Removal of slow-pulsing artifacts in in-phase 15N relaxation dispersion experiments using broadband 1H decoupling

Removal of slow-pulsing artifacts in in-phase 15N relaxation dispersion experiments using... Understanding of the molecular mechanisms of protein function requires detailed insight into the conformational landscape accessible to the protein. Conformational changes can be crucial for biological processes, such as ligand binding, protein folding, and catalysis. NMR spectroscopy is exquisitely sensitive to such dynamic changes in protein conformations. In particular, Carr–Purcell–Meiboom–Gill (CPMG) relaxation dispersion experiments are a powerful tool to investigate pro- tein dynamics on a millisecond time scale. CPMG experiments that probe the chemical shift modulation of N in-phase magnetization are particularly attractive, due to their high sensitivity. These experiments require high power H decoupling 15 15 during the CPMG period to keep the N magnetization in-phase. Recently, an improved version of the in-phase N-CPMG experiment was introduced, offering greater ease of use by employing a single H decoupling power for all CPMG puls- ing rates. In these experiments however, incomplete decoupling of off-resonance amide H spins introduces an artefactual dispersion of relaxation rates, the so-called slow-pulsing artifact. Here, we analyze the slow-pulsing artifact in detail and demonstrate that it can be suppressed through the use of composite pulse decoupling (CPD). We report the performances of various CPD schemes and show that CPD decoupling based on the 90 –240 –90 element results in high-quality dispersion x y x curves free of artifacts, even for amides with high H offset. Keywords Relaxation dispersion · Decoupling · Protein dynamics · Composite pulse Introduction (~ ms), these excited states can be crucial for biologically important processes such as enzyme catalysis (Hammes Biological macromolecules such as nucleic acids and pro- 1964; Eisenmesser et  al. 2005; Henzler-Wildman et  al. teins are non-rigid entities that can populate a variety of 2007; Palmer 2015; Kim et al. 2017), ligand binding or pro- conformers in their energy landscape (Frauenfelder et al. tein–protein interactions (Sugase et al. 2007; Schneider et al. 1991; Wolynes 2005; Henzler-Wildman and Kern 2007). 2015; Pratihar et al. 2016; Xiao et al. 2016; Zhao et al. 2017; The lowest energy conformation, the ground state, is often Delaforge et al. 2018), and protein folding (Korzhnev et al. able to transiently access higher-energy conformations. Even 2010; Neudecker et al. 2012; Kimsey et al. 2015; Libich when their population is low (< 10%) and life times is short et al. 2015; Franco et al. 2017; Culik et al. 2018). While these states cannot be detected directly due to their transient and lowly populated nature, NMR experiments (Akke and Electronic supplementary material The online version of this Palmer 1996; Fawzi et al. 2010; Kovermann et al. 2016) are article (https ://doi.org/10.1007/s1085 8-018-0193-2) contains uniquely able to provide a detailed, atomistic description of supplementary material, which is available to authorized users. the energy landscape. In particular, relaxation dispersion and * Hugo van Ingen chemical exchange saturation transfer experiments are par- h.vaningen@uu.nl ticularly powerful herein, as they give access to the popula- tion, life times and structures of excited states (Palmer et al. Macromolecular Biochemistry, Leiden Institute of Chemistry, Leiden University, P.O Box 9502, 2001; Vallurupalli et al. 2012, 2017; Sauerwein and Hansen 2300 RA Leiden, The Netherlands 2015; Xue et al. 2015; Lisi 2016; Massi and Peng 2018; Present Address: NMR Group, Bijvoet Center Gopalan et al. 2018). for Biomolecular Research, Utrecht University, Padualaan 8, 3584 CH Utrecht, The Netherlands Vol.:(0123456789) 1 3 70 Journal of Biomolecular NMR (2018) 71:69–77 In Carr–Purcell–Meiboom–Gill (CPMG) relaxation dis- Here, we analyze the slow-pulsing artifact in N persion experiments, the characterization of the minor state CW–CPMG sequences in detail and demonstrate a simple is derived from the major state peaks by measurement of method for its removal. In that, we took inspiration from the their effective transverse relaxation rate R , as a function work of Chakrabarti et al. (2016), where composite pulse 2,eff of the pulsing rate in the CPMG period. Signals of nuclear decoupling (CPD) was used to suppress H off-resonance spins that experience exchange between states with differ - effects in exchange mediated saturation transfer experi- ent chemical shifts are affected by exchange-induced line ments. We investigated the performance of various CPD broadening, an effect that depends on the free precession schemes in CW–CPMG sequences and demonstrate here interval (2τ ) between the refocusing pulses in the CPMG that high power CPD based on the 90 –240 –90 element cp x y x element (Palmer et al. 2001; Sauerwein and Hansen 2015). achieves artifact-free dispersion curves over a wide range Analysis of the resulting relaxation dispersion curve, a plot of H offsets. of the R versus CPMG frequency (1/4τ ), allows deter- 2,eff cp mination of the rate of exchange (k ), population of minor ex state (p ) and the absolute chemical shift difference (|Δϖ|) Materials and methods between the exchanging states. Importantly, since the shape of the dispersion profile depends on Δϖ , data is typically NMR samples acquired at two fields to accurately determine the exchange parameters (Sauerwein and Hansen 2015). NMR experiments were recorded on a sample of 2.5 mM 15 15 13 The N backbone amide spin is the most popular uniformly N/ C-labelled Cu(II) azurin in 25 mM potas- nucleus for CPMG RD experiments, due to the simplicity sium phosphate buffer at pH 5.49 with 5% D O. Labelled of isotope-labeling, the straightforwardness of the two-spin azurin was produced and purified according to a previously 1 15 H– N spin system, and the high sensitivity and resolution published protocol with modifications for incorporating 13 15 afforded by these experiments. A critical aspect of these C-glucose and N-ammoniumchloride (Karlsson et al. experiments is appropriate handling of differences in the 1989). intrinsic R of the in-phase (N ) and anti-phase (2N H ) 2 x,y x,y z N magnetization which are generated in the free evolution NMR experiments periods. Anti-phase terms have higher intrinsic relaxation rates due to a contribution of H spin flips to their decay. Relaxation dispersion experiments, using the ST–CW–CPMG The original implementation of the N CPMG RD experi- sequence, were recorded at 298 K on Bruker Avance III HD ment uses a relaxation-compensation scheme to average the spectrometers operating at 850 and 950 MHz H Larmor fre- N and 2N H relaxation rates (Loria et al. 1999b). The quency and equipped with TCI cryoprobes. The constant-time x,y x,y z N CPMG sequence of Hansen et al. (2008b) (CW–CPMG) CPMG relaxation delay (T ) was set to 40 ms with ν set relax CPMG measures the dispersion profile of pure in-phase N by to 25, 50, 75, 100, 125, 175, 225 (2×), 275, 300, 350, 400 (2×), x,y applying high-power continuous wave (CW) H decoupling 500, 550, 600, 650, 700 (2×), 750, 800 (2×), 850, 900, 950 and during the CPMG train, offering enhanced sensitivity for 1000 Hz respectively, run in an interleaved manner. Duplicates non-deuterated proteins. Recently, Jiang et al. (2015) modi- were used to estimate the error in R . The errors were set to 2,eff −1 15 fied this sequence (ST–CW–CPMG) to use a single CPMG 0.2 s at minimum. The pulse length of the N refocusing train with the Yip and Zuiderweg phase cycle (2004) and pulses in the CPMG train was 90 µs. For H decoupling, either a single CW decoupling power, yielding dispersion curves CW decoupling or a CPD-scheme (GARP, DIPSI, MLEV16, free of off-resonance artifacts for a wider range of N offset WALTZ16, 90 –240 –90 ) was used. This was implemented by x y x frequencies. changing the “cw:f1” statement in the pulse program to read Both CW–CPMG sequences are nevertheless sensitive “cpds1:f1” (pulse program available upon request). In either 1 1 to artifacts from H off-resonance effects (Hansen et  al. case, the decoupling field strength was 14.7 kHz (17 µs H 90° 1 1 2008b; Yip and Zuiderweg 2004). Amide H spins that pulse), applied at 8.2 ppm H offset. A total of 3072/120 points 1 15 are far off-resonance from the CW decoupling field are not were acquired in the H/ N dimension with an acquisition fully decoupled from the N spin, resulting in generation time of 90/27.85 ms and a relaxation delay of 2 s and 4 scans of 2N H magnetization through the residual J-coupling. per FID. A reference spectrum, without the relaxation delay, x,y z Consequently, higher R values will be measured for low was also recorded. NMR data were processed with NMRPipe 2,eff ν values, for which free precession periods are long and (Delaglio et al. 1995), using linear prediction in the N dimen- CPMG more of the antiphase terms will be generated. This so-called sion and Lorentz-to-Gauss window functions. Peak volumes slow-pulsing artifact shows up as an artefactual dispersion were obtained by peak fitting using FuDa (Hansen, http://www. curve, interfering with accurate extraction of minor-state bioch em.ucl.ac.uk/hanse n/fuda/), and subsequently converted parameters. into effective relaxation rates via R (ν ) = − 1/T ·ln(I 2,eff CPMG relax 1 3 Journal of Biomolecular NMR (2018) 71:69–77 71 −1 (ν )/I ), where I is the peak intensity in a reference spec- of 13.6 and 26.7 s respectively. Dispersion experiments CPMG 0 0 trum recorded without the relaxation delay T . The R val- were simulated with T set to 40 ms, and ν values relax 2,eff relax CPMG ues measured using the ST–CW–CPMG sequence were cor- ranging from 25 to 1000 Hz, the N 180° refocusing pulse rected for R -contribution according to the formula described was set to 90 µs, H CW decoupling field strength was set to −1 1 by Jiang et al. (2015) using an estimate of 0.95 s R - and 14.7 kHz (17 µs H 90°). −1 10.5 s for R -contribution for all residues. Dispersion curves obtained with either CW or CPD decoupling were compared by calculating the RSMD between the curves for all residues: Results Measurement of in-phase N CPMG relaxation dispersion CPD,i CW,i RMSD = R − R 1 2,eff 2,eff profiles critically relies on H decoupling to measure the i=1 pure in-phase N relaxation rate without contamination by x,y the anti-phase relaxation rate. As pointed out in the work where i is the index of a particular ν value and the sum- CPMG of Jiang et al. (2015), the decoupling field strength has a mation runs over the N recorded points, equal to the num- practical limit of roughly 14 kHz, resulting in a residual J ber of points per dispersion curve (M) times the number coupling interaction for amide protons at non-zero offset to of residues. The systematic difference between the CW or the decoupling field. This interaction causes slow intercon- CPD-based dispersion curves was calculated from the aver- version of in-phase and anti-phase magnetization during the age point-by-point difference per residue and is tabulated in CPMG period, which will lead to undesired averaging of the Table S1. To compensate for these systematic differences, an in-phase and anti-phase relaxation rates (Loria et al. 1999b; “R -offset compensated” RMSD was calculated by replac - Hansen et al. 2008b). To first approximation, this averaging ing the CPD-based R values with the offset compensated 2,eff can be described by the equation derived by Palmer et al. values: (1992) for calculating the effective relaxation rate in spin echo sequences. Here, it is adapted and reformulated to CPD,compensated,i CPD,i CPD,i CW,i R = R − R − R express to the size of the slow-pulsing artifact A: 2,eff 2.eff 2,eff 2,eff i=1 anti in A = R − R 1 − sincJ 2 (1) r cp 2 2 Simulation of  N CW–CPMG dispersion profiles in anti 15 where R and R are the N in-phase and anti-phase 2 2 To evaluate the magnitude of the slow-pulsing artifact in transverse relaxation rates, J is the residual J-coupling, and relaxation dispersion profiles, numerical simulations of 2τ is the inter-pulse delay in the CPMG pulse train. In the cp 1 15 a two-spin H– N system were carried out, assuming a limit of perfect decoupling J ≈ 0, the sinc factor approaches non-deuterated protein. The evolution of magnetization 1 and A ≈ 0 for all τ values. For non-zero J , A approaches cp r in this spin system was calculated for the CPMG part of zero in the limit of fast pulsing where τ is very small. For cp the CPMG–CW and CPMG–ST–CW sequence, including slow pulsing, however, there is a non-zero artifact, with a the flanking N 90° pulses. Simulations in the absence theoretical limit of 0.5 (R − R ) for infinitely slow puls - 2,anti 2,in of exchange and neglecting pulse imperfections were per- ing. In practice, J can be as much 16 Hz (for 3 ppm H offset formed in operator space by solving the complete homogene- at 850  MHz) and 2τ is typically at most 20  ms, which cp ous master equation as described by Allard et al. (1998) and would generate a maximum artifact of roughly 10% of the Helgstrand et al. (2000) using the open source computing difference between the anti-phase and in-phase relaxation language GNU Octave (http://www.gnu.org/softw are/octav rate. e/) (Eaton et al. 2008). All simulation used the parameters To assess more precisely how the slow pulsing artifact detailed below unless noted otherwise. The N spin was is manifested in N CPMG–CW and ST–CW experiments, assumed to be on-resonance. The magnetic field strength numerical simulations of these sequences were performed was set to 19.9 T, corresponding to H Larmor frequency in Liouville space for a non-exchanging two spin N–H sys- of 850 MHz. Relaxation rates were calculated using overall tem. Figure 1a compares the obtained dispersion profiles rotational correlation time τ of 9 ns, a value of 0.85 for the for the two experiments with the predicted curve based on squared generalized order parameter, 100 ps for the correla- Eq. 1, for a N–H system with 3 ppm H offset at an 850 MHz tion time for internal motions, and − 172/+10 ppm for the spectrometer. Whereas a flat curve is expected for a non- 15 1 N/ H chemical shift anisotropy. Relaxation due to neigh- exchanging system, systematically increased R values 2,eff boring protons was included as described in ref. (Allard are measured in the slow pulsing regime for both pulse et al. 1998) by including a virtual proton at 1.85 Å, resulting sequences. While Eq. 1 is derived for periods of free evolu- in R values of in-phase and anti-phase N magnetizations tion in absence of a decoupling field, the curvature of the 1 3 72 Journal of Biomolecular NMR (2018) 71:69–77 Fig. 1 Impact of the slow-pulsing artifact on simulated relaxation dis- netic field strength for 1 and 2  ppm H offset for proteins of 4 and persion profiles. a Simulated slow-pulsing artifact caused by incom- 9  ns tumbling times. The gray area indicates the typical experimen- −1 1 plete J decoupling in the CW–CPMG and the ST–CW–CPMG tal error in range of 0.1–0.3  s . d The typical accessible H offset NH 15 15 1 implementation of the in-phase N CPMG experiment. Solid lines ranges, color coded into a N– H HSQC spectrum. Assuming the are fits obtained using the program CATIA (Hansen, http://www. H CW field is centered at 8  ppm, the blue region is accessible up bioch em.ucl.ac.uk/hanse n/catia /) assuming two-site exchange. The to the highest magnetic fields, orange is accessible up to 600  MHz, artifact expected based on Eq. 1 is shown for comparison. The boxed and the red region is inaccessible. In a–c simulated profiles are shown region is expanded in the inset. The H offset from the decoupling for both CW–CPMG (open triangle) and ST–CPMG (asterisk) pulse field was set to 3  ppm, assuming an 850  MHz spectrometer. b, c sequences; color coding indicated in the figure. All simulations are Maximum size of the artifact (ΔR ) as a function of (b) H offset based on a non-exchanging N–H spin system. Simulation parameters 2,eff for proteins of 4, 6.5 and 9 ns correlation times at 850 MHz; c mag- are given in “Materials and Methods” section, unless noted otherwise slow-pulsing artifact matches the predicted sinc dependence in Fig. 1a), illustrating the potential impact on the extracted on the pulsing rate. The size of the artifact is somewhat exchange parameters. underestimated by Eq. 1. The original CW sequence shows Since the size of the slow pulsing artifact is governed by slightly lower sensitivity to the artifact than the ST–CW the relaxation difference between in-phase and anti-phase experiment. This difference can be traced back to presence magnetization, it is dependent on protein size. Large proteins 15 1 1 of the N refocussing pulse in between the two halves of have more efficient H– H spin flips which increase the anti- the total CPMG period in the CW experiment. Importantly, phase relaxation rate. Figure 1b compares the magnitude of since the shape of the artifact is virtually indistinguishable the artifact for three different protein sizes as function of H from a bona-fide dispersion profile, the artefactual R val- offset from the decoupling field. For larger proteins, where 2,eff ues can be fitted to an actual dispersion curve (see solid lines the chance of finding amide protons at high offset is also 1 3 Journal of Biomolecular NMR (2018) 71:69–77 73 −1 higher, the artifact can be well above 1 s . At offsets larger effect is small. Both DIPSI2 and GARP use particularly long than ~ 1000 Hz the slow pulsing artifact will be higher than composite pulses (corresponding to the length of 2590° and −1 the typical experimental error (on the order of 0.1–0.3 s ) 1054° rotation, respectively), which aggravates the impact (Hansen et al. 2008a), as also noted by Jiang et al. (2015). of the timing mismatch, in particular at high pulsing rates, Since relaxation dispersion data need to be acquired where the effects from each τ period are compounded. cp at two magnetic fields in order to extract accurate protein Indeed, use of WALTZ (540° duration) (Shaka et al. 1983) dynamics parameters, we compared the size of the slow and MLEV (360° duration) (Levitt and Freeman 1981; Lev- pulsing artifact for amide groups at 1 and 2 ppm H offset itt et al. 1982a, b) with shorter duration of the composite as function of magnetic field strength in Fig.  1c. High field pulse did not cause such high spikes. We next applied the strengths are not only attractive because of the sensitivity 90 –240 –90 CPD scheme, which was recently used to sup- x y x and resolution they afford, but also because they are more press artifacts from incomplete H decoupling in exchange sensitive to exchange processes as they increase the fre- mediated saturation transfer experiments (Chakrabarti et al. quency difference between states, Δϖ. However, for a given 2016). The 90 –240 –90 CPD sequence has a short over- x y x resonance, the offset from the decoupling field, and thus the all duration (420° rotation) and offers relatively broadband slow pulsing artifact, will increase with increasing magnetic inversion, free from off-resonance effects without relying on field strength. Strikingly, the artifact will already be signifi- supercycles (Levitt et al. 1982a; Levitt 1982). Gratifyingly, cant at 1 ppm offsets for medium-sized proteins in a future the 90 –240 –90 sequence effectively eliminated the arti- x y x 1.2 GHz spectrometer. To illustrate the impact of the slow facts without causing appreciable spikes or scatter in R 2,eff pulsing artifact, generated by the inability of CW irradiation values (Fig. 2d). The requirement for a short duration of the to decouple the full width of the amide spectrum, the HSQC CPD element also means that the broadband performance can be divided in three areas: a narrow region ± ~1 ppm of CPD decoupling cannot be used to reduce the decoupling around the carrier frequency of the decoupling field that will power. Tests showed that reducing the decupling power to be free of significant artifacts, the region beyond ± ~2 ppm 7 kHz (34 µs decoupling pulse) resulted in spurious artifacts in which significant artifacts will already occur at the lowest dominating the dispersion curves at high CPMG pulsing typical field strength, and the intermediate region (Fig.  1d). rates (data not shown). To confirm the results obtained from simulations, While successful in suppressing the slow-pulsing arti- we experimentally demonstrated the problem using the facts, the use of composite pulse sequences for decoupling ST–CW–CPMG pulse sequence on a sample containing results in systematic differences in R values compared to 2,eff azurin, a 16 kDa electron transfer metalloprotein (Adman those obtained using CW decoupling. This is most apparent 1991). A small subset of residues in azurin have been from the WALTZ data in Fig. 2b, showing systematically reported to undergo conformational exchange on the mil- reduced R values compared to the CW reference data. 2,eff lisecond timescale (Korzhnev et al. 2003). To emphasize the Such offsets between the CPD-derived and CW-derived dis- slow pulsing artifact, we purposely centered the H decou- persion curve are also found for MLEV and 90 –240 –90 x y x pling field at 16 ppm such that the dispersion profiles are decoupling, although typically much smaller. When using dominated by the artifact (Fig. 2a). Using this setup, we next the 90 –240 –90 sequence, the average offset over all resi- x y x −1 screened several broadband decoupling sequences for their dues was found to be ~ 0.3 s with 90% of the profiles hav - −1 ability to suppress the artifact. These sequences rely on com- ing offsets below 0.6 s (see Supplemental Table S1). Since posite pulses to offer good population inversion even in the this offset is small and the absolute value of R is not of 2,eff presence of off-resonance effects (Shaka and Keeler 1987), importance when fitting dispersion curves, it will have neg- and thus should be able to suppress the artifact in theory. As ligible impact on the usefulness of the data obtained with can be seen in Fig. 2b, a wide range of CPD schemes indeed CPD decoupling schemes. suppressed the artifact. Notably, the use of GARP (Shaka Having established that WALTZ, MLEV and 90 –240 –90 x y x et al. 1985) and DIPSI2 (Shaka et al. 1988) results in spuri- decoupling sequences are able to suppress the slow pulsing ous elevated R values at high pulsing rates, rendering the artifact, we further tested their efficacy in a regular experimen- 2,eff dispersion curves unusable. These spikes originate from the tal setup with the decoupling field centered at 8.2 ppm. The timing mismatch between the continuous train of (compos- obtained R values were compared point-by-point between 2,eff ite) 180° pulses on the H channel on the one hand and the the CPD and the CW data-set, and the root-mean-square repetition of free-evolution and 180° refocusing pulses on deviation (RMSD) between data sets was calculated with and the N channel on the other hand. This mismatch results without compensating for the systematic offset in R values 2,eff in incomplete decoupling at the end of each τ period and between the two datasets (Fig. 3a). Clearly, the 90 –240 –90 cp x y x thus elevated R values (Fig. 2c). As noted by Jiang et al. sequence performs best with an average RMSD to the refer- 2,eff −1 (2015), the duration of the mismatch is short when using ence CW data set of 0.17 s , which is on the order of the adequately high power CW H decoupling, and thus the experimental error. The high quality of the data is visible from 1 3 74 Journal of Biomolecular NMR (2018) 71:69–77 Fig. 2 Suppression of slow pulsing artifacts by composite pulse at the end of each composite pulse block (CPD) and the point (dot- 1 15 based broadband H decoupling. a Experimental relaxation disper- ted line) of complete N inversion. As a result, decoupling is incom- sion curves for three residues measured using the ST–CW–CPMG plete and resulting in elevated R values. d Experimental dispersion 2,eff 1 1 sequence with the H decoupling field centered at 16  ppm. Dotted curves for residue D62 using CW and 90 –240 –90 H decoupling x y x lines are fits to Eq.  1 to guide the eye. b Experimental dispersion schemes, both centered at 16  ppm. In b, d between brackets are the curves for residue D62 using the indicated H decoupling schemes, average root-mean-square deviation (RMSD) to a straight line over all centered at 16  ppm. c When the H decoupling power is fixed, all 114 residues in azurin. Note that the RMSD obtained with CW 1 −1 there is an inevitable timing mismatch between complete H inversion decoupling centered at 8.2 ppm was 0.7 s comparison of profiles obtained for residues with negligible Discussion 1 1 H offset, such as shown in Fig.  3b. At high H offset from the decoupling field, the CW data suffers from the slow pulsing We have investigated the impact of the slow-pulsing artifact artifact, which is absent when using the 90 –240 –90 CPD in in-phase N relaxation dispersion experiments by theo- x y x sequence, as exemplified for T52 in Fig.  3c. Notably, this retical considerations, numerical simulations and experi- residue shows the slow pulsing artifact superimposed on a ments. We show that the artifact can be removed by using genuine dispersion of R values. From the comparison to the CPD-based H decoupling during the CPMG period. Out of 2,eff CPD-based experiment, it becomes clear that the data point the tested CPD sequences, the 90 –240 –90 sequence offers x y x at 25 Hz ν pulsing rate is strongly affected by the slow the best performance: the artifact is fully suppressed, while CPMG −1 pulsing artifact with R value spuriously elevated by ~ 1 s . retaining shape of the dispersion curve obtained using CW 2,eff As a final experiment, we recorded both CW and CPD-based decoupling within experimental error. Notably, this is done dispersion profiles at the national ultra-high field NMR Facil- without introducing spurious spikes in R values at high 2,eff ity at 950 MHz. At this field, the resonance with the highest pulsing rates, and with minimal offset to the CW-based dis- 1 −1 H offset shows a slow-pulsing artifact of ~ 1.5 s in the CW persion profiles. Critical to its performance seems to be short experiment, which is effectively suppressed when using the duration of the composite pulse combined with relatively 90 –240 –90 decoupling sequence (Fig. 3d). high quality of off-resonance performance. x y x 1 3 Journal of Biomolecular NMR (2018) 71:69–77 75 Fig. 3 The 90 –240 –90 decoupling scheme offers high-quality an unassigned Arg sidechain resonance (Rsc) both with significant x y x dispersion curves free of slow-pulsing artifact. a Average RMSD H offset. Data for panels b, c recorded at 850  MHz. Data for panel between CPD- and CW-based dispersion curves over all analyzed d recorded at 950  MHz. Dotted lines are best-fit dispersion curves residues in azurin. Open/closed bars refer to the RMSD without/with obtained using CATIA (Hansen, http://www.bioch em.ucl.ac.uk/hanse compensating for the offset between the curves. b–d Experimental n/catia /). The CPD data were corrected for the systematic offset to the dispersion curves for both CW and 90 –240 –90 based experiments CW data before plotting x y x for amide resonances of A92 (no H offset, panel b), and T52, and The cause of the slight offset between the CPD and CW- water polarization. Experimental tests (Hiller et al. 2005) based dispersion profiles is unclear. Closer inspection shows demonstrate this effect and show that after a 2 s delay, cor - that the magnitude of the offset shows no correlation to responding to the recycle delay to the next proton excita- either the N, H , or H chemical shift and that both reference tion pulse, there is minimal difference between the water N α (no CPMG delay) and dispersion experiment (with CPMG polarization in the CPD and CW case (see Supplemental delay) have slightly altered intensities (~ 2–5%) in the CPD Fig. S1). Here, radiation damping caused by the high Q of experiment compared to the CW experiment. The effect on the cryogenic probe likely aids the recovery of the water the reference experiment, where the decoupling block is car- magnetization in the CPD case. Additionally, the low pH ried out before the recycle delay, signifies that the both types of the sample (5.5) will slow down amide-water exchange of decoupling result in a different steady-state magnetiza - and thus additionally dampen the effect of (residual) water tion, presumably both for water and protein protons. saturation. As for the water magnetization, a disadvantage of using In the original implementation of the in-phase dispersion CPD over CW decoupling is the loss of control over its experiment described by Hansen et al. (2008b), the strength state. Whereas in the CW case the water magnetization is of the decoupling field is matched to the CPMG pulsing spin-locked and returned to + z after the CPMG period, con- rate to avoid the timing mismatch as indicated in Fig. 2c. tinuous alteration between x and y-pulse phase during the In principle, such matching could also be done when using 90 –240 –90 CPD element causes dephasing and loss of CPD decoupling schemes, which should result in decreased x y x 1 3 76 Journal of Biomolecular NMR (2018) 71:69–77 Allard P, Helgstrand M, Hard T (1998) The complete homogeneous scatter in the dispersion curves. While simulations indeed master equation for a heteronuclear two-spin system in the basis show such improvement in performance, an experimental of cartesian product operators. J Magn Reson 134:7–16 test showed a severe increase in scatter, presumably due to Chakrabarti KS et al (2016) High-power 1H composite pulse decou- a point-to-point variation in the steady state of the water and pling provides artifact free exchange-mediated saturation trans- fer (EST) experiments. J Magn Reson 269:65–69 aliphatic proton magnetization. Culik RM et al (2018) Effects of maturation on the conformational As noted in Fig. 1, the slow-pulsing artifact will be par- free-energy landscape of SOD1. Proc Natl Acad Sci 115:E2546 ticularly problematic at high magnetic field strengths. At Delaforge E et al (2018) Deciphering the dynamic interaction profile such high fields, it may be better to use TROSY–CPMG of an intrinsically disordered protein by NMR exchange spec- troscopy. J Am Chem Soc 140:1148–1158 sequences (Loria et al. 1999a), which do not suffer from Delaglio F et  al (1995) NMRPipe: a multidimensional spectral the slow-pulsing artifact, even for non-deuterated mod- processing system based on UNIX pipes. J Biomol NMR erately sized proteins. The relative sensitivity of TROSY 6:277–293 and in-phase CPMG experiments is best assessed experi- Eaton JW, Bateman D, Hauberg S (2008) GNU Octave manual version 3. Network Theory Ltd., Bristol, p 568 mentally as it not only depends on magnetic field strength Eisenmesser EZ et al (2005) Intrinsic dynamics of an enzyme underlies but also on protein size, labeling pattern, and temperature. catalysis. Nature 438:117 Next to the absolute sensitivity, one may also consider that Fawzi NL, Ying J, Torchia DA, Clore GM (2010) Kinetics of amyloid lower N relaxation rates during the CPMG period allow β monomer to oligomer exchange by NMR relaxation. J Am Chem Soc 132:9948–9951 the use of longer CPMG delays, increasing the sensitivity Franco R, Gil-Caballero S, Ayala I, Favier A, Brutscher B (2017) Prob- to slow motions (Loria et al. 1999a), as well as spectral ing conformational exchange dynamics in a short-lived protein quality of TROSY spectra (reduced overlap vs. presence folding intermediate by real-time relaxation–dispersion NMR. J of anti-TROSY lines). Additionally, in case data at lower Am Chem Soc 139:1065–1068 Frauenfelder H, Sligar SG, Wolynes PG (1991) The energy landscapes field strength have been recorded using the in-phase CPMG and motions of proteins. Science 254:1598–1603 experiment it may be necessary to record these at high fields Gopalan AB, Hansen DF, Vallurupalli P (2018) CPMG experiments too. for protein minor conformer structure determination. In: Ghose R In conclusion, we show here that the use of broadband H (ed) Protein NMR: methods and protocols. Springer, New York, pp 223–242 decoupling, in particular using the 90 –240 –90 sequence, x y x 15 Hammes GG (1964) Mechanism of enzyme catalysis. Nature 204:342 is a viable and attractive option for recording in-phase N Hansen F. http://www.bioch em.ucl.ac.uk/hanse n/catia /. Accessed 1 relaxation dispersion data. This option is particularly rel- June 2018 evant when the protein spectrum contains resonances far Hansen F. http://www .bioch em.ucl.ac.uk/hanse n/fuda/. Accessed 1 June 2018 from center. It offers artifact-free dispersion profiles without Hansen DF, Vallurupalli P, Lundström P, Neudecker P, Kay LE (2008a) the need for recording data in multiple sets or the need for Probing chemical shifts of invisible states of proteins with relaxa- eliminating of data points, all without compromising data tion dispersion NMR spectroscopy: how well can we do? J Am quality. Chem Soc 130:2667–2675 Hansen DF, Vallurupalli P, Kay LE (2008b) An improved 15N relaxa- tion dispersion experiment for the measurement of millisecond Acknowledgements We thank prof. Daiwen Yang (National University time-scale dynamics in proteins. J Phys Chem B 112:5898–1904 of Singapore) for sharing the pulse sequence code for the ST-CW- Helgstrand M, Härd T, Allard P (2000) Simulations of NMR pulse CPMG experiment. This work was supported by financial support from sequences during equilibrium and non-equilibrium chemical the Dutch Science Foundation NWO by a VIDI grant (723.013.010) exchange. J Biomol NMR 18:49–63 to HvI and grant 184.032.207 to the uNMR-NL National Roadmap Henzler-Wildman K, Kern D (2007) Dynamic personalities of proteins. Large-Scale Facility of the Netherlands. Nature 450:964–972 Henzler-Wildman KA et al (2007) Intrinsic motions along an enzy- Open Access This article is distributed under the terms of the Crea- matic reaction trajectory. Nature 450:838 tive Commons Attribution 4.0 International License (http://creat iveco Hiller S, Wider G, Etezady-Esfarjani T, Horst R, Wüthrich K (2005) mmons.or g/licenses/b y/4.0/), which permits unrestricted use, distribu- Managing the solvent water polarization to obtain improved NMR tion, and reproduction in any medium, provided you give appropriate spectra of large molecular structures. J Biomol NMR 32:61–70 credit to the original author(s) and the source, provide a link to the Jiang B, Yu B, Zhang X, Liu M, Yang D (2015) A (15)N CPMG relaxa- Creative Commons license, and indicate if changes were made. tion dispersion experiment more resistant to resonance offset and pulse imperfection. J Magn Reson 257:1–7 Karlsson BG, Pascher T, Nordling M, Arvidsson RHA, Lundberg LG (1989) Expression of the blue copper protein azurin from Pseu- domonas aeruginosa in Escherichia coli. FEBS Lett 246:211–217 References Kim TH et  al (2017) The role of dimer asymmetry and protomer dynamics in enzyme catalysis. Science. https ://doi.org/10.1126/ Adman ET (1991) Copper protein structures. Adv Protein Chem scien ce.aag23 55 42:145–197 Kimsey IJ, Petzold K, Sathyamoorthy B, Stein ZW, Al-Hashimi HM Akke M, Palmer AG (1996) Monitoring macromolecular motions (2015) Visualizing transient Watson-Crick like mispairs in DNA on microsecond to millisecond time scales by R1ρ–R1 constant and RNA duplexes. Nature 519:315–320 relaxation time NMR spectroscopy. J Am Chem Soc 118:911–912 1 3 Journal of Biomolecular NMR (2018) 71:69–77 77 Korzhnev DM, Karlsson BG, Orekhov VY, Billeter M (2003) NMR anisotropic chemical shift relaxation mechanisms in the measure- detection of multiple transitions to low-populated states in azurin. ment of spin-spin relaxation rates. Mol Phys 75:699–711 Protein Sci 12:56–65 Pratihar S et al (2016) Kinetics of the antibody recognition site in Korzhnev DM, Religa TL, Banachewicz W, Fersht AR, Kay LE (2010) the third IgG-binding domain of protein G. Angew Chem Int Ed A transient and low-populated protein-folding intermediate at 55:9567–9570 atomic resolution. Science 329:1312–1316 Sauerwein A, Hansen D (2015) Relaxation dispersion NMR spectros- Kovermann M, Rogne P, Wolf-Watz M (2016) Protein dynamics and copy. In: Berliner L (ed) Protein NMR. Biological magnetic reso- function from solution state NMR spectroscopy. Q Rev Biophys nance. Springer, Boston 49:e6 Schneider R et al (2015) Visualizing the molecular recognition tra- Levitt MH (1982) Symmetrical composite pulse sequences for NMR jectory of an intrinsically disordered protein using multinuclear population inversion. II. Compensation of resonance offset. J relaxation dispersion NMR. J Am Chem Soc 137:1220–1229 Magn Reson (1969) 50:95–110 Shaka AJ, Keeler J (1987) Broadband spin decoupling in isotropic- Levitt MH, Freeman R (1981) Composite pulse decoupling. J Magn liquids. Prog Nucl Magn Reson Spectrosc 19:47–129 Reson 43:502–507 Shaka AJ, Keeler J, Frenkiel T, Freeman R (1983) An improved Levitt MH, Freeman R, Frenkiel T (1982a) Broadband heteronuclear sequence for broadband decoupling: WALTZ-16. J Magn Reson decoupling. J Magn Reson (1969) 47:328–330 (1969) 52:335–338 Levitt MH, Freeman R, Frenkiel T (1982b) Supercycles for broadband Shaka AJ, Barker PB, Freeman R (1985) Computer-optimized decou- heteronuclear decoupling. J Magn Reson 50:157–160 pling scheme for wideband applications and low-level operation. Libich DS, Tugarinov V, Clore GM (2015) Intrinsic unfoldase/foldase J Magn Reson (1969) 64:547–552 activity of the chaperonin GroEL directly demonstrated using Shaka AJ, Lee CJ, Pines A (1988) Iterative schemes for bilinear multinuclear relaxation-based NMR. Proc Natl Acad Sci USA operators; application to spin decoupling. J Magn Reson (1969) 112:8817–8823 77:274–293 Lisi GP (2016) & Patrick Loria, J. Using NMR spectroscopy to eluci- Sugase K, Dyson HJ, Wright PE (2007) Mechanism of coupled fold- date the role of molecular motions in enzyme function. Prog Nucl ing and binding of an intrinsically disordered protein. Nature Magn Reson Spectrosc 92–93:1–17 447:1021 Loria JP, Rance M, Palmer AG (1999a) A TROSY CPMG sequence Vallurupalli P, Bouvignies G, Kay LE (2012) Studying “invisible” for characterizing chemical exchange in large proteins. J Biomol excited protein states in slow exchange with a major state confor- NMR 15:151–155 mation. J Am Chem Soc 134:8148–8161 Loria JP, Rance M, Palmer AG (1999b) A relaxation-compensated Vallurupalli P, Sekhar A, Yuwen T, Kay LE (2017) Probing conforma- Carr–Purcell–Meiboom–Gill sequence for characterizing chemical tional dynamics in biomolecules via chemical exchange saturation exchange by NMR spectroscopy. J Am Chem Soc 121:2331–2332 transfer: a primer. J Biomol NMR 67:243–271 Massi F, Peng JW (2018) Characterizing protein dynamics with NMR Wolynes PG (2005) Recent successes of the energy landscape theory of R 1ρ relaxation experiments. In: Ghose R (ed) Protein NMR: protein folding and function. Q Rev Biophys 38:405–410 methods and protocols. Springer, New York, pp 205–221 Xiao T, Fan Js, Zhou H, Lin Q, Yang D (2016) Local unfolding of fatty Neudecker P et al (2012) Structure of an intermediate state in protein acid binding protein to allow ligand entry for binding. Angew folding and aggregation. Science 336:362–366 Chem Int Ed 55:6869–6872 Palmer AG 3rd, Kroenke CD, Loria JP (2001) Nuclear magnetic Xue Y et al (2015) Characterizing RNA excited states using NMR resonance methods for quantifying microsecond-to-millisec- relaxation dispersion. Methods Enzymol 558:39–73 ond motions in biological macromolecules. Methods Enzymol Yip G, Zuiderweg E (2004) A phase cycle scheme that significantly 339:204–238 suppresses offset-dependent artifacts in the R 2-CPMG 15 N Palmer AG (2015) Enzyme dynamics from NMR spectroscopy. Acc relaxation experiment. J Magn Reson 171:25–36 Chem Res 48:457–465 Zhao B, Guffy SL, Williams B, Zhang Q (2017) An excited state under - Palmer AG, Skelton NJ, Chazin WJ, Wright PE, Rance M (1992) Sup- lies gene regulation of a transcriptional riboswitch. Nat Chem pression of the effects of cross-correlation between dipolar and Biol 13:968–974 1 3 http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Journal of Biomolecular NMR Springer Journals

Removal of slow-pulsing artifacts in in-phase 15N relaxation dispersion experiments using broadband 1H decoupling

Free
9 pages

Loading next page...
 
/lp/springer_journal/removal-of-slow-pulsing-artifacts-in-in-phase-15n-relaxation-FwvvqLnQMR
Publisher
Springer Journals
Copyright
Copyright © 2018 by The Author(s)
Subject
Physics; Biological and Medical Physics, Biophysics; Biochemistry, general; Spectroscopy/Spectrometry
ISSN
0925-2738
eISSN
1573-5001
D.O.I.
10.1007/s10858-018-0193-2
Publisher site
See Article on Publisher Site

Abstract

Understanding of the molecular mechanisms of protein function requires detailed insight into the conformational landscape accessible to the protein. Conformational changes can be crucial for biological processes, such as ligand binding, protein folding, and catalysis. NMR spectroscopy is exquisitely sensitive to such dynamic changes in protein conformations. In particular, Carr–Purcell–Meiboom–Gill (CPMG) relaxation dispersion experiments are a powerful tool to investigate pro- tein dynamics on a millisecond time scale. CPMG experiments that probe the chemical shift modulation of N in-phase magnetization are particularly attractive, due to their high sensitivity. These experiments require high power H decoupling 15 15 during the CPMG period to keep the N magnetization in-phase. Recently, an improved version of the in-phase N-CPMG experiment was introduced, offering greater ease of use by employing a single H decoupling power for all CPMG puls- ing rates. In these experiments however, incomplete decoupling of off-resonance amide H spins introduces an artefactual dispersion of relaxation rates, the so-called slow-pulsing artifact. Here, we analyze the slow-pulsing artifact in detail and demonstrate that it can be suppressed through the use of composite pulse decoupling (CPD). We report the performances of various CPD schemes and show that CPD decoupling based on the 90 –240 –90 element results in high-quality dispersion x y x curves free of artifacts, even for amides with high H offset. Keywords Relaxation dispersion · Decoupling · Protein dynamics · Composite pulse Introduction (~ ms), these excited states can be crucial for biologically important processes such as enzyme catalysis (Hammes Biological macromolecules such as nucleic acids and pro- 1964; Eisenmesser et  al. 2005; Henzler-Wildman et  al. teins are non-rigid entities that can populate a variety of 2007; Palmer 2015; Kim et al. 2017), ligand binding or pro- conformers in their energy landscape (Frauenfelder et al. tein–protein interactions (Sugase et al. 2007; Schneider et al. 1991; Wolynes 2005; Henzler-Wildman and Kern 2007). 2015; Pratihar et al. 2016; Xiao et al. 2016; Zhao et al. 2017; The lowest energy conformation, the ground state, is often Delaforge et al. 2018), and protein folding (Korzhnev et al. able to transiently access higher-energy conformations. Even 2010; Neudecker et al. 2012; Kimsey et al. 2015; Libich when their population is low (< 10%) and life times is short et al. 2015; Franco et al. 2017; Culik et al. 2018). While these states cannot be detected directly due to their transient and lowly populated nature, NMR experiments (Akke and Electronic supplementary material The online version of this Palmer 1996; Fawzi et al. 2010; Kovermann et al. 2016) are article (https ://doi.org/10.1007/s1085 8-018-0193-2) contains uniquely able to provide a detailed, atomistic description of supplementary material, which is available to authorized users. the energy landscape. In particular, relaxation dispersion and * Hugo van Ingen chemical exchange saturation transfer experiments are par- h.vaningen@uu.nl ticularly powerful herein, as they give access to the popula- tion, life times and structures of excited states (Palmer et al. Macromolecular Biochemistry, Leiden Institute of Chemistry, Leiden University, P.O Box 9502, 2001; Vallurupalli et al. 2012, 2017; Sauerwein and Hansen 2300 RA Leiden, The Netherlands 2015; Xue et al. 2015; Lisi 2016; Massi and Peng 2018; Present Address: NMR Group, Bijvoet Center Gopalan et al. 2018). for Biomolecular Research, Utrecht University, Padualaan 8, 3584 CH Utrecht, The Netherlands Vol.:(0123456789) 1 3 70 Journal of Biomolecular NMR (2018) 71:69–77 In Carr–Purcell–Meiboom–Gill (CPMG) relaxation dis- Here, we analyze the slow-pulsing artifact in N persion experiments, the characterization of the minor state CW–CPMG sequences in detail and demonstrate a simple is derived from the major state peaks by measurement of method for its removal. In that, we took inspiration from the their effective transverse relaxation rate R , as a function work of Chakrabarti et al. (2016), where composite pulse 2,eff of the pulsing rate in the CPMG period. Signals of nuclear decoupling (CPD) was used to suppress H off-resonance spins that experience exchange between states with differ - effects in exchange mediated saturation transfer experi- ent chemical shifts are affected by exchange-induced line ments. We investigated the performance of various CPD broadening, an effect that depends on the free precession schemes in CW–CPMG sequences and demonstrate here interval (2τ ) between the refocusing pulses in the CPMG that high power CPD based on the 90 –240 –90 element cp x y x element (Palmer et al. 2001; Sauerwein and Hansen 2015). achieves artifact-free dispersion curves over a wide range Analysis of the resulting relaxation dispersion curve, a plot of H offsets. of the R versus CPMG frequency (1/4τ ), allows deter- 2,eff cp mination of the rate of exchange (k ), population of minor ex state (p ) and the absolute chemical shift difference (|Δϖ|) Materials and methods between the exchanging states. Importantly, since the shape of the dispersion profile depends on Δϖ , data is typically NMR samples acquired at two fields to accurately determine the exchange parameters (Sauerwein and Hansen 2015). NMR experiments were recorded on a sample of 2.5 mM 15 15 13 The N backbone amide spin is the most popular uniformly N/ C-labelled Cu(II) azurin in 25 mM potas- nucleus for CPMG RD experiments, due to the simplicity sium phosphate buffer at pH 5.49 with 5% D O. Labelled of isotope-labeling, the straightforwardness of the two-spin azurin was produced and purified according to a previously 1 15 H– N spin system, and the high sensitivity and resolution published protocol with modifications for incorporating 13 15 afforded by these experiments. A critical aspect of these C-glucose and N-ammoniumchloride (Karlsson et al. experiments is appropriate handling of differences in the 1989). intrinsic R of the in-phase (N ) and anti-phase (2N H ) 2 x,y x,y z N magnetization which are generated in the free evolution NMR experiments periods. Anti-phase terms have higher intrinsic relaxation rates due to a contribution of H spin flips to their decay. Relaxation dispersion experiments, using the ST–CW–CPMG The original implementation of the N CPMG RD experi- sequence, were recorded at 298 K on Bruker Avance III HD ment uses a relaxation-compensation scheme to average the spectrometers operating at 850 and 950 MHz H Larmor fre- N and 2N H relaxation rates (Loria et al. 1999b). The quency and equipped with TCI cryoprobes. The constant-time x,y x,y z N CPMG sequence of Hansen et al. (2008b) (CW–CPMG) CPMG relaxation delay (T ) was set to 40 ms with ν set relax CPMG measures the dispersion profile of pure in-phase N by to 25, 50, 75, 100, 125, 175, 225 (2×), 275, 300, 350, 400 (2×), x,y applying high-power continuous wave (CW) H decoupling 500, 550, 600, 650, 700 (2×), 750, 800 (2×), 850, 900, 950 and during the CPMG train, offering enhanced sensitivity for 1000 Hz respectively, run in an interleaved manner. Duplicates non-deuterated proteins. Recently, Jiang et al. (2015) modi- were used to estimate the error in R . The errors were set to 2,eff −1 15 fied this sequence (ST–CW–CPMG) to use a single CPMG 0.2 s at minimum. The pulse length of the N refocusing train with the Yip and Zuiderweg phase cycle (2004) and pulses in the CPMG train was 90 µs. For H decoupling, either a single CW decoupling power, yielding dispersion curves CW decoupling or a CPD-scheme (GARP, DIPSI, MLEV16, free of off-resonance artifacts for a wider range of N offset WALTZ16, 90 –240 –90 ) was used. This was implemented by x y x frequencies. changing the “cw:f1” statement in the pulse program to read Both CW–CPMG sequences are nevertheless sensitive “cpds1:f1” (pulse program available upon request). In either 1 1 to artifacts from H off-resonance effects (Hansen et  al. case, the decoupling field strength was 14.7 kHz (17 µs H 90° 1 1 2008b; Yip and Zuiderweg 2004). Amide H spins that pulse), applied at 8.2 ppm H offset. A total of 3072/120 points 1 15 are far off-resonance from the CW decoupling field are not were acquired in the H/ N dimension with an acquisition fully decoupled from the N spin, resulting in generation time of 90/27.85 ms and a relaxation delay of 2 s and 4 scans of 2N H magnetization through the residual J-coupling. per FID. A reference spectrum, without the relaxation delay, x,y z Consequently, higher R values will be measured for low was also recorded. NMR data were processed with NMRPipe 2,eff ν values, for which free precession periods are long and (Delaglio et al. 1995), using linear prediction in the N dimen- CPMG more of the antiphase terms will be generated. This so-called sion and Lorentz-to-Gauss window functions. Peak volumes slow-pulsing artifact shows up as an artefactual dispersion were obtained by peak fitting using FuDa (Hansen, http://www. curve, interfering with accurate extraction of minor-state bioch em.ucl.ac.uk/hanse n/fuda/), and subsequently converted parameters. into effective relaxation rates via R (ν ) = − 1/T ·ln(I 2,eff CPMG relax 1 3 Journal of Biomolecular NMR (2018) 71:69–77 71 −1 (ν )/I ), where I is the peak intensity in a reference spec- of 13.6 and 26.7 s respectively. Dispersion experiments CPMG 0 0 trum recorded without the relaxation delay T . The R val- were simulated with T set to 40 ms, and ν values relax 2,eff relax CPMG ues measured using the ST–CW–CPMG sequence were cor- ranging from 25 to 1000 Hz, the N 180° refocusing pulse rected for R -contribution according to the formula described was set to 90 µs, H CW decoupling field strength was set to −1 1 by Jiang et al. (2015) using an estimate of 0.95 s R - and 14.7 kHz (17 µs H 90°). −1 10.5 s for R -contribution for all residues. Dispersion curves obtained with either CW or CPD decoupling were compared by calculating the RSMD between the curves for all residues: Results Measurement of in-phase N CPMG relaxation dispersion CPD,i CW,i RMSD = R − R 1 2,eff 2,eff profiles critically relies on H decoupling to measure the i=1 pure in-phase N relaxation rate without contamination by x,y the anti-phase relaxation rate. As pointed out in the work where i is the index of a particular ν value and the sum- CPMG of Jiang et al. (2015), the decoupling field strength has a mation runs over the N recorded points, equal to the num- practical limit of roughly 14 kHz, resulting in a residual J ber of points per dispersion curve (M) times the number coupling interaction for amide protons at non-zero offset to of residues. The systematic difference between the CW or the decoupling field. This interaction causes slow intercon- CPD-based dispersion curves was calculated from the aver- version of in-phase and anti-phase magnetization during the age point-by-point difference per residue and is tabulated in CPMG period, which will lead to undesired averaging of the Table S1. To compensate for these systematic differences, an in-phase and anti-phase relaxation rates (Loria et al. 1999b; “R -offset compensated” RMSD was calculated by replac - Hansen et al. 2008b). To first approximation, this averaging ing the CPD-based R values with the offset compensated 2,eff can be described by the equation derived by Palmer et al. values: (1992) for calculating the effective relaxation rate in spin echo sequences. Here, it is adapted and reformulated to CPD,compensated,i CPD,i CPD,i CW,i R = R − R − R express to the size of the slow-pulsing artifact A: 2,eff 2.eff 2,eff 2,eff i=1 anti in A = R − R 1 − sincJ 2 (1) r cp 2 2 Simulation of  N CW–CPMG dispersion profiles in anti 15 where R and R are the N in-phase and anti-phase 2 2 To evaluate the magnitude of the slow-pulsing artifact in transverse relaxation rates, J is the residual J-coupling, and relaxation dispersion profiles, numerical simulations of 2τ is the inter-pulse delay in the CPMG pulse train. In the cp 1 15 a two-spin H– N system were carried out, assuming a limit of perfect decoupling J ≈ 0, the sinc factor approaches non-deuterated protein. The evolution of magnetization 1 and A ≈ 0 for all τ values. For non-zero J , A approaches cp r in this spin system was calculated for the CPMG part of zero in the limit of fast pulsing where τ is very small. For cp the CPMG–CW and CPMG–ST–CW sequence, including slow pulsing, however, there is a non-zero artifact, with a the flanking N 90° pulses. Simulations in the absence theoretical limit of 0.5 (R − R ) for infinitely slow puls - 2,anti 2,in of exchange and neglecting pulse imperfections were per- ing. In practice, J can be as much 16 Hz (for 3 ppm H offset formed in operator space by solving the complete homogene- at 850  MHz) and 2τ is typically at most 20  ms, which cp ous master equation as described by Allard et al. (1998) and would generate a maximum artifact of roughly 10% of the Helgstrand et al. (2000) using the open source computing difference between the anti-phase and in-phase relaxation language GNU Octave (http://www.gnu.org/softw are/octav rate. e/) (Eaton et al. 2008). All simulation used the parameters To assess more precisely how the slow pulsing artifact detailed below unless noted otherwise. The N spin was is manifested in N CPMG–CW and ST–CW experiments, assumed to be on-resonance. The magnetic field strength numerical simulations of these sequences were performed was set to 19.9 T, corresponding to H Larmor frequency in Liouville space for a non-exchanging two spin N–H sys- of 850 MHz. Relaxation rates were calculated using overall tem. Figure 1a compares the obtained dispersion profiles rotational correlation time τ of 9 ns, a value of 0.85 for the for the two experiments with the predicted curve based on squared generalized order parameter, 100 ps for the correla- Eq. 1, for a N–H system with 3 ppm H offset at an 850 MHz tion time for internal motions, and − 172/+10 ppm for the spectrometer. Whereas a flat curve is expected for a non- 15 1 N/ H chemical shift anisotropy. Relaxation due to neigh- exchanging system, systematically increased R values 2,eff boring protons was included as described in ref. (Allard are measured in the slow pulsing regime for both pulse et al. 1998) by including a virtual proton at 1.85 Å, resulting sequences. While Eq. 1 is derived for periods of free evolu- in R values of in-phase and anti-phase N magnetizations tion in absence of a decoupling field, the curvature of the 1 3 72 Journal of Biomolecular NMR (2018) 71:69–77 Fig. 1 Impact of the slow-pulsing artifact on simulated relaxation dis- netic field strength for 1 and 2  ppm H offset for proteins of 4 and persion profiles. a Simulated slow-pulsing artifact caused by incom- 9  ns tumbling times. The gray area indicates the typical experimen- −1 1 plete J decoupling in the CW–CPMG and the ST–CW–CPMG tal error in range of 0.1–0.3  s . d The typical accessible H offset NH 15 15 1 implementation of the in-phase N CPMG experiment. Solid lines ranges, color coded into a N– H HSQC spectrum. Assuming the are fits obtained using the program CATIA (Hansen, http://www. H CW field is centered at 8  ppm, the blue region is accessible up bioch em.ucl.ac.uk/hanse n/catia /) assuming two-site exchange. The to the highest magnetic fields, orange is accessible up to 600  MHz, artifact expected based on Eq. 1 is shown for comparison. The boxed and the red region is inaccessible. In a–c simulated profiles are shown region is expanded in the inset. The H offset from the decoupling for both CW–CPMG (open triangle) and ST–CPMG (asterisk) pulse field was set to 3  ppm, assuming an 850  MHz spectrometer. b, c sequences; color coding indicated in the figure. All simulations are Maximum size of the artifact (ΔR ) as a function of (b) H offset based on a non-exchanging N–H spin system. Simulation parameters 2,eff for proteins of 4, 6.5 and 9 ns correlation times at 850 MHz; c mag- are given in “Materials and Methods” section, unless noted otherwise slow-pulsing artifact matches the predicted sinc dependence in Fig. 1a), illustrating the potential impact on the extracted on the pulsing rate. The size of the artifact is somewhat exchange parameters. underestimated by Eq. 1. The original CW sequence shows Since the size of the slow pulsing artifact is governed by slightly lower sensitivity to the artifact than the ST–CW the relaxation difference between in-phase and anti-phase experiment. This difference can be traced back to presence magnetization, it is dependent on protein size. Large proteins 15 1 1 of the N refocussing pulse in between the two halves of have more efficient H– H spin flips which increase the anti- the total CPMG period in the CW experiment. Importantly, phase relaxation rate. Figure 1b compares the magnitude of since the shape of the artifact is virtually indistinguishable the artifact for three different protein sizes as function of H from a bona-fide dispersion profile, the artefactual R val- offset from the decoupling field. For larger proteins, where 2,eff ues can be fitted to an actual dispersion curve (see solid lines the chance of finding amide protons at high offset is also 1 3 Journal of Biomolecular NMR (2018) 71:69–77 73 −1 higher, the artifact can be well above 1 s . At offsets larger effect is small. Both DIPSI2 and GARP use particularly long than ~ 1000 Hz the slow pulsing artifact will be higher than composite pulses (corresponding to the length of 2590° and −1 the typical experimental error (on the order of 0.1–0.3 s ) 1054° rotation, respectively), which aggravates the impact (Hansen et al. 2008a), as also noted by Jiang et al. (2015). of the timing mismatch, in particular at high pulsing rates, Since relaxation dispersion data need to be acquired where the effects from each τ period are compounded. cp at two magnetic fields in order to extract accurate protein Indeed, use of WALTZ (540° duration) (Shaka et al. 1983) dynamics parameters, we compared the size of the slow and MLEV (360° duration) (Levitt and Freeman 1981; Lev- pulsing artifact for amide groups at 1 and 2 ppm H offset itt et al. 1982a, b) with shorter duration of the composite as function of magnetic field strength in Fig.  1c. High field pulse did not cause such high spikes. We next applied the strengths are not only attractive because of the sensitivity 90 –240 –90 CPD scheme, which was recently used to sup- x y x and resolution they afford, but also because they are more press artifacts from incomplete H decoupling in exchange sensitive to exchange processes as they increase the fre- mediated saturation transfer experiments (Chakrabarti et al. quency difference between states, Δϖ. However, for a given 2016). The 90 –240 –90 CPD sequence has a short over- x y x resonance, the offset from the decoupling field, and thus the all duration (420° rotation) and offers relatively broadband slow pulsing artifact, will increase with increasing magnetic inversion, free from off-resonance effects without relying on field strength. Strikingly, the artifact will already be signifi- supercycles (Levitt et al. 1982a; Levitt 1982). Gratifyingly, cant at 1 ppm offsets for medium-sized proteins in a future the 90 –240 –90 sequence effectively eliminated the arti- x y x 1.2 GHz spectrometer. To illustrate the impact of the slow facts without causing appreciable spikes or scatter in R 2,eff pulsing artifact, generated by the inability of CW irradiation values (Fig. 2d). The requirement for a short duration of the to decouple the full width of the amide spectrum, the HSQC CPD element also means that the broadband performance can be divided in three areas: a narrow region ± ~1 ppm of CPD decoupling cannot be used to reduce the decoupling around the carrier frequency of the decoupling field that will power. Tests showed that reducing the decupling power to be free of significant artifacts, the region beyond ± ~2 ppm 7 kHz (34 µs decoupling pulse) resulted in spurious artifacts in which significant artifacts will already occur at the lowest dominating the dispersion curves at high CPMG pulsing typical field strength, and the intermediate region (Fig.  1d). rates (data not shown). To confirm the results obtained from simulations, While successful in suppressing the slow-pulsing arti- we experimentally demonstrated the problem using the facts, the use of composite pulse sequences for decoupling ST–CW–CPMG pulse sequence on a sample containing results in systematic differences in R values compared to 2,eff azurin, a 16 kDa electron transfer metalloprotein (Adman those obtained using CW decoupling. This is most apparent 1991). A small subset of residues in azurin have been from the WALTZ data in Fig. 2b, showing systematically reported to undergo conformational exchange on the mil- reduced R values compared to the CW reference data. 2,eff lisecond timescale (Korzhnev et al. 2003). To emphasize the Such offsets between the CPD-derived and CW-derived dis- slow pulsing artifact, we purposely centered the H decou- persion curve are also found for MLEV and 90 –240 –90 x y x pling field at 16 ppm such that the dispersion profiles are decoupling, although typically much smaller. When using dominated by the artifact (Fig. 2a). Using this setup, we next the 90 –240 –90 sequence, the average offset over all resi- x y x −1 screened several broadband decoupling sequences for their dues was found to be ~ 0.3 s with 90% of the profiles hav - −1 ability to suppress the artifact. These sequences rely on com- ing offsets below 0.6 s (see Supplemental Table S1). Since posite pulses to offer good population inversion even in the this offset is small and the absolute value of R is not of 2,eff presence of off-resonance effects (Shaka and Keeler 1987), importance when fitting dispersion curves, it will have neg- and thus should be able to suppress the artifact in theory. As ligible impact on the usefulness of the data obtained with can be seen in Fig. 2b, a wide range of CPD schemes indeed CPD decoupling schemes. suppressed the artifact. Notably, the use of GARP (Shaka Having established that WALTZ, MLEV and 90 –240 –90 x y x et al. 1985) and DIPSI2 (Shaka et al. 1988) results in spuri- decoupling sequences are able to suppress the slow pulsing ous elevated R values at high pulsing rates, rendering the artifact, we further tested their efficacy in a regular experimen- 2,eff dispersion curves unusable. These spikes originate from the tal setup with the decoupling field centered at 8.2 ppm. The timing mismatch between the continuous train of (compos- obtained R values were compared point-by-point between 2,eff ite) 180° pulses on the H channel on the one hand and the the CPD and the CW data-set, and the root-mean-square repetition of free-evolution and 180° refocusing pulses on deviation (RMSD) between data sets was calculated with and the N channel on the other hand. This mismatch results without compensating for the systematic offset in R values 2,eff in incomplete decoupling at the end of each τ period and between the two datasets (Fig. 3a). Clearly, the 90 –240 –90 cp x y x thus elevated R values (Fig. 2c). As noted by Jiang et al. sequence performs best with an average RMSD to the refer- 2,eff −1 (2015), the duration of the mismatch is short when using ence CW data set of 0.17 s , which is on the order of the adequately high power CW H decoupling, and thus the experimental error. The high quality of the data is visible from 1 3 74 Journal of Biomolecular NMR (2018) 71:69–77 Fig. 2 Suppression of slow pulsing artifacts by composite pulse at the end of each composite pulse block (CPD) and the point (dot- 1 15 based broadband H decoupling. a Experimental relaxation disper- ted line) of complete N inversion. As a result, decoupling is incom- sion curves for three residues measured using the ST–CW–CPMG plete and resulting in elevated R values. d Experimental dispersion 2,eff 1 1 sequence with the H decoupling field centered at 16  ppm. Dotted curves for residue D62 using CW and 90 –240 –90 H decoupling x y x lines are fits to Eq.  1 to guide the eye. b Experimental dispersion schemes, both centered at 16  ppm. In b, d between brackets are the curves for residue D62 using the indicated H decoupling schemes, average root-mean-square deviation (RMSD) to a straight line over all centered at 16  ppm. c When the H decoupling power is fixed, all 114 residues in azurin. Note that the RMSD obtained with CW 1 −1 there is an inevitable timing mismatch between complete H inversion decoupling centered at 8.2 ppm was 0.7 s comparison of profiles obtained for residues with negligible Discussion 1 1 H offset, such as shown in Fig.  3b. At high H offset from the decoupling field, the CW data suffers from the slow pulsing We have investigated the impact of the slow-pulsing artifact artifact, which is absent when using the 90 –240 –90 CPD in in-phase N relaxation dispersion experiments by theo- x y x sequence, as exemplified for T52 in Fig.  3c. Notably, this retical considerations, numerical simulations and experi- residue shows the slow pulsing artifact superimposed on a ments. We show that the artifact can be removed by using genuine dispersion of R values. From the comparison to the CPD-based H decoupling during the CPMG period. Out of 2,eff CPD-based experiment, it becomes clear that the data point the tested CPD sequences, the 90 –240 –90 sequence offers x y x at 25 Hz ν pulsing rate is strongly affected by the slow the best performance: the artifact is fully suppressed, while CPMG −1 pulsing artifact with R value spuriously elevated by ~ 1 s . retaining shape of the dispersion curve obtained using CW 2,eff As a final experiment, we recorded both CW and CPD-based decoupling within experimental error. Notably, this is done dispersion profiles at the national ultra-high field NMR Facil- without introducing spurious spikes in R values at high 2,eff ity at 950 MHz. At this field, the resonance with the highest pulsing rates, and with minimal offset to the CW-based dis- 1 −1 H offset shows a slow-pulsing artifact of ~ 1.5 s in the CW persion profiles. Critical to its performance seems to be short experiment, which is effectively suppressed when using the duration of the composite pulse combined with relatively 90 –240 –90 decoupling sequence (Fig. 3d). high quality of off-resonance performance. x y x 1 3 Journal of Biomolecular NMR (2018) 71:69–77 75 Fig. 3 The 90 –240 –90 decoupling scheme offers high-quality an unassigned Arg sidechain resonance (Rsc) both with significant x y x dispersion curves free of slow-pulsing artifact. a Average RMSD H offset. Data for panels b, c recorded at 850  MHz. Data for panel between CPD- and CW-based dispersion curves over all analyzed d recorded at 950  MHz. Dotted lines are best-fit dispersion curves residues in azurin. Open/closed bars refer to the RMSD without/with obtained using CATIA (Hansen, http://www.bioch em.ucl.ac.uk/hanse compensating for the offset between the curves. b–d Experimental n/catia /). The CPD data were corrected for the systematic offset to the dispersion curves for both CW and 90 –240 –90 based experiments CW data before plotting x y x for amide resonances of A92 (no H offset, panel b), and T52, and The cause of the slight offset between the CPD and CW- water polarization. Experimental tests (Hiller et al. 2005) based dispersion profiles is unclear. Closer inspection shows demonstrate this effect and show that after a 2 s delay, cor - that the magnitude of the offset shows no correlation to responding to the recycle delay to the next proton excita- either the N, H , or H chemical shift and that both reference tion pulse, there is minimal difference between the water N α (no CPMG delay) and dispersion experiment (with CPMG polarization in the CPD and CW case (see Supplemental delay) have slightly altered intensities (~ 2–5%) in the CPD Fig. S1). Here, radiation damping caused by the high Q of experiment compared to the CW experiment. The effect on the cryogenic probe likely aids the recovery of the water the reference experiment, where the decoupling block is car- magnetization in the CPD case. Additionally, the low pH ried out before the recycle delay, signifies that the both types of the sample (5.5) will slow down amide-water exchange of decoupling result in a different steady-state magnetiza - and thus additionally dampen the effect of (residual) water tion, presumably both for water and protein protons. saturation. As for the water magnetization, a disadvantage of using In the original implementation of the in-phase dispersion CPD over CW decoupling is the loss of control over its experiment described by Hansen et al. (2008b), the strength state. Whereas in the CW case the water magnetization is of the decoupling field is matched to the CPMG pulsing spin-locked and returned to + z after the CPMG period, con- rate to avoid the timing mismatch as indicated in Fig. 2c. tinuous alteration between x and y-pulse phase during the In principle, such matching could also be done when using 90 –240 –90 CPD element causes dephasing and loss of CPD decoupling schemes, which should result in decreased x y x 1 3 76 Journal of Biomolecular NMR (2018) 71:69–77 Allard P, Helgstrand M, Hard T (1998) The complete homogeneous scatter in the dispersion curves. While simulations indeed master equation for a heteronuclear two-spin system in the basis show such improvement in performance, an experimental of cartesian product operators. J Magn Reson 134:7–16 test showed a severe increase in scatter, presumably due to Chakrabarti KS et al (2016) High-power 1H composite pulse decou- a point-to-point variation in the steady state of the water and pling provides artifact free exchange-mediated saturation trans- fer (EST) experiments. J Magn Reson 269:65–69 aliphatic proton magnetization. Culik RM et al (2018) Effects of maturation on the conformational As noted in Fig. 1, the slow-pulsing artifact will be par- free-energy landscape of SOD1. Proc Natl Acad Sci 115:E2546 ticularly problematic at high magnetic field strengths. At Delaforge E et al (2018) Deciphering the dynamic interaction profile such high fields, it may be better to use TROSY–CPMG of an intrinsically disordered protein by NMR exchange spec- troscopy. J Am Chem Soc 140:1148–1158 sequences (Loria et al. 1999a), which do not suffer from Delaglio F et  al (1995) NMRPipe: a multidimensional spectral the slow-pulsing artifact, even for non-deuterated mod- processing system based on UNIX pipes. J Biomol NMR erately sized proteins. The relative sensitivity of TROSY 6:277–293 and in-phase CPMG experiments is best assessed experi- Eaton JW, Bateman D, Hauberg S (2008) GNU Octave manual version 3. Network Theory Ltd., Bristol, p 568 mentally as it not only depends on magnetic field strength Eisenmesser EZ et al (2005) Intrinsic dynamics of an enzyme underlies but also on protein size, labeling pattern, and temperature. catalysis. Nature 438:117 Next to the absolute sensitivity, one may also consider that Fawzi NL, Ying J, Torchia DA, Clore GM (2010) Kinetics of amyloid lower N relaxation rates during the CPMG period allow β monomer to oligomer exchange by NMR relaxation. J Am Chem Soc 132:9948–9951 the use of longer CPMG delays, increasing the sensitivity Franco R, Gil-Caballero S, Ayala I, Favier A, Brutscher B (2017) Prob- to slow motions (Loria et al. 1999a), as well as spectral ing conformational exchange dynamics in a short-lived protein quality of TROSY spectra (reduced overlap vs. presence folding intermediate by real-time relaxation–dispersion NMR. J of anti-TROSY lines). Additionally, in case data at lower Am Chem Soc 139:1065–1068 Frauenfelder H, Sligar SG, Wolynes PG (1991) The energy landscapes field strength have been recorded using the in-phase CPMG and motions of proteins. Science 254:1598–1603 experiment it may be necessary to record these at high fields Gopalan AB, Hansen DF, Vallurupalli P (2018) CPMG experiments too. for protein minor conformer structure determination. In: Ghose R In conclusion, we show here that the use of broadband H (ed) Protein NMR: methods and protocols. Springer, New York, pp 223–242 decoupling, in particular using the 90 –240 –90 sequence, x y x 15 Hammes GG (1964) Mechanism of enzyme catalysis. Nature 204:342 is a viable and attractive option for recording in-phase N Hansen F. http://www.bioch em.ucl.ac.uk/hanse n/catia /. Accessed 1 relaxation dispersion data. This option is particularly rel- June 2018 evant when the protein spectrum contains resonances far Hansen F. http://www .bioch em.ucl.ac.uk/hanse n/fuda/. Accessed 1 June 2018 from center. It offers artifact-free dispersion profiles without Hansen DF, Vallurupalli P, Lundström P, Neudecker P, Kay LE (2008a) the need for recording data in multiple sets or the need for Probing chemical shifts of invisible states of proteins with relaxa- eliminating of data points, all without compromising data tion dispersion NMR spectroscopy: how well can we do? J Am quality. Chem Soc 130:2667–2675 Hansen DF, Vallurupalli P, Kay LE (2008b) An improved 15N relaxa- tion dispersion experiment for the measurement of millisecond Acknowledgements We thank prof. Daiwen Yang (National University time-scale dynamics in proteins. J Phys Chem B 112:5898–1904 of Singapore) for sharing the pulse sequence code for the ST-CW- Helgstrand M, Härd T, Allard P (2000) Simulations of NMR pulse CPMG experiment. This work was supported by financial support from sequences during equilibrium and non-equilibrium chemical the Dutch Science Foundation NWO by a VIDI grant (723.013.010) exchange. J Biomol NMR 18:49–63 to HvI and grant 184.032.207 to the uNMR-NL National Roadmap Henzler-Wildman K, Kern D (2007) Dynamic personalities of proteins. Large-Scale Facility of the Netherlands. Nature 450:964–972 Henzler-Wildman KA et al (2007) Intrinsic motions along an enzy- Open Access This article is distributed under the terms of the Crea- matic reaction trajectory. Nature 450:838 tive Commons Attribution 4.0 International License (http://creat iveco Hiller S, Wider G, Etezady-Esfarjani T, Horst R, Wüthrich K (2005) mmons.or g/licenses/b y/4.0/), which permits unrestricted use, distribu- Managing the solvent water polarization to obtain improved NMR tion, and reproduction in any medium, provided you give appropriate spectra of large molecular structures. J Biomol NMR 32:61–70 credit to the original author(s) and the source, provide a link to the Jiang B, Yu B, Zhang X, Liu M, Yang D (2015) A (15)N CPMG relaxa- Creative Commons license, and indicate if changes were made. tion dispersion experiment more resistant to resonance offset and pulse imperfection. J Magn Reson 257:1–7 Karlsson BG, Pascher T, Nordling M, Arvidsson RHA, Lundberg LG (1989) Expression of the blue copper protein azurin from Pseu- domonas aeruginosa in Escherichia coli. FEBS Lett 246:211–217 References Kim TH et  al (2017) The role of dimer asymmetry and protomer dynamics in enzyme catalysis. Science. https ://doi.org/10.1126/ Adman ET (1991) Copper protein structures. Adv Protein Chem scien ce.aag23 55 42:145–197 Kimsey IJ, Petzold K, Sathyamoorthy B, Stein ZW, Al-Hashimi HM Akke M, Palmer AG (1996) Monitoring macromolecular motions (2015) Visualizing transient Watson-Crick like mispairs in DNA on microsecond to millisecond time scales by R1ρ–R1 constant and RNA duplexes. Nature 519:315–320 relaxation time NMR spectroscopy. J Am Chem Soc 118:911–912 1 3 Journal of Biomolecular NMR (2018) 71:69–77 77 Korzhnev DM, Karlsson BG, Orekhov VY, Billeter M (2003) NMR anisotropic chemical shift relaxation mechanisms in the measure- detection of multiple transitions to low-populated states in azurin. ment of spin-spin relaxation rates. Mol Phys 75:699–711 Protein Sci 12:56–65 Pratihar S et al (2016) Kinetics of the antibody recognition site in Korzhnev DM, Religa TL, Banachewicz W, Fersht AR, Kay LE (2010) the third IgG-binding domain of protein G. Angew Chem Int Ed A transient and low-populated protein-folding intermediate at 55:9567–9570 atomic resolution. Science 329:1312–1316 Sauerwein A, Hansen D (2015) Relaxation dispersion NMR spectros- Kovermann M, Rogne P, Wolf-Watz M (2016) Protein dynamics and copy. In: Berliner L (ed) Protein NMR. Biological magnetic reso- function from solution state NMR spectroscopy. Q Rev Biophys nance. Springer, Boston 49:e6 Schneider R et al (2015) Visualizing the molecular recognition tra- Levitt MH (1982) Symmetrical composite pulse sequences for NMR jectory of an intrinsically disordered protein using multinuclear population inversion. II. Compensation of resonance offset. J relaxation dispersion NMR. J Am Chem Soc 137:1220–1229 Magn Reson (1969) 50:95–110 Shaka AJ, Keeler J (1987) Broadband spin decoupling in isotropic- Levitt MH, Freeman R (1981) Composite pulse decoupling. J Magn liquids. Prog Nucl Magn Reson Spectrosc 19:47–129 Reson 43:502–507 Shaka AJ, Keeler J, Frenkiel T, Freeman R (1983) An improved Levitt MH, Freeman R, Frenkiel T (1982a) Broadband heteronuclear sequence for broadband decoupling: WALTZ-16. J Magn Reson decoupling. J Magn Reson (1969) 47:328–330 (1969) 52:335–338 Levitt MH, Freeman R, Frenkiel T (1982b) Supercycles for broadband Shaka AJ, Barker PB, Freeman R (1985) Computer-optimized decou- heteronuclear decoupling. J Magn Reson 50:157–160 pling scheme for wideband applications and low-level operation. Libich DS, Tugarinov V, Clore GM (2015) Intrinsic unfoldase/foldase J Magn Reson (1969) 64:547–552 activity of the chaperonin GroEL directly demonstrated using Shaka AJ, Lee CJ, Pines A (1988) Iterative schemes for bilinear multinuclear relaxation-based NMR. Proc Natl Acad Sci USA operators; application to spin decoupling. J Magn Reson (1969) 112:8817–8823 77:274–293 Lisi GP (2016) & Patrick Loria, J. Using NMR spectroscopy to eluci- Sugase K, Dyson HJ, Wright PE (2007) Mechanism of coupled fold- date the role of molecular motions in enzyme function. Prog Nucl ing and binding of an intrinsically disordered protein. Nature Magn Reson Spectrosc 92–93:1–17 447:1021 Loria JP, Rance M, Palmer AG (1999a) A TROSY CPMG sequence Vallurupalli P, Bouvignies G, Kay LE (2012) Studying “invisible” for characterizing chemical exchange in large proteins. J Biomol excited protein states in slow exchange with a major state confor- NMR 15:151–155 mation. J Am Chem Soc 134:8148–8161 Loria JP, Rance M, Palmer AG (1999b) A relaxation-compensated Vallurupalli P, Sekhar A, Yuwen T, Kay LE (2017) Probing conforma- Carr–Purcell–Meiboom–Gill sequence for characterizing chemical tional dynamics in biomolecules via chemical exchange saturation exchange by NMR spectroscopy. J Am Chem Soc 121:2331–2332 transfer: a primer. J Biomol NMR 67:243–271 Massi F, Peng JW (2018) Characterizing protein dynamics with NMR Wolynes PG (2005) Recent successes of the energy landscape theory of R 1ρ relaxation experiments. In: Ghose R (ed) Protein NMR: protein folding and function. Q Rev Biophys 38:405–410 methods and protocols. Springer, New York, pp 205–221 Xiao T, Fan Js, Zhou H, Lin Q, Yang D (2016) Local unfolding of fatty Neudecker P et al (2012) Structure of an intermediate state in protein acid binding protein to allow ligand entry for binding. Angew folding and aggregation. Science 336:362–366 Chem Int Ed 55:6869–6872 Palmer AG 3rd, Kroenke CD, Loria JP (2001) Nuclear magnetic Xue Y et al (2015) Characterizing RNA excited states using NMR resonance methods for quantifying microsecond-to-millisec- relaxation dispersion. Methods Enzymol 558:39–73 ond motions in biological macromolecules. Methods Enzymol Yip G, Zuiderweg E (2004) A phase cycle scheme that significantly 339:204–238 suppresses offset-dependent artifacts in the R 2-CPMG 15 N Palmer AG (2015) Enzyme dynamics from NMR spectroscopy. Acc relaxation experiment. J Magn Reson 171:25–36 Chem Res 48:457–465 Zhao B, Guffy SL, Williams B, Zhang Q (2017) An excited state under - Palmer AG, Skelton NJ, Chazin WJ, Wright PE, Rance M (1992) Sup- lies gene regulation of a transcriptional riboswitch. Nat Chem pression of the effects of cross-correlation between dipolar and Biol 13:968–974 1 3

Journal

Journal of Biomolecular NMRSpringer Journals

Published: Jun 2, 2018

References

You’re reading a free preview. Subscribe to read the entire article.


DeepDyve is your
personal research library

It’s your single place to instantly
discover and read the research
that matters to you.

Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.

All for just $49/month

Explore the DeepDyve Library

Search

Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly

Organize

Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.

Access

Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.

Your journals are on DeepDyve

Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.

All the latest content is available, no embargo periods.

See the journals in your area

DeepDyve

Freelancer

DeepDyve

Pro

Price

FREE

$49/month
$360/year

Save searches from
Google Scholar,
PubMed

Create lists to
organize your research

Export lists, citations

Read DeepDyve articles

Abstract access only

Unlimited access to over
18 million full-text articles

Print

20 pages / month

PDF Discount

20% off