Characterization of Salmonella spp. isolated from small turtles and human in Republic of Korea

In 2013, the World Health Organization (WHO) reported that small, pet turtles had caused multistate Salmonella outbreaks in the United States, from where small turtles were subsequently exported into the Republic of Korea. We investigated cases of salmonellosis in South Korea associated with domestic small turtles and analysed genetic characteristics of Salmonella isolates in commercially-available small turtles. We traced six Salmonella serovars, known to have caused human infection in the United States (S. Sandiego, S. Pomona, S. Poona, S. Newport, I 4,(5),12:i:-, and S. Typhimurium), in isolates from suspected Salmonella infection cases in Korea from 2006 to 2015. Additionally, we conducted a pilot study of isolates from small turtles being sold in Korean markets, and performed molecular genetic analysis on the identified strains. S. Pomona was identified in one Salmonella infection case, while all strains isolated from small turtles belonged to either subspecies I (enterica, n = 10, 71.4%) or subspecies IIIb (diarizonae, n = 4, 28.6%). Two serovars (S. Pomona and S. Sandiego) that were highly associated with turtle-to-human transmission were identified with 100% homology to human isolates. Previous to this study, turtle-associated human S. Pomona infections were not well reported in Korea. We report Salmonella infection in small turtles in Korea, and confirm that small turtles should be considered the first infectious agent in S. Pomona infection. We therefore suggest quarantine measures for importing small turtles be enhanced in Korea.


Introduction
Salmonella spp. is a zoonotic pathogen often reported in humans where there was exposure to infected pets and livestock products [1][2][3]. Salmonella colonize the intestine of turtles, which are the main source of human infection due to shedding of bacteria in faeces [4]. In the United States, 15 outbreaks of Salmonella associated with small turtles were reported between 2006 and 2014 [5], with most patients exposed to small turtles (less than 4 inches long) raised on two farms in Louisiana. Thirty one percent of the patients were children under 10 years, of which 70% were under 1 year. In 2013, in accordance with the International Health Regulations, the World Health Organization (WHO) noticed the risk of Salmonella infections to 20 countries, including South Korea that imported turtles from the United States. Therefore, we conducted an epidemiological and microbiological survey on the risk of Salmonella infection from imported small turtles in Korea.

Abstract
In 2013, the World Health Organization (WHO) reported that small, pet turtles had caused multistate Salmonella outbreaks in the United States, from where small turtles were subsequently exported into the Republic of Korea. We investigated cases of salmonellosis in South Korea associated with domestic small turtles and analysed genetic characteristics of Salmonella isolates in commercially-available small turtles. We traced six Salmonella serovars, known to have caused human infection in the United States (S. Sandiego, S. Pomona, S. Poona, S. Newport, S. I 4,(5),12:i:-, and S. Typhimurium), in isolates from suspected Salmonella infection cases in Korea from 2006 to 2015. Additionally, we conducted a pilot study of isolates from small turtles being sold in Korean markets, and performed molecular genetic analysis on the identifi ed strains. S. Pomona was identifi ed in one Salmonella infection case, while all strains isolated from small turtles belonged to either subspecies I (S. enterica, n = 10, 71.4%) or subspecies IIIb (S. diarizonae, n = 4, 28.6%). Two serovars (S. Pomona and S. Sandiego) that were highly associated with turtle-to-human transmission were identifi ed with 100% homology to human isolates. Previous to this study, turtle-associated human S. Pomona infections were not well reported in Korea. We report Salmonella infection in small turtles in Korea, and confi rm that small turtles should be considered the fi rst infectious agent in S. Pomona infections. We therefore suggest quarantine measures for importing small turtles be enhanced in Korea.

Case study on human Salmonella infection
We examined six Salmonella serovars (S. Sandiego, S. Pomona, S. Poona, S. Newport, S. I 4, [5], 12:i:-, and S. Typhimurium) to investigate the possibility of Salmonella spp. infections through small turtle in Korea. We analysed the prevalence of the serovars in isolates from Salmonella infection cases in Korea during the last ten years. The six serovars were classi ied for inclusion in the epidemiological study according to three criteria: 1) the serovar was not reported in Korea before, 2) the serovar was only reported in the last decade, and 3) less than ive isolates of the serovar were reported per year. Serovars not itting any of the criteria were deemed candidates most likely transmitted from small turtles, and the epidemiological survey was then conducted on these.

Samples from domestic turtles
Five small turtles were purchased from markets in four https://doi.org/10.29328/journal.ivs.1001027 randomly selected regions of Korea. Turtles were bred over 5 weeks in separate tanks to prevent cross-contamination, and breeding water was changed every 2-3 weeks. Samples from breeding water were collected weekly to analyse the colonisation of Salmonella in the turtles' intestines. In addition, samples of breeding water from turtles on Jeju Island (Jejusi, Seogwipo-si) were collected weekly, as those turtles were associated with suspected human infection previously

Isolation and identifi cation of Salmonella spp. strains
Isolation and identi ication of Salmonella strains was performed following the Korea Centers for Disease Control and Prevention (KCDC) laboratory guidelines. 1 mL breeding water was inoculated into 9 mL Selenite Cysteine Broth (BD Diagnostic Systems™, Sparks, MD, US). The inoculum was then incubated overnight and a loopful of sample plated on MacConkey agar, Salmonella-Shigella agar, and Xylose Lysine Deoxycholate agar (BD™). The unique colonies were streaked on Kligler Iron agar (BD™), and their identi ication was performed using a biochemical test with API 20E Kit (Biomerieux™, Marcy-1 Etoile, France) and serological testing done as per Lee, et al. [6]. In addition, invA speci ic PCR was performed to con irm the microbiological identity [7]. Genomic DNA was extracted using the boiling method [8] and ampli ied by singleplex PCR with speci ic primers for the invasion of Salmonella into eukaryotic host cells (invA F; 5´-ATT AAT TAT GGA AGC GCT CGC ATT-3´, invA R; 5´-GTA ATG AGA TCC ATC AAA TTA GCG-3´). Serovars were determined by the combination of somatic O and lagellar H antigens according to the Kauffmann-White classi ication scheme [9,10].

PFGE, MLST & antimicrobial susceptibility of Salmonella strains
To investigae the epidemiological relationship, S. Pomona and S. Sandiego isolates (Group IV in our classi ication) were analysed by pulsed-ield gel electrophoresis (PFGE) and multilocus sequence typing (MLST) to compare with human isolates. PFGE was performed according to the PulseNet standard protocol (http://www.pulsenetinternational.org/ protocols/). DNA of each isolate was digested with XbaI and BlnI restriction enzymes and separated by CHEF Mapper XA system (Bio-Rad, CA, USA). Salmonella Braenderup (ATCC BAA664) was used as a reference marker strain. PFGE patterns were analysed using BioNumerics v5.1 software (Applied Maths, Saint-Martens-Lartem, Belgium). The isolates were subjected to MLST by determination of the sequences for seven housekeeping genes: aroC, dnaN, hemD, hisD, purE, sucA, and thrA [11]. Sequencing of the nucleotides was performed with an automated sequencer (ABIPRISM 3730XL, Foster City, CA, USA), and the sequence type was determined using analysis tools on the MLST website (http://mlst.warwick.ac.uk/mlst).

Case study on human Salmonella infection
A total 5,867 Salmonella isolates were collected from diarrheal patients in Korea from 2006 to 2015 (Table 1). A total of 1,144 strains were identi ied belonging to six serovars which were thought to be related with small turtles, and were divided into four groups according to classi ication criteria ( Figure 1). No isolates of Group I (Gr. I), S. Pomona, have been reported in Korea, and this serovar was therefore excluded from the case study. Group II, (Gr. II) including the S. Typhimurium (n = 745) and S. Newport (n = 47) serovars, were excluded from the study as there were too many cases to link logistically to small turtles. Group III (Gr. III) isolates, from the Salmonella I 4, [5],12:i:-, serovar (n = 346), were also excluded; cases of this serotype, a variant strain of S. Typhimurium, have been continuously increasing since it was irst reported in 2008. In addition, Gr. II and Gr. III isolates were excluded for the variety of host ranges. Therefore, Group IV (Gr. IV) isolates, S. Sandiego (n = 4) and S. Pomona (n = 2) serovars, were remaining as possible candidates to link with transmission from turtles; isolates of these serotypes have been reported separately within the last ive years, with no more than ive isolates annually. However, were able to identify just a single suspected case associated with S. Pamona.

A case of S. Pomona infection in Korea
A two-year-old boy began to show symptoms of fever on July, 7. 2013. He visited the emergency room two days later with symptoms of fever, abdominal pain, and diarrhoea. The patient returned home after receiving luid therapy in combination with antispasmodic, sedative, anti-in lammatory, and analgesic therapies. Soon after, he visited the nearby paediatric and juvenile clinic, displaying the symptoms of fever, abdominal pain, and watery diarrhoea again, along with vomiting, poor oral intake, and weight loss (14.8-16 kg). In the evening, he was re-admitted to the emergency room and received intravenous UBACSIN treatment (Jeilpharm, Korea). He remained febrile and had to be hospitalized. Within two days of hospitalization, treatment was changed to cefotaxime and symptoms improved. In microbiological examination, non-typhoidal Salmonella was isolated from patient's faecal samples and serotyping identi ied S. Pomona (M group, 28:y,1,7). During the epidemiological investigation, the patient's parents reported that the child had begun to experience symptoms after touching a turtle while visiting a neighbour. Samples from the neighbour's turtle could not be analyzed as the owner refused to allow testing on the animal, but we purchased two turtles from a nearby traditional market to investigate Salmonella infection in turtles in the relevant area.

Isolation and identifi cation of Salmonella spp. from domestic turtles
A total 14 Salmonella strains were isolated from ive small turtles purchased at randomly chosen commercial markets and two turtles purchased at a traditional market in an area where a patient may have contracted Salmonella from a domestic turtle. Ten of the 14 strains belonged to subspecies

PFGE, MLST, and antimicrobial susceptibility testing
We conducted PFGE and MLST for the characterization of S. Pomona and S. Sandiego which were commonly identi ied through risk assessment and bacterial isolation from turtles. For S. Pomona, human (n = 2) and turtle isolates (n = 5) showed 100% homology on PFGE, while MLST showed the same sequence type as the ST451 strain ( Figure 2). In contrast, S. Sandiego was divided into four clusters as a result of PFGE performed on human (n = 4) and turtle isolates (n = 1). However, human strains (year 2012) and suspected turtle- related strains showed 100% homology on PFGE and the same sequence type as the ST20 strain on MLST (Figure 3). All Salmonella isolates from turtles and humans were susceptible to the 17 antimicrobial agents tested.

Discussion
Salmonella enterica subspecies enterica (I) are mainly isolated from warm-blooded animals; while other subspecies (II, IIIa, IIIb, IV and VI) and Salmonella bongori (V) are mainly isolated from cold-blooded animals and from the environment. However, most isolates from turtles in this study were identi ied as subsp. enterica. In recent years, subsp. enterica has been identi ied in a large number of Salmonella infection cases associated with reptiles [13][14][15]. Generally, turtles carry Salmonella spp. in their intestinal tracts, and continuously release the bacteria through faeces, causing human infection indirectly. Salmonella, which is asymptomatic in reptiles, can cause serious illness in humans. Reptile-associated salmonellosis (RAS) can cause sepsis, meningitis, and bone and joint infections, especially in children [14]. Transmission from turtles accounts for 42% of RAS infections, and children are more affected because they often interact with domestic turtles.
In the United States, at least 473 cases of Salmonella infection due to small turtles were reported from March 2012 through October 2013 [16]. The cases were linked to turtles raised on two farms in Louisiana and exported to more than 20 countries. According to statistics from the Korea Customs Service, 28.76 tons of turtles were imported into the Republic of Korea annually over the past sixteen years, of which 20.17% were imported from the United States [17]. It is therefore concerning that turtles can be imported into Korea without quarantine, making it dif icult to conduct epidemiologic studies to link turtles to cases of disease because the animals are often imported directly from the United States on a small scale.
No cases of Salmonella infection caused by small turtles have been reported in Korea as yet. However several papers reported a possible case of suspected transmission from turtles [18] with cases of rare serovars [19,20]. In the case identi ied here, it was epidemiologically certain that the patient was infected through contact with a turtle, but it was not con irmed experimentally. Additionally, the identi ied serovar, S. Pomona was genetically identical in isolates from the patient and turtles purchased from a regional traditional market. Moreover, S. Pomona strains isolated from additional local patients and further local small turtles were genetically identical to the patient's strain. The case is also similar to two cases previously reported in children younger than 5 years who required treatment [21,22], and we suggest this is the irst con irmed case of Salmonella infection through turtle in Korea.
Antimicrobial agents are used to reduce Salmonella infection in reptiles, which may cause antibiotic resistance [23]. Salmonella isolates from reptiles, especially from turtles, were identi ied to exhibit a variety of multidrug resistance [24]. All isolates in this study were susceptible to the 17 agents used and no antimicrobial resistance genes were detected. However, ampicillin-sulbactam treatment against S. Pomona in the identi ied case did not relieve clinical symptoms, and symptoms only improved after treatment was changed to the cephalosporin, cefotaxime, similar to Kim, et al. [18] who recommended that a third-generation cephalosporin be used in cases of suspected non-typhoidal Salmonella bacteraemia with resistance to ampicillin. This may mean experiencebased antimicrobial therapy is effective when Salmonella is transmitted from small turtle to human.
Small turtles available in the Korean domestic market have Salmonella in their intestines. Although Salmonella infection does not often cause serious illness, high-risk groups such as the immune-compromised, children under 5 years, and pregnant women are advised to avoid contact with small turtles. In addition, quarantine of exotic pets such as turtles should be strengthened as domestic imports continue to increase.

Ethics statement
The authors con irm that the ethical policies of the journal, as noted on the journal's author guidelines page, have been adhered to. No ethical approval was required.

Funding
This research did not receive any speci ic grant from funding agencies in the public, commercial, or not-for-pro it sectors.

Authors' contributions
Su-Jin Chae was responsible for writing the original draft. Jin-Suk Lim was responsible for the methodology and investigation. Deog-Yong Lee was responsible for the conceptualization and supervision of the study.