Short communication

Veterinary Microbiology 137 (2009) 380–383

Contents lists available at ScienceDirect

Veterinary Microbiology

journal homepage: www.elsev ier .com/ locate /vetmic

Secreted sialidase activity of canine mycoplasmas

Meghan May, Daniel R. Brown *

Department of Infectious Diseases and Pathology, College of Veterinary Medicine, University of Florida, Gainesville, FL 32611-0880, USA

A R T I C L E I N F O

Article history:

Received 16 October 2008

Received in revised form 19 December 2008

Accepted 2 January 2009

Keywords:

Dogs

Mycoplasma

Sialidase

A B S T R A C T

Through a survey of the phylogenetic distribution of sialidase among mycoplasmas, we

detected activity secreted by the type strains of 3 of 11 species frequently or first isolated

from dogs. The specific activity of washed cells of the type strains of Mycoplasma canis,

Mycoplasma cynos, and Mycoplasma molare ranged from 5.2 � 0.8 � 10�6 to

1.1 � 0.3 � 10�5 enzymatic units per colony-forming unit (U/CFU). Cells of M. molare strain

H542T had twice the specific activity (P < 0.05) of M. canis strain PG14T or M. cynos strain

H831T. Significant differences in sialidase activity existed among nine clinical isolates of M.

canis, ranging from not detectable to 2.1 � 0.1 � 10�5 U/CFU. The type strains of other species

previously isolated from dogs (Mycoplasma arginini, Mycoplasma bovigenitalium, Mycoplasma

edwardii, Mycoplasma felis, Mycoplasma gatae, Mycoplasma maculosum, Mycoplasma opalescens,

and Mycoplasma spumans) did not exhibit either secreted or cell-associated sialidase activity.

Neither specific nor degenerate PCR primers complementary to the three known mycoplasmal

sialidase alleles were able to amplify orthologs in M. canis, M. cynos, or M. molare, further

evidence that the secreted sialidase of those species is distinct from the strictly cell-associated

sialidases of Mycoplasma alligatoris, Mycoplasma synoviae, and Mycoplasma gallisepticum. This

is the first report of a well-known bacterial virulence factor whose expression varies among

strains of certain Mycoplasma species that infect dogs.

� 2009 Elsevier B.V. All rights reserved.

1. Introduction

Canine mycoplasmosis is associated with multiplespecies of Mycoplasma. The most firmly establishedpathogen is Mycoplasma cynos (Røsendal, 1973), whichwas isolated from fatal infections of puppies, and proven tobe a primary respiratory pathogen by experimentalinfection studies (Røsendal, 1982; Zeugswetter et al.,2007). Mycoplasma canis (Edward, 1955) is best known asan opportunistic pathogen of dogs, and has been recoveredfrom their respiratory, urinary and reproductive tracts.Unlike M. cynos, M. canis often infects other mammalianhosts. For example, M. canis was isolated from the lower

* Corresponding author. Tel.: +1 352 392 2239x3975;

fax: +1 352 392 9704.

E-mail address: BrownD@vetmed.ufl.edu (D.R. Brown).

0378-1135/$ – see front matter � 2009 Elsevier B.V. All rights reserved.

doi:10.1016/j.vetmic.2009.01.009

respiratory tracts of humans with pneumonia (Armstronget al., 1971) and both healthy and pneumonic cattle (Aylinget al., 2004), although studies of M. canis strain C3b as aprimary pathogen of cattle had equivocal results (ter Laaket al., 1993). Mycoplasma molare (Røsendal, 1974) andseveral other species isolated from dogs are also potentialpathogens (Chalker, 2005).

Sialidase is a virulence factor of diverse microorgan-isms. It promotes microbial colonization, tissue invasion,and damage to sialylated host molecules, cell surfaces andthe extracellular matrix (Corfield, 1992). Although Myco-

plasma alligatoris strain A21JP2T, virulent strains ofMycoplasma gallisepticum and Mycoplasma synoviae, andsome strains of other avian mycoplasmas produce cell-associated sialidases (Brown et al., 2004; Bercic et al.,2008a; May and Brown, 2008), the distribution of sialidaseamong other species of Mycoplasma has not beeninvestigated systematically. As part of a broader survey,

M. May, D.R. Brown / Veterinary Microbiology 137 (2009) 380–383 381

we detected secreted sialidase activity in 3 of 11 speciesisolated from dogs. This is the first report of a well-knownbacterial virulence factor occurring in pathogenic speciesof Mycoplasma that infect dogs.

2. Materials and methods

2.1. Mycoplasma isolates and culture techniques

The type strains Mycoplasma arginini G230T, Myco-

plasma bovigenitalium PG11T, M. canis PG14T, M. cynos

H831T, Mycoplasma edwardii PG24T, Mycoplasma felis COT,Mycoplasma gatae CST, Mycoplasma maculosum PG15T, M.

molare H542T, Mycoplasma opalescens MH5408T, andMycoplasma spumans PG13T were obtained from theAmerican Type Culture Collection. Five isolates collectedfrom Shetland sheepdogs with various reproductivedisorders, and four isolates from dogs with unknownclinical history, which were obtained during a prior survey(M.B. Brown, unpublished), were identified as M. canis byPCR–RFLP screening (Spergser and Rosengarten, 2007)confirmed by direct 16S rRNA gene sequencing (May et al.,2007).

M. canis, M. cynos, M. edwardii, M. felis, and M. molare

were propagated in SP-4 medium containing 0.5% (w/v)glucose. M. arginini, M. bovigenitalium, M. gatae, M.

maculosum, M. opalescens, and M. spumans were propa-gated in SP-4 medium containing 0.5% (w/v) glucose plus0.21% (w/v) L-arginine. All cultures were incubated at 37 8Cunder ambient atmospheric conditions. Culture concen-trations were determined by inoculating serial 10-folddilutions in SP-4 broth onto agar. Colonies were countedafter 5 days of incubation.

2.2. Measurement of sialidase activity

Sialidase activity of washed cells was quantitated usingthe fluorogenic substrate 20-(4-methylumbelliferyl)-a-D-N-acetylneuraminic acid (MUAN; Sigma–Aldrich, St. Louis,MO) and the sialidase inhibitor 2-deoxy-2,3-didehydro-N-acetylneuraminic acid (DANA; Sigma–Aldrich) as pre-viously described (May et al., 2007). The positive controlwas M. alligatoris, which expresses cell-associated but notsecreted sialidase. The negative controls were Mycoplasma

Table 1

Sialidase-positive mycoplasmas.

Mycoplasma species Strain Sialidase activitya

M. alligatoris A21JP2T 10�9 U/CFU

M. anseris 1219T Moderate

M. canisb,c PG14T 10�6 U/CFU

M. cloacale 383 Weak

M. corogypsi BV1T Very strong

M. cynos H831T 10�6 U/CFU

M. gallisepticumb,c R 10�8 U/CFU

M. iowae serovar I 695T Weak

M. meleagridisc 17529T Strong

M. molare H542T 10�5 U/CFU

M. synoviaeb,c WVU1853T 10�7 U/CFU

a Highest reported specific activity (enzymatic units per colony-forming unib Species with multiple strains documented to exhibit sialidase activity.c Species with some strains documented to lack sialidase activity.

crocodyli strain MP145T, which does not express anysialidase (Brown et al., 2004), and fresh culture medium.Secreted sialidase activity was detected qualitatively inlate log-phase broth cultures, after the cells wereharvested, by diluting the cell-free conditioned super-natant medium 1:1 with 0.35% (w/v) MUAN, thenobserving the specimen under shortwave ultravioletillumination. The positive control was Type VI purifiedClostridium perfringens sialidase in cell-free M. crocodyli-conditioned medium, and the negative control was plaincell-free M. crocodyli-conditioned medium.

2.3. PCR amplification of sialidase genes

PCR primers complementary to the cell-associatedsialidase gene nanI sequences of M. alligatoris (GenBankAY515695; 50-TGA CAA AAT GCG CTG AAA AA-30 and 50-GCG CCA AAT TTA CAT CCT ACA-30), M. synoviae (GenBankNC_007294; 50-TCT CTT CCT TTT TGA GGG CTA-30 and 50-GCA AAT CAT CTT AAG AAA AGT CAT T-30), and M.

gallisepticum (GenBank NC_004829; 50-TCA GAT CAT TAAACT AGC GCC TAA-30 and 50-CGC ATG ATA CGA TAA CGAAAT G-30), and degenerate primers designed from con-sensus sequences among those alleles (50-GA(CT) G(AG)IGGI (ACT)(AT)I (AT)(GC)I TG(AG)-30 and 50-CAI (GCT)(AT)I(AGT)(CT)I CCI CC(AG) TC-30) were used in attempts todetect the sialidase genes in M. canis, M. cynos, and M.

molare. Genomic DNA templates were purified usingEasyDNA reagents (Invitrogen, Carlsbad CA). The 50-mlreactions included 1 mmol of each primer and 100 ng oftemplate DNA, and consisted of initial template denatura-tion at 94 8C for 2 min, 30 cycles of denaturation at 94 8Cfor 20 s, primer annealing at 48 8C for 20 s, and extension at72 8C for 3 min, with a final extension at 72 8C for 10 min.Reactions using degenerate primers were performed withan annealing temperature of 38 8C. The positive controltemplate was M. gallisepticum strain Rlow genomic DNA.

2.4. Statistical analysis

The effect of species on the specific sialidase activity ofwashed M. canis, M. cynos, and M. molare cells, and theeffect of strain within M. canis, were analyzed by analysesof variance (n = 3 independent replications each), and by

Host class Phylogenetic cluster Reference

Reptile M. synoviae Brown et al., 2004

Bird M. hominis Bercic et al., 2008a

Mammal M. synoviae Current report

Bird M. hominis Bercic et al., 2008a

Bird M. synoviae Bercic et al., 2008a

Mammal M. synoviae Current report

Bird M. pneumoniae May and Brown, 2008

Bird M. muris Bercic et al., 2008a

Bird M. bovis Bercic et al., 2008a

Mammal M. neurolyticum Current report

Bird M. synoviae May et al., 2007

t) or semi-quantitative estimate for washed cells.

Fig. 1. Intraspecific variation in sialidase activity of washed Mycoplasma

canis cells. Bars depict mean � standard error of enzyme units (U) per

colony-forming unit (CFU) of clinical isolates and the type strain PG14T.

Means with a different letter are different (P < 0.05 to P < 0.001) by Fisher’s

protected least significant difference test (n = 3 independent replications

each). The clinical isolate LV had no activity.

M. May, D.R. Brown / Veterinary Microbiology 137 (2009) 380–383382

Fisher’s protected least significant difference test for posthoc comparisons among species or strains when maineffects were significant (May and Brown, 2008). P values<0.05 were considered significant.

3. Results

As part of a survey of the genus Mycoplasma, this studysought to detect the presence of sialidase in primary andopportunistic mycoplasmal pathogens that infect dogs.Sialidase activity was evident in washed cells of the typestrains of M. canis, M. cynos, and M. molare. The specificactivity ranged from 5.2 � 0.8 � 10�6 to 1.1 � 0.3 � 10�5 U/CFU (Table 1). M. molare strain H542T had twice as much

Fig. 2. Secreted sialidase activity of canine mycoplasmas. (A) Sialidase was

detected by release of fluorescent 4-methylumbelliferone from labeled N-

acetylneuraminic acid. Activity was present in washed cells (C) and cell-

free conditioned supernatant SP-4 broth (S) from cultures of Mycoplasma

cynos H831T and Mycoplasma molare H542T. No activity was detected in

the type strains of other species isolated from dogs. Mycoplasma alligatoris

A21JP2T was the positive control for strictly cell-associated activity;

Mycoplasma crocodyli MP145T and fresh SP-4 broth were negative

controls. (B) Sialidase activity was present in Mycoplasma canis PG14T and

eight clinical isolates of M. canis. Isolate LV had no activity.

activity (P < 0.05) as M. canis strain PG14T or M. cynos strainH831T. Nine clinical isolates of M. canis were subsequentlyexamined to sample the extent of intraspecific variation.Washed cells of those isolates ranged in specific activity(P < 0.0001) from 0 (undetectable) for strain LV, to2.1 � 0.1 � 10�5 U/CFU for strain 31 (Fig. 1).

Sialidase activity was also present unexpectedly in cell-free conditioned supernatant broth from cultures of M.

cynos H831T, M. molare H542T (Fig. 2A), and all strains of M.

canis except LV (Fig. 2B). To investigate whether thesecreted activity is conferred by an alternative form of theknown mycoplasmal sialidase alleles, conserved anddegenerate PCR primers were used in attempts to detectorthologous genes, but no amplification was successfulfrom the genomic DNA of M. cynos, M. molare, or any strainof M. canis. No sialidase activity was detected in the otherspecies examined.

4. Discussion and conclusion

Sialidase was formerly considered to be rare amongmycoplasmas (Kahane et al., 1990), but its presence hasnow been documented in 11 species. All mycoplasmascurrently known to produce sialidase either are affiliatedwith the M. synoviae phylogenetic cluster on the basis of16S rRNA gene similarity (Johansson and Pettersson,2002), or they share a common host with one or moresialidase-positive species in that cluster (Table 1). Con-sistent with that observation, M. canis and M. cynos of theM. synoviae cluster share the secreted form of sialidasewith M. molare, a species affiliated with the phylogeneti-cally distant Mycoplasma neurolyticum cluster.

We interpret the activity in washed cells of M. canis, M.

cynos and M. molare to reflect newly synthesized enzymewith the potential to be secreted (Fig. 2), although it seemspossible that distinct secreted and cell-associated siali-dases occur together in some instances. The finding thatsialidase was secreted into the culture medium by 9 of the10 strains of M. canis examined is remarkable because theone unnamed M. canis isolate examined by Zakrajsek(2008) had only membrane-associated activity. In a recentabstract, Bercic et al. (2008b) reported that sialidaseactivity was present in conditioned supernatant mediumfrom a culture of an unspecified Mycoplasma corogypsi

strain, another species from birds that is affiliated with theM. synoviae phylogenetic cluster.

The fourfold variation among M. canis strains in specificsialidase activity of washed cells was statistically significantbut much less than the 65-fold variation between high- andlow-virulence strains of M. synoviae (May et al., 2007). Theabsence of activity in M. canis isolate LV underscores thenecessity to assess multiple strains of each species tocharacterize the phylogenetic distribution of sialidases. Thismay also explain the conflict between earlier reports on theoccurrence of sialidase in Mycoplasma when only singlestrains were examined (Roberts, 1967; Glasgow and Hill,1980), and the equivocal resultsregarding M.canisstrainC3bas a primary pathogen of cattle (ter Laak et al., 1993) if thatstrain lacked sufficient activity to contribute to virulence.

M. cynos and M. canis are the species most consistentlyassociated with canine mycoplasmosis (Chalker, 2005), but

M. May, D.R. Brown / Veterinary Microbiology 137 (2009) 380–383 383

to date no specific virulence factor had been proposed as abasis of their pathogenicity. Dogs inoculated with M. canis

strains S6/L42 or SL1/L45 developed urethritis andprostatitis or metritis (Laber and Holzmann, 1977;Holzmann et al., 1979). M. canis strain A56Hzkl, isolatedfrom the pericardium of a dog, colonized the lung, liver andspleen of mice inoculated intraperitoneally, but did notgenerate lesions (Eberle et al., 1977). Erythrocyte surfacedesialylation reduced hemadsorption by M. canis strainPG14T (Manchee and Taylor-Robinson, 1969), suggestingthat the secreted sialidase might modulate mycoplasmalcytadherence, colonization and transmission (May andBrown, 2008).

In conclusion, we propose that a secreted sialidase is acandidate virulence factor of M. cynos and M. canis. ThoughM. molare is not a definitive canine pathogen, its ability tosecrete sialidase in the urogenital tract may allow it also toact as a primary or co-factor in disease. Further studies willbe required to identify the sialidase genes in thesemycoplasmas, and to examine a possible link betweensialidase expression and pathogenesis during caninemycoplasmosis.

Acknowledgements

Clinical isolates of M. canis were a gift from Dr. Mary B.Brown of the University of Florida. This work was supportedby Public Health Service grant 1R01GM076584 from theNational Institute of General Medical Sciences (DRB).

References

Armstrong, D., Yu, B.H., Yagoda, A., Kagnoff, M.F., 1971. Colonization ofhumans by Mycoplasma canis. J. Infect. Dis. 124, 607–609.

Ayling, R.D., Bashiruddin, S.E., Nicholas, R.A.J., 2004. Mycoplasma speciesand related organisms isolated from ruminants in Britain between1990 and 2000. Vet. Rec. 155, 413–416.

Bercic, R.L., Slavec, B., Lavric, M., Narat, M., Zorman-Rojs, O., Dovc, P.,Bencina, D., 2008a. A survey of avian Mycoplasma species for neur-aminidase enzymatic activity. Vet. Microbiol. 130, 391–397.

Bercic, R.L., Dovc, P., Cizelj, I., Kalan, J., Zakrajsek, T., Narat, M., Bencina, D.,2008b. Neuraminidase enzymatic activity in mycoplasmas of theMycoplasma synoviae cluster. In: 17th International Congress of theInternational Organization for Mycoplasmology (Abstr.).

Brown, D.R., Zacher, L.A., Farmerie, W.G., 2004. Spreading factors ofMycoplasma alligatoris, a flesh-eating mycoplasma. J. Bacteriol. 186,3922–3927.

Chalker, V.J., 2005. Canine mycoplasmas. Res. Vet. Sci. 79, 1–8.Corfield, T., 1992. Bacterial sialidases – roles in pathogenicity and nutri-

tion. Glycobiology 2, 509–521.Eberle, G., Kirchhoff, H., Trautwein, G., 1977. Experimental infection of

mice, gerbils, and rats with mycoplasms from canine pericardium andcardiac valve. Zentralbl. Bakteriol. [Orig. A] 239, 95–103.

Edward, D.G., 1955. A suggested classification and nomenclature fororganisms of the pleuropneumonia group. Int. Bull. Bacteriol.Nomencl. Taxon. 5, 85–93.

Glasgow, L.R., Hill, R.L., 1980. Interaction of Mycoplasma gallisepticum withsialyl glycoproteins. Infect. Immun. 30, 353–361.

Holzmann, A., Laber, G., Walzl, H., 1979. Experimentally induced myco-plasmal infection in the genital tract of the female dog. Theriogenol-ogy 12, 355–370.

Johansson, K.-E., Pettersson, B., 2002. Taxonomy of mollicutes. In: Razin,S., Herrmann, R. (Eds.), Molecular Biology and Pathogenicity of Myco-plasmas. Kluwer Academic/Plenum, New York, pp. 1–30.

Kahane, I., Reisch-Saada, A., Almagor, M., Abeliuck, P., Yatziv, S., 1990.Glycosidase activities of mycoplasmas. Zentralbl. Bakteriol. [Orig. B]273, 300–305.

Laber, G., Holzmann, A., 1977. Experimentally induced mycoplasmalinfection in the genital tract of the male dog. Theriogenology 7,177–188.

Manchee, R.J., Taylor-Robinson, D., 1969. Utilization of neuraminic acidreceptors by mycoplasmas. J. Bacteriol. 98, 914–919.

May, M., Kleven, S.H., Brown, D.R., 2007. Sialidase activity in Mycoplasmasynoviae. Avian Dis. 51, 829–833.

May, M., Brown, D.R., 2008. Genetic variation in sialidase and linkage to N-acetylneuraminate catabolism in Mycoplasma synoviae. Microb.Pathog. 45, 38–44.

Roberts, D.H., 1967. Neuraminidase-like enzyme present in Mycoplasmagallisepticum. Nature 213, 87–88.

Røsendal, S., 1973. Mycoplasma cynos, a new canine Mycoplasma species.Int. J. Syst. Bacteriol. 23, 49–54.

Røsendal, S., 1974. Mycoplasma molare, a new canine Mycoplasma species.Int. J. Syst. Bacteriol. 24, 125–130.

Røsendal, S., 1982. Canine mycoplasmas: their ecological niche and role indisease. J. Am. Vet. Med. Assoc. 180, 1212–1214.

Spergser, J., Rosengarten, R., 2007. Identification and differentiation ofcanine Mycoplasma isolates by 16S-23S rDNA PCR–RFLP. Vet. Micro-biol. 125, 170–174.

ter Laak, E.A., van Dijk, J.E., Noordergraaf, J.H., 1993. Comparison ofpathological signs of disease in specific-pathogen-free calves afterinoculation of the respiratory tract with Ureaplasma diversum orMycoplasma canis. J. Comp. Pathol. 108, 121–132.

Zakrajsek, T., 2008. Nevraminidazna aktivnost bakterije Mycoplasma canis[Neuraminidase activity of the bacterium Mycoplasma canis]. Thesis,University of Ljubljana.

Zeugswetter, F., Weissenbock, H., Shibly, S., Hassan, J., Spergser, J., 2007.Lethal bronchopneumonia caused by Mycoplasma cynos in a litter ofgolden retriever puppies. Vet. Rec. 161, 626–627.