- Email: [email protected]
Detection, isolation and enumeration of Yersinia enterocolitica from raw pork
Available online at www.sciencedirect.com
International Journal of Food Microbiology 123 (2008) 25 – 31 www.elsevier.com/locate/ijfoodmicro
Detection, isolation and enumeration of Yersinia enterocolitica from raw pork J.A. Hudson ⁎, N.J. King, A.J. Cornelius, T. Bigwood, K. Thom, S. Monson Food Safety Programme, Institute of Environmental Science and Research (ESR) Limited, PO Box 29-181, Christchurch 8540, New Zealand Received 11 June 2007; received in revised form 26 November 2007; accepted 27 November 2007
Abstract The methods available for the isolation of Yersinia enterocolitica from foods are generally considered to be less than optimal, and methods for estimation of numbers are lacking. Such methods are needed to understand better the significance of foodborne yersiniosis and to provide data for exposure assessment. We describe a method for the detection and enumeration of Y. enterocolitica containing the pYV virulence plasmid (YeP+) in samples from pork surfaces. The method uses a multiplex PCR targeting the ail and virF genes to detect Y. enterocolitica after incubation of surface swabs in Yersinia enrichment broth according to Ossmer. Enumeration was achieved by adapting the enrichment to a most probable number (MPN) method format. A presumptive result was available within 24 h of sample receipt, and YeP+ isolates were confirmed within four days. The presence/absence and MPN methods were evaluated in a pilot survey of 34 packs of raw pork meat purchased from retail outlets in Christchurch, New Zealand. YeP+ was detected by PCR on meat from 32% of the packs, and YeP+ isolates were obtained from 18% of the samples. YeP+ were present at numbers ranging from 0.30 to 5.42 MPN/cm2. This improved method for the detection and enumeration of YeP+ from meat samples can be used for microbiological surveys to obtain data for assessments of consumer exposure to virulent Y. enterocolitica, and in outbreak investigations. © 2007 Elsevier B.V. All rights reserved. Keywords: Yersinia enterocolitica; Detection method; Enumeration method; Survey; Pork; pYV plasmid
1. Introduction The genus Yersinia comprises 12 species, of which three are known to be human pathogens (Y. pestis, Y. pseudotuberculosis and Y. enterocolitica), while the role of other species (e.g. Y. frederiksenii and Y. kristensenii) in human disease is still under investigation (Sulakvelidze, 2000). Y. enterocolitica is divided into subgroups according to biochemical activities (biogroups) and O antigens (serotypes). The serotype most frequently implicated in human disease worldwide is O:3 and almost all are biogroup 4 (Weynants et al., 1996). Other biogroups associated with human infection include 1B (serotypes O:8 and O:4), biogroup 2 (O:9, O:5,27), biogroup 3 (O:5,27, O:1,2,3) and biogroup 5 (O:2,3). Isolates belonging to biogroup 1A are regarded as avirulent or ‘environmental’, although they may be ⁎ Corresponding author. Tel.: +64 3 351 6019; fax: +64 3 351 0010. E-mail address: [email protected] (J.A. Hudson). 0168-1605/$ - see front matter © 2007 Elsevier B.V. All rights reserved. doi:10.1016/j.ijfoodmicro.2007.11.069
opportunistic pathogens (Bhagat and Virdi, 2007). In New Zealand biogroup 4 strains account for over 90% of cases of yersiniosis (Fenwick and McCarthy, 1995; Wright, 1996). The current incidence of yersiniosis in New Zealand is 13.0 cases per 100,000 people, making it the third most frequently reported, possibly foodborne, enteric disease (ESR, 2007). Little is known about the epidemiology of foodborne yersiniosis in New Zealand, but a case control study identified pork consumption as the only food-related risk factor for yersiniosis (Satterthwaite et al., 1999). This observation is consistent with data from other countries where, for example, pork consumption was a risk factor for sporadic yersiniosis in a Norwegian case control study (Kapperud et al., 1995) and identified as the source of an outbreak of yersiniosis (Grahek-Ogden et al., 2007). Generally, detection methods involve cold enrichment for up to three weeks using non-selective and/or selective broths, followed by plating onto a selective agar (Fredriksson-Ahomaa et al., 1999). Cold enrichment may be replaced by more rapid
26
J.A. Hudson et al. / International Journal of Food Microbiology 123 (2008) 25–31
methods using selective media that exploit the resistance of Yersinia to certain antimicrobials such as irgasan (Toora et al., 1994). The organism's tolerance of high pH may also be used to inhibit growth of or inactivate competing organisms (Aulisio et al., 1980). These relatively rapid culture methods have the advantage of being better suited to outbreak investigations than procedures involving cold enrichment. Pathogenic isolates contain plasmid pYV which mediates phenotypic responses useful for identifying them (Foultier and Cornelis, 2003). In particular the low calcium response results in pinhead sized colonies on low calcium agar, and in addition the cells are able to take up the dye congo red. A low-calcium congo redcontaining agarose (CRBHO) medium has been formulated to distinguish these characteristics (Bhaduri et al., 1991). The presence of the plasmid also presents an opportunity for pathogen-specific genes to be detected by PCR so that pathogenic but not environmental strains are detected. However, the detection, isolation and enumeration of Y. enterocolitica remain problematic. Consequently there is a paucity of data on the prevalences of the organism in foods (Fredriksson-Ahomaa and Korkeala, 2003) and a lack of any quantitative data. The contemporary quantitative risk assessment (QRA) approach to managing food safety risks requires the input of adequate data quality and quantity, so the application of QRA to yersiniosis is currently not possible. Data on the occurrence of Y. enterocolitica in New Zealand foods are needed to identify the extent to which food contributes to yersiniosis, but we are aware of only one published survey of foods in New Zealand, in which Y. enterocolitica was isolated from 3.4% of 203 samples of ready-to-eat fleshfoods (Hudson et al., 1992). Only one of the seven isolates tested was of a human pathogenic type. Therefore, to facilitate quantitative surveys we sought to identify a reliable and sensitive PCR-based method for recovering and enumerating Y. enterocolitica from pork surfaces, and to test the method in a small pilot survey of raw retail pork. 2. Materials and methods 2.1. Bacterial strains and growth conditions Yersinia isolates were obtained from the New Zealand Reference Culture Collection, Medical Section (NZRM) (http://www.esr.cri.nz/competencies/communicabledisease/ nzrcc.htm). Serotype O:3 isolates with the prefix “Z” were also kindly supplied by Dr. Stan Fenwick of Massey University, Palmerston North, New Zealand (Table 1). Other bacteria used for PCR validation (Table 1) were those held by ESR's Christchurch Public Health Laboratory. Y. enterocolitica NZRM 3596 was used as a positive control for all pork analyses. Inocula were prepared by inoculating 2–3 colonies into 5 ml volumes of trypticase soy broth (TSB; Becton, Dickinson and Company, Sparks, USA) which were incubated at 24 °C for 18 h. These cultures were serially diluted in 0.1% peptone water and the numbers were determined from plate counts of colonies on trypticase soy agar (TSA; 15 g agar/L TSB) incubated for 24 h at 37 °C.
Table 1 Detection of the ail and virF genes by PCR in isolates of Yersinia enterocolitica, Yersinia spp. and other bacteria Isolate
Isolate
ail (356 bp) virF (231 bp)
Yersinia enterocolitica, serotype 1(2a,3) Y. enterocolitica, serotype 3 Y. enterocolitica, serotype 9 Y. enterocolitica, serotype O:8
NZRM 767
+
−
+ + −
− − −
+ +
+ +
+
−
+
+
+
−
+
+
+
+
+
+
+
+
+
+
+
+
+
+
− − + − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − −
− − − +? − − − − − − − − − − − − − − − − − − − − − − − − − − − − − −
NZRM 1000 NZRM 1001 NZRM 2603 (ATCC 9610)T Y. enterocolitica, serotype O:3 NZRM 3596 Y. enterocolitica, serotype O:3, z22 pig isolate Y. enterocolitica, serotype O:3, z26 pig isolate Y. enterocolitica, serotype O:3, z27 pig isolate Y. enterocolitica, serotype O:3, z28 pig isolate Y. enterocolitica, serotype O:3, z31 pig isolate Y. enterocolitica, serotype O:3, z32 pig isolate Y. enterocolitica, serotype O:3, z39 pig isolate Y. enterocolitica, serotype O:3, z52 pig isolate Y. enterocolitica, serotype O:3, z53 pig isolate Y. enterocolitica, serotype O:3, z57 pig isolate Y. enterocolitica, serotype O:3, z77 pig isolate Y. pseudotuberculosis NZRM 768 Y. frederiksenii NZRM 2534 Y. kristensenii NZRM 2535 Y. intermedia NZRM 2604 Arcobacter butzleri NZRM 4017 Bacillus cereus NZRM 5 B. subtilus NZRM 143 B. thuringiensis NZRM 3610 Brochothrix campestris NZRM 3569 B. thermosphacta NZRM 3320 Campylobacter coli NZRM 2607 C. fetus NZRM 2398 C. hyointestinalis NZRM 3676 C. jejuni NZRM 2397 C. jejuni subsp. doyley NZRM 3516 C. lari NZRM 2622 C. showae ATCC 51146 C. upsaliensis NZRM 3675 Carnobacterium divergens NZRM 3572 C. gallinarum NZRM 3575 C. mobile NZRM 3576 C. piscicola NZRM 3571 Clostridium perfringens NZRM 20 Enterococcus faecalis NZRM 1106 Enterobacter aerogenes NZRM 798 Escherichia coli NZRM 480 E. coli O157:H7 NZRM 3614 Klebsiella pneumoniae NZRM 482 Listeria grayi NZRM 1088 L. innocua NZRM 3024 L. ivanovii NZRM 797 L. monocytogenes NZRM 2597 L. seeligeri NZRM 3287 L. welshimeri NZRM 3286
J.A. Hudson et al. / International Journal of Food Microbiology 123 (2008) 25–31 Table 1 (continued) Isolate
Isolate
ail (356 bp) virF (231 bp)
Morganella morganii Proteus vulgaris Pseudomonas aeruginosa Saccharomyces cerevisiae Salmonella Menston S. typhi Shigella flexneri S. sonnei Staphylococcus aureus S. epidermidis Streptococcus bovis Vagococcus fluvialis Vibrio cholerae V. parahaemolyticus
NZRM 65 NZRM 67 NZRM 981 NZRM 1207 NZRM 383 NZRM 444 NZRM 3476 NZRM 86 NZRM 917 NZRM 1210 NZRM 2720 NZRM 3573 NZRM 776 NZRM 820
− − − − − − − − − − − − − −
− − +? − − − − − − − − − − −
NZRM, New Zealand Reference Culture Collection, Z22–Z77 supplied by Dr. Stan Fenwick (see Materials and methods). +? Weak PCR product observed.
2.2. Multiplex PCR conditions, validation and sensitivity Phenol/chloroform purified DNA from five Y. enterocolitica reference isolates and 11 isolates from pigs, four other Yersinia spp. and 44 other bacteria were examined for the presence of the chromosomal ail and plasmid-borne virF genes (Harnett et al., 1996), with a minor modification to remove the G from the 5′ end of the VirF-a primer. Each 50 µl reaction mixture contained 25 µl Qiagen (Hilden, Germany) Taq PCR mastermix (1.23 U Taq DNA Polymerase, 200 µM each of dATP, dCTP, dGTP and dTTP, 1.5 mM MgCl2, Qiagen PCR buffer), 0.2 µM of each primer, a varying volume of extracted DNA depending on source, with the balance being sterile water. The PCR was carried out under the following conditions; 35 cycles of 94 °C for 1 min (denaturation), 55 °C for 1 min (annealing) and 72 °C for 1 min (extension), followed by a final extension at 72 °C for 7 min. PCR products were subjected to electrophoresis (70 min, 110 V) using 2% agarose gels in Tris-boric acid-EDTA (TBE) buffer (USB Corporation, Cleveland, USA) containing 0.5 µg ethidium bromide ml−1, and visualised by ultraviolet transillumination. Products were compared to a 100 bp or 1 kb+ ladder (Invitrogen, Carlsbad, California, USA) and purified DNA from Y. enterocolitica NZRM 1000 (pYV-negative). To define the PCR detection limit, fresh broth cultures of Y. enterocolitica isolates NZRM 1000 (O:3 pYV-negative) and NZRM 3596 (O:3 pYV-positive) were centrifuged at 1600 × g for 20 min. The supernatants were discarded and the pellets were resuspended in 5 ml 0.85% NaCl. A mixture of the two cultures was also prepared by adding 1 ml from each into a third tube. The suspensions were serially diluted in 0.1% peptone water and 100 µl from each dilution was processed for PCR by heating for 12 min at 96 °C and then centrifuging for 12 min at 11,750 × g. The supernatant was immediately used for PCR. For this and in all subsequent work the 25 µl PCR reaction mixture contained 12.5 µl Qiagen taq PCR mastermix, 0.2 µM of each primer and 10 µl of the cell extraction. The thermocycling and visualisation conditions were as before. The number of cells added to each reaction mixture was calculated from colony counts on TSA. A cell-free control of uninoculated TSB was
27
processed with the cultures. Purified DNA from Y. enterocolitica NZRM 3596 (pYV-positive) was used as a positive PCR control and nuclease-free ultrapure water was used as a negative (sterility) PCR control. 2.3. Selection of enrichment medium for the isolation and identification of Y. enterocolitica The growth of 16 isolates of Y. enterocolitica serotypes O:1, O:3, O:8 and O:9, with or without the pYV plasmid (YeP+ or YeP−), was monitored by determination of the increase in optical density at 650 nm of cultures in modified tryptic soy broth (MTSB; (Bhaduri et al., 1997)) and Yersinia selective enrichment broth according to Ossmer (YSEO, catalogue No. 1.16701, Merck, Darmstadt, Germany). Comparisons were made of optical densities of triplicate samples incubated at 12, 24 and 30 °C over a period of 48 h, although the 30 °C incubation was discontinued at 24 h having reached a maximum at this time. 2.4. Sampling Retail portions of boneless pork, weighing approximately 400 g were purchased from four supermarkets and three butcher's shops in Christchurch, New Zealand. Portions were purchased from each supermarket on at least two occasions and the butcher's shops once between May and June 2006. Every portion differed with regard to cut type, location or packaging date. Of the 34 portions, 17 were used for presence/absence (P/A) testing and MPN determination, eight were used for P/A testing only and nine were used for MPN determination only. All portions were kept chilled and were processed within 24 h of being purchased. 2.5. Enrichment for P/A testing For P/A tests, a piece of meat was randomly selected from each portion and a segment measuring approximately 5 × 5 cm was aseptically excised from it. The entire upper surface of the segment was swabbed first with a cotton wool swab moistened with a solution of KOH:NaCl (0.25%:0.5% w/v), then with a dry swab. Both swabs were broken into a centrifuge tube containing 9 ml YSEO broth. The tube was shaken vigorously, and then incubated for 18 h at 24 °C. A positive control consisting of 9 ml YSEO broth inoculated with Y. enterocolitica NZRM 3596 was also prepared and incubated under the same conditions, along with a negative (sterility) control containing 9 ml uninoculated YSEO broth. Following enrichment, each tube was inverted 10 times to mix the contents, then the swabs were discarded. 2.6. Enrichment for MPN determinations Segments of meat were obtained and swabbed as for P/A tests. The swabs were broken off into a 50 ml tube containing 20 ml YSEO broth. The tube was vortexed vigorously for 1 min
28
J.A. Hudson et al. / International Journal of Food Microbiology 123 (2008) 25–31
then the volume was made up to 50 ml by the addition of YSEO broth. The tube was shaken vigorously by hand for 10 s, then the rinse fluid was dispensed into a 3 × 3 MPN system consisting of 10 ml of undiluted fluid in each of three 10 ml centrifuge tubes (level A), 1 ml of fluid in 9 ml YSEO broth in each of three 10 ml centrifuge tubes (a 1:10 dilution, level B), and 1 ml of a 1:10 dilution of the fluid in 9 ml YSEO broth in each of three 10 ml centrifuge tubes (a 1:100 dilution, level C). Controls were set up as before. All tubes incubated at 24 °C for 18 h. Segments of meat inoculated with various numbers of strain NZRM 3596 were prepared each time pork segments were sampled with 50 µl of each selected dilution being spread over the upper surface of a randomly selected pork segment. The inoculated segment was refrigerated for at least 40 min before it was sampled to allow attachment of the organism. 2.7. Multiplex PCR screening of pork enrichments A 1 ml volume of each rinse fluid was retained in a sterile tube under refrigeration for subsequent plating. The remaining (8 ml) of fluid was centrifuged at 1600 × g for 20 min, and the supernatant discarded. The pellet was resuspended in 5 ml sterile double-distilled water, and the suspension was centrifuged as before. The supernatant was discarded and the pellet was resuspended in 100 µl of sterile ultrapure water. A 10 µl volume of this suspension was briefly vortexed with 90 µl of sterile ultrapure water and prepared for PCR as described above. The presence or absence of the target PCR products (ail and/or vir-F) was recorded. Samples producing both products were scored as positive for YeP+. All P/A enrichments were screened by PCR before plating. For MPN enrichments, all of the level A tubes were analysed by PCR and tubes from levels A, B and C were tested by PCR for a sub-set of 16 samples. 2.8. Isolation and identification of YeP+ from pork enrichments by culture Each enrichment medium was assessed visually for turbidity and serially diluted accordingly, with a highly turbid samples being diluted to 10− 4, and non-turbid samples being diluted to 10− 2, in buffered peptone water (Merck, Darmstadt, Germany). The exception was the first dilution step where 0.5 ml of enrichment medium was added to 4.5 ml KOH/NaCl solution. After mixing the mixture was further diluted, or 20 µl was spread on CIN agar plates which were incubated for 48 h at 24 °C. Suspect colonies, at least three per enrichment where possible, were each subcultured to plates of TSA and Congo Red Brain Heart Infusion Agarose (CRBHO) (Bhaduri et al., 1991), and spotted to Rhamnose Arabitol McConkey agar (RAM) (Shehee and Sobsey, 2004) using a toothpick. CRBHO plates were incubated at 37 °C, and TSA and RAM plates at 24 °C, all for 18 h. Colonies typical of YeP+ strains on CRBHO are pin head sized and dark orange in colour and give white colonies on RAM. All colonies with these characteristics were selected for confirmation. RAM-positive/CRBHO-negative colonies (presumptively YeP−) were also selected for confirmation.
Isolates from TSA were tested for the production of urease, catalase and oxidase, and for citrate metabolism. Isolates positive for the urease and catalase tests, and negative for the oxidase and citrate tests were confirmed as YeP+ by multiplex PCR. To prepare the PCR template, a colony from a TSA plate was suspended in 100 µl of ultrapure water and processed as described above. The PCR conditions were as described for the 25 µl reactions above, except that only 1 µl of template was used and ultrapure water was added to make the final volume 25 µl. 2.9. Calculation of MPN A spreadsheet was developed using Microsoft Excel to calculate MPN values (Peeler et al., 1992). The spreadsheet used the number of tubes containing YeP+ and values for the surface area of sample tested in each MPN tube. The range over which this 9-tube MPN system operates is between 0.06 MPN/cm2 (LCI of 0.01 with one positive tube at level C; though one positive tube at level A is a more likely result and has a value of 0.07 MPN/cm2) and 8.77 MPN/cm2 (UCI of 39.74; all tubes but one positive at level C). If all tubes were negative the MPN was recorded as b 0.06 MPN/cm2, and if all were positive, the MPN was recorded as N 8.77 MPN/cm2. 3. Results and discussion 3.1. PCR validation and sensitivity All of the Y. enterocolitica isolates tested by the multiplex PCR contained ail, with the exception of NZRM 2603 (ATCC 9610T), a pYV-negative isolate of serotype O:8 isolated from an extra-intestinal infection (Table 1) which produced two PCR products at approximately 800 and 1100 bp. This isolate has been reported to be non-enteropathogenic (Neubauer et al., 2000) and not to contain the ail gene (Thoerner et al., 2003). The virF gene was not detected in every Y. enterocolitica isolate, possibly because they were either non-pathogenic or had lost the pYV plasmid on repeated laboratory culture. Of the other Yersinia species, the ail gene was detected in Y. kristensenii and it appeared that the virF gene was present in Y. intermedia, although subsequent comparison of the virF product from Y. intermedia directly against a control confirmed that the PCR products were of different sizes. Of the non-Y. enterocolitica species tested by Harnett et al. (1996), only Y. pseudotuberculosis produced an amplification product, and only for the virF gene. This was not reproduced for our Y. pseudotuberculosis isolate. We are not aware of any reports of the ail gene being detected in Yersinia species other than Y. enterocolitica (Miller et al., 1989) although similar sequences might be present, at least in Y. intermedia (Robins-Browne et al., 1989). Some of the other bacteria produced weak PCR products, but never of the correct size for both genes. The PCR was therefore specific for pYV-positive Y. enterocolitica (YeP+). The data presented here complement those for 129 pathogenic Y. enterocolitica, 71 non-pathogenic yersiniae and 100 non-Yersinia isolates analysed by Harnett et al. (1996).
J.A. Hudson et al. / International Journal of Food Microbiology 123 (2008) 25–31
29
Fig. 1. Detection limit of the multiplex PCR for a pYV-positive and a pYV-negative Y. enterocolitica, and a 1:1 combination 1.57 × 108 cfu Serial 1:10 dilutions Y. enterocolitica NZRM 1000 (O:3, pYV-negative) 1.44 × 108 cfu Serial 1:10 dilutions Mixture of NZRM 3596 and 1000 (1:1) 9.35 × 107 cfu Serial 1:10 dilutions 1 kb+ DNA ladder (lanes 1 and 32); TSB negative control (lane 29); PCR-positive control (lane 30); PCR-negative control (lane 31)
Lane 2 Lanes 3–10 Lane 11 Lanes 12–19 Lane 20 Lanes 21–28
Y. enterocolitica NZRM 3596 (O:3, pYV-positive)
The detection limit of the PCR was analysed with pYVpositive and pYV-negative Y. enterocolitica reference cultures, alone and in combination (Fig. 1). Under the conditions used, detection occurred at 104 to 105 cfu/ml (103 to 104 cells per 25 µl reaction). This is not a very sensitive method compared to some others (Thisted Lambertz et al., 2007), but is adequate for testing enriched samples. 3.2. Media for the isolation and identification of Y. enterocolitica After 24 h incubation all Y. enterocolitica isolates were approaching or had reached stationary phase in MTSB and YSEO broths at all temperatures. When incubation at 12 °C or 24 °C was extended beyond 24 h, the optical densities changed little. While MTSB yielded higher optical densities than YSEO broth the range of values was greater with MTSB. YSEO broth was selected for further experiments because of the advantage gained from the presence of antibiotics and because of its ability to enrich Y. enterocolitica isolates in a more uniform manner, i.e. not preferentially selecting one serotype. The differences in maximum optical densities reached indicated that the numbers were only 2–3 fold different between the two media.
In other reports 24% of 51 Norwegian pork samples were positive for pathogenic Y. enterocolitica by PCR, but none were positive by culture (Johannessen et al., 2000). Similarly, 25% of 255 minced pork samples from Finland were PCR-positive for pathogenic Y. enterocolitica, but only 2% were positive by culture (Fredriksson-Ahomaa et al., 1999). Lower frequencies have been reported, including 10% of 91, and 11% of 97 Swedish samples positive for pathogenic Y. enterocolitica (Thisted Lambertz and Danielsson-Tham, 2005; Thisted Lambertz et al., 2007). The data obtained from our small survey were therefore comparable to those from PCR studies conducted elsewhere, but the proportion of samples positive by culture was higher. This may be due to a real higher prevalence or an improved ability of the method to detect the viable organism which would, presumably, have resulted from the
Table 2 Most probable number results for all samples positive for pYV-positive Y. enterocolitica by PCR and/or culturing Product
Days old a
Most probable number (MPN/cm2) PCR MPN
3.3. Raw pork survey When the data for all the methods were considered 11 (32%) of the portions of pork were PCR-positive and YeP+ was isolated from six (18%). The prevalence by PCR was 37.5%, 44.4% and 23.5% by presence/absence testing (8 samples), MPN determination (9 samples) and both methods (17 samples) respectively, while by culture the equivalent prevalences were 12.5%, 33.3% and 11.8%.
Schnitzel Steak Schnitzel Chop Steak Schnitzel Schnitzel a
1 b1 1 4 1 3 3
b
+ + + + 0.30 1.52 5.42
Culture UCI
LCI
MPN
UCI
LCI
− − − − 1.34 6.89 24.59
− − − − 0.07 0.34 1.20
0.07 0.41 0.19 b0.06 0.06 0.06 b0.06
0.33 1.87 0.86 − 0.28 0.28 −
0.02 0.09 0.04 − 0.01 0.01 −
Number of days between packaging date and date sampled. +, Y. enterocolitica present but an MPN could not be calculated as only the level A enrichment tubes were tested by PCR. b
30
J.A. Hudson et al. / International Journal of Food Microbiology 123 (2008) 25–31
Table 3 Recovery of pYV-positive Y. enterocolitica in inoculated samples tested by the MPN method MPN/cm2
Product PCR
Steak Schnitzel Schnitzel Schnitzel Steak Schnitzel Schnitzel Steak a
Inoculum (cfu/cm2)
Culture
MPN
UCI
LCI
MPN
UCI
LCI
N8.77 +a + + 1.90 0.57 0.57 1.17
− − − − 8.60 2.59 2.59 5.30
− − − − 0.42 0.13 0.13 0.26
N8.77 N8.77 3.80 0.12 1.90 0.57 0.57 1.17
− − 17.21 0.56 8.60 2.59 2.59 5.30
− − 0.84 0.03 0.42 0.13 0.13 0.26
144.0 22.4 21.9 15.1 10.7 3.0 1.7 1.4
+, Y. enterocolitica present but an MPN could be calculated as only the level A enrichment tubes were tested by PCR.
YSEO enrichment as CIN agar was used commonly in the other studies. Of the seven samples that were PCR-positive by MPN testing, MPN values based on PCR could be derived for only three as, for some samples, only the level A MPN tubes were tested by PCR in an initial screen. For the three PCR-positive samples where an MPN value could be calculated, the concentrations of YeP+ in the three samples were 0.30, 1.52 and 5.42 MPN/cm2 (Table 2). Where comparisons were possible, the MPN values obtained through culture were lower than those calculated from the PCR results (Table 2). When it was not possible to isolate YeP+ from PCR-positive MPN tubes the MPN value was reduced. The plating method uses an alkaline treatment and antibiotics to reduce or inhibit non-Yersinia microflora, but this does not entirely remove background microflora. Identification of a wellisolated Yersinia colony among background microflora on CIN agar plates was often difficult, particularly when older meat was analysed. Since isolation of YeP+ from every PCR-positive MPN tube was challenging, we suggest that the isolation of YeP+ from any PCR-positive MPN tube provides evidence to confirm all PCRpositive results for that sample. False-positive PCR results from dead cells are unlikely as the PCR is conducted on an enrichment and has a relatively poor detection limit. However, when a sample is PCR-positive by MPN, but YeP+ are not isolated, the MPN result should be considered presumptive only. YeP+ were detected by the presence/absence method in all six inoculated samples over the range 1.4 to 17.3 cfu/cm2, though one sample inoculated with 1.7 cfu/cm2 was PCRnegative. No naturally-occurring YeP+ were detected in any of the meat used to prepare the inoculated samples. Of the portions of meat used for inoculated controls in the MPN analyses, two contained naturally-occurring YeP+ probably at numbers less than that of the inoculum. All eight inoculated samples were positive by PCR and isolates were recovered by plating (Table 3). For five of these samples all MPN tubes were analysed by PCR and the MPN values were identical by PCR and plating. YeP+ was detected in inoculated samples over the range 1.4 to 144 cfu/cm2. However, the MPN confidence intervals only encompassed the inoculum in four samples, and two exceeded the computable concentration of this
9-tube MPN system (N8.77 cfu/cm2). This indicates that recovery of YeP+ was good, but that the MPN values only approximated the inoculum levels. Disparities between the known inoculum and the calculated MPN varied, with most in the order of 3–10 fold. These differences might be due to losses through inefficient removal of the organism from the meat or from the filtration and centrifugation steps required for sample preparation. Cotton wool tipped swabs were used because of the small area sampled and the need not to use excision samples should the method be extended to carcass testing. Overall the method successfully detects YeP+ on meat samples when they are present at numbers as low as 2 cfu/cm2. Within 24 h of sample receipt a presumptive PCR result is obtainable. Any suspected Yersinia can be isolated within three days, and YeP+ can be confirmed 24 h later by specific biochemical tests and PCR. This method for the detection and enumeration of YeP+ will allow surveys of meat that can provide data for exposure assessment purposes. The method might also be used in outbreak investigations, though assessment of the method with other food types is required. The method could be improved by better sample preparation to recover more Y. enterocolitica cells, and better techniques are likely to be needed for DNA recovery from samples containing PCR inhibitors. In addition, CIN agar does not distinguish Y. enterocolitica clearly from other bacteria, and the need for a better plating medium remains. Acknowledgements We are grateful to the New Zealand Ministry of Health and New Zealand Food Safety Authority for financial support, and Dr. Stan Fenwick for supplying pig isolates. Thanks are also due to Jane Capill for her capable technical assistance, and to our molecular biology colleagues for their help and forbearance. This paper is dedicated to Surinah Monson who died before her time during the course of this work. References Aulisio, C.C.G., Mehlman, I.J., Sanders, A.C., 1980. Alkali method for rapid recovery of Yersinia enterocolitica and Yersinia pseudotuberculosis from foods. Applied and Environmental Microbiology 39, 135–140.
J.A. Hudson et al. / International Journal of Food Microbiology 123 (2008) 25–31 Bhaduri, S., Turner-Jones, C., Lachica, R.V., 1991. Convenient agarose medium for simultaneous determination of the low-calcium response and congo red binding by virulent strains of Yersinia enterocolitica. Journal of Clinical Microbiology 29, 2341–2344. Bhaduri, S., Cottrell, C., Pickard, A.R., 1997. Use of a single procedure for selective enrichment, isolation, and identification of plasmid-bearing virulent Yersinia enterocolitica of various serotypes from pork samples. Applied and Environmental Microbiology 63, 1657–1660. Bhagat, N., Virdi, J.S., 2007. Distribution of virulence-associated genes in Yersinia enterocolitica biovar 1A correlates with clonal groups and not the source of isolation. FEMS Microbiology Letters 266, 177–183. ESR, 2007. Yersiniosis. New Zealand Public Health Report 5, 4. Fenwick, S.G., McCarthy, M.D., 1995. Yersinia enterocolitica is a common cause of gastroenteritis in Auckland. New Zealand Medical Journal 108, 269–271. Foultier, B., Cornelis, G.R., 2003. DNA sequence and analysis of the pYVa127/90 virulence plasmid of Yersinia enterocolitica strain A127/90. Research in Microbiology 154, 553–557. Fredriksson-Ahomaa, M., Korkeala, H., 2003. Low occurrence of pathogenic Yersinia enterocolitica in clinical, food, and environmental samples: a methodological problem. Clinical Microbiology Reviews 16, 220–229. Fredriksson-Ahomaa, M., Hielm, S., Korkeala, H., 1999. High prevalence of yadA-positive Yersinia enterocolitica in pig tongues and minced meat at the retail level in Finland. Journal of Food Protection 62, 123–127. Grahek-Ogden, D., Schimmer, B., Cudjoe, K.S., Nygård, K., Kapperud, G., 2007. Outbreak of Yersinia enterocolitica serogroup O:9 infection and processed pork, Norway. Emerging Infectious Diseases 13, 754–756. Harnett, H., Lin, Y.P., Krishnan, C., 1996. Detection of pathogenic Yersinia enterocolitica using the multiplex polymerase chain reaction. Epidemiology and Infection 117, 59–67. Hudson, J.A., Mott, S.J., Delacy, K.M., Edridge, A.L., 1992. Incidence and coincidence of Listeria spp., motile aeromonads and Y. enterocolitica on ready-to-eat fleshfoods. International Journal of Food Microbiology 16, 99–108. Johannessen, G.S., Kapperud, G., Kruse, H., 2000. Occurrence of pathogenic Yersinia enterocolitica in Norwegian pork products determined by a PCR method and traditional culturing method. International Journal of Food Microbiology 54, 75–80. Kapperud, G., Ostroff, S.M., Nesbakken, T., Hutwagner, L.C., Bean, N.H., Lassen, J., Tauxe, R.V., 1995. Risk factors for sporadic Yersinia enterocolitica infections in Norway: a case-control study. Contributions in Microbiology and Immunology 12, 25–28. Miller, V.L., Farmer, J.J., Hill, W.E., Falkow, S., 1989. The ail locus is found uniquely in Yersinia enterocolitica serotypes commonly associated with disease. Infection and Immunity 57, 121–131.
31
Neubauer, H., Aleksic, S., Hensel, A., Finke, E.J., Meyer, H., 2000. Yersinia enterocolitica 16 S rRNA gene types belong to the same genospecies but from three homology groups. International Journal of Medical Microbiology 290, 61–64. Peeler, J.T., Houghtby, G.A., Rainosek, A.P., 1992. The Most Probable Number Technique. Compendium of Methods for the Microbiological Examination of Foods. 105–120. American Public Health Association, Washington, D.C. Robins-Browne, R.M., Miliotis, M., Cianciosi, S., Miller, V., Falkow, S., Morris, J., 1989. Evaluation of DNA colony hybridization and other techniques for the detection of virulence in other Yersinia species. Journal of Clinical Microbiology 27, 644–650. Satterthwaite, P., Pritchard, K., Floyd, D., Law, B., 1999. A case control study of Yersinia enterocolitica infections in Auckland. Australian and New Zealand Journal of Public Health 23, 482–485. Shehee, M.W., Sobsey, M.D., 2004. Development of a L-rhamnose and D-arabitol supplemented McConkey agar to identify pathogenic Yersinia enterocolitica among environmental yersinias in swine production wastes. Journal of Microbiological Methods 57, 289–292. Sulakvelidze, A., 2000. Yersiniae other than Y. enterocolitica, Y. pseudotuberculosis and Y. pestis: the ignored species. Microbes and Infection 2, 497–513. Thisted Lambertz, S., Danielsson-Tham, M.L., 2005. Identification and characterization of pathogenic Yersinia enterocolitica isolates by PCR and pulsed-field gel electrophoresis. Applied and Environmental Microbiology 71, 3674–3681. Thisted Lambertz, S., Granath, K., Fredriksson-Ahomaa, M., Johansson, K.E., Danielsson-Tham, M.L., 2007. Evaluation of a combined culture and PCR method (NMKL-163A) for detection of presumptive pathogenic Yersinia enterocolitica in pork products. Journal of Food Protection 70, 335–340. Thoerner, P., Bin Kingombe, C.I., Bögli-Stuber, K., Boissig-Choisat, B., Wassenaar, T.M., Jemmi, T., 2003. PCR detection of virulence genes in Yersinia enterocolitica and Yersinia pseudotuberculosis and investigation of virulence gene distribution. Applied and Environmental Microbiology 69, 1810–1816. Toora, S., Budu-Amoako, E., Ablett, R.F., Smith, J., 1994. Evaluation of different antimicrobial agents used as selective supplements for isolation and recovery of Yersinia enterocolitica. Journal of Applied Bacteriology 77, 67–72. Weynants, V., Jadot, V., Denoel, P.E., Tibor, A., Letesson, J.J., 1996. Detection of Yersinia enterocolitica serogroup O:3 by a PCR method. Journal of Clinical Microbiology 34, 1224–1227. Wright, J., 1996. Gastrointestinal illness: what are the causative organisms? The New Zealand Public Health Report 3, 9–11.
International Journal of Food Microbiology 123 (2008) 25 – 31 www.elsevier.com/locate/ijfoodmicro
Detection, isolation and enumeration of Yersinia enterocolitica from raw pork J.A. Hudson ⁎, N.J. King, A.J. Cornelius, T. Bigwood, K. Thom, S. Monson Food Safety Programme, Institute of Environmental Science and Research (ESR) Limited, PO Box 29-181, Christchurch 8540, New Zealand Received 11 June 2007; received in revised form 26 November 2007; accepted 27 November 2007
Abstract The methods available for the isolation of Yersinia enterocolitica from foods are generally considered to be less than optimal, and methods for estimation of numbers are lacking. Such methods are needed to understand better the significance of foodborne yersiniosis and to provide data for exposure assessment. We describe a method for the detection and enumeration of Y. enterocolitica containing the pYV virulence plasmid (YeP+) in samples from pork surfaces. The method uses a multiplex PCR targeting the ail and virF genes to detect Y. enterocolitica after incubation of surface swabs in Yersinia enrichment broth according to Ossmer. Enumeration was achieved by adapting the enrichment to a most probable number (MPN) method format. A presumptive result was available within 24 h of sample receipt, and YeP+ isolates were confirmed within four days. The presence/absence and MPN methods were evaluated in a pilot survey of 34 packs of raw pork meat purchased from retail outlets in Christchurch, New Zealand. YeP+ was detected by PCR on meat from 32% of the packs, and YeP+ isolates were obtained from 18% of the samples. YeP+ were present at numbers ranging from 0.30 to 5.42 MPN/cm2. This improved method for the detection and enumeration of YeP+ from meat samples can be used for microbiological surveys to obtain data for assessments of consumer exposure to virulent Y. enterocolitica, and in outbreak investigations. © 2007 Elsevier B.V. All rights reserved. Keywords: Yersinia enterocolitica; Detection method; Enumeration method; Survey; Pork; pYV plasmid
1. Introduction The genus Yersinia comprises 12 species, of which three are known to be human pathogens (Y. pestis, Y. pseudotuberculosis and Y. enterocolitica), while the role of other species (e.g. Y. frederiksenii and Y. kristensenii) in human disease is still under investigation (Sulakvelidze, 2000). Y. enterocolitica is divided into subgroups according to biochemical activities (biogroups) and O antigens (serotypes). The serotype most frequently implicated in human disease worldwide is O:3 and almost all are biogroup 4 (Weynants et al., 1996). Other biogroups associated with human infection include 1B (serotypes O:8 and O:4), biogroup 2 (O:9, O:5,27), biogroup 3 (O:5,27, O:1,2,3) and biogroup 5 (O:2,3). Isolates belonging to biogroup 1A are regarded as avirulent or ‘environmental’, although they may be ⁎ Corresponding author. Tel.: +64 3 351 6019; fax: +64 3 351 0010. E-mail address: [email protected] (J.A. Hudson). 0168-1605/$ - see front matter © 2007 Elsevier B.V. All rights reserved. doi:10.1016/j.ijfoodmicro.2007.11.069
opportunistic pathogens (Bhagat and Virdi, 2007). In New Zealand biogroup 4 strains account for over 90% of cases of yersiniosis (Fenwick and McCarthy, 1995; Wright, 1996). The current incidence of yersiniosis in New Zealand is 13.0 cases per 100,000 people, making it the third most frequently reported, possibly foodborne, enteric disease (ESR, 2007). Little is known about the epidemiology of foodborne yersiniosis in New Zealand, but a case control study identified pork consumption as the only food-related risk factor for yersiniosis (Satterthwaite et al., 1999). This observation is consistent with data from other countries where, for example, pork consumption was a risk factor for sporadic yersiniosis in a Norwegian case control study (Kapperud et al., 1995) and identified as the source of an outbreak of yersiniosis (Grahek-Ogden et al., 2007). Generally, detection methods involve cold enrichment for up to three weeks using non-selective and/or selective broths, followed by plating onto a selective agar (Fredriksson-Ahomaa et al., 1999). Cold enrichment may be replaced by more rapid
26
J.A. Hudson et al. / International Journal of Food Microbiology 123 (2008) 25–31
methods using selective media that exploit the resistance of Yersinia to certain antimicrobials such as irgasan (Toora et al., 1994). The organism's tolerance of high pH may also be used to inhibit growth of or inactivate competing organisms (Aulisio et al., 1980). These relatively rapid culture methods have the advantage of being better suited to outbreak investigations than procedures involving cold enrichment. Pathogenic isolates contain plasmid pYV which mediates phenotypic responses useful for identifying them (Foultier and Cornelis, 2003). In particular the low calcium response results in pinhead sized colonies on low calcium agar, and in addition the cells are able to take up the dye congo red. A low-calcium congo redcontaining agarose (CRBHO) medium has been formulated to distinguish these characteristics (Bhaduri et al., 1991). The presence of the plasmid also presents an opportunity for pathogen-specific genes to be detected by PCR so that pathogenic but not environmental strains are detected. However, the detection, isolation and enumeration of Y. enterocolitica remain problematic. Consequently there is a paucity of data on the prevalences of the organism in foods (Fredriksson-Ahomaa and Korkeala, 2003) and a lack of any quantitative data. The contemporary quantitative risk assessment (QRA) approach to managing food safety risks requires the input of adequate data quality and quantity, so the application of QRA to yersiniosis is currently not possible. Data on the occurrence of Y. enterocolitica in New Zealand foods are needed to identify the extent to which food contributes to yersiniosis, but we are aware of only one published survey of foods in New Zealand, in which Y. enterocolitica was isolated from 3.4% of 203 samples of ready-to-eat fleshfoods (Hudson et al., 1992). Only one of the seven isolates tested was of a human pathogenic type. Therefore, to facilitate quantitative surveys we sought to identify a reliable and sensitive PCR-based method for recovering and enumerating Y. enterocolitica from pork surfaces, and to test the method in a small pilot survey of raw retail pork. 2. Materials and methods 2.1. Bacterial strains and growth conditions Yersinia isolates were obtained from the New Zealand Reference Culture Collection, Medical Section (NZRM) (http://www.esr.cri.nz/competencies/communicabledisease/ nzrcc.htm). Serotype O:3 isolates with the prefix “Z” were also kindly supplied by Dr. Stan Fenwick of Massey University, Palmerston North, New Zealand (Table 1). Other bacteria used for PCR validation (Table 1) were those held by ESR's Christchurch Public Health Laboratory. Y. enterocolitica NZRM 3596 was used as a positive control for all pork analyses. Inocula were prepared by inoculating 2–3 colonies into 5 ml volumes of trypticase soy broth (TSB; Becton, Dickinson and Company, Sparks, USA) which were incubated at 24 °C for 18 h. These cultures were serially diluted in 0.1% peptone water and the numbers were determined from plate counts of colonies on trypticase soy agar (TSA; 15 g agar/L TSB) incubated for 24 h at 37 °C.
Table 1 Detection of the ail and virF genes by PCR in isolates of Yersinia enterocolitica, Yersinia spp. and other bacteria Isolate
Isolate
ail (356 bp) virF (231 bp)
Yersinia enterocolitica, serotype 1(2a,3) Y. enterocolitica, serotype 3 Y. enterocolitica, serotype 9 Y. enterocolitica, serotype O:8
NZRM 767
+
−
+ + −
− − −
+ +
+ +
+
−
+
+
+
−
+
+
+
+
+
+
+
+
+
+
+
+
+
+
− − + − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − −
− − − +? − − − − − − − − − − − − − − − − − − − − − − − − − − − − − −
NZRM 1000 NZRM 1001 NZRM 2603 (ATCC 9610)T Y. enterocolitica, serotype O:3 NZRM 3596 Y. enterocolitica, serotype O:3, z22 pig isolate Y. enterocolitica, serotype O:3, z26 pig isolate Y. enterocolitica, serotype O:3, z27 pig isolate Y. enterocolitica, serotype O:3, z28 pig isolate Y. enterocolitica, serotype O:3, z31 pig isolate Y. enterocolitica, serotype O:3, z32 pig isolate Y. enterocolitica, serotype O:3, z39 pig isolate Y. enterocolitica, serotype O:3, z52 pig isolate Y. enterocolitica, serotype O:3, z53 pig isolate Y. enterocolitica, serotype O:3, z57 pig isolate Y. enterocolitica, serotype O:3, z77 pig isolate Y. pseudotuberculosis NZRM 768 Y. frederiksenii NZRM 2534 Y. kristensenii NZRM 2535 Y. intermedia NZRM 2604 Arcobacter butzleri NZRM 4017 Bacillus cereus NZRM 5 B. subtilus NZRM 143 B. thuringiensis NZRM 3610 Brochothrix campestris NZRM 3569 B. thermosphacta NZRM 3320 Campylobacter coli NZRM 2607 C. fetus NZRM 2398 C. hyointestinalis NZRM 3676 C. jejuni NZRM 2397 C. jejuni subsp. doyley NZRM 3516 C. lari NZRM 2622 C. showae ATCC 51146 C. upsaliensis NZRM 3675 Carnobacterium divergens NZRM 3572 C. gallinarum NZRM 3575 C. mobile NZRM 3576 C. piscicola NZRM 3571 Clostridium perfringens NZRM 20 Enterococcus faecalis NZRM 1106 Enterobacter aerogenes NZRM 798 Escherichia coli NZRM 480 E. coli O157:H7 NZRM 3614 Klebsiella pneumoniae NZRM 482 Listeria grayi NZRM 1088 L. innocua NZRM 3024 L. ivanovii NZRM 797 L. monocytogenes NZRM 2597 L. seeligeri NZRM 3287 L. welshimeri NZRM 3286
J.A. Hudson et al. / International Journal of Food Microbiology 123 (2008) 25–31 Table 1 (continued) Isolate
Isolate
ail (356 bp) virF (231 bp)
Morganella morganii Proteus vulgaris Pseudomonas aeruginosa Saccharomyces cerevisiae Salmonella Menston S. typhi Shigella flexneri S. sonnei Staphylococcus aureus S. epidermidis Streptococcus bovis Vagococcus fluvialis Vibrio cholerae V. parahaemolyticus
NZRM 65 NZRM 67 NZRM 981 NZRM 1207 NZRM 383 NZRM 444 NZRM 3476 NZRM 86 NZRM 917 NZRM 1210 NZRM 2720 NZRM 3573 NZRM 776 NZRM 820
− − − − − − − − − − − − − −
− − +? − − − − − − − − − − −
NZRM, New Zealand Reference Culture Collection, Z22–Z77 supplied by Dr. Stan Fenwick (see Materials and methods). +? Weak PCR product observed.
2.2. Multiplex PCR conditions, validation and sensitivity Phenol/chloroform purified DNA from five Y. enterocolitica reference isolates and 11 isolates from pigs, four other Yersinia spp. and 44 other bacteria were examined for the presence of the chromosomal ail and plasmid-borne virF genes (Harnett et al., 1996), with a minor modification to remove the G from the 5′ end of the VirF-a primer. Each 50 µl reaction mixture contained 25 µl Qiagen (Hilden, Germany) Taq PCR mastermix (1.23 U Taq DNA Polymerase, 200 µM each of dATP, dCTP, dGTP and dTTP, 1.5 mM MgCl2, Qiagen PCR buffer), 0.2 µM of each primer, a varying volume of extracted DNA depending on source, with the balance being sterile water. The PCR was carried out under the following conditions; 35 cycles of 94 °C for 1 min (denaturation), 55 °C for 1 min (annealing) and 72 °C for 1 min (extension), followed by a final extension at 72 °C for 7 min. PCR products were subjected to electrophoresis (70 min, 110 V) using 2% agarose gels in Tris-boric acid-EDTA (TBE) buffer (USB Corporation, Cleveland, USA) containing 0.5 µg ethidium bromide ml−1, and visualised by ultraviolet transillumination. Products were compared to a 100 bp or 1 kb+ ladder (Invitrogen, Carlsbad, California, USA) and purified DNA from Y. enterocolitica NZRM 1000 (pYV-negative). To define the PCR detection limit, fresh broth cultures of Y. enterocolitica isolates NZRM 1000 (O:3 pYV-negative) and NZRM 3596 (O:3 pYV-positive) were centrifuged at 1600 × g for 20 min. The supernatants were discarded and the pellets were resuspended in 5 ml 0.85% NaCl. A mixture of the two cultures was also prepared by adding 1 ml from each into a third tube. The suspensions were serially diluted in 0.1% peptone water and 100 µl from each dilution was processed for PCR by heating for 12 min at 96 °C and then centrifuging for 12 min at 11,750 × g. The supernatant was immediately used for PCR. For this and in all subsequent work the 25 µl PCR reaction mixture contained 12.5 µl Qiagen taq PCR mastermix, 0.2 µM of each primer and 10 µl of the cell extraction. The thermocycling and visualisation conditions were as before. The number of cells added to each reaction mixture was calculated from colony counts on TSA. A cell-free control of uninoculated TSB was
27
processed with the cultures. Purified DNA from Y. enterocolitica NZRM 3596 (pYV-positive) was used as a positive PCR control and nuclease-free ultrapure water was used as a negative (sterility) PCR control. 2.3. Selection of enrichment medium for the isolation and identification of Y. enterocolitica The growth of 16 isolates of Y. enterocolitica serotypes O:1, O:3, O:8 and O:9, with or without the pYV plasmid (YeP+ or YeP−), was monitored by determination of the increase in optical density at 650 nm of cultures in modified tryptic soy broth (MTSB; (Bhaduri et al., 1997)) and Yersinia selective enrichment broth according to Ossmer (YSEO, catalogue No. 1.16701, Merck, Darmstadt, Germany). Comparisons were made of optical densities of triplicate samples incubated at 12, 24 and 30 °C over a period of 48 h, although the 30 °C incubation was discontinued at 24 h having reached a maximum at this time. 2.4. Sampling Retail portions of boneless pork, weighing approximately 400 g were purchased from four supermarkets and three butcher's shops in Christchurch, New Zealand. Portions were purchased from each supermarket on at least two occasions and the butcher's shops once between May and June 2006. Every portion differed with regard to cut type, location or packaging date. Of the 34 portions, 17 were used for presence/absence (P/A) testing and MPN determination, eight were used for P/A testing only and nine were used for MPN determination only. All portions were kept chilled and were processed within 24 h of being purchased. 2.5. Enrichment for P/A testing For P/A tests, a piece of meat was randomly selected from each portion and a segment measuring approximately 5 × 5 cm was aseptically excised from it. The entire upper surface of the segment was swabbed first with a cotton wool swab moistened with a solution of KOH:NaCl (0.25%:0.5% w/v), then with a dry swab. Both swabs were broken into a centrifuge tube containing 9 ml YSEO broth. The tube was shaken vigorously, and then incubated for 18 h at 24 °C. A positive control consisting of 9 ml YSEO broth inoculated with Y. enterocolitica NZRM 3596 was also prepared and incubated under the same conditions, along with a negative (sterility) control containing 9 ml uninoculated YSEO broth. Following enrichment, each tube was inverted 10 times to mix the contents, then the swabs were discarded. 2.6. Enrichment for MPN determinations Segments of meat were obtained and swabbed as for P/A tests. The swabs were broken off into a 50 ml tube containing 20 ml YSEO broth. The tube was vortexed vigorously for 1 min
28
J.A. Hudson et al. / International Journal of Food Microbiology 123 (2008) 25–31
then the volume was made up to 50 ml by the addition of YSEO broth. The tube was shaken vigorously by hand for 10 s, then the rinse fluid was dispensed into a 3 × 3 MPN system consisting of 10 ml of undiluted fluid in each of three 10 ml centrifuge tubes (level A), 1 ml of fluid in 9 ml YSEO broth in each of three 10 ml centrifuge tubes (a 1:10 dilution, level B), and 1 ml of a 1:10 dilution of the fluid in 9 ml YSEO broth in each of three 10 ml centrifuge tubes (a 1:100 dilution, level C). Controls were set up as before. All tubes incubated at 24 °C for 18 h. Segments of meat inoculated with various numbers of strain NZRM 3596 were prepared each time pork segments were sampled with 50 µl of each selected dilution being spread over the upper surface of a randomly selected pork segment. The inoculated segment was refrigerated for at least 40 min before it was sampled to allow attachment of the organism. 2.7. Multiplex PCR screening of pork enrichments A 1 ml volume of each rinse fluid was retained in a sterile tube under refrigeration for subsequent plating. The remaining (8 ml) of fluid was centrifuged at 1600 × g for 20 min, and the supernatant discarded. The pellet was resuspended in 5 ml sterile double-distilled water, and the suspension was centrifuged as before. The supernatant was discarded and the pellet was resuspended in 100 µl of sterile ultrapure water. A 10 µl volume of this suspension was briefly vortexed with 90 µl of sterile ultrapure water and prepared for PCR as described above. The presence or absence of the target PCR products (ail and/or vir-F) was recorded. Samples producing both products were scored as positive for YeP+. All P/A enrichments were screened by PCR before plating. For MPN enrichments, all of the level A tubes were analysed by PCR and tubes from levels A, B and C were tested by PCR for a sub-set of 16 samples. 2.8. Isolation and identification of YeP+ from pork enrichments by culture Each enrichment medium was assessed visually for turbidity and serially diluted accordingly, with a highly turbid samples being diluted to 10− 4, and non-turbid samples being diluted to 10− 2, in buffered peptone water (Merck, Darmstadt, Germany). The exception was the first dilution step where 0.5 ml of enrichment medium was added to 4.5 ml KOH/NaCl solution. After mixing the mixture was further diluted, or 20 µl was spread on CIN agar plates which were incubated for 48 h at 24 °C. Suspect colonies, at least three per enrichment where possible, were each subcultured to plates of TSA and Congo Red Brain Heart Infusion Agarose (CRBHO) (Bhaduri et al., 1991), and spotted to Rhamnose Arabitol McConkey agar (RAM) (Shehee and Sobsey, 2004) using a toothpick. CRBHO plates were incubated at 37 °C, and TSA and RAM plates at 24 °C, all for 18 h. Colonies typical of YeP+ strains on CRBHO are pin head sized and dark orange in colour and give white colonies on RAM. All colonies with these characteristics were selected for confirmation. RAM-positive/CRBHO-negative colonies (presumptively YeP−) were also selected for confirmation.
Isolates from TSA were tested for the production of urease, catalase and oxidase, and for citrate metabolism. Isolates positive for the urease and catalase tests, and negative for the oxidase and citrate tests were confirmed as YeP+ by multiplex PCR. To prepare the PCR template, a colony from a TSA plate was suspended in 100 µl of ultrapure water and processed as described above. The PCR conditions were as described for the 25 µl reactions above, except that only 1 µl of template was used and ultrapure water was added to make the final volume 25 µl. 2.9. Calculation of MPN A spreadsheet was developed using Microsoft Excel to calculate MPN values (Peeler et al., 1992). The spreadsheet used the number of tubes containing YeP+ and values for the surface area of sample tested in each MPN tube. The range over which this 9-tube MPN system operates is between 0.06 MPN/cm2 (LCI of 0.01 with one positive tube at level C; though one positive tube at level A is a more likely result and has a value of 0.07 MPN/cm2) and 8.77 MPN/cm2 (UCI of 39.74; all tubes but one positive at level C). If all tubes were negative the MPN was recorded as b 0.06 MPN/cm2, and if all were positive, the MPN was recorded as N 8.77 MPN/cm2. 3. Results and discussion 3.1. PCR validation and sensitivity All of the Y. enterocolitica isolates tested by the multiplex PCR contained ail, with the exception of NZRM 2603 (ATCC 9610T), a pYV-negative isolate of serotype O:8 isolated from an extra-intestinal infection (Table 1) which produced two PCR products at approximately 800 and 1100 bp. This isolate has been reported to be non-enteropathogenic (Neubauer et al., 2000) and not to contain the ail gene (Thoerner et al., 2003). The virF gene was not detected in every Y. enterocolitica isolate, possibly because they were either non-pathogenic or had lost the pYV plasmid on repeated laboratory culture. Of the other Yersinia species, the ail gene was detected in Y. kristensenii and it appeared that the virF gene was present in Y. intermedia, although subsequent comparison of the virF product from Y. intermedia directly against a control confirmed that the PCR products were of different sizes. Of the non-Y. enterocolitica species tested by Harnett et al. (1996), only Y. pseudotuberculosis produced an amplification product, and only for the virF gene. This was not reproduced for our Y. pseudotuberculosis isolate. We are not aware of any reports of the ail gene being detected in Yersinia species other than Y. enterocolitica (Miller et al., 1989) although similar sequences might be present, at least in Y. intermedia (Robins-Browne et al., 1989). Some of the other bacteria produced weak PCR products, but never of the correct size for both genes. The PCR was therefore specific for pYV-positive Y. enterocolitica (YeP+). The data presented here complement those for 129 pathogenic Y. enterocolitica, 71 non-pathogenic yersiniae and 100 non-Yersinia isolates analysed by Harnett et al. (1996).
J.A. Hudson et al. / International Journal of Food Microbiology 123 (2008) 25–31
29
Fig. 1. Detection limit of the multiplex PCR for a pYV-positive and a pYV-negative Y. enterocolitica, and a 1:1 combination 1.57 × 108 cfu Serial 1:10 dilutions Y. enterocolitica NZRM 1000 (O:3, pYV-negative) 1.44 × 108 cfu Serial 1:10 dilutions Mixture of NZRM 3596 and 1000 (1:1) 9.35 × 107 cfu Serial 1:10 dilutions 1 kb+ DNA ladder (lanes 1 and 32); TSB negative control (lane 29); PCR-positive control (lane 30); PCR-negative control (lane 31)
Lane 2 Lanes 3–10 Lane 11 Lanes 12–19 Lane 20 Lanes 21–28
Y. enterocolitica NZRM 3596 (O:3, pYV-positive)
The detection limit of the PCR was analysed with pYVpositive and pYV-negative Y. enterocolitica reference cultures, alone and in combination (Fig. 1). Under the conditions used, detection occurred at 104 to 105 cfu/ml (103 to 104 cells per 25 µl reaction). This is not a very sensitive method compared to some others (Thisted Lambertz et al., 2007), but is adequate for testing enriched samples. 3.2. Media for the isolation and identification of Y. enterocolitica After 24 h incubation all Y. enterocolitica isolates were approaching or had reached stationary phase in MTSB and YSEO broths at all temperatures. When incubation at 12 °C or 24 °C was extended beyond 24 h, the optical densities changed little. While MTSB yielded higher optical densities than YSEO broth the range of values was greater with MTSB. YSEO broth was selected for further experiments because of the advantage gained from the presence of antibiotics and because of its ability to enrich Y. enterocolitica isolates in a more uniform manner, i.e. not preferentially selecting one serotype. The differences in maximum optical densities reached indicated that the numbers were only 2–3 fold different between the two media.
In other reports 24% of 51 Norwegian pork samples were positive for pathogenic Y. enterocolitica by PCR, but none were positive by culture (Johannessen et al., 2000). Similarly, 25% of 255 minced pork samples from Finland were PCR-positive for pathogenic Y. enterocolitica, but only 2% were positive by culture (Fredriksson-Ahomaa et al., 1999). Lower frequencies have been reported, including 10% of 91, and 11% of 97 Swedish samples positive for pathogenic Y. enterocolitica (Thisted Lambertz and Danielsson-Tham, 2005; Thisted Lambertz et al., 2007). The data obtained from our small survey were therefore comparable to those from PCR studies conducted elsewhere, but the proportion of samples positive by culture was higher. This may be due to a real higher prevalence or an improved ability of the method to detect the viable organism which would, presumably, have resulted from the
Table 2 Most probable number results for all samples positive for pYV-positive Y. enterocolitica by PCR and/or culturing Product
Days old a
Most probable number (MPN/cm2) PCR MPN
3.3. Raw pork survey When the data for all the methods were considered 11 (32%) of the portions of pork were PCR-positive and YeP+ was isolated from six (18%). The prevalence by PCR was 37.5%, 44.4% and 23.5% by presence/absence testing (8 samples), MPN determination (9 samples) and both methods (17 samples) respectively, while by culture the equivalent prevalences were 12.5%, 33.3% and 11.8%.
Schnitzel Steak Schnitzel Chop Steak Schnitzel Schnitzel a
1 b1 1 4 1 3 3
b
+ + + + 0.30 1.52 5.42
Culture UCI
LCI
MPN
UCI
LCI
− − − − 1.34 6.89 24.59
− − − − 0.07 0.34 1.20
0.07 0.41 0.19 b0.06 0.06 0.06 b0.06
0.33 1.87 0.86 − 0.28 0.28 −
0.02 0.09 0.04 − 0.01 0.01 −
Number of days between packaging date and date sampled. +, Y. enterocolitica present but an MPN could not be calculated as only the level A enrichment tubes were tested by PCR. b
30
J.A. Hudson et al. / International Journal of Food Microbiology 123 (2008) 25–31
Table 3 Recovery of pYV-positive Y. enterocolitica in inoculated samples tested by the MPN method MPN/cm2
Product PCR
Steak Schnitzel Schnitzel Schnitzel Steak Schnitzel Schnitzel Steak a
Inoculum (cfu/cm2)
Culture
MPN
UCI
LCI
MPN
UCI
LCI
N8.77 +a + + 1.90 0.57 0.57 1.17
− − − − 8.60 2.59 2.59 5.30
− − − − 0.42 0.13 0.13 0.26
N8.77 N8.77 3.80 0.12 1.90 0.57 0.57 1.17
− − 17.21 0.56 8.60 2.59 2.59 5.30
− − 0.84 0.03 0.42 0.13 0.13 0.26
144.0 22.4 21.9 15.1 10.7 3.0 1.7 1.4
+, Y. enterocolitica present but an MPN could be calculated as only the level A enrichment tubes were tested by PCR.
YSEO enrichment as CIN agar was used commonly in the other studies. Of the seven samples that were PCR-positive by MPN testing, MPN values based on PCR could be derived for only three as, for some samples, only the level A MPN tubes were tested by PCR in an initial screen. For the three PCR-positive samples where an MPN value could be calculated, the concentrations of YeP+ in the three samples were 0.30, 1.52 and 5.42 MPN/cm2 (Table 2). Where comparisons were possible, the MPN values obtained through culture were lower than those calculated from the PCR results (Table 2). When it was not possible to isolate YeP+ from PCR-positive MPN tubes the MPN value was reduced. The plating method uses an alkaline treatment and antibiotics to reduce or inhibit non-Yersinia microflora, but this does not entirely remove background microflora. Identification of a wellisolated Yersinia colony among background microflora on CIN agar plates was often difficult, particularly when older meat was analysed. Since isolation of YeP+ from every PCR-positive MPN tube was challenging, we suggest that the isolation of YeP+ from any PCR-positive MPN tube provides evidence to confirm all PCRpositive results for that sample. False-positive PCR results from dead cells are unlikely as the PCR is conducted on an enrichment and has a relatively poor detection limit. However, when a sample is PCR-positive by MPN, but YeP+ are not isolated, the MPN result should be considered presumptive only. YeP+ were detected by the presence/absence method in all six inoculated samples over the range 1.4 to 17.3 cfu/cm2, though one sample inoculated with 1.7 cfu/cm2 was PCRnegative. No naturally-occurring YeP+ were detected in any of the meat used to prepare the inoculated samples. Of the portions of meat used for inoculated controls in the MPN analyses, two contained naturally-occurring YeP+ probably at numbers less than that of the inoculum. All eight inoculated samples were positive by PCR and isolates were recovered by plating (Table 3). For five of these samples all MPN tubes were analysed by PCR and the MPN values were identical by PCR and plating. YeP+ was detected in inoculated samples over the range 1.4 to 144 cfu/cm2. However, the MPN confidence intervals only encompassed the inoculum in four samples, and two exceeded the computable concentration of this
9-tube MPN system (N8.77 cfu/cm2). This indicates that recovery of YeP+ was good, but that the MPN values only approximated the inoculum levels. Disparities between the known inoculum and the calculated MPN varied, with most in the order of 3–10 fold. These differences might be due to losses through inefficient removal of the organism from the meat or from the filtration and centrifugation steps required for sample preparation. Cotton wool tipped swabs were used because of the small area sampled and the need not to use excision samples should the method be extended to carcass testing. Overall the method successfully detects YeP+ on meat samples when they are present at numbers as low as 2 cfu/cm2. Within 24 h of sample receipt a presumptive PCR result is obtainable. Any suspected Yersinia can be isolated within three days, and YeP+ can be confirmed 24 h later by specific biochemical tests and PCR. This method for the detection and enumeration of YeP+ will allow surveys of meat that can provide data for exposure assessment purposes. The method might also be used in outbreak investigations, though assessment of the method with other food types is required. The method could be improved by better sample preparation to recover more Y. enterocolitica cells, and better techniques are likely to be needed for DNA recovery from samples containing PCR inhibitors. In addition, CIN agar does not distinguish Y. enterocolitica clearly from other bacteria, and the need for a better plating medium remains. Acknowledgements We are grateful to the New Zealand Ministry of Health and New Zealand Food Safety Authority for financial support, and Dr. Stan Fenwick for supplying pig isolates. Thanks are also due to Jane Capill for her capable technical assistance, and to our molecular biology colleagues for their help and forbearance. This paper is dedicated to Surinah Monson who died before her time during the course of this work. References Aulisio, C.C.G., Mehlman, I.J., Sanders, A.C., 1980. Alkali method for rapid recovery of Yersinia enterocolitica and Yersinia pseudotuberculosis from foods. Applied and Environmental Microbiology 39, 135–140.
J.A. Hudson et al. / International Journal of Food Microbiology 123 (2008) 25–31 Bhaduri, S., Turner-Jones, C., Lachica, R.V., 1991. Convenient agarose medium for simultaneous determination of the low-calcium response and congo red binding by virulent strains of Yersinia enterocolitica. Journal of Clinical Microbiology 29, 2341–2344. Bhaduri, S., Cottrell, C., Pickard, A.R., 1997. Use of a single procedure for selective enrichment, isolation, and identification of plasmid-bearing virulent Yersinia enterocolitica of various serotypes from pork samples. Applied and Environmental Microbiology 63, 1657–1660. Bhagat, N., Virdi, J.S., 2007. Distribution of virulence-associated genes in Yersinia enterocolitica biovar 1A correlates with clonal groups and not the source of isolation. FEMS Microbiology Letters 266, 177–183. ESR, 2007. Yersiniosis. New Zealand Public Health Report 5, 4. Fenwick, S.G., McCarthy, M.D., 1995. Yersinia enterocolitica is a common cause of gastroenteritis in Auckland. New Zealand Medical Journal 108, 269–271. Foultier, B., Cornelis, G.R., 2003. DNA sequence and analysis of the pYVa127/90 virulence plasmid of Yersinia enterocolitica strain A127/90. Research in Microbiology 154, 553–557. Fredriksson-Ahomaa, M., Korkeala, H., 2003. Low occurrence of pathogenic Yersinia enterocolitica in clinical, food, and environmental samples: a methodological problem. Clinical Microbiology Reviews 16, 220–229. Fredriksson-Ahomaa, M., Hielm, S., Korkeala, H., 1999. High prevalence of yadA-positive Yersinia enterocolitica in pig tongues and minced meat at the retail level in Finland. Journal of Food Protection 62, 123–127. Grahek-Ogden, D., Schimmer, B., Cudjoe, K.S., Nygård, K., Kapperud, G., 2007. Outbreak of Yersinia enterocolitica serogroup O:9 infection and processed pork, Norway. Emerging Infectious Diseases 13, 754–756. Harnett, H., Lin, Y.P., Krishnan, C., 1996. Detection of pathogenic Yersinia enterocolitica using the multiplex polymerase chain reaction. Epidemiology and Infection 117, 59–67. Hudson, J.A., Mott, S.J., Delacy, K.M., Edridge, A.L., 1992. Incidence and coincidence of Listeria spp., motile aeromonads and Y. enterocolitica on ready-to-eat fleshfoods. International Journal of Food Microbiology 16, 99–108. Johannessen, G.S., Kapperud, G., Kruse, H., 2000. Occurrence of pathogenic Yersinia enterocolitica in Norwegian pork products determined by a PCR method and traditional culturing method. International Journal of Food Microbiology 54, 75–80. Kapperud, G., Ostroff, S.M., Nesbakken, T., Hutwagner, L.C., Bean, N.H., Lassen, J., Tauxe, R.V., 1995. Risk factors for sporadic Yersinia enterocolitica infections in Norway: a case-control study. Contributions in Microbiology and Immunology 12, 25–28. Miller, V.L., Farmer, J.J., Hill, W.E., Falkow, S., 1989. The ail locus is found uniquely in Yersinia enterocolitica serotypes commonly associated with disease. Infection and Immunity 57, 121–131.
31
Neubauer, H., Aleksic, S., Hensel, A., Finke, E.J., Meyer, H., 2000. Yersinia enterocolitica 16 S rRNA gene types belong to the same genospecies but from three homology groups. International Journal of Medical Microbiology 290, 61–64. Peeler, J.T., Houghtby, G.A., Rainosek, A.P., 1992. The Most Probable Number Technique. Compendium of Methods for the Microbiological Examination of Foods. 105–120. American Public Health Association, Washington, D.C. Robins-Browne, R.M., Miliotis, M., Cianciosi, S., Miller, V., Falkow, S., Morris, J., 1989. Evaluation of DNA colony hybridization and other techniques for the detection of virulence in other Yersinia species. Journal of Clinical Microbiology 27, 644–650. Satterthwaite, P., Pritchard, K., Floyd, D., Law, B., 1999. A case control study of Yersinia enterocolitica infections in Auckland. Australian and New Zealand Journal of Public Health 23, 482–485. Shehee, M.W., Sobsey, M.D., 2004. Development of a L-rhamnose and D-arabitol supplemented McConkey agar to identify pathogenic Yersinia enterocolitica among environmental yersinias in swine production wastes. Journal of Microbiological Methods 57, 289–292. Sulakvelidze, A., 2000. Yersiniae other than Y. enterocolitica, Y. pseudotuberculosis and Y. pestis: the ignored species. Microbes and Infection 2, 497–513. Thisted Lambertz, S., Danielsson-Tham, M.L., 2005. Identification and characterization of pathogenic Yersinia enterocolitica isolates by PCR and pulsed-field gel electrophoresis. Applied and Environmental Microbiology 71, 3674–3681. Thisted Lambertz, S., Granath, K., Fredriksson-Ahomaa, M., Johansson, K.E., Danielsson-Tham, M.L., 2007. Evaluation of a combined culture and PCR method (NMKL-163A) for detection of presumptive pathogenic Yersinia enterocolitica in pork products. Journal of Food Protection 70, 335–340. Thoerner, P., Bin Kingombe, C.I., Bögli-Stuber, K., Boissig-Choisat, B., Wassenaar, T.M., Jemmi, T., 2003. PCR detection of virulence genes in Yersinia enterocolitica and Yersinia pseudotuberculosis and investigation of virulence gene distribution. Applied and Environmental Microbiology 69, 1810–1816. Toora, S., Budu-Amoako, E., Ablett, R.F., Smith, J., 1994. Evaluation of different antimicrobial agents used as selective supplements for isolation and recovery of Yersinia enterocolitica. Journal of Applied Bacteriology 77, 67–72. Weynants, V., Jadot, V., Denoel, P.E., Tibor, A., Letesson, J.J., 1996. Detection of Yersinia enterocolitica serogroup O:3 by a PCR method. Journal of Clinical Microbiology 34, 1224–1227. Wright, J., 1996. Gastrointestinal illness: what are the causative organisms? The New Zealand Public Health Report 3, 9–11.