Genotypes of multidrug-resistant Salmonella enterica serotypetyphimurium from two regions of Kenya

Samuel Kariuki a;c;*, Joseph O. Oundo a, Jane Muyodi a, Brett Lowe b,E. John Threlfall d, C. Anthony Hart c

a Centre for Microbiology Research, Kenya Medical Research Institute, PO Box 43640, Nairobi, Kenyab Wellcome Trust Research Laboratory, Kenya Medical Research Institute, PO Box 43640, Nairobi, Kenya

c Department of Medical Microbiology and Genito-Urinary Medicine, University of Liverpool, Liverpool L69 3GA, UKd Laboratory of Enteric Pathogens, Central Public Health Laboratory, Colindale, London NW9 5HT, UK

Received 3 April 2000; received in revised form 12 May 2000; accepted 24 May 2000

Abstract

A combination of phage typing and pulsed-field gel electrophoresis of XbaI-digested chromosomal DNA has been used to study theepidemiological relationships of multidrug-resistant Salmonella enterica serotype typhimurium from Nairobi (64 isolates) and Kilifi(40 isolates) collected over the period 1994^1997. Isolates from Nairobi belonged to 11 definitive phage types (DTs) encompassing eightdifferent PFGE patterns. In contrast, isolates from Kilifi were mainly DT 56 (60%) and all fell into a single PFGE pattern. The remainingisolates did not conform to a recognisable phage type. We conclude that multidrug-resistant S. typhimurium infections from Nairobi werecaused by multiple strains while those from Kilifi were likely to be from a microepidemic caused by a single clone. ß 2000 Federation ofEuropean Microbiological Societies. Published by Elsevier Science B.V. All rights reserved.

Keywords: Multidrug-resistant typhimurium ; Kenya; Phage type; Genotype

1. Introduction

Human infections due to non-typhoidal Salmonella spp.have been increasing since the 1980s and have been shownto be related to foodborne outbreaks worldwide. As only afew phage types tend to predominate within a certain geo-graphical region, phage typing has been used with greatsuccess to trace outbreaks [1,2]. For example, recently,multidrug-resistant (MDR) Salmonella enterica serotypetyphimurium (S. typhimurium) de¢nitive phage type (DT)104 has been implicated in most outbreaks caused by con-taminated foods of animal origin in both USA and inEurope [1,3]. However, it has not always been possibleto link an outbreak to a particular phage type. For exam-ple, during the period 1988^1992, DT 193 was the mostcommon typhimurium phage type causing infections in hu-mans in the UK, and most infections were traced to foodanimals, particularly cattle and pigs. However, studies ofthe distribution of insertion sequence IS200 elements con-

¢rmed that DT 193 was a heterogeneous phage type con-taining several distinct clones [4]. In other studies [5,6],pulsed-¢eld gel electrophoresis (PFGE) of macrorestrictedgenomic DNA has proved useful in providing additionalinformation for the epidemiological analysis of outbreak-related and unrelated Salmonella infections. The objectiveof the present study was to use phage typing and PFGE inorder to analyse epidemiological relationships of multipledrug-resistant S. typhimurium isolated from patients fromtwo di¡erent geographical areas of Kenya.

2. Materials and methods

2.1. Patients and bacterial isolates

Patients were febrile adults seen at the main public hos-pital in Nairobi (Kenyatta National Hospital) and chil-dren (1^84 months of age) seen at the District Hospitalin Kili¢, Coast Province some 500 km from Nairobi. Kili¢District Hospital caters for a mainly rural population andhas a specialised children's ward which was selected forthe study. Blood cultures and stool specimens were rou-

0928-8244 / 00 / $20.00 ß 2000 Federation of European Microbiological Societies. Published by Elsevier Science B.V. All rights reserved.PII: S 0 9 2 8 - 8 2 4 4 ( 0 0 ) 0 0 1 8 0 - 2

* Corresponding author. Tel. : +254 (2) 720163; Fax: +254 (2) 711673;E-mail : Skariuki@wtrl.or.ke

FEMSIM 1234 21-8-00

FEMS Immunology and Medical Microbiology 29 (2000) 9^13

www.fems-microbiology.org

tinely obtained from these patients in a sub-study evaluat-ing the aetiology of fever in hospital admissions. A total of64 non-duplicate Salmonella enterica serotype typhimurium(S. typhimurium) (55 from blood and 9 from stools) wereobtained from patients at the Kenyatta National Hospital,Nairobi and 40 isolates (28 from blood and 12 fromstools) were obtained children from Kili¢ District Hospi-tal. These isolates were selected for study on the basis ofbeing multi-drug resistant (resistant to two or more anti-microbials) between 1994^1997, and were frozen at 370³Cin trypticase soy broth with glycerol until analysed. Sero-typing was performed following the Kau¡mann^Whitescheme utilising Salmonella agglutination antisera (Murexdiagnostics, Dartford, UK).

2.2. Phage typing

Bacterial isolates con¢rmed as S. typhimurium were sentto the Laboratory of Enteric Pathogens, Central PublicHealth Laboratory, London, for phage typing.

2.3. Antimicrobial susceptibility

Susceptibility tests were performed on Isosensitest (Ox-oid, Basingstoke, UK) agar by the disk di¡usion tech-nique. Escherichia coli ATCC 25922 was used as the sen-sitive control. The antibiotic disks (Oxoid) used were:ampicillin 10 Wg, tetracycline 30 Wg, trimethoprim 5 Wg,sulfamethoxazole 100 Wg, chloramphenicol 30 Wg, strepto-mycin 10 Wg, gentamicin 10 Wg, co-amoxiclav 20:10 Wg,ceftazidime 30 Wg, cipro£oxacin, 3 Wg, nalidixic acid10 Wg. Disk zone sizes were interpreted according to theNCCLS guidelines [7].

2.4. Conjugation experiments and plasmid extraction

Conjugation experiments were performed in broth bythe method of Walia et al. [8] with E. coli K12 (nalidixicacid-resistant) as recipient. Transconjugants were selectedon MacConkey Agar (Oxoid) supplemented with nalidixicacid 32 mg l31 and ampicillin 32 mg l31. Plasmid DNAwas extracted from the transconjugants by an alkalinelysis method [9]. Plasmids were separated by electrophore-sis on horizontal agarose 0.8% gels at 100 V for 2 h.Plasmid sizes were determined by co-electrophoresis withplasmids of known sizes from E. coli strains V517 (53.7,7.2, 5.6, 3.9, 3.0, 2.7, 2.1 kb) and 39R861 (147, 63, 43.5,6.9 kb). DNA bands were visualised with an ultraviolettransilluminator (UVP, San Gabriel, CA, USA) afterstaining with ethidium bromide 0.05%.

2.5. PFGE of macrorestricted chromosomal DNA

Chromosomal DNA was prepared in agarose plugs asdescribed previously [10] from an overnight culture in Lu-ria broth. Agarose plugs were then equilibrated for 1 h in

0.5 ml of REACT2 bu¡er (Life Technologies, Paisley,UK) and incubated overnight at 37³C in fresh bu¡er(300 Wl), containing 25 units of XbaI. PFGE of agaroseplug inserts was then performed in a CHEF-DR II system(Bio-Rad, Richmond, CA, USA) on a horizontal agarose1% gel for 22 h at 120 V, pulse time of 1^40 s, at 14³C.Lambda DNA concatemers consisting of a ladder (ca.22 fragments) of increasing size from 48.5 kb to approx-imately 1000 kb was included as a DNA size standard.The gel was stained with ethidium bromide and photo-graphed on an UV transilluminator (UVP). The RE digestpatterns were interpreted by considering migration dis-tance and intensity of all visible bands, and by usingguidelines described by Tenover et al. [11]. Banding pat-terns were compared and genetic relatedness was deter-mined by data clustering using the unweighted pair group-ing arithmetic averaging method (UPGMA) (MolecularFingerprinting Program version 1.4.1, Bio-Rad).

3. Results

3.1. Antimicrobial susceptibility

All 64 S. typhimurium from Nairobi and the 40 isolatesfrom Kili¢ were resistant to 2 or more drugs includingampicillin, co-trimoxazole, streptomycin, tetracycline andchloramphenicol. A large proportion of isolates from Nai-robi (36; 56.3%) and Kili¢ (13; 33%) were multiply resis-tant to ampicillin, chloramphenicol, co-trimoxazole andstreptomycin.

3.2. Conjugation experiments and plasmids

A plasmid of ca. 100 kb was present in each of themultidrug-resistant S. typhimurium strains from Nairobi

Table 1Phage types of Salmonella enterica serotype typhimurium from Nairobiand Kili¢, Kenya

Phage type Number of isolates

Nairobi (n = 64) Kili¢ (n = 40)

56 17 24193 7 ^2 6 ^135 5 ^52 1 ^12 2 ^208 1 ^204a 3 ^49 2 ^29 1 ^RDNCa 9 16UNTb 10 ^

aS. typhimurium isolates not speci¢c to any of the standard phages usedfor typing.bUNT, S. typhimurium isolates untypable.

FEMSIM 1234 21-8-00

S. Kariuki et al. / FEMS Immunology and Medical Microbiology 29 (2000) 9^1310

and Kili¢, in addition to several plasmids ranging from 5^42 kb. In conjugation experiments, all 64 multidrug-resis-tant S. typhimurium from Nairobi, and 40 isolates fromKili¢ transferred plasmids of ca. 100 kb to E. coli K12. Ineach case, the 100-kb plasmids transferred resistance toampicillin; in addition resistance to co-trimoxazole, tetra-cycline and chloramphenicol were transferred together in19 (29.7%) and 9 (22.5%) donor-to-E. coli K12 conjuga-tion tests for isolates from Nairobi and Kili¢, respectively.

3.3. Phage types of S. typhimurium

The 64 S. typhimurium from Nairobi encompassed 11DTs. Most were DT 56 (17 isolates; 26.6%), followed byDT 193 (7; 10.9%) and 21 (32.8%) isolates were in eightother smaller phage type groups; 10 (15.6%) were untyp-able (these strains did not react with any of the currentlyavailable typing phages at the PHLS) and 9 (14.1%) wereRDNC (these strains reacted with some typing phages, butdid not conform to a recognised pattern) (Table 1). Incontrast, 24 isolates (60%) from Kili¢ were DT 56 and16 (40%) were RDNC.

3.4. Pulsed-¢eld gel electrophoresis groups

The 64 isolates from Nairobi could be divided into8 PFGE clusters (Table 2). Isolates within each cluster didnot di¡er by more than four bands. Ten (58.8%) isolatesof the most common phage type, DT 56, were in PFGEcluster 2 and the other seven were found to be distributedamong isolates in four other PFGE clusters. More than athird of the isolates from Nairobi (24; 37.5%) were in

Table 2Pulsed ¢eld gel electrophoresis cluster groups and phage types of S. typhimurium from Nairobi and Kili¢

PFGE cluster and (no. of isolates) Phage type and no. of isolates from

Nairobi Kili¢ Nairobi Kili¢

1 8 ^ 56 (2), 135 (1), 204a (1), 193 (1), RDNC (1), UNT (2) ^2 27 40 56 (10), 135 (1), 49 (2), 12 (1), 193 (1), 1 (2) 204a (1),

52 (1), 2 (1), RDNC (5), UNT (2)56 (24), RDNC (16)

3 7 ^ 56 (2), 29 (1), 1 (2), RDNC (1), UNT (1) ^4 2 ^ 56 (1), UNT (1) ^5 9 ^ 56 (2), 204a (2), 193 (2), 135 (1), 12 (1), UNT (1) ^6 8 ^ 193 (2), 135 (2), 204a (1), RDNC (2), UNT (1) ^7 1 ^ UNT (1) ^8 2 ^ 193 (1), UNT (1) ^

CFig. 1. (A) Restriction endonuclease fragment patterns of XbaI-digestedDNA from representative S. typhimurium isolates from Nairobi. Tracks1 and 16 contained the 48.5-kb DNA molecular mass standard. Tracks2^15 contained: 2, N150 (DT 56, cluster 2); 3, N3916 (DT 56, cluster2); 4, N3344 (DT 56, cluster 2); 5, N815 (DT 135, cluster 3); 6, N7830(DT 135, cluster 4); 7, N338 (DT 193, cluster 4); 8, N1487 (DT 135,cluster 4); 9, N592 (DT 1, cluster 5); 10, N710 (DT 1, cluster 5); 11,N617 (DT 12, cluster 3); 12, N844 (RDNC, cluster 4); 13, N2981(RDNC, cluster 4); 14, N7118 (untypable, cluster 3); 15, N235 (untyp-able, cluster 2). (B) Restriction endonuclease fragment patterns of XbaI-digested DNA from representative S. typhimurium isolates from Kili¢.Track 1 contained the 48.5-kb DNA molecular mass standard. Tracks2^21 contained: 2, K1 (DT RDNC); 3, K2 (DT 56); 4, K3 (DT 56);5, K4 (DT RDNC); 6, K5 (DT RDNC); 7, K6 (DT 56); 8, K7 (DT 56);9, K8 (DT 56); 10, K9 (DT 56); 11, K10 (DT 56); 12, K11 (DT 56);13, K12 (DT RDNC); 14, K13 (DT RDNC); 15, K14 (DT 56); 16,K15 (DT RDNC); 17, K16 (DT RDNC); 18, K17 (DT 56); 19, K18(DT 56); 20, K19 (DT 56); and 21, K20 (DT 56).

FEMSIM 1234 21-8-00

S. Kariuki et al. / FEMS Immunology and Medical Microbiology 29 (2000) 9^13 11

PFGE cluster 2. This cluster contained ten di¡erent de¢n-itive phage types. Apart from PFGE cluster 7, which con-tained one untypable isolate and cluster 8 which containedisolates of a single phage type, all other PFGE clusterscontained isolates from more than one phage type.When digested with XbaI, the 100-kb plasmids wereshown to produce fragments of less than 20 kb whichwere not scored during the analysis. PFGE analyses ofthe 40 isolates from Kili¢ showed that the banding pat-terns were similar for both DT 56 and RDNC as they fellinto the same PFGE cluster 2. Thus, S. typhimurium fromKili¢ were closely related, but not identical, to 10 out of17 DT 56 isolates from Nairobi. In total, therefore, 64(61.5%) of S. typhimurium from the present study wereclosely related by PFGE. The fragment patterns of XbaI-digested chromosomal DNA of MDR S. typhimurium iso-lates from Kili¢ and Nairobi are shown in Fig. 1A,B.

4. Discussion

A combination of phage typing and PFGE of XbaI-di-gested chromosomal DNA has been used to study S. ty-phimurium isolates from two geographically unrelated re-gions in Kenya. PFGE con¢rmed that MDR isolates fromNairobi were multiclonal, while those from Kili¢ wereessentially of the same clone. Although many reports ofinfection due to S. typhimurium in developed countries,such as in Europe [2,6], USA [3] and Canada [12] havebeen attributed to a single phage type often of an epidemicnature, our results indicate that this may not always be thecase in developing countries.

Forty S. typhimurium isolates from Kili¢ gave PFGEfragment patterns that were indistinguishable, suggestingthat all isolates represented strains within the same clone.The present study has also shown that S. typhimurium DT56 with a prevalence of 26.6% in Nairobi and 60% inKili¢, was the most common phage type in both areas.In addition, S. typhimurium isolates from Kili¢ that didnot conform to a recognisable phage type (RDNC) pro-duced PFGE fragment patterns that were indistinguishablefrom those of DT 56 strains suggesting that these RDNCstrains were closely related to DT 56. Previous studieshave shown that PFGE is superior to phage typing andantibiogram typing for Salmonella spp. [13]. For example,the acquisition or loss of lysogenic phages might result in achange in phage type. Thus, PFGE was able to furtherdiscriminate between strains of similar phage type. UsingPFGE, our results also indicate that a number of S. typhi-murium phage types are closely related, and that phagetyping is useful in analysing strains that are only epide-miologically closely linked [3]. Combining PFGE patternsand phage typing for S. typhimurium isolates from Kili¢indicates the existence of a MDR clone of micro-epidemicnature that has been previously unrecognised in this fairlystable rural community at the coastal region of Kenya. In

contrast, S. typhimurium isolates from Nairobi yielded atotal of 11 phage types distributed within eight PFGEpatterns indicating the existence of S. typhimurium strainsof a multiclonal nature. It appears from our study that S.typhimurium from Nairobi are more diverse, even withinthe same phage type. Due to the fairly high populationturnover and movement of the general population of Nai-robi it is possible that S. typhimurium isolates come fromdi¡erent parts of the country and that they cause sporadicinfections of a multiclonal nature rather than epidemicoutbreaks.

In developed countries MDR S. typhimurium DT 104with a common resistance phenotype of ampicillin, chlor-amphenicol, streptomycin, sulfonamides and tetracyclinehave been implicated in many cases of epidemic foodbornesalmonellosis [2,3,11]. This phage type has remained fairlysensitive to newer generation cephalosporins and quino-lones, which have been used as alternative drugs of choicefor treatment of MDR S. typhimurium. However, a signi¢-cant proportion of isolates of MDR DT 104 from the UKnow shows decreased sensitivity to cipro£oxacin [14]. Inthe present study, S. typhimurium were resistant to alldrugs commonly available in Kenya including ampicillin,cotrimoxazole, tetracycline, streptomycin and chloramphe-nicol. In addition these resistance phenotypes were self-transmissible on large 100-kb plasmids that was main-tained across all the di¡erent phage types and PFGE clus-ters. The high prevalence of MDR S. typhimurium com-plicates the treatment and management of these infectionsin developing countries including Kenya where e¡ectivealternative drugs for treatment of salmonellosis are oftenuna¡ordable or limited in availability, thus leading totreatment failure. In addition, unlike in developed coun-tries where S. typhimurium infections have often been as-sociated with contaminated foods of animal origin, thenatural reservoirs for non-typhoidal Salmonella spp. indeveloping countries including Kenya are not fully under-stood. Although there has been speculation that person-to-person transmission is important, case control studiesneed to be carried out in order to assess the risk factorsfor salmonellae infections in Kenya.

In conclusion, using phage typing and PFGE thepresent study has demonstrated that Salmonella infectionsdue to MDR S. typhimurium in Kili¢ were likely to beclonal in nature arising from a microepidemic in the re-gion. In contrast, those from Nairobi were likely to befrom sporadic outbreaks caused by a variety of straintypes.

Acknowledgements

The Director of the Kenya Medical Research Institutefor permission to publish this work. S.K. is supported byThe Wellcome Trust Research Development Award inTropical Medicine. We are grateful to Mrs. L.R. Ward

FEMSIM 1234 21-8-00

S. Kariuki et al. / FEMS Immunology and Medical Microbiology 29 (2000) 9^1312

from the Laboratory of Enteric Pathogens, CPHL forphage typing the isolates of S. typhimurium referred toin this study.

References

[1] Threlfall, E.J., Hampton, M.D., Chart, H. and Rowe, B. (1994) Useof plasmid pro¢le typing for surveillance of Salmonella enteritidisphage type 4 from humans, poultry and eggs. Epidemiol. Infect.112, 25^31.

[2] Sandvang, D., Aarestrup, F.M. and Jensen, L.B. (1998) Character-isation of integrons and antibiotic resistance genes in Danish multi-resistant Salmonella enterica Typhimurium DT104. FEMS Microbiol.Lett. 160, 37^41.

[3] Glynn, M.K., Bopp, C., Dewitt, M.S.W., Dabney, P., Mokhtar, M.and Angulo, F.J. (1998) Emergence of multidrug-resistant Salmonellaenterica serotype Typhimurium DT104 infections in the UnitedStates. New Engl. J. Med. 338, 1333^1338.

[4] Hampton, M.D., Threlfall, E.J., Frost, J.A., Ward, L.R. and Rowe,B. (1994) Use of plasmid pro¢le typing for surveillance of Salmonellaenteritidis phage type 4 from humans, poultry and eggs. Epidemiol.Infect. 112, 25^31.

[5] Puohiniemi, R., Heiskanen, T. and Siitonen, A. (1997) Molecularepidemiology of two international sprout-borne Salmonella out-breaks. J. Clin. Microbiol. 35, 2487^2491.

[6] Poirel, L., Guibert, M., Bellais, S., Naas, T. and Sordman, P. (1999)Integron- and carbenicillinase-mediated reduced susceptibility toamoxillin-clavulanic acid in isolates of multidrug-resistant Salmonellaenterica serotype Typhimurium DT104 from French patients. Anti-microb. Agents Chemother. 43, 1098^1104.

[7] National Committee for Clinical Laboratory Standards (1997) Meth-ods for dilution antimicrobial susceptibility tests for bacteria thatgrow aerobically. Approved standard M7-A4. National Committeefor Clinical Laboratory Standards, Wayne, PA.

[8] Walia, S.K., Madhavan, T., Chagh, T.D. and Sharma, K.B. (1987)Characterization of self-transmissible plasmids determining lactosefermentation and multiple antibiotic resistance in clinical strains ofKlebsiella pneumoniae. Eur. J. Clin. Microbiol. Infect. Dis. 7, 279^284.

[9] Birnboim, H.C. and Doly, J. (1979) A rapid alkaline extraction pro-cedure for screening recombinant plasmid DNA. Nucleic Acids Res.7, 1513^1523.

[10] Kariuki, S., Gilks, C., Kimari, J., Muyodi, J., Waiyaki, P. and Hart,C.A. (1999) Analysis of Salmonella enterica serotype Typhimuriumby phage typing, antimicrobial susceptibility and pulsed ¢eld gel elec-trophoresis. J. Med. Microbiol. 48, 1037^1042.

[11] Tenover, F.C., Arbeit, R.D. and Goering, R.V. et al. (1995) Inter-preting chromosomal DNA restriction patterns produced by pulsed-¢eld gel electrophoresis : criteria for bacterial strain typing. J. Clin.Microbiol. 33, 2233^2239.

[12] Ng, L., Mulvey, M.R., Martin, I., Petres, G.A. and Johnson, W.(1999) Genetic characterisation of antimicrobial resistance in Cana-dian isolates of Salmonella serovar Typhimurium DT104. Antimi-crob. Agents Chemother. 43, 3018^3021.

[13] Ling, J.M., Lo, N.W.S. and Ho, Y.M. et al. (2000) Molecular Meth-ods for the Epidemiological Typing of Salmonella enterica SerotypeTyphi from Hong Kong and Vietnam. J. Clin. Microbiol. 38, 292^300.

[14] Threlfall, E.J., Ward, L.R. and Rowe, B. (1999) Resistance to cipro-£oxacin in non-typhoidal salmonellae in England and Wales ^ thecurrent situation. Clin. Microbiol. Infect. 5, 130^134.

FEMSIM 1234 21-8-00

S. Kariuki et al. / FEMS Immunology and Medical Microbiology 29 (2000) 9^13 13