Cationic Antimicrobial Peptide Resistance in Neisseria meningitidis

Abstract

Cationic antimicrobial peptides (CAMPs) are important components of the innate host defense system against microbial infections and microbial products. However, the human pathogen Neisseria meningitidis is intrinsically highly resistant to CAMPs, such as polymyxin B (PxB) (MIC ≥ 512 μg/ml). To ascertain the mechanisms by which meningococci resist PxB, mutants that displayed increased sensitivity (≥4-fold) to PxB were identified from a library of mariner transposon mutants generated in a meningococcal strain, NMB. Surprisingly, more than half of the initial PxB-sensitive mutants had insertions within the mtrCDE operon, which encodes proteins forming a multidrug efflux pump. Additional PxB-sensitive mariner mutants were identified from a second round of transposon mutagenesis performed in an mtr efflux pump-deficient background. Further, a mutation in lptA, the phosphoethanolamine (PEA) transferase responsible for modification of the lipid A head groups, was identified to cause the highest sensitivity to PxB. Mutations within the mtrD or lptA genes also increased meningococcal susceptibility to two structurally unrelated CAMPs, human LL-37 and protegrin-1. Consistently, PxB neutralized inflammatory responses elicited by the lptA mutant lipooligosaccharide more efficiently than those induced by wild-type lipooligosaccharide. mariner mutants with increased resistance to PxB were also identified in NMB background and found to contain insertions within the pilMNOPQ operon involved in pilin biogenesis. Taken together, these data indicated that meningococci utilize multiple mechanisms including the action of the MtrC-MtrD-MtrE efflux pump and lipid A modification as well as the type IV pilin secretion system to modulate levels of CAMP resistance. The modification of meningococcal lipid A head groups with PEA also prevents neutralization of the biological effects of endotoxin by CAMP.

Cationic antimicrobial peptides (CAMPs), which are constitutively present in macrophages and neutrophils and inducibly produced by epithelial cells at mucosal surfaces, play an important role in host defense against microbial infection and are key components of the innate immune response (25). Additionally, certain CAMPs hold effector functions that modulate expression of certain host genes important in host defense, thus providing communication between innate and adaptive immunity (2, 4, 11). Based on their structural characteristics, CAMPs are classified into different categories: β-sheets stabilized by disulfide bonds, amphipathic α-helix, loop structures with a single disulfide bond, cyclic peptides, and boat-like extended structures (25).

Because bacterial pathogens constantly encounter CAMPs during infection, they have developed several mechanisms for conferring intrinsic or inducible resistance. These resistance mechanisms include (i) membrane-bound proteases that degrade peptides (20, 54, 56); (ii) efflux pumps and transport systems that export CAMP from the periplasmic and intracellular compartment (1, 7, 42, 43, 51); (iii) modulation of outer membrane permeability through outer membrane proteins (33); and (iv) alteration of CAMP interaction with lipopolysaccharide (LPS) and lipooligosaccharide (LOS) through various structural modifications (15, 27, 39, 65).

One of the best-characterized mechanisms of resistance to CAMPs is LPS-mediated resistance to polymyxin B (PxB). PxB, a cyclic lipopeptide that binds to lipid A and is lethal to some gram-negative bacteria, has long been used as a model compound to define the mechanisms by which CAMPs kill bacteria and how bacteria develop resistance to their antimicrobial actions. It is believed that PxB primarily interacts with phosphorylated head groups of lipid A, and modification of the lipid A head groups correlates with increased bacterial resistance to PxB. LPS modifications that include alteration of the fatty acid content of lipid A, phosphoethanolamine (PEA) addition to the core and lipid A head groups, and 4-amino-4-deoxy-l-arabinose (Ara4N) addition to the core and lipid A regions have been well studied. Phosphoryl groups linked to the lipid A head groups have been implicated in the formation of a stable outer membrane network enabling adjacent LPS molecules to be cross-linked via divalent cations (19, 47). The esterification of the lipid A head groups by aminoarabinose and PEA (27, 39, 65), which presumably reduces electrostatic interaction between polymyxin and LPS, correlates with an increased resistance to CAMPs in Escherichia coli and Salmonella enterica serovar Typhimurium (14, 21, 22, 39). For example, lipid A isolated from two polymyxin-resistant mutants of E. coli was substituted with PEA at the glycosidic pyrophosphate forming a diphosphate diester linkage at this position in 40% of the total lipid A molecules (39). PmrE and gene products of a 7-gene operon, pmrHFIJKLM (also named pbgP operon [18]), that mediate biosynthesis of arabinose precursors are responsible for the addition of Ara4N to lipid A in E. coli, Salmonella spp. (21), and Pseudomonas aeruginosa (36). The inducible PxB resistance conferred by lipid A modification is under the control of dual two-component regulatory systems, PhoP/Q and PmrA/B (34, 36, 52, 59). Recently, the gene responsible for PEA substitution of the lipid A head groups has been identified in Neisseria meningitidis (10) and S. enterica serovar Typhimurium (31) as lptA and pmrC, respectively. Bacteria that are intrinsically resistant to polymyxin such as Proteus mirabilis (53), Chromobacterium violaceum (26), and Burkholderia cepacia (9) contain near complete substitution of lipid A by Ara4N.

N. meningitidis is highly resistant to the action of PxB, with a MIC that is >100-fold higher than that of Salmonella spp. (21). Genes encoding the proteins responsible for the biosynthesis and attachment of Ara4N are absent in the meningococcal genomes (41, 60). To understand the mechanisms employed by N. meningitidis for conferring intrinsic PxB resistance, we utilized a random transposon insertional library to identify genes involved in modulating PxB resistance. In this report, we demonstrated that both lipid A modification and the MtrC-MtrD-MtrE efflux pump play important roles in meningococcal resistance to PxB and other CAMPs.

MATERIALS AND METHODS

Bacterial strains and media.

Strains used in this study are listed in Table 1. Meningococcal strain NMB (L2/L4 LOS immunotype) is a serogroup B N. meningitidis strain originally isolated from the cerebrospinal fluid of a patient with meningococcal meningitis in Pennsylvania in 1982 (55). All meningococcal strains were grown on GC base agar (Difco Laboratories) supplemented with 0.4% (wt/vol) glucose and 0.68 mM Fe(NO₃)₃ at 37°C with 3.5% (vol/vol) CO₂. Meningococcal mutants exhibiting resistance to kanamycin (Kn) were grown on brain heart infusion base agar (BHI; Becton Dickinson) containing 1.25% fetal bovine serum (GIBCO BRL). Liquid cultures were grown in GC broth with the same supplements and 0.43% (wt/vol) NaHCO₃ at 37°C. E. coli strains were grown in Luria-Bertani (LB) broth (Bethesda Research Laboratories) at 37°C with appropriate antibiotic selection. Antibiotics were used in the following concentrations (in micrograms per milliliter) for meningococci: tetracycline, 5; spectinomycin (Sp), 60; and Kn, 80. Antibiotics used for E. coli were as follows: ampicillin, 100; kanamycin, 50; and spectinomycin, 100.

TABLE 1.

Strains used in this study

Strain	Genotype or description	Source or reference
N. meningitidis
NMB	Wild type, B:2bB:P1.2,5:L2 (CDC8201085)	55
M7	synA::Tn916	58
XZ134	ΔmtrD::Ω(Sp)	This study
M7295	pilM::Ω(Sp); synA::Tn916	This study
NMB285	lpxL2::aphA-3	This study
NMB291	lpxL1::aphA-3	This study
KA2035	lptA::Ω(Sp)	62
KA2033	fbp::Ω(Sp)	This study
NMB319	NMB0596::aphA-3	This study
XZ06	NMB1637::aphA-3	This study
XZdm03	mtrE::Ω(Kn); lptA::Ω(Sp)	This study
M7mtrE	mtrE::Ω(Kn); synA::Tn916	This study
M7295mtrE	pilM::Ω(Sp); mtrE::Ω(Kn); synA::Tn916	This study
XZ036	KA2013 complemented with pXZ032	This study
XZ040	KA2035 complemented with pXZ027	This study
E. coli
DH5α	Cloning host strain	48

mariner transposon mutagenesis.

A high-density transposition mutant bank of strain NMB based upon in vitro mutagenesis by the Himar1 mariner element (45) was developed utilizing the method of Pelicic et al. (45). In brief, 500 ng of the donor plasmid pMM2611-aphA-3, 120 nM purified mariner transposase, and 500 ng of strain NMB chromosomal DNA were mixed in 10% glycerol-2 mM dithiothreitol-250 μg/ml of bovine serum albumin-25 mM HEPES (pH 7.9)-100 mM NaCl-10 mM MgCl₂. After incubation at 30°C for 3 h, the transposase was inactivated by a 10-min exposure at 75°C before purification of the mutagenized DNA by a Qiaex II gel extraction kit (QIAGEN). The single-stranded gaps, introduced upon Himar1 transposition, were repaired by treatment with T4 DNA polymerase followed by overnight ligation with T4 DNA ligase. The mutagenized DNA was transformed into strain NMB, a naturally highly competent strain, and transformants were selected on BHI agar plates containing 80 μg/ml of kanamycin.

PxB susceptibility assays.

Decreased resistance of meningococci to PxB was ascertained by growth on a series of GC agar plates containing doubling amounts of PxB (8 to 256 μg/ml). Colonies were patched onto the plates from the highest to the lowest concentration, whereas mutants with increased PxB resistance were screened by replica plating onto plates containing 1,025 μg/ml of PxB. The plates were incubated for 18 h and then scored for growth of each patch. In addition, broth cultures of mid- to late-exponential phase were diluted to equal cell density, and aliquots of 2 μl of cell suspension were spotted onto PxB containing GC agar plates. Similar results were obtained by both methods.

SSP (single specific primer)-PCR.

The SSP-PCR procedure used has been described by Swartley et al. (57).

Construction of meningococcal isogenic mutants. (i) fbp.

A PCR product of 1.0 kb was generated using primer pair fbp-1 (5′-GCAAACTATGGACACACTGAC-3′) and fbp-2 (5′-GTCTATTTTGCGTGCAGGCGGT-3′) and cloned into the pCR2.1 to yield pKA303. The Ω(Sp) cassette, released from pHP45Ω (46) with SmaI, was cloned into the internal HincII site of pKA303. The resulting plasmid, pKA306, was linearized with ScaI and used to transform N. meningitidis strain NMB by allelic replacement, and spectinomycin-resistant colonies were selected on GC agar plates containing spectinomycin. Plasmid DNA was isolated using Qiaprep spin columns (QIAGEN), and all insertions within the plasmids were confirmed by sequencing. Mutations in NMB were confirmed by PCR using primer pair fbp-1/fbp-2.

(ii) lpxL1.

A PCR product was produced by primers YT103 (5′-CTGCCGTTTGCGCTGCTGC-3′) and YT104 (5′-GTACGCCATTTTCTACGCTTTGCC-3′) and cloned into pCR2.1. The insert was obtained with EcoRI digestion and subcloned into pUC18 to generate pUC-lpxL1. An aphA-3 cassette released with EcoRI and BamHI digestion from pUC18k (35) was blunted with Klenow treatment and then ligated with pUC-lpxL1 that was digested with BssHII and blunted with Klenow to yield pYT291. The orientation of the aphA-3 cassette in transformants was verified by PCR with primers YT103 and KanA (5′-CTTAGCAGGAGACATTCCTTCCG-3′), an outward primer located at the 5′ end of the aphA-3 cassette. The in-frame fusion of the cassette with the 3′ coding sequence of lpxL1 was confirmed by direct sequencing analysis of pYT291. Meningococcal mutant NMB291, with mutation in lpxL1, was obtained by transforming strain NMB with ScaI-linearized pYT291 and selected for kanamycin-resistant colonies. Colony PCR using primers YT103 and YT104 verified the desired mutation.

(iii) lpxL2.

A DNA fragment containing lpxL2 was amplified using primers YT105 (5′-TCTTCAGACGGCATGTTGTATGAAC-3′) and YT106-2 (5′-CAGATACTGCGTCGGAAAACGGCG-3′) and cloned into pCR2.1. The DNA fragment was released by HindIII-EcoRV digestion and subcloned into the vector pUC18 cut with HindIII-HincII to yield pUC-lpxL2. An aphA-3(Kn^R) cassette obtained from pUC18K with SmaI digestion was inserted into the unique BssHII site of pUC-lpxL2 to generate pYT285, and the orientation and in-frame fusion of the aphA-3 cassette was confirmed as described above. ScaI-linearized pYT285 was used to transform strain NMB to obtain kanamycin-resistant colonies. Correct mutants, named NMB285, were confirmed by colony PCR with primers YT105 and YT106-2.

(iv) NMB0596.

An internal fragment of NMB0596 was generated by PCR amplification using primers YT148 (5′-GAGCAGTTCGGGTAGCAGCGG-3′) and YT159 (5′-CGGCGTTTTGGTATCTTCAGGCAC-3′). The PCR product was cloned into pCR2.1 to yield pTA148-159. A 726-bp fragment between two HincII sites was removed and replaced with an aphA-3 cassette that was released with EcoRI and HincII digestion from pUC18k and made blunt with Klenow treatment. The resulting plasmid, pYT319, was verified to have a correct cassette orientation and in-frame fusion by colony PCR and sequencing analyses. Meningococcal strain NMB was transformed with pYT319, and kanamycin-resistant transformants were selected. One correct transformant named NMB319 that was confirmed by colony PCR was saved for future studies.

(v) lptA.

The construction of lptA::Ω(Sp) mutant has been described in Tzeng et al. (62). A lptA/mtrE double mutant, XZdm03, was made by transforming the KA2035 lptA::Ω(Sp) mutant with the chromosomal DNA isolated from a mtrE::Ω(Kn) derivative of a gonococcal strain FA19 (12). Transformants were selected on BHI plates containing spectinomycin and kanamycin and verified by colony PCR confirming the acquisition of both cassettes.

(vi) mtrD.

A DNA fragment containing the mtrD coding sequence was obtained by PCR using primers KH9#16 (5′-GTGATGATTGTGAATGCCCAGG-3′) and mtrE9 (5′-CAGGCAGACAATGCAAAGGC-3′) and subsequently cloned into pCR2.1. An ∼2-kb mtrD internal sequence between two HincII sites was then replaced with an Ω(Sp) cassette released from pHP45Ω by SmaI digestion to generate pXZ013. Spectinomycin-resistant colonies obtained from transformation of the meningococcal strain NMB with linearized pXZ013 were examined by colony PCR and Southern blots to confirm the deletion of the mtrD sequence and the presence of the Ω cassette. The mtr deletion mutant was named XZ134.

(vii) NMB1637.

A PCR product obtained with primers 1638-3 (5′-CACGACCACTATTTCAGCACG-3′) and 1637-1 (5′-CCAGACATTGTCGAATCCCTG-3′) was cloned into pCR2.1 to generate pTA1637. An SmaI-digested aphA-3 cassette was subsequently inserted into the SspI site within the insert to yield pXZ003. The meningococcal mutant with NMB1637 mutation was then obtained by transforming strain NMB with ScaI-linearized pXZ003.

(viii) pilM.

A 1,454-bp DNA fragment containing the 5′ coding sequence of pilM was amplified from strain NMB using primers YT113 (5′-TCCCCATCCGCCGAATAAATAGTC-3′) and YT114 (5′-TCAGTGTCGGCGCATCAAGTTCG-3′) and cloned into pCR2.1. The insert was then released by EcoRI digestion and subcloned into pUC18 vector. The Ω(Sp) cassette excised from pHP45Ω by SmaI digestion was subsequently inserted into the unique EcoRV site present in the pilM coding sequence to generate pYT295. Meningococcal strain NMB was transformed with ScaI-linearized pYT295, and spectinomycin-resistant transformants were selected. Correct mutants were verified by colony PCR using primers YT113 and YT114 and named NMB295.

Complementation of lptA and lpxL1 mutants.

Primers lpxL1-F (5′-TGCAGGTCAAACAGGCGGTAGT-3′) and lpxL1-R (5′-TTCATAGGTTTGCGGTATTTCTTCCA-3′) were used to amplify a 1,327-bp DNA fragment that contains sequence 262 bp upstream of the lpxL1 start codon and the entire lpxL1 coding sequence. PCR amplification with primers 1639-1 (5′-GGCGGCGTTTTGGAGGTGG-3′) and 1637-2 (5′-AAGATCGGCGGTTCGTCAATAATT-3′) yielded a 2,093-bp PCR product that encompasses the 346-bp upstream sequence and the complete lptA coding sequence. The PCR products were cloned into pCR2.1 via TA cloning (Invitrogen). The inserts were released by EcoRI digestion and made blunt ended using Klenow. These DNA fragments were subsequently cloned into the EcoRV site of the meningococcal shuttle vector pYT250 (Em^r) (63) to yield pXZ032 (lpxL1) and pXZ027 (lptA), respectively. The plasmids were methylated with HaeIII methylase (New England Biolabs) according to the manufacturer's protocol and were then used to transform the corresponding KA2013 or KA2035 mutants to yield Kn^r Em^r and Sp^r Em^r transformants, respectively. PCR analyses using chromosome-specific primers and vector-specific primers confirmed the presence of the original mutation at the chromosomal locus and the presence of an intact copy of the complemented gene.

Chromosomal DNA isolation and Southern blots.

Meningococcal chromosomal DNA was prepared according to the method of Nath (38). The Genius 2 DNA labeling and detection system (Boehringer Mannheim) was used to perform DNA hybridization. The digoxigenin-labeled probe for detecting aphA-3 was generated by random primed labeling reaction with the Km6-Km7 (45) PCR product as template. Chromosomal DNA was digested by ClaI overnight and resolved on a 0.7% Tris-acetate-EDTA agarose gel. DNA was transferred to a nylon membrane using a Turboblotter apparatus (Schleicher & Schuell). Hybridization and development of the Southern blots were performed following the manufacturer's protocol.

Disk diffusion assays of growth inhibition.

Overnight plate-grown meningococci were suspended in GC broth to an optical density at 550 nm (OD₅₅₀) of 0.1. After mixing 50 μl of the cell suspension with 5 ml of 0.5% GC agar kept at 40°C, the cell mixtures were then poured onto a plate containing 15 ml of solidified GC agar. Three filter disks (8 mm in diameter) were placed on the top of the solidified soft agar-cell mix. Five microliters of tested agents was spotted on the disk, and the zones of growth inhibition were measured after 24 h of incubation at 37°C.

Antimicrobial peptide sensitivity assays.

The procedure for testing the susceptibility of meningococci to LL-37 and protegrin-1 (PG-1) has been described previously (51). High-performance liquid chromatography-purified synthetic LL-37 and PG-1 were prepared by the Microchemical Facility of Emory University as previously described (51).

Miniscale LOS extraction and tricine SDS-PAGE analysis.

A minigel of 16% (separating gel)-10% (spacer gel)-4% (stacking gel) composition using a tricine sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) system (50) was employed to resolve crude LOS samples prepared by protease K digestion of whole-cell lysates. Briefly, a few single colonies were suspended in distilled water, and the protein concentrations were approximated by the Bradford assays (Bio-Rad) with bovine serum albumin as standard. The cell suspension was then adjusted to 1 μg/μl. A digestion mixture consisting of 8 μl cell suspension, 28 μl of 2% SDS in TE buffer (10 mM Tris, pH 8.0, 1 mM EDTA), and 8 μl of proteinase K (25 mg/ml in 20% glycerol, 500 mM Tris, pH 8.0, 10 mM CaCl₂) was incubated at 60°C overnight and quenched by adding 38 μl of loading dye solution (1 M Tris, 10% glycerol, 2% SDS, 5% β mercaptoethanol, 0.05% bromophenol blue). The samples were heated at 95°C for 5 min before loading. After electrophoresis, the gels were fixed in a solution of 40% ethanol-5% acetic acid overnight and subsequently silver stained (28).

Immunological assays.

Procedures for cell cultures and enzyme-linked immunosorbent assays for tumor necrosis factor alpha (TNF-α), interleukin-1β (IL-1β), and nitric oxide quantification have been previously described (68).

RESULTS

Isolation of PxB-susceptible mutants from mariner libraries of the meningococcal strain NMB.

A random Himar1 mariner transposon insertion library was generated following the in vitro mutagenesis procedure described by Pelicic et al. (45). Approximately 4,100 colonies were patch screened for PxB sensitivity by replica plating colonies from nonselective plates onto GC plates containing 8, 16, 32, 64, and 256 μg/ml PxB; a total of 30 mutants with a decrease of at least fourfold in PxB resistance were obtained (Table 2). Southern blot hybridization analysis illustrated single insertions in the transposon mutants when the chromosomal DNA from the analyzed strains were digested with ClaI and probed for the presence of the aphA-3 cassette (data not shown). Using SSP-PCR analysis, more than half of the insertions were mapped to the mtrCDE operon which encodes proteins forming an efflux pump (24, 49). PCR mapping showed that the mutations in the mtr region were localized to different sites within the mtrCDE operon (Fig. 1A). Three transposon chromosomal insertions were mapped to fructose-1,6-bisphosphatase (fbp). Other chromosomal locations of the transposon were mapped to porB, NMB0596 encoding an integral membrane protein, gshB (glutathione synthetase), fur, and NMB0204, a putative lipoprotein (Table 2).

TABLE 2.

Polymyxin sensitivity of mariner mutants and wild-type meningococcal strain NMB

Strains	MIC (μg/ml)	Insertion site
NMB	512
KA2001	32	mtrE
KA2003	32	porB
KA2005	32	mtrD
KA2006	32	mtrD
KA2007	128	Putative transposase
KA2008	32	mtrC
KA2009	32	mtrD
KA2010	32	mtrE
KA2011	128	NMB1060 (fbp)
KA2012	128	NMB0596
KA2013	64	NMB1419 (ruvC)
KA2014	64	fbp
KA2015	32	NMB1559 (gshB)
KA2016	128	fbp
KA2017	64	NMB0205 (fur)
KA2018	32	mtrD
KA2019	32	mtrD
KA2020	32	mtrD
KA2021	32	mtrD
KA2022	32	mtrD
KA2023	32	mtrD
KA2024	64	mtrD
KA2025	16	NMB0204
KA2026	32	mtrD
KA2027	32	mtrD
KA2028	32	mtrD
KA2029	32	mtrE
KA2030	128	mtrR-mtrC
KA2031	32	mtrE
KA2032	128	mtrR-mtrC
XZ134	32	mtrD
XZS3	8	mtrD, NMB1052
XZS8	8	mtrD, NMB0355
XZS9	8	mtrD, NMB0355
XZS17	8	mtrD, NMB1150
XZS18	2	mtrD, lptA
XZS32	2	mtrD, lptA
XZS38	2	mtrD, lptA

FIG. 1.

Mutations affecting cationic antimicrobial peptide resistance in Neisseria meningitidis. (A) Locations of transposon insertions within the mtrCDE locus as determined by PCR mapping. (B) Locations of the transposon insertion within lptA and the restriction sites (Hc, HincII; Sp, SspI) used in generating specific lptA mutations. (C) PxB-resistant transposon mutants mapped within the pilMNOPQ operon.

To further characterize mechanisms conferring PxB resistance, a second round of transposon mutagenesis was performed in the mtr efflux pump-deficient strain XZ134, which carries a deletion and insertion mutation in mtrD. mtrD encodes the inner membrane transporter protein component of the mtr efflux pump (24, 49). Approximately 4,400 individual transposon mutants were patched onto GC plates containing 8, 16, 32, and 64 μg/ml PxB, and a total of 9 mutants with ≥4-fold decrease in PxB resistance was identified. After confirming that the transposon insertion was associated with the PxB phenotype by backcross transformation (data not shown), SSP-PCR was performed and the insertion sites of seven mutants were mapped (Table 2). Mutants with MICs of 8 μg/ml included XZS8 and XZS9, both with an insertion in NMB0355, XZS3 with an insertion in NMB1052 (dedA), and XZS17 with an insertion in NMB1150 (ilvD; dihydroxy acid dehydratase). All three of these genes are predicted (TopPred II [8] and PSORT [16] programs) to encode inner membrane proteins, and the first two are adjacent to an ATP-binding protein of putative ABC transporter systems. Interestingly, the three mutants (XZS18, XZS32, and XZS38) with the lowest MICs (≤2 μg/ml) were all located within the NMB1638 locus: XZS18 carried the insertion upstream of the NMB1638 coding sequence, possibly within its promoter region, while XZS32 and XZS38 contained transposon insertions within the coding sequence (Fig. 1B).

Characterization of PxB-sensitive mutants.

To further confirm the correlation of PxB sensitivity with the transposon insertion, specific mutations in several genes identified in the library screening, including fbp, NMB0596, lpxL1 (downsteam of ruvC), and lpxL2 (a second lipid A late acyltransferase), were made, and PxB sensitivity was tested. In each case, the isogenic mutants reproduced the PxB sensitivity of the transposon mutants (data not shown).

Levels of PxB susceptibility in gram-negative enteric pathogens have been linked to LPS structures. Accordingly, we examined whether the meningococci expressing decreased PxB resistance had alterations in LOS. LOS produced by these mutants was first isolated by whole-cell proteinase K digestion and then resolved by tricine-SDS-PAGE followed by silver staining. A representative gel displaying the electrophoretic mobility of these LOS samples is shown in Fig. 2. All mutants examined yielded major doublet bands corresponding to the sialylated and the nonsialylated LOS and were indistinguishable from that of the wild-type strain. However, the major LOS doublet species in KA2013 migrated faster than the others, indicating a possible truncation of the LOS structure. An acyltransferase of lipid A, lpxL1, responsible for secondary fatty acid substitution of lipid A, was located downstream of the transposon insertion site in the KA2013 mutant and LOS isolated from the NMB291 (lpxL1::aphA-3) strain migrated similarly to that of the KA2013 mutant. LOS from the KA2033 (fbp::Ω) and NMB291 mutants was isolated in large scale and characterized by matrix-assisted laser desorption-ionization mass spectroscopy analysis. LOS structure changes were not detected in the fbp mutant (data not shown) while, as predicted, the lpxL1 mutant yielded predominantly a pentaacylated lipid A with a minor component of tetra-acylated lipid A structures (Fig. 3). Complementation of the KA2013 mutant with a second copy of lpxL1 in a shuttle vector rescued the PxB-sensitive phenotype (fourfold increase in PxB MIC), confirming that the phenotype of the KA2013 mutant is due to a polar effect on the downstream lpxL1.

FIG. 2.

Silver staining of whole-cell PK digestion of several PxB-sensitive mutants resolved on a 15% tricine-SDS-PAGE gel. Lane 1, 6.5-kDa marker (prestained bovine trypsin inhibitor); 2, wild-type strain NMB; 3, KA2009; 4, KA2010; 5, KA2011; 6, KA2012; 7, KA2013; 8, KA2014; 9, KA2015. The asterisk indicates that the KA2013 mutant with an insertion in ruvC produced a faster LOS migration pattern, likely due to the polar effect on the downstream lpxL1.

FIG. 3.

Matrix-assisted laser desorption-ionization-time-of-flight spectra of lipid A molecules released by mild acid hydrolysis from LOS isolated from strain NMB291. The identities of the various ions are as follows: 1268, P1 GlcNAc2 βOHC12:02 βOHC14:02 (tetraacyl); 1347, P2GlcNAc2 βOHC12:02 βOHC14:02 (tetraacyl); 1390, P1PEA1 GlcNAc2 βOHC12:02 βOHC14:02 (tetraacyl); 1450, P1 GlcNAc2 C12:01 βOHC12:02 βOHC14:02 (pentaacyl); 1530, P2 GlcNAc2 C12:01 βOHC12:02 βOHC14:02 (pentaacyl); 1573, P1PEA1 GlcNAc2 C12:01 βOHC12:02 βOHC14:02 (pentaacyl).

NMB1638 has recently been identified to encode the PEA transferase (LptA) responsible for PEA substitutions of the lipid A head groups in meningococci (10). The 1 and 4′ positions of the N-acetylglucosamine disaccharide of lipid A in several meningococcal LOS structures have been demonstrated to be modified with PEA groups (29, 62). To further confirm the role of LptA-mediated PEA modification of lipid A in mediating PxB resistance, an lptA::Ω(Sp) mutation was created in the wild-type strain NMB to generate strain KA2035. The MIC of this mutant was found to be 2 μg/ml, similar to those of the mariner transposon mutants. The PxB-sensitive phenotype was not due to a possible polar effect on the downstream NMB1637 gene (Fig. 1B), as anNMB1637::aphA-3 mutant yielded a PxB MIC similar to that of the wild type. Thus, the lptA::Ω single mutant yielded the lowest PxB MIC detected in meningococci, similar to those of the mariner mutants (XZS18, XZS32, and XZS38) created in the mtr-deficient background. Complementing the KA2035 (lptA::Ω) mutant with a second copy of lptA rescued the PxB sensitivity (MIC increased from 2 to 256 μg/ml). These data suggest that PEA modification of lipid A head group by LptA is a critical mechanism in conferring high-level PxB resistance in N. meningitidis.

To determine whether transposon insertions indirectly affected the function of the Mtr efflux pump and thus resulted in PxB sensitivity, sensitivities of the mutants to known substrates of Mtr pump (24), Triton X-100 and erythromycin were examined (Fig. 4). As expected, strains with a mutation in the mtrCDE-encoded efflux pump operon (XZ0134 and XZDM03) were more sensitive to both Triton X-100 and erythromycin than the wild-type strain. The lpxL1 (NMB291) and lptA mutants were also susceptible to Triton X-100. Susceptibility of the lpxL1 and lptA mutants to erythromycin was not significantly different from the wild-type parent. The porB (KA2003), fbp (KA2033), and NMB0204 (KA2025) mutants were not susceptible to the Mtr efflux pump substrates, indicating that sensitivity of these mutants to PxB was not caused by a decrease in the efflux function of the Mtr pump.

FIG. 4.

Sensitivity of PxB susceptible mutants to Triton X-100 (A) and erythromycin (B). Each datum is an average of three zones of growth inhibition obtained with 0.5% (white) and 1% (black) Triton X-100 in panel A and 5 μg/ml (white) and 50 μg/ml (black) of erythromycin in panel B after 24 h of incubation at 37°C. Significant difference, as determined by a Student t test, when compared to the wild-type parent strain NMB is indicated by an asterisk (P < 0.01).

PxB-susceptible mutants are also sensitive to the cationic antimicrobial peptides LL-37 and PG-1.

As the action of PxB is similar to those of other CAMPs found in humans and other vertebrates (25), it was predicted that PxB-sensitive mutants would be more susceptible to other CAMPs with different structures. To confirm this prediction and determine the relative contribution of the lipid A modification and the mtr efflux pump (51), the MICs of LL-37 and PG-1 were determined by broth dilution assays (51) for the wild-type strain NMB, the mtrD::Ω mutant XZ134, the lptA::Ω mutant KA2035, and the mtrE::Ω(Kn)/lptA::Ω(Sp) double mutant XZdm03 (Table 3). LL-37 is an amphipathic α-helical antimicrobial peptide, while PG-1 belongs to the cysteine-rich β-sheet structural family (51). LL-37 and PG-1 displayed a MIC of 15.6 μg/ml against the wild-type parental strains, while the mtrD, lptA, and mtrE/lptA double mutant yielded ∼10-fold decreases in MICs (0.98 to 3.9 μg/ml) for both LL-37 and PG-1. Thus, both mechanisms, the reduction of CAMP interaction through the modification of the lipid A head groups and the removal of peptides through an efflux pump, are critical in conferring resistance of meningococci to vertebrate CAMPs. These results also verified the use of PxB in assessing the actions of various structural classes of CAMPs.

TABLE 3.

Minimal growth inhibition concentration (μg/ml) of CAMPs of meningococcal strains

Strain	Genotype	MIC
		LL-37	PG-1	PxB
NMB	Wild type	15.6	15.6	512
XZ134	mtrD::Ω(Sp)	1.95	3.9	32
KA2035	lptA::Ω(Sp)	1.95	1.95	2
XZdm03	lptA::Ω(Sp); mtrE::Ω(Kn)	0.98-1.95	1.95	0.5

PxB neutralizes the LptA LOS more efficiently than the wild-type LOS.

The lptA mutant LOS is devoid of PEA substituents at both the 1 and 4′ positions of lipid A, whereas the wild-type LOS contained PEA modifications of the head groups (10, 62). The drastic reduction in PxB resistance indicated that LOS structural changes caused by the lptA mutation significantly enhanced PxB-LOS interaction. We predicted that this enhanced interaction would be reflected in the ability of PxB to neutralize the endotoxic effect of meningococcal LOSs. To test this hypothesis, LOSs from the wild type and the lptA mutant were extracted and purified (68). Levels of TNF-α and IL-1β induction were quantified in THP-1 human monocytes exposed to the purified LOSs. In addition, the induction of nitric oxide (NO) in RAW 2467 macrophages stimulated either directly with LOS or indirectly with the cell culture supernatants collected from THP-1 cells previously exposed to LOS was determined. Cytokine and NO inductions were then compared in the presence or absence of 5 μg/ml of PxB. As shown in Fig. 5, in all assays, PxB modestly diminished the induction by the wild-type LOS (≤25%), while the similar level of activity stimulated by the lptA LOS was significantly reduced (≥50%) in the presence of PxB. These data demonstrated that PxB was more efficient in neutralizing the TLR4-mediated activation elicited by the lptA LOS than the wild-type LOS.

FIG. 5.

Neutralization of meningococcal LOS isolated from the lptA mutant and the wild-type strain by PxB as measured in macrophages by (A) TNF-α induction, (B) IL-1β induction, (C) direct NO induction, and (D) indirect NO induction. Each assay was performed with 1 ng of highly purified LOS of meningococcal strain NMB in the absence (black) or presence (white) of polymyxin B (5 μg/ml).

Mutants with a defect in pilus biogenesis are more resistant to PxB.

To identify mutants with increased PxB resistance, in vitro mariner transposon mutagenesis reactions were performed using the chromosomal DNA of the nonencapuslated derivative of strain NMB, M7 (MIC = 256 μg/ml), as the target. Seven PxB-resistant colonies (>512 μg/ml) were recovered from a total of ∼10,000 mutants pooled from five independent mutagenesis reactions. Single transposon insertion was confirmed by Southern blots (data not shown). These resistant mutants were reexamined for their PxB susceptibilities and were confirmed to have a twofold increase in PxB resistance. The linkage between the transposon insertion and the PxB resistance phenotype was confirmed by backcross transformation. Five out of the seven mutants retained their enhanced resistance to PxB.

The transposon insertion sites were mapped by an SSP-PCR method, and all five mutations were mapped in the pilMNOPQ locus (Fig. 1C), predicted to be involved in pilus biogenesis (32). PilM is a predicted 371-amino-acid protein, and the transposon was inserted within pilM in mutants YM2, YM5, and YM7 at residues 74, 215, and 333, respectively. YM1 contained insertion within pilO, while the transposon in YM4 was located within the operon's promoter region (Fig. 1C). The meningococcal pilMNOPQ locus is highly homologous to that of Pseudomonas aeruginosa, which has been characterized to be critical in type IV pilin biogenesis (32). The meningococcal pilMNOPQ homologues have not been studied in detail, and only pilP and pilQ have been shown to be involved in pilin biogenesis (13). As pili are crucial for the natural competence of transformation in meningococci, the competency of pilM and pilO mutants were tested. The transformation efficiency of the parent strain M7 was shown to be 10⁻⁴ per μg DNA, whereas mutants YM1, -2, -5, and -7 were unable to be transformed (<10⁻⁸ per μg DNA). These results support that meningococcal pilMNOPQ genes, similar to the homologous genes in P. aeruginosa, are important in the expression of functional pili. To further define the role of these pilin genes in conferring PxB resistance, an isogenic mutant in pilM was made by insertion of an Ω(Sp) cassette. The pilM::Ω mutant consistently produced a twofold increase in PxB resistance. The increase in PxB resistance caused by the pilM mutation was not associated with the function of the Mtr efflux pump as a mtrE/pilM double mutant conferred twofold increases in PxB resistance when compared to the mtrE mutant.

DISCUSSION

N. meningitidis frequently colonizes the human nasopharynx as a commensal but is also a worldwide cause of epidemic meningitis and rapidly fatal sepsis. The human respiratory tract is the only known reservoir of N. meningitidis. On this mucosal surface, meningococci are constantly exposed to human endogenous CAMPs, an important host defense and component of innate immunity. Thus, it is not surprising that mechanisms rendering meningococci resistant to the action of CAMPs have evolved. Using random transposon mutagenesis, two important strategies, PEA modification of the lipid A head groups of LOS and the Mtr efflux pump, of meningococcal resistance to CAMPs were identified.

CAMP resistance via efflux mechanisms has been reported in several bacteria (1, 7, 30, 42, 43, 51). A set of genes, sapABCDF, sapJ, and sapG, from S. enterica serovar Typhimurium has been identified to be involved in resistance to the CAMP protamine (42, 43). SapG shares homology with the NAD⁺ binding protein, TrkA, and other E. coli proteins involved in potassium transport, while SapD and SapF exhibit homology with the ABC transporter family proteins. An efflux pump and potassium antiporter system, RosA and RosB, was shown to mediate resistance to CAMP in Yersinia enterocolitica (1). RosA is homologous to members of the major facilitator superfamily that utilizes proton-motive force for its activity (44), while RosB shows similarity to proteins involved in glutathione-regulated potassium efflux system (37). Similarly, a potassium uptake protein, TrkA, was shown to be required for CAMP resistance in Vibrio vulnificus (7), and a staphylococcal proton-motive force-dependent efflux pump protein, QacA, was shown to confer resistance to a small cationic peptide, thrombin-induced platelet microbicidal protein 1 (30). Finally, the Mtr efflux pump system of N. gonorrhoeae has been shown by Shafer et al. (51) to modulate gonococcal susceptibility to several structurally unrelated CAMPs, such as PG-1 that assumes a β-sheet conformation and LL-37 that folds into an α-helical structure. In this report we show that an Mtr efflux pump in N. meningitidis also decreases meningococcal susceptibility to the cyclic cationic peptide, polymyxin B, the α-helical peptide LL-37, and the β-sheet peptide, PG-1. Interestingly, in the absence of a functional Mtr efflux pump, other transport systems such as the ABC transporter system identified in mutant XZS8 may contribute to CAMP resistance.

Our data indicate that PEA substitutions of the lipid A head groups are a key factor that determines the intrinsic resistance of meningococci to PxB, whereas other structural features of lipid A also contribute. Lipid A species with or without PEA substitution on the head groups have similar biological activity as shown in macrophages (Fig. 5). These events are Toll-like receptor 4 (TLR4) dependent (68). However, lipid A without PEA modifications (produced in the meningococcal lptA mutant) exhibits enhanced neutralization efficiency by PxB compared to wild-type meningococcal lipid A (Fig. 5). In contrast to the lipid A of E. coli and Salmonella enterica, which may be modified by PEA after induction by certain environmental conditions, meningococcal lipid A is constitutively substituted with PEA (29, 31, 62). In S. enterica serovar Typhimurium, Ara4N modification is also important for polymyxin resistance (67). As the meningococcal genome does not encode the machinery to synthesize and attach Ara4N to the 4′-phosphate, constitutive PEA substitution is a critical strategy utilized by N. meningitidis in modifying lipid A head groups to maintain resistance to CAMPs. Differences in the ability of CAMPs to neutralize the wild-type meningococcal lipid A, the cytotoxic component of endotoxin, may correlate with the potent biological activity of meningococcal lipid A in inducing meningococcemia and meningococcal meningitis.

Specific mutations in the late acyl transferases, lpxL1 or lpxL2, responsible for adding the acyloxyacyl laurate chains to the N-linked hydroxymyristates (2 and 2′ positions) of meningococcal lipid A (66) also reduced PxB resistance. The contribution of lipid A acyl chain structures to CAMP resistance has been noted in other bacteria. A 3-O-deacylase encoded by a PhoP/PhoQ-activated gene (pagL) in Salmonella enterica serovar Typhimurium removes the R-3-hydroxymyristate moiety attached at position 3 of certain lipid A precursors (61), and this yields increased PxB resistance. Another S. enterica serovar Typhimurium PhoP/PhoQ-activated gene, pagP, is a palmitoyl transferase required both for biosynthesis of hepta-acylated lipid A species containing palmitate and for increased resistance to cationic antimicrobial peptides (3, 23). Mutation of pagP demonstrates increased outer membrane permeability in response to CAMP. Taken together, our observations and those for S. enterica serovar Typhimurium support the hypothesis that modulation of lipid A acylation can be a CAMP resistance mechanism (3, 23). Other meningococcal LPS structural changes such as truncation of the oligosaccharide outer or inner core did not affect PxB susceptibility of meningococci (data not shown).

The major outer membrane porin, PorB, of N. meningitidis also conferred resistance to polymyxin B, as a porB mutant showed a 16-fold decrease in MIC of PxB. A major outer membrane protein, OmpU, that functions as a porin for iron, phosphate, and sugar transport in Vibrio cholerae has been shown to confer resistance to polymyxin B (33). The increased sensitivity resulting from an ompU mutation is not due to a general defect in outer membrane structures or outer membrane permeability (33). Only certain outer membrane porins appear to influence CAMP resistance, as the major porin proteins, OmpC and OmpF, of E. coli are not involved in CAMP resistance (33). PorB may be involved in an active efflux process or may act in concert with other efflux pumps such as the mtr system. Another meningococcal mutant, KA2025, contains a mutation in a putative outer membrane lipoprotein that is transcribed divergently from the ferric uptake regulator (fur). This lipoprotein shares high homology in amino acid sequence and genetic organization to an outer membrane lipoprotein, OmlA, in Pseudomonas aeruginosa (40). The P. aeruginosa omlA mutant was shown to be susceptible to anionic detergents and various antibiotics, thus a structural role of OmlA in maintaining the cell envelope integrity has been proposed (40).

Mutations in meningococcal genes encoding metabolic enzymes, such as gshB and fbp, may indirectly cause PxB sensitivity. GshB, a glutathione synthetase, is involved in glutathione metabolism. As described above, glutathione-regulated potassium efflux proteins in various bacteria have been implicated in modulating PxB susceptibility (1, 7, 42, 43). Thus, the gshB mutation might have an indirect effect on the glutathione-regulated potassium efflux system present in meningococci (NMB0209). Two independent PxB sensitive mutants were mapped in fbp that encodes the fructose 1,6-bisphosphatase, which is involved in pentose phosphate metabolism and catalyzes the conversion of d-fructose 1,6-bisphosphate to d-fructose 6-phosphate. This enzyme predicted to contain a putative signal peptide by the PSORT program (16) is likely a lipoprotein, and mutations in this metabolic enzyme may reduce the membrane content of lipid A. Decreased levels of lipid A would increase outer membrane permeability and CAMP sensitivity, an effect seen with mutations in the lipid A biosynthesis genes (64); alternatively, the mutation may affect either peptidoglycan or phospholipid biosyntheses leading to alterations in cell permeability (5).

The correlation of the type IV pilin biogenesis apparatus and resistance to certain antibiotic agents has been observed previously. A mutant form of the N. gonorrhoeae pilus secretin protein PilQ has been identified that allows increased entry of antimicrobial compounds such as erythromycin, rifampin, and Triton X-100, while a null mutation in pilQ increases resistance to these compounds (6). In addition, a gonococcal pilin variant that synthesized elongated pilin that is not assembled into pili and a pilin pilE deletion mutant have higher levels of resistance to certain antibiotics such as kanamycin and penicillin (17). Our observation of increased PxB resistance upon inactivation of genes involved in pilin biogenesis suggests that the pilin secretion apparatus may be an entry point for several structurally distinct antimicrobial agents. However, the fact that resistant pilM (encoding a putative ATPase) mutants were recovered but not mutants with transposon insertions within pilQ suggests that the mechanism may be more complex.

In summary, we demonstrate a critical contribution of both PEA modification of meningococcal lipid A and the Mtr efflux pump in the ability of meningococci to resist the action of host cationic antimicrobial peptides. Other changes in outer membrane proteins such as PorB and outer membrane lipoprotein and alteration in pilin secretion or biogenesis can influence meningococcal susceptibility to CAMPs. The constitutive PEA modifications of meningococcal lipid A head groups creates a molecule that not only protects the meningococcus but also may contribute to the potent biological activity of meningococcal endotoxin in vivo.