Noe Interactions Between One Residue And The Next Residue In The Sequence

Using DQF-COSY and TOCSY we can link all of the protons within a single spin system, which corresponds to a single amino acid residue. We can classify each spin system as a pattern of chemical shifts unique to one amino acid or as a member of a class: AMX or five spin. In order to get sequence-specific assignments, however, we have to have some way to correlate protons in one residue to protons in the next residue in the sequence. For unlabeled proteins this is done by NOE interactions: certain protons in one residue are constrained by the peptide bond to be close in space to certain protons in the next residue. These NOE correlations are called sequential or "i, i + 1" because they correlate a proton in residue "i" with a proton in the next residue in the sequence, residue "i + 1." Specifically, we expect to see NOE correlations between Ha of residue i and HN of residue i + 1 (Fig. 12.15) and sometimes between the H protons of residue i and the HN of the next residue. Because the DQF-COSY and TOCSY spectra correlate protons within a residue, we can move from


Figure 12.16

Hn of residue i + 1 to Ha of residue i + 1 via the J coupling. This sets us up for the next sequential NOE "jump" from the Ha of residue i + 1 to the HN of residue i + 2 (Fig. 12.16). This is similar to the "walk" along a carbon chain using COSY data, except that now we alternate between NOE interactions (2D-NOESY) crossing the peptide bond (a dead end for J couplings) and J couplings (2D-COSY) to move to the next residue's Ha. Using this strategy we can "walk" along the polypeptide backbone and directly "read off" the chemical shifts of each Ha and HN, assigning them to specific residues in the protein. These sequential NOEs, known as "a,N" and " j,N" NOEs, are directional and will not be seen in the other direction: Ha (i) ^ HN (i - 1). The geometry of the peptide bond only brings the Ha into proximity with the next residue's HN (i ^ i + 1).

Sequential a,N and j,N NOEs are commonly very strong for extended conformations such as the extended strand of a j sheet (Fig. 12.16). The Ha of residue i is clearly very close to the HN of residue i + 1, but the HN of residue i is farther away from HN of residue i + 1 (the N-H vector points in the opposite direction), so the sequential "N,N" NOE is weak or missing in this conformation (Fig. 12.15). In contrast, the a-helix conformation (Fig. 12.17) orients all of the N-H vectors in the same direction so that the HN of residue i is now close to the HN of residue i + 1 in space. Within an a-helix we can "walk" along the peptide backbone directly using only the NOESY spectrum, using J-coupling information (COSY and TOCSY) to link each HN chemical shift to a specific spin system associated with an amino acid (unique pattern) or amino acid category (AMX or five spin). In the a-helical conformation the Ha of residue i is farther away from the HN of residue i + 1 (Fig. 12.17), so the a,N (and j,N) sequential NOEs are weak or missing. This is a nice consequence of the secondary structure: j -sheet regions will give strong sequential a,N crosspeaks in the NOESY spectrum and weak or missing N, N(HN ^ HN+1) crosspeaks, whereas a-helical regions will give strong sequential N,N crosspeaks and weak or missing a,N crosspeaks. It is always an added bonus if the primary sequential NOE observed (e.g., a,N in a j -sheet) can be confirmed with a weak sequential NOE of the other type (e.g., N,N).

Figure 12.18 shows a portion of the F2 = HN, F\ = Ha region of the NOESY spectrum of B. subtilis HPr, a phosphotransferase that uses an active-site histidine side chain to transfer a high-energy phosphate group. A large number of sequential Ha (i) ^ HN (i + 1) NOE crosspeaks are shown. The HN chemical shifts are indicated by vertical dotted

Figure 12.17

0 0

Post a comment