Characterization of an Acid-Dependent Arginine Decarboxylase Enzyme from Chlamydophila pneumoniae

Abstract

Genome sequences from members of the Chlamydiales encode diverged homologs of a pyruvoyl-dependent arginine decarboxylase enzyme that nonpathogenic euryarchaea use in polyamine biosynthesis. The Chlamydiales lack subsequent genes required for polyamine biosynthesis and probably obtain polyamines from their host cells. To identify the function of this protein, the CPn1032 homolog from the respiratory pathogen Chlamydophila pneumoniae was heterologously expressed and purified. This protein self-cleaved to form a reactive pyruvoyl group, and the subunits assembled into a thermostable (αβ)₃ complex. The mature enzyme specifically catalyzed the decarboxylation of l-arginine, with an unusually low pH optimum of 3.4. The CPn1032 gene complemented a mutation in the Escherichia coli adiA gene, which encodes a pyridoxal 5′-phosphate-dependent arginine decarboxylase, restoring arginine-dependent acid resistance. Acting together with a putative arginine-agmatine antiporter, the CPn1032 homologs may have evolved convergently to form an arginine-dependent acid resistance system. These genes are the first evidence that obligately intracellular chlamydiae may encounter acidic conditions. Alternatively, this system could reduce the host cell arginine concentration and produce inhibitors of nitric oxide synthase.

Chlamydiae are obligately intracellular parasites that co-opt their host cell's metabolic and vesicle-trafficking systems. Infectious chlamydial elementary bodies invade epithelial host cells, differentiate into reticulate bodies inside an inclusion vesicle, and replicate without triggering normal cellular defense mechanisms (2, 9, 44). The bacterium-host interactions that enable chlamydial pathogenesis are not fully understood, but complete genome sequences from representative chlamydial species now provide a scaffold for studying these recalcitrant bacteria.

Despite their different courses of infection, members of the Chlamydiales share a large set of conserved genes. A recent analysis of the genome sequence from the anciently diverged “Candidatus Protochlamydia amoebophila” strain UWE25 (an endosymbiont of Acanthamoeba sp.) found that 711 coding DNA sequences are shared among UWE25 and all of the chlamydiae (20). This conserved set represents approximately 75% of the gene content of the chlamydial genomes, constraining the number of horizontal gene transfer events since the lineages diverged. This result is consistent with the natural barriers to gene transfer imposed by the cells’ life cycles. Therefore, we were surprised to find chlamydial homologs of a gene from nonpathogenic archaea that encodes a pyruvoyl-dependent arginine decarboxylase (PvlArgDC). In Chlamydophila pneumoniae CWL029, this protein is expressed from the CPn1032 locus. All sequenced chlamydial genomes contain orthologs of CPn1032, but it is not found in UWE25.

There are two known functions for arginine decarboxylase in microorganisms. Escherichia coli expresses an arginine-dependent acid resistance system (AR3) that requires arginine decarboxylase (adiA) and arginine-agmatine antiporter (adiC) genes (15, 24, 26). These proteins catalyze arginine uptake, its decarboxylation, and the subsequent excretion of agmatine to raise the cytoplasmic pH and invert the membrane potential (13). Alternatively, E. coli and many bacteria express a biosynthetic arginine decarboxylase (speA) gene. These cells make polyamines by decarboxylating l-arginine to produce agmatine and then hydrolyzing agmatine to urea and putrescine, which is the diamine core for spermidine and spermine synthesis (Fig. 1) (6). Both the AdiA and SpeA arginine decarboxylases require a pyridoxal 5′-phosphate (PLP) coenzyme. The hydrothermal-vent-dwelling archaeon Methanocaldococcus jannaschii uses the same pathway for polyamine biosynthesis, except its PvlArgDC enzyme is unrelated to the bacterial protein and does not require a cofactor (16). Instead, that 17-kDa protein self-cleaves to form a 5-kDa β subunit (derived from the N terminus) and a 12-kDa α subunit (derived from the C terminus). The serine residue at the new N terminus of the α subunit forms a reactive pyruvoyl group that functions analogously to PLP (18).

FIG. 1.

The arginine decarboxylase enzyme (ArgDC; EC 4.1.1.19) catalyzes the production of agmatine and CO₂ from l-arginine. The two known types of arginine decarboxylase use either pyridoxal 5′-phosphate or pyruvoyl cofactors. The agmatine urea hydrolase metalloenzyme (AUH; EC 3.5.3.11) catalyzes the hydrolysis of agmatine to produce putrescine, the core diamine of most polyamines. No homolog of AUH is present in chlamydial genomes.

However, chlamydial cells have no obvious need for an arginine decarboxylase enzyme. They lack homologs of the agmatine ureohydrolase or propylamine transferase enzymes that are required for polyamine biosynthesis. Also, chlamydiae are not believed to encounter acidic environments during their infection cycle (12). Chlamydial inclusions do not fuse with endosomes or lysosomes; they lack lysosomal marker proteins (45) and maintain a neutral pH (17, 43). Therefore, it is possible that the CPn1032 homologs do not use arginine for pH homeostasis. A crystal structure model of M. jannaschii PvlArgDC shows it to be homologous to a previously studied pyruvoyl-dependent histidine decarboxylase (PvlHisDC) that produces histamine (49). This similarity is not apparent at the primary sequence level, where the two enzymes share only a few conserved active site residues. Because the substrate-binding sites of PvlArgDC and PvlHisDC differ only in backbone carbonyl hydrogen bonds to their respective substrates, direct biochemical characterization is required to determine the substrate specificities of the homologs. Each chlamydial genome sequence encodes a homolog of this pyruvoyl protein that shares only 27 to 31% amino acid sequence identity with the M. jannaschii PvlArgDC. Yet the chlamydial sequences share 90% amino acid identity with each other (Fig. 2).

FIG. 2.

Protein sequence alignment of the C. pneumoniae CPn1032 protein (NCBI RefSeq accession no. NP_225226) with homologs from Chlamydia trachomatis (CT373; Swiss-Prot accession no. O84378), Chlorobium tepidum (GenBank accession no. AAM71804), Porphyromonas gingivalis (RefSeq accession no. NP_905146), and Methanocaldococcus jannaschii (Swiss-Prot accession no. Q57764). Conserved amino acid residues are shown in white on a black background. An arrow indicates the site of protein self-cleavage and pyruvoyl group formation. Sequences were aligned using the T-COFFEE program (version 4.96) (32).

To identify the function of the C. pneumoniae CPn1032 homolog of PvlArgDC, we expressed this protein in E. coli and purified it. The heterologous protein self-cleaved at a unique threonine-serine peptide bond to form a thermostable arginine decarboxylase enzyme. This enzyme was most active at an unexpectedly low pH (3.4), suggesting that it is part of an acid resistance system. In support of that model, the chlamydial gene complemented an adiA mutation in E. coli, restoring arginine-dependent acid resistance. The identification of a highly acid-dependent arginine decarboxylase suggests that either chlamydiae encounter an unknown acidic environment during infection or the system has an alternative effect on host metabolism and nitric oxide synthase activity.

MATERIALS AND METHODS

Strains.

Chlamydophila pneumoniae strain Kajaani 6 chromosomal DNA was a generous gift from Benjamin Wizel (University of Texas Health Center at Tyler) (11). Escherichia coli MG1655 (CGSC 7740) was obtained from the E. coli Genetic Stock Center (Yale). E. coli DH5α (Invitrogen) was used as a general cloning host, and E. coli BL21(DE3) (Novagen) was used for protein expression. Bacteriophage P1vir was a gift from Ian Molineux (University of Texas at Austin).

Cloning and molecular biology.

The CPn1032 gene was amplified by PCR from C. pneumoniae genomic DNA. Forward (5′-GGTCATATGGCTTACGGAACTCG-3′) and reverse (5′-GATCGGATCCTTAATTTACCTTAGCTG-3′) oligonucleotide primers for PCR were designed using the published CPn1032 gene sequence from C. pneumoniae strain CWL029 (25). The PCR product was cloned into NdeI and BamHI sites in plasmid pET-19b (Novagen) to create the expression vector pDG122. This NdeI-BamHI fragment was subcloned into compatible sites in pET-11a (Novagen) to create pDG148 and into pGEM-3Zf(+) (Promega) to create pTG01. Forward (5′-CGAGAGCTCACCATGGCTTACGGAACTCG-3′) and reverse (5′-GCGAAGCTTTTAATTTACCTTAGCTG-3′) PCR primers were used to amplify the same gene for cloning between the NcoI and HindIII sites in pBAD/HisA (Invitrogen) to create pDG339. Recombinant DNA was sequenced at the Institute for Cellular and Molecular Biology Core Laboratories DNA sequencing facility (University of Texas at Austin). The resulting gene sequence was identical to that reported for C. pneumoniae strain CWL029 (GenBank accession no. AE017160.1).

The E. coli adiA gene (encoding the protein with Swiss-Prot accession no. P28629) was amplified by PCR from E. coli MG1655 DNA using forward (5′-CGAGAGCTCACCATGGAAGTATTAATTGTTG-3′) and reverse (5′-CGCGGTACCTTACGCTTTCACGCACAT-3′) primers. The amplified DNA was cloned between NcoI and KpnI sites of pBAD/HisA to create pDG352.

Site-directed mutagenesis.

Plasmid pTG01 was mutagenized using the QuikChange II site-directed mutagenesis kit (Stratagene). The forward primer 5′-AATTTTAATATCGTCCCTTATTCATCTGTACTTCCTAAAGAGCTC-3′ and reverse primer (5′-GAGCTCTTTAGGAAGTACAGATGAATAAGGGACGATATTAAAATT-3′) were designed to replace CPn1032 Thr⁵² with Ser. The mutation in the resulting plasmid (pTG02) was confirmed by sequencing. The NdeI-BamHI fragment from this plasmid was subcloned into pET-19b to produce the expression vector pTG04.

Deletion of E. coli adiA.

An adiA null mutation was generated by the gene disruption method of Datsenko and Wanner (8). Forward (5′-GGAAGATACTTGCCCGCAACGAAGATTCCTTCATAACCGTGTAGGCTGGAGCTGCTTC-3′) and reverse (5′-CGTAATGTTATTTAAACAATTACGCCTTCAGCGGAAATTCCGGGGATCCGTCGACC-3′) primers were used to amplify the kanamycin resistance cassette from pKD13 (8). Electrophoresis gel-purified DNA was transformed into E. coli BW25113(pKD46) by electroporation. Recombinant strains were selected by growth on LB agar containing kanamycin (25 μg ml⁻¹). Kan^r strains were screened by PCR using forward (5′-CGAAAAGGCCGGAAGATACT-3′) and reverse (5′-CGACGTAACACCAGCCAAC-3′) external primers. The ΔadiA2::kan allele from a recombinant strain (DEG0116) was transduced into MG1655 using bacteriophage P1vir by standard methods (28). The ΔadiA2::kan allele in one transductant (DEG0121) was confirmed by PCR. A different ΔadiA::kan allele was previously constructed using similar methods (10).

Protein expression and purification.

E. coli BL21(DE3) cells transformed with pDG122, pDG148, or pTG04 were grown in Luria broth containing ampicillin (100 μg ml⁻¹) at 37°C with shaking at 250 rpm. When cultures reached an optical density at 600 nm of 0.6 to 0.7, protein expression was induced by the addition of α-d-lactose (1%, wt/vol). After 2 hours of incubation with the inducer, cells were harvested by centrifugation and stored at −20°C. Heterologous protein expression in whole-cell lysate was analyzed by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) using the Schägger and von Jagow Tris-Tricine system with 12% total and 3.3% cross-linked acrylamide (40). Proteins separated by SDS-PAGE were stained with Coomassie blue R-250 dye or oxidative silver stain (Bio-Rad).

The polyhistidine-tagged protein (His₁₀-CPn1032) was purified and separated from native E. coli proteins by Ni²⁺ affinity chromatography. E. coli BL21(DE3)(pDG122) cells expressing the His₁₀-CPn1032 protein were suspended in 50 mM TES {N-[tris(hydroxymethyl)methyl]-2-aminoethanesulfonate}-NaOH (pH 7.0). These cells were lysed by passage through a French pressure mini cell at 8,000 lb/in² (Thermo Electron) and briefly sonicated on ice using a model 450 sonifier with a microtip (15 W, 30% duty; Branson) to reduce viscosity. Lysates were clarified by centrifugation (14,000 × g for 15 min). The cell extract was then separated by ultracentrifugation (100,000 × g for 30 min at 4°C) to remove insoluble proteins. The supernatant was applied to a 1-ml HisTrap HP column (GE Healthcare) equilibrated with binding buffer that contained 5 mM imidazole, 500 mM sodium chloride, and 20 mM sodium phosphate, pH 7.6. Chromatography was performed using an ÅKTAprime system (GE Healthcare) at a flow rate of 1 ml min⁻¹. Protein was eluted from the column with a linear gradient to 100% elution buffer over 20 min. Elution buffer contained 20 mM sodium phosphate (pH 7.6), 500 mM sodium chloride, and 1 M imidazole. Fractions containing the target protein were identified by absorbance at 280 nm and SDS-PAGE. Affinity-purified His₁₀-Chlamydia pneumoniae arginine decarboxylase protein was desalted using a 5-ml HisTrap Sepharose 25 column (GE Healthcare) equilibrated in 50 mM MES [4-morpholinoethanesulfonate]-NaOH (pH 6). The Thr⁵²Ser variant protein was expressed in an E. coli BL21(DE3)(pTG04) strain and purified as described above for wild-type His₁₀-CPn1032. Total protein concentrations were determined using a Bio-Rad protein assay with bovine serum albumin (fraction V) as a standard.

Anion-exchange chromatography was used to purify untagged CPn1032 from cell extracts of E. coli BL21(DE3)(pDG148). The protein solution was applied at a flow rate of 2 ml min⁻¹ to a MonoQ column (5 by 50 mm; GE Healthcare) that was equilibrated in 20 mM Tris-HCl, pH 8. Protein was eluted from the column by a linear gradient from 0 to 1 M NaCl with 20 mM Tris-HCl (pH 8) over 40 min. Protein in the eluate was detected by its absorbance at 280 nm. Fractions (2 ml) were collected and analyzed by SDS-PAGE.

Measurement of native protein size.

Native masses of purified proteins were measured by analytical size exclusion chromatography in 100 mM sodium phosphate buffer (pH 6.8) at a flow rate of 0.5 ml min⁻¹, using protein standards as described previously (19). Proteins were detected by their absorbance at 220 and 280 nm.

Analysis of protein cleavage.

Protein cleavage was measured by densitometry analysis of SDS-PAGE images using Molecular Imaging software (version 4.0; Eastman Kodak). For matrix-assisted laser desorption ionization-time of flight (MALDI-TOF) mass spectrometry (MS), the purified protein solutions were applied to a Jupiter C₄ column (4.6 by 50 mm, 5 μm; Phenomenex) with a Security Guard C₄ cartridge (4 by 3 mm; Phenomenex) that had been equilibrated in 0.1% (vol/vol) trifluoroacetic acid in water. Protein was eluted from the column with 0.1% trifluoroacetic acid in isopropanol. Fractions containing protein were concentrated under vacuum before application to a target with sinapinic acid matrix. Mass spectra were acquired using a linear MALDI-TOF Voyager instrument (PerSeptive Biosystems), and an external mass calibration profile was built using insulin, cytochrome c, and aldolase proteins. Electrospray ionization (ESI)-MS was performed on affinity-purified His₁₀-CPn1032 protein that was desalted into 50 mM ammonium bicarbonate (pH 7.5). This protein was analyzed using an LCQ instrument (Thermo Finnigan) with an electrospray ion trap at the Analytical Instrumentation Facility Core at the University of Texas at Austin.

To screen for conditions promoting protein self-cleavage, His₁₀-CPn1032 was incubated with 0.8 M potassium chloride or 15 mM l-arginine, spermidine, spermine, or l-lysine for 24 h at 37°C. Cleavage was assayed using SDS-PAGE to compare treated and untreated protein samples.

Identification of amino acid substrates and amine products.

Standard assays for amino acid decarboxylase activity contained 0.5 μg purified protein and 50 mM sodium citrate (pH 5.0) in 50-μl reaction mixtures. Mixtures were preincubated at 37°C for 10 min before reactions were initiated by the addition of 1 mM amino acid substrate. After incubation for 18 min, reactions were stopped by the addition of 50 μl methanol and chilled on ice for 20 min. Mixtures were centrifuged to precipitate enzyme (17,000 × g for 10 min). Control reaction mixtures contained amino acid substrates without enzyme. Primary amines in the supernatant were derivatized with naphthalene 2,3-dicarboxaldehyde and cyanide to produce fluorescent cyanobenz[f]isoindole (CBI) derivatives as described previously (19). The resulting CBI derivatives were applied to a reversed-phase C₁₈ column (4.6 by 250 mm, 5 μm; Axxium) with a Security Guard octyldecyl silane cartridge (4 by 3 mm; Phenomenex) that had been equilibrated in mobile phase containing 50% (vol/vol) acetonitrile, 10 mM ammonium phosphate, 10 mM triethylamine, and water adjusted to pH 2.5 with phosphoric acid. Isocratic elution with this mobile phase was used at a flow rate of 1 ml min⁻¹ at ambient temperature. Under these conditions, retention factors for the CBI derivatives were 0.29 (l-histidine and l-canavanine), 0.40 (d-arginine and l-lysine), 0.41 (l-arginine), 0.43 (N^α-acetyl-l-arginine), 0.46 (l-homoarginine), 0.47 (l-argininamide), 0.52 (l-citrulline), 0.55 (l-glutamine or N^G-methyl-l-arginine), 0.57 (agmatine), 0.70 (histamine), and 0.85 (l-aspartate), with a hold-up time of 2.3 min. Standard curves of CBI derivative fluorescence showed a much lower response for agmatine than for arginine or other α-amino acids (data not shown). Therefore, agmatine production was calculated from peak areas using the method of standard addition.

Arginine decarboxylase assay.

The decarboxylation of radiolabeled arginine was determined using a ¹⁴CO₂ capture assay described previously (16). Standard reaction mixtures (100 μl) contained 40 mM citric acid, 20 mM Na₂HPO₄ (pH 3.5), 10 mM l-arginine-HCl, 100 nCi l-[U-¹⁴C]arginine (305 mCi mmol⁻¹) (Amersham Biosciences), and enzyme in 1.5-ml polypropylene tubes with barium hydroxide-saturated filter disks placed in the top. After 18 min of incubation at 37°C, the reactions were terminated by the addition of 100 μl of 4 M HCl and heated at 70°C for 15 min. Radioactivity collected on the filter was quantified by liquid scintillation counting. One unit of activity catalyzes the decarboxylation of 1 μmol l-arginine per minute. The specific activities of the purified enzyme at various substrate concentrations were then fit to the hyperbolic Michealis-Menten-Henri equation by nonlinear regression (KaleidaGraph version 3.6) to estimate steady-state kinetic parameters.

The effects of pH on decarboxylase activity were determined in standard arginine decarboxylase activity assays at 37°C with mixtures containing 0.36 μg purified enzyme, 10 mM l-arginine, and 100 mM glycine-HCl buffer (pH 1.5 or 2.2) or citrate-phosphate buffer (pH 2.6 to 6.2) (27). Protein thermostability was tested by preincubating 0.25 μg protein at temperatures from 4°C to 100°C for 10 min. The protein was then brought back to room temperature, and arginine decarboxylase activity was determined as described above. The temperature dependence of enzyme activity was determined by preincubating enzyme in reaction buffer at temperatures from 4°C to 70°C for 10 min. Reactions were initiated by the addition of arginine substrate and incubated at the same temperature.

Potential inhibitors were preincubated with 0.2 μg enzyme at 37°C for 10 min before arginine was added to a concentration of 10 mM. Reactions were terminated after 18 min, and CBI derivatives were analyzed by high-pressure liquid chromatography. Inhibitors tested included the arginine analogs l-arginine O-methyl ester, N^α-acetyl l-arginine, l-argininamide, l-citrulline, d-arginine, l-homoarginine, N^G-nitro-l-arginine methyl ester, N^G-methyl l-arginine, l-ornithine, l-histidine, and l-lysine. These compounds were also tested as potential substrates in reactions without arginine. The carbonyl-reactive agents O-methyl hydroxylamine hydrochloride and O-nitrobenzylhydroxyl-amine hydrochloride were tested as mechanism-based inhibitors.

Arginine-dependent acid resistance assay.

A microbiological test for acid resistance was adapted from the method described by Castanie-Cornet et al. (5). E. coli strains were grown for 22 h at 37°C in 5 ml brain heart infusion broth (EMD) and 0.2% (wt/vol) l-arabinose in 18- by 150-mm glass tubes, shaken at 275 rpm in a water bath with a 1-in. stroke length. Cells (approximately 3 × 10⁷ in 20 μl) were added to 2 ml of acid shock medium that had been prewarmed at 37°C. Acid shock medium contained 73 mM KH₂PO₄, 17 mM Na₂HPO₄, 10 mM sodium citrate, and 0.8 mM MgSO₄, supplemented with 0.4% (wt/vol) glucose and 1.5 mM arginine, and was adjusted to pH 2.5 with HCl. Immediately after the addition of cells and at 1- and 2-h intervals, samples were removed for serial dilution. The dilution medium was acid shock medium without glucose or arginine (pH 7.0). Samples from relevant dilutions were spread onto LB agar petri dishes and incubated at 37°C for 12 to 16 h. Viable cells were counted from the number of CFU averaged from several dilutions. Survival efficiency is the percentage of viable cells detected after a fixed interval, relative to the number of viable cells detected immediately after the introduction of cells to acid shock medium.

RESULTS

Expression and purification of CPn1032 protein.

The chlamydial protein encoded by the CPn1032 locus was heterologously expressed in E. coli and fused to an N-terminal decahistidine tag. Although most of the protein was found in the insoluble portion of cell lysate (probably in inclusion bodies), a significant amount of protein remained soluble. The soluble protein was purified by nickel affinity chromatography and analyzed by SDS-PAGE. Three prominent bands corresponding to proteins with apparent molecular masses of 25, 16, and 9 kDa were identified (Fig. 3). These masses match the predicted masses of the proenzyme (π; 24,422 Da), large subunit (α subunit; 15,870 Da), and small subunit (β subunit; 8,543 Da). Gel densitometry indicated that 35% of the purified protein was π, 29% was the α subunit, and 28% was the β subunit. A faint band corresponding to a 21-kDa protein was also observed in protein preparations. In contrast, the insoluble protein fraction contained primarily the π form. Untagged CPn1032 protein also formed a mixture of soluble, partially cleaved protein and insoluble protein (data not shown).

FIG. 3.

The C. pneumoniae CPn1032 protein is expressed in E. coli as a mixture of uncleaved π and α and β subunits produced by self-cleavage. Lane M, protein markers with masses shown at the left; lane 1, affinity-purified His₁₀-CPn1032; lane 2, affinity-purified His₁₀-CPn1032-Thr⁵²Ser variant.

A distantly related histidine decarboxylase from Lactobacillus sp. strain 30a also failed to completely cleave in vivo. That enzyme could be autoactivated by incubation in 0.8 M potassium phosphate (7). However, no increase in protein cleavage was observed when CPn1032 protein was incubated with 0.8 M potassium chloride for 10 h or 15 mM spermidine, spermine, lysine, or arginine for 24 h at 37°C. Detergent-solubilized π from inclusion bodies also failed to cleave after extended incubation with 0.1 M triethanolamine, 10 mM magnesium chloride, 10 mM manganese chloride, 0.1 M methoxyamine, 0.1 M sodium acetate (pH 4.5), or 0.1 M l-glutamate (pH 8.8).

Protein structure and cleavage.

Analytical gel filtration analysis of purified His₁₀-CPn1032 protein identified a single protein peak corresponding to a complex with an apparent mass of 88 kDa and a Stokes radius of 37 Å. This size corresponds to a trimeric complex comprising (αβ)₃, π₃, or a combination of cleaved and uncleaved subunits. The untagged CPn1032 protein eluted with an apparent mass of 73 kDa and a Stokes radius of 35 Å. Because the polyhistidine tag did not affect protein solubility, cleavage, or oligomerization, subsequent experiments used the readily purified His₁₀-CPn1032 protein.

MALDI-TOF MS analysis of His₁₀-CPn1032 protein identified peaks at 8,511 m/z, 15,997 m/z, and 24,459 m/z that correspond to ions of the β, α, and π proteins, respectively. An additional peak at 16,960 m/z could not be assigned. An ESI-MS instrument with higher mass accuracy identified multiple-charge ions corresponding to proteins with masses of 15,875 Da (α-subunit ion) and 24,283 Da (π ion) but did not identify ions of the β subunit. Together, these results confirmed the site of protein cleavage and pyruvoyl group formation between Thr⁵² and Ser⁵³ in CPn1032.

Effect of Thr⁵²Ser substitution on protein cleavage.

In homologous decarboxylases, Thr⁵² is replaced by a serine residue (Fig. 2). The side chain hydroxyl is believed to play an important role in the amide bond cleavage reaction, and a Ser⁸¹Ala variant of the corresponding residue in a PvlHisDC was defective in self-cleavage (14). Therefore, we expressed a Thr⁵²Ser variant of CPn1032 to improve the protein's rate of self-cleavage. However, the variant protein was significantly impaired in cleavage (Fig. 3). Only 12% of this variant protein self-cleaved to form active α and β subunits. The unidentified protein with an apparent mass of 21 kDa was more prominent in this protein preparation than in wild-type CPn1032.

Amino acid decarboxylase activity.

To screen for CPn1032-catalyzed decarboxylation, purified protein was incubated with substrates in discontinuous assays. CBI derivatives of substrate and product amines were separated by reversed-phase high-pressure liquid chromatography and quantified by their fluorescence intensities. The amino acid substrates tested were l-arginine, d-arginine, l-aspartate, l-canavanine, l-citrulline, l-glutamine, l-histidine, l-homoarginine, l-lysine, N^γ-methyl-l-arginine, and l-ornithine. Only l-arginine and its analog l-canavanine were substrates for the enzyme, producing agmatine and γ-guanidinoxypropylamine. The CPn1032 protein required no inorganic or organic cofactors for decarboxylase activity. Additional potassium ions (150 mM potassium chloride) had no effect on activity. Therefore, the CPn1032 protein is a PvlArgDC.

A radioisotopic assay was used to measure arginine decarboxylase activity. Surprisingly, CPn1032 arginine decarboxylase activity was highest under strongly acidic conditions; the protein had maximal specific activity at pH 3.4 (Fig. 4). Activity increased rapidly above pH 1.5 but diminished under more-alkaline conditions, with very low activity detected at neutral pH. The Thr⁵²Ser variant protein had 70% less arginine decarboxylase activity at pH 3.4 than wild-type enzyme; however, the pH profiles of their relative activities are similar (Fig. 4). When its lower extent of cleavage is taken into account, the variant protein has a specific activity similar to that of the wild-type protein.

FIG. 4.

Purified CPn1032 protein catalyzes l-arginine decarboxylation optimally at pH 3.4 (open boxes). The Thr⁵²Ser variant protein has the same pH optimum for l-arginine decarboxylation (closed circles) but a lower specific activity due to the reduced cleavage shown in Fig. 3. Activity assays were performed as described in Materials and Methods, and data were fit to polynomial functions.

The M. jannaschii PvlArgDC is extremely thermostable. It retained 50% of its activity after 20 min of incubation at 121°C (16). To determine whether this is a general property of the PvlArgDC family, we tested the thermostability of the CPn1032 protein by incubating the protein at various temperatures for 10 min before assaying arginine decarboxylase activity at 37°C. This protein retained 48% of its activity after incubation at 50°C and 13% of its activity after incubation at 100°C. The chlamydial protein also had significant decarboxylase activity at high temperatures. Activity increased up to 47°C (8.1 U mg⁻¹); above that temperature, protein denaturation outweighed the increased rate constant (Fig. 5).

FIG. 5.

Temperature dependence of CPn1032 arginine decarboxylase activity. Purified protein has maximal activity at 47°C but retains significant activity up to 80°C.

To determine the steady-state kinetic parameters for arginine decarboxylation, CPn1032 protein was incubated with various concentrations of l-[U-¹⁴C]arginine, and initial rates of CO₂ release were measured. The reaction was first order with respect to l-arginine, and the rates fit the hyperbolic Michaelis-Menten-Henri equation. The K_m value for l-arginine was 5.0 ± 1.0 mM, and the V_max was 9.7 ± 0.80 U mg⁻¹. Using densitometry measurements of protein cleavage to estimate the number of enzyme active sites, the turnover rate (k_cat) was 6.9 s⁻¹.

Inhibitors of decarboxylase activity.

Among the arginine analogs tested as inhibitors of arginine decarboxylation, l-argininamide was the most efficient. A 2 mM l-argininamide concentration reduced decarboxylase activity by 67% in the presence of 5 mM l-arginine. As observed with other pyruvoyl-dependent enzymes, nucleophiles that react with carbonyl groups are excellent inhibitors. O-Methyl hydroxylamine and O-nitrobenzylhydroxylamine reduced enzyme activity by 69% and 75%, respectively.

Complementation of E. coli adiA.

The E. coli arginine-dependent acid resistance system comprises a PLP-dependent arginine decarboxylase (AdiA) and a transmembrane arginine-agmatine antiporter (AdiC) (15, 24, 26). Together, these enzymes catalyze l-arginine uptake, its decarboxylation, its protonation to form agmatine, and the export of agmatine from the cell. The net reaction increases the cytoplasmic pH and aids in the reversal of the transmembrane potential (36). The catalytic properties of the CPn1032 enzyme resemble those of the E. coli AdiA protein, and the adjacent CPn1031 gene encodes a hypothetical amino acid antiporter that may function analogously to the E. coli AdiC transporter. Therefore, we proposed that C. pneumoniae independently evolved an arginine-dependent acid resistance mechanism analogous to the E. coli system. To validate this model, we tested whether CPn1032 expression complements the adiA null mutation in E. coli DEG0121.

The CPn1032 gene was expressed in DEG0121 cells under the control of a P_BAD promoter from plasmid pDG339. As a control, wild-type E. coli adiA was expressed from the same promoter in plasmid pDG352, and the empty vector (pBAD/HisA) served as a negative control. Cells were grown to stationary phase in complex medium supplemented with 0.2% (wt/vol) l-arabinose. The adiA cells had no significant arginine decarboxylase activity; cell extracts from DEG0121 cells had 0.1 ± 0.1 nmol min⁻¹ mg⁻¹ arginine decarboxylase activity, and DEG0121 (pBAD/HisA) cells had 0.2 ± 0.3 nmol min⁻¹ mg⁻¹ activity. In contrast, DEG0121 cells expressing C. pneumoniae arginine decarboxylase had arginine decarboxylase activities similar to those of wild-type cells and DEG0121 cells expressing the native E. coli AdiA protein (Fig. 6A).

FIG. 6.

Expression of CPn1032 complements the adiA mutation in E. coli DEG0121. (A) Mean arginine decarboxylase specific activities of cell extracts from five E. coli strains: DEG0121 (ΔadiA2::kan) (adiA bars), DEG0121 carrying pBAD/HisA (as a vector control), DEG0121 carrying pDG339 (expressing CPn1032), DEG0121 carrying pDG352 (expressing adiA), and MG1655 wild-type cells. Standard deviations are indicated for each strain (n = 3). Decarboxylase activity and total protein assays were performed as described in Materials and Methods. (B) Arginine-dependent acid shock survival rates for the same E. coli strains after 1 h (dark-gray bars) or 2 h (light-gray bars) of incubation at pH 2.5. The mean survival rates and sample standard deviations are shown for each sample (n = 3). Assays were performed with or without 1.5 mM l-arginine (Arg), as indicated. Conditions for cell growth and acid shock are described in the text. Survival rates were less than 0.3% for the adiA mutant and adiA mutant pBAD/HisA strains, as well as in the assays without arginine.

Acid shock experiments tested the ability of CPn1032 protein to restore acid resistance in adiA E. coli cells. Cells from the same cultures used to measure arginine decarboxylase activity were diluted in acid shock medium (pH 2.5) containing 1.5 mM l-arginine. The adiA cells are susceptible to acid shock under these conditions, and arginine does not significantly increase survival (Fig. 6B). Almost none of the DEG0121(pBAD/HisA) cells survived 1 h of treatment in the presence of arginine, compared to 9% of the wild-type MG1655 cells (Fig. 6B). CPn1032 expression rescues cells; survival rates for these cells are significantly higher than wild-type survival rates after 1 h. As expected, plasmid-based expression of adiA also rescues the cells (14% survival). The CPn1032 complementation requires arginine, because survival rates in acid shock experiments without arginine are similar to those of the adiA cells alone. Survival rates after 2 hours of exposure to acid shock medium are slightly lower than survival rates from a 1-hour exposure. These survival rates correlate with the cells’ arginine decarboxylase activities. Therefore, the CPn1032 protein functions in vivo to increase the acid resistance of E. coli cells by decarboxylating l-arginine.

DISCUSSION

Despite their low sequence similarities, the C. pneumoniae and M. jannaschii arginine decarboxylase homologs share similar catalytic activities and remarkable thermostability. Most of the arginine analogs and carbonyl-reactive nucleophiles that inhibited the M. jannaschii enzyme also inhibited the chlamydial enzyme (16). Both enzymes have high K_m values for l-arginine and low turnover rates compared to PLP-dependent biosynthetic decarboxylases (Table 1). Arginine decarboxylation is effectively irreversible, so the enzyme's high K_m may help regulate the reaction to prevent arginine depletion. Higher arginine concentrations lead to higher rates of agmatine production, because the arginine decarboxylase enzymes are not saturated. Inside mouse macrophages, the arginine concentration ranges from 0.3 to 0.6 mM, depending on the cell's state (47); however, little is known about the arginine concentration inside the chlamydial inclusion vesicle or chlamydial cytoplasm.

TABLE 1.

Kinetic properties of arginine decarboxylases and histidine decarboxylases

Enzymea (reference)	Cofactor	pH optimum	Km (mM)	kcat (s−1)	kcat/Km (M−1 s−1)
C. pneumoniae ArgDC	Pyruvoyl	3.4	5.0	6.9	1.4 × 103
M. jannaschii ArgDC (16)	Pyruvoyl	6.0	7.1	2.7	3.8 × 102
Lactobacillus HisDC (7)	Pyruvoyl	4.8	0.4	47	1.2 × 105
E. coli inducible ArgDC (3)	PLP	5.2	0.65	700	1.1 × 106
E. coli biosynthetic ArgDC (50)	PLP	8.4	0.03	19	6.2 × 105
PBCV-1 ArgDC (46)	PLP	8.2	0.45	15	3.3 × 104
Morganella morganii HisDC (48)	PLP	6.5	1.2	51	4.3 × 104

^aArgDC, arginine decarboxylases; HisDC, histidine decarboxylases.

Because chlamydial cells lack the biosynthetic enzymes to produce arginine, they must import it from the host cell. An ABC transporter complex that is regulated in some species by an arginine-responsive transcriptional repressor probably performs this function (41). Surprisingly, cultures of C. pneumoniae and C. trachomatis grown in HEp-2 epithelial cells in the presence of 10 mM arginine had 90%- and 40%-lower progeny infectivity than cultures grown in cells in control medium (1). Therefore, biosynthetic arginine requirements are separate from the catabolic arginine decarboxylase activity described here.

The chlamydial protein is less efficient at self-cleavage than other pyruvoyl-dependent enzymes. Although the highly conserved serine residue adjacent to the site of cleavage is replaced with threonine in the chlamydial homologs, this substitution is not directly responsible for the reduced cleavage since a Thr⁵²Ser variant was even less proficient at cleavage. The pyruvoyl proteins’ cleavage mechanism is not fully understood; the reaction could be facilitated by translational factors.

The chlamydial enzyme has a significantly lower pH optimum than the previously studied decarboxylases (Table 1). Both the inducible E. coli arginine decarboxylase (AdiA) and the Lactobacillus histidine decarboxylase function in acid resistance systems, yet their pH optima are nearly 1.5 pH units higher than that of the CPn1032 protein. At low pH, the Lactobacillus PvlHisDC undergoes a conformational change that stabilizes an α-helix in the active site (42). This feature was not observed in the M. jannaschii arginine decarboxylase structure (49), although we cannot rule out the possibility of pH-dependent structural changes in the CPn1032 protein due to limited sequence similarity between the carboxy termini of M. jannaschii and chlamydial α subunits (Fig. 2).

Consistent with its low pH optimum, the CPn1032 protein can replace the E. coli AdiA protein, which is required by the AR3 arginine-dependent acid resistance system. In acid shock assays, the survival rates for adiA mutants that express CPn1032 were comparable to survival rates for wild-type cells. Both values were significantly higher than the almost undetectable survival rates for adiA mutants. We are investigating whether the CPn1032 protein functions coordinately with a putative arginine-agmatine antiporter (encoded by CPn1031) to form an acid resistance system in chlamydiae.

The E. coli and chlamydial arginine decarboxylation systems evolved convergently; CPn1032 is not homologous to adiA, and the E. coli adiC arginine-agmatine antiporter has no significant sequence similarity to the putative CPn1031 transporter. Yet the CPn1032 gene is highly conserved in all pathogenic chlamydial genomes, and the C. trachomatis ortholog of CPn1032 is expressed in the middle-to-late period of the infection cycle (31). This highly stable protein could be expressed during the differentiation of reticulate bodies into elementary bodies and remain active until infection. During early stages of infection, having a pyruvoyl-dependent form of arginine decarboxylase may be advantageous to the chlamydial cells.

In contrast to enteric bacteria such as E. coli or Shigella sp., chlamydial cells do not pass through an acidic environment during infection. Studies of chlamydial inclusions indicate that these vesicles do not fuse with lysosomes and do not acidify during infection (12). It is possible that the chlamydial arginine decarboxylation system counters vesicle acidification, raising the pH. However, this model does not explain the requirement for vesicular Na⁺,K⁺-ATPase activity to maintain a neutral pH (43) or the lack of lysosomal fusion. Therefore, chlamydiae may encounter an unknown acidic environment prior to endocytosis. Alternatively, this arginine metabolic system could have a function different from pH homeostasis, such as reducing polyamine levels in eukaryotic cells, which inhibits cellular replication and proliferation (38), or inhibiting nitric oxide production.

Arginine is the substrate for inducible nitric oxide synthase (iNOS, or NOS2), and a reduction in host cell arginine concentrations can reduce NO^. production (29). Agmatine inhibits iNOS activity in macrophages, probably after conversion to γ-guanidinobutyraldehyde by diamine oxidase (35, 39). Therefore, a chlamydial system that reduces arginine levels and produces iNOS inhibitors could be a virulence mechanism. NO^. produced chemically or by murine macrophages reduced C. pneumoniae and C. trachomatis infectivity (4, 23). Although iNOS is not required to resolve chlamydial infection in human or mouse models (22, 33), macrophage NO^. production protected mice against high-dose chlamydial lung infection (21) and reduced chronic genital infection and disease (34). However, the role of NO^. in the host response to chlamydiae remains to be determined, due to host-specific differences in the gamma interferon-mediated responses (30) and the complicating effects of NO^. on other host cell processes (37).

The expression level and localization of the arginine decarboxylase described here is not known; it could be exported by a type III secretion system. Yet the presence of a putative arginine/agmatine transporter suggests that the arginine decarboxylase remains inside the chlamydial cell. To distinguish among these roles for the arginine decarboxylase system will require future studies of the expression levels in cell culture and infection models.