Identification and Characterization of a Putative Arginine Kinase Homolog from Myxococcus xanthus Required for Fruiting Body Formation and Cell Differentiation

Abstract

Arginine kinases catalyze the reversible transfer of a high-energy phosphoryl group from ATP to l-arginine to form phosphoarginine, which is used as an energy buffer in insects, crustaceans, and some unicellular organisms. It plays an analogous role to that of phosphocreatine in vertebrates. Recently, putative arginine kinases were identified in several bacterial species, including the social Gram-negative soil bacterium Myxococcus xanthus. It is still unclear what role these proteins play in bacteria and whether they have evolved to acquire novel functions in the species in which they are found. In this study, we biochemically purified and characterized a putative M. xanthus arginine kinase, Ark, and demonstrated that it has retained the ability to catalyze the phosphorylation of arginine by using ATP. We also constructed a null mutation in the ark gene and demonstrated its role in both certain stress responses and development.

INTRODUCTION

In response to nutrient limitation, Myxococcus xanthus cells form a multicellular, three-dimensional, macroscopic structure called a fruiting body (8, 16). During this developmental program, each vegetative cell in the swarm will follow one of three distinct cell fates. Cells within the developing fruiting body can differentiate into metabolically quiescent and environmentally resistant myxospores, and cells excluded from the fruiting body can differentiate into peripheral rod cells. Alternatively, cells can also undergo a programmed cell death pathway, or autolysis, which is hypothesized to provide nutrients and energy to the other two viable differentiating cell types. Depending upon how vegetatively growing cells perceive or encounter nutrient limitation, they can alter the ratios of these three cell fates (myxospores, peripheral rod cells, and autolysis). One unanswered issue in M. xanthus development is the question of what controls the partitioning of these cell fates and whether energy availability can influence these fates.

Phosphagen kinases (PKs) catalyze the reversible transfer of a high-energy phosphoryl group from ATP to a guanidino compound to create a transient buffer of high-energy phosphate in the cell. Arginine kinases (AKs), found predominantly in insects, crustaceans, and some unicellular organisms, are a member of the group of enzymes collectively referred to as phosphagen kinases, which also include creatine kinase (found in vertebrates) and other guanidino compound kinases. The list of organisms containing phosphagen kinases was dramatically expanded when a homolog was identified in the bacterium Desulfotalea psychrophila and shown to phosphorylate arginine, as the first characterized example of a functional PK homolog in a bacterium (1). The same study identified genes for putative AK homologs in the genomes of the social soil bacterium M. xanthus and two other proteobacteria, Sulfurovum sp. NBC37-1 (epsilonproteobacteria) and Moritella sp. PE36 (gammaproteobacteria). However, the role of these proteins in the bacteria and whether they have evolved novel characteristics is still a mystery.

Based on the high degree of similarity of the substrate-binding domains in these bacterial AKs and other known AKs, we hypothesized that the predicted M. xanthus AK-encoding gene (referred to as ark) encodes a functional AK that could phosphorylate arginine. Because M. xanthus development is a highly coordinated process that has a large energy requirement, we further hypothesized that the putative AK would be necessary for development. Here we report the genetic and phenotypic characterization of an ark deletion mutant and show that a functional Ark protein is required for normal fruiting body formation and wild-type levels of sporulation. Previous work in other systems demonstrated that AKs expressed heterologously in systems that naturally lack AKs provide a protective effect against acid and other environmental stresses (25). This led us to also investigate the role of ark in response to nonstarvation stresses. Finally, we report that the M. xanthus Ark enzyme can phosphorylate arginine in vitro, and we have verified its ability to synthesize ATP by using ADP and phosphoarginine as substrates.

MATERIALS AND METHODS

Strains, growth, and developmental conditions.

M. xanthus strain DK1622 was used as the wild-type and parental strain for all subsequent strains. MS2252 is a derivative of DK1622 carrying an in-frame deletion of the ark gene (MXAN2252), with the first and last 6 codons remaining. All strains and plasmids are listed in Table 1.

Table 1.

Strains, plasmids, and primers used for this study

Strain, plasmid, or primer	Description or sequence	Reference
Strains
DK1622	Wild type and parental M. xanthus strain	17
MS2252	DK1622 Δark	This study
Plasmids
pBJ114	M. xanthus cloning vector containing galK and kanamycin resistance gene	15
pJB100	pBJ114 Δark	This study
Primers
MyxoKO A	GAATTCAACCACGGAAGCACCTTCAAC	This study
MyxoKO B	GGGCGCCACGTGCTTGTGGAGCAGCAT	This study
MyxoKO C	CACAAGCACGTGGCGCCCGGGAACTGA	This study
MyxoKO D	GGATCCCTCAACGGGCAAGGCGAAGGAC	This study
MyxoForMBP	TAGGAATTCATGCTGCTCCACAAGCACCT	This study
MyxoRevMBP	TCGTCTAGATCAGTTCCCGGGCGCCACC	This study

Escherichia coli strains were grown at 37°C in LB broth (1.0% tryptone, 0.5% yeast extract, and 0.5% NaCl) or on plates containing LB broth and 1.5% agar. LB broth and LB agar plates were supplemented with 40 μg/ml of kanamycin sulfate or 50 μg/ml of ampicillin as needed. E. coli strains used for protein expression were additionally supplemented with glucose (0.2%) and arginine (20 mM) and induced with isopropyl-β-d-1-thiogalactopyranoside (IPTG) at 0.6 to 1.0 mM.

M. xanthus strains were grown at 32°C in CTTYE broth (1.0% Casitone, 10 mM Tris-HCl [pH 7.6], 1 mM KH₂PO₄, 8 mM MgSO₄, 0.5% yeast extract) with vigorous aeration or on CTTYE plates containing CTTYE broth and 1.5% agar. CTTYE broth and CTTYE agar plates were supplemented with 40 μg/ml of kanamycin sulfate as needed. CTTYE soft agar contained CTTYE broth and 0.3% agar.

Development of M. xanthus strains on TPM (10 mM Tris-HCl [pH 7.6], 1 mM KH₂PO₄, 8 mM MgSO₄) or CF [0.015% Casitone, 0.1% sodium pyruvate, 0.2% sodium citrate, 0.02% (NH₄)SO₄, 10 mM Tris-HCl (pH 7.6), 1 mM KH₂PO₄, 8 mM MgSO₄] agar (to 1.5% [wt/vol]) was performed as previously described (12). Chemically induced sporulation was performed by adding 0.5 M glycerol (final concentration) to vegetatively growing M. xanthus cells in CTTYE broth at a cell density of 5 × 10⁸ cells/ml and incubating them at 32°C with vigorous aeration for 4 h (7).

Construction of in-frame deletion of ark (MXAN2252).

Construction of the ark in-frame deletion followed a previously described method (28). Briefly, a product containing an in-frame deletion of the ark gene with approximately 500 bp of flanking sequence on either end was produced in a two-step PCR. This product was cloned into the plasmid pBJ114 (15) and transformed into E. coli DH5α. The deletion plasmid was electroporated into DK1622, and strains carrying the plasmid were selected for kanamycin resistance; the resulting tandem duplication strain was confirmed by PCR and then plated on CTTYE plates containing 1% galactose. Galactose-resistant colonies were screened for the desired deletion by PCR.

Motility and sporulation assays.

Motility was assayed on 0.3% and 1.5% CTTYE agar as previously described, and the swarm diameter was measured after 3 days of growth (27). Viable spore assays were performed as previously described (33). Spore production was assayed after 3 and 5 days of development on TPM and CF agar. In both cases, cells and fruiting bodies were scraped, collected in 500 μl of deionized distilled H₂O in microcentrifuge tubes, sonicated using a Branson Sonifier 450 3 times for 1 min each at a 70% duty cycle and an output control of 7, and incubated at 50°C for 2 h. Serial 10-fold dilutions were added to molten CTTYE-1% agar, plated onto CTTYE-1.5% agar plates, and incubated at 32°C until colonies were visible for counting (at least 3 days).

For chemical-induced sporulation, aliquots of glycerol-containing culture were taken after 4 h of incubation, and cells were frozen at −20°C. Sonication was then performed and viable spore production assayed as described above.

pH stress, salt and osmotic stress, and oxidative stress assays. (i) pH stress.

Cells were grown to 80 to 100 Klett units, i.e., mid-exponential phase, in CTTYE medium at 32°C, pelleted, and suspended in CTTYE medium at either pH 7.6 or pH 5.0. Suspended cells were incubated at 32°C with vigorous aeration for 1 h and then pelleted and suspended in pH 7.6 CTTYE medium. Growth was recorded and growth rates calculated for the periods before and after the pH shift.

(ii) Salt and osmotic stress.

Cells were grown to mid-exponential phase in CTTYE medium, pelleted, and suspended in CTTYE medium containing 0.2 M NaCl, 0.2 M KCl, or 0.2 M sucrose. Growth was monitored and recorded until cultures reached stationary phase.

(iii) Oxidative stress.

Cells were grown to mid-exponential phase (80 to 100 Klett units) in CTTYE medium, pelleted, and suspended in TPM buffer. Cells were treated with H₂O₂ (0.05% final concentration) or H₂O (equal volume as a control) and incubated for 2 min at room temperature. After exposure, cells were washed and suspended in TPM buffer before plating of serial 10-fold dilutions (as described above for spores). Colonies were counted after 5 to 7 days. Note that several concentrations of H₂O₂ were initially tested, i.e., 0.5%, 0.25%, and 0.1% H₂O₂, all of which produced a >5-log reduction in viable cells. Therefore, for these studies, a starting concentration of 0.05% H₂O₂ was used, which produced a 2-log reduction in viable cells after a 2-min exposure.

Fusion protein construction.

Genomic DNA was isolated from M. xanthus strain DK1622 as described previously (24). The ark gene was cloned by PCR in frame with the maltose binding protein (MBP) in the vector pMAL-c2g, using the EcoRI and XbaI restriction enzyme sites. Clones were screened by diagnostic restriction enzyme digests, and positive clones were verified by DNA sequencing to contain no errors in the cloned ark gene. One positive clone was transformed into competent E. coli strain BL21(DE3) cells for expression.

Fusion protein expression and purification.

A single colony was selected after transformation into BL21(DE3), inoculated into 50 ml of LB containing ampicillin (50 μg/ml), and grown overnight. This culture was transferred to 1.0 liter LB containing 20 mM arginine and 10 mM glucose and allowed to incubate at 37°C with shaking until the optical density at 600 nm reached 0.8 to 1.2 (approximately 4 h). IPTG (1 mM final concentration) was added to induce protein overexpression, and the culture was incubated at 18°C for approximately 16 h. The culture was subjected to centrifugation at 7,500 × g for 20 min at 4°C, and the pellets were collected and stored at −80°C.

The protein was purified as described by the manufacturer's instructions for MBPTrap HP column chromatography (GE Healthcare Bioscience AB, Uppsala, Sweden). Briefly, cell pellets were thawed on ice and resuspended in 40 ml of binding buffer (20 mM sodium phosphate buffer, pH 7.4, 200 mM NaCl, and 1 mM dithiothreitol [DTT]). The suspension was lysed on ice by sonication (6 times, 1-min intervals) and centrifuged at 15,000 × g for 35 min at 4°C. The supernatant was applied to an MBPTrap HP column previously equilibrated in binding buffer and washed with 15 to 20 column volumes of binding buffer. While the absorbance of the flowthrough was monitored at 280 nm, elution buffer (20 mM sodium phosphate buffer, pH 7.4, 200 mM NaCl, 1 mM DTT, and 10 mM maltose) was applied to the column to elute the MBP fusion from the amylose resin, using an elution buffer gradient to 100% over 10 column volumes. Fractions containing the protein were pooled, concentrated using ultrafiltration by nitrogen compression over a selective membrane (YM-10; Amicon), and stored in binding buffer. Concentrations of the MBP-tagged enzyme were estimated by absorbance measurements at 280 nm, using an extinction coefficient (ε₂₈₀ = 93,280) calculated by the method of Pace et al. (22).

Protein samples were further concentrated as needed, using a 10,000-NMWL (nominal molecular weight limit) Millipore ultrafiltration membrane. SDS-PAGE was conducted to verify purification. Purified protein samples were stored on ice at 4°C and used for kinetics analyses within 12 h, unless otherwise noted.

Determination of Ark's phosphagen kinase activity. (i) Determination of phosphagen kinase activity by ³¹P-NMR spectroscopy.

Purified Ark fusion protein was assessed for phosphagen kinase activity by ³¹P-nuclear magnetic resonance (³¹P-NMR) spectroscopy (Bruker 400-MHz spectrometer) as described by Andrews et al. (1). The phosphorylated product formed by Ark upon incubation with MgATP and different guanidino substrates (i.e., arginine, creatine, taurocyamine, or glycocyamine) was identified through comparison to the chemical shift associated with the guanidine-bound phosphate on a standard of phosphocreatine (20 mM) and MgATP (20 mM) in the absence of enzyme. All reactions were carried out in bicine buffer (100 mM, pH 9.0). Experimental reaction mixtures contained MgATP (20 mM), magnesium acetate (1 mM), Ark protein (15 μM), and either l-arginine (100 mM), creatine (100 mM), taurocyamine (100 mM), or glycocyamine (50 mM). Deuterium oxide was added (30% of the final volume) as a locking solvent signal directly before acquiring the initial spectrum. To verify that any phosphoguanidino product formed during the reactions was due to the activity of the added phosphagen kinase, two additional control reaction mixtures were analyzed by ³¹P-NMR. The first contained MgATP and arginine in the absence of enzyme, and the second contained only enzyme in bicine buffer. ³¹P-NMR spectra were obtained after incubation at 25°C for 15 h.

(ii) HPLC biochemical analysis of arginine kinase activity.

Phospho-l-arginine (P-Arg) trisodium salt was purchased from the Toronto Research Center (North York, Ontario, Canada). l-Arginine was purchased from Sigma (St. Louis, MO). Acetonitrile (high-performance liquid chromatography [HPLC] grade) was obtained from Pharmco-AAPER (Chatham, Canada). Stock solutions of 10 mM P-Arg and 10 mM arginine were prepared in separate solutions containing 20 mM KH₂PO₄ at pH 7. A dilution series of 1,000 μM, 500 μM, 250 μM, and 100 μM stock solutions was prepared for each compound and stored at −80°C. A calibration curve was developed by injecting 10 μl of each stock solution in triplicate.

HPLC analysis was conducted as described by Viant et al. (32), using a Hewlett-Packard Agilent 1100 series chromatograph equipped with an injection valve and a UV-visible (UV-Vis) diode array detector set to 205 nm. To obtain separation of arginine and phosphoarginine, a reverse-phase Spherisorb NH₂ amino column (4.6-mm by 200-mm S5 column with 5-μm particle diameter; Milford, MA) was used at room temperature. A solution mixture of 20 mM potassium phosphate (pH 2.6) and acetonitrile (72:28) was used for the isocratic mobile phase. Solvents were pumped at a flow rate of 1 ml/min, with a total analysis time of 15 min.

(iii) Kinetic analysis using a linked assay.

Kinetic analyses of bacterial arginine kinase activity (pH 8.0 and 25°C) were measured in the “reverse” direction (ATP formation) by coupling the initial rate of MgATP formation (from P-Arg and MgADP) to the rate of NADP reduction (followed at 340 nm), using a standard assay containing hexokinase (HK) and glucose-6-phosphate dehydrogenase (G6PDH) (6). Each reaction mixture (1.0 ml) contained bicine buffer (200 mM, pH 8.0), magnesium acetate (5.0 mM), glucose (1 mM), NADP (1 mM), bovine serum albumin (BSA) (0.05 mg/ml), acetate (4.0 mM total, by sufficient addition of sodium acetate), ∼4 units/ml HK/G6PDH, Ark (9 μM), and various concentrations of MgADP and P-Arg. Kinetic assays using a 1-by-8 matrix with a saturating concentration of MgADP (5 mM) versus P-Arg (0.12 to 0.48 mM) were conducted to give 8 initial rates. Kinetic assays using 8.0 mM phosphoarginine and 2.5 mM, 5.0 mM, or 10.0 mM MgADP all resulted in the same maximal rate, confirming that assays run at 5.0 mM MgADP were saturating for the nucleotide substrate.

RESULTS

Identification and phylogenetic characterization of a putative arginine kinase in Myxococcus xanthus.

The ark gene was initially identified by a BLAST search using horseshoe crab (Limulus polyphemus) arginine kinase (GenBank accession no. P51541) queried against all bacterial genomes available. The predicted M. xanthus gene MXAN2252 was a strong match, with an E value of 3e−75. MXAN2252 is a predicted 1,026-bp gene encoding a protein of 341 amino acids. The M. xanthus ark gene either lies at the end of a dicistronic operon or is monocistronic. An AK homolog was also identified in a related Myxococcus species, Myxococcus fulvus. This particular gene was found to be 81% identical in predicted amino acid sequence with the M. xanthus sequence, and moreover, it was in a region that is syntenic between the two species, suggesting that these two homologs were both derived from a common ancestor.

In order to make a preliminary prediction about the substrate specificity of the Ark protein product, an alignment of amino acid sequences was created using phosphagen kinases with different substrate specificities. The MXAN2252 predicted protein product was most similar to the representative AKs used in that alignment (data not shown). A subset of that comparison containing M. xanthus, Desulfotalea, and horseshoe crab AKs and rabbit muscle creatine kinases (CK) is shown in Fig. 1 to illustrate the similarities. As shown in Fig. 1, the translated ark gene encodes key residues known to be important for ATP binding, substrate binding, and catalytic activity (5, 9) in the phosphagen kinase family. In addition, it has a predicted substrate-binding domain that is most similar to that of arginine kinases.

Fig 1.

Alignment of selected creatine and arginine kinase amino acid sequences with those of selected putative eubacterial phosphagen kinase homologs. Numbering is relative to that of the rabbit muscle CK (RmCK) sequence. The dashes in the sequences refer to spaces put in for optimal alignment to a broad range of PK sequences, not all of which are shown here. Residues shaded in gray are conserved differently in AK and CK and served as signature residues for identifying the probable substrate specificity of the bacterial PKs. For clarity, only creatine and arginine kinase sequences are shown here. Regions of functional interest are indicated by the lines as defined in the key. GenBank accession numbers are indicated in Fig. 2. RmCk, rabbit muscle CK; HcAK, horseshoe crab AK; DpAK, Desulfotalea psychrophila LSv54; MxAK, Myxococcus xanthus DK1622 AK.

To further validate this prediction and to learn more about M. xanthus ark and its relationship with other bacterial phosphagen kinase homologs, a phylogenetic analysis was conducted. Figure 2 shows a maximum likelihood tree that was generated using all known predicted bacterial AKs at this time compared with several different protozoan and animal phosphagen kinases. As noted previously, the distribution of the bacterial PK homologs supports the hypothesis that they arrived by horizontal gene transfer (1). The MXAN2252 protein product branches most closely with previously described bacterial AKs (from the deltaproteobacteria Desulfobacterium, Desulfotalea, and M. fulvus) that form a basal branch off a cluster containing bacterial, invertebrate, and protozoan AKs. The cluster contains well-characterized AKs from various invertebrates (shrimp AK1 and -2, sponge AK, and nematode AK) and does not contain any other PKs that utilize different substrates (e.g., taurocyamine kinase [TK] or creatine kinase [CK]).

Fig 2.

Phylogenetic relationships of selected phosphagen kinases determined by maximum likelihood analysis. A sampling of different phosphagen kinases was used to construct a phylogenetic tree highlighting the relationships between the identified bacterial PKs and those of various eukaryotic species, with an emphasis on protozoan species. Microbial species are shown in red, and protozoan species are shown in blue. Bacterial taxonomic information is shown by the corresponding species. Only clades with bootstrap values of >50% are shown. Sequences used (GenBank accession numbers) were as follows: Myxococcus_xanthusAK, M. xanthus AK (YP_630473); Myxococcus_fulvusAK, M. fulvus AK (YP_004666772); ParameciumAK, Paramecium tetraurelia AK (XP_001450164); DesulfotaleaAK, Desulfotalea psychrophila LSv54 AK (YP_064771); DesulfobacteriumAK, Desulfobacterium autotrophicum HRM2 AK (YP_002604207); AhrensiaAK, Ahrensia sp. R2A130 AK (ZP_07375585); ChoanoflagellateAK3, Monosiga brevicollis AK3 (XP_001749532); ChoanoflagellateAK1, Monosiga brevicollis AK1 (XP_001742833); TetrahymenaAK, Tetrahymena thermophila AK (XP_001021675); PhytoplanktonAK, Emiliana huxleyi AK (EH genome database accession no. gm1.400237; protein ID 196115); SipunculanWormHTK, Siphonosoma cumanense HTK (BAE16970); ShrimpAK1, Neocaridina denticulata AK1 (BAH56608); ShrimpAK3, Neocaridina denticulata AK3 (BAH56610); Sponge_AK, Suberites fuscus AK (ABH10974); Sponge_mtCK1, Suberites fuscus mtCK1 (ABH10976); OomyceteHTK, Phytophthora infestans HTK (XP_002901831); SeaUrchinAK, Stichopus japonicas AK (Q9XY07); SeaUrchinCK, Strongylocentrotus purpuratus CK (XP_794336); SeaCucumberAK, Apostichopus japonicas AK (Q9XY07); FeatherdusterWormAK1, Sabellastarte indica AK1 (BAE16968); FeatherdusterWormAK2, Sabellastarte indica AK2 (BAE16969); GiantTubeWormTK, Riftia pachyptila TK (BAE16973); GiantTubeWorm_mtTK, Riftia pachyptila mtTK (BAE16972); RedwormLK, Eisenia fetida LK (BAA22872); LugWormTK, Arenicola brasiliensis TK (BAE16473); LugWorm_mitoTK, Arenicola brasiliensis mitoTK (BAE16474); LanceletCK1, Branchiostoma belcheri CK1 (ABF50761); ZebrafishCK, Danio rerio CK (AAH53305); rabbitCK, Oryctolagus cuniculus CK (NP_001075708); SipunculanWormCK, Siphonosoma sp. ST01 CK (BAD34677); SipunculanWorm_mtCK, Siphonosoma cumanense mtCK (BAE16971); and NereididWorm_mtCK, Namalycastis sp. ST01 mtCK (BAD34675).

Ark can catalyze the reversible transfer of phosphate to arginine.

The bioinformatics and phylogenetic analyses supported the hypothesis that the M. xanthus ark gene encodes a potentially functional AK. To address this hypothesis, an MBP-Ark fusion protein was made by cloning the coding portion of the ark gene into the pMAL-c2g vector as described in Materials and Methods. This construct was used for expression, purification, and biochemical analysis as described below. It was unknown if the gene had evolved to acquire novel functions after incorporation into the M. xanthus genome and/or retained the ability to phosphorylate arginine. To verify that Ark could indeed function as an arginine kinase, we demonstrated by both ³¹P-NMR and HPLC analyses that it was capable of specifically synthesizing phosphoarginine, as predicted. Ark was tested in the presence of a variety of substrates for the ability to transfer phosphate from ATP to a guanidine-containing substrate, using ³¹P-NMR. A chemical shift characteristic of phosphoarginine (singlet at 15.62 ppm with respect to the beta phosphate of ATP [data not shown]) was observed in the ³¹P-NMR spectrum of a reaction mixture containing l-arginine, MgATP, and Ark. This result was confirmed by HPLC analysis, which revealed that a product with a comparable retention time to that of a phosphoarginine standard (7.41 versus 7.42 min) was produced. The specificity of the Ark reaction was demonstrated by the absence of a phosphoguanidinium peak in the ³¹P-NMR spectra of reaction mixtures containing Ark, MgATP, and either creatine, taurocyamine, or glycocyamine, as well as for the controls containing MgATP and arginine without enzyme or containing enzyme only. Taken together, these data suggest that the encoded enzyme has retained the specific capacity to catalyze the transfer of a phosphate from ATP to arginine, producing the phosphagen phosphoarginine.

We next measured the specific activity of the Ark enzyme by using a linked assay to measure the rate of transfer of a phosphate from phosphoarginine to ADP, creating arginine and ATP (Fig. 3). The activity was low (70 nmol min⁻¹ mg⁻¹; observed k_cat = 5.6 ± 0.3 min⁻¹; K_M = 2.1 ± 0.3 mM for phosphoarginine [mean ± standard error of the mean]) but was found to be consistent with a previously described bacterial AK activity (1). We speculate that the low activity was due to the rapid loss of activity that the Ark protein demonstrated during purification (at least a 75% loss after 24 h on ice). Thus, the activity noted here likely significantly underestimates the actual activity in the cell. AK instability relative to that of dimeric PK forms has been noted before (2), and AKs from other bacterial species that we have purified (e.g., that from Desulfobacterium autotrophicum [data not shown]) have also shown a relatively rapid loss of activity. We are currently investigating the cause of the increased instability of the bacterial forms. It is worth pointing out that these experiments verify only that the ark-encoded gene product has retained the ability to catalyze the phosphorylation of arginine and that the physiological role it plays in the M. xanthus cell has yet to be demonstrated.

Fig 3.

Steady-state kinetics analysis of the ability of Ark to catalyze the phosphoryl transfer reaction from phosphoarginine to MgADP. Initial rates were measured with a saturating concentration of MgADP (5 mM) and a protein concentration of 9 μM. Nonlinear regression analysis of the rate data with the Michaelis-Menten equation estimated a k_cat of 5.6 ± 0.3 min⁻¹ and a K_M of 2.1 ± 0.3 mM for phosphoarginine.

Nondevelopmental phenotypic characterization of the Δark mutant.

Phosphagen kinases have been shown to have protective effects against a variety of stressors in many organisms and could potentially affect vegetative growth of M. xanthus under stress conditions, including salt, osmotic, pH, and oxidative stress (3, 4, 25). To determine if M. xanthus ark plays a role in the response to any of these stressors, we constructed an in-frame deletion that deleted 329 codons from the middle of the gene, leaving the first and last 6 codons intact, as described in Materials and Methods. We assayed the survival ability of both the wild-type and Δark strains under stress conditions. First, we examined osmotic and salt stress by monitoring growth in the presence and absence of 0.2 M sucrose and various concentrations of NaCl and KCl. Both the wild-type and Δark strains were grown in CTTYE or CTTYE plus sucrose or salt and monitored for growth, and the results are summarized in Table 2. Under standard laboratory conditions, the Δark mutant grew at a similar rate and to a similar yield as the wild type in all media tested. When cells were grown in the presence of the ionic stressor 0.2 M NaCl or 0.2 M KCl, the generation time was significantly longer for the Δark mutant (P = 0.010 and P = 0.047, respectively), while in the presence of the osmotic stressor 0.2 M sucrose, there was no significant difference between generation times (P = 0.24).

Table 2.

Effects of various stressors on generation times of wild-type and Δark mutant strains^a

Strain	Generation time (h)b
	No stressors	0.2 M sucrose	0.2 M KCl	0.2 M NaCl
Wild type	5.1 ± 0.26	6.6 ± 1.38	6.5 ± 0.95	7.8 ± 0.57
Δark mutant	4.9 ± 1.2	7.5 ± 1.2	9.1 ± 1.6	11.6 ± 1.2

^aCells were grown in liquid CTTYE medium supplemented with each stressor as described in Materials and Methods.

^bGeneration times represent the means ± standard deviations for three or more trials.

In addition to osmotic and ionic stress, AKs have been shown to be protective against pH stress in E. coli engineered to express horseshoe crab (Limulus polyphemus) arginine kinase (4). To determine if ark serves a similar function in M. xanthus, cultures of the wild type and the Δark mutant were grown to mid-exponential phase, moved transiently (1 h) to low-pH CTTYE (pH 5.0), and then returned to normal CTTYE (pH 7.6) for outgrowth. Table 3 shows that after transient pH stress, wild-type cells resumed growth at their original rate within 2 h, while the Δark mutant grew significantly more slowly than its original rate. This effect appeared to persist for multiple generations.

Table 3.

Generation times after pH shift^a

Strain, pH	Generation time after pH shift (h)b
WT, 7.6	5.22 ± 0.47
WT, 5.0	5.28 ± 0.40
Δark mutant, 7.6	4.53 ± 0.14
Δark mutant, 5	7.36 ± 0.53

^aAfter growth in CTTYE at pH 7.6, the wild type and the Δark mutant were shifted to CTTYE at pH 5.0 or pH 7.6 for 1 h and then returned to CTTYE at pH 7.6 and allowed to grow through exponential phase as described in Materials and Methods.

^bGeneration times represent the means ± standard deviations for four or more independent trials.

Oxidative stress is another stressor with energetic implications. In mammalian systems, oxidative stress induced by ischemia-reperfusion experiments with muscle tissue affects cellular energy homeostasis as well as affecting calcium regulation and inducing apoptosis (26). Creatine kinase has been shown to be modified irreversibly by reactive oxygen species in a variety of experimental systems (11, 19, 20). Modified CK or reduced CK function leads to the upregulation of CK gene expression, presumably to compensate for reduced creatine kinase function. While the role of arginine kinase in the cellular response to oxidative stress is less well studied, previous work by Pereira and coworkers has shown that the arginine kinase in Trypanosoma cruzi may play a role in a cell's response to this stressor (21). Their work has shown that oxidative stress can induce expression of the endogenous AK, and moreover, cells that overexpressed a homologous AK were resistant to nutritional, pH, and oxidative stress conditions (21, 23). To investigate the potential role of ark in oxidative stress, we exposed both wild-type and Δark cells to a transient H₂O₂ stress and calculated survivability. We found that unlike E. coli, wild-type M. xanthus was quite sensitive to hydrogen peroxide, as exposure to 0.05% H₂O₂ resulted in a 2- to 3-log killing, on average. Using a variety of concentrations and times, we found no significant difference between wild-type M. xanthus cells and the Δark mutant (data not shown).

Finally, we examined whether the Δark mutant displayed any decrease in motility in order to distinguish between any potential developmental phenotypes and a simple motility defect, as many motility mutants are unable to undergo development due to their inability to move. To examine both the A and S motility systems, wild-type and Δark cells were assayed for movement on both hard (1.5%) and soft (0.3%) agar (27). No significant defect in motility on either 0.3% or 1.5% agar was observed (data not shown).

Developmental phenotypic characterization.

To determine whether or not ark plays a role in M. xanthus fruiting body development or sporulation, a series of developmental assays were performed. First, the wild type and the Δark mutant were spotted onto TPM agar, a standard rapid starvation medium, or CF agar, which provides a gradual starvation medium, or grown under submerged culture conditions using MC7 buffer, another abrupt starvation medium, and assayed for development. The results of these assays are shown in Fig. 4. Under both TPM agar and MC7 submerged culture conditions, the Δark mutant was unable to form the characteristic round, well-defined fruiting bodies of the wild-type strain. Diffuse aggregates were visible at 48 and 72 h, and the mutant did not progress past this point. Surprisingly, development on CF agar was similar to that of the wild type, as the Δark mutant developed larger aggregates with slightly more defined edges than those for fruiting on TPM agar. The Δark mutant also developed a web-like appearance of less-defined aggregation on CF agar.

Fig 4.

Development of the wild-type and Δark mutant strains on various media. TPM and CF agar development is shown after 72 h, while development in MC7 submerged culture is shown after 120 h. TPM and CF agar are shown at a magnification of ×63, and MC7-developed cells are shown at a magnification of ×20.

We next tested for sporulation by assaying viable spore production, and the results are summarized in Table 4. When cells were developed on TPM agar, the viable spore production of the Δark mutant was significantly lowered, to 8.5% of the wild-type level, and in MC7 submerged culture the spore production of the Δark mutant was virtually undetectable (<0.01% of wild-type level). To rule out the issue of cell death as a cause of low myxospore production and viability, we assayed for total cell viability (no heat or sonication treatment) over the 5-day period on TPM agar. Under these conditions, the losses of viability were similar between the Δark mutant and the wild type. In fact, the Δark mutant showed a slight increase in viability over that of the wild type after 3 days. This slight increase can be explained by the onset of aggregation and fruiting body formation in the wild-type population, leading to clumping and the formation of spores and prespores (data not shown). Interestingly, there was no significant difference in viable spore production between the wild type and the Δark mutant on CF agar, which causes a slow starvation. We also tested chemically induced sporulation, a process that bypasses the developmental process of fruiting body formation and rapidly (within 90 to 120 min) allows M. xanthus to form spore-like structures called glycerol spores (13). Though the mechanism of chemical-induced sporulation is unknown, we equate this to a “rapid” induction event because it causes a rapid transformation of cell structure and physiology. While glycerol spores share many characteristics with myxospores, their overall resistances to environmental stresses (such as heat) are not as strong as those of developmentally induced myxospores. In addition, their physiological makeup is different; for example, they lack one of the major spore coat proteins (13). Under these conditions, we found that the Δark mutant produced 8.2% of the wild-type level of viable spores. The data are consistent with the hypothesis that ark is required under abrupt or stringent starvation conditions and plays less of a role in slow nutrient depletion.

Table 4.

Sporulation efficiency of the Δark mutant

Strain	Condition	Viable spore production (% of wild-type level)a
Wild type	TPM	100.0 ± 9
Δark mutant	TPM	9.4 ± 7
Wild type	CF	100.0 ± 42
Δark mutant	CF	92.0 ± 45
Wild type	Glycerol	100.0 ± 10
Δark mutant	Glycerol	8.2 ± 37
Wild type	MC7	100.0 ± 46
Δark mutant	MC7	Undetected (<0.001)

^aSporulation efficiencies are given as percentages of the wild-type level under identical conditions and represent the means ± standard deviations for three or more independent trials.

DISCUSSION

The arginine kinase subfamily is a complex grouping that is based primarily upon the similarity of substrate preferences rather than actual phylogenetic clustering. This is evidenced in our tree by the broad distribution of AKs throughout the tree and the long branch lengths associated with some AK clades. In addition to being very diverse in primary structure, the family also encompasses a range of biochemical characteristics and quaternary structures (14, 30). For example, AKs exist that utilize d-arginine preferentially as a substrate (31), and there are monomeric as well as dimeric forms (29). When characterized biochemically, AKs are generally less active than other PKs and display a range of different kinetic characteristics that include a sometimes broad substrate preference (1, 14, 31).

In this study, we identified and characterized a protein encoded in the M. xanthus genome that bears strong sequence homology with arginine kinases. This is the second characterization of a bacterially encoded AK-like protein, and both proteins display low turnover rates for the synthesis of phosphoarginine (1). Given the lack of eukaryotic-like compartmentalization in bacteria and the need to not deplete free arginine pools in the cytoplasm, this lower activity may be an advantageous adaptation, but this has not been proven. In addition, given the role the protein plays in development, as documented here, the Ark protein may have evolved to perform a signaling role, and thus rapid catalysis may not be optimal for that role. In conclusion, the results of our analyses have established only that the encoded protein has retained the ability to minimally catalyze the synthesis of phosphoarginine; they do not definitively prove that it catalyzes that reaction in the cell as part of maintaining energy homeostasis, as seen in eukaryotes. To underscore this point, it is worth considering that the McsB protein found in several Gram-positive bacteria contains a region with strong homology to the catalytic domain of the phosphagen kinases (18). It lacks the substrate-binding domain and has been shown to function as a protein kinase that can phosphorylate both arginine (10) and tyrosine (18) residues in target proteins.

Our phylogenetic analysis of the bacterially encoded arginine kinase homologs supports the hypothesis that these genes arrived by horizontal gene transfer (Fig. 2). The species that do have arginine kinase homologs are not closely related, and several species that are closely allied to these species do not have an AK homolog. In the case of the ark gene found in M. xanthus, of the closely related Myxococcus species that have completed genome projects, only Myxococcus fulvus has an AK homolog. The related myxobacterial species Anaeromyxobacter dehalogenans, Anaeromyxobacter sp. Fw109-5, Anaeromyxobacter sp. K, Haliangium ochraceum DSM 14365, Plesiocystis pacifica SIR-1, Sorangium cellulosum So ce 56, and Stigmatella aurantiaca DW4/3-1 do not have recognizable AK homologs in their genomes. Interestingly, the M. fulvus AK homolog is 81% identical to and is encoded in a genomic region syntenic with the M. xanthus homolog, suggesting that the gene was acquired prior to the divergence of these two species. The role of the AK in any bacterial species is unknown, but it is tempting to speculate that the addition of an AK gene to a genome can confer a selective advantage immediately. This is supported by work conducted by Sauer and colleagues, who demonstrated that the introduction of an AK gene into both eukaryotic and prokaryotic genomes conferred an increased capacity to withstand environmental stress (3, 4, 25). This is supported by our work in M. xanthus, and thus it is plausible that the introduction of the AK conferred an enhanced ability to tolerate some forms of environmental stress, such as pH, salt, and osmotic stress.

In this study, we found that ark is required for normal development in response to an abrupt depletion of nutrients. The Δark mutant did not form wild-type fruiting bodies under any of the tested conditions; interestingly, the effect was strongest on TPM agar or in submerged culture, in which M. xanthus is subjected to a rapid starvation due to the complete absence of amino acids. CF agar contains low levels of amino acids and thus produces a slower starvation; the Δark mutant still failed to form wild-type fruiting bodies on CF agar, but it did produce larger and slightly better-defined aggregates than those on TPM agar. This suggests that the ark deletion effect can be mitigated to some degree by providing nutrients that slow down the developmental process.

Sporulation of the Δark mutant was negatively affected under most conditions. In TPM-, submerged culture-, and glycerol-induced sporulation, the Δark mutant produced about 10-fold fewer viable spores than the wild type. Growth in 0.5 M glycerol is known to induce sporulation decoupled from fruiting body formation, unlike amino acid starvation. The fact that the Δark mutant was still unable to produce wild-type levels of viable spores indicates that the developmental defect affects not only fruiting body formation but also spore formation. Alternatively, when slowly starved on CF medium, the Δark mutant did not produce measurably different levels of viable spores from those of the wild type.

Taken together, these results suggest that the Δark mutant has more difficulty developing in more stringent starvation modes. This could be due to an inability of cells to store enough reserve energy to undergo development in response to rapid nutrient depletion; alternatively, it is possible that phosphoarginine acts as a signaling molecule and that the observed developmental defects are a result of misregulation of a novel and still undetermined developmental signaling pathway.

Because arginine kinase has been implicated generally in stress tolerance in other organisms, we investigated growth in a variety of ionically, osmotically, and pH-stressed environments. In general exponential growth, the Δark mutant responded more strongly than the wild type to both ionic stressors, with much lower growth rates in either case. Both the wild type and the Δark mutant grew slightly more slowly in CTTYE with sucrose, but their growth rates did not differ significantly from each other. These results suggest that ark has a protective role in tolerance of salt stress but is unlikely to be involved in osmotic stress.

The ark gene also appears to have protective effects against acid stress in M. xanthus, specifically in recovery from low-pH stress. Both the wild type and the Δark mutant tolerated but did not grow in CTTYE at pH 5.0, but only the wild type regained its normal growth rate after being transferred back into the normal CTTYE (pH 7.6) medium. It is possible that a general energy-buffering effect is responsible for the improved response of the wild type to transient pH stress, but it is also possible that a direct H⁺-buffering effect contributes, as the catalyzed reaction also uses H⁺ ions in converting free arginine to phosphoarginine.

Although it is clear that ark plays a role in M. xanthus development and in responses to several stresses, the precise mechanism remains unknown. One possibility is that ark creates a pool of high-energy phosphoarginine that is metabolically inert but able to be converted rapidly back into a useable form for the energy-intensive processes of development and response to stress. A role in direct pH buffering inside the cell cannot be excluded; expression of arginine kinase in other systems (3) has been shown to have such effects and could also play a role in M. xanthus. Finally, it is also possible that phosphoarginine acts as a messenger molecule, affecting gene expression or the activity of another protein in the cell, though this would be a function not previously reported for this molecule in the literature. Further studies will be required to elucidate the roles of both phosphoarginine and Ark in M. xanthus development and differentiation.