Role of Neisseria meningitidis luxS in Cell-to-Cell Signaling and Bacteremic Infection

Abstract

Numerous pathogenic bacteria contain luxS, which is required for autoinducer-2 production. Here, we demonstrate that Neisseria meningitidis contains a functional copy of luxS that is necessary for full meningococcal virulence; strains with a luxS deletion are defective for bacteremia, a prerequisite of meningococcal pathogenesis.

Bacteria have evolved mechanisms to coordinate gene expression in response to population density. Known as quorum sensing, these mechanisms involve the production and detection of signaling molecules (called autoinducers or pheromones), which modulate critical functions including virulence factor production, plasmid conjugal transfer, activation of secretion systems, swarming, swimming and twitching motility, biofilm differentiation, and bioluminescence (for reviews, see references 2, 13, 15, 28, and 30). Thus, bacteria in populations display properties denied individual cells.

Gram-positive and gram-negative bacteria are known to utilize distinct signal molecules for quorum sensing. Gram-positive bacteria export and detect small peptides, while gram-negative bacteria produce membrane-permeative N-acyl homoserine lactones (13, 15). Recently, however, potential quorum sensing systems based on a signal molecule called autoinducer-2 (AI-2) were shown to be present in both gram-positive and gram-negative species (2, 23, 24). The luxS gene product is required for AI-2 production by Salmonella enterica serovar Typhimurium (24), Escherichia coli (19, 23), Shigella flexneri (6), Helicobacter pylori (9, 11), Porphyromonas gingivalis (4, 5), Streptococcus pyogenes (12), and Actinobacillus actinomycetemcomitans (8). The chemical structure of AI-2 appears to be conserved, possibly allowing interspecies communication (2, 3, 18).

Neisseria meningitidis colonizes the human nasopharynx and can cause meningitis and/or septicemia (16). Here, we demonstrate that N. meningitidis possesses a functional luxS necessary for AI-2 production and full meningococcal virulence.

Serogroup B N. meningitidis possesses a functional luxS gene.

Investigation of the MC58 genome sequence (25) using LuxS from Vibrio harveyi (LuxS_Vh) led to the identification of NMB1981 (N. meningitidis serogroup B annotated sequence), a 504-bp open reading frame (ORF) with 80% identity to LuxS_Vh which is annotated as being of unknown function. To establish whether NMB1981 encodes a functional product, the gene was amplified with rTth DNA polymerase XL (Perkin-Elmer, Warrington, United Kingdom), ligated in both orientations into pGEM T-Easy (Promega, Madison, Wis.), and then transferred into Escherichia coli DH5α, which cannot produce AI-2 due to a mutation in luxS (24). Strains were grown at 37°C in Luria-Bertani medium, and AI-2 production was assayed using V. harveyi BB170 as the reporter (3). Exogenous AI-2 induces premature bioluminescence by this reporter strain. Only clones containing NMB1981 expressed through the vector-encoded promoter (pGEMluxS-F) produced culture supernatants with AI-2 activity (Fig. 1A); no activity was detected if the insert was in the opposite orientation (pGEMluxS-R). Identical results were obtained with the NMB1981 homologue from strain SD (data not shown). Thus, NMB1981 complements the defective luxS in DH5α and was designated luxS_Nm.

FIG. 1.

(A) Complementation of E. coli DH5α with luxS_Nm. The growth profiles and extracellular AI-2 activity of E. coli DH5α carrying either pGEM T-Easy or the same vector containing luxS_Nm in the forward or reverse orientation were measured. Growth profiles: empty vector, closed squares; luxS_Nm in forward orientation, closed circles; luxS_Nm in reverse orientation, open circles. AI-2 activity: V. harveyi BB120 (wild type, after 24 h), black bar; luxS_Nm in forward orientation, grey bars; luxS_Nm in reverse orientation, white bars. Activities are presented as the fold increase of bioluminescence in relation to E. coli DH5α carrying an empty vector at the respective time point. (B) Growth profiles and AI-2 activity of N. meningitidis. N. meningitidis strains MC58 (closed circles) and MC58ΔluxS (open circles) were grown, and the optical densities at 600 nm of cultures were measured. Culture supernatants of MC58 (grey bars), MC58ΔluxS (white bars), and V. harveyi BB120 (black bar) were assayed for AI-2 activity. Activities are presented as the fold increase of bioluminescence in relation to the control (sterile medium) at the respective time points. The results for MC58ΔluxS^Ect were identical to those for the wild type and are not shown. For panels A and B, single representative experiments are shown, although similar results were obtained on three occasions for both assays.

Expression of N. meningitidis AI-2 activity is dependent on luxS_Nm.

Next, we investigated whether N. meningitidis initiates luxS_Nm-dependent AI-2 production. A total of 432 bp of luxS_Nm (nucleotides 4 to 436 of the ORF) was deleted by inverse PCR and replaced with the kanamycin resistance gene from Tn903, yielding pGEMΔluxS-kan. The insert of pGEMΔluxS-kan was then cloned into pGIT5.3, a vector containing the Neisseria uptake sequence, and the resulting plasmid (pGIT5.3ΔluxS-kan) was introduced into N. meningitidis by transformation; Southern hybridization confirmed the integration of the deleted allele in MC58ΔluxS and B16/B6ΔluxS. A complemented strain, MC58ΔluxS^Ect, was constructed to enable the attribution of functions to luxS_Nm. luxS_Nm, under the control of an opa promoter, was introduced in the intergenic region between NMB102 and NMB103 (genes oriented in a tail-to-tail fashion), with ermC downstream of luxS. The construct (pYH204) was introduced into MC58ΔluxS by transformation, generating MC58ΔluxS^Ect; Southern analysis confirmed that integration into the NMB102/NMB103 intergenic region had occurred. N. meningitidis bacteria were grown at 37°C on brain heart infusion plates with Levinthal's supplement or in Mueller-Hinton medium for 24 h with shaking for AI-2 production. A list of the bacterial strains and plasmids used in this study is presented in Table 1.

TABLE 1.

Bacterial strains and plasmids used in this study

Strain or plasmid	Genotype and/or details	Source or reference
Bacterial strains
E. coli DH5α	F′ endA1 supE44 thi-1 hsdR17(rK− mK+) recA1 gyrA relA1 Δ(lacIZYA-argF)U169 deoR [φ80d lacΔ(lacZ)M15]	Gibco-BRL
N. meningitidis
MC58	Serogroup B; ET-5	25
B16/B6	Serogroup B; ET-37	17
SD	Serogroup B: B:15:P1.16	1
MC58ΔluxS	Deleted luxSNm replaced with kan	This study
MC58ΔluxSEct	Wild-type luxSNm located in the NMB102/103 intergenic region of MC58ΔluxS	This study
V. harveyi
BB120	Wild type	3
BB170	Sensor AI-1−, sensor AI-2+	3
Plasmids
pGEM T-Easy	Cloning vector	Promega
pGEM luxS-F	1.5-kb fragment encoding luxS of N. meningitidis MC58 and flanking region in pGEM-T Easy, forward orientation	This study
pGEM luxS-R	1.5-kb fragment encoding luxS of N. meningitidis MC58 and flanking region in pGEM-T Easy, reverse orientation	This study
pGEMΔluxS-kan	luxSNm with internal deletion and Tn903 kanamycin resistance gene	This study
pGIT5.3	Derived from pCRII; containing a chloramphenicol resistance cassette and Neisseria DNA uptake sequence	L. Barber and T. Baldwin, unpublished data
pGIT5.3ΔluxS-kan	pGIT5.3 derivative containing the insert of pGEMΔluxS-kan	This study
pYH204	luxSNm in the NMB102/103 intergenic region with ermC	This study

AI-2 activity was detected in MC58 culture supernatants from early logarithmic growth onward (Fig. 1B). While no AI-2 activity was observed in supernatants of MC58ΔluxS, the single ectopic luxS allele in MC58ΔluxS^Ect restored AI-2 activity to wild-type levels. Identical results were obtained with strain B16/B6 and its corresponding mutants (not shown).

N. meningitidis luxS mutants are attenuated for bacteremic infection.

To investigate the influence of luxS_Nm on meningococcal pathogenesis, the ability of MC58ΔluxS to cause disseminated disease was compared directly against that of MC58. Bacteria were grown overnight on brain heart infusion agar and suspended in phosphate-buffered saline, and the number of CFU was determined. Bacteria (100 μl) in phosphate-buffered saline were injected intraperitoneally into five-day-old infant rats (Wistar; Harlan, Bicester, United Kingdom). The virulence of strains was assessed using groups of at least three animals in each of three experiments by determining their competitive index (CI). The CI is the proportion of mutant to wild-type bacteria recovered from the bloodstream following inoculation with a 1:1 ratio of the strains (26); results were compared by using a one-tailed Student's t test. MC58ΔluxS showed decreased survival in infected animals when compared with MC58 (P < 0.01; Fig. 2). Similar results were obtained with B16/B6 (not shown). Providing luxS_Nm in trans (MC58ΔluxS^Ect) restored the virulence to wild-type levels, demonstrating that the loss of luxS, not any polar effect, is responsible for attenuation.

FIG. 2.

MC58ΔluxS is defective for bacteremic infection compared with the parental strain, MC58. This defect is restored by supplying a single, intact copy of luxS in trans to MC58ΔluxS (MC58ΔluxS^Ect); the difference in the virulences of MC58ΔluxS and MC58ΔluxS^Ect is highly significant (P < 0.001). The CI is the proportion of the strains in the bloodstream 21 to 24 h following infection with a 1:1 ratio of the strains.

Complete genome sequences of bacterial pathogens provide an important resource for microbial research (21), though the annotations may miss biologically important homologies. While a luxS homologue was identified in the annotated Z2491 serogroup A strain sequence (14), the annotated MC58 genome (25) does not contain a luxS homologue. However, investigation of the MC58 database revealed an ORF closely related to luxS_Vh. We demonstrate that luxS_Nm complements DH5α and that N. meningitidis expresses AI-2 in a luxS-dependent fashion with maximal production during late exponential growth, consistent with a role in quorum sensing.

Studies with the infant rat model show that luxS_Nm contributes to the ability of the bacterium to cause bacteremia, a prerequisite for meningococcal pathogenicity. This model has been used to demonstrate the influence of other well-characterized virulence determinants on disseminated infection (20, 27). Although CIs provide a sensitive measure of a strain's virulence by eliminating host-to-host variation (22, 26), our findings may underestimate the impact of luxS_Nm on pathogenesis. AI-2 produced by the wild type could partially cross-feed a luxS_Nm mutant in mixed-inoculum experiments. However, the results are also consistent with a metabolic function of the LuxS protein, which is distinct from quorum sensing (18, 29).

Little is known about the regulatory networks governing the expression of N. meningitidis virulence factors apart from that of a LysR homologue, required for entry into host cells (7). Identification of the response elements and the genes controlled by AI-2-dependent signaling may help elucidate this aspect of pathogenesis and identify novel virulence determinants for this important human pathogen. Furthermore, the demonstration that luxS_Nm is required for full virulence suggests that interrupting this pathway may provide a means for preventing meningococcal disease (10).