Incidence of non-typhoidal Salmonella in poultry products in the North West Province , South Africa

HOW TO CITE: Olobatoke RY, Mulugeta SD. Incidence of non-typhoidal Salmonella in poultry products in the North West Province, South Africa. S Afr J Sci. 2015;111(11/12), Art. #2014-0233, 7 pages. http://dx.doi.org/10.17159/ sajs.2015/20140233 This study was conducted to evaluate the incidence of non-typhoidal Salmonella (NTS) serotypes in raw and ready-to-eat (RTE) broiler products in the North West Province of South Africa. A total of 120 raw broiler samples, 40 samples of polonies and 20 samples of smoked viennas were obtained from retail points in major cities and towns in the province. Samples were subjected to aerobic plate count and later screened for the presence of NTS using phenotypic and genotypic techniques. The average bacterial count in raw products was 3.1 x 105 cfu/g whereas bacterial contamination of RTE products was 1.8 x 103 cfu/g. The average recovery rate of NTS species from raw broiler products was 12.5% and the serotypes identified were S. typhimurium (46.4%), S. enteritidis (30.9%) and S. newport (22.9%). No NTS was recovered from the RTE products. However, S. typhimurium was the predominant serotype in whole carcasses whereas S. enteritidis and S. newport were prevalent in chicken parts. Out of the 160 presumptive NTS isolates screened by polymerase chain reaction (PCR), 140 (87.5%) were confirmed for the presence of the Salmonella-specific invA gene. In addition, 115 (82.4%) of the confirmed isolates harboured the plasmid spvC gene. Random Amplified Polymorphic DNA (RAPD) fingerprinting of isolates using RAPD 1 and RAPD 3 primers, revealed some interand intra-serotype genetic diversity among isolates, suggesting varying sources of contamination. The results of this study represent the first report on the incidence and prevalent serotypes of NTS in chicken products in the North West Province of South Africa.


Introduction
Non-typhoidal salmonellosis is an important public health problem worldwide and particularly in sub-Saharan Africa where it commonly manifests as gastroenteritis and/or bloodstream infections in both children and adults. 1 The gastroenteritis form, which may sometimes be self-limiting, is commonly found in industrialised countries.Immunocompromised individuals including patients with HIV, cancer or diabetes, are at higher risk of non-typhoidal Salmonella (NTS) bacteraemia and often develop focal infections such as meningitis, septic arthritis, pneumonia and osteomyelitis. 2 Although more than 2500 serovars of Salmonella enterica have been reported, S. typhimurium and S. enteritidis are identified as the commonest causes of human infection. 3,4A retrospective study in one hospital in the Democratic Republic of Congo from 2002 to 2006, revealed that NTS caused 59% of bacteraemia in children.Salmonella typhimurium and Salmonella enteritidis were responsible for 82% of the cases. 5In Mozambique, NTS were reported to account for 120 cases of childhood bacteraemia per 100 000 persons/year. 6Contaminated poultry meat and eggs, among other factors, have been implicated as vehicles of transmission for these hardy pathogens. 7[9] South Africa has witnessed a tremendous increase in chicken meat consumption. 10Concurrently, the Enteric Disease Reference Unit of the National Institute for Communicable Diseases noted an increasing number of NTS isolates despite the fact that human salmonellosis cases are rarely reported.Evidence of these occurrences are the outbreaks of food-borne illnesses in Mpumalanga Province of South Africa incriminating NTS serotypes.One of the outbreaks involved the consumption of meals prepared with poultry products. 11,12These outbreaks indicate the presence of NTS in South Africa, which may be an issue of public health concern.A few investigations have been conducted in South Africa to ascertain the contamination of chicken carcasses and ready-to-eat foods from retail stores, with various pathogenic bacteria including Salmonella. 13,14However, the incidence of NTS in broiler products in the North West Province (NWP) has not been established despite the fact that the NWP is one of the provinces with the largest production and distribution of broilers in the country.The objective of this study therefore was to screen raw and ready-to-eat broiler products obtained from the NWP of South Africa for incidence of NTS contamination.

Experimental design
The study was cross-sectional with sampling based on two shop types and three product groups.The sampling, which lasted for 6 months, commenced in October 2010 and ended in March 2011, corresponding with the summer months in the study area.

Sample size
Primary population size to be sampled was determined according to the following formula 15 : where SD = p(1-p) Equation 1 where p is the expected prevalence; L is the accepted error and Z a is the value for normally distributed data at a confidence level a.Based on an expected prevalence of 40% and a confidence level of 90%, a total of 57 shops were sampled, irrespective of the source of their products (i.e.either from within or outside the province).A total of 180 samples was obtained for analysis based on the purchase of a minimum of three product groups from each shop, and keeping 5% above the actual sample size to prevent the risk of sample loss during analysis (Table 1).Temperature of samples at point of purchase ranged from -20 °C to 4 °C.For all the procedures involved in the analysis, Salmonella typhimurium ATCC 14028 was used as a positive control strain whereas E. coli ATCC 25922 was used as a negative control.All culture, isolation and biochemical screening techniques were carried out using the MFHPB-20 procedure 16 with appropriate modifications.aThe number of shops visited was dependent on the number of shops available within a particular location.bA minimum of three samples comprising three sample types was purchased from each shop.

Determination of total bacterial count in samples
Briefly, 25 g of each sample was aseptically removed and homogenised with 225 mL of 2% buffered peptone water (BPW) in a stomacher bag (Nasco, Swedesboro, NJ, USA).Then 1 mL of the homogenate was transferred to Eppendorf tubes and serial dilutions were made using 2% BPW.An aliquot (0.1 mL) of each dilution was plated on plate count agar (Merck, SA) by spreading, and incubated at 37 °C for 24 h after which colonies were counted and recorded as cfu/g sample.

Isolation of Salmonella species from broiler carcasses and sausages
For broiler carcasses, each sample was aseptically removed from the package and transferred to a sterile plastic bag.Then 150 mL of 2% BPW was added to each bag and the mixture was shaken constantly for 2 min to obtain carcass rinse.Thereafter, 25 mL of the rinse liquid was transferred to a Whirl-pak bag (Nasco, USA) and another 75 mL of 2% BPW was added.Chicken sausages were aseptically homogenised with 225 mL of double strength (4%) BPW in a stomacher bag (Nasco, USA).Each bag was sealed and incubated at 37 °C for 18 h.Following incubation, contents of each Whirl-pak bag (Nasco, USA) was homogenised and 0.1 mL aliquot of the mixture was transferred to 10 mL of Rappaport-Vassiliadis broth (Merck, SA) for selective enrichment of Salmonella.The broth was incubated at 42 °C for 24 h and then kept at room temperature for 24 h. 17Thereafter, a loopful of each broth sample was streaked on xylose lysine desoxycholate (XLD) agar (Merck, SA) and Salmonella Shigella agar (Merck, SA) and incubated at 37 °C for 24-36 h.Suspected colonies were further purified on XLD agar and thereafter preserved on nutrient agar (Merck, SA) slants for further analysis.

Biochemical screening of presumptive isolates
All suspect colonies from the XLD agar culture were inoculated on triple sugar iron agar (Merck, SA) slant and incubated at 37 °C for 24 h.Presumptive Salmonella isolates from the triple sugar iron test were further confirmed using the API-20E test (bioMerieux Inc., SA) performed according to the manufacturer's instructions.

Serological identification
Isolates showing typical Salmonella biochemical reactions were serotyped by the slide agglutination test using Salmonella specific polyvalent antisera (Davies Diagnostics, SA) for O and H antigens.The test was performed according to the manufacturer's instructions.

DNA extraction
Genomic and plasmid DNA were extracted using the Zymo ® kit (Inqaba Biotech, SA) and alkaline lysis method, 18 respectively.The quantity and purity of all extracted DNA were estimated using a UV visible spectrophotometer (model S-22, Boeco, Germany) after which the integrity of the DNA was checked on standard submarine gel electrophoresis using 0.8% (w/v) agarose.The isolated DNA was stored at -20 °C until use.

PCR of extracted DNA
The 16S rRNA gene fragments of all isolates were amplified, 19 after which presumptive Salmonella isolates were screened for the presence of invasion (invA) and virulence (spvC) genes for identity confirmation. 20ll the oligonucleotide primers used for polymerase chain reactions (PCR) were obtained from Inqaba Biotech, South Africa, and details of the sequences and cycling conditions are shown in Table 2.All PCR reactions were prepared in 25 µL volumes consisting of 1 µg/µL of the template DNA, 50 pmol of each oligonucleotide primer set, 1x PCR master mix, 1U Taq DNA polymerase and nuclease free distilled water.
Amplifications were performed using a Peltier Thermal Cycler (model Dyad™ DNA Engine).A reaction blank containing all the components of the reaction mixture except the template DNA was included in each PCR procedure to check for contamination.Thereafter, 10 µL of the PCR amplicons was analysed by standard submarine gel electrophoresis using 1% (w/v) agarose at 60 V for 6 h.Lambda DNA molecular weight marker was the gene ruler used in electrophoretic analyses.Electrophoresis was conducted in a horizontal Pharmacia biotech equipment system (model Hoefer HE 99X; Amersham Pharmacia Biotech, Sweden) using 1x TAE running buffer (40 mM Tris, 1 mM EDTA and 20 mM glacial acetic acid, pH 8.0).The gels were stained in ethidium bromide (0.1 mg/mL) for 10-15 min and later visualised under UV light. 18A Gene Genius Bio imaging system (Syngene, Synoptics; UK) was used to capture the image using GeneSnap (version 6.00.22) software.Amplicons which showed clear bands for 16S rRNA and invA genes were purified and sequenced at Inqaba Biotech (South Africa), for strain identification of the isolates, using the ABI 3500 XL sequencer.The search for amino acid homology of the sequence results was done using the NCBI Basic Local Alignment Search Tool (BLAST) (http://www.ncbi.nlm.nih.gov/BLAST/).

RAPD-PCR
Confirmed Salmonella isolates identified through the expression of the Salmonella-specific invA gene were further subjected to fingerprinting by random amplified polymorphic DNA (RAPD)-PCR to evaluate the genetic diversity among the isolates.Oligonucleotides of short sequence were used 21,22

Data analysis
To obtain the cfu/g of sample for total bacterial count, colonies on plate count agar were counted and compared with the dilution factor. 23All Salmonella populations were transformed to base 10 logarithms before analysis.Recovery rate of Salmonella isolates was calculated using a previously described equation. 24Means in recovery rates between locations, shop and product types were compared using the Student's t-test of SPSS (SPSS 10.0 for windows, SPSS Chicago, IL, USA).Differences were considered significant at p≤0.05.

Recovery rate =
No of positive samples X 100 Total no of samples Equation 2For cluster analysis, the distance matrices, means and standard deviations were calculated using Statistica and the result was used to construct a phylogenetic relationship among isolates by the neighbour-joining method of Saitou and Nei 25 .

Total bacterial count of samples
On average, total bacterial count of butchery samples (2.2 x 10 5 cfu/g) was higher than that of supermarket samples (1.4 x 10 4 cfu/g).Similarly, raw chicken parts were more contaminated (7.6 x 10 5 cfu/g) than whole chicken carcasses (2.5 x 10 5 cfu/g).RTE products were the least contaminated (1.8 x 10 3 cfu/g) perhaps because the products had been subjected to further processing.However, the level of microbial contamination of all analysed products was within the recommended range for some developed countries such as United Kingdom (10 4-10 5 cfu/g) and Australia (10 6 cfu/g). 26Sources of contamination of raw poultry products vary and could originate from the live birds, processing procedures or from the environment. 27The higher microbial contamination rate of butchery samples and raw chicken portions when compared with supermarket products might be an indication of poor hygienic conditions in the processing environment.This suggests a lack of strict hygiene control measures during product processing and could have a public health implication on the consumers as poultry are usually contaminated with a relatively high frequency of pathogenic bacteria.Indeed, raw poultry products have been reported in quite a number of human food poisoning cases, particularly following handling, undercooking or mishandling of the cooked products. 28,29Furthermore, high level of product contamination, such as recorded in some samples in the current study, could facilitate product spoilage, particularly when organisms such as Pseudomonas are involved.

Incidence of non-typhoidal Salmonella in broiler products
Out of 11 locations sampled, Salmonella was recovered from products obtained from 6 (54.5%) of the locations (Table 1).The difference in Salmonella isolation rates between positive locations was significant (p=0.009) with Mafikeng having the highest rate.Mafikeng is a major town with about the highest number of retail shops in the province.As a result, the number of shops visited and samples obtained from Mafikeng in this study were more than other locations and this could be responsible for the difference in Salmonella isolation rates between locations.Salmonella spp. was not recovered from RTE products whereas average recovery rate in raw broiler products was 12.5%.This rate is lower than the 19.5% reported in the Gauteng Province of South Africa 13 , 20% in the USA 30 , 30% in Canada 31 , 39.7% in Mexico 32 and 60% in Portugal 8 .Other countries such as Zambia 33 , Saudi Arabia 34 , Turkey 7 and Sudan 9 , however, recorded lower incidences of NTS (4.7%, 5.92%, 8% and 9.2%, respectively) in broiler products than noted in the current study.Incidence of Salmonella contamination was higher in whole chicken carcasses (30%) than chicken parts (20%).This was contrary to the observations of Chaiba et al. 35 and Moussa et al. 34 who noted higher incidences of NTS in chicken cuts in Morroco and Sudan respectively.The presence and distribution of NTS serovars vary from region to region depending on the sampling plan and detection limits of the methodology employed. 36tchery products were more contaminated with NTS (32%) than supermarket products (18%).This result, though contrary to the report by Yang et al. 37 , was in agreement with the findings of Chaiba et al. 35 and could be the result of better hygienic standards maintained in supermarkets than butcheries.Frozen products had a higher NTS recovery rate (10.4%) than fresh products (8.8%).This may suggest initial high contamination of the frozen products with NTS.Although previous studies identified storage temperature as an important risk factor for pathogen growth and survival, the result of the current study suggests that freezing temperature range was not sufficient to inactivate the pathogens.It has been documented that freezing may not be regarded as a means of destroying food borne microorganisms.This is because low freezing temperatures of about -20 °C are less harmful to the pathogens than the median temperature ranges. 38

Molecular characterisation of non-typhoidal Salmonella isolates
A total of 160 presumptive Salmonella isolates obtained from analysed broiler products were screened for the presence of invA and spvC genes, following the positive amplification of the 16S rRNA fragment.
The invA gene was detected in 140 (87.5%) of the 160 screened isolates.Sequencing of the 16S rRNA amplicons of the remaining 20 isolates showed that they belonged to three other coliform species namely Klebsiella, Escherichia and Serratia.These could be the contaminants that showed biochemical and serological reactions similar to Salmonella.The invA gene, located on the Salmonella pathogenicity island 1 (SPI 1), is highly conserved in Salmonella species and encodes a type III secretion system (TTSS) that exports proteins in response to bacterial contact with epithelial cells. 39The invA gene operon is essential for full virulence in Salmonella. 40The high detection of invA gene in the current study may be an indication of the potential pathogenicity of the isolated Salmonella strains and could be a cause for public health concern.Torpdahl et al. 40 and Khan et al. 41 similarly detected invA gene in all Salmonella isolates recovered from poultry products and orange juice, respectively.Analysis of the sequenced isolates in the current study revealed three NTS serotypes, namely, S. typhimurium (46.4%; n=65), S. enteritidis (30.7%; n=43) and S. newport (22.9%; n=32).The sequences of the three serotypes were deposited in GenBank with accession numbers JX859913, KC683709 and JX859912, respectively.The rate of isolation of the different NTS serotypes was in agreement with the reports from previous research. 35,36All the identified serotypes were isolated from both supermarket and butchery samples as well as whole carcasses and chicken parts at different rates (Figure 1).This shows that the isolated Salmonella serotypes may be circulating within the environment and possible evidence is the fact that some of the products were contaminated with multiple NTS serotypes, which may also indicate varying sources of product contamination.
Out of the 140 isolates confirmed for the presence of the invA gene, 115 (82.4%) isolates were found to harbour the spvC genes representing 71.9% of all the screened isolates.The spvC gene in Salmonella spp.interacts with the host immune system and is responsible for an increased growth rate in host cells. 41The prevalence of the spvC gene fragment was higher in isolates from butchery samples (71.4%) than supermarket samples (25.9%) and isolates from chicken parts (57.1%) than whole broiler carcasses (42.9%).The detection rate of the spvC gene in the current study agreed with the findings of Bolton et al. 42 and Khan et al. 20 who detected the gene fragment in 97% and 88% of Salmonella typhimurium DT104 strains, respectively.However, Khan et al. 41 did not detect spvC gene in any of the Salmonella serotypes subjected to multiplex PCR.The recovery rate of S. typhimurium in this study was notably higher than that of other serotypes (S. enteritidis and S. newport), as opposed to the WBC, whole broiler carcasses; CP, chicken parts; RTE, ready-to-eat.Bars with different labels are significantly different (p<0.05).

Analysis of RAPD-PCR
The profiles generated by the RAPD-PCR were composed of 1-5 bands ranging between 0.25 kb and 1.4 kb.Both RAPD 1 and RAPD 3 primers each generated four banding profiles for S. typhimurium, three for S. newport and two profiles for S. enteritidis (Figure 2a and 2b).One strain of S. newport (SN1), isolated from chicken parts obtained from the butchery, showed no bands with both RAPD 1 and RAPD 3 primers.Similarly, a strain of S. typhimurium (ST2) isolated from supermarket whole broiler carcass, showed three bands with RAPD 1 but no bands with RAPD 3 primer.However, both strains produced one band of about 0.45 kb as did most of the isolates with RAPD 2 primer although the primer had very poor discriminatory power and was therefore not included in the analysis.Cluster analysis of the RAPD 1 profiles revealed three different RAPD types for S. newport and S. enteritidis whereas S. typhimurium was classified into four types (Figure 3a).RAPD 3 primer on the other hand, grouped S. typhimurium into four, S. newport into three and S. enteritidis into two RAPD types (Figure 3b).Although a few isolates of the same serotype showed either similar or same banding patterns, some level of heterogeneity was generally noticed in the isolates, irrespective of the sources (i.e.geographical location, shop or product type).This may again indicate varying sources of product contamination (e.g.live birds, water, environment or product handler).Furthermore, changes in bacterial genome may occur in the same serovar through plasmid acquisition/genetic mutations without manifesting phenotype alterations. 43Interestingly, a particular strain of S. typhimurium, isolated from a Mafikeng product namely chicken drumstick, was closely clustered with the positive control strain by both RAPD 1 and RAPD 3 primers.This attests to the efficiency of the Salmonella detection method employed in the current study.In spite of the heterogeneity manifested by the isolates however, it may still be concluded that most of the isolates have the same clonal origin.

Conclusion
The results of this study represent the first report on the incidence and prevalent serotypes of NTS in chicken products in the NWP.It is, however, worthy to note that the three NTS serotypes isolated in this study were also the same serotypes identified in 49 NTS cases reported from the NWP to GERM-SA in 2011. 44This indicates that raw chicken products may serve as major vehicles contributing to foodborne salmonellosis in the NWP.Furthermore, multiple contamination of one sample by different serotypes, and the genetic diversity among isolates of the same or different serotypes (as expressed by RAPD fingerprinting) suggest varying sources of product contamination.This is a cause for public health concern and calls for adequate monitoring and establishment of effective strategies to control contamination along the production/ supply chain, in order to safeguard public health.The fact that NTS was not recovered from polonies and viennas (RTE products) in the current study may indicate effectiveness of employed RTE processing techniques in eliminating Salmonella contaminants from raw products.This report should provide confidence and reassurance to both RTE producers and consumers.

Figure 1 :
Figure 1: Percentage incidence of non-typhoidal Salmonella (NTS) serotypes by (a) shop type and (b) product type.

Table 1 :
Number of Salmonella positive samples obtained from different locations