Shigellosis, gastroenteritis caused by Shigella spp., is a major public health problem in both developing and developed countries [1, 2]. Approximately 164.7 million cases of shigellosis occur annually worldwide, resulting in 1.1 million deaths, primarily among children aged < 5 years [3]. Shigellosis occurs mainly in developing countries due to poor hygiene and limited access to clean drinking water; in industrialized countries the disease mainly results from travel to developing countries and exposure to contaminated foods and/or food handlers [1]. In China, Shigella spp. is one of the most frequently isolated gastrointestinal pathogens [4], accounting for up to 1.7 million episodes of shigellosis annually, with up to 200,000 patients admitted to hospitals [5, 6].

Guizhou Province, with nearly 50 million people, is an under-developed province in the southwest of China. Shigellosis has been one of the primary bacterial diseases in Guizhou in past decades, and from 2007 to 2010, 48,222 cases of shigellosis were reported [7]. Four species of Shigella can cause shigellosis but Shigella flexneri is the predominant species in China. Although there has been an observed shift in prevalence from S. flexneri to Shigella sonnei in recent years, S. flexneri is still one of the major etiologic cause of shigellosis in Guizhou Province [7].

Although Shigella has been a major source of disease over the past decades, information on the genetic relationships of circulating S. flexneri isolates from Guizhou Province is lacking. In this study, the molecular techniques multi-locus sequence typing (MLST), pulsed field gel electrophoresis (PFGE) and multi-locus variable-nucleotide tandem-repeat analysis (MLVA) were used to analyze the relationships between S. flexneri isolates recovered from Guizhou during the periods 1972 to 1982 and 2008 to 2010.

In the present study, the genetic characteristics of 60 isolates of S. flexneri recovered from Guizhou Province between 1972 to 1982 and 2008 to 2010 were systematically studied. Thirty of the S. flexneri isolates were from shigellosis cases from 1972 to 1982 and the remaining isolates were recovered from patients with shigellosis from 2008 to 2010. The serotypes of the isolates used in this study included 1a, 2a, 3a 1b, 2b, X, Y, 4av and Yv; isolates were also from seven of the nine prefectures in Guizhou Province. Serotyping results indicated that S. flexneri 1a and 2a were the predominant serotypes from 1972 to 1982 (94%) and 2008 to 2010 (72%), respectively, while 3a isolates were almost equally recovered in both periods. The predominant serotype (2a) in Guizhou Province recovered from 2008 to 2010 are consistent with isolates from Suzhou of Jiangsu, Henan and Shanxi Province [16–18], but are different from isolates of other provinces such as Beijing (4a and 4b) and Jiading of Shanghai Province [19, 20]. It is noteworthy that isolates of serotype 4av and Yv were recovered as early as 1973 and 1978, respectively, indicating the early emergence of these serotypes in China.

Recently a number of genotyping methods with higher discriminatory power than serotyping or biochemical testing such as MLST [21, 22], PFGE [5, 23] and MLVA [10, 11, 24] were introduced to characterize Shigella isolates. These methods are based primarily based on changes in isolate genotype, permitting analysis of phylogenetic relationships. Analysis of the isolates can be helpful for clinical diagnosis, treatment, prevention and control of shigellosis. Choi et al. [22] showed that S. flexneri serotypes 1–5, X and Y clustered together in a group containing many allelic variants while serotype 6 formed a distinct group, as previously established [25, 26]. Wang et al.[10] reported that phylogenetic groupings of 242 S. flexneri isolates recovered from shigellosis cases in Taiwan between 1995 to 2008, based on PFGE and MLVA profiles, correlated with serotype and isolate origin. Two distinct clusters for isolates of serotype 3 were shown but only one distinct cluster for each of the serotype groups 1a/1b/NT, 2a/2b/X/NT, 4a/Y, and 6 were revealed. Serologically different isolates including serotype Y and subserotype 4a; serotype X and subserotype 2b; subserotypes 1a and 1b, and subserotypes 3a and 3b, were genetically more closely related than indicated by serotyping alone.

Ye et al. [9] previously analyzed 37 serotype X and 69 serotype 1a, 2a, 2b, 3a, 4a, 5b, and Y isolates from China; all belonged to ST91 (later renamed ST 100), and concluded that S. flexneri epidemics in China have been caused by a single epidemic clone, ST 100. In this study, 60 isolates of S. flexneri from Guizhou Province separated into 20 STs based on a 15 loci MLST scheme; 18 of the STs were novel. The most common STs from 1972 to 1982 were ST 18, ST 126 and ST 129, however, ST 100 and ST 126 appeared between 2008 to 2010. Our results suggested that the predominant ST was ST 129 from 1972 to 1982, while ST 100 was the the predominant ST during 2008 to 2010, and the predominant ST is consistent with ST of isolates from other provinces of China [9]. Isolates belonging to the same serotype clustered in accordance with the year of isolation using all three genotyping approaches. MST indicated that the 20 STs were divided into 2 CCs, CC 18 and CC 126, and 4 singletons. CC 18 contained isolates expressing serotypes 1a, 2a, 2b, 4b, X and Y, while all the isolates in CC 126 belonged to serotype 3a. In addition, both the cluster tree and the MST, based on MLST data, showed that the isolates of serotype 4av (1973GZ03) was distant from the isolates belonging to other serotypes.

PFGE is a broadly applicable typing method with a high degree of intra- and inter laboratory reproducibility when standardized protocols are followed [23]. It has been shown to be a powerful tool in the laboratory for discriminating Shigella isolates during an outbreak [27]. In this study, PFGE discriminated the 60 isolates of S. flexneri into 39 unique PFGE patterns. Isolates belonging to the same serotypes mainly clustered together; the most closely related isolates were temporally associated with one another was well, suggesting that some drift was associated within each serotype over time. For instance, isolates of serotype 2a in cluster B2 isolated during the period of 1972 to 1982 and 2008 to 2010 were closely clustered, respectively, and similar clustering characteristics were observed for the isolates belonging to serotype 1a (cluster B1) and 3a (cluster A1), but some isolates, such as 2008GZ03, 1982GZ03 and 2009GZ23, were assigned irrespective their isolation time. Additionally, isolates with similar geographic origin were also often grouped by PFGE together as they tended to express the same serotypes.

MLVA is a prominent typing tool which has been used for characterizing S. flexneri; it has also been a useful tool for phylogenetic analysis [10]. In the present study, the S. flexneri isolates were discriminated into 44 different MTs and showed a low (approximately 20%) coefficient of similarity, indicating the high discriminatory power of the MLVA method. This finding is consistent with a previous study showing MLVA exhibited a discriminatory power greater than PFGE [10]. For most of isolates belonging to serotypes 1a, 2a and 3a, MLVA results correlated with serotyping. However, isolates of serotype X and Y were associated with serotype 1a and 2a isolates, respectively, and the serotype 4av isolate was closely related to serotype 3a isolates; this is similar to results observed previously [10]. Similar to PFGE, the majority of isolates belonging to the same serotype were temporally and geographically related.

In this study, isolates 1973GZ03 and 1978GZ01were serologically identified as serotype 4av and Yv respectively. These serotypes differ from 4a and Y because they react with monoclonal antibody MASF IV-1 [13–15]. In theory, serotype 4av and Yv should have originated from serotype 4a and Y, primarily differing only in the acquisition of a 6.8k plasmid carrying a phospheantransferase gene (opt) responding for the MASF IV-1 antigenic determinant in these isolates [13–15]. In this study, isolates of serotype of 4av (1973GZ03) and Yv (1978GZ01) were typed as ST 124 and ST 128 with MLST, respectively. MLST showed that 4av isolate were genetically distinct from isolates belonging to other serotypes, while Yv isolates were relatively close to isolates of serotype 2a. Molecular analysis indicates that isolate 1978GZ01 carries a dysfunctional gtrII gene within genome, and hence it is genetically similar to isolates from serotype 2a (unpublished data). This would explain the observed similarity between isolate 1978GZ01 and the majority of serotype 2a isolates.

In conclusion, phenotypic and molecular profiles of 60 S. flexneri isolates recovered in Guizhou between 1972 to 1982 and 2008 to 2010 were analysed. Nine serotypes (1a, 2a, 3a, 1b, 2b, X, Y, 4av and Yv) were identified, and the predominant serotype has changed from 1a to 2a in Guizhou Province. MLST differentiated the isolates into 20 sequence types (STs); 18 were novel. Four STs, ST 129, ST 100, ST 126 and ST 18, were most abundant, accounting for 65% of the isolates. The predominant ST was ST 129 in 1972 to 1982, while ST 100 was the predominant ST during 2008 to 2010. Thirty-nine NotI-PFGE (pulsotypes, PTs) were observed; eight PTs were represented by more than one isolate with six isolates sharing the PT 13 profile. MLVA analysis recognized 44 different types (MTs); seven MTs were represented by more than one isolate and MT 1 was most commonly encountered. Correlation between genetic relationships and serotypes was observed among the isolates studied; the majority of isolates belonging to the same serotype from different years clustered together based on the molecular data. These clustered isolates were also from similar geographical origins.