Lactate-Induced Dispersal of Neisseria meningitidis Microcolonies Is Mediated by Changes in Cell Density and Pilus Retraction and Is Influenced by Temperature Change

ABSTRACT Neisseria meningitidis is the etiologic agent of meningococcal meningitis and sepsis. Initial colonization of meningococci in the upper respiratory tract epithelium is crucial for disease development. The colonization occurs in several steps and expression of type IV pili (Tfp) is essential for both attachment and microcolony formation of encapsulated bacteria. Previously, we have shown that host-derived lactate induces synchronized dispersal of meningococcal microcolonies. In this study, we demonstrated that lactate-induced dispersal is dependent on bacterial concentration but not on the quorum-sensing system autoinducer-2 or the two-component systems NarP/NarQ, PilR/PilS, NtrY/NtrX, and MisR/MisS. Further, there were no changes in expression of genes related to assembly, elongation, retraction, and modification of Tfp throughout the time course of lactate induction. By using pilT and pptB mutants, however, we found that lactate-induced dispersal was dependent on PilT retraction but not on phosphoglycerol modification of Tfp even though the PptB activity was important for preventing reaggregation postdispersal. Furthermore, protein synthesis was required for lactate-induced dispersal. Finally, we found that at a lower temperature, lactate-induced dispersal was delayed and unsynchronized, and bacteria reformed microcolonies. We conclude that lactate-induced microcolony dispersal is dependent on bacterial concentration, PilT-dependent Tfp retraction, and protein synthesis and is influenced by environmental temperature.

B acterial pathogens encounter constant changes in their environment and must respond quickly to survive and proliferate (1). Neisseria meningitidis is a human-restricted pathogen that is well adapted to survival in the upper respiratory tract. The nasopharyngeal epithelium is the natural reservoir for meningococci; however, bacteria can penetrate the cell barrier and cause sepsis and/or meningitis (2). Initial colonization of the mucosal epithelium is crucial for disease development and is mediated by type IV pili (Tfp). Once Tfp-mediated attachment to the mucosa has been established, meningococci start forming microcolonies. Microcolony formation on host cells stimulates cytoskeletal rearrangement that can contribute to bacterial resistance against physical forces (3). However, to invade the mucosa, bacteria must disperse from the microcolonies and form close contacts with host cells (4)(5)(6).
Several factors have been demonstrated to contribute to the dispersal of microcolonies. The ATPase PilT is responsible for Tfp retraction. A pilT deletion mutant exhibited a hyperpiliated aggregative phenotype unable to disperse and form intimate adhesion to host cells (5). It has been proposed that an initial contact with host cells results in increased expression of pptB encoding pilin phosphotransferase B. Upregulation of pptB results in posttranslational modifications (PTM) by adding phosphoglycerol to

RESULTS
Initiation of N. meningitidis lactate-induced microcolony dispersal is dependent on bacterial concentration but not on AI-2 production. We have previously shown that host-derived lactate induces fast and synchronized microcolony dispersal in N. meningitidis (8). We found that lactate, although required for microcolony dispersal, did not prevent microcolony formation. When bacteria were inoculated in growth medium (10 7 CFU/ml) with 2 mM lactate, it took 2 h to form microcolonies and, after an additional 30 min, the microcolonies dispersed in a synchronized way (8). We therefore examined whether the lactate-mediated dispersal was dependent on cell density. First, we incubated bacteria at different concentrations in the presence of lactate and followed the timing of dispersal by live-cell microscopy. As shown in Fig. 1A, the initiation of microcolony dispersal was dependent on bacterial concentration; the more bacteria were present in the suspension, the earlier the dispersal occurred. Consequently, we speculated that a density-dependent mechanism such as QS might trigger dispersal. To examine the role of quorum sensing, we constructed a luxS-deficient mutant (DluxS). However, the DluxS mutant showed no difference in timing of dispersal from that of the wild-type strain (Fig. 1B). Taken together, these results show that N. meningitidis lactate-induced dispersal occurs in a density-dependent manner; however, it is not affected by disruption of AI-2 production.
The TCS PilR/PilS, NarP/NarQ, NtrY/NtrX, and MisR/MisS are not required for N. meningitidis lactate-induced dispersal. To further identify the potential signaling behind the lactate-induced dispersal, we examined the role of TCS. In Sigurlásdóttir et al. (8), we found no difference in the mRNA level of the TCS-regulator MisR upon lactate-induced dispersal. However, not only the amount of the response regulator but also the phosphorylation level can affect the genes controlled by the systems (30). We therefore attempted to generate deletion mutants of all the TCS regulators encoded by N. meningitidis FAM20 to examine their involvement. We successfully generated DnarP (regulator of NarP/NarQ), DpilR (regulator of PilR/PilS), and DntrX (regulator of NtrY/NtrX) deletion mutants in strain FAM20 and confirmed them by PCR. Despite several attempts, we were unable to create a FAM20 DmisR (regulator of MisR/MisS) deletion mutant, indicating heterogeneity among strains and a possible essential role of MisR in FAM20. Therefore, we used an already existing DmisR mutant in the serogroup C isolate N. meningitidis L91543 (31). In induction assays, we observed that deletion of narP and pilR did not affect the timing of dispersal in the presence of lactate ( Fig. 2A). Although we observed dispersal of DntrX aggregates upon addition of L-lactate, we noted that DntrX mutant gave rise to unusual cell morphology in live-cell images, suggesting that DntrX mutant might suffer abnormalities during cell division (Fig. 2B). We also observed dispersal of both L91543 wild-type and DmisR (previous known as DphoP [31]) aggregates upon addition of L-lactate. However, the dispersal duration was longer than that of FAM20 and the microcolonies could not completely disperse (Fig. 2C). In summary, the data suggest that regulators of NarP/NarQ, PilR/PilS, NtrX/NtrY, FIG 1 Initiation of N. meningitidis lactate-induced microcolony dispersal is dependent on bacterial concentration but not on AI-2 production. (A) Bacteria were resuspended in medium with of 2 mM L-lactate, filtered through a 5-mm filter to break preexisting aggregates, and diluted to suspensions with different concentrations ranging from 2 Â 10 6 to 2 Â 10 7 CFU/ml. The timing of microcolony dispersal was examined by live-cell imaging. (B) Induction assays comparing microcolony dispersal of the wild-type (WT) and DluxS bacteria. Bacteria were resuspended to 10 7 CFU/ml in DMEM containing 1% FBS and allowed to form aggregates for 3 h. Lactate solutions were added to the final concentrations of 10 mM at a 1:1 ratio, and microcolony dispersal was followed by live-cell imaging. A black horizontal line represents the 3-h time point of induction in panel B. DMEM was added as a control. The bars represent the means of three independent experiments. Error bars represent the standard deviations. ns, nonsignificant.

Lactate-Mediated Microcolony Dispersal
Infection and Immunity and MisR/MisS TCS are not involved in lactate-induced N. meningitidis microcolony dispersal. N. meningitidis lactate-induced microcolony dispersal requires PilT retraction and PptB to remain dispersed. Pilus-associated factors are well known to affect clumping and aggregation of the bacteria. We next investigated the transcripts of genes related to assembly (pilE, pilD, and pilQ), elongation (pilF), retraction (pilT), and modification (pptB) of Tfp throughout the time course of lactate induction. However, no significant changes in gene expression were detected at 1, 5, 10, 20, 30, or 60 min after lactate addition (see Fig. S1 in the supplemental material).
Upregulation of pptB upon host-cell contact has previously been shown to hinder Tfp interactions that lead to microcolony dispersal (6). Although mRNA level of pptB indicates that it is not involved in lactate-induced dispersal (8), we wanted to confirm this by using a DpptB mutant. We observed that deletion of pptB in FAM20 resulted in increased microcolony formation as described previously (6; data not shown). No difference in the timing of microcolony dispersal was detected compared to the wild type (Fig. 3A). Shortly after dispersal of the DpptB mutant, we observed reformation of microcolonies, indicating that PptB might be important for bacteria to remain in a planktonic state (Fig. 3B).
Previous studies have shown that a loss of pilT in meningococci results in a hyperpiliated phenotype unable to retract the Tfp (32). Deletion of pilT also inhibits microcolony dispersal and intimate adhesion to host cells. To demonstrate that lactate-induced dispersal in liquid cultures was dependent on PilT-mediated retraction, we used a DpilT mutant in induction assays. The addition of lactate did not induce dispersal in the DpilT mutant, which remained in aggregates throughout the time-lapse experiment (Fig. 4). This suggests that PilT-mediated retraction rather than PTM, similar to phosphoglycerol modification by PptB, mediates lactate-induced dispersal in FAM20. Since Tfp retraction requires protein synthesis (33), we treated meningococci with spectinomycin 30 min before the addition of lactate. Spectinomycin treatment selectively inhibits protein synthesis, and the treatment delayed the dispersal, suggesting that protein synthesis was required (Fig. 5A). Since antibiotics frequently result in bacterial death, we performed a live-dead staining using flow cytometry. Treatment with spectinomycin did not increase bacterial death (Fig. 5B), suggesting that the phenotype was not due to a substantial proportion of the cells being dead compared to the untreated control. These results suggest that microcolony dispersal induced by lactate is dependent on PilT-mediated retraction and protein synthesis. PptB modification is not required for synchronized dispersal but is necessary for meningococci to remain dispersed.
Temperature affects N. meningitidis dispersal and reformation of microcolonies. Reduced temperature has been previously shown to affect protein expression of several virulence factors that enhance microcolony formation (27). Since the temperature in the nasopharynx is approximately 33°C 6 2°C (25), we were interested in examining the timing of dispersal at a lower temperature. We performed induction assays as previously described; bacteria were incubated at 32°C for 3 h, 10 mM lactate was added to the microcolonies, and timing of dispersal was examined. At 37°C, the synchronized dispersal phase occurred consistently at the same time, 30 min after lactate addition,  Fig. 6B (32°C L-lactate lanes), where the bacteria started to disperse after 1 h and after some time re-entered the aggregation phase. Similar results were found using 2 mM lactate (data not shown). Although bacteria formed aggregates at 32°C and responded to lactate, the dispersal was delayed and microcolonies reformed again after 2 to 3 h. These results indicate that not only the lactate concentrations but also the environmental temperature can influence microcolony dispersal in meningococci.

DISCUSSION
Understanding the mechanism behind microcolony dispersal is a major challenge in the field of meningococcal pathogenesis. Recently, we discovered that host-derived lactate induced the dispersal of meningococcal microcolonies (8). The answer to the question of how microcolonies respond to increased lactate concentration in the environment could help us to further understand disease progression. In this study, we focused on further characterization of factors contributing to lactate-induced dispersion. We showed that microcolony dispersal depended on bacterial concentration, although the QS-system AI-2 was not involved. Deletion of the response regulators of the TCS NarP/NarQ, PilR/PilS, NtrY/NtrX, and MisR/MisS had no effect on dispersal. Microcolony dispersal was dependent on pilT-mediated retraction and protein synthesis. Our results showed that the activity of PptB was not responsible for lactate-induced dispersal, but it prevented reaggregation. Finally, we suggest that environmental temperature can regulate microcolony formation and dispersal. We found that dispersal in the presence of lactate was dependent on bacterial density. Little is known about N. meningitidis QS; the bacteria are known to produce functional QS signaling molecule AI-2, and its production is dependent on the luxS gene (12). The effect of QS on meningococcal microcolony dispersal has not been explored yet. Thus far, AI-2 is the only characterized QS system in Neisseria. We observed no differences in the timing of meningococcal microcolony dispersal upon the deletion of luxS. This finding indicates that AI-2-mediated QS does not affect lactate-induced microcolony dispersal. However, QS is involved in biofilm dispersal in several other pathogens, including Pseudomonas aeruginosa, Staphylococcus aureus, and Vibrio cholerae (reviewed in reference 34).
TCS are commonly used by bacterial pathogens to sense metabolite levels and other environmental stimuli such as QS molecules, temperature, pH, antimicrobials,  and oxygen levels (35). Since TCS enable bacteria to sense and respond to extracellular signals, we sought to determine whether they might play a role in the lactate-induced dispersal. Lactate has previously been demonstrated to affect the TCS activity. In Escherichia coli, D-lactate can boost activated ArcB QS-regulator (36,37). Lactate can also activate the three-component system LrbS/LrbA/LrbR in Shewanella putrefaciens. The LrbS/LrbA/LrbR system contains two response regulators that negatively regulate biofilm formation upon activation (38). In addition, pyruvate, the downstream metabolite of lactate, activates three TCS in S. aureus, thereby enhancing its virulence (39). However, inactivation of the meningococcal TCS NarP/NarQ, PilR/PilS, and NtrY/NtrX of MisR/MisS did not prevent the lactate-induced dispersal of microcolonies.
misR and misS deletion mutants have previously been successfully constructed in N. meningitidis serogroups A, B, and C, although not in strain FAM20. Deletion of the genes resulted in phenotypes that were severely growth deficient and sometimes unable to grow at low magnesium levels (15,18,31,40,41). We were not able to generate a misR deletion mutant in FAM20, and differences between strains might be the reason for this. Expression of misR and misS has been shown to be upregulated upon contact with host cells, and the TCS is proposed to be involved in adaptation to growth on cells (18).
When we examined the DntrX mutant in FAM20 by live-cell microscopy, we observed an unusual phenotype suggesting that the mutant was deficient in bacterial cell division. Deletion mutant of ntrX, encoding the regulator of the TCS NtrY/NtrX in Neisseria gonorrhoeae has been described previously. Expression of genes involved in aerobic and anaerobic respiration was downregulated, which resulted in defects in both biofilm formation and survival within human epithelial cells. However, the ntrX mutant was not growth deficient under both aerobic and microaerobic conditions (22). It is not known whether the ntrX mutation affects respiration in meningococci and whether this results in the observed deficiency in cell proliferation.
It is known that the absence of PilF ATPase activity and the presence of PilT ATPase activity in N. meningitidis leads to a rapid disappearance of pili from the bacterial surface and the dispersal of aggregates (42). Therefore, we examined whether the lactateinduced microcolony dispersal was a result of alterations in Tfp biogenesis. However, there were no detectable differences in the transcripts of Tfp-related genes during the process of meningococcal microcolony dispersal.
PptB modifies PilE posttranslationally by conjugating a phosphoglycerol moiety to the serine residue 93, which has a negative impact on pilus-pilus interactions and facilitates the detachment of bacteria from microcolonies (6). We observed that the DpptB mutant responded to L-lactate the same way as the wild type, showing that phosphoglycerol modification of PilE was not required for the induction of rapid and synchronized dispersal. We noted, however, that a loss of PptB activity resulted in subsequent reformation of microcolonies upon dispersal. The expression of pptB has been reported to be induced upon contact with host cells controlled by the CREN regulon (17). However, no increase in pptB expression was observed when microcolony dispersal was induced by lactate in liquid cultures (8). The PTM patterns can vary greatly between the different classes (6,43). The expression level of pptB did not change over time but remained stable up to 60 min after lactate induction. It needs to be taken into account that the serogroup C strain 8013 used by Chamot-Rooke et al. expresses class I Tfp, whereas the FAM20 used in this study expresses class II Tfp.
We did not detect any detachment of bacteria from the microcolonies formed by DpilT mutant throughout the time-lapse (Fig. 4). This suggests that the process is dependent on Tfp retraction rather than PTM that block Tfp bundle formation. Since the mRNA and protein level of PilE did not increase upon lactate-induced dispersal (8), it also indicates that the retraction of the Tfp is needed, not the loss of piliation.
The filamentous prophage MDAU, expressed by meningococcal disease isolates, plays a role in colonization of the nasopharynx. MDAU structure resembles the Tfp and is also assembled through the PilQ porin (44,45). Billie et al. have shown that bacteria in direct contact with epithelial cells mainly express Tfp mediating adhesion, while bacteria not in direct contact express the MDAU form bundles and mediate bacterial interaction (45). The strain used in our experiments, FAM20, is known to possess the MDA island encoding the filamentous prophage (44). The role of MDAU in microcolony dispersal has not been established. However, since lactate-induced dispersal has been reported in N. gonorrhoeae, which does not contain the genetic island (46), MDAU is unlikely to play a role.
It is becoming more evident that meningococcal virulence factors are thermally regulated. Our results suggest that microcolony dispersal is under thermal control. Although dispersal occurred at 32°C, the timing was not synchronized, and we observed aggregate formation shortly after the dispersal phase ended. Lappann et al. have shown that the protein levels of surface factors contributing to a carrier state are induced at 32°C, resembling the environmental temperature in the upper respiratory tract. Major changes were observed in levels of neisserial heparin binding antigen (NhbA), NMB1030, and adhesion complex protein (ACP), involved in meningococcal aggregation, adhesion, and biofilm formation (27). It is tempting to hypothesize that upregulation of these proteins might play a role in the subsequent reformation of the microcolonies upon dispersal.
Collectively, our results suggest a hypothetical model (Fig. 7). Meningococcal microcolony dispersal is dependent on the bacterial concentration but not on the QS system AI-2 and the TCS. N. meningitidis depends on PilT-mediated retraction and protein synthesis for lactate-induced microcolony dispersal. To remain in a dispersed state, meningococci require the PptB activity and an optimal environmental temperature. These findings contribute to the understanding of the complex actions of N. meningitidis during microcolony dispersal. Further research may provide new targets for disease prevention.

MATERIALS AND METHODS
Bacterial strains and growth conditions. The Neisseria meningitidis serogroup C strain FAM20 and its isogenic pilT mutant have been described previously (47). The misR (also called phoP) deletion mutant has been described (31). Prior to experiments, the strains were grown for 16 to 18 h on GC agar supplemented with 1% Kellogg medium at 37°C in 5% CO 2 . Dulbecco modified Eagle medium (DMEM; Thermo Fisher FIG 7 Schematic model of the proposed sequence of events during lactate-induced microcolony dispersal. For meningococci to detach from microcolonies in the presence of lactate, PilT-mediated retraction, protein synthesis, and a certain bacterial density are required. To remain in planktonic phase, meningococci depend on PTM by PptB and optimal environmental temperature. At lower temperatures, meningococci reform microcolonies. Finally, the data suggest that AI-2, PilR/PilS, NarP/NarQ, NtrX/NtrY, and MisR/MisS are not involved. Scientific) containing 1% heat-inactivated fetal bovine serum (FBS; Sigma-Aldrich) was used in all experiments. Kanamycin was used at 50mg/ml for the selection of mutants.
Construction of mutant strains. All primers used for cloning are listed in Table 1. To construct deletion mutants, fusion PCRs were performed in two steps. First, upstream (UHS, primer pair number 1) and downstream (DHS, primer pair number 2) sequences of the genes of interest were amplified from chromosomal DNA. The DNA uptake sequence was incorporated into the primers if not present in the upstream and downstream sequences (Table 1). Antibiotic resistance cassettes were amplified by using primer pair number 3. Primers contained overlapping sequences; therefore, a fusion product of UHS_Resistance-cassette_DHS could be created from PCR products using separate fragments as the templates. The resulting product from the fusion PCR could be inserted into the genome of N. meningitidis FAM20 using homologous allelic replacement with spot transformation. High-Fidelity Phusion DNA polymerase (Thermo Fisher Scientific) was used for all PCRs. The FAM20 DpptB mutant was created using homologous allelic replacement after spot transformation with DpptB mutant chromosomal DNA (6). Mutants were sequenced (MWG Eurofins) to confirm the correct location and sequence.
Live-cell imaging. N. meningitidis FAM20 and its isogenic mutants were resuspended in prewarmed DMEM containing 1% FBS, filtered through 5-mm filters to break existing aggregates, and diluted to preferred bacterial concentrations (2 Â 10 7 , 1 Â 10 7 , 8 Â 10 6 , 6 Â 10 6 , 4 Â 10 6 , and 2 Â 10 6 ) in the presence of 2 mM L-lactate (L7022; Sigma-Aldrich). Microcolony dispersal was observed using time-lapse microscopy (Axiovert Z1; Zeiss). Three positions were chosen per well, and images were acquired every 10 min for 6 h using a 40Â objective. The images were independently evaluated by visual inspection. The aggregation phase was defined as the time from the start of incubation until dispersal of bacteria starts (i.e., when the microcolonies reached maximal size and started to decrease size). The dispersal phase was defined as the time from initiation of bacterial dispersal until microcolonies were no longer visible in the images. The planktonic phase was described as dispersed single bacteria. For wild-type strain and 5 mM lactate, the dispersal was rapid and obvious and occurred from one image to another, i.e., within a 10-min time frame. Induction assay. The induction assays have been described previously (8). Briefly, bacteria were resuspended in prewarmed DMEM containing 1% FBS, filtered through 5-mm filters, and diluted to 10 7 CFU/ml. The bacterial solutions were transferred to 24-well glass-bottom plates (MatTek) and allowed to grow and form microcolonies for 3 h at 32°C or 37°C in a 5% CO 2 environment in a live-cell observer (Axiovert Z1; Zeiss). Prewarmed DMEM supplemented with sodium L-lactate (L7022; Sigma-Aldrich) was added to microcolonies at a 1:1 ratio at final concentrations of 10 mM. During treatment with spectinomycin, the antibiotic was added at a final concentration of 100 mg/ml 30 min before the induction. Three positions were chosen per well, and images were acquired every 5 to 10 min for 4 or 5 h using a 40Â objective. The aggregation phase is described as the start of incubation until the dispersal of bacteria starts. The dispersal phase is described as the initiation of bacterial dispersal until no aggregates are observed. The planktonic phase is described as dispersed single bacteria.
Live-dead staining. Bacteria were treated with 100 mg/ml of spectinomycin 30 min before induction with L-lactate. At 60 min postinduction, bacteria were stained with a Live/Dead BacLight bacterial viability kit (L7012; Invitrogen) according to the manufacturer's protocol. The samples were analyzed by flow cytometry using an LSRFortessa (BD Biosciences) with a PI channel for dead cells and FITC for live cells. For each sample, 15,000 bacterial cells were counted, and data are presented as the percentages of dead or alive bacteria within this population. Data were analyzed with the FlowJo software.
Quantitative real-time PCR analysis. All of the primers used for quantitative real-time PCR analysis are listed in Table 1. Induction assays with L-lactate and control DMEM were performed on FAM20 wild type as indicated above in 1-ml volumes in 2.0-ml Eppendorf tubes. Samples were collected at 1, 5, 10, 20, 30, and 60 min immediately after the addition of L-lactate or DMEM and treated with RNAprotect bacterial reagent (Qiagen). Total RNA was purified using an RNeasy plus minikit (Qiagen) according to the manufacturer's protocol. Reverse transcription was performed with random hexamers using a SuperScript VILO Master Mix (Thermo Fisher). The resulting cDNA was amplified using LightCycler 480 SYBR green I Master mix (Roche) in a LightCycler 480 real-time PCR system. The PCR program was adapted from the manufacturer using an annealing temperature of 55°C. The 30S ribosomal protein rpsJ was used as a reference gene. The target mRNA levels in the samples were normalized to the reference gene and then compared to the value of the noninduced DMEM control sample.
Statistical analysis. For statistical analysis in induction assays, the differences between the lengths of the dispersal phase were analyzed. In Excel, two-tailed and unpaired Student t tests were used to compare differences between two groups. In GraphPad Prism (version 5), analysis of variance with Bonferroni's post hoc test was used to compare differences between multiple groups. P values below 0.05 were accepted as statistically significant.

SUPPLEMENTAL MATERIAL
Supplemental material is available online only. SUPPLEMENTAL FILE 1, PDF file, 0.3 MB.