SURVEILLANCE FOR UPPER RESPIRATORY TRACT DISEASE AND MYCOPLASMA IN FREE-RANGING GOPHER TORTOISES (GOPHERUS POLYPHEMUS) IN GEORGIA, USA Open Access

Gopher tortoise (Gopherus polyphemus) populations have declined throughout much of their historic range in the southeastern US due to habitat loss and fragmentation, human activity, and disease (Auffenberg and Franz 1982; Diemer-Berish et al. 2010). In 1987, the gopher tortoise was federally listed as threatened in the western portion of its range (west of the Mobile and Tombigbee Rivers in Alabama, Louisiana, and Mississippi; USFWS 1987). The tortoise is state-listed in Georgia and is currently a candidate for federal listing in the eastern portion of its range (USFWS 2011).

Upper respiratory tract disease (URTD) in gopher tortoises is caused by several contagious pathogens including Mycoplasma agassizii and Mycoplasma testudineum (Jacobson et al. 2014), herpesvirus (Jacobson et al. 2012), and ranavirus (Westhouse et al. 1996; Johnson et al. 2010), with the former two being the most-commonly associated with URTD clinical signs. Upper respiratory tract disease has been implicated in the reduction of some tortoise populations (e.g., McLaughlin 1997; Gates et al. 2002; Diemer-Berish et al. 2010). However, there is debate about the long-term impacts of URTD in free-ranging tortoise populations (e.g., Seigel et al. 2003; McCoy et al. 2007; Sandmeier 2009). Upper respiratory tract disease is characterized by mild to severe nasal and ocular discharge, conjunctivitis, and swelling of the eyes and nares. The factors that predispose a tortoise to develop clinical disease are unknown. Some gopher tortoise populations in Florida have a high prevalence of clinical URTD, while other populations with a high prevalence of Mycoplasma antibodies have low prevalence of clinical URTD (Diemer-Berish et al. 2010). Because Mycoplasma can cause persistent infections, it is possible that stressors such as habitat disturbance, translocation, or coinfections with other pathogens could cause infected but clinically normal tortoises to develop clinical signs. Also, because multiple species of Mycoplasma infect gopher tortoises, these species, or strains of a species, may vary in pathogenicity (Wendland 2007).

Few investigators have studied the prevalence of Mycoplasma in gopher tortoise populations outside of Florida, USA. In Georgia, USA, studies have been restricted to three isolated populations (Kahn 2006; Tuberville et al. 2008; Hernandez et al. 2010). In a study conducted in 1994, 93 tortoises from Bullock County, Georgia were translocated to St. Catherines Island in Liberty County, Georgia (Tuberville et al. 2008). At the time of translocation, 80% of the relocated tortoises had detectable antibodies to M. agassizii (Tuberville et al. 2008). Currently, over 100 adult tortoises in this population have been tested and are antibody positive, yet clinical disease is rare (T. Norton unpubl. data). Screening of tortoises at Fort Benning, in Muscogee County, Georgia, for Mycoplasma antibodies revealed seasonal variation in exposure rates with only 27% of tortoises (n = 50) tested in the summer being positive and 44% of tortoises tested in the spring being positive (Kahn 2006). Finally, in a 2001 translocation of tortoises from McIntosh County, Georgia to the Savannah River Site in Aiken County, South Carolina, USA, one of 14 tortoises tested was positive for antibodies to M. agassizii and an additional four were suspect (Hernandez et al. 2010). Importantly, none of these studies evaluated the prevalence of antibodies to M. testudineum. Molecular or culture evidence of Mycoplasma infections in Georgia is very limited, with M. agassizii being identified from 10 of 21 tortoises from St. Catherines Island (none of which had clinical signs; Tuberville et al. 2008) and M. agassizii being cultured from a single, moribund tortoise from the Joseph W. Jones Ecological Research Center (JERC) in Baker County (Wendland 2007).

Our purpose was to acquire knowledge of the distribution of Mycoplasma and URTD in gopher tortoise populations. The specific goal was to determine the prevalence of URTD through health assessments and through antibody and PCR testing for M. agassizii and M. testudineum on tortoises from selected populations in Georgia.

Samples were collected from tortoises from 11 sites in Georgia between 1995 and 2013 (Fig. 1 and Table 1). Sites had varied land uses including public use (n = 2), private lands (n = 7), and military installations (n = 2). All properties were reported by land managers to have >30 tortoises, and tortoise population densities were available for five sites: JERC, Reed Bingham State Park (RBSP), Effingham County (EFCO), Orianne Indigo Snake Preserve (OISP), and Fort Gordon Air Force Base (FGAB). The JERC had approximately 7,600 ha of suitable habitat with a tortoise population density of 0.71±0.08 tortoises/ha (L.L.S. unpubl. data). Orianne Indigo Snake Preserve had 95.4 ha of habitat and an estimated density of 1.23±0.21 tortoises/ha (Ballou 2013). Reed Bingham State Park has a high tortoise density of 3.08±0.67 tortoises/ha in 66 ha of habitat (Ballou 2013). Fort Gordon Airforce Base is at the northern extent of the range of the gopher tortoise and has approximately 13,000 ha of suitable habitat but a low tortoise density (0.02±0.004 tortoises/ha; J. Stober unpubl. data). Specific locality data is withheld due to the status of the species and requests of private landowners.

Tortoises were collected using a combination of methods that included opportunistic hand-capture, trapping, and excavation of burrows in an attempt to capture at least 30 adults (see Gonynor 2013) for details about trapping). Juvenile tortoises were trapped at one site (JERC; Ballou 2013). Upon capture, animals were transported to a shaded area or to a laboratory for sample collection. Each tortoise was physically examined and measured (straight-line carapace length, plastron length, gular length, anal notch, and anal fork) and weighed (McRae et al. 1981). At sites with long-term monitoring objectives, tortoises were permanently marked by notching the marginal scutes (Cagle 1939) with a rotary tool (Dremel™, Racine, California, USA). Tortoises were examined for clinical signs suggestive of URTD (e.g., nasal exudates, conjunctivitis, swollen eyes, lethargy, labored/wheezy breathing) and lesions suggestive of chronic URTD (e.g., nasal scarring and asymmetric nares); we recorded whether each of these clinical signs were either present or absent. Gender was determined based on external morphology such as gular length, plastral concavity, and the ratio of anal notch to anal fork width (McRae et al. 1981; Eubanks et al. 2003). Tortoises whose gender could not be confirmed because of ambiguous characters were grouped as “unknown.” Due to difficulty in differentiating between males and females based on shell morphology alone, tortoises less than 230 mm CL were categorized collectively as “juveniles” (McRae et al. 1981).

Between 0.5–4.0 mL of blood (<1% of body weight) were collected from either the caudal vein, brachial vein, or the subcarapacial venous sinus (Wendland 2007). Blood was added to heparinized tubes and centrifuged. Plasma was removed and immediately frozen at −80 C.

In 2010 and 2011, nasal exudates were collected from tortoises exhibiting clinical signs of URTD (i.e., nasal discharge) using sterile rayon swabs (Puritan Medical Products Company LLC, Guilford, Maine, USA; Microbrush International, Grafton, Wisconsin, USA). After collection, swabs were placed in tubes and frozen at −80 C until testing. Tortoises with eroded nares were considered to have chronic URTD with recurring clinical signs.

Tortoises were hydrated in warm water for 15–20 min (Wendland et al. 2009) prior to release at the site of capture. All tortoises were sampled and released within 48 hr of capture. Equipment was disinfected with a 10% bleach solution between tortoises. All methods were reviewed and approved by the University of Georgia's Animal Care and Use Committee (A2010 11-563).

Plasma samples were submitted to the Mycoplasma research laboratory at the University of Florida College of Veterinary Medicine (Gainesville, Florida, USA) for enzyme-linked immunosorbent assay (ELISA) testing. All samples were tested for antibodies to M. agassizii and a subset was tested for antibodies to M. testudineum. Results for both M. agasizzi and M. testudineum were grouped into one of three classes based on antibody titers; positive (titer≥64), negative (titer<32), and suspect (titer = 32–63) (Wendland et al. 2007).

To detect shedding of Mycoplasma spp. in nasal exudates, PCR testing for the 16S rRNA gene of Mycoplasma spp. was conducted. Genomic DNA was extracted from the swabs according to the manufacturer's protocols (Qiagen DNA purification Kit, Germantown, Maryland, USA) and each sample was tested in duplicate using primers GMF-1 (5′-ACACCATGGGAGCTGGTAAT-3′) and GMR-1 (5′-CCTCATCGACTTTCAGACCCAAGGCAT-3′) as described in Lauerman (1998) and Diemer-Berish et al. (2010).

Fisher's exact test was used to analyze differences in antibody prevalence and PCR results among sites and among age classes and between sexes. Linear regression was used to assess the relationship between carapace length and pathogen prevalence.

Samples from at least 30 tortoises were obtained at six sites: JERC, OISP, Cedars Farm Plantation, Decatur County (CFP), EFCO, Telfair County site B (TFB), and Moody Air Force Base, Lowndes and Lanier counties (MAFB) (see Table 1). We captured <30 tortoises at each of the other five sites (already described) because population densities were low. Serum samples were collected from 540 tortoises at the 11 sites (Table 2). Additional tortoises were captured and examined for signs of URTD at JERC; however, serum samples were not collected from these individuals.

Prevalence of antibodies to M. agassizii and M. testudineum varied among sites (Table 3). The prevalence of antibodies to M. agassizii was either very low (≤3%) or very high (92–100%), whereas prevalence of M. testudineum varied from 20–61%. A low prevalence of antibodies (≤3%) to M. agassizii was detected at seven sites (MAFB, OISP, EFCO, TFB, St. Mary's Airport, Camden County [SMAP], Lowndes County [LCO], and Ft. Gordon, Richmond County [FGAB]) and a high prevalence (92–100%) was detected at the remaining four sites (JERC, Forest Lodge Farms, Mitchell County [FLF], Cedars Farm Plantation, Decatur County [CFP], and RBSP). Tortoises (n = 239) at 10 sites were tested for both M. agassizii and M. testudineum. Tortoises with antibodies to both M. testudineum and M. agassizii were detected at five sites: JERC (n = 25/70), FLF (n = 2/5), CFP (n = 5/25), OSIP (n = 1/30), and RBSP (n = 19/33). Tortoises with acute or chronic clinical signs of URTD were only documented at sites with both M. agassizii and M. testudineum in the population (Table 3). We found no relationship between detection of M. testudineum and clinical signs (χ² = 3.190; P = 0.203). Only two sites where M. agassizii antibodies were not detected had evidence of chronic URTD (tortoises with asymmetric nares).

Nasal swabs were collected from 136 tortoises at three sites (JERC, n = 129; FLF, n = 1; and CFP, n = 6) and of these, 50 (37%) swabs were PCR positive for Mycoplasma spp. (Table 3). Based on ELISA results, all three sites also had a high prevalence of M. agassizii antibodies (92–100%). Among the PCR-positive tortoises (n = 50), 39 (29%) had nasal discharge and 38 (28%) had chronic lesions on the nares suggestive of URTD. Of the six CFP swabs tested, two (33%) were PCR positive, both of which were antibody positive to both M. agassizii and M. testudineum. Both PCR-positive tortoises from CFP were antibody-positive to M. agassizii and M. testudineum. The FLF swab was PCR negative, but the tortoise was antibody positive for M. agassizii. Both PCR and serology data are available for 39 tortoises across the three sites (JERC, CFP, OISP); 23 (59%) tortoises had antibody to both M. agassizii and M. testudineum and 13 (33%) were PCR positive. Twenty-nine samples were bidirectionally sequenced at the Georgia Genomics Facility (Athens, Georgia, USA). All samples had >97% identity to M. agassizii (GenBank AY80802).

Prevalence of antibodies to M. agassizii among the 14 juvenile tortoises was significantly lower than among adults (Fisher's exact test, P<0.001), but there was no difference in prevalence between adults and juveniles for M. testudineum (Table 4). A single juvenile from JERC was PCR positive and sequenced as M. agassizii. This tortoise (CL 20.4 cm) was antibody positive to M. testudineum but not to M. agassizii (Table 4). A PCR-negative juvenile (CL 21.2 cm) was suspect for M. agassizii and positive for M. testudineum. One tortoise of unknown gender with a CL of 23 cm was antibody positive to M. agassizii, suspect for M. testudineum, and PCR positive. When data from all sites were combined, there was no difference in prevalence for M. agassizii between males and females (Fisher's exact test, P = 0.148), but the prevalence of M. testudineum was significantly higher for females (43/88) compared with males (33/95) (Fisher's exact test, P = 0.006) (Tables 5 and 6).

Recapture data were available for two sites, MAFB and JERC. Thirty-seven tortoises from MAFB were recaptured and tested between 2000 and 2004, and 21 tortoises at JERC were tested in 1997 (Kahn 2006) and again between 2009 and 2011. At MAFB, serostatus changes were documented for three tortoises. One tortoise was suspect for M. agassizii antibodies in August 2002 and August 2003, positive (titer 64) in September 2003, and was suspect again in September 2004. A second tortoise was antibody negative to M. agassizii in August 2000, suspect in September 2001, and was negative in July 2002 and June 2003. The third tortoise was negative for antibodies to M. agassizii in August 2000, suspect in September 2001, and negative in July 2003. The remaining 34 resampled tortoises at MAFB were negative at each sampling point. At JERC, all 21 resampled tortoises were antibody positive in 1997 and remained positive for M. agassizii in 2009–11. No differences were noted in seroprevalence between sampling years at either site.

Among the gopher tortoise populations tested in Georgia, the prevalence of M. agassizii antibodies was either very high or very low in a population (nearly all or none), with seven of 11 sites (64%) having antibody-positive tortoises. In contrast, antibodies to M. testudineum were detected at all but two sites (82%) and prevalence within the populations varied. Approximately half of the sites had tortoises with antibodies to M. agassizii whereas a higher percentage (73%) of sites had tortoises with antibodies to M. testudineum. Five sites (45%) had tortoises with antibodies to both pathogens and four of these sites had tortoises with clinical signs of URTD. Mycoplasma testudineum can cause clinical disease in the absence of M. agassizii; however, there is some evidence that M. testudineum might be less pathogenic in desert tortoises (Jacobson and Berry 2012). In a Florida study, M. testudineum was documented at three of 11 field sites (27%), all located in the northeastern part of the state (Wendland 2007). To our knowledge, ours is the first report of M. testudineum in Georgia. In contrast to the Florida study, we detected M. testudineum on a greater proportion of sites than M. agassizii. We also observed significantly more female tortoises with antibodies to M. testudineum. Additionally, we only observed clinical signs of URTD in tortoises with antibodies to M. agassizii alone or at sites with tortoises with antibodies to both M. testudineum and M. agassizii. The same study also observed clinical signs (nasal discharge) in tortoises infected with M. testudineum (PCR-positive nasal discharge), but it is unclear how many of those tortoises were coinfected with M. agassizii. Other studies also documented clinical signs at sites with a high prevalence of antibodies to M. agassizii (Wendland 2007).

Contrary to other reports of fluctuating antibody prevalence within tortoise populations over time (Kahn 2006; Diemer-Berish et al. 2010), prevalence at two of our sites, MAFB and JERC, remained stable over the 4-yr and between the 15-yr periods, respectively. Mycoplasma testudineum testing has not been done at MAFB; we do not know if the pathogen is present. Wendland et al. (2007) warned that one positive tortoise detected in a population with low antibody prevalence should be interpreted with caution, as there are occasional false-positive ELISA results. We recommend that land managers perform repeated health assessments to look for clinical signs as well as pathogen surveillance in concert with population monitoring. Additionally, PCR should be utilized when possible to determine current infection because ELISA results only indicate exposure. Interestingly, the positive tortoise from MAFB only developed a low titer and was “suspect” at two other testing times; therefore, the true infection status of this tortoise is unknown. Antibody prevalences to M. agassizii at JERC were similar between two sampling periods (92% in 2009–12 vs. 96% in 1997) even though ELISA sensitivity and specificity had been improved between the sampling periods (Wendland et al. 2010a).

Despite the high prevalence of antibodies to M. agassizii at JERC, the prevalence of current clinical signs and chronic URTD lesions was <30%, suggesting that not all tortoises shed the bacteria and develop clinical signs or that past infections were mild and did not result in visible chronic lesions. Future studies should be conducted to determine if antibody-positive tortoises develop recrudescence of clinical signs.

Several juvenile tortoises from different sites had antibodies or were PCR positive for Mycoplasma spp. Previous investigators have shown a correlation between body size (CL) and antibody prevalence, with tortoises <23 cm CL being antibody-negative (Beyer 1993; Wendland 2007). In our study, the smallest tortoise with detectable antibodies to M. agassizii was from RBSP (CL 13.1 cm), and the smallest tortoise with antibodies to M. testudineum was from JERC (CL 20.4 cm). In previous studies, juvenile tortoises antibody positive to Mycoplasma were from populations undergoing epizootic events (Wendland 2010b); otherwise, most juveniles were negative for both M. agassizii and M. testudineum antibodies (Wendland 2007). The PCR-positive (sequenced as M. agassizii) juvenile tortoise from JERC was interesting in that it was antibody negative for M. agassizii and positive for M. testudineum. It is possible that this tortoise was recently exposed to M. agassizii and had not mounted an immune response. At CFP the only juvenile (CL 20.0 cm) with nares abnormalities was antibody positive for M. agassizii but negative for M. testudineum. The PCR was utilized on samples collected from tortoises with clinical signs. However, nasal flushes (McGuire et al. 2014) would have been helpful on sites where there were no tortoises with clinical signs. Sequencing in concert with serology would be optimal but was not practical on most sites.

We observed lower than expected numbers of tortoises at several sites and, as a result, were unable to sample our target of 30 individuals per population. Despite low sample sizes we detected evidence of both Mycoplasma and chronic URTD at FLF. Similarly at SMAP and FGAB, we detected tortoises with antibodies to M. testudineum. Two tortoises at FGAB had lesions consistent with previous URTD, yet no antibodies to M. agassizzi were detected in the population. These nasal lesions could have been due to M. testudineum or another pathogen. Data from this site highlight the fact that the presence of URTD signs does not indicate that the tortoise or population is infected with M. agassizii. Three of our study populations were subsequently translocated following testing (EFCO, LCO, and TFB). Tortoises at two of these sites were negative for antibodies to M. agassizii but positive for M. testudineum antibodies. Ideally, no Mycoplasma-positive tortoises should be relocated; however, at one site a few native or waif tortoises (tortoises of unknown origin turned in to authorities such as state agencies) of unknown Mycoplasma status were already present. This is concerning because waif tortoises could be at increased risk of infection or be a source of infection to naïve tortoises (McLaughlin 2000).

The gopher tortoise is one of the most commonly translocated animals in North America (Tuberville et al. 2011); therefore, it is important to understand the distribution of pathogens in populations. Multiple strains of M. agassizii exist (Wendland et al. 2010a) and geographic variation in virulence may occur among strains (Seigel et al. 2003). Our study highlights the importance of continued monitoring of tortoise health because our results differed from previous studies in several important ways. In Florida, antibodies to M. agassizii were widespread and prevalences varied considerably (Diemer-Berish et al. 2000; Wendland 2007), whereas in Georgia prevalences of antibodies to M. agassizii were either very high or very low/absent. Mycoplasma testudineum is also a confirmed etiologic agent for URTD (Brown et al. 2004). However, it has been hypothesized that M. testudineum could be less pathogenic than M. agassizii in desert tortoises because tortoises infected with M. testudineum had less-severe lesions than those infected with M. agassizii (Jacobson and Berry 2012). We did not observe clinical signs at sites with tortoises that were antibody positive only to M. testudineum. The relationship between the two pathogens is unclear. However, we observed that sites with both pathogens had tortoises that exhibited signs of clinical disease. Further surveillance for Mycoplasma spp., in addition to other pathogens, in more tortoise populations throughout the range is warranted. Long-term studies need to be initiated, particularly in populations of management concern, to understand the consequences of disease and to fill in knowledge gaps concerning behavior and reproduction.

We thank Elizabeth Miller, Mallory Chronic, Will McGuire, and numerous personnel at the JERC for assistance in the field. We also thank landowners who provided access to their properties and sampling of tortoises. Thanks to Barbara Shock for help with sequencing. Funding for this project was provided by the Morris Animal Foundation (D13ZO-015), Gopher Tortoise Council, Sigma Xi, Sophie Danforth Conservation Biology Fund, Orianne Society, Southeastern Cooperative Wildlife Disease Study through sponsorship with member states, the Joseph W. Jones Ecological Research Center, and the D. B. Warnell School of Forestry and Natural Resources at The University of Georgia.