* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
Download Homologous Recombination in Agrobacterium: Potential
Hybrid (biology) wikipedia , lookup
Artificial gene synthesis wikipedia , lookup
Genome evolution wikipedia , lookup
Metagenomics wikipedia , lookup
Cre-Lox recombination wikipedia , lookup
Site-specific recombinase technology wikipedia , lookup
DNA barcoding wikipedia , lookup
Pathogenomics wikipedia , lookup
Homologous Recombination in Agrobacterium: Potential Implications for the Genomic Species Concept in Bacteria Denis Costechareyre, Franck Bertolla, and Xavier Nesme Université de Lyon, Université Lyon 1, CNRS, and INRA, Ecologie Microbienne UMR 5557, USC 1193, 16 Rue Raphaël Dubois, Domaine Scientifique de La Doua, Villeurbanne, France According to current taxonomical rules, a bona fide bacterial species is a genomic species characterized by the genomic similarity of its members. It has been proposed that the genomic cohesion of such clusters may be related to sexual isolation, which limits gene flow between too divergent bacteria. Homologous recombination is one of the most studied mechanisms responsible for this genetic isolation. Previous studies on several bacterial models showed that recombination frequencies decreased exponentially with increasing DNA sequence divergence. In the present study, we investigated this relationship in the Agrobacterium tumefaciens species complex, which allowed us to focus on sequence divergence in the vicinity of the genetic boundaries of genomic species. We observed that the sensitivity of the recombination frequency to DNA divergence fitted a log-linear function until approximately 10% sequence divergence. The results clearly revealed that there was no sharp drop in recombination frequencies at the point where the sequence divergence distribution showed a ‘‘gap’’ delineating genomic species. The ratio of the recombination frequency in homogamic conditions relative to this frequency in heterogamic conditions, that is, sexual isolation, was found to decrease from 8 between the most distant strains within a species to 9 between the most closely related species, for respective increases from 4.3% to 6.4% mismatches in the marker gene chvA. This means that there was only a 1.13-fold decrease in recombination frequencies for recombination events at both edges of the species border. Hence, from the findings of this investigation, we conclude that—at least in this taxon—sexual isolation based on homologous recombination is likely not high enough to strongly hamper gene flow between species as compared with gene flow between distantly related members of the same species. The 70% relative binding ratio cutoff used to define bacterial species is likely correlated to only minor declines in homologous recombination frequencies. Consequently, the sequence diversity, as a mechanistic factor for the efficiency of recombination (as assayed in the laboratory), appears to play little role in the genetic cohesion of bacterial species, and thus, the genomic species definition for prokaryotes is definitively not reconcilable with the biological species concept for eukaryotes. Introduction Even though there is no generally accepted definition for a bacterial species, according to current taxonomic rules, it is defined by the genomic similarity of its members (Stackebrandt et al. 2002). Presently, to have an accepted description of a new species in journals officially in charge of such publications (e.g., Int J Syst Microbiol Evol), one must provide results of DNA–DNA hybridization studies (or accepted alternative methods) to show that species members globally display a high degree of genome similarity and are clearly different from other known species. According to this strict and formal definition for a bacterial species, the questions are what are ‘‘genomic species’’ and are bacterial species defined in this way comparable to some extent to eukaryote species. A basic and primary difference between prokaryote and eukaryote species is that prokaryotic species must have accumulated enough genetic variations through time to be distinguishable by DNA–DNA hybridization. This implies that, with genomic species, the speciation events took place long ago in the past, although nascent species could not be identified. In addition, contrary to the biological species concept (BSC) for eukaryotes, the purely operational definition for bacterial species is not supported by any formal biological concept. The BSC, as developed by Mayr (1942), described that ‘‘species are groups of actually or potentially interbreeding natural populations, which are reKey words: Agrobacterium, bacterial species, homologous recombination, sexual isolation. E-mail: [email protected]. Mol. Biol. Evol. 26(1):167–176. 2009 doi:10.1093/molbev/msn236 Advance Access publication October 20, 2008 Ó The Author 2008. Published by Oxford University Press on behalf of the Society for Molecular Biology and Evolution. All rights reserved. For permissions, please e-mail: [email protected] productively isolated from other such groups.’’ Considering genetic exchanges as sexuality, this phenomenon is much less restricted in bacteria than in eukaryotes, because genes can be transferred between species. The sexual isolation concept inherent to the BSC is thus likely not applicable to prokaryotes. However, clustering bacteria in discrete groups of strains on the basis of their genome similarities could be related to sexuality as well. The standard whole-genome DNA–DNA hybridization technique was the first genomic method accepted for delineating species (Wayne et al. 1987): ‘‘The phylogenetic definition of a species generally would include strains with approximately 70% or greater DNA-DNA relatedness and with 5°C or less DTm. Both values must be considered.’’ With the development of new genetic technology, methods used to define bacterial species are regularly reevaluated by ad hoc bacterial systematics committees (Stackebrandt et al. 2002; Gevers et al. 2005). Among the methods accepted for species delineation, amplified fragment length polymorphism (AFLP) clearly highlights that bacterial species are genomically similar and well-separated clades of strains (Mougel et al. 2002; Portier et al. 2006). More understandable than the 70% boundary value of DNA–DNA relatedness or relative binding ratio (RBR), AFLP revealed that closely related species could be characterized by maximal intraspecies current genome mispairing or mismatching (CGM) of 11% and a minimal interspecies CGM of 15% (Portier et al. 2006). Consequently, Portier et al. (2006) emphasized that there were gaps in the distribution of genome divergence values, even between closely related species. These gaps currently serve as the basis for the genomic delineation of bacterial species. Similarly, a massive comparative analysis based on genome sequence comparisons of 70 closely related bacteria explored 168 Costechareyre et al. the genetic boundaries of species (Goris et al. 2007). Goris et al. (2007) also revealed that the 70% RBR value corresponds to 5% in average nucleotide divergence values for shared genes. This genetic cohesion of species was also noted in several multilocus sequence studies (Naser et al. 2005; Thompson et al. 2005; Kuhnert and Korczak 2006; Richter et al. 2006; Martens et al. 2008). According to this notion of gaps between species, the pattern of gene diversity within and between taxonomic units was recently explained by Martens et al. (2008). These authors calculated, for each gene and Ensifer species, the amount of ‘‘heterogeneity’’ (highest sequence divergence within a species) and the amount of ‘‘separability’’ (smallest amount of sequence divergence observed between a particular species and other species), although the species displaying the smallest amount of sequence divergence from the particular species is referred to as the closest neighbor taxon. Indeed, biological mechanisms involved in species delineation could be explored at the best with particular species displaying large amounts of heterogeneity and small amounts of separability with their closest neighbors. Given these observations, the genetic cohesion of bacterial species is evidently due to a set of ‘‘cohesive forces.’’ It is quite likely that genetic species boundaries may result from either the existence of ecotypes, corresponding to ecologically isolated populations, and/or sexual isolation caused by DNA sequence divergence. Reinforcing this hypothesis, Majewski (2001) reported substantial sexual isolation (i.e., decrease in recombination frequency) of three orders of magnitude between bacteria showing approximately 10–15% sequence divergence at a marker gene. Accordingly, the 11–15% gap in genome mismatch found by Portier et al. (2006) suggests that this is the innate parameter controlling the physiological mechanism of homologous recombination, leading in turn to species formation by way of efficient genetic isolation. However, until now, sexual isolation reports were only based on genetic divergence without any correlation with whole-genome divergence. Indeed, several studies have explored the effect of genomic sequence divergence on the rate of homologous recombination success in different bacterial models. An earlier study by Matic et al. (1995) involving an Escherichia coli/Salmonella typhimurium model supported the idea that ‘‘the genetic barrier that separates these (and perhaps many other) related species is primarily recombinational.’’ Later studies that explored natural transformation in Bacillus (Roberts and Cohan 1993; Zawadzki et al.1995) or Streptococcus pneumoniae (Humbert et al. 1995; Majewski et al. 2000), as well as Hfr conjugation in E. coli/Salmonella (Vulić et al. 1997), showed that sexual isolation depends, according to a simple mathematical function, on the level of DNA sequence divergence in a log-linear mode. This global relationship was observed at a large scale, but in spite of the hypothesis that ‘‘gaps’’ in genomic divergence could also coincide with gaps in sexual isolation, the effects of nucleotide divergence on sexual isolation at species borders have not been thoroughly examined. Consequently, we address the question as to whether sexual isolation may or may not play a direct role in species cohesion, because this isolation may be measured at the genetic boundaries of genomic species by a significant drop in recombination frequencies. If this is the case, a conceptual link between the BSC and genomic species definition could be feasible. To investigate thishypothesis, we comparedvariations in recombination frequencies with nucleotide sequence divergence in a complex of well-defined genomic species (Popoff et al. 1984; Mougel et al. 2002; Portier et al. 2006) known as Agrobacterium biovar 1, but usually and irrelevantly called ‘‘Agrobacterium tumefaciens.’’ This taxon allowed us to extensively explore sexual isolation at the genetic boundaries of genomic species. Indeed, genomic studies of this taxon highlighted a broad range of intraspecies divergence within very closely related but distinct species. In the present study, a ubiquitous gene, that is, the chromosomal virulence gene chvA, was sequenced to check whether the clear delineation ofall known A. tumefaciens genomic species occursatthe gene scale. Subsequently, the chvA gene, whose sequence divergence was measured, was chosen as the target DNA sequence for recombination. Thus, we characterized the relationship between recombination frequencies and DNA sequence divergence for very closely related species. Finally, we link the sensitivity of recombination to DNA sequence divergence to the occurrence of divergence gaps between species and then discuss the evolutionary implications of genetic recombination on bacterial speciation. Materials and Methods Strains and Plasmids All strains of the A. tumefaciens species complex used in this study are listed in table 1. Escherichia coli strains and plasmids are listed as supplemental data (table 3). Agrobacterium and Sinorhizobium meliloti strains were grown at 28 °C on YPG medium (yeast extract, 5 g/L; Bacto peptone, 5 g/L; glucose, 10 g/L). Escherichia coli strains were grown at 37 °C on Luria Bertani (LB) medium. For recombination experiments, recipient and recombinant strains of A. tumefaciens were grown on a selective AB mannitol medium (ABM) consisted of AB medium salts of Clark and Maaløe (1967) supplemented with 1 g/l mannitol. Media were complemented with appropriate antibiotics: 50 lg/ml rifampicin (Rif50), 25 lg/ml kanamycin (Km25), 25 lg/ml neomycin (Nm25), 50 lg/ml streptomycin (Sm50), and 100 lg/ml ampicillin (Ap100). Generation of Rifampicin-Resistant Mutants Rifampicin-resistant A. tumefaciens strains (C58, TT9, and M2/1) were obtained by spreading 100 ll of overnight cultures on YPG plates supplemented with 50 lg/ml of rifampicin. After confirmation of stable rifampicin resistance (RifR), one clone per strain was arbitrarily selected and stored at 80 °C. DNA Extraction and Purification Genomic and plasmid DNA were, respectively, extracted with the DNeasy Tissue Kit (Qiagen, Hilden, Germany) and the QIAprep Spin Miniprep Kit (Qiagen) according to the manufacturer’s instructions. Polymerase chain reaction (PCR) products were purified with the GFX PCR DNA and Gel Band Purification Kit (GE Healthcare, Little Chalfont, United Kingdom) as recommended by the manufacturer. Sexual Isolation of Bacterial Species 169 Table 1 Agrobacterium Strains Used in This Study and Corresponding chvA GenBank Accession Numbers Reference GenBank chvA Accession Number Crown gall, United States Dahlia sp., Germany Plant Solanum nigrum, root tissue commensal, LCSA, France Hybrid poplar, Leuce section, gall, France Bulk soil, LCSA, France Prunus persica x P. amygdalus gall, cv. GF677, France De Ley et al. (1973) De Ley et al. (1973) Popoff et al. (1984) EU007383 EU007384 EU007386 Portier et al. (2006) Portier et al. (2006) EU007388 Portier et al. (2006) Portier et al. (2006) EU007391 EU007389 Genomic species G2 M 2/1 CFBP 5876, LMG 147 CIP 497-74 CFBP 5494, CFBP 2884 CIP 28-75 CFBP 5495 CIP 43-76 CFBP 5496 R M 2/1 Rif Ditch water, Belgium Human blood, France Human urine, France Human urine, France Rifampicin-resistant mutant Popoff et al. Popoff et al. Popoff et al. Popoff et al. This study (1984) (1984) (1984) (1984) EU007397 EU007398 EU007399 EU007400 Genomic species G3 CIP 107443 CFBP 6623, CIP 493-74 CIP 107442 CFBP 6624, CIP 111-78 Antiseptic flask, France Human, France Popoff et al. (1984) Popoff et al. (1984) EU007401 EU007402 Apple seedling, Iowa, United States/ A. tumefaciens type strain Soil around Prunus dulcis, South Australia Soil/A. radiobacter type strain Popoff et al. (1984) EU007403 De Ley et al. (1973) EU007406 Popoff et al. (1984) EU007408 Patient food, France Human cephalo-rachidian liquid, France Popoff et al. (1984) Popoff et al. (1984) EU007415 EU007416 Genomic species G6 NCPPB 925 CFBP 5499, LMG 225 Dahlia sp., South Africa Popoff et al. (1984) EU007417 Genomic species G7 RV3 CFBP 5500, LMG 317 NCPPB 1641 CFBP 5502, LMG 228 Zutra 3/1 CFBP 6999, LMG 298 No information available Flacourtia indica, United Kingdom Malus sp., Israel Popoff et al. (1984) Popoff et al. (1984) Popoff et al. (1984) EU007419 EU007420 EU007421 Likely hop, United States Humulus lupulus, gall, Victoria, Australia Euonymus alata cv. Compacta, gall, United States Rubus macropetalus, OR, United States Prunus sp. cv. Montmorency (cherry), New York, United States Rifampicin-resistant mutant Rifampicin-resistant mutant Popoff et al. (1984) Popoff et al. (1984) EU007423 EU007426 Portier et al. (2006) EU007427 Portier et al. (2006) EU007428 De Ley (1974) EU007429 This study This study John Innes potting soil, Australia John Innes potting soil, Australia Popoff et al. (1984) Popoff et al. (1984) EU007430 EU007431 Rhizospheric soil from Prunus persicae Portier et al. (2006) EU007432 S. meliloti 2011 parent strain; SmR Meade et al. (1982) AL591793 Strain Name Other Namesa Genomic species G1 TT111 CFBP 5767, LMG 196 NCPPB 396 CFBP 5765, LMG 176, Bo542 S56 CFBP 5491, LMG 321 CFBP 5622 CFBP 2517 CFBP 5771 CFBP 2712 Genomic species G4 (Agrobacterium tumefaciens) B6T CFBP 2413T, LMG 187T Kerr 14 CFBP 5761, LMG 15 ATCC 19358 CFBP 5522, LMG 140 Genomic species G5 CIP 107445 CFBP 6625, CIP 291-77 CIP 107444 CFBP 6626, CIP 120-78 Genomic species G8 TT9 CFBP 5504, LMG 64 Mushin 6 CFBP 6550, LMG 201 LMG 75 CFBP 6549 LMG 46 CFBP 6554 C58 CFBP 1903, LMG 287 C58 RifR TT9 RifR Genomic species G9 0362 CFBP 5507, LMG 26 0363 CFBP 5506, LMG 27 Genomic species G13 CFBP 6927 Other strain (Sinorhizobium meliloti) Sm 1021 a Biological Source, Geographical Origin, and Other Informationb CFBP, Collection Francxaise de Bactéries Phytopathogènes, INRA, Angers, France; LMG, Laboratorium voor Microbiologie, Ghent University, Ghent, Belgium; CIP, Collection de l’Institut Pasteur; ATCC, American Type Culture Collection, Manassas, VA. b SmR, Streptomycin resistance. 170 Costechareyre et al. PCR Amplification and Sequencing Oligonucleotides used as PCR primers were designed according to the genome sequence of strain C58 (http:// depts.washington.edu/agro/). The 1,269-bp sequence used for the phylogenetic study was obtained, depending on the strains, with either primer pair F2044 (ATT AAC CCT CAC TAA AGG GAT TCG GCC GWA TCA TYG ACG C) and F2047 (TAA TAC GAC TCA CTA TAG GGC GAT GAT GAA GGT CGT CC) or primer pair F2880 (CCG TTG TCG TCA TCG CCA) and F2883 (AAC AGG ACG AGA TCG GCA TC), amplifying 1,457 or 1,558 bp, respectively. The 1,401-bp fragments used to clone chvA alleles in pK18mob were obtained with primers F2650 (GGC GGG YTT CGG CGT CT) and F2651 (CGG TGR GCG ATG ATG AAG GT) designed to amplify chvA in Agrobacterium strains as well as the homologous ndvA in S. meliloti. PCRs were performed using Taq polymerase (Invitrogen, Carlsbad, CA) with initial denaturation at 94 °C for 4 min, followed by 34 cycles with denaturation at 94 °C for 1 min, annealing at a strainadjusted temperature (54–62 °C) for 1 min, and extension at 72 °C for 1 min 20 s. Purified PCR products were sequenced (Genoscreen, Lille, France) with primers already used for amplification, except for F2044 and F2047, which have T3 and T7 tails at the 5#-terminus, respectively. Construction of Plasmids for Homologous Recombination chvA PCR products were cloned into the pCR2.1TOPO cloning vector, and then, plasmids were used to transform E. coli DH5a competent cells, according to the manufacturer’s instructions (Invitrogen, Paisley, United Kingdom). Appropriate restriction endonucleases allowed us to select plasmids that contained inserts in the same direction, and then, the sequences of all cloned PCR products were verified using standard M13 forward and M13 reverse primers. Fragments carrying chvA alleles were released with BamHI and XbaI and cloned into a pK18mob vector previously digested with the same enzymes. This generated 14 different pK18-chvA mobilizable vectors that were used to transform competent E. coli S17-1 cells (Simon et al. 1983) using the standard protocol (Sambrook and Russell 2001). Because the tra gene required for pK18mob transfer was present in the chromosome of S17-1, transformations were performed by biparental matings. Transformant cells were selected on LB Km25 Nm25 Ap100 medium. Sequence Analysis Sequence data were manually edited and automatically assembled using Lasergene’s DNAStar SeqMan program (version 7.1). Multiple alignments and determination of the percentage nucleic acid identity were performed using Lasergene’s DNAStar Megalign ClustalW alignment (version 7.1). Phylogenetic trees were constructed using ClustalX software (ftp://ftp-igbmc.u-strasbg.fr/pub/ClustalX/) to obtain a bootstrap neighbor-joining tree (1,000 bootstrap data replicates). DnaSP software (version 4.1) (Rozas and Rozas 1999) was used to determine the G þ C content. The BlastN program (Altschul et al. 1990) on the NCBI Web site was used to compare sequences of plasmid pK18mob and chvA alleles with the Agrobacterium strain C58 genome sequence. Measurement of Recombination Frequencies by the Bacterial Mating System The pK18mob plasmid self-replicates in E. coli, but its maintenance in Agrobacterium is only possible by integrative recombination (Simon et al. 1983; Schäfer et al. 1994). Homologous recombination frequencies, depending only on chvA sequence divergence, were determined using a biparental conjugation system involving E. coli S17-1 (pK18chvA) donor cells and Agrobacterium RifR recipient cells. A single Agrobacterium RifR colony was inoculated into 7 ml of YPG Rif50 and grown at 28 °C until the exponential growth phase at OD600 5 0.8 (approximately 18 h). For E. coli S17-1 (pK18-chvA), a single colony was inoculated into 2.5 ml of LB Km25 Nm25 for overnight growth at 37 °C. Fifty microliters of the preculture was then used to inoculate 4 ml of LB Km25 Nm25 before incubation at 37 °C until OD600 5 0.8 (approximately 5 h). A 1:10 donor:recipient ratio was used for bacterial mating, which corresponds to 108 E. coli S17-1 (pK18-chvA) mixed with 109 Agrobacterium RifR cells. Appropriate volumes of donor and recipient cell cultures were separately washed twice and resuspended in 50 ll of a cold salt solution (NaCl 0.9 g/L). Donor and recipient cells were mixed and incubated at room temperature for 5 min before spotting totally 100 ll on a 0.2-lm Isopore membrane-filter (Millipore) placed on an LB plate. Conjugation mixtures were incubated for 16 h at 28 °C, and then, cells were resuspended in 4 ml of MgSO4 10 mM. Appropriate dilutions in MgSO4 10 mM were plated on selective medium. Escherichia coli donor cells were counted on selective LB Km25 Nm25 medium and incubated for 24 h at 37 °C. Agrobacterium recipient cells were counted on selective ABM Rif50 medium after approximately 72-h incubation at 28 °C. Agrobacterium recombinants were counted on selective ABM Rif50 Km25 Nm25 medium while counterselecting donor and nonrecombinant recipient cells, after approximately 72 h at 28 °C. The recombination frequency was calculated as the number of recombinant cells divided by the number of donor cells. The sexual isolation ratio was determined by dividing the homogamic (i.e., perfect match) recombination frequency by the heterogamic frequency for each trial and each donor recipient. At least four independent crosses were done per mating pair and then averaged. As a control, conjugations between E. coli S17-1 harboring an empty pK18mob and Agrobacterium RifR strains were performed independently in quadruplicate, and no Agrobacterium recombinant cells were obtained. Molecular Screening of Recombinants by PCR and Sequencing The presence of pK18-chvA inserts in the recipient chvA gene was verified by PCR and sequencing for at least one clone of each kind of recombinant. To check for the presence of chvA tandem sequences resulting from Sexual Isolation of Bacterial Species 171 FIG. 1.—Neighbor-joining tree based on chvA sequences (1,299 bp) of Agrobacterium strains. The homologous ndvA gene from Sinorhizobium meliloti 1021 was used as the out-group. The three strains used as recipient cells are indicated by arrows. Strains corresponding to the chvA donor are underlined. Bootstrap % values (70%) obtained from 1,000 data resamplings are shown. integrative recombination, primers F2044 and F5039 (TAA CCG TAT TAC CGC CTT TGA G) were used to amplify the 1,990-bp ‘‘upstream hybrid fragment’’ and primers F2785 (TAA GTT GGG TAA CGC CAG GGT T) and F2787 (GGA ACA GGA CGA GAT CGG CAT) were used to amplify the 1,578-bp ‘‘downstream hybrid fragment’’ (supplemental data, fig. 5). PCR products of hybrid fragments were sequenced as described above. Statistical Analysis Linear regression and covariance analyses were performed using RÓ software (http://www.r-project.org/). Nucleotide Sequence Accession Numbers The sequences have been deposited in GenBank, and the accession numbers are listed in table 1. Results and Discussion Sequence Analysis of chvA The present work was undertaken to evaluate the degree of sexual isolation at genomic species boundaries. For this purpose, chvA was chosen because it is both: 1) not essential for growth and 2) essential for ecological functions in relation to plant interactions (Breedveld and Miller 1994). This gene can therefore be both disrupted for use in recombination studies and found in all agrobacteria for nucleotidic sequence analysis. The nucleotide sequence of chvA was determined for a large set of strains belonging to all known species of the A. tumefaciens complex (table 1). chvA was found to be useful for the delineation of all genomic species (fig. 1) and to have a G þ C content ranging from 59.1% to 61.6%, which is close to the 59.4% G þ C content of the C58 circular chromosome (data not shown). In addition, chvA appeared to be highly correlated with current genome divergence values (CGM), with a broad range of genetic divergences between pairs of strains evenly distributed close to the within-to-between species boundaries (fig. 2). However, although the CGM cutoff for A. tumefaciens species was found to be 15% (Portier et al. 2006), it was found to be 4.5% for chvA in the present study, that is, a value remarkably close to the 5% average nucleotide mismatch cutoff defined by Goris et al. (2007) to delineate bacterial species. Finally, the chvA phylogeny is congruent to those obtained with a variety of other genes (recA, mutS, gyrB, glgC, and gltD; data not shown), indicating that chvA is not particularly prone to recombination between or within species. chvA is thus a relevant alternative for usual whole-genome comparisons carried out for species assignation, and it could also beused to testfinemodificationsin thesensitivity of homologous recombination to gene divergence at boundaries of bacterial species. Sensitivity of Recombination to chvA Sequence Divergence The recombination frequency is indeed variable from one gene to another, ranging for instance from two to more 172 Costechareyre et al. ilar recombination frequencies with the two G8 recipient strains C58 and TT9 and a lower level with the G2 recipient strain M2/1 (supplemental data, fig. 6). This individual strain effect was eliminated after data transformation into sexual isolation values (table 2). As described by Roberts and Cohan (1993), the relationship between sexual isolation and sequence divergence (fig. 3) is given by a log-linear function: log10 ðqÞ 5 /p þ a FIG. 2.—Correlation between chvA sequence mismatch and CGM values in Agrobacterium tumefaciens: ‘‘gaps’’ in their distribution, delineating genomic species, are shaded. CGM values were determined by AFLP (Portier et al. 2006), and sequence mismatches were determined in this study. The graph was drawn by combining data from 225 pairwise comparisons. than four orders of magnitude among genes of the same genome in Ralstonia solanacearum (Fall et al. 2007). In the present study, we showed that chvA recombined up to 102 in homologous conditions. This value is high and comparable to the level noted with other marker genes in previous studies (Roberts and Cohan 1993; Humbert et al. 1995; Zawadzki et al. 1995; Majewski and Cohan 1998; Majewski et al. 2000). As it is easier and more accurate to compare data obtained with genes displaying high recombination frequencies, chvA thus appeared to be a suitable marker gene for studying recombinations. The homologous recombination sensitivity to divergence of chvA was tested relative to genomic species G8 and G2 by using C58 and TT9, and M2/1, respectively, as recipient strains, mated with clones of chvA alleles from a set of strains (fig. 1) displaying wide divergence with respect to recipient strains (table 2). The results showed sim- with q being the sexual isolation, / the sensitivity of sexual isolation to sequence divergence p, and a the y-intercept. No significant differences were found between slopes (F3.17 5 50.22, P 5 0.39 for /C58 and /TT9; F3,16 5 32.34, P 5 0.97 for /C58 and /M2/1; F3.11 5 23.62, P 5 0.57 for /TT9 and /M2/1). The average sensitivity of sexual isolation to sequence divergence in Agrobacterium is / 5 11.58. No significant differences in y-intercept values were found between aC58 and aM2/1 (F2.17 5 51.54, P 5 0.75) and between aTT9 and aM2/1 (F2.12 5 37.29, P 5 0.06). There was a significant difference in the y-intercept value between strains aC58 and aTT9 (F2.18 5 75.85, P 5 0.003). Departure from Exact Log-Linear Regression Paradoxically, the y-intercept in figure 3 is significantly different from zero (P . 0.05), suggesting that sexual isolation could occur even in homogamic conditions. Careful inspection of the curves revealed that the log-linear regression actually fitted better for nucleotide divergence ranging from 1% to 10%. Actually, with the three recipient strains, a dramatic increase in sexual isolation ratios was observed when comparing homologous conditions with heterologous crosses involving the most closely related foreign alleles (i.e., an increase of 4.3 for 1.2% divergence, 3.2 for Table 2 chvA Sequence Divergence, Recombination Frequencies, and Sexual Isolation Measured for Each Donor/Recipient chvA Allele in Agrobacterium tumefaciens Recipient Strains Donor chvA Allele C58 LMG 75 LMG 46 Mushin 6 TT9 NCPPB 925 RV3 B6 Kerr14 M2/1 CFBP 6927 CIP 497-74 CIP 107442 Sm 1021 (ndvA)d a C58 Divergence 0 0.012 0.017 0.036 0.043 0.064 0.086 0.092 0.093 0.101 0.101 0.103 0.103 0.243 TT9 RFa SIb Divergence c 2.00 ± 0.05 0 2.72 ± 0.16 0.63 ± 2.72 ± 0.18 0.70 ± 2.76 ± 0.08 0.83 ± 2.87 ± 0.08 0.91 ± 2.93 ± 0.12 0.96 ± 3.18 ± 0.06 1.21 ± 3.30 ± 0.08 1.29 ± 3.30 ± 0.08 1.33 ± 3.32 ± 0.02 1.39 ± 3.58 ± 0.10 1.61 ± 3.68 ± 0.06 1.71 ± 3.78 ± 0.07 1.81 ± No transformants 0.07 0.09 0.05 0.04 0.08 0.03 0.06 0.05 0.02 0.06 0.03 0.04 0.043 0.036 0.031 0.012 0 0.071 0.081 0.087 0.09 0.101 0.105 0.105 0.097 0.238 RFa M2/1 SIb 2.91 ± 0.04 0.65 ± NT NT 2.71 ± 0.05 0.50 ± 2.23 ± 0.02 0c 3.01 ± 0.09 0.78 ± NT NT 3.35 ± 0.08 1.08 ± NT 3.72 ± 0.11 1.47 ± 3.40 ± 0.05 1.18 ± 3.30 ± 0.07 1.11 ± No transformants Divergence 0.03 0.03 0.06 0.08 0.10 0.03 0.05 0.101 0.10 0.098 0.101 0.101 0.102 0.084 0.089 0.091 0 0.103 0.009 0.094 0.246 RFa SIb 3.64 ± 0.10 1.42 ± NT NT NT NT 3.68 ± 0.09 1.46 ± 3.43 ± 0.10 1.17 ± NT NT 2.24 ± 0.12 0c 4.11 ± 0.04 1.90 ± 3.09 ± 0.07 0.75 ± 3.61 ± 0.09 1.25 ± No transformants 0.02 0.05 0.04 0.02 0.06 0.05 RF 5 Log10(recombination frequency) ± standard error, based on at least four independent experimental assays. NT 5 not tested. SI 5 Log10(sexual isolation ratio) ± standard error. The values are obtained by dividing the frequency of homologous condition by the frequency of heterologous condition for each trial for each combination tested and then averaged. c The sexual isolation ratio in homologous conditions (between a recipient strain and its own chvA allele) is set at zero. d No transformants detected with the use of the ndvA gene, corresponding to a recombination frequency lower than the experimental detection limit, 2 106. b Sexual Isolation of Bacterial Species 173 This length matches MEPS length values reported in the literature: 23–50 bp for E. coli (Shen and Huang 1986; Smith 1988) or inferred by us from reported / estimates, that is, 49 bp for Bacillus subtilis and 42 bp for St. pneumoniae (Majewski and Cohan 1998; Majewski et al. 2000). Although A. tumefaciens (C58) is not known to be naturally competent, the homologous recombination system is efficient, as in other taxa known to be naturally competent, such as Bacillus and St. pneumoniae. Conclusions FIG. 3.—Log10(sexual isolation) versus chvA sequence divergence. Results for the three Agrobacterium recipient strains are represented, C58 (d), TT9 ()), and M2/1 (D). The average value of the three slopes /Agrobacterium 5 11.58 corresponds to the average sensitivity of sexual isolation to DNA sequence divergence in Agrobacterium species. 1.2%, and 5.6 for 0.9%, between C58 LMG 75, TT9 Mushin 6, and M2/1 CIP 497-74, respectively). Conversely, another dramatic increase was observed with matings involving alleles with more than 10% divergence (fig. 3). Molecular Analysis of chvA Recombinant Strains No significant sequence identity of chvA alleles or pK18mob was detected by the BlastN program within the whole C58 genome sequence (except chvA). All Km, Nm, and Rif resistant colonies later obtained with pK18mobchvA were thus primarily assumed to be bona fide recombinant strains resulting from single homologous recombination in the target chvA gene. This was experimentally verified by sequencing the insertion zone of at least one putative recombinant colony per cross. The sequences unambiguously showed that all tested recombinants had integrated pK18mob-chvA into the recipient chvA gene (data not shown). Efficiency of the Homologous Recombination System in A. tumefaciens According to previous studies in E. coli (Rayssiguier et al. 1989; Matic et al. 1995; Vulić et al. 1997; Štambuk and Radman 1998; Vulić et al. 1999), Bacillus (Zawadzki et al. 1995; Majewski and Cohan 1998, 1999), and St. pneumoniae (Humbert et al. 1995; Majewski et al. 2000), the two major factors limiting recombination are the mismatch repair system and the difficulty in heteroduplex formation, mediated by the RecA protein. To initiate recombination, RecA requires a minimum stretch of consecutive base pairs of perfect identity or minimal efficient processing segment (MEPS) between the recipient and donor (Shen and Huang 1986). The MEPS length differs among bacteria due to RecA diversity. As a result, in the present study, the MEPS length (n) was estimated at ;27 bp in Agrobacterium by using the formula of Majewski and Cohan (1998): n 5 / ln 10: Majewski (2001) indicated that ‘‘bacterial species experience a degree of sexual isolation from genetically divergent organisms since recombination occurs more frequently within species than between species.’’ Data reported in this review enabled us to speculate that bacterial species are highly and significantly sexually isolated, because 9% divergence at the marker gene between Haemophilus, Pseudomonas, or Streptococcus results in a drop of three orders of magnitude in recombination efficiency. Nevertheless, the drop in recombination efficiency, that is, the sexual isolation, was only 3-fold between B. subtilis and Bacillus mojavensis displaying only 4.6% divergence at the marker gene, indicating that these two Bacillus species are closely related. We thus assessed the reliability of this result in other taxa by thoroughly examining sexual isolation in the A. tumefaciens genomic species complex. Genomic species are defined by a ‘‘gap’’ in the wholegenome divergence distribution, and such gaps also occur in the distribution of most genes such as chvA, so the multiple loci sequence analysis method is efficient for taxonomy (Gevers et al. 2005). However, there is no such gap or drop in the distribution of sexual isolation values in the G8 model species (fig. 4), likely because the most proximal out-strain, that is, NCPPB 925 from G6, is very closely related to G8. Hence, at the within-to-between species border, for approximately 4.3–6.4% nucleotide divergence, sexual isolations were 8 and 9, respectively. In other words, with the chvA model, there was only a 1.13-fold decrease (i.e., 9/8) in recombination frequency for a nucleotide divergence difference of 2.1% (i.e., 6.4–4.3), thus for a narrow border between species (i.e., maximal intraspecies divergence within the heterogeneous genomic species G8 compared with minimal interspecies divergence between closely related G8 and G6). Indeed, the nucleotide divergence differences at species borders vary according to genes and species. Nevertheless, multilocus sequence analysis (MLSA) performed with A. tumefaciens showed that, except in cases of interspecies gene transfers, the smallest nucleotide divergence gaps at species borders of G8 were 2.3%, 3.3%, 5.4%, 5.0%, and 3.6% for recA, gyrB, mutS, gltD, and glgC, respectively (D. Costechareyre, F. Bertolla and X. Nesme, unpublished data). Assuming that the average sensitivity of sexual isolation to sequence divergence in Agrobacterium (/ 5 11.58) is the same for all genes, it is thus likely that the drop in recombination frequency between closely related but highly heterogeneous species varied from 1.4- to 4.2-fold at most. From a physiologic standpoint as determined in the laboratory and for a single 174 Costechareyre et al. FIG. 4.—Graph representing the nucleotide sequence divergence and the sexual isolation ratio between the reference strain C58 and other related strains. The DNA sequence divergence determined by AFLP (*) or with chvA sequences (), and the sexual isolation ratio (d) are represented. cross-over, declines in recombination frequency from within-to-between species are not different enough to significantly hamper gene flow between closely related species as compared with gene flow between distantly related strains within a species. Nevertheless, sexual isolation of a G8 member (C58) from the other nondistantly related genomic species of the A. tumefaciens complex varied gradually from 16 to 63 for G7 and G5, respectively, thus decreasing the chance of gene flow accordingly. Comparatively, we found a much more substantial drop in recombination efficiency (4.3-fold) with the least mismatching allele displaying 1.2% nucleotide divergence (fig. 3). For recombination involving different strains within the same species instead of different members of the same clone, sexual isolation actually occurs, thus contributing to preserving the genome identities of each clone or strain. A consequence of the genomic species definition for bacteria is that the same genome divergence cutoff is used regardless of the class. It spans around 5% of average nucleotide mismatch values for shared genes as determined with complete genome sequences (Goris et al. 2007). More precisely, MLSA showed that nucleotide divergence patterns within and between species for various genes are very similar in specific comparable genera. For instance, the maximal divergence within Vibrio cholerae was 5.2% for recA, and the minimal interspecies divergence of V. cholerae with its closest neighbor Vibrio mimicus was 7.3% (Thompson et al. 2005) dealing with a minimal nucleotide divergence gap of 2.1%. For the same gene but in the Burkholderia cenocepacia complex, the minimal gap between genomovar III-B and III-C was 1.9% (i.e., 4.5– 2.6; Baldwin et al. 2005). Comparably, in Ensifer/Sinorhizobium spp. (formerly Sinorhizobium spp.), the maximal heterogeneity within species is 1.2%, 1.3%, and 2.8% for dnaK, recA, and gyrB, for Ensifer arboris, Ensifer saheli, and Ensifer fredii, respectively. The respective minimal interspecies divergence of these genes for these species was 4.2%, 7.3%, and 6.7%, with Ensifer meliloti, Ensifer kostiensis, and Sinorhizobium americanum, respectively (Martens et al. 2008), yielding 3%, 6%, and 3.9% minimal divergence gaps, respectively, in this genus. Of course, with less heterogeneous or less related species, divergence gap values can be larger, for example, reaching 2% and 11% for rpoA and pheS, respectively, in Enterococcus spp. (Naser et al. 2005). Nevertheless, as a result, very similar divergence gap values have been found between closely related but heterogeneous species with several genes in several other genera. This suggests a similar minor effect of homologous recombination at species borders in other taxa. However, as the sensitivity of sexual isolation to sequence divergence is generally not known in other genera, it is presently not possible to quantify the corresponding declines in sexual isolation. In conclusion, the present definition of a bacterial species is based on the occurrence of a maximal genome divergence threshold that was empirically found to be almost the same in many taxa (Wayne et al. 1987). The delineation of genomic species in the A. tumefaciens species complex perfectly fits this definition with roughly the same canonical 70% RBR threshold value (Popoff et al. 1984). The present study of patterns that occur at species borders in A. tumefaciens could therefore likely be generalized to other species delineated with the same threshold. Similarly, the log-linear relationship between the recombination frequency and sequence divergence found with A. tumefaciens has also been found in several other models, thus indicating that the results obtained with this taxon could even be generalized to other taxa displaying a similar relationship. In addition, A. tumefaciens allowed us to finely study patterns between very closely related genomic species. The present study performed with a representative marker gene thus suggested that there is very little difference in horizontal gene transfers within and between very closely related but heterogeneous species in bacteria. Sequence diversity as a mechanism factor for the efficiency of the homologous recombination mechanism that is thought to actually control sexuality—whereby genes are integrated in the chromosomal genome and transmitted to the progeny—appears thus to play little role in the genetic cohesion of bacterial species. Our results thus support the conclusions of Hanage et al. (2006) that ‘‘empirical estimates of the relationship between sequence divergence and recombination rate indicate that the decline in recombination is an insufficiently steep function of genetic distance to generate species.’’ Obviously, sexual isolation of species can be experimentally— although often virtually—performed to circumscribe eukaryote species. The present findings indeed demonstrate that this is not the case for bacteria. Moreover, and contrary to what happens in eukaryotes, limited but numerous genetic exchanges occurred during the very long period during which speciation occurs in bacteria. Actually, according to the model of temporal fragmentation speciation of Retchless and Lawrence (2007), genetic isolation of bacteria may be established at different times for different chromosomal regions during speciation, because bacterial recombination involves the occasional transfer of only small DNA fragments instead of complete genomes. Consequently, bacterial speciation may have lasted millions of years following acquisition of the first niche-specific genes, a period during which homologous recombination likely occurred. Consequently, the BSC for eukaryotes defined by Mayr (1942), based on the occurrence of physiological barriers to sexual reproduction, would certainly not be applicable and Sexual Isolation of Bacterial Species 175 adaptable to the present genomic definition for a species described by Wayne et al. (1987) for prokaryotes. Nevertheless, as highlighted by Cohan (2001) and Majewski (2001), behavioral, physical, and ecological barriers can also create effective sexual isolations. In the related genus Sinorhizobium, Bailly et al. (2007) showed by linkage disequilibrium studies that even in the case of bacteria living in the same biotope, gene flow essentially occurs among strains belonging to the same species rather than between closely related but different species. Disconnected ecosystems likely constitute an efficient barrier to gene flow. Therefore, as genomic species are assumed to be ecotypes by some authors (Cohan 2001), it is likely that different ecological specializations determined by niche-specific genes prevent mating and subsequent recombinations between nascent species more surely than the RecA-dependent mechanism. Although the authors generally agree that cohesive evolutionary forces are necessary to generate microbial species, the fate of the current genomic species definition based on methodological approaches or even the identification of ‘‘species-specific’’ niches is questionable. Bacterial species were recently proposed to be metapopulation lineages contrary to a method-based concept (Achtman and Wagner 2008). With this concept, ‘‘species have just to be evolving from other such lineages. Microbes that form distinct groups owing to a cohesive force are metapopulation lineages and thus form species, whereas microbes without limits imposed by a cohesive force do not.’’ However, as stipulated by the authors, microbiologists must still agree on the criteria by which species can be recognized, and further experimental data will be necessary. Supplementary Material Supplementary tables and figures are available at Molecular Biology and Evolution online (http://www.mbe. oxfordjournals.org/). Acknowledgments This work was supported by the ‘‘Unmasking species specific genes and functions involved in species individualisation in ecotypes’’ Bureau des Ressources Génétiques program and a personal grant from INRA-DSPE for X.N.. D.C. received a PhD grant from the French MENESR. We are grateful to Andreas Schlüter for donating plasmid pK18mob, and we also thank Julien Gouré for sending the E. coli S17-1 strain. We are grateful to Philippe Oger, Patrick Mavingui, and Pascal Simonet for helpful discussions and suggestions, and Dr Amy Sapkota from Maryland University and translator Dr D. Manley for reading the manuscript and providing suggestions. We also thank David Chapulliot for skillful technical assistance. Literature Cited Achtman M, Wagner M. 2008. Microbial diversity and the genetic nature of microbial species. Nat Rev Microbiol. 6:431–440. Altschul SF, Gish W, Miller W, Myers EW, Lipman DJ. 1990. Basic local alignment search tool. J Mol Biol. 215:403–410. Bailly X, Olivieri I, Brunel B, Cleyet-Marel JC, G Béna G. 2007. Host-specialization and speciation in the symbiotic nitrogenfixing Sinorhizobium associated with Medicago. J Bacteriol. 189:5223–5236. Baldwin A, Mahenthiralingam E, Thickett KM, Honeybourne D, Maiden MC, Govan JR, Speert DP, Lipuma JJ, Vandamme P, Dowson CG. 2005. Multilocus sequence typing scheme that provides both species and strain differentiation for the Burkholderia cepacia complex. J Clin Microbiol. 43:4665 –4673. Breedveld MW, Miller KJ. 1994. Cyclic beta-glucans of members of the family Rhizobiaceae. Microbiol Rev. 58:145–161. Clark D, Maaløe O. 1967. DNA replication and the division cycle in Escherichia coli. J Mol Biol. 23:99–112. Cohan FM. 2001. Bacterial species and speciation. Syst. Biol. 50:513–524. De Ley J. 1974. Phylogeny of prokaryotes. Taxon. 23:291–300. De Ley J, Tijtgat R, De Smedt J, Michiels M. 1973. Thermal stability of DNA:DNA hybrids within the genus Agrobacterium. J Gen Microbiol. 78:241–252. Falls S, mercier A, Bertolla F, Calteau A, Gueguen L, Perriére G, Vogel TM, Simonet P. 2007. Horizontal gene transfer regulation in bacteria as a ‘‘spandrel’’ of DNA mechanisms. PLoS ONE. 10:e1055. Gevers D, Cohan FM, Lawrence JG, et al. (11 co-authors). 2005. Opinion: re-evaluating prokaryotic species. Nat Rev Microbiol. 3:733–739. Goris J, Konstantinidis KT, Klappenbach JA, Coenye T, Vandamme P, Tiedje JM. 2007. DNA-DNA hybridization values and their relationship to whole-genome sequence similarities. Int J Syst Evol Microbiol. 57:81–91. Hanage WP, Spratt BG, Turner KM, Fraser C. 2006. Modelling bacterial speciation. Philos Trans R Soc Lond B Biol Sci. 361:2039–2044. Humbert O, Prudhomme M, Hakenbeck R, Dowson CG, Claverys JP. 1995. Homeologous recombination and mismatch repair during transformation in Streptococcus pneumoniae: saturation of the Hex mismatch repair system. Proc Natl Acad Sci USA. 92:9052–9056. Kuhnert P, Korczak BM. 2006. Prediction of whole-genome DNA-DNA similarity, determination of G þ C content and phylogenetic analysis within the family Pasteurellaceae by multilocus sequence analysis (MLSA). Microbiology. 152: 2537–2548. Majewski J. 2001. Sexual isolation in bacteria. FEMS Microbiol Lett. 199:161–169. Majewski J, Cohan FM. 1998. The effect of mismatch repair and heteroduplex formation on sexual isolation in Bacillus. Genetics. 148:13–18. Majewski J, Cohan FM. 1999. DNA sequence similarity requirements for interspecific recombination in Bacillus. Genetics. 153:1525–1533. Majewski J, Zawadzki P, Pickerill P, Cohan FM, Dowson CG. 2000. Barriers to genetic exchange between bacterial species: Streptococcus pneumoniae transformation. J Bacteriol. 182: 1016–1023. Martens M, Dawyndt P, Coopman R, Gillis M, De Vos P, Willems A. 2008. Advantages of multilocus sequence analysis for taxonomic studies: a case study using 10 housekeeping genes in the genus Ensifer (including former Sinorhizobium). Int J Syst Evol Microbiol. 58:200–214. Matic I, Rayssiguier C, Radman M. 1995. Interspecies gene exchange in bacteria: the role of SOS and mismatch repair systems in evolution of species. Cell. 80:507–515. Mayr E. 1942. Systematics and the origin of species from the viewpoint of a zoologist. New York: Columbia University Press. 176 Costechareyre et al. Meade HM, Long SR, Ruvkun GB, Brown SE, Ausubel FM. 1982. Physical and genetic characterization of symbiotic and auxotrophic mutants of Rhizobium meliloti induced by transposon Tn5 mutagenesis. J Bacteriol. 149:114–122. Mougel C, Thioulouse J, Perriere G, Nesme X. 2002. A mathematical method for determining genome divergence and species delineation using AFLP. Int J Syst Evol Microbiol. 52:573–586. Naser SM, Thompson FL, Hoste B, Gevers D, Dawyndt P, Vancanneyt M, Swings J. 2005. Application of multilocus sequence analysis (MLSA) for rapid identification of Enterococcus species based on rpoA and pheS genes. Microbiology. 151:2141–2150. Popoff MY, Kersters K, Kiredjian M, Miras I, Coynault C. 1984. Position taxonomique de souches de Agrobacterium d’origine hospitalière. Ann Microbiol. 135A:427–442. Portier P, Fischer-Le Saux M, Mougel C, Lerondelle C, Chapulliot D, Thioulouse J, Nesme X. 2006. Identification of genomic species in Agrobacterium biovar 1 by AFLP genomic markers. Appl Environ Microbiol. 72:7123–7131. Rayssiguier C, Thaler DS, Radman M. 1989. The barrier to recombination between Escherichia coli and Salmonella typhimurium is disrupted in mismatch-repair mutants. Nature. 342:396–401. Retchless AC, Lawrence JG. 2007. Temporal fragmentation of speciation in bacteria. Science. 317:1093–1096. Richter D, Postic D, Sertour N, Livey I, Matuschka FR, Baranton G. 2006. Delineation of Borrelia burgdorferi sensu lato species by multilocus sequence analysis and confirmation of the delineation of Borrelia spielmanii sp. nov. Int J Syst Evol Microbiol. 56:873–881. Roberts MS, Cohan FM. 1993. The effect of DNA sequence divergence on sexual isolation in Bacillus. Genetics. 134: 401–408. Rozas J, Rozas R. 1999. DnaSP version 3: an integrated program for molecular population genetics and molecular evolution analysis. Bioinformatics. 15:174–175. Sambrook J, Russell D. 2001. Molecular cloning: a laboratory manual. Cold Spring Harbor, NY: Cold Spring Harbor Laboratory Press. Schäfer A, Tauch A, Jager W, Kalinowski J, Thierbach G, Puhler A. 1994. Small mobilizable multi-purpose cloning vectors derived from the Escherichia coli plasmids pK18 and pK19: selection of defined deletions in the chromosome of Corynebacterium glutamicum. Gene. 145:69–73. Shen P, Huang HV. 1986. Homologous recombination in Escherichia coli: dependence on substrate length and homology. Genetics. 112:441–457. Simon R, Priefer U, Pehle A. 1983. A broad host range mobilization system for in vitro genetic engineering: transposon mutagenesis in gram negative bacteria. Biotechnology. 1:784–790. Smith GR. 1988. Homologous recombination in procaryotes. Microbiol Rev. 52:1–28. Stackebrandt E, Frederiksen W, Garrity GM, et al. (13 coauthors). 2002. Report of the ad hoc committee for the reevaluation of the species definition in bacteriology. Int J Syst Evol Microbiol. 52:1043–1047. Štambuk S, Radman M. 1998. Mechanism and control of interspecies recombination in Escherichia coli. I. Mismatch repair, methylation, recombination and replication functions. Genetics. 150:533–542. Thompson FL, Gevers D, Thompson CC, Dawyndt P, Naser S, Hoste B, Munn CB, Swings J. 2005. Phylogeny and molecular identification of vibrios on the basis of multilocus sequence analysis. Appl Environ Microbiol. 71:5107–5115. Vulić M, Dionisio F, Taddei F, Radman M. 1997. Molecular keys to speciation: DNA polymorphism and the control of genetic exchange in enterobacteria. Proc Natl Acad Sci USA. 94:9763–9767. Vulić M, Lenski RE, Radman M. 1999. Mutation, recombination, and incipient speciation of bacteria in the laboratory. Proc Natl Acad Sci USA. 96:7348–7351. Wayne LG, Brenner DJ, Colwell RR, et al. (12 co-authors). 1987. Report of the ad hoc committee on reconciliation of approaches to bacterial systematics. Int J Syst Bacteriol. 37: 463–464. Zawadzki P, Roberts MS, Cohan FM. 1995. The log-linear relationship between sexual isolation and sequence divergence in Bacillus transformation is robust. Genetics. 140: 917–932. Andrew Roger, Associate Editor Accepted October 11, 2008