genetic and biochemical diversity among isolates ... · the isolates produced variable results for...
TRANSCRIPT
APPLIED AND ENVIRONMENTAL MICROBIOLOGY,0099-2240/00/$04.0010
Mar. 2000, p. 1098–1106 Vol. 66, No. 3
Copyright © 2000, American Society for Microbiology. All Rights Reserved.
Genetic and Biochemical Diversity among Isolates ofPaenibacillus alvei Cultured from Australian Honeybee
(Apis mellifera) ColoniesSTEVEN P. DJORDJEVIC,* WENDY A. FORBES, LISA A. SMITH, AND MICHAEL A. HORNITZKY
New South Wales Agriculture, Elizabeth Macarthur Agricultural Institute, Camden, New South Wales,Australia 2570
Received 23 June 1999/Accepted 6 December 1999
Twenty-five unique CfoI-generated whole-cell DNA profiles were identified in a study of 30 Paenibacillus alveiisolates cultured from honey and diseased larvae collected from honeybee (Apis mellifera) colonies in geo-graphically diverse areas in Australia. The fingerprint patterns were highly variable and readily discerniblefrom one another, which highlighted the potential of this method for tracing the movement of isolates inepidemiological studies. 16S rRNA gene fragments (length, 1,416 bp) for all 30 isolates were enzymaticallyamplified by PCR and subjected to restriction analysis with DraI, HinfI, CfoI, AluI, FokI, and RsaI. With eachenzyme the restriction profiles of the 16S rRNA genes from all 30 isolates were identical (one restrictionfragment length polymorphism [RFLP] was observed in the HinfI profile of the 16S rRNA gene from isolate17), which confirmed that the isolates belonged to the same species. The restriction profiles generated by usingDraI, FokI, and HinfI differentiated P. alvei from the phylogenetically closely related species Paenibacillusmacerans and Paenibacillus macquariensis. Alveolysin gene fragments (length, 1,555 bp) were enzymaticallyamplified from some of the P. alvei isolates (19 of 30 isolates), and RFLP were detected by using the enzymesCfoI, Sau3AI, and RsaI. Extrachromosomal DNA ranging in size from 1 to 10 kb was detected in 17 of 30 (57%)P. alvei whole-cell DNA profiles. Extensive biochemical heterogeneity was observed among the 28 P. alveiisolates examined with the API 50CHB system. All of these isolates were catalase, oxidase, and Voges-Proskauer positive and nitrate negative, and all produced acid when glycerol, esculin, and maltose were added.The isolates produced variable results for 16 of the 49 biochemical tests; negative reactions were recorded inthe remaining 30 assays. The genetic and biochemical heterogeneity in P. alvei isolates may be a reflection ofadaptation to the special habitats in which they originated.
Paenibacillus alvei is a spore-forming bacterium that swarmsvigorously on routine culture media. P. alvei, Enterococcusfaecalis, Enterococcus faecium, and Achromobacter eurydice areoften recovered from diseased larvae obtained from honeybee(Apis mellifera) colonies infected with Melissococcus pluton, thecausative agent of European foulbrood (EFB) (3). In Austra-lia, P. alvei is the third-most common bacterium detected inhoneybee colonies, and E. faecalis, E. faecium, and A. eurydiceare rarely recovered from EFB-affected colonies (17, 18). P.alvei can produce signs in larvae that are similar to the signsproduced by Paenibacillus larvae subsp. larvae, which causesAmerican foulbrood, the other major bacterial disease of hon-eybees.
Several studies based on comparative analyses of 16S rRNAgene sequences of different Bacillus species revealed five phy-logenetically distinct clusters (groups 1 through 5), which con-firmed that this genus is genetically heterogeneous and in needof extensive taxonomic revision (1, 27). Furthermore, the au-thors suggested that P. alvei (formerly Bacillus alvei) belongs ingroup 3, which comprises 10 species. Members of group 3exhibited comparatively low levels of sequence homology withthe other Bacillus groups and constituted a distinct lineage,and these organisms were transferred to a new genus, thegenus Paenibacillus (2).
Unlike the pathogens M. pluton and P. larvae subsp. larvae,
which are found only in association with honeybees, P. alveioccupies many environmental niches, including the soil (30),milk (24), mosquito larvae (4), the wax moth (11), and humans(26). P. alvei produces alveolysin, a thiol-activated toxin that ishighly homologous to listeriolysin O, perfringolysin O, pneu-molysin, and streptolysin O (10). The role of P. alvei in themicrobiology and ecology of honeybees has received compar-atively little attention. When restriction endonuclease analysis(REA) and immunoblot analysis were used, high levels ofgenetic and antigenic homogeneity were observed among geo-graphically diverse Australian isolates of the primary honeybeepathogens P. larvae subsp. larvae (7, 16) and M. pluton (9).There have been no reports of the levels of genetic heteroge-neity among geographically diverse isolates of P. alvei. In thisstudy, amplified ribosomal DNA restriction analysis(ARDRA) of the 16S rRNA gene was used to investigate thespecies identities of a collection of isolates of P. alvei recoveredfrom foulbrood-affected honeybee colonies in the easternstates of Australia. Whole-cell DNA fingerprint profiles ob-tained by using restriction enzyme CfoI provided a way toinvestigate the amount of genetic heterogeneity in P. alvei andto study the utility of this method for tracing the movement ofP. alvei isolates in epidemiological studies. Moreover, the API50CHB system was used to rapidly confirm the identities of P.alvei isolates and to biochemically characterize the isolates.
MATERIALS AND METHODS
P. alvei strains from larvae. Thirty P. alvei isolates from different geographicregions in the eastern half of Australia were cultured from honey and honeybeelarval samples suspected of having EFB (Table 1). Smears of diseased larvaewere submitted by apiarists or state apiary officers. The methods used to culture
* Corresponding author. Mailing address: NSW Agriculture, Eliza-beth Macarthur Agricultural Institute, Private Mail Bag 8, Camden,NSW 2570, Australia. Phone: 61-246-406426. Fax: 61-246-406384. E-mail: [email protected].
1098
on August 19, 2020 by guest
http://aem.asm
.org/D
ownloaded from
TA
BL
E1.
Biochem
icaldataobtained
byusing
theA
PI50C
HB
systemfor
28geographically
diverseA
ustralianisolates
ofP.alvei
IsolateT
estresults
No.
Designation
SalicinR
iboseA
donitolG
alactoseG
lucoseM
annoseN
-A
cetylglucosamine
Am
ygdalinA
rbutinC
ellobioseM
elibioseSucrose
Trehalose
Raffinose
StarchG
entiobiose
New
SouthW
alesisolates1
97/1133/J12
12
21
21
22
12
21
21
22
97/895/41
11
21
21
21
12
21
21
23
97/1004/6N
11
12
12
11
21
22
11
11
497/1004/IN
11
12
12
11
11
22
12
12
597/1133/E
31
11
11
21
11
11
12
11
2
797/1004/3N
11
12
12
12
21
22
12
12
897/1004/5N
11
12
12
12
21
22
12
12
1497/895/3
11
12
12
12
11
22
12
22
697/895/2
11
12
12
12
21
22
12
12
Tasm
aniaisolates
997/331/D
21
12
11
21
21
12
22
21
111
97/331/C2
11
11
11
11
11
22
12
11
1297/331/C
11
12
11
12
21
22
22
21
113
97/331/B1
11
22
12
12
22
22
12
11
1597/331/1A
11
11
12
11
11
22
12
21
Victoria
isolates16
96/C245/3
11
22
12
12
11
22
22
11
1796/B
852/21
11
21
21
21
12
22
21
118
96/B852/4
11
21
12
12
11
22
22
11
1996/C
245/24A
21
22
12
12
11
22
22
11
2096/C
245/12B
21
12
12
12
11
22
22
11
Queensland
isolates21
96/57421
11
21
21
21
22
21
22
222
96/A345
11
11
12
11
11
21
12
22
SouthA
ustraliaisolates
23K
22171
22
22
11
12
12
21
22
224
K7245/1
11
21
11
11
21
22
22
22
25K
71561
12
11
21
11
12
22
22
226
K6227
11
21
12
11
21
11
21
12
277666/52R
11
21
12
11
11
11
22
22
28K
7463/L8
11
21
12
11
12
22
22
22
29K
81031
22
21
21
11
12
21
22
2Isolates
ofunknow
norigin30
A673/2
31A
294
VOL. 66, 2000 GENETIC AND BIOCHEMICAL HETEROGENEITY OF P. ALVEI 1099
on August 19, 2020 by guest
http://aem.asm
.org/D
ownloaded from
P. alvei have been described previously (19). A mixture of brood material andsaline was streaked onto sheep blood agar, which contained blood agar base no.2 (Oxoid, Basingstoke, United Kingdom) supplemented with 7% citrated ovineblood. Plates were incubated at 37°C for 2 days in air containing 10% CO2.
P. alvei strains from honey. An aliquot (75 ml) of honey was mixed with 75 mlof phosphate-buffered saline (pH 7.2) and centrifuged for 45 min at 3,000 3 g,and most of the supernatant was removed; about 1.5 ml of fluid was left, whichwas mixed with the sediment in the bottle. Each sample was streaked onto sheepblood agar and incubated as described above.
Identification of P. alvei. Swarming colonies were considered to be P. alveicolonies if (i) smears prepared from the colonies and stained by the Grammethod consisted of gram-positive rods that were 2 to 5 mm long and 0.5 to 0.8mm wide, (ii) the organisms produced oval spores, and (iii) the organisms wereVoges-Proskaure and oxidase positive. Other non-carbohydrate-based tests(catalase, nitrate, urease, and indole tests), carried out by using the proceduresdescribed by Cowan and Steel (5), were also used to further characterize theisolates. ARDRA was also used to confirm the identities of the P. alvei isolates.
Carbohydrate acidification. The biochemical characteristics of P. alvei weredetermined with the API 50CHB system, which was used as recommended by themanufacturer. A dendrogram based on the API 50CHB biochemical reactions ofP. alvei isolates was produced with the computer package GENSTAT by usingthe average linkage of cluster analysis applied to similarities based on the match-ing coefficient.
DNA isolation. Total cellular DNA of 30 P. alvei isolates was extracted andpurified by using the procedure developed for isolation of DNA from M. plutonand other gram-positive bacteria (7, 8).
REA of P. alvei DNA. Preliminary experiments revealed that P. alvei DNAdigested with CfoI produced clearly resolved DNA profiles when the DNA wasseparated with 3.5% polyacrylamide gels stained with silver. The methods usedfor electrophoresis and silver staining of DNA fragments have been describedpreviously (6). Gels were loaded with 1 to 2 mg of restricted DNA, and fragmentswere separated by electrophoresis (35 V per gel, 17 to 21 h). Isolates wereconsidered clonal if their restriction profiles exhibited a degree of similarity(DOS) greater than 92% when the following formula was used: DOS 5 100% 2(Nd 3 100)/Ns, where Ns is the total number of bands of two fingerprints beingcompared and Nd is an estimate of the number of bands not shared by the twofingerprints (21, 25).
Extrachromosomal DNA analysis. Total-cell DNA (3 to 5 mg) prepared asdescribed above was loaded onto 1.0% agarose gels and separated for 5 h (50 V)in 0.53 TBE buffer (45 mM Tris-HCl, 45 mM borate, 0.01 mM EDTA; pH 8.0)containing ethidium bromide (0.5 mg ml21).
PCR amplification and restriction fragment length polymorphism (RFLP)analysis of the alveolysin gene. Alveolysin-specific PCR primers ALV100 (59TAAAAAGGGGATGACTGTAT 39; positions 1 to 20) and ALV101 (59 AATGAGGAGATGTTCATACA 39; positions 1555 to 1536) were designed by align-ing the DNA sequences of the thiol-activated toxins alveolysin (EMBL accessionno. M62709), listeriolysin O (X15127), perfringolysin O (M81080), pneumolysin(X52474), and streptolysin O (M18638). The sequences were retrieved fromEMBL and were aligned by using the Pileup program (Genetics ComputerGroup [GCG], University of Wisconsin, Madison, Wis.) accessed via the Aus-tralian National Genome Information Service (University of Sydney, Sydney,Australia). The primers aligned with alveolysin-specific sequences and facilitatedamplification of a 1,555-bp product by PCR. Reactions were performed in 0.2-mltubes, and each 50-ml reaction mixture consisted of 20 pmol of each primer, 200mM dTTP, 200 mM dCTP, 200 mM dATP, 200 mM dGTP, 13 PCR buffer (10mM Tris-HCl, 50 mM KCl; pH 8.3), 0.2 U of Taq polymerase, 1.5 mM magne-sium chloride, and 50 to 100 ng of P. alvei total-cell DNA. Amplifications wereperformed by using Corbett Research thermocyclers (models FTS-960 and PC-960). After an initial denaturation cycle consisting of 93°C for 3 min, the follow-ing conditions were used. A denaturation temperature of 93°C for 30 s and anextension temperature of 72°C for 2.5 min were used throughout. A touchdownprotocol consisting of 10 cycles at 51°C for 1 min followed by 5 cycles at 50°C for1 min, 5 cycles at 49°C for 1 min, 5 cycles at 48°C for 1 min, and then 10 cyclesat 47°C for 1 min was used. A final cycle consisting of 72°C for 3 min completedthe PCR.
Amplification products were separated by electrophoresis through 1.0% aga-rose. The identify of the 1,555-bp amplicon was confirmed by performing aheminested PCR with forward primer ALV102 (59 CTTGAAAGGAAGGAAAGTAC 39; positions 39 to 58) and reverse primer ALV101. The conditions usedfor the PCR were exactly the same as the conditions used for the reactiondescribed above. The heminested reaction amplified a 1,517-bp fragment, whichwas separately digested overnight with CfoI, RsaI, and Sau3A1 in a preliminarystudy to investigate the intragenic genetic heterogeneity in the alveolysin gene.
ARDRA. 16S rRNA genes from representative strains belonging to the Paeni-bacillus cluster (2, 14), including strains of Paenibacillus polymyxa (accession no.X60636 and X57308), Paenibacillus macerans (X60624 and X57306), Paenibacil-lus macquariensis (X60625 and X57305), P. larvae (X60619), Paenibacillus pul-vifaciens (X60636), Paenibacillus amylolyticus (X60606), Paenibacillus pabuli(X60630), Paenibacillus gordonae (X60617), Paenibacillus peoriae (D78476),Paenibacillus lautus (D78473), and P. alvei (X60604, X57304, and M62709), aswell as strains of several Bacillus species, including Bacillus laterosporus (X57307)and Bacillus stearothermophilus (X57309), were aligned by using the Pileup pro-
gram (GCG, University of Wisconsin) accessed through the Australian NationalGenome Information Service. Based on predicted cleavage sites derived by usingthe mapplot program (GCG, University of Wisconsin), a panel of six enzymes(FokI, RsaI, CfoI, DraI, HinfI, and AluI) was chosen and used to differentiateamong these phylogenetically related species. Primers ALV16SF (59 CCTGGCTCAGGACGAACGCT 39; positions 18 to 37) and ALV16SR (59 TTGTAAACTCTCGTGGTGTGACGG 39; positions 1437 to 1414), which exhibited 100%homology to the 16S rRNA gene of P. alvei (accession no. X57304), weredesigned to facilitate amplification of a 1,416-bp fragment by PCR. Each 50-mlreaction mixture consisted of 20 pmol of each primer, 200 mM dTTP, 200 mMdCTP, 200 mM dATP, 200 mM dGTP, 13 PCR buffer, 1 U of Taq polymerase,P. alvei DNA (0.1 to 0.5 mg), and 1.5 mM magnesium chloride. Sterile pure water(Milli Q) was used to bring the final volume to 50 ml. After an initial denaturationcycle (94°C, 3 min), 35 cycles consisting of 94°C for 30 s, 56°C for 45 s, and 72°Cfor 1.5 min followed by a final extension cycle consisting of 72°C for 2 mincompleted the PCR. Amplification products were separately digested with eachof the six restriction enzymes, and the products of digestion were separated byagarose gel electrophoresis.
RESULTS
Identification of P. alvei. All isolates were identified as P.alvei isolates based on their Gram staining characteristics andbiochemical reactions. The ARDRA results for all of the iso-lates were consistent with a P. alvei 16S rRNA gene profile (seebelow).
To eliminate the possibility that isolates comprised mixedcultures of P. alvei, CfoI-generated restriction endonucleasefragment profiles (REFPs) of whole-cell DNA from severaldifferent sites of the advancing front of a swarming colony(produced separately by a subset of five isolates) were ana-lyzed. These profiles were indistinguishable from one anotherand from the profile generated when DNA from the originalDNA extraction was used (data not shown), which confirmedthe homogeneous population structure of the original isolates.
API 50CHB tests. Twenty-eight of the 30 isolates were testedwith the API 50CHB system; (isolates 30 and 31 were used onlyfor the genetic studies). All of the isolates were positive forglycerol, esculin, and maltose reactions and negative for eryth-ritol, D-arabinose, L-arabinose, D-xylose, L-xylose, a-methylxy-loside, D-fructose, L-sorbose, rhamnose, dulcitol, inositol, man-nitol, sorbitol, a-methyl-D-mannoside, a-methyl-D-glucoside,lactose, inulin, melezitose, glycogen, xylitol, D-turanose, D-lyx-ose, D-tagatose, D-fucose, L-fucose, D-arabitol, L-arabitol, glu-conate, 2-ketogluconate, and 5-ketogluconate reactions (Table1). There were 16 biochemical tests in the API 50CHB systemfor which P. alvei strains gave variable results. These were thesalicin, ribose, adonitol, galactose, glucose, mannose, N-acetyl-glucosamine, amygdalin, arbutin, cellobiose, melibiose, su-crose, D-trehalose, raffinose, starch, and gentiobiose tests. Theprofiles for each isolate are shown in Table 1. A dendrogramsummarizing the biochemical relationships among the 28 iso-lates is shown in Fig. 1.
Using API tests, we identified 22 isolates as P. alvei at aconfidence level greater than 99% and two isolates (97/1133/J1and 96/A964/K7156) as P. alvei at a confidence level greaterthan 89%. Four isolates (97/331/C1, 96/B852/2, K2217, andK8103) produced carbohydrate profiles that were more like thecarbohydrate profiles of Bacillus species. However, these iden-tities that were determined tentatively could be ruled out basedon a combination of the ability of the organisms to swarm onlaboratory medium, spore morphology, and supplementarybiochemical test results (see below). Isolates 6, 7, and 8 (Table1) were identical based on the API 50CHB biochemical reac-tions, as were isolates 9 and 18. The final amalgamation in thedendrogram (Fig. 1) occurred at a similarity level of 54%.
Supplementary biochemical assays. All P. alvei isolates werecatalase positive and nitrate negative. All of the isolates exceptthree Victoria isolates (96/B852/4, 96/C245/24A, and 96/C245/
1100 DJORDJEVIC ET AL. APPL. ENVIRON. MICROBIOL.
on August 19, 2020 by guest
http://aem.asm
.org/D
ownloaded from
12B) and one New South Wales isolate (97/1133/J1) were ure-ase positive. Most of the isolates were indole positive; the onlyexceptions were four Tasmania isolates (97/331/B2, 97/331/C1,97/331/C2, and 97/331/B1). Most isolates were o-nitrophenyl-b-D-galactopyranoside positive; the only exceptions were threeNew South Wales isolates (97/C245/3, 97/1133/J1, and 97/1133/E3), one Tasmania isolate (97/331/C1), and one Victoria iso-late (96/B852/2).
ARDRA. Previous studies have shown that ARDRA is asimple and reliable tool for identifying bacterial species (13–15). DNA of all 30 P. alvei isolates was used as templates foramplification of a 1,416-bp portion of the 16S rRNA gene inwhich primers ALV16SF and ALV16SR were used. The1,416-bp amplicon from each of the 30 isolates was separatelydigested with RsaI, CfoI, FokI, HinfI, AluI, and DraI, and therestriction profiles which are representative of each enzyme
are shown in Fig. 2. The 16S rRNA gene restriction profilesobtained with each enzyme were identical, which confirmedthat all of the isolates belonged to the same species. The 16SrRNA gene fragment from isolate 17 produced a slightly dif-ferent HinfI profile (Fig. 2, lane 6) compared to the HinfIprofile of the remaining 29 isolates (Fig. 2, lane 5). Restrictionendonucleases DraI and HinfI produced RFLPs that wereidentical to the RFLPs predicted when the mapplot program(GCG, University of Wisconsin) was used, and the recognitionsites of these two enzymes spanned nucleotide sequenceswhich varied in virtually all of the species in the group 3 cluster,including species that were less closely related phylogenetically(B. stearothermophilus and B. laterosporus). Only Paenibacillusazotofixans could not be reliably differentiated from P. alveiwith the panel of six enzymes used in this study. However, P.azotofixans could easily be differentiated from P. alvei based onits inability to swarm on laboratory agar and its fermentationpattern when API 50CHB tests were performed (Table 2) (28).Collectively, our data confirmed that all 30 isolates were P.alvei isolates. Several restriction endonucleases (RsaI, CfoI,and AluI) generated RFLP profiles that were not similar toprofiles predicted with the mapplot program. The presence ofundesignated nucleotides in the 16S rRNA gene sequencesdeposited in GenBank (accession no. X57304, X60604, andD78317) and minor sequence variations are the most likelyexplanations for these profile variations.
PCR and RFLP analyses of the alveolysin gene. When prim-ers ALV100 and ALV101 were used, a 1,555-bp fragmentrepresenting virtually all of the alveolysin gene was amplifiedfrom 19 of the 30 P. alvei isolates. The identity of the ampli-fication product was confirmed by performing a heminestedPCR with a second internal primer (ALV102), which amplifieda 1,517-bp portion of the alveolysin gene. Primers ALV100 andALV101 or ALV102 did not amplify an alveolysin gene frag-ment when DNA from P. alvei isolates 5, 9, 11 through 13, 16,22, 24, and 26 through 28 were used. Digestion of the 1,517-bpportion of the alveolysin gene with CfoI resulted in identifica-tion of three fragments whose approximate molecular sizeswere 730, 310, and 270 (doublet) bp for all P. alvei isolates (Fig.2, lanes 9 and 11). A fourth fragment at 195 bp (Fig. 2, lane 10)was observed in profiles representing isolates 7, 8, 1, and 14.Five fragments (400, 315, 300, 265, and 240 bp) were predictedby digestion of the alveolysin gene (accession no. M62709) withCfoI.
Digestion of the 1,517-bp alveolysin gene fragment with RsaIproduced four fragments whose approximate sizes were 650,370 (doublet), and 125 bp (Fig. 2, lane 12) for all but threeisolates. RsaI-generated RPLPs were observed in alveolysingene fragments from isolate 17 (630, 420, and 300 bp) (Fig. 2,lane 14) and from isolates 29 and 23 (650, 370, 300, and 125 bp)(Fig. 2, lane 13). Digestion of the alveolysin gene fragmentwith Sau3AI generated four fragments whose approximatesizes were 880, 250 (doublet), and 120 bp (data not shown).These data suggest that there is sequence variability amongalveolysin genes in different isolates of P. alvei.
DNA restriction endonuclease profiles. Digestion of P. alveiDNA with CfoI generated fragments whose molecular sizesranged from approximately 8.5 to 0.3 kb, and these fragmentswere clearly resolved by using 3.5% polyacrylamide gelsstained with silver (Fig. 3). The REFPs of 25 of the 30 P. alveiisolates obtained from five eastern states of Australia could bereadily distinguished from one another. Profile variations wereoften evident throughout the molecular size range (8.5 to 0.3kb). DNA from isolates 27 (South Australia) and 16 (Victoria)produced indistinguishable CfoI-generated REFPs (Fig. 3,lanes 6 and 7) that were clearly different from the CfoI-gener-
FIG. 1. Dendrogram showing the levels of similarity between P. alvei isolatesbased on their API 50CHB biochemical reactions.
FIG. 2. Restriction endonuclease profiles of amplified regions of the 16SrRNA and alveolysin genes of P. alvei. 16S rRNA gene fragments (length, 1,416bp) were separately amplified by PCR and digested with CfoI, RsaI, FokI, HinfI,DraI, and AluI. Representative profiles generated by using CfoI (lane 2), RsaI(lane 3), FokI (lane 4), HinfI (lanes 5 and 6), DraI (lane 7), and AluI (lane 8) areshown. The profiles were generated after digestion of the 16S rRNA genefragment derived from isolate 17 (except in lane 6, which shows the HinfI profileof the 16S rRNA gene fragment derived from isolate 18). CfoI profiles of thealveolysin gene fragment derived from P. alvei isolates 23, 14, and 19 are shownin lanes 9 through 11, respectively. RsaI profiles of the alveolysin gene fragmentderived from P. alvei isolates 30, 29, and 17 are shown in lanes 12 through 14,respectively. The molecular size markers used included a 100-bp ladder (lanes 1and 15) and lambda DNA digested with HindIII (lane 16). Restriction fragmentswere separated by using 1% (wt/vol) agarose and were stained with ethidiumbromide.
VOL. 66, 2000 GENETIC AND BIOCHEMICAL HETEROGENEITY OF P. ALVEI 1101
on August 19, 2020 by guest
http://aem.asm
.org/D
ownloaded from
ated patterns of the other isolates. DNA from isolates 5 (NewSouth Wales) and 9 (Tasmania) (Fig. 3, lanes 2 and 3) pro-duced similar REFPs, although some differences were evident.The CfoI-generated patterns of isolates 19 and 20 (both fromVictoria) were also very similar but could be distinguishedfrom one another (Fig. 3, lanes 13 and 14), and these twoisolates were considered to be clonal. The appearance ofdarkly staining fragments suggested that reiterated sequencesand/or extrachromosomal DNA was present in some isolates.
A longer electrophoretic time (21 h) highlighted the geneticheterogeneity of the P. alvei isolates, particularly in the high-molecular-weight region (7 to 9 kb) (Fig. 4 and 5). The simi-larity in the CfoI-generated profiles of isolates 19 and 20 wasreinforced when the longer electrophoretic time was used (Fig.4A, lanes 5 and 8). The CfoI profiles of DNA from Tasmania
isolates 12 and 13 (Fig. 4B, lanes 3 and 4) were indistinguish-able, and these isolates were considered to be clonal. TheCfoI-derived REFPs in the high-molecular-weight region (7 to9 kb) for isolates 28, 26, 27, and 24 (South Australia), isolate 16(Victoria), isolate 9 (Tasmania), and isolate 5 (New SouthWales) were identical with respect to three fragments (Fig. 3through 5). However, a comparison of the CfoI profilesthroughout the entire molecular size range (8.5 to 0.3 kb) forisolates 28 and 24 revealed RFLPs which distinguished theseprofiles from one another and from the profiles obtained forisolates 27, 26, and 16. The CfoI patterns of isolates 5 and 9were very similar to the CfoI pattern obtained for isolates 27,26, and 16 (Fig. 3 through 5). The close genetic relationships ofthe members of this subgroup of seven P. alvei isolates washighlighted when the CfoI-generated DNA profiles were com-
TABLE 2. Comparison of API 50CHB test results for P. alvei, P. azotofixans, P. macerans, P. polymyxa, P. larvae subsp. larvae, P. peoriae, andP. lautus isolates
Test
% Positive
P. alveia P.azotofixansb
P.maceransc
P.polymyxac
P. larvaesubsp.larvaed
P.peoriaee
P.lautus
Glycerol 100 0 100 100 100 vf 100Esculin 100 100 100 100 100 100 100Maltose 100 100 100 100 0 100 100D-Arabinose 0 0 93 7 0 100 100L-Arabinose 0 0 100 100 0 0 100Ribose 93 0 100 100 100 100 100D-Xylose 0 0 100 100 0 100 100a-Methylxyloside 0 0 100 100 0 100 100Rhamnose 0 0 100 33 0 v 0Dulcitol 0 47 0 0 0 0 0Inositol 0 0 40 0 0 0 0Sorbitol 0 0 73 0 0 0 01-Methyl-D-
mannoside0 9 100 40 0 100 0
Arbutin 68 67 100 100 0 100 100Salicin 89 73 100 100 v 100 100Lactose 0 0 100 100 0 100 100Inulin 0 100 100 87 0 0 vMelezitose 0 100 100 40 0 0 vStarch 64 32 100 100 0 100 100Glycogen 0 32 100 100 0 100 100D-Tagatose 0 41 0 0 v 0 0L-Fucose 0 0 87 0 0 0 vD-Arabitol 0 0 80 7 0 0 0Gluconate 0 9 93 80 0 v v5-Ketogluconate 0 14 93 53 0 0 02-Ketogluconate 0 9 0 0 0 0 0Gentiobiose 39 100 100 100 0 100 100Raffinose 11 100 100 100 0 100 100Trehalose 54 100 100 100 100 0 100Sucrose 14 100 100 100 0 100 100Melibiose 11 100 100 100 0 100 100Cellobiose 86 47 100 100 0 100 100Amygdalin 46 .90 100 100 0 100 100N-Acetylglucosamine 96 15 20 20 100 v 100Mannose 14 100 100 100 100 100 100Glucose 96 100 100 100 100 100 100Galactose 43 100 100 100 v 100 100Adonitol 50 0 100 100 0 0 0D-Fucose 0 0 0 0 0 0 0
a Data from this study.b Data from reference 26.c Data from reference 21.d Data from reference 13.e Data from reference 14.f v, variable number of isolates fermented the carbohydrate.
1102 DJORDJEVIC ET AL. APPL. ENVIRON. MICROBIOL.
on August 19, 2020 by guest
http://aem.asm
.org/D
ownloaded from
pared on the same polyacrylamide gel (Fig. 6). The CfoI pro-files of isolates 16, 26, and 27 were indistinguishable, while theprofiles of isolates of 28, 24, and 9 were identical except for thepresence of several heavily stained fragments. The CfoI profileof isolate 5 was also very similar to the profiles of the other sixisolates, although there were several more distinguishing fea-tures (Fig. 6, lane 12). The high level of genetic similarityamong these seven geographically diverse isolates suggestedthat they are clonally related. The CfoI profiles of DNA fromNew South Wales isolates 14, 8, and 7 (Fig. 6, lanes 1, 4, and5) had many similarities throughout the entire molecular sizerange, although numerous RFLPs made it easier to rapidlydistinguish between the profiles. The profile of isolate 1 (NewSouth Wales) exhibited some minor similarities to the profilesof the other three New South Wales isolates (isolates 14, 7, and8) but was more distantly related. The CfoI pattern of Victoriaisolate 11 was clearly different from the other 11 patternsshown in Fig. 6. The presence of numerous RFLPs in theremaining isolates suggests that P. alvei is a genetically heter-ogeneous species.
Compared with the restriction endonuclease patterns, thebiochemical profiles were unreliable indicators of similaritybetween isolates. Clonal isolates 9 and 16 produced almostidentical fermentation patterns for the 49 carbohydratesexamined (isolate 9 fermented galactose, but isolate 16 didnot ferment galactose), as did genetically indistinguishableisolates 19 and 20 (isolate 20 fermented adonitol, but isolate19 did not ferment adonitol). However, a comparison ofclonal isolate 16 with isolates 26 and 27 revealed seven(galactose, amygdalin, arbutin, melibiose, sucrose, raffinose,and gentibiose) and six (galactose, amygdalin, melibiose,sucrose, starch, and raffinose) pattern differences, respec-tively. Isolates that had very dissimilar CfoI profiles and
originated from geographically separated regions of Austra-lia, such as isolates 22 (Queensland) and 12 (Tasmania),exhibited nine differences in their carbohydrate patterns,while isolates 6 (New South Wales) and 13 (Tasmania) ex-hibited only three differences.
FIG. 3. Silver-stained 3.5% polyacrylamide gel of CfoI-generated REFPs ofgeographically diverse Australian isolates of P. alvei. The gel was electropho-resed for 15 h and shows DNA fragments between 0.45 and 8.5 kb long. Thelanes show the DNA profiles for New South Wales isolate 5, Tasmania isolates9 and 12, South Australia isolates 25 and 27, Victoria isolates 16 through 18,South Australia isolates 29 and 28, Tasmania isolate 15, Victoria isolates 19 and20, and Queensland isolate 21; isolate numbers are indicated at the top. Lane Mcontained molecular size markers (lambda DNA digested with HindIII).
FIG. 4. CfoI-generated REFPs of P. alvei DNA recovered from geographi-cally diverse Australian isolates. Gels (3.5% polyacrylamide) were electropho-resed for 21 h to separate large molecular fragments and were stained with silver.(A) REFPs for isolates 30 and 31 (origin unknown), New South Wales isolates4 and 6, and Victoria isolates 19, 17, 18, and 20. Isolate numbers are indicated atthe top. The positions of molecular size markers (lambda DNA digested withHindIII) are indicated on the left. (B) REFPs for Queensland isolates 22 and 21,Tasmania isolates 12 and 13, South Australia isolates 29 and 24, New SouthWales isolate 2, Tasmania isolate 15, and South Australia isolate 28. Isolatenumbers are indicated at the top.
VOL. 66, 2000 GENETIC AND BIOCHEMICAL HETEROGENEITY OF P. ALVEI 1103
on August 19, 2020 by guest
http://aem.asm
.org/D
ownloaded from
Extrachromosomal DNA. Extrachromosomal DNA ele-ments were observed in the whole-cell DNA (undigested) pro-files of 17 of the 30 P. alvei isolates (isolates 1, 2, 7 through 9,13, 14, 17 through 21, 25, and 28 through 31). The molecular
sizes of these elements ranged from 1 to 15 kb (Fig. 7). Someisolates produced multiple bands, suggesting that more thanone extrachromosomal element may be present in these iso-lates.
DISCUSSION
Although P. alvei was one of the first species described in thegenus Bacillus, comparatively little is known about this organ-ism. Using REA and a range of biochemical tests, we demon-strated that geographically diverse Australian isolates of P.alvei are genetically and biochemically heterogeneous, al-though seven isolates (isolates 5, 9, 16, 24, and 26 through 28)that were obtained from four states of Australia were consid-ered to be clonal. Isolates 19 and 20 (both from Victoria) wereconsidered to be clonal, as were Tasmanian isolates 12 and 13.However, the CfoI patterns that were representative of each ofthe clonal clusters were clearly different from one another andfrom the patterns of the remaining P. alvei isolates. Twenty-fivedifferent CfoI-generated profiles were discerned when 30 Aus-tralian isolates were examined, suggesting that REA may be auseful tool for tracing the movement of P. alvei isolates inepidemiological studies. While 17 of the 30 P. alvei isolatescontained extrachromosomal nucleic acid species which mayhave contributed to differences in the restriction patterns, iso-lates which did not contain extrachromosomal elements alsoexhibited marked genetic heterogeneity. The alveolysin genewas amplified by PCR from only 19 of the 30 isolates, andRFLPs in the alveolysin gene were revealed by using the en-zymes CfoI, RsaI, and Sau3AI. Southern hybridization andimmunoblot studies should be performed to confirm the pres-ence of the alveolysin gene and to determine if this gene isactively expressed in these and other isolates of P. alvei derivedfrom ecologically diverse habitats.
The genetic heterogeneity of Australian P. alvei isolates re-vealed by REA contrasts with the intragenic homogeneity ofthe REFPs of the 16S rRNA gene. The restriction endonucle-ase profiles of the 16S rRNA gene amplified from DNA of the
FIG. 5. CfoI profiles of P. alvei DNA isolated from eastern and southernstates of Australia. The lanes show the profiles for Victoria isolates 18 and 20,South Australia isolates 25 and 27, New South Wales isolates 3 and 5, Tasmaniaisolate 15, and South Australia isolates 28 and 23. Isolate numbers are indicatedat the top. Lane M contained molecular size markers (phage SPPI digested withEcoRI).
FIG. 6. CfoI profiles of a subset of genetically related P. alvei isolates. Thelanes show the profiles for New South Wales isolate 14, Tasmania isolate 11, NewSouth Wales isolates 1, 8, and 7, South Australia isolates 24, 28, 27, and 26,Victoria isolate 16, Tasmania isolate 9, and New South Wales isolate 5. Isolatenumbers are indicated at the top. Lane M contained a 1-kb ladder (Bio-Rad).
FIG. 7. Whole-cell DNA profiles of P. alvei isolates on 1.0% (wt/vol) agarosegels. The lane designations correspond to isolate numbers. Extrachromosomalelements (arrowheads) were observed in the whole-cell DNA profiles of 17isolates. Lane M contained a 1-kb ladder.
1104 DJORDJEVIC ET AL. APPL. ENVIRON. MICROBIOL.
on August 19, 2020 by guest
http://aem.asm
.org/D
ownloaded from
30 isolates were remarkably consistent for each of six enzymes,which confirmed that these isolates were P. alvei isolates. Ashet al. (1) placed P. alvei together with nine other species in aseparate phylogroup (group 3) based on a distance matrixanalysis of 16S rRNA gene sequence comparisons and split thegenus Bacillus into five major clusters representing separategenera. Rossler et al. (27) showed that P. alvei was most closelyrelated to Bacillus macerans, Bacillus macquariensis, and Ba-cillus polymyxa, and the levels of 16S rRNA gene sequencesimilarity ranged from 93 to 94%. The results of RFLP anal-yses of the 16S rRNA gene performed with DraI and HinfIclearly differentiated P. alvei from these phylogenetically re-lated species.
Morphology (swarming ability and spore morphology) andbiochemical tests (API 50CHB tests, as well as several supple-mentary tests) were sufficiently discriminatory to confirm thegenus and species identities of the isolates (Table 2). The API50CHB system was most useful for biochemical characteriza-tion of the species and revealed considerable variability amongthe 28 Australian isolates tested. Logan and Berkeley (22) useda number of morphological and biochemical tests (includingAPI 50CHB tests) to characterize 1,075 strains representingthe genus Bacillus. A comparison of the results obtained for 12P. alvei strains in their study with the results for the 28 strainsused in our study revealed many similarities. Variations wereobserved for 14 carbohydrates (ribose, lactose, melezitose,starch, glycogen, D-turanose, L-fucose, adonitol, D-glucose, ino-sitol, gluconate, a-methyl-D-glucoside, N-acetylglucosamineand 5-ketogluconate). The major differences involved the abil-ity of the majority of the isolates in the study of Logan andBerkeley to ferment a-methyl-D-glucoside, D-turanose, glyco-gen, and inositol; none of the Australian isolates fermentedthese carbohydrates. The remaining differences were minorand involved variability in the percentage of isolates that wereable to ferment some carbohydrates (ribose, N-acetylglu-cosamine, adonitol, and glucose). In addition, a small percent-age of isolates (less than 25%) in the study of Logan andBerkeley fermented lactose, melezitose, gluconate, D-fucose,and 5-ketogluconate; these carbohydrates were not fermentedby any of the Australian isolates. The sources and geographiclocations of the P. alvei isolates used in the study of Logan andBerkeley were not reported, and these isolates may not haveoriginated from honeybee colonies. Collectively, the data de-scribed above confirm the biochemical heterogeneity of P.alvei.
A polyphasic reassessment of the genus Paenibacillus re-vealed limited heterogeneity among seven P. alvei isolates (14).Sodium dodecyl sulfate-polyacrylamide gel electrophoresis,API test, and pyrolysis mass spectrometry results did not groupthe P. alvei isolates into a single cluster, even though theorganisms appeared to be members of a phylogenetically ho-mogeneous species when ARDRA was used. Among Paeniba-cillus species, the ability to produce indole and the ability togenerate acid via fermentation of adonitol are features thatdistinguish P. alvei (14). Of the 28 Australian P. alvei isolatesused in our study, only 50% fermented adonitol, while 24(86%) produced indole. Some of the Australian isolates (11 of28 isolates) fermented gentiobiose, while all seven Europeanisolates fermented this sugar (14). The ability of all 28 Austra-lian isolates to ferment maltose and glycerol, the ability ofsome isolates to ferment salicin and D-trehalose, and the in-ability of the isolates to ferment xylose, arabinose, a-methyxy-loside, D-fructose, mannitol, and methyl-D-mannoside are con-sistent with the findings of Heyndrickx et al. (14).
While morphologic and biochemical analyses may be infor-mative for identification of P. alvei, there was biochemical
variability even among clonal isolates that had very similarDNA profiles. This is perhaps not unexpected since the abilityto ferment a carbohydrate often depends on expression ofseveral functional genes, including genes involved in an enzy-matic reaction(s) and in transport of the carbohydrate acrossthe cell membrane. Genetic homogeneity demonstrated by fin-gerprint profiles is largely a measure of genome stability andmay not reflect the presence of point mutations, deletions, andsmall sequence rearrangements which fail to dramatically alterfingerprint profiles but which may have profound conse-quences for transcription and translation. Chromosomal insta-bility through genetic rearrangement is the likely source of theheterogeneity observed in the restriction patterns of bacteria(12). Invertrons, insertion elements, transposons, and conjuga-tive transposons contribute to site-specific recombinationevents which generate heterogeneity in genomes, and conse-quently, genetic heterogeneity is a feature that is common inbacterial genomes (20).
Genetic diversity is considered essential for bacterial adap-tation to fluctuating environments (31). Investigations of ge-netic heterogeneity among strains of P. polymyxa and P. azoto-fixans, which are phylogenetically closely related to P. alvei,have been described previously (23, 29). A cluster analysis ofphenotypic features revealed that all strains of P. polymyxaexhibited a high degree of similarity but that strain diversitydepended on strain origin. Strains recovered from the rhizo-plane (a close association with wheat roots) exhibited thesmallest amount of variability and clustered into one pheno-typic group, while strains recovered further from the rhizo-plane exhibited greater genetic, serological, and phenotypicdiversity. Strains that were firmly attached to wheat roots fer-mented sorbitol and were genetically homogeneous (23). Sim-ilarly, the genetic diversity within populations of P. azotofixanswas affected by the soil type. Seldin et al. (29) showed thatstrains recovered from different soil types generally belongedto different but related groups which could be distinguished onthe basis of their ability to ferment different carbohydrates inAPI 50CHB tests.
Unlike P. azotofixans and P. polymyxa, which form closeassociations with roots of grain crops, P. alvei not only is re-covered from honeybee colonies with EFB but also is com-monly isolated from very different environmental niches, in-cluding soil, milk, various insect species, and humans (4, 11, 24,26, 30). The high degree of genetic heterogeneity revealed bygenomic profiles and the biochemical variability of this speciessupport the hypothesis that P. alvei is not a primary honeybeepathogen (3). The two primary bacterial pathogens of honey-bee colonies (M. pluton and P. larvae subsp. larvae) exhibitmarked genetic, protein, and antigenic homogeneity, as deter-mined by REA, sodium dodecyl sulfate-polyacrylamide gelelectrophoresis, and immunoblot analyses (7, 9, 16), whichsuggests that these organisms evolved so that they form closehost-parasite relationships with honeybee larvae. Although iso-lates 5, 9, 16, 24, and 26 through 28, which represent 23% ofour isolates, were clonally related, biochemically they wereamong the most diverse isolates. They were not amalgamatedinto a single group until the lowest similarity value for thewhole group of isolates (54%) (Fig. 1). Further studies will benecessary to determine if the disproportionately high occur-rence of this genotype represents the emergence of a predom-inant clone or if a much simpler explanation (e.g., migratorypatterns of Australian beekeepers) is responsible for this ob-servation.
VOL. 66, 2000 GENETIC AND BIOCHEMICAL HETEROGENEITY OF P. ALVEI 1105
on August 19, 2020 by guest
http://aem.asm
.org/D
ownloaded from
ACKNOWLEDGMENTS
This work was funded in part by the Australian Honey Bee Researchand Development Committee.
We acknowledge the efforts of Liz Summerell, who provided tech-nical assistance for the culture of P. alvei. The expert photographicskills of Lowan Turton and the assistance of Paul Nicholls in preparingthe dendrogram are gratefully acknowledged.
REFERENCES
1. Ash, C., J. A. E. Farrow, S. Wallbanks, and M. D. Collins. 1991. Phylogeneticheterogeneity of the genus Bacillus revealed by comparative analysis ofsmall-subunit-ribosomal RNA sequences. Lett. Appl. Microbiol. 13:202–206.
2. Ash, C., F. G. Priest, and M. D. Collins. 1993. Molecular identification ofrRNA group 3 bacilli (Ash, Farrow, Wallbanks and Collins) using a PCRprobe test. Antonie Leewenhoek J. Microbiol. 64:253–260.
3. Bailey, L., and B. V. Ball. 1991. Honey bee pathology, 2nd ed., p. 36–41.Harcourt Brace Jovanovich, Sidcup, United Kingdom.
4. Balaraman, K., U. S. B. Rao, and P. K. Rajagopalan. 1979. Bacterial patho-gens of mosquito larvae—Bacillus alvei (Cheshire and Cheyene) and Bacillusbrevis (Migula) isolated in Pondicherry. Indian J. Med. Res. 70:615–619.
5. Cowan, S. T., and K. J. Steele. 1993. Appendix C: characterisation tests, p.219–238. In G. I. Barrow and R. K. A. Feltham (ed.), Manual for theidentification of medical bacteria. Cambridge University Press, Cambridge,United Kingdom.
6. Djordjevic, S. P., G. J. Eamens, H. Ha, M. J. Walker, and J. C. Chin. 1998.Demonstration that Australian Pasteurella multocida isolates from sporadicoutbreaks of porcine pneumonia are non-toxigenic (toxA2) and display het-erogeneous DNA restriction endonuclease profiles compared with toxigenicisolates from herds with progressive atrophic rhinitis. J. Med. Microbiol.47:679–688.
7. Djordjevic, S. P., M. HoShon, and M. A. Z. Hornitzky. 1994. DNA restrictionendonuclease profiles and typing of geographically diverse isolates of Bacil-lus larvae. J. Apic. Res. 33:95–103.
8. Djordjevic, S. P., K. Noone, L. Smith, and M. A. Z. Hornitzky. 1998. Devel-opment of a hemi-nested PCR assay for the specific detection of Melisso-coccus pluton. J. Apic. Res. 37:165–173.
9. Djordjevic, S. P., L. A. Smith, W. A. Forbes, and M. A. Z. Hornitzky. 1999.Geographically diverse Australian isolates of Melissococcus pluton exhibitminimal genotypic diversity by restriction endonuclease analysis. FEMS Mi-crobiol. Lett. 173:311–318.
10. Geoffroy, C., J. Mengaud, J. E., Alouf, and P. Cossart. 1990. Alveolysin, thethiol-activated toxin of Bacillus alvei, is homologous to listeriolysin O, per-fringolysin O, pneumolysin, and streptolysin O and contains a single cysteine.J. Bacteriol. 172:7301–7305.
11. Gillam, M. 1985. Microbes from apiarian sources: Bacillus spp. in fras of thegreater wax moth. J. Invertebr. Pathol. 45:218–224.
12. Hall, L. M. C. 1994. Are point mutations or DNA rearrangements respon-sible for the restriction fragment length polymorphisms that are used to typebacteria? Microbiology 140:197–204.
13. Heyndrickx, M., K. Vandemeulebroecke, B. Hoste, P. Janssen, K. Kersters,P. De Vos, N. A. Logan, N. Ali, and R. C. W. Berkeley. 1996. Reclassificationof Paenibacillus (formerly Bacillus) pulvifaciens (Nakamura 1984) Ash et al.1994, a later subjective synonym of Paenibacillus (formerly Bacillus) larvae(White 1906) Ash et al. 1994, as a subspecies of P. larvae, with emended
descriptions of P. larvae as P. larvae subsp. larvae and P. larvae subsp.pulvifaciens. Int. J. Syst. Bacteriol. 46:270–279.
14. Heyndrickx, M., K. Vandemeulebroecke, P. Scheldeman, K. Kersters, P. DeVos, N. A. Logan, A. M. Aziz, N. Ali, and R. C. W. Berkeley. 1996. Apolyphasic study of the genus Paenibacillus, reclassification of Bacillus lautus(Nakamura 1984) as Paenibacillus lautus comb. nov. and of Bacillus peoriae(Montefusco et al. 1993) as Paenibacillus peoriae comb. nov., and emendeddescriptions of P. lautus and P. peoriae. Int. J. Syst. Bacteriol. 46:988–1003.
15. Heyndrickx, M., L. Vauterin, P. Vandamme, K. Kersters, and P. De Vos.1996. Applicability of combined amplified ribosomal DNA restriction anal-ysis (ARDRA) patterns in bacterial phylogeny and taxonomy. J. Microbiol.Methods 26:247–259.
16. Hornitzky, M. A. Z., and S. P. Djordjevic. 1992. Sodium dodecyl sulfatepolyacrylamide gel electrophoresis profiles and western blots of Bacilluslarvae. J. Apic. Res. 31:47–49.
17. Hornitzky, M. A. Z., and S. Karlovskis. 1989. A culture technique for thedetection of Bacillus larvae in honey bees. J. Apic. Res. 28:118–120.
18. Hornitzky, M. A. Z., and L. Smith. 1998. Procedures for the culture ofMelissococcus pluton from diseased brood and bulked honey samples. J.Apic. Res. 37:292–294.
19. Hornitzky, M. A. Z., and S. C. Wilson. 1989. A system for the diagnosis of themajor bacterial brood diseases of honey bees. J. Apic. Res. 28:191–195.
20. Kolstø, A.-B. 1997. Dynamic bacterial genome organization. Mol. Microbiol.24:241–248.
21. Kristansen, B. E., B. Sorensen, T. Simonsen, O. Spanne, V. Lund, and B.Bjorvant. 1984. Isolates of Neisseria meningitidis from different sites in thesame patient: phenotypic and genomic studies, with special reference toadherence, piliation, and DNA restriction endonuclease pattern. J. Infect.Dis. 150:389–396.
22. Logan, N. A., and R. C. W. Berkeley. 1984. Identification of Bacillus strainsusing the API system. J. Gen. Microbiol. 130:1871–1882.
23. Mavingui, P., G. Laguerre, O. Berge, and T. Heulin. 1992. Genetic andphenotypic diversity of Bacillus polymyxa in soil and in wheat rhizosphere.Appl. Environ. Microbiol. 58:1894–1903.
24. Mukundan, M., M. Subrahmanyam, and M. N. Parameswaran. 1979. Stud-ies on microflora in boiled milk. J. Vet. Sci. 10:234–239.
25. Orskov, F., and I. Orskov. 1983. From the National Institutes of Health.Summary of a workshop on the clone concept in the epidemiology, taxon-omy, and evolution of the Enterobacteriaceae and other bacteria. J. Infect.Dis. 148:346–357.
26. Reboli, A. C., C. S. Bryan, and W. E. Farrar. 1989. Bacteremia and infectionof a hip prosthesis caused by Bacillus alvei. J. Clin. Microbiol. 27:1395–1396.
27. Rossler, D., L. Wolfgang, K. Schleifer, C. Lin, T. McGill, J. Wisotzkey, P.Jurtshuk, Jr., and G. Fox. 1991. Phylogenetic diversity in the genus Bacillusas seen by 16S rRNA sequencing studies. Syst. Appl. Microbiol. 14:266–269.
28. Seldin, L., and E. G. C. Penido. 1986. Identification of Bacillus azotofixansusing API tests. Antonie Leewenhoek J. Microbiol. 52:403–409.
29. Seldin, L., A. S. Rosado, D. W. da Cruz, A. Nobrega, J. D. van Elsas, and E.Paiva. 1998. Comparison of Paenibacillus azotofixans strains isolated fromrhizoplane, rhizosphere, and non-root-associated soil from maize planted intwo different brazilian soils. Appl. Environ. Microbiol. 64:3860–3868.
30. Smirnov, A. M. 1976. Disinfection of the soil of apiaries infected with thebacteria of bee diseases. Veterinariya (Moscow) 6:32–37.
31. Torsvik, V., J. Goksoyr, and F. L. Daae. 1990. High diversity in DNA of soilbacteria. Appl. Environ. Microbiol. 56:782–787.
1106 DJORDJEVIC ET AL. APPL. ENVIRON. MICROBIOL.
on August 19, 2020 by guest
http://aem.asm
.org/D
ownloaded from