Identification and Characterization of a Novel Shigella flexneri Serotype Yv in China

Shigella flexneri is the major cause of bacterial shigellosis in developing countries. S. flexneri is divided into at least 19 serotypes, the majority of which are modifications of the same basic O-antigen by glucosylation and/or O-acetylation of its sugar residues by phage encoded serotype-converting genes. Recently, a plasmid encoded phosphoethanolamine (PEtN) modification of the O-antigen has been reported, which is responsible for the presence of the MASF IV-1 determinant and results in conversion of traditional serotypes X, 4a and Y to novel serotypes Xv, 4av and Yv, respectively. In this study, we characterized 19 serotype Yv strains isolated in China. A variant of the O-antigen phosphoethanolamine transferase gene opt (formerly called lpt-O) carried by a pSFxv_2-like plasmid was found in serotype Yv strains, which specifies the phosphorylation pattern on the O-antigen of this serotype. For the majority of the O-antigen units, the PEtN modification occurs on RhaIII, while for a minority, modifications occur on both RhaII and RhaIII. Serotype-specific gene detection and PFGE analysis suggested that these serotype Yv isolates were originated from serotypes Y, Xv and 2a by acquisition of an opt-carrying plasmid and/or inactivation of serotype-specific gene gtrII or gtrX. These data, combined with those of serotypes Xv and 4av reported earlier, demonstrate that the plasmid-encoded PEtN modification is an important serotype conversion mechanism in S. flexneri, in addition to glucosylation and O-acetylation.


Introduction
Shigellosis or bacillary dysentery remains an important public health challenge worldwide. It is estimated that there are 125 million people suffer from shigellosis annually, resulting in 14,000 deaths, the majority of which were children under five years old [1]. The causative agent of shigellosis is Shigella spp., nonmotile, nonspore-forming facultative anaerobic Gram-negative bacteria. Based on biochemical and serological properties, the genus Shigella is divided into four species or subgroups, although genetically all are clones of Escherichia coli: S. dysenteriae, S. flexneri, S. sonnei, and S. boydii. Among them, S. flexneri is the predominant species in developing countries [2].
More recently, a novel S. flexneri O-antigen modification, addition of phosphoethanolamine (PEtN) to Rha II was identified in the newly named serotype Xv [9]. This novel PEtN modification was also found in a serotype 4a variant, 4av, and a serotype Y variant, Yv, in which, a PEtN residues was added mainly to Rha III instead of Rha II [9,20,21]. In serotypes Xv, 4av and Yv, there were also some tetrasaccharide units with PEtN groups on both Rha II and Rha III [9,21]. In all the cases, the PEtN modification on the O-antigen confers the MASF IV-1 positive phenotype to the host [9,21]. A single gene, opt, encoding the Oantigen phosphoethanolamine transferase (formerly called lpt-O), which is carried on a 6.8 Kb plasmid pSFxv_2, mediates the PEtN modification in these serotypes [9].
Strains belonging to serotypes Xv and 4av have been isolated from patients in China, Bangladesh, Australia and Russia, and serotype Xv was the predominant serotype in China for several years [7,20,22,23]. Nineteen serotype Yv isolates have also been identified in China [9]. In this study, the phenotypic and genetic characteristics of these 19 serotype Yv isolates were analyzed. Serotype-specific gene detection and PFGE analysis suggest that the serotype Yv isolates were derived from serotypes Y, Xv and 2a on at least three independent occasions through acquisition of a pSFxv_2-like plasmid and/or inactivation of other serotypespecific genes.

Ethics Statement
This study was reviewed and approved by the ethics committee of National Institute for Communicable Disease Control and Prevention, the Chinese CDC. S. flexneri strains were acquired with the written informed consent of the diarrhea patients and with the approval of the ethics committee of National Institute for Communicable Disease Control and Prevention, according to the medical research regulations of Ministry of Health (permit number 2007-17-3).

Strains and Culture Conditions
The details of 19 serotype Yv isolates, including strain code, isolate year and locality, serotype-specific genes are listed in Figure 1. The strains were isolated from diarrheal patients in a surveillance program performed by China CDC during 2000-2011. Serotype Xv strain 2002017 was used as control for the opt gene analysis and strain HN006 was used for plasmid pSFyv_2 extraction and sequencing. S. flexneri strains were routinely grown in a 37uC incubator or orbital shaker in Luria-Bertani broth (LB) supplemented with ampicillin (100 mg ml 21 ) when appropriate.

Serotyping and Biochemical Characterization
The serological features of S. flexneri strains used in this study were determined by slide agglutination using the commercially available monovalent antisera kit (Denka Seiken, Japan) and monoclonal antibody reagents (Reagensia AB, Sweden) specific for all S. flexneri type-and group-factor antigens. The biochemical characteristics of the 19 serotype Yv isolates were identified using API-20E test kits according to the instructions of the manufacturer (Bio-Merieux Vitek Inc., Hazelwood, Mo.).

Serotype-specific Gene Detection
To detect the known serotype-specific genes gtrI, gtrII, gtr1C, gtrX, gtrIV, gtrV, and oac, multiplex-PCR was performed as described previously [25]. Primer pair lpt-O-2 [9], which were designed based on the opt gene (pSFxv_5135) of strain 2002017, was used for the opt gene detection. Primer pair opt-3U: CAACATGGGGTAAAACCGCC and opt-3L: ATCTATT-TATTCAACAACGCCCCCC was used to amplify the whole opt gene. PCR amplification was performed using the standard protocol on a SensoQuest LabCycler (Germany).

Plasmid Profiling and Southern Hybridization
Plasmid profiling and Southern hybridization analysis were performed as described previously [9]. Plasmids isolated from strain 2002017 were used as control. Lambda DNA cleaved with HindIII (TaKaRa, Japan) was used as electrophoresis markers. The DNA product amplified from strain 2002017 using primer pair lpt-O-2 [9] was prepared for a biotin-labeling DNA probe.

Pulsed-field Gel Electrophoresis (PFGE) and Multilocus Sequence Typing (MLST)
PFGE analysis was performed using the method described by Ye et al. [7]. PFGE images were analyzed using the fingerprint analysis software BioNumerics version 4.5 (Applied Maths; Kortrijk, Belgium). A PFGE type was defined as a PFGE pattern with one or more DNA bands different from the others. The dendrogram constructed using the PFGE patterns was generated by the UPGMA algorithm with the Dice-predicted similarity value of two patterns at 1.0% pattern optimization and 0.8% band position tolerance. The virulence plasmid band was removed from comparisons as there is no NotI cut site, and in some cases the band is barely visible. MLST analysis of 15 housekeeping genes was performed as described on the EcMLST website (http://www. shigatox.net/ecmlst). PCR products were sequenced in both directions.

Sequencing of the opt-carrying Plasmid
The opt gene carrying pSFyv_2, a pSFxv_2-like plasmid isolated from strain HN006 was digested with EcoR I and labeled with the ampicillin-resistant gene amplified from plasmid pMD20-T (unpublished data). After replication in E. coli JM109, the labeled plasmid was sequenced directly using primers complementary to the opt sequence. The plasmid sequence obtained was rearranged with the same start position as the linear plasmid sequence of pSFxv_2 (CP001385). Open reading frames (ORFs) were determined using the ORF Finder program (http://www.ncbi.nlm.nih. gov/gorf/gorf.html) and by comparison with pSFxv_2. Sequences obtained in this study have been deposited in GenBank (KC020049).

Characteristics of the 19 S. flexneri Serotype Yv Isolates
Nineteen serotype Yv strains were isolated from Anhui and Henan provinces (Fig. 1), and were identified among 1650 S. flexneri strains collected from 10 provinces of China in a surveillance program performed by China CDC during 2000-2011. All 19 Yv isolates displayed the same biochemical features of Shigella species, with ability to ferment glucose, mannitol, melibiose and arabinose, production of indole, and non-utilization of sorbitol and rhamnose, except for one strain, AH029, which could not use melibiose and arabinose.
All 19 isolates reacted with S. flexneri specific polyvalent antisera B of Denka Seiken, or MASF B monoclonal antibody of Reagensia AB and thus were serologically identified as S. flexneri. However, when sub-typing using typing and grouping antisera of the Denka Seiken and MASF schemes, all 19 isolates showed a common serological feature: agglutinating with both the grouping monoclonal antibody MASF IV-1 and the monovalent antiserum IV (Table 1), and thus were initially typed as serotype 4a or 4. However, these strains did not react with the serotype 4-specific monoclonal antibody MASF IV-2; and the serotype 4-specific gtrIV Figure 1. PFGE profiles of the 19 serotype Yv isolates studied and comparison with closest non-Yv strains. The dendrogram constructed using the NotI-digested PFGE patterns was generated by the UPGMA algorithm with the Dice similarity value of two patterns at 1.0% pattern optimization and 0.8% band position tolerance. Serotype-specific genes were detected using multiplex-PCR as described previously [25]. doi:10.1371/journal.pone.0070238.g001 gene could not be amplified from them by PCR as described in more detail below. Therefore, they did not belong to serotype 4 (4a, 4b or 4av) ( Table 1). In comparison to serotype Xv, the 19 strains are devoid of the 7;8 determinant (Table 1). These 19 strains are also different from classical serotype Y strains by showing the MASF IV-1-positive phenotype (Table 1). Following the naming used for serotypes Xv and 4av of which the v (variant) refers to the MASF IV-1 positivity, we named the 19 strains as serotype Yv, a variant of the typical serotype Y.
The agglutination pattern of serotype Yv was similar to those of several atypical serotypes reported previously. One such strain, which was provisionally named as 4s was reported recently in China [26]. Thirty nine non-typable S. flexneri isolates collected in Bangladesh during 1997-2000 reacted only with the monoclonal antibody MASF IV-1 and were named serotype 4X [22]. Much earlier, 35 isolates with similar serotype features were identified in Bangladesh during 1985-1987 and were named as Y [23]. These data indicate that serotype Yv may have been existed in nature for many years rather than emerged recently.
All but the HN182 isolate was sensitive to both piperacillin (PRL) and Cefepime (FEP). All except 3 isolates (HN053, HN054 and HN055) were sensitive to ciprofloxacin (CIP) and levofloxacin (LEV) ( Table 2). Overall, the antimicrobial patterns of these serotype Yv isolates were similar to those of Chinese S. flexneri isolates reported previously [3,7].

Sequence Variation of the opt Gene Among the Serotype Yv Strains
The plasmid-borne opt gene has been identified to be responsible for the PEtN modification in the MASF IV-1 positive serotypes Xv, 4av and Yv strains [9,21]. However, a significant opt gene variation has been identified between serotype 4av and Xv [9]. To determine the opt gene sequence in serotype Yv isolates, we PCR amplified and sequenced the opt gene from the 19 Yv isolates using primer pair opt-3. The opt gene from six isolates (HN002, HN303, HN343, HN053, HN054 and HN055) was found to be identical to that of 2002017 (serotype Xv), while the opt gene from the remaining 13 Yv isolates was identical to that of serotype 4av strains reported earlier [9], with 11 base changes (243 A-G,  310 G-A, 379 G-A, 687 T-G, 691 T-C, 728 C-T, 772 C-T,  836 C-A, [21]. However, a clear modification difference is observed between them: opt from serotype Xv is better tuned for phosphorylation of Rha II and its variant is better tuned for phosphorylation of Rha III , and hence we named them optII and optIII , respectively [21]. It is intriguing that the opt gene has evolved with a positional preference for PEtN addition. A possible explanation for this evolution may be selection pressure. There is no modification on Rha III in serotypes Y and 4a (and also serotypes 1a, 1b, 1c, 2a, 3b, 4b, 5a, 7b), and the optIII can easily mediate the addition of PEtN onto it. In contrast, in serotype X (and other serotypes 1d, 2b, 3a, 5b), whose Rha III is occupied by a glucosyl residue, the optIII may not effectively compete for the residue with other modifications. Its variant optII can easily modify Rha II to confer the MASF IV-1 antigen to the host. We sequenced the opt genes in another 21 serotype Xv strains, and found that all are the optII (unpublished data).
The plasmid profiles of the 19 serotype Yv isolates were analyzed by PFGE. Similar to that found in serotype Xv strains, all serotype Yv isolates carried a plasmid of about 6.8 kb, the predicted length of pSFxv_2 ( Fig. 2A). Hybridization results suggested that the opt gene was always carried by the 6.8 kb plasmids ( Fig. 2A).

Identification of Defective gtrII and gtrX Genes in Serotype Yv Isolates
The O-antigen modification of S. flexneri (serotype 1-5,7 and X) is generally associated with serotype-converting phages integrated into the host genome, which encode serotype conversion gene modules (gtrABC or oac) and mediate the glucosylation and/Oacetylation of the O-antigen [11,13,14,15,16,17,18,19]. The serotypes of S. flexneri isolates can be identified by PCR detection    of serotype-specific genes (gtrI, gtrII, gtrX, gtrIV, gtrV, oac and gtrIC) [25].Theoretically, strains of serotype Yv should not carry any of the serotype-specific genes mentioned above. We performed multiplex-PCR [25] on the 19 isolates to determine the presence of these genes. Surprisingly, except for 3 isolates (HN002, HN303 and HN343), which had none of the known serotype-specific genes, 13 isolates (AH012, HN068, HN069, AH028, AH029, HN006, HN171, HN182, HN033, HN011, HN106, HN116 and HN049) and 3 isolates (HN053, HN054 and HN055) were found to be positive for gtrII and gtrX genes, respectively (Fig. 3). Genes gtrII and gtrX are encoded in the prophage genomes of SfII and SfX respectively, and responsible for the presence of type II antigen in serotype 2 strains and the group 7;8 determinant in serotypes X, Xv, 1d, 2b, 3a and 5b strains [4,13], respectively. The presence of gtrII and gtrX in these strains indicates that they may have been derived from serotype 2a and serotype X strains, respectively. However, the absence of any of these serotype characteristics in the 16 Yv isolates suggests that the gtrII or gtrX genes are defective. The gtrII genes from the 13 gtrII-carrying Yv isolates were amplified by PCR and sequenced. Except for strain HN006, the gtrII gene was found to be identical in the other 12 Yv isolates, but different from the gtrII gene of serotype 2a strain 2457T (accession no NC_004741) by a single base at position 1222 (GRT), which resulted in an amino acid change at the position 408 (D-Y). The gtrII gene of HN006 carried an additional base change (position 34, CRA), resulting in another amino acid substitution at position 12 (LRI). We further sequenced the other 2 genes of the gtr locus (gtrA and gtrB), as well as the upstream regular elements from the 13 isolates. A nonsynonymous mutation was found in gtrB at position 560 (ARC), which resulted in an amino acid change at the position 187 (QRP). Therefore, either one or both of the mutations in the gtrII (position 1222) and gtrB (position 560) genes are responsible for the defective GtrII function in these serotype Yv strains.
Similarly, sequence analysis of the gtr locus (gtrA, gtrB and gtrX) of the three gtrX-carrying Yv isolates HN053, HN054 and HN055 revealed a single base deletion at position 1125 (T) of the gtrX gene, which resulted in a stop codon at base position 1126-1128 leading to a 10% truncation of the peptide. This mutation is likely to be responsible for the defective GtrX function in these 3 strains.
Such serotype-specific gene inactivating mutations have been found in several serotypes of S. flexneri earlier. Roborts et al. [27] found that the serotype Y vaccine candidate SFL124 contained an insertion sequence (IS1) in the gtrII gene, resulting in a nonfunctional GtrII. Chen et al. [28] also observed the conversion of serotype 2a to serotype Y due to a single base mutation in gtrII which led to a Cys (437) to Tyr substitution. In our previous study [25], we have identified four serotype Y and two serotype X isolates carrying a defective gtrII gene, with one having a four-base deletion (bases 1197 to 1282) and five having a single base deletion (base 1031 or 1024). Defective oac genes with two-base deletion (bases 345 and 346) have been found in S. flexneri serotype 5a strain NCTC 8523 [25]. In light of the various serotype-specific defective genes, we performed a PCR screening of the 1650 S. flexneri isolates in our collection and found only 46 strains carrying defective gtr genes (2.8%) (data not shown), including 16 Yv isolates used in this study. The frequency of serotype-specific gene inactivating mutations seems to be low in the S. flexneri population in China.
However, there seem to be disproportionally more defective gtr genes in serotype Yv isolates, with 16 of the 19 serotype Yv isolates in this category. A possible explanation for this phenomenon is that most serotype Y strains in nature are derived from other serotypes by serotype-specific gene mutation, and the derived serotype Y strains can be easily transformed into serotype Yv by an opt-carrying plasmid. We screened all 35 serotype Y isolates in our collection for defective gtr and oac genes and found 19 isolates with defective gtrII or gtrI (13 gtrII and 6 gtrI). Therefore, 35 (16+19) of the 54 (35+19) serotype Y or Yv isolates evolved actually by back conversion from a modified serotype to the basic O-antigen type. Additionally, the higher number of gtrII mutations may be due to the higher proportion of serotype 2a isolates in nature.
The PFGE patterns of these 19 Yv isolates were further compared with those of over 1600 S. flexneri strains collected from China, and Figure 1 shows the closest strains of other serotypes. It can be seen that the gtrII-carrying Yv isolates share an identical pattern with serotype 2a isolates: HN006 with HN210 and HN018, AH012 with HN090 and HN019, PFGE type CN03 strains with HN322, and PFGE type CN04 strains with GS028, HN049-1 and HN069, respectively (Fig. 1). It should be noted that two serotype Y isolates HN064 and HN126 which showed the same PFGE pattern as those of strains HN006 and AH012, respectively, were also found to carry a defective gtrII gene (Fig. 1); therefore, they were also derived from serotype 2a. Another gtrIIcarrying isolate HN033 also showed the closest similarity to serotype 2a strains HN115 and HN085, with only one band different (Fig. 1).
Similarly, the three gtrX-carrying serotype Yv isolates (HN053, HN054 and HN055) have the same PFGE pattern as serotype Xv strains HN022 and HN056 (Fig. 1). We have also confirmed the 2 serotype Xv strains carry an optII gene. On the other hand, three serotype Yv isolates (HN002, HN343 and HN303), which have none of the other serotype-specific genes, showed the closest similarity to serotype Y strains HN043 and HN064-1, with the same PFGE pattern or only one or two bands different (Fig. 1).
Taking into account the data on the defective gtr genes and the opt gene sequence revealed above, we can conclude that the Yv strains in China have at least three origins (Fig. 4): 1) 3 Yv strains were derived from classical serotype Y strains by gaining an optIIcarrying plasmid; 2) 13 Yv strains were derived from serotype 2a strains by gtrII or gtrB gene inactivation and gaining an optIIIcarrying plasmid; and 3) 3 Yv strains were derived from serotype Xv strains by gtrX gene inactivation. The last 3 strains carry an optII plasmid that also supports their serotype Xv origin.
For the 13 serotype Yv strains converted from serotype 2a, it seems clear that there were at least 5 independent conversions since each PFGE type (CN1 to CN5) contained both serotype 2a and Yv strains (Fig. 1). However the order of the occurrence of the two events (loss of gtrII or gtrB and gain of optIII) is unclear and thus alternatives were presented in Figure 4. It is possible that a serotype 2a strain carrying pSFyv_2 (which would be serotype 2av) lost the GtrII function to become serotype Yv. However, considering that no serotype 2av strains have been detected in nature, it is likely that the inactivation of gtrII or gtrB occurred first and pSFyv_2 was gained subsequently.
This study also further demonstrated the preferential association of pSFxv_2 (optII) with serotype Xv and pSFyv_2 (optIII) with serotype Y or serotype 4a. This preference of optII in Xv is likely to be due to either the avoidance of the competition of modifying the same rhamnose residue [21] since its Rha III is modified by a glucosyl residue and Rha II is unmodified or an advantage to modify the unoccupied Rha II by adding an additional antigenic variation. However, for serotype 4a and Y, both Rha II and Rha III are unoccupied. Therefore both optII and optIII can modify the rhamnose residues. Indeed, we see both types of Yv although there were far more optIII-carrying Yv (13 isolates) than optII carrying isolates (3 isolates). This difference in frequency, however, may well be due to sampling.
Up to now, O-antigen PEtN modification mediated by the opt gene has only been reported in serotypes Xv, 4av and Yv of S. flexneri. As an easily transferable factor, the opt-carrying plasmid theoretically can spread among different strains, and should also been found in other serotypes of S. flexneri. We screened more than 1600 S. flexneri strains collected in our laboratory by PCR amplification of the opt gene. Apart from serotypes Xv, 4av and Yv, opt gene was not detected in other serotypes tested [9]  (unpublished data). However, by transformation under laboratory conditions, we found that all serotypes (serotype 1-5, X, Y) of S. flexneri can be transformed by both plasmids pSFxv_2 and pSFyv_2, and converted into the corresponding MASF IV-1positive variants either retaining or losing the initial serotype characteristics (unpublished data).

Conclusion
In this study, we characterized 19 S. flexneri serotype Yv isolates from China. This new serotype differed from the classical serotype Y by the MASF IV-1 reaction. The optIII gene, which was carried by pSFyv_2, a pSFxv_2-like plasmid, mediates the PEtN modification in serotype Yv strains. Serotype-specific gene and PFGE analyses indicate that these serotype Yv strains originated from serotypes Y, Xv and 2a strains independently by acquisition of the opt-carrying plasmid and/or inactivation of serotype-specific genes. Although the isolation frequency of serotype Yv is low in nature, it may have an advantage over serotypes Xv and 2a to infect humans who had prior exposure to the latter two serotypes; hence, more attention should be paid to the new serotype Yv.