The AmiC/NlpD Pathway Dominates Peptidoglycan Breakdown in Neisseria meningitidis and Affects Cell Separation, NOD1 Agonist Production, and Infection

ABSTRACT

The human-restricted pathogen Neisseria meningitidis, which is best known for causing invasive meningococcal disease, has a nonpathogenic lifestyle as an asymptomatic colonizer of the human naso- and oropharyngeal space. N. meningitidis releases small peptidoglycan (PG) fragments during growth. It was demonstrated previously that N. meningitidis releases low levels of tripeptide PG monomer, which is an inflammatory molecule recognized by the human intracellular innate immune receptor NOD1. In the present study, we demonstrated that N. meningitidis released more PG-derived peptides than PG monomers. Using a reporter cell line overexpressing human NOD1, we showed that N. meningitidis activates NOD1 using PG-derived peptides. The generation of such peptides required the presence of the periplasmic N-acetylmuramyl-l-alanine amidase AmiC and the outer membrane lipoprotein NlpD. AmiC and NlpD were found to function in cell separation, and mutation of either amiC or nlpD resulted in large clumps of unseparated N. meningitidis cells instead of the characteristic diplococci. Using stochastic optical reconstruction microscopy, we demonstrated that FLAG epitope-tagged NlpD localized to the septum, while similarly tagged AmiC was found at the septum in some diplococci but was distributed around the cell in most cases. In a human whole-blood infection assay, an nlpD mutant was severely attenuated and showed particular sensitivity to complement. Thus, in N. meningitidis, the cell separation proteins AmiC and NlpD are necessary for NOD1 stimulation and survival during infection of human blood.

INTRODUCTION

Neisseria meningitidis, also known as the meningococcus, is an obligate colonizer of the human nasopharyngeal space. Meningococci occasionally disseminate to cause invasive disease such as meningitis and septicemia. Invasive disease has a fatality rate of around 11% even with early treatment, and 11 to 19% of survivors develop lifelong sequelae (1–3). N. meningitidis colonizes up to 25% of the population at any one time (3). The carriage rate of N. meningitidis can be as high as 70% in areas of high density such as military barracks and college dormitories and during the Hajj pilgrimage (3, 4); prolonged exposure to carriers increases the risk of invasive meningococcal disease for noncarriers (4). N. meningitidis encodes multiple virulence factors that allow it to invade the bloodstream and cross the blood-brain barrier, cause a large inflammatory response, and evade clearance by the immune system (5, 6). Much of the tissue damage sustained during meningococcal disease is a result of a strong inflammatory immune response (3, 6, 7). One class of immunostimulatory molecules made and released by N. meningitidis is diaminopimelic acid (DAP)-type peptidoglycan (PG) fragments (8).

PG is a structural macromolecule that confers bacterial cell shape and protects the cell against turgor pressure. It is made of repeating subunits of N-acetylglucosamine (GlcNAc) and N-acetylmuramic acid (MurNAc), with peptide chains attached to the N-acetylmuramic acid moiety (9). The third amino acid on this peptide chain is DAP in most Gram-negative bacteria and Gram-positive bacilli and lysine in most other Gram-positive bacteria (10). The breakdown of existing PG strands and the incorporation of newly synthesized PG strands allow cell enlargement and cell separation (11), and this process generates small PG fragments as the bacteria grow (12, 13). Most Gram-negative bacteria release very small amounts of PG fragments; instead, PG fragments are typically contained in the periplasmic space and efficiently taken back into the cytoplasm by a PG fragment permease in order to be reused in cellular processes (13, 14). A limited group of Gram-negative bacteria, including but not limited to N. meningitidis; the closely related pathogen Neisseria gonorrhoeae; the etiological agent of whooping cough, Bordetella pertussis; and the squid symbiont Vibrio fischeri, release cytotoxic PG monomers (GlcNAc-anhydroMurNAc-peptide) with no overt effects on bacterial growth (8, 12, 15, 16). PG monomers released from N. gonorrhoeae cause ciliated cell death in human Fallopian tubes (15, 17). In addition to releasing PG monomers, N. meningitidis also releases PG dimers (two glycosidically linked or peptide-linked monomers) and PG-derived sugars such as GlcNAc-anhydroMurNAc disaccharide and anhydroMurNAc (8). The release of PG-derived peptides by N. meningitidis has not been characterized.

Host organisms such as humans have evolved strategies to detect and respond to microbe-associated molecular patterns to either prevent or resolve an infection. One such intracellular innate immune receptor is human NOD1 (hNOD1). hNOD1 induces a proinflammatory response upon the recognition of PG fragments with a terminal DAP, which includes tripeptide PG monomers (GlcNAc-anhydroMurNAc-l-Ala-d-Glu-mesoDAP) and PG-derived tripeptide (l-Ala-d-Glu-mesoDAP) (18–21). hNOD1 does not respond well to amidated PG fragments; some species of Gram-positive bacteria that have DAP-type PG, such as Bacillus subtilis, amidate the DAP moiety of their PG stems (19, 20, 22). Thus, hNOD1 predominantly detects the presence of Gram-negative bacteria, including N. meningitidis. In this study, we characterized PG-derived peptides released by N. meningitidis and found that meningococci release immunologically relevant amounts of hNOD1 activating PG-derived peptides.

RESULTS

The Ltg and AmiC pathways of sacculus degradation.

Previous studies of PG fragment breakdown and release by N. meningitidis, N. gonorrhoeae, and commensal species indicate that PG can be broken down in two different ways (Fig. 1) (15). In the lytic transglycosylase-dominated pathway, LtgA removes peptide-cross-linked dimers from the sacculus, and an endolytic lytic transglycosylase, likely LtgE, creates glycosidically linked dimers (23, 24). After DacB (penicillin binding protein 3 [PBP3]) cuts cross-links between strands, LtgA and LtgD degrade strands to PG monomers (8, 25). LdcA acts on the tetrapeptide PG monomers to convert them to tripeptide PG monomers, thus making them NOD1 agonists (26, 27). For the AmiC-dependent pathway, strand separation by DacB is followed by stripping of the peptides from the strands by AmiC, with NlpD activation (28, 29). AmiC cannot remove the final peptide (28, 30). LtgC then degrades the nearly naked glycan strand, creating free disaccharides and tetrasaccharide-peptide (31, 32). NagZ breaks down the disaccharides into monosaccharides (33), and NlpC is predicted to break down tetrapeptides (34). Although both pathways are active in all Neisseria species studied, the lytic transglycosylase pathway is slightly favored in N. gonorrhoeae. PG monomers are the most abundant PG fragments released by gonococci, but free tri- and tetrapeptides are almost as abundant as PG monomers (28, 35). In N. meningitidis, experiments monitoring [³H]glucosamine-labeled PG suggest that the AmiC pathway is favored. Free sugars and tetrasaccharide-peptide make up a larger proportion of the PG fragments released by N. meningitidis than do PG monomers and dimers (8). If N. meningitidis extensively uses the AmiC pathway, then this species may make significant amounts of a NOD1 agonist in the form of free tripeptides.

FIG 1.

Peptidoglycan degradation pathways in Neisseria species. During growth, portions of the sacculus are degraded to allow cell wall enlargement or remodeling. (Top) In the lytic transglycosylase-dominated pathway, PG dimers are liberated by LtgA and LtgE, strands are degraded to monomers by LtgA and LtgD, and a majority of PG monomers have their peptides shortened to 3 amino acids by LdcA. (Bottom) In the amidase-dominated pathway, peptides are removed from PG strands by AmiC, which is activated by NlpD. Free peptides may be further degraded to dipeptides by one of the three NlpC enzymes. LtgC degrades the denuded strand to free disaccharides and tetrasaccharide-peptide. NagZ, a cytosolic enzyme, cleaves the disaccharides into monosaccharides. Molecules shown with boxes around them are released from the bacteria into the milieu.

We examined PG fragment release in N. meningitidis using pulse-chase metabolic labeling with [2,6-³H]DAP. PG fragments released into the medium during growth were analyzed by size exclusion chromatography (Fig. 2A). Five major peaks were observed, and these contain PG dimers, tetrasaccharide-peptide, PG monomers, free tri- and tetrapeptides, and free dipeptides (8, 35). The amounts of PG in the free tri- and tetrapeptide and dipeptide fractions were much larger than those in the PG dimer and monomer fractions. These data indicate that N. meningitidis extensively degrades its PG to free peptides and sugars prior to release and that it favors the AmiC-dependent pathway for PG breakdown.

FIG 2.

Peptidoglycan fragment release from growing N. meningitidis cells. (A) [2,6-³H]DAP-labeled PG fragments released by WT N. meningitidis (ATCC 13102 cap) were separated by size exclusion chromatography, resolving into five distinct peaks in the included volume. Cartoons above each peak show the structures of the PG fragments, representing, in order from the left, dimers, tetrasaccharide-peptide, monomers, free tetra- and tripeptides, and free dipeptides. The structures of the dipeptides are not yet chemically confirmed. (B) PG fragment release from an amiC mutant (KL1065) and its complemented strain (EC1020). (C) PG fragment release from an nlpD mutant (KL1072) and its complemented strain (EC1026). Mutation of amiC (B) or nlpD (C) abolished the release of PG-derived peptides. Genetic complementation of the amiC (B) and nlpD (C) mutations partially restored the release of PG-derived peptides. The results depicted are representative of data from three experiments.

Released tri- and tetrapeptides from PG have been characterized in N. gonorrhoeae. Sinha and Rosenthal demonstrated that the peptides were composed of Ala-Glu-DAP and Ala-Glu-DAP-Ala and that they were not cross-linked (35). Lenz et al. found that an N. gonorrhoeae mutant defective for amiC did not release free tri- and tetrapeptides (28). The second PG-derived peptide peak (dipeptides) has been noted previously but not thoroughly investigated, partly because it overlaps somewhat the peak for free single amino acids not incorporated in the labeling process (Fig. 2A) (12, 36). Escherichia coli is known to release the disaccharide DAP-Ala (37). Therefore, we tested whether metabolic labeling with [³H]d-Ala would label the released dipeptide. The incorporation of [³H]d-Ala in this peak was similar to the incorporation of [³H]DAP, indicating that for N. meningitidis, as for E. coli, the released dipeptide fraction may contain DAP-d-Ala.

Mutation of amiC or nlpD abolishes peptide release.

N. meningitidis encodes one periplasmic N-acetylmuramyl-l-alanine amidase, AmiC (38). Meningococcal AmiC is predicted to be a 417-amino-acid zinc-dependent metalloprotease with a Tat signal sequence, an N-terminal Amidase N-terminal (AMIN) domain that binds PG, and a C-terminal catalytic domain that cleaves peptide stems. AmiC homologues in E. coli and N. gonorrhoeae function as cell separation amidases (30, 39). To examine the role of meningococcal AmiC in PG fragment production and release, we generated an in-frame markerless deletion of amiC. Mutation of amiC abolished PG peptide release and resulted in increased PG multimer release (Fig. 2B). If AmiC were able to cleave PG monomers, we would expect to see a proportional increase in PG monomer release with the decrease in peptide release by the amiC mutant. However, there is no significant difference in PG monomer release between the wild type (WT) and the amiC mutant. Complementation of amiC from an ectopic locus restored almost WT distributions of released PG fragments. Our results suggest that meningococcal AmiC liberates peptide stems from the sacculi, and it is necessary for the release of both free tri- and tetrapeptides and free dipeptide.

NlpD is a putative AmiC activator protein and functions to potentiate the amidase activity of AmiC in N. gonorrhoeae and E. coli (28, 29). Meningococcal NlpD is predicted to be a 415-amino-acid outer membrane lipoprotein with two LysM PG binding domains and a degenerate LytM/M23 peptidase domain. Like NlpD homologues from N. gonorrhoeae and E. coli, only 2/4 of the active-site residues in the M23 peptidase domain of meningococcal NlpD are conserved (29, 40). Analyses of N. gonorrhoeae and E. coli indicate that their NlpD homologues do not have enzymatic function but instead act as activator proteins for the amidases (28, 29, 40). Mutation of nlpD in N. meningitidis abolished free peptide release without affecting the release of PG monomers and dimers (Fig. 2C). Complementation of nlpD at an ectopic site partially rescued peptide release. We conclude that AmiC and NlpD are indispensable for PG-derived peptide release in N. meningitidis and that AmiC activity is dependent on activation by NlpD.

PG-derived peptides released by N. meningitidis stimulate NOD1 activation.

HEK293 reporter cell lines overexpressing NOD receptors were used to determine if peptide release by meningococci is necessary for the induction of a NOD-dependent response. In this assay, NOD activation is measured by determining the amount of NF-κB-controlled secreted alkaline phosphatase (SEAP) detected in the medium via a colorimetric assay. HEK293 cells overexpressing human NOD1 were treated with the cell-free supernatant from the WT, the amiC mutant, the nlpD mutant, or induced cultures of the respective amiC or nlpD complementation strains (Fig. 3A). As described previously and demonstrated here, the supernatant from N. meningitidis engenders a NOD1-dependent response (Fig. 3) (8). We found that the supernatant from WT N. meningitidis induced 3.6- or 7.5-times-higher levels of human NOD1 activation than an amiC or an nlpD mutant, respectively (Fig. 3A). The supernatant from the amiC and nlpD complementation strains showed partial phenotypic complementation, with 2- and 3-times-higher NOD1 responses than those in the supernatant from the mutant strains, respectively (P < 0.05) (Fig. 3A).

FIG 3.

Stimulation of human NOD1 or human NOD2 by conditioned media from N. meningitidis mutants that do not release PG-derived peptides. Supernatants from WT (ATCC 13102 cap), amiC mutant (KL1065), amiC complementation (EC1020), nlpD mutant (KL1072), and nlpD complementation (EC1026) strains were harvested, normalized to total cellular protein, and used to treat HEK293 reporter cell lines overexpressing hNOD1 (A) or hNOD2 (B) as well as parental HEK293 cell lines. hNOD activation was determined by measuring the amount of secreted alkaline phosphatase (SEAP) in the media. The values presented are the SEAP levels measured as the A₆₅₀ for the hNOD-overexpressing strain minus those of the parental strain lacking hNOD overexpression, i.e., A₆₅₀ (hNOD1) − A₆₅₀ (NULL1) or A₆₅₀ (hNOD2) − A₆₅₀ (NULL2). Error bars represent the standard deviations. Results are from four independent experiments, with technical triplicates for each biological replicate. Statistical significance was determined by Student’s t test, in which * indicates a P value of <0.05.

We also treated a HEK293 cell line overexpressing human NOD2 in parallel with the same supernatant samples as the ones used for the NOD1 experiments (Fig. 3B). The intracellular innate immune receptor NOD2 recognizes muramyl dipeptide (MDP) and responds to reducing-end PG monomers, and thus, it would not be expected to be differentially affected by the amounts of free peptides (41–43). As predicted, treatment with the supernatant from amiC or nlpD mutants did not alter NOD2 activation levels. Thus, we conclude that PG-derived peptides generated by the action of AmiC and NlpD in N. meningitidis induce a NOD1 response but do not activate NOD2 in a human epithelial cell line.

AmiC and NlpD are required for cell separation.

Periplasmic N-acetylmuramyl-l-alanine amidases that act on the sacculi cleave septal PG to allow proper cell separation (30, 44). To examine the role of meningococcal AmiC and NlpD in cell separation, we performed thin-section transmission electron microscopy (TEM) to visualize the cell morphology of the WT, mutant, and complementation strains. We counted the numbers of cells presenting as monococci, diplococci, tetrads, and clusters of 5 or more cells in 16 to 24 fields/strain. Each field used for quantification had an average of 30 cells; an example of a field for each strain used for quantification is shown in Fig. 4A. A higher-magnification image demonstrating the size of the multicell clumps of the amiC and nlpD mutants is also included.

FIG 4.

Cell separation defects in amiC or nlpD mutants. (A and B) WT, amiC mutant and complementation strains (KL1065 and EC1020), and nlpD mutant and complementation strains (KL1072 and EC1026) (A) as well as strains expressing amiC::FLAG3⁺ (EC1032) and nlpD::FLAG3⁺ (EC1033) (B) were imaged by thin-section transmission electron microscopy. (C) Quantification of the number of cells presenting as monococci, diplococci, tetrads, and clumps of 5 or more cells in 16 to 24 fields of view containing an average of 30 cells each was performed and is presented as a percentage of total cells.

WT cells predominantly presented as diplococci (60%) and monococci (26%), with a few cells being found as tetrads (13%) and clusters of 5 or more cells (1%) (Fig. 4A and C). In contrast, both the amiC and nlpD mutants formed large aggregates of 5 cells or more (94% and 76%, respectively). Mutation of amiC caused a more severe cell separation defect than did mutation of nlpD, as more cells were found in large clusters and fewer cells were found as diplococci or monococci (Fig. 4A and C). Only 3% each of amiC mutant cells presented as diplococci and monococci, while 10% and 4% of nlpD mutants presented as diplococci and monococci, respectively (Fig. 4C). Both the amiC and nlpD complementation strains showed WT-like cell morphology, with most cells presenting as diplococci (57% and 55%, respectively) and monococci (37% and 35%, respectively) (Fig. 4C). We conclude that AmiC and NlpD are required for proper cell separation in N. meningitidis.

C-terminally FLAG3-tagged AmiC displays both septal and distributed localizations, while tagged NlpD is septally localized.

We generated C-terminally triple-FLAG-epitope (FLAG3)-tagged versions of AmiC and NlpD to visualize the localization of these proteins in the cell via stochastic optical reconstruction microscopy (STORM). To determine the functionality of these tagged proteins, we performed TEM to look at the cell morphology of the amiC::FLAG3⁺- and nlpD::FLAG3⁺-expressing strains (Fig. 4B). The nlpD::FLAG3⁺ strain predominantly presented as diplococci (46%) and monococci (24%), although it had more clusters of 5 or more cells than the WT (17% versus 1%). In contrast, amiC::FLAG3⁺ cells were more likely to be found in clumps of 5 or more cells (50%), with only 22% and 18% of cells presenting as diplococci and monococci (Fig. 4C). However, the cell clumps in an amiC::FLAG3⁺ mutant were much smaller and contained fewer cells than the large cell clusters formed by an amiC or an nlpD mutant, suggesting that AmiC::FLAG3 is partially functional for cell separation.

Localizations of AmiC::FLAG3 and NlpD::FLAG3 in N. meningitidis were determined using STORM, and chromosomal DNA was counterstained with 4′,6-diamidino-2-phenylindole (DAPI) and visualized by epifluorescence microscopy to determine the position of the cells. The presence of a signal band between two adjacent DAPI foci (cells) was interpreted as septal localization. We observed two types of localization patterns for AmiC::FLAG3 (Fig. 5). First, AmiC::FLAG3 localized to the putative septum in a small subset of cells (Fig. 5, left). Second, we observed a peripheral distribution of AmiC::FLAG3 for most cells in both diplococcal and larger aggregates. NlpD::FLAG3 was clearly localized to the septum in nearly all cells, with a few cells showing some staining that might extend beyond the septum (Fig. 5, right). Since the FLAG3 tag had some effects on the function of the two proteins (Fig. 4C), more so for AmiC and less so for NlpD, we cannot rule out an effect of the tag on protein localization. However, our observations are similar to the known recruitment of the E. coli AmiC and NlpD homologues to the septum during late cell division (39, 45). Peripheral distributions of AmiC and NlpD homologues in E. coli are also seen in nondividing cells, with a minor enrichment of E. coli AmiC at the poles of such cells (39). Our findings suggest that AmiC::FLAG3 localizes to the septum in dividing meningococcal cells and may be distributed around the cell in nondividing cells.

FIG 5.

Localization of AmiC::FLAG3 and NlpD::FLAG3 in the meningococcal cell. Localizations of AmiC::FLAG3 or NlpD::FLAG3 were determined using STORM. DNA was counterstained with DAPI. The localization of AmiC::FLAG3 is presented in the left panels, and the localization of NlpD::FLAG3 is presented in the right panels. Septal localization of the proteins is seen when they form an approximate line between the two DAPI-stained cells of a diplococcus. As NlpD::FLAG3 exhibited nearly full functionality, a single field with multiple diplococci was easily found and is shown, whereas AmiC::FLAG3 was only partially functional, making it necessary to gather images of diplococci or small clumps from several fields.

Infection of human whole blood.

The severe cell separation defects of the amiC and nlpD mutants might be expected to affect infection, and differences in the release of PG fragments might affect phagocyte function. However, a significant difficulty in testing the infection ability of such mutants is the inability to accurately quantify the number of viable bacterial cells following infection since each CFU for the mutants contains multiple bacterial cells stuck together. To circumvent this difficulty, we used the nlpD mutant carrying an inducible complement construct. In the absence of the anhydrotetracycline inducer, the mutant grows as unseparated cell clusters. However, after 4 h of growth in medium containing anhydrotetracycline, a wild-type level of cell separation is achieved (Fig. 6A). Thus, we could get a reflection of CFU numbers from infection by growing the bacteria from each time point with an inducer and then dilution plating the then-separated bacteria.

FIG 6.

Survival in human whole blood. An nlpD mutant with a tightly regulated complementation construct (EC1026) was used to examine the effect of a cell separation defect on infection. (A) Schematic showing the method used for enumerating bacteria following blood infection. Approximately equal numbers of WT and mutant bacteria were used to infect human blood in vitro, although they give different CFU counts due to the nlpD mutant not separating. At each time point, the bacteria were diluted into GCBL medium containing anhydrotetracycline (aTet) to induce nlpD expression in the complemented strain, and after 4 h of growth, the bacteria were diluted and plated to determine CFU numbers. Thin-section transmission electron microscopy shows that EC1026 exhibits the cell separation defect without induction, but 4 h in the presence of anhydrotetracycline results in separation into individual diplococci and monococci. (B) The WT (ATCC 13102 cap) and nlpD mutant (EC1026) were inoculated into hirudin-treated human blood. Where indicated, CVF was used to deplete complement in the blood before inoculation. Because of the difficulties inherent in obtaining equal numbers of separating and not-separating bacteria for the initial inoculation, the data are plotted as the ratio of final CFU to initial CFU at each time point, on a log scale. Statistical significance was determined by Student’s t test, in which * indicates a P value of <0.05 compared to the WT and # indicates a P value of <0.05 compared to the WT plus CVF. These data are the geometric means ± standard errors of the means from five independent experiments.

We compared the abilities of the WT strain and the nlpD mutant to survive in whole human blood. When we attempted this experiment in an encapsulated strain background, no difference was seen between the WT and the mutant. However, in unencapsulated strains, the nlpD mutant exhibited a substantial defect. Upon the addition of the bacteria to the blood, CFU numbers for the WT strain dropped, reaching ∼50% by 2 h and then growing to 170% by 4 h (Fig. 6B). In contrast, the CFU of the nlpD mutant dropped 45-fold upon addition to blood and continued to decrease. By 4 h, the nlpD mutant CFU had dropped more than 2,000-fold. When we used cobra venom factor (CVF) to deplete complement in the blood before bacterial addition, the initial decrease in CFU numbers for the WT strain was eliminated, and the initial decrease for the nlpD mutant was reduced to only a 2-fold effect. After that point, the WT strain in the CVF-treated blood maintained CFU numbers for the first 2 h and then grew to >9-fold the starting value (Fig. 6B). The nlpD mutant in the CVF-treated blood died at the same rate as the nlpD mutant did in untreated blood albeit with CFU values 20- to 40-fold higher at each time point. A control experiment in which the WT or the nlpD mutant was incubated in diluted tissue culture medium demonstrated that the nlpD mutant is not defective in survival in nutrient-poor medium (see Fig. S1 in the supplemental material). When we incubated the WT or the nlpD mutant with human serum, the survival of the mutant was not significantly different from that of the WT and trended toward better survival for the mutant (Fig. S2). Overall, these results demonstrate that the nlpD mutant has a defect in survival in human blood and that a large part of that defect is mediated by complement but not complement alone. Even when complement is depleted, the nlpD mutant also shows a survival defect, suggesting increased killing by other soluble factors or phagocytes.

DISCUSSION

These studies demonstrated that the major PG fragments released by N. meningitidis are PG-derived peptides, with the most abundant species being the free tripeptide and tetrapeptide. The tri- and tetrapeptides are also released in large amounts by the related species N. gonorrhoeae, although N. meningitidis releases much larger amounts of the peptides than N. gonorrhoeae. The peptides are generated by the action of the periplasmic N-acetylmuramyl-l-alanine amidase AmiC, which is activated by the outer membrane lipoprotein NlpD. Mutation of either amiC or nlpD in N. meningitidis completely abolished PG-derived peptide release (Fig. 2). AmiC and NlpD were also necessary for generating the free dipeptide. The enzyme responsible for dipeptide production has not been characterized in Neisseria spp., but three NlpC family endopeptidases are encoded in the meningococcal genome, and these enzymes generally function to cleave the γ-d-Glu-DAP bond (46). The requirement for AmiC and NlpD for making the dipeptide suggests that the tetrapeptide must first be cleaved from the glycan strand by AmiC before the endopeptidase can act on it. In support of this idea, some NlpC family enzymes have been found to require a free N-terminal l-alanine for peptides to serve as the substrates for these enzymes (34).

Mutation of amiC or nlpD led to a significant reduction in NOD1 activation. This result is consistent with the loss of free tripeptide release in these mutants. It is well established that human NOD1 responds to PG fragments containing γ-d-Glu-DAP and terminating with DAP, which would include the free tripeptide released by meningococci (18–20, 22). In studies of N. gonorrhoeae, abolishing neither free peptide release nor disaccharide-tripeptide monomer release reduced NOD1 activation, indicating that the production of either the tripeptides or the disaccharide-tripeptide monomers was sufficient for stimulating NOD1 (28). Our results with N. meningitidis indicate that the free tripeptides are the major NOD1 agonists for this species and emphasize that meningococci rely on the AmiC-mediated pathway for most PG degradation (Fig. 1 and 2). Meningococci release lower levels of tripeptide PG monomer than gonococci, and the meningococcal supernatant induces lower NOD1 activation in HEK293 cells and interleukin-8 (IL-8) production by Fallopian tube explants than the gonococcal supernatant (8). The increased release of PG monomers by gonococci is partly due to a somewhat defective version of AmpG, the PG fragment permease, leading to a decreased recycling efficiency for anhydro-disaccharide-containing PG fragments. However, ampG is not fully responsible for the differences in fragment release between gonococci and meningococci since the replacement of meningococcal ampG with that of gonococci only somewhat increased PG monomer release, and this meningococcal strain continued to show the more extensive degradation of PG fragments to free disaccharides and anhydro-muramic acid (36). Taken together with these previously reported observations, our results provide support to the hypothesis that N. meningitidis degrades PG fragments more extensively than N. gonorrhoeae and thereby reduces NOD1 activation, which may reduce immune clearance during asymptomatic colonization.

It is not currently clear how PG fragment release contributes to the meningococcal lifestyle. In response to a NOD1 agonist or other inflammatory molecules, oral epithelial cells alter gene expression without inducing an inflammatory response (47). Still, it is not particularly surprising that PG fragments released by meningococci do not induce inflammation in nasopharyngeal tissue since asymptomatic colonization is by definition not overtly damaging to the host on the organismal level. Released PG fragments may play a role in establishing colonization, invading the meninges, exacerbating damage during invasive meningococcal disease, or colonizing new niches such as the urethra (11, 48).

Consistent with the role of AmiC as a cell separation amidase, mutation of amiC or nlpD led to cell separation defects, which manifested as large cell clusters containing more than 10 cells (Fig. 4). STORM analyses revealed that in a small subset of cells, AmiC::FLAG3 formed a banding pattern at the septum, suggesting a septal localization of this protein. In a larger subset of cells, AmiC::FLAG3 was found randomly distributed around the cell. The NlpD::FLAG3 signal was found to localize to the septum. Such localization patterns suggest that NlpD is a septal protein and that AmiC is recruited to the septum during cell division, as observed in other species of bacteria (39, 45, 49).

The growth of the amiC mutant and the nlpD mutant as large aggregates of cells led us to test whether the cell separation deficiency might alter infection proficiency. In other systems, bacterial variants that grow as filaments have been found to have a survival advantage over normal bacilli and to avoid phagocytic killing (50). However, gonococci defective for amiC, while not showing a growth defect, were sensitive to killing by deoxycholate, suggesting membrane instability (30). Also, a meningococcal mutant lacking gna33 (ltgC) showed a cell separation defect and was unable to infect infant rats in a model of septicemia (32). We used a human whole-blood infection model and infected the blood with the WT strain or an nlpD mutant carrying an inducible nlpD expression construct. While the WT strain was able to survive and grow in blood, the nlpD mutant showed a >1,000-fold defect in survival over 4 h. The survival defect was still present, although substantially lessened, when blood depleted for complement was used for infections. Neither incubation in dilute medium nor incubation with human serum recapitulated the decreased survival phenotype. These results indicate that in addition to complement, other factors present in blood, such as phagocytes, are necessary for the increased death of the nonseparating mutant.

Overall, these studies demonstrated that N. meningitidis AmiC and NlpD are necessary for producing the free peptides released by the bacteria and that meningococci favor this pathway for cell wall degradation. AmiC and NlpD are necessary for cell separation, and a cell separation mutant is defective in survival in an ex vivo model of human blood infection.

MATERIALS AND METHODS

Bacterial strains and growth conditions.

All strains used in this study are listed in Table 1. Except where indicated, all meningococcal strains used in the experiments are unencapsulated derivatives of N. meningitidis (ATCC 13102 cap), in which the capsule biosynthesis gene siaD has been interrupted with a chloramphenicol marker (8). These strains also have a point mutation in rpsL conferring streptomycin resistance (8). N. meningitidis strains were grown on gonococcal base medium (GCB) agar (Difco) plates at 37°C with 5% CO₂ or in gonococcal base liquid medium (Difco) with 0.042% bicarbonate (51) and Kellogg’s supplements (52) (cGCBL) at 37°C with aeration. Escherichia coli strains were grown on LB agar (Difco) plates at 37°C or in LB liquid medium at 37°C. When needed, 10 μg/mL chloramphenicol, 10 μg/mL erythromycin, or 80 μg/mL kanamycin was added to N. meningitidis cultures. For E. coli, 100 μg/mL ampicillin, 25 μg/mL chloramphenicol, 500 μg/mL erythromycin, or 40 μg/mL kanamycin was added to the media when appropriate. When necessary, 0.1 mM or 1 mM isopropyl-β-d-thiogalactopyranoside (IPTG) or 2 ng/mL anhydrotetracycline was added to the media to induce the expression of a target gene.

TABLE 1.

Strains and plasmids used in this study

Strain or plasmid	Description	Reference
N. meningitidis strains
ATCC 13102 Str	ATCC 13102 rpsLK43R	56
ATCC 13102 cap	ATCC 13102 rpsLK43R siaD::cat (WT N. meningitidis)	8
KL1060	ATCC 13102 Str transformed with pKL25; ATCC 13102 Str ΔamiC::ermC rpsL	This study
KL1061	KL1060 transformed with pDG005; ATCC 13102 Str ΔamiC	This study
KL1065	KL1061 transformed with pHC10; ATCC 13102 cap ΔamiC	This study
KL1071	ATCC 13102 Str transformed with pKL38; ATCC 13102 Str ΔnlpD::kan	This study
KL1072	KL1071 transformed with pHC10; ATCC 13102 cap ΔnlpD::kan	This study
EC1018	ATCC 13102 cap transformed with pKL50; ATCC 13102 cap PIPTG-amiC	This study
EC1020	EC1018 transformed with pEC140; ATCC 13102 cap amiC::kan PIPTG-amiC	This study
EC1022	ATCC 13102 cap transformed with pKL49; ATCC 13102 cap Ptet-nlpD*-rpsL	This study
EC1024	EC1022 transformed with pKL38; ATCC 13102 cap nlpD::kan Ptet-nlpD*-rpsL	This study
EC1026	EC1024 transformed with PCR-amplified nlpD; ATCC 13102 cap ΔnlpD::kan Ptet-nlpD	This study
EC1032	ATCC 13102 cap transformed with pEC169; ATCC 13102 cap amiC::FLAG3+	This study
EC1033	ATCC 13102 cap transformed with pEC173; ATCC 13102 cap nlpD::FLAG3+-kan	This study
Plasmids
pDG005	Markerless amiC deletion in pIDN3	30
pHC10	Serogroup C siaD::cat	57
pHSS6	Cloning vector; source of aph3 gene (referred to as Kanr)	58
pKC1	Positive/negative selection plasmid for Neisseria spp.; source of rpsL	59
pKH6	Cloning vector derived from pIDN1 and pGCC6	56
pKH99	Cloning vector derived from pHSS6 and pIDN2; source of Kanr gene	56
pMR33	Complementation vector; IPTG inducible	60
pMR68	Complementation vector; anhydrotetracycline inducible	60
pMR100	Cloning vector pIDN3 with a C-terminal FLAG3 sequence	61
pKL24	amiC deletion insert from pDG005 into pKH6	This study
pKL25	ermC rpsL in pKL24; ΔamiC::ermC rpsL in pKH6	This study
pKL28	nlpD and flanking DNA in pKH6	This study
pKL29	Markerless nlpD deletion in pIDN3; ΔnlpD in pIDN3	This study
pKL38	Kanr in pKL29; ΔnlpD::kan in pIDN3	This study
pKL45	nlpD* in pMR68 (has a premature stop codon)	This study
pKL49	nlpD*-rpsL in pMR68	This study
pKL50	amiC in pMR33	This study
pEC140	Kanr from pHSS6 in pDG005; ΔamiC::kan in pIDN3	This study
pEC167	amiC 3′ flank in pMR100	This study
pEC168	nlpD 3′ flank in pMR100	This study
pEC169	Partial amiC coding sequence (amiC′) in pEC167; amiC′::FLAG3+-3′ flank in pIDN3	This study
pEC170	Partial nlpD coding sequence (nlpD′) in pEC168	This study
pEC173	Kanr in pEC170; nlpD′::FLAG3+-Kanr-3′ flank in pIDN3	This study

Strain construction.

N. meningitidis mutants were generated by spot transformation (53). Briefly, 20 μg of linearized plasmid DNA or the PCR product was spotted onto a GCB agar plate and allowed to dry. Subsequently, 5 to 10 piliated colonies were streaked over the DNA spots, and the plate was incubated overnight at 37°C with 5% CO₂. Successful transformants were either selected using antibiotics or identified by PCR screening. All transformants were confirmed by PCR and DNA sequencing. Due to the difficulty in transforming amiC and nlpD mutants, complementation strains were constructed by first transforming the WT with the complementation construct, followed by transforming the resulting strains with the deletion construct. EC1026 is the only strain generated by transformation with a PCR product. This PCR product was generated by PCR amplification of nlpD from WT N. meningitidis chromosomal DNA using primers NlpD DUS F and NlpD DUS R (Table 2), which resulted in a DNA fragment containing the nlpD coding sequence flanked by DNA uptake sequences (DUSs) to facilitate uptake and recombination (54). Transformation with this PCR product corrected the nlpD sequence at the complementation locus to the WT sequence.

TABLE 2.

Primers used in this study

Primer	Sequencea
SacII nlpD F	TAACCGCGGAGGCACTGTGGACATTGTTG
SacI nlpD R	GATTGAGCTCTCACGATGGTTGCGGTCAAAG
BglII nlpD del F	GCCGAGATCTAGCTATATCGCGTTCTGACTTTC
BglII nlpD del R	GCCCAGATCTCATAAGATAACCTTCATGTTCCG
SacI nlpD F2	GCCGGAGCTCATCGGAACATGAAGGTTATC
SpeI nlpD R	GCCGACTAGTGATTTCAGGCAGAAAGTCAG
MluI rpsL F	AATAACGCGTCCGCTCTTGCCGACATGGTG
MluI rpsL R	AATACGCGTTAGGCGGCCGACGTGCCTAATTG
SacI MC amiC F	TAATGAGCTCCCGAACGAGGACGCGAAAGCC
SpeI MC amiC R	GCCGACTAGTACCGCCTTTTCAATCAACCC
Kan F BamHI	CAGGATCCAAAGCCAGTCCGCAGAAACG
Kan R BamHI	CTGGATCCTGGGCGAAGAACTCCAGCAT
NlpD DUS F	ATGCCGTCTGAACTTTGCGCTTGAGAAGAACG
NlpD DUS R	TCTGTACTGTCTGCGGCTTC
SacI MC amiC-sigseq F	GATTAGAGCTCAAAACGGTACGCGCACCAC
EcoRI MC amiC-stop codon R	GTGCAGAATTCACCCCGCTTCAATACGGATG
HindIII amiC down F	GCACGAAGCTTTAGTTGTCGGATGAAGGCAG
XhoI amiC down R	GACGACTCGAGCAACCATTTCAATTCCGTATCC
SacI MC nlpD-sigseq F	GATTAGAGCTCGCTACCCAGCAGCCTGCC
EcoRI MC nlpD-stop codon R	GCGCAGAATTCGAACGCGATATAGCTGTTC
HindIII nlpD down F	GCAGCAAGCTTCTGCCTGAAATCAAGTTGG
XhoI nlpD down R	GATTACTCGAGATCGGCATCCGTGTCGAAGC

^aRestriction enzyme sites are underlined.

Plasmid construction.

All plasmids used in this study are listed in Table 1 and were verified by PCR and sequencing. All primers are listed in Table 2. Plasmids were generated by the ligation of a vector backbone and an insert and transformed into Rapid Trans chemically competent TAM1 cells (Active Motif).

Two plasmids, pKL24 and pKL25, were generated to create an in-frame deletion of amiC. To make pKL24 (amiC deletion in pKH6), pDG005 and pKH6 were digested with XmaI and SacI. The amiC deletion insert from pDG005 was ligated into pKH6, generating pKL24. To generate pKL25 (ermC rpsL in pKL24; final construct, ΔamiC::ermC rpsL), the ermC rpsL positive/negative selection cassette was first excised from pKC1 by digestion with NheI and EcoRV. The insert from pKC1 was blunted with T4 polymerase and ligated into the blunted BspHI site from pKL24 to create pKL25.

Three plasmids were generated to create an nlpD insertional inactivation mutant. To create pKL28 (nlpD and flanking DNA in pKH6), PCR was used to amplify nlpD and flanking DNA with primers SacII nlpD F and SacI nlpD R. The PCR product and pKH6 were digested with SacI and SacII and then ligated. To generate pKL29 (nlpD deletion in pKH6), primers BglII nlpD del F and BglII nlpD del R were used for PCR with pKL28 as a template. The product was digested with BglII and self-ligated to make pKL29, which contains a markerless nlpD deletion with flanking DNA. To build pKL38 (ΔnlpD::kan allelic replacement in pKH6), the kanamycin resistance (Kan^r) marker from pKH99 was excised by digestion with EcoRV and Ecl136II and inserted into the blunted BglII site of pKL29.

Two plasmids, pKL45 and pKL49, were constructed to complement an nlpD mutation at an ectopic site. Plasmid pKL45 (nlpD* in pMR68) was made by first amplifying nlpD from WT N. meningitidis chromosomal DNA with primers SacI nlpD F2 and SpeI nlpD R. The PCR product and pMR68 were subsequently digested with SacI and SpeI and ligated to each other. Despite repeated attempts, the resulting plasmid contained one or more mutations in the nlpD coding sequence. We deduce that the presence of a second copy of nlpD in E. coli is toxic to the bacterium and saved one of the plasmids as pKL45. pKL45 has a nonsense mutation leading to a premature stop codon at the third codon. Plasmid pKL49 (nlpD*-rpsL in pMR68) was generated by the insertion of rpsL into pKL48. rpsL was amplified from pKC1 using primers MluI rspL F and MluI rpsL R; the PCR product and pKL45 were digested with MluI and subsequently ligated to each other to generate pKL49.

Plasmid pKL50 was generated to complement an amiC mutation at an ectopic site. To make pKL50 (amiC in pMR33), amiC was amplified from WT N. meningitidis chromosomal DNA with primers SacI MC amiC F and SpeI MC amiC R and digested with SacI and SpeI. pMR33 was digested similarly and ligated with the digested amiC PCR product.

Plasmid pEC140 was built to create a deletion/insertion mutant where kan replaced amiC. To generate pEC140 (amiC::kan in pIDN3), pDG005 was first digested with BspHI and subsequently blunted with T4 DNA polymerase. The aph3 gene conferring resistance against kanamycin (referred to as the Kan^r gene) was amplified from pHSS6 using primers kan F BamHI and kan R BamHI and blunt ligated with the blunted, BspHI-digested pDG005 to form pEC140.

To epitope tag amiC at the native locus, pEC169 (amiC′::FLAG3⁺-3′ flank in pIDN3) was constructed in multiple steps. First, the amiC 3′-flanking region was amplified using primers HindIII MC amiC down F and XhoI MC amiC down R and WT N. meningitidis chromosomal DNA as a template. pMR100 and the PCR product were digested with HindIII and XhoI and ligated to make pEC167 (amiC 3′ flank in pMR100). The partial coding sequence of amiC lacking the start codon, signal sequence, and stop codon (amiC′) was amplified from WT N. meningitidis DNA using primers SacI MC amiC-sigseq F and EcoRI MC amiC-stop codon R, digested with SacI and EcoRI, and ligated into similarly digested pEC167 to form pEC169.

To epitope tag nlpD at the native locus, pEC173 (nlpD′::FLAG3⁺-Kan^r-3′ flank in pIDN3) was generated with a strategy similar to that for pEC169. The nlpD 3′-flanking region was amplified using primers HindIII MC nlpD down F and XhoI MC nlpD down R instead, digested with HindIII and XhoI, and ligated with similarly digested pMR100 to form pEC168. The partial coding sequence of nlpD lacking the start codon, signal sequence, and stop codon (nlpD′) was amplified with primers SacI MC nlpD-sigseq F and EcoRI MC nlpD-stop codon R, digested with SacI and EcoRI, and ligated into similarly digested pEC168 to form pEC170. pEC170 was digested with HindIII, blunted with T4 polymerase, and ligated with the Kan^r gene to form pEC173.

Metabolic labeling of peptidoglycan and quantitative fragment release.

Metabolic labeling of PG and quantitative fragment release were performed as described previously (55). Strains were pulse-chase labeled with 25 μg/mL [2,6-³H]DAP in Dulbecco’s modified Eagle’s medium (DMEM) lacking cysteine and supplemented with 100 μg/mL methionine and 100 μg/mL threonine for 45 min. A sample of the culture was removed to determine the number of radioactive counts per minute (cpm) by liquid scintillation counting, and the bacteria were diluted to obtain equal cpm in each culture. After a 2-h chase period, the supernatant was harvested by centrifuging the culture at 3,000 × g for 10 min and filter sterilizing the supernatant with a 0.22-μm filter. [2,6-³H]DAP-labeled molecules in the filtered supernatant were then separated using tandem size exclusion chromatography with 0.1 M LiCl as the mobile phase and detected by liquid scintillation counting. The growth media for the amiC and nlpD complementation strains were supplemented with 0.1 mM IPTG and 2 ng/mL anhydrotetracycline, respectively.

Thin-section transmission electron microscopy.

N. meningitidis strains were grown in cGCBL until mid-log phase, when cells were harvested by centrifugation at 17,000 × g for 1 min. Growth medium for the amiC or nlpD complementation strain contained 0.1 mM IPTG or 2 ng/mL anhydrotetracycline, respectively, for gene induction. Cells were washed once with phosphate-buffered saline (PBS) and resuspended in a fixative solution (2% paraformaldehyde and 2.5% glutaraldehyde in 0.1 M phosphate buffer). Thin-section electron microscopy of fixed samples was done at the University of Wisconsin—Madison Medical School Electron Microscope Facility. Enumeration of meningococci in cell clusters was done by counting the number of complete, in-plane cells presenting as monococci, diplococci, tetrads, or clusters of five or more cells in 16 to 24 fields/strain containing an average of 30 cells/field.

Treatment of the HEK293 reporter cell line with the filtered supernatants from meningococcal cultures.

HEK293-Blue cells that encode secreted alkaline phosphatase (SEAP) under the control of an NF-κB promoter (parental cell lines NULL1 and NULL2) and hNOD1 or hNOD2 receptors (NOD and NOD2 cell lines) were grown, maintained, and treated according to the manufacturer’s instructions (InvivoGen). Supernatant samples were harvested from meningococcal cultures started at an optical density at 540 nm (OD₅₄₀) of 0.2, grown for 2 h by centrifugation at 17,000 × g for 1 min, and filter sterilized using a 0.22-μm-pore-size filter. Cultures were supplemented with 1 mM IPTG or 2 ng/mL anhydrotetracycline as needed. Supernatant samples were normalized to total cellular protein. HEK293 cells that were seeded at a density of 2.8 × 10⁵ cells/mL (NOD1/NULL1) or 1.4 × 10⁵ cells/mL (NOD2/NULL2) in 180 μL of DMEM with 4.5 g/L glucose, 2 mM l-glutamine, and 10% heat-inactivated fetal bovine serum (FBS) were treated with 20 μL of the supernatant for 18 h at 37°C with 5% CO₂. A total volume of 20 μL each of cGCBL (medium-only control), 10 μg/mL TriDAP (positive control for NOD1/NULL1) (InvivoGen), 10 μg/mL muramyl dipeptide (MDP) (positive control for NOD2/NULL2) (InvivoGen), 10 μg/mL Tri-Lys (negative control for NOD1/NULL1), or 10 μg/mL MDP-control (MDP-c) (negative control for NOD2/NULL2) (InvivoGen) was used as a treatment control. After 18 h of treatment, 20 μL of the HEK293 culture supernatant was added to 180 μL of QUANTI-Blue reagent (InvivoGen) and incubated at 37°C for 4 h. SEAP activity was determined by measuring the absorbance at 650 nm. To calculate NF-κB-dependent hNOD activation (graphed SEAP levels [A₆₅₀]), A₆₅₀ values from NULL1 or NULL2 controls were subtracted from the corresponding A₆₅₀ values from NOD1 or NOD2 treatments. The results are from four independent experiments, and statistical tests were performed using Student’s t test.

Stochastic optical reconstruction microscopy.

Sample preparation and imaging were done as described previously (23). Meningococcal cultures were grown to mid-log phase and harvested by centrifugation at 17,000 × g for 2 min. Cells were washed once with PBS and suspended in 400 μL of a fixative solution (4% methanol-free formaldehyde in PBS). The cell suspensions were incubated first at room temperature (RT) for 15 min and then on ice for 15 min. The cell suspensions were then centrifuged at 6,000 × g, and the fixative solution was discarded. Cells were quickly washed three times with cold PBS to ensure the thorough removal of the fixative solution, resuspended in cold GTE buffer (50 mM glucose, 1 mM EDTA, 20 mM Tris-HCl [pH 7.5]), and incubated on ice for 10 min. Cells were quickly washed with PBS-T (0.3% Triton X-100 in PBS), suspended in cold methanol, and incubated at −20°C for a minimum of 10 min. Methanol was removed, after which cells were washed with PBS-T for 5 min, blocked with 400 μL 5% goat serum in PBS-T for 15 min, and incubated with 1:150-diluted M2 primary antibody (Sigma) in PBS for 1 h. Cells were subsequently washed with PBS three times and incubated with 200 μL 1:100-diluted Alexa Fluor 647 (Thermo Fisher) secondary antibody, 100 ng/mL DAPI, and 5% goat serum in PBS for 1 h at RT or overnight at 4°C. Cells were washed with PBS five times to remove unbound antibody and dye suspended in 200 μL PBS. Around 2-μL volumes of the samples were spotted onto poly-l-lysine-coated coverslips, mounted with Vectashield H-1000 mounting medium (Vector Laboratories), and visualized using a Nikon N-Storm microscope. Images were analyzed using NIS-Elements AR software with N-Storm analysis. Linear brightness/contrast adjustments and false coloring were performed using FIJI.

Human blood infection assay.

Blood was collected from healthy donors using hirudin-containing vacutainer tubes (48). All human subjects gave written informed consent in accordance with a protocol approved by the University of Wisconsin—Madison Health Sciences Institutional Review Board (protocol number 2017-1179). N. meningitidis was grown overnight on a GCB agar plate and harvested into phosphate-buffered saline. The blood was inoculated with N. meningitidis strains at 1 × 10⁸ CFU/mL bacteria, and the cultures were grown with rotation at 37°C. To enumerate N. meningitidis used as the inoculum at the start of the experiment, the bacteria were inoculated into cGCBL instead of blood and dilution plated. In contrast, time zero was the time point taken immediately after the inoculation of the blood. Where indicated, blood was pretreated with 1 μg/mL cobra venom factor for 1 h prior to inoculation, with the blood culture rotating at 37°C. At each time point, 375 μL of each blood culture was removed and diluted into 1.125 mL of cGCBL medium. Anhydrotetracycline (final concentration of 2 ng/mL) was added to each culture. The diluted cultures were grown at 37°C for 4 h to allow expression from the complementation construct. After 4 h, the bacteria were diluted and plated for CFU determinations. Incubations in dilute medium were performed in the same manner as for blood except that the bacteria were suspended in DMEM diluted 1:8 in phosphate-buffered saline.

Serum resistance assay.

N. meningitidis cells were suspended in cGCBL and grown for 2 h with aeration to reach log phase. Log-phase bacteria were diluted to an OD₅₄₀ of 0.2 in DMEM. Four 1:10 dilutions of the culture were performed, and the fourth dilution was treated with 25% human serum at 37°C. At 0 and 30 min, 25 μL of each culture was plated for CFU counts. To prepare the heat-killed control serum, the serum was incubated at 56°C for 30 min.