Structural and Biochemical Characterization of Staphylococcus aureus Cysteine Desulfurase Complex SufSU

Abstract

In this work, we provide the first in vitro characterization of two essential proteins from Staphylococcus aureus (S. aureus) involved in iron–sulfur (Fe–S) cluster biogenesis: the cysteine desulfurase SufS and the sulfurtransferase SufU. Together, these proteins form the transient SufSU complex and execute the first stage of Fe–S cluster biogenesis in the SUF-like pathway in Gram-positive bacteria. The proteins involved in the SUF-like pathway, such as SufS and SufU, are essential in Gram-positive bacteria since these bacteria tend to lack redundant Fe–S cluster biogenesis pathways. Most previous work characterizing the SUF-like pathway has focused on Bacillus subtilis (B. subtilis). We focus on the SUF-like pathway in S. aureus because of its potential to serve as a therapeutic target to treat S. aureus infections. Herein, we characterize S. aureus SufS (SaSufS) by X-ray crystallography and UV–vis spectroscopy, and we characterize S. aureus SufU (SaSufU) by a zinc binding fluorescence assay and small-angle X-ray scattering. We show that SaSufS is a type II cysteine desulfurase and that SaSufU is a Zn²⁺-containing sulfurtransferase. Additionally, we evaluated the cysteine desulfurase activity of the SaSufSU complex and compared its activity to that of B. subtilis SufSU. Subsequent cross-species activity analysis reveals a surprising result: SaSufS is significantly less stimulated by SufU than BsSufS. Our results set a basis for further characterization of SaSufSU as well as the development of new therapeutic strategies for treating infections caused by S. aureus by inhibiting the SUF-like pathway.

Introduction

Fe–S clusters are ancient and essential metallocofactors present in all kingdoms of life.¹ Their ubiquity in nature stems from the multitude of functions they perform, and some of which include electron transport, cellular respiration, and regulation of gene expression.² Due to the toxicity of free iron and sulfide in cells, Fe–S cluster formation requires dedicated biogenesis pathways in organisms.³ Across the kingdoms of life, there are five such pathways identified: the nitrogen fixation (NIF), mitochondrial iron–sulfur cluster (ISC), cytosolic iron–sulfur cluster assembly (CIA), sulfur mobilization (SUF), and SUF-like pathways.⁴ While the proteins within each pathway are mostly unique, the pathways share the four stages of Fe–S cluster biosynthesis: (1) sulfur abstraction by cysteine desulfurase, (2) sulfide transfer, (3) Fe–S cluster formation, and (4) Fe–S cluster transfer to apo-proteins.⁵

Most organisms contain redundant Fe–S cluster biogenesis pathways. To illustrate, E. coli expresses the ISC and SUF pathways, and humans express the ISC and CIA pathways.^6,7 However, some organisms express only one pathway. This applies to Gram-positive bacteria from the Bacilli class (such as B. subtilis and S. aureus) and some parasites, which exclusively express the SUF-like pathway.⁸ In these organisms, the SUF-like pathway is essential as it is the only pathway yielding Fe–S clusters required for many essential cellular pathways.⁹ Herein, we provide the first in vitro characterization of two proteins from the SUF-like pathway from S. aureus, which is a potential target for the development of new antibacterial therapeutics.⁸⁻¹⁰ Notably, most previous studies on the SUF-like pathway refer to it as the SUF pathway. Herein, we use the less widely used term, the “SUF-like pathway”, to distinguish it from the SUF pathway expressed in Gram-negative bacteria, such as E. coli, since these two pathways have key differences in their component proteins and mechanisms of action.¹¹

The SUF-like pathway (Figure 1) is encoded by the sufBCDSU operon in S. aureus. The pathway begins with the pyridoxal 5′-phosphate (PLP)-dependent cysteine desulfurase, SufS, which abstracts a sulfur atom from a free cysteine substrate. This sulfur atom is captured by a cysteine residue in SufS, forming a persulfide, and alanine is released as a byproduct.^12,13 SufS then forms a transient complex with the Zn²⁺-dependent sulfurtransferase, SufU, through a Zn²⁺-ligand swapping mechanism.¹⁴ When SufS and SufU are in a complex with each other, the sulfur atom transfers from SufS to SufU as a persulfide on cysteine residues.¹⁵ Next, SufU shuttles the sulfur to the SufBCD protein complex, and Fe–S clusters assemble at the SufBD interface.¹⁶ The Fe–S cluster formation is facilitated by ATP hydrolysis by the SufC ATPases.¹⁷⁻¹⁹ Fully formed Fe–S clusters are then transferred to apo-proteins either directly or via carriers.^11,20

Figure 1.

The SUF-like pathway consists of five core proteins: SufS, SufU, SufB, SufC, and SufD. (1) SufS converts a free cysteine substrate to alanine, and the resulting sulfide is captured by a cysteine residue, generating a persulfide. (2) The sulfide moves from SufS to SufU to SufB. (3) Upon Fe donation, Fe–S clusters assemble at the SufB-SufD interface in the SufBCD complex. (4) Assembled clusters are transferred to various apo-target proteins via carriers. The figure was created with BioRender.com.

In this study, we characterize the SufS cysteine desulfurase and SufU sulfurtransferase from the SUF-like pathway in S. aureus (called SaSufS and SaSufU, respectively). Analysis of the evolution of the SUF-like pathway suggests that both these proteins were recruited to the pathway to help protect sulfur-trafficking intermediates from reactive oxygen species after oxidation of the Earth’s atmosphere.²¹ Cysteine desulfurases are classified as either type I or type II. Type II cysteine desulfurases differ from type I by the lack of a flexible 12-residue loop that aids in positioning the catalytic cysteine near the PLP cofactor in the active site of the protein.²² Instead, type II cysteine desulfurases exhibit a rigid loop containing the catalytic cysteine along with a conserved β-hook that has been suggested to aid in persulfide formation and communication between the monomers of the homodimeric structure.²³ SufS homologues from the SUF-like pathway tend to be type II cysteine desulfurases, as demonstrated with SufS from B. subtilis (BsSufS).⁴

Among the different pathways of Fe–S cluster biogenesis, there are different types of binding partners to the cysteine desulfurase enzyme. For example, U-type scaffold proteins (such as NifU and IscU) form a complex with their respective cysteine desulfurase and serve as the assembly site for Fe–S clusters.²⁴⁻²⁹ In contrast, SufE, which exists in the SUF pathway, serves as a sulfurtransferase from SufS to SufBCD and does not assemble Fe–S clusters.^30,31 SufU from B. subtilis (BsSufU) in the SUF-like system is unique because it is structurally homologous to U-type scaffold proteins but functionally similar to SufE. Some evidence suggests that BsSufU can serve as a Fe–S cluster scaffold (like NifU and IscU);¹² however, more recent studies contradict these results and suggest that the sole function of BsSufU is as a sulfurtransferase, like SufE.^14,15BsSufU fulfills this function using an essential Zn²⁺ ion, which facilitates the formation of the SufS–SufU complex for persulfide transfer using a ligand-swapping mechanism.¹⁴ Although SufU has a similar function as SufE, SufE does not use a Zn²⁺ ion.¹⁴ This is an important distinction between the SUF and SUF-like pathways.

Herein, we characterize the SufSU cysteine desulfurase complex in S. aureus by structural and biochemical techniques. We report that SaSufS is a PLP-dependent type II cysteine desulfurase, and we characterized SaSufU as a Zn²⁺-containing sulfurtransferase that does not serve as a scaffold for Fe–S clusters, which is consistent with BsSufS and BsSufU. Our cysteine desulfurase activity assays confirm that SaSufS is a cysteine desulfurase. Unexpectedly, we observed that SaSufS is significantly less stimulated by SaSufU compared to SufS and SufU from B. subtilis and other species that utilize SufS and SufE of the SUF pathway.^11−13,32 Through cross-species activity analysis (i.e., SaSufS with BsSufU and BsSufS with SaSufU), we show that cysteine desulfurase activity levels are dictated by the SufS homologue and BsSufU and SaSufU can be interchanged with minimal or no impact on SufS activity. Overall, our results show that SaSufS and SaSufU are structurally homologous to BsSufS and BsSufU, respectively, and that SaSufS is a cysteine desulfurase. The structural similarity of these SufS and SufU proteins suggest that they have conserved mechanisms of action; however, the unexpected differences in the cysteine desulfurase activity stimulation by SufU are intriguing. This work lays the foundation for continued study of these proteins, including extensive kinetic analysis and additional structural characterization to further understand how the S. aureus SUF-like pathway compares to other SUF and SUF-like pathways. This work also provides structural and functional information for rationally designing inhibitors that target the active site of SaSufS for the development of new antibacterial therapeutics.

Materials and Methods

Protein Generation

SaSufS and SaSufU Plasmid Construction

SufS and SufU were amplified from S. aureus USA300_LAC chromosomal DNA using the following primer pairs and Phusion DNA polymerase: sufSNheI GGGGCTAGCGCCGAACACTCATTTGACGTTAATGAAGTAATCCTCGAG and sufSXhoI GGGCTCGAGTTAAAATTCATAAGAGAAAAACTCCTTCGTTTGTTTCAAGGC and sufU_NheI GGGGCTAGCAATTTTAATAATCTAGATCAATTATATAGATCTGTCAT and sufU_XhoI GGGCTCGAGCTATTCTTCTTCAGTCGTACCTTCTGCTTTACC.

The PCR product was gel-purified and digested with XhoI and NheI and ligated into similarly digested pET28a before being transformed into Escherichia coli (E. coli) strain DH5α. The resulting pET28a_SaSufS and pET28a_SaSufU plasmids were sequence-verified (Azenta; South Plainfield, NJ). Restriction enzymes and Quick Ligase were purchased from New England Biolabs.

BsSufS and BsSufU Plasmid Construction

pET28a_BsSufS and pET28a_BsSufU plasmids were purchased from GenScript (Piscataway, NJ) using SufS and SufU sequences from the BSubCyc genome database. Plasmids contained a TEV-cleavable N-terminal 6× His tag, and sequences were verified (Azenta; South Plainfield, NJ) before transformation.

Protein Expression

All four protein plasmids (SaSufS, SaSufU, BsSufS, and BsSufU) were independently expressed and purified using similar protocols. First, the plasmids were transformed into BL-21(DE) E. coli in the presence of kanamycin for expression. Cells were grown in LB Broth containing 50 μg/mL kanamycin at 37 °C until an OD₆₀₀ of 0.8–1 was obtained. Expression was initiated with 0.2 mM IPTG and left for 20 h at 20 °C. The media for SaSufS and BsSufS expression were supplemented with 100 μM and 1 mM PLP after induction, respectively, to obtain >95% PLP occupancy. Cells were then harvested by centrifugation and lysed in a binding buffer (50 mM Tris–HCl, pH 7.8, 500 mM KCl, 20 mM imidazole) in the presence of a protease inhibitor. The cell lysate was centrifuged at 17,000 RPM at 4 °C for 40 min. The supernatant was collected, pooled, and filtered through 0.45 μm nitrocellulose before loading onto a HisTrap HP column (Cytiva). After washing the column with five column volumes of the binding buffer, the protein was eluted with a linear gradient of 0–100% elution buffer (50 mM Tris–HCl, pH 7.8, 500 mM KCl, 250 mM imidazole). Fractions containing SufS or SufU were then concentrated, and the buffer was exchanged for the SEC purification buffer. The protein was loaded onto a HiLoad 26/600 Superdex 200 pg column (Cytiva) and eluted with a running buffer (50 mM HEPES, pH 8.0, 300 mM NaCl). Fractions containing SufS or SufU were then concentrated and divided into two pools: one was buffer-exchanged into the assay storage buffer (100 mM MOPS, pH 8.0, 10% glycerol), and one was buffer-exchanged into the crystallography storage buffer (50 mM HEPES, pH 8.0, 300 mM NaCl, 10% glycerol). Both pools were aliquoted, flash-frozen, and stored at −80 °C.

Sequence Alignments

Homologous proteins from differing genera were identified using the NCBI Basic Local Alignment Search Tool (BLAST) by excluding results from the S. aureus species and setting the max target sequences parameter to 1000.³³ Alignments were executed using Clustal Omega, and data were processed for figure generation in Jalview.^34,35

SufS Structure

X-ray Crystallography

The purified SaSufS protein for crystallization was stored at ∼60 mg/mL at −80 °C in a buffer consisting of 300 mM NaCl, 50 mM HEPES (pH 8.0), and 10% glycerol. Crystallization experiments were set up in 24-well plates with a 2:1.5 ratio of the protein solution to a reservoir solution. The reservoir solution consisted of 100 mM Tris (pH 8.35) and 1.8–1.9 M ammonium sulfate. Protein crystals grew by vapor diffusion at room temperature in sitting drops. Large yellow crystals with a hexagonal morphology appeared within 24 h and reached their full size after 3–4 days with little or no protein precipitation. Crystals were typically 100–300 microns long. Crystals were cryo-protected with a solution of one part ethylene glycol and one part glycerol. This solution was added directly to the crystallization drop at a 1:1 ratio of the cryo-protectant solution volume to the reservoir solution volume. After applying the cryo-protection solution, crystals were immediately harvested from the crystallization drop and then flash-frozen in liquid nitrogen. Crystals were stored in liquid nitrogen until data collection. Data were collected at 100 K on a Bruker D8 Venture diffractometer with a MetalJet X-ray generator, Helios MX optics, and Photon III C14 detector at a wavelength of 1.34 Å. Data were integrated, scaled, and merged using the Bruker APEX3 software package (Bruker, 2017). Phasing was determined by molecular replacement against a previously published BsSufS structure (PDB 5J8Q)³⁶ using PHASER.³⁷ The structure was refined with REFMAC5,³⁸ and figures were made in PyMOL.³⁹ A crystal structure of SaSufS at 1.39 Å was deposited into the Protein Data Bank (PDB) with accession code 8D8S.

UV–vis Absorbance Spectroscopy

The PLP occupancy of SaSufS and BsSufS was quantified using an adapted form of the method developed by Blahut et al.¹¹ SufS was diluted to 30–60 μM in 800 μL samples containing 50 mM HEPES at pH 8.0 and 300 mM NaCl. Next, 200 μL of 5 M NaOH was added to the samples, and after which, they were incubated at 75 °C for 10 min. After incubation, 85 μL of 12 M HCl was added to the samples, and then they were centrifuged at 15,000 × g for 5 min. The supernatant was transferred to a plastic cuvette for analysis. Measurements were made on an Agilent 8453 UV–vis absorption spectrophotometer in a range of 300–500 nm. The PLP content was determined using a standard curve of free PLP absorbance at 390 nm under identical experimental conditions, and the occupancy was evaluated by comparing molar ratios of PLP and monomeric SufS.

SufU Structure

Small Angle X-ray Scattering (SAXS)

SaSufU and BsSufU were diluted to ∼25 mg/mL in 50 mM Tris at pH 7.8, 500 mM KCl, and 10% glycerol. SAXS measurements were made using a PANalytical X-ray generator with a Cu target (at 50 mA, 40 kV) and an Anton-Paar SAXSess system, which is an updated model of the Kratky camera.⁴⁰ The signal was acquired with an electronic CMOS detector over a scattering angle range (2θ) from 0.14 to 10°. Data processing was performed with SAXSquant software from Anton-Paar and consisted of desmearing to correct for the line slit geometry of the Kratky method.⁴¹

FluoZin-3 Fluorescence Spectroscopy

The Zn²⁺ occupancy of SaSufU was quantified by comparing the fluorescence to that of a ZnCl₂ standard curve. SufU was diluted to 2 μM in 100 mM MOPS at pH 8.0 and then alkylated in the dark at room temperature for 1 h in the presence of 20 mM iodoacetamide and 10 mM TCEP. Samples were then loaded into a Corning black 96-microwell plate and incubated in the dark at room temperature for 30 min in the presence of 10 μM FluoZin-3. Measurements were made using a BioTek Synergy Neo2 multimode plate reader with an excitation of 494 nm and emission of 516 nm. The Zn²⁺ content was determined using a standard curve of ZnCl₂ under identical experimental conditions, and the occupancy was evaluated by comparing molar ratios of Zn²⁺ and monomeric SufU.

Protein Activity

Cysteine Desulfurase Assays

Cysteine desulfurase activity was evaluated based on a method reported by Selbach et al.¹³ Using this method, alanine production resulting from cysteine desulfurase activity was quantified by measuring the amount of a fluorescent alanine–NDA adduct formed. Cysteine desulfurase reactions were performed at room temperature in 60 μL volumes in Corning black 96-microwell plates containing 100 mM MOPS at pH 8.0, 2 mM TCEP, 0.5 mM cysteine, 1.6 μM SufS, and 8.88 μM SufU. Reactions were initiated by adding 48 μL of 625 μM cysteine to 12 μL of protein and were carried out for 7 min before quenching with 12 μL of 9% TCA for a final concentration of 1.5% TCA. Next, a developing solution of 2.2 mM NDA, 21.4 mM KCN, and 411 mM sodium borate at pH 9.0 was freshly mixed from stock solutions of NDA in MeOH and sodium borate/KCN in nano-pure water. A volume of 28 μL of the developing solution was added to each well, yielding working concentrations of 0.61 mM NDA, 6 mM KCN, and 115 mM sodium borate. Samples were incubated for 20 min in the dark, allowing the fluorescent alanine–NDA to form. The amount of alanine–NDA adducts formed in the presence of excess cysteine and the reducing agent was measured using a BioTek Synergy Neo2 multimode plate reader with an excitation of 390 nm and emission of 440 nm. The amount of the alanine–NDA adduct was quantified using a standard curve that was produced under identical reaction conditions, including 100 mM MOPS at pH 8.0, 2 mM TCEP, 0.5 mM cysteine, and alanine at varying concentrations. Standards and samples within a given 96-well plate were run in triplicate and averaged as one data set. Three separate plates were performed from which reported activity and stimulation values and error propagation were determined. As described in the book “Fe–S Proteins: Methods and Protocols”, this plate reader assay can be difficult to manage due to issues with high background and a subsequent 10-fold decrease in sensitivity to alanine.⁴¹ We were able to achieve satisfactory signal-to-noise ratios by increasing the alanine detection range to 1–50 nmol. In this range, the fluorescence of the alanine–NDA adduct does not follow a linear response to the concentration, so a binomial fit was employed. We included our raw data and calculations in the Supporting Information.

Fe–S Cluster Reconstitution

Scaffold capabilities of SaSufU were tested using an adapted form of the method developed by Selbach et al.¹⁵ Reactions were initiated in a quartz cuvette in a vinyl anaerobic chamber (Coy) and capped before exiting the chamber for analysis. Reaction mixtures consisted of 50 mM HEPES at pH 8.0 and 300 mM NaCl (reaction buffer), 4 μM SufS, 140 μΜ SufU, 420 μΜ Fe²⁺, 420 μM l-cysteine, and 420 μΜ DTT. Two solutions were prepared in the anaerobic chamber prior to reaction initiation: one solution contained 840 μM DTT in the reaction buffer, and the other contained 840 μM Fe²⁺ and 840 μM cysteine from solid FeCl₂·4H₂O and solid l-cysteine, respectively, in the reaction buffer. The protein was added to the cuvette first followed by the DTT solution then followed by the Fe²⁺/cysteine solution. Absorbance spectra and pictures of the reactions were obtained at 45 min intervals for 3 h of reaction time. Measurements were made using an Agilent 8453 UV–vis absorption spectrophotometer in a range of 250–600 nm.

Results and Discussion

SaSufS Characterization

Homology to Type II Cysteine Desulfurases

Initial sequence comparisons of SaSufS with BsSufS indicated high sequence identity (61%). To further investigate the cysteine desulfurase type of SaSufS, we performed a primary sequence comparison of SaSufS with other cysteine desulfurases (Figure S1). Our results show that SaSufS lacks the flexible loop and has interfacial β-hook regions, and both of which are characteristics of type II cysteine desulfurases. This is consistent with other SUF-like systems of Fe–S cluster biogenesis, which also possess type II cysteine desulfurases.⁸ Structure and activity data subsequently presented further supports the classification of SaSufS as a type II cysteine desulfurase. Furthermore, we note important residues conserved among type I and type II cysteine desulfurases that are present in and around the active site, including the catalytic cysteine responsible for the formation of the persulfide intermediate (Cys389 in SaSufS), the lysine that covalently harbors PLP (Lys250), and the histidine that acts as a base in the sulfur acquisition reaction from the cysteine substrate (His147).

PLP Occupancy

The cysteine desulfurase activity of SufS is mediated by its active site cofactor, PLP, which is maintained at a >90% occupancy.^42,43 Since overexpression of proteins can result in partial occupancy of enzyme cofactors, we sought to quantify the PLP occupancy in our purified SaSufS and BsSufS proteins using UV–vis absorbance measurements (Figures S3 and S4 and Table S1). In these experiments, the PLP was released from its covalent attachment to Lys250 by the addition of NaOH and the protein was precipitated. We quantified the amount of PLP by comparing the absorption band of our samples at 390 nm to a standard curve of free PLP, which has a signature absorption band at 390 nm. Our initial occupancy data showed that ∼60% of SaSufS and BsSufS contained PLP post-purification. Upon 100 μM PLP supplementation into the growth media,⁴⁴ a >95% occupancy was observed in SaSufS. BsSufS required 1 mM PLP supplementation to achieve a >95% PLP occupancy. All subsequent protein characterization experiments described herein were performed with protein purified from growth media with supplemental PLP to ensure a >95% PLP occupancy in the resulting purified protein (except X-ray crystallography as described below).

Cysteine Desulfurase Activity Profile

SaSufS cysteine desulfurase activity was confirmed by the detection of alanine formation by NDA derivatization in the presence of excess cysteine and was quantified by an alanine–NDA fluorescence standard curve.⁴¹SaSufS is active on its own in extracting sulfur from cysteine (independent of SaSufU) and subsequently releases approximately 4 nmol alanine over a 7 min reaction time in the presence of the excess reducing agent (Table 2). This is comparable to activity levels of BsSufS (independent of BsSufU), which releases approximately 3 nmol of alanine over a 7 min reaction time in the presence of the excess reducing agent (Table 2).

Table 2. Cysteine Desulfurase Activity Summary for SaSufS and BsSufS.

	no SufU added (basal SufS activity)	SaSufU added		BsSufU added
	alanine produced (nmol)	alanine produced (nmol)	stimulation factor	alanine produced (nmol)	stimulation factor
SaSufS activity	4 ± 2	6 ± 2	1.5	4 ± 1	1
BsSufS activity	2 ± 1	30 ± 10	15	30 ± 10	15

Crystal Structure

To assess the structure of SaSufS, we crystallized the protein and used X-ray crystallography. We achieved a crystal structure resolved to 1.39 Å, which is the first structure of SaSufS as well as the highest resolution structure for a SufS homologue reported to date. A summary of data collection and refinement statistics is provided in Table 1.

Table 1. X-ray Crystallographic Data Collection and Refinement Statistics of SaSufS.

PDB ID	8D8S
data collection statistics
space group	P3221
cell dimensions: a, b, c (Å); α, β, γ (°)	93.85, 93.85, 101.46; 90, 90, 120
resolution (Å)	31.74–1.39 (1.41–1.39)a
Rmerge	0.015 (0.415)a
I/σ(I)	22.8 (2.1)a
completeness (%)	100.0 (97.6)a
no. of unique reflections	104,403 (4995)a
redundancy	22.6 (1.9)a
refinement statistics
Rwork/Rfree	0.168/0.191
average B-factor	21.00
r.m.s. bond lengths (Å)	0.016
r.m.s bond angles (°)	2.05

^aMetrics for the highest resolution shell are given in parentheses.

Visual inspection of our structure shows that the secondary and tertiary structures of SaSufS are nearly identical to those of BsSufS (Figure 2). To quantitatively assess the structure similarity between SaSufS and BsSufS, we performed a structure alignment between our SaSufS structure and a 1.7 Å structure of BsSufS in the resting state (PDB ID 5J8Q)³⁶ using a rigid-body superposition tool to align the C_α of both structures. The resulting alignment has a root mean square deviation (RMSD) of 0.59 Å and a template modeling score of 0.97. (The template modeling score is given on a scale of 0–1 where 1 is perfect alignment.) Both the RMSD and template modeling score confirm high structural similarity between SaSufS and BsSufS. It follows that the SaSufS structure is also highly similar to that of E. coli SufS, IscS, NifS, and CsdA.^{22,36,45−49}

Figure 2.

(A) Overlay of our SaSufS structure (green cartoon; PDB ID 8D8S) with BsSufS (gray cartoon; PDB ID 5J8Q(36)). The SufU binding site (residues 368–371 in SaSufS and 340–343 in BsSufS) and β-hook (residues 279–291 in SaSufS and 253–265 in BsSufS) are colored in blue and pink, respectively. (B) Overlay of the active sites of SaSufS and BsSufS. Residues are numbered relative to SaSufS in bold, and the corresponding BsSufS residues are shown in parentheses. (C) Overlay of the SufU binding sites of SaSufS and BsSufS. The His-Pro-His-Asp residues of the SufU binding motif are shown as blue and gray sticks for SaSufS and BsSufS, respectively. (D) Electron density (2F_o – F_c) map of SaSufS active site residues, contoured to 1.5 σ. (E) Front and side views of the PLP electron density contoured to 1.5 σ. The front view shows the unknown electron density bulge protruding from the phosphorus atom of the PLP phosphate group. The side view shows the apparent trigonal planar arrangement of the terminal P–O bonds in the PLP phosphate group. Images were made in PyMOL.

There are several notable features of SufS homologues that are also present in SaSufS. First, the β-hook latches onto another SufS monomer in the crystallographic structure (residues 279–291 in SaSufS and 253–265 in BsSufS; colored in pink in Figure 2).³⁶ Second, the His-Pro-His-Asp residues of the SufU binding site are strictly conserved, including the side-chain orientation (residues 368–371 in SaSufS and 340–343 in BsSufS; colored in blue in Figure 2).¹⁴ Of these residues, we predict that His370 in SaSufS binds to the Zn²⁺ ion in SufU, as demonstrated in the BsSufS system with corresponding residue His342.¹⁴ Third, the position and orientation of key residues in the active site are also conserved, including the catalytic Cys389 (Cys361 in BsSufS), His147 (His121 in BsSufS), and the PLP-bound Lys250 (Lys224 in BsSufS).

Our SaSufS crystal structure shows an unexpected electron density bulge protruding from the phosphate group of the PLP cofactor (Figure 2E). Typically, the phosphate group of the PLP adopts a tetrahedral geometry. Instead, we observe what appears as a trigonal bipyramidal geometry of the phosphate group. This observed geometry results from a trigonal planar arrangement of the terminal P–O bonds, and the axial positions are occupied by the fourth P–O bond and an unidentifiable bulge in the electron density. Crystallographic refinement suggests that this bulge in the electron density is due to a diatomic species covalently bonded to the central atom. The reason for this perturbation is unclear; however, we suspect that it is a result of unknown species co-crystallized into the structure with partial occupancy. It is unlikely due to radiation damage since there are no other signs of radiation damage in the data. Furthermore, we collected datasets on several different crystals using a wide range of data collection parameters with resolutions ranging from 1.39 to 1.85 Å, and all structures displayed the same perturbation to the PLP. It is also unlikely a reaction intermediate since the phosphate group of the PLP does not participate in the cysteine desulfurase mechanism. We suspect another species is present because our initial protein samples used for the crystal structures contain a <100% PLP occupancy. Based on crystallographic refinement, the PLP occupancy in all our structures is ∼75%, so an unknown pentacoordinate species could be occupying the phosphate binding site in the other ∼25% of protein molecules in the structure. It is unclear what other pentacoordinate species could be present in the structure since no chemicals used in the protein purification or crystallization buffers include a pentacoordinate species. We are exploring this phenomenon further in our laboratory.

SaSufU Characterization

Homology to U-Type Proteins

Similar to SaSufS, we found that SaSufU is homologous to BsSufU with high sequence identity (70%). Unsurprisingly, our primary sequence comparison of SaSufU to other U-type proteins showed the presence of conserved residues that are likely to coordinate Zn²⁺ based on conservation in BsSufU (Figure S2). In SaSufU, these residues are numbered Cys43, Asp45, Cys68, and Cys130. Assuming SaSufU functions like BsSufU and abides by the same Zn²⁺ ligand-swapping mechanism to execute the persulfide transfer, then Cys43 in SaSufU is the residue that releases its bond to the Zn²⁺ ion and accepts the persulfide from SufS.¹⁴

Fold Verification via SAXS

Proper folding of SaSufU post-purification was confirmed by comparing SAXS data from SaSufU and BsSufU (Figure 3). Overall, the particle volume distribution (blue lines in Figure 3) shows that the proteins have a globular structure with no denaturation or unfolding occurring under buffered conditions.⁵⁰ The linear portion at the upper part of the curve indicates slight aggregation, which is expected. The mean particle diameter (histograms in Figure 3) was also consistent between species at 53 Å as well as a less intense signal around 100 Å, which aligns with the size of the SufU monomer and dimer, respectively. These results confirm proper folding of our purified SaSufU protein.

Figure 3.

(A) SAXS diffraction data of SaSufU in solution. (B) SAXS diffraction data of BsSufU in solution. Data from both species show proper folding due to the shape of the particle volume distribution curves (blue lines) and have almost identical particle diameter distributions (histograms) indicating SaSufU is in a highly similar structural condition as BsSufU.

Zn²⁺ Occupancy

Since BsSufU has been shown to be a Zn²⁺-dependent sulfurtransferase, we sought to confirm and quantify the presence of Zn²⁺ in SaSufU. To do this, we alkylated SaSufU, which releases the Zn²⁺ from the protein by effectively capping the cysteines that otherwise coordinate Zn²⁺.⁵¹ The solution was then incubated with the Zn²⁺-fluorophore FluoZin-3, which binds to free Zn²⁺ in solution (K_d = 15 nM). The amount of Zn²⁺ released from the protein was quantified against a Zn²⁺ standard curve. Using this method, we report a 1:1 molar ratio of SaSufU:Zn²⁺, as expected. The data (Table S2) and FluoZin-3-Zn²⁺ standard curve (Figure S5) can be found in the Supporting Information.

Fe-S Cluster Reconstitution

As described earlier, there are contradicting reports if BsSufU can serve as a scaffold for Fe-S clusters. We sought to investigate if SaSufU can form Fe–S clusters in vitro using similar methods.¹⁵ To do so, we incubated SaSufU with SaSufS, cysteine, Fe²⁺, and DTT (reducing agent) for 3 h in an anaerobic environment. Cluster formation was monitored by UV–vis spectroscopy in anaerobic conditions. Absorbance spectra of the reaction mixture indicate no presence of Fe–S cluster formation as shown by the lack of bands around 320 nm (Figure 4A). Furthermore, previous reports show that the reaction solution will darken due to the formation of iron sulfides if SufU does not act as a scaffold for cluster synthesis.¹⁵ In our experiments, we observed a slight darkening of the solution, which is consistent with previous results from BsSufU (Figure 4B,C). From this, we conclude that SaSufU is homologous to BsSufU: both are Zn²⁺-dependent sulfurtransferases that likely do not serve as scaffolds for Fe-S cluster synthesis when Zn²⁺ is present.

Figure 4.

(A) Absorbance spectra of the Fe–S cluster reconstitution solution in 45 min intervals. (B) SaSufSU Fe-S cluster reconstitution solution immediately after initiating the reaction. (C) Fe–S cluster reconstitution solution after 3 h.

Stimulation of SaSufS Activity by SaSufU

BsSufU from the SUF-like system has been shown to stimulate SufS activity between 40 and 100 times and has been characterized as a substrate of SufS.^12,13 Furthermore, SufU from E. faecalis has been shown to stimulate its SufS activity 37-fold.³² Prior to our work, these were the only Gram-positive species investigated for SufS stimulation by SufU, and both demonstrate significant SufS activity stimulation when SufU is present. We sought to investigate the extent by which SaSufU stimulates SaSufS activity, and we directly compare this to our own measurements of BsSufS activity and its stimulation by BsSufU.

Our activity analysis shows that SaSufU mildly stimulates SaSufS with a stimulation factor of 1.5 (Table 2). This is 1 order of magnitude less than the stimulation factor of BsSufU with BsSufS, which we measured to be 15. Our observed decrease in stimulation of BsSufS by SufU compared to literature values is likely caused by the reaction time and substrate concentration chosen for these experiments. Our 7 min reaction time lies outside of the linear range of product formation for both cysteine desulfurase systems, which allows us to compare the efficiency of each system in terms of the total substrate turnover. We chose a substrate concentration of 500 μM to supply the systems with excess substrate while staying well below the concentration at which substrate inhibition begins to occur. For the BsSufSU system, substrate inhibition occurs at ≥1 mM cysteine.¹³ Our observed basal activity for BsSufS is approximately 1 order of magnitude higher than that reported previously with a sulfide detection method and DTT as the reducing agent in the presence of excess cysteine.¹³ However, this difference is attributed to the use of TCEP in our assays since TCEP has been shown to be a more effective reducing agent in regenerating cysteine desulfurases for a subsequent substrate turnover.⁵²

Our experiments show that BsSufS quickly exhausts the substrate supply in the 7 min reaction period tested when in the presence of SufU (the theoretical maximum alanine production is 30 nmol). We expected SaSufS to behave similarly; however, we found that under the same experimental conditions, SaSufS consumes only ∼20% of the substrate supply in the 7 min reaction period when in the presence of SufU. It is unlikely that the lack of stimulation by SaSufU is due to structural issues from SaSufU, such as misfolding or sub-stoichiometric ratios of Zn²⁺, since we confirmed its folded structure and Zn²⁺ occupancy. Also, our characterization of SaSufU shows no evidence of it acting as a scaffold, which would otherwise decrease its stimulatory ability.⁵³ Furthermore, it is unlikely that the lack of stimulation comes from an inherent decrease in cysteine desulfurase activity since SaSufS has a strong homology and comparable basal activity to BsSufS.

To further investigate this lack of stimulation, we conducted cross-species activity analysis by measuring the production of alanine by SaSufS and BsSufS with SufU from the opposing organism in identical reaction conditions and molar ratios of SufS:SufU. Our results show that SaSufS activity is stimulated less than twofold by both SaSufU and BsSufU, and BsSufS is stimulated 15-fold by SaSufU and BsSufU. These intriguing results show that cysteine desulfurase activity levels are dictated by the SufS homologue, and BsSufU and SaSufU are interchangeable. These results also demonstrate that SaSufU acts in an identical manner to BsSufU as a sulfurtransferase as evidenced by comparable stimulation of both BsSufS and SaSufS. To further compare these systems and characterize the unique features of the SaSufSU system, we are conducting thorough kinetic investigations in a separate study.

The results presented herein demonstrate that, although there is high structure similarity between SaSufS and BsSufS, there is a notable decrease in the efficiency of SaSufS to turnover its cysteine substrate when in the presence of SufU. We repeated these experiments on several different days using fresh solutions and protein samples from several different purifications. In all cases, we observed similar stimulation factors and cross-species activity. Given the high sequence identity and structural similarity of SaSufS and BsSufS, it is so far unclear why SaSufS is less stimulated by SufU than BsSufS. One possibility is that the Fe–S cluster scaffold complex SufBCD may be required to stimulate cysteine desulfurase activity. To illustrate, it has been shown in the SUF pathway from E. coli that SufS activity is stimulated eightfold in the presence of its sulfur acceptor SufE, but activity is further stimulated 32-fold by the presence of SufBCD.⁵⁴ Cysteine desulfurase activity stimulation in the presence of SufBCD has not yet been investigated for the SUF-like pathway and may influence activity levels in a similar manner. Another possibility is that the underlying mechanism of the persulfide transfer from SaSufS to SaSufU may be responsible for the lack of stimulation observed in the S. aureus SufSU cysteine desulfurase system. Further investigations are underway in our laboratory.

Conclusions

In this study, we provide the first in vitro characterization of the cysteine desulfurase SaSufS and sulfurtransferase SaSufU proteins from the SUF-like pathway from S. aureus. Our work shows that SaSufS and SaSufU fall under the categories of type II cysteine desulfurase and Zn²⁺-dependent sulfurtransferase, respectively. Both SaSufS and SaSufU show high structural congruence to their B. subtilis homologues, and the basal activity of SaSufS is also comparable to that of BsSufS. We observed an irregular electron density blob at the site of the phosphate group in the PLP cofactor in our SaSufS crystal structure, which we suspect is an unknown pentacoordinate species co-crystallized with the protein. Also, unexpectedly, our work shows that SaSufS activity is significantly less stimulated than BsSufS by their respective SufU proteins with stimulation factors of 1.5 and 15, respectively. Interestingly, the stimulation factors remain unchanged when the SufU species are swapped (i.e., SaSufS with BsSufU and BsSufS with SaSufU), indicating that cysteine desulfurase activity levels are dictated by the SufS homologue and SufU homologues may be interchangeable. Overall, the results of this work set a foundation for further kinetic characterization of SaSufS and SaSufU and investigating these proteins as potential targets for antibacterial development against S. aureus.

Glossary

Abbreviations

SUF

sulfur mobilization

SaSufS

S. aureus SufS

SaSufU

S. aureus SufU

BsSufS

B. subtilis SufS

BsSufU

B. subtilis SufU

SAXS

small-angle X-ray scattering

NIF

nitrogen fixation

ISC

mitochondrial iron–sulfur cluster

CIA

cytosolic iron–sulfur cluster assembly

PLP

pyridoxal 5′-phosphate

IPTG

isopropyl β-d-1-thiogalactopyranoside

SEC

size exclusion chromatography

MOPS

3-morpholinopropane-1-sulfonic acid

HEPES

2-[4-(2-hydroxyethyl)piperazin-1-yl]ethanesulfonic acid

TCEP

tris(2-carboxyethyl)phosphine

DTT

dithiothreitol

TCA

trichloroacetic acid

Supporting Information Available

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acsomega.2c05576.

• Sequence alignments for SaSufS to cysteine desulfurases; sequence alignments for SaSufU to U-type proteins; absorbance spectra, standard curve, and data table for PLP analysis of SaSufS; standard curve and data table for Zn²⁺ quantitation of SaSufU; and raw data and calculated alanine production for cysteine desulfurase assays (PDF)

Accession Codes

Atomic coordinates and structure factors of SaSufS have been deposited in the Protein Data Bank under accession code ID 8D8S.

Author Contributions

The manuscript was written through contributions of all authors. All authors have given approval to the final version of the manuscript. J.D.H., A.E.B., and C.N.M. conceived the project, performed experiments, and analyzed data. E.T.S. participated in experimental development and protein preparation. T.A. and J.M.B. developed SaSufS and SaSufU plasmids and aided in editing the manuscript. J.D.H. and C.N.M. prepared the manuscript.

This work was supported by Colorado School of Mines (for J.D.H., A.E.B., E.T.S., and C.N.M.) and National Science Foundation Award no. 1750624 (for T.A. and J.M.B.)

The authors declare no competing financial interest.