JOURNAL OF CLINICAL MICROBIOLOGY, Oct. 2010, p. 3666–3674 Vol. 48, No. 10 0095-1137/10/$12.00 doi:10.1128/JCM.00866-10 Copyright ? 2010, American Society for Microbiology. All Rights Reserved.
Cholera between 1991 and 1997 in Mexico Was Associated with
Infection by Classical, El Tor, and El Tor Variants
of Vibrio cholerae
111112Munirul Alam,Suraia Nusrin,Atiqul Islam,Nurul A. Bhuiyan,Niaz Rahim,Gabriela Delgado, 22234Rosario Morales,Jose Luis Mendez,Armando Navarro,Ana I. Gil,Haruo Watanabe, 451Masatomo Morita,G. Balakrish Nair,and Alejandro Cravioto*
1International Center for Diarrhoeal Disease Research, Bangladesh (ICDDR,B), Mohakhali, Dhaka 1212, Bangladesh; Department of 2Public Health, Faculty of Medicine, Universidad Nacional Auto?noma de Me?xico, Mexico, D.F., Mexico; Instituto de 3Investigacio?n Nutricional, Lima, Peru; National Institute of Infectious Diseases, Shinjuku-ku, Tokyo, 45Japan; and National Institute of Cholera and Enteric Diseases, Kolkata, India
Received 29 April 2010/Returned for modi；cation 9 June 2010/Accepted 17 July 2010
Vibrio cholerae O1 biotype El Tor (ET), the cause of the current 7th pandemic, has recently been replaced in Asia and Africa by an altered ET biotype possessing cholera toxin (CTX) of the classical (CL) biotype that originally caused the ;rst six pandemics before becoming extinct in the 1980s. Until recently, the ET prototype was the biotype circulating in Peru; a detailed understanding of the evolutionary trend of V. cholerae causing endemic cholera in Latin America is lacking. The present retrospective microbiological, molecular, and phylogenetic study of V. cholerae isolates recovered in Mexico (n 91; 1983 to 1997) shows the existence of the pre-1991 CL biotype and the ET and CL biotypes together with the altered ET biotype in both epidemic and endemic cholera between 1991 and 1997. According to sero- and biotyping data, the altered ET, which has shown predominance in Mexico since 1991, emerged locally from ET and CL progenitors that were found
coexisting until 1997. In Latin America, ET and CL variants shared a variable number of phenotypic markers, CLCLCLwhile the altered ET strains had genes encoding the CL CTX (CTX) prophage, ctxBand rstR, in addition ETto resident rstR, as the underlying regional signature. The distinct regional ;ngerprints for ET in Mexico and Peru and their divergence from ET in Asia and Africa, as con;rmed by subclustering patterns in a pulsed-;eld gel electrophoresis (NotI)-based dendrogram, suggest that the Mexico epidemic in 1991 may have been a local event and not an extension of the epidemics occurring in Asia and South America. Finally, the CL biotype
reservoir in Mexico is unprecedented and must have contributed to the changing epidemiology of global
cholera in ways that need to be understood.
Voges-Proskauer [VP] test results) (15) and genotypic traits In 1991, when cholera reemerged after being absent from (ctxB, acfB, tcpA, and rstR). In addition, the CL and ET biotype Latin America for about a century, millions of people were affected, with nearly 9,000 dying in 1993 alone (13). Following strains differ in two major genomic regions, namely, the Vibrio its ；rst appearance along the coast of Peru in January 1991, seventh pandemic pathogenicity island I (VSP-I) and VSP-II, cholera rapidly spread to all countries in Latin America except that are unique to the 7th pandemic ET biotype (10). Biotype Uruguay, reaching Mexico in June of the same year (13, 19, 25, CL caused the ；rst six of the seven cholera pandemics re- 32). Since then, there has been a great interest in understand- corded between 1817 and 1923 (26), with ；ve of these affecting ing the source and transmission of cholera in Latin America. the American continents. Ever since the ；rst cases were de- Limited genetic analysis of Vibrio cholerae O1 biotype El Tor tected in the Americas in the 1830s, endemic cholera contin- (ET) strains identi；ed from the epidemic showed that they ued to be prevalent until 1895 (18). have a unique signature, distinguishing them from 7th pan- Historically, cholera has been endemic in Asia for centuries demic ET strains (25). However, later studies showed the (26), with Asia being at the center of each of the seven cholera clonal nature of the bacterium, suggesting that the Latin pandemics (10, 26). Although the ET biotype was ；rst reported American epidemic was an extension of the 7th pandemic in 1905 and the 6th pandemic, caused by the CL biotype, lasted caused by ET strains from the Western Hemisphere (34). until 1923, it was not until the early 1960s that the ET biotype V. cholerae O1 and O139 are the two serogroups known to displaced the CL biotype in Asia and became the causal agent cause cholera. V. cholerae O1 has two biotypes, classical (CL) of the 7th pandemic (33). The CL biotype maintained a low and ET, which differ in speci；c phenotypic traits (hemolysis of pro；le in its last recognized niche in the coastal ecosystem of sheep erythrocytes, agglutination of chicken erythrocytes, sen- the Bay of Bengal, before it disappeared as a causal agent of sitivity to polymyxin B [PMB], phage susceptibilities, and clinical disease in the 1980s (33). Over the past few years, the ET biotype causing Asiatic cholera has shown remarkable changes in its phenotypic and * Corresponding author. Mailing address: ICDDR,B, GPO Box 128, genetic characteristics (23). Recent molecular analysis of ET Dhaka 1000, Bangladesh. Phone: 880-2-8823031. Fax: 880-2-8823116. strains causing acute watery diarrhea in Bangladesh shows E-mail: email@example.com. them to be hybrids because they possess phenotypic and ge- Published ahead of print on 28 July 2010.
VOL. 48, 2010 VARIANTS ASSOCIATED WITH CHOLERA IN MEXICO 3667
TABLE 1. Phenotypic and related genetic characteristics of V. cholerae O1 biotypes ET and CL and hybrid variants of both abiotypes isolated in Mexico from 1983 to 1997
Phenotypic properties Genetic screening by PCR Yr of Sensitivity No. of Deduced rstR isolation or biotype ctxA ctxB strains Source Serotype CCA strain PMB CL-speci；c ET-speci；c type b typetcpA rtxC (50 U) phage IV phage V 1 Clin Inaba CL CL CL S S R 1983 CL 1991 6 Clin Inaba R R S ET ET ET ET ET 1991 2 Clin Ogawa R R S CL ET ET ET, CL Alt-ET 1991 19 Clin Ogawa R R S CL ET ET ET, CL Alt-ET 1991 2 Clin Ogawa S R S CL ET ET ET, CL Alt-ET 1992 3 Clin Inaba R R S ET ET ET ET ET 1992 1 Env Inaba S R S ET ET ET ET ET 1992 1 Clin Inaba R R R ET ET ET ET ET c1992 1Env Inaba S R R CL CL CL 1993 6 Clin Ogawa R R S CL ET ET ET, CL Alt-ET 1993 1 Clin Ogawa S R S ET ET ET ET 1993 1 Clin Ogawa S R S CL ET ET ET, CL Alt-ET 1994 1 Clin Inaba R R S ET ET ET ET 1994 1 Env Inaba R R S CL ET ET ET, CL Alt-ET 1995 1 Clin Ogawa S R R CL CL CL CL 1995 1 Clin Inaba R R S ET ET ET ET ET 1995 1 Env Inaba R R R CL CL CL CL 1995 1 Clin Inaba R R S CL ET ET ET, CL Alt-ET 1995 2 Clin Ogawa R R S CL ET ET ET, CL Alt-ET 1995 1 Env Ogawa R R R CL ET ET ET, CL Alt-ET 1995 1 Clin Inaba S R S CL ET ET ET, CL Alt-ET 1997 8 Clin Ogawa R R S CL ET ET ET, CL Alt-ET 1997 26 Clin Ogawa R R S CL ET ET ET, CL Alt-ET 1997 1 Env Inaba S S R CL CL CL CL 1997 1 Clin Inaba S S R CL CL CL CL 1997 1 Env Inaba R R S CL ET ET ET, CL Alt-ET N16961 Control Inaba R R S ET ET ET ET El Tor O395 Control Ogawa S S R CL CL CL CL
a A total of 91 strains were evaluated. Abbreviations: Clin, clinical; Env, environmental; R, resistant; S, sensitive; Alt, altered. b Determined by MAMA-PCR (21). c Serologically nonreactive to antibodies speci；c for O1 or O139.
MATERIALS AND METHODS notypic traits of the CL biotype against an ET background (23). Subsequent retrospective studies showed that all of the Bacterial strains. The V. cholerae serogroup O1 strains characterized and
compared in the present study are shown in Table 1, together with their sources O1 ET strains isolated in Bangladesh since 2001 have been and years of isolation. hybrids of both the CL and ET biotypes, while those isolated Con;rmation of V. cholerae strain types. The Latin American isolates of V. before 2001 contained all the attributes of the 7th pandemic V. cholerae used in the present study were obtained from the Department of Public cholerae O1 ET strains (22). Such genetic changes among ET Health, Faculty of Medicine, National Autonomous University of Mexico (UNAM). These strains were of patient and surface water origin and were strains causing cholera in Latin America were also evident isolated as part of the nationwide cholera surveillance conducted between 1983 from the changing serotypes, electrophoretic types, ribotypes, and 1997 (3). The bacterial strains were shipped in soft agar, their identities were and pulsed-；eld gel electrophoresis (PFGE) types (3, 9, 11, con；rmed by standard culture methods, and their serogroups and biotypes were 27). While the ET prototype has been completely replaced by identi；ed by a combination of biochemical and molecular methods, as described previously (1). an altered ET in Asia and Africa (29), recent data show the Serogrouping. The serogroups of the V. cholerae isolates that were identi；ed nature of the ET prototype (7th pandemic) of V. cholerae O1 using biochemical and molecular methods were con；rmed serologically by a slide isolated in Peru between 1991 and 2003 (24). The Peruvian ET agglutination test using speci；c polyvalent antisera to V. cholerae O1 and O139, strains that seem to be closely linked clonally to the Asian and followed by a monoclonal antibody that is speci；c for both serogroups (1). Biotyping. Biotyping involved a number of phenotypic tests: chicken erythro- African ET prototype strains were shown to have a distinct cyte agglutination, sensitivity to polymyxin B, and Mukerjee CL phage IV and region in VSP-II that differentiates them from the ET proto- Mukerjee ET phage V tests (15). To complement the biotype characterization by type strains isolated in other continents (24). To better under- phenotypic traits, PCR assays were carried out using previously described pro- stand the dynamics of the cholera epidemic in Latin America, cedures that were targeted to detect tcpA (CL and ET variants) and the type of the present study analyzed 91 V. cholerae O1 strains isolated the rstR gene encoding the phage transcriptional regulator (16). Storage of strains. V. cholerae strains whose types were con；rmed by biochem- from both clinical and environmental sources in Mexico be- ical, serological, and molecular methods were subcultured on gelatin agar (GA) tween 1983 and 1997. V. cholerae O1 ET strains from Peru plates; and a single representative colony from the GA was aseptically inoculated (1991 to 1999), Bangladesh (1985 to 2007), and Zambia (1996 into T1N1 broth (1% Trypticase, 1% NaCl), incubated at 37?C for 3 to 4 h, and to 2004) were also included for comparative purposes. stored at 80?C with 15% glycerol until it was required.
3668 ALAM ET AL. J. CLIN. MICROBIOL.
Genomic DNA preparation. Genomic DNA extraction was carried out follow- the V. cholerae strains showing biotype ET-speci；c phenotypic ing previously described methods (24). traits, such as chicken cell agglutination (CCA), sensitivity to Con;rmation of serogrouping by PCR assays. The subtypes of all the strains ET-speci；c phage V, and resistance to both PMB and CL- were recon；rmed using a V. cholerae species-speci；c ompW PCR (24). The speci；c phage IV, were primarily recognized to be biotype ET serogroups of these strains were recon；rmed using polyvalent O1 and monova- lent Inaba and Ogawa antisera and by multiplex PCR targeted to identify genes (Table 1). Three of the CCA-negative strains matched biotype encoding O1 (wbe) and O139 (wbf)-speci；c O biosynthetic genes and the cholera CL reference strain O395 in being sensitive to both PMB and toxin (CTX) gene (ctxA) (14). CL-speci；c phage IV but not to ET-speci；c phage V, thus MAMA-PCR for determination of ctxB gene type. The mismatch ampli；cation being recognized primarily as biotype CL (Table 1). Most of mutation assay (MAMA) was recently developed to detect the sequence poly- CLETmorphism between the CL and ET ctxB genes (ctxBand ctxB, respectively) the V. cholerae strains were phenotypic variants of ET because by focusing on nucleotide position 203 of the ctxB gene (21). MAMA-PCR was they had major ET traits but failed to show all the properties performed to test for the presence of the ctxB genes speci；c for the CL and ET typical of the ET reference strain, N16961 (Table 1). biotypes. A conserved forward primer (Fw-con, 5 -ACTATCTTCAGCATATG Although V. cholerae O1 strains in Mexico varied signi；- CACATGG-3 ) and two allele-speci；c polymorphism detection primers, Rv-cla (5 -CCTGGTACTTCTACTTGAAACG-3 ) and Rv-elt (5 -CCTGGTACTTCT cantly in their biotype-speci；c, phenotypic markers, 85 of the ACTTGAAACA-3 ), were used. PCR conditions were as follows: after initial 91 phenotypically biotype ET strains ampli；ed the primers for denaturation at 96?C for 2 min, 25 cycles of denaturation at 96?C for 10 s, the rtxC gene, an ET-speci；c marker, con；rming their ET trait. annealing at 50?C for 10 s, extension at 72?C for 30 s, and a ；nal extension at 72?C Fourteen of the strains that were primarily identi；ed to be ET for 2 min. The resulting V. cholerae O1 isolates, O395 CL and N16961 ET, were ETETampli；ed the primers for the tcpAand rstRgenes but not used as standard reference strains. CLCLPFGE. The whole agarose-embedded genomic DNA from V. cholerae was for the tcpAand rstRgenes, further con；rming their typ- prepared. PFGE was carried out with a contour-clamped homogeneous electrical ical ET identity (Table 1). The remaining 71 ET strains were ；eld (CHEF-DRII) apparatus (Bio-Rad), according to procedures described variants, since they possessed various numbers of major phe- elsewhere (6). The conditions used for separation were as follows: 2 to 10 s for notypic markers for the CL biotype, in addition to having 13 h, followed by 20 to 25 s for 6 h. An electrical ；eld of 6 V/cm was applied at an included ；eld angle of 120?. Genomic DNAs of the test strains were digested major ET traits and amplifying the primers for the ET-speci；c by the NotI restriction enzyme (Gibco-BRL, Gaithersburg, MD), and Salmonella ETETalleles of the tcpAand rstRgenes (Table 1). All of these enterica serovar Braenderup was digested by XbaI, with the fragments being used ET variants also ampli；ed the gene encoding the CTX pro- as molecular size markers. The restriction fragments were separated in 1% CLETphage, rstR, in addition to amplifying the resident rstR2, pulsed-；eld-certi；ed agarose in 0.5 TBE (Tris-borate-EDTA) buffer. In the postelectrophoresis gel treatment step, the gel was stained and destained. The and were primarily designated altered ET. DNA was visualized using a UV transilluminator, and images were digitized via In sharp contrast, the template DNA of the six remaining V. a one-dimensional gel documentation system (Bio-Rad). cholerae strains, most of which varied in their major phenotypic Image analysis. The ；ngerprint pattern in the gel was analyzed using a com- traits, including the one that was not recognized by O1- or puter software package, Bionumeric (Applied Maths, Belgium). After back- ground subtraction and gel normalization, the ；ngerprint patterns were sub- O139-speci；c antibodies but that reacted to Inaba antisera, did jected to typing on the basis of banding similarity and dissimilarity using Dice not amplify the primers for biotype ET-speci；c marker gene similarity coef；cient and unweighted-pair group method using average linkages rtxC, suggesting possible CL attributes (Table 1). Of these, the (UPGMA) clustering methods, as recommended by the manufacturer; these ；ve serologically O1 strains ampli；ed the primers for the were graphically represented as dendrograms. CLCLETtcpAand rstRgenes but not those for the tcpAand ET rstRgenes. This con；rms that these ；ve strains, one of which
RESULTS was a pre-1991 strain isolated in Mexico, belong to the CL
biotype, which disappeared as a cause of cholera in the 1980s. Microbiological and serological tests. All tested strains (n The serologically non-O1/O139 strain, which possessed certain 91) produced characteristic colonies typical of V. cholerae when CL attributes, such as being sensitive to both CCA and PMB, they were grown on taurocholate tellurite gelatin agar despite being resistant to both CL- and ET-speci；c phages, (TTGA). The resulting colonies gave biochemical reactions CLampli；ed the primers for the biotype-speci；c gene tcpAal- typical of V. cholerae, and all except one strain reacted to the ETlele but not for the tcpAallele or any of the primers for the monoclonal antibody speci；c for O1 but not to that for O139. gene encoding CTX phage, ctxA or rstR. These results suggest All V. cholerae O1 strains, including the serologically nonre- that the serologically non-O1/O139 strain may have originally acting strain, reacted to monovalent Inaba and Ogawa anti- been toxigenic V. cholerae O1 and also biotype CL but had its sera, suggesting that all belonged to serogroup O1. CTX prophage excised. Ampli;cation of primers speci;c for V. cholerae serogroup MAMA-PCR. MAMA-PCR using primers speci；c for the O1 and ctxA by PCR assays. All strains (n 91) ampli；ed the CL or ET biotype offers a precise and accurate method for primers for the V. cholerae species-speci；c gene ompW, and all determining the type of CTX that V. cholerae produces. The V. except the strain serologically determined to be non-O1/O139 cholerae O1 strains, including the CL (O395) and ET (N16961) ampli；ed the primers speci；c for the O biosynthesis gene wbe reference strains, were analyzed to determine their CTX type but not those speci；c for wbf. In addition, all the strains sero- by MAMA-PCR. As shown in Table 1, V. cholerae O1 strains logically con；rmed to be O1 ampli；ed the primers for the
that were primarily identi；ed to be ET ampli；ed the ET- cholera toxin gene ctxA, con；rming that all strains were ctx- ETCLspeci；c ctxBgene but not the ctxBgene, con；rming their positive V. cholerae O1.
prototype ET traits. In contrast, all of the ET strains possessing Phenotypic and related genotypic characteristics. The re- CLCLthe gene encoding CTXphage, rstR, in other words, the sults of the major phenotypic and related genetic characteriza-
tions of the V. cholerae O1 strains (n 91) are presented in altered ET strains, including the ；ve CL biotype strains, am- CLETTable 1. The V. cholerae O1 strains varied greatly in their pli；ed the ctxBgene but not the ctxBgene, again con；rm-
phenotypic and related genetic characteristics. Twenty-two of ing their CTX to be that of the CL biotype. However, the
VOL. 48, 2010 VARIANTS ASSOCIATED WITH CHOLERA IN MEXICO 3669
FIG. 1. Dendrogram showing genomic ；ngerprints of V. cholerae O1 strains isolated in Mexico (1983 to 1997), North America. The dendrogram was prepared by Dice similarity coef；cient and UPGMA clustering using the PFGE patterns of NotI-digested genomic DNA. The scale bar at the top shows the correlation coef；cient (%). Two major clusters separating the CL biotype from the ET biotype strains show the respective lineages. The CL cluster, which includes the North American CL strains with the CL reference control strain (strain 569B), shows high degrees of divergence among the strains. The major ET cluster shows two subclusters that separate the prototype ET (Pro-ET) (including reference control strain N16961) from the altered ET (Alt-ET) strains, suggesting two different lineages for them. IN, Inaba; OG, Ogawa; Clin, clinical; Env, environ- mental; Edo., estado (state).
serologically non-O1/O139 strain failed to amplify any of the PFGE and cluster analysis. The NotI-digested genomic
two biotype-speci；c alleles by MAMA-PCR. DNAs of both clinical and environmental V. cholerae O1
The altered ET that has been con；rmed in Mexico along strains of different spatiotemporal origins and the resulting with the progenitor ET and CL biotype strains presented direct different biotype categorizations, including one pre-1991 CL evidence of a genetic transition from the ET prototype to the strain isolated in Mexico in 1983, were subjected to PFGE to altered ET. Of the seven environmental strains analyzed in the determine their genetic relatedness and clonal origin. The present study, three, including the serologically nonreacting NotI restriction enzyme digested the genomic DNA of both the strain, were identi；ed to be of the CL biotype, one was iden- test and the reference control strains into 20 to 23 fragments ti；ed to be of the ET biotype, and the remaining three were (Fig. 1); the molecular sizes of the DNA fragments ranged identi；ed to be of the altered ET biotype (Table 1), suggesting from 20 to 350 kb. V. cholerae O1 strains of the ET and CL
an aquatic reservoir for them. biotypes presented overall banding patterns that were charac-
3670 ALAM ET AL. J. CLIN. MICROBIOL.
teristic of their respective ET and CL reference controls (Fig. DISCUSSION
1). The serologically non-O1 but genotypically con；rmed O1 The 1991 epidemic in Peru was shown to be a clonal expan- CL strain had the signature banding pattern of the CL refer- sion of the 7th ET biotype pandemic currently being seen in ence control, suggesting a CL ancestry. Asia (3, 5, 20). A signi；cant and recent development has been Cluster analysis, which was performed with dendrograms the emergence of an altered ET possessing the gene encoding (prepared by Dice similarity coef；cient and UPGMA cluster- the CTX that is speci；c for the CL biotype. This is interesting, ing methods) obtained from the PFGE patterns of NotI-di- because the CL biotype caused the ；rst six pandemics before gested genomic DNA, separated both the CL biotype strains being replaced by the ET biotype, and therefore, the CL bio- (cluster A) from the ET biotype strains (cluster B) and the type was considered to be extinct or to have at least disap- altered ET biotype strains from the progenitor ET biotype peared as a cause of clinical disease. While the altered ET has strains, irrespective of their source (Fig. 1). This analysis fur- already replaced the prototype 7th pandemic ET in Asia (11) ther con；rmed the biotype categorization of V. cholerae O1 and Africa (17, 21), according to a recent study, the ET causing strains made in the present study, thereby indicating their endemic cholera in Peru up until 2003 was of the 7th pandemic genetic ；ngerprints. A great deal of variation was observed in ET prototype (18). The present retrospective study of the the clustering patterns of the genetically patchy strains belong- phenotypic, molecular, and phylogenetic characterizations of ing to the two biotypes. Subclustering was identi；ed in the V. cholerae O1 isolated in Mexico shows that the CL biotype major biotype-speci；c clusters, which separated the prototype was circulating in Mexico before and after 1991. The study also CL strains from their variants and the progenitor ET (cluster reveals that both the ET and CL biotypes, together with their B1) from the altered ET (cluster B2), suggesting that each hybrid variants, were consistently present in the aquatic envi- followed a different lineage. Both clinical and environmental ronment and were the causal agents for the 1991 epidemic and strains of a particular biotype category clustered together, sug- the subsequent endemic cholera in Mexico. gesting that they are clonally related. The results of the microbiological culture and the biochem- To further understand the clonal link between the V. chol- ical and serological tests primarily con；rmed that the V. chol-
erae O1 strains occurring in Latin America, a separate cluster erae isolates in Mexico were serogroup O1, while one V. chol-
analysis was performed using the PFGE (NotI) images of the erae strain did not react serologically to antiserum speci；c for
genomic DNA of O1 strains isolated in Mexico, together with O1 or O139. These primary microbiological results were fur- those of selected ET strains causing endemic cholera in Peru ther complemented by simplex and multiplex PCR assays,
which included V. cholerae species-speci；c gene (ompW) ctxA (1991 to 1999). As shown in Fig. 2, two major biotype-speci；c
(encoding subunit A of CTX) and the gene encoding sero- clusters separated the CL biotype strains (cluster A) from the
group O1 (wbe) (14), further con；rming that the V. cholerae ET biotype strains (cluster B), irrespective of their spatiotem-
strain in Mexico was toxigenic and belonged to serogroup O1 poral origin, suggesting biotype-speci；c lineages for them. (4, 19). The serologically nonreacting environmental strain Subclustering was observed under the major ET cluster (clus- that reacted to monovalent Inaba antiserum and that ampli；ed ter B), which separated the Peruvian ET strains (cluster B1) the primers for the ompW and wbe genes (14) but not the from the prototype ET strains (cluster B2) and the altered ET primer for ctxA suggested that it was ctx-negative V. cholerae strains (cluster B3) of Mexico, showing divergence and re- O1. The ctx-negative V. cholerae O1 strain, which occurs in gional signatures. aquatic environments, has been shown to arise following the Since the ET prototype causing the 1991 epidemic in loss of the CTX prophage (1). Mexico showed a regional signature different from that of V. cholerae O1 biotypes CL and ET can be distinguished by the ET biotype that caused the Peruvian epidemic at the the differences in their phenotypic and genotypic characteris- same time, cluster analysis was performed by dendrogram tics (10, 15). The results of the phenotypic and genetic screen- analysis with the PFGE images of the V. cholerae O1 strains ing in the present study indicate that V. cholerae O1 strains obtained in the present study to compare them with those of circulating in Mexico included both CL and ET strains to- the CL and ET (altered ET) strains of Asia and Africa. gether with the variants of both biotypes. This ；nding is un- Again, two major clusters (Fig. 3) separated the CL strains precedented, since it shows that the CL biotype was involved in (cluster A) from the ET strains (cluster B), irrespective of endemic cholera caused by biotype ET at a time when the CL their origin. However, a great deal of divergence between biotype was thought to be extinct (33). Furthermore, until the strains of the two biotypes was observed, as seen in the recently, only strains of the ET prototype, such as those iso- similarity indices. Despite being divergent, the pre- and lated in Peru, had been shown to be the cause of the Latin post-1991 prototype CL strains isolated in Mexico clustered American epidemic, where the CL biotype has never been together with the Asian CL strains (cluster A), suggesting detected (4, 19, 34). Therefore, Mexico was clearly an impor- the same lineage for all of these strains. The ET cluster tant regional habitat for V. cholerae O1 strains that were dif- (cluster B), which also displayed considerable heterogene- ferent from those isolated in South America at the same time. ity, underwent further subclustering that separated the al- The pre- and post-1991 CL biotype reservoir in Mexico must tered ET strains of Mexico (cluster B1) from those of Asia have contributed to the changing epidemiology of global chol- and Africa (cluster B4) and that separated the ET prototype era in ways that still need to be understood. strain of Mexico (cluster B2) from that of Peru (cluster B3), In the present study, the altered nature of the ET strains suggesting type-speci；c lineages and regional signatures for isolated in Mexico was con；rmed, showing that their ctxB gene
V. cholerae O1. was of the CL biotype, the prototype of which was reported in
VOL. 48, 2010 VARIANTS ASSOCIATED WITH CHOLERA IN MEXICO 3671
FIG. 2. Dendrogram showing genomic ；ngerprints of V. cholerae O1 isolated in Mexico (1983 to 1997) and Peru (1991 to 1998), Latin America. The dendrogram (prepared by Dice similarity coef；cient and UPGMA clustering) was based on the PFGE (NotI) images of genomic DNA. The two major clusters, showing the degree of similarity (%), separated the CL from the ET biotype strains, suggesting respective biotype-speci；c lineages. The CL biotype cluster, which includes strains exhibiting signature PFGE patterns for CL reference control strain 569B, shows high degrees of divergence among the strains. The major ET cluster shows subclusters, separating the Peruvian prototype ET (ET1) subcluster from the rest while separating the Mexican prototype ET (ET2) subcluster from the Mexican altered ET (Alt-ET), suggesting regional and type-speci；c ；ngerprints. IN, Inaba; OG, Ogawa; Clin, clinical; Env, environmental; Edo., estado (state).
3672 ALAM ET AL. J. CLIN. MICROBIOL.
VOL. 48, 2010 VARIANTS ASSOCIATED WITH CHOLERA IN MEXICO 3673
Bangladesh (22, 31). Therefore, the question is, why did the differed from the variants in Asia (30) and Africa (2), with all CLPeruvian V. cholerae O1 ET strains not switch from the pro- the Latin American altered ET strains having ctxBand CLETtotype (24) to the altered type until recently, considering that rstRgenes, in addition to the resident rstRgene. There-
the ET strains, such as those in Asia, Africa (2, 28, 29), and fore, these results show the unique regional signature for the Mexico, did so in as early as 1991? The answer may not be Latin American altered ET strains, further substantiating their precise, but it is likely that this change that took place in independent emergence from the coexisting ET and CL bio- Mexico, but not in Peru (24), may have been prompted by the type strains in Mexico.
copresence of the CL and ET biotype strains in the aquatic Originally, the V. cholerae O1 ET that caused the 1991 ep- environment, which was reported in Bangladesh up until the idemic in Latin America was shown to be homogeneous (4, 19, 1980s (33) but which was a phenomenon not seen in Peru (24). 27). It was only later that divergence in other serotypes was There are two different possibilities for the emergence and reported through the establishment of different electro- global spread of the altered ET biotype in Asia and Africa (28, phoretic types, ribotypes, and PFGE types (3, 9, 11, 22); and 29). The ；rst one is that altered ET was the result of clonal this has been corroborated more recently, with heterogeneity expansion of a single ancestral ET that had acquired the ctxB being reported among the ET strains isolated in Peru (24). gene of the CL biotype in one of the regions where cholera is Differences in the PFGE patterns of the CL and ET biotype endemic. The second one is that the emergence of the altered strains and their variant strains were also observed in Mexico ET was not a single event but consisted of multiclonal events in the present study. Although the genetic basis for this diver- occurring independently in each region. Direct evidence sup- gence remains unknown, it may be due to CL and ET biotype porting these hypotheses is lacking, however. A recent study strains sharing a niche in the aquatic ecosystem, as demon- has proposed that the transition took place between 1990 and strated in the present study, which allows interbiotype and 1994 for Southeast Asia, with all O1 ET strains isolated before cross-serotype genetic recombination, resulting in genetic re- 1990 being of the ET prototype (28). The present study is the assortment (3, 9, 11).
；rst to provide direct evidence of the altered ET occurring with V. cholerae O1 ET strains causing endemic cholera in Peru an array of other minor phenotypic variant strains of both have been shown to be different from the 7th pandemic ET biotypes together with the progenitor CL and ET strains that strains in Asia (24, 25). The subclustering in the PFGE (NotI)- were involved in the 1991 epidemic and the subsequent en- based dendrogram in the present study, indicating the separa- demic cholera in Mexico until 1997. The background epidemi- tion of the Mexican ET from the Peruvian ET and the Asian ological data and the patient selection criteria were not avail- and African altered ET from the Latin American altered ET, able for the clinical V. cholerae strains included in the present further con；rms the presence of speci；c regional signatures.
retrospective study; therefore, the possibility that the overrep- These regional DNA ；ngerprints suggest that the 1991 epi- resented altered ET (of the Ogawa serotype) strains in 1991 demic in Mexico may have been a local event rather than an and 1997 were not disproportionately picked up from cases in extension of the epidemics that occurred in Peru and Asia at well-de；ned time-space clusters cannot be ruled out. Nonethe- that time (3, 19, 24, 25, 27, 34).
less, the evidence of a genetic shift from prototype ET to While the Latin American epidemic was considered to be altered ET in Latin America provided here suggests that this imported via bilge water from ships coming from areas in Asia was a local event that occurred in Mexico and that was unre- or Africa where cholera is endemic (20), sporadic but consis- lated to what was happening in Asia and Africa at the time. tent cases of cholera on the Gulf Coasts of the United States The hybrid variants of V. cholerae O1, designated Matlab and Mexico between 1965 and 1991 suggest a local source (4, variants, which were biotype ET but which had the phenotypic 35). Following a cholera case that was reported in Cancun, and genotypic traits of CL biotypes, were ；rst reported from Mexico, in 1983 (3), immediate studies failed to isolate any V.
patients with acute secretory diarrhea in Matlab, Bangladesh cholerae O1 strains, but diarrhea caused by non-O1 strains was (23), and then in Mozambique (2). Such genetic ET variants shown to be highly prevalent (12). However, these results do CLreported from Bangladesh had the ctxBgene in combination not rule out a pre-1991 niche for V. cholerae in Latin America, ETCLETCLas the serogroup O1 strains, which seldom respond to culturing with either rstRonly, rstRonly, or both rstRand rstR,
media (7), can arise from non-O1 through seroconversion (5). irrespective of whether their tcpA alleles were of the CL or the
This assumption may be further substantiated by con；rmation ET biotype (30). The ET variants in Mozambique, on the other CLCLof the V. cholerae O1 ET, CL, and altered ET biotypes, includ- hand, had only the CTXprophage-related genes, ctxBand CLing the serologically non-O1 but phenotypically and genetically rstR, in pairs (2). The altered ET in the present study may be
an analogue of the Matlab variant (23, 30) or Mozambique O1 CL variant isolated from the aquatic environments of Mex-
ico in the present study, con；rming their aquatic reservoir and variants (2) of V. cholerae O1 ET strains, although the former
FIG. 3. Dendrogram showing genomic ；ngerprints of V. cholerae O1 isolates recovered in Mexico (North America) and Peru (South America) and their comparison with V. cholerae O1 isolates recovered in Bangladesh (Asia) and Zambia (Africa). Dendrogram (prepared by Dice similarity coef；cient and UPGMA clustering) was based on the PFGE (NotI) patterns of genomic DNA. Two major clusters that separated the CL from the ET biotype strains, showing the degree of similarity (%), suggest biotype-speci；c lineages. The cluster CL includes the genetically divergent Latin American pre- and post-1991 CL strains with Asian CL strains and their reference control strain (strain 569B), suggesting the same clonal lineage for them. The major ET cluster divided the strains into four subclusters, separating the Mexican altered ET (Alt-ET1) strains from the altered ET of Asia and Africa (Alt-ET2) and Mexican prototype ET (ET1) strains from prototype ET (ET2) of Peru, depicting different signatures. IN, Inaba; OG, Ogawa; Clin, clinical; Env, environmental; Edo., estado (state).
3674 ALAM ET AL. J. CLIN. MICROBIOL.
Sepulveda, D. Bessudo, and P. A. Blake. 1987. Non-O1 Vibrio cholerae showing that non-O1 serogroup strains can also arise from O1 infections in Cancun, Mexico. Am. J. Trop. Med. Hyg. 36:393–397. serogroup strains. Therefore, even if the ecological niche for V. 13. Guthmann, J. P. 1995. Epidemic cholera in Latin America: spread and cholerae is not well de；ned for Latin America, the rapid spread routes of transmission. J. Trop. Med. Hyg. 98:419–427. 14. Hoshino, K., S. Yamasaki, A. K. Mukhopadhyay, S. Chakraborty, A. Basu, of cholera in 1991 was presumably due to a preexisting, small S. K. Bhattacharya, G. B. Nair, T. Shimada, and Y. Takeda. 1998. Devel- population of plankton-bound O1 that exists mainly in coastal opment and evaluation of a multiplex PCR assay for rapid detection of waters (4, 7, 35). The pre-1991 existence of the CL biotype, as toxigenic Vibrio cholerae O1 and O139. FEMS Immunol. Med. Microbiol. 20:201–207. con；rmed in the present study, and its copresence with the 15. Kaper, J. B., J. G. Morris, Jr., and M. M. Levine. 1995. Cholera. Clin. prototype ET biotype and the phenotypic and genetic variants Microbiol. Rev. 8:48–86. of both biotypes between 1991 and 1997 in Mexico may indi- 16. Kimsey, H. H., G. B. Nair, A. Ghosh, and M. K. Waldor. 1998. Diverse CTXphis and evolution of new pathogenic Vibrio cholerae. Lancet 352:457– cate a persistent niche for V. cholerae in Latin America. 458. The association of V. cholerae with plankton in the coastal 17. Lizarraga-Partida, M. L., and M. L. Quilici. 2009. Molecular analyses of Vibrio cholerae O1 clinical strains, including new nontoxigenic variants iso- waters of Peru and Mexico has been well documented (7, 8, lated in Mexico during the cholera epidemic years between 1991 and 2000. 18). Although V. cholerae strains isolated in Mexico after 1997 J. Clin. Microbiol. 47:1364–1371. were not available for analysis in the present study, a recent 18. Lizarraga-Partida, M. L., E. Mendez-Gomez, A. M. Rivas-Montano, E. Var- gas-Hernandez, A. Portillo-Lopez, A. R. Gonzalez-Ramirez, A. Huq, and report shows the involvement of serogroup O1 in endemic R. R. Colwell. 2009. Association of Vibrio cholerae with plankton in coastal cholera in the Gulf of Mexico coast after 1997 (17). Although areas of Mexico. Environ. Microbiol. 11:201–208. this publication does not include the full genetic characteristics 19. Mata, L. 1994. Cholera El Tor in Latin America, 1991–1993. Ann. N. Y. Acad. Sci. 740:55–68. of V. cholerae O1 strains isolated after 1997, it does indicate a 20. McCarthy, S. A., and F. M. Khambaty. 1994. International dissemination of change in the ribosomal pattern that separates strains into two epidemic Vibrio cholerae by cargo ship ballast and other nonpotable waters. Appl. Environ. Microbiol. 60:2597–2601. different groups before and after 1997. However, it would not 21. Morita, M., M. Ohnishi, E. Arakawa, N. A. Bhuiyan, S. Nusrin, M. Alam, be possible here to predict if the CL and ET progenitors that A. K. Siddique, F. Qadri, H. Izumiya, G. B. Nair, and H. Watanabe. 2008. were found coexisting with the variant strains of both biotypes Development and validation of a mismatch ampli；cation mutation PCR assay to monitor the dissemination of an emerging variant of Vibrio cholerae until 1997 still exist or were replaced by the altered ET that O1 biotype El Tor. Microbiol. Immunol. 52:314–317. predominated in Mexico. 22. Nair, G. B., F. Qadri, J. Holmgren, A. M. Svennerholm, A. Safa, N. A. Bhuiyan, Q. S. Ahmad, S. M. Faruque, A. S. Faruque, Y. Takeda, and D. A. Sack. 2006. Cholera due to altered El Tor strains of Vibrio cholerae O1 in ACKNOWLEDGMENTS Bangladesh. J. Clin. Microbiol. 44:4211–4213. This research was partially supported by NIID, Tokyo, Japan, and 23. Nair, G. B., S. M. Faruque, N. A. Bhuiyan, M. Kamruzzaman, A. K. Sid- dique, and D. A. Sack. 2002. New variants of Vibrio cholerae O1 biotype El the ICDDR,B. The ICDDR,B is supported by donor countries and Tor with attributes of the classical biotype from hospitalized patients with agencies, which provide unrestricted support to the center for its op- acute diarrhea in Bangladesh. J. Clin. Microbiol. 40:3296–3299. eration and research. 24. Nusrin, S., A. I. Gil, N. A. Bhuiyan, A. Safa, M. Asakura, C. F. Lanata, E. Hall, H. Miranda, B. Huapaya, G. C. Vargas, M. A. Luna, D. A. Sack, S. REFERENCES Yamasaki, and G. B. Nair. 2009. Peruvian Vibrio cholerae O1 El Tor strains 1. Alam, M., M. Sultana, G. B. Nair, A. K. Siddique, N. A. Hasan, R. B. Sack, possess a distinct region in the Vibrio seventh pandemic island-II that dif- D. A. Sack, K. U. Ahmed, A. Sadique, H. Watanabe, C. J. Grim, A. Huq, and ferentiates them from the prototype seventh pandemic El Tor strains. R. R. Colwell. 2007. Viable but nonculturable Vibrio cholerae O1 in bio；lms J. Med. Microbiol. 58:342–354. in the aquatic environment and their role in cholera transmission. Proc. Natl. 25. Olsvik, O. 1992. The cholera epidemic in Latin America. Tidsskr. Nor. Acad. Sci. U. S. A. 104:17801–17806. Laegeforen 2:1843–1846. (In Norwegian.) 2. Ansaruzzaman, M., N. A. Bhuiyan, B. G. Nair, D. A. Sack, M. Lucas, J. L. 26. Politzer, R. 1959. Cholera. WHO Monograph Series No. 43. World Health Deen, J. Ampuero, and C. L. Chaignat. 2004. Cholera in Mozambique, Organization, Geneva, Switzerland. variant of Vibrio cholerae. Emerg. Infect. Dis. 10:2057–2059. 27. Popovic, T., C. Bopp, O. Olsvik, and K. Wachsmuth. 1993. Epidemiologic 3. Beltran, P., G. Delgado, A. Navarro, F. Trujillo, R. K. Selander, and A. application of a standardized ribotype scheme for Vibrio cholerae O1. J. Clin. Cravioto. 1999. Genetic diversity and population structure of Vibrio cholerae. Microbiol. 31:2474–2482. J. Clin. Microbiol. 37:581–590. 28. Raychoudhuri, A., T. Patra, K. Ghosh, T. Ramamurthy, R. K. Nandy, Y. 4. Blake, P. A. 1994. Endemic cholera in Australia and United States, p. Takeda, G. Balakrish-Nair, and A. K. Mukhopadhyay. 2009. Classical ctxB 309–319. In I. K. Wachsmuth, P. A. Blake, and O. Olsvik (ed.), Vibrio in Vibrio cholerae O1, Kolkata, India. Emerg. Infect. Dis. 15:131–132. cholerae and cholera: molecular to global perspectives. ASM Press, Wash- 29. Safa, A., J. Sultana, P. Dac Cam, J. C. Mwansa, and R. Y. Kong. 2008. Vibrio ington, DC. cholerae O1 hybrid El Tor strains, Asia and Africa. Emerg. Infect. Dis. 5. Blokesch, M., and G. K. Schoolnik. 2007. Serogroup conversion of Vibrio 14:987–988. cholerae in aquatic reservoirs. PLoS Pathog. 3:e81. 30. Safa, A., N. A. Bhuiyan, M. Alam, D. A. Sack, and G. B. Nair. 2005. Genomic 6. Cameron, D. N., F. M. Khambaty, I. K. Wachsmuth, R. V. Tauxe, and T. J. relatedness of the new Matlab variants of Vibrio cholerae O1 to the classical Barrett. 1994. Molecular characterization of Vibrio cholerae O1 strains by and El Tor biotypes as determined by pulsed-；eld gel electrophoresis. pulsed-；eld gel electrophoresis. J. Clin. Microbiol. 32:1685–1690. J. Clin. Microbiol. 43:1401–1404. 7. Colwell, R. R., and W. M. Spira. 1992. The ecology of Vibrio cholerae and 31. Safa, A., N. A. Bhuyian, S. Nusrin, M. Ansaruzzaman, M. Alam, T. Ham- cholera, p. 107–127. In D. Barua and W. B. Greenough III (ed.), Cholera. abata, Y. Takeda, D. A. Sack, and G. B. Nair. 2006. Genetic characteristics Plenum Publishing Corp., New York, NY. of Matlab variants of Vibrio cholerae O1 that are hybrids between classical 8. Craig, J. P. 1996. Cholera: outlook for the twenty-；rst century. Caduceus and El Tor biotypes. J. Med. Microbiol. 55:1563–1569. 12:25–42. 32. Seas, C., J. Miranda, A. I. Gil, R. Leon-Barua, J. Patz, A. Huq, R. R. Colwell, 9. Dalsgaard, A., M. N. Skov, O. Serichantalergs, P. Echeverria, R. Meza, and and R. B. Sack. 2000. New insights on the emergence of cholera in Latin D. N. Taylor. 1997. Molecular evolution of Vibrio cholerae O1 strains isolated America during 1991: the Peruvian experience. Am. J. Trop. Med. Hyg. in Lima, Peru, from 1991 to 1995. J. Clin. Microbiol. 35:1151–1156. 62:513–517. 10. Dziejman, M., E. Balon, D. Boyd, C. M. Fraser, J. F. Heidelberg, and J. J. 33. Siddique, A. K., A. H. Baqui, A. Eusof, K. Haider, M. A. Hossain, I. Bashir, Mekalanos. 2002. Comparative genomic analysis of Vibrio cholerae: genes and K. Zaman. 1991. Survival of classic cholera in Bangladesh. Lancet that correlate with cholera endemic and pandemic disease. Proc. Natl. Acad. 337:1125–1127. Sci. U. S. A. 99:1556–1561. 34. Wachsmuth, I. K., G. M. Evins, P. I. Fields, O. Olsvik, T. Popovic, C. A. 11. Evins, G. M., D. N. Cameron, J. G. Wells, K. D. Greene, T. Popovic, S. Bopp, J. G. Wells, C. Carrillo, and P. A. Blake. 1993. The molecular epide- Giono-Cerezo, I. K. Wachsmuth, and R. V. Tauxe. 1995. The emerging miology of cholera in Latin America. J. Infect. Dis. 167:621–626. diversity of the electrophoretic types of Vibrio cholerae in the Western Hemi- 35. Weber, J. T., W. C. Levine, D. P. Hopkins, and R. V. Tauxe. 1994. Cholera sphere. J. Infect. Dis. 172:173–179. in the United States, 1965–1991. Risks at home and abroad. Arch. Intern. 12. Finch, M. J., J. L. Valdespino, J. G. Wells, G. Perez-Perez, F. Arjona, A. Med. 154:551–556.