Csd Chemical Shift Deviation

9.0 8.0 7.0 6.0 5.0 4.0 3.0 2.0 1.0 ppm protein because of the regular repeating units and the multiple copies of identical monomers in a biopolymer. The chemical-shift dispersion due to specific environments in a protein makes sequence-specific assignment possible. Some proteins are more difficult than others due to the lack of chemical-shift dispersion even in a fully folded structure. A sequence with few or no aromatic amino acids will be a challenge from an NMR perspective because the aromatic rings are a major source of chemical-shift dispersion. A protein with primarily a-helix secondary structure is more difficult than a ^-sheet protein because the ^-sheet produces larger deviations in chemical shift.

12.5.2 Secondary Structure

The variation in chemical shifts for the multiple copies of a particular amino acid is the result of two factors: one is the essentially random effect of aromatic rings oriented in a precise relationship to the proton in question (Chapter 2, Figure 2.15 ), shifting it upfield (shielding region above and below the ring) or downfield (deshielding region in the plane of the ring). This is a result of the tertiary structure of the protein: the precise three-dimensional folding of the polypeptide backbone and all of the side chains. The second factor is the relative position of the backbone carbonyl groups of the peptide bonds. This is also a through-space (anisotropic) effect due to unsaturation, but it correlates with certain common medium-range folding motifs in proteins: the a-helix (Fig. 12.12) and the (antiparallel) ^-sheet (Fig. 12.13). In the figures, hydrogen bonds are shown as dotted lines. These motifs are called secondary structure, and they constitute the structural building blocks of proteins. Note that in the ^-sheet the carbonyls and the NH groups alternate direction along the strand,


Figure 12.12

whereas in the a-helix all carbonyls point in one direction and all NH groups point in the opposite direction. These repeating geometric arrangements lead to a correlation between secondary structure and chemical shift. Both the HN and the Ha chemical shifts move downfield from their random coil values when the residue is in a ^-sheet structure, and upfield when it is in an a-helix. For example, of the seven Lys residues in Heregulin-a EGF domain (Fig. 12.11), the HN shift of K35 (9.6) is shifted far downfield from the lysine random-coil value (8.4) and the HN shift of K11 is shifted far upfield (7.6 ppm). In fact, K35 is located in a ^-sheet and K11 is part of an a-helical portion of the protein structure. The same patterns are seen for the Ha protons: Ha of F40 (5.6) is shifted downfield from the phenylalanine random-coil Ha value (4.7), and Ha of F13 (4.1) is shifted upfield. In the 3D structure, F40 is in a ^-sheet and F13 is in the single a-helix. Note that we have to compare sequence-specific chemical shifts to the random-coil shifts for the same amino acid because there is some dependence on the amino acid type even when there is no specific conformation (Fig. 12.9). This predictive tool can be formalized by subtracting the random-coil shift from the actual shift for each residue to obtain the "chemical-shift deviation" (CSD) for that residue. For example, the Ha CSD is -0.6 for F14 in Heregulin-a EGF domain (4.1-4.7 ppm). To simplify the prediction even further, one can define a "chemical-shift index," which is +1 if the CSD is greater than or equal to 0.1 ppm, —1 if CSD is less than or equal to —0.1, and zero if it is between +0.1 and —0.1. A bar graph of these CSI indicators versus the residue number will show a string of +1 values for a strand of a ^-sheet and a string of —1 values for an a-helix. Thus with nothing more than sequence-specific assignments and a list of chemical shifts, we can begin to identify the secondary structure building blocks of the 3D protein structure. Figure 12.14 shows the CSI values for Heregulin-a EGF domain with the ^-strands and a-helix identified from the final 3D structure. A fairly long "run" of negative CSI values (residues 5-13) even with a few gaps, identifies an a-helix. Long runs of positive CSI values (residues 19-24 and

32-37) indicate extended strands of a ^-sheet. Blank regions (25-31 and 54-61) may be unstructured or at least lacking in regular secondary structure. Because we are ignoring the "random" factors—the orientation of nearby aromatic rings—which can significantly affect chemical shifts and which have no relationship to secondary structure, the CSI is only a broad indicator and will not correlate perfectly with the final structure. Still, it is extremely useful and requires nothing more than sequence-specific assignments.

Ideally, every proton in the protein structure would have a unique chemical shift, different from every other proton. In NMR, we can only identify a proton by its chemical shift, so if two protons have the same chemical shift all of the information associated with them—NOE distances and dihedral angles—becomes ambiguous. One way to overcome this "overlap" problem is to measure the chemical shifts of heavy atoms: nitrogen and carbon. If we replace every N and C (normally 14N and 12C) with spin-^ 15N and 13C, we can now distinguish two protons with identical chemical shifts by the (usually) different chemical shifts of the carbon or nitrogen they are connected to. This has led to the possibility of studying larger proteins with a complex array of new experiments in which magnetization is "tossed around" between 1H, 15N, and 13C nuclei. These experiments, mostly 3D, are called "triple-resonance" experiments because they include pulses on all three of these nuclei.

0 0

Post a comment