Identification of PamA as a PII-binding Membrane Protein Important in Nitrogen-related and Sugar-catabolic Gene Expression in Synechocystis sp. PCC 6803 *

The PII signaling protein plays a pivotal role in the coordination of carbon and nitrogen metabolism in a wide variety of bacteria, Archaea,andplantchloroplasts.Byusingayeasttwo-hybridscreen-ing system, we identified a transmembrane protein, designated PamA (encoded by sll0985 ), as a PII-binding protein in Synechocystis sp. PCC 6803. The interaction between PII and PamA was confirmed in vitro , and the interaction was inhibited in the presence of ATP and 2-oxoglutarate, whereas the interaction was not influ-enced by the phosphorylation status of PII. Northern blot analyses revealed that the transcripts of a set of nitrogen-related genes, including nblA , nrtABCD , and ureG , were decreased in a pamA deletion mutant. The mRNA and protein levels of a group 2 (cid:1) factor SigE were also reduced by the pamA mutation, and transcripts for sugar catabolic genes, such as gap1 , zwf , and gnd that are under the control of SigE, were consequently decreased in the pamA mutant. In addition, the pamA mutant was found to be unable to grow in glucose-containing media. These results indicate that PamA has a role in the transcript control of genes for nitrogen and sugar metabolism in Synechocystis sp. PCC 6803.

ATP and 2-OG binding, Synechococcus and Synechocystis PII proteins are phosphorylated at a serine residue under nitrogen starvation (7,8). This phosphorylation level is also affected by carbon status (9), and recent analyses revealed that PII dephosphorylation specifically responded to intracellular 2-OG concentrations (10).
A PII-deficient mutant (MP2) of Synechococcus sp. PCC 7942 does not inhibit the nitrate/nitrite transporter in the presence of ammonium, which is the preferred nitrogen source for unicellular cyanobacteria; however, it inhibits the nitrate/nitrite transporter in the wild-type strain (11). High affinity bicarbonate transporters are constitutively activated regardless of the ambient carbon status in a PII-deficient mutant of Synechocystis sp. PCC 6803 (8). Thus, it was suggested that carbon metabolism is also regulated by PII in this cyanobacterium. However, the underlying molecular mechanisms by which the nitrate/nitrite and bicarbonate transporters are regulated by PII remain unknown.
The expression of nitrogen-related genes in cyanobacteria is activated by the global nitrogen regulator NtcA, a transcription factor that belongs to the CRP family (12,13). It was recently shown that 2-OG binds directly to NtcA and thereby promotes its interaction with target promoters (14 -16). Thus, both PII and NtcA are two independent 2-OG sensors among cyanobacteria. With regard to the potential role of PII in transcriptional regulation, it was shown that the level of expression of NtcA-regulated genes, such as ntcA, nirA, glnA, and amt1, was reduced in MP2 cells, implicating PII in transcriptional activation of NtcA regulons (17). Aldehni et al. (18) also revealed that PII was involved in regulating the in vivo activity of the NtcA regulons in Synechococcus sp. PCC 7942 (18). However, the underlying mechanism for this PII function also remains unclear.
At present, there are two proteins known to interact with PII in unicellular cyanobacteria. Irmler and Forchhammer (19) identified a PP2Ctype protein phosphatase PphA of Synechocystis sp. PCC 6803 that dephosphorylates phosphorylated PII proteins in vivo and in vitro (19). A second PII interacting protein was recently discovered when it was found that PII formed a complex with N-acetyl glutamate kinase, a key enzyme of arginine biosynthesis, and controlled the activity in Synechococcus sp. PCC 7942 (20,21). Despite these findings, it remains to be elucidated how PII is involved in transcriptional regulation, and no interacting protein involved in transcriptional regulation has been discovered so far.
In this study we used the yeast two-hybrid approach and identified a putative membrane protein PamA that binds PII in Synechocystis sp. PCC 6803. Transcripts for a set of nitrogen-related and sugar catabolic genes were reduced in a pamA mutant, suggesting that PamA might play a role in the control of their transcript abundances.

EXPERIMENTAL PROCEDURES
Bacterial Strains and Culture Conditions-The glucose-tolerant strain of Synechocystis sp. PCC 6803 (GT), isolated by Williams (22), and its pamA deletion mutant were grown on BG-11 plates or in BG-11 liquid medium (5 mM NH 4 Cl as the nitrogen source) at 30°C under continuous white light (ϳ70 mol of photons m Ϫ2 s Ϫ1 ) (23). For photomixotrophic growth, BG-11 plates were supplemented with 5 mM glucose.
Yeast Two-hybrid Library Screening-The full-length PII gene, glnB (encoded by open reading frame ssl0707), of Synechocystis sp. PCC 6803 was amplified by PCR with specific primers (see TABLE ONE) and subcloned into the pAS2-1-AscI bait vector, consisting of the pAS2-1 vector (Clontech) with an introduced AscI restriction site. Details of the two-hybrid screening procedure will be described elsewhere. 3 Affinity Purification of GST-PamA-(475-680) and Cellulose Binding Domain (CBD)-PII Fusion Proteins-A region of the Synechocystis sp. PCC 6803 genome encoding a COOH-terminal portion of PamA (amino acids 475-680) was amplified by PCR with specific primers (see TABLE ONE), digested with ScaI and XhoI, and inserted into the SmaI-XhoI sites of pGEX5X-1 (Amersham Biosciences). Full-length ssl0707 was amplified with specific primers (see TABLE ONE), digested with StuI and SacI, and inserted into the ScaI-SacI sites of pET35b (Novagen). The constructed plasmids encoding GST-PamA-(475-680) or CBD-PII were introduced separately into E. coli BL21 Codon Plus (Novagen) by transformation, expression of the encoded proteins was induced by the addition of 1 mM isopropyl-␤-D-thiogalactopyranoside (Wako) to 1 liter of LB medium, and the cells were subsequently cultured overnight at 37°C. The cells were then collected by centrifugation and lysed in a solution containing 40 mM Tris-HCl (pH 8.0), 5% glycerol, 5 mM EDTA, and 4.5% Triton X-100, and an insoluble fraction containing the recombinant protein was isolated by centrifugation of the cell 3 S. Sato, A. Muraki, M. Kohara, Y. Nakamura, and S. Tabata, manuscript in preparation. lysate at 17,400 ϫ g for 20 min at 4°C. The insoluble fraction was washed with cell lysis buffer, suspended in sterilized water, and solubilized by the addition of half a volume of a solution containing 8 M urea, 50 mM Tris-HCl (pH 8.0), and 10 mM dithiothreitol and incubated for 1 h at 37°C. The solubilized GST-PamA-(475-680) and CBD-PII proteins were dialyzed against PBS or CBinD buffer (20 mM Tris-HCl (pH 7.5), 50 mM NaCl) and purified with glutathione-Sepharose 4B (Amersham Biosciences) or CBinD 100 resin (Novagen), respectively. The total yields of GST-PamA-(475-680) and CBD-PII were ϳ4 and 3.5 mg, respectively. Protein concentration was determined with the Bio-Rad protein assay.
Production of Anti-PII and PamA Antiserum and Immunoblotting-For the production of anti-PII and PamA antiserum, fusion proteins were purified as described above. The protein concentrations and purities were examined by Bio-Rad protein assay and SDS-PAGE with Coomassie Brilliant staining, respectively. Each 2 mg of purified PII and PamA protein was injected into rabbits, and the antiserum production was performed by Qiagen (Tokyo, Japan). Immunoblotting was performed as described previously (24).
Far Western Blotting-3.0 g of GST or GST-PamA-(475-680) protein was subjected to SDS-PAGE with a 12% gel, and blotted onto Immobilon-P transfer membrane (Millipore). Cyanobacterial cell extracts containing nonphosphorylated or full-phosphorylated PII protein were prepared from an ammonia-grown GT cell culture or a GT cell culture deprived of nitrogen for 4 h. Cells were disrupted by sonication (30 s 5 times) in ice-cold PBS-T (PBS containing 0.1% Tween 20 (Sigma)), and debris was removed by centrifugation (17,400 g ϫ 10 min). In previous experiments we found that the amount of PII protein was increased 2.1 times under this nitrogen deprivation condition (data not shown). Subsequently, 100 g (for ammonia grown cells) or 48 g (for nitrogen-deprived cells) of protein was used for the reaction with the membrane in 10 ml of PBS-T for 3 h at 4°C. After washes with PBS-T (10 min 5 times), binding proteins were cross-linked by 0.5% formaldehyde in PBS-T at room temperature for 30 min. The reactions were quenched by incubation with 2% glycine in PBS-T for 10 min and then washed with PBS-T (10 min 2 times). The membranes were incubated in 3% bovine serum albumin, and PII associated with GST or GST-PamA-(475-680) was detected by the anti-PII antiserum. For Far Western analysis with endogenous PamA, 3.0 g of CBD-PII protein was similarly electrophoresed by SDS-PAGE and blotted onto a membrane as described above. To prepare cell extracts, cells were disrupted by sonication (30 s 5 times), and PamA proteins were solubilized with 1% Triton X-100(Sigma-Aldrich) in PBS and incubated for 90 min at 4°C. Debris was removed after centrifugation at 17,400 g for 10 min. 1.6  . Far Western analysis with the recombinant or endogenous PamA proteins. A, 3.0 g of GST or GST-PamA-(475-680) was electrophoresed with a 12% SDS-PAGE gel and blotted onto Immobilon-P membrane. Cell extracts containing nonphosphorylated PII or full-phosphorylated PII were incubated with the membranes, and the associated PII proteins were detected by anti-PII antiserum. Signals were visualized by reaction with alkali phosphatases with 5-bromo-4-chloro-3-indolyl phosphate/nitro blue tetrazolium as a substrate. As a control reaction, 1.5 g of GST and GST-PamA-(475-680) was resolved in the same gel and detected by anti-GST antibody (Molecular Probe). B, Far Western analysis with the endogenous PamA protein. 3.0 g of purified CBD-PII protein was resolved with a 12% SDS-PAGE gel. GT cell extracts containing the endogenous PamA protein was incubated with the membrane, and the PamA protein was detected with anti-PamA antiserum. Anti-SigE antiserum was used as a negative control. Detection of 0.6 g of the recombinant PII protein with the anti-PII antiserum is shown in the right panel for comparison.
mg of protein was incubated with the membrane in 10 ml of ice-cold PBS containing 1% Triton X-100 for 3 h at 4°C followed by washing with the same buffer (10 min 3 times). The subsequent steps were performed as mentioned above.
Construction of a pamA Deletion Mutant-A 5.5-kbp HpaI fragment of cosmid cs0120 (25) containing pamA was subcloned into pTZ18R (Amersham Biosciences ) that had been digested with HincII, ensuring that the orientation of pamA was the same as that of the lacZ gene of the vector. The resulting plasmid, p0985, was digested with SmaI and XhoI, rendered blunt-ended with a DNA Blunting Kit (TaKaRa), and selfligated to eliminate the BamHI site in the polylinker region. The new plasmid was digested with BamHI, resulting in the deletion of an ϳ2.4kbp fragment (from the ϩ165 position of the pamA-coding region to the 3Ј end of the downstream transposase gene sll0986), and the deleted portion was replaced with a kanamycin resistance cassette (26) to yield p0985B. Synechocystis sp. PCC 6803 cells were transformed with p0985B and selected on BG-11 plates containing kanamycin (10 g/ml).
Isolation of RNA and Northern Blot Analysis-Cells of mid-exponential phase cultures of Synechocystis sp. PCC 6803 (A 750 , 0.5-0.7) grown in BG-11 medium were collected by filtration and resuspended in BG-11 0 medium (BG-11 without NH 4 Cl). After culturing for 0 or 4 h, the cells were collected and subjected to RNA extraction by the acid phenol-chloroform method as previously described (27). Northern blot analysis of the isolated RNA was performed as described previously (28). For the construction of gene-specific probes, we amplified the corresponding coding region by PCR with specific primers (see TABLE ONE) and labeled the PCR products with digoxigenin with the use of a DIG DNA Labeling Kit (Roche Applied Science).
Complementation of the pamA Mutant-A DNA fragment including the pamA gene was obtained by digestion of the p0985 plasmid with SacII and EcoRV. The resulting 3.3-kbp fragment was blunt-ended with a DNA blunting kit (TaKaRa) and inserted into the cyanobacterial autonomous replication plasmid pVZ322 (29) digested with SmaI. The resultant complementation vectors, namely, pVZ322:pamAC, were introduced into GN10 cells by triparental gene transfer (29).

RESULTS AND DISCUSSION
Identification of a PII-binding Protein by Yeast Two-hybrid Screening-Yeast two-hybrid screening of 3.8 ϫ 10 6 independent clones of a Synechocystis genomic library with Synechocystis PII as bait yielded four positive clones. One of these four clones contained the PII gene (glnB) itself, consistent with the trimeric nature of PII, and the other three clones were found to harbor the 3Ј terminal region of a predicted gene sll0985. The potential protein encoded by sll0985 comprised 680 amino acids and contained seven membrane-spanning segments, as predicted by the SOSUI program (30) (Fig. 1A). Analysis of sequence similarity with the BLAST program indicated that this protein belonged to the protein family defined by the MSC mechanosensitive ion chan-  (TABLE ONE). B and C, Northern blot analysis of nitrogen-related genes in GT and GN10 cells. Total RNA was isolated from GT or GN10 cells 0 or 4 h after nitrogen deprivation and was subjected to Northern blot analysis (8 g of RNA/lane) with probes specific for the indicated genes. The positions of molecular size markers are indicated in kilobases. As a control, rRNA was stained with methylene blue. GT and GN10 cells were transferred to nitrogen-deprived medium and collected at the indicated times. Proteins were obtained by disrupting the cells by sonication. Total protein (7.5 g) was subjected to immunoblotting with anti-serum to SigE proteins. Production of antiserum to SigE was as described (39). C, down-regulation of gap1, zwf, and gnd in GN10 cells. Total RNA was isolated from GT or GN10 cells at 0 or 4 h after nitrogen deprivation and was subjected to Northern blot analysis (8 g of RNA per lane) with probes specific for the indicated genes. The positions of molecular size standards are indicated in kilobases.
nel of E. coli (31). The yeast two-hybrid analysis showed that PII interacted with amino acids 571-680 of Sll0985, located in the COOH-terminal tail region (amino acids 498-680) of the protein, a region that is well conserved among several cyanobacterial proteins (Fig. 1B). We designated the hypothetical gene sll0985 as pamA (PII-associated membrane protein A).
Interaction of PII with PamA in Vitro-To confirm in vitro the interaction between PII and PamA, as detected by the yeast two-hybrid screening, we expressed a GST fusion protein containing the COOHterminal region (residues 475-680) of PamA in E. coli and purified the recombinant protein to homogeneity by glutathione affinity chromatography ( Fig. 2A). PII was similarly expressed as a CBD fusion protein and purified to homogeneity. Either GST-PamA-(475-680) or GST bound to glutathione-Sepharose 4B beads was then incubated with CBD-PII. The beads were isolated by centrifugation, and bound proteins were subjected to immunoblot analysis with antibodies to the CBD tag. CBD-PII was found to coprecipitate with GST-PamA-(475-680) but not with GST (Fig. 2B). Furthermore, although neither ATP nor 2-OG alone affected the interaction between PamA and PII in vitro, the presence of both 1 mM ATP and 1 mM 2-OG prevented the association of the two proteins (Fig. 2B). In Synechococcus sp. PCC 7942, the binding of 2-OG to PII has previously been shown to require the presence of ATP (5). Our results may suggest that PII binds with PamA under nitrogen-repleted conditions and dissociates under nitrogen-limited conditions.
To further analyze the interaction between PII and PamA, we performed Far Western blotting analyses. GST or GST-PamA-(475-680) protein was resolved in a 12% SDS-PAGE gel and blotted onto a mem-brane. To examine the effect of PII phosphorylation on the PII-PamA interaction, cell extracts were prepared from ammonia-grown or nitrogen-starved GT cells and incubated with the membrane, and bound PII was detected with PII-specific antiserum. Both nonphosphorylated (ammonia-grown) and fully phosphorylated (nitrogen-starved) PII could interact with GST-PamA-(475-680) but not with GST (Fig. 3A). This indicates that phosphorylation states do not affect the PII-PamA interaction. 2-OG was increased immediately after nitrogen starvation (increased ϳ2-fold, 15 min after nitrogen depletion), whereas phosphorylation of PII proteins only took place 1 h after nitrogen depletion (14). Thus, these results imply that PII and PamA proteins may be promptly dissociated under nitrogen starvation by 2-OG accumulation.
Additionally, we examined the effect of PamA truncation on the PII interaction (Fig. 3B). In brief, purified CBD-PII was electrophoresed and blotted onto a membrane. Cell lysates prepared from GT cells were incubated with this filter, and endogenous PamA, interacting with the filter bound PII, was detected with the PamA antiserum. The results

Sequences of PCR primers
YTH, yeast two-hybrid screening; RP, recombinant protein production; Pr, probe generation for Northern blot analysis; Conf, confirmation of knockout of pamA; Cp, confirmation of complementation of GN10.

Gene Purpose
Forward or reverse Sequence showed that native PamA could also interact with PII, denying an artificial interaction by the PamA truncation.
Transcripts of Nitrogen-related Genes Were Decreased by the pamA Mutation-To analyze the functions of PamA, we constructed a pamA deletion mutant (GN10) of Synechocystis by transforming cells with a targeting plasmid in which a genomic fragment from the 5Ј region of pamA to the 3Ј end of the downstream transposase gene sll0986 was replaced by a kanamycin resistance cassette (Fig. 4A). This pamA deletion mutant grew equally well as the parental GT strain (data not shown). However, Northern blot analysis of the RNA expression of a set of nitrogen-related genes revealed that the transcripts of nblA, nrt-ABCD, and ureG were greatly reduced in GN10 compared with the GT strain (Fig. 4B). nblA (whose product is responsible for phycobilisome degradation) and the nrtABCD operon (which encodes the nitrate/nitrite transporter) are known to be regulated by NtcA (32)(33)(34)(35). On the other hand, the transcripts of other NtcA-regulated genes such as glnA (which encodes glutamine synthetase type I), glnB, amt1 (which encodes the ammonium transporter), and glnN (which encodes glutamine synthetase type III) were less affected by the pamA mutation (Fig.  4C). These genes, except for ureG, are known to be controlled by NtcA (36). The ureG gene, which encodes a urease accessory protein, was shown to be regulated by NtcA in Prochlorococcus marinus PCC 9511 (37) and, therefore, could also be regulated by NtcA in Synechocystis sp. PCC 6803. These results indicate that only a set of NtcA-regulated genes was affected by PamA disruption, suggesting that genes included in NtcA regulons might be sub-grouped by differential controls. Thus, the pamA function is involved in the expression of a part of the NtcAregulated genes.
The Expression of sigE and Sugar Catabolic Genes Was Decreased by the pamA Mutation-Because a group 2 factor sigE is regulated by NtcA and induced by nitrogen depletion (38), we examined the expression of the sigE gene in the pamA mutant. Northern analyses and immunoblotting revealed that the sigE transcript and the SigE protein levels were decreased by the pamA mutation (Fig. 5, A and B). Recently, we found that SigE activated transcription of sugar catabolic genes (39). Subsequently, we tested the transcript levels of gap1 (encoding glyceraldehyde-3-phosphate dehydrogenase), zwf (encoding glucose-6-phosphate dehydrogenase), and gnd (encoding 6-phosphogluconate dehydrogenase) and found that the mRNAs of gap1, zwf, and gnd were decreased by the pamA mutation (Fig. 5C). Consistent with the transcript levels, the enzyme activities of glucose-6-phosphate dehydrogenase and 6-phosphogluconate dehydrogenase were confirmed to be reduced by the pamA mutation (data not shown).
Complementation of the pamA Mutant-To genetically confirm that the observed characteristics of GN10 actually resulted from the pamA  . Immunoblot analysis with antiserum to PII. GT and GN10 cells were grown in the presence of ammonium and then transferred to a medium lacking combined nitrogen. Cells were collected at indicated times and disrupted by sonication. A, the amount of PII proteins in GT and GN10. Total protein (7.0 g) was subjected to electrophoresis with a 15% SDS-PAGE gel and immunoblotting with antiserum to PII. B, phosphorylation states of PII in GT and GN10. Total protein (5.0 g) was subjected to nondenaturing polyacrylamide gel electrophoresis and immunoblotting with antiserum to PII. The differently modified PII proteins were separated according to their increased negative charge. P 0 indicates nonphosphorylated PII trimer, and P 1 , P 2 , and P 3 represent PII trimers containing one, two, and three phosphorylated PII proteins, respectively. Interacting with PII, PamA positively regulates expression of nitrogen-related genes (nblA, nrtABCD, and ureG) and the group 2 factor SigE. This factor, in turn, activates sugar catabolic genes, such as gap1, zwf, and gnd. deficiency but not from a secondary mutation, we performed a complementation test with a plasmid containing the wild-type pamA gene, pVZ322:pamAC (Fig. 6A). After the introduction, the plasmid maintenance was confirmed by PCR ( Fig. 6B) with pamA-specific primers (TABLE ONE). Northern analysis revealed that the wild-type pamA gene restored the transcript levels of gap1 (Fig. 6C) and other genes (data not shown), indicating the decreased mRNAs of nitrogen-related and sugar catabolic genes were caused by the pamA mutation.
The pamA Mutant Exhibited Glucose-sensitive Phenotype-In search of phenotypes conferred by the pamA deletion, we found that GN10 cells were not able to grow on BG-11 plates supplemented with 5 mM glucose (Fig. 7A), whereas the parental glucose-tolerant strain could grow in the presence of glucose at concentrations of up to at least 30 mM (data not shown). The glucose-sensitive phenotype in GN10 was also observed in liquid cultures with 10 mM glucose (Fig. 7B). Under the same photomixotrophic condition, the sigE mutant did not show comparable sensitivity to glucose (data not shown), indicating that the phenotype was not caused only by the decreased amount of SigE protein in GN10. Koksharova et al. (40) demonstrated that a Gap1-deficient mutant of Synechocystis was unable to grow in glucose-containing medium (40), suggesting that the reduced transcription of glucose catabolic genes, including gap1, in GN10 cells might be responsible for their observed glucose sensitivity. The transcript levels of gap1 in the pamA mutant and the sigE mutant were decreased up to 20 -30 or 60 -70% of the parental GT strain, respectively. Thus, the lesser transcript accumulation of gap1 in the pamA mutant than in the sigE mutant could be a reason for the differential glucose sensitivity. Consistently, when glucose concentration in the medium was increased to 20 mM glucose in BG-11 liquid medium, the sigE mutant also showed glucose sensitivity (data not shown). We also confirmed that the complementation by the wild-type pamA gene with pVZ322:pamAC restored the glucose sensitivity in GN10 (data not shown).
The Hypothetical Function of PamA-At present we have no reasonable explanation on the exact role of PamA in nitrogen-related transcriptional regulation. Immunoblot analysis with antiserum to PII revealed that the amount of PII proteins was decreased in GN10 (Fig.  8A), whereas the phosphorylation states were identical to the parental strain (Fig. 8B). These results may suggest that the decreased amount of PII proteins accounts for the reduced gene expressions in GN10 (Fig. 9), as shown in Synechococcus sp. PCC 7942 (17). Structurally, PamA belongs to MSCs that respond to the hypoosmotic shock (31). When cells are suddenly shifted from high to low osmotic conditions, the osmotic mechanical tension results in the opening of MSCs and induces solvent effluxes to avoid cell bursting. Accordingly, cell viabilities were compared between the wild type and the pamA mutant after the osmotic down shifts, but the pamA deficiency did not result in any difference in viability (data not shown). Moreover, although MscS family proteins are widely conserved, apparent Sll0985 orthologs are not encoded by all cyanobacterial genomes, suggesting that this protein may be important only in Synechocystis and a few other cyanobacteria. Therefore, the significance of the MSC-related structure is unclear. In any cases, elucidation of the exact role of PamA should surely reveal a new aspect of PII function in this unicellular cyanobacterium