An evaluation of the combined use of nuclear magnetic resonance and distance geometry for the determination of protein conformations in solution

An evaluation of the potential of nuclear magnetic resonance (n.m.r.) as a means of determining polypeptide conformation in solution is performed with the aid of a new distance geometry program which is capable of computing complete spatial structures for small proteins from n.m.r. data. Ten sets of geometric constraints which simulate the results available from n.m.r. experiments of varying precision and completeness were extracted from the crystal structure of the basic pancreatic trypsin inhibitor, and conformers consistent with these constraints were computed. Comparison of these computed structures with each other and with the original crystal structure shows that it is possible to determine the global conformation of a polypeptide chain from the distance constraints which are available from n.m.r. experiments. The results obtained with the different data sets also provide a standard by which the quality of protein structures computed from n.m.r. data can be evaluated when no crystal structure is available, and indicate directions in which n.m.r. experiments for protein structure determination could be further improved.     

Nuclear magnetic resonance solution and X-ray structures of squash trypsin inhibitor exhibit the same conformation of the proteinase binding loop

A comparison of the solution nuclear magnetic resonance (n.m.r.) structures of squash trypsin inhibitor from seeds of the squash Cucurbita maxima with the X-ray structure of a trypsin complex of the inhibitor shows that the n.m.r. and X-ray structures are similar in terms of the global folding and secondary structure. The average atomic root-mean-square difference between the 36 n.m.r. structures on the one hand and the X-ray structure is 0.96 A for the backbone atoms and 1.95 A for all heavy atoms. The n.m.r. and X-ray structures exhibit extremely similar conformations of the primary proteinase binding loop. Despite the overall similarity, there are small differences between the mean computed structure and the X-ray structure. The n.m.r. structures have slightly different positions of the segments from residues 16 to 18, and 24 and 25. The n.m.r. results show that the inclusion of stereospecific assignments and precise distance constraints results in a significant improvement in the definition of the n.m.r. structure, making possible a detailed analysis of the local conformations in the protein.     

Experimental determination of torsion angles in the polypeptide backbone of the gramicidin A channel by solid state nuclear magnetic resonance.

An analytical method for the determination of torsion angles from solid state 15N nuclear magnetic resonance (n.m.r.) spectroscopic data is demonstrated. Advantage is taken of the 15N-1H and 15N-13C dipolar interactions as well as the 15N chemical shift interaction in oriented samples. The membrane-bound channel conformation of gramicidin A has eluded an atomic resolution structure determination by more traditional approaches. Here, the torsion angles for the Ala3 site are determined by obtaining the n.m.r. data for both the Gly2-Ala3 and Ala3-Leu4 peptide linkages. Complete utilization of the orientational constraints derived from these orientation-dependent nuclear spin interactions in restricting the conformational space is most effectively achieved by utilizing spherical trigonometry. Two possible sets of torsion angles for the Ala3 site are obtained (phi, psi = -129 degrees, 153 degrees and -129 degrees, 122 degrees), both of which are consistent with a right-handed beta-helix. Other functional and computational evidence strongly supports the set for which the carbonyl oxygen atom of the Ala3-Leu4 linkage is rotated into the channel lumen.     

Nuclear magnetic resonance study of the solution structure of alpha 1-purothionin. Sequential resonance assignment, secondary structure and low resolution tertiary structure

The solution structure of the 45-residue plant protein, alpha 1-purothionin, is investigated by nuclear magnetic resonance (n.m.r.) spectroscopy. Using a combination of two-dimensional n.m.r. techniques to demonstrate through-bond and through-space (less than 5 A) connectivities, the 1H n.m.r. spectrum of alpha 1-purothionin is assigned in a sequential manner. The secondary structure elements are then delineated on the basis of a qualitative interpretation of short-range nuclear Overhauser effects (NOE) involving the NH, C alpha H and C beta H protons. There are two helices extending from residues 10 to 19 and 23 to 28, two short beta-strands from residues 3 to 5 and 31 to 34 which form a mini anti-parallel beta-sheet, and five turns. In addition, a number of long-range NOE connectivities are assigned and a low resolution tertiary structure is proposed.     

Human interleukin 4. The solution structure of a four-helix bundle protein.

Heteronuclear 13C and 15N three-dimensional nuclear magnetic resonance (n.m.r.) techniques have been used to determine the solution structure of human interleukin 4, a four-helix bundle protein. A dynamical simulated annealing protocol was used to calculate an ensemble of structures from an n.m.r. data set of 1735 distance restraints, 101 phi angle restraints and 27 pairs of hydrogen bond restraints. The protein structure has a left-handed up-up-down-down topology for the four helices with the two long overhand loops in the structure being connected by a short section of irregular antiparallel beta-sheet. Analysis of the side-chains in the protein shows a clustering of hydrophobic residues, particularly leucines, in the core of the bundle with the side-chains of charged residues being located on the protein surface. The solution structure has been compared with a recent structure prediction for human interleukin 4 and with crystal structures of other helix bundle proteins.     

Protein structures in solution by nuclear magnetic resonance and distance geometry. The polypeptide fold of the basic pancreatic trypsin inhibitor determined using two different algorithms, DISGEO and DISMAN

A set of conformational restraints derived from nuclear magnetic resonance (n.m.r.) measurements on solutions of the basic pancreatic trypsin inhibitor (BPTI) was used as input for distance geometry calculations with the programs DISGEO and DISMAN. Five structures obtained with each of these algorithms were systematically compared among themselves and with the crystal structure of BPTI. It is clear that the protein architecture observed in single crystals of BPTI is largely preserved in aqueous solution, with local structural differences mainly confined to the protein surface. The results confirm that protein conformations determined in solution by combined use of n.m.r. and distance geometry are a consequence of the experimental data and do not depend significantly on the algorithm used for the structure determination. The data obtained further provide an illustration that long intramolecular distances in proteins, which are comparable with the radius of gyration, are defined with high precision by relatively imprecise nuclear Overhauser enhancement measurements of a large number of much shorter distances.     

The fine structure of two DNA dodecamers containing the cAMP responsive element sequence and its inverse. Nuclear magnetic resonance and molecular simulation studies.

1H and 31P n.m.r. (nuclear magnetic resonance) spectroscopy have been used in conjunction with molecular simulation to determine the structure of two DNA dodecamers. The first of these, CATGACGTCATG, contains the octameric sequence CRE (cAMP responsive element), while the second is the reversed sequence, GTACTGCAGTAC. Structure determination was based on both NOESY (nuclear Overhauser spectroscopy) derived distances and COSY (correlated spectroscopy) dihedral angle data. Access to the 31P spectra also allowed the epsilon backbone angles to be determined. Considerable care was taken in deriving structural parameters from the n.m.r. data and an excellent level of agreement is obtained with the simulated conformations. Both dodecamers are found to belong to the B-DNA family; however, there is a striking difference between the CRE sequence and its inverse, the former conformation alone showing a strong structural heterogeneity.     

Dihydrofolate reductase. 1H resonance assignments and coenzyme-induced conformational changes

Lactobacillus casei dihydrofolate reductase has been studied in solution by one and two-dimensional 1H nuclear magnetic resonance (n.m.r.) spectroscopy at 500 MHz. By using a combination of n.m.r. methods in conjunction with the crystal structure of the enzyme-methotrexate-NADPH complex, resonances have been assigned for 32 of the 162 residues of the enzyme. These are widely distributed throughout the structure of the protein, and include all the histidine and tyrosine residues, as well as several valine, leucine, isoleucine and phenylalanine residues. The assignments have been made for the enzyme-methotrexate and enzyme-methotrexate-NADP+ complexes as well as the enzyme-methotrexate-NADPH complex. Comparison of assigned resonances in the spectra of the three complexes has permitted a preliminary assessment of structural differences between them. The beta-sheet "core" of the protein is unaffected by coenzyme binding, but two regions of the structure that undergo coenzyme-induced conformation changes have been identified. These are the loop comprising residues 13 to 23, and alpha-helix C (residues 42 to 49).     

Determination of the nuclear magnetic resonance solution structure of the DNA-binding domain (residues 1 to 69) of the 434 repressor and comparison with the X-ray crystal structure.

The DNA-binding domain of the phage 434 repressor consisting of N-terminal residues 1 to 69 (434 repressor(1-69)), was expressed in Escherichia coli with natural isotope abundance, uniform 15N-labeling and biosynthetically directed fractional 13C-labeling in extent of about 10%. With these protein preparations the three-dimensional structure was determined in solution. The techniques used were nuclear magnetic resonance (n.m.r.) spectroscopy for the collection of conformational constraints, calculation of the protein structure from the n.m.r. data with the program DIANA and structure refinements by restrained energy minimization with a modified version of the program AMBER. A group of 20 conformers characterizes a well-defined structure for residues 1 to 63, with an average of 0-6 A for the root-mean-square deviations (RMSD) calculated for the backbone atoms of the individual conformers relative to the mean co-ordinates. The spatial structure of C-terminal residues 64 to 69 is not defined by the n.m.r. data. The molecular architecture of the 434 repressor(1-69) in solution includes five alpha-helices extending from residues 2 to 13, 17 to 24, 28 to 35, 45 to 52 and 56 to 60, which enclose a well-defined hydrophobic core. The n.m.r. structure is closely similar to the reported crystal structure of the 434 repressor(1-69), with an RMSD value of 1.1 A for the backbone atoms of residues 1 to 63. Small differences between the two structures in regions of the first helix and the loop between helices 3 and 4 were analyzed relative to possible correlations with protein-protein contacts in the crystal lattice and the different milieus of pH and ionic strength in the crystals and n.m.r. samples. Further systematic comparisons of local conformational features indicated that there are correlations between amino acid types, local precision of the structure determination by both techniques and local differences between the structures in the crystals and in solution. Overall, hydrophobic residues are most precisely characterized and agree most closely in the two environments.     

Molecular dynamics simulation of the DNA triplex d(TC)5.d(GA)5.d(C+T)5.

A molecular dynamics simulation of the DNA triple helix d(TC)5.d(GA)5.d(C+T)5 is described (C+ represents a protonated cytosine residue). The simulation has been performed using the program AMBER 3.1 and includes counterions and explicit solvent under periodic boundary conditions. Both the dynamic and time-averaged behaviour of the system has been analysed. Considerable deviations from the fibre-diffraction model for DNA triple helix structure are observed, including the repuckering of the purine strand sugars that has been identified in some nuclear magnetic resonance (n.m.r.) studies. The simulation suggests that this conformational change may be driven by the possibility of improved interactions between the phosphate groups of this strand and both the solvent and counterions. Several examples of a particular conformational transition are observed, involving correlated changes in the backbone angles alpha and gamma. These transitions provide a possible explanation for some unusual n.m.r. data that have been reported. The structure of the triple helix major groove also suggests an explanation for the observed stabilization of DNA triplexes by polyvalent cations, and their ability to interact with drugs that bind in the minor groove of DNA duplexes.     

Identification and description of beta-structure in horse muscle acylphosphatase by nuclear magnetic resonance spectroscopy

Nuclear magnetic resonance spectra of acylphosphatase were searched for signs of beta-structure, i.e. characteristic nuclear Overhauser enhancement patterns displayed in the two-dimensional spectra, typical chemical shifts, coupling constants and slow 2H-H exchange. The results provided identification of the main-chain resonances of amino acid residues involved in the beta-structure. The full sequential assignment of this region was gained by identification of some amino acid spin systems and their alignment with the primary sequence. The assignment of the side-chains was virtually completed subsequently and a list produced of nuclear magnetic resonance (n.m.r.) constraints derived from the spectra. The beta-structure consists of a beta-sheet with four antiparallel chains, one attached parallel chain, three tight turns and a beta-bulge. The conformation of the beta-sheet was determined by distance geometry calculation using the n.m.r. constraints (174 intraresidual, 107 sequential and 226 long-range distances, 32 torsion angles, phi, and 28 hydrogen bonds) as input. Observation of some interactions between the sheet and previously identified alpha-helical regions made it possible to give an outline of the three-dimensional structure of the enzyme.     

Folding of peptide fragments comprising the complete sequence of proteins. Models for initiation of protein folding. II. Plastocyanin.

In an attempt to understand the earliest events in the protein folding pathway, the complete sequence of French bean plastocyanin has been synthesized as a series of short peptide fragments, and the conformational preferences of each peptide examined in aqueous solution using proton n.m.r. methods. Plastocyanin consists largely of beta-sheet, with reverse turns and loops between the strands of the sheet, and one short helix. The n.m.r. experiments indicate that most of the peptides derived from the plastocyanin sequence have remarkably little propensity to adopt folded conformations in aqueous solution, in marked contrast to the peptides derived from the helical protein, myohemerythrin (accompanying paper). For most plastocyanin peptides, the backbone dihedral angles are predominantly in the beta-region of conformational space. Some of the peptides show weak NOE connectivities between adjacent amide protons, indicative of small local populations of backbone conformations in the a region of (phi,psi) space. A conformational preference for a reverse turn is seen in the sequence Ala65-Pro-Gly-Glu68, where a turn structure is found in the folded protein. Significantly, the peptide sequences that populate the alpha-region of (phi,psi) space are mostly derived from turn and loop regions in the protein. The addition of trifluoroethanol does not drive the peptides into helical conformations. In one region of the sequence, the n.m.r. spectra provide evidence of the formation of a hydrophobic cluster involving aromatic and aliphatic side-chains. These results have significance for understanding the initiation of protein folding. From these studies of the fragments of plastocyanin (this paper) and myohemerythrin (accompanying paper), it appears that there is a pre-partitioning of the conformational space sampled by the polypeptide backbone that is related to the secondary structure in the final folded state.     

Solution of the phase problem in the X-ray diffraction method for proteins with the nuclear magnetic resonance solution structure as initial model. Patterson search and refinement for the alpha-amylase inhibitor tendamistat

Patterson search calculations using the three-dimensional structure of the alpha-amylase inhibitor from Streptomyces tendae obtained from experimental nuclear magnetic resonance (n.m.r.) data were performed to study the possibility of solving the phase problem in the X-ray diffraction method with protein structures determined by n.m.r. Using all heavy atoms (C, N, O, S) of the residues 5 to 73 in the best n.m.r. structure of the alpha-amylase inhibitor (520 out of the 558 heavy atoms in the complete polypeptide chain), the maximum of the rotation function corresponded to the correct solution obtained by the previous independent determination of the crystal structure. However, additional local maxima, which are not significantly lower than the global maximum, also showed up. Performing the Patterson search with a model containing the backbone atoms and the heavy atoms of only the interior side-chains (399 atoms), which are much better defined by the n.m.r. data, the correct maximum was significantly higher than all other maxima. A translation search for the best orientation of the latter model yielded the correct solution. The energy-restrained crystallographic refinement was performed with this model to an R-factor of 26%. This corresponds approximately to the R-factor calculated for the X-ray crystal structure previously determined using the isomorphous replacement technique, if the residues 1 to 4 and 74 and all localized solvent molecules were removed from this structure. During the refinement the root-mean-square deviation between the two structures decreased from 1.03 A to 0.26 A for the polypeptide backbone and from 1.64 A to 0.73 A for all heavy atoms. There are no major local conformational differences between the two structures, with the single exception of the side-chain of Gln52.     

Determining local conformational variations in DNA. Nuclear magnetic resonance structures of the DNA duplexes d(CGCCTAATCG) and d(CGTCACGCGC) generated using back-calculation of the nuclear Overhauser effect spectra, a distance geometry algorithm and constrained molecular dynamics

Two-dimensional nuclear magnetic resonance (n.m.r.) spectroscopy and a variety of computational techniques have been used to generate three-dimensional structures of the two DNA duplexes d(CGCCTAATCG) and d(CGTCACGCGC). The central six base-pairs in these two decamers contain all ten dinucleotide pairs in DNA and thus, represent a model system for investigating how the local structure of DNA varies with base sequence. Resonance assignments were made for the non-exchangeable base protons and most of the C-1'-C-4' sugar protons in both decamers. Three-dimensional structures were generated using a distance geometry algorithm and these initial structures were refined by optimizing the fit of back-calculated spectra against the experimental two-dimensional nuclear Overhauser effect (NOE) spectra. This back-calculation procedure consists of calculating NOE cross relaxation rates for a given structure by solution of the Bloch equations, and directly accounts for spin diffusion effects. Use of this refinement procedure eliminates some assumptions that have been invoked when generating structures of DNA oligomers from n.m.r. data. Constrained energy minimization and constrained quenched molecular dynamics calculation were also performed on both decamers to help generate energetically favorable structures consistent with the experimental data. Analysis of the local conformational parameters of helical twist, helical rise, propeller twist, displacement and the alpha, beta, gamma, epison and zeta backbone torsion angles in these structures shows that these parameters span a large range of values relative to the X-ray data of nucleic acids. However, the glycosidic and pseudorotation angles are quite well defined in these structures. The implications that these results have for determination of local structural variations of DNA in solution, such as those predicted by Callidine's rules, are discussed. Our results differ significantly from some previous studies on determining local conformations of nucleic acids and comparisons with these studies are made.     

N.m.r. and biochemistry

The role and historical progress of n.m.r. applications in biochemistry are briefly outlined. Technical advances over the years have made n.m.r., at last, a technique which can give valuable information about a wide range of biochemical topics, from enzyme kinetics in vivo to the structure of protein-DNA complexes. Emphasis here is placed on studies of proteins, especially those made up from mosaics of modules. It is shown that n.m.r. can readily give detailed structural information about individual protein modules and that valuable information about the structure and function of the intact mosaic protein can be inferred.     

Prediction of protein folding pathways.

Recent 1H nuclear magnetic resonance (n.m.r.) hydrogen exchange experiments on five different proteins have delineated the secondary structures formed in trapped, partially folded intermediates. The early forming structural elements are identifiable through a technique described in this work to predict folding pathways. The method assumes that the sequential selection of structural fragments such as alpha-helices and beta-strands involved in the folding process is founded upon the maximal burial of solvent accessible surface from both the formation of internal structure and substructure association. The substructural elements were defined objectively by major changes in main-chain direction. The predicted folding pathways are in complete correspondence with the n.m.r. results in that the formed structural fragments found in the folding intermediates are those predicted earliest in the pathways. The technique was also applied to proteins of known tertiary structure and with fold similar to one of the five proteins examined by 1H n.m.r. The pathways for these structures also showed general consistency with the n.m.r. observations, suggesting conservation of a secondary structural framework or molten globule about which folding nucleates and proceeds.     

The folding of an enzyme. V. H/2H exchange-nuclear magnetic resonance studies on the folding pathway of barnase: complementarity to and agreement with protein engineering studies.

Two major methods are currently being used to characterize transient intermediates during protein folding at the level of individual residues. Nuclear magnetic resonance (n.m.r.) measurements on the protection of peptide NH hydrogens against exchange with solvent during refolding can provide information about secondary structure formation. Protein engineering and kinetics can provide direct information about intramolecular interactions of protein side-chains and indirect evidence on secondary structure. These procedures have provided the most complete pictures so far about protein folding intermediates. Both methods have been applied to the characterization of an intermediate in the refolding of barnase. Although the two methods give complementary information, there are some regions of the protein where the methods overlap well. We show that, with one possible exception that is obscure, n.m.r. and protein engineering give identical results for those interactions that can be analysed by both methods. This suggests that these are valid approaches for the study of protein folding intermediates in the case of barnase and that the combination of the methods is a powerful analytical procedure. Information provided by n.m.r. data that is complementary to the protein engineering experiments is: (1) early formation of the C terminus of helix2; (2) early formation of helix3; (3) early formation of several beta-turns (46-49, 101-104 in loop5); and (5) partial formation of loop5. Confirmatory evidence of protein engineering data on the intermediate is: (1) helix1 is complete from residues 10 to 18; (2) the interactions between all beta-strands are present; (3) part of loop2 is not formed; (4) part of loop3 is formed; and (5) some specific tertiary interactions are not made. For some interactions the protein engineering and H/2H exchange methods overlap directly. The information obtained for direct overlap is self consistent.     

Glycera dibranchiata hemoglobin. Structure and refinement at 1.5 A resolution

The coelomic cells of the common marine bloodworm Glycera dibranchiata contain several hemoglobin monomers and polydisperse polymers. We present the refined structure of one of the Glycera monomers at 1.5 A resolution. The molecular model for protein and ordered solvent for the deoxy form of the Glycera monomer has been refined to a crystallographic R-factor of 12.7% against an X-ray diffraction dataset at 1.5 A resolution. The positions of 1095 protein atoms have been determined with a maximum root-mean-square (r.m.s.) error of 0.13 A, and the r.m.s. deviation from ideal bond lengths is 0.015 A and from ideal bond angles is 1.0 degree. The r.m.s. deviation of planar groups from their least-squares planes is 0.007 A, and the r.m.s. deviation for torsion angles is 1.2 degrees for peptide groups and 16.8 degrees for side-chains. A total of 153 water molecules has been located, and they have been refined to a final average occupancy of 0.80. Multiple conformations have been found for five side-chains, and a change has been suggested for the sequence at five residues. The heme group is present in the "reverse" orientation that differs only in the positions of the vinyl beta-carbons from the "normal" orientation. The doming of the heme towards the proximal side, and the bond distances and angles of the heme and proximal histidine are typical of most deoxy globin structures. The substitution of leucine for the distal histidine residue (E7) creates an unusually hydrophobic heme pocket.     

Computing the geometry of a molecule in dihedral angle space using n.m.r.-derived constraints. A new algorithm based on optimal filtering.

We have developed a method based on optimal filtering to determine the three-dimensional structure of a protein from n.m.r.-derived constraints, using the dihedral angle internal representation of the molecule. It differs from currently proposed methods in that it directly produces estimates of errors on the parameters that are refined, hence providing an image of the minimum that has been found. A similar algorithm had already been proposed using cartesian co-ordinates as independent parameters, encoded in PROTEAN2. We found that using dihedral angles significantly reduces the computational burden of the technique, and provides better control over a priori informations that can be used, such as geometric restrictions for proline residues and informations from vicinal coupling constants. Performance of the method, encoded in FILMAN, is demonstrated by application to the folding of a ten-residue alanine polypeptide, to the geometric cyclization of an 11-residue peptide, as well as on the folding of a medium size protein, i.e. tendamistat. The validity of the error estimates on the dihedral angles produced by FILMAN is discussed.