Synechococcus sp. PCC7002 Uses Peroxiredoxin to Cope with Reactive Sulfur Species Stress.

Cyanobacteria are a widely distributed group of microorganisms in the ocean, and they often need to cope with the stress of reactive sulfur species, such as sulfide and sulfane sulfur. Sulfane sulfur refers to the various forms of zero-valent sulfur, including persulfide, polysulfide, and element sulfur (S8). Although sulfane sulfur participates in signaling transduction and resistance to reactive oxygen species in cyanobacteria, it is toxic at high concentrations and induces sulfur stress, which has similar effects to oxidative stress. In this study, we report that Synechococcus sp. PCC7002 uses peroxiredoxin to cope with the stress of cellular sulfane sulfur. Synechococcus sp. PCC7002 contains six peroxiredoxins, and all were induced by S8. Peroxiredoxin I (PrxI) reduced S8 to H2S by forming a disulfide bond between residues Cys53 and Cys153 of the enzyme. A partial deletion strain of Synechococcus sp. PCC7002 with decreased copy numbers of the prxI gene was more sensitive to S8 than was the wild type. Thus, peroxiredoxin is involved in maintaining the homeostasis of cellular sulfane sulfur in cyanobacteria. Given that peroxiredoxin evolved before the occurrence of O2 on Earth, its original function could have been to cope with reactive sulfur species stress, and that function has been preserved. IMPORTANCE Cyanobacteria are the earliest microorganisms that perform oxygenic photosynthesis, which has played a key role in the evolution of life on Earth, and they are the most important primary producers in the modern oceans. The cyanobacterium Synechococcus sp. PCC7002 uses peroxiredoxin to reduce high levels of sulfane sulfur. That function is possibly the original role of peroxiredoxin, as the enzyme evolved before the appearance of O2 on Earth. The preservation of the reduction of sulfane sulfur by peroxiredoxin5-type peroxiredoxins may offer cyanobacteria an advantage in the complex environment of the modern oceans.

INTRODUCTION

Cyanobacteria are one of the most important microbial groups; they provided the first source of O2 on Earth via oxygenic photosynthesis (1, 2). However, some environments that cyanobacteria inhabit periodically experience decreased oxygen levels. Cyanobacterial mats are one environment with periodically anoxic conditions, in which cyanobacteria perform oxygenic photosynthesis in the daytime and turn to respiration in the dark. The insufficient diffusion of O2 into the mat makes the mat turn anoxic. As a result, heterotrophic bacteria in the mat perform sulfate respiration and produce hydrogen sulfide (H2S). Cyanobacteria can use H2S as an electron donor to perform anoxygenic photosynthesis when oxygenic photosynthesis is inhibited by high concentrations of H2S. Tons of sulfane sulfur may be produced by the oxidation of H2S via sulfide:quinone oxidoreductase (SQR) in the mats (3, 4). Sulfane sulfur refers to the various forms of zero-valent sulfur, including persulfide, polysulfide, and elemental sulfur (S8). Cyanobacteria inhabiting oxygen minimum zones (OMZs), where H2S is sporadically accumulated, also face low-O2 and sulfidic conditions. Moreover, cyanobacteria encounter sulfur in the photic zones above OMZs, which are even visible as “clouds” on satellite images. Sulfane sulfur is also likely to be abundant in the benthic realm (5, 6). H2S and sulfane sulfur are two of the most important reactive sulfur species (RSS) that tend to be present in sulfidic conditions. RSS are a diverse class of sulfur-containing compounds and functional groups with important roles in chemical biology and bioinorganic chemistry (7–10). Therefore, cyanobacteria need to cope with the RSS stress caused by high concentrations of H2S and the accumulation of sulfane sulfur in the environments discussed above (11, 12).

Sulfane sulfur, including persulfide forms (RSSH and HSSH), polysulfide forms (RSSnH, RSSnR, and H2Sn, n ≥ 2), and elemental sulfur (S8), is commonly present in the cytoplasm of living organisms and plays important roles in maintaining intracellular redox homeostasis and metabolic regulation (7, 13–15). However, a high concentration of sulfane sulfur is toxic to cells and causes protein persulfidation and disulfide bond formation (16, 17). Inorganic polysulfides generated from organosulfur compounds inhibit several types of pathogenic and drug-resistant bacteria (18). Elemental sulfur is used as a potential antifungal agent (19). Consequently, cells have various enzymes and regulatory systems to protect against excessive levels of sulfane sulfur (20).

In some sulfidic environments, cyanobacteria can perform anoxygenic photosynthesis by using H2S as an electron donor, and the key enzyme in this process is SQR (1, 21–23). Cyanobacteria SQR and peroxidase dioxygenase (PDO) can work together to detoxify H2S (24). PDO is normally involved in the oxidation of the sulfane sulfur that is produced by SQR, but it functions at high levels of cellular sulfane sulfur. Other pathways that help to maintain the homeostasis of sulfane sulfur in cyanobacteria remain to be explored, considering the important role of sulfane sulfur in cellular signaling (25).

Microorganisms have evolved a series of mechanisms by which to maintain the homeostasis of intracellular sulfane sulfur. Besides PDO (26, 27), thioredoxins (Trx) and glutaredoxins (Grx) also reduce sulfane sulfur to H2S (28–31). Furthermore, some cyanobacteria are capable of sulfur respiration, using elemental sulfur as an electron acceptor in dark and anoxic conditions (3, 32, 33). Moreover, catalase, which typically catalyzes the disproportionation of H2O2 to H2O and O2 (34, 35), also has the ability to oxidize inorganic persulfide (H2S2), which is structurally similar to H2O2 (36). Because peroxiredoxin (Prx) also uses H2O2 as a substrate (37–39), an immediate question is whether Prx can metabolize sulfane sulfur.

Prxs are ubiquitous in plants, animals, and bacteria (38, 40–43). Their active Cys residue is oxidized to sulfonic acid by H2O2 and organic peroxides. Depending on whether one or two Cys residues are involved in the process of recycling sulfonic acid back to the thiol form, they can be divided into three categories: typical 2-Cys Prxs, atypical 2-Cys Prxs, and 1-Cys Prxs (40, 44, 45). However, this classification was not unequivocally accepted. Kimberly et al. developed a method that used the Deacon Active Site Profiler tool to extract functional site (PXXXTXXCP) profiles from structurally characterized Prxs and classify the Prxs into six distinct subclasses (46, 47): alkyl hydroperoxide reductase subunit C (AhpC-Prx), bacterioferritin comigratory protein (BCP-PrxQ), alkyl hydroperoxide reductase subunit E (AhpE), peroxiredoxin 5 (Prx5), peroxiredoxin 6 (Prx6), and thiol peroxidase (Tpx). Because inorganic polysulfide (H2Sn) and H2O2 are structural analogs, we hypothesize that Prx can also metabolize H2Sn. From the perspective of evolution, the origin of Prxs precedes the appearance of O2 on Earth (48). Therefore, Prx might have been used to manage intracellular RSS before the appearance of O2 and may offer cyanobacteria the advantage of being able to move in and out of hypoxic areas in the modern ocean (49–51).

Synechococcus sp. PCC7002 (PCC7002) contains six hypothetical Prxs: PrxI (ACA98797.1), PrxII (ACA98565.1), PrxIII (ACA99108.1), PrxIV (ACA98330.1), PrxV (ACA98124.1), and PrxVI (ACA99379.1). Among these, the role of PrxI in H2O2 metabolism has been confirmed (52). Here, we report that the Prxs in PCC7002 are all induced by S8. However, only PrxI was able to reduce S8 to H2S. The Cys53 and Cys153 residues of PrxI play critical roles in the reduction of sulfane sulfur to H2S via the formation of a disulfide bond. PrxI, which belongs to the Prx5 subfamily, was distinct from the other Prxs of PCC7002 in a phylogenetic analysis. These results improve our understanding of the sulfane sulfur metabolic pathway of cyanobacteria, provide some explanation for the widespread distribution of cyanobacteria in the modern ocean, and provide a new perspective from which to explore the important role of cyanobacteria in the early evolution of life on Earth.

RESULTS

Sulfane sulfur upregulates the expression of prxs in PCC7002.

Sulfane sulfur plays an important role in the regulation of the gene expression associated with photosynthesis in PCC7002, but it is toxic at high concentrations (24). S8 at the concentrations of 500 μM and 1 mM were fatal to PCC7002 (Fig. S1). Then, 100 and 250 μM S8 were used to induce PCC7002, and the expression of prxs were detected. All six prxs in PCC7002 were upregulated after induction by S8, as determined by a quantitative polymerase chain reaction (qPCR) analysis at all tested concentrations (Fig. 1A). At the concentration of 100 μM, the expression levels of prxIV and prxVI were upregulated by approximately 5-fold. The expression of prxI was increased notably (>10-fold). Furthermore, the expressions of prxII, prxIII, and prxV levels were also increased notably (from 20-fold to 30-fold). All prxs were also significantly upregulated at the concentration of 200 μM, although the amplitudes were not as high as those observed at 100 μM. Under reactive oxygen species (ROS) pressure, the expression of prxs changed slightly, by a maximum of about 2-fold after H2O2 induction (Fig. 1B) or incubation under 2%, 10%, or 20% O2 (Fig. 1C).

FIG 1

The effects of S8, H2O2, and O2 on the expression of prxs in PCC7002. The expression levels of prxI, prxII, prxIII, prxIV, prxV, and prxVI were measured using RT-qPCR after induction by S8 (A) and H2O2 (B) for 3 h. (C) The expression of prxs in PCC7002 incubated under 2%, 10%, and 20% O2 in the gas phase. To determine the expression levels of prxs, relative quantitative PCR was used. The relative gene expression represented the prx expression levels, standardized by the reference gene rpnA. All data are averages from three samples with standard deviations shown (error bars). The experiment was repeated at least three times. *, P value < 0.1; **, P value < 0.01; ***, P value < 0.001; ****, P value < 0.0001; ns, not significant (paired t-test).

The effect of S8 on the growth of PCC7002. (A) The growth curve of PCC7002 in the presence of S8. S8 at the concentrations of 0, 100, 200, and 500 μM were added to the PCC7002 culture at the beginning of the culturing. (B) The cell morphology of PCC7002 after incubation with 500 μM and 1 mM S8. PCC7002 at an OD730nm of 0.05 was cultured for 3 days after the addition of S8. All data are averages from three samples with standard deviations shown (error bars). The experiment was repeated at least three times. Download FIG S1, TIF file, 2.0 MB.

PrxI metabolized S8 and produced H2S.

To compare the functions of Prxs, H2S production by recombinant Prxs fused to the C-terminus of MBP was detected (Fig. 2). Prxs with His-tags were found to be insoluble. First, 100 μg/mL purified Prx–MBP fusion protein (Fig. S2A) were incubated with 200 μM elemental sulfur (S8) and 100 μM dithiothreitol (DTT) for 5 min at 30°C in 50 mM HEPES buffer (pH 7.0). About 90 μM H2S was released by Prx–MBP, while the H2S production by PrxII through PrxVI was not significantly different from that of a control that contained only 200 μM S8 and 100 μM DTT in HEPES buffer (Fig. 2A). Second, 200 μM S8 were added to cell lysates of recombinant E. coli BL21 expressing Prx–MBP fusion (10 mg of protein mL−1). An SDS-PAGE analysis showed a similar amount of the fused proteins in each sample (Fig. S2B). The lysate of recombinant E. coli BL21 expressing the PrxI fusion protein released about 130 μM H2S in 5 min of incubation. The control E. coli BL21 with empty vector pMal-C2X released only 90 μM H2S. Compared to the control, no more H2S was released by the lysates containing PrxII through PrxVI (Fig. 2B). Third, the ability of resting cells expressing Prx–MBP fusion protein to metabolize S8 and produce H2S was also measured, and only the resting cells with the PrxI-MBP fusion protein produced more H2S than did the control cells with the empty vector (Fig. 2C). Thus, PrxI clearly reduced sulfane sulfur to H2S.

FIG 2

The metabolism of S8 by Prxs. (A) The production of H2S by purified PrxI-MBP, Prx II-MBP, Prx III-MBP, Prx IV-MBP, Prx V-MBP, and Prx VI-MBP. 100 μg/mL purified Prx-MBP was incubated with 200 μM elemental sulfur (S8) and 100 μM DTT for 15 min at 30°C in 50 mM HEPES buffer (pH 7.0). CKA is represented the HEPES buffer with 200 μM elemental sulfur (S8) and 100 μM DTT. (B) The production of H2S by the lysates of the recombinant E.coli BL21 (DE3) (pMal-C2X) (CKB), E.coli BL21 (DE3) (pMal-prxI) (Prx I), E.coli BL21 (DE3) (pMal-prxII) (Prx II), E.coli BL21 (DE3) (pMal-prxIII) (Prx III), E.coli BL21 (DE3) (pMal-prxIV) (Prx IV), E.coli BL21 (DE3) (pMal-prxV) (Prx V), and E.coli BL21 (DE3) (pMal-prxVI) (Prx VI), with a total protein concentration of 10 mg/mL. 200 μM S8 was used as the source of sulfane sulfur, and the treatment time was 5 min. (C) The production of H2S by recombinant E.coli BL21 (DE3) (pMal-C2X) (CKC) and E.coli BL21 (DE3) (pMal-prxI-VI) cells. The recombinant E.coli BL21 strains were harvested and resuspended to an OD600 nm of 10, and then 200 μM S8 was added to initiate the reaction. The treatment time was 15 min. All data are averages from three samples with standard deviations shown (error bars). The experiment was repeated at least three times. *, P value < 0.1; **, P value < 0.01; ***, P value < 0.001; ****, P value < 0.0001; ns, not significant (paired t-test).

SDS-PAGE analysis of the purified Prxs-MBP and the lysates of recombinant E. coli BL21 (DE3) expressing Prxs-MBP. (A) Prxs were fused with MBP on the vector pMal-C2X and expressed in E. coli BL21 (DE3). Prxs-MBP were purified by the amylose resin and resolved by SDS-PAGE. The sample loading was 5 μg of total protein per lane. (B) The cell lysate of the recombinant E. coli BL21 (DE3) was analyzed by SDS-PAGE. The sample loading was 10 μg of total protein per lane. CK, E.coli BL21 (pMal-C2X); PrxI, E.coli BL21 (pMal-prxI); PrxII, E.coli BL21 (pMal-prxII); PrxIII, E.coli BL21 (pMal-prxIII); PrxIV, E.coli BL21 (pMal-prxIV); PrxV, E.coli BL21 (pMal-prxV); PrxVI, E.coli BL21 (pMal-prxVI). Download FIG S2, TIF file, 2.9 MB.

A previously reported CstR-mKate reporter (53) was adapted to analyze the function of PCC7002 PrxI and its Cys residues. The reporter system included CstR and mKate, in which CstR inhibits the expression of mkate. Sulfane sulfur could relieve the inhibitory effect of CstR. Thus, the fluorescence intensity of mKate in the E. coli host cells could serve as an indicator of the levels of intracellular sulfane sulfur (Fig. 3A1), reaching a maximum when E. coli cells entered the early stationary phase (25). When prxI was cloned behind mkate, the mKate fluorescence was decreased because of the metabolism of sulfane sulfur by PrxI (Fig. 3A2). PrxI contains three cysteine residues (Cys53, Cys78 and Cys153), and they were individually mutated to serine (Ser). The mKate fluorescence intensity in the modified reporter system with PrxI C78S was slightly higher than that in the system with wild-type PrxI. However, the mKate fluorescence intensities with PrxI C53S and PrxI C153S were significantly enhanced compared with those for the construct containing PrxI, and the control without PrxI had the highest mKate fluorescence (Fig. 3A2). Thus, the mutation of Cys53 and Cys153 destroyed the ability of PrxI to reduce sulfane sulfur.

FIG 3

The metabolism of S8 by PrxI and its mutants. (A) The function of PrxI detected by the CstR-reporter system. (A1) The schematic diagram of the CstR-reporter system. The CstR reporter system was built to access the activities of Prxs, in which H2Sn could relieve the repression of mkate by CstR. The mKate fluorescence was used to characterize the abilities of Prxs to metabolize sulfane sulfur. (A2) The fluorescence intensities of mKate in the CstR-reporter system coupled with PrxI and its cysteine mutants. (B) The production of H2S by purified PrxI, PrxI C53S, PrxI C78S, and PrxI C153S. (C) The production of H2S by lysates of recombinant E.coli BL21 (DE3) cells expressing prxI and its cysteine mutants. (D) Nonreducing SDS-PAGE of PrxI, PrxI C53S, PrxI C78S, and PrxI C153S after S8 and DTT treatment. The proteins were cleaved from the MBP-fusion proteins by Factor Xa at room temperature for 24 h, and 6 μg of PrxI proteins were loaded. All data are averages from three samples with standard deviations shown (error bars). The experiment was repeated at least three times. *, P value < 0.1; **, P value < 0.01; ns, not significant (paired t-test).

Furthermore, the Cys residues of PrxI in pMal-C2X were also individually mutated to Ser. Purified PrxI-MBP C53S and PrxI-MBP C153S produced less H2S than did wild-type PrxI-MBP and PrxI-MBP C78S, indicating the importance of Cys53 and Cys153 (Fig. 3B and Fig. S3A). Meanwhile, the lysate of E. coli BL21 cells expressing PrxI-MBP C53S or PrxI-MBP C153S also produced less H2S from added S8 than did cells expressing wild-type PrxI-MBP or PrxI-MBP C78S (Fig. 3C). The cell lysates of the E. coli BL21 expressing PrxI-MBP and its mutants were standardized by protein concentration and were confirmed to contain similar amounts of proteins by an SDS-PAGE analysis (Fig. S3B).

SDS-PAGE of the lysates of the purified PrxI, PrxI C53S, PrxI C78S, PrxI C153S, and the recombinant E. coli BL21 (DE3) expressing PrxI and its mutants. PrxI and its mutants were fused with MBP on the vector pMal-C2X and expressed in E. coli BL21 (DE3). (A) PrxsI, PrxI C53S, PrxI C78S, and PrxI C153S were purified by the amylose resin and resolved by SDS-PAGE. The sample loading was 5 μg of total protein per lane. (B) The cell lysate of the recombinant E. coli BL21 (DE3) was analyzed by SDS-PAGE. The sample loading was 10 μg of total protein per lane. PrxI, E.coli BL21 (DE3) (pMal-prxI); C53S, E.coli BL21 (DE3) (pMal-prxI C53S); C78S, E.coli BL21 (DE3) (pMal-prxI C78S); C153S, E.coli BL21 (DE3) (pMal-prxI C153S). Download FIG S3, TIF file, 2.6 MB.

We tested whether the Cys53 and Cys153 of PrxI formed a disulfide bond. The MBP fusion proteins were purified and cleaved by Factor Xa to release PrxI, PrxI C53S, PrxI C78S, and PrxI C153S. The released Prx proteins were analyzed by non-reducing SDS-PAGE. In the SDS-PAGE, untreated PrxI and PrxI C78S showed two bands, with the upper band being dominant. The upper band was converted to the lower band upon treatment with 250 μM S8. PrxI C53S and PrxI C153S showed only the upper band, and S8 treatment did not affect it (Fig. 3D). The upper band represented the PrxI protein without an intramolecular disulfide bond, while the lower band represented the protein with an intramolecular disulfide bond. All modifications were converted back to thiols by treatment with DTT. Hence, the Cys53 and Cys153 of PrxI are involved in reducing S8 to H2S, and they form an intramolecular disulfide bond.

PrxI enhanced the survival of PCC7002 after sulfane sulfur exposure.

The deletion of prxI affected the survival of PCC7002 after sulfane sulfur exposure. We tried to construct a single deletion strain by homologous recombination. However, prxI could only be partially knocked out (to give strain PCC7002ΔprxI-p), as the kanamycin-resistant mutant contained both the intact prxI gene and the kanamycin resistance gene when checked by PCR (Fig. S4). Cyanobacteria often have multiple chromosomes per cell (54), and many critical genes cannot be completely deleted from all chromosomes, as that would be fatal to the cell. Because prxI could not be completely deleted, PrxI is likely to play an essential physiological role in PCC7002. Even though not all copies of prxI were knocked out, the mutant showed a distinct response to S8 exposure compared to that of the wild type. PCC7002 and PCC7002ΔprxI-p cells were treated with various amounts of S8 for 6 h and then placed on A+ agar. After culturing for 7 days, PCC7002 grew well at the S8 concentration of 250 μM (Fig. 4A), while the growth of PCC7002ΔprxI-p was largely inhibited at that concentration (Fig. 4B). At 500 μM S8, PCC7002 could grow in small colonies, while PCC7002ΔprxI-p was completely inhibited. Furthermore, the growth curves of PCC7002 and PCC7002ΔprxI-p were monitored in the presence of 0, 100, 250, and 500 μM S8 (Fig. 4C). The growth curve of PCC7002 and PCC7002ΔprxI-p was similar in the absence of S8. PCC7002 grew better than PCC7002ΔprxI-p did in the presence of 100, 250, and 500 μM S8. The above results indicate that PrxI plays a critical role in the survival of PCC7002 after exposure to S8.

FIG 4

The tolerance of PCC7002 and PCC7002ΔprxI-p to S8-treatment. The growth of the wild type PCC7002 (A) and its prxI partial deletion mutant, PCC7002ΔprxI-p, (B) on the A+ agar plate after incubation with 100, 250, and 500 μM S8. The cells with an OD730nm of 0.05 were treated with S8 under 30°C and 50 μmol photons m−2 s−1illumination for 6 h. Then, the treated cells were diluted with A+ medium to 100, 10−1, and 10−2, then placed on the A+ plate for a further cultivation of 7 days under 30°C and 50 μmol photons m−2 s−1 illumination. (C) The growth curve of PCC7002 and PCC7002ΔprxI-p in the presence of S8. PCC7002 showed a higher resistance to S8 treatment than did PCC7002ΔprxI-p. All data are averages from three samples with standard deviations shown (error bars). The experiment was repeated at least three times.

The partial deletion of prxI was verified by PCR. The prxI gene was partially deleted from the PCC7002 genome, displaying two lanes in the agarose gel. Download FIG S4, TIF file, 2.3 MB.

Phylogenetic analysis of Prxs in PCC7002.

We conducted a phylogenetic analysis of the six Prxs in PCC7002 (Fig. 5A). Based on an analysis using the Deacon Active Site Profiler tool, Prxs are classified into six subfamilies (47). Here, representative sequences from each subfamily were selected to analyze the classification of the Prxs in PCC7002 (Table S3). The Prxs in PCC7002 belonged to five subfamilies: PrxI belonged to the Prx5 subfamily, PrxII belonged to the AhpC-Prx1 subfamily, PrxIII and PrxV belonged to the BCP-PrxQ subfamily, PrxVI belonged to the AhpE subfamily, and PrxIV belonged to the Prx6 subfamily. There is no Prx in PCC7002 belonging to the Tpx subfamily. Although the Prx5 subfamily was significantly different from the other subfamilies, as evidenced by it occupying a separate branch in the phylogenetic tree (Fig. 5A), the sequence around the active site of Prx5 subfamily members (PXXXTXXCP, where CP is the Cys53 of PrxI) is highly conserved (Fig. S5A). Based on the above findings, we deduced that the sequence specificity of PrxI determined its activity.

FIG 5

Phylogenetic analysis of Prxs in Cyanobacteria. (A) The genetic diversity of Prxs in PCC7002. The six Prxs in PCC7002 were divided into five subclasses: PrxI PCC7002 belonged to the Prx5 family, PrxII PCC7002 belonged to the AphC-Prx1 family, PrxIII PCC7002 and PrxV PCC7002 belonged to the BCP-PrxQ family, PrxIV PCC7002 belonged to the Prx6 family, and PrxVI PCC7002 belonged to the AphE family. The tree was an unrooted one. The candidates were analyzed by using ClustalW for alignment and MEGA 7.0 for neighbor-joining tree building with the following parameters: pairwise deletion, p-distance distribution, and bootstrap analysis of 1,000 repeats. (B) The fractions of cyanobacteria that carry the Prx5-type Prx. 127 of the 198 cyanobacteria genomes encoded Prx5-type Prx.

The queries used in the phylogenetic analysis. Download Table S3, DOCX file, 0.02 MB.

The consensus sequence and phylogenetic analysis of Prx5-type Prxs in cyanobacteria. (A) Prediction of the conserved sequence of the Prx5-type Prxs active sites by WebLogo. (B) There were 129 Prx5-type Prxs in PCC7002. All of the candidates were analyzed by using ClustalW for alignment and MEGA 7.0 for neighbor-joining tree building with the following parameters: pairwise deletion, p-distance distribution, and bootstrap analysis of 1,000 repeats. Trxs (Thioredoxin) in PCC7002 were used as the outgroup of Prx. Download FIG S5, TIF file, 2.3 MB.

Currently, 198 genomes of cyanobacteria have been sequenced, and we searched them for Prxs, using the queries in Table S3. There were 1,272 probable Prxs in these cyanobacteria, of which 194 belonged to the AhpC-Prx subfamily, 148 to the Prx6 subfamily, 189 to the AhpE subfamily, 612 to the BCP-PrxQ subfamily, and 129 to the Prx5 subfamily (Table S4). No Tpx family members were found in these cyanobacteria. The 129 Prx5 subfamily members were distributed across 127 cyanobacteria, and 65.5% of the sequenced cyanobacteria encoded at least one Prx5 (Fig. 5B and Fig. S5 and Table S5).

The distribution of Prxs in cyanobacteria. Download Table S4, DOCX file, 0.02 MB.

The information of all Prx5-type Prxs in cyanobacteria. Download Table S5, XLSX file, 0.03 MB.

DISCUSSION

Here, we report the participation of PrxI in sulfane sulfur metabolism in cyanobacteria (Fig. 6). S8 significantly induced the expression of all six prxs in PCC7002 (Fig. 1). Among them, we demonstrated that PrxI reduced S8 to H2S (Fig. 2) via the formation of an intramolecular disulfide bond between its Cys53 and Cys153 residues (Fig. 3). When prxI was partially inactivated, the PCC7002 mutant became more sensitive to S8 (Fig. 4). These results support the idea that PCC7002 uses PrxI to deal with RSS stress.

FIG 6

PeroxiredoxinI metabolized S8 and produced H2S. The expressions of prxs in PCC7002 are induced by S8. Among these, PrxI reduces S8 to H2S by donating an electron, thereby generating a disulfide bond between Cys53 and Cys153. Thioredoxin (Trx) then reduces the disulfide bond. As a result, PrxI helps PCC7002 to cope with the reactive sulfur species stress in the living environments.

Three experiments were designed to prove the function of Prxs in PCC7002: one using the purified Prxs-MBP (Fig. 2A), one using the cell lysate of E. coli expressing Prxs-MBP (Fig. 2B), and one using the resting cells of recombinant E. coli expressing Prxs-MBP (Fig. 2C). All three experiments indicated that only PrxI had the ability to reduce S8. DTT was used as a reductant in the purified protein experiment. Even though DTT can directly react with S8 to produce H2S, the existence of PrxI in the reaction produced more H2S (Fig. 2A). The maximum production of H2S by purified PrxI was at 15 min, while that of the cell lysate was at 5 min. This may be due to the fact that DTT was a chemical reductant which could be much lower than the physiological reductant in the cell lysate (55). The maximum production of H2S by the recombinant E. coli with PrxI-MBP was also at 15 min, which may be due to the slow transformation of S8 to cells. Furthermore, the lysates of E. coli expressing PrxII-MBP and PrxIV-MBP, as well as the resting cells expressing PrxIV-MBP and PrxVI-MBP, had lower H2S production. That may be due to the interaction of Prx with cellular components, as H2S production was not decreased in the experiment with using purified proteins (Fig. 2A). Prxs may have the ability to metabolize H2S in the presence of cellular components. It has been reported that Cu/Zn superoxide dismutase (SOD) catalyzed H2S oxidation to form polysulfide (56). We deduced that Prx may also have that ability, and this needs to be explored in a further study.

Prxs are antioxidant enzymes that play an important role in redox homeostasis and in redox regulation (28, 29). The mechanism of H2O2 metabolism by Prxs has been well-studied (38, 43). The “peroxidative” cysteine of the catalytic site (CP) attacks H2O2 and is oxidized to sulfenic acid (CP–SOH) in the first step of the catalytic cycle. Then, the resolving cysteine (CR) attacks the (CP–SOH) to release an H2O molecule and form a disulfide bond (CP–CR). Prxs are divided into three classes based on the way the sulfenic acid (CP–SOH) is recycled back to a thiol (CP–SH): typical 2-Cys Prxs, atypical 2-Cys Prxs, and 1-Cys Prxs. In the typical 2-Cys Prxs, the CP-SOH from one subunit is attacked by the CR from the other subunit, resulting in the formation of an inter-subunit disulfide bond. In the atypical 2-Cys Prxs, both the CP and the CR are contained in the same subunit, and the condensation reaction results in the formation of an intramolecular disulfide bond. The 1-Cys Prxs contain only CP and are without CR. The CP and CR residues of PCC7002 PrxI are Cys53 and Cys153, and they formed an intramolecular disulfide bond. A probable mechanism of sulfane sulfur reduction by PCC7002 PrxI is also proposed (Fig. 6) based on that of H2O2 metabolism, in which Cys53 reacts with sulfane sulfur, such as S8, to produce a persulfide (Cys53–SSH), and Cys153 attacks Cys53-SSH to form an intramolecular disulfide bond (Cys53–Cys153) and release H2S. In summary, we deduced that PCC7002 PrxI belongs to the atypical 2-Cys Prx family based on its mechanism (37, 40, 43), while it belongs to the Prx5 subfamily based on the analysis by the Deacon Active Site Profiler tool (Fig. 5A).

All six Prxs in PCC7002 were induced by S8 (Fig. 1). However, in this study, only PrxI had the S8 reduction activity. According to a phylogenetic analysis (Fig. 5), PrxI belongs to a separate branch from the other Prxs in PCC7002, the Tpx subfamily, while the sequences near the CP of the Tpx family are highly conserved (Fig. S5B), suggesting that this region may be the key site for sulfane sulfur reduction, which is also vital for H2O2 reduction. It remains to be investigated whether other Prxs, especially the Prx5-type Prxs, reduce sulfane sulfur. Our analysis showed that 65% of cyanobacteria encode Prx5-type Prx (Fig. 5B). These results indicate the widespread and important roles of Prx5-type Prx in cyanobacteria.

Prxs are most likely the primary enzymes responsible for maintaining intracellular sulfane sulfur homeostasis in cyanobacteria in anoxic or hypoxic conditions (21, 57). In aerobic conditions, PDO oxidizes sulfane sulfur to sulfite (24). Because the RSS stress is more severe in hypoxic conditions (5, 12), Prxs may play important roles in sulfane sulfur metabolism in anoxic or hypoxic conditions. The action of Prxs against RSS may have been preserved through evolution, as Prxs existed long before oxygen became abundant on Earth (48).

Prxs are known to participate in ROS metabolism (58). Here, we report that they are also involved in RSS metabolism. Cyanobacteria are the oldest surviving microorganisms, and they have experienced the transformation from an anaerobic environment on Earth to an aerobic environment (59, 60). Given the long history of cyanobacterial sulfur exposure, Prxs were most likely first used to resist RSS (5, 60). Interestingly, many of the strategies used against ROS stress, such as catalase, superoxide dismutase, and OxyR, have also been shown to be involved in coping with RSS stress (20, 36, 56). Here, the expression levels of prxs in PCC7002 were not as sensitive to H2O2 induction as they were to induction by S8 (Fig. 1C), which might be due to of the activity of other H2O2 mitigating enzymes, such as catalase and superoxide dismutase. Furthermore, sulfane sulfur might also disturb H2O2 homeostasis by downregulating catalase, thereby affecting the expression pattern of prx (Fig. 1). In summary, there is a close relationship between the strategies for coping with ROS and RSS (15, 61).

The maintenance of sulfane sulfur homeostasis in cyanobacteria is of great importance. In aerobic conditions, sulfane sulfur is an important intracellular signaling molecule that is involved in the regulation of critical photosynthesis genes in cyanobacteria. Sulfane sulfur reduction by Prx would help maintain the normal physiology and photosynthesis of cyanobacteria (24). In hypoxia and darkness, elemental sulfur can be used as an electron receptor for sulfur-dependent respiration, which enables cyanobacteria to yield ATP via the fermentation of endogenous stored glycogen (3, 33). However, high concentrations of sulfane sulfur in this environment can also be toxic to cells, so Prx-mediated sulfane sulfur reduction is a key pathway for detoxification, as photosynthesis ceases and oxidation by PDO is excluded in hypoxia and darkness.

In summary, here, a Prx is shown for the first time to act as a sulfur reductase that reduces S8 to H2S. Also, cyanobacteria may use Prxs to deal with RSS stress. S8 induced the expression of prxs in PCC7002, and PrxI worked effectively to reduce S8 to H2S, thereby improving the tolerance of PCC7002 to S8. The conserved sequence of PrxI near residue CP appears to be important for the activity of PrxI. Sulfane sulfur metabolism by Prxs could be the main strategy by which ancient cyanobacteria coped with RSS stress, which facilitated the survival of cyanobacteria in complex environments, especially oxygen-limited areas in the modern oceans.

MATERIALS AND METHODS

Strains and culture conditions.

PCC7002 and its mutants were grown in conical flasks containing medium A (62), supplemented with 1 mg of NaNO3 mL−1 (designed as medium A+) under continuous illumination by 50 μmol photons m −2 s −1 at 30°C. Kanamycin (50 μg/mL) was used to select the prxI mutant. To explore the effect of O2 concentrations, we cultured PCC7002 by bubbling with a mixture of O2 and N2, with O2 contents of 2%, 10%, and 20%. Escherichia coli (E. coli) was cultured in Luria-Bertani (LB) medium at 37°C. The strains and plasmids used in this paper are listed in Table S1.

Strains and plasmids used in this study. Download Table S1, DOCX file, 0.02 MB.

Induction, RNA extraction, and qRT-PCR analysis.

PCC7002 cells in logarithmic growth with an OD730 nm value of 0.6 to 0.7 were induced by S8, H2O2, and O2 under continuous illumination by 50 μmol photons m−2 s−1 at 30°C for 3 h. The cells were then harvested by centrifugation at 10,000 × g at 4°C for 10 min. Total RNA was isolated using the TaKaRa MiniBEST Universal RNA Extraction Kit, and the concentration of RNA was verified using a Qubit 4 instrument (Thermo Fisher). cDNA was produced using the Prime Script RT Reagent Kit with gDNA Eraser (TaKaRa, Beijing, China). The SYBR Premix Ex Taq II Kit (TaKaRa) was used for a quantitative reverse transcriptase polymerase chain reaction (qRT-PCR), and the reactions were run in a Light Cycler 480 II sequence detection system (Roche, Shanghai, China). Primers for target genes are given in Table S2. rnpA (SYNPCC7002_A0989), encoding the protein subunit of RNase P (RNase P), was used as the reference gene (63). The results were analyzed according to the 2−ΔΔCT method (64).

Primers used in this study. Download Table S2, DOCX file, 0.02 MB.

Overexpression of Prxs and enzyme activity determination.

Recombinant Prxs were fused to the C-terminus of maltose binding protein (MBP) and were overexpressed using the vector pMal-C2X (65, 66). Whole fragments encoding prxI–VI were amplified from PCC7002 genomic DNA using primers pMal-prxI–VI-F/R. Then, the fragments were ligated with pMal-C2X and transformed into E. coli DH5α. The resulting plasmids were transformed into E. coli BL21 (DE3) to overexpress the recombinant Prx fusion proteins. E. coli BL21 (pMal-C2X) and E. coli BL21 (pMal-prxs) were cultured in LB at 37°C to an OD600 nm of 0.6. Next, 0.5 mM isopropyl β-D-1-thiogalactopyranoside (IPTG) was added, and the cells were further cultivated at 30°C for 6 h. For resting cell analysis, cells were collected and resuspended in phosphate-buffered saline (PBS; 50 mM, pH 7.4) at an OD600 nm of 10. Then, S8 (200 μM) was added to initiate the reaction, and the release of H2S was determined by the methylene blue method (67). Here, the S8 was made by dissolving sulfur powder in acetone, in which it was soluble in the range of concentrations we used. For the analysis of cell lysates, the collected cells in PBS were disrupted using a pressure cell homogenizer (SPCH-18; Stansted Fluid Power Ltd., United Kingdom). The total protein content in the cell lyses was adjusted to 20 mg mL−1, and SDS-PAGE was used to verify whether the lysates contained similar amounts of recombinant Prx. Again, 200 μM S8 was added to start the reaction, and the release of H2S was determined by the methylene blue method. For the analysis of the purified protein, induced cells were harvested and resuspended in binding buffer (20 mM Tris-HCl, 200 mM NaCl, 1 mM EDTA). Then, the cells were disrupted using a pressure cell homogenizer, and the mixture was centrifuged at 20,000 × g for 20 min to acquire crude cell extract. The crude extract was loaded onto amylose resin, and the target protein was eluted using the binding buffer containing 10 mM maltose. The eluted protein solution was then loaded onto a PD-10 desalting column (GE) for buffer exchange to desalting buffer (20 mM NaH2PO4, 10% glycerol, pH 7.6). The purified proteins were then resolved by SDS-PAGE. The reaction mixtures contained 100 μg/mL Prx, 100 μM DTT, 200 μM S8, and 50 mM HEPES-NaOH (pH 7.0). The control containing DTT and S8 but no Prx was included. The cysteine to serine mutants of PrxI were generated using the primer pairs prxI-C53S-F/R, prxI-C78S-F/R, and prxI-C153S-F/R with a modified QuikChange site-directed mutagenesis method (68, 69). The reduction of S8 by cell lysates of E. coli BL21 (pMal-prxI C53S), E. coli BL21 (pMal-prxI C78S), and E. coli BL21 (pMal-prxI C153S) was detected as described above.

Non-reducing SDS-PAGE.

PrxI, PrxI C53S, PrxI C78S, and PrxI C153S with the MBP tag were purified in the same way as described above. PrxI, PrxI C53S, PrxI C78S, and PrxI C153S were released from the fusion with MBP by using Factor Xa at room temperature for 24 h. The released proteins were treated with 250 μM S8 at 25°C for 30 min. After the S8 treatment, 1 mM DTT was added to convert the modified thiols back to reduced thiols. No treatment and treatment with only 1 mM DTT were used as controls. The samples were then resolved by nonreducing SDS-PAGE, in which the loading buffer contained no DTT or other reducing agents.

Construction of the CstR reporter system.

A CstR-based reporter plasmid was constructed by following a reported protocol to assess the ability of PrxI to metabolize sulfane sulfur (53). In Staphylococcus aureus, CstR (Copper-sensing operon repressor [CsoR]-like sulfurtransferase repressor) is a transcriptional repressor that represses the expression of the cst operon, which encodes a putative sulfide oxidation system, by binding to the OP1 and OP2 sites of the cst promoter (70). Here, CstR and the cst promoter with OP1 and OP2 sites were used to regulate the expression of mkate (encoding a red fluorescent protein, mKate). CstR represses the expression of mkate, but sulfane sulfur can act on CstR and depress the repression. In this way, the fluorescence intensity of mKate could be used to characterize the concentration of intracellular sulfane sulfur. We constructed plasmids with the prxI gene expressed, coupled behind the mKate-encoding gene. The prxI gene was cloned using primers cstR-mkate-prxI-F/R that contained 20-bp extensions overlapping the vector fragment. Then, the segments were connected with the CstR-OP1-mKate vector by using a TEDA assembly. The three cysteines in PrxI were all individually mutated to serine, using primer pairs prxI-C53S-F/R, prxI-C78S-F/R, and prxI-C153S-F/R, by a QuikChange site-directed mutagenesis to assess their roles. Correct CstR reporter plasmids were transformed into E. coli BL21 for experiments.

Construction of PCC7002 mutants.

A prxI mutant of PCC7002 (PCC7002ΔprxI-p) was constructed by homologous recombination as previously reported (24). Briefly, the primer sets prx-del-1/prx-del-2 and prx-del-5/prx-del-6 (Table S2) were used to acquire the upstream and downstream segments of the prxI gene by PCR from the genomic DNA of PCC7002. The lengths of the segments were about 1,000 bp.

The kanamycin resistance cartridge was amplified from pET30a using primers prx-del-3/prx-del-4. Long fragments coupling the upstream segment, the kanamycin resistance cartridge, and the downstream segment were obtained by fusion PCR. The fused fragment was connected with the pJET1.2 blunt vector by using the TEDA method (71), and then the resulting vector was transformed into E. coli DH5α by electroporation. Correct transformants were verified by PCR and by sequencing. The correct plasmid was then transformed into PCC7002 by natural transformation. The final transformants were selected using kanamycin (50 μg/mL) and confirmed by PCR.

Toxicity analysis of sulfane sulfur.

PCC7002 and PCC7002ΔprxI-p in the logarithmic growth phase with an OD730 nm of 0.6 to 0.7 were treated with S8 for 6 h in sealed centrifugation tubes. After incubation, the cells were washed and resuspended in fresh A+ medium. The cells were diluted with A+ medium to an OD730 nm of 0.05, and 10 μL were spread on an A+-agar plate. Differences between strains PCC7002 and PCC7002ΔprxI-p were observed after cultivation at 30°C under continuous illumination by 50 photons m−2 s−1 for about 7 days. Furthermore, the growth curves of PCC7002 and PCC7002ΔprxI-p were also monitored. S8 was added to the medium at the beginning of the culturing.

Phylogenetic analysis.

198 cyanobacterial genomes were downloaded from the NCBI database (updated 17 December 2021). The query sequences of the Prxs were based on reported data (Table S3) (46). The Prx candidates in PCC7002 were analyzed by using ClustalW software for sequence alignment and MEGA 7.0 to build neighbor-joining phylogenetic trees. The parameters were: pairwise deletion, p distance distribution, and bootstrap analysis with 1,000 repeats.

Data availability.

We will provide any strain and materials used in this study upon request.