JOURNAL OF CLINICAL MICROBIOLOGY,0095-1137/99/$04.0010

July 1999, p. 2262–2269 Vol. 37, No. 7

Copyright © 1999, American Society for Microbiology. All Rights Reserved.

Upper Respiratory Tract Disease in the Gopher TortoiseIs Caused by Mycoplasma agassizii†

M. B. BROWN,1* G. S. MCLAUGHLIN,2,3 P. A. KLEIN,4 B. C. CRENSHAW,1

I. M. SCHUMACHER,4 D. R. BROWN,1 AND E. R. JACOBSON2

Department of Pathobiology and Division of Comparative Medicine1 and Department of Small Animal ClinicalSciences,2 College of Veterinary Medicine, Department of Wildlife Ecology and Conservation, Institute of

Food and Agricultural Sciences,3 and Department of Pathology, Immunology, and LaboratoryMedicine, College of Medicine,4 University of Florida, Gainesville, Florida

Received 19 November 1998/Returned for modification 18 February 1999/Accepted 31 March 1999

Upper respiratory tract disease (URTD) has been observed in a number of tortoise species, including thedesert tortoise (Gopherus agassizii) and the gopher tortoise (Gopherus polyphemus). Clinical signs of URTD ingopher tortoises are similar to those in desert tortoises and include serous, mucoid, or purulent discharge fromthe nares, excessive tearing to purulent ocular discharge, conjunctivitis, and edema of the eyelids and ocularglands. The objectives of the present study were to determine if Mycoplasma agassizii was an etiologic agent ofURTD in the gopher tortoise and to determine the clinical course of the experimental infection in a dose-re-sponse infection study. Tortoises were inoculated intranasally with 0.5 ml (0.25 ml/nostril) of either sterile SP4broth (control group; n 5 10) or 108 color-changing units (CCU) (total dose) of M. agassizii 723 (experimentalinfection group; n 5 9). M. agassizii caused clinical signs compatible with those observed in tortoises with natu-ral infections. Clinical signs of URTD were evident in seven of nine experimentally infected tortoises by 4 weekspostinfection (p.i.) and in eight of nine experimentally infected tortoises by 8 weeks p.i. In the dose-responseexperiments, tortoises were inoculated intranasally with a low (101 CCU; n 5 6), medium (103 CCU; n 5 6),or high (105 CCU; n 5 5) dose of M. agassizii 723 or with sterile SP4 broth (n 5 10). At all time points p.i. inboth experiments, M. agassizii could be isolated from the nares of at least 50% of the tortoises. All of the exper-imentally infected tortoises seroconverted, and levels of antibody were statistically higher in infected animalsthan in control animals for all time points of >4 weeks p.i. (P < 0.0001). Control tortoises in both experimentsdid not show clinical signs, did not seroconvert, and did not have detectable M. agassizii by either culture orPCR at any point in the study. Histological lesions were compatible with those observed in tortoises with natu-ral infections. The numbers of M. agassizii 723 did not influence the clinical expression of URTD or the anti-body response, suggesting that the strain chosen for these studies was highly virulent. On the basis of the re-sults of the transmission studies, we conclude that M. agassizii is an etiologic agent of URTD in the gopher tortoise.

Gopher tortoises (Gopherus polyphemus) are found in thesoutheastern United States, with the major population concen-trations found in Florida and southern Alabama and Georgiaand only remnant populations found in South Carolina, Mis-sissippi, and Louisiana (9). The gopher tortoise is legally pro-tected in all states within the range (Alabama, Mississippi,Louisiana, Georgia, South Carolina, and Florida) and is listedin Appendix II of the Convention on International Trade inEndangered Species of Wild Fauna and Flora, which requirespermits for the exportation of the species from the UnitedStates to any signatory nation or for reexportation (20). Go-pher tortoises are an important element in the ecosystems inwhich they are found and are considered by many ecologists tobe a keystone species. Gophers are the most fossorial of thefour North American species of tortoises, digging burrows thatmay extend 5 m down from the surface and 15 m in length (9,12). The burrows provide a microclimatically stable environ-ment not only for the tortoises but also for numerous com-mensal species. Approximately 60 vertebrate species, fromsnakes to birds, and over 300 invertebrates including spiders,

crickets, and beetles have been found in tortoise burrows orhave been observed using them as permanent homes or refugesfrom heat, cold, fire, and predators (14, 22, 36). Several speciesthat either exclusively or frequently use tortoise burrows havelegal protection in Florida and other parts of their ranges (6).Thus, tortoises are of critical importance to the ecosystem.

Upper respiratory tract disease (URTD) has been observedin a number of tortoise species (15, 16, 19), including the deserttortoise (Gopherus agassizii) and the gopher tortoise. Clinicalsigns of URTD have been observed in a number of importedcaptive tortoise species (19) and in tortoises submitted to theVeterinary Medical Teaching Hospital (VMTH), University ofFlorida (UF), including the red-footed tortoise (Geochelonecarbonaria), leopard tortoise (Geochelone pardalis), Indian startortoise (Geochelone elegans), and radiated tortoise (Geoch-elone radiata). Numerous wild and captive gopher tortoiseshave been submitted to VMTH with clinical signs consistentwith URTD.

Clinical signs of URTD in gopher and desert tortoises aresimilar and include serous, mucoid, or purulent discharge fromthe nares, excessive tearing to purulent ocular discharge, con-junctivitis, and edema of the eyelids and ocular glands (16, 31).Individual infected tortoises vary in the suite of signs that theyhave, and the severity can vary from day to day. Nares maybecome occluded with caseous exudate, preventing externallyvisible nasal discharge. Tortoises may become lethargic and

* Corresponding author. Mailing address: Department of Pathobi-ology, Box 110880, College of Veterinary Medicine, University ofFlorida, Gainesville, FL 32611-0880. Phone: (352) 392-4700, ext. 3970.Fax: (352) 846-2781. E-mail: mbbrown@nersp.nerdc.ufl.edu.

† Journal series article R-06916 of the Florida Agriculture Experi-ment Station.

2262

on Septem

ber 6, 2018 by guesthttp://jcm

.asm.org/

Dow

nloaded from

anorectic, leading to dehydration, emaciation, and eventualdeath from cachexia.

In a previous study, we fulfilled Koch’s postulates and dem-onstrated that an etiologic agent for URTD in the deserttortoise is Mycoplasma agassizii proposed sp. novum (2, 4).Histologically, the lesions from experimentally infected deserttortoises were consistent with those seen in naturally infectedtortoises (4, 16). In the desert tortoise, we have shown that thepresence of clinical signs of URTD was positively related to thepresence of specific antibody to M. agassizii (31). Additionalwork led to the development of a PCR test for detection of thebacteria in nasal lavage and swab samples (2).

We isolated M. agassizii from the nasal passages of clinicallyill gopher tortoises submitted to VMTH, UF. The similarity ofthe clinical signs and histological lesions between experimen-tally infected and naturally infected tortoises and the isolationof M. agassizii from naturally infected gopher tortoises sug-gested that URTD in this species might also be of mycoplasmalorigin. The objectives of this study were to determine if M.agassizii was an etiologic agent of URTD in the gopher tortoiseand to determine the clinical course of the experimental infec-tion in a dose-response infection study.

MATERIALS AND METHODS

Tortoises. Gopher tortoises were transferred under Florida Game and FreshWater Fish Commission permit nos. WX93227, issued to E. R. Jacobson, andWX94037, issued to M. B. Brown, from a development site in central Florida inApril, July, and August 1994 and April 1995 and were processed on the dayfollowing arrival. Tortoises were examined for clinical signs of URTD: nasal andocular discharge, palpebral edema, and conjunctivitis. Tortoises were weighed tothe nearest 10 g, and ketamine hydrochloride (Ketaset; Fort Dodge Laborato-ries, Inc., Fort Dodge, Iowa) was administered at 20 mg/kg of body weight. Ablood sample (2 to 3 ml) was drawn from the jugular or brachial vein and wasplaced in a Vacutainer tube (Becton Dickinson and Company, Rutherford, N.J.)containing lithium heparin. Blood was centrifuged, and an aliquot of plasma wasremoved for specific antibody screening by an enzyme-linked immunosorbentassay (ELISA). After cleansing of the area around the nares with alcohol-dampened gauze, nasal lavage samples were collected by flushing with approxi-mately 0.5 ml of sterile SP4 broth with a 1-ml syringe without a needle. Calciumalginate-tipped swabs were gently inserted into the nares, and a sample wasobtained and streaked onto SP4 agar plates (33).

Husbandry. Tortoises were housed individually in outdoor pens at the UFAnimal Resource Farm. There were four groups of 10 pens in a larger enclosuresurrounded by a chain-link fence. Individual pens were approximately 21 m2,were constructed of a wooden frame with sheet metal extending vertically ap-proximately 0.7 m above and below the ground, and were partially covered byshade cloth. The tortoises were provided an artificial burrow, a water dish, anda cement feeding stone. Because tortoises burrow, the risk of cross contamina-tion was too great to allow randomization of treatment groups within pen groups.Each treatment group of 10 pens was separated from the other treatment pensby .200 m. The tortoises were fed a salad of mixed vegetables three times perweek, and fruit was provided on an occasional basis. Water was provided asneeded. Husbandry personnel wore gloves for all procedures requiring handlingof food, feeding stones, or water dishes. Entry into pens and handling of tortoiseswere restricted to research personnel. Any person handling a tortoise wore cleangloves, which were changed as necessary and before handling of a differenttortoise.

Infection groups. Animals were allowed to acclimate for a minimum of 1month prior to initiation of the experimental transmission trials. Tortoises wereblocked on the basis of size and sex prior to random assignment to experimentaland control groups. Tortoises in the control group (n 5 10) were sham inocu-lated intranasally with 0.5 ml (0.25 ml/nostril) of sterile SP4 broth. Tortoises inthe experimental infection group (n 5 9) were inoculated intranasally with 0.5 ml(0.25 ml/nostril; 108 color-changing units [CCU] [total dose]) of M. agassizii 723.Isolate 723 was obtained from a clinically ill tortoise from Sanibel Island, LeeCounty, Fla. M. agassizii 723 was grown in SP4 broth and was only two passagesfrom the primary isolation. Aliquots of strain 723 were frozen at 280°C, and allinfections were done with a common stock of M. agassizii. The purity of theisolate was determined by growth inhibition and 16S rRNA sequence analysis(2).

Dose-response study. After the initial experiments which demonstrated thatM. agassizii caused URTD in the gopher tortoise, an experiment was designed todetermine the effect of dose on clinical expression of disease and immuneresponse. Tortoises were inoculated intranasally with a low (101 CCU; n 5 6),medium (103 CCU; n 5 6), or high (105 CCU; n 5 5) dose of M. agassizii.Tortoises in the control group (n 5 10) were sham inoculated intranasally with

0.5 ml (0.25 ml/nostril) of sterile SP4 broth. Monitoring and housing wereidentical to those in the initial experimental infection study.

Postinfection monitoring of tortoises. The tortoises were observed from 3 to 7days each week, with some days including multiple observations. Because ofindividual behavior patterns, not every tortoise was observed at each daily ob-servation time point. At specific time points (usually 2- to 4-week intervals,depending on the study), tortoises were captured by hand or with wire cage-typetraps (Tomahawk Live Trap Company, Tomahawk, Wis.) that were covered withbrown paper to protect the animals from the weather. The traps were cleaned,sprayed with bleach solution, and allowed to air dry following each use. Thepaper was discarded, and fresh paper was used for the next trapping effort. Eachtortoise was placed in a plastic, lidded container (LEWISystems; Menasha Cor-poration, Watertown, Wis.) for transport and holding. Containers were bleached,scrubbed, and washed in an automatic cage washer before reuse. Tortoises wereexamined for clinical signs of URTD: nasal and ocular discharge, palpebraledema, and conjunctivitis. A photographic record consisting of right, left, and fullface views was made for each tortoise at each time point when the tortoises werecaptured. The signs were graded individually on a scale of from 0 to 3, whichindicated none, minimal, mild, and severe signs, respectively. Visual grading ofsigns was confirmed by independent observation of the photographic record.Serum was obtained for quantitation of specific antibody. Nasal swabs andlavages were obtained for culture and PCR testing.

Culture. Mycoplasmal cultures were performed as described previously (4). A100-ml aliquot of the lavage sample was used for PCR analysis; the remainingsample was serially diluted 10-fold to 1022 and was incubated at 30°C for amaximum of 3 weeks or until it was determined to be positive or contaminated.In some cases, an aliquot of the broth culture of both lavages and swabs wasremoved after 24 to 48 h and was used for PCR to confirm growth of M. agassizii.Twenty microliters of each dilution was placed on SP4 agar, and the plates wereincubated at 30°C in 5% CO2. The swabs were streaked onto the surface of anSP4 plate. The plates were examined regularly for a maximum of 6 weeks todetect the growth of mycoplasma.

PCR. Nasal aspirate lavage specimens were analyzed for the presence of M.agassizii DNA on the basis of PCR amplification of the 16S rRNA gene (2). Nasallavage specimens and selected culture samples obtained at between 24 and 48 hwere centrifuged at 16,000 3 g for 60 min at 4°C, and the supernatant wasaspirated. The pellets were resuspended in 3 to 4 ml of 20 mg of proteinase K(Sigma, St. Louis, Mo.) per ml in 20 ml of lysis buffer (100 mM Tris [pH 7.5], 6.5mM dithiothreitol, 0.05% Tween 20), and the mixture was incubated at 37°C for8 to 16 h. After denaturation of the proteinase K at 97°C for 15 min, 5 ml of eachsample was removed and was added to 45 ml of a reaction solution containing twoprimers for the 16S rRNA gene at 1 mM each, deoxynucleoside triphosphates at200 mM, 2.0 mM MgCl2, and 2.5 U of Taq polymerase (Promega, Madison,Wis.). The primers were complementary to sequences found in the V3 variableregion of the 16S rRNA gene (sense strand nucleotides [nt] 471 to 490 [59-CCTATATTATGACGGTACTG-39]) and a Mycoplasma genus-specific region (an-ti-sense strand nt 1055 to 1031 [59-TGCACCATCTGTCACTCTGTTAACCTC-39]) (2, 34). The samples were subjected to 50 cycles of template denaturation for45 s at 94°C, primer annealing for 1 min at 55°C, and polymerization for 45 s at72°C, followed by 10 min at 72°C. Positive samples yielded 576-bp products thatwere visualized by combining 15 ml of product with 2 ml of bromphenol blue in50% glycerol solution and electrophoresing on ethidium bromide-stained 1.5%agarose gels in Tris-borate-EDTA buffer. Positive control samples with 250 ng ofpurified M. agassizii DNA as the template and negative control samples withwater in place of a template were included with each amplification run. Amolecular weight marker, an HaeIII digest of phage fX174 DNA, was includedon each gel.

Restriction fragment length polymorphism analysis was conducted with atleast one isolate from each tortoise in order to confirm conclusively that theisolates obtained from naturally and experimentally infected tortoises were M.agassizii (2). Twenty-microliter samples of products from the amplification pro-cedure described above were incubated with 10 to 20 U of the endonuclease AgeI(New England Biolabs, Inc., Beverly, Mass.), which cuts the M. agassizii ampli-fication product at nt 613, and 5 ml of reaction buffer at 25°C for 1 h, and theproducts were electrophoresed as described above. The procedure resulted inproducts of 434 and 142 bp from M. agassizii-positive samples, and all mycoplas-mal isolates in the study were confirmed to be M. agassizii.

ELISA. Specific antibody to M. agassizii was determined by ELISA as de-scribed previously (30). Ninety-six-well microliter plates (Maxisorp F96; Nunc,Kamstrup, Denmark) were coated with 50 ml of a whole-cell lysate of M. agassizii723 at 10 mg/ml in phosphate-buffered saline (PBS) with 0.02% azide (PBS-AZ).The plates were incubated overnight at 4°C, washed four times with PBS-AZ plus0.05% Tween 20 (PBST) in an automatic plate washer (EL403; Bio-Tek Instru-ments, Inc., Winooski, Vt.), and blocked overnight at 4°C with 250 ml of PBSTcontaining 5% nonfat dry milk (PBS-TM) per well. Following washing, 50 ml ofplasma diluted appropriately for the specific study with PBS-TM was added toindividual wells in duplicate or triplicate, and the plates were incubated at roomtemperature for 60 min. The plates were washed, 50 ml (per well) of a biotinyl-ated monoclonal antibody (monoclonal antibody HL673) against the light chainof desert tortoise immunoglobulins Y and M at 1 mg/ml in PBS-TM was added,and the plates were incubated for 60 min. Following washing, a conjugate ofalkaline phosphatase and streptavidin (Zymed Laboratories, Inc., San Francisco,

VOL. 37, 1999 M. AGASSIZII CAUSES URTD IN GOPHER TORTOISE 2263

on Septem

ber 6, 2018 by guesthttp://jcm

.asm.org/

Dow

nloaded from

Calif.) at 1:2,000 in PBS-AZ was added at 50 ml/well, and the plates wereincubated for 60 min and washed. The substrate, p-nitrophenyl phosphate diso-dium (pNPP; Sigma), was prepared at 1 mg/ml in 0.01 M sodium bicarbonate(pH 9.6) with 2 mM MgCl2 and was added to the wells at 100 ml/well. The plateswere incubated for 60 min in the dark and were then read at 405 nm on amicroplate reader (EAR 400 AT; SLT Labinstruments, Salzburg, Austria). Themean for two or three wells coated with antigen and incubated with conjugateand substrate only was used as the blank. Seroconversion was defined as anantibody level greater than 2 standard deviations above the values for normalcontrol serum. A positive control, which consisted of plasma from a naturallyinfected gopher tortoise from Sanibel Island, Fla., and a negative control, whichconsisted of plasma from an uninfected tortoise from Orange County, Fla., wereincluded on each plate.

Pathology. After a minimum of 16 weeks postinfection (p.i.), all diseased andselected healthy tortoises were killed with a combination of drugs. Ketamine wasadministered intramuscularly at 60 to 80 mg/kg followed by administration of aconcentrated barbiturate solution (Socumb; The Butler Company, Columbus,Ohio) intracoelomically at 1 ml/kg. Once the tortoises showed complete musclerelaxation and were unresponsive to painful stimulation, they were exsangui-nated with a 23-gauge butterfly catheter inserted into the carotid artery and thendecapitated. Lavage and swab samples were collected as described previously,and then the head was bisected longitudinally with an electric saw. Followingbisection, the cartilage over each nasal cavity was reflected aseptically, and lavageand swab specimens from both left and right nasal cavities were collected.

For those tortoises selected for complete necropsy, the plastron was removedfrom the carapace, and viscera within the coelomic cavity were exposed. A grossnecropsy was conducted. For histopathologic studies, the heads were fixed in10% neutral buffered formalin, decalcified, embedded in paraffin, sectionedlongitudinally at 5 to 6 mm, and stained with hematoxylin and eosin. Sectionswere examined by light microscopy and were classified on a scale of from 0 to 5,with 0 being normal and 5 exhibiting severe inflammation and/or changes.Changes in the epithelium and submucosa were recorded separately.

The following criteria were used for the grading of lesions: (i) normal (score 50), occasional small subepithelial lymphoid aggregates, rare heterophils in thelamina propria, no changes in mucosal or glandular epithelium, and no edema;(ii) mild (score 5 1), multifocal small subepithelial lymphoid aggregates; multi-focally, small numbers of heterophils, lymphocytes, and plasma cells in the lam-ina propria; mild edema in the lamina propria; and minimal changes in mucosalepithelium; (iii) moderate (score 5 2 or 3), multifocal to focally extensivelymphoid aggregates; diffusely, moderate numbers of heterophils, lymphocytes,and plasma cells in the lamina propria, occasionally infiltrating the overlyingmucosal epithelium; moderate edema in the lamina propria; and proliferationand disorganization of the basal epithelium; (iv) severe (score 5 4 or 5), focallyextensive to diffuse bands of lymphocytes and plasma cells subjacent to and ob-scuring the overlying mucosal epithelium; large numbers of heterophils in laminapropria and infiltrating the overlying mucosal epithelium; marked edema of thelamina propria; degeneration, necrosis, and loss of the mucosal epithelium withoccasional erosion; proliferation of the basal cells of the epithelium with meta-plasia of the mucous and olfactory epithelium to a basaloid epithelium; andoccasional squamous metaplasia.

Statistical analysis. Binomial data were analyzed by the chi-square test. Con-tinuous data were analyzed by t test (two-group comparisons) or analysis ofvariance (multiple comparisons), followed by Fisher’s least-squares-differenceanalysis. For clinical sign score data, the nonparametric Mann-Whitney (two-group comparison) or the Kruskal-Wallis (comparison of more than two groups)test was used. A P value of ,0.05 was accepted as significant.

RESULTS

Clinical disease outcome: experimentally infected tortoises.M. agassizii caused clinical signs compatible with those ob-served in animals with natural infection (Table 1; see also Fig.2). Clinical signs of URTD were evident in seven of nineexperimentally infected tortoises by 4 weeks p.i. and in eight ofnine experimentally infected tortoises by 8 weeks p.i. (Table 1).The one tortoise which failed to show clinical signs did sero-convert, indicating that the animal had been colonized suffi-ciently to stimulate a host immune response. At all time pointsp.i., M. agassizii could be isolated from the nares of at least50% of the tortoises (Table 1). After 8 weeks p.i., the ELISAwas the most reliable method of detection, with 100% of theinfected tortoises testing positive for antibodies to M. agassizii(Table 1). Control tortoises which were sham inoculated withsterile broth did not show clinical signs and did not serocon-vert, and M. agassizii was not detected in these tortoises byeither culture or PCR at any point in the study.

The cumulative scores for infected and control tortoises

were different at all time points p.i. (P , 0.001) (Fig. 1A).Nasal discharge was statistically greater in infected tortoisesthan in control tortoises at 4 (P 5 0.004), 8 (P 5 0.01), 12 (P 50.004), and 16 (P 5 0.01) weeks p.i. (Fig. 1B). Ocular dischargewas statistically greater in infected tortoises than in controltortoises at 4 (P 5 0.01), 12 (P 5 0.004), and 16 (P 5 0.005)weeks p.i. (Fig. 1B). Palpebral edema was statistically greaterin infected tortoises than in control tortoises at 4 (P 5 0.04), 8(P 5 0.004), and 12 (P 5 0.04) weeks p.i. (Fig. 1B). Conjunc-tivitis was statistically greater in infected tortoises than in con-trol tortoises at 12 (P 5 0.04) weeks p.i. (Fig. 1B).

The overall cumulative severity of clinical signs increasedand then reached a relative plateau (Fig. 1A). However, theindividual clinical signs comprising the cumulative scoresshowed significantly more variability (Fig. 1B). The severity ofpalpebral edema and conjunctivitis remained relatively con-stant throughout the 16-week observation period. The severityof nasal and, to a lesser extent, ocular discharge did increasewith time following infection (Fig. 1B).

Considerable variability in the expression of clinical signsamong individual animals occurred (data not shown). Someanimals showed a classical plateau response, while othersclearly demonstrated intermittent clinical signs. Some individ-ual animals had very severe clinical signs, while others (n 5 3)had clinical signs which had relatively low scores and oneanimal showed no clinical signs.

Infection with M. agassizii resulted in detectable antibodyresponses by week 8 p.i. (Fig. 2). All of the experimentallyinfected tortoises seroconverted. Levels of antibody were sta-tistically higher in infected animals than in control animals forall time points .4 weeks p.i. (P , 0.0001) (Fig. 2). No antibodyresponse was detected in any control animal at any time point.

Clinical disease outcome: dose-response study. The num-bers of M. agassizii used to infect the tortoises initially did notinfluence the clinical expression of URTD. In most instances,the most severe clinical signs were observed at 8 weeks p.i.,regardless of the infective dose. As was seen in the earlierinfection trial, there was considerable variability in the expres-sion of clinical signs by individual animals (data not shown).The antibody response (Fig. 3), like the expression of clinicalsigns, was not affected by the infection dose used. The antibodyresponse in infected animals was first detectable at 6 weeks p.i.and was statistically different from that in the control animalsat all time points thereafter (P , 0.001).

Histology: experimentally infected tortoises. The nasal cav-ities of control tortoises consisted of a dorsal multilayeredolfactory epithelium (Fig. 4A) and a ventral mucous epithe-lium consisting of mucus cells intercalated with ciliated epithe-lial cells (Fig. 4B). Eight of nine experimentally infected tor-toises had changes in the nasal epithelia and submucosa (Fig.

TABLE 1. Outcome of transmission study to determinepathogenicity of M. agassizii in gopher tortoisesa

Wk

No. of animals positive for the following/total no. of animals tested (%):

Clinical signs ELISA PCR Culture

0 0/9 (0) 0/9 (0) 0/9 (0) 0/9 (0)4 7/9 (77.8) 3/6 (33.3) NDb 8/8c (100)8 8/9 (88.9) 8/9 (88.9) 6/9 (66.7) 6/9 (66.7)

12 8/9 (88.9) 9/9 (100) 3/9 (33.3) 8/8c (100)16 8/9 (88.9) 9/9 (100) ND 8/9 (88.9)

a Tortoises (n 5 9) were infected intranasally with 108 CCU of M. agassizii.b ND, not done.c Denotes one missing sample.

2264 BROWN ET AL. J. CLIN. MICROBIOL.

on Septem

ber 6, 2018 by guesthttp://jcm

.asm.org/

Dow

nloaded from

5). The epithelium was intact in all nine tortoises, and noulcerations were present. One tortoise had mild to moderatechanges, five tortoises had moderate changes, and three tor-toises had changes that were characterized as moderate tosevere.

Histology: dose-response study. Changes were observed inthe nasal cavity epithelia of experimentally infected tortoises.Eight tortoises were selected for full necropsy. The changes

FIG. 1. Clinical signs of URTD in gopher tortoises experimentally infectedwith M. agassizii. Tortoises (n 5 9) were infected intranasally with 108 CCU of M.agassizii. No clinical signs were seen for control tortoises (n 5 10), which receivedsterile broth. (A) Results are expressed as the mean cumulative score, which wascalculated as the nasal discharge score plus the mean for the three ocular scores.Infected tortoises (F) had greater cumulative signs than control tortoises (E) atall time points p.i. (P 5 0.001). (B) Results are expressed as the mean clinicalsign score 1 standard error on a scale of from 0 (none) to 3 (severe). At 4 weeksp.i., nasal (P 5 0.004) and ocular (P 5 0.01) discharges as well as palpebraledema (P 5 0.04) were greater in infected tortoises than in control tortoises. At8 weeks p.i., nasal discharge (P 5 0.01) and palpebral edema (P 5 0.004) weregreater in infected tortoises than in control tortoises. At 12 weeks p.i., nasal (P 50.004) and ocular (P 5 0.004) discharges as well as palpebral edema (P 5 0.04)and conjunctivitis (P 5 0.04) were greater in infected tortoises than in controltortoises. No clinical signs were seen for control tortoises (n 5 10), whichreceived sterile broth (data not shown). The clusters of four bars for each weekrepresent palpebral edema, conjunctivitis, nasal discharge, and ocular dischargefrom left to right, respectively.

FIG. 2. Specific antibody to M. agassizii in gopher tortoises experimentallyinfected with M. agassizii (■) and control gopher tortoises (h). Levels of anti-body were higher in infected animals than in control animals for all times .4weeks p.i. (P , 0.0001).

FIG. 3. Specific antibody to M. agassizii in gopher tortoises experimentallyinfected with different doses of M. agassizii (1, low dose; o, medium dose; ■,high dose) and control gopher tortoises (h).

VOL. 37, 1999 M. AGASSIZII CAUSES URTD IN GOPHER TORTOISE 2265

on Septem

ber 6, 2018 by guesthttp://jcm

.asm.org/

Dow

nloaded from

observed, like the clinical signs, were variable and appeared tobe independent of infective dose. In at least one tortoise,changes were different in the right versus left nasal cavity,suggesting that intra-animal variation also occurred. In the

high-dose group (n 5 2), one animal had mild to moderateinflammatory changes; the other tortoise showed moderatechanges. Three tortoises in the medium-dose group were se-lected for necropsy. One tortoise that received the medium

FIG. 4. Photomicrograph of the nasal cavity of a control gopher tortoise. (A) Area of mucous and ciliated epithelial cells (M) overlying a lamina propria submucosa(S) primarily consisting of connective tissue and small vessels. (B) Area of multilayered olfactory mucosa (O) overlying a lamina propria submucosa (S) consisting ofconnective tissue, vessels, and melanophores. Hematoxylin and eosin stain was used.

2266 BROWN ET AL. J. CLIN. MICROBIOL.

on Septem

ber 6, 2018 by guesthttp://jcm

.asm.org/

Dow

nloaded from

infective dose had moderate changes, and two tortoises in thisgroup had moderate to severe changes. Three tortoises in thelow-dose group were selected for necropsy. One tortoise thatreceived the low infective dose had moderate changes, onetortoise in this group had moderate to severe changes, and thefinal tortoise in this group had severe changes.

DISCUSSION

The most stringent requirements for definitive proof of acausative relationship between an infectious agent and a dis-ease is the fulfillment of the Henle-Koch-Evans postulates (10,11). In a free-ranging, wild animal which is also legally pro-tected, this is a daunting challenge. The M. agassizii isolateused in these studies was obtained from an animal with clinicaldisease. In the present experimental infection studies, this iso-late was cultured in vitro and was inoculated into clinicallyhealthy animals which were free of mycoplasma infection for aperiod of several months as determined by culture, PCR, andserology. The experimentally infected animals developed clin-ical signs of disease (eight of nine animals) and producedspecific antibody against the infectious agent (nine of nineanimals). Both the appearance of clinical disease and the se-roconversion occurred within reasonable time periods relativeto the time of inoculation of the infectious agent. The lesions

observed in the animals with the experimental infection weresimilar to those observed both in desert tortoises with naturaland experimental infections (4, 16) and in gopher tortoiseswith natural infections (unpublished data). M. agassizii wasrecovered from the experimentally infected animals at varioustimes p.i. Thus, we have fulfilled the Henle-Evans-Koch’s pos-tulates and conclude that M. agassizii causes URTD in thegopher tortoise.

In this study, the patterns of the experimental infections aswell as of the natural infections of gopher tortoises that wereobserved are consistent with existing knowledge of mycoplas-mal respiratory infections in other species. The variability ob-served in the clinical expression of disease among individualanimals is common with other mycoplasmal respiratory infec-tions (32). In most animals, respiratory mycoplasmosis is typ-ified as a slowly progressing, chronic, and seemingly clinicallysilent infection which may be exacerbated by environmentalfactors, stress, or other microbial agents (5, 29, 32, 35). Mosthosts have difficulty in eliminating the mycoplasma, even in thepresence of a strong immune response. In fact, the host im-mune response is critical for the development of lesions (32).Although overt clinical signs may be inapparent, lesions canrange from microscopic to gross, with eventual loss of thenormal respiratory epithelium architecture (4, 5, 16, 32). Theincreased numbers of inflammatory cells, particularly in foci,and the lymphoid hyperplasia observed in the lesions of exper-imentally infected gopher tortoises are consistent with respi-ratory mycoplasmosis in other species, including the deserttortoise (16, 32).

The relative insensitivity of PCR versus the sensitivity ofculture was somewhat surprising. Electron micrographic stud-ies have identified preferentially colonized sites on the mucosalsurfaces of ventrolateral depressions in tortoise upper respira-tory passages, which may not be accessible by swabbing orlavage (data not shown). Since a prolonged incubation is re-quired for culture, small numbers of M. agassizii may be ex-panded to a detectable level as a result of microbial growth.Despite its relative insensitivity, PCR still can play an impor-tant role in the diagnosis of URTD. A positive PCR result canbe obtained with broth cultures after 24 to 48 h of growth, eventhough the initial PCR with the lavage specimen was negative.An additional problem encountered in diagnosis is the qualityof samples. Since gopher tortoises are burrowing animals andmany sample collections are done in the field under less thanideal conditions, the samples are often contaminated with bac-teria and fungi. Many of these, especially fungi, rapidly over-grow the cultures or alter the medium beyond the pH rangetolerated by mycoplasmas for growth. Culture in SP4 broth for48 h before taking an aliquot for PCR enhances detection ofviable mycoplasma, and contamination with other sources ofDNA does not interfere with the PCR (data not shown). Iso-lation of M. agassizii from broth or agar can take up to 6 weeks.If broth cultures are grown for 24 to 48 h and then tested byPCR, we can detect positive culture samples faster, which maybe of primary importance if animals are being quarantined orheld prior to relocation as part of recommended health sur-veillance to prevent the spread of disease.

A wide variety of potential virulence factors have been sug-gested for mycoplasmas, including superantigen production,surface antigenic variation, and host immunomodulation (32).It is not uncommon to see a wide variety in the virulence ofdifferent strains of the same species within a specific host (7, 8,17, 28). The strain selected for our studies was obtained froma very ill, naturally infected tortoise. This particular strainappears to be highly virulent, as evidenced by the fact thatinitial infective doses of only 10 CFU were capable of causing

FIG. 5. Representative moderate to severe changes observed in the upperrespiratory tracts of experimentally infected gopher tortoises. The epithelium (E)is hyperplastic, and there are diffuse accumulations of mixed inflammatory cells(IC) in the lamina propria. Hematoxylin and eosin stain was used.

VOL. 37, 1999 M. AGASSIZII CAUSES URTD IN GOPHER TORTOISE 2267

on Septem

ber 6, 2018 by guesthttp://jcm

.asm.org/

Dow

nloaded from

both clinical disease and severe lesions. We have preliminaryevidence that other strains of M. agassizii do not cause overtclinical disease, even when relatively high infective doses areused. These observations of strain variability are similar tothose observed for respiratory mycoplasmosis in rodents andpoultry (7, 8, 28).

For the most part, any given mycoplasmal species has arelatively limited range of host specificity. Because M. agassiziihas a limited temperature range for growth and does not growat 35°C (3a), it is highly unlikely that it represents a threat tohumans or other mammals. Conversely, evidence suggests thatother chelonians (turtles and tortoises) may be susceptible toM. agassizii (15, 16, 19, 31). In the past several years, we haveseen clinical cases of URTD in tortoises and turtles fromzoological collections as well as private collections. It is notuncommon for these reptiles to be housed together, withoutquarantine or determination of health status. In at least oneinstance, confiscated star tortoises were sent to a zoo and werelater found to have URTD. We have demonstrated the pres-ence of M. agassizii in these animals by serology, culture, andPCR. Treatment with antibiotics alleviates clinical signs butdoes not eliminate the infection (14a). Appropriate quaran-tine, screening, and health surveillance of reptiles in collectionswill be needed to protect animals from URTD. This is partic-ularly important when confiscated shipments of animals withunknown health histories are distributed to zoological settings.Both the gopher and desert tortoises have, to various extents,legal protection due to their decreasing populations. In manycases, the management tools used are relocation of animals towildlife sanctuaries or other suitable habitats. Until recently,wildlife management decisions have not included consider-ations of infectious diseases and the possible impact of thesediseases on population health. The full impact of relocationefforts involving ill animals is yet to be determined and willrequire long-term monitoring studies.

Research on the impact of mycoplasmosis on wildlife hasbeen limited, but reports of recent disease outbreaks in differ-ent wildlife species are provoking interest in mycoplasmosis asa newly emerging (or at least newly recognized) disease threatfor wildlife. The most publicized disease outbreak has beenseen in Mycoplasma gallisepticum infection of house finches,goldfinches, and blue jays (21, 23, 27). In 1993 an epizootic ofpolyarthritis occurred in juvenile farmed crocodiles (Crocody-lus niloticus) in Zimbabwe (18, 25). A mycoplasma was isolatedfrom the joints and lungs of affected crocodiles, was deter-mined to be a previously unrecognized species, and was namedMycoplasma crocodyli (18, 25). In 1995, a systemic infection ofcaptive adult American alligators at a private facility in Floridawas associated with a different species of mycoplasma that hadalso been previously unrecognized. Unlike the outbreak incrocodiles, the disease in alligators was characterized by a veryhigh mortality rate (.70%) and widespread dissemination ofthe infectious agent within the tissues of infected animals (3).

Infectious diseases are an ever present risk to wildlife, par-ticularly during situations in which animals are removed fromtheir natural habitat for captive breeding programs or duringconditions of stress such as release into new habitats, translo-cation, ecosystem perturbation, and encroachment on theirhabitats by urbanization (13, 24, 26, 27). Infectious diseases,their implications for population health, and their impact onthe success of conservation and management plans are rarelyconsidered in management issues. URTD is an excellent ex-ample of the importance of wildlife diseases in populationbiology. Coupled with habitat destruction and environmentalstress factors such as drought, this disease is believed to be amajor factor in declines of desert tortoise populations in the

Mojave Desert (1, 15, 16). Increasing awareness of the role ofinfectious diseases has resulted in inclusion of disease moni-toring and assessment of population health in managementdecisions in both the desert and tortoise populations.

ACKNOWLEDGMENTS

This work was supported by a grant from the Walt Disney WorldCorporation.

We acknowledge the technical assistance of Diane Dukes, MichaelLao, Alyssa Whitmarsh, John Hutchison, and Sylvia Tucker.

REFERENCES1. Berry, K. H. 1997. Demographic consequences of disease in two desert

tortoise populations in California, USA, p. 91–99. In J. Van Ebbema (ed.),Proceedings: conservation, restoration, and management of tortoises andturtles—an international conference. Wildlife Conservation Society TurtleRecovery Program and the New York Turtle and Tortoise Society, NewYork, N.Y.

2. Brown, D. R., B. C. Crenshaw, G. S. McLaughlin, I. M. Schumacher, C. E.McKenna, P. A. Klein, E. R. Jacobson, and M. B. Brown. 1995. Taxonomy ofthe tortoise mycoplasmas Mycoplasma agassizii and Mycoplasma testudinis by16S rRNA gene sequence comparisons. Int. J. Syst. Bacteriol. 45:348–350.

3. Brown, D. R., M. B. Brown, and J. G. Tully. Comparison of mycoplasmaisolated from alligators and crocodiles, abstr. G-10, p. 281. In Abstracts ofthe 97th General Meeting of the American Society for Microbiology 1997.American Society for Microbiology, Washington, D.C.

3a.Brown, M. B. Unpublished data.4. Brown, M. B., I. M. Schumacher, P. A. Klein, K. Harris, T. Correll, and E. R.

Jacobson. 1994. Mycoplasma agassizii causes upper respiratory tract diseasein the desert tortoise. Infect. Immun. 62:4580–4586.

5. Cassell, G. H., W. A. Clyde, and J. K. Davis. 1985. Mycoplasmal respiratorymycoplasmosis, p. 69–107. In S. Razin and M. F. Barile (ed.), The Myco-plasmas, vol. IV. Academic Press, Inc., New York, N.Y.

6. Cox, J., D. Inkley, and R. Kautz. 1987. Ecology and habitat protection needsof gopher tortoise (Gopherus polyphemus) populations found on lands slatedfor large-scale development in Florida. Technical report no. 4. Florida Gameand Fresh Water Fish Commission Nongame Wildlife Program, Tallahassee.

7. Davidson, M. K., J. K. Davis, J. R. Lindsey, and G. H. Cassell. 1988.Clearance of different strains of Mycoplasma pulmonis from the respiratorytract of C3H/HeN mice. Infect. Immun. 56:2163–2168.

8. Davidson, M. K., J. R. Lindsey, R. F. Parker, J. G. Tully, and G. H. Cassell.1988. Difference in virulence for mice among strains of Mycoplasma pulmo-nis. Infect. Immun. 56:2156–2162.

9. Diemer, J. E. 1986. The ecology and management of the gopher tortoise inthe United States. Herpetologica 42:125–133.

10. Evans, A. S. 1976. Causation and disease: the Henle-Koch postulates revis-ited. Yale J. Biol. Med. 49:175–195.

11. Evans, A. S. 1977. Limitations of Koch’s postulates. Lancet ii:1277–1278.12. Hansen, K. L. 1963. The burrow of the gopher tortoise. J. Fl. Acad. Sci.

26:353–360.13. Hutchins, M., and U. S. Seal. 1991. The role of veterinary medicine in

endangered species conservation. J. Zoo Wildl. Med. 22:277–281.14. Jackson, D. R., and E. G. Milstrey. 1989. The fauna of gopher tortoise

burrows, p. 86–98. In J. E. Diemer, D. R. Jackson, J. L. Landers, J. N. Layne,and D. A. Wood (ed.), Gopher Tortoise Relocation Symposium proceedings.Technical report no. 5. Florida Game and Fresh Water Fish Commission,Tallahassee.

14a.Jacobson, E. R. Unpublished data.15. Jacobson, E. R., M. B. Brown, I. M. Schumacher, B. R. Collins, R. K. Harris,

and P. A. Klein. 1995. Mycoplasmosis and the desert tortoise (Gopherusagassizii) in Las Vegas Valley, Nevada. Chel. Cons. Biol. 1:280–284.

16. Jacobson, E. R., J. M. Gaskin, M. B. Brown, R. K. Harris, C. H. Gardiner,J. L. LaPointe, H. P. Adams, and C. Reggiardo. 1991. Chronic upper respi-ratory tract disease of free-ranging desert tortoises, Xerobates agassizii. J.Wildl. Dis. 27:296–316.

17. Jones, G. E., J. S. Gilmour, and A. G. Rae. 1982. The effects of differentstrains of Mycoplasma ovipneumoniae on specific pathogen-free and conven-tionally-reared lambs. J. Comp. Pathol. 92:267–272.

18. Kirchhoff, H., K. Mohan, R. Schmidt, M. Runge, D. R. Brown, M. B. Brown,C. M. Foggin, P. Muvavirirwa, H. Lehmann, and J. Fossdorf. Mycoplasmacrocodyli sp. nov., a new species from crocodiles. Int. J. Syst. Bacteriol.47:742–746.

19. Lawrence, K., and J. R. Needham. 1985. Rhinitis in long-term captive Med-iterranean tortoises (Testudo graeca and T. hermanni). Vet. Rec. 117:662–664.

20. Levell, J. P. 1995. A field guide to reptiles and the law. Serpent’s Tale,Excelsior, Minn.

21. Ley, D. H., J. E. Berkhoff, and J. M. McLaren. 1996. Mycoplasma gallisepti-cum isolated from house finches (Carpodacus mexicanus) with conjunctivitis.Avian Dis. 40:480–483.

2268 BROWN ET AL. J. CLIN. MICROBIOL.

on Septem

ber 6, 2018 by guesthttp://jcm

.asm.org/

Dow

nloaded from

22. Lips, K. R. 1991. Vertebrates associated with tortoise (Gopherus poly-phemus) burrows in four habitats in south-central Florida. J. Herpetol. 25:477–481.

23. Luttrell, M. P., J. R. Fischer, D. E. Stallknecht, and S. H. Kleven. 1996. Fieldinvestigation of Mycoplasma gallisepticum infections in house finches (Car-podacus mexicanus) from Maryland and Georgia. Avian Dis. 40:335–341.

24. Miller, R. E. 1992. Zoo veterinarians—doctor’s on the ark? J. Am. Vet. Med.Assoc. 5:642–647.

25. Mohan, K., C. M. Foggin, P. Muvavarirwa, J. Honywill, and A. Pawandiwa.1995. Mycoplasma-associated polyarthritis in farmed crocodiles (Crocodylusniloticus) in Zimbabwe. Ond. J. Vet. Res. 62:45–49.

26. Nettles, V. F. 1992. Wildlife diseases. J. Am. Vet. Med. Assoc. 5:648–652.27. Nettles, V. F. 1996. Reemerging and emerging infectious diseases: economic

and other impacts on wildlife. ASM News 62:589–591.28. Rodriguez, R., and S. H. Kleven. 1980. Pathogenicity of two strains of

Mycoplasma gallisepticum in broilers. Avian Dis. 24:800–807.29. Schoeb, T. R., K. C. Kervin, and J. R. Lindsey. 1985. Exacerbation of murine

respiratory mycoplasmosis in gnotobiotic F344/N rats by Sendai virus infec-tion. Vet. Pathol. 22:272–282.

30. Schumacher, I. M., M. B. Brown, E. R. Jacobson, B. R. Collins, and P. A.Klein. 1993. Detection of antibodies to a pathogenic mycoplasma in deserttortoises (Gopherus agassizii) with upper respiratory tract disease. J. Clin.Microbiol. 31:1454–1460.

31. Schumacher, I. M., D. B. Hardenbrook, M. B. Brown, E. R. Jacobson, andP. A. Klein. 1997. Relationship between clinical signs of upper respiratorytract disease and antibodies to Mycoplasma agassizii in desert tortoises fromNevada. J. Wildl. Dis. 33:261–266.

32. Simecka, J. W., J. K. Davis, M. K. Davidson, S. E. Ross, C. T. K. H.Stadtlander, and G. H. Cassell. 1992. Mycoplasmas which infect animals, p.391–415. In J. Maniloff, R. N. McElhaney, L. R. Finch, and J. B. Baseman(ed.), Mycoplasmas: molecular biology and pathogenesis. American Societyfor Microbiology, Washington, D.C.

33. Tully, J. G., D. L. Rose, R. F. Whitcomb, and R. P. Wenzel. 1979. Enhancedisolation of Mycoplasma pneumoniae from throat washings with a newlymodified culture medium. J. Infect. Dis. 139:478–482.

34. Van Kuppeveld, F. J. M., J. T. M. Van Der Logt, A. F. Angulo, M. J. VanZoest, W. G. V. Quint, H. G. M. Niesters, J. M. D. Galama, and W. J. G.Melchers. 1992. Genus- and species-specific identification of mycoplasmasby 16s rRNA amplification. Appl. Environ. Microbiol. 58:2602–2615.

35. Weinach, O. M., G. H. Snoeyenbos, C. F. Smeyser, and A. S. Soerjude-Liem.1985. Influence of Mycoplasma gallisepticum, infectious bronchitis, and cy-clohexamide on chickens protected by native intestinal microflora againstSalmonella typhimurium or Escherichia coli. Avian Dis. 28:416–4425.

36. Witz, B. W., D. S. Wilson, and M. D. Palmer. 1991. Distribution of Gopheruspolyphemus and its vertebrate symbionts in three burrow categories. Am.Mid. Nat. 126:152–158.

VOL. 37, 1999 M. AGASSIZII CAUSES URTD IN GOPHER TORTOISE 2269

on Septem

ber 6, 2018 by guesthttp://jcm

.asm.org/

Dow

nloaded from