Sulfane Sulfur Posttranslationally Modifies the Global Regulator AdpA to Influence Actinorhodin Production and Morphological Differentiation of Streptomyces coelicolor

ABSTRACT The transcription factor AdpA is a key regulator controlling both secondary metabolism and morphological differentiation in Streptomyces. Due to its critical functions, its expression undergoes multilevel regulations at transcriptional, posttranscriptional, and translational levels, yet no posttranslational regulation has been reported. Sulfane sulfur, such as hydro polysulfide (HSnH, n ≥ 2) and organic polysulfide (RSnH, n ≥ 2), is common inside microorganisms, but its physiological functions are largely unclear. Here, we discovered that sulfane sulfur posttranslationally modifies AdpA in Streptomyces coelicolor via specifically reacting with Cys62 of AdpA to form a persulfide (Cys62-SSH). This modification decreases the affinity of AdpA to its self-promoter PadpA, allowing increased expression of adpA, further promoting the expression of its target genes actII-4 and wblA. ActII-4 activates actinorhodin biosynthesis, and WblA regulates morphological development. Bioinformatics analyses indicated that AdpA-Cys62 is highly conserved in Streptomyces, suggesting the prevalence of such modification in this genus. Thus, our study unveils a new type of regulation on the AdpA activity and sheds a light on how sulfane sulfur stimulates the production of antibiotics in Streptomyces.

IMPORTANCE Streptomyces species produce a myriad of natural products with (potential) clinical applications. While the database of biosynthetic gene clusters is quickly expanding, their regulation mechanisms are rarely known. Sulfane sulfur species are commonly present in microorganisms with unclear functions. Here, we discovered that sulfane sulfur increases actinorhodin (ACT) production in S. coelicolor. The underlying mechanism is that sulfane sulfur specifically reacts with AdpA, a global transcription factor controlling both ACT gene cluster and morphological differentiation-related genes, to form sulfhydrated AdpA. This modification changes the dynamics of AdpAcontrolled gene networks and leads to high expression of ACT biosynthetic genes. Given the wide prevalence of AdpA and sulfane sulfur in Streptomyces, this mechanism may represent a common regulating pattern of all AdpA-controlled biosynthetic pathways. Thus, this finding provides a new strategy for mining and activating valuable biosynthetic gene clusters.
KEYWORDS sulfane sulfur, AdpA, polyketides, actinorhodin, Streptomyces S treptomyces spp. are Gram-positive bacteria with a filamentous form which colonize a wide range of terrestrial and aquatic niches. The most famous characteristic of Streptomyces is the ability to produce a myriad of secondary metabolites, including antibiotics, antifungals, antivirals, anthelmintic agents, antitumoral drugs, antihypertensives, herbicides, and valuable pigments (1)(2)(3). Much effort has been spent on searching, identifying, and modifying the gene clusters responsible for biosynthesis of these secondary metabolites (4). In contrast, much less energy has been invested in illustrating the transcriptional/translational regulation of these gene clusters. One reason is that Streptomyces have a complex life cycle that includes sporulation, a vegetative or substrate state, and aerial mycelial growth. The biosynthesis of secondary metabolites is closely linked to the stages of the life cycle (5,6), which makes relative studies challenging.
AdpA is a transcriptional regulator universally present in Streptomyces (7). It is located in the second layer of the A-factor-dependent transcriptional network in Streptomyces griseus; the first layer is the A-factor receptor, which activates AdpA expression at the presence of A-factor (g-butyrolactone, a quorum sensing hormone). Therefore, AdpA expression is indirectly controlled by the quorum sensing signal. Aside from A-factor, there are at least four other players in AdpA expression regulation-the master developmental regulator BldD regulating at the transcriptional level (8,9), the cis-antisense RNA regulating at the posttranscriptional level (10), the rare tRNA (tRNA UUA Leu )-encoding gene bldA, and the posttranscriptional tRNA modifications regulating at the translational level (11,12). It was also reported that AdpA can be transcriptionally self-inhibited (13). One reason why regulation of AdpA expression is so complicated is that AdpA is a key regulator of both secondary metabolism and morphological differentiation (14). Considering the critical functions it conducts, whether there are other players regulating at different levels on AdpA expression or activity is unclear but worthy of further investigation.
Sulfane sulfur-containing compounds, such as persulfide (HSSH and RSSH) and polysulfide (HSS n H, S n , RSS n H, RSS n R, n $ 2), are commonly present in both eukaryotic and prokaryotic cells (15). In the past 2 decades, intensive studies of sulfane sulfur have been performed with mammalian cells because it was found that sulfane sulfur is involved in the regulation of diverse physiological and pathological processes, including apoptosis, carcinogenesis, and redox maintenance (16)(17)(18)(19). On the other hand, studies of microorganism sulfane sulfur are traditionally focused on its metabolism and its role in the global sulfur cycle (20,21). In recent years, the physiological functions of sulfane sulfur in microorganisms also got attentions. For instance, Peng et al. (22) found that sulfane sulfur regulates the expression of virulence factors in Staphylococcus aureus, and Liu et al. (23) reported that sulfane sulfur is involved in photosynthesis regulation in Synechococcus. Although they have been noticed, the functions of sulfane sulfur in microorganisms are largely obscure.
In a previous study, we discovered that sulfane sulfur functions as a signal to activate actinorhodin (ACT) production in S. coelicolor M145, a model strain of Streptomyces. In addition, the spore formation process is accelerated by endogenously accumulated sulfane sulfur (24). These phenomena suggest that sulfane sulfur affects both secondary metabolism and the cell cycle in S. coelicolor M145. Based on these findings, we studied the underlying mechanism of how sulfane sulfur performs such functions. We found that AdpA is the key medium of sulfane sulfur signaling. AdpA senses the level of intracellular sulfane sulfur and adjusts ACT production and spore formation. Even the expression of AdpA itself is affected by sulfane sulfur; i.e., sulfane sulfur is a new regulator of AdpA. Thus, this study unveils one way via which sulfane sulfur signals in Streptomyces.

RESULTS
AdpA is a key regulator of ACT production and morphological development in S. coelicolor. Previous studies demonstrated that AdpA is involved in the regulation of ACT production and morphological development in S. coelicolor (25,26). Here, we constructed an adpA-disrupted S. coelicolor M145 strain (DadpA). It exhibited a phenotype of no ACT but high undecylprodigiosin (RED) production when cultured on yeast-beefpeptone (YBP) agar medium (Fig. 1A). Complementary expression of the adpA gene using a plasmid, pMS82-adpA (DadpA::adpA), restored ACT production, while the control, DadpA harboring an empty plasmid (DadpA::pMS82), showed no change. In addition, we noticed that both DadpA and the control showed a bald and nonspore form on YBP medium (Fig. 1A), while DadpA::adpA restored the spore formation. These results verified that AdpA controls ACT production and morphological development in S. coelicolor strain M145.
Sulfane sulfur performing ACT activation requires the presence of AdpA. Since the DadpA strain displayed opposite phenotypes as that of the sulfane sulfur-treated strain (24), we suspected that AdpA had interwound functions with sulfane sulfur. We performed sulfane sulfur induction experiments using the S. coelicolor M145 (wild type [WT]), DadpA, and DadpA::adpA strains. The strains were spread on YBP medium containing 1 mM thiosulfate or 0.1 mM cysteine, which can be converted to sulfane sulfur in vivo (27), and cultured at 30°C for 10 days. For the WT, the production of ACT was significantly increased by thiosulfate/cysteine treatment ( Fig. 1B and C). For DadpA, no Phenotypes of the WT, DadpA, DadpA::adpA, and DadpA::pMS82 strains grown on YBP medium at 30°C. Images were taken at the indicated times. (B) First, 1 mM thiosulfate or 0.1 mM cysteine was added to YBP agar plates before inoculation. The plates were incubated at 30°C for 10 days, and images were captured from the reverse side of the plates. (C) Quantitative determination of ACT produced by the wt, DadpA, and DadpA::adpA strains on YBP containing thiosulfate or cysteine. The plates were incubated at 30°C for 10 days. Data are from three independent repeats. (D and E) WT and DadpA strains were grown on YBP liquid medium. At 36 h, 400 mM HS n H or S 8 was added, and after 1 h of induction, RNA samples were isolated. Real-time PCR data are from three independent repeats and shown as the average 6 standard deviation (SD).
Sulfane Sulfur Modifies AdpA and Affects Its Activity mBio production of ACT was observed with or without thiosulfate/cysteine treatment. For DadpA::adpA, the induction effects were similar to those in the WT ( Fig. 1B and C). These results indicated that AdpA is required for sulfane sulfur to execute the ACT production-activating function. ActII-4 is the ACT production "pathway-specific" activator (26,28). We analyzed transcription of actII-4 using the real-time quantitative reverse transcription-PCR (RT-qPCR) method. The WT and DadpA strains were treated with two sulfane sulfur-containing chemicals, hydrogen polysulfides (HS n H, n $ 2) and sublimed sulfur (S 8 ). For the WT, the transcription level of actII-4 was much higher in the treated strain than that in the untreated one ( Fig. 1D and E), whereas for DadpA, the transcription level of actII-4 had no obvious change after sulfane sulfur treatment ( Fig. 1D and E). These results indicated that sulfane sulfur can increase ActII-4 expression, which subsequently activates ACT production, but this process requires the presence of AdpA.
Sulfane sulfur affects the interaction between AdpA and its cognate promoters. AdpA controls the transcription of actII-4 and wblA (whiB-like gene A, which controls morphological development in S. coelicolor) via binding to their promoters (29). Using these two promoters and an enhanced green fluorescence protein-encoding gene (egfp), we constructed two reporter systems ( Fig. 2A and B). These reporter systems Data are from three independent repeats and shown as the average 6 SD. (C and D) EMSA analysis of the AdpA affinity to P actII-4 promoter DNA (C) and the P wblA promoter DNA (D). All lanes contained 20 nM probe DNA, lanes 2 to 5 contained protein with the indicated concentration, and lanes 3 to 5 contained the HS n H (left) or S 8 (right). The black arrow indicates the free DNA probe, and red arrows indicate the P actII-4 -AdpA or P wblA -AdpA complex.
were introduced into the WT and DadpA strains. HS n H (400 mM) was used to treat the strains containing the reporter systems. After 30 min of treatment, the mycelium was collected by centrifugation, and the fluorescence was read by a fluorophotometer. For the WT strain, HS n H treatment enhanced the strength of both the actII-4 promoter (P actII-4 ) and wblA promoter (P wblA ), evidenced by the increased EGFP expression, whereas, in the DadpA strain, EGFP expression was not increased but decreased after HS n H treatment, indicating that HS n H treatment lost the enhancing effect on these promoters. These results suggested that sulfane sulfur may affect the interaction between AdpA and its cognate promoters.
We then performed electrophoretic mobility shift assays (EMSA) to investigate the interaction. The AdpA protein was expressed in Escherichia coli BL21(DE3) and purified. The DNA probes of the wblA and actII-4 promoters were obtained by PCR. When AdpA was mixed with the P actII-4 or P wblA DNA probe, it bound to them ( Fig. 2C and D). When HS n H (200 mM) or S 8 (200 mM) was also added, the fraction of the AdpA-probe complexes decreased ( Fig. 2C and D). These results indicated that sulfane sulfur decreased the affinity of AdpA to P actII-4 and P wblA .
To test whether the influence of sulfane sulfur can be reversed by a reductant, we added dithiothreitol (DTT) into the mixture of sulfane sulfur (1 mM), AdpA, and the P wblA probe (the DTT dosage was 2-fold of HS n H/S 8 ). After DTT treatment, AdpA restored the high affinity with the P wblA probe (Fig. 3), which had been attenuated by sulfane sulfur. These phenomena demonstrated that the affinity attenuation of AdpA to its cognate DNA caused by sulfane sulfur was reversible.
Sulfane sulfur also affects the transcription of AdpA itself. The adpA gene is transcriptionally self-controlled (30). There are five AdpA binding sites in the P adpA promoter (Fig. 4A). We designed a pair of primers (adpA-wt) from the undeleted part of the adpA gene ( Fig. 4B) and used these primers to analyze the transcription change of adpA in the WT and DadpA. The transcription level of the undeleted part was ;30-fold higher in DadpA than that in the WT, indicating that in the absence of AdpA, the strength of P adpA was higher, i.e., AdpA acted as a repressor for its own transcription (Fig. 4C).
To test whether sulfane sulfur can affect this self-repression, we compared the transcription levels of adpA in the WT stain and the Dpdo strain. In the latter, intracellular sulfane sulfur is accumulated due to a lack of the persulfide oxidation gene (pdo) (24). Results showed that the adpA transcription levels were higher in Dpdo than in the WT (Fig. 4D). We then used exogenous sulfane sulfur to treat the WT strain and found that both HS n H (400 mM) and S 8 (400 mM) can increase adpA transcription ( Fig. 4E and F). EMSA showed that sulfane sulfur (100 to 200 mM) also reduced the affinity of AdpA to P adpA probe, as the unbound probe increased after the addition of HS n H and S 8 (Fig. 5A). Fluorescence polarization (FP) analysis was performed, and the results showed that HS n H (500 mM) obviously increased the K D value (the equilibrium dissociation constant) of AdpA to the P adpA probe, as well as to the P wblA probe ( Fig. 5B and C), indicating that the affinities of AdpA to these promoters were attenuated by HS n H. Simulating the dynamics of the AdpA-controlled promoter system with a simplified model. For the regulation of AdpA on P adpA strength, the logic is easy to understand; P adpA and AdpA compose a classic closed negative-feedback loop. AdpA is a repressor of P adpA . When AdpA is abundant, it binds with P adpA to turn it off/down. The off/down state lasts until the AdpA concentration becomes low due to degradation, and then P adpA turns on/up again. Therefore, without other interference, the strength of P adpA fluctuates, leading to a wave-like expression pattern of AdpA. Since AdpA is an activator of P act-4 , the expression of Act-4 also fluctuates following the concentration wave of AdpA. The principle of these dynamics can be simulated with a simplified mathematical model (Fig. 6A).
For the regulation of AdpA on P actII and P wblA , there is a paradoxical phenomenon; reporter system and RT-qPCR experiments indicated that AdpA enhanced the strength of these two promoters in the presence of sulfane sulfur, but EMSA and FP experiments indicated that sulfane sulfur decreased the AdpA affinities to them. There are several possible reasons for this paradox: 1. There is another player, possibly a transcription factor (TF), involved in this system. This unknown TF takes the place of AdpA in the presence of sulfane sulfur and then further increases P actII /P wblA strength. 2. Sulfane sulfur leads to increased AdpA production, and when the concentration of AdpA is higher than that of sulfane sulfur, free AdpA is more abundant than the sulfane sulfur-modified AdpA. In this case, free AdpA binds to P actII /P wblA and enhances their strength. 3. The affinities of sulfane sulfur-modified AdpA to P adpA , P actII , and P wblA are different.
These differences lead to variations in the transcriptional levels of these genes.
Previously, we observed that the concentration of intracellular sulfane sulfur of S. coelicolor changed along with expression levels of its metabolic genes (24). Therefore, it is highly possible that the ratio of sulfane sulfur to AdpA (S 0 /AdpA) is dynamic (a scenario in item 2). To understand how S 0 /AdpA influences the strength of P actII-4 and P adpA , we developed another mathematical model; at the initial stage, S 0 /AdpA is high, so the sulfane sulfur-reacted AdpA (AdpA-S) is the dominant form (more abundant than apo AdpA), which leads to enhanced expression of adpA but not actII-4 (Fig. 6B). Before S 0 is completely consumed, AdpA is continuously produced, leading to a higher level of AdpA than that in the wave-like expression pattern. However, along with consumption of S 0 , S 0 /AdpA gradually reduces, and finally apo AdpA becomes the dominant form; then the AdpA system returns to its closed negative-feedback loop as shown in Fig. 6A. Based on this simulation, we proposed that sulfane sulfur can temporarily break the self-inhibition in AdpA expression, allowing AdpA to accumulate to a higher level for a longer period (compared with the no-sulfane sulfur scenario), which finally leads to more ActII-4 expression.
The cysteine residue Cys 62 is critical for AdpA sensing sulfane sulfur. Sulfane sulfur can react with cysteine residues of certain proteins to change their configurations (31,32). AdpA contains four cysteine residues, Cys 62 , Cys 126 , Cys 187 , and Cys 307 . To find out which cysteine residue involves in the AdpA-sulfane sulfur interaction, we made a cysteine-to-serine mutation on each cysteine residue of AdpA. The mutated adpA genes were introduced into the DadpA strain. When growing in YBP agar medium, the DadpA::adpA C126S , DadpA::adpA C187S , and DadpA::adpA C307S strains did not show obvious difference from the WT and DadpA::adpA strains. However, the DadpA::adpA C62S strain was distinct from the others. It lost the ability to generate spores, and its ACT production was also apparently lower (Fig. 7A). These phenotype changes indicated that Cys 62 was critical for AdpA performing its regulatory function. EMSA was then performed to examine whether the C62S mutation affects the binding of AdpA to its cognate promoter. AdpA C62S still bound to the P wblA DNA fragment, and two main P wblA -AdpA C62S complexes with different molecular weights (MW) were observed. The complex with lower MW was no longer influenced by sulfane sulfur even when sulfane sulfur was added at high concentrations (.1,000-fold higher than that of AdpA C62S ) (Fig. 7B), whereas the complex with higher MW disappeared when high concentrations of sulfane sulfur were added. Since the lower-MW complex was the most abundant one formed by AdpA C62S and the P wblA DNA fragment, we proposed that the sulfane sulfur sensing ability was at least partially impaired by the C62S mutation.
To check how sulfane sulfur reacts with AdpA, purified AdpA was treated with HS n H (200 mM) or DTT (200 mM). The treated-AdpA was labeled with iodoacetamide (IAM) and then subjected to trypsin digestion, followed by LTQ-Orbitrap tandem mass spectrometry analysis. For the HS n H-treated AdpA, two peptides (1 and 2, Fig. 8) were identified. In peptide 1 (1,299.67 Da), the Cys 62 residue was directly blocked by IAM to form Cys 62 -AM (acetamide) (Fig. 8 and Fig. S1). In peptide 2 (1,331.64 Da), a mass increase of 32 (132 MW) was identified. A secondary mass spectrometry(MS 2 ) spectrum indicated that the 132 MW happened on the thiol group of Cys 62 to form peptide-S-AM ( Fig. 8 and Fig. S2). The MS 1 signal intensity ratio of peptide 1/peptide 2 was 17%. As the control, only a peptide with Cys 62 -AM (1,299.67 Da, peptide 3) was identified from DTTreacted AdpA, corresponding to a direct blockage of IAM on the Cys 62 residue (Fig. 8   FIG 6 Modeling principles of how AdpA regulates adpA and actII-4 expression. (A) In the absence of sulfane sulfur, P adpA is self-repressed by AdpA to form a negative-feedback loop, and hence, both P adpA strength and AdpA amount show a wave-like pattern. (B) Sulfane sulfur temporarily breaks the negative-feedback loop, which leads to a higher and longer expression of AdpA but not ActII-4 at the initial stage. After sulfane sulfur is consumed, the high AdpA level will lead to high expression of ActII-4. Equations and related parameters used for modeling are provided in Text S1. and Fig. S3). These results indicated that sulfane sulfur can modify Cys 62 -SH to form Cys 62 -SSH.
The thiol group of Cys62 is accessible to solution due to its location on the AdpA 3D structure. The 3D structure of AdpA was modeled by using AlphaFold 2 (https://www.hpc.caltech.edu/documentation/software-and-modules/alphafold-2). The crystal structure of a truncated AdpAsg containing only the DNA binding domain, which is from Streptomyces griseus, is available in the PDB database (PDB: 3w6v). We aligned AdpAsg with the modeled AdpA, and the alignment parameter RMSD was 0.365, indicating a high confidence of the predicted structure of AdpA ( Fig. 9A and B).
We then analyzed the locations of the four cysteine residues in AdpA (Fig. 9C).   Fig. S1 to S3. Sulfane Sulfur Modifies AdpA and Affects Its Activity mBio fold into the interior of AdpA, and hence they are protected from sulfane sulfur attack. In contrast, Cys 62 is located near the protein surface, and its thiol group is exposed to solution, which may explain why Cys 62 can be modified by sulfane sulfur. It is noteworthy that Cys 126 is located on the interior surface of a tunnel through the dimerization domain. Therefore, sulfane sulfur compounds may not enter this tunnel to react with its thiol group. In addition, the distance between any two cysteine residues is too far to form a disulfide (S-S) or tri-sulfide (S-S-S) bond. The cysteine residues are conserved in Streptomyces AdpAs. We analyzed AdpA and its homologues in the Streptomyces genus. In the PATRIC database, 2,752 of 3,033 sequenced Streptomyces strains contain AdpA, accounting for a 90.73% prevalence. We selected some representative AdpA sequences to construct a phylogenetic tree. The results revealed that AdpA homologues were not on one evolutionary branch (Fig. S4). However, when we performed multiple sequence comparisons with them, we found that their four cysteine residues were highly conserved, including Cys 62 (Fig. S5). These results suggested that using cysteine residues to sense sulfane sulfur may be a common mechanism for AdpA functioning in Streptomyces.

DISCUSSION
The global transcription factor AdpA plays an important role in regulation of secondary metabolism and morphological development in the Streptomyces genus (33)(34)(35)(36)(37). Its own expression is controlled by multiple factors. In this study, we discovered that sulfane sulfur affects AdpA activity via the posttranslational modification. After reacting with sulfane sulfur, the affinity of AdpA to its cognate promoters, P adpA , P actII-4 , and P wblA , is attenuated. We constructed a simplified model to help understand the Sulfane Sulfur Modifies AdpA and Affects Its Activity mBio effect of sulfane sulfur on the AdpA-controlled promoters. As shown in our simulation (Fig. 6), P adpA is under the control of a negative feedback loop of self-repression. Without the presence of sulfane sulfur and/or other disturbing factors, activation of AdpA on actII-4 and wblA expression cannot last long due to the negative feedback. Sulfane sulfur modifies AdpA to temporarily break the self-repression, and hence, AdpA can accumulate to a higher level for a longer time until sulfane sulfur is consumed. The accumulated AdpA finally activates expression of actII-4 and wblA. Thus, the effect of sulfane sulfur on the AdpA regulon may represent a fine-tuned regulation for the production of antibiotics and morphological development. Furthermore, we found that Cys 62 is critical for AdpA sensing sulfane sulfur. Our in vitro experiments showed that sulfane sulfur treatment can lead to a sulfhydration modification in Cys 62 (Cys 62 -SSH), and this modification was not observed in the other three cysteine residues of AdpA. A limitation of this work is that such modification has not been examined in vivo due to the lack of a trustable method. The AlphaFold 2 predicted structure shows that the thiol of Cys 62 is accessible to solution, while the other thiols are not, which may explain why Cys 62 is easily sulfhydrated by sulfane sulfur. However, we also noticed that AdpA and AdpA C62S formed different complexes with the P wblA DNA fragment even in the absence of sulfane sulfur, suggesting that the binding pattern was affected by C62S mutation. Since Cys 62 is located in the dimerization domain, it may affect the dimerization and then alter the AdpA binding pattern. Therefore, C62S mutation may result in multiple influences, including both DNA binding and sulfane sulfur sensing.
It is noteworthy that the AdpA complemented strain (DadpA::adpA) shows a peculiar pattern of ACT production-the center of the plate lacks the characteristic blue color-in the plates containing 0.1 M cysteine (Fig. 1B). A similar pattern was described in a recent report (38), and the authors linked the pattern formation to AdpA. Therefore, the peculiar pattern observed in our experiment may be caused by both AdpA expression alteration and cysteine addition. In the same report the authors discovered that expression of adpA and other genes controlled by it displayed a spatiotemporally separated wave-like pattern when S. coelicolor was cultured in solid medium and found that this pattern was driven by a combination of physiological gradients and regulatory network architecture (38). The finding is consistent with our simulation. From a genetic architecture viewpoint, the negative feedback loop inevitably leads to wave-like expression of adpA. However, the frequency (or wavelength) of the "wave" can be altered by environmental factors such as sulfane sulfur or siderophore. Understanding how the pattern forms and its determinants surely are important for interpreting the complicated differentiation process of Streptomyces and hence worth further study.
Streptomyces mainly exist in terrestrial soils, but they also have been detected in extreme environments such as deep seas, the north and south poles, hydrothermal fluids, hot springs, etc. In some environments (such as sea/lake bed) sulfane sulfur levels can be high, up to ;400 mM. Therefore, the possibility that Streptomyces live in sulfane sulfur-rich conditions cannot be excluded. In addition, sulfane sulfur has been recognized as a common intracellular chemical nowadays, and its concentration varies from 10 mM to ;500 mM (39). Hence, both self-produced and environmental sulfane sulfur may affect secondary metabolism and morphological development of Streptomyces.
In recent years, a few transcription factors that can be modified by sulfane sulfur have been identified from different microorganisms (40,41). These transcription factors can be categorized into two groups. Group I consists of specific regulators for genes related to sulfur metabolism, including BigR (42), CstR (43), FisR (44), CsoR (24), and SqrR (45,46). They control the expression of sulfane sulfur oxidation enzyme PDO and sulfane sulfur transferase RhoD (BigR in this case). Since H 2 S oxidation enzyme SQR is often located in the same operon with PDO and RhoD, they also control SQR expression (CstR, FisR, and SqrR in this case) (47). Group I regulators can sense the intracellular level of sulfane sulfur via their cysteine residues; when the sulfane sulfur level is high, cysteine residues are sulfhydrated to form an RS n H (n $ 2) or RS n R (n $ 3) bond, which leads to a configuration change of the regulator and subsequently high expression of PDO and other genes, and then the sulfane sulfur level is decreased through being oxidized to sulfite (20). Therefore, group I regulators mainly function as managers to maintain the homeostasis of intracellular sulfane sulfur.
Group II includes global or multifunctional transcription factors currently including MgrA (22), MexR (48), and OxyR (40). MgrA is a global virulence regulator of Staphylococcus aureus. It senses the intracellular level of sulfane sulfur to regulate the expression of virulence factors (22). MexR controls the multiple-antibiotic resistance process in Pseudomonas aeruginosa, and it senses intracellular sulfane sulfur to regulate the expression of the mexAB-oprM multidrug efflux operon (48). OxyR is a global antioxidation regulator in many bacteria. Recently, it was found that OxyR also senses sulfane sulfur and controls the expression of sulfane sulfur-reducing enzymes (40). Like group I, group II regulators also sense sulfane sulfur via their cysteine residues.
AdpA is deemed a group II regulator since it senses sulfane sulfur and accordingly adjusts the ACT production and spore formation in S. coelicolor. Bioinformatics analyses indicated that AdpA and its Cys residues are highly conserved in Streptomyces spp. Further investigation of this protein and its homologues should provide insights into how sulfane sulfur regulates the production of secondary metabolites and morphological developments in this genus. The widespread existence of AdpA implies that sulfane sulfur may play a wide range of regulatory functions in Streptomyces, providing unlimited possibilities for sulfane sulfur working as a signal molecule to stimulate increased production of important secondary metabolites, such as antibiotics, antitumor drugs, immunosuppressants, and antibiotics.

MATERIALS AND METHODS
Bacterial strains, plasmids, and growth conditions. All strains and plasmids used or constructed in this work are summarized in Table S1.
Preparation of sulfane sulfur species and other sulfur-containing compounds. Sodium hydrosulfide (NaHS, H 2 S donor), cysteine, sulfur power, and thiosulfate were purchased from Sigma-Aldrich. S 8 solution was prepared by dissolving excess sulfur powder in acetone to saturation. The concentration of saturated acetone sulfur is determined as 17 mM as reported previously (51). The stock solution of HS n H was prepared by mixing sulfur powder, NaOH, and NaHS (40 mM each chemical) in degassed distilled water at 30°C until the powder was completely dissolved as previously described (48,52). The concentrations of HS n H were determined with the cyanolysis method (53) and calibrated by using thiosulfate as the standard. Specifically, pipetting 550 mL 1% boric acid into a 1.5-mL Eppendorf (EP) tube and removing dissolved oxygen by putting the EP tube in boiling water for 1 min and then adding 250 mL sample and 200 mL 1 M potassium cyanide. After boiling in a water bath (100°C) for 1 min, the EP tube was taken out and cooled down to room temperature, and 100 mL ferric nitrate color solution was added to form Fe(SCN) 3 . The A 460nm absorbance value was detected. Thiosulfate was used to make a standard curve.
Construction of S. coelicolor DadpA. All primers used in this experiment are listed in Table S2. The strain DadpA was constructed using a homologous recombination method (54). Briefly, a 939-bp region was deleted from the open reading frame (ORF) of adpA, leaving the upstream 168 bp (relative to the start codon) and the downstream 90-bp (relative to the stop codon) coding sequence of adpA. The knockout region was replaced by the apramycin resistance gene. The conjugation transfer was accomplished using the methylation-sensitive strain E. coli ET12567/pUZ8002 (containing the mutant plasmid pJTU-adpA) and S. coelicolor M145 following a previously reported protocol (55). The deletion mutant was verified by resistance screening and colony PCR with the primers VeradpA-F/R. Construction of DadpA::adpA, DadpA::pMS82, DadpA::adpA C62S , DadpA::adpA C126S , DadpA:: adpA C187S , and DadpA::adpA C307S . A DNA fragment carrying the adpA ORF (1,197 bp) and its promoter (500 bp) was obtained using PCR amplification and was connected to the UBT1 integrative vector pMS82 (56) to generate pMS82-adpA plasmid (Table S1). This plasmid was then integrated into the attP site of the DadpA genome by intergeneric conjugation. To construct the negative-control strains, empty pMS82 vector was also transformed into DadpA; these derivative strains were selected and confirmed by PCR and DNA sequencing.
AdpA protein overexpression, purification, and mutation. To construct the AdpA expression strain, the coding sequence of adpA was amplified from WT genomic DNA with the primers ExadpA-F/R. The PCR product was purified and ligated into the pET15b vector with a C-terminal His tag to create plasmid pET-AdpA by using the ClonExpress II one-step cloning kit (TaKaRa). The plasmid was transformed into E. coli BL21(DE3) cells, which were grown in LB medium at 37°C to an optical density at 600 nm (OD 600 ) of 0.6, and then a total of 0.5 mM isopropyl b-D-1-thiogalactopyranoside (IPTG) was added and an additional overnight cultivation was continued at 16°C. Cultures were collected by centrifugation and disrupted though a pressure cell homogenizer (SPCH-18) in sonication buffer (50 mM NaH 2 PO 4 , 250 mM NaCl, 20 mM imidazole, pH 8.0); 1 mM DTT was added before breaking the cells. Purification of the AdpA His-tagged proteins was performed with a Ni-NTA-Sefinose column (Sangon) as described previously (24). The protein purification process was conducted in an anaerobic glove box, which was filled with mixed gas (N 2 , 85%; H 2 , 10%; CO 2 , 5%). The purity of the protein was assessed by SDS-PAGE gel, and its concentration was determined using the bicinchoninic acid (BCA) protein assay reagent (Thermo Fisher Scientific). The same method was used for purification of AdpA mutants.
Electrophoretic mobility shift assay (EMSA). The DNA probes containing AdpA binding sequences were amplified from genomic DNA. Different sulfane sulfur compounds were reacted with purified AdpA (and its mutants) in the binding buffer (20 mM Tris-HCI, 2 mM EDTA, 20 mM KCI, 0.5 mM dithiothreitol [DTT], pH 8.0) at room temperature for 20 min. Then DNA probe was added, and the binding reaction was performed at 30°C for 20 min. The binding complexes were separated on an 8% nondenaturing polyacrylamide gel at 120 V for 2 h in ice (58). The gel was dyed with SYBR green I (Sangon) for 20 min (44). All images were captured with a FluorChemQ system (Alpha Innotech).
RNA preparation, RT-PCR, and RT-qPCR. To extract RNA, spores (2 Â 10 7 ) of WT and DadpA strains were inoculated into the liquid YBP medium and incubated at 30°C with shaking (220 rpm) for 36 h to the mid-exponential phase. HS n H (400 mM) or S 8 (400 mM) was added. After another 30-min cultivation, these mycelia were harvested and ground into powder with liquid nitrogen. Similarly, the cultures of WT, DadpA, and Dpdo were collected at the indicated times. All RNAs were isolated with a SteadyPure universal RNA extraction kit (Accurate Biology) following the manufacturer's instructions, and their quality and concentration were determined using a NanoDrop ND-1000 device (Thermo Fisher). RT-PCR was carried out using a reverse transcriptase kit (Invitrogen) and SYBR premix Ex Taq (TaKaRa) following the manufacturers' recommendations. The Roche LightCycler 480 thermal cycler was used (59). The expression of hrdB mRNA was used as the internal standard to normalize the relative quantities of cDNA. The relative expression abundance of the target gene was analyzed using a relative quantification method (2 DCT , test gene-hrdB). Three independent replicates were performed.
Phenotypic analysis and ACT production assay. S. coelicolor strains were cultured on solid YBP medium at 30°C for phenotypic analysis. ACT production was determined following a previously reported method (24,60,61). Briefly, Streptomyces strains were incubated on YBP medium for 7 or 10 days, and mycelia were harvested from the plate. KOH (1 M final concentration) was added to treat the mycelia for 4 h. Then the mixtures were centrifuged. The ACT concentration in the supernatant was determined by a spectrophotometer. Three independent biological experiments were replicated.
Construction and testing of EGFP reporter systems. To construct the reporter plasmids, promoter fragments (2400 to 21 upstream of actII-4 and 2460 to 21 upstream of wblA) were amplified using primers pMS82-actII-4p-egfp S1-F/R and pMS82-wblAp-egfp S1-F/R (Table S2). Then these promoter fragments and a DNA fragment encompassing the egfp gene were cloned into the pMS82 vector to generate pMS82-actII-4p-egfp and pMS82-wblAp-egfp. Next, we introduced these reporter plasmids into the WT and DadpA.
Strains containing reporter plasmids were precultured in liquid YBP medium for 36 h at 30°C. Subsequently, equal amounts mycelia of each strain were transferred to the fluted bottle, and inducer (400 mM HS n H or 400 mM S 8 ) was added. After 60 min of induction, the bacteria were collected by centrifugation, and mycelia were resuspended in 200 mL phosphate-buffered saline (PBS) buffer (OD 450 , 2). EGFP fluorescence was measured using the microplate reader Synergy H1. The excitation wavelength and emission wavelength were set to 485 nm and 515 nm, respectively. The EGFP fluorescence intensity was normalized against cell density (fluorescence/OD 450 of mycelia).
LC-MS/MS analysis of AdpA. The analysis was performed following a previous report (24). Freshly purified protein AdpA (,100 mg) was treated with 10-fold amounts of HS n H (200 mM) or DTT (200 mM). After reacting at room temperature for 40 min. The reacted protein was treated with denaturing buffer (0.5 M Tris-HCl, 2.75 mM EDTA, 6 M guanidine-HCl, pH 8.0) containing 1 M iodoacetamide (IAM). The treatment was carried out in the dark for 1 h, and then the sample was digested with trypsin (1:25, wt/ wt) at 37°C for 20 h. The digestion products were filtered by C 18 Zip-Tip (Millipore) and vacuum-dried. The obtained peptides were resuspended in 10 mL double-distilled water (ddH 2 O).
The Prominence nano-LC system (Shimadzu) equipped with a custom-made silica column (75 mm by 15 cm) packed with 3 mm ReproSil-Pur 120 C 18 -AQ was used. Positive electrospray ionization was performed, and the ions were scanned with an LTQ-Orbitrap Velos Pro CID mass spectrometer (Thermo Scientific); the data were analyzed using a data-dependent acquisition mode with Xcalibur 2.2.0 software (Thermo Scientific). Full-scan MS spectra (from 400 to 1,800 m/z) were detected and assessed with the Orbitrap at a resolution of 60,000 at 400 m/z.
AdpA structure modeling. The AlphaFold 2 algorithm (62) was used to predict the tertiary structure of AdpA. This method used the custom multiple sequence alignment (MSA) option and was accessed via the Colab server on GitHub (https://github.com/sokrypton/ColabFold). The structural model of AdpA was analyzed and visualized with PyMOL.
Fluorescence polarization (FP) analysis. FP analysis experiments were performed following a reported protocol (63). DNA probes were amplified by PCR and labeled by 596-FAM (carboxyfluorescein) (Sangon). Purified AdpA (treated with 1 mM HS n H for 10 min or not) was diluted to different concentrations (0.01 mM to ;22.5 mM). The reaction buffer contained 10 mM Tris-HCl and 75 mM NaCl, pH 7.5. After mixing diluted AdpA and labeled DNA in the reaction buffer, the solution was incubated at 37°C for 15 min in the dark. The fluorescence was detected with a BioTek Synergy HT instrument. The K D value was calculated using GraphPad Prism 5 software.

SUPPLEMENTAL MATERIAL
Supplemental material is available online only. TEXT S1, PDF file, 0.3 MB.