Abstract
Biovar 1 of the genus Agrobacterium consists of at least nine genomic species that have not yet received accepted species names. However, rapid identification of these organisms in various biotopes is needed to elucidate crown gall epidemiology, as well as Agrobacterium ecology. For this purpose, the AFLP methodology provides rapid and unambiguous determination of the genomic species status of agrobacteria, as confirmed by additional DNA-DNA hybridizations. The AFLP method has been proven to be reliable and to eliminate the need for DNA-DNA hybridization. In addition, AFLP fragments common to all members of the three major genomic species of agrobacteria, genomic species G1 (reference strain, strain TT111), G4 (reference strain, strain B6, the type strain of Agrobacterium tumefaciens), and G8 (reference strain, strain C58), have been identified, and these fragments facilitate analysis and show the applicability of the method. The maximal infraspecies current genome mispairing (CGM) value found for the biovar 1 taxon is 10.8%, while the smallest CGM value found for pairs of genomic species is 15.2%. This emphasizes the gap in the distribution of genome divergence values upon which the genomic species definition is based. The three main genomic species of agrobacteria in biovar 1 displayed high infraspecies current genome mispairing values (9 to 9.7%). The common fragments of a genomic species are thus likely “species-specific” markers tagging the core genomes of the species.
Agrobacteria are common soil and root inhabitants. They are generally inoffensive plant commensals. However, agrobacteria may cause crown gall or hairy root diseases in many plants, including economically important crops, when they harbor a large Ti or Ri plasmid (8, 36). Ti and Ri plasmids can be experimentally transferred by conjugation (13) to numerous bacterial species, most of which are in the family Rhizobiaceae (17, 34, 35). In most cases, new transconjugant species or genera become pathogenic (5, 17, 34, 35). Probably due to in situ plasmid transfer, pathogenic populations of agrobacteria involved in crown gall outbreaks are characterized by a high degree of chromosomal background diversity in terms of genotype, serotype, ribotype, and biovar, while the same Ti plasmids can be associated with various chromosome backgrounds (25, 26, 28, 30). Therefore, elucidating crown gall epidemiology and Ti plasmid ecology requires identification of genuine or potential bacterial Ti plasmid reservoirs in soil and plants at various taxonomic levels.
The current consensus for bacterial species identification (40) is based on DNA-DNA hybridization of whole genomes and indicates that “the phylogenetic definition of a species generally would include strains with approximately 70% or greater DNA-DNA relatedness and with 5°C or less ΔTm.” This definition has recently been supported by an international committee (32), but the committee also indicated that delineation of genomic species can now be obtained with alternative molecular methods that sample parts of the genome instead of DNA-DNA hybridization of whole genomes alone. Among the alternative genome sampling methods, AFLP (38) easily identifies genomic species (24). Multilocus sequence typing (22) or, more generally, multilocus sequence analysis (MLSA) has also been proposed for identification of bacterial species (14). AFLP has the advantage of identifying bacterial species on a genomic basis. The AFLP method is based on analysis of markers which are massively and randomly sampled along the entire genome. Thus, AFLP targets not only genes in the conserved core genome but also accessory genes that have likely ecological relevance for the species. This is important for the concept of ecological species (6, 37), which have been described as clusters of strains adapted to the same ecological niches (e.g., ecotypes). In addition, the AFLP methodology provides high resolution and reveals differences between strains not distinguished by MLSA (1).
Agrobacterium taxonomy is still strongly debated. First, it was proposed that so-called “biovars 1 and 2,” which are defined by biochemical characteristics (18, 19, 20), should be redefined at the species level as A. radiobacter and A. rhizogenes, respectively (31). However, based on the International Code of Nomenclature of Bacteria (20a), Bouzar (2) indicated that A. tumefaciens was the type species of the genus and thus should be retained. Controversially, in a request for an opinion, Young et al. (44) decided that the epithet “radiobacter” was proposed before the epithet “tumefaciens” and that A. radiobacter has priority over A. tumefaciens. Moreover, although A. rhizogenes (i.e., “biovar 2”) is a true genomic species, as revealed by DNA-DNA hybridization, and therefore a bona fide species, “biovar 1” was found to be a complex of several genomic species (9, 10, 29). Researchers identified nine genomic species or genomospecies in biovar 1 by DNA-DNA hybridization, which were subsequently designated genomic species G1 to G9 (24). The splitting of “biovar 1” into several genomic species does not solve the terminology problem. Actually, the type strain of A. tumefaciens (strain B6 [= ATCC 23308]) belongs to genomic species G4; thus, only genomic species G4 corresponds to the bona fide species A. tumefaciens sensu stricto. Unfortunately, the type strain of A. radiobacter (ATCC 19358) also belongs to the same genomic species (genomic species G4). A choice between A. tumefaciens and A. radiobacter must be made for the name of this taxon. Meanwhile, strain C58, which was completely sequenced twice (15, 41), belongs to genomic species G8, which has not received an accepted species name yet. Genomic species G1 is another unnamed genomic species frequently found in crown gall outbreaks (unpublished results). Second, there is also a long-standing quarrel about the existence of the genus Agrobacterium itself (11, 42, 43). Briefly, Young et al. (42) proposed that the genus Agrobacterium should be eliminated by including all agrobacteria in Rhizobium, essentially to solve the rrs (i.e., 16S rRNA gene) polyphyly problem encountered with the former classification. Farrand et al. (11) recognized Agrobacterium as a definable genus, leaving the rrs polyphyly question pending. The subcommittee on the taxonomy of Agrobacterium and Rhizobium suggested that it is up to individual authors to choose which name they want to use (21). A compromise position is used in this paper. We refer to Agrobacterium as the rrs monophyletic genus, which includes “biovar 1” members and three closely related species, A. larrymoorei (3), A. vitis (formerly known as biovar 3 [27]), and A. rubi (33). Based on rrs phylogeny, strain NCPPB1650 belonging to an undefined related species should also be included in the genus Agrobacterium. The remote “biovar 2” is considered Rhizobium rhizogenes.
The present study was done to facilitate high-throughput identification of agrobacteria using the AFLP methodology, which is useful for genomic species identification and infraspecific typing of a large number of strains. Catalogues of specific AFLP genomic markers, which are conserved among all genomic species members, were established using a set of 52 agrobacterial strains from international collections, and there was a particular emphasis on the three most frequent genomic species of Agrobacterium biovar 1, genomic species G1, G4, and G8. This approach was validated by subsequent DNA-DNA hybridization of some environmental isolates identified as members of genomic species G1, G4, and G8 by AFLP.
MATERIALS AND METHODS
Bacterial strains.
The strains analyzed (Table 1) represent almost all the genomic species richness of the genus Agrobacterium known so far. We placed special emphasis on genomic species G1, G4, and G8 because the numbers of strains of these genomic species identified by DNA-DNA hybridization were greatest (9, 11, and 5 strains, respectively) and because these three species are so closely related that they cannot be distinguished by conventional biochemical tests, leading to their inclusion in a single biovar called “biovar 1.” All strains are available from the Collection Française de Bactéries Phytopathogènes, INRA, Angers, France; most are also available from the Laboratorium voor Microbiologie, Ghent University, Ghent, Belgium. The clinical isolates of agroabacteria are available from the Collection de l'Institut Pasteur.
TABLE 1.
Genomic species and strains of Agrobacterium used in this study
| Strain | Other designation(s)a | Biological source, geographical origin, and other information | DNA-DNA hybridization results (reference)b |
|---|---|---|---|
| Genomic species G1c | |||
| TT111d | CFBP 5767, LMG 196 | Crown gall, United States | 10 |
| ATCC 4720 | CFBP 5493, LMG 182 | Black raspberry, Iowa | 29 |
| NCPPB 396 | CFBP 5765, LMG 176 | Dahlia sp., Germany | 10 |
| S377 | CFBP 5768, LMG 326 | Plant | 29 |
| S56 | CFBP 5491, LMG 321 | Plant | 29 |
| S4 | CFBP 5492, LMG 318 | Plant | 29 |
| CFBP 5622 | Solanum nigrum, root tissue commensal, LCSA,e France | This study | |
| CFBP 2517 | Hybrid poplar, Leuce section, gall, France | This study | |
| CFBP 5771 | Bulk soil, LCSA, France | This study | |
| CFBP 2712 | Prunes persica × Prunus amygdalus cv. GF677 gall, France | ND | |
| Genomic species G2 | |||
| CIP 497-74 | CFBP 5494, CFBP 2884 | Human blood, France | 29 |
| Bernaerts M2/1 | CFBP 5876, LMG 147 | Ditch water, Belgium | 29 |
| CIP 28-75 | CFBP 5495 | Human urine, France | 29 |
| CIP 43-76 | CFBP 5496 | Human urine, France | 29 |
| Genomic species G3 | |||
| CIP 107443 | CFBP 6623, CIP 493-74 | Antiseptic flask, France | 29 |
| CIP 107442 | CFBP 6624, CIP 111-78 | Human, France | 29 |
| Genomic species G4 (A. tumefaciens sensu stricto) | |||
| B6T | CFBP 2413T, LMG 187T | Apple seedling, Iowa; A. tumefaciens type strain | 29 |
| CIP 67-1 | CFBP 2413T, LMG 187T | Other designation for B6 | 29 |
| ATCC 4452 | CFBP 5766, LMG 181 | Rubus idaeus, Iowa | 10 |
| ATCC 4718 | CFBP 5764, LMG 139 | Soil, United States | 10 |
| Kerr 14 | CFBP 5761, LMG 15 | Soil around Prunus dulcis, South Australia | 10 |
| Hayward 0322 | CFBP 5770, LMG 1687 | Prunus persica stock gall, South Australia | 10 |
| ATCC 19358 | CFBP 5522, LMG 140 | Soil; A. radiobacter type strain | 29 |
| LMG 340 | CFBP 5769, ICPB TT11 | Librocedrus sp. gall, United States | 10 |
| LMG 62 | No information available | 10 | |
| CFBP 5621 | Lotus corniculata, root tissue commensal, LCSA, France | This study | |
| CFBP 5627 | Bulk soil, LCSA, France | This study | |
| CFBP 2514 | Vitis vinifera gall, Spain | ND | |
| Genomic species G5 | |||
| CIP 107444 | CFBP 6626, CIP 120-78 | Human cephalorachidian liquid, France | 29 |
| CIP 107445 | CFBP 6625, CIP 291-77 | Patient food, France | 29 |
| Genomic species G6 | |||
| NCPPB 925 | CFBP 5499, LMG 225 | Dahlia sp., South Africa | 29 |
| Zutra F/1 | CFBP 5877, LMG 296 | Dahlia sp., Israel | 29 |
| Genomic species G7 | |||
| Zutra 3/1 | CFBP 6999, LMG 298 | Malus sp., Israel | 29 |
| RV3 | CFBP 5500, LMG 317 | No information available | 29 |
| NCPPB 1641 | CFBP 5502, LMG 228 | Flacourtia indica, United Kingdom | 29 |
| Genomic species G8 | |||
| C58 | CFBP 1903, LMG 287 | Prunus sp. cv. Montmorency (cherry), New York | 9 |
| TT9 | CFBP 5504, LMG 64 | Likely hop, United States | 29 |
| T37 | CFBP 5503, LMG 332 | Juglans sp. gall, California | 29 |
| Mushin 6 | CFBP 6550, LMG 201 | Humulus lupulus gall, Victoria, Australia | 29 |
| LMG 75 | CFBP 6549 | Euonymus alata cv. Compacta gall, United States | ND |
| LMG 46 | CFBP 6554 | Rubus macropetalus, Oregon | ND |
| AW137 | LMG R-10181 | Transmitted by A. Willems, LMG | ND |
| J-07 | CFBP 5773 | Bulk soil, LCSA, France | This study |
| Genomic species G9 | |||
| Hayward 0363 | CFBP 5506, LMG 27 | John Innes potting soil, Australia | 29 |
| Hayward 0362 | CFBP 5507, LMG 26 | John Innes potting soil, Australia | 29 |
| Agrobacterium sp. strain NCPPB 1650 | CFBP 4470, LMG 230 | Rosa sp., South Africa | ND |
| A. larrymoorei AF 3.10T | CFBP 5473, LMG 21410 | Ficus benjamina; type strain | 3 |
| A. rubi LMG 17935T | CFBP 5509T | Rubus ursinus var. Loganobaccus, United States; type strain | 20 |
| A. vitis K309T | CFBP 5523T | Vitis vinifera; type strain | 27 |
| Rhizobium rhizogenes | |||
| LMG 150T | CFBP 2408T, CFBP 5520T | Apple tree; type strain | 29 |
| K84 | CFBP 1937, LMG 138 | Prunus persica, soil around gall, Australia | ND |
| CFBP 2519 | Hybrid poplar gall, Leuce section, France | ND |
CFBP, Collection Française de Bactéries Phytopathogènes, INRA, Angers, France; LMG, Laboratorium voor Microbiologie, Ghent University, Ghent, Belgium; CIP, Collection de l'Institut Pasteur.
Reference for strain assignment to a genomic species as determined by DNA-DNA hybridization. ND, DNA-DNA hybridization not done.
Genomic species G1 to G9 based on the nomenclature of Mougel et al. (24).
The first strain of each genomic species is the reference strain or the accepted type strain of the described species.
LCSA, maize field at La-Côte-Saint-André, Insère, France.
Eight environmental strains isolated from galls, bulk soil, or root tissue identified as members of biovar 1 after conventional biochemical tests (production of 3-keto-lactose) were included in the study. The AFLP classification in genomic species for six of them was later confirmed by DNA-DNA hybridization.
DNA extraction and purification.
DNA used in AFLP experiments was extracted with a DNeasy tissue kit (QIAGEN, Hilden, Germany) by following the manufacturer's instructions. Genomic DNAs used in DNA-DNA hybridization were extracted as described by Brenner et al. (4).
AFLP analyses.
The AFLP methodology used in this study was adapted from that of Vos et al. (38). The EcoRI and MseI endonucleases were used for genomic DNA restriction as previously described (24), using adaptors and PCR primers shown in Table 2. The digestion-ligation step was performed for 2 h at 37°C with an 11-μl (final volume) mixture by incubating 55 ng of genomic DNA with EcoRI (5 U), MseI (5 U), T4 DNA ligase (1 U), the appropriate quantity of T4 DNA ligase buffer (Boeringer-Manheim, Germany), 0.5 μg of bovine serum albumin, 50 μM NaCl, the EcoRI-specific adaptor (0.18 μM) prepared with F1363-adEco+ hybridized with F1931-adEco−, and the MseI-specific adaptor (1.8 μM) prepared with F1365-adMse+ hybridized with F1931-adMse−. Each adapted DNA (4 μl) was then subjected to a nonselective PCR performed in a 20-μl (final volume) mixture containing 15 μl of the AFLP amplification CoreMix (Perkin-Elmer Applied Biosystems, Foster City, Calif.), 0.25 μM primer F1247-coreEco, and 0.25 μM primer F1248-coreMse. A PE-9600 thermocycler (Perkin-Elmer) was used with the following PCR program: denaturation at 94°C for 20 s, annealing at 56°C for 30 s, and elongation at 72°C for 2 min for 20 cycles. The quality of nonselective PCRs was controlled by agarose gel electrophoresis before storage of well-amplified products at −20°C. A well-amplified product was then typically diluted 1/30 before it was used as a template (1.5 μl) in a selective PCR mixture (final volume, 10 μl) with 7.5 μl of AFLP amplification CoreMix, primer F1248-coreMse (0.25 μM), and fluorescently labeled primers (0.05 μM). The following fluorescently labeled primers were designed with the F1247-coreEco sequence plus two discriminant nucleotides at the 3′ end: F1598-EcoCA-FAM, F1599-EcoCC-HEX, F1601-EcoCG-HEX, and F1915-EcoCT-FAM for selective nucleotides CA, CC, CG and CT, respectively. Selective PCRs were performed using a touchdown procedure consisting of denaturation at 94°C for 20 s, annealing at temperatures ranging from 66 to 57°C (the temperature was decreased 1°C per cycle) for 30 s, and elongation at 72°C for 2 min for 10 cycles, followed by a conventional PCR consisting of denaturation at 94°C for 20 s, annealing at 56°C for 30 s, and elongation at 72°C for 2 min for 20 cycles.
TABLE 2.
AFLP oligonucleotides used to construct adaptors and to prime PCRs
| Oligonucleotide | Sequence |
|---|---|
| EcoRI-specific adaptors | |
| F1363-adEco+ | CTCGTAGACTGCGTACC |
| F1931-adEco− | AATTGGTACGCAGTCTAC |
| MseI-specific adaptors | |
| F1365-adMse+ | GACGATGAGTCCTGAG |
| F1931-adMse− | TACTCAGGACTCAT |
| Core primers | |
| F1247-coreEco | GACTGCGTACCAATTC |
| F1248-coreMse | GATGAGTCCTGAGTAA |
| Selective primers | |
| F1598-EcoCA-FAM | GACTGCGTACCAATTCCA |
| F1599-EcoCC-HEX | GACTGCGTACCAATTCCC |
| F1601-EcoCG-HEX | GACTGCGTACCAATTCCG |
| F1915-EcoCT-FAM | GACTGCGTACCAATTCCT |
The total selective PCR products (10 μl) were purified on a Sephadex G-50 column (Amersham Biosciences, Orsay, France) before separation of the AFLP fragments by electrophoresis with a capillary sequencer (MegaBACE 1000; Amersham Pharmacia Biotech Europe, Orsay, France). The device automatically calculated the length of the fluorescent fragments by comparison to the MegaBACE ET-900-R size standard (Amersham). A genetic profiler (model 1.5; Molecular Dynamics Inc., Sunnyvale, Calif.) was used to display the results and to export data in text format. Data were transferred to a spreadsheet with the program Thresholdfilter 1.3 (Yann Legros, Amersham). A threshold fluorescence value of 200 arbitrary units was generally used to eliminate the background before subsequent bioinformatic treatment.
For each strain, experiments under the four AFLP conditions (EcoRI+CA/MseI+0, EcoRI+CC/MseI+0, EcoRI+CT/MseI+0, and EcoRI+CG/MseI+0) were performed in duplicate. AFLP fragments were coded as in the following example: the CA109 fragment corresponded to a 109-bp fragment experimentally obtained under the Eco+CA/Mse+0 AFLP conditions.
Predictive AFLP.
A predictive AFLP analysis was performed with the complete genome sequence of C58 essentially as previously described by Mougel et al. (24) by simulating digestion with EcoRI and with MseI and then selecting restriction fragments based on the selective nucleotides added to selective primers. The lengths of the predicted AFLP fragments corresponded to the lengths of the restriction fragments plus 27 bp for the adaptors.
Phylogenomic analyses.
The LecPCR and Align2 programs were used to transform raw data into tabular binary data, and the DistAFLP program was used to calculate the evolutionary genome divergence (rate of nucleotide substitution) and current genome mispairing (CGM) essentially as described by Mougel et al. (24), except that fragments were placed in length classes for genomic species instead of classes for all the strains together, which was done in the previous study. The LecPCR, Align2, and DistAFLP programs are accessible at http://pbil.univ-lyon1.fr/ADE-4/microb/. Dendrograms were calculated with the Neighbor/UPGMA module of the PHYLIP package (12) using evolutionary genome divergence data as distance data and were displayed with NJ-Plot (http://pbil.univ-lyon1.fr/software/njplot.html). Bootstrap values were calculated by using the bootstrap option of DistAFLP and the Neighbor/UPGMA and Consense modules of the PHYLIP package.
Determination of common AFLP fragments.
A newly developed program, AlignK (http://pbil.univ-lyon1.fr/ADE-4/microb/), was used to align fragment patterns in length classes in a spreadsheet by gathering fragments with the same length ± 0.5 bp in the same line (one line for each fragment length class). After visual correction, the program calculated how frequent a fragment of a given length was, its average length, and the length standard deviation. The program was used to determine the AFLP fragments common to all members of the same species.
DNA-DNA hybridization.
Native DNAs were labeled in vitro by random priming with tritium-labeled nucleotides using the Megaprime DNA labeling systems (Amersham Biosciences). DNA-DNA hybridization was performed by using the S1 nuclease-trichloroacetic acid method (7, 16). Reassociation was performed at 70°C in 0.42 M NaCl. DNA-DNA hybridizations were carried out using labeled DNAs from strains TT111, C58, and B6T.
RESULTS
Experimental AFLP analyses.
The average numbers of fragments per AFLP pattern obtained using the EcoRI+CA/MseI+0, EcoRI+CC/MseI+0, EcoRI+CT/MseI+0, and EcoRI+ CG/MseI+0 conditions were 69, 45, 76, and 66 fragments, respectively. The phylogenomic analysis separated groups of strains on long branches supported by significant bootstrap values (>80%) (Fig. 1). These clusters are in accordance with genomic species assignments described previously (29). The maximum CGM value obtained for two strains in a genomic species was 10.8% (for Zutra 3/1 and RV3 in genomic species G7) (Fig. 1). The three major genomic species, genomic species G1, G4, and G8, exhibited comparable high levels of diversity; the greatest CGM values within these genomic species were 9.7%, 9.7%, and 9.0%, respectively, and the average CGM values were 7.5%, 6.5%, and 7.2%, respectively. A relevant finding is that the smallest CGM value found for comparisons of genomic species was 15.2% (for strain Zutra 3/1 in genomic species G7 and strain Hayward 0362 in genomic species G9) (Fig. 1).
FIG. 1.
Phylogenomic relatedness of Agrobacterium strains, species, and genomic species, calculated by the neighbor-joining method using pooled data for four AFLP conditions. The values are significant bootstrap values (i.e., >80%) obtained from 100 data resamplings. The unit of evolutionary distance is the rate of nucleotide substitution per nucleotide site calculated with DistAFLP.
Confidence interval for AFLP fragment length.
The AFLP fragment lengths provided by the sequencer were subject to variation due to experimental conditions and sequence differences. The rrs gene was used to estimate the variation in fragment length, because fragment lengths are identical within all members of a given genomic species but there are small sequence differences between genomic species G1, G4, and G8 (24). The lengths of two fragments (CA109 and CG221) flanking the EcoRI site occurring in the rrs genes of genomic species G1, G4, and G8 strains (accession numbers of reference strains, AJ389895, AJ389904, and AJ012209, respectively) were studied within and between genomic species. The lengths and sequences of the CG221 fragment, predicted from rrs sequences, were identical in all members of the three genomic species. The lengths of the experimental rrs fragments were 221.40 ± 0.04, 221.38 ± 0.03, and 221.38 ± 0.04 bp (averages ± standard deviations) in genomic species G1, G4, and G8, respectively (Table 3), indicating that the lengths of fragments with identical sequences could differ by a maximum of 0.10 bp (P < 0.05). The sequence of the rrs CA109 fragment differed by four nucleotides at positions 522, 523, 533, and 534 in genomic species G8 and G1 or G4 (5% mispairing), but the fragments were identical within genomic species. The lengths of the experimental rrs fragments were 108.37 ± 0.05, 108.40 ± 0.02, and 107.95 ± 0.05 bp (averages ± standard deviations) for genomic species G1, G4, and G8, respectively (Table 3), indicating that there was a difference of about 0.60 bp (P < 0.05) between the estimated lengths of experimental fragments having about 5% different nucleotides. Taking this uncertainty into account and considering that the maximum level of whole-genome mispairing within a genomic species is greater (9 to 10%) (see above), we assumed that the estimated lengths of fragments originating from the same genome region in different strains could differ by 1 bp (i.e., average length ± 0.5 bp).
TABLE 3.
AFLP fragments common to all members of genomic species G1, G4, and G8
| Genomic species | Fragment size (bp)
|
|||||||
|---|---|---|---|---|---|---|---|---|
| EcoRI+CA/MseI+0
|
EcoRI+CC/MseI+0
|
EcoRI+CG/MseI+0
|
EcoRI+CT/MseI+0
|
|||||
| Mean | SD | Mean | SD | Mean | SD | Mean | SD | |
| G1 | 52.10 | 0.07 | 48.94 (G4)b | 0.12 | 48.16 | 0.18 | 65.70 (G8) | 0.08 |
| 65.91 | 0.23 | 64.35 | 0.09 | 55.90 (G4, G8) | 0.09 | 128.12 | 0.08 | |
| 67.04 | 0.04 | 89.22 (G4, G8) | 0.08 | 63.11 | 0.07 | 150.23 | 0.07 | |
| 89.31 | 0.08 | 161.35 | 0.12 | 91.53 | 0.29 | 156.55 | 0.12 | |
| 108.37 (rrs)a | 0.05 | 184.75 | 0.21 | 94.68 | 0.25 | 171.60 (G8) | 0.22 | |
| 116.93 | 0.30 | 573.97 (G8) | 0.13 | 108.51 (G8) | 0.04 | 186.06 | 0.16 | |
| 131.91 | 0.05 | 176.49 | 0.12 | 191.17 (G4) | 0.12 | |||
| 136.59 | 0.08 | 195.83 | 0.17 | 204.87 | 0.17 | |||
| 146.81 | 0.33 | 213.68 | 0.08 | 246.84 | 0.63 | |||
| 163.08 (G8) | 0.26 | 221.40 (rrs) | 0.04 | 386.72 | 0.28 | |||
| 177.57 (G4) | 0.19 | 232.70 | 0.12 | 525.49 (G4) | 0.12 | |||
| 191.68 | 0.28 | 236.35 | 0.29 | 528.16 | 0.19 | |||
| 271.39 | 0.49 | 244.49 | 0.13 | 589.69 (G4, G8) | 0.23 | |||
| 295.09 | 0.12 | 250.73 | 0.11 | |||||
| 339.75 | 0.43 | 253.91 | 0.18 | |||||
| 418.95 | 0.21 | 260.71 (G8) | 0.14 | |||||
| 423.55 | 0.21 | 299.40 | 0.16 | |||||
| 504.42 | 0.30 | 308.22 | 0.11 | |||||
| 525.96 | 0.18 | 542.74 | 0.15 | |||||
| G4 | 81.06 | 0.30 | 49.13 (G1) | 0.07 | 50.81 (G8) | 0.06 | 62.35 | 0.06 |
| 108.40 (rrs) | 0.02 | 89.48 (G1, G8) | 0.06 | 55.74 (G1, G8) | 0.05 | 74.74 | 0.06 | |
| 144.36 | 0.05 | 167.22 | 0.17 | 71.53 | 0.08 | 93.73 | 0.07 | |
| 171.92 | 0.03 | 219.43 | 0.18 | 104.90 | 0.28 | 148.39 | 0.25 | |
| 174.63 | 0.03 | 271.83 | 0.08 | 124.80 | 0.13 | 191.11 (G1) | 0.06 | |
| 178.03 (G1) | 0.05 | 286.36 | 0.10 | 131.20 | 0.06 | 197.20 | 0.05 | |
| 183.05 | 0.09 | 333.43 (G8) | 0.30 | 157.89 | 0.04 | 201.10 | 0.03 | |
| 190.48 | 0.29 | 372.56 | 0.37 | 199.50 | 0.05 | 212.16 | 0.26 | |
| 210.70 | 0.15 | 433.68 | 0.23 | 204.49 | 0.11 | 251.25 | 0.21 | |
| 238.33 | 0.07 | 454.53 | 0.17 | 206.99 | 0.04 | 435.75 | 0.15 | |
| 243.79 (G8) | 0.30 | 518.13 | 0.08 | 221.38 (rrs) | 0.03 | 524.95 (G1) | 0.07 | |
| 263.43 | 0.26 | 588.61 | 0.15 | 245.76 (G8) | 0.38 | 589.77 (G1, G8) | 0.37 | |
| 270.03 | 0.42 | 255.65 | 0.06 | |||||
| 379.76 | 0.33 | 321.66 | 0.07 | |||||
| 422.31 | 0.25 | 430.58 | 0.14 | |||||
| 426.37 | 0.12 | |||||||
| G8 | 36.62 | 0.45 | 82.60 | 0.07 | 40.70 | 0.16 | 40.78 | 0.13 |
| 68.20 | 0.30 | 84.90 | 0.13 | 50.80 (G4) | 0.06 | 66.80 (G1) | 0.07 | |
| 95.63 | 0.26 | 89.30 (G1, G4) | 0.07 | 55.70 (G1, G4) | 0.14 | 88.48 | 0.08 | |
| 106.20 | 0.31 | 213.00 | 0.14 | 103.92 | 0.08 | 136.80 | 0.07 | |
| 108.00 (rrs) | 0.05 | 334.30 (G4) | 0.05 | 108.78 (G1) | 0.05 | 172.70 (G1) | 0.27 | |
| 121.54 | 0.06 | 520.10 | 0.16 | 120.00 | 0.04 | 266.91 | 0.54 | |
| 164.40 (G1) | 0.32 | 574.10 | 0.16 | 180.50 | 0.30 | 292.30 | 0.21 | |
| 210.24 | 0.10 | 190.70 | 0.13 | 300.60 | 0.24 | |||
| 231.34 | 0.24 | 221.38 (rrs) | 0.04 | 302.09 | 0.45 | |||
| 233.23 | 0.05 | 223.53 | 0.26 | 336.00 | 0.14 | |||
| 244.37 (G4) | 0.05 | 246.23 (G4) | 0.32 | 374.34 | 0.09 | |||
| 355.6 | 0.07 | 252.80 | 0.13 | 441.70 | 0.59 | |||
| 463.25 | 0.11 | 260.56 | 0.33 | 590.46 (G1, G4) | 0.22 | |||
| 507.26 | 0.15 | 275.40 | 0.06 | |||||
| 627.13 | 0.11 | 318.10 | 0.08 | |||||
| 357.18 | 0.12 | |||||||
| 363.03 | 0.12 | |||||||
| 370.14 | 0.09 | |||||||
| 374.54 | 0.30 | |||||||
| 391.20 | 0.19 | |||||||
| 400.30 | 0.26 | |||||||
| 510.40 | 0.21 | |||||||
rrs, fragments whose lengths corresponds to the lengths of fragments predicted to be liberated from rrs genes in all members of genomic species G1, G4, and G8.
G1, G4, and G8, fragments also found in genomic species G1, G4, and G8 members, respectively.
AFLP markers of genomic species G1, G4, and G8.
The fragments common to all members of a genomic species were determined by aligning the AFLP patterns of all members of the genomic species with AlignK (Table 3). The total numbers of common fragments for the four AFLP conditions were 57, 55, and 57 for genomic species G1, G4, and G8, respectively. The genomic markers of a genomic species were generally found to be common only to members of that genomic species, even if they could be found in some strains belonging to another genomic species (data not shown). Thus, combinations of these genomic markers could be used as discriminative characteristics for genomic species. Nevertheless, as expected from the rrs sequence analysis, predicted fragments CA109 and CG221 flanking the EcoRI site found in the rrs genes of genomic species G1, G4, and G8 strains were detected in all members of the three genomic species, and the sizes of the fragments detected ranged from 108.00 to 108.47 bp for CA109 and from 221.38 to 221.40 bp for CG221 (Table 3). Three other fragments, CC89 (range, 89.22 to 89.48 bp), CG56 (range, 55.70 to 55.90 bp), and CT590 (range, 589.69 to 590.46 bp), were found to be common to the three genomic species, and some other fragments were common to pairs of genomic species. Moreover, the two putative rrs fragments (CA109 and CG221) and the CT590 fragment were found in all biovar 1 members, while fragments CC89 and CG56 were common to most but not all biovar 1 members (CC89 was not found in genomic species G3 and G9; CG56 was not found in genomic species G2 and G9). Thus, the presence of all five of these fragments is unique for genomic species G1, G4, G5, G6, G7, and G8, and these fragments did not occur outside biovar 1, while a combination of CT590, CA109, and CG221 was found both in all biovar 1 members and also in A. vitis (data not shown).
An attempt was made to identify fragments conserved in biovar 1 by predictive AFLP. This investigation revealed that CA109 and CG221 probably originated from the four rrs copies, ATU0053, ATU2547, ATU3937, and ATU4180, since there are four ribosomal operons in C58 with identical rrs copies (15, 41). The other conserved fragments, CG56, CC89, and CT590, presumptively originated from ATU0223, ATU2571, and a region spanning the ATU3980 and ATU3981 genes, respectively. Remarkably, ATU0223, ATU2571, and ATU3980 are not distributed at random over the genomes but are located in the vicinity of the ATU0053, ATU2547, and ATU3937 rrs gene copies at 170 kbp, 25 kbp, and 53 kbp, respectively.
Confirmation of AFLP species identification by DNA-DNA hybridization.
Eight strains isolated from plants or soil samples whose classification in genomic species was not previously determined were used in this study. The AFLP patterns of these strains unambiguously placed them in genomic species G1, G4, or G8. The AFLP placement of six of them (CFBP 5622, CFBP 2517, CFBP 5771, CFBP 5621, CFBP 5627, and J-07) in genomic species G1, G4, and G8 was confirmed a posteriori by DNA-DNA hybridizations performed with TT111, B6T, and C58, respectively (Table 4).
TABLE 4.
Genomic species assignment of environmental agrobacteria by DNA-DNA hybridization
| Source of unlabeled DNA | Genomic species | Relative binding ratio (%) with 3H-labeled DNA from strain:
|
||
|---|---|---|---|---|
| TT111 | B6T | C58 | ||
| TT111 | G1 | 100 | 45 | 45 |
| CFBP 5622 | G1 | 77 | NDa | 45 |
| CFBP 2517 | G1 | 78 | ND | 45 |
| CFBP 5771 | G1 | 75 | ND | 41 |
| B6T | G4 | 42 | 100 | 52 |
| CFBP 5621 | G4 | ND | 76 | ND |
| CFBP 5627 | G4 | 37 | 75 | ND |
| C58 | G8 | 39 | 58 | 100 |
| CFBP 5773 | G8 | ND | ND | 87 |
ND, hybridization not done.
DISCUSSION
The AFLP methodology clearly delineates all genomic species presently known in the genus Agrobacterium when enough AFLP conditions are used to obtain clusters supported by significant bootstrap values (24). Some genomic species, such as genomic species G7 and G9, display great infraspecies diversity. The DNA-DNA hybridization values for these genomic species are close to or even less than 70%, and to ascertain the genomic species status, it is necessary to verify that differences in melting temperatures for members of a species are less than 5°C (29). These highly diverse genomic species were significantly separated from their close neighbors if up to four AFLP conditions were used (bootstrap values, >80%), while only one AFLP condition was enough to correctly place a strain in a less diverse genomic species (data not shown). In all cases, the AFLP approach rapidly and unambiguously determined the genomic species status of agrobacteria isolated from various biotopes, as well as strains deposited in international bacterial collections. Moreover, the results were unambiguously confirmed by further DNA-DNA hybridizations for six environmental strains. As a consequence, AFLP has been proven to be reliable and has rendered DNA-DNA hybridization unnecessary for placing environmental agrobacteria into genomic species.
In spite of the power of AFLP for species identification, AFLP data are not easily compared by phylogenetic methods when they are obtained with different sequencers (gel versus capillary; different molecular weight markers; different data extraction software). It would be easier to identify a species by a set of well-defined molecular markers. In order to facilitate the analysis and to examine the universal applicability of the method, AFLP fragments common to all members of a genomic species were determined for the three major genomic species of agrobacteria, genomic species G1, G4, and G8 (Table 2). We verified that most of these “species-specific” markers were present in AFLP profiles obtained previously with another sequencer and with smaller sets of strains (http://pbil.univ-lyon1.fr/ADE-4/microb/IJSEM2001), provided that a larger confidence interval for the fragment size determination was used since different molecular weight ladders were used. Thus, the combination of the fragments listed Table 3 is sufficient to place an isolate into genomic species G1, G4, or G8. Thus, we describe here a reliable and portable method for identifying the three major Agrobacterium genomic species even in the absence of databases and phylogenomic analyses. With the present list of “species-specific” markers, this approach is accessible to all laboratories without a need for phylogenetic analyses.
At higher taxonomic levels, only five AFLP fragments, including two fragments of rrs, were found to be common to all or almost all biovar 1 members. This small number of markers is not enough to allow significant and robust identification of a new isolate as a member of biovar 1 by AFLP. Nevertheless, these conserved fragments likely originated from the core genome of biovar 1 strains and are good candidates for definition of markers for sequence-based approaches, such as PCR-restriction fragment length polymorphism analysis or MLSA. Identification of the fragments conserved in biovar 1 was attempted. We used predictive AFLP because fragments detected by fluorescent AFLP (which provides only virtual images of fragments) are difficult to clone and sequence. Remarkably, all the biovar 1 markers were presumptively found to originate from genes located close to the ribosomal operons. Housekeeping genes, such as recA, gyrB, groEL, or mutS, which are frequently used in MLSA, were found in the vicinity of the ribosomal operons. This confirmed that AFLP markers readily tag the core genome of biovar 1 strains and that standard housekeeping genes are the most relevant genes for MLSA.
One of the advantages of AFLP is that it measures diversity by estimating maximal and average current genome mispairing of populations (23). This provided a minimal set of strains which displayed the highest diversity in a given species and which were the strains that were best suited for determining the most conserved markers in a given species. Using strains from various collections, the three main genomic species of agrobacteria in biovar 1 displayed infraspecies CGM values of 9 to 9.7%, which were close to the maximum value found in another genomic species in biovar 1 (10.8%). As a result, the list of markers in Table 3 is likely to be “species specific,” and the markers tag the core genomes of the genomic species.
Current genome mispairing values are key parameters for species delineation as they are highly correlated to the data obtained by the conventional DNA-DNA hybridization procedure (24). In the present study, the maximum CGM value found for two strains in a genomic species was 10.8%, while the smallest CGM value found for two genomic species was 15.2%. This emphasizes the gap in the distribution of genome divergence values on which the current genomic species definition is based. These results establish the CGM threshold values that could be used to delineate genomic species. The values are actually slightly different but likely much more accurate than the values reported previously (24), because they were calculated by using a more relevant procedure for assignment of AFLP fragments to fragment length classes.
In order to determine whether the CGM threshold values determined for Agrobacterium spp. are relevant for delineating genomic species in other taxa, it is necessary to perform extensive studies before the method can be generalized. However, as an indication, the minimal interspecies CGM value for agrobacteria as determined by AFLP (15.2%) appears to be close to the value for whole-genome mispairing estimated by Vulic et al. (39) using DNA-DNA hybridization data for Escherichia coli and Salmonella enterica (16%). These two species are very closely related but are sexually isolated, as measured by the dramatic drop in the rate of homologous recombination within and between species (39). This strengthens the idea that the genome divergence threshold, which delineates genomic species, including agrobacterial genomic species (70% for DNA-DNA, corresponding to 15% genome mispairing), is a critical parameter of the sexual isolation of bacterial species.
In this study, we developed an original method to place rapidly an agrobacterial strain into one of the three major Agrobacterium genomic species (genomic species G1, with reference strain TT111; genomic species G4 or bona fide A. tumefaciens, including type strain B6; and genomic species G8, including sequenced strain C58) by determining lists of common AFLP markers detected in all strains of the three species. This procedure is thought to allow high-throughput identification and genomic species classification of agrobacteria isolated from field samples.
Acknowledgments
We thank Anne Willems and Moniek Gillis of the Laboratorium voor Microbiologie, Ghent University, and Martine Kedidjian of Institut Pasteur, who kindly provided strains for this study; Marie-André Poirier for technical help; and Tim Vogel for reading the manuscript and providing suggestions. AFLP analyses were performed using the sequencing facilities of the DTAMB at the Université Claude Bernard-Lyon 1.
Perrine Portier received a Ph.D. grant from MENRT. This work was financially supported by a personal grant from INRA-DSPE to Xavier Nesme.
Footnotes
Published ahead of print on 25 August 2006.
REFERENCES
- 1.Boudon, S., C. Manceau, and J.-L. Nottéghem. 2005. Structure and origin of Xanthomonas arboricola pv. pruni populations causing bacterial spot of stone fruit trees in Western Europe. Phytopathology 95:1081-1088. [DOI] [PubMed]
- 2.Bouzar, H. 1994. Request for a judicial opinion concerning the type species of Agrobacterium. Int. J. Syst. Bacteriol. 44:373-374. [Google Scholar]
- 3.Bouzar, H., and J. B. Jones. 2001. Agrobacterium larrymoorei sp. nov., a pathogen isolated from aerial tumours of Ficus benjamina. Int. J. Syst. Evol. Microbiol. 51:1023-1026. [DOI] [PubMed] [Google Scholar]
- 4.Brenner, D. J., A. C. McWhorter, J. K. Leete Knutson, and A. G. Steigerwalt. 1982. Escherichia vulneris: a new species of Enterobacteriaceae associated with human wounds. J. Clin. Microbiol. 15:1133-1140. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Broothaerts, W., H. J. Mitchell, B. Weir, S. Kaines, L. M. Smith, W. Yang, J. E. Mayer, C. Roa-Rodriguez, and R. A. Jefferson. 2005. Gene transfer to plants by diverse species of bacteria. Nature 433:629-633. [DOI] [PubMed] [Google Scholar]
- 6.Cohan, F. M. 2001. Bacterial species and speciation. Syst. Biol. 50:513-524. [DOI] [PubMed] [Google Scholar]
- 7.Crosa, J. H., D. J. Brenner, and S. Falkow. 1973. Use of a single-strand-specific nuclease for analysis of bacterial and plasmid deoxyribonucleic acid homo- and heteroduplexes. J. Bacteriol. 115:904-911. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.De Cleene, M., and J. De Ley. 1976. The host range of crown gall. Bot. Rev. 42:389-466. [Google Scholar]
- 9.De Ley, J. 1974. Phylogeny of procaryotes. Taxon 23:291-300. [Google Scholar]
- 10.De Ley, J., R. Tijtgat, J. De Smedt, and M. Michiels. 1973. Thermal stability of DNA:DNA hybrids within the genus Agrobacterium. J. Gen. Microbiol. 78:241-252. [Google Scholar]
- 11.Farrand, S. K., P. B. Van Berkum, and P. Oger. 2003. Agrobacterium is a definable genus of the family Rhizobiaceae. Int. J. Syst. Evol. Microbiol. 53:1681-1687. [DOI] [PubMed] [Google Scholar]
- 12.Felsenstein, J. 2004. PHYLIP: Phylogeny Inference Package, version 3.6. Department of Genome Sciences, University of Washington, Seattle. [Online.] http://evolution.genetics.washington.edu/phylip.html.
- 13.Genetello, C., N. Van Larebeke, M. Holsters, A. De Picker, M. Van Montagu, and J. Schell. 1977. Ti plasmids of Agrobacterium tumefaciens as conjugative plasmids. Nature 265:561-563. [DOI] [PubMed] [Google Scholar]
- 14.Gevers, D., F. M. Cohan, J. G. Lawrence, B. G. Spratt, T. Coenye, E. J. Feil, E. Stackebrandt, Y. van De Peerqq, P. Vandamme, F. L. Thompson, and J. Swings. 2005. Reevaluating prokaryotic species. Nat. Microbiol. Rev. 3:733-739. [DOI] [PubMed] [Google Scholar]
- 15.Goodner, B., G. Hinkle, S. Gattung, N. Miller, M. Blanchard, B. Qurollo, B. S. Goldman, Y. Cao, M. Askenazi, C. Halling, L. Mullin, K. Houmiel, J. Gordon, M. Vaudin, O. Iartchouk, A. Epp, F. Liu, C. Wollam, M. Allinger, D. Doughty, C. Scott, C. Lappas, B. Markelz, C. Flanagan, C. Crowell, J. Gurson, C. Lomo, C. Sear, G. Strub, C. Cielo, and S. Slater. 2001. Genome sequence of the plant pathogen and biotechnology agent Agrobacterium tumefaciens C58. Science 294:2323-2328. [DOI] [PubMed] [Google Scholar]
- 16.Grimont, P. A. D., M. Y. Popoff, F. Grimont, C. Coynault, and M. Lemelin. 1980. Reproducibility and correlation study of three deoxyribonucleic acid hybridization procedures. Curr. Microbiol. 4:325-330. [Google Scholar]
- 17.Hooykaas P. J. J., P. M. Klapwijk, M. P. Nuti, R. A. Schilperoort, and A. Rorsh. 1977. Transfer of the Agrobacterium tumefaciens Ti-plasmid to avirulent agrobacteria and to Rhizobium ex-planta. J. Gen. Microbiol. 98:477-484. [Google Scholar]
- 18.Keane, P. J., A. Kerr, and P. B. New. 1970. Crown gall on stone fruits: identification and nomenclature of Agrobacterium isolates. Aust. J. Biol. Sci. 23:585-595. [Google Scholar]
- 19.Kersters, K., J. De Ley, P. H. A. Sneath, and M. Sackin. 1973. Numerical taxonomic analysis of Agrobacterium. J. Gen. Microbiol. 78:227-239. [Google Scholar]
- 20.Kersters, K., and J. De Ley. 1984. Agrobacterium Conn 1942, p. 244-254. In N. R. Krieg and J. G. Holt (ed.), Bergey's manual of systematic bacteriology, vol. 1. Williams & Wilkins, Baltimore, Md. [Google Scholar]
- 20a.Lapage, S. P., P. H. A. Sneath, E. F. Lessel, V. B. D. Skerman, H. P. R. Seeliger, and W. A. Clark (ed.). 1992. International code of nomenclature of bacteria. American Society for Microbiology, Washington, D.C. [PubMed]
- 21.Lindström, K., and M. E. Martinez-Romero. 2002. International Committee on Systematics of Prokaryotes, Subcommittee on the Taxonomy of Agrobacterium and Rhizobium, minutes of the meeting, 4th July 2001, Hamilton, Canada. Int. J. Syt. Evol. Microbiol. 52:2337. [Google Scholar]
- 22.Maiden, M. C., J. A. Bygraves, E. Feil, G. Morelli, J. E. Russell, R. Urwin, Q. Zhang, J. Zhou, K. Zurth, D. A. Caugant, I. M. Feavers, M. Achtman, and B. G. Spratt. 1998. Multilocus sequence typing: a portable approach to the identification of clones within populations of pathogenic microorganisms. Proc. Natl. Acad. Sci. USA 95:3140-3145. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Mougel, C., S. Teyssier, C. d'Angelo, K. Groud, M. Neyra, K. Sidi-Boumedine, A. Cloeckaert, M. Peloille, S. Baucheron, E. Chaslus-Dancla, S. Jarraud, H. Meugnier, F. Forey, F. Vandenesch, G. Lina, J. Etienne, J. Thioulouse, C. Manceau, P. Robbe, R. Nalin, J. Briolay, and X. Nesme. 2001. Experimental and theoretical evaluation of typing methods based upon random amplification of genomic restriction fragments (AFLP) for bacterial population genetics. Genet. Sel. Evol. 33(Suppl. 1):S319-S338. [Google Scholar]
- 24.Mougel, C., J. Thioulouse, G. Perriere, and X. Nesme. 2002. A mathematical method for determining genome divergence and species delineation using AFLP. Int. J. Syst. Evol. Microbiol. 52:573-586. [DOI] [PubMed] [Google Scholar]
- 25.Nesme, X., M. F. Michel, and B. Digat. 1987. Population heterogeneity of Agrobacterium tumefaciens in galls of Populus L. from a single nursery. Appl. Environ. Microbiol. 53:655-659. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Nesme, X., C. Ponsonnet, C. Picard, and P. Normand. 1992. Chromosomal and pTi genotypes of Agrobacterium strains isolated from Populus tumors in two nurseries. FEMS Microbiol. Ecol. 101:189-196. [Google Scholar]
- 27.Ophel, K., and A. Kerr. 1990. Agrobacterium vitis sp. nov. for strains of Agrobacterium biovar 3 from grapevines. Int. J. Syst. Bacteriol. 40:236-241. [Google Scholar]
- 28.Pionnat, S., H. Keller, D. Héricher, A. Bettachini, Y. Dessaux, X. Nesme, and C. Poncet. 1999. Ti plasmid from Agrobacterium characterizes rootstock clones that initiate spread of crown gall disease in Mediterranean countries. Appl. Environ. Microbiol. 65:4197-4206. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Popoff, M. Y., K. Kersters, M. Kiredjian, I. Miras, and C. Coynault. 1984. Taxonomic position of Agrobacterium strains of hospital origin. Ann. Microbiol. 135:427-442. [PubMed] [Google Scholar]
- 30.Raio, A., R. Peluso, X. Nesme, and A. Zoina. 2004. Chromosomal and plasmid diversity of Agrobacterium populations isolated from Ficus benjamina tumors. Eur. J. Plant Pathol. 109:163-173. [Google Scholar]
- 31.Sawada, H., H. Ieki, H. Oyaizu, and S. Matsumoto. 1993. Proposal for rejection of Agrobacterium tumefaciens and revised descriptions for the genus Agrobacterium and for Agrobacterium radiobacter and Agrobacterium rhizogenes. Int. J. Syst. Bacteriol. 43:694-702. [DOI] [PubMed] [Google Scholar]
- 32.Stackebrandt, E., W. Frederiksen, G. M. Garrity, P. A. Grimont, P. Kampfer, M. C. Maiden, X. Nesme, R. Rossello-Mora, J. Swings, H. G. Truper, L. Vauterin, A. C. Ward, and W. B. Whitman. 2002. Report of the Ad Hoc Committee for the Re-evaluation of the Species Definition in Bacteriology. Int. J. Syst. Evol. Microbiol. 52:1043-1047. [DOI] [PubMed] [Google Scholar]
- 33.Starr, M. P., and J. E. Weiss. 1943. Growth of phytopathogenic bacteria in a synthetic asparagin medium. Phytopathology 33:314-318. [Google Scholar]
- 34.Teyssier-Cuvelle, S., C. Mougel, and X. Nesme. 1999. Direct conjugal transfers of Ti plasmid to soil microflora. Mol. Ecol. 8:1273-1284. [DOI] [PubMed] [Google Scholar]
- 35.Teyssier-Cuvelle S., P. Oger, C. Mougel, K. Groud, S. Farrand, and X. Nesme. 2004. A highly selectable and highly transferable Ti plasmid to study conjugal host range and Ti plasmid dissemination in complex ecosystems. Microb. Ecol. 48:10-18. [DOI] [PubMed] [Google Scholar]
- 36.Van Larebeke, N., G. Engler, M. Holsters, S. Van den Elsacker, I. Zaenen, R. A. Schilperoort, and J. Schell. 1974. Large plasmid in Agrobacterium tumefaciens essential for crown gall-inducing ability. Nature 252:169-170. [DOI] [PubMed] [Google Scholar]
- 37.Van Valen, L. 1976. Ecological species, multispecies, and oaks. Taxon 25:233-239. [Google Scholar]
- 38.Vos, P., R. Hogers, M. Bleeker, M. Reijans, T. Van De Lee, M. Hornes, A. Frijters, J. Pot, J. Peleman, and M. Kuiper. 1995. AFLP: a new technique for DNA fingerprinting. Nucleic Acids Res. 23:4407-4414. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39.Vulic, M., F. Dionisio, F. Taddei, and M. Radman. 1997. Molecular keys to speciation: DNA polymorphism and the control of genetic exchange in enterobacteria. Proc. Natl. Acad. Sci. USA 94:9763-9767. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 40.Wayne, L. G., D. J. Brenner, R. R. Colwell, P. A. D. Grimont, O. Kandler, M. I. Krichevisky, L. H. Moore, W. E. C. Moore, R. G. E. Murray, E. Stackebrandt, M. P. Starr, and H. G. Truper. 1987. Report of the Ad Hoc Committee on Reconciliation of Approaches to Bacterial Systematics. Int. J. Syst. Bacteriol. 37:463-464. [Google Scholar]
- 41.Wood, D. W., J. C. Setubal, R. Kaul, D. E. Monks, J. P. Kitajima, V. K. Okura, Y. Zhou, L. Chen, G. E. Wood, N. F. Almeida, Jr., L. Woo, Y. Chen, I. T. Paulsen, J. A. Eisen, P. D. Karp, D. Bovee, Sr., P. Chapman, J. Clendenning, G. Deatherage, W. Gillet, C. Grant, T. Kutyavin, R. Levy, M. J. Li, E. McClelland, A. Palmieri, C. Raymond, G. Rouse, C. Saenphimmachak, Z. Wu, P. Romero, D. Gordon, S. Zhang, H. Yoo, Y. Tao, P. Biddle, M. Jung, W. Krespan, M. Perry, B. Gordon-Kamm, L. Liao, S. Kim, C. Hendrick, Z. Y. Zhao, M. Dolan, F. Chumley, S. V. Tingey, J. F. Tomb, P. G. Milton, V. O. Maynard, and W. E. Nester. 2001. The genome of the natural genetic engineer Agrobacterium tumefaciens C58. Science 294:2317-2323. [DOI] [PubMed] [Google Scholar]
- 42.Young, J. M., L. D. Kuykendall, E. Martinez-Romero, A. Kerr, and H. Sawada. 2001. A revision of Rhizobium Frank 1889, with an emended description of the genus, and the inclusion of all species of Agrobacterium Conn 1942 and Allorhizobium undicola De Lajudie et al. 1998 as new combinations: Rhizobium radiobacter, R. rhizogenes, R. rubi, R. undicola and R. vitis. Int. J. Syst. Evol. Microbiol. 51:89-103. [DOI] [PubMed] [Google Scholar]
- 43.Young, J. M., L. D. Kuykendall, E. Martinez-Romero, A. Kerr, and H. Sawada. 2003. Classification and nomenclature of Agrobacterium and Rhizobium—a reply to Farrand et al. 2003. Int. J. Syst. Evol. Microbiol. 53:1689-1695. [DOI] [PubMed] [Google Scholar]
- 44.Young, J. M., S. R. Pennycook, and D. R. W. Watson. 2006. Proposal that Agrobacterium radiobacter has priority over Agrobacterium tumefaciens. Request for an opinion. Int. J. Syst. Evol. Microbiol. 56:491-493. [DOI] [PubMed] [Google Scholar]