Oxygen Reactivity of PutA from Helicobacter Species and Proline-Linked Oxidative Stress

Proline is converted to glutamate in two successive steps bythe proline utilization A (PutA) flavoenzyme in gram-negativebacteria. PutA contains a proline dehydrogenase domain thatcatalyzes the flavin adenine dinucleotide (FAD)-dependent oxidationof proline to ${Delta}$ ¹-pyrroline-5-carboxylate (P5C) and a P5C dehydrogenasedomain that catalyzes the NAD⁺-dependent oxidation of P5C toglutamate. Here, we characterize PutA from Helicobacter hepaticus(PutA_Hh) and Helicobacter pylori (PutA_Hp) to provide new insightsinto proline metabolism in these gastrointestinal pathogens.Both PutA_Hh and PutA_Hp lack DNA binding activity, in contrastto PutA from Escherichia coli (PutA_Ec), which both regulatesand catalyzes proline utilization. PutA_Hh and PutA_Hp displaycatalytic activities similar to that of PutA_Ec but have higheroxygen reactivity. PutA_Hh and PutA_Hp exhibit 100-fold-higherturnover numbers ( $~$ 30 min^–1) than PutA_Ec (<0. 3 min^–1)using oxygen as an electron acceptor during catalytic turnoverwith proline. Consistent with increased oxygen reactivity, PutA_Hhforms a reversible FAD-sulfite adduct. The significance of increasedoxygen reactivity in PutA_Hh and PutA_Hp was probed by oxidativestress studies in E. coli. Expression of PutA_Ec and PutA fromBradyrhizobium japonicum, which exhibit low oxygen reactivity,does not diminish stress survival rates of E. coli cell cultures.In contrast, PutA_Hp and PutA_Hh expression dramatically reducesE. coli cell survival and is correlated with relatively lowerproline levels and increased hydrogen peroxide formation. Thediscovery of reduced oxygen species formation by PutA suggeststhat proline catabolism may influence redox homeostasis in theecological niches of these Helicobacter species.

INTRODUCTION

Top

Introduction

All organisms convert proline to glutamate in two successivesteps catalyzed by proline dehydrogenase (PRODH) and ${Delta}$ ¹-pyrroline-5-carboxylate(P5C) dehydrogenase (P5CDH). In gram-negative bacteria, PRODHand P5CDH activities are combined within the proline utilizationA (PutA) flavoenzyme (1, 26, 33, 46, 53, 61). The first stepof proline oxidation is performed by the flavin adenine dinucleotide(FAD)-dependent PRODH domain, which couples the 2e^– oxidationof proline to the reduction of the electron chain transportsystem of the cytoplasmic membrane. The product of the PRODHreaction, P5C, is subsequently hydrolyzed to ${gamma}$ -glutamic semialdehyde,which is then oxidized to glutamate by the NAD-dependent P5CDHdomain of PutA. Glutamate that is formed by the oxidation ofproline eventually enters the tricarboxylic acid cycle via ${alpha}$ -ketoglutarate.Proline catabolism has been shown to be an important energypathway in various gram-negative bacteria such as Bradyrhizobiumjaponicum, Sinorhizobium meliloti, and Helicobacter pylori,with proline serving as a preferred respiratory substrate invarious environments (12, 25, 38, 58, 59).

Although the PRODH and P5CDH activities of PutA are conservedin gram-negative bacteria, the regulation of proline utilizationis divergent. In certain prokaryotes such as Escherichia coli,Pseudomonas putida, and Salmonella enterica serovar Typhimurium,PutA is an autogenous transcriptional repressor responsiblefor regulating the expression of the putA and putP (encodesa high-affinity Na⁺-proline transporter) genes (7, 34, 43, 60).The ability of PutA in these bacteria to switch from a DNA-boundtranscriptional repressor to a membrane-bound enzyme is dependenton proline availability and the redox state of the FAD cofactor.In the absence of proline, PutA binds to the intergenic DNAcontrol region of the putA and putP genes to repress transcription(7, 43). Upon proline reduction, PutA translocates from thecytosol to a position on the membrane which results in transcriptionalactivation of the put genes (37, 42). Thus, put gene expressionis regulated by the reduction of the FAD cofactor and subsequentPutA-membrane binding (42, 56, 64, 68). The DNA binding domainof PutA from E. coli has been identified as a ribbon-helix-helix(RHH) motif located at the N terminus (residues 1 to 47) (17).Primary sequence analysis predicts whether PutA proteins fromother gram-negative bacteria also contain the RHH domain. Prokaryotesin which PutA lacks transcriptional repressor activity, suchas Agrobacterium tumefaciens and Rhodobacter capsulatus, regulateproline utilization via transcriptional factors such as PutR,an Lrp-type activator protein which activates putA gene expressionin response to increased proline levels (11, 22, 24).

Besides contributing to cellular energy needs, proline metabolismhas been shown to impact intracellular redox homeostasis ina variety of organisms such as fungal pathogens, yeast, andmammalian cells (10, 13, 36). Upregulation of the mitochondrialPRODH enzyme in mammalian cells leads to high proline catabolicflux, increased reactive oxygen species (ROS) production, andeventual cell death via apoptosis (13, 32, 44). Proline alsoappears to function as an antioxidant, scavenging ROS to minimizeintracellular oxidative damage during various abiotic and bioticstresses (10). In yeast, disruption of the PUT1 gene (PRODH)increases oxidative stress tolerance (36). Thus, proline metabolismnot only contributes to the energy demands of the cell but alsocan affect the overall intracellular redox environment (18).

Although the metabolic function of PutA is well known, its rolein proline-mediated ROS formation has not been investigated.Here, we explore the impact of proline metabolism in the genusHelicobacter by characterizing PutA enzymes from the murineenterohepatic pathogen Helicobacter hepaticus and the humangastric pathogen H. pylori (52). This work is motivated by twoindependent studies that previously reported that proline wasa preferred energy substrate of H. pylori (38, 58). In fact,10-fold-higher levels of the amino acid proline were detectedin the gastric juice of patients infected with H. pylori relativeto noninfected individuals, suggesting that in the gut environment,proline serves as an energy source for opportunistic gastrointestinalpathogens (38). Because of the multifaceted roles of proline,proline not only may serve as an opportunistic source of energybut may also be involved in setting the local redox environmentof infection. ROS changes accompany events in gastrointestinaldiseases and seem to occur mainly via the host inflammatoryresponse requiring H. pylori to survive an onslaught of ROSduring infection (2, 19, 23, 63). However, H. pylori is alsoknown to increase ROS levels in host tissues, leading to epithelialcell damage and increased persistence in the gut (9). H. hepaticusinfection also causes oxidative stress in the liver (51). Consequently,proline obtained from host tissue damage may serve as an essentialenergy source and may support persistent colonization and growthof Helicobacter species (38). In certain environments, PutAcould have a critical role in colonization and development ofchronic disease that has not been previously recognized.

Primary structure analysis of PutA from H. hepaticus (PutA_Hh)and PutA from H. pylori (PutA_Hp) predicts that PutA from theseHelicobacter species lacks DNA binding activity and is strictlya bifunctional catabolic enzyme. Unexpectedly, PutA_Hh and PutA_Hpexhibit reactivity with molecular oxygen during catalytic turnoverwith proline that is much higher than that of PutA from E. coli(PutA_Ec), leading to the formation of reduced oxygen speciessuch as hydrogen peroxide (H₂O₂). PutA_Hh and PutA_Hp appear tohave different FAD environments, relative to PutA_Ec, that facilitateproline oxidase activity. The impact of Helicobacter PutA reactivitywith oxygen was assessed by oxidative stress studies in an E.coli model system that demonstrated that the enzyme action ofPutA_Hh and PutA_Hp is toxic to E. coli.

MATERIALS AND METHODS

Top

Materials and Methods

Chemicals, bacterial strains, and culture conditions. All chemicals and buffers, unless otherwise noted, were purchasedfrom Fisher Scientific and Sigma-Aldrich, Inc. DL-P5C was synthesizedas described previously and quantitated using o-aminobenzaldehyde(o-AB) (35, 54). Restriction endonucleases and T4 DNA ligaseswere purchased from Fermentas and Promega, respectively. PutA_Ecand PutA from Bradyrhizobium japonicum (PutA_Bj) were preparedas N-terminal His tag fusion proteins using the vectors PutA_Ec-pET14band PutA_Bj-pK8AH as previously described (26, 68). Terrificbroth was used for the expression and purification of PutA proteins,while LB and M9 minimal salts media were used for in vivo oxidativestress studies. E. coli strains BL21(DE3)(pLyS) and BL21(DE3)were used for the purification of the recombinant PutA proteinsand the oxidative stress studies, respectively. Promoter regionsof the putA gene from H. hepaticus and H. pylori were amplifiedfrom genomic DNA by PCR using primer sets 5'-GTTTTGATTTTAAGCTAAAATTTGTTTATTTG-3'and 5'-CACTTGCTCCTGCACTCAGGGCGGAAAC-3' and 5'-GCAGAGCTTACCTTTTATTTAAGAATTTGGCTT-3'and 5'-GTGAGTTGTTGCTACAAAATTAAAATTCAAGCG-3', respectively.

Cloning and purification of PutA_Hh and PutA_Hp enzymes. H. hepaticus strain ATCC 51449 genomic DNA was obtained fromGerald Duhamel, University of Nebraska, Lincoln, Nebr. The putAgene from H. hepaticus was PCR amplified using primers 5'-CAGGGCCATATGATGCAAGAAATTATCCAAGAA-3'and 5'-GTTAGCAGCCGGATCCTTACTTCAAAACTCTTG-3', which incorporatedNdeI and BamHI sites into the PCR product for subsequent cloninginto a pET14b vector. The putA gene from H. pylori was clonedsimilarly by using H. pylori strain J99 genomic DNA, which wasa generous gift from Joseph Barycki, Department of Biochemistry,University of Nebraska, Lincoln, Nebr. The putA gene was PCRamplified using primers 5'-CTTAAATAAAAGGTCATATGATGCAAAAAATCATT-3'and 5'-CTTATTTAATAGGATCCTTATTTTTCAGCACAGCA-3', which incorporatedrestriction NdeI and BamHI sites that were subsequently usedto clone the putA gene into pK8AH. The resulting PutA_Hh-pET14band PutA_Hp-pK8AH constructs were confirmed by nucleic acid sequencingof the entire putA genes.

PutA_Hh and PutA_Hp were overexpressed as N-terminal His₆ andHis₈ tag fusion proteins, respectively, in E. coli strain BL21(DE3)(pLysS)at 25°C as described previously for PutA_Bj, except that0.5% Triton X-100 was included during the disruption of theresuspended E. coli cells by sonication at 4°C (26). ThePutA proteins were purified using Ni²⁺-nitrilotriacetic acidand anion-exchange column chromatography as described previously(26, 68). The purity of the protein preparations was >90%as judged by sodium dodecyl sulfate-polyacrylamide gel electrophoresis,and the identity of the PutA_Hh and PutA_Hp proteins was confirmedby mass spectral analysis performed in the Metabolomics CoreFacility in the Redox Biology Center at the University of Nebraska-Lincoln.The oligomeric properties of PutA_Hh and PutA_Hp were characterizedby gel filtration chromatography (Superdex-200 column) as describedpreviously (17). PutA_Hh and PutA_Hp were stored in 50 mM potassiumphosphate (pH 7.5) containing 10% glycerol at –70°C.

Characterization of PutA proteins. PRODH activity was analyzed using the proline:dichlorophenolindophenol(DCPIP) oxidoreductase assay, and P5CDH activity was determinedby the reduction of NAD⁺ at 340 nm as described previously (6).Proline:DCPIP oxidoreductase assays were performed in 50 mMTris buffer, pH 7.5, with various amounts of proline (0 to 400mM). One unit of PRODH activity is the quantity of enzyme thattransfers electrons from 1 µmol of proline to DCPIP permin at 25°C. Proline:O₂ activity was determined in air-saturatedMOPS (morpholinepropanesulfonic acid) buffer (20 mM MOPS, pH7.5) at 25°C with various amounts of proline (0 to 400 mM).Rates of the proline:O₂ activity were determined by followingthe production of H₂O₂ or P5C with one unit of activity definedas the amount of PutA that generates 1 µmol of H₂O₂ orP5C per min at 25°C. H₂O₂ was quantitated by analyzing aliquotsfrom the oxidase reaction mixture by a catalase colorimetricassay kit (Sigma) using a standard curve of known H₂O₂ concentrations.The catalase assay was performed as recommended by the manufacturer.Superoxide anion detection during the PutA oxidase reactionswas performed as previously described by monitoring cytochromec reduction at 550 nm in the absence and presence of 30 U ofbovine superoxide dismutase (Sigma) (41). For detecting P5C,4 mM (final concentration) of o-AB was included in the proline:O₂assay mixture, and the formation of the o-AB-P5C yellow complexwas monitored at 443 nm ( ${epsilon}$ = 2,900 M^–1 cm^–1) (35).All enzyme assays were initiated by adding PutA enzyme and wereperformed in triplicate. The kinetic parameters K_m and k_catwere estimated by nonlinear regression analysis of the initialreaction velocity versus proline concentration using the Michaelis-Mentenequation. Turnover numbers were calculated using molecular massvalues of 135.5 kDa and 137.4 kDa for the monomeric speciesof PutA_Hh and PutA_Hp, respectively. The apparent inhibitionconstant (K_i) for the inhibition of PutA_Hh and PutA_Hp by L-tetrahydro-2-furoicacid (L-THFA) was determined as described previously for PutAfrom E. coli (69).

The concentrations of PutA proteins were determined using bicinchoninicacid reagents (Pierce) with bovine serum albumin as the standardand spectrophotometrically using molar extinction coefficientsat 450 nm of 13,373 ± 281 cm^–1 M^–1 and 12,944± 166 cm^–1 M^–1 for PutA_Hh and PutA_Hp, respectively(6, 62). The molar extinction coefficients for PutA_Hh and PutA_Hpat 450 nm were determined as previously described (6). Becauseit became apparent during this work that purified PutA_Hh wasrelatively more stable than PutA_Hp, further biochemical characterizationof PutA from Helicobacter species focused on PutA_Hh. Prolinetitrations of PutA_Hh were performed at 20°C in 50 mM potassiumphosphate buffer (pH 7.5) containing 5% glycerol in a glovebox (Belle Technology) under a nitrogen atmosphere (26). Theprotein-proline mixtures were equilibrated 10 min prior to therecording of each spectrum, and the data were analyzed as previouslydescribed, assuming the formation of a reduced PutA-P5C equilibriumcomplex (6). Titrations of PutA_Hh with sodium sulfite were performedanaerobically under a nitrogen atmosphere in a glove box usingsodium sulfite stock concentrations of 2, 20, 200, and 1,000mM. Spectra were recorded after equilibrating PutA_Hh with differentconcentrations of sodium sulfite. The enzyme solution was allowedto equilibrate after each addition of sodium sulfite until nofurther changes in the absorbance were observed, $~$ 15 to 20 min.To test the reversibility of the FAD-sulfite adduct, L-THFAwas added incrementally to the sulfite-PutA_Hh mixture up toa final concentration of 4 mM L-THFA. After each addition ofL-THFA, the PutA solution was equilibrated for 30 min beforethe spectrum was recorded.

Oxidative stress studies. Oxidative stress studies were performed in E. coli strain BL21(DE3)in M9 minimal salts medium according to a previously describedprotocol (16, 67). Cells harboring the putA genes from E. coli,B. japonicum, H. hepaticus, and H. pylori were grown in 20-mlcultures in LB broth at 37°C, and expression of the PutAproteins was induced for 3 h with 0.5 mM isopropyl-ß-D-thiogalactopyranoside(IPTG) when the culture reached an optical density of $~$ 0.8. Expressionof the PutA proteins was confirmed by PRODH assays of cell extractsfrom each culture with average values of about 0.33 (PutA_Hh),0.43 (PutA_Hp), 1.2 (PutA_Bj), and 1.2 (PutA_Ec) U/mg of totalprotein, indicating an almost fourfold variation in PRODH activitylevels. After treatment with IPTG, cell cultures (20 ml) werespun down at 9,000 x g, and the resulting cell pellets wereresuspended in 2 ml of M9 minimal salts medium and incubatedwith various reagents at 37°C for 1 h. H₂O₂, proline, andthe competitive inhibitor L-THFA were added to the 2-ml culturesat a final concentration of 5 mM. After the 1-h incubation,cell cultures were serially diluted and plated onto LB agarplates containing ampicillin (50 µg/ml) and grown overnightat 37°C. The next day ( $~$ 16 h), colonies were counted andmultiplied by the dilution factor to estimate the number ofsurviving colonies. Results are presented as the percent survivalrate, which is calculated as follows: [(number of colonies afteraddition of reagent)/(number of colonies with no addition ofreagent)] x 100. The reported percent survival rates are theaverage values from at least six independent experiments.

Proline and H₂O₂ levels were determined in 2-ml cultures supplementedwith no proline, 5 mM proline, or a mixture of 5 mM each ofproline and L-THFA as described above. For proline measurements,the 2.0-ml cultures in the M9 minimal salt medium were pelletedby centrifugation, resuspended in 0.5 mM of sterile water, andlysed by boiling for 10 min. The lysed samples were then centrifugedfor 5 min at 9,000 x g. The resulting supernatant was incubatedwith 46 mM ninhydrin and 0.2 ml glacial acetic acid for 1 hat 100°C (5). The reactions were stopped by placing thesamples on ice. After the mixtures were extracted with toluene(0.4 ml), the absorbance at 520 nm was recorded (5). A prolinestandard curve ranging from 0 to 3.0 mM proline was used todetermine the proline level of each sample. The intracellularproline levels were then estimated by assuming a cytosolic volumeof 0.47 µl per ml of culture (optical density, $~$ 1.0) (21).H₂O₂ in the medium of the cell cultures was monitored by analyzingaliquots at different time points from 0 to 60 min during theincubation as described above. Cells were pelleted by centrifugation,and H₂O₂ in the supernatant was measured using an Amplex Redhydrogen peroxide/peroxidase assay kit (Molecular Probes). ExtracellularH₂O₂ generated in a control medium of E. coli cells devoid ofrecombinant PutA was subtracted from the H₂O₂ measurements ofE. coli cells containing different PutA enzymes to correct fornonspecific H₂O₂ formation. The estimates of proline and H₂O₂are average values of at least three independent determinations.

RESULTS

Top

Results

Functional domain analysis and properties of PutA_Hh and PutA_Hp. The PutA enzymes from H. hepaticus (GenBank accession numberNC_004917) and H. pylori (GenBank accession number NP_222770)are polypeptides of 1,166 and 1,185 amino acids, respectively,share 63% sequence identity, and have a low sequence identityof 18 to 19% with PutA_Ec (1,320 amino acids) (3, 55). Domainalignment of Helicobacter PutA enzymes with PutA_Ec indicatesthat PutA_Hh and PutA_Hp share the PRODH and P5CDH domains butlack the RHH domain of PutA_Ec. The absence of the RHH domainwas confirmed by testing the DNA binding activities of PutA_Hhand PutA_Hp by gel mobility shift assays using the correspondingpromoter regions of the putA genes from H. hepaticus and H.pylori (6). PutA_Hh and PutA_Hp displayed no DNA binding activityeven with up to a 100-fold excess of PutA protein compared toDNA (data not shown as negative results). X-ray crystal structures(2.0 to 2.1 Å) of the PutA_Ec PRODH domain complexed withcompetitive inhibitors L-lactate and L-THFA have been solved,allowing active-site residues within 5 Å of the noncovalentlybound FAD cofactor and the inhibitors to be identified (66).Among the active-site residues in the PRODH domain of PutA_Ec,52% are identical in PutA_Hh and PutA_Hp, indicating a similarPRODH domain fold. Size exclusion chromatography shows thatPutA_Hh and PutA_Hp purify as dimer-monomer mixtures with thepredominate species being a dimer. PutA_Ec purifies mainly asdimeric species in which the dimer is stabilized by intermolecularinteractions in the RHH domain (17).

The proline:DCPIP oxidoreductase activities of PutA_Hh and PutA_Hpare summarized in Table 1. Although slightly higher K_m valuesfor proline were observed, similar PRODH turnover numbers relativeto PutA_Ec were determined for PutA_Hh and PutA_Hp. The PRODH activesites in PutA_Hh and PutA_Hp were explored by evaluating the inhibitionof PRODH activity by the competitive inhibitor L-THFA. L-THFAcompetitively inhibits PutA_Ec, with a K_i of 0.2 mM (69). PutA_Hhand PutA_Hp were also shown to be competitively inhibited byL-THFA, with apparent K_i values of 0.33 and 0.35 mM, respectively,supporting similar substrate binding (data not shown). SpecificP5CDH activities in PutA_Hh and PutA_Hp were 0.25 and 0.19 U/mg,respectively, which are near the P5CDH specific activity level(0.3 U/mg) for PutA_Ec (6).

View this table:
[in this window]
[in a new window]
TABLE 1. Kinetic parameters for PutA_Ec, PutA_Hh, and PutA_Hp enzymes^a

The UV-visible absorbance spectrum of oxidized PutA_Hh is shownin Fig. 1 (curve 1), with the main absorbance peaks at 450 and367. A molar extinction coefficient of 13,373 ± 281 cm^–1M^–1 at 450 nm was determined for FAD bound to PutA_Hh withabout 1 mol of FAD bound per polypeptide, similar to PutA_Ec(6). The same ratio of FAD to polypeptide was found for PutA_Hp,with a molar extinction coefficient of 12,944 ± 166 cm^–1M^–1 at 450 nm. Figure 1 also shows a proline titrationof PutA_Hh demonstrating that proline can fully reduce the boundFAD cofactor with an equilibrium constant for the reductionof the enzyme of 14.3 ± 0.43 mM^–1 proline. Theobservation that proline (reduction potential [E_m], $~$ –0.124V for the proline/P5C couple [pH 7.5]) can fully reduce theFAD cofactor indicates that the reduction potential of the FADcofactor in PutA_Hh is more similar to the reduction potentialof the FAD cofactor in PutA_Ec (E_m, –0.077 V [pH 7.5])than to that in PutA_Bj (E_m, –0.132 V [pH 7. 5]), whichis only partially reduced by the substrate proline (6, 26, 62,69). Proline-reduced PutA_Hh is fully reoxidized upon mixingwith an equal volume of air-saturated buffer with a half-lifeof $~$ 12 min. The oxygen reactivity of PutA_Hh is about 10-foldfaster than that of proline-reduced PutA_Ec, which reacts veryslowly with air (half-life, $~$ 120 min) (6). The sluggish oxygenreactivity of proline-reduced PutA_Ec protects the substrate-derivedreducing equivalents until the flavoenzyme reacts with the correctphysiological partner, which is presumably ubiquinone, in theelectron transport chain.

View larger version (23K):
[in this window]
[in a new window]
FIG. 1. Proline reduction of PutA_Hh. PutA_Hh was titrated with proline under anaerobic conditions in 50 mM potassium phosphate (pH 7.5) at 25°C. Spectra were recorded 10 min after the addition of proline. Curves 1 to 9, 0, 0.02, 0.04, 0.05, 0.07, 0.37, 0.97, 1.97, and 2.97 mM proline, respectively.

Proline:O₂ activity. To assess reactivity with molecular oxygen during catalyticturnover with proline, the rates of proline oxidation by PutA_Hh,PutA_Hp, and PutA_Ec were measured by monitoring H₂O₂ formationin air-saturated buffer. The kinetic parameters are summarizedin Table 1 and show that PutA_Hh and PutA_Hp display significantcatalytic turnover consistent with the high oxygen reactivityof the proline-reduced enzymes relative to PutA_Ec. The proline:O₂specific activities for PutA_Hh and PutA_Hp correspond to apparentturnover numbers of 29 and 31 min^–1, respectively, whichare >100-fold higher than the estimated turnover number forPutA_Ec (<0.3 min^–1). The K_m values for proline determinedfor PutA_Hh and PutA_Hp are the same in both the proline:DCPIPoxidoreductase and the proline:O₂ assays. Evaluation of thekinetic data in Table 1 also shows that PutA_Ec has a significantpreference for the artificial electron acceptor during catalyticturnover with proline based on the >2,500-fold activity inthe proline:DCPIP oxidoreductase assays. In contrast, PutA_Hhand PutA_Hp show a more modest preference for the artificialelectron acceptor, with about 20-fold-higher activity in theproline:DCPIP oxidoreductase assays than in the proline:O₂ assays.The diminished preference for the artificial electron acceptoris due not to lower proline:DCPIP oxidoreductase activity butrather to a noticeable increase in oxygen reactivity. Similardifferences between PutA_Ec and PutA_Hh in catalytic turnoverwith proline using oxygen as the electron acceptor were alsoobserved in o-AB chromogenic assays, which monitor the formationof P5C. Turnover numbers of 17 min^–1 and <0.3 min^–1were determined for PutA_Hh and PutA_Ec, respectively. To testwhether the higher proline:O₂ activity may be generally associatedwith bifunctional PutA enzymes, we also evaluated the bifunctionalPutA_Bj enzyme, which shares 37% amino acid identity with PutA_Ec(26). PutA_Bj displayed low proline:O₂ activity (1 min^–1),similar to PutA_Ec, indicating that the increased reactivityof PutA_Hh and PutA_Hp with molecular oxygen during catalyticturnover with proline is not necessarily a general propertyof bifunctional PutA enzymes. Formation of a superoxide anionwas detected during PutA_Hh and PutA_Hp oxidase reactions by cytochromec reduction, indicating that superoxide is generated duringcatalytic turnover with proline (data not shown). Superoxidewas not detected in assays with PutA_Ec, consistent with muchlower oxygen reactivity.

To further characterize the increased proline:O₂ activity ofHelicobacter PutA enzymes, we compared the reactivity of theFAD cofactor with sodium sulfite in PutA_Hh with that in PutA_Ec.Flavoenzymes that react with molecular oxygen during catalyticturnover typically form reversible sulfite adducts at the N-5position of the isoalloxazine ring of FAD, which is observedby bleaching of the flavin absorbance at 450 nm (31). Figure2 (top) shows that the addition of sodium sulfite to PutA_Hhdecreases the absorbance of the FAD cofactor at 450 nm. Fromthe change in the absorbance at 450 nm, a dissociation constantof 400 ± 60 µM was estimated for the formationof a FAD-sulfite adduct in PutA_Hh. In contrast, PutA_Ec doesnot react with sulfite even at up to 50 mM concentrations ofsodium sulfite (data not shown). The FAD-sulfite adduct wasstable in air but was disrupted by the addition of L-THFA tothe PutA_Hh-sulfite mixture. Figure 2 (bottom) shows that L-THFArestores the oxidized spectrum of PutA_Hh, indicating that sulfiteis displaced from the PRODH active site. It is evident in Fig.2 that L-THFA is bound to the PRODH active site, as a prominentshoulder in the FAD absorbance spectrum at 474 nm is observed(69). Thus, PutA_Hh shows biochemical characteristics that areconsistent with higher oxygen reactivity during turnover withsubstrate.

View larger version (23K):
[in this window]
[in a new window]
FIG. 2. Reaction of PutA_Hh with sodium sulfite. (Top) PutA_Hh was incubated with different concentrations of sodium sulfite in 50 mM potassium phosphate buffer (pH 7.5) at 25°C for 15 min. Curves 1 to 6, 0, 0.05, 0.1, 0.5, 1, and 4 mM sodium sulfite, respectively. Inset is a plot of the decrease in absorbance ( ${Delta}$ Abs) at 451 nm versus sulfite concentration. The data are fit to a single-binding-site isotherm with a dissociation constant value of 0.4 ± 0.06 mM sulfite. (Bottom) Spectral changes that occur upon the addition of L-THFA to the sulfite-PutA_Hh mixture. Curves 1 to 3, 0, 1, and 5 mM L-THFA, respectively.

Stress tolerance in Escherichia coli. To test whether the proline oxidase activities of PutA_Hh andPutA_Hp have any impact on redox status in bacteria, we comparedthe stress tolerance levels of E. coli cultures expressing PutA_Ec,PutA_Bj, PutA_Hh, and PutA_Hp. Figures 3 and 4 show the survivalof E. coli cells after treatment with 5 mM H₂O₂. Expressionof PutA_Ec and PutA_Bj, which display minimal oxidase activity,has little effect on the tolerance of E. coli cells to oxidativestress relative to control E. coli cells that contain only avector (i.e., pET14b). In stark contrast, acute oxidative stresstreatment is basically lethal to E. coli cells harboring PutA_Hhand PutA_Hp, with survival rates of <0.05%. To explore thetoxicity of PutA_Hh and PutA_Hp further, E. coli cultures wereincubated with 5 mM proline. As described above, PutA_Ec andPutA_Bj had little effect relative to the control E. coli cells.However, in the presence of proline, expression of PutA_Hh andPutA_Hp is toxic to E. coli, with survival rates of around 0.5%.It appears that the enzyme action of PutA_Hh and PutA_Hp killsthe E. coli cells, perhaps due to the formation of reduced oxygenspecies. The PRODH activity of PutA_Hh and PutA_Hp was then blockedby supplementing E. coli cultures with the competitive inhibitorL-THFA (5 mM) in the presence of proline (5 mM). Figures 3 and4 show that L-THFA rescues E. coli cells expressing PutA_Hh andPutA_Hp and increases survival rates to 60 to 70%. Thus, theenzyme action of PutA_Hh and PutA_Hp is the causative agent ofproline toxicity in E. coli.

View larger version (50K):
[in this window]
[in a new window]
FIG. 3. Acute stress treatment of E. coli cell cultures expressing the various PutA enzymes. E. coli cells were incubated with 5 mM H₂O₂, 5 mM proline, or a mixture of 5 mM each of proline and L-THFA in M9 minimal medium for 1 h at 37°C. After the incubation, cells were serially diluted and plated onto LB agar plates containing ampicillin and grown at 37°C. Images were recorded after 16 h of growth. The control cells contain the pET14b vector alone.

View larger version (42K):
[in this window]
[in a new window]
FIG. 4. Cell survival rates of E. coli cells expressing different PutA enzymes. E. coli cells were incubated with 5 mM H₂O₂ (black), 5 mM proline (gray), or a mixture of 5 mM proline and 5 mM L-THFA (diagonal stripes) in M9 minimal medium for 1 h at 37°C. After incubation, cells were serially diluted and plated onto LB agar plates and grown at 37°C for 16 h. Colonies were then counted to estimate the percent survival rates, which are the average values from at least six independent experiments. Percent survival was calculated as follows: [(number of colonies after addition of reagent)/(number of colonies with no addition of reagent)] x 100. EcPutA, PutA_Ec; BjPutA, PutA_Bj; HhPutA, PutA_Hh; HpPutA, PutA_Hp.

To gain further insights into how the expression of the HelicobacterPutA enzymes affects cell survival, proline and H₂O₂ levelswere measured after E. coli cells were incubated in M9 minimalsalts medium for 1 h at 37°C with no supplement, with 5mM proline, or with a mixture of 5 mM proline and 5 mM L-THFA.Figure 5 shows that without proline supplementation, intracellularproline levels are about twofold lower in E. coli cells (2 to2.3 mM) expressing the PutA enzymes than in control cells (5.2± 0.3 mM proline). With proline supplementation in themedium, proline levels were also lower in E. coli cells expressingPutA_Ec and PutA_Bj ( $~$ 17 mM) and cells expressing PutA_Hh and PutA_Hp( $~$ 9 mM) than in control cells (33.2 ± 1 mM). These observationsdemonstrate that the catabolic activity of recombinant PutAenzymes depletes intracellular proline pools, with larger decreasesobserved with the Helicobacter PutA enzymes. The estimated intracellularlevels of proline are within the range of previous measurementsof proline pools in E. coli from studies that evaluated prolineuptake and accumulation in response to osmotic stress (4, 39,65). As expected, the addition of 5 mM L-THFA prevents the breakdownof intracellular proline, as levels in E. coli cells harboringPutA enzymes (27 to 36 mM proline) are in the range of prolinecontent measured for the control cultures (34.5 ± 0.9mM).

View larger version (37K):
[in this window]
[in a new window]
FIG. 5. Proline levels in E. coli cell suspensions expressing different PutA enzymes. E. coli cells were incubated and supplemented with no proline (black), 5 mM proline (gray), or a mixture of 5 mM proline and 5 mM L-THFA (diagonal lines) in M9 minimal medium for 1 h at 37°C. Proline levels were determined from lysed cell supernatants and are reported as estimates of intracellular concentrations. The values are an average of three independent experiments. EcPutA, PutA_Ec; BjPutA, PutA_Bj; HhPutA, PutA_Hh; HpPutA, PutA_Hp.

Next, H₂O₂ released into the medium from E. coli cells expressingthe different recombinant PutA enzymes was measured. Figure6A shows that the amount of H₂O₂ detected in the medium by 60min was higher in E. coli cells expressing Helicobacter PutA( $~$ 1.5 µM) than in cells expressing PutA_Bj (0.32 ±0.05 µM) and PutA_Ec (0.28 ± 0.04 µM). AlthoughPutA_Ec and PutA_Bj decrease the intracellular proline pool, similarto PutA_Hh and PutA_Hp (Fig. 5), less H₂O₂ is generated duringproline degradation by these enzymes. Figure 6B demonstratesthat proline supplementation increases H₂O₂ levels in the externalmedium of each culture, with estimates at 60 min of 7.3 ±0.6 µM and 7.2 ± 0.3 µM for PutA_Hh and PutA_Hp,respectively. In E. coli cells containing PutA_Bj and PutA_Ec,H₂O₂ concentrations of 1.6 ± 0.1 µM and 0.7 ±0.1 µM were detected, respectively. Thus, the additionof proline to E. coli cells expressing Helicobacter PutA elevatesextracellular H₂O₂ levels nearly 5-fold higher than cells containingPutA_Bj and $~$ 10-fold higher than cells with PutA_Ec. It is evidentfrom the lower proline levels found in cells harboring PutA_Hhand PutA_Hp (Fig. 5) that Helicobacter PutA enzymes create ahigher proline catabolic flux than PutA_Ec and PutA_Bj that isapparently due to increased oxygen reactivity. Figure 6C showsthat the addition of L-THFA restores H₂O₂ levels, demonstratingthat the higher amounts of extracellular H₂O₂ measured fromcells expressing PutA_Hh and PutA_Hp are derived from prolineoxidation.

View larger version (15K):
[in this window]
[in a new window]
FIG. 6. Time course of H₂O₂ released from E. coli cells harboring different PutA enzymes. Extracellular H₂O₂ was measured in E. coli cell cultures containing PutA_Ec ( ${blacksquare}$ ), PutA_Bj ( ${blacklozenge}$ ), PutA_Hh (•), and PutA_Hp ( ${blacktriangleup}$ ) in M9 minimal medium. Cultures were supplemented with no proline (A), 5 mM proline (B), or a mixture of 5 mM proline and 5 mM L-THFA (C). The values are an average of three independent experiments.

DISCUSSION

Top

Discussion

PutA enzymes can be classified according to whether they lack(i.e., bifunctional) or are endowed with (i.e., trifunctional)transcriptional regulator activity. Sequence alignments of PutA_Ecwith PutA enzymes from different sources have identified >20PutA proteins that contain the RHH domain and are thus predictedto function as autogenous transcriptional regulators. For bacteriain which PutA does not function as an autogenous transcriptionalregulator, proline utilization requires other regulatory proteinssuch as PutR and PruR. PutR is an Lrp-type transcriptional repressorthat has been identified in R. capsulatus, and PruR is a AraC/XylSfamily regulator that has been shown to activate expressionof the putAP operon in Pseudomonas aeruginosa (24, 40). In thefood-borne pathogen Vibrio vulnificus, neither PutR nor PruRappears to be involved; instead, the cyclic AMP receptor proteinhas been postulated to regulate the putAP genes (28). Genomeanalysis of H. hepaticus and H. pylori reveals that the putAand putP genes are arranged in an operon similar to those ofP. aeruginosa and V. vulnificus. Potential homologues of PutRand PruR in the genomes of H. hepaticus and H. pylori, however,are not apparent, and a cyclic AMP receptor protein bindingsite (TGTGAN_6-8TCACA) is not present in the promoter regionsof the putAP operons, suggesting further divergence in mechanismsfor the regulation of proline utilization in Helicobacter species.

Surprisingly, PutA_Hh and PutA_Hp have significantly higher reactivitywith molecular oxygen than PutA_Ec during catalytic turnoverwith proline. Accordingly, PutA_Hh forms a FAD-sulfite adduct,while PutA_Ec is unreactive, consistent with the notion thatflavoproteins that react with sulfite generally have higheroxidase activity. Flavin dehydrogenase enzymes typically reactslowly with oxygen during catalytic turnover with substrateto help ensure the efficient transfer of electrons to the correctphysiological acceptor for maximum energy utilization of thesubstrate (29). A notable example includes the mammalian acylcoenzyme A dehydrogenases, which profoundly depress oxygen reactivityin the product-bound form, preventing formation of reactiveoxygen species in the mitochondria (14). Similarly, the free-dihydroflavinform of PutA_Ec reacts rapidly with oxygen but shows significantlylower oxygen reactivity in the substrate/product-bound form(6).

The structural determinants of oxygen reactivity in flavoenzymesare not fully understood, but it seems that the electrostaticenvironment of the active site is a critical component (30).The first step in the reaction of reduced flavins with molecularoxygen is the formation of a radical pair involving superoxide(O₂^–) and a flavin radical (29). The formation of thesuperoxide anion is favored by overall positive electrostaticenvironments, whereas active sites with lower dielectric environmentsare thought to hinder superoxide formation (14, 15, 30). Structuralstudies of the conversion of xanthine dehydrogenase to xanthineoxidase indeed show differently charged active sites (15). Xanthinedehydrogenase is a metalloflavoenzyme, which, during catalyticturnover with xanthine, has a preference for NAD⁺ as an electronacceptor, while the xanthine oxidase form prefers molecularoxygen (20). It is clear from the structural analysis of bothforms of the enzyme that the electrostatic environment is morepositive in the xanthine oxidase form than in the xanthine dehydrogenaseform (15). Kinetic studies of glucose oxidase also suggest thatpositively charged residues in the active-site environment areimportant for accelerating oxygen reactivity of FAD cofactors(45). Differences between the electrostatic environments ofthe PRODH active sites of PutA_Ec and Helicobacter PutA are notobvious by primary structure analysis alone. PutA_Hh and PutA_Hpshare 52% of the active-site residues in PutA_Ec that have beenidentified by X-ray structure determination of the PutA_Ec PRODHdomain (66). The predicted variations in the active-site residuesof PutA_Hh and PutA_Hp from PutA_Ec do not involve changes in chargedresidues. The only relevant substitution may be that Tyr437in PutA_Ec, which is conserved in PutA_Bj as Tyr345, is replacedby Asn292 and Asn291 at the analogous positions in PutA_Hh andPutA_Hp, respectively (26, 66). From the structural analysisof the PRODH domain of PutA_Ec, it appears that Tyr437 separatesthe active site from bulk solvent (66). Thus, the replacementof Tyr with a smaller side-chain residue such as Asn in HelicobacterPutA may increase solvent accessibility in the active site,which would favor higher oxidase activity. In the mammalianmedium-chain acyl coenzyme A dehydrogenase, it has been proposedthat product binding desolvates the active site, resulting inlower oxygen reactivity relative to the free enzyme (14). Likewise,proline/P5C binding may more effectively desolvate the activesite in PutA_Ec than in PutA_Hh or PutA_Hp. A future goal is todetermine the three-dimensional structure of PutA from Helicobacterspecies to compare the flavin active-site environment and oxygenreactivity with those of PutA_Ec.

The proline-linked formation of reduced oxygen species by PutA_Hhand PutA_Hp was shown to be toxic to E. coli, severely inhibitingcell survival. Blocking PRODH activity rescued cell survival,demonstrating that the enzyme action of Helicobacter PutA isharmful under high-proline conditions. Increased proline catabolismin general, however, was not lethal, as PutA_Ec and PutA_Bj expressionin E. coli had no impact on cell survival despite the detectionof increases in H₂O₂ levels upon the addition of proline tothe medium. Therefore, the higher oxidase activity of HelicobacterPutA appears to be the culprit for inducing cell death. Alkylhydroperoxide reductase, the main scavenger of endogenous H₂O₂in E. coli, and catalase are responsible for keeping intracellularH₂O₂ concentrations below toxic levels ( $~$ 2 µM) (48-50).Apparently, the H₂O₂ generated by PutA_Bj and PutA_Ec does notreach toxic levels, while the higher proline oxidase activityof PutA_Hh and PutA_Hp is enough to overwhelm the scavenging systemand produce toxic effects under our experimental conditions.Perhaps more importantly, though, is the formation of superoxideduring the proline oxidase reaction, which likely has a considerablerole in the lethal effect of proline catabolism via HelicobacterPutA and in the acute sensitivity to oxidative stress treatment.

The physiological significance of PutA oxidase activity in Helicobacterspecies is presently not clear. One simple explanation is thatthe higher oxidase activity of PutA_Hh and PutA_Hp is an outcomeof the microaerophilic environment of Helicobacter species,which may not require robust protection from oxygen. On theother hand, PutA_Bj, which is also from a microaerophilic bacterium,resembles PutA_Ec, with low oxidase activity and no deleteriouseffect on E. coli cell survival. A potential role of the PutAoxidase activity in Helicobacter species may be that prolinemetabolism contributes to the redox environment of infection.It has been shown that proline is a preferred respiratory substratein cultured H. pylori cells and that patients infected withH. pylori have considerably higher levels of proline in thegut (38). Thus, proline not only may be used for energy butmay also affect ROS levels which impact colonization and/orthe persistence of Helicobacter species in the host. Prolinemetabolism has been shown to be critical in different ecologicalniches, with proline having multifaceted roles in protein chaperoning,abiotic stress protection, and energy utilization (8, 10, 18,27, 36, 38, 47, 52, 57). To address the pathophysiological roleof PutA and proline metabolism in Helicobacter species, futurework will include the analysis of PutA expression and prolinemetabolic flux in host infection animal models.

ACKNOWLEDGMENTS

This research was supported in part by NIH GM061068, the Universityof Nebraska Biochemistry Department and Redox Biology Center,the Layman Foundation, and the Nebraska Agricultural ResearchDivision, Journal Series no. 14693. This publication was alsomade possible by NIH grant number P20 RR-017675-02 from theNational Center for Research Resources.

The contents of this paper are solely the responsibility ofthe authors and do not necessarily represent the official viewsof the NIH.

FOOTNOTES

* Corresponding author. Mailing address: Department of Biochemistry, University of Nebraska, N258 Beadle Center, Lincoln, NE 68588. Phone: (402) 472-9652. Fax: (402) 472-7842. E-mail: dbecker3{at}unl.edu.

REFERENCES

Top