|Home | About | Journals | Submit | Contact Us | Français|
The nucleotide sequence from the genome of Moraxella catarrhalis ATCC 43617 was annotated and used both to assess the metabolic capabilities and limitations of this bacterium and to design probes for a DNA microarray. An absence of gene products for utilization of exogenous carbohydrates was noteworthy and could be correlated with published phenotypic data. Gene products necessary for aerobic energy generation were present, as were a few gene products generally ascribed to anaerobic systems. Enzymes for synthesis of all amino acids except proline and arginine were present. M. catarrhalis DNA microarrays containing 70-mer oligonucleotide probes were designed from the genome-derived nucleotide sequence data. Analysis of total RNA extracted from M. catarrhalis ATCC 43617 cells grown under iron-replete and iron-restricted conditions was used to establish the utility of these DNA microarrays. These DNA microarrays were then used to analyze total RNA from M. catarrhalis cells grown in a continuous-flow biofilm system and in the planktonic state. The genes whose expression was most dramatically increased by growth in the biofilm state included those encoding a nitrate reductase, a nitrite reductase, and a nitric oxide reductase. Real-time reverse transcriptase PCR analysis was used to validate these DNA microarray results. These results indicate that growth of M. catarrhalis in a biofilm results in increased expression of gene products which can function not only in energy generation but also in resisting certain elements of the innate immune response.
Moraxella catarrhalis is a gram-negative, unencapsulated bacterium that can colonize the mucosal surface of the human nasopharynx, most frequently in infants and very young children (22). When this organism traverses the eustachian tube in these very young individuals, it can cause otitis media (7). Alternatively, in colonized adults, M. catarrhalis gains access to the bronchi and there causes exacerbations of chronic obstructive pulmonary disease (48). This organism can also infrequently cause other types of infections (for reviews, see references 39 and 73).
Information about the virulence mechanisms employed by this organism in the production of disease is still very limited, although a number of putative virulence factors have been identified, including proteins located in or attached to the outer membrane (1, 6, 9, 23, 24, 26, 32-34, 40, 44, 49, 51, 54, 57, 65, 71), as well as lipooligosaccharide (42, 59). Validation of the actual involvement of these different gene products in disease processes has been severely hindered by the lack of an appropriate animal model for M. catarrhalis disease (39). Similarly, little is known about how M. catarrhalis colonizes the nasopharynx, and while several M. catarrhalis adhesins which function in vitro have been identified (33, 40, 61, 71), the relative importance of these macromolecules in the colonization process in vivo remains to be determined. Recent studies aimed at addressing this issue have included the use of reverse transcriptase PCR (RT-PCR) to detect mRNA species expressed by M. catarrhalis in the nasopharynx of young children (46). Finally, little is known about the regulation of gene expression in M. catarrhalis, although recent studies indicate that both phase variation (41, 47, 57, 64) and temperature (30) can affect expression of some M. catarrhalis proteins.
Nasopharyngeal colonization by other pathogens that cause otitis media, especially Streptococcus pneumoniae and Haemophilus influenzae, has been studied in considerable detail, and bacterial gene products likely involved in this process have been identified through the use of relevant animal models (17, 63, 76). In the nasopharynx, it is reasonable to assume that M. catarrhalis, similar to other bacteria that colonize the nasopharynx (37, 52), exists in a biofilm, and it was recently reported that M. catarrhalis, like both H. influenzae and S. pneumoniae, can form a biofilm on the middle ear mucosa of children with otitis media (29). At least for H. influenzae, there have been extensive studies that have identified H. influenzae gene products involved in or affected by growth in the biofilm state in vitro (27, 50, 52, 68, 77). Whether there may be preferential expression of certain gene products in the biofilm state remains to be determined for M. catarrhalis, but recent studies with Neisseria meningitidis, another gram-negative pathogen that often colonizes the nasopharynx and can form biofilms in vitro (78), indicates that certain genes are expressed or up-regulated in response to contact with human cells in vitro (28). To date, however, there have been only a few reports describing biofilm formation by M. catarrhalis (13, 58), and only one of these involved a continuous-flow biofilm system (13).
In this study, we used the available nucleotide sequence from the genome of M. catarrhalis ATCC 43617 to facilitate investigation of different aspects of this pathogen. First, some of the basic metabolic capabilities and limitations of M. catarrhalis were inferred from the predicted gene products. Second, we developed and validated an M. catarrhalis DNA microarray for measuring global gene expression by this organism and then used this microarray to identify changes in gene expression and potential metabolic changes resulting from growth of M. catarrhalis in a continuous-flow biofilm system in vitro.
M. catarrhalis strains ATCC 43617 and ATCC 25238 were obtained from the American Type Culture Collection, Manassas, VA. M. catarrhalis strain ETSU-9 was obtained from Steven Berk, Quillen-Disher College of Medicine, East Tennessee State University, Johnson City. The base medium employed in this study was brain heart infusion (BHI) (Difco, Detroit, MI) medium, and broth cultures were incubated at 37°C with aeration. BHI medium was supplemented with kanamycin (15 μg/ml) or spectinomycin (15 μg/ml) when appropriate. All BHI agar plates were incubated at 37°C in an atmosphere containing 95% air and 5% CO2.
Iron limitation was achieved by adding the chelator deferoxamine mesylate (Desferal; Novartis, East Hanover, NJ) to BHI medium (3). Planktonic growth for iron limitation studies was obtained by inoculating bacterial growth from a BHI agar plate (grown overnight) into 20 ml of BHI broth (with or without Desferal) in a 500-ml flask. The cultures were grown at 37°C with shaking (200 rpm). Cell growth was measured by use of a Klett-Summerson photoelectric colorimeter (VWR International, West Chester, PA). For Western blot analysis, whole-cell lysates were prepared from bacteria grown in broth for 6 h. Briefly, the cells were pelleted by centrifugation and then resuspended in phosphate-buffered saline to a density of 300 Klett units. A 5-ml portion of the suspension was then subjected to centrifugation, and the resultant cell pellet was used to prepare a whole-cell lysate as described previously (56). Western blot analysis was performed using the M. catarrhalis CopB protein-specific monoclonal antibody 10F3 (31).
The Sorbarod cellulose filter-based continuous-flow system (13) for growing M. catarrhalis biofilms was used essentially as described previously (57). Bacterial cells from a stationary-phase BHI broth culture were used as the inoculum for the Sorbarod system. Briefly, a 3-ml portion of this culture was added to the cellulose filter, and then the flow of medium into the system was initiated. After 3 days of growth, the system was disassembled and the layer of bacterial growth that extended upward along the silicone tubing away from the cellulose filter was harvested for RNA extraction. Bacterial growth within or adherent to the Sorbarod filter itself was not used in the experiments. To obtain planktonic cells for RNA extraction, cells grown to the mid-logarithmic phase in BHI broth were harvested.
M. catarrhalis strain ATCC 43617 was originally derived from a patient with chronic bronchitis. A total of 1.9 Mb of nucleotide sequence from the genome of M. catarrhalis ATCC 43617 (World patent WO0078968) was obtained from the European Network of Patent Databases (http://ep.espacenet.com) as 41 different contigs. The same contigs can also be found at NCBI (GenBank accession numbers AX067426 to AX067466). The genome of M. catarrhalis strain ATCC 25238 was estimated to contain 1.75 to 1.94 Mb (25, 53), and therefore, the available nucleotide sequence from M. catarrhalis ATCC 43617 likely represents most if not all of this strain's genome. GLIMMER 2.02 software, obtained from The Institute for Genomic Research, was used to predict the open reading frames (ORFs) in the contigs (18). The 41 contigs were concatenated into one contiguous sequence (see Table S1 in the supplemental material) for use with the GLIMMER program. To minimize the possibility that GLIMMER would erroneously predict false ORFs spanning the contig junctions, a nucleotide sequence (5′-TTAACTAACTAG-3′) containing translation termination codons in all possible reading frames was placed between each pair of contigs. It should be noted that this assembly did not reflect the actual order of the DNA fragments in the M. catarrhalis chromosome and was done only to permit annotation. GLIMMER identified 1,761 putative ORFs (encoding proteins with at least 50 amino acids) in this contiguous sequence; these ORFs were translated using the standard genetic code and annotated by aligning the amino acid sequences against the nonredundant protein database at NCBI using BLASTP (see Table S2 in the supplemental material). The 1,761 annotated ORFs were used to design 70-mer oligonucleotides that would specifically anneal to each ORF. Because genome-directed primers (10, 19) that annealed within the 3′ 30% of each ORF were used to amplify the mRNA, each 70-mer probe was designed to anneal to a region within the 5′ 70% of each ORF. Following synthesis of the 70-mers (QIAGEN, Valencia, CA), the probes were spotted in triplicate on Corning UltraGAP II slides by Microarrays, Inc. (Nashville, TN). Each slide also contained three irrelevant 70-mers as negative controls.
RNA samples were isolated from planktonic or biofilm-derived cells of M. catarrhalis ATCC 43617 by using a QIAGEN RNeasy midi kit (QIAGEN) and following the manufacturer's protocol. For use in DNA microarray analysis, RNA samples were treated with QIAGEN RNase-free DNase. For real-time RT-PCR analysis, RNA samples were further treated with a MessageClean kit (GenHunter Corp., Nashville, TN) by following the manufacturer's protocol. For the DNA microarray experiments involving iron restriction, four independent experiments were performed to obtain four sets of RNA samples. For the biofilm-related DNA microarray experiments, five independent experiments were performed to obtain five sets of RNA samples.
A CyScribe postlabeling kit (GE Healthcare, Piscataway, NJ) was used according to the manufacturer's protocol, except that M. catarrhalis genome-directed primers (10, 19) were used to synthesize cDNA in the presence of amino allyl-dUTP. Twenty micrograms of total RNA and 3 μg of genome-directed primers were used for cDNA synthesis. Amino allyl-labeled cDNA was purified by using a QIAGEN QIAquick gel extraction kit. CyDye-labeled cDNA samples were purified by using Microcon YM30 columns (Millipore Corp., Bedford, MA). “Dye swap” experiments were done three times with cDNAs from the iron-related experiments and twice with cDNAs from the biofilm-related experiments. A NanoDrop spectrophotometer (NanoDrop Technologies, Wilmington, DE) was used to determine the concentration of nucleic acids in RNA, cDNA, and CyDye-labeled cDNA samples.
Cy3- and Cy5-labeled cDNA samples were mixed, lyophilized, and then dissolved in 14 μl of distilled H2O (dH2O). Following denaturation at 94°C for 3 min, these samples were kept on ice until they were used for DNA hybridization. A 40.5-μl mixture containing 10 μl of 4× hybridization buffer (GE Healthcare), 16 μl of formamide, 0.5 μl of yeast tRNA (10 mg/ml), and 14 μl of the mixture of Cy3- and Cy5-labeled cDNA was added to a DNA microarray slide and incubated at 50°C overnight (~16 h). The slides were then washed twice in 6× SSPE containing 0.01% Tween 20 at 50°C for 5 min, twice in 0.8× SSPE containing 0.001% Tween 20 at 50°C, and twice in 0.8× SSPE at room temperature (1× SSPE is 0.18 M NaCl, 10 mM NaH2PO4, and 1 mM EDTA [pH 7.7]). The slides were dried by centrifugation at 900 × g in loosely capped 50-ml conical tubes for 3 min. Microarray slides were scanned by using a GenePix 4100A microarray reader (Molecular Devices, Sunnyvale, CA) and analyzed with GenePix and Acuity 4.0 software (Molecular Devices).
Data were processed using the Acuity 4.0 software package as follows. Raw data were first normalized using a ratio-based normalization method to equalize the means and medians of the features to 1 and exclude ratios that were less than 0.1 or greater than 10. Additionally, the features that were flagged by the software as “bad,” “absent,” or “not found” were also excluded from further analysis. The array provided results for 86.8 and 85.7% of the ORFs represented on the array for iron- and biofilm-related experiments, respectively. This indicates that the majority of the genome was being interrogated by the array. To identify the genes that exhibited an altered expression profile, we utilized a threshold of twofold-increased or -decreased expression over the selected baseline. In the case of the iron-limiting growth conditions, the iron-replete growth was used as the baseline, and in the biofilm experiments, the planktonic growth was used as the baseline. We then applied two additional constraints on the data: (i) the gene had to demonstrate the altered expression profile in at least five of the six DNA microarrays and (ii) a P value of <0.05 as calculated by the one-sample t test. The data were utilized to guide the selection of targets for real-time RT-PCR analysis, and in all cases real-time RT-PCR confirmed the trends observed in the microarray data, suggesting that the metrics used for analysis were valid.
Oligonucleotide primer pairs for quantitative RT-PCR (qRT-PCR) (Table (Table1)1) were designed using PrimerExpress software (Applied Biosystems, Foster City, CA). Each 25-μl qRT-PCR mixture contained 12.5 μl of SYBR green PCR Master Mix (Applied Biosystems), 0.5 μl each of two oligonucleotide primers (from 2.5 μM stock solutions), 6.375 μl of dH2O, 0.125 μl of MultiScribe reverse transcriptase (Applied Biosystems), and 5 μl of total RNA (5 ng/μl). Controls lacking reverse transcriptase or RNA template contained the appropriate amount of dH2O in place of the enzyme or template. This one-step qRT-PCR method, including both the controls lacking reverse transcriptase and the no-template controls, was used to test freshly isolated RNA preparations for DNA contamination and the presence of “primer dimers,” respectively, using a 7500 real-time PCR system (Applied Biosystems) with the dissociation step. After RNA samples were tested for DNA contamination, one-step relative qRT-PCR (ΔΔCT) was performed using the same machine. All relative qRT-PCR experiments involved the use of two independently isolated RNA preparations. The results were analyzed by using Relative Quantification Study software (Applied Biosystems).
An approximately 980-bp DNA fragment containing some upstream DNA and the extreme 5′ end of the M. catarrhalis narG gene (MCORF923) and an approximately 870-bp DNA fragment containing the majority of the narH gene (MCORF924) were PCR amplified by using the WW191-WW192 and WW193-WW194 oligonucleotide primer pairs (Table (Table1),1), respectively, with M. catarrhalis ATCC 43617 genomic DNA as the template. The two PCR amplicons were used in overlapping extension PCR (36) with oligonucleotide primers WW191 and WW194 (Table (Table1)1) to obtain a 1.8-kb PCR product that was subsequently digested with both BamHI and SacI. The DNA fragment was ligated into the M. catarrhalis cloning vector pWW115 (74), which itself had been digested with the same two restriction enzymes. This ligation reaction was used to electroporate M. catarrhalis ATCC 25238, and a plasmid purified from a spectinomycin-resistant recombinant was designated pWW119. The kan gene from pAC7 (75) was amplified by PCR using the WW195-WW196 primer pair (Table (Table1)1) and was ligated into the SmaI site within the cloned amplicon in pWW119. The resultant plasmid, pWW120, with the kan cartridge inserted in the same direction as the narG gene, was used as the template for a PCR with the WW191-WW194 primer pair to obtain a 2.9-kb PCR product containing the plasmid insert. This PCR product was used to electroporate M. catarrhalis ETSU-9, and transformants (i.e., putative narGH deletion mutants) were selected with kanamycin.
A competitive index approach was used to determine whether the narGH mutant had a reduced ability to form a biofilm in the Sorbarod continuous-flow biofilm system. Briefly, the kanamycin-resistant narGH mutant (3 × 107 CFU) was mixed with a streptomycin-resistant mutant of the ETSU-9 parent strain (3 × 107 CFU) and used to inoculate the Sorbarod system. After 3 days of growth, the biofilm was harvested as described above, and the resultant cell suspension was serially diluted and plated onto both BHI agar plates containing kanamycin and BHI agar plates containing streptomycin to determine the relative numbers of viable wild-type and mutant organisms present in the biofilm.
The raw data from the DNA microarray experiments were deposited in the NCBI Gene Expression Omnibus (GEO) database (http://www.ncbi.nlm.nih.gov/geo/) under accession numbers GSE4346 (for the iron-limited growth studies) and GSE4348 (for the biofilm growth studies).
We deduced the metabolic capabilities of M. catarrhalis ATCC 43617 from the genomic sequence data for two reasons: to validate the completeness of our M. catarrhalis DNA microarrays for measuring global gene expression in this organism and to determine likely changes in metabolism, if any, that might result from growth in the biofilm state. To assess the completeness of the DNA microarray, we determined the probable functions of the ORF-encoded protein products, deduced the metabolic capabilities of the organism from these proposed functions, and compared these capabilities to known metabolic characteristics of M. catarrhalis (Table (Table22).
An unusual feature of the deduced metabolism of M. catarrhalis is its apparent inability to utilize any exogenous carbohydrate. This bacterium appears to lack intact glycolytic pathways, all carbohydrate catabolic enzymes, all carbohydrate transport systems, and all anapleurotic reactions (Table (Table2).2). This is consistent with data in Bergey's Manual of Determinative Bacteriology (35) which indicate that M. catarrhalis does not produce acid from glucose and does not utilize any carbohydrate tested (for a review, see reference 16). The inability to utilize exogenous carbohydrates suggests that gluconeogenesis is essential, and the apparent presence of all the gluconeogenic enzymes is consistent with this conclusion.
M. catarrhalis appears to possess all of the enzymes required for aerobic energy metabolism. Both subunits of succinyl coenzyme A synthetase are missing, but otherwise this bacterium has an intact citric acid cycle, glyoxylate cycle, electron transport system, and ATP synthase. The utilization of acetate by M. catarrhalis (35) is consistent with a functional glyoxylate cycle. Enzymes associated with anaerobic metabolism (nitrate reductase, nitrite reductase, nitric oxide reductase, and acetate kinase-phosphotransacetylase) are also present and perhaps could be used in a low-oxygen or microaerophilic environment. The reported utilization of nitrate by M. catarrhalis (16) is consistent with the presence of nitrate reductase.
The genome sequence also suggests several unusual aspects of nitrogen metabolism. First, M. catarrhalis appears to have both pathways for ammonia assimilation: a glutamate dehydrogenase pathway and a glutamate synthase-glutamine synthetase pathway. However, despite the presence of genes for these enzymes, it is not apparent that M. catarrhalis can assimilate ammonia. It lacks all regulators that are associated with nitrogen limitation in organisms such as Escherichia coli: RpoN, PII, NtrB, and NtrC. Furthermore, the observation that an ammonium salt did not stimulate M. catarrhalis growth in a defined medium (38) implies that M. catarrhalis cannot assimilate even high ammonia concentrations. Although the genes for enzymes of ammonia assimilation are present, it is not clear that they function in this capacity. Their sole function may be in glutamate and glutamine synthesis, and they may be insufficiently expressed to participate in ammonia assimilation. Second, M. catarrhalis has intact pathways for the synthesis of all amino acids except proline and arginine. The latter finding is consistent with the report that M. catarrhalis is an arginine auxotroph (38). Proline auxotrophy has not been reported for M. catarrhalis. It is possible that proline can be synthesized from arginine (via ornithine), and all the enzymes required for this unconventional pathway appear to be present.
The deduced metabolic capabilities of M. catarrhalis suggest an unusual and unpredicted pattern of metabolism. Nonetheless, all of the conclusions described above are consistent with published observations. Most importantly, every pathway predicted to be complete is intact. For example, M. catarrhalis is a histidine prototroph, and all of the numerous different enzymes required for histidine synthesis are encoded by the genome. The same is true for processes such as protein synthesis, since we were able to identify all of the tRNA synthetases. Furthermore, the few pathways predicted to have deficiencies (e.g., arginine synthesis) are missing only one or two enzymes. Therefore, the oligonucleotide probes used in the M. catarrhalis DNA microarray described in this study account for the vast majority, if not all, of the genes encoding metabolic enzymes of M. catarrhalis and, by extrapolation, perhaps all of the genes of this pathogen.
To establish the utility of the DNA microarray developed by using the genome information from M. catarrhalis ATCC 43617, we first performed studies using this DNA microarray to analyze the transcriptomes from M. catarrhalis cells grown under iron-replete and iron-limiting conditions. Previous studies with other M. catarrhalis strains had shown that the chelator Desferal could be used to effectively limit iron availability in vitro (3), and the identities of several different M. catarrhalis proteins whose expression is affected by the availability of iron in the growth environment have been well established (3, 11, 14, 43, 44). To determine the concentration of Desferal necessary to effect iron-limited growth of M. catarrhalis strain ATCC 43617, this strain was grown in BHI broth in the presence of increasing concentrations of this chelator (Fig. (Fig.1A).1A). It was found that 30 μM Desferal effectively limited growth of this strain (Fig. (Fig.1A).1A). When protein expression by the cells was examined, it was found that the cells grown in the presence of both 30 and 50 μM Desferal exhibited increased expression of the CopB protein (Fig. (Fig.1B).1B). Expression of CopB in other M. catarrhalis strains has previously been shown to be regulated by the availability of iron (3, 15). These results indicated that growth of M. catarrhalis ATCC 43617 in the presence of 30 μM Desferal was iron limited.
Having established iron-limiting growth conditions for this strain, we proceeded to extract total RNA from M. catarrhalis ATCC 43617 cells grown into logarithmic phase in BHI broth with and without 30 μM Desferal. DNA microarray analysis revealed that the expression of over 100 different genes was affected at least twofold by iron limitation (42 genes with increased expression and 65 genes with decreased expression [Table [Table3])3]) . The genes whose expression was most highly up-regulated by iron limitation, as determined by DNA microarray analysis, included lbpB, lbpA, tbpB, copB, and several other genes predicted to encode proteins involved in iron uptake (Table (Table3).3). The first four genes listed above have previously been shown to be up-regulated by iron limitation (3, 12, 14, 43). The majority of the genes whose expression was most adversely affected by iron limitation encoded ribosomal proteins (Table (Table3);3); this is likely due to the growth inhibition caused by iron limitation (Fig. (Fig.1A).1A). These DNA microarray data demonstrating up-regulation of M. catarrhalis genes known to be affected by iron limitation indicated that this DNA microarray can be used to assess global gene expression by this pathogen.
M. catarrhalis ATCC 43617 was grown in BHI broth into the mid-logarithmic phase of growth or in the Sorbarod biofilm system. Total RNA extracted from both the broth-grown and biofilm-grown cells was subjected to DNA microarray analysis. For 83 genes, expression was affected at least twofold by growth in the biofilm state relative to growth in the planktonic state (54 genes with increased expression and 29 genes with decreased expression [Table [Table4])4]) . The genes whose expression was increased the most by growth in a biofilm were predicted to encode components of the nitrate respiratory chain (Table (Table4).4). These genes included genes encoding enzymes involved in nitrate reduction, a nitrite reductase, and a nitric oxide reductase. Other up-regulated genes included genes predicted to encode the Lon protease and DnaK (Table (Table4).4). The genes whose expression was most reduced by growth in a biofilm were mainly the genes predicted to encode ribosomal proteins (Table (Table44).
Real-time qRT-PCR analysis was used to confirm the validity of the DNA microarray results (Table (Table4).4). Twenty different M. catarrhalis ATCC 43617 genes were selected for this analysis. These genes included 13 genes whose expression appeared to be up-regulated by growth in a biofilm, as measured by DNA microarray analysis, 3 genes that appeared to be down-regulated, and 4 genes representative of the genes that were apparently unaffected (Table (Table4).4). The values obtained by qRT-PCR correlated well with those obtained from DNA microarray analysis (Fig. (Fig.22).
The M. catarrhalis genes that were proportionally up-regulated the most by growth in the Sorbarod biofilm system included narG, narH, narI, and narJ (Table (Table4).4). In other organisms, including E. coli (72) and Mycobacterium tuberculosis (66), these four genes form an operon or are grouped together. In M. catarrhalis ATCC 43617, these genes are similarly located together, and RT-PCR indicated that these four ORFs are cotranscribed in M. catarrhalis to form a single mRNA transcript (data not shown). To determine whether lack of expression of the narGHIJ operon affected the ability of M. catarrhalis to form a biofilm, the narG and narH genes were inactivated in M. catarrhalis strain ETSU-9. The ETSU-9 strain was used for mutant construction because we have been unsuccessful to date in constructing mutants of any type in M. catarrhalis ATCC 43617 (data not shown). A narGH mutant was constructed by deleting the majority of the narG ORF and the 5′ end of the narH gene and replacing this DNA with a kan cartridge. When tested for its ability to form a biofilm in the Sorbarod continuous-flow biofilm system, however, the mutant did not exhibit any apparent deficiency in biofilm formation (data not shown).
The deduced metabolism of M. catarrhalis was used to assess the completeness of the available nucleotide sequence from the genome of M. catarrhalis ATCC 43617 and to provide a context for interpreting transcriptome profiles. In addition, a comparison of the metabolism of M. catarrhalis with that of two other gram-negative colonizers of the nasopharynx, H. influenzae and N. meningitidis, has the potential to provide some insight into metabolic approaches used by these three organisms to survive in the environment of the nasopharynx. Table Table55 presents some key metabolic features of these three pathogens, with an emphasis on significant differences.
At the outset, it is readily apparent that these organisms utilize different metabolic strategies for survival in the nasopharynx. All three organisms grow readily under aerobic conditions, while S. pneumoniae is an aerotolerant anaerobe. H. influenzae can grow anaerobically in the apparent absence of alternative respiratory substrates (e.g., nitrate) (20, 21), and both growth experiments (21) and analysis of encoded protein products of the H. influenzae Rd genome (69) indicated that this organism might prefer growth under reducing conditions. N. meningitidis cannot grow under strictly anaerobic conditions but has effective systems for using both nitrite and nitric oxide as respiratory substrates (5, 62). Similar to the meningococcus, M. catarrhalis cannot grow anaerobically (4, 35). However, in addition to containing the genes encoding predicted nitrite and nitric oxide reductases (Table (Table4),4), M. catarrhalis also possesses the ability to reduce nitrate to nitrite (16, 35). Interestingly, during growth in a biofilm, these genes encoding the enzymatic machinery necessary to reduce nitrate to the level of nitrous oxide were among those that were most highly expressed (Table (Table44).
M. catarrhalis differs substantially from the other two pathogens with respect to central metabolic pathways. M. catarrhalis is unable to utilize exogenous carbohydrates (16), apparently lacking both glycolytic pathways and sugar transport systems (Table (Table5).5). In contrast, both H. influenzae and N. meningitidis can utilize a limited number of carbohydrates (16, 35), which is consistent with the presence of intact glycolytic pathways and complete phosphotransferase sugar transport systems in these two organisms. In aggregate, these organisms can utilize only a limited variety of carbohydrates, which suggests that the nasopharynx may be restricted in carbohydrate diversity and perhaps availability. Only N. meningitidis appears to have a complete citric acid cycle. M. catarrhalis appears to lack both subunits of succinyl coenzyme A synthetase, whereas H. influenzae is missing several genes encoding citric acid cycle enzymes. Finally, M. catarrhalis has a glyoxylate cycle, whereas the other two organisms do not (Table (Table55).
Each of these organisms has certain nutritional deficiencies, which is not surprising considering their relatively small genomes. All three bacteria are missing crucial components for high-affinity ammonia assimilation but appear to possess low-affinity ammonia assimilation capability via the activity of glutamate dehydrogenase. However, M. catarrhalis has been reported to be unable to assimilate ammonia (38), which implies that M. catarrhalis is effectively a glutamate auxotroph. The absence of high-affinity ammonia assimilation in N. meningitidis and H. influenzae also implies that there is glutamate auxotrophy in the nasopharynx, unless a high concentration of ammonia is present. Given that M. catarrhalis requires at least arginine for growth and is unable to utilize carbohydrates, it seems likely that amino acids are available in the nasopharynx for the amino acid auxotrophies and energy, although the source of amino acids in the nasopharynx is not readily apparent. This conclusion is reinforced by the similar properties of N. meningitidis and H. influenzae: a limited ability to utilize carbohydrates, at least one requirement for an amino acid, and a potential glutamate auxotrophy in the absence of sufficient ammonia.
When total RNA extracted from M. catarrhalis cells grown under both iron-replete and iron-limiting conditions was subjected to DNA microarray analysis using probes derived from M. catarrhalis ATCC 43617, the genes that were markedly up-regulated included those previously shown by protein expression measurements (3, 12, 14, 43) to be affected by the availability of iron in the growth environment. The DNA microarrays were then used to identify M. catarrhalis genes whose expression was affected by growth in a biofilm. To date, there are only very limited data available about biofilm development by M. catarrhalis (13, 58). Similarly, there is limited information about biofilm formation by N. meningitidis (78), although genes induced or up-regulated by contact of this pathogen with human cells in an in vitro system have been identified by DNA microarray analysis (28). Studies of biofilm formation by H. influenzae are more extensive, and several gene products of this pathogen which are involved in or affected by biofilm growth have been identified (27, 50, 52, 68, 77). One of the H. influenzae gene products that is up-regulated by growth in a biofilm is the thiol-dependent peroxidase peroxiredoxin-glutaredoxin, and isogenic H. influenzae mutants unable to express this protein were shown to be deficient in biofilm formation in vitro (52). A similar peroxiredoxin was also up-regulated during growth of M. catarrhalis in a biofilm (MCORF783) (Table (Table4),4), although mutant analysis of this gene product was not performed in the present study.
In a preliminary effort to extend our findings with the DNA microarray-derived data and determine whether genes maximally up-regulated during biofilm growth were essential for this mode of growth, we inactivated two of the genes (narG and narH) which were among those most highly up-regulated by growth in a biofilm (Table (Table4).4). However, when tested in a competitive index experiment in the Sorbarod continuous-flow biofilm system, the narGH mutant did not appear to have a deficiency in the ability to form biofilms (data not shown). This result suggests that nitrate reductase activity, while substantially up-regulated during biofilm growth, is not essential for biofilm development in this model system as used in this study. The up-regulation of these particular genes may instead reflect some type of sensing of reduced oxygen tension in the biofilm state.
We also noted that expression of the ORFs predicted to encode nitrite reductase and nitric oxide reductase was highly up-regulated in the biofilm (Table (Table4).4). The ability of N. meningitidis to survive in nasopharyngeal tissue has been shown to be enhanced by nitric oxide detoxification systems (67), and in Pseudomonas aeruginosa nitric oxide is involved in signaling biofilm dispersal (8). The predicted ability of M. catarrhalis to reduce nitrate to the level of nitrous oxide may provide an alternative means for energy generation by this organism under oxygen-limited conditions. In addition, the ability to reduce nitric oxide may also provide M. catarrhalis with some level of protection against macrophage-generated nitric oxide. In this context, it is interesting to note that there has been one report describing the selection of M. catarrhalis variants or mutants that were more resistant to nitric oxide than their wild-type parent strain (45), but the identity of the relevant gene product(s) was not determined.
In summary, this study used nucleotide sequence data from the genome of M. catarrhalis ATCC 43617 to provide a preliminary analysis of M. catarrhalis metabolism and to construct DNA microarrays that were used to evaluate global gene expression under defined conditions in vitro. The resultant data indicate that growth of M. catarrhalis ATCC 43617 in a biofilm in vitro differentially affected the expression of genes in only a relatively few categories. The genes whose expression was most highly up-regulated were associated with energy generation involving the reduction of nitrate, nitrite, and nitric oxide. Expression of ribosomal genes was down-regulated, which is consistent with slower growth, while some heat shock genes had increased expression, which would be consistent with stress. More extensive genetic analyses are required to determine which M. catarrhalis genes are specifically required for biofilm formation in the continuous-flow system.
This study was supported by U.S. Public Health Service grant AI36344 to E.J.H. M.M.P. was supported by U.S. Public Health Service training grant 5-T32-AI007520.
We thank Steven Berk for supplying the ETSU-9 isolate of M. catarrhalis used in this study.
Editor: D. L. Burns
Published ahead of print on 9 July 2007.
†Supplemental material for this article may be found at http://iai.asm.org/.