Lactate Acquisition Promotes Successful Colonization of the Murine Genital Tract by Neisseria gonorrhoeae

Abstract

Previous studies on Neisseria gonorrhoeae have demonstrated that metabolism of lactate in the presence of glucose increases the growth rate of the bacterium and enhances its resistance to complement-mediated killing. Although these findings in vitro suggest that the acquisition of lactate promotes gonococcal colonization, the significance of this carbon source to the survival of the gonococcus in vivo remains unknown. To investigate the importance of lactate utilization during Neisseria gonorrhoeae genital tract infection, we identified the gene lctP, which encodes the gonococcal lactate permease. A mutant that lacks a functional copy of lctP was unable to take up exogenous lactate and did not grow in defined medium with lactate as the sole carbon source, in contrast to the wild-type and complemented strains; the mutant strain exhibited no growth defect in defined medium containing glucose. In defined medium containing physiological concentrations of lactate and glucose, the lctP mutant demonstrated reduced early growth and increased sensitivity to complement-mediated killing compared with the wild-type strain; the enhanced susceptibility to complement was associated with a reduction in lipopolysaccharide sialylation of the lctP mutant. The importance of lactate utilization during colonization was evaluated in the murine model of lower genital tract infection. The lctP mutant was significantly attenuated in its ability to colonize and survive in the genital tract, while the complemented mutant exhibited no defect for colonization. Lactate is a micronutrient in the genital tract that contributes to the survival of the gonococcus.

Gonorrhea is a sexually transmitted disease caused by Neisseria gonorrhoeae that elicits an intense inflammatory response within the human genital tract. The bacterium causes a purulent discharge in the urethra and/or cervix containing polymorphonuclear leukocytes (PMNs), cell debris, and serum components (5). N. gonorrhoeae must survive and multiply within this hostile environment. Lactate, along with glucose and pyruvate, is one of the few carbon energy sources that can be utilized by pathogenic Neisseria, and the acquisition of lactate has been implicated in the virulence of this species (28). The first indication that lactate metabolism might contribute to the pathogenesis of gonococcal infection was the demonstration that lactate from human neutrophils stimulated oxygen consumption by gonococci, which in turn could impair oxygen-dependent bactericidal mechanisms (1). Additionally, work on bacteria recovered directly from exudates from the lower genital tract or grown in vitro showed that the gonococcus can incorporate host-derived cytidine 5′-monophosphate-N-acetyl neuraminic acid (CMP-NANA) into its lipopolysaccharide (LPS), thereby promoting resistance to killing by complement and by phagocytes (13, 36). It has also been shown that lactate from blood cell extracts enhances LPS sialylation of the gonococcus as it emerges from lag phase during growth in a medium containing glucose (23, 30, 31); this results from a stimulation of metabolism, leading to a more rapid emergence from lag phase, a 20% increase in the rate of growth with enhanced LPS, and protein production (12). The reason for the stimulation is that lactate does not have to contribute to gluconeogenesis in the bacterium in the presence of glucose and so is solely used as a source of energy (42, 43).

These observations suggest that, in the human genital tract, where glucose is present in millimolar concentrations, lactate metabolism could enhance the successful colonization of the gonococcus. However, none of these studies provide conclusive evidence that lactate availability and utilization contribute to gonococcal pathogenesis. Rigorous testing of this hypothesis requires the construction of mutants that are specifically unable to utilize this carbon and/or energy source and analysis of the mutants in vivo. One approach to interrupt lactate metabolism is to delete genes encoding lactate dehydrogenases. However, N. gonorrhoeae has three lactate dehydrogenase isoforms (two bound to the bacterial inner membrane with the other present in the cytoplasm) (7, 8, 11), and a strain deficient in all three enzymes has not been constructed. The aim of this work was to understand the contribution of lactate on the biology of gonococcal infection. We describe the identification of the gene encoding the gonococcal lactate permease (lctP) and the impact of lactate acquisition on gonococcal colonization in vivo.

MATERIALS AND METHODS

Bacterial strains and growth media.

N. gonorrhoeae was grown on GC media with Kellogg's supplements (21) or chocolate agar at 37°C in the presence of 5% CO₂ or in supplemented liquid GC broth with 0.005 M sodium bicarbonate with agitation as described previously (19). A spontaneous streptomycin-resistant (Str^r) derivative of N. gonorrhoeae F62 (kindly provided by Dan Stein, University of Maryland) was isolated by plating 10⁸ CFU on GC agar with 100 μg/ml streptomycin. Defined medium was prepared as described previously (2), except without glycerin, sodium acetate, Tween 80, polyvinyl alcohol, and spermine. Glucose or lactate was added to media as required. Escherichia coli was cultured in Luria Bertani (LB) media. Antibiotics were added at the following concentrations: kanamycin, 50 and 75 μg/ml; chloramphenicol, 20 and 0.5 μg/ml; or erythromycin, 300 and 0.3 μg/ml for Escherichia coli and N. gonorrhoeae, respectively. Isopropyl-β-d-thiogalactopyranoside (IPTG, final concentration, 0.1 mM; Sigma) was used to induce lctP transcription.

Molecular methods and construction of strains.

Genomic and plasmid DNA were obtained from bacteria by standard methods. To construct an insertionally inactivated allele of lctP, a 1,796-bp fragment was amplified from N. gonorrhoeae F62Sm^r with oligonucleotides F1-lctP and R1-lctP (Fig. 1); SalI recognition sequences were incorporated to facilitate subsequent manipulation. Details of the oligonucleotides used in this study are given in Table 1. The PCR product was ligated into PCR-Blunt (Invitrogen) to generate pHW901. A 993-bp AgeI-XbaI fragment containing a chloramphenicol acetyltransferase gene (cat) was excised from pGCC5 and ligated into the unique BstEII site 876 nucleotides downstream from the start codon of lctP to insertionally inactivate the gene and generate pHW901a-Cat. The inactivated lctP allele was amplified with F1-lctP and R1-lctP, and the product was introduced into N. gonorrhoeae by transformation to generate GP900 (15). Allelic exchange was confirmed by PCR.

FIG. 1.

Construction of the F62Sm^r lctP mutant, GP900, and complemented mutant, GP922. (A) Diagram of lctP inactivated by the introduction of the Cat^r gene. The position of relevant PCR primers is shown. (B) PCR confirmation of allelic exchange in GP900 (F62Sm^r ΔlctP). In lanes 1 and 2, the products of reactions with primers F1-lctP and R1-lctP were loaded; as predicted, a 2.787-kb band was amplified from GP900 (due to the insertion of the cat gene) and a 1,796-bp band from the wild-type strain. Primers RCAT and R1-lctP (lanes 3 and 4) were used to amplify a 1,060-bp product from GP900 but not the wild-type strain. Sizes of a molecular marker in kb are shown. (C) Map of pGCC4-lctP used for complementation.

TABLE 1.

Oligonucleotide primers used in this study

Primer	Sequence (5′-3′)a
F1-lctP	ACGCGTCGACAAAGGATTCGTTATGGCACT
R1-lctP	ACGCGTCGACTACTACGCCTGAATGCAAAC
F2-lctP	GCGTCTTACCAAACGCTGTA
R2-lctP	GGAGAAGAACGCACCGATC
RCAT	TTTACGATGCGATTGGGATATA

^aSalI recognition sites are underlined.

To complement GP900, a wild-type allele of lctP was introduced into an intergenic region as described previously (26). Briefly, the lctP open reading frame was excised from pHW901 by digestion with SalI and then ligated into the ScaI site of pGCC4, which has an inducible lac promoter (Fig. 1C). The resulting plasmid, pGCC4-lctP, was linearized and introduced into GP900 by transformation, generating GP922. PCR analysis confirmed that recombination resulted in a wild-type copy of lctP integrated downstream of the disrupted chromosomal allele (not shown).

Lactate uptake.

Radioactive lactate uptake assays were carried out as described previously (34). Bacteria grown to mid-exponential phase were harvested, rinsed in phosphate-buffered saline (PBS, pH 7.2), and resuspended in modified Jyssums minimal media (22) with 1.4 mM arginine, 0.06 mM cysteine, 1 mM glutamate, and 1 mM glycine at 30 to 40 mg cellular protein ml⁻¹. A Clarke oxygen electrode maintained at 37°C was filled with 2 ml modified Jyssums minimal media-11 mM sodium bicarbonate. Bacteria were added to a concentration of 0.2 mg cell protein ml⁻¹ and allowed to equilibrate. The assay was started by adding 5 μM (48.1 KBq) [l-¹⁴C]lactate (specific activity, 4.85 MBq μmol⁻¹). Samples were removed, the reaction was stopped by the addition of 20 mM dl-lactate in meningococcal chemically defined medium, the samples were filtered onto nitrocellulose, and scintillation was counted.

Reverse transcription (RT)-PCR of lctP.

RNA was prepared using the TRIzol reagent (Invitrogen) from bacteria grown on solid media. RNA was purified with the RNeasy mini kit (QIAGEN) and quantified by measuring the A₂₆₀/A₂₈₀ ratio of samples. DNA was eliminated by treatment with DNase I (Sigma). cDNA was synthesized with reverse transcriptase (ThermoScript RT-PCR system kit; Invitrogen), and PCR amplification with primers F2-lctP and R2-lctP (yielding a 297-bp product) was used to monitor lctP expression.

LPS and fluorescence-activated cell sorter (FACS) analysis.

Bacteria were grown for 2 h in chemically defined liquid media with glucose (10 mM), lactate (2 mM), and exogenous CMP-NANA (final concentration, 50 μM), harvested, and then suspended and boiled in sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) loading buffer at a concentration of 10¹⁰ CFU ml⁻¹. For LPS analysis, whole-cell extracts were prepared in LPS loading buffer (10), treated with proteinase K, and then subjected to Tricine SDS-PAGE analysis. LPS was visualized by silver staining (Pierce).

For FACS analysis, strains were grown as above, collected, fixed in 3% paraformaldehyde for 1 h, and then washed three times with PBS. Next, bacteria (10⁷ cells) were incubated with the monoclonal antibody 3F11 (kind gift from Michael Apicella, University of Iowa) at a 1 in 10 dilution in PBS for 2 h at 37°C, washed twice, then resuspended in PBS-0.1% Tween 20 containing a fluorescein isothiocyanate-conjugated donkey anti-mouse polyclonal antibody (1:200 dilution; Jackson ImmunoResearch Laboratories, Inc., Bath, United Kingdom), and incubated for 1 h on ice. After washing in PBS-0.1% Tween 20, fluorescence was measured using a FACSCalibur analyzer (Becton Dickson), recording at least 40,000 events. The gate was set at around 1% of cells after incubation with PBS. Results were calculated as the mean fluorescence index (calculated as the geometric mean multiplied by the percentage of positive cells).

Serum resistance and murine genital infection.

N. gonorrhoeae was grown in liquid media as above, and 10⁴ bacteria were incubated with serial dilutions of human serum from a healthy donor for 1 h, and the proportion of bacteria surviving was determined by plating to solid media as described previously (10). Control assays included bacteria with sera that had been inactivated by heating at 56°C for 30 min. Assays were performed on four separate occasions.

Infection of female 17-β-estradiol-treated BALB/c mice (6 to 8 weeks old) (National Cancer Institute, Bethesda, MD) was performed as described previously (18), except that mice received streptomycin sulfate (0.24 mg twice daily) and no vancomycin. For competitive infections, mice received suspensions of 10⁶ CFU of equal numbers of wild-type and mutant bacteria in 20 μl PBS. Bacteria were recovered by vaginal swabbing on alternate days for 14 days following inoculation. The competitive index (C.I.) was calculated as (the number of mutants recovered/number of wild type recovered)/(the number of mutants in the inoculum/number of wild type in the inoculum). Genital secretions were collected by placing a polyester swab into the mouse vagina for 1 min. Swabs from 5 to 10 mice were combined into a single 0.5-ml tube and centrifuged for 5 min at 10,000 × g. The concentrations of glucose and lactate in the supernatants were measured spectrophotometrically in triplicate using the Glucose Kit (diagnostics procedure no. 115; Sigma) or Lactate Kit (diagnostics procedure no. 735; Sigma) as described by the manufacturer. All animal experiments were conducted under a protocol approved by the Uniformed Services University's Institutional Animal Care and Use Committee.

RESULTS

Identification of the N. gonorrhoeae lactate permease.

We recently identified the gene lctP, encoding the Neisseria meningitidis lactate permease (10). The gene encoding the permease was originally identified as being required for meningococcal bacteremia; a mutant with a transposon insertion in lctP was attenuated in the infant rat model of bacteremia (37). Blast searches of the N. gonorrhoeae genome sequence (http://www.genome.ou.edu/gono.html) with the deduced amino acid sequence of the meningococcal LctP revealed an identical sequence (NGO1449) except for a single, predicted amino acid difference (Val¹⁶⁷ to Ala¹⁶⁷ for the meningococcal and gonococcal sequences, respectively). The gonococcal sequence is expected to encode a protein with 14 transmembrane domains, consistent with an inner membrane permease.

First we constructed GP900, a deletion mutant of NGO1449 in N. gonorrhoeae F62Sm^r, and complemented the mutant by introducing a single copy of the gene on the chromosome by transformation with pGCC4-lctP (Fig. 1), resulting in GP922. To examine the influence of NGO1449 on lactate metabolism, lactate uptake studies were performed on F62Sm^r, GP900 and GP922 (Fig. 2A). F62Sm^r rapidly accumulated [¹⁴C]lactate, assimilating 9.1 nmol per mg of protein within 1 min, whereas uptake by GP900 was negligible on two independent occasions (Fig. 2A and data not shown); the complemented strain, GP922, acquired exogenous lactate with kinetics similar to those of the wild-type isolate.

FIG. 2.

(A) Uptake of radiolabeled l-lactate by N. gonorrhoeae. Bacteria were incubated in 5 μM [l-¹⁴C]lactate, and uptake was measured over time. The mutant lacking NGO1449 (GP900) does not assimilate [l-¹⁴C]lactate, while the parental and complemented strains rapidly take up this carbon source; the strains in all panels are shown in the key. Growth kinetics of the wild-type strain (F62Sm^r), the lctP mutant (GP900), and the complemented mutant (GP922) in defined media supplemented with glucose (panel B, 28 mM) or lactate (panel C, 28 mM) as the sole carbon sources are shown. The lctP mutant failed to grow with lactate as the sole carbon source but showed no replication defect when grown with glucose.

Next we examined the growth characteristics of the strains in media with defined carbon energy sources. The wild-type strain F62Sm^r, GP900, and GP922 exhibited equivalent growth kinetics in defined media supplemented with glucose as the sole carbon source (Fig. 2B). In contrast, GP900 was unable to grow in media supplemented with lactate alone (Fig. 2C). The growth of the mutant was restored to wild-type levels in strain GP922 (Fig. 2C), demonstrating that loss of NGO1449 is solely responsible for the growth defect observed. Even though the complementing lctP allele was downstream of the lac promoter, growth of GP922 in lactate as the sole carbon source did not require the presence of the inducer, IPTG (Fig. 2C). Therefore, functional complementation of NGO1449 in GP922 is not dependent on exogenous IPTG. Indeed, the transcript of the lctP homologue was detected in the wild-type strain and the complemented mutant in the absence of IPTG, indicating that the lac promoter is active without IPTG induction (Fig. 3). Taken together, the results of uptake and growth demonstrate that NGO1449 is required for lactate acquisition and utilization by N. gonorrhoeae, and this gene was therefore designated lctP.

FIG. 3.

RT-PCR analysis of lctP transcription with primers R2-lctP and F2-lctP in the absence of IPTG, indicated by a 297-bp product in the wild-type (F62Sm^r) and complemented (GP922) strains but not the lctP mutant (GP900). RNA was extracted from bacteria grown without IPTG. The sizes of a molecular marker (bp) are shown.

The gonococcal lactate permease is necessary for the lactate-mediated stimulation of growth.

Millimolar concentrations of lactate enhance the metabolism of the gonococcus when added to cultures containing glucose. This is most obviously seen as a rapid emergence from lag phase (12), an effect that may be important during the initial stages of gonococcal colonization at mucosal surfaces. Therefore, we examined whether this stimulation of metabolism was abolished in a strain unable to utilize lactate though inactivation of its cognate permease. Strains were cultured for 2 h in defined liquid medium with physiological concentrations of glucose and lactate (10 and 2 mM, respectively), and the extent of growth was monitored by measuring the optical density of cultures. There was a reproducible difference in the degree of growth of the wild-type strain (F62Sm^r) and the lactate permease mutant (GP900) over this time (Fig. 4A), with the wild-type reaching a higher optical density (OD) than the mutant on each occasion (average percent difference, 28.8%; P < 0.01, Student's t test); the complemented strain had a similar early growth advantage over the lctP mutant (Fig. 4A).

FIG. 4.

(A) The lactate permease mutant, GP900, has a reduced initial growth rate compared with the parental strain, F62Sm^r, and the complemented strain. The mean percent increases in OD at the A₆₀₀ of cultures of F62Sm^r and the complemented strain GP922 were compared with the lctP mutant after 2 h of growth in defined medium containing glucose (10 mM) and lactate (2 mM). The initial OD of all cultures was 0.1, and the error bars show the standard errors of the means. (B) Serum resistance of gonococcal strains. Bacteria were incubated with a series of dilutions of NHS or HIS for 1 h. The strains and serum used are shown in the key. The percent survival of the strains compared with the inoculum is shown. GP900 is more sensitive to human complement than the wild-type strain; this defect is restored in the complemented strain GP922.

lctP enhances gonococcal serum resistance.

Previous studies have shown that lactate, when added to cultures containing glucose, enhances CMP-NANA-dependent resistance of N. gonorrhoeae against complement (23, 30, 31). The ability to resist complement-mediated killing through LPS sialylation is an important feature of gonococcal pathogenesis, particularly for strains such as F62 that do not have inherently high levels of serum resistance due to the type of porin expressed. Therefore, we compared the survival of the lactate permease mutant against the wild-type strain in the presence of normal human serum. Bacteria were grown for 2 to 3 h in media containing glucose, lactate, and CMP-NANA, and then exposed to either normal human serum (NHS) or heat-inactivated serum (HIS, 56°C for 30 min) at a final concentration of 1 to 5%. Although there was interexperiment variation in the absolute numbers of bacteria recovered, on four independent occasions, the wild-type strain was reproducibly recovered at significantly higher numbers than GP900 (F62Sm^r ΔlctP) in NHS (Fig. 4B), while the introduction of the wild-type copy of the gene into the mutant in GP922 partially restored resistance against complement; the survival of the strains was identical in HIS (Fig. 4B).

The N. gonorrhoeae lactate permease contributes to LPS sialylation.

Sialylation of LPS contributes to the ability of N. gonorrhoeae to avoid complement-mediated killing, and LPS sialylation is increased by the addition of lactate to the gonococcus grown in the presence of glucose (31). Therefore, we next examined whether the enhanced sensitivity of GP900 (F62Sm^r ΔlctP) to complement is associated with alterations in LPS sialylation. Bacteria were collected after growth in the presence of glucose, lactate, and CMP-NANA to allow exogenous sialylation of LPS, and FACS analysis was performed using the monoclonal antibody 3F11, which recognizes unsialylated LPS (41). As found previously with N. gonorrhoeae grown with or without lactate (23, 30), there was reduced sialylation of GP900 compared with the wild-type strain, as detected by increased binding of 3F11 (mean fluorescence index: F62Sm^r, 18.4; F62Sm^r ΔlctP, 86.3; P = 0.0021, Student's t test).

LctP promotes successful colonization of the genital tract.

The biological relevance of lactate utilization for the survival of N. gonorrhoeae in vivo was investigated in the murine model of genital tract infection (18). Initially, we determined the concentrations of glucose and lactate in murine serum and vaginal secretions. The results show that both glucose and lactate are detected in millimolar concentrations in both hosts (Table 2 and reference 32). Therefore, the bacterium is exposed to mixed carbon sources in the murine and human genital tracts, and the murine model is relevant for assessing the impact of lactate availability on gonococcal colonization.

TABLE 2.

Serum and vaginal concentrations of glucose and lactate^a

Sample	Concn (mM) of:
	Glucose	Lactate
Human serum	3.5-5.5	0.7-1.4
Murine serum	10.6 (0.2)	3.6 (0.09)
Human vaginal fluidb	3.3	6.2
Murine vaginal fluid	4.4 (1.4)	8.4 (2.5)

^aStandard deviations are shown in parentheses.

^bValues obtained from reference 32.

The survival of the lctP mutant and wild-type strains was directly compared in mixed-inoculum infections (4). The lctP mutant was at a substantial disadvantage in vivo compared with the parental strain, with a decrease in recovery of GP900 within 2 to 12 days in 6 of 7 animals (Fig. 5A). Furthermore, the lctP mutant was eliminated from the genital tract significantly faster than the wild-type strain (Fig. 5B). As phenotypic and transcript analyses demonstrated that the wild-type lctP allele is expressed in GP922 even in the absence of IPTG induction, we examined the colonization capacity of the complemented strain in vivo. Complementation restored the ability of the mutant to survive in the genital tract to wild-type levels. The duration of colonization of the complemented mutant, GP922, was equivalent to the wild-type strain (Fig. 5B). Additionally, in 6 of 7 animals, GP922 did not have a significant colonization defect compared to the wild-type strain through the course of experiments, and in 1 animal, there was only a transient decrease in the C.I. of GP922 on days 6 and 8, which returned to >0.5 on days 10 and 12 (not shown). Therefore lctP promotes the survival of N. gonorrhoeae in vivo.

FIG. 5.

The gonococcal lactate permease promotes effective vaginal colonization in the murine model. (A) Bacteria were recovered from mice by vaginal swabbing on alternate days after inoculation with a 1:1 mixture of wild-type and mutant strains. The relative numbers of the wild-type and mutant strains were determined by plating to medium with and without chloramphenicol. The C.I. of the lctP mutant (GP900) for individual animals are shown as filled circles, and the mean C.I. for all animals is given at each time point. A low C.I. indicates attenuation of the mutant. (B) Duration of carriage of the wild-type, lactate permease mutant (GP900), and complemented (GP922) strains. The difference in length of colonization of GP900 is significantly shorter than that of the other strains (P < 0.01, Student's t test).

DISCUSSION

At all stages of the disease process, bacteria must acquire nutrients from the surrounding microenvironment to sustain growth and replication. Recent findings from genetic screening methods to identify factors required for bacterial pathogenesis in vivo have served to emphasize the essential role of microbial physiology during pathogenesis (25, 35).

N. gonorrhoeae has stringent nutrient requirements for growth. Only a limited repertoire of carbon energy sources, including glucose and lactate, are effectively utilized by this bacterium (2). Both are available in the genital tract, the normal habitat of N. gonorrhoeae. While biochemical evidence has demonstrated the effect of additional lactate on gonococcal metabolism and resistance against complement, this is the first description of a strain of N. gonorrhoeae that is specifically unable to utilize this carbon source. This was achieved by taking advantage of the recent identification of the gene encoding the lactate permease in the related pathogen N. meningitidis (10). The corresponding gene in N. gonorrhoeae was inactivated, and the mutant was defective for the uptake and utilization of exogenous lactate and failed to grow with lactate as the sole carbon source, confirming that its product functions as the lactate permease. Complementation was successfully achieved with a single, chromosomal copy of lctP under the control of the lac promoter, which allows low-level expression even in the absence of the inducer IPTG (Fig. 3).

Infection with N. gonorrhoeae leads to a marked inflammatory response in the genital tract, and the levels of lactate may increase during gonorrhea as PMNs actively metabolize glucose and generate lactate (1). However, under normal conditions, lactobacilli are the main source of lactate in the female lower genital tract, and inflammation is not necessary for the appearance of this carbon source in the vagina and cervix. While several host-specific features may limit the capacity of mice to completely mimic the human infection (20, 29), there are many physiological similarities between the lower genital tracts of female mice and women, including the presence of commensal lactobacilli and mixed carbon sources for the gonococcus (Table 2); this supports the use of the murine model for the study of pathogenicity factors involved in bacterial survival in vivo.

The presence of lactate in secretions from the human genital tract and the addition of lactate to N. gonorrhoeae grown in glucose in vitro enhance resistance against complement-mediated lysis through increased LPS sialylation by enhanced synthesis of LPS and Lst. The additional lactate acts as an effective energy source via the tricarboxylic acid cycle, allowing the incorporation of glucose carbon into cell wall lipids rather than catabolism via the Entner Doudoroff pathway (43). Our genetic studies with a mutant lacking LctP are consistent with these previous biochemical observations. The lctP mutant was significantly more sensitive to lysis through the activity of human complement than the wild-type strain. This correlated with the reduced LPS sialylation of F62Sm^r ΔlctP compared with the parental strain. The addition of sialic acid to LPS leads to increased serum resistance by blocking deposition of complement components, restricting access of antibody to the bacterial cell surface (38, 39), and/or inhibition of the alternative complement pathway by binding factor H (33). In keeping with this, lactate utilization also promotes LPS sialylation in the meningococcus (10), although this occurs via the sialic acid biosynthesis pathway (10), which is absent from N. gonorrhoeae. Our findings do not exclude the possibility that changes aside from LPS sialylation are responsible for the enhanced sensitivity of F62Sm^r ΔlctP to complement. Indeed, the addition of lactate to cultures of the gonococcus grown in physiological concentrations of glucose is associated with changes in the fatty acid and carbohydrate composition of LPS that are not detected by SDS-PAGE or Western analyses (41, 42). Alterations in both the lipid and carbohydrate components of LPS have both been associated with changes in the virulence of pathogenic Neisseria (6, 38).

The loss of lctP in strain F62Sm^r markedly impaired the colonization capacity of N. gonorrhoeae, reducing both the density and the duration of carriage. There are several potential explanations for the attenuation. First, we found that the lctP mutant has a growth defect in a defined medium containing physiological concentrations of lactate and glucose. This may be critical during colonization when bacterial replication in the genital tract mucosa might be limited by an inability to utilize lactate as an effective carbon energy source (12). Second, the mutant was sensitive to killing by human complement, although this might not be of direct relevance to survival in the murine genital tract. Although N. gonorrhoeae incorporates host-derived CMP-NANA during growth in the murine genital tract at a level that confers resistance to killing by human serum, LPS sialylation does not contribute to the avoidance of bacteriolysis mediated by soluble components of the murine complement system (40). This finding is consistent with recent evidence for host specificity in the interactions between N. gonorrhoeae and certain complement factors, such as C4bp and factor H (29). Instead, lactate utilization may be important during interactions between the bacterium and murine PMNs. Sialylated gonococci are more resistant to killing by murine neutrophils in vitro, which probably reflects significantly reduced uptake of sialylated strains. Sialylated gonococci also induce a weaker respiratory burst in murine PMNs (40). It has also been shown that the gonococcal metabolism of lactate produced by PMNs reduces the oxygen available for the phagocytic respiratory burst (1). Further studies are under way to define the mechanisms that underlie the attenuation of the lctP mutant in vivo.

Recent studies have emphasized the importance of carbon metabolism to the persistence of bacterial pathogens in vivo. Enzymes in the Entner-Doudoroff pathway are essential for the colonization of the gastrointestinal tract by E. coli (3); this is the major glycolytic pathway in N. gonorrhoeae (16). Furthermore, the expression of isocitrate lyase, an enzyme in the glyoxylate cycle, is upregulated by Mycobacterium tuberculosis in phagocytes (14) and is necessary for survival in these cells (24). However the glyoxylate cycle (which replenishes the tricarboxylic acid cycle) is not present in pathogenic Neisseria. With regard to exogenous carbon energy sources, the acquisition of lactate is also necessary for Haemophilus influenzae to cause bacteremic infection (17), while the meningococcal lactate permease is required for both nasopharyngeal colonization and disseminated infection (9, 10). Our results demonstrate that successful uptake of this carbon source from the environment also contributes to the survival of N. gonorrhoeae in a model of lower genital tract infection. Further studies on gonococcal strains unable to acquire other carbon sources, such as glutamate (27), or which have defects in other metabolic pathways should provide further information about the nutritional status of this important pathogen in vivo.