Reactive sulfur species mediated persulfidation cascade regulation mechanism of novel transcription factor SscR^Ac in extremophile Acidithiobacillus caldus under extreme copper stress

Abstract

Severe environmental conditions enhance the resilience of biomining microorganisms to elevated metal ion concentrations. The mechanism of how biomining microorganisms resist metal ions is poorly understood. We identified a novel reactive sulfur species (RSS)-sensitive MarR family transcription factor (SscR^Ac) in Acidithiobacillus caldus by persulfidation proteomics and observed increase in RSS and protein persulfidation levels under 250 mM Cu²⁺ stress. The deletion of sscR^Ac gene via CRISPR–Cas9 and conjugative transfer technology enhanced copper sensitivity in A. caldus. ChIP-seq/qRT-PCR revealed that SscR^Ac regulates copper detoxification by blocking efflux pumps and stimulating RSS metabolism. LC-MS/MS analysis revealed that both Cys74 and Cys78 in SscR^Ac interact with RSS and undergo persulfidation, resulting in the dissociation of the protein from the promoter–DNA of target genes. Upstream signaling analysis indicated that copper-sensitive repressor CsoR^Ac, regulated by SscR^Ac, inversely regulates SscR^Ac, thereby jointly enabling copper–RSS signal transduction. In conclusion, we identify SscR^Ac as the first RSS-dependent transcriptional switch directly linking copper toxicity with the persulfidation signaling pathway in extremophiles.

Graphical Abstract

Graphical Abstract.

Introduction

The depletion of high-grade ore sources, coupled with persistent industrial demand growth, has heightened interest in biomining for metal extraction from low-grade ores [1]. Acidithiobacillus caldus, the predominant strain in biomining environments, is essential to the biomining process. As a chemolithotrophic bacteria, A. caldus predominantly utilizes sulfur metabolism for energy generation. The sulfur oxidation process generates sulfuric acid, which aids in preventing the formation of passivation layers on mineral surfaces [2–4]. Prolonged exposure to elevated levels of metal ions facilitates the development of metal tolerance in A. caldus. Through long-term adaptive evolution, A. caldus can survives under 250 mM copper stress [5]. This suggests that A. caldus possesses significant application potential, as the concentration of Cu²⁺ in the environment will progressively rise throughout the later stages of bioleaching as the ore dissolves. For instance, in leachate of heap bioleaching, the concentration of Cu²⁺ typically ranges from 2 to 6 g/l (30–90 mM). In the stirred tank reactor utilized for processing sulfide concentrates, the copper concentration may attain 19 g/l (∼300 mM) or exceed this level [6]. The remarkable copper-resistant adaptability of A. caldus designates it as an exemplary model organism for investigating microbial metal resistance mechanisms. Therefore, a thorough analysis of A. caldus tolerance mechanisms under extreme copper stress holds considerable significance for comprehending biological limits and variety within extremophilic systems.

Like neutrophilic bacteria such as Escherichia coli, A. caldus employs both active and passive mechanisms for heavy metal protection [7]. Active mechanisms primarily include Cop-system ATPase efflux pumps, metallochaperones, and proton gradient driven resistance-nodulation-division (RND) systems [8]. Inorganic polyphosphate polymers may enhance efflux systems via chelation-mediated extracellular copper transport [7]. Passive processes primarily depend on extracellular polymeric substance (EPS) barriers and the protection of membrane potential [9]. The EPS-formed physical barrier reduces copper ion permeability [10], whereas the acidophile-specific electrochemical gradient generated via membrane potential polarization effectively inhibits metal influx [11]. Although these mechanisms partially explain microbial copper tolerance in biomining systems, the genetic and molecular basis underlying their efficient coordination under extreme environmental stress remains poorly characterized. Genomic analysis reveals that A. caldus ATCC 51756 contains a 13.4-kb copper resistance gene cluster (ACAty_RS08430-ACAty_RS15375) comprising genes that encode copper detoxification proteins and potential transcriptional regulators [5]. Notably, the upstream region of this cluster contains several genes that encode thiol (SH)-related proteins, such as glutathione S-transferase (GST, ACAty_RS08500), protein disulfide reductase (ACAty_RS08525), and sulfur dioxygenase (SDO, ACAty_RS08530). Homologs of these proteins frequently participate in reactive sulfur species (RSS)-mediated persulfidation processes during microbial detoxification pathways [12–14], with SDO being directly involved in RSS metabolic regulation [15]. Excessive copper and other metal ions are known to catalyze the production of reactive oxygen species (ROS). Endogenous hydrogen sulfide (H₂S) mitigates oxidative damage induced by the overaccumulation of ROS in cells [16]. Recent studies demonstrate that H₂S biological effects are fundamentally driven by the potent nucleophilic properties of its metabolic derivatives RSS [14, 17]. Despite the substantial disruption of intracellular RSS homeostasis caused by copper stress, the prevailing agreement asserts that copper stress signaling and RSS signaling in microorganisms are governed by two separate systems. The classic example pertains to copper-sensitive operon repressor (CsoR) family proteins, classified as CsoR [18] and CsoR-like sulfur transferase repressor (CstR) [19] according to their unique stimulus recognition capacities. CsoR and CstR in Staphylococcus aureus act as paralogs within the same cell, independently regulating copper homeostasis and sulfur metabolism [19]. As of now, no significant regulatory elements directly connecting copper stress signaling to persulfidation have been found in biomining microorganisms.

A. caldus, as an inorganic sulfur compound (ISC)-oxidizing acidophile, with a complex sulfur metabolic network [20]. Our research team earlier isolated a copper stress-adapted strain, A. caldus CCTCC M 2018727 (called A. caldus M), using adaptive laboratory evolution under copper selection pressure. Transcriptomic research demonstrated substantial activation of sulfur metabolic pathways in A. caldus M under 3 g/l (≈47 mM) copper stress [21], which is essential for bacterial survival in extreme conditions. Two copies of the SDO-encoding genes, sdo1 (ACAty_RS08530) and sdo2 (ACAty_RS10625), demonstrated significant transcriptional downregulation [21]. In mammalian systems, the SDO homolog persulfide dioxygenase (PDO) cooperate with sulfide: quinone oxidoreductase (SQR) to facilitate the oxidation of H₂S, thus alleviating H₂S toxicity [22]. In A. caldus, SDO and SQR mediate catalytic reactions as shown in Equations (1) and (2) [23]. Transcriptional suppression of sdo1/sdo2 under copper stress may lead to RSS accumulation (e.g. glutathione persulfide, GSSH). Therefore, by augmenting the nucleophilic reaction's capacity to neutralize deleterious chemicals, one might diminish the toxicity of metals.

(1)

(2)

Here, we focused on the extremely copper tolerant strain A. caldus M, identifying a novel regulatory element and a copper–RSS signaling pathway that links the copper homeostasis maintenance abilities of A. caldus M with RSS metabolism. This study confirmed markedly increased RSS levels in A. caldus M under severe copper stress (250 mM). To comprehend the strategies for responding to copper stress mediated by RSS and to clarify their regulatory mechanisms under extreme copper stress, we examined the differences in protein persulfidation before and after stress using post-translational modification proteomics (results interpreted with reference to the model strain A. caldus ATCC 51756 whole-genome sequence). A MarR-family transcription factor (ACAty_RS16455) linked to copper resistance and RSS metabolism was identified. Given its sensitivity to RSS, we called it SscR^Ac (RSS-sensitive transcription factor in copper stress response). We conducted genetic and biochemical investigations to examine the functional role of SscR^Ac in A. caldus M during adaptation to copper stress. The results indicated that SscR^Ac improves copper resistance in A. caldus M by simultaneously controlling copper ion efflux and RSS metabolism. Furthermore, RSS disrupted SscR^Ac binding to specific promoter DNA. We established a copper–RSS signaling path through the functional coupling between SscR^Ac and the copper-sensitive transcription factor CsoR^Ac, wherein CsoR^Ac and SscR^Ac work together to transform and amplify copper stress signals into RSS signals that are detectable by extensive protein networks.

Materials and methods

Strains, culture media, and growth conditions

All bacterial strains, plasmids, and primers used in this study are detailed in Supplementary Tables S1 and S2. A. caldus strain CCTCC M 2018727 was isolated from Zijin Copper Mine (Fujian, China) and was being adaptively domesticated to survive under 250 mM CuSO₄. The strain was cultured at 40°C and 180 rpm in modified Starkey medium supplemented with trace elements, as previously described [24]. E. coli strains were grown in LB medium at 37°C and 200 rpm, with antibiotics (50 μg/ml kanamycin or 100 μg/ml ampicillin) added as required. Since the genome sequencing of A. caldus CCTCC M 2018727 has not been completed, genomic data from the model strain A. caldus ATCC 51756 were used as the reference. All genomic DNA and plasmids were extracted using commercial kits from Vazyme Biotech Co., Ltd.

Persulfidation proteomics

This study utilized a label-free quantitative persulfidation proteomics approach [25]. A. caldus M was cultured in modified Starkey liquid medium until reaching the mid-log phase (biomass, 0.5 × 10⁸ cells·ml^-1), then resuspended and adjusted to a biomass concentration of 1 × 10⁶ cells ml^-1 before being reinoculated into fresh Starkey medium. The experimental group was supplemented with anhydrous CuSO₄ to a final copper ion concentration of 250 mM. Upon reaching the stationary phase, three biological replicates from both the control and experimental groups were collected. Culture aliquots were subjected to low-speed centrifugation (367 × g, 4°C, 3 min) to remove solid particles, followed by centrifugation at 3000 × g (4°C, 5–10 min) to pellet cells, which were washed three times with ice-cold phosphate buffered saline (PBS). Subsequent protein digestion, peptides enrichment, and identification, as well as analysis were performed by Shanghai Bioprofile Biotechnology. Raw data were analyzed using MaxQuant 1.6.1.0 against the A. caldus ATCC 51756 reference genome.

Preparation of H₂S_n and determination of RSS

The preparation of H₂S_n followed a previously reported method with minor modifications [26]. Briefly, sulfur powder, NaOH, and NaHS were mixed in a 1:1:1 molar ratio, dissolved in distilled water under an argon atmosphere, and sealed in a bottle incubated at 37°C until complete dissolution of sulfur. For RSS detection via the sulfite method [27], reaction buffer (50 mM Tris–HCl pH 9.5, 1% Triton X-100, 50 μM DTPA, 1 mM sulfite) and control buffer (without sulfite) were prepared. Samples were separately mixed with reaction or control buffer, incubated at 95°C for 10 min to convert RSS to thiosulfate. A 50 μl aliquot of each sample was treated with 5 μl of 25 mM monobromobimane (mBBr) in acetonitrile, incubated in the dark at room temperature for 30 min to derivatize thiosulfate into thiosulfate-bimane. Proteins insoluble in organic solvents were precipitated by adding an equal volume of acetic acid/acetonitrile (1:9, v/v), followed by centrifugation at 16 000 × g for 3 min to remove debris. The supernatant was analyzed by HPLC with fluorescence detection. The RSS content was calculated as the difference in thiosulfate concentrations between the reaction and control buffers.

Protein expression, purification, and reaction with H₂S_n

The wild-type SscR^Ac and its mutants containing an N-terminal His-tag were expressed in E. coli BL21(DE3) harboring the pET28a vector carrying the sscR^Ac gene or its mutant variants. Cells were cultured in LB medium supplemented with 50 μg/ml kanamycin at 37°C until reaching an OD_{600 nm} of 0.6, followed by induction with 0.5 mM isopropyl-β-d-thiogalactopyranoside (IPTG). Protein expression was performed at 16°C for 20 h. Harvested cells were centrifuged at 8 000 × g (4°C, 10 min), washed with Buffer A (50 mM Tris, 300 mM NaCl, pH 7.8) containing 1 mM EDTA, and resuspended in Buffer A supplemented with 1 mM dithiothreitol (DTT). Cell lysis was achieved via ultrasonication for 30 min, followed by filtration through a 0.22-μm membrane. The clarified lysate was loaded onto a 5-ml HisTrap nickel-NTA FastFlow column pre-equilibrated with Buffer A. Recombinant proteins were eluted using a linear imidazole gradient on an ÄKTA avant protein purification system, followed by further purification via a G-25 column equilibrated with Buffer B (20 mM Tris, 100 mM NaCl). Protein purity was verified by 15% Tricine–SDS–PAGE, and concentrations were determined using the Coomassie brilliant blue assay. The CsoR^Ac protein was purified using the same protocol. For protein–H₂S_n interaction studies, reactions were conducted in an anaerobic glove box with a molar ratio of 1:10 (reduced SscR^Ac or CsoR^Ac:H₂S_n). After 30-min incubation, unreacted H₂S_n was removed by dialysis. The reacted proteins were sealed and transferred out of the glove box for subsequent experiments.

Constructing mutants of SscR^Ac

Mutants of SscR^Ac were generated using a Mut Express II fast mutagenesis kit V2 (Vazyme). Cys74, 78 were selected as the mutation sites, and mutants were generated by polymerase chain reaction (PCR) using pET-28a-SscR^Ac as the template (primers see Supplementary Table S2). After PCR and DNA purification, the linearized pET-28a vector was digested by DpnI and cyclized by Exnase II (Vazyme). E. coli BL21(DE3) competent cells were transformed with the generated vectors and were confirmed by DNA sequencing analysis. The methods of expression and purification for SscR^Ac mutants were described as above.

Size-exclusion chromatography and circular dichroism spectroscopy

The purified SscR^Ac (25 μM) in 20 mM sodium phosphate buffer (pH 7.6) was incubated with 200 μM H₂S_n for 20 min under anaerobic conditions prior to size-exclusion chromatography (SEC) analysis. Untreated and H₂S_n-treated protein samples were analyzed using an ÄKTA avant system equipped with a Superdex-75 10/300 GL column. Proteins were eluted with 20 mM Tris–HCl (pH 7.8) at a flow rate of 0.5 ml/min. Molecular weight estimation was performed using a standard curve generated from bovine serum albumin (67.0 kDa), ovalbumin (45.0 kDa), chymotrypsin (22.5 kDa), and cytochrome c (12.4 kDa). The secondary structure of SscR^Ac was analyzed using a MOS-450 circular dichroism (CD) spectrometer (Biologic, France). Spectral data were collected in the 190–250 nm wavelength range with a continuous scanning mode at 100 nm/min. The control buffer (20 mM Tris–HCl, pH 7.8) baseline was subtracted from sample measurements. The average residue molar ellipticity ([θ], in millidegrees) was calculated from the raw CD data using the instrument’s detected values and the protein’s residue concentration.

Construction of CRISPR–Cas9 system

The broad-host-range plasmid pJRD215 was used as the backbone to construct the CRISPR–Cas9 knockout vector [28]. To enable antibiotic selection, a streptomycin resistance gene (Sm^r) was first introduced into pJRD215, generating the modified plasmid pJRD-Sm^r. Since the donor strain E. coli SM10 carries a chromosomal kanamycin resistance gene (Km^r), Sm^r was selected for recipient screening. The guide RNA (gRNA) target sequence (20 bp) specific to the sscR^Ac gene was designed using Cas-Designer (http://www.rgenome.net/cas-designer). The N20 spacer sequence in the original pTarget plasmid was replaced with the sscR^Ac-specific N20 sequence. The Cas9 gene and homologous arms (upstream/downstream flanking regions of sscR^Ac) were amplified from the pCas9 vector and A. caldus CCTCC M2018727 genomic DNA, respectively. The cysteine mutation sequences were amplified from the pet-28a-C74S and pet-28a-C78S plasmids. Promoters, terminators, and single-guide RNA (sgRNA) fragments were synthesized by Genewiz. These components were assembled via overlap extension PCR. The fused fragments were ligated into the XbaI/HindIII-digested pJRD-Sm^r plasmid using a one-step cloning kit, yielding the CRISPR–Cas9 knockout plasmid pJRD-ΔsscR^Ac and cysteine mutant plasmid pJRD-p-C74S, pJRD-p-C78S. The plasmid was chemically transformed into E. coli SM10 competent cells. A modified conjugation method [29] was employed to transfer pJRD-ΔsscR^Ac from E. coli SM10 to A. caldus M. Donor cells were grown in LB medium with antibiotics at 37°C to late exponential phase, while recipient cells were cultured in modified Starkey medium to stationary phase. Cells were pelleted by centrifugation and washed twice with 4°C Starkey mineral salt wash solution. Donor and recipient cells were mixed at a 1:2 ratio, spotted onto a filter membrane placed on Starkey-Na₂S₂O₃ solid plates, and incubated at 37°C for 3 days for conjugation. Cells were resuspended from the filter, washed with Starkey mineral salt solution, centrifuged, serially diluted, and plated on selective agar containing antibiotics. Plates were incubated at 37°C for 7–10 days. Transformants were validated by colony PCR.

Analysis of main growth parameters of A. caldus

The culture was incubated in a shaking incubator at 40°C and 180 rpm. Biomass was measured every 2 days using the method described by Feng et al. [21]. Briefly, 1.0 ml of culture was centrifuged at 367 × g for 3 min to remove solid particles. The supernatant was serially diluted appropriately. A 0.5 ml aliquot of the diluted sample was transferred to a centrifuge tube, mixed with 0.5 ml of 5% trichloroacetic acid (TCA), and heated in an 80°C water bath for 25 min, followed by ice cooling. The OD at 260 nm was measured using 5% TCA as a blank. Biomass concentration was calculated based on a pre-established standard curve correlating microscopic cell counts with OD₂₆₀. Sulfate (SO₄²⁻) concentration was quantified using the barium chromate method. A 10 ml aliquot of appropriately diluted sample was transferred to a test tube, mixed with 0.2 ml of 2.5 M HCl, and boiled for 5 min. Then, 0.5 ml of 0.1 M barium chromate suspension was added, and the mixture was boiled for an additional 5 min. A (1:1) ammonia solution was slowly added dropwise until the solution turned lemon-yellow. After cooling at room temperature for 3 min, the mixture was filtered through slow-speed qualitative filter paper, and the filtrate was collected in a 10 ml clean colorimetric tube. If turbidity persisted, the filtration step was repeated until the filtrate became clear. The filtrate was diluted to the mark with distilled water, and the absorbance at 420 nm was measured to calculate SO₄²⁻ concentration using a standard calibration curve. The total incubation period was 26 days. To compensate for daily evaporation losses, 2 ml of sterile distilled water was added to the culture daily.

A. caldus intracellular REDOX status and determination of copper content

Cells were harvested by centrifugation (367 × g, 3 min) to remove solid particles, followed by three washes with PBS. The washed cells underwent five freeze–thaw cycles and were further disrupted via sonication (ice bath, 200 W power, 3 s pulse/5 s interval, and total duration 15 min). The lysate was centrifuged at 8000 × g for 10 min to remove cell debris and pellets, and the supernatant was collected and stored at 4°C for subsequent assays. Total glutathione (GSH), oxidized glutathione (GSSG), total antioxidant capacity (T-AOC), and 2,2-Diphenyl-1-picrylhydrazyl (DPPH) free radical scavenging capacity were quantified using commercial kits (Sangon Biotech, China) according to the manufacturer’s protocols. The intracellular copper content was determined with modifications to a previously reported method [30]. The cell lysate supernatant was diluted with ultrapure water to an appropriate concentration, acidified with high-purity HNO₃ to achieve 2% (v/v) acidity, and subjected to microwave digestion. Total copper was measured using a flame atomic absorption spectrophotometer (AA-240; Varian, USA) equipped with a copper-specific hollow cathode lamp. Absorbance was recorded at 324.8 nm with a slit width of 0.5 nm. Calibration curves were generated using certified copper standard solutions.

Bioleaching experiments

Hundred milliliters of Starkey medium (pH = 2.5) and 2 g of sterilized chalcopyrite mineral particles (≤48 μm) were introduced into a 250 ml conical flask. Various amounts of Cu²⁺ were introduced to replicate the stress induced by the buildup of Cu²⁺ during the latter phase of the real bioleaching process. A. caldus M, A. caldus ΔsscR^Ac, and A. caldus-p-sscR^Ac strains were individually injected into the culture medium containing copper ore. An abiotic control group, designated G1, was established. The exact experimental conditions are shown in Table 1. The bioleaching duration was 40 days, conducted in a shaking incubator at 40°C and 180 r·min^-1. Daily, 2.0 ml of sterilized distilled water was supplied to compensate for evaporation loss. Upon completion of the leaching process, the slag underwent nitrification treatment, and the concentration of Cu²⁺ in the slag was determined by atomic absorption method [31].

Table 1.

Conditions of bioleaching experiments

Condition	Cu2+ (mM)	Initial inoculum (10 ml)	Copper ore (g)
Group
G1	0	\	2
G2	0	A. caldus M	2
G3	0	A. caldus ΔsscRAc	2
G4	0	A. caldus -p-sscRAc	2
G5	125	\	2
G6	125	A. caldus M	2
G7	125	A. caldus ΔsscRAc	2
G8	125	A. caldus -p-sscRAc	2
G9	250	\	2
G10	250	A. caldus M	2
G11	250	A. caldus ΔsscRAc	2
G12	250	A. caldus -p-sscRAc	2

ChIP-seq

The recombinant plasmid p-sscR^Ac-Flag, expressing SscR^Ac fused with a 3 × Flag tag, was constructed via molecular cloning and transformed into the donor strain E. coli SM10. The A. caldus recombinant strain (A. caldus-p-sscR^Ac-Flag) was generated using the aforementioned conjugation method. The strain was cultured to early exponential phase, and formaldehyde was added to the culture at a final concentration of 1% for 10 min to induce crosslinking. The reaction was quenched by adding glycine to a final concentration of 125 mM, followed by an additional 5-min incubation. Fixed cells were harvested by centrifugation (8000 × g, 10 min), and ChIP-seq sample preparation was performed as previously described [32]. Purified DNA was sequenced on an Illumina NovaSeq 6000 platform (SeqHealth Tech, China). Sequencing reads were aligned to the A. caldus ATCC 51756 reference genome, and peaks were identified using MACS2 [33]. Data visualization was conducted with IGV (Integrative Genomics Viewer), and DNA motifs within enriched regions were analyzed using MEME [34].

ChIP-qPCR

Acquire the ChIP samples via the previously described approach. Primers were designed with Primer3 Plus (Supplementary Table S2) and were chosen to amplify fragments around the ChIP-seq peak summit regions and with a length between 150 and 250 bp. The gene ACAty_RS16305 (nonenriched in ChIP-seq data) was chosen as a negative control for ChIP-quantitative polymerase chain reaction (ChIP-qPCR) experiments. Fold-enrichment calculations were performed with the 2^−ΔΔCt method and the A. caldus M strain was used as mock experiment.

qRT-PCR

Appropriate cultures were centrifuged (367 × g, 3 min) to remove particulate debris. Wet cell pellets were collected, flash-frozen in liquid nitrogen, and stored for subsequent use. Total RNA was extracted using the TRIzol reagent kit (Vazyme, China), and residual DNA was removed. All quantitative real-time PCR (qRT-PCR) oligonucleotides (Supplementary Table S2) were designed with Primer3 Plus software. qRT-PCR analysis was performed on an ABI 7500 Real-Time PCR System using SGExcel FastSYBR Mixture (Sangon Biotech, China) for amplification and fluorescence detection.

Isolation of nascent transcripts

The wild-type strain and the mutant strain with the deletion mutation were cultivated to the logarithmic growth phase. Simultaneously, 250 mM Cu²⁺ was introduced, and 200 μM 4sU was included into the culture media at the final concentration, allowing for one further hour of cultivation to label the freshly transcribed molecules. Subsequent to tagging, the cells were promptly rinsed with PBS solution, and total RNA was extracted utilizing Magzol. The total RNA was extracted and carefully combined with 5 mM N-[6-(Biotinamido)hexyl]-3’-(2’-pyridyldithio)propionamide (Biotin-HPDP). Subsequent to the reaction, an equivalent amount of phenol/chloroform solution (pH 6.7) was included, vigorously agitated, and equilibrated at ambient temperature for 2–3 min. Subject to centrifugation at 20 000 × g for a duration of 5 min. Subsequently, 1/10 volume of 5 M NaCl and equivalent amount of isopropanol was includedl, ensuring thorough mixing by inversion. RNA was centrifuged at 20 000 × g for 20 min to achieve sedimentation. The supernatant was carefully removed. A suitable volume of pre-chilled 75% ethanol was added. The mixture was centrifuged at 20 000 × g for 10 min. The supernatant was again removed. The precipitate was allowed to equilibrate at ambient temperature for 5–10 min. A suitable volume of diethyl pyrocarbonate (DEPC) water was added to solubilize the RNA precipitate. The aforementioned biotinylated RNA samples were incubated at 65°C for 10 min for denaturation. The denatured RNA was rapidly cooled on ice for 5 min. For 100 g of RNA, 100 μl of streptavidin-biotinylated beads were utilized and incubated at ambient temperature for 15 min. The beads were subsequently washed three times with pre-warmed Wash buffer at 65°C. The beads were then washed an additional three times at ambient temperature to eliminate nonspecifically bound unlabeled 4sU RNA. The washing solution was removed. Hundred microliters of 100 mM DTT solution was added and incubated for 3–5 min. The aforementioned procedures were repeated to elute the surplus RNA. The eluate was collected twice. 100% anhydrous ethanol (pre-cooled to −20°C) was added for overnight precipitation. The sample was centrifuged at 20 000 × g for 15 min. A single wash with 75% ethanol at ambient temperature was performed and centrifuged at 20 000 × g for 5 min. The supernatant was removed. The pellet was allowed to equilibrate at ambient temperature for 5–10 min. A suitable volume of DEPC water was added. The RNA concentration was quantified using a microplate spectrophotometer. The RNA was reverse-transcribed into cDNA using Protoscript II reverse transcriptase. Quantitative PCR was subsequently performed to assess the abundance of newly produced transcripts of the SscR^Actarget gene.

Fluorescence reporter strain construction and fluorescence detection

The fluorescent reporter plasmid was constructed using the backbone vector pJRD-Sm^r. The enhanced green fluorescent protein (EGFP) reporter gene egfp and the AmpR promoter were amplified from the pET28a-EGFP and pUC19 plasmids, respectively. The SscR^Ac coding gene and its target DNA-binding sequence (∼300 bp, identified via ChIP-Seq) were amplified from the A. caldus M genome. Using overlap extension PCR, the AmpR promoter was fused to sscR^Ac, and the target DNA-binding sequence was linked to the egfp reporter gene. The fused fragments were cloned in opposite orientations into linearized pJRD-Sm^r. The recombinant plasmid was conjugated into A. caldus M to generate the fluorescent reporter strain. A CsoR^Ac fluorescent reporter strain was similarly constructed, with potential CsoR^Ac target promoters predicted using the online tool BacPP (http://bacpp.bioinfoucs.com/home).

Fluorescent reporter strains were cultured in modified Starkey medium containing antibiotics at 40°C with shaking (180 rpm). For fluorescence measurement, 1 ml of culture was centrifuged (367 × g, 3 min) to remove particulate debris. Cells were washed three times with PBS and resuspended. Aliquots (200 μl) were transferred to a black 96-well microplate for fluorescence quantification using a Synergy H4 Hybrid Microplate Reader (excitation: 485 nm, emission: 528 nm, gain: 50). Biomass was independently confirmed using the TCA method.

Electrophoretic mobility shift assays

Electrophoretic mobility shift assays (EMSAs) were performed as previously described [5]. A ∼300 bp DNA probe (identified by ChIP-Seq) was amplified from the A. caldus M genome. For binding reactions, 200 ng of the DNA fragment was incubated with purified transcription factors in binding buffer (10 mM HEPES, 150 mM NaCl, 1 mM EDTA, 1 mM DTT, pH 7.5) for 30 min, followed by the addition of 2 μl of EMSA/Gel-Shift Loading Buffer (Beyotime, China). Samples (20 μl total volume) were loaded onto a pre-electrophoresed 6% polyacrylamide gel (PAGE) in 0.5× Tris-borate-EDTA (TBE) buffer. Electrophoresis was conducted at 100 V for 1.5 h. The gel was subsequently scanned using an ImageQuant LAS 4000 mini imaging system (GE Healthcare) to visualize DNA–protein complexes.

Isothermal titration calorimetry

Isothermal titration calorimetry (ITC) was performed as previously described [35]. All protein and DNA samples were dialyzed overnight at 4°C in a buffer containing 20 mM sodium phosphate (pH 7.0), 150 mM NaCl, and 1 mM EDTA. Protein and DNA concentrations were quantified by UV absorbance measurements at 260 nm (DNA) and 280 nm (protein). Prior to experiments, samples were degassed for 10 min. Titrations were conducted at 25°C using a VP-ITC isothermal titration calorimeter (Malvern Panalytical, UK), with 18 successive injections (2 μl each) administered at 180 s intervals under constant stirring at 750 rpm. Data were processed by subtracting buffer control heat signals and applying baseline corrections. Raw data integration, normalization, and binding curve fitting were performed using MicroCal PEAQ-ITC Analysis Software.

SscR^Ac LC-MS/MS

SscR^Ac was chemically treated using a previously reported protocol [36]. Briefly, SscR^Ac (0.5 mg/ml) after reaction with H₂S_n was alkylated with iodoacetamide (IAM) and digested with trypsin. The digested peptides were desalted using ZipTip C18 pipette tips (Millipore, USA) and reconstituted in 20 μl of dissolution buffer (0.1% formic acid). The mixture was vortexed thoroughly, centrifuged (17 000 × g, 4°C, 20 min), and the supernatant was transferred to an injection vial. A 4 μl aliquot was loaded onto a Thermo EASY nLC 1000 system (Thermo Fisher Scientific, USA) equipped with a reversed-phase C18 analytical column (75 μm inner diameter × 15 cm length). Peptides were eluted over a 60-min gradient from 0% to 100% solvent B (0.1% formic acid in 98% acetonitrile) at a flow rate of 350 nl/min, with solvent A consisting of 0.1% formic acid in 2% acetonitrile. Eluted peptides were ionized via electrospray and analyzed using an LTQ-Orbitrap Velos Pro CID mass spectrometer (Thermo Fisher Scientific, USA) operated in data-dependent acquisition mode with Xcalibur 2.2.0 software (Thermo Fisher Scientific, USA).

Bioinformatics analysis

Homologous sequences and putative functions of SscR^Ac were analyzed using the NCBI BLASTP program (https://blast.ncbi.nlm.nih.gov/Blast.cgi). Potential promoter regions were predicted via the online tool BacPP (http://bacpp.bioinfoucs.com/home). Amino acid sequence alignments were performed using Jalview software. A neighbor-joining phylogenetic tree of multiple SscR^Achomologs was constructed with MEGA 11 software using the Poisson model and 1000 bootstrap replicates. The secondary structure composition was predicted using the SOPMA online tool. DNA-binding motifs were analyzed by scanning the A. caldus ATCC 51756 genome with the MEME FIMO tool. Monomeric structural models of SscR^Ac and CsoR^Ac proteins were obtained from the AlphaFold Protein Structure Database (https://alphafold.com/). The CsoR^Ac–DNA binding model was generated using the online program AlphaFold3. All simulations were visualized with PyMOL 2.5.

Statistics and reproducibility

Statistical analyses were conducted using SPSS 22 (IBM Corp.) and Origin 2021 (OriginLab Corporation). Error bars in all figures represent the standard deviation (SD). Statistical significance was defined as an adjusted P < 0.05.

Results

Global analysis of protein persulfidation in A. caldus before and after extreme copper stress

To confirm whether RSS participate in the defense of A. caldus M against copper ion stress, we quantified intracellular RSS levels utilizing a mBBr-based sulfite conversion technique. The results indicated that 125 mM copper stress did not cause significant alterations in RSS levels; however, under 250 mM copper stress, RSS levels increased by 350.60% to 371.63 nmol·l^-1·10⁸ cells^-1 (Fig. 1A). A. caldus ATCC 51756 genome contains two paralogs each for SQR [sqr1 (ACAty_RS05975), sqr2 (ACAty_RS07185)] and SDO [sdo1 (ACAty_RS08530), sdo2 (ACAty_RS10625)]. Transcriptional analysis of the SDO- and SQR-encoding genes at various copper stress levels revealed that at 125 mM copper stress, the transcriptional levels of the sdo1 and sdo2 genes decreased by 20.13% and 23.73%, respectively, compared to the control group, while SQR genes exhibited no significant alterations (Fig. 1B). Under extreme 250 mM stress, these transcriptional levels of all genes were downregulated, with SDO-encoding genes demonstrating stronger suppression, sdo1 and sdo2 decreased by 65.47% and 55.47%, respectively (Fig. 1B). The results shown that A. caldus M increases intracellular RSS levels under extreme copper stress via inhibiting the SQR–SDO pathway. RSS serve dual functions in the antioxidant hierarchy of sulfur metabolic networks via separate but complementary methods. Primarily, RSS facilitates cysteine persulfidation to avert irreversible protein oxidation under stress circumstances [37, 38]. Secondarily, RSS directly neutralizes ROS via rapid and enzyme-independent quenching [39]. The detected RSS buildup in A. caldus M likely serves as a critical defensive mechanism against copper-induced oxidative damage. Considering that the transcriptional program may display varying stress responses throughout time. A series of repeated tests at different time intervals were done to examine the dynamic alterations in RSS levels under copper stress. The experimental results indicate that the RSS concentration levels in all three groups demonstrated a quick increase, followed by stability, and then a modest decline (Fig. 1C). In the 250 mM copper stress group, the intracellular RSS concentration attained 424.89 nmol·l^-1·10⁸ cells^-1 in the latter stage, which was 3.29 times more than that of the control group. This outcome aligns with Fig. 1A. Under 125 mM copper stress, the intracellular RSS level rose, although the change was not substantial. The reason may be that A. caldus M can withstand 125 mM copper stress by gradually engaging its copper resistance system.

Figure 1.

Analysis of protein persulfidation in A. caldus M proteome. (A) Intracellular RSS concentrations in A. caldus M under varying levels of copper stress. Error bars indicate SD of the mean. * 0.05 < P < 0.1, **0.01 < P < 0.05, ***P < 0.01 (Student’s t-test). (B) qRT-PCR analysis of transcriptional changes in the SQR–SDO system of A. caldus M under varying levels of copper stress. Error bars indicate SD of the mean. * 0.05 < P < 0.1, **0.01 < P < 0.05, ***P < 0.01 (Student’s t-test). (C) Dynamic changes in intracellular RSS concentration of A. caldus M under different concentrations of copper stress. The A. caldus M strain was divided into three replicates. After being cultivated to the logarithmic growth phase, 0, 125, and 250 mM copper stress was added, respectively. Cells were collected every 4 h. (D) Workflow for identifying cysteine residues undergoing persulfidation modifications in proteins. Red and black‘S’denote unmodified thiol groups (-SH), while blue‘SS’represents persulfidated thiol groups (-SSH). (E) Plot of σ^R (range 0–1): each point denotes a protein, with its color and size reflecting the aggregated intensity value of all peptides associated with that protein. The bar chart in the lower-left corner shows the average number of persulfidated peptides and proteins before and after copper treatment. Error bars indicate SD of the mean. * 0.05 < P < 0.1, **0.01 < P < 0.05, ***P < 0.01 (Student’s t-test). (F) The mass spectrometry intensity bubble plot for all peptides enriched by proteins with a σ^R value of 1. Each point denotes a peptide, with its color and size representing the adjusted mass spectrometry intensity of the peptide. All peptides, except those in the other or function-unknown categories, are designated by their corresponding protein names or numbers .

To systematically investigate RSS-mediated copper stress responses and regulatory mechanisms in A. caldus M under extreme copper stress (250 mM), we performed comparative persulfidation proteomics before and after stress exposure. Our research focused on transcription factors related to RSS regulate and components associated with copper resistance, utilizing the experimental procedure outlined in Fig. 1D. All data were examined in comparison to the A. caldus ATCC 51756 reference genome, with quality control metrics verifying ≥90.9% peptide detection consistency across biological replicates (≥2 replicates), <5.0% false discovery rate (FDR), and >95.0% overall confidence level. Under 0 mM Cu²⁺ conditions, A. caldus M exhibited 863 ± 54 persulfidated peptides corresponding to 606 ± 42 modified proteins. Following exposure to 250 mM copper, these values climbed to 1144 ± 121 peptides and 692 ± 33 proteins (Supplementary Table S3). The baseline persulfidation coverage in A. caldus M significantly exceeded levels reported in Acinetobacter baumannii [40] and S. aureus [41], indicating that A. caldus M possesses a more robust RSS-responsive protein network. Copper stress resulted in a 14.19% increase of modified proteins, a degree of amplification seldom recorded in other investigations, indicating an inherent connection between copper sensing and RSS metabolism. Bioinformatics analysis indicated that persulfidation targets included various functional modules related to copper tolerance, such as the heavy metal-translocating P-type ATPase CopA (involved in metal efflux), polyphosphate kinase (PPK), and exopolyphosphatase (PPX) from the inorganic polyphosphate (polyP) system (which sequester metal cations), GST, and glutathione peroxidase (Gpx) (which scavenge ROS), key intermediates SoxAX and SoxB from the sulfur oxidation system (SOX), carboxysome shell protein CsoS2, and activation protein CbbQ (related to carbon fixation), along with proteins involved in extracellular polysaccharide secretion and cell adhesion (Fig. 1E). To investigate the putative signaling pathways, we concentrated on the transcription factors susceptible to modified by persulfidation (Supplementary Table S4). Including RpoD (the primary sigma factor of RNA polymerase), RpoS (induced under stress conditions) and metal-sensitive transcription factors such as CsoR^Acand ArsR. These regulatory elements may represent possible targets for RSS signaling. Their changed persulfidation states under extreme copper stress indicate the presence of a copper–RSS signaling path in the defensive mechanism against copper toxicity in A. caldus M.

Comparative analysis employed an established semi-quantitative method [41] to compute σ^R values for quantifying alterations in persulfidation levels. σ^R is defined as the ratio of the total number of cysteines modified by persulfidation in the experimental group to the total number of cysteines modified by persulfidation in both the experimental and control groups. A σ^R value of 1 signifies that the peptide chain has not experienced persulfidation modification prior to copper stress, whereas a σ^R value of 0 shows that such modifications occur solely in the absence of copper stress. It should be noted that σ^R values specifically reflect the stress-induced fraction of overall persulfidation capability, irrespective of absolute protein expression levels (Fig. 1E). We concentrate on the 116 proteins with a σ^R value of 1. The 130 peptides were enriched from these 116 proteins (Supplementary Table S5). They are categorized into 12 classifications based on their functionalities (Fig. 1F). Numerous peptides from proteins including the efflux RND transporter permease subunit (ACAty_RS03870), metalloregulator ArsR/SmtB family transcription factor (ACAty_RS08535), and thiol peroxidase (ACAty_RS05920) were enriched, indicating the potential presence of multiple persulfidation modification sites. To elucidate the disparities in mass spectrometry intensity among the peptides, the intensities were normalized and thereafter compared horizontally. The findings are illustrated in Fig. 1F. In transcription factors with σ^R = 1, the peptides from the two ArsR family transcription factors (ACAty_RS08535, ACAty_RS07410) exhibit high mass spectrometry intensity. The ArsR family of transcription factors was formerly thought to respond solely to metal ions; however, new evidence indicates that several members, including SoxR [42] and HlyU [43], may also detect RSS.

Notably, among these proteins, an uncharacterized transcription factor, SscR^Ac (encoded by ACAty_RS16455), exhibited copper stress-dependent persulfidation. Its genomic location in A. caldus ATCC 51756 is distinct. The upstream region harbors a gene cluster linked with copper resistance, comprising a P-type ATPase (ACAty_RS08475), an RND efflux system (ACAty_RS08440-ACAty_RS08450), and a copper transporter (ACAty_RS08465), all of which are components of the same copper-responsive regulatory module (Supplementary Fig. S1). Our previous studies revealed several transcription factors within the copper resistance gene cluster that contribute to copper efflux to varied extents [5, 44]. The downstream region comprises genes that encode glutathione, GST, and protein disulfide reductase—products that may participate in RSS-mediated antioxidant mechanisms. The persulfidation selectivity of SscR^Ac, along with the functional context of its chromosomal vicinity, strongly indicates that this transcription factor may serve as a pivotal element in the copper–RSS signaling cascade. Moreover, supplementary localization analysis was performed on the SscR^Ac homologous protein-coding genes among the genomes of several bioleaching microorganisms (Supplementary Fig. S1). The results show that the genome positioning of these identical genes is also of considerable significance. For instance, In A. ferridurans ATCC33020, the SscR^Ac homologue gene exists in two copies (HF561_RS09505 and HF561_RS09755), with HF561_RS09505 flanked by genes associated with sulfur metabolism, whereas the downstream gene of HF561_RS09755 pertains to copper ion excretion. In A. thiooxdans ATCC19377, the SscR^Ac homolog gene exists in two copies (GCD22_RS06870 and GCD22_RS08165), both of which are associated with a cluster of genes involved in sulfur metabolism. In A. ferrooxidans ATCC53993, the analogous gene SscR^Ac (LFERR_RS15705) corresponds to ACAty_RS16455 as discussed in this study, and it is situated inside a gene cluster associated with sulfur metabolism and metal detoxification. The results demonstrate that the genomic positioning of the SscR^Achomologous protein-coding genes in the genomes of bioleaching microorganisms is conserved and functionally interconnected, further corroborating the hypothesis that SscR^Ac plays a role in regulating copper excretion and active sulfide metabolism.

Functional validation of SscR^Ac using CRISPR–Cas9 system and conjugative transfer technology

Although the ACAty_RS16455 gene being classified as a MarR family transcription factor, its amino acid sequence exhibits minimal similarity with extensively studied MarR family proteins, such as E. coli MarR [45] and S. Typhimurium SlyA (Supplementary Fig. S2) [46]. The secondary structure prediction of SscR^Ac indicated a predominance of α-helices and random coils (Supplementary Fig. S2), which was corroborated by CD spectroscopy (Supplementary Fig. S3B). The secondary structure of the SscR^Ac protein consists of 64% α-helix, 4% β-sheet, and 32% random coil. SEC analysis showed that the recombinant protein eluted at a volume equivalent to 82.76 kDa (Supplementary Fig. S3A). Given its theoretical monomeric molecular weight of 21.5 kDa, it indicates that SscR^Ac adopts a tetrameric structure in solution, contrasting with the dimeric assembly characteristic of the MarR family. Phylogenetic study utilizing BLAST alignment of its amino acid sequence against the NCBI database revealed that its closest homologs are solely present in biomining microorganisms (Supplementary Fig. S4), suggesting potential functional conservation of this transcription factor in these organisms. These data collectively indicate that SscR^Ac may constitute a novel subclass within the MarR family, characterized by unique structural and functional attributes.

Biomining microorganisms exhibit prolonged development cycles, significant genetic barriers, and an absence of effective gene-editing tools. This study established the inaugural CRISPR–Cas9 gene-editing system for A. caldus, utilizing the broad host range plasmid pJRD-215 as the backbone to further validate the function of SscR^Ac (Fig. 2A). Utilizing conjugation transfer [29], we successfully generated the sscR^Ac knockout strain (A. caldus ΔsscR^Ac) and SscR^Ac-overexpressing strain (A. caldus-p-sscR^Ac) (Supplementary Fig. S5). The growth performance and sulfate production were evaluated under normal and copper-stressed situations among four strains, A. caldus M, empty vector control A. caldus-pJRD215, A. caldus ΔsscR^Ac, and A. caldus-p-sscR^Ac (Fig. 2B–G). In the absence of copper stress, all four strains exhibited similar biomass accumulation and sulfate concentrations (Fig. 2B and E), indicating that the deletion or overexpression of sscR^Ac does not influence basal growth or sulfur oxidation in A. caldus M. Under 125 mM copper stress, however, A. caldus ΔsscR^Ac demonstrated a 14.37% decrease in stationary-phase biomass relative to controls, with its maximum specific growth rate postponed until the 15.88th day, approximately 2 days later than the control (13.97th day) (Fig. 2C). Simultaneously, the sulfate concentration in A. caldus ΔsscR^Ac stabilized at 1.35 g·l^-1, reflecting a 21.05% reduction compared to controls, and its peak specific sulfate production rate was postponed by 2.92 days (Fig. 2F). Under 250 mM copper stress, all strains experienced considerable biomass reduction (Fig. 2D), A. caldus ΔsscR^Ac was significantly compromised, attaining merely 0.15 × 10⁸ cells·ml^-1 in stationary-phase biomass and 0.82 g·l^-1 sulfate, markedly lower than the 1.62 g·l^-1 of A. caldus M, with a maximum specific sulfate production rate of 0.08 × 10^–8 g (d·cell)^-1 (Fig. 2G). In contrast, A. caldus-p-sscR^Ac sustained biomass and growth rates similar to controls under 250 mM copper stress, while generating sulfate concentrations 0.33–0.42 g·l^-1 higher and attaining the earliest maximum specific sulfate production rate (Fig. 2G). The rise in sulfate is not a consequence of cell proliferation but is instead attributable to the targeted stimulation of the sulfur oxidation pathway due to the overexpression of sscR^Ac. To express this conclusion more intuitively, we employed the formula as follows: . This ratio removes the influence of cell amount and directly indicates the sulfur conversion efficiency of each bacterial cell. Results indicate that A. caldus-p-sscR^Ac may more effectively convert elemental sulfur into sulfate under copper stress, hence demonstrating its enhanced sulfur oxidation efficiency in such conditions (Fig. 2H). Under 250 mM Cu²⁺ stress, the elemental sulfur conversion efficiency of all strains, except A. caldus ΔsscR^Ac, considerably enhanced compared to the absence of copper stress, suggesting that A. caldus M can withstand extreme copper stress by activating the sulfur metabolism pathway. However, the deletion of the sscR^Ac gene inhibits the specific activation of this pathway. The phenotypic data indicate that altering SscR^Ac expression levels directly affects A. caldus M ability to endure high copper stress.

Figure 2.

Phenotypic and physiological characterization of A. caldus strains under copper stress. (A) Schematic of the CRISPR–Cas9 system in A. caldus and the construction process of the A. caldus ΔsscR^Ac mutant strain. (B–D) Growth curves and specific growth rates of four strains (A. caldus M, A. caldus ΔsscR^Ac, A. caldus-p-sscR^Ac, and A. caldus-pJRD215) under 0 mM (B), 125 mM (C), and 250 mM (D) copper stress. Error bars indicate SD of the mean. (E–G) Sulfate production curves and specific sulfate production rates of the four strains under 0 mM (E), 125 mM (F), and 250 mM (G) copper stress. Error bars represent biological variation for triplicates (standard deviations). (H) Sulfate concentration per unit cell under different concentrations of copper stress. Error bars indicate SD of the mean; * 0.05 < P < 0.1, **0.01 < P < 0.05, ***P < 0.01 (Student’s t-test). (I) Intracellular RSS levels in A. caldus M, A. caldus ΔsscR^Ac, A. caldus-p-sscR^Ac. Error bars indicate SD of the mean; * 0.05 < P < 0.1, **0.01 < P < 0.05, ***P < 0.01 (Student’s t-test). (J) Intracellular copper ion concentrations in A. caldus M, A. caldus ΔsscR^Ac, andA. caldus-p-sscR^Ac. Error bars indicate SD of the mean; * 0.05 < P < 0.1, **0.01 < P < 0.05, ***P < 0.01 (Student’s t-test). (K) DPPH and ABTS radical scavenging rates of A. caldus M, A. caldus ΔsscR^Ac, and A. caldus-p-sscR^Ac. Error bars indicate SD of the mean. * 0.05 < P < 0.1, **0.01 < P < 0.05, ***P < 0.01 (Student’s t-test).

We expected that SscR^Ac may alleviate high copper stress by regulating copper efflux or RSS metabolism, based on the functional annotation of genes adjacent to sscR^Ac in the A. caldus ATCC51756 genome (Supplementary Fig. S1). To evaluate this, intracellular copper and RSS concentrations were quantified in A. caldus ΔsscR^Ac, A. caldus-p-sscR^Ac, and A. caldus M under different copper stress conditions (Fig. 2I and J). Under nonstressed conditions, intracellular copper concentrations showed no significant differences among strains. At 125 mM copper stress, A. caldus ΔsscR^Ac demonstrated a 22.31% decrease in intracellular copper (90.35 nM·10⁸ cells^-1) relative to A. caldus M, while A. caldus-p-sscR^Ac showed a 42.09% increase (165.25 nM·10⁸ cells^-1), indicating that SscR^Ac negatively regulates copper transporter activity. Under 250 mM copper stress, intracellular copper concentrations equalized among strains, suggesting that excessive copper accumulation surpassed the regulatory capability of efflux mechanisms. For RSS metabolism, under 250 mM copper stress, A. caldus-p-sscR^Ac exhibited a 2.62-fold enhancement in RSS (418.68 nmol·l^-1·10⁸ cells^-1) relative to unstressed conditions, whereas A. caldus ΔsscR^Ac demonstrated the least increase (1.47-fold, 203.72 nmol·l^-1·10⁸ cells^-1), significantly lower than both A. caldus M and overexpressing strain. This suggests that SscR^Ac regulates RSS metabolism in A. caldus M, perhaps utilizing RSS to mitigate oxidative damage induced by copper excess. To validate this, redox states were assessed using the GSH/(GSH + GSSG) ratio (Supplementary Fig. S6A). In the absence of copper or under 125 mM stress, all strains displayed comparable ratios (∼97.95% and ∼96.14%, respectively); however, A. caldus ΔsscR^Ac demonstrated reduced GSH levels (Supplementary Fig. S6B and C), indicating compromised GSH synthesis. Under 250 mM copper stress, the GSH/(GSH + GSSG) ratio decreased to 69.64% (A. caldus M), 56.91% (A. caldus ΔsscR^Ac), and 77.02% (A. caldus-p-sscR^Ac), with A. caldus ΔsscR^Ac exhibiting the most pronounced oxidative stress. These results correspond with the DPPH and ABTS radical scavenging experiments (Fig. 2K). At 125 mM copper, A. caldus ΔsscR^Ac exhibited 5.68% DPPH scavenging (compared to 7.02% in A. caldus M), whereas A. caldus-p-sscR^Ac surpassed controls by 27.38% (DPPH) and 19.10% (ABTS). Under 250 mM stress, the scavenging capacities of A. caldus ΔsscR^Ac iminished significantly, but A. caldus-p-sscR^Ac surpassed the controls by 28.40% (DPPH) and 25.88% (ABTS). These findings collectively indicate that SscR^Achelps A. caldus M in surviving extreme copper stress by regulating intracellular copper and RSS levels.

A. caldus-p-sscR^Ac demonstrated exceptional sulfur conversion efficiency under 250 mM Cu²⁺ stress, indicating its significant potential for industrial use. We reproduced the real bioleaching circumstances and performed tests to explore the bioleaching performance of A. caldus-p-sscR^Ac. A nonbiological control group was established to mitigate extraneous interference. The experimental results indicate that, with the exception of the G11 group, the slag concentration in the other groups was markedly lower than that of the respective nonbiological control groups. In the absence of copper stress, the bioleaching efficiency of the three strains exhibited no significant variation. The ultimate concentration of copper ions in the slag fluctuated between 159.02 and 169.91 mg/l. In the 125 mM Cu²⁺ stress system, the leaching rates of each group were influenced to a degree. The group most impacted was G7, exhibiting a copper concentration in the slag of 204.85 mg/l, which was 10.60% greater than that of G6, while the leaching rate markedly diminished. Under 250 mM Cu²⁺ stress, the copper ion concentration in G11 slag was 224.49 mg/l, exhibiting no significant difference from the nonbiological control. This signifies that the bioleaching impact of A. caldus ΔsscR^Ac at this concentration was significantly impeded. In comparison to G10, the copper ion concentration in G12 slag diminished by 12.70 mg/l, leading to a total of 190.00 mg/l. The SscR^Ac overexpression strain demonstrated an improvement in bioleaching efficiency under 250 mM Cu²⁺ stress (Supplementary Fig. S7).

Identification of SscR^Ac target genes via ChIP-seq analysis

This study employed ChIP combined with next-generation sequencing to comprehensively analyze the regulatory network of SscR^Ac by identifying its DNA-binding sites on a genome-wide scale. Due to the lack of ChIP-grade antibodies suitable for A. caldus M, a Flag-tagged fusion expression vector (p-sscR^Ac-Flag) was incorporated into A. caldus M to construct a ChIP-compatible engineered strain (A. caldus-p-sscR^Ac-Flag). ChIP-seq reads were aligned to the reference genome of A. caldus ATCC51756, and MACS 2.0 detected 685 putative binding peaks (Fig. 3A), including 122 high-confidence peaks (P= 1 × 10^-5) localized to promoter regions (Supplementary Table S6). Functional annotation indicated that of the 122 genes, 4 were directly involved with copper tolerance and 7 were linked to RSS-mediated oxidative stress (Table 2). To verify the reliability of the ChIP-seq data, independent ChIP-qPCR experiments were performed using a separate set of bacterial cultures. ChIP-seq peak information was used to design qPCR probes for the 11 binding sites mentioned above and the promoter of the sscR^Ac gene itself (Supplementary Table S2). SscR^Ac binding was detected at all sites (Fig. 3B). qRT-PCR was employed to confirm the regulatory function of SscR^Ac on these 11 genes. The findings indicated that copper resistance-associated genes ACAty_RS08430 (which encodes the copper-sensitive transcription factor CsoR^Ac) and ACAty_RS08445 (which encodes the efflux RND transporter periplasmic adaptor subunit CusB) were increased by 2.64- and 6.04-fold, respectively (Fig. 3C), in A. caldus ΔsscR^Ac. Within the context of RSS-mediated oxidative stress-related genes, SQR (ACAty_RS05975, sqr1) and SDO (ACAty_RS10625, sdo2) exhibited downregulation of 42.74% and 65.72%, respectively, in A. caldus ΔsscR^Ac, whereas the thioredoxin family gene ACAty_RS08565 (designated dsbG) demonstrated a 5.40-fold upregulation (Fig. 3C). The levels of steady-state RNAs are influenced at the post-transcriptional level by mechanisms including RNA degradation, splicing, and transport [47, 48]. To accurately elucidate the regulatory influence of SscR^Ac on the transcriptional activity of target genes under 250 mM Cu²⁺ stress, we employed the methodologies of Watson et al. [49] and Barrass et al. [50] utilizing the 4sU labeling strategy to investigate the transcriptional regulatory mechanism of SscR^Ac by capturing newly synthesized RNAs. The results showed that under copper stress, csoR^Ac, cusB, and dsbG exhibited a mild increase (∼1.5-fold) upon A. caldus ΔsscR^Ac in the nascent sample (Supplementary Fig. S8). This may be because under copper stress, A. caldus increases the intracellular RSS level and weakens the inhibitory effect of SscR^Ac. In addition, SscR^Ac may also affect post-transcriptional regulation. Considering the functional attributes of these target genes, we suggest that SscR^Ac, presumably a multifunctional global regulator in A. caldus M, sustains cellular copper homeostasis by governing copper ion efflux and influences cellular redox status via RSS metabolism. Furthermore, SscR^Ac likely functions as a negative regulator by inhibiting the promoter activity of csoR^Ac, cusB, and dsbG, in accordance with the conventional mechanism of most MarR-family transcription factors [51]. Nonetheless, it may demonstrate unconventional activation effects on sqr1 and sdo2.

Figure 3.

Potential regulatory sites and mechanisms of SscR^Ac. (A) Genome-wide binding profile of SscR^Ac revealed by ChIP-seq. Peaks are annotated by genomic location and signal intensity, with six highlighted peaks associated with copper tolerance and RSS-mediated oxidative stress. (B) Fold enrichment of 12 candidate genes promoter region validated by ChIP-qPCR. Error bars indicate SD of the mean; * 0.05 < P < 0.1, **0.01 < P < 0.05, ***P < 0.01 (Student’s t-test). (C) Relative expression levels of 11 candidate genes validated by qRT-PCR. Error bars indicate SD of the mean; * 0.05 < P < 0.1, **0.01 < P < 0.05, ***P < 0.01 (Student’s t-test). (D–  I) EMSA showing SscR^Ac binding to ∼300 bp DNA probes. Protein concentrations are labeled above each lane. Sequence motifs shared among peaks (identified via HOMER motif analysis) are depicted below the corresponding EMSA results.

Table 2.

SscR^Ac ChIP-seq peak calling results

Gene ID	Product a	Start	End	P-value	Distance to TSSb
ACAty_RS01620	Extracellular solute-binding protein	329 207	329 565	3.44E-11	307
ACAty_RS03605	Redox-regulated ATPase YchF	720 984	721 292	2.19E-06	553
ACAty_RS05280	NAD(P)H-hydrate dehydratase	1 056 529	1 057 100	4.57E-07	107
ACAty_RS05580	Efflux RND transporter periplasmic adaptor subunit	1 128 490	1 128 796	4.50E-13	84
ACAty_RS05975	Sulfide:quinone oxidoreductase SQR	1 213 799	1 214 249	2.50E-29	424
ACAty_RS08430	Metal-sensitive transcriptional repressor CsoRAc	1 740 577	1 740 878	1.94E-07	468
ACAty_RS08445	Efflux RND transporter periplasmic adaptor subunit CusB	1 745 548	1 745 932	1.66E-08	276
ACAty_RS08565	Thioredoxin fold domain-containing protein DsbG	1 762 781	1 762 980	1.97E-12	170
ACAty_RS10625	Persulfide dioxygenase SDO	2 185 613	2 186 087	1.94E-06	548
ACAty_RS11190	TlpA family protein disulfide reductase	2 296 724	2 297 011	5.02E-10	462
ACAty_RS11925	NADH-quinone oxidoreductase NQO	2 443 061	2 443 340	7.70E-08	83
ACAty_RS16455	MarR family transcriptional regulator	1 755 685	1 755 913	1.49E-86	273

^ahomology annotation of genes in uniprot database; ^bthe distance from peak to TSS is the middle of peak minus the value of TSS.

To validate the dual regulation mode of SscR^Ac, this study constructed an EGFP-based in vivo reporter system in A. caldus M. Utilizing the plasmid pJRD-215 as the template, sscR^Ac and egfp were inserted in opposing orientations. The 5′ end of sscR^Ac was fused to the ampicillin promoter, while the 5′ end of egfp was linked to six tested promoters, generating a series of reporter plasmids, p-sscR^Ac-PsscR^Ac/PcsoR^Ac/PcusB/PdsbG/Psqr1/Psdo2-egfp. A control plasmid lacking sscR^Ac (p-egfp) was introduced into A. caldus M to generate the control strain A. caldus-p-egfp. It was presented that EGFP expression driven by Psqr1 and Psdo2 promoters increased by 99.37% and 88.72%, respectively, compared to the control (Supplementary Fig. S9A). Conversely, the remaining four promoters (PsscR^Ac, PcsoR^Ac, PcusB, and PdsbG) demonstrated differing levels of repression (Supplementary Fig. S9A). Additionally, to examine the interaction between SscR^Acand DNA under copper stress conditions, we performed a fluorescence reporter gene experiment at 250 mM copper stress. The experimental results show that, compared with the without copper stress, the deletion of the sscR^Ac gene leads to a considerable rise in the expression level of EGFP driven by PcsoR^Ac and PcusB, while the expression level of EGFP driven by PdsbG and Psdo2 diminishes. Notably for Psdo2, the fluorescence intensity reduced by 55.97% (Supplementary Fig. S9B). This result signifies that copper ions can influence the activity of these promoters. Enhancing the activities of PcsoR^Ac and PcusB enhances the excretion of copper ions, whereas diminishing the activity of Psdo2 fosters the accumulation of RSS. For the SscR^Ac negative regulatory genes (sscR^Ac, csoR^Ac, cusB, and dsbG), the presence of sscR ^Ac leads to a reduction in the fluorescence disparity between the experimental and control groups with the addition of copper ions, to varied extents. It is indicated that the addition of copper ions mitigated the inhibitory influence of SscR^Ac on them. For the SscR^Ac positive-regulating genes (sqr1, sdo2), the introduction of copper stress did not substantially influence the activation of Psqr1, but it resulted in a diminished activation of Psdo2 (Supplementary Fig. S9B). These findings closely correspond with prior qRT-PCR data, validating that SscR^Ac exerts a bidirectional regulatory mechanism of “activation-repression” on its target genes. EMSAs were used to further analyze the in vitro interaction between SscR^Ac and DNA (Fig. 3D). At protein concentrations of 2.5–3.75 μM, all examined promoter fragments exhibited distinct mobility shift bands, indicating that SscR^Acdirectly interacts with the promoter regions of its target genes (Fig. 3D–I).

Mechanisms of RSS-triggered dissociation of SscR^Ac from target genes

Proteomic modification studies indicated that SscR^Ac undergoes persulfidation in response to copper stress. To clarify the effect of this modification on its protein-DNA binding characteristics, we treated SscR^Acwith varying concentrations of inorganic RSS (H₂S_n). A negative control experiment was performed utilizing the T7 promoter (PT7), which is not regulated by SscR^Ac (Supplementary Fig. S10). Following H₂S_n treatment, the affinity of SscR^Ac for its own promoter fragment PsscR^Ac was markedly diminished. At an H₂S_n concentration of 0.6 μM, the free DNA probe was observed (Fig. 4A). Subsequent examinations of additional target promoters revealed that H₂S_n treatment eliminated binding of SscR^Ac to PcsoR^Ac, PcusB, and PdsbG (Fig. 4B–D). However, SscR^Ac maintained consistent binding to Psqr1 and Psdo2, unaffected by H₂S_n (Fig. 4E and F). ITC was used to measure changes in protein–DNA binding affinity before and after H₂S_n treatment. Unmodified SscR^Acdemonstrated dissociation constants (K_D) ranging from roughly 0.11 to 0.63 μM for all promoter segments (Fig. 4G–L). After H₂S_n treatment, the K_D values for Psqr1 and Psdo2 were 0.348 ± 0.199 μM and 0.570 ± 0.197 μM, respectively (Fig. 4K and L), whereas binding signals for other promoters were completely absent (Supplementary Table S7). Given that Cu²⁺ and H₂O₂ may function as signaling molecules [45], we employed EMSA to ascertain the impact of Cu²⁺ and H₂O₂ on the binding affinity of SscR^Ac to Psqr1 and Psdo2. The results verified that treatment with 20 mM Cu²⁺ or 10 mM H₂O₂ did not interfere with the binding of SscR^Ac to Psqr1 or Psdo2 (Supplementary Fig. S11).

Figure 4.

Effects of H₂S_n on SscR^Ac–DNA binding. (A–F) EMSA demonstrating the attenuated binding capacity of SscR^Ac to promoter regions PsscR^Ac (A), PcsoR^Ac (B), PcusB (C), and PdsbG (D) following H₂S_n treatment. Distinct free DNA bands became evident at 0.6 μM H₂S_n, indicative of reduced protein–DNA complex formation. SscR^Ac binding to Psqr1 (E) and Psdo2 (F) remained unaffected by H₂S_n exposure. H₂S_n concentrations are labeled above each lane. (G–L) ITC analysis of SscR^Ac binding to DNA probes under H₂S_n treatment. The top panels show a corrected heat rate, and the bottom panels have normalized heat. Fits of the single-site binding model function are shown as solid lines. Black squares represent untreated control data points, red circles denote H₂S_n-treated data points. Dissociation constants (K_D) are annotated; “N/A” indicates no detectable binding.

In order to observe the changes in the binding of SscR^Ac to the promoter in vivo as a result of RSS effect. We utilized the fluorescence reporting system to supply NaHS to the cells under both copper stress and nonstress conditions, subsequently measuring the fluorescence signal intensity. The results indicate that in the absence of copper stress, NaHS treatment significantly reduced the affinity of SscR^Ac for the promoters of negative regulatory genes. This impact remained unchanged under copper stress settings (Fig. 5A). In vitro, the introduction of copper ions to SscR^Acwill result in a white precipitate (Fig. 5B). Utilizing CsoR^Acas the control, different concentrations of copper ions were added to the protein solutions of 5 μM SscR^Ac and CsoR^Ac. Upon reaching a final concentration of 40 mM of copper ions, significant precipitation was seen in SscR^Ac. The CsoR^Ac group remained relatively clear. A reddish-brown precipitate emerged following the addition of 0.6 μM H₂S_n to SscR^Ac group (Fig. 5C). The bottle with the lower copper concentration (No. 1) likewise exhibited reddish-brown sediment (Fig. 5C). We suggest that Cu²⁺ may induce nonspecific aggregation of SscR^Ac, whereas H₂S_n can safeguard SscR^Ac from the effects of copper ions by chelating Cu²⁺ and persulfidation modification on cysteine. To validate this hypothesis, we performed nonreducing SDS–PAGE. The results indicated that in the control group, an increase in copper ion concentration led to the emergence several nonspecific bands. The incorporation of H₂S_n inhibited the emergence of nonspecific bands (Fig. 5D). We employed the EMSA experiment to ascertain if Cu²⁺ would influence the dissociation of H₂S_n-induced SscR^Acfrom DNA. Figure 5E shows that, when the concentration of copper ions increases, a free band consistently exists and no blocking band appeared. This signifies that copper ions did not influence SscR^Ac to sense H₂S_n.

Figure 5.

I n vivo and in vitro verification that H₂S_n protects SscR^Ac from Cu²⁺. (A) Difference in signal intensity of the SscR^Ac fluorescence reporter system after cells were treated with Cu²⁺ and NaHS. Error bars indicate SD of the mean; * 0.05 < P < 0.1, **0.01 < P < 0.05, ***P < 0.01 (Student’s t-test). (B) Different concentrations of Cu²⁺ were added to 5 μM SscR^Ac and CsoR^Ac protein solutions. When Cu²⁺ was excessive, SscR^Ac showed a white precipitate. (C) Regardless of the presence of SscR^Ac, Cu²⁺ and H₂S_n produced a reddish-brown precipitate. (D) Nonreducing SDS–PAGE verified that copper ions caused many nonspecific bands to appear in SscR^Ac, while the addition of H₂S_n avoided the generation of nonspecific bands. Lane 1, untreated SscR^Ac control. Lanes 2–5, SscR^Ac treated with 5, 10, 20, and 40 mM Cu²⁺, respectively. Lanes 6–9, SscR^Ac treated with corresponding concentrations of Cu²⁺ and 0.6 μM H₂S_n. (E) EMSA analysis of SscR^Ac–PsscR^Ac binding under Cu²⁺ and H₂S_n treatment. H₂S_n concentrations are 0.6 μM, and Cu²⁺ concentrations are labeled above each lane.

Prior research has indicated that MarR-family transcription factors, which can detect RSS, generally employ cysteine residues to interact with RSS, thus influencing their DNA-binding activity [52]. In this study, we obtained the SscR^Ac monomer model from the AlphaFold Protein Structure Database. The model identified two cysteine residues (Cys74 and Cys78) in SscR^Ac, roughly 6.8 Å apart, a distance akin to the typical lengthy intersubunit Sγ-Sγ lengths (7–9 Å) noted in recognized primary RSS sensors (Supplementary Fig. S12A) [53]. To examine the functional roles of Cys74 and Cys78, mutant strains A. caldus C74S and A. caldus C78S were performed via the CRISPR–Cas9 system (Supplementary Fig. S5C). The intracellular copper ion and RSS concentrations of the mutant strains were analyzed using atomic absorption spectroscopy and the mBBr-based sulfite conversion method. The results showed that under 125 mM copper stress, the intracellular copper ion concentrations in the C74 and C78 mutant strains were 33.71% and 21.41% greater than those in A. caldus M, respectively (Fig. 6A). It is demonstrated that these two cysteine residues are indeed related to the extrusion of Cu²⁺. However, in a 250 mM Cu²⁺ stress situation, the intracellular copper ion concentration exhibited no significant change. This signifies that the copper buildup resulting from such severe copper stress beyond the regulating capability of the A. caldus M efflux system (Fig. 6A). The intracellular RSS levels of the mutant strains diminished by 12.16% and 15.01% correspondingly under 250 mM Cu²⁺ stress compared to A. caldus M (Fig. 6B). The cysteine mutation may have diminished SscR^Ac capacity to alleviate its own transcriptional repression under severe copper stress, thereby impacting the intracellular RSS level. To further verify the roles of Cys74 and Cys78 in the H₂S_n sensing process in vitro, we purified the mutant proteins (SscR^Ac–C74S and SscR^Ac–C78S) and performed EMSA analysis. EMSA experiments indicated that both mutants maintained their affinity for the PsscR^Ac promoter fragment, implying that these mutations had minimal effect on intrinsic DNA-binding ability. However, H₂S_n treatment did not affect the DNA-binding activity of the mutants, so indicating that Cys74 and Cys78 are essential for RSS detection (Supplementary Fig. S12B).

Figure 6.

Molecular mechanism of RSS-sensing by cysteine residues in SscR^Ac. (A) Intracellular copper ion concentrations in A. caldus M, A. caldus C74S, and A. caldus C78S. Error bars indicate SD of the mean; * 0.05 < P < 0.1, **0.01 < P < 0.05, ***P < 0.01 (Student’s t-test). (B) Intracellular RSS levels in A. caldus M, A. caldus C74S, and A. caldus C78S. Error bars indicate SD of the mean; * 0.05 < P < 0.1, **0.01 < P < 0.05, ***P < 0.01 (Student’ t-test). (C and D) LTQ-Orbitrap tandem mass analysis of SscR^Ac under H₂S_n treatment (C) and control (D). A schematic diagram of the fragmentation pattern is illustrated in the upper-right corner of the figure. (E) Proposed mechanistic model of SscR^Ac-mediated copper stress resistance in A. caldus M. Green circles denote Cu⁺. Proteins regulated by SscR^Ac are marked with colored shapes. The SscR^Ac regulatory system is depicted in the lower half of the figure, where thiol groups (-SH) on cysteine residues of S SscR^Ac are labeled, and red “S” represents the inserted sulfur atom after persulfidation modification. The gray light bulb represents transcriptional inhibition, and the yellow light bulb represents either inhibition release or transcriptional activation. The red question mark denotes putative unidentified molecular component(s) mediating the transduction of copper stress signaling into RSS signaling pathways. The black arrows denote the orientation of a chemical reaction or the path of molecular transfer. The red solid lines denote the activation impact. The dotted lines illustrate the inhibitory impact, with black dotted lines denoting persistent inhibition and red dotted lines signifying lifted inhibition.

The aforementioned studies have respectively demonstrated in vivo and in vitro that Cys74 and Cys78 play critical roles in the regulation of copper excretion function and the sensing of H₂S_n in SscR^Ac. To clarify the molecular foundation of RSS sensing, we performed high-resolution LC-MS/MS analysis of H₂S_n-treated SscR^Ac, using DTT-treated SscR^Ac as a control. Due to the proximity of Cys74 and Cys78, tryptic digestion yielded a single peptide containing both residues. In the DTT-treated control, this peptide exhibited a triply charged peak (m/z: 698.94) with a 156.08 Da mass increase, corresponding to N-terminal acetylation (+42.01) and IAM alkylation (+57.02*2) (Fig. 6C). In the H₂S_n-treated group, the triply charged peak were also detected with acetylation and alkylation (m/z: 719.77), accompanied by a ∼62 Da mass shift, indicative of the persulfidation of both cysteines (Cys74-SSH and Cys78-SSH) (Fig. 6D). The results demonstrate that both cysteines are essential for RSS detection, and H₂S_n induces persulfidation by inserting a sulfur atom into the sulfhydryl groups of Cys74 and Cys78, respectively. Although previous research indicates that H₂S_n may detach transcription factors from DNA by modifying oligomeric states [54], SEC revealed that SscR^Ac retains its tetrameric form both prior to and following H₂S_n treatment (Supplementary Fig. S13). This indicates that RSS-induced conformational changes arise from localized sulfur-bond restructuring rather than global oligomeric reorganization. The putative mechanism by which SscR^Ac assists A. caldus M in resisting copper stress is summarized in Fig. 6E. When the intracellular copper ions in A. caldus M begin to rise, metal-sensitive transcription factors such as AcsR and CsoR^Ac will dissociate from the promoter DNA of their target genes upon sensing the metal ions. This initiates the metal efflux process, and the copper ions are subsequently transported from the cell by RND efflux pumps or other transport systems. However, the intracellular RSS level begins to rise when the degree of stress increases and the intracellular copper ion accumulation exceeds the regulatory capacity of the A. caldus M efflux mechanism. On the one hand, RSS can alleviate the oxidative damage caused by ROS. On the other hand, it can act as a signaling molecule, stimulating SscR^Ac, which in turn removes the transcriptional restriction on its own and other copper excretion-related genes, thereby further activating the SQR–SDO system. RSS was further metabolized into sulfites until the copper stress was relieved.

CsoR^Ac transduces copper stress signal into RSS signal

Although SscR^Ac activates the transcription of sqr1 and sdo2 genes, this activation occurs independently of RSS signals, copper ions, or ROS, indicating that the connection between copper stress and RSS signaling need other regulatory elements. In previous studies from our group, the copper-sensitive transcription factor CsoR^Ac possesses a high affinity for copper [44] and demonstrated a substantial transcriptional upregulation (log2 Fold change = 2.412) under 47 mM copper stress [21]. Interestingly, while another branch of the CsoR family, CstR, cannot sense copper but responds to RSS signals [55], our proteomic analysis revealed that CsoR^Ac undergoes persulfidation under copper stress (Fig. 1E), suggesting its possible dual function in sensing both copper and RSS signals. To determine whether CsoR^Ac acts as a channel for transforming copper stress into RSS signals, we first tried to identify its downstream targets. We refined the previously predicted CsoR^Ac-binding consensus sequence to ATRCCCSNNNSGGGYAT (R = A/G; S = C/G; Y = C/T; N = any nucleotide) and scanned the promoter–TSS regions of the copper resistance determinant cluster and sulfur metabolism-related genes in the reference strain A. caldus ATCC 51756 using FIMO [44, 56]. This analysis identified four high-confidence possible targets (Table 3), ACAty_RS16455 (sscR^Ac), ACAty_RS08430 (csoR^Ac), ACAty_RS08530 (sdo1), and the functionally uncharacterized gene ACAty_RS08435.

Table 3.

Putative CsoR^Acoperator targets

Gene ID	Product	Score a	P-value	Sequence
ACAty_RS08530	Persulfide dioxygenase SDO	13.4545	4.23E-06	ATGCGCGAGTTGGGCAT
ACAty_RS08430	Metal-sensitive transcriptional repressor CsoRAc	12.9212	8.68E-06	ATACACATAGGGGGTAT
ACAty_RS16455	MarR family transcriptional regulator	8.73 939	3.21E-05	ATGCCCGCCCGCGGGAA
ACAty_RS08435	Hypothetical protein	7.96 364	7.36E-05	ATTGCCGTCTCGGGTAC

^aThe motif match score of a position in a sequence is computed by summing the appropriate entry from each column of the position-dependent scoring matrix that represents the motif.

Prior studies have revealed that CsoR^Ac binds to its own promoter. We utilized EMSA to evaluate its capacity to interact with three other promoters: PsscR^Ac (The motifs identified by SscR^Ac are in proximity to the motifs identified by CsoR^Ac. The PsscR^Ac sequence is roughly 300 bp and contains these two motifs. Therefore, it is still named PsscR^Ac), Psdo1, and P08435. Increasing CsoR^Ac concentrations progressively shifted the mobility of PsscR^Ac and Psdo1 DNA probes, indicating dose-dependent binding (Fig. 7A and B), whereas P08435 showed no binding (Supplementary Fig. S14A). This confirms that CsoR^Ac specifically recognizes PsscR^Ac and Psdo1 in vitro. In addition, a csoR^Ac overexpression strain A. caldus-p-csoR^Ac was constructed, and qRT-PCR experiments on sscR^Ac and sdo1 were performed. The results showed that compared with the control, the transcription levels of sscR^Ac and sdo1 in the overexpression strain were reduced by 64.01% and 51.26%, respectively (Supplementary Fig. S15). This indicates that overexpression of csoR^Ac significantly reduced the transcription levels of sscR^Ac and sdo1. Given that CsoR^Ac is a copper-sensitive transcriptional repressor, we investigated whether Cu²⁺ influences its DNA-binding activity. Unexpectedly, the pre-incubation of 20 μM CsoR^Ac with escalating doses of Cu²⁺ did not result in the anticipated release of DNA binding (Supplementary Fig. S14B and C). Reducing CsoR^Ac to 2 μM demonstrated that Cu²⁺ increases DNA binding in a dose-dependent manner (Fig. 7C and D), indicating that copper induces conformational changes that stabilize CsoR^Ac–DNA complexes. Given that CsoR^Ac undergoes persulfidation under copper stress (Fig. 1E), we treated CsoR^Ac with H₂S_n and repeated EMSA. H₂S_n inhibited the formation of the CsoR^Ac–DNA complex (Fig. 7E and F), indicating that Cu²⁺ and H₂S_n have opposing actions on CsoR^Ac, copper enhances binding, whereas RSS interferes with it. Then, we supplemented an additional EMSA experiment to determine the effect of Cu²⁺ on CsoR^Ac sensing RSS. The results indicated that with an increase in copper ion concentration, the free bands of PsscR^Ac and Psdo1 persisted (Fig. 7G and H), and there was no phenomenon of the disappearance of the free bands as shown in Fig. 7C and D. To verify these in vitro results, we developed fluorescent reporter strain (A. caldus-p-csoR^Ac-PsscR^Ac-egfp and A. caldus-p-csoR^Ac-Psdo1-egfp) for dynamic observation. Under 250 mM Cu²⁺, fluorescence intensity either stabilized or diminished between 8 and 28 h, indicating that copper-enhanced CsoR^Ac–DNA binding inhibits reporter expression (Supplementary Fig. S16). After 28 h, fluorescence increased and stabilized at levels beyond the unstimulated control, consistent with a dynamic equilibrium in which prolonged copper stress may enhance endogenous RSS synthesis to mitigate CsoR^Ac-mediated suppression. Furthermore, to obtain more direct evidence that RSS triggers the dissociation of CsoR^Ac from PsscR^Ac and Psdo1 in vivo. The CsoR^Ac fluorescent reporter strain was treated with 0.2 mM NaHS with or without 250 mM Cu²⁺ stress, and the fluorescence signal intensity was measured. The results of the study indicate that irrespective of the presence of Cu²⁺, the addition of 0.2 mM NaHS can mitigate this inhibitory impact (Fig. 7I). To explore the structural foundation of Cu²⁺-dependent regulation, we employed AlphaFold3 to predict the interaction between CsoR^Ac with PsscR^Ac and Psdo1 promoters [57]. The CsoR^Ac monomer structure (obtained using AlphaFold DB) showed that Cu²⁺ reduces the spatial distance between CsoR^Ac and DNA, with interaction confidence (ipTM) stabilizing at ∼0.7 (Fig. 7J). This supports copper-induced conformational tightening as a mechanism for enhanced DNA binding.

Figure 7.

CsoR^Ac transduces copper stress signals into RSS signal. (A and B) EMSA validating CsoR^Ac binding to target gene promoters. Protein concentrations are labeled above the lanes. (C and D) EMSA of 2 μM CsoR^Ac after copper treatment. Copper concentrations are labeled above the lanes. (E and F) Effects of H₂S_n on CsoR^Ac–DNA binding. H₂S_n concentrations are labeled above each lane. (G and H) EMSA analysis of CsoR^Ac–DNA binding after Cu²⁺ and H₂S_n treatment. H₂S_n concentrations are 0.8 μM, and Cu²⁺ concentrations are labeled above each lane. (I) Difference in signal intensity of the CsoR^Ac fluorescence reporter system after cells were treated with Cu²⁺ and NaHS. Error bars indicate SD of the mean; * 0.05 < P < 0.1, **0.01 < P < 0.05, ***P < 0.01 (Student’s t-test). (J) Left, AlphaFold3-predicted binding of CsoR^Ac to the sdo1 (top) and sscR^Ac (bottom) promoter fragment in the presence of copper. Right, mechanistic model of how Cu²⁺ and H₂S_n modulate CsoR^Ac activity. The gray light bulb represents transcriptional inhibition or inhibition enhancement, and the yellow light bulb represents either inhibition release.

Discussion

This study reveals the RSS-mediated response strategy to extreme copper stress in A. caldus M and determines the functional mechanism of the major transcription factor SscR^Ac in this context. Our results indicate that A. caldus M undergoes extensive protein persulfidation under extreme copper stress, driven by a copper–RSS signaling pathway. SscR^Ac is crucial for regulating copper homeostasis and redox balance in A. caldus M, although it directly detects RSS instead of copper ions or hydrogen peroxide. Further studies indicate that SscR^Ac regulates copper efflux and RSS metabolism genes via dual mechanisms of repression and activation. RSS-triggered persulfidation of SscR^Ac diminishes its affinity for the promoters of copper efflux and disulfide isomerase-related genes but still preserves its binding to the sqr1 and sdo2 promoters. Additionally, CsoR^Ac demonstrates contrasting regulatory functions in response to copper ions compared to RSS stimulation, highlighting its role in transmuting copper stress signals into RSS signals. In conclusion, we present a molecular model of the SscR^Ac–CsoR^Ac signaling pathway in A. caldus M under extreme copper stress (Fig. 8), wherein CsoR^Ac detects copper ions, intensifies the repression of sdo1, and increases intracellular RSS levels. An elevation in RSS concentration can both induce SscR^Ac persulfidation and protect the essential elements of a broader copper resistance functional module. This modification occurs not just in the enzymes associated with sulfur metabolism but also in numerous other regulators or enzymes that are sensitive to redox reactions. For instance, the Fe-S cluster, functioning as a protein cofactor, demonstrates distinctive characteristics in redox processes owing to the delocalized dispersion of electrons across the Fe and S ions [58]. Consequently, the Fe-S cluster is exceptionally adept for redox sensing, catalysis, or electron transfer [59]. This also implies that they are extremely vulnerable to harm from metal ions or ROS, which could lead to a loss of function. However, the integrity and functionality of the iron–sulfur clusters are preserved by the reversible persulfidation modification. The reversibility of the persulfidation modification makes it a temporary switch under oxidative stress conditions. Once the stress ends, the Trx and GSH systems will restore it to its reduced form [13]. Furthermore, the development of biofilm might be related to redox homeostasis. For example, dysregulation in biofilm production is shown in uropathogenic E. coli that lacks GSH. When exogenous thiol is added, this problem is resolved [60]. This study found that the enzymes involved in c-di-GMP metabolism in A. caldus are also potential targets for RSS modification. This modification may affect the formation of bacterial biofilms by regulating the level of c-di-GMP. Simultaneously, SscR^Ac and CsoR^Ac collaboratively regulate intracellular RSS levels, allowing A. caldus to reduce copper-induced stress and oxidative damage while averting lethal RSS buildup.

Figure 8.

Mechanism of RSS-mediated resistance to extreme copper stress in A. caldus M. The central core illustrates the copper–RSS signaling transduction mechanism, encircled by eight functional copper resistance modules (copper efflux, sulfur metabolism, Fe–S cluster, CO₂ fixation, cell membrane, cell attachment, motility chemotaxis, and ROS scavenge). The copper–RSS signal transduction pathway is divided into seven steps. The gray light bulb represents transcriptional inhibition or inhibition enhancement, and the yellow light bulb represents either inhibition release or transcriptional activation. Proteins marked with colored shapes exhibit increased persulfidation modification under copper stress, while gray shapes represent proteins critical to the module but without elevated persulfidation levels.

This study revealed that SscR^Ac regulates copper efflux but does not directly detect copper ions. Instead, its activity is regulated via RSS-induced persulfidation. This observation highlights an inherent relationship between copper tolerance and RSS signaling in A. caldus M. RSS, acknowledged as the operational mediators of H₂S signaling, have been thoroughly investigated in mammals for their functions in activating tumor suppressors, ion channels, antioxidation, and cytoprotection [61, 62]. However, their chemically diverse nature introduces high complexity in signal transduction [14]. The physiological importance of RSS in prokaryotes, especially in biomining microbes, is inadequately defined. Prior research indicated that hydrogen peroxide stimulates H₂S production in E. coli, resulting in sulfide buildup. Excess sulfide can bind Fe²⁺, thus alleviating ROS-induced damage [63]. This study demonstrates that A. caldus M bypasses Fenton reactions by directly transforming copper stress signals into RSS signaling. This pathway initiating with copper stress sensing, progressing through persulfidation-mediated signal transduction, and ultimately achieving gene regulation, circumvents secondary oxidative damage caused by Fenton reactions. Furthermore, the RSS-responsive thiol proteome of A. caldus M is more extensive than those documented in S. aureus [41] and A. baumannii [40], indicating wider physiological functions for RSS in A. caldus M.

Through prolonged adaptive evolution in natural extreme copper-rich environments, biomining microorganisms may possess a more extensive array of copper resistance elements in their genomes compared to neutrophilic species [64]. Accurately regulating these resistance mechanisms in oligotrophic settings such as acid mine drainage probably entails a more complex procedure [65]. Metal induced transcriptional regulators facilitate hierarchical responses to regulate metal ion uptake, storage, and efflux mechanisms. These regulators are typically responsive to particular metal ions (e.g. Fur [66] and MntR [67]) or metal metabolism byproducts (e.g. the iron-responsive regulator Irr [68]). They are substantially recruited to the genome during stress, yet maintain only modest binding to chromatin. Moderate binding of metal-induced transcription factors to chromatin facilitates more economical adaptation to environmental stress. Some researchers designed a chromium ion stress exchange method (60 mg/l CdCl₂ stress medium to normal medium) to verify the flexibility of transcriptional regulation of metal-induced transcription factors [69]. In Pseudomonas chenduensis strain MBR, metal-resistant genes such as the metal-inducible transcription factor arsR and the RND efflux pump (czcABC) are activated in the presence of cadmium to resist cadmium stress, but their transcription levels quickly return to pre-stress levels after cadmium removal [69]. This is crucial for maintaining intracellular metal homeostasis. The flexibility of metal-induced transcription factors in binding to chromatin may be attributed to the specific binding of metal ions [70]. This regulatory mechanism enables transcription factors to dynamically adjust their binding strength to DNA according to the concentration of available metal ions in the cell, thereby precisely regulating gene expression. This dynamic balance is an important strategy for bacteria to adapt to changes in the metal ion environment. Some MarR family members, including SscR^Ac, are not typical metal-induced transcription factors [71]. They also have a high degree of regulatory flexibility. The flexibility may be attributed to the redox sensitivity of the cysteine residue. Although it has been reported that Cu²⁺ can trigger the dissociation of E. coli MarR from its DNA binding site, the nature of this is that Cu²⁺ induces redox modification of Cys 80 rather than specific binding to Cu²⁺ [45]. RSS also impairs DNA binding of several MarR family transcriptional regulators through a similar mechanism [52]. Examples include Cys12 in S. aureus MgrA [41] and Cys121 in P. aeruginosa OhrR [72]. Although SscR^Ac is not a metal-induced transcription factor, it can also achieve flexible regulation similar to metal-induced transcription factors by virtue of the reversible redox modification of cysteine residues. This dynamic binding mode ensures a rapid response to stress while avoiding overactivation of downstream gene expression, better balancing the limited resources and energy consumption of A. caldus in the process of resisting environmental stress. This is particularly important for the survival of A. caldus in extremely barren conditions and under multiple environmental stresses [9].

The functions of SscR^Ac as a regulator may be diverse. In addition to exerting regulatory activity as a transcription factor, it may also directly/indirectly affect post-transcriptional regulation. Among the potential target genes of SscR^Ac identified by ChIP-seq (Supplementary Table S6), ACAty_RS16385 encodes an Rne/Rng family ribonuclease, which not only plays a core role in ribosomal RNA (rRNA) processing, messenger RNA (mRNA) decay, and possibly transfer RNA (tRNA) processing but also plays an important role in bacterial sRNA-mediated gene expression regulation [73]. In Brucella abortus, RNase E coordinates expression of mRNAs and small RNAs, and is critical for the virulence. Loss of the RNase E encoding gene affected the levels of a range of RNAs, with the most downregulated being genes involved in metal homeostasis, including the metal-induced transcription factor Fur, the outer membrane protein OmpW, and the manganese transporter MntH [74]. This suggests that in B. abortus, RNase E is involved in the post-transcriptional regulation of metal homeostasis proteins, helping to maintain intracellular metal balance. Additionally, among the potential targets of SscR^Ac, there are diguanylate cyclase/phosphodiesterase (ACAty_RS01330, ACAty_RS0144) that are involved in c-di-GMP metabolism. In addition to affecting bacterial phenotypes such as biofilm formation, motility, and virulence, c-di-GMP may also affect post-transcriptional regulation. For example, in Pseudomonas fluorescens, c-di-GMP promotes the post-translational modification of the ribosomal protein RpsF by directly binding to RimK and enhancing its enzymatic activity. This process affects ribosome function and leads to a decrease in the abundance of Hfq [75, 76]. In addition to regulating porin expression at the post-transcriptional level [77], Hfq also affects the ability of the metal-sensitive transcription factor CzcR to bind to the oprD promoter, thereby preventing the toxic effects of excessive metals on cells [78].

Supplementary data

Supplementary data is available at NAR online.

Conflict of interest

None declared.

Funding

This study was supported by grants from the National Key Research and Development Program of China (2022YFC3401304/001), and the National Natural Science Foundation of China (32471540, 32371540, 21878128, and 31701582). This work was funded by the Basic Research Program of Jiangsu and supported by the Jiangsu Basic Research Center for Synthetic Biology (grant no. BK20233003). Funding to pay the Open Access publication charges for this article was provided by the National Key Research and Development Program of China (2022YFC3401304/001).

Data availability

The experimental strain A. caldus CCTCC M 2018727 is deposited at the China Center for Type Culture Collection (CCTCC; NCBI BioSample ID: SAMN10907965). All raw data for the ChIP-seq study presented in this work have been deposited in the Gene Expression Omnibus (GEO) databank with accession codes GSE293048. All data needed to evaluate the conclusions in the paper are present in the paper and/or the Supplementary Materials.