Dependence of amino acid side chain 13C shifts on dihedral angle: application to conformational analysis

Abstract

Chemical shift data from the BiomagResDataBank and conformational data derived from the protein data bank have been correlated in order to explore the conformational dependence of side chain (13)C resonance shifts. Consistent with predictions based on steric compression, upfield shifts for Cgamma resonances of Thr, Val, Ile, Leu, Met, Arg, Lys, Glu, and Gln residues correlate with both the number of heavy atom (nonproton) gamma-substituents and with gauche conformational orientations of gamma-substituents. The (13)C shift/conformation correlations are most apparent for Cgamma carbons but also can be observed at positions further from the backbone. Intraresidue steric conflict leads to a correlation between upfield-shifted side chain (13)C resonances and statistically lower probabilities in surveys of protein side chain conformation. Illustrative applications to the DNA pol lambda lyase domain and to dihydrofolate reductase are discussed. In the latter case, (13)C shift analysis indicates that the conformation of the remote residue V119 on the betaF-betaG loop is correlated with the redox state of the bound pyridine nucleotide cofactor, providing one basis for discrimination between substrate and product. It is anticipated that (13)C shift data for protein sidechains can provide a useful basis for the analysis of conformational changes even in large, deuterated proteins. Additionally, the large dependence of the leucine methyl shift difference, deltaCdelta1-deltaCdelta2, on both chi1 and chi2 is sufficient to allow this parameter to be used as a restraint in structure calculations if stereospecific assignment data are available.

Figure 1

(a) Three stable rotameric conformations of threonine. The significant steric interactions are indicated by red arrows. (b) Threonine ¹³Cγ shifts corresponding to the three stable rotameric states. Shift data were included for ±20° ranges.

Figure 2

Analysis of leucine methyl shifts in malate synthase G. (A) Variation of ΔCδ₁₂ with χ2 and (b) correlation with χ1. As is apparent from the figure, most of the observed conformations correspond to χ1 = -60°, χ2 = 180°, with a smaller proportion corresponding to χ1 = 180°, χ2 = 60°. Data were analyzed for 54 leucine residues for which both NMR shift data and crystallographic χ1/χ2 data were available.

Figure 3

(a) ¹H-¹³C HSQC spectrum of [ϵ-¹³C]lysine polymerase λ lyase domain (residues 242-327) obtained on a Unity INOVA 800. Three of the residues, K287, K291, and K312, show inequivalent ϵ protons under the conditions of the study. Other sample conditions were as follows: 100 mM KCl, 2 mM deuterated DTT, 0.1% NaN₃ (w/v), 90% H₂O and 10% ²H₂O (v/v), and 50 μM DSS, used as a chemical shift standard. (b) Structural data showing salt bridge interactions for K287-E261 and K291-E298, taken from structure 1RZT.

Figure 4

Structure of E. coli dihdyrofolate reductase showing the Met20 and βF-βG loops. (a) Structure 1RA2 corresponding to the DHFR•Folate•NADP⁺ complex. The Met20 loop is indicated in cyan, and the βF-βG loop, in orange. NADP⁺ is shown in magenta, and folate, in yellow. Sidechains for V119 on the βF-βG loop and several residues in the Met20 loop are also explicitly indicated. (b) Superposition of active site loops in structures 1RX2 (corresponding to a ternary DHFR•Folate•NADP⁺ complex in which the Met20 loop adopts a closed conformation) and 1RX7 (corresponding to a binary DHFR•Folate complex with the Met20 loop in an occluded conformation). Both structures show residues 9-24 of the Met 20 loop and 117-122 of the βF-βG loop, as well as the sidechains for P21 and V119. The proximity of the P21-V119 sidechains is significantly greater in the closed conformation of the Met20 loop, which is presumably the basis for the observed shift perturbation and corresponding conformational change of V119.