Contribution of the Cyclic Nucleotide Phosphodiesterases PdeA and PdeB to Adaptation of Myxococcus xanthus Cells to Osmotic or High-Temperature Stress

Abstract

A tBLASTn search of the Myxococcus xanthus genome database at The Institute for Genomic Research (TIGR) identified three genes (pdeA, pdeB, and pdeC) that encode proteins homologous to 3′,5′-cyclic nucleotide phosphodiesterase. pdeA, pdeB, and pdeC mutants, constructed by replacing a part of the gene with the kanamycin or tetracycline resistance gene, showed normal growth, development, and germination under nonstress conditions. However, the spores of mutants, especially the pdeA and pdeB mutants, placed under osmotic stress germinated earlier than the wild-type spores. The phenotype was the opposite of that of the receptor-type adenylyl cyclase (cyaA or cyaB) mutant. Also, pdeA and pdeB mutants were found to have impaired growth under the condition of high-temperature stress. Intracellular cyclic AMP (cAMP) levels of pdeA or pdeB mutant cells under these stressful conditions were about 1.3-fold to 2.0-fold higher than those of wild-type cells. These results suggest that PdeA and PdeB may be involved in osmotic adaptation during spore germination and temperature adaptation during vegetative growth through the regulation of cAMP levels.

Myxococcus xanthus, a gram-negative gliding bacterium, feeds on other microorganisms and decaying organic matter. In response to nutritional stress, M. xanthus forms a multicellular aggregate, which evolves into a fruiting body (7, 8, 21). About 10⁵ to 10⁶ cells migrate to an aggregation focus, where they construct a fruiting body. During the formation of the aggregate and fruiting body, the cell shape changes from a rod shape (0.5 to 1 μm by 5 to 10 μm) to a sphere (about 2 μm in diameter).

Osmotic stress is a common challenge encountered by eukaryotic and prokaryotic cells. We demonstrated that two receptor-type adenylyl cyclases (CyaA and CyaB) of M. xanthus act as osmosensors. CyaA and CyaB function as osmosensors mainly during development and spore germination and throughout the life cycle of M. xanthus, including the vegetative stage, respectively (9, 10). We also report that an M. xanthus CbpB containing two cyclic AMP (cAMP)-binding domains is involved in osmotic and temperature tolerance (11). The cbpB mutant was more sensitive to osmotic stress than wild-type cells were, and when cultured under high- or low-temperature conditions, the cbpB mutant exhibited a marked reduction in growth compared to wild-type cells. These results suggest that the cAMP-mediated signal transduction pathway plays an important role in osmotic and temperature adaptation in M. xanthus.

In the social amoeba Dictyostelium discoideum, cAMP acts as a morphoregulatory signal, which controls chemotaxis, gene expression, and cell differentiation during development (17, 20). The cellular concentration of cAMP in M. xanthus increases rapidly during starvation- and glycerol-induced development (5, 22). Campos and Zusman also observed that the formation of fruiting bodies was stimulated by the addition of cAMP to agar containing low levels of nutrients (1). They suggested that cAMP may trigger transcription at differentiation-specific promoter sites, in a manner similar to the regulation of β-galactosidase. On the other hand, Manoil and Kaiser demonstrated that cAMP induces fruiting body formation indirectly by causing intracellular nutritional imbalances (15).

The intracellular level of cAMP is regulated by its synthesis through adenylyl cyclases, degradation, and efflux. The degradation of cAMP is achieved by 3′,5′-cyclic nucleotide phosphodiesterases, which catalyze cleavage of 3′,5′-cAMP to 5′-cAMP. However, little is known about the biophysical function of cyclic nucleotide phosphodiesterases in bacteria, with a few exceptions (2, 6, 14). A cAMP phosphodiesterase gene (cpdA)-disrupted Escherichia coli or Salmonella enterica serovar Typhimurium strain showed about 2- or 1.3-fold higher intracellular cAMP than the wild-type strain, respectively (2, 6). In Haemophilus influenzae, a cyclic nucleotide phosphodiesterase mutation increased cAMP-dependent sugar fermentation and competence development (14). In this study, we constructed three 3′,5′-cyclic nucleotide phosphodiesterase homologous gene disruption mutants (pdeA, pdeB, and pdeC mutants) and discussed the function of the cyclic nucleotide phosphodiesterases in M. xanthus on the basis of the phenotypes of these mutants.

Analysis of pdeA, pdeB, and pdeC genes.

We searched the database of The Institute for Genomic Research (TIGR) with the tBLASTn program and found three genes (pdeA, pdeB, and pdeC) that encode proteins with homology to bacterial 3′,5′-cyclic nucleotide phosphodiesterases (Fig. 1A). The predicted pdeA, pdeB, and pdeC gene products contain 256, 284, and 366 amino acids, respectively, and showed 25, 22, and 20% identity to the E. coli cAMP phosphodiesterase (CpdA), respectively (6) (Fig. 2). The 3′,5′-cyclic nucleotide phosphodiesterases are classified by primary structure into three major classes (classes I, II, and III) (19). Almost all bacterial cAMP phosphodiesterases belong to class III. PdeA, PdeB, and PdeC contain the conserved sequence motifs, DXH-X_n-GD-X_n-GNH(E/D)-X_n-H-X_n-GHXH, where X is any residue, of bacterial 3′,5′-cyclic nucleotide phosphodiesterases. These residues can bind a binuclear Fe³⁺-Me²⁺ that is essential for catalysis (19). The conserved sequence motifs of phosphodiesterases are very similar to those of purple acid phosphatases (12, 19). A tyrosine residue is conserved in region II (GDXXY) of purple acid phosphatases, but it is not present in the class III phosphodiesterases. Glutamic acid residues of PdeA and PdeC or an alanine residue of PdeB were located in the position of a tyrosine in region II of purple acid phosphatases. In addition, the secondary structure predicted with PSIPRED V2.4 showed that a longer protein sequence between β-sheet 5 and α-helix 6 of purple acid phosphatase from pigs, which is the most obvious difference between the phosphatase and the phosphodiesterase in alignment, was not present in PdeA, PdeB, and PdeC (data not shown) (19). These findings suggest that M. xanthus PdeA, PdeB, and PdeC have characteristics of cyclic nucleotide phosphodiesterases. We could not find homologues of eukaryotic cyclic nucleotide phosphodiesterase in the M. xanthus database.

FIG. 1.

Restriction maps and RT-PCR analyses. (A) Restriction maps of the pdeA, pdeB, and pdeC genes of M. xanthus. Lines with arrows indicate the orientation. (B) RT-PCR analysis for polar effects. Total RNA was prepared from each of the strains, mutant and wild-type (WT) strains. The RNA (0.2 μg) was treated with (+) or without (−) reverse transcriptase and subjected to PCR with appropriate primers. (C) Cotranscription of pdeA and orfB or pdeB and orfC in the wild-type strain. Total RNA prepared from vegetative cells of the wild type was either treated (+) or not treated (−) with reverse transcriptase. PCRs were performed using pdeA-orfB primers and pdeB-orfC primers. The positions of molecular size markers (base pairs) are indicated to the left of the gels.

FIG. 2.

Amino acid sequence alignment of homologous regions in E. coli 3′,5′-cyclic nucleotide phosphodiesterase (CpdA) and M. xanthus PdeA, PdeB, and PdeC. Amino acid residues in agreement for more than three residues are indicated by white letters on black background. Five conserved regions (I to V) in class III 3′,5′-cyclic nucleotide phosphodiesterases are overlined (19). Gaps introduced to maximize alignment are indicated by dashes.

Construction of pde mutants.

To investigate the biological function of PdeA, PdeB, and PdeC, we constructed pdeA, pdeB, and pdeC single mutants. The oligonucleotide primers pdeA1 (5′-GTGAGGACTTGGATGAGCAGG-3′) and pdeA2 (5′-TGTTCGACGACCAGCTTCCCG-3′), pdeB1 (5′-CCTTCGTCGGAGAGTCCG-3′) and pdeB2 (5′-CCGCTGTCGTTCCACACG-3′), and pdeC1 (5′-TTCCCAGCAGTCGTTGGC-3′) and pdeC2 (5′-GGGCGGCTTCGTTTATCC-3′) were used to amplify the pdeA, pdeB, and pdeC genes, respectively, from the M. xanthus genome. The PCR products were ligated into the pT7BlueT vector to construct pT7pdeA, pT7pdeB, and pT7pdeC, respectively. pdeA::Km^r and pdeC::Km^r were constructed by inserting a kanamycin resistance gene cassette from Tn5 (4) into the StuI-PmaCI sites of pT7pdeA and EcoRV-PmaCI sites of pT7pdeC, respectively. pdeB::Tet^r was constructed by the following method. For deletion of an SmaI site of pBluescript II, pT7pdeB was digested with HindIII and BamHI, and a 2.3-kb HindIII-BamHI fragment containing pdeB was recovered from an agarose gel and ligated into pBluescript II previously treated with the same enzymes. The pBluescript II plasmid containing pdeB was digested with SmaI, and a 1.5-kb tetracycline resistance cassette was inserted between the SmaI sites of the plasmid. The disrupted genes constructed as described above were amplified by PCR using the above oligonucleotides. The PCR products thus obtained were introduced into M. xanthus FB (3) by electroporation (18).

Using PCR and restriction enzyme analyses, we confirmed that the kanamycin or tetracycline resistance gene was inserted into the pdeA or pdeC gene or the pdeB gene, respectively, on the chromosomes of M. xanthus mutants in the same orientation. We also confirmed that a downstream gene, orfB, orfC, or orfE, was transcribed in the pdeA, pdeB, or pdeC mutant, respectively, using reverse transcription-PCR (RT-PCR) (Fig. 1B). Total RNA was isolated from exponentially growing wild-type and mutant cells at 30°C (13). Contaminating DNA was removed by digestion with DNase I. For RT-PCR, 0.2 μg of RNA was used for cDNA synthesis with BcaBEST polymerase in accordance with the manufacturer's directions (Takara Bio). PCR was performed with Bca-optimized Taq polymerase using an orfB primer set (5′-GACTTCTTCACCTTCACG-3′ and 5′-CCGTGAGCTGAGCATCG-3′), an orfC primer set (5′-ATGACGCCGTAGGACAGG-3′ and 5′-TCACGACGTCTTCACGC-3′), and an orfE primer set (5′-AGAACGTGGAGCTGAGCG-3′ and 5′-TTCTCATCGTTGCCCTGC-3′). The expected 106-bp, 158-bp, or 121-bp product containing each part of the downstream gene, orfB, orfC, or orfE, respectively, was amplified from total RNA of each mutant when treated with reverse transcription, suggesting that the phenotypes of these mutants are not due to polar effects.

We examined whether the pdeA and orfB or pdeB and orfC genes were cotranscribed in wild-type strain using RT-PCR. RT-PCR was carried out by using primers (pdeA-orfB primer set [5′-TCCATGCGCACGATGTCC-3′ and 5′-ACGTACCAGGGTGACGCC-3′] linked pdeA and orfB and pdeB-orfC primer set [5′-CTGGCTCATCGACGTGGC-3′ and 5′-AGGCTGTCCTCCAAGGCG-3′] linked pdeB and orfC) that amplified a 213-bp fragment of pdeA-orfB, or a 178-bp fragment of pdeB-orfC (Fig. 1C). The results of these experiments suggested that pdeA and orfB or pdeB and orfC were transcribed as a single transcript.

Phenotypes of pde mutants (i) Cell growth.

The pdeA, pdeB, and pdeC mutants showed normal growth in Casitone-yeast extract (CYE) medium (1a) at optimal (30°C) or low (20°C) temperature. However, when cultured at a high temperature (37°C), the pdeA and pdeB mutants showed a significant reduction in growth compared to the wild type, where the cell density of the mutants in stationary phase decreased by about 50% (Fig. 3A). The deletion of pdeA or pdeB did not affect the growth rate. No significant difference was observed in growth at high temperature of the wild type and pdeC mutant.

FIG. 3.

(A) Growth of wild-type and mutant cells at high temperature (37°C). The wild type (circles) and pdeA (squares), pdeB (triangles), and pdeC (diamonds) mutants were grown in CYE medium at 30°C (open symbols) and 37°C (closed symbols). Experiments were repeated twice with similar results, and typical results are shown. (B) Intracellular cAMP levels in the wild-type cells and pdeA, pdeB, and pdeC mutant cells growing at 30°C and 37°C. Wild-type cells (W) and pdeA (A), pdeB (B), and pdeC (C) mutant cells were grown in CYE medium at 30°C (open bars) and 37°C (closed bars) to an optical density at 600 nm (1-cm path length) of 0.5 and harvested by centrifugation. The cells were washed with 50 mM Tris-HCl buffer (pH 7.5), resuspended in 0.2 ml of the buffer, and then immediately heated to 95°C for 5 min. After centrifugation, supernatants were used for assays with a commercial immunoassay kit (Amersham Pharmacia). Experiments were repeated two times. The standard deviations are shown by error bars.

To determine the intracellular concentration of cAMP under temperature stress, cell extracts were prepared from wild-type and mutant cells growing at an optimal (30°C) or high temperature (37°C), and the cAMP concentration in cell extracts was measured using the cAMP enzyme immunoassay kit (Fig. 3B). The cAMP levels of pdeA, pdeB, and pdeC mutant cells grown at 30°C were similar to that of the wild-type cells. The wild-type cells cultured at 37°C showed about 50% lower levels of intracellular cAMP than those cultured at 30°C. The pdeA and pdeB mutants cultured at 37°C also exhibited a decrease in cAMP levels, but the levels were still about 1.8 and 1.5 times higher than that of wild-type cells, respectively. We previously reported that an M. xanthus cAMP-binding protein gene (cbpB)-disrupted mutant also showed a significant reduction in growth compared to the wild type when cultured at high temperature. M. xanthus cultured at high temperature shows a reduction in intracellular cAMP levels that may be important for adaptation to growth under high-temperature stress.

Incubation at high osmolarity of M. xanthus cyaB mutant shows a similar growth rate and final cell density but displays a prolonged lag phase compared to wild type (10). The pdeA, pdeB, and pdeC mutants grew as well as the wild type did under osmotic stress (data not shown), suggesting that PdeA, PdeB, or PdeC is not involved in the adaptation to osmotic stress during vegetative growth.

(ii) Development.

To study the functions of PdeA, PdeB, and PdeC in development, the wild-type strain and these mutant strains were cultured on starvation medium (clone fruiting [cF] agar [4a]). When wild-type cells were allowed to develop on CF agar, aggregations and fruiting bodies formed at approximately 12 to 24 h and 48 to 72 h, respectively. The pdeA, pdeB, and pdeC mutants showed a normal developmental process, and the morphology of the fruiting body of each mutant strain was similar to that of the wild-type strain. The numbers of total and viable spores of the pdeA, pdeB, and pdeC mutants counted at day 6 of development were similar to those of the wild-type strain (data not shown). Ho and McCurdy reported that the cellular concentration of cAMP in M. xanthus during development increased rapidly about 20-fold during the first 15 h of incubation and then decreased 75% by 25 h, and the rapid drop in cAMP concentration corresponds to an increase in cAMP phosphodiesterase activity which exhibited a maximum at 40 h (5). McCurdy et al. also reported that exogenous cyclic nucleotide phosphodiesterase accelerates fruiting body formation in M. xanthus (16). These results suggest that cyclic nucleotide phosphodiesterases may be involved in the development of M. xanthus, however, ablation of the single cyclic nucleotide phosphodiesterase gene (pdeA, pdeB, or pdeC) has no noticeable effect on development. When cultured on CF plates containing 0.2 M NaCl or sucrose, the mutants also developed as well as the wild type did.

(iii) Spore germination.

Under nonstress conditions, spores of all mutants germinated normally in CYE medium (Fig. 4A). The mutant spores began to germinate after about 24 h of incubation and elongated into rod-shaped cells. The germination of wild-type and mutant spores was delayed by osmotic stress, but the mutant spores, especially pdeA and pdeB mutant spores, under osmotic stress germinated earlier than the wild-type spores (Fig. 4A). The pdeA and pdeB mutants incubated in CYE medium containing 0.2 M NaCl or sucrose germinated about 10 to 12 h earlier than the wild type. The phenotype was the opposite of that of M. xanthus receptor-type adenylyl cyclase (cyaA and cyaB) mutants (9, 10). The cyaA or cyaB mutant showed a considerable delay (about 12 to 28 h) in the germination of spores under high osmotic stress (0.2 M NaCl or sucrose). When not under osmotic stress, pdeA, pdeB, and pdeC mutant spores germinated slightly earlier than wild-type spores did. There was no difference in germination at 36°C between the wild type and these mutant strains (data not shown).

FIG. 4.

(A) Spore germination of the wild type (circles) and pdeA (squares), pdeB (triangles), and pdeC (diamonds) mutants in CYE medium containing 0.2 M NaCl or sucrose. Spores were harvested from 6- to 8-day-old fruiting bodies on CF agar plates, sonicated for 2 min, and heated (60°C for 15 min). The spores were inoculated to 1 × 10⁷ cells/ml in CYE medium alone (open symbols) or CYE medium containing 0.2 M NaCl or sucrose (closed symbols) and incubated at 30°C with continuous shaking. The number of ungerminated spores in each culture was counted with a hemacytometer. The percent germinated spores {[(number of inoculated spores − number of ungerminated spores)/number of inoculated spores] × 100 } is the mean of duplicate experiments. (B) Intracellular cAMP levels in the wild type and pdeA, pdeB, and pdeC mutants at germination under osmotic stress. Wild-type spores (W) and pdeA (A), pdeB (B), and pdeC (C) mutant spores were incubated in CYE medium alone (open bars) or CYE medium containing 0.2 M NaCl or sucrose (closed bars) at 30°C with continuous shaking until almost 30% of spores in the medium were germinated. Each culture was harvested and sonicated with 0.1 g of 0.4-mm glass beads for 4 min. After centrifugation, the supernatant was used for cAMP assays. Experiments were repeated two or three times. The standard deviations are shown by error bars.

We next determined intracellular cAMP levels in wild-type and mutant cells at germination with no osmotic stress or under high osmotic stress. When about 30% of spores germinated in CYE medium or CYE medium containing 0.2 M NaCl or sucrose, culture samples were taken and sonicated with glass beads. The supernatant was obtained by centrifugation and used to assay for cAMP. Under nonosmotic conditions, there were no significant differences in intracellular cAMP levels between wild-type and mutant strains. In contrast, the pdeA and pdeB mutants under high osmolarity showed 1.7- to 2.0-fold and 1.3- to 1.5-fold-higher levels than the wild type, respectively (Fig. 4B). These results may suggest that in M. xanthus spores under osmotic stress, intracellular cAMP levels are lowered by cyclic nucleotide phosphodiesterases, such as PdeA and PdeB, and the decrease in cAMP results in a delay of germination under osmotic conditions.

In conclusion, we suggested that PdeA and PdeB may have roles in regulating intracellular cAMP levels of M. xanthus cells during vegetative growth at high temperature and germination under osmotic stress and contribute to the adaptation of cells to these forms of stress.