Staphylococcus aureus SufT: an essential iron-sulfur cluster assembly factor in cells experiencing a high-demand for lipoic acid

Summary

S. aureus SufT is composed solely of the domain of unknown function 59 (DUF59) and has a role in the maturation of iron-sulfur (Fe-S) proteins. We report that SufT is essential for S. aureus when growth is heavily reliant upon lipoamide-utilizing enzymes, but dispensable when this reliance is decreased. LipA requires Fe-S clusters for lipoic acid (LA) synthesis and a ΔsufT strain had phenotypes suggestive of decreased LA production and decreased activities of lipoamide-requiring enzymes. Fermentative growth, a null clpC allele, or decreased flux through the TCA cycle diminished the demand for LA and rendered SufT non-essential. Abundance of the Fe-S cluster carrier Nfu was increased in a ΔclpC strain and a null clpC allele was unable to suppress the LA requirement of a ΔsufT Δnfu strain. Over-expression of nfu suppressed the LA requirement of the ΔsufT strain. We propose a model wherein SufT, and by extension the DUF59, is essential for the maturation of holo-LipA in S. aureus cells experiencing a high demand for lipoamide-dependent enzymes. The findings presented suggest that the demand for products of Fe-S enzymes is a factor governing the usage of one Fe-S cluster assembly factor over another in the maturation of apo-proteins.

Abbreviated Summary

Cells contain iron-sulfur (Fe-S) cluster utilizing proteins and proteins that facilitate cluster assembly. Lipoic acid (LA) is a product of the Fe-S cluster dependent enzyme LipA. The SufT Fe-S cluster assembly factor is essential in cells experiencing a high demand for lipoamide-dependent enzymes. We propose that the demand for products of Fe-S enzymes is a factor governing the usage of one Fe-S assembly factor over another in the maturation of apo-proteins.

Introduction

Iron (Fe) is necessary for the viability of most organisms. Fe is not readily available in environments and bacteria utilize multiple means to secure Fe. Once ferric iron passes across the cellular membrane, it is reduced to ferrous iron and is used to metallate proteins that require the Fe atom for function or to build Fe-containing prosthetic groups (reviewed here (Hammer & Skaar, 2011)). Much of the intracellular Fe is used to assemble inorganic prosthetic groups called iron-sulfur (Fe-S) clusters. Fe-S cluster assembly occurs in three steps (reviewed here (Py & Barras, 2010)). First, iron, sulfur, and electrons are combined on a molecular scaffolding protein to produce an Fe-S cluster. The Fe-S cluster is then trafficked to a target apo-protein by carrier molecules and subsequently inserted to produce the holo-protein.

Iron-sulfur cluster synthesis in S. aureus occurs upon the SufBCD molecular scaffold and it is an essential process ((Mashruwala et al., 2015) and Roberts et al., in preparation). The sulfur for Fe-S cluster synthesis is provided by the SufS cysteine desulfurase and the SufU sulfur carrier (Selbach et al., 2014, Selbach et al., 2010). The Fe and electron donors are unknown. SufA and Nfu are Fe-S cluster carriers that are responsible for shuttling the preformed Fe-S clusters to target apo-proteins and facilitating their maturation to holo-proteins (Rosario-Cruz et al., 2015, Mashruwala et al., 2015). The genes encoding for SufA and Nfu are dispensable for S. aureus viability. A strain lacking the low-molecular weight thiol bacillithiol has phenotypes that mimic strains lacking Fe-S cluster carrier molecules and these phenotypes are suppressed by overexpression of sufA, suggesting a role for bacillithiol in the maturation of Fe-S proteins (Rosario-Cruz & Boyd, 2015). S. aureus strains defective in Fe-S cluster assembly have decreased virulence and fitness in in vivo and in vitro models of infection (Mashruwala et al., 2015, Valentino et al., 2014, Mashruwala et al., 2016). Many of the factors that S. aureus uses to maturate Fe-S proteins are not present in higher eukaryotes suggesting that Fe-S cluster assembly is a viable antimicrobial target.

S. aureus SufA and Nfu can both transfer Fe-S clusters to apo-aconitase (apo-AcnA) at comparable rates in vitro (Mashruwala et al., 2015, Rosario-Cruz et al., 2015). However, in vivo Nfu is utilized for the maturation of aconitase during aerobic growth while SufA is dispensable (Mashruwala et al., 2016). Further experimentation found that sufA is negatively epistatic to nfu suggesting that the cell can utilize SufA for AcnA maturation in vivo. Therefore, S. aureus preferentially utilizes Nfu instead of SufA to maturate AcnA. The forces governing the usage of one Fe-S maturation factor over another remain to be fully defined and are an area of active research (Vinella et al., 2009).

We have recently identified an additional protein (SufT) that has a role in Fe-S cluster assembly in S. aureus (Mashruwala et al., 2016). The S. aureus SufT is composed entirely of a domain of unknown function (DUF) 59 (Luo et al., 2012, Weerapana et al., 2010, Hausmann et al., 2005). DUF59 domains are ~90 amino acids in length and contain one highly conserved cysteine (Mashruwala et al., 2016). In S. aureus, SufT is an auxiliary factor involved in the maturation of proteins that utilize Fe-S cluster cofactors (AcnA, LeuCD and IlvD) and its role increases under conditions of high-demand for Fe-S clusters.

Eukaryotic proteins containing DUF59 domains also have roles in Fe-S cluster assembly. Arabidopsis thaliana HCF101 contains a N-terminal DUF59 domain, is localized in the chloroplast and is required for photosystem I function (Lezhneva et al., 2004). HCF101 is also capable of binding a [Fe₄-S₄] cluster and transferring the cluster to an apo-protein resulting in the hypothesis that HCF101 is an Fe-S cluster carrier (Schwenkert et al., 2010). Human Fam96a and Fam96b (CIA2A and CIA2B in Saccharomyces cerevisiae) also contain DUF59 domains and are involved in the assembly of Fe-S clusters upon cytosolic and nuclear proteins (Stehling et al., 2013, Chen et al., 2012). The A. thaliana AE7 protein, which is a CIA2 homologue, is also necessary for the maturation of nuclear and cytosolic Fe-S proteins (Luo et al., 2012). Phylogenetic analyses found that nearly all eukaryotes possess a homologue of CIA2 (Tsaousis et al., 2014). The role(s) of the DUF59 domains in these proteins is unknown.

Lipoic acid synthase (LipA) is a s-adenosylmethionine-radical enzyme that generates lipoic acid using s-adenosylmethionine and octanoic acid as substrates (Busby et al., 1999, Ollagnier-de Choudens & Fontecave, 1999). LipA binds two [Fe₄-S₄] clusters (Harmer et al., 2014). One of the LipA Fe-S clusters act as a substrate for lipoic acid synthesis by providing the two sulfur atoms (Cicchillo & Booker, 2005, Lanz et al., 2015, McLaughlin et al., 2016). S. aureus has four known enzymes that require lipoamide to function: pyruvate dehydrogenase (Pdh), 2-oxoglutarate dehydrogenase (Suc), branched chain keto-acid dehydrogenase (Bck), and the glycine cleavage system (Gcv).

This study was initiated to enhance our understanding of the roles of SufT, and by extension the DUF59 domain, in Fe-S cluster assembly and cellular physiology. We found that SufT was essential for growth under conditions that impose a high demand for lipoamide-utilizing enzymes, but dispensable under low demand for these enzymes. A ΔsufT strain had decreased Pdh activity and displayed phenotypes consistent with decreased activities of LipA, Bck and Gcv. We propose that SufT is essential for the maturation of holo-LipA during conditions of high demand for lipoic acid. These findings have led to a model wherein the demand for the product(s) of the Fe-S cofactor dependent protein target(s) is a factor governing the usage of one Fe-S cluster assembly factor over another for the maturation of the apo-protein.

Results

SufT is essential for growth in cells experiencing a high demand for lipoamide-requiring enzymes

A S. aureus ΔsufT strain is capable of growth in liquid chemically defined medium containing the 20 canonical amino acids (AA) and glucose as a carbon source (hereafter 20AA medium) (Mashruwala et al., 2016). In contrast, the ΔsufT strain was unable to grow upon the same medium in solid form (Figure 1A). Supplementation of the solid 20 AA medium with lipoic acid fully corrected the growth of the ΔsufT strain (Figure 1A). Returning the sufT gene in trans to the ΔsufT strain also corrected growth in the absence of lipoic acid (Figure 1A). The initial experiments were conducted using strains cultured to stationary phase (18 hours; optical density ~9.5 (A₆₀₀)). The ΔsufT strain pre-cultured to post-exponential growth (6 hours; optical density ~5 (A₆₀₀)) also required lipoic acid for growth on solid 20AA medium verifying that the phenotypes of the ΔsufT strain did not arise as an artifact of stationary phase growth (data not shown).

Figure 1. SufT is essential for growth in cells experiencing a high demand for lipoamide-requiring enzymes.

Panel A) A S. aureus ΔsufT strain is incapable of growth upon solid medium lacking lipoic acid. Growth of the WT (JMB1100) carrying empty vector (pCM28) and the ΔsufT strain (JMB2512) carrying either empty vector or the complementing plasmid (pCM28_sufT) upon chemically defined solid media containing the 20 canonical amino acids (hereafter 20AA media) in the presence or absence of lipoic acid. Panel B) Growth of S. aureus upon solid 20AA medium is reliant upon the lipoic acid requiring enzymes pyruvate dehydrogenase (Pdh) and branched chain keto-acid dehydrogenase (Bck). Growth of the WT (JMB1100), suc::Tn (JMB2015), gcv::Tn (JMB7911), bck::Tn (JMB7888), pdh::Tn (JMB7891), and ΔsufT (JMB2512) strains upon solid 20AA medium. Panel C) Reliance of S. aureus upon Pdh and Bck for growth is decreased during culture in liquid 20AA medium. Growth profiles for WT (JMB1100), bck::Tn (JMB7888), pdh::Tn (JMB7891) and ΔsufT (JMB2512) strains are shown. Data in Panels A and B are representative of at least three independent experiments. Data in Panel C represent the average of biological duplicates and error bars represent standard deviations.

One explanation underlying the ability of the ΔsufT strain to grow in liquid defined medium in the absence of lipoic acid supplementation was that there was a decreased reliance upon lipoamide-requiring enzymes. The growth of the WT, ΔsufT, bck::Tn, pdh::Tn, suc::Tn and gcv::Tn strains was assessed upon solid and liquid 20 AA media. The suc::Tn and gcv::Tn mutants grew similar to the WT on solid and liquid defined media (Figure 1B and data not shown). There was no visible growth noted for the ΔsufT, bck::Tn and pdh::Tn mutants upon solid medium (Figure 1B), while all three strains were capable of growth in liquid defined medium (Figure 1C). However, the pdh::Tn and bck::Tn strains displayed growth defects during the latter phases of growth in liquid medium (Figure 1C). These findings verified that growth of S. aureus cells upon solid defined medium has an increased reliance upon Pdh and Bck.

A strain lacking SufT displays phenotypes consistent with decreased activities of multiple lipoamide-dependent enzymes

WT S. aureus permitted growth of a lipA::Tn mutant when the two strains were cultured in close proximity on solid medium, suggesting that S. aureus releases lipoic acid or an alternate metabolite into the extracellular milieu that bypasses the need for LipA (data not shown). To quantify this effect, the ΔsufT and WT strains were cultured in liquid 20AA medium (lacking lipoic acid), the spent medium supernatants were isolated, sterilized and used to support the growth of a lipA::Tn mutant. The spent medium from the ΔsufT strain supported growth to a lower final optical density than the spent medium from the WT strain (Figures 2A). The lipA::Tn strain was incapable of growth in the absence of exogenously supplied lipoic acid (Figures 2A and S1). Assuming that lipoic acid was the only metabolite present in the supernatants permitting growth of the lipA::Tn mutant, the concentration of lipoic acid in the WT and ΔsufT supernatants was ~80 and ~46 pg/mL, respectively.

Figure 2. The ΔsufT strain has phenotypes consistent with decreased activities of multiple lipoamide-dependent enzymes.

Panel A) Spent medium metabolites isolated from the ΔsufT strain are deficient in supporting growth of a lipA::Tn mutant in media that otherwise lacks lipoic acid. The lipA::Tn mutant (JMB7156) was cultured in liquid 20 AA media supplemented with no spent media or spent media from the WT (JMB1100) or ΔsufT (JMB2512) strains and the final optical densities are presented. Panel B) Supplementing solid 20AA medium with branch chain fatty acids (BCFAs) allows the bck::Tn and ΔsufT strains to grow in the absence of lipoic acid. Growth of the WT (JMB1100), bck::Tn (JMB7888), pdh::Tn (JMB7891) and ΔsufT (JMB2512) strains upon solid 20AA media with and without 5 μM each of methyl butyric acid, isobutyric acid, and isovaleric acid. Panel C) A strain lacking SufT is deficient for growth upon liquid 19AA medium lacking serine (hereafter 19AA-serine) and this deficiency is corrected by lipoic acid. Growth profiles for the WT (JMB1100), gcv::Tn (JMB7911) and ΔsufT (JMB2512) strains cultured in liquid 19AA-serine media in the presence or absence of lipoic acid (LA). Panel D) Activity of the lipoamide-dependent enzyme pyruvate dehydrogenase (Pdh) is decreased in a ΔsufT strain. Pdh activity was assessed in cell-free lysates generated from the WT (JMB1100) and ΔsufT (JMB2512) strains cultured in media imposing a low (20AA) or increased (19AA-serine) demand for lipoic acid. Panel E) Transcriptional activities of sufC and sufT are increased upon culture in serine dropout medium. Transcriptional activities of sufC and sufT were assessed in the WT (JMB1100) strain after culture in 20AA or 19AA-serine media. Data represent the average value of biological triplicates (Panels A, C and E) or biological duplicates (Panel D) and error bars represent standard deviations. Data in panel B are representative of at least three independent experiments.

Bck is required for the synthesis of branch-chain fatty acids (BCFAs) and defects in Bck activity result in decreased growth in media lacking BCFAs (Keeney et al., 2009). The ΔsufT, bck::Tn and pdh::Tn strains were incapable of growth upon solid 20AA medium. Supplementing the medium with BCFAs fully restored growth of the bck::Tn mutant and partially restored growth of the ΔsufT strain (Figure 2B). The growth of the pdh::Tn mutant was not altered by BCFAs.

Serine can be synthesized from glycine using the glycine cleavage system (Gcv) (Stauffer et al., 1989) and decreased Gcv function result in a growth deficiency in media lacking serine (Plamann et al., 1983). The ΔsufT strain was deficient for growth in a defined liquid glucose medium lacking serine (hereafter 19AA-serine or serine dropout medium) and growth was restored by lipoic acid supplementation (Figure 2C). The gcv::Tn mutant grew poorly without serine and this growth defect was not alleviated by lipoic acid supplementation.

The ΔsufT strain has decreased activities of the Fe-S cluster dependent enzymes AcnA, LeuCD and IlvD (Mashruwala et al., 2016). Growth in defined media lacking leucine and isoleucine and containing glutamate as a carbon source (hereafter 18AA glutamate medium) is reliant upon AcnA, LeuCD and IlvD (Mashruwala et al., 2016). As previously noted, the ΔsufT strain was highly deficient for growth in liquid 18AA glutamate medium (Figure S2) (Mashruwala et al., 2016). Supplementing the 18AA glutamate medium with lipoic acid did not rescue the growth defect of the ΔsufT strain. This suggested that the phenotypic correction imparted by lipoic acid is specific to growth conditions imposing a requirement for lipoamide-dependent enzymes, and does not provide a general correction of SufT dependent phenotypes.

Pyruvate dehydrogenase is a lipoamide-dependent enzyme. Growth in liquid 20AA medium and liquid serine dropout medium imposes a low and an increased demand for lipoic acid, respectively. Pdh activity in the ΔsufT strain was decreased by ~15% in cells cultured in 20AA media and ~40% when cells were cultured in serine dropout media (Figure 2D).

Catalase (Kat) activity was measured in the same lysates as those used to measure Pdh activity shown in Figure 2D. Kat activity was comparable in the WT and ΔsufT strains (Figure S4). These data allowed us to rule out a general defect in enzyme activity in the ΔsufT strain cultured in liquid serine dropout medium.

LipA requires Fe-S clusters for function (Lanz et al., 2015). Therefore, conditions that impose an increased demand for lipoic acid should impose an increased demand for Fe-S cluster synthesis and assembly. Consistent with this prediction, the transcriptional activities of sufT and sufC were increased during culture in liquid serine dropout medium relative to liquid 20AA medium (Figure 2E).

The data in Figure 2 suggested that the ΔsufT strain has decreased LipA, Pdh, Bck and Gcv activities in a manner that is dependent upon the cellular demand for lipoic acid. Further, a mechanism exists that increases the transcription of sufT and sufC upon increased demand for LipA and lipoic acid.

Methionine concentration influences the reliance upon lipoamide-requiring enzymes for growth

The initial phenotypic analyses presented, as well as the suppressor analyses discussed below, were conducted upon solid media that was prepared and stored at 4 °C prior to use. We noted that if the solid defined media was used on the same day as it was prepared (hereafter, freshly prepared media) the ΔsufT strain was capable of growth in the absence of lipoic acid supplementation (Figure 3A). We hypothesized that a component of the solid defined medium decreased in bioavailability over time.

Figure 3. Methionine concentration influences the reliance upon lipoamide-requiring enzymes for growth.

Panel A) The lipoic acid requirement of the ΔsufT strain is influenced by the presence of methionine (Met) in the growth medium. Growth of the WT (JMB1100) and ΔsufT (JMB2512) strains upon freshly prepared 19AA media lacking Met (hereafter 19AA-Met) in the presence or absence of lipoic acid supplementation or solid 20AA medium. Panel B) Extracellular metabolites from the ΔsufT strain are deficient in supporting growth of the lipA::Tn strain when cultured upon freshly prepared solid 20AA medium. The lipA::Tn mutant (JMB7156) was cross-struck on a freshly prepared solid 20AA medium plate lacking lipoic acid and the WT (JMB1100) or ΔsufT (JMB2512) strains were perpendicularly struck, bisecting the lipA::Tn streak. Panel C) Methionine concentration alters the reliance upon Pdh for growth. Growth of the WT (JMB1100), ΔsufT (JMB2512), bck::Tn (JMB7888) and pdh::Tn (JMB7891) strains on freshly prepared solid 20AA and 19AA-Met media. Data in panels A and C are representative of at least three independent experiments.

Individual amino acids were spread upon solid 20AA medium that had been previously prepared and stored at 4 °C. Supplementing the solid media with methionine (Met) permitted growth of the ΔsufT strain (data not shown).

Growth of the ΔsufT strain was assessed upon freshly prepared solid 20AA media or media lacking Met (hereafter 19AA-Met or Met dropout medium) and/or lipoic acid supplementation. The ΔsufT strain was incapable of growth upon Met dropout medium unless it was supplemented with lipoic acid (Figure 3A).

The effect of Met supplementation upon the ability of the ΔsufT strain to permit growth of the lipA::Tn strain was examined. The lipA::Tn mutant was cross-struck upon freshly prepared solid 20AA medium lacking lipoic acid. The ΔsufT and WT strains were individually and perpendicularly cross-struck, bisecting the lipA::Tn streak. The WT strain permitted growth of the lipA::Tn strain when the cells were proximal to the WT streak, but not when the cells were distal (Figure 3B). The growth of the lipA::Tn mutant was visibly decreased when cross-struck with the ΔsufT strain. These data suggested that although Met supplementation allowed growth of the ΔsufT strain it did not correct the deficiency this strain has in producing the extracellular metabolite (presumably lipoic acid) that permits growth of the lipA::Tn mutant.

The effect of Met supplementation upon the requirement for Pdh or Bck for growth was assessed. The pdh::Tn mutant was capable of growth upon freshly prepared solid 20AA medium, but not upon Met dropout medium (Figure 3C). Methionine supplementation had no effect on the growth of the bck::Tn mutant. These findings suggested that solid media lacking Met had an increased requirement for Pdh, and therefore, an increased demand for lipoic acid.

The minimum concentration of Met necessary to limit growth of the ΔsufT strain was determined. Supplementing freshly prepared solid growth media with concentrations greater than 1.65 pM Met supported growth of the ΔsufT strain (data not shown). Unless specifically stated, for the remainder of this manuscript solid media were supplemented with 1.65 pM Met (defined as 20AA limiting Met (LM) medium) and were used on the same day as they were prepared.

The lipoic acid requirement of a sufT mutant is suppressed by mutations in either clpC or upstream of fur

Suppressor analyses were conducted to further dissect the role of SufT in lipoic acid dependent growth. Approximately ~1e+9 cells of the ΔsufT (JMB2512) strain were plated on solid tryptic soy broth (TSB) medium containing rifampicin or solid chemically defined medium lacking lipoic acid. Colonies arose at a frequency of ~1e–8 upon TSB rifampicin medium and ~1e–6 upon the defined medium. One colony from each defined medium plate was retained for further analysis. Further experimentation verified that seven of the ten strains did not require lipoic acid for growth.

Whole genome sequencing was conducted to identify putative single nucleotide polymorphisms (SNP) that result in suppression. SNPs were verified in four of the seven genomes sequenced and were located in one of two locations. One strain had a mutation upstream of the fur gene at −90 base pairs from the translational start site (USA300_FPR3757 genome) (SAUSA300_1448*; fur*) (Table 2). Three strains had individual SNPs within the clpC gene (SAUSA300_0510*; hereafter a general descriptor for these alleles is clpC*) (Table 2). Preliminary analyses found that the suppression by the fur* and clpC* mutations are likely to be imparted through different mechanisms, and therefore, in the current report we elected to focus on the suppression imparted by the clpC* mutations.

Table 2.

Properties of sufT suppressor mutations.

Strain	Location	Mutation type	Mutation	Effect
JMB4974	clpC	Missense	G672→R	dominant
JMB4980	clpC	Amber	E362→stop	recessive
JMB4982	fur promoter	Base-pair change	−90 promoter thy→cyt	Not tested
JMB4986	clpC	Frame shift	Y611SVIL→YFCDFNstop	recessive

Altered ClpC functionality suppresses the lipoic acid requirement of the ΔsufT strain

Two of the clpC* mutations were nonsense mutations whereas the third was a missense mutation (Table 2) leading to the hypothesis that these alleles result in altered ClpC function. Puromycin causes premature chain termination during translation resulting in an increased demand for the proteolytic machinery (Nathans, 1964). S. aureus strains lacking ClpC are sensitive to puromycin (Frees & Ingmer, 1999). The ΔsufT clpC* strains and a ΔsufT ΔclpC strain displayed increased sensitivity to puromycin (Figure 4A). All strains displayed similar growth upon solid TSB medium lacking puromycin (data not shown).

Figure 4. Altered ClpC function suppresses the lipoic acid requirement of a ΔsufT mutant.

Panel A) Strains with clpC* alleles are sensitive to puromycin toxicity and phenocopy a ΔclpC strain. Growth of the WT (JMB1100), ΔsufT (JMB2512), ΔsufT ΔclpC (JMB7121), ΔsufT clpC_G672R (JMB4974), ΔsufT clpC_E362→stop (JMB4980), and ΔsufT clpC_{Y611SVIL→YFCDFNstop} (JMB4986) strains upon solid TSB medium with 7 μg mL⁻¹ puromycin is shown. Panel B) Introduction of a null clpC allele suppresses the lipoic acid requirement of the ΔsufT strain. Growth of the WT (JMB1100), ΔsufT (JMB2512), ΔclpC (JMB5674), ΔsufT ΔclpC (JMB7121), and ΔsufT clpC_E362→stop (JMB4980) strains upon solid 20AA LM media with and without lipoic acid supplementation is shown. Data in both panels are representative of at least three independent experiments.

The ability of a clpC null allele to suppress the lipoic acid requirement of the ΔsufT mutant was assessed. The ΔsufT ΔclpC strain was capable of growth upon solid 20AA LM medium whereas the ΔsufT mutant was not (Figure 4B). These data verified that decreased ClpC activity suppressed the lipoic acid dependent growth defect of the ΔsufT mutant.

The clpC_G672R allele is dominant negative

We sought to determine whether the effects of the clpC* alleles were recessive or dominant. A wild type clpC allele was returned to the ΔsufT clpC* strains, as well as the ΔsufT and WT strains via a plasmid (pclpC) and growth upon solid 20AA LM medium was monitored. The presence of pclpC abolished suppression in the ΔsufT clpC_E362→stop strain whereas it had an intermediate effect upon the ΔsufT clpC_{Y611SVIL→YFCDFNstop} strain (Figure 5A). Notably, the ΔsufT clpC_G672R strain containing pclpC remained incapable of growth in the absence of lipoic acid.

Figure 5. The clpC_G672R allele is dominant negative.

Panel A) Genetic complementation reveals the presence of both recessive, as well as non-recessive clpC* alleles. Growth of the WT (JMB1100), ΔsufT (JMB2512), ΔsufT clpC_G672R (JMB4974), ΔsufT clpC_E362→stop (JMB4980) and ΔsufT clpC_{Y611SVIL→YFCDFNstop} (JMB4986) strains containing either empty vector (pEPSA5) or pclpC upon solid 20AA LM medium lacking lipoic acid is shown. Panel B) The clpC_G672R allele exerts a dominant effect upon the lipoic acid requirement of a ΔsufT mutant strain. Growth of the WT (JMB1100) and ΔsufT (JMB2512) strains containing either empty vector (pEPSA5), pclpC or pclpC_G672R upon solid 20AA LM medium containing 1% xylose to induce transcription. Panel C) The clpC_G672R allele exerts a dominant effect upon puromycin resistance in the WT strain. Growth of the WT (JMB1100) containing empty vector (pEPSA5), pclpC or pclpC_G672R, as well as the ΔclpC mutant (JMB7384) containing empty vector (pEPSA5), upon TSB media with and without 6 μg mL⁻¹ puromycin is shown. Data in all panels are representative of at least three independent experiments.

The clpC_G672R allele was introduced into the ΔsufT strain via plasmid (pclpC_G672R) and growth upon solid 20AA LM medium was monitored. The presence of pclpC_G672R, but not pclpC allowed growth of the ΔsufT strain (Figure 5B). We next assessed the effect of the clpC_G672R allele upon protein processing. The growth of the WT strain carrying empty vector, pclpC or pclpC_G672R upon solid TSB media with or without puromycin was monitored. The WT strain carrying pclpC_G672R displayed increased sensitivity towards puromycin relative to the WT strain carrying empty vector or pclpC (Figure 5C). As expected, the clpC strain carrying empty vector was sensitive to puromycin.

Strains lacking ClpP or TrfA mimic the suppression imparted by strains with decreased ClpC function

The Gly at position 672 of ClpC is located in an exposed disordered loop annotated as the IGF-loop. The IGF-loop facilitates interaction between ClpC and ClpP, which is the core peptidase of the ClpCP proteolytic complex (Singh et al., 2001, Kim et al., 2001). The IGF-loop is conserved in Clp AAA+ ATPases (ClpC and ClpX) that interact with ClpP. ClpX variants containing residue changes in the IGF-loop are active chaperones, but are defective in proteolysis because of altered interactions with ClpP (Singh et al., 2001, Kim et al., 2001).

We examined whether the absence of ClpP would suppress the lipoic acid requirement of the ΔsufT strain. The ΔsufT clpP::Tn strain was capable of growth on solid 20AA LM medium in the absence of lipoic acid (Figure 6). The strain displayed a slow growth phenotype, but this growth defect was similar to the growth defect of the isogenic clpP::Tn mutant and occurred irrespective of lipoic acid supplementation (data not shown). S. aureus contains an alternate peptidase (ClpQ) that interacts with AAA+ ATPases to process proteins (Frees et al., 2005). Introduction of a clpQ::Tn mutation did not suppress the lipoic acid dependent growth of the ΔsufT strain (Figure 6).

Figure 6. The ΔsufT strains lacking ClpC, ClpP or TrfA do not require lipoic acid supplementation for growth.

Growth of the WT (JMB1100), ΔsufT (JMB2512), ΔsufT clpQ::Tn (JMB7729), ΔsufT clpP::Tn (JMB7709), ΔsufT ΔclpC (JMB7121) and ΔsufT trfA::Tn (JMB7703) strains upon solid 20AA LM media with and without lipoic acid supplementation. Data are representative of at least three independent experiments.

Small adaptor proteins facilitate regulated proteolysis by aiding recognition of target peptides. The TrfA adaptor is proposed to interact with ClpCP (Donegan et al., 2014, Jousselin et al., 2013). Introduction of a trfA::Tn mutation allowed for growth of the ΔsufT strain in the absence of lipoic acid supplementation (Figure 6). The data in Figure 6 allowed for the conclusion that decreased protein processing by ClpCP, as well as decreased protein targeting to ClpCP by TrfA, bypassed the lipoic acid requirement of the ΔsufT strain.

Decreased ClpC function lowers the requirement for lipoic acid

ClpC has a regulatory role in directing metabolic flux in S. aureus (Chatterjee et al., 2011, Chatterjee et al., 2005, Chatterjee et al., 2009). We examined the premise that the clpC* alleles were suppressing the lipoic acid dependent growth defect of the ΔsufT strain by redirecting metabolic flux, and thereby, decreasing cellular demand for lipoic acid.

Individual top agar overlays containing the ΔsufT lipA::Tn and ΔsufT clpC_E362→stop lipA::Tn strains were poured upon solid 20AA medium and 1.5 pmol lipoic acid was spotted in the center of each plate. The amount of growth was determined after 48 hours of incubation. Growth was enhanced in the strain carrying the clpC_E362→stop allele (Figures 7A and 7B), suggesting that this allele allowed for growth with less lipoic acid.

Figure 7. Strains with decreased ClpC function have a decreased requirement for lipoic acid.

Panels A and B) Top-agars containing the ΔsufT lipA::Tn (JMB7750), ΔsufT clpC _E362→stop lipA::Tn (JMB7746), lipA::Tn (JMB7156), or ΔclpC lipA::Tn (JMB7385) were overlaid upon 20AA LM medium plates before 1.5 pmol of lipoic acid was spotted in the center. Representative photographs of zones of growth are displayed in Panel A. Quantification of the zones of growth are displayed in Panel B and the data represent the average of biological triplicates and error bars represent standard deviations.

The ability of a null clpC allele to lower the requirement for lipoic acid was examined. The absence of ClpC also lowered the requirement for lipoic acid (Figure 7B).

Decreased TCA cycle function or fermentative growth decreases the demand for lipoic acid and for SufT

Strains lacking ClpC have reduced AcnA activity and decreased carbon flux through the TCA cycle (Chatterjee et al., 2005, Chatterjee et al., 2009). The expression of genes encoding for Pdh, Bck and Suc are decreased upon fermentative culture (Fuchs et al., 2007). We examined whether modulating carbon flux through the TCA cycle or fermentative growth decreased the requirement for lipoic acid and bypassed the need for SufT.

Aconitase (AcnA) is an enzymatic gatekeeper for flux through the TCA cycle in S. aureus (Somerville et al., 2002). Top agar overlays containing the lipA::Tn or ΔacnA lipA::Tn strains were poured upon solid 20AA medium and 1.5 pmol of lipoic acid was spotted. The growth of the ΔacnA lipA::Tn strain was significantly greater than that of the lipA::Tn mutant (Figure 8A).

Figure 8. Decreased TCA cycle function results in a lower demand for lipoic acid and renders SufT non-essential.

Panel A) Strains lacking aconitase (AcnA) have a decreased requirement for lipoic acid. Top agar overlays containing the lipA::Tn (JMB7156) or lipA::Tn ΔacnA (JMB7825) strains were overlaid upon 20AA LM medium and 1.5 pmol of lipoic acid was spotted in the center. Zones of growth were measured after incubation. Panel B) A strain with decreased ClpC function has decreased AcnA activity. The WT (JMB1100), ΔsufT (JMB2512) and ΔsufT clpC _E362→stop (JMB4980) strains were cultured in liquid 20AA medium and AcnA activity was determined in cell-free lysates. Panel C) A ΔsufT mutant with decreased AcnA expression is capable of growth upon solid 20AA LM medium and this phenotype is suppressed upon induction of acnA transcription. Growth of the acnA::Tn (JMB2475) and acnA::Tn ΔsufT (JMB3539) strains carrying pacnA, which contains acnA under the transcriptional control of a xylose inducible promoter, were cultured overnight in the presence or absence of varying concentrations of xylose. The strains were subsequently serial diluted and spot plated upon solid 20AA LM media with and without 1% xylose. Data in Panel A and B represent the average values of biological triplicates and error bars represent standard deviations. Data in Panel C are representative of at least three independent experiments.

AcnA activity was monitored in the WT, ΔsufT, and ΔsufT clpC_E362→stop strains after culture in liquid 20AA LM medium. AcnA activity was decreased in the ΔsufT clpC_E362→stop strain relative to the ΔsufT strain (Figure 8B).

The effect of modulating carbon flux through the TCA cycle was analyzed. We utilized acnA::Tn and ΔsufT acnA::Tn strains that carried acnA upon a plasmid under the transcriptional control of a xylose inducible promoter (pacnA) allowing for modulation of acnA transcription. The acnA::Tn and ΔsufT acnA::Tn strains containing pacnA were pre-cultured in liquid TSB media containing 0, 0.5 or 1% xylose to induce transcription of acnA and subsequently spot plated upon solid 20AA LM solid media supplemented with and without 1% xylose. In the absence of induction, the ΔsufT acnA::Tn strain containing pacnA grew on 20AA LM medium (Figure 8C). Induction of acnA transcription resulted in decreased growth on 20AA medium.

The WT and lipA::Tn strains were cultured aerobically, microaerobically, and fermentatively (inside an anaerobic chamber) in liquid 20AA medium lacking lipoic acid and final culture optical densities were measured. The requirement for LipA was maximal during respiratory growth and minimal during fermentative growth (Figure 9A).

Figure 9. Fermentative growth imposes a decreased demand for lipoic acid and renders SufT non-essential.

Panel A) Requirement of lipoic acid for growth is decreased upon fermentative culture. Growth of the WT (JMB1100) and lipA::Tn (JMB7156) strains grown in liquid 20AA medium lacking lipoic acid under aerobic, microaerobic or fermentative (growth inside an anaerobic chamber) conditions. Panel B) SufT is dispensable for growth in serine dropout medium during fermentative culture. Growth of the WT (JMB1100) and ΔsufT (JMB2512) strains after growth in liquid 19AA-serine medium in the presence or absence (fermentative growth) of dioxygen. Panel C) SufT is dispensable for growth upon solid media lacking lipoic acid during fermentative culture. Growth of the WT (JMB1100) and ΔsufT (JMB2512) strains upon solid medium lacking lipoic acid and in the presence or absence of dioxygen is shown. Data in Panel A and B represent the average values of biological triplicates and error bars represent standard deviations. Data in Panel C are representative of at least three independent experiments. For data in Panels A and B final optical densities were assessed and to facilitate comparative analyses the data were normalized with respect to the WT strain under each growth condition.

The activity of Pdh was examined in WT cells cultured aerobically and fermentatively in liquid 20AA medium. Pdh activity was lower during fermentative growth (Figure S3). The change in Pdh activity upon fermentative growth was similar to that reported for Escherichia coli and Enterococcus faecalis (Kaiser & Sawers, 1994, Snoep et al., 1990).

The WT and ΔsufT strains were cultured in liquid serine dropout medium aerobically or fermentatively. The ΔsufT strain was deficient for growth in the presence of dioxygen, but not during fermentative growth (Figure 9B). Likewise, SufT was essential for growth on solid 20AA LM medium lacking lipoic acid in the presence of dioxygen, but dispensable during fermentative growth (Figure 9C).

Nfu can compensate for the absence of SufT for lipoic acid-dependent growth

The epistatic relationships between sufT, nfu and sufA were investigated by phenotypically examining mutant strains lacking one, two or all three maturation factors. Growth was examined in liquid 20AA and serine dropout media, which impose varying demands for lipoic acid. The ΔsufA strain did not display a growth deficiency in either medium (Figures 10A and 10B). The ΔsufA ΔsufT double mutant phenocopied the ΔsufT strain in 20AA medium, but displayed a phenotype that was more severe than the phenotypes of the parental strains in serine dropout medium. The Δnfu ΔsufA double mutant phenocopied the Δnfu strain in both media. The phenotypes of the Δnfu and ΔsufT mutations displayed synergism and the Δnfu ΔsufT double mutant displayed a severe growth defect in both media (Figures 10A and 10B). The Δnfu ΔsufT ΔsufA triple mutant strain phenocopied the Δnfu ΔsufT strain (Figures 10A and 10B).

Figure 10. nfu is synergistic with sufT, while sufA is epistatic to sufT.

Panels A and B) sufT displays synergy with nfu and sufA is epistatic to sufT. Growth was assessed for the WT (JMB1100), ΔsufT (JMB1146), Δnfu (JMB1165), ΔsufA (JMB2223), ΔsufA ΔsufT (JMB2224), Δnfu ΔsufA (JMB6834), Δnfu ΔsufT (JMB2514), and the Δnfu ΔsufT ΔsufA (JMB6835) strains in 20AA (Panel A) or 19AA-serine (Panel B) media. Panel C) The phenotypes of the nfu and sufT mutations display synergism for lipoic acid dependent growth. The WT (JMB1100), ΔsufT (JMB2512), Δnfu (JMB1165) and ΔsufT Δnfu (JMB2514) strains were plated on solid 20AA media with and without lipoic acid. Panel D) nfu and sufT display synergy with respect to succinate dehydrogenase (Sdh) activity. Sdh activity was assessed in cell-free lysates generated from the same strains used in Panel C. Data represent the average value of biological triplicates (Panel D) or biological duplicates (Panel A and B) and error bars represent standard deviations. Data in panel C are representative of at least three independent experiments.

We further examined the phenotypes of the Δnfu ΔsufT strain relative to the single mutants. Supplementing liquid 20AA or liquid serine dropout media with lipoic acid largely rescued the growth of the Δnfu ΔsufT strain (Figures S5A and S5B). The Δnfu ΔsufT strain has widespread metabolic defects including multiple growth deficiencies apart from those shown in the current study (Mashruwala et al., 2016). Growth of the Δnfu ΔsufT strain in liquid 18AA glutamate medium was not rescued by lipoic acid supplementation (Figure S5C), confirming that the Δnfu ΔsufT strain displays both lipoic acid dependent, as well as lipoic acid independent defects. Growth upon solid 20AA medium was also examined. The WT, ΔsufT and Δnfu strains were capable of growth. However, the Δnfu ΔsufT double mutant strain was incapable of growth. Supplementation of medium with lipoic acid rescued the growth of the Δnfu ΔsufT strain (Figure 10C).

The ΔsufT and Δnfu mutations display synergy with regard to AcnA activity (Mashruwala et al., 2016). Dehydratase enzymes contain solvent-accessible Fe-S clusters that can be inactivated by stress such as reactive oxygen species insult resulting in Fe-S cluster turnover (Varghese et al., 2003, Jang & Imlay, 2010, Mashruwala et al., 2016). One of the Fe-S clusters upon LipA is a substrate for lipoic acid synthesis (Lanz et al., 2015, Cicchillo & Booker, 2005). By inference, the maintenance of Fe-S clusters in dehydratases and LipA necessitate an increased demand for Fe-S cluster assembly. Succinate dehydrogenase also requires Fe-S clusters for function. However, the Sdh Fe-S clusters are protected by the polypeptide providing increased stability (Jang & Imlay, 2010, Yankovskaya et al., 2003). Sdh activity was decreased by ~40% in the Δnfu strain, increased by ~200% in the ΔsufT strain and decreased by ~80% in the Δnfu ΔsufT strain, relative to the WT (Figure 10D).

Strains lacking genes encoding for proteins with functional overlap display synergistic (super-additive) phenotypic effects (Perez-Perez et al., 2009). The synergistic and epistatic relationships amongst the Fe-S assembly factors led us to examine the effect of over-expression of nfu (pnfu) or sufA (psufA) on the lipoic acid requirement of the ΔsufT strain. Growth was first examined upon solid 20AA LM medium, which imposes a high demand for lipoic acid. The ΔsufT strain carrying pnfu was capable of growth whereas the ΔsufT strain with empty vector was not. However, the presence of pnfu did not rescue growth to the levels observed in the WT strain (Figure 11). The ΔsufT strain carrying psufA did not grow without lipoic acid supplementation. Growth was also analyzed upon a medium that imposed a decreased demand for lipoic acid (containing 6.6 pM Met; 20AA intermediate Met (IM) medium). Solid 20AA IM medium allowed for intermediate growth of the ΔsufT strain carrying the empty vector (Figure 11). The ΔsufT strain carrying pnfu grew similar to the WT strain containing empty vector. The presence of pnfu did not have a notable effect on the growth of the WT strain.

Figure 11. Nfu can compensate for the absence of SufT for lipoic acid dependent growth.

Overexpression of nfu, but not sufA, permits growth of the ΔsufT strain on solid media lacking lipoic acid. The ΔsufT (JMB2512) and WT (JMB1100) strains carrying pEPSA5 (empty vector), pnfu or psufA were pre-cultured for 6 hours with 1% xylose, serial diluted and spot plated upon solid 20AA LM medium, 20AA LM medium with lipoic acid or 20AA intermediate methionine (IM) medium containing 1% xylose. Data are representative of at least three independent experiments.

Suppression imparted by the clpC_E362→stop allele requires Nfu

Nfu could compensate for SufT during decreased demand for lipoiac acid. We tested the hypothesis that a lower demand for lipoic acid in strains with decreased ClpC function is accompanied by an increased abundance of Nfu and these two mechanisms simultaneously aid in phenotypic suppression of the ΔsufT strain. We created a Nfu-Gfp translational fusion that was present upon an episome and complementation studies found that it functioned in vivo (Figure S6). GFP is highly resistant to proteolysis and it is stable in S. aureus and is not a substrate for ClpCP unless specifically manipulated to contain a SsrA degradation signal (Bokman & Ward, 1981, Donegan et al., 2014). The fluorescence attributed to the Nfu-Gfp translational fusion was increased by ~10-fold in the ΔclpC strain relative to the parental strain (Figure 12A).

Figure 12. Nfu-Gfp abundance is increased in a ΔclpC mutant and Nfu is required for the suppression imparted by the clpC_E362→stop allele on the lipoic acid requirement of the ΔsufT strain.

Panel A) The abundance of Nfu-Gfp is increased in a clpC mutant strain. Abundance of Nfu-Gfp was assessed by monitoring fluorescence in the Δnfu (parent; JMB3403) and ΔclpC Δnfu mutant (ΔclpC; JMB8246) strains containing the pLL39_nfu_gfp episome. Panel B) Nfu is required for the suppression imparted by the clpC_E362→stop allele on the lipoic acid requirement of the ΔsufT strain. Growth of the WT (JMB1100), ΔsufT (JMB2512), Δnfu (JMB1165), ΔsufT clpC_E362→stop (JMB4980), ΔsufT clpC_E362→stop nfu::Tn (JMB8121) and Δnfu ΔsufT (JMB2514) strains cultured on solid chemically defined 20AA LM media with and without lipoic acid supplementation is shown. Data in Panel A represent the average value of biological triplicates and error bars represent standard deviations. Data in panel B are representative of at least three independent experiments.

We next examined whether the clpC_E362→stop allele could suppress the lipoic acid requirement of the ΔsufT strain in the absence of Nfu. Growth of the ΔsufT clpC_E362→stop nfu::Tn triple mutant strain and its parental strains was examined upon solid 20AA LM media with and without lipoic acid supplementation. The ΔsufT clpC_E362→stop strain was capable of growth whereas the ΔsufT clpC_E362→stop nfu::Tn strain was not (Figure 12B). Both strains grew in the presence of lipoic acid.

Discussion

The S. aureus SufT is composed entirely of the DUF59 domain. This study was initiated to further investigate the role of SufT, and by extension the DUF59 domain, in Fe-S cluster assembly and cellular physiology. SufT was essential for growth upon solid chemically defined medium (with limiting Met) in the absence of exogenous lipoic acid supplementation. Further experimentation found that S. aureus growth upon solid chemically defined media was highly reliant upon the lipoamide-requiring enzymes pyruvate dehydrogenase (Pdh) and branched chain keto-acid dehydrogenase (Bck). The reliance upon Pdh and Bck for growth was decreased upon culture in liquid defined media. The reliance on SufT for lipoic acid dependent growth was also decreased in liquid media and it was rendered non-essential.

Lipoic acid synthesis requires LipA, which utilizes two Fe-S clusters for catalysis (Ollagnier-de Choudens & Fontecave, 1999, Harmer et al., 2014). The ΔsufT strain displayed multiple phenotypes that could be attributed to decreased activities of enzymes requiring lipoate. A ΔsufT strain also had decreased activity of the lipoamide-dependent enzyme Pdh. In addition, supplementation of solid growth medium with branched-chain fatty acids, which are synthesized via the Bck enzyme, partially rescued growth of the ΔsufT strain in the absence of lipoic acid supplementation. The finding that SufT was essential when cellular reliance upon Pdh and Bck was increased led to a working model wherein SufT is required for the maturation of LipA under growth conditions that impose a high demand for lipoic acid.

The bioavailability of methionine in the solid defined media used in this study decreased upon storage. This fact led to the finding that the lipoic acid dependent growth defect of the ΔsufT strain was suppressed by excess methionine (Met). It is currently unclear why this occurs. Reports from numerous studies show that solvent accessible Fe-S clusters can be protected from damage by interactions with substrates (Varghese et al., 2003, Gardner & Fridovich, 1991, Dougherty & Downs, 2006, Downs, 2006, Macomber & Imlay, 2009). Met is a precursor for s-adenosylmethionine (SAM) synthesis. It is possible that increasing intracellular Met results in an increased concentration of SAM, which provides protection to the Fe-S clusters of LipA, and thereby, decreases the need for SufT for the assembly of holo-LipA. Growth of S. aureus upon solid defined media containing excess Met had a decreased requirement for Pdh. Therefore, a more likely explanation is that Met supplementation decreased the demand for lipoic acid, and thereby, rendered SufT non-essential.

Strains deficient in ClpC activity were capable of suppressing the lipoic acid dependent growth defect of the ΔsufT mutant. ClpC acts either independently as a molecular chaperone or in conjunction with the ClpP peptidase and the TrfA adaptor (Frees et al., 2003, Jousselin et al., 2013). The ΔsufT mutant strains lacking either ClpP or TrfA were capable of growth in the absence of lipoic acid. These findings led to a model wherein altered protein targeting by TrfA and decreased protein processing by ClpCP resulted in the accumulation of proteins that facilitated suppression of the lipoic acid requirement of the ΔsufT strain.

The manifest question arising from the suppressor analyses was how do the mutant clpC* alleles alter the cellular requirement for lipoic acid. Strains lacking ClpC decreased the cellular demand for lipoic acid. ClpC positively influences the TCA cycle and decreased ClpC function results in reduced activity of the fulcrum TCA cycle enzyme aconitase (AcnA) (Chatterjee et al., 2005, Chatterjee et al., 2009). An ΔacnA strain displayed a decreased demand for lipoic acid. A ΔsufT ΔacnA strain was capable of growth in the absence of lipoic acid while increased expression of acnA in this strain decreased growth. Fermentative growth also decreases flux through the TCA cycle (Ledala et al., 2014) and the transcription of pdh and bck are decreased during fermentative growth (Fuchs et al., 2007). Fermentative growth resulted in a decreased need for lipoic acid synthesis and it also rescued the lipoic acid dependent phenotypic abnormalities of the ΔsufT strain.

Conditions that imposed a low demand for lipoic acid rendered SufT non-essential for growth. We inferred that an alternate cellular factor involved in Fe-S cluster assembly was compensating for the absence of SufT during these growth conditions. The phenotypes associated with the nfu and sufT mutations were synergistic with respect to their lipoic acid dependent growth phenotypes. When cultured on media where SufT and Nfu were individually dispensable for growth (20AA medium) a Δnfu ΔsufT double mutant strain was incapable of growth. Overexpression of nfu in the ΔsufT strain also permitted growth in the absence of lipoic acid supplementation. Moreover, Nfu abundance was increased in a ΔclpC strain and Nfu was required for the suppression imparted to the ΔsufT strain by the clpC _E362→stop allele. These data lead to a model wherein decreased ClpC function lowers the demand for lipoic acid and increases the abundance of Nfu, which act simultaneously to bypass the lipoic acid requirement of the ΔsufT strain.

The phenotypes of nfu and sufT are synergistic and genes encoding for proteins with functional overlap often display synergistic phenotypic effects when the gene products are absent or non-functional (Perez-Perez et al., 2009, Mashruwala et al., 2016). Further, sufA was negatively epistatic to sufT for lipoic acid dependent growth in liquid serine dropout media. Since Nfu and SufA function as Fe-S cluster carriers, it is tempting to speculate that SufT also functions in this capacity. However, further biochemical evidence is necessary to draw this conclusion.

S. aureus encodes for multiple factors involved in the maturation of Fe-S proteins including Nfu, SufA and SufT (Rosario-Cruz et al., 2015, Mashruwala et al., 2015). Thus far, it has been unclear what factors favor the use of one maturation factor over another. SufT is essential during conditions imposing a high demand for lipoic acid while Nfu appears to compensate for its absence during low demand for lipoic acid. By inference, the selective usage of SufT over alternate assembly factors for the maturation of LipA is a direct outcome of lipoic acid demand imposed by growth conditions. We propose a model wherein the utilization of an assembly factor for the maturation of a Fe-S protein is dictated by the cellular demand for the metabolite synthesized by the target apo-protein (increased need for lipoic acid, in the case of SufT).

It is currently unclear why SufT is required during growth conditions imposing a high demand for lipoic acid. However, the rate of apo-protein maturation, and thence, the rate of product formation is likely to be a critical factor in cells experiencing a high demand for the enzyme product. Moreover, it has been hypothesized that one of the two Fe-S clusters of LipA provides the two sulfur atoms for the synthesis for lipoic acid (Harmer et al., 2014, Lanz et al., 2015, McLaughlin et al., 2016). Under such a scenario, the Fe-S cluster serving as the substrate must be either repaired or synthesized de novo prior to each round of catalysis. Both scenarios place an additional demand on the Fe-S protein maturation machinery. It is worth noting that results from our recent study suggest that SufT is not necessary for the repair of the Fe-S cluster in AcnA (Mashruwala et al., 2016).

In summation, SufT is essential for S. aureus under growth conditions that impose a high demand for lipoic acid. SufT functions in Fe-S cluster assembly in S. aureus (Mashruwala et al., 2016). The data presented are consistent with a model wherein SufT, and by extension the DUF59 domain, functions in the maturation of holo-LipA.

Experimental Procedures

Materials

Restriction enzymes, quick DNA ligase kit, deoxynucleoside triphosphates and Phusion DNA polymerase were purchased from New England Biolabs. The plasmid mini-prep and gel extraction kits were purchased from Qiagen. DNase I was purchased from Ambion. Lysostaphin was purchased from Ambi products. Oligonucleotides were purchased from Integrated DNA Technologies and sequences are listed in Table S1. Tryptic Soy broth (TSB) was purchased from MP biomedical. Difco BiTek agar was added (15 g L⁻¹) for solid medium. Unless specified, all chemicals were purchased from Sigma-Aldrich and were of the highest purity available.

Bacterial strains, media, and growth conditions

Unless specified, the S. aureus strains used in this study (Table 1) were isogenic and constructed in the community-associated S. aureus strain USA300_LAC that was cured of the native plasmid pUSA03 that confers erythromycin resistance (Boles et al., 2010). S. aureus were cultured at 37°C with shaking at 200 rpm aerobically in 10 or 30 mL culture tubes containing either 1 mL or 5 mL of TSB media, respectively. Antibiotics where added to the following concentrations: 150 μg mL⁻¹ ampicillin; 30 μg mL⁻¹ chloramphenicol (Cm); 10 μg mL⁻¹ erythromycin (Erm); 3 μg mL⁻¹ tetracycline (Tet); rifampicin 1.25 μg mL⁻¹; streptonegrin 15 ng mL⁻¹; puromycin 5–7 μg mL⁻¹. To maintain plasmids, the media was supplemented with 15 μg mL⁻¹ or 5 μg mL⁻¹ of chloramphenicol or erythromycin, respectively.

Table 1.

Strains and plasmids used in this study.

Strains used in this study
Strain name	genotype	Source/reference
JMB1100	USA300_LAC	A.R. Horswill (Boles et al., 2010)
RN4220	Restriction minus	NARSA (Kreiswirth et al., 1983)
JMB1146	ΔsufT (SAUSA300_0875)	This study
JMB2512	ΔsufT::tetM	This study
JMB1432	Δfur::tetM	(Torres et al., 2010)
JMB2475	acnA::TN (ermR) (SAUSA300_1246)	NARSA (Fey et al., 2013)
JMB3539	acnA::TN (ermR) ΔsufT	This study
JMB5674	clpC::TN (ermR) (SAUSA300_0510)	NARSA (Fey et al., 2013)
JMB7121	ΔsufT::tetM clpC::TN (ermR)	This work
JMB4982	ΔsufT::tetM fur*	This work
JMB4974	ΔsufT::tetM clpC*G672→R	This work
JMB4980	ΔsufT::tetM clpC*E362→stop	This work
JMB4986	ΔsufT::tetM clpC*Y611SVIL→YFCDFNstop	This work
JMB7384	ΔclpC::tetM	This work
JMB7156	lipA::TN (ermR) (SAUSA300_0829)	NARSA (Fey et al., 2013)
JMB7750	lipA::TN (ermR) ΔsufT::tetM	This work
JMB7746	lipA::TN (ermR) ΔsufT::tetM clpC*E362→stop	This work
JMB7825	lipA::TN (ermR) ΔacnA::tet	this study
JMB7385	ΔclpC::tetM lipA::TN (ermR)	This work
JMB1165	Δnfu (SAUSA300_0839)	(Mashruwala et al., 2015)
JMB2223	ΔsufA::tetM (SAUSA300_0843)	(Rosario-Cruz et al., 2015)
JMB2224	ΔsufA::tetM ΔsufT	(Mashruwala et al., 2016)
JMB6834	ΔsufA::tetM Δnfu::kanR	(Mashruwala et al., 2016)
JMB6835	ΔsufA::tetM Δnfu::kanR ΔsufT	(Mashruwala et al., 2016)
JMB3403	Δnfu pLL39_gfp_nfu	This work
JMB8246	Δnfu clpC::TN (ermR) pLL39_gfp_nfu	This work
JMB2015	sucA::TN (ermR) (SAUSA300_1306)	NARSA (Fey et al., 2013)
JMB2916	pdh::TN (ermR) (SAUSA300_0955)	NARSA (Fey et al., 2013)
JMB7505	bckA::TN (ermR) (SAUSA300_1465)	NARSA (Fey et al., 2013)
JMB7714	clpQ::TN(ermR) (SAUSA300_1147)	NARSA (Fey et al., 2013)
JMB7911	gcv::TN (ermR) (SAUSA300_1496)	NARSA (Fey et al., 2013)
JMB7729	ΔsufT::tetM clpQ::TN (ermR)	This work
JMB4898	clpP::TN (ermR) (SAUSA300_0752)	NARSA (Fey et al., 2013)
JMB7709	ΔsufT::tetM clpP::TN (ermR)	This work
JMB7702	trfA::TN (ermR) (SAUSA300_0899)	NARSA (Fey et al., 2013)
JMB7703	ΔsufT::tetM trfA::TN (ermR)	This work
JMB2514	ΔsufT::tetM Δnfu	This work
JMB8121	ΔsufT::tetM clpC*E362→stop nfu::TN (ermR)	This work
JMB8123	ΔsufT::tetM clpC*Y611SVIL→YFCDFNstop nfu::TN (ermR)	This work
JMB8124	ΔsufT::tetM clpC*G672→R nfu::TN (ermR)	This work
Escherichia coli PX5	Used for gene cloning	Protein Express
Plasmids used in this study
Plasmid name	Insert	Function and Reference
pEPSA5	None	Genetic complementation (Forsyth et al., 2002)
pEPSA5_clpC	SAUSA300_0510	clpC overexpression; this work
pEPSA5_ clpCG672→R	SAUSA300_0510G672→R	clpC overexpression; this work
pEPSA5_acnA	SAUSA300_1246	acnA overexpression (Mashruwala et al., 2015)
pEPSA5_nfu	SAUSA300_0839	nfu overexpression (Mashruwala et al., 2015)
pEPSA5_sufA	SAUSA300_0839	sufA overexpression (Rosario-Cruz et al., 2015)
pCM28	None	Genetic complementation (Pang et al., 2010)
pCM28_sufT	SAUSA300_0875	sufT complementation; This work
pJB38	None	Mutant construction; (Bose et al., 2013)
pJB38_ΔsufT	ΔSAUSA300_0875	(Mashruwala et al., 2016)
pJB38_ΔsufT::tetM	ΔSAUSA300_0875	Mutant construction; This work
pJB38_ΔclpC	ΔSAUSA300_0510	Mutant construction; This work
pJB38_ΔclpC::tetM	ΔSAUSA300_0510	Mutant construction; This work
pCM11	None	Promoterless gfp (Malone et al., 2009)
pCM11_sufC	sufC promoter	Transcriptional activity (Mashruwala et al., 2016)
pCM11_sufT	sufT promoter	Transcriptional activity (Mashruwala et al., 2016)
pLL39		(Luong & Lee, 2007)
pLL39_native_nfu_FLAG	SAUSA300_0839	Translational fusion; this work
pLL39_nfu_gfp	SAUSA300_0839 and Gfp	Translational fusion; this work
pTET	None	Replace ermR with tetM (Bose et al., 2013)

A defined minimal medium was used for phenotypic analyses, which has been described previously (Mashruwala et al., 2015). The three liquid growth medium formulations utilized for nutritional analyses were: 1) 20AA glucose medium, containing the 20 canonical amino acids and 14 mM glucose as a source of carbon; 2) 19AA glucose medium, lacking serine and 14 mM glucose as a source of carbon; and 3) 18AA glutamate medium, containing 18 lacking leucine and isoleucine and 44 mM glutamate as a source of carbon. Methionine was supplemented at 1.65 pM in limiting methionine (LM) media and 6.6 pM in intermediate methionine (IM) media. When provided at a non-limiting concentration, lipoic acid was supplemented at 0.5 μg mL⁻¹. Methyl butyric acid, isobutyric acid, and isovaleric acid were each supplied at 5 μM.

Top-agar overlays were created using 0.5% agar prepared in water as follows. Cells were cultured overnight before pelleting by centrifugation. Cells were resuspended in an equal volume of phosphate buffered saline (PBS) and diluted 1:100 in PBS. One-hundred μL of cells were added to 3.5 mL of 0.5% molten agar before laying the soft-agar over the top of pre-warmed solid media.

Anaerobic growth was conducted in a COY anaerobic chamber. Culture media and plastic ware was allowed to reach anaerobiosis by incubating in the chamber overnight.

Growth analyses

Nutritional requirements were assessed in 200 μL cultures in 96-well plates using a BioTek 808E Visible absorption spectrophotometer with medium shake speed at 37 °C. Culture optical density was monitored at 600 nm. Strains cultured overnight in TSB were used to inoculate staphylococcal chemically defined media to a final optical density of 0.025 (A₆₀₀) units. Before assessing nutritional requirements, cultures were harvested by centrifugation and the cell pellets were washed to prevent carryover of rich media components.

For culture on solid media, strains were cultured overnight in TSB and cells were harvested by centrifugation. Cell pellets were washed to prevent carryover of rich media components, serial diluted in PBS and 5 μL aliquots were spotted upon solid media plates.

Suppressor analyses

Ten independent cultures of JMB2512 were propagated for 18 hours in 10 mL culture tubes containing 1 mL of TSB. The cells were pelleted and resuspended in an equal volume of phosphate buffered saline. One-hundred μL of each replicate was plated onto both TSB-Rif plates and chemically defined media plates lacking lipoic acid containing amino acids (ARNDCQEGHLKMFPSTWYV). Aliquots of each replicate were also serially diluted and plated on TSB agar plates to determine the total colony forming units plated. Mutational frequency was defined as the ratio of colonies present on selective media to total CFUs plated. One colony from each chemically defined media plate was retained for phenotypic analyses and genome sequencing.

Genome sequencing

The genomes of strains JMB2512, JMB4974, JMB4980, JMB4982 and JMB4986 were sequenced using an IonTorrent PGM as previously outlined (Mitra et al., 2014). Briefly, genomic DNA (1 μg) was sheared using the Ion Xpress™ Plus Fragment Library Kit to generate libraries for 200 bp sequencing. Following purification and size selection on an E-gel Size select 2% gel (Invitrogen), DNA was quantified using a Bioanalyzer High-sensitivity DNA chip (Agilent). Libraries were diluted and template-positive Ion Sphere Particles (ISPs) prepared using the Ion OneTouch 200 template kit (Life Technologies). ISPs were sequenced using the Ion PGM 200 Sequencing kit. Whole genome sequencing data was exported from the Ion Torrent Server and analyzed using the CLC Genomics Workbench software package (Qiagen). Reads were aligned to the S. aureus genome, using the USA300 FPR3757 genome sequence as a reference. Quality-based variant detection was then performed to identify polymorphisms in each strain. A minimum threshold detection frequency of 80% was employed. The lists of polymorphisms generated for each suppressor mutant strain (JMB4974, JMB4980, JMB4982 and JMB4986) were cross-referenced against the parental strain (JMB2512). Common polymorphisms were eliminated (as were polymorphisms in homopolymeric nucleotide tracts) resulting in the identification of specific genetic variations between the suppressor strains and parental strain.

SNP verification

The SNPs identified by whole chromosome sequencing were individually verified by sequencing of PCR fragments. The clpC alleles from strains JMB2512, JMB4974, JMB4980 and JMB4986 were amplified using the clpCup5EcoRI and clpCdwn3salI primers. The fur gene and promoter region were amplified from strains JMB2512 and JMB4982 using the primers 1448upbamHI and 1448dwnPstI. The amplicons were gel purified and sequenced (Genewiz, South Plainfield, NJ) using the clpCabrseq or clpCglyseq primers.

Recombinant DNA and genetic techniques

Escherichia coli PX5 was used as a cloning host for plasmid constructions. All clones were passaged through S. aureus RN4220 (Kreiswirth et al., 1983) and subsequently transduced into the appropriate S. aureus strains using bacteriophage 80α (Novick, 1991). All S. aureus mutant strains and plasmids were verified by phenotypic analysis, analyzing PCR products, or by sequencing PCR products or plasmids (Genewiz, South Plainfield, NJ).

Creation of plasmids and mutant strains

The clpC chromosomal deletion was constructed as previously described (Rosario-Cruz et al., 2015). Briefly, the up and down-stream regions corresponding to the clpC gene were amplified using the clpCup5EcoRI/clpCup3fuse and clpCdwn5fuse/clpCdwn3salI primer pairs, respectively. The amplicons were joined using PCR using the clpCup5EcoRI and clpCdwn3salI primers and the resulting amplicon was digested with EcoRI and SalI and ligated into similarly digested pJB38 creating pJB38_ΔclpC. The tetM gene with its native promoter was amplified from strain JMB1432 using the G+tet nheI and G+tet mluI primers. The G+tet amplicon was digested with NheI and MluI and ligated into similar digested pJB38_ΔclpC creating pJB38_ΔclpC::tetM. The pJB38_sufT::tetM chromosomal deletion construct was constructed using pJB38_ ΔsufT as outlined above. The procedure to create chromosomal deletions is described elsewhere (Mashruwala et al., 2015). The lipA::Tn (tetM) strain was created using pTET as described elsewhere (Bose et al., 2013).

The pclpC complementation vectors were created by amplifying the desired clpC alleles from S. aureus chromosomes using the ClpCRBSEcoRI and ClpC3BamHI primers. The amplicons were gel purified, digested with EcoRI and BamHI before ligating into similarly digested pEPSA5. Note that the clpC clones had an engineered sodA ribosomal binding site. The psufT complementation vector was created by amplifying the sufT gene and the native promoter using the primers sufT5BamHI and sufT3SalI before digesting the amplicon with BamHI and SalI and ligating into similarly digested pCM28 resulting in pCM28_sufT.

The pLL39_native_nfu_FLAG plasmid was constructed using yeast recombinational cloning (YRC) as previously described (Mashruwala & Boyd, 2016, Joska et al., 2014). Amplicons were generated using the following primer pairs: pLLYCC5 and YCC_Nfu pro Rev; YCC_Nfu pro for and NfuCter_flag Rev; NfuCter_flag For and flag_pLLRev. The primers were designed to incorporate a NheI recognition sequence upstream of the FLAG sequence and PstI and HindIII recognition sequences downstream of the flag sequence and nfu was under the transcriptional control of its native promoter. The nfu allele and the upstream promoter region were amplified from the LAC chromosome and the pLL39 vector was linearized using SalI. The resultant pLL39_native_nfu_FLAG plasmid contained additional NheI and HindIII recognition sequences that were native to the pLL39 vector backbone. We removed these native restriction sites using primers pairs pLL39_3_no NheI and pLL39_5_no HindIII using YRC. The resultant plasmid, pLL39_nfu_FLAG was digested with NheI and PstI to remove the FLAG sequence. The gfp allele was amplified from the pCM11 plasmid (Malone et al., 2009) using the sGFPnhe15 and sGFPHindIII3 primer pair. The amplicon was digested with NheI and HindIII and ligated into similarly digested pLL39_nfu_FLAG. The resultant pLL39_nfu_gfp plasmid was integrated as an episome into the chromosome of the Δnfu strain (JMB1165).

Transcriptional reporter fusion and translational fusion assays

WT containing pCM11_nfu, or pCM11_sufT were grown TSB-Erm and the Δnfu strain containing pLL39_nfu_gfp or pLL39 were grown in TSB overnight. The cultures were diluted to 0.1 A₆₀₀ in TSB and grown to exponential or stationary growth as described in the figure legend. Two-hundred μL of cells were removed and assayed for culture optical density (A₆₀₀) and fluorescence using a Perkin Elmer Spectrometer HTS 7000 plus bioassay plate reader. GFP was excited at 485 nm and emission was read at 535 nm. For data analyses, fluorescence readings were normalized with respect to a strain not carrying a fluorescent plasmid and thereafter to culture optical density. Subsequent normalizations were as mentioned in the figures and/or figure legends.

Enzyme assays

Aconitase (AcnA) assays

AcnA assays were conducted as previously described (Mashruwala et al., 2016). Briefly, S. aureus cells cultured overnight in TSB media were washed with PBS and used to inoculate 1 mL of liquid chemically 20AA medium. Strains were cultured in 10 mL culture tubes at 37 °C for 18 hours with shaking at 200 rpm. Cells were harvested by centrifugation and cell pellets were frozen and stored at −80 °C until assed. Cell pellets were resuspended in anaerobic buffer within a COY anaerobic chamber and lysed by the addition of lysostaphin and DNAse. AcnA activity was assessed as previously described (Mashruwala et al., 2015).

Pyruvate dehydrogenase (Pdh) assays

S. aureus cells cultured overnight in TSB media were pelleted by centrifugation, washed with PBS, and subsequently used to inoculate 2 mL of liquid chemically defined 20AA or 19AA-serine media. Strains were cultured in 30 mL culture tubes for 8 hours (20AA medium) or 25 hours (19AA-serine medium) to ensure all strains were in similar growth phases. Where mentioned anaerobic growth was achieved as described earlier (Mashruwala et al., 2016). Cell pellets were harvested and lysates prepared as for aconitase assays. Pdh activity was assessed using the assay kit from BMR services (University of Buffalo) and as per manufacturer’s instructions.

Catalase assays

Cells were cultured, harvested, and cell pellets obtained as described above for Pdh assays. The cell lysate was further diluted 10-fold in lysis buffer and catalase activity was assayed by the addition of two μL of the diluted extract to 985 μL of assay buffer A (50 mM Tris, pH 7.5, 150mM NaCl, and 18 mM H₂O₂). The decomposition of H₂O₂ was monitored spectrophotometrically, as described elsewhere (Beers & Sizer, 1952).

Succinate dehydrogenase assays

Strains were cultured in TSB medium for 18 hours and cell lysates obtained as described above for AcnA assays. Sdh activity was determined as described earlier, with minor changes (Spencer & Guest, 1973, Enos-Berlage & Downs, 1997). The reaction mixture contained PBS (pH 7.5), 2 mM KCN, 742 nM XTT, 19.5 nM phenazine methosulfate and 130 mM sodium succinate (pH 7.1). The reaction was initiated by the addition of cell-free extract and reduction of XTT was monitored at 490 nm.

Protein concentration determination

Protein concentrations was determined using a copper/bicinchoninic acid based colorimetric assay modified for a 96-well plate (Olson & Markwell, 2007).

Quantification of lipoic acid in spent culture media

S. aureus strains were cultured overnight in TSB and cells were harvested, washed twice with PBS and diluted into 20AA medium to a final optical density of 0.025 (A₆₀₀). Strains were cultured to stationary phase, optical densities were recorded and the cells were partitioned from the spent media supernatant using centrifugation. The spent media supernatants were filter sterilized by passage over a 0.22 μM micron filter and standardized to culture optical density using sterile water. The lipA::Tn mutant was cultured overnight in TSB before the cells were harvested by centrifugation and cell pellets were washed using PBS. The lipA::Tn strain was subsequently cultured in fresh 20AA media that was supplemented with filter-sterilized spent media supernatant from the WT or ΔsufT strains in lieu of lipoic acid supplementation. The lipA::Tn strain was cultured for 8 hours before optical densities (A₆₀₀) were recorded. The concentration of lipoic acid in spent media supernatants was estimated using a standard curve developed using the lipA::Tn strain. The lipA::Tn strain was prepared as described above before being inoculated into 20AA media that had been supplemented with varying concentrations of lipoic acid (0–250 pg/mL) (see Figure S1).