Skip to main content
Advertisement
Browse Subject Areas
?

Click through the PLOS taxonomy to find articles in your field.

For more information about PLOS Subject Areas, click here.

  • Loading metrics

Microevolution of Pandemic Vibrio parahaemolyticus Assessed by the Number of Repeat Units in Short Sequence Tandem Repeat Regions

  • Katherine García,

    Affiliation Instituto de Nutrición y Tecnología de los Alimentos (INTA), Universidad de Chile, Santiago, Chile

  • Ronnie G. Gavilán,

    Affiliation Instituto de Acuicultura, Universidad de Santiago de Compostela, Campus Universitario Sur, Santiago de Compostela, Spain

  • Manfred G. Höfle,

    Affiliation Department of Vaccinology and Applied Microbiology, Helmholtz Centre for Infection Research (HZI), Braunschweig, Germany

  • Jaime Martínez-Urtaza,

    Affiliation Instituto de Acuicultura, Universidad de Santiago de Compostela, Campus Universitario Sur, Santiago de Compostela, Spain

  • Romilio T. Espejo

    romilio.espejo@gmail.com

    Affiliation Instituto de Nutrición y Tecnología de los Alimentos (INTA), Universidad de Chile, Santiago, Chile

Abstract

The emergence of the pandemic strain Vibrio parahaemolyticus O3:K6 in 1996 caused a large increase of diarrhea outbreaks related to seafood consumption in Southeast Asia, and later worldwide. Isolates of this strain constitutes a clonal complex, and their effectual differentiation is possible by comparison of their variable number tandem repeats (VNTRs). The differentiation of the isolates by the differences in VNTRs will allow inferring the population dynamics and microevolution of this strain but this requires knowing the rate and mechanism of VNTRs' variation. Our study of mutants obtained after serial cultivation of clones showed that mutation rates of the six VNTRs examined are on the order of 10−4 mutant per generation and that difference increases by stepwise addition of single mutations. The single stepwise mutation (SSM) was deduced because mutants with 1, 2, 3, or more repeat unit deletions or insertions follow a geometric distribution. Plausible phylogenetic trees are obtained when, according to SSM, the genetic distance between clusters with different number of repeats is assessed by the absolute differences in repeats. Using this approach, mutants originated from different isolates of pandemic V. parahaemolyticus after serial cultivation are clustered with their parental isolates. Additionally, isolates of pandemic V. parahaemolyticus from Southeast Asia, Tokyo, and northern and southern Chile are clustered according their geographical origin. The deepest split in these four populations is observed between the Tokyo and southern Chile populations. We conclude that proper phylogenetic relations and successful tracing of pandemic V. parahaemolyticus requires measuring the differences between isolates by the absolute number of repeats in the VNTRs considered.

Introduction

Diarrhea associated with seafood consumption is caused primarily by pathogenic Vibrio parahaemolyticus. This species includes a large number of marine strains, only a few of which are pathogenic in humans [1]. Cases of diarrhea related to seafood consumption increased worldwide with the emergence of pandemic strain O3:K6, which was originally observed in Southeast Asia [2]. Isolates in this group are commonly recognized by genetic markers, including the O3:K6 antigens, which determine the serovar; the presence of genes toxRS/new [3], orf8 [4], and tdh; and the absence of trh, which is found in some pathogenic strains. Isolates with this genotype correspond to a clonal complex, whose independently obtained isolates are rarely differentiated. The identity of the genomes is assessed by multilocus sequence typing (MLST) [5], [6], genome restriction fragment length polymorphism–pulsed field gel electrophoresis (RFLP-PFGE) [7], direct genome restriction enzyme analysis (DGREA) [8], and arbitrarily primed polymerase chain reaction (AP-PCR) [2], [3]. Some variants have been observed occasionally, namely in serotype [9] but also in RFLP-PFGE or AP-PCR patterns [2], [10], and others lacking toxRS/new or orf8 [10]. Microarray-based comparative genomic hybridization (M-GCH) of 4021 genes allowed the differentiation of 39 pandemic strains into five subgroups [11]. The M-GCH data (log2 ratios) for these five subgroups vary in the genomic islands and O/K antigen genes. Despite the resolving power of these last methods, however, the discrimination is not sufficient for tracing the isolate in seafood outbreaks caused by pandemic V. parahaemolyticus. Multiple locus variable analysis (MLVA) of the variable number tandem repeats (VNTRs) has been developed for most medically relevant bacterial species and can be used effectively for tracing outbreaks or other forms of bacterial dissemination [12], [13]. VNTRs consist of short sequences, known as repeat units or motifs, that are repeated in tandem and have been shown to vary in repeat copy number by the insertion or deletion of one or more repeat units. Although recombination that produces large differences also occurs, it is less frequent [14]. This system has been highly successful for epidemiological studies of genetically homogeneous bacterial pathogens, such as Yersinia pestis [15], V. cholerae O1 and O139 [16], Escherichia coli O157:H7 [17], Bartonella henselae [18], and Mycobacterium leprae [19], providing useful genetic discrimination whether the populations were worldwide, regional, or from a local outbreak. MLVA of VNTRs has also been employed to study within-host evolution of Burkholderia pseudomallei infection [20]. The high-resolution power of MLVA was recently shown for pandemic V. parahaemolyticus; eight VNTRs in 28 pandemic V. parahaemolyticus O3:K6 strains isolated from human cases produced 28 distinct VNTR patterns [21]. Analysis of 36 pandemic isolates belonging to the clonal complex isolated in Chile produced 26 patterns [22]. Multiple loci VNTR analysis of a number of representative pandemic V. parahaemolyticus strains from Asia, Peru, and Chile using seven polymorphic loci divided the populations into two genetically distinct groups [23]. One of them grouped with the original Asiatic population and strains arriving in Peru and Chile in 1997 comprised one group. Thus, MLVA seems useful for tracing and studying the phylogeny within highly homogenous bacterial species subgroups. However, a more exact interpretation of the results requires knowing the rate and mechanism of variation of the VNTRs. Here, we show that the variation of VNTRs is better explained by single stepwise mutation (SSM) and that a more credible phylogeny by comparing VNTRs is obtained when data is analyzed considering that variation between VNTRs occurs by SSM. We also show that the analysis of the relation between variants generated by SSM show plausible results using the Minimum Spanning Trees (MST) method with the Manhattan category which builds the tree considering the sum of the differences in repeat units of any two mutants [24]. MST is a mathematical topology tool that applies the maximum parsimony principle, applicable for population modeling (micro-evolution) and epidemiology. When a set of distances is given between n samples, a minimum spanning tree is the tree that connects all samples in such a way that the summed distance of all branches of the tree is minimized.

Materials and Methods

Clones and Native Bacterial Strains

Four different clones of pandemic V. parahaemolyticus were used for serial cultivation and subsequent MLVA analysis. Clones are designated those cultures obtained from single colonies of strains VpKX and PMC57.5 used for in vitro study of the mutation rates. Native strains were isolates with pandemic V. parahaemolyticus characteristics, and were obtained from clinical and environmental samples from different geographical sites. Clone KX-1 was obtained from a sample of the reference prototype strain RIMD2210633 (VpKX), isolated in 1996 and sequenced [25]. It was received in 2005 from Professor Takeshi Honda (Research Institute for Microbial Diseases, Osaka University) and kept in soft agar from this year [4]. Clones PMC57.5-1, -2, and -3 were obtained from PMC57.5, isolated in 2005 from a clinical sample in Puerto Montt, Region de Los Lagos, and kept in soft agar at room temperature. The pandemic strains (between 1995 and 1998) from Southeast Asia used in this study were received from Mitsuaki Nishibuchi (Center for Southeast Asian Studies, Kyoto University). Chilean native strains were collected from the coast of Antofagasta (1998) in the north and from Region de Los Lagos in the south (between 2004 and 2009). Most of the isolates were described in previous publications [8], [26], [27], [28], [29]. Data for the pandemic strains from Tokyo correspond to those reported by Kimura et al [21]. These strains are described in Table S1.

Serial Subcultures

Serial subculturing was performed as described previously [30]. Briefly, clones KX-1) and PMC57.5-1 were suspended in 5 mL of LB with 3% NaCl and incubated at 37°C in a rotary shaker overnight. Each culture was serially propagated into 20 subcultures diluted 1∶100 plus 80 subcultures diluted 1∶10000. After 100 subcultures, performed in 100 days, 60 colonies from each culture were picked up for MLVA.

Parallel Serial Passage Experiments (PSPE)

PSPE were performed as described previously [15]. Two single colonies from strain PMC57.5 obtained in LB agar 3% NaCl (clones PMC57.5-2 and PMC57.5-3) were grown independently in 5 mL of LB with 3% NaCl, and a single colony from each of these cultures was serially transferred for 10 passages. In both types of in vitro mutants, from subcultures and from PSPE, the number of total generations was calculated as follows: number of colonies x number of duplications for each colony per subculture x number of subcultures.

DNA Extraction and MLVA of bacterial clones, in vitro mutants and native bacterial strains

DNA was extracted with the DNA Wizard Genomic kit (Promega, Madison,WI). PCR reactions and analysis of the amplicon size were performed as described previously [22] except that only six VNTR loci (VNTRs 1, 3, 5, 6, 7 and 8) were selected for determination of mutation rates (Table 1) and each VNTR was amplified independently. When 8 VNTR loci (VNTRs 1 to 8) were amplified for the phylogenetic study of the native strains three multiplex reactions were performed. MultiA (VNTR 2 and 7), MultiB (VNTR 1 and 8) and MultiC (VNTR 2, 3, 4, and 6).

thumbnail
Table 1. Number of repeat units of six VNTR loci in four clones obtained from pandemic Vibrio parahaemolyticus strains VpKX and PMC57.5 and mutation rates after prolonged in vitro culturing.

https://doi.org/10.1371/journal.pone.0030823.t001

Mutation Rate and Analysis of Mutation Type

Mutation rate was calculated as: number of mutants/number of generations. The distribution of VNTR mutants with 1, 2, 3, or more repeat units deleted or inserted was calculated from the combined data for the populations from the four clones of pandemic V. parahaemolyticus examined. A geometric distribution was obtained with the experimental data of the number of mutants that differed in one repeat unit from the closest neighbor that was coincident with the theoretical expected distribution, according to the equation P(X = n)  =  P(1-P)n-1 [31], where P is the probability of a single repeat mutation, and n is number of repeats involved in a mutation. Only mutants that differed in less than three repeat units from their closer neighbors were considered. Mutations that occur simultaneously in more than one repeat unit were assumed to be produced by recombination [32].

Genetic Distance

Genetic distance, (δµ)2, of allele size in populations computed as δµ2  =  (µA–µB)2 where µA y µB are the variances and means, respectively, was defined previously for microsatellite loci, incorporating the features of stepwise mutation model [33] by including the difference in each VNTR, by using Microsatellite Analyzer (MSA) version 4.05 software. The neighbor-joining tree-building method of MEGA4 [34] software was used to infer the phylogenetic tree based on genetic distance.

Phylogenetic Relationship analysis of in vitro Mutants or Native Variants

The minimum spanning tree (MST) was calculated for the different populations of the mutants obtained in vitro, and from the native bacterial strains, by using the Manhattan coefficient (offset = 0, saturation = 3) in Bionumerics (v 5.10, Applied Maths, Sint-Martens-Latern, Belgium) [24].

Results

Diversity of VNTRs among Clones of the Same Strain and Mutation Rates

We measured the mutation rates of six VNTRs for four clones, obtained by culturing single colonies of V. parahaemolyticus pandemic strains. One of these clones, KX-1, was obtained from a sample of strain RIMD2210633 (VpKX), which was received in 2005 and has been sequenced [25]. This clone (KX-1) showed the same repeat numbers reported for the published genome sequence in each analyzed VNTR, except in VNTR3, which contained seven repeat units instead of the six reported in the genome sequence. The other three clones were derived from strain PMC57.5, which was obtained from a clinical sample in southern Chile in 2005 and purified twice by selecting a single colony after plating [8]. Although strain PMC57.5 is indistinguishable from VpKX by the common genetic markers of the pandemic strain, as well as by RFLP-PFGE and DGREA patterns [8], examined clones from this strain differed from VpKX and among themselves in several VNTRs. Ten clones derived from PMC57.5 were randomly selected to explore their possible variation in VNTRs; among them, eight contained 34 and 14 repeat units in VNTRs 1 and 6, respectively; one (PMC57.5-2) contained 38 repeat units in VNTR1; and another (PMC57.5-3) contained six repeat units. This last clone also differed from the other clones of the same strain by the presence of 18 and 5 repeat units in VNTRs 6 and 8, respectively, instead of the 14 and 7 repeat units found in the other clones (Table 1). Because VpKX and the PMC57.5 strains had been kept in sealed stab agar at room temperature since 2005, the observed diversity among the clones derived from the same sample may be due to growth during storage; in some Salmonella spp., such phenomenon has been observed [35], [36]. Mutation rates were measured in experiments performed by either serial subculturing [2] or a series of PSPEs [15]. Mutation rates differed between both VNTRs and clones, which was apparently related to the number of repeats in the VNTR in each clone (Figure 1). The relationship between the number of repeats and mutation rate was described previously [31], [32], but this is not precise when different VNTR are considered. The relationship became more precise when the same VNTR with different numbers of repeat units is considered (note the values of R presented in the legend of Figure 1 for each situation). The low number of repeat units in VNTR1 of clone PMC57.5-3 is the probable cause of the much smaller mutation rate observed for VNTR1 in this clone.

thumbnail
Figure 1. Mutation rates of VNTRs of pandemic V parahaemolyticus, in relation to the number of repeats in each VNTR.

White circles correspond to the mutation rates observed for any VNTR, circles in black correspond to the mutation rates observed for VNTR1 in the different clones. The best fit equation for VNTR1 was y = 0.1736×–1.1245 (R2 = 0.9674) while that for all the VNTRs was y =  0.1598×– 0.7923 (R2 = 0.8052).

https://doi.org/10.1371/journal.pone.0030823.g001

Diversity in the Population Obtained after Serial Cultivation

The distribution of mutation types observed for the different pandemic V. parahaemolyticus clones was consistent with the distributions observed in E. coli O157:H7 [17] and Y. pestis [32]. The frequency of mutants with 1, 2, 3, or 4 repeat unit changes roughly followed a geometric distribution. Though deviations were observed in some experiments, the mean distribution of the four clones showed fairly good agreement with geometric distribution (Table 2, Figure 2). As observed in previous studies [2], [30], we found both an unexpected high proportion of insertions and a small number of mutants with large changes in repeat units in the VNTRs (Figure 2). As previously postulated for clonal populations including similar generation numbers, mutants with changes smaller than four repeat units may be generated by SSM in which mutations occur by the insertion or deletion of single units, as initially proposed for microsatellites in eukaryotes [37] and later for bacteria [32]. On the other hand, mutants with large changes in repeat units, unlikely to be observed if the distribution were ideally geometric are probably generated by recombination. Overall, our results support the theory advanced for E. coli O157:H7 [31] and Y. pestis [32], that mutations in VNTRs mostly occur by SSM.

thumbnail
Figure 2. Frequency of mutants with increasing differences in the number of repeat units.

Black bars correspond to percent of predicted mutants. Combined bars correspond to percentage of observed mutants, white sections of bars correspond to deletions and gray to insertions.

https://doi.org/10.1371/journal.pone.0030823.g002

thumbnail
Table 2. Frequency of observed and expected mutants according to a geometric distribution, with 1, 2, 3, or 4 repeat unit changes for each of the clones examined.

https://doi.org/10.1371/journal.pone.0030823.t002

Evolutionary Models and Phylogenetic Trees of in vitro mutants

Taking into consideration that most mutants emerge by SSM, we constructed phylogenetic trees of the mutants observed after prolonged cultivation by considering the absolute differences in the number of repeats in the different VNTRs. For this reason we employed Minimum Spanning Tree (MST) using the Manhattan category which builds the tree considering the sum of the differences in repeat units of any two mutants. If the absolute value of repeat units is not considered, using by example a categorical approach in MST, or the categorical approach of multilocus sequence typing (MLST), mutant pairs with different number of repeat units appear equally distant. The difference in the phylogenetic trees when these different coefficients are employed is evident in Figure 3 which shows the MST tree obtained for the KX-1 population of 75,600 generations using either the Manhattan (Figure 3A) or the categorical approach (Figure 3B). Cluster B correspond to the founder. However, when the absolute difference in the number of repeat units is not considered, the founder is not properly identified and distantly related mutants are shown as closely related or considered direct variants (Figure 3B). In contrast, the tree obtained by considering the absolute difference in repeat units shows only mutants differing in a single repeat unit as direct variants, as would be expected if mutants are generated by SSM. Because one potential application of the MLVA of V. parahaemolyticus is the tracing of outbreaks, we analyzed the combined four populations obtained from each the four original clones after serial passages in the laboratory using the Manhattan approach. Despite the small differences in VNTRs among the founders, we correlated almost every mutant with its actual founder, distinguishing between populations with different but closely related founders (Figure 4). If the analysis is performed with the categorical approach, only a few mutants derived from different founders were distinguished (data not shown).

thumbnail
Figure 3. Phylogenetic trees for KX-1 population obtained by MST-Manhattan (A) and MST categorical (B).

Each circle corresponds to the different cluster of mutants. Numbers in the lines correspond to the differences in VNTRs repeat units between clusters.

https://doi.org/10.1371/journal.pone.0030823.g003

thumbnail
Figure 4. Phylogenetic tree of the mutants from the populations of the four different clones obtained after serial subculturing of pandemic V. parahaemolyticus.

Tree was obtained by MST-Manhattan. The founder of each cluster is indicated above the clusters.

https://doi.org/10.1371/journal.pone.0030823.g004

Phylogeny of Native Populations

Another useful application of MLVA is population modeling (microevolution) and epidemiology [23], [38]. We used the MST approach to study the relationships between a large number of native strains of V. parahaemolyticus. Eight VNTR loci were used for this analysis (Table S4). Because pandemic clinical strains might represent a subpopulation of the pandemic strain population found in shellfish, we compared 29 clinical strains and 21 environmental strains collected in southern Chile between 2004 and 2009. The analysis did not differentiate between populations of clinical and environmental strains. No differences were observed between strains isolated in different years. We also compared the strains collected in southern Chile; northern Chile, where an outbreak caused by the pandemic strain was observed in 1997 [28]; and Southeast Asia to determine if MLVA (Manhattan approach) allows us to distinguish between strains according to the geographic origin and to establish possible relationships among these populations. We also included the data from a collection of 28 strains isolated in Tokyo and analyzed for the same 8 VNTRs by Kimura et al [21]. Using this approach, almost every mutant from these four populations were clustered according to their geographic origin (Figure 5). In accordance with previous results [23], populations from the outbreak in northern Chile seem to be more closely related to the Southeast Asian population than the population from southern Chile. Interestingly, the Japanese population consisting of mostly clinical strains isolated from single patients in Tokyo between 1996 and 2003 [21] clustered separately from isolates obtained in different Southeast Asia locations.

thumbnail
Figure 5. Phylogenetic tree of native isolates of pandemic V. parahaemolyticus obtained worldwide.

Phylogenetic tree generated by MST-Manhattan for native isolates obtained from southern Chile (blue) northern Chile (red), Southeast Asia (brown) and Tokyo (yellow).

https://doi.org/10.1371/journal.pone.0030823.g005

Genetic distance (δµ)2 has been used to date divergence between mammal populations [39]. We used this parameter that corresponds to the differences between the mean lengths of the VNTRs of each population in order to date the possible split between the four native populations analyzed by MST. Because strains from Tokyo and southern Chile were isolated during a period of 8 and 6 years, respectively, we divided these populations into two (ancient and recent isolates, obtained in the first and second half of the period, respectively). The calculated genetic distances (Table S2) allowed us to construct a tree by neighbor-joining, showing the apparent split of the populations of pandemic V. parahaemolyticus from different geographic regions (Figure 6).

thumbnail
Figure 6. Neighbor Joining tree for different populations of pandemic V. parahaemolyticus.

Neighbor joining tree based on (δµ)2 genetic distance for population of clinical isolates from southern Chile (PMCa isolated from 2004–2006 and PMCr isolated from 2007–2009), northern Chile (ATC isolated from 1997–1998), Southeast Asia (isolated from 1996–1998) and Tokyo (TOKYOa isolated 1996–1999 and TOKYOr from 2000–2003).

https://doi.org/10.1371/journal.pone.0030823.g006

Discussion

The analysis of VNTRs with large mutation rates offers an opportunity to study evolution in a clonal population, such as pandemic V. parahaemolyticus. However, interpretation of the results for microevolution and population epidemiology requires knowledge of the mutation rates and understanding how VNTRs change with time. The mutation rates we measured in pandemic V. parahaemolyticus were similar to those observed in E. coli O157:H7 and Y. pestis by PSPE; values observed were from 3.4×10−6 to 4.0×10−4 for E. coli and 8.5×10−6 to 3.7×10−4 for Y. pestis [31], [32]. The differences in the number of repeat units between the clone of RIMD2210633 (VpKX) we analyzed, and that from other clones whose genome was sequenced [25] are probably due to subsequent propagation of the strain. More illustrative of the changes in VNTRs is the finding of different clones from purified colonies of V. parahaemolyticus stored in agar stabs. The emergence of the variants probably occurred during growth, which is known to occur in bacteria stored in stabs [35], [36]. These observations indicate the need to store strains at a low temperature to avoid growth and to consider the possible consequences of strains stored in stab gels in MLVA results. This consideration should also be taken into account when sequencing the whole genome of stored strains, and in the interpretation of differences between sequenced bacterial genomes.

In this study we determined the relationship between mutation rates and the number of repeats in VNTRs, which was already described in E. coli O157:H7 and Y. pestis [31], [32], but in this case we are able to demonstrate that this relationship is more evident and precise when the mutation rates observed of the same VNTR from clones with different numbers of repeats are compared. However, it cannot be discarded that the relationship is more precise because the high number of repeat units in the VNTR compared (VNTR1). The mechanism of VNTR variation has been more deeply explored and discussed regarding the variation of tandem repeats in microsatellites in eukaryotes. The accepted mutational mechanism leading to changes in microsatellite length is polymerase template slippage [40], [41]. During the replication of a repetitive region, DNA strands may dissociate and then re-associate incorrectly. Renewed replication in this misaligned state leads to the insertion or deletion of repeat units, altering allele length. The oldest model, and probably the simplest, is SSM, in which the number of repeat units is equally likely to increase or decrease by one at a rate independent of the microsatellite length [37]. Since then a number of models have been proposed to account for the effect of increasing microsatellite length, length limitations, an increase or decrease of more than one repeat in a single mutation event, point mutations that interrupt the repeat chain, and others. One of these models considers SSM and proportional slippage and point mutations [39]. One of the strongest arguments for the SSM model is that the distribution of mutants with 1, 2, or 3 repeat unit differences follows a geometric distribution. Similar to Vogler et al [31], [32], we also found a geometric distribution for mutants with 1, 2, or 3 repeat unit differences. Using a different approach, we previously validated the SSM model for a large set of native strains of pandemic V. parahaemolyticus by showing that at least 71% of the allelic changes between closest relatives were related to differences in one repeat [23]. Despite the general acceptability of this model, some MLVAs consider differences in each allele (VNTR) in a binary or categorical system without pondering the differences in the number of repeat units. This consideration is only later integrated for deeper analysis of the results, though it has been extensively used for analyzing microsatellite variation [39], [42]. According to a SSM model of VNTR variation, an appropriate method for analyzing MLVA data is the MST of Bionumerics using the Manhattan category; in this way the absolute value of the difference in the number of repeat units is incorporated for construction of the more parsimonious tree. Because Bionumerics is not freely available, goeBURST can be employed with similar results to MST using Manhattan category if each informative repeat unit, instead of each VNTR, is introduced as a locus or allele. goeBURST [43] (http://goeburst.phyloviz.net/) is a modified version of eBURST, a parsimony-based method commonly used to determine the genetic relatedness of bacterial populations that have diverged over short evolutionary time spans using either MLVA or MLST. goeBURST will consider absolute difference of repeat units if each unit is a locus or allele instead of the VNTR, as is usually done. Table S3 contains the data from the KX-1 experiment displayed as should be uploaded in goeBURST. The alternative tree obtained when the absolute difference in repeat units is not considered (categorical, Figure 3B) seems not reliable at first because it shows a similar relationship for pairs differing in one of the VNTRs independent of the difference in the number of repeat units between these pairs. A more objective comparison of the plausibility of both trees was obtained by calculating the overall probability for these alternative trees. When calculated according to Vogler et al [32] as P  =  Пni = 1µi, where µi is the probability of a given mutation and n is the number of mutational steps, the results were 3.2×10−35 for the categorical MST tree versus 8.8×10−31 for the Manhattan MST tree, assuming SSM.

Previous Manhattan-MST analysis of 69 strains from Asia, Peru, and Chile distinguished two groups. One group included all strains from Asia and some from Peru, and northern Chile. The second group was composed of strains from Peru and southern Chile [23]. Using a different set of VNTRs and a larger collection of 98 strains, that comprise the strains from Tokyo analyzed by Kimura et al [21], we broaden the population epidemiology of pandemic V. parahaemolyticus in the present study. The MST-Manhattan showed, as in the previous publication [23], that strains from northern Chile cluster together with Southeast Asia strains and that this cluster is clearly differentiated from strains isolated in southern Chile. Unexpectedly, we found that the strains from Tokyo clearly differentiate from the strains of Southeast Asia. However, the MST tree is probably better interpreted together with the Neighbor -Joining tree based on genetic distance. This parameter, defined as (äµ)2, and incorporates features of the SSM, has been used to infer population structure and demographic history in mammals, including humans, according to microsatellite variation [33], [42]. The MST-Manhattan and (δµ)2 trees suggest that the strains that caused the outbreaks in northern Chile are closely related to Southeast Asia strains and, thus, probably arrived in northern Chile from Asia. The closeness of the strains found in southern and northern Chile suggests that strains in the south were probably derived from northern strains and do not correspond to an independent introduction of Southeast Asia strains. On the other hand, strains from Tokyo are more closely related to Southeast Asia strains. When interpreting the trees it should be considered that beside geographical origin the strains differ by date of isolation. Differences in the date of isolation seemed to be more clearly observed in strains from Tokyo isolated from 1996 to 1999 (TOKYOa) and from 2000 to 2003 (TOKYOr) (Figure 6). The deepest split in the pandemic populations we analyzed seems to have occurred between the southern Chile and Tokyo populations. Goldstein et al [33], [42] used (δµ)2 to estimate divergence times in primates according to a relationship between the average mutation rate of microsatellites and the generation time, according to the equation: Eg [ (δµ)2 (T)]  = 2β. Assuming that the isolates found in southern Chile between 2004 and 2009 descend from a population that split from the Southeast Asia population isolated between 1996 and 1998 and hence evolved independently for about 10 years, pandemic V. parahaemolyticus can be estimated to have been reproducing with a generation time of 3.6 days. Assuming that the genetic distances underlying the tree have a linear variation with time, the split of the Tokyo and southern Chile populations from the Southeast Asia population occurred around the same time. Similarly, the southern Chile population would have split from the northern Chile population approximately 5 years ago.

Supporting Information

Table S1.

Native strains and clones (*) of pandemic Vibrio parahaemolyticus used in this work.

https://doi.org/10.1371/journal.pone.0030823.s001

(XLS)

Table S2.

Genetic distance (δµ)2 between populations of pandemic V. parahaemolyticus isolated in different years and geographical location.

https://doi.org/10.1371/journal.pone.0030823.s002

(XLS)

Table S3.

Number of repeats in the different VNTRs of the KX-1 population, ready for uploading in goeBURST.

https://doi.org/10.1371/journal.pone.0030823.s003

(XLS)

Table S4.

VNTR profiles of native strains used in this work.

https://doi.org/10.1371/journal.pone.0030823.s004

(XLS)

Acknowledgments

It is a pleasure to acknowledge the helpful review of the manuscript by Paulina Uribe and the excellent technical assistance of Rafael Torres.

Author Contributions

Conceived and designed the experiments: KG RTE MGH. Performed the experiments: KG. Analyzed the data: KG RTE JMU MGH RGG. Contributed reagents/materials/analysis tools: RTE MGH JMU. Wrote the paper: KG RTE.

References

  1. 1. Iida T, Park KS, Honda T (2006) Vibrio parahaemolyticus. In: Thompson FL, Austin Swings BJ, editors. The Biology of Vibrios. Washington D.C.: ASM Press. pp. 340–348.
  2. 2. Okuda J, Ishibashi M, Hayakawa E, Nishino T, Takeda Y, et al. (1997) Emergence of a unique O3:K6 clone of Vibrio parahaemolyticus in Calcutta, India, and isolation of strains from the same clonal group from Southeast Asian travelers arriving in Japan. J Clin Microbiol 35: 3150–3155.
  3. 3. Matsumoto C, Okuda J, Ishibashi M, Iwanaga M, Garg P, et al. (2000) Pandemic spread of an O3:K6 clone of Vibrio parahaemolyticus and emergence of related strains evidenced by arbitrarily primed PCR and toxRS sequence analyses. J Clin Microbiol 38: 578–585.
  4. 4. Nasu H, Iida T, Sugahara T, Yamaichi Y, Park KS, et al. (2000) A filamentous phage associated with recent pandemic Vibrio parahaemolyticus O3:K6 strains. J Clin Microbiol 38: 2156–2161.
  5. 5. Chowdhury NR, Stine OC, Morris JG, Nair GB (2004) Assessment of evolution of pandemic Vibrio parahaemolyticus by multilocus sequence typing. J Clin Microbiol 42: 1280–1282.
  6. 6. Gonzalez-Escalona N, Martinez-Urtaza J, Romero J, Espejo RT, Jaykus LA, et al. (2008) Determination of molecular phylogenetics of Vibrio parahaemolyticus strains by multilocus sequence typing. J Bacteriol 190: 2831–2840.
  7. 7. Wong HC, Liu SH, Wang TK, Lee L, Chiou CS, et al. (2000) Characteristics of Vibrio parahaemolyticus O3:K6 from Asia. Appl Environ Microbiol 66(9): 3981–6.
  8. 8. Fuenzalida L, Hernandez C, Toro J, Rioseco ML, Romero J, et al. (2006) Vibrio parahaemolyticus in shellfish and clinical samples during two large epidemics of diarrhea in southern. Chile Environ Microbiol 8: 675–683.
  9. 9. Nair GB, Ramamurthy T, Bhattacharya SK, Dutta B, Takeda Y, et al. (2007) Global dissemination of Vibrio parahaemolyticus serotype O3:K6 and its serovariants. Clin Microbiol Rev 20: 39–48.
  10. 10. Chowdhury A, Ishibashi M, Thiem V, Tuyet D, Tung T, et al. (2004) Emergence and serovar transition of Vibrio parahaemolyticus pandemic strains isolated during a diarrhea outbreak in Vietnam between 1997 and 1998. Microbiol Inmunol 48: 319–327.
  11. 11. Han H, Wong HC, Kan B, Guo Z, Zeng X, et al. (2008) Genome plasticity of Vibrio parahaemolyticus: microevolution of the ‘pandemic group’ BMC. Genomics 9: 570.
  12. 12. Van Belkum A (2007) Tracing isolates of bacterial species by multilocus variable number of tandem repeat analysis (MLVA) FEMS. Immunol Med Microbiol 49: 22–27.
  13. 13. Lindstedt BA (2005) Multiple-locus variable number tandem repeats analysis for genetic fingerprinting of pathogenic bacteria. Electrophoresis 26: 2567–2582.
  14. 14. Eisen J (1999) Mechanistic basis for microsatellite instability. In: Goldstein DB, Schötterer C, editors. Microsatellites: evolution and applications. New York: Oxford University Press. pp. 34–48.
  15. 15. Girard JM, Wagner DM, Vogler AJ, Keys C, Allender CJ, et al. (2004) Differential plague-transmission dynamics determine Yersinia pestis population genetic structure on local, regional, and global scales. Proc Natl Acad Sci USA 101: 8408–8413.
  16. 16. Ghosh R, Nair GB, Tang L, Morris JG, Sharma NC, et al. (2008) Epidemiological study of Vibrio cholera using variable number of tandem repeats FEMS. Microbiol Lett 288: 196–201.
  17. 17. Pei Y, Terajima J, Saito Y, Suzuki R, Takai N, et al. (2008) Molecular characterization of enterohemorrhagic Escherichia coli O157:H7 isolates dispersed across Japan by pulsed-field gel electrophoresis and multiple-locus variable-number tandem repeat analysis Jpn. J Infect Dis 61: 58–64.
  18. 18. Monteil M, Durand B, Bouchouicha R, Petit E, Chomel B, et al. (2007) Development of discriminatory multiple-locus variable number tandem repeat analysis for Bartonella henselae. Microbiology 153: 1141–1148.
  19. 19. Hall BG, Salipante SJ (2010) Molecular epidemiology of Mycobacterium leprae as determined by structure-neighbor clustering. J Clin Microbiol 48: 1997–2008.
  20. 20. Price EP, Hornstra HM, Limmathurotsakul D, Max TL, Sarovich DS, et al. (2010) Within-host evolution of Burkholderia pseudomallei in four cases of acute melioidosis. PLoS Pathog 6: e1000725.
  21. 21. Kimura B, Sekine Y, Takahashi H, Tanaka Y, Obata H, et al. (2008) Multiple-locus variable-number of tandem-repeats analysis distinguishes Vibrio parahaemolyticus pandemic O3:K6 strains. J Microbiol Methods 72: 313–320.
  22. 22. Harth-Chu E, Espejo RT, Christen R, Guzmán CA, Höfle MG (2009) Multiple locus variable-number tandem-repeat analysis for clonal identification of Vibrio parahaemolyticus isolates by using capillary electrophoresis. Appl Environ Microbiol 75: 4079–4088.
  23. 23. Ansede-Bermejo J, Gavilan RG, Trinanes J, Espejo RT, Martinez-Urtaza J (2010) Origins and colonization history of pandemic Vibrio parahaemolyticus in South America. Mol Ecol 19: 3924–3937.
  24. 24. Minimum spanning trees for population modeling (2008) The Bionumerics Manual Version 5.1. Cap. 4.11. pp. 263–270. Available online in www.applied-maths.com. Accessed 2011 Nov 15.
  25. 25. Makino K, Oshima K, Kurokawa K, Yokoyama K, Uda T, et al. (2003) Genome sequence of Vibrio parahaemolyticus: a pathogenic mechanism distinct from that of V cholera. Lancet 361: 743–749.
  26. 26. Fuenzalida L, Armijo L, Zabala B, Hernandez C, Rioseco ML, et al. (2007) Vibrio parahaemolyticus strains isolated during investigation of the summer 2006 seafood related diarrhea outbreaks in two regions of Chile. Int J Food Microbiol 117: 270–275.
  27. 27. Garcia K, Torres R, Uribe P, Hernandez C, Rioseco ML, et al. (2009) Dynamics of clinical and environmental Vibrio parahaemolyticus strains during seafood-related summer diarrhea outbreaks in southern. Chile Appl Environ Microbiol 75: 7482–7487.
  28. 28. Gonzalez-Escalona N, Cachicas V, Acevedo C, Rioseco ML, Vergara JA, et al. (2005) Vibrio parahaemolyticus diarrhea, Chile 1998 and 2004. Emerg Infect Dis 11: 129–131.
  29. 29. Harth E, Matsuda L, Hernandez C, Rioseco ML, Romero J, et al. (2009) Epidemiology of Vibrio parahaemolyticus outbreaks, southern. Chile Emerg Infect Dis 15: 163–168.
  30. 30. Papadopoulos D, Schneider D, Meier-Eiss J, Arber W, Lenski RE, et al. (1999) Genomic evolution during a 10,000-generation experiment with bacteria. Proc Natl Acad Sci USA 96: 3807–3812.
  31. 31. Vogler AJ, Keys C, Nemoto Y, Colman RE, Jay Z, et al. (2006) Effect of repeat copy number on variable-number tandem repeat mutations in Escherichia coli O157:H7. J Bacteriol 188: 4253–4263.
  32. 32. Vogler AJ, Keys CE, Allender C, Bailey I, Girard J, et al. (2007) Mutations, mutation rates, and evolution at the hypervariable VNTR loci of Yersinia pestis. Mutat Res 616: 145–158.
  33. 33. Goldstein DB, Ruiz LA, Cavalli-Sforza LL, Feldman MW (1995) Genetic absolute dating based on microsatellites and the origin of modern humans. Proc Natl Acad Sci USA 92: 6723–6727.
  34. 34. Tamura K, Dudley J, Nei M, Kumar S (2007) MEGA4: Molecular Evolutionary Genetics Analysis (MEGA) software version 4.0. Mol Biol and Evol 24: 1596–1599.
  35. 35. Rabsch W, Helm RA, Eisenstark A (2004) Diversity of phage types among archived cultures of the Demerec collection of Salmonella enterica serovar Typhimurium strains. Appl Environ Microbiol 70: 664–669.
  36. 36. Sutton A, Buencamino R, Eisenstark A (2000) rpoS mutants in archival cultures of Salmonella enterica serovar typhimurium. J Bacteriol 182: 4375–4379.
  37. 37. Kimura M, Ohta T (1978) Stepwise mutation model and distribution of allelic frequencies in a finite population. Proc Natl Acad Sci USA 75: 2868–2872.
  38. 38. Li Y, Cui Y, Hauck Y, Platonov ME, Dai E, et al. (2009) Genotyping and phylogenetic analysis of Yersinia pestis by MLVA: insights into the worldwide expansion of Central Asia plague foci. PLoS ONE 4: e6000. doi:10.1371/journal.pone. 0006000 p.
  39. 39. Calabrese PP, Durrett RT, Aquadro CF (2001) Dynamics of microsatellite divergence under stepwise mutation and proportional slippage/point mutation models. Genetics 159: 839–852.
  40. 40. Schlotterer C, Tautz D (1992) Slippage synthesis of simple sequence DNA. Nucleic Acids Res 20: 211–215.
  41. 41. Strand M, Prolla TA, Liskay RM, Petes TD (1993) Destabilization of tracts of simple repetitive DNA in yeast by mutations affecting DNA mismatch repair. Nature 365: 274–276.
  42. 42. Goldstein DB, Roemer GW, Smith DA, Reich D, Bergman A, et al. (1999) The use of microsatellite variation to infer population structure and demographic history in a natural model system. Genetics 151: 797–801.
  43. 43. Francisco AP, Bugalho M, Ramirez M, Carrico JA (2009) Global optimal eBURST analysis of multilocus typing data using a graphic matroid approach.BMC Bioinformatics 10: 152.