|Home | About | Journals | Submit | Contact Us | Français|
We isolated 326 Yersinia enterocolitica strains from 5,919 specimens from patients with diarrhea at outpatient clinics, livestock, poultry, wild animals, insect vectors, food, and the environment in the cities of Nantong and Xuzhou in Jiangsu Province, China, from 2004 to 2008. The results showed that the 12 pathogenic strains were of the O:3 serotype. Six strains were isolated from domestic dogs (Canis familiaris) belonging to farmers and were found to be the primary carriers of pathogenic Y. enterocolitica strains, especially in Xuzhou. Pulsed-field gel electrophoresis analysis of the pathogenic strains from dogs belonging to farmers showed that they shared the same patterns as strains from diarrhea patients isolated in 1994. This indicates that the strains from domestic dogs have a close correlation with the strains causing human infections.
The pathogenesis of Yersinia enterocolitica, a food-borne pathogen (5) that causes various enteric and systemic syndromes (3), is mediated by virulence factors encoded on the chromosome and plasmids (2, 4, 12, 15, 16). Infection with Y. enterocolitica causes a broad spectrum of self-limiting clinical manifestations, such as diarrhea, enteritis, enterocolitis, and mesenteric lymphadenitis. Of concern are the complications (e.g., septicemia, which is usually lethal) and sequelae of Y. enterocolitica infections, such as reactive arthritis, erythema nodosum, and infected mycotic aneurysms (1, 3). Recently, Y. enterocolitica has become of concern worldwide, and infections have been reported in hundreds of countries. Y. enterocolitica outbreaks have occurred in Finland, Japan, the United States, and Brazil (9, 22, 23). Two outbreaks causing more than 500 infections were reported in China in the 1980s (24). Yersinia enterocolitica has a broad animal reservoir, with swine being the most common source of infection for humans (10, 11, 13, 18). Y. enterocolitica is distributed among a diverse range of animals in China. Strains are isolated from more than 10 kinds of animals, with pigs, cattle, goats, dogs, and mice being the more common sources. These findings are similar to those of studies conducted in other countries. Swine are the primary source of human infection with Y. enterocolitica throughout China; however, dogs belonging to farmers may be a potential source of human infection with pathogenic Y. enterocolitica in Jiangsu Province, as determined in an investigation of livestock and poultry from that area whose results are reported here. The threat for human infection from dogs may be more serious than that from swine, as dogs have more opportunities to come into contact with humans.
An investigation aimed at evaluation of the topography of Y. enterocolitica in the cities of Nantong and Xuzhou of Jiangsu Province, China, from 2004 to 2008 was performed. We collected 4,919 specimens of feces from patients with diarrhea who visited outpatient clinics at hospitals, fresh feces from domestic and wild animals, insects, and raw and cooked meats using an Y. enterocolitica isolation method reported previously (5, 24). Strain enrichment was performed by using phosphate-buffered saline with sorbitol and bile salts (PSB) at 4°C for 21 days, and then the strains were inoculated onto Yersinia selective agar (cefsulodin-Irgasan-novobiocin [CIN] agar; Difco). Colonies with a typical bulls-eye appearance (deep red centers surrounded by an outer transparent zone) on CIN agar were inoculated onto Kligler iron and urea medium. Yersinia enterocolitica strains were preliminarily identified by the use of biochemical tests, i.e., fermentation of glucose (no gas) and hydrolysis of urea.
Reference strains Y. enterocolitica O:3 strain 52203 (Ye134) and O:9 strain 52212 (Ye383) were purchased from the Institute for Chinese Medicine and Biological Standardization. Y. enterocolitica O:3/3 strain 9401 and O:9/3 strain Liuzhongli were isolated from the feces of diarrhea patients from Xuzhou in 1994 and 1999, respectively. Fifteen pathogenic Y. enterocolitica strains were isolated from dogs in other areas of China (Ningxia, four strains; Jilin, five strains; and Henan, six strains).
The isolates were serotyped by slide agglutination. Commercial serotype identification test kits for Y. enterocolitica were purchased from D. Seiken Co., Ltd., Japan, and the Institute of Chinese Biomedicine. The biotypes of all the strains were identified by using the scheme reviewed by Bottone (1) (Table (Table11).
The relative virulence genes (ail, ystA, ystB, virF, and yadA) from the chromosome and plasmids were amplified by the methods described previously (21, 24). Pathogenic strains were positive for ail (ail+), ystA (ystA+), virF (virF+), and yadA (yadA+); but some pathogenic strains lost the plasmid virulence genes (ail+, ystA+, virF negative [virF−], and yadA negative [yadA−]) in the course of bacterial culture. Some strains that did not carry the ail, ystA, virF, and yadA genes on the chromosome but that did carry on the chromosome the ystB gene, which encodes the heat-stable enterotoxin B, similar to YstA, were named ystB+ strains (ail negative, ystA negative, ystB+, virF−, and yadA−).
The pulsed-field gel electrophoresis (PFGE) method was used to analyze all of the pathogenic strains isolated in this study by use of the procedures described by Wang et al. (24). The plugs digested with 25 U NotI were electrophoresed with a pulse time of from 2 s to 20 s, and those digested with 30 U FseI were electrophoresed with a pulse time of from 2 s to 35 s; both types of plugs were electrophoresed for 18 h at 20 V. For data analysis, .tiff images of the gels were imported into the database of the PFGE patterns of Y. enterocolitica strains from China. We performed a cluster analysis for serotypes O:3 and O:9. The patterns of those pathogenic strains isolated from the Xuzhou and Nantong areas were compared with those of two strains from patients in the Xuzhou area and other pathogenic strains in China, especially those pathogenic isolates from dogs.
The PulseNet Gel Analysis Guidelines (PND04) were used for analysis of the PFGE gel images. The isolates were tested twice when the faint bands were inconsistent on the repeat test with the isolate, and bands whose intensities were less than the faintest band on the standard were not included in the analysis. Clustering of the band patterns was performed with BioNumerics software (version 4.6) and by use of the unweighted-pair group method using average linkages (UPGMA) with the Dice coefficient and 1.5% tolerance. All patterns were visually inspected after computer analysis. The patterns identified as indistinguishable by computer and visual inspections were assigned a pattern designation.
We used an API 20E commercial rapid identification system, which is accurate for the identification of presumptive isolates of Y. enterocolitica. We found the 326 isolates listed in Table Table2.2. Y. enterocolitica strains were isolated from diarrhea patients, animals, food, and insect vectors in Xuzhou, where the rate of occurrence was higher in pigs (5.87%) and dogs (5.48%); but Y. enterocolitica strains were isolated from animals and insect vectors in the Nantong area, where the rates of occurrence were high in all animal hosts. The sample isolate collection and strains are shown in Table Table22.
In this study, the primary serotypes isolated were O:8 and O:6,30 in Xuzhou and O:12,26, O:7,8, and O:5 in Nantong, where O:3 was the primary serotype isolated from dogs and swine. The serotype distributions of 326 Y. enterocolitica strains isolated from Xuzhou and Nantong are shown in Table Table33.
Y. enterocolitica serotypes O:3 and O:9 are the primary pathogenic serotypes in China. Of 13 O:3 isolates from Xuzhou, 12 were pathogenic and 1 strain isolated from a goat in 2008 was nonpathogenic. Twelve Y. enterocolitica strains, six from dogs, five from swine, and one from a goat in Xuzhou, were identified as pathogenic strains by the detection of the virulence genes (ail, ystA, virF, and yadA) (Table (Table4).4). There were no pathogenic strains isolated from Nantong.
Among the 326 isolates, biotype 1A was the primary biotype isolated. In Xuzhou, all pathogenic strains were biotype 3 and most of the nonpathogenic strains (111/116, 95.7%) were biotype 1A, except that 1 strain was biotype 2, 3 were biotype 3, and 1 was biotype 5. In Nantong, the rate of occurrence of biotype 1A strains was 77.8%; and other strains were biotype 2 (16 strains), biotype 3 (26 strains), biotype 4 (1 strain), and biotype 5 (1 strain) (Table (Table55).
A portion of the nonpathogenic biotype 1A strains carried ystB. The proportions of ystB+ strains among all isolates from the different hosts in Xuzhou and Nantong are shown in Fig. Fig.1.1. ystB was carried by 81.25% of the isolates from patients and 58.33% of the isolates from dogs in Xuzhou. The rates of Y. enterocolitica carriage were high in the various animal hosts in Nantong; however, no pathogenic strains were found.
All 12 pathogenic strains obtained during the investigation were found in Xuzhou, Jiangsu Province, with 6 strains being isolated from dogs, 5 strains being isolated from pigs, and 1 strain being isolated from a goat. In the database of the PFGE patterns of Y. enterocolitica strains in China, all pathogenic strains from the dogs, one from a goat, and most strains from swine in Xuzhou had indistinguishable PFGE patterns, which was the K6GN11C30021 pattern; and that pattern was different from the PFGE pattern of one serotype O:3 pathogenic strain isolated from a pig in 2007. One O:3 pathogenic strain from a diarrhea patient in Xuzhou, Jiangsu Province, isolated in 1994 and some O:3 pathogenic strains isolated from dogs near Suixian (in Henan Province) had indistinguishable PFGE patterns (pattern K6GN11C30021), whereas their PFGE patterns were different from the PFGE patterns of strains which were isolated from dogs in other provinces (Fig. (Fig.2).2). This shows that K6GN11C30021 is one of the primary PFGE patterns of O:3 pathogenic strains in China. Most of the strains sharing this pattern were distributed in Xuzhou, Jiangsu Province, during the past 10 years and spread among animals. One O:9 pathogenic strain isolated from a patient in Xuzhou had the K6GN11C90007 pattern. We also used FseI digestion to confirm the reliability of the patterns, and the result was consistent with that obtained by digestion with NotI (Fig. (Fig.33).
The distribution of pathogenic Y. enterocolitica strains has a close correlation with geographic and climate factors but with evident regional disparity. The distribution of pathogenic Y. enterocolitica strains in China (25) is similar to that in other countries (8, 14). This was confirmed by using epidemiology and laboratory isolation, which found that typical pathogenic Y. enterocolitica strains carrying the ail, ystA, yadA, and virF virulence genes are found in the cold northern city of Xuzhou in Jiangsu Province and are only of serotypes O:3 and O:9, whereas no pathogenic Y. enterocolitica strains were found in the warmer southern coastal city of Nantong in Jiangsu Province. No pathogenic strains were isolated by the research team of Enshu Yu in Nantong in the 1980s, and the distribution of serotypes was similar to our previous findings (E. Yu et al., unpublished data). It is interesting that Y. enterocolitica O:8, a common and important pathogenic serotype in the United States, is nonpathogenic in China, even though many O:8 strains were isolated from humans and animals (24).
The Y. enterocolitica ystB gene encodes a heat-resistant enterotoxin similar to YstA (17). Strains carrying ystB are considered nonpathogenic and lack the pathogenic marker genes ail, ystA, yadA, and virF encoded on the chromosome and plasmid pYV. Some researchers believe that these strains are pathogenic for humans and can cause local outbreaks (19, 20). A total of 157 strains carrying the ystB gene were isolated during this study, and a high proportion of the isolates from diarrhea patients in Xuzhou belonged to biotype 1A. Our results show that it is difficult to detect pathogenic strains with these characteristics compared to the ease of detection of typical pathogenic strains. We consider strains without these virulence genes to be unable to infect hosts, as they cannot break through the immune barrier; therefore, the enterotoxin encoded by ystB cannot induce pathogenesis. This requires further study.
It is well-known that swine are the primary source of pathogenic Y. enterocolitica strains throughout the world, and this is also the case in most provinces of China. However, the investigation in Xuzhou, Jiangsu Province, during 2004 and 2008 showed that dogs belonging to farmers may be a source of pathogenic Y. enterocolitica infections in these areas (Table (Table4).4). Y. enterocolitica is extensively distributed among animals in the two areas investigated, where most strains are nonpathogenic. Some pathogenic serotype O:3 strains were isolated from dogs and swine belonging to farmers.
The primary PFGE patterns of the serotype O:3 strains in the Chinese Y. enterocolitica isolate bank (with more than 10 strains) were K6GN11C30012 (54 strains), K6GN11C30021 (31 strains), K6GN11C30015 (23 strains), and K6GN11C30016 (11 strains). Strains isolated from dogs belonging to farmers shared the same PFGE pattern (pattern K6GN11C30021), showing that they may have come from the same clone. Strains isolated from humans and livestock in Henan Province and the city of Tianjin also had the same PFGE pattern (pattern K6GN11C30021) (Fig. (Fig.2).2). Strains isolated from dogs in recent years have the same PFGE patterns as strains isolated from diarrhea patients in 1994, indicating that the strains have persisted for more than 10 years in humans and dogs. Another restriction enzyme instead of NotI, FseI, was chosen for use in PFGE analysis of these strains, and the results confirmed that the six strains isolated from dogs had the same PFGE patterns as strains isolated from diarrhea patients in 1994 (Fig. (Fig.3).3). This suggests that Y. enterocolitica strains from humans and dogs belonging to farmers are closely correlated and provides molecular evidence that the same pathogen transmitted between humans and dogs came from the same source, e.g., pigs. Domestic dogs (Canis familiaris) belonging to farmers may be infected through pigs or other infected livestock, and the source of the organism for the farmers is thus the dogs belonging to those farmers because of the close contact of dogs with people. As the PFGE patterns of isolates from both sources were indistinguishable, we considered the transmission of the pathogen to be from animals (domestic dogs belonging to farmers) to humans, but it also may be possible that the reverse route of transmission, that is, from humans to animals, occurs. Fredriksson-Ahomaa et al. reported that raw pork can be an important source of Yersinia enterocolitica 4/O:3 infections in dogs and cats, and these pets may be one source of human infection, especially in young children (6, 7).
The primary PFGE patterns of serotype O:9 strains in the Chinese Y. enterocolitica isolate bank were K6GN11C90008 (58 strains), K6GN11C90010 (54 strains), K6GN11C90018 (47 strains), and K6GN11C90004 (25 strains). The pattern of the O:9 strain from a diarrhea patient in 1999 from Xuzhou was K6GN11C90007. At the same time, one O:3 strain isolated from a pig had a different PFGE pattern. These results show the polymorphism of the pathogenic strain distribution and that more transmission strands exist in Jiangsu.
Unfortunately, Y. enterocolitica pathogenic strains have not been isolated from specimens from outpatient clinics. We speculate that the reason is that the symptoms of diarrhea caused by Y. enterocolitica are often mild and most cases are self-limiting. Farmers with mild diarrhea seldom visit outpatient clinics due to poverty, and it is too late for strain isolation when the patients show complications and visit the hospital. Additionally, it is rare for humans to be infected through food due to Chinese cooking and eating habits, even though many domesticated animals in China carry pathogenic Y. enterocolitica isolates. Chinese seldom consume raw food, and when they cook food, they completely kill any Y. enterocolitica isolates that may be present. Our surveillance program for Y. enterocolitica was adjusted in 2009. The investigation includes households in the community and countryside, and fecal samples from patients with mild diarrhea are used to isolate Y. enterocolitica to obtain a more complete understanding of Y. enterocolitica infections among the local population.
In conclusion, our data show that pathogenic Y. enterocolitica strains are of serotype O:3, and the distribution of pathogenic strains showed a notable regional disparity, with such strains primarily occurring in the northern cold city of Xuzhou, similar to the case in other countries (8, 14). Although it is difficult to determine the pathway of transmission of the pathogen between human and domestic dogs (Canis familiaris) belonging to farmers, resulting in the zoonotic disease of yersiniosis, some strains isolated from human and dogs in Xuzhou had the same subtype. Therefore, we believe that dogs belonging to farmers may be another primary carrier of pathogenic Y. enterocolitica in local communities and may be a potential source for human infections.
We thank Jim Nelson for critical reading of the manuscript.
Published ahead of print on 24 February 2010.
‡The authors have paid a fee to allow immediate free access to this article.