ABSTRACT
We describe herein the use of a 2H-labeling strategy to achieve specific assignments of considerably overlapped cross peaks in the 1H-NMR spectrum of a DNA trinucleotide repeat sequence. Our strategy focuses on site-specific 2H-labeling of base moieties to simplify the NMR spectral regions which contain the major portion of the structural information. To achieve efficient preparation of 2H8- or 2H6-labeled DNA and RNA nucleosides and nucleotides, the existing synthetic and purification procedures were significantly improved. Our experiments demonstrate that pyrimidine H6 deuteration reactions may be carried out using non-deuterated base reagents with DMSO-d6 as a 2H donor. These reactions are simple and economic to perform and produce base deuterated nucleosides and nucleotides in high yield. The 2H-labeled residues have been incorporated into oligonucleotides with minor modifications of the existing reaction conditions. Using the homologous CGG repeat sequence, d(CGG)5, as an example, the effectiveness of the site-specific base deuteration strategy is demonstrated. In the otherwise extensively overlapped spectra of d(CGG)5, 2H-labeling has permitted unambiguous identification of a sequential connectivity at a central CG step and confirmation of several other NOE assignments. This information is critical for elucidation of the structure and the folding of the CGG repeat sequences and will contribute to the intensive effort to understand the mechanisms of triplet expansion, which has been implicated in the development of a number of hereditary neurodegenerative diseases. In addition to the two dimensional spectral simplification in a key spectral region using site-specific 2H8/2H6-labeling, the potential applications of the prescribed strategy in homonuclear three dimensional experiments are also discussed.
For several decades, perdeuteration and selective deuteration have been a useful approach for simplification of NMR spectra and for other structural studies of large biomolecules (1 -3 ). In the last few years, in particular, driven by the progress in multinuclear multidimensional NMR spectroscopy, deuteration of nucleic acids and proteins has especially found wide applications in the solution NMR studies of these complex molecular systems (4 -7 ). The major recent applications of uniform or selective deuterium (2H) labeling are in several areas: (i) uniform or fractional 2H-labeling in 13C- and 15N-labeled large proteins and oligonucleotides for suppression of spin diffusion and for deceleration of spin relaxation. 2H-labeled samples, therefore, exhibit enhanced signal sensitivities compared to their 1H counterparts. (ii) Selective 2H-labeling a portion of a residue or certain types of residues in a molecule for simplification of overcrowded 1H NMR spectra. These 2H-labeled molecules help reduce the spectral complexity and make it possible for specific assignments. Protein 2H-labeling has been carried out biosynthetically to produce uniformly labeled molecules, while various biosynthetic and chemical routes for preparation of 2H-labeled oligonucleotides are available (1 ,3 ,5 ,6 ,8 -15 ). In these studies, the NMR spectral comparisons between natural and the deuterated sequences clearly demonstrate the power of selective deuteration methods applied to resolve spectral overlaps of large, complex molecules. To advance these applications, we describe in this report two major improvements: efficient synthesis of oligonucleotides containing site-specific incorporation of deuterated base residues and use of deuteration to selectively simplify the fingerprint spectral regions (i.e. base to sugar proton region) of oligonucleotides. Our work has been focused on the efficient syntheses of nucleosides and nucleotides containing 2H-labeled base residues. As shown in the literature and particularly in this report, these reactions and the subsequent purification are easier to perform and much more cost effective than multi-step chemical and biochemical procedures for sugar deuteration (6 ,16 -18 ). We recognized that in a typical two dimensional (2D) NOESY spectra of oligonucleotides, structural information is primarily contained in the region that connects base with sugar protons (Figs 1 A and 2 A). If the selected base protons, especially H8 in purines and H6 in pyrimidines, are replaced with 2H, all the pertinent cross peaks correlating the base to sugar protons will disappear. Our strategy of site-specific incorporation of base deuteration is, therefore, aimed at effective elimination of selected 1H signals in this important spectral region to provide well-defined spectral information for specific sites in the molecule studied.
For purine residues, upon completion of the reaction, the mixture of D2O and TEA was rotatory-evaporated in vacuo. Deuterated nucleosides thus obtained were essentially pure and contained 99% 2H8 as assessed by one dimensional (1D) NMR spectral comparison with the commercial dA or dG (expanded regions shown in Fig. 3 ). The products were directly protected and derivatized into phosphoramidites according to typical procedures described in the literature (23 ). For pyrimidine nucleosides, after the reaction, the dC solution was diluted with cold H2O, neutralized using HCl (0.5 M) and rotatory-evaporated in vacuo. The residue was then dissolved in a mixture of CH3OH:CH2Cl2 (1:1) and the resultant solution was filtered and concentrated in vacuo. The subsequent ion exchange column chromatography (OH-, Dowex 1 × 2-400) using H2O as eluant followed by concentration produced crystalline 2H6-dC (~95% yield, 97-99 and 50-78% 2H6 and 2H5, incorporation, respectively). 2H6-dC was also purified using reverse phase C18 chromatography using 0-10% CH3OH/H2O to yield the crystalline product. The T reaction was quenched in a way similar to that used for dC. Purification on a silica gel column (Silica gel 60, 230-400 mesh) using 10-15% CH3OH/CH2Cl2 furnished crystalline 2H6-T with 95% 2H incorporation in overall >95% yield.
NMR experiments were conducted on either a GE 300 MHz or a Bruker AMX-II 600 MHz spectrometer. The disappearance of the 1H signal at deuterated positions was monitored using 1D 1H NMR. Both 1H and 31P 1D NMR were used to monitor the reactions of nucleotides. The degree of 2H substitution was derived from the integrated residual 1H peak intensities of base proton resonances compared to that of H1' (100%). These spectra were recorded with a spectral window of 6 kHz (12 p.p.m.) and a digital resolution of 0.375 Hz per point. The spectra of d(CGG)5 were recorded in D2O or 10% D2O, 90% H2O contained 0.1 M NaCl, 10 mM phosphates and 0.1 mM EDTA, pH 6.5 at 25 and 42°C. NMR data were processed using the UXNMR (Bruker Instruments, Inc.) and the Felix 95.0 programs (Molecular Simulations, Inc.). A representative NOESY spectrum recorded at 42°C is shown in Figure 1 . The cross peak assignments of the oligonucleotide were obtained by tracing the sequential connectivities and by comparing to the shorter sequence analogs, such as d(CGG)3, which have been analyzed in great detail (21 ,24 ). The complete analysis of the spectral information of d(CGG)5 will be reported separately.
We report here improved procedures for 2H incorporation into the base residues of all four DNA or RNA nucleosides and potentially for the corresponding nucleotides. New reaction conditions have been developed, under which the desired products can be obtained (i) within a shorter period of time, (ii) with a higher level of deuteration and (iii) higher yield, and (iv) using simpler work-up procedures compared to literature methods.
Purines. It is well known that purine H8 can be easily substituted by 2H under various basic conditions (9 ,10 ). While most of these reactions used inorganic salt conditions, this work applied 1-5 equivalent triethylamine (TEA) in D2O for preparation of 2H8-purine nucleosides at 60-70°C (Table 1 ). TEA has the advantage of being volatile and thus, can be easily removed without using any chromatography. 2H8-dA and 2H8-dG were, therefore, not purified after the reaction and were directly converted into phosphoramidites in good yields and purity which is comparable to that of commercial products. These conditions result in nearly complete deuteration at the 8-position of the A and G bases (Table 1 and Fig. 3 ) and quantitative yield for both deoxy and ribonucleosides. The reactions for dG took 15-30 h (small scale reaction required less time), while the reactions for dA required 50-60 h. Adenine H2, which, as well as H8, can simultaneously be exchanged into 2H in the presence of PtCl2/NaBD4 (14 ), does not exchange with 2H under these conditions. A small amount of DMSO-d6 was added to the reactions of purine ribonucleosides to improve their solubility. When the reactions of comparable scales are compared the required reaction times tend to increase in the order of deoxy-purines, ribo-purines to purine nucleotides. The same reaction conditions were applied to ATP and GTP on small scale reactions. These 2H8-nucleotides can then be used in in vitro transcription of RNA oligomers. NMR showed the formation of the deuterated products (Table 1 ), and 31P 1D spectra showed that the phosphate linkages in ATP and GTP are stable over the course of the reaction.
Pyrimidines. More vigorous reaction conditions were required for base 2H6 incorporation (9 ,15 ). Although it is possible to discriminate H6 and H5 substitutions, this work presents no effort in this aspect since H5 often exhibits much fewer NOEs and thus, the related spectra are less crowded. In general, two sets of reaction conditions have been applied (reagents were either 1H or 2H forms, Table 1 ): methanol/sodium methoxide/DMSO-d6 (applied to C or U nucleobases) or water/sodium hydroxide/DMSO-d6 at 60-135°C (applied to T, CMP and UMP). We have found that it is not necessary to use deuterated base reagents as long as DMSO-d6 is present in excess, although a slightly longer reaction time is needed (Table 1 ). In these reactions, we have also been able to reduce the amount of the base reagents used to one or two equivalents. This resulted in milder reaction conditions to afford, after chromatography purification [rather than charcoal filtration as reported in the literature (9 ,15 )], products in improved yield [>95% compared to 70-80% in the literature (9 ,15 )] and 2H6 incorporation [>95% 2H6 in T and U compared to 90% in the literature (9 ,15 )]. The 1D NMR spectra of the deuterated dC and T (5-methyl is untouched) are displayed in Figure 3 . Deuteration of CTP and UTP under similar conditions has not been successful due to solubility problems of these compounds in DMSO. However, CMP and UMP may be treated with NaOD (3 equiv.) in a mixture of D2O and DMSO-d6 (vol 2:3) at elevated temperatures to give deuterated products containing ~75% 2H6 (Table 1 ). The reactions potentially can be allowed to proceed for a longer time to achieve more complete deuteration at the expense of the overall reaction yield (Table 1 ).
In an effort to understand the molecular basis of triple nucleotide expansions observed in certain neurological regulatory gene sequences, we have embarked on studies of all four types of CXG (X = A, C, G or T) repeat sequences (22 ). As we extend our studies from short, model sequences to longer repeats, rather monotone appearances in the NMR spectra were observed because of repeated structural units in these molecules. Interesting NMR signals from the residues of structurally different portions of the molecule are likely to be buried under the clusters of the major signals. Therefore, it is critical to use isotope labeling methods in order to observe the behavior of an individual trinucleotide unit in a long repeat sequence. At the present time, 13C and/or 15N-labeling of DNA fragments (25 ) of a long sequence is rather cost prohibitive. Furthermore, such a labeling scheme does not allow separation of one repeat unit from another and thus, would not, in any way, resolve the spectral overlaps in a repeat sequence. The site-specific labeling strategy, however, should be particularly well suited for our applications. Nucleotide base 2H-labeling can be achieved at lower cost and in a shorter period of time. More importantly, NOE connectivities involving base proton resonances are the major source of conformational and structural information (Fig. 2 A and B) and it is often that the spectral regions containing these connectivities [i.e. between base (H6 or H8) and sugar proton (H1', H2', H2'' and H3')] resonances are congested as illustrated in Figures 1 A and 2 C. 2H-labeled base residues permit selectively removal of proton resonances in this region and the resultant spectrum can be unambiguously assigned (Fig. 2 D).
Our previous studies of the CGG triplet repeats have identified the formation of an antiparallel homo-duplex by d(CGG)3 at ~30°C (21 ). In this duplex, the two strands of d(CGG)3 align in a staggered fashion to form two Watson-Crick C-G base pairs in each of the 5'-GGC-GGC-3' repeat units, whereas the middle G is mismatched and conformationally flexible. We have further examined, using UV and NMR, a series of d(CGG)n (n = 3-20) in order to define the structure and solution behavior of the longer CGG repeat sequences. However, the NOESY spectra of these sequences are too overlapped to be meaningful. Even for a five repeat sequence, the five H8s of the ten G residues are overlapped (Fig. 1 A). To take advantage of site-specific deuteration, we have synthesized d(CGG)5 containing 2H8-G at 2, 3, 5, 6, 11, 12 and 14 positions [d(C1GGCGGC7G8G9C10GGCGG15) or 2H-d(CGG)5, G residues containing H8 are underlined]. The 3'-terminal G on solid support was not 2H-labeled and this resonance is, nevertheless, resolved (Fig. 1 A). The H8 in the central G8 and G9 residues will permit the assignments of the related connectivities. In this sequence, we are mostly interested in the conformation of the central CGG residues, which potentially may be the point of folding (Huang and Gao, unpublished UV and pulsed field gradient NMR experiments).
2H-d(CGG)5 was synthesized and deprotected using a ND4OD aqueous solution at room temperature to minimize 2H8 to H8 exchange at deuterated sites. Although ND4OD may cause reverse exchange (H8 to 2H8) at non-deuterated sites, this exchange is more forgiven since it does not generate NMR signal. In the end of the reaction, it was found that H8s maintained the majority of their population during the ammonolysis (Fig. 1 B). A potential problem associated with base 2H-labeled oligonucleotides is that H8 of purine residues, especially that in G residues, may slowly exchange with D2O under neutral pH. This would cause gradual loss of H8 signals. We have experienced this unfavorable exchange over the time when 2D and 3D NMR spectra were recorded at 42°C for more than 10 consecutive days. The flexible terminal G H8 first disappeared, followed by the internal G8 and G9 H8 resonances. However, normally, this exchange may not be as threatening since NMR spectra of oligonucleotides are usually recorded at room temperatures or below. Under these conditions H and 2H exchange is sufficiently slow.
The appearance of the NOESY spectrum of 2H-d(CGG)5, displayed in Figure 1 B, is identical to the non-labeled d(CGG)5 (Fig. 1 A) except that the NOE patterns related to the five overlapped G residues (G6, G8, G9, G11 and G14) are greatly simplified. The isolated signals have provided important structural information. In this spectral region, the C H6 resonances fall in the ~7.2 p.p.m. range (except the terminal C1 H6 at 7.6 p.p.m.) and these resonances exhibited NOEs to the intra-residue sugar H2' and H2'' protons (1.7-2.4 p.p.m. along the F1 axis) and to the 5'-preceding G sugar H2' and H2'' protons (2.4-2.9 p.p.m. along the F1 axis, box I in Fig. 1 A). The latter set of NOEs identifies the four 5'-GC connectivities in d(CGG)5. All 10 G H8 resonances fall in a narrow 7.5-7.8 p.p.m. range, which exhibited NOEs to intra-residue sugar H2' and H2'' protons (2.4-2.9 p.p.m. along the F1 axis) and to the 5'-preceding C sugar H2' and H2'' protons (1.7-2.4 p.p.m. along the F1 axis, squared boxed regions II-IV). There are five 5'-CG steps; several related NOEs are discernible from the spectrum shown (Fig. 1 A, boxes II-IV). However, because five G H8 resonances are clustered in narrow spectral region, the 5'-C7G8 sequential NOEs cannot be unambiguously identified. This is, in particular, an important structural feature related to the questions concerning the folding and structural continuity of the longer CGG repeat sequences.
The above spectral ambiguity has been resolved in the spectrum of 2H-d(CGG)5. The peaks in boxes II and III completely disappeared (Fig. 1 B), identified these NOEs belonging to the sequential connectivities from G2, G5, G11 and G14 to the 5'-preceeding C residues. The spectral comparison also ruled out a connectivity between 5'-C7 H2' and G8 H8. However, a weak peak was detected between C7 H2'' and G8 H8 (Fig. 1 B, box IV), providing definitive evidence that C7 and G8 are weakly stacked. Theoretically, we should be able to unambiguously identify the connections between G9 and C10 using homonuclear 3D NOESY-NOESY spectral analysis (26 ). In a 2D plane located at the G9 H8 frequency, cross peaks are expected linking G9 H8 with intra- and inter-residue sugar protons on the NOE lines (Fig. 1 C). There should also be cross peaks from these sugar protons, such as G9, to base protons, such as C10 H6, through 3D NOE magnetization transfer. The 3D NOEs will thus separate the overlaps in the 2D spectral region, where C7 and C10 H6 fall in the same resonance frequency (Fig. 1 B). Therefore, the 3D NOEs can be unambiguously assigned to G6/C7 and/or G9/C10 sequential connectivities. However, after exhaustive 2D and 3D experiments in D2O we discovered that the internal G8 and G9 H8s, albeit exchanging with solvent D2O much slower than the 3'-terminal G H8, have almost completely exchanged into 2H8. The peak intensity in the 3D data was too weak to provide definitive information. It is also unfortunate that H8 resonances of G8 and G9 are overlapped and thus, the sequential connectivity between G8 and G9 cannot be determined. Work is in progress in this laboratory to prepare alternative 2H-labeled d(CGG)5 to further resolve spectral overlaps.
In summary, we have modified and improved reactions and purification procedures so that the base 2H8 and 2H6 labeled DNA and RNA nucleosides and nucleotides can be prepared efficiently and economically. Using the homologous CGG repeat sequence, d(CGG)5, as an example, we demonstrate an efficient strategy using site-specific 2H-labeling to simplify the spectral region which is most important for nucleic acid structure elucidation. We propose that the advantage of selective 2H-labeling may be further explored using homonuclear 3D NMR spectroscopy (Fig. 1 C). NOESY-NOESY 3D spectra (26 ) will be most useful when the planes at the selected base proton frequency are reviewed. In these planes, sequential NOE connectivities may be further resolved. NOESY-TOCSY 3D spectra (27 ), if also reviewed at the selected base proton frequency, will reveal sugar proton scalar connectivities for only those which have NOE contacts with the non-deuterated base protons. This will significantly reduce the overlaps in the sugar proton-sugar proton correlation spectral region without the requirement of the lengthy synthesis of deuterated sugars. These experiments are presently in progress in this laboratory.
The 600 MHz NMR spectrometer at the University of Houston is funded by the W.M.Keck Foundation. Acknowledgment is made to NIH (R01GM54652, R29GM59957) and the Robert A.Welch Foundation (E-1270) for financial support and to the W.M.Keck Center for Computational Biology for computer resource support. Lillian Pierson was a high school Welch Summer Scholar and has contributed to some of the deuteration chemical reactions.
This article has been cited by other articles:
Synthesis of base 2H-labeled nucleosides and nucleotides. All deuterated reagents and solvents were purchased from Cambridge Isotope Laboratories, Inc. The deuteration procedures reported in the literature (9 ,15 ) were modified to derive 2H8 and 2H6/2H5-labeled dA, dC, dG, T, A, C, G and U. ATP, GTP, CMP and UMP were each treated in an NMR tube using the reaction conditions similar to those used for the synthesis of the corresponding nucleoside. Typical synthetic conditions and results of these reactions are summarized in Table 1 .
Nucleosidesa
Reaction conditionsb
Duration (h)
% Substitutionc
1
dA (4 mmol)
TEA (1)/D2O (50 ml), 60°C
48
H8: 99%, H2: 0%
2
dC
DOCD3/NaOCD3 (8.5)/DMSO-d6, 60°C
18
H6: 99%, H5: 55%
3
dC
HOCH3/NaOCH3 (8.5)/DMSO-d6, 60°C
18
H6: 92%, H5: 48%
4
dC
DOCD3/NaOCD3 (8.5)/DMSO, 60°C
20
H6: <5%, H5: 0%
5
dC
DOCD3/NaOCD3 (2.5)/DMSO-d6, 60°C
17
H6: 99%, H5: 30%
6
dC (1 mmol)
DOCD3/NaOCD3 (2.0)/DMSO-d6 (2.2 ml), 60°C
24
H6: 98%, H5: 78%
7
dC (1 mmol)
DOCD3/NaOCD3 (1.0)/DMSO-d6 (2.5 ml), 60°C
48
H6: 97%, H5: 50%
8
dG (4 mmol)
TEA (1)/D2O (95 ml), 65°C
30
H8: 99%
9
T (1 mmol)
NaOH (2.5)/DMSO-d6 (3.0 ml), 135°C
120
H6: 95%
10
T (1 mmol)
NaOD (2.5)/DMSO-d6 (3.0 ml), 135°C
90
H6: 95%
11
A
TEA (4.5)/D2O/DMSO-d6, 60°C
57
H8: 95%, H2: 0%
12
C
DOCD3/NaOCD3 (3.0)/DMSO-d6, 60°C
24
H6: 95%, H5: 65%
13
G
TEA (1.7)/D2O:DMSO-d6, 60°C
<16
H8: 99%
14
U
DOCD3/NaOCD3 (1.2)/DMSO-d6, 135°C
<22
H6: 98%, H5: 98%
15
ATP
TEA (4.0)/D2O, 60°C
130
H8: 95%
16
GTP
TEA (4.0)/D2O, 60°C
21
H8: 95%
17
CMP
NaOD (3.0)/D2O:DMSO-d6, 60°C
>15
H6: >75%, H5: <10%
18
UMP
NaOD (3.0)/D2O:DMSO-d6, 60°C
>20
H6: >66%, H5: <10%
REFERENCES
CiteULike 
Connotea 
Del.icio.us What's this?
A. Adhikary, A. Y. S. Malkhasian, S. Collins, J. Koppen, D. Becker, and M. D. Sevilla
UVA-visible photo-excitation of guanine radical cations produces sugar radicals in DNA and model structures
Nucleic Acids Res.,
October 4, 2005;
33(17):
5553 - 5564.
[Abstract]
[Full Text]
[PDF]
This Article 
Abstract
Print PDF (187K)
Alert me when this article is cited
Alert me if a correction is posted
Services
Email this article to a friend
Similar articles in this journal
Similar articles in ISI Web of Science
Similar articles in PubMed
Alert me to new issues of the journal
Add to My Personal Archive
Download to citation manager
Search for citing articles in:
ISI Web of Science (19)
Request Permissions
Commercial Re-use Guidelines
for Open Access NAR Content
Google Scholar
Articles by Huang, X.
Articles by Gao, X.
Search for Related Content
PubMed
PubMed Citation
Articles by Huang, X.
Articles by Gao, X.
Social Bookmarking
What's this?