Physiological Characterization of Haemophilus influenzae Rd Deficient in Its Glutathione-dependent Peroxidase PGdx*

The chimeric peroxidase PGdx of Haemophilus influenzae Rd belongs to a recently identified family of thiol peroxidases capable of reducing hydrogen peroxide as well as alkylhydroperoxides by means of glutathione redox cycling. In the present study, we constructed a H. influenzae Rd strain, deficient in its PGdx encoding gene (open reading frame HI0572). The mutant was shown by disk inhibition and liquid culture growth assays to exhibit increased susceptibility to organic hydroperoxides. The hampered growth was restored by complementing the interrupted gene on the genome with a replicating plasmid bearing an intact copy of the gene, hereby rejecting the possible influences of polar effects. Elevated levels of hydrogen peroxide scavenging activity, due to the catalase HktE, were measured in the absence of a functional pgdx gene rendering the mutant more resilient against hydrogen peroxide. On the other hand, after initiation of the stationary phase, aerobic cultures of the pgdx mutant were practically devoid of living cells, whereas wild-type counterparts retained viability. This observed feature was alleviated by complementation with the functional gene or with the addition of catalase.

Organisms inhabiting aerobic environments unavoidably encounter the downside effect of oxygen. That is, inadequate metabolic reduction of molecular oxygen to water results in the production of reactive oxygen species, including the superoxide anion radical (O 2 . ), hydrogen peroxide (H 2 O 2 ), and the hydroxyl radical (OH ⅐ ). Furthermore, these species are able to react with other cellular components causing damage and/or generating oxidizing derivatives such as alkylhydroperoxides.
The fastidious organism Haemophilus influenzae, etiological agent of inflictions such as meningitis and otitis media, not only has to endure oxidative stress resulting from its own aerobic metabolism, but together with macrophages in which it may reside (1), neighboring strains of peroxidogenic Streptococci (2) also add in the continuous exposure of the bacterium to reactive oxygen species.
Recently, we reported the initial characterization of H. influenzae Rd PGdx 1 (encoded by open reading frame HI0572) as a glutathione-dependent peroxidase capable of reducing both H 2 O 2 and alkylhydroperoxides (3). 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 (3,4,5). Previously, we have also shown that glutathione and the only catalase, HktE, provide overlapping protection in H. influenzae Rd against respiratory-generated H 2 O 2 (6). To investigate whether this GSH-based mitigation of H 2 O 2 is mainly attributable to PGdx, we constructed a knock-out in the pgdx gene.
Physiological studies, including growth assays and pro-oxidant sensitivity assays, allow us to conclude that PGdx acts as a major antioxidant in vivo. We also show that PGdx plays an essential role during stationary-phase growth of the bacterium, as observed in overnight grown aerobic cultures.

EXPERIMENTAL PROCEDURES
Materials-Restriction endonucleases were obtained from New England Biolabs (Beverly, MA). DNA purification from gel or solution was carried out using either the Qiaquick DNA Extraction or PCR Purification Kit (Qiagen, Crawley, UK). Ligations were performed using T 4 DNA ligase (Promega, Madison, WI). Plasmid DNA was prepared by the alkaline lysis method on either a small scale (7) or a 30-ml scale using the Qiagen plasmid purification kit. H 2 O 2 , tert-butyl hydroperoxide (t-BOOH), cumene hydroperoxide (COOH), methyl viologen and bovine liver catalase were obtained from Sigma.
Bacterial Strains, Media, and Growth Conditions-Escherichia coli MC1061 and TOP10 (Invitrogen, Paisley, UK) were used as hosts for cloning. All E. coli strains were cultured at 37°C in Luria-Bertani medium on an orbital shaker rotating at 200 rpm. When appropriate, 100 g of carbenicillin, 25 g of chloramphenicol, or 25 g of kanamycin were added per ml of E. coli culture media. H. influenzae Rd (KW20) was obtained from ATCC (Manassas, VA; number 51907). H. influenzae Rd was grown at 37°C under a 3% CO 2 atmosphere (candle extinction jar method) on an orbital shaker rotating at 180 rpm (unless otherwise stated). H. influenzae Rd medium consisted of brain-heart infusion (BHI) liquid (Difco) supplemented with ␤-NAD and hemin (Fluka, Glossop, UK). When appropriate, 2 g of chloramphenicol or 7 g of kanamycin were added per ml of H. influenzae Rd media. Solid media for all strains were prepared by adding agar to the liquid media to a final concentration of 1.8%. See Table I for list of bacterial strains and plasmids.
Construction of pgdx Mutant-A 4.0-kb DNA fragment, including the pgdx gene (HI0572) and its 1.4-kb upstream and 2.0-kb downstream region, was amplified by PCR from H. influenzae Rd genomic DNA (prepared as described elsewhere (8)) using Expand Long Template DNA polymerase (Roche Applied Science) and the following primers: forward primer (5Ј-CGC GGA TCC TGC CTG AAC TTT CGC GTA  ATA-3Ј) and reverse primer (5Ј-CGC GGA TCC TGT TTG ATT TGG  CGG ATG TA-3Ј), both containing a BamHI (underlined) restriction site. The obtained PCR product was purified and cloned into the TOPO-XL vector (Invitrogen), generating plasmid pSG4.0. The * This work was supported by the Institute for the Promotion of Innovation by Science and Technology in Flanders (IWT, Grant 3072) (to F. P.) and by the Fund for Scientific Research-Flanders (FWO, Grant 3G003601) (to J. J. V. B.). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
chloramphenicol acetyltransferase (Cam r ) gene was amplified from pACYC184 (New England Biolabs) together with its native constitutively expressed promoter using AccuPrime Taq DNA polymerase (Invitrogen), the forward primer 5Ј-AAC TGC AGT AGA CAG CAC CTC AAA AAC ACC ATC ATA CAC-3Ј and the reverse primer 5Ј-AAT ACG TAG TCT ACC AGG CGT TTA AGG GCA CCA ATA ACT-3Ј. Both primers contain an AccI (underlined) restriction site. The amplified resistance gene was cloned into pGEM-T (Promega, Madison, WI) prior to digestion with AccI. The digested fragment was inserted into the AccI digested construct pSG4.0, yielding plasmid pSGKO in which the pgdx gene was disrupted by the Cam r gene. The pgdx::Cam r cassette was amplified by PCR from pSGKO and used to replace the functional pgdx gene by homologous recombination (transformation of H. influenzae Rd was performed as described elsewhere (8), and positive transformants were scored by chloramphenicol resistance). Functional Complementation-Functional complementation of pgdx was accomplished as follows. A fragment containing the complete pgdx gene and its upstream and downstream region was generated by PCR and the following primers: forward primer (5Ј-CAA CGT TGA CCA ATT GTT CTA ATA ATT GAT GC-3Ј) containing a HincII restriction site (underlined), reverse primer (5Ј-AAA ACT GCA GCA ATT GCG TTA TAT GG-3Ј) containing a PstI restriction site (underlined). After digestion with both HincII and PstI, the fragment was ligated into a HincII/ PstI linearized pACYC177 (New England Biolabs) vector producing the complementation vector pCOMP. This complementation vector was transformed into a pgdx mutant of H. influenzae Rd, and positive transformants were selected by kanamycin resistance.
Southern Blot Analysis-Purified H. influenzae Rd chromosomal DNA was overnight digested with MfeI and ScaI and subjected to agarose gel (w/v 1%) electrophoresis. The DNA was transferred to nitrocellulose paper and probed by Southern blot analysis as previously described by Ref. 9. Digoxigenin-labeled DNA probes were constructed by PCR using the forward primer Probe A 5Ј-TTC AAT TAA AAG AAT AGA AT-3Ј, reverse primer Probe A 5Ј-ATT TTT TGA ATA CTG GCG CT-3Ј, forward primer Probe B 5Ј-AAC ATC TCT TTC ATT CCA GA-3Ј, reverse primer Probe B 5Ј-TTT ATC GTG TAA AAG TTG TT-3Ј.
Disk Diffusion Assay-Standing overnight grown cultures were diluted 1:20 in freshly prepared supplemented BHI (sBHI) medium without antibiotics and incubated without shaking at 37°C up to the midexponential phase. Cells were confluently plated onto sBHI plates containing no antibiotics using a sterile swab. In the center, a paper disk (diameter 5 mm, Whatman) was placed containing either 3% H 2 O 2 , 7% t-BOOH, 10% COOH, or 5 mM methyl viologen. Inhibition zone diameters were measured after incubating the plates for 48 h under a 3% CO 2 atmosphere at 37°C.
Growth Curves-Aerobic cultures were diluted 1:20 from standing overnight grown precultures in loosely capped culture tubes that fit tightly into the cuvette holder of a Shimadzu 1240 Mini Single-beam UV-visible spectrophotometer (Shimadzu, Duisburg, Germany) and were then incubated at 37°C on a shaker rotating at 180 rpm. Absorbance at 600 nm was measured at regular intervals over a period of 8.5 h. At the early exponential phase (A 600 0.25), cultures were stressed with different concentrations of hydroperoxides. The percent ratio growth rate of the treated samples to the untreated sample was calculated from these growth curves. For anaerobic experiments, cultures were prepared in a Coy chamber (Coy Laboratory Products, Inc.) under an atmosphere of 85% N 2 /10% H 2 /5% CO 2 . Culture tubes were closed with a silicone stopper prior the removal from the Coy chamber to preserve anaerobiosis during incubation.
Determination of H 2 O 2 Scavenging Activity-H 2 O 2 scavenging activity in whole-cells was either measured in the presence of low concentrations (1. For the experiments with high concentrations of H 2 O 2 cells were grown aerobically, pelleted during different stages of growth, washed twice with phosphate-buffered saline, and resuspended at an A 600 of 0.14. H 2 O 2 was added to a final concentration of 20 mM and the decrease in absorbance was monitored at 240 nm (⑀ 240 43.6 M Ϫ1 cm Ϫ1 ).
Catalase Activity Staining-Cell extracts of aerobically grown midexponential phase wild-type and mutant strains (15 g) were electrophoretically separated on a native polyacrylamide gel. Catalase activity was measured, using a ferricyanide-negative stain, as follows: the gel was soaked in 0.003% H 2 O 2 for 15 min, rinsed with distilled H 2 O, and stained with a 50:50 solution of 2% K 3 Fe(CN) 6 and 2% FeCl 3 ⅐6H 2 O.

Construction of a Mutant Deficient in PGdx Expression-To
investigate the in vivo relevance of PGdx as an antioxidant enzyme, we made a pgdx mutant utilizing the integrative disruption method (Fig. 1A). A 4.0-kb subgenomic fragment, encompassing the gene for pgdx (HI0572) and a 1.4-kb upstream  Fig. 1B; the mutant amplicon was enlarged with around 1,000 base pairs of the antibiotic resistance cassette) as well as by Southern hybridization (Fig. 2 A; probes used for Southern hybridization are depicted in Fig. 1A). Western blot analysis with mouse polyclonal antibody raised against recombinant PGdx confirmed the absence of peroxidase protein in mutant whole cell extracts (Fig. 2B). Phenotypic Characterization of the pgdx Mutant-The pgdx mutant and its wild-type parent were cultured both aerobically and anaerobically to distinguish for possible differences in growth due to environmental oxygen. As such, no dissimilarities were recorded when plotting the data of the growth curve (data not shown).
The sensitivity of the H. influenzae Rd pgdx mutant to var- ious pro-oxidants was examined by disk diffusion (Fig. 3A) and liquid culture assays (Fig. 3, B-D). Compared with the wildtype strain the mutant displayed enhanced sensitivity to t-BOOH and COOH, with inhibition zone differences ranging from 12 to 13%, respectively. When stressed in liquid culture with increasing concentrations of hydroperoxides, such as t-BOOH and COOH, we observed the expected, more rapid decline in growth rates in the case of the pgdx mutant (Fig. 3, C  and D). In both assays, resistance against the organic hydroperoxides was restored in a strain complemented in trans with a copy of the intact pgdx gene (Fig. 3, A-D; replication of the complementing plasmid and expression of PGdx in the complemented strain was confirmed; results not shown). However, above observations are not akin for H 2 O 2 . Here, the mutant shows more resilience than the wild-type or the complemented strain in both the disk diffusion assays and the liquid culture assays (Fig. 3B). In the case of methyl viologen, an indirect inducer of H 2 O 2 (11), we observed similar results (Fig. 3A). exponential phase followed by an attenuation when entering the stationary phase. However, when comparing absolute activities, the PGdx-deficient mutant exhibited around 2 orders of magnitude higher H 2 O 2 detoxification throughout growth compared with the wild-type strain. The complemented strain had its activity restored to wild-type levels (Fig. 4A). To determine whether at low micromolar concentrations mutant cells continue to turn over H 2 O 2 more efficiently compared with its wild-type counterpart, we analyzed scavenging activities at 1.2 M H 2 O 2 . These results reflect the ones when high concentrations were assayed (Fig. 4B).

Up-regulation of H 2 O 2 Scavenging Activity in pgdx Mu
The up-regulation of scavenging activity in the mutant most probably results from an increase in catalase activity. To assess this premise, cell extracts of exponential-phase cultures were resolved by native polyacrylamide gel. The gel was subsequently stained for catalase activity using a ferricyanide-negative stain. In the case of the mutant, we observed a higher in-gel activity derived from the catalase HktE (Fig. 5).
PGdx Is Essential for Stationary Survival-Much to our surprise, we found that when streaked onto a plate, aerobic cultures of the pgdx mutant exhibited no or barely detectable growth after 12 h of stationary phase while the anaerobically grown culture did (Fig. 6A and Table II). The wild-type parent and complemented mutant showed a moderate decrease in plating efficiency overnight in liquid culture. In contrast, the pgdx mutant rapidly lost viability within hours (Fig. 6B).
This aerobic growth defect in the mutant was restored in the complemented strain (Fig. 6, A and B) or with the addition of catalase during the mid-exponential phase and the early stationary phase (Fig. 6A and Table II). DISCUSSION The chimeric enzyme peroxiredoxin/glutaredoxin, or PGdx, from H. influenzae Rd constitutes a recently characterized glutathione-based removal system that reduces H 2 O 2 as well as alkylhydroperoxides at significant rates (3). Based on former work on the involvement of glutathione in oxidative stress, we already found that this low molecular weight thiol, when omitted from the growth medium, (i) has no effect on the growth rate of H. influenzae, (ii) increases the sensitivity of the bacterium for alkylhydroperoxides while decreasing it for H 2 O 2 , and (iii) induces a doubling in catalase activity (12). In addition we demonstrated that glutathione-based scavenging is crucial for metabolizing endogenously generated H 2 O 2 , since it complements for the absence of a functional catalase (6). Hence, from these data and from kinetic parameters (3), it was inferred that H. influenzae PGdx, in analogy with the NAD(P)H-dependent peroxidase Ahp from E. coli (13), fulfils a role as major peroxidase for low concentrations of H 2 O 2 . To assess the quantitative importance of this antioxidant enzyme against hydroperoxides, we constructed a H. influenzae Rd knock-out in the PGdx encoding gene HI0572 and studied its effect.
Characterization of the pgdx Mutant-No differences in growth rate between the pgdx mutant and its wild-type parent were observed, both under aerobic or anaerobic conditions. This confirms the apparent redundancy of GSH-based scavenging during logarithmic growth. However, as will be discussed below, the absence of PGdx is actually compensated for, thus concealing the loss in antioxidant capacity.
Our results also demonstrate that H. influenzae Rd, deficient in PGdx, is considerably hampered in the ability to reduce organic hydroperoxides such as t-BOOH and COOH. Polar effects can be ruled out here because sensitivities are restored when complementing in trans with an intact copy of the disrupted gene. However, comparison of the in vitro measured catalytic efficiencies for t-BOOH (k cat /K m 5.67 ϫ 10 4 s Ϫ1 ⅐M Ϫ1 ) and H 2 O 2 (k cat /K m 5.01 ϫ 10 6 s Ϫ1 ⅐M Ϫ1 ), in terms of substrate preference, puts alkylhydroperoxides second in line after H 2 (15), Pseudomonas aeruginosa (16), and Bacillus subtilis (17), where disruption of ahp causes an induction of catalase activity to compensate for its loss. Transformation with the functional gene complemented for the variation in resistance, as was also the case for Xanthomonas and Bacillus (14,17). We can conclude that this response of H. influenzae, inherent to the loss of the enzyme, underscores the importance of PGdx in H 2 O 2 scavenging.
Induction of hktE in the Absence of PGdx-At high H 2 O 2 concentrations, whole cells of the pgdx mutant scavenged nearly twice as much H 2 O 2 than its isogenic wild-type strain, which correlates well with the determined HktE activity in GSH-devoid medium (76 Ϯ 13 units/mg protein) compared with GSH-replete medium (36 Ϯ 7 units/mg of protein) (12). At low concentrations the mutant continued to scavenge at a higher rate, again reflecting our previous results when comparing a wild-type Rd strain in GSH-depleted medium with a catalasedeficient hktE mutant in GSH-replete medium (6). In concert with the results of the catalase activity gel staining we conclude that HktE is indeed the major factor governing the increased H 2 O 2 scavenging.
The absence of a functional PGdx is liable to induce the expression of the catalase HktE to compensate for the loss of antioxidant capacity. Activation of OxyR requires endogenous H 2 O 2 concentrations as low as ϳ10 Ϫ7 M (18). Since both PGdx and HktE (19) contain an OxyR-binding consensus sequence in their promoter, deficiency in glutathione-based scavenging by the cell could easily result in the upholding of sufficient H 2 O 2 levels, to low for efficient reduction by HktE but adequate enough for OxyR oxidation and subsequent catalase up-regulation. Moreover, we have shown that glutathione-based scavenging is more effective at scavenging very low concentrations of H 2 O 2 (5). These results not only support the foregoing explanation, but also classify H. influenzae PGdx as major antioxidant for low, metabolically generated H 2 O 2 .
Stationary-phase Dependence of H. influenzae on PGdx-Rather than producing a growth defect during log phase, the presence of a dysfunctional pgdx gene became apparent in the stationary phase. Only in the presence of exogenously added catalase, or when cultured anaerobically, did the PGdx-deficient mutant continue to grow after prolonged stationary growth.
Stationary-phase survival is often associated with virulence (20). In addition, a recent paper from Kraiss et al. (21) concludes that invasiveness of H. influenzae relies on the ability of the bacterium to grow aerobically. Several reports exist where stationary-phase induction of catalase plays a critical role (e.g. Caulobacter crescentus (22), Legionella pneumophila (23), B. fragilis (15)). From our observations, however, it seems that PGdx, rather than HktE, provides protection during stationary phase. This is likely because the latter becomes repressed at this point (19) and does not play a role in virulence (24). While in E. coli aerobic respiration was reported to decrease during stationary phase (18), other sources may continue to be an oxidative threat. With antioxidant power in H. influenzae being scarce due to the combined repression of catalase and the deficiency of PGdx, these threats become emphasized in the stationary phase, resulting in cells that succumb.
In short, we have demonstrated that, besides reducing organic hydroperoxides, the primary action of PGdx is to scavenge metabolically generated H 2 O 2 as well as to safeguard the cell from oxidative stress during the stationary phase. How exactly PGdx is involved in stationary-phase survival remains to be determined. Therefore, we are currently evaluating the virulence properties of a pathogenic variant of H. influenzae defective in pgdx, as well as the mechanisms behind the regulation of the gene.