of 6/6
JOURNAL OF BACTERIOLOGY, 0021-9193/97/$04.0010 May 1997, p. 3244–3249 Vol. 179, No. 10 Copyright © 1997, American Society for Microbiology Genome Structure and Phylogeny in the Genus Brucella SYLVIE MICHAUX-CHARACHON, GISELE BOURG, ESTELLE JUMAS-BILAK, PATRICIA GUIGUE-TALET, ANNICK ALLARDET-SERVENT, DAVID O’CALLAGHAN, AND MICHEL RAMUZ* Unite ´ 431, Faculte ´ de Me ´decine, Institut National de la Sante ´ et de la Recherche Me ´dicale, 30900 Nı ˆmes, France Received 21 October 1996/Accepted 7 March 1997 PacI and SpeI restriction maps were obtained for the two chromosomes of each of the six species of the genus Brucella: B. melitensis, B. abortus, B. suis, B. canis, B. ovis, and B. neotomae. Three complementary techniques were used: hybridization with the two replicons as probes, cross-hybridization of restriction fragments, and a new mapping method. For each type strain, a unique I-SceI site was introduced in each of the two replicons, and the location of SpeI sites was determined by linearization at the unique site, partial digestion, and end labeling of the fragments. The restriction and genetic maps of the six species were highly conserved. However, numerous small insertions or deletions, ranging from 1 to 34 kb, were observed by comparison with the map of the reference strain of the genus, B. melitensis 16M. A 21-kb SpeI fragment specific to B. ovis was found in the small chromosome of this species. A 640-kb inversion was demonstrated in the B. abortus small chromo- some. All of these data allowed the construction of a phylogenetic tree, which reflects the traditional phenetic classification of the genus. Brucella is a small gram-negative bacterium pathogenic for animals and occasionally for humans. The genus is divided into six species on the basis of cultural, metabolic, and antigenic characteristics: B. melitensis, B. abortus, B. suis, B. ovis, B. canis, and B. neotomae (6). However, it was shown by DNA-DNA hybridization that the genus is more probably monospecific (42). In a previous work (1), we found that each of the so- called species exhibited a specific macrorestriction pattern by digestion of genomic DNA with XbaI, with only minor varia- tions between biovars within a species. Recently, a physical map was proposed for the genus type strain, B. melitensis 16M, by a cross-hybridization method (30). We showed that the Brucella genome was constituted by two circular chromosomes with sizes of about 2.05 and 1.15 Mb. In this study, we established the PacI and SpeI restriction maps of the two chromosomes of the other reference strains in the genus Brucella: B. abortus 544, B. suis 1330, B. canis 62/290, B. ovis 63/290, and B. neotomae 5K33. Three different but complementary techniques were employed: hybridization with the two replicons as probes, cross-hybridization of restriction fragments, and a new mapping strategy using I-SceI. I-SceI is an endonuclease encoded by a group I intron of the Saccha- romyces cerevisiae mitochondrial 21S rRNA gene (31). Re- cently, by using artificially inserted I-SceI sites, a physical map of chromosome XI of S. cerevisiae (40) and a physical map of YAC inserts (5) were made. Since bacterial genomes lack natural I-SceI sites, we showed that the introduction of a unique restriction site recognized by I-SceI could be a useful tool for the study of the genome organization (20). In this study, we used an indirect end-labeling method to localize the SpeI sites, derived from a previously described procedure in- volving partial digestion of linearized DNA, followed by South- ern blotting and hybridization with two end probes (37). Comparison of the six physical maps showed a rather high conservation of restriction sites among the six Brucella species genomes, even though a large inversion was revealed in the B. abortus 544 small chromosome. MATERIALS AND METHODS Bacterial strains. Bacterial strains used in the study are listed in Table 1. Brucella strains were grown on a low-salt 2YT medium (10 g of yeast extracts, 10 g of tryptone, 5 g of NaCl per liter) and Escherichia coli on LB medium. Conjugation experiments. Introduction of a unique restriction I-SceI site in each replicon of the different Brucella species was described elsewhere (20). Briefly, the suicide vector pGF2, containing the Tn5Map with the 18-bp I-SceI recognition site (18), was transferred from E. coli SM10 lpir into Brucella mu- tants resistant to either streptomycin or nalidixic acid. Transposition events were selected by kanamycin (25 mg/ml) and streptomycin resistance or kanamycin and nalidixic acid resistance according to the resistant Brucella mutant strain used. For mapping experiments, we used five insertion mutants for B. melitensis, two for B. abortus, five for B. suis, and eight for B. ovis. Endonuclease digestion of Brucella genomic DNA and pulsed-field gel electro- phoresis (PFGE) conditions. Intact Brucella genomic DNA was prepared in agarose plugs and digested with rare-cutting endonucleases as previously de- scribed (30). For SpeI partial digestion, a 120-min diffusion period at 4°C fol- lowed by a 10-min digestion at 25°C gave optimal results with 3 U of endonu- clease per plug (2 mg of DNA). For double digestions, plugs were first partially digested with SpeI, incubated overnight at 37°C in 0.5 M EDTA (pH 8) contain- ing proteinase K (1 mg/ml) and lauroyl sarcosine (1%), and then washed for 1 h in Tris-EDTA-phenylmethylsulfonyl fluoride (TE-PMSF) and three times in TE. Digestion with I-SceI (Boehringer Mannheim) was then performed as described elsewhere (20). PFGE of intact and digested DNA was performed in a contour- clamped homogeneous electric field (CHEF) apparatus (CHEF-DRII; Bio-Rad) as previously described (30). Separation of double-digested DNA was carried out for 40 h in 1% agarose gels at 170 V by using two sets of pulse conditions, in separate runs, in order to estimate the fragment sizes more precisely: 100 to 12 s for separating the large bands and 50 to 10 s for separating the small bands. Hybridization experiments. Southern blots of PFGE agarose gels were pre- pared on Nytran membranes (Schleicher & Schuell) by using the VacuGene XL Blotting System (Pharmacia LKB) with 203 SSC (13 SSC is 0.15 M NaCl plus 0.015 M sodium citrate) as the eluant. Preparation of probes for cross-hybrid- ization of restriction fragments or bands separated from intact DNA was done as previously described (30). To obtain the 1.1- and 2.7-kb probes, corresponding to the ends of the I-SceI-linearized Tn5Map, plasmid ColE1::Tn5Map DNA (18) was successively digested by SphI, NotI, and then I-SceI. After separation on 1% low-melting-temperature agarose (SeaPlaque; FMC), the 1.1- and 2.7-kb restric- tion fragments were excised from the gel and labeled by a nonisotopic method (digoxigenin-DNA labeling and detection kit; Boehringer). The Brucella DNA size markers were revealed with a total genomic DNA probe. Several genes were located on the maps by hybridizing SpeI restriction fragments with cloned or amplified (PCR) genes as described previously (30). Probes for pal, bfr, and P16.5 (8, 16, 41) were kindly provided by J. J. Letesson; for omp19 by A. Tibor; for brrP (phoP homolog), brrN (ntrC homolog), htrA (38), and recA (39) by B. Wren; for ery (36) by J. Aguerro; and for galE by L. Petrovska. A probe for dnaA was obtained by amplification using oligonucleotides calculated from the sequence of this gene in Rhizobium meliloti (29). brrA is the gene of a response regulator amplified in our laboratory which is similar to ctrA described in Caulobacter crescentus (unpublished data). * Corresponding author. Phone: (33) 66 23 46 79. Fax: (33) 66 23 66 52. 3244 on October 21, 2020 by guest http://jb.asm.org/ Downloaded from

Genome Structure and Phylogeny in the Genus …Brucella genome was constituted by two circular chromosomes with sizes of about 2.05 and 1.15 Mb. In this study, we established the PacI

  • View
    1

  • Download
    0

Embed Size (px)

Text of Genome Structure and Phylogeny in the Genus …Brucella genome was constituted by two circular...

  • JOURNAL OF BACTERIOLOGY,0021-9193/97/$04.0010

    May 1997, p. 3244–3249 Vol. 179, No. 10

    Copyright © 1997, American Society for Microbiology

    Genome Structure and Phylogeny in the Genus BrucellaSYLVIE MICHAUX-CHARACHON, GISELE BOURG, ESTELLE JUMAS-BILAK, PATRICIA GUIGUE-TALET,

    ANNICK ALLARDET-SERVENT, DAVID O’CALLAGHAN, AND MICHEL RAMUZ*

    Unité 431, Faculté de Médecine, Institut National de la Santé et de la Recherche Médicale,30900 Nı̂mes, France

    Received 21 October 1996/Accepted 7 March 1997

    PacI and SpeI restriction maps were obtained for the two chromosomes of each of the six species of the genusBrucella: B. melitensis, B. abortus, B. suis, B. canis, B. ovis, and B. neotomae. Three complementary techniqueswere used: hybridization with the two replicons as probes, cross-hybridization of restriction fragments, and anew mapping method. For each type strain, a unique I-SceI site was introduced in each of the two replicons,and the location of SpeI sites was determined by linearization at the unique site, partial digestion, and endlabeling of the fragments. The restriction and genetic maps of the six species were highly conserved. However,numerous small insertions or deletions, ranging from 1 to 34 kb, were observed by comparison with the mapof the reference strain of the genus, B. melitensis 16M. A 21-kb SpeI fragment specific to B. ovis was found inthe small chromosome of this species. A 640-kb inversion was demonstrated in the B. abortus small chromo-some. All of these data allowed the construction of a phylogenetic tree, which reflects the traditional pheneticclassification of the genus.

    Brucella is a small gram-negative bacterium pathogenic foranimals and occasionally for humans. The genus is divided intosix species on the basis of cultural, metabolic, and antigeniccharacteristics: B. melitensis, B. abortus, B. suis, B. ovis, B. canis,and B. neotomae (6). However, it was shown by DNA-DNAhybridization that the genus is more probably monospecific(42). In a previous work (1), we found that each of the so-called species exhibited a specific macrorestriction pattern bydigestion of genomic DNA with XbaI, with only minor varia-tions between biovars within a species. Recently, a physicalmap was proposed for the genus type strain, B. melitensis 16M,by a cross-hybridization method (30). We showed that theBrucella genome was constituted by two circular chromosomeswith sizes of about 2.05 and 1.15 Mb.

    In this study, we established the PacI and SpeI restrictionmaps of the two chromosomes of the other reference strains inthe genus Brucella: B. abortus 544, B. suis 1330, B. canis 62/290,B. ovis 63/290, and B. neotomae 5K33. Three different butcomplementary techniques were employed: hybridization withthe two replicons as probes, cross-hybridization of restrictionfragments, and a new mapping strategy using I-SceI. I-SceI isan endonuclease encoded by a group I intron of the Saccha-romyces cerevisiae mitochondrial 21S rRNA gene (31). Re-cently, by using artificially inserted I-SceI sites, a physical mapof chromosome XI of S. cerevisiae (40) and a physical map ofYAC inserts (5) were made. Since bacterial genomes lacknatural I-SceI sites, we showed that the introduction of aunique restriction site recognized by I-SceI could be a usefultool for the study of the genome organization (20). In thisstudy, we used an indirect end-labeling method to localize theSpeI sites, derived from a previously described procedure in-volving partial digestion of linearized DNA, followed by South-ern blotting and hybridization with two end probes (37).

    Comparison of the six physical maps showed a rather highconservation of restriction sites among the six Brucella speciesgenomes, even though a large inversion was revealed in the B.abortus 544 small chromosome.

    MATERIALS AND METHODS

    Bacterial strains. Bacterial strains used in the study are listed in Table 1.Brucella strains were grown on a low-salt 2YT medium (10 g of yeast extracts,10 g of tryptone, 5 g of NaCl per liter) and Escherichia coli on LB medium.

    Conjugation experiments. Introduction of a unique restriction I-SceI site ineach replicon of the different Brucella species was described elsewhere (20).Briefly, the suicide vector pGF2, containing the Tn5Map with the 18-bp I-SceIrecognition site (18), was transferred from E. coli SM10 lpir into Brucella mu-tants resistant to either streptomycin or nalidixic acid. Transposition events wereselected by kanamycin (25 mg/ml) and streptomycin resistance or kanamycin andnalidixic acid resistance according to the resistant Brucella mutant strain used.For mapping experiments, we used five insertion mutants for B. melitensis, twofor B. abortus, five for B. suis, and eight for B. ovis.

    Endonuclease digestion of Brucella genomic DNA and pulsed-field gel electro-phoresis (PFGE) conditions. Intact Brucella genomic DNA was prepared inagarose plugs and digested with rare-cutting endonucleases as previously de-scribed (30). For SpeI partial digestion, a 120-min diffusion period at 4°C fol-lowed by a 10-min digestion at 25°C gave optimal results with 3 U of endonu-clease per plug (2 mg of DNA). For double digestions, plugs were first partiallydigested with SpeI, incubated overnight at 37°C in 0.5 M EDTA (pH 8) contain-ing proteinase K (1 mg/ml) and lauroyl sarcosine (1%), and then washed for 1 hin Tris-EDTA-phenylmethylsulfonyl fluoride (TE-PMSF) and three times in TE.Digestion with I-SceI (Boehringer Mannheim) was then performed as describedelsewhere (20). PFGE of intact and digested DNA was performed in a contour-clamped homogeneous electric field (CHEF) apparatus (CHEF-DRII; Bio-Rad)as previously described (30). Separation of double-digested DNA was carried outfor 40 h in 1% agarose gels at 170 V by using two sets of pulse conditions, inseparate runs, in order to estimate the fragment sizes more precisely: 100 to 12 sfor separating the large bands and 50 to 10 s for separating the small bands.

    Hybridization experiments. Southern blots of PFGE agarose gels were pre-pared on Nytran membranes (Schleicher & Schuell) by using the VacuGene XLBlotting System (Pharmacia LKB) with 203 SSC (13 SSC is 0.15 M NaCl plus0.015 M sodium citrate) as the eluant. Preparation of probes for cross-hybrid-ization of restriction fragments or bands separated from intact DNA was done aspreviously described (30). To obtain the 1.1- and 2.7-kb probes, corresponding tothe ends of the I-SceI-linearized Tn5Map, plasmid ColE1::Tn5Map DNA (18)was successively digested by SphI, NotI, and then I-SceI. After separation on 1%low-melting-temperature agarose (SeaPlaque; FMC), the 1.1- and 2.7-kb restric-tion fragments were excised from the gel and labeled by a nonisotopic method(digoxigenin-DNA labeling and detection kit; Boehringer). The Brucella DNAsize markers were revealed with a total genomic DNA probe. Several genes werelocated on the maps by hybridizing SpeI restriction fragments with cloned oramplified (PCR) genes as described previously (30). Probes for pal, bfr, and P16.5(8, 16, 41) were kindly provided by J. J. Letesson; for omp19 by A. Tibor; for brrP(phoP homolog), brrN (ntrC homolog), htrA (38), and recA (39) by B. Wren; forery (36) by J. Aguerro; and for galE by L. Petrovska. A probe for dnaA wasobtained by amplification using oligonucleotides calculated from the sequence ofthis gene in Rhizobium meliloti (29). brrA is the gene of a response regulatoramplified in our laboratory which is similar to ctrA described in Caulobactercrescentus (unpublished data).

    * Corresponding author. Phone: (33) 66 23 46 79. Fax: (33) 66 23 6652.

    3244

    on October 21, 2020 by guest

    http://jb.asm.org/

    Dow

    nloaded from

    http://jb.asm.org/

  • Phylogenetic analysis. Similarity between each pair of strains was calculated asthe Dice coefficient as described by Grothues and Tümmler (13), with one mod-ification: two fragments with the same size were considered as equivalent only ifthey cohybridized. Strains were clustered by using the Phylip program (10).

    RESULTS

    Comparative PFGE-based macrorestriction analysis anddetermination of genome size. The genomic DNA from refer-ence strains of the six Brucella species (Table 1) was digestedwith the rare-cutting endonucleases PacI (giving 9 to 10 frag-ments) and SpeI (26 to 28 fragments) and subjected to PFGE(Fig. 1). It could be seen that each species exhibited a specificrestriction pattern with the two endonucleases, except for B.suis 1330 and B. canis, which had the same patterns. PFGE ofintact genomic DNA revealed, as for B. melitensis 16M (30),the presence of two chromosomes in the other species (datanot shown). To recognize to which replicon belong the PacIand SpeI restriction fragments, the bands corresponding to thetwo chromosomes of each Brucella species undigested genomicDNA were excised from the PFGE gel, labeled, and used asprobes for hybridizing Southern blots of the same DNA di-gested with PacI or SpeI. The addition of restriction fragmentsizes allowed a good estimation of the two replicon sizes foreach Brucella species, and these were always close to 2.05 and1.15 Mb, except for the small chromosomes of B. suis 1330, B.canis, and B. neotomae, which were 50 kb longer. Then, foreach species, the PacI restriction fragments were used as link-ing probes on Southern blots of genomic DNA digested withSpeI. As the large PacI restriction fragments, ranging from 900to 260 kb, hybridized with more than two SpeI restrictionfragments, another technique was needed to localize all theSpeI sites within each PacI fragment.

    SpeI physical mapping by introducing a unique I-SceI site ineach replicon. We used a Tn5 derivative, Tn5Map, to intro-duce a unique I-SceI site into both replicons of each of the sixBrucella species, as described previously (20). For each species,Tn5Map insertion mutant clones were selected and the geno-mic DNA was digested with SpeI, separated by PFGE, andtested by Southern hybridization for the presence of Tn5Mapusing labeled ColE1::Tn5Map as a probe (data not shown).

    For mapping experiments, agarose plugs of genomic DNAfrom each Tn5Map insertion mutant were first partially di-gested with SpeI and then I-SceI digestion was performed.DNA products of double digestion were separated by PFGE induplicate, on both sides of the Brucella DNA size markers.After Southern blotting of the gel, each half of the nylonmembrane, containing a sample of I-SceI-SpeI digested DNA,was hybridized with one of the two probes containing the endsof Tn5Map. The SpeI restriction sites were directly localizedon each side of the insertion of the I-SceI site, so that thelabeled fragments read from the bottom to the top of themembrane, allowing the construction of the SpeI physical map.

    We show an example of the two hybridization patterns inFig. 2 for the B. ovis clone 63/290v7, with the I-SceI siteinsertion in the 130-kb SpeI fragment on the small chromo-some. The sizes of the DNA fragments labeled with the 1.1-kbprobe (Fig. 2, lane 1) were, from the bottom to the top, 80, 85,105, 250, 300, 480, and 500 kb, and with the 2.7-kb probe (lane2) they were 45, 65, 80, (85), (105), (250), (300), 325 and 440 kb(parentheses indicate fragments with fainter signals). In fact,all the restriction fragments labeled with the 1.1-kb probe werealso found to hybridize with the 2.7-kb probe, although somehad fainter signals. The 2.7-kb probe hybridized weakly withbands labeled by the 1.1-kb probe. We therefore needed toignore the bands which also hybridized with the 1.1-kb probe.Starting from this 130-kb fragment, the order of the SpeI re-striction fragments was 5, 20, 145, 50, 180, and 20 kb on the1.1-kb end probe side and 20, 15, 245, and 115 kb on the 2.7-kbend probe side.

    From these results, when constructing the SpeI restrictionmap of the B. ovis 63/290 small chromosome, the main diffi-culty encountered in aligning the fragments was the existenceof fragments of slightly different sizes. To separate by PFGEthe fragments obtained by partial digestion, large pulses wereneeded. In these conditions, the determination of the sizes ofthe restriction fragments from the hybridization patterns wasnot sufficiently accurate to identify the fragments precisely,particularly in the higher part of the run. To overcome thisproblem, we used Brucella genomic DNA digested with SpeI orPacI instead of the usual markers such as lambda concatemerDNA or S. cerevisiae chromosomes. This gave a ladder rangingfrom 900 to 3 kb (Fig. 2, lanes S and P). Furthermore, whennecessary, we used two different pulse ramps in separate runsfor a better size calculation. Data from cross-hybridizationexperiments between SpeI and PacI restriction fragments andthe use of several mutants with different Tn5Map insertionsites further confirmed the map.

    Using these methods, we constructed physical maps of the

    FIG. 1. PFGE. SpeI digestion of genomic DNA from six Brucella speciesreference strains. Lane 1, B. melitensis 16M; lane 2, B. abortus 544; lane 3, B. ovis63/290; lane 4, B. suis 1330; lane 5, B. canis RM/666; lane 6, B. neotomae 5K33.Size markers in kilobases are at the left of the figure.

    TABLE 1. Brucella strains used in the study

    Species Straina Host Geographicorigin

    B. melitensis 16M (ATCC 23456) Goat United StatesB. abortus 544 (ATCC 23448) Cattle EnglandB. suis 1330 (ATCC 23444) Swine United StatesB. canis RM/666 (ATCC 23365) Dog United StatesB. ovis 63/290 (ATCC 25840) Sheep AfricaB. neotomae 5K33 (ATCC 23459) Desert wood rat

    a Type strains and biovar reference strains are listed by their FAO/OMSdesignations and their American Type Culture Collection numbers.

    VOL. 179, 1997 PHYSICAL MAPPING AND GENOME ORGANIZATION IN BRUCELLA 3245

    on October 21, 2020 by guest

    http://jb.asm.org/

    Dow

    nloaded from

    http://jb.asm.org/

  • small and large chromosomes for the different Brucella species:B. abortus 544, B. suis 1330, B. canis 62/290, and B. ovis 63/290.The B. neotomae 5K33 map was deduced from cross-hybrid-ization of restriction fragments (Fig. 3). The construction ofthe B. melitensis 16M physical map by this method allowed usto test the validity of the new mapping technique using I-SceIand to confirm the existing map (30). The restriction map of B.suis 1330 was also confirmed by hybridization of SpeI frag-ments with linking clones from a B. suis 1330 DNA cosmidlibrary: one to three of 68 clones linked two adjacent SpeIfragments, except for three sites.

    Comparison of the physical maps. Since a physical map hasbeen established for the genus type strain, B. melitensis 16M(biovar 1) (M1), we arbitrarily decided it was the referencemap and we compared the three other maps to it. The sixphysical maps are shown in Fig. 3.

    The most frequent differences are small insertions or dele-tions (indels) ranging from 1 to 34 kb. These events are mostnoticeable in the small chromosome. We cannot, however,exclude the possibility that small indels exist in the large SpeIfragments, which are predominantly found in the large chro-mosome, since size determination is less accurate for the largefragments.

    A 21-kb fragment was found between two SpeI sites in thesmall chromosome of B. ovis which did not hybridize to DNAs

    from the other Brucella species (data not shown). Althoughother strain-specific nucleotide sequences were not found,their existence cannot be ruled out.

    For B. abortus 544, a 640-kb inversion with respect to M1was found in the small chromosome between two SpeI restric-tion sites. The comparison of cross-hybridization results be-tween SpeI and PacI restriction fragments of M1 DNA andthose of these B. abortus strains confirmed this large inversion.However, this inversion was found only in biovars 1, 2, 3, and4, the small chromosome of the other biovars being very sim-ilar to that of M1 (unpublished results).

    The map of B. suis 1330 was again related to that of M1.However, modifications of restriction fragment length and thecreation or loss of restriction sites were more often observed inthis species than in the other. The small chromosome is 50 kblonger than that of M1. B. canis 62/690 had a physical mapidentical to that of B. suis 1330. B. neotomae 5K33 has a smallchromosome that is very similar to that of B. suis 1330, while itslarge chromosome is more similar to that of M1.

    Genomic organization. In order to compare the genomicorganization of the two chromosomes of the four Brucellaspecies, each of the SpeI restriction fragments of B. melitensis16M was taken as a probe and hybridized with those of theother species. At this macrorestriction level, the genetic infor-mation seemed to be in the same order in the four genomesexcept for the B. abortus 544 small chromosome, showing alarge inversion (see below). Only a few genes have been clonedin Brucella, and some of them were located on the maps byhybridization with SpeI restriction fragments (Fig. 3). The lo-calization of rrn loci, omp1 and omp2, BCSP 31, dnaK, groE,and IS6501 copies has already been published for B. melitensis(30) and was confirmed for the other species. The localizationof the other genes was obtained as described in Materials andMethods. All the genes were found in the corresponding SpeIrestriction fragments in the different species. We also identi-fied the SpeI restriction fragments containing at least one copyof the insertion sequence IS6501 described in the Brucellagenome (33). We found 16 of them in B. ovis 63/290 and only5 in the three other strains (Fig. 4). But, by hybridization withEcoRI-digested DNA of these Brucella strains, it was shownthat the number of IS6501 copies was around 30 in the B. ovisgenome and between 5 and 10 in the three others (33). Finally,three rrn loci were found in all the strains, located in thecorresponding SpeI fragments.

    Phylogeny within the Brucella genus. Restriction fragmentlength polymorphism has been used to study the phylogenyof Pseudomonas aeruginosa (13). We employed a relatedmethod to study the phylogeny within the genus Brucella, using16 type strains representing all the species and biovars. Thedendrogram is shown in Fig. 4. This tree is unrooted, sincemacrorestriction patterns are very specific for Brucella anddefinition of a related extragroup to place the root is artificial.The isolates were scattered in several groups. The two B.abortus groups (biovars 1 to 4 and biovars 5 to 9) form a tightlyrelated branch. The three B. melitensis biovars are regroupedto the latter but are more remote from one another. B. ovis ismore or less related to B. melitensis biovar 1 by the methodused for the construction of the tree. The four B. suis biovarsare localized at the opposite extremities of the tree and arerelatively separated from one another. B. canis is not repre-sented in a separate branch, since its physical map is thesame as that of B. suis biovar 1. Finally, B. neotomae is anintermediate between the B. melitensis-B. abortus group andthe B. suis biovar group.

    FIG. 2. Hybridization of I-SceI-linearized and partially SpeI-digested geno-mic DNA of B. ovis clone 63/290v7 with the end probes. The electrophoreticconditions were optimized for the fragments below 450 kb. Lane 1, hybridizationwith the 1.1-kb probe; lane 2, hybridization with the 2.7-kb probe; lanes S and P,SpeI digestion and PacI digestion, respectively, of B. ovis DNA labeled with atotal genomic DNA probe. The sizes of the PacI restriction fragments are givenin kilobases.

    3246 MICHAUX-CHARACHON ET AL. J. BACTERIOL.

    on October 21, 2020 by guest

    http://jb.asm.org/

    Dow

    nloaded from

    http://jb.asm.org/

  • DISCUSSIONThe first striking observation concerning our results is that

    the restriction maps of the six Brucella species are very similar.Genome mapping of several isolates belonging to the samebacterial species or genus has been performed for few bacteria,and comparison showed different patterns of genome conser-vation. First, chromosomal mapping could reveal a high con-servation of the restriction site or gene order (2, 22, 24, 34).The second pattern associated the conservation of gene orderwith a high restriction polymorphism (4, 25, 32). In a thirdpattern, comparison of genome mapping revealed a mosaicstructure showing the presence of conserved and modifiedregions (3, 11, 23). In the most extreme case, the restriction siteor gene order differed markedly (19, 43). The close relation-ship found among the six physical maps in our study allows usto assign the Brucella genus to the first group. Small indels and

    creation or loss of restriction sites are frequently observed,most often in the small chromosome. Variability is localized tocertain regions, similar to the polymorphic hot spots observedin Clostridium perfringens (2).

    A 640-kb inversion was found in the small chromosome of B.abortus 544 (biovar 1). PacI restriction polymorphism showedthat this inversion is also present in biovars 2, 3, and 4 of thesame species but not in biovars 5, 6, and 9. It is unlikely that thephenotypic differences between the two groups of biovars canbe explained by this chromosomal rearrangement. Similar in-versions with a size of several hundred kilobases have also beendescribed in other bacterial species (21, 22, 25–28, 35, 43).

    It has been demonstrated that large genomic rearrange-ments in E. coli and in Salmonella serovars occur by homolo-gous recombination at the rrn loci (17, 27, 28) or repeatednucleotide sequences, such as insertion sequences (7, 12, 21).

    FIG. 3. Physical maps of the two chromosomes of B. melitensis 16M (B. m), B. abortus 544 (B. a), B. suis 1330 (B. s), B. canis RM/666, B. ovis 63/290 (B. o), andB. neotomae (B. n). The two circular chromosomes are represented in linear form, each one starting from a conserved SpeI fragment. (A) large chromosomes; (B) smallchromosomes. For each chromosome map, SpeI sites are located above and PacI sites are located below. Sizes are given in kilobases for all the restriction fragmentsof the B. melitensis chromosomes, and only for those different from B. melitensis for the other species. Several genes are located (see Materials and Methods). rrn lociare represented by open squares, IS6501 copies by stars, and Tn5Map insertions by open circles. The inversion in the small chromosome of B. abortus is shown by twolines.

    VOL. 179, 1997 PHYSICAL MAPPING AND GENOME ORGANIZATION IN BRUCELLA 3247

    on October 21, 2020 by guest

    http://jb.asm.org/

    Dow

    nloaded from

    http://jb.asm.org/

  • An insertion sequence, IS711 (15) or IS6501 (33), was de-scribed in Brucella species. These mechanisms cannot explainthe rearrangement in the small chromosome of B. abortus,since rrn loci or IS copies were not located at the edges of theinversion. It is also noteworthy that the B. ovis genome con-tains three to four times more copies of IS6501 than the otherBrucella species, but its restriction map is very similar to that ofthese species, without evidence of major chromosomal rear-rangements. Other repeated sequences exist in the Brucellagenome which may also have played a role in the genomicrearrangements observed (14).

    The dendrogram deduced from the maps grossly reflects thetraditional phenetic classification of the genus Brucella, con-firming a previous study in which we found that Brucella spe-cies exhibited different XbaI restriction patterns, that biovarswithin a species are less polymorphic, and that, except for B.ovis isolates (33), field strains within a biovar have exactly thesame pattern (1). This result was not expected since this divi-sion of the genus into species and biovars is based solely upona small number of metabolic traits, two lipopolysaccharideepitopes, and sensitivity to a set of phages which are host-rangevariants of the same ancestor phage (6). A recent study, usingsequence data from the omp2 locus from different Brucellaspecies, confirms the existence of species-specific lineages (9).As in our study, a close relationship was found between thesame strains (between B. suis and B. canis or between B.melitensis and B. abortus, the latter being divided into the sametwo subgroups), but an extreme divergence of B. ovis and B.neotomae from the other species was also observed. However,the authors suggested the existence of a gene conversion forthis locus in these two species, explaining the divergence withthe other species.

    The nucleotide sequence similarity between all the Brucellastrains, measured by DNA-DNA hybridization, is so high thatVerger et al. (42), following the international recommenda-tions for bacterial classification, proposed that the genus ismonospecific. However, although virulence is not recognized

    as a taxonomic criterion, the division of Brucella into severalspecies has been always more or less influenced by the restric-tion of the virulence of each so-called species to one or a smallnumber of mammalian hosts. With view to a phylogeneticclassification, such a proposal can be made only if the different“species” are evolutive lineages restricted to a narrow niche.This is the case for Brucella for three reasons. First, because oftheir fastidious growth requirements, Brucella organisms can-not actively multiply in the environment but only in infectedanimals which harbor huge numbers of bacteria in their tissues.Second, since mechanisms of genetic exchange, such as plas-mid, temperate bacteriophage, or transformation, have neverbeen demonstrated to occur naturally in Brucella, each speciesis genetically isolated, despite the frequent coexistence of dif-ferent strains infecting different animals in the same farm.Third, virulence of the different species is restricted to a smallnumber of hosts. For example, B. melitensis infects goats andsheep while B. abortus infects cattle. While both species arealso pathogenic for humans and B. melitensis occasionally in-fects cattle, the “abnormal” host is always contaminated via thenatural host and the infection is an epidemiological dead end.A narrower restriction of the virulence is observed for B. canis,B. ovis, and B. neotomae, which are, respectively, pathogenicfor dogs, sheep, and desert wood rats. Finally, B. suis is essen-tially virulent for swine. The similarity of the maps argues fora monospecific genus, as suggested by Verger et al., and thetree for the existence of subspecies corresponding to evolutivelineages (pathovars) adapted to a particular host.

    Liu and Sanderson have recently compared the physical andgenetic maps of several different Salmonella serovars (27). Themaps of the different serovars of Salmonella enterica (e.g., S.typhimurium and S. enteritidis) are closely related. S. entericainfections in humans range from gastroenteritis to fatal septi-cemia, but the organism also infects a wide range of warm- andcold-blooded animals. The genomic organization of the ubiq-uitous E. coli was also similar (but not identical) to that of S.enterica. However, the physical map of S. typhi, the agent of

    FIG. 4. Phylogenetic tree of the six Brucella species.

    3248 MICHAUX-CHARACHON ET AL. J. BACTERIOL.

    on October 21, 2020 by guest

    http://jb.asm.org/

    Dow

    nloaded from

    http://jb.asm.org/

  • typhoid fever in humans, is very different in both its organiza-tion and content. Moreover, when four wild-type S. typhi iso-lates were tested, all were found to have identical physicalmaps. Since S. typhi is pathogenic only for humans and has noother animal hosts, the authors proposed that the restrictedhost range has exerted specific selective constraints, conduciveto the selection of a particular clone.

    The very close homology between the different species,shown by the comparison of macrorestriction maps, requires astill more precise analysis of the Brucella genome in order tounderstand the differences in the virulence restriction of thedifferent species. High-resolution maps are now being con-structed in our laboratory.

    REFERENCES

    1. Allardet-Servent, A., G. Bourg, M. Ramuz, M. Pages, M. Bellis, and G.Roizes. 1988. DNA polymorphism in strains of the genus Brucella. J. Bacte-riol. 170:4603–4607.

    2. Canard, B., B. Saint-Joanis, and S. T. Cole. 1992. Genomic diversity andorganization of virulence genes in the pathogenic anaerobe Clostridium per-fringens. Mol. Microbiol. 6:1421–1429.

    3. Carlson, C. R., D. A. Caugant, and A. B. Kolstø. 1994. Genotypic diversityamong Bacillus cereus and Bacillus thuringiensis strains. Appl. Environ. Mi-crobiol. 60:1719–1725.

    4. Casjens, S., M. Delange, H. L. Ley III, P. Rosa, and W. M. Huang. 1995.Linear chromosomes of Lyme disease agent spirochetes: genetic diversityand conservation of gene order. J. Bacteriol. 177:2769–2780.

    5. Colleaux, L., C. Rougeulle, P. Avner, and B. Dujon. 1993. Rapid physicalmapping of YAC inserts by random integration of I-SceI sites. Hum. Mol.Genet. 2:265–271.

    6. Corbel, M. J., and W. J. Brinley-Morgan. 1984. Genus Brucella, p. 377–388.In N. R. Krieg and J. G. Holt (ed.), Bergey’s manual of systematic bacteri-ology, vol. 1. The Williams & Wilkins Co., Baltimore, Md.

    7. Daniels, D. L. 1990. The complete AvrII restriction map of the Escherichiacoli genome and comparisons of several laboratory strains. Nucleic AcidsRes. 18:2649–2651.

    8. Denoel, P., M. S. Zygmunt, V. Weynants, A. Tibor, B. Lichtfouse, P. Britteuil,J. N. Limet, and J. J. Letesson. 1995. Cloning and sequencing of the bacte-rioferritin gene of Brucella melitensis 16M strain. FEMS Microbiol. Lett.361:238–242.

    9. Ficht, T. A., H. S. Husseinen, J. Derr, and S. W. Bearden. 1996. Species-specific sequences at the omp2 locus of Brucella type strains. Int. J. Syst.Bacteriol. 46:329–331.

    10. Ficht, W. M., and E. Margoliash. 1967. Construction of phylogenetic trees.Science 155:279–284.

    11. Fonstein, M., T. Nikolskaya, and R. Haselkorn. 1995. High-resolution align-ment of a 1-megabase-long genome region of three strains of Rhodobactercapsulatus. J. Bacteriol. 177:2368–2372.

    12. Galas, I., and M. Chandler. 1989. Bacterial insertion elements, p. 109–162.In D. E. Berg and M. M. Howe (ed.), Mobile DNA. American Society forMicrobiology, Washington, D.C.

    13. Grothues, D., and B. Tümmler. 1991. New approaches in genome analysis bypulsed-field gel electrophoresis: application to the analysis of Pseudomonasspecies. Mol. Microbiol. 5:2763–2776.

    14. Halling, S. M., and B. J. Bricker. 1994. Characterization and occurrence oftwo repeated palindromic DNA elements of Brucella spp.: Bru-RS1 andBru-RS2. Mol. Microbiol. 14:681–689.

    15. Halling, S. M., F. M. Tatum, and B. J. Bricker. 1993. Sequence and char-acterization of an insertion sequence, IS711, from Brucella ovis. Gene 133:123–127.

    16. Hemmen, F., V. Weynants, T. Scarcez, J.-J. Letesson, and E. Saman. 1995.Cloning and sequence analysis of a newly identified Brucella abortus geneand serological evaluation of the 17-kilodalton antigen that it encodes. Clin.Diagn. Lab. Immunol. 2:263–267.

    17. Hill, C. W., and B. W. Harnish. 1981. Inversions between ribosomal RNAgenes of Escherichia coli. Proc. Natl. Acad. Sci. USA 78:7069–7072.

    18. Hiller, B., B. Frey, and W. Schumann. 1994. Tn5Map, a transposon for therapid mapping of restriction sites in plasmids. FEMS Microbiol. Lett. 115:151–156.

    19. Jiang, Q., K. Hiratsuka, and D. E. Taylor. 1996. Variability of gene order indifferent Helicobacter pylori strains contributes to genome diversity. Mol.Microbiol. 20:833–842.

    20. Jumas-Bilak, E., C. Maugard, S. Michaux-Charachon, A. Allardet-Servent,A. Perrin, D. O’Callaghan, and M. Ramuz. 1995. Insertion of a unique

    restriction site in the bacterial chromosomes for study of genome organiza-tion. Microbiology 141:2425–2432.

    21. Komoda, Y., M. Enomoto, and A. Tominaga. 1991. Large inversion in Esch-erichia coli K12 1485IN between inversely oriented IS3 elements near lac andcdd. Genetics 129:639–645.

    22. Ladefoged, S. A., and G. Christiansen. 1992. Physical and genetic mapping ofthe genomes of five Mycoplasma hominis strains by pulsed-field gel electro-phoresis. J. Bacteriol. 174:2199–2207.

    23. Leblond, P., M. Redenbach, and J. Cullum. 1993. Physical map of theStreptomyces lividans 66 genome and comparison with that of the relatedstrain Streptomyces coelicolor A3(2). J. Bacteriol. 175:3422–3429.

    24. Le Bourgeois, P., M. Lautier, M. Mata, and P. Ritzenthaler. 1992. Physicaland genetic map of the chromosome of Lactococcus lactis subsp. lactisIL1403. J. Bacteriol. 174:6752–6762.

    25. Le Bourgeois, P., M. Lautier, L. van den Berghe, M. J. Gasson, and P.Ritzenthaler. 1995. Physical and genetic map of the Lactococcus lactis subsp.cremoris MG1363 chromosome: comparison with that of Lactococcus lactissubsp. lactis IL1403 reveals a large genome inversion. J. Bacteriol. 177:2840–2850.

    26. Liu, S. L., A. Hessel, and K. E. Sanderson. 1993. The XbaI-BlnI-CeuI geno-mic cleavage map of Salmonella enteritidis shows an inversion relative toSalmonella typhimurium LT2. Mol. Microbiol. 10:655–664.

    27. Liu, S.-L., and K. E. Sanderson. 1995. Rearrangements in the genome of thebacterium Salmonella typhi. Proc. Natl. Acad. Sci. USA 92:1018–1022.

    28. Liu, S.-L., and K. E. Sanderson. 1995. The chromosome of Salmonellaparatyphi A is inverted by recombination between rrnH and rrnG. J. Bacte-riol. 177:6585–6592.

    29. Margolin, W., D. Bramhill, and S. R. Long. 1995. The dnaA gene of Rhizo-bium meliloti lies within an unusual gene arrangement. J. Bacteriol. 177:2892–2900.

    30. Michaux, S., J. Paillisson, M. J. Carles-Nurit, G. Bourg, A. Allardet-Servent,and M. Ramuz. 1993. Presence of two independent chromosomes in theBrucella melitensis 16M genome. J. Bacteriol. 175:701–705.

    31. Monteilhet, C., A. Perrin, A. Thierry, L. Colleaux, and B. Dujon. 1990.Purification and characterization of the in vitro activity of I-SceI, a novel andhighly specific endonuclease encoded by a group I intron. Nucleic Acids Res.18:1407–1413.

    32. Ojaimi, C., B. E. Davidson, I. Saint Girons, and I. G. Old. 1994. Conserva-tion of gene arrangement and an unusual organization of rRNA genes in thelinear chromosomes of the Lyme disease spirochaetes Borrelia burgdorferi, B.garinii, and B. afzelii. Microbiology 140:2931–2940.

    33. Ouahrani, S., S. Michaux, J. Sri Widada, G. Bourg, R. Tournebize, M.Ramuz, and J. P. Liautard. 1993. Identification and sequence analysis ofIS6501, an insertion sequence in Brucella spp: relationship between genomicstructure and the number of IS6501 copies. J. Gen. Microbiol. 139:3265–3273.

    34. Perkins, J. D., J. D. Heath, B. R. Sharma, and G. M. Weinstock. 1993. XbaIand BlnI genomic cleavage maps of Escherichia coli K-12 strain MG1655 andcomparative analysis of other strains. J. Mol. Biol. 232:419–445.

    35. Riley, M., and K. E. Sanderson. 1990. Comparative genetics of Escherichiacoli and Salmonella typhimurium, p. 85–95. In K. Drlica and M. Riley (ed.),The bacterial chromosome. American Society for Microbiology, Washing-ton, D.C.

    36. Sangari, F. J., J. M. Garcia-Lobo, and J. Aguerro. 1994. The Brucella abortusvaccine strain B19 carries a deletion in the erythritol catabolic genes. FEMSMicrobiol. Lett. 121:337–342.

    37. Smith, H. O., and M. L. Birnstiel. 1976. A simple method for DNA restric-tion mapping. Nucleic Acids Res. 3:2387–2398.

    38. Tatum, F. M., N. F. Cheville, and D. Morfitt. 1994. Cloning, characterizationand construction of htrA and htrA-like mutants of Brucella abortus and theirsurvival in BALB/c mice. Microb. Pathog. 17:23–36.

    39. Tatum, F. M., D. C. Morfitt, and S. Halling. 1993. Construction of a Brucellaabortus RecA mutant and its survival in mice. Microb. Pathog. 14:177–185.

    40. Thierry, A., and B. Dujon. 1992. Nested chromosomal fragmentation in yeastusing the meganuclease I-SceI: a new method for physical mapping of eu-caryotic genomes. Nucleic Acids Res. 20:5625–5631.

    41. Tibor, A., V. Weynants, P. Denoel, B. Lichtfouse, X. De Bolle, E. Saman, J. N.Limet, and J. J. Letesson. 1994. Molecular cloning, nucleotide sequence, andoccurrence of a 16.5-kilodalton outer membrane protein of Brucella abortuswith similarity to PAL lipoprotein. Infect. Immun. 62:3633–3639.

    42. Verger, J. M., F. Grimont, P. A. D. Grimont, and M. Crayon. 1985. Brucella,a monospecific genus as shown by deoxyribonucleic acid hybridization. Int. J.Syst. Bacteriol. 35:292–295.

    43. Zuerner, R. L., J. L. Herrmann, and I. Saint Girons. 1993. Comparison ofgenetic maps for two Leptospira interrogans serovars provides evidence fortwo chromosomes and intraspecies heterogeneity. J. Bacteriol. 175:5445–5451.

    VOL. 179, 1997 PHYSICAL MAPPING AND GENOME ORGANIZATION IN BRUCELLA 3249

    on October 21, 2020 by guest

    http://jb.asm.org/

    Dow

    nloaded from

    http://jb.asm.org/