School of Molecular and Microbial Biosciences, Bldg G08, University of Sydney, NSW 2006, Australia
Correspondence
Peter R. Reeves
reeves{at}angis.usyd.edu.au
![]() |
ABSTRACT |
---|
![]() ![]() ![]() ![]() ![]() ![]() |
---|
The GenBank accession numbers for the sequences reported in this paper are AY238992AY239002.
Present address: Hypertension and Stroke Research Laboratories, Department of Physiology, University of Sydney and Department of Neurosurgery, Royal North Shore Hospital, St Leonards, NSW 2065, Australia.
![]() |
INTRODUCTION |
---|
![]() ![]() ![]() ![]() ![]() ![]() |
---|
Apart from its medical significance, O antigen is also a good candidate for studying lateral gene transfer. The genes responsible for O antigen biosynthesis are generally clustered at a specific locus of the chromosome. In the case of Salmonella enterica, a strongly clonal species, transfer of housekeeping genes between subspecies occurs rarely (Selander et al., 1996), yet O antigen forms are generally distributed among the subspecies, suggesting relatively frequent transfer of O antigens between subspecies (Reeves, 1997
). It has been proposed, based on atypical GC content, that O antigen gene clusters of S. enterica and Escherichia coli are derived from polysaccharide gene clusters of other species (Reeves, 1993
). Lateral transfer of O antigen gene clusters has also been proposed to account for the generation of specific clones such as the epidemic V. cholerae strain, O139 Bengal (Bik et al., 1995
; Comstock et al., 1995
; Manning et al., 1994
) and the O157 : H7 clone of E. coli (Tarr et al., 2000
).
The genes in O antigen gene clusters generally fall into three classes: pathway genes for the biosynthesis of nucleotide sugars, glycosyltransferase genes for linking the sugars, and processing genes for polymerization and transportation of O units. Compared with glycosyltransferase genes and processing genes, which are commonly heterogeneous due to the different linkages involved, pathway genes are in general homologous regardless of species. For this reason they can be particularly useful in studies of lateral gene transfer in O antigen gene clusters, or studies on evolutionary relationships of polysaccharide gene clusters from different species.
L-Rhamnose is commonly present in O antigens of Gram-negative bacteria and capsular polysaccharides (CPS) of Gram-positive bacteria. dTDP-L-rhamnose is the activated form for the rhamnose moiety in both O antigen and CPS synthesis. The synthesis of dTDP-L-rhamnose from D-glucose 1-phosphate is catalysed by four enzymes encoded by rmlA, rmlB, rmlC and rmlD, which are generally grouped together within O antigen or CPS gene clusters. The rml gene sets have been sequenced from a number of species and are clearly homologous although the gene order may vary (Jiang et al., 1991; Mitchison et al., 1997
; Stevenson et al., 1994
; Yoshida et al., 1998
).
The rml genes offer an excellent opportunity to follow the movement of gene clusters subject to lateral transfer. In contrast to pathogenicity islands, the distribution of O antigen gene clusters is well known as it corresponds to the distribution of the O antigens of serotyping. The rml genes are always present if rhamnose is in the structure and provide a good basis for determining origins and relationships. We have recently reported on sequence variation of rml genes in S. enterica (Li & Reeves, 2000). All the sequenced S. enterica rml gene sets are located at the 5' end of the O antigen gene cluster, and we observed a strong polarity in the level of sequence similarity and in the nature of variation within rml genes. The 5' rml genes are highly similar and subspecies-specific while the 3' rml genes are much more variable and exhibit O antigen-specific sequences. We proposed that rml-containing O antigen gene clusters commonly transfer between subspecies by recombination in the 3' end of the rml gene set.
In this paper we report the results of a similar study of rml gene sets of V. cholerae, which has a large number of O antigen specificities. At the time this work started, 155 O antigens were recognized, of which 33 had been shown to contain L-rhamnose (Kondo et al., 1997). We sequenced and analysed the full rml gene set from 11 O antigens with different extents of divergence among the four genes. Our results suggest that rml genes of V. cholerae have come from diverse sources. A polarity of sequence similarity was also observed in these rml gene sets, suggesting that as for S. enterica, recombination within rml genes is important in lateral transfer of the O antigen gene clusters, with substantial impact on the levels of variation in the shared rml genes.
![]() |
METHODS |
---|
![]() ![]() ![]() ![]() ![]() ![]() |
---|
|
|
|
Construction of rmlC strain P5435.
When we started this study, no rml genes from V. cholerae had been sequenced. A method for cloning rml genes from plasmid gene banks was devised by modification of the method described by Clarke & Whitfield (1992).
We first constructed two E. coli K-12 strains with deletions of either all four rml genes or just rmlC. Only the rmlC mutant worked and was used in this study. The K-12 recD strain DB1316 (P5416) (Wertman et al., 1986) was used for construction of the
rmlC strain. Like most K-12 strains, it has the wbaN50 mutation that inactivates synthesis of the rhamnosyltransferase (Liu & Reeves, 1994
). pPR1474 has a 5 kb segment including a functional wbaN gene and adjacent DNA of K-12 (Liu & Reeves, 1994
). We isolated the segment by cutting pPR1474 with EcoRI and SphI and transformed it into DB1316. The culture was then spread on plates pre-seeded with 105 p.f.u. of bacteriophage Ffm (Wilkinson et al., 1972
). Phage Ffm lyses E. coli strains with rough LPS (R-LPS) (Schmidt et al., 1974
). The strain P5418, with a functional rhamnosyltransferase gene recovered by homologous recombination, was isolated by resistance to bacteriophage Ffm. This strain has the complete O antigen gene cluster of K-12.
A rmlC mutation was then generated by homologous recombination between P5418 chromosomal DNA and a DNA segment which extended from 2·4 kb upstream of rmlC to 2·25 kb downstream of rmlC of the E. coli K-12 O antigen gene cluster but had rmlC replaced by a 0·85 kb kanamycin-resistance cassette (Fig. 1). The construction of this segment was as follows. Primers 904 and 905, with EcoRI and BamHI sequences respectively attached, were used for amplification of the 2·4 kb upstream fragment, and primers 789 and 790, which include BamHI and XbaI sequences, respectively, were used for amplification of the 2·25 kb downstream fragment. The downstream fragment was cloned into pGEM7zf(+) BamHI and XbaI sites. The upstream fragment was cloned into the resulting plasmid using EcoRI and BamHI. A 0·85 kb kanamycin-resistance gene cassette which had been amplified from Kanamycin Resistance GenBlock (Pharmacia) by primers 787 and 788 with attached BamHI sites was inserted into the BamHI site between the upstream and the downstream fragments. One plasmid with correct orientation was named pPR1964.
|
Cloning rml genes from V. cholerae using E. coli strain P5435.
Chromosomal DNA from each V. cholerae strain was partially digested with Sau3A. DNA fragments of 48 kb were collected from an agarose gel and ligated to BamHI-digested pGEM7zf(+) vector. The ligation mixture was transformed into P5435. The bacteria transformed with a plasmid containing a rml gene should make smooth LPS, which would prevent the bacteria from being lysed by phage Ffm; such bacteria were obtained by selection for resistance to bacteriophage Ffm (105 p.f.u. phage pre-seeded on a plate), and further examined by agglutination with O16 antisera. Several Ffm phage-resistant clones were isolated from each gene bank and their plasmids were prepared. The ends of the cloned fragments were sequenced using M13 forward and reverse primers to screen rml-containing plasmids and those that contained all or most of the rml gene sets were chosen for further sequencing.
The O66 rmlC gene could not be cloned using the phage Ffm selection method. The O66 rmlB gene was first amplified and sequenced using primers based on the O6 sequence. To obtain genes downstream of rmlB, we used suppression PCR as described previously (Lan & Reeves, 1998). Two micrograms of O66 chromosomal DNA was digested completely with 30 units of BglI, BglII, ClaI, DraI, HpaI, KpnI, EcoRI or XbaI. Overhanging ends were filled using 10 pmol dNTPs and 1 unit Klenow. The reaction mixtures were extracted with phenol/chloroform, precipitated and dissolved in 20 µl TE. Ligation to the adaptor (Table 3
) (Lan & Reeves, 1998
) was carried out in a 20 µl reaction which contains 200 ng of digested fragment, 1 µM adaptor, 15 % PEG, 1x ligation buffer and 1 unit of T4 DNA ligase, incubated at 16 °C overnight. The ligation mix was heat-inactivated at 65 °C for 15 min, diluted 10-fold and 1 µl added to 50 µl PCR reaction using two outer primers 1513 and 604 (Lan & Reeves, 1998
), which were based on the rmlB gene and the adaptor, respectively. To confirm the specificity of the PCR fragment, the PCR product was diluted 10 000-fold and 1 µl used for PCR amplification with two internal primers 1317 and 605 (Lan & Reeves, 1998
).
LPS analysis and SDS-PAGE.
Preparation of LPS, SDS-PAGE and silver staining were carried out as described by Franco et al. (1998).
Computer analysis.
DNA sequence data were assembled and edited using programs from ANGIS (The Australian National Genomic Information Service) at the University of Sydney. Pairwise comparisons and polymorphism analysis of DNA sequence data were conducted using the MULTICOMP package (Reeves et al., 1994), which incorporates a number of programs for DNA sequence and phylogeny analysis. Multiple protein sequence alignment were conducted using PILEUP (Feng & Doolittle, 1986
). Phylogenetic trees were constructed by the neighbour-joining method (Saitou & Nei, 1987
) using PHYLIP (version 3.4 written by Joseph Felsenstein, Department of Genetics, University of Washington, Seattle, USA).
![]() |
RESULTS AND DISCUSSION |
---|
![]() ![]() ![]() ![]() ![]() ![]() |
---|
|
PCR amplification using primers based on the O1 and O139 hldD sequences, and the O6 rmlB sequence, showed that the O6 rml gene set is located at the 5' end of the O antigen gene cluster downstream of hldD, with a 237 bp intergenic region that included a JUMPstart sequence (Hobbs & Reeves, 1994) 5794 bp upstream of rmlB.
Probes specific for the O6 rml genes were used to hybridize Southern blots of HindIII- or PstI-digested chromosomal DNA of the remaining 32 strains of those representing the 33 rhamnose-containing O antigens (O6 was included as a control) (Table 4). Thirteen strains showed strong hybridization signals for all four rml genes. Another five showed strong signals for one to three rml genes, but weak or no signal for the others, with the frequency and level of signal decreasing from rmlB to rmlC in map order from 5' to 3' end. DNA from the remaining 14 strains did not hybridize to any of the probes.
|
For O96 and O66, PCR was also performed with primers based on hldD and downstream sequences of the three identified rml genes, to exclude the possibility that non-contiguous fragments had been cloned. Fragments of the size expected were generated from both strains and partial sequencing confirmed the gene order observed for the cloned fragments.
The location of rml gene sets in O antigen gene clusters
The location of the rml gene set in the O antigen gene cluster was also determined for each strain by PCR amplification of the region between rmlB and hldD, followed by full or partial sequencing of these PCR products. The O12, O14, O19, O66 and O151 rml gene sets were located downstream of hldD with a short intergenic sequence as for O6. For O11, O96 and O54, there is a wzz gene between the hldD and the rml gene sets, each with considerable similarity to wzz of Vibrio anguillarum (AF025396).
For O149 and O146 there is approximately 5 kb and 10 kb DNA respectively between hldD and rmlB. Sau3A digestion of the O149 PCR product, followed by cloning and sequencing some of the fragments, showed that this region contains genes that appear to be involved in nucleotide sugar pathways. One ORF has similarity to UDP-GlcNAc-2-epimerase genes and another to sugar dehydrogenases. Partial sequencing of the O146 PCR product revealed a 996 bp wzz gene immediately upstream of rmlB and three ORFs that are 5467 % identical to wbfA, wbfB and wbfE genes of the V. cholerae O139 CPS/O antigen gene cluster at protein level. These three genes of no assigned function are also present between wzz and hldD genes at the 5' end of the O antigen gene clusters of O139, O22, O69 and O141 gene clusters (Bik et al., 1996). The corresponding region for O146 also contains an ORF that is 39 % identical to insertion element IS1358 found in V. cholerae O139 and a number of serogroups (Bik et al., 1996
; Stroeher et al., 1998
).
The chimeric structure of rml gene sets in V. cholerae
The rml genes in V. cholerae are always in the order rmlB, rmlA, rmlD, rmlC, although in two cases only three are present. The 11 rml gene sequences were compared with each other, and with those in databases. All four rml genes of O146 and O149 are considerably divergent from those of any other strain, but the remaining nine strains have chimeric structures, with part of the sequence highly similar, but part very divergent. The level of similarity decreases from 5' to 3' end and the change from conserved to divergent sequence was found to be abrupt and at different positions for different strains (Fig. 3).
|
|
|
The O146 and O149 rml genes in the middle of the gene cluster
The O146 and O149 gene clusters were unusual in that the rml gene sets were not at the end of the gene clusters. They also lacked the chimeric structures as levels of identity were low for all rml genes. For O146, the levels were 69·472·2 %, 67·373·8 %, 45·967·9 % and 56·563·3 % for rmlB, rmlA, rmlD and rmlC, respectively in pairwise comparisons with the other 10 strains sequenced in this study, and for O149 the levels were 67·169·3 %, 45·660·2 %, 64·675·2 % and 59·562·6 % respectively. BLAST searches did not find close counterparts in other species.
Possible role of rml genes in lateral transfer of V. cholerae O antigen gene clusters
Nine of the 11 rml gene sets are related in that the sequences at the 5' end are very similar, although the sequences at the 3' end are divergent and strain-specific (Fig. 3). The changes in level of similarity coincided with changes in GC content. The differences in GC content strongly indicate that the two parts of the rml gene sets of these strains are of different origins, and that each rml gene set arose by recombination. The GC content of the conserved 5' parts of the rml gene set is very close to that of the V. cholerae genome (4648 %) (Bik et al., 1995
), and we suggest that this DNA has been in the species for a long time. In contrast, the 3' divergent segments of the rml gene sets are of lower GC content, suggesting derivation from other species. Moreover the low level of similarity among them suggests that they were derived from divergent sources. For some O antigens (for example O66), the extremely low GC content indicates that the genes have only recently been acquired by V. cholerae, as after lateral transfer the GC content will drift towards that of V. cholerae (Lawrence & Ochman, 1997
).
We can compare the above situation with that found in O antigen gene clusters of S. enterica (Li & Reeves, 2000), where the rml gene sets are also at the 5' end of the O antigen gene cluster. In this case also the 5' ends of the rml gene sets were generally similar, while genes at the 3' end were generally very different. S. enterica has a well-defined subspecies structure, with housekeeping genes generally having subspecies-specific sequences. It was found that not only were the 5' ends of the rml gene clusters very similar, but the variation was subspecies-specific. The pattern and level of variation between the 5' rml genes for different subspecies was similar to that for housekeeping genes, suggesting a similar evolutionary history, with this part of the rml gene set diverging with the subspecies. The 3' end of these rml gene sets was generally extremely divergent, but for strains with the same O antigen it was near identical and thus O antigen-specific.
The situation then is that for both S. enterica and V. cholerae the 5' end of the rml gene set has all the characteristics of DNA of housekeeping genes, whereas the 3' end has the characteristics of DNA of genes that have arrived by lateral transfer. The O antigen gene clusters are thought to transfer quite commonly between strains in each species, including transfer between subspecies for S. enterica. This transfer within a species was generally thought to occur by homologous recombination in flanking DNA (Reeves & Wang, 2002), but the sequence data suggest that the rml genes at the 5' end of the gene clusters can also be involved. In effect when a rhamnose-containing O antigen transfers to a strain which itself had a (different) rhamnose-containing O antigen, then the recombination site at the 5' end can be within the rml gene set. It seems clear that in both species this happens in a reasonable proportion of such cases. Further it seems clear that in both species, there is a host form of the rml genes. Each time a rhamnose-containing O antigen gene cluster is transferred in this way, a part of the rml gene set may, by recombination, be replaced by the host form of the gene. There is probably a minor advantage in the host form with its host pattern of codon usage, and over time the gene cluster loses all but the 3' end of the 3' gene of the rml gene set as present in the gene cluster when it arrived in the species. On this hypothesis the often small segment of low GC content DNA at the 3' end is all that has remained with the central serogroup-specific region during subsequent transfer within the species. Recombination downstream of the rml gene is generally not possible as the genes are different. The reason for the remnant of original rml DNA that seems to be present in all rhamnose-containing O antigen gene clusters is presumably that recombination is unlikely at the very junction of homologous and non-homologous DNA. The V. cholerae data discussed in this paper can be interpreted in the same way as originally done for S. enterica, but because V. cholerae has a much higher frequency of recombination, with a non-clonal or weakly clonal population structure (Byun et al., 1999
), it will lack a subspecies structure, so we cannot look for such a correlation.
The situation for the Streptococcus pneumoniae CPS gene clusters is interesting in yet another way, because in this case the rml gene sets are located at the 3' end of the CPS gene clusters (Boulnois & Jann, 1989; Coffey et al., 1998
; Jiang et al., 2001
; Morona et al., 1997
; Munoz et al., 1997
). Comparison of Strep. pneumoniae CPS rml gene sets (Coffey et al., 1998
; Jiang et al., 2001
; Morona et al., 1997
) revealed a gradient in the level of the variation among the genes. In this case the higher level of variation is at the 5' end of the rml gene sets, which for Strep. pneumoniae CPS gene clusters is the end adjacent to the central region that differs among the capsule types. Again the explanation is that the highly variable segments of the rml gene sets are proximal to the specificity-determining regions and have been associated with them since acquisition by lateral transfer, whereas the more distal part with much less variation has undergone recombination with genes resident in the species for a much longer time.
The O146 and O149 rml gene sets are in the middle of their gene clusters, and thus not able to be involved in mediating recombination between O antigen gene clusters as proposed for the V. cholerae strains and S. enterica described above. The fact that all genes in these two gene sets are different from each other and from the O6-related rml gene sets supports the role of recombination in generating the patterns observed for the O6-related rml gene sets. The O146 and O149 rml gene sets are most likely directly descended from the genes transferred from other species along with serogroup-specific genes.
Origin of the V. cholerae rml genes
We constructed gene trees based on the currently available rml gene sequences from different species (Fig. 6). For each rml gene, those strains with a gene that is closely related to the O6 gene are grouped in one branch. However for strains with chimeric rml gene sets, the genes at the 3' of the junction (Fig. 6
, O54_3', for example), and all four rml genes of O146 and O149, are not only divergent, but sometimes group with those of other species. This presumably reflects their various origins, as they are thought to represent the genes as they first entered V. cholerae by lateral gene transfer. However none was found to be close enough to a gene from another species to indicate a specific origin. This might simply be due to the limited number of rml gene sequences available so far, or changes accumulated over time in V. cholerae may prevent identification of the original host. It is also interesting that, with the exception of the O66 rmlD gene, their branches are among those of the families Enterobacteriaceae (e.g. Salmonella enterica, E. coli and Serratia marcescens) or Pasteurellaceae (e.g. Actinobacillus actinomycetemcomitans). These two families, like the Vibrionaceae, belong to subgroup three of the
subclass of the Proteobacteria (superfamily I) and have close evolutionary relationships (Logan, 1994
), suggesting that most V. cholerae O antigen gene clusters may have been captured from related species.
|
There is generally a sharp boundary within the rml gene set, between DNA of high and low levels of similarity (see Fig. 3). The nine V. cholerae gene clusters that fit this pattern range from O96, in which only rmlB has DNA resembling that of housekeeping genes, to O11 and O6, in which all but a small part of rmlC has these characteristics. As for S. enterica, the DNA of the more divergent segment has generally lower GC content and in this resembles most other genes of the CPS and O antigen gene clusters.
The DNA at the ends of the clusters has presumably been in the species for a long time. Although the rml genes are part of the O antigen gene cluster, they are at one end so that movement of the gene cluster within the species can involve recombination within the rml gene set, if both donor and recipient have a rhamnose-containing structure. We propose that this leads over time to substitution of the host form for the incoming form for much of the rml gene set. The presence of rml genes at the end of the gene cluster probably greatly facilitates movement of an incoming gene cluster to the normal locus for O antigen or CPS gene clusters of that species, and so provides an advantage to such gene clusters in getting established within a species, which could account for the finding that most rml genes are found at the end of the gene cluster, and that within any species they are at the same end.
The two rml gene sets of V. cholerae that are located at the centre of the gene cluster are of low GC content throughout and do not have any DNA similar in sequence to that in any of the other rml gene sets. These genes fit the pattern for other O antigen and CPS genes in being of low GC and are presumably unable to participate in the recombination events that affect the majority of the rml gene clusters, as that would lead to loss of the genes upstream of the rml genes. The properties of these internal rml genes sets, not seen in S. enterica or Strep. pneumoniae, further supports the hypothesis discussed above.
![]() |
ACKNOWLEDGEMENTS |
---|
![]() |
REFERENCES |
---|
![]() ![]() ![]() ![]() ![]() ![]() |
---|
Bik, E. M., Gouw, R. D. & Mooi, F. R. (1996). DNA fingerprinting of Vibrio cholerae strains with a novel insertion sequence element: a tool to identify epidemic strains. J Clin Microbiol 34, 14531461.[Abstract]
Boulnois, G. J. & Jann, K. (1989). Bacterial polysaccharide capsule synthesis, export and evolution of structural diversity. Mol Microbiol 3, 18191823.[Medline]
Byun, R., Elbourne, L. D. H., Lan, R. & Reeves, P. R. (1999). Evolutionary relationships of the pathogenic clones of Vibrio cholerae by sequence analysis of four house keeping genes. Infect Immun 67, 11161124.
Clarke, B. R. & Whitfield, C. (1992). Molecular cloning of the rfb region of Klebsiella pneumoniae serotype O1 : K20: the rfb gene cluster is responsible for synthesis of the D-galactan I O polysaccharide. J Bacteriol 174, 46144621.[Abstract]
Coffey, T. J., Enright, M. C., Daniels, M., Wilkinson, P., Berron, S., Fenoll, A. & Spratt, B. G. (1998). Serotype 19A variants of the Spanish serotype 23F multiresistant clone of Streptococcus pneumoniae. Microb Drug Resist 4, 5155.[Medline]
Comstock, L. E., Maneval, D., Jr, Panigrahi, P., Joseph, A., Levine, M. M., Kaper, J. B., Morris, J. G., Jr & Johnson, J. A. (1995). The capsule and O antigen in Vibrio cholerae 0139 Bengal are associated with a genetic region not present in Vibrio cholerae O1. Infect Immun 63, 317323.[Abstract]
Feng, D. & Doolittle, R. F. (1986). Progressive sequence alignment as a prerequisite to correct phylogenetics trees. J Mol Evol 25, 351360.
Franco, V. A., Liu, D. & Reeves, P. R. (1998). The Wzz (Cld) protein in Escherichia coli: amino acid sequence variation determines O antigen chain length specificity. J Bacteriol 180, 26702675.
Hobbs, M. & Reeves, P. R. (1994). The JUMPstart sequence: a 39 bp element common to several polysaccharide gene clusters. Mol Microbiol 12, 855856.[Medline]
Jekel, M. & Wackernagel, W. (1994). Location of the endA gene coding for endonuclease I on the physical map of the Escherichia coli K-12 chromosome. J Bacteriol 176, 15501551.[Medline]
Jiang, X. M., Neal, B., Santiago, F., Lee, S. J., Romana, L. K. & Reeves, P. R. (1991). Structure and sequence of the rfb (O antigen) gene cluster of Salmonella serovar Typhimurium (strain LT2). Mol Microbiol 5, 695713.[Medline]
Jiang, S.-M., Wang, L. & Reeves, P. R. (2001). Molecular characterization of Streptococcus pneumoniae type 4, 6B, 8, and 18C capsular polysacharide gene clusters. Infect Immun 69, 12441255.
Kondo, S., Kawamata, Y., Sano, Y., Iguchi, T. & Hisatune, K. (1997). A chemical study of the sugar of the polysaccharide portion of lipopolysaccharides isolated from Vibrio cholerae non O1 from O2 to O155. Syst Appl Microbiol 20, 111.
Lan, R. & Reeves, P. R. (1998). Recombination between rRNA operons created most of the ribotype variation observed in the seventh pandemic clone of Vibrio cholerae. Microbiology 144, 12131221.[Abstract]
Lawrence, J. G. & Ochman, H. (1997). Amelioration of bacterial genomes: rates of change and exchange. J Mol Evol 44, 383397.[Medline]
Li, Q. & Reeves, P. R. (2000). Genetic variation of dTDP-L-rhamnose pathway genes in Salmonella enterica. Microbiology 146, 22912307.
Liu, D. & Reeves, P. R. (1994). Escherichia coli regains its O antigen. Microbiology 140, 4957.[Abstract]
Logan, N. A. (1994). Bacterial Systematics, 1st edn. London: Blackwell Scientific Publications.
Maloy, S. R. & Nunn, W. D. (1981). Selection for loss of tetracycline resistance by Escherichia coli. J Bacteriol 145, 11101111.[Medline]
Manning, P. A., Stroeher, U. H. & Morona, R. (1994). Molecular basis for O-antigen biosynthesis in Vibrio cholerae O1: Ogawa-Inaba switching. In Vibrio cholerae and Cholera, Molecular to Global Perspectives, pp. 7779. Edited by I. K. Wachsmuth, P. A. Blake & Ø. Olsvik. Washington, DC: American Society for Microbiology.
Mitchison, M., Bulach, D. M., Vinh, T., Rajakumar, K., Faine, S. & Adler, B. (1997). Identification and characterization of the dTDP-rhamnose biosynthesis and transfer genes of the lipopolysaccharide related rfb locus in Leptospira interrogans serovar Copenhageni. J Bacteriol 179, 12621267.[Abstract]
Morona, J. K., Morona, R. M. & Paton, J. C. (1997). Characterization of the locus encoding the Streptococcus pneumoniae type 19F capsular polysaccharide biosynthetic pathway. Mol Microbiol 23, 751763.[Medline]
Munoz, R., Mollerach, M., Lopez, R. & Garcia, E. (1997). Molecular organization of the genes required for the synthesis of type 1 capsular polysaccharide of Streptococcus pneumoniae: formation of binary encapsulated pneumococci and identification of cryptic dTDP-rhamnose biosynthesis genes. Mol Microbiol 25, 7992.[Medline]
Reeves, P. R. (1993). Evolution of Salmonella O antigen variation by interspecific gene transfer on a large scale. Trends Genet 9, 1722.[CrossRef][Medline]
Reeves, P. R. (1997). Specialized clones and lateral transfer in pathogens. In Ecology of Pathogenic Bacteria: Molecular and Evolutionary Aspects, pp. 237254. Edited by B. A. M. van der Zeijst, W. P. M. Hoekstra, J. D. A. van Embden & A. J. W. van Alphen. Amsterdam: Elsevier.
Reeves, P. R. & Wang, L. (2002). Genomic organization of LPS-specific loci. In Pathogenicity Islands and the Evolution of Microbes (Current Topics in Microbiology and Immunology, vol. 1), pp. 109135. Edited by J. Hacker, J. B. Kaper, W. Compans & M. Cooper: Berlin: Springer.
Reeves, P. R., Farnell, L. & Lan, R. (1994). MULTICOMP: a program for preparing sequence data for phylogenetic analysis. Comput Appl Biosci 10, 281284.[Abstract]
Saitou, N. & Nei, M. (1987). The neighbor-joining method: a new method for reconstructing phylogenetic trees. Mol Biol Evol 4, 406425.[Abstract]
Sambrook, J., Fritsch, E. F. & Maniatis, T. (1989). Molecular Cloning: a Laboratory Manual. Cold Spring Harbor, NY: Cold Spring Harbor Laboratory.
Schmidt, G., Jann, B. & Jann, K. (1974). Genetic and immunochemical studies on Escherichia coli O14 : K7 : H-. Eur J Biochem 42, 303309.[Medline]
Selander, R. K., Li, J. & Nelson, K. (1996). Evolutionary genetics of Salmonella enterica. In Escherichia coli and Salmonella Cellular and Molecular Biology, 2nd edn, pp. 26912707. Edited by F. C. Neidhardt and others. Washington, DC: American Society for Microbiology.
Stevenson, G., Neal, B., Liu, D., Hobbs, M., Packer, N. H., Batley, M., Redmond, J. W., Lindquist, L. & Reeves, P. R. (1994). Structure of the O-antigen of E. coli K-12 and the sequence of its rfb gene cluster. J Bacteriol 176, 41444156.[Abstract]
Stroeher, U. H., Jedani, K. E. & Manning, P. A. (1998). Genetic organization of the regions associated with surface polysaccharide synthesis in Vibrio cholerae O1, O139 and Vibrio anguillarum O1 and O2: a review. Gene 223, 269280.[CrossRef][Medline]
Tarr, P. I., Schoening, L. M., Yea, Y. L., Ward, T. R., Jelacic, S. & Whittam, T. S. (2000). Acquisition of the rfb-gnd cluster in evolution of Escherichia coli O55 and O157. J Bacteriol 182, 61836191.
Wertman, K. F., Wyman, A. R. & Botstein, D. (1986). Host/vector interactions which affect the viability of recombinant phage lambda clones. Gene 49, 253262.[CrossRef][Medline]
Wilkinson, R. G., Gemski, P., Stocker, J. & Stocker, B. A. D. (1972). Non-smooth mutants of Salmonella typhimurium: differentiation by phage sensitive and genetic mapping. J Gen Microbiol 70, 527554.[Medline]
Yoshida, Y., Nakano, Y., Yamashita, Y. & Koga, T. (1998). Identification of a genetic locus essential for serotype b-specific antigen synthesis in Actinobacillus actinomycetemcomitans. Infect Immun 66, 107114.
Received 1 April 2003;
revised 9 June 2003;
accepted 9 June 2003.