INTRODUCTION

Top Abstract Introduction Materials & Methods Results Discussion References

Protein phosphorylation or dephosphorylation is a prominent mechanism found in all living organisms for mediating signal transmission. For many years, protein kinases and phosphoprotein phosphatases that catalyze protein phosphorylation and dephosphorylation, respectively, were thought to be different in prokaryotes and eukaryotes: protein phosphorylation occurs mainly on histidine and aspartic acid residues in prokaryotes but on serine, threonine, and tyrosine residues in eukaryotes. During the last few years, however, eukaryotic-type protein kinases and phosphatases have been reported in several bacteria, and homologs of prokaryotic protein kinases have been discovered in eukaryotes (for reviews, see references 1, 2, 16, 19, and 30). These studies suggest that eukaryotes and prokaryotes have similar mechanisms for signal transmission.

The relative activities of protein kinases and phosphatases determine the extent of phosphorylation at a particular site, which can in turn modulate the function of target proteins (16). Although initial studies on protein phosphorylation were concentrated on protein kinases, recent development in the understanding of protein phosphatases has firmly established that phosphatases are as important as kinases in signal transduction. Protein phosphatases in eukaryotes can be divided into two main groups based on their enzymatic specificity: Ser/Thr phosphatases and Tyr phosphatases (10, 26). The Ser/Thr-specific phosphatases show broad and overlapping specificities in vitro and have been grouped into four classes, PP1, PP2A, PP2B, and PP2C, according to their sensitivities to different inhibitors and their dependence on ions (3, 10, 17, 26). Amino acid sequence comparison indicates that PP1, PP2A, and PP2B are members of the same gene family, whereas PP2C represents a distinct gene family (3). The activities of PP1 and PP2A are not dependent on divalent cations, but those of PP2B and PP2C are affected by the presence of Ca²⁺ and Mg²⁺, respectively. PP1 is sensitive to nanomolar concentrations of inhibitor 1 and 2, while PP2A, PP2B, and PP2C are resistant (3, 17).

Anabaena sp. strain PCC 7120 (referred to hereafter as Anabaena) is a filamentous cyanobacterium that can modulate its cellular activity, including the differentiation of heterocysts devoted to nitrogen fixation, in response to environmental factors. After combined-nitrogen deprivation, heterocyst differentiation occurs following a semiregular pattern along each filament, with interheterocyst spacing ranging between 10 and 20 vegetative cells. The semiregular pattern is maintained during subsequent cell growth on dinitrogen, as new heterocysts arise between two existing ones (for reviews, see references 7 and 27). Mutations that disturb the heterocyst pattern usually affect cell growth in the absence of a combined-nitrogen source as well (5, 6, 8, 14, 20, 21). Late in heterocyst development, the nitrogenase responsible for nitrogen fixation begins to be synthesized. Heterocysts have a thick envelope with a glycolipid layer surrounded by polysaccharides. This structure ensures a microanaerobic environment within the heterocyst, preventing the inactivation of nitrogenase by oxygen (7, 27).

The presence of a family of Ser/Thr kinase-like proteins, which are well-documented signal transduction components in eukaryotes, has been demonstrated in Anabaena (29-31). This prompted us to search for cognate protein phosphatases within this strain. Here we report the isolation and characterization of a gene cluster with two open reading frames (ORFs) separated by only 301 bp that encode a phosphoprotein phosphatase and a protein kinase.

	MATERIALS AND METHODS

Top Abstract Introduction Materials & Methods Results Discussion References

Bacterial strains and growth conditions. Anabaena is grown, as already described (24, 29), in BG11 with either 18 mM of NaNO₃ or 5 mM of NH₄Cl (buffered with 10 mM HEPES [pH 7.5]) as the combined nitrogen source or with BG11₀, in which 18 mM of NaCl replaces the combined nitrogen. Escherichia coli strains used for cloning and conjugation are already described (12, 29).

PCR. To search for protein phosphatases PP1, PP2A, and PP2B, the two following PCR primers were used (R as purine, Y as pyrimidine, N as any of the four nucleobases): RW191, CTTGGATCCGGNGANRTNCAYGG (from the conserved motif GD I/V HG [11]); RW193, CTTGGATCCTCRTGRTTNCCNC (from the conserved motif RGNHE [11]). PCR was performed as described in reference 29, except that the annealing temperature was 57°C. The clones obtained after insertion into the BamHI site of pBluescript SK were analyzed by DNA sequencing and digestion with suitable restriction enzymes. The screening of the genomic library is described in reference 29.

Protein overexpression and immunotechnology. Both prpA and pknE genes were amplified with PCR primers tagged with appropriate restriction sites. The amplified catalytic domain-encoded region of PrpA corresponds to the first 517 amino acid residues of the entire PrpA protein. The inserts were cloned into either pET15b (Novagen) or pGEX-2T (Pharmacia). The pET15b constructs give rise to His-tagged recombinant proteins, whereas the pGEX-2T constructs give rise to glutathione S-transferase (GST) fusion.

Assay of phosphatase activity. To test phosphatase activity, 20 mM p-nitrophenyl phosphate (pNPP) was incubated with purified GST-PrpA fusion or with purified GST as a control in TBS buffer at 30°C in the presence of different ions or without ions. The reaction was then monitored by measuring the optical density at 410 nm (22).

Construction of prpA and pknE mutants. A 4.4-kb HindIII fragment bearing the prpA gene and a partial pknE gene (Fig. 1) was cloned into the HindIII site of pBluescript SK, giving rise to plasmid pHE10 (Fig. 1). The neomycin-resistant gene cassette was PCR amplified from pUC4K (Pharmacia) by a pair of primers bearing EcoRV/NdeI sites. The PCR products were digested either with NdeI and inserted into the unique NdeI site of pHE10 or with EcoRV and inserted into the double EcoRV sites of the pknE gene of pHE10. The former construct (pNEOA) disrupted the prpA gene, while the latter (pNEO8) disrupted the pknE gene (Fig. 1). The inserts of these two constructs were then cloned respectively into the PstI/XhoI sites of the conjugative vector pRL271 (12). The prpA and pknE mutants were obtained by conjugation as described in reference 12. Southern hybridization experiments confirmed that these two genes were completely inactivated (data not shown).

View larger version (16K): [in this window] [in a new window]   FIG. 1.   Restriction map of the prpA-pknE region. The coding regions of the two genes are shown by arrow bars below the restriction map. Shaded regions correspond to the catalytic domains of PrpA and PknE. DNA fragments (pHE10, pPPE3, pp-1, and pk-1) relevant to the text are also shown. The strategies used to inactivate either the prpA gene or the pknE gene are illustrated above the restriction map (for details, see Materials and Methods).

Nitrogenase assays and analysis of glycolipids. The nitrogenase assay was performed by measuring the acetylene reduction by gas chromatography coupled with mass spectroscopy as described previously (13, 31). Glycolipids were extracted and analyzed by thin-layer chromatography as described by Black et al. (4).

Nucleotide sequence accession number. The nucleotide sequence reported in this study has been submitted to the EMBL data bank under accession no. AJ224354.

	RESULTS

Top Abstract Introduction Materials & Methods Results Discussion References

Search for Ser/Thr phosphatase homologs in Anabaena. Two oligonucleotides have been designed against conserved regions of PP1-, PP2A-, and PP2B-type Ser/Thr phosphatases (11). PCR with these degenerate oligonucleotides as primers in the presence of 0.3 µg of chromosome DNA from Anabaena gave rise to a predominant DNA fragment of about 270 bp. This PCR product was cloned into the pBluescript vector, and five clones (PCR-PRP1 to PCR-PRP5) were sequenced. Their deduced amino acid sequences were identical and showed significant similarity to those of PP1-, PP2A-, and PP2B-type protein phosphatases. In particular, the sequence motif FLGDLVDR, found in the middle of the deduced sequence, was conserved in all PP1-, PP2A-, and PP2B-type protein phosphatases (data not shown). We designated this novel protein phosphatase PrpA. In order to determine if different protein phosphatases were also amplified by the two degenerate primers, 19 more clones were analyzed by comparing their restriction patterns with that of PCR-PRP1 and by DNA sequencing. However, all the clones were the same as those previously identified (data not shown).

Cloning and sequencing of prpA. The PCR-PRP1 insert was used as a probe to screen a genomic bank of Anabaena, and one positive clone was identified. Results from Southern hybridization with plasmid DNA isolated from this clone indicated that the prpA gene is located on a 4.4-kb HindIII fragment at one end of the insert.

View larger version (61K): [in this window] [in a new window]   FIG. 2.   Amino acid sequence comparison between PrpA and Ser/Thr phosphatases (A) and PknE and Ser/Thr kinases (B). Identical residues are in bold-faced type. Gaps (indicated by dashes) were introduced to optimize sequence alignment. PP1-AT, protein phosphatase 1 in A. thaliana (23); PP2A-hu, human protein phosphatase 2A (25); PP2B-DM, protein phosphatase 2B in D. melanogaster (15); PknA, protein kinase in Anabaena (29); Raf-rat, the Ser/Thr kinase Raf in rats (18).

A protein kinase gene (pknE) is located downstream of prpA. Three hundred one bases downstream of the prpA gene is a second ORF, whose deduced amino acid sequence is similar to protein Ser/Thr kinases (Fig. 1 and 2). This ORF, designated pknE, represents a putative protein kinase different from those previously identified by PCR (29, 31). The pknE gene predicts a protein of 475 residues with a molecular weight of 53,359. The region containing 280 residues at the N-terminal part of PknE is 37% identical to the corresponding region of PknA in Anabaena (29) and 15.5% identical to the catalytic domain of the mammalian oncogene Raf kinase (18). The C-terminal part of PknE does not show sequence similarity to any proteins in the data banks.

Biochemical characterization of PrpA. Both prpA and pknE were overexpressed in E. coli as His-tagged fusion proteins with the vector pET15b under the control of the T7 RNA polymerase. The sizes of the expressed proteins correlated well with the molecular weights predicted from the DNA sequence. Both the His-tagged, recombinant PrpA and PknE proteins were purified and used for the production of antisera in order to examine the protein level expressed under different conditions (see below).

View larger version (14K): [in this window] [in a new window]   FIG. 3.   Hydrolysis of pNPP by the recombinant GST-PrpA fusion protein. The affinity-column-purified proteins (GST and GST-PrpA fusion) were incubated at 30°C in the presence of different ions. The hydrolysis of pNPP was followed by measurement of the change in OD410 (22). , GST-PrpA with 2 mM Mn2+; , GST with 2 mM Mn2+; &atyp0220;, GST-PrpA with 2 mM Ca2+; and , GST-PrpA with 2 mM Mg2+.

Regulation of prpA and pknE expression during heterocyst development. The polyclonal antisera prepared against PrpA and PknE were used to detect the expression of the two proteins under various conditions. The specificities of the antisera were first controlled with total proteins extracted from an E. coli strain expressing the antigen. Total proteins of Anabaena were extracted from cells first cultured in the presence of nitrate and then transferred to combined nitrogen-limiting medium for heterocyst induction. One antiserum against PrpA detected a protein from the wild-type Anabaena that was absent from the prpA mutant A.3 (see below and Fig. 4A). The estimated size of the protein, approximately 95 kDa, is in good agreement with the value predicted from the DNA sequence. The level of PrpA expression, after correction for the relative amount of total proteins estimated from Coomassie blue staining, shows little variation during heterocyst differentiation. The antiserum also detected a protein slightly smaller in size than PrpA that requires further characterization.

View larger version (52K): [in this window] [in a new window]   FIG. 4.   Expression of PrpA and PknE examined by Western blotting. Total proteins were separated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis and blotted onto Immobilon membrane for Western analysis with antibodies against either PrpA (A and C) or PknE (B and D). To check the protein expression during heterocyst development (A and B), similar amounts of total proteins were loaded onto each lane of the gel. Protein samples were prepared from cells grown in the presence of nitrate (lanes 0, NO3) or transferred to nitrate-free medium (N2) for 1 to 96 h (lanes 1 to lanes 96).

Are PrpA and PknE coexpressed? A DNA fragment encoding the C-terminal region of PrpA and the entire PknE protein (pPPE3, Fig. 1) was inserted into the expression vector pET3a under the control of an IPTG-inducible promoter in E. coli. After induction, only a polypeptide of about 36 kDa corresponding to the C-terminal region of PrpA was detected and no trace of the 53-kDa PknE protein was visible (data not shown). This result suggests that, at least in E. coli, prpA and pknE may not form an operon.

Genetic analysis of prpA and pknE mutants. The prpA and pknE genes were inactivated in vivo. The prpA mutants (A.3 and B.2) were obtained by the insertion of a Neo^r cassette into one NdeI site within the prpA coding region (Fig. 1), and the pknE mutants (2.1, 2.2, and 8.1) were prepared by inserting the Neo^r cassette in the place of a 0.5-kb EcoRV fragment within the pknE gene (Fig. 1). Southern hybridization experiments confirmed that the mutants obtained were completely segregated (see Materials and Methods). A.3 and 8.1 were used for further analysis.

View larger version (14K): [in this window] [in a new window]   FIG. 5.   Growth curves of the wild type (), the prpA mutant A.3 (), and the pknE mutant 8.1 () under N2-fixing conditions. Cells were incubated in a combined nitrogen-free medium (BG110) at 30°C in air. Cell growth was monitored by measuring the OD700.

Nitrogenase activity and glycolipid synthesis in A.3 and 8.1 mutants. To determine whether nitrogen fixation activity was impaired in A.3 and 8.1 mutants, we performed nitrogenase assays with both mutants and the wild type. Under aerobic conditions, nitrogenase activity in the A.3 mutant is only about 2.2 and 3.4% of that of the wild type 2 and 4 days after combined-nitrogen stepdown, respectively (Table 1). The nitrogenase activity in the pknE mutant 8.1 is also affected, as its nitrogenase activity is 13.1 and 10.6% of that of the wild type 2 and 4 days after the cells were transferred into N₂-fixing conditions, respectively. Exposing the cells to anaerobic conditions only partly restores the nitrogenase activity in the A.3 and 8.1 mutants, as the nitrogenase activity reaches about 21.7 and 26.9%, respectively, of that of the wild type under similar conditions (Table 1).

	DISCUSSION

Top Abstract Introduction Materials & Methods Results Discussion References

The protein phosphatase PrpA characterized here, although homologous to the entire catalytic domains of PP1, PP2A, and PP2B, is not easily classified within any group of protein phosphatases based on the sequence similarity of the catalytic domains, since the degree of sequence similarity between PrpA and PP1 is close to that between PrpA and PP2A or between PrpA and PP2B (Fig. 2A). We were able to clone only one protein phosphatase gene from Anabaena by the approach used here, and no PP2C-type protein phosphatase could be detected by a similar approach (data not shown). However, by analogy with other organisms, a family of protein phosphatase genes can also be expected from Anabaena, in which a family of Ser/Thr kinases has been identified (29-31). The antiserum against PrpA reacts strongly with another protein of 92 kDa, which may be a PrpA homolog in Anabaena. Curiously, one gene encoding a protein similar to the catalytic domains of Ser/Thr phosphatases of types PP1, PP2A, and PP2B is found on the entire genome of Synechocystis sp. strain PCC 6803 (Cyanobase data bank). This protein, however, has no C-terminal regulatory sequence, as does PrpA, and no polypeptide similar to the C-terminal region of PrpA was detected in the Cyanobase. This result may suggest that prpA reported in this study is involved in the regulation of a biological process in Anabaena, such as nitrogen fixation or cell differentiation, that is absent from the unicellular strain Synechocystis sp. strain PCC 6803.

The prpA and pknE genes are closely linked on the chromosome, although their products have rather antagonistic actions. This unusual genetic organization offers an opportunity to study how the coordinated processes of phosphorylation and dephosphorylation ensure a balanced response to informational change in signal transduction. We are aware of only one example from Bacillus subtilis in which a close linkage between genes encoding protein kinases and phosphatases was reported (28). Two pairs of protein kinases and protein phosphatases encoded by a gene cluster in B. subtilis interact with each other by a partner-switching mechanism to transmit signals of environmental stress (28). It was also shown in human neurons that the cyclic AMP-dependent protein kinase A and the protein phosphatase 2B bind to a common anchoring protein, thereby forming a single protein complex to ensure their enzymatic specificity toward the common target (9). The inactivation of either prpA or pknE affects similar biological processes (heterocyst structure development, cell growth under N₂-fixing conditions, and nitrogen fixation activity), strongly suggesting that these two genes are involved in the regulation of similar targets in Anabaena. Both prpA and pknE mutants produce heterocysts with aberrant structures, which may account for their poor performance in nitrogen fixation and, ultimately, their inability to sustain growth in the absence of combined nitrogen. It is, however, difficult to pinpoint the defects of heterocyst structures of these mutants under the light microscope. The heterocyst-specific glycolipid synthesis, which is critical for nitrogen fixation (7, 27), is not affected by the inactivation of either prpA or pknE.

Since prpA and pknE are closely linked on the chromosome, we used different approaches to attempt to determine their genetic organization. Our results support the possibility that they are not necessarily cotranscribed, although additional experiments will be required to substantiate this conclusion. Our data suggest that prpA is constitutively expressed under the conditions tested here, while the expression of pknE is regulated during the process of heterocyst development. Immunoblotting and RNA/DNA hybridization experiments indicate that pknE is strongly expressed in the presence of nitrate, with its expression decreasing during the first few hours of nitrate deprivation and then increasing again. These observations suggest that the equilibrium between protein phosphorylation and dephosphorylation in response to environmental changes is regulated at the level of the protein kinase. The inactivation of a protein kinase or a protein phosphatase leads to unchecked protein dephosphorylation or phosphorylation, respectively, resulting in a misregulation of cell activity (16). The results of our studies demonstrate the importance of maintaining protein phosphorylation at a correct level, which requires the actions of both protein kinase and protein phosphatase.