Impact of two-bond ¹⁵N-¹⁵N scalar couplings on ¹⁵N transverse relaxation measurements for arginine side chains of proteins

Abstract

NMR relaxation of arginine (Arg) ¹⁵N_ε nuclei is useful for studying side-chain dynamics of proteins. In this work, we studied the impact of two geminal ¹⁵N-¹⁵N scalar couplings on measurements of transverse relaxation rates (R₂) for Arg side-chain ¹⁵N_ε nuclei. For 12 Arg side chains of the DNA-binding domain of the Antp protein, we measured the geminal ¹⁵N-¹⁵N couplings (²J_NN) of the ¹⁵N_ε nuclei and found that the magnitudes of the ²J_NN coupling constants were virtually uniform with an average of 1.2 Hz. Our simulations, assuming ideal 180° rotations for all ¹⁵N nuclei, suggested that the two ²J_NN couplings of this magnitude could in principle cause significant modulation in signal intensities during the Carr-Purcell-Meiboom-Gill (CPMG) scheme for Arg ¹⁵N_ε R₂ measurements. However, our experimental data show that the expected modulation via two ²J_NN couplings vanishes during the ¹⁵N CPMG scheme. This quenching of J modulation can be explained by the mechanism described in Dittmer and Bodenhausen (2006) ChemPhysChem 7, 831–836. This effect allows for accurate measurements of R₂ relaxation rates for Arg side-chain ¹⁵N_ε nuclei despite the presence of two ²J_NN couplings. Although the so-called recoupling conditions may cause overestimate of R₂ rates for very mobile Arg side chains, such conditions can readily be avoided through appropriate experimental settings.

Introduction

For a mechanistic understanding of protein function, it is important to gain knowledge about dynamic behaviors of protein side chains. Conformational dynamics are directly relevant to enzymatic activities (Lisi and Loria 2017), molecular recognition by proteins (Baldwin and Kay 2009), and often make important entropic contributions to binding free energy (Wand 2013). Arginine (Arg) side chains play important roles in electrostatic interactions and hydrogen bonding at molecular interfaces. NMR spectroscopy has been used to study dynamic behaviors of Arg side chains of proteins (Berglund et al. 1995; Diehl et al. 2010; Esadze et al. 2016; Gerecht et al. 2017; Nguyen et al. 2017; Nieto et al. 1997; Trbovic et al. 2009; Werbeck et al. 2013; Wilkinson et al. 2000; Wilkinson et al. 2004).

The key methodology for such investigations is the ¹⁵N relaxation measurements for N_εH groups of Arg side chains. Arg side-chain guanidinium moieties are planar, comprising three nitrogen (N_ε, N_η1, and N_η2), one carbon (C_ζ), and five hydrogen (H_ε, H_η11, H_η12, H_η21, and H_η22) atoms on the same plane. Hindered 180° rotations along the N_ε-C_ζ, C_ζ-N_η1, and C_ζ-N_η2 bonds occur, causing inter-conversions between two N_ηH₂ groups and between two hydrogen atoms within the same N_ηH₂ groups. NMR signals from Arg N_ηH₂ groups are usually broad and difficult to analyze, primarily because these inter-conversions tend to occur in the intermediate exchange regime on the ¹H and ¹⁵N chemical shift timescale at physiological temperature (Gerecht et al. 2017; Nieto et al. 1997; Yamazaki et al. 1995). However, ¹H and ¹⁵N resonances of Arg N_εH groups are unaffected by the hindered 180° bond rotations and exhibit relatively sharp NMR signals when hydrogen exchange is slow enough. Thus, N_εH groups are convenient for NMR investigations of Arg side-chain guanidinium moieties.

Typically when ¹⁵N NMR relaxation is analyzed for Arg side-chain N_εH groups of ¹⁵N-labeled proteins, the relaxation theory for an AX spin system has been used. Although this treatment is reasonable, it should be noted that ¹⁵N_ε nuclei interact with ¹⁵N_η1 and ¹⁵N_η2 nuclei through geminal ¹⁵N-¹⁵N scalar couplings (²J_NN). In principle, these two ²J_NN couplings can significantly modulate signal intensities during the Carr-Purcell-Meiboom-Gill (CPMG) experiments for measuring¹⁵N transverse relaxation rates (R₂), because chemical shifts of ¹⁵N_ε (~85ppm) and ¹⁵N_η1/¹⁵N_η12 (~72ppm) nuclei are close enough for typical ¹⁵N refocusing pulses to affect simultaneously.

In this work, using a small (7 kDa) protein for which Arg ¹⁵N_ε relaxation is slow enough to precisely analyze the impact of ²J_NN modulation, we have investigated the extent to which the two ²J_NN couplings influence the CPMG-based measurements of ¹⁵N R₂ relaxation rates for Arg side chains. For the 12 Arg side chains of this protein, we measured ²J_NN couplings using a new pulse sequence that is not affected by rapid transverse relaxation of ¹⁵N_η1/¹⁵N_η2 nuclei. Then, based on the measured ²J_NN couplings, we examined the impact of two ²J_NN couplings on the ¹⁵N transverse relaxation measurements using the CPMG scheme. Although a considerable impact was anticipated from the ²J_NN coupling data, our experimental data show that the two ²J_NN couplings give virtually no impacts on ¹⁵N CPMG experiments for Arg side-chain ¹⁵N_ε nuclei. The quenching of ²J_NN modulation during the ¹⁵N CPMG scheme can be explained by the mechanism proposed by Bodenhausen and co-workers (Aeby and Bodenhausen 2008; Dittmer and Bodenhausen 2006; Gopalakrishnan et al. 2007).

Materials and Methods

Sample preparation

The DNA-binding domain (60 residues; homeodomain) of the fruit fly Antp protein was expressed in Escherichia coli cultured in a ¹⁵N medium and was purified as previously described (Zandarashvili et al. 2015). A 370 µl solution of 1.6 mM ¹⁵N-labeled Antp homeodomain in a buffer of 20 mM succinate-d₄•KCl (pH 5.8), 100 mM KCl, and 0.4 mM NaF (as a preservative) was sealed into a coaxial NMR tube with a stem insert containing 100 µl D₂O for NMR lock. D₂O was separately sealed to avoid any adverse effects arising from partial deuteration of Arg guanidinium moieties (e.g., ¹⁵N chemical shift changes due to ¹H/²H exchange during the CPMG scheme). The coaxial NMR tube (outer diameter 5 mm, the stem diameter 2 mm) was purchased from Norell, Inc.

NMR experiments

The NMR experiments were conducted at 25°C using a Bruker Avance III spectrometer equipped with a TCI cryogenic probe operated at the ¹H frequency of 800 MHz. Resonances of the Arg side chains of the Antp homeodomain under the same conditions were assigned in our previous work (Nguyen et al. 2017). The HNN experiment (Löhr and Rüterjans 1998) was conducted using 1-ms rSNOB pulses (Kupče et al. 1995) for ¹⁵N 180° pulses at ¹H-¹⁵N INEPT schemes to selectively observe signals from Arg side chains. The spin-echo ²J_NN modulation constant-time HISQC experiment was conducted as described in the figure caption. The ¹⁵N CPMG experiment for Arg side chains was conducted as previously described (Esadze et al. 2016). During the CPMG scheme, ¹H continuous wave (CW) was applied in the manner of the CW-CPMG method (Hansen et al. 2008). The NMR spectra were processed by the NMR-Pipe program (Delaglio et al. 1995) and analyzed by the NMR-View program (Johnson and Blevins 1994). Numerical calculations and nonlinear least-squares fittings were conducted with MATLAB software (MathWorks, Inc.).

Results and Discussion

We used ¹⁵N-labeled Antp homeodomain to investigate the ²J_NN couplings and their impact on Arg ¹⁵N R₂ measurements. Relatively slow ¹⁵N_ε transverse relaxation of Arg side chains in this 60-residue protein facilitated our quantitative analysis of ²J_NN couplings and their impact on transverse relaxation measurements for Arg side chains.

Observation of ²J_NN couplings between Arg ¹⁵N_ε and ¹⁵N_η1/¹⁵N_η2 nuclei

To observe ²J_NN couplings, we recorded the HNN spectrum (Löhr and Rüterjans 1998) for the Antp homeodomain. In this experiment, coherence transfers between ¹⁵N_ε and ¹⁵N_η nuclei via ²J_NN couplings give rise to [¹H_ε,¹⁵N_η1] and [¹H_ε,¹⁵N_η2] cross peaks. The HNN spectrum clearly showed these cross peaks in addition to [¹H_ε, ¹⁵N_ε] cross peaks (Figure 1). The ¹⁵N_η1/¹⁵N_η2 resonances in this spectrum represent direct evidence of sizable geminal ¹⁵N-¹⁵N couplings in Arg side chains. In principle, the ²J_NN couplings can be measured from the ratio of signal intensities of these cross peaks (Löhr and Rüterjans 1998). For quantitative analysis, however, many ¹⁵N_η1/¹⁵N_η2 resonances in the HNN spectrum were too broad, presumably due to hindered 180° bond rotations causing inter-conversions between ¹⁵N_η1 and ¹⁵N_η2 nuclei in the intermediate exchange regime. 10 out of 12 Arg side chains exhibited ¹⁵N_η1/¹⁵N_η2 resonances that are broadened and partially degenerated presumably due to the intermediate exchange. The broad ¹⁵N_η1/¹⁵N_η2 resonances make it difficult to determine the ²J_NN couplings, although this spectrum provides useful information of ¹⁵N_η1/¹⁵N_η2 resonances for individual Arg N_εH groups.

Figure 1.

The HNN spectrum (Löhr and Rüterjans 1998) recorded for the Arg side chains of ¹⁵N-labeled Antp homeodomain at 25°C. Two 1-ms ¹⁵N 180° rSNOB pulses (Kupče et al. 1995) were used for the ¹H-¹⁵N INEPT scheme in the HNN experiment to selectively observe signals from Arg side chains. The length of each ¹⁵N-¹⁵N coherence transfer scheme was 126 ms in this experiment. Positive and negative contours are shown in black and red, respectively. Cross peaks with ¹H_ε and ¹⁵N_η1/¹⁵N_η2 resonances arises from coherence transfers via geminal ²J_NN couplings between Arg ¹⁵N_ε and ¹⁵N_η1/¹⁵N_η2 nuclei.

Magnitudes of ²J_NN couplings in Arg side chains

To resolve this problem in measuring the ²J_NN couplings between Arg ¹⁵N_ε and ¹⁵N_η1/¹⁵N_η2 nuclei, we used a new pulse sequence, which is shown in Figure 2A. This is a variation of spin-echo J-modulation constant-time ¹H-¹⁵N HISQC experiments (Anderson et al. 2013; Zandarashvili et al. 2011). J-modulation for in-phase single-quantum ¹⁵N transverse term N_x for Arg ¹⁵N_ε nuclei is used in this experiment. Two sub-experiments are conducted in an interleaved manner. In one sub-experiment, ¹⁵N_η-selective IBURP-2 inversion pulses and ¹⁵N_ε-selective REBURP refocusing pulses (Geen and Freeman 1991) are applied so that the net evolution time for ²J_NN couplings throughout the constant-time period (T_c) becomes zero (i.e., ²J_NN modulation OFF). In the other sub-experiment, these shaped ¹⁵N pulses are applied so that the net evolution time for ²J_NN couplings becomes equal to the length T_c (i.e., ²J_NN modulation ON). The ratio of the signal intensities in the sub-spectra recorded through these sub-experiments is given by:

$$I_{\mathit{\text{on}}}/I_{\mathit{\text{off}}}={\text{cos}}^2\pi J_{\mathit{\text{NN}}}{T}_c,$$ (1)

where I_on and I_off represent the signal intensities in the sub-spectra with ²J_NN modulation on and off, respectively. This equation assumes the presence of two identical ²J_NN couplings between ¹⁵N_ε and ¹⁵N_η1 and between ¹⁵N_ε and ¹⁵N_η2 nuclei. This assumption is reasonable, considering the covalent geometry of Arg guanidinium. Unlike the HNN experiment, this experiment is not directly impacted by the hindered 180° rotations of N_ε-C_ζ bonds because transverse magnetizations of ¹⁵N_η1 and ¹⁵N_η2 nuclei are not utilized.

Figure 2.

Measuring the magnitudes of ²J_NN coupling constants (i.e., |²J_NN|) without being influenced by broadening of ¹⁵N_η1/¹⁵N_η2 resonances. (A) Pulse sequence for the Arg-selective constant-time spin-echo ²J_NN modulation ¹H-¹⁵N HISQC. Thin and bold bars in black represent hard rectangular 90° and 180° pulses, respectively. Unless indicated otherwise, pulse phases are along x. Short-bold bars represent water-selective soft-rectangular ¹H 90° pulses (1.2 ms) and a half-bell shape represents a water-selective half-Gaussian 90° pulse (2.1 ms). Lengths of the ¹⁵N shaped pulses were as follows:¹⁵N rSNOB 180° pulse, 1.03 ms; ¹⁵N_ε-selective REBURP, 6 ms; and ¹⁵N_η1/η2-selective IBURP2, 6 ms.. WALTZ RF strengths: ¹H, 4.5 kHz; and ¹⁵N, 1kHz. Delays: τ_a = 2.3 ms; δ = 5.4 ms; and T_c = 50–252 ms. Phase cycles: ϕ₁ = [x, −x], ϕ₂ = [2x, 2y, 2(−x), 2(−y)], ϕ₃ = x, and receiver = [x, 2(−x), x]. Quadrature detection in the t₁ domain was achieved using States-TPPI for ϕ₁. Pulsed field gradients were optimized to minimize the water signal. The sub-experiments with ²J_NN modulation on and off were conducted in an interleaved manner. (B) The constant-time spin-echo ²J_NN modulation ¹H-¹⁵N HISQC data for the Antp homeodomain. (C) Intensity ratio I_on/I_off recorded for R24 as a function of T_C. The best-fit curve obtained with Eq. 1 is also shown. (D) |²J_NN | determined for 12 Arg side chains of the Antp homeodomain. Uncertainties in the determined |²J_NN | values were estimated to be less than 5%.

We conducted the spin-echo ²J_NN modulation CT-HISQC experiment for the ¹⁵N-labeled Antp DNA-binding domain at the ¹H frequency of 800 MHz. Figure 2B shows the spectra recorded in an interleaved manner. As expected, the ¹H_ε/¹⁵N_ε cross peaks in the sub-spectrum with ²J_NN modulation on were found to be significantly weaker than in the other sub-spectrum. The intensity ratio I_on / I_off changed as a function of T_C (Figure 1C). Using Eq. 1, we calculated |²J_NN| from the peak intensity ratio data. This experiment allowed us to determine the |²J_NN| values for all 12 Arg side chains (Figure 2D). The determined |²J_NN| values ranged from 1.1 Hz to 1.3 Hz with an average of 1.2 Hz. These are slightly larger than what Löhr and Rüterjans measured for Arg side chains in flavodoxin (0.9–1.1 Hz) using the HNN spectrum (Löhr and Rüterjans 1998), presumably because intensities of the signals arising from the ²J_NN couplings were underestimated due to the broadening of ¹⁵N_η1/¹⁵N_η2 resonances in the HNN spectrum. The magnitudes of the ²J_NN couplings in the 12 Arg side chains of the Antp homeodomain were virtually uniform, although some Arg side chains form hydrogen bonds with other side chains.

Impacts of two ²J_NN couplings on Arg ¹⁵N_ε CPMG

Observing sizable ²J_NN couplings between ¹⁵N_ε and ¹⁵N_η1/¹⁵N_η2 nuclei motivated us to investigate the impact of these couplings on transverse relaxation measurements for Arg ¹⁵N_ε nuclei. Usually, transverse relaxation rates are determined through mono-exponential fitting using:

$$I(T_{\mathit{\text{CPMG}}})=A\text{exp}(-R_{\mathit{2}}{T}_{\mathit{\text{CPMG}}})$$ (2)

where A corresponds to the initial intensity (i.e., I(0) = A) and T_CPMG represents the length of the CPMG scheme. If each ¹⁵N 180° pulse in the CPMG scheme cause an exact 180° rotation along the x axis for ¹⁵N_ε (~85ppm) and ¹⁵N_η1/¹⁵N_η2 (~72ppm) simultaneously, the two ²J_NN couplings to ¹⁵N_η1 and ¹⁵N_η2 nuclei will be active and modulate the magnitude of N^ε_x term by a factor of cos²πJ_NNT_CPMG through the CPMG scheme. Although the difference between ¹⁵N_ε and ¹⁵N_η1/¹⁵N_η2 chemical shifts may be comparable to the CPMG frequency, it is unlikely that Hartmann-Hahn transfer occurs from ¹⁵N_ε to ¹⁵N_η1/¹⁵N_η2 nuclei during the CPMG scheme. The Hartmann-Hahn transfer through the strong coupling effect requires the condition that the difference between the effective fields for the coupled nuclei is comparable to or smaller than πJ (Müller and Ernst 1979). In the current case of the CPMG experiment on the Arg ¹⁵Nε nuclei, however, this condition is not satisfied because the effective field Ω_eff for ¹⁵N_η is significantly greater than that for ¹⁵N_ε nuclei and |Ω_eff(¹⁵N_η) − Ω_eff(¹⁵N_ε)| >> |πJ_NN|. Thus, the assumption of the weak coupling limit is valid in the current case, and the signal intensity can be approximated by:

$$I(T_{\mathit{\text{CPMG}}})=A\text{exp}(-R_{\mathit{2}}{T}_{\mathit{\text{CPMG}}}){\text{cos}}^2(\pi J_{\mathit{\text{NN}}}{T}_{\mathit{\text{CPMG}}})$$ (3)

Strictly speaking, the signal intensity should also depend on the relaxation rates of the anti-phase terms 2N^ε₊N^η_z and 4N^ε₊N^η_zN^η_z. The relaxation rates of the 2N^ε₊N^η_z and 4N^ε₊N^η_zN^η_z terms should be slightly faster than that of the N^ε₊ term. However, our simulations based on the density operator evolution show that Eq. 3 is valid when the relaxation rates of these terms are comparable to that of the in-phase term N^ε₊ (see the Supporting Information).

Although |²J_NN| = 1.2 Hz may seem to be small, the simultaneous presence of two ²J_NN couplings of this magnitude can in principle cause considerable reduction in signal intensity of Arg ¹⁵N_ε nuclei during the CPMG scheme of typical length. To illustrate this potential effect, Figure 3 shows expected intensity modulations calculated using Eq. 2 and Eq. 3 for the cases with R₂ = 2, 7, and 24 s⁻¹. The impact of the ²J_NN couplings on CPMG data depends on the R₂ relaxation rate. When the R₂ relaxation rate is as fast as 24 s⁻¹, the overall intensity profile calculated with Eq. 3 is virtually indistinguishable from that calculated with Eq. 2. However, when the R₂ relaxation rate is slower, Eq. 3 and Eq. 2 give significantly different intensity profiles, as seen in Figure 3. When the R₂ relaxation rate is as slow as 2 s⁻¹, the intensity profile calculated with Eq. 3 appears more obviously non-exponential and slightly concave downward (see the red solid curve in Figure 3), as opposed to concave upward for that calculated with Eq. 2. For Arg side chains that exhibit such slow ¹⁵N_ε transverse relaxation (i.e., highly mobile Arg side chains), fitting with Eq. 3 should yield better agreement with experimental data, if ²J_NN couplings actually modulate signal intensities throughout the CPMG scheme.

Figure 3.

Intensity modulation calculated with Eqs. 2 and 3. Blue curves were obtained using Eq. 2 and the indicated R₂ rates. Red curves were obtained using Eq. 3, the indicated R₂ rates, and ²J_NN = 1.2 Hz. The gray curve shows modulation by cos²(πJ_NNT_CPMG).

Despite this expectation, our experimental CPMG data for Arg ¹⁵N_ε nuclei in the Antp homeodomain showed obviously better fitting results with Eq. 2 than those with Eq. 3 even for the side chains that exhibit very slow ¹⁵N transverse relaxation. Figure 4 shows the ¹⁵N CPMG relaxation data for the R10, R28 and R29 ¹⁵N_ε nuclei. These side chains are highly mobile and exhibit very slow ¹⁵N transverse relaxation (Nguyen et al. 2017). Solid curves in this figure show the best-fit curves obtained through 1) two-parameter fitting to Eq. 2 (blue) and 2) two-parameter fitting to Eq. 3 together with experimentally obtained ²J_NN couplings (red). As displayed in Figure 4, Eq. 3 with experimental ²J_NN data gave only poor and biased fitting results, compared to Eq. 2. This bias is obvious in the plots of fitting residuals. Although the best-fit curves obtained with Eq. 3 for these Arg side chains are concave downward, the experimental data do not appear to be so. The statistics of these fitting calculations are shown in the insets within Figure 4. The minimized sum of squared differences (SS_min) and the coefficient of determination (r²) clearly indicate that Eq. 2 is a better model than Eq. 3 together with the experimental ²J_NN constant. We also did three-parameter fitting using Eq. 3, optimizing ²J_NN within a constraint of |²J_NN | < 1.5 Hz. These calculations gave |²J_NN| < 0.1 Hz and the resultant R₂ rates were virtually identical to those obtained with Eq. 2. These results clearly show that modulation via two ²J_NN couplings somehow vanished during the ¹⁵N CPMG scheme for Arg side chains.

Figure 4.

¹⁵N_ε transverse relaxation data from the CPMG experiment for R10, R28, and R29 side chains of ¹⁵N-labeled Antp homeodomain at 25°C. Solid curves represent the best-fit curves obtained through nonlinear least-squares fitting using Eq. 2 (blue) or Eq.3 with ²J_NN = 1.2 Hz (red). The minimized sum of squared differences (SS_min), coefficient of determination (r²), and fitting residuals are also shown for each data. The RF strength (γB₁/2π) of rectangular ¹⁵N 180° pulses used in the CPMG scheme was 3.2 kHz; the CPMG field strength (ν_CPMG) was 417 Hz; and the ¹⁵N carrier position was 86ppm.

Quenching of homonuclear J modulation in CPMG through the SITCOM effect

The apparent absence of ²J_NN modulation during the ¹⁵N CPMG for Arg side chains can be explained by the mechanism that Bodenhausen’s group described (Aeby and Bodenhausen 2008; Dittmer and Bodenhausen 2006; Gopalakrishnan et al. 2007). They dubbed this mechanism SITCOM, which stands for “stabilization by interconversion with a triad coherences under multiple refocusing”. In CPMG schemes using 180° pulses of moderate radio-frequency (RF) strength, the SITCOM effect arises from conversion of homonuclear anti-phase coherence 2I_yS_z into multiple-quantum terms through many non-ideal 180° pulses for off-resonance spin S. Because the effective field is tilted for the off-resonance spin S, the refocusing pulses generate multiple-quantum terms 2I_yS_x and 2I_yS_y, which effectively hinders the anti-phase term 2I_yS_z from building up and locks the in-phase coherence I_x of the on-resonance spin I. Bodenhausen’s group extensively studied this effect for homonuclear ¹⁵N-¹⁵N (Dittmer and Bodenhausen 2006), ¹³C-¹³C (Aeby and Bodenhausen 2008; Gopalakrishnan et al. 2007), and ¹H-¹H (Baishya et al. 2009; Segawa et al. 2008) couplings. Quenching of J-modulation through the SITCOM mechanism is virtually complete for a wide variety of conditions, but disappears when the following conditions (“recoupling conditions”) are satisfied (Aeby and Bodenhausen 2008):

$${\mathrm{\Omega }}_S/(2\pi )=4n{\nu }_{\mathit{1}}{\nu }_{\mathit{\text{CPMG}}}/(2{\nu }_{\mathit{1}}+{\nu }_{\mathit{\text{CPMG}}})$$ (4)

where n is an integer; Ω_S/(2π), the offset in Hz for off-resonance spin S; ν₁, the RF strength in Hz for refocusing pulses; ν_CPMG, the CPMG field frequency in Hz. When ν₁ >> ν_CPMG, Eq. 4 becomes Ω_S/(2π) ≈ 2nν_CPMG. When the difference between Ω_S/(2π) and the nearest recoupling condition is less than 10% of ν_CPMG, significant amplitude modulate occurs and the apparent R₂ rate becomes larger than the actual R₂ relaxation rate. Not surprisingly, this adverse effect is weaker for a smaller magnitude of the homonuclear J coupling. When Ω_S/(2π) is more significantly different from the recoupling conditions, virtually complete quenching of J-modulation occurs, allowing accurate measurements of the R₂ rates.

Recoupling conditions can readily be avoided through appropriate settings of the CPMG frequency ν_CPMG, the RF strength ν₁, and the ¹⁵N carrier position. In our current case with ν_CPMG = 417 Hz and ν₁= 3200 Hz, the recoupling conditions are Ω_S/(2π) = 783n Hz. Since the ¹⁵N carrier position was set to 86ppm, ¹⁵N_η1/¹⁵N_η2 resonances are 70–75ppm, no Arg side chain satisfies the recoupling conditions at the ¹H frequency of 800 MHz. Therefore, the homonuclear ²J_NN couplings were effectively quenched through the SITCOM mechanism during the ¹⁵N CPMG experiment for Arg side chains.

Concluding Remarks

Although the two geminal ¹⁵N-¹⁵N couplings can in principle affect Arg ¹⁵N_ε transverse relaxation measurements, our current study shows that the impact of the ²J_NN couplings on CPMG-based transverse relaxation measurements for Arg ¹⁵N_ε nuclei is negligible. This is due to the SITCOM effect (Dittmer and Bodenhausen 2006) and relatively small magnitudes of the ²J_NN couplings. Use of selective refocusing pulses (e.g., REBURP) in the CPMG scheme was previously proposed to avoid modulation through homonuclear scalar couplings (Ishima et al. 2004; Yuwen and Skrynnikov 2014). However, the use of long shaped pulses during the CPMG scheme allows only small ν_CPMG frequencies, which may not completely cancel exchange contribution to measured R₂ relaxation rates. Our current study shows that such use of ¹⁵N_ε-selective shaped pulses during the CPMG scheme is unnecessary because the CPMG scheme with typical rectangular ¹⁵N 180° pulses effectively quench ²J_NN couplings in Arg side chains through the SITCOM mechanism. If the recoupling conditions (Eq. 4) are accidentally met, ¹⁵N_ε R₂ rates may be overestimated for very mobile Arg side chains. However, such conditions can readily be avoided through appropriate settings of the ¹⁵N carrier position, the RF strength, and the CPMG frequency ν_CPMG.