MATERIALS AND METHODS

Sample collection: A total of 510 samples comprising poultry blood (255) and faeces (255) from 255 birds were collected from retail outlets of Bareilly city, Uttar Pradesh, India. Ten milliliter of poultry blood was collected aseptically in sterile tubes containing anticoagulant (1.5 mg EDTA mL^-1 blood) and transported to the laboratory under chilled condition (±4°C). The caecum from the same birds were aseptically collected into sterile polythene bags, transported under chilled condition (±4°C) and processed immediately.

Conventional cultural method: Isolation and identification of Salmonella spp. from the 510 samples was carried out by following standard protocol (Agarwal et al., 2003). Samples were pre-enriched in Buffered Peptone Water (BPW) in 1:10 ratio (37°C, 16 h). Selective enrichment (42°C, 24 h) was done in tetrathionate broth (TT) and selective plating was performed on Hektoen Enteric Agar (HEA) at 37°C for 24 h. Smooth, transparent colonies with greenish periphery with or without black centre on HEA were picked up and confirmed biochemically (Barrow and Feltham, 1993). Biochemical identification was carried out on 3-5 isolated colonies. Initially, the isolates were inoculated in Triple Sugar Iron (TSI) agar and urea broth at 37°C for 24 h. The colonies showing a TSI result of alkaline slant (pink) and acidic butt (yellow) with or without H₂S production (blackening) and urease negative were considered as positive for Salmonella presumptively. The isolates which were positive for Salmonella by these tests were further subjected to the primary and secondary biochemical identification tests (Table 1).

Serotype confirmation of the biochemically positive isolates was carried out by tube agglutination test using polyvalent antiserum and tube agglutination test using somatic and flagellar group specific and factor antiserum available in the laboratory.

Table 1:	Biochemical characterization of Salmonella isolates

Table 2:	Oligonucleotide primers used in the study

PCR method
Oligonucleotide primers: Primers used in the study (Table 2) were custom synthesized from Genuine Chemical Corporation (GCC), New Delhi, India.

Template preparation: DNA was isolated from pre-enriched samples (in buffered peptone water) by Snap-chill method. Briefly, 2 mL of the pre-enriched samples were centrifuged at 6,000 rpm for 10 min to pellet the bacterial cells and washed with sterile Phosphate Buffered Saline (PBS) (10 mM; pH 7.4) once. The pellet was resuspended in 50 μL of nuclease free water and kept in a boiling water bath (100°C) for 10 min. This was then transferred immediately to -20°C for 10 min. After incubation, the suspension was centrifuged again at 6,000 rpm for 10 min and the supernatant collected which was used as template for PCR assay.

Genus specific PCR: The PCR was standardized for genus specific detection of Salmonella from poultry blood and faecal samples (510) targeting invA gene (Galan et al., 1992). The PCR reaction mixture was finally optimized to contain 2.5 μL of 10X PCR buffer, 2.5 mM MgCl₂, 0.2 mM dNTP mix, 10 pmol of each forward and reverse primers, 1 U of Taq DNA polymerase, 4 μL of the template prepared by boiling and snap chilling method and sterile deionised water upto 25 μL.

The reaction was performed in Eppendorf gradient thermocycler with preheated lid (lid temperature 105°C). The cycling condition comprised of initial denaturation at 94°C for 5 min, followed by 34 cycles each of denaturation at 94°C for 1 min, primer annealing at 50°C for 1 min, elongation at 72°C for 1 min and finally a single extension step at 72°C for 7 min. The bands were visualized on 1.5% agarose gel electrophoresis and photographed by the gel documentation system. All the samples were screened for the presence of invA gene.

Serotype specific PCR: All the isolates recovered by cultural method were subjected for serotypic identification of Salmonella Typhimurium and Salmonella Enteritidis by targetting typh and ent genes, respectively, as per Alvarez et al. (2004). The PCR mixture and cyclic condition were same as that for invA gene except that the annealing temperature for typh gene was 57°C for 1 min. Electrophoretic analysis of PCR amplified product was performed on 1.5% agarose gel and results were recorded.

Statistical analysis: Evaluation sensitivity, specificity, positive predictive value and negative predictive value and concordance of diagnostic tests, viz., PCR and cultural method were done as per Thrusfield (2005).

RESULTS

A total of 510 field samples (poultry blood and faeces) were subjected to conventional cultural isolation and PCR methods for detection of Salmonella with special reference to Typhimurium and Enteritidis serotypes. Samples were pre-enriched in BPW, followed by enrichment in TT broth and selective plating on HEA plates. The suspected colonies were confirmed by biochemical reactions. Out of 510 samples, Salmonella spp. was isolated from a total of 26 (5.09%) samples, of which 17 (6.66%) were from blood samples and 9 (3.52%) from faecal samples. On serotyping, 15 isolates (57.69%) were recognized as Salmonella Typhimurium, 5 isolates (19.23%) as Salmonella rough type, 4 isolates (15.38%) as Salmonella Enteritidis and 2 (7.69%) isolates were untypable. Out of 15 Salmonella Typhimurium isolates, 11 isolates (73.33%) were from poultry blood samples and remaining 4 isolates (26.66%) were from poultry faeces samples. Four isolates of Salmonella Enteritidis isolates were from blood and out of 5 rough strains 4 were isolated from faeces samples and the remaining one was from blood sample. The untypable isolates (1 each) were from blood and faeces (Table 3).

PCR assay for the detection of Salmonella from pre-enriched samples was standardized employing a set of primer of invA gene. Electrophoretic analysis of the PCR product revealed the specific amplification of a 284 bp fragment (Fig. 1). No non-specific products were detected on agarose gel electrophoresis. Similarly, serotype specific PCR for detection of Typhimurium and Eneteritidis were successfully standardized, resulting amplification of 401 bp (Fig. 2) and 304 bp (Fig. 3) products, respectively.

Table 3:	Serotypes of Salmonella recovered from poultry samples

Fig. 1:	PCR amplification of invA gene, Lane M: 1.5+100 bp DNA ladder, Lane 1: Salmonella Typhimurium and Lane 2: Salmonella Enteritidis

Fig. 2:	PCR amplification of typh gene, Lane M: 100 bp DNA ladder and Lane 1, 2, 3: Salmonella Typhimurium

Fig. 3:	PCR amplification of ent gene, Lane M: 100 bp DNA ladder and Lane 1, 2, 3: Salmonella Enteritidis

Analysis of 510 poultry samples (255 each of blood and faeces) by PCR employing invA gene revealed that, of the 255 blood samples, 22 (8.62%) were positive but only 7 (2.74%) of the 255 faecal samples gave positive amplification. In all 29 samples were positive by PCR, whereas only 26 samples were positive by cultural method.

All the 26 isolates of Salmonella were also checked by serotype specific PCR to confirm the presence of Salmonella Typhimurium and Salmonella Enteritidis. The analysis revealed that the 15 (57.69%) isolates found positive for Salmonella Typhimurium by serotyping were also found positive for typh gene indicating them to be Salmonella Typhimurium. However, 8 (30.76%) isolates were found to be positive for ent gene (Table 3) but by serotyping only 4 isolates were found to be Eneritidis and remaining 4 were recognized as rough type. On further probing of these 4 rough isolates, it was found that although they lacked O antigen but all of them were positive for g, m flageller antigen. Interestingly, Salmonella Enteritidis serotype also has g, m as flageller antigen.

The results of the PCR were used to estimate diagnostic test-characteristics, viz., sensitivity, specificity, positive predictive value, negative predictive value and concordance (Thrusfield, 2005) with respect to conventional cultural method which was considered as standard method (Table 4).

Diagnostic sensitivity of PCR was 100% and specificity was 97.89% for poultry blood samples. The diagnostic sensitivity from faecal samples was 77.77%, while specificity was found to be 100%. The positive predictive values of PCR from blood and faecal samples were found to be 77.27 and 100% with a concordance of 98.03 and 99.21%, respectively. The negative predictive values from blood and faecal samples were found to be 100 and 99.19% (Table 5).

Table 4:	Comparative efficacy of PCR and conventional isolation for the detection of Salmonella from poultry blood and faeces samples

Table 5:	Diagnostic test characteristics of PCR from poultry blood and faeces samples

DISCUSSION

Salmonellosis has remained a significant public health problem causing food poisoning in humans. Poultry, its products and eggs, represents an important source of Salmonella organism for consumer health (Altekruse et al., 2006; Kabir, 2010; Barrow et al., 2012; Dhama et al., 2013). The conventional method of cultivation used in the detection of Salmonella is reliable but slow as it includes stages of pre-enrichment, selective enrichment, cultivation in selective agars, biochemical characterisation of suspected isolates and a final serological confirmation (OIE, 2008). Moreover, it may not identify all Salmonella infected flocks because of the intermittent nature of Salmonella excretion (Hassan et al., 1990; Nicholas and Cullen, 1991; Van Zijderveld et al., 1992). Skov et al. (2002) demonstrated that the number of fecal excretors declined rapidly with time in experimental chickens, down to 6% in 16 weeks for Salmonella Typhimurium and down to a similar level within the first 8 weeks for Salmonella Enteritidis. The relatively long time required to carry out analysis (4-7 days) as well as labour intensive nature of conventional identification procedure have stimulated the development of faster detection methods. Recent advances in biotechnology and molecular biology have provided molecular detection methods like of Polymerase Chain Reaction (PCR) and its allied versions which have proven to be highly sensitive, specific and provide a rapid and confirmatory diagnosis of Salmonella (Kataria et al., 2005; Menghistu et al., 2011; Batista et al., 2013; Sokolov and Sokolov, 2013). The important criteria in the development of a nucleic acid based detection assay for Salmonella is the ability to detect all the diverse serotypes of the organism and PCR has been employed to replace conventional serotyping methods. PCR-based serotypings depend on specific virulence genes and have provided high specificity (Jarquin et al., 2009).

In this study, PCR method was compared with that of conventional cultural isolation. For the purpose, 255 samples each of poultry blood and faeces, were analysed. Salmonella spp. was isolated by conventional cultural methods from 17 (6.66%) blood samples and 9 (3.52%) faecal samples. The low rate of faeces culture positive birds compared to blood was in agreement with the findings by earlier researchers (Hassan et al., 1990; Nicholas and Cullen, 1991; Van Zijderveld et al., 1992). The low isolation rate from the faecal samples may be because of the intermittent nature of Salmonella excretion (Barrow, 1992). A negative faecal culture result may not necessarily indicate that a bird is not infected (OIE, 2008). In this investigation, no records were available about antibiotic treatments of flocks which may also reduce the likelihood of cultivating Salmonella from seropositive birds as reported by Feld et al. (2000). Wide variation in prevalence rate of Salmonella in poultry has been reported. Peplow et al. (1999) reported a Salmonella prevalence of 43% from fresh preplacement samples and 61% from fresh preslaughter samples of poultry environmental samples. Freezing of samples resulted in drop of prevalence to 13% for preplacement samples and to 23% for preslaughter samples. Leon-Velarde et al. (2004) reported 5.5% Salmonella prevalence from poultry house environmental samples. It indicates that the nature and condition of sample plays an important role in recovery of organism.

Kumar (2009) reported a 2% isolation of Salmonella from poultry faeces and 1% from chicken samples. This value was lower than the isolation rate obtained in this study. However, Oscar (2004) reported a higher isolation rate of 22.2% from chicken feacal samples. Similarly, about 16-21% prevalence rate of Salmonella have been reported from chicken meat (Roberts, 1991; Plummer et al., 1995). The percentage of Salmonella-positive birds and feacal samples on farms has ranged from 5 to 100% in U.S. (Carraminana et al., 1997; Bailey et al., 2002). Tapchaisri et al. (1999) reported the prevalence rate from chicken samples by 7% with the DNA amplification method and conventional culture method.

The PCR results revealed that 22 (8.62%) blood samples and 7 (2.74%) faecal samples were positive for invA gene. Eyigor et al. (2005) reported 5.87 and 4.10% Salmonella prevalence rate in Turkey between 2000 and 2001 by real-time polymerase chain reaction and bacteriology, respectively, out of a total of 1242 samples. This prevalence rate was considerably higher than the present findings, especially with reference to faecal samples. This may be due to the presence of inhibitory substances present in the faecal matter or due to the overgrowth of the competitive natural microflora present in the faeces than Salmonella (Schrank et al., 2001).

The PCR employed for detection of Salmonella from poultry blood samples was found to be 100% sensitive, giving a positive predictive value of 77.27%, when compared with the culture method. The specificity from blood samples was found to be 97.89% by PCR. PCR was found to be 100% specific and positively predictive with regard to faeces samples, giving a concordance of 99.21%. Whereas, the two false negative results given by the PCR test from faeces samples (0.78%) directly influenced the sensitivity (77.77%). One of the major problems in using PCR for detection of pathogenic organisms from clinical and environmental samples is the presence of inhibitory substances to the polymerase reaction (Schrank et al., 2001). The negative predictive values from blood and faecal samples were found to be 100 and 99.19%. Eyigor et al. (2007) reported a PCR-ELISA to detect Salmonella DNA from selective primary enrichment culture of chicken intestinal samples using invA primers and reported 100% relative sensitivity and specificity when compared to bacteriology. Perelle et al. (2004) compared PCR-ELISA and real-time PCR assays for detecting Salmonella species in milk and meat samples by amplifying invA gene and reported 100% concordance with the bacteriological reference method. Tapchaisri et al. (1999) reported a sensitivity, specificity, efficacy and positive and negative predictive values of 100, 91.58, 92, 65.21 and 100%, respectively, when they compared the DNA amplification and the culture method. Compared to their observation, the present study values were higher, especially with respect to specificity and positive predictive values from both the blood and faeces samples.

Serotyping of the positive isolates in this study revealed 15 isolates (57.69%) of Salmonella Typhimurium, 4 isolates (15.38.76%) of Salmonella Enteritidis and 5 rough isolates (3.84%) of Salmonella rough type. Among Salmonella Typhimurium isolates, 73.33% were from poultry blood samples and 26.66% were from poultry faeces samples. Salmonella Enteritidis isolates (4) were obtained from blood samples. Salmonella Typhimurium has been reported to be most frequently associated with food poisoning, followed by Salmonella Enteritidis in Uruguay (Betancor et al., 2004, 2010). Salmonella Typhimurium was also most frequently identified from human cases in Great Britain (Cook, 2003). The present study also correlate with these reports, where Salmonella Typhimurium was the most common. However, Schneid et al. (2006) reported 88.6% of Salmonella Enteritidis isolates, out of a total of 35 Salmonella positive samples. The prevalence of Salmonella on chicken farms and colonizing birds varied considerably with the interflock studies. Poppe et al. (1992) isolated 45 and 35% of Salmonella Enteritidis from pooled feacal samples of two flocks. Culture results of individual organs from the two flocks indicated that the ceca were the predominant site of Salmonella Enteritidis infection in birds. Since clusters of Salmonella are not evenly distributed within an affected faecal mass, subsamples may not always have Salmonella present. This variation could have been reduced by homogenizing the faecal mass before obtaining subsamples (Cannon and Nicholls, 2002). In the present study, the faeces was also collected directly from the caecum and homogenised before using. Croci et al. (2004) also reported the presence of various Salmonella serotypes, viz., Salmonella Enteritidis, Salmonella Give and Salmonella Newrochelle from poultry samples. Eyigor et al. (2005) reported Salmonella Enteritidis as dominant Salmonella serovar from poultry and environmental samples.

On PCR analysis of 26 isolates for Typhimurium and Enteritidis specific genes (typh and ent, respectively), it was observed that 15 Typhimurium isolates were positive for typh gene. However, four of the rough isolates showed the presence of ent gene, indicating them to be Salmonella Enteritidis. Interestingly, these four rough isolated possessed g, m flagller factor. The rough strains lack immunoreactive O-chain (Guard-Petter et al., 1999) which may happen due to mutagenesis of the O-antigen gene cluster involved in O-antigen synthesis (Ochoa-Reparaz et al., 2005). Thus, serotype specific PCR has advantage in identifying serotypes among rough strains.

CONCLUSION

It may be concluded from the study that PCR is reliable and fast technique in detecting Salmonella in blood and poultry samples and thus can be used in prevalence studies.