HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Abstract Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via CrossRef Citing Articles via Google Scholar Google Scholar Articles by Schmidt-Ott, R. Articles by Groß, U. Search for Related Content PubMed PubMed Citation Articles by Schmidt-Ott, R. Articles by Groß, U. Agricola Articles by Schmidt-Ott, R. Articles by Groß, U.

Improved serodiagnosis of Campylobacter jejuni infections using recombinant antigens

Ruprecht Schmidt-Ott¹, Felicitas Brass¹, Christiane Scholz¹, Carola Werner² and Uwe Groß¹

^1,2Institute of Medical Microbiology¹ and Department for Medical Statistics², University of Göttingen, D-37075 Göttingen, Germany

Correspondence Ruprecht Schmidt-Ott ruprecht.schmidt-ott{at}gsk.com

Received February 8, 2005
Accepted April 26, 2005

Campylobacter jejuni is a frequent cause of infectious diarrhoea and is increasingly recognized as a trigger for late-onset complications. The poor standardization of commonly used serological tests might explain the conflicting results regarding the frequency of antecedent C. jejuni infections in defined patient groups. In order to obtain reliable epidemiological data as to the role of C. jejuni in causing late-onset complications, a highly specific and sensitive diagnostic tool for the epidemiological investigation of C. jejuni-associated diseases was developed. It was shown that recombinant proteins encoded by the C. jejuni genes cj0017 (P39) and cj0113 (P18) are specifically recognized by antibodies in sera from patients with C. jejuni enteritis. An ELISA using recombinant P18 and P39 as antigens was 91.9 % sensitive and 99.0 % specific, with positive and negative predictive values of 97.1 % and 97.0 %, respectively, comparing favourably with the 27.0 % sensitivity of a routinely used serological assay.

Abbreviations: CFA, complement-fixation assay; ROC, receiver operating characteristic.

	INTRODUCTION

TOP INTRODUCTION Methods RESULTS DISCUSSION ACKNOWLEDGEMENTS REFERENCES

 
Campylobacter jejuni is a very important cause of diarrhoeal disease in developing and industrialized countries (Friedman et al., 2004), ranking second in incidence after Salmonella enterica in the USA and Germany (CDC, 2003; Stark & Alpers, 2004). There is an increasing concern that C. jejuni infections might trigger a number of late-onset complications, which so far has best been documented for Guillain-Barré syndrome (Hughes & Rees, 1997) and HLA B27-associated reactive arthritis (Colmegna et al., 2004). Recent data suggest a link between C. jejuni and immunoproliferative small intestinal disease (Lecuit et al., 2004), a primary intestinal lymphoma with poor prognosis that accounts for one-third of primary small intestinal lymphomas in Mediterranean countries (Salem et al., 1986).

The aetiological link between these diseases and C. jejuni infections has mostly been established on the basis of serological data, since cultural diagnosis of C. jejuni after cessation of diarrhoea is rarely successful (Svedhem & Kaijser, 1980). Several serological methods have been used to detect C. jejuni-specific antibodies, including complement-fixation assays (CFAs), immunoblotting and ELISAs. So far, these tests have not been standardized with regard to antigens used or to stringent criteria defining seropositivity. Crude antigenic preparations are commonly applied, including acid-glycine extracts (Blaser & Duncan, 1984), heat-stable antigens (Kaldor & Speed, 1984), whole-cell sonicates (Boucquey et al., 1991), outer-membrane protein preparations (Enders et al., 1993) and phenol/water extraction of C. jejuni LPS (Jacobs et al., 1998). This lack of standardization might explain the conflicting results regarding the frequency of preceding C. jejuni infections in patients with Guillain-Barré syndrome (Hughes & Rees, 1997) or reactive arthritis (Bremell et al., 1991; Eastmond et al., 1983; Kosunen et al., 1980; Locht & Krogfelt, 2002). Moreover, the specificity of these assays might be low due to antigenic cross-reactivities, in particular with the closely related Helicobacter pylori and with other Gram-negative bacteria that are associated with clinical manifestations similar to C. jejuni, such as enteropathogenic Yersinia species (Colmegna et al., 2004). In order to obtain reliable epidemiological data as to the role of C. jejuni in causing late-onset complications, specific serological markers of past C. jejuni infections are essential.

So far only a few immunoreactive components in crude C. jejuni extracts have been identified. Although the immunodominance of flagellin (FlaA) during human infection is well established (Nachamkin & Hart, 1985; Wenman et al., 1985), the diagnostic exploitation of this antigen is hampered by immunodominant epitopes of FlaA being serotype specific (Newell & Nachamkin, 1992). Seroconversions to the major cell-adherence factor Cj0921 (Peb1) and the glycoprotein Cj0289 (Peb3) have been observed in a small number of convalescent patients (Pei et al., 1991), and the putative peptidoglycan-associated protein Cj0113 showed a reaction with sera from some patients with arthritis of suspected C. jejuni etiology (Burnens et al., 1995). The C. jejuni flagellar hook protein Cj1729c (FlgE2) and the major outer-membrane protein Cj1259 (PorA) are immunogenic in humans (Blaser et al., 1984). However, as for FlaA, the surface-exposed epitopes of FlgE2 seem to be serotype specific (Power et al., 1992) and antibodies reacting with PorA are highly prevalent in the healthy population (Blaser et al., 1984). Taken together, low sensitivity or specificity of FlaA, FlgE2 and PorA limit their use as reliable serological markers for C. jejuni infection.

We show that the proteins encoded by strain NCTC 11168 genes cj0017 (P39) and cj0113 (P18) are valuable antigens for the serological diagnosis of past C. jejuni infections. Using purified recombinant P39 and P18, we established a highly specific and sensitive serological test for C. jejuni infections, which will be a valuable tool for future epidemiological studies on late-onset complications of C. jejuni enteritis.

	Methods

TOP INTRODUCTION Methods RESULTS DISCUSSION ACKNOWLEDGEMENTS REFERENCES

 
Bacterial strains.

The C. jejuni strain NCTC 11168 was obtained from the National Collection of Type Cultures (NCTC, London, UK). Escherichia coli DH10B was purchased from Invitrogen.

For the recombinant expression of C. jejuni DNA we used Epicurian Coli BL21-CodonPlus from Stratagene adapted for the expression of AUA- and AGA-rich coding sequences, since a bias regarding these codons seems to be responsible for the reported difficulties in expressing C. jejuni DNA in E. coli (Ketley, 1997).

Sera.

Patients attending the University Hospital of Göttingen between January 2003 and December 2004 with bloody or watery diarrhoea were routinely tested for C. jejuni in their faeces. Campylobacter strains were identified at the species level using a commercial biochemical differentiation kit (API CAMPY, bioMérieux). Post-infection sera from 37 patients (age, mean ± SD, 48 ± 20 years) were collected within 4 weeks (3–24 days) of the cultural diagnosis of C. jejuni enteritis. Follow-up sera were obtained from four patients at various intervals after the onset of diarrhoea. Control sera were sampled from 57 healthy blood donors (age, mean ± SD, 47 ± 12 years), from 19 patients with arthritis and serological evidence for a recent Yersinia enterocolitica infection (Y. enterocolitica IgG/IgA ELISA, Virotech) (age, mean ± SD, 38 ± 22 years), and from 22 patients who were tested positive for IgG in a H. pylori-specific immunoblot (Karvar et al., 1997) (age, mean ± SD, 27 ± 20 years). All sera were stored at –70 °C until used.

Generation of a genomic expression library.

Genomic DNA of the C. jejuni reference strain NCTC 11168 was prepared using Qiagen Genomic-tip 20/G (Qiagen) according to the manufacturer's instructions and partially digested with Sau3A (New England Biolabs). DNA fragments sized 1–2 kb were gel purified with the MinElute Gel Extraction Kit from Qiagen. The DNA fragments were cloned in-frame into a BamHI restriction site of the pET3a, b or c expression vectors (Novagene) under the control of the inducible T7 promoter. The resulting three genomic pET3 libraries were electroporated into E. coli DH10B. Transformants were selected on LB agar containing ampicillin (50 µg ml^–1) and chloramphenicol (30 µg ml^–1). Approximately 10 000–15 000 transformants were pooled and plasmid DNA was extracted with the Miniprep Kit from Qiagen. Plasmid DNA (100 ng) was subsequently transformed into electrocompetent Epicurian Coli BL21-CodonPlus. Transformants were grown overnight on LB agar containing ampicillin (50 µg ml^–1) and chloramphenicol (30 µg ml^–1). Ten thousand colonies were pooled in LB broth and aliquots were kept at –80 °C until they were used for immunoscreening.

Colony immunoscreen of the genomic C. jejuni expression library.

The expression library was plated on nitrocellulose membranes (ó 82 mm) (Optitran BA-S 85, Schleicher & Schuell) at a density of 500 c.f.u. per membrane, and grown overnight at 37 °C on LB agar containing ampicillin (50 µg ml^–1) and chloramphenicol (30 µg ml^–1) (master membrane). Replica membranes were constructed as previously described (Sambrook et al., 1989). Protein synthesis was induced by exposing the master membrane to 1 mM IPTG for 3 h at 37 °C. Bacteria were lysed in situ in TBST, lysozyme (40 µg ml^–1) and DNase I (1 µg ml^–1) for 2 h at room temperature. After blocking the membranes for 60 min with 5 % (w/v) skimmed milk in TBST (Tris-buffered saline with 0.05 % v/v Tween 20; blocking solution), a colony immunoscreen was performed, as described previously (Sambrook et al., 1989), using pooled convalescent phase sera from five patients with culture-confirmed C. jejuni enteritis [diluted 1 : 200 in TBST and 1 % (w/v) skimmed milk] and an alkaline phosphatase-conjugated rabbit anti-human-IgG secondary antibody (Dianova) diluted 1 : 5000 in TBST. The immunoreactivity of selected library clones was confirmed by immunoblot analysis, according to the instruction manual of the pET system (Novagen). Briefly, SDS-PAGE of induced bacterial culture was performed under denaturing conditions according to the Laemmli method (Sambrook et al., 1989), followed by immunoblotting on Optitran BA-S 85 nitrocellulose (Schleicher & Schuell) with patient sera as described above.

Identification, recombinant expression and purification of immunogenic C. jejuni proteins.

The plasmids of immunoreactive library clones were purified (Miniprep Kit, Qiagen) and the pET3 vector was used as a template for DNA sequencing reactions with the T7-promoter primer. The resulting sequences were blasted against the published C. jejuni NCTC 11168 genome sequence (www.sanger.ac.uk/Projects/C_jejuni/), revealing the coding DNA for the immunoreactive proteins.

For the recombinant expression of the immunoreactive proteins, the genes cj0017, cj0113 and cj1339 were PCR amplified from NCTC 11168 genomic DNA using the KOD Hot Start DNA polymerase (Novagene) and the primer pairs FB7/FB8 (5'-GGGATCCGCCTGTAAGATTT AGTTTAAA-3'/5'-CGGGATCCGTTAGTTTAAAGTATAAAGCTTG-3'), FB1/FB2 (5'-CCGCTCGAGGTTATTAGTGGTTGTAGCA-3'/ 5'-CCGCTCGAGTTATCTTGATAATTTAAATTC-3') and FB3/FB4 (5'-CGGGATCCGGGATTTCGTATTAACACCA-3'/5'-CGGGATCC CTACTGTAGTAATCTTAAAAC-3'), respectively. The PCR was performed in a volume of 50 µl 1 x PCR buffer for KOD polymerase (20 ng of C. jejuni NCTC 11168 genomic DNA; 200 µM each of dATP, dCTP, dGTP and dTTP; 0.3 µM each primer; 1 mM MgSO₄; 1 U of KOD polymerase). PCR conditions were as follows: an initial melting temperature of 95 °C for 2 min; 35 cycles of 95 °C for 1 min, 50 °C for 1 min and 72 °C for 2 min; final extension at 72 °C for 10 min. The resulting PCR products were cloned in-frame in the pET15b expression vector (Novagen). After transformation into Epicurian Coli BL21-CodonPlus, expression of the 6 x His fusion proteins was induced with 1 mM IPTG following the pET System manual (Novagen). The 6 x His fusion proteins were purified by Ni-NTA affinity chromatography (Ni-NTA agarose, Qiagen) under denaturing conditions according to the manufacturer's instructions. The immunoreactivity of the purified fusion proteins was confirmed in an immunoblot with post-infection serum samples as described above.

Recombinant immunoblot and ELISA.

SDS-PAGE of affinity-purified recombinant P18 and P39 was performed under denaturing conditions. Subsequently, the proteins were blotted onto Optitran BA-S 85 nitrocellulose (Schleicher & Schuell) and incubated with sera diluted 200-fold in TBST. Bound antibodies were detected using alkaline phosphatase-conjugated rabbit human-IgG- or IgA-specific secondary antibody (Dianova) diluted 1 : 5000 in TBST.

ELISA microtitre plates (Greiner) were coated with recombinant affinity-purified P39 and P18. The proteins were pooled in coating buffer (0.015 M Na₂CO₃, 0.035 M NaHCO₃, pH 8.4) to a final concentration of 0.3 µg ml^–1 of each. Each well of the ELISA microtitre plates was coated with 100 µl antigen solution (4 °C, overnight). After blocking with 150 µl ELISA blocking reagent (Boehringer) per well for 30 min, 100 µl diluted patient serum [1 : 100 in PBS-0.05 % v/v Tween 20 (PBST)] was added to each well for 2 h at 37 °C. Subsequently, 50 µl horseradish peroxidase (HRP)-conjugated secondary antibody was added to each well and the plates were incubated for 1 h at 37 °C. The conjugates were diluted in PBST as follows: anti-human-IgG–HRP (DAKO), 1 : 6000; anti-human-IgA–HRP (DAKO), 1 : 4000. Five washing steps with washing buffer (0.35 M NaCl, 0.05 M Tris/HCl, pH 7.4, 0.1 % v/v Tween 20) were performed between each step of the ELISA. Substrate was prepared with 14 mg 1,2-phenylenediamine dihydrochloride (DAKO), 12 ml H₂O and 5 µl 30 % H₂O₂. One hundred microlitres of the substrate was added to each well and the plate was incubated for 15 min. The reaction was stopped with 100 µl 0.25 M H₂SO₄, and the optical density (OD) was measured at a wavelength of 490 nm (reference filter, 620 nm). All serum samples were tested in duplicate.

Complement-fixation assay.

For the C. jejuni-specific CFA a commercial kit (Virion Serion) was used according to the instructions of the manufacturer. As recommended by the manufacturer, sera were tested at a dilution of 1 : 30.

Statistical analysis.

To define the diagnostic value of the P18/P39 ELISA, receiver operating characteristic (ROC) curve analyses were performed (Zhou et al., 2002). For the estimation of the accuracy of the P18/P39 ELISA, i.e. the ability to discriminate between individuals with and without disease, the areas under the ROC curves were calculated for IgA- and IgG-antibodies. The accuracy of a test tends to 1 as the ROC curve tends to perfect distinction. The areas under the ROC curves for the IgA- and IgG-specific P18/P39 ELISA were compared with an extension of the Wilcoxon-Mann-Whitney test (Kaufmann et al., 2005) to see if one of the assays has a significantly higher accuracy, i.e. is more appropriate for detecting a preceding infection. Since our intention was to optimize the assay for both sensitivity and specificity, the cut-off values were determined according to the Youden index (Zhou et al., 2002), which maximizes the sum of both.

	RESULTS

TOP INTRODUCTION Methods RESULTS DISCUSSION ACKNOWLEDGEMENTS REFERENCES

 
Identification of immunoreactive C. jejuni proteins

Fourteen out of 21 000 library clones tested expressed a C. jejuni protein that reacted with post-infection sera and were therefore retained for further analysis. Sequencing of the corresponding DNA identified the genes cj0017, cj0113 and cj1339 in four, four and six of these clones, respectively.

The gene cj0017 encodes a putative protein of unknown function, designated P39. The hydropathy profile of the protein sequence reveals several hydrophobic domains in the N-terminal part of the protein, suggesting a membraneous location, which is supported by the analysis with the Transmembrane Hidden Markov model (www.cbs.dtu.dk/services/TMHMM-2.0), predicting the presence of five sequential transmembrane protein domains. The two cysteines in the first periplasmic domain are in a Cys-X-Y-Cys configuration, which is characteristic of the active site of proteins involved in disulphide bond formation (Bardwell et al., 1991). The scanning of the Prosite database (www. expasy.ch/tools/scanprosite/) reveals a putative ATP/GTP binding site (P-loop) within the hydrophilic C-terminal part of the cj0017-encoded protein. Homologues to the cj0017-encoded protein are only found in closely related H. pylori (54 % similarity, 37 % identity) and in Corynebacterium diphtheriae (45 % similarity, 28 % identity), which are both pathogenic to humans. Wolinella succinogenes, a non-pathogenic member of the Campylobacterales, has no homologue to cj0017.

The immunogenic protein P18, which is encoded by cj0113, is homologous to the peptidoglycan-associated protein (Pal) of E. coli (54 % similarity, 36 % identity). Pal-related proteins have been described in many bacterial species (Engleberg et al., 1991; Hemila, 1991; Lazzaroni & Portalier, 1992; Nelson et al., 1988), in which they generally play an important role in the host immune response. In Haemophilus influenzae, the Pal homologue might be an important antigen for the induction of protective immunity (Murphy et al., 1992). The gene cj1339 encodes the structural protein FlaA of flagella. The proteins encoded by the genes cj0113 and cj1339 have been shown to be immunogenic by others also (Burnens et al., 1995; Nachamkin & Hart, 1985), thus confirming the validity of our systematic approach.

P18/P39 immunoblot

P18, P39 and P62 were fused to 6 x His tags at their N-terminus, omitting N-terminal signal sequences and hydrophobic domains to facilitate subsequent expression and purification of the fusion protein. The fusion proteins P18, P39 and P62, corresponding to the genes cj0113, cj0017 and cj1339, respectively, were recombinantly expressed at high levels in Epicurian Coli BL21-CodonPlus (Fig. 1) and were immunoreactive with post-infection serum samples. Affinity purification of 6 x His-tagged P18 and P39 was successful, whereas the preparation of 6 x His-tagged P62 appeared to be impure (Fig. 1). Prevalence of P18- and P39-specific antibodies in post-infection sera (n = 27) and in control sera from healthy blood donors (n = 26) and patients with positive Yersinia serology (n = 22) was tested in an immunoblot. IgA and IgG antibodies in post-infection sera reacted more strongly and significantly more often with P18 and P39 than those in control sera (Fig. 2, Table 1). Combining the immunoblot data of P18 and P39 for both antibody classes resulted in a sensitivity of 88.8 % and a specificity of 72.9 %, with positive- and negative-predictive values of 64.9 % and 92.1 %, respectively.

View larger version (22K): [in this window] [in a new window]   Fig. 1. Coomassie-stained SDS-PAGE of purified recombinant fusion proteins.

View larger version (11K): [in this window] [in a new window]   Fig. 2. IgG-specific immunoblot of P18 and P39 with convalescent-phase sera from patients with C. jejuni enteritis and healthy blood donors. Pooled convalescent-phase serum from five patients with culture-confirmed C. jejuni enteritis was used as positive control (K+). Alkaline phosphatase-conjugated rabbit anti-human-IgG was used as secondary antibody (see Methods).

P18/P39 ELISA and CFA

On the basis of the immunoblot data, we developed a P18/P39 ELISA. Fig. 3 shows the results of the ELISA with sera from healthy blood donors, patients with positive Y. enterocolitica serology, patients with positive H. pylori serology and post-infection sera from patients with C. jejuni enteritis.

View larger version (14K): [in this window] [in a new window]   Fig. 3. Scattergrams showing (a) IgG- and (b) IgA-specific OD490 values measured in sera with the P18/P39 ELISA. The cut-off values indicated by the dotted lines were determined according to the Youden index (Zhou et al., 2002).

Considering all controls, the accuracy of the IgA-specific ELISA (0.981) was significantly higher (P < 0.002) than the accuracy of the IgG-specific ELISA (0.855), which indicates that the IgA-specific ELISA is more appropriate for discriminating between diseased and non-diseased individuals. The cut-off OD₄₉₀ values defining positivity were determined according to the Youden index and were 0.614 for IgG and 0.444 for IgA. The results of the serological tests are combined in Table 2. Specificity and sensitivity were highest in the IgA-specific ELISA (99 % and 92 %, respectively). Twelve post-infection sera were negative in the IgG-specific ELISA, 10 of which were positive in the IgA-specific assay. Follow-up sera (see below) from two out of four patients analysed indicated that titres of IgA antibodies peaked before IgG antibodies. Most of the IgG-negative sera were sampled in the acute phase and we assume therefore that IgG antibodies had not yet reached detectable levels. The maximum sensitivity of the P18/P39 ELISA (94.6 %) for our study group is achieved by combining the data for both antibody classes (Table 2).

Ten post-infection sera (27 %) were positive in the standard CFA (titre 1 : 30), with titres above 1 : 160 in five of these. All of the CFA-positive sera were positive in both the IgA- and IgG-specific P18/P39 ELISA. However, CFA titres did not correlate with the OD₄₉₀ values measured in the ELISA. None of the control sera was positive in the CFA.

Time-course of the P18/P39 serological response

Sequential sera of four patients with C. jejuni enteritis were tested with the P18/P39 ELISA. In all patients, specific IgA antibodies (titres 1 : 128) and IgG antibodies (titres = 1 : 2048) were detected. In all patients the IgA antibody titres dropped below the cut-off within 4 weeks after cultural diagnosis, whereas IgG antibodies remained positive (titres 1 : 256) for at least 3 months (Fig. 4).

View larger version (22K): [in this window] [in a new window]   Fig. 4. Antibody titres measured in the P18/P39 ELISA in follow-up sera from four patients with culture-confirmed C. jejuni enteritis. The sera were sampled at various intervals after the onset of diarrhoea.

	DISCUSSION

TOP INTRODUCTION Methods RESULTS DISCUSSION ACKNOWLEDGEMENTS REFERENCES

 
C. jejuni is a major cause of bacterial diarrhoeal disease (CDC, 2003; Stark & Alpers, 2004) and there is increasing concern that C. jejuni is involved in the triggering of numerous other diseases, such as Guillain-Barré syndrome (Hughes & Rees, 1997), reactive arthritis (Kosunen et al., 1980) and malignant disease (Lecuit et al., 2004). The reported frequency of antecedent C. jejuni infections in these diseases varies (Bremell et al., 1991; Colmegna et al., 2004; Eastmond et al., 1983; Kosunen et al., 1980; Locht & Krogfelt, 2002), which might reflect not only geographical differences but also the lack of standardized specific assays for C. jejuni serodiagnosis. Moreover, the sensitivities and specificities of the various serological assays can be contested since the antigen mixtures used for antibody detection include C. jejuni antigens that have been shown either to be strain-specific (Newell & Nachamkin, 1992; Power et al., 1992) or to lack specificity (Blaser et al., 1984).

Therefore, the goal of this study was to develop a reliable diagnostic tool for future serological studies dealing with C. jejuni as a potential cause of late-onset diseases. Our first step was to identify and purify conserved antigens that could be used as serological markers of C. jejuni infections. In a systematic screen of our C. jejuni genomic expression library with post-infection sera from patients with C. jejuni enteritis, we identified three immunogenic proteins encoded by the genes cj0017, cj0113 and cj1339. We succeeded in purifying the recombinant proteins P39 and P18, encoded by cj0017 and cj0113, respectively. These recombinant proteins were both specifically recognized by antibodies in post-infection sera. The high seroconversion rate to P18 and P39 in patients presumably infected with different C. jejuni strains indicates that the immunoreactive epitopes of these proteins are conserved in C. jejuni and not restricted to the infecting strain. Conservation of P18 in C. jejuni strains of different geographical origins has been shown previously (Burnens et al., 1995; Pawelec et al., 2000). Taken together, the results suggest that the recombinant proteins P39 and P18 are suitable candidates as antigens for serological studies.

Our immunoblot analysis indicated that the combined use of P18 and P39 increases the sensitivity compared to each protein used alone. Consequently, we combined both proteins as antigens in the ELISA for our serological investigations. The detection of specific IgA antibodies was most discriminatory for a recent C. jejuni infection, reaching a 100 % specificity with regard to healthy blood donors and patients with serological evidence of a Y. enterocolitica infection and 96 % with regard to H. pylori-infected controls. Taking together the results for all patient sera and controls, the specificity of the IgA-specific ELISA was 99.0 %, the sensitivity was 91.9 % and the positive- and negative-predictive values were 97.1 % and 97.0 %, respectively. Detection of C. jejuni-specific IgG antibodies was also highly specific (92.9 %) for a previous C. jejuni infection, but the sensitivity (67.6 %) was lower than for IgA, which might be due to the very early sampling of some patient sera. The time-course of the P18/P39-specific antibody response in single patients showed that, in most patients, the specific IgA antibody titre fell below the assay cut-off within 14 days after onset of diarrhoea, whereas IgG antibody titres persisted for at least 3 months. The combined data of both antibody classes maximizes the sensitivity of the assay. In addition, the IgA-specific testing is useful for narrowing down the time-point of infection. To assign the role of C. jejuni in causing late-onset complications, we suggest the combination of data for both antibody classes.

Y. enterocolitica and C. jejuni are both aetiological agents of reactive arthritis and it is therefore noteworthy that the P18/P39 ELISA did not cross-react with sera from patients with serological evidence of a preceding Yersinia infection. The specificity of the assay is further confirmed by the negative test results of sera from patients chronically infected with H. pylori, a species closely related to C. jejuni. The discrimination between C. jejuni and H. pylori infections is particularly relevant considering the high prevalence of chronic H. pylori infection in the general population (Dunn et al., 1997).

In conclusion, we show that the use of selected recombinant C. jejuni proteins improves the serodiagnosis of C. jejuni infection. We have established a recombinant ELISA with excellent specificity and sensitivity for a preceding C. jejuni infection. This assay will enable reliable epidemiological data as to the role of C. jejuni in causing late-onset complications to be generated.

	ACKNOWLEDGEMENTS

TOP INTRODUCTION Methods RESULTS DISCUSSION ACKNOWLEDGEMENTS REFERENCES

 
We thank Professor Dr M. Köhler from the department of Transfusion Medicine of the University of Göttingen for providing sera from healthy blood donors and Friederike Fisher and Regina Schmidt-Ott for the critical reading of the manuscript. The study was approved by the ethics committee of the University of Göttingen (No. 5/9/03).

	REFERENCES

TOP INTRODUCTION Methods RESULTS DISCUSSION ACKNOWLEDGEMENTS REFERENCES