Acceleration of Natural-Abundance Solid-State MAS NMR Measurements on Bone by Paramagnetic Relaxation from Gadolinium-DTPA

Abstract

Reducing the data collection time without affecting the signal intensity and spectral resolution is one of the major challenges for the widespread application of multidimensional nuclear magnetic resonance (NMR) spectroscopy, especially in experiments conducted on complex heterogeneous biological systems such as bone. In most of these experiments, the NMR data collection time is ultimately governed by the proton spin-lattice relaxation times (T₁). For over two decades, gadolinium(III)-DTPA (Gd-DTPA, DTPA = Diethylenetriamine pentaacetic acid) has been one of the most widely used contrast-enhancement agents in magnetic resonance imaging (MRI). In this study, we demonstrate that Gd-DTPA can also be effectively used to enhance the longitudinal relaxation rates of protons in solid-state NMR experiments conducted on bone without significant line-broadening and chemical-shift-perturbation side effects. Using bovine cortical bone samples incubated in different concentrations of Gd-DTPA complex, the ¹H T₁ values were calculated from data collected by ¹H spin-inversion recovery method detected in natural-abundance ¹³C cross-polarization magic angle spinning (CPMAS) NMR experiments. Our results reveal that the ¹H T₁ values can be successfully reduced by a factor of 3.5 using as low as 10 mM Gd-DTPA without reducing the spectral resolution and thus enabling faster data acquisition of the ¹³C CPMAS spectra. These results obtained from ¹³C-detected CPMAS experiments were further confirmed using ¹H-detected ultrafast MAS experiments on Gd-DTPA doped bone samples. This approach considerably improves the signal-to-noise ratio per unit time of NMR experiments applied to bone samples by reducing the experimental time required to acquire the same number of scans.

1. Introduction

Over the past decade or so, solid-state NMR (ssNMR) spectroscopy has advanced substantially to become the primary probe for the molecular structure and dynamics of a wide variety of systems, ranging from small inorganic and organic materials to large biological macromolecules like crystalline peptides and proteins, amyloid fibrils and membrane-bound systems [1–5]. In addition, numerous recent studies have demonstrated that ssNMR using magic angle spinning (MAS) is particularly well-suited for mapping the structural and dynamic features of heterogeneous non-crystalline biomaterials that are difficult to characterize by solution NMR spectroscopy and diffraction methods, such as bone and related connective tissues [6–17]. It is well known that bone represents one of nature’s most challenging and exciting biological materials [18–20]. This is due to the complex hierarchical structural organization of bone which, at the macroscopic level, is composed of a rigid inorganic mineral (mainly poorly crystalline carbonated hydroxyapatite) and a soft organic matrix (largely type I collagen, along with ~10% of various non-collagenous proteins, lipids, polysaccharides, and citrates) that mutually interact and contribute to bone biomechanical properties and functions [8, 10, 20–22]. Water is the third major component of bone, and is found at multiple locations: it occupies the Ca²⁺ coordination sites within the mineral crystal lattice; it is bound to collagen fibrils; and a significant portion exists as free bulk water that fills the microscopic pores in the calcified matrix [16, 17, 23]. Thus, understanding the molecular interactions among the different constituents and the specific role of each is of prime significance in order to gain high-resolution insights into the structure-function relationship in bone. In this regard, one- and two-dimensional ¹H, ¹³C, and ³¹P MAS ssNMR have been the most widely used methods for the atomic-level structural elucidation of bones [6–17].

In spite of the wealth and diversity of information that it can provide about molecular structure and dynamics, one of the major challenges for the widespread application of ssNMR spectroscopy on macromolecular complex systems is its intrinsic poor sensitivity that originates mainly from the very small spin population difference between magnetically induced nuclear spin states in thermal equilibrium [24, 25]. This issue of low sensitivity is particularly evident in the context of ssNMR spectroscopy of bone due to its complex heterogeneous and low-crystalline structure, which is reflected as low resolution of the obtained NMR line shapes relative to the chemical shift range. Thus, several hours or even days of signal averaging are required to obtain an acceptable signal-to-noise ratio (S/N) in multidimensional experiments that involve detection of dilute nuclei like ¹³C through cross-polarization from the more abundant protons. In these measurements, the total NMR experimental time is determined largely by the inter-scan recycle delay, which in turn depends on the ¹H spin-lattice (or longitudinal) relaxation time (T₁) and is necessarily set long enough (typically ~5 times ¹H T₁) in order to fully relax the proton magnetization to thermal equilibrium and also to minimize the duty cycle of the NMR probe and to avoid probe arcing and sample overheating due to continuous radiofrequency (rf) irradiation [26–28]. Generally, several approaches can be implemented in order to improve the signal intensity in NMR experiments [29, 30]: increasing the sample quantity, isotopic enrichment, using ultra-high field NMR spectrometers, lowering the temperature, or utilizing the cutting-edge dynamic nuclear polarization (DNP)-based experiments.

One alternative simple approach for enhancing the signal, known as paramagnetic relaxation enhancement (PRE), is to introduce a small amount of a paramagnetic reagent into the targeted macromolecular lattice that will enable faster recovery of the ¹H equilibrium magnetization, which allows the use of shorter recycle delays and subsequently faster repetition rates, which result in a higher S/N per unit time [31]. Transition metals (e.g., Cu²⁺, Ni²⁺, Mn²⁺), inner-transition metals (e.g., lanthanides) and the stable nitroxide free radicals that cause considerable PRE have proven to be particularly useful for fulfilling this desired effect [32].

In light of the above, signal enhancement through the longitudinal PRE effects of various paramagnetic dopants have been successfully explored in solution [26, 33–37] and in solid-state NMR studies [27, 38–48] on microcrystalline peptides, proteins, and membrane-bound systems. For instance, Ishii and co-workers have developed the “paramagnetic relaxation–assisted condensed data collection” (PACC) method [27, 40], which relies on a combination of copper(II)-EDTA (Cu-EDTA) doping, very fast MAS (~ 40 kHz), and fast recycling of the low-power rf pulses to significantly accelerate multidimensional NMR experiments on hydrated microcrystalline proteins. Reif and co-workers have demonstrated that the combined use of Cu-EDTA doping and perdeuteration of proteins can induce further signal enhancement in MAS ssNMR experiments without significant loss of resolution [42, 43]. Our group and others have shown that a very low concentration of copper-chelated lipid is sufficient enough to substantially reduce the proton T₁ values for faster data acquisition in a lipid bilayer or isotropic bicelles containing a membrane protein with minimum effect on spectral resolution [38, 39, 44]. In a very recent study, Glaubitz and co-workers have shown that gadolinium paramagnetic complexes can be used to significantly reduce the measurement time in MAS ssNMR experiments conducted on a 7TM membrane protein in lipid bilayers [48].

While the previous ssNMR studies have successfully explored the use of paramagnetic dopants to enhance the longitudinal relaxation in MAS NMR measurements of crystalline macromolecules, the PRE effects induced by these reagents on complex heterogeneous biomolecules like bone have not been investigated in depth. In a recent work [49], we have demonstrated that by immersing bovine cortical bone samples into a 30 mM Cu-EDTA solution, it was feasible to reduce the ¹H T₁ values by a factor of 2.2, thus allowing for faster acquisition of natural-abundance ¹³C CPMAS (cross-polarization magic angle spinning) [50] spectra without significant line-broadening side effects nor chemical shift perturbations. Here, we extend our efforts toward investigating the possibility of further shortening the ¹H T₁ values in natural-abundance ¹³C CPMAS ssNMR spectra of bone via the incorporation of more suitable paramagnetic metal ions, for the ultimate goal of increasing the sensitivity of these NMR spectra. In this context, chelated complexes of the gadolinium ion (Gd³⁺) are well suited to address the desired goal, owing to the presence of seven unpaired electrons in the inner 4f orbitals of Gd³⁺ and its large magnetic moment [51, 52]. By virtue of its isotropic magnetic susceptibility tensor, Gd³⁺ possesses unique paramagnetic properties in that while it causes larger PRE than other lanthanides, it does not cause any perturbations in the NMR chemical shifts [32]. These favorable paramagnetic properties, along with its relatively long electronic spin relaxation times (in the range of nano to micro seconds) due to its symmetric S-state, make Gd³⁺ an attractive choice as a relaxation enhancement agent in carbon-detected MAS ssNMR experiments with magnetization starting on protons. For these reasons, gadolinium-based chelates have been widely used as contrast-enhancement agents in medical magnetic resonance imaging (MRI), as a tool for clinical diagnosis of organ and tissue abnormalities [51]. Among these complexes, [Gadolinium(III)-DTPA]²⁻ (henceforth referred to as Gd-DTPA, DTPA = diethylene triamine pentaacetic acid) stands out as the first contrast agent to be approved for clinical use in 1988 [51]. In this work, we have conducted a comprehensive concentration-dependent study to demonstrate that Gd-DTPA (Figure 1) can also be effectively used to enhance the longitudinal relaxation rates of protons in natural-abundance ¹³C CPMAS ssNMR experiments on bone tissues without significant line-broadening side effects and chemical shift perturbations in the ¹³C NMR line shapes. Using bovine cortical bone samples incubated in solutions with different concentrations of Gd-DTPA complex, the ¹H T₁ values were calculated from a series of data collected by ¹H spin-inversion recovery method detected in ¹³C CPMAS NMR experiments. Our results reveal that the ¹H T₁ time constants can be successfully reduced by a factor of 3.5 using as low as 10 mM Gd-DTPA without any loss of spectral resolution and thus enabling faster data acquisition of the ¹³C CPMAS spectra at natural abundance. We further investigated the combined effect of very fast MAS and Gd-DTPA doping on the sensitivity in proton-detected solid-state NMR experiments applied to the bone samples. Despite the reduced sample quantity used in the ultrafast MAS experiments, we observed about 3-fold gain in overall S/N per unit time of the ¹H MAS NMR spectra in the presence of 10 mM Gd-DTPA at 50 kHz MAS, which illustrates the ability for much faster data acquisition on extremely limited sample quantities.

Figure 1.

Chemical structure of the Gd-DTPA complex used as a paramagnetic dopant in this study to shorten the spin-lattice relaxation times of protons from bone samples.

2. Experimental Details

Sample Preparation

Powdered bovine cortical bone samples, collected from fresh bovine femora, were prepared and stored according to our previously published procedure [49]. Gd–DTPA solutions with different concentrations were prepared by dissolving the required amount of powder gadopentetic acid (Diethylene triamine pentaacetic acid gadolinium(III) dihydrogen salt, Sigma Aldrich, St. Louis, MO, USA) in standard PBS buffer. Bone samples were soaked with Gd–DTPA solutions for about 30 minutes, and filtered for each NMR experiment.

Prior to proceeding with the NMR experiments for this study, it was crucial to confirm that the treatment of bone samples with Gd-DTPA complex would not produce undesired effects on the structure and stability of the mineral crystal lattice in bone due to the possible substitution of Gd³⁺ for Ca²⁺ ions within the mineral crystal lattice and/or in the hydrated surface layer of bone. Theoretically, the formation constants (log K) for the Ca-DTPA and Gd-DTPA complexes are 9.8 and 22.2, respectively [53]; the gadolinium complex is thus favored by over 12 orders of magnitude compared to the calcium complex. Conditional formation constants (pH dependent) of the two DTPA complexes are also expected to differ by the same order, and therefore the probability of exchanging gadolinium for calcium in the Gd-DTPA complex is practically null. This argument was confirmed experimentally by measuring the Ca²⁺ concentration in our bone samples without and with Gd–DTPA using inductively coupled plasma – optical emission spectroscopy (ICP–OES). The results show that the Ca²⁺ concentration in bone does not practically change in the presence of 60 mM Gd–DTPA ([Ca²⁺] = 144 ± 39 μg/L with 60 mM Gd–DTPA, and 154 ± 58 μg/L without Gd–DTPA). The addition of the Gd–DTPA complex to bone therefore does not cause any substitution of Gd³⁺ for Ca²⁺.

Solid-State NMR Experiments

All ¹³C-detected solid-state NMR experiments were conducted at 11.75 T on an Agilent VNMRJ 500 MHz NMR spectrometer operating at a resonance frequency of 125.6 MHz for ¹³C and 499.44 MHz for ¹H, and using a 3.2-mm triple-resonance MAS probe. The ¹³C MAS spectra were acquired under 10 kHz MAS conditions at 25 °C using linear ramped cross-polarization (CP) pulse sequence [54] with a ¹H excitation pulse length of 2.5 μs and a CP contact time of 2 ms. For each experiment, 10,000 scans were accumulated, with recycle delays ranging from 1 to 3.5 s that were optimized for each sample. To decouple protons during signal acquisition, a 65 kHz two-pulse phase-modulation (TPPM) scheme was applied [55]. The ¹³C NMR chemical shifts were referenced with respect to tetramethylsilane (TMS) using adamantane as a secondary external standard. The ¹H T₁ values were calculated from a series of data collected by ¹³C-detected ramped CPMAS inversion recovery experiments, in which a ¹H π-pulse followed by an inversion recovery delay were introduced prior to the CP sequence. Seven data points with increasing inversion recovery delays were recorded for each sample. The recycle delays were set to 4 s for samples containing Gd–DTPA and 6 s for those without Gd–DTPA, and 2000 scans were accumulated for each inversion recovery data point. The integral signal intensities were measured for peaks in the aliphatic (10–70 ppm) and the carboxyl/carbonyl (165–185 ppm) regions to calculate the ¹H T₁ values for the individual amino acid residues. A mono-exponential function was used to calculate and analyze the ¹H T₁ values from the inversion recovery data using a script developed in the Ramamoorthy laboratory. The average of the ¹H T₁ values determined from all peaks was used as the ¹H T₁ value of the sample of interest.

All ¹H-detected MAS NMR experiments were acquired at 14.1 T (599.8 MHz ¹H resonance frequency) on an Agilent/Varian VNMRS 600 MHz solid-state NMR spectrometer, using a 1.2-mm triple-resonance ultrafast MAS probe at 25 and 50 kHz spinning speeds with a 1.2 μs 90° ¹H excitation pulse. The ¹H T₁ values for the bone samples were calculated from a series of data collected by ¹H-detected MAS inversion recovery experiments. The ¹H single-pulse and rotor-synchronized spin-echo (with a 0.6 ms inter-pulse delay) spectra were recorded with 64 scans each and an optimized recycle delay in the range of 1.2–5 s depending on the sample.

3. Results and Discussion

In this study, the concentration-dependent paramagnetic relaxation effects of Gd³⁺ ions on the proton intrinsic spin-lattice relaxation times measured in MAS NMR experiments on fresh bone samples are investigated for the purpose of accelerating the ssNMR data acquisition rate using the optimal dopant concentration. The ¹H T₁ relaxation times as a function of Gd-DTPA concentration were measured via ¹H spin inversion recovery method detected in ¹H to ¹³C CPMAS NMR experiments (Supplementary Information). Figure 2 shows the variation of the average ¹H T₁ for the bone samples as a function of the Gd-DTPA concentration. Overall, we observe that increasing the concentration of Gd-DTPA yields a gradually shorter T₁ time constants, and thereby faster relaxation rates. However, the T₁ relaxation times do not exhibit linear dependence on the Gd-DTPA concentration. A considerable reduction in ¹H T₁ is observed at lower Gd-DTPA concentrations (5 and 10 mM), while a much slower decrease in T₁ is achieved at higher Gd-DTPA concentrations. Accordingly, a 3.5-fold shortening in the ¹H T₁ time is observed with the addition of 10 mM Gd-DTPA. Further increase in the Gd-DTPA concentration to 30 mM and then to 60 mM results in only an additional minor decrease (< 10 %) in the ¹H T₁ time, which implies that the longitudinal PRE becomes almost saturated once this range of high concentrations is reached. This can be explained by realizing that ¹H PREs are mediated by a flexible ¹H–¹H spin-diffusion mechanism: the unpaired electrons of Gd³⁺ ions interact strongly with the close ¹H nuclei in the sample, which will significantly speed up the longitudinal relaxation of these nearby protons. This effect is further transferred to remote protons in the organic matrix via a ¹H-¹H spin-diffusion mechanism; this efficient diffusion process, by virtue of the high gyromagnetic ratio of protons, results in a uniform enhancement of the longitudinal relaxation rates for all protons in the sample [56]. At higher Gd-DTPA concentrations, it is most likely that this spin-diffusion process does not transfer the longitudinal PREs efficiently and uniformly throughout the organic-mineral interface in bones. It is therefore expected that since Gd-DTPA is hydrophilic, the ¹H magnetization is recovered more slowly at higher Gd-DTPA concentrations by paramagnetic longitudinal relaxation around the water-accessible organic-mineral interface.

Figure 2.

Average ¹H spin-lattice relaxation time (T₁) of powdered cortical bone as a function of Gd-DTPA concentration. The ¹H T₁ values were determined from the ¹H spin-inversion-recovery method detected in ¹³C ramped-CPMAS NMR experiments, and the reported errors were estimated from the best-fitting of experimental data.

Meanwhile, it is instructive to compare the present results with those obtained in our previous study on the longitudinal PRE effect of Cu-EDTA on bone samples [49], in which a 2.2-fold enhancement in the ¹H relaxation rates was observed when bovine cortical bone samples were doped with 30 mM Cu-EDTA. Gd-DTPA is obviously a more efficient relaxation enhancement dopant than Cu-EDTA for bone. The better performance of Gd-DTPA might be explained in terms of the differences/similarities in the magnetic and structural properties exhibited by the two paramagnetic ions and their complexes. Gd³⁺ ion has seven unpaired electrons whereas Cu²⁺ possesses only one, with the unpaired electron spins of both ions relaxing relatively slowly (on the nanosecond time scale). Moreover, the magnetic susceptibility tensor of Gd³⁺ is totally isotropic, whereas that of Cu²⁺ is not totally isotropic [32]. Gd-DTPA is a nine-coordinate complex in which eight binding sites of the inner coordination sphere around the Gd center are occupied by the (DTPA)⁵⁻ ligand while the ninth coordination site is occupied by a nearby water molecule [51]. Cu-EDTA, on the other hand, is a six-coordinate complex in which the (EDTA)⁴⁻ ligand occupies all six binding sites at the Cu center, so that water is excluded as a first-coordination sphere ligand [52]. The presence of one water molecule directly coordinated to Gd, and the lack of it around Cu, might be one of the factors that contribute to the more efficient longitudinal PRE of Gd-DTPA over Cu-EDTA. It is quite plausible that the ¹H-¹H spin diffusion process is facilitated in the case of Gd-DTPA through the directly bonded water molecules at the bone mineral surface, and subsequently transferred to the water-bound collagen via the free bulk water within the bone pores. A second contributing factor is the larger size (mass) of Gd-DTPA molecules compared to Cu-EDTA. The larger molecular size of the Gd³⁺ complex reduces its mobility, which leads to enhanced relaxation rates through the increase of the rotational correlation time, which is one of the parameters that affect PRE rates.

Shown in Figure 3 are the natural-abundance ¹³C CPMAS NMR spectra for the doped bone samples in comparison with the undoped bone. These spectra were all collected under identical experimental conditions, except for the recycle delay, which was set to ~ 5 × ¹H-T₁ average value calculated from the spin inversion recovery experiments for each sample. It is evident that the recycle delay can be reduced considerably without loss of spectral resolution, resulting in a better S/N per unit time by shortening the data collection time required to acquire the same number of scans. In the presence of 5 mM Gd-DTPA, a 2-fold faster data acquisition is obtained, whereas a 3.5-fold faster acquisition is obtained upon the addition of 10 mM Gd-DTPA and more, which can achieve a two-fold improvement in the S/N per unit instrumental time. Not surprisingly, the NMR line shapes in all these spectra are dominated by carbon resonances of the amino-acid residues in type I collagen, the major organic component in bones. Most of these resonances can be effortlessly identified, and hence assigned based on previously reported values [12, 15, 57], as shown in Fig. 4a for the ¹³C CPMAS spectrum of untreated fresh bone. It is apparent from Fig. 3 that no major perturbations in the ¹³C NMR chemical shifts are caused upon addition of various amounts of Gd–DTPA. It is well known that in addition to PRE, paramagnetic ions can cause the so-called pseudo-contact shifts (PCSs), which are reflected as changes in NMR chemical shift values of the nuclear spins that sense the paramagnetic center. PCSs originate mainly from through-space dipolar interactions between the nuclear spins and the relaxing unpaired electrons, and they depend on the anisotropy of the magnetic susceptibility tensor for the paramagnetic ion, which describes the interaction of the paramagnetic dipole moment with the external magnetic field. Specifically, PCSs are proportional to the anisotropy of this tensor but are independent of its isotropic component [32]. In the case of Gd³⁺, no PCSs are induced in the NMR line shapes due to the isotropic nature of its magnetic susceptibility tensor. Notably, this confirms the assertion that no undesired changes has occurred in the secondary structure and conformational dynamics of the protein and the protein-mineral surface layer in the investigated bone samples due to the addition of Gd-DTPA.

Figure 3.

. Proton-decoupled ¹³C CPMAS spectra recorded at 10 kHz MAS for (a) undoped bone, and (b-e) bone doped in different concentrations of Gd-DTPA (as indicated for each spectrum). All spectra were recorded under identical experimental conditions, except for the recycle delays that were set to ~ 5 × ¹H T₁ in an experimental time of: (a) 9.8 h, (b) 4.9 h and (c, d, e) 2.8 h.

Figure 4.

(a) The ¹³C NMR chemical shift assignments of various amino acids in the ¹³C CPMAS spectrum of fresh cortical bovine bone. Ala, alanine; Leu, leucine; Pro, proline; Glu, glutamic acid; Hyp, hydroxyproline; Gly, glycine; Arg, arginine; CO, carbonyl. (b) Spin-lattice ¹H relaxation times (¹H T₁) of the various residues in powdered cortical bone doped in different concentrations of Gd-DTPA. The T₁ values were determined from ¹H-spin-inversion recovery detected in ¹³C CPMAS NMR experiments, and the reported errors were estimated from the best-fitting of experimental data. A, alanine; L, leucine; P, proline; E, glutamic acid; O, hydroxyproline; G, glycine; CO, carbonyl. The signals from (Pα, Oα) and (Oδ, Eα) overlap in the ¹³C NMR spectrum.

To gain a deeper insight into the site-specific paramagnetic effect of Gd-DTPA on longitudinal relaxation, the calculated ¹H T₁ relaxation times for the individual peaks (from amino acids) are shown in Figure 4b for the doped bone samples as well as for the undoped one. In the absence of Gd-DTPA, it is observed that protons located at different positions in various amino acids exhibit different T₁ relaxation times. Most likely, these differences in the ¹H T₁ relaxation times are due to different proton-proton dipolar interactions arising from internal structural fluctuation and molecular mobility. Despite that the general trend of relaxation times as a function of the amino-acid residues is largely retained upon incubation with Gd-DTPA, the site-specific ¹H T₁ relaxation times are gradually shortened with increasing dopant concentration. For most of the amino acids, however, the ¹H T₁ values are only marginally shortened in the presence of Gd-DTPA with concentrations higher than 10 mM. Moreover, a thorough observation of the measured ¹H T₁ delays reveals that the reduction in T₁ times is not homogeneous among the ensemble of amino acid residues. Since paramagnetic relaxation enhancement by Gd³⁺ (and other ions with long electronic lifetimes) is modulated by dipolar coupling between the unpaired electronic spins and the nuclear spin, and hence is proportional to the inverse sixth power of the metal-nucleus distance [34, 58], then there exists a correlation between T₁ shortening levels and residue proximity to the paramagnetic Gd³⁺ ion, which obviously implies that collagen residues located nearby the Gd³⁺ ion will exhibit more shortening in their ¹H T₁ values than the remote ones.

To examine the effect of Gd–DTPA doping on spectral resolution, the ¹³C line widths, measured as the full width at half-maximum (FWHM) of the amino acid signals, were plotted as a function of Gd–DTPA concentration and are shown in Figure 5. Overall, we observe that the average ¹³C line width increases with increasing dopant concentration. However, only a marginal broadening in the line shapes is observed especially in the low-concentration range, as the average line width increases by <5% upon addition of 10 mM Gd–DTPA, and it increases by ~ 17% upon doping with 60 mM Gd–DTPA, which can still be considered a modest broadening. This indicates that Gd–DTPA concentrations up to 10 mM cause no pronounced effect on the ¹³C line width, whereas higher concentrations induce only a modest broadening in the ¹³C line shapes. These modest line broadenings can be well explained in terms of the transverse (T₂) PRE effects of Gd³⁺, which is known to cause such broadenings through ¹H transverse relaxation enhancements [26]. Our experiments, however, have been designed to start with proton excitation and end with carbon detection. This implies that since the dipolar contribution of the paramagnetic transverse relaxation enhancement rates (R₂) depends on the square of gyromagnetic ratio (γ) of the observed nucleus [26, 58], then the expected line broadenings caused by paramagnetic transverse relaxation are much less substantial for the ¹³C line shapes compared to those in the ¹H line shapes (by a factor of ~1/16 since γ_H ≈ 4 γ_C). Moreover, the ¹³C-¹³C spin-diffusion process is much less efficient than in the case of protons (due also to the smaller γ_C compared to γ_H), which slows down and restricts the homogeneous transfer of ¹³C magnetization throughout the entire macromolecular lattice [48].

Figure 5.

(a) Carbon-13 full-width at half-maximum (FWHM) values obtained from ¹³C ramped-CPMAS NMR spectra of bone samples under 10 kHz MAS conditions for powdered bovine cortical bone samples without Gd–DTPA (black) and with various concentrations of Gd-DTPA (red, blue, pink, green). A, alanine; L, leucine; P, proline; E, glutamic acid; O, hydroxyproline; G, glycine; CO, carbonyl. The signals from (Pα, Oα) and (Oδ, Eα) overlap in the ¹³C NMR spectrum. (b) Gd–DTPA concentration-dependence of the average ¹³C FWHM values calculated from (a).

For the purpose of obtaining ¹H MAS NMR spectra of the bone samples at the highest possible resolution, we further investigated the effect of Gd-DTPA doping on bone NMR spectra under very fast MAS conditions. With such very high spinning frequencies, it becomes feasible to significantly reduce the line broadenings caused by the strong homonuclear ¹H-¹H dipolar couplings in the ¹H NMR line shapes. However, the issue of poor sensitivity arises in this case due to the very limited sample amounts required by the tiny rotors (~ 2.5 – 3 mg for the bone specimens used in this study). Thus, the use of Gd-DTPA doping in conjunction with very fast MAS rates is beneficial to increase the sensitivity of the NMR spectra of bones. Figure 6 shows the ¹H T₁ time constants measured as a function of Gd-DTPA concentration at two different spinning frequencies (25 and 50 kHz) using the ¹H spin-inversion-recovery method in ¹H-detected experiments. At the magnetic field of 14.1 T, the average T₁ values are slightly shorter at 50 kHz than at 25 kHz for all of the dopant concentrations used. In accord with the ¹H T₁ data obtained from the ¹³C-detected spectra, the 10 mM Gd-dTPA concentration is the optimum concentration that gives the maximum T₁ reduction in the ¹H-detected experiments, where a 4-fold reduction is obtained at 50 kHz MAS. Compared to the 10 mM concentration, higher Gd-DTPA concentrations produced longer ¹H T₁ relaxation times. This may be attributed to an increase in the sample temperature due to a combination of high dopant concentrations and very fast spinning speeds.

Figure 6.

Average ¹H spin-lattice relaxation time (T₁) of powdered cortical bone as a function of Gd-DTPA concentration. The ¹H T₁ values were determined by ¹H spin-inversion-recovery methods in ¹H-detected NMR experiments with 25 (red) and 50 (blue) kHz MAS rates. The reported errors were estimated from the best-fitting of experimental data.

Using optimized recycle delays obtained from the measured ¹H T₁ values for each of the samples and with 25 and 50 kHz MAS rates, the 1D ¹H single-pulse and spin-echo MAS NMR spectra were acquired for the doped and undoped bone samples (Figure 7). The ¹H single-pulse spectra of bone exhibit three main peaks, which are dominated by the intense signal from water centered at ~5 ppm. In addition to this intense water signal, the other broader peaks in the ¹H single-pulse NMR spectra originate from overlapping signals that correspond to protons in the organic and mineral components of bone. It is obvious that the sensitivity is considerably improved in the spectra acquired with higher spinning speed of 50 kHz compared to those at 25 kHz. The sizeable gain in signal intensity due to Gd-DTPA doping is also obvious in the single-pulse spectra at both spinning speeds. Specifically, among the doped samples, the one doped in 10 mM GD-DTPA gives spectra with the largest increase in S/N per unit time (~ 2-fold increase) compared to the undoped one. Due to different transverse dephasing rates of the signals, better improvement in both resolution and sensitivity can be achieved in the spin-echo NMR spectra, which exhibit more resolved peaks compared to the single-pulse spectra. In the spin-echo spectra acquired with 50 kHz MAS rate, for example, more than ~ 3-fold gain in overall S/N is achieved for bone doped in 10 mM Gd-DTPA. All in all, our ¹H-detected MAS NMR results demonstrate the additional benefits offered by paramagnetic doping, when combined with ultrafast spinning rates, for the sensitivity enhancement in biomolecular NMR studies of complex heterogeneous solids, particularly for samples that are only available in limited amounts.

Figure 7.

¹H MAS NMR spectra of bovine cortical bone samples at MAS rates of (a) 25 kHz and (b) 50 kHz. The experiments were acquired with either a ¹H one-pulse excitation (top row) or a rotor-synchronized spin echo (bottom row), using a 1.2 mm MAS HXY probe with a 1.2 μs π/2¹H-pulse. For the spin-echo experiments, an interpulse delay of 0.6 ms was used. All spectra were recorded with 64 transients, and the recycle delays were set to ~ 5 × ¹H T₁ for each sample as follows: 5 s for the undoped sample; 1.2 s for samples doped in 5 mM and 10 mM Gd-DTPA; and 2 s for samples doped in 30 mM and 60 mM Gd-DTPA.

4. Conclusions

Paramagnetic doping is arguably the simplest and most practical approach for sensitivity enhancement in multidimensional solid-state NMR spectroscopy. This robust approach has been applied here on bone samples, and the obtained results demonstrate that the widely used MRI contrasting agent Gd-DTPA is also beneficial for enhancing the proton longitudinal relaxation rates, and thereby reducing the measurement time, with minimal line broadenings and no chemical shift perturbations in ¹³C CPMAS ssNMR of heterogeneous complex systems like bones. Doping of finely powdered bones in as low as 10 mM Gd-DTPA solution results in 3.5-fold enhancement in the intrinsic spin-lattice relaxation rates of protons in carbon-detected experiments, which is approximately 1.6 times faster compared to bone samples doped in 30 mM Cu-EDTA. Under the experimental conditions used in this study, the S/N improvement per unit time afforded by Gd-DTPA is ~ 1.3 times better than for Cu-EDTA, while a 3-fold lower concentration is needed. This lower dopant concentration is favorable for reducing complications that may arise from sample heating and undesired dopant-bone interactions. In addition to Gd-DTPA, a number of gadolinium-based MRI contrast-enhancement agents have been successfully used in NMR for the structural elucidation of biomolecular systems [32, 34, 36, 48]. Larger paramagnetic relaxation enhancements than with Gd-DTPA can be achieved in future studies on bone samples using other paramagnetic relaxation agents to further maximize signal enhancements of the bone organic matrix resonances. We believe that our results presented in this study would facilitate multidimensional solid-state NMR experiments for high-resolution structural and dynamic studies of various biomolecules in bone and other challenging macromolecular systems.

Highlights.

• Gd-DTPA was successfully used to speed up solid-state NMR measurements on bone.

• ¹H-detected ultrafast and ¹³C CP MAS were used to evaluate longitudinal PREs.

• A 3.5-fold reduction in NMR time was obtained using as low as 10 mM Gd-DTPA.

• No line-broadening or peak shift was observed on bone in the presence of Gd-DTPA.