Application of Recombinant FimbrialProtein for the Specific Detection ofSalmonella enteritidis Infection inPoultry

Gireesh Rajashekara, Shirin Munir,Chinta M. Lamichhane, Alberto Back,Vivek Kapur, David A. Halvorson, andKakambi V. Nagaraja

A number of disease outbreaks of Salmonella enterica sero-type enteritidis (SE) in humans have been traced to the con-sumption of SE-contaminated egg and egg products. A rapid,specific, and inexpensive method of detecting SE infection inpoultry is necessary to reduce human outbreaks. We evaluatedrSEF14 fimbrial antigen of SE for specific detection of SE-infected birds in latex agglutination test and enzyme-linkedimmunosorbent assay. rSEF14 antigen was highly specific inidentifying birds infected with SE. The sera from birds infectedwith closely related serogroup-D Salmonella and other avian

pathogens did not react with rSEF14 antigen. The rSEF14antigen identified antibodies in serum of 88% of birds duringthe first 2 weeks of infection, and 100% of the birds subse-quently. The SE-specific antibodies were detected in egg yolkas early as 6 days post-infection in rSEF14-enzyme-linked im-munosorbent assay. Our results suggest that rSEF14-basedassays could be used as screening tests for detection of SE an-tibodies and would overcome the cross reactions observed withexisting serological tests. © 1998 Elsevier Science Inc.

INTRODUCTION

Salmonellosis is the most common foodborne infec-tious disease of humans reported in the USA. In1996, among 39,072 culture-confirmed cases of sal-monellosis reported, Salmonella enterica serotype en-teritidis (SE) was the number one cause of humansalmonellosis (Angula 1997). The incidence of casesof SE in humans in the USA began to rise in the late1970s, and in 1990 it became the leading serotyperesponsible for food poisoning in the USA (Rodrigueet al. 1990). A number of disease outbreaks from SE

infection in humans have been traced to the con-sumption of SE-contaminated eggs (Hedberg et al.1993; St. Louis et al. 1988). In the USA, 67 billion eggswere produced in 1996 and to the producers of theseeggs, SE is an important economic as well as publichealth issue. It is estimated that 0.01% of all shelleggs contain SE, and the percentage may be evenhigher in some parts of the USA (Mason and Ebel1992). The presence of SE in chickens and its trans-mission through table eggs into human food chainhighlights the need for specific test to identify SE-infected chickens and to eliminate them from theflock.

Preharvest food safety is the most important com-ponent of an effective prevention and control strat-egy for SE infection. Many control strategies havebeen implemented to reduce the prevalence of SEinfection in poultry with the ultimate aim of reduc-ing SE outbreaks in humans. Successful implemen-

From the Department of Veterinary PathoBiology (GR, SM,AB, VK, DAH, KVN), University of Minnesota, St. Paul, MN55108 and Kirkegaard and Perry Laboratories (CML), Gaithers-burg, MD 20879.

Address reprint requests to Dr. Kakambi V. Nagaraja, Depart-ment of Veterinary PathoBiology, 1971 Commonwealth Avenue,University of Minnesota, St. Paul, MN 55108.

Received 11 May 1998; revised and accepted 21 July 1998.

DIAGN MICROBIOL INFECT DIS 1998;32:147–157© 1998 Elsevier Science Inc. All rights reserved. 0732-8893/98/$19.00655 Avenue of the Americas, New York, NY 10010 PII S0732-8893(98)00091-1

tation of control measures requires rapid, specific,and inexpensive methods of detection of birds in-fected with SE. Current methods for the detection ofSE infection in chickens rely primarily on conven-tional bacteriologic culture. This procedure is rela-tively slow, often taking up to 3 to 4 days to provideeven a presumptive diagnosis. More importantly,birds infected with Salmonella may not excrete theorganism everyday and with this intermittent shed-ding, traditional bacteriologic culture methods gen-erally have low sensitivity.

The detection of antibodies specific to the bacte-rium can provide an alternate method for assessingthe presence of infection. Several serological meth-ods such as serum plate agglutination (Chart et al.1990a, 1990b), microagglutination (Williams andWhittemore 1971), microantiglobulin tests (Cooper etal. 1989), ELISA (Kim et al. 1991; Timoney et al. 1990;Van Zijderveld et al. 1992) have previously beenused for the detection of SE infection in poultry. Themajor problem with these assays is that they lackeither the sensitivity or the specificity to detect SE-infected birds, or the tests are too difficult to performin a routine laboratory or at field setting (Barrow1994). This has precluded the widespread applica-tion of these tests for the detection of SE infection.

Egg yolk has been recognized as an excellentsource of antibodies (Jensenius et al. 1981; Polson etal. 1980). Only a small proportion of eggs from in-fected flocks may be positive for Salmonella on cul-ture and the examination of large number of eggs forevidence of infection by culture is impractical. Alter-natively, detection of antibodies in eggs from SE-infected birds does not involve physical handling ofbirds and is economical.

SE produces at least four distinct fimbriae: SEF14,SEF17, SEF18, and SEF21 (Clouthier et al. 1994; Col-linson et al. 1991; Muller et al. 1991; Thorns et al.1990). The SEF14 fimbriae is a uniquely specificstructure consists of repeating subunits with a mo-lecular mass of 14,300 Da (Thorns et al. 1990). Thegene sefA, encoding SEF14 has been shown to havelimited distribution among Salmonella serotypes be-longing to serogroup-D. SEF14 fimbriae are presentin S. enteritidis, Salmonella dublin, Salmonella moscow,and Salmonella blegdam (Thorns et al. 1992). Though,many of the serogroup-D Salmonella such as Salmo-nella typhi, Salmonella gallinarum, and Salmonella pul-lorum possess the intact gene, they fail to expressSEF14 fimbriae (Turcotte and Woodward 1993).

Here we describe the in vitro expression of SEF14fimbrial antigen and the development of specific se-rological assays using the recombinant SEF14 fim-brial antigen for the detection of SE-specific antibod-ies in serum and egg yolk in birds exposed to SE.

MATERIALS AND METHODS

Isolation of SE Genomic DNA

S. enteritidis UMN4 was grown overnight at 37°C inLuria-Bertani (LB) broth and the genomic DNA wasextracted as described (Koeuth et al. 1995).

Oligonucleotide Primer Selection andSynthesis

Oligonucleotide primers corresponding to an inter-nal fragment (64–498 bp) of the sefA gene were usedfor PCR amplification. Additional bases were addedto the 59 end of each primer to confer a recognitionsequence for either EcoRI (forward primer) or XhoI(reverse primer). The oligonucleotide primers wereobtained from Integrated DNA technologies Inc.(Ames, IA). The DNA sequences for the forward andreverse primers are 59 GGGAATTCGCTGGCTTT-GTTGGTAACA and 59 GGGCTCGAGTTAGTTTT-GATACTGAACGTA. Additional nucleotides addedto the 59 end of the primers are underlined.

PCR Amplification and Cloning of sefA GeneFragment

Amplification reactions were performed in 30 mLvolumes with 30 pmol of each primer and 5 mMMgCl2. The reagents and enzymes used for PCRwere obtained from Perkin Elmer (Foster City, CA).SE genomic DNA (100 ng) was used as a template forPCR amplification with the following parameters: aninitial denaturation at 94°C for 5 min, followed by 35cycles of denaturation (94°C for 1.5 min), annealing(52°C for 1 min), extension (72°C for 2 min), and afinal extension (72°C for 15 min). Amplification re-actions were performed in a DNA thermal cycler(Perkin-Elmer Cetus. Model 480). The PCR productswere analyzed on 1% agarose gel, stained withethidium bromide (0.5 mg/mL), and photographedunder UV light. PCR products were gel extracted(Qiagen Inc., Chatsworth, CA), quantitated spectro-phometrically at 260 nm, and cloned into pGEM-Tvector (Promega, Madison, WI). After ligation, 2 mLof the reaction products were transformed into Esch-erichia coli DH5a cells (Gibco BRL, Gaithersburg,MD) by heat shock method. Recombinant colonieswere selected on ampicillin/IPTG-Xgal-containingplates and screened for the presence of the appropri-ate insert by restriction analysis.

Nucleotide Sequence Analysis

A bacterial colony containing the recombinant plas-mid with the sefA gene fragment was grown in LB-ampicillin media, and the plasmid was extracted us-

148 Rajashekara et al.

ing plasmid extraction kit (Qiagen). The nucleotidesequence of the insert was determined using oligo-nucleotide primers specific to the vector sequence byautomated DNA sequencing (Advanced GeneticAnalysis Center, University of Minnesota St. Paul,MN). The sefA gene insert was sequenced in its en-tirety in both orientations, and the amino acid se-quence deduced using the standard genetic code(DNA*, Madison, WI).

SefA Gene Fragment Expression

The pGEM-T plasmid carrying sefA fragment wasdigested with EcoRI and XhoI, and the digested prod-ucts were gel purified (Qiagen) and cloned into EcoRIand XhoI digested pET/ABC expression vectors (No-vagen Inc., Madison, WI). Ligation products (2 mL)from each of the reactions were transformed into E.coli BL21(DE3) or E. coli BL21PlysS cells by heatshock method. The transformed mixture was platedon kanamycin and/or chloramphenicol containingplates, and the recombinant clones were selectedbased on restriction enzyme analysis and analyzedfor sefA expression as described by the manufacturer(Novagen).

SDS-PAGE and Western Blot Analysis

The E. coli cells expressing rSEF14 were lysed and thelysates were analyzed by SDS-polyacrylamide gelelectrophoresis (PAGE) for the presence of therSEF14 by mixing with an equal volume of 2x SDSsolubilization buffer separating on 12% polyacryl-amide gel, and staining with Coomassie blue. Forwestern blot the proteins from polyacrylamide gel(12%) were transferred onto a nitrocellulose mem-brane using Transblot apparatus (Bio-Rad laborato-ries, Hercules, CA). After transfer, the membranewas blocked with 3% bovine serum albumin (BSA) inphosphate-buffered saline (PBS) and stained witheither T7 tag antibody (Novagen) or rabbit SEF14-specific antibody (kindly provided by Dr. W. W.Kay, University of Victoria, BC, Canada). The mem-brane was washed and stained with anti-rabbit IgG/horse radish peroxidase (HRP) conjugate and treatedwith developing reagent (Amersham life sciencesInc., USA) for 1 min, exposed to X-ray film, and theradiograph developed.

Purification of rSEF14 Protein by ColumnChromatography and Electroelution

The recombinant SEF14 protein was purified usingnickel columns as described by the manufacturer(Novagen). The protein was quantitated using a Bio-Rad protein assay kit (Bio-Rad), and analyzed bySDS-PAGE. Because the column purified recombi-

nant protein contained traces of non-specific pro-teins, the rSEF14 protein was further purified byelectroelution (Bio-Rad).

Development of rSEF14 Protein-based LatexAgglutination Test (rSEF14-LAT)

The electroeluted rSEF14 protein was coupled to0.5-mm blue-dyed latex beads by gluteraldehydemethod (Polysciences Inc., Warrington, PA). TherSEF14-coated latex beads were examined as testantigens in a slide agglutination test with knownpositive and negative SE sera from chickens. A totalvolume of 7.5 mL of rSEF14-coated latex beads weremixed with an equal volume of serum. The presenceor absence of agglutination was recorded after 1 minof observation.

Development of rSEF14 Protein-basedEnzyme-linked Immunosorbent Assay(rSEF14-ELISA)

ELISA was developed following the standard proce-dures (Bullock and Walls 1977; Kim et al. 1991) withslight modifications. The 96-well microtiter plateswere coated with 100 mL of 1:200 diluted rSEF14antigen (6.6 mg/mL) using carbonate/bicarbonatebuffer overnight at 4°C. The coated plates wereblocked and stabilized using KPL blocking buffer(KPL, Gaithersburg, MD). The plates were exposedto test serum (1:100 dilution in KPL proflock dilutionbuffer) for 30 min at room temperature (RT), theplates were washed with KPL washing buffer (3min 3 3) and were incubated with 100 mL ofBiotinylated-goat anti chicken IgG (1:100 dilution)for 30 min at RT. After incubation the plates werewashed with KPL washing buffer (3 min 3 3) andexposed to 100 mL of streptavidin-HRP (1:100 dilu-tion) for 30 min at RT. The plates were washed withKPL washing buffer (3 min 3 3) and developed withKPL TMB substrate for 15 min at RT. The reactionwas stopped with KPL TMB stop solution and theoptical density was measured at 630 nm in a microELISA reader. The data were analyzed using KPLproflock software. The ELISA for yolk samples wasconducted with the same protocol as above, exceptthe purified yolk samples were used at 1:20 dilution.

Specificity of rSEF14 Antigen for Detectionof SE Antibodies in LAT and ELISA

Five 4-week-old specific pathogen free (SPF) whiteleghorn chickens (Hy-vac, Adel, IA) were experi-mentally exposed to one of the following Salmonellaserotypes: S. enteritidis, S. pullorum, S. typhimurium, S.gallinarum, S. berta, and S. dublin. An overnight broth

S. enteritidis Immunodiagnosis 149

culture of Salmonella was centrifuged, the bacterialcell pellet was resuspended in PBS and inactivedwith 0.3% formalin. Birds were inoculated intramus-cularly with formalin-killed Salmonella cultures atweekly intervals for up to 5 weeks. The sera frombirds exposed to Salmonella cultures were examinedin LAT and ELISA using rSEF14 antigen. In addition,rSEF14 antigen was also tested in ELISA on a panelof antisera (Table 1) against various avian pathogensobtained from National Veterinary Services Labora-tories (NVSL, Ames, IA).

Sensitivity of rSEF14 Antigen for Detectionof SE Antibodies in LAT and ELISA

The ability of rSEF14 antigen in LAT and ELISA todetect the birds exposed to two different doses of SEcells was examined. Briefly, 55-week-old white leg-horn layer chickens were divided into three groups(A, B, and C) of 10 birds each. These birds weretested negative for SE by culture and had no anti-bodies to SE when tested with serum plate and mi-croagglutination tests. The birds in Groups A and Bwere given orally SE culture of 104 cfu and 1010 cfuper bird, respectively. Birds in Group C were given 1mL PBS. The serum samples were collected at

weekly intervals from birds in all groups for up to 7weeks, at which point the experiment was termi-nated. Eggs from all birds were collected daily for 49days. In addition, cloacal swabs were taken atweekly intervals for 7 weeks from all the birds andcultured for SE to monitor fecal shedding of SE. Theserum samples and eggs were analyzed for the pres-ence of SE-specific antibodies using LAT and ELISA.The samples were also analyzed by serum plate test(SPT) using S. pullorum whole-cell antigen and mi-croagglutination test (MT) using SE whole-cell anti-gen for comparison.

Extraction of Egg Yolk Antibodies

Antibodies from egg yolk were extracted using poly-ethylene glycol (PEG) precipitation method (Polsonet al. 1980). Eggs collected every third day were usedfor antibody extraction. Briefly, eggs were allowed towarm at RT, yolk was aseptically collected and trans-ferred into a clean, sterile 50-mL flask, and mixed.Ten mL of yolk from each egg was used for extrac-tion of antibodies. Yolk was stirred slowly on a mag-netic stirrer and mixed with 3 volumes of solution A(4.67% PEG8000, 10mMK2HPO4, pH 7.4, and 0.1MNaCl) for 5 min. The precipitate was centrifuged at9,0003 g for 15 min at 4°C, the supernatant wastransferred to another flask by filtering through fourlayers of gauze. The sample was precipitated againwith solution B (36% PEG8000, 10 mM K2HPO4, pH7.4, and 0.1M NaCl) for 5 min by gentle stirring. Thefinal precipitate was centrifuged at 9,0003 g for 20min at 4°C and the supernatant was discarded, thepellet containing antibody was suspended in storagebuffer (0.01M K2HPO4, pH 7.4, 0.1M NaCl, 50mg/mL gentamicin sulfate), and stored at 4°C untilfurther use.

RESULTS

Expression of S. enteritidis SEF14 FimbrialProtein

Using DNA sequences corresponding to sefA geneencoding mature SEF14 fimbrial protein was clonedand expressed in pET expression systems. The SEF14protein was expressed in E. coli BL21DE3 or PlysScells harbouring pETA/sefA plasmid as a fusion pro-tein in a soluble form of molecular mass approxi-mately 19 kDa (Figure 1A). The fusion part of therSEF14 at the amino terminus contained amino acidresidues that facilitated the purification (His tag) andidentification (T7 tag) of the expressed protein (Fig-ure 1B). The rSEF14 antigen expressed in E. coli wasconfirmed by western blot using antibodies to fusion

TABLE 1 Specificity of the rSEF14 Protein for theDetection of Antibodies to SE

Species

Serological tests

SPTa MTb LATc ELISAd

Salmonella enteritidis 1 1 1 1Salmonella gallinarum 1 1 2 2Salmonella pullorum 1 1 2 2Salmonella dublin 1 1 1 1Salmonella berta 1 1 2 2Salmonella typhimurium 2 1 2 2Salmonella arizonae 2 2 2 2Escherichia coli 2 2 2 2Mycoplasma synoviae NT NT NT 2Mycoplasma gallisepticum NT NT NT 2Pox virus NT NT NT 2Reo virus NT NT NT 2Reticuloendothelial virus NT NT NT 2Marek’s disease virus SB-1 NT NT NT 2Infectious bursal disease virus NT NT NT 2Infectious laryngotracheitis virus NT NT NT 2Lymphoid leukosis virus A NT NT NT 2Lymphoid leukosis virus B NT NT NT 2Newcastle disease virus NT NT NT 2Chicken anaemia virus NT NT NT 2Herpes virus of turkeys NT NT NT 2Infectious bronchitis virus NT NT NT 2

a Serum plate test using pullorum whole-cell antigen; b micro-agglutination test using enteritids whole-cell antigen; c LATusing rSEF14 antigen; and d ELISA using rSEF14 antigen.NT, not tested; 1, positive; 2, negative.

150 Rajashekara et al.

peptide as well as the SEF14 antigen itself (Figure 2Aand B).

Specificity of rSEF14-LAT and rSEF14-ELISA

The rSEF14 antigen was evaluated in LAT as well asin ELISA for the detection of SE-specific antibodies.

The rSEF14 antigen-based LAT and ELISA specifi-cally identified antibodies in serum from birds in-fected with SE. Serum samples obtained from birdsinfected with other serogroup-D Salmonella serotypesdid not show antibodies to SEF14 antigen except S.dublin (Table 1). In contrast, SPT using pullorumwhole-cell antigen and MT using SE whole-cell anti-gen cross reacted with the antibodies to all theserogroup-D Salmonella serotypes tested. In addition,antibodies to various avian pathogens did not reactwith rSEF14 in ELISA.

Sensitivity of rSEF14-LAT and rSEF14-ELISA

The ability of rSEF14-LAT and rSEF14-ELISA to de-tect the antibodies in the birds exposed to differentconcentrations of SE cells was examined.

Serum Antibody Response. The results of rSEF14-LAT are shown in Figure 3. The rSEF14-LAT identi-fied antibodies to SE in 88% of the birds at 1 and 2weeks post-inoculation (PI) in the group given 1010

cfu of SE. Thereafter, the sensitivity of detectionreached 44% by 7 weeks PI. In contrast, rSEF14-LATidentified SE antibodies in 55% and 66% of the birdsin the group given 104 cfu of SE at 1 and 2 weeks PI,respectively. No bird was found positive for SE an-tibodies by rSEF14-LAT from 6 weeks PI onwards.

The results of rSEF14-ELISA for antibodies to SEin serum and egg yolk from birds exposed to SE areshown in Figure 4. In birds exposed to 1010 cfu of SE,the serum antibodies to SE were detected in 88% ofbirds at 1 and 2 weeks PI and in 100% of birds fromWeek 3 PI onwards. With the birds given 104 cfu ofSE, 55% of the birds were positive up to 4 weeks PI.The percentage of positive birds increased to 77% by7 weeks PI. The serum samples were also analyzed

FIGURE 1 A. SDS-PAGE of soluble extracts fromBL21DE3 E. coli cells expressing rSEF14 protein. Lane 1,cell extract before purification; Lane 2, cell extract afterpurification by nickel column; Lane 3, MW marker; andLane 4, electroeluted nickel column purified rSEF14 pro-tein: B. The deduced amino acid sequence of rSEF14 pro-tein. Additional amino acid residues at the amino terminusare underlined.

FIGURE 2 Western blot of rSEF14 expressed in E. coli BL21DE3 or PlysS cells. A. The soluble extracts from E. coliBL21DE3 or PlysS cell expressing rSEF14 probed with T7-tag monoclonal antibody. Lane 1, BL21DE3-(pETA/sefAplasmid); Lane 2, PlysS-(pETA/sefA plasmid); Lane 3, BL21DE3-(pETA/fimA plasmid) plasmid control; Lane 4, BL21DE3-(pETA/sefA plasmid) uninduced control; Lane 5, PlysS (pETA/sefA plasmid) uninduced control; Lane 6, BL21DE3(pETB/sefA plasmid) induced control; Lane 7, BL21DE3 (pETC/sefA plasmid) induced control; and Lane 8, BL21DE3(pETA/fimA plasmid) uninduced control. B. Western blot of nickel column purified rSEF14. Lane 1, probed withmonospecific polyclonal anti-SEF14 antibody; and Lane 2, probed with T7-tag monoclonal antibody.

S. enteritidis Immunodiagnosis 151

by conventional tests. Figure 5 and 6 shows the se-rum antibody response of the birds detected by SPTand MT.

Egg Yolk Antibody Response. The SE specific anti-body was also detected in egg yolk by rSEF14-ELISA.The SE-specific antibodies were detected as early as 6days PI in both the 104 cfu (25%) as well as 1010 cfu(28%) of SE-exposed groups. The percentage of eggspositive for SE antibodies increased to 100% by Day12 PI in birds given 1010 cfu of SE. However, in birdsgiven 104 cfu of SE only, 80% of the eggs were foundpositive for SE antibodies. The yolk samples werealso analyzed in SPT and MT. In SPT no positivereaction was observed in eggs until 15 days PI andonly 25% of the eggs were found positive throughoutthe course of the experiment (Figure 5). On the otherhand, MT test detected antibodies as early as 12 daysPI and up to 50% of the eggs were positive for SEantibodies throughout the course of the experiment(Figure 6).

Fecal Shedding of SE

The fecal shedding of SE was monitored by culturingcloacal swabs from SE-exposed chickens. Figure 7shows the fecal shedding of SE post-exposure. SEwas isolated from all the chickens administered with1010 cfu of SE only during the first week of PI. Inchickens administered with 104 cfu of SE, no SE wasisolated at 1 week PI and only 25% of birds werepositive for SE by the second week of PI. Thereafter,no isolation of SE was made from the cloacal swabsin these birds.

DISCUSSION

The infection of chickens with S. enteritidis and itsspread through table eggs into human food chain,highlights the need for procedures to identify SE-infected chickens to intervene in its transmission.Testing of primary and multiplier breeder flocks ofchickens for SE has been suggested by the UnitedStates Department of Agriculture (USDA), NationalPoultry Improvement Plan (NPIP), Centers for Dis-ease Control (CDC), and other regulatory agencies toprevent the spread of SE to commercial egg layingbirds. The control efforts for SE have focused ontesting birds suspected for SE infection and the ini-tiation of intervention programs. The increase in bac-teriological monitoring of poultry flocks for SE hasrequired an evaluation of alternative screening pro-cedures to reduce the time and costs involved. Themost likely alternative approach is serological test-ing. Currently used serologic test for monitoring SEinfection include agglutination tests or ELISA. Theagglutination tests such as SPT using S. pullorumwhole-cell antigen and MT using SE whole-cell anti-gen are not able to detect SE infection specificallyand give false positive reactions with other Salmo-nella serotypes (Barrow 1994). The ELISAs commonlyused to detect SE infection involve use of eitherlipopolysccharide (LPS) or flagellar antigen of SE.Though these tests are more sensitive than aggluti-nation tests, the major problem with these ELISAsare that they are not specific. The antigenic determi-nents of LPS and flagellar antigens of SE are alsoshared by other Salmonella serotypes that result incross reactions in the field (Barrow 1994). In addition,the group D Salmonella includes two very importantserotypes, S. gallinarum and S. pullorum, which arehighly egg transmitted and are not pathogenic tohumans, unlike SE, which belongs to the samegroup. In the poultry industry there is an eradicationprogram for S. pullorum and S. gallinarum adminis-tered by USDA. The antigen used in this eradicationprogram is a whole-cell antigen made from S. pul-lorum. This antigen does not distinguish SE, S. pul-lorum, and S. gallinarum infection of chickens. If andwhen there is a positive reaction with this antigen,the ability to differentiate it from SE infection is veryimportant to avoid confusion before declaring theflock positive for S. pullorum or S. gallinarum.

The expression of SEF14 fimbrial antigen by SEensures its differentiation from other Salmonella se-rotypes including closely related serogroup-D Salmo-nella. In this study we investigated the application ofrSEF14 fimbrial antigen for the specific serodiagnosisof chickens infected with SE. In our studies therSEF14 antigen was highly specific in LAT andELISA in identifying birds infected with SE (Table 1).Although, the sefA gene, which encodes SEF14 pro-

FIGURE 3 Detection of antibodies to SE in the serumwith rSEF14 antigen in LAT.

152 Rajashekara et al.

tein, is present in many serogroup-D Salmonella, onlyS. enteritidis, S. dublin, S. moscow, and S. blegdamexpress SEF14 antigen (Turcotte and Woodward1993). Because S. dublin, S. moscow, and S. belgdam arerarely found in birds, a presumptive identification ofSE infection in chickens is possible using SEF14 an-tigen. Even though S. blegdam, S. dublin, and S. mos-cow are very closely related antigenically to SE, theyare extremely rare in chickens. There were no isola-tions of S. blegdam and S. moscow from chickens re-ported by NVSL since 1984 except for 10 isolations ofS. blegdam reported in the year 1994. Although S.dublin isolations though occasionally reported inchickens, only a total of 10 isolations have been re-ported since 1984. Similarly, no examples of S. bleg-dam, S. dublin, and S. moscow isolations have beenreported from Central Veterinary Laboratories (CVL)

since the Zoonoses Order (1975) started in the UnitedKingdom in 1976 (Thorns et al. 1992).

We demonstrated the use of latex agglutinationtest to specifically detect SE infection. This test can bea cost-effective alternate procedure to conventionalculturing methods, which often have very low sen-sitivity because of intermittent shedding of the or-ganisms in the feces. In our study 55 to 88% of thebirds were positive by rSEF14-LAT by first-week PIdepending on the amount of SE cells inoculated (Fig-ure 3). Subsequently less number of birds was foundpositive in both the groups. This can be attributed tothe type of antibody response that the infected chick-ens will have in the early phase of infection and thefact that the test mainly detects IgM antibodies.However, in SPT and MT tests 75 to 100% of the birdsremained positive throughout course of the experi-

FIGURE 4 Detection of antibodies to SE with rSEF14 antigen in ELISA: serum antibodies (top panel) and egg yolkantibodies (bottom panel).

S. enteritidis Immunodiagnosis 153

ment in the 1010 cfu of SE-inoculated group. This isprobably because of the ability of these tests to detectantibody to variety of SE antigens including SEF14.Where as rSEF14-LAT detects antibodies to SEF14fimbrial antigens alone and thus identify birds spe-cifically infected with SE. In the group of birds ex-posed to 104 cfu of SE, less number of birds wasfound positive for SE antibodies in all three tests. Ithas been suggested that the serotypes, which havebeen associated with human food poisoning, unlikethe host-adapted strains such as S. pullorum, producelow levels of agglutinin antibodies (Cullen and Nich-olas 1991). In an examination of nearly 50 flocks witha history of SE infection the SPT detected antibody inonly 50% of flocks. More revealingly, SPT detectedantibody in only 30 of 92 naturally infected birds

from which SE was isolated from organs (Cullen andNicholas 1991). In addition, the use of SPT has manydisadvantages, it produces cross reaction with otherSalmonella serotypes including group B Salmonellasuch as S. typhimurium (Chart et al. 1990a, 1990b).

IgG antibodies can be detected in serum and eggyolk long after an active infection with SE has ceased(Barrow 1992). Several ELISA tests, based on a vari-ety of Salmonella antigens or on Salmonella monoclo-nal antibodies (van Zijderveld et al. 1992) have beendeveloped to detect antibodies to Salmonella infectionin poultry. In the present study the rSEF14-ELISAwas able to detect 55% (104 cfu) and 88% (1010 cfu) ofthe SE infected birds during first 2 weeks PI; how-ever, subsequently 77% (104 cfu) and 100% (1010 cfu)of birds remained positive for SE antibodies through-

FIGURE 5 Detection of antibodies to SE in serum plate agglutination test using pullorum whole-cell antigen; serumantibodies (top panel) and egg yolk antibodies (bottom panel).

154 Rajashekara et al.

out the course of the experiment (Figure 4). Thorns etal. (1993) have also shown that the birds infectedwith SE readily seroconvert within 7 days of infec-tion and the SEF14 IgG response persists at least 8weeks thereafter. Several investigators have usedLPS from SE in ELISA for detection of SE infection.This has often resulted in positive reaction with otherSalmonella serotypes particularly against S. gallina-rum and S. typhimurium (Desmidt et al. 1996). Thisresult is expected because both serotypes share so-matic antigens with SE (Le Minor 1984). Some inves-tigators have used flagellar antigen (gm) of SE(Timoney et al. 1990) to overcome the cross-reactionobserved with LPS-ELISA. One has to recognize thatthe serotypes sharing the same gm flagellar antigens

would result in cross reaction and may conceivablypose a problem. In the field cross reactions seem tooccur with other Salmonella serotypes (Barrow 1994).Therefore, the use of fimbrial antigen specific to SEwould overcome the drawbacks associated with theexisting ELISAs.

The results of our investigation suggested thatantibodies to SE could be detected in egg yolk usingthe rSEF14 antigen. Eggs provide a convenient andinexpensive source of antibodies. The antibodies inthe egg yolk can be detected using the same methodsused for detecting serum antibodies without affect-ing the biosecurity of the flock. In the present studythe antibodies to SEF14 antigen in the egg yolk ofSE-infected birds was observed as early as 6-day

FIGURE 6 Detection of antibodies to SE in microagglutination test using SE whole-cell antigen: serum antibodies (toppanel) and egg yolk antibodies (bottom panel).

S. enteritidis Immunodiagnosis 155

post-infection in rSEF14-ELISA and the number ofpositive eggs increased subsequently. Moreover,higher number of birds was found positive for SEF14antibodies in eggs than the serum from 104 cfu SE-inoculated group. It has been suggested that theantibody level in the egg yolk seems to exceed thosein the serum (Kasper et al. 1991; Patterson et al. 1962;Rose and Orlans 1981). Investigators in the UK havesimilarly identified antibodies to Salmonella species

at high titers in the eggs of experimentally infectedhens (Dadrast et al. 1990). However, the SPT and MTdid not detect antibodies to SE in eggs until 15 daysand 12 days post-inoculation, respectively. In addi-tion, only up to 25% and 50% of the eggs were foundpositive in SPT and MT, respectively, throughout thecourse of the experiment. The rSEF14-ELISA wasvery efficient in identifying the eggs from SE-infected birds than SPT and MT, both in terms ofpercentage of eggs giving positive reaction as well asdetecting SE antibodies in early stages of infection.

In the birds infected with either 1010 or 104 cfu ofSE, the isolation of SE from cloacal swab was veryinconsistent (Figure 7). This suggests the poor sensi-tivity of bacteriologic culture for the detection of SEinfected birds. The serological tests using rSEF14antigen identified birds exposed to SE even though,they were found negative on culture of cloacalswabs.

In conclusion, the results of the present investiga-tion suggest that rSEF14-based serological assays(LAT and ELISA) provide a sensitive and specificidentification tool for diagnosis of SE infection. Webelieve that the rSEF14-based assays will have wideapplication as screening tests in the control of SEinfection in poultry and the subsequent reduction ofhuman illnesses. Further studies to evaluate therSEF14 based serological assays for commercial useare in progress.

REFERENCES

Angula F (1997) The emergence of multidrug-resistant Sal-monella typhimurium DT104 in humans in the UnitedStates. Proceedings of Research on salmonellosis in the FoodSafety Consortium. Eds B. P. Smith and K. V. Nagaraja.US Animal Health Association meeting. Louisville,Kentucky, October 20.

Barrow PA (1992) ELISA and the serological analysis ofSalmonella infections in poultry: a review. EpidemiolInfect 109:361–369.

Barrow PA (1994) Serological diagnosis of Salmonella sero-type enteritidis infection in poultry by ELISA and othertests. Int J Food Microbiol 21:55–68.

Bullock SL, Walls KW (1977) Evaluation of some of theparameters of the enzyme-linked immunosorbent as-say. J Infect Dis 136:S279–285.

Chart H, Rowe B, Baskerville A, Humphrey TJ (1990a)Serological tests for Salmonella enteritidis in chickens. VetRec 126:20.

Chart H, Rowe B, Baskerville A, Humphrey TJ (1990b)Serological tests for Salmonella enteritidis in chickens. VetRec 126:92.

Clouthier SC, Collinson SK, Kay WW (1994) Uniquefimbriae-like structures encoded by sefD of the SEF14fimbrial gene cluster of Salmonella enteritidis. Mol Micro-biol 12:893–903.

Collinson SK, Emody L, Muller KH, Trust TJ, Kay WW(1991) Purification and characterization of thin, aggre-gative fimbriae from Salmonella enteritidis. J Bacteriol173:4773–4781.

Cooper GL, Nicholas RAJ, Bracewell CD (1989) Serologicaland bacteriological investigations of chickens fromflocks naturally infected with Salmonella enteritidis. VetRecord 125:567–572.

Cullen GA, Nicholas RA (1991) Serological analysis forantibodies to S. enteritidis. Vet Record 128:387–388.

Dadrast H, Hesketh R, Taylor DJ (1990) Egg-yolk antibodydetection in identification of Salmonella infected poul-try. Vet Rec 126:219.

Desmidt M, Ducatelle R, Haesebrouck F, De Groot PA,Verlinden M, Wijffels R, Hinton M, Bale JA, Allen VM(1996) Detection of antibodies to Salmonella enteritidis insera and yolks from experimentally and naturally in-fected chickens. Vet Record 138:223–226.

Hedberg CW, David MJ, White KE, MacDonald KL, Oster-holm MT (1993) Role of egg consumption in sporadicSalmonella enteritidis of phage type 4 for chickens andSalmonella typhimurium infections in Minnesota. J InfectDis 167:107–111.

Jensenius JC, Andersen I, Hau J, Crone M, Koch C (1981)

FIGURE 7 The fecal shedding of SE by culturing cloacalswab. The birds were inoculated with 104 cfu and 1010 cfuof SE orally and cloacal swabs were collected at weeklyintervals.

156 Rajashekara et al.

Eggs: conveniently packaged antibodies. Methods forpurification of yolk IgG. J Immunol Meth 46:63–68.

Kaspers B, Schranner I, Losch U (1991) Distribution ofimmunoglobulins during embryogenesis in the chicken.J Vet Med A38:73–79.

Kim CJ, Nagaraja KV, Pomeroy BS (1991) Enzyme-linkedimmunosorbent assay for the detection of Salmonellaenteritidis infection in chickens. Am J Vet Res 52:1069–1074.

Koeuth T, Versalovic J, Lupski JR (1995) Differential sub-sequence conservation of interspersed repetitive Strep-tococcus pneumoniae BOX elements in diverse bacteria.Genome Research 5:408–418.

Le Minor (1984) Genus 111. Salmonella Lignieres 1900. InBergey’s Manual of Systematic Bacteriology. Eds Krieg NRand Holt JG.. 1st Ed, Baltimore: Williams & Wilkins, p.447.

Mason J, Ebel E (1992) USDA-APHIS Salmonella enteritidiscontrol program. In: Proceedings of the Symposium on theDiagnosis and Control of Salmonella. Ed, GH Snoeyenbos.Richmond, Virginia: US Animal Health Association, p.78.

Muller KH, Collinson SK, Trust TJ, Kay WW (1991) Type 1fimbriae of Salmonella enteritidis. J Bacteriol 173:4765–4772.

Patterson R, Youngner JS, Weigle WO, Dixon FJ (1962)Antibody production and transfer to egg yolk in chick-ens. J Immunol 89:272–278.

Polson A, Von Wechmar MB, Van Regenmortel MHV(1980) Isolation of viral IgY antibodies from yolks ofimmunized hens. Immunol Commun 9:475–493.

Rose ME, Orlans E (1981) Immunoglobulins in the egg,embryo and young chick. Dev Comp Immunol 5:15–20.

Rodrigue DC, Tauxe RV, Rowe B (1990) International in-crease in Salmonella enteritidis: A new pandemic? Epide-miol Infect 105:21–27.

St. Louis ME, Morse DL, Potter ME, DeMelfi TM, Guze-wich JJ, Tauxe RV, Blake PA (1988) The emergence ofgrade A eggs as a major source of Salmonella enteritidisinfection. J Am Med Assoc 259:2103–2107.

Thorns CJ, Sojka MG, Chase D (1990) Detection of a novelfimbrial structure on the surface of Salmonella enteritidisby using a monoclonal antibody. J Clin Microbiol 28:2409–2414.

Thorns CJ, Sojka MG, McLaren IM, Dibb-Fuller M (1992)Characterization of monoclonal antibodies against afimbrial structure of Salmonella enteritidis and certainother serogroup D salmonellae and their application asserotyping reagents. Res Vet Sci 53:300–308.

Thorns CJ, Nicholas RAJ, Bell MM, Chrisnall SC (1993)Preliminary results on the use of SEF14 fimbrial antigenin an ELISA for the serodiagnosis of Salmonella enterit-idis infection in chickens. Proceedings of the EC workshopon ELISA for serological diagnosis of Salmonella infection inpoultry. Brussels, 7–9 June. Directorate General for Ag-riculture, Commission of the European communitiesBrussels.

Timoney JF, Sikora N, Shivaprasad HL, Opitz M (1990)Detection of antibody to Salmonella enteritidis by a gmflagellin-based ELISA. Vet Rec 127:18–19.

Turcotte C, Woodward MJ (1993) Cloning, DNA nucleo-tide sequence and distribution of the gene encodingSEF14, a fimbrial antigen of Salmonella enteritidis. J GenMicrobiol 139:1477–1485.

Van Zijderveld FG, Van Bemmel AMV, Anakotta J (1992)Comparison of four different enzyme-linked immu-nosorbent assays for serological diagnosis of Salmonellaenteritidis infections in experimentally infected chickens.J Clin Microbiol 30:2560–2566.

Williams JE, Whittemore AD (1971) Serological diagnosisof pullorum disease with the microagglutination sys-tem. Appl Microbiol 21:394–399.

S. enteritidis Immunodiagnosis 157