Reptile Reservoirs and Seasonal Variation in the Environmental Presence of Salmonella in an Island Ecosystem, Stephens Island, New Zealand

The transmission of pathogens and expression of disease in wild animal populations is a complex interaction of host, pathogen, and environmental factors (Spurgin et al. 2012). Measuring the natural levels of infection in wildlife populations and determining how infection varies between sympatric species is a fundamental step in understanding the potential impacts of pathogens in host populations (Spurgin et al. 2012). Pathogenic, Salmonella species are distributed worldwide and affect animals across a broad taxonomic spectrum. The clinical outcomes of Salmonella infections are dependent on host immunity, pathogenicity of the Salmonella serotype, and environmental stressors, and many animals act as asymptomatic carriers that shed the bacteria intermittently (Pfleger et al. 2003).

Stephens Island is home to the largest density of the New Zealand endemic reptile, tuatara (Sphenodon punctatus) with numbers estimated up to 50,000 (Newman 1987). An estimated one million fairy prions (Pachyptila turtur), a small seabird, use the island as a breeding ground (Harper 1985). Tuatara and fairy prions share burrows during the fairy prion breeding season (October–January), but tuatara will also prey on fairy prion eggs, chicks, and occasionally adults (Newman 1987). The 150-ha island is also habitat for four species of native skinks; Oligosoma lineoocellatum, Oligosoma infrapunctatum, Oligosoma nigriplantare polychrome, and Oligosoma zelandicum. Several other seabirds, such as the sooty shearwater (Puffinus griseus) and little blue penguin (Eudyptula minor), as well as introduced and native passerines also inhabit the island, but none interact directly with tuatara.

In New Zealand, Salmonella has been isolated from many species of native and introduced lizards (Middleton et al. 2010; Kikillus et al. 2011), birds, and domestic animals (Clark et al. 2002). Salmonella in populations of native skinks on Stephens Island was found at a test prevalence of 4.0% (estimated true prevalence 0–12%; Middleton et al. 2010). Three serotypes of Salmonella were isolated from skinks during the 2010 study Salmonella Saintpaul, Salmonella Typhimurium DT135, and a serovar of antigen determination 4,12:-1,2. Serovars S. Saintpaul and S. Typhimurium DT135 have been responsible for outbreaks of salmonellosis in humans and animals in New Zealand (Environmental Science and Research 2010). Despite Salmonella being isolated from skinks living in close proximity to tuatara, no intestinal shedding of Salmonella has been detected in tuatara (Gartrell et al. 2007), raising the question of whether the lack of detectable Salmonella in tuatara is due to a lack of exposure or whether there are immune factors involved. We investigated the prevalence of Salmonella in tuatara on Stephens Island and compared levels with those in syntopic species.

We conducted sampling in Keepers Bush on Stephens Island (New Zealand, 40°40′S, 174°0′E) during eight trips from October 2009 to October 2011 (Table 1). We captured 275 skinks (O. lineoocellatum, O. infrapunctatum, O. nigriplantare polychrome, and O. zelandicum) using pitfall traps and artificial cover objects (corrugated bitumen roof tiles) on five of the trips. Following capture, each skink was marked with a silver marker to avoid repeat sampling in a trip.

We captured 620 tuatara during five trips; of these, 120 were recaptured on more than one trip. Tuatara were identified to prevent resampling by passive integrated transponders, nuchal crest beads, or marking using a nontoxic black marker. The tuatara were then observed over the duration of the field trip to locate their primary burrow at the time of the study. Samples were taken from tuatara and from skinks and fairy prions inhabiting burrows occupied by tuatara. Soil samples were also collected from tuatara-inhabited burrows and burrows ≤1 m from tuatara burrows. Fairy prions are generally only present at their colonies during their breeding season (October–January); hence, none were caught during the March or June field trips.

Soil samples were collected repeatedly from the same 50 burrows, plus a random selection of additional burrows as time permitted (Table 1). In June 2010, June 2011, and October 2011, only soil samples were collected. At each burrow, approximately 2 g of soil was collected (using sterile gloves) from the end of the burrow. When the end of the burrow could not be reached, soil was collected at one arm's length into the burrow (∼1 m from the entrance).

A single cloacal sample was collected from each animal caught, using a sterile, dry Minitip swab (Copan Diagnostics Inc., Corona, California, USA) and stored in Amies agar gel with charcoal (Copan Diagnostics Inc.). The swabs were immediately refrigerated at 4 C and stored for up to 20 days until transportation was available to the laboratory. A 70% ethanol solution was used to sterilize our hands and any equipment coming into contact with the animals.

Salmonella isolation was achieved by inoculating selenite F enrichment broths (Fort Richard Laboratories, Auckland, New Zealand) with either swabs or 0.5 g of soil. Enrichment broths were then subcultured onto xylose lysine deoxycholate and evaluated using biochemical tests (Middleton et al. 2010). Samples positive for Salmonella were sent to Environmental Science and Research Services (Porirua, New Zealand) for confirmation and serotyping by antigenic determination.

Comparisons were made among the frequencies of Salmonella carriage by species using the Pearson's chi-square test. Where the frequencies involved were small, analyses were conducted using the Fisher's exact test. For the purpose of analyses, the four species of skink were pooled into the category “skink.” Statistical analyses were performed using SPSS (version 18) for Windows (SPSS, Inc., Chicago, Illinois, USA).

The estimated true prevalence of Salmonella was calculated using test prevalence P^T (Thrusfield 2005) and assuming that the Salmonella culture techniques were 50% sensitive and 98% specific (Bager and Petersen 1991). We calculated 95% confidence intervals of the true prevalence (Thrusfield 2005). In those cases in which no Salmonella was detected, the detection of disease component of WinEpiScope 2.0 (Facultad de Veterinaria, Zaragoza, Spain) was used to calculate the maximum possible prevalence at the 95% confidence level. Estimates of population sizes are required for this calculation, and populations were estimated at 50,000 tuatara (Newman 1987), one million fairy prions (Harper 1985), 500,000 skinks (Stephens 2007), and one million burrows.

Salmonella was not isolated from any of the 620 tuatara or 190 fairy prion samples investigated (Table 1). Of the Salmonella-positive skinks, 80% were captured within 1 m of a burrow also found positive for Salmonella. In contrast, only 32.3% of Salmonella-negative skinks were <1 m from a Salmonella-positive burrow (32.3%; χ²₍₁₎ = 19.1, P<0.0001). Finally, the prevalence of Salmonella in soil samples (8.4%) was not significantly higher than in the skinks (χ²₍₁₎ = 0.642, P<0.423; Table 1).

The prevalence of Salmonella isolated from skinks varied from 0–19% among sampling occasions. During the three study trips March 2010, January 2011, and March 2011, Salmonella prevalence in skinks was not significantly different (χ²₍₂₎ = 2.6, P = 0.28; Table 1). There was significantly more Salmonella detected in soil in June 2010 than in any other sampling period (χ²₍₄₎ = 21.58, P<0.001). If the June 2010 outlier is excluded, there is no difference in the prevalence of Salmonella between soil and skinks (χ²₍₁₎ = 0.225, P = 0.635). Fifty-two Salmonella isolates were cultured, serotyped, and identified as S. Saintpaul (n = 44) and S. Mississippi (n = 8). Salmonella was detected in skinks and soil. Salmonella Mississippi was isolated from skink samples only. We did not detect any seasonal or species differences in the distribution of serovars. Salmonella Saintpaul, isolated previously from geckos (Hoplodactylus maculatus) and skinks (O. infrapunctatum, O. nigriplantare polychrome, and O. zelandicum) on Stephens Island (Middleton et al. 2010), is a common serovar isolated from a range of taxa globally (Taylor et al. 2010). Salmonella Mississippi has also been isolated from native reptiles in New Zealand but not previously from Stephens Island (Middleton et al. 2010).

Salmonella was found in Oligosoma skinks at a test prevalence of approximately 6.5% (estimated true prevalence 10%), similar to other studies of reptiles in New Zealand and overseas (4.0% skinks and geckos, Stephens Island [Middleton et al. 2010]; 11% lizards, Australia [Parsons et al. 2010]). Other studies have reported higher prevalences of Salmonella within members of the order Squamata (98% of land iguanas [Conolophus subcristatus], Galapagos Islands [Franco et al. 2011]).

We propose that, in addition to the skink host, environmental contamination likely plays a role in the persistence of Salmonella on Stephens Island. Salmonella is exposed to fluctuating temperatures and pH and limited nutrient availability and osmotic stress outside their hosts, but in contrast to many enteric pathogens, it can survive and multiply for up to a year outside of the host environment (Winfield and Groisman 2003). Salmonella was not isolated during October 2009 and 2011, suggesting another reservoir of Salmonella on the island, or the seasonal reintroduction by an unmeasured carrier. Alternatively, this period of time may correspond to unfavorable environmental conditions in the soil that causes Salmonella to cease reproducing or enter a nonculturable state (Winfield and Groisman 2003).

Given the often high prevalence of Salmonella isolated from soil collected from the interior of tuatara/fairy prion burrows and from skinks caught in proximity to these burrows, it was surprising that Salmonella was not isolated from tuatara or fairy prions. Further research is needed to understand immunity to and factors affecting susceptibility and carriage of Salmonella and to understand the disease ecology of this organism and its cross-species spread to human, domestic animals, and wildlife hosts.

We acknowledge the New Zealand Department of Conservation and Ngāti Koata no Rangitoto ki te Tonga for support and permits to conduct field work on Stephens Island. Funding for this research was provided by the Allan Wilson Centre for Molecular Ecology and Evolution and Victoria University of Wellington. All handling and sampling of animals was approved by the Victoria University Animal Ethics committee (permit 2009R12) and the New Zealand Department of Conservation (permit NM26339-FAU).