Gonococcal Lipooligosaccharide Sialylation Prevents Complement-

Transcription

1 INFECTION AND IMMUNITY, Jan. 1992, p Vol. 6, No /92/139-5$2./ Copyright 1992, American Society for Microbiology Gonococcal Lipooligosaccharide Sialylation Prevents Complement- Dependent Killing by Immune Sera LEE M. WETZLER,* KEVIN BARRY, MILAN S. BLAKE, AND EMIL C. GOTSCHLICH Laboratory of Bacterial Pathogenesis and Immunology, The Rockefeller University, New York, New York 121 Received 14 August 1991/Accepted 15 October 1991 Previous investigators have demonstrated that a sialic acid residue is added to the terminal galactose moiety of gonococcal lipooligosaccharide (LOS) when incubated with 5'-CMP-N-acetylneuraminic acid. When this in vitro sialylation occurs, gonococci become resistant to the bactericidal activity of normal human serum. This is believed to result because the added sialic acid residue blocks the binding of bactericidal anti-los antibodies present in normal human serum. We extend these studies by demonstrating that sialylated gonococci also become resistant to the bactericidal effect of immune sera containing antibodies that recognize exposed components of the outer membrane besides LOS. Prevention of antibody binding to the organism was not the cause, since the same percentage of bactericidal antibodies to the major outer membrane protein, Protein I, can be absorbed with sialylated organisms as with wild-type organisms. In addition, gonococcal sialylation prevents opsonophagocytosis by antigonococcal antisera. The negative effect of sialic acid on the complement pathway might be the reason for the findings in this study. Pathogenic bacteria need to exist in many different environments in order to spread and propagate. It is not surprising that these organisms have evolved to evade the various protective systems of the host and take advantage of the milieu in which they exist. Some organisms have evolved to use host substances to alter their surface to mask themselves from the immune system or mimic host structures (3). The gonococcus mainly exists in human mucosal areas and occasionally in the bloodstream and is able to alter itself and evade the immune response it induces. Ward et al. (24) demonstrated in 197 that freshly isolated gonococci are resistant to the bactericidal effect of normal human serum (NHS) and rabbit hyperimmune sera, and this resistance was lost on subculture. Smith and his colleagues discovered that there is a low-molecular-weight dialyzable substance present in human sera (11, 23), leukocytes (18), erythrocytes (12, 15, 16, 19), and guinea pig sera (12, 19, 22) that renders the organism resistant to the bactericidal activity of NHS. They eventually determined that this substance was 5'-CMP-N-acetylneuraminic acid (CMP-NANA). Incubation with purified CMP-NANA made gonococci resistant to NHS (12, 14, 15), increased the molecular weight of their lipooligosaccharides (LOS) (15), and did not alter the structure of any of the gonococcal outer membrane proteins (17). Smith et al. and other investigators demonstrated that incubating gonococci with CMP-NANA added a sialic acid residue to the terminal galactose of the gonococcal LOS structure Galp1-+4GlcNac (4, 8, 15). Recently, Apicella and his colleagues demonstrated that freshly isolated gonococci are sialylated in the same manner; the binding of an LOS monoclonal antibody that recognizes the Gal,B1-+4GlcNac epitope is blocked in these bacteria but binds to them when they have been treated with neuraminidase (1). This is also true for in vitro-sialylated gonococci (8). Only intracellular gonococci were examined for in vivo sialylation (1). More recently, meningococcal LOS has been demonstrated to undergo both endogenous and exogenous sialylation (7). The phenotypic change of gonococci from sensitivity to * Corresponding author. 39 resistance to the bactericidal effect of nonimmune human sera has been assumed to be the significant function of gonococcal sialylation. We show that these strains are also resistant to the bactericidal activity of sera that were raised to other specific epitopes present on the gonococcal surface. This finding appreciably changes the scope of the pathogenic significance of gonococcal LOS sialylation. MATERIALS AND METHODS Bacterial strains. Gonococcal strains (kindly provided by Charles Brinton, University of Pittsburgh, Pittsburgh, Pa.) and UU1 (kindly provided by Zell McGee, University of Utah, Salt Lake City, Utah) were used. These strains were described and used previously (25, 27). Mutant gonococcal strain APIII, lacking Rmp in its outer membranes (27), was also used. Nonpiliated transparent organisms lacking gonococcal Protein II (Opa) were used in all experiments. The absence of pili and Opa was confirmed by using a dissecting microscope to observe the nonpiliated transparent phenotype of gonococci grown on gonococcal typing agar (21). Sera. NHS samples were obtained from volunteers with no history of gonococcal infections. Rabbit antisera were raised to the gonococcal major outer membrane protein, Protein I (PI), by using PI inserted in liposomes as described previously (25). Sera recognizing most gonococcal surface components were obtained from rabbits immunized with either or UU1 emulsified with complete Freund's adjuvant and then with similar boosters in incomplete Freund's adjuvant. CMP-NANA treatment. CMP-NANA was obtained from Sigma. All organisms were treated with CMP-NANA in the same manner. Gonococci were grown for 18 h overnight on gonococcal agar plates in a 5% CO2 incubator at 37 C as described previously (25). Fresh gonococci were lifted from the plate with a Dacron swab and resuspended in GC liquid medium (1.5% proteose peptone,.9% NaCl, 1 mm phosphate buffer [ph 7.2], 1% IsoVitaLex enrichment [BBL, Baltimore, Md.]) (25) to an OD16. of.1 (approximately 18 organisms per ml) in polystyrene test tubes (16 by 125 mm). Downloaded from on October 17, 218 by guest

2 4 WETZLER ET AL. One milliliter of these organisms was added to 4 ml of the GC liquid medium. Various amounts of CMP-NANA in sterile water were added to the mixture, which was then rotated slowly for 3 h at 37 C. The OD6. of the organisms in the liquid medium after this growth phase was measured, and the organisms were then used for the bactericidal assays, opsonophagocytic assays, or absorption assays or prepared for polyacrylamide gel electrophoretic (PAGE) analysis. Gonococci were also incubated with CMP-NANA as above, but with diminishing concentrations of each, with the ratio between these two variables remaining constant. These organisms were then used in bactericidal assays. In some experiments, these CMP-NANA-treated organisms were incubated with Vibrio cholerae neuraminidase (Calbiochem, La Jolla, Calif.) to remove the sialic acid moiety. After the 3-h incubation, 1 U of neuraminidase was added to 5 ml of medium containing gonococci and CMP-NANA, and the mixture was rotated for an additional hour. Control organisms were handled in the same manner but without the addition of CMP-NANA and/or neuraminidase. The organisms were used in the bactericidal and opsonophagocytic assays without subsequent washes. LOS electrophoresis and silver staining. Organisms were suspended in sodium dodecyl sulfate (SDS)-PAGE solubilization buffer (6). Proteinase K (1 mg/ml) was added to each aliquot to degrade the majority of the protein present in the organism (5). After a 6-min incubation at 37 C, each aliquot was separated on a 15% polyacrylamide gel by a variation of the method described by Laemmli (6). After the dye front had run 8% of the length of the gel, the gel was silver stained with a commercially available kit (Bio-Rad Laboratories, Richmond, Calif.). Bactericidal assay. Organisms were grown for 3 h with or without CMP-NANA and then were diluted to an OD6. of.1. Bactericidal assays were performed with these organisms. Bactericidal reaction mixtures contained 125 RI of organisms, 25 RI of complement source, 25 RId of heatinactivated (56 C for 3 min) antiserum, and 75 pul of Hanks' buffered saline solution. The complement source was human serum absorbed with glutaraldehyde-fixed gonococci to remove antigonococcal antibodies (25, 26). The reaction mixtures were rotated for 1 h at 37 C. Aliquots were removed at the beginning and end of the incubation, and the percent survival was determined by dilution plating on gonococcal agar and compared with the percent survival in reaction mixtures which did not contain complement. Control reaction mixes included (i) organisms that had equivalent amounts of CMP-NANA added only after the 3-h incubation, just prior to their inclusion in the bactericidal assays, and (ii) organisms that were not incubated with CMP-NANA and were treated only with neuraminidase. These last two controls were included to ensure that these two substances, by themselves, would not affect the bactericidal activity of the antiserum and/or complement and that the change in the bactericidal sensitivity of the organisms was due solely to the alteration of the gonococcal surface components. No differences in the results of the bactericidal assays were observed when these controls were used instead of normal untreated organisms (data not shown). Opsonophagocytosis assay. The opsonophagocytosis assay was performed as reported previously (25). Organisms were grown for 3 h with or without CMP-NANA and then were diluted to an OD6 of.1. This mixture was then diluted 1:1 in Hanks' balanced saline solution; the final concentration of organisms was approximately 17 CFU/ml. The complement source lacked the eighth component of complement (C8D; -a 1,) mhi CMP-NANA (pg/mi) FIG. 1. CMP-NANA titration. Strain, containing PIB, was incubated with CMP-NANA at various concentrations. After the 3-h incubation, the organisms were used in routine bactericidal assays. As shown in this figure, the effect of CMP-NANA treatment and subsequent gonococcal LOS sialylation on bacterial sensitivity to NHS is titratable. The end point of bactericidal protection was seen at concentrations of CMP-NANA as low as 2,ug/ml. The methods are described in the text. Quidel, Inc., San Diego, Calif.), and therefore killing was caused by opsonophagocytosis and not the complete complement cascade. Polymorphonuclear cells (PMN) were isolated from normal human volunteers and diluted to a concentration of 16/ml. Opsonophagocytic reaction mixtures contained 1 pul of organisms, 1 pi of PMN (PMN/ bacteria ratio, 1:1), 25 pu1 of complement source, and 25,ul of heat-inactivated (56 C for 3 min) antiserum. Controls were reaction mixtures without complement. Additional controls were included as described for the bactericidal assay method. Whole-organism absorption. The whole-organism absorption experiments were performed to measure the percentage of anti-pi antibodies that could be absorbed with sialylated and control organisms by a previously described procedure (26). RESULTS INFECT. IMMUN. Gonococci were incubated with various amounts of CMP- NANA, and bactericidal assays were performed with rabbit preimmune sera. Figure 1 demonstrates that 1% survival was seen with the addition of 2 or 4 pug of CMP-NANA per ml. This is higher than the concentration of CMP-NANA in human blood (.4 pug/ml) (9) but much lower than other investigators have used for their assays (8, 14). For the remaining experiments, we used 2 jig of CMP-NANA per ml Ḋecreasing concentrations of gonococcal strain and decreasing concentrations of CMP-NANA (at the same ratio) were incubated together for 3 h: 18 organisms with 2,ug of CMP-NANA per ml, 5 x 17 organisms with 1,ug of CMP-NANA per ml, and 2 x 17 organisms with.4 pug of CMP-NANA per ml. The gonococci, after incubation with CMP-NANA, were reacted with rabbit Protein IB (PIB) antiserum in bactericidal assays as described in Materials and Methods. There was 1% protection of the organisms when incubated with diminishing amounts of CMP-NANA as long as the ratio of the number of organisms to CMP- NANA concentration remained constant. An LOS SDS-PAGE gel stained with silver was made to Downloaded from on October 17, 218 by guest

3 VOL. 6, 1992 GONOCOCCAL LIPOOLIGOSACCHARIDE SIALYLATION 41 uu1 'a D uui FIG. 2. Gonococcal LOS electrophoresis. Gonococcal strains and UU1 were incubated with CMP-NANA (2,ug/ml), resuspended in the SDS loading buffer, treated with proteinase K (1,g/ml), and separated on a 15% acrylamide gel with a 4% stacking gel by the method of Laemmli (6). Control organisms were handled in the same manner. After the gel was run, it was silver stained with a commercial kit (Bio-Rad Laboratories). There is an appreciable shift in the LOS molecular weight after treatment with CMP-NANA (lanes +), consistent with the addition of a sialic acid residue. determine whether shifts could be observed in the LOS molecular weight when strains and UU1 were treated with 2,ug of CMP-NANA per ml. As shown in Fig. 2, there is a shift in the molecular weight consistent with the addition of a sialic acid residue to the LOS. This shift has been demonstrated to occur with other gonococcal strains when incubated with CMP-NANA (8, 14, 15), but these investigators incubated the organisms with 25 to 5 times the amount used in the experiments described in this article. There are gonococcal strains that cannot be sialylated, and it is believed that these strains lack the Gal,l-+GlcNAc moiety, the LOS structure that is normally sialylated. These two strains, and UU1, appear to contain this structure, allowing sialylation of the terminal galactose moiety and the shift in the LOS molecular weight. Various antigonococcal antisera were used in bactericidal assays to discern the amount of protection that incubation with CMP-NANA gave a PIB-containing strain,, and a Protein IA (PIA)-containing strain, UU1. Anti-PIA or anti-pib antiserum could no longer kill these organisms. When the sialylated organisms were treated with neuraminidase, their sensitivity to the anti-pi antisera returned. These data are represented graphically in Fig. 3. In addition, CMP-NANA incubation of both strains rendered them resistant to antisera recognizing multiple gonococcal surface components obtained from a rabbit immunized with intact gonococci in Freund's adjuvant, NHS, and convalescentphase serum from a patient with a history of disseminated gonococcal infection (data not shown). Gonococcal mutants that lack Protein III (27) were also used in this system, and the same results were found; CMP-NANA treatment rendered them resistant to the bactericidal activity of anti-pi and anti-whole organism antisera (data not shown). The effect of gonococcal LOS sialylation on opsonophagocytosis of the organisms was also measured. When strain was incubated with CMP-NANA, rabbit anti-pi antisera could no longer opsonize and kill the organism (Fig. 4). Controls without complement did not kill the wild-type or CMP-NANA-treated organisms, but once complement was added, the wild type was killed but the sialylated strains were not. A possible explanation for the lack of a bactericidal effect of the anti-pi antisera could be that the sialylation of the LOS prevented proper binding of the PI antibodies to PI on the organisms. To examine this hypothesis, absorptions assays were performed. Various dilutions of gonococcal strain incubated with CMP-NANA were used to U a No Sera With Sera NANA NANA/Nour. FIG. 3. Bactericidal protection by CMP-NANA incubation. Strains and UU1 were incubated with CMP-NANA (2,.g/ml), and then a small aliquot was treated with neuraminidase (Neur.,.2 U/mi) to remove added sialic acid. The sera used in these experiments were NHS, anti-pi rabbit serum, and anti-whole organism rabbit serum (see Materials and Methods). All three sera gave similar results, and a typical experiment is displayed. This strain was protected from the bactericidal activity of the sera, and this protection was negated by treatment with neuraminidase, demonstrating that the addition of the sialic acid residue is related to this phenotypic change. No sera, reaction mixture with normal organisms without any antisera; with sera, reaction mixture with normal organisms and antisera; NANA, reaction mixture with antisera and organisms treated with CMP-NANA; NANA/Neur., reaction mixtures with antisera and CMP-NANA-treated organisms which were then treated with neuraminidase. absorb PI antibodies from immune rabbit sera. The amount of PI antibodies that remained was measured by enzymelinked immunosorbent assay and compared with the absorption of PI antibodies by the wild-type organism. As demonstrated in Fig. 5, sialylated organisms absorbed equal amounts of PI antibodies as did the wild-type strains, disproving this hypothesis. DISCUSSION When gonococci are incubated with CMP-NANA, the donor of sialic acid groups in human cells, their LOS is sialylated and increases in molecular weight equivalent to - a- cn cl 1 Pgh-NANA Minus C8D [: Plus C8D 3 FIG. 4. Effect of CMP-NANA incubation on opsonophagocytosis. CMP-NANA treatment of also protected this strain from the opsonophagocytic ability of NHS, anti-pi rabbit sera, and anti-whole organism rabbit serum (average survival from three experiments with anti-pi antisera is displayed)., wild type; Pgh-NANA, CMP-NANA-treated organisms; C8D, complement depleted of the eighth component. Downloaded from on October 17, 218 by guest

4 42 WETZLER ET AL. L. 1 ) L.. 2 Pgh-NANA Reciprocal Dilution of Organisms FIG. 5. PI antibody absorption. To demonstrate that the bactericidal protection afforded gonococci by CMP-NANA incubation is not due to prevention of binding of gonococcus-specific antibodies, wild-type gonococcal strain and treated with CMP-NANA (Pgh-NANA) were incubated with anti-pi rabbit antisera (n = 4). The amount of PI antibodies remaining after absorption was quantitated by enzyme-linked immunosorbent assay. Purified PI (2,ug/ml) was used as the sensitizing antigen. CMP-NANAincubated organisms were able to absorb the same percentage of PI antibodies as the control wild-type organisms. Therefore, the protection from the bactericidal action of these antisera is not due to prevention of PI antibody binding to the organism. the addition of one sialic acid residue. These sialylated gonococci are resistant to the bactericidal activity of nonimmune human sera. We have shown that gonococcal strains (PIB) and UU1 (PIA), when sialylated, are resistant not only to the bactericidal effect of rabbit or human nonimmune sera, but also to the bactericidal effect of immune sera, which recognize the whole organism or specifically the gonococcal major outer membrane protein, PI. This protection is abrogated when CMP-NANA-treated organisms are incubated with neuraminidase, which hydrolyzes sialic acid from the LOS. The exact mechanism for the protection of gonococci from the bactericidal activity of immune antisera is not known. Anti-LOS antibodies in NHS are responsible for their ability to kill serum-sensitive gonococci (2, 2). Seminal work in this field, performed by Smith and his colleagues, demonstrated that gonococcal LOS sialylation prevents killing by NHS. They demonstrated that wild-type gonococci can absorb the bactericidal anti-los antibodies present in NHS, but not following CMP-NANA treatment (14). We found that, in addition, the bactericidal effect of antibodies on antigens other than LOS is also prevented by gonococcal sialylation, and in the case of anti-pi bactericidal antibodies, we have shown that this is not due to interference with the binding of the PI-specific antibodies. CMP-NANA can be limiting, and the ratio of CMP-NANA concentration to number of organisms is important. This was demonstrated in two experiments. First, we found that there is diminishing bactericidal resistance when the same number of organisms are incubated with decreasing concentrations of CMP-NANA (Fig. 1). It appears that a certain threshold number of gonococcal LOS molecules need to be sialylated in order to observe resistance to the bactericidal effect of serum. When the concentration of CMP-NANA and the number of organisms are decreased at a constant ratio, 1% INFECT. IMMUN. resistance to the bactericidal ability of anti-pi antisera was observed. The concentration of CMP-NANA in human serum is approximately 4 ng/ml (12), but we needed to use 5 times this concentration to induce resistance, and others have used up to 5, times this concentration (1, 8). The findings in these experiments might explain why it is necessary to use superphysiologic concentrations of CMP-NANA to observe a protective effect. Relatively large numbers of organisms are used in the in vitro experiments, and the actual ratio of organisms to concentration of CMP-NANA might not be much greater than what is found in vivo, in either serum or PMN (1). Investigators have demonstrated that sialic acid present on glycolipids or bacterial capsules can inhibit the alternative complement cascade (1, 13), which is believed to be important in increasing the intensity of complement deposition even when the cascade is initiated via the classical pathway. The bactericidal and opsonophagocytic abilities of antigonococcal antibodies are dependent on the complement cascade, and both appear to be equally affected by LOS sialylation. This suggests that the negative effect of gonococcal sialylation on the bactericidal and opsonophagocytic abilities of antigonococcal antisera is caused by attenuation of the complement cascade by sialic acid. Whatever the purpose of gonococcal LOS sialylation, it appears that this change occurs in vivo and has direct phenotypic effects on the organism. Whether the prevention of killing by immune sera is the main purpose waits to be discovered, but resistance to the bactericidal and opsonic activity of immune sera occurs upon the addition of sialic acid to the terminus of gonococcal LOS. We propose that this phenomenon is not always due to the blocking of antibody binding by steric hindrance, but is most likely related to the effect of the addition of sialic acid on the complement cascade. REFERENCES 1. Apicella, M. A., R. E. Mandrell, M. Shero, M. E. Wilson, J. M. Griffiss, G. F. Brooks, C. Lammel, J. F. Breen, and P. A. Rice Modification by sialic acid of Neisseria gonorrhoeae lipooligosaccharide epitope expression in human urethral exudates: an immunoelectron microscopic analysis. J. Infect. Dis. 162: Apicella, M. A., J. Westerink, S. A. Morse, H. Schneider, P. A. Rice, and J. M. Griffiss Bactericidal antibody response of normal human serum to the lipooligosaccharide of Neisseria gonorrhoeae. J. Infect. Dis. 153: Brown, M. R. W., and P. Williams The influence of environment on envelope properties affecting survival of bacteria in infections. Annu. Rev. Microbiol. 39: Griffiss, J. M., H. Schneider, R. E. Mandrell, R. Yamasaki, G. A. Jarvis, J. J. Kim, B. W. Gibson, R. Hamadeh, and M. A. Apicella Lipooligosaccharide: the principal glycolipids of the neisserial outer membrane. Rev. Infect. Dis. 1:S287-S Hitchcock, P. J Analyses of gonococcal lipopolysaccharide in whole-cell lysates by sodium dodecyl sulfate-polyacrylamide gel electrophoresis: stable association of lipopolysaccharide with the major outer membrane protein (protein I) of Neisseria gonorrhoeae. Infect. Immun. 46: Laemmli, U. K Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature (London) 227: Mandrell, R. E., J. J. Kim, C. M. John, B. W. Gibson, J. V. Sugai, M. A. Apicella, J. M. Griffiss, and R. Yamasaki Endogenous sialylation of the lipooligosaccharides of Neisseria meningitidis. J. Bacteriol. 173: Mandrell, R. E., A. J. Lesse, A. Sugai, M. Shero, J. M. Griffiss, J. A. Cole, N. J. Parsons, H. Smith, S. A. Morse, and M. A. Apicella In vitro and in vivo modification of Neisseria Downloaded from on October 17, 218 by guest

5 VOL. 6, 1992 GONOCOCCAL LIPOOLIGOSACCHARIDE SIALYLATION 43 gonorrhoeae lipooligosaccharide epitope structure by sialylation. J. Exp. Med. 171: Mandrell, R. E., H. Schneider, M. A. Apicella, W. D. Zollinger, P. A. Rice, and J. M. Griffiss Antigenic and physical diversity of Neisseria gonorrhoeae lipooligosaccharides. Infect. Immun. 54: Markham, R. B., A. Nicholson-Weller, G. Schiffman, and D. L. Kasper The presence of sialic acid on two related bacterial polysaccharides determines the site of the primary immune response and the effect of complement depletion on the response in mice. J. Immunol. 128: Martin, P. M. V., P. V. Patel, N. J. Parsons, and H. Smith Induction of phenotypically determined resistance of Neisseria gonorrhoeae to human serum factors in human serum. J. Gen. Microbiol. 127: Nairn, C. A., J. A. Cole, P. V. Patel, N. J. Parsons, J. E. Fox, and H. Smith Cytidine 5'-monophosphate-N-acetylneuraminic acid or a related compound is a low Mr factor from human red blood cells which induces gonococcal resistance by human serum. J. Gen. Microbiol. 134: Okada, N., T. Yasuda, and H. Okada Restriction of alternative complement pathway activation by sialosylglycolipids. Nature (London) 299: Parsons, N. J., J. R. C. Andrade, P. V. Patel, J. A. Cole, and H. Smith Sialylation of lipopolysaccharide and loss of absorption of bactericidal antibody during conversion of gonococci to serum resistance by cytidine 5'-monophospho-N-acetyl neuraminic acid. Microb. Pathogen. 7: Parsons, N. J., P. V. Patel, E. L. Tan, J. R. C. Andrade, C. A. Nairn, M. Goldner, J. A. Cole, and H. Smith Cytidine 5'-monophospho-N-acetyl neuraminic acid and a low molecular weight factor from human blood cells induce lipopolysaccharide alteration in gonococci when conferring resistance to killing by human serum. Microb. Pathogen. 5: Patel, P. V., P. M. V. Martin, M. Goldner, N. J. Parsons, and H. Smith Red blood cells, a source of factors which induce Neisseria gonorrhoeae to resistance to complement-mediated killing by human serum. J. Gen. Microbiol. 13: Patel, P. V., P. M. V. Martin, E. L. Tan, C. A. Nairn, N. J. Parsons, M. Goldner, and H. Smith Protein changes associated with induced resistance of Neisseria gonorrhoeae to killing by human serum are relatively minor. J. Gen. Microbiol. 134: Patel, P. V., N. J. Parsons, J. R. C. Andrade, C. A. Nairn, E. L. Tan, M. Goldner, J. A. Cole, and H. Smith White blood cells including polymorphonuclear phagocytes contain a factor which induces gonococcal resistance to complement-mediated serum killing. FEMS Microbiol. Lett. 5: Patel, P. V., D. R. Veale, J. E. Fox, P. M. V. Martin, N. J. Parsons, and H. Smith Fractionation of guinea pig serum for an inducer of gonococcal resistance to killing by human serum: active fractions containing glucopeptides similar to those from human red blood cells. J. Gen. Microbiol. 13: Schneider, H., J. M. Griffiss, G. D. Williams, and G. A. Pier Immunologic basis of serum resistance of Neisseria gonorrhoeae. J. Gen. Microbiol. 128: Swanson, J. L Studies on gonococcus infection. XII. Colony color and opacity variants of gonococci. Infect. Immun. 19: Tan, E. L., P. V. Patel, N. J. Parsons, P. M. V. Martin, and H. Smith Lipopolysaccharide alteration is associated with induced resistance of Neisseria gonorrhoeae to killing by human serum. J. Gen. Microbiol. 132: Veale, D. R., C. W. Penn, and H. Smith Factors affecting the induction of phenotypically determined serum resistance of Neisseria gonorrhoeae grown in media containing serum or its diffusable component. J. Gen. Microbiol. 122: Ward, M. E., P. J. Watt, and A. A. Glynn Gonococci in urethral exudates possess a virulence factor lost on subculture. Nature (London) 227: Wetzler, L. M., M. S. Blake, and E. C. Gotschlich Characterization and specificity of antibodies to protein I of Neisseria gonorrhoeae produced by injection with various protein I-adjuvant preparations. J. Exp. Med. 168: Wetzler, L. M., M. S. Blake, and E. C. Gotschlich Protein I (Por) of Neisseria gonorrhoeae as an immunogen: liposomes, proteosomes and the lack of blocking antibodies, p In Transactions of the American Association of Physicians. Waverly Press, Baltimore, Md. 27. Wetzler, L. M., E. C. Gotschlich, M. S. Blake, and J. M. Koomey The construction and characterization of Neisseria gonorrhoeae lacking protein III in its outer membrane. J. Exp. Med. 169: Downloaded from on October 17, 218 by guest