Ureaplasma Urealyticum adopted species

INFECTION AND IMMUNrrY, June 1977, P. 734-737 Copyright © 1977 American Society for Microbiology

Vol. 16, No. 3 Printed in U.S.A.

Immunological Analysis of Plasma Membranes of a T-Strain of Mycoplasma (Ureaplasma urealyticum) NINO ROMANO,* ROSARIA LA LICATA, AND GIULIANA SCARLATA

Istituto d'Igiene "G. D'Alessandro" dell'Universitd di Palermo, 90127 Palermo, Italy

Received for publication 29 December 1976

The cell membranes of a T-strain of mycoplasma, obtained by ultrasonic disruption, were as effective as whole organisms in eliciting metabolism-inhibit- ing and complement-fixing antibodies. The soluble fraction separated from cell membranes by centrifugation at 35,000 x g showed a minor ability to elicit an antibody response as measured by metabolism inhibition and complement fixa- tion tests. After a further centrifugation at 100,000 x g, the immunogenic activity of the soluble fraction was completely lost. Immunogenic determinants in mycoplasma membranes could also be demonstrated by adsorption tests: cell membranes were more effective than soluble fractions in adsorbing antibody capacity from the immune sera against whole cells. It has been shown by further experiments that cell membranes have at least two major antigenic determi- nants, which differ either in chemical nature or in capacity to adsorb and evoke antibodies, characterized by different serological behaviors.

Membranes of several mycoplasma species have been isolated by different procedures and examined for their antigenic properties (8).

In spite of the difficulties in separating the cell membrane from the cytoplasm, localization studies of the antigenic activities have shown that immunogens eliciting the production of metabolism-inhibiting (MI) and growth-in- hibiting as well as complement-fixing (CF) an- tibodies (8) are located in the cell membrane. Chemical analysis of mycoplasma membranes shows that, like other biological membranes, they are essentially built of protein and lipid (7). A previous report demonstrated that the CF

activity ofa T-strain of mycoplasma was associ- ated with its own lipid component (10), and since virtually all the lipid content is located in the membrane (7), it is likely that this activity is membrane bound. This possibility prompted a study ofimmuno-

logical characterization of the T-strain myco- plasma membrane. The results of our investigations are pre-

sented in this paper.

MATERIALS AND METHODS Organisms and growth conditions. The T-strain

of mycoplasma studied was P108, isolated from hu- man vagina in our laboratory. This strain can be classified as Ureaplasma urealyticum on the basis of its habitat, cultural and biochemical properties, and serological cross-reactivity with serotype VI (strain Pi, ATCC 27818) of U. urealyticum. The organisms, grown in Trypticase soy medium

(11) containing 5% (vol/vol) horse serum, were har- vested after 16 to 20 h of incubation at 3700, washed three times in 0.25 M NaCl in the cold, and resus- pended in a small volume of the same solution.

Isolation of cell membranes and soluble cell pro- teins. Soluble proteins and cell membranes of strain P108 were obtained by ultrasonic disruption of the organisms. An organism suspension in 0.25 M NaCl was treated in a Biosonik II ultrasonic disintegrator (18.6 kHz/s) at 40C for 15 1-min periods, with 1-min intervals to avoid heating. The soluble fraction was separated from the cell membranes by centrifuga- tion at 35,000 x g for 2 h at 40C. Part of the soluble fraction was subjected to an additional centrifuga- tion at 100,000 x g for 2 h at 40C. All the soluble fractions were freeze-dried and stored at -20°C. The cell membranes were washed three times in

0.05 M NaCl in 0.01 M phosphate buffer, suspended in the same buffer to a concentration of 3 to 4 mg of membrane protein per ml, and stored at -20°C until used.

Analytical methods. Protein was determined by the Folin phenol method of Lowry et al. (4) against a standard of crystalline bovine plasma. Membrane lipids were extracted with chloroform-methanol (2:1) (11). Defatted membrane proteins. The techniques of

Fleischer et al. (2), Rega et al. (9), and extraction of lipids with chloroform-methanol (2:1) were em- ployed to free membrane proteins from lipids.

Preparation of antigens. P108 organisms washed three times in 0.25 M NaCl and suspended in the same solution were used as whole-cell antigen. The preparations of antigens containing cell mem- branes, soluble cell proteins, and membrane pro- teins free from lipids were as described above. Anti- gens for immunization were prepared by dilution of the above preparations to 1 mg of protein per ml of

734

IMMUNOLOGICAL ANALYSIS OF MYCOPLASMA MEMBRANES 735

phosphate-buffered saline, pH 7.4. To prepare lipid antigen, lipid amounts equivalent to 500 ,zg were pipetted into 25-ml flasks. The chloroform-methanol (2:1) was removed in vacuo by using a rotary evapo- rator. To the thin, dry lipid film, 1 ml of deionized water containing 1 mg of human serum albumin at 37°C was added. A gentle shaking was followed by sonic treatment for 4 to 5 min in a Biosonik II ultrasonic disintegrator. The immunization sched- ule was as previously described (10).

Adsorption of antisera. Antisera were adsorbed against membranes by addition of 0.1 ml of antise- rum to the sedimented antigen (containing 1 mg of protein). Adsorption against soluble cell protein and membrane proteins free from lipid antigens was done by mixing equal volumes of antiserum and protein solution (10 mg/ml). The mixture was shaken at 37°C for 1 h, transferred at 4°C overnight, and then centrifuged at 25,000 x g for 15 min. The supernatant fluid was separated and adsorbed against another portion of the antigen, as described above.

For adsorption of antisera against membrane lip- ids, dried lipids, obtained from membranes contain- ing 3 mg of protein per ml, were emulsified in 0.3 ml of antiserum, and adsorption was carried out as described above. The adsorbed sera were kept at -20°C until used.

Serological tests. The CF test was carried out in a microtitration system, using the whole-cell suspen- sion as antigen (10). The MI test was carried out as described by Pur-

cell et al. (6).

RESULTS Immunogenicity of cell membranes and sol-

uble protein fractions. The antisera obtained by immunization of rabbits with P108 orga- nisms, cell membranes, and soluble protein fractions were tested for CF and MI with P108 organisms. Heterologous controls were also used, i.e., Mycoplasma hominis (strain PG21), grown in the medium with 1 g of L-arginine per 100 ml, and horse serum, to assay the reactivity of the T-strain antigen with horse serum or medium components.

Table 1 shows that cell membranes of strain P108 were as effective as whole cells in eliciting MI and CF antibodies. The soluble fraction showed a minor ability to elicit an antibody response, as measured by the above-mentioned tests. Furthermore, the soluble fraction sub- jected to a further centrifugation at 100,000 x g failed to elicit any significant antibody re- sponse, suggesting that the supernatant frac- tions were contaminated by minute membrane particles derived from fragmentation by ultra- sonic disruption. However, the rabbit antisera were also

slightly cross-reactive in the CF test with het- erologous controls. The reactivity was com- pletely abolished when the immune sera were previously adsorbed with horse serum; there- fore, for the subsequent experiments only im- mune sera previously adsorbed with horse se- rum were used.

Localization of immunogenic determinants in mycoplasma membranes could also be dem- onstrated by adsorption tests. CF and MI anti- bodies of serum against whole cells were com- pletely adsorbed by washed membranes (Table 2). The activity ofthe soluble fractions (1 and 2) in adsorbing the antibodies of the antiserum to whole cells could be detected only with the CF test, since the MI test could not be performed

TABLE 2. Adsorption of MI and CF antibodies by various cell fractions ofP108 mycoplasma

Reciprocal of anti- Antiserum body titer

CF MI

To P108 whole cells 160 2,048 After adsorption with: Membranes 16 32 Soluble protein fraction 1 32 _ a Soluble protein fraction 2 160 a _, MI test could not be performed due to the

urease activity present in the adsorbed sera.

TABLE 1. Serological response of rabbits to cell fractions ofP108 mycoplasma and control preparations as measured by CF and MI tests

Reciprocal of antibody titer

Soluble protein Antigen Whole cells Cell membranes

Fraction la Fraction 2b

CFe MI CF MI CF MI CF MI

P108 320 2,048 160 2,048 80 256 10 4 M. hominis 20 <4 18 <4 10 <4 10 <4 Horse serum 40 <4 30 <4 20 <4 10 <4

a Fraction separated after centrifugation at 35,000 x g for 30 mm. b Fraction separated after centrifugation at 100,000 x g for 2 h. e CF antigen was a whole-cell suspension.

VOL. 16, 1977

736 ROMANO, LA LICATA, AND SCARLATA

due to the urease activity present in the ad- sorbed serum, which caused a rapid pH shift of the medium to alkalinity.

Table 2 shows again that only soluble frac- tion 1 was able to adsorb some CF antibodies, thus confirming the hypothesis that this solu- ble fraction was contaminated with minute membrane fragments. Adsorption of antibody by membrane frac-

tions. To establish the nature of the immuno- gens, lipids extracted from membranes and sev- eral defatted membrane protein preparations were examined for MI and CF antibody-adsorb- ing capacity. Lipids extracted from membranes reduced the CF antibody titer but did not affect the MI antibody titer (Table 3). The lipid-free membrane preparations obtained after chloro- form-methanol or 10% water in acetone extrac- tion showed a low capacity to adsorb either MI or CF antibody, whereas defatted membrane protein preparations with cold n-butyl alcohol did so most effectively (Table 3). Immunogenicity of membrane fractions.

Since lipids extracted with chloroform-metha- nol from P108 membranes could adsorb only the CF antibodies, and defatted membrane prepa- rations with cold n-butanol could adsorb both MI and CF antibodies, our subsequent experi- ments were planned to determine whether an antibody response could be elicited by the injec- tion of lipids and defatted membrane proteins into rabbits. However, as the serologically ac- tive lipids can be regarded as haptens, we tried to evoke an immune response to lipids in rab- bits by the injection of a mixture consisting of lipids extracted with chloroform-methanol (2:1) with human serum albumin.

Lipid antigen and defatted membranes ob- tained by n-butanol extraction retained most of the ability of whole membranes to elicit anti- bodies in rabbits (Table 4). Moreover, only CF

TABLE 3. Adsorption of MI and CF antibodies by membrane fractions ofP108 mycoplasma

Reciprocal of anti- Antiserum body titer

CF MI

Nonadsorbed 160 2,048 Adsorbed with: 10% water-acetone-extracted 64 512 membranes

n-Butanol-extracted mem- 16 8 branes (aqueous phase)

Chloroform-methanol-ex- 64 288 tracted membranes

Membrane lipids (chloroform- 16 1,536 methanol-extracted)

TABLE 4. MI and CF antibody titer to various membrane fractions

No. Antibody titer (mean and range of Antigen of an- reciprocal)

imals MI CF

Lipid antigen 3 <5 66 (40-80) Lipid alone 2 <5 <15 Membrane 3 298 (128-512) 53 (40-80)

(n-bu- tanol-ex- tracted)

Cell mem- 2 1,536 (1,024-2,048) 160 (160) branes

antibodies could be detected after the injection of lipid-protein complex, whereas defatted membrane proteins elicited both MI and CF antibodies.

DISCUSSION Our results show that animals immunized

with membranes produce antibodies reacting with P108 organisms in CF and MI tests. A comparison of the titers induced by soluble

and membrane fractions shows that immuniz- ing antigens are located in the membrane, just as the major immunogens of other organisms tend to be located on the surface. That anti- genic determinants are located in the myco- plasma membrane can also be demonstrated by adsorption tests. Thus, the antibodies from an- tiserum to whole cells were most effectively adsorbed by membranes, but hardly at all by the soluble fraction. It should be noticed, how- ever, that the urease activity still present in the anti-P108 organism antiserum after adsorp- tion with the soluble fraction did not allow the performance of the MI test. The adsorption tests performed with different

membrane preparations have also shown that the cell membrane is the site of at least two major antigenic determinants: one, extracted by chloroform-methanol, could eliminate only the CF antibodies, and the second one, obtained in the defatted membrane, could adsorb both MI and CF antibodies. However, the defatted protein fractions varied largely in their capac- ity to adsorb MI and CF antibodies. There is no doubt that solubilizing membrane proteins without causing denaturation or conforma- tional changes is a very difficult problem. Ex- traction of the lipids with cold n-butanol was gentle enough for the membrane proteins to retain their MI and CF adsorbing capacity. Moreover, this defatted fraction retained a sig- nificant part of its immunogenic ability. The current intensive study of biological

membranes will probably sweep away the tech-

INFECT. IMMUN.

IMMUNOLOGICAL ANALYSIS OF MYCOPLASMA MEMBRANES 737

nical difficulties that now impede the charac- terization of membrane proteins in general and the serologically active one in particular. Some of the new methods based on the use of

proteolytic enzymes and labeling agents com- bined with electrophoretic analysis of mem- brane polypeptides in polyacrylamide gels, re- cently used to study protein disposition in the membrane of Acholeplasma laidlawii (1), may be used to determine the exact position of the protein antigen in or on the membrane. Our results demonstrate, too, that the lipids

are able to adsorb and evoke CF antibodies. Such results are in agreement with our pre- vious report (10), which showed the lipids' abil- ity to react in the CF test with an antiserum against the whole organism. However, since they cannot adsorb MI anti-

bodies or elicit an MI response, their disposition on the membrane surface seems to be different from that of the lipids of Mycoplasma pneumo- niae, which can adsorb and evoke both MI and CF antibodies. Although immunochemical data are availa-

ble for only some of the known mycoplasma species, the generalization seems warranted that in most species proteins are the major antigens responsible for eliciting the produc- tion of MI and CF antibodies (8). Lipids play a minor role, if any, in the antigenic activity of the mycoplasmas tested so far. The exceptions are M. pneumoniae and Mycoplasma fermen- tans (3), in which the major antigen was found in the lipid fraction.

In our Ureaplasma strain, it seems that both chemical compounds, lipids and proteins, are involved in a different way in antibody forma- tion. Such information indicates that the orga-

nisms in the order Mycoplasmatales are sur-

prisingly heterogeneous, and the chemistry of the major determinants may eventually pro- vide further criteria for classification.

LITERATURE CITED 1. Amar, A., S. Rottem, and S. Razin. 1974. Characteriza-

tion of the Mycoplasma membrane proteins. IV. Dis- position of proteins in the membrane. Biochim. Bio- phys. Acta 352:228-244.

2. Fleischer, S., B. Fleischer, and W. Stockenius. 1967. Fine structure of lipid-depleted mitochondria. J. Cell Biol. 32:193-208.

3. Kenny, G. E. 1967. Heat lability and organic solvent solubility of Mycoplasma antigens. Ann. N.Y. Acad. Sci. 143:676-81.

4. Lowry, 0. H., N. J. Rosebrough, A. L. Farr, and R. J. Randall. 1951. Protein measurement with the Folin phenol reagent. J. Biol. Chem. 103:265-275.

5. Ne'man, Z., I. Kahane, J. Kovartovsky, and S. Razin. 1972. Characterization of Mycoplasma proteins. III. Gel filtration and immunological characterization of Acholeplasma laidlawii membrane proteins. Bio- chim. Biophys. Acta 266:255-268.

6. Purcell, R. H., D. Taylor-Robinson, D. Wong, and R. M. Chanock. 1966. Color test for the measurement of antibody to T-strain mycoplasmas. J. Bacteriol. 92:6- 12.

7. Razin, S. 1973. Physiology of mycoplasmas, p. 180. In A. M. Rose and D. W. Tempest (ed.), Advances in microbial physiology, vol. 10. Academic Press Inc., London.

8. Razin, S., I. Kahane, and J. Kovartovsky. 1972. Immu- nochemistry ofmycoplasma membranes, p. 93-112. In Ciba Foundation Symposium on Pathogenic Myco- plasmas. Elsevier, Excerpta Medica North Holland, Amsterdam.

9. Rega, A. F., R. I. Weed, C. F. Reed, G. G. Berg, and A. Rothstein. 1967. Changes and properties of human erythrocyte membrane protein after solubilization by butanol extraction. Biochim. Biophys. Acta 147:297- 312.

10. Romano, N., and G. Scarlata. 1974. Serological activity of lipids of a T strain ofMycoplasma. Infect. Immun. 9:1062-1065.

11. Romano, N., P. F. Smith, and W. R. Mayberry. 1972. Lipids of a T strain of Mycoplasma. J. Bacteriol. 109:565-569.

VOL. 16, 1977