Conformational exchange of aromatic side chains characterized by L-optimized TROSY-selected 13C CPMG relaxation dispersion
- 1.1k Downloads
Protein dynamics on the millisecond time scale commonly reflect conformational transitions between distinct functional states. NMR relaxation dispersion experiments have provided important insights into biologically relevant dynamics with site-specific resolution, primarily targeting the protein backbone and methyl-bearing side chains. Aromatic side chains represent attractive probes of protein dynamics because they are over-represented in protein binding interfaces, play critical roles in enzyme catalysis, and form an important part of the core. Here we introduce a method to characterize millisecond conformational exchange of aromatic side chains in selectively 13C labeled proteins by means of longitudinal- and transverse-relaxation optimized CPMG relaxation dispersion. By monitoring 13C relaxation in a spin-state selective manner, significant sensitivity enhancement can be achieved in terms of both signal intensity and the relative exchange contribution to transverse relaxation. Further signal enhancement results from optimizing the longitudinal relaxation recovery of the covalently attached 1H spins. We validated the L-TROSY-CPMG experiment by measuring fast folding–unfolding kinetics of the small protein CspB under native conditions. The determined unfolding rate matches perfectly with previous results from stopped-flow kinetics. The CPMG-derived chemical shift differences between the folded and unfolded states are in excellent agreement with those obtained by urea-dependent chemical shift analysis. The present method enables characterization of conformational exchange involving aromatic side chains and should serve as a valuable complement to methods developed for other types of protein side chains.
KeywordsRelaxation dispersion Conformational exchange Protein folding Aromatic side chain TROSY
Conformational transitions are intimately tied to protein function. Numerous reports have shown that transiently populated high-energy states play important roles in enzyme catalysis (Cole and Loria 2002; Eisenmesser et al. 2002; Sprangers et al. 2005; Boehr et al. 2006) or ligand binding by conformational selection (Malmendal et al. 1999; Brüschweiler et al. 2009). Intermittent transitions between different conformations generally lead to modulation of NMR parameters, such as the chemical shift (Gutowsky and Saika 1953) or residual dipolar couplings (Igumenova et al. 2007; Vallurupalli et al. 2007), resulting in exchange contributions to transverse relaxation rates. Biologically relevant exchange correlation times often fall in the range of milliseconds, which can be probed by NMR relaxation dispersion methods, such as the R 1ρ (Akke and Palmer 1996) or Carr-Purcell-Meiboom-Gill (CPMG) experiment (Carr and Purcell 1954; Meiboom and Gill 1958) and variants thereof (Loria et al. 1999a, b; Igumenova and Palmer 2006). These experiments provide a powerful means of characterizing exchange processes in terms of the exchange rate, k ex, the relative populations of the exchanging states, p i, and the difference in chemical shift, Δω, or residual dipolar coupling between them. In addition to the applications mentioned above, relaxation dispersion methods have proven very successful in studying protein folding, including characterization of transition states as well as intermediate states (Grey et al. 2006; Neudecker et al. 2006; Teilum et al. 2006b; O’Connell et al. 2009; Korzhnev et al. 2010). To date, experiments have been designed to probe conformational exchange at specific sites in proteins, including the backbone (Akke and Palmer 1996; Loria et al. 1999a, b; Hill et al. 2000; Mulder and Akke 2003; Lundström and Akke 2005a, b; Igumenova and Palmer 2006; Lundström et al. 2008, 2009a) and side-chain aliphatic (Lundström et al. 2009b; Hansen et al. 2012), carbonyl/carboxyl (Paquin et al. 2008; Hansen and Kay 2011), and methyl groups (Mulder et al. 2002; Brath et al. 2006; Baldwin et al. 2010; Otten et al. 2010).
Aromatic residues are prevalent in protein binding interfaces, where they contribute significantly to the binding free energy (Bogan and Thorn 1998; Lo Conte et al. 1999; Birtalan et al. 2010). His and Tyr also play prominent roles in enzyme catalysis (Bartlett et al. 2002). Furthermore, aromatic side chains make up a significant proportion of the protein interior, and therefore provide an attractive means of probing the dynamics of the hydrophobic core in the native state (Wüthrich and Wagner 1975), as well as the formation of the core in protein folding reactions. Thus, there is great incentive to extend the existing repertoire of CPMG-type experiments to probe also aromatic side chains in proteins.
We applied the L-TROSY-CPMG experiment to characterize fast folding–unfolding of cold-shock protein B (CspB) from Bacillus subtilis (Schindler et al. 1995) under native conditions. CspB was expressed in Escherichia coli cultured on medium containing 1-13C1-glucose as the sole carbon source, resulting in enrichment of 13C into the δ positions of Phe, δ1 and ε3 of Trp, and δ2 and ε1 of His. CspB contains 7 Phe, 1 Trp, 1 His, but no Tyr.
Significant dispersion profiles were obtained for 9 out of 11 aromatic residues in CspB. The data are adequately represented by a global two-state (folded/unfolded) model using the Carver-Richards equation (Carver and Richards 1972; Palmer et al. 2001), with k ex = 528 ± 52 s−1 and a population of the unfolded state of p U = 2.3 ± 0.2 % (Fig. 2a–d). By contrast, the relaxation data obtained using the rc-CPMG sequence (Fig. 2a, inset; Fig. S1) suffer severely from the fast transverse relaxation of the averaged 13C doublet (roughly a factor of two greater than that of the spin-state selective TROSY line), and are clearly not of sufficient quality to warrant further interpretation.
In the context of the L-TROSY-CPMG experiment, it is critical to recognize that the decay of antiphase coherences (e.g. 2CyHz) can be affected by strong coupling between the 1H spin covalently attached to the monitored 13C spin and neighboring 1H spins, which makes the effective R 2 rates depend on the refocusing frequency ν cp also in the absence of conformational exchange (van Ingen et al. 2009). In the case of aromatic spin systems, such aberrant dispersions can become significant if strong scalar coupling occurs between vicinal 1H spins (e.g. 1Hδ1 and 1Hε1 in Phe) for which 3 J HH ≈ 8 Hz, whereas weaker coupling constants (e.g. 4 J HH ≈ 2 Hz, across the ring) do not cause any significant problems irrespective of the frequency difference between the coupled 1H spins, in agreement with previous results (van Ingen et al. 2009). His δ2 and ε1 and Trp δ1 1H spins are always weakly coupled to their vicinal 1H neighbors, which are covalently attached to nitrogen and therefore resonate far downfield. In the case of Phe and Tyr residues, rapid ring flipping makes it impossible to determine a priori whether the weak coupling limit applies. However, numerical simulations using QSim (Helgstrand and Allard 2004) demonstrate that artifactual dispersion decays caused by strong scalar coupling level out completely for ν cp ≥ 100 Hz, where the 13C–1H scalar interaction is effectively decoupled. Furthermore, the artifactual dispersion magnitudes are generally limited to R 2(ν cp = 0)−R 2(ν cp → ∞) ≤3 s−1. These two criteria serve as useful guidelines for establishing the accuracy of relaxation dispersions measured for Phe and Tyr.
In CspB, H29ε1 and W8δ1 both show dispersion steps of approximately 10 s−1 (Fig. 2c–d). Two out of seven Phe exhibit dispersions of 4–6 s−1 (as exemplified by F15δ, Fig. 2a), which also can be safely attributed to conformational exchange. Three remaining Phe rings show smaller dispersion steps of 2 s−1 (as exemplified by F38δ, Fig. 2b), and 2 rings do not show any appreciable relaxation dispersion. All aromatic 13C dispersion curves of CspB reach a plateau at ν cp > 200 Hz, clearly indicating that the dispersions arise from conformational exchange, rather than from effects of strong coupling. Thus, we conclude that the present data are unhampered by any strong coupling between protons and accurately reflect exchange between folded and unfolded states.
The unfolding rate, k U = 12 ± 3 s−1 (derived from the fitted parameters: k U = k ex·p U) matches perfectly with that (12 ± 7 s−1) determined previously from stopped-flow data (Schindler et al. 1995), whereas the folding rates differ by a factor of two. Notably, it has been observed for CspB that k U is almost independent of urea concentration, whereas the folding rate (k F) exhibits a strong urea dependence (Schindler et al. 1995), indicating that the former is considerably more robust with respect to variations in solvent conditions, such as the difference in buffer composition between the present (10 mM HEPES) and previous (20 or 100 mM sodium cacodylate; Schindler et al. 1995; Zeeb and Balbach 2005) experiments. At the level of individual 13C sites, no deviations from two-state folding behavior were observed, in agreement with previous results (Schindler et al. 1995; Zeeb and Balbach 2005). In the present case, ring flips are expected to be significantly faster than the folding–unfolding kinetics. Consequently, the two processes are time-scale separated, such that CPMG dispersion measurements report only on the folding–unfolding process, while exchange due to ring flips in CspB is too fast to be probed by current refocusing rates.
We assessed the accuracy of the chemical shift differences between the folded and unfolded states determined from the CPMG dispersions by comparing with shift differences measured from 1H–13C HSQC spectra of the folded and unfolded states. 13C chemical shifts of the unfolded state under native conditions were obtained by monitoring the urea-dependence of the 1H–13C HSQC spectrum from 0.5 to 2.5 M urea and extrapolating linearly to 0 M (Fig. S2). Residue-specific assignments of Phe 13Cδ resonances in the unfolded state were not necessary, because they all merged to a single overlapped signal.
In conclusion, the L-TROSY-CPMG relaxation dispersion experiment for aromatic 13C spins provides accurate information on conformational exchange, including the aromatic chemical shifts of the transiently populated high-energy state, and should serve as a valuable complement to experiments developed for other types of side chains.
We thank Prof. Jochen Balbach for kindly providing the CspB clone. This research was supported by the Swedish Research Council (621-2010-4912; 822-2005-2915), the Göran Gustafsson Foundation for Research in Natural Sciences and Medicine, and the Knut and Alice Wallenberg Foundation. U.W. was supported by an EMBO long-term fellowship. M.R. was supported by a postdoctoral fellowship funded by the Swedish Research Council (623-2009-800).
This article is distributed under the terms of the Creative Commons Attribution License which permits any use, distribution, and reproduction in any medium, provided the original author(s) and the source are credited.
- Carver JP, Richards RE (1972) A general two-site solution for the chemical exchange produced dependence of T2 upon the carr-purcell pulse separation. J Magn Reson 6:89–105Google Scholar
- Geen H, Freeman R (1991) Band-selective radiofrequency pulses. J Magn Reson 93:93–141Google Scholar
- Grey MJ, Tang YF, Alexov E, McKnight CJ, Raleigh DP, Palmer AG (2006) Characterizing a partially folded intermediate of the villin headpiece domain under non-denaturing conditions: contribution of His41 to the pH-dependent stability of the N-terminal subdomain. J Mol Biol 355:1078–1094CrossRefGoogle Scholar
- Lundström P, Akke M (2005a) Microsecond protein dynamics measured by rotating-frame 13Ca spin relaxation. Chem Bio Chem 6:1685–1692Google Scholar
- Lundström P, Teilum K, Carstensen T, Bezsonova I, Wiesner S, Hansen F, Religa TL, Akke M, Kay LE (2007) Fractional 13C enrichment of isolated carbons using [1–13C]- or [2–13C]-glucose facilitates the accurate measurement of dynamics at backbone Cα and side-chain methyl positions in proteins. J Biomol NMR 38:122–199Google Scholar
- Neudecker P, Zarrine-Afsar A, Choy WY, Muhandiram DR, Davidson AR, Kay LE (2006) Identification of a collapsed intermediate with non-native long-range interactions on the folding pathway of a pair of Fyn SH3 domain mutants by NMR relaxation dispersion spectroscopy. J Mol Biol 363:958–976CrossRefGoogle Scholar
- Palmer AG, Cavanagh J, Wright PE, Rance M (1991) Sensitivity improvement in proton-detected two-dimensional heteronuclear correlation NMR spectroscopy. J Magn Reson 93:151–170Google Scholar
- Shaka AJ, Barker PB, Freeman R (1985) Computer-optimized decoupling scheme for wideband applications and low-level operation. J Magn Reson 64:547–552Google Scholar
- van Ingen H, Korzhnev DM, Kay LE (2009) An analysis of the effects of 1HN-1HN dipolar couplings on the measurement of amide bond vector orientations in invisible protein states by relaxation dispersion NMR. J Phys Chem B 113:9968–9977Google Scholar
- Weininger U, Diehl C, Akke M (2012) 13C relaxation experiments for aromatic side chains by longitudinal- and transverse-relaxation optimized NMR spectroscopy. J Biomol NMR. doi: 10.1007/s10858-012-9650-5