The Transcriptional Repressor PerR Senses Sulfane Sulfur by Cysteine Persulfidation at the Structural Zn2+ Site in Synechococcus sp. PCC7002

Cyanobacteria can perform both anoxygenic and oxygenic photosynthesis, a characteristic which ensured that these organisms were crucial in the evolution of the early Earth and the biosphere. Reactive oxygen species (ROS) produced in oxygenic photosynthesis and reactive sulfur species (RSS) produced in anoxygenic photosynthesis are closely related to intracellular redox equilibrium. ROS comprise superoxide anion (O2●−), hydrogen peroxide (H2O2), and hydroxyl radicals (●OH). RSS comprise H2S and sulfane sulfur (persulfide, polysulfide, and S8). Although the sensing mechanism for ROS in cyanobacteria has been explored, that of RSS has not been elucidated. Here, we studied the function of the transcriptional repressor PerR in RSS sensing in Synechococcus sp. PCC7002 (PCC7002). PerR was previously reported to sense ROS; however, our results revealed that it also participated in RSS sensing. PerR repressed the expression of prxI and downregulated the tolerance of PCC7002 to polysulfide (H2Sn). The reporter system indicated that PerR sensed H2Sn. Cys121 of the Cys4:Zn2+ site, which contains four cysteines (Cys121, Cys124, Cys160, and Cys163) bound to one zinc atom, could be modified by H2Sn to Cys121-SSH, as a result of which the zinc atom was released from the site. Moreover, Cys19 could also be modified by polysulfide to Cys19-SSH. Thus, our results reveal that PerR, a representative of the Cys4 zinc finger proteins, senses H2Sn. Our findings provide a new perspective to explore the adaptation strategy of cyanobacteria in Proterozoic and contemporary sulfurization oceans.


Introduction
The environment on Earth transformed from anaerobic to aerobic during the evolution of life [1]. Cyanobacteria, some of the oldest microorganisms on Earth that can perform both anoxygenic and oxygenic photosynthesis, were a key driving force during evolution [2]. Life is created, regulated, and sustained by reduction-oxidation (redox) reactions, and ROS and RSS are two critical kinds of signal intracellular molecules associated with the redox balance [3]. In Proterozoic oceans, cyanobacteria perform anoxygenic photosynthesis, using alternative reduced electron donors, such as hydrogen sulfide (H 2 S) [4]. Therefore, RSS should be the major participant in the regulation of the intracellular redox balance in cyanobacteria. Oxygenic photosynthesis, which uses solar energy to pry electrons from water, became a major part of the Earth's ecosystems as it succeeded in oxygenating the atmosphere and the biosphere more than 3 billon years (Ga) ago [5]. The main player in redox regulation became ROS. Although the modern ocean is aerobic, there are still some areas that lack oxygen, such as oxygen-minimum zones [6] and microbial mats [7]. binds to the promoter region of prx to regulate its expression in response to peroxide stress [9]. Ludwig et al. found that prx expression in PCC7002 was also regulated by PerR [40]. However, these studies did not resolve the specific regulation mechanism of PerR. The mechanism by which PerR senses H 2 O 2 in Bacillus subtilis has been reported in detail. B. subtilis PerR contains a structural metal ion (Zn 2+ ) binding site and a regulatory metal ion (Fe 2+ or Mn 2+ ) binding site. In the presence of excess H 2 O 2 or O 2 , the two histidines that constitute the binding site of regulatory metal ions are oxidized, and the inhibitory effect of PerR is released. This is the main mechanism by which PerR senses H 2 O 2 [38]. In addition, four conserved cysteines combine with Zn 2+ to form a Cys 4 :Zn 2+ structure, which also plays a key role in the process of redox regulation [39]. Based on this mechanism of H 2 O 2 sensing and considering that the site of action of sulfane sulfur is cysteine [41], we speculated that the active site of sulfane sulfur on PerR may be the Cys 4 :Zn 2+ structure. It has not been reported how, or indeed whether, the Cys 4 :Zn 2+ structure is affected by sulfane sulfur, and thus, the mechanism needs to be explored in greater depth.
Here, we report that Synechococcus PerR senses sulfane sulfur and regulates the expression of prxI. PerR effectively decreases the tolerance of PCC7002 to sulfane sulfur by altering the expression of prxI. Sulfane sulfur modifies Cys 19 and Cys 121 to form Cys 19 -SSH and Cys 121 -SSH; as a result, the zinc atom is released from the Cys 4 :Zn 2+ site, destroying the function of PerR. The discovery that sulfane sulfur acts on the Cys 4 :Zn 2+ site of regulators is significant. Our findings reveal a new sulfane sulfur sensing mechanism, and provide a new perspective for exploring the adaptive mechanism of cyanobacteria in the evolution from an anaerobic environment to an aerobic one on Earth and the contemporary anoxic environment.

Strains and Culture Conditions
PCC7002 and its mutants (PCC7002∆perR and PCC7002∆prxI∆perR) were grown in conical flasks containing medium A + [42] under continuous illumination of 50 µmol photons m −2 ·s −1 , at 30 • C. To sustain normal growth, 30 µg/mL chloramphenicol was added to the medium of PCC7002∆perR, and 50 µg/mL kanamycin and 30 µg/mL chloramphenicol were added to the medium of PCC7002∆prxI∆perR. Escherichia coli strains were cultured in LB medium, at 37 • C. All strains and plasmids are listed in Table S1.

Construction of PCC7002 Mutants
The PCC7002∆prxI mutant was constructed in our previous study [31]. PCC7002∆perR and PCC7002∆prxI∆perR were constructed by natural transformation and homologous recombination according to a previously reported method [24]. The plasmid used in perR deletion was constructed as follows. First, two segments,~1000-bp long, immediately upstream and downstream of the perR gene, were acquired using the primer sets perRdel-1/perR-del-2 and perR-del-5/perR-del-6 (Table S2) by PCR from genomic DNA of PCC7002. The chloramphenicol resistance cartridge was amplified with the primers perRdel-3/perR-del-4. Second, the above three segments were fused by PCR, and they were connected with the pJET1.2 blunt vector by the TEDA method [43]. Then, the product was transformed into E. coli DH5α by electroporation, and correct transformants were verified by PCR and sequencing. For PCC7002∆perR, the correct plasmid was transformed into PCC7002 by natural transformation. For PCC7002∆prxI∆perR, the correct plasmid was transformed into PCC7002∆prxI. Here, 30 µg/mL chloramphenicol was used to select for correct transformants. Finally, the mutants PCC7002∆perR and PCC7002∆prxI∆perR were verified by PCR and sequencing.
2.3. The Toxicity of H 2 S n against PCC7002, PCC7002∆perR, and PCC7002∆prxI∆perR H 2 S n , at concentrations of 1, 3, and 5 mM, was added to the sealed centrifugation tubes containing PCC7002, PCC7002∆perR, and PCC7002∆prxI∆perR cells at log phase with an OD 730nm of 0.6-0.7. H 2 S n was prepared according to a previously reported method with minor modification [15]. Briefly, sulfur powder, NaOH, and NaHS were mixed in a 1:1:1 molar ratio and dissolved in distilled water under argon gas in sealed bottle. Then, the bottle was incubated, at 37 • C, till sulfur was completely dissolved. After 6 h incubation, at 30 • C, under continuous illumination of 50 µmol photons·m −2 ·s −1 , cells were washed and resuspended in fresh A + medium. Then, 10 µL of cells was placed on the A + agar plate after diluting with A + medium to 10 0 , 10 −1 , and 10 −2 . The plates were cultivated at 30 • C under continuous illumination of 50 µmol photons·m −2 ·s −1 for 7 days.
2.4. Induction, RNA Extraction, and qRT-PCR Analysis PCC7002 and PCC7002∆perR cells at log phase with an OD 730 nm of 0.6-0.7 were incubated with or without H 2 S and H 2 S n (at concentrations of 250 and 500 µM) in sealed centrifuge tubes for 3 h, at 30 • C, and 50 µmol photons·m −2 ·s −1 illumination. H 2 S was prepared according to the previous report [44] and experimental requirements, and the preparation method was as follows: 56.06 mg NaHS was dissolved into 1 mL of buffer (containing 50 mMTris.HCL and 100 µM DTPA), which had been degassed with N 2 prior to NaHS powder solubilization, and diluted according to the desired concentration. Then, the induced cells were harvested by centrifugation at 10,000× g, and 4 • C for 10 min. Total RNA was isolated using the TaKaRa MiniBEST Universal RNA Extraction Kit, and the concentration and quality of RNA were verified by Qubit 4 (Invitrogen, Carlsbad, CA, USA). The cDNA was acquired using the Prime Script™ RT reagent kit with gDNA Eraser (TaKaRa, Dalian, China). qRT-PCR was performed using the CFX96 Touch Real-Time PCR Detection System (Bio-Rad, Hercules, CA, USA) with the SYBR ® Premix Ex Taq™ II kit (TaKaRa, Dalian, China). The primers used for the target genes are shown in Table S1. The reference gene rnpA (SYNPCC7002_A0989) was also included [45]. The results were analyzed according to the 2 −∆∆CT method [46].

Construction of the perR-Repressed Reporter System
A perR-repressed reporter system in E. coli BL21 was constructed to assess the regulatory role of PerR on prxI expression. The plasmid pBBR-perR-PprxI-egfp was constructed as follows: The perR gene was expressed under the control of the lacI promoter, and the egfp gene was expressed under the control of the prxI promoter. PerR could act on the promoter region of prxI, thus influencing the fluorescence of GFP. The effect of H 2 S n and S 8 on PerR were evaluated by the changes in fluorescence intensities. The plasmid was transformed to E. coli BL21 for further study. E. coli BL21 (pBBR-perR-PprxI-egfp) was cultured in LB media, at 37 • C, to logarithmic phase (OD 600nm = 0.6) and 0.5 mM of isopropyl β-D-thiogalactoside (IPTG) was added to induce PerR expression. Then, H 2 S, H 2 S n , and S 8 (at concentrations of 0, 150, and 300 µM) were added to the medium and the cells were cultured for another 2 h. Finally, the cells were collected and washed twice with 50 mM PBS (pH 7.4) to detect the fluorescence of GFP at excitation and emission wavelengths of 482 nm and 515 nm.

Construction, Overexpression, and Purification of PerR
PerR was fused to the C-terminus of maltose binding protein (MBP) and overexpressed in the vector pMal-C2X [48]. To achieve this, the perR fragment was amplified from the PCC7002 genome using the primer pair pMal-perR-F/R, ligated to pMal-C2X by the TEDA method, and transformed into E. coli DH5α. Verified plasmid was then transformed into E. coli BL21(DE3), and the resulting pMal-perR cells were cultured in LB medium, at 37 • C, to an OD 600nm of 0.6. Then, 0.5 mM IPTG was added for an additional 6 h incubation, at 30 • C. The cells were disrupted by a pressure cell homogenizer (SPCH-18; Stansted Fluid Power Ltd., London, UK). The cell debris was removed by centrifugation at 13,000× g and 4 • C for 20 min. PerR protein with the MBP (MBP-PerR) was separated by Amylose Resin Column (Invitrogen, Carlsbad, CA, USA) according to the supplier's recommendations. PerR was released from the fusion with MBP using Factor Xa, at room temperature, for 24 h.

Zn 2+ Release Assay
PAR could bind to Zn 2+ and the Zn 2+ -PAR complex had maximum absorption at 494 nm; thus, absorption was used to indicate the amount of Zn 2+ . Here, 5 µM of purified PerR was treated with 10 mM H 2 S n or 10 mM H 2 O 2 in the presence of 100 µM PAR at 25 • C, and released Zn 2+ ions were measured by monitoring the Zn 2+ -PAR complex at 494 nm every 1 s for 10 min. PerR without treatment was used as a control.

Phylogenetic Analysis
Cyanobacterial genomes were downloaded from the NCBI database. The sequences in Table S3 were used as queries to obtain PerR candidates. The candidates were obtained by searching the database with the standalone BLASTP algorithm, using conventional criteria (E value of ≥1 × 10 −5 coverage of ≥45%, and identity of ≥30%) [53]. PerR candidates were aligned using MAFFT version 7.490 [54] with the option "-auto-maxiterate 1000", and ambiguously aligned regions were removed using trimAl version 1.4 [55] with the "gappyout" option. Phylogenetic analysis was performed based on maximum likelihood methods using IQ-TREE [56] with automatic detection of the best-fit model with the "-MFP" option using ModelFinder [57] under the Bayesian information criterion (BIC). The topological robustness of the tree was evaluated by 1000 ultrafast bootstrap replicates. PerR proteins from Staphylococcus epidermidis, Staphylococcus haemolyticus, and Staphylococcus aureus, detailed in Table S3 were used as an outgroup.

Phylogenetic Analysis of PerR in Cyanobacteria
To investigate the distribution of PerR in cyanobacteria, we performed a BLASTsearch of the 198 cyanobacteria genomes (downloaded from the NCBI database on 17 December 2021) with the queries (Table S3) to find PerR candidates ( Figure 1). PerR genes were identified using phylogenetic tree analysis ( Figure 1A). In total, 68 PerR-encoding genes were distributed among 64 cyanobacteria genomes (Table S4). The cyanobacteria PerRs were distributed among five orders, including 25 Synechococcales, 27 Nostocales, 3 Gloeobacteria, 9 Oscillatoriales, and 4 Pseudanabaenales ( Figure 1B). In Gloeobacteria, which was believed to be the early diverging lineage of cyanobacteria, all three of the published genomes within this order were encoded as perR. Furthermore, the proportions of species that contained perR in Oscillatoriales and Nostocales were 81.8% and 56.3%, respectively. For Synechococcales, the proportion was 27.5%, even though the total number of PerR genes was 25. For Pseudanabaenales, the proportion was only 23.5%. (Table S3) to find PerR candidates ( Figure 1). PerR genes were identified using phylogenetic tree analysis ( Figure 1A). In total, 68 PerR-encoding genes were distributed among 64 cyanobacteria genomes (Table S4). The cyanobacteria PerRs were distributed among five orders, including 25 Synechococcales, 27 Nostocales, 3 Gloeobacteria, 9 Oscillatoriales, and 4 Pseudanabaenales ( Figure 1B). In Gloeobacteria, which was believed to be the early diverging lineage of cyanobacteria, all three of the published genomes within this order were encoded as perR. Furthermore, the proportions of species that contained perR in Oscillatoriales and Nostocales were 81.8% and 56.3%, respectively. For Synechococcales, the proportion was 27.5%, even though the total number of PerR genes was 25. For Pseudanabaenales, the proportion was only 23.5%.  Table S3. PerRs from Staphylococcus epidermidis, Staphylococcus haemolyticus and Staphylococcus aureus in Table S3 were used as the outgroup. (B) The distribution of PerR-encoding genes in cyanobacteria genomes. In total, 68 predicted PerR-encoding genes were detected among 64 cyanobacteria genomes, including 25 Synechococcales, 27 Nostocales, 3 Gloeobacteria, 9 Oscillatoriales, and 4 Pseudanabaenalles.

PerR Deletion Increases the Tolerance of PCC7002 to High H 2 S n
To investigate the effect of PerR on the tolerance of PCC7002 to sulfane sulfur, we constructed a single-deletion strain PCC7002∆perR, and double-deletion strain PCC7002∆prxI ∆perR by homologous recombination ( Figure S1). The mutation was verified by PCR ( Figure  S1A). Then, the tolerance of PCC7002, PCC7002∆perR, and PCC7002∆prxI∆perR to sulfane sulfur was tested ( Figure 2). PCC7002∆perR grew better than the wild-type after induction with 5 mM H 2 S n , indicating that the deletion of perR increased tolerance (Figure 2A,B). However, the double-deletion mutant (PCC7002∆prxI∆perR) showed decreased tolerance to H 2 S n , and growth inhibition was apparent after induction with 3 mM H 2 S n ( Figure 2C). These results indicated that PerR and PrxI were all involved in H 2 S n tolerance of PCC7002.
Antioxidants 2023, 12, x FOR PEER REVIEW 8 of 18 Figure 2. PerR decreases the tolerance of PCC7002 to H2Sn. The deletion of perR (PCC7002ΔperR) (B) increased H2Sn tolerance, while the double deletion of perR and prxI (PCC7002ΔprxIΔperR) (C) decreased H2Sn tolerance compared with the wild-type (A). PCC7002, PCC7002ΔperR, and PCC7002ΔprxIΔperR cells at log phase with an OD730 nm of 1 were treated with 1, 3, and 5 mM H2Sn, at 30 °C, and 50 μmol photons m −2 ·s −1 illumination for 6 h. Then, cells were diluted with A + medium to 10 0 , 10 −1 , and 10 −2 , and plated onto the A + solid medium and cultured for 7 days, at 30 °C, and 50 μmol photons·m −2 ·s −1 illumination. The tolerance of PCC7002ΔperR and PCC7002ΔprxIΔperR to H2Sn was opposite to that of the wild-type. The expression of prxI was largely upregulated by H2Sn (D) and H2S (E) in PCC7002, while H2Sn (F) and H2S (G) showed little influence on its expression in PCC7002ΔperR. PCC7002 and PCC7002ΔperR cells at log phase were induced by H2Sn and H2S with concentrations of 250 μM and 500 μM for 3 h, and the expression of prxI was measured. The Y-axis is the fold change in prxI calculated by relative quantitative qPCR, based on the 2 −ΔΔCT method, with rnp as the reference gene. All data are averages from three samples with standard deviations (error PerR decreases the tolerance of PCC7002 to H 2 S n . The deletion of perR (PCC7002∆perR) (B) increased H 2 S n tolerance, while the double deletion of perR and prxI (PCC7002∆prxI∆perR) (C) decreased H 2 S n tolerance compared with the wild-type (A). PCC7002, PCC7002∆perR, and PCC7002∆prxI∆perR cells at log phase with an OD 730 nm of 1 were treated with 1, 3, and 5 mM H 2 S n , at 30 • C, and 50 µmol photons m −2 ·s −1 illumination for 6 h. Then, cells were diluted with A + medium to 10 0 , 10 −1 , and 10 −2 , and plated onto the A + solid medium and cultured for 7 days, at 30 • C, and 50 µmol photons·m −2 ·s −1 illumination. The tolerance of PCC7002∆perR and PCC7002∆prxI∆perR to H 2 S n was opposite to that of the wild-type. The expression of prxI was largely upregulated by H 2 S n (D) and H 2 S (E) in PCC7002, while H 2 S n (F) and H 2 S (G) showed little influence on its expression in PCC7002∆perR. PCC7002 and PCC7002∆perR cells at log phase were induced by H 2 S n and H 2 S with concentrations of 250 µM and 500 µM for 3 h, and the expression of prxI was measured. The Y-axis is the fold change in prxI calculated by relative quantitative qPCR, based on the 2 −∆∆CT method, with rnp as the reference gene. All data are averages from three samples with standard deviations (error bars). The experiment was repeated at least three times. ****, p < 0.0001; ns, not significant (paired t test).

PerR Senses H 2 S n and Regulates the Expression of prxI
PerR acts as a transcriptional repressor in the regulation of prxI expression, as qPCR analysis showed that the transcript level of prxI was upregulated~100-fold in PCC7002∆perR compared with PCC7002 ( Figure S1B). Furthermore, the expression levels of prxI were analyzed in PCC7002 and PCC7002∆perR after induction with H 2 S n and H 2 S to verify whether PerR is involved in the regulation of H 2 S n metabolism. The expression of prxI increased 1.5-fold following 250 µM H 2 S n induction and 3-fold following 500 µM H 2 S n induction ( Figure 2D); this effect was concentration dependent. H 2 S induction could also increase prxI expression by 2.5-fold at concentrations of 250 µM and 4-fold at concentrations of 500 µM ( Figure 2E). However, neither H 2 S nor H 2 S n could induce the expression of prxI in PCC7002∆perR ( Figure 2F,G), which was different from the wild-type. Based on the above results, we deduced that PerR played a critical role in H 2 S n sensing, thus regulating the expression of prxI.
Meanwhile, a PerR-repressed reporter system in E. coli BL21 was constructed to further assess the effect of H 2 S n on prxI expression regulated by PerR (Figure 3). In the reporter system, the perR gene is controlled by the lacI promoter (PlacI), and the egfp gene is controlled by the prxI promoter (P prxI ) ( Figure 3A). When the expression of PerR was induced by IPTG, GFP fluorescence decreased significantly, indicating that the expressed PerR could act on the prxI promoter and inhibit the expression of GFP ( Figure S2). Having verified that PerR acts on the promoter of prxI to inhibit its expression, the effects of H 2 S n and S 8 were tested. H 2 S n induction caused an increase in fluorescence intensity (Figure 2A), and S 8 , another form of sulfane sulfur, had a similar effect ( Figure 2B).
To test the critical role of Cys residues in PerR, all six Cys residues (Cys 19 , Cys 121 , Cys 123 , Cys 137 , Cys 160 , and Cys 163 ) were individually mutated to Ser. The mutation of Cys 19 and Cys 137 (C19S and C137S) resulted in decreased fluorescence intensity but did not affect their response to H 2 S n . Cys 121 , Cys 124 , Cys 160 , and Cys 163 were important components of the Cys 4 :Zn 2+ site, and their mutation to Ser (C121S, C124S, C160S, and C163S) resulted in increased fluorescence intensities compared with the wild-type, indicating the inactivation of PerR ( Figure 3C). As a result, PerR no longer acted on the prxI promoter to inhibit the expression of egfp, and it no longer responded to H 2 S n induction. The mutation of His had no effect on PerR ( Figure 3D), indicating H 2 S n did not act on the Fe 2+ /Mn 2+ site. We concluded that the expression of prxI was regulated by PerR, and the induction of S 8 and H 2 S n enhanced prxI expression by acting on PerR. Meanwhile, the Cys 121 , Cys 124 , Cys 160 , and Cys 163 residues of PerR played crucial roles in H 2 S n sensing.

Sulfane Sulfur Acts on the Cys 4 :Zn 2+ Site of PerR
Furthermore, we measured the rate at which Zn 2+ ions were released from PerR:Zn in the presence of H 2 S n (Figure 4). An amount of 1 µM PerR contains about 0.125 µM Zn, 0.033 µM Fe and 0.006 µM Mn, as detected by ICP-MS. As the previous result ( Figure 3D) confirmed that H 2 S n did not act on the Fe 2+ /Mn 2+ site, the effect on Cys 4 :Zn 2+ site was monitored here. The formation of the colored Zn 2+ -PAR complex, whose absorption maximum was observed at 494 nm, was used to monitor the release of Zn 2+ . Thus, the Zn 2+ release result showed that all selected concentrations of H 2 S n induced the release of Zn 2+ . Based on the above results, we deduced the mechanism by which H 2 S n acts on PerR, which involves H 2 S n acting on the Cys 4 :Zn 2+ site to dissociate Zn 2+ from the active site, thus destroying the normal function of PerR.
increased fluorescence intensities compared with the wild-type, indicating the inactivation of PerR ( Figure 3C). As a result, PerR no longer acted on the prxI promoter to inhibit the expression of egfp, and it no longer responded to H2Sn induction. The mutation of His had no effect on PerR ( Figure 3D), indicating H2Sn did not act on the Fe 2+ /Mn 2+ site. We concluded that the expression of prxI was regulated by PerR, and the induction of S8 and H2Sn enhanced prxI expression by acting on PerR. Meanwhile, the Cys 121 , Cys 124 , Cys 160 , and Cys 163 residues of PerR played crucial roles in H2Sn sensing.

Sulfane Sulfur Acts on the Cys4:Zn 2+ Site of PerR
Furthermore, we measured the rate at which Zn 2+ ions were released from PerR:Zn in the presence of H2Sn (Figure 4). An amount of 1 μM PerR contains about 0.125 μM Zn, 0.033 μM Fe and 0.006 μM Mn, as detected by ICP-MS. As the previous result ( Figure 3D) confirmed that H2Sn did not act on the Fe 2+ /Mn 2+ site, the effect on Cys4:Zn 2+ site was monitored here. The formation of the colored Zn 2+ -PAR complex, whose absorption maximum was observed at 494 nm, was used to monitor the release of Zn 2+ . Thus, the Zn 2+ release result showed that all selected concentrations of H2Sn induced the release of Zn 2+ . Based on the above results, we deduced the mechanism by which H2Sn acts on PerR, which involves H2Sn acting on the Cys4:Zn 2+ site to dissociate Zn 2+ from the active site, thus destroying the normal function of PerR.  Finally, we explored the mechanism by which H 2 S n acts on the Cys 4 :Zn 2+ site by LTQ-Orbitrap tandem mass spectrometry ( Figure 5). Cys 19 -SH of Peptide 1 in H 2 S n -treated PerR was modified to Cys 19 -SSH ( Figure 5A), while Cys 19 -SH of Peptide 2 in DTT-treated PerR was directly modified by acetamide (CAM) ( Figure 5B). Similarly, Cys 121 -SH of Peptide 3 in H 2 S n -treated PerR was modified to Cys 121 -SSH ( Figure 5C), while Cys 121 -SH of Peptide 4 in DTT-treated PerR was also directly modified by acetamide (CAM) ( Figure 5D). Among these, Cys 121 was an important constituent of the Cys 4 :Zn 2+ site, thus indicating that H 2 S n acts on Cys 121 to inhibit the activity of PerR. Notably, Cys 19 was also modified by H 2 S n , although it was not a component of the Cys 4 :Zn 2+ site. In summary, the persulfide modification of Cys 121 in the Cys 4 :Zn 2+ site by H 2 S n was the mechanism that affected PerR activity.
in H2Sn-treated PerR was modified to Cys 121 -SSH ( Figure 5C), while Cys 121 -SH of Peptide 4 in DTT-treated PerR was also directly modified by acetamide (CAM) ( Figure 5D). Among these, Cys 121 was an important constituent of the Cys4:Zn 2+ site, thus indicating that H2Sn acts on Cys 121 to inhibit the activity of PerR. Notably, Cys 19 was also modified by H2Sn, although it was not a component of the Cys4:Zn 2+ site. In summary, the persulfide modification of Cys 121 in the Cys4:Zn 2+ site by H2Sn was the mechanism that affected PerR activity. The purified PerR (5 mg/mL) was treated with 1 mM H2Sn and 1 mM DTT for 30 min, at 25 °C. After denaturing and incubating with IAM, the reacted protein was digested by trypsin. The generated peptides were detected by LTQ-Orbitrap tandem mass spectrometry.

Discussion
The data from our study revealed that PerR senses H2Sn and regulates the expression of prxI (Figure 6). The deletion of perR increased the tolerance for H2Sn in PCC7002 ( Figure 2B), although the enhanced tolerance was not observed in the dual mutant PCC7002Δper-RΔprxI ( Figure 2C), indicating that PerR functioned by acting on PrxI. Meanwhile, the induction effect of H2Sn on prxI transcription levels was only observed in the presence of PerR (Figure 2), a result which was further confirmed by the PerR-repressed reporter

Discussion
The data from our study revealed that PerR senses H 2 S n and regulates the expression of prxI ( Figure 6). The deletion of perR increased the tolerance for H 2 S n in PCC7002 ( Figure 2B), although the enhanced tolerance was not observed in the dual mutant PCC7002∆perR∆prxI ( Figure 2C), indicating that PerR functioned by acting on PrxI. Meanwhile, the induction effect of H 2 S n on prxI transcription levels was only observed in the presence of PerR (Figure 2), a result which was further confirmed by the PerR-repressed reporter system (Figure 3), indicating that PerR acted on the promoter region to inhibit prxI expression. H 2 S had similar effect with H 2 S n on prxI transcription levels ( Figure 2B). This effect may be caused by H 2 S n , which derive from H 2 S solution prepared from NaHS, as considerable levels of H 2 S n may present [44]. Meanwhile, H 2 S may be converted to H 2 S n by SQR [24]. Certainly, H 2 S may also act on PerR with a new mechanism directly, which needs to be further verified. Furthermore, H 2 S n acted on the Cys 4 :Zn 2+ site to release Zn 2+ , thus removing the inhibition (Figure 4) and allowing prxI to be expressed in large quantities to clear the excess sulfane sulfur. H 2 S n modified Cys 121 to form Cys 121 -SSH ( Figure 5), destroying the structure of the Cys 4 :Zn 2+ site and causing the release of Zn 2+ . Cys 19 could also be modified by H 2 S n . Thus, H 2 S n acted on the zinc figure structure of PerR, which represents a new type of mechanism for sulfane sulfur sensing. [24]. Certainly, H2S may also act on PerR with a new mechanism directly, which needs to be further verified. Furthermore, H2Sn acted on the Cys 4 :Zn 2+ site to release Zn 2+ , thus removing the inhibition (Figure 4) and allowing prxI to be expressed in large quantities to clear the excess sulfane sulfur. H2Sn modified Cys 121 to form Cys 121 -SSH ( Figure 5), destroying the structure of the Cys 4 :Zn 2+ site and causing the release of Zn 2+ . Cys 19 could also be modified by H2Sn. Thus, H2Sn acted on the zinc figure structure of PerR, which represents a new type of mechanism for sulfane sulfur sensing. Figure 6. PerR senses H2Sn and regulates the expression of prxI in PCC7002. H2Sn induces the expression of prxI via PerR. PerR binds to Zn 2+ and acts on the promoter of prxI to inhibit its expression, a process that can be disinhibited by H2Sn. H2Sn acts on the Cys4:Zn 2+ site of PerR to relieve Zn 2+ , destroying the zinc finger structure. C 121 -SH in the Cys4:Zn 2+ site is modified by H2Sn and forms C 121 -SSH.
Zinc-binding proteins are among the most abundant transcriptional regulators in eukaryotes, harboring at least one common motif, the zinc finger, which contributes to proper protein structure and function [58,59]. Zinc finger proteins are also found in prokaryotic genomes, such as Bacillus PerR [39] and Synechococcus PerR [9]. Zinc-binding proteins display variable secondary structures and vast functional diversity, and can be classified into three classes based on their distinct structural properties: Cys2His2 (C2H2) zinc finger proteins (Class I), Cys4 (C4) zinc finger proteins (Class II), and Cys6 (C6) zinc finger proteins (Class III). Class I proteins are often referred to as the classical zinc finger [60]. Class II proteins contain four cysteine residues bound to one zinc atom [61], whereas Class III proteins contain six cysteine residues bound to two zinc atoms [62]. Thus, Synechococcus PerR belongs to Class II, and this is the first report of a zinc-binding protein being involved in sulfane sulfur sensing.
OxyR and PerR are two representative regulators that can sense both H2O2 and H2Sn. In total, 68 PerR proteins were identified among 198 sequenced cyanobacteria genomes ( Figure 5), whereas only 9 OxyR proteins were identified (Table S5) [30], indicating that PerR may be the key player in cyanobacteria. Furthermore, PerR is a transcriptional inhibitor, whereas OxyR is a transcriptional activator, and their sensing mechanisms for H2O2 and H2Sn are quite different. The exact mechanism for the OxyR sensing of H2O2 is still under debate. The formation of a disulfide bond between Cys 199 and Cys 208 or the oxidization of Cys 199 to C 199 -SOH in E. coli are two of the proposed mechanisms [63][64][65]. For H2Sn sensing, the Cys 199 of E. coli OxyR is modified to Cys 199 -SSH [30]. Bacillus PerR Figure 6. PerR senses H 2 S n and regulates the expression of prxI in PCC7002. H 2 S n induces the expression of prxI via PerR. PerR binds to Zn 2+ and acts on the promoter of prxI to inhibit its expression, a process that can be disinhibited by H 2 S n . H 2 S n acts on the Cys4:Zn 2+ site of PerR to relieve Zn 2+ , destroying the zinc finger structure. C 121 -SH in the Cys4:Zn 2+ site is modified by H 2 S n and forms C 121 -SSH.
Zinc-binding proteins are among the most abundant transcriptional regulators in eukaryotes, harboring at least one common motif, the zinc finger, which contributes to proper protein structure and function [58,59]. Zinc finger proteins are also found in prokaryotic genomes, such as Bacillus PerR [39] and Synechococcus PerR [9]. Zinc-binding proteins display variable secondary structures and vast functional diversity, and can be classified into three classes based on their distinct structural properties: Cys 2 His 2 (C2H2) zinc finger proteins (Class I), Cys 4 (C4) zinc finger proteins (Class II), and Cys 6 (C6) zinc finger proteins (Class III). Class I proteins are often referred to as the classical zinc finger [60]. Class II proteins contain four cysteine residues bound to one zinc atom [61], whereas Class III proteins contain six cysteine residues bound to two zinc atoms [62]. Thus, Synechococcus PerR belongs to Class II, and this is the first report of a zinc-binding protein being involved in sulfane sulfur sensing.
OxyR and PerR are two representative regulators that can sense both H 2 O 2 and H 2 S n . In total, 68 PerR proteins were identified among 198 sequenced cyanobacteria genomes ( Figure 5), whereas only 9 OxyR proteins were identified (Table S5) [30], indicating that PerR may be the key player in cyanobacteria. Furthermore, PerR is a transcriptional inhibitor, whereas OxyR is a transcriptional activator, and their sensing mechanisms for H 2 O 2 and H 2 S n are quite different. The exact mechanism for the OxyR sensing of H 2 O 2 is still under debate. The formation of a disulfide bond between Cys 199 and Cys 208 or the oxidization of Cys 199 to C 199 -SOH in E. coli are two of the proposed mechanisms [63][64][65]. For H 2 S n sensing, the Cys 199 of E. coli OxyR is modified to Cys 199 -SSH [30]. Bacillus PerR senses H 2 O 2 by metal-catalyzed oxidation [38], where one oxygen atom is incorporated into histidine 37 or histidine 91, which coordinates the bound Fe 2+ . Cysteines in the Cys 4 :Zn 2+ site may also be oxidized by H 2 O 2 [39]. Our results revealed that Cys 121 in the Cys 4 :Zn 2+ site of Synechococcus PerR could be modified by H 2 S n to form Cys 121 -SSH (Figure 4), releasing one zinc atom and destabilizing the structure. Normally, PerR and OxyR do not exist in the same microbial strain, but there are some exceptions [66,67]. In cyanobacteria, PerR and OxyR did not coexist (Tables S4 and S5). Thus, although PerR and OxyR are considered functionally complementary, the mechanisms by which they function are different.
In addition to OxyR and PerR, two-component systems play an important role in H 2 O 2 signal transduction in cyanobacteria [68]. A microarray-based study in Synechocystis PCC6803 revealed that His kinases (Hiks), namely, Hik33, Hik34, Hik16, and Hik42, are involved in the expression of a large number of H 2 O 2 -inducible genes [10]. Among the four Hiks, Hik33 was the main contributor and was responsible for the regulation of more H 2 O 2inducible genes than PerR. Furthermore, the response of Synechocystis to H 2 O 2 treatment also relied on Group 2 sigma factors, namely, SigB and SigD [69]. The lack of Group 2 sigma factors meant that the strain was unable to sustain its growth under oxidative stress. Taken together, the signaling of H 2 O 2 -induced oxidative stress is based on the coordinated action of several regulators and dedicated alternative sigma factors. Whether these regulators participate in sulfane sulfur sensing requires further investigation.
The distribution of PerR proteins in cyanobacteria was also investigated. Gloeobacterales are early-branching photosynthetic cyanobacteria that are used as model species to study the physiology of early oxygenic phototrophs [70]. Gloeobacterales contain reduced photosystems that lack thylakoids and a circadian clock. However, our results revealed that all three species with published genomes within this order encoded perR, which may offer insight into the important role of PerR in primitive cyanobacteria and the evolution of oxygenic photosynthesis. Meanwhile, 81.8% of the species in Oscillatoriales and 56.3% of those in Nostocales contained PerR. Oscillatoriales and Nostocales are bloom-forming cyanobacteria that dominate among the cyanobacterial biomass of shallow polymictic eutrophic lakes [71]. The high proportion of PerR proteins among the two orders may provide insight into the survival strategies of cyanobacteria in hypoxic and sulfidic environments.
The finding that PerR senses H 2 S n in cyanobacteria is significant. First, cyanobacteria have to tolerate the accumulation of sulfane sulfur in living environments. In Proterozoic oceans [2] and modern oxygen minimum zones [72], the environments in which cyanobacteria thrive are anoxic and sulfidic, and as a result, sulfane sulfur might accumulate. In cyanobacteria mats, a typical habitat for these microorganisms, the cyanobacteria are intermittently exposed to sulfane sulfur [73]. Although cyanobacteria can perform sulfur respiration and provide ATP for growth under dark and anoxic conditions by reducing sulfane sulfur [74], excess sulfane sulfur is fatal to cells [23]. Therefore, PerR sensing of H 2 S n provides the opportunity for cyanobacteria to activate the expression of metabolic genes in time to scavenge excess sulfane sulfur, thus ensuring survival in such environments. Second, cyanobacteria perform anoxygenic photosynthesis under low-O 2 and sulfidic conditions, using H 2 S as the electron donor [8,75]. In addition, cyanobacteria could produce some sulfur-containing histidine such as ergothioneine and ovothiols, which can also be used as electron donors [76,77]. As a result, sulfane sulfur was generated during the process of H 2 S oxidation by SQR during anoxygenic photosynthesis. Sulfane sulfur is a signal that participates in the regulation of physiology and critical gene expression in photosynthesis [24]. The PerR sensing of H 2 S n may help cyanobacteria to maintain normal signal transduction and photosynthesis. Third, a previous study reported that the composition and stability of the photosynthetic machinery and the cell division process were affected by the overexpression of PerR [78], indicating that the effect of sulfane sulfur on the function of PerR may also affect the above process. Thus, the association between PerR, sulfane sulfur, photosynthesis, and cell division provides a new perspective on the significance of the PerR sensing of sulfane sulfur. In brief, the ability of PerR to sense H 2 S n may ensure that cyanobacteria respond to intracellular and extracellular sulfane sulfur in a timely manner, allowing them to maintain normal photosynthesis and cell division, and adapt to environmental conditions.

Conclusions
In summary, we showed that PerR in PCC7002 senses H 2 S n and regulates the expression of prxI. PCC7002 was able to respond in a timely manner to excess H 2 S n in the environment with the help of PerR, enhancing its tolerance. H 2 S n modified Cys 121 of PerR to form Cys 121 -SSH, thus releasing Zn 2+ from the Cys 4 :Zn 2+ site, revealing a new mechanism of sulfane sulfur sensing in cyanobacteria. This is also the first report of a zinc-binding protein that participates in sulfane sulfur sensing. Our findings offer new insight into the mechanism of sulfane sulfur sensing and provide a new perspective for understanding the adaptation mechanism of cyanobacteria in anaerobic and sulfidic environments.

Supplementary Materials:
The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/antiox12020423/s1; Figure S1. The deletion of perR was verified by PCR and its effect on the transcriptional level of prxI; Figure S2. The expressed PerR could act on the prxI promoter and inhibit the expression of GFP; Table S1. Strains and plasmids used in this study; Table S2. Primers used in this study; Table S3. The queries used in the phylogenetic analysis of PerR; Table S4. The information of PerRs in cyanobacteria; Table S5. The OxyRs in cyanobacteria.