Two-dimensional 17O multiple quantum magic-angle spinning NMR of organic solids

Article abstract of DOI:10.1021/ja0102181

We report two-dimensional (2D) ¹⁷O multiple-quantum magic-angle spinning (MQMAS) NMR spectra for four ¹⁷O-labeled organic compounds: [¹⁷O₂]-D-alanine (1), potassium hydrogen [¹⁷O₄]dibenzoat

Full text of DOI:10.1021/ja0102181

J. Am. Chem. Soc. 2001, 123, 9119-9125
9119
17
Two-Dimensional O Multiple Quantum Magic-Angle Spinning
NMR of Organic Solids
Gang Wu* and Shuan Dong
Contribution from the Department of Chemistry, Queen’s UniVersity, Kingston, Ontario, Canada K7L 3N6
ReceiVed January 25, 2001. ReVised Manuscript ReceiVed May 23, 2001
Abstract: We report two-dimensional (2D) 17O multiple-quantum magic-angle spinning (MQMAS) NMR spectra
for four 17O-labeled organic compounds: [17O2]-D-alanine (1), potassium hydrogen [17O4]dibenzoate (2), [17O4]-
D,L-glutamic acid‚HCl (3) and [2,4-17O2]uracil (4). The high spectral resolution observed in the 2D 17O MQMAS
NMR spectra allows extraction of precise 17O NMR parameters for all crystallographically distinct oxygen
sites. We demonstrate that rotor synchronization is important in obtaining high-quality 17O MQMAS spectra
for organic compounds. Several issues related to the potential of 17O MQMAS NMR for large biomolecular
systems are also discussed.
Introduction
exploring the potential of solid-state 17O NMR in the study of
organic and biological compounds.7-11 Recent work from this
laboratory has expanded considerably the realm of solid-state
17O NMR for organic compounds.12-18 We were particularly
interested in the effect of intermolecular hydrogen bonding
interactions on 17O NMR tensors. In addition to the experimental
characterization of 17O NMR tensors in important functional
groups, we have also performed extensive quantum chemical
calculations on 17O NMR properties and established a basis for
further understanding 17O NMR parameters.
Multinuclear nuclear magnetic resonance (NMR) spectros-
copy is an important technique for studying molecular structure
and dynamics of chemical and biological systems. Most
successful NMR studies have been based on observation of 1H,
13C, and 15N nuclei. Although oxygen is also an abundant
element in organic and biological molecules, 17O NMR studies
are not common.1 The scarcity of 17O NMR studies arises from
the fact that 17O is a quadrupolar nucleus (spin 5/2) with a very
low natural abundance (0.037%). The major obstacle of obtain-
ing NMR spectra for quadrupolar nuclei is the large size of the
nuclear quadrupole interaction (typically of the order of 106 Hz).
In addition, the unique nature of nuclear quadrupole interactions
also makes it difficult to record NMR spectra with sufficient
spectral resolution to resolve chemically or crystallographically
different sites. For these reasons, the development of 17O NMR
spectroscopy has been quite slow in the past 50 years.
One major limitation of the aforementioned solid-state 17O
NMR studies is that spectral analyses were based exclusively
on observation of “low-resolution” NMR spectra obtained for
either magic-angle spinning (MAS) or stationary samples. This
has made it very difficult to study chemical systems with more
than one type of oxygen atoms. The intrinsic difficulty of
obtaining high-resolution NMR spectra for 17O arises from the
fact that the second-order quadrupole interaction cannot be
completely averaged by the traditional MAS technique. Al-
though new techniques such as dynamic angle spinning (DAS)19
and double rotation (DOR)20 were developed for achieving high-
resolution for half-integer quadrupolar nuclei including 17O,
Solid-state 17O NMR studies of organic compounds were
pioneered by several groups in the past decade. In particular,
Fiat and co-workers attempted to record solid-state 17O NMR
signals for crystalline amino acids.2 Oldfield and co-workers
applied solid-state 17O NMR to probe hemoproteins and model
compounds.3 Ando and co-workers used solid-state 17O NMR
to study H-bonding interactions in polypeptides.4 In addition,
two single-crystal 17O NMR studies were also reported dealing
with organic molecules.5,6 As relatively high magnetic fields
are becoming available, there has been a renewed effort in
(6) Zhang, Q. W.; Zhang, H. M.; Usha, M. G.; Wittebort, R. J. Solid
State Nucl. Magn. Reson. 1996, 7, 147.
(7) Takahashi, A.; Kuroki, S.; Ando, I.; Ozaki, T.; Shoji, A. J. Mol. Struct.
1998, 442, 195.
(8) Salzmann, R.; McMahon, M. T.; Godbout, N.; Sanders, L. K.;
Wojdelski, M.; Oldfield, E. J. Am. Chem. Soc. 1999, 121, 3818.
(9) Godbout, N.; Sanders, L. K.; Salzmann, R.; Havlin, R. H.; Wojdelski,
M.; Oldfield, E. J. Am. Chem. Soc. 1999, 121, 3829.
(10) Yamauchi, K.; Kuroki, S.; Ando, I.; Ozaki, T.; Shoji, A. Chem.
Phys. Lett. 1999, 302, 331.
(11) Kaupp, M.; Rovira, C.; Parrinello, M. J. Phys. Chem. B 2000, 104,
5200.
(12) Dong, S.; Yamada, K.; Wu, G. Z. Naturforsch. 2000, 55A, 21.
(13) Wu, G.; Hook, A.; Dong, S.; Yamada, K. J. Phys. Chem. A. 2000,
104, 4102.
(14) Wu, G.; Yamada, K.; Dong, S.; Grondey, H. J. Am. Chem. Soc.
2000, 122, 4215.
(15) Yamada, K.; Dong, S.; Wu, G. J. Am. Chem. Soc. 2000, 122, 11602.
(16) Dong, S.; Ida, R.; Wu, G. J. Phys. Chem. A 2000, 104, 11194.
(17) Wu, G.; Dong, S. Chem. Phys. Lett. 2001, 334, 265.
(18) Wu, G.; Dong, S.; Ida, R. Chem. Commun. 2001, 891.
(19) (a) Llor, A.; Virlet, J. Chem. Phys. Lett. 1988, 152, 248. (b) Meuller,
K. T.; Sun, B. Q.; Chingas, G. C.; Zwanziger, J. W.; Terao, T.; Pines, A.
J. Magn. Reson. 1990, 86, 470.
* Corresponding author. Phone: 613-533-2644. Fax: 613-533-6669.
E-mail: gangwu@chem.queensu.ca.
(1) Boykin, D. W. 17O NMR Spectroscopy in Organic Chemistry; CRC
Press: Boca Raton, FL, 1991.
(2) (a) Goc, R.; Ponnusamy, E.; Tritt-Goc, J.; Fiat, D. Int. J. Peptide
Protein Res. 1988, 31, 130. (b) Goc., R.; Tritt-Goc., J.; Fiat, D. Bull. Magn.
Reson. 1989, 11, 238.
(3) (a) Augspurger, J. D.; Dykstra, C. E.; Oldfield, E. J. Am. Chem. Soc.
1991, 113, 2447. (b) Oldfield, E.; Lee, H. C.; Coretsopoulos, C.; Adebodum,
F.; Park, K. D.; Yang, S.; Chung, J.; Phillips, B. J. Am. Chem. Soc. 1991,
113, 8680. (c) Park, K. D.; Guo, K.; Adebodun, F.; Chiu, M. L.; Sligar, S.
G.; Oldfield, E. Biochemistry 1991, 30, 2333.
(4) (a) Kuroki, S.; Ando, I.; Shoji, A.; Ozaki, T. J. Chem. Soc., Chem.
Commun. 1992, 433. (b) Kuroki, S. Takahashi, A.; Ando, I.; Shoji, A.;
Ozaki, T. J. Mol. Struct. 1994, 323, 197.
(5) Scheubel, W.; Zimmermann, H.; Haeberlen, U. J. Magn. Reson. 1985,
63, 544.
10.1021/ja0102181 CCC: $20.00 © 2001 American Chemical Society
Published on Web 08/21/2001

9120 J. Am. Chem. Soc., Vol. 123, No. 37, 2001
Wu and Dong
vacuum line. Potassium hydrogen [17O4]dibenzoate was prepared as
these techniques are difficult to apply to organic compounds
for various reasons. To our knowledge, there has been only one
preliminary 17O DAS NMR study where an organic compound
was examined.21 In 1995, Frydman and co-workers22,23 intro-
duced a new approach known as the multiple-quantum magic-
angle spinning (MQMAS) method for completely averaging
second-order quadrupole interactions. This new technique
immediately attracted considerable attention because it can be
readily implemented on most commercial NMR spectrometers.
Over the past several years, the MQMAS method has been
applied to a large number of chemical systems containing
common half-integer quadrupolar nuclei. The first 17O MQMAS
study was demonstrated by Wu et al.24 Subsequently, several
research groups have reported 17O MQMAS studies for many
inorganic systems where 17O quadrupole coupling constants are
reasonably small.25-32 However, 17O MQMAS NMR for organic
compounds is still an unexplored area. In this contribution, we
report for the first time 17O MQMAS NMR results for four 17O-
enriched organic compounds, [17O2]-D-alanine (1), potassium
hydrogen [17O4]dibenzoate (2), [17O4]-D,L-glutamic acid‚HCl (3),
and [2,4-17O2]uracil (4). The primary objective of the present
study is to evaluate the feasibility of 17O MQMAS NMR for
studying organic compounds at a moderate field strength (11.75
T) and with a reasonable 17O enrichment level (ca. 40%). In
addition, this study will attempt to address the question as to
whether the sensitivity of the 17O MQMAS experiment is
suitable for studying biological systems.
1
described previously.12 All products were confirmed by H and 13C
NMR. The exact level of 17O enrichment in compounds 1-4 was not
determined in this study; however, previous studies12,34,35 have yielded
estimates for the 17O enrichment level in these systems: 1 (40%), 2
(50.8%), 3 (40%), and 4 (14% at O2 and 24% at O4).
Solution 17O NMR. All solution 17O NMR experiments were
performed on Bruker Avance-400 and Avance-500 NMR spectrometers
with 5-mm broadband probes. Chemical shifts were referenced to an
external sample of H2O. The solution 17O NMR spectrum of [17O2]-
D-alanine exhibits a single peak at 257 ppm at acidic pH values. KH
[17O4]dibenzoate in aqueous solution shows a single peak at 253 ppm.
The solution 17O NMR spectrum of [17O4]-D,L-glutamic acid‚HCl
exhibits two peaks at 251.0 and 263.5 ppm which can be assigned to
the R-COOH and γ-COOH groups, respectively. These values are in
agreement with previous solution 17O NMR studies.33 The 17O NMR
spectrum of [2,4-17O]uracil in DMSO exhibits two peaks at 329 and
247 ppm with a relative intensity ratio of 1 (O4) to 0.61 (O2). This is
in good agreement with the literature values reported by Fiat and co-
workers.35
Solid-State 17O NMR. All solid-state 17O NMR spectra were
recorded on a Bruker Avance-500 spectrometer operating at 500.13
1
and 67.78 MHz for H and 17O nuclei, respectively. Polycrystalline
samples were packed into zirconium oxide rotors (4 mm o.d.). A Bruker
4-mm MAS probe was used for 17O MAS and MQMAS experiments.
The sample spinning frequency was controlled to be 14 500 ( 4 Hz.
The z-filter pulse sequence proposed by Amoureux et al.36 was used:
P1(φ1)-t1-P2(φ2)-τ-P3(φ3)-ACQ(t2,φ4) where φ1 ) (0°), φ2 )
(0, 0, 60, 60, 120, 120, 180, 180, 240, 240, 300, 300°), φ3 ) (0, 180°),
φ4 ) (0, 180, 180, 0°), and τ ) 20 µs. The optimized excitation (P1)
and conversion (P2) pulse widths were 5.5 and 2.0 µs, respectively.
The radio frequency field strength at the 17O frequency was ap-
proximately 80-90 kHz. The pulse width of the selective 17O 90° pulse
(P3) was 27 µs. The hypercomplex data method37 was used for obtaining
pure-phase 2D spectra. An external sample of H2O (20% 17O atom)
was used for RF calibration and chemical shift referencing. Other
experimental details are given in the figure captions.
Experimental Section
Sample Synthesis. [17O2]-D-Alanine, [17O4]-D,L-glutamic acid‚HCl,
and [2,4-17O]uracil were prepared by acid-catalyzed exchange with
H217O (40.9 atom % 17O, LOT No. IM1378-14, purchased from
ISOTEC Inc., Miamisburg, OH) following the literature procedures.33-35
For example, 150 mg of D,L-glutamic acid monohydrate was dissolved
in 0.45 mL of H217O in a 1.0 mL glass vial. The solution was saturated
with dry HCl gas, sealed, and then heated to 95-100 °C for
approximately 12 h. It was noted that heating over approximately 9
days was required to exchange both O2 and O4 of uracil. Upon
completion of a reaction, excessive H217O was recovered using a high
Results and Discussion
[17O2]-D-Alanine. Figure 1 shows the 2D 17O MQMAS
spectra of compound 1 obtained without and with rotor
synchronization for the t1 increments. Because the 2D spectrum
shown in Figure 1A was obtained without synchronizing the t1
increment with the rotor period, strong spinning sidebands are
present along the F1 dimension. The advantage of this approach
is that the F1 spectral width is not limited by the sample spinning
frequency. However, the presence of F1 spinning sidebands in
2D spectra often results in low peak intensities. In contrast, the
2D spectrum shown in Figure 2B was obtained with the t1
increment being synchronized with the rotor period, TR (i.e.,
∆t1 ) TR). Under such circumstances, the spectral width along
the F1 dimension is equal to the spinning frequency and
consequently all F1 spinning sidebands are folded onto the
central bands. As previously demonstrated by Massiot,38 the
advantage of t1 rotor synchronization is that this approach
enhances the overall sensitivity of MQMAS experiments,
especially for sites associated with large quadrupole coupling
constants. Because the origin of F1 spinning sidebands in 2D
MQMAS spectra is quite complex,39,40 it is often desirable to
obtain sideband-free MQMAS spectra. On the other hand, as
(20) (a) Samoson, A.; Lippmma, E.; Pines, A. Mol. Phys. 1988, 65, 1013.
(b) Chmelka, B. F.; Mueller, K. T.; Pines, A.; Stebbins, J.; Wu, Y.;
Zwanziger, J. W. Nature 1989, 339, 42. (c) Wu, Y.; Sun, B. Q.; Pines, A.;
Samoson, A.; Lippmaa, E. J. Magn. Reson. 1990, 89, 296.
(21) Gann, S. L.; Baltisberger, J. H.; Wooten, E. W.; Zimmermann, H.;
Pines, A. Bull. Magn. Reson. 1994, 16, 68.
(22) Frydman, L.; Harwood, J. S. J. Am. Chem. Soc. 1995, 117, 5367.
(23) Medek, A.; Harwood, J. S.; Frydman, L. J. Am. Chem. Soc. 1995,
117, 12779.
(24) Wu, G.; Rovnyak, D.; Sun, B. Q.; Griffin, R. G. Chem. Phys. Lett.
1996, 249, 210.
(25) Wu, G.; Rovnyak, D.; Huang, P. C.; Griffin, R. G. Chem. Phys.
Lett. 1997, 277, 79.
(26) Dirken, P. J.; Kohn, S. C.; Smith, M. E.; van Eck, E. R. H. Chem.
Phys. Lett. 1997, 266, 568.
(27) Stebbins, J. F.; Xu, Z. Nature 1997, 390, 60.
(28) Xu, Z.; Maekawa, H.; Oglesby, J. V.; Stebbins, J. F. J. Am. Chem.
Soc. 1998, 120, 9894.
(29) Amoureux, J.-P.; Bauer, F.; Ernst, H.; Fernandez, C.; Freude, D.;
Michel, D.; Pingel, U.-T. Chem. Phys. Lett. 1998, 285, 10.
(30) Pingel, U.-T.; Amoureux, J.-P.; Anupold, T.; Bauer, F.; Ernst, H.;
Fernandez, C.; Freude, D.; Samoson, A. Chem. Phys. Lett. 1998, 294, 345.
(31) Lee, S. K.; Stebbins, J. F. J. Phys. Chem. B 2000, 104, 4091.
(32) Bull, L. M.; Bussemer, B.; Anupo˜ld, T.; Samoson, A.; Sauer, J.;
Cheetham, A. K.; Dupree, R. J. Am. Chem. Soc. 2000, 122, 4948.
(33) Gerothanassis, I. P.; Hunston, R.; Lauterwein, J. HelV. Chim. Acta
1982, 65, 1764.
(36) Amoureux, J.-P.; Fernanderz, C.; Steuernagel, S. J. Magn. Reson.
Ser. A 1996, 123, 116.
(37) States, D. J.; Haberkorn, R. A.; Ruben, D. J. J. Magn. Reson. 1982,
48, 286.
(34) (a) Steinschneider, A.; Valentine, B.; Burgar, M. I.; Fiat, D. Magn.
Reson. Chem. 1985, 23, 104. (b) Steinschneider, A.; Burgar, M. I.; Buku,
A.; Fiat, D. Int. J. Peptide Protein Res. 1981, 18, 324.
(35) Burgar, M. I.; Dhawan, D.; Fiat, D. Org. Magn. Reson. 1982, 20,
184.
(38) Massiot, D. J. Magn. Reson. Ser. A 1996, 122, 240.
(39) Amoureux, J.-P.; Pruski, M.; Lang, D. P.; Fernanderz, C. J. Magn.
Reson. 1998, 131, 170.
(40) Charpentier, T.; Fermon, C.; Virlet, J. J. Chem. Phys. 1998, 109,
3116.

17O MQMAS NMR of Organic Solids
J. Am. Chem. Soc., Vol. 123, No. 37, 2001 9121
Figure 2. Experimental and simulated 1D 17O MAS spectra of [17O2]-
D-alanine. The simulated total spectrum (O1+O2) is a sum of O1 and
O2 sub-spectra with a 1:1 intensity ratio. The sample spinning frequency
was 14.5 kHz. A total of 512 transients were collected with a recycle
time of 10 s.
Figure 1. Contour plot of 17O MQMAS spectra of [17O2]-D-alanine
obtained without (A) and with (B) rotor synchronization. Corresponding
1D MAS spectra and the F1 projections of the 2D spectra are shown at
the top and on the side, respectively. The sample spinning frequency
was 14.5 kHz. The t1 increment was 16.67 and 68.97 µs for parts A
and B, respectively. The number of t1 increments was 84 and 42 for
parts A and B, respectively. In both parts A and B, a total of 240
transients were accumulated for each t1 increment with a recycle delay
of 10 s.
Figure 3. Molecular structure and atomic numbering for L-analine.41
The relevant H-bond distances and angles are the following: N-H(1)‚‚‚
O(1), 2.853 Å, 160.9°; N-H(2)‚‚‚O(2), 2.813 Å, 168.1°; N-H(3)‚‚‚O(2),
2.832 Å, 163.7°.
and 1D MAS spectra yielded the following parameters for
compound 1: O1, δiso ) 275 ( 5 ppm, CQ ) 7.60 ( 0.02
MHz, η ) 0.60 ( 0.01; O2, δiso ) 262 ( 5 ppm, CQ
)
6.40 ( 0.02 MHz, η ) 0.65 ( 0.01. The observed and simulated
1D 17O MAS spectra of compound 1 are shown in Figure 2.
Spectral assignment can be readily achieved on the basis of the
crystal structure of L-alanine.41 As seen in Figure 3, the alanine
molecule exists in its zwitterionic form and the two oxygen
atoms have quite different H-bonding environments. In par-
ticular, O1 is involved in only one CdO‚‚‚H-N hydrogen bond,
but O2 is involved in two hydrogen bonds. The stronger
H-bonding environment at O2 also results in a small but
significant lengthening of the CdO(2) bond length, 1.258 Å,
in comparison with CdO(1), 1.242 Å. It has been previously
established that the values of both 17O isotropic chemical shift
and quadrupole coupling constant decrease with increasing
H-bonding strength.12-16,42 On the basis of this general trend
and the aforementioned structural difference, it is straightforward
to assign O1 to the oxygen atom with larger values of δiso and
CQ. It is also noted that our 17O NMR parameters for compound
1 are in reasonably good agreement with the data reported by
Gann et al.21 from an analysis of 17O DAS spectra. However,
the MQMAS approach is clearly easier to implement than the
the sample spinning frequency on modern NMR spectrometers
can be as high as 35 kHz, it is usually not difficult to achieve
a sufficiently large spectral width along the F1 dimension.
Comparison of the two NMR spectra shown in Figure 1
confirms that rotor synchronization can indeed lead to a dramatic
sensitivity enhancement. In the remainder of this contribution,
we will focus only on rotor-synchronized MQMAS experiments.
As also seen from Figure 1B, two isotropic peaks are observed
in the 17O MQMAS spectrum of compound 1, indicating that
the two oxygen atoms of the alanine molecule experience
different chemical environments. Furthermore, each isotropic
peak in the F1 dimension is related to a sub-spectrum along the
F2 dimension, from which information about isotropic chemical
shift (δiso), quadrupole coupling constant (CQ ) e2qQ/h), and
asymmetry parameter (η) can be estimated. However, it should
be emphasized that, because F2 sub-spectra often exhibit line
shape distortion and do not contain quantitative information
about peak intensities, they were used only as a starting point
and restraints in the simulations of 1D MAS spectra. Final NMR
parameter extraction was achieved by comparing spectral
features (peaks and shoulders) in experimental and simulated
1D MAS spectra. Such a combined analysis of the 2D MQMAS
(41) Lehmann, M. S.; Koetzle, T. F.; Hamilton, W. C. J. Am. Chem.
Soc. 1972, 94, 2657.
(42) Bulter, L. G.; Brown, T. L. J. Am. Chem. Soc. 1981, 103, 6541.

9122 J. Am. Chem. Soc., Vol. 123, No. 37, 2001
Wu and Dong
Figure 5. Experimental and simulated 1D 17O MAS spectra of
potassium hydrogen [17O4]dibenzoate. The simulated total spectrum
(O1+O2) is a sum of O1 and O2 sub-spectra with a 1:1 intensity ratio.
The sample spinning frequency was 14.5 kHz. A total of 128 transients
were collected with a recycle time of 10 s.
2 are somewhat different from those determined for the same
compound by Poplett et al.46 in a nuclear quadrupole resonance
(NQR) study (CQ ) 6.165 MHz, η ) 0.591). Nevertheless, it
is also well-known that determination of η was not accurate
from NQR studies. In addition, Poplett et al.46 were unable to
determine the 17O quadrupole parameters for O1, due to the
lack of 1H-17O interactions. From this example, it is quite clear
that solid-state 17O NMR is more advantageous than NQR in
yielding detailed information for all oxygen sites.
Figure 4. (A) Molecular structure and atomic numbering of potassium
hydrogen dibenzoate.43 Hydrogen atoms are not reported in the X-ray
study. The O(2)‚‚‚O(2′) distance is 2.51 ( 0.04 Å. (B) Contour plot of
the 17O MQMAS spectrum of potassium hydrogen [17O4]dibenzoate.
Corresponding 1D MAS spectrum and the F1 projection of the 2D
spectrum are shown at the top and on the side, respectively. A total of
240 transients were accumulated for each t1 increment with a recycle
delay of 4 s. A total of 88 t1 increments were collected.
It is also interesting to note from Figure 4B that the isotropic
peak for O2 is broader than that for O1. This is due to the strong
1H-17O dipolar interaction at the O2 position. Because hetero-
DAS method and has also produced more accurate NMR results
for this compound.
1
nuclear H-17O dipolar interactions are amplified three times
Potassium Hydrogen [17O4]Dibenzoate. The crystal structure
of KH dibenzoate was reported by Skinner et al.43 in an early
X-ray diffraction study. As illustrated in Figure 4A, the two
benzoate groups are related by an inversion center of symmetry.
Although the hydrogen atom positions were not determined in
the early X-ray study, the proton must lie on the midpoint
between O(2) and O(2′) required by the crystallographic
symmetry. Therefore, the O(2)‚‚‚H‚‚‚O(2′) hydrogen bond is
symmetric with the distance between O(2) and O(2′) being
2.51 ( 0.04 Å.
in a 3QMAS experiment, large dipolar interactions are difficult
to completely remove by continuous-wave (CW) high-power
proton decoupling. We recently demonstrated that insufficient
CW proton decoupling can result in additional broadening or
even doubling of the isotropic peaks in 17O MQMAS NMR
spectra.47
[17O4]-D,L-Glutamic Acid‚HCl. Figure 6 shows the molec-
ular structure and the 2D 17O MQMAS spectrum of compound
3. In the crystal lattice, both R-COOH and γ-COOH groups
are in free acid form and all four oxygen atoms are crystallo-
graphically distinct.48 Each oxygen atom of the molecule is also
involved in H-bonding. In particular, O1 and O3 are involved
in hydrogen bonds of the type O-H‚‚‚O as donor and acceptor,
respectively. O2 forms a hydrogen bond from one of the amino
hydrogen atoms. O4 is associated with an O-H‚‚‚Cl- hydrogen
bond, which is quite different from other three-hydrogen bonds.
As seen from Figure 6B, the 17O MQMAS spectrum of
compound 3 exhibits well-resolved isotropic peaks. Compared
with the complex 1D 17O MAS line shape observed for
compound 3, the resolving power of the 2D 17O MQMAS
approach is striking. However, only three isotropic peaks are
observed in Figure 6B. Careful comparison between experi-
mental and simulated 1D MAS spectra revealed that the sub-
As shown in Figure 4B, the 2D 17O MQMAS spectrum of
compound 2 exhibits two isotropic peaks in the F1 dimension.
A combined analysis of 2D MQMAS and 1D MAS spectra
yielded the following 17O NMR parameters for compound 2:
O1, δiso ) 285 ( 5 ppm, CQ ) 8.30 ( 0.02 MHz, η ) 0.15 (
0.01; O2, δiso ) 215 ( 5 ppm, CQ ) 5.90 ( 0.02 MHz, η )
0.90 ( 0.01. The observed and simulated 1D 17O MAS NMR
spectra are presented in Figure 5. Again, the above spectral
assignment was based upon the H-bonding difference between
O1 and O2. The 17O quadrupole parameters determined for O2
are similar to those found in potassium hydrogen diacetate
(CQ ) 5.960 MHz, η ) 0.980)44 and in potassium hydrogen
diformate (CQ ) 5.462 MHz, η ) 0.833). In these latter
compounds, the O‚‚‚O distance, 2.44 Å,45 is similar to that in
compound 2. It is, however, noted that our data for compound
(46) Poplett, I. J. F.; Sabir, M.; Smith, J. A. S. J. Chem. Soc., Faraday
Trans. 2 1981, 77, 1651.
(47) Wu, G. Chem. Phys. Lett. 2000, 322, 513.
(48) Sequeira, A.; Rajagopal, H.; Chidambaram, R. Acta Crystallogr.
1972, B28, 2514.
(43) Skinner, J. M.; Stewart, G. M. D.; Speakman, J. C. J. Chem. Soc.
1954, 180.
(44) Seliger, J.; Zagar, V.; Blinc, R. Chem. Phys. Lett. 1989, 164, 405.
(45) Larsson, G.; Nahringbauer, I. Acta Crystallogr. 1968, B24, 666.

17O MQMAS NMR of Organic Solids
J. Am. Chem. Soc., Vol. 123, No. 37, 2001 9123
Figure 7. Experimental and simulated 1D 17O MAS spectra of [17O4]-
D,L-glutamic acid‚HCl. The simulated total spectrum (all sites) is a sum
of O1, O4, and O2+O3 sub-spectra. Note that the O2+O3 sub-spectrum
is twice as intense as O1 and O4 sub-spectra. The sample spinning
frequency was 14.5 kHz. A total of 1713 transients were collected with
a recycle time of 5 s.
Figure 6. (A) Molecular structure and atomic numbering for L-glutamic
acid‚HCl.49 The relevant H-bond distances and angles are the follow-
ing: O(1)-H(1)‚‚‚O(3′), 2.637 Å, 176.0°; N-H(2)‚‚‚O(2), 2.904 Å,
158.6°; O(4)-H(2)‚‚‚Cl, 3.043 Å, 173.4°. (B) Contour plot of the17O
MQMAS spectrum of [17O4]-D,L-glutamic acid‚HCl. Corresponding 1D
MAS spectrum and F1 projection of the 2D spectrum are shown on
the top and on the side, respectively. A total of 960 transients were
accumulated for each t1 increment with a recycle delay of 5 s. A total
of 28 t1 increments were collected.
and C-O-H from a carboxylic acid functional group, very few
data are available in the literature regarding 17O chemical shifts
for C-O-H groups in carboxylic acids. Therefore, the reason
for the observation of such an unusual chemical shift for O4 in
compound 3 is not obvious at this time. It would be of interest
to see whether ab initio shielding calculations can shed light
on this problem.
[2,4-17O2]Uracil. Uracil is one of the common nucleic acid
bases. The two oxygen atoms of the pyrimidine ring are
chemically different: O2 is related to a urea functional group
and O4 is similar to an amide oxygen. The crystal structure of
uracil indicates that O2 and O4 have quite different H-bonding
environments.49 As shown in Figure 8A, O4 is involved in two
nearly linear CdO‚‚‚H-N hydrogen bonds in a Y-shape fashion,
meanwhile O2 is free of any direct H-bonding interaction. The
two O(4)‚‚‚N distances are 2.865 and 2.864 Å. As will be
discussed later, these structural features have a profound effect
on the observed 17O NMR parameters for uracil.
The 2D 17O MQMAS spectrum of compound 4 clearly shows
two well-resolved peaks. Similar to the discussions described
above, analyses of both the 17O MQMAS and MAS spectra yield
the following 17O NMR parameters for compound 4: O2,
δiso ) 240 ( 5 ppm, CQ ) 7.62 ( 0.02 MHz, η ) 0.50 (
0.01; O4, δiso ) 275 ( 5 ppm, CQ ) 7.85 ( 0.02 MHz, η )
0.55 ( 0.01. The experimental and simulated 1D 17O MAS
spectra of compound 4 are shown in Figure 9. In compound 4,
the value of CQ for the urea-type oxygen, O2, is considerably
larger than that found in crystalline urea, 7.24 MHz.16 As we
noted recently, the small value of CQ in crystalline urea is related
to the very strong H-bonding interactions. The carbonyl oxygen
atom of the urea molecule is hydrogen bonded to four N-H
groups; such strong H-bonding interactions also lead to a
considerable lengthening of the CdO bond, 1.265 Å.50 In
spectrum associated with the central isotropic peak must be twice
as intense to achieve best agreement. This indicates that the
17O NMR signals from two oxygen sites overlap severely. On
the basis of this procedure, we obtained the following 17O NMR
parameters for compound 3: O1, δiso ) 170 ( 5 ppm, CQ
7.20 ( 0.02 MHz, η ) 0.20 ( 0.01; O2 and O3, δiso ) 250 (
5 ppm, CQ ) 6.80 ( 0.02 MHz, η ) 0.58 ( 0.01; O4, δiso
)
)
320 ( 5 ppm, CQ ) 8.20 ( 0.02 MHz, η ) 0.00 ( 0.01. The
experimental and simulated 1D 17O MAS spectra of compound
3 are shown in Figure 7. The observed 17O NMR parameters
for the two carbonyl oxygen atoms, O2 and O3, are essentially
identical. It is noted that the C-O(2) and C-O(3) bond
distances are practically the same, 1.221 and 1.225 Å, despite
the different H-bonding environments. In contrast, the two
C-O-H groups, O1 and O4, exhibit remarkably different 17
O
NMR parameters. In particular, the value of CQ for O4 is about
1 MHz larger than that for O1, and the 17O chemical shifts for
the two seemingly similar oxygen atoms differ by 150 ppm.
Close inspection of the crystal structure of compound 3 indeed
reveals considerable differences between the two C-O-H
groups. The C-O(4) bond length, 1.315 Å, is substantially
longer than that of the C-O(1) bond, 1.296 Å.48 Furthermore,
the O(4)-H(2) bond length, 0.981 Å, is significantly shorter
than that of the O(1)-H(1) bond, 1.017 Å. As mentioned earlier,
O4 is also involved in an unusual hydrogen bond, O-H‚‚‚Cl-.
Despite these structural features, it still seems surprising that
O4 exhibits such a large chemical shift value, which corresponds
to a very deshielded environment. Because it is usually not
possible to detect separate solution 17O NMR signals for CdO
(49) Stewart, R. F.; Jensen, L. H. Acta Crystallogr. 1967, 23, 1102.

9124 J. Am. Chem. Soc., Vol. 123, No. 37, 2001
Wu and Dong
smaller than those observed for secondary amides, 8.50-8.97
MHz.15 In the secondary amides, the carbonyl oxygen is
typically involved in one CdO‚‚‚H-N hydrogen bond with the
O‚‚‚N distance varying between 2.930 and 3.112 Å. As seen
from Figure 8A, O4 of the uracil molecule clearly experiences
a much stronger H-bonding environment, which results in a
much smaller 17O quadrupole coupling constant. These general
trends are perfectly consistent with the correlation between 17
O
quadrupole parameters and H-bonding environment as we
illustrated in several recent studies.13-16,18
It is also interesting to compare the 17O chemical shifts
measured for uracil in the solid state with those measured for
a solution sample. The solution NMR spectrum of [2,4-17O]-
uracil in DMSO exhibits two peaks at 329 and 247 ppm assigned
for O4 and O2, respectively. In the solid state, although the
17O isotropic chemical shift for O2, 240 ppm, is essentially the
same as that in solution, the 17O chemical shift for O4 is shifted
to the direction of shielding by 54 ppm! This large change in
the 17O chemical shift must be related to the strong H-bonding
interaction at O4 in the solid state. As mentioned earlier, O4 is
strongly hydrogen bonded to the two N-H groups of the
neighboring molecules whereas O2 is free of any direct
H-bonding interaction. In solution, all these intermolecular
H-bonding interactions are essentially absent. Therefore, the 17
O
chemical shift values measured in solution correspond to
situations where the uracil molecules are isolated from one
another. In the solid state, because strong intermolecular
H-bonding interactions are present, the 17O chemical shifts are
profoundly changed. A similar effect was also observed for
thymine.18 The implication of our observation is that 17O NMR
is potentially useful for detecting base pairing in nucleic acid
structures. Extensive ab initio calculations for nucleic acid bases
confirmed that the 17O NMR tensors are remarkably sensitive
to the H-bonding environment; a detailed analysis of solid-state
17O NMR parameters as a function of hydrogen bond geometry
will be published elsewhere.
Figure 8. (A) Molecular and H-bonding structures of uracil.50 (B)
Contour plot of the 17O MQMAS spectrum of [2,4-17O2]uracil.
Corresponding 1D MAS spectrum and the F1 projection of the 2D
spectrum are shown at the top and on the side, respectively. A total of
960 transients were accumulated for each t1 increment with a recycle
delay of 5 s. A total of 24 t1 increments were collected.
Resolution and Sensitivity of 17O MQMAS Experiments.
As is clearly seen from the above MQMAS spectra, the complete
removal of second-order quadrupolar broadening has increased
the spectral resolution dramatically. To make quantitative
comparison, we define the MQ evolution time as t1 and label
the isotropic dimension, F1, in frequency units (hertz) after the
shearing transformation.51 Using this convention, we can
describe the peak positions along the isotropic dimension of
5
the 3QMAS spectra for S ) /2 by
10
17
3QMAS
iso
CS
iso
(2)
iso
ν
) -ν
-
ν
(1)
CS
where ν is the frequency contribution from the chemical
iso
(2)
shift and ν is the isotropic second-order quadrupole shift
iso
defined as:
2
CQ
1
500
(2)
iso
Figure 9. Experimental and simulated 1D 17O MAS spectra of
[2,4-17O2]uracil. The simulated total spectrum (O2+O4) is a sum of
O2 and O4 sub-spectra with a 7:12 intensity ratio. This ratio was
determined by solution 17O NMR. A total of 445 transients were
collected with a recycle time of 5 s.
ν
)
(3 + η2)
(2)
νL
The observed line widths for the isotropic peaks in the 17O
MQMAS spectra were also reported in Table 1. Compared with
peaks in the 1D MAS spectra, the line width of the isotropic
peaks in the F1 dimension is reduced by at least an order of
magnitude. At this stage, factors that will ultimately determine
the resolution limit in 17O MQMAS spectra for organic
compounds deserve comment. As seen from Table 1, the line
comparison, the CdO(2) bond length in compound 4 is
significantly shorter, 1.215 Å. On the other hand, the amide-
type oxygen atom of compound 4, O4, exhibits a CQ value much
(50) Swaminathan, S.; Craven, B. M.; McMullan, R. K. Acta Crystallogr.
1984, B40, 300.
(51) Man, P. P. Phys. ReV. B 1998, 58, 2764.

17O MQMAS NMR of Organic Solids
J. Am. Chem. Soc., Vol. 123, No. 37, 2001 9125
Table 1. Experimental Solid-State 17O NMR Parameters for
Compounds 1-4
Finally, several new methods54-58 have recently been demon-
strated to improve the sensitivity of MQMAS experiments by
severalfold. Combining all these factors on a conservative basis
leads to the conclusion that it is indeed possible to apply 17O
MQMAS NMR to biomolecular systems with significant
molecular weights. It should be mentioned that “low resolution”
solid-state 17O NMR spectra have been successfully recorded
for large biomolecular systems such as [17O2]myoglobin (ca.
17 kDa) and [17O2]hemoglobin (ca. 68 kDa) as reported by
Oldfield and co-workers.3
3QMAS
δiso
ppm MHz
((5) ((0.02) ((0.01) ((0.2) ((0.10)
/
CQ/
ν
/
∆ν1/2
/
iso
η
Hza
kHzb
compd
[17O2]Ala (1)
O1 275
O2 262
O1 285
O2 215
7.60
6.40
8.30
5.90
7.20
6.80
6.80
8.20
7.62
7.85
0.60
0.65
0.15
0.90
0.20
0.58
0.58
0.00
0.50
0.55
14.5
17.4
18.0
23.0
20.2
14.5
14.5
8.5
0.70
0.82
1.17
1.60
0.70
0.94
0.94
0.59
0.94
1.12
[17O4]PHB (2)
[17O4]Glu‚HCl (3) O1 170
O2 250
O3 250
O4 320
O2 240
O4 275
Conclusion
[17O2]uracil (4)
16.2
13.8
We have presented high-resolution 17O MQMAS NMR results
for 17O-enriched organic compounds ranging from free amino
acids to a nucleic acid base. Using an 17O enrichment level of
ca. 40% and a moderate magnetic field (11.75 T), we were able
to obtain high-quality 17O MQMAS NMR spectra for organic
compounds with 17O quadrupole coupling constants greater than
8 MHz. The high-resolution 17O MQMAS spectra revealed
crystallographically distinct oxygen sites, which makes it
possible to extract precise information about 17O chemical shifts
and quadrupole parameters for each of the oxygen sites.
Accumulation of such site-specific NMR information will
ultimately lead to a better understanding of the relationship
between 17O NMR parameters and molecular structure. The
present study is a continuation of our recent effort in exploring
the potential of solid-state 17O NMR spectroscopy. Although
we have examined only small organic molecules in the present
study, the basic H-bonding features in the systems studied are
similar to those observed in proteins (e.g., side chain conforma-
tion) and nucleic acids (e.g., base pairing). The demonstrated
sensitivity of 17O MQMAS experiments suggests that high-
resolution 17O MQMAS NMR can potentially become a
practical technique for chemists to study organic and biological
systems. The results of the present study are important in that
they represent the first actual demonstration of 17O MQMAS
NMR for organic compounds. In addition, the new experimental
solid-state 17O NMR parameters reported in this study will be
useful as testing cases for future quantum chemical computa-
tions. The challenge of 17O NMR spectroscopy for biological
solids remains to lie in the synthesis of 17O-enriched biological
molecules and in the development of new methodologies to
further enhance NMR sensitivity.
a Isotropic peak positions in MQMAS spectra. b Full width at the
half-height of the isotropic peaks.
width observed for organic solids is typically between 0.5 and
1.5 kHz. In contrast, much smaller line widths (0.1-0.2 kHz)
were observed in 17O MQMAS spectra of inorganic systems
where 1H-17O dipolar interactions are weak.24,25 It is thus
believed that a main contributor to the residual line width in
1
17O MQMAS spectra is the incomplete removal of H-17O
dipolar interactions. Because new decoupling sequences such
as TPPM52 have been demonstrated to be more effective than
the CW decoupling scheme, it is anticipated that the residual
line width in 17O MQMAS spectra can be further reduced. Other
factors that may contribute to the residual line width include
poor crystallinity of the sample and inaccurate setting of the
magic angle. The advantage of 17O MQMAS over traditional
solution NMR is that the residual line width in solid-state NMR
spectra is not restricted by the molecular weight of the system
under observation. Another application of eq 1 is to confirm
the values of δiso, CQ, and η obtained from 1D MAS spectral
analyses. This can be done simply by comparing the observed
isotropic peak positions in the F1 axis with the calculated values
from eq 1 for each of the oxygen sites.
In this study, the molecular weight of the largest molecular
system is on the order of 200 and the 17O enrichment level is
generally less than 40%. It is only natural to question whether
the sensitivity of 17O MQMAS experiments can be sufficiently
high to make studies of biological macromolecules possible.
Although the sensitivity of NMR experiments depends on many
factors, it is still meaningful to use the current 17O MQMAS
NMR results to shed light on the feasibility of the approach for
larger molecular systems. For 17O MQMAS experiments, the
key factors for sensitivity include (1) the level of 17O enrichment
of the sample, (2) the strength of the principal B0 magnetic field,
and (3) the efficiencies for both MQ generation and MQ-to-1Q
conversion. Suppose that the sensitivity of 17O central-transition
NMR experiments depends on the applied magnetic field as
B07/2. With a maximum level of 17O enrichment, 100%, the
sensitivity of 17O MQMAS experiments at 18.8 T (800 MHz
Acknowledgment. We wish to thank the Natural Sciences
and Engineering Research Council (NSERC) of Canada for
research and equipment grants. This research was partially
supported by a grant from the Advisory Research Committee
of Queen’s University. G.W. thanks Queen’s University for a
Chancellor’s Research Award (2000) and the Government of
Ontario for a Premier’s Research Excellence Award (2000). We
are also grateful to Alan Wong and Ramsey Ida for assistance
in preparing several figures.
1
for H) can be increased by an order of magnitude compared
JA0102181
with that reported in this study (40% enrichment at 11.75 T).
This will make 17O MQMAS experiments feasible for molecular
systems of a few kDa with the methodologies used in the present
study. The overall sensitivity of the MQMAS experiment can
be further increased by an order of magnitude with a new data
acquisition scheme utilizing a train of quadrupolar Carr-
Purcell-Meiboom-Gill refocusing pulses (QCPMG).53 This
additional sensitivity enhancement may extend the MQMAS
method to biomolecular systems of considerably larger sizes.
(53) (a) Vosegaard, T.; Larsen, F. H.; Jakobsen, H. J.; Ellis, P. D.;
Nielsen, N. C. J. Am. Chem. Soc. 1997, 119, 9055. (b) Larsen, F. H.; Nielsen,
N. C. J. Phys. Chem. 1999, 103, 10825.
(54) Madhu, P. K.; Goldbourt, A.; Frydman, L.; Vega, S. Chem. Phys.
Lett. 1999, 307, 41.
(55) Goldbourt, A.; Madhu, P. K.; Vega, S. Chem. Phys. Lett. 2000,
320, 448.
(56) Alemany, L. B.; Callender, R. L.; Barron, A. R.; Steuernagel, S.;
Iuga, D.; Kentgens, A. P. M. J. Phys. Chem. B 2000, 104, 11612.
(57) Iuga, D.; Scha¨fer, H.; Verhagen, R.; Kentgens, A. P. M. J. Magn.
Reson. 2000, 147, 192.
(52) Bennett, A. E.; Rienstra, C. M.; Auger, M.; Lakshmi, K. V.; Griffin,
R. G. J. Chem. Phys. 1995, 103, 6951.
(58) Vosegaard, T.; Florian, P.; Massiot, D.; Grandinetti, P. J. J. Chem.
Phys. 2001, 114, 4618.