Immunization of guinea pigs with Neisseria gonorrhoeae: strain specificity and mechanisms of immunity.

Infection of subcutaneusly implanted chambers in guinea pigs conferred immunity against homologous infection of other chambers in the same animals. However, attempts to immunize guinea pigs by subcutaneous injection of filtered fluid from infected chambers, or with small doses of formalin-killed, chamber gonococci were not successful. Thus, neither organisms grown in vivo nor their extracellular products appeared to be exceptionally immunogenic. In immunizing tests with different isolates of gonococci adapted to growth in guinea-pig chambers, cross-immunity to chamber infection with low challenge doses was detected only between two of six isolates. The killing of gonococci in chambers of immunized animals, which occurred only after homologous challenge or with the heterologous strain showing cross-immunity, was not due primarily to humoral factors in the chamber fluid but probably to an enhanced effectiveness of phagocytosis. The serum of immunized animals was bactericidal for homologous strains and for the strain showing cross-immunity but not for strains showing no cross-immunity. Hence, serum bactericidal activity might be a useful indicator for investigating the specificity of immunity produced by different gonococcal strains.     

Phenotypic changes in the resistance of Neisseria gonorrhoeae to killing by normal human serum.

Gonococci adapted to growth in chambers implanted subcutaneously into guinea pigs are resistant to killing by human serum. This resistance is lost after a few generations in vitro both in culture medium and in fluid taken from guinea-pig chambers. The rate of loss is too rapid to occur by mutation and selection. Furthermore, the resistance is regained after a few generations when bacteria from the first in vitro culture are inoculated back into guinea-pig chambers in vivo. Hence the loss of serum resistance in vitro and the gain in vivo are probably due to phenotypically controlled events. Such events could be important in the pathogenicity of Neisseria gonorrhoeae.     

Lipopolysaccharide alteration is associated with induced resistance of Neisseria gonorrhoeae to killing by human serum

On SDS-PAGE, solubilized and proteinase K treated preparations of Neisseria gonorrhoeae strain BS4 (agar) showed differences in silver stained lipopolysaccharide (LPS) patterns, before and after induction to resistance to serum killing by incubation for 3 h at 37 degrees C with low Mr fractions from lysates of guinea pig red blood cells. Preparations from the original serum susceptible gonococci and LPS purified from such bacteria showed two components, but the preparations from the serum resistant gonococci were deficient in the higher Mr component. Furthermore, on immunoblotting with fresh human serum (FHS), the two LPS components of the susceptible gonococci reacted strongly with IgM. With preparations from the serum resistant gonococci there was no reaction in the area corresponding to the higher Mr component and a weaker reaction with the component of low Mr. Purified LPS from the susceptible gonococci neutralized the bactericidal activity of FHS against N. gonorrhoeae strain BS4 (agar) probably by reacting with the relevant antibody, since heated FHS was no longer bactericidal when mixed with a source of complement (human placental serum) after prior reaction with the LPS. These neutralization tests coupled with the results of immunoblotting strongly suggest that increased serum resistance is due to the lack of the high Mr LPS moiety.     

Selection from gonococci grown in vitro of a colony type with some virulence properties of organisms adapted in vivo.

Gonococci from subcutaneously implanted chambers in guinea pigs produced, on agar, more than 95% small colonies showing a "double highlight" (DH) effect in oblique reflected light combined with transmitted light. Laboratory strains of gonococci produced some DH colonies, but other showed a single highlight (SH) or no highlight (NH). Selection of DH colonies and comparison of their organisms with gonococci grown in vivo and with those from SH colonies, showed that the DH character was associated with high infectivity for guinea-pig chambers, resistance to killing by human phagocytes and heavy pilation. Furthermore, DH colonies were found in the first culture of three fresh samples of urethral pus. Thus, the DH colony characteristic may be a more reliable criterion of pathogenicity of gonococcal isolates than systems used previously. There were, however, some differences between the gonococci grown in vivo and the DH colony types. The gonococci grown in vivo and cultured once on solid medium possessed one or two antigens which differed from those of DH (or SH) colonies. They also formed smooth suspensions (which separated slowly) in saline, compared with the rough suspensions (which separated quickly) formed by gonococci from DH (or SH) colonies. Finally, the organisms grown in vivo were resistant to killing by human serum whereas the DH (and SH) colony types were susceptible; the resistance of the organisms grown in vivo was lost during one subculture on agar suggesting that the property is a phenotypic characteristic. Hence, in addition to selecting DH colony types the conditions in vivo produce organisms which differ, probably phenotypically, from cultured organisms.     

Ampicillin treatment of Neisseria gonorrhoeae in vivo. An experimental study in rabbits

The effect of ampicillin on gonococci was investigated in chambers subcutaneously implanted in rabbits. An intramuscular injection of ampicillin resulted in a rapid increase of the ampicillin concentration in serum, whereas the diffusion of ampicillin into the fluid of a non-infected chamber was comparatively slow. The ampicillin concentration was, however, maintained in the chamber fluid during a prolonged period of time as compared to ampicillin in serum. The concentration profile of ampicillin in the infected chamber was similar to that of the non-infected chamber, though at a lower level. No viable gonococci were detected 120 minutes after the injection of ampicillin. In contrast, the same concentration of ampicillin in a liquid culture resulted in slower reduction in the viability of the gonococcal strain. Even after 300 minutes a small population of gonococci was viable. Thus, a difference in the activity of ampicillin could be observed between the in vitro and the in vivo test situation.     

Virulence and immunogenicity of types 1 and 3 Neisseria gonorrhoeae in guinea pig subcutaneous chambers.

The relative virulence and immunogenicity of type 1 (T1) and type 3 (T3) cells of Neisseria gonorrhoeae were determined by tests with two different kinds of subcutaneous chambers in guinea pigs. In tests with a tissue nonencapsulated (NE) chamber, T1 gonococci were found to be greater than 1000 times more virulent as well as about 1000 times more immunogenic than T3 cells of the same gonococcal strain. However, T1 and T3 cells were found to be equally virulent for a tissue encapsulated (TE) chamber in guinea pigs. Analysis of fluids from the two types of chambers in a complement-dependent bactericidal assay revealed that the NE chamber fluid contained a substantially higher level of complement activity than fluid from TE chambers. The decline in complement level of chamber fluids due to tissue encapsulation was also confirmed by quantitation with rocket gel electrophoresis. A greater resistance of T1 cells to the bactericidal effects of complement appeared to provide a mechanism by which the T1 cells were most virulent than T3 cells for subcutaneous chambers in guinea pigs. Consequently, the NE chamber implant would appear to provide a more relevant environment for studying the virulence, as well as immunological characteristics of gonococcal strains and experimental immunogens.     

Protein changes associated with induced resistance of Neisseria gonorrhoeae to killing by human serum are relatively minor

Serum-susceptible (SS) Neisseria gonorrhoeae were induced to resistance (SR) to complement-mediated killing by fresh human serum (FHS) by a small-Mr factor(s) from guinea-pig blood in 3 h at 37 degrees C, but not in the presence of bacteriostatic concentrations of chloramphenicol or neomycin, indicating that proteins mediated the acquisition of resistance. SDS-PAGE protein profiles of lysates of equal numbers of gonococci showed only two qualitative differences between SR and SS organisms, both in minor components (a protein A of about 205 kDa in the former and not the latter and vice versa for a protein B of about 16 kDa). Many proteins, however, including the three principal outer-membrane proteins, were present in larger amounts in SR gonococci. The lack of major changes in proteins when resistance is acquired was confirmed by immunoblotting the two protein profiles with the IgG of hyper-immune rabbit anti-SR and anti-SS sera, of rabbit anti-SR serum after absorption by SS organisms and of FHS used alone and after absorption with SS organisms. The IgM of FHS, which is responsible for most of the bactericidal activity, showed only faint reactions with a few proteins common to both SS and SR gonococci and no reactions when the FHS was absorbed with SS gonococci. This is in contrast to the strong and different reactions given with lipopolysaccharide (LPS) components of SS and SR organisms, which, prepared from the former organisms, neutralize the bactericidal activity of FHS. Hence, the relatively small protein changes accompanying induction are less likely to be directly responsible for serum resistance than the more profound LPS changes.     

Human factor H interacts selectively with Neisseria gonorrhoeae and results in species-specific complement evasion

Complement forms a key arm of innate immune defenses against gonococcal infection. Sialylation of gonococcal lipo-oligosaccharide, or expression of porin 1A (Por1A) protein, enables Neisseria gonorrhoeae to bind the alternative pathway complement inhibitor, factor H (fH), and evade killing by human complement. Using recombinant fH fragment-murine Fc fusion proteins, we localized two N. gonorrhoeae Por1A-binding regions in fH: one in complement control protein domain 6, the other in complement control proteins 18-20. The latter is similar to that reported previously for sialylated Por1B gonococci. Upon incubation with human serum, Por1A and sialylated Por1B strains bound full-length human fH (HufH) and fH-related protein 1. In addition, Por1A strains bound fH-like protein 1 weakly. Only HufH, but not fH from other primates, bound directly to gonococci. Consistent with direct HufH binding, unsialylated Por1A gonococci resisted killing only by human complement, but not complement from other primates, rodents or lagomorphs; adding HufH to these heterologous sera restored serum resistance. Lipo-oligosaccharide sialylation of N. gonorrhoeae resulted in classical pathway regulation as evidenced by decreased C4 binding in human, chimpanzee, and rhesus serum but was accompanied by serum resistance only in human and chimpanzee serum. Direct-binding specificity of HufH only to gonococci that prevents serum killing is restricted to humans and may in part explain species-specific restriction of natural gonococcal infection. Our findings may help to improve animal models for gonorrhea while also having implications in the choice of complement sources to evaluate neisserial vaccine candidates.     

Natural immunity to murine gonococcal bacteremia: roles of complement, leucocytes, and sex

The roles of the serum bactericidal system, inflammatory cells, and sex in resisting gonococcal infection were studied in a murine model of gonococcal bacteremia. The role of serum killing in defense was investigated with complement component 5 deficient (C5-deficient) (B1O.D2/OSN) and normal (B1O.D2/NSN) mice. No significant differences were found between LD50's with either murine serum-sensitive or serum-resistant gonococci in those two mouse strains. However, in vitro experiments revealed a heat-stable factor in mouse serum which killed gonococci. Thus it appeared that the C5-deficient mouse is not a good model for the study of the role of C-mediated killing in resistance to gonococcal infection. Mice with Chediak-Higashi disease were used to study the role of phagocytes and natural killer cells. The difference in LD50's between affected mice (C57B1/6J beige J) and controls (C57B1/6J) was significant. The CBA/N mice, which have a B-cell maturation defect, were no more resistant to infection than control mice, which was taken as further evidence that B cells were less important than other leucocytes in innate immunity to gonococcal infection. Finally, male mice were significantly more resistant than female mice to gonococcal bacteremia. Thus, in this study the two most important determinants of resistance to gonococcal infection were inflammatory cells and sex.     

Influence of iron-limited and replete continuous culture on the physiology and virulence of Neisseria gonorrhoeae

Neisseria gonorrhoeae strains P9-2 (PenS) and KW2 (PenR) were grown in chemostats of nonferrous design at constant growth rate, pH and dissolved oxygen tension. Iron limitation (micromax 0.1 h-1) was imposed by omitting iron salts from the defined medium and titrating increasing concentrations of the non-metabolizable iron chelators ovotransferrin and Desferal, to progressively decrease the growth yield. Metabolic activity during iron limitation was very high, with a qGlc which was 2- or 11-fold greater than during cystine- or glucose-limited growth, respectively. More aspartate and isoleucine were metabolized during cystine-limited growth, while more glutamate, proline and serine were metabolized during glucose- or iron-limited growth. Significant concentrations of alanine or valine were excreted during cystine- or glucose-limited growth, respectively. Iron-limited growth of an initial inoculum of non-piliated, transparent colony-forming (P-O-) gonococci resulted in the selection of 100% piliated bacteria. Initial inocula of P+O- gonococci retained this phenotype for over 100 generations. Iron-limited gonococci were extremely virulent in the guinea-pig subcutaneous chamber model and inocula of even 12 bacteria grew rapidly and persisted. By contrast, cystine-limited (iron-replete) gonococci retained piliation but did not survive in the chambers. Transition from iron-limited to glucose-limited growth resulted in marked loss of piliation but the bacteria remained virulent. Loss of virulence did not correlate with susceptibility to killing by normal human serum, nor with changes in the content or composition of lipooligosaccharide, which contained 2.9, 3.7, 4.3 and 4.8 kDa moieties. Additional proteins were detectable in Sarkosyl-purified outer membranes of iron-limited gonococci but several proteins with molecular masses similar to those described in the literature for iron-restricted gonococci were detectable in cystine- or glucose-limited bacteria.     

Temperature Effects on Legionella pneumophila Killing by and Multiplication in Phagocytes of Guinea Pigs

We examined the effects of temperature on the interaction between Legionella pneumophila and phagocytes of guinea pigs. The body temperatures of guinea pigs infected with a sublethal dose (1.2 × 10⁴ CFU) or a lethal dose (1.0 × 10⁵ CFU) of L. pneumophila elevated from 38.4±0.15 C to 40.2±0.42 C or 40.3 ± 0.62 C, respectively. The intracellular bacterial killing by and bacterial proliferation in the phagocytes were examined at 33, 37, 40, and 42 C, using in vitro culture systems of peritoneal macrophages or polymorphonuclear leukocytes (PMN) of guinea pigs. In all the macrophages incubated at different temperatures, significant intracellular bacterial killings were observed at 4 hr after in vitro phagocytosis. After 24 hr of incubation, there was about a 100-fold increase of CFU and the number reached a maximum after 48 hr of incubation in the macrophages incubated at 42 C as well as 37 and 40 C, suggesting that macrophages support the intracellular bacterial growth in hyperthermia. In the PMN, L. pneumophila CFU 4 hr or 12 hr after the infection were significantly lower at 42 C than those at 37 C (P<0.05), indicating that the bactericidal capacity of PMN was enhanced at 42 C compared to 37 C. However, in all the PMN incubated at different temperatures, there were about 10-fold increases of CFU 24 hr after the infection, suggesting that PMN as well as macrophages support intracellular bacterial growth in hyperthermia. The extracellular bacterial growth was examined at 33, 37, 40, and 42 C in buffered yeast extract (BYE) broth or RPMI 1640 medium containing 50% guinea pig serum as a permissive or non-permissive liquid medium for the bacterial growth, respectively. Inhibition of bacterial growth in BYE broth at 42 C, and a decrease of CFU in RPMI 1640 medium containing 50% guinea pig serum at 42 C were observed. In conclusion, hyperthermia may be beneficial by restricting extracellular bacterial survival, but it exerts no beneficial effect on the restriction of intracellular bacterial growth in phagocytes, though PMN showed enhanced initial killing at 42 C. These results suggest that fever, or hyperthermia itself, may not largely contribute as a nonspecific host defense early in the course of legionellosis.     

Maintenance of Virulent Gonococci in Laboratory Culture

WE have previously shown¹ that gonococci in urethral exudates are resistant to the bactericidal action of complement plus natural or immune antibodies, yet after subculture the same strains are rapidly killed. Earlier workers² could infect volunteers more readily with gonococci from urethral pus than with the same strains grown in the laboratory. Together, these findings suggest that on subculture gonococci may lose both virulence and resistance to serum killing in parallel. Nevertheless, Kellogg et al.^{3, 4} believe that gonococci can be maintained in a virulent form on culture and that the virulent can be distinguished from avirulent forms by their different colonial morphology on their colonial type medium (CT medium). Freshly isolated gonococci had colony types referred to as 1 and 2; on random subculture different colonies known as types 3 and 4 appeared. Even after repeated subculture one strain, F62, in the type 1 colony form was able to infect volunteers, whereas the so-called avirulent type 4 colonies had lost this ability after some sixty-nine passages in vitro. Nevertheless the dose of type 1 organisms needed to set up infection was about 1.5 × 10¹⁰ microbes⁵. These results need not conflict with our¹ impression that virulence, if correlated with resistance to antibody plus complement, is lost after even one subculture. If one postulates that loss of virulence occurs in two steps type 1 organisms might be phenotypically avirulent while the type 4 colonies might have actually lost the genetic information required for virulence. The vast dose used to challenge volunteers would enable a few gonococci to survive in the host long enough to revert to the phenotypically virulent form.     

Red blood cells, a source of factors which induce Neisseria gonorrhoeae to resistance to complement-mediated killing by human serum

Lysates of guinea pig or human red blood cells (RBC) contain far more of the factors that induce resistance in gonococci to complement-mediated killing by fresh human serum that do plasma or serum. As was previously found with serum, most of the resistance-inducing activity of guinea pig RBC lysates was found in ultrafiltrates with molecular weights of less than 5000. In contrast, and as with human serum, most of the resistance-inducing activity of human RBC lysates did not pass ultrafilters which removed molecules of less than 5000 daltons, although some active material of low molecular weight was present.     

Resistance to human serum of gonococci in urethral exudates is reduced by neuraminidase

Gonococci examined directly from urethral exudates are resistant to killing by human serum, but most strains become susceptible on subculture. Previous work with gonococci grown in vitro indicates that resistance in vivo is due to sialylation of gonococcal lipopolysaccharide (LPS) by a host factor, cytidine 5'-monophospho-N-acetylneuraminic acid (CMP-NANA) or a related compound present in urogenital secretions and blood cells including phagocytes, which exude during inflammation. This sialylation inhibits the reaction between bactericidal IgM in serum and its target LPS sites. Here, we confirm the indication by using gonococci grown in vivo. Crucial to the above conclusions was the marked reduction of CMP-NANA-conferred serum resistance when gonococci were treated with neuraminidase to remove sialyl groups from their LPS. We now show that the serum resistance of gonococci in urethral exudates was reduced by treatment with neuraminidase from more than 95% (calculated in relation to controls incubated with heated serum) to 2-11% according to sample and incubation time. Subculture of the gonococci also reduced resistance to 9-11% but resistance was restored to more than 95% by incubation with CMP-NANA. This work is the culmination of an investigation that underlines the need to identify specific host factors and the virulence determinants they induce in vivo in future studies of pathogenicity.     

Morphological, biological and antigenic properties of Neisseria gonorrhoeae adapted to growth in guinea-pig subcutaneous chambers.

Gonococci (strain BS3) passaged three times and harvested directly from plastic chambers implanted subcutaneously in guinea pigs were compared with the parent strain (BS) grown in vitro. The strain grown in vivo produced smaller colonies than that grown in vitro and when examined directly in chamber fluid was sometimes not pilated. It was more resistant to the bactericidal action of human serum and more infective for guinea-pig chambers. In gel diffusion, extracts of the organisms adapted in vivo and cultured once on agar appeared to contain one or two antigens that were different from those in extracts of the in vitro grown organisms; and on polyacrylamide gels, electrophoresis of similar extracts showed one or more protein components for strain BS3 which were not seen for strain BS. Gonococci grown in guinea-pig subcutaneous chambers appear to be suitable for studies on the determinants of gonoccal pathogenicity.     

Trypanosoma (Schizotrypanum) dionisii: influence of mouse peritoneal macrophages and calf sera on extracellular growth in vitro at 37 degrees C

Macrophages and certain batches of sera were essential for extracellular multiplication of Trypanosoma dionisii in vitro in medium 199 with 20% (v/v) calf serum at 37 degrees C. In mixtures of 'good' and 'bad' batches of sera, multiplication increased as the proportion of the former was increased. Mixtures of 'bad' and 'intermediate' sera permitted virtually no growth. Phagocytosis of parasites by macrophages was unaffected by different batches of sera, though reduced in medium 199 alone. Replacement of supernatant medium by medium containing 'bad' serum reduced the macrophage infection rate no more than did transfer to medium containing 'good' serum. Daily addition of medium conditioned by prior contact with macrophages in vitro to cultures of trypanosomes without macrophages resulted in growth at least as good as that in the presence of macrophages. Extracellular replication in medium 199 at 37 degrees C apparently required at least two factors: 'M factor' provided by macrophages, but not required at 28 degrees C; and 'S+ factor' present in some batches of calf serum, essential also at 28 degrees C but not required by intracellular parasites. Some sera appeared to contain an inhibitory 'S- factor'.     

Lysis of tumor cells by antibody and complement. VI. Enhanced killing of enzyme-pretreated tumor cells.

The ascites form of a chemically induced guinea pig hepatoma, line-10, was resistant to killing in vitro by xenogeneic antibody and guinea pig complement. Pretreatment of line-10 cells with certain proteolytic enzymes rendered tham susceptible to the killing action of antibody and guinea pig complement. The effects of enzyme pretreatment were dependent on enzyme concentration, temperature, and could be blocked by addition of competitive or non-competitive inhibitors. The effect of the enzyme treatment could reversed by incubating the treated cells at 37 degrees C (but not at 0 degrees C), in the absence of the enzyme. Effective enzymes included ficin, bromelain, pronase, elastase, papain, trypsin, collagenase, lipases type I and type VI, and the neuraminidase preparation isolated from Clostridium perfringens. The activity of the lipase preparations and the neuraminidase preparation isolated from Clostridium perfringens appeared to be caused by proteolytic enzyme contamination. Enzyme preparations that proved ineffecitve in rendering the line-10 cells sensitive to killing by antibody and guinea pig complement included DNase, RNase, beta-glucuronidase type 6A or type B10, hyaluronidase type V or type VI, and pectinesterase.     

A surface polysaccharide forms when gonococci are converted to serum resistance by cytidine 5''-monophospho-N-acetyl neuraminic acid

A serum-susceptible, guinea-pig chamber-passaged, laboratory strain (BS4 (agar)) of Neisseria gonorrhoeae was converted to serum resistance by incubation with cytidine 5-monophospho-N-acetyl neuraminic acid (CMP-NANA) and examined by electron microscopy after staining with ruthenium-red. In contrast to serum susceptible gonococci incubated without CMP-NANA, the majority (60-70%) of the serum resistant organisms showed a surface accumulation of polysaccharide. This surface polysaccharide was enhanced on all the resistant gonococci after incubation with fresh human serum. Control susceptible gonococci were devoid of the polysaccharide after incubation with heated human serum. Identical results were obtained with a fresh gonococcal isolate which had lost serum resistance on subculture but which, in common with 3 other isolates, was restored to serum resistance by incubation with CMP-NANA.     

Hyperthermia,antipyretics and function of polymorphonuclear leukocytes.

Whether hyperthermia (temperature, 40 degrees C), salicylates, acetaminophen or phenacetin has an adverse effect on polymorphonuclear leukocyte (PMNL) function was examined. Migration experiemnts were carried out in Boyden chambers with bacterial chemotactic factor as the attract, and bactericidal assays were done with Staphylococcus aureus and serum from an AB blood group donor as a source of opsonins. PMNL viability was determined by the trypan blue exclusion method. Neither hyperthermia nor any of the drugs tested affected PMNL viability adversely, but sodium salicylate and phenacetin suppressed PMNL migration. Early staphylococcal killing was greater at 40 degrees C; however, after 2 hours the converse was true. Bactericidal activity was suppressed by acetylsalicylic acid, sodium salicylate and phenacetin. Hence it appears PMNL function is similar at 37 degrees and 40 degrees C but that some commonly used antipyretics have an adverse effect on PMNL activity.     

Physiology and virulence determinants of Neisseria gonorrhoeae grown in glucose-, oxygen- or cystine-limited continuous culture

Piliated Neisseria gonorrhoeae forming small, transparent colonies (P+O-) on clear typing agar have been grown in prolonged continuous culture to ascertain how different growth environments might affect gonococcal physiology and the expression of virulence determinants. Virulence of the penicillin-sensitive P9-2 and the penicillin-resistant KW1 strains was assessed by their ability to survive in polypropylene chambers implanted into the flanks of guinea pigs. Initial continuous culture experiments in the defined medium of Manchee et al. (FEMS Microbiology Letters 7, 115-118, 1980) indicated that growth was actually cystine-limited, rather than the anticipated glucose-limited. Surprisingly, cysteine was not completely metabolized and ammonium salts remained in excess. The molar growth yield on glucose (YGlc) was 65 g dry wt mol-1 and 45% of the glucose carbon metabolized was converted to biomass. Gonococci, whilst retaining the P+O- phenotype for over 100 generations of growth, did not survive in the subcutaneous chambers when inoculated at a variety of doses. When the cystine and glucose concentrations were increased and decreased respectively, growth became glucose-limited, the YGlc increased to 108 g mol-1 for strain KW1 and 75% of the metabolized glucose carbon was converted to biomass. After 17 generations of growth, however, only 2% of the gonococci retained the P+O- phenotype and P-O- bacteria predominated. Nevertheless, these bacteria were virulent in the chamber model, as was strain P9-2, which also retained only 2% of the P+O- phenotype during glucose-limited continuous culture. By contrast, the P+O- phenotype was retained during prolonged cystine- or oxygen-limited growth but only the latter was virulent. SDS-PAGE of membrane extracts confirmed that opaque colonies (O+) selected from the glucose-limited cultures contained a heat-modifiable protein (protein II) whereas transparent colony types lacked such proteins. The initial phenotype of virulent gonococci recovered from the subcutaneous chambers was P+O- but opaque variants dominated after several days. A 40 kDa outer-membrane protein was apparently induced during oxygen-limited continuous culture whereas a 44 kDa protein was absent during cystine-limited growth.