ABSTRACT
There are a paucity of data concerning gene products that could contribute to the ability of Moraxella catarrhalis to colonize the human nasopharynx. Inactivation of a gene (mesR) encoding a predicted response regulator of a two-component signal transduction system in M. catarrhalis yielded a mutant unable to grow in liquid media. This mesR mutant also exhibited increased sensitivity to certain stressors, including polymyxin B, SDS, and hydrogen peroxide. Inactivation of the gene (mesS) encoding the predicted cognate sensor (histidine) kinase yielded a mutant with the same inability to grow in liquid media as the mesR mutant. DNA microarray and real-time reverse transcriptase PCR analyses indicated that several genes previously shown to be involved in the ability of M. catarrhalis to persist in the chinchilla nasopharynx were upregulated in the mesR mutant. Two other open reading frames upregulated in the mesR mutant were shown to encode small proteins (LipA and LipB) that had amino acid sequence homology to bacterial adhesins and structural homology to bacterial lysozyme inhibitors. Inactivation of both lipA and lipB did not affect the ability of M. catarrhalis O35E to attach to a human bronchial epithelial cell line in vitro. Purified recombinant LipA and LipB fusion proteins were each shown to inhibit human lysozyme activity in vitro and in saliva. A lipA lipB deletion mutant was more sensitive than the wild-type parent strain to killing by human lysozyme in the presence of human apolactoferrin. This is the first report of the production of lysozyme inhibitors by M. catarrhalis.
INTRODUCTION
Moraxella catarrhalis is a Gram-negative coccobacillus that can cause disease in both the upper and lower respiratory tracts of humans (1). In infants and young children, this bacterium is a significant cause of acute otitis media (i.e., middle ear infection) (2–4). In adults, M. catarrhalis can cause infectious exacerbations of chronic obstructive pulmonary disease (COPD) (5–7) and is likely responsible for approximately 4 million exacerbations of COPD annually in the United States (6). The latter disease has global significance because it has been predicted that by 2020, COPD will become the third leading cause of death worldwide (reviewed in reference 8). In addition, M. catarrhalis can cause sinusitis, pneumonia, and, more rarely, bacteremia (1, 9).
M. catarrhalis colonization of the human nasopharynx is apparently asymptomatic and, at least in infancy, can be correlated with an increased risk of otitis media (10). In this anatomic niche, M. catarrhalis likely forms a biofilm, together with the normal bacterial flora of the nasopharynx (11, 12). Once established in the nasopharynx, this bacterium can spread to the middle ear (causing otitis media) or to the lower respiratory tract (resulting in an infectious exacerbation of COPD). A number of putative M. catarrhalis colonization or virulence factors have been described in the past decade (13–20), but identification of those bacterial gene products that are truly essential for nasopharyngeal colonization has proceeded more slowly. To date, the type IV pilus is the only known adhesin shown to be involved in the colonization ability of M. catarrhalis in vivo in an animal model (18). More recently, expression of both M. catarrhalis open reading frame (MCORF) 1550 (21) and MCORF 113 (22) was shown to be necessary for the optimal persistence of M. catarrhalis O35E in the chinchilla nasopharynx, although the specific function of the products encoded by the genes remains to be determined.
To date, only a few regulatory systems in M. catarrhalis have been described in any detail, and none of these have any demonstrated role in controlling the production of proteins or other gene products proven to be involved in nasopharyngeal colonization. Early studies of gene regulation in M. catarrhalis focused on slipped-strand mispairing in homopolymeric and heteropolymeric nucleotide repeats that affected expression of several different M. catarrhalis genes (23–28). The first study looking at regulatory proteins in M. catarrhalis involved the Fur protein and its control of the expression of certain outer membrane proteins (29). More recently, the ability of reduced temperature to influence the expression of numerous different M. catarrhalis gene products was reported (30), and small regulons controlled by the OxyR and NsrR proteins have been described (31, 32). The likely presence of two-component signal transduction systems in M. catarrhalis was first noted in a PCR-based study which detected the presence of a gene encoding an OmpR family member that most closely resembled the PhoB response regulator of Pseudomonas aeruginosa (33). Analysis of the nucleotide sequence of the genome of M. catarrhalis ATCC 43617 (34) indicated the presence of at least four different two-component systems, a finding verified by subsequent analysis of the genomes of M. catarrhalis BBH18 (35) and 10 other M. catarrhalis strains (36, 37).
In the present study, generalized transposon-mediated mutagenesis yielded a mutant unable to grow in liquid media. This mutant had a transposon insertion in a gene, designated mesR, that encodes an OmpR family response regulator. Inactivation of the mesS gene encoding the cognate sensor (histidine) kinase also resulted in a mutant that could not grow in liquid media. DNA microarray analysis of a mesR mutant revealed that at least two open reading frames (ORFs) previously demonstrated to be involved in the ability of M. catarrhalis to colonize the chinchilla nasopharynx were upregulated in the absence of MesR. More importantly, two other genes (lipA and lipB) that were upregulated in this mesR mutant encoded previously undescribed proteins shown in the present study to function as inhibitors of human lysozyme activity. Recombinant versions of the M. catarrhalis LipA and LipB proteins inhibited the activity of both purified human lysozyme and lysozyme present in human saliva, and an M. catarrhalis mutant lacking the ability to express these two proteins exhibited increased sensitivity to killing by human lysozyme in vitro. This is the first report of lysozyme inhibitor production by M. catarrhalis.
MATERIALS AND METHODS
Bacterial strains and culture conditions.The M. catarrhalis and Escherichia coli strains used in this study are listed in Table 1. M. catarrhalis strains were routinely cultured in brain heart infusion (BHI) medium (Difco). BHI medium was supplemented with kanamycin (15 μg/ml) or spectinomycin (15 μg/ml) when appropriate. BHI agar plates were incubated at 37°C in an atmosphere containing 95% air and 5% CO2, while BHI broth cultures were incubated at 37°C with aeration. Addition of methyl cellulose (Sigma-Aldrich Chemicals) was used to increase the viscosity of BHI broth for one set of experiments. Growth curves were performed in the absence of antibiotics. E. coli strains were routinely grown in LB medium, with ampicillin (100 μg/ml) being added when appropriate.
Bacterial strains and plasmids used in this study
Disk diffusion assays.M. catarrhalis strains grown overnight on BHI agar plates were suspended to an optical density at 600 nm (OD600) of 1.0 in fresh BHI broth and then diluted 1/100 in BHI medium. A 100-μl portion containing approximately 105 CFU was spread on a BHI agar plate. Sterile disks (diameter, 6 mm) cut from Whatman no. 40 filter paper were saturated with 1 mg of SDS, 30 μg of polymyxin B, or 10 μl of 88 mM hydrogen peroxide and then applied to the top of the agar. The diameter of the zone of growth inhibition around each disk was averaged for three plates each in two independent experiments.
Attachment assay.The ability of wild-type and mutant strains of M. catarrhalis to attach to an immortalized human bronchial epithelial cell line (16HBE14o-) (38) in vitro was measured as described previously (22).
Transposome mutagenesis and identification of mutants of M. catarrhalis.M. catarrhalis strain ETSU-9 was mutagenized with an EZ::TN<KAN-2> transposome kit (Epicentre). Briefly, following the growth of M. catarrhalis in BHI broth to an OD600 of 0.7, cells were pelleted by centrifugation (8,000 × g for 10 min at 4°C), washed once with 10% (vol/vol) glycerol, and electroporated with the transposome as described previously (39). After recovery on BHI medium plates, cells were resuspended in BHI broth and plated on BHI agar containing kanamycin. The resultant colonies were patched onto BHI agar with kanamycin. Swatches from these patches were used to inoculate 2 ml BHI broth in 24-well plates.
RNA purification, DNA microarray analysis, and real-time RT-PCR.Bacterial cells harvested from BHI agar plates incubated overnight at 37°C in a 95% air–5% CO2 incubator were used to prepare total RNA using a RiboPure kit (Ambion, Austin, TX) following the manufacturer's instructions. The quality of the RNA samples was assessed by the Genomics Core Facility at the University of Texas (UT) Southwestern Medical Center using an Agilent 2100 bioanalyzer (Agilent Technologies). The M. catarrhalis custom DNA microarrays and the hybridization protocol used in these studies have been previously described in detail (34). Briefly, a 5-μg quantity of total RNA was used to obtain aminoallyl cDNA, which was subsequently posttranscriptionally labeled with Cy3 or Cy5, and each biological replicate was reverse labeled to avoid dye bias (i.e., a dye swap). Differential expression was defined as a 2-fold or larger change in the level of expression in the M. catarrhalis ETSU-9ΔmesR mutant relative to that in the wild-type strain. The data shown in Table 2 and Table S1 in the supplemental material include the expression profiles only of ORFs that had a P value of ≤0.05 after a one-sample t-test analysis. To validate the DNA microarray results, a set of 21 genes was selected from the final microarray data set and real-time reverse transcriptase PCR (RT-PCR) was used to confirm their relative transcription levels. The primer sequences used in these assays are listed in Table S2 in the supplemental material. Assays were performed on three independent biological replicates, using MCORF 1452 (gyrB) to normalize the amount of cDNA per sample. Additional real-time RT-PCR analyses were performed with wild-type strain O35E and a mesR mutant derived from it, again using BHI agar plate-grown cells for RNA extraction. The fold change of in the level of expression of each gene was calculated using the 2−ΔΔCT threshold cycle (CT) method.
RT-PCR for examining transcriptional linkage.A TaqMan reverse transcription reagent kit (Applied Biosystems) was used to synthesize cDNA from 500 ng of DNase-treated RNA extracted from BHI agar plate-grown wild-type M. catarrhalis ETSU-9 cells. Reaction mixtures were assembled in a 20-μl volume and included no-reverse transcriptase controls. A 2-μl aliquot of cDNA or a 1-μl aliquot of ETSU-9 genomic DNA (80 ng) was added to a PCR mix containing 2× AmpliTaq Gold 360 master mix (Applied Biosystems), 125 nM forward and reverse primers, and molecular biology-grade water for a total volume of 25 μl. The PCR was performed using an initial denaturation step of 95°C for 10 min, followed by 30 cycles of denaturation at 95°C for 30 s, annealing at 55°C for 30 s, and extension at 72°C for 30 s.
Construction of M. catarrhalis mesR and mesS deletion mutants.The mesR gene was disrupted using amplicons generated by overlapping extension PCR (40). First, using chromosomal DNA from M. catarrhalis ETSU-9 as the template, primers P1 and P2 (see Table S2 in the supplemental material) were used to amplify ∼1 kb of DNA upstream of mesR, including the mesR ATG translational start codon, while primers P3 and P4 (see Table S2 in the supplemental material) were used to amplify ∼1 kb of DNA downstream of mesR, including the mesR translational stop codon. Primers P2 and P3 include sequences that allow overlapping extension PCR with a nonpolar kanamycin resistance cartridge (41) using primers P1 and P4. The final amplicon was purified and used for transformation of M. catarrhalis ETSU-9, and mesR mutants were selected on BHI agar containing kanamycin. All mutants were confirmed by PCR and DNA sequence analyses. A mesR deletion mutant of M. catarrhalis strain O35E was also constructed, again with the nonpolar kanamycin cartridge.
The mesS gene was disrupted with a kanamycin cartridge fused in frame in a deletion between the nucleotides encoding amino acids 34 and 491. Primer sets P13/P14 and P15/P16 were used to amplify 750-bp fragments containing upstream and downstream flanking sequences and portions of the mesS ORF, using ETSU-9 chromosomal DNA as the template. Primers P14 and P15 append a sequence that permits overlapping extension PCR with the nonpolar kanamycin cartridge (41) using primers P13 and P16. This amplicon was purified and used to transform M. catarrhalis ETSU-9 to obtain a mesS deletion mutant that was confirmed by PCR and DNA sequence analysis.
Complementation of the mesR and mesS mutants.Primers P1 and P4 were used to PCR amplify the wild-type mesR locus from ETSU-9. After digestion with BamHI, this amplicon was ligated into the BamHI site of pWW115 (42). ETSU-9 was transformed with the ligation reaction mixtures, and transformants were selected on BHI medium containing spectinomycin. PCR and DNA sequence analyses were used to confirm insertion of the mesR locus into the plasmid vector. This plasmid (pSNJ224) was subsequently purified from ETSU-9 and used to transform ETSU-9ΔmesR. Complementation of mesS was accomplished similarly, with primers P13 and P16 being used to amplify the wild-type mesS locus from ETSU-9, which was then digested with BamHI and ligated into the BamHI site of pWW115 to obtain pSNJ225.
Construction of lipA and lipB mutants.The nonpolar kanamycin cartridge was amplified from pUC18K3 (41) using primers np kan 5′ and np kan 3′. Primers CP1 and CP2 were used to amplify a 1,027-bp fragment upstream of and extending 46 bp from the translation start codon into lipA from M. catarrhalis O35E chromosomal DNA. Primer set CP3 and CP4 was used to amplify a 1,069-bp fragment downstream of and extending 27 bp into lipA. These amplicons were combined with the kanamycin cartridge amplicon and subjected to overlapping extension PCR (43) using primers CP1 and CP4. The gel-purified product was then used to transform strain O35E. A kanamycin-resistant transformant was confirmed to be a lipA deletion mutant by PCR and DNA sequence analyses.
To construct a lipA lipB double mutant, primers CP5 and CP6 were used to amplify a 1,032-bp fragment downstream of and extending 54 bp into lipB. This amplicon was combined with the kanamycin cartridge and CP1-CP2 amplicons described immediately above and used in overlapping extension PCR with primers CP1 and CP6. The gel-purified final DNA product was used to transform O35E. A kanamycin-resistant transformant was confirmed to be a lipA lipB deletion mutant by PCR and DNA sequence analyses.
A lipB deletion mutation was also constructed in strain O35E. Primers WL1 and WL2 were used to amplify a 687-bp fragment upstream of and extending 43 bp into lipB, primers WL3 and WL4 were used to amplify a 609-bp fragment downstream of and extending 52 bp into lipB, and primers WW195 and WW196 were used to amplify the kanamycin cartridge from pAC7 (44). These three amplicons were used in overlapping extension PCR with primers WL1 and WL4. The final PCR product was used to transform O35E. A kanamycin-resistant transformant was confirmed to be a lipB deletion mutant by PCR and DNA sequence analyses.
Purification of recombinant MesR protein.Primers P11 and P12 were used to amplify the mesR coding sequence from M. catarrhalis ETSU-9 with BamHI ends appended. Following restriction digestion, this fragment was ligated into pQE30 (Qiagen). The subsequent plasmid was confirmed by DNA sequence analysis, designated pSNJ320, and introduced into chemically competent E. coli strain M15. To purify recombinant His6-MesR, 1 liter of LB broth containing ampicillin was inoculated with E. coli M15(pSNJ320) and induced with 1 mM IPTG (isopropyl-β-d-thiogalactopyranoside) when the cells reached mid-log phase (OD600 = 0.5). After 3 h of growth, cells were harvested by centrifugation (8,000 × g for 10 min at 4°C). The pellet was resuspended in buffer A (50 mM Na2PO4, 200 mM NaCl, 10 mM imidazole, pH 8) containing protease inhibitors and frozen at −20°C. The cells were disrupted by sonication, and the soluble crude extract was obtained by centrifugation (15,000 × g for 10 min at 4°C). The supernatant was filtered through a 0.22-μm-pore-size filter before using Ni2+-chelate chromatography to purify the His-tagged MesR protein. The purified protein was dialyzed against three 500-ml-volume changes of buffer A over 24 h, divided into multiple portions, and frozen.
Antiserum production.The use of commercial vendors for antiserum production was approved by the Institutional Animal Care and Use Committee at UT Southwestern Medical Center. MesR antiserum was produced using the purified His-tagged MesR protein described above to immunize rats (Rockland Immunochemicals). A MesS peptide conjugate was used for MesS antiserum production. The Protein Technology Center at the University of Texas Southwestern Medical Center synthesized an oligopeptide (CVRVSEEGNDEIAVLAHGFNQSAQK) encompassing amino acids 201 to 225 of the predicted M. catarrhalis ATCC 43617 MesS protein. This oligopeptide was covalently coupled to maleimide-activated keyhole limpet hemocyanin (Thermo Scientific) and used to immunize rats (Rockland Immunochemicals). LipA antiserum was produced using a purified His-tagged LipA protein to immunize mice (Cocalico Biologicals). LipB antiserum was produced using the purified maltose-binding protein (MBP)–LipB protein described below to immunize rabbits (Immunization Services, Animal Resources Center, UT Southwestern Medical Center).
Whole-cell lysate preparation and Western blot analysis.M. catarrhalis cells were scraped off the plates and resuspended to an OD600 of 0.75 to 1.0 in 5 ml. This suspension was subjected to centrifugation at 7,600 × g for 7 min, and the pellet was resuspended in 5 ml phosphate-buffered saline. A 200-μl portion of this suspension was mixed with 100 μl 3× digestion buffer (0.1875 M Tris-HCl, pH 6.5, 30% glycerol, 6% SDS, a small amount of pyronin Y for color). Samples were boiled for 5 min prior to use in SDS-PAGE. Western blotting-based immunodetection of specific proteins was performed using the different polyclonal antisera described above and the CopB-specific monoclonal antibody (MAb) 10F3 (45) as primary antibodies. Chemiluminescence detection was accomplished using the appropriate secondary antibodies with a SuperSignal West Pico kit (Thermo Scientific).
Purification of recombinant LipA and LipB proteins.Primer sets 994 F NdeI/994 R BamHI and 992 F NdeI/992 R BamHI were used to amplify the requisite lipA and lipB nucleotide sequences, respectively, from M. catarrhalis O35E chromosomal DNA. These sequences lacked the DNA encoding the putative signal peptides. These amplicons were digested with both NdeI and BamHI and ligated into similarly digested pMAL-p5X (New England BioLabs). The ligation mixes were transformed into E. coli TOP10 cells made competent by the use of calcium chloride. Plasmid constructions were confirmed by sequencing. To eliminate background maltose-binding protein expression from the E. coli malE gene, the recombinant plasmids were then transformed into the E. coli malE deletion mutant JW3994-2 (46).
A 5-ml portion of an overnight culture of the recombinant strain was inoculated into 1 liter LB broth in a 2.8-liter Fernbach flask. After ∼3 h growth at 37°C, 0.3 mM IPTG was added to induce gene expression. The culture was grown for an additional 3 h and then harvested, and the cell pellets were stored at −20°C overnight. To prepare a periplasmic extract, cells were resuspended in a total of 400 ml of 30 mM Tris-HCl (pH 8.0, containing 20% sucrose and 1 mM EDTA) and incubated for 10 min with shaking. After centrifugation at 8,000 × g for 20 min at 4°C, the cells were resuspended in a total of 400 ml cold 5 mM MgSO4, shaken for 10 min at 4°C, and centrifuged as described above. The supernatant was poured into a T150 tissue culture flask and adjusted to contain a 20 mM final concentration of Tris-HCl (pH 7.4). After the addition of washed amylose resin (New England BioLabs), the flask was rocked at 4°C for 2 h. The supernatant plus resin was subsequently centrifuged at 500 × g at 4°C for 5 min, followed by three washes with 20 mM Tris-HCl, and poured into a column. Protein fractions were eluted with 20 mM Tris-HCl (pH 7.4) containing 10 mM maltose, examined by Coomassie blue staining after SDS-PAGE, and concentrated in an Amicon Ultra centrifugal filter (3,000-Da-molecular-mass cutoff) (Millipore). After the addition of glycerol to a final concentration of 20%, purified proteins were stored at −80°C. Purified MBP was prepared in the same fashion, using empty pMAL-p5X vector. Protein concentrations were determined by using a Bio-Rad DC protein assay kit (Bio-Rad).
Measurement of lysozyme activity in the presence of inhibitors.Inhibition of lysozyme enzymatic activity was determined by the method of Callewaert et al. (47), using a 96-well plate-based assay. Purified MBP, MBP-LipA, or MBP-LipB (1 μM each) was added to 20 mM potassium phosphate buffer (pH 7.4) containing a suspension of freeze-dried Micrococcus lysodeikticus (final concentration, 0.5 mg/ml; Sigma). To start the reaction, 20 U of recombinant human lysozyme (Sigma) or 40 U of hen egg white lysozyme (Sigma) in 20 mM potassium phosphate buffer (pH 7.4) was added to the suspension, for a total volume of 300 μl. Measurements were recorded using an Epoch spectrophotometer (Bio Tek) at time zero and every 15 min thereafter, for a total of 2 h at room temperature. Each experiment used samples in triplicate for each reaction mixture.
Inhibition of lysozyme activity in saliva.Pooled human saliva (Lee Biosolutions) was sterilized through a 0.22-μm-pore-size filter and stored at −80°C. Before each assay, an aliquot of saliva was thawed at room temperature. Each lysozyme inhibition assay contained 0.5 mg/ml freeze-dried M. lysodeikticus cells (Sigma), potassium phosphate buffer (pH 7.4), 50 μl of human saliva, and 1 μM purified MBP, MBP-LipA, or MBP-LipB in a total volume of 300 μl in a 96-well plate. Additional wells containing only M. lysodeikticus cells and buffer or containing M. lysodeikticus cells, buffer, and saliva were included as negative and positive controls, respectively. Lysozyme inhibition was measured as described immediately above. Each experiment used samples in triplicate for each reaction mixture.
Sensitivity of M. catarrhalis strains to killing by lysozyme in the presence of lactoferrin.Bacterial cells from overnight growth on BHI agar plates were suspended in BHI medium to an OD600 of ∼0.7. Recombinant human lysozyme (Sigma) was dissolved in BHI medium at a concentration of 1.2 mg/ml and filter sterilized. The assay mixture contained 20 μl bacterial cell suspension, 50 μl recombinant human lysozyme (final concentration, 300 μg/ml), 80 μl BHI medium, and 50 μl human apolactoferrin (final concentration, 2 mg/ml; Lee Biosolutions) in a total volume of 200 μl per well in a 96-well plate that was incubated at 37°C in a 95% air–5% CO2 incubator. For each assay, 10-μl aliquots were removed from each well, serially diluted, and plated on BHI agar at time zero and 6 h. Each experiment involved two wells each of the wild-type O35E parent strain and the O35EΔlipAlipB mutant.
Microarray data accession number.The raw data have been deposited in the NCBI Gene Expression Omnibus (GEO) database (http://www.ncbi.nlm.nih.gov/geo/) under accession number GSE50229.
RESULTS
Identification of a mutant deficient in growth in liquid media.Generalized transposon-mediated mutagenesis of M. catarrhalis ETSU-9 yielded a mutant unable to grow in liquid media; this mutant was shown to have a transposon insertion in MCORF 725, encoding a predicted protein with homology to bacterial OmpR-family regulatory proteins (listed in Table S2 in the supplemental material in reference 34). This gene was designated mesR (Fig. 1A), for the reasons explained below.
Characterization of wild-type M. catarrhalis ETSU-9 and the mesR, mesS, and complemented mutants of M. catarrhalis ETSU-9. (A) The mesRS locus in wild-type ETSU-9; (B) the mesR mutant ETSU-9ΔmesR; (C) the mesS mutant ETSU-9ΔmesS. (D) Western blot analysis of MesR and CopB protein production by wild-type ETSU-9 (lane 1), ETSU-9ΔmesR (lane 2), ETSU-9ΔmesR(pSNJ224) (lane 3), and ETSU-9ΔmesR(pWW115) (lane 4). (E) Western blot analysis of MesS and CopB protein production by wild-type ETSU-9 (lane 1), ETSU-9ΔmesS (lane 2), ETSU-9ΔmesS(pSNJ225) (lane 3), and ETSU-9ΔmesS(pWW115) (lane 4). Detection of the CopB outer membrane protein (22) was used as a loading control to ensure that equivalent amounts of whole-cell lysates were present in each lane. Size position markers (in kDa) are present on the left side of panels D and E. (F) RT-PCR-based analysis of transcriptional linkage between mesR and mesS. Primers 724 F and 725 R were designed to allow amplification of a 233-bp product that spans from mesR to mesS. Primers 990 F and 990 R bind inside the oppA ORF (105), were designed to allow amplification of a 396-bp product, and were used here as a positive control. Lanes 1 and 4, samples from RT-PCR with wild-type ETSU-9 RNA as the template and RT added; lanes 2 and 5, samples from RT-PCR with wild-type ETSU-9 RNA as the template and no RT; lanes 3 and 6, samples from RT-PCR with wild-type ETSU-9 chromosomal DNA as the template. Size position markers (in 100-bp increments) are presented on the left side.
Construction and characterization of an isogenic mesR mutant.A nonpolar deletion mutation was constructed in the mesR gene in M. catarrhalis ETSU-9 (Fig. 1B and D, lane 2). Like the mesR transposon insertion mutant, this new ETSU-9 mesR mutant was unable to grow in broth (Fig. 2A). Provision of a wild-type mesR gene in trans (Fig. 1D, lane 3) eliminated this broth growth deficiency (Fig. 2A). Despite being unable to grow in liquid BHI medium, the mesR mutant readily formed colonies on BHI agar plates (data not shown).
Growth characteristics of wild-type M. catarrhalis ETSU-9 and mesR, mesS, and complemented mutants of M. catarrhalis ETSU-9. (A) Broth growth curves for wild-type ETSU-9 (filled diamonds), ETSU-9ΔmesR (open squares), ETSU-9ΔmesR(pSNJ224) (open circles), and ETSU-9ΔmesR(pWW115) (filled triangles). (B) Broth growth curves for wild-type ETSU-9 (filled diamonds), ETSU-9ΔmesS (open squares), ETSU-9ΔmesS(pSNJ225) (open circles), and ETSU-9ΔmesS(pWW115) (filled triangles). (C) Inhibition of the growth of wild-type ETSU-9, the mesR mutant, and complemented mutants on BHI agar by different agents, including SDS, polymyxin B, and hydrogen peroxide; data were analyzed by using a 2-way analysis of variance with a Sidak correction (**, P < 0.0022; ***, P < 0.0009; ****, P < 0.001). All growth data are the results of two independent experiments.
Efforts to enable the growth of this mesR mutant in liquid media included the use of BHI broth with increased osmolality, accomplished by the addition of increasing amounts of potassium chloride or sucrose independently, and were unsuccessful (data not shown). Similarly, increasing the viscosity of BHI broth medium by the inclusion of increasing concentrations of methyl cellulose (48) did not allow growth of this mesR mutant (data not shown). The growth phenotype of these mutants suggested that this particular response regulator was part of a system required for growth in one environment (i.e., in liquid media) but not for growth in a different physical environment (i.e., on solidified media). We designated this gene Moraxella environmental system R (mesR).
Construction and characterization of an isogenic mesS mutant.Inspection of the M. catarrhalis locus containing mesR revealed the presence of a 1.6-kb ORF located immediately downstream (Fig. 1A) which encoded a protein with homology to sensor (histidine) kinases of bacterial two-component signaling systems (TCSs). This gene was designated mesS. An ETSU-9 mesS mutant (Fig. 1C and E, lane 2) was constructed and shown to be unable to grow in liquid media (Fig. 2B). As with the mesR mutant, the growth deficiency of the mesS mutant in liquid media could be remedied by provision of the wild-type mesS gene in trans (Fig. 1E, lane 3, and 2B). RT-PCR analysis showed that mesS and mesR were transcriptionally linked (Fig. 1F).
Structural biology of the M. catarrhalis MesS and MesR proteins.The protein encoded by M. catarrhalis mesS is likely to be a 521-amino-acid (aa) sensor (histidine) kinase (HK), based on conserved domain searches and sequence homology to other HKs. Secondary structure prediction (49), transmembrane helix predictions (50), and structure-based comparison (51) to the crystal structure of the cytoplasmic portion of HK853 from Thermotoga maritima (52) (see Fig. S1A in the supplemental material) suggest that MesS is a transmembrane protein; residues 38 to 169 comprise a putative periplasmic domain that is bracketed by transmembrane α helices. The cytoplasmic portion of the protein (residues 190 to 521) likely comprises the HK apparatus of the protein. This analysis does not readily indicate which particular signaling pathway that this HK might belong to, although BLAST searches revealed homologies to both CpxA and BaeS, both of which are involved in responding to cell envelope stress (53, 54).
M. catarrhalis mesR encodes a polypeptide containing 240 aa. A conserved domain search clearly identified MesR to be a response regulator protein in a two-component signaling system. The predicted effector domain of MesR (residues 136 to 240) probably binds to DNA because its sequence is similar to the sequences of effector domains that are known to bind DNA as transcription factors. For example, the MesR effector domain shares 33% identity with that of the PhoB transcriptional regulator from E. coli (see Fig. S1B in the supplemental material). The conservation of amino acid residues making specific contacts between E. coli PhoB and its target DNA implies that the DNA sequence recognized by MesR may be similar to that of E. coli PhoB (55).
Sensitivity of the mesR mutant to detergent and other stressors.The genomes of the Rhizobiaceae encode a TCS comprised of the ChvG sensor kinase and the ChvI response regulator that has been shown to be important for symbiosis between these bacteria and leguminous plants (56). A recent study of a mutant of Rhizobium leguminosarum unable to express the ChvG sensor kinase indicated that this mutation rendered this legume symbiont unable to grow in liquid media (57). Additional experimentation indicated that this chvG mutant was also more sensitive than its wild-type parent strain to some stressors (57). The ETSU-9 mesR mutant was tested for its sensitivity to several different agents, including SDS, polymyxin B, and hydrogen peroxide, in disk diffusion assays (Fig. 2C). This mesR deletion mutant was more sensitive than the wild-type parent strain to all three of these agents, and this increased sensitivity was eliminated by the provision of the wild-type mesR gene in trans (Fig. 2C).
Initial investigation of a MesRS regulon.To identify genes whose expression might be controlled by a MesRS TCS, DNA microarray analysis was performed with wild-type M. catarrhalis ETSU-9 and the ETSU-9ΔmesR mutant. A number of genes were expressed differently in this mesR mutant than in the parent strain (i.e., 134 genes with increased expression and 305 genes with decreased expression; see the data for GEO accession number GSE50229 at www.ncbi.nlm.nih.gov/geo/). The M. catarrhalis ETSU-9 genes whose expression was increased or decreased at least 4-fold in the mesR mutant are listed in Table 2. Real-time RT-PCR analysis of selected genes identified in these DNA microarray experiments showed a good correlation between the two different sets of data (Fig. 3).
M. catarrhalis genes whose expression is most affected by the absence of MesR, measured by DNA microarray analysis
Comparison of DNA microarray and real-time RT-PCR (quantitative RT-PCR [qRT-PCR]) data for selected M. catarrhalis ETSU-9 genes whose expression was affected by inactivation of mesR. (A) Expression levels of 21 selected genes in ETSU-9ΔmesR cells relative to those in wild-type ETSU-9 cells were measured by DNA microarrays (black bars) or quantitative RT-PCR analysis (white bars), as described in Materials and Methods. These data are the means of the results from three independent DNA microarray experiments and three independent quantitative RT-PCR experiments. (B) Plot of the correlation between the log2 values from the DNA microarray and quantitative RT-PCR analyses. The diagonal line represents the power trend line.
Among the upregulated ORFs were two previously identified as being relevant in vivo in an animal model. Inactivation of MCORF 1550 decreased the ability of M. catarrhalis O35E to colonize the chinchilla nasopharynx (21), and a lack of expression of the outer membrane lipoprotein encoded by MCORF 113 resulted in the decreased persistence of M. catarrhalis in the chinchilla nasopharynx (22). Expression of both MCORF 1550 (which increased 6-fold; Table 2) and MCORF 113 (which increased 1.7-fold; see Table S1 in the supplemental material) was upregulated in the mesR mutant. The increased expression of both of these ORFs was subsequently confirmed by real-time RT-PCR (Fig. 3). Other upregulated ORFs (Table 2) included several encoding surface-exposed or outer membrane proteins. These included MCORFs 976 and 295, encoding the MchB outer membrane protein transporter (20) and the associated MhaB1 filamentous hemagglutinin-like protein (58), respectively. In addition, MCORF 995, encoding the M35-like porin protein (59), was upregulated 8-fold (Table 2).
Located close to MCORF 995, but on the opposite DNA strand (Fig. 4A), were two genes whose expression was increased at least 2-fold in this mesR mutant. MCORF 994 (whose expression was increased 4.2-fold; Table 2) and MCORF 992 (whose expression was increased 2.1-fold; see Table S1 in the supplemental material) encode predicted proteins that were originally annotated (see Table S2 in the supplemental material in reference 34) as being most like a conserved hypothetical protein encoded by Psychrobacter sp. strain PRwf-1 and the hypothetical protein NGO1981 from Neisseria gonorrhoeae FA 1090, respectively. In the present study, these two proteins were designated LipA and LipB, respectively, for reasons explained below. Examination of the M. catarrhalis genome indicated that these two ORFs were separated by 280 nucleotides (Fig. 4A).
Characterization of wild-type M. catarrhalis O35E and lipA, lipB, and lipA lipB mutants of M. catarrhalis O35E. (A) The lipA and lipB loci in wild-type O35E; (B) the lipA mutant O35EΔlipA; (C) the lipB mutant O35EΔlipB; (D) the O35EΔlipAlipB double mutant. MCORF 995 is labeled “M35-like.” (E) Western blot analysis of LipA, LipB, and CopB protein production by wild-type O35E (lane 1), O35EΔlipA (lane 2), O35EΔlipB (lane 3), and O35EΔlipAlipB (lane 4). Detection of the CopB outer membrane protein (22) was used a loading control to ensure that equivalent amounts of whole-cell lysates were present in each lane. Size position markers (in kDa) are present on the left side of panel E. (F) Attachment of wild-type O35E and O35EΔlipAlipB to 16HBE14o- cells in vitro; the attachment assay was performed as described previously (22). These results are the mean of two independent experiments (Mann-Whitney test; P = 0.85). ns, no statistically significant difference.
Properties of the M. catarrhalis LipA and LipB proteins.The predicted LipA protein contained 136 aa and had a predicted signal peptidase I cleavage site (i.e., A-H-A; identified using SignalP [version 4.0] software) located 23 residues from the N terminus. The predicted LipB protein had 118 residues, with a signal peptidase I cleavage site (i.e., A-N-A) being predicted just after residue 22. Disregarding the signal sequences in each protein, an alignment of the putative mature forms of LipA and LipB demonstrates that they are approximately 37% identical and 69% similar (Fig. 5). BLAST (60) homology searches using LipA and LipB returned numerous proteins from the bacterial genera Kingella, Neisseria, Acinetobacter, and Psychrobacter. Many of the returned proteins were annotated as “adhesins” or “components of adhesin complex,” apparently (61) on the basis of homology to the LecA protein from Eikenella corrodens (62). A significant difference between the M. catarrhalis proteins and LecA is the fact that the latter has two domains, but the former have only one (data not shown). Nonetheless, the one-domain Acp protein encoded by the Neisseria meningitidis NMB2095 genome has adhesin-like activity (63) and strong sequence homologies to LipA and LipB (82% and 79% similarities, respectively).
Alignment of the M. catarrhalis LipA and LipB proteins with homologous proteins. A structure-based sequence alignment of LipA, LipB, the N. meningitidis Acp adhesin (NmAcp), the two E. corrodens LecA domains (EcN, the E. corrodens amino-terminal domain; EcC, the E. corrodens carboxyl-terminal domain), the B. abortus lysozyme inhibitor PliC (PliC-Ba), and the S. Typhimurium lysozyme inhibitor PliC (PliC-St) is shown. Residue numbering of the full-length proteins is shown at the right and left. The secondary structure (106) from the known structure of S. Typhimurium PliC (65) is shown, with blue arrows depicting β strands and black lines showing loop regions; all strands are labeled, as are relevant loops. S. Typhimurium PliC residues homologous to those known to contact hen egg white lysozyme in the P. aeruginosa MliC-hen egg white lysozyme structure (66) are boxed. Also shown in boxes are residues from B. abortus PliC that contact human lysozyme in that complex structure (68). The two absolutely conserved cysteines are highlighted in yellow, and the absolutely conserved serine is highlighted in blue. Other conserved residues are highlighted in purple.
Using hidden Markov methods (64), we identified three proteins that have high probabilities of being structurally similar to the M. catarrhalis proteins: PliC from Salmonella enterica serovar Typhimurium (PDB accession number 3OE3; 60% and 67% similarities to LipA and LipB, respectively) (65), MliC from P. aeruginosa (PDB accession number 3F6Z; 59% and 64% similarities, respectively) (66), and MliC from E. coli (PDB accession number 2F09; 60% and 53% similarities, respectively) (67). Also, the recently reported structure (68) of PliC from Brucella abortus is structurally similar to the structures of LipA and LipB (58% and 71% similarities, respectively). All of these proteins are either known or suspected inhibitors of C-type vertebrate lysozymes (47). These proteins share a topology with eight β strands arranged as a barrel. The structures comport with secondary structure predictions for LipA and LipB (69), which feature only β strands and loop regions (not shown). Additionally, all of the known structures feature an intramolecular disulfide bond between cysteine residues in the first and eighth β strands; these residues are conserved in the M. catarrhalis proteins (Fig. 5). Thus, LipA and LipB have strong sequence homologies to apparent bacterial adhesins but a high probability of having structures that are very similar to those of lysozyme inhibitors.
Effect of lipA and lipB mutations on adherence ability of M. catarrhalis.The amino acid sequence similarities and structural biology considerations described immediately above raised two possibilities: the LipA and LipB proteins were either adhesins or lysozyme inhibitors. To determine whether these two M. catarrhalis proteins might be involved in adherence, the lipA and lipB genes were deleted in M. catarrhalis strain O35E. The O35E strain was used for mutant construction because its genome has been sequenced (36) and because we have extensive experience in using the O35E strain in attachment assays (22, 70). Western blot analysis confirmed that individual lipA and lipB mutants (Fig. 4B and C, respectively) each failed to express the LipA or LipB protein (Fig. 4E) and that the lipA lipB double mutant (Fig. 4D) expressed neither protein (Fig. 4E). When used in attachment assays with the human bronchial epithelial cell line 16HBE14o- (38), the lipA lipB mutant adhered to these epithelial cells at levels equivalent to those obtained with the wild-type parent strain (Fig. 4F) (P = 0.85).
Inhibition of lysozyme activity by LipA and LipB.The structural biology-derived prediction that both LipA and LipB were similar to bacterial lysozyme inhibitors also indicated that these proteins were likely localized to the periplasmic compartment in M. catarrhalis (71). To increase the likelihood of obtaining recombinant proteins that were folded properly, LipA and LipB were expressed as fusion proteins with a maltose-binding protein (MBP) that is secreted into the periplasm of E. coli. Both of these fusion proteins and MBP itself were purified by amylose column chromatography.
Incubation of purified human lysozyme and MBP together with cells of M. lysodeikticus resulted in lysis of these Gram-positive bacterial cells, as evidenced by a continuing decrease in the density of the cell suspension (Fig. 6A). However, when either MBP-LipA or MBP-LipB was present in the reaction mixture, there was much less lysis of M. lysodeikticus cells (Fig. 6A), with MBP-LipA apparently inhibiting lysis more effectively than MBP-LipB on a molar basis. On the basis of this inhibitory activity of the purified recombinant M. catarrhalis proteins, the proteins encoded by MCORFs 994 and 992 were designated lysozyme inhibitor protein A (LipA) and LipB, respectively. Additional experiments were performed with LipA, the more potent of the two inhibitors. When hen egg white lysozyme was used in place of human lysozyme, MBP-LipA did not inhibit lysis of the target cells (Fig. 6B).
The M. catarrhalis LipA and LipB proteins inhibit lysozyme activity. (A) Purified recombinant MBP-LipA (closed circles), MBP-LipB (open circles), and MBP (closed squares) (1 μM each) were incubated with dried cells of M. lysodeikticus and recombinant human lysozyme (20 U) as described in Materials and Methods. Lysozyme activity was measured by determination of the decrease in the OD600 over time. A 2-way analysis of variance with a Tukey correction for multiple comparisons indicated that the extent of lysis obtained in the presence of MBP-LipA and MBP-LipB was different (P < 0.05) from that obtained with MBP by the 15-min time point. By the 75-min time point, the extent of lysis obtained with MBP-A was different (P < 0.05) from that obtained with MBP-LipB. These are the data from two independent experiments each involving triplicate sample wells. (B) Purified recombinant MBP-LipA (open circles) and MBP (closed squares) (1 μM each) were incubated with dried cells of M. lysodeikticus and hen egg white lysozyme (40 U). A 2-way analysis of variance with a Sidak correction for multiple comparisons indicated no difference in the extent of lysis obtained with MBP-LipA and MBP at any time point during the assay. These are the data from two independent experiments each involving triplicate sample wells. (C) Inhibition of lysozyme activity in pooled normal human saliva. Dried cells of M. lysodeikticus were incubated with buffer (closed triangles), saliva (open squares), saliva and MBP (closed squares), saliva and MBP-LipA (closed circles), and saliva and MBP-LipB (open circles) as described in Materials and Methods. Data were analyzed by using a 2-way analysis of variance for multiple comparisons with a Tukey correction. By 15 min, lysis caused by saliva was more extensive than that obtained with the buffer control (P < 0.0001). By 15 min, the extent of lysis obtained in the presence of saliva with either MBP-LipA or MBP-LipB was different (P < 0.0001) from that obtained with saliva and MBP. By 15 min, the extent of lysis obtained with saliva and MBP-LipA was different (P < 0.0001) from that obtained with saliva and MBP-LipB. These are the data from two independent experiments each involving triplicate sample wells. (D) Killing of the O35E wild-type strain (black columns) and the O35E ΔlipA ΔlipB mutant (gray columns) by human lysozyme in the presence of human apolactoferrin. Data analysis with a 2-way analysis of variance for multiple comparisons with a Sidak correction indicated that the extent of killing of the mutant was different (P < 0.0001) from that of the wild-type strain. These are the data from three independent experiments each involving duplicate sample wells.
Next, pooled human saliva was used to provide a more relevant environment for testing the ability of LipA and LipB to inhibit lysozyme activity. The pooled human saliva contained readily detectable lysozyme activity (Fig. 6C). When MBP-LipA and MBP-LipB were independently added to pooled human saliva, both of these recombinant proteins inhibited lysozyme enzymatic activity (Fig. 6C). Once again, the fusion protein containing LipA was more effective than MBP-LipB in inhibiting lysozyme enzymatic activity (Fig. 6C).
Inactivation of lipA and lipB renders M. catarrhalis more susceptible to killing by human lysozyme.Lack of expression of lysozyme inhibitors can reduce the resistance of Gram-negative bacteria to killing by lysozyme, especially in the presence of lactoferrin (47, 72, 73), which increases the permeability of the outer membrane (74, 75). Incubation of the wild-type M. catarrhalis O35 strain and the O35E lipA lipB double mutant in the presence of both lactoferrin (2.0 mg/ml) and human lysozyme (300 μg/ml) for 6 h resulted in a decrease in the viability of both strains, with the mutant being killed more extensively than the wild-type parent strain (Fig. 6D; P < 0.0001).
Expression of lipA and lipB in an M. catarrhalis O35E mesR mutant.The ETSU-9 mesR mutant described above exhibited increased levels of transcripts from both lipA and lipB. To eliminate the possibility that this was a strain-specific result, real-time RT-PCR analysis was performed with RNA prepared from an O35E mesR mutant. This O35E-derived mutant was unable to grow in broth media, just like the ETSU-9 mesR mutant (data not shown). Compared to the wild-type O35E strain, this mesR mutant exhibited increased expression of both lipA and lipB (Fig. 7). In addition, this O35E mesR mutant showed increased expression of MCORF 995, MCORF 1550, and MCORF 113 (Fig. 7). These results indicate that the increased expression of these same five genes in the ETSU-9 mesR mutant (Fig. 3A) is not a strain-specific effect.
Expression of selected genes measured by real-time RT-PCR in an M. catarrhalis O35E mesR mutant. Expression levels of lipA, lipB, MCORF 995, MCORF 1550, and MCORF 113 in the O35E mesR mutant relative to those in wild-type O35E were determined. MCORF 1452 (gyrB) was used to normalize the amount of cDNA per sample. The results depicted are representative of those from two independent experiments.
DISCUSSION
The inability of both the M. catarrhalis mesR mutant and the mesS mutant to grow in liquid media (Fig. 2A and B) could be due to downregulation of one or more genes whose normal expression is essential for growth in liquid media. Perusal of the list of genes whose expression was most downregulated in the mesR mutant (see Table S1 in the supplemental material) reveals many different enzymes involved in metabolism (e.g., pyridine nucleotide-disulfide oxidoreductase, l-aspartate oxidase, quinolinate synthetase). Alternatively, it is possible that increased production of certain gene products in these mutants resulted in their failure to survive and grow in a liquid environment. The precise defect(s) in the mesS and mesR mutants that prohibits these TCS mutants from growing in a liquid environment remains to be determined, and identification of the relevant gene products will likely require the use of complementation analysis or another genetic selection method in future studies.
To the best of our knowledge, previous reports of TCS mutations that resulted in strains that could not grow in liquid media are limited to plant symbionts in the family Rhizobiaceae. Inactivation of the chvG sensor kinase gene in R. leguminosarum (57) and the very similar exoS sensor kinase gene in Sinorhizobium meliloti (56) resulted in mutants that did not grow in a liquid medium. The R. leguminosarum chvG mutant was sensitive to the same stressors (i.e., SDS, polymyxin B, and H2O2) used in the present study (Fig. 2C). In this Rhizobium mutant, provision in trans of a gene encoding the RopB outer membrane protein restored the ability of this mutant to resist killing by these stressors. In R. leguminosarum, expression of RopB is linked to changes in lipid A structure (76). Use of the RopB sequence in a BLASTp search of the proteins encoded by the M. catarrhalis BBH18 genome (35) yielded only four hits (excluding signal peptides), one of which was a 63-aa sequence with 29% identity to RopB (data not shown).
Whether MesR functions in a positive or negative manner in M. catarrhalis remains to be determined, but there are ample data to indicate that some TCS response regulators can act as both transcriptional activators and repressors (77–79). Examination of data from a previous study that identified M. catarrhalis genes whose expression was upregulated when broth-grown bacteria were inoculated into the chinchilla nasopharynx (21) showed that there was little or no change in the level of mesR transcription, even though the levels of lipA and lipB transcripts increased approximately 4-fold and 1.5-fold, respectively. In M. catarrhalis bacteria that attached to human bronchial epithelial cells in vitro (80), there were very modest (i.e., 1.4- to 1.5-fold) increases in mesR, lipA, and lipB transcription levels. Further interpretation of the available regulatory data is limited by the fact that the phosphorylation, or a lack thereof, of a response regulator (i.e., MesR) greatly affects its ability to stimulate or inhibit transcriptional activity (81).
An unexpected consequence of our preliminary investigation of the MesRS regulon was the identification of M. catarrhalis genes encoding two different lysozyme inhibitors. Lysozyme, a member of the bacterial cell wall hydrolase family, possesses muramidase enzymatic activity which hydrolyzes the β-1,4 glycosidic linkages between N-acetylmuramic acid and N-acetyl-d-glucosamine residues in the peptidoglycan of the bacterial cell wall. Hydrolysis of the peptidoglycan can lead to lysis of the bacterium because of turgor pressure. Equally important is the fact that peptidoglycan fragments released by this muramidase activity can be recognized by pattern recognition receptors crucial in innate immunity (82). Lysozyme is a relatively small (14.5-kDa) and basic (pI 9.3) protein that is produced by numerous different cells and is present in many human bodily fluids (for a review, see reference 83). Humans produce c-type lysozyme, so named for chickens, which are the source of the hen egg white lysozyme widely used in biological and biochemical studies. There are also g-type (“g” for goose) and i-type (“i” for invertebrate) lysozymes; the amino acid identities among the three types is low (16% to 24%) (for a review, see reference 84).
Evidence for the protective effect of lysozyme against bacterial infection in the lower airways was first obtained in experiments where the production of lysozyme was either ablated or enhanced in transgenic animals (85–87). More recent experiments with Streptococcus pneumoniae in mice (88, 89) showed that lysozyme is also an important component of innate immune defense in the upper respiratory tract. Other experiments involving nasal fluids (90) or airway secretions (91) suggested the involvement of lysozyme in antibacterial activity, and the use of purified lysozyme against a selected panel of airway pathogens (i.e., S. pneumoniae, Haemophilus influenzae, and M. catarrhalis) showed that these bacteria differed in their resistance to growth inhibition by lysozyme (92). However, the antibacterial effect of lysozyme is not solely dependent on its ability to cleave peptidoglycan (93). Catalytically inactive lysozyme was shown to still be able to kill some Gram-positive bacteria in vitro (93), and small (e.g., 9- to 15-aa) peptides from lysozyme were shown to cross the Gram-negative bacterial outer membrane and disrupt inner membrane function (94). A subsequent study showed that the muramidase activity was not essential for the bactericidal activity of lysozyme in vivo (95).
The concentration of lysozyme in human airway fluids is impressively high, on the order of 500 μg/ml (90, 96), and bacteria have evolved two very different mechanisms for protecting themselves against lysozyme. The first mechanism is used by both Gram-positive and Gram-negative bacteria and involves the modification of peptidoglycan structure, typically through O-acetylation of N-acetylmuramic acid or N-deacetylation of N-acetyl-d-glucosamine residues (97). The second method is restricted to Gram-negative bacteria and involves the production of proteins that directly inhibit lysozyme enzymatic activity. The first of these to be identified was the E. coli Ivy (i.e., inhibitor of vertebrate lysozyme) protein (98) that inhibits only c-type lysozymes. Four other families of lysozyme inhibitors were subsequently identified; these differ in their localization (i.e., periplasmic [PliC] versus membrane bound [MliC]) and in their specificity for the different lysozyme types (i.e., c type, g type, or i type) (71).
Whereas it may seem counterintuitive that lysozyme can breach the outer membrane and attack the peptidoglycan layer, evidence for the importance of lysozyme inhibitors in vivo was first obtained with a Gram-negative organism. An MliC-deficient mutant of an avian-pathogenic E. coli strain exhibited reduced virulence in a chick model of lethal infection (99). Subsequent to this study, lysozyme inhibitor-deficient mutants of both Yersinia pestis (100) and Edwardsiella tarda (101) were shown to have reduced virulence in mouse and fish infection models, respectively, reinforcing the relevance of lysozyme to innate immune defense.
In the present study, we showed that the M. catarrhalis LipA and LipB proteins can inhibit both purified human lysozyme (Fig. 6A) and lysozyme present in human saliva (Fig. 6C). Furthermore, inactivation of both lipA and lipB resulted in a M. catarrhalis mutant that was killed more readily than its wild-type parent strain by purified human lysozyme (Fig. 6D). These lysozyme killing experiments were performed in the presence of human apolactoferrin, another important component of airway fluids (102) that can increase the sensitivity of Gram-negative bacteria to killing by lysozyme (47, 72, 103) by affecting the permeability of the outer membrane (74). Among the 12 M. catarrhalis strains whose genomes have been sequenced (35, 36), LipA proteins have 99 to 100% identity, with LipB proteins being 97 to 100% identical. Expression of these two genes was controlled, directly or indirectly, by the MesRS TCS described in the present report, and the effect of the mesR mutation on upregulation of these genes was shown to not be strain specific (Fig. 3A and 7). While the data in the present study indicate that both lipA and lipB are upregulated in the absence of MesR, we were not able to demonstrate a specific interaction between MesR and the promoter regions of lipA and lipB by an electrophoretic mobility shift assay (data not shown). Therefore, it is possible that other control elements may influence the production of these inhibitor proteins and that they could be regulated differently.
The M. catarrhalis LipA and LipB proteins appear to have structures that most closely resemble the structure of the periplasmic PliC lysozyme inhibitor produced by S. Typhimurium (47). This supposition is based on three observations: (i) the hidden Markov matches of LipA and LipB are closer to PliC than to the membrane-bound MliC lysozyme inhibitors; (ii) PliC proteins tend to have a longer loop (loop 5) between β strands 5 and 6 (65), and this region aligns well with both LipA and LipB (Fig. 5); and (iii) the MliC proteins are lipoproteins, but there is no apparent lipidation signal in either LipA or LipB. Thus, we expect that LipA and LipB have tertiary structures that are similar to the tertiary structure of S. Typhimurium PliC; however, because of the low identities between these two Moraxella proteins and this PliC protein (ca. 21%), the probability of building an accurate model of LipA or Lip 2 is low. Furthermore, LipA and LipB both have a loop 6 (between β strands 6 and 7) that is longer than the corresponding structure in S. Typhimurium PliC and represents a significant deviation from the PliC structure.
Despite the high probability that the folds of LipA and LipB most closely resemble the fold of the S. Typhimurium PliC protein, the structure of the B. abortus lysozyme inhibitor PliC (ca. 20% sequence identity to both LipA and LipB) deserves special scrutiny. This is because the B. abortus PliC protein and LipA may specifically inhibit human lysozyme. A cocrystal structure revealed the residues that are important for the B. abortus PliC protein to contact human lysozyme (68). Our alignments suggest that only 1 or 2 of these residues is conserved between the M. catarrhalis proteins and this B. abortus protein. Indeed, the residues that align with the human lysozyme-contacting K97 residue of B. abortus PliC are aspartates (D102 and D88, respectively) in LipA and LipB; this represents an inversion of the charge at these putative contact points (Fig. 5). These observations imply that LipA and LipB may employ a unique mode of human lysozyme inhibition.
Interestingly, the inhibitory activity of LipA (the more potent of these two M. catarrhalis proteins) did not extend to hen egg white lysozyme (Fig. 6B), the type of lysozyme almost universally used in studies of bacterial lysozyme inhibitors (68). The paucity of commercially available purified lysozyme proteins from other animal species makes a survey of the specificity of the inhibitory activity of LipA and LipB very difficult, if not impossible. A BLAST search for lysozyme proteins revealed that the chinchilla (Chinchilla lanigera) genome encodes both a predicted c-type lysozyme and a predicted g-type lysozyme. The potential expression of two different lysozymes by this rodent makes it problematic to use the chinchilla model for nasopharyngeal colonization by M. catarrhalis (21, 104) to evaluate the different lysozyme inhibitor mutants constructed in the present study.
The three bacterial organisms most frequently involved in causing acute otitis media are S. pneumoniae, nontypeable H. influenzae, and M. catarrhalis (9). As noted above, proteinaceous lysozyme inhibitors are not produced by Gram-positive organisms, which, instead, escape lysis by lysozyme by altering their peptidoglycan structure (97). To date, there are no published reports describing lysozyme inhibitors produced by H. influenzae. Another Gram-negative pathogen that can asymptomatically colonize the human nasopharynx is Neisseria meningitidis, but a recent study indicated that the genomes of Neisseria species do not encode lysozyme inhibitors (71). The apparent rarity of lysozyme inhibitors produced by other important Gram-negative pathogens in this environmental niche makes the production of two different lysozyme inhibitors by M. catarrhalis even more remarkable. Demonstration of the relative importance of the LipA and LipB proteins in vivo must necessarily await appropriate experimentation.
ACKNOWLEDGMENTS
This study was supported by U.S. Public Health Service grant no. AI036344 to E.J.H. S.N.J. was supported by U.S. Public Health Service training grant no. 5-T32-AI007520 and 5-T32-AI005284.
We thank Wei Liu for expert technical assistance.
FOOTNOTES
- Received 15 August 2014.
- Returned for modification 14 September 2014.
- Accepted 9 October 2014.
- Accepted manuscript posted online 13 October 2014.
Supplemental material for this article may be found at http://dx.doi.org/10.1128/IAI.02486-14.
- Copyright © 2015, American Society for Microbiology. All Rights Reserved.
REFERENCES
- 1.↵
- 2.↵
- 3.↵
- 4.↵
- 5.↵
- 6.↵
- 7.↵
- 8.↵
- 9.↵
- 10.↵
- 11.↵
- 12.↵
- 13.↵
- 14.↵
- 15.↵
- 16.↵
- 17.↵
- 18.↵
- 19.↵
- 20.↵
- 21.↵
- 22.↵
- 23.↵
- 24.↵
- 25.↵
- 26.↵
- 27.↵
- 28.↵
- 29.↵
- 30.↵
- 31.↵
- 32.↵
- 33.↵
- 34.↵
- 35.↵
- 36.↵
- 37.↵
- 38.↵
- 39.↵
- 40.↵
- 41.↵
- 42.↵
- 43.↵
- 44.↵
- 45.↵
- 46.↵
- 47.↵
- 48.↵
- 49.↵
- 50.↵
- 51.↵
- 52.↵
- 53.↵
- 54.↵
- 55.↵
- 56.↵
- 57.↵
- 58.↵
- 59.↵
- 60.↵
- 61.↵
- 62.↵
- 63.↵
- 64.↵
- 65.↵
- 66.↵
- 67.↵
- 68.↵
- 69.↵
- 70.↵
- 71.↵
- 72.↵
- 73.↵
- 74.↵
- 75.↵
- 76.↵
- 77.↵
- 78.↵
- 79.↵
- 80.↵
- 81.↵
- 82.↵
- 83.↵
- 84.↵
- 85.↵
- 86.↵
- 87.↵
- 88.↵
- 89.↵
- 90.↵
- 91.↵
- 92.↵
- 93.↵
- 94.↵
- 95.↵
- 96.↵
- 97.↵
- 98.↵
- 99.↵
- 100.↵
- 101.↵
- 102.↵
- 103.↵
- 104.↵
- 105.↵
- 106.↵
