Vol. 54, No. 11

Stability of Bradyrhizobium japonicum Inoculants afterIntroduction into Soil

BRIGITTE BRUNEL,'* JEAN-CLAUDE CLEYET-MAREL,2 PHILIPPE NORMAND,3 AND RENE BARDIN'

Laboratoire d'Ecologie Microbienne CNRS U.A. 697, UniversiW Lyoni I, 43 Bd dii i1 noveinbre 1918, 69622 Villeur-banneCedex, France'; Laboratoire de Recherche siur les Symbiotes des Racines INRA, 34060 Montpellier Cedex, France2; and

Centre de Recherche en Biologie Forestie're, Universi Laval, Qu(ebec GIK 7P4, Canada3

Received 23 May 1988/Accepted 15 August 1988

Bradyrhizobium japonicum USDA 125-Sp, USDA 138, and USDA 138-Sm had been used as inoculants forsoybean (Glycine max (L.) Merr.) in soils previously free of B. japonicum. At 8 to 13 years after their release,these strains were reisolated from soil samples. A total of 115 isolates were obtained through nodules, and sevencolonies were obtained directly by a serological method. The stability of the inoculants was confirmed bycomparing the reisolated cultures with their respective parental strains which had been preserved by beinglyophilized or stored on a yeast extract-mannitol agar slant at 4°C. Comparisons were made on morphologicaland serological characters, carbon compound utilization (8 tested), intrinsic antibiotic resistance (9 tested), andenzymatic activity (19 tested). Mucous and nonmucous isolates of serogroup 125 were analyzed for symbioticeffectiveness and restriction fragment hybridization with a DNA probe. Our data suggest that the B. japonicuminoculants have survived for up to 13 years in the soils without significant mutation except for two reisolateswith a slightly increased kanamycin resistance level.

Strains of rhizobia are commonly added to agriculturalsoils around the world because of the economic benefitsprovided by the nitrogen fixed during their symbiotic asso-

ciation with host legumes. A rhizobial inoculant must com-

pete with persistent and well-adapted indigenous microor-ganisms during its saprophytic life and during nodulation ofan appropriate host plant. Other interacting soil factors alsoapply selective pressure on rhizobial populations, includingabiotic factors such as temperature, pH, humidity, and soilminerals and biotic factors such as plants and predators (13).In addition, soil bacteria, appear to be capable of substantialgene transfer with homologous bacteria and with phyloge-netically distant microbes (26, 28, 31). There are, therefore,several possible mechanisms which could lead to significantgenetic modifications in a microbial population introducedinto the field.

In the laboratory the problem of genetic stability inrhizobia has usually been assessed by studying symbioticproperties. Upon storage on agar media or as a result oftreatment with acriflavine, UV radiation, or heavy metalions, the loss of nodule-forming ability occurs strictly infast-growing rhizobia (33), where symbiotic genes are lo-cated on plasmids. Concerning slow-growing rhizobia,Ozawa et al. (20) have shown that a soil population ofsoybean-nodulating Bradyrhizobiuim japoniclum decreasedafter incubation at 30°C for 20 days or at 35 to 40°C for 15days.

In the field the stability of rhizobia has been studied less.Van Rensburg and Strijdom (34) showed that the symbioticeffectiveness of rhizobial inoculants (R. meliloti, R. legumi-nosarum bv. trifolii, B. japonicum, and a Rhizobiium sp.)generally did not change significantly from that of theirparental strains. The exceptions were that 60% of thereisolated cultures of R. trifolii were less effective and 11%of those of R. meliloti were more effective than theirrespective lyophilized parental cultures. Lotononis rhizobia

* Corresponding author.

introduced into soil had retained their antigenic propertiesover a 12-year period, but some of the reisolated cultureshad an N2-fixing ability lower than that of the initial inocu-lant, and the antibiotic sensitivity of the bacteria did notappear to be very stable (5).

B. japonicurm, which had been introduced into Frenchsoils for soybean crops, is normally absent in them; thesedeliberate introductions permitted us to follow the intro-duced bacterial population because serologically distinct B.japonicium strains had been used as inoculants. These strainsalso have other well-defined biological, biochemical, ge-netic, and physiological characteristics. The objectives ofour study were (i) to reisolate B. japonicium which had beenintroduced into French soils 8, 10, and 13 years ago and (ii)to determine whether these rhizobia had retained phenotypicproperties (morphology, serology, sugar utilization, antibi-otic resistance, and enzymatic activity). In addition, isolateswhich showed changing properties were also checked forgenetic and symbiotic variation.

MATERIALS AND METHODS

Bacterial strains. B. japoniclum USDA 125, USDA 138,and USDA 110 were obtained from the U.S. Department ofAgriculture, Beltsville, Md., and B. japonicum GMB1 was

obtained from IRA, Madagascar. Strains were grown on

yeast extract-mannitol (YEM) medium (35). The specti-nomycin-resistant (500 jig. ml-') strain USDA 125-Sp, thestreptomycin-resistant (1,000 jig- ml-') strains USDA 138-Sm and USDA 110-Sm, and the kanamycin-resistant (100jig ml-') strain GMB1-Km were spontaneous mutants ob-tained as described by Obaton (19). B. japonicum USDA125-Sp, USDA 138, USDA 138-Sm, GMB1-Km, and USDA110-Sm were the strains introduced in the field.Growth conditions. Cultures were grown at 28°C on YEM

(35) and stored on YEM agar slants at 4°C or as YEM liquidsuspensions containing 20% (wt/vol) glycerol at -80°C.Strains USDA 125-Sp and USDA 138 were lyophilized inYEM-horse serum (50% [wt/vol]) to give the lyophilized

2636

APPLIED AND ENVIRONMENTAL MICROBIOLOGY, Nov. 1988, p. 2636-26420099-2240/88/112636-07$02.00/0Copyright ©D 1988, American Society for Microbiology

on June 5, 2020 by guesthttp://aem

.asm.org/

Dow

nloaded from

STABILITY OF B. JAPONICUM INOCULANTS 2637

strains USDA 125-L-Sp and USDA 138-L. The parentalstrains were also stored on YEM slants at 4°C to give USDA125-Sp-G and USDA 138-Sm-G.

Field experiments. Field inoculations of commercial soy-bean were done on three sites. Site A (1,000 m2) was locatedon an Institut National de la Recherche Agronomique(INRA)-Montpellier (France) field. The physical and chem-ical properties of this soil were as follows: sand, 327mg g-'; silt, 512 mg g'1; clay, 162 mg. g-1 (pH 8.2);organic matter, 8 mg g-1; and N, 0.89 mg g-1. B. japoni-cum USDA 125-Sp and GMB1-Km were introduced by brothcultures into this site at a 50:50 ratio in 1977. The totalnumber of rhizobia were 5 x 104 g of dried soil-'. Sites Band C were located on an INRA-Toulouse (France) fieldwhich had the following soil properties: sand, 209 to 385mg g-1; silt, 348 to 422 mg. g-1; clay, 203 to 312 mg .-(pH 7.4 to 7.6); organic matter, 16 to 20 mg g-1; and N,1.14 to 1.45 mg. g-1. Site B (2,000 m2) was inoculated withonly USDA 138 in 1975 by peat cultures (1 x 106rhizobia. seed-'; 4 x 101" rhizobia. ha-'), whereas site C(150 m2) received broth cultures of USDA 138-Sm, USDA125-Sp, and USDA 110-Sm (ratio of 37:39:24) in 1972. Thetotal number of rhizobia was 8.5 x 104 g of dried soil-'. Noother soybean crop had been grown on the three sites sincethe rhizobial strain introductions.

Reisolation of introduced rhizobia by using host plants. Soilsamples were collected from the three sites at a depth of 5 to10 cm in 1985, at 8, 10, and 13 years after rhizobialintroduction into sites A, B, and C, respectively. Soybean(Glycine max (L.) Merr. cv. Kingsoy) seeds were surfacesterilized for 10 min in 3% (wt/vol) calcium hypochlorite,washed five times with sterile water, and soaked in water for2 h before germination on water agar plates (7 g- liter-').Seeds were germinated during 2 days in a growth chamber at28°C. Seedlings were transferred and cultivated in 5-literpots containing test soil from the three sites. Plants weregrown in a greenhouse with a 16-h photoperiod. Lighting wassupplemented with two 400-W mercury vapor lamps. Thedaytime and nighttime temperatures were 25 and 15°C,respectively. Individual pots contained four seeds, and 10replicates were performed for each soil type. Plants werewatered with a nitrogen-free nutrient solution containing 22mg of KH2PO4, 155 mg of KCl, 250 mg of MgSO4 7H20,215 mg of CaCl2. 2H20, 1 mg of MnSO4 2H20, 0.25 mg ofZnSO4 7H20, 0.25 mg of CuSO4- 5H20, 0.25 mg ofH3BO3, and 1 ml of Sequestrene (containing 6% iron;CIBA-GEIGY Corp., Rueil Malmaison, France) per liter ofdistilled water. Nodules were collected from 6-week-oldplants, surface sterilized in a 0.1% (wt/vol) mercuric chloridesolution for 90 s, and then rinsed repeatedly with steriledistilled water. Each nodule was squashed in a test tubecontaining 2 ml of saline solution (0.85% [wt/vol] NaCI), andthe resultant suspension was streaked onto a YEM agarplate. Plates were incubated at 28°C for 7 days, and asingle-colony isolate was kept from each nodule.

Reisolation of introduced rhizobia directly from soil. Weused a method of immunoisolation described previously byHranizky et al. (9) and modified as follows. Soil extractswere obtained by suspending 10 g of soil in 100 ml of sterilewater with a Waring blender. After 1 h the soil suspensionwas centrifuged at 600 x g for 10 min to remove most soilparticles, and the supernatant was subsequently centrifugedat 5,000 x g (30 min) to pellet fine particulates. The pelletwas rinsed three times with phosphate-buffered saline (PBS;pH 7.2) and then suspended in PBS-bovine serum albumin(1% [wt/vol]). The suspension was incubated for 1 h with

agitation at room temperature to saturate nonspecific anti-genic sites on soil particulates. The suspension was thenincubated in PBS overnight at 4°C with the filter-sterilizedanti-B. japonicum USDA 125 serum (dilution, 1/10; titer, 1/1,280). The treated samples were rinsed with PBS by cen-trifugation and incubated with 1 ml of protein A-Sepharose6MB beads (Pharmacia, Inc., Piscataway, N.J.) washed withPBS. This incubation was performed for 2 h at room tem-perature with gentle agitation. Rapidly sedimenting beadswere subsequently washed five times with PBS and platedwithout any dilution on YEM agar containing the following(in micrograms milliliter-'); spectinomycin, 1,000; cyclo-heximide (Sigma Chemical Co., St. Louis, Mo.), 50; andbenomyl (E.I. du Pont de Nemours & Co., Inc., Wilmington,Del.), 50. Colonies obtained on plates after incubation peri-ods of 1 to 2 weeks at 28°C were identified by serology (25).

Nodulation tests. All isolates from nodules or soil weretested for their ability to nodulate soybean (G. max cv.Kingsoy), using the Gibson tube technique (6). A total of 108late-log-phase bacteria cultivated in YEM broth were addedper plant. Plants were grown under the same greenhouseconditions as described above and watered with sterile plantmedium (6).

Serological identification. The identity of single-colonyisolates was determined serologically with fluorescent anti-sera prepared against strains USDA 125, USDA 138, andGMB1-Km as described by Schmidt (25).

Utilization of simple carbon compounds. The basal mini-mal medium contained 0.5 g of (NH4)2S04, 100 jig ofcyanocobalamine, 500 ,ug of nicotinic acid, 500 ,ug of pyro-doxin, 100 jig of biotin, 100 jig of panthotenic acid, 1.5 mg ofH3B03, 0.1 mg of MgSO4 5H20, 0.02 mg of(NH4)6Mo7024 4H20, 0.001 mg of CoS04 * 7H20, 10 mg ofcitric acid, 10 mg of ferric citrate, and 15 g of Noble agar ineach liter of distilled water. A total of 106 5-day-old cellscultivated on YEM agar were suspended in saline solution(0.85% [wt/vol] NaCI) and streaked on the minimal medium.Carbon sources (mannitol, arabinose, galactose, dextrose,inositol, maltose, sucrose, and lactose) were supplied assterile Glucidiscs (BioMerieux, Lyon, France) and depositedon the surface of agar plates. Bacterial growth after 15 daysat 28°C was compared with controls without sugar andscored for visible growth.

Antibiotic sensitivity testing. Cells from 5-day-old YEMagar-grown cultures were streaked onto the surface of YEMplates containing one of the following antibiotics: kanamycinsulfate, streptomycin sulfate, erythromycin, rifampin, tetra-cycline hydrochloride (Boehringer Mannheim Biochemicals,Indianapolis, Ind.), ampicillin sodium salt, spectinomycin(Sigma), and thiostrepton (a gift from S. J. Luciania, E. R.Squibb & Sons, Princeton, N.J.). These antibiotics weretested at concentrations ranging from 0.1 to 1,000 jg ml ofYEM-1. Antibiotic plates were incubated at 28°C for 5 to 7days, and bacterial growth was compared with their growthon YEM control plates.Enzymatic tests. Bacteria were collected from 5-day-old

YEM liquid cultures and washed three times with salinesolution (0.85% NaCl [wt/vol]). Both the bacterial pelletsand the culture supernatants were stored at -20°C. Thepellet was sonicated in ice water for 3 min at 50 W (Lab-sonic; Poly Labo Paul Block & Cie, Strasbourg, France) andsubsequently centrifuged for 20 min at 5,000 x g. Thesupernatant was assayed for activity of intracellular en-zymes, whereas the culture supernatant was assayed forexcreted enzymes. The enzymatic tests (alkaline phospha-tase, esterase, esterase lipase, lipase, leucyl arylamidase,

VOL. 54, 1988

on June 5, 2020 by guesthttp://aem

.asm.org/

Dow

nloaded from

APPL. ENVIRON. MICROBIOL.

valine arylamidase, cystine arylamidase, trypsin, ot-chymo-trypsin, acid phosphatase, naphthol-AS-BI-phosphohy-drolase, ox-galactosidase, P3-galactosidase, 3-glucuronidase,ox-glucosidase, 3-glucosidase, N-acetyl-43-glucosaminidase,cx-mannosidase, and (x-fucosidase) were done with a com-

mercial system (API ZYM, Api System, S.A., Montalieu,Vercieu, France) used according to the instructions ofthe manufacturer with enzyme reactions performed at 37°Cfor 3 h.

Effectiveness testing. Symbiotic effectiveness was mea-

sured with eight isolates chosen from site A; four were

mucous and four were nonmucous on YEM agar, and one ofthem grew on kanamycin (10 p.g ml-'). Tests were per-

formed on soybean plants (G. mnax cv. Kingsoy) grown in thegreenhouse under the same conditions as described aboveexcept that 5-liter pots contained only sterilized sand withone seedling per pot. Plants were watered with the nitrogen-free nutrient solution described above. The different strainswere grown to the early stationary phase in YEM mediumand adjusted by optical density at 620 nm to obtain 107bacteria seed-1. Four replicates for each isolate or strainwere done, and the dry weights of inoculated plants (60 daysafter inoculation) were determined and compared with con-

trols. The greenhouse experiments were arranged in a ran-

domized complete block experimental design, and plantweight data were subjected to an analysis of variance (F test

for the ratio of variances) according to Dagn6lie (4).Southern hybridizations. Total DNA was isolated from 10

serogroup 125 isolates from nodules and parental strainsUSDA 125-Sp-L and USDA 125-Sp-G by a modification ofthe Marmur technique (15). Cultures were grown to thestationary phase in YEM medium. The cells (10 ml) were

centrifuged, washed in TE medium (50 mM Tris, 20 mMdisodium EDTA; pH 8.0), and suspended in 10 ml of TEmedium containing lysozyme (5 mg ml-'; Sigma). Sodiumdodecyl sulfate (0.5% [wt/vol]) and pronase B (Sigma; 5mg ml-'; predigested for 1 h at 37°C in TE medium) were

added, and the suspension was incubated for 1 h at 37°C andthen for 15 min at 65°C. The viscous solution was extractedtwice by adding an equal volume of phenol-chloroform-isoamyl alcohol (24/24/1, vol/vol/vol). DNA was precipitatedin the presence of 1 volume of isopropanol and 0.03 Macetate sodium, pelleted by centrifugation, washed (first in70% ethanol and then in 95% ethanol), and dried undervacuum at room temperature. The dried DNA was sus-

pended in 0.1 ml of TE buffer. Each DNA preparation was

digested with restriction enzymes Ba,nHI, Pstl, and SstIaccording to the instructions of the manufacturer and sub-sequently electrophoresed at 1 V cm-' for 10 h in a 0.8%(wt/vol) agarose horizontal gel in tris borate buffer (14). The

DNA in the gel was then transferred to nitrocellulose (29).Purified restriction fragments carrying the n?ifD gene, iso-lated from agarose gels by a freeze-squeeze procedure (30),were labeled with -32P]dCTP (11.1 x 1013 Bq mmolV;Amersham International, Amersham, England) by nicktranslation (21). Specific activities of 18.5 x 1012 Bq pmolof DNA-' were obtained. Prehybridizations (65°C, 1 h) andhybridizations (55°C, 15 h) were performed essentially as

described by Simonet et al. (27). X-Omat-AR film (Eastman

Kodak Co., Rochester, N.Y.) was exposed to the dried

nitrocellulose filter with an intensifying screen at -80°C for

15 h.Cloning of the B. japonicum DNA fragment homologous to

nWMD. Approximately 50 pug of DNA from B. japonicumUSDA 125 was partially digested with Sau3A1 to yieldpredominantly 10- to 20-kilobase (kb) fragments as described

by Maniatis et al. (14). These fragments were ligated to theBainHI-digested EMBL4 vector in an overnight reaction at15°C, using 1 U of T4 DNA ligase (Boehringer Mannheim).The ligation mixture was in vitro packaged into phageparticles which were subsequently used to transfect cells ofEscherichia oli P2392 (Vector Cloning Systems; Genofit,Geneva, Switzerland). From the 2,000 plaques screened, onephage was isolated which contained DNA that hybridizedwith a i[fJD probe from Franslkia sp. (P. Normand, P. Simo-net, and R. Bardin, Mol. Gen. Genet., in press). Theisolation of recombinant phages and DNA isolation werecarried out as described by Maniatis et al. (14).

RESULTSReisolation of B. japonicum after field inoculation. A total

of 115 isolates were obtained from soil samples by nodula-tion of soybean. These B. japonwiun isolates belonged toeither serogroup 125 or serogroup 138. At site A, weobtained 45 isolates, all belonging to serogroup 125. GMB1-Km was not reisolated from this site. At site B, 44 coloniesbelonging to serogroup 138 were isolated. At site C, weisolated 25 serogroup 138 colonies and only 1 serogroup 125colony. No serogroup 110 isolate was obtained.The use of the antibiotic-resistant strains did not permit

the reisolation of B. japoniculn directly from the soil of sitesA and C. From soil samples of site A, none of the 290 Sp"bacteria and none of the 90 Km'- bacteria belonged toserogroup 125 or GMB1. Furthermore, none of the 310 Sprbacteria or the 280 Sm'r bacteria isolated from the soil of siteC belonged to serogroup 125 or 138. However, at site A weobtained by immunoisolation seven serogroup Sp'r colonieswhich were still able to nodulate soybean.

Utilization of simple carbon compounds. All of the eightcarbon compounds tested were used by the isolates in amanner similar to the parental strains which had beenconserved lyophilized or on YEM slants at 4°C (15 clones ofeach tested). Compared with parental strains, growth of theisolates was not detectably retarded. Arabinose, galactose,and dextrose were used for all 122 isolates and the 60parental clones. Inositol, maltose, sucrose, and lactose werenot used. The only observed variation was in mannitolutilization: all isolates from sites B and C grew on mannitolminimal medium, whereas 39% of the isolates from site Agrew well and the rest grew poorly on mannitol minimalmedium. The colonies of these two types of isolates weremorphologically different on YEM agar plates, the formerbeing mucous and the latter being less mucous or nonmu-cous. This variation was also detected in parental strainUSDA 125-Sp-G, where only 3 of the 15 clones tested grewwell on mannitol.

Determination of intrinsic antibiotic resistance. Nine anti-biotics belonging to nine different antibiotic families weretested. The intrinsic antibiotic resistance concentrations forall of the field isolates in serogroups 125 and 138 werevirtually identical to those of their respective parent strains(Table 1). The mutant strains had also retained their respec-tive antibiotic resistance marker. The resistance of ampicil-lin permitted the discrimination of mucous and nonmucousserogroup 125 subpopulations on YEM agar, the formerhaving a higher resistance than the latter. Finally, we de-tected two isolates belonging to serogroup 125 (one mucousand one nonmucous) which were unusually tolerant tokanamycin. Furthermore, USDA 125 and USDA 138 couldbe differentiated with tetracycline, rifampin, and ampicillin.

Enzymatic activities. All 122 field isolates were tested for19 enzymatic activities by using the semiquantitative API

2638 BRUNEL ET AL.

on June 5, 2020 by guesthttp://aem

.asm.org/

Dow

nloaded from

STABILITY OF B. JAPONICUM INOCULANTS 2639

TABLE 1. Intrinsic antibiotic resistance of B. japoniciulin isolates belonging to serogroups 125 and 138

No. of Antibiotic resistance concn (,ug ml of YEM1)"'Serogroupisolates Sm Sp Km Tc Rif Amp Thi Cm Ery

Site A (125) 17"' 1 1,000 1' 10 1 250 1 10 1028"/ 1 1,000 1' 10 1 10 1 10 107" 1 1,000 1 10 1 10 1 10 10

Site B (138) 44 1 10 10 100 10 100 1 10 10

Site C125 1 1 1,000 1 10 1 250 1 10 10138 25 1,000 10 10 100 10 100 1 10 10

Parental strainsUSDA 125-Sp-G' 3b 1 1,000 1 10 1 250 1 10 10

12" 1 1,000 1 10 1 10 1 10 10USDA 125-Sp-Lf 15 1 1,000 1 10 1 250 1 10 10USDA 138-Sm-G 15 1,000 10 10 100 10 100 1 10 10USDA 138-L 15 1 10 10 100 10 100 1 10 10

" Sm, Streptomycin; Sp, spectinomycin; Km. kanamycin. Tc. tetracycline- Rif, rifampin: Amp. ampicillin: Thi. thiostrepton: Cm. chloramphenicol: Ery.erythromycin.

t Mucous isolates on YEM agar.'*One isolate grew on 10 Ftg of kanamycin- mld Nonmucous isolates on YEM agar.Isolates obtained directly from soil.

f Parental strains conserved on YEM slants at 4'C (G) and lyophilized (L).

ZYM method. They presented enzymatic activity profilesthat were the same as those of the parental clones (Table 2).Phosphatase, esterase, lipase, arylamidase, and phosphohy-drolase were the intracellular enzymes detected after 5 daysof growth on YEM broth. The most intense activity wasalways that of esterase, with a higher activity level detectedin USDA 138 than in USDA 125. No activity was detectedfor the other enzymes tested, i.e., cystine arylamidase,x-chymotrypsin, ox-galactosidase, 3-galactosidase, 3-gluc-uronidase, oa-glucosidase, ,3-glucosidase, N-acetyl-,B-gluco-saminidase, cx-mannosidase, and cx-fucosidase. Enzymaticassays of culture supernatants revealed the presence of thesame enzymes detected in intracellular extracts.

Genetic analysis. Genomic DNAs from 10 serogroup 125isolates were digested with three restriction enzymes. As

TABLE 2. Enzymatic activity of B. japonhiuin isolatesas determined by API ZYM

Enzyme activity"

Isolates of Parental strainsEnzyme serogroup:

USDA USDA125" 138' 125-Sp-G 138-Sm-G

and -L" and -L"

Alkaline phosphatase + + + +Esterase +++ +++ +++ +++Esterase lipase ++ ++ ++ ++Lipase + + + +Leucyl arylamidase + + + +Valine arylamidase + + + +Trypsin 0 + 0 +Acid phosphatase + + + +Naphthol-AS-BI-phosphohydrolase + + + +

' 0, No activity detected; +, <15 U; + +, approximately 20 U; + + +. >30U (where 1 U is the release of 1 nmol of product by 10' bacteria ml-' in 3 h).

b Sites A and C; 53 isolates."Sites B and C; 69 isolates.d Parental strains lyophilized (L) and conserved on YEM slants at 4'C (G);

30 clones.

expected, similar banding patterns were found in the fieldisolates and the parental strains (Fig. 1) with all threeenzymes. Southern blots of DNAs were hybridized to a32P-labeled fragment cloned from parental strain USDA125-Sp-L. This cloned 15-kb fragment hybridized with the6.6-kb EcoRI fragment of pSA30 from Klebsiella pnieiumtio-niiae (1), a fragment which contains the structural genesniifHDK (data not shown). Figure 1 shows the positivelyhybridizing fragments observed on the Southern blots withthe nif gene probe. Results (Fig. 1, bottom) indicate that allof the isolates had hybridization patterns similar to those oftheir respective parental strains. For BamtiHI, each isolatehad 7.4- and 2.6-kb hybridizing fragments; for PstI, eachisolate had 1.5- and 1.1-kb bands; and for SstI, each isolatehad three major bands of 5.9, 4.7, and 2.9 kb.

Symbiotic properties. The statistical measure of variabil-ity, the F test for the ratio of variances, indicated nosignificant phenotypic differences (at the P = 0.05 level)between original cultures and eight isolates (four mucous[no. 2, 3, 4, and 6] and four nonmucous [no. 1, 5, 7, and 8] onYEM plates [one of them, isolate 8, grew on 10 ,ug ofkanamycin ml-1) of B. japoniculn serogroup 125 with re-spect to symbiotic nitrogen-fixing ability with soybean.Furthermore, the seven isolates obtained directly from soilhad retained the ability to nodulate the host plant and formednodules as rapidly as the parental strains did. At 60 daysafter inoculation, the plant dry weights (in grams) were asfollows for the serogroup 125 isolates (values are means +

standard error): 1, 21.0 + 3.5; 2, 22.3 + 5.4; 3, 21.3 ± 4.2; 4,18.9 + 4.2; 5, 19.2 ± 0.8; 6, 19.4 ± 4.1; 7, 23.1 ± 3.7; and 8,23.9 ± 3.6. For controls, the values were as follows: USDA125-Sp-G, 17.1 ± 1.8; USDA 125-Sp-L, 19.6 + 4.4; anduninoculated, 2.3 ± 2.3.

DISCUSSION

The long-term phenotype and genetic stability of bacteriaused as a soil inoculant is an important parameter in micro-bial ecology as well as for agricultural purposes. It is

VOL. 54, 1988

on June 5, 2020 by guesthttp://aem

.asm.org/

Dow

nloaded from

APPL. ENVIRON. MICROBIOL.

1 2 3 4 5 6 7 8 9 10 11 12

1 2 3 4 5 6 7 8 9 10 11

23

9.5

6.6

4.4

2.3

2.0

i .

FIG. 1. Restriction enzyme and Southern hybridization analysisof 12 serogroup 125 isolates. Genomic DNAs of five mucous (lanes1 to 5) and five nonmucous (lanes 6 to 10) serogroup 125 field isolatesand of USDA 125-Sp-G (lane 11) and USDA 125-Sp-L (lane 12)parental strains were digested with restriction enzyme SstI andseparated by electrophoresis (top). The DNAs were transferred ontonitrocellulose and hybridized to the B. japonicinm USDA 125-Sp(lyophilized parental culture) symbiotic region (bottom). Molecularsize markers (in kilobase pairs) are on the left.

essential to know whether these bacterial populations are

able to persist and whether these strains conserve theirproperties. Few long-term experiments testing stability havebeen reported, and those that have been mainly focused on

nitrogen fixation (5, 34). In this study we checked thestability of three B. japonicium populations released byreisolating them and testing 38 phenotypic characters(morphology, serology, sugar utilization, antibiotic resis-tance, and enzymatic activity). No consistent differencecould be observed among isolates belonging to the same

population. A more detailed investigation of symbiotic andgenetic stability, restricted to mucous and nonmucous sero-

group 125, did not permit the detection of any furtherdifference.The major variation detected was in colony morphology

on YEM. Of the 45 serogroup 125 isolates from nodules ofsite A, 39% had large, slimy colonies and 61% had small,nonslimy colonies. This seemed to be related to mannitolutilization because the nonslimy colonies grew poorly onminimal medium with mannitol. The differences in mannitol-utilizing ability have been shown to be related to D-mannitoldehydrogenase activity, slimy colonies having higher levelsof constitutive D-mannitol dehydrogenase (11, 12, 17). How-ever, the observed variation is not a consequence of theintroduction of B. japonicurn in the soil because a similarvariation was also detected among subcolonies of parentalstrain USDA 125-Sp-G. These morphological characteristicswere stable for at least 2 years in the laboratory.The mucous and nonmucous derivatives (mannitol utiliz-

ing and nonutilizing, respectively) had no significant differ-ence in the ability to fix nitrogen on soybean plants. In otherstudies a correlation between mannitol-utilizing ability andpoor N2 fixation was found in B. japonicium USDA 110 (11)and B. japonicum 61A76, USDA 76, and USDA 140 (32).However, Mathis et al. (18) isolated a mannitol-utilizingsubclone of B. japonicum USDA 110 which fixed N2 moreefficiently than the parental strain did. Furthermore, Her-ridge and Roughley (8) had previously reported a correlationbetween the nonmucous character and N2-fixing activity forother species of rhizobia. Thus, mannitol utilization in B.japonicum may not necessarily be correlated with symbioticeffectiveness.

Reports on various successful methods for discriminatingamong Rhizobilum species and strains have appeared in theliterature (10, 16, 22-24, 26, 32). We used different methodsto differentiate B. japonicirm strains isolated from the field.The specific recognition of USDA 125 and USDA 138bradyrhizobia by their respective antisera was conserved inthese field isolates; similar results have been shown withLotononis rhizobia (5). Intrinsic levels of antibiotic resis-tance of these isolates remained stable for the nine testedantibiotics, and those for tetracycline, rifampin, and ampi-cillin could be used to discriminate between the two B.japonicum strains studied. The antibiotic marker in themutant strains also remained stable, spectinomycin forUSDA 125-Sp and streptomycin for USDA 138-Sm. Further-more, the enzymatic activity profiles corroborated strainstability among 122 B. japonicum isolates when they werecompared with their respective parental strains. Two iso-lates, however, which were able to grow on YEM containing10 Vig of kanamycin ml-1 were found. Diatloff (5), in hisstudy of Lotononis rhizobia, mentions that antibiotic resis-tance was one of the most unstable characters that he tested.A homologous nifD gene probe was used to screen several

serogroup 125 isolates. Hybridization patterns of fragmentsof their genomic DNA did not indicate any genetic variationamong these isolates. However, it is known that the struc-tural nif gene region is conserved; therefore, it might beinteresting to use more variable regions to detect mutations.

It has been shown previously that after 8 to 13 years,introduced B. japonicum populations persist in French soilseven without soybean crops at 103 to 104 rhizobia g of driedsoil-' from site C (3). From site A, 3.7 x 103 strain USDA125 rhizobia (unpublished results) were counted by a fluo-rescent-antibody technique on a membrane filter (2). Theysurvive at a level (<104 CFU g of dried soil-') which is toolow for reisolation directly from soil, even with the use ofantibiotic-resistant markers. This failure may be due to high

2640 BRUNEL ET AL.

on June 5, 2020 by guesthttp://aem

.asm.org/

Dow

nloaded from

STABILITY OF B. JAPONICUM INOCULANTS 2641

levels of contaminants or the development of quiescentforms (7) of B. japoni(cum, whose physiological state mayhave hindered colony formation under the conditions used.However, we have succeeded in obtaining seven isolatesdirectly from soil by using the specific serological propertiesof strain USDA 125, and these isolates nodulated soybean.We therefore suggest that high levels of other microorgan-isms prevented the direct isolation of B. japoniuitin bacteriafrom soil on YEM agar, even in the presence of spectinomy-cin, cycloheximide, and benomyl.Our data indicate that inoculant strains can persist for

more than 10 years in French soils. This persistence was theresult of the saprophytic competence in B. japonhium be-cause soybean had not been grown on the sites after the yearof release, cereal crops having been cultivated in the inter-vening period. We did not detect a genetical instability ofrhizobia in soil, as has been noted for R. legiuiminosaruitin bv.tiifolii by Schofield et al. (26). However, a greater instabilityof symbiotic genes in fast-growing rhizobia may be explainedby their occurrence on large plasmids. Therefore, it appearsthat once introduced into a suitable soil without indigenousB. jtiponhiklin populations, B. japoniCuin inoculants willintegrate into the indigenous soil communities without sig-nificant modifications.

ACKNOWLEDGMENTS

We thank N. Deperrier and D. Bernillon for technical assistance,M. Obaton (INRA-Montpellier) for providing bacterial strains andfor his comments on the manuscript, T. Huguet (INRA-Toulouse)for the gift of plasmids, and K. Wong (CRBF, Canada) for readingthe manuscript and correcting the English. We are also grateful tothe society LIPHA (Lyon) for the production of B. japonicluminoculants.

This study was financed by the Etablissement Public de la RegionRh6ne-Alpes. B.B. received a fellowship from the Fondation Scien-tifique de Lyon et du Sud Est, and P.N. received a Natural Sciencesand Engineering Research Council of Canada grant (no. a-3501).

LITERATURE CITED1. Cannon, F. C., G. E. Riedel, and F. M. Ausubel. 1979. Overlap-

ping sequences of Kiebsiella puel1nonicie nif DNA cloned andcharacterised. Mol. Gen. Genet. 174:59-66.

2. Crozat, Y., J. C. Cleyet-Marel, and A. Corman. 1987. Use of thefluorescent antibody technique to characterize equilibrium sur-vival concentrations of Bracdvrhizohiuln japonicuin strains insoil. Biol. Fertil. Soils 4:85-90.

3. Crozat, Y., J. C. Cleyet-Marel, J. J. Giraud, and M. Obaton.1982. Survival rates of Rhizobiu,n j(aponlicumll populations intro-duced into different soils. Soil Biol. Biochem. 14:401-405.

4. Dagnelie, P. 1970. L'analyse de la variance a deux criteres, p.153-209. In Theorie et m6thodes statistiques. Applicationsagronomiques, vol. 2. Les Presses Agronomiques de Gembloux,Gembloux, Belgium.

5. Diatloff, A. 1977. Ecological studies of root-nodule bacteriaintroduced into field environments. 6. Antigenic and symbioticstability in Lotonon1is rhizobia over a 12-year period. Soil Biol.Biochem. 9:85-88.

6. Gibson, A. H. 1963. Physical environment and symbiotic nitro-gen fixation. I. The effect of root temperature on recentlynodulated Trifoliutn subterranucin. L. plants. Aust. J. Biol. Sci.16:28-42.

7. Hattori, T. 1982. Analysis of plate count data of bacteria innatural environments. J. Gen. Appl. Microbiol. 28:13-22.

8. Herridge, D. F., and R. S. Roughley. 1975. Variation in colonycharacteristics and symbiotic effectiveness of Rlii.obiin. J.Appl. Bacteriol. 38:19-27.

9. Hranitzky, K. W., A. D. Larson, D. W. Ragsdale, and R. J.Siebeling. 1980. Isolation of 01 serovars of Vibrii (cholerae fromwater by serologically specific method. Science 210:1025-1026.

10. Josey, D. P., J. L. Beynon, A. W. B. Johnston, and J. E.Beringer. 1979. Strain identification in Rhizobiumr using intrinsicantibiotic resistance. J. Appl. Bacteriol. 46:343-350.

11. Kuykendall, L. D., and G. H. Elkan. 1976. Rlhizobiuln japoni-(cur derivatives differing in nitrogen-fixing efficiency and carbo-hydrate utilization. AppI. Environ. Microbiol. 32:511-519.

12. Kuykendall, L. D., and G. H. Elkan. 1977. Some features ofmannitol metabolism in Rlii.ohiiurn japonuicu. J. Gen. Micro-biol. 98:291-295.

13. Lowendorf, H. S. 1980. Factors affecting survival of Rhliiobiirin soil. Adv. Microb. Ecol. 4:87-124.

14. Maniatis, T. E., E. F. Fritsch, and J. Sambrook. 1982. Molecularcloning. laboratory manual. Cold Spring Harbor Laboratory.Cold Spring Harbor, N.Y.

15. Marmur, J. 1961. A procedure for the isolation of DNA frommicroorganisms. J. Mol. Biol. 3:208-218.

16. Masterson, R. V., R. K. Prakash, and A. G. Atherly. 1985.Conservation of symbiotic nitrogen fixation gene sequences inRhlii-obiuirn japonicumn and Bradvr-hli obiumn japonicuml. J. Bac-teriol. 163:21-26.

17. Mathis, J. N., W. M. Barbour, T. B. Miller, D. W. Israel, andG. H. Elkan. 1986. Characterization of a mannitol-utilizing.nitrogen-fixing Bradvrlhizobiumn japonwiciin USDA 110 deriva-tive. Appl. Environ. Microbiol. 52:81-85.

18. Mathis, J. N., L. D. Kuykendall, and G. H. Elkan. 1986.Restriction endonuclease and niif homology patterns of Brady-rhizobiuni japoniculn USDA 110 derivatives with and withoutnitrogen fixation competence. Appl. Environ. Microbiol. 51:477-480.

19. Obaton, M. 1971. Utilisaition de mutants spontanes resistantsaux antibiotiques pour l'tude ecologique des Rllizobiuilm. C.R.Acad. Sci. Ser. D 272:2630-2633.

20. Ozawa, T., A. Takemi, and M. Yamaguchi. 1984. Decrease inthe proportion of cells capable of inducing nodules to thepopulation of Rlhizobiuiiii japoniculn introduced into soil. SoilSci. Plant Nutr. 30:479-484.

21. Rigby, P. W. J., M. Dieckmann, D. Rhodes, and P. Berg. 1977.Labelling deoxyribonucleic acid to high specific activity in vitroby nick translation with DNA polymerase 1. J. Mol. Biol. 113:237-251.

22. Sadowsky, M. J., B. B. Bohlool, and H. H. Keyser. 1987.Serological relatedness of Rhizobiuni ftedii to other rhizobia andto the bradyrhizobia. Appl. Environ. Microbiol. 53:1785-1789.

23. Sadowsky, M. J., H. H. Keyser, and B. B. Bohlool. 1983.Biochemical characterization of fast- and slow-growing rhizobiathat nodulate soybeans. Int. J. Syst. Bacteriol. 33:716-722.

24. Sadowsky, M. J., R. E. Tully, P. B. Cregan, and H. H. Keyser.1987. Genetic diversity in Bra-tdvrhi-izobiiumtnjaponicilurin serogroup123 and its relation to genotype-specific nodulation of soybean.Appl. Environ. Microbiol. 53:2624-2630.

25. Schmidt, L. E. 1974. Quantitative autoecological study of mi-croorganisms in soil by immunofluorescence. Soil Sci. 118:141-149.

26. Schofield, P. R., A. H. Gibson, W. F. Dudman, and J. M.Watson. 1987. Evidence for genetic exchange and recombina-tion of Rhliizobjiiin symbiotic plasmids in a soil population. Appl.Environ. Microbiol. 53:2942-2947.

27. Simonet, P., J. Haurat, P. Normand, R. Bardin, and A. Moiroud.1986. Localization of nifgenes on a large plasmid in Franktia sp.strain ULQ132105009. Mol. Gen. Genet. 204:492-495.

28. Slater, J. H. 1984. Genetic interactions in microbial communi-ties, p. 87-93. In M. J. Klug and C. A. Reddy (ed.). Currentperceptives in microbial ecology. American Society for Micro-biology, Washington, D.C.

29. Southern, E. M. 1975. Detection of specific sequences amongDNA fragments separated by gel electrophoresis. J. Mol. Biol.98:503-517.

30. Tautz, D., and M. Renz. 1983. An optimized freeze-squeezemethod for the recovery of DNA fragments from agarose gels.Anal. Biochem. 132:14-19.

31. Trevors, J. T., T. Barkay, and A. W. Bourquin. 1987. Genetransfer among bacteria in soil and aquatic environments: a

review. Can. J. Microbiol. 33:191-198.

VOL. 54, 1988

on June 5, 2020 by guesthttp://aem

.asm.org/

Dow

nloaded from

2642 BRUNEL ET AL. APPI. ENVIRON. MICROBIOL.

32. Upchurch, R. G., and G. H. Elkan. 1977. Comparison of colonymorphology, salt tolerance, and effectiveness in Rl2izohi7u,japonicum. Can. J. Microbiol. 23:1118-1122.

33. Van Rensburg, H. J., and B. W. Strijdom. 1971. Stability ofinfectivity in strains of Rliizobiiun japonicum. Phytophylactica3: 125-130.

34. Van Rensburg, H. J., and B. W. Strijdom. 1985. Effectiveness ofRlizobiin71 strains used in inoculants after their introductioninto soil. Appl. Environ. Microbiol. 49:127-131.

35. Vincent, J. M. 1970. A manual for the practical study ofroot-nodule bacteria. IBP handbook 15. Blackwell ScientificPublications, Ltd., Oxford.

on June 5, 2020 by guesthttp://aem

.asm.org/

Dow

nloaded from