Protein-DNA interaction has been a central theme in studies of transcription regulation. A novel sequence specific DNA-binding motif has been defined by thirteen recently identified DNA-binding proteins including Mrf-2 which do not share sequence homology with previously characterized DNA-binding motifs. Structural studies of the Mrf-2 DNA-binding domain has been carried out using nuclear magnetic resonance methods (NMR) and 15N- and 13C/15N- labeled samples. Three-dimensional structures of the protein have been calculated with more than 1200 NOEs and 71 dihedral angle constraints. Twenty-one hydrogen bonding constraints were introduced based on slowly exchanging amide protons and secondary structures. The structures have been calculated using the program DYANA/v1.4 and have an RMSD. of approximately 1.0 angstron between backbone atoms of residues in the helical regions. More than 70% of the residues are located within the most favored regions of the Ramanchandran plot. The protein contains six helices but no beta-sheet. Information on the flexibility of this protein has been provided by amide proton exchange data.
The UGA codon signals translational termination in the universal genetic code. One way in which termination can be circumvented is to decode UGA for the insertion of selenocysteine. Translational recoding requires four trans-acting factors in Eukaryotes. Of these, one is a stem-loop structure situated within the 3'-UTR of mRNAs that code for selenoproteins. This structure is called the selenocysteine insertion element (SECIS), and has been implicated in protein binding that orchestrates the recoding event. We present NMR data that destinguishes betweeen the two models for the secondary structure of this RNA. The defined stem regester creates a tandem G-A region flanked by two U-U baspairs. We present 2-4D NMR data that will be used for detailed structural characterization of this RNA. Developement of a new 4D 15N/13C-NOESY experiment for RNA, and amino edited-3D NOESY data has enabled assignment of basepairing. The use of (H)CCH-TOCSY and HC(C)H-TOCSY correlations allowed us to use the complete sugar spin system to make sugar-base correlations.
Phospholipase A2 (PLA2, EC 3.1.1.4) are calcium-dependent lipolytic enzymes that catalyzes the hydrolysis of the sn-2 ester bond of membrane phospholipids. The solution structure of bovine pancreatic PL A2 was determined by multidimensional NMR at pH 6.0 and 310 K. It is represented by an ensemble of 20 conformers calculated with X-PLOR using 1601 distance and 108 dihedral angle restraints. The values of root-mean-square deviation are 0.78 and 1.35 angstrom for backbone (N, Ca, C) atoms and all heavy atoms, respectively. The NMR structure is virtually similar to the one obtained by X-ray crystallography, except that alpha-helix B (19-21) is absent and alpha-helix D appears shorter (59-62). The structures in calcium binding loop (25-36) and helix-D location (60-73) are not well defined with bigger RMSD than average. The evidence of hydrogen-bonding network is incomplete. The protein backbone dynamics was investigated by T1, T2 15N-relaxation measurements with modification on the phase cycling in the pulse sequence. The results reveal slow internal motions (microseconds to seconds) for the residues in the interfacial binding site and the active site. These intramolecular motions are probably important for enzyme catalysis.
The 39 kD E. coli protein Ada is a critical regulator of the bacterial adaptive response toward DNA methylation. Ada comprises two independently functioning domains. The 19 kD C-terminal domain repairs the O6-methylguanine and O4-methylthymine lesions by irreversible transfer of the methyl group to Cys 321. The N-terminal domain of Ada repairs the Sp diastereomer of methylphosphotriesters through irreversible transfer of the methyl group to another cysteine residue, Cys 69. This cysteine residue is activated for nucleophilic attack on the methyl group by coordination to a tightly bound zinc. Upon methyl transfer, Ada acquires the ability to bind DNA sequence specifically, and thereby to induce genes that function collectively to protect E. coli from the mutagenic and toxic effects of methylation agents. This sequence specific DNA binding function is located in the second subdomain (amino acids 91-139) of the N-terminal domain of Ada. The first subdomain (amino acids 1-90) is responsible for the methyl transfer reaction. In order to investigate the conformational switch mechanism by which Ada changes its function from a repair protein to a transcription factor we have determined the NMR structure of an N-terminal 16 kD fragment comprising both subdomains in its methylated form in complex with a 18 bp DNA. These investigations have shown, that the two subdomains of N-Ada16 are connected only by a flexible linker in the unmethylated protein. Methylation of Cys 69 in the repair subdomain leads to the build up of an interface between the two subdomains in which the methyl group provides a crucial contact to the second subdomain. The structure of the complex will be presented and the switch mechanism will be discussed.
Cell division is controlled by a number of positive and negative regulators. Failure of these regulators may contribute to cancer by allowing excessive cell proliferation. P16 is a negative regulator which inhibits cyclin-dependent kinases cdk4 and cdk6, thus regulating cell cycle progression from g1 to S phase. Mutations in p16 have been found in > 70 different types of tumor cells to date. Furthermore, accumulating biochemical and genetic data clearly demonstrate p16 to be one of the major tumor suppressors. The structure of the tumor suppressor p16 was determined by NMR spectroscopy. A total of 1370 distance and 58 dihedral angle restraints were used for the simulated annealing calculations using the XPLOR program. The final 16 structures show the average r.m.s. differences from the mean structures of 0.97 A for the backbone atoms, and 1.34 A for all the non-hydrogen atoms, respectively. None of the structures show distance restraint violation bigger than 0.5 A and diheral angle restraints bigger than 5.0 degree. The tertiary structure of p16 is characterized by repeating structural units from the four ankyrins in p16: the central part of each ankyrin forms helix-loop-helix (H-L-H) structure and the two ends which have close contacts form extended conformations. The functionally and structurally important residues are proposed based on the p16 structure and the cdk4 inhibition assay data from both p16 and cdk4 mutants.
Glutaredoxin is an essential enzyme for de novo DNA biosynthesis. Besides its role as a source of reducing equivalents for ribonucleotide reductase, it has been shown to play a number of important roles in cells, including the reduction of dehydroascorbate, sulfate, and arsentate, regeneration of oxidatively damaged proteins and acceleration of protein folding. Human glutaredoxin is a basic protein with 105 amino acids containing five cysteine residues. 3D 15N-edited NOESY and 13C -edited NOESY spectra were recorded to derive NOE constraints. Dihedral angle constraints were derived from 3D HNHA spectra for phi and 3D HACAHB-COSY spectra for chi1. A totally 1160 useful distance constraints and 187 dihedral angle constraints were obtained and used as the input for structure calculations by DYANA. Stereospecific assignments were obtained for betaCH2 groups using program HABAS and for methyl groups with program GLOMSA. Assignments for 62 pairs of diastereotopic substituents were obtained, including 34 of the betaCH2 groups and 8 methyl groups of Val and Leu. The r.m.s.d. of the 20 best conformers relative to the mean coordinates is 0.53 for all the backbone atoms. The overall fold of the structure and analysis of the structure with respect to its functional properties will be presented.
The electronic ground states of pheophytin cofactorst in the bacterial photosynthetic reaction center have been investigated through a characterization of the electronic densities at individual atomic positions of pheophytin a from 13C chemical shift data. A new experimental approach involving multispin 13C labelling and 2-D NMR is presented. Bacterial photosynthetic reaction centers of Rb. Sphaeroides R 26 were reconstituted with uniformly 13C biosynthetically labelled plant Pheo a in the two pheophytin binding sites from the multispin labeled samples. 1-D and 2-D solid-state 13C magic angle spinning NMR spectra could be obtained and used to characterize the pheophytin a ground state in the Rb. sphaeroides R 26 RCs without a necessity for time-consuming selective labeling strategies involving organic synthesis. From the 2-D solid state 13C-13C correlation spectra collected with spinning speeds of 8 and 10 kHz with mixing times of 1 and 0.8 ms, many 13C resonances of the [U-13C] Pheo a molecules reconstituted in the RCs could be assigned in a single set of experiments and used to study the electronic structure, polarization and protonation/hydrogen-bonding state. Parts of the pheophytins interacting with the protein at the level of 13C shifts modified by binding could be identified. Small reconstitution shifts are detected for the 172 side chain of ring IV. In contrast there is no evidence for electrostatic differences between the two Pheo a, for instance, due to a possibly strong selective electrostatic interaction with Glu L104 on the active branch. The protonation states appear the same and the NMR suggests a strong overall simliarity between the ground states of the two Pheo a which is of interest in view of the asymetry of the electron transfer in the photosynthetic RC.
SH2 Domains are key protein modules in cellular regulation. SH2s respond to tyrosine phosphorylation by binding tyrosine-phosphorylated sequences. Structures of different SH2 domains exhibit a central beta-sheet which is flanked by two alpha-helices. Binding of phosphopeptides was first described as a prong-and-socket model. By comparing chemical shifts of amide resonances in different phosphopeptide complexes we determined the role of individual amino acids in the protein and in the phosphopeptide for specific interaction. From HSQC spectra recorded upon peptide titration we calculated differential kinetic parameters for individual amino acids of the protein. This analysis shows that the life time of the protein-peptide interaction varies by almost two orders of magnitude for different positions of the protein. The binding reaction follows a second order kinetic for all amino acids. For some residues we observe a multi-step binding reaction which may be correlated to rearrangements required for ligand binding. Binding kinetics are compared for different ligands and various SH2 mutants. Changes in the nature of the ligand are reflected in altered local binding kinetics. For mutants with altered binding specificity effects on kinetic parameters of individual amino acids are observed. Analysis of NMR chemical shifts in combination with a kinetic analysis provides information on protein-ligand interaction which should be particularly useful for the design of synthetic inhibitors.
Core binding factor (CBF) was originally identified as a DNA-binding protein that specifically binds to the asymmetric sequence PyGPyGGT, corresponding to the highly conserved "core" site in mammalian type C retrovirus enhancers. CBF binding sites have subsequently been identified in a number of T-cell specific genes, providing evidence for the the role of CBF as a T-cell transcription factor. Isolation and subsequent cloning of CBF showed the protein to be a heteromer consisting of an a and a b subunit. The a subunit contacts the DNA directly, whereas the b subunit does not, as indicated by the lack of any changes in the number of phosphate contacts made by a in the presence of b. Binding of the b subunit to the a subunit increases the affinity of the a subunit for the DNA sixfold without altering the sequence specificity. Glutathione S-transferase (GST) fusion proteins with the DNA-binding domain alone have shown this domain is responsible for both the DNA-binding and b-binding capabilities of the a subunit. The lack of any resemblance of the DNA-binding domain or CBFb to any known structural motifs makes them important targets for structure determination. We are carrying out a structure determination of a 20 KDa DNA-binding domain fragment bound to an 18 bp DNA oligonucleotide containing a core site. Due to the size of the resulting complex (31 KDa), we are utilizing partial deuteration (50%) to improve the performance of the triple resonance spectra while still being able to record high quality NOESY data. Backbone heteronuclear resonance assignments obtained from a number of triple resonance spectra and a preliminary analysis of the secondary structure of the protein will be presented.
Glutaredoxins provide an external cofactor to ribonucleotide reductase in the de novo biosynthesis of DNA. A theme ubiquitous throughout numerous species and within two large DNA viruses, vaccinia and T4 phage, is the encoding and packaging of redox-active glutaredoxin(s). Vaccinia Glutaredoxin-1 (Grx-1), a product of the o2L gene, is a functional thioltransferase and is expressed after the onset of DNA replication. Our goal is understanding the structural/dynamical differences and similarities among these evolutionary conserved redox proteins. Herein, we report progess toward the structure determination of Vaccinia Grx-1. An array of multidimensional heteronuclear NMR experiments were recorded on15N and 13C/15N labeled Vaccinia Grx-1 samples. Complete resonance assignments have been obtained and secondary structural elements identified on the basis of NOE information and CSI data.
The first committed step in de novo DNA biosynthesis is the direct reduction of ribonucleotides to form deoxyribonucleotides. Ribonuleotide Reductase (RR) is the enzyme that catalyzes this step.The reduction equivalents needed for the production of deoxiribonucleotides are provided by Glutaredoxin (Grx) or Thioredoxin (Trx), with NADPH as the primary donnor.The primary location for transfer of reducing equivalents from Grx to RR is at the C-terminus of the RR B1 subunit. We have prepared a mixed disulfide between Grx and a 25-mer peptide corresponding to the C-terminus of ribonucleotide reductase. Uniformelly labeled 15N and 13C/15N Grx and peptide were overexpresed in a bacterial system. Samples in which either the protein or the peptide were selectively 15N or 13C/15N labeled were prepared and characterized. With the use of several 2D and 3D homo and heteronuclear experiments and half-filtered experiments, we have obtained the complete assignments of the protein and the peptide. Here we present the structure of the complex based on the obtained intra- and inter-molecular NOE and dihedral angle constraints. We have determined the residues involved in binding and mapped this interactions on the protein. 15N dynamics of the peptide, in agreement with the intermolecular NOE data information, will also be presented.
Core binding factor (CBF) are heteromeric transcription factors essential for several developmental processes, including hematopoiesis. CBFs contain a DNA-binding CBFa subunit and a non-DNA binding CBFb subunit which increases the affinity of CBFa for DNA. Knockouts of the genes encoding either subunit of CBF in mice result in embryonic lethality and a profound block in hematopoietic development. Herein, we show the progress made toward the structural determination of the CBFb subunit. NMR experiments were performed on uniformly 15N- and 15N/13C-labeled CBFb(141) using a variety of heteronuclear NMR experiments. The complete assignments for the protein have been obtained and the secondary structure elements have been determined. The tertiary structure of CBFb(141) will be presented. Additionally, the site for binding of CBFa on CBFb has been mapped by comparison of amide NH chemical shifts for CBFb(141) in a ternary complex with the DNA-binding domain of CBFa and DNA.
The non-receptor tyrosine kinase c-Src is an important component of intracellular signal transduction. The proto-oncogene encoding Src has been implicated in several carcinomas, including breast and colon cancer. Regulation of Src is implemented through its two modular binding domains, SH3 and SH2. Allosteric interactions between contiguous SH3-SH2 domains have been established for Src(1,2) as well as two Src family members (3,4,5). The occupancy of the SH2 domain with a high affinity ligand (EPQpYEEIPIYL) induced significant amide 1H and 15N chemical shift changes in the SH3 domain and the intervening linker(6). Interdomain communication may finely tune the specificity and binding kinetics of Src. This additional level of regulation would directly affect targetting of Src's catalytic domain, modulating the high degree of specificity and critical timing necessary to prevent cross-talk in the signal transduction circuitry. Further characterization of the interdomain communication between SH3 and SH2 using high resolution solution state NMR coupled with titration microcalorimetry will be presented, focussing on the structural effects of binding the high affinity SH3 domain ligand RALPPLPRY. 3) Panchamorthy, G. et. al. (1994) Mol. and Cel. Biology, 14, 461-472. 4) Nam, H.-J. et. al. (1996) Structure, 4, 1105-1114. 5) Cowburn, D. et al. (1997) J. Biol. Chem., 270, 26738-26741. 6) Tessari, M. (1997) Biochemistry, 36, 14561-14571.
Bacteriorhodopsin (bR) is a 26 kD retinal protein that functions as a light-activated proton pump. Various biophysical techniques have shown that bR undergoes chromophore isomerizations and protein conformational changes as it progresses through its photocycle. Detailed structural characterization is an important step toward understanding the pump mechanism. Previous one-dimensional solid-state NMR studies have provided chemical shift assignment for various bR photointermediates by selective labeling. In addition, unambiguous measurements have been made of internuclear distances in isolated spin pairs. However, these piecemeal processes and time-consuming and costly. A method capable of simultaneously assigning multiple chemical shifts and/or measuring multiple distances is required to effectively determine protein structures. We have applied two-dimensional correlation spectroscopy to a bR sample regenerated with multiply-labeled retinals and assigned the chemical shifts of each spin. The application of novel, double-quantum polarization transfer techniques with increased transfer efficiency has afforded the signal-to-noise required to obtain quality multi-dimensional spectra of this large protein.
Melanoma Inhibitory Activity (MIA) is a 12 kD protein that is secreted from several malignant melanoma cell lines. In vitrostudies indicate that MIA inhibits the growth of these cell lines, and thus MIA is predicted to be a growth factor of unknown biological function. Structural prediction methods place the fold of MIA into the class of SH3 domains. Since this would be the first example of a cytokine with an SH3 fold, we have undertaken a full structural determination of MIA using NMR. We present assignments and compare MIA's secondary structure to that of SH3 domains.
Chemokines help direct the trafficking of leukocytes in the immune response and are believed to play key roles in both normal host defense and the pathogenesis of a variety of diseases including cancer, atherosclerosis, abnormal inflammation, and AIDS. Most chemokines are small (8-11kDa), secreted proteins that belong to one of 3 classes (CXC, CC, or C) depending on the number and arrangement of N-terminal cysteines. Recently, a protein constituting a fourth chemokine class, fractalkine, has been identified. Fractalkine has a unique CX3C spacing of the characteristic motif and is a transmembrane protein, with a chemokine module that is attached to the membrane via an extended mucin-like stalk. Membrane-bound fractalkine is upregulated in endothelial cells and causes adhesion of T-cells and monocytes in vitro. The molecule can also be released as a soluble, 95kDa glycoprotein that causes in vitro migration of T-cells and monocytes. We present the high resolution solution structure and backbone dynamics of the N-terminal chemokine domain (residues 1-76) of fractalkine. The structure was determined using ambiguous distance restraints and floating stereospecific assignment methods. We show that the protein is monomeric and compare its structure to those from both CXC and CC families.
ATP-utilizing enzymes occur in a number of critical cellular processes. Our goal is to determine the structures of enzyme-bound reaction complexes in a group of these enzymes in order to understand the mechanisms of the reactions. All ATP-utilizing enzymes require an obligatory divalent cation, Mg(II), and most of them are activated by Mn(II). Measurement of the paramagnetic effects on T1 allows the determination of distances of selected substrate nuclei from the Mn(II). In this abstract, we present 13C relaxation measurements of [UL_13C] nucleotides bound to two enzymes, yeast 3-phosphoglycerate kinase and E. coli adenylate kinase. These measurements were performed at 125 MHz and 75 MHz as a function of Mn(II) concentration and temperatures between 5oC and 25oC. From an analysis of the frequency and temperature dependence of the relaxation rates, we determined the distances of all the ten carbons on the adenosine moeity from the cation in the enzyme bound state. An isolated ATP molecule has several internal mobilities. Therefore, considerable structural data is necessary to establish its conformation in the enzyme bound complex. We have previously determined the chelation pattern of the cation to the phosphate chain of the nucleotide by 31P relaxation measurements and the glycosidic torsion and the sugar pucker from TRNOESY measurements. These results, taken along with the 13C relaxation measurements reported here, allow a detailed characterization of the nucleotide conformation in the enzyme-bound reaction complexes by the use of molecular modelling computations. Results of these analyses will be presented.
The RNA Polymerase II 14.4 kDa subunit, h-RPB6, was solved by solution NMR methods. RPB6 is conserved among the eukaryotes and shown to be essential for the cell survival. Specific interactions between RPB6 and others RNA Pol-II components, such as RPB5 and RPB3, have been described. Nevertheless, the specific function of RPB6 is still unknown. The h-RPB6 structure consists of a triple stranded antiparallel beta-sheet and two alpha-helices. The beta-sheet strands, the C-terminal region and residues 52-59 flanking the N-termini of the alpha1-helix form the core of the molecule. Relaxation experiments show that the N-terminal portion of the molecule is not well defined. Consistently, the N-terminal region of RPB6 family is not conserved, whereas the C-termini, which form part of the molecule core, is highly conserved. The application of the J-Doubling method to obtain coupling constants between Ha-HN from 15N-HSQC traces and further incorporation of the given fi angles restraints refined loops and the alpha2-helix. RPB6 represents a novel protein fold.
In adaptation to the osmolarity stress in growth environment, Echerichia coli<cells modulate the expression of the porin genes OmpC and OmpF through the function of EnvZ and OmpR at the transcriptional level. This regulation system is called histidyl-aspartyl phosphorelay system. EnvZ, the osmosensor, is a 450AA membrane-located protein which forms homodimer and exhibits kinase activity specific to OmpR at the C-terminus. Meanwhile osmoregulation, the EnvZ mediated phosphorylation of OmpR is a crucial event and the autophosphorylation at conserved His243 is the starting signal. In this study, multidimentional heteronuclear NMR techniques has been applied to an EnvZ fragment (residues 223-289), which is responsible for the dimer formation of EnvZ and contains the autophosphorylation site, His243. Chemical shifts index, hidrogen bonds, and other NMR data indicate that each subunit of the homodimer contains two alpha helices in large. The orientation of two subunits was determined by inter-subunit NOE contacts observed in [13C/F3]-filtered [13C/F1]edited HMQC-NOESY experiment for the 1:1 mixture of 13C-labelled and of unlabelled EnvZ(223-289) samples. Symmetrically orientated two subunits forms 4 helix bundle and this structural information will be used to discuss the EnvZ -OmpR interaction mechanism.
The Ras superfamily of GTP-binding proteins are involved in a number of cellular signaling events including, but not limited to, tumorigenesis, intracellular trafficking, and cytoskeletal organization. The Rho subfamily, of which Cdc42Hs is a member, is involved in cell morphogenesis through a GTPase cascade which regulates cytoskeletal changes. Cdc42Hs has been shown to stimulate DNA synthesis as well as to initiate a protein kinase cascade that begins with the activation of the p21-activated serine/threonine kinases (PAKs). We have determined previously the solution structure of Cdc42Hs (Feltham et al. (1997) Biochemistry, 36:8755) using NMR spectroscopy. The binding domain of PAK on Cdc42Hs was identified by producing successively smaller constructs surrounding the CRIB (Cdc42/Rac-interactive-binding) sequence. A minimal binding domain of 46 amino acids was identified (PBD46) which bound to Cdc42Hs with a KD of approximately 20 nM and inhibitied the GTP dissociation and hydrolysis. The binding interface was mapped by producing a 100% deuterated sample of 15N-Cdc42Hs bound to PBD46. An 1H,15N-NOESYHSQC spectrum demonstrated that the binding surface consists of a portion of the loop between alpha1 and beta2 (switch I) and beta2. A complex of PBD46 bound to 15N-Cdc42Hs(GMPPCP) exhibited extensive chemical shift changes in the 1H,15N-HSQC spectrum which mapped to the beta1, beta2, beta3, alpha1, alpha5, switch I and switch II (loop between beta3 and beta4) regions of the protein. Thus, PBD46 likely produces structural changes in Cdc42Hs which are not limited to the binding interface, consistent with its effects on GTP dissociation and hydrolysis. These results suggest that the kinase binding domain on Cdc42Hs is similar to but more extensive than the c-Raf binding domain on the ras antagonist, Rap1 (Nassar et al. (1995) Nature 375:554).
We are investigating the contribution of hydrophobic core packing to the stability, structure, and uniqueness of proteins. We have previously designed and characterized nine hydrophobic core variants of the protein ubiquitin. Results indicate that the proteins differ considerably in their thermodynamic and solubility properties, yet all appear to retain the same fold as the WT and all possess a significant level of conformational uniqueness. The current study aims to characterize selected variants at a more detailed level. We are using heteronuclear NMR spectroscopy to determine the high resolution structure and investigate the dynamics of three of the ubiquitin core variants. Two of these variants have designed cores with 7 and 8 mutations respectively, and the third variant has a randomized core with 7 mutations and is used as a control protein. These three variants and WT possess significantly different physical properties, yet all have comparable core volumes. As a consequence we are able to interpret structural and dynamic differences solely in terms of packing interactions.
KSI, a homodimer with 125 residues per subunit catalyzes the conversion of 5- to 4-3-ketosteroids via a dienolic intermediate. Tyr-14 and Asp-38 function as acid and base catalysts. The structure of the KSI-NTHS complex was solved by heteronuclear multidimensional NMR methods using 2064 distance restraints. The secondary structure consists of 3 helices, 7 beta-strands, and 5 turns (Biochem, 36, 34-58). The beta-strands form a mixed beta-sheet. The 3 helices are packed onto the concave surface of the beta-sheet with a groove between the helices and sheet into which NTHS binds at a site defined by 30 intermolecular proximities. Tyr-14 from alpha-helix 1 approaches Asp-99 and the 3-keto group of NTHS. Asp-38 from beta-strand 1 approaches C4 and C6 of the alpha-face of NTHS (Biochem, 36, 14616-14626). The dimer interface, defined by 9 NOEs obtained from the 3D 13C-edited NOESY spectrum on a 13C/12C hetero-labeled dimer, is between the convex surface of the beta-sheets. Intersubunit hydrophobic interactions occur between V74-L115', V74-V40', A79-L115', A79-I98', A75-V40' and A75-G41'. Preliminary comparison of the enzyme-steroid complex with the free enzyme (Wu et al, Science, 276, 415) reveals small changes in backbone structure near the steroid binding site.
Scalar and dipolar coupling data derived restraints are included with NOE derived restraints during the dynamics and simulated annealing phases of NMR structure calculation. The improvement in the accuracy and precision of structures calculated using scalar and dipolar coupling restraints is compared and contrasted using structures calculated without scalar and dipolar coupling data for Human Ubiquitin. Improvement in agreement with the crystallographic structure and in quality of the Ramachandran map are demonstrated. In addition, the dipolar couplings themselves are used to cross validate the improvement in the quality of the NMR structures upon addition of dipolar and scalar coupling restraints.
Hepatocyte growth factor (HGF) is a plasminogen-related multipotent growth factor that transduces a wide range of biological signals, including mitogenesis. motogenesis and morphogenesis. HGF is also implicated in the growth, invasion, and metastasis of tumor cells. The amino terminal domain (N) of HGF is important for receptor-binding. The N domain is also the primary binding site for heparin or heparan sulfate, which modulates receptor-dependent mitogenesis. Using multi-dimensional NMR techniques, including 3D and 4D experiments, we have solved the first structure of HGF, a high-resolution solution structure of the N domain. The structure reveals a novel folding topology with a distinct pattern of charge distribution and indicates a possible heparin-binding site. A hairpin loop region, which contains two disulfide bonds and is found in all HGF family of proteins, plays a major role in stabilizing the structure and contributes to the heparin-binding site. This binding site was confirmed by heparin-binding studies and a binding constant was extracted. Research sponsored in part by the National Cancer Institute, DHHS, under contract with ABL.
The active transport of a wide variety of substances through the cytoplasmic membrane of gram-negative bacteria is accomplished by transport systems consisting of a binding protein located in the periplasmic space, and a membrane-bound protein complex. The periplasmic binding proteins share a smilar two-domain hinged structure; however, each one exhibits high affinity and specificity for a particular ligand. The structural basis for ligand binding, affinity, and specificity in this class of proteins is being investigated by ongoing multidimensional NMR studies of liganded glutamine-binding protein (GlnBP) of Escherichia coli. The polypeptide backbone resonances of this 25 kDa protein were previously assigned using several 3D-NMR experiments performed on uniformly and specifically (15N,13C)-labeled samples. Assignments are extended to the amino acid side-chains using C(CO)NH, H(CCO)NH, and HCCH-TOCSY experiments. A hinge-bending conformational change similar to that seen in other periplasmic binding proteins is applied to the x-ray structure of unliganded GlnBP to generate a model structure of the closed-form, liganded GlnBP. This model structure is used to aid the assignment of 3D 15N- and 13C- edited NOESY-HMQC and HMQC-NOESY-HMQC cross peaks. These NOEs provide distance constraints which are used to generate initial structures by the distance geometry / simulated annealing methods implemented in the XPLOR software package. The initial stuctures are then refined using restrained molecular dynamics. Progress in determining both the solution structure and dynamic behavior of liganded GlnBP will be presented. This work is supported by NIH grants (GM-28874 and HL-24525).
Chemokines are a family of small secreted proteins that play an important role in inflammatory responses by recruiting leukocytes to inflammatory loci. Inhibition of chemokine activity may be a powerful strategy to treat inflammatory diseases. Development of chemokine inhibitors requires a detailed understanding of structure-function. We are using NMR spectroscopy and a variety of other biophysical and biochemical methods to determine the structure and structure-function relationships of human eotaxin, a recently identified chemokine which can attract eosinophils and basophils. Imbalance PCR with six oligonucleotides was used to synthesize a gene encoding human eotaxin with codons optimized for E. coli expression. A variety of expression constructs were made in order to optimize expression levels and yield a protein with the native N-terminus which is important for activity. The differences between the expression levels and activities of the various expressed proteins will be described along with characterization of the self-association properties using analytical ultracentrifugation and current progress towards the NMR structure
HMG-D is a non-histone chromosomal protein from Drosophila melanogaster which contains a high mobility group (HMG) DNA binding domain. Although HMG-D is a non-sequence specific DNA binding protein, it preferentially binds to both deformable DNA sites and DNA that is inherently bent, such as the bent DNA formed by the anti-tumor drug, cisplatin. Previous nuclear magnetic resonance (NMR) structure determinations have shown that HMG proteins, including HMG-D, adopt a conserved 'L' shaped three helix motif. Structures of two sequence-specific HMG proteins (LEF-1 and SRY) complexed with DNA have also been determined, but the information gained from the examination of the protein/DNA interactions present in these cases cannot be extended directly to the non-specific class of protein exemplified by HMG-D. Until now, obtaining structural information from complexes of non-sequence specific HMG proteins with DNA, has been elusive. We have designed a stable, well-behaved complex of HMG-D with a 16 basepair strand of duplex DNA which has proven to be suitable for structural analysis. The quality of this samples binding characteristics and preliminary NMR spectra lead us to construct a sample of 100% 15N, 13C, 60% 2H labeled HMG-D complexed with the crosslinked DNA. This sample has been used in a series of multi-dimensional NMR experiments designed to facilitate the assignment of all backbone chemical shifts in the protein. We will show how the complex was designed, discuss the strategy for a full structure determination, and present current NMR data.
Telomeres are the protein-DNA complexes that protect the ends of eukaryotic linear chromosomes from degradation and fusion. Mammalian telomeres are composed of long tandem arrays of the double stranded telomeric repeat TTAGGG packaged by the telomeric protein, telomeric-repeat binding factor, TRF1. The human TRF1 consists of 439 amino acids, containing three functional domains: an N-terminal acidic domain, a central TRF-specific dimerization domain and a C-terminal DNA-binding domain. Here, we have determined the NMR structure of the hTRF1 DNA-binding domain consisting of 53 amino acids. It contains three helices, whose architecture is very similar to that of each of three repeats of the c-Myb DNA-binding domain and also to that of each of two subdomains in the DNA-binding domain of yeast telomeric protein, RAP1. The second and third helices of the hTRF1 DNA-binding domain form a helix-turn-helix (HTH) variant motif containing a longer turn than the corresponding turn in prototypical HTH proteins, c-Myb repeats and RAP1 subdomains. Although the each c-Myb repeat alone has no sequence specific DNA-binding ability: the second and third repeats of c-Myb are essential and sufficient for the sequence specific DNA-binding, the single c-Myb repeat homologous domain of hTRF1 seems to bind specifically with a DNA fragment containing a telomeric DNA sequence, TTAGGG. Base on the present structure, a model of the hTRF1 DNA-binding domain in a complex with human telomeric DNA was constructed and the DNA-binding mode of hTRF1 is discussed.
Villin [1] is a member of the actin severing protein family, which binds and severs actin filament or caps actin monomers in a Ca2+/PIP2 regulated manner. We present the structure of the Ca2+ bound form of the N-terminal domain of villin (villin14t, 14 kD) and discuss local conformational changes compared to the apo form[2]. We present results of the Ca2+ titration on the two N-terminal domains (villin28t, 28 kD) and discuss them in respect to the crystal structure of the homologous protein gelsolin[3]. [1] W.L.Bazari et al., Proc. Natl. Acad. Sci. USA 85, 4986 4990 ( 1988 ) [2] M.M.Markus et al., Protein Science 6, 1197 1209 ( 1997) [3] L.D. Burtwick et al., Cell 90, 661-670 ( 1997) This work was funded by the DFG and the NIH
The GATA family of transcription factors possess tandem repeats of a highly conserved DNA-binding domain (DBD) comprised of a Cys2-Cys2 type IV zinc finger followed by a basic tail. The C-terminal zinc finger module of these tandem repeats is capable of tight, specific binding to the consensus target (A/T)GATA(A/G). Although it was previously found that the N-terminal finger of GATA-1 did not independently recognize GATA targets, recent results show that the N-terminal fingers of GATA-2 and GATA-3 demonstrate strong, independent binding to palindromic GATC core sequences. These N-terminal DBDs are further distinguished by the presence of two basic arms which flank the zinc finger core and are required for specific, high affinity binding. The utilization of two basic arms for binding is a variation of the known GATA DBD motif and may represent a mechanism for selective transcriptional control when numerous GATA factors are present during cellular development. In order to further investigate the molecular mechanisms which modulate individual GATA factor affinity and specificity, the three-dimensional structure of the N-terminal GATA-2 DBD complexed with an appropriate DNA target site is presented.
The T cell receptor (TCR) is a multi-subunit, membrane-bound complex that upon recognition of foreign antigens initiates the transduction of signals leading to an immune response. The antigen-binding specificity of the TCR is determined by the alpha and beta polypeptides of the complex. Single chain T cell receptors (scTCRs) contain the two variable domains from the alpha and beta chains of the TCR expressed as a single polypeptide. An scTCR derived from the murine D10 TCR was expressed in E. coli and shown to be functional through specific recognition of antigen bound to major histocompatibility complex (MHC). The solution structure of the 28 kilodalton D10 scTCR was determined using heteronuclear, multi-dimensional nuclear magnetic resonance experiments. Protein samples uniformly labeled with 13C, 15N and partially labeled with deuterium were used in the assignment and structure determination. The two domains in the scTCR adopt immunoglobulin-like folds and their interface is well defined. The packing of the two domains is similar to that observed in other immunoglobulins. Differences between the structure of the D10 scTCR and other TCRs occur in regions of greatest sequence variability. Nitrogen relaxation studies were used to study the backbone dynamics of the D10 scTCR. TCRs interact with MHC/antigen complexes through specific TCR loop regions. These loops in the D10 scTCR exhibit differential mobility on the nanosecond and picosecond time scales. The mobility differences are correlated with the sequences of the loops and with their biological functions.
The c-myc gene encodes a protein critical to the regulation of cell growth and differentiation. A number of upstream binding proteins influence the expression of this gene. One such factor, human nuclear ribonucleoprotein K (hnRNP K), contains three homologous units appropriately labeled KH1, KH2, and KH3 that are responsible for the binding affinity of this protein. These highly conserved KH sequences are also present in a variety of RNA-binding proteins. Our work reveals the structure of the C-terminal KH domain of hnRNP K, KH3. The KH3 construct utilized spans residues 379-463 of hnRNP K. The NMR structure was derived using NOE, dihedral, and internuclear dipolar coupling constraints. Dipolar couplings were measured in a dilute aqueous discotic liquid-crystalline medium that produced the requisite molecular alignment. A modified version of X-PLOR was utilized to incorporate the dipolar coupling information into the structure calculation. The results of several dynamics studies will also be presented. These findings suggest that KH3 forms a dimer under our NMR sample conditions.
RhoGDI plays a role as negative regulator of Rho-family GTPases involved in intracellular signal transduction pathway. It decreases the intrinsic rates of GDP dissociation and GTP hydrolysis, inhibits the actions of GTPase activating proteins (GAPs) and guanine nucleotide exchange factors (GEFs), and extracts RhoGTPases from membranes through binding to their C-terminal isoprene. Two structurally distinct regions of RhoGDI have been identified and the structure of the C-terminal domain has been determined using triple resonance NMR spectroscopy. The solution structure of the C-terminal domain appears as a beta-sandwich motif with a deep hydrophobic pocket that binds isoprenes, and the exposed surface interacts with the protein portion of Cdc42, a member of the RhoGTPase family. The isoprene binding site was mapped through chemical shift perturbation upon addition of farnesylated and geranylgeranylated peptides and confirmed by intermolecular NOEs between RhoGDI and the isoprene ligand. The determination of the complex structure is in progress and the results will be disscussed.
Conformationally sensitive solid state 13C NMR chemical shifts were used to determine the fraction of beta-sheet conformations in the alanine residues of artificially spun silk fibers. These fibers were spun from the dissolved silk of the silkworm, Bombyx mori, using a novel microspinneret that was fabricated using silicon etching technology. The NMR results were correlated with the mechanical properties of the artificial fibers. Results show that a relatively high fraction of the alanine residues in the silks must be in the beta-sheet conformation (>65%) in order to produce the highest stress fibers. However, the fraction of alanine residues in the beta-sheet conformation does not uniquely determine the maximum stress of a fiber; it is suggested that orientation of these beta-sheets is also an important parameter.
Protein Inhibitor of neuronal Nitric Oxide Synthase (nNOS) selectively binds to nNOS and destablize the dimer of nNOS, a conformation necessary for its biological activity. Nitric oxide, the product of nNOS, is an important neurotransmitter in the nervous system, regulating the cGMP level in brain and playing a role in the forming of memory. PIN, as the protein inhibitor of nNOS was first identified and cloned by Jaffrey, S. R. & Snyder, S. H. (1996, Science, 274, 774-777). We successfully subcloned a modified version of PIN (PINe) into pET-15b expression system and expressed in E. coli bacterium strain BL21(DE3). PINe inhibits nNOS as well as the original PIN. Gelfiltration analysis indicates that PINe is a dimer under our NMR experiment conditions. Both one and two dimensional NMR spectra indicate PINe is a folded protein. Two preliminary 15N-HSQC spectra of the PINe and PINe complexed with a 13mer peptide from nNOS display notable changes in chemical shifts, indicating that PINe may undergo a conformation change upon binding to the nNOS 13mer.
The development of human leukemias and lymphomas is typically associated with specific non-random chromosomal rearrangements that alter the normal regulated expression of specific genes. These changes can lead to the over-expression of regulatory proteins that alter cell growth or survival, this in turn contributes to the formation of malignancies. The t(X;14) chromosomal translocation of the MTCP-1 gene into the TCR locus drastically up-regulates its expression in pre-malignant T-cells. The over expression of the 108 residue MTCP-1 protein ( mature T cell proliferation-1) in T cells is a critical step in the progression to certain leukemias. Our goal is to investigate the solution structure of MTCP-1 and the related TCL-1 protein, members of a unique family of oncogenic proteins. This poster will report on progress towards the determination of the three dimensional structure of MTCP-1 using Nuclear Magnetic Resonance (NMR). Protocols for the recombinant expression, refolding and purification of large amounts of MTCP-1 have been developed. Uniformly labeled 15N MTCP-1 has been produced using a modified minimal media. Structural information will aid in elucidating MTCP-1's role in normal and malignant human T cells at the molecular level. The aim of this research complements the Kimmel Cancer Institute's initiative to develop a better understanding of the expression, structure and function of both genes and proteins important in human cancers
The F1Fo ATP Synthases use a transmembrane proton gradient to drive the synthesis of ATP from ADP and Pi during oxidative or photo-phosphorylation. Subunit c, and specifically its buried DCCD-reactive carboxyl residue Asp61, are directly involved in the proton transport. Changes in the ionization state of Asp61 of subunit c initiate the conformational changes leading to the release of ATP product at the catalytic sites of the enzyme. Subunit c of F1Fo ATPase folds normally as a pair of interacting anti-parallel helices in a mixed solvent (4:4:1 CHCl3:CH3OH:H2O). The 3D structure of subunit c at pH 6.0 (Asp61 protonated) was presented during ENC97 (WOpm). Complete 1H, 13C and 15N resonance assignments for the protein at pH 8.0 (Asp61 ionized) have been obtained from a set of triple resonance experiments. The secondary structure of subunit c at pH 8.0 and comparative structural studies of subunit c at pH 6.0 and at pH 8.0 will be presented, focussing on the structural changes in the polar loop and the rearrangement of the alpha helices.
Myotrophin is a novel 12.5 kDa protein recently identified at elevated levels from hypertrophic human hearts, which is composed of ANK repeats, the ubiquitous 33 amino acid motifs known to be involved in a wide range of protein-protein interactions. Myotrophin has been shown to stimulate protein synthesis in cardiac myocytes and induce early genes such as c-fos and c-jun, and hypertrophy markers such as b-MHC and b-MHC ANF, providing evidence that it may play an important role in the initiation of cardiac hypertrophy. To understand the molecular mechanism of myotrophin in the process of cardiac hypertrophy as well as to gain insight into the structural basis of the ubiquitous and multifunctional ANK repeat motif, we have undertaken the solution structure determination of myotrophin by multidimensional heteronuclear NMR spectroscopy. The structure determination was based on 2840 experimental NMR restraints, and the precision of the coordinates for the of final 45 simulated annealing structures is 0.43 Angstroms for the backbone and 0.87 Angstroms for all heavy atoms. The structure of myotrophin is well-defined and is ellipsoida, approximately 46 Angstroms long and 21 Angstroms wide. The elongated shape of myotrophin arises from four sequential helix-turn-helix motifs stacked as helical bundles, which are primarily stabilized by compact hydrophobic cores. The ANK repeats are characteristic of a protruding loop followed by a V-shaped helix-turn-helix motif. ANK repeats, which constitute the main part of the myotrophin structure, appear to adopt the similar folding pattern to those of other ANK repeat-bearing proteins, despite their loose consensus sequence. A comparison with the recently solved crystal structure of another functionally different ANK repeat-bearing protein, 53BP2 in complex with p53, indicates that the protruding loops of ANK repeats, which lie along the same side of the myotrophin structure, form the binding interface to the target protein, whereas the variations of sequence and conformational flexibility in the loop regions of the two proteins likely confer the different functional specificity.
Cellular Retinol Binding Protein II (CRBP II) is an abundant intestinal 15.6 kDa protein, which is felt to play an important role in the intestinal uptake and metabolism of vitamin A. To selectively observe various components of the CRBP II retinol complex, time-shared [13C,15N] double-half-filtered NOE experiment was performed on uniformly 13C, 15N enriched recombinant rat CRBPII complexed with unlabeled all-trans-retinol. Multiple proton resonances were observed for each vinyl proton on the polyene chain, suggesting that the CRBPII-bound retinol assumes multiple conformations. These results are consistent with the results of previous 19F NMR studies of fluororetinols complexed with CRBP II. The NOESY data are more consistent with a 6-s cis, torsion angle between the ring and the polyene chain , than the 6-s trans conformation described for retinol bound to crystalline CRBP and CRBP II. Of note the 6-s cis conformation is observed for all-trans-retinoic acid complexed to CRABP II based on both NMR and crystallography data. The results of the 8-, 10- and 12- single bonds in the polyene chain were in the s- trans configuration and are therefore consistent with the x-ray structure of crystalline CRBPII-retinol. In summary, the NMR studies provides a dynamic picture of CRBP II bound retinol in solution.
Calmodulin (CaM) is an ubiquitous Ca2+ binding protein which plays a pivotal role in the signal transduction chain at the cell membrane in eucariotic cells. Free CaM has a dumbbell type structure with two globular domains connected by a flexible linker. Binding of CaM to target proteins induces a dramatical conformational change in which the dumbbell like structure of free CaM collapses to a globular structure. In this study we used NMR spectroscopy to determine the difference in binding of CaM with two peptides (C20W and C21W) resembling overlapping regions of the binding domain of the plasma membrane Ca2+ pump [1-6]. We employed CaM enriched in both 13C and 15N obtained from expession in E. coli, and unlabled peptides obtained by the procedure of Vorherr [4]. Complete Sequence assignments of the complexes were obtained by a number of 3-D experiments (HNCACB, CBCA(CO)NH, HBHA(CO)NH, HCCH-TOCSY). Structural information was obtained by NOE measurments within CaM (NH-NOESY-HSQC and CH-NOESY-HSQC), within the peptide and between the two (filtered NOESY). Information about backbone dynamics was obtained using 15N T1 relaxation and hetero-NOE measurements. As a result C20W binds only to the C-terminal domain of CaM, leaving the general structure of CaM mainly intact. C21W on the other hand interacts with both domains of CaM. The resulting differences in structure and dynamics between the two complexes will be discussed. [1] P. James et al. J. Biol. Chem. 1988, 263, 2905. [2] D. B: Heidorn et al. Biochemistry 1989, 28, 6757. [3] J. Trewhella et al. Biochemistry 1990, 29, 9316. [4] T. Vorherr et al. Biochemistry 1990, 29, 355. [5] M. Kataoka et. al. Biochemistry 1991, 30, 6247. [6] T. Vorherr et al. Biochemistry 1992, 31, 8245.
Calcyclin is an S100 EF-hand calcium-binding protein that has an implied role in the regulation of cell-growth or division, exhibits deregulated expression in association with cell transformation and is found in high abundance in certain breast cancer cell lines. The novel homodimeric structural motif first identified for apo calcyclin raised the possibility that S100s recognize target proteins in a manner that is distinctly different from the current calcium sensor paradigm based on calmodulin. The NMR solution structure of calcium-loaded calcyclin was determined to identify the calcium-induced changes in structure and to obtain insights into the mechanism for calcium-triggered target protein recognition. 1H, 13C and 15N resonance assignments and secondary structure information have been obtained from a series of 2D and 3D homonuclear and heteronuclear NMR experiments. The solution structure of calcium-loaded calcyclin has been determined using experimental constraints derived from 2D homonuclear NOESY, 3D15N-1H NOESY-HSQC and 3D 13C-1H NOESY-HSQC experiments. Calcium-calcyclin has the same symmetric homodimeric fold as observed for the apo protein. Only very modest calcium-induced changes are observed in the structure of calcyclin, in sharp contrast to the domain-opening that occurs in calmodulin and related calcium sensor proteins. The solution structure of calcium-bound calcyclin will be presented and implications for Ca2+-signal transduction by S100 proteins will be discussed.
Calcyclin, a member of the S100 family, is a Ca2+-binding protein implicated in calcium-mediated signal transduction pathways. The protein is a symmetrical homodimer and is comprised of two 90 amino acid residue monomers. Each monomer subunit contains two helix-loop-helix motifs, known as EF-hands. The three-dimensional solution structure for apo calcyclin will be presented. Nearly complete 1H, 15N and 13C resonance assignments have been obtained. Evaluation of 3D 13C NOESY-HSQC, 4D 13C/13C HMQC-NOESY-HMQC as well as 2D homonuclear NOESY experiments at different mixing times has allowed for a large number of distance restraints to be obtained. Backbone and side-chain torsion angle restraints have been derived from the analysis of various 3D heteronuclear correlation experiments. However, the symmetry of the dimer makes it difficult to distinguish inter- and intra-molecular NOE contacts. The dimer interface is defined by an antiparallel alignment of helices I and I' and of helices IV and IV', bringing these helices in close contact both at the monomer as well as at the dimer level of the structure. The use of 2D and 3D select-filtered NOESY experiments on an asymmetrically labeled dimer (produced from a mixture of labeled and unlabeled protein) were essential for obtaining inter-molecular restraints for structure calculations. Strategies for sample preparation, characterization and data acquisition will be discussed. The high resolution of the strucure at the dimer level has been made possible by the large number of inter-subunit NOEs obtained with the above mentioned strategy. The present structure allows for detailed comparison of structural features with the Ca2+-bound state of calcyclin and with those of other homologous S100 calcium-binding proteins.
The protein Rhodniin that can be purified from the assassi bug Rhodnius prolixus is a thrombin-specific inhibitor [1] with a molecular mass of 11 kDa. The amino acid sequence (103 residues) is homologous to Kazal-type inhibitors from egg white or the saliva of predatory animals. Backbone and alipahtic side chain assignment of the 13C/15N labeled protein was obtained from a combination of HNCO, CBCACO(N)H, CBCANH, (H)CCCO(N)H, H(CC)CO(N)H and HCCH-TOCSY experiments. In order to assign resonances of the aromatic ring systems which are of vital importance for the evaluation of hydrophobic cores a (HB)CB(Car)H experiment [2] has been recorded that correlates the resonances of the aromatic protons with the CB resonances. NOE's were measured with simultaneous edited 13C/15N- HSQC-NOESY and 4D-HMQC-NOESY-HSQC experiments. Slow conformational exchange of the disulfide bridges lead to doubling of the resonances of amino acids close to the cysteins in the NMR spectra. Studies on this slow conformational equilibrium was done by measurement of heteronuclear longitudinal 2-spin order. Angles between pairs of spins can directly be determined by measurement of cross correlated relaxation rates [3] without need for a "Karplus" calibration. These rates were implemented in structure calculations using deviations from the theoretical rate curve as restraints in XPLOR. 1700 NOEs, 150 angle constraints and 40 hydrogen bridges were used in a structure calculation using XPLOR. The structure is charactarized by two homologous domains brigded by a 6 amino acid linker. Each domain contains a helix and a three-strand beta-sheet. Information on the binding sites in Rhodniin is derived from chemical shift differences of the free protein and the complex with Thrombin. The NMR results will be compared to the data derived from X-ray crystallography [4]. [1] T.Friedrich et al., J.Biol.Chem. 268, 16216 (1993) [2] T. Carlomagno et al., J.Biomol.NMR 8, 161 (1996) [3] B. Reif et al., Science 276, 1230 (1997) [4] A. van de Locht et al., EMBO Journal 21, 5149 (1995)
The largest contributors of fixed N2 in the biosphere are root nodules formed by the symbiotic relationship between legumes and Rhizobial bacteria. Two Rhizobial nodulation proteins, nodE and nodF, are both necessary and sufficient for synthesis of host-specific fatty acids that trigger nodule formation. NodF, a soluble protein of 92 amino acids, has a 4'-phosphopantetheine prosthetic group that carries the host-specific fatty acid during synthesis. NodF has sequence homology to acyl carrier proteins (ACPs) around the site of attachment of the prosthetic group. However, the structural homology between NodF and ACP seems to be general (Ghose et al. (1996) FEBS Letters 388, 66-72). Here, we present a solution structure determination of NodF from Rhizobium leguminosarum based on 1H/15N/13C double and triple resonance NMR data. The structural homology with ACP from E. coli is discussed.
The cellular mechanism and etiology of chronic inflammatory processes are poorly understood. Macrophages, which act predominently in inflammatory processes, have been shown to express two specific calcium-binding proteins which form a heterodimer in solution. This heterodimeric protein is a member of the S-100 subfamily of EF-hand calcium-binding proteins, and is comprised of a 10kD (MRP8) and a 14kD (MRP14) polypeptide chain. 15N-HSQC experiments indicate that homogenous solutions of MRP8 and MRP14 suffer from severe aggregation problems leading to limited signal dispersion and extreme linebroadening. However, the addition of the other monomeric unit leads to the formation of a more stable heterodimeric complex. The 15N-HSQC spectra of 15N-MRP8/unlabeled MRP14 or unlabeled MRP8/15N-MRP14 exhibit marked improvement in both signal dispersion and linewidths. Although this protein is a heterodimer, each subunit of the heterodimer can be characterized separately using a heteronuclear filtering strategy, labeling one chain and not the other. Using isotope-edited NMR experiments, the labeled protein can be selectively studied and facilitate the analysis of intermolecular contacts across the heterodimer interface. Select-filter strategies are also being inplemented to specifically identify NOEs at the interface between the two subunits.
Janos Borbely*, Gyula Batta, Katalin E. Kover and Agnes Krecz Kossuth L. University. Department of Colloid Chemistry, H-4010 Debrecen, HUNGARY NMR studies of proteins built up from repeating units 1 and 2 will be described. -[(AG)4EG]- (1) -[(AG)4PEG]- (2) These repatative proteins were prepared via bacterial expression of artificial genes, as part of a broadly based exploration of the conformational and crystallization properties of periodic polymer chains. Alanylglycine-rich proteins, including Bombix mori silk, are known to adopt extended sheet-like structures in the solid state. The purpose of the present study is to establish the effects of the chain sequence and processing conditions on the crystal structures of this class of polymers. Acknowledgement. Bacterial strains were taken from the University of Massachusetts, Polymer Science Department, Amherst, MA.
Caltractin (also known as centrin), a 19.5 kDa phosphoprotein, is of considerable interest due to its implicated role in the cell cycle and its ubiquitous distribution in eukaryotes. Caltractin belongs to the EF-hand superfamily of calcium-binding proteins, and shares approximately 50% sequence similarity with calmodulin. However, caltractin displays distinct functional, physical and spectroscopic properties not shared by other members of the EF-hand superfamily. Of particular interest, caltractin is phosphorylated in vivo. We are interested in examining the interplay of calcium- and phosphorylation- dependent structural properties of caltractin, since both signaling pathways are known to have central roles in regulation of the cell cycle. Our experimental approach is to study the two domains of the protein independently, and then apply the results to subsequent studies of the intact protein. Preliminary evidence suggests that this "isolated domain" strategy is valid for studies of caltractin. Isotopically labeled fragments of caltractin have been expressed in E. coli and purified, and multidimensional NMR experiments have been performed on the Ca2+-loaded form of the N-terminal domain of caltractin.
Neurocalcin-delta is a 22-kDa neuron-specific calcium-binding protein originally found in the bovine brain. Neurocalcin is thought to regulate a signal transduction in the central nervous system, because its homologue in retina, recoverin, is found to regulate a phototransduction system. Both neurocalcin and recoverin has four EF-hands in primary structure, but their calcium-binding abilities are different: neurocalcin binds three calcium ions using EF-2, EF-3, and EF-4, while recoverin binds two with EF-2 and EF-3. The tertiary structure of neurocalcin, especially in the C-terminal region that has EF-4 and binds the third calcium ion may be different from that of recoverin. We have started NMR analyses to understand the structural characteristics of neurocalcin in behavior against calcium ions. Uniformly 15N- or 13C/15N-labelled protein was overexpressed with N-myristoyl transferase and purified for NMR studies. The backbone assignments have been made by analyzing four triple-resonance experiments, (HB)CBCA(CO)NNH, HNCACB, (HB)CBCACO(CA)HA, and HNCO. Assignments of the side-chain resonances was carried out using the 3D HCCH-TOCSY, CCC-TOCSY, and 13C-edited NOESY-HMQC spectra. The secondary structure based on chemical shift index and medium-range NOEs detected in the 15N-edited NOESY spectrum suggested that neurocalcin has a similar structure in the N-terminal half with recoverin but has a longer C-terminal helix. The significance of the differences in 3D structure will be discussed.
Homeodomain containing transcription factors play important roles in development. Many mutations in homeodomains have been found to be lethal or cause diseases. In this study, an embryonically lethal single amino acid mutant, vnd/NK2-A35T, of the vnd/NK2 homeodomain that is important in determining the formation of a part of the central nervous system of Drosophila in over-expressed in E-coli and purified to homogeneity. Circular dichroism (CD) and NMR studies show that this protein in the free state is unfolded even at 0 C. It binds a vnd/NK2 target DNA with a binding constant 50 fold lower than that of the wild type vnd/NK2 homeodomain. The DNA bound vnd/NK2-A35T has a higher helix content and shows characteristic wild type vnd/NK2's folding. Significant differences between 3d structures of the DNA bound mutant vnd/NK2-A35T and wild type vnd/NK2 are observed.
After budding, the Human Immunodeficiency Virus (HIV) must "mature" into an infectious viral particle. Viral maturation requires proteolytic processing of the Gag polyprotein at the matrix-capsid (MA-CA) junction, which liberates the central capsid domain to condense from the spherical protein coat of the immature virus into the conical core of the mature virus. NMR spectroscopic studies 1 have revealed that the amino-terminal end of the mature capsid protein forms a b-hairpin that is stabilized by formation of a salt bridge between the processed amino-terminus (Pro1) and a highly conserved aspartate residue (Asp51). We hypothesize that the b-hairpin structure forms after Gag proteolysis, and that this refolding event causes capsid to switch from a spherical assembly (in the immature particle) to a conical core (in the mature virus). Consistent with this model, a recombinant capsid protein with just 4 N-terminal matrix residues assembles into spheres in vitro, whereas the protein forms cylinders when the matrix residues are removed2. NMR spectroscopic studies are being used to investigate structural details of this capsid maturation switch. Mulation of CA Asp51 to Ala unfolds the N-terminal capsid b-hairpin, demonstrating the importance of the salt bridge in stabilizing the mature capsid structure. Moreover, addition of the four terminal matrix amino acids to the N-terminal domain of capsid dramatically alters backbone chemical shifts throughout the b-hairpin and helices 1 and 2. The solution structure of this 150 amino acid MA-CA fusion protein is currently being determined and assingments have been obtained for 90/138 backbone amide protons. Comparison of the MA-CA structure with our previously determined solution structure of the mature CA protein will reveal the structural changes that accompany proteolysis at the MA-CA junction of HIV-1 Gag. 1Gitti, R. K., Lee, B. M., Walker, M. F., Summers, M. F., Yoo, S., Sundquist, W. I. (1996) Science 273, 231 2Von Schwedler, U.K., Stemmler, T. L., Klishko, V. Y., Albertine, K. H., Davis, D. R., Sundquist, W. I. (1997) submitted.
The advances made in recent years to provide complete resonance assignments and high-resolution solution structures for small and medium size proteins have been remarkable. Nevertheless, there is still a wide range of large to very large proteins [30 kDa and larger] which are very important biologically and are difficult to study by NMR methods. The usual problems are exacerbated further if the proteins are highly alpha-helical, which results in limited spectral resolution for both assignments and NOE measurements. To attack these problems we have utilized a combination of techniques: 1) uniform, 100% triple labeling with 15N/13C/2H combined with quadruple resonance experiments, 2) very high efficiency methyl protonation within the uniform >95% 15N/13C/2H background [lLV-labeling], 3) new high resolution 13C-edited NOE experiments, and 4) the use of dipolar couplings in partially aligned samples [Tjandra and Bax, Science, 278:1111-1114 (1997)]. This combination is enabling us to determine the solution structure (medium resolution) of a 30 kDa component of a STAT protein. The approach should be applicable to a wide range of large proteins. This research was sponsored by the National Cancer Institute, Department of Health and Human Services (DHHS) under contract with ABL.
Two peptides (23- and 25-residue), corresponding to the cytoplasmic domain of 13.7 kDa Adenoviral E3 protein which downregulates EGFR and other receptors have been investigated. These peptides contain three different sorting motifs. Their structures in water and dodecyl-phosphocholine micelles have been studied by NMR spectroscopy and circular dichroism (CD). The first peptide was N-acetylated, while the second one was investigated in N-myristoylated form. This study showed that the first peptide has mostly random structure in aqueous solution, but adopts a mostly a-helical structure when bound to micelles. The presence of a myristoyl anchor was not observed to be needed for tight association with micelles and did not affect the structre of the surface-bound polypeptide. Furthermore, the affinity of both peptides for a qwitterionic membrance surface is high. The addition of negatively charged lipids was not required for high affinity. Despite of presence of His in two of the sorting motifs, binding of these peptides to the membranc mimics was not pH-dependent. However, changing pH from acidic to neutral did affect slightly the second sorting motif and several adjacent residues on C-terminal side, decreasing their probability of being in helical conformation. Measurements made using POPC vesicles instead of micelles suggested that the mode of association of these peptides with a true bilayer surface is akin to that observed for micelles. Our major conclusion is that all three sorting motifs present in these peptides are closely associated with membrance surface. Two (YLRH and LL) lie at the interface as parts of amphipathic helices, while the third (HPQY) appears to "dimple" just above the surface, set between two anchored amphipathic helices. The high affinity and intimacy of membrance interactions observed for the E3 cytoplasmic domains raises the question of whether homologous sorting motifs found in various receptors might often be closely membrane associated, a hypothesis which has profound implication for the molecular methanisms of cellular protein sorting. As a preliminary test of this hypothesis the sequences adjacent to sorting motifs in a
Human acidic fibroblast growth factor aFGF is involved in a broad spectrum of physiological processes including cell-proliferation, cell differentiation and other hormone like activities. In vitro experiments have demonstrated that binding of heparin or other proteoglycan analogues increases the structural stability and mitogenic activity of aFGF. Low levels of expression, proteolytic processing, labile nature and its marked tendency to aggregate under physiological conditions have rendered high resolution structural studies extremely difficult. In this context, we have recently devised conditions to clone and express aFGF in E. coli in high yields. This aspect has permitted us to solve the solution structure of the full-length FGF using multi-dimensional, homo and heteronuclear NMR techniques. Recently, aFGF has been shown to bind to various nucleotides and binding of the nucleotides is shown to profoundly modulate the mitogenic activity of aFGF. In the present study, we also investigate the effect of binding of adenosine triphosphate on the three-dimensional structure of aFGF. In this context, we have studied the solution structure of the aFGF-ATP complex using various multidimensional NMR techniques. The details of the structural features and the functional significance of the aFGF-ATP complex would be discussed.