Sulfur Mobilization for Fe-S Cluster Assembly by the Essential SUF Pathway in the Plasmodium falciparum Apicoplast and Its Inhibition

Abstract

The plastid of the malaria parasite, the apicoplast, is essential for parasite survival. It houses several pathways of bacterial origin that are considered attractive sites for drug intervention. Among these is the sulfur mobilization (SUF) pathway of Fe-S cluster biogenesis. Although the SUF pathway is essential for apicoplast maintenance and parasite survival, there has been limited biochemical investigation of its components and inhibitors of Plasmodium SUFs have not been identified. We report the characterization of two proteins, Plasmodium falciparum SufS (PfSufS) and PfSufE, that mobilize sulfur in the first step of Fe-S cluster assembly and confirm their exclusive localization to the apicoplast. The cysteine desulfurase activity of PfSufS is greatly enhanced by PfSufE, and the PfSufS-PfSufE complex is detected in vivo. Structural modeling of the complex reveals proximal positioning of conserved cysteine residues of the two proteins that would allow sulfide transfer from the PLP (pyridoxal phosphate) cofactor-bound active site of PfSufS. Sulfide release from the l-cysteine substrate catalyzed by PfSufS is inhibited by the PLP inhibitor d-cycloserine, which forms an adduct with PfSufS-bound PLP. d-Cycloserine is also inimical to parasite growth, with a 50% inhibitory concentration close to that reported for Mycobacterium tuberculosis, against which the drug is in clinical use. Our results establish the function of two proteins that mediate sulfur mobilization, the first step in the apicoplast SUF pathway, and provide a rationale for drug design based on inactivation of the PLP cofactor of PfSufS.

INTRODUCTION

The relic plastid of the malaria parasite, the apicoplast, has long been thought to be its “Achilles' heel,” as it contains pathways of bacterial origin that are not found in the human host. The apicoplast is essential for parasite survival (1, 2). It harbors unique pathways such as the type II fatty acid biosynthesis pathway, the DOXP pathway of isoprenoid synthesis, and the heme biosynthesis pathway (3–5), that offer potential novel sites for drug intervention at different stages of the infection cycle. In addition, experimental evidence of the existence of the sulfur mobilization (SUF) pathway for biogenesis of Fe-S clusters on organellar proteins was provided by our laboratory (6), and this pathway has recently been shown to be essential for apicoplast maintenance in erythrocytic stages of Plasmodium falciparum (7).

Fe-S clusters are believed to be among the first catalysts in nature; they act as versatile cofactors for essential biological functions in all living organisms by serving as catalytic centers in electron transfer enzymes and functioning in redox and nonredox catalysis (8). Fe-S clusters are most commonly assembled onto apoproteins as rhombic 2Fr-2S or cubic 4Fr-4S forms with Fe ions in the cluster being coordinated by cysteinates from the protein (9). Fe-S cluster assembly involves largely conserved core pathways. In bacteria, three distinct sets of factors or enzymes comprising the NIF (nitrogen fixation), ISC (iron-sulfur cluster formation), and SUF (sulfur mobilization) pathways assemble Fe-S clusters (10). NIF and ISC are housekeeping Fe-S assembly systems for bacteria, whereas the SUF pathway functions under conditions of oxidative stress or iron starvation (11, 12). In eukaryotes, the ISC pathway is the mitochondrial assembly system, while the SUF pathway is found in plastid-containing species (13, 14).

Components of the three pathways first mobilize sulfur atoms from l-cysteines and assemble them onto scaffold components that also receive iron from iron donors. The Fe-S cluster on the scaffold is subsequently transferred to the target apoprotein via an A-type carrier (ATC) (Fig. 1). The SUF pathway employs the cysteine desulfurase SufS and its partner SufE for release of sulfur from l-cysteine. A SufBC₂D (or SufB₂C₂) scaffold complex provides the chemical and structural environment for the assembly of Fe-S clusters. The SufC ATPase in the complex is required to bring the sulfur and iron into the scaffold protein SufB (15). The assembled Fe-S clusters on the scaffold are transferred to the ATC protein SufA, which in turn transfers them to a wide range of targets (16).

FIG 1.

Schematic representation of the steps involved in Fe-S cluster assembly on target apoproteins by the SUF pathway.

Plasmodium spp. encode the ISC and SUF pathways (8, 17), some of whose constituent proteins have recently been shown to partition to the parasite mitochondrion and apicoplast, respectively (6, 7). The 35-kb genome of the apicoplast carries the gene encoding SufB, while all of the other proteins of the SUF pathway are encoded by the nucleus. In P. falciparum, the targeting of SufC, SufS, and SufE to the apicoplast and the interaction of apicoplast-encoded SufB with SufC have been demonstrated (6, 7). Nucleus-encoded SufA and SufD homologues carrying apicoplast-targeting signal and transit peptide sequences have been identified in P. falciparum (6, 17). In P. berghei, a scaffold NifU homologue has been shown to target to the apicoplast but is proposed to play only an auxiliary or redundant role in Fe-S formation (18).

The essentiality of the SUF pathway for apicoplast maintenance and parasite survival in erythrocytic stages, as recently confirmed from disruption of the P. falciparum sufC gene (7), identifies it as a leading potential target for antimalarial drug discovery. The absence of any known inhibitors of Suf proteins underlines the need for biochemical and structural characterization of Plasmodium Suf proteins that would enable target-specific drug design and testing for inhibition. We report apicoplast-specific localization and functional characterization of PfSufS and PfSufE, which mobilize sulfur in the first step of cluster biogenesis. Using an inhibitor of the PLP (pyridoxal phosphate) cofactor, we also identify an approach for inhibiting the cysteine desulfurase activity of PfSufS.

MATERIALS AND METHODS

Parasite culture.

P. falciparum strains (3D7 and D10 ACP_leader-GFP [green fluorescent protein]) were cultured by standard methods (19). Infected red blood cells (RBCs) were maintained in RPMI 1640 supplemented with 0.5% Albumax (Invitrogen). Total RNA was isolated with TRIzol (Invitrogen), and cDNA synthesis was performed with the SuperScript III first-strand synthesis system for reverse transcription-PCR (Invitrogen).

The SYBR green assay (20) for antimalarial activity was performed with dual synchronized parasites treated with various concentrations of d-cycloserine (DCS) in ring stages at 0.5% parasitemia and 1% hematocrit. Readings were taken at 48 and 96 h on a Biotek FLX800 instrument (excitation at 485 nm, emission at 530 nm). The use of human RBCs from healthy volunteers for P. falciparum culture was approved by the CSIR-Central Drug Research Institute (CDRI) Institutional Ethics Committee (Human Research) (CDRI/IEC/CEM/21-07-2010). Written informed consent was obtained from voluntary donors for the use of this sample in research.

Protein expression and purification.

The DNA sequences encoding the predicted processed forms of PfSufS (PlasmoDB [plasmodb.org] accession no. PF3D7_0716600) and PfSufE (PlasmoDB accession no. PF3D7_0206100) were PCR amplified with P. falciparum total cDNA as the template. The sequence encoding amino acids (aa) 96 to 546 of PfSufS was amplified with forward (5′-CGCGGATCCGAGAAGATGAGTGAGTTTTAT-3′) and reverse (5′-CCCAAGCTTTTTTTCAATTTTCATTTCATTTAA-3′) primers; the gene sequence encoding aa 101 to 249 of PfSufE was amplified with primers 5′-CGCGGATCCGATGAATACAACTTAACACCAAAATTGAAA-3′ (forward) and 5′-CCCAAGCTTATTGTCCATATTCTTCAAGATATTGGTGCA-3′ (reverse) carrying BamHI and HindIII restriction sites (underlined). The PCR products were cloned into the pQE-30 vector (Qiagen), and clones were confirmed by DNA sequencing. pQE30-sufS and pQE30-sufE were cotransformed with the RIG plasmid (gift from W. G. J. Hol, University of Washington) into Escherichia coli TG1 cells. Cultures were grown at 37°C until the A₆₀₀ reached 0.4 to 0.6. Following this, the expression of recombinant proteins His₆-PfSufS and His₆-PfSufE was induced with 0.5 mM isopropyl-β-d-thiogalactopyranoside (IPTG) for 16 h at 20°C. Recombinant PfSufS (∼53 kDa) and PfSufE (∼18 kDa) were purified by affinity chromatography with Ni-nitrilotriacetic acid agarose (Qiagen). Proteins were eluted with 250 mM imidazole. Eluted PfSufS was dialyzed against a buffer containing 50 mM Tris (pH 7.5), 200 mM NaCl, and 10% glycerol, and PfSufE was dialyzed against a buffer containing 50 mM Tris (pH 8.0), 200 mM NaCl, and 10% glycerol. Purified proteins were stored at −80°C and quantitated with Bradford reagent (Sigma).

Generation of Abs.

Antibodies (Abs) against purified recombinant PfSufS and PfSufE were raised in rabbits. Approval for animal use was given by the Institutional Animal Ethics Committee, CDRI, Lucknow, India. Abs against the antigen were purified (21) and used for Western blotting and immunofluorescence microscopy.

Mutagenesis of PfSufS.

The PfSufS Cys497A mutant protein was generated with pQE30-sufS as the template. The QuikChange XL site-directed mutagenesis kit (Stratagene) was used to mutate the PfsufS gene. The mutation was confirmed by DNA sequencing. Purification of the mutant protein was carried out as for wild-type recombinant PfSufS.

Pulldown assay.

For investigation of in vitro complex formation of recombinant PfSufS and PfSufE, anti-SufS Abs were cross-linked to protein A Sepharose CL4B beads (GE Healthcare) with dimethyl pimelimidate (Sigma) cross-linker (22). PfSufS and PfSufE were coincubated for 1 h at 4°C; this was followed by further incubation of the mixture with anti-SufS Ab-cross-linked beads for 12 h at 4°C. Beads were washed with wash buffer (50 mM Tris, 1% Triton X-100, 1 mM EDTA, 150 mM NaCl, 0.25% bovine serum albumin [BSA]) and phosphate-buffered saline (PBS), and proteins were eluted in 1× Laemmli buffer. Proteins were separated on a 12% SDS-polyacrylamide (PA) gel, followed by Western blotting with horseradish peroxidase-tagged anti-His Ab (Calbiochem).

For coimmunoprecipitation from parasite lysate, parasites were lysed in buffer containing 50 mM Tris, 1% Triton X-100, 0.6 M KCl, protease inhibitor cocktail, and 1 mM phenylmethylsulfonyl fluoride; this was followed by preclearing with protein A Sepharose CL4B beads cross-linked to anti-PfSufS Ab as described in reference 23. The precleared lysate was incubated with fresh cross-linked Sepharose beads for 12 h at 4°C. The beads were then washed three times with 0.1% BSA in 1× PBS. The beads were boiled in 2× Laemmli buffer for 5 min, and the samples were loaded onto a 12% SDS-PA gel; this was followed by Western blotting with anti-PfSufS and anti-PfSufE Abs, respectively.

Chemical cross-linking.

In vivo cross-linking was carried out with dithiobis(succinimidyl propionate) (DSP; Calbiochem), a reduction-sensitive, cell-permeating cross-linker. Cross-linking was performed as described in reference 24. Briefly, infected RBCs were pelleted at 400 × g for 15 min and washed with PBS. Cells were resuspended in PBS containing 2 mM DSP and incubated at room temperature for 30 min. The cross-linking reaction was stopped by incubation with a final concentration of 25 mM Tris (pH 7.5) for 15 min at room temperature. Parasites were released from RBC by saponin lysis, resuspended in 1× NRSB buffer (0.05 M Tris-Cl, 10% glycerol, 2 mM EDTA, 2% SDS, bromophenol blue) with increasing concentrations of dithiothreitol (DTT), and then incubated at 80°C for 5 min. The samples were electrophoresed in an SDS-PA gel and then Western blotted with rabbit anti-SufS and anti-SufE Abs.

Cysteine desulfurase assay.

PfSufS desulfurase activity was determined by the sulfide detection method as described in reference 25. Enzymatic assays were carried out anaerobically in 100 μl of the final reaction mixture. Reaction mixtures contained 5 μM PfSufS (or E. coli SufS) and various molar ratios of PfSufE in 25 mM Tris-Cl–100 mM NaCl buffer (pH 7.4)–100 μM PLP–2 mM DTT. Reactions were initiated by the addition of l-cysteine to a final concentration of 1 mM and allowed to proceed for 20 min. Reactions were stopped by the addition of 12 μl of 20 mM N,N-dimethyl-p-phenylenediamine (Sigma) in 7.2 M HCl and 12 μl of 30 mM ferric chloride (Sigma) in 1.2 M HCl. Sulfide production was quantified colorimetrically by measuring absorbance at 670 nm. Inhibition by DCS was assayed as described above, except that PfSufS and PfSufE were preincubated with various concentrations of DCS for 1 h at 4°C prior to initiation of the reaction with l-cysteine.

PLP binding assay.

UV-visible absorption spectra of PfSufS and EcSufS were taken in 50 mM Tris-Cl–200 mM NaCl buffer (pH 7.5). Absorption spectra were recorded at wavelengths of 250 to 500 nm. When a sample was reduced by the addition of 2 mM sodium borohydride (NaBH₄₎, the spectrum was recorded 1 min after the addition of NaBH₄ to the enzyme. The UV-VIS spectrum in the presence of DCS was recorded after the incubation of PfSufS with 5 or 0.5 mM DCS for 3 h at 4°C.

Confocal microscopy.

Immunofluorescence assays were carried out as described by Tonkin et al. (4). For mitochondrial staining, P. falciparum 3D7 cells were incubated with 50 ng/ml Mitotracker Red CMXRos (Invitrogen) prior to fixation. Cells were washed with PBS and fixed with 4% paraformaldehyde and 0.0075% glutaraldehyde in PBS for 30 min. After one wash with PBS, cells were permeabilized with 0.1% Triton X-100 in PBS for 10 min. After another PBS wash, cells were treated with 0.1 mg/ml NaBH₄ in PBS for 10 min. Cells were washed again with PBS and blocked in 3% BSA–PBS for 1 h. For PfSufS localization, cells were incubated overnight with rabbit anti-PfSufS serum (1:25 dilution) and mouse anti-PfHU (nucleus-encoded bacterial histone-like protein) serum (1:100 dilution) (23) at 4°C. After five washes with PBS, cells were probed with Alexa Fluor 568-tagged anti-rabbit Ab (1:1,000 dilution; Invitrogen) and Alexa Fluor 488-tagged anti-mouse Ab (1:1,000 dilution; Invitrogen) for 2 h at room temperature and allowed to settle onto poly-l-lysine-coated coverslips. For localization of PfSufE, P. falciparum D10-ACP_leader-GFP (26) cells were fixed as described above and parasites were probed with anti-PfSufE serum (1:25 dilution) and mouse anti-GFP Ab (1:100 dilution; Roche) as primary Abs and Alexa Fluor 488-tagged anti-mouse Ab (1:1,000 dilution; Invitrogen) and Alexa Fluor 568-tagged anti-rabbit Ab (1:1,000 dilution; Invitrogen) as secondary Abs. For costaining with Mitotracker Red, P. falciparum 3D7 cells were fixed and probed with anti-PfSufE serum as the primary Ab and Alexa Fluor 514-tagged anti-rabbit secondary Ab. Slides were scanned in a confocal laser scanning microscope (Zeiss LSM-510) under a 63× oil immersion lens.

Molecular modeling.

To explore the three-dimensional structures and interaction of PfSufS and PfSufE, homology models were constructed. BLASTP of the PfSufS and PfSufE sequences was performed against the sequences in the Protein Data Bank (PDB) to identify suitable templates. Crystal structures of the E. coli CsdB NifS homologue (1C0N) and E. coli SufE (1MZG) were selected as the templates for PfSufS and PfSufE, respectively, on the basis of identity and query coverage. Models were generated with Modeler9v10 (27). Models were then subjected to the SAVeS server (http://nihserver.mbi.ucla.edu/SAVES/) for structural validation by the Ramachandran plot.

To get an insight into the PfSufS-PfSufE interaction, docking of PfSufS and PfSufE was performed with the GRAMMX server (http://vakser.bioinformatics.ku.edu/resources/gramm/grammx/) (28), which employs smoothed potentials, extensive refinement, and knowledge-based scoring with fast Fourier transform global search methodology to predict the structure of a protein complex.

RESULTS AND DISCUSSION

PfSufS and PfSufE are processed and targeted to the apicoplast.

Homologs of bacterial SufS and SufE are encoded by the Plasmodium nuclear genome (PlasmoDB accession no. PF3D7_0716600 and PF3D7_0206100, respectively). Although these proteins have low overall sequence identity with their E. coli counterparts, they carry conserved domains and residues that are critical for cofactor binding and function (see Fig. S1 and S2 in the supplemental material). PfSufS contains a PLP-binding motif (SGHK) with a conserved Lys residue that forms an internal aldimine with PLP; other residues that interact with PLP in the internal aldimine and substrate-bound external aldimine forms (29) are also conserved (see Fig. S1). PfSufS also has the conserved Cys₄₉₇ residue (corresponding to E. coli Cys₃₆₄) in a typical consensus region, RXGHHCA, found in NifS-like proteins classified as group II cysteine desulfurases (30). On the basis of sequence alignment (see Fig. S2), PfSufE is predicted to have a 102-aa N-terminal apicoplast-targeting leader sequence (31) and carries a conserved Cys₁₅₄ residue (corresponding to Cys₅₁ in E. coli SufE) that is known to receive sulfur from SufS in the bacterial system (32).

In order to initiate biochemical characterization of PfSufS and PfSufE homologs, we carried out recombinant expression of their predicted processed forms in E. coli. Expression of the predicted processed form of PfSufS (aa 60 to 546) from a codon-optimized synthetic gene with an N-terminal 6×His or maltose-binding protein tag was attempted. However, the protein was highly unstable and a full-length form could not be purified. A functional form of PfSufS retaining the essential conserved motifs and covering aa 96 to 546 was subsequently expressed as a 6×His-tagged protein, purified, and used in subsequent experiments (Fig. 2Ai). Some lower-molecular-weight degraded forms of the protein were also obtained in addition to major full-length 53-kDa PfSufS. PfSufE (aa 101 to 249, ∼20 kDa) was expressed as a soluble protein with an N-terminal 6×His tag (Fig. 2Bi). Polyclonal Abs against purified PfSufS and PfSufE were raised in rabbits and used to detect the proteins in P. falciparum lysates. A 61.5-kDa band corresponding to the size expected for a signal-cleaved form and a 56-kDa band corresponding to the expected size of signal- and transit-cleaved PfSufS were detected with the anti-PfSufS serum in a Western blot assay (Fig. 2Aii). The anti-PfSufE serum also detected a major 18-kDa band corresponding to the processed PfSufE protein and a 27-kDa band that would represent the signal-cleaved form. Two other bands that might represent degraded or intermediate cleaved forms of the 27-kDa protein were also detected. These results demonstrate that the two proteins predicted to play a role in sulfur mobilization in the SUF pathway are expressed in the parasite.

FIG 2.

Expression of recombinant PfSufS and PfSufE and their detection in the parasite. (A) Coomassie-stained SDS-PAGE of purified recombinant PfSufS expressed as a 6×His-tagged protein in E. coli (i) and detection of the protein in P. falciparum lysate (ii) by Western blotting with preimmune serum (Pre-I) and anti-SufS serum (I). (B) Purified recombinant PfSufE seen in a Coomassie-stained SDS-PA gel (i) and detection of the protein in parasite lysate by Western blotting with anti-PfSufE Ab (I) (ii).

The subcellular localization of PfSufS and PfSufE in the parasite was investigated by immunofluorescence assays with anti-PfSufS and anti-PfSufE serum. PfSufS colocalized with PfHU, an apicoplast DNA condensation protein (23), and no overlapping signals for PfSufS were seen with the mitochondrial marker Mitotracker Red in P. falciparum 3D7 cells (Fig. 3A). PfSufE was localized in the P. falciparum D10-ACP_leader-GFP line that targets GFP to the apicoplast (26). Specific PfSufE signals overlapping apicoplast-targeted GFP were observed, and no PfSufE signal overlap was seen with Mitotracker Red (Fig. 3B). The localization of PfSufS and PfSufE exclusively to the apicoplast was thus confirmed. Localization of these two proteins to the apicoplast has also recently been demonstrated with a C-terminal GFP tag (7). Unlike Arabidopsis thaliana, where SufE is dually localized to the plastid and mitochondria (33), P. falciparum SufE is targeted exclusively to the apicoplast. As the apicoplast localization of the PfSufC ATPase and its interaction with apicoplast-encoded scaffold protein PfSufB have been reported earlier (6), we conclude that the first two steps of the SUF pathway, namely, sulfur mobilization and Fe-S assembly on the scaffold, are present in the apicoplast. The targeting of the third predicted scaffold component PfSufD and the Fe-S transfer proteins PfSufU and PfSufA remains to be confirmed.

FIG 3.

Immunofluorescence localization of PfSufS and PfSufE. (A) Immunofluorescence assay of P. falciparum 3D7 cells with anti-PfSufS Ab with anti-PfHU Ab as an apicoplast marker (i) or Mitotracker Red (ii) indicates targeting of PfSufS to the apicoplast. (B) Colocalization of PfSufE and apicoplast-targeted GFP observed in the P. falciparum D10 ACP_leader-GFP cell line with anti-PfSufE and anti-GFP Abs (i). No overlap of PfSufE signal was observed with Mitotracker Red in P. falciparum 3D7 cells (ii), indicating apicoplast localization of PfSufE. DIC, differential interference contrast; DAPI, 4′,6-diamidino-2-phenylindole.

The cysteine desulfurase activity of PfSufS is enhanced by PfSufE.

Cysteine desulfurases are classified as class V aminotransferases of fold type 1 (34). They are PLP-dependent enzymes with PLP bound to a Lys residue in a highly conserved HK motif that is also found in PfSufS. The binding of PLP to concentrated purified PfSufS was indicated by its light yellow color, which is typical of PLP-binding enzymes. The UV-VIS absorption spectrum of purified recombinant PfSufS displayed a peak centered around 420 nm in addition to the major protein peak at 280 nm. The 420-nm peak is indicative of bound pyridoxal in an aldimine linkage in the enzyme (Fig. 4A). Addition of the reducing agent sodium borohydride to PfSufS resulted in reduction of the 420-nm peak and the appearance of a peak close to 335 nm (reduced PLP), an observation consistent with its known effect on PLP-binding proteins (30) (Fig. 4A). A similar shift in the UV-VIS absorption spectrum in the presence of sodium borohydride was also seen with E. coli SufS, which is known to be a PLP-dependent enzyme (35) (Fig. 4A, inset).

FIG 4.

PfSufS is a PLP cofactor-dependent desulfurase. (A) UV-VIS absorption spectrum of PfSufS with and without treatment with NaBH₄. The inset shows the effect of NaBH₄ on the absorption spectrum of E. coli SufS. The 420-nm peak of PfSufS-bound PLP shifts to a 335-nm peak of reduced PLP after treatment with NaBH₄. (B) Enhancement of cysteine desulfurase activity of PfSufS with increasing concentrations of PfSufE. All of the reaction mixtures except the control (20 μM PfSufE alone) contained 5 μM PfSufS. (C) Effect of the PfSufS Cys₄₉₇Ala mutation on sulfide release in the presence of PfSufE.

The ability of recombinant PfSufS to function as a cysteine desulfurase was measured in the presence of the l-cysteine substrate (25). PfSufS alone exhibited very low cysteine desulfurase activity (Fig. 3B). The activity of PfSufS was greatly enhanced by PfSufE (up to ∼17-fold) in a concentration-dependent manner (Fig. 4B). PfSufE by itself had no cysteine desulfurase activity. Even though PfSufE has very low sequence identity with E. coli SufE (∼23%), it could enhance the activity of E. coli SufS (see Fig. S3 in the supplemental material), indicating that critical interactions between the two proteins are conserved in Plasmodium and bacteria. Our results demonstrate that the P. falciparum apicoplast-targeted PfSufS homolog is indeed a cysteine desulfurase that requires PfSufE for efficient mobilization of sulfur from l-cysteine.

The sulfane sulfur released from the l-cysteine substrate by cysteine desulfurases is transferred to generate a protein-bound persulfide intermediate at the SufS active site (32), and the stimulatory effect of SufE is dependent on transfer of the sulfur atom from the active site persulfide to a conserved cysteine of SufE (25). The SufS active site cysteine in E. coli is Cys₃₆₄, which corresponds to Cys₄₉₇ in PfSufS. In order to confirm the role of this residue in cysteine desulfurase activity, we generated a PfSufS Cys₄₉₇Ala mutant protein. The mutant protein PfSufS was unable to mobilize sulfur, and no desulfurase activity could be detected in the presence of PfSufE, thus establishing that Cys₄₉₇ is the critical residue involved in sulfur mobilization from the cysteine substrate (Fig. 4C).

PfSufS-PfSufE complexation is detected in the parasite.

The interaction of recombinant PfSufS and PfSufE was investigated by a pulldown assay with beads cross-linked to anti-PfSufS Ab (Fig. 5A). After coincubation and washing, PfSufE was recovered together with PfSufS (Fig. 5A, lane 2) and did not bind the beads in the absence of its interacting partner (Fig. 5A, lane 3). In vitro complexation of the two proteins was thus observed.

FIG 5.

PfSufS and PfSufE complex in vitro and in vivo. (A) Coimmunoprecipitation of recombinant PfSufS and PfSufE with PfSufS Abs cross-linked to protein A Sepharose beads, followed by Western blotting with anti-His Ab. Arrows indicate ∼53-kDa PfSufS and ∼20-kDa PfSufE (lane 2). No signal was observed in beads cross-linked with preimmune serum (lane 1) or in the presence of PfSufE alone (lane 3). Lane 5 is purified PfSufE used as size marker. (B) Coimmunoprecipitation of PfSufE with PfSufS from parasite lysates with beads cross-linked to anti-PfSufS Ab. Coprecipitated PfSufS (i) and PfSufE (ii) were detected in Western blot assays with anti-PfSufS and anti-PfSufE Abs, respectively. (C) Parasite proteins were cross-linked in vivo by DSP and then treated with increasing concentrations of DTT to break the complex(s). SDS-PA gels of the samples were probed with anti-PfSufE Ab (i) or anti-PfSufS Ab (ii) in Western blot assays to detect the complex and released proteins.

Immunoprecipitation from parasite lysates with beads cross-linked to anti-PfSufS Ab resulted in the coprecipitation of PfSufS and PfSufE, as revealed by Western blotting (Fig. 5B). Formation of the PfSufS-PfSufE complex in the parasite was confirmed by chemical cross-linking by DSP, followed by reduction of the cross-linked products by DTT. A cross-linked complex (∼74 kDa) and its DTT-mediated breakage products were detected by Western blotting with anti-PfSufE and anti-PfSufS Abs (Fig. 5C). The intensity of this complex decreased slightly in the presence of DTT (up to 200 μM) with increasing intensity of released PfSufE (Fig. 5Ci) or PfSufS (Fig. 5Cii). An ∼37-kDa band, possibly representing a degraded product of PfSufS, was also detected in addition to the 56-kDa processed form of the protein (Fig. 5Cii). The coimmunoprecipitation of PfSufS and PfSufE from parasite lysates, as well as the size of the DSP-cross-linked complex (74 kDa), suggests that monomers of PfSufS (56 kDa) and PfSufE (18 kDa) interact within the parasite apicoplast to mobilize sulfur release from l-cysteine.

Structural modeling of the PfSufS-PfSufE interaction.

Homology modeling of PfSufS and PfSufE was carried out with crystal structures of E. coli CsdB/SufS (PDB accession no. 1C0N) and E. coli SufE (PDB accession no. 1MZG) as the templates. Processed PfSufS and PfSufE have only 30 and 23% identity, respectively, with their E. coli homologs (see Fig. S1 and S2 in the supplemental material). Even with its very low identity, PfSufE was modeled very well on the template (see Fig. S4), with Ramachandran plot analysis indicating 96.1, 3.9, 0.0, and 0.0% of its residues in the core, allowed, generously allowed, and disallowed regions, respectively. Although major structural folds of PfSufS were also conserved, the large 66-aa insertion in its C-terminal half could not be modeled as it lacked identity with any known protein and appeared as a loop in the structure (see Fig. S4). Overall, the modeled PfSufS structure had 86.6, 9.9, 2.4, and 1.1% of its residues in the core, allowed, generously allowed, and disallowed regions in Ramachandran plot analysis. Structural superimpositions with templates showed root mean square deviations of 0.598 and 0.387 Å for SufS and SufE, respectively.

If the sulfur atom of the persulfide on Cys₄₉₇ of PfSufS (corresponding to Cys₃₆₄ of EcSufS) has to be transferred to Cys₁₅₄ of PfSufE (corresponding to Cys₅₁ of EcSufS) through a direct transpersulfuration reaction (32), the two cysteines of the interacting partners must be in proximity in the PfSufS-PfSufE complex. Since the structure of a SufS-SufE complex is not available, we predicted the binding interaction between PfSufS and PfSufE by protein-protein docking with the GRAMM-X docking server. Fifty possible solutions for the PfSufS-PfSufE interaction were generated. One of the docked complexes was selected on the basis of predicted proximity of the critical cysteine residues of PfSufS and PfSufE and maximal representation in the solutions (Fig. 6). Cys₄₉₇ of PfSufS and Cys₁₅₄ of PfSufE are present in the protein-protein interface in the docked solution. The distance between the sulfur atoms of the two cysteines is 8.477 Å, which would be further reduced by the additional sulfur atom of the persulfide on the PfSufS cysteine, thus making direct transpersulfuration to PfSufE plausible.

FIG 6.

Structural model of PfSufS-PfSufE interaction. (A) Protein docking with the GRAMMX server identified a complex with proximal positioning of the critical cysteine residues of the interacting proteins. A large insertion in PfSufS cannot be modeled and appears as a loop. (B) Enlarged portion of panel A that shows the active site and the PfSufS-PfSufE interface. The PLP cofactor is in black.

SufS from bacteria is a homodimeric enzyme with a well-separated and distinct active site in each subunit with some residues of the other subunit contributing to the active site (30, 36). The size of the complex obtained by in vivo cross-linking in P. falciparum suggests that apicoplast PfSufS is a monomer that interacts with PfSufE. Recombinant PfSufS was also purified as a monomer, and a dimeric form was not detected in gel filtration assays (data not shown). The presence of the large 66-aa insertion in PfSufS might interfere with dimerization of the protein, although this does not seem to inhibit the functionality of its active site.

A PLP-binding drug inhibits PfSufS and exhibits antimalarial activity.

Since no inhibitors of components of the SUF pathway are known, we investigated the possibility of identifying putative inhibitors that would act either on the PfSufS active site or at the PfSufS-PfSufE interface. As shown in the PfSufS-PfSufE interaction model (Fig. 6), the protein-protein interface lacks a distinct pocket into which a putative inhibitor might be docked. Thus, the PfSufS active site that includes the bound PLP cofactor was considered next. Inhibitors of PLP-dependent enzymes that form adducts with the cofactor have been used as irreversible inhibitors of enzyme activity (37). One such inhibitor, DCS, is in human clinical use as a second-line drug against Mycobacterium tuberculosis (38). DCS acts on enzymes with PLP fold type 1 (aspartate aminotransferase family) (37, 39, 40). Since cysteine desulfurases are PLP fold type 1 proteins (41), we explored the effect of DCS on PLP interactions in PfSufS. A reduction of the 420-nm peak of the internal aldimine between PLP and Lys291 (corresponding to Lys226 of E. coli) was seen in the UV-VIS spectrum of PfSufS in the presence of DCS (Fig. 7A). The reduction of the 420-nm peak was accompanied by increased absorbance at ∼320 nm that corresponds to the peak expected for the PLP-DCS adduct (3-hydroxyisoxazole-pyridoxamine derivative) as reported for other PLP-dependent enzymes inhibited by DCS (39, 40). DCS is thus capable of forming an adduct with the PLP cofactor bound to PfSufS.

FIG 7.

DCS inhibits PfSufS activity. (A) UV-VIS absorption spectrum of PfSufS in the presence of DCS showing a reduction in the 420-nm bound PLP peak accompanied by the appearance of a peak at 320 nm representing a PLP-DCS adduct. (B) Effect of increasing concentrations of DCS on cysteine desulfurase activity of PfSufS. The IC₅₀ of DCS was determined to be 29.2 ± 2.9 μM.

The effect of DCS on the cysteine desulfurase activity of recombinant PfSufS was investigated. Dose-dependent inhibition of PfSufS activity by DCS was observed in the presence of PfSufE. The 50% inhibitory concentration (IC₅₀) was calculated to be ∼29 μM (Fig. 7B). The antimalarial action of DCS in P. falciparum blood stage parasites was assayed by SYBR green whole-cell screening (20). Compounds that exhibit apicoplast-specific action are known to cause a delayed-death effect in which parasite growth is inhibited primarily in the second cycle of infection (42, 43). The IC₅₀ of DCS for P. falciparum strains was thus determined for the first and second infection cycles (at 48 and 96 h posttreatment). The IC₅₀s of DCS for the growth of P. falciparum 3D7 were ∼62 and 68 μM in the first and second cycles, respectively (Table 1). The IC₅₀s of DCS for chloroquine-resistant P. falciparum K1 were determined to be ∼62 and 71 μM in first and second cycles. These values are close to the reported IC₅₀ of DCS for M. tuberculosis H37Rv culture growth (58 μM) (44).

TABLE 1.

IC₅₀ of DCS for the first and second infection cycles

Time (h)	Avg IC50 ± SD (μM)
	P. falciparum 3D7	P. falciparum K1
48	62.4 ± 0.9	62.2 ± 5.2
96	67.9 ± 1.5	71.16 ± 0.3

The inhibition of PfSufS by DCS provides evidence for the possibility of designing inhibitors targeting the PLP moiety of this important enzyme of the apicoplast Fe-S biogenesis pathway. The absence of a delayed-death effect upon DCS treatment of blood-stage parasites suggests that the drug also has additional, nonapicoplast targets in the parasite. There are more than a dozen PLP-dependent enzymes in P. falciparum (http://priweb.cc.huji.ac.il/malaria/maps/vitaminB6metpath.html) at least eight of which are fold type 1 enzymes that could also serve as targets of DCS (see Table S1 in the supplemental material); these include PfIscS, the cysteine desulfurase of the ISC pathway of the parasite mitochondrion. However, PfSufS would be a major target of DCS in vivo, as also supported by the fact that the IC₅₀ for parasite growth is higher than the IC₅₀ for PfSufS desulfurase activity.

The apicoplast SUF pathway is believed to play a critical role in parasite biology by mediating the assembly of Fe-S clusters on target proteins that include ferredoxin (Fd)/ferredoxin reductase (FNR), lipoic acid synthase (LipA), (dimethylallyl)adenosine tRNA methylthiotransferase (MiaB), and the last two enzymes of the nonmevalonate DOXP pathway of isoprenoid biosynthesis [(E)-4-hydroxy-3-methylbut-2-en-1-yl diphosphate synthase (IspG) and (E)-4-hydroxy-3-methylbut-2-en-1-yl diphosphate reductase (IspH)] (45). The antimalarial effect of inhibition of Fe-S assembly on target proteins in the apicoplast has been recently demonstrated through the generation of dominant negative SufC mutant parasites (7). These parasites lost the apicoplast and the organellar genome and were viable only when supplemented with isopentenyl pyrophosphate, an essential product of the apicoplast DOXP pathway that supports parasite survival in blood culture even when the apicoplast is lost (46). Our results demonstrate that it is possible to inhibit the apicoplast SUF pathway with compounds targeting the PLP cofactor of its constituent cysteine desulfurase. Unlike its enantiomer l-cycloserine, DCS is a natural product with low toxicity in humans (38). Other PLP inhibitors with demonstrated effects on PLP cofactor-dependent target enzymes in other organisms (37) can also be tested for inhibition of Plasmodium SufS in vitro and as well as for their antimalarial activity.

The SUF pathway is not the default pathway for Fe-S biogenesis in bacteria, where the ISC and NIF pathways play more significant roles and SUF proteins are activated only under conditions of stress. This redundancy of the SUF pathway in bacteria is abolished in Plasmodium spp., where ISC and SUF components are partitioned to the mitochondrion and apicoplast, respectively, thus conferring all Fe-S biogenesis function on critical apicoplast targets to the SUF pathway. In addition, the SUF pathway is absent from humans, making its components attractive targets for inhibitor identification. The design of novel chemical inhibitors and generation of suitable assay systems with recombinant Plasmodium SUF proteins and their complexes is the obvious next step.