High Resolution ¹⁷O NMR Spectroscopy of Structural Water

Abstract

The importance of studying site-specific interactions of structurally similar water molecules in complex systems is well known. We demonstrate the ability to resolve four distinct bound water environments within the crystal structure of lanthanum magnesium nitrate hydrate via ¹⁷O solid state nuclear magnetic resonance (NMR) spectroscopy. The approach utilizes high-resolution multi-dimensional ¹⁷O NMR experiments at high magnetic fields (18.8 – 35.2 T) where each individual water environment was resolved. The quadrupolar coupling constants and asymmetry parameters of the ¹⁷O of each water were determined to be between 6.6 and 7.1 MHz, and 0.83 and 0.90. The resolution of the four unique, yet similar, structural waters within a hydrated crystal via ¹⁷O NMR spectroscopy demonstrates the ability to decipher the unique electronic environment of structural water within a single hydrated crystal structure.

Graphical Abstract

INTRODUCTION

The influence of water on the structure, stability, function, and dynamics of complex biological and inorganic systems underlines the importance of understanding the detailed structure of individual water molecules in such systems.¹ For example, it is well known that water is intimately involved in the intra- and intermolecular hydrogen bonding of proteins, membranes, and nucleic acids^2–4 and is therefore important in determining the secondary and tertiary structure of these macromolecular systems. Recently, in a combined cryo-electron microscopy and magic-angle spinning (MAS) NMR study of the structure of amyloid fibrils from our group, we observed a water bilayer that is thought to be essential to stabilizing the structure of the paired twisted β-sheets.^5–10 In addition to the impact of water on the structure of biochemical systems water is known to influence other systems,^11–19 one such system in particular is the formation of organic rosette nanotubes.^18–19 These and other results have stimulated the study of the structure of water in multiple different systems using solid-state NMR^{10, 19–27} with many studies focusing on the mobile waters of hydration. However, the resolution of different water molecules, and therefore the site-specific study of these waters, has proved challenging because of the small chemical shift range present in ¹H NMR spectra which is the primary probe of the water molecules.^{10, 20–25} In contrast, the sizeable chemical shift range (~2,000 ppm) and quadrupolar nature of ¹⁷O (I = 5/2) makes it attractive for the study of intra- and intermolecular hydrogen bonding interactions involving water. However, ¹⁷O NMR is inherently insensitive due to the low natural abundance (~0.037%) and small gyromagnetic ratio (~1/7^th of ¹H). In addition, ¹⁷O NMR spectra are broadened by the second-order quadrupolar interaction which is not averaged by MAS.²⁸ Despite these shortcomings, ¹⁷O has been used extensively to study biological and inorganic systems, by utilizing isotopic enrichment,^29–33 and high magnetic field strengths (≥ 16.4 T).^34–39 Approaches improving the resolution of ¹⁷O MAS NMR spectra by attenuating the second-order quadrupolar interaction, such as multiple quantum MAS (MQMAS)⁴⁰ and satellite transition MAS (STMAS),⁴¹ have been shown to yield well-resolved isotropic quadrupolar spectra. And, despite the poor efficiency (~5%) of these techniques in the presence of large quadrupolar coupling constants (>5 MHz),^42–43 they are used to successfully study ¹⁷O enriched biological samples and produced promising results.^44–47 For example, the increased resolution that is present in the MQMAS experiment allowed studies of inorganic glasses and minerals that could not be achieved with traditional MAS experiments.^{32–33, 48–58}

In recent studies of H₂¹⁷O structurally bound to organic and inorganic crystals,^{11, 59–60} we observed ¹⁷O chemical shifts dispersed over ~50 ppm. This result suggests that ¹⁷O NMR can be used to distinguish bound water in complex biological and inorganic systems via the dispersion of oxygen chemical shifts. Here we report ¹⁷O spectra of lanthanum magnesium nitrate hydrate [La₂Mg₃(NO₃)₁₂ • 24H₂¹⁷O] (LMN), a hydrated crystal containing four distinct water environments (each site comprised of six individual equivalent waters),⁶¹ as a model system that demonstrates this possibility. As illustrated in Figure 1, each of the three Mg²⁺ ions is coordinated to six water molecules and the remaining waters in the crystal exist in a layer between the lanthanum nitrate and one of the magnesium hydrate subunits. The four crystallographically distinct waters are indicated on the molecular unit in Figure 1(a) with the average O-H bond distance and ∠HOH angle for each water site given in Figure 1(b).

Figure 1:

LMN crystal structure (a) showing the four crystallographically distinct water environments labeled as A-D, and corresponding (b) interatomic distances and angles determined by neutron diffraction.⁶¹

EXPERIMENTAL

a.). Materials and Synthesis:

Lanthanum magnesium nitrate hydrate, La₂Mg₃(NO₃)₁₂ • 24H₂¹⁷O, samples were synthesized by dissolving lanthanum nitrate hexahydrate, La(NO₃)₃ • 6H₂O (Sigma Aldrich (SA), St. Louis, MO), and magnesium nitrate hexahydrate, Mg(NO₃)₂ • 6H₂O (SA), in excess ¹⁷O labeled water (90%-H₂¹⁷O, Cambridge Isotopes Laboratories (CIL), Andover, MA), and recrystallizing in a sealed eppendorf tube over the course of two to 14 days. The crystals were then air-dried and ground into a fine powder using an agate mortar and pestle.

b.). Nuclear Magnetic Resonance Spectroscopy

Oxygen-17 NMR experiments were performed at 18.8 (Francis Bitter Magnet Laboratory – Massachusetts Institute of Technology, FBML-MIT), 18.8 (National High Magnetic Field Laboratory, NHMFL), and 21.1 (FBML-MIT) T using a Bruker Avance II or III spectrometer. Oxygen-17 NMR experiments were also performed at 35.2 T on the series-connected hybrid magnet at the NHFML using a Bruker Avance NEO console and a single-resonance 3.2 mm MAS probe designed and constructed at the NHMFL.⁶² A recycle delay of between 0.5 and 1 second was used for all ¹⁷O experiments, unless otherwise noted. Between 2,400 and 8,192 scans with γB₁/2π (¹⁷O) = 27.7 to 140 kHz were utilized for MAS ¹⁷O NMR experiments. A spinning frequency (ω_R/2π) of 20 kHz or 23 kHz was used for ¹⁷O MAS experiments at 18.8, and 21.1 T, respectively. One-dimensional MAS NMR experiments were acquired using a Hahn-echo (π/2–τ–acquire) pulse sequence.

Two-dimensional ¹⁷O shifted-echo triple quantum magic-angle spinning (3QMAS) spectra were acquired at 18.8, 21.1, and 35.2 T with 128, 92, 44 (80) rotor-synchronized t₁ increments with an increment of 62.5, 43.48, and 100 μs with 2,016, 2,400, and 1,152 (960) scans and ω_R/2π = 16, 23, and 10 kHz. The 3QMAS experiments at 18.8 and 35.2 T were performed with 3Q excitation and conversion pulses of 3 and 1 μs (γB₁/2π = 100 kHz), and π/2 and π pulses of 2.5 and 5 μs (γB₁/2π = 33.3 kHz). The 3QMAS experiment at 21.1 T were performed with 3Q excitation and conversion pulses of 4.6 and 2.8 μs (γB₁/2π = 27.7 kHz), and π/2 and π pulses of 3 and 6 μs (γB₁/2π = 27.7 kHz). Spectra were referenced to liquid water, ¹⁷O (18% H₂¹⁷O, 0 ppm), via the substitution method.⁶³

c.). Spectral Processing and Simulations

All spectra were processed by RNMR (Dr. D. Ruben, FBML-MIT), TOPSPIN (Bruker, Billerica, USA), or MATLAB (MathWorks, Natick, MA, USA) with between 10 and 500 Hz of exponential apodization. ¹⁷O spectral simulations were performed using either the WSolids,⁶⁴ DMFit,⁶⁵ or SIMPSON⁶⁶ software packages.

RESULTS AND DISCUSSION

At high magnetic fields (18.8, 21.1, and 35.2 T⁶²) multiple unique bound water environments were identified by ¹⁷O NMR spectroscopy of a crystalline sample of LMN prepared from H₂¹⁷O. The 3QMAS spectra illustrated in Figure 2 reveal four resolved peaks in the 2D landscape corresponding to four unique bound waters within the LMN crystals. The projection of the isotropic dimension at 18.8 T was found to have an integrated intensity ratio of 1:1:1:1 for the four environments (Figure 3) suggesting that they are the same bound water species that were identified by neutron diffraction.⁶¹ The observed isotropic frequency of the MQMAS spectrum is a linear combination of the isotropic chemical shift and the 2^nd order quadrupolar shift. However, while the chemical shifts increase as ω₀/2π, the 2^nd order quadrupolar shifts²⁸ decrease as (1/(ω₀/2π)²) resulting in a change in the relative isotropic frequencies in spectra recorded at different static magnetic fields (Figure 4). There are also differences in the breadth of the peaks in the isotropic dimension that could arise from differences in structural heterogeneity or T₂ relaxation arising from the molecular motions (twofold flips) of the different environments.

Figure 2:

Oxygen-17 2D MQMAS NMR spectra at (a) 35.2 T (ω_0H/2π = 1500 MHz) and (b) 18.8 T (ω_0H/2π = 800 MHz). Four distinct water sites are resolved in each spectrum. The projection of the isotropic dimension is shown to the right of each spectrum.

Figure 3:

Experimental and fit Isotropic projections of ¹⁷O 2D MQMAS spectra at (a) 35.2 T (ω_0H/2π = 1500 MHz) and (b) 18.8 T (ω_0H/2π = 800 MHz). Four distinct water environments are resolved with isotropic frequencies of (a) 0.4 ± 1, 3.1 ± 1, 5.6 ± 1, and 7.8 ± 1 ppm and (b) 13.1 ± 1, 16.0 ± 1, 17.1 ± 1, and 19.9 ± 1 ppm. The Lorenztian fits of each environment indicate a ratio of populations of 1:1:1:1.

Figure 4:

Magnetic field dependence (B_o = 18.1 to 35.2 T) of the observed ¹⁷O isotropic frequency shown using the (a) observed isotropic frequency from the 2D MQMAS spectra, and (b) using the equation${\delta }_{iso}={\delta }_{3Q}+\frac{3}{850}\frac{P_Q^2}{{\nu }_0^2}x{10}^{-6},$where$P_Q=C_Q\sqrt{1+\frac{{\eta }^2}{3}}$and v₀ is the Larmor frequency of ¹⁷O and is calculated from the values in Table 1. A projection of the isotropic dimension at each field is shown to the left of the isotropic frequencies for each field in (a). The dashed lines (Black, Red, Blue and Purple) are linear fitted lines to the isotropic frequencies of each site.

The resolution of multiple water environments with such a small chemical shift difference demonstrates the ability to use ¹⁷O NMR to study bound water in complex systems where the water environments are expected to be similar. The MQMAS spectrum at 21.1 T, Figure S1, shows only three resolved peaks. The field dependence of the 2^nd order quadrupolar shift that causes the two center peaks that are resolved at 18.8 and 35.2 T to be nearly indistinguishable at 21.1 T (Figure 4). The determined ratio of the three peaks is 1:2:1, which is in line with the expected ratio considering that the center peak is two unresolved water environments. Using iterative spectral simulations of the 3Q-filtered MAS lineshape of each of the four peaks, the electric field gradient (EFG) and chemical shift anisotropy (CSA) tensor parameters were determined and are given in Table 1.

Table 1:

¹⁷O EFG and CSA tensor parameters determined from MAS and MQMAS NMR

17O site	CQ (± 0.3) / MHz	ηQ (± 0.15)	δiso / ppm	Ω (± 30) / ppm	κ (± 0.5)	δ3Q (±1) / ppm at 18.8 T	δ3Q (±1) / ppm at 21.1 T	δ3Q (±1) / ppm at 35.2 T
1	7.1	0.85	−5 ± 2	50	−1	13.1	9.1	0.4
2	7.1	0.83	−2 ± 2	70	−1	17.1	n.r.	3.1
3	6.6	0.83	0 ± 2	70	−1	16.0	n.r.	5.6
4	6.8	0.9	3 ± 2	80	−1	19.9	16.5	7.8

n.r. – not resolved, one peak at 12.8 ±1 ppm.

The breadth and asymmetry of the second-order quadrupolar lineshape of the central transition of half-integer quadrupolar nuclei, as measured by the quadrupole coupling constant, C_Q, and the asymmetry parameter, η_Q, depend on the local structure of the oxygen nucleus. Therefore, the differences in the EFG and CSA tensor parameters could yield a better understanding of the structure of each of the water environments.

The quadrupole coupling constants were found to be between 6.6 and 7.1 MHz. Although previous work has shown that the asymmetry parameter is a sensitive indicator of the ∠HOH bond angle,⁶⁷ the fast limit twofold flipping of the bound water in the fast limit at room temperature averages the EFG tensor,^{11, 59–60} rendering the ability to use the asymmetry parameter as a target for assignment of the water environments is ill-advised. The experimentally determined asymmetry parameters for the four water environments in LMN were found to be between 0.83 and 0.9. The differences found in the EFG tensors likely reflect both differences in the local structure and the room temperature dynamics of the bound waters. Both the EFG and CSA tensor values were found to be in good agreement with previous results for bound water within hydrated crystals.^{11, 59–60} While the one-dimensional MAS NMR spectra at 18.8 and 21.1 T (Figure 5) indicated the presence of multiple unique oxygen environments, the 2D MQMAS experiments were necessary to identify and characterize the four oxygen environments (Figures 2 and 5). Although inclusion of the ¹⁷O chemical shift parameters did not affect the 1D MAS NMR lineshape, it was necessary to include the CSA tensor parameters (Table 1) for the simulations of the 3QMAS lineshapes at 35.2 T (Figure 6). The 3QMAS dimension was analyzed from the projection of the direct dimension of the 2D MQMAS experiment.

Figure 5:

Experimental (solid) and simulated (dashed) ¹⁷O MAS NMR spectra at (a) 21.1 T (ω_0H/2π = 900 MHz), and (b) 18.8 T (ω_0H/2π = 800 MHz) with ω_R/2π = 23, and 20 kHz, respectively. EFG tensor parameters for the spectral simulations are given in Table 1.

Figure 6:

Projections of the 3Q-filtered lineshape of each ¹⁷O resonance from Figure 2 at (a) 35.2 T (ω_0H/2π = 1500 MHz) and (b) 18.8 T (ω_0H/2π = 800 MHz) with spectral simulations displayed as dashed lines. The EFG and CSA tensor parameters of the simulations are given in Table 1.

Unique assignments of the NMR resolved peaks to the crystal structure water environments were not performed in this study due to the complex dynamics of such environments.^{11, 59–60, 68–70} The librational modes of the bound water sites have been shown to average the EFG tensors and therefore complicate the relationship between local and hydrogen bonding structure and the experimentally determined EFG tensors at room temperature. While a dependence of the isotropic chemical shift and the OH bond distance and ∠HOH bond angle was reported in crystalline amino acid and dipeptide hydrates,⁶⁰ a study of crystalline inorganic hydrates did not demonstrate the same dependence.¹¹ Either correlation spectroscopy between the ¹⁷O and nearby nuclei or low temperature NMR could be utilized to assist in determining which NMR environments correspond to their neutron diffraction counterparts.

CONCLUSION

Notwithstanding a unique assignment of the H₂¹⁷O lines, the resolution of multiple structurally similar environments indicates the benefit of high magnetic fields for the study of structurally important bound water. Due to the nature of the crystal structure studied here, correlation spectroscopy (e.g., ¹³C-¹⁷O, ¹⁵N-¹⁷O)^{47, 71–74} was not performed to better identify the oxygen environments that were resolved. However for complex biological systems, such as GNNQQNY or TTR_105–115,^6–9 where more accessible correlation spectroscopy is available an identification of resolved structural water molecules may be possible. The ability to use high resolution ¹⁷O MAS NMR to study unique, yet similar, structural waters in a single inorganic crystal structure indicates the possibility of analyzing structural waters in complex inorganic and biological systems. The addition of dynamic nuclear polarization to the high resolution ¹⁷O MAS NMR^75–76 will enhance the ability to study the small number of structural water molecules in these complex systems in comparison to the inorganic crystal in this study.