Microbial Pathogenesis Unit, Department of Microbiology and Immunology, University of Adelaide, Adelaide, South Australia 50051
Author for correspondence: Paul A. Manning. Tel: +1 781 839 4000. Fax: +1 781 839 4500. e-mail: paul.manning{at}astrazeneca.com
![]() |
ABSTRACT |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
Keywords: gene capture, site-specific recombination, repetitive sequences, evolution
Abbreviations: 59-be, 59-base element; 5-CS, 5 conserved sequence; VCCI, Vibrio cholerae chromosomal integron; VCR, Vibrio cholerae repeat
The GenBank accession numbers for the sequences determined in this work are AF025662, AF055586, X64097, AF179593 and AF179596.
a Present address: AstraZeneca R&D Boston, 35 Gatehouse Drive, Waltham, MA 02451, USA.
![]() |
INTRODUCTION |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
Nucleotide sequencing of the original clone (Barker et al., 1994 ) revealed that each individual gene, with two minor exceptions, was oriented in the same direction and flanked by directly repeated copies of VCR. Both exceptions encoded small operons. This pattern is reminiscent of antibiotic-resistance gene cassette arrays of integrons, as recognized by Recchia & Hall (1995
, 1997
) who observed that the outer ends of the VCR consensus are related to the core and inverse core sites of 59-base elements (59-be) found in integron cassette arrays. Integron is a term applied to a genetic unit encoding an integrase gene intI, and immediately upstream, an attachment site attI with the core site, G
TTRRRY, defining its 3' end. IntI-mediated site-specific recombination occurs at the nick site between GT residues of the core sites of attI and the 59-be in gene cassettes (reviewed by Collis & Hall, 1995
; Recchia & Hall, 1995
, 1997
). The mobile elements termed gene cassettes, each of which carries a 59-be containing a core site, are defined as the region between two successive core nick sites within an array or as closed circles containing a single 59-be (Collis et al., 1993
).
We and others have recently reported the presence in V. cholerae of a new integrase (Clark et al., 1997 ; Manning et al., 1999
; Mazel et al., 1998
), homologous to the IntI family of integrases. It is divergently transcribed from the genes contained within the VCR region, in the manner of antibiotic-resistant integrons. The VCR-associated region was the first example of such a chromosomal locus containing all the components of an integron and was dubbed the Vibrio cholerae mega-integron (Clark et al., 1997
), but will now be referred to as the Vibrio cholerae chromosomal integron (VCCI). The sequence of an intI4 was described along with the presence of VCR cassettes in Vibrio metschnikovii and the ability of a VCR cassette to be integrated into an intI1 integron (Mazel et al., 1998
). In this study we present a general means of cloning the ends of this and similar arrays, and define the extent of the region. The extent of the region is defined and further evidence of its chromosomal variation is shown using both Southern blot and DNA sequence analysis in a variety of Vibrio spp. isolates. Together with evidence demonstrating that VCRs act as 59-bes in recombination assays, we show that the region functions as a new class of integron.
![]() |
METHODS |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
PCR.
Amplification using Amplitaq DNA polymerase (Hoffman-La Roche) was carried out by standard protocols with the oligo-deoxynucleotide primers 343 (5'-GTGGTTDCGGTTGTTGTGTG-3'), 922 (5'-CCCCTTAGGCGGGCGTTA-3'), 923 (5'-CCCTCTTGAGGCGTTTGTTA-3'), 2355 (5'-TAACGCCCGCCTAAGGGGCT-3') and 2583 (5'-GTACAACCGGTAACTTTCGTTCTAG-3').
DNA sequencing procedures.
The nucleotide sequences from strains 569B, O134, H1 and V800 were obtained using either Applied Biosystems 373 or 377 automated DNA sequencers. Dye-labelled terminators were used in the sequencing reactions. The accession numbers for the individual sequences are listed in the legend to Fig. 3.
|
Mating-out assay for site-specific recombination.
Donor strains were constructed by introducing all three plasmids at once into Escherichia coli S17-1 by electroporation. These were mated with E. coli DH5 as previously described (Hansson et al., 1997
) and plated on agar plates supplemented with nalidixic acid and chloramphenicol or nalidixic acid and kanamycin.
![]() |
RESULTS |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
|
To confirm that the VCR elements are in fact all direct repeats, as with all 59-bes studied so far, single-primer PCR was performed with either oligo-deoxynucleotide 922 or 923 using DNA from a variety of O1 and non-O1 serotype strains. The inability to obtain a PCR product with the individual primers suggests that in each case the VCR sequences are present as direct repeats within the arrays.
Identification of the 5 restriction fragment flanking the VCR region by Southern hybridization
The conserved nature of the VCR sequence and the apparent organization into one contiguous region (Barker et al., 1994 ) provided the means of identifying sequences immediately flanking this region. The VCR sequences contain four 6 bp restriction sites: HincII, KspI, MluI and Bsu36I, which are conserved in 28, 26, 21 and 28, respectively, of the 32 VCR elements identified and sequenced. Considering that the mean distance between these VCRs is 0·7 kb, the size of these restriction fragments from within the array would be expected on average to be significantly smaller than those from the immediate region flanking the array. This is particularly true of KspI as it has a G+C content of 100 mol%, whereas that of the V. cholerae genome is 4749 mol% (Baumann et al., 1984
). Genomes with a similar G+C content to that of V. cholerae, such as E. coli, have a mean KspI restriction fragment size of greater than 20 kb. Thus, the 5' restriction fragment of the array could be expected to be large and carrying only the 5' part of the first VCR. Using each enzyme separately, the genomic DNA from a variety of V. cholerae non-O1, and El Tor and Classical O1 serotype strains was digested and probed with oligo-deoxynucleotide 923, which is derived from the 5' end of the VCR. In Fig. 1
, hybridization of the probe to the 5' end of the VCR in chromosomal KspI digests reveals a series of bands, which we have divided into two groups of bands one corresponding to fragments above 7 kb and the other to fragments below 5 kb. In the range 4·40·72 kb there is a ladder of more than 20 discrete bands in each track, but when higher percentage agarose gels were used, numerous bands down to 0·3 kb were observed (data not shown), each probably corresponding to an individual VCR cassette. Throughout the size range there are a number of conserved restriction fragments among the O139 and all of the O1 serotype strains except for C31, which is known to have a major deletion in this region of the chromosome (Barker et al., 1994
; Fig. 2
).
|
Cloning of the V. cholerae integrase gene intI4 and defining the class 4 integron
After identifying the potential 5' end of the VCR region, the presence of an integrase immediately upstream of the first was sought. Genomic DNAs of V. cholerae strains 569B and H1 were digested with KspI and fragments of approximately 20 kb were isolated by agarose-gel purification and further digested with 20 different six-base recognition restriction enzymes found within the polylinker of pBluescript KS. A bank for each strain was constructed from 20 separate chromosomal digests ligated to pBluescript KS cut with the corresponding enzyme. Screening each ligation by vector-anchored PCR amplification with oligo-deoxynucleotide 923 and pBluescript reverse primer revealed a 1 kb product in the ClaI ligations with both 569B and H1, with an additional 2 kb fragment for H1. Both fragments were cloned. Oligo-deoxynucleotide 2583 probe, derived from the 569B 1 kb clone sequence, hybridized to a cloned 2·24 kb HindIII fragment from a 23 kb 569B subgenomic library. After sequencing, an integrase identical to IntI4 (Mazel et al., 1998 ) of the IntI family was found in a transcriptionally divergent direction to the first VCR cassette in 569B and the first three cassettes in H1. This is characteristic of integrons. The two sequenced VCRs of H1 do not possess KspI sites, thus explaining the two amplicons as well as some of the potential 5' array end fragments still hybridizing with oligo-deoxynucleotide 343. Comparison of intI4 from 569B and H1 showed nucleotide sequence identity from the ClaI site within the gene to 218 nt upstream of the intI4 initiation site. The point of divergence appeared to be the 5' boundary of the first cassette of the array. PCR and sequence analyses of V. cholerae O134 and V. mimicus using primers 2583 and 2355 were conducted. The DNA of strain O134 was 98% identical up to the same point of divergence in both sequences. The point of divergence from the 5' conserved sequence (5'-CS) in all three V. cholerae strains immediately follows 8 bp that conform to the core site consensus (see Fig. 4
), thus defining the 3' end of the attI4 site and thereby defining the class 4 integron of V. cholerae. Southern hybridization to the intI4 probes, oligo-deoxynucleotides 2593 and 2594, shows that the strains surveyed for array 5' flanking fragments can be divided into two groups with a KspI fragment of either 20 kb or 13·5 kb in size (Fig. 2
). Furthermore, in all but one instance the largest VCR-hybridizing Bsu36I fragments contained the 5' end of the array (data not shown).
|
Mapping the extent of the VCR region in 569B
In addition to the lipocalin-encoding gene vlpA, which we have cloned and sequenced (Barker & Manning, 1997 ), a second copy was identified on a 4·5 kb BamHI fragment, which contains an array of seven VCRs. Southern data, probing with oligo-deoxynucleotide 922, indicated the presence of three KpnI restriction fragments each with a similar size to the whole bacteriophage lambda genome. A 43 kb fragment was cloned into the cosmid vector c2XMCS (Reilly & Silva, 1993
) from a KpnI genomic library and screened by PCR with oligo-deoxynucleotide 922 and 923 primers. The 1·3 kb and 6·2 kb KpnI/SacI ends of the insert were subcloned and sequenced (Fig. 3
). The remainder of the cosmid insert had VCR-specific signals in a number of different digests in a Southern hybridization probed with 922, suggesting a continuous VCR cassette array (data not shown). Southern blots of 569B genomic DNA using both single and double restriction enzyme combinations were probed with oligo-deoxynucleotide 923, then stripped and reprobed with the probes indicated in Fig. 3
. Combining these sources, a restriction map of the region showing all the KpnI fragments hybridizing with VCRs and the position of sequenced regions was constructed (Fig. 3
). The first two KpnI fragments could not be linked. The two regions of the cassette array total 120 kb of VCR-containing DNA in the V. cholerae 569B chromosome.
Analysis of gene cassettes
Nucleotide sequence analysis and examination of the encoded ORFs showed that the average cassette of 693 nt generally contained very little non-coding DNA in addition to the VCR (Fig. 3). Of the 28 cassettes with an identifiable ORF in the same orientation as the VCR, 20 had insufficient space for a promoter between the 5' end of the VCR and the translational initiation site. Typical of 59-bes (Reilly & Silva, 1993
), most ORFs finish within 20 nt of a VCR, and in fact, many use the TAA of the inverse core as the stop codon. We have given ORFs within cassettes that do not have clear database homologues the name Vco for Vibrio cassette ORF. The associated numbers are the same as previously published ORF numbers where possible; otherwise they have been numbered in the order in which they appeared in the NCBI database. Subsequently identified vco homologues with greater than 95% identity are numbered in the order of discovery; for example, vco13.2.
An unusual feature is the presence of ORFs transcribed in the opposite direction to that of the array such as the gene pairs vco3AB, vco21AB and vco27AB, as well as vco24B which is convergent with vco24A (Fig. 3). All of these ORFs have sufficient space for a promoter between the proposed start and the VCR upstream of it. IS1360 is a new IS element with 77% nucleotide identity with IS5 (accession no. X13668) and was found inserted into the first complete cassette on the cosmid clone C169. IS1359 apparently disrupts ORF23, the gene after the last cassette (Fig. 3
), and is a new member of the IS3 family with 64% identity at the nucleotide level to the functional IS911 of Shigella dysenteriae. IS1359 appears to mark the end of the chromosomal VCR array, as the last 4·8 kb of the cosmid is devoid of VCR elements.
mccF1, a new cassette-associated ORF of 344 residues, has 55% identity at the nucleotide level over its full length with the E. coli microcin C7 immunity protein MccF (Gonzalez-Pastor et al., 1995 ). A second homologue was found to lie on a 7·3 kb BamHI fragment within the second KpnI fragment by probing with mccF1 (data not shown). The vco13 cassette is yet another with homologues: the vco20 cassette is 67% identical. The vlpA and vlpa2 genes were nearly identical.
![]() |
DISCUSSION |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
The HindIII and KspI restriction polymorphisms that were detected imply that the VCR array in Vibrio is plastic and potentially provides a rapid means of demonstrating whether strains are closely related. The biotype El Tor strains are considered to be a clonal group for some loci (Byun et al., 1999 ). However, no two strains had identical Southern blot profiles, although most did have many fragments in common. The absence of common fragments in strain C31 confirms the previous suggestions that it has undergone a major deletion within the region of the chromosome encoding the mrhAB locus (Barker et al., 1994
; Van Dongen et al., 1987
).
Restriction enzyme sites within the VCR elements also provided a means of identifying the 5'-CS of the region in many strains. The approach presented here for cloning the 5' end of the array may be generally applicable to cloning the ends of other large arrays of highly conserved repetitive elements. Cosmid C169 appears to contain the 3' end of the array since 8·8 kb of VCR-free DNA is located downstream of its last VCR. Furthermore, the sequence up to the end of IS1359 of region E from 569B (Fig. 3) is nearly identical to a similar region devoid of VCR on the chromosome of the V. cholerae El Tor strain N16961 that has been sequenced by The Institute for Genomic Research (see http://www.tigr.org) (Fig. 4
). The sequence in 569B after IS1359 is not homologous to that of N16961, and preliminary PCR analysis indicates that El Tor and Classical strains differ at this point, suggesting that IS1359 may be a recombinational hotspot. The location of IS elements at the 3' ends of arrays is quite typical of other classes of integrons (Hall, 1995
).
The size of the integron array here is defined as approximately 120 kb, consistent with a previous estimate in which strain 569B was estimated by densitometry to have 90100 copies of VCR (Barker et al., 1994 ), suggesting that there could be in excess of 150 gene cassettes, vastly greater than the few genes detected in the average antibiotic-resistance integron. Recent data from the incomplete genome sequence of V. cholerae El Tor strain N16961 have identified 147 VCR elements within a 147 kb chromosomal integron (see http://www.tigr.org) confirming that this complexity is not restricted to Classical strains.
We have compared the organization of the sequenced regions of the 569B VCCI and the genes within them with the available data for N16961 (Fig. 4). It is immediately apparent that the spacing between the regions differs and that region B does not exist, at least in the same organization seen in 569B. In addition, there are more copies of IS1359, three in each chromosome of N16961 but only two associated with the VCCI in 569B. Also there are two copies of vlpA in the VCCI in 569B, but four copies in N16961, only one of which is identical and in the same position as 569B, namely that in region C. The additional copies all have a few mismatches within the gene (12, 13 and 13 bp). Furthermore, although both strains have two copies of mccF, the spacing differs. These data suggest that there is continual microevolution occurring within VCCI leading to gain, loss or duplication of genes, presumably all mediated by the specific integrase.
The comparison of strains 569B, H1 and O134 implies that the 5'-CS satisfies all of the definitional requirements for the fourth class of integron and also provides the first direct evidence that attI core-site-specific recombination typical of integrons has occurred in V. cholerae. Recently, it has been shown that a VCR cassette can be incorporated into the attI site of the class 1 integron on R388 (Mazel et al., 1998 ). Using a cointegration assay, we have been able to demonstrate that the VCR functions as a 59-be and that it can act as a substrate for site-specific recombination using IntI1 (unpublished data). The presence of IntI1 led to a 350-fold higher recombination rate, which was confirmed by sequence analysis to correspond to site-specific recombination events that are consistent with typical recombination involving a 59-be core, as has been shown previously for a 59-be (Collis et al., 1993
; Martinez & de La Cruz, 1990
).
There are three main differences between the VCCI and the plasmid-borne antibiotic-resistant integrons: size, mobility and homogeneity of the 59-be. The expression of the antibiotic-resistance cassettes within integrons derived solely from the Pant promoters located in the 5'-CS (Lesesque et al., 1994 ) seems to be one of the limitations on their size unless most of the array is silent. However, this arrangement would be inadequate within the VCCI. This is of particular concern since some 59-bes may act as weak terminators (Collis & Hall, 1995
) as has been shown for VCRs within region C (A. Barker,S. G. Williams & P. A. Manning, unpublished). This implies therefore that either large regions of the VCCI are silent or there are numerous internal promoters.
While the presence of promoters within VCRs is an intriguing possibility, this activity, as yet, has not been demonstrated. An alternative is that promoters are provided by some of the cassettes with enough space 5' of the start codon to encode a promoter, as has been shown for mrhAB (Barker et al., 1994 ). Clearly, the presence of four cassettes, with ORFs in the opposite orientation with respect to the array, and which has not been seen in other integrons, further indicates the necessity for endogenous promoters, and each of these cassettes has sufficient coding capacity. This ad hoc arrangement might well explain the need for multiple copies of genes, such as in the case of the promoterless vlpA, which is present in up to five copies in some strains (P. Kaewrakon, P. A. Manning and T. Focareta, unpublished)
It has been suggested (Mazel et al., 1998 ) that the high degree of homology of VCR elements indicates that each of the previously variant 59-be represents a single member of similar vast arrays found in many different bacterial genera. It is indeed likely that there are other chromosomal integrons and that they are an important means of gene capture and chromosomal building. However, there are a number of factors that may contribute to drift in 59-bes outside these integrons. For example, the permanent residency of the integron provides a stable genetic environment for co-evolution with the host not likely to be duplicated, as in the mobile plasmid-associated integrons.
To date, there appears to be a significant difference in the types of genes acquired by the VCCI, with the possible exception of mccF1 should it prove to act as a microcin-immunity protein. This difference is likely to reflect the antibiotic-resistance bias in sampling of the plasmid-based integrons examined, rather than any fundamental differences between the two systems. The pool for gene capture is likely to be the converse. That is, the VCCI samples genes from episomes passing through. For example, the large plasmid carrying mccF (Gonzalez-Pastor et al., 1995 ) might be the ancestral source for mccF1. In this way the VCCI captures useful genes and eliminates the burden of maintaining the episomes. It is tempting to speculate that the chromosomal integrons, at least within the genus Vibrio, act as a large gene pool for horizontal gene transfer between Vibrios. This could potentially occur via natural transformation of VCR cassette circles or by plasmid-integron intermediates such as that found recently in an El Tor isolate (Fablo et al., 1999
).
![]() |
ACKNOWLEDGEMENTS |
---|
![]() |
REFERENCES |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
Barker, A. & Manning, P. A. (1997). VlpA of Vibrio cholerae O1: the first bacterial member of the 2-microglobulin lipocalin superfamily. Microbiology 143, 1805-1813.[Abstract]
Barker, A., Clark, C. A. & Manning, P. A. (1994). Identification of VCR, a repeated sequence associated with a locus encoding a hemagglutinin of Vibrio cholerae O1. J Bacteriol 176, 5450-5458.[Abstract]
Baumann, P., Furniss, A. L. & Lee, J. V. (1984). Genus I. Vibrio Pacini 1854, 411AL. In Bergeys Manual of Systematic Bacteriology , pp. 518-538. Edited by N. R. Krieg & J. G. Holt. Baltimore, MD:Williams & Wilkins.
Byun, R., Elbourne, L. D. H., Lan, R. & Reeves, P. R. (1999). Evolutionary relationships of pathogenic clones of Vibrio cholerae by sequence analysis of four housekeeping genes. Infect Immun 67, 1116-1124.
Clark, C. A., Purins, L., Kaewrakon, P. & Manning, P. (1997). VCR repetitive sequence elements in the Vibrio cholerae chromosome constitute a mega-integron. Mol Microbiol 26, 1137-1138.[Medline]
Collis, C. M. & Hall, R. M. (1995). Expression of antibiotic resistance genes in the integrated cassettes of integrons. Antimicrob Agents Chemother 39, 155-162.[Abstract]
Collis, C. M., Grammaticopoulos, G., Briton, J., Stokes, H. W. & Hall, R. M. (1993). Site-specific insertion of gene cassettes into integrons. Mol Microbiol 9, 41-52.[Medline]
Fablo, V., Caratolli, A., Tosini, F., Pezzella, C., Dionisi, A. M. & Luzzi, I. (1999). Antibiotic resistance conferred by a conjugative plasmid and a class I integron in Vibrio cholerae O1 El Tor strains isolated in Albania and Italy. Antimicrob Agents Chemother 43, 693-696.
Franzon, V. L., Barker, A. & Manning, P. A. (1993). Nucleotide sequence encoding the mannose-fucose-resistant hemagglutinin of Vibrio cholerae O1 and construction of a mutant. Infect Immun 61, 3032-3037.[Abstract]
Gonzalez-Pastor, J. E., San Milan, J. L., Angeles Castilla, M. & Moreno, F. (1995). Structure and organization of plasmid genes required to produce the translation inhibitor microcin C7. J Bacteriol 177, 7131-7140.[Abstract]
Hall, R. M. (1995). Integrons: the movers and shakers of antibiotic resistance. Todays Life Sci 7, 38-44.
Hansson, K., Skold, O. & Sundstrom, L. (1997). Non-palindromic attI sites of integrons are capable of site-specific recombination with one another and with secondary targets. Mol Microbiol 26, 441-453.[Medline]
Higgins, C. F., Ames, G. F.-L., Barnes, W. M., Clement, J.-M. & Hofnung, M. (1982). A novel intercistronic regulatory element of prokaryotic operons. Nature 298, 760-762.[Medline]
Hulton, C. S. J., Higgins, C. F. & Sharp, P. M. (1991). ERIC sequences: a novel family of repetitive elements in the genomes of Escherichia coli, Salmonella typhimurium and other enterobacteria. Mol Microbiol 5, 825-834.[Medline]
Kang, J. H., Lee, J. H., Park, J. H., Huh, S. H. & Kong, I.-S. (1998). Cloning and identification of a phospholipase gene from Vibrio mimicus. Biochim Biophys Acta 1394, 85-89.[Medline]
Kim, G.-T., Lee, J.-Y., Huh, S. H., Yu, J.-H. & Kong, I.-S. (1997). Nucleotide sequence of the vmha gene encoding hemolysin from Vibrio mimicus. Biochim Biophys Acta 1360, 102-104.[Medline]
Kita-Tsukamoto, K., Ozayaizu, H., Nanba, K. & Simidu, U. (1993). Phylogenetic relationships of marine bacteria, mainly members of the Vibrionaceae, determined on the basis of 16S rRNA sequences. Int J Syst Bacteriol 43, 8-19.[Abstract]
Lesesque, C., Brassard, S., Lapointe, J. & Roy, P. H. (1994). Diversity and relative strength of tandem promoters for the antibiotic-resistance genes of several integrons. Gene 142, 49-54.[Medline]
Manning, P. A., Heuzenroeder, M. W., Yeadon, J., Leavesley, D. I., Reeves, P. R. & Rowley, D. (1986). Molecular cloning and expression in Escherichia coli K-12 of the O antigens of the Inaba and Ogawa serotypes of the Vibrio cholerae O1 lipopolysaccharides and their potential for vaccine development. Infect Immun 53, 272-277.[Medline]
Manning, P. A., Clark, C. A. & Focareta, T. (1999). Gene capture in Vibrio cholerae. Trends Microbiol 7, 93-95.[Medline]
Martinez, E. & de la Cruz, F. (1990). Genetic elements involved in Tn21 site-specific integration, a novel mechanism for the dissemination of antibiotic-resistance genes. EMBO J 9, 1275-1281.[Abstract]
Mazel, D., Dychinco, B., Webb, V. A. & Davies, J. (1998). A distinctive class of integron in the Vibrio cholerae genome. Science 280, 605-608.
Recchia, G. D. & Hall, R. M. (1995). Gene cassettes: a new class of mobile element. Microbiology 141, 3015-3027.[Medline]
Recchia, G. D. & Hall, R. M. (1997). Origins of the mobile gene cassettes found in integrons. Trends Microbiol 5, 388-394.
Reilly, J. D. & Silva, R. F. (1993). A double-cos-site vector containing a multiple cloning site flanked by T7 and T3 RNA polymerase promoters. Gene 124, 135-136.[Medline]
Sambrook, J., Fritsch, E. F. & Maniatis, T. (1989). Molecular Cloning: a Laboratory Manual, 2nd edn. Cold Spring Harbor, NY: Cold Spring Harbor Laboratory.
Shangkuan, Y.-H., Lin, H.-C. & Wang, T.-M. (1997). Diversity of DNA sequences among Vibrio cholerae O1 and non-O1 isolates detected by whole-cell repetitive element sequence-based polymerase chain reaction. J Appl Microbiol 82, 335-344.[Medline]
Southern, E. M. (1975). Detection of specific sequence among DNA fragments separated by gel electrophoresis. J Mol Biol 98, 503-517.[Medline]
Van Dongen, W. M. A. M., Vlerken, V. & De Graaf, F. K. (1987). Nucleotide sequence of a DNA fragment encoding a Vibrio cholerae haemagglutinin. Mol Gen (Life Sci Adv) 6, 8591.
Received 5 January 2000;
revised 29 June 2000;
accepted 6 July 2000.