|Home | About | Journals | Submit | Contact Us | Français|
Moraxella catarrhalis O35E was shown to synthesize a 105-kDa protein that has similarity to both acid phosphatases and autotransporters. The N-terminal portion of the M. catarrhalis acid phosphatase A (MapA) was most similar (the BLAST probability score was 10−10) to bacterial class A nonspecific acid phosphatases. The central region of the MapA protein had similarity to passenger domains of other autotransporter proteins, whereas the C-terminal portion of MapA resembled the translocation domain of conventional autotransporters. Cloning and expression of the M. catarrhalis mapA gene in Escherichia coli confirmed the presence of acid phosphatase activity in the MapA protein. The MapA protein was shown to be localized to the outer membrane of M. catarrhalis and was not detected either in the soluble cytoplasmic fraction from disrupted M. catarrhalis cells or in the spent culture supernatant fluid from M. catarrhalis. Use of the predicted MapA translocation domain in a fusion construct with the passenger domain from another predicted M. catarrhalis autotransporter confirmed the translocation ability of this MapA domain. Inactivation of the mapA gene in M. catarrhalis strain O35E reduced the acid phosphatase activity expressed by this organism, and this mutation could be complemented in trans with the wild-type mapA gene. Nucleotide sequence analysis of the mapA gene from six M. catarrhalis strains showed that this protein was highly conserved among strains of this pathogen. Site-directed mutagenesis of a critical histidine residue (H233A) in the predicted active site of the acid phosphatase domain in MapA eliminated acid phosphatase activity in the recombinant MapA protein. This is the first description of an autotransporter protein that expresses acid phosphatase activity.
Initially thought to be a harmless commensal organism, Moraxella catarrhalis has gradually gained repute as an etiologic agent of at least two significant diseases in humans. This gram-negative, unencapsulated bacterium has been shown to colonize the upper airways of infants and very young children (14, 15) and is one of the three most prominent causes of otitis media (45). Additionally, adults with chronic obstructive pulmonary disease are at risk for infectious exacerbations caused by M. catarrhalis (45, 66). A recent study indicates that each year in the United States, as many as four million chronic obstructive pulmonary disease exacerbations may be attributed to M. catarrhalis (46).
The secretion of proteins by gram-negative bacteria is a function necessary for numerous metabolic and physiologic processes. Five different secretion systems have been well characterized in bacteria (13), and a sixth has recently been described (44, 56). The type V secretion system has received increased attention in recent years (11, 27, 35, 38). The absence of a requirement for energy coupling or accessory factors for successful protein secretion has resulted in this class of proteins being described as autotransporters. In gram-negative bacteria, autotransporters make up the largest family of outer membrane porins involved in protein translocation (12). The autotransporter secretion system was first described for the immunoglobulin A1 protease of Neisseria gonorrhoeae (54, 55), and subsequently, numerous autotransporters have been described for other gram-negative bacteria (25, 27). A three-domain model for type V secretion systems has emerged, comprising (i) an amino-terminal leader peptide or signal sequence, (ii) the secreted mature protein (or passenger domain), and (iii) a C-terminal translocation domain responsible for the formation of a pore in the outer membrane to allow passage of the passenger domain to the cell surface (26).
The passenger domains of previously described autotransporter systems have been shown to have widely different functions in gram-negative bacteria, including but not limited to proteolytic, adhesive, and cytotoxic activities (27). M. catarrhalis has been shown to synthesize at least three proteins (i.e., UspA1, UspA2, and Hag) that have been classified as trimeric autotransporters and one additional protein that is considered a conventional autotransporter (i.e., McaP) (for reviews, see references 11, 19, and 35). These four previously characterized M. catarrhalis autotransporters have been shown to be involved in adherence (1), serum resistance (6), binding of immunoglobulin D (18), autoaggregation (51), and lipolysis (68).
Acid phosphatases catalyze the hydrolysis of phosphomonoesters at an acidic pH (9). Bacterial nonspecific acid phosphatases (NSAP) are subdivided into three classes (for a review, see reference 67). Class A acid phosphatases typically are secreted enzymes with broad substrate specificity and usually consist of relatively small (25 kDa to 27 kDa) polypeptides that may occur as monomers or oligomers. In contrast, class B acid phosphatases are polymeric metalloproteins that also demonstrate broad substrate specificity. Class C acid phosphatases are the most recently identified and are lipoproteins that bear some resemblance to class B acid phosphatases. Typically, the enzymatic activity of an acid phosphatase helps to catalyze the two-step hydrolysis of phosphomonoesters. The enzymes utilize a covalent inorganic phosphoenzyme complex as a short-lived intermediate (71).
The biological function of these bacterial nonspecific acid phosphatases has been defined in only a few instances. It has been assumed that these enzymes are involved in cleaving organic phosphoesters, such as nucleotides and sugar phosphates, into inorganic phosphate and a dephosphorylated product that can be transported across the cytoplasmic membrane (63). In addition, it has been suggested that intracellular pathogens can use acid phosphatases to affect signaling pathways controlling the respiratory burst (58, 64). In Salmonella enterica, the periplasmic AphA acid phosphatase removes phosphate from the nicotinamide mononucleotide so that the resultant nicotinamide ribonucleoside can be taken up by a transport system in the cytoplasmic membrane (21). The Haemophilus influenzae lipoprotein e (P4) has been reported to utilize NADP as a substrate (57) and has been studied in great detail (34, 49, 59, 61). A Legionella pneumophila acid phosphatase was shown to be secreted by a type II secretion system but was reported to not be essential for intracellular replication of L. pneumophila in macrophages (4). Most recently, a respiratory burst-inhibiting acid phosphatase from Francisella tularensis has been crystallized and studied in considerable detail (7, 17, 17, 58, 60).
In the present report, we describe the identification and characterization of a novel M. catarrhalis autotransporter (MapA) that exhibits acid phosphatase activity. When expressed in recombinant form in Escherichia coli, the mapA gene product retained its acid phosphatase activity. Site-directed mutagenesis was used to confirm the identity of the active site in the acid phosphatase domain, and the functional ability of the C-terminal translocation domain was proven experimentally. We demonstrated that this autotransporter protein is localized to the M. catarrhalis outer membrane and is highly conserved among strains of this pathogen. To the best of our knowledge, this is the first detailed description of an autotransporter protein that expresses acid phosphatase activity.
The bacterial strains and plasmids used in this study are listed in Table Table1.1. M. catarrhalis strains were grown as described previously (5) using brain heart infusion (BHI) broth (Difco/Becton Dickinson, Sparks, MD); this medium was supplemented with spectinomycin (15 μg/ml) when appropriate. E. coli strains were grown on Luria-Bertani (LB) medium as described previously (65); this medium was supplemented with either ampicillin (100 μg/ml) or spectinomycin (100 μg/ml) when required.
LB agar was supplemented with phenolphthalein diphosphate (1 mg/ml) and methyl green (50 μg/ml) as an indicator of phosphatase activity (62). Phosphatase activity was evidenced by the appearance of green colonies on this medium, whereas phosphatase-negative colonies appeared yellow or colorless.
The predicted amino acid (aa) sequence of the MapA protein from M. catarrhalis strain O35E was used to design a small oligopeptide for immunization purposes. A 20-aa sequence (KAYDGISHIYQDIETTTQDK; aa 378 to 397 from the O35E MapA protein) was synthesized with a cysteine residue added to the N terminus of this peptide. This peptide was covalently coupled to Imject maleimide-activated mariculture keyhole limpet hemocyanin (Pierce, Rockford, IL) and used to immunize mice. The hybridoma fusion procedure was carried out by the Monoclonal Antibody Center at UT Southwestern Medical Center. The MapA-reactive monoclonal antibody (MAb) 1H12 was identified by screening hybridoma culture supernatant fluids in an enzyme-linked immunosorbent assay using the MapA-derived peptide described above as the antigen.
Agar plate-grown M. catarrhalis and E. coli cells were used to prepare whole-cell lysates as described previously (37). Outer membrane vesicles were prepared from broth-grown M. catarrhalis cells as described previously (47). Concentrated culture supernatant fluids were prepared from overnight broth cultures of M. catarrhalis as described previously (73), except that the Amicon Ultra centrifugal filter device (Millipore Corp., Billerica, MA) used in this procedure had a nominal cutoff of 10,000 Da. Western blot analysis was performed as described previously (5) by using MAb 1H12 to detect MapA and MAb 10F3 to detect the CopB outer membrane protein (22). The lipooligosaccharide (LOS) present in proteinase K-treated whole-cell lysates was resolved by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and stained with silver (69) or probed by Western blot analysis with the M. catarrhalis LOS-specific MAb 8E7 as described previously (6).
Acid phosphatase activity was measured by using p-nitrophenyl phosphate (PnPP) (Sigma Chemical, St. Louis, MO) as the substrate in an assay procedure modified from that described previously by Reilly and colleagues (60). Recombinant E. coli strains were grown overnight in LB broth supplemented with ampicillin. M. catarrhalis strains were grown overnight in BHI broth. The bacterial suspensions were subjected to centrifugation at 7,600 × g for 8 min. The bacterial cell pellet was suspended in 0.4 M sodium acetate buffer (pH 6.0) to obtain a final suspension that yielded a reading of 300 Klett U on a Klett-Summerson colorimeter (VWR Scientific). A 5-ml portion of this suspension was then subjected to centrifugation at 1,860 × g for 5 min. The final cell pellet was suspended in 1 ml of the sodium acetate buffer. Portions (100 μl) of this cell suspension and serial 10-fold dilutions were added in duplicate into individual wells in a 96-well, clear-bottom, black-sided microtiter plate (Corning International, Corning, NY). The standard assay mixture consisted of 100 μl of bacterial suspension, 80 μl H2O, and 20 μl of the PnPP substrate (20 mM stock solution). The reaction was allowed to proceed for 30 min at 37°C. The microtiter plate was then subjected to centrifugation at 96 × g for 5 min. A 100-μl portion of the resultant supernatant fluid was then transferred to an empty well, and a 100-μl volume of 0.5 M glycine (pH 10.0) was added to the well to stop residual enzyme activity. The absorbance was then measured at 405 nm in a SpectraFluor Plus microplate reader (Tecan, Research Triangle Park, NC). Analysis of variance methods were used for statistical analysis of enzyme activity levels (36).
A 500-ml overnight culture of M. catarrhalis cells was harvested by centrifugation and suspended in 6 ml of phosphate-buffered saline (PBS). This very dense suspension was sonicated using a Branson sonifier (model 450; Branson Ultransonics, Danbury, CT) for three 1-min cycles at an output of 50% with a duty cycle of 5. The sonicate was then subjected to centrifugation at 12,000 × g for 10 min at 4°C to remove whole cells and large debris, and the resultant supernatant fluid was transferred to another centrifuge tube. This supernatant fluid was subjected to centrifugation at 100,000 × g for 90 min at 4°C. The final supernatant fluid was carefully removed by pipetting without disturbing the pellet; this fluid represented the soluble cytoplasmic fraction. The pellet was resuspended in PBS and represented the cell envelope fraction. The well-established method of Murphy and Loeb (47) was used to prepare outer membrane vesicles from broth-grown M. catarrhalis cells.
Standard molecular biology and recombinant DNA techniques were performed as described previously (65). ExTaq DNA polymerase (PanVera, Madison, WI) was used for PCR-based amplification of DNA fragments for cloning purposes and for overlapping extension PCR (28). Taq DNA polymerase (New England Biolabs) was used for colony PCRs. The Easy-DNA kit (Invitrogen, Carlsbad, CA) was used to prepare chromosomal DNA from M. catarrhalis. The Qiaprep Spin Miniprep kit (Qiagen, Valencia, CA) was used to purify plasmid DNA from both E. coli and M. catarrhalis.
The oligonucleotide primers AT01 (5′-ATAGGATCCGCACCAGCCTCATCAAAT-3′, with the BamHI site underlined) and AT02 (5′-AATGGATCCTTGTGCCAGTGCCATTT-3′, with the BamHI site underlined) were used in PCR with chromosomal DNA from M. catarrhalis O35E to obtain a 3.5-kb fragment containing the mapA gene, which was ligated into plasmid pCR2.1 (Invitrogen, Carlsbad, CA) and used to transform E. coli TOP10 cells (Invitrogen). An ampicillin-resistant transformant was shown to contain the desired recombinant plasmid, which was designated pKW04. As a negative control, a 108-nucleotide (nt) fragment of M. catarrhalis O12E chromosomal DNA, carrying part of the promoter region of the uspA2 gene and part of the uspA2 open reading frame (ORF), was cloned into pCR2.1 and designated pTH10. Similarly, the mapA genes from four other M. catarrhalis strains (7169, ETSU-9, O12E.44, and V1120) were amplified by PCR for nucleotide sequence analysis.
The primers TH22 (5′-TTGGATCCGACCTGCCAGCACGATCAAG-3′, with the BamHI site underlined) and TH20 (5′-TGA CTTGTCACGCCCGGGCATCAAGATGTTGATACC-3′, with the SmaI site underlined) were used to amplify a 1.2-kb fragment containing the extreme 5′ end of the O35E mapA gene and upstream flanking DNA, whereas primers TH21 (5′-GGGCGTGACAAGTCAAATCAA-3′, with a partial SmaI site underlined) and TH23 (5′-TACCGAGCTCGATGATAACGGGCGTGTA-3′, with the SacI site underlined) were used to obtain a 0.8-kb fragment containing the extreme 3′ end of the mapA ORF and downstream flanking DNA. These two amplicons were mixed and used as the templates in overlapping extension PCR (28, 29) with primers TH22 and TH23 to generate an approximately 2.0-kb fragment that was cloned into pCR2.1 and used to transform E. coli TOP10; recombinant clones were identified by blue-white screening on LB agar containing X-Gal (5-bromo-4-chloro-3-indolyl-β-d-galactopyranoside) (30 μg/ml). Colony PCR was used to confirm the presence of the 2-kb amplicon, and one of these recombinant plasmids, designated pTH03, was selected for further use. This plasmid was digested with SmaI, and the spectinomycin resistance cassette from pSPECR (74) was blunt end ligated into this site to obtain plasmid pTH04. Plasmid pTH04 was used to transform the wild-type M. catarrhalis strain O35E as described previously (52), and transformants were selected for spectinomycin resistance. One of these transformants, designated O35EΔmapA, was selected for further study. Additional mapA deletion mutants were constructed in M. catarrhalis strains 7169 and ETSU-9 by the same method.
The use of plasmid pWW115 (72) for complementation analysis requires spectinomycin as the selective marker. This necessitated the reconstruction of the M. catarrhalis O35EΔmapA mutant to eliminate the spectinomycin resistance cassette. M. catarrhalis O35EΔmapA was used in the plate transformation system (52) with the 2-kb amplicon described above (i.e., that obtained with primers TH22 and TH23), which had the large mapA deletion. Potential transformants were patched onto both BHI agar and BHI-spectinomycin agar to identify those transformants in which allelic exchange had resulted in the removal of the spectinomycin resistance cassette. Subsequent nucleotide sequence analysis of the relevant chromosomal region from one of these transformants, designated O35EΔmapA-9, confirmed the successful removal of the spectinomycin resistance cassette from the disrupted mapA gene.
Chromosomal DNA from M. catarrhalis strain O35E was used in PCR with primers AT01 and TH31 (5′-CAAGAGCTCATCTATCGGCAGATGCTTGAAT-3′, with the SacI site underlined) to obtain a 3.4-kb amplicon containing the entire mapA gene. This amplicon was digested with both BamHI and SacI and then ligated into pWW115 (72), which had also been digested with the same two enzymes. This ligation reaction mixture was used to electroporate M. catarrhalis O35E, and transformants were selected on BHI-spectinomycin agar. A plasmid from one of these transformants was shown to contain the mapA amplicon and was designated pTH13.
M. catarrhalis O35EΔmapA-9 was electroporated with either pTH13 or pWW115 (i.e., vector-only control). Spectinomycin-resistant recombinant clones were selected on BHI-spectinomycin agar. Nucleotide sequence analysis of the plasmid recovered from this recombinant strain [M. catarrhalis O35EΔmapA-9(pTH13)] confirmed that pTH13 was successfully cloned into this mutant. Additional nucleotide sequence analysis of the relevant chromosomal region of this complemented mutant confirmed the presence of the mapA deletion.
Approximately 1.5 kb of DNA encoding the leader peptide, predicted passenger (lipase) domain, and 17 residues of the predicted α-helical linker region of the McaP protein (39) of M. catarrhalis strain O35E was amplified by PCR using the primers TH75 (5′-GCGCGGATCCCATTGCGGTAACT-3′, with the BamHI site underlined) and TH77 (5′-GCTTTGTTGACCATGTTTAATCAGA-3′), with O35E chromosomal DNA as the template. A second PCR utilizing primers TH78 (5′-CATGGTCAACAAAGCATGGGAAGCTTGTATACATTA-3′) and TH76 (5′-GCGCGAGCTCCGCAGACACAGAA-3′, with the SacI site underlined) was performed to amplify the nucleotide sequence (approximately 1.1 kb) encoding the entire predicted translocation module (i.e., autotransporter domain) of the O35E MapA protein, together with 10 aa immediately upstream of this module. These two amplicons were mixed and used as the templates in overlapping extension PCR with primers TH75 and TH76 to generate an approximately 2.6-kb fragment. This amplicon was subsequently digested with both SacI and BamHI and ligated into pWW115 (72). This ligation mixture was used to electroporate the M. catarrhalis mcaP mutant O35E.M (39), and transformants were selected on BHI-spectinomycin agar. A plasmid isolated from one of these transformants was shown to contain the desired fusion product and was designated pTH34. Nucleotide sequence analysis of the 2.6-kb DNA insert in the pTH34 plasmid indicated a few nucleotide changes, one of which caused a single predicted amino acid change (i.e., Y105H) in the McaP passenger domain. There were no predicted amino acid changes in the MapA autotransporter domain.
Detection of the McaP passenger domain on the surface of the recombinant strain M. catarrhalis O35E.M(pTH34) was accomplished by using polyclonal murine antibodies to aa 51 to 650 of the O35E McaP protein (39) as the primary antibody in flow cytometry. Briefly, bacterial cells were grown overnight on BHI agar plates and suspended in PBS to an optical density at 600 nm of 0.35. Portions (100 μl) of these suspensions were subjected to centrifugation at 18,500 × g and resuspended in 100 μl of PBS containing 1% (wt/vol) bovine serum albumin (PBS-BSA) in which the polyclonal antibody to the McaP protein had been diluted 1:60. These tubes were incubated at room temperature for 20 min and then washed three times with 500 μl of PBS-BSA. The bacteria were then incubated in the dark with 1 μg of fluorescein isothiocyanate-conjugated goat anti-mouse antibody (Abcam, Cambridge, MA) for 20 min at room temperature. The bacteria were next washed three times with 500 μl PBS-BSA, resuspended in 1 ml of filter-sterilized PBS, and analyzed by flow cytometry utilizing a FACScan flow cytometer (Becton Dickinson).
The QuikChange II site-directed mutagenesis kit (Stratagene, La Jolla, CA) was utilized to generate a specific mutation in the putative active site of the M. catarrhalis MapA protein. By using pKW04 as the template, we used primers TH48 (5′-AGCCGAGTCATTGTGGGTGCGGCTTTTCCAACAGATACCATGACTTCT-3′, with altered bases underlined) and TH49 (5′-AGAAGTCATGGTATCTGTTGGAAAAGCCGCACCCACAATGACTCGGCT-3′, with altered bases underlined) in a PCR to change the histidine at position 233 in the MapA protein to an alanine. The plasmid encoding the H233A mutation was designated pTH24, and nucleotide sequence analysis confirmed that this was the only change. In addition, as a control, a second histidine was mutated at a position predicted to be unlikely to be involved in the enzymatic activity of the MapA protein. The H110A change was accomplished by using pKW04 as the template, together with primers TH52 (5′-CCTTAAAGGAGTTTGCTCCGGCTATAACTGATGAACAGTTTGTAAATATC-3′, with altered bases underlined) and TH53 (5′-GATATTTACAAACTGTTCATCAGTTATAGCCGGAGCAAACTCCTTTAAGG-3′, with altered bases underlined). The plasmid encoding the H110A mutation was designated pTH26.
The nucleotide sequences of these mapA genes were deposited in GenBank and assigned the following accession numbers: EF186006 (for O35E), EF192600 (for 7169), EF192601 (for ETSU-9), EF192599 (for O12E.44), and EF192602 (for V1120).
Previous efforts in this laboratory involved the M. catarrhalis Hag protein (51, 52), which is encoded by a gene located in contig 32 (GenBank accession number AX067457) from the sequenced genome of M. catarrhalis ATCC 43617. An examination of the chromosomal loci flanking the hag gene revealed the presence of an ORF located approximately 2 kb upstream from hag, which encoded a predicted autotransporter protein (Fig. (Fig.1).1). More extensive BLAST-based analyses revealed that the N-terminal region of this predicted autotransporter protein resembled several bacterial nonspecific acid phosphatases (described in detail below). We tentatively designated this protein as M. catarrhalis acid phosphatase A (MapA).
The mapA ORF contained 2,820 nt encoding a predicted protein with 940 aa. The mapA gene was flanked upstream by an ORF, designated prmA, encoding a predicted ribosomal protein L11 methyltransferase (70) and downstream by a gene encoding a predicted penicillin-binding protein 1B (Fig. (Fig.1).1). The mapA gene appears not to be transcriptionally linked to any other ORFs. Nucleotide sequence analysis of the mapA gene from five additional M. catarrhalis strains showed that the encoded proteins were very highly conserved with 97% identity (Fig. (Fig.2).2). Hydrophobicity analysis of the N-terminal region and the use of the SignalP predictor program (SignalP V2.0.b2 [www.cbs.dtu.dk/services/SignalP/]) showed the likely presence of a signal peptide with a predicted signal peptidase I cleavage site between residues 23 and 24 (Fig. (Fig.2).2). The calculated mass of the predicted mature MapA protein was 103 kDa.
To confirm the presence of acid phosphatase activity in the MapA protein, the mapA gene from M. catarrhalis O35E was cloned into E. coli, yielding the recombinant plasmid pKW04. When E. coli DH5α(pKW04) was streaked onto phenolphthalein-methyl green agar, this recombinant strain produced the deep green color associated with phosphatase activity (Fig. (Fig.3A,3A, lane 1). In contrast, when a control strain containing pCR2.1 with an irrelevant M. catarrhalis DNA insert (i.e., pTH10) was streaked onto the same medium, it did not produce a color change (Fig. (Fig.3A,3A, lane 2). Additional enzymatic analysis using a liquid-phase assay with whole cells of E. coli DH5α(pKW04) (Fig. (Fig.3B,3B, bar 1) showed the presence of acid phosphatase activity at levels at least 10-fold greater than those obtained with the negative control strain (Fig. (Fig.3B,3B, bar 2) (P < 0.0001).
To facilitate investigation of the expression of MapA in its native background, we constructed the O35EΔmapA deletion mutant as described in Materials and Methods. PCR and nucleotide sequence analysis were used to confirm the presence of the desired deletion mutation (Fig. (Fig.4A).4A). Western blot analysis using the MapA-specific MAb 1H12 showed that the wild-type parent strain (Fig. (Fig.4B,4B, lane 1) expressed an antigen that bound this MAb and migrated just below the 105-kDa marker, whereas the O35EΔmapA mutant (Fig. (Fig.4B,4B, lane 2) did not. Use of the liquid-phase enzymatic assay with whole cells of these two strains showed that the wild-type M. catarrhalis parent strain (Fig. (Fig.4C,4C, bar 1) had levels of acid phosphatase activity that were higher than those expressed by the O35EΔmapA mutant (Fig. (Fig.4C,4C, bar 2) (P < 0.0001). The protein compositions of outer membrane vesicles obtained from both the wild-type parent strain and the O35EΔmapA mutant appeared to be the same, as determined by SDS-PAGE, followed by Coomassie blue staining. It must be noted that the MapA protein is expressed at a level that could not be detected by staining with Coomassie blue (data not shown). Both silver staining and Western blot analysis with the M. catarrhalis LOS-reactive MAb 8E7 (30) showed no apparent difference in the LOS expressed by these two strains (data not shown). Similarly, there were no apparent differences in the growth rates of the wild-type parent strain and this mutant in BHI broth (data not shown).
The wild-type mapA gene from M. catarrhalis O35E was cloned into pWW115 to obtain pTH13. However, this recombinant plasmid containing the mapA gene could not be used with the O35EΔmapA mutant because this mutant contained the spectinomycin resistance cassette. Therefore, transformation and allelic exchange were used to excise the spectinomycin resistance cassette from O35EΔmapA, resulting in a new mapA deletion mutant designated O35EΔmapA-9. When pTH13 was introduced into O35EΔmapA-9, the resultant recombinant M. catarrhalis strain expressed the MapA protein (Fig. (Fig.4B,4B, lane 3) at a level much greater than that expressed by the wild-type parent strain (Fig. (Fig.4B,4B, lane 1). A control construct in which the pWW115 plasmid was introduced into the O35EΔmapA-9 mutant did not express any MapA protein (Fig. (Fig.4B,4B, lane 4). As expected, acid phosphatase activity expressed by the recombinant strain containing the cloned mapA gene in trans (Fig. (Fig.4C,4C, bar 3) was much greater than that expressed by the same mutant containing the empty plasmid vector (Fig. (Fig.4C,4C, bar 4) (P < 0.0001).
MAb 1H12 was used to detect the presence of MapA in other M. catarrhalis isolates. The MapA protein was readily detectable in whole-cell lysates of these 14 additional strains (Fig. (Fig.5A,5A, lanes 3 to 16), which included both strains isolated from patients with disease and strains isolated from the nasopharynxes of healthy individuals. The positive and negative controls in this screen included the wild-type M. catarrhalis strain O35E (Fig. (Fig.5A,5A, lane 1) and the O35EΔmapA mutant (Fig. (Fig.5A,5A, lane 2).
When whole-cell lysates and cell envelope preparations of M. catarrhalis O35E and the M. catarrhalis O35EΔmapA mutant were probed by Western blot analysis with MAb 1H12, the MapA antigen was readily detectable in the wild-type parent strain (Fig. (Fig.5B,5B, lanes 1 and 3), but not in the mapA deletion mutant (Fig. (Fig.5B,5B, lanes 2 and 4). We were unable to detect the presence of the MapA protein either in the soluble cytoplasmic fraction derived from the wild-type parent strain (Fig. (Fig.5B,5B, lane 5) or in the concentrated culture supernatant fluid of the same strain (Fig. (Fig.5B,5B, lane 7).
The localization of the MapA protein to the cell envelope fraction (Fig. (Fig.5B,5B, lane 3), together with its predicted autotransporter domain (described above), indicated that this protein should be present in the outer membrane of M. catarrhalis. To address this localization issue directly, outer membrane vesicles were extracted from three wild-type M. catarrhalis strains (O35E, ETSU-9, and 7169) and their corresponding mapA deletion mutants as described in Materials and Methods. Western blot analysis with MAb 1H12 showed that the MapA protein was present in the outer membrane vesicles from wild-type O35E (Fig. (Fig.6,6, lane 1), ETSU-9 (Fig. (Fig.6,6, lane 3), and 7169 (Fig. (Fig.6,6, lane 5) and missing from outer membrane vesicles of the mapA deletion mutants (Fig. (Fig.6,6, lanes 2, 4, and 6, respectively) constructed from these three wild-type strains.
To prove directly that MapA was an autotransporter protein, the predicted MapA translocation domain (designated AT [Fig. [Fig.7A])7A]) was used to replace the predicted translocation domain of the McaP protein (39) as described in Materials and Methods. The resultant mcaP-mapA fusion construct, encoding the McaP passenger domain (designated lipase [Fig. [Fig.7A])7A]) and the MapA translocation domain, was cloned into pWW115 to obtain pTH34 and expressed in the M. catarrhalis mcaP mutant O35E.M. When probed by Western blot analysis with a polyclonal antiserum previously shown to bind surface epitopes of McaP (39), the wild-type O35E strain (Fig. (Fig.7B,7B, lane 1) expressed a readily detectable McaP protein. Neither the mcaP mutant (Fig. (Fig.7B,7B, lane 2) nor the mcaP mutant containing the vector pWW115 (Fig. (Fig.7B,7B, lane 3) expressed any detectable protein reactive with this antiserum. In contrast, the recombinant M. catarrhalis strain O35E.M(pTH34) (Fig. (Fig.7B,7B, lane 4) expressed an McaP-MapA fusion protein of the predicted size. To confirm that the McaP passenger domain of this fusion protein had been translocated to the surface of M. catarrhalis by the MapA translocation domain, the same four strains were subjected to flow cytometry using the McaP polyclonal antiserum as the primary antibody. Both the wild-type strain O35E (Fig. (Fig.7C,7C, panel 1) and the mcaP mutant expressing the McaP-MapA fusion protein (Fig. (Fig.7C,7C, panel 4) showed similar levels of positive reactivity with these polyclonal antibodies, whereas the mcaP mutant (Fig. (Fig.7C,7C, panel 2) and the mcaP mutant containing the vector pWW115 (Fig. (Fig.7C,7C, panel 3) exhibited very little reactivity with the same antibodies.
An analysis of the deduced amino acid sequence of MapA from M. catarrhalis ATCC 43617 using BLAST (3) revealed that the N-terminal 273 aa of this protein have homology with several bacterial orthologs previously identified as class A NSAP. Of the top eight homologous sequences, BLAST probability scores ranged from 8 × 10−10 for the Pseudomonas fluorescens ortholog (27% amino acid identity over 237 residues) to 4 × 10−11 for the Enterobacter aerogenes ortholog (30% amino acid identity over 213 residues). Included in the list of MapA orthologs is the nonspecific class A acid phosphatase from Escherichia blattae (29% amino acid identity over 213 residues; the BLAST probability score was 6 × 10−10), an enzyme for which the three-dimensional structure (31) has been elucidated (Protein Data Bank [PDB] code 1D2T). Results from ClustalW sequence alignment of the acid phosphatase portion of MapA with these eight orthologs are shown in Fig. Fig.8,8, together with the signature motif of the NSAP class A family as described previously by Thaller and colleagues (67).
The aforementioned signature motif for acid phosphatases (67) and other related phosphatases has two conserved histidine residues (Fig. (Fig.8).8). The second corresponds to H233 of the MapA protein. By analogy to other enzymes (i.e., glucose 6-phosphatase and vanadium-containing chloroperoxidase) that contain this signature sequence, H233 putatively serves to attack a phosphorylated substrate, thereafter forming a covalent phosphoenzyme intermediate (24, 50). In the acid phosphatase PiACP from Prevotella intermedia, the analogous histidine (H209) is essential for catalytic activity (10). To determine whether H233 of MapA was involved in acid phosphatase activity, site-directed mutagenesis was used to convert this histidine to an alanine. The resultant recombinant MapA (H233A) protein expressed by pTH24 (Fig. (Fig.3A,3A, lane 3, and B, bar 3) had little or no phosphatase activity compared to that expressed by the wild-type protein encoded by pKW04 (Fig. (Fig.3A,3A, lane 1, and B, bar 1) (P < 0.0001). Alteration of a different histidine residue (H110) outside the predicted active site in pTH26 had no apparent effect on phosphatase activity as measured by the indicator plate assay (Fig. (Fig.3A,3A, lane 4). However, the level of enzymatic activity obtained with pTH26 (Fig. (Fig.3B,3B, bar 4) was slightly less than that obtained with the wild-type enzyme expressed by pKW04 (Fig. (Fig.3B,3B, bar 1) (P < 0.0104). These data indicate that H233 is essential for the phosphatase activity of the M. catarrhalis MapA protein.
The N-terminal portion of the MapA protein (aa 24 to 279 [Fig. [Fig.7A])7A]) is most similar to class A NSAP molecules as described above. Sequence alignment of this region of the MapA protein with the NSAP of E. blattae demonstrated that the two proteins share about 23% identity over this region. It therefore appears to be likely that the MapA amino terminus will be significantly structurally similar to the NSAP of E. blattae. The latter is a compact, all-α-helical protein (Fig. (Fig.9A),9A), and it is the only NSAP for which a crystal structure is available. The structure of the NSAP of E. blattae demonstrated that the aforementioned signature sequence contains residues that are crucial to the functioning of the enzyme's active site. MapA also contains the class A NSAP signature sequence with the crucial histidine residue located at position 233.
The middle portion of the MapA polypeptide does not harbor strong sequence homology to proteins in the data banks. However, there is a region (aa 414 to 639 [Fig. [Fig.7A])7A]) that demonstrates limited homology to pertactin-like passenger domains described for other autotransporter proteins (41). The E value (42) of this homology is 0.003; for comparison, the much stronger homology described above for the acid phosphatase region has an E value of 7 × 10−24. Pertactin-like domains can have a variety of activities, including proteolysis and cell adhesion. However, no protease or cell-adhesion signature sequences were located in the MapA sequence. The pertactin-like region of MapA may act as a linker between the acid phosphatase domain and the carboxy-terminal translocation domain.
The carboxy-terminal domain (aa 672 to 940 [Fig. [Fig.7A])7A]) of MapA is homologous to the β domains of conventional autotransporters (E value of 6 × 10−10). A PROMALS-generated alignment showed strong correspondence between the predicted β-strands (33) of MapA (Fig. (Fig.9B)9B) and the experimentally observed β-strands of the crystalline NalP autotransporter domain (48).
The identification of an autotransporter-associated acid phosphatase in M. catarrhalis is a novel finding. Previous work from this laboratory and others has demonstrated that M. catarrhalis synthesizes at least four other autotransporters (1, 6, 51, 68). The MapA protein appears to be most similar to conventional autotransporters. The prototypical signal sequence for a bacterial autotransporter usually consists of 18 to 26 aa; the predicted signal peptide for MapA has 23 residues (Fig. (Fig.2).2). The passenger domain of the typical autotransporter molecule contains the region responsible for the biological function of the translocated protein and can exceed 100 kDa (11, 27, 35). In the M. catarrhalis MapA protein, the predicted passenger domain (Fig. (Fig.7A)7A) has similarity to the pertactin family of passenger domains but has no readily identifiable functional activity. Instead, it may serve to link the acid phosphatase domain to the MapA translocation domain, such that the passenger component of MapA includes both the acid phosphatase domain and the pertactin-like domain.
Based on their translocation domains, autotransporters can be divided into two types. The C-terminal region of conventional autotransporters, including MapA, usually consists of 250 to 300 aa and forms a β-barrel structure with transmembrane β-strands (11, 12, 27). A trimeric autotransporter also has a β-barrel structure, but it comprises three sets of four β-strands. In both cases, the β-barrel secondary structure, once inserted into the outer membrane, allows for the movement of the effector portion to the bacterial cell surface (12). Initially, autotransporter proteins had been predicted to contain 14 membrane-spanning β-strands arranged as a barrel with a central pore (12). However, more recently, a crystal structure of the β-domain of the NalP autotransporter from Neisseria meningitidis (48) showed that this protein comprises only 12 β-strands, with a single α-helix residing in the pore formed by the strands. This result was controversial because the NalP β-domain protein used for crystallization had to be solubilized and refolded from inclusion bodies. However, a subsequent crystal structure of the β-domain of a natively folded trimeric autotransporter, that of Hia from Haemophilus influenzae, has been determined (43). This structure confirmed the 12-stranded organization of the β-domain of autotransporters.
We aligned the amino acid sequences of the β-domain of the M. catarrhalis MapA protein and NalP, which is the only conventional autotransporter whose structure has been determined by X-ray crystallography (48). The alignment revealed that these domains have only limited identity (about 9%). However, the alignment algorithm that was used (PROMALS) employs a new hidden Markov model that scores aligned positions based on both amino acid similarity and the correspondence of secondary structures predicted by PSIPRED (33, 53). Based on the available crystal structures, the sequential homologies, and the putative structural homologies, we conclude that the β-domain of MapA is likely to be a 12-stranded membrane-spanning pore with an α-helix occupying the center of the pore. Proof that this domain was functional in translocation of a passenger domain to the M. catarrhalis cell surface was obtained by using the McaP-MapA fusion protein (Fig. (Fig.77).
BLAST analysis revealed the presence of four putative autotransporters in the databases with E scores of less than 8 × 10−17 that also have a class A acid phosphatase-like domain near their N termini. These include predicted proteins from Pseudomonas syringae pv. phaseolicola 1448A (GenBank accession number AAZ35404) (32), P. syringae pv. syringae B728A (GenBank accession number YP_233446) (16), P. syringae pv. tomato DC3000 (GenBank accession number NP_794931) (8), and Bradyrhizobium sp. strain BTAi1 (GenBank accession number ZP_00861929). No descriptions of these proteins have been published to date.
The conservation of the MapA protein (97% identity) among the six M. catarrhalis strains used in this study was very high (Fig. (Fig.2).2). In addition, Western blot analysis of 13 additional M. catarrhalis isolates with the MapA-reactive MAb 1H12 (Fig. (Fig.5A)5A) showed that at least the MapA epitope bound by this MAb is present in all of these strains. Our experiments indicate that the MapA protein is localized to the outer membrane and, by analogy with other autotransporters, is likely exposed on the surface of the bacterium. The identification of the putative active site of the MapA acid phosphatase domain (Fig. (Fig.8)8) and the subsequent mutagenesis of its critical histidine residue served to confirm the enzymatic activity of this protein (Fig. (Fig.3).3). The fact that MapA is expressed at a relatively low level in comparison to other M. catarrhalis autotransporters (e.g., UspA2) may be related to the fact that its enzymatic function is catalytic and does not require abundant protein expression. The cleavage and release of phosphate from organic sources by this acid phosphatase may be necessary for the uptake of an essential nutrient. The wild-type parent strain and the mapA deletion mutant had similar growth rates in vitro, but these conditions clearly do not mimic those encountered in the nasopharynx, where certain nutrients and other environmental factors may be limiting in the absence of MapA activity. Taken together, these data suggest that the mapA gene product is likely important to M. catarrhalis, although the specific biological function of this acid phosphatase, like that of most other bacterial nonspecific acid phosphatases, remains to be determined.
This study was supported by U.S. Public Health Service grant no. AI36344 to E.J.H.
We thank John Nelson, Anthony Campagnari, David Goldblatt, Steven Berk, Frederick Henderson, and Merja Helminen for the isolates of M. catarrhalis used in this study and Jason Huntley for assistance with bioinformatics. We also thank Eric Lafontaine for providing the mcaP mutant O35E.M and the McaP polyclonal antiserum.
Published ahead of print on 7 December 2007.