|
| Chazin Home | Ca-binding Protein DB | Vanderbilt Home |
|
| Research Description | Publications | Wisdom | Search |
Barbara C.M. Potts, Göran Carlström, Katsuo Okazaki, Hiroyoshi Hidaka
and Walter J. Chazin
Protein Science 5, 2162-2174 (1994) Abstract
Sequential Assignment Strategy. The procedure used to obtain the sequence-specific 1H NMR resonance assignments involves identification of the system of resonances (spin system) for each amino acid followed by sequential assignment of these spin systems (reviewed in Wüthrich (1986)). The short-hand notation of Wüthrich et al. (1984) is adopted to identify proton-proton distances. Amino acid spin systems were identified by the strategy described in Chazin et al. (1988). In this approach, an emphasis is placed on scalar correlation experiments acquired from H2O solution, with the aim of observing relayed connectivities from the sidechain protons to the backbone amide proton (Chazin & Wright, 1987). Relayed connectivities to CaH and protons of the sidechain termini (e.g. leucine CdH3) provide supplementary information to complete the assignments for those spin systems that cannot be identified solely from relayed connectivities to the backbone amide proton. Multiple quantum (MQ) and MQ-filtered COSY spectra are then utilized to confirm these assignments and unambiguously distinguish the protons of each sidechain. MQ spectra are also used to unambiguously identify instances of resonance degeneracy.
Certain modifications to the basic strategy were required in this study due to the sensitivity limitations associated with the limited quantity of sample and the large resonance linewidths. Thus, the efficiency of relayed coherence transfer experiments was not as high as has been observed for polypeptides of similar length. As a result, the sequential resonance assignment procedure (Billeter et al., 1982) was begun when less than half of the amide based spin systems had been classified. The assignment process thus involved several iterations through the scalar correlated and NOESY spectra. Although the spin system identification steps were not as complete as in some of the previous studies in this laboratory, resonance assignments were made for all but one backbone proton and some of the sidechain protons of the 5-spin and lysine residues.
Spin System Identification. For a 90 amino acid protein containing one proline, crosspeaks from 88 residues should be observed in the backbone fingerprint region of a COSY (Figure 1) or 2Q spectrum. For apo rabbit lung calcyclin at pH 7.0 and 300 K, only 81 discrete crosspeaks were observed. Three cases of near degeneracy were resolved by raising the temperature to 310 K (Glu41/Glu67) or by lowering the pH (Ala79/Lys18; Leu5/Glu52). The Ser46 backbone amide proton was observed only at pH values < 6.2. CaH-NH correlations for Lys64 and Phe70 were observed only in NOESY spectra, and the backbone amide proton resonance of Ser20 could not be identified in any of the experiments.
Less than 50% of all possible correlations from CbH to backbone amide protons were observed at 300 K, but this figure improved to 70% after analysis of the TOCSY acquired at 310 K. The vast majority of these correlations was subsequently verified and many additional CbH resonances (including cases of degenerate CbH resonances) were identified in 2Q spectra, bringing the CbH assignments to 90% completion. Correlations from other sidechain protons to the backbone amide proton are discussed below on a per-residue basis.
(A) Assignment of residues with unique spin systems. Of the 7 glycine residues, 6 were characterized by their unique multiplet structure in COSY. The tentative identification of Gly90 at the C-terminus was made on the basis of the particularly strong intensity and characteristic multiplet structure of the CaH-NH crosspeak. The complete assignments of all 7 glycines were obtained from remote peaks in the 2Q spectrum at w1=CaH + CaH', w2=NH, including the degenerate CaH, CaH' protons of Gly90.
Of the 6 alanine residues, 4 showed correlations from the methyl proton to the backbone amide proton in the R-COSY spectrum. Assignments for the Ala51 and Ala81 spin systems were completed by identification of CβH3-CaH correlations in 2Q and 2QF-COSY spectra. The 3 threonine spin systems were assigned in a stepwise fashion because correlations from CgH3 to NH were not observed. The Thr43 backbone amide proton showed only a single correlation in the CaH/CbH region (at 4.29 ppm) in COSY, R-COSY, and DR-COSY spectra, but the 2Q spectrum revealed that the CaH and CbH resonances of Thr43 are degenerate.
Of the 54 expected methyl resonances for the 2 valine, 13 leucine, and 8 isoleucine residues, 50 discrete COSY crosspeaks were observed. The assignment of this region was possible after an in-depth analysis revealed that one of the Leu5 CdH-CgH crosspeaks was almost entirely obscured due to spectral crowding, and that the methyl resonances of Leu11, Ile15, and Ile44 are nearly or completely degenerate. Only Val68, Leu29, and Leu42 exhibited complete spin system correlation to the backbone NH proton in TOCSY spectra. However, correlations from the methyl resonances to the CaH resonances were observed in TOCSY spectra for all Leu and Val residues except Leu60. For Leu60, relayed coherence transfer from the methyl resonances extended only to CbH, and this CbH was not correlated to the CaH or the backbone NH. These observations strongly suggest that values of both 3Jab are very small. The complete Leu60 spin system assignment could only be confirmed via analysis of NOESY spectra. The CgH3 and Cb protons of the 8 isoleucine residues showed correlations to CaH or the backbone NH in R-COSY or TOCSY spectra. The CdH3-CgH2 spin subsystems were identified in COSY, 2QF-COSY, and MQ spectra. These two Ile spin subsystems were correlated in TOCSY spectra, except for Ile15, Ile38, and Ile53, for which assignments could only be confirmed from the analysis of NOESY spectra.
The complete Arg62 spin system was identified via relayed coherence transfer to the backbone amide proton. Of the 11 lysine residues, complete sidechain proton assignments were obtained only for Lys89, for which correlations from the backbone NH through to CeH2 were observed. Overlap of relayed coherence transfer connectivities between backbone NH-based and CeH2-based spin subsystems (Chazin et al., 1987) provided assignments for Lys40, Lys47, Lys55 and Lys59. Assignments for the sidechains of the other 6 lysine residues were extended only to the Cb protons, due to resonance degeneracies and limited coherence transfer.
(B) Assignment of Residues with 3-Spin Side Chains. Twelve of the 22 residues with 3-spin sidechains were identified by discrete connectivities from CaH and both CbH to the backbone amide proton. For eight other residues, the assignments of NH, CaH and CbH2 could be obtained after the complementary analysis of the MQ, MQF-COSY, and TOCSY spectra. The specific assignment of a 3-spin system to serine was made on the basis of the low field chemical shifts of the Cb protons (d>3.8 ppm). No scalar correlations between CbH and CaH or NH were observed for Ser30, but the resonances could be tentatively assigned from the NOESY spectra. The CaH-CbH-CbH' spin subsystem of Ser20 could be identified by scalar correlations and was assigned by default as the last remaining serine spin system. This was confirmed in the NOESY spectra, although the backbone amide proton could not be identified (vide supra). In the aromatic fingerprint region of the COSY spectrum, the 3 tyrosine and 3 phenylalanine ring spin systems were readily identified. The aromatic ring systems, including the CdH resonances of the histidines, and the 3 asparagine sidechain amide resonances were correlated to the remainder of their spin systems by NOEs to the Cb and/or Ca protons.
(C) Assignment of Residues with 5-Spin Side Chains. For the 15 5-spin systems (excluding Met1, discussed below), at least one correlation from CbH2 and CgH2 to the backbone NH or CaH could be identified, except for Gln49. The identification of Cb protons was confirmed from remote peaks in the 2Q spectrum, including 4 instances of degeneracy or near degeneracy. The CgH resonances are assumed to be degenerate if only one scalar correlation was observed for the two resonances. The tentative assignment of CgH2 for Glu41 is based on the correlation to the backbone amide proton in NOESY spectra. The CgH2 resonances for Gln49 are not assigned. The 5 glutamine residues were specifically identified by NOEs between the sidechain amide protons and CgH, CbH, and/or CaH. Two very narrow resonances at d = 2.03 and 2.18 ppm were attributed to methionine methyl groups and were assigned to Met57 and Met82, respectively, on the basis of NOEs to the Cg protons.
Identification of Met1 and Pro4 was deferred until the last stages of analysis, when there remained few unassigned scalar correlated spin systems in the MQ and MQF-COSY spectra. The CaH-CbH2 portion of the Met1 sidechain could be identified at low contour level in 2Q and TOCSY spectra, but the Cg protons could not be identified. The CeH3 resonance was identified by a strong NOE to the Ala2 backbone amide proton, indicating that it is degenerate with Met57 CeH3. The Pro4 spin system was identified by correlations from the CaH-CbH2 and Cd-Cd' spin subsystems to the same CgH resonance (at 2.31 ppm) in the TOCSY and 3Q spectra. The spin system was assigned to Pro4 based on the characteristic chemical shifts.
Sequence-Specific Assignments. The sequence-specific assignments were obtained using the standard sequential assignment procedure of Billeter et al. (1982). The assignments for residues in helical conformation were largely obtained from dNN connectivities. In many cases, these assignments were confirmed by the presence of less intense daN connectivities. For residues in extended conformation, the sequence-specific assignments are based on strong daN(i, i + 1) connectivities. A summary of the sequential connectivities is included in Figure 2. At least one dNN(i, i + 1) or daN(i, i + 1) connectivity is observed for each pair of consecutive residues with the following exceptions: Met1/Ala2, Ala8/Ile9, Tyr19/Ser20, Ser20/Gly21, Gly45/Ser46, Ser46/Lys47, and Phe76/Leu77. All but Ser20/Gly21 and Gly45/Ser46 are associated with multiple resonance degeneracies. In two of these instances (Ser46/Lys47 and Phe76/Leu77), the sequential assignment could be made on the basis of dbN connectivities. The lack of connectivities involving Ser20 and Ser46 reflects the inability to identify their backbone amide protons in NOESY spectra. The daN connectivity for Ser20/Gly21 may be lost due to the coincidence of the Ser20 CaH resonance with the solvent signal. The sequential NOE connectivities between Ser3 and Pro4 include correlations from the Ser3 backbone NH and CaH to the Pro4 Cd protons, indicating that the Pro4 peptide bond has adopted the trans geometry.
Billeter M, Braun W, Wüthrich K. 1982. Sequential resonance assignments in protein 1H nuclear magnetic resonance spectra. J Mol Biol 155:321-346.
Chazin WJ, Rance M, Wright PE. 1987. Complete assignment of lysine resonances in 1H NMR spectra of proteins as probes of surface structure and dynamics. FEBS Lett 222:109-114.
Chazin WJ, Rance M, Wright PE. 1988. Complete assignment of the 1H nuclear magnetic resonance spectrum of French bean plastocyanin. Application of an integrated approach to spin system identification in proteins. J Mol Biol 202:603-622.
Chazin WJ, Wright PE. 1987. A modified strategy for identification of 1H spin systems in proteins. Biopolymers 26:973-977.
Wüthrich K. 1986. NMR of Proteins and Nucleic Acids. New York: John Wiley & Sons.
Wüthrich K, Billeter M, Braun W. 1984. Polypeptide secondary structure determination by nuclear magnetic resonance observation of short proton-proton distances. J Mol Biol 180:715-740.