Coordination of the Fe-S cluster biogenesis network by the sRNA RyhB in E. coli

ABSTRACT

Iron (Fe) plays critical roles as enzyme cofactor involved in key biological processes but can also lead to toxicity by catalysing the formation of highly damaging reactive oxygen species. To stabilize Fe and perform catalysis, most organisms rely on Fe-S clusters, which are fundamental and evolutionary ancient cofactors. In E. coli, two distinct pathways for the biosynthesis of Fe-S cluster exist: the three-part iscR-SUA-hscBA-fdx-iscX (ISC-HSC) operon and the sufABCDSE (SUF) operon. The iscR-SUA section of the ISC-HSC operon is regulated at the promoter level by the IscR transcription factor and post-transcriptionally by the small RNA (sRNA) RyhB. The SUF operon is regulated by a combination of transcription factors, including the Fe-sensing Fur, the Fe-S using IscR, and the oxidative stress responsive OxyR. Here, we show evidence that the sRNA RyhB regulates the hscBA-fdx-iscX part of the ISC-HSC operon as well as part of the SUF operon. RyhB orchestrates a complex pattern of expression of the iscR-SUA-hscBA-fdx-iscX operon during Fe starvation. This results in increased level of iscR and constant expression of iscSUA, encoding the scaffold for Fe-S cluster formation. However, the third part of the operon, hscBA-fdx-iscX, encoding a chaperone that facilitates Fe-S cluster transfer, is repressed by RyhB during Fe starvation. Furthermore, RyhB represses part of the sufABCDSE transcript, which counteracts Fur derepression. Overall, RyhB represses both ISC and SUF systems under iron starvation, to reduce Fe-S biogenesis under such limiting conditions.

SIGNIFICANCE STATEMENT

Fe-S cluster biosynthesis proceeds through various pathways, including the ISC-HSC (iscR-SUA-hscBA-fdx-iscX) and SUF (sufABCDSE) machineries in bacteria. Although the expression of these Fe-S cluster machineries has been mostly studied at the promoter level by the Fe-sensing transcriptional factor Fur, our findings provide evidence that both ISC-HSC and SUF operons are regulated post-transcriptionally by a regulatory RNA, RyhB. Specifically, RyhB strongly represses the HSC proteins of the ISC-HSC operon to the benefit of ISC and SUF. Overall, this suggests that the sRNA RyhB could play a key role in potentially modulating the trafficking of Fe-S clusters by ISC and SUF Fe-S cluster machineries.

Introduction

Iron (Fe) is one of the most abundant elements on the surface of the earth [1]. It serves as a critical cofactor in numerous essential enzymes involved in vital processes, including oxygen transport, ATP production, gene expression and DNA synthesis [2]. Fe is so critical for life that mammalian hosts use lactoferrin and transferrin to restrict Fe access, effectively preventing bacterial growth [3].

Due to its ability to exist under different oxidation states, i.e. Fe²⁺ and Fe³⁺, Fe can act as an electron donor and acceptor. For example, Fe is the cofactor for some ribonucleotide reductases, which converts ribonucleotides into deoxyribonucleotides [4]. Furthermore, Fe is key for catalase and some peroxidase enzymes, which detoxify the normal by-product H₂O₂ into H₂O and O₂ [5]. Notably, in the presence of high levels of H₂O₂, Fe itself becomes toxic by catalysing the formation of highly reactive and damaging hydroxyl radical (·OH) through the Fenton reaction (Fe²⁺ + H₂O₂ → Fe³⁺ + OH- + ·OH). Accumulation of ·OH can lead to oxidative stress, resulting in damage to DNA, RNA, proteins and lipids [6].

To prevent excess Fe accumulation and associated toxicity, bacterial cells utilize several proteins sequestering Fe such as ferritin and bacterioferritin [7,8]. Moreover, polyphosphate was also recognized as an effective Fe chelator that shields against Fenton reaction [9]. Another defence strategy involved the production of potent antioxidants such as glutathione, which can scavenge reactive oxygen species and prevent further damage [10].

Fe-S clusters are versatile protein prosthetic groups known to perform four major functions in cells: electron transport, catalysis, sulphur (S) donation and environmental sensing and are fundamental and evolutionarily ancient cofactors used by most organisms [11]. Fe-S clusters play a pivotal role in transferring electron through various mechanisms [12]. In E. coli, there are about 140 proteins using Fe-S clusters for many biological processes, including respiration and DNA repair [13–15].

One of the remarkable advantages of Fe-S clusters is their ability to interact with free radicals and other highly reactive species and undergo rapid and reversible structural changes. This flexibility allows Fe-S clusters to serve as a sensor to changes in the local environment, such as alterations in pH or redox potential [16]. This flexibility makes Fe-S clusters highly versatile and indispensable in a wide variety of biological processes. Detailed account of Fe-S cluster assembly has been extensively reviewed previously [15,17–19].

Another strategy to regulate Fe homoeostasis involves the Fe-dependent transcription factor IscR, which is part of the iscR-SUA-hscBA-fdx-iscX (ISC-HSC) operon that encodes genes involved in Fe-S cluster biogenesis. Under condition where Fe is present, the majority of cluster formation is due to Isc-Hsc system [12]. The IscS protein, a cysteine desulfurase, and IscU, a scaffolding protein are responsible for clustering. The transfer of clusters to recipient proteins is facilitated by the chaperones HscA and HscB. These chaperones, together with IscU, coordinate the folding of the client proteins and cluster assembly in a stepwise cycle [20]. IscR is an Fe-S protein that can bind to DNA sequences in the regulatory regions of target genes and modulate their expression. IscR has the ability to function as a repressor or activator and depends on the presence/absence of [2Fe-2S] clusters [21,22]. The iscR locus is located in the iscR-SUA operon and can be separated from the remainder of the transcript through the action of the small RNA RyhB [23]. The iscRSUA transcript is also regulated by two additional sRNAs, FnrS and OxyS during anaerobiosis and oxidative stress, respectively [24].

When cells are exposed to oxidative stress, the other Fe-S cluster formation system, sufABCDSE (SUF), potentially assumes the task of Fe-S cluster formation [18,25,26]. The SUF system can be the only Fe-S biogenesis pathway in some bacteria or protozoan, such as M. tuberculosis and P. falciparum [26,27], and it is absent in humans.

The sRNA RyhB is regulated by Fur, the main regulator of Fe homoeostasis in many bacteria [28]. In Fe-rich conditions, the Ferric uptake regulator (Fur) binds to Fe²⁺ and not only inhibits the transcription of ryhB but also represses numerous other genes involved in Fe uptake [29]. It has also been recently suggested that Fur could bind a Fe-S cluster to regulate intracellular Fe homoeostasis [30,31]. Upon Fe starvation, Fur becomes inactive, thereby relieving RyhB repression [32]. As Fe becomes limited in the medium, RyhB is expressed and forms base pairs that suppresses the translation of mRNA targets responsible for non-essential Fe-utilizing proteins [33]. This initiates an ‘Fe sparing’ response, redirecting the utilization of Fe within the cell.

We present new evidence suggesting the sRNA RyhB directly regulates the hscBA-fdx-iscX section of the ISC-HSC operon, as well as the complete SUF operon particularly under conditions of Fe starvation. Strikingly, RyhB organizes a three-fold pattern of expression of the iscR-SUA-hscBA-fdx-iscX operon. The sRNA promotes (i) increased levels of iscR and (ii) constant expression of iscSUA, which encodes the scaffold for Fe-S cluster formation in the absence of Fe. However, the third part of the operon, hscBA-fdx-iscX, which encodes a chaperone that facilitates Fe-S cluster transfer to receptor Apo-proteins, is (iii) considerably repressed by RyhB. Additional western blots and enzymatic assays indicate that RyhB coordinates the HSC-dependent shift from HscBA-IscU group of proteins to IscA-dependent proteins. Furthermore, RyhB represses the entire sufABCDSE transcript under low Fe conditions, which partially impedes Fur derepression of the operon. Overall, our data suggest that the sRNA RyhB is essential for ensuring the adaptative expression of ISC-HSC machineries according to Fe levels and helps to switch to SUF machinery during Fe depletion conditions or oxidative stress.

Results

Exploration of RyhB targetome related to Fe-S cluster machineries

Although we previously demonstrated that the iscR-SUA transcript was regulated by the sRNA RyhB [23], a growing body of evidence pointed to additional regulation of both hscBA-fdx-iscX and sufABCDSE operons by RyhB. Previously, a transcriptomic approach had identified 56 genes encoded by 20 mRNAs as targets of RyhB [34], including hscBA-fdx-iscX and sufABCDSE. Other groups studied the RyhB targetome by combining experimental and computational approaches [25,35]. They compared results from CopraRNA analysis [36,37], MAPS (MS2-affinity purification coupled with RNaseq) [38], and RIL-seq (RNA interaction by ligation and sequencing) [39] interactome and found hscB and sufB transcripts as putative targets of RyhB. Additionally, CopraRNA analyses suggested that sufB RNA sequence could potentially interact with RyhB [25]. Moreover, a ribosome profiling analysis indicated sufB, sufA and hscA transcripts as putative RyhB targets [40]. Prompted by this, we used Genome Browser to reanalyse our own data from a previously performed MS2 pull-down with RyhB [38] and found promising interactions between hscB (Figure 1(A) and S1A) and sufB (Figure 1(B) and S1B) with the RyhB sRNA [41]. Enrichment can be observed from sections within the iscRSUA transcript (6- to 19-fold), the hscBA-fdx-iscX transcript (3- to 14-fold) and sufABCDSE (1- to 8-fold) (Table 1). The weaker expression of the suf operon in Fe-rich LB medium conditions might explain the relatively lower enrichment as compared to iscRSUA and hscBA-fdx-iscX transcripts. These observations suggest that the whole machinery in charge of Fe-S cluster biogenesis (see Figure 1(C,D)) could be regulated by the sRNA RyhB.

Figure 1.

MS2-RyhB RNAseq (MAPS) data visualized using UCSC genome browser. Schematic representation of the (A) hscBA mRNA and (B) sufAB mRNA. Description of (C) iscRSUA and hscBA-fdx-iscX operons and (D) sufABCDSE operon. Newly characterized RyhB binding sites indicated by dashed red lines. Green lines indicate activation and red lines indicate repression. Blues and yellow circles represent Fe and S elements, respectively.

Table 1.

Short list of co-purified mRNAs using MS2-RyhB as bait in a rne131 ΔryhB background (ratio MS2-RyhB/RyhB control). Cells were harvested in both exponential (OD_600 nm = 0.5) and early stationary (OD_600 nm = 1.0) phases of growth. New putative target mRNAs are highlighted in grey. A list of enriched candidates at least 3X is available in Supplemental Table S1.

Gene	Ratio MS2- RyhB/Ctrl	Function/Activity	Re ference
iscR	12	DNA-binding trancriptionnal dual regulator	23
iscS	12	Cysteine desulfurase	23
iscS_iscR	17	Intergenic region	23
iscU	15	Scaffold protein Fe-S cluster assembly	23
iscU_iscS	6	Intergenic region	23
iscA	15	Fe-S cluster insertion protein	23
iscA_iscU	19	Intergenic region	23
hscB_iscA	9	Intergenic region	This study
hscB	14	Fe-S cluster biosynthesis co-chaperone	This study
hscA_hscB	14	Intergenic region	This study
fdx	7	Reduced ferredoxin	This study
iscX_fdx	5	Intergenic region	This study
iscX	3	Intergenic region	This study
sufA	6	Fe-S cluster insertion protein	This study
sufB_sufA	8	Intergenic region	This study
sufB	4	Fe-S cluster scaffold complex subunit	This study
sufC	1	Fe-S cluster scaffold complex subunit	This study
sufD	0	Fe-S cluster scaffold complex subunit	This study
sufS	0	L-cysteine desulfurase	This study
sufE	1	Sulphur carrier protein	This study

RyhB sRNA represses both hscBA-fdx-iscX and sufABCDSE transcripts

To validate these 2 new candidate targets of RyhB, we overproduced RyhB with a pBAD promoter from a plasmid in both ΔryhB and ΔryhBΔfur strains and used northern blots for RNA detection. Because the transcriptional regulator Fur blocks sufABCDSE transcription initiation in LB, we used ΔryhBΔfur strains to constitutively express the sufABCDSE transcript (Figure S8A). In the presence of overproduced RyhB, we observed a rapid decrease of hscBA-fdx-iscX (Figure 2(A); Figure S2A) and sufABCDSE (Figure 2(B); Figure S2B) mRNAs, suggesting a negative regulation at the post-transcriptional level. Next, we used the lacZ reporter gene fused in-frame with the coding sequence of hscBA and sufAB (HscBA-LacZ and SufAB-LacZ fusions; Figure S3) based on RyhB potential binding site found by IntaRNA or CopraRNA respectively. β-galactosidase assays with pBAD-ryhB confirmed that RyhB represses translation of HscBA-LacZ (80%) and SufAB-LacZ (30%) translational fusions (Figure 2(C)).

Figure 2.

Repression of two mRNAs involved in Fe-S cluster biogenesis by RyhB. Northern blot analysis of hscBA-fdx-iscX and sufABCDSE mRNAs. The expression of RyhB from a pBAD promoter was induced by addition of 0.1% arabinose (Ara) when cells reached an OD_600 nm = 0.5 (A) hscBA-fdx-iscX (hscA probe) mRNA in ΔryhBΔara background, and (B) sufABCDSE (sufE probe) mRNA in ΔfurΔryhBΔara background. The empty vector pNM12 was used as a control. Both 16S and 5S rRnas were used as a loading controls. Data are representative of two independent experiments. (C) β-galactosidase assays using HscBA-lacZ and SufAB-LacZ (translational) fusions with expression of RyhB from a pBAD-ryhB vector (gray, (+)) and with an empty vector (white, (-)) in a ΔryhBΔara or ΔfurΔryhBΔara background respectively. Arabinose (0.1%) was added when cells reached an OD_600 nm of 0.1. Samples (N = 3, mean ± SD) were taken at OD_600 nm = 1.0. ****p < 0.0001, **p < 0.0012, unpaired two-tailed Student’s t test. Lead acetate (PbAc) probing of (D) γ-hscBA (+486 to +685) or (F) γ-sufAB with RyhB. γ-hscBA or γ-sufAB were incubated for 15 min in the absence (-) or in the presence of (+) RyhB (0.1 μM) prior to addition of PbAc. Ctrl; non-reacted samples, OH; alkaline ladder, T1; RNase T1 ladder. Numbers on the left indicate nucleotide position relative to + 1 transcriptional start site. Putative RyhB pairing site is indicated with a purple (binding site 1 on hscBA) or blue line (binding site 2 on hscBA or binding site on sufAB). AUG start site is indicated in green. (E) in vitro validated pairing between hscBA and both pairing site 1 and site 2 for RyhB sRNA. hscA AUG start site is indicated in green and Shine Dalgarno (SD) is underlined. (G) in vitro validated pairing between sufAB and RyhB sRNA. sufB AUG start site is indicated in green.

To identify RyhB sRNA binding sites on both 5’-radiolabeled hscBA and sufAB, we performed in vitro probing assay with lead acetate (PbAc) in the absence or presence of RyhB. In the case of the hscBA transcript (+486 to + 685 from hscB promoter [42], γ-hscBA), we observed two sites protected by RyhB. Binding site 1 (BS1, Figure 2(D); Figure S4A) located between nucleotides +40 to + 52 shows a weaker protection than the binding site 2 (BS2, Figure 2(D); Figure S4B), which displays a clear protection for nucleotides +92 to + 99 containing the Shine-Dalgarno sequences (SD). Both binding sites are represented in Figure 2(E). To validate that the protection is due to RyhB base pairing to hscBA, we introduced mutations in RyhB seed pairing to hsc to obtain RyhB_GGAA (Figure S5A). This mutation on RyhB should disrupt binding on BS1 and BS2 because the same RyhB nucleotides can bind on both sites. Next, we performed electrophoretic mobility shift assays (EMSA) of γ-hscBA and results showed that γ-hscBA interacts with RyhB but not with RyhB_GGAA (Figure S5B; S5E). This was confirmed in vivo by β-galactosidase assays by using a translational HscBA-LacZ fusion (Figure S5F) with an empty vector and overexpression of RyhB or RyhB_GGAA. We observed that pBAD-ryhB_GGAA (Figure S5F; _GGAA) could not block hscBA translation contrary to pBAD-ryhB WT (Figure S5F; +). Notably, the same region on hscBA was identified by RIL-Seq [39]. We then introduced mutations in hscBA to obtain hscBA_mutBS1 and hscBA_mutBS2 (Figure S5A) and EMSA were performed on γ-hscBA_mutBS1 (Figure S5C; S5E) and γ-hscBA_mutBS2 (Figure S5D; S5E) in the presence of RyhB. The pairing of γ-hscBA_mutBS1 with RyhB was similar to γ-hscBA, indicating that RyhB does not bind at the site 1. In contrast, the pairing of γ-hscBA_mutBS2 with RyhB was completely abolished, suggesting a base pairing-dependent interaction between hscBA BS2 and RyhB. Taken together, these results confirm that RyhB negatively regulates hscBA-fdx-iscX operon probably through blocking translation by base pairing at the Shine-Dalgarno region.

In the case of the sufAB transcript (+366 to + 548 from sufA promoter, γ-sufAB), we observed a clear protection by RyhB binding to nucleotides +51 to + 74 located in the coding sequence of sufB (Figure 2(F); Figure S6). The binding site is represented in Figure 2(G) (BS). To validate that the protection depends on RyhB pairing to sufAB, we introduced mutations in RyhB to obtain RyhB_mut (Figure S7A) and performed electrophoretic mobility shift assays (EMSA) of γ-sufAB. Results showed that γ-sufAB interacts with RyhB but not with RyhB_mut (Figure S7B; S7D). This was confirmed in vivo by β-galactosidase assays using SufAB-lacZ fusion (Figure S7E) with an empty vector and overexpression of RyhB or RyhB_mut. We observed that pBAD-ryhB_mut did not block sufAB translation contrary to pBAD-ryhB (Figure S7E; mut and +). Moreover, the same RyhB-binding region on sufAB was identified by RIL-Seq [39]. These results suggest that RyhB regulates sufABCDSE operon by blocking translation. We then introduced mutations in sufAB to obtain sufAB_mut (Figure S7A) and EMSA were performed on γ-sufAB_mut in the presence of RyhB (Figure S7C; S7D). Results indicated the shift of γ-sufAB_mut with RyhB is clearly delayed compared to γ-sufAB, but this mutation is not sufficient to totally impair RyhB binding. This suggested that RyhB can interact with sufAB through base pairing in vitro.

Regulation of hsc and suf by Fe and transcriptional factors

Next, we wanted to determine the effect of endogenous RyhB on the expression of both hscBA-fdx-iscX and sufABCDSE operons. To do so, endogenous RyhB was induced with the Fe chelator 2,2’-dipyridyl (DIP) followed by β-galactosidase assays, which showed that endogenous RyhB is repressing translation fusions of HscBA-LacZ and SufAB-LacZ by 54% and 70%, respectively (Figure 3(A)). To investigate the effect of endogenous RyhB on the mRNA level, we conducted northern blots assays with hscA-specific, sufE-specific, and RyhB-specific probes. We observed a rapid decrease of hscBA-fdx-iscX mRNA after 5 min of RyhB induction by DIP in the WT strain and an increase after 20 min (Figure 3(B), lanes 1 to 5). In contrast, the ΔryhB strain showed no degradation of hscBA-fdx-iscX after 5 min of RyhB induction and a large increase after 20 min (Figure 3(B), lanes 6 to 10). The expression of hscBA-fdx-iscX mRNA was increased 1.6-fold after 30 min induction with DIP in the absence of RyhB as compared to WT background.

Figure 3.

Regulation by RyhB under Fe starvation conditions. (A) β-galactosidase assays using HscBA-LacZ and SufAB-LacZ translational fusions in WT (gray) or ΔryhB (white) background in LB after the addition of 250 µM DIP at an OD_600 nm of 0.1. Samples (N = 3, mean ± SD) were taken at OD_600 nm = 1.0. **p < 0.0019, ****p < 0.0001, unpaired two-tailed Student’s t test. (B) Northern blot analysis of hscBA-fdx-iscX (hscA probe), sufABCDSE (sufE probe) mRNA and RyhB, in LB after the addition of 250 µM DIP at an OD_600 nm of 0.5 for 10 min. (C) Northern blot analysis of sufABCDSE (sufE probe) mRNAs and RyhB in the presence of an Fe chelator. DIP was added in LB at an OD600_nm of 0.5 for 10 min, in WT, Δfur, ΔiscR and ΔoxyR in the presence (+) or in the absence (-) of ryhB allele. Both 16S rRNA and 5S rRnas were used as loading controls. Data are representative of two independent experiments.

As expected, the sufABCDSE mRNA increased slightly in WT strain 5 min after we added DIP and remained stable afterwards (Figure 3(B)). Notably in ΔryhB strain, there was a stronger increase of the suf mRNA after induction with DIP (Figure 3(B); lanes 6 to 10). The Suf system is known to be expressed under Fe depletion and oxidative stress conditions [22,43–45]. To coordinate this adaptative response, three transcriptional regulators, Fur, IscR and OxyR control the expression of the suf operon [46–49]. Fur represses sufABCDSE in the presence of Fe, as we can observe at time 0 before adding DIP (Figure 3(B), lane 1). However, during growth under Fe starvation, the repression of sufABCDSE transcription is lost due to the inactivation of apo-Fur (Figure 3(B) lanes 2 to 5). Moreover, RyhB represses sufABCDSE mRNA by a 3-fold factor if we compare the WT strain and ΔryhB (Figure 3(B) lanes 5 and 10 and Figure 3(C)). Notably, transcription of sufABCDSE is positively regulated by apo-IscR during Fe starvation condition. We previously demonstrated that RyhB decreases the stability by cleaving the iscRSUA mRNA without affecting iscR part of the mRNA [23]. The intergenic region between iscR and iscS forms a strong secondary structure that is responsible for the RyhB-dependent accumulation of iscR transcript [23]. We then investigated the steady-state expression of the sufABCDSE transcript in wild-type, Δfur, ΔiscR and ΔoxyR backgrounds with or without RyhB, cells grown in LB media in the presence of DIP for 10 min, which allows RyhB expression (Figure 3(C)). RyhB can repress sufABCDSE in WT (3.2-fold), Δfur (3.8-fold), ΔiscR (3.5-fold) and ΔoxyR (2.3-fold) if we compare RyhB + and - (Figure 3(C) lanes 3 to 8). Under these conditions, Fur is not active on sufABCDSE (compare Figure 3(C) lanes 1 and 3). We can observe a small increase of sufABCDSE mRNA without OxyR (see Figure 3(C) lane 1 and 7, and 2 and 8). These results support the model in which RyhB regulates sufABCDSE mRNA.

Next, we investigated OxyR activation by H₂O₂, leading to the induction of the sufABCDSE operon [44]. To do so, we added 600 µM H₂O₂ to our WT, ΔryhB, ΔoxyR and ΔoxyRΔryhB strains to induce oxidative stress and then monitored the sufABCDSE mRNA by northern blot (Figure S8C). As expected, sufABCDSE mRNA increased after 5 min with H₂O₂ treatment in WT and ΔryhB strains (Figure S8C lanes 2 and 7) and a smaller increase in ΔoxyR and ΔoxyRΔryhB strains (Figure S8C lanes 12 and 17). It has been suggested that H₂O₂ may inactivate E. coli Fur, possibly by oxidation of the bound Fe²⁺ [50]. This can explain the increase in RyhB expression after addition of H₂O₂ after 5 min. In contrast, OxyR can also activate transcription of fur gene during oxidative stress [51], which can possibly explain the 1.8-fold increase of RyhB in the ΔoxyR strain (Figure S8C lane 12). We observed that the sufABCDSE mRNA signal decreases 10 min after the addition of H₂O₂ (Figure S8C lanes 3, 8, 13 and 18). This could be due to the action of catalases (KatG), (KatE) and NADH peroxidase (AhpCF) that are responsible for the reduction of H₂O₂ levels [52]. We could observe an increase in sufABCDSE mRNA in ΔryhB and ΔoxyRΔryhB strains (Figure S8C lanes 8 and 18). These results suggest that RyhB can slightly regulate the sufABCDSE mRNA during oxidative stress.

Controlling Fe-S cluster formation across varied Fe levels

Next, we determined the expression of iscRSUA, hscBA-fdx-iscX and sufABCDSE during growth under different Fe concentrations. We measured the β-galactosidase activity of 4 translational fusions; IscR-LacZ; IscRS-LacZ; HscBA-LacZ; and SufAB-LacZ in both WT and ΔryhB backgrounds grown in minimum M63 medium with increasing FeSO₄ concentrations (0; 0.1; 0.5; 1 µM). We also monitored the expression of RyhB in these growth conditions (Figure S8B). We first investigated the expression of iscRSUA operon, which is self-repressed under Fe-rich conditions. In the presence of Fe, IscR acquires a [2Fe-2S] cluster to form holo-IscR and binds to the iscR promoter to repress its transcription [53,54]. As expected, we observed that IscR-LacZ expression decreases under high Fe concentration (Figure 4(A)). We then used the same IscR-LacZ fusion in the absence of RyhB (ΔryhB) and observed the same decrease in the expression of IscR, as compared to WT. We can see a difference in IscR-LacZ expression within the WT vs. ΔryhB, but the largest difference is in the iron-free media (Figure 4(A); WT vs. ΔryhB). The decreased activity of the IscR-LacZ fusion in the absence of RyhB correlates with increased IscRS-LacZ expression (6-fold increase), indicating that IscSUA proteins are more produced in the absence of RyhB (Figure 4(B); WT vs. ΔryhB). Expressing more IscSUA could produce more Fe-S clusters, resulting in an increase in holo-IscR, which subsequently leads to a repression of IscR-LacZ. These observations support our previous description that RyhB promotes the expression of the iscR transcript while concurrently reducing the iscRSUA transcript [23]. This also confirms the previously observed increased expression of IscS in ΔryhB strain growing in the absence of Fe [23,55]. Remarkably, we detected a strong consistency of the β-galactosidase activities in IscRS-LacZ from the WT strain even when we increased Fe concentration from no Fe to 1 µM FeSO₄ (Figure 4(B), grey bars). Our data suggest a double repression of IscRS-LacZ, by RyhB at low Fe level (0 and 0.1 uM) and by IscR feedback at higher Fe level (0.5 and 1 uM).

Figure 4.

Fe-S cluster Biogenesis in varied Fe levels. β-galactosidase assays using translational fusions in M63 0,2% glucose in the absence (-) or in the presence of FeSO₄ at 0.1 µM, 0.5 µM, or 1 µM, in a WT or ΔryhB background. (A) IscR-LacZ, (B) IscRS-LacZ (C), HscBA-LacZ and (D) SufAB-LacZ translational fusions. Samples were taken at an OD_600 nm of 0.7–0.9. Data are relativized to the absence (-) of FeSO₄ in WT background. Data represent the mean of three independent experiments ± SD. ****p < 0.0001, ***p = 0.0002, ns: p> 0.05, two-way ANOVA test.

The chaperone/cochaperone system HscA and HscB plays a role with IscU assembly scaffold. HscA and HscB are required for the assembly of Fe-S clusters as their inactivation reduces the formation of Fe-S cluster proteins [13,56,57]. For HscBA-LacZ translational fusion we observe a 3-fold increase between conditions without Fe and 0.5–1 µM FeSO₄ (Figure 4(C), grey bars), which is consistent with the β-gal in overexpression (Figure 2(C)) and with DIP (Figure 3(A)). While there is an increase of the activity in the ΔryhB strain if we compare with the WT strain (Figure 4(C)), there was nearly no variation through several Fe concentrations in the ΔryhB strain.

To complete the Fe-S cluster biogenesis regulation, we examined the SUF system using the SufAB-LacZ translational fusion. The most dramatic changes are observed in ΔryhB strain where SufAB-LacZ is 4-fold more expressed than WT in the absence of Fe (Figure 4D), which is consistent with results with DIP experiments in Figure 3(A,B,C). There is also a negative regulation by Fur on sufABCDSE transcription in the presence of Fe in the media (Figure 4(D); compare WT and ΔryhB at 0.5 and 1 uM FeSO₄). Thus, both RyhB and Fur repress the sufABCDSE operon at different levels of Fe concentration. This confirms previous observation of higher amounts of SufA, SufB, SufC, SufD, SufS and SufE proteins in poorer medium as compared to a rich medium [58].

Discussion

In the present study, we show new evidence that RyhB directly regulates the hscBA-fdx-iscX section of the ISC-HSC operon, as well as the complete SUF operon during Fe starvation. These findings provide valuable insights into the intricate regulation of Fe-S cluster biogenesis when cells are in a state of low Fe availability. We previously demonstrated the discoordinated regulation of the iscRSUA transcript through RyhB binding and subsequent iscR accumulation for cells growing under low Fe conditions [23]. Notably, this involves an increase in IscR levels, while IscS, IscU and IscA remained constant. This may serve as a mechanism to uphold cellular Fe-S homoeostasis under conditions of Fe limitation. These findings are in agreement with previous research demonstrating the pivotal role of RyhB for the maintenance of Fe homoeostasis during Fe scarcity [59].

One of the most striking observations from these results is the shifting behaviour of IscR-LacZ, IscRS-LacZ, HscBA-LacZ and SufAB-LacZ across various Fe conditions. Although IscS remains relatively stable across most Fe conditions, HscA increased with rising Fe concentration, underscoring its crucial role in regulating Fe-S biogenesis within the ISC-HSC system (Figure 4(C,E)). The strong repression of hscBA-fdx-iscX by RyhB suggests that RyhB may be involved in modulating the trafficking of Fe-S clusters into Apo-proteins. More experiments are required to address this specific question. These findings provide novel insights into the mechanism of RyhB-mediated regulation of Fe-S cluster biogenesis through alteration of the Fe-S fluxes.

Surprisingly, the repression of the whole sufABCDSE transcript under low Fe conditions by RyhB partially counteracts Fur derepression (Figure 4(D), Figure S9). This is a significant finding as it suggests that RyhB is involved in the transition from ISC-HSC to SUF operon during Fe depletion conditions or oxidative stress. The regulation of Fe-S cluster biogenesis is critical for the survival of the cell under stress conditions, and the role of RyhB in this process emphasizes its importance in the adaptive response of the cell to environmental stress. The induction of the Suf system during oxidative stress or Fe limitation is likely due to its ability to maintain Fe-S cluster biogenesis under these conditions. The SufB scaffold appears to be more resistant to destabilization by hydrogen peroxide, oxygen and Fe chelators than the IscU scaffold, making it a more suitable scaffold for Fe-S cluster assembly during oxidative stress [22,43]. Additionally, the Suf pathway may be well suited to acquire Fe when it is scarce, and its specific activity in vitro is higher than that of IscS alone or in complex with IscU at low L-cysteine concentrations and after exposure to hydrogen peroxide [60,61].

However, even in the presence of high levels of the SUF machinery, it appears that some Fe-S proteins do not fully mature in the absence of the ISC pathway [62,63]. This suggests that the SUF pathway may only meet the minimal Fe-S biogenesis requirements for essential Fe-S proteins during stress conditions. The regulation of the ISC-HSC and SUF operons is inherently intricate, likely dependent on several variables, including Fe concentration and the kinetics of each operon’s response. The repression of the entire sufABCDSE transcript under low Fe conditions by RyhB partially counteracts Fur-mediated derepression, suggesting that RyhB is involved in the transition from ISC-HSC to SUF operon during gradual Fe depletion or oxidative stress. This is reminiscent of another study showing an antagonistic effect on the expression of the erpA, a mRNA targeted by RyhB [63]. In this specific case, RyhB acts both as a repressor and an inducer or erpA by alleviating holo-IscR repression of erpA promoter [64]. In this case, RyhB acts as a repressor of erpA at low Fe level while IscR represses at high Fe level. Because of the combined actions of RyhB and IscR actions, erpA will be specifically activated at intermediary Fe level.

In summary, these findings underline the importance of both the ISC-HSC and SUF pathways in maintaining cellular Fe-S homoeostasis, as well as the complex regulatory mechanisms governing their interplay. Further research is imperative to fully elucidate the exact mechanism of the RyhB-mediated regulation of Fe-S cluster biogenesis, particularly in response to environmental stress such as oxidative stress and Fe limitation. Such insights have significant implications for the survival of the cell. Considering the absence of the SUF system in humans and its critical role in certain bacterial pathogens, further investigation in this area is also highly warranted.

Materials and methods

Strains and growth conditions

Derivatives of E. coli K-12 MG1655 strains were used in all experiments. Strains constructed by P1 transduction were selected for the appropriate antibiotic-resistant marker. Except as otherwise indicated, for cells carrying pNM12 and pBAD-ryhB, ampicillin was used at a final concentration of 50 µg/mL. See Supplemental Tables S2 and S3 for a complete description of strains, plasmids, and oligonucleotides used in this study. Cells were grown in rich medium (LB) or in M63 (20 mM KH₂PO₄, 40 mM K₂HPO₄, 15 mM (NH₄)₂SO₄, 1 mM MgSO₄, 2 g/L glucose) medium with no FeSO₄ or with 0.1 µM, 0.5 µM, or 1 µM FeSO₄.

MS2-affinity purification coupled with RNA sequencing

MS2 Affinity Purification Coupled with RNA-Seq experiment was performed as described previously [38]. Briefly, the MS2 aptamer was fused to the 5’-end of RyhB sRNA. After MS2 affinity purification, cDNA libraries were prepared with ScriptSeq v2 RNA-Seq Library Preparation Kit (Epicenter). Samples were then sequenced on a MiSeq Sequencing System (Illumina). We have used previously published data (GEO: GSE66519) and reanalysed them (National Library of Medicine: PRJNA1273173) with the Fragments Per Million mapped reads (FPM) normalization method [65]. We used Galaxy Project [66] and UCSC Microbial GenomeBrowser [67] to analyse and visualize data. The enrichment ratio is the MS2-RyhB/RyhB. Copurified targets from Fe-S cluster biogenesis are presented in Table 1. A list of genes enriched at least 3X is also available in Supplemental Table S1.

RNA extraction and northern blot analysis

Total RNA was extracted using the hot-phenol procedure as described previously [68]. When using M63 medium, cells were washed prior to the extraction. To induce gene expression from the pBAD vector, 0.1% arabinose was added when indicated. For the northern blot experiment, 5–10 µg of total RNA were loaded on a polyacrylamide gel (5–10% acrylamide 29:1, 8 M urea) or 20 µg on an agarose gel (1%, 30 mM Tricine, 30 mM Triethanolamine, 0.4 M Formaldehyde [69]). Then, RNA was electro-transferred to a Hybond-XL membrane (Amersham Bioscience) for a polyacrylamide gel or transferred by capillarity on a Biodyne B membrane (Pall) for an agarose gel. Cross-linking was performed by UV (1200 J). Prehybridization was performed in Church buffer [70] for radiolabeled DNA probes. Transcription of radiolabeled RNA probes for detection of the gene of interest was performed as previously described [23]. Oligonucleotides used for DNA and RNA probes are described in Supplemental Table S3. Membranes were then exposed to phosphor storage screens and analysed using a Typhoon Trio (GE Healthcare) instrument. Image Studio Lite software (LICOR) was used for densitometry analysis when applicable. Results reported here correspond to data from at least two independent experiments.

Lead acetate probing assays

Lead acetate probing assays were performed as previously described [71]. Briefly, 0.2 pmol of 5’-radiolabeled hscBA or sufAB transcripts (in vitro transcription described in Supplementary Data) was incubated in the presence or in the absence of 1 μM RyhB sRNA for 15 min at 37°C before addition of 5 mM PbAc for 2 min. H₂O was used for controls. Reactions were stopped by adding 10 μl of Loading Buffer II (LBII: 95% formamide, 18 mM EDTA, 0.025% SDS, 0.025% xylene cyanol-bromophenol blue) to the mix. Samples were loaded on polyacrylamide gels (8% acrylamide:bisacrylamide 19:1, 8 M urea) and migrated in TBE 1X at 38 W. Gels were dried and then exposed to phosphor screens for visualization using a Typhoon Trio (GE Healthcare) device.

β-galactosidase assays

β-Galactosidase assays were done as described previously [72]. Cells from overnight cultures were diluted 1/1000 in LB medium and 1/50 (~OD_600 nm = 2.0) in M63 0.2% glucose (no FeSO₄, 0.1 µM 0.5 µM and 1 µM) and grown at 37°C with vigorous agitation. RyhB was expressed endogenously in M63 medium or in LB in the presence of 2,2’-dipyridyl (DIP, 250 µM), added at an OD_600 nm = 0.5, as specified. RyhB is also expressed from a BAD promoter induced by addition of 0.1% arabinose at OD_600 nm = 0.5, as specified. Samples are taken at OD₆₀₀ ~1.0. Data represent the mean ± Standard Deviation (mean ± SD) of the V_max/OD_600 nm calculated value. See Supplementary Figure S3 and Supplementary Data for Materials and Methods details on the construction of lacZ fusions.

Western blot

Western blots were performed as previously reported [73]. Proteins were separated on SDS-PAGE gel and transferred to nitrocellulose membrane (Cytiva Life SciencesTM AmershamTM ProtranTM). The mouse monoclonal ANTI-FLAG® M2 antibody (Millipore Sigma) was used at a dilution of 1:1000. The IRDye 800CW-conjugated goat anti-mouse secondary antibody (Li-Cor Biosciences) was used at a dilution of 1:15,000. Western blots were revealed on an Odyssey infrared imaging system (Li-Cor Biosiences), and quantification was performed using the Image Studio Lite Version 5.2. The results reported represent data of three independent experiments.

Bacterial growth curve measurement

Bacterial strains were grown overnight in 3 mL of LB medium. Cultures were washed once in M63 minimal medium without iron. Cells were then diluted to a concentration of 6 x10⁸ cells/mL in a medium containing 10 mM glutamate and another batch without glutamate. The assay was performed in a Microtest plate, 96-well, flat base, polystyrene, sterile (Sarstedt), and growth was monitored using an Epoch 2 Microplate Spectrophotometer reader (BioTek) with the following settings: OD = 600 nm, Temperature = 37°C, Reading = every 10 min for 24 hr., continuous shaking.

Funding Statement

This work was funded by an operating grant [BMB 389354] from the Canadian Institutes of Health Research (CIHR) to EM.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Author contributions

E.M., K.P., D.L. designed the experiments. K.P., M-H.N., C-D.V-V, A-L.B., and S.P. performed the experiments. K.P. and T.C analysed MAPS data. E.M., K.P., C-D.V-V., and D.L. wrote the paper. All authors discussed the results and commented on the manuscript. E.M. supervised the project. All authors have read and agreed to the published version of the manuscript.

Data availability statement

The data that support the findings of this study are openly available in GEO at https://www.ncbi.nlm.nih.gov/geo/, reference number GSE66519. Supplementary materials can be found at https://doi.org/10.5281/zenodo.16846489.

Supplementary material

Supplemental data for this article can be accessed online at https://doi.org/10.1080/15476286.2025.2570040