Anabaena 7120 cultures were grown in BG-11 liquid medium, pH 7.2, with combined nitrogen (17 mM NaNO₃) under continuous illumination (30 μEm^–2 s^–1) without or with shaking (100 rpm) at 27°C ± 2°C (Allen, 1968). Escherichia coli cultures were grown in Luria–Bertani (LB) medium at 37°C (DH5α, HB101) or 30°C (SHuffle) with shaking at 120 rpm. The neomycin antibiotic was used for recombinant Anabaena cultures in BG-11 liquid medium (15 μg ml^–1) or BG-11 agar plates (25 μg ml^–1), whereas chloramphenicol (34 μg ml^–1) or carbenicillin (100 μg ml^–1) was used for E. coli cultures. Primers, plasmids, E. coli, and Anabaena strains used in this study are listed in Table 1.

Alignment of DNA and protein sequences was determined using ClustalW (Thompson et al., 1994) and Clustal Omega (Madeira et al., 2019), respectively. Jalview was used to visualize and edit aligned protein sequences (Waterhouse et al., 2009). A phylogenetic tree was constructed using MEGA version X (Kumar et al., 2018) by the maximum likelihood method. The I-TASSER software was used to predict the tertiary structure of AzuR and metal-binding residues (Zhang, 2008; Roy et al., 2010; Yang et al., 2015). Pattern search analysis of conserved sequences was carried out using the online tool Pattern Locator (Mrázek and Xie, 2006). The −10 and −35 boxes of the upstream region of nmtA were predicted from BPROM (Salamov and Solovyevand, 2011).

The azuR ORF (363 bp) was PCR amplified from Anabaena 7120 genomic DNA and cloned into pTwin1 vector at NdeI–SapI sites. The resulting construct pTwinazuR was confirmed by sequencing and transformed into an E. coli SHuffle strain. Overexpression of chitin-binding domain (CBD)-tagged AzuR was induced by the addition of 0.5 mM IPTG. The protein purification was carried out by chitin affinity chromatography as per the manufacturer’s protocol (New England Biolabs). The protein was cleaved from its tag and eluted following incubation with 40 mM DTT at 4°C for 3 days. CBD was also eluted as the contaminating protein. This eluate was loaded onto the fresh chitin resin after DTT removal. The flow-through was collected, which contained purified Anabaena AzuR without CBD. The purified protein band following electrophoresis on 15% SDS-PAGE was excised and processed for LC-MS/MS analysis (Q Exactive Plus BioPharma High-Resolution Orbitrap MS system, Thermo Fischer Scientific) at the Sophisticated Analytical Instrument Facility (SAIF), IIT Bombay, India. Spectrum was acquired in positive ion mode in a mass range from 350 to 2,000 m/z. The resultant spectrum was used for peptide identification using the Anabaena 7120 protein database available at UniProt.

Determination of the oligomeric status of AzuR was done by glutaraldehyde cross-linking of protein in the native state. Purified AzuR was incubated with 10 mM glutaraldehyde at room temperature (RT) for 10–15 min in 10 mM Tris, pH 7.5. To this, a cracking buffer without or with DTT (50 mM) was added. The resulting cross-linked protein was analyzed by 15% SDS-PAGE. The native molecular mass of AzuR was determined by size-exclusion chromatography (AKTA FPLC system, GE Healthcare) using the GE Superdex 75 column equilibrated with 20 mM Tris, 100 mM NaCl, pH 7.5 at 25°C at a flow rate of 0.5 ml min^–1. The column was previously calibrated using a set of gel filtration markers [bovine serum albumin (66 kDa), ovalbumin (44 kDa), carbonic anhydrase (44.3 kDa), and cytochrome c (29 kDa)] (GE Healthcare).

Analysis of the secondary structure of AzuR was performed by circular dichroism (CD) spectroscopy (MOS-500 Biologic CD spectrometer equipped with a Peltier-type thermostatic cell holder) at 25°C. The CD spectrum was recorded in the wavelength range of 200–260 nm using a cuvette with a path length cell of 0.1 mm. The samples were prepared in 10 mM Tris buffer, pH 7.5. The alpha helical content was calculated using the online tool K2D2 (Perez-Iratxeta and Andrade-Navarro, 2008). CD spectra were also recorded for titrations of AzuR with increasing concentrations of zinc (molar equivalents ranging from 1 to 10).

Total RNA was isolated from Anabaena 7120 treated with 10 μM cadmium for 1 h as described earlier (Divya et al., 2018). cDNA was synthesized with 0.5 μg of total RNA using ReadyScript cDNA Synthesis Mix (Sigma-Aldrich). Following dA tailing of cDNA by terminal transferase (Roche), PCR was performed with the oligo(dT)-anchor primer and nmtA primer as listed in Table 1. The PCR product was then sequenced.

The putative promoter region (100 bp DNA sequence upstream of nmtA ORF) was PCR amplified (primers listed in Table 1) and end-labeled with DIG-ddUTP as per manufacturer’s instructions (Roche). Two nanograms of a DIG-labeled probe (P_nmtA) was incubated with various concentrations of AzuR protein in a total reaction volume of 20 μl containing 20 mM Tris–Cl (pH 7.5) and 1 mM EDTA at RT for 30 min. The DNA–protein complexes were resolved on 10% native PAGE in 0.5× TBE. Separated complexes were electroblotted onto a nylon membrane, cross-linked with UV, and stored at 4°C. It was probed with an anti-DIG antibody and developed using a colorimetric substrate, NBT-BCIP, according to the manufacturer’s protocol (DIG High Prime DNA Labeling and Detection Starter Kit I, Roche). The bands were quantified by the ImageJ software, and the data were fitted to Hill’s equation. Each experiment was repeated three times. In order to evaluate the specificity of interaction of the DNA-AzuR protein binding, electrophoretic mobility shift assay (EMSA) was performed with 100 ng of AzuR (360 nM) with either 20 ng of P_nmtA or 20 ng of non-specific DNA (nmtA gene). For protein specificity, 20 ng of P_nmtA with non-specific proteins like AnLexA, BSA, or NmtA, each at 360 nM concentration, was taken for EMSA. The DNA–protein complexes were resolved on 10% native PAGE and visualized by ethidium bromide staining. To evaluate whether different divalent metal ions affect the binding of AzuR to the target 100 bp DNA, EMSAs were carried out in the presence of 100 μM of various metals. The metal salts used in the study were ZnSO₄.7H₂O, CdCl₂.1/2H₂O, CuSO₄.5H₂O, Co(NO₃)₂.6H₂O, NiSO₄.7H₂O, MnCl₂.4H₂O, and Pb(NO₃)₂. EMSAs were also carried out in the presence of 1 mM DTT (Erbe et al., 1995) for reactions containing all the aforesaid metals.

Overexpression of azuR gene in its native strain was achieved by triparental conjugation (Divya et al., 2018). The azuR gene was cloned downstream to the light-inducible psbA1 promoter in the pFPN vector at NdeI and BamHI sites. A SalI–XmaI fragment from pFPNazuR was excised and cloned into the E. coli/Anabaena shuttle vector pAM1956 upstream of the promoterless gfpmut2 gene. pAMazuR was then transferred into Anabaena 7120. The recombinant Anabaena strain was designated as AnazuR⁺. In a similar way, AnnmtA⁺ (Anabaena strain overexpressing NmtA) was also generated. AnpsbA⁺ (Anabaena harboring pAM1956 with constitutive expression of GFP) was generated by excising the P_psbA1 fragment from the pFPN vector and cloning it into the vector pAM1956 upstream of the promoterless gfpmut2 gene and transferred conjugally into Anabaena 7120. The recombinant Anabaena strains were repeatedly subcultured and maintained under the selective pressure of neomycin (Nm¹⁵). Visualization of GFP fluorescence in the recombinant cells confirmed the expression of the azuR gene placed upstream of the gfpmut2 gene.

For RT-PCR, 1 μg RNA was used for cDNA synthesis (ReadyScript cDNA Synthesis Mix, Sigma-Aldrich). RT-PCR was carried out with azuR-specific primers (Table 1) with 16S rRNA serving as the internal control. RT-PCR products were resolved by electrophoresis on 1% agarose gel and detected by staining with ethidium bromide. For quantification of nmtA transcripts, real-time PCR was performed with nmtA-specific primers in Qiagen rotor-Gene Q real-time PCR cycler. 16S rRNA was used as the internal control.

Exponential phase cultures (3-day-old cultures) of AnpsbA⁺, AnnmtA⁺, and AnazuR⁺ were inoculated in BG-11 N⁺ (Nm¹⁵) liquid medium at a chlorophyll a (Chla) density of ∼4 μg ml^–1 and incubated for 10 days under illumination without or with cadmium at 10 and 20 μM concentrations. Growth was assessed by measuring Chla content at regular intervals. For spot assays, exponentially growing cultures of AnpsbA⁺, AnnmtA⁺, and AnazuR⁺ were spotted onto BG-11 N⁺ (Nm²⁵) agar plates without or with cadmium (10, 20, and 40 μM) at the chlorophyll density mentioned in the figure and incubated under continuous illumination for 7 days.

Bright-light and fluorescence microscopy (FM) images were taken at ×600/×1,500 magnification on a Carl Zeiss Axioscope 40 microscope with a charge coupled device (CCD) AxioCam MRc camera (Zeiss). Green fluorescence of GFP was visualized using a Hg-arc lamp (excitation BP: 450–490 nm, emission LP: 515 nm). Chla fluorescence of Anabaena was visualized with green light excitation (excitation BP: 546/12, emission LP: 590 nm). It should be noted here that the microscopic settings for GFP fluorescence used the emission filter (λ_emission: 515 nm) that could detect both GFP and Chla fluorescence. For scanning electron microscopy (SEM), exponential-phase cells of WT, AnpsbA⁺, AnnmtA⁺, and AnazuR⁺ were harvested by centrifugation, and the resulting cell pellets were washed with 0.9% NaCl and fixed with 2.5% glutaraldehyde at 4°C for 1–2 h. Post fixation, the cells were serially dehydrated in 20, 30, 50, 70, 90, and 100% ethanol. The dehydrated sample was then gold coated with a sputtering device (Q 150R ES, Quorum) and visualized using SEM (EVO 18 Research, Carl Zeiss, United Kingdom).

Growth experiments were repeated three times. Average values with standard deviations are shown for a representative experiment. For determination of cell size, data are represented as average values ± standard deviation. One-way ANOVA was employed for calculating the significance of the difference in cell size between WT, AnpsbA⁺, AnnmtA⁺, and AnazuR⁺ cultures.

The genome of Anabaena PCC 7120 harbors two proteins belonging to the ArsR-SmtB family of proteins, All7621 and Alr0831. The ArsR-SmtB family of transcriptional metalloregulators represses the expression of genes/operons involved in maintaining metal homeostasis or toxic metal detoxification (Osman and Cavet, 2010). Among the 15 characterized metal binding motifs (Saha et al., 2017), the metal-sensing members of the regulators include two structurally diverse metal-binding sites, namely, α3N, and α5 (Busenlehner et al., 2003). All7621 in Anabaena 7120 encodes for AztR, a regulator of AztA [Zn(II)/Pb(II) CPx-ATPase efflux pump] (Liu et al., 2005), and belongs to the α3N group of proteins. The α3N site consists of cysteine thiolate ligands—two from the α3 helix with signature motifs Cx_1–2C or Cx₂CD and one or two cysteine ligands derived from the amino-terminus (Saha et al., 2017). The sequence analysis of the yet-uncharacterized Alr0831 (AzuR) revealed the absence of a functional α3N site in AzuR as it contained only one cysteine residue each in the α3 helix and at the amino-terminus (Figure 1A). Protein sequence alignment of AzuR with the Synechococcus transcriptional repressor SmtB showed 60% sequence identity, and the key amino acids in the α5 site important for metal sensing, i.e., His, Glu, and Asp in SmtB (VanZile et al., 2000, 2002), were found to be conserved in AzuR (Figure 1A). It is likely that the function of AzuR is similar to that of SmtB owing to the high degree of sequence identity. Tertiary structure prediction of AzuR using the software I-TASSER showed the presence of all the secondary structural folds (α1–α5, β1, and β2) similar to that of SmtB (Figure 1B, i). Structural modeling predicted zinc-binding residues Asp102 and His104 (Figure 1B, ii) of AzuR comparable to that of Staphylococcus aureus CadC as well as His115 and Glu118 (Figure 1B, iii) similar to that of Synechococcus SmtB. Hence, AzuR could possibly be grouped into α5 SmtB/ArsR metalloregulators with the signature motif DxHx₁₀Hx₂E present in the α5 helix (Figure 1A). Phylogenetic analysis of representative sequences from SmtB/ArsR family members showed that AzuR shared maximum identity to BxmR (67%), which contain both α3N and α5 sites (Figure 1C). It also showed that SmtB (α5) and proteins belonging to different groups—ZiaR (α3N, α5), AztR (α3N), BxmR (α3N, α5), and AzuR (α5)—evolved independently but were linked to a common ancestor (Figure 1C).

Several metal-responsive proteins and their repressors of SmtB/ArsR family members have been shown to exist as operons. For example, BmtA (MT of Oscillatoria brevis) and its repressor BxmR (Liu et al., 2004), ZiaA (Zn efflux protein of Synechocystis PCC 6803) and its repressor ZiaR (Thelwell et al., 1998), and AztA (Zn²⁺-translocating CPx-ATPase) and its repressor AztR (Liu et al., 2005) are organized in operons. In Synechococcus PCC 7942, the smtB gene and smtA gene are separated by 100 bp, forming a divergon (Huckle et al., 1993). However, there is a deviation in the genetic organization of Anabaena MT, which is not organized in an operon. The nmtA ORF (located between positions 3938083 and 3937925) is present within a larger ORF of an unknown protein, asr3266, but in the opposite orientation (Bose et al., 2006). Similarly, the putative regulator azuR is not placed adjacent to the nmtA locus but is present within the ZnuABC operon (Figure 1D). Alr0831 is positioned at 956795→957157 between alr0830 (ZnuC, ABC transporter permease protein) and alr0832 (ZnuA, ABC transporter ATP binding protein) in the opposite orientation. Similar to AzuR, the SmtB ortholog has been identified within an operon with an ABC-type transporter system in other cyanobacteria like Nodularia and Anabaena variabilis (Blindauer, 2008). Analysis of the genomic organization of other prokaryotic MTs like Pseudomonas MT also revealed an absence of regulatory protein adjacent to the Pseudomonas fluorescens Q2-87 MT locus. Also, the genes adjacent to the Pseudomonas MT gene code for proteins of unknown function (Habjanič et al., 2020). Genomic arrangement of MT and its regulator as operons apparently is not mandatory as such regulators function as trans-acting factors on cis-regulatory elements.

To characterize the regulatory role of AzuR, the corresponding gene (alr0831) was cloned in the pTwin1 vector. The resulting construct pTwinazuR was expressed in the E. coli SHuffle strain. Induction with IPTG expressed a ∼42 kDa protein corresponding to CBD-tagged AzuR (Figure 2A). The cloning at NdeI–SapI sites ensured that no extra amino acids were incorporated in the purified protein following removal of the tag. AzuR was purified by chitin affinity chromatography, and the removal of the CBD tag was achieved by thiol-induced cleavage with 40 mM DTT at 4°C. The purified AzuR was visualized on SDS-PAGE as a monomer under reducing conditions with a molecular weight of ∼13.9 kDa (Figure 2B), which was further confirmed with LC-MS/MS analysis. The MS analysis identified six unique peptides, showing 77% coverage of the Anabaena AzuR protein sequence.

The SmtB/ArsR family of proteins binds to the regulatory DNA sequences as homodimers (Osman and Cavet, 2010). To ascertain the native form of AzuR, the oligomeric status of AzuR was evaluated by glutaraldehyde cross-linking. The protein was predominantly found to be present in the dimeric state as observed by glutaraldehyde cross-linking (Figure 2C). The dimeric state was also confirmed with size-exclusion chromatography (Figure 2D). This is in agreement with the previously characterized SmtB/ArsR family of prokaryotic metalloregulatory transcriptional repressors that existed as stable dimers in solution (Busenlehner et al., 2001; Liu et al., 2005, 2008). It was observed that AzuR existed as a monomer under reducing conditions and dimer under non-reducing conditions (Figure 2C). These observations suggested the involvement of cysteine residues in AzuR dimerization. Secondary structure analysis by CD showed that AzuR is composed of 67.5% α helical content (Figure 2E), suggesting that the purified recombinant AzuR protein was properly folded. This is in agreement with the theoretical secondary structure prediction of AzuR using the SOPMA software, which projected 67% α helical content followed by 17% random coil and 11% extended strand.

The SmtB/ArsR family of transcriptional regulators binds to 12–2–12 inverted repeats present upstream or within the genes that they regulate (Erbe et al., 1995; Turner et al., 1996). RACE analysis with total RNA isolated from the cadmium-treated (IC₅₀ 10 μM) Anabaena 7120 showed an ∼200 bp cDNA product (Figure 3A). Sequence analysis of the product identified the transcriptional start site (TSS) to be at 23 nt upstream of the translational start of the nmtA ORF (Figure 3B). The palindromic sequence (12–2–12 imperfect inverted repeat), corresponding to the consensus of the α3N and α5 groups of SmtB/ArsR-binding sites (Saha et al., 2017), was found to be located 36 nt upstream of the nmtA translation start site (Figure 3B). Its position overlaps with the theoretical prediction of the −35 element of the promoter. It is shown that the cis-regulatory element of metal-inducible operons is composed of one or two inverted 12–2–12 repeats present in the vicinity or overlapping the transcriptional start site of the gene under regulation. For example, one of the two such inverted 12–2–12 repeats found in Synechococcus 7942 was essential for the regulation of smtA expression by its repressor, SmtB (Turner et al., 1996). Similarly, the Synechocystis zia O/P region has a single 12–2–12 inverted repeat between the −10 box and the translational start site of ziaA, which is regulated by a divergently transcribed repressor, ziaR (Thelwell et al., 1998). Pattern search analysis was performed with the conserved bases in the 12–2–12 imperfect repeat along the entire Anabaena 7120 genome. Similar repeats were found at sites upstream and within other genes that include all1178, which codes for a two-component hybrid sensor and regulator, alr7622 (also designated as aztA), encoding for cation-transporting ATPase and other hypothetical proteins (Table 2). The conserved 12–2–12 inverted repeat of SmtB/ArsR-regulated O/Ps are shown in Figure 3C. Although nmtA and its putative regulator azuR do not constitute an operon in Anabaena 7120, the inverted 12–2–12 imperfect repeat could be located at the appropriate upstream distance from the nmtA translation start site. Although the azuR ORF is present within the znuABC operon, a detailed search for conserved bases in the 12–2–12 imperfect repeat following global search analysis by PATLOC in the close vicinity of the znuABC operon (corresponding to the 500 bp upstream region to 500 bp downstream of the operon) and within the operon did not show any such repeat sequence in the entire analyzed region. The regulation of the znuABC operon by the zur (all2473)/furB regulator has been demonstrated previously in Anabaena 7120 (Napolitano et al., 2012). Zur (zinc uptake regulator), known to be the master regulator for zinc homeostasis in Anabaena 7120, regulated the expression of genes involved in zinc homeostasis like alr0830 (ZnuC), alr0833 (ZnuA), and all7621 (AztR). On analysis, we did not find zur-binding sequences upstream of the azuR ORF, indicating that the global regulator of zinc homeostasis, Zur, did not regulate azuR expression.

Anabaena 7120 AztA is transcriptionally regulated by AztR (belonging to the SmtB/ArsR family) by recognizing and binding to the inverted 12–2–12 imperfect repeat region. EMSA studies done with AztR and the nmtA/bmtA upstream region showed its binding in vitro (Tottey et al., 2007). Similar inverted repeat sequences identified by AztR and AzuR indicate that AztR and AzuR might be sharing the function of regulating AztA and NmtA. As described above, AztR belongs to the α3N group and AzuR to the α5 group of the SmtB/ArsR family. The α5 group members sense physiologically important metals like Zn²⁺, Cu²⁺, Co²⁺, and Ni²⁺, while the α3N group prefers larger, more thiophilic metal ions like Cd²⁺ or Pb²⁺ (Busenlehner et al., 2003). It is possible that AzuR and AztR preferred different groups of metal ions but could regulate both MT and efflux proteins, thus enabling the cell to respond to a wide range of metal ions.

Electrophoretic mobility shift assays (EMSAs) were done in order to identify the AzuR-DNA binding site using a 100 bp fragment (P_nmtA) upstream of the nmtA gene (probe) containing the 12–2–12 inverted repeat sequence. The results showed that AzuR could bind and form complexes with P_nmtA in a concentration-dependent manner (Figure 4A). The Hill coefficient of AzuR binding to DNA was calculated to be 2.48 ± 1.14 (>1) (Figure 4B), which indicated positive cooperative binding (Hill, 1910). SmtB has been shown to bind to the smt O/P in a multimeric state (Erbe et al., 1995). The positive cooperative binding suggested that AzuR bound to the target DNA as an oligomer similar to that of SmtB. The specificity of DNA–protein binding was confirmed by using the nmtA gene or DNA-binding protein LexA from Anabaena 7120 (AnLexA) or other proteins like BSA and NmtA. No retardation in the mobility of P_nmtA was observed in the presence of AnLexA. Also, AzuR could not bind to the nmtA gene sequence, confirming that AzuR regulated nmtA expression by binding to the upstream sequence and not to its internal region (Figure 4C). Our results established the specific binding of P_nmtA with the AzuR protein.

SmtB senses metal ions through the α5 site. Zinc binding to residues present in this site allosterically regulates the DNA binding activity of SmtB to the smtA O/P region (VanZile et al., 2002) similar to other reported SmtB/ArsR repressors (Busenlehner et al., 2003). The bound Zn²⁺ changes the conformation of the protein, which inhibits the DNA binding. Since AzuR contains a similar α5 site, the conformational changes in AzuR as a result of metal binding was assessed by CD spectra of the protein in the presence of various concentrations of zinc (Figure 4D). The degree of the alpha helical region progressively decreased with increasing concentrations of zinc, indicating the changes in the secondary structure of AzuR in the presence of zinc. To further confirm whether zinc or other metal ions interfered with the AzuR DNA binding ability, EMSA was carried out in the presence of various metal ions. Dissociation of the DNA–AzuR complex was clearly evident with increasing concentrations of Zn²⁺ (Figures 4E,F). The interaction of Cd²⁺ with AzuR also disrupted the binding with P_nmtA (Figure 4F); however, the disruption was more prominent in the presence of DTT (Figure 4G), emphasizing the requirement of free sulfhydryls for Cd²⁺ binding to AzuR in vitro. It was interesting to see the reversal of AzuR binding to P_nmtA in the presence of other divalent metal ions like Cu²⁺, Co²⁺, Ni²⁺, Pb²⁺, and Mn²⁺ (Figure 4H), suggesting that AzuR not only senses toxic metal ions like Cd²⁺ and Pb²⁺ but also is capable of sensing essential metal ions like Zn²⁺, Cu²⁺, Co²⁺, Ni²⁺, and Mn²⁺. EMSAs attempted with metals other than Zn²⁺ and Cd²⁺ in the presence of DTT showed visible precipitates in the binding reaction and hence were not included here.

We have previously observed the induction of nmtA in the presence of Cd²⁺, Zn²⁺, and Cu²⁺ (Divya et al., 2018). AzuR, therefore, can be proposed as a negative regulator of nmtA as it binds to regulatory DNA sequence in the absence of the metals and the repression is relieved in the presence of metal ions. In view of our results, it can be suggested that AzuR might have a larger role in the metal resistance system of Anabaena 7120.

Overexpression of transcriptional regulators has been previously studied in Anabaena sp. (Wu et al., 2007). To gain insights into the effect of AzuR on various characteristics or phenotype of Anabaena 7120, we constructed a recombinant strain of Anabaena 7120 that overexpressed AzuR. The azuR gene was cloned and overexpressed constitutively in Anabaena 7120 from a strong light-inducible promoter, P_psbA. GFP fluorescence of the downstream reporter gene was the first indication of successful azuR gene expression (Figure 5A). GFP fluorescence was visualized in Anabaena harboring an empty vector with P_psbA upstream of the gfpmut2 gene, AnpsbA⁺ and Anabaena overexpressing nmtA, and AnnmtA⁺ (Figure 5A). WT cells did not show any such GFP fluorescence (Figure 5A). The observation of few cells appearing red in the filaments of recombinant cells under FM and blue-light excitation (BLE) conditions could be due to partial or reduced GFP expression (Figure 5A). The filament length in AnazuR⁺, AnpsbA⁺, and AnnmtA⁺ was comparable to that of WT Anabaena cells. The uniformity of the cell stacking in AnazuR⁺ filaments appeared to be compromised as compared to those in the filaments of WT, AnpsbA⁺, and AnnmtA⁺. However, the Chla fluorescence in AnazuR⁺ cells was intact and equivalent to that observed for WT, AnpsbA⁺, or AnnmtA⁺ cells (Figure 5A). A substantial increase in azuR transcript level was seen in RT-PCR performed with RNA isolated from AnazuR⁺ as compared to AnpsbA⁺ and AnnmtA⁺, thus confirming the overexpression of the regulator in vivo (Figure 5B).

Scanning electron microscopy (SEM) analysis of exponential-phase cells of AnazuR⁺ revealed a significant decrease in cell size with the cells showing spherical and globular morphology in contrast to AnpsbA⁺, AnnmtA⁺, and WT cells (Figure 5A). The average cell size of AnazuR⁺ cells was found to be 2.70 ± 0.26 μm as compared to 3.92 ± 0.50 μm for AnpsbA⁺ and 3.67 ± 0.34 μm for AnnmtA⁺. The cell size of AnazuR⁺ was lesser than the WT cells (3.214 ± 0.34 μm) (Figure 5C). Similar morphological changes regarding cell stacking and cell size were observed following overexpression of the global transcriptional regulator FurA in Anabaena 7120 (González et al., 2010). The elongated cell phenotype seen in AnpsbA⁺ and AnnmtA⁺ cells could be because of stress owing to neomycin and heterologous GFP overexpression. The gross morphological changes in AnazuR⁺ as compared to the empty vector AnpsbA⁺ indicate the possible involvement of AzuR in the regulation of genes involved in functions other than metal homeostasis. Chromatin immunoprecipitation (ChIP) studies need to be done in the future to identify direct binding of targets of AzuR in the Anabaena genome.

DNA binding studies by EMSA showed that AzuR bound to the upstream region of the nmtA ORF in vitro. Evaluation of nmtA expression levels in AnazuR⁺ by qRT-PCR with 16S rRNA as internal control showed the downregulation of nmtA expression in AnazuR⁺ by ∼32-fold as compared to its empty vector AnpsbA⁺. These results are in agreement with the negative regulation of nmtA transcription by AzuR in vivo.

Previously, overexpression of NmtA in Anabaena 7120 had conferred tolerance to cadmium stress (Divya et al., 2018). Since the negative regulation of nmtA transcription by AzuR was observed here, we were interested to see the effect of the overexpression of AzuR on the cadmium tolerance ability of Anabaena 7120. We compared the response of cadmium stress in AnazuR⁺, AnpsbA⁺, and AnnmtA⁺ cultures. Spot assays showed increased sensitivity of AnazuR⁺ cells to cadmium stress following 7 days of exposure (Figure 6A). Growth of AnazuR⁺ assessed in terms of Chla content showed a substantial decrease even at concentrations of 10 μM cadmium as compared to AnpsbA⁺ (Figure 6B). Growth kinetics studies in the presence of 20 μM cadmium resulted in almost complete bleaching of cultures of both AnpsbA⁺ and AnazuR⁺ after 10 days of exposure to the stress (Figure 6C) including extensive cell lysis in AnazuR⁺ culture (Figure 6D). In contrast, filaments of AnnmtA⁺ appeared intact, long, and healthy on exposure to cadmium (Figure 6D). The spot assays and growth studies assessed in terms of Chla contents (Figures 6A–C) of AnnmtA⁺ also supported the microscopy observations, which are in agreement with our previous results showing superior tolerance of AnnmtA⁺ against cadmium stress (Divya et al., 2018). The GFP and Chla fluorescence were found to be unaffected in AnnmtA⁺ similar to AnazuR⁺ and AnpsbA⁺ in the presence of cadmium (Figure 6D). The toxic effects of cadmium on the photosynthetic machinery have been studied extensively in Synechocystis PCC 6803 (Tóth et al., 2012). The major proteins involved in photosynthetic machinery include zinc-containing enzymes like carbonic anhydrase and sulfhydryl groups in ribulose-5-phosphate kinase among others that lose their activity by replacement with cadmium (Tóth et al., 2012). MTs play a key role in metal detoxification by directly binding to the toxic metal, which results in lesser bioavailability (Klaassen et al., 1999). This protects the essential metalloproteins from the toxic metal. Since AzuR overexpression leads to a decrease in basal nmtA expression, the protective role of NmtA in imparting cadmium tolerance could be obliterated, resulting in the susceptibility of AnazuR⁺ to cadmium stress, which was evident from its decreased growth and increased cell lysis. The susceptibility of AnazuR⁺ to cadmium stress confirms the negative regulation of nmtA expression at the physiological level in Anabaena.