Iron–Sulfur Clusters toward Stresses: Implication for Understanding and Fighting Tuberculosis

Abstract

Tuberculosis (TB) remains the leading cause of death due to a single pathogen, accounting for 1.5 million deaths annually on the global level. Mycobacterium tuberculosis, the causative agent of TB, is persistently exposed to stresses such as reactive oxygen species (ROS), reactive nitrogen species (RNS), acidic conditions, starvation, and hypoxic conditions, all contributing toward inhibiting bacterial proliferation and survival. Iron–sulfur (Fe-S) clusters, which are among the most ancient protein prosthetic groups, are good targets for ROS and RNS, and are susceptible to Fe starvation. Mtb holds Fe-S containing proteins involved in essential biological process for Mtb. Fe-S cluster assembly is achieved via complex protein machineries. Many organisms contain several Fe-S assembly systems, while the SUF system is the only one in some pathogens such as Mtb. The essentiality of the SUF machinery and its functionality under the stress conditions encountered by Mtb underlines how it constitutes an attractive target for the development of novel anti-TB.

1. Introduction

Tuberculosis (TB) is the 13th leading cause of death worldwide. In 2020, it was anticipated that TB will rank as the second leading cause of death from a single infectious agent, after COVID-19 [1]. TB accounts for 1.5 million deaths annually on the global level. This situation is worsened by the alarming steady increase in the number of drug-resistant TB cases, with nearly half a million such cases in 2020 according to the WHO report (Global Tuberculosis report 2021).

Mycobacterium tuberculosis (Mtb), the causative agent of TB, has co-evolved with its human host to establish the continual loop of inhalation, infection, dormancy and transmission to other individuals. Throughout its life in the human host, the stresses to which Mtb is constantly exposed include reactive oxygen species (ROS), reactive nitrogen species (RNS), acidic conditions, starvation, and hypoxic conditions, all of which contribute to inhibiting the proliferation and survival of the bacterium. Indeed, during its pathogenic cycle, in the course of transmission from lungs to aerosols, Mtb experiences hypoxia and nutrient starvation. Once aerosols are inhaled inside the lungs, before encountering macrophages, Mtb is exposed to a hypophase containing short chain fatty-acids, surfactant proteins and glutathione. Upon phagocytosis by naïve macrophages, Mtb experiences increased levels of ROS. Inside activated macrophages, it experiences ROS, RNS and acidic pH. Finally, upon granuloma formation (organized aggregates of macrophages, often with characteristic morphological changes, and other immune cells), it experiences hypoxia, RNS and ROS [2]. Therefore, ROS and RNS are very often encountered by the pathogen Mtb.

Iron–sulfur (Fe-S) clusters, inorganic cofactors of proteins, are excellent targets for ROS and RNS and are susceptible to Fe starvation. Mtb was reported to contain at least 50 Fe-S containing proteins that harbor different structural types of Fe-S clusters participating mainly in energy production and conversion, transcription regulation and amino-acid transport and metabolism [3]. Fe-S cluster biogenesis in prokaryotes is achieved via complex protein systems (ISC, SUF, NIF), which construct a nascent Fe-S cluster on a scaffold protein and then transfer the cluster to recipient proteins [4]. Many microorganisms contain several Fe-S assembly systems, while the SUF system is the only one in some pathogens, such as Mtb [5]. In order to better assess how to target Mtb and fight TB by interacting with the Fe-S metabolism (biogenesis and reactivity) of the pathogen, we report in this review (i) the stresses that target Fe-S cluster proteins within the host macrophages and granulomas; (ii) the putative Fe-S proteins/enzymes in Mtb, (iii) the way Mtb Fe-S cluster proteins are affected by these stresses, and (iv) how some Mtb Fe-S proteins sense and respond to such stresses to adapt to survive, including Fe-S biogenesis/repair. From this analysis, it is reasonable to believe that targeting Fe-S metabolism through the Fe-S biogenesis process could be a good strategy for TB control.

2. Stress Experienced by Mtb

During the course of infection, Mtb passes through a series of intracellular and extracellular locations, from alveolar macrophages to granulomas lesions, where Mtb is exposed to various types of stress. Mtb must therefore cope with a variety of host-mediated stresses, such as cell wall attacks, intra-phagosomal low pH, oxidative and nitrosative stress, hypoxia, iron limitation, nutrient starvation, and DNA damage [6,7].

2.1. Attacks on Mtb Cell Wall

The mycobacterial cell wall is considered as a key element in mediating resistance to host-related stresses as well as in intrinsic antibiotic resistance [8,9]. It is composed of a thick peptidoglycan layer, mycolic acid and arabinogalactan [9]. Hence, Mtb is subject to numerous attacks on its cell surface by host antimicrobial peptides present in the respiratory tract. Indeed, endogenous host defense peptides (HDPs) constitute the first-line defense against pathogens. Mammalian defensins (β-Defensin), granulysin, ubiquitin, and cathelicidins are cell-wall-targeting HDPs that exert an antimycobacterial activity by disturbing membrane permeability [10].

2.2. Harsh pH Conditions in Phagosomes

After phagocytosis by macrophages, Mtb remains in the phagosomes [11]. The host nuclear factor kappa B (NF-_KB) regulates the lysosomal hydrolytic enzymes, and acids release inside the phagosomes in order to destruct the pathogen’s intracellular content (DNA, polysaccharides, lipids, proteins). NF-_KB also increases the production of membrane transport molecules responsible for the phagosome–lysosome fusion that could lead to pathogen elimination [12].

2.3. ROS and RNS

The host’s interferon gamma (IFN-γ) and the tumour necrosis factor alpha (TNF-α) activate the functions of macrophages that generate reactive oxygen species (ROS) and reactive nitrogen species (RNS) in the mM range [11,13,14,15]. Reactive nitrogen species (sometimes referred as reactive nitrogen intermediates) include nitric oxide (•NO), peroxynitrite (ONOO-) and its reaction products, such as nitrogen dioxide (•NO₂) and other types of chemically reactive free radicals (Figure 1). Peroxynitrite is formed from nitric oxide (produced via the enzymatic activity of inducible nitric oxide synthase (iNOS) expressed primarily in macrophages) and superoxide (O₂^•−) (produced by NADPH phagocyte oxidase gp91^phox and gp47^phox) [16,17]. Molecular dioxygen is not the direct toxin of the reactive oxygen species (ROS). Because oxygen can easily take electrons, high oxygen concentrations favor the rapid formation of intracellular superoxide (O₂^•−) and hydrogen peroxide (H₂O₂), which are poisons to cellular molecules. H₂O₂ is capable of giving rise, via “Fenton reaction”-type reactions, to the most deleterious radical species of oxidative stress, the hydroxyl radical OH^• (Figure 1) [18]. ROS and RNS damage DNA and proteins, in particular metalloproteins. Nitric oxide can readily react with transition metals relevant to biological processes including Fe, Cu and Mn; a well-known ‘historical’ example includes the reaction of NO with hemoproteins such as hemoglobin [19], soluble guanylate cyclase [20] and cytochrome c oxidase [21]. NO or NO-derived reactive species also yield adducts with non-heme iron proteins such as Fe-S cluster proteins (see below). ROS damage some iron metalloenzymes such as mononuclear Fe(II) proteins, which are demetallated and subsequently mismetallated by alternative divalent metal ions [22], and also iron–sulfur cluster enzymes (see below). Although ROS and RNS damage numerous targets in a microbial cell, including metal centers, thiols, protein tyrosines, nucleotide bases and lipids, Fe-S clusters are exceptionally susceptible cellular cofactors [23,24].

2.4. Starvation

Human necrotic granulomas are a pathologic hallmark and one of the main features of the host’s immune response to Mtb [25]. The environment inside granulomas is poor in nutrients such as leucine, arginine, tryptophan and phosphate [26,27,28,29]. Another major host immune response to Mtb infection is iron starvation. Iron is an essential element for Mtb as it has a crucial structural and catalytic role in transcription factors and enzymes [30]. Fe is vital for the survival and persistence of Mtb [31]. The host lowers serum iron levels by overproducing iron-binding molecules such as lactoferrin, transferrin, and ferritin, as well as initiating iron withdrawal from the circulation by macrophages [32]. Therefore, there is restricted iron availability within the host.

2.5. Hypoxia

Mtb is an obligate aerobe microorganism that preferentially replicates and resides in the most oxygen-rich regions in the human body [33], in agreement with its presence within patients’ open airways. The granuloma environment is known for its harsh conditions and limited oxygen, making the microenvironment no longer permissive for bacterial persistence and replication. Under hypoxic conditions, the expression of up to 100 genes involved in DNA synthesis, protein synthesis, and cellular division is altered [34].

Despite these impressive stresses, Mtb has the ability to adopt strategies to counteract these detrimental conditions to persist in the host ⁶. We will discuss Mtb responses to ROS, RNS, and Fe limitation, which directly and adversely impact a family of metalloproteins that we focus on in this review, namely the iron–sulfur proteins (see below).

3. Fe-S Clusters Proteins in Mtb

The first protein with an Fe-S cluster was discovered in the early 1960s, due to the presence of an EPR signal at g = 1.9 in the NADH dehydrogenase that had never been observed before [35]. Iron and inorganic sulfur determinations, as well as structure determinations, revealed a completely new chemical composition. Since then, the number of identified proteins containing an Fe-S center has been increasing due to the combination of spectroscopic, biochemical, crystallographic and chemical studies (synthesis of structural analogues of active sites) [36,37,38,39,40,41]. Fe-S clusters are inorganic prosthetic groups composed exclusively of iron and inorganic sulfides contained in proteins [42]. In general, Fe-S clusters are coordinated by thiolate from cysteine residues of the proteins and are involved in the redox reaction, and in particular in electron transport. The most common are Fe₂S₂ and Fe₄S₄ clusters, although more complex architectures are known [43]. The chemical and structural versatility of Fe-S clusters is uniquely achieved by combining the individual chemical properties of Fe and S, which has enabled Fe-S proteins to undertake many important, sometimes essential, functions other than electron transport, such as substrate activation, environmental sensors, gene regulation and structural elements [43,44,45]. A bioinformatic analysis has revealed that Mtb contains at least 50 Fe-S proteins [3] which is half of the known Fe-S proteins in E. coli [46] (Mtb and E. coli have comparable genome sizes). The majority of these proteins coordinate a Fe₄S₄ cluster, but some other types of Fe-S exist, such as FeS₄ (rubredoxin), Fe₂S₂, and Fe₃S₄ clusters. A detailed analysis of these putative Fe-S proteins (their sequence and function), completed by an analysis of the literature on known Fe-S proteins in Mtb, and by a thorough study of characterized bacterial Fe-S proteins whose genes exist in Mtb, made it possible to draw up a list of 58 Fe-S proteins (

Table 1. Predicted Fe-S proteins in Mtb and their essentiality. Genome-wide gene expression tuning using CRISPR interference system. Vulnerability index (VI) and gene essentiality (yes/no) are given (data extracted from DataS2 in [52]). For comparison, genes encoding the targets of first-line TB therapy (rpoB, inhA, and embAB) have a VI value of −9.5, −9.9, −5.85 and −6.45, respectively [52]. Gene for in vitro growth of H37Rv by analysis of saturated Himar1 transposon libraries [47]. Gene for in vitro growth of H37Rv in a MtbYM rich and Mtb minimal media by Tn-seq studies [50]. Gene by Himar1 transposon mutagenesis in H37Rv and CDC1551 strains [48,49,51]. Essentiality is defined as a gene essential for growth and/or survival. Even though NuoI (Rv3153) displays a vulnerability index value (VI: −2.127) in the range of a potential vulnerable gene, the high variability between the lower and higher bounds of the vulnerability index makes this gene likely non-vulnerable.

Gene Name	Functional Category	Name	Predicted Fe-S	Essentiality	VI/ Essentiality [52]
Rv0247c	Intermediary metabolism and respiration	Succinate dehydrogenase	Fe4S4	No [47,48,50]	1.083/No
Rv0252	Intermediary metabolism and respiration	Nitrite reductase NAD(P)H large subunit [FAD flavoprotein] NirB	Fe2S2	No [47,49,50]	1.106/No
Rv0338c	Intermediary metabolism and respiration	Heterodisulfide reductase IspQ	Fe2S2 + Fe4S4	No [50], Yes [47,48,49]	−5.232/Yes
Rv1162	Intermediary metabolism and respiration	Respiratory nitrate reductase NarH	Fe4S4	No [47,48,49,50,51]	1.259/No
Rv1161	Intermediary metabolism and respiration	Respiratory nitrate reductase NarG	Fe4S4	No [47,49,50]	1.345/No
Rv1177	Intermediary metabolism and respiration	Ferredoxin FdxC	Fe4S4 Fe3S4	Yes [47,48,49,50]	-4.792/Yes
Rv1465	Intermediary metabolism and respiration	Nitrogen-fixation-related protein	Fe2S2	Yes [47,48,49,50]	−1.14/No
Rv1475c	Intermediary metabolism and respiration	Aconitase Acn	Fe4S4	Yes [47,48,49,50]	−2.18/Yes
Rv1553	Intermediary metabolism and respiration	Fumarate reductase FrdB	Fe4S4	No [47,48,49,50,51]	−1.393/No
Rv2195	Intermediary metabolism and respiration	Rieske protein QcrA	Fe2S2	Yes [47,48,49,50]	−8.875/Yes
Rv2776c	Intermediary metabolism and respiration	Oxidoreductase	Fe2S2	No [47,48,49,50]	0.734/No
Rv3230c	Intermediary metabolism and respiration	Oxidoreductase	Fe2S2	No [47,49,50] Yes [48]	0.848/No
Rv3151	Intermediary metabolism and respiration	NADH dehydrogenase I NuoG	Fe4S4	No [47,50]	0.727/No
Rv3153	Intermediary metabolism and respiration	NADH dehydrogenase I NuoI	Fe4S4	No [47,49,50]	−2.127/No
Rv3250c	Intermediary metabolism and respiration	Rubredoxin RubB	FeS4	No [47,48,49]	0.681/No
Rv3251c	Intermediary metabolism and respiration	Rubredoxin RubA	FeS4	No [47,48,49,50,51]	0.22/No
Rv3316	Intermediary metabolism and respiration	Succinate dehydrogenase SdhC	Fe4S4	No [47,48,49,50]	0.688/No
Rv3318	Intermediary metabolism and respiration	Succinate dehydrogenase SdhA	Fe4S4	No [47,49,50,51]	1.32/No
Rv3436c	Intermediary metabolism and respiration	Glucosamine-fructose-6-phosphate aminotransferase GlmS	Fe4S4	Yes [47,48,49,50]	−8.478/Yes
Rv3319	Intermediary metabolism and respiration	Succinate dehydrogenase SdhB	Fe4S4	No [47,48,49,50]	1.147/No
Rv3554	Intermediary metabolism and respiration	FdxB	Fe2S2	No [47,48,49,50,51]	1.141/No
Rv2007c	Intermediary metabolism and respiration	FdxA	Fe4S4 and Fe3S4	No [47,48,50]	0.494/No
Rv3674c	Information pathway	Endonuclease III Nth	Fe4S4	No [47,48,49,50,51]	0.789/No
Rv1259	Information pathway	Adenine DNA glycosylase	Fe4S4	No [47,48,49,50]	1.236/No
Rv3571	Intermediary metabolism and respiration	Reductase component of 3-ketosteroid-9-alpha-hydroxylase KshB	Fe2S2	No [47,48,49,50,51]	1.112/No
Rv1485	Intermediary metabolism and respiration	Ferrochelatase (HemZ)	Fe2S2	Yes [47,48,49,50]	−3.427/Yes
Rv1594	Intermediary metabolism and respiration	Quinolinate synthase (NadA)	Fe4S4	Yes [47,48,49,50]	−8.86/Yes
Rv1937	Intermediary metabolism and respiration	Oxygenase	Fe2S2	No [47,49,50,51]	0.808/No
Rv0022c	Regulatory protein	Transcriptional regulatory protein WhiB5	Fe4S4	No [47,48,49,50]	0.687/No
Rv1287	Regulatory protein	Conserved hypothetical protein Rrf family	Fe4S4 or Fe2S2	No [47,48,49,50]	−0.197/No
Rv3197a	Regulatory protein	Transcriptional regulatory protein WhiB-like WhiB7	Fe4S4	No [47,48,50]	−0.17/No
Rv3219	Regulatory protein	Transcriptional regulatory protein WhiB-like WhiB1	Fe4S4	Yes [47,50] No [49]	−10.916/Yes
Rv3260c	Regulatory protein	Transcriptional regulatory protein WhiB2	Fe4S4	Yes [47,50]	−7.373/Yes
Rv3416	Regulatory protein	Transcriptional regulatory protein WhiB3	Fe4S4	No [47,48,49,50]	−0.205/No
Rv3681	Regulatory protein	Transcriptional regulatory protein WhiB4	Fe4S4	No [47,48,49,50]	0.69/No
Rv3862c	Regulatory protein	Transcriptional regulatory protein WhiB6	Fe4S4	No [47,48,49,50,51]	0.309/No
Rv0189c	Intermediary metabolism and respiration	Dihydroxy-acid dehydratase IlvD	Fe2S2	Yes [47,48,49] No [50]	−7.985/Yes
Rv0492c	Intermediary metabolism and respiration	Oxidoreductase GMC-type	Fe4S4	No [47,48,49,50]	0.821/No
Rv0886	Intermediary metabolism and respiration	NADPH adrenodoxin oxidoreductase FprB	Fe4S4	No [47,49,50,51]	0.962/No
Rv2391	Intermediary metabolism and respiration	Sulfite reductase SirA	Fe4S4	No [47] Yes [49,50]	0.514/No
Rv2392	Intermediary metabolism and respiration	Adenosine 5’-phosphosulfate reductase (CysH)	Fe4S4	Yes [47,48,49,50]	−0.34/No
Rv2988c	Intermediary metabolism and respiration	3-isopropylmalate dehydratase (LeuC)	Fe4S4	Yes [48]; No [47,50]	−2.19/No
Rv0808	Intermediary metabolism and respiration	Amido-phosphoribosyl transferase PurF	Fe4S4	Yes [47,48,49] No [50]	−7.635/Yes
Rv1616	Cell wall and cell process	Rubredoxin	FeS4	No [47,49,50]	0.517/No
Rv0863	Conserved hypothetical	Unknown	Fe4S4	No [47,48,49,50]	0.183/No
Rv3242c	Virulence, detoxification, adaptation	Phosphoribosyl transferase	Fe4S4	No [47,48,49,50,51]	0.984/No
Rv3161c	Intermediary metabolism and respiration	Dioxygenase	Fe2S2	No [47,48,49,50]	0.714/No
Rv3858c	Intermediary metabolism and respiration	Glutamate synthase GltD	Fe4S4	Yes [47,48,49]	−9.128/Yes
Rv3859c	Intermediary metabolism and respiration	Glutamate synthase GltB	Fe4S4	Yes [47,48,49]	−11.539/Yes
Rv0322	Intermediary metabolism and respiration	UDP-glucose 6-dehydrogenase UdgA	Fe3S4	No [47,48,49,50]	0.898/No
Rv0423c	Intermediary metabolism and respiration	Phosphomethyl Pyrimidine synthase ThiC	Fe4S4	Yes [47,48,49]	−5.557/Yes
Rv3109	Intermediary metabolism and respiration	Molybdenum cofactor biosynthesis MoaA	Fe4S4	No [47,50] Yes [48,49]	1.107/No
Rv1173	Intermediary metabolism and respiration	F420 biosynthesis FbiC	Fe4S4	No [47,50] Yes [48]	0.962/No
Rv2733c	Translation, ribosomal structure and biogenesis	tRNA 2-methylthio-N(6)-dimethylallyl-adenosine synthase MiaB	Fe4S4	No [47,48,49,50]	−1.184/No
Rv1589	Intermediary metabolism and respiration	Biotin synthase BioB	Fe4S4 and Fe2S2	No [47,49,50,51] Yes [48]	−0.169/No
Rv2218	Intermediary metabolism and respiration	Lipoate Synthase LipA	2x Fe4S4	Yes [47,48,49,50]	−5.037/Yes
Rv1110	Cell wall and cell process	LytB2/IspH	Fe4S4	Yes [47,48,50]	−9.422/Yes
Rv2204c	Conserved hypothetical	Fe-S insertion protein	Fe2S2	No [47,49]	−1.674/No

). Predicted functions were associated with these putative Mtb Fe-S proteins. Among these are intermediary metabolism and respiration, transcriptional regulation (WhiB proteins), cell wall and cell process, information pathway, virulence (Table 1), showing that Mtb exploits Fe-S clusters containing proteins for respiration, metabolism, DNA repair, antibiotic resistance and persistence. Obviously, Fe-S proteins appear to be critical to Mtb’s life. Then, based on data available in the literature (systematic genome-wide studies in RvH37 Mtb strain, including studies using Tn-seq in rich and minimal media) [47,48,49,50,51], we report essential Mtb Fe-S proteins in vitro (essential = indispensable for growth and/or survival) (Table 1, column 5) [47,48,49,50,51]. A set of 20 Fe-S proteins, out of the total 58, are predicted to be essential (Table 1, column 5). Recently, genome-wide gene expression tuning has revealed diverse vulnerabilities of Mtb. A CRISPR interference system, using Mtb as a model organism, has been used to titrate gene expression and uncover gene vulnerability, redefining the concept of essential genes and identifying antimicrobial targets [52]. The vulnerability of a gene was defined by its vulnerability index (VI). In simple terms, the analysis of the gene expression tuning enables a VI value to be obtained; if negative, this means that the gene is essential (Essentiality: Yes). On the contrary, a positive value of the VI (or a low negative value of the VI) indicates that the gene is not essential (Essentiality: No) (Table 1, column 6). Combining all these data (columns 5 and 6), among the 58 predicted Mtb Fe-S proteins, 16 are predicted as essential, both highlighted using the CRISPR interference system and the Himar1 transposon mutagenesis. Rv1465, Rv3109, Rv2391 and Rv2392, which have been predicted to be essential from in vitro growth using Tn-seq, are predicted not to be essential from the CRISPR system with a vulnerability index (VI) of −1.142, 1.107, 0.514 and −0.34, respectively [52]. Therefore, they are not considered as essential.

Among the essential Fe-S proteins, both highlighted by in vitro growth using Tn-seq studies and CRISPR technique, are aconitase Acn (Rv1475c), ferredoxin FdxC (Rv1177), quinolinate synthase NadA (Rv1594), glucosamine-fructose 6-phosphate aminotransferase GlmS (Rv3436c), WhiB1 (Rv3219) and WhiB2 (Rv3260c), IlvD (Rv0189), QcrA (Rv2195), ferrochelatase HemZ (Rv1485), amido-phosphoribosyl transferase PurF (Rv0808), 4-hydroxy-3-methylbut-2-enyl diphosphate reductase 2 (Rv1110), Lipoate synthase (Rv2218), Phosphomethylpyrimidine synthase (Rv0423c), heterodisulfide reductase (Rv0338c), and glutamate synthases (Rv3858c and Rv3859c) (Table 1). Although these proteins/enzymes represent interesting therapeutic targets, few have been characterized at a molecular level as Fe-S proteins.

Mtb Aconitase (AcnA) functions as a tricarboxylic acid (TCA) cycle enzyme converting isocitrate to cis-aconitate, likely using its Fe-S cluster as a catalyst. Upon iron depletion (apo form), it behaves like an iron-responsive protein (IRP), binding to the selected iron-responsive elements (IREs) in vitro. Indeed, the apo form of AcnA (no Fe-S cluster) functions as an RNA-binding regulatory protein that binds to selected IRE-like sequences, present within the UTRs (untranslated regions) of 3’ thioredoxin (trxC) and 5’ iron-dependent repressor and activator (IdeR) mRNA [53]. These two activities of Mtb Acn are mutually exclusive, pointing to its role in iron homeostasis.

Mtb IlvD is a dihydroxy-acid dehydratase, a key enzyme involved in branched-chain amino acid (BCAA) biosynthesis, which catalyzes the synthesis of 2-ketoacids from dihydroxyacids [54]. Recently, the IlvD crystal structure was solved at 1.88 Å resolution, revealing an Fe₂S₂ cluster coordinated by three cysteine residues and one exchangeable water molecule or hydroxide [55]. Spectroscopic studies suggested that the substrate binds to the cluster, strongly suggesting that it acts as a Lewis acid similarly to E. coli IlvD [56]. IlvD (Rv0189c) has no mammalian counterpart and therefore constitutes a very interesting therapeutic target against TB. Mtb-IlvD is inhibited by a herbicide, aspterric acid, that might be a potential lead compound for the design of novel anti-TB drugs [55].

Ferredoxin (FdxC) is a protein that transfers electrons, being possibly involved in a wide variety of metabolic reactions. From its amino-acid sequence, it could bind both Fe₃S₄ and Fe₄S₄ clusters, even though this has to be demonstrated in Mtb.

D-fructose 6-phosphate aminotransferase (GlmS) is involved in the carbohydrate derivative biosynthetic process, catalyzing the first step in hexosamine metabolism and converting fructose-6P into glucosamine-6P using glutamine as a nitrogen source [57]. Mtb GlmS has not yet been characterized in vitro like GlmS enzymes from other microorganisms.

Ferrochelatase HemZ is involved in coproporphyrin-dependent heme b biosynthesis [58]. It catalyzes the insertion of ferrous iron into protoporphyrin IX to form protoheme, which serves as a prosthetic group in a wide array of metabolic pathways; well-known enzymes include respiratory cytochromes, hemoglobin, cytochrome P450s, catalases, and other hemoproteins. Mtb ferrochelatase does not use protoporphyrin as an intermediate but, instead, inserts ferrous iron into coproporphyrin, resulting in the formation of coproheme [58]. This enzyme incorporates a Fe₂S₂ cluster ligated by four cysteine residues, and the cluster is essential for the stability and function of the protein, even though its role is not yet known [59].

QcrA is an iron–sulfur subunit of the cytochrome bc1 complex, an essential component of the respiratory electron transport chain required for ATP synthesis [60]. The bc1 complex catalyzes the oxidation of menaquinol and the reduction of cytochrome c in the respiratory chain. The bc1 complex operates through a Q-cycle mechanism that couples electron transfer to generation of the proton gradient that drives ATP synthesis. Mtb QcrA was identified as a high-confidence drug target [61] and proposed to contain an Fe₂S₂ cluster, even though this has yet to be demonstrated.

Amidophosphoribosyl transferase, PurF, is involved in purine metabolism, catalyzing the formation of phosphoribosylamine from phosphoribosylpyrophosphate (PRPP) and glutamine [62]. Even though it was not characterized at a molecular level, Mtb PurF is proposed to bind a Fe₄S₄ cluster like other bacterial amido-phosphoribosyl transferases [63].

Mtb synthesises the isoprenoid precursor, isopentenyl diphosphate (IPP), via the non-mevalonate or 1-deoxy-D-xylulose 5-phosphate (MEP/DOXP) pathway, in contrast to the human mevalonate pathway [64,65]. LytB (IspH) (Rv1110) is a 4-hydroxy-3-methylbut-2-enyl diphosphate reductase involved in the terminal step of the MEP/DOXP pathway catalyzing the conversion of (E)-4-hydroxy-3-methyl-but-2-enyl pyrophosphate (HMB-PP) into IPP and dimethylallyl pyrophosphate (DMAPP). The LytB structure has been determined in several species and the enzyme contains an iron–sulfur Fe₄S₄ cluster [66,67,68]. Mtb has two homologs of lytB (Rv3382c and Rv1110). Only LytB2 is essential in Mtb, confirming that LytB1 is not able functionally to complement the loss of lytB2 [69]. Mtb LytB2 contains cysteine residues that bind Fe-S clusters in E. coli LytB [70], strongly suggesting that LytB2 is an Fe-S enzyme.

Lipoic acid is a simple cofactor consisting of an eight-carbon fatty acyl chain containing sulfur atoms at C6 and C8. It is synthesized de novo in a two-step process. The first step, catalyzed by octanoyl transferase (LipB), is the transfer of an n-octanoyl chain from octanoyl-acyl carrier protein (ACP) to a lypoyl carrier protein [71,72]. The second enzyme, LipA, catalyzes the stepwise insertion of sulfur at C6 and C8 of the octanoyl chain, yielding the final cofactor. Mechanistic studies have focused on LipA from E. coli. LipA is a member of the radical SAM superfamily of enzymes, which use a [4Fe−4S]⁺ cluster to catalyze the reductive cleavage of S-adenosyl-L-methionine (SAM) to methionine, and a reactive 5′-deoxyadenosyl 5′-radical (5′dA•). The role of the 5′-dA• is to initiate radical catalysis by abstracting target hydrogen atoms (H•), from unactivated or poorly activated carbon centers. LipA harbors two Fe₄S₄ clusters essential for activity [73]. One cluster is coordinated by the three cysteines of the canonical CX3CX2C motif and the SAM coordinates the fourth iron site; this is the cluster where the radical chemistry occurs (RS cluster). The second cluster is ligated by cysteines of the CX4CX5C motif and a serine residue, and it acts as a sulfur donor for lipoic acid formation [74,75]. Lipoyl synthase from Mtb (Rv2218) contains two Fe₄S₄ clusters and converts an octanoyl peptide substrate to the corresponding lipoyl peptide product via the same C6-monothiolated intermediate as that observed in the E. coli LipA reaction [76]. Moreover, LipA from Mtb can complement a lipA mutant of E. coli, demonstrating the commonalities of the two enzymes [76].

The thyamine pyrimidine synthase (Rv0423c) catalyzes the complex rearrangement of aminoimidazole ribonucleotide (AIR) to 4-amino-5-hydroxymethyl-2-methylpyrimidine phosphate (HMP-P) in the thiamin biosynthetic pathways of bacteria and plants [77] and is an example of a non-canonical radical SAM enzyme with the CX2CX4C motif coordinating the Fe-S cluster [78]. ThiC contains only one Fe₄S₄ cluster that is essential for its activity [79]. Structures of ThiC with Fe₄S₄ clusters were solved [80], leaving no doubt that ThiC is an Fe-S enzyme. Structural and functional characterization of Mtb ThiC remains to be carried out.

IspQ (iron–sulfur protein Q) (Rv0338c) was proposed (from Tuberculist) to encode a putative iron–sulfur heterodisulfide reducatase. This is a membrane-bound redox enzyme recently reported to be involved in energy metabolism or redox sensing [81]. This enzyme is the primary target of two 6,11-Dioxobenzo[f]pyrido[1,2-a]indoles (DBPI), molecules that kill Mtb [81]. Even though this protein is suggested to bind two Fe-S clusters (based on heterodisulfide reductase from Methanococcus thermolithotrophicus) [81], this has to be demonstrated through biochemical and structural approaches.

Glutamate synthases carry out the synthesis of two molecules of L-glutamate from L-glutamine and 2-oxoglutarate (2-OG) by catalyzing the reductive transfer of the glutamine amide group to the C(2) carbon of 2-OG. On the basis of the primary structures of glutamate synthases from different sources and the known biochemical properties, three distinct classes of enzymes can be distinguished [82]: (class I) Ferredoxin-dependent glutamate synthase is an iron–sulfur (a Fe₃S₄) and FMN-containing enzyme that has been detected in chloroplasts of higher plants, cyanobacteria and algae; (class II) NADPH-dependent glutamate synthase is mostly found in bacteria; the enzyme is composed of two tightly bound dissimilar subunits, which form the αβ holoenzyme containing one flavin adenine dinucleotide (FAD) and one FMN cofactor, and three distinct iron–sulfur clusters, one Fe₃S₄ center and two Fe₄S₄ clusters; and (class III) NADH-dependent glutamate synthase is poorly characterized and is mainly found in fungi, lower animals, seeds and roots of plants [83]. In all glutamate synthases studied so far, Fe-S clusters serve to transfer electrons. Mtb glutamate synthase, both small and large subunits (GltB and GltD, respectively) were expressed in E. coli and purified, unfortunately without its Fe-S clusters [84].

Quinolinate synthase (NadA) catalyzes a unique condensation reaction between iminoaspartate and dihydroxyacetone phosphate, affording quinolinic acid, a central intermediate in the biosynthesis of nicotinamide adenine dinucleotide (NAD). NadA from Mtb contains a Fe₄S₄ cluster coordinated by three conserved cysteine residues, the fourth catalytic iron site being coordinated by a water molecule [85,86,87]. Implicated in the biosynthesis of an essential cofactor, NadA has been suggested to be a noteworthy target for the design of antibacterial agents against Mtb, like the other genes involved in NAD biosynthesis and recycling. The catalytic mechanism has been well studied using biochemical and structural approaches [88,89,90] allowing the design of NadA inhibitors [91,92]. Unfortunately, these inhibitors have not been proven to be active in cellulo on Mtb NadA.

The WhiB transcriptional regulators (WhiB1-WhiB7). WhiB proteins are members of the WhiB-like (Wbl) family, which contain proteins characterized by a small size (80–140 residues) and the presence of four invariant conserved cysteine residues (Cys(x)nCys(x2)Cys(x5)Cys) involved in the binding of a Fe₄S₄ cluster essential for their function. They share another conserved sequence of five residues (G [V/I]WGG), located at one end of a loop that follows the last conserved cysteine. Structural predictions suggest this loop to be a key element for interacting with cellular molecules [93]. The Mtb genome encodes seven paralogues of Wbl family proteins, WhiB1 (Rv3219), WhiB2 (Rv3260c), WhiB3 (Rv3416), WhiB4 (Rv3681c), WhiB5(Rv0022c), WhiB6 (Rv3862c), and WhiB7 (Rv3197A). All WhiB proteins are transcriptional regulators and, except for WhiB2, they all have been described to display a disulfide reductase activity (Figure 2) [94]. All Mtb WhiB proteins are important for Mtb survival within the host, even though only WhiB1 and WhiB2 are essential proteins. Mtb WhiB proteins are able to detect stress signals (through their Fe-S cluster), allowing Mtb to adjust its metabolism to survive under stress conditions in macrophages and to enter dormancy [3]. In the whole, Mtb WhiB proteins are involved in regulation of virulence (WhiB1, WhiB3, WhiB4, WhiB6), antibiotic resistance (WhiB7), and the regulation of cell division (WhiB2) [93,95]. Mtb WhiB proteins have low sequence homology but all contain an O₂- and NO-sensitive Fe₄S₄ cluster, raising the question as to how they can support distinct cellular function within Mtb. Recent studies shed light on the biochemical function of WhiB as transcriptional factors and sensors of NO and O₂, and suggest that WhiB evolution has created diversity in protein–protein interactions, Fe-S sensitivity and ability to bind DNA [93].

4. How ROS, RNS and Fe Starvation Stresses Affect Iron-Sulfur (Fe-S) Cluster Proteins

We discussed previously (part 2) the different stresses which Mtb has to battle. Fe-S clusters are good targets for ROS and RNS and are susceptible to Fe starvation. Here, we will describe at a molecular level how these stresses can disrupt Fe-S clusters in bacteria. In some cases, the disruption is deleterious and leads to the inactivation of the protein by breaking the cluster. In other cases, the cluster disrupted by the stress (leading to its degradation or its modification) is used as a sensor, as in the case of transcriptional regulators.

4.1. Fe-S Clusters Sensors of Reactive Oxygen Species

Fe₄S₄ clusters are sensitive to oxidative degradation, in particular solvent-accessible Fe₄S₄ clusters, and as such are used at times as sensors of oxidative stress. In E. coli, a well-known case is the fumarate and nitrate reductase FNR, a transcriptional regulator that controls the expression of a large regulon of more than 100 genes in response to changes in oxygen availability [96,97]. FNR is active when it acquires, under anaerobic conditions, a Fe₄S₄ cluster in the +2 oxidation state. The presence of this cluster per monomer promotes protein dimerization and site-specific DNA binding, facilitating activation or repression of target promoters in response to anaerobiosis [98,99]. FNR directly senses O₂ through the lability of its Fe₄S₄ cluster. As oxygen levels rise, the oxygen-labile Fe₄S₄ cluster of FNR is oxidized and converted to a variety of protein-bound clusters, including a [Fe₃S₄]¹⁺ and a [Fe₂S₂]²⁺, along with an unexpected Fe₃S₃ cluster [100,101]. Conversion of the [Fe₄S₄]²⁺ cluster to a lower cluster nuclearity presumably alters FNR’s conformation, unable to bind DNA. Recent structural analysis of FNR from Aliivibrio fischer explained the extremely fine-tuned dimer–monomer equilibrium, giving insights into the DNA dissociation mechanism [102]. In E. coli, the redox-sensitive transcriptional regulator SoxR also responds to ROS. SoxR is activated under oxidative conditions by the oxidation of its [Fe₂S₂]¹⁺ cluster into the [Fe₂S₂]²⁺ form [103,104]. The recent structure of SoxR from E. coli shows that the Fe₂S₂ is completely solvent-exposed, explaining how it can sense small molecules [105] such as superoxide or redox-cycling drugs such as paraquate, menadione and phenazine methosulfate [106]. Once the SoxR cluster is oxidized, SoxR activates the expression of SoxS, which in turn activates a large number of genes (superoxide dismutase, DNA repair nuclease, oxidation resistant enzymes), thereby restoring some essential cellular functions [107,108,109,110].

4.2. Fe-S Cluster Sensors of Reactive Nitrogen Species

The interaction of nitric oxide with Fe-S cluster proteins results in degradation and breakdown of the cluster to generate dinitrosyl iron complexes (DNICs). Like for oxidation of clusters, the formation of DNICs from Fe-S clusters can lead to activation of a regulatory pathway or the loss of enzyme activity. At low concentrations, NO is a signaling molecule that several bacterial proteins sense; among the Fe-S containing proteins, the best studied is the transcriptional regulator NsrR (nitric-oxide-sensitive response regulator) that functions as a regulator of NO-induced stress response in many bacterial species. In E. coli, ≥60 genes are under NsrR control, the principal target being the NO stress response (hmp, which encodes a flavohaemoglobin that converts NO to nitrate under aerobic conditions and nrfA) and general stress response (sodB) [111,112]. NsrR is an Fe₄S₄ protein [113] with a unique coordination of the cluster at the interface of the two subunits of a dimer, by three Cys residues of one subunit and an Asp residue of the other. In its Fe₄S₄ form, NsrR is able to bind DNA and thus represses the cell’s response to NO stress. Upon exposure to NO, the cluster undergoes a rapid, complex, nitrosylation reaction resulting in the loss of DNA-binding (because of conformational change and/or change in affinity for target DNA) and the formation of a mixture of Fe-nitrosyl species [114]. SoxR and FNR are partially destroyed by NO [115,116], showing that they play a minor role as NO sensors.

4.3. The Case of Mtb WhiB Proteins

Mtb does not contain redox sensors such as SoxR, NsrR or FNR. Instead, Mtb contains WhiB proteins as transcriptional regulators, which sense NO and O₂ stresses. Of the seven WhiB-like proteins of Mtb, only WhiB1, WhiB3, WhiB4 and WhiB6 have been investigated at a molecular level. As mentioned before, the Fe₄S₄ clusters of WhiB proteins (for those which were characterized) have different sensitivity towards NO and ROS. WhiB1 expression is regulated by cAMP-CRP [117,118] (Figure 3). Under its holo-form (Fe₄S₄), WhiB1 cannot bind DNA; its cluster is stable to O₂ and remarkably sensitive to NO, reacting 10⁴-fold faster than its reaction with O₂ [119]. During Mtb infection, the Fe₄S₄-WhiB1 protein detects NO in the microenvironment through its Fe₄S₄ cluster [119]; it reacts with eight NO molecules, leading to the cluster disassembly by nitrosylation, releasing the apo-WhiB1 and nitrosylated-WhiB1, able to bind specific DNA sequences, acting as a transcriptional regulator [120]. The nitrosylation and cluster degradation are thought to follow the following mechanism: [Fe^II₂Fe^III₂S₄(Cys)₄]²⁻ + 8NO → 2(Fe^I₂(Cys)₂(NO)₄) + S²⁻ + 3S° [121]. The apo-form of WhiB1, as well as nitrosylated holo-WhiB1, bind to the whiB1 promoter region to repress its own transcription. In addition, WhiB1 regulates the secretion of ESX-1 [120] (a virulence factor that interrupts innate immune mechanisms) and represses transcription of groEL2 (an essential chaperonin), assisting entry into the dormant state of Mtb [122]. The apo-form of WhiB1 is also endowed with a disulfide reductase activity of the α(1,4)-glucan branching enzyme GlgB [123] and was reported to interact with the principal sigma factor SigA [124,125], which is essential for virulence in Mtb [126]. By similarity with other WhiB proteins, Mtb WhiB2 likely contains a Fe₄S₄ cluster, although its presence remains to be demonstrated. WhiB2 is regulated in response to antibiotics [127], and to ROS and heat-shock through SigB protein (Figure 3) [128,129,130]; a direct interaction between ROS and the WhiB2 cluster has not been reported. WhiB2 was recently reported to interact with SigA factor [124]. Mtb WhiB3 is positively regulated by PhoP [131] in response to acidic environments both in vitro and in macrophages [132,133], by phosphorylated RegX3 in response to low pH [134] and by MprAB through upregulation of Rv0081, predicted to induce the whiB3 operon [117]. WhiB3 harbors a Fe₄S₄ cluster that reacts with NO and slowly with O₂ [135], likely through the mechanism described above for WhiB1. Once the cluster is lost, cysteine residues, initially involved in Fe-S coordination, form intramolecular disulfide bonds that enable WhiB3 to interact strongly with DNA. Upon NO stress, WhiB3 induces lipid (polyketides, pks) production [136], which are responsible for arresting the host cell cycle of infected macrophages at the G₀/G₁ phase, thus manipulating the immune system [137], pathogenicity (through SigA interaction) [124,138] and lipid anabolism to modulate macrophage response (through fbpA, necessary for glycolipid trehalose dimycolate production) [136]. WhiB4 is over-expressed in the host macrophages and thus plays a role in the in vivo persistence. As-purified WhiB4 contains a Fe₂S₂ cluster, while after reconstitution it shields an Fe₄S₄ cluster [139]. Both Fe₂S₂ and Fe₄S₄ clusters are extremely NO- and O₂-sensitive [139,140]. The Fe₄S₄ cluster seems more sensitive to O₂ than those reported for other WhiB such as WhiB3 and WhiB1, suggesting a specific role of WhiB4 in redox sensing [140]. Holo-WhiB4 is not competent to bind to DNA, whereas apo-WhiB4 (with oxidized cysteines) binds strongly to DNA. Apo-WhiB4 binds to the promotor of ahpC, which encodes a protein responsive in suppressing oxidative stress [121]. The apo-form display also a disulfide reductase activity [140]. Like most of the WhiB proteins, WhiB4 interacts with SigA [124]. WhiB5 is regulated by oxidative stress and regulates the transcription of a number of genes, including sigM (alternative RNA polymerase sigma factor) and the genes for two type VII secretion systems, ESX-2 and ESX-4 [141]. The presence of an Fe-S cluster and the molecular mechanism in response to the oxidative stress remain to be discovered. When the Fe₄S₄-containing form of WhiB6 interacts with ROS and NO, the cluster undergoes degradation and the apo-WhiB6 and WhiB6-DNIC are formed, respectively [95]. WhiB6-DNIC suppresses ESX-1 expression, while WhiB6-DNIC and reduced apo-WhiB6 induce the DosR regulon in order to maintain the integrity of the granuloma and prepare the cell for latency [95]. WhiB6 also regulates the expression of espACD [95]. Like for WhiB2 and WhiB5, almost nothing is known regarding WhiB7, beyond its expression profile and its protein disulfide reductase activity [93]. It plays a role in antibiotics resistance [142]. WhiB7 is reported to be activated by oxidative stress, although the mechanism is not known. Like all WhiB proteins, it contains the four conserved cysteine residues that are able to bind the redox active Fe-S cluster, suggesting that it is also able to shield a Fe-S cluster. Recently, structural studies have demonstrated the presence of a Fe₄S₄ cluster in WhiB7 [143]. WhiB7 was reported to regulate a significant number of genes, such as cut2, erm, hflX, tap, Rv0263, Rv1257c, Rv1988, and Rv1473 [144], and there are several evidences of its interaction with SigA [124,143,145,146]. In summary, unlike E. coli, which has specialized proteins to respond to O2 and NO stresses, Mtb relies on a single family of proteins, WhiB, to sense these stresses. Through WhiB proteins, one notices how Fe-S clusters are critical for Mtb virulence and persistence.

4.4. Fe-S Enzymes and ROS/RNS

At high concentrations, or if conditions persist, ROS and NOS can be deleterious to Fe-S cluster enzymes; the resulting apo-enzyme may lead to protein degradation and cell death in case of essential enzymes. Sometimes, the apo-enzyme can be repaired by adding Fe²⁺ and a reducing agent [147]. Superoxide and hydrogen peroxide can inactivate Fe₄S₄ enzymes, including dehydratases and radical-SAM enzymes. The common point of these enzymes is their solvent-exposed Fe₄S₄ cluster; it is necessary for their function, but at the same time it endangers them. Many years ago, Fridovich and Flint discovered that superoxide rapidly inactivates the Fe₄S₄ family of dehydratases, including key enzymes of the branched-chain and TCA pathways (dehydratases, aconitase and fumarase and isopropylmalate isomerase) [148,149,150]. This is also true for α,β-dihydroisovalerate and dehydrogenase 3-Deoxy-d-Arabinoheptulosonate 7-Phosphate Synthase [22,151]. The damage occurs when superoxide directly oxidizes the Fe-S cluster, converting the [Fe₄S₄]²⁺ form to an unstable [Fe₄S₄]³⁺, which releases iron (Fe²⁺). The resultant [Fe₃S₄]¹⁺ cluster lacks the catalytic iron atom, so that the enzyme is inactive and the pathway with which it is involved fails. Most of the time, the [Fe₃S₄]¹⁺ cluster is converted to [Fe₂S₂]²⁺ that can be ultimately degraded. After exposure to air, the quinolinate synthase [Fe₄S₄]²⁺ cluster is converted to a stable [Fe₂S₂]²⁺ [152]. The rate constants with which dehydratase clusters react with superoxide and hydrogen peroxide are extremely high: 3 × 10⁶ M⁻¹ s⁻¹ and 4 × 10³ M⁻¹ s⁻¹, respectively [153]. It is not surprising, as these oxidants are small and can enter the active sites of dehydratases which are solvent-exposed. SAM-superfamily enzymes—which use also a solvent exposed Fe₄S₄ cluster to bind SAM and/or substrate—are also rapidly inactivated when they are exposed to oxygen. In vitro spectroscopic studies on Fe₄S₄ enzymes exposed to air or titration with H₂O₂ indicate stable or semi-stable breakdown intermediates. The Fe₄S₄ cluster of MiaB from Thermotoga maritima degrades into an [Fe₂S₂]²⁺ cluster [154]. An [Fe₂S₂]²⁺ cluster with cysteine persulfides was also observed by X-ray crystall ography in HydE enzyme [155]. None of the Fe-S-containing enzymes in Mtb (Table 1) have been investigated upon exposure to ROS and/or NO at a molecular level, but we can hypothesize that similar degradation mechanisms occur and therefore that essential metabolic pathways involving Fe-S enzymes may be hampered.

4.5. Iron Starvation and Fe-S Clusters

Fe-S clusters are formed from iron (Fe²⁺ and/or Fe³⁺) and inorganic sulfide. Therefore, iron starvation in bacteria has a direct impact on the Fe-S cluster biogenesis machineries (dedicated to assembling Fe-S clusters from L-cysteine and iron) [156,157,158]. Nevertheless, Mtb contains only the SUF system, which is functional under iron limitation (as discussed above). Whether this is due to higher affinity of the SUF components to small iron chelate or iron-binding proteins is not known yet. Mtb manages with this stress with the synthesis of mycobactin siderophore, which chelates ferric iron from host storage proteins in the phagosome [31].

5. Activation of Mtb Fe-S Biogenesis and Metabolism upon ROS, NO and Fe Starvation Stresses

Mtb alleviates ROS and RNS partially through Fe-S cluster proteins involved in redox sensing, gene regulation and DNA repair for persistence, such as WhiB regulators [132,135,136,140,159]. While these findings underscore the importance of Fe-S clusters in ensuring stress tolerance and survival of Mtb, what do we know about Fe-S biogenesis in the human pathogen Mtb?

In bacteria, three pathways for Fe-S cluster biogenesis and delivery have been identified: the ISC, NIF and SUF systems [160]. The ISC system is considered as the housekeeping system, SUF acts as a backup in stress conditions, and NIF is specialized in the maturation of nitrogenase in N₂-fixing bacteria. Although made of different components, these three systems facilitate Fe-S cluster biogenesis following the same basic principles: a cysteine desulfurase takes sulfur from L-cysteine and transfers it as a persulfide onto a scaffold protein, which also receives Fe²⁺ and electrons to build an Fe-S cluster that is then transferred to cellular apo-protein targets via carrier proteins (Figure 4) [160]. Whereas some bacteria such as E. coli contain several Fe-S biogenesis machineries (SUF and ISC), interestingly, some pathogens, such as Staphylococcus aureus and Mtb, contain only the SUF system as Fe-S assembly machinery (Figure 4) [5,161]. An iscS gene (cysteine desulfurase, Rv3025c) is present outside the suf locus but corresponds to a separate ORF, and it is not surrounded by other isc genes [5]. Therefore, the SUF machinery appears to be the primary Fe-S system in Mtb. Based on extensive work performed on the SUF system of E. coli (sufABCDSE), it is known that (i) the SufS-SufE complex provides sulfur from L-cysteine; (ii) the SufB-SufC-SufD complex acts as a scaffold for the assembly of Fe-S clusters, and that (iii) the SufA protein is a Fe-S transporter [162] (Figure 4). However, despite intensive studies, the nature and ligand of the Fe-S cluster on the E. coli SufBC₂D complex is still unknown and mysterious. The SUF system from the Gram(+) Bacillus subtilis (sufCDSUB) is also well studied, in particular the SufS and SufU proteins, which are implicated in the sulfur production of L-cysteine. As-purified SufU contains one zinc ion per protein essential to enhance the cysteine desulfurase activity of SufS [163]. SufU from B. subtilis was also proposed as an Fe-S cluster containing protein after reconstitution, although the FeS-SufU cannot promote activation of the SufS cysteine desulfurase activity [164]. In this study, only the apo-SufU (no Zn, no Fe-S) can activate SufS, in striking contrast to Selbach’s data [164]. The characterization of the SufB, SufC and SufD proteins, which likely form a complex based on E. coli data, suffers from SufB protein instability, preventing its function from being determined. Interestingly, the E. coli and B. subtilis SUF systems show a series of differences, from genetic organization to genetic composition, and were recently shown to exhibit different efficiency in maturing heterologous Fe-S targets [165,166]. In Mtb, the suf operon is composed of seven genes, namely Rv1460(sufR), Rv1461(sufB), Rv1462(sufD), Rv1463(sufC), Rv1464(sufS), Rv1465(sufU) and Rv1466(sufT). From its genetic composition, it more resembles the B. subtilis SUF system than that of the E. coli one, even though SUF from Mtb contains the additional sufT gene. sufT is not unique to Mtb or mycobacteria. Indeed, in prokaryotes, bioinformatics analysis revealed that 70% of genomes encoding SufBC also encoded a SufT, and 49% of these are associated with the suf operon [167]. In the last decade, Mtb SUF proteins have been poorly explored at a molecular level, but recently some exciting studies have been reported. The SufR homologue Rv1460 is a repressor of the Mtb suf operon. It contains, after purification, an Fe₂S₂ cluster, proposed to be coordinated by three cysteine residues (Cys203, C216, and C244) [168]. A recent study suggests that SufR contains instead a Fe₄S₄ cluster that is a sensor of NO to support persistence by reprogramming Fe-S cluster metabolism and the bioenergetics of Mtb [169]. SufT is proposed to act as an accessory factor in Fe-S biogenesis in Mtb, since it is dispensable for growth of Mtb under standard culture conditions, required under conditions of iron limitation [170], and SufT loss does not increase susceptibility to oxidative stress [170]. Very recently, Mtb SufT protein was shown to interact with SufS and SufU and to maintain the activity of Fe-S cluster proteins during normal growth conditions, and under environmental settings that enforce a high demand of Fe-S clusters [171]. Characterizations of Mtb SufS and SufU will certainly provide additional information and interesting insights into the sulfur production. SufB, SufC and SufD proteins from Mtb interact in cellulo [5]; however, the structural and functional characterization of these proteins remain also to be carried out in order to determine whether, like the E. coli SufBC₂D, they play a scaffold role.

Mtb suf operon is essential for the viability of Mtb under normal growth conditions [5,48,49]. More recently, Rv1461(sufB), Rv1462(sufD), Rv1463(sufC) and Rv1464(sufS) genes were proved to be vulnerable [52] (

Table 2. Genes predicted to be implicated in Fe-S biogenesis in Mtb and their essentiality.

Essentiality Identification Method	Rv1460	Rv1461	Rv1462	Rv1463	Rv1464	Rv1465	Rv1466
Through transposon [47,48,49,50]	No	Yes	Yes	Yes	Yes	Yes	Yes
Through CRISPR interference/VI[52]	No VI: −3.9	Yes VI: −7.78	Yes VI: −7.60	Yes VI: −9.29	Yes VI: −5.38	No VI: −1.14	No? VI: 0.08

). The Mtb suf operon is under both transcriptional and post-translational regulation by SufR [168] and SufB, respectively, even though the mechanisms of these regulations are not well understood. SufR is involved in the regulation of the Mtb suf operon, acting as a transcriptional repressor of its own expression in Mtb, and also as a repressor of the suf operon [168]. The Fe-S cluster of Rv1460 is required for its function as cysteine-to-serine mutants alter SufR function. SufB from Mtb possesses an intein whose splicing appears to have a critical role in modulating interaction with SufC and SufD [172]. Recently, the Mtb SufB intein was shown to possess high sensitivity for oxidative and nitrosative stress when expressed in E. coli [173]. High levels of ROS and NO inhibit SufB splicing. The splicing inhibition is proposed to be an immediate, post-translational regulatory response that can be either reversible, by inducing precursor accumulation, or irreversible, by inducing N-terminal cleavage, which may potentially channel Mtb into dormancy under extreme NO and oxidative stress [173].

The Mtb Suf operon is upregulated under nitrosative (NO) and oxidative conditions (H₂O₂), stressors of the innate immune response [14,174]. Additionally, microarray-based expression profiles of the suf Fe-S cluster assembly genes show that Rv1460-Rv1466 genes are highly upregulated upon exposure to H₂O₂, NO, and within macrophages infection [3]. The upregulation of the suf operon is likely to compensate for the loss of Fe-S clusters due to oxidative stress in Fe-S enzymes such as aconitase, quinolinate synthase and other Fe-S enzymes. Moreover, Mtb SUF components are induced during iron starvation [175], a process experienced by the pathogen in host tissues, indicating that Fe-S assembly and therefore Fe-S metabolism may be important in the establishment of latent infection. Intriguingly, suf genes are downregulated during hypoxia, a situation encountered in granulomas [3]. Therefore, interfering with Fe-S metabolism (SUF Fe-S biogenesis system) in Mtb offers a powerful strategy for eliminating TB (Figure 5). First, this will impose a pleiotropic effect on Mtb’s ability to synthesize the Fe-S proteins, some of which are essential in bioenergetics and central metabolism. Second, this will disarm Mtb by removing its ability to synthesize Fe-S for detecting environmental stress signals (through WhiB proteins). The loss (or inactivity) of the redox sensors WhiB would prevent the bacterium from mounting metabolic adaptations to enter a latent state. Importantly, the SUF system is absent in humans where Fe-S biogenesis is sustained by the ISC machinery [176]. Therefore, disrupting Fe-S cluster metabolism by targeting the Mtb SUF system is likely an exceptional route for developing novel anti-TB agents. So far, no inhibitors of the Mtb SUF machinery have been discovered.