Case report
Peer reviewed

Diagnosis of Yersinia enterocolitica serovar O:9 in a commercial 2400-sow farm with false-positive Brucella suis serology using western blot, competitive ELISA, bacterial isolation, and whole genome sequencing

Ross A. Free, DVM; Mark Ladd, DVM; Randy Capsel, PhD; Linda Cox, MS; Jessica Hicks; Kristina Lantz, PhD; Michael J. Neault, DVM; James Kittrell, DVM; Barry J. Meade, DVM


Despite eradication of swine brucellosis from US commercial swine, Brucella suis still exists in feral swine. Therefore, brucellosis surveillance occurs to detect and eliminate any disease introduction from feral swine to domestic swine. As serology for swine brucellosis has imperfect specificity, false-positive serological reactions (FPSRs) occur and true brucellosis infection must be ruled out. In this case report, we detail a process to rule out B suis infection in a commercial sow herd using additional diagnostics including bacterial culture, whole genome sequencing, western blot, and competitive enzyme-linked immunosorbent assay. It was determined Yersinia enterocolitica serovar O:9 caused the FPSRs.

Keywords: swine, Brucella suis, false-positive, Yersinia enterocolitica, serology

Received: May 7, 2019
Accepted: October 7, 2019

Resumen — Diagnóstico de Yersinia enterocolitica serovar O:9 en una granja comercial de 2400 cerdas con un falso-positivo de serología de Brucella suis usando Western blot, ELISA competitiva, aislamiento bacteriano, y secuenciación del genoma complet

A pesar de la erradicación de la brucelosis porcina de los cerdos comerciales de EE UU, Brucella suis todavía existe en los cerdos salvajes. Por lo tanto, la vigilancia de la brucelosis se usa para detectar y eliminar cualquier introducción de enfermedad de cerdos salvajes a cerdos domésticos. Como la serología para la brucelosis porcina tiene una sensibilidad imperfecta, se producen reacciones serológicas falsas-positivas (FPSRs), y se debe descartar una verdadera infección por brucelosis. En este reporte de caso, detallamos un proceso para descartar la infección por B suis en una piara de cerdas comerciales utilizando diagnósticos adicionales que incluyen cultivo bacteriano, secuenciación del genoma completo, western blot y ensayo competitivo de inmunoadsorción ligado a enzimas. Se determinó que Yersinia enterocolitica serovar O:9 causó las FPSRs.

Résumé — Détection de Yersinia enterocolitica serovar O:9 dans une ferme commerciale de 2400 truies présentant des résultats faux-positifs à Brucella suis par sérologie en utilisant l’immunobuvardage, un ELISA compétitif, l’isolement bactérien, et le séquençage du génome entier

Malgré l’éradication de la brucellose porcine chez les porcs américains commerciaux, Brucella suis est présent chez les porcs sauvages. Ainsi, la surveillance pour la brucellose porcine existe afin de détecter et d’éliminer toute transmission de la maladie des porcs sauvages aux porcs domestiques. Étant donné que le test sérologique pour la brucellose a une sensibilité imparfaite, des réactions faussement-positives (FPSRs) se produisent, et une véritable infection brucellique doit être exclue. Dans le présent rapport de cas, nous détaillons un processus pour exclure l’infection à B suis dans un troupeau commercial de truies en utilisant des méthodes diagnostiques additionnelles incluant la culture bactérienne, le séquençage du génome complet, l’immunobuvardage, et une épreuve ELISA compétitive. Il fut déterminé que Yersinia enterocolitica O:9 était responsable des FPSRs.

Swine brucellosis was eradicated in the US commercial swine herd in 2011 when Texas was added as the final validated brucellosis-free state.1 In spite of this eradication success, Brucella suis continues to exist in a wildlife carrier, feral swine.2,3 Brucella suis presents a risk of disease re-introduction to domestic swine via contact with feral swine and presents an ongoing risk of zoonotic disease to people who have contact with blood or other body fluids from infected swine.4 Therefore, swine brucellosis disease surveillance programs exist at US slaughter plants to allow prompt detection and removal of infected domestic swine and to provide assurance to international trading partners that US commercial swine herds are brucellosis-free.

False-positive serological reactions (FPSRs) are common when testing for swine brucellosis, and Yersinia enterocolitica serovar O:9 appears to be the most common cause of these false-positive tests due to the similar lipopolysaccharide (LPS) O-antigens in both organisms.5,6 Additionally, Y enterocolitica serovar O:9 has also been shown to cause FPSRs in cattle that are serologically tested for Brucella abortus for the same reason.7 Many researchers have sought to create serologic tests7-9 that cancel out cross-reactivity and either prevent or rule out these FPSRs.

In spite of these efforts, there is still no dependable serological test for the diagnosis of swine brucellosis in an individual animal.10 Hence, ruling out a true swine brucellosis infection in a seropositive animal or herd comes at a considerable cost to the swine producer due to time spent under quarantine and to the state or federal government due to required additional testing to ensure a herd is not infected with B suis.11,12 In the absence of an alternative method, a serologic surveillance program with specificity less than 100% will continue to be used and FPSRs will need to be investigated. This case report details a diagnostic work-up to rule out swine brucellosis in a herd with FPSRs, and Y enterocolitica serovar O:9 was isolated and deemed to be the cause of the FPSRs.

Case description

Initial herd investigation

In February 2017, the National Veterinary Services Laboratories (NVSL) notified the North Carolina Veterinary Services office of a swine brucellosis reactor animal found by slaughter surveillance. Serology results from the cull sow collected at slaughter revealed fluorescent polarization assay (FPA) values of 85/80 Delta millipolarization units (mP; each sample was analyzed twice for comparison and reported as two values, eg, 85/80; negative reference range: < 10 Delta mP, suspect reference range: 10-20 Delta mP, and positive reference range: > 20 Delta mP) and complement fixation (CF) value of 2+ at a 1:10 dilution (negative reference range: no complement fixation occurs at a 1:10 dilution). This animal was traced to a 2400-sow farm in North Carolina. The source herd did not have clinical signs suggesting swine brucellosis infection. The herd was kept in closed buildings and potential exposure to feral swine was considered negligible. Pigs that were weaned from the source farm were destined for market production only after shipment to a nursery and then to a finishing unit. No females weaned from this farm were kept as replacement gilts. Serological testing of 160 breeding females was conducted within the source herd. The herd was placed under quarantine due to the positive herd test serology and replacement females could not enter and cull sows could not leave during the investigation period. The finishing units that ultimately received pigs from the sow farm flow could move pigs to slaughter under permit during the investigation period.

In order to differentiate an FPSR situation from a truly infected swine brucellosis herd, the North Carolina Department of Agriculture, the US Department of Agriculture, and the herd veterinarian agreed that 4 of the sows with high titers should be humanely euthanized by the herd veterinarian and necropsied at the North Carolina Veterinary Diagnostic Laboratory System (NCVDL). Because the herd had no clinical signs of swine brucellosis, more sows were not sacrificed for tissue collection, thus preventing unnecessary loss to the producer. Tissues sampled from each euthanized sow were submitted to the NVSL for culture. Tissue samples included mandibular lymph nodes, retropharyngeal lymph nodes, hepatic lymph nodes, internal iliac lymph nodes, superficial inguinal lymph nodes, mesenteric lymph nodes, kidney, and tonsil. These tissues were examined in order to maximize the likelihood of isolating B suis if it was present in the animals. Three of the 4 euthanized sows were pregnant and fetal lung, amniotic fluid, and placenta samples were submitted for culture (Table 1).

Table 1: Serologic titers and culture results from four sows that were euthanized and tissues collected to determine swine brucellosis status

Sow ID (Parity) Card test FPA, Delta mP CF value (dilution) MLN RLN Tonsil ALN Kidney* AF* PL* FL*
1979 (5) Pos 98/92 2+ (1:60) YE NI YE NI NI NI NI NI
5218 (0) Pos 94/96 4+ (1:80) YE YE YE YE NI NI NI NI
2870 (4) Pos 52/47 3+ (1:10) NI NI YE NI NI NS NS NS
2672 (5) Pos 91/100 2+ (1:10) NI NI YE NI NI NI NI NI

*   Brucella isolation attempt only, no Yersinia isolation attempt.

ID = identification; FPA = fluorescent polarization assay; mP = millipolarization units; CF = complement fixation (cold) test; MLN = mandibular

lymph node; RLN = retropharyngeal lymph node; ALN = additional lymph nodes; AF = amniotic fluid; PL = placenta; FL = fetal lung;

Pos = positive; YE = Yersinia enterocolitica; NI = no isolation of Brucella suis or Y enterocolitica; NS = not submitted (sow not pregnant).

If B suis was isolated from the collected tissues, whole-herd depopulation and further tissue collection would be the likely outcome. More samples for culture (placenta and milk) would become available as sows farrowed, which would provide further evidence of a negative herd status and further prevent the need for sacrificing additional animals. Resampling of the remaining seropositive sows in the source herd was accomplished 39 days after initial samples were taken and titers were compared (Table 2). One-time milk and placenta samples were collected from 8 sows with titers when they farrowed and were submitted for isolation of B suis at the NCVDL (Table 3). A third set of serological testing was completed on 8 of the seropositive sows on the source farm between 88 to 104 days after the initial herd test (Table 2).

Table 2: Chronological decrease of titers in sows seropositive for swine brucellosis

Sow ID Initial FPA, Delta mP* Follow-up FPA, Delta mP* Final FPA, Delta mP* Initial CF value (dilution) Follow-up CF value (dilution) Final CF value (dilution) cELISA, %I Follow-up cELISA, %I
5128 21/20 25/26 NS 2+ (1:80) 2+ (1:40) NS 31.5 NS
4613 14/15 34/31 NS 3+ (1:10) 1+ (1:40) NS 40.1 NS
5053 117/113 72/74 36/31 2+ (1:160) 1+ (1:80) 2+ (1:10) 76.0 28.8
4505 21/19 9/10 NS 2+ (1:10) Neg (1:10) NS 20.2 NS
5097 16/16 10/9 7/6 3+ (1:10) Neg (1:10) Neg (1:10) 12.0 -4.9
4029 42/40 19/20 NS Neg (1:10) Neg (1:10) NS 28.6 NS
3647 34/35 22/30 NS 1+ (1:10) 1+ (1:10) NS 29.8 NS
4177 32/25 20/19 NS 1+ (1:40) 1+ (1:10) NS 28.7 NS
2623 29/28 17/16 8/7 Neg (1:10) Neg (1:10) Neg (1:10) 25.9 4.65
4146 86/89 39/51 NS 4+ (1:40) 1+ (1:20) NS 54.1 NS
3284 41/46 27/23 NS 3+ (1:10) 1+ (1:10) NS 32.3 NS
2169 24/27 13/13 -1/-2 4+ (1:20) 1+ (1:10) Neg (1:10) 23.6 10.8
4026 34/29 32/28 NS 2+ (1:10) 2+ (1:10) NS 52.0 NS
5004 70/66 54/53 NS 4+ (1:40) 1+ (1:20) NS 40.6 NS
4259 54/52 25/22 NS 3+ (1:20) 1+ (1:10) NS 26.1 NS
5142 50/50 29/26 10/11 3+ (1:20) Neg (1:10) Neg (1:10) 21.7 -20.7
4539 25/21 15/13 5/5 4+ (1:40) 1+ (1:10) Neg (1:10) 29.0 1.6
5191 117/115 79/70 NS 2+ (1:160) 1+ (1:40) NS 84.1 NS
4530 61/61 41/37 NS 3+ (1:40) 3+ (1:10) NS 81.2 NS
4167 18/18 42/41 NS 2+ (1:10) 3+ (1:20) NS 61.8 NS
3160 49/42 24/26 NS 2+ (1:10) 3+ (1:10) NS 25.6 NS
3163 50/47 55/52 NS 3+ (1:10) 2+ (1:80) NS 64.5 NS
3153 73/74 69/63 NS Neg (1:10) 1+ (1:10) NS 48.8 NS
4216 47/50 22/19 NS 4+ (1:20) 2+ (1:10) NS 35.7 NS
4702 23/19 27/28 5/4 Neg (1:10) 2+ (1:10) Neg (1:10) 20.1 10.0
5217 19/23 19/18 6/6 2+ (1:10) 2+ (1:10) Neg (1:10) 20.0 -1.9

*   FPA reference ranges: < 10 Delta mP = negative; 10-20 Delta mP = suspect; > 20 Delta mP = positive.

†   cELISA reference ranges: ≥ 30% inhibition = positive; < 30% inhibition = negative.

ID = identification; FPA = fluorescence polarization assay; mP = millipolarization units; CF = complement fixation (cold) test;

cELISA = competitive enzyme-linked immunosorbent assay; I = inhibition; NS = not submitted; Neg = negative.

Table 3: Initial serologic titers and culture results from sows immediately post-partum

Sow ID Initial FPA, Delta mP Initial CF value (dilution) Milk culture Placenta culture
2623 29/28 Neg (1:10) NI NI
5053 117/113 2+ (1:160) NS NI
4539 25/21 4+ (1:40) NI NI
2169 24/27 4+ (1:20) NI NI
5217 19/23 2+ (1:10) NI* NS
5142 50/50 3+ (1:20) NI* NS
5097 16/16 3+ (1:10) NI* NS
4702 23/19 Neg (1:10) NI* NI

*   Isolation of Yersinia enterocolitica attempted in addition to Brucella suis isolation attempt.

ID = identification; FPA = fluorescent polarization assay; mP = millipolarization units; CF = complement fixation (cold) test; NI = no isolation of

Brucella suis; NS = not submitted.

Serological sampling of breeding females in the source herd and from swine in epidemiologically linked herds was conducted to approximate a 95% confidence level of finding an infected animal assuming a 2% herd prevalence and 90% diagnostic test sensitivity. The brucellosis card test (NVSL SOP-SERO-0020) was used for sample screening and FPA (NVSL SOP-SERO-0021) and CF (NVSL SOP-SERO-0015) were used as confirmatory tests. In addition, for selected secondary sow samples, a competitive enzyme-linked immunosorbent assay (cELISA; NVSL SOP-SERO-0023; Boehringer Ingelheim Svanova) was performed as a potential highly specific differential test. All serological testing was conducted using standard operating procedures administered by the NVSL which are controlled documents and available through the NVSL Quality Assurance program section (

Testing of the 160 breeding females at the source herd identified 35 animals as card positive, and these positive serum samples were sent to the NVSL for confirmatory testing. Of these 35 card-positive animals, 23 animals were positive in both the FPA and the CF, 3 animals were suspect in the FPA and positive in CF (Table 2). Four animals were positive in the FPA and negative in the CF, and the remaining 5 animals were negative in both the FPA and the CF. Of the 26 sows resampled from the source farm, 21 had a decrease in the mean FPA value and 20 had a decrease or no change in the CF value (Table 2). Half of the 26 animals were negative in the cELISA (Table 2).

Serologic investigation of epidemiologically linked herds

The source herd received replacement gilts from a single 2400-sow multiplier herd. Serological testing was conducted on 164 animals from the multiplier herd. Gilts from the multiplier were sent to a nursery and then a finisher before arriving at the source herd for breeding. The multiplier finisher (7920-head farm) that supplied gilts to the source farm also had serum collected from 167 gilts. The source herd had no boars, but had received semen from 2 boar studs in the previous 12 months. The 2 boar studs, which housed 430 and 532 boars, had 143 and 150 animals sampled, respectively.

During quarantine, farms within 2.4 km of the quarantined herd were identified by the North Carolina Department of Agriculture. Five farms in this radius were commercial finishing units, and 2 were backyard swine producers. The backyard swine producers each had 1 breeding female on their respective farms.

All boars tested from the first boar stud were negative in the card test. Two boars from the second boar stud were positive in the card test; these samples were shipped to the NVSL for confirmatory serological testing and both samples were negative in the FPA and the CF test.

Serological testing of 164 animals from the multiplier herd indicated 32 of the 164 breeding females were positive using the card test. These positive samples were sent to the NVSL for confirmatory serological testing, and all 32 samples were negative using the FPA and the CF tests. Serum samples collected from the 167 gilts at the finisher farm resulted in 2 samples being positive using the card test. These samples were sent to the NVSL for confirmatory serological testing and were both negative using the FPA and the CF test. The 2 breeding females from the backyard swine producers were found to be serologically negative for brucellosis. All epidemiologically linked herds were considered negative for swine brucellosis.

Brucella culture testing

At the NVSL, culture for B suis was performed as previously described,13 with a modification for the use of a blender to homogenize tissues. At the NCVDL, tissues were aseptically placed in a sterile plastic bag with trypticase-soy broth and macerated for up to 10 minutes. A sterile swab was used to inoculate the following media: 1) Brucella serum tryptose agar plate (made in-house) composed of horse serum (5 mL/500 mL of prepared media), polymixin B (1.5 mL/500 mL of prepared media), cyclohexamide (2.5 mL/500 mL of prepared media), and bacitracin (1 mL/500 mL of prepared media); 2) Brucella crystal violet tryptose agar plate (made in-house) composed of 1% crystal violet solution (0.7 mL/500 mL of prepared media), polymixin B (1.5 mL/500 mL of prepared media), cycloheximide (2.5 mL/500 mL of prepared media), and bacitracin (1 mL/500 mL of prepared media); and 3) Brucella selective tryptose agar plate composed of heat inactivated horse serum (25 mL/500 mL of prepared media) and Brucella Selective Supplement (Oxoid Brucella Selective Supplement, ThermoFisher Scientific; 10 mL/500 mL of prepared media) containing 2500 IU of polymyxin B, 12,500 IU of bacitracin, 50 mg of cycloheximide, 2.5 mg of nalidixic acid, 50,000 IU of nystatin, and 10 mg of vancomycin.

The plates were incubated at 37°C in 5% to 7% CO2. Plates were examined daily for 14 days. Any colonies with a morphology consistent with Brucella species would have been subcultured to a blood agar plate and examined by Gram stain. Isolates exhibiting the typical Brucella Gram stain (gram-negative coccobacilli, or short rods) would have been further tested by performing a Koster’s stain, an oxidase test, and inoculating a triple sugar iron (TSI) slant and a urea slant. If presumptive tests were positive for Brucella species, the isolate would have been forwarded to the NVSL.

Yersinia culture testing

Bacterial culturing for Yersinia at the NVSL was conducted by cutting tissues into 1 to 2 mm pieces with sterile scissors or sterile scalpels and put into peptone sorbitol bile broth (PSBB; made in-house) in a 1:10 ratio and thoroughly vortexed. The PSBB consisted of 8.23 g sodium phosphate dibasic anhydrous (Sigma-Aldrich), 1.2 g sodium phosphate monobasic monohydrate (Avantor), 1.5 g bile salts mixture (Becton, Dickinson and Co), 5 g sodium chloride (Fisher Scientific), 10 g D-sorbitol (Sigma-Aldrich), 5 g Bacto peptone (Becton, Dickinson and Co), and was brought to 1000 mL with sterile water.14 The PSBB was incubated at 10°C for 10 to 12 days.

After incubation was complete, the PSBB was thoroughly vortexed. A swab was used to sample the PSBB and then plated directly onto MacConkey (MAC; Remel) and Yersinia Selective agar (cefsulodin-irgasan-novobiocin; CIN; Remel) and streaked for isolation. Also, 0.1 mL PSBB was transferred to 0.9 mL of 5% potassium hydroxide (Sigma-Aldrich) in normal saline and vortexed. This was plated onto MAC and CIN agar using a swab and streaked for isolation. Another 0.1 mL PSBB was transferred to 0.9 mL normal saline and swabbed on MAC and CIN agar and streaked for isolation. All plates were incubated at 30°C for 1 to 2 days.

After incubation the plates were read and suspect colonies were streaked on trypticase soy agar with 5% sheep blood agar plates (Remel) which were incubated at 30°C for 1 to 2 days. Isolated colonies were identified by Matrix Assisted Laser Desorption/Ionization-Time of Flight (MALDI-TOF) using Bruker Biotyper software (Bruker Daltonics) on a Bruker Autoflex MALDI-TOF (Bruker Daltonics).

For bacterial culturing for Yersinia at the NCVDL, tissues were aseptically placed in a sterile plastic bag with trypticase-soy broth and macerated using a stomacher for up to 10 minutes. A sterile swab was used to inoculate a MAC agar plate and a CIN agar plate. The plates were incubated at 30°C in ambient air for 48 hours. The swab was also used to inoculate a sterile tube containing phosphate buffered saline (PBS; pH 7.4). This tube was stored at 2°C to 8°C for up to 21 days with weekly subcultures to MAC and CIN agar plates which were also incubated at 30°C for 48 hours.

Original plates and plates from weekly subcultures were observed for colonies exhibiting morphologies consistent with Yersinia species. Suspicious colonies, if they had been found, would have been further tested by inoculating biochemicals including a TSI slant, a urea slant, and two sulfide, indole, motility tubes (one at 30°C and one at 37°C). Oxidase and indole tests would also have been performed. If presumptive tests were consistent with Yersinia, an Analytical Profile Index 20E (bioMérieux, Inc) would have been set up.

Yersinia isolate sequencing and serovar determination

One isolate of Y enterocolitica from sow 1979 and sow 5218 were streaked on blood agar plates and incubated at 37°C for 18 to 24 hours. Genomic DNA was extracted from each isolate using the Promega Maxwell RSC 48 instrument with the Maxwell RSC whole blood DNA kit (Promega). Isolates were sequenced on an Illumina MiSeq (Illumina) using 2 × 250 paired end chemistry and the NexteraXT (Illumina) library preparation kit. Each isolate was aligned using the Burrows-Wheeler Aligner-MEM algorithm to reference genomes for Y enterocolitica serovar O:3 strain Y11 (GenBank accession NC_017564), Y enterocolitica serovar O:8 strain 8081 (GenBank accession NC_008800), and Y enterocolitica serovar O:9 strain 105.5R(r) (GenBank accession CP002246). Alignments and annotation were viewed using Integrative Genomics Viewer version 2.3.97. Samtools was used to output depth of coverage at each position, which was used to determine percent coverage of the O-antigen clusters. In addition, the Genome Annotation Toolkit’s Unified Genotyper was used to call single-nucleotide polymorphisms for determining percent identity of the O-antigen clusters.

Western blot testing

Twelve serum samples from seropositive sows were subjected to western blot testing to differentiate between Yersinia and Brucella antibodies. Antigens were prepared from B abortus strain 2308 and strain RB51, and from Y enterocolitica serovar O:8 (Y enterocolitica subspecies enterocolitica ATCC 51871) and serovar O:9 (Y enterocolitica subspecies enterocolitica ATCC 55075), using a cell lysis extraction kit (CellLytic B cell lysis solution, Sigma-Aldrich) according to the manufacturer’s directions followed by centrifugation at 5018g. The supernatant was retained with subsequent filtration using a 0.2 µm syringe filter. The antigen preparations were a crude extract containing outer membrane and cytoplasmic proteins. Resulting suspensions were tested for inactivation. Precast 4% to 12% Novex Bis-Tris gels (12 well, 1 mm thickness, ThermoFisher Scientific) were used for SDS-PAGE separation of proteins. Respective protein suspensions were prepared by the addition of 60 µL of sample buffer (4x NuPage LDS Sample Buffer, ThermoFisher Scientific) to 180 µL of antigen. Preparations were heated at 70°C for 10 minutes prior to loading 15 µL into pre-assigned gel lanes. The approximate protein concentrations for each respective antigen well was B abortus 2308 = 9 mg; B abortus RB51 = 8 mg; Y enterocolitica serovar O:8 = 40 mg; Y enterocolitica serovar O:9 = 25 mg.

Electrophoresis was conducted in an Invitrogen XCell SureLock Mini-Cell system (ThermoFisher Scientific) at a constant current of 125 mA for 35 minutes. A control gel to be used as a western blot comparative standard was prepared by including Invitrogen SeeBlue Plus2 prestained molecular standard (ThermoFisher Scientific) to serve as a marker for molecular weight determination in one lane of the respective gel.

Electrophoretic transfer of proteins onto nitrocellulose was performed using the Invitrogen XCell II Blot Module (ThermoFisher Scientific) and Invitrogen NuPAGE transfer buffer (ThermoFisher Scientific) at 160 mA for 1 hour. After transfer, membranes were blocked with PBS (pH 7.0) with 0.5% Tween 20 plus 2% bovine serum albumin (PBST+BSA) at room temperature for 1 to 2 hours with rocking. Membranes were washed 3 times with PBS plus 0.5% Tween 20 (PBST). Nitrocellulose sheets were then cut into 3 sections, with each section containing duplicate antigen lanes, for incubation with swine sera. Swine sera were diluted at either 1:50 or 1:200 in PBST+BSA and incubated with the membranes at room temperature on a rocker platform for approximately 60 minutes. Membranes were washed 3 times with PBST.

Membranes were incubated for approximately 3 minutes at room temperature on a rocker with Pierce peroxidase conjugated Protein A (ThermoFisher Scientific) diluted 1:20,000 in PBST+BSA. Membranes were then washed 3 times with PBST. Membranes were developed in Sigma TMB Substrate (Sigma-Aldrich) according to the manufacturer’s directions.

Bacterial isolation results

Three of the 4 sows at the source farm that were euthanized after the herd quarantine were pregnant, and none of the 4 sows had gross lesions on necropsy. Brucella suis was not isolated from any tissue sample from the 4 euthanized sows. Yersinia enterocolitica was isolated from all 4 sows, with tonsil being the most common tissue of successful isolation (4 of 4 animals). One animal also yielded Y enterocolitica from the mandibular lymph node. Yersinia enterocolitica was isolated from multiple lymph nodes of the fourth animal including mandibular, supra-pharyngeal, internal iliac, and superficial inguinal nodes (Table 1).

Alignment to the O-antigen cluster of Y enterocolitica serovar O:3 had 18% coverage with 99.6% identity, Y enterocolitica serovar O:8 had 58% coverage with 97.8% identity, and Y enterocolitica serovar O:9 had 100% coverage with 99.97% identity. Alignments of both isolates with the O-antigen cluster are consistent with an identification as Y enterocolitica serovar O:9 as previously described.15 The regions of Y enterocolitica serovar O:3 and Y enterocolitica serovar O:8 O-antigen clusters with sequence coverage correspond directly with genes that are homologous to genes present in the Y enterocolitica serovar O:9 O-antigen cluster. Unique regions of the O-antigen clusters showed no sequence coverage, consistent with absence of the O:3 and O:8 O-antigen clusters.

Western blot evaluation

The 1:50 serum dilution resulted in an overload of antibody preventing clear interpretation of the blot results. There was excessive smearing observed at the bottom of the Yersinia antigen lanes and across other lanes on the blot. Multiple protein band reactivity against B abortus strain RB51 antigen was observed with the 1:50 serum dilutions and is normally not observed. This was attributed to non-specific binding due to the overload of antibody. A 1:200 serum dilution improved the ability to decipher banding patterns and reduce smearing and nonspecific binding (Figure 1). However, due to very high antibody levels to Yersinia the incubation time was kept to a minimum, with the reactivity resulting in heavy staining with moderate smearing between the 38 and 14 kDa molecular weight ranges in both the Y enterocolitica serovar O:8 and Y enterocolitica serovar O:9 antigen lanes. Immunoreactivity was observed against multiple protein bands in both the Y enterocolitica serovar O:8 and Y enterocolitica serovar O:9 antigen lanes with strong reactivity noted at bands of approximately 35, 28, 20, 12, and 5 kDa molecular weight. Moderate to strong immunoreactivity was also observed in both the Y enterocolitica serovar O:8 and Y enterocolitica serovar O:9 antigen lanes corresponding to molecular weights of approximately 98, 62, 60, 58, and 50 kDa. Moderate staining intensity accompanied by smearing was observed against multiple proteins of the B abortus strain 2308 antigen in ranges between 28 and 90 kDa, but of less intensity than observed against both the Yersinia antigen proteins.

Figure 1: Western blot of 2 sow samples (sows 4539 and 4529) tested using 2 Brucella and 2 Yersinia antigen preparations. Lanes 1and 5: Brucella abortus strain 2308 antigen; Lanes 2 and 6: B abortus strain RB51 antigen; Lanes 3 and 7: Yersinia enterocolitica serotype O:8 antigen; Lanes 4 and 8: Y enterocolitica serotype O:9 antigen. Sow serum was diluted 1:200 in blot blocker. Note strong reactivity in Yersinia antigen lanes corresponding to approximate 35, 28, 20, 12, and 5 kD molecular weights. Moderate reactivity in Yersinia antigen lanes corresponding to approximate molecular weights ranging between 48 and 62 kD. A strong band of reactivity is noted in the B abortus RB51 antigen lane corresponding to an approximately 38 kD protein band.

Reactivity to a single protein band (approximately 38 kDa) within the B abortus strain RB51 antigen was consistently observed for all sow samples. Stronger immunoreactivity against both the Y enterocolitica serovar O:8 and Y enterocolitica serovar O:9 low and mid-molecular weight antigens in comparison to lower reactivity observed within the two Brucella antigen lanes were indicative of positive Yersinia antibody reactivity. In addition, comparing results obtained with a control blot using brucellosis positive bovine field samples, bovine positive control serum, and Y enterocolitica serovar O:8 and Y enterocolitica serovar O:9 control serum, there was a lack of strong reactivity of the sow serum to low to medium molecular weight proteins (3 to 28 kDa) against the B abortus strain 2308 antigen (Figure 2). The sow serum also resulted in a greater number of protein bands staining within both Yersinia antigen lanes as compared to results observed with the brucellosis control sera to the Yersinia antigen preparation. The reactivity of the sow sera to a single protein band in the B abortus strain RB51 antigen at approximately the 38 kDa molecular weight range was also consistent with that observed in the control blot using the Y enterocolitica serovar O:8 and Y enterocolitica serovar O:9 control serum.

Figure 2: Western blot using control sera for evaluation of expected results for analysis. Lanes 1, 5, and 9: Yersinia enterocolitica serotype O:8 antigen; Lanes 2, 6, and 10: Brucella abortus strain RB51 antigen; Lanes 3, 7, and 11: B abortus strain 2308 antigen; Lanes 4, 8, and 12: pre-stained molecular weight marker. Control sera used were: Lanes 1-4: B abortus bovine field sample; Lanes 5-8: B abortus 12-H (high positive control serum); Lanes 9-12: Yersinia enterocolitica O:8 positive rabbit control serum. Serum was diluted 1:100 in blot blocker. Strong homologous reactivity was evident in lane 9 of the Yersinia control serum at the 38, 28, and 3-5 kD range. Specific reactivity of the Yersinia control serum was noted at the 38 kDa protein band to the B abortus RB51 antigen and moderate reaction at the 28 kDa range for both the B abortus RB51 and Strain 2308 antigens. Of significant interest was the distinction noted of the B abortus field sample and control serum reacting to the 49 kDa protein of the RB51 antigen. Multiple bands of reactivity to high molecular weight proteins visible against the homologous Yersinia control serum evident in Lane 9 that is not as prevalent in Lanes 1 and 5 with the Brucella control serum. Strong contrast of reactivity is noted in the 30-50 kDa range between Lanes 3, 7, and 11 with the varying control serum, indicating strong reactions of the Brucella positive serum and lack of reaction of the Yersinia control serum.

As was noted in the control blot the Brucella control and field serum samples react with a higher molecular weight RB51 antigen at approximately 49 kDa. This higher molecular weight RB51 protein band was not visible from the swine sera tested on the immunoblot procedure. Strong reactivity to both Yersinia antigens, the lack of similar reactivity to the B abortus strain 2308 antigen, and specific reactivity to the B abortus strain RB51 antigen 38 kDa protein band indicated the sow sera contained high levels of Yersinia antibody.

After culture results became available on the euthanized sows, coupled with the western blot results and declining serologic titers, the herd received a partial quarantine release that enabled the herd to move cull sows direct to slaughter but not to buying stations and to receive replacement gilts. Brucella was not isolated from any milk or placental samples taken from farrowing sows between 88 and 104 days from the initial herd test (Table 3). Yersinia was not isolated from any of these samples (Table 3). Once culture results became available on milk and placenta samples, the herd received a full quarantine release.


This report describes the difficulties associated with FPSRs for swine brucellosis. These FPSRs cause significant economic costs to both the producer and the state government due to time spent under quarantine, labor for follow-up testing, and costs associated with confirmatory diagnostic tests including serology and culture. This case reveals potential methods for dealing with this situation in the future. Serologic titers in this case report did decline over time and can be used as evidence for FPSRs as has previously been discussed.5 However, this is not ideal as the herd must remain under quarantine during the waiting period between serial sampling. Yersinia enterocolitica was readily cultured from the tissues of sows with swine brucellosis titers, but this requires the sacrifice of productive females from the herd.

As Y enterocolitica has been isolated from bovine raw milk samples,16,17 attempts were made to isolate the organism from some of the post-partum milk samples. Swabbing the tonsils of swine has been shown to be a possible method of isolating Y enterocolitica from carrier swine.18,19 Reports of similar cases in species other than swine have cultured Y enterocolitica from the feces of infected animals.20,21 However, these methods would not rule out the potential for dual infection with Yersinia and Brucella, and therefore would not be a suitable test for ruling out FPSRs. It should be noted that newer cell-mediated assays11,12,22 have shown promise when used to rule out FPSRs, however, they were not utilized for this investigation.

The use of a developmental western blot assay in this investigation added to the evidence that the herd was not infected with B suis. However, interpretation is somewhat subjective and does not provide ample evidence by itself for a diagnosis of FPSR and subsequent quarantine release. Western blot results from our study were consistent with previous studies 23,24 using Brucella positive bovine control serum resulting in intense protein band staining between 29 and 68 kDa against a smooth Brucella antigen preparation. In addition, during development of the western blot assay Brucella positive bovine field samples and Brucella positive bovine control serum were evaluated against Y enterocolitica serovar O:8 antigen. The brucellosis positive bovine samples had limited reactivity detected at approximately 38, 48, and 98 kDa against the Y enterocolitica O:8 antigen. These results are similar to a previous study25 indicating little to no reaction of B abortus positive serum against Y enterocolitica O:8 LPS and proteins, whereas there was greater cross-reactivity against Y enterocolitica O:9 LPS and proteins. The use of Y enterocolitica O:8 control serum against the Y enterocolitica O:8 antigen during assay development consistently resulted in blot staining detected at locations corresponding to approximately 12 to 15, 20, 28, 35, 48, 62, and 98 kDa molecular weight proteins. These developmental test results indicated that positive brucellosis bovine serum samples would have limited reaction to the Y enterocolitica O:8 antigen. In contrast, cross-reactions of either Yersinia O:8 or O:9 antibody would be expected against both Yersinia antigens. Inclusion of the O:8 antigen in our study allowed us an additional component to decipher the level and characteristics of possible cross reactions if Brucella antibody was present. The high level of reactivity observed against both Yersinia antigens supported the presence of high level of Yersinia antibody. Additional supporting information for the lack of Brucella antibody was indicated by the lack of strong reactivity in the mid-molecular weight range (28-49 kDa) against the B abortus strain 2308 antigen as is observed with the use of positive control serum. In our study, an additional higher molecular weight band (approximately 98 kDa) was observed with the NVSL B abortus strain 2308 antigen not previously reported with immunoblot procedures.

Use of the B abortus strain RB51 antigen in this study provided information related to possible antibody reactions against core Brucella proteins. Results indicated specific differences between reactions of Yersinia antibody reacting at approximately 38 kDa versus Brucella antibody which indicated reactivity with a protein band at approximately 49 kDa. This variance may provide additional support in the future for differentiating Yersinia from Brucella immunological reactions in these situations.

One difficulty associated with use of western blot is unknown antibody titers that may be present in field samples. During antigen standardization trials this variable antibody titer of field samples continued to result in difficulties establishing antigen concentrations that would provide clear blot results and yet ensure adequate sensitivity. Decreasing protein concentrations of the Brucella antigens allowed better delineation of banding patterns from Yersinia-positive samples, but still results in variable smearing. Initially, higher concentrations of the Yersinia antigens proved useful for low titer brucellosis serum samples but does present continued difficulties when encountering Yersinia field samples containing high antibody titers. This may result in having to repeat immunoblot testing if serum samples were over- or under-diluted during initial testing and may add time onto the testing period. As further work proceeds with immunoblot procedures it may be possible to determine an initial serum dilution based upon a correlation with brucellosis serological results.

The amount of additional diagnostics performed in this investigation was extensive since the implications for the company and the state pork industry would have been immeasurable if the herd would have truly been infected with swine brucellosis. Therefore, the efforts were necessary to rule out swine brucellosis infection and to prevent unnecessary depopulation of the herd.


  • Due to imperfect specificity, other diagnostics were used to rule out B suis infection.
  • A joint effort was needed to determine herd status and relieve the burden of quarantine.
  • Several diagnostic tools helped confirm FPSR for B suis and remove the herd quarantine.


We thank the North Carolina Department of Agriculture and Consumer Services Veterinary Diagnostic Laboratory System for performing diagnostic necropsies, tissue collection, bacteriology and serology to assist in this investigation. We also thank Christine Quance at NVSL for performing the Brucella cultures for this investigation. Finally, we thank Suelee Robbe-Austerman for reviewing the manuscript.

Conflict of interest

None reported.


Scientific manuscripts published in the Journal of Swine Health and Production are peer reviewed. However, information on medications, feed, and management techniques may be specific to the research or commercial situation presented in the manuscript. It is the responsibility of the reader to use information responsibly and in accordance with the rules and regulations governing research or the practice of veterinary medicine in their country or region.


*1. Animal and Plant Health Inspection Service. Brucellosis in swine; Add Texas to list of validated brucellosis-free states. Fed Reg. 2011;76:97.

2. Corn J, Cumbee JC, Barfoot R, Erickson GA. Pathogen exposure in feral swine populations geographically associated with high densities of transitional swine premises and commercial swine production. J Wildl Dis. 2009;43:713-721.

3. Gresham C, Gresham CA, Duffy MJ, Faulkner CT, Patton S. Increased prevalence of Brucella suis and Pseudorabies virus antibodies in adults of an isolated feral swine population in coastal South Carolina. J Wildl Dis. 2002;38:653-656.

4. Glynn MK, Lynn TV. Brucellosis. J Am Vet Med Assoc. 2008;233(6):900-908.

5. Jungersen G, Sørensen V, Giese S, Stack J, Riber U. Differentiation between serological responses to Brucella suis and Yersinia enterocolitica serotype O:9 after natural or experimental infection in pigs. Epidemiol Infect. 2006;134:347-357.

6. Lindberg AA, Haeggman S, Karlson K, Carlsson HE, Mair N. Enzyme immunoassay of the antibody response to Brucella and Yersinia enterocolitica 09 infections in humans. J Hyg. 1982;88:295-307.

7. Weynants V, Tibor A, Denoel PA, Saegerman C, Godfroid J, Thiange P, Letesson JJ. Infection of cattle with Yersinia enterocolitica O:9 a cause of the false positive serological reactions in bovine brucellosis diagnostic tests. Vet Microbiol. 1996;48:101-112.

8. Nielsen K, Smith P, Yu WL, Elmgren C, Halbert G, Nicoletti P, Perez B, Conde S, Samartino L, Nicola A, Bermudez R, Renteria T. Validation of a second generation competitive enzyme immunoassay (cELISA) for the diagnosis of brucellosis in various species of domestic animals. Vet Immunol Immunopathol. 2008;125:246-250.

9. Nielsen K, Smith P, Yu W, Nicoletti P, Jungersen G, Stack J, Godfroid J. Serological discrimination by direct enzyme immunoassay between the antibody response to Brucella sp. and Yersinia enterocolitica O:9 in cattle and pigs. Vet Immunol Immunopathol. 2006;109:69-78.

10. Ragan V, Vroegindewey G, Babcock S. International standards for brucellosis prevention and management. Rev Sci Tech. 2013;32:189-198.

11. Dieste-Pérez L, Blasco JM, de Miguel MJ, Marin CM, Barberán M, Conde-Alvarez R, Moriyón I, Muñoz PM. Performance of skin tests with allergens from B melitensis B115 and rough B abortus mutants for diagnosing swine brucellosis. Vet Microbiol. 2014;168:161-168.

12. Dieste-Perez L, Blasco JM, de Miguel MJ, Moriyón I, Muñoz PM. Diagnostic performance of serological tests for swine brucellosis in the presence of false positive serological reactions. J Microbiol Methods. 2015;111:57-63.

*13. Alton GG, Jones LM, Angus RD, Verger JM. Techniques for the brucellosis laboratory. Paris, France: INRA Publications; 1988:13-61.

14. Atlas, R. Handbook of microbiological media. 4th ed. Boca Raton, FL: CRC Press; 2010:1365.

15. Garzetti D, Susen R, Fruth A, Tietze E, Heesemann J, Rakin A. A molecular scheme for Yersinia enterocolitica patho-serotyping derived from genome-wide analysis. Int J Med Microbiol. 2014;304:275-283.

16. Jamali H, Paydar M, Radmehr B, Ismail S. Prevalence, characterization, and antimicrobial resistance of Yersinia species and Yersinia enterocolitica isolated from raw milk in farm bulk tanks. J Dairy Sci. 2015;98:798-803.

17. Jayarao BM, Donaldson SC, Straley BA, Sawant AA, Hegde NV, Brown JL. A survey of foodborne pathogens in bulk tank milk and raw milk consumption among farm families in Pennsylvania. J Dairy Sci. 2006;89:2451-2458.

18. Van Damme I, Berkvens D, De Zutter L. Effect of sampling and short isolation methodologies on the recovery of human pathogenic Yersinia enterocolitica from pig tonsils. Foodborne Pathog Dis. 2012;9:600-606.

19. Van Damme I, Berkvens D, Baré J, De Zutter L. Influence of isolation methods on the occurrence of plasmid-carrying Yersinia enterocolitica serotype O:3 in slaughter pig tonsils, faeces and carcass surface swabs. Int J Food Microbiol. 2013;164:32-35.

20. Chenais E, Bagge E, Lambertz ST, Artursson K. Yersinia enterocolitica serotype O:9 cultured from Swedish sheep showing serologically false-positive reactions for Brucella melitensis. Infect Ecol Epidemiol. 2012;2:1-7.

21. O’Grady D, Kenny K, Power S, Egan J, Ryan F. Detection of Yersinia enterocolitica serotype O:9 in the faeces of cattle with false positive reactions in serological tests for brucellosis in Ireland. Vet J. 2016;216:133-135.

22. Riber U, Jungersen G. Cell-mediated immune responses differentiate infections with Brucella suis from Yersinia enterocolitica serotype O:9 in pigs. Vet Immunol Immunopathol. 2007;116:13-25.

23. Edmonds MD, Ward FM, O’Hara TM, Elzer PH. Use of western immunoblot analysis for testing moose serum for Brucella suis biovar 4 specific antibodies. J Wildl Dis. 1999;35:591-595.

24. Schumaker BA, Mazet JA, Gonzales BJ, Elzer PH, Hietala SK, Ziccardi MH. Evaluation of the western immunoblot as a detection method for Brucella abortus exposure in elk. J Wildl Dis. 2010;46:87-94.

25. Gu WP, Wang X, Qiu HY, Luo X, Xiao YC, Tang LY, Kan B, Xu JG, Jing HQ. Comparison of lipopolysaccharide and protein immunogens from pathogenic Yersinia enterocolitica bio-serotype 1B/O:8 and 2/O:9 using SDS-PAGE. Biomed Environ Sci. 2012;25(3):282-290.

* Non-refereed references.

RAF, BJM: US Department of Agriculture, Animal and Plant Health Inspection Service, Veterinary Services, Raleigh, North Carolina.

ML: Smithfield Foods, Inc, Rose Hill, North Carolina.

RC, LC, JH, KL: US Department of Agriculture, Animal and Plant Health Inspection Service, National Veterinary Services Laboratories, Ames, Iowa.

MJN, JK: North Carolina Department of Agriculture and Consumer Services Veterinary Division, Raleigh, North Carolina.

Corresponding author: Dr Ross Free, USDA APHIS Veterinary Services, 920 Main Campus Drive Suite 200, Raleigh, North Carolina, 27606; Tel: 919-855-7712; Email:

Free RA, Ladd M, Capsel R, Cox L, Hicks J, Lantz K, Neault MJ, Kittrell J, Meade BJ. Diagnosis of Yersinia enterocolitica serovar O:9 in a commercial 2400-sow farm with false-positive Brucella suis serology using western blot, competitive ELISA, bacterial isolation, and whole genome sequencing. J Swine Health Prod. 2020;28(1):21-30.

Portable Document Format PDF version

RIS citation RIS citation (download for bibliography management software)