Multidimensional Heteronuclear NMR Research Articles

Unusual helix-containing greek keys in development-specific Ca(2+)-binding protein S. 1H, 15N, and 13C assignments and secondary structure determined with the use of multidimensional double and triple resonance heteronuclear NMR spectroscopy.

Multidimensional heteronuclear NMR spectroscopy has been used to determine almost complete backbone and side-chain 1H, 15N, and 13C resonance assignments of calcium loaded Myxococcus xanthus protein S (173 residues). Of the range of constant-time triple resonance experiments recorded, HNCACB and CBCA(CO)NH, which correlate C alpha and C beta with backbone amide resonances of the same and the succeeding residue respectively, proved particularly useful in resolving assignment ambiguities created by the 4-fold internal homology of the protein S amino acid sequence. Extensive side-chain 1H and 13C assignments have been obtained by analysis of HCCH-TOCSY and 15N-edited TOCSY-HMQC spectra. A combination of NOE, backbone amide proton exchange, 3JNH alpha coupling constant, and chemical shift data has been used to show that each of the protein S repeat units consists of four beta-strands in a Greek key arrangement. Two of the Greek keys contain a regular alpha-helix between the third and fourth strands, resulting in an unusual and possibly unique variation on this common folding motif. Despite similarity between two nine-residue stretches in the first and third domains of protein S and one of the Ca(2+)-binding sequences in bovine brain calmodulin [Inouye, S., Franceschini, T., & Inouye, M. (1983) Proc. Natl. Acad. Sci. U.S.A. 80, 6829-6833], the protein S topology in these regions is incompatible with an EF-hand calmodulin-type Ca(2+)-binding site.

Localization of bound water in the solution structure of a complex of the erythroid transcription factor GATA-1 with DNA.

The erythroid specific transcription factor GATA-1 is responsible for the regulation of transcription of erythroid-expressed genes and is an essential component required for the generation of the erythroid lineage. GATA-1 binds specifically as a monomer to the asymmetric consensus target sequence (T/A)GATA-(A/G) found in the cis-regulatory elements of all globin genes and most other erythroid specific genes that have been examined. We have previously determined the solution structure of the complex of the zinc-containing DNA-binding domain of chicken GATA-1 with its cognate DNA target site by multidimensional heteronuclear NMR. From previous studies of complexes between proteins and DNA, water appears to play an important role in DNA-protein recognition by mediating bridging hydrogen bonds between functional groups on the protein and DNA bases. Solvation free energy calculations, however, suggest that hydrophobic interactions should exclude water from parts of the GATA-1:DNA interface. Using water-selective two-dimensional heteronuclear magnetic resonance spectroscopy, we have identified the location of bound water molecules in the specific complex of chicken GATA-1 with DNA. A number of water molecules could be detected between the protein and the phosphate backbone, as well as at the solvent exposed surface of the protein. However, no water molecules could be observed at the interface of the protein with the bases of the DNA. With only one exception, the bound water molecules have a residency time > 200-300 ps. Unlike other protein-DNA complexes, the majority of specific interactions between GATA-1 and the DNA bases in the major groove are hydrophobic in nature. The exclusion of water from the protein-DNA base interface in the major groove supports the view that the specific binding energy is indeed dominated by hydrophobic effects.

Structural Characterization of the FK506 Binding Protein Unfolded in Urea and Guanidine Hydrochloride

Characterizing the structure properties of unfolded proteins is important for understanding the stability and folding of native proteins. However, little structural information is available for the unfolded state. Using recently developed heteronuclear multi-dimensional NMR techniques, the 1H, 13 C and 15N chemical shift assignments of the FK506 binding protein (FKBP) unfolded in concentrated urea and guanidine hydrochloride (GuHCl) solutions have been obtained, and the structural properties of unfolded FKBP have been characterized. FKBP displays extensive conformational averaging when unfolded in urea and GuHCl, but defined regions of secondary structure are present. Subtle differences regarding the location and stability of the secondary structures exist between the two solvents. Secondary structure formation in unfolded FKPB was correlated with statistical and thermodynamic predictions of helix formation as well as with the three-dimensional structure of folded FKBP determined by NMR and X-ray crystallography. Residues involved in secondary structures in unfolded FKPB are generally found in the same type of secondary structure in the folded protein. An exception to this was found at the C terminus of FKBP, which forms a different secondary structure in the unfolded and folded states.

Unique Structure of the DNA Binding Domain of Interferon Regulatory Factor 2 Determined by NMR Spectroscopy.

The secondary structure elements and the folding topology of the DNA-binding domain of mouse interferon regulatory factor 2 [IRF-2(113)] were determined by heteronuclear multidimensional NMR spectroscopy. The sequential NOE (nuclear Overhauser effect) connectivities indicated the presence of three α-helical regions and four short β-strands connected by relatively long loops. The long range NOEs indicated the four strands form an antiparallel β-sheet and the three α-helices form a bundle on the sheet. The arrangement of the secondary structure elements and the overall folding topology resemble those of the DNA binding domains of bacterial activator CAP, heat shock transcription factors and fork-head family transcription factors, although there is no sequence homology among them.

Sequence-specific assignments of the backbone 1H, 13C, and 15N resonances of the MutT enzyme by heteronuclear multidimensional NMR.

The MutT protein, a 129-residue enzyme from Escherichia coli which prevents A.T-->C.G mutations, catalyzes the hydrolysis of nucleoside triphosphates (NTP) to nucleoside monophosphates (NMP) and pyrophosphate [Bhatnagar, S. K., Bullions, L. C., & Bessman, M. J. (1991) J. Biol. Chem. 266, 9050-9054], by a mechanism involving nucleophilic substitution at the rarely attacked beta-phosphorus of NTP [Weber, D. J., Bhatnagar, S. K., Bullions, L. C., Bessman, M. J., & Mildvan, A. S. (1992a) J. Biol. Chem. 267, 16939-16942]. The bacterial MutT gene was inserted into the plasmid pET-11b under control of the T7 promoter and overexpressed in minimal media, permitting labeling of MutT with 13C and/or 15N. The yield after purification of the soluble fraction was approximately 35 mg of homogeneous MutT/L with physical and enzymatic properties indistinguishable from those of the originally isolated enzyme. Essentially complete sequence-specific assignments of the backbone HN, N, C alpha, H alpha, and CO resonances of the free enzyme (1.5 mM) were made at pH 7.4 and 32 degrees C, by heteronuclear double- and triple-resonance experiments using a modified Bruker AM 600 NMR spectrometer. Specifically, 1H[15N]HSQC, 1H[15N]TOCSY-HMQC, and 1H[15N]NOESY-HMQC experiments were done with uniformly 15N-labeled enzyme. A 1H[15N] HSQC experiment was done with selective [alpha-15N]Lys-labeled enzyme. Also HNCA, HN(CO)CA, HNCO, constant time 1H[13C]HSQC, HCACO, and HCA(CO)N experiments were done with uniformly 13C- and 15N-labeled enzyme. Sequence-specific assignments were initiated from HN and 15N chemical shifts of Gly residues and of selectively labeled Lys residues in 1H[15N]HSQC experiments. They were confirmed by C alpha chemical shifts of Ala residues uniquely identified by residual coupling to C beta resonances in constant time 1H[13C]HSQC experiments. The sequence-specific assignments proceeded bidirectionally, terminating at Pro residues and at residues with undetectable NH signals, and the segments were linked to complete the backbone assignments. The backbone assignments reported here have permitted the interpretation of NOEs in the elucidation of the solution secondary structure of MutT, and the C alpha and H alpha chemical shifts have provided an independent approach to identifying secondary structural elements and to define their extent [Weber, D. J., Abeygunawardana, C., Bessman, M. J., & Mildvan, A. S. (1993) Biochemistry (following paper in this issue)].

Secondary structure and zinc ligation of human recombinant short-form stromelysin by multidimensional heteronuclear NMR.

Stromelysin-1, a member of the matrix metalloendoprotease family, is a zinc protease involved in the degradation of connective tissue in the extracellular matrix. As a step toward determining the structure of this protein, multidimensional heteronuclear NMR experiments have been applied to an inhibited truncated form of human stromelysin-1. Extensive 1H, 13C, and 15N sequential assignments have been obtained with a combination of three- and four-dimensional experiments. On the basis of sequential and short-range NOEs and 13C alpha chemical shifts, two helices have been delineated, spanning residues Asp-111 to Val-127 and Leu-195 to Ser-206. A third helix spanning residues Asp-238 to Gly-247 is characterized by sequential NOEs and 13C alpha chemical shifts, but not short-range NOEs. The lack of the latter NOEs suggests that this helix is either distorted or mobile. Similarly, sequential and interstrand NOEs and 13C alpha chemical shifts characterize a four-stranded beta-sheet with three parallel strands (Arg-100 to Ile-101, Ile-142 to Ala-147, Asp-177 to Asp-181) and one antiparallel strand (Ala-165 to Tyr-168). Two zinc sites have been identified in stromelysin [Salowe et al. (1992) Biochemistry 31, 4535-4540]. The NMR spectral properties, including chemical shift, pH dependence, and proton coupling of the imidazole nitrogens of six histidine residues (151, 166, 179, 201, 205, and 211), invariant in the matrix metalloendoprotease family, suggest that these residues are zinc ligands. NOE data indicate that these histidines form two clusters: one ligates the catalytic zinc (His-201, -205, and -211), and the other ligates a structural zinc (His-151, -166, and -179). Heteronuclear multiple quantum correlated spectra and specific labeling experiments indicate His-151, -179, -201, -205, and -211 are in the N delta 1H tautomer and His-166 is in the N epsilon 2H tautomer.

DNA and redox state induced conformational changes in the DNA-binding domain of the Myb oncoprotein.

The DNA-binding domain of the oncoprotein Myb comprises three imperfect repeats, R1, R2 and R3. Only R2 and R3 are required for sequence-specific DNA-binding. Both are assumed to contain helix-turn-helix (HTH)-related motifs, but multidimensional heteronuclear NMR spectroscopy revealed a disordered structure in R2 where the second HTH helix was predicted [Jamin et al. (1993) Eur. J. Biochem., 216, 147-154]. We propose that the disordered region folds into a 'recognition' helix and generates a full HTH-related motif upon binding to DNA. This would move Cys43 into the hydrophobic core of R2. We observed that Cys43 was accessible to N-ethylmaleimide alkylation in the free protein, but inaccessible in the DNA complex. Mutant proteins with charged (C43D) or polar (C43S) side chains in position 43 bound DNA with reduced affinity, while hydrophobic replacements (C43A, C43V and C43I) gave unaltered or improved DNA-binding. Specific DNA-binding enhanced protease resistance dramatically. Fluorescence emission spectra and quenching experiments supported a DNA-induced conformational change. Moreover, reversible oxidation of Cys43 had an effect similar to the inactivating C43D mutation. The highly oxidizable Cys43 could function as a molecular sensor for a redox regulatory mechanism turning specific DNA-binding on or off by controlling the DNA-induced conformational change in R2.

Secondary structure of the DNA-binding domain of the c-Myb oncoprotein in solution. A multidimensional double and triple heteronuclear NMR study.

The DNA-binding domain of the c-Myb oncoprotein contains two repeats, R2 and R3, both of which have been proposed to be related to the helix-turn-helix (HTH) motif. As a first step towards determination of the three-dimensional structure of this domain and of the mode of interaction with the DNA, we have undertaken multidimensional heteronuclear NMR studies using uniformly 15N-labeled and 13C, 15N double-labeled R2R3 and, a selectively 15N-enriched sample on all lysine, histidine and leucine residues of R2R3. We present almost complete assignments of the backbone 1H, 15N and 13C" atoms and determine the secondary structure of R2R3 in solution. The R3 repeat is composed of three helices (residues 62-75, 78-85 and 91-100) while for the R2 repeat only two helices are found (residues 10-23 and 28-34). The remaining C-terminal part of the R2 repeat, predicted to be helical and part of the HTH motif, undergoes intermediate conformational exchange processes. Stabilization of this segment might occur upon binding to DNA.

The three-dimensional structure of acyl-coenzyme A binding protein from bovine liver: structural refinement using heteronuclear multidimensional NMR spectroscopy.

The 3D structure of bovine recombinant acyl-coenzyme A binding protein has been determined using multidimensional heteronuclear magnetic resonance spectroscopy in a study that combines investigations of 15N-labeled and unlabeled protein. The present structure determination is a refinement of the structure previously determined (Andersen, K.V. and Poulsen, F.M. (1992) J. Mol. Biol., 226, 1131-1141). It is based on 1096 distance restraints and 124 dihedral angle restraints of which 69 are for phi-angles and 8 for chiral centers and 47 for prochiral centers. The new experimental input for the structure determination has provided an increase of 263 distance restraints, 5 phi-angle restraints, and 32 chi-angle restraints in 2 chiral centers, and 31 prochiral centers restraining an additional 23 chi 1, 8 chi 2, and 1 chi 3 angles. The increase of 300 distance and dihedral angle restraints representing an additional 30% of input parameters for the structure determination has been shown to be in agreement with the first structure. A set of 29 structures has been calculated and each of the structures has been compared to a mean structure to give an atomic root mean square deviation of 0.44 +/- 0.12 A (1 A is 0.1 nm) for the backbone atoms C, C alpha, and N in the four alpha-helices A1, residues 4-15, A2, residues 21-36, A3, residues 51-62 and A4, residues 65-84. The loop-region of residues Gly45-Lys50 could not be defined by the restraints obtained by NMR. The program PRONTO has been used for the spectrum analysis, assignment of the individual nuclear Overhauser effects, the integration of the cross peaks, and the measurement of the coupling constants. The programs DIANA, X-PLOR and INSIGHT have been used in the structure calculations and evaluations.

NMR structure-based drug design

NMR is a useful tool for rapidly determining the conformations of receptor-bound ligands and identifying those portions of the ligand in contact with the receptor. In addition, the complete 3D structures of receptors and ligand/receptor complexes can be obtained using recently developed heteronuclear multi-dimensional NMR techniques. This NMR-derived structural information is potentially useful for aiding in The design of improved pharmaceutical agents. Approaches for utilizing the NMR-derived structural information along with the computational tools that facilitate this process are discussed.

Aliphatic 1H and 13C resonance assignments for the 26-10 antibody VL domain derived from heteronuclear multidimensional NMR spectroscopy.

Extensive 1H and 13C assignments have been obtained for the aliphatic resonances of a uniformly 13C- and 15N-labeled recombinant VL domain from the anti-digoxin antibody 26-10. Four-dimensional triple resonance NMR data acquired with the HNCAHA and HN(CO)CAHA pulse sequences [Kay et al. (1992) J. Magn. Reson., 98, 443-450] afforded assignments for the backbone HN, N, H alpha and C alpha resonances. These data confirm and extend HN, N and H alpha assignments derived previously from three-dimensional 1H-15N NMR studies of uniformly 15N-labeled VL domain [Constantine et al. (1992), Biochemistry, 31, 5033-5043]. The identified H alpha and C alpha resonances provided a starting point for assigning the side-chain aliphatic 1H and 13C resonances using three-dimensional HCCH-COSY and HCCH-TOCSY experiments [Clore et al. (1990), Biochemistry, 29, 8172-8184]. The C alpha and C beta chemical shifts are correlated with the VL domain secondary structure. The extensive set of side-chain assignments obtained will allow a detailed comparison to be made between the solution structure of the isolated VL domain and the X-ray structure of the VL domain within the 26-10 Fab.

Side chain and backbone assignments in isotopically labeled proteins from two heteronuclear triple resonance experiments

Two multidimensional heteronuclear NMR experiments are described for assigning the resonances in uniformly 15N- and 13C-labeled proteins. In one experiment (HCNH-TOCSY), the amide nitrogen and proton are correlated to the side-chain protons and carbons of the same and preceding residue. In a second triple resonance experiment (HC(CO)NH-TOCSY), the amide nitrogen and proton of one residue is correlated exclusively with the side-chain proton and carbon resonances of the preceding residue by transferring magnetization through the intervening carbonyl. The utility of these two experiments for making sequential resonance assignments in proteins is illustrated for [U- 15N, 13C]FKBP (107 residues) complexed to the immunosuppressant, ascomycin.

A practical method for uniform isotopic labeling of recombinant proteins in mammalian cells.

A method to obtain uniformly isotopically labeled (15N and 15N/13C) protein from mammalian cells is described. The method involves preparation of isotopically labeled media consisting of amino acids isolated from bacterial and algal extracts supplemented with cysteine and enzymatically synthesized glutamine. The approach is demonstrated by producing 15N-labeled and 15N/13C-labeled urokinase from Sp2/0 cells and successfully growing Chinese hamster ovary (CHO) cells on the labeled media. Thus, using the procedures described, isotopically labeled proteins that have been expressed in mammalian cells can be prepared, allowing them to be studied by heteronuclear multidimensional NMR techniques.

Three-dimensional heteronuclear NMR studies of RNA.

Multidimensional heteronuclear NMR has revolutionized solution structure determinations of proteins. But this technique has not been applied to nucleic acids because of difficulties in the synthesis of isotopically (13C and/or 15N) labelled molecules. Here we report the application of three-dimensional heteronuclear NMR to the study of a uniformly 13C/15N or 15N-labelled RNA duplex of defined sequence. These experiments simplify resonance assignment and the analysis of proton-proton nuclear Overhauser effects (and therefore distance information) in the molecule. Our results show that it is now possible to determine the structures of larger and more complex RNAs using multidimensional heteronuclear NMR.

Secondary structure and side-chain proton and carbon-13 resonance assignments of calmodulin in solution by heteronuclear multidimensional NMR spectroscopy

Heteronuclear 2D and 3D NMR experiments were carried out on recombinant Drosophila calmodulin (CaM), a protein of 148 residues and with molecular mass of 16.7 kDa, that is uniformly labeled with 15N and 13C to a level of greater than 95%. Nearly complete 1H and 13C side-chain assignments for all amino acid residues are obtained by using the 3D HCCH-COSY and HCCH-TOCSY experiments that rely on large heteronuclear one-bond scalar couplings to transfer magnetization and establish through-bond connectivities. The secondary structure of this protein in solution has been elucidated by a qualitative interpretation of nuclear Overhauser effects, hydrogen exchange data, and 3JHNH alpha coupling constants. A clear correlation between the 13C alpha chemical shift and secondary structure is found. The secondary structure in the two globular domains of Drosophila CaM in solution is essentially identical with that of the X-ray crystal structure of mammalian CaM [Babu, Y., Bugg, C. E., & Cook, W.J. (1988) J. Mol. Biol. 204, 191-204], which consists of two pairs of a "helix-loop-helix" motif in each globular domain. The existence of a short antiparallel beta-sheet between the two loops in each domain has been confirmed. The eight alpha-helix segments identified from the NMR data are located at Glu-6 to Phe-19, Thr-29 to Ser-38, Glu-45 to Glu-54, Phe-65 to Lys-77, Glu-82 to Asp-93, Ala-102 to Asn-111, Asp-118 to Glu-127, and Tyr-138 to Thr-146. Although the crystal structure has a long "central helix" from Phe-65 to Phe-92 that connects the two globular domains, NMR data indicate that residues Asp-78 to Ser-81 of this central helix adopt a nonhelical conformation with considerable flexibility.

Uniform labeling of a recombinant antibody Fv-fragment with 15N and 13C for heteronuclear NMR spectroscopy

The expression of functional antibody fragments in Escherichia coli enables a detailed analysis by NMR spectroscopy. This is demonstrated with the uniform labeling of an Fv-fragment (25 kDa) comprising the antigen binding site of an anitbody against 2-phenyloxazolone with 15N and 13C. The antigen-complexed Fv-fragment was analysed for a potential assignment by heteronuclear multi-dimensional NMR spectroscopy. For almost all backbone amides 15N/ 1H crosspeaks and for 80% of them TOCSY crosspeaks were observed. In a 13C-edited-HCCH-2D experiment 17 out of 18 threonine spin-systems were identified. Thus detailed assignments are possible, but some amino acid specific labeling in addition to uniform labeling will be required for complete assignments of Fv-fragments.