Immunization against a Saccharide Epitope Accelerates Clearance of Experimental Gonococcal Infection

The emergence of ceftriaxone-resistant strains of Neisseria gonorrhoeae may herald an era of untreatable gonorrhea. Vaccines against this infection are urgently needed. The 2C7 epitope is a conserved oligosaccharide (OS) structure, a part of lipooligosaccharide (LOS) on N gonorrhoeae. The epitope is expressed by 94% of gonococci that reside in the human genital tract (in vivo) and by 95% of first passaged isolates. Absence of the 2C7 epitope shortens the time of gonococcal carriage in a mouse model of genital infection. To circumvent the limitations of saccharide immunogens in producing long lived immune responses, previously we developed a peptide mimic (called PEP1) as an immunologic surrogate of the 2C7-OS epitope and reconfigured it into a multi-antigenic peptide, (MAP1). To test vaccine efficacy of MAP1, female BALB/c mice were passively immunized with a complement-dependent bactericidal monoclonal antibody specific for the 2C7 epitope or were actively immunized with MAP1. Mice immunized with MAP1 developed a TH1-biased anti-LOS IgG antibody response that was also bactericidal. Length of carriage was shortened in immune mice; clearance occurred in 4 days in mice passively administered 2C7 antibody vs. 6 days in mice administered control IgG3λ mAb in one experiment (p = 0.03) and 6 vs. 9 days in a replicate experiment (p = 0.008). Mice vaccinated with MAP1 cleared infection in 5 days vs. 9 days in mice immunized with control peptide (p = 0.0001 and p = 0.0002, respectively in two replicate experiments). Bacterial burden was lower over the course of infection in passively immunized vs. control mice in both experiments (p = 0.008 and p = 0.0005); burdens were also lower in MAP1 immunized mice vs. controls (p<0.0001) and were inversely related to vaccine antibodies induced in the vagina (p = 0.043). The OS epitope defined by mAb 2C7 may represent an effective vaccine target against gonorrhea, which is rapidly becoming incurable with currently available antibiotics.


Introduction
Neisseria gonorrhoeae infection is the second most common bacterial sexually transmitted infection (STI); the worldwide incidence is 106 million cases per year [1]. Gonococci cause a broad spectrum of diseases [2]; HIV co-infection in men enhances risk of HIV transmission to female sex-partners [3]. Recent, widespread emergence of resistance to currently used antimicrobials [4] and the potential for spread of resistant gonococci threaten to herald an era of untreatable disease, worldwide. Uniform vaccination of persons at greatest risk would be an effective deterrent.
Development of safe effective vaccines against gonococcal infection is challenging because the correlates of immune protection are not fully known [5]. Furthermore, gonococcal surface molecules that may be appropriate targets often are antigenically variable. Unfortunately, adaptive immune responses that target highly conserved gonococcal antigens fail to elicit protection [6].
N. gonorrhoeae lipooligosaccharide (LOS) is an important component of the gonococcal outer membrane [7]. Antibodies directed against LOS engage complement to kill N. gonorrhoeae directly [8] and also promote opsonophagocytosis [9]. LOS antibodies may also contribute to protection against re-infection with the homologous strain in experimental infection of human male volunteers [10].
Despite antigenic heterogeneity of LOS, we have identified a common oligosaccharide structure within gonococcal LOS that is recognized by a murine monoclonal antibody (mAb), called 2C7 [9,11]. This structure (Figure 1) requires the substitution of lactose onto HepII and at a minimum, substitution of lactose on HepI [12]. The 2C7 epitope was identified directly in the genital secretions of 94% of 68 culture-positive subjects and on 95% of 101 strains of N. gonorrhoeae isolated from infected subjects [9].
Human antibodies against the 2C7 epitope also mediate complement-dependent bacterial killing and opsonophagocytosis. Compared to purified LOS, the 2C7 epitope selectively elicited a greater antibody response after gonococcal endometritis and disseminated infection [9]. Male volunteers immunized with a gonococcal outer membrane vaccine that contained LOS harboring the 2C7 epitope developed a 10-fold excess of 2C7 antibody compared to a rise in antibody against whole LOS [9], thereby confirming superior immunogenicity of the 2C7 epitope in a human vaccine trial.
Carbohydrates are poor immunogens and induce T cellindependent immune responses that may not lead to full affinity maturation and are poor immunogens [13]. A promising approach in vaccine design uses peptides that are the structural and/or functional mimics of carbohydrate antigens [14,15,16]. Peptide mimics elicit cross-reactive immune responses to the nominal carbohydrates [17]; when used for immunization they can elicit an immune response against carbohydrate antigens and lead to effective immunity [17].
Previously, we selected a peptide mimic of the 2C7 epitope by screening a random peptide library with mAb 2C7 [11]. We reconfigured the peptide into a multi-antigenic form, called MAP1 [11]. Immunization of mice elicited cross-reactive anti-LOS antibodies that possessed dose-responsive direct complementdependent bactericidal activity against gonococci [11]. Here, we further characterized the MAP1 induced antibody responses in mice and determined the efficacy of both passive immunization with mAb 2C7 and active vaccination with MAP1 in attenuating infection in mice following experimental vaginal challenge with N. gonorrhoeae.

Results
Characterization of LOS structure of gonococcal strains: FA1090wt and FA1090lgtG 2 Figure 1 illustrates gonococcal LOS structure identifying the innermost oligosaccharide structures that bind mAb 2C7 (shaded gray) and the phase-variable LOS glycosyl transferases (lgt) biosynthetic genes. Loss of lactose (Gal-Glc) substitution on HepII in the FA1090lgtG 2 (lgtG 2 ) mutant was confirmed by compositional analysis and by loss of reactivity with mAb 2C7 in western blot and whole-cell ELISA (data not shown). The ''wild-type'' (wt) and lgtG 2 mutant expressed similar HepI LOS substituted glycan extensions, revealed by whole-cell ELISA and western blot using 3 mAbs that recognize distinct HepI glycan extensions ( Figure 1): 3F11 (lacto-N-neotetraose), L8 (lactose) and L1 (GalRGalRGlc, or the P K -like structure [18,19,20,21] (data not shown). Furthermore, DNA sequencing indicated that there was no variation in the homopolymeric regions of lgtA, lgtC and lgtD ( Figure 1) between the wt and lgtG 2 mutant suggesting that expression of HepI glycan extensions in the wt and mutant were similar (data not shown).

Loss of lgtG expression attenuates N. gonorrhoeae FA1090 infection
The 2C7 epitope is expressed on most gonococci that infect humans [9] and we hypothesized that the phase-variable lgtG gene genes that are involved in LOS biosynthesis and are subject to phase variation are indicated in grey italic font. Each transferase catalyzes the substitution of the next more distal hexose onto the LOS backbone. The grey shaded box represents the minimum OS structure expressed by naturally occurring gonococcal strains that is required for mAb 2C7 binding. doi:10.1371/journal.ppat.1003559.g001

Author Summary
Neisseria gonorrhoeae cause a common sexually transmitted infection (ca. 106 million cases per year, worldwide). Gonococcal organisms have become resistant to almost all previously effective antibiotics, creating an urgent need for preventive vaccines. Extreme variability of the gonococcal surface has precluded identification of common components that might be present on a wide variety of isolates/ strains; these might serve as effective vaccine targets. We identified a common sugar structure, present on 95% of gonococcal organisms; this structure elicits a large and specific antibody response in women after natural gonococcal infection. Because peptides are more effective than sugars when used as vaccines, we identified peptide ''lookalikes'' that mimicked the sugar; first by using a monoclonal antibody that recognized the sugar, then by screening trillions of small peptides to identify those also recognized by the antibody. Mice that were vaccinated with one of these peptides cleared infection more quickly than animals vaccinated with an irrelevant peptide. We also found that administration of antibody from immune to normal, unimmunized, mice, prior to infection, subsequently hastened clearance of infection, indicating that the antibody administered was the protective agent. This study represents an important step in developing a vaccine to protect humans from infection caused by a wide variety of gonococcal strains.
remains in-frame (on) because it provides the organism a survival advantage in its natural (human) host. Growth and viability of FA1090wt and FA1090lgtG 2 were similar in vitro as revealed by similar growth curves (doubling time and final optical density) when the strains were grown separately or together in co-culture (data not shown). To establish the role of lgtG in vivo, mice were inoculated with a mixture of strains FA1090wt (1.5610 5 colony forming units [CFU]) and FA1090lgtG 2 (1.6610 5 CFU). Five of 5 inoculated mice became infected. Kaplan Meier analysis of time to median clearance of FA1090lgtG 2 was 4 days; only one mouse cleared FA1090wt infection by 10 days, the length of the experiment (P = 0.04), ( Figure S1A). A lower longitudinal trend in mean log 10 CFU colonization over time was measured for the lgtG 2 mutant ( Figure S1B) (P,0.001). Cumulative CFU expressed as area under the curve (AUC) also showed a lower bacterial burden for the lgtG 2 mutant strain (P = 0.04), ( Figure S1C). Based on the in vivo co-culture results, we hypothesized that although selective (antibody) pressure targeting the 2C7 epitope may serve to select for phase variation events that yield a 2C7 epitope negative gonococcal infection, the reduced fitness associated with loss of the 2C7 epitope would not favor continued infection.

mAb 2C7 accelerates clearance of HepII lactose bearing gonococci
We examined whether passive administration of mAb 2C7 also shortens time of FA1090wt infection in mice. Mice were administered mAb 2C7 or control IgG3l mAb intraperitoneally (ip) daily for 3 days prior to and including the day of challenge. In Group 1 (Figure 2), using a vaginal inoculum of 5610 5 CFU, 16 of 20 (80%) mice administered mAb 2C7 and 16 of 20 mice administered control IgG3l mAb became infected. In Group 2 (Figure 2), using a 3.6610 5 CFU inoculum; 18 of 20 (90%) mice administered mAb 2C7 and 18 of 20 mice given control IgG3l mAb became infected.
Kaplan Meier analysis showed that mAb 2C7-treated animals cleared gonococci faster than IgG3l mAb-treated control animals in both experiments (Experiment 1: median time to clearance; 4 days in mAb 2C7 administered vs. 6 days in the IgG3l mAb administered group, P = 0.03 [ Figure 2A, left panel]; Experiment 2: median time to clearance; 6 days in mAb 2C7 administered vs. 9 days in the IgG3l mAb administered group, P = 0.008 [ Figure 2A, right panel]). A diminished longitudinal trend in mean log 10 CFU colonization over time was measured in mAb 2C7 administered mice (P = 0.008 and P = 0.0005 in Experiments 1 and 2, respectively [ Figure 2B]). Cumulative CFU expressed as area under the curve (AUC) was also lower in mAb 2C7 administered mice (P = 0.014 and P = 0.0114 in Experiments 1 and 2, respectively [ Figure 2C]). All gonococcal colonies isolated from the 2 mice that remained infected at 7 days in Experiment 2 after inoculation (75 and 40 colonies per mouse) were detected by mAb 2C7 in colony blots further emphasizing that the 2C7 structure is important for prolonged survival of FA1090 in the mouse model.
Clearance of the FA1090lgtG 2 mutant occurred similarly in single infection in the presence of either mAb 2C7 or control IgG3l mAb ( Figure S2). There were no differences between the two groups in median time to clearance (5 days in both groups; P = 0.86 by Kaplan Meier analysis [ Figure S2A]), mixed model analysis (P = 0.95) ( Figure S2B) and AUC analysis (P = 0.56) ( Figure S2C).

Antibody elicited following MAP immunization
Sera from mice actively immunized with MAP1 or MAPcontrol (a peptide that does not bind 2C7), each administered with a Monophosphoryl Lipid A (MPL) containing adjuvant, were evaluated for total IgG antibody elicited against LOS of FA1090wt. Two weeks following the third booster of MAP1 (given in week 14), the median serum anti-LOS IgG antibody level had risen to 1.018 mg/ml (range 0.353 to 2.204 mg/ml) ( Figure 3A). A mixed model analysis of mean anti-LOS antibody levels over time showed significant increases between weeks designated for antibody testing (P,0.001) ( Figures 3A and S3). MAP-control immunization did not yield anti-LOS IgG responses (,0.001 mg/ml) but showed IgG antibody responses against MAPcontrol itself (data not shown). MAP1 immune sera did not contain detectable IgG antibody (,0.001 mg/ml) against LOS prepared from FA1090lgtG 2 , confirming that the LOS antibody response following MAP1 immunization was 2C7 specific (data not shown). Samples obtained from vaginal swabs showed median anti-LOS IgG levels of 0.008 mg/ml (range 0.005-0.015 mg/ml) from MAP1 immunized mice at 14 weeks (data not shown).
Serum anti-LOS IgG subclass levels, anti-LOS IgM and IgA levels in the 32 mice immunized with MAP1 were quantitated at week 14 after immunization ( Figure 3B). As expected, MAPcontrol immunized mice did not yield detectable anti-LOS antibodies. Subclass analysis of the anti-LOS response suggested a T H 1-biased response (T H 1:T H 2 index response as defined by the ratio [(IgG2a+IgG3)60.5]/IgG1 was 1.75; ratios .1 and ,1 indicate T H 1 and T H 2 responses, respectively [22]).

Immune anti-LOS IgG mediates anti-gonococcal bactericidal activity
Immunization of the 32 mice with MAP1-MPL elicited a 3.8fold higher IgM antibody response against LOS compared to IgG ( Figure 3B). Using 17% added human complement, we measured the relative contributions of anti-LOS IgG and anti-LOS IgM to serum bactericidal activity against FA1090wt in serum from 6 separate mice that had been immunized with MAP1-MPL but not challenged with N. gonorrhoeae. Non-challenged animals were immunized identically to their challenged counterparts (same lot, time of immunization and schedule); the mean 6 SE anti-LOS IgG, IgG subclasses, IgA and IgM levels in the sera of these 6 mice at week 14 were similar to the corresponding Ig levels in mice used for challenge (Table S1). Likewise, the mean differences in Ig levels between the two groups were similar (Table S2). Immune sera from the 6 mice also lacked IgG antibody against FA1090lgtG 2 LOS.
Each of the 6 sera was passed over protein A/G agarose; the flow-through contained IgM and IgA and was devoid of IgG. IgG eluted from the protein A/G column contained IgM at 10%-20% and IgA at 2-15% of serum levels (Table S3). IgG fractions ( Figure 4A; labeled IgG (Eluate)) were more bactericidal (29% median survival) than corresponding sera devoid of IgG ( Figure 4A; labeled IgM + IgA (Flow-through); 87% median survival), (P = 0.0001). Percent survival of gonococci in each of the serum fractions was expressed as a function of the nanomolar concentrations of anti-LOS IgG in the IgG fraction (Eluate) or anti-LOS IgM present in the IgM + IgA fraction (Flow-through; Figure 4B); IgA is not bactericidal. On a nanomolar basis, the anti-LOS IgG fraction ( Figure 4B; labeled IgG (Eluate)) was ,2.9-fold more bactericidal than the anti-LOS IgM fraction ( Figure 4B; labeled IgM + IgA (Flow-through)). Both anti-LOS IgG and IgM + IgA fractions showed a significant inverse correlation with bacterial survival (R 2 = 0.974 for IgG, P = 0.0003 and R 2 = 0.877 for IgM, P = 0.006).
Having shown that the majority of the killing activity in each of the 6 immune sera was mediated by anti-LOS IgG antibody, we next compared killing by polyclonal immune IgG antibody with killing by mAb 2C7. In the presence of human complement, mAb 2C7 also killed FA1090wt in a concentration-dependent manner ( Figure 4B; R 2 = 0.9661, P = 0.0004,). Using 50% survival in the bactericidal assay as a measure to discriminate bactericidal activity in the different fractions ( Figure 4B), on a nanomolar basis, the polyclonal IgG fraction in sera from mice immunized with MAP1-MPL was ,10-12.5-fold more bactericidal against FA1090wt than mAb 2C7 ( Figure 4B).

Gonococcal infection is attenuated in mice immunized with MAP1
In Groups 1 and 2 (Experimental Design shown in Figure 5), MAP1 and MAP-control immunized mice (n = 16 in each group in Group 1 and n = 10 and n = 7 respectively, in Group 2) were selected/pre-treated as described in Materials and Methods and then challenged intravaginally with FA1090wt (5.     Table S3). A. Bactericidal activity of immune Whole serum, IgG-depleted immune serum (labeled IgM + IgA (Flow through)) and immune IgG fraction (labeled IgG (Eluate)) from 6 mice immunized with MAP1-MPL. Results are expressed as % survival of FA1090wt CFUs surviving incubation for 30 min. in a reaction mixture containing 20% serum (each of the 6 immune sera were tested separately) or an amount of each of the corresponding chromatographic fractions that contained the equivalent of the amount of IgG or IgM in 20% of the parent serum. Complement (17% (v/v)) was added separately. Results are expressed as survival at 30 min.

Discussion
N. gonorrhoeae possess numerous mechanisms to subvert innate immune mechanisms [23,24] and adaptive immune responses following natural infection (reviewed in [25]). Lack of serum bactericidal activity in infected persons [26] may result from the presence of ''blocking'' (subversive) antibodies against Rmp, a conserved antigenic structure in the gonococcal outer membrane [27]. In addition, engagement of carcinoembryonic antigen cellular adhesion molecule 1 (CEACAM1) on lymphocytes may suppress CD4 T cell responses [28] that can augment or suppress the development of protective immunity. In a murine infection model [29], gonococci incite a T H 17 response that promotes an influx of PMNs that may serve as sanctuary for live gonococci that gain entry via opsonic mechanisms [30]. There are a numerous host factors in humans that influence gonococcal infection and may restrict infection in the mouse model of infection [31] A major restriction that is relevant to the IgG antibody protection model includes complement regulators, such as factor H and C4bbinding protein (C4BP) that down regulate IgG mediated complement activation in humans but do not bind to N. gonorrhoeae in the mouse and therefore do not exhibit regulator function [32,33].
In vitro correlates of protection of gonococcal vaccines have not been defined. Antibodies against serovar defining porin (Por) types may be associated with a decreased risk of re-infection with N. gonorrhoeae [34], although infected subjects were not fully protected from re-infection with the same Por type [35]. Antibodies against gonococcal opacity associated (Opa) proteins, which are phasevariable and lack serovar specificity [36,37], were associated with diminished rates of gonococcal salpingitis [37]. In contrast, antibodies against Rmp, which is conserved, were associated with enhanced likelihood of gonococcal infection in commercial sex workers in Kenya [27]. It is noteworthy that anti-Rmp antibodies block killing of gonococci by bactericidal antibodies [38].
It is no surprise that a safe and effective vaccine against N. gonorrhoeae has remained elusive. Early efforts using killed bacteria did not confer protection in humans [39]. Subsequently, pilin (Pil) derived from a single strain was effective in protecting male human volunteers from experimental urethral infection, but only against the homologous strain [40]. This Pil vaccine failed to protect U.S. military personnel stationed in Korea because of the high degree of antigenic variation in Pil across strains [41]. An outer membrane vaccine enriched with porin elicited antibody responses against Rmp [42] and 2C7 [9] that were disproportionately increased compared to porin antibody. The vaccine failed to protect against experimental challenge with the strain used to prepare the vaccine, principally, it was thought, because of the excessive/subversive Rmp response [42] Nevertheless, immunogenicity of the 2C7 epitope in this vaccine trial and also in natural infection [9] spurred our efforts to develop a peptide mimic of 2C7 as a vaccine candidate.
The 2C7 mimic displayed as a multi-antigen peptide elicited a $4-fold increase in anti-LOS titers in 87% of immunized BALB/c mice [11]. In the current study, we showed that either passive immunization with mAb 2C7 that targets the epitope or immunization with the 2C7 mimic, MAP1, shortens the course of gonococcal infection in the mouse model of genital infection. Vaginal levels of vaccine induced 2C7 immune antibodies correlated significantly with reduction in bacterial burden. These results strongly support an antibody-mediated effect that was dependent on the presence of local IgG antibody in mouse vaginas.
Greater specific bactericidal activity of IgG antibody in immune sera compared to mAb 2C7 ( Figure 4B) may have been related to a predominant complement-activating IgG subclass (IgG2a) response (Table S3) resulting from T H 1-biased immune stimulation elicited by MAP-1 used in combination with MPL, a known T H 1 adjuvant [43]. Simultaneous elicitation of IgG3 LOS antibody indirectly confirmed that the peptide immunogen indeed acted as a carbohydrate mimotope [13].
Immunization with MAP-1 MPL also elicited an anti-LOS IgM response. Specific bactericidal activity of anti-LOS IgM was ,3fold less than anti-LOS IgG but accounted for one-third of bactericidal activity in vaccinated mice. Pentavalent IgM is the most potent complement activating class of antibody. We speculate that the lesser role of anti-LOS IgM bactericidal function, compared to anti-LOS IgG, may have resulted from the recruitment of complement fixing IgG2b LOS antibodies (in addition to the predominant IgG2a) by MAP1-MPL. Diminished bactericidal function of IgM anti-LOS antibody may have also resulted from lower affinity of immune IgM to whole bacteria and/or diminished access of a bulky IgM molecule to the 2C7 epitope, which may be partially obscured by glycose extension that occurs beyond lactose substitution of HepI within the LOS core ( Figure 1).
2C7 antibodies may shorten infection by enhancing complement activation on bacteria. Small amounts of complement [44] on mucosal surfaces may not be sufficient to kill gonococci directly but killing is amplified when complement receptors on professional phagocytes, mostly PMNs, are targeted in combination with Fc receptors to enhance uptake of opsonized gonococci [45,46]. Although controversy remains about the fate of gonococci taken up by PMNs, gonococci that resist killing by PMNs are ingested principally via non-opsonic mechanisms [30]; gonococci that are taken up via opsonic mechanisms are readily killed [47].
Gonococci possess numerous phase-variable genes [48] that are responsible for antigenic diversity, a major obstacle for vaccine development. The 2C7 epitope requires expression of the phasevariable lgtG gene product (Figure 1). However, the majority of clinical isolates of N. gonorrhoeae express the 2C7 epitope [9] suggesting a critical role for this structure in gonococcal survival and pathogenesis. Our experiments showed durability/persistence of the 2C7 epitope despite selective immune pressure; FA1090 colonies isolated throughout the course of infection in immune mice always expressed the 2C7 epitope. Indeed, an lgtG deletion mutant of FA1090wt was more rapidly cleared in the murine infection model.
We acknowledge that mice immunized with MAP1 were not resistant to intravaginal challenge; rather they showed enhanced clearance of infection and diminished bacterial burdens. We note that fewer than 50% of immune animals cleared infection in the first 5 days and accelerated decreases in bacterial loads did not occur until after the third day of infection.
In conclusion, a vaccine that targets the 2C7 LOS epitope represents a significant step forward in the development of a vaccine against gonococcal infection. We are entering an era where conventional antibiotics against this infection have ceased to be universally effective. Public health measures, including immunization of high-risk individuals with a safe and effective vaccine to curb the spread of this disease are critical.

Bacterial strains
N. gonorrhoeae FA1090wt (PorB1B, streptomycin resistant, serum resistant [SR]) [49], was used for mouse experimental challenge. To evaluate the contribution of the 2C7 LOS structure to gonococcal fitness in the mouse, an isogenic mutant that lacked the 2C7 epitope, FA1090lgtG 2 , lacking HepII glycans and consequently the 2C7 epitope [12], was constructed and used for comparison. A 1.2 kb fragment of DNA containing lgtG was amplified from N. gonorrhoeae strain 398079 [50] using the primers lgtGF530 (CGCATTACCCTACCCCCTCACGCAC) and lgtGR1729 (TCTGTACGACGTTTTGAAAATTGC). The resultant amplicon was cloned into pCR2.1-TOPO (Invitrogen, Carlsbad, CA). The plasmid (pRYGW1) that was recovered was digested with StuI and ligated to a blunt 2.6 kb tetracycline resistance cassette containing tetM that had been excised from pHSX-tetM-lacIOP [51] using HincII. Linearized plasmid DNA was used to transform FA1090wt as described previously [52]. Transformants were selected on GC agar base media (Remel, Lenexa, KS) containing Isovitalex equivalent and 0.2 mg/ml tetracycline. PCR was used to confirm the lgtG::tet genotype. FA1090wt and FA190lgtG 2 LOS structures were assessed by: silver staining of LOS separated by SDS-PAGE; western blot using mAbs 3F11, L1, L8 and 2C7 [18,19,20,21]; compositional analysis by negative ion spectrometry and DNA sequencing of the homopolymer repeats present in the phase-variable lgtA, lgtC, lgtD and lgtG genes.

Human serum
Normal human sera (NHS) obtained from 13 healthy adult human volunteers were equally distributed into a pool, aliquoted and stored at 270uC.

Passive immunization of mice
Female BALB/c mice 5-6 wks of age (Jackson Laboratories) in the diestrus phase of the estrous cycle were started on treatment (that day) with 0.5 mg. of water soluble 17b-estradiol (Sigma) given subcutaneously on each of three days; 22, 0 and +2 days (before, the day of and after inoculation) to prolong the estrus phase of the cycle and promote susceptibility to N. gonorrhoeae infection. Antibiotics ineffective against N. gonorrhoeae were also used to reduce competitive microflora [53,54] . Estradiol treated mice were given either mAb 2C7 or control IgG3l mAb intraperitoneally (ip) at a dose of 20 mg (10 mg twice a day) for 3 consecutive days: 2 days prior to infection and the day of infection. Dosing of 2C7 mAb was determined a priori and yielded serum mAb levels of 1.0460.06 mg/ml on the day of infection and the subsequent 4 days and vaginal levels of 0.0160.004 mg/ml on the day of infection, the next day and 0.00660.004 mg/ml on the third day (data not shown). IgG3l mAb was administered using the same schedule.

Active immunization of mice
Two groups of 32 female BALB/c mice (Jackson Laboratories), aged 5-6 weeks, were immunized intraperitoneally (ip) on the same day with either 50 mg of MAP1 or 50 mg of MAP-control, each emulsified in 75 mg of Monophosphoryl Lipid A (MPL)containing adjuvant (Sigma Adjuvant System). Primary immunization was followed by 3 booster doses at 3 weekly intervals ( Figure 2) and blood was collected 2 weeks following each boost for serology. These mice were used for studies of experimental infection.

Fractionation of mouse immune serum
We passed aliquots (250 ml) of sera from 6 additional uninfected mice immunized with MAP1 over Protein A/G-agarose (Pierce). Flow through and eluent fractions were dialyzed (0.9% saline; Amicon 15 ultra device [30 kDa cutoff]) and concentrated back to original volumes. Anti-LOS IgG, IgM and IgA concentrations were determined by ELISA.
Measurement of anti-LOS antibodies (mouse serum, vaginal washings and mAb 2C7) by ELISA Mouse immune serum or mAb 2C7 each diluted in PBS containing 0.05% Tween 20 was dispensed into microtiter wells (Immulon 1B.) coated with whole bacterial lysates of FA1090wt (or mutant F1090lgtG 2 ) and ELISA performed [11]. Mouse vaginal cavities were swabbed with sterile polyester tipped applicators (Solon Manufacturing Company) that were rinsed/eluted in 100 ml normal saline and dispensed onto microtiter wells coated with whole bacterial (FA1090wt) lysates and ELISA performed [11].

Measurement of LOS IgG subclass antibodies
LOS IgG subclass antibodies were quantified using a mouse isotype-specific ELISA kit (Bethyl Laboratories). Diluted immune serum was added to microtiter wells coated with whole bacterial lysates of FA1090wt and ELISA performed [11].

Serum bactericidal assays
Serum bactericidal assays against FA1090wt [55] were performed with: i) mAb 2C7, ii) sera from 6 MAP1 immunized mice not challenged with N. gonorrhoeae, iii) immune serum IgG fractions eluted from protein A/G-agarose and iv) flow through fractions devoid of immune containing residual IgM + IgA. Pooled NHS was used as a source of complement.

Intravaginal challenge
Non-immune mice. A dual/competitive infection experiment was performed using a mixture of FA1090wt and FA1090lgtG 2 to determine if the 2C7 LOS structure provided gonococci a survival advantage. Female BALB/c mice 5-6 wks of age (Jackson Laboratories) in the diestrus phase of the estrous cycle were treated with water soluble 17b-estradiol (Sigma) and antibiotics as described previously [53,54] and the following day were vaginally inoculated with a mixture of FA1090wt (1.5610 5 CFU) and FA1090lgtG 2 (1.6610 5 CFU) on Day 0 and colony forming counts (CFUs) performed daily by duplicate plating. 100 ml vaginal swab rinses were quantitatively cultured for N. gonorrhoeae daily (43) onto GC agar supplemented with vancomycin, colistin, nystatin, trimethoprim and streptomycin (GC-VCNTS) and GC-VCNTS plus tetracycline (0.2 mg/ml); the latter permitted growth of FA1090 lgtG 2 but not the wt strain.
2C7 mAb passively immunized animals. Female BALB/c mice 5-6 wks of age (Jackson Laboratories) in the diestrus phase of the estrous cycle were treated with water soluble 17b-estradiol (Sigma) and antibiotics as described above [53]. Mice were passively immunized with mAb 2C7 or IgG3l mAb and were challenged intravaginally with FA1090wt (5.7610 5 CFU) on the following day (Day 0) and colony forming counts (CFUs) determined daily by duplicate plating of vaginal swab rinses as described above. A replicate experiment was performed on a second group a week later using the same number of animals and protocol used in the first group; animals were challenged with 3.6610 5 CFU of FA1090wt. Results in the two groups were analyzed separately. To assess durability of the 2C7 epitope under antibody pressure, colony blots were performed on all gonococcal colonies isolated from animals that remained infected in the replicate experiment on day 7 after inoculation. Colony containing agar plates were overlayed with nitrocellulose and transfer of colony material was allowed to proceed for at least 2 minutes or until the membranes were totally wet. Then the membranes were blotted gently with kimwipes 3 times using a fresh kimwipe each time. Membranes were then blocked with PBS-1% dry milk for 30 min at 24C and incubated with mAb 2C7 for 15 h at 4uC. Membrane-bound mAb 2C7 was incubated for one hour with alkaline phosphatase-conjugated anti-mouse IgG (Sigma) used at 1:1000 dilution in PBS-1% dry milk followed by 5-bromo-4chloro-3-indolyl phosphate/nitro blue tetrazolium (BCIP/NBT) for detection.
To test mAb 2C7 specificity, passively immunized mice inoculated with mutant FA1090lgtG 2 were also examined.
MAP1 actively immunized animals. At fourteen weeks after immunization, we identified immunized mice that had entered the diestrus phase the day before challenge was to be performed; diestrus is asynchronous and not all of the animals entered the diestrus phase at the same time. Immune mice that were in the diestrus phase of the cycle were treated with water soluble 17b-estradiol (Sigma) as above to prolong the estrous phase of the cycle and promote susceptibility to N. gonorrhoeae. Antibiotics were also used to reduce competitive microflora [53,54]. Half [16 animals] of each group that had been immunized with MAP1 or MAP-control, were challenged at this time (Group 1 in Figure 5) and colony forming counts (CFUs) determined daily by duplicate plating of vaginal swab rinses as described above. A week later, using the identical protocol that identified and treated Group 1 animals, we challenged as many of the remaining immunized mice; 10 immunized with MAP1 and 7 with MAP-control, that had entered the diestrus phase the day before challenge (Group 2 in figure 5). Results in the two groups were analyzed separately. By performing challenge experiments at two separate times, we used more animals, thereby gaining maximum usage of the immunized animals.

Statistical analysis
Experiments that compared time to clearance of N. gonorrhoeae CFU in two independent groups of mice estimated and tested three characteristics of the data: Time to clearance, longitudinal trends in mean log 10 CFU and the cumulative CFU as area under the curve (AUC). Statistical analyses were performed using mice that initially yielded bacterial colonies on Days 1 and/or 2. Median time to clearance was estimated using Kaplan-Meier survival curves; the times to clearance were compared between groups using a log-rank test. Mean log 10 CFU trends over time were compared between groups using a linear mixed model with mouse as the random effect using both a random intercept and a random slope. A quadratic function in time was determined to provide the best fit; random slopes were also quadratic in time. A likelihood ratio test was used to compare nested models (with and without the interaction term of group and time) to test whether the trend differed over time between the two groups. The mean area under the curve (log 10 CFU vs. time) was computed for each mouse to estimate the bacterial burden over time (cumulative infection); the means under the curves were compared between groups using the nonparametric rank sum test because distributions were skewed or kurtotic.
In the experiment that used unimmunized animals challenged with strains FA1090 wt and FA1090 lgtG 2 simultaneously and in passive and active immunization experiments that used strains individually the data were paired. Individual Kaplan-Meier curves were estimated with median time to clearance. Paired times to clearance were tested using a signed rank to test the paired differences. Mean log 10 CFU trends over time were compared by taking the difference between wt and mutant within each mouse at each day and modeling the difference over time, testing if the difference was significantly different than zero using a mixed model with mouse as a random effect. The mean area under the curve for wt and mutant were computed separately and paired differences were compared using a signed rank test.
Measurement and comparison of serum IgG anti-LOS levels at weeks 5, 8, 11 and 14 after immunization with MAP1 were carried out using linear mixed models with mouse as the random effect (random intercept and slope). A Wald chi-square tested the overall significance of antibody levels at each of the weeks when measurements were made; estimates of differences in levels between adjacent times (weeks) used the coefficients of the model.
The association of burden of infection as measured by AUC of log 10 CFU vs. time with vaginal antibody concentration in the combined experiments was tested using linear regression analysis stratified by experiment. Table S1 Comparison of serum immunoglobulin isotype-specific anti-LOS concentrations of immunized mice used for challenge versus identically immunized mice used only for anti-LOS measurements and bactericidal assays.

(DOC)
Table S2 Mean Differences (and 95% CI) of serum immunoglobulin isotype-specific anti-LOS concentrations between immunized mice used for challenge versus identically immunized mice used only for anti-LOS measurements and bactericidal assays.