Ammonium/Methylammonium Permeases of a Cyanobacterium

Ammonium is an important nitrogen source for many microorganisms and plants. Ammonium transporters whose activity can be probed with [14C]methylammonium have been described in several organisms including some cyanobacteria, and amtgenes encoding ammonium/methylammonium permeases have been recently identified in yeast, Arabidopsis thaliana, and some bacteria. The unicellular cyanobacterium Synechocystis sp. PCC 6803 exhibited a [14C]methylammonium uptake activity that was inhibited by externally added ammonium. Three putativeamt genes that are found in the recently published complete sequence of the chromosome of strain PCC 6803 were inactivated by insertion of antibiotic resistance-encoding gene-cassettes. The corresponding mutant strains were impaired in uptake of [14C]methylammonium. Open reading framesll0108 (amt1) was responsible for a high affinity uptake activity (K s for methylammonium, 2.7 μm), whereas open reading frames sll1017(amt2) and sll0537 (amt3) made minor contributions to uptake at low substrate concentrations. Expression of the three amt genes was higher in nitrogen-starved cells than in cells incubated in the presence of a source of nitrogen (either ammonium or nitrate), but amt1was expressed at higher levels than the other two amtgenes. Transcription of amt1 was found to take place from a promoter bearing the structure of the cyanobacterial promoters activated by the nitrogen control transcription factor, NtcA.

Ammonium is an important nitrogen source for many microorganisms and plants. Ammonium transporters whose activity can be probed with [ 14 C]methylammonium have been described in several organisms including some cyanobacteria, and amt genes encoding ammonium/methylammonium permeases have been recently identified in yeast, Arabidopsis thaliana, and some bacteria. The unicellular cyanobacterium Synechocystis sp. PCC 6803 exhibited a [ 14 C]methylammonium uptake activity that was inhibited by externally added ammonium. Three putative amt genes that are found in the recently published complete sequence of the chromosome of strain PCC 6803 were inactivated by insertion of antibiotic resistance-encoding gene-cassettes. The corresponding mutant strains were impaired in uptake of [ 14 C]methylammonium. Open reading frame sll0108 (amt1) was responsible for a high affinity uptake activity (K s for methylammonium, 2.7 M), whereas open reading frames sll1017 (amt2) and sll0537 (amt3) made minor contributions to uptake at low substrate concentrations. Expression of the three amt genes was higher in nitrogen-starved cells than in cells incubated in the presence of a source of nitrogen (either ammonium or nitrate), but amt1 was expressed at higher levels than the other two amt genes. Transcription of amt1 was found to take place from a promoter bearing the structure of the cyanobacterial promoters activated by the nitrogen control transcription factor, NtcA.
Ammonium is a key compound in the assimilation of nitrogen in numerous biological systems because it is the inorganic form of nitrogen that is incorporated, usually via the glutamine synthetase/glutamate synthase cycle, into carbon skeletons. Ammonium present in the environment can be assimilated by many bacteria, yeast, fungi, algae, and higher plants. Ammonium solutions always contain ammonia (pK a [ammonium/ammonia], 9.25), which can diffuse through biological membranes (1,2). Diffusion of ammonia followed by trapping of intracellular ammonium by glutamine synthetase can represent a significant process for nitrogen acquisition, especially in organisms like some bacteria, which can grow in alkaline media. This process would be less important in fungi that grow in acidic media. On the other hand, there is evidence for the presence of ammonium transport systems in numerous organisms (2). The first report of active ammonium uptake, by Hackette et al. (3), concerned the fungus Penicillium chrysogenum. These authors introduced the use of [ 14 C]methylammonium as a probe for the activity of the ammonium permease, a technique that has proven useful to study ammonium transport in many other biological systems including bacteria (4) (the pK a of methylammonium/methylamine is 10.65). Bacterial ammonium/methylammonium permeases are commonly repressed by high concentrations of ammonium in a process that, in the enterobacteria, involves the nitrogen control Ntr system (Refs. 5 and 6; reviewed in Ref. 7).
Cyanobacteria are organisms that belong to the bacteria (or eubacteria) group (20) and are characterized by their ability to perform oxygenic photosynthesis. Sources of nitrogen used by many cyanobacteria include nitrate, dinitrogen, urea, and ammonium (21). In cyanobacteria, incorporation of ammonium into carbon skeletons takes place mainly through the glutamine synthetase/glutamate synthase cycle (Ref. 22; reviewed in Ref. 21). The pH values of cyanobacterial growth media are usually above neutral (23). Diffusion of ammonia through cyanobacterial cytoplasmic membranes has been demonstrated (24) and can provide, pulled by glutamine synthetase, a mechanism for the net uptake of ammonium (see e.g. Ref. 25). Expression of an ammonium/methylammonium transport activity in some cyanobacterial strains, including Synechococcus sp. PCC 7942, Anabaena azollae, and Anabaena variabilis, has also been reported (26,27). In Synechococcus sp. PCC 7942, the ammonium/methylammonium transport activity is repressed by growth in ammonium-containing medium (28) and requires an intact ntcA gene to be expressed (29). NtcA is a cyanobacterial transcriptional regulator, homologous to Crp from E. coli, that activates the expression of ammonium-repressible genes in the absence of ammonium (30,31). The DNA target to which NtcA binds in the promoter of the regulated genes has been characterized in Synechococcus sp. PCC 7942 and contains the sequence signature GTAN 8 TAC located about 22 bp 1 upstream from a Ϫ10, Pribnow box in the form TAN 3 T (31). This promoter structure for NtcA-regulated genes is predicted to be valid for other cyanobacteria as well, because the putative helix-turn-helix motif for binding to DNA is identical in the NtcA polypeptides from different sources, including Synechocystis sp. PCC 6803 and Anabaena sp. PCC 7120 (32).
The whole sequence of the chromosome of the unicellular cyanobacterium Synechocystis sp. PCC 6803 has been determined (33) 2 and shown to contain three ORFs, namely sll0108, sll0537, and sll1017, that would encode polypeptides homologous to the AMT1 and MEP gene products cited above. Another ORF (sll0895) described by Kaneko et al. (33) as ammonium transport protein would actually encode a homologue of CysQ (see above). In this work, we show that the three putative amt genes of Synechocystis sp. PCC 6803 actually provide the cells with the capability to take up [ 14 C]methylammonium from the extracellular medium and are subjected to repression by ammonium.

EXPERIMENTAL PROCEDURES
Strains and Growth Conditions-Synechocystis sp. strain PCC 6803 was grown axenically in BG11 (nitrate-containing) medium (23) or in BG11 0 (nitrogen-free) medium supplemented with 5 mM NH 4 Cl and 10 mM TES-NaOH buffer (pH 7.5). For plates, the medium was solidified with 1%, separately autoclaved agar (Difco). Cultures were grown at 30°C in the light with shaking (80 -90 rpm) for liquid cultures. Synechocystis sp. PCC 6803 mutants carrying gene-cassette C.K3 (34) were routinely grown in medium supplemented with 30 -50 g of Km⅐ml Ϫ1 , and mutants carrying gene-cassette C.C1 (34) were grown in medium supplemented with 10 -80 g of Cm⅐ml Ϫ1 . Other cyanobacterial strains used in this work (Anabaena sp. PCC 7120, Nostoc sp. PCC 7107, Nostoc sp. PCC 7413, Calothrix sp. PCC 7601, Fischerella muscicola UTEX 1829, Pseudanabaena sp. PCC 6903, and Synechococcus sp. PCC 7942) were grown in BG11 medium. Cultures used for RNA isolation and for some of the uptake assays shown in Fig. 1, as indicated, were grown in BG11C (BG11 supplemented with 0.84 g⅐l Ϫ1 NaHCO 3 ) or BG11 0 C medium supplemented with 15 mM NH 4 Cl and 30 mM TES-NaOH buffer (pH 7.5) bubbled with a mixture of CO 2 (1.5% v/v) and air. To achieve nitrogen starvation of cultures, nitrate-grown cells (ammonium-grown cells in the experiment shown in Fig. 7) were harvested at room temperature, washed with and resuspended in BG11 0 medium (or BG11 0 C medium, in the case of cultures used for RNA isolation), and incubated for 6 h under culture conditions. Cyanobacterial cell mass was estimated by measuring the concentration of Chl of the cultures. Chl was determined in methanolic extracts (35).
E. coli strains DH5␣, GM48, and BL21 were grown in LB medium with, when necessary, 50 g of Ap⅐ml Ϫ1 , 50 g of Km⅐ml Ϫ1 , or 25 g of Cm⅐ml Ϫ1 .
Methylammonium Uptake Assays-Wild-type or mutant Synechocystis cells grown in nitrate-or ammonium-containing medium or incubated in the absence of any nitrogen source for 6 h were harvested by low speed centrifugation at room temperature, washed with 20 mM KH 2 PO 4 , 10 mM NaHCO 3 -NaOH buffer (pH 7.1) and resuspended in the same buffer. After a preincubation at 30°C in the light (100 watt⅐m Ϫ2 , white light) for 5 to 30 min, the assays were started by mixing the suspension of cells (4 to 15 g of Chl⅐ml Ϫ1 ) with a solution of [ 14 C]CH 3 NH 2 ⅐HCl (50 Ci⅐mol Ϫ1 ; ICN Pharmaceuticals, Inc.) in phosphate-bicarbonate buffer. After incubation for the time periods indicated in each experiment, 0.1-to 1-ml samples were filtered (0.45-m pore size Millipore HA filters were used) and washed with 2 to 5 ml of phosphate-bicarbonate buffer. The filters carrying the cells were then immersed in scintillation mixture, and their radioactivity was measured. Retention of radioactivity by boiled cells was used as a blank. In some experiments, as indicated, 1 mM L-methionine-D,L-sulfoximine was added to the cell suspension 25 min before the assay was started.
Intracellular Accumulation of Labeled Methylammonium-Filters containing cells that had been used in uptake assays as described above were immediately immersed in 2 or 3 ml of boiling water and incubated at 100°C for 5 min. The filters were removed, and the resulting suspensions were centrifuged. Samples from the supernatants were lyophilized and dissolved in a small volume of water. Metabolites present in these samples were resolved by thin layer chromatography using 0.1-mm cellulose plates (20 by 20 cm; Merck). The solvent used was isopropanol/formic acid/water (40:2:10, v/v) (26). The resulting radioactive areas were quantified in an InstantImager scanner for ␤ particles (Packard). To calculate the intracellular concentration of [ 14 C]methylammonium, an intracellular volume of 125 l⅐mg Ϫ1 Chl was assumed (36,37).
DNA and RNA Isolation and Manipulation-Isolation of genomic DNA from cyanobacteria was carried out as described previously (38), except for strain UTEX 1829 cells, which were frozen with liquid air and broken by grinding with glass beads in a mortar. Isolation of total RNA from Synechocystis sp. PCC 6803 was made as described previously (39). Other molecular biology manipulations were carried out by standard procedures (40).
PCR products were cloned in pGEM-T vector (Promega). Plasmids containing PCR products generated with primers Tr18-Tr19, Tr2a-Tr2b, and Tr3a-Tr3b were named pCSX23, pCSX53, and pCSX52, respectively. Plasmid pCSX23 was digested with ApaI and PstI to isolate the insert that was then cloned between the ApaI and PstI sites of pBluescript SK(ϩ), generating plasmid pCSX46. The 1.1-kb Km r genecassette C.K3 (34) excised with SmaI was inserted into a unique HincII site that is present in the strain PCC 6803-derived insert of pCSX46 to generate plasmids pCSX47a and pCSX47b (both orientations). The 1.9-kb Cm r gene-cassette C.C1 (34) digested with HincII was inserted into the unique MscI site of pCSX53 and into the unique MscI site of pCSX52 (these MscI sites are present in the strain PCC 6803-derived inserts of the plasmids), rendering plasmids pCSX57 and pCSX56, respectively. The orientation of C.C1 in the inserts of pCSX56 and pCSX57 was not determined. Sequences of the inserts of pCSX23, pCSX53, and pCSX52 were verified by using a T7 Sequencing kit (Amersham Pharmacia Biotech) and [␣-35 S]thio-dATP.
Transformation of Synechocystis sp. PCC 6803 with pCSX47a, pCSX47b, pCSX56, or pCSX57 as well as transformation of insertional mutant strain CSX47a with pCSX56 or pCSX57 was carried out as described previously (41). Transformants were selected in BG11 solid medium supplemented with 30 g of Km⅐ml Ϫ1 or 10 g of Cm⅐ml Ϫ1 . Km r and Cm r transformants were then grown in medium supplemented with 50 g of Km⅐ml Ϫ1 or 20 -80 g of Cm⅐ml Ϫ1 . To test whether the resulting mutant strains were homozygous for the mutant chromosomes, PCR amplification with genomic DNA from the mutants as templates and the corresponding primers was carried out; additional testing was made by hybridization.
Southern and Northern Blotting and Hybridization-Southern analysis was carried out as described previously (41) using GeneScreen Plus membranes (DuPont). For Northern blots, 10 to 15 g of RNA were loaded/lane and electrophoresed in 1% agarose denaturating formaldehyde gels. Transfer to nylon membranes (Hybond N-plus, Amersham Pharmacia Biotech; GeneScreen Plus, NEN Life Science Products), prehybridization, and washes were carried out following the manufacturer's recommendations; hybridization was performed at 42°C in the presence of 50% formamide, 5ϫ SSPE (1ϫ SSPE is 0.18 M NaCl, 10 mM sodium phosphate, and 1 mM EDTA (pH 7.4)), 5ϫ Denhardt's solution (40), 0.1% SDS, and 50 g of herring sperm DNA⅐ml Ϫ1 . Filters were washed twice at room temperature with 2ϫ SSPE and 0.1% SDS for 10 min and once with 1ϫ SSPE and 0.1% SDS at 65°C for 15 min. Total cpm of radioactive areas in Northern blot hybridizations were determined with an InstantImager scanner.
DNA probes used in the hybridizations were obtained by PCR using pCSX23, pCSX53, and pCSX52 as templates and the corresponding oligonucleotides as primers. Probes were labeled with a DNA labeling kit (Ready to Go, Amersham Pharmacia Biotech) and [␣-32 P]dCTP.
Primer Extension Analysis-A 525-bp fragment putatively containing the promoter region of ORF sll0108 was amplified using primers Am1 (5Ј-GGGAGCCACTAAAGTTCACAGG-3Ј; corresponds to positions Ϫ236 to Ϫ215 with respect to the translational start of ORF sll0108) and Am2 (5Ј-CGGCTATCAAAATCCAGATGGC-3Ј; complementary to nucleotides ϩ289 to ϩ268 with respect to the translational start of ORF sll0108). The PCR product was cloned in pGEM-T vector, rendering plasmid pCSX49. This plasmid was used to generate dideoxy-sequencing ladders for primer extension analysis. Oligonucleotides used for primer extension were Am2 (see above) and Am3 (5Ј-GGAACACAGGC-CAACCAGGGAG-3Ј; complementary to nucleotides ϩ146 to ϩ125 with respect to the translational start of ORF sll0108). Oligonucleotides were end-labeled with T4 polynucleotide kinase (Boehringer Mannheim) and [␥-32 P]dATP as described (45)  Reaction mixtures were then treated with RNase A (DNase-free, Boehringer Mannheim) and extracted with phenol, and the DNA was precipitated with sodium acetate and ethanol and resuspended in formamide loading dye. These preparations were loaded onto 6% polyacrylamide sequencing gels next to the corresponding dideoxy-sequencing ladders.
Band-shift Assays-A 382-bp fragment putatively containing the promoter region of ORF sll0108 was amplified by PCR using pCSX49 as template and primers Am1 and Am3 (see above and Fig. 8A). This fragment was digested with BstXI, and the resulting 138-and 244-bp fragments were isolated and used in nonradioactive band-shift assays. Binding assays were carried out in a final volume of 20 l containing 6 mM HEPES-NaOH buffer (pH 8.0), 7 mM Tris-HCl (pH 8.0), 42.5 mM KCl, 0.5 mM EDTA, 0.5 mM dithiothreitol, 0.375 mM MgCl 2 , 0.0025% gelatin, 0.001% Tween 20, 0.05 g⅐l Ϫ1 bovine serum albumin, 8% glycerol, 75-100 ng of the DNA fragment to be tested, 75 ng of a control DNA fragment, and 5 l of a diluted E. coli BL21 extract in which the NtcA protein from Anabaena sp. PCC 7120 had been overproduced from the isopropyl-␤-D-thiogalactopyranoside-inducible promoter of vector pQE9 (Quiagen, Inc). 3 Extract of E. coli BL21 was prepared according to the manufacturer's recommendations, and 1:100 and 1:200 dilutions of an extract containing 56.8 g⅐l Ϫ1 total protein were used in the experiments. The DNA fragment used as nonrelated, control DNA was a 223-bp fragment of pBluescript SK(ϩ) obtained by PCR amplification using the Universal M13 -20 and Reverse primers. Assay mixtures were incubated at room temperature for 30 min, loaded in a nondenaturing 6% polyacrylamide gel, and electrophoresed at 200 V. Gels were stained with ethidium bromide for visualization of DNA bands.
Glutamine Synthetase Activity-Glutamine synthetase activity was measured by using the ␥-glutamyltransferase assay as described previously (46). One activity unit corresponds to one mol of ␥-glutamylhydroxamate produced/min.

RESULTS
Transport of Labeled Methylammonium-The cellular activity level of uptake of [ 14 C]methylammonium in Synechocystis sp. PCC 6803 was higher in cells starved for nitrogen than in nonstarved cells (grown with either nitrate or ammonium). Additionally, growth in CO 2 -enriched air allowed higher [ 14 C]methylammonium uptake activities (Fig. 1). Determination of the rate of [ 14 C]methylammonium uptake in 1-min assays using a range of substrate concentrations from 1 to 75 M showed one kinetic component with the following parameters: K s , 2.7 M; V max , 169 nmol⅐min Ϫ1. (mg of Chl) Ϫ1 . Methylammonium does not serve as a nitrogen source in strain PCC 6803 (not shown), and as is the case in some other cyanobacteria (26,27), intracellular radioactivity derived from [ 14 C]methylammonium was distributed into two metabolites (Fig. 2). One of these metabolites was identified as methylammonium itself by co-chromatography with authentic [ 14 C]methylammonium. Production of the second metabolite was hampered (96.4 and 97.5% inhibition in two independent experiments, respectively) by treatment of the cells with L-methionine-D,L-sulfoximine, an inhibitor of glutamine synthetase, putatively identifying that metabolite as ␥-glutamylmethylamide (47,48). In experiments where [ 14 C]methylammonium was supplied at 1 M, intracellular methylammonium concentrations of about 50 M, which represented, at the time of sampling, an intracellular to extracellular methylammonium ratio of close to 60, were detected. The process of [ 14 C]methylammonium uptake was completely inhibited by the addition of ammonium (Fig. 3). Inhibition by ammonium was reversible, and the lag period before [ 14 C]methylammonium uptake commenced was almost proportional to the added concentration of ammonium. Assuming that [ 14 C]methylammonium uptake started when ammonium was exhausted from the incubation medium, an ammonium uptake rate of 780 nmol⅐min Ϫ1. (mg of Chl) Ϫ1 can be calculated that is similar to the figures obtained in direct determinations of ammonium uptake (not shown).
Glutamine synthetase and the capability to take up 1 M [ 14 C]methylammonium were determined in cells incubated for 6 h in BG11 0 medium for each of the mutants described above and the wild-type strain PCC 6803. Similar glutamine synthetase activity levels, 40 to 50 units⅐(mg of Chl) Ϫ1 , were found in all the strains. On the other hand, inactivation of ORF sll0108 abolished [ 14 C]methylammonium uptake by about 97%, whereas inactivation of sll0537 or sll1017 had only minor effects (Table I). Nonetheless, time course assays carried out with low substrate concentrations (5 to 270 nM) confirmed that inactivation of sll0537 or sll1017 further impaired [ 14 C]methylammonium uptake in the sll0108-inactivated mutant (see data for 50 nM [ 14 C]methylammonium in Fig. 5). The three investigated ORFs appear therefore to encode permeases that mediate methylammonium uptake. We have named them as amt1 (sll0108), amt2 (sll1017), and amt3 (sll0537), respectively.
The effect of the concentration of [ 14 C]methylammonium on the rate of uptake was studied in mutant strains CSX47a,  Analysis of Expression of the amt Genes-The expression of each amt gene was studied by Northern analysis using total RNA isolated from cells of strain PCC 6803 grown with nitrate or ammonium as the nitrogen source or grown with nitrate and incubated in the absence of any source of nitrogen for 6 h. As a probe, a DNA fragment generated by PCR as described above (see also Fig. 4) was used for each amt gene.
A single mRNA of about 2 kb was detected for amt1 (Fig. 6A). The relative levels of this transcript in the tested nitrogen regimes were 1:1.76:30 (ammonium:nitrate:minus nitrogen). An mRNA of about 1.5 kb was observed for amt2 only in nitrogen-starved cells (Fig. 6B). (A band of about 2.6 kb was also detected with the amt2 probe and RNA from the three different nitrogen regimes. It is possible, however, that this band represents nonspecific hybridization with some rRNA.) Although hardly visible, a 2.6-kb mRNA was detected only in nitrogen-starved cells for amt3 (Fig. 6C). These results showed that the three amt genes of strain PCC 6803 are expressed and that, in the three cases, expression is activated under nitrogen stress. Indeed, activation of expression took place quickly in response to nitrogen starvation, as shown in Fig. 7 for amt1 (similar results were obtained with an amt2 probe, whereas the amt3 transcript was hardly visible at any incubation time (not shown)).
Because the amt1 transcript was more readily detectable than the amt2 or amt3 transcripts, amt1 appears to be expressed at higher levels than the other two genes. To corroborate this observation, we performed experiments where filterbound, PCR-generated DNA fragments from the three amt genes were hybridized to 32 P-labeled total RNA. The RNA was isolated from cells of strain PCC 6803 grown with ammonium or nitrate as the nitrogen source or grown with nitrate and starved for nitrogen for 6 h. With RNA isolated from nitrogenstarved cells, hybridization was detected for the three genes, and the relative level of labeling was 1:6:54 (amt3:amt2:amt1) (Table II). With RNA isolated from ammonium-or nitrategrown cells, only hybridization to DNA corresponding to the amt1 gene was detected.
Possible NtcA-dependent Transcription Start Point for the amt1 Gene-The possible tsp of amt1 was investigated by primer extension analysis. Two oligonucleotides, Am2 and Am3, complementary to sequences located close to the 5Ј end of the amt1 gene (Fig. 8A) were used as primers. An extension product whose 3Ј end corresponded to a T doublet located 142 nucleotides upstream from the amt1 start codon was detected with the Am3 primer. This extension product was much more efficiently obtained with RNA isolated from nitrogen-starved cells than from cells grown with nitrate and more with RNA from nitrate-than from ammonium-grown cells (Fig. 8B). The possible tsp of the amt1 gene can thus be localized to the A doublet indicated in Fig. 8A that is preceded by a putative Ϫ10 box in the form TTGAAT. Seventeen nucleotides upstream from this box a TACAGA hexamer is found that might be considered a poor Ϫ35 promoter box. However, 20 nucleotides upstream from the Ϫ10 box a nucleotide sequence, AAAAGTAN 8 TAC, is found that would represent a perfect NtcA-binding site (Fig. 8A).
Binding of NtcA to that putative NtcA-binding site was tested by using an extract of an E. coli strain carrying a plasmid bearing the Anabaena sp. PCC 7120 ntcA gene as a source of NtcA and a 138-bp DNA fragment corresponding to the first 138 nucleotides shown in Fig. 8A. Binding to a 244-bp fragment containing the putative NtcA-binding site that is located immediately upstream of the amt1 gene was also tested (see Fig.  8A). In both cases, a nonrelated, control DNA was included in the assay. As shown in Fig. 8C, retardation was only observed with the 138-bp DNA fragment containing the putative amt1 promoter. On the other hand, no retardation was observed in binding assays carried out with extracts of an E. coli strain bearing expression vector pQE9 (not shown). These results suggest that the 138-bp DNA fragment bears a real NtcAbinding site, whereas the significance, if any, of the putative NtcA-binding site present in the 244-bp DNA fragment remains to be investigated.
Occurrence of amt Homologous Sequences in Other Cyanobacteria-The same DNA fragments corresponding to the three amt genes used to probe Northern blots were used in Southern blot analyses to investigate the presence of putative amt homologues in some other cyanobacteria (not shown). Hybridization with the amt1 probe was observed for every cya-

TABLE II
Hybridization of 32 P-labeled RNA to DNA fragments of the strain PCC 6803 amt genes Total RNA was isolated from ammonium-or nitrate-grown cells or from cells grown with nitrate and starved for nitrogen for 6 h, labeled with 32 P, and hybridized to filter-bound PCR-generated DNA fragments (2 pmol) corresponding to each amt gene, as described under "Experimental Procedures." Note that data for different nitrogen regimes correspond to different hybridizations. Therefore, only data for the different genes under a given nitrogen regime can be strictly compared. Data are radioactivity associated to a DNA fragment.  8. Analysis of the amt1 promoter. A, nucleotide sequence of the 5Ј end and sequences upstream of amt1 (taken from Kaneko et al. (33)). The location of primers Am1, Am2, and Am3 as well as of the possible tsp (ϩ1) and promoter elements (Ϫ10 box and NtcA-binding site) of the amt1 gene is indicated. Note a second putative NtcA-binding site (GTAN 8 TAC) just in front of the ORF start. B, primer extension using the Am3 primer and total RNA isolated from cells grown with ammonium (1) or nitrate (2) or grown with nitrate and incubated for 6 h in the absence of any nitrogen source (3). The arrow points to the extension product identifying the putative tsp. C, band-shift assay of a DNA fragment containing the putative promoter of the amt1 gene (138-bp fragment, from the start of the sequence shown in A to the BstXI site) (lanes 1 to 3) or a DNA fragment carrying the 5Ј end and some sequences upstream of the amt1 gene (244-bp fragment, corresponding to nucleotides 139 through 382 of the sequence shown in A, i.e. from the BstXI site to the Am3 primer) (lanes 4 to 6). In addition, as a control, all assays contained a nonrelated DNA fragment from pBluescript SK(ϩ) (223-bp fragment). As a source of NtcA protein, a cell-free extract from an E. coli strain carrying an Anabaena sp. PCC 7120 ntcA clone was used (see "Experimental Procedures"). Cell-free extract added: 0 g of protein (lanes 1 and 4), 1.4 g of protein (lanes 2 and 5), and 2.8 g of protein (lanes 3 and 6). muscicola UTEX 1829). Appreciable hybridization with the amt2 probe was observed with DNA from all the strains but UTEX 1829 and PCC 7120. A clear, though weak, hybridization signal with the amt3 probe was only observed with DNA from UTEX 1829. DISCUSSION Synechocystis sp. PCC 6803 shows an activity of [ 14 C]methylammonium uptake that exhibits characteristics similar to those of methylammonium uptake in other cyanobacteria including Synechococcus sp. PCC 7942 and A. variabilis ATCC 29413 (26,27). Thus, [ 14 C]methylammonium uptake would result from an initial transport of methylammonium followed by further transport and metabolism via glutamine synthetase. Under our experimental conditions, an accumulation of [ 14 C]methylammonium within the cells representing an intracellular to extracellular ratio of methylammonium of up to about 60 was observed. Ratio values of up to 200 or 40 have been reported for strains PCC 7942 and ATCC 29413, respectively (26,27). An intracellular to extracellular ratio of 60 would correspond, under the incubation conditions used in this work, to a free energy change for [ 14 C]methylammonium transport equivalent to ϩ107 mV. Because the membrane potential of photosynthetically active cyanobacterial cells is known to be in the range of Ϫ110 to Ϫ130 mV (49,50), the observed accumulation of methylammonium is close to that which would be permitted by the membrane potential. The effects of some metabolic inhibitors on methylammonium uptake have also been interpreted in terms of methylammonium transport being dependent on the membrane potential of the cells (1,16,26,27). Interestingly, methylammonium influx has been shown to decrease the membrane potential in some algae (51,52). It should be noted, however, that the observed accumulation of [ 14 C]methylammonium may represent an underestimation of the methylammonium transport activity of the cells, because some of the [ 14 C]methylammonium taken up may be leaking out from the cells by diffusion as [ 14 C]methylamine (24).
Inactivation of each of the three putative amt genes of Synechocystis sp. PCC 6803, namely sll0108 (amt1), sll0537 (amt3), and sll1017 (amt2), impairs [ 14 C]methylammonium uptake (Table I, Fig. 5). Tight inhibition by ammonium of [ 14 C]methylammonium uptake (see Fig. 3 and Refs. 3,12,16,26,53,54) suggests that ammonium is the natural substrate for the permease(s) taking up methylammonium, especially in organisms for which methylammonium is not a nutrient. The affinity of those permeases for ammonium would be higher than for methylammonium (K s can be 40 to 100 times lower for ammonium than for methylammonium (Refs. 3,11,53)). The Amt1 permease, which appears to be responsible for the K s 2.7 M methylammonium uptake kinetic component exhibited by strain PCC 6803, would thus represent a transport system with a very high affinity for ammonium and would therefore be able to mediate the uptake of ammonium that may be found at very low concentrations in some natural habitats. Amt2 and Amt3, on the other hand, make only a limited contribution to uptake of methylammonium in the M range and could therefore represent permeases with a lower affinity for ammonium/methylammonium than Amt1. The Synechocystis amt mutants isolated in this work, including strains that bear only one functional amt gene (strain CSX200: amt1 Ϫ , amt2 ϩ , amt3 Ϫ ; strain CSX201: amt1 Ϫ , amt2 Ϫ , amt3 ϩ ), are however still able to grow using ammonium as a nitrogen source (not shown). It is currently unknown whether diffusion of ammonia or transport of ammonium via Amt2 or Amt3 is responsible for ammoniumdependent growth of those mutants. An impairment in ammonium-dependent growth has only been reported in a strain of S. cerevisiae lacking the three MEP genes present in this orga-nism (11) and in an E. coli amtB mutant when the cells were incubated at pH values below 7 (19).
The regulatory pattern of expression of the amt genes suggests a major role for the Amt permeases in uptake of ammonium present at low concentrations in the extracellular medium. [ 14 C]Methylammonium uptake activity, representing transport plus metabolism via glutamine synthetase, is repressed by ammonium and is maximal in cells starved for nitrogen in CO 2 -enriched air (Fig. 1). Consistently, the three amt genes are preferentially expressed in nitrogen-starved cells (Fig. 6). Expression of amt1, however, is much higher than expression of amt2 or amt3 under any of the tested nitrogen regimes. This resembles the situation in S. cerevisiae where the gene encoding the permease with highest affinity for ammonium, MEP2, is expressed at a much higher level than the MEP1 and MEP3 genes encoding lower affinity permeases (11).
We have further studied the expression of amt1 and have found that transcription of this gene takes place from a promoter that shows the structure of the cyanobacterial NtcAactivated promoters (31). Additionally, a DNA fragment carrying this promoter binds NtcA in vitro (Fig. 8). Like some other genes characterized in Synechococcus sp. PCC 7942 and Anabaena sp. PCC 7120 that are subjected to repression by ammonium (29,42,(55)(56)(57)(58), amt1 appears to belong to the NtcA regulon. A Synechocystis sp. PCC 6803 ntcA mutant is not yet available to confirm NtcA-dependent gene expression in this cyanobacterium. Nonetheless, NtcA-mediated nitrogen control can be important also in Synechocystis sp. PCC 6803, because a number of genes of this strain have been shown to be transcribed from NtcA-type promoters that bind NtcA in vitro. These include, in addition to amtl, icd coding for isocitrate dehydrogenase (59), glnA for glutamine synthetase (60), and glnB for the cellular nitrogen status signaling protein P II (39).
The sizes of the transcripts for the three Synechocystis amt genes (Fig. 6) were analyzed in relation to the DNA sequences surrounding these genes (Ref. 33; see also Fig. 4). The amt1 gene, which is composed of 1521 bp, would be transcribed as a monocistronic mRNA (observed transcript size, 2 kb). For amt2, a gene consisting of 1326 bp, the observed 1.5-kb hybridization band would also correspond to a monocistronic transcript. The amt3 gene, which is composed of 1623 bp, is found downstream of an ORF (sll0536) with which it shows a 4-nucleotide overlap. A transcript containing both ORFs (amt3 and sll0536) would have a size of, at least, 2693 nucleotides, which roughly corresponds to the size of the mRNA detected with the amt3 probe (2.6 kb). Interestingly, sll0536 would encode a polypeptide that shares homology with a putative potassium channel protein of E. coli (accession number P31069). Whether this implies a coordinated function of Amt3 and a potassium channel remains to be investigated.
The Synechocystis Amt proteins as well as their homologues from other biological sources are highly hydrophobic polypeptides that show 10 to 12 putative membrane-spanning regions. They appear to constitute monocomponent permeases whose activity would not depend on ATPase subunits or periplasmicbinding proteins. Consistently, the methylammonium uptake activity of Synechococcus sp. PCC 7942 is preserved in spheroplasts (26). As deduced from data found in currently available data banks, the Amt family would comprise proteins from very diverse biological groups. In addition to the three Amt permeases from Synechocystis sp. PCC 6803, some other members of this protein family are the three MEP proteins from yeast (accession numbers P40260, P41948, and P53390, respectively), AMT1 from A. thaliana (P54144), Amt from C. glutamicum (P54146), A. brasilense (AF005275), A. vinelandii (U91902), and E. coli (P37905), and putative ammonium permeases of B. subtilis (NrgA; Q07429), Mycobacterium tuberculosis (Q10968), Methanococcus jannaschii (Q58739 and Q60366), and Caenorhabditis elegans (P54145). Phylogenetic analyses of these proteins can be found elsewhere (11,17). Interestingly, the three Synechocystis Amt proteins show a higher identity degree to each other (40 to 43% identity) than to any of their homologues from other organisms listed above (27 to 37% identity).