Search

Empirical shielding surfaces are most commonly used to predict chemical shifts in proteins from known backbone torsion angles, phi and psi. However, the prediction of (15)N chemical shifts using this technique is significantly poorer, compared to that for the other nuclei such as (1)H(alpha), (13)C(alpha), and (13)C(beta). In this study, we investigated the effects from the preceding residue and the side-chain geometry, chi(1), on (15)N chemical shifts by statistical methods. For an amino acid sequence XY, the (15)N chemical shift of Y is expressed as a function of the amino acid types of X and Y, as well as the backbone torsion angles, phi and psi(i-1). Accordingly, 380 empirical 'Preceding Residue Specific Individual (PRSI)' (15)N chemical shift shielding surfaces, representing all the combinations of X and Y (except for Y=Pro), were built and used to predict (15)N chemical shift from phi and psi(i-1). We further investigated the chi(1) effects, which were found to account for differences in (15)N chemical shifts by approximately 5 ppm for amino acids Val, Ile, Thr, Phe, His, Tyr, and Trp. Taking the chi(1) effects into account, the chi(1)-calibrated PRSI shielding surfaces (XPRSI) were built and used to predict (15)N chemical shifts for these amino acids. We demonstrated that (15)N chemical shift predictions are significantly improved by incorporating the preceding residue and chi(1) effects. The present PRSI and XPRSI shielding surfaces were extensively compared with three recently published programs, SHIFTX (Neal et al., 2003), SHIFTS (Xu and Case, 2001 and 2002), and PROSHIFT (Meiler, 2003) on a set of ten randomly selected proteins. A set of Java programs using XPRSI shielding surfaces to predict (15)N chemical shifts in proteins were developed and are freely available for academic users at http://www.pronmr.com or by sending email to one of the authors Yunjun Wang (yunjunwang@yahoo.com).

In this study, we report nearest neighbor residue effects statistically determined from a chemical shift database. For an amino acid sequence XYZ, we define two correction factors, Delta((X)Y)n,s and Delta(Y(Z))n,s, representing the effects on Y's chemical shifts from the preceding residue (X) and the following residue (Z), respectively, where X, Y, and Z are any of the 20 naturally occurring amino acids, n stands for (1)H(N), (15)N, (1)H(alpha), (13)C(alpha), (13)C(beta), and (13)C' nuclei, and s represents the three secondary structural types beta-strand, random coil, and alpha-helix. A total of approximately 14400 Delta((X)Y)n,s and Delta(Y(Z))n,s, representing nearly all combinations of X, Y, Z, n, and s, have been quantitatively determined. Our approach overcomes the limits of earlier experimental methods using short model peptides, and the resulting correction factors have important applications such as chemical shift prediction for the folded proteins. More importantly, we have found, for the first time, a linear correlation between the Delta((X)Y)n,s (n = (15)N) and the (13)C(alpha) chemical shifts of the preceding residue X. Since (13)C(alpha) chemical shifts of the 20 amino acids, which span a wide range of 40-70 ppm, are largely dominated by one property, the electron density of the side chain, the correlation indicates that the same property is responsible for the effect on the following residue. The influence of the secondary structure on both the chemical shifts and the nearest neighbor residue effect are also investigated.

The general theory of the Linderstrøm-Lang model, which, in simplified form, is widely used for the interpretation of NMR data on backbone-proton exchange in proteins, is systematically discussed. An experimental protocol for testing the applicability of the customary simplifications is described and experimental data which require the application of the general, rather than of the simplified, theory are presented. The appearance of pH-dependent biexponential magnetization recovery for any amide group in a protein is an unequivocal indication that the conventional simplified versions of the model do not apply.

We tested the dependence of the accuracy and precision of calculated NMR structures on the errors of the distance constraints using sequential simulated annealing and found that: (1) the accuracy of the family of structures depends mainly on the quality of the data, but is no better than about 1 A even if the errors in distance constraints are smaller than +/- 1 A. (2) The precision of the calculated structures, on the other hand, is nearly insensitive to the quality of the data. With present methods, the accuracy of NMR structures is at best of the order of 1 to 2 A, although a precision of 0.4 to 0.7 A is readily attainable. Comparisons with recent studies of this problem also brought out the importance of distinguishing between correct and incorrect definitions of accuracy when reporting numerical estimates. Using an incorrect definition of the term accuracy can lead to an artificially favorable estimate of its numerical value.

Modern NMR has revitalized the study of protein dynamics. Multidimensional spectra and the heteronuclear spectroscopy allow a substantial gain in resolution. Dynamics can be analyzed at individual sites and data on segmental and sequence-dependent flexibility are accumulating. This review summarizes the wide variety of NMR approaches for observing internal motions, including the folding processes, and the attempts to correlate dynamics to the biological activity of proteins. The implications of mobility on structure determination by NMR is also discussed.

The effects of selective deuteration on calculated NOESY intensities have been analyzed for the structure of the E. coli trp aporepressor, a 25 kDa protein. It is shown that selectively deuterated trp aporepressor proteins display larger calculated NOESY intensities than those for the same interproton distances in the natural abundance protein. The relatively larger magnetization transfer is demonstrated by a comparison of the NOE build-up curves for specific proton pairs, and for the calculated NOE intensities of short-range NOEs to backbone amide protons. This increase in intensity is especially pronounced for the NHi-NHi+1 cross peaks in the alpha-helical regions, and particularly for amide protons of two sequential deuterated residues. The effect is shown to be further intensified for longer mixing times. It is also shown that in all cases, each amide proton exhibits stronger NOEs to its own side chain, with an enhanced effect for deuterated derivatives. This theoretical analysis demonstrates that an evaluation of the relative NOE intensities for different selectively deuterated analogs may be an important tool in assigning NMR spectra of large proteins. These results also serve as a guide for the interpretation of NOEs in terms of distances for structure calculations based on data using selectively deuterated proteins.

Non-crystallographic approaches to the determination of protein structure must solve the problem of insufficient and low information content experimental data. Most successful methods augment experimentation with theoretical constraints (for example, potential energy functions or optimization error metrics). We believe it is important to separate the contributions of experimentation and theory in the construction of protein structure. The PROTEAN system defines protein topology on the basis of experimental data alone. Its performance on three data sets, derived from the lac-repressor headpiece of E. coli, sperm whale myoglobin, and domain 1 of bacteriophage T4 lysozyme, indicates that there may be families of related conformations that are consistent with the experimental data. These conformations provide insight into the strengths and weaknesses in the data sets. They also provide a set of structures with which to begin theoretical refinements. We outline here a strategy which maintains a clear distinction between refinements based on theory and those based on experiment, and thus allows a careful analysis of the properties of such refinement methods.

Nuclear Overhauser enhancement (NOESY) spectra were theoretically generated by solving the generalized Bloch equations with the appropriate initial conditions. The input to the equations were the coordinates of the protons of two similar crystal structures of basic pancreatic trypsin inhibitor. The two NOESY spectra obtained were compared to published experimental spectra of the protein in solution. It was found that the two crystal structures of basic pancreatic trypsin inhibitor give different theoretical spectra. The solution of the Bloch equations is very sensitive to small variations in the distance between protons (approx. 0.2 A), and to differences in the surrounding configurations. The method allows a detailed comparison of the crystal and solution structures of proteins. The structure of the trypsin inhibitor in solution was found to be similar to either one or the other crystal forms in different regions of the molecule.