Rice University

Department of Biochemistry and Cell Biology


Kevin R. MacKenzie

Laboratory Members

Publications

Research

Address

3D HNCACB spectrum

Backbone resonances in proteins can be unambiguously assigned to the amino acid sequence using methods that correlate resonances from adjacent residues using J coupling (through bonds). These pulse sequences require doubly labeled (15N,13C) samples and the ability to pulse three RF channels at once.  The correlations available from one such experiment, the HNCACB, are indicated schematically below.  The HNCACB resolves amide proton/nitrogen correlations in the same fashion as the 2D 1H-15N HSQC.  The third dimension of the spectrum contains the 13C chemical shifts of the Ca and Cb resonances of a given residue and of the residue before it in the sequence.  The experiment exploits the fact that each amide nitrogen is coupled to its own Ca (1Jcn ~ 12 Hz) more strongly than to the Ca behind it (2Jcn ~ 7 Hz).  As a result, the strip for each N-H contains intense peaks for its own Ca and Cb resonances and weaker peaks for the Ca and Cb resonances of the previous residue; this is indicated schematically above by the heavily and lightly dotted lines.  The way that magnetization is transferred between Ca and Cb causes the correlations for these carbons to have opposite signs; this is indicated by the use of black and red lines.  Aligning the weak peaks in one strip with the strong peaks from another establishes that the corresponding N-H pairs are adjacent in the sequence. 

Using only the Ca correlations, some of these strips can be placed in the correct order, but overlap of the strong and weak peaks in two strips interrupt the alignment.  Distinguishing between isoleucines of such similar Ca shifts is not possible in this spectrum.  Happily, this ambiguity is resolved by looking at the Cb correlations.  Neither Ca nor Cb correlations suffice by themselves, but looking at both sets allows the correct order to be deduced.  Characteristic chemical shifts can then be used to map the strips onto the protein sequence.

The strip for A82 contains correlations from another N-H pair just as in the 3D NOESY-HSQC strip.  The potential effect of overlap in this spectrum is to confuse the assignment of the backbone resonances.