37°C and after 48

h of incubation small colonies were observed. The

colonies were then analyzed for amoxicillin suscepti-

bility, and it was found that the transformed isolates

were fully resistant to amoxicillin (data not shown).

Figure 2. Agarose gel electrophoresis of plasmid DNA of resistant isolates. E. coli PDK-9 is used as a marker; CHR indicates the banding position of the chromosomal DNA.

Table 3: The minimum inhibitory concentrations (MIC) and minimum bactericidal concentrations (MBC) of the isolates

Antibiotics Number of isolates Suspected microorganisms MIC(μg/ml) MBC(μg/ml)

Amoxicillin

12

141516

E.coliS. dysenteriae

E.coliK. pneumonia

M. lucteus

3264323264

641286464128

Cefixime

345671718

K. pneumoniaK. pneumonia

S. TyphimuriumK. pneumonia

E.coliM. lucteusM. lucteus

16161616321616

64323264643264

Ciprofloxacin

8910111920

Not identifiedE.coli

P. vulgarisE. coli

S. aureusS. lactis

221221

484484


From plasmid profile, curing and transformation

study, it can be concluded that resistance to amoxi-

cillin is plasmid mediated but for cefixime and cip-

rofloxacin resistance was not plasmid mediated.

The presence of plasmid mediated resistance for

amoxicillin has also been reported by Cattoir et al.

(2008). The presence of a similar plasmid in cefixime

resistant bacteria has been reported by Kalantar et

al. (2011). The ciprofloxacin resistant isolates plas-

mid profiles shows similarity to the data reported

by Ferrero et al. (1995). These study results demon-

strated that the plasmid is one of the important ways

to spread resistance but chromosomal mutation by

environmental selection might also responsible for

resistance. Acquisition of antibiotic resistant proper-

ties of a bacterium could be its inherent properties

or occurs due to chromosomal mutation(s) or by ac-

quiring extra-chromosomal DNA plasmid (Mandal et

al., 2004).

CONCLUSIONS

In conclusion, the result of this study demonstrat-

ed that the fresh vegetable samples collected from

local markets were heavily contaminated with resis-

tant bacteria and is of special concern for human

consumption. Plasmid profile, curing and transfor-

mation study results demonstrated that resistance to

amoxicillin is plasmid mediated but for cefixime and

ciprofloxacin were not. These study results demon-

strated that the plasmid is one of the important ways

to spread resistance but chromosomal mutation by

Table 4. Antibiotic sensitivity pattern of the 22 isolates against 8 commonly used antibiotics.

Isolate No. SuspectedMicroorganisms Resistant Intermediate Sensitive

1 E.coli Amp, R, OT - Cip, Cef, S,C

2 S. dysenteriae Amp, OT - Cip, Cef, C,S,OT,R

3 K. pneumonia Amp, C, Cef, R, S,OT CFM Cip

4 K. pneumonia Amp, C, Cef, R,OT CFM Cip, S

5 S. Typhimurium Amp, Cef, CFM C Cip,R, S,OT

6 K. pneumonia Amp, R, CFM - C, Cef, S, OT, Cip

7 E .coli Amp, Cip, Cef, OT, CFM S C,R

8 Not identified Cef - Amp, S, C, Cip, R, OT

9 E. coli Amp, C, Cef, Cip, R,OT - S

10 P. vulgaris - S Amp, C, Cef, Cip, R,OT

11 E.coli Amp, S - C, Cef, Cip, R, OT,

12 M. lucteus Amp, Cef, OT S C, Cip, R

13 C. xerosis C, Cip, S Amp Cef, R,OT

14 E. coli Amp, Cip, Cef, OT - C,S,R

15 K. pneumonia Amp, C, Cef, R,S - Cip, OT

16 M. lucteus Amp, Cef, OT R S, Cip, C

17 M. lucteus Amp, C, Cef, Cip, S,OT, CFM R

18 M. lucteus CFM - Amp, C, Cef, Cip, S, R, OT

19 S. aureus R,OT Cef Amp,C,Cip,S

20 S. lactis OT S,R Amp, Cef, C, Cip

21 S. Typhimurium Amp, Cef, C,Cip, OT - C,S

22 S. aureus Cef, R,OT - Amp, C, Cip, S

Abbreviations: S= Streptomycin, Cip= Ciprofloxacin, A = Ampicillin, R = Rifampicin, OT= Oxytetracyclin, Cef= Cephalo-sporin, CFM= Cefixime and C= Chloramphenicol .


environmental selection might also responsible for

resistance.

Further research is needed to determine all ma-

jor sources of antibiotic resistant food borne patho-

gens in fresh produce. Data on the prevalence and

types of antibiotic resistance in microorganisms iso-

lated from fresh produce may help explain the role

of foods in the transmission of antibiotic-resistant

strains to human populations.

REFERENCES

American Society of Microbiology. 1981. Manual for

general bacteriology. Ourfood.com. http://our-

food.com/general_bacteriology.html. Accessed

July, 2010.

Bauer, A. W., W. M. Kirby, J. C. Sheris, M. Turck. 1996.

Antibiotic susceptibility testing by a standardized

single disc method. Am. J. Clin. Pathol. 45:149-158.

Beuchat, L. R. 1996. Pathogenic microorganisms as-

sociated with fresh produce. J. Food Prot. 59:204–

216.

Birnboim, H. C. and J. Doly. 1979. A rapid alkaline

extraction procedure for screening recombinant

plasmid DNA. Nucleic Acids Res. 7:1513–1523.

Cappuccino, J. G., and N. Sherman. 1990. Microbiol-

ogy: A laboratory manual. 4th Edition. The Benja-

min-Cumming Publishing Company, Inc., Menlo-

park, CA. 186 p.

Cattoir, V., L. Poirel, P. Nordmann. 2008. Plasmid-me-

diated quinolone resistance pump QepA2 in an

Escherichia coli isolate from France. Antimicrob.

Agents Chemother. 52:3801–3804.

CDC. 2005. National Antimicrobial Resistance Moni-

toring System (NARMS) Frequently Asked Ques-

tions (FAQ) About Antibiotic Resistance - Why is

antibiotic resistance a food safety problem? Cdc.

gov. http://www.cdc.gov/narms/faq_pages/5.htm.

Accessed January, 2011.

CDC. 2006. Progress report: Implementation of a

public health action plan to combat antimicrobial

resistance. Cdc.gov. http://www.cdc.gov/drugre-

sistance/actionplan/2005report/CDC_6-21-06.pdf,

Accessed January, 2011.

Cimitile, M. 2009. Crops absorb livestock antibiot-

ics, science shows. Environmental Health News.

Environmentalhealthnews.org. Available at http://

www.environmentalhealthnews.org/ehs/news/an-

tibiotics-in-crops. Accessed January, 2011.

Ferrero, L., B. Cameron, J. Crouzet. 1995. Analysis

of gyrA and grlA mutations in stepwise-selected

ciprofloxacin-resistant mutant of Staphylococcus

aureus. Antimicrob. Agents Chemother. 39:1554–

1558.

Kalantar, E., F. Soheili, H, Salimi, D. M. Soltan. 2011.

Frequency, antimicrobial susceptibility and plas-

mid profiles of Escherichia coli pathotypes ob-

tained from children with acute diarrhea. Jundis-

hapur J. Microbiol. 4:23-28.

Kilonzo Nthenge, A., E. Rotich, C. Thompson. 2009.

Antibiotic-resistant bacteria isolated from organic

and convectional fresh produce. National Institute

of Food Technology p. 49. http://www.reeis.usda.

gov/web/crisprojectpages/213697.html Accessed

on May 2011.

Kumar, K., S. C. Gupta, S. K. Baidoo, Y. Chander, C. J.

Rosen. 2005. Antibiotic uptake by plants from soil

fertilized with animal manure. J. Environ. Qual. 34:

2082-2085.

Kümmerer, K. and A. Henninger. 2004. Promoting

resistance by the emission of antibiotics from hos-

pitals and households into effluents. Clin Microbiol

Infect. 9:1203–1214.

Kümmerer, K. 2009. Antibiotics in the aquatic en-

vironment – A review – Part II. Chemosphere 75:

435-44.

Klech, W. J. and L. S. Lee. 1978. Antibiotic resistance

patterns of Gram-negative bacteria isolated from

environmental sources. Appl. Environ. Microbiol.

36:450-456.

Mandal, S., M. D. Mandal, N. K. Pal 2004. Plasmid-

encoded multidrug resistance of Salmonella typhi

and some enteric bacteria in and around Kolkata,

India: a preliminary study. Online J. Health Allied

Scs. 4:1-7.

NACMCF (National Advisory Committee on Mi-

crobiological Criteria for Foods). 1999. Microbio-

logical safety evaluation and recommendations on

fresh produce. Food Control 10:117–143.

Nemi, M., M. Sibakov, S. Niemela. 1983. Antibiotic


resistance among different species of fecal coli-

forms isolated from water samples. Appl. Environ.

Microbiol. 45:79-83.

Sambrook, J., E. F. Fritsch, T. Maniatis 1989. Molecu-

lar cloning: A laboratory manual. Second edition,

Cold Spring Harbor Laboratory Press. 1:74-135.

Sanches, N. R., D. A. Garcia, M.S. Schiavini, C. V. Na-

kamura, B. P. D. Filho. 2005. An evaluation of an-

tibacterial activities of Psidium guajava. Brazillian

Arch. Bio and Technol. 48:429-436.

Schwartz, T., T. Kohnen, B. Jansen. 2003. Detection

of antibiotic-resistant bacteria and their resistance

genes in wastewater, surface water, and drinking

water biofilms. FEMS Microbiol. Ecol. 43:325–35.

Tomeda, M., M. Inuzuka, N. Kubo, S. Nakamura.

1968. Effective elimination of drug resistance in

sex factors in Escherichia coli by sodium dodecyl

sulfate. J. Bacteriol. 95:1078-1089.

Turnidge, J. D., M. J. Ferraro, J. H. Jorgensen. 2003.

Susceptibility test methods: General consider-

ations. In: P. R. Murray, E. J. Baron, J. H. Jorgensen,

M. A. Pfaller, R. H. Yolken. Manual of clinical micro-

biology. 8th Ed. American Society of Clinical Mi-

crobiology, ASM Press, Washington. p 1102-1107.

Willis, C. 2000. Antibiotics in the food chain: their

impact on the consumer. Rev. Med. Microbiol.

11:153-160.

Woodford, N. and M. J. Ellington. 2007. The emer-

gence of antibiotic resistance by mutation. Clin.

Microbiol. Infect. 13:5–18.




ABSTRACT

Melons (Cucumis melo L.) have been associated with many foodborne illness outbreaks. Current melon

handling recommendations do not require cool temperature storage to maintain the safety of whole mel-

ons, yet storage temperature often influences microbial growth. While calcium chloride (CaCl2) dips have

been shown to reduce post-harvest decay and increase shelf-life of many different fruits, the anti-microbial

effect is unknown. It is also unclear whether production method (organic or conventional) impacts micro-

bial growth on fresh produce. This study evaluated the impact of a 20 min 0.08 M CaCl2 post-harvest dip

treatment and two storage temperatures (10˚C and 21˚C) on total aerobic and Enterobacteriaceae bacterial

counts present on the surface of organically and conventionally grown melons (cultivar ‘Arava’) stored for

10 d. Storage temperature significantly impacted microbial growth, as higher levels of aerobic and Entero-

bacteriaceae bacteria grew overall on melons stored at 21˚C vs. 10˚C (p<0.05). Storage time did not impact

bacterial counts within the 10 d. Organic melons had overall more aerobic growth than conventional mel-

ons (p<0.05), with non-dipped organic melons stored at 21˚C having the highest aerobic bacterial counts.

Organic melons treated with CaCl2 and stored at 21˚C had lower (p<0.05) Enterobacteriaceae levels than

non-dipped organic melons stored at 21˚C. Based on the results of this study, storing whole ‘Arava’ melons

at cooler temperatures is best to slow microbial growth, and the use of a CaCl2 dip treatment may be a

useful method to minimize growth on organic melons.

Keywords: post-harvest storage, calcium chloride, food safety, aerobic bacteria, Enterobacteriaceae,

organic production, melons, storage temperature

INTRODUCTION

Foodborne illness outbreaks associated with fresh

produce have increased in the United States dur-

Correspondence: Patricia Kendall, [email protected], Tel: +1 -970-491-0799

ing the last thirty years (Sivapalasingam et al., 2004;

Doyle and Erickson, 2008). Fresh fruits and vegetables

are especially vulnerable to causing foodborne out-

breaks due to the fact that they are often consumed

raw or minimally processed and each step, from plant-

ing through post-harvest handling, may contribute to

the microbial load (Johnston et al., 2005).

BRIEF COMMUNICATIONSImpact of Calcium Chloride Dip and Temperature on Microbial

Quality of Organically and Conventionally Grown Melons H. T. Aldrich1, L. Goodridge2, M. Bunning1, C. Stushnoff3 and P. Kendall1*

1Department of Food Science and Human Nutrition 2Department of Animal Sciences

3Department of Horticulture and Landscape ArchitectureColorado State University, Fort Collins, CO 80523



Melons (Cucumis melo L.) are a popular fruit in

the United States that have been associated with

numerous foodborne illness outbreaks. There were

at least 28 documented outbreaks associated with

cantaloupes and muskmelons between 1984 and

2003 (Bowen et al., 2006), indicating a strong need to

increase the safety of fresh melons. The foodborne

agents implicated in the contamination events in-

cluded Salmonella enterica, Campylobacter jejuni,

Escherichia coli O157:H7, and norovirus, with both

domestic and imported melons implicated in past

outbreaks.

There is some concern among public health ex-

perts that organic production methods increase the

risk for microbial contamination due to increased

manure use, compared to using chemical fertilizers

in conventional farming (Stephenson, 1997a,b). Be-

cause organic foods are one of the fastest growing

food categories with sales increasing nearly 20%

each year since 1990 (Winter and Davis, 2006), the

impact of organic production practices on food

safety remains an important issue to evaluate. Stud-

ies assessing consumer perceptions of organic pro-

duce have often found that consumers think organic

produce is safer, more nutritious, and better tasting

than conventionally grown produce (Torjusen et al.,

2001; Magnusson et al., 2003; Shepherd et al., 2005;

Yiridoe et al., 2005). However, research comparing

such attributes has produced inconsistent or incon-

clusive results, most likely due to unparallel growing

conditions, cultivar choices, and other uncontrolled

variables (Harker, 2004; Lester, 2006). Well-controlled

studies are needed to better understand the impact

of organic and conventional growing methods on

microbial growth on fresh produce.

Calcium plays an important role in maintaining

cell wall structure in fruit by interacting with pectic

acid to form calcium pectate, which has a firming

effect on cell walls; thus, calcium deficiency during

the growing process has been shown to cause a vari-

ety of physiological disorders in produce (Poovaiah,

1986). Because of the role calcium has in maintaining

cell walls, researchers have investigated the use of

post-harvest calcium treatments to increase the qual-

ity and shelf-life of fruit after harvest (Martin-Diana

et al., 2007). One such calcium treatment showing

positive results on fruits such as strawberries, lem-

ons, peaches, and melons has been calcium chloride

(CaCl2) dips (Garcia et al., 1996; Lester and Grusak,

2001; Tsantili et al., 2002; Manganaris et al., 2007;

Martin-Diana et al., 2007; Lysiak et al., 2008). Con-

centrations of CaCl2 used in the above studies have

ranged from 0.045-0.27 M, with the recommended

concentrations falling within the 0.06-0.09 M range,

depending on the fruit being studied and whether

the fruit was treated whole or fresh-cut (Garcia et al.,

1996; Lester and Grusak, 2001; Tsantili et al., 2002;

Manganaris et al., 2007; Martin-Diana et al., 2007;

Lysiak et al., 2008). The optimal CaCl2 concentration

found to slow senescence without any negative side

effects in a study using whole honeydew melons was

0.08 M (Lester and Grusak, 2001).

To our knowledge, the microbial effects of using

CaCl2 as a post-harvest treatment have not been

studied. Research by Chikthimmah et al. (2005)

found that using CaCl2 in mushroom irrigation wa-

ter during crop production resulted in lower levels of

microbial growth during storage compared to mush-

rooms grown without CaCl2 added to the irrigation

water, indicating the potential of such a treatment to

impact microbial levels following harvest.

This study was designed to assess selected post-

harvest effects on aerobic and Enterobacteriaceae

microflora commonly found on fresh produce. Test-

ing for aerobic bacteria provided an indicator of the

impact of the post-harvest treatments on all bacteria

that grow in oxygenated conditions, including spoil-

age as well as pathogenic organisms. Evaluating

Enterobacteriaceae bacterial levels provided infor-

mation on how the post-harvest treatments affected

bacteria in this Gram-negative family, which includes

common food pathogens, such as Salmonella spp.,

Shigella spp., and enterohemorrhagic Escherichia

coli (Varnam and Evans, 1991).

According to current melon handling recommen-

dations (Fleming and Pool, 2005), cool temperature

storage is not required to maintain the safety of

whole melons. However, a study evaluating the ef-

fect of temperature on inoculated whole cantaloupe

melons found significantly more microbial growth


occurred at 19˚C than 4˚C (Annous et al., 2004). Most

whole melon storage temperature recommenda-

tions are based on preventing chilling injury among

commonly grown cultivars (Salunkhe and Kadam,

1998; Thompson, 2003).

There is a lack of research available on optimal

storage conditions for specialty melon cultivars such

as ‘Arava.’ This project evaluated the impact of stor-

age temperature and a CaCl2 dip treatment on total

aerobic and Enterobacteriaceae bacterial concen-

trations present on organically and conventionally

grown ‘Arava’ melons stored for up to 10 d.


Plant Material

Galia melons (cultivar ‘Arava’) (Johnny’s Selected

Seeds, Winslow, ME) were grown at the Colorado

State University Horticulture Field Research Center in

Fort Collins, CO during the summer of 2007. ‘Arava’

melons have a netted rind that is a grey-green color

while growing and turns light yellow when ripe. Fruit

flesh of ‘Arava’ melons is light green and the melons

require approximately 80-90 d to reach maturity.

Melons were grown simultaneously on organic

and conventional plots spaced 50 m apart. Soil at the

Horticulture Research Center is classified as Nunn

clay with a pH of 7.8 and the organic plots have been

USDA certified organic since 2001.

Prior to planting, soil tests were conducted on

both plots. Compared to the conventional plots, the

certified organic plots contained 2.0-2.4% higher

levels of organic matter derived from green manure

plough-down of legume and cereal cover crops and

from thoroughly composted chicken manure. The ni-

trogen, phosphorus, and potassium contents of the

two plots were made approximately equivalent us-

ing organic or conventional fertilizers prior to field

planting. For the organic plot, ‘Evergreen’ poultry

compost (A1 Organics, Eaton, CO) was applied with

a Millcreek spreader (Millcreek Manufacturing Co,

Lancaster County, PA) and rototilled into the soil. To

match nutrient levels in the organic fertilizer, urea

(45-0-0) and triple superphosphate (0-20.1-0) were

applied to the conventional plot using a broadcast

spreader.

Melon plants were started in the Colorado State

University Plant Environmental Research Center’s

greenhouses in 3-in. peat pots using Sunshine Or-

ganic Basic planting media (Sun Gro Horticulture,

Bellvue, WA) with 20% vermicompost (local source).

After four weeks, the melons were transplanted to

the field, spaced evenly in black plastic mulched

beds (rows 24 in. apart and beds 50 in. apart).

Crops were irrigated using drip irrigation with mu-

nicipal water. Irrigation levels were determined using

‘Watermark’ granular matrix sensors (Irrometer Com-

pany, Riverside, CA). Irrigation levels were monitored

to ensure the melons were watered adequately to

prevent plant moisture stress throughout the grow-

ing season.

During the growing season, pest management

practices were used to minimize cucumber beetle

(Acalymma vittatum) pressure on the melons. Syn-

thetic insecticide Permethrin (Loveland Products

Inc., Greeley, CO) was applied to the conventional

plots and naturally derived pyrethrum (MGK Co.,

Golden Valley, MN) was used on the organic plots.

Once the ‘Arava’ melons reached peak maturity (as

indicated by light yellow rind and nearly full slip off

the vine), they were harvested manually early in the

morning, then transported at ambient temperature

to the laboratory for processing within 30 min.

Treatments

Organically and conventionally grown melons (17

± 1-cm diam) were randomly assigned into CaCl2 dip

and non-dip (control) groups (total n=144). Any vis-

ible soil was brushed off melons using paper tow-

els. Half of all organically and conventionally grown

melons were dipped in a 0.08 M CaCl2 solution (8.8

g CaCl2 per L of water) and half were left untreated.

The CaCl2 concentration was chosen based on fa-

vorable results to increase shelf-life of fruit in other

studies (Garcia et al., 1996; Lester and Grusak, 2001;

Tsantili et al., 2002; Manganaris et al., 2007) as well as

preliminary research conducted in our lab.

For the dip, food grade CaCl2 (DOW Chemical


Company, Midland, MI) was mixed with tap water

(21 ± 1˚C) in 68 L plastic tubs (Sterile, Townsend,

MA) until dissolved. Each growing method/storage

time treatment set of melons (n=12) was completely

immersed in the CaCl2 solution for 20 min, then re-

moved, allowed to air dry on paper towels for 1 h and

randomly assigned to room or refrigerated storage.

Fresh dip solution was made for each set of dipped

melons. All melons and dipping solutions were kept

at ambient temperature (21± 1˚C) throughout the

dipping and drying steps.

Melons were then individually wrapped loosely

in tissue paper labeled with sample ID information

and placed into new 30.5×38.1×25.4-cm cardboard

boxes (Weyerhaeuser, Federal Way, WA), keeping

treatment groups separate. Melons were stored at

21± 1˚C (relative humidity 30 ± 5%) or 10 ± 1˚C (rela-

tive humidity 70 ± 5%). At d 1, 5, and 10, total aerobic

and Enterobacteriaceae counts were determined,

with six melons evaluated individually per treatment

group as test replications.

Microbial Analysis

Microbial testing was based on methods used

by the United States Department of Agriculture’s

(USDA) Microbiological Data Program (MDP, 2002-

2006; MDP, 2003-2007) and modified for our lab (K.

McCallum (Colorado Department of Agriculture,

Denver, CO, personal communication)).

Each whole melon was placed in a 38.1×50.8-

cm sterile bag (VWR, West Chester, PA) to which

300 mL of a solution of Universal Pre-Enrichment

Broth (UPEB) (Difco, Sparks, MD) and 0.1% Tween

80 (Fisher Scientific, Fair Lawn, NJ) were added. The

bags were sealed with a twist tie. The bagged mel-

ons were shaken for 20 up and down strokes and 20

side to side strokes to assure the UPEB solution ad-

equately “washed” all surfaces of the melon, then

stored with the melon remaining in the bag at 5˚C for

18-24 h. A 5 mL sample was taken from the bagged

UPEB solution and transferred to a sterile Falcon

tube (BD Falcon™, Franklin Lakes, NJ). Six, 10-fold

serial dilutions were made using buffered peptone

water (Difco, Sparks, MD). From each dilution, 1 mL

samples were plated on aerobic and Enterobacte-

riaceae Petrifilm™ (3M, St. Paul, MN) according to

manufacturer’s instructions. All Petrifilm™ plates

were incubated overnight (37˚C). Colonies for each

sample were counted on plates that contained 25-

250 colonies and expressed as CFU/mL.

Data Analysis

Results were transformed into log scale and ana-

lyzed using SAS Proc Mixed (Version 9.1, Cary, NC).

A factorial analysis of variance was performed with

differences between means assessed using a signifi-

cance of p<0.05 with the Tukey-Kramer adjustment

for multiple comparisons. Fixed effects included

time, temperature, dip, and growing method; repli-

cation was included as a random effect.

RESULTS

Aerobic Counts

The mean aerobic bacterial counts recovered from

the melon rinds ranged from 4.56 to 8.28 log CFU/

mL. Storage time did not impact aerobic growth

(p>0.05), but storage temperature, dip treatment,

and growing method did have significant overall af-

fects on aerobic bacterial levels (p<0.05). Since over-

all storage time was not significant, results from d 1,

5, and 10 were combined (Figure 1).

Overall, aerobic counts were lower (p<0.01) on

CaCl2 dipped melons compared to non-dipped con-

trol melons (6.16 vs. 6.54 log CFU/mL, respectively).

Also, melons stored at 10˚C had overall lower bac-

terial growth (p<0.0001) than those stored at 21˚C

(5.98 vs. 6.72 log CFU/mL, respectively). Within each

temperature/growing method combination, treat-

ment with CaCl2 did not significantly impact aero-

bic counts, but CaCl2 dipped melons stored at 10˚C

had the lowest bacterial counts (p<0.05), regardless

of growing method. Growing method impacted the

presence of aerobic bacteria (p<0.05), with organi-

cally grown melons having higher overall mean aero-

bic counts than conventional melons (6.49 vs. 6.21

log CFU/mL, respectively). Much of this difference


was due to organic control (no dip) melons stored

at 21˚C having the highest aerobic counts (Figure 1).

Enterobacteriaceae Counts

The Enterobacteriaceae bacterial counts ranged

from 4.19 to 7.15 log CFU/mL, and were significantly

affected overall by temperature and dip treatment

(p<0.05). Again, storage time was not significant

(p>0.05), so results from d 1, 5, and 10 were com-

bined (Figure 1).

Like the aerobic counts, melons stored at 10˚C

had lower (p<0.0001) Enterobacteriaceae counts

than melons stored at 21˚C (5.20 vs. 6.06 log CFU/

mL, respectively). Also, melons dipped in CaCl2 had

lower (p<0.001) overall Enterobacteriaceae counts

compared to non-dipped control melons (5.38 vs.

5.87 log CFU/mL, respectively). The combination

of the CaCl2 dip and 10˚C storage temperature also

produced the lowest Enterbacteriaceae counts (Fig-

ure 1).

Growing method did not impact Enterobacteria-

ceae counts overall (p>0.05). When looking at the

results by dip treatment and storage temperature

(Figure 1), organic and conventional melons did

not follow a similar pattern. Organic CaCl2-dipped

Aerobic10 ˚C 10 ˚C 21 ˚C 21 ˚COrganic Conventional Organic Conventional

CaCl2 Dip 5.67 5.64 6.69 6.65No Dip 6.45 6.16 7.15 6.41

SE=0.19

10 ˚C 10 ˚C 21 ˚C 21 ˚CCaCl2 Dip No Dip CaCl2 Dip No Dip

Organic 5.67 6.45 6.69 7.15Conventional 5.64 6.16 6.65 6.41

Enterobac10 ˚C 10 ˚C 21 ˚C 21 ˚COrganic Conventional Organic Conventional

CaCl2 Dip 4.74 5.07 5.56 6.18No Dip 5.87 5.11 6.50 6.00


Organic 4.74 5.87 5.56 6.5Conventional 5.07 5.11 6.18 6

012345678

Organic Conventional Organic Conventional

10 ˚C 10 ˚C 21 ˚C 21 ˚C

Aerobic Results

Log

CFU

/mL

CaCl2DipNo Dip

012345678


10 ˚C 10 ˚C 21 ˚C 21 ˚C

Enterobacteriaceae Results

Log

CFU

/mL

CaCl2DipNo Dip

Aerobic10 ˚C 10 ˚C 21 ˚C 21 ˚COrganic Conventional Organic Conventional

CaCl2 Dip 5.67 5.64 6.69 6.65No Dip 6.45 6.16 7.15 6.41

SE=0.19


Organic 5.67 6.45 6.69 7.15Conventional 5.64 6.16 6.65 6.41

Enterobac10 ˚C 10 ˚C 21 ˚C 21 ˚COrganic Conventional Organic Conventional

CaCl2 Dip 4.74 5.07 5.56 6.18No Dip 5.87 5.11 6.50 6.00


Organic 4.74 5.87 5.56 6.5Conventional 5.07 5.11 6.18 6

012345678


10 ˚C 10 ˚C 21 ˚C 21 ˚C

Aerobic Results

Log

CFU

/mL

CaCl2DipNo Dip

012345678


10 ˚C 10 ˚C 21 ˚C 21 ˚C

Enterobacteriaceae Results

Log

CFU

/mL

CaCl2DipNo Dip

Figure 1. Aerobic and Enterobacteriaceae bacterial counts (log CFU/mL) of organically and con-ventionally grown ‘Arava’ melons. Melons were immersed in a 0.08 M CaCl2 solution for 20 minutes or not treated then stored at 10 or 21˚C. Each bar represents the mean bacterial counts of organic or conventional melons by storage treatment (n=18 per bar). Error bars indicate standard error. Within each graph, bars with similar lowercase letters are not significantly different (p<0.05).

c

d

cabc

abc cd cdab aba

bcd

bc ab ab abca


melons had significantly lower (p<0.05) Enterobacte-

riaceae counts than non-dipped control melons at a

given temperature, whereas CaCl2 treatment did not

impact conventionally grown melons. In contrast,

conventionally-grown melons were significantly im-

pacted (p<0.05) by storage temperature, with lower

Enterobacteriaceae counts on melons stored at 10˚C

than at 21˚C. Enterobacteriaceae counts on the or-

ganic melons were not impacted by storage tem-

perature (Figure 1).

DISCUSSION

Using CaCl2 dips on fresh whole fruit has been

shown to be an effective treatment for prolonging

the shelf-life as well as improving quality character-

istics during storage compared to fruit dipped in a

water control (Garcia et al., 1996; Lester and Grusak,

2001; Tsantili et al., 2002; Manganaris et al., 2007)

or left dry (Garcia et al., 1996; Lysiak et al., 2008).

CaCl2 must be dissolved in water before being ap-

plied to fresh fruit; therefore, it is not possible to

use a CaCl2 application without submerging fruit in

a water-based dip. For the control in this study, we

chose to use non-dipped melons, since that is how

small-scale farmers in the Rocky Mountain region

currently handle their fresh specialty melons (F. Ston-

aker (Specialty Crops Program Coordinator, Colora-

do State University Department of Horticulture and

Landscape Architecture, personal communication)).

CaCl2 dip treatments have been found to have many

benefits compared to fruit left untreated or dipped

in a water control (Lester and Grusak, 2001; Manga-

naris et al., 2007; Martin-Diana et al., 2007; Lysiak et

al., 2008), so we wanted to compare the use of such

a dip to the existing post-harvest handling meth-

ods used by small-scale farmers. The results of this

study warrant further research with additional pro-

duce, with additional controls, and with inoculated

samples to establish the role CaCl2 dips may have in

minimizing microbial growth.

The CaCl2 dip in this study appears to be a prom-

ising option for decreasing the bacterial counts on

the surface of whole ‘Arava’ melons, especially when

the melons are grown organically and stored either

at room temperature, which is a common practice

among consumers, or at cooler postharvest storage

temperatures. Since the Enterobacteriaceae family

contains many common foodborne pathogens such

as Shigella spp., Salmonella spp. and Enterohemor-

rhagic E. coli (Varnam and Evans, 1991), CaCl2 may

potentially increase the safety of organic melons

during storage. CaCl2 is approved as a processing

aid for organic food production (Code of Federal

Regulations, 2008), is relatively inexpensive, and eas-

ily accessible (Dow-Chemical, 2007), making it a fea-

sible solution for farmers.

The USDA has monitored for the presence of sev-

eral foodborne pathogens on cantaloupe as well as

alfalfa sprouts, pre-cut bagged lettuce, spinach, and

tomatoes since 2002 as part of the Microbiologi-

cal Data Program (MDP). Among the thousands of

samples tested for this program between 2002-2006,

only 0.17% of cantaloupe samples tested were posi-

tive for Salmonella spp. and 0.24% for pathogenic

E. coli (K. McCallum (Colorado Department of Ag-

riculture, Denver, CO, personal communication)

(MDP, 2002-2006). Though this shows pathogens can

be found on cantaloupe, due to the very low rates

of contamination, it would be difficult to determine

post-harvest treatment effects on specific foodborne

pathogens without inoculating the produce.

The effect of storage temperature on microbial

growth seen in this study is consistent with the results

from other studies showing lower bacterial growth

at lower temperatures for inoculated whole melons

(Annous et al., 2004), as well as for fresh-cut produce

(Zagory, 1999; Francis and O’Beirne, 2001). Storing

melons at cooler temperatures may be an effective

solution for limiting post-harvest microbial growth.

Yet, many melons are susceptible to chilling injury

when storage temperatures are decreased, and the

specific temperature causing chilling injury symp-

toms varies greatly by cultivar (Miccolis and Saltveit,

1995). Most storage recommendations have been

based on commonly-grown commercial cantaloupe

or honeydew cultivars (Salunkhe and Kadam, 1998;

Thompson, 2003). We did not test for chilling injury

in this study, but it would be useful to determine the

effect of lower storage temperatures on quality and


sensory characteristics of specialty cultivars such as

‘Arava.’

In this study, organic melons were observed to

have higher levels of aerobic bacteria compared to

conventionally grown melons, primarily due to signif-

icantly higher counts on non-dipped control organic

melons stored at 21˚C. This indicates organic ‘Arava’

melons may be more conducive than conventionally

grown melons to higher levels of microbial growth

after storage at room temperature for several days.

This is a potential concern because growers as well

as consumers may store melons at room tempera-

ture before the fruit is consumed. Enterobacteria-

ceae counts were not significantly different between

the two growing methods, yet the growing method

affected how the melons responded to dip treat-

ment and storage temperature in this study.

Other studies comparing organic and conven-

tional produce safety have used produce from differ-

ent farms (Magkos et al., 2006), which would greatly

confound the results. Based on our research, it ap-

pears that growing method may impact the general

microflora of ‘Arava’ melons when stored at room

temperature as well as impacting possible methods

to minimize Enterobacteriaceae bacteria. Additional

well-controlled research is needed to determine the

growing method effect on other produce as well as

on specific microorganisms.

Overall, time (up to 10 d) was not a significant fac-

tor impacting microbial growth in this study. This is

encouraging as melons are often stored by growers

or consumers for several days before consumption.

However, other variables, such as temperature and

growing method may eventually allow microbial lev-

els to significantly increase over time.

The sensory impact of using a CaCl2 treatment

should also be addressed. Research on whole mel-

ons dipped in a 0.08 M CaCl2 solution indicate there

may be no negative sensory effects (Lester and Gru-

sak, 2001), yet another study found higher bitterness

and lower melon flavor scores for fresh cut melons

dipped in 1% and 2.5% CaCl2 solutions (approxi-

mately 0.09 and 0.23 M, respectively) compared to

the control and other treatments (Luna-Guzman and

Barrett, 2000). Since the latter study used fresh cut

melons the results may not be the same for whole

melons, as the physiology of cut melons has been

shown to be distinctly different than that of whole

fruit (Lamikanra et al., 2003). Further assessment of

the value of dipping whole melons in CaCl2 as well as

sensory effects of CaCl2 on different melon cultivars

is important to assure consumer acceptance.

CONCLUSIONS

Based on the results of this study, the common

practice of storing melons at room temperature may

pose a greater microbial risk for melons grown using

organic compared to conventional production meth-

ods. This risk may be minimized through the use of a

CaCl2 dip at harvest. Regardless of growing method,

storing whole ‘Arava’ melons at cooler temperatures

is an effective method for slowing bacteria growth.

Additional research should be conducted to explore

the potential for a CaCl2 dip to decrease specific

foodborne pathogens on melons and other produce

as well as treatment and storage effects on sensory

and quality characteristics.

ACKNOWLEDGEMENTS

This project was supported by the National Re-

search Initiative of the USDA Cooperative State

Research, Education and Extension Service, grant

number 2005-55618-15634. The assistance of Frank

Stonaker and the horticultural crew at the CSU Horti-

culture Field Research Center in growing the melons

is gratefully acknowledged.

REFERENCES

Annous, B. A., A. Burke, and J. E. Sites. 2004. Surface

pasteurization of whole fresh cantaloupes inocu-

lated with Salmonella poona or Escherichia coli. J.

Food Prot. 67:1876-1885.

Bowen, A., A. Fry, G. Richards, and L. Beuchat. 2006.

Infections associated with cantaloupe consump-

tion: a public health concern. Epidemiol. Infect.

134:675-685.

Chikthimmah N., L. F. LaBorde, and R. B. Beelman.


2005. Hydrogen peroxide and calcium chloride

added to irrigation water as a strategy to reduce

bacterial populations and improve quality of fresh

mushrooms. J. Food Sci. 70:M273-M278.

Code of Federal Regulations. 2008. Available from:

http://ecfr.gpoaccess.gov/cgi/t/text/text-idx?c=e

cfr&sid=4410780cb6a0c74813c0c51973cca3ff&rgn

=div5&view=text&node=7:3.1.1.9.31&idno=7. Ac-

cessed May, 2008.

Dow-Chemical. 2007. Calcium Chloride Product in-

formation. Available from: http://www.dow.com/

calcium/. Accessed June, 2007.

Doyle, M .P., and M. C. Erickson. 2008. Summer

meeting 2007 - the problems with fresh produce:

an overview. J. Appl. Microbiol. 105:317-330.

Fleming ,P., and B. Pool. 2005. Commodity Spe-

cific Food Safety Guidelines for the Melon Sup-

ply Chain. Available from: http://www.fda.gov/

downloads/Food/FoodSafety/Product-SpecificIn-

formation/FruitsVegetablesJuices/GuidanceCom-

plianceRegulatoryInformation/UCM168625.pdf.

Accessed April, 2011.

Francis, G. A., and D. O’Beirne. 2001. Effects of veg-

etable type, package atmosphere and storage

temperature on growth and survival of Escherichia

coli O157:H7 and Listeria monocytogenes. J. Ind.

Microbiol. Biot. 27:111-116.

Garcia, J. M., S. Herrera, and A. Morilla. 1996. Effects

of postharvest dips in calcium chloride on straw-

berry. J. Agr. Food Chem. 44:30-33.

Harker, F. R. 2004. Organic food claims cannot be

substantiated through testing of samples inter-

cepted in the marketplace: a horticulturalist’s opin-

ion. Food Qual. Prefer. 15:91-95.

Johnston, L. M., L. A. Jaykus, D. Moll, M. C. Martinez,

J. Anciso, B. Mora, and C. L. Moe. 2005. A field

study of the microbiological quality of fresh pro-

duce. J. Food Prot. 68:1840-1847.

Lamikanra, O., B. Juaraez, M. A. Watson, and O .A.

Richard. 2003. Effect of cutting and storage on

sensory traits of cantaloupe melon cultivars with

extended postharvest shelf life. J. Sci. Food Agric.

83:702-708.

Lester, G. E. 2006. Organic versus conventionally

grown produce: Quality differences, and guide-

lines for comparison studies. HortScience. 41:296-

300.

Lester, G. E., and M. A. Grusak. 2001. Postharvest

application of chelated and nonchelated calcium

dip treatments to commercially grown honeydew

melons: Effects on peel attributes, tissue calcium

concentration, quality, and consumer preference

following storage. HortTechnology. 11:561-566.

Luna-Guzman, I., and D. M. Barrett. 2000. Compari-

son of calcium chloride and calcium lactate effec-

tiveness in maintaining shelf stability and quality

of fresh-cut cantaloupes. Postharvest Biol. Tech.

19:61-72.

Lysiak, G., W. J. Florkowski, and S. E. Prussia. 2008.

Postharvest calcium chloride application and mois-

ture barrier influence on peach fruit quality. Hort-

Technology. 18:100-105.

Magkos, F., F. Arvaniti, and A. Zampelas. 2006. Or-

ganic food: Buying more safety or just peace of

mind? A critical review of the literature. Crit. Rev.

Food Sci. Nutr. 46:23-56.

Magnusson, M. K., A. Arvola, U. K. K. Hursti, L. Aberg,

and P. O. Sjoden. 2003. Choice of organic foods

is related to perceived consequences for human

health and to environmentally friendly behaviour.

Appetite. 40:109-117.

Manganaris, G. A., M. Vasilakakis, G. Diamantidis,

and I. Mignani. 2007. The effect of postharvest cal-

cium application on tissue calcium concentration,

quality attributes, incidence of flesh browning and

cell wall physicochemical aspects of peach fruits.

Food Chem. 100:1385-1392.

Martin-Diana, A. B., D. Rico, J. M. Frias, J. M. Barat,

G. T. M. Henehan, and C. Barry-Ryan. 2007. Cal-

cium for extending the shelf life of fresh whole and

minimally processed fruits and vegetables: a re-

view. Trends Food Sci. Tech. 18:210-218.

MDP. 2002-2006. Microbiological Data Program,

Program Reports. Available from: http://www.ams.

usda.gov/science/MPO/Download.htm. Accessed

November, 2007.

MDP. 2003-2007. Microbiological Data Program

Standard Operating Procedures. Available from:

http://www.ams.usda.gov/science/MPO/SOPs.

htm. Accessed November, 2007.


Miccolis, V., and M. E. Saltveit. 1995. Influence of

storage period and temperature on the posthar-

vest characteristics of 6 melon (Cucumis melo L.,

Inodorus Group) cultivars. Postharvest Biol. Tech.

5:211-219.

Poovaiah, B.W. 1986. Role of calcium in prolonging

storage life of fruits and vegetables. Food Tech.

40:86-89.

Salunkhe, D. K., and S. S. Kadam. 1998. Handbook

of Vegetable Science and Technology: Production,

Composition, Storage, and Processing. Marcel

Dekker, New York, NY.

Shepherd, R., M. Magnusson, and P. O. Sjoden.

2005. Determinants of consumer behavior related

to organic foods. Ambio. 34:352-359.

Sivapalasingam, S., C. R. Friedman, L. Cohen, and

R. V. Tauxe. 2004. Fresh produce: A growing cause

of outbreaks of foodborne illness in the United

States, 1973 through 1997. J. Food Prot. 67:2342-

2353.

Stephenson, J. 1997a. Public health experts take aim

at a moving target: Foodborne infections. JAMA.

277:97-98.

Stephenson, J. 1997b. New approaches for detect-

ing and curtailing foodborne microbial infections.

JAMA. 277:1337-1339.

Thompson, A. K. 2003. Fruit and Vegetables: Har-

vesting, Handling, and Storage. Blackwell Pub.

Co., Ames, IA.

Torjusen, H., G. Lieblein, M. Wandel, and C. A. Fran-

cis. 2001. Food system orientation and quality

perception among consumers and producers of

organic food in Hedmark County, Norway. Food

Qual. Prefer. 12:207-216.

Tsantili, E., K. Konstantinidis, P. E. Athanasopoulos,

and C. Pontikis. 2002. Effects of postharvest cal-

cium treatments on respiration and quality attri-

butes in lemon fruit during storage. J. Hortic. Sci.

Biotech. 77:479-484.

Varnam, A. H., and M. G. Evans. 1991. Foodborne

Pathogens: An Illustrated Text. Wolfe Pub., Lon-

don.

Winter, C. K., and S. F. Davis. 2006. Organic foods. J.

Food Sci. 71:R117-R124.

Yiridoe, E .K., S. Bonti-Ankomah, and R. C. Martin.

2005. Comparison of consumer perceptions and

preference toward organic versus conventionally

produced foods: A review and update of the litera-

ture. Renew. Agr. Food Syst. 20:193-205.

Zagory, D. 1999. Effects of post-processing handling

and packaging on microbial populations. Posthar-

vest Biol. Tech. 15:313-321.




ABSTRACT

Listeria monocytogenes is the one of most lethal of all food pathogens. It has a high fatality rate among

immune compromised individuals and has been shown to contaminate ready-to-eat (RTE) deli meats,

which have been linked to several outbreaks of listeriosis. Unfortunately, the incidence of listeriosis has not

decreased significantly since 2001 and the 2010 Healthy People goal of 2.4 cases of listeriosis per million

consumers has not been met. More than 8 out of 10 of the listeriosis cases linked to delis were attributed to

RTE luncheon meats sliced in retail stores, which has led risk assessors to suggest that cross-contamination

from the retail deli environment is likely responsible for the stubborn resistance in reducing listeriosis.

Research has also shown that most food borne illnesses result from food handler error, which may be

minimized when employees are properly trained and transfer their training to their jobs. There is a need

for training that is specifically focused on the deli employee which could have a measurable impact in de-

creasing the risk of L. monocytogenes cross-contamination. Proper motivation and training of employees

and managers is vital to keep consumers safe. Thus, there is a crucial need to fill gaps in the knowledge

base for designing effective training for newly hired and hard-to-reach employees in a retail food service

environment.

Keywords: Worker training, delicatessen, Listeria monocytogenes

INTRODUCTION

Every year, food borne illnesses devastate the lives

of millions of Americans and take a significant eco-

nomic toll, considering the costs of medical treat-

ment and lost wages in addition to the costs of food

Correspondence: Philip G. Crandall, [email protected]: +1 -479-575-7686 Fax: +1-479-575-6936

recalls much of which is paid for by the retailer and

manufacturer. Scharff (2010) estimated the health

care burden for the U.S. economy to be $152 billion

for just the costs of the acute illness and a few of

the long-term health-related costs, certainly not the

“bottom-line” for the total cost to the U.S. economy.

Scharff (2010) continued separating out these costs

for individual pathogens. Listeria monocytogenes

has the highest cost to the long-term “quality of life”

REVIEWMinimizing the Risk of Listeria monocytogenes in Retail Delis by

Developing Employee Focused, Cost Effective TrainingP. G. Crandall1*, J. A. Neal Jr.2, C. A. O’Bryan1, C. A. Murphy3, B. P. Marks4 and S. C. Ricke1

1Address: Department of Food Science and Center for Food Safety, University of Arkansas, Fayetteville, AR 727042College of Hotel and Restaurant Management, University of Houston, Houston, TX 772043Department of Curriculum and Instruction, University of Arkansas, Fayetteville, AR 72704

4Department of Biosystems and Agricultural Engineering, Michigan State University, East Lansing, MI 48824



and is second only to Vibrio for the highest cost to

treat an individual case. L. monocytogenes is 3rd be-

hind Campylobacter and Salmonella, both of which

rarely result in mortality, in the total costs for treat-

ments from a single food borne pathogen.

One reason that L. monocytogenes accounts for

such a disproportionately large portion of this cost

burden is because when invasive listeriosis invades

the brain or spinal cord, the fatality rates can ap-

proach 30%. Survivors may be left with neurologi-

cal impairments requiring lifelong, constant care.

Pregnant women account for approximately 30%

of the cases of listeriosis, resulting in miscarriages,

still- or premature births, and 20% of the surviving

infants require lifelong care (Roberts and Pinner,

1990). Despite a wholesale commitment on the part

of the food industry and Federal regulators to mini-

mize this problem, the incidence of listeriosis has

not decreased significantly since 2001. The current

estimate of listeriosis in the U.S. is 2.7 cases/mil-

lion persons, which is considerably higher than the

2010 Healthy People goal of 2.4 cases/million (CDC,

2009). Kause (2009) has stated that although FSIS

in-plant testing programs have reported a dramatic

decline in L. monocytogenes from deli ready-to-eat

(RTE) meat and poultry products from 1990 to 2008,

a corresponding decline in cases of listeriosis has not

occurred. Furthermore, 83% of the listeriosis cases

attributed to delis were attributed to RTE luncheon

meats sliced at retail deli stores, which led Kause

(2009) to suggest that cross-contamination from the

retail environment is likely responsible for the pla-

teau in reducing listeriosis in the U.S.

Therefore, training that is specifically designed for

deli managers and employees could have a measur-

able impact in decreasing the risk of L. monocyto-

genes cross-contamination. Cost-effective training

can be accomplished through cooperative training

efforts involving the industry, regulators and univer-

sity researchers and educators. Cleaning a cutting

board or sanitizing a floor drain may seem to be mun-

dane tasks, but they are critically important. Proper

motivation and training of employees and managers

is vital to keep consumers safe (Bricher, 2007). How-

ever, two issues must be addressed to make prog-

ress in reducing L. monocytogenes cross-contami-

nation in the retail deli. First, all currently available

educational approaches need to be critically evalu-

ated, including: 1) determining the strengths, weak-

nesses, knowledge content and usability of currently

available Computer Based Trainings (CBT) for deli

managers and hourly employees and 2) adapting

the traditional face-to-face deli employee trainings

to include newly developed techniques to minimize

L. monocytogenes cross-contamination. Employers

are required by he latest edition of the FDA Model

Food Code (FDA, 2009) that serves as the basis of

most health regulations for delis, to train and keep

written documentation of training for employees

handling food.

Secondly, there is a need to improve L. monocyto-

genes sanitation procedures by 1) filling knowledge

gaps in current quantitative risk assessment for RTE

foods by developing transfer coefficients for L. mono-

cytogenes cross-contamination that can result from

the of behavior of employees working improperly in

potentially L. monocytogenes contaminated deli en-

vironments, 2) using a visual indicator for meat and

fat residues to verify the cleaning of slicers and food

contact surfaces prior to sanitation, 3) introducing

a lethal kill step capable of reducing L. monocyto-

genes contamination by 5 logs and 4) modeling the

risk reduction by verifying comprehension and per-

manent changes in employee behavior. Thirdly, new

sanitation techniques need to be incorporated into

the training and documentation must be made for

changes in behavior of new and hard-to-reach em-

ployees, who may not have been effectively trained

using previous interventions proposed to reduce L.

monocytogenes cross contamination. Achieving the

twin goals of cost-effective education of deli employ-

ees and improved sanitation procedures will require

a coordinated research and education effort with the

prospect of a verifiable decrease in the risk from L.

monocytogenes cross-contamination in retail delis.

Regulations

Retail delis are unique in that they may have to

be knowledgeable of and compliant with regulatory


statutes from multiple and sometimes confusing ju-

risdictions including: three levels of health depart-

ments--county, municipal, and state health depart-

ments, and requirements set out by the FDA Model

Food Code. New regulations targeting reduced risk

from L. monocytogenes in RTE foods must be based

on quantitative risk assessment models. Currently,

a significant number of assumptions have had to be

incorporated into these risk models, because there

is a lack of research data on L. monocytogenes trans-

fer rates from contaminated surfaces to foods and

attachment and cross-contamination from food con-

tact surfaces. Additionally, the impact of practices

and behaviors of retail deli employees must also be

targeted to collect actual data to replace risk asses-

sors’ assumptions. This could be termed the “work-

er effect” for risk assessment models.

This review will assess the characteristics,

strengths, weaknesses, completeness of information

and the level of knowledge content, style of presen-

tation and usability of current retail food safety train-

ing platforms for deli managers and hourly employ-

ees. As part of employees’ “hands-on” training, the

use of fluorescent compounds has been shown to

be an effective indicator of potential behaviors that

can lead to cross-contamination. The use of these

fluorescent compounds can be extended to esti-

mate the rates of L. monocytogenes cross-contam-

ination, document the effectiveness of new sanitiza-

tion procedures and identify potential opportunities

for modeling supplemental risk assessment data by

adding the “worker effect” component to the vari-

ability to the risk transfer model. Finally, new online

training modules on sanitation for deli employees

and current research that focuses on best practices

to reduce or eliminate L. monocytogenes will be dis-

cussed, along with the impact of deli specific train-

ing targeting new employees and the potential for

long-term changes in behavior.

L. MONOCYTOGENES AND THE DELI

First, we will discuss a systematic approach to le-

verage newly published information on the locations

of persistent strains of L. monocytogenes isolated

from retail delis (Chen et al., 2001; Sauders et al.,

2004; Sauders et al., 2009). We will then discuss the

requirements for training retail deli managers and

employees. In a recent survey of the prevalence of

L. monocytogenes in RTE brands of sliced luncheon

meats in Europe, significant differences were found

between deli meat that was sliced and packaged in

a retail deli (8.5% positive) compared to commer-

cially manufactured, sliced and packaged product

that had only about 1/3 the incidence, 2.7% positive

(Garrido et al., 2009). These researchers went on to

publish pulsed-field gel electrophoresis (PFGE) that

revealed the same L. monocytogenes pulse-type

was repeatedly recovered from the same sliced RTE

products from the same retail store. While these

pulse-types were not recovered from patients with

listeriosis, it is compelling evidence for the need to

improve worker hygiene and cleaning/disinfection

programs in conjunction with more effective clean-

Listeria monocytogenesfrom a variety of sources: cutting boards, shoes, gloves, floors, drains, aerosols, coolers, mats, etc.

L. monocytogenes levels in final deli product.

Source of Listeria Contamination

Effectiveness of sanitation

Worker effect

Transfer coefficients for L. monocytogenes

Cross-contamination potential factors

Figure 1. Factors influencing Listeria monocytogenes Contamination of Deli Products


ing practices and constant food worker education

for retail delis.

L. monocytogenes cross-contamination sources

are quite complex and can originate from a variety

of sources (Figure 1). Cross-contamination can occur

from deli meats to commercial cutting boards, in-

cluding wood or high density polyethylene (HDPE),

to 304 grade stainless steel and to latex, vinyl, or

polyethylene in gloves. Consequently there is a

need to quantify transfer coefficients to or from soles

of shoes, gloves and all types of equipment and en-

vironmental surfaces, including stainless steel, ce-

ment, ceramic tile and deli flooring material. As an

example of factors influencing sequential-contact

transfer rates, Campos et al. (2009) demonstrated

that the accumulation of food product residue on a

slicer blade influenced L. monocytogenes transfer

rates to and from deli meat products. In general, the

data consistently exhibited a biphasic tendency, with

a distinctive log-linear phase (less than 10 contacts),

followed by a long tailing phase. The duration of

first log-linear phase was product-specific, but cor-

responded roughly to the point where accumulated

product residue exceeded approximately 5 microns

in thickness on the blade. This would be consistent

with a hypothesis that the mechanisms of bacterial

transfer between a food product and equipment sur-

face are significantly different when there is a food

barrier thicker than the cells being transferred. This

is an example of why any model forms generated for

L. monocytogenes should be based on a phenom-

enological framework for the transfer process being

modeled. Phenomenological evidence is based on

observations or the results of experiments rather

than derived theory basis and the boundaries may

be referred to as “fuzzy.”

Candidate transfer models to quantify the num-

ber of bacteria transferred include: (a) log-linear, (b)

Weibull-type, and (c) a biphasic-like model, illustrat-

ed here with the example of a two-part, first-order

kinetic analogy:

where N0 is the number of bacteria on the donor sur-

face prior to a contact event, Nc is the number of

bacteria on the donor surface after the contact event,

and ki and n are model parameters. Those param-

eters can be estimated by using non-linear regres-

sion to minimize the sum of squared errors between

the experimental data (for a given transfer scenario)

and the model-predicted counts. Although all three

candidate models are essentially empirical, the third

is based on prior evidence of the impact of multiple,

sequential contact events on the transfer rate, due

to concurrent transfer of food product residue on

contact. Given that each of the candidate models

has a different number of parameters, the best mod-

el should be selected utilizing both qualitative and

quantitative criteria, such as Akaike’s Information

Criterion (AIC) (Motulsky and Christopoulos, 2004).

However, before models can be verified, biological

data must be generated to provide input on trans-

fer coefficients. The following section addresses the

use of indicators for incorporation into food systems.

Fluorescent indicators for estimating cross-contamination

The current state-of-the-art in modeling of sur-

face-to-surface bacterial transfer in food processing

environments is still essentially empirical. Because

of commercial concerns of contaminating food pro-

cessing plants with a non-pathogenic surrogate, such

as Listeria innocua, it would be ideal to use a model

fluorescent indicator as a surrogate for L. monocyto-

genes when assessing delicatessen-type transfer of L.

monocytogenes and cross-contamination. In order to

test the suitability of a fluorescent indicator surrogate

for any system such as a simulated deli, it is not essen-

tial that the model parameters be identical between

the Listeria and fluorescent indicator results. Rather,

it is sufficient if a replicable, functional relationship

can be identified between the L. monocytogenes and

the fluorescence indicator even in a nonlinear rela-

tionship, as long as the functional form is consistent

across the tested cross-contamination transfer sce-

narios, in order to ensure robustness. The fluorescent

indicators can be quantified fluorometrically using the

(a) (b) (c) 1


maximum excitation wavelength, then measuring the

responding fluorescent wavelength compared to a

standard curve for this fluorescent compound. Ide-

ally, these measurements would yield a series of trans-

fer coefficients relating the amount of fluorescence to

the number of L. monocytogenes cells that would be

transferred under identical conditions.

A cocktail of L. monocytogenes strains isolated

from retail delis together with a fluorescent indicator

would allow for simultaneous assessment of the quan-

titative transfer of both L. monocytogenes and fluo-

rescent indicator from deli meats (turkey, ham, salami)

to and from a deli slicer. For example, a retail chub

of delicatessen turkey, ham and salami can simulate

surface contamination following thermal processing

by the manufacturer by directly dip-inoculating in the

L. monocytogenes cocktail containing the fluorescent

indicator and allowing it to air-dry. Using the fluo-

rescent indicator Glo-Germ®, it has been possible

to identify 5 product contact surfaces on deli slicers

that were cross-contaminated following the slicing

of fluorescent indicator -contaminated cooked, RTE

turkey chubs (Vorst et al., 2006). This work served as

the basis for the ability to quantify surface-to-product

and product-to-surface transfer of L. monocytogenes

during slicing of various deli meats (Keskinen et al.,

2008; Vorst et al., 2006). More recently, this same

research group examined quantitative transfer of E.

coli O157:H7 during processing of leafy greens, using

Glo-Germ-inoculated head lettuce to identify 23 key

product contact surfaces on a commercial-scale flume

tank, shaker table, and dewatering centrifuge (Buch-

holz et al., 2008) for subsequent sampling (Buchholz

et al. 2009a,b).

Significantly more data must be collected before

practical working applications can be created, en-

compassing a wider range of contact scenarios to

build the foundation for phenomenological models,

which then can be applied across multiple pathogens

and multiple foods. New Standard Sanitary Oper-

ating Procedures (SSOPs) will need to take into ac-

count the recent FSIS 2009 Listeria Risk Assessment

and findings, which identify critical cleaning points to

be targeted in a retail deli for deli employee specific

training (Chen et al., 2001; Sauders et al., 2004, 2009).

Modeling the “worker-effect” on the variability in bacterial transfer events

The recent draft FSIS risk assessment for L. mono-

cytogenes in deli meats pointed to the fact that

there is need for better assessment of how contami-

nation of deli meats at retail occurs (FSIS, 2009). In

addition, FDA has assessed foodborne illness risk

factors in foodservice operations and the main risk

factors determined are listed in Table 1 (FDA, 2010).

Deli employee data generated, in a sense of experi-

mental simulations, would provide extremely novel

evidence of the interaction among transfer events in

this environment. Those data could subsequently be

utilized to model the “worker effect” on the overall

transfer outcome and the variability in that outcome.

This type of analysis typically follows the overall ap-

proach (Chen et al., 2001; Schaffner, 2004) to quantify

variability in cross-contamination rates among the

components of the system (i.e., the food product,

the environmental surfaces, and the worker) and in

quantifying the impact of the worker (and the impact

of the proposed training interventions on that fac-

tor). Frequency distributions of aggregate transfer

can be generated for a given contaminated food

contact surface and non-contaminated surfaces in

the deli environment, and the distributions for each

test case (pre- and post-transfer) can be statistically

compared to test for impact.

There have been various studies on worker com-

pliance with practices that reduce foodborne illness

risk. Lynch et al. (2005), observed gloves were used

by 46% of workers in fast food restaurants. They also

Table 1. Risk factors for foodborne illness at foodservice establishments

Risk factorsContaminated equipment

Food from unsafe source

Improper holding time/temperature

Inadequate cooking

Poor personal hygiene

Hand washing

Glove use


observed that 93% of workers in states that prohibit-

ed bare hand contact of food used gloves, while just

less than 5% used gloves in states that did not pro-

hibit bare hand contact (Lynch et al., 2005). Lubran

et al (2010) in an observational study did not observe

bare-hand contact with RTE foods at any time during

their study in Maryland, another state that forbids

bare hand contact of food. Observance of workers

by customers was also cited as an incentive to wear-

ing gloves (Lubran et al., 2010). The FDA (2010) re-

ported that food employees in the deli departments

of retail stores were compliant with proper, adequate

hand washing procedures 48% of the time. In con-

trast, Lubran et al. (2010) found that employees at

chain stores washed their hands 17% of the recom-

mended times and food employees at independent

stores washed their hands 2% of recommended

times. Lubran et al. (2010) also found that employees

wiped surfaces of equipment with cloths dipped in

a sanitizer solution without washing and rinsing the

surface first. Of course, sanitizer effectiveness is re-

duced when it is used on unclean surfaces such as

slicer blades, which often contain greasy food resi-

dues from slicing luncheon meats and cheeses with

high fat content (Vorst et al., 2006).

Modeling transfer coefficients for Listeria monocytogenes

Under SSOPs for deli operations, it is vital that

any piece of equipment be clean prior to sanitizing

(FDA, 2009). However, to systematically assess 1) the

vast array of cleaning compounds and 2) cleaning

protocols, in addition to variations in equipment and

food contact surfaces, known concentrations, mix-

tures of blended deli meat, and fluorescent indicator

would have to be used to purposefully contaminate

a number of deli slicers from various manufactur-

ers followed by cleaning according to accepted

protocols. After proper disassembly and cleaning

the slicer, we found in our lab that testing could be

followed by staining any meat residue with a dilute

water solution of food grade dye, FD&C 3 and 40,

followed by re-cleaning any meat residues prior to

quantifying the amount of remaining meat residues

by fluorescence (Crandall et al., 2009). This process

would also allow assessment of cross-contamination

of the sink area.

Thermal inactivation of Listeria monocy-togenes on food equipment

In addition to the FDA mandate to clean all room

temperature food contact surfaces “to sight and

feel” every 4 hours, additional treatments are re-

quired to sanitize potential niches, Listeria harborag-

es, that are difficult to reach with traditional cleaning

and sanitizing treatments. Research conducted in

our laboratory established threshold time, tempera-

ture and relative humidity to achieve a 5 log reduc-

tion of L. innocua, a more heat stable surrogate for

L. monocytogenes (Crandall et al., 2010). Crandall et

al. (2010) determined on dissembled slicer food con-

tact surfaces that 77º C for 3 hours under saturated

moisture conditions could produce a 5 log reduction

of the more heat resistant L. innocua.

However, two questions remain: it must be deter-

mined whether repeated moist heat treatment of an

entire slicer, including the electrical components, im-

mediately followed by an extensive hot drying cycle

is practical in a retail deli setting. Secondly, what are

the heat transfer rates from a commercial convection

oven or moist-heating bread proofing cabinet into

the potential L. monocytogenes harborages on a

commercial deli slicer?

DELI EMPLOYEE TRAINING AND ROLE IN SANITATION

Many of the most important issues on sanitation,

employee hygiene and training have been recently

summarized for deli workers and will only briefly dis-

cussed here (Clayton and Griffith, 2008; Endrikat et

al., 2010; Gapud, 2009). There are several factors

to consider. First, while isolated incidences of food

borne illness outbreaks (which involve more than a

single individual) garner national publicity, the vast

majority of food-borne illnesses are sporadic, where

an individual person contracts a life-threatening ill-

ness that is not part of a recognized, multiple-per-


son outbreak (Jones et al., 2004). Secondly, almost

half (44%) of American adults eat out daily and more

than 40% of the food borne illness outbreaks where

a source had been identified were linked to retail

food service between 1993 and 1997. Improving

the safety of retail environments is an area where a

significant impact can be achieved by improving the

overall safety of RTE deli foods. Third, there is an

ever-increasing, world-wide population who are at

increased risk from listeriosis: the elderly, immuno-

compromised, the very young and pregnant women.

In the seven years from 1993 to 2000, the elderly por-

tion of the U.S. population (those over 65) grew 20%,

and the elderly currently represent about 1 person

in 8 living in the U.S. (U.S. Census Bureau, 2005). In

addition, persons 60 and over have been shown to

have little awareness of the preventative measures

to take to minimize their risk of listeriosis (Cates et

al., 2006), suggesting a dangerous combination of

higher vulnerability and lack of knowledge.

In addition, Dan Engeljohn, Deputy Assistant

Administrator for the Food Safety and Inspection

Service (FSIS), has been quoted as saying that his

agency has “found that deli departments generally

have insanitary conditions, which raises the risk that

an outbreak of Listeria monocytogenes would oc-

cur.” The FSIS’s report estimates that “a person is

7 times more likely to die from listeriosis after eating

deli meat sliced by a retailer than in a federal (in-

spected) plant.” (Johnston, 2009). Current L. mono-

cytogenes risk assessment models have identified

RTE deli meats as posing one of the highest risks for

consumers to contract listeriosis. The USDA/FDA/

CDC collaborative risk assessment has declared that

RTE deli meats pose the single greatest danger for

listeriosis. The public spends billions of dollars each

year eating from delis, where sliced meats account

for almost 40% of deli sales (Anonymous, 2000). In

1998, consumers were found to purchase deli foods

an average of 2.5 times per week, an increase from

1.2 times per week just 5 years before in 1994 (Anon-

ymous, 2000). However, FSIS lacks both the legisla-

tive authority and staff to inspect the approximately

1 million deli and food service establishments em-

ploying an estimated 9 million workers.

A key reason to focus on deli employee training

is that in the majority of sporadic cases of listerio-

sis it is not possible to identify the specific contami-

nated food due to long L. monocytogenes incuba-

tion times, loss of samples and most consumers’

inability to accurately recall what they have eaten

and where they ate during the past several days. For

these reasons, many experts believe that improving

employee behavior in retail establishments, many of

which may also contain environmental L. monocyto-

genes, could be the best strategy for reducing the

incidence of sporadic cases of listeriosis (Varma et

al., 2007). However, any form of non-targeted, gen-

eralized training could be problematic for broad

spectrum implementation in the deli retail industry.

Almost half (43%) of deli employees are between the

ages of 16 and 24, earning an average of only $215 /

week, which ranks these employees right at the bot-

tom pay rate of all industries (Bureau of Labor Statis-

tics, 2009). One in 8 front line supervisors are Hispan-

ic, more than ¼ of the employees speak a language

other than English at home, and 1/5 of adults read

at or below the 5th grade (Sneed and Strohbehn,

2008). There is potential to significantly improve this

situation by conducting culturally appropriate edu-

cation in partnership with deli retailers and health

departments. Only a small number of studies have

investigated the behaviors of employees working in

potentially L. monocytogenes contaminated retail

delis. Furthermore, none of these have focused on

comparative evaluation of deli training materials, tar-

geting known L. monocytogenes harborages, devel-

oping new risk assessment modeling information, or

deli training specific for minority employees.

Effectiveness of current training

Although there have been previous evaluations

of training materials, none have been as complete

as is possible with more recently developed com-

puter based training platforms. For example, Frash

et al. (2006) measured how well current face-to-face

food safety training for food service managers was

being transferred to their employees in 1,000 stores

across 8 states. The extent of the transfer of food


safety training was assessed by tracking the number

of health code violations reported by individual in-

store inspections from local health inspectors. They

hypothesized that greater knowledge and more

training would translate into fewer violations. Un-

fortunately they found no significant reduction in

the number of violations in stores where the man-

ager had greater food safety knowledge. In a second

study published more recently, Noble et al. (2009)

surveyed more than 150 stores in Canada comparing

1) those that required food safety training for their

managers and employees to 2) stores without re-

quired food safety training for managers. They also

categorized 863 health code violations from 1,417

health department inspections. Once again, there

was no significant difference between retail stores

requiring food safety training and those not requir-

ing training. They concluded that food handlers did

not put their knowledge into practice, as evidenced

by the numerous violations of health code regula-

tions that were specifically covered in the training.

It can be assumed that the training reported in this

research of both the managers and the deli work-

ers was in the traditional face-to-face manager train-

ing new employee. What is known is that in these

Canadian stores all the trained managers and food

handlers had passed an exam and obtained a Food

Handler Certification. Consequently there is a need

contribute additional understanding to enhancing

the “transfer of training” to the target audience in

the retail deli.

Training requirements

Whether evaluating face-to-face or computer

based training, all training has at least two principal

instructional elements that must be considered: 1)

the information, the facts that get transmitted to the

knowledge realm of the employee and 2) the pre-

sentation, the way in which the knowledge is trans-

mitted. For a training program to be effective, it

must not only transmit the appropriate knowledge

that is needed, but must also present the knowledge

in a learner appropriate manner. When performing

an evaluation of existing food training programs, the

first thing to evaluate is whether there is a univer-

sally accepted set of standards by which the current

knowledge base that food service and especially

deli employees are being trained. The Model Food

Code provides generally accepted minimums, but

there are multitudes of varying health department

and company regulations.

Need for more targeted training ap-proaches

As of 2005, 48 of 56 U.S. state and territories’

health department food safety rules were based on

one of the five versions of FDA’s Food Code, begin-

ning with the 1993 edition (FDA, 2009). The latest re-

vision, due out in early 2011, requires documentation

by management of the managers’ food safety train-

ing of employees. This complex regulatory situation

is well illustrated by the Texas Department of Health

Guidelines for Food Establishments (2009), which

contains almost 70 references to various health de-

partmental requirements at that must be part of deli

employee training and their certification exams.

The Retail Food Service Food Safety Consor-

tium was formed to assist in this area and is com-

posed of faculty from 5 universities and 3 national

associations. In addition, third party auditors, such

AIB, have their own food safety employee training

manuals, which are more than 100 pages long. Third

party auditors are instructed to “observe employee

behavior and ask questions” as the basis of deter-

mining the effectiveness of employee food safety

training (O’Bryan, 2010). This morass of regulations

and lack of an accepted method to verify learning on

the part of the employee may contribute to putting

consumers at greater risk. Secondly, there is a need

to know how food employees are being trained. An

informal poll of a group of food processor HACCP

managers meeting in a focus group setting, con-

ducted by the authors, revealed several important

points to be considered. They were asked “How do

you make the time and how do you train rank and file

employees plus new employees when their turn-over

rate exceeds 150%?” For the most part, there was

consensus that training had to be constant with con-


current explanations provided on correct safety pro-

cedures. In addition, group meetings, “hands-on”

approaches and belonging to a team were deemed

as important. In short, all of these solutions were

all traditional face-to-face approaches. While these

face-to-face methods are appropriate, they are not

always feasible with increased time pressures and

dwindling training resources (McEvoy and Buller,

1990). However, utilizing this established training

information together with training methods that are

supported by technologies may result in better train-

ing in a more time efficient manner.

Potential barriers for current training approaches

Immigrants are a substantial and growing seg-

ment of the U.S. labor force. In 2008, 24.1 million

persons (15.6%) of the U.S. labor force were foreign-

born (Bureau of Labor Statistics, 2009). In 2000, 25

million of the 31 million foreign-born people resid-

ing in the U.S. indicated that they spoke a language

other than English at home. (Shin and Bruno, 2003).

Despite the language barrier, millions of non-English

speaking immigrants work in U.S. labor force and

many of them in retail delis. Currently, the food-

service industry is one of the largest employers of

Non-English Speaking Individuals (NESI) in the U.S.

(National Restaurant Association, 2006). In New York

City, home to the most restaurants in the U.S., 67.5%

of restaurant workers were immigrants (Restaurant

Opportunities Center of New York, 2003).

Language barriers are but one of the great im-

pediments to smooth integration of immigrants

into a workforce (Loosemore and Lee, 2001; Victor,

1992). The process of communication has multiple

variables, including culture, norms, attitudes, social

organizations, non-verbal behavior and language

(Victor, 1992). Foodservice managers often struggle

to communicate with Non-English speakers (Lee and

Chon, 2000).

Food safety training platforms

Education is considered the teaching of facts and

knowledge, in contrast to training, which actually

allows the employee to experience “hands-on” in-

teraction with the knowledge (Yiannas, 2008). There-

fore, the first task in successful training is to evaluate

the content (knowledge) contained in current retail

food service training platforms. The four most com-

monly used computer-based food safety employee

training programs are listed in Table 2. These four

retail food safety training platforms are considered

representative of currently used training programs.

A complete summary of the characteristics for

each platform, including contact information, price,

training target audience, numbers of modules/units,

certifications, and other parameters, are being re-

viewed by a panel of experts in training . Alchemy

Systems has a training library focused on HACCP for

foodservice and frontline workers in food processing

plants. They also design and build eLearning mod-

ules for their clients and already have several food

processors as customers. ServSafe offers a $15 on-

line training program, Starters Training and Assess-

ment, which is advertised as a complete solution to

deliver food safety training. The Food Marketing

Institute (FMI) in cooperation with Learnovation, LLC

developed SuperSafeMark for training employees of

supermarkets and food warehouses; the employees’

text and on-line training is $15, and a complete kit

for trainers, including slides, videos, and manuals is

$200 for non-FMI members. The Training Achieve-

ment Program, TAP, offers online training for food

Table 2. Selected computer based food safety employee training programs.

System Website

Alchemy Systems

http://www.alchemysystems.com

ServSafe http://www.servsafe.com/foodsafety

FMI Super Safe Mark program

http://www.fmi.org/supersafemark/

Training Achievement Program, TAP

http://www.tapseries.com/


handlers at $15 per student. Several of these train-

ing platforms have been accepted by U.S. state and

county health departments as meeting their require-

ments for Food Handler Certification.

To adapt these general food safety training plat-

forms specifically for deli operations for ideal imple-

mentation: 1) an Evaluation Matrix should be de-

veloped that is specific for individual delis, which

identifies the amount of knowledge, including the

elements of the FDA’s Model Food Code, plus de-

tails from the best from the training sites. Develop-

ing an individual Evaluation Matrix allows the incor-

poration of generally accepted practices and the

local cultural norms of “this is how we do that around

here”. Once the evaluation is complete, the current

training regime needs to be compared against this

newly developed Evaluation Matrix. In addition,

the content should be classified using the revised

Bloom’s Taxonomy (Anderson and Krathwohl, 2001)

to determine the instructional level of the content in

relation to each standard that will be taught. Some

of the supporting research that deli managers will

need has already been done. Pisik (1997) normal-

ized an instrument for evaluating on-line courses for:

content, appropriate for learners (appropriate exam-

ples), transfer of learning to job, design / packaging

and usability issues such as ease of navigation and

logical layout of the materials.

The imperative knowledge that must be included

for complete deli employee training in a particular

retail establishment can be garnered from 1) the list

of standards for knowledge content that regulatory

bodies require enhanced with 2) additional knowl-

edge content determined to be critical by the resi-

dent deli subject matter experts.

This list of standards could be incorporated in the

Evaluation Matrix to compare what experts feel must

be included in the training versus what is actually

included in each of the current training programs.

For example, to our knowledge, none of the current

training utilizes the most recent research findings

that have shown that persistent strains of L. monocy-

togenes reside in specific niches in deli environments

(Sauders et al. 2004, 2009). These L. monocytogenes

harborages must be identified and addressed in deli

specific food safety training programs (Table 2). To

identify missing or more specific standards, holding

Delphi panels composed of experts in deli opera-

tions to identify food safety knowledge specific to

individual retail deli operations may be needed. The

Delphi technique is a means for obtaining group

input for ideas to develop a number of alternatives

for a situation (Linstone and Turoff, 1975). In order

to solicit participation from a variety of subject ex-

perts, a modified Delphi technique using the World-

Wide-Web is a possible approach. A prospective

Delphi panel for delis would be composed of ex-

perts including corporate supermarket personnel,

supermarket deli managers, hourly supermarket deli

employees, independently owned deli managers,

Table 3. Food Safety Training Programs and Opportunities for Deli Workers

Categories Current Training Available Potential training opportunities

Programs Computer based, more general food safety training platforms

Delphi techniques to identify specific training/education objectives specific for deli workers.

Evaluation Limited to available platforms with deli-specific information, some integrate stan-dards and permits ranking in an evalua-tion matrix

New training modules suitable for deli workers designed in cooperation with commercial suppliers and professional multi-media staff

Knowledge Base Limited knowledge of experts, deli man-agers and employees

Usability testing by new employees for the deli specific training

Outcome Minimal deli-specific knowledge and un-der-motivated usability by new food ser-vice students

Motivation of short-term learning and long-term behavior change


hourly deli employees, health department inspec-

tors and food safety professionals. An appropriate

number of content experts currently working in del-

is are considered to be 15 to 20 to complete each

round (Linstone and Turoff, 1975).

A three-round Delphi Panel Design can be used

to better define panel member views of what food

safety topics are specific to deli operations. Initially,

the process involves supplying a list of deli specific

trainings in addition to general sanitation training

(Gapud, 2009), and asking the Delphi panel mem-

bers for other factors that may not have been listed.

Online survey instruments can be a useful to rapidly

facilitate this process.

Secondly, panel members could be required to

narrow the list to the top 10 training foci and then

rank those 10 according to importance when exam-

ining food safety priorities for delis. Finally, panel

members would be asked whether they agree with

the rankings in Round 2 or re-rank order the factors

again. They will be asked if they feel any of the top

10 items would reflect differently depending on indi-

vidual needs of the deli where this training will take

place.

The Delphi panel results can also be integrated

into the general food safety and sanitation training

to create integrated standards, which could be built

into the Evaluation Matrix to identify the breadth

and depth of knowledge found in the existing train-

ing programs. This integrated standards list could

also be used to identify gaps in content and inform

the development of new deli specific online training

materials. Once expert investigators identify which

integrated content standards are addressed, to what

extent and at which level of learning for each of the

four online training platforms, the Evaluation Ma-

trix can be used to assess the knowledge content

presented in the online training programs that are

under consideration for purchase, against the inte-

grated standard developed after being reviewed by

deli managers and employees. This would include

reviewing all programs and calculating inter-rater

reliabilities to determine the level of agreement

among experts and also would allow outliers to be

reassessed.

It is important in this process to assess the deliv-

ery and usability of these retail food safety training

platforms to potential deli managers and their em-

ployees. Computer based training software, such as

those listed in Table 2, would be an ideal place to

start. However, “hands-on” evaluation by individuals

actually working in the food service industry should

be an essential follow up step. Prior to evaluating

the knowledge content of the food safety training

platforms, participants would have to be given a pre-

test to determine their current level of food safety

knowledge. Short-term learning could be assessed

by completing a post-test after these participants

have used various training platforms. Once com-

pleted, the training platform could be further evalu-

ated using the checklist for evaluating online courses

described previously. In addition, several of these

platforms are fairly versatile in also having programs

in Spanish, Vietnamese, Chinese, Korean or Asian

(Urdu, Hindi or Bengali). In summary, evaluating cur-

rent and computer based training and measuring

learning outcomes addresses two of Kirkpatrick’s

(1998) four levels of evaluation (Reaction, Learning,

Behavior, and Results). Achieving thorough evalu-

ation should result in critical comparative data that

food safety instructors and deli managers could use

to make informed choices in selecting training mate-

rials suitable for their particular needs.

Developing deli specific training

The U.S. Census Bureau noted that by 2015, the

Hispanic population will be double the size it was in

1990 (U.S. Census Bureau, 2009). As one of the larg-

est employers of Hispanics, the restaurant industry

must focus its efforts on training non-English speak-

ers in food safety. Offering training material in Span-

ish may be a limited start but may not be a com-

plete solution. By doing so, the assumption is made

that all workers (not just Hispanic workers) are visual

learners and that all have similar reading levels. In

addition, young people entering the workforce may

be more familiar with computers and online training.

By combining the best practices for training Non-

English Speaking Individuals (NESI) workers with on-


line training, deli operations can be more successful

in their food safety training, therefore reducing the

risk of foodborne illness from L. monocytogenes.

Usability testing is defined as the degree of suc-

cess a user has in learning and using a product to

achieve their goals, as well as satisfaction of the user

with the product (Dumas and Redish, 1993). Usability

testing is not a test of the specific training materials,

but rather an evaluation of the interactions between

the user-learner and the training materials. Rubin

(1994) suggested that the most logical approach for

deli managers with sufficient lead time is to conduct

a literature search, write and review training ideas

along the way and then create learning objectives

that are specific for the current retail deli operation.

Each of the new training modules are typically bro-

ken down into clear, concise plans for testing. The

major concepts pertinent for testing deli workers

include: (1) short-term learning by the user, as mea-

sured using a questionnaire prior to and after com-

pletion, (2) user’s completion times, as measured by

overall completion time for each module, the num-

ber of videos used by each user and number of links

properly accessed, (3) user completion of tasks, as

outlined in the virtual orientation and (4) overall user

satisfaction with each of the design elements, as

measured by an oral question and answer session of

each user after completion of the virtual tour.

Rubin’s (1994) design for a usability test is focused

on user’s reactions to the major elements. Data col-

lection methods include “think aloud” user test-

ing (Travis, 2003), which consists of recording users

talking out loud as they perform prescribed tasks.

The usability data collection can be combined with

a contemporaneous record of the users’ thoughts

from the think aloud sessions while they are com-

pleting their virtual tours. O’Bryan et al. (2009, 2010)

have demonstrated that this combination gives

a complete evaluation by each user with minimal

crossover among users and minimal external influ-

ences. Nielsen (2000) determined that 5 users rep-

resenting the target population will discover 85% of

the usability flaws of a web site design, with 12 to 15

users finding close to 100% of problems.

Typically, each participant is given a written pre-

test knowledge survey covering basic knowledge

of retail deli food safety, given brief instructions for

launching the web-based learning and provided

with a task list. Participants are encouraged to talk

out loud (think out loud) during the entire time they

are using the learning module. Their conversations

can be observed by a test monitor and/or video-

taped during their entire session. After completion

of the training modules, participants are usually

given a written post-test knowledge survey and a

profile questionnaire to obtain demographic infor-

mation. After the post-test, participants are asked

their opinions of how usable the computer based

module was. The test monitor also provides a way

for these learners to offer their opinions and sugges-

tions for improvement. By giving participants a pre-

test and post-test assessment of learning, in addi-

tion to the usability testing, two of Kirkpatrick’s four

levels of evaluation (Reaction, Learning, Behavior,

and Results) can be addressed, giving a more robust

and accurate evaluation. Objective data can be com-

plied on completion times, number of tasks success-

fully completed and overall user satisfaction with the

design elements can be recorded. The summary of

the think-aloud comments into similar groupings

can be compiled and analyzed for common trends

or themes that can be used to inform revisions to the

training program.

These suggested approaches can yield produc-

tion of new online training programs targeted spe-

cifically at diverse deli workers and incorporating the

latest methods and knowledge concerning effective

sanitation techniques for the reduction of L. mono-

cytogenes harborages. The Delphi Panel provides a

ranked list of factors that can be used for develop-

ing deli food safety modules. The evaluation of the

new training materials is focused on two of Kirkpat-

rick’s four levels of evaluation (Reaction and Learn-

ing). While it is desirable to address all four levels,

the timeframe required to address all four levels is

usually beyond the scope of ongoing program de-

velopment. However, data obtained from this evalu-

ation still provides a foundation to facilitate future

assessment of the remaining two levels (Behavior

and Results).


Certified training programs

There are also advantages of having a certifica-

tion program associated with any training. However,

new certification programs are usually not needed

because 1) there already are several nationally (and

internationally) recognized certification programs for

food safety trainers and managers and 2) many of

these certifications have already met the current le-

gal requirements imposed by local and state health

officials. The recent review by Gapud (2009) covered

these. Sites such as ServSafe’s give an overview of

these varying legal and certification requirements.

Secondly, consumer in-home preparation is the final

link in the food safety chain. However, existing pro-

grams, such as Partnership for Food Safety Educa-

tion – Food Marketing Institute and Fightbac have

extensive cooperative networks and a track record

of proven successes in educating consumers in safe

food handling practices.

CONCLUSIONS

A carefully planned training program that includes

assessing impact of training on both short and long

term changes in behavior can lead to a decrease

in the number of consumers contracting listeriosis.

Noble et al. (2009) have developed methods to com-

pare the public health inspection records, especially

looking for violations of hand hygiene before and af-

ter training of employees to see if increased inspec-

tion without additional training decreased the num-

ber of critical violations. If sanitation techniques are

better able to clean, sanitize and thermally treat po-

tential L. monocytogenes harborages on food con-

tact surfaces, equipment and environmental surfac-

es, then it can be assumed that employees in these

environments will be less likely to cross-contaminate

RTE foods with L. monocytogenes. Again, fewer L.

monocytogenes harborages and minimal cross-con-

tamination will decrease the hazard and decrease

the risk of listeriosis. Extending these findings to

impact deli specific training targeting Hispanics and

other minority deli employees can decrease one of

the principal factors in reducing risk, minimizing or

eliminating the hazard. Short term assessment of

training for deli employees to minimize behaviors

that cause cross-contamination is fairly achievable.

However, long-term behavior changes are difficult

to measure in employees where there are high-turn-

over rates, such as in a deli. Deli specific training

with pre-test and post-test assessment of learning, in

addition to the usability testing at the Reaction and

Learning levels, could make these newly developed

training modules part of the solution.

ACKNOWLEDGEMENTS

This review was supported by a USDA National

Integrated Food Safety Initiative Grant to the au-

thors, an American Meat Institute grant to Drs. Phil

Crandall and Steve Ricke and a USDA CREES Food

Safety Consortium grant to Steve Ricke.

REFERENCES

Anderson, L. W. and D. R. Krathwohl. 2001. A taxono-

my for learning, teaching and assessing: A revision

of Bloom’s Taxonomy of educational objectives.

Longman, New York, NY. 384 p.

Anonymous. 2000. Dollars from the Deli. Findarticles.

com. http://findarticles.com/p/articles/mi_m3289/

is_2_169/ai_60071462/ Accessed Nov, 2009.

Bricher, J. L. 2007. Protecting your customers while

safeguarding your business. Food Saf. 13:1-40.

Buchholz, A. L., Z. Yan, E. T. Ryser. 2008. Glo-germ as

a cross contamination indicator during processing

of leafy greens. IAFP 95th Annual Meeting, p. 14

(Abst. T4-11).

Buchholz, A. L., G. R. Davidson, B. P. Marks, E. C.

D. Todd, E. T. Ryser. 2009. Transfer of Escherichia

coli O157:H7 from equipment surfaces to iceberg

and romaine lettuce during simulated commercial

processing., IAFP 95th Annual Meeting, p. 17 (Abst.

T5-04).

Buchholz, A. L., G. R. Davidson, D. T. Campos, B. P.

Marks, E. C. D. Todd, E. T. Ryser. 2009. Quantifica-

tion of Escherichia coli O157:H7 transfer to equip-

ment during commercial production of fresh-cut


leafy greens. IAFP 96th Annual Meeting, p. 34

(Abst. P3-22).

Bureau of Labor Statistics. 2009. Foreign-born work-

ers: Labor force characteristics in 2008. Bls.gov.

http://www.bls.gov/news.release/pdf/forbrn.pdf

Accessed December, 2009.

Campos, D. T., B. P. Marks, K. L. Vorst, L. E. Kes-

kinen, Z. Yan, Y. Zhang, P. Sirmeyer, E. T. Ryser.

2009. Accounting for product residue effects when

modeling bacterial transfer between processing

equipment and meat products. IAFP 96th Annual

Meeting, p. 3 (Abst. P1-01).

Cates, S. C., R. A. Morales, S. A. Karns, L. A. Jaykus,

K. M. Kosa, T. Teneyck, C. M. Moore, P. Cowen.

2006. Consumer knowledge, storage, and han-

dling practices regarding Listeria in frankfurters

and deli meats: Results of a web-based survey. J.

Food Prot. 69:1630-1639.

CDC, Centers for Disease Control and Prevention.

2009. Preliminary FoodNet Data on the Incidence

of Infection with Pathogens Transmitted Common-

ly Through Food --- 10 States, 2008. Morb. Mortal.

Wkly. Rep. 58:333-337.

Chen, Y. H., K. M. Jackson, F. P. Chea, D. W. Schaff-

ner. 2001. Quantification and variability

analysis of bacterial cross-contamination rates in

common food service tasks. J. Food Prot. 64:72-80.

Clayton, D. and C. Griffith. 2008. Efficacy of an ex-

tended theory of planned behavior model for

predicting caterers’ hand hygiene practices. Int. J.

Environ. Health Res. 18:83-98.

Crandall, P. G., C. A. O’Bryan, E. M. Martin, S. Pend-

leton, E. Shannon, J. Marcy, and S. C. Ricke. 2010.

Dry heat thermal inactivation of Listeria innocua on

deli slicer components. Food Prot. Trends 30:589-

592.

Crandall, P. G., J. Marcy, S. C. Ricke, M. Johnson, C.

A. O’Bryan, and E. M. Martin. 2009. Cost effective

treatments to minimize Listeria monocytogenes

cross contamination of ready-to-eat meats by the

in-store deli meat slicer. Available at: http://amif.

org/ht/a/GetDocumentAction/i/52743 Accessed

28 June 2011.

Dumas, J. S. and J. C. Redish. 1993. A practical

guide to usability testing. Ablex Publishing Corp,

Norwood, NJ. 304 p.

Endrikat, S., D. Gallagher, R. Pouillot, H. Quesen-

berry, D. LaBarre, C. Schroeder, J. Kause. 2010. A

comparative risk assessment for Listeria monocy-

togenes in prepackaged versus retail-sliced deli

meat. J. Food Prot. 73:612-619.

FDA. 2009. Food code 2009. Fda.gov. http://www.

fda.gov/Food/FoodSafety/RetailFoodProtection/

FoodCode/FoodCode2009/default.htm Accessed

May, 2010.

FDA. 2010. FDA trend analysis report on the occur-

rence of foodborne illness risk factors in selected

institutional foodservice, restaurant, and retail

food store facility types (1998-2008) http://www.

fda.gov/Food/FoodSafety/RetailFoodProtection/

FoodborneIllnessandRiskFactorReduction/Retail-

FoodRiskFactorStudies/ucm223293.htm Accessed

May, 2011

Frash, R., M. Binkley, D. Nelson, B. Almanza. 2006.

Transfer of training efficacy in US food safety ac-

creditation. J. Culinary Sci. Technol. 4:7-38.

FSIS. 2009. FSIS Comparative Risk Assessment for

Listeria monocytogenes in ready-to eat meat

and poultry deli meats. Fsis.gov. http://www.fsis.

usda.gov/Science/risk_assessments/index.asp Ac-

cessed May, 2010.

Gapud, V. 2009. Food safety trends in retail and

foodservice. Food Saf. 12:28-35.

Garrido, V., A. I. Vitas, I. Garcia-Jalon. 2009. Survey of

Listeria monocytogenes in ready-to-eat products:

Prevalence by brands and retail establishments

for exposure assessment of listeriosis in Northern

Spain. Food Control. 20:986-991.

Johnston, T. 2009. FSIS to focus on retailers to im-

prove meat safety. Meatingplace.com. http://www.

meatingplace.com/MembersOnly/webNews/de-

tails.aspx?item=11846 Accessed May, 2010.

Jones, T., B. Imhoff, M. Samuel, P. Mshar, K. G. Mc-

Combs, M. Hawkins, V. Deneen, M. Cambridge,

S. J. Olsen. 2004. Limitations to successful inves-

tigation and reporting of foodborne outbreaks: an

analysis of foodborne disease outbreaks in Food-

Net catchment areas, 1998–1999. Clin. Infect. Dis.

38:297–302.

Kause, J. 2009. Interagency retail Listeria monocy-


togenes risk assessment. www.fsis.usda.gov/PPT/

Lm_Kause_062309.ppt Accessed May, 2010.

Keskinen, L. A., E. C. D. Todd, E. T. Ryser. 2008. Trans-

fer of surface-dried Listeria monocytogenes from

stainless steel knife blades to roast turkey breast.

J. Food Prot. 71:176-181.

Kirkpatrick, D. L. 1998. Another look at evaluating

training programs. American Society for Training &

Development, Alexandria, VA. 257 p.

Lee, C. and K. S. Chon. 2000. An investigation of

multicultural training practices in the restaurant

industry: the training cycle approach. Int. J. Con-

temp. Hospital Manage. 12:126-134.

Linstone, H. and M. Turoff. 1975. The Delphi Meth-

od: Techniques and Applications. Addison- Wes-

ley Pub. Co., Upper Saddle River, NJ. 618 p.

Loosemore, M. and P. Lee. 2001. Communication

problems with ethnic minorities in the construction

industry. Int. J. Project Manage. 20:517-524.

Lubran, M. B., Pouillot, R., Bohm, S., Calvey, E. M.,

Meng, J., Dennis, S. 2010. Observational study of

food safety practices in retail deli departments. J.

Food Prot. 73:1849-1857.

Lynch, R. A., M. L. Phillips, B. L. Elledge, S. Hanu-

manthaiah, and D. T. Boatright. 2005. A preliminary

evaluation of the effect of glove use by food han-

dlers in fast food restaurants. J. Food Prot. 68:187–

190.

McEvoy, G. M. and P. F. Buller. 1990. Five uneasy

pieces in the training evaluation puzzle. Training

and Development Journal 44:39-42.

Motulsky, H. and A. Christopoulos. 2004. Fitting

models to biological data using linear and nonlin-

ear regression: a practical guide to curve fitting.

Oxford University Press, Cary, NC. 350 p.

National Restaurant Association. 2006. Restaurant

industry facts. Restaurant.org. http://www.restau-

rant.org/research/ind_glance.cfm. Accessed Jan,

2010.

Nielsen, J. 2000. Why you only need to test with

5 users. Useit.com. http://www.useit.com/alert-

box/20000319.html Accessed April, 2007.

Noble, S., M. Griffiths, S. Thompson, T. Maclaurin.

2009. Frequency and type of food safety infrac-

tions in food establishments with and without cer-

tified food handlers. Food Prot. Trend. 29:840-848.

O’Bryan, C. A., P. G. Crandall, K. Shores-Ellis, D. M.

Johnson, S. C. Ricke, J. Marcy. 2009. Optimizing

web based instruction — A case study using poul-

try processing unit operations. J. Food Sci. Educ.

8:93-100.

O’Bryan, C. A., D. M. Johnson, K. Shores-Ellis, P. G.

Crandall, J. A. Marcy, S. C. Seideman, S. C. Ricke.

2010. Designing an affordable usability test for eL-

earning modules. J. Food Sci. Educ. 9:6-10.

Pisik, G. B. 1997. Is this course instructionally sound?

A guide to evaluating online training courses. Edu-

cational Technol. 7:50-59.

Restaurant Opportunities Center of New York. 2003.

The New York City restaurant industry analysis: The

Quantitative Report. ROC-NY, New York, NY. 71 p.

Roberts, T. and R. Pinner. 1990. Economic impact of

disease caused by Listeria monocytogenes. In: A.

J. Miller, J. L. Smith, G. A. Somkuti. Eds. Foodborne

Listeriosis. Elsevier Scientific Publications, Amster-

dam, UK. p 137-149.

Rubin, J. 1994. Handbook of usability testing: How

to plan, design, and conduct effective tests. John

Wiley & Sons Inc, New York, NY. 384 p.

Sauders, B. D., K. Mangione, C. Vincent, J. Schermer-

horn, C. M. Farchione, N. B. Dumas, D. Bopp, L. Ko-

rnstein, E. D. Fortes, K. Windham, M. Wiedmann.

2004. Distribution of Listeria monocytogenes mo-

lecular subtypes among human and food isolates

from New York State shows persistence of human

disease-associated Listeria monocytogenes strains

in retail environments. J. Food Prot. 67:1417-1428.

Sauders, B. D., M. D. Sanchez, D. H. Rice, J. Corby,

S. Stich, E. D. Fortes, S. E. Roof, M. Wiedmann.

2009. Prevalence and molecular diversity of Lis-

teria monocytogenes in retail establishments. J.

Food Prot. 72:2337-2349.

Schaffner, D. W. 2004. Mathematical frameworks for

modeling Listeria cross-contamination in food-

processing plants. J. Food Sci. 69:155-159.

Scharff, R. L. 2010. Health-related costs from

foodborne illness in the United States. Produc-

esafetyproject.org.http://www.producesafetypro-

ject.org/admin/assets/files/Health-Related-Food-

borne-Illness-Costs-Report.pdf-1.pdf Accessed


May, 2010.

Shin, H. B., and R. Bruno. 2003. Language use

and English-speaking ability: 2000, Census

2000 Brief. Census.gov. http://www.census.gov/

prod/2003pubs/c2kbr-29.pdf Accessed May, 2010.

Sneed, J., and C. H. Strohbehn. 2008. Trends impact-

ing food safety in retail foodservice: implications

for dietetics practice. J. Am. Diet. Assoc. 108:1170-

1177.

Texas Department of Health. 2009. Guidelines for

Food Establishments. Dshs.state.tx.us. http://

www.dshs.state.tx.us/foodestablishments/rules.

shtm Accessed May, 2010.

Travis, D. 2003. Discount usability: time to turn back

the pendulum? Userfocus.co.uk. http://www.user-

focus.co.uk/articles/discount.html Accessed June,

2009.

US Census Bureau. 2005. 65+ in the United States:

2005. Census.gov. http://www.census.gov/

prod/2006pubs/p23-209.pdf Accessed Novem-

ber, 2009.

Varma, J. K., M. C. Samuel, R. Marcus, R. M. Hoek-

stra, C. Medus, S. Segler, B. J. Anderson, T. F.

Jones, B. Shiferaw, N. Haubert, M. Megginson, P.

V. McCarthy, L. Graves, T. Van Gilder, F. J. Angulo.

2007. Listeria monocytogenes infection from foods

prepared in a commercial establishment: a case-

controlled study of potential sources of sporadic

illness in the United States. Clin. Infect. Dis. 44:521-

528.

Victor, D. A. 1992. International business communi-

cations. Harper Collins, New York, NY. 280 p.

Vorst, K. L., E. C. D. Todd, E. T. Ryser. 2006. Transfer

of Listeria monocytogenes during mechanical slic-

ing of turkey breast, bologna, and salami. J. Food.

Prot. 69:619-626.

Yiannas, F. 2008. Food Safety Culture: Creating a

behavior-based food safety management system.

Springer, New York, NY. 96 p.




ABSTRACT

The objective of this study was to compare the ability of Salmonella to survive in organic and conven-

tional broiler feeds as affected by temperature (11, 25, and 38°C), water activity (aw 0.75, 0.55, and 0.43) and

storage time (up to 80 days). Feeds were inoculated with a mixture of five Salmonella serotypes at high

and low populations (6 and 3 log CFU/g, respectively), and populations and presence (by enrichment) were

monitored over time. Although the number of Salmonella in organic feed for the majority of temperature-

by-aw combinations was significantly lower (P ≤ 0.05) compared to the number in conventional feed over

the 80-day storage period, differences in mean populations were less than 1 log CFU/g. The odds-ratio

(OR) for presence of Salmonella was significantly higher (P ≤ 0.05) in conventional feed than in organic feed

containing high and low inocula (OR = 4.76 and 2.92, respectively). Based on these findings, we generally

conclude that there were no biologically significant differences in survival of Salmonella in organic and

conventional poultry feeds.

Keywords: Salmonella, conventional poultry feed, organic poultry feed, storage temperature, wa-

ter activity, aw

INTRODUCTION

Salmonella is one of the most common patho-

gens known to cause foodborne disease in the Unit-

ed States and worldwide. According to Scallan et al.

(2011), an estimated 1.03 million people suffer from

Correspondence: Walid Q. Alali , [email protected]: +1 - 770-467-6066 Fax: +1-770-229-3216

salmonellosis annually in the U.S. Nontyphoidal Sal-

monella is considered to be the leading cause of

foodborne illness-related hospitalizations and death

in the U.S. Poultry and poultry products are consid-

ered to be an important source of Salmonella (Tauxe,

1991; Bryan and Doyle, 1995) and contaminated feed

is considered to be one of the main sources of Sal-

monella infection in broiler birds (Jones et al., 1991;

Maciorowski et al., 2004).

Survival of Salmonella in Organic and Conventional Broiler Feed as Affected by Temperature and Water Activity

A. Petkar1, W. Q. Alali1*, M. A. Harrison1,2, and L. R. Beuchat1

1Center for Food Safety and Department of Food Science and Technology, University of Georgia, Griffin, Georgia 30223-1797

2Athens, Georgia 30602-2610



The presence of Salmonella in poultry feed as

well as feed ingredients such as grain, oilseed meal,

feathers, fishmeal, and meat by-products has been

documented (Williams 1981a; Cox et al., 1983; Stu-

art 1984; Veldman et al., 1995). After contaminated

feed is consumed by broilers, Salmonella can mul-

tiply in the gastrointestinal tract of the bird and po-

tentially be shed in the feces during grow out (Cason

et al. 1994). Several intervention methods, including

heating and pelleting at 70°C and 90°C, irradiation

(gamma rays), and addition of chemicals (organic ac-

ids, organic salts, formaldehyde, and bacterial mem-

brane disruptors such as terpenes and essential oils)

have been applied to poultry feed and feed ingredi-

ents to control Salmonella (Wilder, 1969; Hinton and

Linton, 1988). Despite the use of these interventions,

poultry feed can be recontaminated with Salmonella

post-production, e.g., during storage at the feed

mill, transportation, and storage at the farm (Hinton

and Linton, 1988).

Storage of poultry feed under various environ-

mental conditions may influence the survival and

growth of Salmonella. Reports from several studies

show that survival of Salmonella is influenced by fac-

tors such as the presence of antimicrobials, moisture

content, and storage temperature (Himathongkham

et al., 1996; Halls and Tallentire 1978; Furuta et al.,

1980; McCapes et al., 1989). Juven et al. (1984) re-

ported that survival of Salmonella is greater at a

water activity (aw) of 0.43 than at 0.75. The survival

and heat resistance of Salmonella in poultry feed has

been reported to be inversely related to moisture

content and relative humidity (% relative humidity =

aw x 100), except at a moisture content that allows

growth (Liu et al., 1969; Carlson and Snoeyenbos,

1970; Juven et al., 1984).

In a recent study, it was found that conventional

poultry feed was contaminated with Salmonella,

whereas USDA-certified organic feed was Salmonel-

la-free (Alali et al., 2010). It is unclear whether the

organic feed was contaminated with Salmonella

which subsequently did not survive or was present

at a level below the limit of detection, or whether

the feed was not contaminated. The objective of this

study was to determine the ability of Salmonella to

survive in USDA-certified organic broiler feed versus

conventional broiler feed stored at different temper-

atures and aw over an 80-day period.


Salmonella serotypes

Five Salmonella enterica serotypes were obtained

from the Poultry Diagnostic Research Center, Univer-

sity of Georgia, Athens, GA. These serotypes were S.

Typhimurium, S. Heidelberg, S. Enteritidis, S. Monte-

video, and S. Gaminara. All serotypes were grown in

tryptic soy broth (TSB) (Difco, BD; Sparks, MD) for 24 h

at 37°C. A loopful of TSB culture of each serotype was

streaked on a double-layered agar medium consist-

ing of tryptic soy agar (TSA) (Difco, BD) and xylose

lysine tergitol (XLT4) agar (Difco, BD), and incubated

at 37°C for 24 h. The TSA-XLT4 agar was prepared

as described in Kang and Fung (2000). This agar has

been used for direct plating to support recovery and

enumeration of injured Salmonella. The agar is com-

posed of a bottom layer of XLT4 agar as a selective

medium and thin top layer of nonselective TSA.

Salmonella inoculum

To prepare the Salmonella inoculum, 6-ml quanti-

ties of nutrient broth (NB) (Difco, BD) were inoculated

with cells from single colonies of Salmonella formed

on TSA-XLT4 agar plates incubated at 37°C for 24 h.

Each serotype was cultured separately. The NB inocu-

lum concentration was adjusted to approximately 8

log CFU/ml using a spectrophotometer (Spectronic

20; Bausch and Lomb, Rochester, NY) at an opti-

cal density of 0.5 to 0.6, as described by Kaiser et al.

(2002). This suspension was used to prepare a high-

population dry chalk powder inoculum. Ten-fold se-

rial dilutions were made in Nutrient Agar (NA) to pre-

pare a low-population chalk inoculum.

Preparation of dry chalk inocula

Chalk was used as a carrier for preparing a dry Sal-

monella inoculum (Okelo et al., 2008). Prior to use,


Crayola chalk (Code #51-0320, Binney and Smith,

Easton, PA) was sterilized by autoclaving. The chalk

did not possess antimicrobial properties. The pH of

the chalk was 7.0, as measured by a pH meter (Fisher

Scientific, Pittsburgh, PA). Chalk (4.3-g pieces) was

submerged in the five-serotype mixture of high- or

low-population NB suspension for 12 h at 37°C, then

placed in sterile petri dishes and dried for 72 h at

37°C. The inoculated, dried chalk was pulverized us-

ing a food processor (Hamilton Beach Food Proces-

sor, model 70590, type FP 11, Southern Pines, NC) in a

laminar air flow chamber to obtain powdered inocula

containing Salmonella at approximately 7 log CFU/g

(high inoculum) and 4 log CFU/g (low inoculum) of

chalk. The aw of chalk powder inocula was 0.22.

Enumeration of Salmonella in dry chalk inocula

Enumeration of viable Salmonella was done by

adding 10 g of high- or low-population chalk inoc-

ulum to 90 ml of phosphate buffered saline (PBS),

vigorously shaking the suspension, and making 10-

fold serial dilutions in PBS. Samples (0.1 ml) were

surface plated in duplicate on TSA-XLT4 agar plates

and incubated at 37°C for 24 h before colonies were

counted.

Feeds used

Conventional pelleted broiler feed formulated for

grower birds (Table 1) was purchased from two con-

ventional poultry companies (companies A and B).

The formulations listed for these feeds were similar.

Organic mash feed formulated for grower birds (Ta-

ble 1) was obtained from two organic poultry com-

panies (companies C and D). Prior to inoculation,

feed samples were tested (as described later under

Salmonella analysis-selective enrichment) to ensure

that they were negative for Salmonella.

Preparation of feeds with desired aw

Saturated salt solutions (potassium carbonate,

sodium bromide, and sodium chloride) were placed

Table 1. Composition of conventional and organic broiler-grower feeds.

IngredientConventional

feed (%)Organic feed (%)6

Corn meal 72.79 65.80

Soybean meal 15.80 23.77

Alfalfa meal (dehydrated) - 1.20

Poultry meat & bone meal 6.00 -

Fat 2.70 5.00

Dicalcium phosphate 0.20 1.24

Limestone 0.66 1.55

Salt 0.39 0.31

Methionine 0.29 0.23

Vitamin premix2 0.25 0.12

Lysine 0.65 0.70

Trace mineral premix3 0.08 0.08

Coban4 0.05 -

BMD5 0.05 -

L-Threonine 0.07 -

TBCC5 0.02 -1Vitamin mix contained the following: vitamin A, vitamin D3, vitamin E, vitamin B12, riboflavin, niacin, d-Pantothenic acid, choline, menadione, folic acid, thiamine, pyridoxine, biotin, and ethoxyquin.

2Trace mineral mix contained the following: calcium, iron, magnesium, manganese, zinc, copper, iodine, selenium.

3Coban: coccidiostat

4BMD: bacitracin methylene disalycilate

5TBCC: tribaic copper chloride

6Specific ingredients listed for organic broiler feed: mono-sodium phosphate, organic kelp meal, diatomaceous earth, , ferrous sulfate, organic apple cider vinegar, , zinc sulfate, , organic potato starch, organic dehydrated eggs, organic dried tomato pomace, organic dried whole milk, organic linseed meal, organic aloe vera gel concentrate, organic soybean oil, organic oat flour, lecithin, organic wheat middlings, organic sugar, potassium chloride, at-tapulgite clay, pyridoxine hydrochloride, folic acid, ferric choline citrate complex, zinc choline citrate complex, car-otene, ascorbic acid, yeast culture, cobalt sulfate, Lacto-bacillus acidophilus, Lactobacillus casei, Bifidobacterium thermophilum, Enterococcus faecium, organic sources of (cayenne pepper, peppermint, garlic, parsley, dandelion root extract, elder flowers, dandelion extract, ginger ex-tract, German chamomile, lemon grass extract, thyme, sweet fennel extract, sweet basil, sage, cloves), and natural tocopherols.


inside each cabinet to attain the desired aw (Rock-

land, 1960). Three cabinets at three different relative

humidities were each kept at three different temper-

atures (11, 25, and 38°C). Conventional and organic

feeds were stored inside secador desiccator cabinets

(Structure Probe, Inc., West Chester, PA) for 3 weeks

to reach target aw values of approximately 0.43, 0.55

and 0.75 prior to inoculating with Salmonella. The aw

of feeds was measured over time using an Aqualab-

aw meter (Decagon Devices, Inc., Pullman, WA).

Experimental design

All experiments were replicated twice. Feeds from

companies A and C (conventional and organic, re-

spectively) were used in the first trial, whereas feeds

from companies B and D were used in the second

trial. A 3 x 3 factorial design was used to conduct

the study. For each feed containing either high- or

low inoculum, three storage temperatures (11, 25,

and 38°C) and three aw levels (0.75, 0.55 and 0.43)

were tested to determine their effects on survival of

Salmonella in conventional and organic broiler feeds

over a 80-day period. For each temperature/aw com-

bination, triplicate 9-g samples of conventional and

organic feed in sterile glass test tubes were inocu-

lated with 1 g of powdered chalk inoculum to obtain

populations of approximately 6 log (high) and 3 log

(low) CFU/g. The inoculated feeds were mixed, de-

posited in tubes, and placed inside the cabinets.

Analysis of feeds for Salmonella

Triplicate 10-g samples of inoculated feed stored

for 0, 3, 7, 14, 21, 28, 35, 50, 65, and 80 days at each

temperature/aw combination were analyzed for

populations and presence (by enrichment) of Sal-

monella. For the high-inoculum samples, 10 g were

suspended in a 90 ml of Luria Bertani (LB) (Sigma Al-

drich Corp., St. Louis, MO) broth in a 250-ml flask,

followed by shaking for 1 h on a rotary shaker (New

Brunswick Scientific, Edison, NJ). Ten-fold serial dilu-

tions of the suspension were made in sterile micro-

centrifuge tubes containing 0.9 ml of (PBS). Aliquots

(0.1 ml) from each dilution (10-1 to 10-3) were spread

plated on TSA-XLT4 agar plates. After incubating

plates for 24 h at 37°C, colonies presumptive for

Salmonella were enumerated. For the low-inoculum

samples, the procedure for enumerating Salmonella

was similar with the exception that suspensions were

not diluted. Undiluted LB/feed homogenates were

surface-plated on TSA-XLT4 agar. The LB/feed sus-

pensions were incubated for 24 h at 37°C.

Selective enrichment

When direct plating of diluted feed samples was

negative for Salmonella, 1 ml and 0.1 ml of preen-

riched LB/feed mixture were inoculated into 10 ml

of tetrathionate (TT) broth (Difco, BD) and 10 ml of

Rappaport Vassiliadis broth (RV) broth (Difco, BD),

respectively. After incubating broths for 24 h at 35°C

(TT broth) and 42°C (RV broth), a loopful of the cul-

ture was streaked on TSA-XLT4 agar. After incubat-

ing plates for 24 h at 37°C, cells from colonies pre-

sumptive for Salmonella were inoculated into triple

sugar iron (TSI) agar slants (Difco, BD), incubated

with caps loosened at 35°C for 18 to 24 h, and exam-

ined for carbohydrate fermentation, gas production,

and hydrogen sulfide production. Based on these

observations, feed samples were judged as positive

or negative for Salmonella.

Statistical analysis

Populations of Salmonella (log CFU/g) deter-

mined by direct plating samples of conventional and

organic broiler feeds were compared among each

temperature/aw combination, by inoculum level, and

at each storage day using repeated measures of

analysis of variance ANOVA in General Linear Model

(GLM) in SAS software version 9.1.3 (GLM procedure,

SAS Inst., Inc., Cary, NC). Salmonella counts were

logarithmically transformed by use of log base 10 to

approximate normality. Data from the two replicate

experiments were tested using Levene’s test for ho-

mogeneity of variances. The variances between the

two replicate experiments were not significantly dif-

ferent (P > 0.05). Based on these findings, data from


both replicate trials were pooled to obtain a set of

six observations for each sampling day. In addition,

the proportion of samples positive for Salmonella

was compared among each temperature-by-aw com-

bination, by inoculum level, and at each sampling

time using a GLM with a binomial distribution and a

logit link in SAS (GENMOD procedure). The report-

ed odds ratios (OR) from GLM model was comparing

the odds of Salmonella in one type of feed a group

to the odds of Salmonella in the other type of feed

(Dohoo et al., 2003). A sample was considered posi-

tive if direct plating was positive or if enrichment was

positive when direct plating was negative.

RESULTS

All uninoculated feed samples were negative for

Salmonella. Salmonella populations in inoculated

conventional and organic broiler feeds were deter-

mined over an 80-day period. Although the differ-

ences between populations in the two types of feeds

held at various temperatures/aw conditions were

small (generally < 1 log CFU/g), these differences

were statistically significant (P ≤ 0.05) within both

inoculum levels. For the feeds containing a high

inoculum, mean log CFU of Salmonella/g over the

80-day storage period were 4.71 ± 0.09 and 4.36 ±

0.09 for conventional and organic feed, respectively.

For feeds containing the low inoculum, the mean log

CFU of Salmonella/g were 2.88 ± 0.08 and 2.38 ±

0.08 for conventional and organic feed, respectively.

We considered differences in Salmonella popula-

tions between organic and conventional feeds to be

biologically meaningful if they were significantly dif-

ferent and >1 log CFU/g.

At day 0, the mean populations of Salmonella re-

Table 2. Mean Salmonella populations (log10 CFU/g) in conventional and organic poultry feed ini-tially containing high or low inoculum and at different temperature and aw combinations over an 80-day storage period.

Salmonella (log CFU/g)

Inoculum level aw Temp (Co) Conventional feed Organic feed P-valuea

High 0.43 11 6.17 5.93 <0.001

25 5.29 5.05 0.140

38 2.69 2.52 0.600

0.55 11 6.21 5.93 0.055

25 5.55 5.09 0.030

38 2.86 2.44 0.150

0.75 11 6.31 6.16 0.030

25 5.67 4.79 0.001

38 2.44 2.22 0.380

Low 0.43 11 4.29 4.12 0.020

25 3.26 2.64 <0.001

38 1.10 0.80 0.009

0.55 11 4.58 4.24 <0.001

25 3.89 3.15 <0.001

38 1.39 1.06 0.011

0.75 11 4.43 4.08 <0.001

25 2.96 1.86 <0.001

38 0.93 0.81 0.003aP-values were considered significant at ≤ 0.05 at different temperatures and aw.


covered from inoculated conventional and organic

feeds containing high and low inocula levels were

approximately 6 and 3 log CFU/g, respectively, indi-

cating an approximate 1-log CFU/g reduction based

on the number of Salmonella added to the feeds via

chalk inocula. The effects of different temperature/

aw combinations by feed type and inoculum level on

survival of Salmonella over an 80-day storage period

are shown in Table 2. The mean Salmonella popula-

tions in organic and conventional feeds initially con-

taining high and low inocula and stored at different

temperature/aw combinations for up to 80 days are

shown in Figures 1 and 2, respectively. There were

statistically significant differences (P ≤ 0.05) and >

1 log CFU/g difference in high-inoculum feeds (aw

0.75) stored at 25°C for 50 and 65 days (Fig. 1, F).

There were significant differences (> 1 log CFU/g)

in low-inoculum feeds stored at 25°C and aw 0.43,

0.55, and 0.75 during the study period (Fig. 2, D-F).

Regardless of inoculum level, Salmonella in organic

and conventional feeds decreased to an undetect-

able level (2 log CFU/g by direct plating; 1 CFU/10 g

by enrichment) in feeds stored at 38°C, regardless of

the aw (Fig. 1, G-I; Fig. 2, G-I).

Considering all temperature/aw combinations, the

OR for the presence of Salmonella in conventional

feed was significantly (P ≤ 0.05) higher than that in

organic feed containing a high inoculum (OR = 4.76

[95% confidence interval {CI}], 2.66 to 22.57). The OR

for presence of Salmonella in low-inoculum conven-

Figure. 1. Mean number of Salmonella (log CFU/g) in conventional and organic poultry feeds [high inoculum level (106 CFU/g)] over a storage period of 80 days at various temperature/aw conditions: A to C = 11oC, aw 0.43, 0.55, and 0.75; D to F = 25oC, aw 0.43, 0.55, and 0.75; and G to I = 38oC, aw 0.43, 0.55, and 0.75. Each data point represents the mean of values from three replicate samples for two experiments (n = 6) per treatment.

11˚C

25˚C

38˚C

0.43

A

D

G

0.55

B

E

H

0.75

C

F

I


tional feed was significantly higher (P ≤ 0.05) than for

organic feed (OR = 2.92 [95% CI, 2.16 to 4.53]). The

GLM model comparing Salmonella populations in

conventional feed versus organic feed (as the refer-

ence category) failed to converge due to the sparse

amount of data (no or few positives) for feed stored

at most temperature/aw combinations, regardless of

the inoculum level. The percentage of Salmonella-

positive conventional feed samples containing a

high inoculum level decreased by 50% at the end of

80-day storage period as compared to 43% in organ-

ic feed (Fig. 3). In contrast, at a low inoculum level,

the percentage of Salmonella-positive conventional

feed samples decreased 58% by the end of the 80-

day storage period as compared to 40% in organic

feed (Fig. 3).

DISCUSSION

The ability of Salmonella to survive in organic

versus conventional broiler feed as affected by stor-

age temperature and aw has not been reported, al-

though in the past decades, poultry feed has been

considered to be an important vehicle for transmis-

sion of Salmonella to broiler birds. Salmonella has

been detected in poultry feed and feed ingredi-

ents (Williams, 1981a; Williams, 1981 b; Cox et al.,

1983; Stuart, 1984; Veldman et al., 1995). In our ex-

periments, a dry inoculation technique rather than

a liquid-suspension inoculation was used to distrib-

ute Salmonella in feeds. The dry inoculum did not

substantially alter the aw of the feed. Compared to

a wet inoculum, distribution of Salmonella from dry

inocula is homogenously distributed and changes

in background microbiota are expected to be mini-

mum (Hoffmans and Fung, 1993). In addition, dry

inoculation more closely mimics contamination that

might occur via dust and other dry materials in grow-

out facilities.

We evaluated three temperatures and three aw

values covering ranges existing in storage facilities

at feed mills, during transportation, and on-farm

during different seasons. Poultry feeds are marketed

at aw ranging from about 0.45 to 0.75 and exposed

to various storage temperatures, (e.g., 11°C in win-

ter and fall, 25°C in spring, and 38°C in summer) in

southeastern U.S. These conditions were selected

to simulate environmental conditions for handling

and storage of poultry feed along the production-

to-consumption chain. Furthermore, to investigate

differences in contamination levels, poultry feeds

were inoculated with high and low populations of

Salmonella. Since poultry feeds are generally stored

for less than 8 to 10 weeks from production to con-

sumption, the period of 80 days was chosen to eval-

uate the survivability of Salmonella.

Although most of the differences in mean Salmo-

nella populations between the two types of feed, at

each temperature/aw combination, and by inoculum

population were statistically significant (Table 2),

the numerical differences were too small (<1.0 log

CFU/g) to be considered biologically meaningful.

The proportion of samples positive for Salmonella

was significantly higher in conventional feed com-

pared to organic feed for both inoculum levels over

all temperature/aw combinations. Differences in com-

position of organic and conventional feeds (Table 1)

did not appear to impact the survival of Salmonella;

however, feed composition may have impacted the

percentage of samples positive for Salmonella in the

two types of feed. A possible explanation for this dif-

ference may be attributable to the absence of ani-

mal protein meals (blood and bone) and presence

of natural antimicrobials such as garlic, clove, ginger,

and basil in organic feed. Leuschner and Zamparini

(2002) tested the effect of different natural antimicro-

bials such as garlic, ginger, mustard, and cloves on

growth and survival of Escherichia coli O157 and S.

enterica serotype Enteritidis in broth model systems.

Garlic and clove showed bacteriostatic and bacterio-

cidal effects on both foodborne pathogens. Clove

was found to be more effective than garlic. Mustard

and ginger also exhibited bacteriostatic activities

against both bacteria.

When examining survival of Salmonella as affect-

ed by storage time (Figs. 1 and 2), a few differences

>1 log CFU/g were observed, mainly in feed con-

taining a low inoculum and stored at 25°C. Salmonella

in both high- and low-inoculum feeds stored at 38°C

decreased to populations below the limit for detection


by direct plating (2 log CFU/g). Similar observations in

conventional poultry feed were reported by Williams

and Benson (1978), where S. Typhimurium survived

much longer at 11 and 25°C than at 38°C. In our study,

in feed containing a high inoculum, the mean num-

ber of Salmonella recovered from both types of feed

ranged from 4 to 5 log CFU/g after storage for 80 days

at 11 and 25°C, regardless of aw. This observation con-

curs with the findings of Davies and Wray (1996) show-

ing that S. Typhimurium declined to 3 log CFU/g of

conventional poultry feed over a period of 3 months.

These researchers inoculated feed using a suspension

of Salmonella that may have resulted in osmotic shock

of cells, thus immediately lowering the number of vi-

able Salmonella as compared to what may occur in the

feed to which a dry inoculum is added. The inoculation

procedure and aw of feeds have been reported to af-

fect the survivability of Salmonella (Liu et al., 1969; Carl-

son and Snoeyenbos, 1970; Juven et al., 1984).

Previous studies show that Salmonella does not

grow but survives well in low-aw foods such as peanut

butter, infant formula, cereals, and dry aniseed (GMA,

2009). These and other low-aw products have been

associated with outbreaks of salmonellosis. Investi-

gations suggest that factors such as poor sanitation

practices, poor equipment design, improper mainte-

Figure. 2. Mean number of Salmonella (log CFU/g) in conventional and organic poultry feeds [low inoculum level (103 CFU/g)] over a storage period of 80 days at various temperature/aw conditions: A to C = 11oC, aw 0.43, 0.55, and 0.75; D to F = 25oC, aw 0.43, 0.55, and 0.75; and G to I = 38oC, aw 0.43, 0.55, and 0.75. Each data point represents the mean of values from three replicate samples for two experiments (n = 6) per treatment.

25˚C

38˚C

11˚C

0.43

A

D

G

0.55

B

E

H

0.75

C

F

I


nance, and poor ingredient control were contributing

factors in these outbreaks. During storage, reduction

in Salmonella populations occurs but is dependent on

multiple factors, including temperature and aw. Feeds

and foods containing different ingredients but similar

moisture content can have different aw values, depend-

ing on the type and amount of solute present, leading

to different rates of inactivation (Duncan and Adams,

1972). In our study, there was no biologically meaning-

ful difference in the number of Salmonella recovered

from conventional and organic feeds stored at most

temperature/aw combinations, regardless of the inocu-

lum level, throughout the 80-day storage period.

Several researchers have shown that survival of Sal-

monella in dried products (e.g., fish meals, dry milk,

poultry feed, cocoa powder, pecans, and meat and

bone meal) is influenced by aw (Beuchat and Mann,

2010; Doesburg et al., 1970; Juven et al., 1984). Inter-

actions between aw and environmental factors such as

temperature may play an important role in affecting

the survival of Salmonella (Banwart and Ayres, 1956;

Corry, 1976; Doesburg et al., 1970). Juven et al. (1984)

showed that Salmonella can survive at higher popula-

tions in feeds at aw 0.43 and 0.53 compared to feeds at

aw 0.75. We observed a similar pattern for Salmonella

in conventional and organic broiler feeds. At aw 0.43

and 0.55, Salmonella survived at higher levels in both

types of feed as compared to survival at aw 0.75. Reduc-

tions were more rapid as the storage temperature was

increased.

ACKNOWLEDGEMENTS

We thank Dr. John Maurer from Poultry Diagnos-

tic Research Center, University of Georgia, Athens,

GA for providing the Salmonella isolates used in this

study. We also thank Rebekah Turk for her technical

assistance in the laboratory.

REFERENCES

Alali, W. Q., S. Thakur, R. D. Berghaus, M. P. Martin,

and W. Gebreyes. 2010. Prevalence and distribu-

tion of Salmonella in organic and conventional

broiler poultry farms. Foodborne Path. Dis. 7:1363-

1371.

Banwart, G. J. and J. C Ayres. 1956. The effect of

high temperature on the content of Salmonella

and on functional properties of dried egg white.

Food Technol. 10:68-73.

Beuchat, L. R. and D. A. Mann. 2010. Factors affect-

ing infiltration and survival of Salmonella in in-shell

pecans and nutmeats. J. Food Prot. 73:1257-1268.

Bryan, F. L. and M. P. Doyle. 1995. Health risks and

consequences of Salmonella and Campylobacter

jejuni in raw poultry. J. Food Prot. 58:326-344.

Carlson, V. L. and G. H. Snoeyenbos. 1970. Effect

of moisture on Salmonella populations in animal

feeds. Poult. Sci. 49:717-725.

Cason, J. A., N. A. Cox, and J. S. Bailey. 1994. Trans-

mission of Salmonella Typhimurium during hatch-

ing of broiler chicks. Avian Dis. 38:583–588.

Corry, J. E. L. 1976. The safety of intermediate mois-

ture foods with respect to Salmonella. In: R. Da-

vies, G. G. Birch, and K. J. Parker. Ed. Intermediate

moisture foods. Applied Science Publishers, Lon-

don. p 215-238.

Cox, N. A., J. S. Bailey, J. E. Thomson, and B. J. Ju-

ven. 1983. Salmonella and other Enterobacteria-

ceae found in commercial poultry feed. Poult. Sci.

62:2169-2175.

Figure. 3. Percentage of Salmonella-positive samples in conventional and organic poultry feeds initially containing a high- and low inocu-la and stored up to 80 days.


Davies, R. H. and C. Wray. 1996. Persistence of Sal-

monella Enteritidis in poultry units and poultry

food. British Poult. Sci. 37:589–596.

Doesburg, J. J., E. C. Lamprecht, and M. Elliot. 1970.

Death rates of Salmonellae in fish meals with dif-

ferent water activities. I. During storage. J. Sci.

Food Agric. 21:632- 635.

Dohoo, I. R., W. Martin, and H. Stryhn. 2003. Veteri-

nary Epidemiologic Research. Charlottetown, Can-

ada: University of Prince Edward Island.

Duncan, M. S. and A.W. Adams. 1972. Effects of a

chemical additive and of formaldehyde-gas fumi-

gation on Salmonella in poultry feeds. Poult. Sci.

51:797–802.

Furuta, K., I. Oku, and S. Sato. 1980. Bacterial con-

tamination in feed ingredients, formulated chicken

feed and reduction of viable bacteria by pelleting.

Lab. Anim. 14:221-224.

Grocery Manufacturers Association. 2009. Control

of Salmonella in low moisture foods. Gmaonline.

org. http://www.gmaonline.org/science/Salmonel-

laControlGuidance.pdf. Accessed May, 2009.

Halls, N. A. and A. Tallentire. 1978. Effects of pro-

cessing and gamma radiation on microbiologi-

cal contaminants of a laboratory animal diet. Lab

Anim. 12:5-10.

Hinton, M. and A. H. Linton. 1988. Control of Sal-

monella infections in broiler chickens by the acid

treatment of their feed. Vet. Rec. 123:416–421.

Himathongkham, S., M. Das Gracas Pereira, and H.

Riemann. 1996. Heat destruction of Salmonella in

poultry feed: effect of time, temperature and mois-

ture. Avian Dis. 36:625-631.

Hoffmans, C. M. and D. Y. C. Fung. 1993. Effective

method for dry inoculation of bacterial cultures. J.

Rapid. Meth. Auto. Microbiol. 1:287-294.

Jones, F. T., R. C. Axtell, D. V. Rives, S. E. Scheideler,

F. R. Tarver, R. L. Walker, and M. J. Wineland. 1991.

A survey of Salmonella contamination in modern

broiler production. J. Food Prot. 54:502-507.

Juven, B. J., N. A. Cox, J. S. Bailey, J. E. Thomson,

O. W. Charles, and J. V. Shutze. 1984. Survival of

Salmonella in dry food and feed. J. Food Prot.

47:445-448.

Kaiser, M. G., N. Lakshmanan, T. Wing, and S. J.

Lamont. 2002. Salmonella enterica serovars Enter-

itidis burden in broiler breeder chicks genetically

associated with vaccine antibody response. Avian

Dis. 46:25-31

Kang, D. H. and D. Y. C. Fung. 2000. Application

of thin agar layer method for recovery of injured

Salmonella Typhimurium. Int. J. Food Microbiol.

54:127-132.

Leuschner, R. G. K. and J. Zamparini. 2002. Effects of

spices on growth and survival of Escherichia coli

O157 and Salmonella enterica serovar Enteritidis

in broth model systems and mayonnaise. Food

Control. 13:399-404.

Liu, T. S., G. H. Snoeyenbos, and V. L. Carlson.1969.

Thermal resistance of Salmonella Senftenberg 775

W in dry animal feeds. Av. Dis. 13:611-631.

Maciorowski, K. G., F. T. Jones, S. D. Pillai, and S. C.

Ricke. 2004. Incidence, sources, and control of

food-borne Salmonella spp. in poultry feeds. W.

Poult. Sci. J. 60:446-457.

McCapes, R. H., H. E. Ekperigin, W. J. Cameron, W.

L. Ritchie, J. Slagter, V. Stangeland, and K.V. Nagra-

ja. 1989. Effect of new pelleting process on level of

contamination of poultry mash by Escherichia coli

and Salmonella. Av. Dis. 41:58-61.

Okelo, P. O., S. W. Joseph, D. D. Wagner, F. W. Whea-

ton, L. W. Douglass, and L. E. Carr. 2008. Improve-

ments in reduction of feed contamination: An al-

ternative monitor of bacterial killing during feed

extrusion. J. Appl. Poult. Res. 17:219-228.

Rockland, L. B. 1960. Saturated salt solutions for

static control of relative humidity between 5° and

40°C. Anal. Chem. 32:1375.

Scallan, E., R. M. Hoekstra, F. J. Angulo, R. V. Tauxe,

M. A. Widdowson, S. L. Roy. 2011. Foodborne ill-

ness acquired in the United States—major patho-

gens. Emerging Infect. Dis. 17:7-15.

Stuart, J. C. 1984. The introduction of Salmonella

into poultry flocks. J. Sci. Food Agric. 35:632-633.

Tauxe, R. V. 1991. Salmonella: A postmodern patho-

gen. J. Food Prot. 54:563-568.

Veldman, A., H. A. Vahl, G. J Borggreve, D. C. and

Fuller. 1995. A survey of incidence of Salmonella

species and Enterobacteriaceae in poultry feeds

and feed components. Vet. Res. 136:169-172.


Wilder, O. H. M. 1969. A method of destroying Sal-

monella. J. Am. Oil Chem. Soc. 46:233–234.

Williams, J. E. and S. T. Benson. 1978. Survival of Sal-

monella Typhimurium in poultry feed and litter at

three temperatures. Avian Dis. 22:742-747.

Williams, J. E. 1981a. Salmonella in poultry feeds-

a worldwide review. Part I: Introduction. World’s

Poult. Sci. J. 37:6-19.

Williams, J. E. 1981b. Salmonella in poultry feeds - a

worldwide review. Part III. Methods in control and

elimination. World’s Poult. Sci. J. 37:97-105.




ABSTRACT

Two of nine acetogenic ruminal isolates screened for plasmids were found to contain plasmid DNA.

Five plasmids ranging in size from 4.5 to 32 kilobase pairs (kb) were observed in isolate H3HH while

a single 35 kb plasmid was observed in isolate H4. The smallest of the plasmids from isolate H3HH,

estimated at 4.5 kb, was isolated using gel electrophoresis followed by electroelution. Of the 13 re-

striction endonucleases tested, this plasmid was cut once by EcoRV, SinI and HindIII and cut twice

BglII. The physiological functions of the individual plasmids are unknown. However, a plasmid-

free derivative (H3HP) of isolate H3HH displayed increased sensitivity to several antibiotics.

Keywords: Acetogens, plasmids, rumen, genetics, bacteria, isolation, characterization

INTRODUCTION

Interest in acetogenic bacteria stems from their

possible utilization as competitors against metha-

nogens as a hydrogen sink in the ruminal microbial

community, for acetate production from inexpensive

feedstocks, and their ability to utilize methoxylated

phenolics and phenolic acrylates (Drake, 1992; Hes-

pell, 1987; Ragsdale, 1991). Methane production rep-

resents a 5 to 15% loss of digestible feed energy to

ruminants (Blaxter and Clapperton, 1965). Shifting

any portion of the hydrogen used to form methane

towards acetate production could result in a concom-

itant increase in animal productivity.

Correspondence: J. Patterson, [email protected] Tel: +1 -765-494-4826 Fax: +1-765-494-9347

Although isolation and characterization of plas-

mids in other ruminal bacteria (e.g., Prevotella rumi-

nicola, Butyrivibrio fibrisolvens, Selenomonas rumi-

nantium) has been reported (Asmundson and Kelly,

1987; Dean et al., 1989; Flint et al., 1988; Flint and

Stewart, 1987; Mann et al., 1986; Martin and Dean,

1989; Ricke et al., 1996; Teather, 1982;) the genetic

systems of ruminal acetogenic bacteria remain un-

known (Forsberg et al., 1986). Analysis of plasmid

DNA in acetogens isolated from ruminal contents

is of interest because plasmids provide a means to-

wards developing understanding of acetogen physi-

ology and potentially for genetically manipulating

acetogens (Dean et al., 1989). This report describes

the isolation and initial characterization of plasmid

DNA from a ruminal acetogenic bacterium.

BRIEF COMMUNICATIONIsolation and Initial Characterization of Plasmids in

an Acetogenic Ruminal Isolate

R. S. Pinder1,2 , and J. A. Patterson1

1Department of Animal Sciences, Purdue University, West Lafayette, IN 47907-10262Current address: 7855 South 600 East, Brownsburg, IN 46112




Organisms and Media

Nine H2-utilizing, acetate producing isolates (H1

through H9) that had been isolated previously from

an H2 limited continuous culture system (Boccazzi et

al., 2011) were grown anaerobically in acetogen me-

dium which contained: (mg/L) K2HPO4, 300; KH2PO4,

300; (NH4)2SO4, 300; NaCl, 600; MgSO4 · 7 H2O, 123;

CaCl2· H2O, 80; NH4Cl, 540; HSeO3, 0.2; NiCl2 · 6 H2O,

20; hemin, 0.01; resazurin, 1; clarified rumen fluid, 5

ml; trace mineral solution (Greening and Leedle,

1989), 10 ml; yeast extract, 500. After the pH of the

medium was adjusted to 6.8 with HCl, the medium

was prepared using basic anaerobic techniques. The

medium was boiled in an appropriately sized reagent

bottle while being gassed with oxygen-free CO2. The

bottle was sealed with a rubber stopper, wired in

place and autoclaved (121ºC, 20 min). After cooling,

the flask was transferred into an anaerobic chamber

(Coy Laboratory Products, Ann Arbor, MI) and the fol-

lowing sterile anaerobic reagents were added asepti-

cally (in mg/liter): Na2CO3, 4000; cysteine · HCl, 250;

Na2S · 7 H2O, 250 and vitamin solution (Greening and

Leedle, 1989), 10 ml. Twenty ml of media was asep-

tically dispensed into sterile serum bottles (125 ml)

which were then capped with butyl stoppers (Bellco

Glass Co., Vineland, NJ).

Cultivation

When H2-grown cells were desired, a 10% inocu-

lum (from a culture at mid-log phase growing on H2/

CO2) was aseptically added to serum bottles which

were then flushed (for 30 sec) and pressurized to 200

kPa with a 80:20 mixture of H2/CO2. These bottles

were incubated at least 48 h with vigorous shaking

at 37ºC. If cells were to be used for plasmid isola-

tion, the medium was supplemented with 0.5% glu-

cose, inoculated (10% v/v), and incubated overnight

at 37ºC with gentle shaking (100 rpm). When larger

amounts of plasmid DNA were needed, the isolates

were grown anaerobically in brain heart infusion

broth (Difco Laboratories, Detroit, MI) overnight.

Plasmid Isolation

Plasmid DNA was extracted from isolates by the

procedure of Sanders and Klaenhammer (1983), with

slight modifications: Ten ml of culture were transferred

to sterile polypropylene centrifuge tubes (17 x 100

mm, Fisher Scientific Co., Pittsburgh, PA). The cells

were harvested by centrifugation (6,000 x g, 7 min,

0ºC), washed once with 10 ml of TES buffer (30 mM

Tris, pH 8; 5 mM EDTA; 50 mM NaCl), recentrifuged

and resuspended in 1 ml of sucrose buffer (50 mM

Tris, pH 7.5; 5mM EDTA; 25% (w/v) sucrose). Freshly

prepared lysozyme solution (10 mg/ml in 50 mM tris,

pH 8) was added to a final concentration of 1 mg/

ml followed by incubation (1 h, 0ºC). Sphaeroplasts

were separated from cell wall debris by centrifugation

(4,500 x g, 10 min, 0ºC) and resuspended in 0.5 ml of

glucose lysis buffer (50 mM Tris; 5 mM EDTA; 50 mM

glucose; 3% (w/v) sodium dodecyl sulfate; mixed with

4.3 μl of 10 N NaOH immediately prior to use). After

adding proteinase K (final conc. 0.1mg/ml) the pellet

was disrupted using a plastic disposable pipette and

the lysate incubated (1 h, 62ºC). The lysate was slowly

cooled to room temperature, 70 μl of 2 M Tris, pH 7 was

added, followed by 70 μl of 5 M NaCl. The lysate was

transferred to a 1.5 ml microcentrifuge tube along with

0.5 ml of phenol (saturated with 3 % NaCl), emulsified

with a vortexer (Genie mixer; Scientific Products, Inc.,

McGaw Park, IL) for 3 sec and incubated (5 min, RT). To

aid in separation of the phases, 0.3 ml of chloroform

was mixed in and the emulsion centrifuged (13,000 x

g, 5 min, RT). The upper phase (approx 0.6 ml) was re-

moved to a second microcentrifuge tube and 0.6 ml

of chloroform:isoamyl alcohol (24:1) added. After a

second centrifugation, 0.5 ml of the upper phase was

removed to a third tube. DNA was precipitated by

adding 2 vol of ice-cold (-20ºC) ethanol and incubat-

ing (>1 h, -60ºC). The precipitated DNA was recovered

by centrifugation (13,000 x g, 5 min, RT) and dissolved

in 30 μl of sterile distilled water. To reduce RNA inter-

ference, 1 μl of DNAse free RNAse (10 mg/ml) solution

was added, followed by incubation (20 min, 37ºC).

When larger quantities (greater than 1 mg) of plas-

mid DNA were required (i.e., for isolation of individual

plasmids), a procedure similar to that described by


Anderson and McKay (1983) was used. Three liters

of culture were centrifuged (8,000 x g, 7 min, 0ºC).

Pelleted cells were resuspended in 120 ml of STE

buffer (6.7% (w/v) sucrose; 50 mM Tris, pH 8; 1 mM

EDTA, pH 8). The cell suspension was divided into

four 250-ml polypropylene centrifuge bottles. After

adding 7.5 ml of fresh lysozyme solution (20 mg/ml

in 50 mM Tris, pH 8) to each bottle, the cells were

incubated on ice for 1 h, then at 37ºC for 10 min.

Next, 3.75 ml of chelating buffer (0.25 M EDTA, pH

8; 50 mM Tris, pH 8) was added to each bottle and

the cell suspension swirled for 10 sec. Cell lysis was

initiated by adding 2.25 ml of lysing solution (20 %

SDS, 50 mM Tris; 20 mM EDTA; pH 8) to each bottle

followed by vigorous mixing by hand for 15 sec. Nu-

clease activity in the cell suspensions was reduced

by adding 0.5 ml of proteinase K solution (10 mg/ml

in 50 mM Tris, pH 8) to each bottle and incubating

for 10 min at 37ºC. To complete cell lysis and irre-

versibly denature chromosomal DNA, 0.6 ml of 10 N

NaOH was added slowly while swirling the bottles

by hand. The bottles were swirled for an additional

10 min using a gyratory shaker (Model G76; New

Brunswick Scientific Co., New Brunswick, NJ).

In order to neutralize the pH of the lysates, 4 ml

of 2 M Tris, pH 7 was added to each bottle and swirl-

ing continued for 3 additional minutes. After mixing

in 5.7 ml of 5 M NaCl, the preparations were centri-

fuged (9,000 x g, 15 min, 0ºC). The supernatant from

each bottle was transferred to a second 250-ml cen-

trifuge bottle along with 55 ml of phenol (saturated

with 3% NaCl in distilled H2O). The preparations were

swirled for 10 minutes using the gyratory shaker, after

which 55 ml of chloroform was added into each bot-

tle. After centrifugation (6,000 x g, 10 min, 0ºC), the

upper aqueous phase was removed to a clean 250-ml

centrifuge bottle with an inverted 25 ml pipette and

set aside. Twenty-five ml of TES buffer was added to

the phenol:chloroform mixture and vigorously mixed

for 30 sec to extract additional DNA from the organic

phase. Following centrifugation (6,000 x g, 10 min,

0ºC), the upper aqueous phase was removed with an

inverted 25 ml pipette and mixed with the aqueous

fraction previously set aside. After adding an equal

volume of chloroform:isoamyl alcohol (24:1), the two

phases were mixed together, followed by centrifuga-

tion (6,000x g, 10 min, 0ºC). The upper aqueous layer

was removed and DNA was precipitated by adding

2 volumes of ice-cold (-20ºC) ethanol and incubation

overnight at -20ºC. The precipitated DNA was recov-

ered by centrifugation (9,000 x g, 45 min, -5ºC) and

resuspended in sterile distilled water and pooled.

The plasmid DNA preparation was kept at 4ºC until

used.

Individual plasmids from isolate H3HH were iso-

lated as follows: DNA was electrophoresed through

1 % agarose in TAE buffer (40mM tris-acetate, pH

7.8; 2mM EDTA, pH 7.8) in a horizontal electropho-

resis apparatus (Model H5, BRL, Inc., Gaithersburg,

MD). The plasmid bands were visualized by incubat-

ing the gel in an ethidium bromide solution (0.5 μg/

ml) for 10 min and briefly illuminating the gel with

ultraviolet light. The gel areas containing the de-

sired plasmid DNA were separated from the rest of

the gel with a razor blade. Plasmid DNA was extract-

ed from these gel strips by electroelution as follows:

The gel strip was inserted into dialysis tubing (Spec-

tro-por, 12000 to 14000 MWCO, Spectrum Medi-

cal Systems, Houston, TX). After removing most of

the buffer surrounding the gel slice, the tubing was

clamped shut and placed in the electrophoresis unit

filled with TAE buffer. The tubing was arranged so

that the gel slice was parallel to the electrodes. The

plasmid DNA was moved out of the agarose gel

slice by applying 5 V/cm (100 V) for 2 h. Following

a 1 min polarity reversal of the electric current, the

tubing was removed from the unit, unclamped and

the gel slice carefully removed. The contents of the

dialysis tubing were transferred to a polypropylene

centrifuge tube (17x 100 mm, Fisher Scientific Co.,

Pittsburgh, PA). The inside of the dialysis tubing was

rinsed with a small amount of sterile distilled water

which was then transferred to the centrifuge tube

containing the plasmid DNA. After an initial centrif-

ugation (13,000 x g, 5 min, 0ºC) to remove agarose

particles, the plasmid DNA was transferred to a mi-

crocentrifuge tube, precipitated with ethanol, and

pelleted by centrifugation (13,000 x g, 5 min, -5ºC).

The DNA was dissolved in 50 ul of sterile distilled wa-

ter and stored at 4ºC until used.


Restriction enzyme analysis

Plasmid DNA was digested by restriction endo-

nucleases under the conditions (e.g., salt concentra-

tion) specified by the supplier of each enzyme. The

plasmid preparations were incubated for 3 h with the

appropriate amount of enzymes. When digestions

with two different enzymes were performed simulta-

neously, conditions were adjusted appropriately to

optimize activity of both restriction endonucleases.

The size of plasmid DNA fragments was estimated by

electrophoresis through horizontal 1 % agarose gels

using TAE buffer. A 1 kb linear DNA ‘ladder’ (Gibco

BRL, Inc, Gaithersburg, MD) served as the molecular

weight marker and was used to estimate the size of

plasmid DNA fragments. After electrophoresis, plas-

mid DNA bands were visualized by soaking the aga-

rose gel in ethidium bromide solution for 20 minutes.

The gels were illuminated with a UV transiluminator

and photographed with a Polaroid MP-4 Land cam-

era system using type 57 Polaroid film.

Antibiotic sensitivity

The sensitivity of isolates H3HH and H3HP to vari-

ous antibiotics was determined as follows: 0.1 ml of an

overnight culture of isolate H3HH was spread across

the surface of 15 cm diameter petri dishes contain-

ing brain heart infusion broth (Difco Laboratories,

Detroit, MI) solidified with 2% agar. Once the surface

had dried, the appropriate antibiotic diffusion disks

(Difco Laboratories, Detroit, MI) containing one of

the following: novobiocin, vancomycin, kanamycin,

chloramphenicol, penicillin G, rifampicin, polymixin B,

streptomycin, erythromycin, tetracycline, gentamycin

or nalidixic acid were aseptically placed on the agar

surface separated by at least at least 3 cm. The plates

were incubated anaerobically for 24 h at 37ºC and the

diameter of the zones of inhibition measured.

Reagents

Restriction endonucleases were obtained from

Boeringer Mannheim (Indianapolis, IN) and Prome-

ga, Inc (Madison, WI). Tris, phenol, proteinase K, ly-

sozyme, and ethidium bromide were obtained from

Sigma Chemical Co. (St Louis, MO). SeaKem ME aga-

rose was obtained from FMC Corp (Rockland, ME).

Gases (H2, H2/CO2, N2, N2/CO2, CO2) were obtained

from Matheson Gas Products (Joliet, IL). All other re-

agents were of the highest purity commercially avail-

able.


The underlying objective for determining if

ruminal acetogenic isolates contain plasmid DNA

was to obtain the genetic machinery (i.e. origins of

replication) necessary for construction of genetic

transfer systems (i.e., shuttle vectors). Although shut-

tle vectors have been inserted into a non-ruminal

acetogen, we felt that use of replicons obtained from

plasmids found in acetogens isolated from ruminal

contents would enable construction of stable and

efficient shuttle vectors between the acetogens iso-

lated from ruminal contents and organisms with well-

researched genetic systems such as E. coli.

Several procedures, including those described by

Birnboim and Doly (1979), Kado and Liu (1981), and

Dean et al., (1989) were tested but resulted in poor

recovery (both quantitative and qualitative) of ace-

togen plasmid DNA (data not shown). The methods

described herein were adapted from published pro-

cedures (Anderson and McKay, 1983; Sanders and

Klaenhammer, 1983) designed for use with strepto-

cocci and lactococci.

Plasmid DNA was difficult to obtain from these

isolates due to high nuclease activity as well as a very

tough cell membrane. Deoxyribonuclease activity is

not unusual among ruminal bacteria and could pose

a barrier to development of gene transfer systems

(Flint and Thomson, 1990; Javorsky and Vanat, 1992).

However, in isolate H3HH, the deoxyribonuclease

activity could be reduced by treating the lysate with

proteinase K. Alternatively, when the lysate was heat-

ed to 95ºC for 10 min deoxyribonuclease activity was

reduced, but the reproducibility was variable and at

times plasmid DNA could not be visually detected in

ethidium bromide (EtBr) stained agarose gels. Incu-


bation of the cell suspension with lysozyme was very

important as only minimal lysis was achieved if this

step was omitted. A 1 h incubation at 0ºC was used

for both small- and large-scale preparations but an

additional 10 min incubation at 37ºC was necessary

during the large scale preparation to increase DNA

yields. Incubation with mutanolysin, another cell-

wall degrading enzyme, was not effective (data not

shown).

Isolate H3HH contained plasmid DNA ranging

in size from 32 to 4.5 kilobase (kb) (data not shown)

while a second isolate (H4) contained a single 35 kb

plasmid (data not shown). The presence of plasmid

DNA suggests that these isolates may be capable of

maintaining foreign DNA, a prerequisite for the ex-

ploration of the genetic systems of these organisms.

Because of the number of plasmid bands obtained

from isolate H3HH (which could be covalently closed

circular (CCC), open circular (OC) and linear forms

of a few plasmids or CCC form of many plasmids), it

was impossible to determine the actual number of

plasmids using conventional (unidimensional) aga-

rose gel electropheresis. However, using the tech-

nique of Hintermann et al. (1981) where plasmid

DNA was subjected to two dimensional agarose

gel electrophoresis with ultraviolet light irradiation

preceeding the second dimension electrophoresis,

five plasmid bands were identified in DNA obtained

from isolate H3HH (data not shown). The presence

of multiple plasmids in one organism is not unusual

and has been reported previously in other bacterial

species including: Escherichia coli V517 (Macrina et

al., 1978), Lactococcus lactis (Sanders and Klaen-

hammer, 1983).

Because of apparent high copy number (estimat-

ed by the relative band brightness of EtBr stained

gels) and relatively small size, the smallest plasmid

of isolate H3HH (named pRSP5) was selected for

characterization . A restriction map of the plasmid

was constructed by performing single and double

digestions with 13 restriction endonucleases (Fig. 1).

Although this plasmid was cut by SinI, EcoRV, Hin-

dIII and BglI, this plasmid did not appear to contain

cleavage sites for PstI, SalI, EcoRI, PvuI, PvuII, BamHi,

XhoI, SphI, and EcoRI. Analysis of the digested DNA

fragments suggested that pRSP5 is a circular mol-

ecule of approximately 4.5 kb.

The function of the individual plasmids has not

been established because sub-isolates containing a

single plasmid band have not been isolated. Howev-

er, the plasmids are unstable and disappear if isolate

H3HH is repeatedly transferred in rich media (such

as brain heart infusion). A plasmid-free derivative

(H3HP) of isolate H3HH (obtained after more than 10

transfers in brain heart infusion) displayed increased

sensitivity to antibiotics, suggesting the presence of

Figure 1. Physical map of pRSP5 obtained from isolate H3HH. Plasmid DNA was electrophoresed through a 0.7% agarose gel and briefly stained with ethidium bromide (0.5 μg/ml). The appropriate band was excised from the gel and the DNA electroeluted out of the agarose. Aliquots of the isolated plasmid DNA were subjected to restriction enzyme diges-tion and the resulting fragments electrophoresed through a 1% agarose gel , stained with ethidium bromide (0.5 μg/ml) and photographed. The length of the fragments was estimated by comparison with a 1 kb DNA ladder that was also electrophoresed through the gel. The fragments were aligned based on the digestion pattern obtained by simutaneous digestion with two restriction enzymes.


antibiotic resistance genes on these plasmids (Table

1). However, isolate H3HP is capable of growth on

H2/CO2 at rates similar to isolate H3HH (data not

shown), suggesting that these plasmids do not carry

genes essential for chemolithoautotrophic growth.

The work described here may open the door for

the potential genetic manipulation of acetogens iso-

lated from the rumen. Further research will be need-

ed to determine the genetic information (i.e., origin

of replication, promoters, open reading frames,

and other components) encoded in pRSP5. The re-

striction map of pRSP5 will prove helpful for further

analysis of this plasmid as well as providing sites for

insertion of genes, antibiotic resistance markers or

other vectors. Nevertheless, before further work on

preparation of shuttle vectors for acetogenic bacte-

ria is performed, a more thorough understanding of

the physiology (i.e., determination of control points

of the acetyl CoA pathway) and genetics (i.e., analy-

sis of the genes that encode enzymes as well as ac-

cessory proteins) of these bacteria is needed. Most,

if not all, of the enzymes involved in the acetyl-CoA

pathway have been purified to homogeneity and

characterized (Ragsdale, 1991). More recently, the

complete sequence has become available for cer-

tain nonruminal acetogens such as Moorella ther-

moacetica (f. Clostridium thermoaceticum) which

should aid genetic modifications (Pierce et al., 2008).

REFERENCES

Anderson, D. G. and L. L. McKay. 1983. Simple and

rapid method for isolating large plasmid DNA

from lactic streptococci. Appl. Environ. Microbiol.

46:549-552.

Asmundson, R. V. and W. J. Kely. 1987. Isolation and

characterization of plasmid DNA from Ruminococ-

cus. Curr. Microbiol. 16:97-100.

Birnboim, H. C. and J. Doly. 1979. A rapid alkaline

extraction procedure for screening recombinant

plasmid DNA. Nucleic Acids Res. 7:1513-1523.

Blaxter, K. L. and J. L. Clapperton. 1965. Prediction

of the amount of methane produced by ruminants.

Br. J. Nutr. 19:511-522.

Boccazzi, P., and J. A. Patterson. 2011. Using hydro-

gen limited anaerobic continuous culture to iso-

late low hydrogen threshold ruminal acetogenic

bacteria. Agric. Food Anal. Bacteriol. 1:33-44.

Dean, R. G., S. A. Martin, and C. Carver. 1989. Isola-

tion of plasmid DNA from the ruminal bacterium

Selenomonas ruminantium HD4. Lett. Appl. Micro-

biol. 8:45-48.

Drake, H.L. 1992. Acetogenesis and Acetogenic

bacteria. In: J. Lederberg. Ed. Encyclopedia of Mi-

crobiology. Academic Press, Inc. San Diego, CA.

p 1-15.

Flint, H. J. and C. S. Stewart. 1987. Antibiotic resis-

tance patterns and plasmids of ruminal strains of

Bacteroides ruminicola and Bacteroides multiaci-

dus. Appl. Microbiol. Biotechnol. 26:450-455.

Flint, H. J., S. H. Duncan, J. Bisset and C. S. Stewart.

1988. The isolation of tetracycline-resistant strains

of strictly anaerobic bacteria from the rumen. Lett.

Table 1. Antibiotic sensitivity of isolates H3HH and H3HP

Growth inhibition zone (mm)

antibiotic amount / disk

isolate H3HH

isolate H3HP

Chloramphenicol 30 μg 29a 37

Erythromycin 15 μg 27 36

Gentamycin 10 μg 17 31

Kanamycin 30 μg 10 26

Nalidixic acid 30 μg 8 8

Novobiocin 30 μg 24 43

Penicillin G 2 units 29 39

Polymixin B 300 units 8 23

Rifampicin 5 μg 19 22

Streptomycin 10 μg 7 20

Tetracycline 30 μg 35 48

Vancomycin 30 μg 23 34

a Overnight broth cultures (0.1 ml) were spread on 100 mm diameter plates containing brain heart infusion medium + 1% agar. Antibiotic disks were placed on the dried surface. Plates were incubated for 48 h inside an anaerobic incu-bator at 37oC. Values reported are mean (in mm) of two different zones, each in a different plate.



Flint, H. J. and A. M. Thomson. 1990. Deoxyribonu-

clease activity in rumen bacteria. Lett. Appl. Micro-

biol. 11:18-21.

Forsberg, C. W., B. Crosby, and D. Y. Thomas. 1986.

Potential for manipulation of the rumen fermenta-

tion through the use of recombinant DNA tech-

niques. J. Anim. Sci. 63:310-325.

Greening, R. C., and J. A. Z. Leedle. 1989. Enrich-

ment and isolation of Acetitomaculum ruminis,

gen. nov., sp. nov.: acetogenic bacteria from the

bovine rumen. Arch. Microbiol. 151:399-407

Hespell, R. A. 1987. Biotechnology and modifications

of the rumen microbial ecosystem. Proc. Nutr. Soc.

46:401-413.

Hintermann, G., H. M. Fischer, R. Crameri, and R.

Hutter. 1981. Simple procedure for distinguishing

CCC, OC, and L forms of plasmid DNA by agarose

gel electrophoresis. Plasmid 5:371-373.

Javorsky, P. and I. Vanat. 1992. Deoxyribonuclease

activity in Streptococcus bovis. Lett. Appl. Micro-

biol. 14:108-110.

Kado, C. I. and S.-T. Liu. 1981. Rapid procedure for

detection and isolation of large and small plas-

mids. J. Bacteriol. 145:1365-1373.

Macrina, F. L., D. J. Kopecko, K. R. Jones, D. J. Ayers,

and S. M. McCowen. 1978. A multiple plasmid-

containing Escherichia coli strain: convenient

source of size reference plasmid molecules. Plas-

mid 1:417-420.

Mann, S. P. ,G. P. Hazlewood, and C. G. Orpin. 1986.

Characterization of a cryptic plasmid (pOM1) in

Butyrivibrio fibrisolvens by restriction endonucle-

ase analysis and its cloning in Escherichia coli.

Curr. Microbiol. 13:17-22.

Martin, S. A. and R. G. Dean. 1989. Characterization

of a plasmid from the ruminal bacterium Sele-

nomonas ruminantium. Appl. Environ. Microbiol.

55:3035-3038.

Pierce, E., G. Xie, R. D. Barabote, E. Saunders, C.

S. Han, J. C. Detter, P. Richardson, T. S. Brettin, A.

Das, L. G. Ljungdahl, and S. W. Ragsdale. 2008. The

complete genome sequence of Moorella thermo-

acetica (f. Clostridium thermoaceticum). Environ.

Microbiol. 10:2550-2573.

Ragsdale, S. W. and L. G. Ljundahl. 1984. Hydrog-

enase from Acetobacterium woodii. Arch. Micro-

biol. 139:361-365.

Ricke, S. C., S. A. Martin, D. J. Nisbet. 1996. Ecology,

metabolism, and genetics of ruminal selenomo-

nads. Crit. Rev. Microbiol. 22:27-56.

Sanders, M. E. and T. R. Klaenhammer. 1983. Char-

acterization of phage-sensitive mutants from a

phage-insensitive strain of Streptococcus lactis:

evidence for a plasmid determinant that pre-

vents phage absorption. Appl. Environ. Microbiol.

46:1125-1133.

Teather, R. M. 1982. Isolation of plasmid DNA from

Butyrivibrio fibrisolvens. Appl. Environ. Microbiol.

43:298-302.


MANUSCRIPT SUBMISSION

Authors must submit their papers electronically ([email protected]). According to instructions pro-

vided online at our site: www.afabjournal.com. Authors who are unable to submit electronically should contact

the editorial office for assistance by email at [email protected].

• Aerobic microbiology

• Aerobiology

• Anaerobic microbiology

• Analytical microbiology

• Animal microbiology

• Antibiotics

• Antimicrobials

• Bacteriophage

• Bioremediation

• Biotechnology

• Detection

• Environmental microbiology

• Feed microbiology

• Fermentation

• Food bacteriology

• Food control

• Food microbiology

• Food quality

• Food Safety

• Foodborne pathogens

• Gastrointestinal microbiology

• Microbial education

• Microbial genetics

• Microbial physiology

• Modeling and microbial kinetics

• Natural products

• Phytoceuticals

• Quantitative microbiology

• Plant microbiology

• Plant pathogens

• Prebiotics

• Probiotics

• Rumen microbiology

• Rapid methods

• Toxins

• Veterinary microbiology

• Waste microbiology

• Water microbiology

CONTENT OF MANUSCRIPT

We invite you to consider submitting your research and review manuscripts to AFAB. The journal serves

as a peer reviewed scientific forum for to the latest advancements in bacteriology research on Agricultural

and Food Systems which includes the following fields:

INSTRUCTIONS TO AUTHORS


With an open access publication model of this

journal, all interested readers around the world can

freely access articles online. AFAB publishes origi-

nal papers including, but not limited to the types

of manuscripts described in the following sections.

Papers that have been, or are scheduled to be, pub-

lished elsewhere should not be submitted and will

not be reviewed. Opinions or views expressed in pa-

pers published by AFAB are those of the author(s)

and do not necessarily represent the opinion of the

AFAB or the editorial board.

MANUSCRIPT TYPES

Full-Length Research Manuscripts

AFAB accepts full-length research articles con-

taining four (4) figures and/or tables or more. AFAB

emphasizes the importance of sound scientific ex-

perimentation on any of the topics listed in the focus

areas followed by clear concise writing that describes

the research in its entirety. The results of experi-

ments published in AFAB must be replicated, with

appropriate statistical assessment of experimental

variation and assignment of significant difference.

Major headings to include are: Abstract, Introduc-tion, Materials and Methods, Results, Discussion (or Results and Discussion), Conclusion, Acknowl-edgements (optional), Appendix for abbreviations (optional), and References.

Manuscripts clearly lacking in language will be re-

turned to author without review, with a suggestion

that English editing be sought before the paper is

reconsidered. AFAB offers a fee based language

service upon request. Please contact [email protected] for more information about our fees

and services.

Rapid Communications

Under normal circumstances, AFAB aims for re-

ceipt-to-decision times of approximately one month or less. Accepted papers will have priority for publi-

cation in the next available issue of AFAB. However,

if an author chooses or requires a much more rapid

peer review, the journal editorial office has the capa-

bility to shorten the review timing to one week or less.

Any type of manuscript whether it be a full length

manuscript, brief communication or review paper can

be submitted as a rapid communication. There will be

additional costs for processing and page charges will

be double the normal rate. Authors who choose this

option must select Rapid Communications as the pa-

per type when submitting the paper and the editors

will judge whether a rapid review is possible and let

the author know immediately.

Brief Communications

Brief communications are short research notes giv-

ing the results of complete experiments but are con-

sidered less comprehensive than full-length articles

with three (3) figures and/or tables or less. Manuscripts

should be prepared with the same subheadings as full

length research papers. The running head above the

title of the paper is “Brief Communications.”

Unsolicited Review Papers

Review papers are welcome on any topic listed in

the focus section and have no page limits. Reviews

are assessed the same pages charges as all other

manuscripts. All AFAB guidelines for style and form

apply. Major headings to include are: Abstract, In-troduction, Main discussion topics and appropri-ate subheadings, Conclusions, Acknowledgements (optional) and References. Review papers shorter

than 20 pages of double spaced text and references

will be considered mini-reviews with the subhead-

ing above the title on the first page. The running

head above the title of the paper is either “Review”

or “Mini-review”.

Solicited Review Papers

Solicited reviews will have no page limits. The

editor-in-chief will send invitations to the authors

and then review these contributions when they are

submitted. Nominations or suggestions for potential

timely reviews are welcomed by the editors or edito-


rial board members and should be sent to submit@

afabjournal.com. There will be no page charges for

solicited review papers but the solicitation must origi-

nate from the editor-in-chief or one of the editors. Re-

quests from authors will automatically be classified as

unsolicited review papers. The running head above

the title of the paper will be “Invited Review.”

Conference and Special Issues Reviews

AFAB welcomes opportunities to publish papers

from symposia, scientific conference, and/or meet-

ings in their entirety. Conference organizers need

simply to contact AFAB at [email protected]

and a rapid decision is guaranteed. If in agreement,

the conference organizers must guarantee delivery

of a set number of peer reviewed manuscripts within

a specified time and submitted in the same format

as that described for unsolicited review papers. Con-

ference papers must be prepared in accordance with

the guidelines for review articles and are subject to

peer review. The conference chair must decide

whether or not they wish to serve as Special Issue

Editor and conduct the editorial review process. If

the conference chair/organizer chooses to serve as

special issue editor, this will involve review of the pa-

pers and, if necessary, returning them to the authors

for revision. The conference organizer then submits

the revised manuscripts to the journal editorial of-

fice for further editorial examination. Final revisions

by the author and recommendations for acceptance

or rejection by the chair must be completed by a

mutually agreed upon date between the editor and

the conference organizer. Manuscripts not meeting

this deadline will not be included in the published

symposium proceedings but if submitted later can

still be considered as unsolicited review papers. Al-

though offprints and costs of pages are the same

as for all other papers, the symposium chair may be

asked to guarantee an agreed upon number of hard

copies to be purchased by conference attendees. If

the decision is not to publish the symposium as a

special issue, the individual authors retain the right

to submit their papers for consideration for the jour-

nal as ordinary unsolicited review manuscripts.

Book Reviews

AFAB publishes reviews of books considered to

be of interest to the readers. The editor-in-chief ordi-

narily solicits reviews. Book reviews shall be prepared

in accordance to the style and form requirements of

the journal, and they are subject to editorial revision.

No page charges will be assessed solicited reviews

while unsolicited book reviews will be assigned the

regular page charge rate.

Opinions and Current Viewpoints

The purpose of this section will be to discuss, cri-

tique, or expand on scientific points made in articles

recently published in AFAB. Drafts must be received

within 6 months of an article’s publication. Opinions

and current perspectives do not have page limits.

They shall have a title followed by the body of the

text and references. Author name(s) and affiliation(s)

shall be placed between the end of the text and list

of references. If this document pertains to a par-

ticular manuscript then the author(s) of the original

paper(s) will be provided a copy of the letter and of-

fered the opportunity to submit for consideration a

reply within 30 days. Responses will have the same

page restrictions and format as the original opinion

and current viewpoint, and the titles shall end with

“Opinions.” They will be published together. Letters

and replies shall follow appropriate AFAB format

and may be edited by the editor-in-chief and a tech-

nical editor. If multiple letters on the same topic are

received, a representative set of opinions concern-

ing a specific article will be published. A disclaimer

will be added by the editorial staff that the opinion

expressed in this viewpoint is the authors alone and

does not necessarily represent the opinion of AFAB

or the editorial board.

COPYRIGHT AGREEMENT

The copyright form is published in AFAB as space

permits and is available online (www.afabjournal.com).


AFAB grants to the author the right of re-publication

in any book of which he or she is the author or edi-

tor, subject only to giving proper credit to the original

journal publication of the article by AFAB. AFAB re-

tains the copyright to all materials accepted for pub-

lication in the journal. If an author desires to reprint

a table or figure published from a non-AFAB source,

written evidence of copyright permission from an au-

thority representing that source must be obtained by

the author and forwarded to the AFAB editorial office.

PEER REVIEW PROCESS

Authors will be requested to provide the names

and complete addresses including emails of five (5) potential reviewers who have expertise in the research

area and no conflict of interest with any of the authors.

Except for manuscripts designated as Rapid Commu-

nication each reviewer has two (2) weeks to review

the manuscript, and submit comments electronically

to the editorial office. Authors have three (3) weeks

to complete the revision, which shall be returned to

the editorial office within six (6) weeks after which the

authors risk having their manuscript removed from

AFAB files if they fail to ask the editorial office for

an extension by email. Deleted manuscripts will be

reconsidered, but they will have to be processed as

new manuscripts with an additional processing fee as-

sessed upon submission. Once reviewed, the author

will be notified of the outcome and advised accord-

ingly. Editors handle all initial correspondence with

authors during the review process. The editor-in chief

will notify the author of the final decision to accept or

reject. Rejected manuscripts can be resubmitted only

with an invitation from the editor or editor-in chief. Re-

vised versions of previously rejected manuscripts are

treated as new submissions.

PRODUCTION OF PROOFS

Accepted manuscripts are forwarded to the edito-

rial office for technical editing and layout. The manu-

script is then formatted, figures are reproduced, and

author proofs are prepared as PDFs. Author proofs

of all manuscripts will be provided to the correspond-

ing author. Author proofs should be read carefully and

checked against the typed manuscript, because the

responsibility for proofreading is with the author(s).

Corrections must be returned by e-mail. Changes

sent by e-mail to the technical editor must indicate

page, column, and line numbers for each correction

to be made on the proof. Corrections can also be

marked using “track changes” in Microsoft Word or

using e-annotation tools for electronic proof correc-

tion in Adobe Acrobat to indicate necessary chang-

es. Author alterations to proofs exceeding 5% of the

original proof content will be charged to the author. All

correspondence of proofs must be agreed to by the

editorial office and the author within 48 hours or proof

will be published as is and AFAB will assume no re-

sponsibility for errors that result in the final publication.

PUBLICATION CHARGES

AFAB has two publication charge options: conven-

tional page charges and rapid communication. The

current charge for conventional publication is $25 per printed page in the journal. There is no additional

charge for the publication of pages containing color

images, micrographs or pictures. For authors who

wish to have their papers processed as a rapid com-

munication, authors will pay the rapid communication

fee when proofs are returned to the editorial office

in addition to twice the conventional page charges.

Charges for rapid communications are $1000 per manuscript for guaranteed peer review within one

week and $100 per journal page.

HARD COPY OFFPRINTS

If you are wishing to obtain a physical hard copy of

the AFAB journal, offprints are available in any quan-

tity at an additional charge: $100/page for black-white

and $150/page for color prints. You may order your

offprints at any time after publication on our website.

Scientific conference organizers may be expected to

agree to a set number of offprints as a part of their

agreement with AFAB.


MANUSCRIPT CONTENT REQUIREMENTS

Preparing the Manuscript File

Manuscripts must be written in grammatically

correct English. AFAB offers a fee based language

service upon request ([email protected]).

Manuscripts should be typed double-spaced, with

lines and pages numbered consecutively. All docu-

ments must be submitted in Microsoft Word (.doc or

.docx, PC or Mac). All special characters (e.g., Greek,

math, symbols) should be inserted using the sym-

bols palette available in this font. Tables and figures

should be placed in separate sections at the end of

the manuscript (not placed in the text). Failure to fol-

low these instructions will cause delays of the pro-

cessing and review of the manuscript.

Title Page

At the very top of the title page, include a title of

not more than 100 characters. Format the title with

the first letter of each word capitalized. No abbre-

viations should be used. Under the title, the authors

names are listed. Use the author’s initials for both first

and middle names with a period (full-stop) between

initials (e.g., W. A. Afab). Underneath the authors, a

list affiliations must be listed. Please use numerical

superscripts after the author’s names to designate

affiliation. If an authors address has changed since

the research was completed, this new information

must be designated as “Current address:”. The cor-

responding author should be indicated with an aster-

isk e.g., * Corresponding author. The title page shall

include the name and full address of the correspond-

ing author. Telephone and e-mail address must also

be provided for the corresponding author, and email-addresses must be provided for all authors.

Editing

Author-derived abbreviations should be defined

at first use in the abstract and again in the body of

the manuscript. If abbreviations are extensive au-

thors may need to provide a list of abbreviations

at the beginning of the manuscript. In vivo, in vitro

and bacterial names must be italicized (obligatory).

Authors must avoid single sentence paragraphs and

merge such paragraphs appropriately. Authors must

not begin sentences with “Figure or Table shows…”

as these are inanimate objects and cannot “show”

anything. When number are reported in text or in ta-

bles, always put a zero in front of decimal numbers:

“0.10” instead of “.10”.

MANUSCRIPT SECTIONS

Abstract

The abstract provides an abridged version of the

manuscript. Please submit your abstract on a sepa-

rate page after the title page. The abstract should

provide a justification of your work, objectives, meth-

ods, results, discussion and implications of study or

review findings . Your abstract must consist of com-

plete sentences without references to other work or

footnotes and must not exceed 250 words. On the

same page as your abstract, please provide at least ten (10) keywords to be used for linking and index-

ing. Ideally, these keywords should include signifi-

cant words from the title.

Introduction

The introduction should clearly present the foun-

dation of the manuscript topic and what makes the

research or the review unique. The introduction

should validate why this topic is important based on

previously published literature, and the relevance of

the current research. Overall goals and project ob-

jectives must be clearly stated in the final sentence

of the last paragraphs of the introduction.

Materials and Methods

Information on equipment and chemicals used

must include the full company name, city, and state

(country if outside the United States or Province if

in Canada) [i.e., (Model 123, ACME Inc., Afab, AR)].


Variability, Replication, and Statistical Analysis

To properly assess biological systems indepen-

dent replication of experiments and quantification

of variation among replicates is required by AFAB.

Reviewers and/or editors may request additional

statistical analysis depending on the nature of the

data and it will be the responsibility of the authors

to respond appropriately. Statistical methods com-

monly used in the bacteriology do not need to be

described in detail, but an adequate description

and/or appropriate references should be provided.

The statistical model and experimental unit must

be designated when appropriate. The experimen-

tal unit is the smallest unit to which an individual

treatment is imposed. For bacterial growth stud-

ies, the average of replicate tubes per single study

per treatment is the experimental unit; therefore,

individual studies must be replicated. Repeated

time analyses of the same sample usually do not

constitute independent experimental units. Mea-

surements on the same experimental unit over time

are also not independent and must not be consid-

ered as independent experimental units. For analy-

sis of time effects, assess as a rate of change over

time. Standard deviation refers to the variability

in the biological response being measured and is

presented as standard deviation or standard error

according to the definitions described in statistical

references or textbooks.

Results

Results represent the presentation of data in

words and all data should be described in same

fashion. No discussion of literature is included in

the results section.

DiscussionThe discussion section involves comparing the

current data outcomes with previously published

work in this area without repeating the text in the

results section. Critical and in-depth dialogue is

encouraged.

Results and Discussion

Results and discussion can be under combined or

separate headings.

Conclusions

State conclusions (not a summary) briefly in one

paragraph

Acknowledgments

Acknowledgments of individuals should include

institution, city, and state; city and country if not U.S.;

and City or Province if in Canada. Copies being re-

viewed shall have authors’ institutions omitted to re-

tain anonymity.

References

a) Citing References In Text

Authors of cited papers in the text are to be pre-

sented as follows: Adams and Harry (1992) or Smith

and Jones (1990, 1992). If more than two authors of

one article, the first author’s name is followed by the

abbreviation et al. in italics. If the sentence structure

requires that the authors’ names be included in pa-

rentheses, the proper format is (Adams and Harry,

1982; Harry, 1988a,b; Harry et al., 1993). Citations to a

group of references should be listed first alphabeti-

cally then chronologically. Work that has not been

submitted or accepted for publication shall be listed

in the text as: “G.C. Jay (institution, city, and state,

personal communication).” The author’s own un-

published work should be listed in the text as “(J.

Adams, unpublished data).” Personal communica-

tions and unsubmitted unpublished data must not

be included in the References section. Two or more

publications by the same authors in the same year

must be made distinct with lowercase letters after

the year (2010a,b). Likewise when multiple author ci-

tations designated by et al. in the text have the same

first author, then even if the other authors are differ-

ent these references in the text and the references

section must be identified by a letter. For example


“(James et al., 2010a,b)” in text, refers to “James,

Smith, and Elliot. 2010a” and “James, West, and Ad-

ams. 2010b” in the reference section.

b) Citing References In Reference Section

In the References section, references are listed in

alphabetical order by authors’ last names, and then

chronologically. List only those references cited in the

text. Manuscripts submitted for publication, accepted

for publication or in press can be given in the refer-

ence section followed by the designation: “(submit-

ted)”, “(accepted)’, or “(In Press), respectively. If the

DOI number of unpublished references is available,

you must give the number. The year of publication fol-

lows the authors’ names. All authors’ names must be

included in the citation in the Reference section. Jour-

nals must be abbreviated. First and last page num-

bers must be provided. Sample references are given

below. Consult recent issues of AFAB for examples

not included in the following section.

Journal manuscript:

Examples:

Chase, G. and L. Erlandsen. 1976. Evidence for a

complex life cycle and endospore formation in the

attached, filamentous, segmented bacterium from

murine ileum. J. Bacteriol. 127:572-583.

Jiang, B., A.-M. Henstra, L. Paulo, M. Balk, W. van

Doesburg, and A. J. M. Stams. 2009. A typical

one-carbon metabolism of an acetogenic and

hydrogenogenic Moorella thermioacetica strain.

Arch. Microbiol. 191:123-131.

Book:

Examples:

Hungate, R. E. 1966. The rumen and its microbes.

Academic Press, Inc., New York, NY. 533 p.

Book Chapter:

Examples:

O’Bryan, C. A., P. G. Crandall, and C. Bruhn. 2010.

Assessing consumer concerns and perceptions

of food safety risks and practices: Methodologies

and outcomes. In: S. C. Ricke and F. T. Jones. Eds.

Perspectives on Food Safety Issues of Food Animal

Derived Foods. Univ. Arkansas Press, Fayetteville,

AR. p 273-288.

Dissertation and thesis:

Maciorowski, K. G. 2000. Rapid detection of Salmo-

nella spp. and indicators of fecal contamination

in animal feed. Ph.D. Diss. Texas A&M University,

College Station, TX.

Donalson, L. M. 2005. The in vivo and in vitro effect

of a fructooligosacharide prebiotic combined with

alfalfa molt diets on egg production and Salmo-

nella in laying hens. M.S. thesis. Texas A&M Uni-

versity, College Station, TX.

Van Loo, E. 2009. Consumer perception of ready-to-

eat deli foods and organic meat. M.S. thesis. Uni-

versity of Arkansas, Fayetteville, AR. 202 p.

Web sites, patents:

Examples:

Davis, C. 2010. Salmonella. Medicinenet.com.

http://www.medicinenet.com/salmonella /article.

htm. Accessed July, 2010.

Afab, F. 2010, Development of a novel process. U.S.

Patent #_____

Author(s). Year. Article title. Journal title [abbreviated].

Volume number:inclusive pages.

Author(s) [or editor(s)]. Year. Title. Edition or volume (if

relevant). Publisher name, Place of publication. Number

of pages.

Author(s) of the chapter. Year. Title of the chapter. In:

author(s) or editor(s). Title of the book. Edition or vol-

ume, if relevant. Publisher name, Place of publication.

Inclusive pages of chapter.

Author. Date of degree. Title. Type of publication, such

as Ph.D. Diss or M.S. thesis. Institution, Place of institu-

tion. Total number of pages.


Abstracts and Symposia Proceedings:

Fischer, J. R. 2007. Building a prosperous future in

which agriculture uses and produces energy effi-

ciently and effectively. NABC report 19, Agricultural

Biofuels: Tech., Sustainability, and Profitability. p.27

Musgrove, M. T., and M. E. Berrang. 2008. Presence

of aerobic microorganisms, Enterobacteriaceae and

Salmonella in the shell egg processing environment.

IAFP 95th Annual Meeting. p. 47 (Abstr. #T6-10)

Vianna, M. E., H. P. Horz, G. Conrads. 2006. Options

and risks by using diagnostic gene chips. Program

and abstracts book , The 8th Biennieal Congress of

the Anaerobe Society of the Americas. p. 86 (Abstr.)

Data Presentation in Tables and Figures

Figures and tables to be published in AFAB must

be constructed in such a fashion that they are able

to “stand alone” in the published manuscript. This

means that the reader should be able to look at

the figure or table independently of the rest of the

manuscript and be able to comprehend the experi-

mental approach sufficiently to interpret the data.

Consequently, all statistical analyses should be very

carefully presented along with variation estimates

and what constitutes an independent replication

and the number of replicates used to calculate the

averages presented in the table or figure.

Each table and figure must be on a separate

page in the submitted paper. If your manuscript

is accepted for publication, you will need to sub-

mit all data for charts, tables and figures in Excel

spreadsheet format.

All figures should be clearly presented with well

defined axis and units of measurement. Symbols,

lines, and bars must be made distinct as “stand

alone” black and white presentations. Stippling,

dashed lines etc. are encouraged for multiple com-

parison but shades of gray are discouraged. Color

images, micrographs, pictures are recommended

and there is no additional fee for their submission.

AFAB Online Edition is Now Available!

www.AFABjournal.com

• Free Access

• Print PDFs

• Flip Through Issues

• Search Article Archives

• Order Reprints

• Submit a Paper

Online Publication: www.AFABjournal.com

afab-vol.1-issue 2

Documents

kwon university of arkansas

university of houston

hangcornell university

kirkhamkansas state

dunkley university of

shabaturafayetteville

crandalluniversity of

rickeuniversity of arkansas