Succinyl-DOTOPA: An effective triradical dopant for low-temperature dynamic nuclear polarization with high solubility in aqueous solvent mixtures at neutral pH

Abstract

We report the synthesis of the nitroxide-based triradical compound succinyl-DOTOPA and the characterization of its performance as a dopant for dynamic nuclear polarization (DNP) experiments in frozen solutions at low temperatures. Compared with previously described DOTOPA derivatives, succinyl-DOTOPA has substantially greater solubility in glycerol/water mixtures with pH > 4 and therefore has wider applicability. Solid state nuclear magnetic resonance (ssNMR) measurements at 9.39 T and 25 K, with magic-angle spinning at 7.00 kHz, show that build-up times of DNP-enhanced, cross-polarized ¹³C ssNMR signals are shorter and that signal amplitudes are larger for glycerol/water solutions of L-proline containing succinyl-DOTOPA than for solutions containing the biradical AMUPol, with electron spin concentrations of 15 mM or 30 mM, resulting in greater net sensitivity gains from DNP. In similar measurements at 90 K, AMUPol yields greater net sensitivity, apparently due to its longer electron spin-lattice and spin-spin relaxation times. One- and two-dimensional ¹³C ssNMR measurements at 25 K on the complex of the 27-residue peptide M13 with the calcium-sensing protein calmodulin, in glycerol/water with 10 mM succinyl-DOTOPA, demonstrate the utility of this compound in DNP-enhanced ssNMR studies of biomolecular systems.

Graphical Abstract

Introduction

The introduction of nitroxide-based biradicals as dopants for dynamic nuclear polarization (DNP) experiments on frozen solutions in 2004 [1, 2] was a major contributing factor for the growth of DNP-related activity in the solid state nuclear magnetic resonance (ssNMR) community. A variety of biradical dopants have been developed subsequently with improved performance under various experimental conditions [3–17]. In 2010, we reported that the nitroxide-based triradical compound DOTOPA has favorable properties for low-temperature DNP, under certain conditions producing larger DNP-enhanced signals and more rapid build-up of nuclear spin polarizations under microwave irradiation [18]. We subsequently described a series of DOTOPA derivatives that have greater solubility at neutral or basic pH in water/glycerol mixtures than does the original DOTOPA compound [19]. Favorable properties of DOTOPA have also been reported by others [20].

Although we have used DOTOPA and DOTOPA derivatives in several published studies [21–26], their solubility is still a limitation, leading to sample-to-sample variations in DNP-enhanced signals and build-up times. Our previously described DOTOPA derivatives are also not useful in experiments where low glycerol-to-water ratios are required, such as studies of transient intermediates that are prepared by rapid mixing and freeze-trapping [27].

In this paper, we report the synthesis and DNP properties of a new derivative, succinyl-DOTOPA, which is highly soluble at neutral and basic pH. In experiments on frozen glycerol/water solutions of ¹³C-enriched L-proline at 9.4 T, 25 K, and magic-angle spinning (MAS) frequencies up to 7 kHz, we show that succinyl-DOTOPA produces greater DNP-enhanced, cross-polarized ¹³C NMR sensitivity than does the commercially available biradical AMUPol [6] at equivalent electron spin concentrations. At 90 K, greater sensitivity is obtained with AMUPol, apparently because the 1.5 W microwave power available to us in these experiments is not sufficient to saturate electron spin transitions in succinyl-DOTOPA. Measurements of temperature-dependent electron spin relaxation times T_1e and T_2e provide insight into the DNP properties of these compounds.

Materials and methods

Synthesis of 4-{N,N-di-[2-(succinate, sodium-trimethylamine salt)-3-(TEMPO-4’-oxy)-propyl]}-amino TEMPO (succinyl-DOTOPA)

Succinyl-DOTOPA was prepared by reacting DOTOPA with succinic anhydride (Fig. 1). Succinic anhydride and sodium bicarbonate were purchased from Sigma–Aldrich, and Mallinckrodt, respectively. DOTOPA was synthesized as described previously [18]. All solvents were reagent grade and used without further purification. Silica column chromatography was performed on silica gel (130–270 mesh, 60 Å pore size, Sigma-Aldrich). Liquid chromatography-mass spectrometry (LC-MS) data were recorded on a Thermo Scientific UltiMate 3000 LC-MS system, equipped with a MSQ Plus single quadrupole mass spectrometer with electrospray ionization.

Fig. 1:

Synthesis of succinyl-DOTOPA (iii) by reaction of DOTOPA (i) with succinic anhydride (ii).

Succinic anhydride (1.0 g, 10 mmol) was mixed with sodium bicarbonate (1.0 g, 12 mmol) and DOTOPA (314 mg, 0.5 mmol) in 7.5 ml ethyl acetate as described by Lugemwa et al. [28]. The mixture was stirred at room temperature overnight. After filtration, the residue was washed with ethyl acetate (5.0 ml) three times and dried under nitrogen gas. After loading the filtrate on the silica gel column using CH₂Cl₂ with 0.5% triethylamine, 1–3% methanol with 0.5% triethylamine was applied to elute the final product. After removing the organic solvents and drying under vacuum overnight, we obtained a 90% yield of red oily product. LC-MS data (Figs. S1a and S1b) indicated a mass of 828.10 Da (C₄₁H₇₁N₄O₁₃³*, theoretical average mass = 828.02 Da).

The solubility of succinyl-DOTOPA in glycerol/water (40:60 volume ratio) as a function of pH at 24° C was determined by preparing a solution with [succinyl-DOTOPA] = 48 mM at pH 9.5, then reducing the pH in steps of 0.5 units by addition of HCl. After each step, insoluble material was removed by centrifugation and the concentration was determined by optical absorbance at 430 nm, using an extinction coefficient for nitroxide groups equal to 11 cm⁻¹ M⁻¹. Results in Fig. S1c indicate a solubility above 15 mM at pH < 4.0.

Sample preparations

¹³C₅,¹⁵N-L-proline was dissolved at a concentration of 100 mM in ²H₂O:¹²C₃,²H₈-glycerol:¹H₂O (50:40:10 ratio of initial volumes) with 50 mM phosphate buffer, pH 7.5. ¹²C₃,²H₈-glycerol was purchased from Cambridge Isotope Laboratories, Inc. Stock solutions of 75 mM succinyl-DOTOPA and 754 mM AMUPol (based on absorbance at 430 nm, with an extinction coefficient for nitroxide groups of 11 cm⁻¹ M⁻¹) were prepared in ²H₂O with 100 mM phosphate buffer, pH 8, and were included in the glycerol/water mixtures to achieve the desired dopant concentrations. An additional sample of ¹³C₅,¹⁵N-L-proline in the same solvent mixture was prepared with 5 mM DyCl₃ and 5 mM disodium ethylenediaminetetraacetate (5 mM Dy-EDTA).

For measurements on the calmodulin/M13 complex, 20 μl of a 11 mM solution of unlabeled calmodulin in ²H₂O and 9.2 μl of a 12 mM solution of selectively ¹⁵N,¹³C-labeled M13 in ²H₂O were combined with 3.5 μl of 326 mM CaCl₂ in ²H₂O, 7.5 μl of 200 mM Tris buffer in ¹H₂O (pH 8.0), 5.0 μl of 75 mM succinyl-DOTOPA in ²H₂O, and 30 μl of ¹²C₃,²H₈-glycerol. M13 was synthesized on the 0.1 mmol scale by automated solid-phase methods (Protein Technologies Tribute synthesizer) using standard fluorenylmethyloxycarbonyl (Fmoc) chemistry and a Fmoc-Met-Wang resin (0.53 mEq/g, Peptides International, Inc.), and was purified by high-performance liquid chromatography, using a water/acetonitrile gradient with 0.1% trifluoroacetic acid and a preparative C18 reverse-phase column. Bovine calmodulin was obtained as a lyophilized powder from Ocean Biologics.

ssNMR measurements

DNP-enhanced ssNMR measurements were performed at 9.4 T (400.9 MHz and 100.8 MHz for ¹H and ¹³C NMR frequencies, respectively), with a Bruker Avance III spectrometer console and the home-built DNP-MAS ssNMR probe described previously [25, 29], which uses zirconia rotors (O’Keefe Ceramics, Inc.) with 4.0 mm outer diameters and maximum sample volumes of 80 μl. Rotor caps and drive tips were obtained from Revolution NMR, LLC. An extended interaction oscillator (Communications and Power Industries, LLC) and quasi-optical interferometer (Thomas Keating Ltd.) provided circularly polarized microwaves at 263.9 GHz with 1.5 W power [25].

Sample solutions in MAS rotors were frozen by immersion in liquid nitrogen, then loaded into the pre-cooled DNP-MAS ssNMR probe below the NMR magnet. The probe was then raised into the magnet, with the transfer line (carrying helium or nitrogen) attached and with the sample spinning (and remaining frozen) throughout this process.

One-dimensional (1D) ¹³C ssNMR spectra were obtained with ¹H-¹³C cross-polarization (CP), using a ¹³C radio-frequency (RF) field strength of 29 kHz and ¹H RF field strengths of 43 kHz, 37 kHz, and 33 kHz at ν_MAS values of 7.00, 4.00 and 2.00 kHz, respectively. Two-pulse phase modulated (TPPM) ¹H decoupling [30] was applied during ¹³C signal acquisition, with an 80 kHz decoupling field. ¹H polarization build-up times T_DNP (or spin-lattice relaxation times T₁ for the Dy³⁺-doped sample described below) were measured with the saturation-recovery method, using a train of 16 π/2 pulses with 10 ms delays to saturate the initial nuclear spin polarizations of both ¹H and ¹³C. In all other measurements, recycle delays were then set to 1.26T_DNP (or 1.26T₁). DNP signal enhancement factors ε for ¹³C are ratios of signal areas with and without microwave irradiation, from spectra obtained with ¹H-¹³C CP. Signal areas were obtained by integrating the full¹³C spectra, including MAS sidebands.

Sample cooling and temperature control

For measurements at a nominal sample temperature of 25 K, samples were cooled with helium as previously described [25, 29, 31]. The liquid helium tank gauge pressure (2.5 psi) and helium transfer line needle valve opening were adjusted to produce a temperature of 25 ± 2 K for a sample of KBr powder at MAS frequency ν_MAS = 7.00 kHz, as determined from the temperature-dependent ⁷⁹Br spin-lattice relaxation rate [32]. Sample temperatures were not significantly different at ν_MAS = 2.0 kHz or 4.0 kHz. The same conditions, which correspond to a liquid helium consumption rate of approximately 2.0 liters per hour, were then used for DNP measurements on frozen proline solutions.

For measurements at a nominal sample temperature of 90 K, a liquid helium tank (Cryofab model CMSH-10) was filled with liquid nitrogen instead of liquid helium. The same transfer line was used (Janis Research model FHT-ST) and no other modifications were required for liquid nitrogen cooling. At 90 K and ν_MAS = 7.0 kHz, liquid nitrogen consumption was 0.5 liters per hour.

During measurements at 90 K, sample temperatures were determined from the temperature-dependent peak wavelength of photoluminescence (PL) from a thin film of CdSeS/ZnS core/shell quantum dots (QDs), which was painted on the outer surface of the MAS rotor as previously described [33]. QDs were purchased from Ocean NanoTech (product number QSP-665) and co-dissolved with varnish (Lake Shore Cryotronics, VGE-7031) in toluene, with approximately 5 μl of varnish and 0.15 mg of QDs in 100 μl toluene. The QD solution was applied to the MAS rotor surface and dried in air to create the desired film. PL spectra were excited by a battery-powered laser operating at 405 nm with 200 mW power and detected with a fiber optic spectrometer (StellarNet, Inc., model CXR-100). To carry the excitation and PL emission light, two multimode optical fibers (400 μm diameter, 0.39 numerical aperture, Thorlabs part number FT400UMT) were inserted into the MAS module of our DNP-MAS ssNMR probe through holes in the Teflon fitting that connects the transfer line to the MAS model (see Fig. 1 of ref. [25]). In this way, the ends of the two optical fibers were positioned between turns of the solenoidal radio-frequency (RF) sample coil, approximately 2 mm from the MAS rotor surface.

Fig. S2 shows the temperature dependence of the PL peak wavelength, calibrated against sample temperatures that were determined simultaneously from ⁷⁹Br spin-lattice relaxation rates of a spinning KBr powder sample, cooled with liquid helium. Although the QDs used in these measurements have longer PL wavelengths than the QDs used in the original demonstration of this method for determining sample temperatures under MAS [33], the sensitivity to temperature is approximately the same (i.e., 13 nm change in PL peak from 50 K to 250 K). Based on the PL spectrum, which was monitored continuously during DNP measurements, sample temperatures were maintained at 90 ± 2 K.

EPR measurements

For temperature-dependent electron paramagnetic resonance (EPR) measurements, dopants were dissolved in ²H₈-glycerol/²H₂O (40:60 ratio of initial volumes) at 0.1 mM concentrations. Electron spin-lattice (T_1e) and spin-spin (T_2e) relaxation times were measured from 20 K to 90 K in 10 K increments with the magnetic field at 1.1956 T, using a Bruker Elexsys EPR spectrometer. T_1e measurements used an echo-detected inversion-recovery pulse sequence (π − trec − π/2 − δt − π − δt, with δt = 0.6 μs and incremented t_rec). T_2e measurements used a standard spin echo sequence (π/2 − t_echo/2 − π − t_echo/2 with incremented t_echo).

Results

Fig. 2 shows 1D ¹³C and ¹H ssNMR spectra of a frozen ¹³C₅,¹⁵N-L-proline solution containing 10 mM succinyl-DOTOPA at 25 K with ν_MAS= 2.00, 4.00, and 7.00 kHz. Spectra are shown with and without microwave irradiation. For comparison, spectra of a solution containing 15 mM AMUPol, so that the total electron spin concentration c_e,tot is also 30 mM, are shown in Fig. S3. Spectra of these samples at 90 K are shown in Figs. S4 and S5. Spectra of frozen ¹³C₅,¹⁵N-L-proline solutions with c_e,tot = 15 mM and c_e,tot = 45 mM (5 mM and 15 mM succinyl-DOTOPA, 7.5 mM and 22.5 mM AMUPol) were also obtained (data not shown).

Fig. 2:

¹³C (left) and ¹H (right) ssNMR spectra of a frozen glycerol/water solution of ¹⁵N,¹³C-L-proline containing 10 mM succinyl-DOTOPA at 25 K, with MAS frequencies of 7.00 kHz (a), 4.00 kHz (b), and 2.00 kHz (c). The magnetic field strength was 9.39 T. Blue spectra were obtained with microwave irradiation and with two scans, using 1.5 W of circularly polarized microwaves at 263.9 GHz. Red spectra were obtained without microwave irradiation and with 32 scans.

Fig. 3 shows the DNP build-up times T_DNP for all six samples at the two temperatures and three values of ν_MAS, obtained from saturation-recovery measurements with ¹H-¹³C CP and ¹³C signal detection. The saturation-recovery data and single-exponential fits are shown in Figs. S6 and S7. Values of T_DNP are roughly inversely proportional to dopant concentration, consistent with the simplistic idea that (on the average) each dopant molecule polarizes a volume around itself that is inversely proportional to the dopant concentration. For the same value of c_e,tot, T_DNP is smaller when succinyl-DOTOPA is used, consistent with the idea that the fraction of dopant molecules that participate in the cross-effect DNP mechanism is larger for the triradical succinyl-DOTOPA than for the biradical AMUPol. Values of T_DNP at 25 K and 90 K are nearly equal and do not depend strongly on ν_MAS, especially for samples that contain succinyl-DOTOPA.

Fig. 3:

Dependence of the build-up time for cross-polarized ¹³C ssNMR signals under microwave irradiation (T_DNP) on MAS frequency and dopant concentration at 25 K and 90 K, for frozen glycerol/water solutions of ¹⁵N,¹³C-L-proline containing either succinyl-DOTOPA (left) or AMUPol (right). Values of T_DNP were determined from single-exponential fits to build-up curves in Figs. S6 and S7.

The DNP-enhanced ¹³C ssNMR sensitivity is proportional to the absolute ¹³C signal area with microwave irradiation and inversely proportional to T_DNP^0.5. Signal areas divided by T_DNP^0.5 are plotted as functions of c_e,tot in Fig. 4, for the three values of ν_MAS, two temperatures, and two dopants. Under our experimental conditions, succinyl-DOTOPA produces significantly greater cross-polarized ¹³C ssNMR sensitivity at 25 K, with c_e,tot = 15 mM or c_e,tot = 30 mM. At 90 K, AMUPol produces greater or equal sensitivity. In addition, for samples containing succinyl-DOTOPA, sensitivities at 25 K are greater than at 90 K by factors of 9–12 at ν_MAS= 7.00 kHz, factors of 9–18 at ν_MAS = 4.00 kHz, and factors of 14–26 at ν_MAS = 2.00 kHz. For samples containing AMUPol, sensitivities at 25 K are greater than at 90 K by factors of 3–6 at ν_MAS = 7.00 kHz, factors of 3–11 at ν_MAS = 4.00 kHz, and factors of 9–10 at ν_MAS = 2.00 kHz. Leavesley et al. have also reported that dopants with multiple unpaired electrons yield smaller T_DNP values and greater overall sensitivity in DNP experiments at 4 K without MAS [20].

Fig. 4:

Dependence of the net DNP-enhanced, cross-polarized ¹³C sensitivity (¹³C signal area in arbitrary units with microwave irradiation divided by T_DNP^0.5) on electron spin concentration and temperature for frozen glycerol/water solutions of ¹⁵N,¹³C-L-proline containing either succinyl-DOTOPA (blue or green symbols at 25 K or 90 K, respectively) or AMUPol (red or yellow symbols at 25 K or 90 K, respectively). Data are shown for MAS frequencies of 7.00 kHz (left), 4.00 kHz (middle), and 2.00 kHz (right).

Data in Fig. 4 show that the differences in ¹³C ssNMR sensitivity at 25 K between samples containing succinyl-DOTOPA or AMUPol become smaller at higher electron spin concentrations. At an electron spin concentration of 45 mM, the average volume per electron spin is 36.8 nm³ = (3.33 nm)³. Given that the diameter of a succinyl-DOTOPA molecule is roughly 1.5 nm, one expects the distinction between triradical and biradical dopants to become less pronounced at such high concentrations.

These measurements of sensitivity do not include the temperature dependence of noise in the spectra. In practice, noise in our measurements does not change significantly between 90 K and 25 K because only the sample coil is cooled by the cryogens, whereas other components of the probe circuit remain near 180 K.

Tables 1 and 2 contain the complete results for ¹H and ¹³C signals, DNP enhancement factors ε (ratio of signal areas with and without microwave irradiation), T_DNP values, and ¹³C sensitivities for samples containing succinyl-DOTOPA or AMUPol. For comparison, results are also given for a frozen glycerol/water solution of ¹³C₅,¹⁵N-L-proline containing 5 mM Dy-EDTA, which reduces the ¹H T₁ value but does not produce DNP [31]. Compared with this Dy³⁺-doped sample, samples with 5 mM or 10 mM succinyl-DOTOPA exhibit roughly 30-fold greater cross-polarized ¹³C sensitivity with microwave irradiation at 25 K and ν_MAS = 7.00 kHz.

Table 1:

Summary of measurements of ¹³C and ¹H signal amplitudes, DNP enhancement factors s, and build-up times T_DNP at 25 K. Although signal amplitudes are in arbitrary units, all ¹³C signal amplitudes in Tables 1 and 2 are on the same scale and can be directly compared with one another. All ¹H signal amplitudes are also on the same scale and can be directly compared with one another. The ¹³C sensitivity is defined as the ¹³C signal amplitude with microwaves on divided by the square root of T_DNP For the sample doped with Dy-EDTA, build-up times are spin-lattice relaxation times T₁, independent of microwave irradiation. ¹³C T_DNP values are build-up times with ¹H-¹³C cross-polarization.

sample	MAS frequency (kHz)	13C signal, μwaves on, 2 scans	13C signal, μwaves off, 32 scans	13C ε	13C TDNP (or T1)	1H signal, μwaves on, 2 scans	1H signal, μwaves off, 32 scans	1H ε	1H TDNP (or T1)	relative 13C sensitivity
5 mM succinyl-DOTOPA	2.0	611	158	61.9	5.40 ± 0.49	138	83	26.7	4.56 ± 0.01	263
	4.0	526	106	79.0	5.37 ± 0.11	118	52	36.6	4.92 ± 0.01	227
	7.0	401	78	82.8	5.46 ± 0.07	139	41	54.9	5.54 ± 0.01	172
10 mM succinyl-DOTOPA	2.0	555	86	102.9	2.90 ± 0.15	129	50	41.7	2.26 ± 0.02	326
	4.0	381	72	84.1	2.62 ± 0.06	111	28	64.3	2.78 ± 0.01	235
	7.0	260	45	92.1	2.58 ± 0.14	100	23	68.3	3.19 ± 0.01	161
15 mM succinyl-DOTOPA	2.0	393	62	100.8	1.34 ± 0.08	117	24	78.7	0.72 ± 0.01	339
	4.0	307	53	92.3	2.08 ± 0.02	98	21	73.0	1.16 ± 0.02	213
	7.0	173	32	85.3	1.99 ± 0.03	89	17	86.5	1.51 ± 0.01	123
7.5 mM AMUPol	2.0	479	154	49.7	7.35 ± 0.19	114	59	31.1	5.38 ± 0.01	177
	4.0	419	103	65.2	9.17 ± 0.17	103	52	31.4	6.34 ± 0.01	138
	7.0	244	30	128.5	10.32 ± 0.26	89	25	57.7	8.83 ± 0.01	76
15 mM AMUPol	2.0	418	132	50.7	3.32 ± 0.08	116	57	32.9	3.08 ± 0.01	229
	4.0	447	99	71.9	3.78 ± 0.09	104	49	33.7	3.36 ± 0.01	230
	7.0	244	30	128.5	4.63 ± 0.03	96	26	59.3	3.83 ± 0.01	114
22.5 mM AMUPol	2.0	570	80	114.1	3.00 ± 0.23	257	54	76.5	2.93 ± 0.01	329
	4.0	364	67	87.2	3.12 ± 0.07	105	33	50.9	3.25 ± 0.02	206
	7.0	222	38	98.8	3.00 ± 0.14	97	20	76.4	3.32 ± 0.01	128
5 mM Dy-EDTA	2.0		167	–	4.17 ± 0.02		96	–	5.08 ± 0.01	5.1
	4.0		153	–	3.00 ± 0.03		95	–	4.85 ± 0.01	5.5
	7.0		164	–	3.31 ± 0.01		74	–	4.11 ± 0.01	5.6

Table 2:

Same as Table 1, but for measurements at 90 K. ¹³C T_DNP values are build-up times with ¹H-¹³C cross-polarization.

sample	MAS frequency (kHz)	13C signal, μwaves on, 2 scans	13C signal, μwaves off, 32 scans	13C ε	13CTDNP (or T1)	1H signal, μwaves on, 2 scans	1H signal, μwaves off, 32 scans	1H ε	1H TDNP (or T1)	Relative 13C sensitivity
5 mM succinyl-DOTOPA	2.0	44	26	26.9	5.49 ± 0.60	16	20	13.5	4.85 ± 0.01	19
	4.0	56	31	29.1	5.34 ± 0.08	20	18	17.8	5.21 ± 0.01	24
	7.0	44	25	28.5	5.38 ± 0.03	20	18	18.2	5.26 ± 0.01	19
10 mM succinyl-DOTOPA	2.0	36	24	23.7	2.78 ± 0.11	14	18	13.1	2.80 ± 0.02	21
	4.0	27	22	19.5	2.42 ± 0.43	14	17	13.9	2.74 ± 0.01	16
	7.0	21	18	19.3	2.54 ± 0.12	17	16	17.3	2.34 ± 0.02	13
15 mM succinyl-DOTOPA	2.0	15	18	12.9	1.56 ± 0.05	10	17	10.1	1.37 ± 0.01	13
	4.0	18	18	15.6	1.61 ± 0.12	12	16	12.2	1.42 ± 0.01	12
	7.0	15	15	15.8	1.76 ± 0.03	13	15	14.1	1.51 ± 0.01	10
7.5 mM AMUPol	2.0	55	18	48.0	8.59 ± 0.05	25	14	29.3	8.98 ± 0.01	20
	4.0	55	12	75.7	8.09 ± 0.08	27	14	31.4	8.03 ± 0.02	18
	7.0	84	15	89.2	8.50 ± 0.03	44	13	51.9	8.21 ± 0.01	26
15 mM AMUPol	2.0	42	16	43.0	4.24 ± 0.11	22	14	24.6	5.35 ± 0.02	23
	4.0	135	17	129.2	4.29 ± 0.02	61	14	67.8	4.62 ± 0.01	69
	7.0	60	11	83.0	4.04 ± 0.03	49	13	58.9	4.81 ± 0.02	28
22.5 mM AMUPol	2.0	66	15	72.4	2.48 ± 0.24	26	7	60.8	2.85 ± 0.01	38
	4.0	32	9	56.7	1.72 ± 0.15	15	11	21.9	2.52 ± 0.01	18
	7.0	36	7	84.6	2.47 ± 0.15	21	10	35.0	2.69 ± 0.01	21
5 mM Dy-EDTA	2.0	–	46	–	5.37 ± 0.02	–	25	–	4.53 ± 0.01	1.2
	4.0	–	35	–	5.92 ± 0.01	–	19	–	5.23 ± 0.02	0.9
	7.0	–	30	–	6.17 ± 0.02	–	20	–	5.29 ± 0.02	0.8

Fig. 5 compares cross-polarized ¹³C signal areas without microwave irradiation for samples containing succinyl-DOTOPA, AMUPol, and Dy-EDTA. At 25 K, samples containing either succinyl-DOTOPA or AMUPol show strong depolarization effects [26, 34], i.e., reduced signal areas relative to the sample containing Dy-EDTA. The extent of depolarization increases with increasing c_e,tot and with increasing ν_MAS With c_e,tot = 15 mM and c_e,tot = 30 mM and with ν_MAS = 7.00 kHz, samples containing AMUPol show larger depolarization effects than do samples containing succinyl-DOTOPA. The opposite is true under certain other conditions.

Fig. 5:

Dependences of cross-polarized ¹³C signal areas, without microwave irradiation, on MAS frequency for frozen glycerol/water solutions of ¹⁵N,¹³C-L-proline containing succinyl-DOTOPA (blue symbols), AMUPol (red symbols), or Dy-EDTA (black symbols) with the indicated concentrations. Data are shown for measurements at 25 K (a) and 90 K (b).

Differences between samples containing succinyl-DOTOPA or AMUPol may be attributed in part to differences in electron spin relaxation properties. Fig. 6 shows the temperature dependences of T_1e and T_2e for the two dopants in frozen glycerol/water solutions, measured at an EPR frequency of 33.6 GHz. T_1e values for AMUPol are significantly larger than those for succinyl-DOTOPA. Assuming that this is also true at 263.9 GHz, the larger T_1e value at 25 K may contribute to the stronger dependence of depolarization on ν_MASfor AMUPol in Fig. 5a.

Fig. 6:

Temperature-dependences of the electron spin-lattice (T_1e, left ) and spin-spin (T_2e, right) relaxation times in frozen glycerol/water solutions of succinyl-DOTOPA (blue symbols) and AMUPol (red symbols) at 1.1956 T.

At 25 K and c_e,tot = 15 mM, depolarization effects at ν_MAS= 2.00 kHz are small, suggesting that T_1e values for both AMUPol and succinyl-DOTOPA are less than approximately 0.5 ms at 263.9 GHz, since depolarization is expected to be negligible when ν_MAST_1e < 1 [26]. At 25 K, c_e,tot = 15 mM, and ν_MAS = 7.00 kHz, greater depolarization is observed for the sample containing AMUPol, suggesting that T_1e for AMUPol is approximately 100 μs, while T_1e for succinyl-DOTOPA is somewhat shorter. With c_e,tot = 30 mM or 45 mM, samples containing either AMUPol or succinyl-DOTOPA show significant depolarization at ν_MAS= 2.00 kHz, consistent with simulations showing that intermolecular electron-electron couplings can enhance depolarization [26]. The observation that samples containing succinyl-DOTOPA exhibit greater depolarization (i.e., smaller cross-polarized ¹³C ssNMR signals) with c_e,tot = 30 mM or 45 mM and with ν_MAS= 2.00 kHz or 4.00 kHz is not readily explained. Apparently, the quantitative aspects of depolarization can be sensitive to multiple parameters, including T_1e, ν_MAS, dopant concentration, and dopant structure (e.g., intramolecular electron-electron distances and nitroxide orientations). Reductions in ssNMR signals due to other paramagnetic effects [35] may also contribute to the results in Fig. 5.

At 90 K, T_2e for AMUPol is also significantly larger than T_2e for succinyl-DOTOPA. The relatively rapid electron spin-lattice and spin-spin relaxation in succinyl-DOTOPA at 90 K may reduce the achievable level of saturation of electron spin transitions, leading to the lower DNP-enhanced ¹³C sensitivity for samples containing succinyl-DOTOPA at 90 K under our experimental conditions. We estimate that the Rabi frequency for electron spins in our experiments is approximately 200 kHz (corresponding to 0.5 W of circularly polarized microwave power on a 1 cm² area). Higher microwave powers are likely to produce greater DNP-enhanced ¹³C sensitivity for samples containing succinyl-DOTOPA at 90 K. In earlier experiments on a frozen solution containing a different DOTOPA derivative (DOTOPA-ethanol), we found that DNP-enhanced ¹³C signals became nearly independent of microwave power above 0.8 W at 25 K, but continued to increase beyond 1.2 W at 100 K [25].

Discussion

Data presented above show that succinyl-DOTOPA is a useful DNP dopant for ssNMR measurements on biological or organic molecules in frozen glycerol/water solutions. Of particular interest for experiments in our laboratory [23, 27], DNP build-up times at 25 K are consistently shorter in samples that contain succinyl-DOTOPA than in samples that contain AMUPol (Fig. 3), and cross-polarized ¹³C sensitivity at ν_MAS = 7.00 kHz is greater by factors of 2.3 and 1.4 at electron spin concentrations of 15 and 30 mM, respectively (Fig. 4). The higher sensitivity under these experimental conditions is due to both the smaller T_DNP values and the larger DNP-enhanced ¹³C signals, which reflect a weaker depolarization effect for succinyl-DOTOPA (Fig. 5). The weaker depolarization effect apparently results at least in part from a shorter T_1e (Fig. 6).

At 90 K, samples containing AMUPol show greater cross-polarized ¹³C sensitivity at ν_MAS = 7.00 kHz (Fig. 4). This observation is probably a consequence of lower saturation of electron spin transitions, due to the short T_1e and T_2e values of succinyl-DOTOPA at 90 K (Fig. 6).

At ν_MAS = 7.00 kHz, the highest ¹³C sensitivity in our DNP-enhanced measurements at 25 K (Table 1, 5 mM succinyl-DOTOPA) is greater than the highest ¹³C sensitivity at 90 K (Table 2, 15 mM AMUPol) by a factor of 6.1. This corresponds to a reduction in the time required to achieve a given signal-to-noise ratio by a factor of 37. Because experiments at 25 K require cooling with liquid helium, in practice our experiments at 25 K are typically limited to total measurement times of 12 h or less [23, 27]. Comparable experiments at 90 K would require measurement times of about 18 days. At 25 K and ν_MAS = 7.00 kHz, the ¹³C sensitivity for a sample containing 5 mM succinyl-DOTOPA is greater than the ¹³C sensitivity for a Dy³⁺-doped sample by a factor of 31, meaning that use of DNP can reduce total measurement times by factors greater than 900.

In a previous publication, we described several DOTOPA derivatives (DOTOPA-Ethanol, DOTOPA-NH, DOTOPA-4OH, DOTOPA-3OH-Methoxy) with solubilities slightly above 10 mM in glycerol/water mixtures at pH 7.4 [19]. At 25 K, ν_MAS ≈ 6.7 kHz, and 10 mM dopant concentration, these compounds produced T_DNP values in the 2.6–3.8 s range and ε values in the 94–128 range (for cross-polarized ¹³C ssNMR signals of melittin). These T_DNP and ε values are very similar to those of succinyl-DOTOPA under similar experimental conditions. Thus, the DNP performance of DOTOPA derivatives is largely independent of variations in their chemical structures. The primary advantage of succinyl-DOTOPA is its greater solubility above pH 5 (Fig. S1c). This is an important practical advantage. Experiments with the previously-described DOTOPA derivatives required that they be dissolved in dimethyl sulfoxide (DMSO) before addition to the glycerol/water solution, resulting in final DMSO contents of roughly 2–5%. The previously-described DOTOPA derivatives were also not useful in experiments with relatively low glycerol concentrations [27], due to insufficient solubility.

Finally, to illustrate the applicability to systems of genuine biochemical interest, Fig. 7 shows 1D and 2D ¹³C-¹³C spectra of a frozen solution of the complex of calmodulin with the 27-residue peptide M13 [36], with uniform ¹⁵N and ¹³C labeling of four amino acids in M13. This sample contained 1.5 mM M13, 3.0 mM calmodulin, and 15 mM CaCl₂ and was doped with 10 mM succinyl-DOTOPA. The sample volume was 75 μl, corresponding to 113 nmol of M13. Spectra were recorded at 25 K with ν_MAS = 7.00 kHz. The value of T_DNP for cross-polarized ¹³C signals was determined to be 4.6 s at 25 K. Comparison of signal areas in 1D spectra with and without microwave irradiation (Fig. 7a) indicates an apparent DNP enhancement factor of ε ≈ 140. A 2D spectrum with good signal-to-noise was obtained in 40 min (Fig. 7b).

Fig. 7:

(a) 1D ¹³C NMR spectra at 25 K of the M13/Ca²⁺/calmodulin complex in frozen glycerol/water containing 10 mM succinyl-DOTOPA. Spectra were recorded with microwave irradiation and 4 scans (blue) or without microwave irradiation and 128 scans (red). The sample contains 1.5 mM of the 27-residue M13 peptide, uniformly ¹⁵N,¹³C-labeled at Phe8, Ile9, Ala10, and Val11. The MAS frequency was 7.00 kHz. (b) 2D ¹³C-¹³C NMR spectrum of the same sample at 25 K with microwave irradiation, using a 25 ms spin diffusion mixing period between t₁ and t₂ periods. A 1D slice at 65 ppm is shown above the 2D spectrum. The 2D spectrum was acquired in 40 min, using a 5.9 s recycle delay, 100 complex t₁ points, and a 40.0 μs t₁ increment. Contour levels increase by successive factors of 1.3.

• The nitroxide-based triradical succinyl-DOTOPA is highly soluble in glycerol/water solutions and performs well in low-temperature dynamic nuclear polarization experiments.

• At 25 K, 9.39 T, and 7.00 kHz magic-angle spinning, succinyl-DOTOPA gives higher overall cross-polarized ¹³C solid state NMR sensitivity than does the biradical AMUPol.

• At 90 K, AMUPol gives higher overall sensitivity, attributable to its longer electron spin relaxation times.

• Experimental results for a peptide/calmodulin complex in glycerol/water at 25 K verify the utility of succinyl-DOTOPA.