Transcription of hupSL in Anabaena variabilis ATCC 29413 Is Regulated by NtcA and Not by Hydrogen

Abstract

Nitrogen-fixing cyanobacteria such as Anabaena variabilis ATCC 29413 use an uptake hydrogenase, encoded by hupSL, to recycle hydrogen gas that is produced as an obligate by-product of nitrogen fixation. The regulation of hupSL in A. variabilis is likely to differ from that of the closely related Anabaena sp. strain PCC 7120 because A. variabilis lacks the excision element-mediated regulation that characterizes hupSL regulation in strain PCC 7120. An analysis of the hupSL transcript in a nitrogenase mutant of A. variabilis that does not produce any detectable hydrogen indicated that neither nitrogen fixation nor hydrogen gas was required for the induction of hupSL. Furthermore, exogenous addition of hydrogen gas did not stimulate hupSL transcription. Transcriptional reporter constructs indicated that the accumulation of hupSL transcript after nitrogen step-down was restricted primarily to the microaerobic heterocysts. Anoxic conditions were not sufficient to induce hupSL transcription. The induction of hupSL after nitrogen step-down was reduced in a mutant in the global nitrogen regulator NtcA, but was not reduced in a mutant unable to form heterocysts. A consensus NtcA-binding site was identified upstream of hupSL, and NtcA was found to bind to this region. Thus, while neither hydrogen gas nor anoxia controlled the expression of hupSL, its expression was controlled by NtcA. Heterocyst differentiation was not required for hupSL induction in response to nitrogen step-down, but heterocyst-localized cues may add an additional level of regulation to hupSL.

Cyanobacteria may possess up to three different enzymes with hydrogenase activity: NiFe uptake hydrogenase, NiFe bidirectional hydrogenase, and nitrogenase (32). Nitrogenases produce hydrogen gas as an obligate by-product of the reduction of nitrogen to ammonia (32). To recapture the electrons used to make this hydrogen, most (but not all [19]) nitrogen-fixing cyanobacteria have an uptake hydrogenase encoded by the hup genes (30, 31). This uptake hydrogenase comprises a small and a large subunit (hupS and hupL, respectively) and contains an NiFe cofactor. The uptake hydrogenase captures much of the hydrogen that is produced by nitrogenase, and eliminating this enzyme results in a threefold increase in hydrogen production in heterocyst-forming cyanobacteria (11, 30, 42).

Because hydrogen is produced under N₂-fixing conditions in cyanobacteria, hupSL transcription is coordinately regulated with nitrogenase. In Anabaena variabilis ATCC 29413 (11), Anabaena sp. strain PCC 7120, Nostoc punctiforme ATCC 73102 (17), and Gloeothece sp. strain ATCC 27152 (23), the transcription of hupSL increases in response to nitrogen step-down. In a medium lacking a source of fixed nitrogen, approximately 10% of the cells in some filamentous cyanobacteria differentiate into microaerobic cells called heterocysts, where the oxygen-labile nitrogenase functions under aerobic growth conditions. The transcription of hupSL in these cyanobacteria has been reported to occur concomitantly with the differentiation of heterocysts (11). The expression of hupSL in Anabaena sp. strain PCC 7120 is restricted to heterocysts because there is an excision element in hupL that is removed during heterocyst development (4). However, other species, such as A. variabilis, lack this excision element, so the role of heterocyst differentiation in the regulation of hupSL is unknown.

Factors that may control hupSL in A. variabilis include hydrogen, nitrogen status or other heterocyst development signals, anoxia (15), and Ni availability (1). In several cyanobacterial species, uptake hydrogenase activity increases in the presence of hydrogen. In A. variabilis, exposure to hydrogen resulted in a slight transient increase in uptake hydrogenase activity in whole cells (39). Uptake hydrogenase activity in Nostoc muscorum and Anabaena cylindrica was up-regulated in response to hydrogen, but only after exposure for several days to the elevated hydrogen concentrations (34). In Anabaena sp. strain PCC 7120 and N. punctiforme ATCC 73102, the addition of hydrogen to aerobically grown cells caused an increase in uptake hydrogenase activity (14, 24). The addition of hydrogen had no effect on the reversible hydrogenase activity of Anabaena sp. strain PCC 7120 (14).

The induction of hupSL transcription by hydrogen in some bacteria is controlled by a hydrogen-sensing and the transcriptional regulation system, HupUV, HupT, and HupR (7). While the sequenced genomes of cyanobacteria lack genes similar to this sensory hydrogenase, transcriptional up-regulation of hupSL by hydrogen has been reported for Nostoc muscorum and Nostoc punctiforme (1).

In addition to regulation by hydrogen, nitrogen status may control the transcription of hydrogenase genes in nitrogen-fixing species. The promoters of several cyanobacterial uptake hydrogenase genes contain the consensus sequence for binding the global regulator of nitrogen metabolism, NtcA, suggesting potential regulation by nitrogen status cues (16, 17, 23).

In an attempt to further define the signals controlling hupSL expression in A. variabilis, we tested the effect of hydrogen, anoxia, heterocyst differentiation, and NtcA on transcription of hupSL. We demonstrated that neither hydrogen nor anoxic conditions affected the expression of hupSL. However, NtcA, but not heterocyst differentiation, per se, was required for the induction of hupSL after nitrogen step-down.

MATERIALS AND METHODS

Bacterial strains and growth conditions.

A. variabilis strain FD was derived from A. variabilis ATCC 29413 by selecting for increased growth at 40°C (6). Strain Avm13 was kindly provided by the laboratory of Thomas Happe (11). All cyanobacterial strains were grown photoautotrophically on agar-solidified Allen and Arnon medium (AA) or in a liquid medium composed of an eightfold dilution of AA (AA/8) (35) either with or without 5.0 mM NH₄Cl and 10 mM N-Tris(hydroxymethyl)methyl-2-aminoethanesulfonic acid (TES; pH 7.2), unless otherwise noted. Cultures were routinely grown at 30°C with shaking (200 rpm) and illuminated with cool white fluorescent bulbs at 50 to 80 μE m⁻² s⁻¹. The following antibiotics were added to cyanobacterial strains at the indicated concentrations: neomycin (5 μg/ml for liquid, 40 μg/ml for agar plates), spectinomycin (0.3 μg/ml for liquid, 3 μg/ml for agar plates), or streptomycin (0.3 μg/ml for liquid, 3 μg/ml for agar plates).

In experiments requiring cells to be shifted from AA/8 supplemented with NH₄/TES to AA/8 lacking fixed nitrogen, cells were washed three times in AA/8. The anaerobic induction was performed on cells grown for several transfers in medium containing 5 mM fructose. Cells were washed as described above to remove fixed nitrogen from the medium and resuspended in AA/8 containing 5 mM fructose and 10 μM 3-(3,4-dichlorophenyl)-1,1-dimethylurea. Cultures were grown in the light and were bubbled vigorously with nitrogen or argon gas to remove trace amounts of oxygen. Cells exposed to hydrogen were grown in 250 ml AA/8 in a 500-ml flask bubbled with air alone or air augmented with 10% hydrogen gas. Samples were withdrawn at the indicated time points.

Cultures of Escherichia coli were grown overnight in Luria-Bertani (LB) liquid medium or LB agar medium supplemented, when appropriate, with ampicillin (50 μg/ml), kanamycin (50 μg/ml), chloramphenicol (25 μg/ml), or spectinomycin (50 μg/ml). Cloning techniques were performed according to protocols described previously in reference 28.

Construction of plasmid pBP283.

To make a hupL::lacZ fusion, a 5.2-kb SalI-to-EcoRI fragment containing hupSL was removed from pAAWY14832, a clone used for sequencing the A. variabilis genome by the Joint Genome Institute. This fragment was ligated into pBR322, digested with the same enzymes (replacing part of Tet^r), and renamed pBP273. After being digested with EcoRV, pBP273 was ligated with a 1.9-kb SmaI-digested fragment containing the aadA gene from pRL5801 and named pBP275. To make pBP283, a 5-kb fragment, containing the promoterless lacZ gene from Escherichia coli and the neomycin/kanamycin resistance gene C.K3 (Nm^r), was removed from pPE20 by using SmaI (9). This 5-kb fragment was ligated to HpaI-digested pBP275 and was named pBP283. This plasmid was conjugated into A. variabilis strains FD and NF76, selecting for single recombination, to make strains BP283/FD and BP283/NF76, respectively.

Hydrogen and acetylene reduction assays.

For the detection of hydrogen, 1 ml of culture was injected into a 10-ml serum vial, sealed with an air-tight rubber stopper that had been sparged with argon gas, shaken at 30°C with illumination as described above, and sampled after 1 h. Samples from the headspace (1 ml) were injected into a 5890 series II (HP) gas chromatograph fitted with a Carboxen 1010 Plot fused-silica capillary column (30 m by 0.53 mm; Sigma), and hydrogen was detected by thermal conductivity using argon as the carrier gas. Acetylene reduction assays were performed as described previously in reference 21.

RNA extractions.

Cells were grown as described above and centrifuged, and all liquid medium was removed. Cell pellets were flash frozen with liquid nitrogen and stored at −80°C until extraction. Extractions were performed using TRIzol (Sigma) as described previously (25). DNA was removed using the Turbo DNA-Free kit (Ambion) following the manufacturer's instructions. RNA was assayed for traces of DNA by performing PCR as described below for reverse transcription-PCR (RT-PCR) but using Taq as the sole polymerase with RnpB-L and RnpB-R primers (Table 1).

TABLE 1.

Bacterial strains, plasmids, and primers used in this work

Strain, plasmid, or primer	Description	Reference or source
Bacterial strains
A. variabilis ATCC 29413	Wild type	C. P. Wolk
NF76	Point mutant unable to form heterocysts	5
NF76/BP283	lacZ-Nmr inserted into hupL by single crossover	This work
A. variabilis ATCC 29413 strain FD	Wild-type strain selected for growth at 40°C	6
Avm13	Nmr gene replacing hupS-hupL genes	11
MM3	Nmr gene interrupting the ntcA gene	38
JE21	Nmr gene replacing the nifUII-nifDII region of the nifII genes	37
JE9	Nmr gene replacing the xisA-nifEI region of the nifI genes	37
BP283/FD	lacZ-Nmr inserted into hupL by single crossover	This work
Plasmid
pBP283	pBR322-derived vector with lacZ-Nmr inserted into hupL	This work
Primers
qNifH-2L	5′-TTAGCCAAGAAAATCATCAACAAC
qNifH-2R	5′-CTGCGTGCTTAGAATCATCATC
qHupSL-2L	5′-CCCCCGCTACTTCTATGGATTGTT
qHupSL-2R	5′-GTCACGCCGAGTTCATAGGATTTG
qRnpB-2L	5′-AGCAAGGCCGAAGGAACTATGGT
qRnpB-2R	5′-ATTGCTTTACACGAGGGCGATTAT
HupSRT-L2	5′-TTAATGCCCCCAACGGTACA
HupSRT-R2	5′-GCGGTTGCAGGAAGAATGAG
NifH1-L	5′-CGGCATGACCTATTGGTAGC
NifH1234-R	5′-GGTGARATGATGGCGATGTAYGC
NifH2-L	5′-TAGACTACTACTGCTTTGCC
RnpB-L	5′-AGAGTAGGCGTTGGCGGTTGC
RnpB-R	5′-ATTGCTTTACACGAGGGCGATTAT
HupSupntcA-L	5′-ATATAAGCGTGGCTACAACTAACCAA
HupSdownntca-L	5′-CTGGCAGGATCAAAAAAGCCTCA
HupSntcAtest-R	5′-GACGCTTGTAGAGGTTCACCTGT
HC1	5′-TAGTACATCGGTGAGGGGTG
HC2	5′-TGTGAGCAACATCGACATCTG
HupSL-RRT	5′-TCGCATACTGTCGGTTCTTC
HupSL/R	5′-ATATACCCGGGTAGTTTTAGTGGCATTGGCTATTAGTTTGTATATTG
PE	5′-CTAGTACTCCGCTATTGCGGTACCCTTGTACGCCTGTTTTATA
P1	5′-GGTATTGCGGTACCCTTGTAC

Semiquantitative and QRT-PCR.

Semiquantitative RT-PCR was performed as described previously in reference 25. Primers were as follows: for hupSL, HupSRT-L2 and HupSRT-R2; for nifH1, NifH1-L and NifH1234-R; for nifH2, NifH2-L and NifH1234R; and for rnpB, RnpB-L and RnpB-R. Quantitative RT-PCR (QRT-PCR) was performed using the Superscript III Platinum one-step QRT-PCR kit with ROX (Invitrogen) in 20-μl reaction mixtures according to the manufacturer's recommended protocol. Primers (qNifH-2L and qNifH-2R, qHupSL-2L and qHupSL-2R, and qRnpB-2L and qRnpB-2R) were used at a final concentration of 150 nM each. Assays were performed on a Prism 7700 (ABI) using the following thermal cycle: 55°C for 10 min, 95°C for 5 min, 95°C for 10 s, 58°C for 30 s, 72°C for 30 s, and 40 cycles of steps 3 to 5. The QRT-PCR cycle was immediately followed by a melting curve: 95°C for 15 s, 60°C for 20 s, ramp to 95°C over 19:59 min with data collection, and a final 95°C hold for 15 s. Triplicate reactions were performed for all samples subjected to QRT-PCR. Reaction efficiencies were approximately 1.8, as measured by regression (26). Thus, approximate differences can be calculated by raising the efficiency, 1.8, to the difference in threshold cycle (C_T) value.

β-Galactosidase-mediated in situ localization.

In situ localization was performed essentially as described previously (38). Briefly, cells were fixed in 0.01% glutaraldehyde, washed, and exposed to C-12-fluorescein-β-d-galactoside (C₁₂-FDG, a substrate that fluoresces after cleavage by β-galactosidase). Filaments were viewed with an epifluorescence microscope (Zeiss) fitted with a fluorescein filter set and a short-pass filter to block biliprotein fluorescence (excitation, 450 to 490 nm; dichroic mirror, 510 nm; barrier filter, 520 nm). Images were acquired using a Retiga EXi (QImaging) cooled, charge-coupled device camera with IP Labs 4.0 software (BD Biosciences). Exposure time was 1.0 s for fluorescence pictures and about 0.05 s for bright-field images.

RNA ligase-mediated RT-PCR.

The transcription start site of hupSL was determined essentially as described previously by reference 2. Briefly, total RNA was extracted from cultures of A. variabilis strain FD grown in AA/8 in the absence of fixed nitrogen, and DNA was removed as described above. RNA (5 to 10 μg) was treated with 25 U tobacco acid phosphatase (Epicentre Biotechnologies) according to manufacturer's instructions or was left untreated as a control. RNA was extracted with 1 volume of phenol-chloroform and precipitated with ethanol. The RNA oligonucleotide PE (Table 1) was ligated to the free 5′ ends of RNA with T4 RNA ligase (120 U; New England Biolabs). The ligated RNA was extracted and precipitated again as described above, and the RNA was reverse transcribed by using 1 pmol of the primer hupSL-RRT and Superscript III (100 U; Invitrogen) at 50°C for 1 h according to the manufacturer's instructions. DNA from the reverse transcribed (2-μl) reaction mixture was amplified by PCR using the nested primers P1 and hupSL/R (Table 1) in the cycle of 95°C for 5 min, 95°C for 30 s, 56°C for 30 s, and 72°C for 60 s and 30 cycles of steps 3 to 5. A small band (ca. 200 bp) observed only in samples treated with tobacco acid phosphatase was excised from the gel, cleaned (PCR cleanup kit; Qiagen), and cloned using the pCR2.1 TA cloning kit (Invitrogen) according to the manufacturer's instructions. Plasmid from five colonies was purified and sequenced using the M13 forward and reverse primers to determine the 5′ transcription start site.

NtcA-binding assay.

NtcA-binding reactions were carried out essentially as described previously (20). PCR products of the 431-bp and 301-bp hupSL upstream regions of A. variabilis (created using the primer hupSntcAtest-R with hupSupntcA-L and hupSdownntcA-L, respectively), a 282-bp PCR product of the hetC promoter region of Anabaena sp. strain PCC 7120 (using primers HC1 and HC2 [described previously in reference 22]) or a ca. 500-bp PCR product of the rnpB gene from A. variabilis (obtained with the rnpB-L and rnpB-R primers [Table 1]) was end labeled with [γ-³²P]ATP and incubated for 30 min at room temperature with native purified NtcA (1.25 to 3.75 ng/μl final concentration) in 15 μl binding buffer (HEPES-NaOH, 12 mM; Tris-HCl, 4 mM; KCl, 60 mM; EDTA, 1 mM; dithiothreitol, 1 mM; pH 8.0) that contained salmon sperm DNA (0.05 mg/ml), 8% glycerol, 2-oxoglutarate (0.6 mM), and bovine serum albumin (0.05 mg/ml). The DNA fragments were separated on a 6% polyacrylamide gel. Some experiments included the addition of 20 ng (25-fold excess) unlableled PCR product.

RESULTS

Analysis of hupSL transcription.

In the filamentous cyanobacterium A. variabilis ATCC 29413, hupSL expression is up-regulated during the differentiation of heterocysts after nitrogen step-down (11). The expression of hupSL began to increase at 12 and 14 h after nitrogen step-down (Fig. 1). The timing of the increase in mRNA agrees with the induction of uptake hydrogenase activity reported elsewhere (11, 39). The expression of nifH1 followed a pattern similar to that of hupSL, with transcript increasing between 12 and 14 h after nitrogen step-down (Fig. 1).

FIG. 1.

Induction of hupSL and nifH1 after nitrogen step-down. Time course for hupSL, nifH1, and rnpB transcription measured by RT-PCR for 30 h after nitrogen step-down. Hours after nitrogen step-down are indicated above each lane.

Effect of hydrogen on hupSL transcription.

To understand whether the induction of hupSL transcript is controlled by heterocyst-specific developmental cues or is regulated by the hydrogen made by nitrogenase, we examined hupSL expression by real-time QRT-PCR in the strain JE9. This mutant, in which part of the nifD1 excision element and the 3′ end of nifD1 were replaced by an Nm^r gene, lacks the Mo-nitrogenase and thus produces no hydrogen (Fig. 2B and C) (36). Hydrogen produced from other nitrogenases (the anaerobically induced Nif2 nitrogenase and the Mo-repressed V nitrogenase) was not observed, because these alternative nitrogenases are not expressed under aerobic conditions in the presence of Mo (35, 37). While strain JE9 lacks nitrogenase activity and dies after 2 or 3 days due to nitrogen starvation (Fig. 2A), it still differentiates heterocysts (36). Despite the lack of nitrogenase activity and its concomitant hydrogen production, hupSL transcript increased in abundance in strain JE9 (Fig. 2D). This finding suggests that hydrogen was not required for the induction of hupSL transcription and that in A. variabilis, other cues are involved in hupSL regulation.

FIG. 2.

Hydrogen production by A. variabilis. Growth curves (A), acetylene reduction (B), and hydrogen production (C) of strains FD (diamonds) and JE9 (squares) during 48 h after nitrogen step-down. (D) Expression of hupSL measured by QRT-PCR in two strains (FD [gray bars] and JE9 [black bars]) expressed as change in C_T value from 0 to 24 h or 0 to 30 h after nitrogen step-down. Thus, each unit represents an approximately twofold difference in the amount of RNA (see Materials and Methods). Error bars indicate standard deviations from the mean for at least three independent experiments. OD₇₂₀, optical density at 720 nm.

While hydrogen is not required to activate hupSL expression in A. variabilis, hupSL expression might still be modified by hydrogen. To test whether hupSL expression is affected by hydrogen, we exposed A. variabilis grown in AA/8 to air augmented with 10% hydrogen and measured the abundance of hupSL transcript by QRT-PCR. At 8 h, the average change in C_T value relative to a t value of 0 for hupSL was 0.22 (±0.19) or 0.38 (±0.10) for cells exposed to air alone or to air plus 10% hydrogen, respectively. At 24 h, the average change in C_T value relative to a t value of 0 for hupSL was 1.44 (±0.63) or 1.22 (±0.85) for cells exposed to air alone or to air plus 10% hydrogen, respectively. Thus, after 8 h or 24 h of exposure to H₂, no increase in hupSL transcript was observed that could be attributed to the presence of hydrogen.

Effect of anoxia on hupSL transcription.

The microaerobic conditions in heterocysts are not established until late in development; hence, the timing of the induction of hupSL and nifH1 transcription might reflect a response to the microaerobic conditions in the heterocyst. Thus, growth under anaerobic conditions in the absence of fixed nitrogen might induce hupSL. Cells were washed to remove ammonium from the growth medium and then exposed to either anaerobic or aerobic conditions. Anaerobic conditions were confirmed by acetylene reduction assay at 6 h, showing that the anaerobically induced Nif2 system was active (data not shown) (36). Aerobic cultures had no acetylene reduction activity at 6 h. The transcription of hupSL in anaerobic cultures did not increase compared to that in aerobic cultures at either 8 or 24 h after nitrogen step-down (Fig. 3). As a control for anaerobic induction, nifH2 was found to be expressed strongly at 8 h, with slightly less transcript present at 24 h after nitrogen step-down. Thus, oxygen status does not appear to be the heterocyst-linked cue regulating the transcription of hupSL.

FIG. 3.

Effect of anoxia on expression of hupSL. Expression of hupSL, nifH1, and nifH2 under aerobic and anaerobic conditions 8 and 24 h after nitrogen step-down. Cells were grown in the presence of NH₄-TES (5 mM) and fructose (5 mM), washed, and exposed to aerobic or anaerobic conditions. Samples were harvested at the indicated times for RT-PCR analysis. Lane 1, aerobic conditions at 8 h after nitrogen step-down; lane 2, anaerobic conditions at 8 h after nitrogen step-down; lane 3, aerobic conditions at 24 h after nitrogen step-down; lane 4; anaerobic conditions at 24 h after nitrogen step-down 4. Experiments were repeated at least three times, and representative figures are shown.

Effect of heterocysts and NtcA on hupSL transcription.

To test whether hupSL expression is controlled by heterocyst developmental cues, two mutant strains of A. variabilis that are unable to form heterocysts were tested for hupSL expression. The first strain, MM3, bears a mutation in ntcA, a gene involved in the global regulation of nitrogen status in cyanobacteria (38). Strain MM3 does not differentiate heterocysts and has the delayed phycobilisome degradation following the removal of fixed nitrogen from the growth medium that is typical of ntcA mutants (29). Transcripts of hupSL and nifH1 did not accumulate in strain MM3 24 h after nitrogen step-down (Fig. 4A). The low-level transcription of hupSL observed when cells are grown in the presence of ammonium was not affected in the ntcA mutant (data not shown). Thus, it appears that either NtcA or the downstream signals initiated by this protein during heterocyst differentiation are required for hupSL induction in response to nitrogen step-down.

FIG. 4.

(A) Induction of hupSL and nifH1 in heterocyst-deficient mutants. The amount of transcript of hupSL (gray bars) and nifH1 (black bars) in FD and mutants NF76 and MM3 was measured by QRT-PCR 24 h after nitrogen step-down. Error bars indicate standard deviations from the mean for at least three independent experiments. (B) Diagram showing insertion of lacZ:Nm^r into hupL in the plasmid pBP283. (C and D) In situ localization of β-galactosidase in BP283/FD (C) and BP283/NF76 (D). Cells were grown in AA/8 supplemented with NH₄-TES (5 mM), washed, and resuspended in AA/8. Cells were imaged 24 h after nitrogen step-down after incubation with C₁₂-FDG as described in Materials and Methods. Images at left and right are bright-field and epifluorescence microscopy, respectively. Arrowheads indicate heterocysts.

The second mutant strain, NF76, is an uncharacterized mutant that does not form heterocysts or fix nitrogen (5). NF76 has a functional ntcA gene, however, since it can use nitrate as a nitrogen source (data not shown), while ntcA mutants cannot (41). In contrast to the results for the ntcA mutant, hupSL was induced after nitrogen step-down in strain NF76 (Fig. 4A). The induction of nifH1 was also observed in NF76 but was reduced approximately 15-fold compared to the induction of the wild type (Fig. 4A). Despite the induction of nifH1, no acetylene reduction activity was measured for NF76 24 h after nitrogen step-down (data not shown). Thus, mature heterocysts are not essential for hupSL transcription, but either NtcA or some signal derived from NtcA activity is required for hupSL induction.

In situ localization of hupSL expression.

To determine in which cells the induction of hupSL occurred in the NF76 mutant, reporter strains with lacZ inserted within hupL (plasmid pBP283) (Fig. 4B) were constructed. Spatial distribution of β-galactosidase activity along the filament was determined by observing the fluorescence from the intracellular cleavage of the substrate C₁₂-FDG by β-galactosidase. In strain BP283/FD, fluorescence was strongest in heterocysts (Fig. 4C). β-Galactosidase activity was also observed at a much lower level in the vegetative cells and is consistent with reports of hupSL expression in these cells (3, 39). Fluorescence in strain BP283/NF76 was observed throughout the vegetative cell filaments, and no localization to any spaced, cryptic proheterocyst cells was observed (Fig. 4D).

Binding of NtcA to the hupSL promoter.

The regulation of hupSL by NtcA suggested that an NtcA-binding site might be present in the promoter. We identified a consensus NtcA-binding site 427 bp upstream of the transcriptional start site (TSS) reported by Happe et al. (11) (Fig. 5C). The ability of the NtcA protein to bind to this site was tested using two PCR products: a 431-bp product that spanned the putative binding site in the hupSL promoter and a 301-bp product that contained the hupSL promoter but lacked the binding site (Fig. 5C). NtcA was found to bind the fragment of the hupSL promoter that included the NtcA site. The presence of this site resulted in much stronger binding compared to that of the shorter fragment of the hupSL promoter lacking the NtcA site (Fig. 5A). The addition of the unlabeled product that included the NtcA site in 25-fold excess reduced the binding of the labeled fragment (Fig. 5B), and a negative control fragment from the rnpB gene did not bind to the protein (Fig. 5B). Because NtcA-binding sites are typically, but not exclusively, located approximately −40 bp before the TSS (13), we investigated whether an alternative TSS was present. By using RNA ligase-mediated RT-PCR, we identified one TSS 77 bp upstream of the TSS reported by Happe et al. (11) (Fig. 5C), but there was no evidence of a TSS near the NtcA-binding site.

FIG. 5.

NtcA binding to the hupSL upstream region. (A) Lanes 1 and 2, a 431-bp region upstream of hupSL containing a consensus NtcA-binding site incubated without (lane 1) or with (lane 2) NtcA (1.25 ng/μl). Lanes 3 and 4, a 301-bp region upstream of hupSL lacking a consensus NtcA-binding site incubated without (lane 3) or with (lane 4) NtcA. Lanes 5 and 6, a 282-bp region upstream of hetC containing a strong NtcA-binding site (22) incubated without (lane 5) or with (lane 6) NtcA. (B) Lanes 1 to 5, 431-bp region upstream of hupSL containing a consensus NtcA-binding site incubated with the following: no protein (lane 1), 2.5 ng/μl NtcA (lane 2), 3.125 ng/μl NtcA (lane 3), 3.75 ng/μl NtcA (lane 4), and 3.75 ng/μl NtcA with 25-fold cold competitor 431-bp PCR product (lane 5). Lanes 6 to 9, 500-bp fragment of the rnpB gene incubated with the following: no protein (lane 6), 3.125 ng/μl NtcA (lane 7), 3.75 ng/μl NtcA (lane 8), and 3.75 ng/μl NtcA with 25-fold cold competitor rnpB PCR product (lane 9). (C) Upstream region of hupSL. Shown are primers used to amplify the products for the NtcA-binding assay (underlined), a putative NtcA-binding site (dashed box), the TSS published previously by Happe et al. (11) (solid circle), a newly identified TSS (dashed circle), and a consensus −10 site corresponding to the new TSS (gray box). Line numbers indicate nucleotides before the beginning of the coding region. Bold uppercase letters indicate the hupSL open reading frame.

DISCUSSION

Among cyanobacterial strains, the regulatory cue signaling hupSL induction varies considerably (31, 33). For example, Anabaena sp. strain PCC 7120 contains an excision element in hupL that is removed during heterocyst maturation (4), while hupSL in Gloeothece sp. strain ATCC 27152 is regulated by the global nitrogen regulator NtcA (23). Because hupSL regulatory signals vary by species and had not been characterized for A. variabilis, we examined several potential hupSL regulatory cues, such as hydrogen, anoxia, heterocyst differentiation, and the global nitrogen regulator NtcA.

The induction of hupSL transcription in response to nitrogen step-down, but not the basal level of expression observed in cells grown with ammonium, depended on a functional ntcA gene, indicating that NtcA directly, or indirectly via heterocyst differentiation, controls the increased transcription of hupSL after nitrogen step-down. The induction of hupSL transcript to the same degree in both wild-type cells and the heterocyst-defective mutant NF76 suggested that heterocyst differentiation was not essential for hupSL induction. The mutation in NF76 is unknown; however, the strain grows with nitrate, indicating a functional NtcA. The induction of hupSL in NF76 might be due to an increase in spaced cells destined for heterocyst differentiation; however, the expression of β-galactosidase in all cells in the BP283/NF76 strain suggested that the increase in hupSL transcript occurred in all cells and not in cryptic proheterocysts. Thus, the increase in hupSL transcription, on a per cell basis, was much lower in NF76 than in the wild-type strain, suggesting that additional regulation from a heterocyst-related cue is necessary for the full induction observed in heterocysts. One possibility is that the mutated gene in NF76 may play a role in vegetative cell gene repression during the response to nitrogen step-down.

We identified an NtcA-binding site in the hupSL promoter 428 bp upstream of the hupSL TSS published by Happe et al. (Fig. 5C) (11). This binding site bears the consensus sequence (TGT-N_9or10-ACA) shown to be present in other genes controlled by NtcA, such as xisA, glnA, and rbcL (27). Considerable variation in NtcA-binding sequences has been observed, but the minimal consensus (GT-N₁₀-AC) fits the sequence found upstream of A. variabilis hupSL.

Binding sites for NtcA in genes positively regulated by NtcA are typically found −40 bp upstream of a TSS, suggesting the possibility that an alternate, NtcA-regulated TSS may exist for hupSL (12). We examined this possibility through RNA-ligase-mediated RT-PCR in cells grown without fixed nitrogen. Our finding of a TSS 77 bp upstream of the previously reported TSS may be a result of the different methods used to obtain the TSS. While Happe et al. employed 5′ extension, our method depended upon an intact 5′ triphosphate to which an RNA linker could be ligated and thus selected for an intact 5′ sequence (2). Despite the identification of a new TSS, the consensus NtcA site remains 350 bp upstream of the new TSS. NtcA binding sites at distances greater than the prototypical distance of −40 bp have been demonstrated for several cyanobacterial genes, such as cox2, cox3, and furA (13, 18).

The binding of NtcA to the hupSL promoter region was shown to be much stronger in a fragment of the promoter that included the NtcA consensus binding site than in a promoter fragment that lacked this site (Fig. 5C). This finding, when combined with the absence of hupSL induction in the ntcA mutant, suggests that NtcA positively regulates hupSL. Some binding to the smaller fragment that did not contain the consensus NtcA binding site was observed and may indicate that a second nonconsensus NtcA binding site may also be present closer to the TSS. Because uptake hydrogenases function to recapture hydrogen produced during nitrogen fixation, our finding that NtcA controls hupSL expression in A. variabilis is consistent with the role of NtcA in nitrogen acquisition.

The transcription of nifH in Nostoc sp. strain PCC 7120 after nitrogen step-down was detected by Northern analysis between 18 and 24 h (10) or as early as 12 h using more sensitive RT-PCR (40). Consistent with this result, we observed the induction of nifH1 transcription between 12 and 14 h after nitrogen step-down in A. variabilis. The transcription of hupSL in A. variabilis began at the same time and perhaps even slightly earlier than nifH1 (Fig. 1). This further supports the idea that hydrogen produced by nitrogenase is not required for the induction of hupSL, because the induction occurs before nitrogenase can produce hydrogen. Thus, hupSL belongs in the same category of “late development” genes as does nifH1 in the heterocyst differentiation process.

A partial fumarate nitrate-reductase regulator site was previously identified in the promoter of hupSL, pointing to anoxia as a potential regulator of transcription (11). However, anoxic conditions did not affect hupSL transcription. It has been reported that the transcription of nifH1 could not be induced by anaerobic conditions in vegetative cells (8, 37). Thus, the anaerobic conditions found in the heterocyst are not sufficient to induce hupSL expression. In addition, no difference was observed when the ratio of hydrogen production to acetylene reduction was compared after anaerobic induction in wild-type FD and AVM13, a hupSL mutant strain (data not shown). This result suggests that the normal basal expression of hupSL in vegetative cells is not part of a strategy to allow the rapid production of uptake hydrogenase for the anaerobically induced Nif2 nitrogenase that is produced in vegetative cells of this strain.

The control of hupSL in the absence of fixed nitrogen is regulated by NtcA, a transcriptional regulator that controls cyanobacterial genes involved in nitrogen metabolism. While it is reasonable to suggest that hupSL transcription in A. variabilis might be regulated by hydrogen, our studies indicate that this is not the case. The nature of the reported hydrogen-dependent increases in hydrogenase activity is a subject for future study (1).