Sequence Analysis of the rfb Loci, Encoding Proteins Involved in the Biosynthesis of the Salmonella enterica O17 and O18 Antigens: Serogroup-Specific Identification by PCR

We report sequencing of the O antigen encoded by the rfb genecluster of Salmonella enterica serotype Jangwani (O17) and Salmonellaserotype Cerro (O18). We developed serogroup O17- and O18-specificPCR assays based on rfb gene targets and found them to be sensitiveand specific for rapid identification of Salmonella serogroupsO17 and O18.

INTRODUCTION

Top

Introduction

Serotyping of Salmonella enterica based on O and H surface antigenshas proven to be extremely useful for understanding the hostrange and disease spectrum of this pathogen (20) and has beenan invaluable typing method for epidemiological investigations(2, 3, 19, 20). It does, however, have limitations; productionand quality control of the hundreds of antisera required forserotyping is difficult and time consuming, and not all Salmonellastrains are typeable. Many of the enzymes involved in O-antigenbiosynthesis are encoded by an rfb gene cluster which variesin size for each serogroup depending on the sugar compositionand complexity of the O antigen (24). rfb gene clusters havebeen characterized from a growing number of gram-negative bacteria(23), including 13 Salmonella serogroups (9, 22, 26, 28). Theuse of PCR targeting regions within rfb gene clusters for O-antigenidentification of various organisms (6, 7, 11, 27), includingSalmonella (9, 10, 16), has been reported as an alternativeto O-group detection by traditional serotyping. The rfb geneclusters of the more common Salmonella serogroups have beenthe most extensively studied at the molecular level to date.Little is known about the rfb gene clusters of other SalmonellaO serogroups. As part of a larger project to develop a DNA-basedapproach for serotyping Salmonella, we report here the sequenceanalysis of the rfb gene cluster of two serogroups, O17 fromSalmonella enterica serotype Jangwani and O18 from Salmonellaserotype Cerro, and the development of serogroup-specific PCRdetection assays based on the rfb gene targets wzy and wzx,respectively.

Salmonella isolates 00-0073 (Salmonella serotype Jangwani, 17:a:1,5)and 99-0087 (Salmonella serotype Cerro, 18:z₄,z₂₃:[1,5]) wereselected for rfb sequence analysis. Serotyping was performedaccording to the Kaufmann-White scheme (2, 3). For evaluationof the serogroup O17- and O18-specific PCR assays, 443 strainsrepresenting all of the subspecies/serogroup combinations recognizedin the Kaufmann-White scheme were evaluated. The rfb regionsof O17 and O18 were amplified by PCR using previously describedprimers (14). A combination of subcloning and primer walkingwas used to sequence the rfb regions in both directions on anautomated DNA sequencer (PE Applied Biosystems, California).DNA sequence data were assembled and analyzed using Lasergene99 (DNAStar) software.

O17 rfb gene cluster.

Top

O17 rfb gene cluster.

The O17-antigen structure is unknown. Chemical analyses suggestthat the O17 antigen contains a polysaccharide core comprisedof galactose, glucose, glucosamine, heptose, and 3-deoxy-oct-2-ulosonicacid but has no reported additional sugars in its O17-antigen-specificside chain (15). Analysis of the DNA sequence of the rfb genecluster from Salmonella serotype Jangwani revealed 9.130 kbpflanked by JUMPstart and gnd and eight open reading frames (ORFs)(Fig. 1A). All ORFs were analyzed using BLAST, and the geneswere named in accordance with Bacterial Polysaccharide GeneDatabase (BPGD) nomenclature (21). All genes had the same transcriptionaldirection and have low G+C content (Table 1).

View larger version (14K):
[in this window]
[in a new window]

FIG. 1. Organization and G+C content of (A) O17 and (B) O18, O6,14, and C₁ rfb gene clusters. The percent G+C ratios were calculated and plotted for every 100 bases. Putative ORFs are represented by arrows, with the corresponding assignment of the gene name.

View this table:
[in this window]
[in a new window]

TABLE 1. S. enterica serogroup O17 antigen biosynthetic proteins

We identified the gene mnaA, encoding a UDP-N-acetylglucosamine(GlcNAc)-2-epimerase. When analyzed using Pfam protein analysis,the predicted protein exhibited significant identity to theUDP-N-acetylglucosamine-2-epimerase (PN 00235) protein family.This enzyme catalyzes the reversible epimerization of UDP-N-acetylglucosamine(UDP-GlcNAc) and thereby provides the cell with UDP-N-acetylmannosamine(UDP-ManNAc), the activated donor of ManNAc residues. The presenceof this gene in the O17 rfb gene cluster suggests that ManNAcmay be one of the sugars in the Salmonella O17 antigen. orf7showed the highest identity (65%) at the amino acid level withGlf, a UDP-galactopyranose mutase from Streptococcus suis. Salmonellaserogroup B and D isolates have also been shown to have geneswith identity to glf (Table 1), but they are not located inthe rfb gene cluster. These enzymes catalyze the interconversionof the pyranose (UDP-Galp) and furanose (UDP-Galf) forms ofgalactose (18). The presence of this gene in the O17 rfb genecluster suggests that galactofuranose is one of the sugars inthe O17 antigen. The O17 rfb gene cluster contains four ORFspredicted to encode putative transferases. orf3, -5, -6, and-8 shared various degrees of similarity with glycosyltransferasesfrom a range of other organisms (Table 1) and have been namedwfbQ, wfbR, wfbS, and wfbU. Three of four putative O17 transferases(wfbQ, wfbR, and wfbS) are part of the GT4 family of retainingglycosyltransferases, and wfbU is a putative inverting GT2 familyglycosyltransferase based on the carbohydrate-active enzymedatabase (CAZy) (http://www.cazy.org) classification (4, 5).We identified two ORFs encoding hydrophobic proteins with 10and 11 predicted transmembrane domains. These are characteristicfeatures for the O-antigen polymerase and transporter proteins,and we named these genes wzy and wzx. The putative O17 Wzx proteinhad similarities to several putative polysaccharide repeat unittransporters (Table 1).

O18 rfb gene cluster.

Top

O18 rfb gene cluster.

The O18-antigen-specific O-subunit structure has been determinedusing nuclear magnetic resonance and chemicals methods (25).Analysis of the DNA sequence from the Salmonella enterica O18rfb gene cluster revealed 9.018 kbp flanked by JUMPstart andthe gnd gene and contained seven open reading frames (Table2). Five of the seven genes had a low G+C content ranging from28.27% to 31.21% (Fig. 1B), a feature characteristic of otherSalmonella rfb gene clusters. The two genes, manC and manB,had G+C contents of 47.46% and 55.29%, respectively. All ofthe genes have the same transcriptional direction. Since theO18 antigen contains mannose, it was anticipated that the rfbcluster would contain manB, encoding phosphomannomutase, andmanC, encoding mannose-1-phosphate guanylytransferase, the twogenes responsible for the biosynthesis of GDP-mannose from mannose-6-phosphate(12). orf3 and orf4 were identified as manB and manC, respectively,based on their high levels of identity to other GDP-mannosebiosynthesis genes (12). The deduced gene products of orf1,orf2, and orf7 shared identity to glycosyltransferases, andwe have named these genes wfbV, wfbW, and wfbX. wfbV and wfbWhave specific homology to mannosyltransferases and retain glycosyltransferasesfrom family GT4 (5). The remaining O:18 transferase, encodedby wfbV, could not be assigned to 1 of the 65 different glycosyltransferasefamilies previously described and was left unclassified (5).A Pfam search (1) suggests it to be related to an acyltransferasefamily (PF01757). orf5 and orf6 are predicted to encode hydrophobicintegral membrane proteins with 12 and 11 transmembrane domains,respectively, and have a 45% and 48% amino acid identity tothe Wzy and Wzx proteins of Salmonella enterica O6,14 (9). Wehave named these genes wzy and wzx, respectively.

View this table:
[in this window]
[in a new window]

TABLE 2. S. enterica serogroup O18 antigen biosynthetic proteins

O17- and O18-specific PCR.

Top

O17- and o18-specific pcr.

A total of 46 pools of genomic DNA (50 ng/µl per isolate)were made, corresponding to the 46 O serogroups; each pool containedDNA from 1 to 12 isolates. The O17- and O18-specific primers(O17, 5'-GGC TGG GTT GTG GCT TTT T-3' and 5'-CTT CCG AAA TCAATA GAA AAA TCA A-3'; and O18, 5'-CTC TAG GAT CAA CTG AAG GTGGTC-3' and 5'-CAA CCC AGC AAT AAA GCA GAA-3') were derived fromsequences within wzy and wzx to generate a 565-bp and a 370-bpPCR amplicon. PCR amplification was carried out with the followingparameters: 96°C for 2 min, 25 cycles of 94°C for 30s, 58°C for 30 s, and 72°C for 45 s, and 72°C for10 min. When a positive PCR result was seen for a particulargenomic DNA pool, individual DNA samples from isolates representedin that pool were individually retested by PCR. A 565-bp ampliconwas generated only with the pool containing DNA from isolatesfrom different serotypes belonging to serogroup O17, and a 370-bpamplicon was generated only with the pool containing DNA fromisolates from different serotypes belonging to serogroup O18.All other pools, containing DNA from 398 strains representingthe other 44 serogroups, were negative in the PCR assays. DNAsamples from the 12 strains in each of the O17 and O18 poolswere individually tested in the PCR, and all generated a PCRproduct of the expected size (Fig. 2). An additional 19 O17strains and 20 O18 strains were tested, and all generated theexpected 565- and 370-bp PCR products, respectively.

View larger version (38K):
[in this window]
[in a new window]

FIG. 2. (A) PCR amplification products with serogroup O17-specific primers using template DNA from different S. enterica serogroups. Lanes 1 and 5, serogroup O17 strains; lane 2, serogroup C₁; lane 3, serogroup B; lane 4, serogroup D₁; lane 6, 100-bp ladder (Gibco). (B) PCR amplification products with serogroup O18-specific primers using template DNA from different S. enterica serogroups. Lanes 1 and 7, 100-bp ladders (Gibco); lanes 2, 4, 5, and 6, serogroup O18 strains; lane 3, serogroup B strain.

Conclusions.

Top

Conclusions.

DNA sequence analysis of the rfb gene cluster from Salmonellaserogroup O17 had limited identity to previously characterizedrfb gene clusters and appears to have been acquired by lateraltransfer from a low-G+C-content species. Evolution of most ofthe O18 rfb gene cluster also occurred within a low-G+C-contentspecies prior to transfer to S. enterica, but the cluster appearsto have been assembled from more than one source as the manCand manB genes at least appear to have been derived from theSalmonella cps cluster. The O18 rfb gene cluster is most relatedto Salmonella serogroups O6,14 and C1. The PCR assays describedhere form the basis for the development of molecular detectionassays for the identification of Salmonella serogroups O17 andO18 based on genes directly involved in the formation of theO antigens. Together with the serogroup PCR assays that havepreviously been described for serogroups A/D, B, C₂, and D (16)and, more recently, O6,14 (9), we intend to identify additionalrfb targets specific for other Salmonella O antigens associatedwith the top 100 serotypes of human Salmonella infection. Ouroverall goal is to combine this serogroup identification systemwith DNA targets specific for H antigens (17) to establish acomprehensive DNA-based scheme for the identification of themajor Salmonella serotypes (8). Currently, we are exploringthe utility of additional technologies, including Luminex, asa potential platform to combine the O- and H-antigen molecularassays on a single platform. This will allow a rapid and convenientalternative for the definitive identification of Salmonellaserotypes attainable by nonspecialized laboratories.

Nucleotide sequence accession numbers.

Top

Nucleotide sequence accession...

The DNA sequences have been deposited in GenBank and assignedaccession numbers EF032634 and EF032635.

ACKNOWLEDGMENTS

We thank Bernard Henrissat (University of Marseille, France)for classifying the glycosyltransferases from Salmonella serogroupsO17 and O18 using the CAZy scheme and Kate Wilson and PeterReeves (University of Sydney, Australia) for assigning BPGDgene names for the Salmonella O17 and O18 rfb gene clusters.

The use of trade names is for identification only and does notimply endorsement by the Public Health Service or by the U.S.Department of Health and Human Services.

FOOTNOTES

* Corresponding author. Mailing address: Foodborne and Diarrheal Diseases Branch, Division of Bacterial and Mycotic Diseases, NCID, Centers for Disease Control and Prevention, 1600 Clifton Road NE, Atlanta, GA 30333. Phone: (404) 639-0838. Fax: (404) 639-3333. E-mail: chf3{at}cdc.gov.

Published ahead of print on 20 October 2006.

REFERENCES

Top