High-Efficiency Low-Power ¹³C-¹⁵N Cross Polarization in MAS NMR

Abstract

Biomolecular solid-state magic angle spinning (MAS) NMR spectroscopy frequently relies on selective ¹³C-¹⁵N magnetization transfers, for various kinds of correlation experiments. Introduced in 1998, spectrally induced filtering in combination with cross polarization (SPECIFIC-CP) is a selective heteronuclear magnetization transfer experiment widely used for biological applications. At MAS frequencies below 20 kHz, commonly used for ¹³C-detected MAS NMR experiments, SPECIFIC-CP transfer between amide ¹⁵N and ¹³C^α atoms (NCA) is typically performed with radiofrequency (rf) fields set higher than the MAS frequency for both ¹³C and ¹⁵N channels, and high-power ¹H decoupling rf field is simultaneously applied. Here, we experimentally explore a broad range of NCA zero-quantum (ZQ) SPECIFIC-CP matching conditions at the MAS frequency of 14 kHz and compare the best high- and low-power matching conditions with respect to selectivity, robustness, and sensitivity at lower ¹H decoupling rf fields. We show that low-power NCA SPECIFIC-CP matching condition gives rise to 20% sensitivity enhancement compared to high-power conditions, in 2D NCA spectra of microcrystalline assemblies of HIV-1 CA_CTD-SP1 protein with inositol hexakis-phosphate (IP6).

Graphical Abstract

1. Introduction

In solid-state NMR, cross polarization (CP) is a fundamental step in almost every experiment, used for sensitivity enhancement and correlation spectroscopy [1, 2]. Spin-lock radiofrequency (rf) fields of matching magnitudes (the Hartmann-Hahn condition) are applied on two nuclei simultaneously, to induce a magnetization transfer through dipolar coupling. Under magic-angle spinning (MAS), the Hartmann-Hahn matching condition is:

$$v_1^I\pm v_1^S=n\cdot v_r;\left|n\right|=1,2,3,\dots$$

where ν₁ is the applied rf field magnitude of spins I or S in Hz, ν_r is the MAS frequency in Hz, and n is an integer greater or smaller than zero [3]. The matching conditions for which the difference or the sum of the rf fields equals the MAS frequency times a non-zero integer are termed zero-quantum (ZQ) and double-quantum (DQ) conditions, respectively.

In biomolecular MAS NMR spectroscopy, a double cross polarization (DCP) experiment [4] is often performed by introducing the second magnetization transfer step between ¹³C and ¹⁵N as adiabatic SPECIFIC-CP for selectivity [5–7]. For example, in peptides and proteins the commonly used SPECIFIC-CP steps would entail transfer of magnetization between amide ¹⁵N and ¹³C^α (NCA) or ¹³C’ (NCO) atoms. At MAS frequencies below 20 kHz, a ZQ (n=±1) condition with rf fields of about 5/2 and 3/2 times the MAS frequency is commonly exploited for ¹⁵N-¹³C transfers in NCA/NCO experiments [8]. Since during the SPECIFIC-CP step strong ¹H decoupling field has to be applied (exceeding the magnitude of the rf fields on the heteronuclei, to avoid any Hartmann-Hahn matching), MAS probes with diameters of 3.2 mm and larger are often at their rf limits on every channel, which is undesirable as this could cause probe arcing and/or shorten the sample lifespan.

Recently, an efficient NCA SPECIFIC-CP matching condition was discovered, utilizing relatively weak rf fields, ca. 1/2 and 3/2 of the MAS frequency for ¹⁵N and ¹³C, respectively, at the MAS frequency of 14 kHz [9]. Similar to the higher-power condition, it is a ZQ (n=±1) matching condition, but with roughly 80% weaker ¹⁵N rf field. Of note, a similar condition was also used in a handful of other studies. For example, Gelenter and Hong utilized 10/12.5 and 30/32.5 kHz rf fields on ¹³C and ¹⁵N, respectively, at the MAS frequency of 20 kHz for NCA SPECIFIC-CP in a tripeptide N-formylmethionine-leucyl-phenylalanine (f-MLF) and microcrystalline β1 immunoglobulin binding domain of protein G (GB1) [10]. Böckmann et al. used 15 and 5 kHz on ¹³C and ¹⁵N, respectively, at the MAS frequency of 10 kHz for NCA SPECIFIC-CP in microcrystalline histidine-containing phosphocarrier protein (Crh) dimer of Bacillus subtilis [11]. Donovan et al. applied 8 and 28 kHz rf fields on ¹³C and ¹⁵N, respectively, at the MAS frequency of 20 kHz for both NCA and NCO SPECIFIC-CP in amyloid β_1-42 fibrils [12]. With these notable exceptions, most of the biomolecular MAS NMR experiments reported in the literature have employed SPECIFIC-CP conditions for ¹³C-¹⁵N magnetization transfer where both rf fields are higher than the MAS frequency.

Herein, we explore all experimentally feasible NCA ZQ (n=±1) SPECIFIC-CP matching conditions at the MAS frequency of 14 kHz and magnetic field of 14.1 T. We investigated the performance of the two best matching conditions, at rf fields of 35 and 21 kHz as well as 5 and 19 kHz, on the ¹⁵N and ¹³C channels, respectively. We found that a lower-power NCA SPECIFIC-CP is more sensitive and selective, likely due to improved robustness to B₁ inhomogeneity, as well as lowering the minimally required ¹H decoupling field strength during the magnetization transfer. We anticipate that low-power SPECIFIC-CP will be broadly applied in many biomolecular MAS NMR experiments, in particular those requiring mAs rotors of diameters of 3.2 mm and wider, where probe rf capabilities are more limited.

2. Material and Methods

All experiments were carried out on a 14.1 T Bruker Avance III HD NMR spectrometer equipped with a 1.3 mm HCN probe. The ¹H, ¹³C, and ¹⁵N Larmor frequencies are 599.8, 150.8, and 60.8 MHz, respectively. The MAS frequency was set to 14 kHz, controlled by a Bruker MAS III unit to within ±5 Hz. The probe temperature was set to 268 K, controlled by a Bruker BCU II unit to within ±0.1 K, corresponding to a sample temperature of about 277 K. The pulse sequences were designed to scan all SPECIFIC-CP matching conditions by adjusting the rf fields in units of kHz or MAS rate multipliers instead of amplifier output power in W or dB, as described in [13].

All experiments were acquired on microcrystalline U-¹³C,¹⁵N-labeled f-MLF and microcrystalline assemblies of U-¹³C,¹⁵N-labeled HIV-1 capsid protein (CA) C-terminal domain (CTD) encompassing the spacer-peptide 1 (SP1) assembled at hexakisisophoshate (IP6) concentration of 1.6 mM (CA_CTD-SP1/IP6), prepared as described elsewhere [14].

2D NCA spectra of CA_CTD-SP1/IP6 were acquired by accumulating 256 transients for each of the 96 t₁ increments; States-TPPI phase sensitive detection was applied [15], corresponding the ¹⁵N spectral width of 3.5 kHz (57.6 ppm).

¹H continuous-wave (CW) decoupling field during SPECIFIC-CP was set to 105 kHz, unless stated otherwise, to ensure our results are solely affected by the choice of the SPECIFIC-CP matching conditions, i.e. the ¹³C and ¹⁵N rf fields.

All other experimental and processing parameters are detailed in the Supplementary Information, Tables S1 and S2, and SPECIFIC-CP rf fields are indicated in the text below. Data analysis was carried out using Bruker TopSpin v3.6.3, Bruker Dynamics Center v2.8.3, and Matlab R2023a [16].

Spin simulations of ¹⁵N-¹³C SPECIFIC-CP were carried out using SIMPSON v4.2.1 [17–19], see the script and related files in the Supplementary Information. Simulated transfer efficiencies were calculated and scaled with respect to ¹⁵N nutation experiment as described in [20], acquired using U-¹³C,¹⁵N-labeled glycine by accumulating two transients for each of the 110 t₁ increments of 3.5 μs, corresponding to a maximum ¹⁵N pulse length of 385 μs. The same shaped pulses were applied in the spin simulation as in our experiments, i.e., rectangular pulse on the ¹⁵N channel, and a tangentially shaped pulse on the ¹³C channel.

3. Results and Discussion

We explored a broad range of NCA ZQ (n=±1) SPECIFIC-CP matching conditions on f-MLF, setting the rf field during the CP period to be higher for either ¹³C (Figure 1a) or ¹⁵N (Figure 1b), and changing both rf fields simultaneously in increments of 0.7 kHz (5% of the MAS frequency). The results presented in Figure 1 clearly indicate two distinct matching conditions, summarized in Table 1. We denote these NCA SPECIFIC-CP matching conditions as high-power (HP) and low-power (LP) SPECIFIC-CP, for the remainder of the text.

Figure 1.

NCA ZQ (n=±1) SPECIFIC-CP matching conditions at either lower (left) or higher (right) ¹⁵N rf spinlock field with respect to the ¹³C rf field. ω₁ and ω_r are the rf field strength and MAS frequency, respectively. Y axes represent the relative transfer efficiency (left) or the absolute transfer efficiency (right), as compared to a ¹H-¹³C CP spectrum. Each datapoint was calculated by integrating the CA region from the corresponding spectrum acquired by accumulating a total of 32 transients. The spinlock fields are tangential and rectangular for ¹³C and ¹⁵N, respectively. The optimal matching conditions at either low or high ¹⁵N rf fields are colored in magenta and blue, respectively. RF field strengths of the optimal condition are given at the bottom of each bar graph in units of the MAS frequency (14 kHz).

Table 1.

NCA ZQ (n=±1) SPECIFIC-CP optimal matching conditions as identified in Figure 1 in either units of kHz or multipliers of the MAS frequency (ω_r). The MAS frequency during the experiments was 14 kHz.

Matching condition		Rf field ω1
		(kHz)	(Multiplies of ωr)
HP	13C	21	1.50
	15N	35	2.50
LP	13C	19	1.35
	15N	5	0.35

A similar scan of the NCO ZQ (n=±1 and n=±2) SPECIFIC-CP matching conditions did not show a high-efficiency low-power matching condition, see Supplementary Information Figure S1, likely due to the large magnitude of the carbonyl chemical shift tensor (δ_σ ca. −78 to −83 ppm [21]) and the use of weak ¹³C rf fields. When ¹³C rf field strength is less than half of the MAS frequency, around 5 kHz, the effective field for the aliphatic/carbonyl region is ca. 19 kHz, when the carrier is set on the carbonyl/aliphatic region with B₀=14.1 T. This is very close to the ¹⁵N rf field applied to establish the LP NCA SPECIFIC-CP transfer. It is therefore noteworthy that the LP matching condition where ¹³C and ¹⁵N rf fields are 5 and 19 kHz, respectively, may work well at different magnetic fields, and so may the LP NCO SPECIFIC-CP at weak ¹³C rf fields.

We next examined the two HP and LP NCA SPECIFIC-CP matching conditions with respect to several experimental parameters: CP contact time, ¹H decoupling rf field, ¹⁵N chemical shift offset and rf mis-set (Figure 2). The ¹³CA magnetization builds up similarly for both matching conditions (Figure 2a).

Figure 2.

Comparison of HP (black, circles) and LP (magenta, triangles) NCA SPECIFIC-CP for microcrystalline f-MLF. a) NCA SPECIFIC-CP magnetization buildup. b,c,d) NCA SPECIFIC-CP transfer efficiency dependence on experimental conditions: ¹H decoupling rf field during SPECIFIC-CP (b), ¹⁵N chemical shift offset (c), and ¹⁵N rf mis-set (d). Each datapoint was calculated by integrating the CA region from the corresponding spectrum acquired by accumulating a total of 8 transients. The data in each panel is independently scaled to only compare the behavior of each NCA SPECIFIC-CP matching condition.

The HP matching condition requires a ¹H rf field of 95 kHz or above, to avoid its being Hartmann-Hahn matched with the higher ¹⁵N rf fields of ~35 kHz (Figure 2b). In contrast, for the LP matching condition, ¹H rf fields less than 84 kHz can be used without causing polarization transfer to ¹³C for which rf field is set at ~19 kHz. In cases where the maximal available ¹H rf field is below ~85 kHz, the LP matching condition is significantly more efficient. As expected, both NCA SPECIFIC-CP matching conditions work well if high ¹H rf fields exceeding 85 kHz are available.

The effect of ¹⁵N chemical shift offset and rf field mis-set on both matching conditions is as expected, see Figure 2c,d. The use of a lower ¹⁵N rf field leads to a narrower excitation bandwidth, filtering out the sidechain nitrogen atoms of lysine residues as well as reducing the transfer efficiency for those of arginine sidechains. Since both matching conditions utilize a similar ¹³C rf field, the improved selectivity is only with respect to ¹⁵N chemical shifts. Conversely, off-resonance ¹³C spin-lock field during LP ZQ NCA SPECIFIC-CP may result in a competitive DQ (n=±2) or ZQ (n=±2) NCO transfer, where the ¹³CO effective field is about 1.6-1.7 or 2.3-2.4 times the MAS frequency, respectively, see Supplementary Information Figure S2. Both the MAS frequency and magnetic field strength conjointly affect these competitive magnetization transfers.

The use of a lower ¹⁵N rf field also renders NCA SPECIFIC-CP largely insensitive to the rf field mis-set (Figure 2d). These results suggest that the LP matching condition is more robust towards ¹⁵N rf inhomogeneity, possibly explaining why its sensitivity is slightly higher compared to the HP matching condition. These results are also supported by spin simulations for both matching conditions when considering the ¹⁵N nutation experiment of the probe used in this study, see Supplementary Information Figure S3. Previous studies showed the adverse impact of B₁ inhomogeneity on CP efficiency both theoretically and experimentally [20, 22].

We also tested whether the LP NCA SPECIFIC-CP matching condition may be used in experiments designed with a bi-directional magnetization transfer [23]. Comparison of the 1D CAN DCP spectra with both LP and HP matching condition shows that the LP matching condition provides a similar enhancement for magnetization transfer in the opposite direction as expected, see Supplementary Information Figure S3.

Finally, we evaluated the transfer efficiencies and selectivity of the two NCA SPECIFIC-CP matching conditions on microcrystalline assembly of HIV-1 U-¹³C,¹⁵N-CA_CTD-SP1 with inositol hexakis-phosphate (IP6) (Figure 3). Comparison of all assigned cross peak intensities in the two 2D NCA spectra reveals a sensitivity enhancement of about 20% in favor of the LP matching condition (Figure 3b). Intra-sidechain transfer efficiencies are also affected, given the higher transfer selectivity, as expected from our observations in Figure 2c. With the LP matching condition, arginine Nε-Cδ cross peak is weaker (see the outlier in Figure 3b) while lysine Nζ-Cε cross peaks are absent altogether (see the aliased cross peaks in Figure 3a).

Figure 3.

2D NCA DCP spectra of HIV-1 U-¹³C,¹⁵N-CA_CTD-SP1/IP6 microcrystalline assembly utilizing either high (black) or low (magenta) power SPECIFIC-CP for NCA magnetization transfer. a) Overlay of the 2D NCA spectra and resonance assignment according to BMRB ID 30930. Aliased peaks are shown separately in an inset at the top left side of the panel. b) Comparison of NCA assigned cross peaks height using either HP or LP NCA SPECIFIC-CP matching condition. Data points, outlier, and linear fit are colored in grey, yellow, and red, respectively.

All glycine and proline NCA signals were enhanced even though the low ¹⁵N rf field results in a narrower excitation bandwidth. Glycine and proline NCA transfer efficiency may not be enhanced by the LP matching condition at higher magnetic fields (>14.1 T). One solution may be to increase the MAS frequency (up to 25-30 kHz), which would lead to a higher optimal ¹⁵N rf field at the LP matching condition and consequently broaden the ¹⁵N excitation bandwidth.

4. Conclusions

¹³C-¹⁵N SPECIFIC-CP is an essential part of almost every study of proteins by MAS NMR spectroscopy. The low-power matching condition explored here results in a higher transfer efficiency than its high-power counterpart, likely due to improved robustness towards B₁ inhomogeneity. In MAS probes where the available rf field strengths are limited, such as those with rotor diameter of 3.2 mm and above, the low-power NCA SPECIFIC-CP may result in an even higher sensitivity enhancement due to its higher efficiency at weaker ¹H decoupling rf fields. We expect the low-power SPECIFIC-CP matching condition to be broadly utilized for biomolecular MAS NMR studies.

Highlights.

• A weak 15N rf field during NCA SPECIFIC-CP leads to 20% higher efficiency.

• Higher efficiency explained by robustness towards B1 inhomogeneity.

• Low-power matching condition narrower bandwidth improves selectivity.

• Weaker 1H decoupling fields can be applied without causing interfering recoupling.