Dynamic Nuclear Polarization of Deuterated Proteins

Magic-angle spinning nuclear magnetic resonance (MAS NMR) spectroscopy has evolved as a robust and widely applicable technique for investigating the structure and dynamics of biological systems.^[1–3] It is in fact rapidly becoming an indispensable tool in structural biology studies of amyloid,^{[4, 5]} nanocrystalline,^{[6, 7]} and membrane proteins.^[8] However, it is clear that the low sensitivity of MAS experiments to directly detected ¹³C and ¹⁵N signals limits the utility of the approach, particularly when working with systems which are difficult to obtain in large quantities. This limit provides the impetus to develop methods to enhance the sensitivity of MAS experiments, the availability of which will undoubtedly broaden the applicability of the technique. Remarkable progress towards this goal has been achieved by incorporating high-frequency dynamic nuclear polarization (DNP) into the MAS NMR technique.^[9–17] The DNP method exploits the microwave-driven transfer of polarization from a paramagnetic center, such as nitroxide free radical, to the nuclear spins, and has been demonstrated to produce uniformly polarized macromolecular samples. In principle signal enhancements, ε = (γ_e/γ_I) ≈ 660 can be obtained for ¹H and recently signal enhancements of ε = 100–200 were observed in model compounds. However, in applications of DNP to MAS spectra of biological systems, including studies of lysozyme,^[18] and bacteriorhodopsin,^{[16, 19, 20]} the enhancements have been smaller, ε = 40–50. An exception is the amyloidogenic peptide GNNQQNY_7–13 which forms nanocrystals for which the proton T₁ time is long and ε ≈ 100.^[21]

Almost a decade ago in studies of model systems, it was observed that deuteration of the solvent resulted in significant increases in ε^[22] and subsequently many DNP experiments have employed ²H-labelled glasses, such as [D₆]DMSO or [D₈]glycerol/D₂O/H₂O in an approximately 6:3:1 ratio.^[23–25] The approximately 90% ²H concentration level slows the relaxation among protons, while the approximately 10% ¹H concentration level is sufficient to ensure that ¹H–¹H spin diffusion distributes the enhanced polarization uniformly through the sample. The validity of this explanation explains the success of the DNP experiments on GNNQQNY even though the peptide is protonated.

Despite the success of deuteration in improving DNP enhancements, to date it has not been employed in studies of proteins. Herein, we demonstrate that deuteration of the protein itself results in three to five times larger DNP enhancements in its ¹³C MAS spectra. This is a very significant increase in the efficiency of DNP and may well become the preferred means of performing DNP-MAS experiments in biological systems.

For the experiments reported herein we used samples of the SH3 domain of the protein α-spectrin in which at all the amino acids were deuterated, the samples were then recrystallized in appropriate H₂O/D₂O buffers to adjust the ¹H/²H ratio at the exchangeable sites. Subsequently the protein was dispersed in a [D₈]glycerol/D₂O/H₂O matrix. Figure 1 shows a comparison of one-dimensional (1D) ¹³C MAS NMR spectra recorded with microwave irradiation using the pulse sequences shown in Figure 1A and B, with cross-polarization (CP) for the protonated SH3 samples (Figure 1C) and deuterated (Figure 1D) SH3 samples. Figures 1E and F show MAS spectra recorded with direct ¹³C excitation at different recycle delays (RDs). The DNP enhancement observed in the ¹³C CP-MAS spectrum from a protonated and fully ¹³C/¹⁵N labeled sample is ε ≈ 31 (Figure 1C), comparable to enhancements previously reported.^{[16, 19, 20]}

Figure 1.

The pulse sequences used to record the DNP enhanced ¹³C spectra with A) CP-MAS and B) direct ¹³C excitation, at approximately 98 K and approximately 9 kHz MAS. C) ¹³C CP-MAS spectrum of protonated SH3 with DNP. DNP enhanced D) ¹³C CP-MAS and E),F) MAS spectra of the deuterated-SH3 with 50% exchangeable proton content. A relaxation delay (RD) of 2 s (C–E) and 12 s (F) were used. Continuous-wave microwave irradiation was used while acquiring the DNP enhanced spectra. For calculation of the DNP enhancement, the spectra were recorded with microwave irradiation and compared to the spectra recorded without microwave irradiation under exactly same experimental conditions. The spectra are plotted with the same noise level, to allow direct comparison.

In the deuterated protein, the efficiency of the DNP enhancement of the ¹³C CP-MAS experiments increases by a factor of approximately 3.9 (ε = 120, Figure 1D) compared to protonated samples. Furthermore, by using direct ¹³C excitation on the deuterated sample, the enhancement is further increased by a factor of around 4.8 to ε = 145 (Figure 1F) compared to the ¹³C CP-MAS experiment on the fully protonated SH3.

The sensitivity increase in a ¹³C DNP-CP-MAS experiment is determined by the enhancement of the proton spin reservoir which is subsequently transferred to ¹³C and is limited by the ratio (γ_e/γ_¹H). Similarly, in a ¹³C direct excitation MAS experiment, the enhancement depends on γ_e/γ_¹³C which is a factor of four larger. Thus, the enhancements in direct excitation experiments are expected to be larger, although the required recycle delays could be longer because of slower spin diffusion in the ¹³C reservoir. These ideas were recently confirmed experimentally^[26] and it was also demonstrated that the maximum in the ¹H enhancement field profile is identical for ¹³C. However, because of the lower value of γ_¹³C, the optimal field for the direct ¹³C enhancements is on the opposite side of the profile. Nevertheless, in the case of SH3, the ¹³C T₁ times are short and in the MAS DNP experiment we observe ε ≈ 148 (Figure 1E). This enhancement is significantly larger than the enhancement obtained from the CP experiment (Figure 1D). However, the ¹³C DNP enhancement observed for the fully protonated SH3 sample recorded with direct ¹³C excitation is ε ≈ 8. These observations strongly suggest that protein deuteration, as well as direct ¹³C excitation, is responsible for the further increase in ε. Supporting this hypothesis is the fact that the direct ¹⁵N DNP enhancement is ε = 207 for deuterated SH3 with protons at 50% of the exchangeable sites. To quantify the signal per unit time, we recorded a ¹³C spectrum with direct excitation and a short relaxation delay of 2 s for a deuterated protein with 50% protons at the exchangeable sites (Figure 1E). The intensity at the CO and C_glycerol signal is reduced compared to the CP spectrum of the fully protonated SH3 sample (Figure 1C), whereas, the intensity in the aliphatic region is slightly increased.

Figure 2A shows the dependence of the DNP enhancement on the exchangeable proton content in the protein and buffer. The enhancements obtained for various nuclei (¹H, ¹³C, and ¹⁵N) and by using different experimental approaches (MAS and CP-MAS), depend strongly on the exchangeable proton content. A gradual increase of the ¹³C and ¹⁵N DNP enhancement is observed by increasing the exchangeable proton content from 15% to approximately 50%. For all types of experiments, the fully protonated SH3 has lower DNP enhancements. The data suggests that a plethora of protons attenuates the enhancement, and a paucity interferes with the distribution of polarization through spin diffusion causing a reduced signal enhancement.^[23–25]

Figure 2.

The dependence of the A) DNP enhancement and B) T₁ relaxation time (seconds) on the overall ¹H content. For comparison, a fully protonated (at exchangeable and non-exchangeable sites) and three different deuterated SH3 (at non-exchangeable sites) proteins were used. The proton content at the exchangeable sites of the protein was tuned by recrystallization of SH3 in buffers containing different H₂O/D₂O ratios. The H₂O contents were set around 15, 25, 50 and 100% to be able to cover the full range of exchangeable proton content. The spectra were recorded at a temperature of approximately 98 K, an MAS frequency of approximately 9 kHz, and with approximately 5 W microwave irradiation in zirconium rotors. The experimental data points are connected with lines as a guide to the eye, the data points are discontinuous.

The fully protonated SH3 sample shows slightly shorter T₁ values (except for ¹⁵N) than the perdeuterated SH3 samples, which all show similar values. Using higher concentrations of biradical in deuterated proteins can circumvent this problem, provided that the radical is sufficiently bulky and does not diffuse into the crystal lattice and broaden the ¹³C lines. In addition, the possibility of using ²H as an initial polarization transfer source could enhance the absolute sensitivity in deuterated proteins and help to exploit the increase in polarization enhancement further.

The resolution observed at the cryogenic temperature of MAS-DNP experiments at 400 MHz is currently not sufficient for complete assignment of the signals of a fully labeled protein. Accordingly, it is of interest to increase the temperature to achieve higher resolution, partially sacrificing enhanced DNP sensitivity.^[27] Nevertheless, there might be a compromise temperature where there is sufficient resolution and a sufficiently high DNP enhancement. To study the temperature effect, we measured the ¹H, ¹³C, and ¹⁵N enhancements at elevated temperatures, from 98 K up to 200 K, for fully protonated and perdeuterated SH3 samples (Figure 3). In this temperature range, an increase of 20 K results in a decrease of around 30–40% in the enhancement. Above 160 K, it becomes impractical to perform DNP-MAS NMR for the fully protonated or the perdeuterated protein with 15% protonation level, since the DNP enhancements decrease dramatically. For the SH3 sample with a 50% protonation level, the enhancement decreases by 90% between 98 and 178 K, nevertheless, the DNP enhancements are still ε ≈ 10 and ε ≈ 15 in ¹³C CP and direct-excitation MAS-NMR spectra. Thus, this sample is suitable for high-temperature DNP.

Figure 3.

The dependence of the ¹H, ¹³C, and ¹⁵N DNP enhancements on temperature. Results for the protonated and perdeuterated (15 and 50% exchangeable proton content) SH3 are shown. Enhancement values are calculated at each temperature for each type of experiment. For the protonated SH3, only the ¹³C CPMAS DNP enhancement values are shown.

In conclusion, we have shown that perdeuteration of a protein has remarkable effects on the observed DNP enhancements. Superior DNP enhancements are obtained for perdeuterated SH3 samples of up to 3.9 and 18.5 times for ¹³C CP-MAS, and ¹³C MAS experiments, respectively, compared to the same type of experiments in fully protonated SH3. The optimum exchangeable proton content is found to be approximately 50% which results in the maximum enhancement of ε ≈ 148 in a ¹³C MAS NMR spectrum obtained using a ZrO₂ rotor. By taking into account the 20% increase in enhancement by using sapphire rotors, higher ¹³C DNP enhancement of ε ≈ 180 can be expected. Moreover, by using the deuterated SH3 protein with 50% proton content at the exchangeable sites, it is possible to increase the temperatures at which DNP experiments still yield considerable enhancements. We expect that the use of perdeuterated proteins in DNP-MAS NMR will open new possibilities in the application of these techniques to difficult biological problems.

Experimental Section

Details of the sample preparation by unfolding, exchanging, and refolding of perdeuterated and protonated SH3 are described elsewhere.^{[28, 29]} The samples for the DNP-MAS measurements were prepared by dissolving the protein in 1:3:6 vol% H₂O/D₂O/glycerol solution which forms a stable glassy matrix and cryoprotects the protein.^[30] Totapol biradical^[31] is additionally dissolved in this solution at a concentration of 20 mm, corresponding to an electron concentration of 40 mm.

All solid-state DNP-MAS NMR experiments were performed on a commercial Bruker DNP spectrometer operating at a ¹H frequency of 400 MHz and microwave frequency of 263 GHz. Spectra were recorded using a triple resonance, low-temperature, HCN DNP probe employing 3.2 mm ZrO₂ rotors. Cryogenic temperatures were achieved and controlled with Bruker low-temperature MAS accessory. The signal enhancement is achieved in situ, directly at the magnetic field inside the probe. The millimeter wave power, approximately 5 watts, is generated by Bruker gyrotron oscillator.

All of the DNP enhanced ¹³C MAS spectra were recorded at ω_r/2π =8888 Hz and using π/2 pulses of 4 and 5 µs for ¹H and ¹³C, respectively, and a CP contact time of 2 ms. A sapphire rotor was used to determine the enhancement difference between zirconia and sapphire rotors and found that use of sapphire rotor results in an approximately 20% increase in the observed DNP enhancement values. In addition, we note that the microwave irradiation reduces the apparent T₁ relaxation times by 30%, most probably because of sample heating.