Moraxella catarrhalis uses a twin-arginine translocation system to secrete the β-lactamase BRO-2

Background Moraxella catarrhalis is a human-specific gram-negative bacterium readily isolated from the respiratory tract of healthy individuals. The organism also causes significant health problems, including 15-20% of otitis media cases in children and ~10% of respiratory infections in adults with chronic obstructive pulmonary disease. The lack of an efficacious vaccine, the rapid emergence of antibiotic resistance in clinical isolates, and high carriage rates reported in children are cause for concern. Virtually all Moraxella catarrhalis isolates are resistant to β-lactam antibiotics, which are generally the first antibiotics prescribed to treat otitis media in children. The enzymes responsible for this resistance, BRO-1 and BRO-2, are lipoproteins and the mechanism by which they are secreted to the periplasm of M. catarrhalis cells has not been described. Results Comparative genomic analyses identified M. catarrhalis gene products resembling the TatA, TatB, and TatC proteins of the well-characterized Twin Arginine Translocation (TAT) secretory apparatus. Mutations in the M. catarrhalis tatA, tatB and tatC genes revealed that the proteins are necessary for optimal growth and resistance to β-lactams. Site-directed mutagenesis was used to replace highly-conserved twin arginine residues in the predicted signal sequence of M. catarrhalis strain O35E BRO-2, which abolished resistance to the β-lactam antibiotic carbanecillin. Conclusions Moraxella catarrhalis possesses a TAT secretory apparatus, which plays a key role in growth of the organism and is necessary for secretion of BRO-2 into the periplasm where the enzyme can protect the peptidoglycan cell wall from the antimicrobial activity of β-lactam antibiotics.


Background
Moraxella catarrhalis is a Gram-negative bacterium primarily associated with otitis media in children and respiratory infections in adults with compromised lung function, particularly patients with Chronic Obstructive Pulmonary Disease (COPD). The organism is also readily isolated from the upper respiratory tract of healthy individuals and thus was considered a commensal bacterium until relatively recently. The rate of colonization by M. catarrhalis varies depending on many factors such as age, socioeconomic status, geography, and overall health condition. It has been reported that~2/3 of children are colonized in their first year of life and 3-5% of adults carry the organism asymptomatically. Following initial colonization, there is a high rate of turnover, indicating continual clearance and re-colonization by new strains .
Moraxella catarrhalis is often refractory to antibiotic treatment. Over 90% of isolates have been shown to possess a beta-lactamase, making them resistant to penicillinbased antibiotics [44][45][46][47][48][49][50][51], which are typically prescribed first to treat otitis media. The genes specifying this resistance appear to be of Gram-positive origin [52,53], suggesting that M. catarrhalis can readily acquire genes conferring resistance to additional antibiotics via horizontal transfer. Additionally, recent evidence has shown that M. catarrhalis persists as a biofilm in vivo, giving it further protection from antibiotic treatment and the host immune response [54][55][56][57][58].
The bacterial twin-arginine translocation (TAT) system mediates secretion of folded proteins across the cytoplasmic membrane. The TAT apparatus typically consists of three integral membrane proteins, namely TatA, TatB, and TatC. TatA forms the pore through which TAT substrates are secreted whereas TatB and TatC are important for binding and directing the substrates to the TatA pore. TatC acts as the gatekeeper for the secretion apparatus and specifically recognizes TAT substrates via a well-conserved signal sequence [59][60][61][62]. This signal sequence is similar to that of the general secretory system in that it possesses N-(neutral), H-(hydrophobic), and C-(charged) regions, but also contains a TAT-recognition motif (S/T-R-R-x-F-L-K) exhibiting nearly invariable twin arginines (RR) followed by two uncharged residues [59][60][61][62]. Proteins secreted via the TAT system are often, but not limited to, proteins that bind cofactors in the cytoplasm prior to transport, such as those involved in respiration and electron transport, and proteins that bind catalytic metal ions [59][60][61][62].

Results and discussion
Identification of tatA, tatB and tatC genes in M. catarrhalis Analysis of the patented genomic sequence of M. catarrhalis strain ATCC43617 using NCBI's tblastn service (http://blast.ncbi.nlm.nih.gov/Blast.cgi) identified an ORF (nucleotides 267,266 to 266,526 of GenBank accession number AX06766.1) that encodes a protein similar to the tatC gene product of Pseudomonas stutzeri [75] (expect value of 7e -56 ). TatC is the most highly-conserved component of the TAT system among organisms known (or predicted) to utilize this particular secretion apparatus [59][60][61][62]. TatC is located in the cytoplasmic membrane, typically contains 6 membrane-spanning regions, and plays a key role in recognizing the twinarginine motif in the signal sequence of molecules secreted by the TAT system. The M. catarrhalis ATCC43617 tatC-like ORF specifies a 27-kDa protein of 247 amino acids, and analysis using the TMPred server (http:// www.ch.embnet.org/software/TMPRED_form.html) revealed that it contains 6 potential membrane-spanning domains (data not shown).
Sequence analysis upstream of the M. catarrhalis tatC ortholog identified gene products similar to other conserved components of the TAT system, TatA and TatB ( Figure 1). The ORF immediately upstream encodes a 178-residue protein with a molecular weight of 20-kDa that resembles TatB of Providencia stuartii [76] (expect value of 3e -8 ). Upstream of the M. catarrhalis tatB-like gene, we identified an ORF specifying a 9-kDa protein of 77 aa that is most similar to TatA of Xanthomonas oryzae [77] (expect value of 2e -5 ). TatA and TatB are cytoplasmic proteins anchored to the cytoplasmic membrane via hydrophobic N-termini. TatB forms a complex with TatC often referred to as the twin-arginine motif recognition module, while TatA oligomerizes and forms a channel that is used to secrete TAT substrates [59][60][61][62]. Both M. catarrhalis ATCC43617 TatA (aa 4-21) and TatB (aa 5-21) orthologs are predicted to contain hydrophobic membrane-spanning domains in their N-termini using TMPred (data not shown). Together, these observations suggest that M. catarrhalis possesses a functional TAT system.
To assess the presence and conservation of the tat genes in other M. catarrhalis isolates, we amplified and sequenced these genes from strains O35E, O12E, McGHS1, V1171, and TTA37. The encoded gene products were then compared using ClustalW (http://www. ebi.ac.uk/Tools/msa/clustalw2/). Of note, the annotated genomic sequence of the M. catarrhalis isolate BBH18 has been published [78] and the predicted aa sequence of the TatA (MCR_0127, GenBank accession number ADG60399.1), TatB (MCR_0126, GenBank accession number ADG60398.1) and TatC (MCR_0125, GenBank accession number ADG60397.1) proteins were included in our comparative analyses. Overall, the TatA and TatC proteins are perfectly conserved. The TatB proteins divide the strains into two groups where O35E, McGHS1, TTA37, ATCC43617, and BBH18 are 100% identical to each other, while O12E and V1171 both contain the same aa substitution at residue 38 (S in lieu of G). We also noted that in all isolates examined, the tatA and tatB ORFs overlap by one nucleotide. A similar onenucleotide overlap is also observed for the tatB and tatC coding regions. This observation suggests that the M. catarrhalis tatA, tatB, and tatC genes are transcriptionally and translationally linked.
The M. catarrhalis tatA, tatB and tatC genes are necessary for optimal growth To study the functional properties of the Tat proteins in M. catarrhalis, we constructed a panel of isogenic mutant strains of isolate O35E in which the tatA, tatB and tatC genes were disrupted with a kanamycin resistance (kan R ) marker. Each mutant was also complemented with a plasmid encoding a wild-type (WT) copy of the mutated tat gene and/or with a plasmid specifying the entire tatABC locus. A growth defect was immediately noted in the tat mutants as~40-hr of growth at 37°C was necessary for isolated colonies of appreciable size to develop on agar plates, compared to~20-hr for the parent strain O35E. Hence, we compared the growth of the tat mutants to that of the WT isolate O35E in liquid medium under aerobic conditions. This was accomplished by measuring the optical density (OD) of cultures over a 7-hr period. In some of these experiments, we also plated aliquots of the cultures to enumerate colony forming units (CFU) as a measure of bacterial viability.
As shown in Figure 2A, the tatA, tatB and tatC mutants carrying the control plasmid pWW115 had lower OD readings than their progenitor strain O35E throughout the entire course of the experiments. Significant differences in the number of CFU were also observed between mutants and WT strains ( Figure 2B). Together, these results demonstrate that mutations in the tatA, tatB and tatC genes substantially reduce the growth rate of M. catarrhalis cells. Complementation of the tatA ( Figure 3A) and tatB ( Figure 3B) mutants with plasmids encoding WT tatA (i.e. pRB.TatA) or tatB (i.e. pRB.TatB) did not rescue the growth phenotype of these strains. However, the construct pRB.TAT, which specifies the entire tatABC locus, restored growth of the tatA and tatB mutants to WT levels ( Figure 3A and B). These results support the hypothesis that the tatA, tatB and tatC genes are transcriptionally and translationally linked due to the one nucleotide overlaps between the tatA and tatB, as well as the tatB and tatC ORFs. For the tatC mutant, O35E.TC, introduction of the plasmid pRB. TatC, which encodes only the tatC gene, is sufficient to restore growth to WT levels ( Figure 3C). This finding is consistent with the above observations since tatC is located downstream of tatA and tatB (Figure 1), thus it is unlikely that a mutation in tatC would affect the expression of either the tatA or tatB gene product. A tatC mutation was also engineered in the M. catarrhalis isolate O12E. The resulting strain, O12E.TC, exhibited a growth defect comparable to that of the tatC mutant of strain O35E, and this growth defect was rescued by the plasmid pRB.TatC (data not shown). These results demonstrate that the importance of the TAT system to M. catarrhalis growth is not a strain-specific occurrence. Of note, all tat mutants carrying the control plasmid pWW115 grew at rates comparable to the mutants containing no plasmid (data not shown).
The tatA, tatB and tatC genes are necessary for the secretion of β-lactamase by M. catarrhalis TAT-deficient mutants of E. coli [79] and mycobacteria [72][73][74]80] have been previously shown to be hypersensitive to antibiotics, including β-lactams. Moreover, the β-lactamases of M. smegmatis (BlaS) and M. tuberculosis (BlaC) have been shown to possess a twin-arginine motif in their signal sequences and to be secreted by a TAT system [74]. More than 90% of M. catarrhalis isolates are resistant to β-lactam antibiotics [44][45][46][47][48][49][50][51]. The genes responsible for this resistance, bro-1 and bro-2, specify lipoproteins of 33-kDa that are secreted into the periplasm of M. catarrhalis where they associate with the inner leaflet of the outer membrane [52,53]. Analysis of the patented genomic sequence of M. catarrhalis strain ATCC43617 with NCBI's tblastn identified the bro-2 gene product (nucleotides 8,754 to 7,813 of GenBank accession number AX067438.1), which is predicted to encode a protein of 314 residues with a predicted MW of 35-kDa. The first 26 residues of the predicted protein were found to specify characteristics of a signal sequence (i.e. n-, h-, and c-region; see Figure 4A). Analysis with the LipoP server (http://www.cbs.dtu.dk/services/LipoP/) indicated a signal sequence cleavage site between residues 26 and 26 (i.e. TG 26▼ C 27 K) of BRO-2 (arrowhead Figure 1 Schematic representation of the M. catarrhalis tatABC locus. The relative position of tat-specific oligonucleotide primers (P1-P8) used throughout the study is shown (see Methods section for details).
in Figure 4A), which would provide a free cysteine residue for lipid modification of this lipidated β-lactamase [52]. Of significance, the putative signal sequence of BRO-2 contains the highly-conserved twin-arginine recognition motif RRxFL (Figure 4), thus suggesting that the gene product is secreted via a TAT system. Of note, analysis of M. catarrhalis BRO-1 sequences available through the NCBI database indicates that the molecules also contain the twin-arginine recognition motif (data not shown). Based on these findings, we compared the ability of our panel of WT and tat mutant strains to grow in the presence of the β-lactam antibiotic carbenicillin. This was accomplished by spotting equivalent numbers of bacteria onto agar plates supplemented with the antibiotic. For comparison, bacteria were also spotted onto agar plates without carbenicillin. These plates were incubated for 48-hr at 37°C to accommodate the slower growth rate of tat mutants. In contrast to WT M. catarrhalis O35E, which is resistant to carbenicillin, the tatA ( Figure 5A), tatB ( Figure 5B), and tatC ( Figure 5C) mutants were sensitive to the antibiotic. The introduction of plasmids containing a WT copy of tatA (i.e. pRB. TatA, Figure 5A) and tatB (i.e. pRB.TatB, Figure 5B) did not restore the ability of the tatA and tatB mutants to grow in the presence of carbenicillin, respectively. Resistance to the β-lactam was observed only when the tatA and tatB mutants were complemented with the plasmid specifying the entire tatABC locus (see pRB.TAT in Figure 5A and B), which is consistent with the results of the growth experiments presented in Figure 3. Introduction of the plasmid encoding only the WT copy of tatC (i.e. pRB.TatC) in the strain O35E.TC was sufficient to restore the growth of this tatC mutant on medium supplemented with carbenicillin ( Figure 5C). Of note, the tatC mutant of strain O12E was tested in this manner and the results were consistent with those obtained with O35E.TC (data not shown). In order to provide an appropriate control for these experiments, an isogenic mutant strain of M. catarrhalis O35E was constructed in which the bro-2 gene was disrupted with a kan R marker. The mutant, which was designated O35E.Bro, grew at the same rate as the parent strain O35E in liquid medium ( Figure 3C). As expected, the bro-2 mutant did not grow on agar plates containing carbenicillin ( Figure 5C). The effect of tat mutations on the β-lactamase activity of M. catarrhalis was quantitatively measured using the chromogenic β-lactamase substrate nitrocefin. These assays were performed using suspensions of freshly plate-grown bacteria placed into the wells of a 48-well tissue culture plate. A solution containing nitrocefin was added to these suspensions and the change of color from yellow to red (indicative of cleavage of the β-lactam ring) was monitored by measuring the absorbance of well contents at a wavelength of 486 nm. Substantially less β-lactamase activity was observed for the tatA, tatB and tatC mutants compared to the WT strain O35E Growth of O35E is compared to that of its tatA isogenic mutant strain, O35E.TA, carrying the plasmid pWW115 (control), pRB.TatA (specifies a WT copy of tatA), and pRB.TAT (harbors the entire tatABC locus). Panel B: Growth of O35E is compared to that of its tatB isogenic mutant strain, O35E.TB, carrying the plasmid pWW115, pRB.TatB (specifies a WT copy of tatB), and pRB.TAT. Panel C: Growth of O35E is compared to that of its tatC isogenic mutant strain, O35E.TC, carrying the plasmid pWW115 and pRB.TatC (contains a WT copy of tatC). Growth of the bro-2 isogenic mutant strain O35E.Bro is also shown. Results are expressed as the mean OD ± standard error. Asterisks indicate a statistically significant difference in the growth rates of mutant strains compared to that of the WT isolate O35E.
( Figure 6). Complementation of the tatA and tatB mutants with plasmids containing only the WT copies of the inactivated genes did not restore β-lactamase activity, as expected based on the results of the experiments depicted in Figures 3 and 5. The plasmid pRB. TAT, which specifies the entire tatABC locus, restored the ability of the mutants O35E.TA ( Figure 6A) and O35E.TB ( Figure 6B) to hydrolyze nitrocefin. The plasmid pRB.TatC was sufficient to rescue β-lactamase activity in the tatC mutant strain O35E.TC to near WT levels ( Figure 6C). The tatC mutant of strain O12E was tested in this manner and the results were consistent with those obtained with O35E.TC (data not shown). The control strain, O35E.Bro, was impaired in its ability to hydrolyze nitrocefin at levels comparable to those of the tatA, tatB and tatC mutants ( Figure 6A, B and C). Taken together, these results suggest that the M. catarrhalis tatABC locus is necessary for secretion of the β-lactamase BRO-2 into the periplasm where the enzyme can protect the peptidoglycan cell wall from the antimicrobial activity of β-lactam antibiotics.
To conclusively demonstrate that M. catarrhalis BRO-2 is secreted by the TAT system, we cloned the bro-2 gene of strain O35E in the plasmid pWW115 (pTS.Bro) and used site-directed mutagenesis to replace the twinarginine (RR) residues in BRO-2's predicted signal sequence ( Figure 4A) with twin lysine (KK) residues (pTS. BroKK). Similar conservative substitutions have been engineered in TAT substrates of other bacteria to demonstrate the importance of the RR motif in TATdependent secretion [74]. These plasmids were introduced in the mutant O35E.Bro and the recombinant strains were tested for their ability to hydrolyze nitrocefin. As shown in Figure 7A, expression of the mutated BRO-2 from plasmid pTS.BroKK did not restore the ability to hydrolyze nitrocefin. These results establish that the M. catarrhalis β-lactamase BRO-2 is secreted into the periplasm by the TAT system. Interestingly, the mutation in the RR motif of BRO-2 also interfered with secretion of the β-lactamase by recombinant Haemophilus influenzae DB117 bacteria ( Figure 7B).

Identification of other M. catarrhalis gene products potentially secreted by the TAT system
To identify other M. catarrhalis molecules that may be secreted via the TAT apparatus, the annotated genomic sequences of strains ATCC43617 [81] and BBH18 [78] were analyzed with prediction algorithms available through the TatFind 1.4 [82] and TatP 1.0 [83] servers. Although these servers are designed for the same purpose (i.e. identify proteins secreted by the TAT system), the algorithms used for each differ and as such proteins identified as TAT substrates do not overlap 100% between the two prediction algorithms.
Six ORFs were predicted to be TAT substrates in strain ATCC43617, only one of which was identified by both algorithms (Figure 8). The TatP 1.0 server identified MCORF 312 and MCORF 1197 as proteins potentially secreted by the TAT system, but no twin-arginine motif was found within the signal sequences of these gene products. Conversely, the TatFind 1.4 server identified MCORF 1917 as a TAT substrate and a twin-arginine motif was observed between residues 18 and 23. Although the encoded protein does not specify characteristics of a prokaryotic signal sequence (i.e. n-, h-, c-region), a potential lipoprotein signal sequence cleavage site was identified using the LipoP server. Interestingly, the MCORF 1197 and MCORF 1199 gene products resemble cytochrome c molecules involved in the electron transport chain. Cytochromes have been predicted, as well as demonstrated, to be TAT substrates in several bacterial species [84][85][86][87]. MCORF 1917 exhibits similarities to iron-dependent peroxidases, which is consistent with the previously reported role of the TAT system in the secretion of enzymes that bind metal ions, while MCORF 518 resembles the phosphate ABC transporter inner membrane protein PstA [88]. MCORF 838 shows similarities to a family of C-terminal processing peptidases and contains important functional domains including a post-translational processing, maturation and degradation region (PDZ-CTP), and a periplasmic protease Prc domain described as important for cell envelope biogenesis. Six putative TAT substrates were identified in the genome of M. catarrhalis strain BBH18, five of which overlapping those predicted in ATCC43617 (Figure 8). Strain BBH18 specifies the unique TAT substrate MCR_920, which is predicted to be a highly-conserved phosphatase (Figure 8). The MCORF 1659 of strain ATCC43617 encodes a gene product that is 96.8% identical to this putative phosphatase, but neither of the TatFind 1.4 and TatP 1.0 servers identified the ORF as a TAT substrate, likely due to significant amino acid divergence in the signal sequence (data not shown  The ability of tat mutants to grow in the presence of carbenicillin (cab) was tested by spotting equivalent numbers of bacteria onto Todd-Hewitt agar plates supplemented with the antibiotic (TH + cab). As control, bacteria were also spotted onto agar plates without carbenicillin (TH). These plates were incubated for 48 hrs at 37°C to accommodate the slower growth rate of the tat mutants. Panel A: Growth of O35E is compared to that of its tatA isogenic mutant strain, O35E.TA, carrying the plasmid pWW115 (control), pRB.TatA (specifies a WT copy of tatA), and pRB.TAT (harbors the entire tatABC locus). Panel B: Growth of O35E is compared to that of its tatB isogenic mutant strain, O35E.TB, carrying the plasmid pWW115, pRB.TatB (specifies a WT copy of tatB), and pRB.TAT. Panel C: Growth of O35E is compared to that of its tatC isogenic mutant strain, O35E.TC, carrying the plasmid pWW115 and pRB.TatC (contains a WT copy of tatC). Growth of the bro-2 isogenic mutant strain O35E.Bro is also shown. The results are shown as a composite image representative of individual experiments that were performed in duplicate on at least 3 separate occasions.
BBH18 specifies a putative C-terminal processing peptidase (MCR_1063) that is 98.1% identical to the putative TAT substrate MCORF 838 of ATCC43617. Like the MCORF 838 of ATCC43617, the BBH18 gene product lacked a TAT motif in its signal sequence (data not shown). Despite these similarities, the BBH18 MCR_1063 C-terminal processing peptidase was not identified as a being potential TAT substrate by the TatFind 1.4 or TatP 1.0 algorithms.

Conclusions
This report is the first characterization of a secretory apparatus for M. catarrhalis. Our data demonstrate that the TAT system mediates secretion of β-lactamase and is necessary for optimal growth of the bacterium. Moraxella catarrhalis is a leading cause of otitis media worldwide along with Streptococcus pneumoniae and non-typeable Haemophilus influenzae (NTHi), and is often found in mixed infections with these organisms [1][2][3][4][5][6][7][8]89]. In contrast to M. catarrhalis, most S. pneumoniae and NTHi isolates are susceptible to β-lactam antibiotics [90]. In a set of elegant studies, Schaar et al. demonstrated that outer membrane vesicles produced by M. catarrhalis contain β-lactamase and function as a long-distance delivery system to confer antimicrobial resistance for β-lactamase negative isolates of S. pneumoniae and NTHi [91]. This constitutes a novel mechanism by which M. catarrhalis  compounds that selectively inhibit TAT secretion in M. catarrhalis could be used in concert with β-lactam antibiotics as β-lactamase inhibitors. This hypothesis is supported by the recent discovery that the compounds N-phenyl maleimide and Bay 11-7782 specifically interfere with TAT-dependent secretion of the Pseudomonas aeruginosa phospholipase C PlcH [92].

Recombinant DNA techniques
Standard molecular biology techniques were performed as described elsewhere [97]. Moraxella catarrhalis genomic DNA was obtained using the Easy-DNA™ kit (Invitrogen™ Life Technologies™) per the manufacturer's instructions. Plasmid DNA was purified with the QIAprep Spin Miniprep system (QIAGEN). Polymerase chain reactions were performed using Taq DNA Polymerase (Invitrogen™ Life Technologies™) unless otherwise specified. A 1,018-nt fragment containing the tatC gene was amplified with primers P1 (5′-AAA GCCAAGCCAACGGACTT-3′) and P2 (5′-ACCTCC AAGAAACCCACGCTATCA-3′) using genomic DNA from M. catarrhalis strain O12E (see Figure 1 for more details regarding primers). This PCR product was cloned into the vector pCC1 using the Copy Control™ PCR Cloning Kit (Epicentre® Illumina®) per the manufacturer's instructions. This process yielded plasmid pRB.TatC.2, which was sequenced to verify that mutations were not introduced in the tatC gene during cloning. The β-lactamase activity of H. influenzae DB117 carrying plasmids pWW115, pTS.Bro, and pTS.BroKK is compared. Sterile broth was used as a negative control in these experiments. The results are expressed as the mean ± standard error A 486 . Asterisks indicate that the reduction in the β-lactamase activity of strains lacking expression of BRO-2, or expressing a mutated BRO-2 that contains two lysine residues in its signal sequence instead of 2 arginines, is statistically significant (P < 0.05) when compared to bacteria expressing a WT copy of the bro-2 gene.

Mutant construction
To create a tatC mutation in M. catarrhalis, the plasmid pRB.TatC.2 was mutagenized with the EZ-TN5™ < KAN-2 > Insertion Kit (Epicentre® Illumina®) and introduced into Transformax™ EPI300™ electrocompetent cells. Chloramphenicol resistant (cam R , specified by the vector pCC1) and kanamycin resistant (kan R , specified by the EZ-TN5 < KAN-2 > TN) colonies were selected and plasmids were analyzed by PCR using the pCC1-specific primer, P11 (5′-TACGCCAAGCTATTTAGGTGAGA-3′), and primers specific for the kan R marker, P12 (5′-ACCT ACAACAAAGCTCTCATCAACC-3′) and P13 (5′-GC AATGTAACATCAGAGATTTTGAG-3′). This strategy identified plasmid pRB.TatC:kan, in which the EZ-TN5 < KAN-2 > TN was inserted near the middle of the tatC ORF. The disrupted tatC gene was then amplified from pRB.TatC:kan with the pCC1-specific primers P11 and P14 (5′-TAATACGACTCACTATAGGG-3′) using Pla-tinum® Pfx DNA Polymerase. This 2.3-kb PCR product was purified and electroporated into M. catarrhalis strains O12E and O35E to create the kan R isogenic mutant strains O12E.TC and O35E.TC via homologous recombination. Allelic replacement was confirmed by PCR with the tatC primers P1 and P2 using Platinum® Pfx DNA Polymerase. These primers yielded PCR products in the mutant strains that were 1.2-kb larger than the amplicons obtained in the wild-type (WT) isolates O35E and O12E due to the presence of the EZ-TN5 < KAN-2 > TN in tatC. To construct mutations in the tatA and tatB genes of M. catarrhalis O35E, the plasmid pRB.Tat.1 was first mutagenized with the EZ-TN5™ In-Frame Linker Insertion Kit (Epicentre® Illumina®) and introduced into Transformax™ EPI300™ electrocompetent cells. Plasmid DNA was isolated from several cam R (specified by the vector pCC1) and kan R (specified by the EZ-TN5 < Not I/KAN-3 > TN) clones and sequenced to determine the sites of insertion of the TN. This approach identified the plasmids pRB.TatA:kan and pRB.TatB:kan, which contained the EZ-TN5 < Not I/KAN-3 > TN at nt 90 of the tatA ORF and nt 285 of the tatB ORF, respectively. These plasmids were then introduced into M. catarrhalis strain O35E by natural transformation as previously described [34]. The resulting kan R strains were screened by PCR using primers specific for tatA (P3 and P4) and tatB (P5 and P6), which produced DNA fragments that were 1.2-kb larger in size in mutant strains when compared to the WT strain O35E because of the insertion of the EZ-TN5 < Not I/KAN-3 > TN in tatA and tatB. This strategy yielded the mutant strains O35E. TA and O35E.TB.
To construct a mutation in the bro-2 gene of M. catarrhalis O35E, plasmid pRN.Bro11 was mutagenized with the EZ-TN5™ In-Frame Linker Insertion Kit as described above. Plasmids were isolated from kan R cam R colonies and sequenced to identify constructs containing the EZ-TN5 < Not I/KAN-3 > TN near the middle of the bro-2 ORF. This approach yielded the construct pRB. Bro:kan, which was introduced in M. catarrhalis O35E by natural transformation. Transformants were selected for resistance to kanamycin and then tested for their ability to grow on agar plates containing the β-lactam antibiotic carbenicillin. Kan R and carbenicillin sensitive (cab S ) strains were further analyzed by PCR using primers P9 and P10 to verify allelic exchange of the bro-2 gene. These primers produced a 1-kb DNA fragment in the WT strain O35E and a 2.2-kb in the mutant O35E. Bro, which is consistent with insertion of the 1.2-kb EZ-TN5 < Not I/KAN-3 > TN in bro-2.
Site-directed mutagenesis of the M. catarrhalis bro-2 gene The bro-2 ORF of M. catarrhalis O35E harbored by plasmid pRN.Bro11 was mutated using the QuikChange Lightning Site-Directed Mutagenesis Kit (Agilent Technologies) according to the manufacturer's instructions. The mutagenesis primers, P15 (5′-AAGGGGATAATG ATGCAAAAGAAGCATTTTTTA-3′) and P16 (5′-GG TTTTTTGTAAAAAATGCTTCTTTTGCAT CAT-3′), were used to replace two arginine residues at position 4 and 5 of BRO-2 with two lysines, yielding plasmid pTS. BroKK.Ec. This plasmid was sequenced to verify that only the intended mutations were introduced in the bro-2 ORF.

Complementation of mutants
The construction of plasmids to complement tat and bro2 mutant strains was achieved as follows. Plasmid DNA (pRB.TatA.5, pRB.TatB.1, pRB.TatC.2, pRB.Tat.1, pRN.Bro11, pTS.BroKK.Ec) was digested with BamHI to release the cloned M. catarrhalis genes from the vector pCC1. Gene fragments were purified from agarose gel slices using the High Pure PCR Product Purification Kit (Roche Applied Science), ligated into the BamHI site of the M. catarrhalis/Haemophilus influenza-compatible shuttle vector pWW115 [95], and electroporated into H. influenzae strain DB117. Spectinomycin resistant (spc R ) colonies were screened by PCR using the pWW115-specific primers P17 (5′-TACGCCCTTTTATACTGTA G-3′) and P18 (5′-AACGACAGGAGCACGATCAT-3′), which flank the BamHI cloning site, to identify clones containing inserts of the appropriate size for the tat and bro2 genes. This process produced plasmids pRB.TatA, pRB.TatB, pRB.TatC, pRB.TAT, pTS.Bro, and pTS.BroKK. The O35E.TA mutant was naturally transformed with plasmids pWW115, pRB.TatA, and pRB.TatABC. The plasmids pWW115, pRB.TatB, and pRB.TAT were introduced in the O35E.TB mutant by natural transformation. The tatC mutants O35E.TC and O12E.TC were naturally transformed with the vector pWW115 and plasmid pRB. TatC. The plasmids pWW115, pTS.Bro, and pTS.BroKK were electroporated into the bro-2 mutant strain O35E. Bro. The successful introduction of these plasmids into the indicated strains was verified by PCR analysis of spc R transformants with the pWW115-specific primers P17 and P18, and by restriction endonuclease analysis of plasmid DNA purified from each strain.

Growth rate experiments
Moraxella. catarrhalis strains were first cultured onto agar plates supplemented with appropriate antibiotics. These plate-grown bacteria were used to inoculate 500-mL sidearm flasks containing 20-mL of broth (without antibiotics) to an optical density (OD) of 50 Klett units. The cultures were then incubated with shaking (225-rpm) at a temperature of 37°C for 7-hr. The OD of each culture was determined every 60-min using a Klett™ Colorimeter (Scienceware®). These experiments were repeated on at least three separate occasions for each strain. In some experiments, aliquots were taken out of each culture after recording the optical density, diluted, and spread onto agar plates to determine the number of viable colony forming units (CFU).

Carbenicillin sensitivity assays
Moraxella catarrhalis strains were first cultured onto agar plates supplemented with the appropriate antibiotics. These plate-grown bacteria were used to inoculate sterile Klett tubes containing five-mL of broth (without antibiotics) to an OD of 40 Klett units. Portions of these suspensions (25 μL) were spotted onto agar medium without antibiotics as well on plates supplemented with carbenicillin, and incubated at 37°C for 48-hr to evaluate growth. Each strain was tested in this manner a minimum of three times.

Nitrocefin assays
Moraxella catarrhalis and H. influenzae strains were tested for their ability to cleave the chromogenic β-lactamase substrate nitrocefin as previously described [98]. Bacterial strains were first cultured onto agar plates supplemented with appropriate antibiotics. These plate-grown cells were suspended to an OD of 300 Klett units in 5-mL of broth, and aliquots (50 μL,~10 7 CFU) were transferred to duplicate wells of a 48-well tissue culture plate; control wells were seeded with broth only. To each of these wells, 325 μL of a nitrocefin (Calbiochem®) solution (250 μg/mL in phosphate buffer) was added and the absorbance at a wavelength of 486 nm (A 486 ) was immediately measured using a μQuant™ Microplate Spectrophotometer (BioTek®) and recorded as time "0". The A 486 of the samples was then measured after a 30-min incubation at room temperature. These experiments were repeated a minimum of three times for each strain.

Statistical analyses
The GraphPad Prism Software was used for all statistical analyses. Growth rate experiments and nitrocefin assays were analyzed by a two-way analysis of variants (ANOVA), followed by the Bonferroni post-test of the means of each time point. Asterisks indicate statistically significant differences where P < 0.05.