Dap1p, a heme-binding protein that regulates the cytochrome P450 protein Erg11p/Cyp51p in Saccharomyces cerevisiae

Alkylating agents chemically modify DNA and cause mutations that lead to cancer. In the budding yeast Saccharomyces cerevisiae, resistance to the alkylating agent methyl methanesulfonate (MMS) is mediated in part by Dap1p (damage resistance protein 1). Dap1p is related to cytochrome b5, which activates cytochrome P450 proteins, elevating the metabolism of lipids and xenobiotic compounds. We have found that Dap1p, like cytochrome b5, binds to heme and that Dap1p targets the cytochrome P450 protein Erg11p/Cyp51p. Genetic analysis indicates that Erg11p acts downstream of Dap1p. Furthermore, Dap1p regulates the stability of Erg11p, and Erg11p is stabilized in dap1Delta cells by the addition of heme. Thus, Dap1p utilizes heme to stabilize Erg11p, which in turn regulates ergosterol synthesis and MMS resistance. Dap1p homologues have been identified in numerous eukaryotes, including mammals, suggesting that the Dap1p-cytochrome P450 protein pathway is broadly conserved in eukaryotic species.     

Spectroscopic and biochemical characterization of heme binding to yeast Dap1p and mouse PGRMC1p

Yeast damage-associated response protein (Dap1p) and mouse progesterone receptor membrane component-1 protein (mPGRMC1p) belong to a highly conserved class of putative membrane-associated progesterone binding proteins (MAPR), with Dap1p and inner zone antigen (IZA), the rat homologue of mPGRMC1p, recently being reported to bind heme. While primary structure analysis reveals similarities to the cytochrome b(5) motif, neither of the two axial histidines responsible for ligation to the heme is present in any of the MAPR proteins. In this paper, EPR, MCD, CD, UV-vis, and general biochemical methods have been used to characterize the nature of heme binding in both Dap1p and a His-tagged, membrane anchor-truncated mPGRMC1p. As isolated, Dap1p is a tetramer which can be converted to a dimer upon addition of 150 mM salt. The heme is noncovalently attached, with a maximal, in vitro, heme loading of approximately 30%, for both proteins. CD and fluorescence spectroscopies indicate a well-ordered structure, suggesting the low level of heme loading is probably not due to improperly folded protein. EPR confirmed a five-coordinate, high-spin, ferric resting state for both proteins, indicating one axial amino acid ligand, in contrast to the six-coordinate, low-spin, ferric state of cytochrome b(5). The MCD spectrum confirmed this conclusion for Dap1p and indicated the axial ligand is most likely a tyrosine and not a histidine, or a cysteine; however, an aspartic acid residue could not be conclusively ruled out. Potential axial ligands, which are conserved in all MAPRs, were mutated (Y78F, D118A, and Y138F) and purified to homogeneity. The Y78F and D118A mutants were found to bind heme; however, Y138F did not. This result is consistent with the MCD data and indicates that Tyr138 is most likely the axial ligand to the heme in Dap1p.     

Regulation of iron homeostasis mediated by the heme-binding protein Dap1 (damage resistance protein 1) via the P450 protein Erg11/Cyp51

Fungal infections arise frequently in immunocompromised patients, and sterol synthesis is a primary pathway targeted by antifungal drugs. In particular, the P450 protein Erg11/Cyp51 catalyzes a critical step in ergosterol synthesis, and the azole class of antifungal drugs inhibits Erg11. Dap1 is a heme-binding protein related to cytochrome b5 that activates Erg11, so that cells lacking Dap1 accumulate the Erg11 substrate and are hypersensitive to Erg11 inhibitors. Heme binding by Dap1 is crucial for its function, and point mutants in its heme-binding domain render Dap1 inactive for sterol biosynthesis and DNA damage resistance. Like Dap1, the human homologue, PGRMC1/Hpr6, also regulates sterol synthesis and DNA damage resistance. In the present study, we demonstrate that the Dap1 heme-1 domain is required for growth under conditions of low iron availability. Loss of Dap1 is suppressed by elevated levels of Erg11 but not by increased heme biosynthesis. Dap1 localizes to punctate cytoplasmic structures that co-fractionate with endosomes, and Dap1 contributes to the integrity of the vacuole. The results suggest that Saccharomyces cerevisiae Dap1 stimulates a P450-catalyzed step in sterol synthesis via a distinct localization from its homologues in Schizosaccharomyces pombe and mammals and that this function regulates iron metabolism.     

Interactions of oxidosqualene cyclase (Erg7p) with 3-keto reductase (Erg27p) and other enzymes of sterol biosynthesis in yeast

In Saccharomyces cerevisiae and Candida albicans, two enzymes of the ergosterol biosynthetic pathway, oxidosqualene cyclase (Erg7p) and 3-keto reductase (Erg27p) interact such that loss of the 3-keto reductase also results in a concomitant loss of activity of the upstream oxidosqualene cyclase. This interaction wherein Erg27p has a stabilizing effect on Erg7p was examined to determine whether Erg7p reciprocally has a protective effect on Erg27p. To this aim, three yeast strains each lacking the ERG7 gene were tested for 3-ketoreductase activity by incubating either cells or cell homogenates with unlabeled and radiolabeled 3-ketosteroids. In these experiments, the ketone substrates were effectively reduced to the corresponding alcohols, providing definitive evidence that oxidosqualene cyclase is not required for the 3-ketoreductase activity. This suggests that, in S. cerevisiae, the protective relationship between the 3-keto reductase (Erg27p) and oxidosqualene cyclase (Erg7p) is not reciprocal. However, the absence of the Erg7p, appears to affect other enzymes of sterol biosynthesis downstream of lanosterol formation. Following incubation with radiolabeled and non-radiolabeled 3-ketosteroids we detected differences in hydroxysteroid accumulation and ergosterol production between wild-type and ERG7 mutant strains. We suggest that oxidosqualene cyclase affects Erg25p (C-4 sterol oxidase) and/or Erg26p (C-3 sterol dehydrogenase/C-4 decarboxylase), two enzymes that, in conjunction with Erg27p, are involved in C-4 sterol demethylation.     

The ERG28-encoded protein, Erg28p, interacts with both the sterol C-4 demethylation enzyme complex as well as the late biosynthetic protein, the C-24 sterol methyltransferase (Erg6p)

In Saccharomyces cerevisiae, the C-24 sterol methyltransferase (Erg6p) converts zymosterol to fecosterol, an enzymatic step following C-4 demethylation of 4,4-dimethylzymosterol. Our previous study showed that an endoplasmic reticulum (ER) transmembrane protein, Erg28p, functions as a scaffold to tether the C-4 demethylation enzymatic complex (Erg25p-Erg26p-Erg27p) to the ER. To determine whether Erg28p also interacts with other ergosterol biosynthetic proteins, we compared protein levels of Erg3p, Erg6p, Erg7p, Erg11p and Erg25p in three pairs of erg28 and ERG28 strains. In erg28 strains, the Erg6p level in the ER fraction was decreased by about 50% relative to the wild-type strain, while ER protein levels of the four other ergosterol proteins showed no significant differences. Co-immunoprecipitation experiments, using an erg28 strain transformed with the epitope-tagged plasmid pERG28-HA and proteins detected with anti-HA and anti-Erg6p antibodies, indicated that Erg6p and Erg28p reciprocally co-immunoprecipitate. Further, the split ubiquitin yeast membrane two-hybrid system designed to detect protein interactions between membrane bound proteins also indicated an Erg28p-Erg6p interaction when pERG6-Cub was used as the bait and pERG28-NubG was used as the prey. We conclude that Erg28p may not only anchor the C-4 demethylation enzyme complex to the ER but also acts as a protein bridge to the Erg6p enzyme required for the next ergosterol biosynthetic step.     

Inversion of axial coordination in myoglobin to create a "proximal" ligand binding pocket

A ligand binding pocket has been created on the proximal side of the heme in porcine myoglobin by site-directed mutagenesis. Our starting point was the H64V/V68H double mutant which has been shown to have bis-histidine (His68 and His93) heme coordination [Dou, Y., Admiraal, S. J., Ikeda-Saito, M., Krzywda, S., Wilkinson, A. J., Li, T., Olson, J. S., Prince, R. C., Pickering, I. J., George, G. N. (1995) J. Biol. Chem. 270, 15993-16001]. The replacement of the proximal His93 ligand by noncoordinating Ala (H64V/V68H/H93A) or Gly (H64V/V68H/H93G) residues resulted unexpectedly in a six-coordinate low-spin species in both ferric and ferrous states. To test the hypothesis that the sixth coordinating ligand in the triple mutants was the imidazole of His97, this residue was mutated to Phe, in the quadruple mutants, H64V/V68H/H93A/H97F and H64V/V68H/H93G/H97F. The ferric quadruple mutants show a clear water/hydroxide alkaline transition and high cyanide and CO affinities, characteristics similar to those of wild-type myoglobin. The nu(Fe-CO) and nu(C-O) stretching frequencies in the ferrous-CO state of the quadruple mutants indicate that the "proximal" ligand binding heme pocket is less polar than the distal pocket in the wild-type protein. Thus, we conclude that the proximal heme pocket in the quadruple mutants has a similar affinity for exogenous ligands to the distal pocket of wild-type myoglobin but that the two pockets have different polarities. The quadruple mutants open up new approaches for developing heme chemistry on the myoglobin scaffold.     

Important roles of Tyr43 at the putative heme distal side in the oxygen recognition and stability of the Fe(II)-O2 complex of YddV, a globin-coupled heme-based oxygen sensor diguanylate cyclase

YddV from Escherichia coli (Ec) is a novel globin-coupled heme-based oxygen sensor protein displaying diguanylate cyclase activity in response to oxygen availability. In this study, we quantified the turnover numbers of the active [Fe(III), 0.066 min(-1); Fe(II)-O(2) and Fe(II)-CO, 0.022 min(-1)] [Fe(III), Fe(III)-protoporphyrin IX complex; Fe(II), Fe(II)-protoporphyrin IX complex] and inactive forms [Fe(II) and Fe(II)-NO, <0.01 min(-1)] of YddV for the first time. Our data indicate that the YddV reaction is the rate-determining step for two consecutive reactions coupled with phosphodiesterase Ec DOS activity on cyclic di-GMP (c-di-GMP) [turnover number of Ec DOS-Fe(II)-O(2), 61 min(-1)]. Thus, O(2) binding and the heme redox switch of YddV appear to be critical factors in the regulation of c-di-GMP homeostasis. The redox potential and autoxidation rate of heme of the isolated heme domain of YddV (YddV-heme) were determined to be -17 mV versus the standard hydrogen electrode and 0.0076 min(-1), respectively. The Fe(II) complexes of Y43A and Y43L mutant proteins (residues at the heme distal side of the isolated heme-bound globin domain of YddV) exhibited very low O(2) affinities, and thus, their Fe(II)-O(2) complexes were not detected on the spectra. The O(2) dissociation rate constant of the Y43W protein was >150 s(-1), which is significantly larger than that of the wild-type protein (22 s(-1)). The autoxidation rate constants of the Y43F and Y43W mutant proteins were 0.069 and 0.12 min(-1), respectively, which are also markedly higher than that of the wild-type protein. The resonance Raman frequencies representing ν(Fe-O(2)) (559 cm(-1)) of the Fe(II)-O(2) complex and ν(Fe-CO) (505 cm(-1)) of the Fe(II)-CO complex of Y43F differed from those (ν(Fe-O(2)), 565 cm(-1); ν(Fe-CO), 495 cm(-1)) of the wild-type protein, suggesting that Tyr43 forms hydrogen bonds with both O(2) and CO molecules. On the basis of the results, we suggest that Tyr43 located at the heme distal side is important for the O(2) recognition and stability of the Fe(II)-O(2) complex, because the hydroxyl group of the residue appears to interact electrostatically with the O(2) molecule bound to the Fe(II) complex in YddV. Our findings clearly support a role of Tyr in oxygen sensing, and thus modulation of overall conversion from GTP to pGpG via c-di-GMP catalyzed by YddV and Ec DOS, which may be applicable to other globin-coupled oxygen sensor enzymes.     

Characterization of a mutation that results in independence of oxidosqualene cyclase (Erg7) activity from the downstream 3-ketoreductase (Erg27) in the yeast ergosterol biosynthetic pathway

In yeast, deletion of ERG27, which encodes the sterol biosynthetic enzyme, 3-keto-reductase, results in a concomitant loss of the upstream enzyme, Erg7p, an oxidosqualene cyclase (OSC). However, this phenomenon occurs only in fungi, as mammalian Erg27p orthologues are unable to rescue yeast Erg7p activity. In this study, an erg27 mutant containing the mouse ERG27 orthologue was isolated that was capable of growing without sterol supplementation (FGerg27). GC/MS analysis of this strain showed an accumulation of squalene epoxides, 3-ketosterones, and ergosterol. This strain which was crossed to a wildtype and daughter segregants showed an accumulation of squalene epoxides as well as ergosterol indicating that the mutation entailed a leaky block at ERG7. Upon sequencing the yeast ERG7 gene an A598S alteration was found in a conserved alpha helical region. We theorize that this mutation stabilizes Erg7p in a conformation that mimics Erg27p binding. This mutation, while decreasing OSC activity still retains sufficient residual OSC activity such that the strain in the presence of the mammalian 3-keto reductase enzyme functions and no longer requires the yeast Erg27p. Because sterol biosynthesis occurs in the ER, a fusion protein was synthesized combining Erg7p and Erg28p, a resident ER protein and scaffold of the C-4 demethyation complex. Both FGerg27 and erg27 strains containing this fusion plasmid and the mouse ERG27 orthologue showed restoration of ergosterol biosynthesis with minimal accumulation of squalene epoxides. These results indicate retention of Erg7p in the ER increases its activity and suggest a novel method of regulation of ergosterol biosynthesis.     

Genetic analyses involving interactions between the ergosterol biosynthetic enzymes, lanosterol synthase (Erg7p) and 3-ketoreductase (Erg27p), in the yeast Saccharomyces cerevisiae

Protein-protein interaction studies in the Saccharomyces cerevisiae ergosterol biosynthetic pathway suggest that enzymes in this pathway may act as an integrated multienzyme complex. The yeast sterol 3-ketoreductase (Erg27p) required for C-4 demethylation of sterols has previously been shown to also be required for the function of the upstream oxidosqualene cyclase/lanosterol synthase (Erg7p); thus, erg27 mutants accumulate oxidosqualenes as precursors rather than 3-ketosterones. In the present study, we have created various mutations in the ERG27 gene. These mutations include 5 C-terminal truncations, 6 internal deletions, and 32 point mutants of which 14 were obtained by site-directed mutagenesis and 18 by random mutagenesis. We have characterized these ERG27 mutations by determining the following: Erg27 and Erg7 enzyme activities, presence of Erg27p as determined by western immunoblots, ability to grow on various sterol substrates and GC sterol profiles. Mutations of the predicted catalytic residues, Y202F and K206A, resulted in the endogenous accumulation of 3-ketosterones rather than oxidosqualenes suggesting retention of Erg7 enzyme activity. This novel phenotype demonstrated that the catalytic function of Erg27p can be separated from its Erg7p chaperone ability. Other erg27 mutations resulted in proteins that were present, as determined by western immunoblotting, but unable to interact with the Erg7 protein. We also classify Erg27p as belonging to the SDR (short-chain dehydrogenase/reductase) family of enzymes and demonstrate the possibility of homo- or heterodimerization of the protein. This study provides new insights into the role of Erg27p in sterol biosynthesis.     

Critical roles of Leu99 and Leu115 at the heme distal side in auto-oxidation and the redox potential of a heme-regulated phosphodiesterase from Escherichia coli

The heme-regulated phosphodiesterase from Escherichia coli (Ec DOS), which is a heme redox-dependent enzyme, is active with a ferrous heme but inactive with a ferric heme. Global structural changes including axial ligand switching and a change in the rigidity of the FG loop accompanying the heme redox change may be related to the dependence of Ec DOS activity on the redox state. Axial ligands such as CO, NO, and O2 act as inhibitors of Ec DOS because they interact with the ferrous heme complex. The X-ray crystal structure of the isolated heme-bound domain (Ec DosH) shows that Leu99, Phe113 and Leu115 indirectly and directly form a hydrophobic triad on the heme plane and that they should be located at or near the ligand access channel of the heme iron. We generated L99T, L99F, L115T, and L115F mutants of Ec DosH and examined their physicochemical characteristics, including auto-oxidation rates, O2 and CO binding kinetics, and redox potentials. The Fe(III) complex of the L115F mutant was unstable and had a Soret absorption spectrum located 5 nm lower than those of the wild-type and other mutants. Auto-oxidation rates of the mutants (0.049-0.33 min(-1)) were much higher than that of the wild-type (0.0063 min(-1)). Furthermore, the redox potentials of the former three mutants (23.1-34.6 mV versus SHE) were also significantly lower than that of the wild-type (63.9 mV versus SHE). Interaction between O2 and the L99F mutant was different from that in the wild-type, whereas CO binding rates of the mutants were similar to those of the wild-type. Thus, it appears that Leu99 and Leu115 are critical for determining the characteristics of heme iron. Finally, we discuss the roles of these amino-acid residues in the heme electronic states.     

Crystal structures of Parasponia and Trema hemoglobins: differential heme coordination is linked to quaternary structure

Hemoglobins from the plants Parasponia andersonii (ParaHb) and Trema tomentosa (TremaHb) are 93% identical in primary structure but differ in oxygen binding constants in accordance with their distinct physiological functions. Additionally, these proteins are dimeric, and ParaHb exhibits the unusual property of having different heme redox potentials for each subunit. To investigate how these hemoglobins could differ in function despite their shared sequence identity and to determine the cause of subunit heterogeneity in ParaHb, we have measured their crystal structures in the ferric oxidation state. Furthermore, we have made a monomeric ParaHb mutant protein (I43N) and measured its ferrous/ferric heme redox potential to test the hypothesized link between quaternary structure and heme heterogeneity in wild-type ParaHb. Our results demonstrate that TremaHb is a symmetric dimeric hemoglobin similar to other class 1 nonsymbiotic plant hemoglobins but that ParaHb has structurally distinct heme coordination in each of its two subunits that is absent in the monomeric I43N mutant protein. A mechanism for achieving structural heterogeneity in ParaHb in which the Ile(101(F4)) side chain contacts the proximal His(105(F8)) in one subunit but not the other is proposed. These results are discussed in the context of the evolution of plant oxygen transport hemoglobins, and other potential functions of plant hemoglobins.     

Structural and spectroscopic analysis of the F393H mutant of flavocytochrome P450 BM3.

In the preceding paper in this issue [Ost, T. W. B., Miles, C. S., Munro, A. W., Murdoch, J., Reid, G. A., and Chapman, S. K. (2001) Biochemistry 40, 13421-13429], we have established that the primary role of the phylogenetically conserved phenylalanine in flavocytochrome P450 BM3 (F393) is to control the thermodynamic properties of the heme iron, so as to optimize electron-transfer both to the iron (from the flavin redox partner) and onto molecular oxygen. In this paper, we report a detailed study of the F393H mutant enzyme, designed to probe the structural, spectroscopic, and metabolic profile of the enzyme in an attempt to identify the factors responsible for causing the changes. The heme domain structure of the F393H mutant has been solved to 2.0 A resolution and demonstrates that the histidine replaces the phenylalanine in almost exactly the same conformation. A solvent water molecule is hydrogen bonded to the histidine, but there appears to be little other gross alteration in the environment of the heme. The F393H mutant displays an identical ferric EPR spectrum to wild-type, implying that the degree of splitting of the iron d orbitals is unaffected by the substitution, however, the overall energy of the d-orbitals have changed relative to each other. Magnetic CD studies show that the near-IR transition, diagnostic of heme ligation state, is red-shifted by 40 nm in F393H relative to wild-type P450 BM3, probably reflecting alteration in the strength of the iron-cysteinate bond. Studies of the catalytic turnover of fatty acid (myristate) confirms NADPH oxidation is tightly coupled to fatty acid oxidation in F393H, with a product profile very similar to wild-type. The results indicate that gross conformational changes do not account for the perturbations in the electronic features of the P450 BM3 heme system and that the structural environment on the proximal side of the P450 heme must be conformationally conserved in order to optimize catalytic function.     

The FG loop of a heme-based gas sensor enzyme, Ec DOS, functions in heme binding, autoxidation and catalysis

Ec DOS is a heme-based gas sensor enzyme that catalyzes conversion from cyclic-di-GMP to linear-di-GMP in response to gas molecules, such as oxygen, CO and NO. Ec DOS contains an N-terminal heme-binding PAS domain and C-terminal phosphodiesterase domain. Based on crystal structures of the isolated heme-binding domain, it is suggested that the FG loop is involved in intra-molecular signal transduction to the catalytic domain. We generated nine full-length proteins mutated at ionic and non-ionic polar residues between positions 83 and 96 corresponding to the F-helix and FG loop, and examined the heme binding properties, autoxidation rates, and catalytic activities of mutant proteins. N84A and R85A mutant proteins displayed lower heme binding affinities, consistent with the finding that Asn84 interacts with propionate of protoporphyrin IX, and Arg85 with Asp40 on the heme proximal side. Autoxidation rates (0.058-0.54 min⁻¹) of R91A, S96A and K89A/R91A/E93A mutant proteins were significantly higher than that (0.0053 min⁻¹) of wild-type protein, suggesting that these residues in the FG loop form heme distal architecture conferring stability to the Fe(II)-O₂ complex. Catalytic activities of N84A and R85A mutant proteins with low heme affinity were significantly higher than those of wild-type protein in the absence of gas molecules. Accordingly, we propose that loss of heme binding enhances basal catalysis without the gas molecule, consistent with previous reports on heme inhibition of Ec DOS catalysis.     

Effects of heme ligand mutations including a pathogenic variant, H65R, on the properties of human cystathionine beta-synthase

Human cystathionine beta-synthase is a hemeprotein that catalyzes a pyridoxal phosphate (PLP)-dependent condensation of serine and homocysteine into cystathionine. Biophysical characterization of this enzyme has led to the assignment of the heme ligands as histidine and cysteinate, respectively, which has recently been confirmed by crystal structure determination of the catalytic core of the protein. Using site-directed mutagenesis, we confirm that C52 and H65 represent the thiolate and histidine ligands to the heme. Conversion of C52 to alanine or serine results in spectral properties of the resulting hemeprotein that are consistent with the loss of a thiolate ligand. Thus, the Soret peak blue-shifts from 428 to 415 and 417 nm in the ferric forms of the C52S and C52A mutants, respectively, and from 450 to 423 nm in the ferrous states of both mutants. Addition of CO to the dithionite-reduced ferrous C52 mutants results in spectra with Soret peaks at 420 nm. EPR spectroscopy of the ferric C52 variants reveals the predominance of a high-spin species. The H65R mutant, a variant described in a homocystinuric patient, has Soret peaks at 424, 421, and 420 nm in the ferric, ferrous, and ferrous CO states, respectively. EPR spectroscopy reveals predominance of the low-spin species. Both C52A and C52S mutations lead to protein with substoichiometric heme (19% with respect to wild type); however, the PLP content is comparable to that of wild-type enzyme. The heme and PLP contents of the H65R mutant are 40% and 75% that of wild-type enzyme. These results indicate that heme saturation does not dictate PLP saturation in these mutant enzymes. Both H65 and C52 variants display low catalytic activity, revealing that changes in the heme binding domain modulate activity, consistent with a regulatory role for this cofactor.