Veterinary Microbiology, 17 (1988) 45-57 45 Elsevier Science Publishers B.V., Amsterdam w Printed in The Netherlands

Recognition of Mycoplasma arthritidis Membrane Antigens by Rats and Rabbits: Comparison by Immunoblotting and Radioimmunoprecipitation

LEIGH RICE WASHBURN "1'2, JAMES R. RAMSAY ''2 and MARY BETH ANDREWS 2

l University of South Dakota School o[ Medicine, Department of Microbiology, Vermillion, SD 57069 (U.S.A.) 2University of Utah School of Medicine, Division of Rheumatology, 50 North Medical Drive, Salt Lake City, UT 84132 ( U.S.A.)

( Accepted for publication 28 September 1987)

ABSTRACT

Washburn, L.R., Ramsay, J.R. and Andrews, M.B., 1988. Recognition ofMycoplasma arthritidis membrane antigens by rats and rabbits: comparison by immunoblotting and radioimmunopre- cipitation. Vet. Microbiol., 17: 45-57.

Sera from rats convalescing from infection with Mycoplasma arthritidis were tested for their ability to react with M. arthritidis membrane antigens by immunoblotting and radioimmunopre- cipitation. The absence of metabolism-inhibition (MI) antibody activity in these sera suggested that rats might fail to recognize those membrane antigens involved in eliciting MI antibodies therefore rabbit antisera, which are strongly MI positive for M. arthritidis, were used for compar- ison. Antigenic recognition patterns of M. arthritidis surface and membrane antigens were not identical for rats and rabbits. The most striking and reproducible difference was the failure of rats to produce IgG antibodies against a surface antigen migrating in the 47 000-50 000 molecular weight range on SDS-polyacrylamide gels. However, rats recognized at least 2 antigens which we had previously shown to be "MI antigens", therefore the inability to express MI antibodies prob- ably cannot be explained by their inability to recognize M. arthritidis "MI antigens".

INTRODUCTION

Mycoplasma arthritidis is a natural pathogen for rats and also causes arthri- tis in rabbits under experimental conditions ( Woglom and Warren, 1938; Col- lier, 1939; Cole et al., 1977; Washburn et al., 1980a). Immunized and previously infected animals of both species are protected from challenge (Collier, 1939; Hershberger et al., 1960; Howell and Jones, 1963; Cole et al., 1969; Hannan

*Author to whom correspondence should be addressed at: University of South Dakota School of Medicine, Department of Microbiology, Vermillion, SD 57069, U.S.A.

0378-1135/88/$03.50 © 1988 Elsevier Science Publishers B.V.

46

and Hughes, 1971; Washburn et al., 1985a). In rabbits, protection is probably due to the presence of neutralizing antibodies in synovial fluids at the time of challenge (Washburn et al., 1985a). In rats, protection is also antibody me- diated (Cole et al., 1969 ); however, one of the most interesting aspects of the rat disease is the apparent inability of rats to produce or express metabolism- inhibition (MI), growth-inhibition (GI), or opsonizing antibodies against this organism (Woglom and Warren, 1938; Collier, 1939; Cole et al., 1969, 1970; Cole and Ward, 1973; Hill and Dagnall, 1975). Thus, the mechanisms of pro- tection remain unknown.

Cahill et al. (1971) suggested that the absence of MI and other neutralizing antibodies in rats could be explained by lack of recognition of certain M. ar- thritidis membrane antigens which cross-reacted with rat tissues. The purpose of this study was to identify those M. arthritidis membrane antigens recognized by convalescent rat sera, to compare them with the antigens recognized by neutralizing antibody-positive rabbit sera, and to determine whether any of the antigens which convalescent rats failed to recognize corresponded to an- tigens previously shown to be involved in the MI reaction (Washburn et al., 1985b). Such a comparison should begin to provide some insight into which antigens are important in both rat and rabbit arthritis, as well as which are involved in the neutralizing and opsonizing antibody responses. This infor- mation should have application beyond M. arthritidis-induced arthritis, since absent or impaired neutralizing antibody responses have also been reported in a number of other mycoplasma host-parasite systems, including M. pulmonis, M. neurolyticum, M. synoviae, M. gallisepticum and M. hyosynoviae infections (Taylor-Robinson and Berry, 1969; Cole et al., 1970; Levisohn and Razin, 1973; Jordan, 1975; Zimmerman and Ross, 1977).

MATERIALS AND METHODS

Mycoplasmas and mycoplasmal membrane preparation

Mycoplasma arthritidis strains 158p10, 14152p13 (Golightly-Rowland et al., 1970) and 158p10p9 (Cole et al., 1971) were grown in a soy peptone/casein/ yeast extract dialysate medium modified from Kenny (1967) and supple- mented with a dialysate of PPLO Broth (Difco Laboratories, St. Louis, MO) and with boiled gammaglobulin-free horse serum (Washburn et al., 1985a). This medium is referred to as SPCY. In addition, one group of rats was injected with strain 158p10 which had been grown in a second medium modified from Chanock et al. (1962) and referred to as modified Hayflick broth. This medium was supplemented with 10% (vol/vol) whole horse serum. Mycoplasmal mem- branes were prepared as described previously (Washburn et al., 1985b).

Rat and rabbit antisera

47

Antisera against M. arthritidis were prepared in male LEW rats weighing an average of 220 g (Harlan Sprague-Dawley, Indianapolis, IN) and in female New Zealand white rabbits obtained from a local supplier. Injection protocols for rats and rabbits, arthritis scores and final titers of the antisera used in this study are detailed in Table I. Intra-articular (i.a.) injections for the primary rabbit immunizations and intravenous (i.v.) injection for rats were used be- cause these are the most effective methods of inducing mycoplasmal arthritis in these two species ( Cole and Washburn, 1983). Prior to injection, mycoplas- mas were suspended in fresh medium from which the serum supplement had been omitted. For collection of sera, rats and rabbits were exsanguinated by cardiac puncture, rats in conjunction with CO2 euthanasia and rabbits under heavy pentobarbital anesthesia.

Prior to injection, none of the rats and rabbits used in this study showed detectable antibodies by enzyme immunoassay (FLISA) against either M. ar-

T A B L E I

Injection protocols, arthritis scores and final antibody titers for tl e animals used in antiserum production in this study

Antiserum Vs. Injection protocol Peak arthritis Antibody titer a produced in strain a score c

ELISA MI

Rats 1-5 158p10 i.v. w i th4×106c fu 14.2_+ 7.5 11.7_+0.9 <3.32 g - 1, collected at 6 weeks b

Ra t s6 -9 158p10 Asfor 1-5 15.2-+ 2.2 9 .1 ' ,.0 <3.32 Rats 10-17 158p10p9 As for 1-5 18.9_+ 4.6 10.9+0.7 <3.32 Rats 18-26 14152p13 As for 1-5 21.6_+26.8 9.8-+0.9 <3.32

Rabbit R1 158p10 i.a. with 8 × 10 s cfu, ND 8.32 7.32 collected at 8 weeks

Rabbit R2 158p10 As for R1 ND 9.32 10.32 Rabbit L2 158p10 i.a. with 1 × 109 cfu, ND 11.32 10.32

i.v. at 5 weeks with 3 × 109 cfu, collected at 6 weeks

Rabbit L3 158p10p9 As for L2 ND 11.32 11.32 Rabbit L4 14152p13 As for L2 ND 11.32 11.32

aRats 1-5 were injected with strain 158p10 grown in modified Hayflick medium; all other animals received mycoplasmas grown in SPCY. bcfu = colony-forming units. CMean + standard deviation; ND = not determined. dAntibody titers are expressed as - log2 of the end-point dilution _+ standard deviation.

48

thritidis or M. pulmonis. In addition, only one of the resulting antisera reacted by ELISA with whole horse serum (rabbit L4, with a titer of 1:20), a medium component which often contaminates mycoplasmal antigen preparations (Kenny, 1979). Moreover, it was previously shown that rat and rabbit antisera prepared as described do not react appreciably with horse serum by immuno- blotting, nor does commercial anti-horse serum antiserum immunoprecipitate 12~I-labeled M. arthritidis antigens (Washburn et al., 1985b, 1988), therefore, it is unlikely that medium antigens played a significant role in the present study.

Serologic assays

Enzyme immunoassays were performed as described (Washburn et al., 1982). Mycoplasmal antigens for ELISA were prepared from washed organisms grown in SPCY. Metabolism-inhibition assays were performed according to the method of Purcell et al. (1966), as modified by Washburn (1983). Antibody titers are expressed as - log2 of the end-point dilution.

Immunoblotting ( IB ) and radioimmunoprecipitation (RIP)

Immunoblott ing and RIP assays were performed as described (Washburn et al., 1985b). All mycoplasmas used in these assays were grown in SPCY. For IB, membranes were electrophoresed on SDS-containing 10% (w/v) poly- acrylamide gels (SDS-PAGE), blotted onto nitrocellulose and exposed first to rat or rabbit antisera diluted 1:40, then to peroxidase-conjugated anti-rat or anti-rabbit IgG antiserum (heavy and light chain specific, Cooperbiomedical, Malvern, PA) and finally to a substrate consisting of H202 and 4-chloro-1- naphthol.

For RIP, intact cells were labeled by the lactoperoxidase-glucose oxidase method of Hubbard and Cohn (1972) so that only surface antigens were la- beled. Cells were then solubilized with sodium deoxycholate and exposed to a 1:13.5 dilution of rat or rabbit antiserum. Antigen-antibody complexes were precipitated by protein A-bearing staphylococci, disrupted by incubation in an SDS-dithioerythritol-containing solubilizing solution, and electrophoresed by SDS-PAGE (Cullen and Schwartz, 1976). Labeled bands were visualized by autoradiography.

RESULTS

Recognition of M. arthritidis membrane antigens by IB and RIP

The reactivity by IB of anti-158p10 antisera from Rats 6-9 (Table I) was compared with the three rabbit anti-158p10 antisera, R1, R2 and L2, for

49

SDS-PAGE-separa ted M. arthritidis 158p10 membrane antigens (Fig. 1). ( Similar results were obtained with antisera from Rats 1-5, not shown. ) There were both quanti tat ive and qualitative differences between rats and rabbits as well as among individual animals of the same species. A total of 45 antigens appeared on this blot. The rats recognized 31-35 and the rabbits, 38-45 pep- tides. In this and replicate experiments, the most consistent difference be- tween rats and rabbits was in their response to an antigen migrating in the 47 000-50 000 molecular weight ( M W ) range, designated MAP28. Rabbit sera reacted strongly with this antigen, while rat sera reacted very weakly or not at all. As described earlier in Materials and methods, all rat and rabbit sera were

MV~

K¢

_92

_66

_45

-31

_21 -14

Fig. 1. Immunoblot ofM. arthritidis 158p10 membrane proteins electrophoretically separated on an SDS-containing 10% polyacrylamide gel and exposed to rabbit anti-158p10 antisera R1, R2 and L2 (Lanes 1-3) and rat anti-158p10 antisera nos. 6-9 (Lanes 4-7). Positions of the MW markers are shown to the left (KD=kilodaltons). These rats failed to react with an antigen, designated MAP28 (arrow), migrating in the 47 000-50 000 MW range.

50

used at a dilution of 1:40 in IB experiments so tha t the effect of different an- tibody titers could be observed. The rat sera shown in Fig. 1 had ELISA titers ranging from 8.32 to 10.32 and the rabbit from 8.32 to 11.32 (Table I) ; there- fore, antibody titer did not seem to be a factor in the inability of rats to react with MAP28.

Differences in reactivity to MAP28 could also be demonstrated by RIP. We labeled surface antigens from three M. arthritidis strains, 158p10, 158p10p9 and 14152p13, with 125I and immunoprecipitated each labeled preparation with rabbit anti-158p10 (L2) , anti-158p10p9 (L3) and anti-14152p13 (L4) sera and with pooled rat anti- 158p10 (from Rats 6-9), anti-158p10p9 (Rats 10-17) and anti- 14152p13 (Rats 18-26) sera. Immunoprecipitation of strain 14152p13 antigens by these antisera is shown in Fig. 2; results were essentially the same

Fig. 2. Autoradiograph of 12SI-labeled M. arthritidis strain 14152p13 surface antigens immuno- precipitated by pooled rat anti-158p10 (Lane 2), anti-158p10p9 (Lane 3) and anti-14152p13 (Lane 4) antisera andby rabbit anti-158p10 (Lane 5), anti-158p10p9 (Lane 6) and anti-14152p13 (Lane 7 ) antisera. Total labeled surface proteins are shown in Lane 1. MW markers are shown to the right ( KD = kilodaltons). All three pooled rat antisera failed to react with an antigen migrat- ing similarly to MAP28 (arrows).

51

with regard to antigen MAP28 when labeled antigens from the other 2 M. ar- thritidis strains were used, as well as when the rat sera comprising each of the pools were tested individually ( not shown). All three pooled rat antisera failed to react with a surface antigen migrating in the 47 000-50 000 MW range which appeared to correspond to the antigen previously designated MAP28. In addi- tion, in this experiment, rabbits reacted more strongly with a wide band mi- grating in the 20 000-24 000 MW range, although this was not always reproducible. Also shown in Fig. 2 are total 125I-labeled surface proteins from strain 14152p13. It is interesting that the MAP28 peptide, which seems to be a major antigen for rabbits, appeared as only a faint band on autoradiographs of total labeled proteins.

Recognition by rats of M. arthritidis "MI antigens"

In an earlier report, we described the characterization of those M. arthritidis antigens responsible for eliciting MI antibodies in rabbits. During that study,

/

M W x 1 0 0 0

g 3 _

8 6 _

4 5 _

31.-

21-- . 14-- .

I g

Q O

1 2 3 4 5 6 7 8 9

Fig. 3. Autoradiograph of 125I-labeled M. arthritidis membrane antigens immunoprecipitated by rat anti- 158p 10 antisera nos. 2, 4, 5, 7, 8 and 9 (Lanes 1-6 ), MI-active monoclonal antibody E11.6 (Lane 7) and rabbit anti-158p10 antisera L2 and R2 {Lanes 8 and 9). Both rats and rabbits reacted with an antigen migrating identically with the 45 000 MW "MI antigen" identified by monoclonal antibody E11.6 (large arrow}. The MAP28 antigen is identified by the small arrow and is clearly distinct from the "MI antigen".

52

Fig. 4. Immunoblot of electrophoretically separated total M. arthritidis 158~10 cellular proteins exposed to MI-active monoclonal antibody F1l.10 (Lane 1) and rat anti-158plO antisera nos. 2, 4, 5, 7, 8, and 9 (Lanes 2-7) and then to peroxidase-conjugated anti-IgG. The “MI antigen” recognized by Fl 1 .lO is identified by the arrow.

we produced 2 IgG monoclonal antibodies with MI activity. One, designated E11.6, recognized a 45 000 MW surface antigen and the other, F1l.lO, an 89 000 MW antigen present in whole cells and the cyloplasmic fraction, but not de- tectable in membrane preparations (Washburn et al., 198513).

In order to determine whether rats recognized the 45 000 MW “MI antigen”, we labeled M. arthritic& 158~10 membrane antigens with lz51 and exposed

53

them to 6 individual rat anti-158p10 antisera (nos. 2, 4, 5, 7, 8 and 9 ), 2 rabbit anti-158p10 antisera (L2 and R2) and the MI-active monoclonal antibody El l .6 (Fig. 3). The rat antiserum portion of this autoradiograph was under- exposed to avoid overexposure of the rabbit lanes, but it can be clearly seen that rats reacted strongly with a 45 000 MW antigen migrating identically with the 45 000 MW "MI antigen" identified by Ell .6. It is also clear from this figure that this "MI antigen" is distinct from MAP28.

Figure 4 shows an immunoblot comparing reactivity of the same 6 rat anti- sera and the monoclonal antibody Fll .10 against total cellular antigens from strain 158p10. Again, these rats reacted with antigens migrating identically with those recognized by the MI-active monoclonal antibody.

DISCUSSION

The purpose of this study was to compare reactivities of rat and rabbit anti- sera to M. arthritidis membrane antigens. These animals were chosen because they are both susceptible to M. arthritidis arthritis, but show considerably dif- ferent antibody responses (Woglom and Warren, 1938; Collier, 1939; Wash- burn et al., 1980b). There are a number of possible explanations for the differences between rat and rabbit antigen recognition patterns seen in this study, ranging from technical limitations inherent in each assay system to spe- cies-based differences in reactivity. With regard to the former, qualitative dif- ferences in the peroxidase-conjugated anti-rat and anti-rabbit IgG antisera used as second antibodies in the IB assay could have resulted in less intense staining of lanes treated with rat serum. However, this does not apply to RIP, in which no second antibody is used. A major limitation of the RIP technique is the use of staphylococcal Protein A to precipitate antigen-antibody com- plexes, since Protein A binds rabbit IgG much more effectively than rat ( Lan- gone, 1978; Medgyesi et al., 1978; Richman et al., 1982). This could have profoundly influenced which antigens were immunoprecipitated by rat anti- sera, although the IB assay, which was under no such restraint, gave identical results with respect to antigen MAP28. For these reasons, it was important to use both IB and RIP analysis in this comparative study in order to confirm the observations and minimize the effect of assay-related deficiencies.

Antigen recognition patterns did not differ dramatically between rats and rabbits, although in general rats recognized fewer antigens by both IB and RIP. In addition to species-related differences, there were also minor qualitative and quantitative differences among individual sera from the same species. The lat- ter did not seem to be related to arthritis score (Table I), although differences in antibody titer were probably at least partly responsible. However, with one exception, all M. arthritidis membrane and surface antigens recognized by rab- bits were also recognized by at least some of the rats. The exception was the single antigen, designated MAP28, which migrated in the 47 000-50 000 MW

54

range on SDS-polyacrylamide gels. This did not seem to be an artifact of the assay system, since the MAP28 antigen was missing from rat serum lanes on both immunoblots and autoradiographs. Nor was it due to lower IgG antibody titers in rat sera, since even the highest-titered rat antisera failed to recognize this antigen ( Table I, Figs. 1-3 ). Finally, this pattern was not confined to the 9 rats injected with strain 158p10, but also appeared in 16 additional 158p10p9- and 14152p13-injected animals (Fig. 2). Moreover, we have now shown that LEW rats injected with M. arthritidis strains PG6 and H606 also fail to rec- ognize MAP28 (L.R. Washburn, unpublished observations).

The MAP28 peptide appeared as only a faint band on autoradiographs of total labeled surface proteins from the 3 strains used in this study, compared to the intense band precipitated from these strains by rabbit antisera. Figure 2 shows only strain 14152p13, but the other 2 strains behaved similarly. Ex- planations for this include the possibility that MAP28 may be partly covered by other surface structures, or that it is a minor membrane component or de- ficient in exposed tyrosine residues and thus only poorly labeled by 12~I. More- over, if it is indeed highly antigenic for rabbits, then the immunoprecipitation process might be expected to result in its selection and concentration. The absence of IgG antibodies against MAP28 in rat sera suggests that this antigen may be involved in the neutralizing or opsonizing antibody response; however, its exact function and location on the mycoplasma surface cannot be deter- mined from the present data. We have shown that MAP28 does not correspond with either previously identified M. arthritidis "MI antigen" (Washburn et al., 1985b). However, since M. arthritidis clearly possesses at least two such an- tigens, we cannot rule out the possibility that others exist as well.

Cahill et al. (1971) suggested that rats fail to produce neutralizing antibod- ies against M. arthritidis because they cannot recognize certain membrane an- tigens. They also suggested that this might be due to cross-reactivity between M. arthritidis and certain rat tissues. This idea is supported by additional re- ports of M. arthritidis-rat tissue cross-reactivity from Kirchhoff et al. (1984) and Thirkill et al. (1981), although our own search for cross-reacting antigens has been unsuccessful (Washburn et al., 1988). In the present study, LEW rats produced IgG antibodies against both previously identified "MI antigens". This is consistent with the observation by Cole et al. (1970) that low levels of MI activity could be detected in Sprague-Dawley rats after repeated injections of M. arthritidis. In addition, Kirchhoff et al. (1983) reported low-titered MI antibodies in Sprague-Dawley rats after a single injection of M. arthritidis Strain ISR 1. These results, plus our own observations, suggest that it is un- likely that the failure of rats to express MI activity is due to an inability to recognize "MI antigens". Alternative explanations include the possibility that rat antibodies recognize different epitopes on the "MI antigens" than are in- volved in the MI reaction, or that these rat antibodies are unable to fix com- plement, a reaction thought to be important for optimal expression of MI

55

ac t iv i ty ( K e n n y , 1979; T a y l o r - R o b i n s i o n , 1983) . T e s t i n g of b o t h these hy- po the se s is c u r r e n t l y u n d e r w a y in our l abora to ry .

ACKNOWLEDGEMENTS

T h i s work was s u p p o r t e d in p a r t by g r a n t s f rom the N a t i o n a l I n s t i t u t e s of H e a l t h ( A M 2 9 8 5 8 ) , the Ar th r i t i s F o u n d a t i o n , the D e a n ' s R e s e a r c h Advisory C o m m i t t e e of the U n i v e r s i t y of U t a h School of Medic ine , the Gene ra l Re- sea rch F u n d of the U n i v e r s i t y of S o u t h . D a k o t a a n d the P a r s o n ' s E n d o w m e n t F u n d of the U n i v e r s i t y of S o u t h D a k o t a School of Medic ine .

T h e au tho r s would like to t h a n k D e b r a A s b u r y a n d Ei leen W r i g h t for excel- len t t echn ica l ass i s tance .

REFERENCES

Cahill, J.F., Cole, B.C., Wiley, B.B. and Ward, J.R., 1971. Role of biological mimicry in the path- ogenesis of rat arthritis induced by Mycoplasma arthritidis. Infect. Immun., 3: 24-35.

Chanock, R.M., Hayflick, L. and Barile, M.F., 1962. Growth on artificial medium of an agent associated with atypical pneumonia and its identification as a PPLO. Proc. Natl. Acad. Sci. U.S.A., 48: 41-49.

Cole, B.C. and Ward, J.R., 1973. Interaction of Mycoplasma arthritidis and other mycoplasmas with murine peritoneal macrophages. Infect. Immun., 7: 691-699.

Cole, B.C. and Washburn, L.R., 1983. Evaluation of arthritogenic properties of mycoplasmas for small laboratory animals. In: J.G. Tully and S. Razin { Editors), Methods in Mycoplasmology, Vol. II, Diagnostic Mycoplasmology. Academic Press, New York, pp. 305-310.

Cole, B.C., CahiU, J.F., Wiley, B.B. and Ward, J.R., 1969. Immunological responses of the rat to Mycoplasma arthritidis. J. Bacteriol., 98: 930-937.

Cole, B.C., Golightly-Rowland, L., Ward, J.R. and Wiley, B.B., 1970. Immunological response of rodents to murine mycoplasmas. Infect. Immun., 2: 419-425.

Cole, B.C., Ward, J.R., Jones, R.S. and Cahill, J.F., 1971. Chronic proliferative arthritis of mice induced by Mycoplasma arthritidis. I. Induction of disease and histopathological characteris- tics. Infect. Immun., 4: 344-355.

Cole, B.C., Griffiths, M.M., Eichwald, E.J. and Ward, J.R., 1977. New models of chronic synovitis in rabbits induced by mycoplasmas: Microbiological, histopathological, and immunological observations on rabbits injected with Mycoplasma arthritidis and Mycoplasma pulmonis. In- fect. Immun., 16: 382-396.

Collier, W.A., 1939. Infectious polyarthritis of rats. J. Pathol. Bacteriol., 48: 579-589. Cullen, S.E. and Schwartz, B.D., 1976. An improved method for isolation of H-2 and Ia alloanti-

gens with immunoprecipitation induced by protein A-bearing staphylococci. J. Immunol., 117: 136-142.

Golightly-Rowland, L., Cole, B.C., Ward, J.R. and Wiley, B.B., 1970. Effect of animal passage on arthritogenic and biological properties of Mycoplasma arthritidis. Infect. Immun., 1: 538-545.

Hannan, P.C.T. and Hughes, B.O., 1971. Reproducible polyarthritis in rats caused by Mycoplasma arthritidis. Ann. Rheum. Dis., 30: 316-321.

Hershberger, L.G., Hansen, L.M. and Calhoun, D.W., 1960. Immunization to experimental ar- thritis in rats. Arthritis Rheum., 3: 387-394.

56

Hill, A. and Dagnall, G.J.R., 1975. Experimental polyarthritis in rats produced by Mycoplasma arthritidis. J. Comp. Pathol., 85: 45-52.

Howell, E.V. and Jones, R.S., 1963. Factors influencing pathogenicity of Mycoplasma arthritidis (PPLO). Proc. Soc. Exp. Biol. Med., 112: 69-72.

Hubbard, A.L. and Cohn, Z.A., 1972. The enzymatic iodination of the red cell membrane. J. Cell. Biol., 55: 390-405.

Jordan, F.T.W., 1975. Avian mycoplasma and pathogenicity-a review. Avian Pathol., 4: 165-174. Kenny, G.E., 1967. Heat-lability and organic solvent-solubility of mycoplasma antigens. Ann.

N.Y. Acad. Sci., 143: 676-681. Kenny, G.E., 1979. Antigenic determinants. In: M.F. Barile and S. Razin {Editors), The Myco-

plasmas. Vol. I, Cell Biology. Academic Press, New York, pp. 351-384. Kirchhoff, H., Heitmann, J., Mielke, H., Dubenkropp, H. and Schmidt, R., 1983. Studies of po-

lyarthritis caused by Mycoplasma arthritidis in rats. II. Serological investigation of rats ex- perimentally infected with M. arthritidis ISR 1. Zentralbl.Bakteriol. Parasitenkd. Infektionskr. Hyg., Abb. 1 Orig. Reihe A, 254: 275-280.

Kirchhoff, H., Heitmann, J., Dubenkropp, H. and Schmidt, R., 1984. Antigenic cross-reactions between Mycoplasma arthritidis and rat tissues. Vet. Microbiol., 9: 237-248.

Langone, J.J., 1978. (~25I) protein A: A tracer for general use in immunoassay. J. Immunol. Meth- ods, 24: 269-285.

Levisohn, S. and Razin, S., 1973. Isolation, ultrastructure, and antigenicity of Mycoplasma gaUi- septicum membranes. J. Hyg., 71: 725-737.

Medgyesi, G.A., Ftist, G., Gergely, J. and Bazin, H., 1978. Classes and subclasses of rat immuno- globulins: Interaction with the complement system and with staphylococcal protein A. Im- munochemistry, 15: 125-129.

Purcell, R.H., Taylor-Robinson, D., Wong, D.C. and Chanock, R.M., 1966. A color test for the measurement of antibody to the non-acid-forming human mycoplasma species. Am. J. Epi- demiol., 84: 51-66.

Richman, D.D., Cleveland, P.H., Oxman, M.N. and Johnson, K.M., 1982. The binding of staph- ylococcal protein A by the sera of different animal species. J. Immunol., 128: 2300-2305.

Taylor-Robinson, D., 1983. Metabolism inhibition tests. In: S. Razin and J.G. Tully (Editors), Methods in Mycoplasmology, Vol. I, Diagnostic Mycoplasmology. Academic Press, New York, pp. 411-417.

Taylor-Robinson, D. and Berry, D.M., 1969. The evaluation of the metabolic-inhibition technique for the study of Mycoplasma gaUisepticum. J. Gen. Microbiol., 55: 127-137.

Thirkill, C.E., Muchmore, H.G., Hyde, R.M. and Scott, L.V., 1981. Immunological reactions of rabbit anti-Mycoplasma arthritidis serum with in vitro-cultivated rat synovial cells. In Vitro, 17: 405-411.

Washburn, L.R., 1983. Spectrophotometric method for determining titers of antimycoplasma me- tabolism inhibition antibody. Appl. Environ. Microbiol., 45: 343-346.

Washburn, L.R., Cole, B.C., Gelman, M.I. and Ward, J.R., 1980a. Chronic arthritis of rabbits induced by mycoplasmas. I. Clinical, microbiologic, and histologic features. Arthritis Rheum., 23: 825-836.

Washburn, L.R., Cole, B.C. and Ward, J.R., 1980b. Chronic arthritis of rabbits induced by my- coplasmas. II. Antibody response and the deposition of immune complexes. Arthritis Rheum., 3: 837-845.

Washburn, L.R., Cole, B.C. and Ward, J.R., 1982. Chronic arthritis of rabbits induced by myco- plasmas. III. Induction with nonviable Mycoplasma arthritidis antigens. Arthritis Rheum., 25: 937-946.

Washburn, L.R., Cathey, W.J., Cole, B.C. and Ward, J.R., 1985a. Chronic arthritis of rabbits induced by mycoplasmas. IV. Protective effect of immunization and prior infection. Clin. Exp. Rheumatol., 3: 195-204.

57

Washburn, L.R., Ramsay, J.R. and Roberts, L.K., 1985b. Characterization of the metabolism inhibition antigen of Mycoplasma arthritidis. Infect. Immun., 49: 357-364.

Washburn, L.R., Cole, B.C. and Ward, J.R., 1988. Expression of metabolism-inhibition antibodies against Mycoplasma arthritidis in rats. Am. J. Vet. Res., 49: 52-57.

Woglom, W.H. and Warren, J., 1938. A pyogenic filterable agent in the albino rat. J. Exp. Med., 68: 513-528.

Zimmermann, B.J. and Ross, R.F., 1977. Modified metabolic-inhibition test for detection of an- tibodies to Mycoplasma hyosynoviae in swine serum. Am. J. Vet. Res., 38: 2075-2076.