Original Article

Structural Variation of the Superintegron in the Toxigenic Vibrio cholerae O1 El Tor*

GAO Yan^§,$, PANG Bo^§, WANG Hai Yin, ZHOU Hai Jian, CUI Zhi Gang, and KAN Biao^#

State Key Laboratory for Infectious Disease Prevention and Control, National Institute for Communicable Disease Control and Prevention, Chinese Center for Disease Control and Prevention, Beijing 102206, China

 

INTRODUCTION

ntegrons were first discovered in the characterization of multiple-resistance-encoding plasmids and transposons^[1-2]. The major components of an integron are an attC site (associated with gene cassette), intI gene (encoding an integrase), attI site, and a promoter^[3]. The integrase can integrate and excise gene cassettes by catalyzing recombination between attI and attC, and recombination between attCs^[4]. A distinct type of integron, superintegron (SI), was first identified in a Vibrio cholerae strain, with the characteristics of a large number of cassettes that are clustered, spaced by attC sites, and specific integrase^[5]. Comparison of SI structures among the Vibrio species and the different isolates within the same species may provide valuable data for the analyses of SI nature, gene flow, and evolution of SIs and hosts. It has been shown that a low number of cassette counterparts are shared among different Vibrio species, which suggests a wide range of species source for the entrapped genes and an active cassette assembly process^[6-8]. SI structures have been found in many bacterial genomes, including gammaproteobacteria, betaproteobacteria, and deltaproteobacteria, besides Vibrionaceae^[6-7]. SI is a potential gene capture system and may play a role in bacterial adaptation and evolution^[9]. Although the functions of most of the encoded genes in SI are unknown, some of the SI open reading frames (ORFs) encode adaptive functions including pathogenicity and antibiotic resistance determinants^[7,10-12].

Historically seven cholera pandemics have been recorded. Toxigenic V. cholerae serogroup O1 of biotype El Tor caused the seventh cholera pandemic since 1961^[13], and O139 cholera emerged in Bangladesh and India in 1992 and caused epidemics in Southeast Asia^[14]. In V. cholerae N16961, an SI is located in the small chromosome and carries 216 ORFs^[15]. attC is called the Vibrio cholerae repeat sequence (VCR) in the SI of V. cholerae^[6]. The comparative genome hybridization (CGH) and whole genome PCR scanning have suggested that the content variance of SIs in different lineages of V. cholerae is distinct^[16-17]. With three whole genome sequences of V. cholerae (including the toxigenic/notoxigenic O1 El Tor strains and the toxigenic O1 classical strain), substantial changes in the SI of these strains have been revealed^[18]. Studies based on PCR and hybridization have also revealed plastic structures of SIs between different serogroups^[7,19-21].

However, CGH and other PCR-based genotyping analyses have not revealed the structural details of the SI. In some studies, PCR was performed with primers based on the conservative sequence of VCR, and then the amplicons were analyzed with electrophoresis, Southern hybridization and sequencing to compare the ORF content of the SI in different V. cholerae strains^[19-21]. The ORFs arrangement in SIs, and deletion and insertion of the ORFs in SIs remain unclear. Many repeat sequences exist in SIs, which also make it impossible to assemble the SI sequence and to study the gene arrangement in SIs. PCR scanning^[17], which uses the overlapping amplicons to study the arrangement of genes, is valuable for the genome fragment assembly and especially for those containing many repeat sequences.

In this study, a detailed PCR scanning strategy combined with sequencing was used to analyze the strain-to-strain genetic organization variance of the SI in 60 toxigenic V. cholerae O1 El Tor strains during the seventh cholera pandemic in China. The structure of the SIs in the test strains was basically syntenic; however, diversity and decadal signatures were also observed, characterized by successive ORF deletion, ORF insertion, and a few potential ORF translocations. Homologous recombination based on repeat sequence and VCR played roles in the gene flows of SIs.

Materials and methods

Strains

From 1961 to 2008, three epidemiologically defined cholera epidemics occurred in China. In this study, 60 toxigenic O1 El Tor strains isolated in different epidemic periods and inter-epidemic periods in different geographic regions were selected for analysis (Figure 1). The details of experimental strains are shown in Table 1.

Figure 1. Temporal and geographic distribution of the 60 tested V. cholerae strains. The curve indicates the cholera cases reported in China between 1961 and 2008. The spots denote the number of strains: the small to big spots denote 1, 2, 3, and 5 strains, respectively.

Table 1. Characteristics of the Toxigenic O1 El Tor Strains Used in This Study

Note. ^*Strains with deletion of VCA0291-VCA0293; ^#strains with deletion of VCA0395-VCA0436; +minimum number of ORFs in the SI was uncertain, because some possible insertion fragments were not amplified and the exact ORFs could not be determined.

Preparation of Chromosomal DNA

All strains were grown in 5 mL Luria–Bertani broth to OD₆₀₀ of 0.8. Chromosomal DNA was prepared with NucleoSpin Tissue kit (Macherey–Nagel, Duren, Germany), according to the manufacturer’s protocol.

Primer Design

Genome sequence of strain N16961 was used as the template. From VCA0290 (gene coding is followed with N16961 genome), every two adjacent ORFs were amplified by a pair of primers designed in the two ORFs respectively. Every two adjacent amplicons overlapped with each other. More ORFs were included in one amplicon if no primers could be designed with this principle. Based on the published genome sequence of N16961^[15], 213 pairs of primers were designed using Oligo6.0 primer analysis software (Molecular Biology Insights, Cascade, CO, USA) (Supplemental Table). In these primers, 175, 32, five and one pairs of primers were used to amplify two, three, four and five adjacent ORFs, respectively. Primers ranged from 18 to 22 nt (most being 22 nt), and the average expected size of the amplicons was 885 bp. The maximum and minimum amplicon sizes obtained were 1998 and 153 bp, respectively. Overlaps between adjacent segments ranged from 1 to 625 bp, with the average size being 69 bp. When an amplification reaction failed, the primers flanking this negative fragment were used, and new primers locating the flanking positive regions were redesigned once the tentative PCR was still negative. If an amplicon with different size from strain N16961 was obtained, then the amplicon was sequenced and the ORF and VCR were determined.

PCR Scanning

PCR reactions were performed on a DNA Engine Tetrad^？ 2 Peltier Thermal Cycler (Bio-Rad, Hercules, CA, USA). N16961 was used as the positive control. Amplification reactions were carried out in a 20-μL volume using rTaq DNA polymerase system (TaKaRa, Dalian, China) according to the manufacturer’s instructions. PCR was performed under the following conditions: initial denaturation at 94 °C for 5 min, 30 cycles of denaturation at 94 °C for 30 s, annealing at 55-58 °C for 30 s, extension at 72 °C for 1 min; and a single final extension at 72 °C for 7 min. The PCR products were analyzed by electrophoresis through a 1% agarose gel in a Bio-Rad Wide Mini-Sub Cell GT system. Repeated PCR reactions were performed for the negative amplification reaction to assure the reliability of non-amplifiable samples.

Multiple Loci Variable Number of Tandem Repeats (VNTRs) Analysis (MLVA)

Five previously described VNTR markers (VC0147, VC0437, VC1650, VCA0171, and VCA0283) were used^[22]. The primers used for amplification of these five loci were the same as those reported initially^[22], except for VC0147, for which a new primer set (VC0147-F: 5’-CAA ACG CAG GAT GAA CCA-3’, VC0147-R: 5’-AAG AAG CCA GCG CCA ATA-3’) was designed to yield bigger PCR products, thus allowing better distinction from PCR products of other loci when analyzed by capillary separation. The PCR products were analyzed by capillary separation along with an internal size standard (GeneScana ROX-500 size standard, PerkinElmer Applied Biosystems, San Jose, CA, USA. ) on a PE Applied Biosystems ABI Prisma 3 730 instrument. The data were conserved as *.fsa files and processed by GeneMarker version 1.71 software (SoftGenetics LLC, State College, PA, USA). The size bins had an error range of ±0.5 bp. If any product sizes were situated outside this interval, an error message was returned. The products sizes were exported and transformed to allele profiles in Excel files, and were entered into BioNumerics software (Applied Maths, Sint-Martens-Latem, Belgium) as character values. Dendrograms were clustered and constructed by using the unweighted pair group method using arithmetic averages (UPGMA).

Pulsed-field Gel Electrophoresis (PFGE)

PFGE was performed according to the PulseNet standardized PFGE protocol for V. cholerae subtyping^[23]. Genomic DNA from all isolates was prepared in agarose plugs and digested with the restriction enzyme NotI, separated in a 1% agarose gel in 0.5× Tris-borate-EDTA at 14 °C using a CHEF-DRIII apparatus (Bio-Rad, Hercules, CA, USA). The pulse time ranged from 2 to 10 s for 13 h, and from 20 to 25 s for 6 h, both at 6 V/cm. After visualization, the PFGE patterns were analyzed using BioNumerics. The similarity between two patterns was expressed as a Dice coefficient^[24]. Dendrograms were clustered and constructed by using the UPGMA with a tolerance of 1.0%.

RESULTS

Structure Map of SIs in Test V.cholerae El Tor Strains

In total, 12 780 PCR scanning reactions covering the SIs of 60 toxigenic El Tor strains were performed, including N16961 as the control. All amplicons were obtained with N16961 genome DNA as the template. Of all the reactions, 11 070 were positive and 1 710 (13.4%) were negative, suggesting that there was not much variation among these strains. In the negative reactions, at least 97.4% (1 665/1 710) were successive in two or more overlapping amplicons (SI of N16961 as the reference). For each negative reaction, primers were redesigned in the positive flanking regions to determine whether the negative results were caused by deletion or new fragment insertion. In this way, ORFs which existed in the test strains but not in N16961 and deletions were obtained by PCR and sequencing (Table 2). However, for some regions in certain strains (e.g. fragments VCA0292-VCA0324 and VCA0327-VCA0329 in strain 7751), redesigned primers located in the positive flanking boundaries still failed to yield amplicons, which suggested that these regions were replaced by new fragments, but were too long to be amplified in our PCR tests. We did not obtain the sequence information for these regions. This group included strains 83535, 196153, 1961119, 62110, 2000180, and 78599.

Table 2. Sequence Analysis of Continuous Negative Results of PCR

Note. ^*and ^#as in Table 1.

Within the test strains, the PCR scanning results of nine test strains were similar to those of N16961, which indicated that 216 ORFs also existed in those strains (Supplemental Figure 1, from the bottom, lines 1-4 and 6-11). Only ORF deletion (compared to N16961) was detected in the SIs of the 45 test strains if translocation was not taken into account (Supplemental Figure S1, from the bottom, lines 12-55). The number of deleted ORFs ranged from 1 to 49, which indicated that 167-215 ORFs existed in those strains (Supplemental Figure S1, from the bottom, lines 12-55). In two test strains (78599 and 19612540), inserted fragments were obtained by PCR with primers located in boundary regions and sequenced, and were found to comprise 15 ORFs, which replaced the corresponding 15 ORFs in N16961. In the left four test strains (8212, 829, 19612533, and 7751), both ORF deletions and new fragment insertions were detected. A total of 140 ORFs in each were found in the SIs of strains 8 212, 829, and 19612533. The number of ORFs in strain 7751 was uncertain because of PCR failure of insertions in some SI regions. In summary, the number of ORFs in the SIs of the test strains ranged from 140 to 216 (Supplemental Figure S1 and Table 1). 

SI gene maps of the 60 test toxigenic El Tor strains were constructed, except for strains 7751, 83535, 196153, 1961119, 62110, 2000180, and 78599, due to undetermined insertions, based on the PCR scanning and sequencing (Supplemental Figure S1). Most of the variations in the SIs of the test strains clustered in the first two-thirds of the SI (Figure 2). The variation regions did not randomly distribute in the SI, which indicated that certain sites within the SI were hotspots for gene gain or loss. The structure of SIs in the test strains seemed to be syntenic, except for ORFs deletion and replacement.

Figure 2. Variation frequency of ORFs in the SIs of the 60 tested V. cholerae strains based on the result of PCR scanning, compared to N16961. Most of the variation regions clustered in the first two-thirds of the SI (5’ to 3’).

Clustering Analysis of SIs from Test Strains Based on Gene Content

Based on the gene presence and absence in each SI, a minimum spanning tree was reconstructed for the test strains (Figure 3). The undetermined insertion genes were ignored. We distinguished the test strains into three groups: 1960s, 1970s-1980s, and 1990s and after, according to the epidemic curves and years since the start of the seventh cholera pandemic. SIs with the same content were grouped together. Different SI clustering characteristics for each epidemic period were found. In the tree, the strains isolated in the 1990s and after were grouped closely, whereas the 1970s-1980s strains formed a large dispersed cluster. No distinct clustering was found even if these strains were divided into two decades. The strains isolated in the 1960s were much more dispersed than the others, suggesting their divergence in SI content.

Gene Deletion, Insertion and Uncertain Transfer in SIs of the Test V. cholerae El Tor Strains

VCA0292-0293 Fragment Insertion   No amplicons were obtained in 49 of the 60 test strains (81.7%) with primer pair 2F/3R (Supplemental Table), which detected VCA0292 and VCA0293. These strains spanned from 1961 to 2008, whereas all but one 1980s strain had these two genes. Amplicons of 1.4 kb were obtained in all of those 49 strains with primer pair 2F/3R, and sequencing confirmed the absence of a fragment spanning nt 310943-312311 (in N16961 genome), which included ORFs of VCA0292 and VCA0293 and the adjacent sequence (Figure 4a). These two genes are the closest genes to IntI4 of SI. attI^[10] was found in the upstream region of VCA0292. It has been reported that another integrase, IntI1, can catalyze recombination (attI×attC, attC×attC, and attI×attI), whereas recombination between attI and attC is the most efficient^[25]. Moreover, it has been shown that the deletion frequencies at which IntI1 and the integrase of Nitrosomonas europaea delete cassettes in the attI sites are low^[26-27]. Therefore, we deduced that the fragment of VCA0292-0293 was integrated into the 11 test strains compared to those 49 strains, rather than deleted. The integration was possibly a recombination between attI and VCR mediated by integrase.

 

Figure 3. Minimal spanning tree of 26 subtypes of SIs based on the presence and absence data of the ORFs in the SIs of 60 toxigenic O1 El Tor V. cholerae. Subtypes are indicated by circles, whose diameter increases as the number of strains increases. The numbers on the black lines which connect two circles denote the number of ORFs difference between the two connected SI subtypes. The smallest and the second smallest circles denote subtypes which were only represented by one or two SIs. The strain names in which the SIs existed are shown. The circles with capital letters denote subtypes in which three or more SIs were included. The different capital letters represent strains in which different SI subtypes existed: A, including 94068, 93014, 90-26, 1994400, 1993005, 1993050, 2004DD534, 84159, 2005125, 2005035, 2004152, 1998146, 64193, 1991225, 1992031, 1994A20, 2001058, 2002001, and 2005122; B, including 7925, 80954, 79192, 62048, and 65930; C, including 751130, 781079, and 661673; D, including 85079, 8788, 8593, 83101, GDL-ETV.760, 801298, 19861071, 1999001, and 83795.

All the strains except for 801290, which possessed VCA0292 and VCA0293, had an identical SI genetic structure (named 80s-SI). We also performed MLVA and PFGE for these test strains. Most of the 11 strains possessing VCA0292 and VCA0293 (10 and 9 strains, respectively, in these two analyses) clustered together (Supplemental Figures S2 and S3), which suggested that these strains have evolved from one common clone. Some divergence might have occurred, for example, in strain 801290, fragment VCA0362-0374 was probably deleted. Within this group, one exceptional strain, 1999001, had the same SI structure and similar MLVA and PFGE patterns, suggesting re-emergence of this clone in the 1990s.

VCA0395-0436 Deletion   In the majority (95.7%, 22/23) of the test strains isolated in 1990 and after, and three strains isolated before 1990 (19612540, 64193, and 84159), most of the ORFs between VCA0395 and VCA0436 could not be detected (Supplemental Figure S1, from the top, lines 6-30). A previous study also has shown that an ~24-kb fragment deletion occurred in some O1 El Tor and O139 strains^[20]. Fragments of 2.4 kb were obtained in those strains in our study with the primer pair A. (same to primer pair MGC86F/MGC128R in ref. 20). Sequencing revealed that 23 967 bp were deleted compared to the corresponding region in N16961. We called these SIs with absence of a 24-kb fragment 90s-SI. In N16961, 552-bp direct repeat sequences were detected in the both of the up and down stream flanking sequences of the 23 966-bp fragment; therefore, it is most likely that absence of this ~24-kb fragment was a deletion event caused by recombination between 552-bp repeats (Figure 4b). Additionally, the same deletion was also detected in three strains (19612540, 64193, and 84159) isolated before 1990 (Supplemental Figure S1).

In 96% (24/25) of the strains with 90s-SI amplicons could be obtained with primer pair 86 (detecting VCA0404 and VCA0405), although the result indicated that VCA0404 and VCA0405 were missing. No repeat copy of VCA0404 and VCA0405 was found in N16961 genomes. Therefore, we speculated these two genes translocated to other positions in the genomes of those strains. Similarly, some other dispersed translocations of genes within 90s-SIs were observed, such as VCA0396-0397 in strain 64193, VCA0406-0408 in 2002001, and VCA0363-0364 in 2000031. The absence of other scattered gene clusters was also found in several strains (Supplemental Figure S1).

In the minimum spanning tree of SIs, the 1990s strains with 90s-SI structure, which had the same or similar genetic components clustered together, and strains 64193, 84159, and 19612540 were also included (Figure 3). Most of the 1990s isolates clustered closely in MLVA and PFGE dendrogram trees (two clusters presented in Supplemental Figures S2 and S3), suggesting their discriminatory clones with other strains. Furthermore, strains 19612540 and 84159 had close similarity for PFGE patterns with the 1990s strains, and 84159 also had a similar MLVA pattern, suggesting the possible clone that appeared in the epidemics before 1990.

 

 

Figure 4. Schematic representation of sequence comparison between the tested strains and N16961. (a) Partial sequence comparison between strain with VCA0292-VCA0293 (N16961) and that without VCA0292-VCA0293 (LN1993050). The 1351-bp deleted sequence in LN1993050 included complete VCA0292, VCA0293, and the intergenic sequence between VCA0292 and VCA0293 and the adjacent sequence. The upstream region of the 1351-bp sequence was 223 bp away from IntI4 (VCA0291), which is the position of the attI site (indicated as a yellow rectangle). (b) Partial sequence comparison between strain with VCA0395-VCA0436 (N16961) and that without VCA0395-VCA0436 (BJ1994400). In total, 23 966 bp were not detected in strain BJ1994400 in this region. The 9372-bp sequence included the complete sequence from VCA0395 to VCA0436 and the adjacent intergenic sequence. A 552-bp direct repeat sequence which was indicated as a purple rectangle was found in both of the flanking sequences of the up- and downstream region of fragment VCA0395-VCA0436. Fragment VCA0395-VCA0436 was not detected in most of the tested strains isolated after 1990 and three strains isolated before 1990. (c) Partial sequence comparison between strain with VCA0337-VCA0351 (N16961) and those without VCA0337-VCA0351 (73364 and 73448). The 9372-bp deletion sequence in those 3 test strains included the complete sequence from VCA0337 to VCA0351 and the adjacent intergenic sequence. A 562-bp direct repeat sequence which was indicated as purple rectangle was found in both of the flanking sequences of the up- and downstream regions of fragment VCA0337-VCA0351. (d) Partial sequence comparison between strain with fragment VCA0303-VCA0317 (N16961) and those in which fragment VCA0303-VCA0317 was replaced by a new sequence (GD19612540 and 78599). In those strains, fragment VCA0303-VCA0317 was replaced by a ~9-kb sequence that included 14 ORFs. VCRs were found in both of the flanking sequences of the up- and downstream regions of fragment VCA0303-VCA0317 in N16961. (e) Partial sequence comparison between strain with fragment VCA0354-VCA0370 (N16961) and those in which fragment VCA0354-VCA0370 was replaced by a new sequence (829, 8212 and GD19612533). In strains 829, 8212 and GD19612533, fragment VCA0354-VCA0370 was replaced by a ~4-kb sequence that included eight ORFs. Fragment VCGLB07-VCGLB11 in strains 829, 8212 and GD19612533 was identical to fragment VCM66A0392-VCM66A0396 in strain M66-2.

Deletions of VCA0292-0324, VCA0338-0351, VCA0375-0382, VCA0452-0457, and VCA0460-0467    All these mutations, possibly caused by deletion, occurred in the SIs of strains 19612533, 829, and 8212 [VCA0323-0324 was found to translocate to other positions in the genome but not within the SI (Supplemental Figure S1)]. Genes VCA0381 and VCA0382 in fragment VCA0375-0382, and genes VCA0462-VCA0464 in fragment VCA0460-0467 were still amplified individually, suggesting their translocation in the chromosome out of the SI. The same SI contents existed in these strains (Supplemental Figure S1 and Figure 3). These three strains also formed a tight cluster in MLVA and PFGE UPGMA trees (Supplemental Figures S2 and S3), suggesting their high clonality. We speculated that this was a non-predominant clone that caused a cholera epidemic, but it presented itself for >20 years.

Similar to GD19612533, 829 and 8212, VCA0338-0351 and VCA0375-0382 were also absent in strain 7751 (Supplemental Figure S1). Additionally, VCA0452-0457 and VCA0460-0467 were detected in strain 7751, although they were absent in strains 19612533, 829 and 8212. In the PFGE or MLVA UPGMA tree, 7751 was far removed from the cluster formed by 19612533, 829, and 8212.

VCA0337-0351 Deletion   Negative results were obtained with primer pairs 11-48 (Supplemental Table 1) in strains 73364 and 73448, which covered VCA0337 to VCA0351. 19612533, 829, and 8212 also had similar fragment absence with one gene difference. A fragment of 1 204 bp was obtained in these two strains with primer pair 34F/49R. Sequencing confirmed a 9371-bp fragment was absent compared to SI of N16961. A 562-bp direct repeat sequence existed in both the up- and downstream flanking sequence of the absent fragment, suggesting the deletion event was mediated by homologous recombination of the 562-bp repeat sequences (Figure 4c). These two strains had the same SI composition (Figure 3), similar MLVA pattern (Supplemental Figure S2) but different PFGE patterns (Supplemental Figure S3), showing the possible difference in genome structure.

Cassettes Replacement   The corresponding sequence of VCA0303-0317 (in N16961) in strains 19612540 and 78599 was replaced by a new fragment consisting of 15 ORFs (Figure 4d). The counterparts of these ORFs were also found in the genomes of several serogroups of V. cholerae and V. mimicus strains (Table 3), whereas in the genomes of these strains, these ORFs existed individually. VCA0395-0436 deletion in the 90s-SI was also seen in strain 19612540. The MLVA and PFGE patterns of these two strains differed greatly, which suggested that the VCA0303-0317 deletion happened independently in these two strains.

The counterpart of fragment VCA0354-0370 in N16961 was replaced by a fragment consisting of eight ORFs (Figure 4e) in strains GD19612533, 829, and 8212. Some of these ORFs also existed in V. cholerae classical strain O395 and El Tor strain M66-2. VCGLB03-VCGLB10 in this fragment had a corresponding counterpart in strain O395 (VC395_A0382-0389), which was also clustered (Figure 4e). VCGLB07-VCGLB11 also had corresponding counterparts in El Tor strain M66-2 (VCM66_A0392-0396). At the same site, a different fragment containing 10 ORFs replaced VCA0354-0370 in strain 7751 (Supplemental Figure S1).

However, although fragments VCA0292-0324, VCA0327-0329 and VCA0415-0416 were not detected in strain 7751, no amplicons were obtained with primers located in the flanking sequences of these fragments, suggesting large fragments might have been inserted into these regions. Such examples also included VCA0411-0418 in strain GD1961119, VCA0355-0356 in 62110, and VCA0375-0379 in 2000180.

Table 3. Function and the Best Hit of the ORFs in the 9-kb Insertion Sequence

 

Other Dispersive Absence of Gene Fragments   Some other dispersed gene deletions compared to N16961 were also observed, such as: VCA0364-0365 absence in strains 751130, 75714, 781079, 74449, and 661673, where all the strains except for 74449 had identical or similar SI structure, 751130 and 75714 had the same MLVA and similar PFGE patterns, whereas others had similar MLVA but different PFGE patterns; VCA0364-0366 absence in strains 63244 and 642345, which had similar SI structures and MLVA patterns but different PFGE patterns; and VCA0365-0366 absence in strains 62110 and 61226, which also had similar SI structures and MLVA patterns, but different PFGE patterns. Other dispersive deletions included VCA0362-0374 in strain 801290 and VCA0294-0295 in 642345, and VCA0356 absence in 74449. These absences suggest sporadic genetic events occurred in different strains, but some clones with the same or different SI genetic composition and similar or different MLVA and PFGE patterns may also be observed, suggesting divergent genetic clones within these 1960s and 1970s strains.

Some insertions were also observed, although others still could not be discovered by PCR (Supplemental Figure S1), probably because the fragment length was much longer. It is interesting that the insertions were often accompanied by gene deletions in the integrating sites.

SI Content Types of Strains in Different Decades

Based on the cassette content sketching, we calculated the types of the SIs in the strains from different decades, on which the different gene contents were based. As shown in Table 4, the ratios of types/strains in different decades decreased from the 1960s to 1990s and after. Even the strains from the 1970s and 1980s were merged; the ratio was 0.52 (13/25) in the 1970s-1980s group. It seems that the gradually simplified diversity of SIs was observed, which may represent the SI diversity of toxigenic El Tor isolates from China, a regional cholera epidemic area.

Table 4. Ratios (No. of types/No. of strains) of the SIs in the Strains from Different Decades^*

Note. ^*Ratio was 0.60 (15/25) when the strains from 1970s and 1980s were grouped together.

Gene Flows in SIs

Based on the cassettes content and repeat sequences analysis, a sketch map of hypothetical gene flows among the different SI clades in the minimum spanning tree was constructed, which showed possible generation of the SIs in the different strains by gene deletion, insertion and replacement (Figure 5). The ancestor SI (or even the

Figure 5. Proposed hypothetical gene flows in SIs in the 60 tested toxigenic O1 El Tor V. cholerae strains. Probable insertions and deletions of ORFs found in 60 tested V. cholerae strains are indicated by red and black arrows, respectively, along the minimum spanning tree based on the PCR scanning data. Hypothetical keypoint SIs are indicated by yellow circles. The names of the strains in which the corresponding SIs existed are shown in the circles. Nineteen strains comprised 94068, 93014, 90-26, 1994400, 1993005, 1993050, 2004DD534, 84159, 2005125, 2005035, 2004152, 1998146, 64193, 1991225, 1992031, 1994A20, 2001058, 2002001, and 2005122; five strains comprised 7925, 80954, 79192, 62048, and 65930; and nine strains comprised 85079, 8788, 8593, 83101, L-ETV.760, 801298, 19861071, 1999001, and 83795.

host strain) could not be predicted because of its complicated structure and active recombination, and the sampling of the strains used in this study, therefore no arrow was used in the link line between two SIs in Figure 5. However, frequent lateral gene transfer could be seen among the SIs with different gene content. The largest number of strains used in this study possessed 90s-SI, and some gene indels still occurred within the 1990s strains. Twelve strains isolated in the 1960s had the most variant SI cassette contents, whereas the 23 1990s strains had only seven SI patterns with different cassette contents (Figure 3).

DISCUSSION

The SI is inclined to integrate and excise ORFs frequently, which results in variance of its content and structure^[28]. In this study, we analyzed the SI components based on PCR scanning and sequencing in isolates from the seventh cholera pandemic. By using this strategy, the deletions, insertions and rearrangement of cassettes were found within the different SIs. We showed that the SIs in the strains isolated from a span of nearly 50 years exhibited syntenic character but still certain diversity, especially those isolated before 1990.

An integron is an active gene capture system for its distinctive structure, especially the SI, and confers upon bacteria new abilities, such as survival, and drives bacterial evolution by integrating exogenous genes and converting them to expression^[28]. With the collection of epidemic El Tor isolates between 1961 and 2008, multiple deletions/insertions and replacements were found to contribute to the diversity of SIs. It was particularly interesting that successive ORF variations were common in SI structural mutations, while deletions or insertions of single ORFs were rare, which shows clearly the mobilization of the cassettes in clusters. The SI integrase can integrate and excise ORFs through recombination between attI sites and VCRs or between different VCRs^[29]. In the SIs of the test strains, VCRs were present in the flanking sequences of most of the variation regions, which suggested that these gene flows were caused by recombination between VCRs, which was catalyzed by integrase. Homologous recombination between the 550-bp repeat sequences flanking the VCA0395-0436 fragment possibly resulted in the characteristic ~24-kb deletion, which generated the 90s-SI structure. It also resulted in VCA0337-0351 deletion in two 1970s strains. Whether loss of these genes is advantageous for environmental and host adaption, or just a random deletion event during clone development, needs further functional studies. Within the SI of N16961, 24 copies of ~550-bp repeat sequences exist^[10], and the 550-bp repeat has high identity to the 552 bp that mediated deletion of VCA0395-436 and the 562 bp that mediated deletion of VCA0337-0351. Thus, we speculate that 24 copies of ~550 bp may provide possible sites of cassette flow in clusters mediated by homologous recombination.

The ORFs in the SI of bacteria may be transferred from virus, bacteria and even eukaryotes^[10], which is one of the sources of genome diversity and novel phenotypes in bacteria^[30]. Within three different insertion fragments in the SIs of some strains, we identified with PCR scanning and sequencing some genes that were also found in different serogroups of V. cholerae and even V. mimicus, which is more closely related to V. cholerae than other Vibrios. The SI integrases of these two species and the VCRs and V. mimicus repeats were closer than those between V. cholerae and other Vibrios^[31]. Thus, it seems that the genes captured by SIs of El Tor strains were preferentially those within the same species and the closest phylogenetic species. These autochthonous bacteria in the coastal and estuarine environment share common ecosystems, and lateral gene transfer may occur in their environmental niches^[13,31-32]. The genes shared with V. mimicus in these insertion fragments are clustered in V. cholerae, whereas in V. mimicus SI, they are located at different sites. This suggests that SI cassettes can transfer both in individual and cluster manners among different hosts.

In our study, although the strains were analyzed at a regional level, the distinct structural variance of SIs in the host strains was observed for the time span of the seventh cholera pandemic. The SIs from the strains of the 1960s epidemic wave showed multiplicity in their contents and much dispersal in the minimum spanning tree. Continuous epidemic waves occurred in the 1970s and 1980s. The SIs in the strains from the 1970s and 1980s showed lower dispersal than those from the 1960s, whereas it seems that complex clones still contemporaneously existed. When the strains from these two decades were separated into two groups, some SIs from the 1980s strains showed a similar SI structure, however multiplex SIs were continuously dispersed in the 1970s and 1980s strains. In comparison, a highly similar SI structure with a characteristic ~24-kb deletion existed in the strains from the 1990s and after, although some epidemic peaks appeared in this decade. Interestingly, the same ~24-kb deletion in 90s-SI in the El Tor strains has also been observed in the SI of O139 strain^[20], which emerged and caused a cholera epidemic in 1992^[14]. Some cassette contents of SIs and even their host strains are probably sustained for decades, because the same SI structure and similar MLVA and PFGE patterns were obtained from the strains from different decades. The 1990s clone with 90s-SI might have appeared in the 1980s, because strain 84159 had the 90s-SI and similar PFGE and MLVA patterns. Whether strain 64193 isolated in 1964 belongs to the 1990s clone with 90s-SI or not, needs further evolutionary analysis based on whole genome information. A similar situation was also observed for some 1960s, 1970s and 1980s strains, and further studies are required.

The gene content change in SIs is still obvious among strains from epidemics in different decades; however, the divergence is based on syntenic structure of SIs in these El Tor strains. From this study, it seems that the SI clades of the pandemic strains are undergoing simplification: the number of SI structural patterns in the strains grouped by decade decreased. Whether this is only the drift in clones in regional epidemics, or the result of natural stress selection needs to be further determined. However, it should be mentioned that SI diversity analysis may rely on extensive sampling. Clones of V. cholerae appearing in epidemics may be influenced by many factors, including their different environmental survival and their opportunity to invade and spread in the human population. Additionally, health care, economic factors and social development can influence the clonal spread by affecting epidemic severity. Therefore, the strains used in the evolutionary analysis may be only a part of the clones in the environment. The simplification may be only the SI structure dynamic character of the V. cholerae isolated in China. In addition, the panorama of gene flow and evolution of SIs could be described much clearly if the non-toxigenic El Tor strains are included, although it is difficult to obtain good samples from the strains surviving in the environment. In this study, only V. cholerae strains isolated in China were used. The use of strains from other countries and regions may give a global view of SI variation.

Drift in the genomes of the seventh pandemic V. cholerae strains is derived mainly from lateral gene transfer and is ongoing^[30], and the diversity of SI structure observed in this study may provide proof of genome diversity. As a special genomic structure in V. cholerae, the SI confers genome diversity and probably environmental adaptation for evolution of this bacterium. Such a variable genome structure can be used in the strain evolution analysis in different scales and even in clone definition, and may be used in the molecular epidemiological surveillance and source tracing in outbreak investigations. Here, the detailed gene content of SIs from the different V. cholerae El Tor strains was analyzed. With the whole genome sequences of many strains, the interaction between SIs and the genome in V. cholerae evolution can also be conducted, and SI gene function may even be discovered.

ACKNOWLEDGEMENTS

We sincerely thank Professor LIANG Wei Li for helpful comments and discussions.

REFERENCES

1.   Martinez E, de la Cruz F. Genetic elements involved in Tn21 site-specific integration, a novel mechanism for the dissemination of antibiotic resistance genes. The EMBO journal, 1990; 9, 1275-81.

2.   Stokes HW, Hall RM. A novel family of potentially mobile DNA elements encoding site-specific gene-integration functions: integrons. Mol Microbiol, 1989; 3, 1669-83.

3.   Hall RM, Collis CM. Mobile gene cassettes and integrons: capture and spread of genes by site-specific recombination. Mol Microbiol, 1995; 15, 593-600.

4.   Sundstrom L. The potential of integrons and connected programmed rearrangements for mediating horizontal gene transfer. APMIS Suppl, 1998; 84, 37-42.

5.   Mazel D, Dychinco B, Webb VA, et al. A distinctive class of integron in the Vibrio cholerae genome. Science, 1998; 280, 605-8.

6.   Barker A, Clark CA, Manning PA. Identification of VCR, a repeated sequence associated with a locus encoding a hemagglutinin in Vibrio cholerae O1. J Bacteriol, 1994; 176, 5450-8.

7.   Dean A. Rowe-Magnus MZ, and Didier Mazel. The adaptive genetic arsental of pathogenic Vibrio species: the role of integrons. Washington, DC, ASM Press, 2006; 95-111.

8.   Chen CY, Wu KM, Chang YC, et al. Comparative genome analysis of Vibrio vulnificus, a marine pathogen. Genome Res, 2003; 13, 2577-87.

9.   Mazel D. Integrons: agents of bacterial evolution. Nat Rev Microbiol, 2006; 4(8), 608-20.

10.       Rowe-Magnus DA, Guerout AM, Mazel D. Super-integrons. Res Microbiol, 1999; 150, 641-51.

11.       Ogawa A, Takeda T. The gene encoding the heat-stable enterotoxin of Vibrio cholerae is flanked by 123-base pair direct repeats. Microbiol Immunol, 1993; 37, 607-16.

12.       Franzon VL, Barker A, Manning PA. Nucleotide sequence encoding the mannose-fucose-resistant hemagglutinin of Vibrio cholerae O1 and construction of a mutant. Infect Immun, 1993; 61, 3032-7.

13.       Kaper JB, Morris JG, Jr., Levine MM. Cholera. Clin Microbiol Rev, 1995; 8, 48-86.

14.       Ramamurthy T, Garg S, Sharma R, et al. Emergence of novel strain of Vibrio cholerae with epidemic potential in southern and eastern India. Lancet, 1993; 341, 703-4.

15.       Heidelberg JF, Eisen JA, Nelson WC, et al. DNA sequence of both chromosomes of the cholera pathogen Vibrio cholerae. Nature, 2000; 406, 477-83.

16.       Pang B, Yan M, Cui Z, et al. Genetic diversity of toxigenic and nontoxigenic Vibrio cholerae serogroups O1 and O139 revealed by array-based comparative genomic hybridization. J Bacteriol, 2007; 189, 4837-49.

17.       Pang B, Zheng X, Diao B, et al. Whole genome PCR Scanning reveals the syntenic genome structure of toxigenic Vibrio cholera strains in the O1/O139 population. PLoS ONE, 2011; 6, e24267

18.       Feng L, Reeves PR, Lan R, et al. A recalibrated molecular clock and independent origins for the cholera pandemic clones. PLoS ONE, 2008; 3, e4053.

19.       Castaneda NC, Pichel M, Orman B, et al. Genetic characterization of Vibrio cholerae isolates from Argentina by V. cholerae repeated sequences-polymerase chain reaction. Diagn Microbiol Infect Dis, 2005; 53, 175-83.

20.       Labbate M, Boucher Y, Joss MJ, et al. Use of chromosomal integron arrays as a phylogenetic typing system for Vibrio cholerae pandemic strains. Microbiology , 2007; 153, 1488-98.

21.       Clark CA, Purins L, Kaewrakon P, et al. The Vibrio cholerae O1 chromosomal integron. Microbiology , 2000; 146, 2605-12.

22.       Danin-Poleg Y, Cohen LA, Gancz H, et al. Vibrio cholerae strain typing and phylogeny study based on simple sequence repeats. J Clin Microbiol, 2007; 45, 736-46.

23.       Cooper KL, Luey CK, Bird M, et al. Development and validation of a PulseNet standardized pulsed-field gel electrophoresis protocol for subtyping of Vibrio cholerae. Foodborne Pathog Dis, 2006; 3, 51-8.

24.       Dice LR. Measures of the amount of ecological association between species. Ecology, 1945; 26, 6.

25.       Collis CM, Recchia GD, Kim MJ, et al. Efficiency of recombination reactions catalyzed by class 1 integron integrase IntI1. J Bacteriol, 2001; 183, 2535-42.

26.       Hansson K, Sundstrom L, Pelletier A, et al. IntI2 integron integrase in Tn7. J Bacteriol, 2002; 184, 1712-21.

27.       Leon G, Roy PH. Excision and integration of cassettes by an integron integrase of Nitrosomonas europaea. J Bacteriol, 2003; 185, 2036-41.

28.       Rowe-Magnus DA, Guerout AM, Biskri L, et al. Comparative analysis of superintegrons: engineering extensive genetic diversity in the Vibrionaceae. Genome Res, 2003; 13, 428-42.

29.       MacDonald D, Demarre G, Bouvier M, et al. Structural basis for broad DNA-specificity in integron recombination. Nature, 2006; 440, 1157-62.

30.       Holmes AJ, Gillings MR, Nield BS, et al. The gene cassette metagenome is a basic resource for bacterial genome evolution. Environ Microbiol, 2003; 5, 383-94.

31.       Rowe-Magnus DA, Guerout AM, Ploncard P, et al. The evolutionary history of chromosomal super-integrons provides an ancestry for multiresistant integrons. Proc Natl Acad Sci USA, 2001; 98, 652-57.

32.       Chun J, Grim CJ, Hasan NA, et al. Comparative genomics reveals mechanism for short-term and long-term clonal transitions in pandemic Vibrio cholerae. Proc Natl Acad Sci USA, 2009; 106, 15442-7.

attention: Supplemental Table and Supplemental Figure S1, S2, and S3 can be found in the whole text on www.besjournal.com, 2011; 24(6).