OUP user menu

Expression of a peroxiredoxin-glutaredoxin by Haemophilus influenzae in biofilms and during human respiratory tract infection

Timothy F. Murphy, Charmaine Kirkham, Sanjay Sethi, Alan J. Lesse
DOI: http://dx.doi.org/10.1016/j.femsim.2004.12.008 81-89 First published online: 1 April 2005

Abstract

Evidence is mounting that nontypeable Haemophilus influenzae grows as a biofilm in the middle ear of children with otitis media and the airways of adults with chronic obstructive pulmonary disease. To begin to assess antigens expressed by H. influenzae in biofilms, cell envelopes of bacteria grown as a biofilm were compared to those grown planktonically. A ~30 kDa peroxiredoxin–glutaredoxin was present in greater abundance during growth in biofilms. Mutants deficient in expression of peroxiredoxin–glutaredoxin were constructed by homologous recombination in four clinical isolates. The mutants showed a 25–50% reduction in biofilm formation compared to the corresponding parent strains. To study in vivo expression of peroxiredoxin–glutaredoxin during human respiratory tract infection, paired pre- and post-exacerbation serum from adults with chronic obstructive pulmonary disease and H. influenzae in sputum were assayed using an enzyme-linked immunosorbent assay and purified recombinant peroxiredoxin–glutaredoxin. Eight from 18 (44.4%) paired serum samples showed a significant increase in antibody to peroxiredoxin–glutaredoxin from pre- to post-infection. These results indicate that (1) peroxiredoxin–glutaredoxin is present in greater abundance in H. influenzae biofilms compared to planktonically grown bacteria; (2) peroxiredoxin–glutaredoxin is involved in biofilm formation by H. influenzae and the degree of involvement varies among strains; and (3) peroxiredoxin–glutaredoxin is expressed by H. influenzae during infection of the human respiratory tract and is recognized by the human immune system.

Keywords
  • Haemophilus influenzae
  • Biofilms
  • Peroxiredoxin
  • Glutaredoxin
  • Chronic obstructive pulmonary disease

1 Introduction

Clinicians and microbiologists who deal with recurrent and chronic bacterial infections have defined a new category of infectious diseases that differs in a variety of ways from acute infections caused by free floating bacterial pathogens [1]. These infectious diseases are characterized by a less aggressive clinical course, persistence for months to years, and alternating periods of quiescence and acute exacerbations [13]. When grown on conventional media, the bacteria appear to be sensitive to antibiotics. These antibiotics fail to resolve infections but provide some benefit during acute exacerbations [1]. In some instances, it is not possible to culture bacteria using standard culture techniques. However, bacteria are present in a metabolically active form and can be observed microscopically in matrix-enclosed aggregates. Many such infections are caused by bacteria that grow as biofilms. A biofilm is a structured community of bacterial cells enveloped in a polymeric matrix and adherent to an inert or living surface [2,3]. Examples of human infections caused by bacterial biofilms include pulmonary infections in cystic fibrosis, chronic prostatitis, intravascular catheter infections, endocarditis, chronic osteomyelitis and chronic otitis media. The list of human infections caused by biofilms is growing rapidly. New approaches involving biofilm science are being applied to better understand the mechanisms of biofilm infection and to devise better ways to prevent and manage these difficult infections.

Nontypeable Haemophilus influenzae is an important cause of otitis media in children and lower respiratory tract infections in adults with chronic obstructive pulmonary disease (COPD) [48]. Several indirect lines of evidence indicate that H. influenzae grows in the form of a biofilm during colonization and infection of the human respiratory tract. Otitis media with effusion refers to the presence of fluid in the middle ear in the absence of symptoms of acute infection. Many middle ear fluids from children with otitis media with effusion are sterile by culture. However, DNA of nontypeable H. influenzae is present in a substantial proportion of culture-negative middle ear fluids [912]. Furthermore, bacterial mRNA is present as well, indicating the presence of viable bacteria [13]. These observations suggest that bacteria are present in a physiological state different from that of bacteria growing planktonically in a free floating state. Furthermore, H. influenzae biofilms have been observed microscopically in the middle ear of chinchillas [14]. In view of these observations in human samples and studies in animal models, several authors have proposed that H. influenzae is present in the form of a biofilm in the middle ear of children with otitis media [1317].

Observations of the dynamics of colonization of the respiratory tract in adults with COPD provide a growing interest in assessing the potential role of biofilms in chronic colonization by H. influenzae of a subset of these patients. In an ongoing study of bacterial infection in adults with COPD at the Buffalo Veterans Affairs Medical Center, we have observed that, while active turnover of strains occurs, selected strains have the ability to persist for many months [1821]. Monthly sputum cultures reveal intermittent negative cultures in spite of continuous colonization by the same strain [21]. These observations suggest that H. influenzae grows as a biofilm in the respiratory tract of adults with COPD, reminiscent of Pseudomonas aeruginosa and its well established propensity to grow as a biofilm in patients with cystic fibrosis [22].

Bacteria undergo marked metabolic alteration during growth as biofilms, including changes in expression of surface antigens. Characterizing the surface molecules expressed during growth as a biofilm will be important to understand the host immune response to bacteria that cause colonization and infection in the form of biofilms. Such information will have important implications in elucidating the targets of the human immune response and in designing strategies to prevent and treat infections caused by bacterial biofilms.

Nontypeable H. influenzae forms biofilms in vitro [23]. Strains of H. influenzae from middle ear fluids of children with otitis media and from sputum of adults with COPD show substantial variability in their ability to form biofilms [23]. Previous studies of cell envelopes during growth of H. influenzae as a biofilm demonstrated an increased abundance of a ~30-kDa protein [23]. The goal of this study is to evaluate this protein further, identified here as a peroxiredoxin–glutaredoxin (PGdx) [2426], that is expressed by H. influenzae during biofilm growth and to assess the potential expression of this protein during human respiratory tract infection.

2 Materials and methods

2.1 Bacterial strains and growth conditions

H. influenzae strains 6P8H1, 13P24H1, 49P5H1, 56P41H1, 67P38H1 and 69P4H1 were isolated from the sputum of six different adults with COPD followed in the COPD Study Clinic at the Buffalo VA Medical Center (see below). Strain 12 was kindly provided by Dr. Joseph St. Geme.

H. influenzae strains were grown on chocolate agar or, in experiments comparing biofilms and planktonic growth, on brain heart infusion (BHI) agar supplemented with hemin and nicotinamide adenine dinucleotide both at 10 µg ml+1. When grown in liquid media, BHI supplemented with hemin and NAD was used.

2.2 Biofilm growth assay

The assay to grow and quantitate nontypeable H. influenzae biofilms has been described previously [23] and is essentially the same assay which has been used for several other bacterial species [2730]. An overnight broth culture was diluted 1:200 in fresh broth and 200 µl was inoculated into the wells of a 96-well Linbro tissue culture plate (ICN Biomedical, Inc., Aurora, OH). The plates were incubated at 37 °C on a nutator for 24 h. Before biofilm quantitation, growth was assessed measuring the optical density at 490 nm (OD490) in a BioRad plate reader. To quantitate biofilm formation, 20 µl of Difco crystal violet (Becton Dickinson, Sparks, MD) was added to each well and incubated at room temperature for 15 min. Wells were washed vigorously with distilled water and the plate was air dried. A volume of 230 µl of 95% ethanol was added to each well and the OD570 was measured. Each plate included four wells, which contained sterile broth instead of bacteria but were treated identically otherwise. The OD570 was standardized against these wells. Strains were tested in quadruplicate wells. The mean and standard deviation were calculated from values for each strain. Comparison of biofilm formation between mutant and parent strains was performed using the unpaired t-test.

2.3 Purification of cell envelopes

To study the cell envelopes of nontypeable H. influenzae, bacteria were grown as a biofilm overnight in a 24-well Linbro tissue culture plate. Broth was aspirated and the wells were washed three times with 1 ml of 0.01 M HEPES, pH 7.4. Bacteria in the biofilm were harvested by scraping the wells with the tip of a 1-ml micropipette tip. To prepare cell envelopes, cells were suspended in 0.01 M HEPES, pH 7.4 and sonicated on ice using a Branson Sonifier (small tip, setting 7) [31]. The suspension was centrifuged at 12,000g for 2 min at 4 °C to remove unbroken cells and debris. The supernatant was recovered and centrifuged at 12,000g for 45 min. The resulting pellet was suspended in sample buffer and subjected to sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS–PAGE) and immunoblot assay.

2.4 SDS–PAGE and immunoblot assays

Cell envelopes were subjected to SDS–PAGE and Coomassie blue staining using previously described methods [31]. Immunoblot assays were performed as described previously [32]. After incubating with a murine antiserum against PGdx, blots were developed with peroxidase conjugated secondary antibody and color developer [32].

2.5 Cloning the gene that encodes PGdx

Oligonucleotide primers corresponding to 5′ and 3′ termini of the gene that encodes PGdx (open reading frame HI0572 in the genome sequence of H. influenzae Rd [33]) were designed with BamHI and EcoRI sites (Table 1). The primers were used in PCR reactions to amplify the gene from strain 67P38H1 using standard methods. The DNA fragments were directionally cloned into pRSETB (Invitrogen, San Diego, CA) that had been digested with BamHI and EcoRI. The sequence of the entire insert of the resulting plasmid, p572Sb was confirmed by sequencing.

View this table:
Table 1

Oligonucleotide primer sequences

PrimerSequence
30kd5′GCGCGCGGATCCATGGAAGGAAAAAAAAGTCC
30kd3′ATATATGAATTCTGCAAAGTATTTTTCTAAATC
HI05′GAGGATCCAAGTGCGGTCACTTGCCACACGCTCATAACG
HI03′TAGGTACCACCGCACTTGCACCAGCACCTTGTTCGCC
  • Restriction enzyme sites are underlined. H. influenzae specific uptake sequence is in bold.

2.6 Expression and purification of recombinant PGdx

To express recombinant PGdx containing a 6-histidine tag on its amino terminus, a single colony of clone p572Sb in E. coli BL21(DE3) was inoculated into 2-ml TB broth containing 200 µg ml+1 carbenicillin and the culture was incubated at 37 °C until the cells were in logarithmic phase. Cells were centrifuged and resuspended in 2 ml of fresh TB broth; a volume of 0.1 ml was used to inoculate 8 ml of TB broth containing 500 µg ml+1 of carbenicillin. This culture was incubated at 37 °C until cells were in the logarithmic phase. Cells were centrifuged for 5 min at 6000g. Cells were resuspended in 50 ml TB broth containing 500 µg ml+1 carbenicillin and 1 mM isopropylthiogalactopyranoside (IPTG). After incubating at 30 °C for 2 h, cells were harvested by centrifugation at 6000g for 10 min at 4 °C. The cells were resuspended in 5 ml of 50 mM NaH2PO4, 10 mM Tris–HCl, 6 M guanidine, 100 mM NaCl, 1 mM Pefabloc™ (Boehringer Mannheim), pH 8.0 (lysis buffer) and mixed on a nutator for 20 min at room temperature to lyse cells. The suspension was centrifuged at 10,000g for 10 min at 4 °C; and the supernatant, which represented the bacterial lysate, was saved.

An aliquot of Talon™ resin (Clontech, Palo Alto, CA) was centrifuged at 750g for 5 min at 4 °C. The resulting pellet was suspended in 10 volumes of lysis buffer. After mixing, the resin was centrifuged and the supernatant was discarded. The resin was resuspended in the bacterial lysate (1-ml resin per 75 ml of culture) and mixed on a nutator for 20 min at room temperature. The suspension was centrifuged at 750g for 5 min at 4 °C and the resin was saved. The resin was washed four times in 10 volumes of lysis buffer and once in 0.02 M Tris–HCl, 0.5 M NaCl, 1%β-octylglucoside, pH 8.0 (TON buffer) by sequential centrifugation and resuspension. To elute recombinant PGdx, the washed resin was resuspended in TON buffer containing 0.05 M EDTA and incubated at room temperature on a nutator for 20 min. The resin was removed by centrifugation and the elution was repeated two additional times. The eluted protein was defiltrated with TON buffer to remove EDTA. The protein was tested by SDS–PAGE and the protein concentration was determined by bicinchoninic acid (BCA) assay (Pierce, Rockford, IL).

2.7 Development of antiserum to PGdx

Balb/c mice were immunized intraperitoneally with 25 µg recombinant PGdx emulsified with incomplete Freund's adjuvant on days 0, 14 and 31. Blood was drawn on day 38 and serum was collected by centrifugation after the blood was allowed to clot. Serum from 10 mice was pooled.

2.8 Construction of PGdx deficient mutants

To construct PGdx deficient mutants, oligonucleotide primers were designed to correspond to a sequence 1030 base pairs upstream and 1130 downstream of HI0572 (Table 1). These primers, which included KpnI and BamHI sites, were used to amplify by PCR a ~3 kb-fragment that included HI0572 and ~1 kb upstream and ~1 kb downstream using genomic DNA of strain 12. The amplicon was ligated into pRSETB, transformed into E. coli and the resultant plasmid was isolated. A kanamycin cassette amplified by PCR from pACYC177 was ligated into an AccI site located 200 bp downstream of the start codon of HI0572. The resulting plasmid was linearized with NheI and used to transform H. influenzae strains 6P8H1, 13P24H1, 49P5H1, 56P41H1 and 67P38H1 after the induction of competence by the M IV method [34]. Mutants were selected on BHI supplemented with hemin and NAD and containing 15 µg ml+1 ribostamycin, an aminoglycoside antibiotic that inhibits the growth of kanamycin-resistant transformants of H. influenzae while minimizing the selection of spontaneous “breakthrough” kanamycin resistance.

2.9 COPD study clinic

H. influenzae strains and serum samples were obtained from patients followed in the COPD Study Clinic at the Buffalo Veterans Affairs Medical Center which has been described previously [18,35,36]. Briefly, patients with chronic bronchitis are followed prospectively and evaluated monthly. Clinical criteria are used to define exacerbations [18]. Eighteen exacerbations were identified during which H. influenzae was isolated from a sputum sample and following which a new serum IgG response was detected to the homologous isolate [37]. Pre- and post-exacerbation serum samples from these 18 episodes of infection were used to study antibody responses to PGdx.

2.10 Serum samples

Blood was obtained at monthly clinic visits using venipuncture technique and was allowed to clot. Serum was obtained by centrifugation and was stored at +80 °C. Pre-exacerbation serum samples were obtained approximately four weeks prior to the exacerbation (range two to eight weeks). Post-exacerbation samples were obtained four to eight weeks following the exacerbation.

2.11 Enzyme-linked immunosorbent assay

Wells of a 96-well microtiter plate (Immulon 4; Dynatech) were coated overnight at room temperature with 10 µg ml+1 recombinant PGdx in 0.1 M sodium carbonate, 0.1 M sodium bicarbonate (pH 9.6). The wells were washed three times between each step with PBS containing 0.05% Tween 20 (PBS-Tween). Wells were blocked by adding 3% nonfat dry milk in PBS-Tween and incubated at room temperature for 1 h. After washing the wells, serum diluted in 1% nonfat dry milk in PBS-Tween was added to the wells and incubated at 37 °C for 2 h. Horseradish peroxidase-conjugated rabbit anti-human IgG (Kirkegaard & Perry Laboratories, Gaithersburg, MD) diluted 1:3000 in 3% goat serum was added to the wells and incubated for 1 h at room temperature. After washing the wells, color was developed by adding 0.1 mg of 3,3′,5,5′-tetramethylbenzidine-dimethyl sulfoxide–0.02% hydrogen peroxide per ml in 0.1 M sodium acetate adjusted to pH 4.5 with citric acid. After 15 min of incubation, the reaction was stopped by the addition of 4 N H2SO4. The optical density at 450 nm was monitored.

All samples were run in duplicate. A negative control included wells coated with PGdx and assayed with buffer in place of serum. A second negative control included wells that were “sham coated” with buffer in place of PGdx with each dilution of serum. The percentage change in OD between the pre-exacerbation samples and post-exacerbation samples for each dilution was calculated using the following formula: [(OD post-exacerbation serum + OD pre-exacerbation serum)/OD pre-exacerbation serum] ×100. To determine the percentage change which represents a significant increase in the antibody response from a pre-exacerbation to a post-exacerbation sample, Enzyme-linked immunosorbent assay (ELISA) was performed on 11 paired serum samples, collected two months apart from subjects in the COPD Study Clinic who had negative sputum cultures for H. influenzae. The percentage change in OD was calculated as described above.

3 Results

3.1 Identification of PGdx

To assess the expression of antigens by H. influenzae during biofilm growth, cell envelopes of sputum isolates grown as biofilms in supplemented BHI in wells of a 24-well microtiter plate were purified. Simultaneously, cell envelopes of the same isolates were purified using the same method from cells grown planktonically on supplemented BHI agar plates. Fig. 1 is a Coomassie stained SDS gel that shows the presence of a ~30-kDa band in the cell envelopes of biofilm grown cells in four isolates of H. influenzae; the band is absent or clearly less prominent in planktonically grown cells.

Figure 1

Coomassie blue stained SDS gel of cell envelope preparations. H. influenzae strain numbers are noted at the bottom. The lanes marked “P” were grown planktonically and the lanes marked “B” were grown as biofilms. The arrows denote the ~30-kDa band that is present in increased abundance in biofilms. Molecular mass standards are noted in kDa on the left.

Cell envelopes of strain 67P38H1 grown as a biofilm were subjected to SDS–PAGE, transferred to polyvinylidene difluoride and the band was excised. The amino-terminal sequence was identified as: MEGKKVPQVTFRTRQGDK. The sequence was queried versus the nonredundant Genbank protein database using the Web based BLASTP program (http://www.ncbi.nlm.nih.gov:80/BLAST/) with parameters set for short, nearly identical matches. This 18 amino acid query was a 100% match to amino acids 4–21 of locus HI0572 of the H. influenzae Rd KW20 genome (Genbank accession number U32739). We conclude that the PGdx protein encoded by HI0572 was present in increased abundance during growth of H. influenzae as a biofilm.

3.2 Analysis of sequence

Sequence analysis of open reading frame (ORF) HI0572 in the H. influenzae Rd genome revealed that the gene codes for a peroxiredoxin hybrid composed of an N-terminal peroxiredoxin fused to a C-terminal glutaredoxin [26,38]. However, the database protein is annotated to begin with an ATG codon 9 bp upstream of the codon corresponding to the N-terminal amino acid sequence determined experimentally above. This annotation likely represents the common annotation bias to include the longest ORF beginning with a methionine in proteins without experimental evidence of the actual start codon.

Oligonucleotide primers were used to amplify by PCR the gene from genomic DNA of strain 67P38H1 (Table 1); the sequence was determined. Using the methionine corresponding to the experimentally determined N-terminal amino acid sequence as the start codon, the sequence of the gene from strain 67P38H1 revealed a 714-bp ORF that encodes a 238 amino acid protein with a calculated molecular mass of 26,417 Da with a calculated isoelectric point of 5.50. The nucleotide sequence is 98% identical to HI0572 in H. influenzae Rd and the translated protein is 98% identical and 99% similar to the protein in H. influenzae Rd. The protein is in the family of peroxiredoxin and glutaredoxin hybrid molecules that contain both an N-terminal alkyl hydroperoxide reductase conserved domain fused to a C-terminal GrxC, glutaredoxin conserved domain [26,38,39].

3.3 Characterization of mutants deficient in PGdx

Mutants deficient in expression of PGdx were constructed by homologous recombination in strains 6P8H1, 13P24H1, 56P41H1 and 67P38H1. Fig. 2 (top panel) shows an SDS gel of cell envelopes of three mutants and corresponding parent strains. The ~30-kDa band is absent in the mutants compared to the parent strains. No other differences in band patterns are apparent. An immunoblot assay using antiserum against recombinant PGdx confirms the absence of the 30-kDa band in the mutants (Fig. 2 (bottom panel)). Genomic DNA from mutant and corresponding parent strains was cut with EcoRI and subjected to agarose gel electrophoresis and Southern blot assay using a probe corresponding to a 200-bp fragment of the kanamycin cassette used to construct the mutants. Fig. 3 shows that a single kanamycin cassette is present in each of the mutants on a fragment of identical size (~6 kb) in each mutant.

Figure 2

Coomassie stained SDS gel (top panel) and immunoblot assay (bottom panel) of cell envelope preparations of PGdx mutants and corresponding parent strains. The immunoblot assay was probed with polyclonal mouse antiserum to PGdx. Lanes a, 6P8H1; b, 6P8H1 PGdx mutant; c, 56P4H1; d, 56P4H1 PGdx mutant; e, 67P38H1; f, 67P38H1 PGdx mutant. Arrows show the location of the PGdx band in the parent strains. Molecular mass standards are noted in kDa on the left.

Figure 3

Southern blot assay of genomic DNA cut with EcoRI and probed with a labeled fragment of the kanamycin cassette inserted into mutants. Lanes contain DNA from strains: a, 6P8H1; b, 6P8H1 PGdx mutant; c, 13P24H1; d, 13P24H1 PGdx mutant; e, 49P5H1; f, 49P5H1 PGdx mutant; g, 56P41H1; h, 56P41H1 PGdx mutant; i, 67P38H1; j, 67P38H1 PGdx mutant. Molecular size standards are shown on the right in kilobases.

Fig. 4 shows the genes surrounding pxdx in H. influenzae Rd. To identify the genes surrounding pgdx in the clinical isolates and mutants, oligonucleotide primers HI05′ and HI03′ were used to amplify by PCR the region surrounding the gene. All six mutant strains yielded an ~4-kb amplicon as predicted by the genome sequence of Rd. The nucleotide sequence of the entire amplicon of the four mutant strains was determined and showed the identical arrangement of genes as depicted in Fig. 4 and confirmed that the kanamycin cassette was present in pgdx.

Figure 4

Diagram of HI0572 (peroxiredoxin–glutaredoxin) and surrounding genes in H. influenzae Rd [33]. The numbers at the top indicate location in the genome. The gene designations are annotated below the line. Arrows indicate gene direction.

No differences between mutant and parent strains were observed in colony morphology on chocolate agar, Gram stain appearance or growth rate in broth with shaking. The ability of mutants to form biofilms was compared with the corresponding parent strains. The four PGdx mutants showed approximately 25–50% reduction in biofilm formation compared to their corresponding parent strains (Fig. 5, top). When strains were grown in 96-well plates under the conditions to quantitate biofilm formation, the mutants tended to show reduced growth indicated by turbidity in wells compared to parent strains. To compensate for this growth difference, a second series of experiments to compare biofilm formation between mutant and parent strains was performed. Wells were inoculated with double the number of bacteria with mutants compared to parent strains. The level of growth in broth measured by turbidity of mutant and parent strains was comparable under these conditions. Fig. 5 (bottom) shows that the statistically significant reduction in biofim formation by mutant strains of 6P8H1, 13P24H1 and 56P41H1 was retained but the difference was lost for the mutant of strain 67P38H1 compared to parent strains. We conclude that PGdx is involved in biofilm formation in some strains of H. influenzae and that the degree of involvement varies among strains.

Figure 5

Results of biofilm growth assays of PGdx mutants and corresponding parent strains. Y-axes are OD570 indicating level of biofilm formation. Strain numbers are shown on the X-axes. Cross-hatched bars represent PGdx mutants and solid bars represent corresponding parent strains. Top graph shows results with wells inoculated with a 1:200 dilution of broth culture for all strains. Bottom graph shows results in which mutant strains were inoculated with double the inoculum of the parent strains to compensate for slower growth of mutants in microtiter wells. Values denoted by bars represent the mean of quadruplicate values. Error bars represent standard deviation. P values show differences between corresponding mutant and parent strains calculated using the unpaired t-test.

3.4 Purification of recombinant PGdx

Based on the sequence of strain Rd, oligonucleotide primers were designed and the gene that encodes PGdx was amplified from genomic DNA of H. influenzae 67P38H1. The gene was cloned into pRSETB and the protein was expressed with an amino-terminal six-histidine tag. The recombinant protein was purified by elution from a Talon column and was subjected to SDS–PAGE to assess purity. The purified recombinant PGdx separates as a single band at a molecular mass slightly higher than the native protein owing to the six-histidine tag. The purified recombinant protein was used to immunize mice and to perform ELISAs to measure the human immune response.

3.5 Serum IgG response to PGdx following infection with H. influenzae

To assess the extent to which some of the new antibodies in samples from patients who developed new serum IgG to their homologous infecting strain following exacerbations were directed at PGdx, an ELISA using purified recombinant PGdx was performed. Serum samples collected one month prior to and one month following an exacerbation during which H. influenzae was isolated from sputum were assayed. The percentage change in OD between the pre-exacerbation samples and post-exacerbation samples for each dilution was calculated using the following formula: [(OD post-exacerbation serum + OD pre-exacerbation serum)/OD pre-exacerbation serum] ×100.

To determine the cutoff value for a significant percent change in serum IgG levels to PGdx from pre- to post-exacerbation samples, 11 pairs of control serum samples collected two months apart from subjects in the COPD Study Clinic who had negative sputum cultures for H. influenzae were subjected to ELISA with recombinant PGdx. These paired control serum samples demonstrated a 7.03%± 3.07 (mean ± standard deviation) change when tested with PGdx. A change in OD of 14.9% represented the 99% confidence interval for the control samples. Therefore, any change in OD of >14.9% between pre- and post-exacerbation serum samples was regarded as a significant change.

Fig. 6 shows the results of ELISAs with PGdx in 18 paired pre- and post-exacerbation samples known to contain new antibodies to the infecting strain. Eight of the samples showed significant increases in OD from pre- to post-exacerbation samples. The level of antibody increase from pre- to post-exacerbation sera in the eight serum pairs ranged from 24% to 250% increase. The results indicate that in these eight exacerbations, patients made new serum IgG to PGdx in response to infection by H. influenzae.

Figure 6

Results of ELISA with pre- and post-exacerbation serum samples with purified recombinant PGdx. X-axis: 18 exacerbations; Y-axis: % increase in optical density from pre- to post-exacerbation. The cutoff (14.9%) for significant increase from pre- to post-exacerbation increase is shown.

4 Discussion

PGdx is present in increased abundance in H. influenzae biofilms in vitro compared to planktonically grown cells (Fig. 1). This observation led us to perform experiments to test the hypothesis that the PGdx is expressed during human infection. Analysis of paired serum samples from well-characterized episodes of respiratory tract infection caused by H. influenzae demonstrated that eight of 18 patients made antibodies to PGdx following infection (Fig. 6). These immunoassays were designed to detect an increase from pre-infection to post-infection levels, thus controlling for background levels of antibodies. Therefore, the assays detected antibodies to PGdx made specifically during the episode of infection. These results indicate that PGdx is expressed during human respiratory tract infection.

H. influenzae PGdx belongs to a family of thiol-dependent peroxidases which are typified by a chimeric structure consisting of an N-terminal peroxiredoxin domain and a C-terminal glutaredoxin domain. While many bacterial species have separate proteins that encode a peroxiredoxin and/or a glutaredoxin, the chimeric molecule is seen in a number of human pathogens in addition to H. influenzae, including Yersinia pestis, Vibrio cholerae, Pasteurella multocida, Neisseria meningitidis, Bordetella pertussis and others. These molecules are implicated in the defense of bacterial cells against oxidative stress as well as in other regulatory processes, including transcription, apoptosis and cell signaling [24,38,39]. One might speculate that expression of PGdx by H. influenzae in biofilms represents part of a stress response induced by conditions in the human respiratory tract. Indeed, the airways in patients with COPD are characterized by oxidative stress and depletion of anti-oxidative defenses [40,41].

PGdx-deficient mutants demonstrated similar colony morphology and Gram stain appearance compared to parent strains. While pairs of mutant and parent strains showed similar growth rates in shaking broth cultures, mutants showed a reduction in biofilm formation, indicating that PGdx is important in biofilm formation. The reduction in biofilm formation observed with disruption of pgdx varied somewhat among strains. In view of the observation that multiple genes are likely involved in biofilm formation, we speculate that clinical isolates of H. influenzae vary in the presence of genes that may have redundant functions with PGdx, accounting for variable effects of disruption of pgdx with regard to biofilm formation. This speculation is consistent with the recent observations that the genomes of various strains of H. influenzae show variability in the presence or absence of multiple genes [42,43].

We have not absolutely excluded the possibility that disruption of pgdx might have polar effects. However, analysis of the sequence of the region surrounding pgdx revealed that the downstream gene (HI0573, slyX) is oriented in the reverse direction (Fig. 4). This orientation suggests that transcription of downstream genes is unlikely to be affected by disruption of pgdx. Furthermore, analysis of cell envelopes revealed that the only apparent difference between mutant and parent strains was the absence of the ~30-kDa PGdx band. These observations support the conclusion that PGdx is involved in biofilm formation.

The PGdx protein has been expressed, purified and studied by Hwang et al. [26]. Kim et al. [38] determined the crystal structure of the protein. The molecule forms a tightly associated tetramer in which active sites of different monomers interact with each other. The gene that encodes PGdx lacks a signal sequence, suggesting a cytoplasmic location for the enzyme. However, the observation that the protein is present in cell envelope preparations suggests that PGdx may be associated with the cytoplasmic membrane. Further studies are required to characterize the cellular location of the protein.

H. influenzae forms a biofilm in vitro. H. influenzae biofilms have been observed microscopically in the middle ear of chinchillas [14]. The work described here shows increased abundance of PGdx by H. influenzae grown as a biofilm in vitro, and expression in vivo of PGdx by H. influenzae during infection of the human respiratory tract. While these two observations do not prove that H. influenzae grows as a biofilm during human infection, they represent another intriguing indirect evidence, suggesting that biofilm formation by H. influenzae is important in the human respiratory tract. Furthermore, our data clearly establishes that PGdx is expressed by H. influenzae during human respiratory tract infection and that humans develop antibody responses to the molecule.

Acknowledgements

This work was supported by NIH grant AI 19641 and by the Department of Veterans Affairs.

References

  1. [1].
  2. [2].
  3. [3].
  4. [4].
  5. [5].
  6. [6].
  7. [7].
  8. [8].
  9. [9].
  10. [10].
  11. [11].
  12. [12].
  13. [13].
  14. [14].
  15. [15].
  16. [16].
  17. [17].
  18. [18].
  19. [19].
  20. [20].
  21. [21].
  22. [22].
  23. [23].
  24. [24].
  25. [25].
  26. [26].
  27. [27].
  28. [28].
  29. [29].
  30. [30].
  31. [31].
  32. [32].
  33. [33].
  34. [34].
  35. [35].
  36. [36].
  37. [37].
  38. [38].
  39. [39].
  40. [40].
  41. [41].
  42. [42].
  43. [43].
View Abstract