Strain UCMJ was isolated from a swab of the choana used to inoculate fortified commercial (FC) broth (11), incubated at 37°C for 48 h, and then plated onto FC agar plates incubated in a humidified environment at 37°C. Filter cloning of the wild-type isolate was carried out three times in Frey's medium (4) using 0.22-μm membrane filters (15). Strain UCMJ was isolated in and subcultured on Frey's broth and agar, respectively, and cultures were routinely incubated at 37°C in ambient air and humidity. For ultrastructural studies, the organism was pelleted from a mid-log-phase broth culture and fixed in 1.5% formaldehyde and 1.5% glutaraldehyde with 3 mM MgCl₂ in either 0.1 M cacodylate buffer at pH 7.2 or 0.1 M HEPES. Samples for scanning electron microscopy were postfixed in 1% osmium tetroxide in 0.1 M cacodylate buffer, subjected to three 10-min washes with distilled ultrafiltered water, dehydrated in a graded ethanol series, critical point dried for 3.5 h, mounted on metal stubs using silver paint, sputter-coated with gold-palladium, and examined using a Zeiss DSM 982 Gemini field emission scanning electron microscope. Samples for transmission electron microscopy were centrifuged at 8,000 × g for 15 min and then resuspended in fixative. Fixed cells were centrifuged at 12,000 × g for 10 min, washed three times in 0.1 M cacodylate buffer, postfixed in 1% OsO₄ in 0.1 M cacodylate buffer, washed three times in distilled water, and then gently resuspended in 1% agar in 0.1 M cacodylate buffer. The agar block was allowed to harden at 4°C and then was cut into pieces depending on the apparent sample. Each piece was stained en bloc with 1% uranyl acetate in 0.1 M maleate buffer, dehydrated in ethanol, and embedded in LR White (SPI Supplies, West Chester, Pa.) using a microwave technique (6, 7) and a Pelco model 3440 microwave processor with variable wattage, a vacuum chamber, and a water recirculation table (ColdSpot; TedPella Inc., Redding, Calif.). Ultrathin sections were stained with 1% lead citrate and viewed using a Philips 300 electron microscope at 80 kV.

For the hemadsorption assay, two milliliters of a 0.5% suspension of washed sheep red blood cells (Quad Five, Ryegate, Mont.) in 1× phosphate-buffered saline were added to plates containing well-separated colonies of either M. gallisepticum strain R (positive control) or strain UCMJ and then incubated at 37°C for 30 min. Plates were washed with 2 ml of 1× phosphate-buffered saline to remove excess red blood cells, and hemadsorption results were determined after viewing the cells through an inverted compound microscope. The requirement of sterol for growth was determined by the digitonin disk inhibition procedure of Poveda (12). Serological identification was assayed by a standard disk growth inhibition technique (2) using 19 antisera to the following mycoplasmas within the Mycoplasma hominis cluster: M. hominis, M. alkalescens, M. bovis, M. bovigenitalium, M. felifaucium, M. fermentans, M. iners, M. leopharyngis, M. lipofaciens, M. opalescens, M. spermatophilum, M. equigenitalium, M. gypis, M. lipophilum, M. neurolyticum, M. pulmonis, M. sualvi, M. synoviae, and M. hyopharyngis (obtained fromMaureen K. Davidson, curator of the Mollicutes Collection, Purdue University, West Lafayette, Ind.). The G+C content of the organism's genomic DNA was determined by the fluorometric dye-binding method described by Gerhardt et al. (5) using Escherichia coli DH5α (50.8%), Pseudomonas aeruginosa 27853 (66.6%), and Mycoplasma gallisepticum R strain (31%) as known standards.

DNA for phylogenetic analyses was extracted from a mid-log-phase culture of the fourth passage of a clonal isolate using the Easy DNA kit (Invitrogen, Carlsbad, Calif.). The 16S rRNA gene was amplified using primers SG170 and SG171 (3), cloned into the pCR 2.1 TOPO vector (Invitrogen), and then transformed into E. coli Top10 competent cells. Plasmid DNA was purified using the QIAprep Spin Miniprep kit (QIAGEN Inc., Santa Clarita, Calif.), and sequencing was performed using a Beckman CEQ sequencer at the University of Connecticut Bioservices Center (Storrs, Conn.). The resultant 16S rRNA gene sequence was compared to archived genetic sequences using standard nucleotide-nucleotide BLAST searches made of the GenBank database (1). The 16S rRNA gene sequence of strain UCMJ was included in an alignment constructed using ClustalX v1.81 (14) and comprised of 16S rRNA gene sequences representative of known mycoplasmas, members of recognized mycoplasma groupings, and members of the genera Aneroplasma, Acholeplasma, and Lactobacillus. Phylogenetic trees were inferred by distance and parsimony algorithms using PAUP* 4.0b10 (13). The data set consisted of 46 taxa with 1,670 total characters, of which 827 were constant, 161 were variable but parsimony uninformative, and 682 were informative in parsimony analysis. A bootstrapped maximum-parsimony tree was generated using heuristic search parameters and 1,000 bootstrap resamplings with a starting tree obtained via random stepwise addition and tree bisection-reconnection branch swapping. A distance-based phylogenetic tree was reconstructed using the criterion of minimum evolution optimality and the HKY85 substitution model (8), with among-site rate variation being assumed to follow a gamma distribution of 0.5 and 1,000 bootstrap resamplings of the data set. Heuristic search parameters were used, and the starting tree(s) was obtained via neighbor-joining, followed by tree bisection-reconnection branch swapping.

Colonies with a typical fried-egg appearance were apparent on FC agar plates after 5 days of incubation in a humidified environment at 37°C. Scanning electron photomicrographs revealed that strain UCMJ in mid-log-phase culture was present as spherical, flask-shaped, and elongate forms ranging from 0.3 to 0.6 μm in width, with short, unipolar tip structures evident in a small number of organisms examined (Fig. 1A). Transmission electron photomicrographs revealed that the cytoplasmic margins of cells were formed by a single lipid bilayer, which is typical of members of the Mycoplasmatales, e.g., Mycoplasma sturni (3), and no walled forms were evident (Fig. 1B). Strain UCMJ was determined to be hemadsorption negative. Growth was inhibited in the presence of 1.5% digitonin, and zones of inhibition ranged from 6 to 10 mm, indicating a biochemical requirement for sterols characteristic of members of the Mycoplasmatales and not Acholeplasmatales. Strain UCMJ did not react with any antisera tested and was therefore determined to be serologically unique. The G+C content of strain UCMJ was determined to be 28%. The 16S rRNA gene sequence of strain UCMJ was unique, having the greatest nucleotide sequence identity with Mycoplasma hyopharyngis (94%). Phylogenetic analyses using both parsimony and distance algorithms were concordant. Strain UCMJ associated with the mycoplasmas but branched separately from other known mycoplasmas. Although it did not form a distinct cluster with any other species, it did occasionally appear as a sister taxon to M. hyopharyngis (Fig. 2).

FIG. 1. Mid-log-phase cultures of Mycoplasma sphenisci. (A) Scanning electron photomicrograph demonstrating the spherical, flask-shaped, and elongate forms encountered, some of which have tip structures at one pole. Bar = 1 μm. (B) Transmission (more ...)

FIG. 2. Phylogram using 16S rRNA gene sequences from Mycoplasma sphenisci, other mycoplasmas, anaeroplasmas, and acholeplasmas, with Lactobacillus sp. as a designated outgroup. Numbers at nodes indicate confidence values after 1,000 bootstrap replicates for parsimony (more ...)

Based on standards for the description of new species of mycoplasmas (9), this organism qualifies as a unique, previously unidentified and undescribed mycoplasma species. The name Mycoplasma sphenisci sp. nov. is proposed in recognition of the genus Spheniscus, the genus of penguin that includes the jackass penguin (Spheniscus demersus) from which this mycoplasma was isolated. Characterization of M. sphenisci from a jackass penguin broadens our understanding of the range of avian species capable of hosting mycoplasmas. Although the original source of M. sphenisci was not determined, its isolation was associated with recurrent mucocaseous choanal discharge, and no other bacterium was recovered from the choanal swab that yielded the mycoplasma isolate. However, inferences suggesting that the recurrent choanal discharge should be attributable to M. sphenisci must be viewed speculatively, as Koch's postulates have not been fulfilled. Isolation of Mycoplasma sphenisci from a jackass penguin represents a heretofore-unrecognized bacterium-host interaction that may impact clinical and diagnostic concerns of microbiologists, veterinarians, and husbandry specialists participating in the care of zoological, aquarium, and wildlife collections.

The 16S rRNA gene sequence of strain UCMJ is available through the GenBank database under accession number AY756171.

We thank Andrew Routh, Leslie Boerner, and Gary Egerie for submission of culture swabs leading to isolation and the Husbandry and Veterinary Services staff of New England Aquarium for assistance with records and historical information. We are grateful to Terri Wheeler (IDEXX Laboratories, North Grafton, Mass.) for sharing the initial isolate with us. We thank Stephen Daniels of the Electron Microscopy Laboratory, University of Connecticut (Storrs, Conn.) for assistance with transmission electron photomicroscopy.