Comparison of Campylobacter jejuni Lipooligosaccharide Biosynthesis Loci from a Variety of Sources
http://www.100md.com
微生物临床杂志 2005年第6期
Produce Safety and Microbiology Unit, WRRC, USDA, ARS, 800 Buchanan St., Albany, California 94710
Institute for Biological Sciences, National Research Council of Canada, 100 Sussex Dr., Ottawa, Ontario K1A 0R6
Bacteriology and Enteric Diseases Program, National Microbiology Laboratory, Winnipeg, Canada
ABSTRACT
Campylobacter jejuni strains exhibit significant variation in the genetic content of the lipooligosaccharide (LOS) biosynthesis loci with concomitant differences in LOS structure. The C. jejuni LOS loci have been grouped into six classes based on gene content and organization. Utilizing PCR amplifications of genes from these loci, we were able to classify a majority (80%) of the LOS biosynthesis loci from 123 strains of C. jejuni that included 39 of the Penner serotype reference strains. We found that a particular LOS class was not always associated with a specific Penner serotype, and 14 of 16 Guillain-Barre syndrome-associated isolates tested in this study shared the same LOS class. The remaining isolates that could not be classified were often distinguishable from each other based on the results of gene-specific PCR and the lengths of their LOS biosynthesis loci as determined by long (XL) PCR. Sequence analysis of two of these unique XL PCR products demonstrated two new LOS classes. These results support the hypothesis that the LOS locus is a hot spot for genetic exchange and rearrangements. Analysis of the LOS biosynthesis genes by PCR assays can be used for typing C. jejuni and offers the advantage of inferring potential LOS structures.
INTRODUCTION
Campylobacter jejuni is the leading cause of acute bacterial gastroenteritis worldwide. Campylobacteriosis, in rare cases, is a likely antecedent to the development of peripheral neuropathies, Guillain-Barre syndrome (GBS), and Miller Fisher syndrome (29, 30). The lipooligosaccharides (LOS) of several C. jejuni strains have been shown to exhibit molecular mimicry with gangliosides concentrated in peripheral nerves, and it has been speculated that the peripheral neuropathies are related directly to autoimmune mechanisms following infection (20). However, not all strains of C. jejuni exhibit ganglioside mimicry, and it is estimated that fewer than 1 per 1,000 Campylobacter infections are followed by GBS (1, 18).
A cluster of genes involved in C. jejuni LOS biosynthesis was identified by analysis of the complete genome sequence of C. jejuni NCTC 11168; it extends from Cj1131c (galE) to Cj1151c (rfaD) (16, 17, 24). The results of both sequencing of LOS biosynthesis loci from other C. jejuni strains (5, 7, 10, 11) and microarray analysis (2, 14, 15, 25) indicated that the LOS biosynthesis locus is one of the more variable regions of the C. jejuni genome. Based on gene content and organization of the LOS biosynthesis loci, LOS classes A, B, and C have been described (7). Along with class-specific glycosyltransferases, these classes possess neuBCA genes that are required for sialic acid biosynthesis and a cst gene that encodes a sialic acid transferase (7, 10, 17). Thus, these three classes encode genes responsible for the production of sialylated LOS that are ganglioside mimics (7). Moreover, it was demonstrated that the LOS structures did not always correlate with a particular Penner serotype.
The Penner serotyping system relies on differences in Campylobacter heat-stable antigens, originally proposed to be LOS and/or lipopolysaccharide-type molecules (26, 27). Recently, capsular polysaccharides were shown to account for the Penner serotype specificity of several serotypes (12). To gain a greater understanding with regard to the diversity of the C. jejuni LOS loci and the potential relationship to Penner serotypes, we examined the composition of over 100 C. jejuni LOS biosynthesis loci. We were able to classify approximately 80% of the LOS loci and observed that over 60% of these loci belong to class A, B, or C. We also determined the genetic composition of two loci that were not classified.
MATERIALS AND METHODS
Bacterial strains and growth conditions. Bacterial strains used in this study are listed in Table 1. C. jejuni strains were cultured at 42°C under microaerophilic conditions (5% O2, 10% CO2, and 85% N2) on Brucella agar supplemented with 0.025% (wt/vol) FeSO4 · 7H2O, 0.025% (wt/vol) sodium metabisulfite (anhydrous), and 0.025% (wt/vol) sodium pyruvate (anhydrous).
Preparation of C. jejuni genomic DNA. C. jejuni cells were scraped from a plate and resuspended in 1.5 ml 10% (wt/vol) sucrose-50 mM Tris (pH 8.0), to which was added 250 μl of a 10-mg ml–1 lysozyme solution (in 250 mM Tris, pH 8.0), followed by 600 μl of 0.1 M EDTA. The suspension was incubated for 10 min on ice, and 300 μl of a 5% (wt/vol) sodium dodecyl sulfate solution was added and then briefly vortexed to clarify the solution. RNaseA (1 mg ml–1) and proteinase K (10 mg ml–1) were added, sequentially, and the lysates incubated for 30 min and 1 h at 37°C, respectively. Sodium acetate and ethanol were added, and DNA was removed by spooling onto a hooked Pasteur pipette. DNA was resuspended in Tris-EDTA (pH 8.0), extracted twice with phenol-chloroform (1:1, vol/vol) and once with chloroform, and concentrated by ethanol precipitation. Genomic DNA was also isolated from bacteria by using the DNeasy tissue kit (QIAGEN, Valencia, CA).
PCR. Conventional PCR reagents were supplied by Epicentre (Madison, WI). Each PCR consisted of 1x MasterAmp Taq PCR buffer, 1x MasterAmp Taq Enhancer, 2.5 mM MgCl2, 200 μM each deoxynucleoside triphosphate, forward and reverse primers at 0.2 μM each, 0.2 U of MasterAmp Taq DNA polymerase (Epicentre) or Taq DNA polymerase (NEB, Beverly, MA), and approximately 50 ng of genomic DNA (final reaction volume, 25 μl). The LOS gene-specific primers are described in Table 2. Amplification conditions for all LOS gene-specific products occurred under the following parameters: 30 cycles of 25 s at 94°C, 25 s at 52°C, and 1 min at 72°C and a final extension at 72°C for 5 min. Long PCR (XL PCR) reagents were supplied by Epicentre (Madison, WI). Each XL PCR consisted of 1x MasterAmp Extra-Long PreMix 5, LOSXL primers (Table 2) at 0.2 μM each, 2.5 U of MasterAmp Extra-Long DNA polymerase, and 250 ng of genomic DNA (final reaction volume, 50 μl). The cycling conditions for XL PCR products were 25 cycles of 30 s at 94°C, 45 s at 52°C, and 15 min at 68°C and a final extension at 68°C for 15 min. Thermal cycling was performed with a Tetrad thermal cycler (MJ Research, Waltham, MA). All PCR products were analyzed by agarose gel electrophoresis. Positive samples were identified based on the presence of bands of anticipated sizes. Primers were purchased from either Operon Technologies (Alameda, CA) or Sigma-Genosys (The Woodlands, TX).
DNA sequencing. Cycle sequencing reactions were performed using the ABI PRISM BigDye terminator cycle sequencing kit (version 3.0). All extension products were purified on DyeEx spin columns (QIAGEN, Valencia, CA) or Centri-Sep spin columns (Princeton Separations, Princeton, NJ). DNA sequencing was performed using an ABI PRISM 3100 Genetic Analyzer (Applied Biosystems, Foster City, CA) with the POP-6 polymer and ABI PRISM Genetic Analyzer Data Collection and ABI PRISM Genetic Analyzer Sequencing Analysis software. Sequencing oligonucleotide primers were purchased from Operon Technologies (Alameda, CA).
Nucleotide sequence accession numbers. Nucleotide sequences have been deposited with GenBank and assigned accession numbers as follows: RM1170 LOS loci, AY434498; RM1555 (ATCC 42437) LOS loci, AY436358; RM1047 (ATCC 43431), AY800272.
RESULTS AND DISCUSSION
Specificity of LOS biosynthesis locus class-specific PCR primers. Based on gene content and organization of the LOS biosynthesis loci from 15 different strains of C. jejuni (3, 5, 7, 8, 10, 11, 22, 23), we grouped the loci into six major LOS classes: A, B, C, D, E, and F (Fig. 1). Classes A, B, and C, involved in sialylated LOS synthesis, were previously described (7). The sequences for classes D, E, and F were obtained from GenBank (3, 8, 22, 23) Each locus possesses a number of novel putative glycosyltransferase genes and genes with unknown function but lacks the neuBCA and cst genes and likely directs the synthesis of nonsialylated LOS structures. We evaluated 42 primer pairs (Table 2) specific to the genes within the six classes designated in Fig. 1 by PCR amplification of genomic DNA from strains representing each defined LOS class (Table 3). All C. jejuni LOS classes were amplified with the waaM (orf2) and waaV (orf12) primer sets. Strains RM1048, RM1556, RM1050, and RM1052 (classes A and B) share many amplification products, including cstII (orf7ab), neuB1-ab (orf8ab), neuC1-ab (orf9ab), neuA1-ab (orf10ab), and orf11ab. The cgtA2 (orf5bII) product, however, is specific for class B and can be used to distinguish class A and class B. Previous work (7) demonstrated sequence divergence of a region spanning the cgtA (orf5abI) and cgtB (orf6ab) genes (1,200 bp), creating two sets of alternative alleles: orf5abI1 and orf6ab1 or orf5abI2 and orf6ab2. These alternative alleles are present in both class A and class B. We designed allele-specific primers for these genes, such that we amplified orf5abI1 and orf6ab1 from strains RM1048 (ATCC 43432) (class A1) and RM1052 (ATCC 43456) (class B1) but not orf5abI2 or orf6ab2. This allowed us to designate the alleles for class A (A1 and A2) and class B (B1 and B2).
The products for Cj1137c (orf14c), Cj1138 (orf15c), Cj1139c(orf6c), cstIII (orf7c), neuB1-c (orf8c), neuC1-c (orf9c),cgtAneuA1-c (orf5c/10c), and orf16c were amplified specifically from the class C loci. Classes D and F share many amplification products, including orf18d, orf19df, orf20df, and orf16df. The orf17d product, specific for class D, can distinguish classes D and F. Only class E C. jejuni strains produced PCR products for the orf21e, 22e, 25e, 26e, 27e, 28e, 29e, 30e, 31e, 32e, 33e, and 34e genes. Surprisingly, strain 81116 (class E) produced amplification products with neuB1-ab (orf8ab) and neuC1-ab (orf9ab). The identities of these products were confirmed by sequencing and suggest the presence of these genes elsewhere in the genome.
LOS biosynthesis locus diversity of the Penner serotype reference strains. From the class-specific primer pairs, we identified 12 primer pairs representing genes from the six classes (orf6ab1, orf6ab2, orf7ab, orf8ab, orf5bII, orf6c, orf7c, orf8c, orf17d, orf18df, orf3e, and orf27e) that would allow putative class determination and distinguish between the two alleles of classes A and B. As a positive control for the presence of a C. jejuni LOS locus, we amplified the orf12 (waaV) product. Thirty-nine Penner serotype reference strains were examined. Twenty-five of the reference strains (HS:1 to HS:8, HS:10, HS:11, HS:13, HS:16, HS:18, HS:19, HS:23, HS:27, HS:29, HS:32, HS:35, HS:36, HS:41, HS:44, HS:45, HS:62, and HS:64) showed PCR-amplification patterns consistent with possessing class A1, A2, B1, B2, C, D, E, or F (Table 4). For these strains, we attempted to verify the LOS classes by additional amplifications of genes within the presumed class. Most strains had patterns consistent with their LOS class designation; however, we identified a number of strains with patterns that diverged from the expected class patterns. Most presumptive class E strains—RM3407 (HS:3), RM3411 (HS:7), RM3428 (HS:37), RM3429 (HS:38), and RM3432 (HS:45)—failed to produce two class E primer amplicons, orf26e and orf28e, while RM3423 (HS:27) failed to produce the orf28e amplicon. The sizes of the XL PCR products for the LOS loci of these strains (extending from orf2 [waaM] to orf12 [waaV]) were approximately 1 kb smaller than those of the class E strains. Similarly, the presumptive class D strain RM3418 (HS:17) and the presumptive class F strain RM3437 (HS:58) failed to produce the orf19df product and had XL PCR products that were slightly larger than those of the class D and class F strains (data not shown).
The remaining 15 Penner serotype reference strains (HS:9, HS:17, HS:22, HS:32, HS:37, HS:38, HS:43, HS:50, HS:52, HS:53, HS:57, HS:58, HS:60, HS:63, and HS:65) failed to produce PCR amplification patterns consistent with the six classes by use of the 12 primers. Some of these unclassified LOS loci possessed LOS genes from known classes. As with RM3407, these loci may represent a new class that is derived from a known class. The LOS loci of these strains are likely novel, or they possess extensive sequence divergence that prevents PCR amplification. Interestingly, strains RM3434 (HS:52) and RM3442 (HS:65) appeared to exhibit multiple LOS classes: B2 and C for RM3434 and A2 and F for RM3442. The XL PCR products for RM3434 and RM3442 are similar in size to the C and A2 XL PCR products, respectively. This suggests that these LOS loci are not likely a combination of the two classes. DNA sequencing will be required to determine the exact LOS class for each of these isolates and to address where the genes from the other LOS class are.
LOS classes of isolates from clinical and environmental sources. To investigate any associations between LOS classes and isolate sources, we screened 70 clinical and environmental isolates with the 12 primer pairs. We surveyed 16 GBS-associated isolates and found that 14 were positive for class A1 products, whereas the other 2 isolates were positive for ganglioside-associated classes A2 and C (Table 5). This is consistent with a study by Nachamkin et al. (21), who previously demonstrated a strong association of cst-II (orf7a,b), cgtA (orf5), and cgtB (orf6) with GBS isolates. It should be noted that most of the GBS-associated strains in this study (12 of 16 GBS strains) were HS:19 or HS:41, which have been shown to be genetically clonal (13, 28). This overrepresentation certainly increases the occurrence of the class A1 locus in this study. Examining additional non-HS:19 and non-HS:41 GBS-associated strains would provide a better indication of the relationship between LOS class and this syndrome. Indeed, Godschalk et al. recently examined 17 GBS-associated isolates and found that 53% possessed the class A locus (9). Interestingly, they also found that 18% of the GBS-associated strains possessed LOS loci with no known sialic acid transferase gene. These strains may possess a unique sialic acid transferase that allows the synthesis of ganglioside mimics. It is also possible that these strains coinfected with other C. jejuni strains capable of synthesizing a ganglioside mimic and are merely the strains recovered by culturing.
For other clinical isolates and for environmental isolates (Table 6), we observed strains with putative LOS locus classes A, B, C, D, E, and F, and we were also able to reclassify the LOS locus for several the putative class E isolates to class H (described below) by using the orf26e primer pair. As with the Penner serotyping reference strains, the LOS locus class of several isolates could not be determined by the PCR screen. The most striking finding was that 64% (35 of 55) of the non-GBS-associated isolates possessed an LOS locus class (A, B, or C) capable of synthesizing a ganglioside mimic. Similarly, of the 21 isolates associated with enteritis, Godschalk et al. found that 62% possessed a class A, B, or C locus (9). It is not clear why the incidence of strains capable of producing ganglioside mimics from all sources is so high, although these findings seem to suggest that these ganglioside-mimicking LOS structures are advantageous to C. jejuni for colonization or infection of various hosts. What is clear is that the production of ganglioside-mimicking LOS structures alone is not sufficient to elicit GBS; certain host factors and/or other bacterial factors are required.
Distribution of LOS classes within a group of Penner serotype isolates. Although there is strong evidence suggesting that the capsule rather than LOS is responsible for many Penner serotypes (12), it is possible that there remains a correlation between LOS genotype and Penner serotype. Indeed, all HS:19 and HS:41 strains examined share the same class A1 locus (see Tables 5 and 6). With only a limited number of strains possessing the same Penner type, we observed that Penner type and LOS class are not consistently associated. The HS:10 isolates (RM1556, RM1413, and RM1507) contained class A1 or A2, and the HS:2 isolates (RM1046, RM1188, RM1478, RM1480, and RM1862) contained class A1, B2, or C. Horizontal transfer of the LOS locus (or the capsular locus) provides a plausible mechanism for different LOS classes between strains with the same Penner serotype, as evidenced by the characterization of the class A1 LOS locus of GB11 (HS:2) (6). However, alternative mechanisms would be required to maintain the A1 LOS locus class in all the HS:19 and HS:41 strains examined in this study.
Identification of two new LOS locus classes. To verify novel LOS loci, we sequenced the XL PCR products from strains RM1555 (ATCC 43437) (GenBank accession number AY436358) and RM1047 (ATCC 43431) (GenBank accession number AY800272). The amplification pattern of RM1555 was inconsistent with the known LOS classes; it was positive for orf3e and waaV (orf12) products and negative for all other products. The sequence of the 5,537-bp region between waaM (Cj1134) and waaV (Cj1146c) establishes a seventh LOS locus class, G (Fig. 2). This locus contains four newly described open reading frames (ORFs) that encode putative glycosyltransferases based on BLASTX scores. A fifth potential ORF contains an intergenic homopolymeric G tract with 8 G bases (versus 9 G bases) that disrupts the reading frame after 336 nucleotides. By pairwise alignment, these 336 nucleotides show 93.5% identity to the first 322 nucleotides of the 888-nucleotide orf16d from RMLIO87 and 92.9% identity across the entire corresponding 888-nucleotide region. orf16f from RM1170 is similarly truncated, with 8 G bases in the homopolymeric G tract, and is 93.8% identical by pairwise alignment. The orf16df PCR product was not amplified from RM1555 due to significant sequence differences at the binding site of primer orf16df (data not shown). Interestingly, orf16d and orf16f are located adjacent to orf12d and orf12f in LIO87 and RM1170, respectively, but orf16g is separated from orf12g by orf38g, a pattern that presumably resulted from an insertion event.
The previous PCR results suggested that the LOS locus of RM1047 (ATCC 43431) differed from the class E locus but was similar to the presumptive class H locus. Indeed, the XL PCR product from RM1047 (ATCC 43431) was the same size as the class H LOS product (data not shown). Sequence analysis verified that this LOS locus represents a new class, termed class H, that is distinct from class E (Fig. 2). The orf28e gene is clearly missing from this locus, but the 3' end of the orf26e gene is still present (Fig. 2). Furthermore, these changes appear to be the results of two different events. The absence of orf26e seems to be the result of an insertion-deletion event, with the 3' end of orf25e still present and the 5' region replaced by a putative butyryltransferase gene (orf39h). The second event appears to have been a deletion event that resulted in the complete removal of orf28e from RM1047, with minor changes to the 3' end of orf27e and the 5' end of orf29e. Considering these differences, we were able to reassign many of the isolates that were initially identified as class E to class H based on the PCR results for orf26e and orf28e (Tables 4 and 6). Furthermore, we identified hypervariable homopolymeric G tracts within orf23e (9 or 10 G bases) and orf25e (8 or 9 G bases). Therefore, all classes of LOS loci examined so far appear to have this common mechanism for creating LOS variability.
Conclusions. This study reports a method to examine the diversity and classification of the LOS biosynthesis loci of C. jejuni. The PCR method identified the presence of genes from six LOS classes and was rapid, sensitive, and specific. The LOS structure of a particular isolate is determined not only by the genes of the locus but also by the presence of various mutations. Thus, strains sharing the same LOS class do not necessarily express the same LOS structure (7). Unfortunately, this method is unable to detect certain insertions, deletions, and point mutations that affect LOS structure. Nevertheless, we can infer LOS structures once an isolate's LOS class has been determined, particularly the presence or absence of ganglioside mimicry. Considering the potential role of C. jejuni ganglioside-mimicking structures in eliciting GBS or Miller Fisher syndrome, we believe that the LOS class of clinical strains provides valuable information and should be incorporated into epidemiological studies. Indeed, we found that all GBS-related strains and 64% of the other clinical and environmental isolates examined in this study belonged to an LOS class (LOS class A, B, or C) that allowed the synthesis of a sialylated LOS. The overrepresentation of HS:19 and HS:41 isolates likely biased the occurrence of LOS class A1 among GBS-associated strains. However, Godschalk et al. (9) similarly observed a significant association of LOS class A with GBS-associated isolates. Clearly, additional analysis with a more diverse group of GBS-associated strains is required to determine the significance of the LOS class.
Although not all LOS loci could be classified at this time, we demonstrated the ability to identify new loci, and we characterized the sequences of two of these, increasing the number of LOS classes to eight. For these new LOS classes, class G and class H, we observed evidence of genetic rearrangements (deletions or insertions) and homopolymeric G tracts that can lead to LOS structural variability. Specifically, the LOS synthesized by the LOS class H strain RM1047 could differ from those of other class H strains due to differences in the lengths of G tracts within biosynthesis genes. Also, LOS class H appears to be a derivative of class E, and therefore the LOS from these classes may share structural features. Additional analysis of C. jejuni LOS structures may allow potential structures to be inferred once an LOS class has been identified. Furthermore, the other unclassified LOS loci certainly substantiate the need for additional studies, and we envision the development of a microarray-based method that allows simultaneous analysis for distinguishing C. jejuni LOS classes.
ACKNOWLEDGMENTS
We thank S. Abbott, M. Englen, P. Guerry, W. Johnson, A. Lastovica, R. Meinersmann, I. Nachamkin, M. Nicholson, L. Stanker, and I. Wesley for providing strains. We also thank J. Klena for critical reading of the manuscript.
This work was supported by the United States Department of Agriculture, Agricultural Research Service CRIS project 5325-42000-041. M.G. was supported by the NRC Genome Sciences and Health Related Research Initiative.
REFERENCES
Allos, B. M. 1997. Association between Campylobacter infection and Guillain-Barre syndrome. J. Infect. Dis. 176(Suppl. 2):S125-S128.
Dorrell, N., J. A. Mangan, K. G. Laing, J. Hinds, D. Linton, H. Al-Ghusein, B. G. Barrell, J. Parkhill, N. G. Stoker, A. V. Karlyshev, P. D. Butcher, and B. W. Wren. 2001. Whole genome comparison of Campylobacter jejuni human isolates using a low-cost microarray reveals extensive genetic diversity. Genome Res. 11:1706-1715.
Fry, B. N., V. Korolik, B. A. van der Zeijst, and P. J. Coloe. 1998. A gene cluster from Campylobacter jejuni involved in inner core and lipid A synthesis. NCBI accession number AJ131360. Applied Biology and Biotechnology, Royal Melbourne Institute of Technology University, Melbourne, Australia.
Gilbert, M., J. Bellefeuille, M. F. Karwaski, A. M. Cunningham, and W. W. Wakarchuk. 2001. Partial sequence of the lipooligosaccharide biosynthesis locus of Campylobacter jejuni O:3. NCBI accession number AF411225. Institute for Biological Sciences, National Research Council of Canada, Montreal, Canada.
Gilbert, M., J. R. Brisson, M. F. Karwaski, J. Michniewicz, A. M. Cunningham, Y. Wu, N. M. Young, and W. W. Wakarchuk. 2000. Biosynthesis of ganglioside mimics in Campylobacter jejuni OH4384. Identification of the glycosyltransferase genes, enzymatic synthesis of model compounds, and characterization of nanomole amounts by 600-mHz 1H and 13C NMR analysis. J. Biol. Chem. 275:3896-3906.
Gilbert, M., P. C. Godschalk, M. F. Karwaski, C. W. Ang, A. van Belkum, J. Li, W. W. Wakarchuk, and H. P. Endtz. 2004. Evidence for acquisition of the lipooligosaccharide biosynthesis locus in Campylobacter jejuni GB11, a strain isolated from a patient with Guillain-Barre syndrome, by horizontal exchange. Infect. Immun. 72:1162-1165.
Gilbert, M., M. F. Karwaski, S. Bernatchez, N. M. Young, E. Taboada, J. Michniewicz, A. M. Cunningham, and W. W. Wakarchuk. 2002. The genetic basis for the variation in the lipo-oligosaccharide of the mucosal pathogen, Campylobacter jejuni. Biosynthesis of sialylated ganglioside mimics in the core oligosaccharide. J. Biol. Chem. 277:327-337.
Gilbert, M., R. K. Singh, M.-F. Karwaski, and W. W. Wakarchuk. 2002. Sequencing of the lipo-oligosaccharide biosynthesis locus of Campylobacter jejuni LIO87. NCBI accession number AF40069. Institute for Biological Sciences, National Research Council of Canada, Ottawa, Ontario, Canada.
Godschalk, P. C., A. P. Heikema, M. Gilbert, T. Komagamine, C. W. Ang, J. Glerum, D. Brochu, J. Li, N. Yuki, B. C. Jacobs, A. van Belkum, and H. P. Endtz. 2004. The crucial role of Campylobacter jejuni genes in anti-ganglioside antibody induction in Guillain-Barre syndrome. J. Clin. Investig. 114:1659-1665.
Guerry, P., C. P. Ewing, T. E. Hickey, M. M. Prendergast, and A. P. Moran. 2000. Sialylation of lipooligosaccharide cores affects immunogenicity and serum resistance of Campylobacter jejuni. Infect. Immun. 68:6656-6662.
Guerry, P., C. M. Szymanski, M. M. Prendergast, T. E. Hickey, C. P. Ewing, D. L. Pattarini, and A. P. Moran. 2002. Phase variation of Campylobacter jejuni 81-176 lipooligosaccharide affects ganglioside mimicry and invasiveness in vitro. Infect. Immun. 70:787-793.
Karlyshev, A. V., D. Linton, N. A. Gregson, A. J. Lastovica, and B. W. Wren. 2000. Genetic and biochemical evidence of a Campylobacter jejuni capsular polysaccharide that accounts for Penner serotype specificity. Mol. Microbiol. 35:529-541.
Lastovica, A. J., E. A. Goddard, and A. C. Argent. 1997. Guillain-Barre syndrome in South Africa associated with Campylobacter jejuni O:41 strains. J. Infect. Dis. 176(Suppl. 2):S139-S143.
Leonard, E. E., II, T. Takata, M. J. Blaser, S. Falkow, L. S. Tompkins, and E. C. Gaynor. 2003. Use of an open-reading frame-specific Campylobacter jejuni DNA microarray as a new genotyping tool for studying epidemiologically related isolates. J. Infect. Dis. 187:691-694.
Leonard, E. E., II, L. S. Tompkins, S. Falkow, and I. Nachamkin. 2004. Comparison of Campylobacter jejuni isolates implicated in Guillain-Barre syndrome and strains that cause enteritis by a DNA microarray. Infect. Immun. 72:1199-1203.
Linton, D., M. Gilbert, P. G. Hitchen, A. Dell, H. R. Morris, W. W. Wakarchuk, N. A. Gregson, and B. W. Wren. 2000. Phase variation of a -1,3 galactosyltransferase involved in generation of the ganglioside GM1-like lipo-oligosaccharide of Campylobacter jejuni. Mol. Microbiol. 37:501-514.
Linton, D., A. V. Karlyshev, P. G. Hitchen, H. R. Morris, A. Dell, N. A. Gregson, and B. W. Wren. 2000. Multiple N-acetyl neuraminic acid synthetase (neuB) genes in Campylobacter jejuni: identification and characterization of the gene involved in sialylation of lipo-oligosaccharide. Mol. Microbiol. 35:1120-1134.
McCarthy, N., Y. Andersson, V. Jormanainen, O. Gustavsson, and J. Giesecke. 1999. The risk of Guillain-Barre syndrome following infection with Campylobacter jejuni. Epidemiol. Infect. 122:15-17.
Miller, W. G., A. H. Bates, S. T. Horn, M. T. Brandl, M. R. Wachtel, and R. E. Mandrell. 2000. Detection on surfaces and in Caco-2 cells of Campylobacter jejuni cells transformed with new gfp, yfp, and cfp marker plasmids. Appl. Environ. Microbiol. 66:5426-5436.
Moran, A. P. 1997. Structure and conserved characteristics of Campylobacter jejuni lipopolysaccharides. J. Infect. Dis. 176(Suppl. 2):S115-S121.
Nachamkin, I., J. Liu, M. Li, H. Ung, A. P. Moran, M. M. Prendergast, and K. Sheikh. 2002. Campylobacter jejuni from patients with Guillain-Barre syndrome preferentially expresses a GD1a-like epitope. Infect. Immun. 70:5299-5303.
Oldfield, N. J., A. P. Moran, L. A. Millar, M. M. Prendergast, and J. M. Ketley. 2002. Characterization of the Campylobacter jejuni heptosyltransferase II gene, waaF, provides genetic evidence that extracellular polysaccharide is lipid A core independent. J. Bacteriol. 184:2100-2107.
Parker, C. T. 2004. Diversity in the lipooligosaccharide biosynthesis locus of Campylobacter jejuni. NCBI accession number AY434498. Produce Safety and Microbiology Unit, United States Department of Agriculture, Agriculture Research Service, Albany, Calif.
Parkhill, J., B. W. Wren, K. Mungall, J. M. Ketley, C. Churcher, D. Basham, T. Chillingworth, R. M. Davies, T. Feltwell, S. Holroyd, K. Jagels, A. V. Karlyshev, S. Moule, M. J. Pallen, C. W. Penn, M. A. Quail, M. A. Rajandream, K. M. Rutherford, A. H. van Vliet, S. Whitehead, and B. G. Barrell. 2000. The genome sequence of the food-borne pathogen Campylobacter jejuni reveals hypervariable sequences. Nature 403:665-668.
Pearson, B. M., C. Pin, J. Wright, K. I'Anson, T. Humphrey, and J. M. Wells. 2003. Comparative genome analysis of Campylobacter jejuni using whole genome DNA microarrays. FEBS Lett. 554:224-230.
Penner, J. L., and J. N. Hennessy. 1980. Passive hemagglutination technique for serotyping Campylobacter fetus subsp. jejuni on the basis of soluble heat-stable antigens. J. Clin. Microbiol. 12:732-737.
Penner, J. L., J. N. Hennessy, and R. V. Congi. 1983. Serotyping of Campylobacter jejuni and Campylobacter coli on the basis of thermostable antigens. Eur. J. Clin. Microbiol. 2:378-383.
Prendergast, M. M., A. J. Lastovica, and A. P. Moran. 1998. Lipopolysaccharides from Campylobacter jejuni O:41 strains associated with Guillain-Barre syndrome exhibit mimicry of GM1 ganglioside. Infect. Immun. 66:3649-3655.
van Koningsveld, R., R. Rico, I. Gerstenbluth, P. I. Schmitz, C. W. Ang, I. S. Merkies, B. C. Jacobs, Y. Halabi, H. P. Endtz, F. G. van der Meche, and P. A. van Doorn. 2001. Gastroenteritis-associated Guillain-Barre syndrome on the Caribbean island Curacao. Neurology 56:1467-1472.
Willison, H. J., and G. M. O'Hanlon. 1999. The immunopathogenesis of Miller Fisher syndrome. J. Neuroimmunol. 100:3-12.(Craig T. Parker, Sharon T)
Institute for Biological Sciences, National Research Council of Canada, 100 Sussex Dr., Ottawa, Ontario K1A 0R6
Bacteriology and Enteric Diseases Program, National Microbiology Laboratory, Winnipeg, Canada
ABSTRACT
Campylobacter jejuni strains exhibit significant variation in the genetic content of the lipooligosaccharide (LOS) biosynthesis loci with concomitant differences in LOS structure. The C. jejuni LOS loci have been grouped into six classes based on gene content and organization. Utilizing PCR amplifications of genes from these loci, we were able to classify a majority (80%) of the LOS biosynthesis loci from 123 strains of C. jejuni that included 39 of the Penner serotype reference strains. We found that a particular LOS class was not always associated with a specific Penner serotype, and 14 of 16 Guillain-Barre syndrome-associated isolates tested in this study shared the same LOS class. The remaining isolates that could not be classified were often distinguishable from each other based on the results of gene-specific PCR and the lengths of their LOS biosynthesis loci as determined by long (XL) PCR. Sequence analysis of two of these unique XL PCR products demonstrated two new LOS classes. These results support the hypothesis that the LOS locus is a hot spot for genetic exchange and rearrangements. Analysis of the LOS biosynthesis genes by PCR assays can be used for typing C. jejuni and offers the advantage of inferring potential LOS structures.
INTRODUCTION
Campylobacter jejuni is the leading cause of acute bacterial gastroenteritis worldwide. Campylobacteriosis, in rare cases, is a likely antecedent to the development of peripheral neuropathies, Guillain-Barre syndrome (GBS), and Miller Fisher syndrome (29, 30). The lipooligosaccharides (LOS) of several C. jejuni strains have been shown to exhibit molecular mimicry with gangliosides concentrated in peripheral nerves, and it has been speculated that the peripheral neuropathies are related directly to autoimmune mechanisms following infection (20). However, not all strains of C. jejuni exhibit ganglioside mimicry, and it is estimated that fewer than 1 per 1,000 Campylobacter infections are followed by GBS (1, 18).
A cluster of genes involved in C. jejuni LOS biosynthesis was identified by analysis of the complete genome sequence of C. jejuni NCTC 11168; it extends from Cj1131c (galE) to Cj1151c (rfaD) (16, 17, 24). The results of both sequencing of LOS biosynthesis loci from other C. jejuni strains (5, 7, 10, 11) and microarray analysis (2, 14, 15, 25) indicated that the LOS biosynthesis locus is one of the more variable regions of the C. jejuni genome. Based on gene content and organization of the LOS biosynthesis loci, LOS classes A, B, and C have been described (7). Along with class-specific glycosyltransferases, these classes possess neuBCA genes that are required for sialic acid biosynthesis and a cst gene that encodes a sialic acid transferase (7, 10, 17). Thus, these three classes encode genes responsible for the production of sialylated LOS that are ganglioside mimics (7). Moreover, it was demonstrated that the LOS structures did not always correlate with a particular Penner serotype.
The Penner serotyping system relies on differences in Campylobacter heat-stable antigens, originally proposed to be LOS and/or lipopolysaccharide-type molecules (26, 27). Recently, capsular polysaccharides were shown to account for the Penner serotype specificity of several serotypes (12). To gain a greater understanding with regard to the diversity of the C. jejuni LOS loci and the potential relationship to Penner serotypes, we examined the composition of over 100 C. jejuni LOS biosynthesis loci. We were able to classify approximately 80% of the LOS loci and observed that over 60% of these loci belong to class A, B, or C. We also determined the genetic composition of two loci that were not classified.
MATERIALS AND METHODS
Bacterial strains and growth conditions. Bacterial strains used in this study are listed in Table 1. C. jejuni strains were cultured at 42°C under microaerophilic conditions (5% O2, 10% CO2, and 85% N2) on Brucella agar supplemented with 0.025% (wt/vol) FeSO4 · 7H2O, 0.025% (wt/vol) sodium metabisulfite (anhydrous), and 0.025% (wt/vol) sodium pyruvate (anhydrous).
Preparation of C. jejuni genomic DNA. C. jejuni cells were scraped from a plate and resuspended in 1.5 ml 10% (wt/vol) sucrose-50 mM Tris (pH 8.0), to which was added 250 μl of a 10-mg ml–1 lysozyme solution (in 250 mM Tris, pH 8.0), followed by 600 μl of 0.1 M EDTA. The suspension was incubated for 10 min on ice, and 300 μl of a 5% (wt/vol) sodium dodecyl sulfate solution was added and then briefly vortexed to clarify the solution. RNaseA (1 mg ml–1) and proteinase K (10 mg ml–1) were added, sequentially, and the lysates incubated for 30 min and 1 h at 37°C, respectively. Sodium acetate and ethanol were added, and DNA was removed by spooling onto a hooked Pasteur pipette. DNA was resuspended in Tris-EDTA (pH 8.0), extracted twice with phenol-chloroform (1:1, vol/vol) and once with chloroform, and concentrated by ethanol precipitation. Genomic DNA was also isolated from bacteria by using the DNeasy tissue kit (QIAGEN, Valencia, CA).
PCR. Conventional PCR reagents were supplied by Epicentre (Madison, WI). Each PCR consisted of 1x MasterAmp Taq PCR buffer, 1x MasterAmp Taq Enhancer, 2.5 mM MgCl2, 200 μM each deoxynucleoside triphosphate, forward and reverse primers at 0.2 μM each, 0.2 U of MasterAmp Taq DNA polymerase (Epicentre) or Taq DNA polymerase (NEB, Beverly, MA), and approximately 50 ng of genomic DNA (final reaction volume, 25 μl). The LOS gene-specific primers are described in Table 2. Amplification conditions for all LOS gene-specific products occurred under the following parameters: 30 cycles of 25 s at 94°C, 25 s at 52°C, and 1 min at 72°C and a final extension at 72°C for 5 min. Long PCR (XL PCR) reagents were supplied by Epicentre (Madison, WI). Each XL PCR consisted of 1x MasterAmp Extra-Long PreMix 5, LOSXL primers (Table 2) at 0.2 μM each, 2.5 U of MasterAmp Extra-Long DNA polymerase, and 250 ng of genomic DNA (final reaction volume, 50 μl). The cycling conditions for XL PCR products were 25 cycles of 30 s at 94°C, 45 s at 52°C, and 15 min at 68°C and a final extension at 68°C for 15 min. Thermal cycling was performed with a Tetrad thermal cycler (MJ Research, Waltham, MA). All PCR products were analyzed by agarose gel electrophoresis. Positive samples were identified based on the presence of bands of anticipated sizes. Primers were purchased from either Operon Technologies (Alameda, CA) or Sigma-Genosys (The Woodlands, TX).
DNA sequencing. Cycle sequencing reactions were performed using the ABI PRISM BigDye terminator cycle sequencing kit (version 3.0). All extension products were purified on DyeEx spin columns (QIAGEN, Valencia, CA) or Centri-Sep spin columns (Princeton Separations, Princeton, NJ). DNA sequencing was performed using an ABI PRISM 3100 Genetic Analyzer (Applied Biosystems, Foster City, CA) with the POP-6 polymer and ABI PRISM Genetic Analyzer Data Collection and ABI PRISM Genetic Analyzer Sequencing Analysis software. Sequencing oligonucleotide primers were purchased from Operon Technologies (Alameda, CA).
Nucleotide sequence accession numbers. Nucleotide sequences have been deposited with GenBank and assigned accession numbers as follows: RM1170 LOS loci, AY434498; RM1555 (ATCC 42437) LOS loci, AY436358; RM1047 (ATCC 43431), AY800272.
RESULTS AND DISCUSSION
Specificity of LOS biosynthesis locus class-specific PCR primers. Based on gene content and organization of the LOS biosynthesis loci from 15 different strains of C. jejuni (3, 5, 7, 8, 10, 11, 22, 23), we grouped the loci into six major LOS classes: A, B, C, D, E, and F (Fig. 1). Classes A, B, and C, involved in sialylated LOS synthesis, were previously described (7). The sequences for classes D, E, and F were obtained from GenBank (3, 8, 22, 23) Each locus possesses a number of novel putative glycosyltransferase genes and genes with unknown function but lacks the neuBCA and cst genes and likely directs the synthesis of nonsialylated LOS structures. We evaluated 42 primer pairs (Table 2) specific to the genes within the six classes designated in Fig. 1 by PCR amplification of genomic DNA from strains representing each defined LOS class (Table 3). All C. jejuni LOS classes were amplified with the waaM (orf2) and waaV (orf12) primer sets. Strains RM1048, RM1556, RM1050, and RM1052 (classes A and B) share many amplification products, including cstII (orf7ab), neuB1-ab (orf8ab), neuC1-ab (orf9ab), neuA1-ab (orf10ab), and orf11ab. The cgtA2 (orf5bII) product, however, is specific for class B and can be used to distinguish class A and class B. Previous work (7) demonstrated sequence divergence of a region spanning the cgtA (orf5abI) and cgtB (orf6ab) genes (1,200 bp), creating two sets of alternative alleles: orf5abI1 and orf6ab1 or orf5abI2 and orf6ab2. These alternative alleles are present in both class A and class B. We designed allele-specific primers for these genes, such that we amplified orf5abI1 and orf6ab1 from strains RM1048 (ATCC 43432) (class A1) and RM1052 (ATCC 43456) (class B1) but not orf5abI2 or orf6ab2. This allowed us to designate the alleles for class A (A1 and A2) and class B (B1 and B2).
The products for Cj1137c (orf14c), Cj1138 (orf15c), Cj1139c(orf6c), cstIII (orf7c), neuB1-c (orf8c), neuC1-c (orf9c),cgtAneuA1-c (orf5c/10c), and orf16c were amplified specifically from the class C loci. Classes D and F share many amplification products, including orf18d, orf19df, orf20df, and orf16df. The orf17d product, specific for class D, can distinguish classes D and F. Only class E C. jejuni strains produced PCR products for the orf21e, 22e, 25e, 26e, 27e, 28e, 29e, 30e, 31e, 32e, 33e, and 34e genes. Surprisingly, strain 81116 (class E) produced amplification products with neuB1-ab (orf8ab) and neuC1-ab (orf9ab). The identities of these products were confirmed by sequencing and suggest the presence of these genes elsewhere in the genome.
LOS biosynthesis locus diversity of the Penner serotype reference strains. From the class-specific primer pairs, we identified 12 primer pairs representing genes from the six classes (orf6ab1, orf6ab2, orf7ab, orf8ab, orf5bII, orf6c, orf7c, orf8c, orf17d, orf18df, orf3e, and orf27e) that would allow putative class determination and distinguish between the two alleles of classes A and B. As a positive control for the presence of a C. jejuni LOS locus, we amplified the orf12 (waaV) product. Thirty-nine Penner serotype reference strains were examined. Twenty-five of the reference strains (HS:1 to HS:8, HS:10, HS:11, HS:13, HS:16, HS:18, HS:19, HS:23, HS:27, HS:29, HS:32, HS:35, HS:36, HS:41, HS:44, HS:45, HS:62, and HS:64) showed PCR-amplification patterns consistent with possessing class A1, A2, B1, B2, C, D, E, or F (Table 4). For these strains, we attempted to verify the LOS classes by additional amplifications of genes within the presumed class. Most strains had patterns consistent with their LOS class designation; however, we identified a number of strains with patterns that diverged from the expected class patterns. Most presumptive class E strains—RM3407 (HS:3), RM3411 (HS:7), RM3428 (HS:37), RM3429 (HS:38), and RM3432 (HS:45)—failed to produce two class E primer amplicons, orf26e and orf28e, while RM3423 (HS:27) failed to produce the orf28e amplicon. The sizes of the XL PCR products for the LOS loci of these strains (extending from orf2 [waaM] to orf12 [waaV]) were approximately 1 kb smaller than those of the class E strains. Similarly, the presumptive class D strain RM3418 (HS:17) and the presumptive class F strain RM3437 (HS:58) failed to produce the orf19df product and had XL PCR products that were slightly larger than those of the class D and class F strains (data not shown).
The remaining 15 Penner serotype reference strains (HS:9, HS:17, HS:22, HS:32, HS:37, HS:38, HS:43, HS:50, HS:52, HS:53, HS:57, HS:58, HS:60, HS:63, and HS:65) failed to produce PCR amplification patterns consistent with the six classes by use of the 12 primers. Some of these unclassified LOS loci possessed LOS genes from known classes. As with RM3407, these loci may represent a new class that is derived from a known class. The LOS loci of these strains are likely novel, or they possess extensive sequence divergence that prevents PCR amplification. Interestingly, strains RM3434 (HS:52) and RM3442 (HS:65) appeared to exhibit multiple LOS classes: B2 and C for RM3434 and A2 and F for RM3442. The XL PCR products for RM3434 and RM3442 are similar in size to the C and A2 XL PCR products, respectively. This suggests that these LOS loci are not likely a combination of the two classes. DNA sequencing will be required to determine the exact LOS class for each of these isolates and to address where the genes from the other LOS class are.
LOS classes of isolates from clinical and environmental sources. To investigate any associations between LOS classes and isolate sources, we screened 70 clinical and environmental isolates with the 12 primer pairs. We surveyed 16 GBS-associated isolates and found that 14 were positive for class A1 products, whereas the other 2 isolates were positive for ganglioside-associated classes A2 and C (Table 5). This is consistent with a study by Nachamkin et al. (21), who previously demonstrated a strong association of cst-II (orf7a,b), cgtA (orf5), and cgtB (orf6) with GBS isolates. It should be noted that most of the GBS-associated strains in this study (12 of 16 GBS strains) were HS:19 or HS:41, which have been shown to be genetically clonal (13, 28). This overrepresentation certainly increases the occurrence of the class A1 locus in this study. Examining additional non-HS:19 and non-HS:41 GBS-associated strains would provide a better indication of the relationship between LOS class and this syndrome. Indeed, Godschalk et al. recently examined 17 GBS-associated isolates and found that 53% possessed the class A locus (9). Interestingly, they also found that 18% of the GBS-associated strains possessed LOS loci with no known sialic acid transferase gene. These strains may possess a unique sialic acid transferase that allows the synthesis of ganglioside mimics. It is also possible that these strains coinfected with other C. jejuni strains capable of synthesizing a ganglioside mimic and are merely the strains recovered by culturing.
For other clinical isolates and for environmental isolates (Table 6), we observed strains with putative LOS locus classes A, B, C, D, E, and F, and we were also able to reclassify the LOS locus for several the putative class E isolates to class H (described below) by using the orf26e primer pair. As with the Penner serotyping reference strains, the LOS locus class of several isolates could not be determined by the PCR screen. The most striking finding was that 64% (35 of 55) of the non-GBS-associated isolates possessed an LOS locus class (A, B, or C) capable of synthesizing a ganglioside mimic. Similarly, of the 21 isolates associated with enteritis, Godschalk et al. found that 62% possessed a class A, B, or C locus (9). It is not clear why the incidence of strains capable of producing ganglioside mimics from all sources is so high, although these findings seem to suggest that these ganglioside-mimicking LOS structures are advantageous to C. jejuni for colonization or infection of various hosts. What is clear is that the production of ganglioside-mimicking LOS structures alone is not sufficient to elicit GBS; certain host factors and/or other bacterial factors are required.
Distribution of LOS classes within a group of Penner serotype isolates. Although there is strong evidence suggesting that the capsule rather than LOS is responsible for many Penner serotypes (12), it is possible that there remains a correlation between LOS genotype and Penner serotype. Indeed, all HS:19 and HS:41 strains examined share the same class A1 locus (see Tables 5 and 6). With only a limited number of strains possessing the same Penner type, we observed that Penner type and LOS class are not consistently associated. The HS:10 isolates (RM1556, RM1413, and RM1507) contained class A1 or A2, and the HS:2 isolates (RM1046, RM1188, RM1478, RM1480, and RM1862) contained class A1, B2, or C. Horizontal transfer of the LOS locus (or the capsular locus) provides a plausible mechanism for different LOS classes between strains with the same Penner serotype, as evidenced by the characterization of the class A1 LOS locus of GB11 (HS:2) (6). However, alternative mechanisms would be required to maintain the A1 LOS locus class in all the HS:19 and HS:41 strains examined in this study.
Identification of two new LOS locus classes. To verify novel LOS loci, we sequenced the XL PCR products from strains RM1555 (ATCC 43437) (GenBank accession number AY436358) and RM1047 (ATCC 43431) (GenBank accession number AY800272). The amplification pattern of RM1555 was inconsistent with the known LOS classes; it was positive for orf3e and waaV (orf12) products and negative for all other products. The sequence of the 5,537-bp region between waaM (Cj1134) and waaV (Cj1146c) establishes a seventh LOS locus class, G (Fig. 2). This locus contains four newly described open reading frames (ORFs) that encode putative glycosyltransferases based on BLASTX scores. A fifth potential ORF contains an intergenic homopolymeric G tract with 8 G bases (versus 9 G bases) that disrupts the reading frame after 336 nucleotides. By pairwise alignment, these 336 nucleotides show 93.5% identity to the first 322 nucleotides of the 888-nucleotide orf16d from RMLIO87 and 92.9% identity across the entire corresponding 888-nucleotide region. orf16f from RM1170 is similarly truncated, with 8 G bases in the homopolymeric G tract, and is 93.8% identical by pairwise alignment. The orf16df PCR product was not amplified from RM1555 due to significant sequence differences at the binding site of primer orf16df (data not shown). Interestingly, orf16d and orf16f are located adjacent to orf12d and orf12f in LIO87 and RM1170, respectively, but orf16g is separated from orf12g by orf38g, a pattern that presumably resulted from an insertion event.
The previous PCR results suggested that the LOS locus of RM1047 (ATCC 43431) differed from the class E locus but was similar to the presumptive class H locus. Indeed, the XL PCR product from RM1047 (ATCC 43431) was the same size as the class H LOS product (data not shown). Sequence analysis verified that this LOS locus represents a new class, termed class H, that is distinct from class E (Fig. 2). The orf28e gene is clearly missing from this locus, but the 3' end of the orf26e gene is still present (Fig. 2). Furthermore, these changes appear to be the results of two different events. The absence of orf26e seems to be the result of an insertion-deletion event, with the 3' end of orf25e still present and the 5' region replaced by a putative butyryltransferase gene (orf39h). The second event appears to have been a deletion event that resulted in the complete removal of orf28e from RM1047, with minor changes to the 3' end of orf27e and the 5' end of orf29e. Considering these differences, we were able to reassign many of the isolates that were initially identified as class E to class H based on the PCR results for orf26e and orf28e (Tables 4 and 6). Furthermore, we identified hypervariable homopolymeric G tracts within orf23e (9 or 10 G bases) and orf25e (8 or 9 G bases). Therefore, all classes of LOS loci examined so far appear to have this common mechanism for creating LOS variability.
Conclusions. This study reports a method to examine the diversity and classification of the LOS biosynthesis loci of C. jejuni. The PCR method identified the presence of genes from six LOS classes and was rapid, sensitive, and specific. The LOS structure of a particular isolate is determined not only by the genes of the locus but also by the presence of various mutations. Thus, strains sharing the same LOS class do not necessarily express the same LOS structure (7). Unfortunately, this method is unable to detect certain insertions, deletions, and point mutations that affect LOS structure. Nevertheless, we can infer LOS structures once an isolate's LOS class has been determined, particularly the presence or absence of ganglioside mimicry. Considering the potential role of C. jejuni ganglioside-mimicking structures in eliciting GBS or Miller Fisher syndrome, we believe that the LOS class of clinical strains provides valuable information and should be incorporated into epidemiological studies. Indeed, we found that all GBS-related strains and 64% of the other clinical and environmental isolates examined in this study belonged to an LOS class (LOS class A, B, or C) that allowed the synthesis of a sialylated LOS. The overrepresentation of HS:19 and HS:41 isolates likely biased the occurrence of LOS class A1 among GBS-associated strains. However, Godschalk et al. (9) similarly observed a significant association of LOS class A with GBS-associated isolates. Clearly, additional analysis with a more diverse group of GBS-associated strains is required to determine the significance of the LOS class.
Although not all LOS loci could be classified at this time, we demonstrated the ability to identify new loci, and we characterized the sequences of two of these, increasing the number of LOS classes to eight. For these new LOS classes, class G and class H, we observed evidence of genetic rearrangements (deletions or insertions) and homopolymeric G tracts that can lead to LOS structural variability. Specifically, the LOS synthesized by the LOS class H strain RM1047 could differ from those of other class H strains due to differences in the lengths of G tracts within biosynthesis genes. Also, LOS class H appears to be a derivative of class E, and therefore the LOS from these classes may share structural features. Additional analysis of C. jejuni LOS structures may allow potential structures to be inferred once an LOS class has been identified. Furthermore, the other unclassified LOS loci certainly substantiate the need for additional studies, and we envision the development of a microarray-based method that allows simultaneous analysis for distinguishing C. jejuni LOS classes.
ACKNOWLEDGMENTS
We thank S. Abbott, M. Englen, P. Guerry, W. Johnson, A. Lastovica, R. Meinersmann, I. Nachamkin, M. Nicholson, L. Stanker, and I. Wesley for providing strains. We also thank J. Klena for critical reading of the manuscript.
This work was supported by the United States Department of Agriculture, Agricultural Research Service CRIS project 5325-42000-041. M.G. was supported by the NRC Genome Sciences and Health Related Research Initiative.
REFERENCES
Allos, B. M. 1997. Association between Campylobacter infection and Guillain-Barre syndrome. J. Infect. Dis. 176(Suppl. 2):S125-S128.
Dorrell, N., J. A. Mangan, K. G. Laing, J. Hinds, D. Linton, H. Al-Ghusein, B. G. Barrell, J. Parkhill, N. G. Stoker, A. V. Karlyshev, P. D. Butcher, and B. W. Wren. 2001. Whole genome comparison of Campylobacter jejuni human isolates using a low-cost microarray reveals extensive genetic diversity. Genome Res. 11:1706-1715.
Fry, B. N., V. Korolik, B. A. van der Zeijst, and P. J. Coloe. 1998. A gene cluster from Campylobacter jejuni involved in inner core and lipid A synthesis. NCBI accession number AJ131360. Applied Biology and Biotechnology, Royal Melbourne Institute of Technology University, Melbourne, Australia.
Gilbert, M., J. Bellefeuille, M. F. Karwaski, A. M. Cunningham, and W. W. Wakarchuk. 2001. Partial sequence of the lipooligosaccharide biosynthesis locus of Campylobacter jejuni O:3. NCBI accession number AF411225. Institute for Biological Sciences, National Research Council of Canada, Montreal, Canada.
Gilbert, M., J. R. Brisson, M. F. Karwaski, J. Michniewicz, A. M. Cunningham, Y. Wu, N. M. Young, and W. W. Wakarchuk. 2000. Biosynthesis of ganglioside mimics in Campylobacter jejuni OH4384. Identification of the glycosyltransferase genes, enzymatic synthesis of model compounds, and characterization of nanomole amounts by 600-mHz 1H and 13C NMR analysis. J. Biol. Chem. 275:3896-3906.
Gilbert, M., P. C. Godschalk, M. F. Karwaski, C. W. Ang, A. van Belkum, J. Li, W. W. Wakarchuk, and H. P. Endtz. 2004. Evidence for acquisition of the lipooligosaccharide biosynthesis locus in Campylobacter jejuni GB11, a strain isolated from a patient with Guillain-Barre syndrome, by horizontal exchange. Infect. Immun. 72:1162-1165.
Gilbert, M., M. F. Karwaski, S. Bernatchez, N. M. Young, E. Taboada, J. Michniewicz, A. M. Cunningham, and W. W. Wakarchuk. 2002. The genetic basis for the variation in the lipo-oligosaccharide of the mucosal pathogen, Campylobacter jejuni. Biosynthesis of sialylated ganglioside mimics in the core oligosaccharide. J. Biol. Chem. 277:327-337.
Gilbert, M., R. K. Singh, M.-F. Karwaski, and W. W. Wakarchuk. 2002. Sequencing of the lipo-oligosaccharide biosynthesis locus of Campylobacter jejuni LIO87. NCBI accession number AF40069. Institute for Biological Sciences, National Research Council of Canada, Ottawa, Ontario, Canada.
Godschalk, P. C., A. P. Heikema, M. Gilbert, T. Komagamine, C. W. Ang, J. Glerum, D. Brochu, J. Li, N. Yuki, B. C. Jacobs, A. van Belkum, and H. P. Endtz. 2004. The crucial role of Campylobacter jejuni genes in anti-ganglioside antibody induction in Guillain-Barre syndrome. J. Clin. Investig. 114:1659-1665.
Guerry, P., C. P. Ewing, T. E. Hickey, M. M. Prendergast, and A. P. Moran. 2000. Sialylation of lipooligosaccharide cores affects immunogenicity and serum resistance of Campylobacter jejuni. Infect. Immun. 68:6656-6662.
Guerry, P., C. M. Szymanski, M. M. Prendergast, T. E. Hickey, C. P. Ewing, D. L. Pattarini, and A. P. Moran. 2002. Phase variation of Campylobacter jejuni 81-176 lipooligosaccharide affects ganglioside mimicry and invasiveness in vitro. Infect. Immun. 70:787-793.
Karlyshev, A. V., D. Linton, N. A. Gregson, A. J. Lastovica, and B. W. Wren. 2000. Genetic and biochemical evidence of a Campylobacter jejuni capsular polysaccharide that accounts for Penner serotype specificity. Mol. Microbiol. 35:529-541.
Lastovica, A. J., E. A. Goddard, and A. C. Argent. 1997. Guillain-Barre syndrome in South Africa associated with Campylobacter jejuni O:41 strains. J. Infect. Dis. 176(Suppl. 2):S139-S143.
Leonard, E. E., II, T. Takata, M. J. Blaser, S. Falkow, L. S. Tompkins, and E. C. Gaynor. 2003. Use of an open-reading frame-specific Campylobacter jejuni DNA microarray as a new genotyping tool for studying epidemiologically related isolates. J. Infect. Dis. 187:691-694.
Leonard, E. E., II, L. S. Tompkins, S. Falkow, and I. Nachamkin. 2004. Comparison of Campylobacter jejuni isolates implicated in Guillain-Barre syndrome and strains that cause enteritis by a DNA microarray. Infect. Immun. 72:1199-1203.
Linton, D., M. Gilbert, P. G. Hitchen, A. Dell, H. R. Morris, W. W. Wakarchuk, N. A. Gregson, and B. W. Wren. 2000. Phase variation of a -1,3 galactosyltransferase involved in generation of the ganglioside GM1-like lipo-oligosaccharide of Campylobacter jejuni. Mol. Microbiol. 37:501-514.
Linton, D., A. V. Karlyshev, P. G. Hitchen, H. R. Morris, A. Dell, N. A. Gregson, and B. W. Wren. 2000. Multiple N-acetyl neuraminic acid synthetase (neuB) genes in Campylobacter jejuni: identification and characterization of the gene involved in sialylation of lipo-oligosaccharide. Mol. Microbiol. 35:1120-1134.
McCarthy, N., Y. Andersson, V. Jormanainen, O. Gustavsson, and J. Giesecke. 1999. The risk of Guillain-Barre syndrome following infection with Campylobacter jejuni. Epidemiol. Infect. 122:15-17.
Miller, W. G., A. H. Bates, S. T. Horn, M. T. Brandl, M. R. Wachtel, and R. E. Mandrell. 2000. Detection on surfaces and in Caco-2 cells of Campylobacter jejuni cells transformed with new gfp, yfp, and cfp marker plasmids. Appl. Environ. Microbiol. 66:5426-5436.
Moran, A. P. 1997. Structure and conserved characteristics of Campylobacter jejuni lipopolysaccharides. J. Infect. Dis. 176(Suppl. 2):S115-S121.
Nachamkin, I., J. Liu, M. Li, H. Ung, A. P. Moran, M. M. Prendergast, and K. Sheikh. 2002. Campylobacter jejuni from patients with Guillain-Barre syndrome preferentially expresses a GD1a-like epitope. Infect. Immun. 70:5299-5303.
Oldfield, N. J., A. P. Moran, L. A. Millar, M. M. Prendergast, and J. M. Ketley. 2002. Characterization of the Campylobacter jejuni heptosyltransferase II gene, waaF, provides genetic evidence that extracellular polysaccharide is lipid A core independent. J. Bacteriol. 184:2100-2107.
Parker, C. T. 2004. Diversity in the lipooligosaccharide biosynthesis locus of Campylobacter jejuni. NCBI accession number AY434498. Produce Safety and Microbiology Unit, United States Department of Agriculture, Agriculture Research Service, Albany, Calif.
Parkhill, J., B. W. Wren, K. Mungall, J. M. Ketley, C. Churcher, D. Basham, T. Chillingworth, R. M. Davies, T. Feltwell, S. Holroyd, K. Jagels, A. V. Karlyshev, S. Moule, M. J. Pallen, C. W. Penn, M. A. Quail, M. A. Rajandream, K. M. Rutherford, A. H. van Vliet, S. Whitehead, and B. G. Barrell. 2000. The genome sequence of the food-borne pathogen Campylobacter jejuni reveals hypervariable sequences. Nature 403:665-668.
Pearson, B. M., C. Pin, J. Wright, K. I'Anson, T. Humphrey, and J. M. Wells. 2003. Comparative genome analysis of Campylobacter jejuni using whole genome DNA microarrays. FEBS Lett. 554:224-230.
Penner, J. L., and J. N. Hennessy. 1980. Passive hemagglutination technique for serotyping Campylobacter fetus subsp. jejuni on the basis of soluble heat-stable antigens. J. Clin. Microbiol. 12:732-737.
Penner, J. L., J. N. Hennessy, and R. V. Congi. 1983. Serotyping of Campylobacter jejuni and Campylobacter coli on the basis of thermostable antigens. Eur. J. Clin. Microbiol. 2:378-383.
Prendergast, M. M., A. J. Lastovica, and A. P. Moran. 1998. Lipopolysaccharides from Campylobacter jejuni O:41 strains associated with Guillain-Barre syndrome exhibit mimicry of GM1 ganglioside. Infect. Immun. 66:3649-3655.
van Koningsveld, R., R. Rico, I. Gerstenbluth, P. I. Schmitz, C. W. Ang, I. S. Merkies, B. C. Jacobs, Y. Halabi, H. P. Endtz, F. G. van der Meche, and P. A. van Doorn. 2001. Gastroenteritis-associated Guillain-Barre syndrome on the Caribbean island Curacao. Neurology 56:1467-1472.
Willison, H. J., and G. M. O'Hanlon. 1999. The immunopathogenesis of Miller Fisher syndrome. J. Neuroimmunol. 100:3-12.(Craig T. Parker, Sharon T)