Global Distribution of Shigella sonnei Clones

To investigate global epidemiology of Shigella sonnei, we performed multilocus variable number tandem repeat analysis of 1,672 isolates obtained since 1943 from 50 countries on 5 continents and the Pacific region. Three major clonal groups were identified; 2 were globally spread. Type 18 and its derivatives have circulated worldwide in recent decades.

To investigate global epidemiology of Shigella sonnei, we performed multilocus variable number tandem repeat analysis of 1,672 isolates obtained since 1943 from 50 countries on 5 continents and the Pacifi c region. Three major clonal groups were identifi ed; 2 were globally spread. Type 18 and its derivatives have circulated worldwide in recent decades.
S higella sonnei is the most commonly isolated species among the 4 Shigella species in industrialized countries (1,2). Transmission of S. sonnei across geographic boundaries is frequently linked to international travel and cross-border food trade (3,4). S. sonnei is a monomorphic organism and therefore requires a highly discriminative sequence-based method for investigating its clonal structure and the geographic distribution of clones.
A total of 26 variable number tandem repeats (VNTRs) have been used to type S. sonnei isolates (5). Because VNTRs have a wide range of variability, they are useful markers for investigating clonal relationships among strains that have evolved over different times (6). In this study, we analyzed 1,672 S. sonnei isolates obtained since 1943 from 50 countries on 5 continents (Africa, Asia, Europe, North America, and South America) and the Pacifi c region by multilocus VNTR analysis (MLVA) to investigate the global epidemiology of S. sonnei.

The Study
Isolates were obtained from 50 countries on 5 continents and the Pacifi c region (online Appendix Table, www.cdc.gov/EID/content/17/10/10-1486-appT.htm). Of these isolates, 31 were obtained during 1943-1983 (from Cameroon, Denmark, France, Senegal, and Sweden) and 1,641 were obtained during 1994-2008 from 48 countries. Isolates were lyophilized, kept in stab culture medium, or stored in 15%-20% glycerol at -75°C for long-term storage. The isolates were not repeatedly subcultured before this study. MLVA26, an MLVA assay based on analysis of 26 VNTRs, classifi ed the 1,672 isolates into 620 MLVA26 types. With only 2 exceptions, no common MLVA26 type was shared among isolates from different countries. One isolate from Malaysia (1999) and 2 isolates from Vietnam (2006) shared a common MLVA26 type, and an isolate from Chad and an isolate from France (both isolated in 2007) shared a common MVLA26 type.
The high resolving power of MLVA is primarily caused by highly diverse VNTRs (6). Clustering analysis of the MLVA26 types using a minimum spanning tree (MST) algorithm grouped the 1,672 isolates into 3 large clusters (A, B, and C), 1 small cluster (D), and 1 singleton (E). Each cluster was defi ned to include MLVA26 types differing at <7 loci among the 26 loci. The 3 large clusters displayed distinct allelic diversity features. Eight loci (SS1, SS3, SS6, SS9, SS10, SS11, SS12, and SS23) had Simpson diversity values >0.5 for >1 of the 3 large clusters (Table). Differences in diversity values >0.3 among the 3 clusters were observed for 9 of the 26 VNTRs. The largest difference was in 2 hypervariable VNTRs (SS1 and SS6). These 2 VNTRs displayed a high degree of allelic diversity in cluster A, but were invariant in cluster B. SS1 was invariant but SS6 displayed a high degree of diversity in cluster C.
Of 1,672 isolates, 66% (1,100) were obtained from patients who acquired infections in Taiwan. Most isolates belonged to an insertion element IS1 interspacer 1 clone (6,7) that had slightly lower diversity values for some VNTRs than diversity values for total isolates obtained from a panel of 620 isolates representing the 620 MLVA26 types. However, a large number of clonal isolates from Taiwan did not affect relative magnitudes of diversity of the 26 VNTRs.
Although highly variable VNTRs are useful markers in distinguishing closely related strains, they are less useful for investigating clonal relationships among strains that have evolved over time (6). MLVA18 profi les, which excluded the data of 8 highly variable VNTRs (SS1, SS3, SS6, SS9, SS10, SS11, SS12, and SS23) from the 26-locus panel, were used to investigate the clonal structure of the isolates. On the basis of 18-locus profi les, 105 MLVA18 types were identifi ed.
A simplifi ed MST was created by analysis of a subset of 200 isolates selected by obtaining 1 MLVA18 type among those identifi ed in each of 50 countries. As shown in the MST (Figure), cluster A was further divided into subclusters A1 and A2 and singleton A3, and cluster C was divided into subclusters C1 and C2.
Cluster A consisted of 46 MLVA18 types, which represented 1,382 isolates obtained since 1943 in 40 countries on 5 continents and the Pacifi c region (online Appendix Table). Several MLVA18 types within cluster A were widespread. SS18.2 was detected in 23 countries on 5 continents. SS18.2 had 14 single-locus variants (SLVs) at which genotypes differed only at 1 of the 18 loci; 3 (SS18.80, SS18.1, and SS18.60) of the SLVs were detected in 13, 10, and 7 countries, respectively. SS18.2 and its SLVs represented 1,290 isolates obtained in during 1995-2008 from 38 countries on 5 continents. Four MLVA18 types were detected in samples obtained in 1943-1983. These isolates shared identical MLVA18 profi les or differed at 1-2 loci from recently obtained isolates. Subcluster A2 consisted of 4 MLVA18 types found in isolates from Argentina only. Singleton A3, which was found in isolates obtained in Vietnam in 2008, was distantly separate from subclusters A1 and A2.
Cluster B consisted of 12 MLVA18 types representing 75 isolates, which were obtained in 8 countries in Africa, Asia, and Europe. SS18.6 had the highest number of SLVs in cluster B and was detected in 5 countries in Asia and Africa. Five types were detected in isolates obtained in 1943-1974; they shared identical MLVA18 profi les or differed at 1-2 loci from recently recovered isolates.
Cluster C was relatively diverse; it consisted of 44 MLVA18 types, which represented 212 isolates obtained in 21 countries on 5 continents and the Pacifi c region. SS18.8 had the highest number of SLVs and was found in isolates obtained during 1974-2007 in 8 countries on 5 continents. SS18.4, the largest SLV of SS18.8, was found in isolates from 5 countries on 5 continents. Subcluster C2 consisted of isolates from Argentina obtained in 2002 and Sweden and Denmark in 1943. These isolates emerged in 1943-1974 and were genetically similar to recently obtained isolates. Clusters (clonal groups) A and C were globally spread, and clonal group B was found in countries in Africa, Asia and Europe only.

Conclusions
Genetic analysis using MLVA presented a simple clonal structure for 1,672 S. sonnei isolates obtained since 1943 from 50 countries on 5 continents and the Pacifi c region. Three large clonal groups were identifi ed; they displayed distinct allelic diversity features, particularly for 2 hypervariable VNTRs (SS1 and SS6). Clonal groups A and C were globally spread. One MLVA18 type (SS18.2) and several of its SLVs were widely distributed over 5 continents in the past 10 years. Dr Filliol-Toutain is deputy director of the National Reference Center for Escherichia coli and Shigella at Institut Pasteur, Paris, France. Her research interests are surveillance of Shigella strains and Shiga toxin-producing E. coli strains circulating in France, detection of outbreaks, surveillance of the circulating serovars, and surveillance of antimicrobial drug resistance.