PMCCPMCCPMCC

Search tips
Search criteria 

Advanced

 
Logo of nihpaAbout Author manuscriptsSubmit a manuscriptHHS Public Access; Author Manuscript; Accepted for publication in peer reviewed journal;
 
J Am Chem Soc. Author manuscript; available in PMC 2012 December 4.
Published in final edited form as:
PMCID: PMC3513771
NIHMSID: NIHMS424092

15N-1H Scalar Coupling Perturbation: An Additional Probe for Measuring Structural Changes Due to Ligand Binding

Abstract

Chemical shift perturbation mapping of backbone amides is one of the most widely employed techniques in biomolecular NMR, providing residue-by-residue information on interaction interfaces, ligand binding and chemical modification sites, even for samples where poor solubility, short lifetime, or large size precludes more sophisticated experimental approaches. Significant changes can also occur in the amide one-bond 15N-1H scalar coupling constants for glutamine binding protein (GlnBP) due to ligand binding. Like chemical shift perturbations, large changes (> 1 Hz) are seen near the site of glutamine binding, though perturbations also occur distant to the site. The coupling constant perturbations correlate with significant structural changes, especially changes in backbone hydrogen bonding. Thus, amide scalar coupling perturbation can serve as an adjunct to chemical shift perturbation, providing additional information on both short-range and longer-range, allosteric structural changes.

Chemical shift perturbation mapping of backbone amides is one of the most widely employed techniques in biomolecular NMR1,2. It provides residue-by-residue information on interaction interfaces, ligand binding and chemical modification sites. It is applicable even for samples where poor solubility, short lifetime, or large size precludes more sophisticated experimental approaches. Here, we show that significant changes can also occur in the backbone amide 15N-1H scalar coupling constants for glutamine binding protein (GlnBP) due to ligand binding. Like chemical shift perturbations, large changes (> 1 Hz) are seen near the site of glutamine binding, though perturbations also occur distant to the site. The coupling constant perturbations correlate with significant structural changes, especially changes in backbone hydrogen bonding36. More importantly some of the changes in the J-couplings do not correspond to changes in the chemical shifts. Thus, amide scalar coupling perturbation can serve as an adjunct to chemical shift perturbation, providing additional information on both short-range and longer-range, allosteric structural changes.

Nuclear spin-spin scalar coupling between nuclei is relayed via their interactions with the electrons of the system7. The interaction is dominated by the four Ramsey terms, the diamagnetic and paramagnetic spin-orbit, spin-dipole, and Fermi contact mechanisms8. The first two, diamagnetic and paramagnetic spin-orbit mechanisms are related to the diamagnetic and paramagnetic shielding terms that lead to chemical shift perturbation, thus the perturbation in coupling values arising from these terms should resemble the pattern seen for chemical shift perturbation. The second two terms have no shielding counterparts, and for one-bond scalar coupling, the Fermi contact term dominates. The largest contribution arises from s-orbital spin density centered at one nucleus overlapping with the coupled nucleus, and thus has a strong dependence on bond length. The HN bond length depends, in turn, on its immediate electronic environment, on hydrogen bonding in particular9,10.

Figures 1A and 1B compare plots of the HN scalar coupling (ΔJHN) and the weighted average chemical shift differences (ΔδHN, calculated by equation, i.e., {[(ΔH)2 + (ΔN/5)2]/2}1/2)11,12 between the bound and the free form GlnBP. Figures 1C and 1D compare the largest chemical shift and J-coupling changes mapped onto the glutamine-bound GlnBP structure13b. The coupling constant perturbations are more dispersed through the structure compared to the chemical shift changes. Two among the top five ΔJHN values, D157 (−1.6 Hz) and T11 (1.4 Hz), correspond to two of the largest chemical shift changes, and both lie adjacent to the bound glutamine. Three of the top five ΔJHN, T11, V121 (−1.3 Hz) and D157, undergo formation or breaking of hydrogen bonds upon ligand binding. Overall, the J-coupling changes for amides undergoing hydrogen bond breakage or formation are twice as large (± 0.5 Hz) as those which remain bonded (± 0.25 Hz). Also notable is K131 (ΔJHN 0.7 Hz), which lies adjacent one of the most structurally labile loops of the GlnBP protein (Figure 2A). Backbone amide residual dipolar couplings (RDCs) predicted from the x-ray structures show poor agreement with the measured RDC values for this region (data not shown), thus figure 2A might under-represent the structural changes in solution for K131. One of the largest coupling constant perturbations appears unrelated to hydrogen bonding, that of Asp28 (ΔJHN −1.0 Hz). This amide lies near two aromatic rings (< 4 Å), Phe16 and Tyr27, illustrating the significance of the paramagnetic spin-orbit Ramsey term to the coupling constant perturbation (figure 2B).

Figure 1
NMR-observed scalar coupling difference (A) and the weighted average chemical shift differences profile (B) of GlnBP13 between bound and free form. Residues with significant J-coupling and backbone amide chemical shift change are shaded green, while the ...
Figure 2
Expanded small region of superimposed free13a (PDB code 1GGG, colored grey) and bound13b (PDB code 1WDN, colored light blue) crystal structure of GlnBP around residues K131 (A) and D28 (B). The hydrogen bonding of residue K131 in the bound form is shown ...

The ΔJHN values shown in figure 1 were obtained using the NMR signal intensity based J-modulation technique for greater accuracy. By this approach the interference effect arising from dipole/dipole cross-correlated relaxation is almost eliminated by refocusing spin evolution15,16. Repeated measurements indicated these values were reproducible to less than 0.08 Hz. Residues with significant resonance overlap were excluded from the analysis. The J-modulation experiments were repeated on a 600 MHz spectrometer, and the resulting ΔJHN values matched the 800 MHz results with a standard deviation of 0.15 Hz, providing an estimate of the accuracy of the technique. For comparison, ΔJHN values were also measured at 800 MHz using the in-phase and anti-phase IPAP technique17, yielding a standard deviation of 0.20 Hz compared to the 800 MHz J-modulation values. The larger deviation reflects dipole-dipole cross-correlated relaxation and unresolved E.COSY effects18. The spread in ΔJHN values seen for GlnBP is −1.6 to 1.4 Hz, so the IPAP technique should be adequate for identifying large ΔJHN values, while the J-modulation experiment will be needed for detecting more subtle perturbations near 0.2 Hz.

Measurement of free GlnBP ΔJHN values at two salt concentrations (0 and 100mM NaCl) showed no significant changes (± 0.07 Hz), and the ΔJHN values for free GlnBP showed no significant changes (± 0.07 Hz) in the presence of a non-binding substrate analog, D-glutamic acid (1.2 mM). Comparison of ΔJHN values at two temperatures (31 and 41 °C) showed a slight increase (0.08Hz) on average at the lower temperature, as expected from the relation between molecular tumbling correlation time and the dynamic frequency shift19, but showed no significant changes when the overall increase was subtracted (see in Supporting Information).

The differences in J-coupling and chemical shift changes (Figures 1A and 1B) suggest their values have different sensitivities to local perturbations despite sharing some very similar interaction terms in their hamiltonians. Moreover these changes vary in sign depending on the local geometry. Even in cases where the dominant terms leading to perturbation are the same for both J-coupling and chemical shift, the various contributions to the chemical shift might cancel, while the contributions to the J-coupling might add, and vice versa. Thus, while it is natural to suspect that the J-couplings and chemical shifts will be perturbed similarly upon a conformational change, it is not surprising to find discrepancies.

Comparison of ΔJHN with 1H and 15N chemical shift differences reveals modest correlations, with a coefficient of 0.44 between ΔJHN and ΔN and −0.31 between ΔJHN and ΔH. The magnitude of changes |ΔJHN| and ΔδHN (Figure 1B) also show some correlation (0.40), comparable to the correlation seen between |ΔH| and |ΔN| (0.46). No correlation was seen between either |ΔJHN| or ΔδHN and backbone r.m.s. deviation between free and bound x-ray structures. Attempts to compare ΔJHN to linear combinations of backbone (ϕ, ψ, ω) and side chain (χ1, χ2) dihedral angles also produced no significant correlations. Even with hydrogen bond length and angle information included, no simple correlation was seen. This concurs with calculations of ΔJHN using ab initio density functional theory which predict that sign of the variation of ΔJHN as a function of hydrogen bond length depends also on other variables, such as HCO angle9. Thus, HN coupling constant perturbations provide valuable information on where structural changes occur but do not reveal what specifically those structural changes are.

On a final note, the results here also represent a cautionary message. When determining changes in residual dipolar couplings, between free and bound states, for instance, changes in the isotropic HN J-coupling constants must be taken into account as well. This also applies to residual dipolar couplings of minor population states deduced through R2 dispersion techniques20,

Supplementary Material

supporting info

Acknowledgments

This work was supported by the Intramural Research Program of the NIH, National Heart, Lung, and Blood Institute.

Footnotes

Supporting Information Available: Correlation plot between J-coupling and chemical shift differences for GlnBP; Chemical shift and J-coupling changes mapped onto the glutamine-free structure of GlnBP; J-coupling at two temperatures; salt and non-binding amino acid dependence; table of J-coupling values of bound and free GlnBP.

References

1. Farmer BT, Constantine KL, Goldfarb V, Friedrichs MS, Wittekind M, Yanchunas J, Robertson JG, Mueller L. Nat Struct Biol. 1996;3:995–997. [PubMed]
2. Williamson RA, Carr MD, Frenkiel TA, Feeney J, Freedman RB. Biochemistry. 1997;36:13882–13889. [PubMed]
3. Blake PR, Park J-B, Adams MW, Summers MF. J Am Chem Soc. 1992;114:4931–4933.
4. Cordier F, Grzesiek S. J Am Chem Soc. 1999;121:1601–1602.
5. Cordier F, Barfield M, Grzesiek S. J Am Chem Soc. 2003;125:15750–15751. [PubMed]
6. Bouvignies G, Bernado P, Meier S, Cho K, Grzesiek S, Brueschweiler R, Blackledge M. Proc Natl Acad Sci USA. 2005;102:13885–13890. [PubMed]
7. Helgaker T, Watson M, Handy NC. J Chem Phys. 2000;113:9402–9409.
8. Ramsey NF, Purcell EM. Phys Rev. 1952;85:143–144.
9. Tuttle T, Kraka E, Wu AA, Cremer D. J Am Chem Soc. 2004;126:5093–5107. [PubMed]
10. Sahakyan AB, Shahkhatuni AG, Shahkhatuni AA, Panosyan HA. J Phys Chem A. 2008;112:3576–3586. [PubMed]
11. Garrett DS, Seok YJ, Peterkofsky A, Clore GM, Gronenborn AM. Biochemistry. 1997;36:4393–4398. [PubMed]
12. Chen K, Bachtiar I, Piszczek G, Bouamr F, Carter C, Tjandra N. Biochemistry. 2008;47:1928–1937. [PMC free article] [PubMed]
13. (a) Hsiao CD, Sun YJ, Rose J, Wang BC. J Mol Biol. 1996;262:225–242. [PubMed](b) Sun YJ, Rose J, Wang BC, Hsiao CD. J Mol Biol. 1998;278:219–229. [PubMed](c) Hsiao CD, Sun YJ, Rose J, Cottam PF, Chien H, Wang BC. J Mol Biol. 1994;240:87–91. [PubMed]
14. Koradi R, Billeter M, Wüthrich K. J Mol Graphics. 1996;14:51–55. [PubMed]
15. de Alba E, Tjandra N. J Magn Reson. 2006;183:160–165. [PubMed]
16. de Alba E, Tjandra N. J Biomol NMR. 2006;35:1–16. [PubMed]
17. Ottiger M, Delaglio F, Bax A. J Magn Reson. 1998;131:373–378. [PubMed]
18. Yao L, Ying J, Bax A. J Biomol NMR. 2009;43:161–170. [PMC free article] [PubMed]
19. Tjandra N, Grzesiek S, Bax A. J Am Chem Soc. 1996;118:6264–6272.
20. Vallurupalli P, Hansen DF, Kay LE. Proc Natl Acad Sci USA. 2008;105:11766–11771. [PubMed]