Improved accuracy of 15N-1H scalar and residual dipolar couplings from gradient-enhanced IPAP-HSQC experiments on protonated proteins

Abstract

The presence of dipole-dipole cross-correlated relaxation as well as unresolved E.COSY effects adversely impacts the accuracy of (1)J(NH) splittings measured from gradient-enhanced IPAP-HSQC spectra. For isotropic samples, the size of the systematic errors caused by these effects depends on the values of (2)J(NHalpha), (3)J(NHbeta) and (3)J(HNHalpha). Insertion of band-selective (1)H decoupling pulses in the IPAP-HSQC experiment eliminates these systematic errors and for the protein GB3 yields (1)J(NH) splittings that agree to within a root-mean-square difference of 0.04 Hz with values measured for perdeuterated GB3. Accuracy of the method is also highlighted by a good fit to the GB3 structure of the (1)H-(15)N RDCs extracted from the minute differences in (1)J(NH) splitting measured at 500 and 750 MHz (1)H frequencies, resulting from magnetic susceptibility anisotropy. A nearly complete set of (2)J(NHalpha) couplings was measured in GB3 in order to evaluate whether the impact of cross-correlated relaxation is dominated by the (15)N-(1)H(alpha) or (15)N-(1)H(beta) dipolar interaction. As expected, we find that (2)J(NHalpha) < or = 2 Hz, with values in the alpha-helix (0.86 +/- 0.52 Hz) slightly larger than in beta-sheet (0.66 +/- 0.26 Hz). Results indicate that under isotropic conditions, N-H(N)/N-H(beta) cross-correlated relaxation often dominates. Unresolved E.COSY effects under isotropic conditions involve (3)J(HNHalpha) and J(NHalpha), but when weakly aligned any aliphatic proton proximate to both N and H(N) can contribute.

Figure 1

The impact of a third spin, H^R, on the measurement of one-bond ¹H-¹⁵N J splittings. (A) Interference between ¹H^R-¹⁵N and ¹H-¹⁵N dipolar relaxation contributions cause differential line widths and amplitudes for the components of the ¹⁵N multiplet. For visual purposes, a larger than realistic value is used for J_N-HR (20 Hz). The lower multiplet is simulated in the absence of relaxation interference, with a line width of 15 Hz. The top spectrum corresponds to the case where J_N-HR and J_N-H have the same sign, and P₂(cosθ) < 0, with exaggerated relaxation interference and outer/inner line widths of 17/33 Hz line width. (B) Effect of the magnitude of the error in ¹J_NH splitting measurement ΔJ_app, calculated for τ_c = 10 ns, r_NH = 1.04 Å, r_N-HR = 2.5 Å, ¹⁵N CSA = 170 ppm, 600 MHz ¹H frequency, as a function of P₂(cosθ), where θ is the H-N-H^R angle, in the absence of internal dynamics. Results are shown for three values of J_N-HR: 1.5 Hz (solid line), 3 Hz (dotted) and 4.5 Hz (dashed). The two extreme relaxation rates, corresponding to P₂(cosθ) = 1, are 13.9 ± 1.4 s⁻¹. The time domain data were apodized with an exponential function to yield a final linewidth of ~12 Hz prior to peak picking. Splittings were determined from the distance between the highest points of the two ¹⁵N-{¹H^N} doublet components. (C) Schematic diagram for the E.COSY character of ¹H-¹⁵N HSQC correlations obtained with the gradient-enhanced ¹⁵N → ¹H transfer scheme, with the passive spin H^R not being attached to ¹⁵N. The appearance of the E.COSY effect is also impacted by the ¹⁵N line width, shown larger for the upfield ¹⁵N-{¹H} component than for the downfield one. Note that in practice the J_HNHR and J_N-HR couplings are often not resolved, and the effect simply manifests itself as a skewed peak shape (see arrows in Figure 4).

Figure 2

Measurement of ²J_NHα values in the K4AK19EV42E mutant of GB3. (A) Superimposed regions of six ¹⁵N-¹³C^α cross-sections taken through the 3D HNCA[HA] E.COSY spectrum. The spectrum has been recorded at high resolution in the ¹⁵N dimension, such as to enhance the precision at which the ²J_NHα (and ³J_NHα) can be measured. The pulse sequence used for generating this spectrum is available as Supplementary Material. The intense correlations correspond to intraresidue connectivities; the weaker correlations (e.g., E24) correspond to sequential connectivities to ¹³C^α of the preceding residue. (B) ²J_HNCα values as a function of residue number. Error in the measurement is estimated at ±0.2 Hz.

Figure 3

Pulse scheme of the gradient-enhanced 2D BSD-IPAP HSQC experiment. The pulses in the box are only applied for generating the antiphase (AP) spectrum and are omitted for generating the in-phase (IP) spectrum. Narrow and wide pulses correspond to 90° and 180° flip angles respectively. The ¹H 90° water flip back pulse is sine-bell shaped, has a duration of 1.5 ms, and is only applied for the IP experiment. Solid shaped ¹H pulses are of the IBURP2 type (Geen and Freeman 1991) and serve to decouple H^α,β from ¹⁵N. They are centered at 2.4 ppm, with a duration adjusted to invert ±2.8 ppm. A 600 μs 180° hyperbolic secant shaped pulse (Silver et al. 1984), centered at 116 ppm, is used to decouple C^α and C′. Rance-Kay t₁ quadrature detection is used by alternating the phase of the ¹⁵N 90° pulse after G₆ between x and −x, in concert with alternating the polarity of G₄ and G₅ (Kay et al. 1992). Pulsed field gradients G_1,2,4,5,8 are sine-bell shaped 6.6 (G₁), 9.0 (G₂), 28.2 (G₄), −28.2 (G₅) and 28.2 (G₈) G/cm. G₃, G₆ and G₇ are rectangular with strengths of 16.2 (G₃), 0.6 (G₆) and −0.6 (G₇) G/cm. Gradient durations: G_{1,2,3,4,5,6,7,8} = 1.9, 2.65, 1.7, 1.0, 1.0, t₁/4, t₁/4, 0.203 ms. Delay durations: δ = 2.65 ms; ε = 0.55 ms. A more balanced way of recording IPAP spectra can be achieved by reducing the relaxation difference between IP and AP, executing both IP and AP spectra with the boxed region of Fig.3 in place, but inserting ¹H 180° pulses at the center of each δ delay (and removal of the 180 ¹H pulse applied between the two δ delays, see Supplementary Figure S2 for details).

Figure 4

Small regions extracted from the IPAP-HSQC spectra, displaying the downfield doublet components for a small spectral region of (A,B) isotropic and (C,D) Pf1-aligned GB3. (A,C) Recorded in the absence and (B,D) in the presence of band-selective ¹H decoupling during t₁. The arrows indicate the skewing of the cross peak shape resulting from unresolved E.COSY patterns (see main text and Fig.1C).

Figure 5

Comparison of backbone ¹J_NH splittings measured for the U{¹H,¹³C,¹⁵N} and U{²H,¹³C,¹⁵N} K4AK19EV42E mutant of GB3. ¹J_NH splittings measured for the protonated and deuterated samples using the gradient-enhanced IPAP-HSQC experiment (A) in the absence and (B) in the presence of band-selective decoupling. The measurements were performed at 750 MHz ¹H frequency.

Figure 6

Correlation between experimental (x axis) and predicted magnetic field dependence of the ¹J_NH splittings in GB3. The predicted values are obtained from an SVD fit of differences between ¹J_NH splittings measured at 750 and 500 MHz ¹H frequencies to the N-H vector orientations of GB3 (Yao et al. 2008b), with the dynamic frequency shift treated as a variable which is assumed to be uniform across all amides. The fitted alignment tensor amplitude (D_a) is −0.121 Hz, the rhombicity is 0.576 and the three Euler angles are 18°, 314°, 117° (z-y-z convention) relative to the frame of the PDB coordinates (entry 2OED). The fitted dynamic frequency shift equals 0.106 Hz, in good agreement with an expected value of 0.102 Hz, calculated for τ_c = 3.3 ns (Hall and Fushman 2003)when neglecting internal dynamics and using the constants and approximations of Tjandra et al. (1996). The root mean square deviation (RMSD) between the experimental and best-fitted ¹J_NH⁷⁵⁰ - ¹J_NH⁵⁰⁰ values is 0.037 Hz.