Intracellular location of mycoplasmas in cultured cells ... - NCBI

Int. J. Exp. Path. (1991) 72, 705-7I4

Intracellular location of mycoplasmas in cultured cells demonstrated by immunocytochemistry and electron microscopy D. Taylor-Robinson*, H.A. Daviest, P. Sarathchandrat and P.M. Furr*

*Division of Sexually Transmitted Diseases and tSection of Electron Microscopy, Clinical Research Centre, Harrow, Middlesex, UK

Received for publication 6 June i 99 i Accepted for publication I 3 August I 99 I

Summary. Mycoplasmafermentans (strain 'incognitus') was incubated with HeLa cells for up to 96 h. After 24 h, mycoplasma organisms were demonstrated intracellularly by immunocytochemistry using mule anti-M. fermentans antiserum and gold labelling on ultrathin sections of both Lowicryl K4M and Araldite-embedded HeLa cells, the latter being treated with hydrogen peroxide. The Araldite-embedded cells were fixed with glutaraldehyde and osmium tetroxide in the presence of ruthenium red to stain the mucopolysaccharide surface components ofboth the procaryotic and eucaryotic cells. Intracellular localization of some M. fermentans organisms was confirmed by exclusion of ruthenium red from their membranes. Various numbers of mycoplasma organisms were seen per cell and occasionally some were within vacuoles, the membranes of which were also unstained by ruthenium red. The PG I 8 strain of M. fermentans and a strain of M. hominis were also detected intracellularly using similar methodology and homologous mule or rabbit antisera. The occasional presence of both apparently normal and some denser degenerate mycoplasmas in the same cell may indicate gradual degradation by phagolysosomal digestion. Keywords: Intracellular mycoplasmas, Mycoplasma fermentans, Mycoplasma hominis, HeLa cells

Mycoplasmas are the smallest free-living organisms. Unlike viruses, they do not require eucaryotic cells in which to multiply. Although taken up by phagocytic cells (polymorphonuclear leucocytes; macrophages) (Zucker-Franklin et al. I966), mycoplasmas are generally considered to remain at or attached to the surface of epithelial cells, whether in vivo or in cell cultures. Nevertheless, there are a few reports of their intracytoplasmic existence based on light microscopy

(Hayflick & Stinebring 1 9 5 5, I 960; Shepard 1958; Shedden & Cole I966; Fogh & Fogh I967) and electron microscopy (Edwards & Fogh I 9 60; Hummeler et al. I 9 6 5; Anderson & Manaker, I966; Horne 1972). Interpretation of the light microscope findings is difficult because it is not possible to distinguish between mycoplasmas that are on the surface of the cells and those that might be intracellular. The electron microscope observations also lead to conflicting interpreta-

Correspondence: Professor D. Taylor-Robinson, Division of Sexually Transmitted Diseases, Clinical Research Centre, Watford Road, Harrow, Middlesex HAI 3UJ, UK.

705

706

D. Taylor-Robinson et al. tion, some (Zucker-Franklin et al. I966) cally and the medium was replaced by i ml of believing that the mycoplasmas are within maintenance medium before the cultures neither the cell cytoplasm nor vacuoles but were inoculated with mycoplasmas. are at the bottom of crypts formed by invagination of the cell membrane. The report based on immunocyto- Mycoplasmas chemistry and electron microscopy of a The 'incognitus' strain of M. fermentans was mycoplasma, termed Mycoplasma incognitus, received from Dr S. Lo (Armed Forces Instiin large numbers in non-phagocytic cells of tute of Pathology, Washington, USA) and patients with the acquired immune defi- was passed twice in vitro in the Clinical ciency syndrome (AIDS) (Lo et al. I989) has Research Centre (CRC) laboratory before use. stimulated renewed interest in intracellular The PG i 8 prototype strain of M. fermentans localization. Since this report, the mycoplas- was obtained from Dr R.M. Chanock (Natioma has been characterized further and found nal Institutes of Health, Bethesda, USA) and to be a strain of M. fermentans (Saillard et al. passed once in the CRC laboratory. A strain I 990), a species of human genital origin that of M. hominis was received from the late Dr B. hitherto has been isolated rarely. To assess Andrews (Public Health Laboratory, Northe extent, if at all, to which this mycoplasma wich). It had been recovered from the blood becomes internalized, we have examined of a patient with post-partum fever and was HeLa cell cultures infected experimentally passed twice in vitro in the CRC laboratory with the 'incognitus' strain of M. fermentans before being used. by immunocytochemical and electron microscopic techniques; M. hominis was also used to determine whether any effect Mycoplasma medium and propagation observed was unique to M. fermentans. The medium comprised PPLO broth base supplemented with i o% (v/v) fresh yeast Materials and methods extract (25% W/V), 20% (v/v) horse serum, iooo IU penicillin/ml, o.s% thallium acetate Medium for HeLa cells and 0.002% phenol red as an indicator; Growth medium comprised minimal essen- either glucose or arginine was added at a tial medium (MEM) containing Io% foetal final concentration of o. I%, the pH value of calf serum, 3% sodium bicarbonate, I% the glucose-containing medium being glutamine, I% Hepes buffer, o.3% sodium adjusted to 7.8 and that of the argininehydroxide and I 00 IU penicillin/ml. The cells containing medium to 7.0 (Taylor-Robinson were maintained in the same medium except & Furr I 98 I). Growth of M.fermentans in the that it contained only 2% foetal calf serum. former medium was accompanied by a decrease in the pH of the medium resulting in a change in colour from red to yellow, while HeLa cells growth of M. hominis in the latter medium A I.0 ml volume of a suspension of cells at a caused a change from yellow to red. To concentration of 3 x Io4/ml in growth estimate the number of viable mycoplasmas medium was seeded onto glass coverslips of in a suspension, it was diluted serially from 13.0 mm diameter contained in flat-botio-' to io-10 in the appropriate medium tomed plastic bottles of 5 ml capacity. The and the dilutions incubated at 370C. The cell cultures were incubated at 3 70C in an highest dilution at which a colour change atmosphere of 5% CO2 in air. Confluent occurred was deemed to contain one colourmonolayers of cells were produced within changing unit (ccu) (Taylor-Robinson & 24 h. The cultures were checked microscopi- Furr I98I).

M. fermentans and M. Inoculation of cell cultures and procedure The cell monolayer cultures were inoculated with i x io7 ccu of the 'incognitus' strain of M. fermentans, I X 106 ccu of strain PG i8 of M.fermentans, and i x io7 ccu of M. hominis. These cultures and uninoculated controls were incubated at 3 70C and examined 24, 48, 72, 96, I 20, 144 and i68 h after inoculation. At these times, the number of viable mycoplasmas in an aliquot of supernatant medium was estimated as described above. The monolayers were then washed and fixed for electron microscopy as described below. Antisera An antiserum produced in mules to the PGi8 strain of M. fermentans, and received from Dr J.G. Tully (National Institutes of Health, Frederick, USA), was used to examine HeLa cells infected with the PGi8 strain or the 'incognitus' strain of M. fermentans. Two antisera to the prototype (PG2 I) strain of M. hominis were used; one was prepared in mules (received from Dr J.G. Tully) and the other in rabbits.

Fixation and embedding of cells The maintenance medium was removed and the cell cultures were washed three times with O.OI M phosphate-buffered saline 'A' containing NaCl, KCI, Na2HP04 and KH2PO4 (PBSA; pH 7.4) to remove non-specifically bound mycoplasmas. Half of the cultures were fixed in 3% paraformaldehyde in o. i M phosphate buffer, pH 7.4, for 30 min at room temperature. Cultures at 24 and 48 h were not fixed. The cells were scraped off the coverslips and processed in microcentrifuge tubes. The samples were quenched with 0. 5 M ammonium chloride in O.I M phosphate buffer for 30 min and processed using the method of progressive lowering of temperature (Roth et al. I98I). The cells were dehydrated in 30 and 50% methanol at 40C, 8o and 90% methanol at - 20°C and infil-

hominis in HeLa cells 707 trated with Lowicryl K4M resin before final embedding in Lowicryl K4M and polymerization by long-wave ultraviolet light (365 nm) in the microcentrifuge tubes at - 30°C. The other half of the cell cultures was fixed in 3% glutaraldehyde in o. I M sodium cacodylate buffer, pH 7.4, containing O.I% ruthenium red (RR) for 2 h at room temperature. The cells were post-fixed in I% osmium tetroxide in cacodylate buffer containing RR for 2 h at room temperature, scraped off the coverslips, and processed in microcentrifuge tubes. Dehydration through a graded series of acetone was followed by infiltration with Araldite resin and final polymerization at 6o0C for 24 h. Ultrathin sections of all the polymerized blocks were cut on a Reichert Ultracut E ultramicrotome and collected on carbon/ Formvar films on 200 mesh copper grids. Immunocytochemistry Ultrathin sections of cells embedded in Araldite 24, 48, 72 and 96 h after mycoplasmal inoculation of the cultures were immunolabelled, as were cells that had been embedded in Lowicryl resin after 72 and 96 h. Sections of the former were incubated with I0% hydrogen peroxide (H202) in distilled water for I0 min at room temperature (Baskin et al. 1979) and washed in distilled water before undertaking the immunocytochemical procedures. The immunolabelling method was similar for both the H202-treated Araldite sections and the Lowicryl K4M sections. All procedures were carried out with 20 u1 of each reagent. After preincubation of the sections with 5% bovine serum albumin (A30IO Sigma) in PBSA (BSA/PBSA), they were treated for i h at room temperature with antiserum that was diluted to yield the optimal labelling density. M. fermentans antiserum was used at dilutions of I/5o and i / I00 on Lowicryl sections of cells infected with strains incognitus and PG i 8 of M. fermentans, respectively, and at dilutions of I/I0 and I/30 on sections of Araldite-

7o8

D. Taylor-Robinson et al. embedded infected cells. M. hominis antiimmunolabelling experiment. Uninfected serum was used at a dilution of i / I00 on cel!s were treated with relevant mycoplassections of Lowicryl-embedded cells infected mal antiserum and normal rabbit serum was with M. hominis and at I/5 on Araldite included at a dilution equivalent to that of sections. To reduce non-specific background the antiserum. labelling with the M. hominis antiserum on Araldite sections, i% Tween 20 was added to Results the BSA/PBSA. After washing with BSA/PBSA, the Aral- Examination of ultrathin Araldite sections of dite and Lowicryl sections were treated for glutaraldehyde/osmium/RR-treated HeLa I h at room temperature with a I/io or i/ cells, 24, 48, 72 and 96 h after infection with I00 dilution, respectively, of protein A/gold M. hominis and the two strains of M. fermenin BSA/PBSA to give optimal labelling dentans, showed mycoplasmas of each species on sity without non-specific background label- the outside of the cells at 72 and 96 h. The ling. Conjugation of protein A (Sigma) with membranes of both the HeLa cells and the colloidal gold (Aurobeads GI5, RPN4 75, mycoplasmas were stained with RR; M. Amersham International) was undertaken hominis-infected cells are illustrated in Fig. i. at the CRC using the method of Roth (1989 ). All three mycoplasma strains were assoAfter further washing in BSA/PBSA and ciated closely with the HeLa cell membrane stream-washing in distilled water, the sec- and there was indentation of the membrane tions were air-dried and stained with beneath a few of the associated organisms aqueous uranyl acetate and lead citrate. (Fig. i). Approximately one in 30 cells had Control sections were included with every mycoplasmas associated with them. How-

Fig. i. M. hominis-infected HeLa cells at 24 h after infection. Ruthenium red staining of the mycoplasma membranes and the HeLa cell membrane is apparent (arrows). x 52 500.

M. fermentans and M. hominis in HeLa cells

709

Fig. 2. Lowicryl ultrathin section of a HeLa cell 72 h after infection with M. fermentans (strain PG i 8), reacted with M. fermentans antiserum and I5 nm protein A/gold. Intracellular (large arrow) and extracellular (small arrow) mycoplasmas are seen. x 30 000.

ever, it was not possible to recognize intracellular mycoplasmas with certainty. Immunocytochemical examination of sections of Lowicryl-embedded HeLa cells 72 and 96 h after infection revealed that there was medium to heavy dense gold labelling of the extracellular mycoplasmas of all three strains. Structures of mycoplasmal appearance that were associated with discrete areas of specific label were detected within some of the cells; such an area of labelling in a cell infected with M. fermentans (strain PG i 8) is shown in Fig. 2. The areas of label that were perinuclear, and thus at a distance from the cell membrane, were judged to be intracellular. All three mycoplasma strains were detected by immunocytochemical examination also of the H202-treated Araldite sections of glutaraldehyde/osmium/RR HeLa cells. The density of gold labelling was about 50% less than that seen in Lowicryl-embedded cells. The membranes of both the HeLa cells and the mycoplasmas were stained with

RR, as in Fig. i. Organisms of the three mycoplasma strains were seen outside the cell in close apposition to a flat or, occasionally, invaginated cell membrane. In addition, some of the mycoplasmas were seen within the cytoplasm. A vacuolar membrane that was not stained with RR was usually visible around the internalized mycoplasmas, one or several organisms being present in a single vacuole (Fig. 3). M. fermentans (strain PG I 8) was detected both extracellularly and intracellularly 24, 48, 72 and 96 h after infection of the cells; small numbers only were detected in a few cells at 24 h. M. fermentans (strain incognitus) was detected at 48, 72 and 96 h and M. hominis at 72 and 96 h after infection. The largest number of organisms of all three mycoplasma strains was seen 72 h after infection. The labelling density for the two strains of M. fermentans was about I5 gold particles per mycoplasma organism (Fig. 3a, b), that for M. hominis being less at five to i o gold particles per organism (Fig. 3c).

7IO

D. Taylor-Robinson et al.

M. fermentans and M. hominis in HeLa cells

71II

Fig. 4. M. fermentans-infected HeLa cells 48 h after infection, reacted with M. fermentans antiserum and I5 nm protein A/gold. The vacuolar membrane without ruthenium red staining is visible (arrows). x 35 000.

Occasionally, groups of very dense mycoplasmas of any of the three strains were seen in vacuoles (Fig. 4) and both dense and healthy-looking mycoplasmas were present in some cells (Fig. 5). Discussion

Specific attachment of mycoplasmas to eucaryotic cells is a well known phenomenon (Razin I985). For example, M. pneumoniae possesses a specific attachment protein (Hu et al. I977) and adhesion is mediated through a sialic acid receptor on the host cell

(Loomes et al. I984). Attachment of M. hominis to HeLa and other cells has been demonstrated but the nature of adhesin and receptor is defined less clearly (Manchee & Taylor-Robinson I969). What happens to mycoplasmas following attachment has been in doubt, and the observations of Lo et al. (I989) stimulated our interest in determining whether these organisms invade epithelial cells. None of the previous observations based on light microscopy allows such a conclusion to be drawn, the only possible chance of success being with the use of electron microscopy. However, even in the

Fig. 3. Immunolabelling of hydrogen peroxide-treated Araldite ultrathin sections of HeLa cells. Ruthenium red staining is visible on the HeLa cell membrane (small arrows) but not on the vacuolar membrane (large arrows). a, 24 h after infection with M. fermentans (strain PGi 8); cells treated with M.

fermentans antiserum and 15 nm protein A/gold. b, 72 h after infection with M. fermentans (strain incognitus); cells treated with M. fermentans antiserum and 1 5 nm protein A/gold. c, 72 h after infection

with M. hominis; cells treated with M. hominis antiserum and I 5 nm protein A/gold, in the presence of Tween 20. x 40 000.

712

D. Taylor-Robinson et al.

Fig.

A

:f

firs' i

-.e^, }

Fig. 5. M. fermentans (strain incognitus) infected HeLa cells 48 h after infection, reacted with M.

fermentans antiserum and I 5 nm protein A/gold. Both healthy looking (small arrows) and degenerate mycoplasmas (large arrows) are present in intracellular vacuoles. x 35 000.

present study, although specific labelling was shown in Lowicryl-embedded cells, internalization of mycoplasmas was not possible to demonstrate with certainty. Only where the mycoplasma was detected in the perinuclear region away from the cell membrane was there a probability of it being truly intracellular. Otherwise, invagination of the cell membrane which was not present in the plane of the section may have produced the spurious appearance of an internal mycoplasma. This, of course, was a criticism of previous electron microscope studies. However, in the current investigation we also used ruthenium red staining and a specific gold label to seek evidence of internalization. Labelling the coat of HeLa cells with ruthenium red enables differentiation to be made between the internal and external membranes. Ruthenium red reacts with anionic sites on the cell coats of both procaryotic and

eucaryotic cells but does not penetrate the cells (Luft 197ia, b; Erdos I986). Material that had been embedded for morphological examination was used. This required the unmasking of antigens from osmium tetroxide/RR fixed and epoxy resin-embedded cells with hydrogen peroxide (Baskin et al. I 9 79). Opinions differ about the mode of action of hydrogen peroxide in this context; some consider that it 'etches' the resin (Bendayan & Zollinger I983), while others regard it as removing the powerful cross-linking fixative osmium tetroxide (Baskin et al. 1 9 79). In this study, the exposure time of resin sections to hydrogen peroxide was assessed to maintain some osmium tetroxide/RR product in the sections but unmask sufficient antigen to produce a good density of labelling. Exclusion of ruthenium red from both the mycoplasma membrane and the HeLa vacuolar membrane, in addition to localization of

M. fermentans and M. specific gold label, was clear evidence for internalization. This occurred not only with the incognitus strain of M. fermentans, consistent with the reported intracellular occurrence of this mycoplasma in vivo, but also with the prototype strain and a strain of M. hominis. The HeLa cells are functioning probably as non-professional phagocytes. It has been shown (Hale & Bonventre 1979; Hale et al. I 9 79) that phagocytosis of Shigella flexneri by Henle embryonic intestinal cells in vitro requires active eucaryotic cells and viable procaryotic cells. We have not determined whether viable mycoplasmas are necessary, nor whether there is mycoplasmal multiplication within the cell. The large numbers of intracellular organisms observed by Lo et al. (I989) suggest that this is probable but the relatively small numbers of organisms that we were able to demonstrate make it unlikely in our study. The question, therefore, is what happens to intracellular mycoplasmas? We noted that some of the intracellular mycoplasmas were dense and appeared degenerate and may have been undergoing digestion within a phagolysosome, whereas others in the same or a separate cell appeared healthy. This may be a function of the non-synchronous infection of the cells and reflect a progression of uptake and digestion where the early engulfed mycoplasmas show degeneration, or mean that some mycoplasmas are able to evade phagolysosome digestion. Moulder (i 98 5) reviewed intracellular parasitism of eucaryotic cells by procaryotes and considered that three methods of escaping phagolysosomal digestion are available to procaryotes. First, escape from the phagosome after uptake into the cell; second, resistance to lysosomal enzymes; and third, prevention of phagosome-lysosome fusion. In Mycobacterium avium infection of mouse bone macrophages (Frehel et al. I 986), some of the bacteria are taken into the phagolysosomes and digested while others are not digested and thus exhibit resistance to lysosomal enzymes. Prevention of phagosome-lysosome fusion is achieved in different ways. For example,

hominis in HeLa cells

713

Chlamydia psittaci (Friis 1972) modifies the lysosomal membrane, and bacterial products of Mycobacterium tuberculosis (Goren et al. 1976) interact with the lysosomal membranes which renders them non-fusible. In the current study, escape from the phagosome is unlikely as mycoplasmas were observed in phagosomes or phagolysosomes late in the infection process. Thus, the other methods of escape are more likely. The intracellular localization of mycoplasmas, if only for a short period, may help to protect them against the effects of antibody and antibiotic and, to some extent, account for the difficulty of eradicating them from cell cultures.

References ANDERSON D.R. & MANAKER R.A. (i966) Electron microscopic studies of Mycoplasma (PPLO strain 88o) in artificial medium and in tissue culture. J. Nat. Cancer Inst. 36, 1 39-1 54. BASKIN D.G., ERLANDSEN S.L. & PARSONS J.A. (I 9 79) Influence of hydrogen peroxide or alcoholic sodium hydroxide on the immunocytochemical detection of growth hormone and prolactin after osmium fixation. J. Histochem. Cytochem. 27, 1290-I292. BENDAYAN M. & ZOLLINGER M. (I983) Ultrastructural localisation of antigenic sites on osmiumfixed tissues applying the protein A-gold technique. J. Histochem. Cytochem. 31, 101-IO9. EDWARDS G.A. & FOGH J. (I960) Fine structure of pleuropneumonia-like organisms in pure culture and in infected tissue culture cells. J. Bacteriol. 79, 267-276. ERDos G.W. (I986) Localisation of carbohydratecontaining molecules. In Ultrastructure Techniques for Microorganisms. Eds H.C. Aldrich & W.J. Todd. Plenum Press. pp. 399-419. FOGH J. & FOGH H. (I967) Morphological and quantitative aspects of mycoplasma-human

cell relationships. Proc. Soc. Exp. Biol. Med. 125, 423-430. FREHEL C., DE CHASTELLIER C., LANG T. & RASTOL N.

(I986) Evidence for inhibition of fusion of lysosomal and prelysosomal compartments with phagosomes in macrophages infected with pathogenic Mycobacterium avium. Infect. Immun. 52, 252-262. FRIIS R.R. (I972) Interaction of L cells and Chlamydia psittaci: entry of the parasite and host

714

D. Taylor-Robinson et al.

responses to its development. J. Bacteriol. i8o,

706-72I. GOREN M.B., HART P.D., YOUNG M.R. & ARMSTRONG J.A. (1I 9 76) Prevention of phagosome-lysosome fusion in cultured macrophages by sulfatides of Mycobacterium tuberculosis. Proc. Nat] Acad. Sci. USA 73, 2510-2514. HALE T.L. & BONVENTRE P.F. (I979) Shigella infection of Henle intestinal epithelial cells: role of the bacterium. Infect. Immun. 24, 879-886. HALE T.L., MORRIS R.E. & BONVENTRE P.F. (1979) Shigella infection of Henle intestinal epithelial cells: role of the host cell. Infect. Immun. 24, 887-894. HAYFLICK L. & STINEBRING W.R. ( 1 9 5 5) Intracellular growth of pleuropneumonia-like organisms. Anat. Rec. 121, 477-478. HAYFLICK L. & STINEBRING W.R. (i 960) Intracellular growth of pleuropneumonia-like organisms (PPLO) in tissue culture and in ovo. Ann. N.Y. Acad. Sci. 79, 433-449. HORNE R.W. (1972) Comparison between the structure of animal and plant mycoplasmas: extracellular and intracellular morphology. In Pathogenic Mycoplasmas. A Ciba Foundation Symposium, Excerpta Medica, North Holland. PP. 39-57. Hu P.C., COLLIER A.M. & BASEMAN J.B. (I977) Surface parasitism by Mycoplasma pneumoniae of respiratory epithelium. J. Exp. Med. 145, 1328-I343. HtIMMELER K., ARMSTRONG D. & ToMASSINI N. (I965) Cytopathogenic mycoplasmas associated with two human tumors. II Morphological aspects. J. Bacteriol. 90, 5 II-5 I 6. Lo S.-C., DAWSON M.S., WONG D.M., NEWTON P.B. III, SONODA M.A., ENGLER W.F., WANG R.Y.-H., SHIH J.W.-K., ALTER H.J. & WEAR D.J. (I989) Identification of Mycoplasma incognitus infection in patients with AIDS: an immunohistochemical, in situ hybridization and ultrastructural study. Am. J. Trop. Med. Hyg. 41, 6oi6i6. LooMEs L.M., UEMURA K., CHILDS R.A., PAULSON J.C., ROGERS G.N., SCUDDER P.R., MICHALSKI J-C., HOIJNSELL E.F., TAYLOR-ROBINSON D. & FEIZI T. (I984) Erythrocyte receptors for Mycoplasma pneumoniae are sialylated oligosaccharides of Ii antigen type. Nature 307, 560-563. LIJFT J.H. ( 197ia) Ruthenium red and violet. I

Chemistry, purification, methods of use for electron microscopy and mechanism of action. Anat. Rec. 171, 347-368. LUFT J.H. (197Ib) Ruthenium red and violet. II Fine structural localization in animal tissues. Anat. Rec. 171, 369-4I6. MANCHEE R.J. & TAYLOR-ROBINSON D. (I[969) Studies on the nature of receptors involved in attachment of tissue culture cells to mycoplasmas. Br. 1. Exp. Path. 50, 66-75. MOULDER J.W. (I985) Comparative biology of intracellular parasitism. Microbiol. Rev. 49, 298-337. RAZIN S. (I985) Mycoplasma adherence. In The Mycoplasmas. Volume 4. Eds S. Razin & M.F. Barile. New York: Academic Press. pp. i6i202. ROTH J. (I989) The colloidal gold marker system for light and electron microscopic cytochemistry. In Techniques in Immunocytochemistry. Volume 2. Eds G.R. Bullock & P.Petrusz. London: Academic Press. pp. 217-284. ROTH J., BENDAYAN M., CARLEMALM E., VILLIGER W. & GARAVITO M. (I98I) Enhancement of structural preservation and immunocytochemical staining in low temperature embedded pancreatic tissue. I. Histochem. Cytochem. 29,

663-67I. SAILLARD C., CARLE P., Bovt J.M., BtBtAR C., Lo S.-C., SHIH J.W.-K., WANG R.Y.-H., ROSE D.L. & TULLY J.G. (1990) Genetic and serologic relatedness between Mycoplasma fermentans strains and a mycoplasma recently identified in tissues of AIDS and non-AIDS patients. Res. Virol. 14I, 385-395. SHEDDEN W.I.H. & COLE B.C. (I966) Rapid method for demonstrating intracellular pleuropneumonia-like organism in a strain of hamster kidney cell (BHK2 IC1'3). Nature 2IO, 868. SHEPARD M.C. (I958) Growth and development of T strain pleuropneumonia-like organisms in human epidermoid carcinoma cells (HeLa). 1. Bacteriol. 71, 35 1-355. TAYLOR-ROBINSON D. & FURR P.M. (198I) Recovery and identification of human genital tract mycoplasmas. Isr. J. Med. Sci. I7, 648-653. ZIJcKER-FRANKLIN D., DAVIDSON M. & THOMAS L. (I966) The interaction of mycoplasmas with mammalian cells. I HeLa cells, neutrophils, and eosinophils. J. Exp. Med. 124, 521-532.