cyAbrB Transcriptional Regulators as Safety Devices To Inhibit Heterocyst Differentiation in Anabaena sp. Strain PCC 7120

Spore formation in Bacillus subtilis and Streptomyces has been extensively studied as models of prokaryotic nonterminal cell differentiation. In these organisms, many cells/hyphae differentiate simultaneously, which is governed by a network in which one regulator stands at the top. Differentiation of heterocysts in Anabaena sp. strain PCC 7120 is unique because it is terminal, and only 5 to 10% of vegetative cells differentiate into heterocysts. In this study, we identified CalA/cyAbrB1 as a repressor of two genes that are essential for heterocyst formation independently of HetR, a master activator for heterocyst differentiation. This finding is reasonable for unique cell differentiation of Anabaena because CalA/cyAbrB1 could suppress heterocyst differentiation tightly in vegetative cells, while only cells in which HetR is overexpressed could differentiate into heterocysts.

ABSTRACT

Cyanobacteria are monophyletic organisms that perform oxygenic photosynthesis. While they exhibit great diversity, they have a common set of genes. However, the essentiality of them for viability has hampered the elucidation of their functions. One example of these genes is cyabrB1 (also known as calA in Anabaena sp. strain PCC 7120), encoding a transcriptional regulator. In the present study, we investigated the function of calA/cyabrB1 in the heterocyst-forming cyanobacterium Anabaena sp. PCC 7120 through CRISPR interference, a method that we recently utilized for the photosynthetic production of a useful chemical in this strain. Conditional knockdown of calA/cyabrB1 in the presence of nitrate resulted in the formation of heterocysts. Two genes, hetP and hepA, which are required for heterocyst formation, were upregulated by calA/cyabrB1 knockdown in the presence of combined nitrogen sources. These genes are known to be induced by HetR, a master regulator of heterocyst formation. hetR was not induced by calA/cyabrB1 knockdown. hetP and hepA were repressed by direct binding of CalA/cyAbrB1 to their promoter regions in a HetR-independent manner. In addition, the overexpression of calA/cyabrB1 abolished heterocyst formation upon nitrogen depletion. Also, knockout of calB/cyabrB2 (a paralogue gene of calA/cyabrB1), in addition to knockdown of calA/cyabrB1, enhanced heterocyst formation in the presence of nitrate, suggesting functional redundancy of cyAbrB proteins. We propose that a balance between amounts of HetR and CalA/cyAbrB1 is a key factor influencing heterocyst differentiation during nitrogen stepdown. We concluded that cyAbrB proteins are essential safety devices that inhibit heterocyst differentiation.

IMPORTANCE Spore formation in Bacillus subtilis and Streptomyces has been extensively studied as models of prokaryotic nonterminal cell differentiation. In these organisms, many cells/hyphae differentiate simultaneously, which is governed by a network in which one regulator stands at the top. Differentiation of heterocysts in Anabaena sp. strain PCC 7120 is unique because it is terminal, and only 5 to 10% of vegetative cells differentiate into heterocysts. In this study, we identified CalA/cyAbrB1 as a repressor of two genes that are essential for heterocyst formation independently of HetR, a master activator for heterocyst differentiation. This finding is reasonable for unique cell differentiation of Anabaena because CalA/cyAbrB1 could suppress heterocyst differentiation tightly in vegetative cells, while only cells in which HetR is overexpressed could differentiate into heterocysts.

INTRODUCTION

Cyanobacteria are ancient and monophyletic prokaryotes, which are characterized by a capacity to perform oxygenic photosynthesis. They are found in diverse habitats, including freshwater, marine water, hot springs, frozen lakes, soil, and deserts (1). Specific responses to environmental changes enable them to adapt to their habitats (2). In addition, they exhibit a great diversity of morphology and cell arrangements. Moreover, some cyanobacteria can differentiate into specific cell types in response to environmental stimuli, which is one type of stress response. The most studied differentiated cell type in cyanobacteria is the heterocyst. At semiregular intervals, some filamentous cyanobacteria can differentiate into larger and round cells called heterocysts, which are cells specialized for nitrogen fixation, which enables heterocystous cyanobacteria to inhabit nitrogen-poor environments.

Anabaena sp. strain PCC 7120 (hereafter Anabaena) has been extensively studied as a model for heterocyst differentiation (3, 4). Upon the depletion of combined nitrogen, 5 to 10% of vegetative cells that perform oxygenic photosynthesis differentiate into heterocysts. A transcriptional regulator, NtcA, widely conserved in cyanobacteria, perceives nitrogen deficiency as an increase of a metabolite, 2-oxoglutarate (5–9). Subsequently, NtcA indirectly induces HetR, a master regulator of heterocyst differentiation (10). Accumulation of HetR spatially initiates a specific developmental program and enables patterned heterocyst formation (11–13). During differentiation, deposition of exopolysaccharide and glycolipid layers results in morphological changes in the cells. In addition, cellular metabolism is dynamically altered by the inactivation of oxygenic photosystem II and enhancement of respiration (3, 4). Such changes enable heterocysts to protect oxygen-labile nitrogenase from oxygen.

Despite their great diversity, cyanobacteria have a core set of genes that are conserved across the phylum (14, 15). Many of the conserved genes have been found to be associated with core biological processes such as DNA replication, transcription, translation, photosynthesis, the Calvin cycle, and various metabolic pathways (15). Therefore, many of their functions can be predicted. However, the functions of some of the conserved genes are yet to be elucidated. Although the study of such genes could offer novel insights into cyanobacterial biology, the essentiality of the core genes (15, 16) has hampered such investigations. An example of these genes is cyabrB1 (17).

cyabrB, encoding a transcriptional regulator, is conserved among cyanobacteria (17). The DNA-binding domain of cyAbrB located at the C terminus belongs to an AbrB-like family (Pfam14250), which is unique to cyanobacteria and is not conserved in other organisms, including melainabacteria, a nonphotosynthetic sister phylum to the cyanobacteria. Each cyanobacterium has at least two copies of cyabrB genes, cyabrB1 and cyabrB2 (also known as calA and calB in Anabaena, respectively [18]; in this article, we refer to each Anabaena cyAbrB gene as calA/cyabrB1 and calB/cyabrB2, respectively, to avoid confusion). Because some attempts to disrupt calA/cyabrB1 in Anabaena and the unicellular cyanobacterium Synechocystis sp. strain PCC 6803 (hereafter Synechocystis) have failed so far (17–20), this gene should be essential for cyanobacteria. Conversely, gene disruption of cyabrB2 is possible in Synechocystis. Studies on cyabrB2 mutants have revealed that cyAbrB2 is involved in acclimation to changes in carbon and nitrogen availability (17, 21–23). While those studies demonstrated that cyabrB2 had specific functions, some evidence suggested a functional overlap between cyAbrB1 and cyAbrB2 (24).

Recent studies toward the photosynthetic production of useful chemicals have rapidly developed tools for artificial gene regulation systems for cyanobacteria (25–27). Among them, a gene knockdown technology, CRISPR interference (CRISPRi), has attracted much attention because this system exhibited repression over a wide dynamic range in an inducer concentration-dependent manner (28–32). Target genes can be repressed by the formation of a complex consisting of a nuclease-deficient Cas9 (dCas9); a single guide RNA (sgRNA), which corresponds to the target DNA sequence; and target DNA (33). While CRISPRi has been successfully applied to enhance the production of desired products in cyanobacteria (28–30, 34), its basic scientific applications are still awaited.

The question of whether a transcriptional regulator, CalA/cyAbrB1, conserved in cyanobacteria regulates core genes or specific genes in Anabaena motivated us to study the function of calA/cyabrB1. In the present study, we created Anabaena strains in which calA/cyabrB1 is conditionally knocked down through CRISPRi technology. CalA/cyAbrB1 amounts were significantly repressed under any condition tested and in any genetic background tested when the CRISPRi system was induced. Repression of calA/cyabrB1 resulted in the formation of heterocysts even in the presence of nitrate. Not hetR but two direct target genes of HetR, hetP and hepA, which are required for heterocyst development (4), were induced by calA/cyabrB1 knockdown in the presence of combined nitrogen in a HetR-independent manner. Overexpression of calA/cyabrB1 abolished heterocyst formation under nitrogen-depleted conditions. Therefore, we concluded that CalA/cyAbrB1 is essential for the suppression of heterocyst differentiation and propose a model where CalA/cyAbrB1 offers HetR an appropriate threshold for the induction of heterocyst development.

RESULTS

Heterocyst formation by calA/cyabrB1 knockdown in the presence of nitrate.

A conditional calA/cyabrB1 knockdown strain, C104, was constructed by integrating a plasmid containing P_petE-tetR, P_L03-dcas9, and P_J23119-sgRNA targeting calA/cyabrB1 (see Fig. S1 in the supplemental material) into the neutral site cyaA (35). In this strain, an inducer, anhydrotetracycline (aTc), derepresses the L03 promoter by binding to TetR, and dcas9 is induced. sgRNA is constitutively expressed. Therefore, the addition of aTc switches on repression by CRISPRi. A negative-control strain, C100, without sgRNA was also constructed.

Strain C104 was grown in nitrate-containing medium and bubbled with air containing 1% (vol/vol) CO₂. When the expression of calA/cyabrB1 was repressed for 48 h, 0.6% heterocysts were formed (Fig. S2A and Table S1). In contrast, when calA/cyabrB1 was not repressed, few heterocysts (<0.1%) were formed. Heterocyst frequencies of the control strain C100 and the wild-type strain were not changed by the inducer (<0.1%). To facilitate clearer observation of the phenotype, C104 was cultured in nitrate-containing medium and bubbled with air containing 5% (vol/vol) CO₂, in which carbon is in excess of nitrogen in the cells. Under these conditions, the wild-type strain, C104, and C100 formed 0.6, 0.5, and 0.4% heterocysts, respectively, in the absence of the inducer. In the presence of the inducer, C104 formed 2.9% heterocysts, but the wild-type strain and C100 did not form any heterocysts (Fig. 1A and Table S1) (<0.1%). These results indicate that calA/cyabrB1 downregulation stimulates heterocyst formation. Knockdown of CalA/cyAbrB1 was confirmed by Western blot analysis (Fig. 1B and C). While the addition of aTc to control strain C100 did not repress CalA/cyAbrB1 expression, in the case of C104, the addition of aTc considerably repressed CalA/cyAbrB1 expression (less than 10%), particularly at 48 and 72 h. Repression of calA/cyabrB1 did not lead to heterocyst formation when C104 was cultured in ammonium-containing medium (Fig. S2B). During 5 days of calA/cyabrB1 repression, C104 hardly showed a growth defect. However, when C104 cells that had been cultivated for 5 days with aTc were transferred into fresh medium containing aTc, the growth of C104 cells was significantly inhibited, while no growth defect was observed when the cells were transferred into fresh medium without aTc (data not shown). Control strain C100 showed no growth defect under the same experimental conditions, confirming that calA/cyabrB1 is essential for the viability of Anabaena. We focused on the reason why heterocysts were formed by knockdown of calA/cyabrB1 even in the presence of nitrate in the present study.

FIG 1.

Promotion of heterocyst formation by knockdown of calA/cyabrB1 in the presence of nitrate. A control strain lacking sgRNA (C100) and a calA/cyabrB1 knockdown strain (C104) were cultured in nitrate-containing medium in the absence or presence of the inducer aTc (50 ng/ml). Cultures were bubbled with air containing 5% (vol/vol) CO₂. (A) Formation of heterocysts by calA/cyabrB1 knockdown in the presence of nitrate. Images were photographed after 48 h of cultivation. Arrowheads indicate heterocysts. (B) Confirmation of CalA/cyAbrB1 knockdown. After 48 h of cultivation, total protein was extracted, and Western blotting using anti-cyAbrB1 antibody was performed. Different amounts of total proteins from C100 cultured in the absence of aTc were loaded to show the linearity of the results. (C) Time course analysis of CalA/cyAbrB1 knockdown. Each strain was cultivated for the indicated times in the presence of 50 ng/ml aTc. Subsequently, total protein was extracted, and Western blotting using anti-cyAbrB1 antibody was performed.

To rule out the possibility that the observed phenotype is due to off-target effects of CRISPRi, we constructed C105 and C106, which retain sgRNAs targeting different sites of calA/cyabrB1. The formation of heterocysts was promoted in both C105 and C106 by calA/cyabrB1 repression, similarly to C104, when cells were cultured in nitrate-containing medium and bubbled with air containing 5% (vol/vol) CO₂ (Fig. S3). These results suggest that CalA/cyAbrB1 is required for the suppression of heterocyst formation in the presence of nitrate.

Repression of hetP and hepA by CalA/cyAbrB1 in the presence of nitrogen sources.

To elucidate the cause of heterocyst formation in the presence of nitrate, RNA was extracted from cells of C100 and C104 cultured for 48 h in nitrate-containing medium with 5% CO₂ in the absence or presence of the inducer, and reverse transcription-quantitative PCR (RT-qPCR) analysis was performed. The expression of four genes that are induced at early stages of heterocyst development, hetR, hetP, hetZ, and hepA, was investigated (Fig. 2A). hetR encodes a master regulator of heterocyst differentiation, and its overexpression causes heterocyst formation in the presence of nitrogen sources (13). Both hetP and hetZ are directly upregulated by HetR, and ectopic expression of either gene leads to heterocyst formation in the presence of nitrogen sources (4, 36–40). hepA encodes a component of an ABC transporter required for the construction of the heterocyst exopolysaccharide layer (41), which is the first step during morphological differentiation (4). HetR also directly induces the expression of hepA (4, 42). While the expression of hetR and hetZ did not change significantly, the expression of hetP and hepA was significantly induced when calA/cyabrB1 was repressed in C104. In C100, the expression of hetP and hepA was not changed by the addition of aTc (Fig. 2A), indicating that aTc itself and the expression of dCas9 had no effects on the expression of these genes. Similar results were observed for C105 and C106 (Fig. 2A). We confirmed that the calA/cyabrB1 transcript and the CalA/cyAbrB1 protein in C104, C105, and C106 were repressed in the presence of the inducer (Fig. S4A and B, respectively).

FIG 2.

Upregulation of some genes related to heterocyst differentiation following knockdown of calA/cyabrB1. After 48 h of cultivation in the absence or presence of 50 ng/ml aTc, RNA was extracted, and RT-qPCR was performed. rnpB was used for normalization. Data represent means ± standard deviations (SD) (n = 3 from independent cultures). Amounts of each gene relative to those in C100 cells grown in nitrate-containing medium bubbled with air containing 5% (vol/vol) CO₂ in the absence of the inducer are shown. (A) Each strain was grown in nitrate-containing medium and bubbled with air containing 5% (vol/vol) CO₂. (B) Strain C104 was grown in nitrate-containing medium bubbled with air containing 1% (vol/vol) CO₂ or in ammonium-containing medium bubbled with air containing 5% (vol/vol) CO₂.

Subsequently, we extracted RNA from C104 cultured in the presence of nitrate with 1% CO₂ bubbled or in the presence of ammonium with 5% CO₂ bubbled (Fig. 2B). Although no heterocysts were formed under the latter condition, hetP and hepA were greatly induced under both conditions when calA/cyabrB1 was repressed, similar to what is shown in Fig. 2A. We confirmed that the calA/cyabrB1 transcript and the CalA/cyAbrB1 protein were repressed under both conditions in the presence of the inducer (Fig. S4C and D, respectively). The results indicate that the upregulation of hetP and hepA by calA/cyabrB1 knockdown induced heterocyst formation, rather than hetP and hepA being induced following the initiation of heterocyst development, and that heterocyst formation was suppressed by an unknown mechanism in the presence of ammonium.

The expression of nifH, encoding a subunit of nitrogenase, was quantified to determine whether the heterocysts formed by calA/cyabrB1 repression were mature. Repression of calA/cyabrB1 in C104 in the presence of nitrate bubbled with 5% CO₂ induced nifH but not under conditions of bubbling with 1% CO₂ (Fig. S5), suggesting that maturation of heterocysts depends on the C-N balance inside the cells and does not directly depend on CalA/cyAbrB1. Therefore, we concluded that CalA/cyAbrB1 is essential for repression of hetP and hepA in the presence of nitrogen sources.

Induction of hetP and hepA by calA/cyabrB1 knockdown independently of HetR.

To clarify whether the upregulation of hetP and hepA is independent of or dependent on HetR, we constructed a strain, C104h, in which calA/cyabrB1 could be knocked down through the CRISPRi system in the hetR mutant background (43) (Fig. S1). RNA was extracted from C104h cells cultured in the presence of nitrate bubbled with 5% CO₂ in the absence or presence of aTc. RT-qPCR analysis revealed that hetP and hepA were upregulated by calA/cyabrB1 repression in the hetR-deficient background (Fig. 3A), indicating that CalA/cyAbrB1 regulates the expression of the two genes independently of HetR. We confirmed that the calA/cyabrB1 transcript and the CalA/cyAbrB1 protein in C104h were repressed similarly to those in C104 in the presence of the inducer (Fig. 3A and B).

FIG 3.

Knockdown of calA/cyabrB1 in the hetR mutant background. Cells were cultured in nitrate-containing medium bubbled with air containing 5% (vol/vol) CO₂ in the absence or presence of aTc for 48 h. (A) Effect of hetR inactivation on expression of calA/cyabrB1 and genes related to heterocyst formation. RNA was extracted, and RT-qPCR was performed. rnpB was used for normalization. Data represent means ± SD (n = 3 from independent cultures). Amounts of each gene relative to those in C100 cells cultured in nitrate-containing medium bubbled with air containing 5% (vol/vol) CO₂ in the absence of the inducer are shown. (B) Confirmation of CalA/cyAbrB1 knockdown. Total protein was extracted, and Western blotting using anti-cyAbrB1 antibody was performed. (C) No heterocyst formation by calA/cyabrB1 knockdown in a hetR-deficient background. Cells of C104h were microphotographed.

Heterocysts were not formed following the repression of calA/cyabrB1 in C104h (Fig. 3C). Although the expression of hetP in C104h was induced by calA/cyabrB1 in the presence of aTc compared to that in the absence of the inducer, the expression levels of hetP in the presence of the inducer in C104h were low compared to those in C104 (Fig. 2A and Fig. 3A), possibly due to the effect of hetR disruption. This result could explain why heterocysts were not formed even when hetP and hepA were induced in C104h (Fig. 3C).

Specificity and redundancy of cyAbrB proteins.

Subsequently, we constructed a calB/cyabrB2 knockout mutant, DR2080, and a calA/cyabrB1 knockdown, calB/cyabrB2 knockout mutant, C104B2 (Fig. S1), to examine the specificity and redundancy of CalA/cyAbrB1 and CalB/cyAbrB2. In C104B2, calA/cyabrB1 knockdown caused 4.5% heterocyst formation in nitrate-containing medium bubbled with 1% CO₂ (Fig. 4A and Table S1). In contrast, the calA/cyabrB1 knockdown mutant C104 and the calB/cyabrB2 knockout mutant DR2080 produced only 0.6 and 0% heterocysts under similar conditions, respectively (Fig. 4A and Fig. S2A). RT-qPCR analysis revealed that hetP and hepA were similarly induced by calA/cyabrB1 knockdown in C104 and C104B2 (compare Fig. 2B and Fig. 4B). The expressions of hetR and hetZ were not induced by calA/cyabrB1 knockdown in C104B2, similarly to C104 (Fig. 4B). Deletion of calB/cyabrB2 hardly influenced the expression of hetR, hetP, hetZ, and hepA. Repression of the calA/cyabrB1 transcript and the CalA/cyAbrB1 protein in C104B2 was confirmed (Fig. S6A and B). C104B2 did not form heterocysts in ammonium-containing medium (Fig. S6C). Comparison of the results with calA/cyabrB1 knockdown, calB/cyabrB2 knockout, and cyabrB1 knockdown/cyabrB2 knockout mutants revealed that CalA/cyAbrB1 but not CalB/cyAbrB2 specifically regulates the expression of hetP and hepA. However, with regard to heterocyst formation in the presence of nitrate, CalA/cyAbrB1 and CalB/cyAbrB2 could be redundant since the double mutant produced more heterocysts than did the single mutants.

FIG 4.

Specific and redundant functions of cyabrB genes. Cells were cultured in nitrate-containing medium bubbled with air containing 1% (vol/vol) CO₂ in the absence or presence of aTc for 48 h. (A) Formation of heterocysts by calA/cyabrB1 knockdown and calB/cyabrB2 knockout. Cells of C104B2 or DR2080 were microphotographed. (B) RNA was extracted, and RT-qPCR was performed. rnpB was used for normalization. Data represent the means ± SD (n = 3 from independent cultures). Amounts of each gene relative to those in C100 cells grown in nitrate-containing medium bubbled with air containing 5% (vol/vol) CO₂ in the absence of the inducer are shown.

Direct binding of CalA/cyAbrB1 to promoters of hetP and hepA.

To test whether the expression of hetP and hepA was directly regulated by CalA/cyAbrB1, we expressed a recombinant His-CalA/cyAbrB1 protein in Escherichia coli and purified it. The purified His-CalA/cyAbrB1 protein had an apparent molecular weight of 18,000, which was largely consistent with the theoretical value (Fig. 5A). We performed a gel mobility shift assay using His-CalA/cyAbrB1 and the Cy3-labeled DNA probe PhetP, which includes the hetP promoter region (Fig. 5B). His-CalA/cyAbrB1 retarded the mobility of the probe. Thereafter, we examined the specificity of the interaction using a competition assay. The addition of a 5- or 10-fold molar excess of nonlabeled DNA probes PhetP and PhepA (hepA promoter region) eliminated the retardation, but that of cyabrB1RT (internal region of calA/cyabrB1) did not (Fig. 5B). These results indicate that CalA/cyAbrB1 binds promoter regions of hetP and hepA and that the two genes are directly repressed by CalA/cyAbrB1.

FIG 5.

Direct binding of CalA/cyAbrB1 to promoters of hetP and hepA. (A) Purification of His-CalA/cyAbrB1. Purified His-CalA/cyAbrB1 was subjected to 15% SDS-PAGE. Lane M, protein molecular weight marker; lane His-CalA/cyAbrB1, purified His-CalA/cyAbrB1. (B) Electrophoretic mobility shift assay with His-CalA/cyAbrB1. His-CalA/cyAbrB1 was mixed with 3 nM DNA probe (promoter region of hetP). Nonlabeled DNAs (PhetP, PhepA, or the internal region of calA/cyabrB1 [cyabrB1RT]) were added.

Inhibition of heterocyst formation by overexpression of calA/cyabrB1.

We investigated whether or how CalA/cyAbrB1 participates in heterocyst development in the absence of nitrogen sources. calA/cyabrB1 was knocked down upon removal of combined nitrogen sources in C104. After 24 or 48 h of nitrogen stepdown, heterocysts were formed in the strain in the absence or presence of aTc (Fig. S7A and B). Vegetative cell intervals between heterocysts after 48 h of nitrogen stepdown were not significantly affected by the presence or absence of the inducer (Fig. S7C). Figure S7D confirms that CalA/cyAbrB1 was repressed in the presence of the inducer in C104 at 48 h in the absence of nitrogen sources. Gradual repression of CalA/cyAbrB1 (Fig. 1C) might make it difficult to observe a clear phenotype.

Next, we constructed a calA/cyabrB1 overexpression strain, T121 (Fig. S1). In this strain, aTc induced the expression of calA/cyabrB1 driven by P_L03. Strain C100, in which aTc induced the expression of dcas9 but not calA/cyabrB1, was used as the control strain. C100 or T121 was transferred from a nitrate-containing medium to a nitrogen-free medium and grown in the absence or presence of aTc. While C100 grew regardless of the absence or presence of aTc, T121 did not grow at all in the presence of the inducer (Fig. 6A). In contrast, the overexpression of calA/cyabrB1 only minimally inhibited the growth of T121 in the presence of nitrate, as previously demonstrated (44). Microscopic observations revealed that the addition of aTc to T121 abolished heterocyst formation 24 h after nitrogen depletion (Fig. 6B). Expression levels of hetR, hetP, and hepA were measured after depletion of nitrogen sources for 8 h, at which time the genes were upregulated in the wild-type strain (45). While the expression of hetP was not repressed, the expression of hepA was repressed in T121 in the presence of aTc compared to that in the absence of aTc or in C100 in the absence or presence of aTc (Fig. 6C). Expression of hetR was also repressed in T121 following the addition of aTc, although it was minimal compared to the expression of hepA, for an unknown reason. We confirmed that the addition of aTc led to the accumulation of calA/cyabrB1 transcripts and CalA/cyAbrB1 protein (Fig. 6C). The results suggested that overproduction of CalA/cyAbrB1 inhibited the transcription of hepA. The reason why the expression of hetP was not inhibited is discussed below.

FIG 6.

Effect of calA/cyabrB1 overexpression on heterocyst development. Cells of the negative-control strain C100 and the calA/cyabrB1 overexpression strain T121 were cultured in the absence or presence of 200 ng/ml aTc. (A) Impaired growth under nitrogen-depleted conditions by calA/cyabrB1 overexpression. Cells were cultured in nitrogen-free or nitrate-containing medium, and the OD₇₅₀ was monitored. Data represent means ± SD (n = 3 from independent cultures). (B) Effect of calA/cyabrB1 overexpression on heterocyst formation. Cells of C100 and T121 cultured for 24 h in the absence or presence of the inducer were microphotographed. (C) Effect of calA/cyabrB1 overexpression on expression of genes related to heterocyst formation. After 8 h of nitrogen depletion, RNA was extracted, and RT-qPCR was performed. rnpB was used for normalization. Data represent means ± SD (n = 3 from independent culture). Amounts of each gene relative to those in C100 cells grown in nitrate-free medium in the absence of the inducer are shown. Overexpression of CalA/cyAbrB1 was confirmed by Western blotting.

We quantified the amounts of CalA/cyAbrB1 after nitrogen depletion in the wild-type strain. Cells of the wild-type strain were washed with nitrogen-free medium, transferred to nitrate-containing or nitrogen-free medium, and cultured for 8 h. Total proteins were extracted from cells before and after their cultivation. Western blotting using anti-cyAbrB1 antibody revealed that the depletion of nitrogen sources did not alter the amounts of CalA/cyAbrB1 (Fig. S8A and B). We also quantified the amounts of CalA/cyAbrB1 in mature heterocysts after nitrogen depletion for 24 h relative to those in whole filaments. The amounts of CalA/cyAbrB1 were marginally smaller in heterocysts (Fig. S8C and D). These results suggested that high-level induction of HetR in proheterocysts (11), rather than decreasing CalA/cyAbrB1 amounts, was a limiting step in the upregulation of hetP and hepA during heterocyst development in the wild-type strain following combined nitrogen stepdown.

Effect of calA/cyabrB1 repression or overexpression on expression of alr0947.

It has previously been shown that calA/cyabrB1 and alr0947 constitute an operon (44). Therefore, we evaluated the polar effect of calA/cyabrB1 repression. While the expression of calA/cyabrB1 was highly repressed following the addition of aTc in C104, C105, and C106 under some conditions (Fig. S4A and C), the expression of alr0947 was only slightly repressed (Fig. S9A and B). It was demonstrated that the overexpression of CalA/cyAbrB1 repressed the expression of alr0947, since CalA/cyAbrB1 inhibits the transcription of the calA/cyabrB1-alr0947 operon (44). When calA/cyabrB1 was overexpressed, alr0497 was repressed marginally (Fig. S9). While both repression and induction of calA/cyabrB1 resulted in a slight decrease in alr0947 expression, repression induced heterocyst formation, and induction inhibited heterocyst formation, indicating that the observed phenotypes in the present study were caused by changes in expression levels of calA/cyabrB1 rather than those of alr0947.

DISCUSSION

In the present study, we revealed that CalA/cyAbrB1 is essential for the repression of hetP and hepA in the presence of nitrogen sources and that heterocysts are formed even in the presence of nitrate when calA/cyabrB1 is knocked down. Since the first description of cyabrB1 in cyanobacteria (19), its essentiality has hampered investigation of the function of the gene. We overcame this challenge by applying CRISPRi, which is a recently developed technology that has been employed to facilitate photosynthetic production of desired chemicals in some model cyanobacteria (28–30). Here, we demonstrated that this technology is also very useful for basic research. Under all the conditions tested, and in any genetic background, CRISPRi facilitated the robust knockdown of calA/cyabrB1. The inhibition of heterocyst formation by the overexpression of calA/cyabrB1 in the absence of nitrogen sources has been overlooked so far. Agervald et al. (44) used the nirA promoter, which overexpresses only in the presence of nitrate, while He and Xu (20) used the petE promoter, which might not be strong enough to hinder heterocyst formation. Consequently, the development of a variety of tools to regulate gene expression (27) is fundamental and could broaden the range of basic research and biotechnological applications.

Balance between CalA/cyAbrB1 and HetR determines heterocyst differentiation.

Some evidence indicated that CalA/cyAbrB1 hinders heterocyst formation in the presence of nitrogen sources by repressing the expression of hetP and hepA through specific binding to promoter regions of the genes, which are direct targets of HetR. The possibility that the upregulation of the genes by calA/cyabrB1 knockdown is due to an increase in HetR protein or its activity (e.g., decreased level of PatS, a HetR inhibitor [46]) should be rejected based on the results that the effects of calA/cyabrB1 knockdown on gene expression are independent of HetR (Fig. 3A) and that expression of hetZ, a direct target of HetR (38), was not induced following calA/cyabrB1 knockdown (Fig. 2). Upon nitrogen stepdown, amounts of CalA/cyAbrB1 did not significantly decrease (see Fig. S8 in the supplemental material). We propose that strong induction of HetR in proheterocysts (11) outcompetes the repression by CalA/cyAbrB1 during heterocyst differentiation in the wild-type strain. Indeed, the overexpression of CalA/cyAbrB1 hampered the induction of hepA but not hetP. This observation could be explained by the fact that the binding affinity of HetR for the hetP promoter is much higher than that for the hepA promoter (42). Decreased CalA/cyAbrB1 activity attributed to glutathionylation (47) in proheterocysts is a potential reason. However, this possibility is less likely because overexpression of calA/cyabrB1 resulted in repression of hepA under nitrogen-depleted conditions. Otherwise, amounts of CalA/cyAbrB1 might significantly and transiently decrease in proheterocysts. Taken together, these data lead us to propose that a balance between the amounts of CalA/cyAbrB1 and HetR is a key determinant for the initiation of heterocyst differentiation, although further evidence should be required with respect to hetP induction.

It is typical that both a global activator and a repressor regulate prokaryotic cell differentiation. In Streptomyces, a global transcriptional activator, AdpA, and repressor, BldD, control morphological differentiation (48, 49), while in Bacillus subtilis, an activator, Spo0A, and a repressor, AbrB, regulate spore formation (50). Notably, both the DNA-binding domain of cyAbrB (AbrB-like family [Pfam14250]) and that of AbrB from B. subtilis (MazE_antitoxin family [Pfam04014]) belong to the same clan (AbrB [CL0132]), although the former is located at the C terminus and the latter is located at the N terminus. However, the relationship between the activator and the repressor is different between Anabaena and B. subtilis as well as Streptomyces. The expression of AdpA is regulated by BldD in Streptomyces (49, 51), and the expression of AbrB is regulated by Spo0A in B. subtilis (50). In contrast, our results demonstrated that HetR does not regulate CalA/cyAbrB1 and vice versa. Many cells/hyphae differentiate in B. subtilis and Streptomyces simultaneously. Hence, the fact that one regulator governs the whole network is a practical strategy for cell differentiation in these organisms. In contrast, only 1/10 of cells differentiate into heterocysts, and the remaining vegetative cells maintain viability and photosynthetic activity in Anabaena. In addition, heterocyst differentiation is terminal (nonreversible), while cell differentiation in B. subtilis and Streptomyces is nonterminal (reversible). Therefore, robust inhibition of heterocyst differentiation by CalA/cyAbrB1, whose functioning is independent of HetR, could be an essential safety device in Anabaena, in concert with the HetR inhibitors PatS (46) and HetN (52, 53) and regulation by HetR phosphorylation (54).

Repression of heterocyst formation in ammonium medium.

While hetP was induced in the presence of both ammonium and nitrate by calA/cyabrB1 knockdown in C104, heterocysts were formed only under the latter conditions. These results suggest the existence of an unidentified mechanism that regulates heterocyst differentiation because a previous study demonstrated that the overexpression of hetP induced heterocyst formation even in the presence of ammonium (37). The inconsistency between our results and those of the previous study could be explained by the unidentified mechanism, which would be CalA/cyAbrB1 dependent.

Perspectives.

In the present study, we could not identify target genes of CalA/cyAbrB1 other than hetP and hepA, which are not conserved in many nonheterocystous cyanobacteria. Upregulation of the genes in a CalA/cyAbrB1 knockdown mutant could not explain why calA/cyabrB1 is essential in Anabaena. A transcriptome analysis would identify other target genes comprehensively and would answer the question of whether cyAbrB1, conserved in cyanobacteria, regulates core genes (14, 15). If CalA/cyAbrB1 regulated core genes, whether CalA/cyAbrB1 is involved in the reconstruction of metabolism during heterocyst development, such as inactivation of photosystem II (55) or enhancement of photosystem I (56), would be an interesting question.

Our results demonstrated that CalA/cyAbrB1, but not CalB/cyAbrB2, specifically regulates the expression of hetP and hepA. However, heterocyst formation in the presence of nitrate was enhanced in the double mutant C104B2 compared to that in the single mutant C104 (Fig. 4). These results indicate that cyAbrB proteins can function both specifically and redundantly/cooperatively in Anabaena, as has been suggested for Synechocystis (24), although the underlying mechanism remains to be elucidated.

It has been demonstrated that cyAbrB2 of Synechocystis functions in concert with a global nitrogen regulator, NtcA (17), whose activity is regulated by 2-oxoglutarate (5–9), and the inorganic carbon regulators CmpR and NdhR (21, 23). Nitrogen sources (ammonium or nitrate) and/or the CO₂ concentration should affect the 2-oxoglutarate concentration (57) and NtcA activity. However, irrespective of nitrogen sources and CO₂ concentrations, hetP and hepA were similarly induced by calA/cyabrB1 knockdown (Fig. 2), indicating that the regulation of hetP and hepA by CalA/cyAbrB1 is independent of NtcA (and/or 2-oxoglutarate) and PacR, a regulator of inorganic carbon assimilation in Anabaena (58). In the future, it should be addressed whether another target gene(s) of CalA/cyAbrB1 and target genes of CalB/cyAbrB2 are coregulated by NtcA or PacR in Anabaena by a transcriptome analysis of C104B2 and C104. Moreover, comparison of transcriptome data for cyabrB knockdown/knockout mutants of Anabaena and Synechocystis would shed more light on the evolution and adaptation of cyanobacteria.

MATERIALS AND METHODS

Bacterial strains and growth conditions.

Anabaena strains were cultured at 30°C under 30 to 35 μmol photons m⁻² s⁻¹ in BG11 medium (59) (17.6 mM sodium nitrate as the nitrogen source). BG11₀ (lacking nitrogen sources) or BG11_a (5 mM ammonium chloride as the nitrogen source) medium was used after washing the cells twice with BG11₀ medium, where indicated. Each medium was supplemented with 20 mM HEPES-NaOH (pH 7.5). Two micrograms per milliliter each of spectinomycin and streptomycin and/or 25 μg/ml neomycin sulfate was added when required. Liquid cultures were bubbled with air containing 1.0% (vol/vol) CO₂ unless otherwise stated. Expression of dcas9 or calA/cyabrB1 was induced by the addition of aTc when cells were inoculated.

Plasmid construction.

Plasmids for the knockdown of calA/cyabrB1 by CRISPRi or the overexpression of calA/cyabrB1 were constructed using the hot fusion method (60). DNA fragments were inserted between the BamHI and KpnI sites of a genome-integrating vector, pSU102-cyaA (30). Schematic representations of them are shown in Fig. S1, and detailed sequences are described in the supplemental material. Inactivation of calB/cyabrB2 was accomplished by replacing a 200-bp portion of the cyabrB2 coding region with a spectinomycin resistance cassette, as follows. Upstream and downstream regions of the calB/cyabrB2 gene were amplified by PCR using the primer pair 2080-5F and 2080-5R and the primer pair 2080-3F and 2080-3R, respectively. The spectinomycin cassette was inserted between upstream and downstream regions, and the resultant construct was cloned between SacI and XhoI sites of pRL271 (61) to construct pR2080S. To construct the expression plasmids for the hexahistidine-tagged CalA/cyAbrB1 protein, a DNA fragment containing the calA/cyabrB1 coding regions was amplified by PCR using the primer pair 0946-F and 0946-R. The amplified DNA fragment was cloned between NdeI and BamHI sites of the pET-28a expression vector (EMD Millipore) to construct pEcyAbrB1.

RNA extraction and RT-qPCR analysis.

Extraction and purification of total RNA, synthesis of cDNA, and qPCR were performed as described previously (62). Primers used in qPCR are listed in Table 1. The significance of the difference was assessed by Student’s t tests (P ≤ 0.05).

TABLE 1.

Primers used in this study

^aEMSA, electrophoretic mobility shift assay.

Western blot analysis.

Extraction of total protein and Western blotting were performed as described previously (62). CalA/cyAbrB1 was detected using a rabbit polyclonal antibody raised against His-cyAbrB1 from Synechocystis (24).

Enrichment of heterocysts.

The enrichment of heterocysts from Anabaena filaments was performed as described previously (30).

Expression and purification of His-CalA/cyAbrB1.

E. coli BL21(DE3) harboring pEcyAbrB1 was grown at 37°C in 250 ml of Luria-Bertani medium. The recombinant gene was expressed in exponentially growing cells (optical density at 600 nm [OD₆₀₀] of 0.6) by adding 1 mM isopropyl-β-d-thiogalactopyranoside. After 5 h of incubation, the cells were harvested by centrifugation. His-CalA/cyAbrB1 was purified with the Ni-nitrilotriacetic acid (NTA) Fast Start kit (Qiagen). The elution fractions containing the purified protein were loaded onto a PD MidiTrap G-25 column (GE Healthcare) equilibrated with 20 mM phosphate buffer (pH 7.4) containing 0.5 M NaCl and 10% glycerol, and the protein was eluted with the same buffer.

Gel mobility shift assay.

The hetP promoter region was amplified by PCR using primer pair PhetP-F and PhetP-R and cloned into the EcoRV site of PBluescript II KS⁺ to construct pBPhetP. A Cy3-labeled probe, PhetP, was prepared by PCR using the Cy3-labeled M13-F primer and PhetP-R, with pBPhetP as a template. His-CalA/cyAbrB1 was incubated with a Cy3-labeled probe (3 nM) in 20 μl of incubation buffer (20 mM Tris-HCl [pH 7.5], 10 mM MgCl₂, 1 mM dithiothreitol, 40 ng/μl bovine serum albumin [BSA], and 5% glycerol) for 30 min at room temperature. The mixtures were subjected to electrophoresis on a native 5% polyacrylamide gel, and Cy3-labeled probes were detected on an FLA-9000 imaging system (Fuji Film). The nonlabeled DNA probes PhetP, PhepA, and cyabrB1RT were prepared by PCR using primer pair PhetP-F and PhetP-R, primer pair PhepA-F and PhepA-R, and primer pair RTcyabrB1-F and RTcyabrB1-R, respectively.