MAPK/ERK2 phosphorylates ERG at serine 283 in leukemic cells and promotes stem cell signatures and cell proliferation

Aberrant ERG (v-ets avian erythroblastosis virus E26 oncogene homolog) expression drives leukemic transformation in mice and high expression is associated with poor patient outcomes in acute myeloid leukemia (AML) and T-acute lymphoblastic leukemia (T-ALL). Protein phosphorylation regulates the activity of many ETS factors but little is known about ERG in leukemic cells. To characterize ERG phosphorylation in leukemic cells, we applied liquid chromatography coupled tandem mass spectrometry and identified five phosphorylated serines on endogenous ERG in T-ALL and AML cells. S283 was distinct as it was abundantly phosphorylated in leukemic cells but not in healthy hematopoietic stem and progenitor cells (HSPCs). Overexpression of a phosphoactive mutant (S283D) increased expansion and clonogenicity of primary HSPCs over and above wild-type ERG. Using a custom antibody, we screened a panel of primary leukemic xenografts and showed that ERG S283 phosphorylation was mediated by mitogen-activated protein kinase/extracellular signal-regulated kinase (MAPK/ERK) signaling and in turn regulated expression of components of this pathway. S283 phosphorylation facilitates ERG enrichment and transactivation at the ERG +85 HSPC enhancer that is active in AML and T-ALL with poor prognosis. Taken together, we have identified a specific post-translational modification in leukemic cells that promotes progenitor proliferation and is a potential target to modulate ERG-driven transcriptional programs in leukemia.

Allosteric activation of protein phosphatase 2C by D-chiro-inositol-galactosamine, a putative mediator mimetic of insulin action

Insulin-stimulated glucose disposal in skeletal muscle proceeds predominantly through a nonoxidative pathway with glycogen synthase as a rate-limiting enzyme, yet the mechanisms for insulin activation of glycogen synthase are not understood despite years of investigation. Isolation of putative insulin second messengers from beef liver yielded a pseudo-disaccharide consisting of pinitol (3-O-methyl-d-chiro-inositol) beta-1,4 linked to galactosamine chelated with Mn(2+) (called INS2). Here we show that chemically synthesized INS2 has biological activity that significantly enhances insulin reduction of hyperglycemia in streptozotocin diabetic rats. We used computer modeling to dock INS2 onto the known three-dimensional crystal structure of protein phosphatase 2C (PP2C). Modeling and FlexX/CScore energy minimization predicted a unique favorable site on PP2C for INS2 in a surface cleft adjacent to the catalytic center. Binding of INS2 is predicted to involve formation of multiple H-bonds, including one with residue Asp163. Wild-type PP2C activity assayed with a phosphopeptide substrate was potently stimulated in a dose-dependent manner by INS2. In contrast, the D163A mutant of PP2C was not activated by INS2. The D163A mutant and wild-type PP2C in the absence of INS2 had the same Mn(2+)-dependent phosphatase activity with p-nitrophenyl phosphate as a substrate, showing that this mutation did not disrupt the catalytic site. We propose that INS2 allosterically activates PP2C, fulfilling the role of a putative mediator mimetic of insulin signaling to promote protein dephosphorylation and metabolic responses.

Membrane structure and fusion-triggering conformational change of the fusion domain from influenza hemagglutinin

The N-terminal domain of the influenza hemagglutinin (HA) is the only portion of the molecule that inserts deeply into membranes of infected cells to mediate the viral and the host cell membrane fusion. This domain constitutes an autonomous folding unit in the membrane, causes hemolysis of red blood cells and catalyzes lipid exchange between juxtaposed membranes in a pH-dependent manner. Combining NMR structures determined at pHs 7.4 and 5 with EPR distance constraints, we have deduced the structures of the N-terminal domain of HA in the lipid bilayer. At both pHs, the domain is a kinked, predominantly helical amphipathic structure. At the fusogenic pH 5, however, the domain has a sharper bend, an additional 3(10)-helix and a twist, resulting in the repositioning of Glu 15 and Asp 19 relative to that at the nonfusogenic pH 7.4. Rotation of these charged residues out of the membrane plane creates a hydrophobic pocket that allows a deeper insertion of the fusion domain into the core of the lipid bilayer. Such an insertion mode could perturb lipid packing and facilitate lipid mixing between juxtaposed membranes.

An optimized PCR-based procedure for production of 13C/15N-labeled DNA

We have substantially improved a procedure that we previously described for producing 13C/15N-labeled DNA (Chen et al., FEBS Lett. 436, 372-376, 1998) to provide an economical and straightforward approach to the preparation of labeled DNA. The conditions for the PCR reactions have been optimized to permit the use of low concentrations of the costly labeled dNTPs (50 microM for each). In addition, a rapid and high-yield purification procedure has been developed that allows us to obtain a high yield of very pure labeled DNA. These modifications to our original procedure permit us to obtain 1.9 mg of an 18 bp DNA oligomer from 20 mg of dNTPs (ca. 10% yield from the starting dNTPs). This is sufficient material for the preparation of 0.4 mM sample in a volume of 400 microl. In summary, this procedure is a cost-effective, time-efficient procedure for the production of labeled DNA for NMR studies.

The NOESY jigsaw: automated protein secondary structure and main-chain assignment from sparse, unassigned NMR data

High-throughput, data-directed computational protocols for Structural Genomics (or Proteomics) are required in order to evaluate the protein products of genes for structure and function at rates comparable to current gene-sequencing technology. This paper presents the JIGSAW algorithm, a novel high-throughput, automated approach to protein structure characterization with nuclear magnetic resonance (NMR). JIGSAW applies graph algorithms and probabilistic reasoning techniques, enforcing first-principles consistency rules in order to overcome a 5-10% signal-to-noise ratio. It consists of two main components: (1) graph-based secondary structure pattern identification in unassigned heteronuclear NMR data, and (2) assignment of spectral peaks by probabilistic alignment of identified secondary structure elements against the primary sequence. Deferring assignment eliminates the bottleneck faced by traditional approaches, which begin by correlating peaks among dozens of experiments. JIGSAW utilizes only four experiments, none of which requires 13C-labeled protein, thus dramatically reducing both the amount and expense of wet lab molecular biology and the total spectrometer time. Results for three test proteins demonstrate that JIGSAW correctly identifies 79-100% of alpha-helical and 46-65% of beta-sheet NOE connectivities and correctly aligns 33-100% of secondary structure elements. JIGSAW is very fast, running in minutes on a Pentium-class Linux workstation. This approach yields quick and reasonably accurate (as opposed to the traditional slow and extremely accurate) structure calculations. It could be useful for quick structural assays to speed data to the biologist early in an investigation and could in principle be applied in an automation-like fashion to a large fraction of the proteome.

Structure of outer membrane protein A transmembrane domain by NMR spectroscopy

We have determined the three-dimensional fold of the 19 kDa (177 residues) transmembrane domain of the outer membrane protein A of Escherichia coli in dodecylphosphocholine (DPC) micelles in solution using heteronuclear NMR. The structure consists of an eight-stranded beta-barrel connected by tight turns on the periplasmic side and larger mobile loops on the extracellular side. The solution structure of the barrel in DPC micelles is similar to that in n-octyltetraoxyethylene (C(8)E(4)) micelles determined by X-ray diffraction. Moreover, data from NMR dynamic experiments reveal a gradient of conformational flexibility in the structure that may contribute to the membrane channel function of this protein.

Energetic and functional contribution of residues in the core binding factor beta (CBFbeta ) subunit to heterodimerization with CBFalpha

Core-binding factors (CBFs) are a small family of heterodimeric transcription factors that play critical roles in several developmental pathways, including hematopoiesis and bone development. Mutations in CBF genes are found in leukemias and bone disorders. CBFs consist of a DNA-binding CBFalpha subunit (Runx1, Runx2, or Runx3) and a non-DNA-binding CBFbeta subunit. CBFalpha binds DNA in a sequence-specific manner, whereas CBFbeta enhances DNA binding by CBFalpha. Recent structural analyses of the DNA-binding Runt domain of CBFalpha and the CBFbeta subunit identified the heterodimerization surfaces on each subunit. Here we identify amino acids in CBFbeta that mediate binding to CBFalpha. We determine the energy contributed by each of these amino acids to heterodimerization and the importance of these residues for in vivo function. These data refine the structural analyses and further support the hypothesis that CBFbeta enhances DNA binding by inducing a conformational change in the Runt domain.

CBF--a biophysical perspective

Core binding factor (CBF) is a heterodimeric transcription factor consisting of a DNA-binding subunit (Runx, also referred to as CBFA, AML 1, PEBP2alpha) and a non-DNA-binding subunit (CBFB). Biophysical characterization of the two proteins (and their interactions is providing a detailed understanding of this important transcription factor at the molecular level. Measurements of the relevant binding constants are helping to elucidate the mechanism of leukemogenesis associated with altered forms of these proteins. Determination of the 3D structures of CBFB and the DNA- and CBFB-binding domain of Runx, referred to as the Runt domain, are providing a structural basis for the functioning of the two proteins of CBF.

Biophysical characterization of interactions between the core binding factor alpha and beta subunits and DNA

Core binding factors (CBFs) play key roles in several developmental pathways and in human disease. CBFs consist of a DNA binding CBFalpha subunit and a non-DNA binding CBFbeta subunit that increases the affinity of CBFalpha for DNA. We performed sedimentation equilibrium analyses to unequivocally establish the stoichiometry of the CBFalpha:beta:DNA complex. Dissociation constants for all four equilibria involving the CBFalpha Runt domain, CBFbeta, and DNA were defined. Conformational changes associated with interactions between CBFalpha, CBFbeta, and DNA were monitored by nuclear magnetic resonance and circular dichroism spectroscopy. The data suggest that CBFbeta 'locks in' a high affinity DNA binding conformation of the CBFalpha Runt domain.

The Ig fold of the core binding factor alpha Runt domain is a member of a family of structurally and functionally related Ig-fold DNA-binding domains

BACKGROUND

CBFA is the DNA-binding subunit of the transcription factor complex called core binding factor, or CBF. Knockout of the Cbfa2 gene in mice leads to embryonic lethality and a profound block in hematopoietic development. Chromosomal disruptions of the human CBFA gene are associated with a large percentage of human leukemias.

RESULTS

Utilizing nuclear magnetic resonance spectroscopy we have determined the three-dimensional fold of the CBFA Runt domain in its DNA-bound state, showing that it is an s-type immunoglobulin (Ig) fold. DNA binding by the Runt domain is shown to be mediated by loop regions located at both ends of the Runt domain Ig fold. A putative site for CBFB binding has been identified; the spatial location of this site provides a rationale for the ability of CBFB to modulate the affinity of the Runt domain for DNA.

CONCLUSIONS

Structural comparisons demonstrate that the s-type Ig fold found in the Runt domain is conserved in the Ig folds found in the DNA-binding domains of NF-kappaB, NFAT, p53, STAT-1, and the T-domain. Thus, these proteins form a family of structurally and functionally related DNA-binding domains. Unlike the other members of this family, the Runt domain utilizes loops at both ends of the Ig fold for DNA recognition.

Haploinsufficiency of CBFA2 causes familial thrombocytopenia with propensity to develop acute myelogenous leukaemia

Familial platelet disorder with predisposition to acute myelogenous leukaemia (FPD/AML, MIM 601399) is an autosomal dominant disorder characterized by qualitative and quantitative platelet defects, and propensity to develop acute myelogenous leukaemia (AML). Informative recombination events in 6 FPD/AML pedigrees with evidence of linkage to markers on chromosome 21q identified an 880-kb interval containing the disease gene. Mutational analysis of regional candidate genes showed nonsense mutations or intragenic deletion of one allele of the haematopoietic transcription factor CBFA2 (formerly AML1) that co-segregated with the disease in four FPD/AML pedigrees. We identified heterozygous CBFA2 missense mutations that co-segregated with the disease in the remaining two FPD/AML pedigrees at phylogenetically conserved amino acids R166 and R201, respectively. Analysis of bone marrow or peripheral blood cells from affected FPD/AML individuals showed a decrement in megakaryocyte colony formation, demonstrating that CBFA2 dosage affects megakaryopoiesis. Our findings support a model for FPD/AML in which haploinsufficiency of CBFA2 causes an autosomal dominant congenital platelet defect and predisposes to the acquisition of additional mutations that cause leukaemia.

Binding specificity and mechanistic insight into glutaredoxin-catalyzed protein disulfide reduction

The reduction equivalents necessary for the ribonucleotide reductase (RNR)-catalyzed production of deoxyribonucleotides are provided by glutaredoxin (Grx) or thioredoxin (Trx). The initial location for transfer of reducing equivalents to RNR is located at the C terminus of the B1 subunit and involves the reduction of a disulfide between Cys754 and Cys759. We have used a 25-mer peptide corresponding to residues 737-761 of RNR B1 (C754-->S) to synthesize a stable mixed disulfide with Escherichia coli Grx-1 (C14-->S) resembling the structure of an intermediate in the reaction. The high-resolution solution structure of the mixed disulfide has been obtained by NMR with an RMSD of 0.56 A for all the backbone atoms of the protein and the well-defined portion of the peptide. The binding interactions responsible for specificity have been identified demonstrating the importance of electrostatic interactions in this system and providing a rationale for the specificity of the Grx-RNR interaction. The disulfide is buried in this complex, implying a solely intra-molecular mechanism of reduction in contrast to the previously determined structure of the glutathione complex where the disulfide was exposed; mutagenesis studies have shown the relevance of intermolecular reduction processes. Substantial conformational changes in the helices of the protein are associated with peptide binding which have significant mechanistic implications for protein disulfide reduction by glutaredoxins.

Solution structure of core binding factor beta and map of the CBF alpha binding site

The core binding factor beta subunit (CBF beta) is the non-DNA binding subunit of the core-binding factors, transcription factors essential for multiple developmental processes including hematopoiesis and bone development. Chromosomal translocations involving the human CBFB gene are associated with a large percentage of human leukemias. The N-terminal 141 amino acids of CBF beta contains the heterodimerization domain for the DNA-binding CBF alpha subunits, and is sufficient for CBF beta function in vivo. Here we present the high-resolution solution structure of the CBF beta heterodimerization domain. It is a novel alpha/beta structure consisting of two three-stranded beta-sheets packed on one another in a sandwich arrangement, with four peripheral alpha-helices. The CBF alpha binding site on CBF beta has been mapped by chemical shift perturbation analysis.

A PCR-based method for uniform 13C/15N labeling of long DNA oligomers

A polymerase chain reaction (PCR)-based method is described for uniform 13C/15N labeling of DNA duplexes. In this method, multiple copies of a blunt-ended duplex are cloned into a plasmid with each copy containing the sequence of interest and the restriction HincII sequences at the 5' and 3' ends. PCR with uniformly 13C/15N-labeled dNTP precursors results in a labeled DNA duplex containing multiple copies of the sequence of interest. Use of bi-directional primers, instead of self-priming [Louis et al. (1998) J. Biol. Chem. 273, 2374-2378], produces a DNA fragment of unique length. Twenty-four cycles of PCR of this purified product followed by restriction and purification gives (with 30% yield) the uniformly 13C/15N-labeled duplex sequence for multi-nuclear magnetic resonance spectroscopy.

The NMR solution structure of human glutaredoxin in the fully reduced form

The determination of the nuclear magnetic resonance (NMR) solution structure of fully reduced human glutaredoxin is described. A total of 1159 useful nuclear Overhauser effect (NOE) upper distance constraints and 187 dihedral angle constraints were obtained as the input for the structure calculations for which the torsion angle dynamics program DYANA has been utilized followed by energy minimization in water with the AMBER force field as implemented in the program OPAL. The resulting 20 conformers have an average root-mean-square deviation value relative to the mean coordinates of 0.54 A for all the backbone atoms N, Calpha and C', and of 1.01 A for all heavy atoms. Human glutaredoxin consists of a four-stranded mixed beta-sheet composed of residues 15 to 19, 43 to 47, 72 to 75 and 78 to 81, and five alpha-helices composed of residues 4 to 9, 24 to 34, 54 to 65, 83 to 91, and 94 to 100. Comparisons with the structures of Escherichia coli glutaredoxin-1, pig liver glutaredoxin and human thioredoxin were made. Electrostatic calculations on the human glutaredoxin structure and that of related proteins provide an understanding of the variation of pKa values for the nucleophilic cysteine in the active site observed among these proteins. In addition, the high-resolution NMR solution structure of human glutaredoxin has been used to model the binding site for glutathione and for ribonucleotide reductase B1 by molecular dynamics simulations.

Preparation, characterization, and complete heteronuclear NMR resonance assignments of the glutaredoxin (C14S)-ribonucleotide reductase B1 737-761 (C754S) mixed disulfide

The first committed step in de novo DNA biosynthesis involves the conversion of ribonucleotides to the corresponding deoxyribonucleotides catalyzed by the enzyme ribonucleotide reductase. Reduction of disulfides in ribonucleotide reductase is essential and is catalyzed by the protein disulfide reductants glutaredoxin or thioredoxin. The interaction region between Escherichia coli glutaredoxin-1 and E. coli ribonucleotide reductase has been localized to the C-terminal end of the B1 subunit of ribonucleotide reductase. We have demonstrated that a 25-residue peptide corresponding to this C-terminal sequence is a very good substrate for glutaredoxin via a fluorescence assay and that this peptide binds in a specific manner via isothermal titration calorimetric measurements. By selectively mutating the two cysteines in the peptide, we have identified the electrophilic cysteine as C759 (B1 numbering) and prepared a mixed disulfide between E. coli glutaredoxin-1 (C14 --> S) and the C759 monothiol form of the peptide. The peptide and the protein have been labeled with 13C and 15N, and complete heteronuclear NMR resonance assignments have been completed for both the peptide and the protein in the complex. By using half-filtered NOESY spectra, intermolecular NOEs between the protein and the peptide have been identified and the binding site on glutaredoxin has been mapped. The electrostatic charge distribution of the protein in this region is very positive, thus providing an excellent match for the highly negatively charged peptide. In addition, the electrostatic potential of the peptide provides a rationale for the observed cysteine selectivity in the reaction between glutaredoxin and the B1 peptide.

Overexpression, purification, and biophysical characterization of the heterodimerization domain of the core-binding factor beta subunit

Core-binding factors (CBF) are heteromeric transcription factors essential for several developmental processes, including hematopoiesis. CBFs contain a DNA-binding CBF alpha subunit and a non-DNA binding CBF beta subunit that increases the affinity of CBF alpha for DNA. We have developed a procedure for overexpressing and purifying full-length CBF beta as well as a truncated form containing the N-terminal 141 amino acids using a novel glutaredoxin fusion expression system. Substantial quantities of the CBF beta proteins can be produced in this manner allowing for their biophysical characterization. We show that the full-length and truncated forms of CBF beta bind to a CBF alpha DNA complex with very similar affinities. Sedimentation equilibrium measurements show these proteins to be monomeric. Circular dichroism spectroscopy demonstrates that CBF beta is a mixed alpha/beta protein and NMR spectroscopy shows that the truncated and full-length proteins are structurally similar and suitable for structure determination by NMR spectroscopy.

Comparison of backbone dynamics of reduced and oxidized Escherichia coli glutaredoxin-1 using 15N NMR relaxation measurements

NMR-based structure determination of Escherichia coli glutaredoxin-1 in its reduced and oxidized forms revealed only subtle structural differences between the two forms. In an effort to characterize the role dynamics may play in the functioning of the protein, the backbone dynamics of both the reduced and oxidized forms of E. coli glutaredoxin-1 have been characterized using inverse-detection two-dimensional 15N-1H NMR spectroscopy. Longitudinal (T1) and transverse (T2) 15N relaxation time constants and steady-state [1H]-15N NOEs were measured for a majority of the protonated backbone nitrogen atoms. These data were analyzed by using a model-free formalism to determine the generalized order parameter (S2), the effective correlation time for internal motions (tau(e)), 15N exchange broadening contributions (R(ex)), and the overall molecular rotational correlation time (tau(m)). Sedimentation equilibrium measurements showed the reduced protein to be monomeric whereas the oxidized form could be fit to a monomer-dimer equilibrium. In order to try and assess the effect of dimerization on the dynamical parameters, the measurements on the oxidized protein have been carried out at two concentrations with very different monomer/dimer ratios. There is increased motion on both nano-picosecond and micro-millisecond time scales in the reduced form relative to the oxidized form, consistent with a more rigid oxidized protein. The increase in motion in the reduced protein correlates with its decreased thermodynamic stability. The role of the observed differences in the dynamic behavior in the two forms, particularly in the active site, in glutaredoxin-1's role as a protein disulfide reductant is discussed.

Complete 1H, 13C, and 15N NMR resonance assignments and secondary structure of human glutaredoxin in the fully reduced form

Human glutaredoxin is a member of the glutaredoxin family, which is characterized by a glutathione binding site and a redox-active dithiol/disulfide in the active site. Unlike Escherichia coli glutaredoxin-1, this protein has additional cysteine residues that have been suggested to play a regulatory role in its activity. Human glutaredoxin (106 amino acid residues, M(r) = 12,000) has been purified from a pET expression vector with both uniform 15N labeling and 13C/15N double labeling. The combination of three-dimensional 15N-edited TOCSY, 15N-edited NOESY, HNCA, HN(CO)CA, and gradient sensitivity-enhanced HNCACB and HNCO spectra were used to obtain sequential assignments for residues 2-106 of the protein. The gradient-enhanced version of the HCCH-TOCSY pulse sequence and HCCH-COSY were used to obtain side chain 1H and 13C assignments. The secondary structural elements in the reduced protein were identified based on NOE information, amide proton exchange data, and chemical shift index data. Human glutaredoxin contains five helices extending approximately from residues 4-10, 24-36, 53-64, 83-92, and 94-104. The secondary structure also shows four beta-strands comprised of residues 15-19, 43-48, 71-75, 78-80, which form a beta-sheet almost identical to that found in E. coli glutaredoxin-1. Complete 1H, 13C, and 15N assignments and the secondary structure of fully reduced human glutaredoxin are presented. Comparison to the structures of other glutaredoxins is presented and differences in the secondary structure elements are discussed.

The CBFbeta subunit is essential for CBFalpha2 (AML1) function in vivo

The CBFbeta subunit is the non-DNA-binding subunit of the heterodimeric core-binding factor (CBF). CBFbeta associates with DNA-binding CBFalpha subunits and increases their affinity for DNA. Genes encoding the CBFbeta subunit (CBFB) and one of the CBFalpha subunits (CBFA2, otherwise known as AML1) are the most frequent targets of chromosomal translocations in acute leukemias in humans. We and others previously demonstrated that homozygous disruption of the mouse Cbfa2 (AML1) gene results in embryonic lethality at midgestation due to hemorrhaging in the central nervous system and blocks fetal liver hematopoiesis. Here we demonstrate that homozygous mutation of the Cbfb gene results in the same phenotype. Our results demonstrate that the CBFbeta subunit is required for CBFalpha2 function in vivo.

Biochemical and biophysical properties of the core-binding factor alpha2 (AML1) DNA-binding domain

The Runt domain is the DNA-binding domain defining a small family of transcription factors that are involved in important developmental processes. Developmental pathways controlled by Runt domain proteins include sex determination, neurogenesis, segmentation, and eye development in Drosophila and hematopoiesis in mammals. In addition to binding DNA, the Runt domain also mediates heterodimerization with another subunit called the core-binding factor beta (CBFbeta) subunit. In this study we overexpress the Runt domain from the mouse CBFalpha2 (AML1) protein in Escherichia coli, and purify it from the insoluble fraction. We determine the equilibrium constants for Runt domain binding to two different DNA sequences by surface plasmon resonance technology. Circular dichroism spectroscopy demonstrates that the Runt domain is a folded beta-domain with essentially no alpha-helical content. The single tryptophan residue in the CBFalpha2 Runt domain at amino acid 79 is shown by tryptophan fluorescence spectroscopy to reside in a polar environment. Finally, we demonstrate that ATP can be UV cross-linked to the Runt domain and that ATP binding is sensitive to an amino acid substitution in the putative Kinase-1a motif (P-loop).

The nuclear magnetic resonance solution structure of the mixed disulfide between Escherichia coli glutaredoxin(C14S) and glutathione

The determination of the nuclear magnetic resonance (NMR) solution structured of the mixed disulfide between the mutant Escherichia coli glutaredoxin Grx(C14S) and glutathione (GSH), Grx(C14S)-SG, is described, the binding site for GSH on Grx(C14S) is located, and the non-bonding interactions between -SG and the protein are characterized. Based on nearly complete sequence-specific NMR assignments, 1010 nuclear Overhauser enhancement upper distance constraints and 116 dihedral angle constraints were obtained as the input for the structure calculations, for which the distance geometry program DIANA was used followed by energy minimization in a waterbath with the AMBER force field in the program OPAL. The -SG moiety was found to be localized on the surface of the protein in a cleft bounded by the amino acid residues Y13, T58, V59, Y72, T73 and D74. Hydrogen bonds have been identified between -SG and the residues V59 and T73 of Grx(C14S), and the formation of an additional hydrogen bond with Y72 and electrostatic interactions with the side-chains of D74 and K45 are also compatible with the NMR conformational constraints. Comparison of the reduced and oxidized forms of Grx with Grx(C14S)-SG shows that the mixed disulfide more closely resembles the oxidized form of the protein. Functional implications of this observation are discussed. Comparisons are also made with the related proteins bacteriophage T4 glutaredoxin and glutathione S-transferase.

Biosynthetic 15N and 13C isotope labelling of glutathione in the mixed disulfide with Escherichia coli glutaredoxin documented by sequence-specific NMR assignments

A biosynthetic procedure for obtaining 13C-15N doubly labelled glutathione from readily available precursor molecules is described. Isolation of the mutant Escherichia coli [C14S]glutaredoxin from E. coli cultures grown on 15N-13C doubly labelled media in the absence of reducing agents yields the mixed disulfide labelled in both the protein and the glutathione. 15N NMR assignments for glutathione obtained from two-dimensional [15N,1H]-correlation spectroscopy (COSY), and 13C NMR assignments for the entire mixed disulfide obtained from combined use of three-dimensional ct-HA[CAN]HN experiments and HCCH-total correlation spectroscopy ([HCCH]-TOCSY) demonstrated unequivocally that the glutathione is uniformly labelled with both 15N and 13C. This result also supports earlier suggestions that the intracellular glutaredoxin activity is sensitive to the glutathione redox status of the cell. Complete sets of 1H, 13C and 15N chemical shifts of both components in the mixed disulfide of [C14S]glutaredoxin and glutathione were obtained from the sequence-specific NMR assignments.

3D 13C-15N-heteronuclear two-spin coherence spectroscopy for polypeptide backbone assignments in 13C-15N-double-labeled proteins

The pulse sequence of a new constant-time 3D triple-resonance experiment, ct-HA[CAN]HN, is presented. This experiment delineates exclusively scalar connectivities and uses 13C alpha-15N heteronuclear two-spin coherence to overlay the chemical shift evolution periods of the 13C alpha and 15N nuclei, thereby providing the four resonance frequencies of the alpha-proton, the alpha-carbon, the amide nitrogen, and the amide proton of a given amino acid residue in three dimensions. This experiment promises to be a valid alternative to 4D experiments, providing the same information on intraresidue polypeptide backbone connectivities in 13C-15N-double-labeled proteins.

Structural and functional characterization of the mutant Escherichia coli glutaredoxin (C14----S) and its mixed disulfide with glutathione

Glutaredoxin is essential for the glutathione (GSH)-dependent synthesis of deoxyribonucleotides by ribonucleotide reductase, and in addition, it displays a general GSH disulfide oxidoreductase activity. In Escherichia coli glutaredoxin, the active site contains a redox-active disulfide/dithiol of the sequence Cys11-Pro12-Tyr13-Cys14. In this paper, we have prepared and characterized the Cys14----Ser mutant of E. coli glutaredoxin and its mixed disulfide with glutathione. The Cys14----Ser mutant of glutaredoxin is shown to retain 38% of the GSH disulfide oxidoreductase activity of the wild-type protein with hydroxyethyl disulfide as substrate but to be completely inactive with ribonucleotide reductase, demonstrating that dithiol glutaredoxin is the hydrogen donor for ribonucleotide reductase. The covalent structure of the mixed disulfide of glutaredoxin(C14S) with GSH prepared with 15N-labeling of the protein was confirmed with nuclear magnetic resonance (NMR) spectroscopy, establishing a basis for NMR structural studies of the glutathione binding site on glutaredoxin.

NMR structure of oxidized Escherichia coli glutaredoxin: comparison with reduced E. coli glutaredoxin and functionally related proteins

The determination of the NMR structure of oxidized Escherichia coli glutaredoxin in aqueous solution is described, and comparisons of this structure with that of reduced E. coli glutaredoxin and the related proteins E. coli thioredoxin and T4 glutaredoxin are presented. Based on nearly complete sequence-specific 1H-NMR assignments, 804 nuclear Overhauser enhancement distance constraints and 74 dihedral angle constraints were obtained as the input for the structure calculations, for which the distance geometry program DIANA was used followed by simulated annealing with the program X-PLOR. The molecular architecture of oxidized glutaredoxin is made up of three helices and a four-stranded beta-sheet. The three-dimensional structures of oxidized and the recently described reduced glutaredoxin are very similar. Quantitative analysis of the exchange rates of 34 slowly exchanging amide protons from corresponding series of two-dimensional [15N,1H]-correlated spectra of oxidized and reduced glutaredoxin showed close agreement, indicating almost identical hydrogen-bonding patterns. Nonetheless, differences in local dynamics involving residues near the active site and the C-terminal alpha-helix were clearly manifested. Comparison of the structure of E. coli glutaredoxin with those of T4 glutaredoxin and E. coli thioredoxin showed that all three proteins have a similar overall polypeptide fold. An area of the protein surface at the active site containing Arg 8, Cys 11, Pro 12, Tyr 13, Ile 38, Thr 58, Val 59, Pro 60, Gly 71, Tyr 72, and Thr 73 is proposed as a possible site for interaction with other proteins, in particular ribonucleotide reductase. It was found that this area corresponds to previously proposed interaction sites in T4 glutaredoxin and E. coli thioredoxin. The solvent-accessible surface area at the active site of E. coli glutaredoxin showed a general trend to increase upon reduction. Only the sulfhydryl group of Cys 11 is exposed to the solvent, whereas that of Cys 14 is buried and solvent inaccessible.

Sequence-specific 1H n.m.r. assignments and determination of the three-dimensional structure of reduced Escherichia coli glutaredoxin

The determination of the nuclear magnetic resonance structure of reduced E. coli glutaredoxin in aqueous solution is described. Based on nearly complete, sequence-specific resonance assignments, 813 nuclear Overhauser effect distance constraints and 191 dihedral angle constraints were employed as the input for the structure calculations, for which the distance geometry program DIANA was used followed by simulated annealing with the program X-PLOR. The molecular architecture of reduced glutaredoxin is made up of three helices and four-stranded beta-sheet. The first strand of the beta-sheet (residues 2 to 7) runs parallel to the second strand (32 to 37) and antiparallel to the third strand (61 to 64), and the sheet is extended in an antiparallel fashion with a fourth strand (67 to 69). The first helix with residues 13 to 28 and the last helix (71 to 83) run parallel to each other on one side of the beta-sheet, with their direction opposite to that of the two parallel beta-strands, and the helix formed by residues 44 to 53 fills space available due to the twist of the beta-sheet and the reduced length of the last two beta-strands. The active site Cys11-Pro-Tyr-Cys14 is located after the first beta-strand and occupies the latter part of the loop connecting this strand with the first helix.

Investigation of an octapeptide inhibitor of Escherichia coli ribonucleotide reductase by transferred nuclear Overhauser effect spectroscopy

Several peptides contained within the C-terminal sequence of the B2 subunit of Escherichia coli ribonucleotide reductase (RNR) were investigated for their ability to inhibit the enzyme, presumably by interfering with association of the B1 and B2 subunits. AcYLVGQIDSE, corresponding by sequence homology to a nonapeptide that inhibits herpes simplex RNR [Gaudreau et al. (1987) J. Biol. Chem. 262, 12413] shows no inhibition of the E. coli enzyme (IC50 greater than 3 mM), whereas AcDDLSNFQL, the C-terminal octapeptide of the E. coli B2 subunit, is a noncompetitive inhibitor (Ki = 160 microM). Neither bradykinin (RPPGFSPFR) nor the pentapeptide AcSNFQL inhibits the E. coli enzyme. Transferred nuclear Overhauser enhancement spectroscopy was used to probe the conformation of AcDDLSNFQL when it is bound to the B1 subunit. These experiments suggest that the peptide adopts a turn in the region of Asn5 and Phe6 and that a hydrophobic cluster of the phenylalanine and leucine side chains is involved in the interaction surface.