Recovery of Salmonella isolated from eggs and the commercial layer farms

Background Salmonella is recognized as a common bacterial cause of foodborne diarrheal illness worldwide, and animal or its food products have been the most common vehicles of the Salmonella infections. This study aimed to investigate the distribution of Salmonella in two commercial layer farms and to determine the genetic relatedness between these strains. The Salmonella isolates were serotyped by slide agglutination using commercial antisera and analyzed for genetic relatedness using pulsed-field gel electrophoresis (PFGE). Results The internal environment had the highest prevalence of Salmonella (14/15, 93.3%), followed by external environment (60/96, 62.5%) and egg samples (23/84, 27.3%). The prevalence of Salmonella in the environment was significantly higher than that in egg samples (p < 0.05). The occurrence of Salmonella in the internal environment (93.3%) was relatively higher than in the external environment (55.6–77.2%). The 111 isolates were distributed among 15 PFGE types, and the PFGE results suggested that there existed cross-contamination between these strains not only from eggs, but also from the environments. Conclusions The findings indicated ongoing Salmonella cross-contamination inside or outside of the layer farms, and that Salmonella could also spread along the egg production line. Electronic supplementary material The online version of this article (10.1186/s13099-017-0223-8) contains supplementary material, which is available to authorized users.


Background
Salmonella enterica is recognized as one of the most common bacterial cause of human diarrheal illness worldwide, which has been a considerable burden to public health and economic loss. There are 1.4 million occurrences of human salmonellosis in the USA annually, 95% of which were foodborne [1]. Accordingly, Salmonella-related infections have been associated with $365 million in annual direct medical costs [2]. In the European Union, 99,020 food-borne cases caused by Salmonella were reported in 2010 [3]. The estimated incidence of Salmonella gastroenteritis in East Asia circa 2006 was 3980 cases per 100,000 person-years (compared to a global incidence of 1140 per 100,000 person-years) [4]. Salmonella outbreaks are commonly associated with consumption of contaminated food, such as poultry meat and eggs, which have been identified as the important vehicle for human salmonellosis [5,6].
Salmonella has been frequently recovered from laying hen house environments, suggesting that the environment of the poultry farm can act as a reservoir for Salmonella and contribute to the horizontal dissemination of Salmonella via animal-to-animal contact and contaminated feed [7][8][9]. In addition to feed, the water, feces, dust, cages and litter contaminated with Salmonella are important sources of infection [9][10][11][12][13][14][15]. Many studies focused on the distribution of Salmonella among different sample origins in poultry environments, or on antibiotic resistance, virulence, and control strategies [9,14,16,17]. However, there have been few investigations of the

Open Access
Gut Pathogens *Correspondence: zhanganyun@scu.edu.cn; zoulikou@sicau.edu.cn 1 College of Resources, Sichuan Agricultural University, Chengdu 611130, People's Republic of China 5 College of Life Sciences, Sichuan University, Chengdu 610065, People's Republic of China Full list of author information is available at the end of the article association between Salmonella isolates recovered from the internal and external poultry environment and the relationship between isolates obtained from sequential points along the production chain. Recognition of these aspects is important in controlling the spread of Salmonella and reducing the prevalence of Salmonella in production settings.
Although all serotypes may be regarded as potential human pathogens, the majority of infections are caused by a very limited number of serotypes, of which Salmonella Enteritidis and Salmonella Typhimurium are the two most common ones associated with gastrointestinal disease of humans [18,19].
There have been increasing concerns over the past 30 years about the worldwide emergence of multi-drug resistant phenotypes among Salmonella serotypes, in particular S. Typhimurium. Other non-typhoidal Salmonella serovars, such as S. Braenderup, S. Derby, S. Jerusalem and S. Bovismorbificans [15,20], have caused outbreaks but they do not frequently occur or rarely outbreaks. Several serotypes can colonize the digestive tract of chickens and be excreted in feces, which can persist in the environment and may lead to vertical and horizontal transmission in chickens, ultimately contaminating the processing chain and retail poultry products [18].
Previous studies have investigated the distribution and prevalence of Salmonella in broiler chickens and the processing environments [9,14]. In Guangdong Province, China, the most frequent serotypes isolated from raw retail poultry meat were S. Enteritidis, Salmonella Indiana and S. Typhimurium [21] and from live chickens, the most frequent serotypes were unidentified, followed by S. Typhimurium and S. Enteritidis [22]. However, very little is known about the distribution and relatedness of Salmonella from layer farms and the variation of that distribution along the production processing chain, particularly in China. Therefore, the objective of this study was to investigate the distribution of Salmonella at each particular link in the internal and external environments of two commercial layer farms and to analyze the relatedness of the prevalent strains along with the egg production chain.

Sampling farms
Sampling was conducted on two layer farms, one built in 1999 and the other in 2006, which both belong to the same commercial egg production company and are 5 km apart. The old layer farm has a capacity for 250 thousand birds, while the new layer farm can hold 500 thousand birds. Additionally, the technical characteristics of the old layer farm are different from those of the new layer farm. Due to the construction time, the facilities of the old layer farm are relatively outdated, and the new layer farm has modern farming technology with fully automated management, egg harvesting, egg washing, disinfection, ultraviolet (UV) irradiation and packaging.

Egg collection
Eighty-four egg samples were collected from three main parts of the production chain, including cages (placed at the front, middle, and posterior of the internal henhouse) of the old layer farm, egg belt (in the front, middle, and posterior of the belt) and egg collection conveyer of the new layer farm. In addition, retail eggs traced to their farm of origin were collected from the supermarket within their original packaging. All eggs were aseptically transferred into individual sterile plastic bags and transported to the laboratory for cultivation and isolation within 6 h.

Environmental sample collection
The environmental sampling sites included all major points of the egg processing line and environments, which can be divided into two main parts, the internal and external environments. These sampling sites included disinfectant system, soil, feces, dust and gutter samples of the outdoor environment, and wet curtain cooling system, cage, egg nest samples of the indoor environment of the old layer farm. Samples from the washing room, washing water, irradiation room, package room, storage room and feces of the outdoor environment, and cage, egg belt and egg collection conveyer samples of the indoor environment were collected in the new layer farm (Tables 1 and 2). A total of 126 environmental samples were collected using a sterile cotton swab. After collection, all samples were transported to the laboratory in an ice chest and stored at a 4 °C cooler for bacterial isolation within 6 h.

Salmonella enrichment
Modified enrichment methods were used based on the preliminary data on Salmonella growth from eggs described previously [18]. Briefly, a swab technique was used to sample intact egg surfaces. Swabs were directly inoculated into 10 mL Buffer Peptone Water (BPW) for pre-enrichment and incubated at 37 °C for 18-24 h. After eggshell surface processing, each eggshell surface was dipped into 3:1 solution consisting of three parts of 70% alcohol to one part iodine tincture solution for 1 min to kill any bacteria on the outside of the shell and was allowed to air dry in a biosafety cabinet. The eggs were cracked open and egg contents were placed into a sterile container. The internal eggshell was washed with sterile phosphate buffered saline (PBS) to rinse off the adhering egg albumen. The internal shell and membranes from eggs were crushed into pieces, transferred to a 250 mL sterile triangular flask with BPW and incubated at 37 °C for 18-24 h. Subsequently, mixed egg content was inoculated into BPW and incubated at 37 °C for 18-24 h. The ratio of egg shell to BPW and content sample to BPW was maintained at 1:10 by volume. The environmental samples were also pre-enriched in BPW at 35 °C for 18-24 h.

Salmonella confirmation
One millilitre of overnight culture was inoculated to 100 mL Rappaport-Vassiliadis (RV) Broth and 10-100 mL Tetrathionate Broth Base (TTB) (Beijing Land Bridge Technology Co, Ltd., Beijing, China) [23] and the inoculated broth was then incubated at 42 °C for 24 h. A full loop of each of the enrichment RV and TTB broth was streaked on Brilliant Green Sulfadiazine agar (BGS) and Xylose Lysine Desoxycholate (XLD) plates, and then incubated at 37 °C for another 24 h. Presumptive Salmonella colonies from each plate were stabbed into Triple Sugar Iron agar (TSI) and urea-agar slants (Beijing Land Bridge Technology Co, Ltd., Beijing, China) [24]. After 24 h of incubation at 37 °C. Isolates with typical Salmonella phenotypes were confirmed by PCR. The PCR assays for identification of Salmonella were previously described [25]. A 284-bp PCR product targeting invA was amplified using primers invA139 (5′-GTGAAATTATCGCCACGTTCGGGCAA-3′) and invA141 (5′-TCATCGCACCGTCAAAG GAACC-3′). Confirmed isolates were stored in TSB containing 20%

Pulsed-field gel electrophoresis
All isolates were compared using pulsed-field gel electrophoresis (PFGE) analysis according to the PulseNet protocol. The XbaI-digested DNA fragments were separated in a 1% agarose gel using a CHEF MAPPER electrophoresis system (Bio-Rad, Hercules, California, USA). The electrophoresis conditions were as previously described [26]. S. enterica serovar Braenderup H9812 was used as a marker. PFGE results were analyzed by BioNumerics software (Applied Maths, Kortrijk, Belgium), and banding patterns were compared by using Dice coefficients with a 1.5% band position tolerance.

Statistical analysis
Frequency differences among the isolates were analyzed using SPSS v.12 (SPSS Inc. 1989-2003, and the Chi square test was used to determine the significance of the differences. A p value less than 0.05 was considered statistically significant.
As shown in Table 1, four Salmonella serotypes were present among 44 Salmonella isolates in the old layer farm. S. Derby was most frequently recovered serotype (81.8%, n = 36), followed by S. Enteritidis and S. Jerusalem (both 6.8%, n = 3), and S. Bovismorbificans (2.3%, n = 1). S. Jerusalem was found in soil and fecal samples. Interestingly, S. Enteritidis was only present in the soil and S. Bovismorbificans was detected only in wet curtain system samples. One isolate with an unidentified serotype was recovered from a gutter sample.
Four Salmonella serotypes were also found in the new layer farm among the 44 Salmonella isolates, of which S. Derby was also the predominant serotype (52.3%, n = 23), followed by S. Jerusalem (25.0%, n = 11), S. Bovismorbificans (13.6%, n = 6), and S. Enteritidis (2.3%, n = 1), respectively. S. Derby was found in most samples but absent from irradiation room (UV) and storage room samples. S. Jerusalem was present in samples from the egg collection conveyer, washing room, washing water, irradiation room, storage room and feces. S. Bovismorbificans was detected in washing water and feces. Strangely, S. Enteritidis was only present in one fecal sample. Additionally, unidentified isolates were isolated from washing water, storage room and fecal samples.

Frequency of Salmonella in egg production chain
On the new layer farm, four points of the production chain were sampled (Table 3). Twenty-three of the 45 farm-level samples (51.1%), 27.3% (n = 6) at the processing level, 33.3% (n = 4) at the storage level and 50.0% (n = 6) at the retail level were positive for Salmonella contamination. Farm-level samples included those from cage, egg belt, egg (from belt), egg collection conveyer and egg (from conveyer), processing level included washing and irradiation room and package room samples, storage level included the storage room and retail level included retail eggs.

PFGE typing
To determine genetic similarity among isolates from different origins, we defined PFGE types as having similarity index equal to or greater than 75%. Overall, a total of 15 distinct PFGE types were identified among the 111 Salmonella isolates (Additional file 2: Table S2, Additional file 3: Figure S1). Interestingly, isolates of PFGE type 1 were found in diverse samples including feces, dust, cage, egg nest, egg belt and washing room. Isolates 1 (dust), 2 (feces) from the external environment of the old layer farm had high similarity to isolate 3 (cage) and 4 (egg nest) of the internal environment of the old layer farm as well as to isolates 10, 11, and 12 (cages) from the internal environment and isolates 13 and 14 (both from feces) from the external environment of the new layer farm. The same results were also found in isolate 34, 35 and 36, 37 in different environment of the old layer farm. Additionally, the isolate 23 and 24, 44 and 45 of the old layer farm and 57 and 58, 68, 69 and 70 of the new layer farm in external environment also showed highly similarity to each other. There were also existed genetically relationship between isolates 6, 7 and 8 from internal environment. The isolates 6, 7 (from egg belt) of the new layer farm and 8 (from egg nest) of the old layer farm both in internal environment samples had been found a highly similarity to each other, respectively. Besides, the isolate 27, 28 (both from feces samples) of the new layer farm had also been found genetically related to the isolate 29 isolated from disinfection room samples of the old layer farm. The same results were also found between the isolates 57 (washing water), 58 (package room) of new layer farm and 59 (soil) of old layer farm as well as 75 (storage room) and 76 (feces), respectively. Indistinguishable S. Jerusalem isolates were also from different farms as were the S. Jerusalem isolates within type 5. The same results were also found in S. Bovismorbificans in type 9.
Additionally, the isolates from egg origins were also related to the Salmonella isolates from environment samples of the two layer farms. Two S. Braenderup isolates, one (94) isolated from the retail, another (95) from egg samples of belt, had a highly similarity of more than 85%. The same results had also been found in four un-identified isolates (107 and 108, 110 and 111) with different origins.

Discussion
Human salmonellosis has been consistently associated with the consumption of poultry products worldwide [5,27]. S. Derby was most frequently observed in layer farm environments while S. Jerusalem, S. Braenderup and S. Derby were the predominant serotypes in egg samples. Derby was one of the main serotypes in the present study, but only small outbreaks have been associated with this serotype according to the reports of the CDC website of USA. This finding may be attributable to the inherent physiological characteristics of Derby which lacks pathogenicity islands 13 and 14, the fimbrial lpf operon, and other regions that encode metabolic functions [28]. S. Jerusalem isolates from a chicken farm have also been reported previously [29]. Braenderup is reportedly a major cause of outbreaks in America [30]. Compared to environmental samples, the incidence of Salmonella was lower in egg samples, which was slightly lower (16.7% in old layer farm) or consistent (33.3%) with another report of Salmonella isolation from eggshells (34%) [31]. In this study, Salmonella isolates were recovered not only from eggshells, but also from egg content. Previous studies revealed that under normal conditions of storage and moisture, Salmonella contaminating eggshells could migrate to the egg content [27], which might result in human infections.
Notably, the prevalence of Salmonella contamination in environment of the layer farm was somewhat higher than that in eggs in this and other studies [14,27]. This prevalence was also higher compared with reported prevalence in live broiler chicken samples [14]. The environment of the layer farm was considered as a reservoir for Salmonella and could contribute to the horizontal/vertical dissemination of Salmonella [14,32], since Salmonella had the ability to persist in both host and non-host environments for its enhanced survival capabilities [33].
Additionally, the incidence of Salmonella in the internal environment (93.3%) was somewhat higher than in the external environment (55.6-77.2%) of both layer farms. High similarities between these isolates were also found, which suggested that cage, egg belt and egg nest were the important reservoirs for Salmonella in the internal environments and that transmission of Salmonella occurs readily between locations in the internal environments. Furthermore, direct contact between egg belt and egg nest eggs were considered to be efficient mechanisms for the transmission of Salmonella [9,11]. Although not assessed in this study, it is plausible that insects and mice play a role as vectors of Salmonella in internal environments of layer farms [34,35]. Contaminated laying hens could spread Salmonella to nearby hens by direct contact or could disseminate Salmonella in egg forming by its reproductive tract [27]. Thus, these factors indicate complex network of potential cross-contamination of Salmonella in the internal environments of layer farms.
There also was high similarity among Salmonella isolates from the external environment. Feces, dust, water, and soil were the main source reservoirs for Salmonella in the external environments. Feces played an important role in Salmonella dissemination, as contaminated feces excreted into the environment could then be a source of the bacteria to naive hosts, perpetuating its survival over the layer farm environment [9,11,14,32]. Salmonella was also detected in rinsing water for egg washing. Contaminated washing water flowed along with the gutter and may be used for irrigation water, which could be a major route of Salmonella contamination for crops and produce [36,37]. Dust has also been considered as a vector for Salmonella spread through potential airborne transmission [34]. Salmonella in dust could also contaminate pelleted feed [34,38]. Contaminated soil could act as a persistent source of Salmonella difficult to disinfect [34,37]. A moist floor associated with daily rinsing with water for cleaning and the spillage of water from the drinkers could provide favorable condition for survival of Salmonella [14].
The difference in prevalence of Salmonella in the external environment of the new layer farm to that of the old layer farm (55.6% [30/54] versus 71.4% [30/42]) supports the use of newer farming technology as helpful for controlling Salmonella contamination. The lower frequency of Salmonella contamination in the disinfection system showed that disinfectant application and washing of eggs was contributing to preventing or reducing bacteria. To prevent Salmonella contamination in external environment, litter, feces and dust should be removed frequently and disinfectant applied to surfaces.
Salmonella isolates from internal and external environmental samples were also highly similar to each other. This might be explained by cross contamination between internal and external environments when birds or feces are removed [39]. Dust and waste drinking water also have the potential ability to spread Salmonella by airborne and waterborne transmission. Other vectors, such as mice, insects and wild birds could introduce Salmonella from external to internal environment [40,41].
The prevalence of Salmonella changed dynamically along the egg production chain. The farm level had the highest prevalence of Salmonella, followed by retail level, storage and processing level. The reduced recovery of Salmonella at processing level might be owing to the strategies for prevention of egg contamination. Egg washing and disinfection were efficient ways to wash bacteria off of egg surfaces. Sealed packages also prevented contamination of eggs. However, during prolonged storage period, the risk of Salmonella contamination may increase and result in decreased quality of the egg products. Isolates from individual eggs were genetically similar to each other, which suggested that the Salmonella contamination of eggs at the farm level could persist into the retail level. Salmonella has been confirmed as having the capacity to colonize the reproductive tract of the laying hens and thereby contaminating forming eggs [34,42,43]. We found that isolates from different parts of the production chain were highly similar. Isolates from three different parts of the production chain (farm level, egg production processing part and retail level) had more than 85% similarity. The same results had also been found in isolates 94 (retail level) and 95 (farm level), 110 (retail level) and 111 (storage level). These results indicated that the pathogens could spread along with the poultry breeding to and production chain.
Cross-contamination between two layer farms was also evident from the PFGE results. The most probable explanation could be that breeder chickens upstream of the production chain contaminated both farms. The two layer farms had a same origin of layer chicks, which could introduce Salmonella [44]. Contaminated layer hens could transmit Salmonella to the forming egg within the reproductive tract [27], which may be the reason the two farms had the same frequency of Salmonella contamination (93.3%) in internal environment. Another essential factor might owe to the exchange workers, equipment or managers between two layer farms. Humans and equipment as mechanical vectors could introduce Salmonella to each other indirectly [34,39].
The findings presented herein indicated that there was a significant difference in contamination of Salmonella serotypes among egg samples and environmental samples. S. Enteritidis was absent in egg samples but present in environmental samples, which was different from previous studies [14,18]. In general, S. Enteritidis was confirmed as strongly associated with shell eggs and egg containing products [45]. The results might be due to strategies applied at the feeding and production processing line, such as disinfection, washing and UV radiation. Disinfection and UV radiation have the capacity to reduce or kill microorganisms on egg shell surfaces [34,46,47]. Egg washing was also used to reduce the bacterial contamination and to prevent penetration of bacteria to the egg contents [34]. Additionally, S. Braenderup was only present in egg samples, but not in environmental samples. Serotype-specific characteristics may explain their own niche preferences within poultry environments [48,49].
This study showed that Salmonella contamination is common in the layer farms that we studied. S. Derby was most frequently observed in layer farm environments while S. Jerusalem, S. Braenderup and S. Derby were the predominant serotypes in egg samples. The prevalence of Salmonella in environment of the layer farm was higher than that in egg samples. The incidence of Salmonella in internal environment was relatively higher than in external environment in both layer farms. Salmonella could be disseminated not only between internal and external environment, but also between different layer farms. It could also spread along the egg production processing chain. Measures, such as cleaning and disinfection routinely etc., should be taken to prevent or reduce the dissemination of Salmonella in layer farm environment.

Conclusions
The findings indicated ongoing Salmonella cross-contamination inside or outside of the layer farms, and that Salmonella could also spread along the egg production line.