Synthesis and studies on the effect of 2-thiouridine and 4-thiouridine on sugar conformation and RNA duplex stability
Synthesis and studies on the effect of 2-thiouridine and 4-thiouridine on sugar conformation and RNA duplex stability
Raju K.
Kumar
and
Darrell R.
Davis*
Department of Medicinal Chemistry, University of Utah,
Salt Lake City
, Utah 84112,
USA
Received October 23, 1996;
Revised and Accepted January 20, 1997
ABSTRACT
In order to understand the effect of 2-thiouridine (s
2
U) substitution on RNA structure and the potential for stabilization of tRNA
codon-anticodon interactions through s
2
U-34 modification, a pentamer RNA sequence, Gs
2
UUUC, was synthesized and characterized by NMR spectroscopy. The single strand
contains the UUU anticodon sequence of tRNA
Lys
with flanking GCs to increase duplex stability. Regiochemical effects of
uridine thiolation were determined by comparing the structure and stability of
the 2-thiouridine containing oligonucleotide with an identical sequence
containing 4-thiouridine (s
4
U) and also the normal uridine nucleoside. Circular dichroism spectrum indicated
an A-form helical conformation for Gs
2
UUUC which was further confirmed by 2D ROESY NMR experiments. The duplex
stability of the three pentamers complexed with a 2
'
-O-methyl-ribonucleotide complementary strand, G
m
A
m
A
m
A
m
C
m
, was determined by UV thermal melting studies and by
1
H NMR spectroscopy. The duplex containing s
2
U has a
T
m of 30.7
o
C compared to 19.0
o
C for the unmodified control and 14.5
o
C for the s
4
U containing duplex. The results from UV experiments were corroborated by imino
proton NMR studies that show proton exchange rates, chemical shift differences, and NH proton linewidths indicative of the stability order s
2
U > U > s
4
U. The magnitude of the effect of s
2
U in our model system is comparable to the 20
o
C stabilization observed by Grosjean and co-workers for 2-thiolation in a codon-anticodon model system composed of two tRNAs with complementary anticodon sequences [Houssier, C., Degee, P., Nicoghosian, K. and
Grosjean, H. (1988)
J. Biomol. Struct. Dyn
., 5, 1259-1266].
INTRODUCTION
Transfer RNAs are unique among the naturally occurring RNAs in that a very high
percentage of the nucleosides are chemically modified (
1
,
2
). Despite the wide distribution of modified nucleosides, with few exceptions,
the functional role of many of the modified nucleosides is unknown. Reviews
published in recent years delineate the potential role of the modified nucleosides in many biological systems (
3
-
5
). Among the uridine modifications, 2-thiouridine (s
2
U) and its C-5 derivatives are found predominantly in the wobble position of tRNAs (
3
). For s
2
U, it has been shown that sulfur substantially stabilizes the 3'-endo sugar conformation at the nucleoside and dinucleotide level (
6
-
9
). It has also been shown that 2-thioribothymidine (s
2
T) stabilizes tRNAs from extreme thermophiles (
10
-
12
), presumably through stabilization of the 3'-endo sugar conformation. Direct evidence for this in intact tRNA
was not obtained, and the stabilizing properties of s
2
U have not been confirmed for oligoribonucleotides mainly due to the synthetic
difficulties involved in site-specific incorporation of s
2
U and s
2
T. Methods for s
2
U incorporation involving base protection add additional synthetic steps to the
phosphoramidite chemistry and therefore have not received wide use. Early
attempts to incorporate s
2
U phosphoramidite without base protection into RNA using the conventional
reagents in the automated solid-phase phosphoramidite method have failed (
13
,
14
). As described previously, the difficulty can be overcome in a straightforward
fashion by using
tert
-butyl hydroperoxide instead of I
2
/water for the oxidation step in solid phase oligonucleotide synthesis (
15
-
17
). In the present study, we have used this method to synthesize oligonucleotides
containing s
2
U and characterized the resulting RNA stabilization using NMR, UV and CD
spectroscopy.
As part of our project to investigate the influence of RNA modification on codon-anticodon interaction, we synthesized the RNA sequence Gs
2
UUUC as the minimal model: the central pyrimidines form the anticodon
trinucleotide modeled after tRNA
Lys
and the GCs at the end were inserted in order to impart reasonable stability for the duplex formed with the complementary strand G
m
A
m
A
m
A
m
C
m
. The selection of the complementary strand G
m
A
m
A
m
A
m
C
m
instead of GAAAC was favored since the 2'-O-methyl containing sequence did not aggregate at the low temperature necessary to form stable duplex, whereas
aggregation was a problem in the case of GAAAC. The 2'-O-methyl containing complementary strand also added stability
to the RNA duplexes (
18
,
19
). The sequence GUUUC was used as a control for comparison with s
2
U, and experiments with Gs
4
UUUC helped to establish that only s
2
U, and not s
4
U, stabilizes Watson-Crick base-pairs in RNA duplexes. The observation that s
2
U modification improves codon-anticodon fidelity in the wobble position of tRNAs (
20
) is explained by our observations of significant duplex stabilization upon substitution of U with s
2
U.
MATERIALS AND METHODS
UV spectroscopy
Samples for UV melting studies contained ~8 * 10
-4
M of each duplex at one of two different salt concentrations, 100 mM NaCl and 1 M NaCl, in 25 mM phosphate buffer, 0.5 mM EDTA, pH 7.0. The
T
m
experiments were done on an HP 8452A spectrophotometer equipped with a 890090
Peltier temperature control unit; the temperature was increased in 1oC increments with a 1 min equilibration time. Measurements were repeated
three times for each duplex, and the melting transitions were determined from
the first derivative of the absorbance versus temperature curve. Thermodynamic
parameters were determined from UV melting curves using the MeltFit program (
21
).
CD spectroscopy
Samples for CD spectroscopy were prepared at a concentration of ~35 [mu]M in 25 mM phosphate buffer, pH 7.0, containing 100 mM NaCl and 0.5 mM
EDTA. The CD experiments were done on a JASCO J-720 CD spectrometer, and the temperature was controlled with a Neslab RTE-110 refrigerated water bath and a flow-through CD cell. The temperature was allowed to equilibrate for 5 min at each temperature.
NMR spectroscopy
NMR experiments for the nucleoside derivatives were done using a Bruker AC200,
200 MHz NMR spectrometer. All NMR experiments on RNA oligonucleotides were
carried out using a Varian Unity 500 MHz NMR spectrometer and a Nalorac ID500
indirect detection probe. The NMR samples contained ~2 mM RNA in 400 [mu]l of 25 mM phosphate buffer, pH 7.0, containing 100 mM NaCl, 0.05 mM
EDTA (for single strands) and in 220 [mu]l of 50 mM phosphate buffer, pH 7.0, containing 200 mM NaCl, 0.1 mM EDTA
(for duplexes). Shigemi (Allison Park, PA) microsample NMR tubes were used for
the 220 [mu]l duplex samples. Samples in D
2
O were prepared by twice lyophilizing, then redissolving the aqueous samples in
99.96% D
2
O and finally redissolving in either 400 [mu]l (for single strands) or 220 [mu]l (for duplexes) of 99.996% D
2
O. 2D TOCSY, ROESY, PECOSY and heteronuclear COSY experiments were collected as hypercomplex data sets using TPPI-States phase cycling (
22
) and processed with Varian VNMR software. The ROESY experiments were collected
using a time-shared spinlock field of ~3500 Hz (
23
,
24
). The F
2
, acquisition, data were collected with 4096 total data points and 400 t
1
increments were zero-filled to 2048 total points in F
1
. PECOSY (
25
) spectra were collected with WALTZ-16
31
P decoupling and a decoupler field strength of ~500 Hz to minimize sample heating. The PECOSY data were collected with a
spectral width of 2300 Hz, centered at the water frequency, with 4096 data
points collected in F
2
domain and 400 t
1
increments zero-filled to 2048 data points in F
1
. The proton-detected heteronuclear COSY (
26
) data were collected with 64 transients per increment, spectral widths of 1300
and 250 Hz in F
2
and F
1
, respectively and 64 t
1
increments; the t
1
data was zero-filled to 1024 points.
Mass spectrometry
The FAB mass spectra were recorded on a Finnigan MAT 95 high resolution gas
chromatography/mass spectrometer with Finnigan MAT ICIS II operating system.
Electrospray ionization mass spectra were acquired using a PE Sciex (Norwalk,
CT) API III
+
quadrupole mass spectrometer.
Synthesis
Thin layer chromatography was performed with Merck silica gel 60F-254 and Merck silica gel 60F (230-400 mesh) was used for flash chromatography. Chemical reagents were
purchased from Aldrich except where noted and used without further
purification. Spectral grade acetonitrile (Baxter) was used for HPLC.
5
'
-O-(4,4
'
-Dimethoxytrityl)-2-thiouridine
(
4
). 2-Thiouridine (3.25 g, 12.5 mmol, 1 equiv.) was dried by evaporation from dry pyridine (2 * 15 ml), then dissolved in 30 ml of dry pyridine under Ar atmosphere. To this was added 4,4'-dimethoxytrityl chloride (5.0 g, 15 mmol, 1.2 equiv.) and the reaction mixture was
stirred overnight at room temperature. Solvent was removed
in vacuo
and then coevaporated with toluene to remove residual pyridine. The residue was
dissolved in 100 ml of methylene chloride and extracted with saturated sodium
bicarbonate followed by sodium chloride. The solvent was dried over anhydrous sodium sulfate, solvent removed
and the residue was purified by flash chromatography, using chloroform:methanol (95:5) to yield 5 g (70%) of compound
4
as a slightly yellow foam. m.p. 95-100oC;
1
H NMR (DMSO-d
6
) 12.70 (s, 1H), 8.00 (d, 1H), 7.30 (m, 9H), 6.90 (d, 4H), 6.50 (d, 1H), 5.60 (d, 1H), 5.40 (d, 1H), 5.20 (d, 1H), 4.10 (m, 3H), 3.70 (s, 3H), 3.30 (m, 2H); FAB MS, found: 561.1708 (calculated for C
30
H
30
N
2
O
7
S
1
, MH
-
: 561.1695).
5
'
-O-(4,4
'
-Dimethoxytrityl)-2
'
-O-(tert-butyldimethylsilyl)-2- thiouridine
(
5
)
.
Compound
4
(2.8 g, 5.0 mmol, 1 equiv.) was dried by evaporation from dry pyridine (2 * 15 ml), then dissolved in 45 ml of dry pyridine under Ar atmosphere. To
this was added imidazole (1.4 g, 20 mmol, 4 equiv.) and
tert
-butyldimethylsilyl chloride (0.9 g, 6 mmol, 1.2 equiv.). The solution was
stirred for 9 h at room temperature. Solvent was evaporated and then further
coevaporated with toluene to remove the residual pyridine. The residue was then
dissolved in 100 ml of methylene chloride and extracted with saturated sodium
bicarbonate followed by sodium chloride. The organic layer was separated and
dried over anhydrous sodium sulfate. The solvent was removed
in vacuo
. The crude product which was a mixture of both 2' and 3' isomers was subjected to flash chromatography. The 2' isomer was separated and solvent evaporated to yield 0.65
g (50%) of compound
5
as a white foam. M.p. 110-115oC;
1
H NMR (DMSO-d
6
) 12.70 (s, 1H), 8.05 (d, 1H), 7.30 (m, 9H), 6.90 (d, 4H), 6.50 (d, 1H), 5.45 (d, 1H), 5.30 (d, 1H), 4.20 (m, 3H), 3.75 (s, 3H), 3.40 (m,
2H), 0.90 (s, 9H), 0.10 (d, 6H); FAB MS, found 675.2526 (calculated for C
36
H
44
N
2
O
7
S
1
Si
1
, MH
-
: 675.256).
5
'
-O-(4,4
'
-Dimethoxytrityl)-2
'
-O-(tert-butyldimethylsilyl)-2- thiouridine-3
'
-(cyanoethyl N,N-diisopropylphosphoramidite)
(
6
). Compound
5
(1.0 g, 1.5 mmol, 1.0 equiv.) was dissolved in 30 ml of dry THF under Ar
atmosphere. To this was added DMAP (0.035 g, 0.3 mmol, 0.2 equiv.), and diisopropylethylamine (0.5 ml, 3.0 mmol, 2.0 equiv.). The solution was stirred while adding 2-cyanoethyl N,N-diisopropylphosphonamidic chloride (0.67 ml, 3.0 mmol, 2.0 equiv.).
After 2 h, another 1.0 equivalent of 2-cyanoethyl N,N-diisopropylphosphonamidic chloride was added and the stirring was
continued for 5 h. The reaction was quenched by adding 100 ml of ethyl acetate,
the reaction mixture was extracted with saturated sodium bicarbonate followed
by sodium chloride. The organic layer was separated, dried over anhydrous
sodium sulfate and the solvent was removed
in vacuo
. The residue was purified by flash chromatography on silica gel to yield 0.83 g
(65%) of compound
6
as a white foam. The material exists as 1:1 ratio of stereoisomers about
phosphorous and two chemical shifts are observed for some of the resonances. The secondary shifts are indicated in parentheses. m.p. 120-125oC;
1
H NMR (Acetone-d
6
) 11.25 (s, 1H), 8.20(
8
.05) (d, 1H), 6.80-7.60 (m, 13H), 6.70 (s, 1H), 5.05(
4
.95) (d, 1H), 4.50 (m, 8H), 4.00 (m, 3H), 3.60 (m), 2.50-2.90 (m), 0.80-1.30 (m), 0.15 (d);
31
P NMR (Acetone-d
6
) 151.12, 150.94; FAB MS, found: 877.3797 (calculated for C
45
H
61
N
4
O
8
S
1
Si
1
P
1
, MH
+
: 877.3795).
Oligonucleotide synthesis
The oligoribonucleotides were synthesized on an Applied Biosystems 394
oligonucleotide synthesizer on a 10 [mu]mol scale using 0.05 M acetonitrile solutions of Perseptive Biosearch
Expedite phosphoramidites; the s
2
U amidite concentration was 0.1 M. Controlled Pore Glass supports were obtained
from Glen Research, all other reagents were standard solutions (Applied
Biosystems, Foster City, CA). Amidites were coupled for 30 min and for the
synthesis of Gs
2
UUUC, the
tert
-butyl hydroperoxide (10% solution in acetonitrile) oxidation step was 2 * 6 minutes. Synthesis of other sequences were carried out following
the standard ABI procedure.
Deprotection
The CPG-bound RNA sequences were transferred from the column to a screw cap glass
vial, to this was added 2.0 ml of NH
4
OH:EtOH (3:1 v/v) and the solution kept at room temperature for 3 h. The supernatant was decanted, the support material washed with an
additional 1.0 ml of NH
4
OH:EtOH and the combined solutions heated at 55oC for 6 h, then lyophilized on a Speed-Vac concentrator. The dried material was treated with 1.5 ml of neat Et
3
N.3HF and the solution stirred at room temperature for 8 h. The reaction was
quenched with 0.3 ml of water and the RNA precipitated by adding 15 ml of n-butanol and allowing the solution to stand at -20oC for 6 h. The precipitated RNA was recovered by
centrifugation and dried
in vacuo
.
Purification and analysis
Scheme 1.
The oligonucleotides were desalted using a Waters C18 Sep-Pak cartridge following the procedure of Khare and Orban (
27
). The G
m
A
m
A
m
A
m
C
m
, Gs
2
UUUC and Gs
4
UUUC oligos were further purified by anion-exchange HPLC as described by Vinayak and co-workers (
28
). The RNA oligonucleotides were analyzed by electrospray mass spectrometry and
by enzymatic digestion followed by HPLC/mass spectrometry (
15
,
29
) and also sequenced by ESI tandem mass spectrometry (
30
).
RESULTS
Synthesis of modified nucleoside phosphoramidites and oligonucleotides
Synthesis of 2-thiouridine (Scheme
1
) was achieved in high yields by the coupling of the bis silyl derivative of 2-thiouracil with 1-O-acetyl-2,3,5-tri-O-benzoyl-[beta]-D-ribofuranose
following the procedure of Vorbruggen (
31
); the hydroxyl protecting groups were subsequently removed with 2 M NH
3
in MeOH. The sugar protection and phosphoramidite synthesis follow standard literature procedures (
32
). The 5'-hydroxyl was then protected by dimethoxytritylation using DMT-Cl to give compound
4
in 70% yield. Treatment of compound
4
with TBDMS-Cl gave a mixture of 2' and 3'-TBDMS derivatives, which on purification by flash
chromatography yielded the pure 2'-isomer compound
5
in 50% yield. Reaction of compound
5
with 2-cyanoethyl-N,N-diisopropyl phosphonamidic chloride for 5 h at room temperature afforded
the desired s
2
U phosphoramidite compound
6
in 65% yield. The amidite
6
was then incorporated into RNA using the modified oxidation protocol as
reported (
15
), which employs
tert
-butyl hydroperoxide instead of the conventional aqueous iodine reagent, for the oxidation step in solid phase automated oligonucleotide synthesis (
33
). The stepwise yield for the s
2
U amidite coupling was >95% as evidenced by trityl assays and verified by HPLC.
Deprotection and purification were carried out following published RNA protocols (
28
,
34
-
36
). Synthesis of the 4-triazolouridine phosphoramidite required to make the 4-thiouridine containing RNA using post-synthetic modification was accomplished by following the reported procedures starting from the
commercially available uridine phosphoramidite (
37
,
38
). The 4-triazolouridine phosphoramidite was incorporated into RNA using the
standard methods and the RNA was then treated with thiolacetic acid followed by
the deprotection using 10% DBU/MeOH (
37
), which yielded the s
4
U containing RNA oligonucleotide. The identity of the modified nucleosides was
confirmed by electrospray mass spectrometry on the oligonucleotide, as well as
by nuclease digestion of the RNA to nucleosides followed by LC/MS (
29
)
NMR resonance assignments
The 1D
1
H NMR spectra and the
31
P NMR spectra of the three oligoribonucleotides GUUUC, Gs
2
UUUC and Gs
4
UUUC are shown in Figures
1
and
2
, respectively. Resonance assignments were made using 2D NMR experiments
recorded at 27oC and 30oC; the data is summarized in Table
1
. The H1' proton of the s
2
U residue in Gs
2
UUUC was easily identified from its characteristic downfield shift due to the sulfur substitution at C2. The H5 and H6 protons of
the uridines were differentiated from the cytidine H5 and H6 protons by their
different coupling constant (for uridines,
3
J
H5-H6
~8.0 Hz and for cytidine
3
J
H5-H6
~7.6 Hz). Also, the H5 resonance of cytidine was shifted downfield, and was
clearly resolved from the uridine H5 resonances and the H1' resonances (Fig.
1
). For Gs
4
UUUC, sulfur substitution at C4 resulted in a downfield shift of the H5
resonance for s
4
U, while the H6 resonance experienced an upfield shift compared to other
residues. The 5' and 3' terminal GCs were identified, respectively, by their
characteristic downfield and upfield shifts of the H2' and H3' protons; it has been reported that the downfield shift of the H2' and H3' resonances of the 5' terminal residue is attributed to the
absence of ring current effect and the upfield shift of the corresponding
resonances of the 3' terminal nucleotide is observed due to the absence of a terminal
phosphate (
39
). On the basis of all these starting points, assignment of the resonances for the individual residues was accomplished by a combination of TOCSY, PECOSY,
ROESY and
1
H-
31
P hetero-COSY experiments. Assignment of most of the resonances for GUUUC and Gs
2
UUUC was achieved; however, extensive overlap of the central uridine resonances
in Gs
4
UUUC prevented complete assignment of this sequence.
Measurement of
3
JH1
'
-H2
4
scalar couplings
It has been well established by Lee and Tinoco that the percentage 3'-endo(N-type) sugar conformation of a nucleoside in RNA is an
indication of the extent of stacking for that nucleoside (
40
,
41
). Increased stacking is also predictive of a tendency to form stronger RNA
duplexes when an oligo interacts with its complement (
42
). We measured the
3
J
H1'-H2'
scalar couplings for all the residues of Gs
2
UUUC from the 1D spectrum at 30oC (Fig.
1
), where all the H1' resonances are resolved. For GUUUC, measurements were made using the 1D
spectra as well as the
31
P decoupled PECOSY experiment; for Gs
4
UUUC, measurements for the terminal GCs were possible from the PECOSY, but this
data had little further utility and therefore we could not make a meaningful
comparison with the other sequences. The
3
J
H1'-H2'
couplings were used to calculate the percentage of the 3'-endo sugar conformation (Table
2
). The limiting value of the sum of the
3
J
H1'-H2'
and the
3
J
H3'-H4'
couplings was set to 10 Hz based on the measurements from the
31
P decoupled PECOSY made at 27oC and 40oC, where both the
3
J
H1'-H2'
and the
3
J
H3'-H4'
cross-peaks were resolved for GUUUC and Gs
2
UUUC. Examination of the data in Table
2
clearly indicates that the 3'-endo conformation for the s
2
U residue is significantly stabilized compared to the comparable uridine in
GUUUC. For the s
2
U residue, the H1' proton also remains resolved throughout the entire temperature range
from 15 to 40oC, therefore the calculation of the percentage of 3'-endo pucker for this particular residue was straightforward at
all the temperatures studied. Below 15oC only a singlet was observed for the H1' proton suggesting that the sugar pucker approaches 100% 3'-endo, although we can only state definitively that the
3
J
H1'-H2'
coupling is less than the 1 Hz based on a comparison of the H1' and H8/H2 linewidths. The high percentage of the 3'-endo sugar conformation measured for the s
2
U residue suggests that s
2
U confers conformational rigidity to the RNA structure throughout the temperature range from 15 to 40oC.
.
Percentage 3'-endo (N) conformer and equilibrium constants (N/S) at 30oC for GUUUC and Gs
2
UUUC
%N
%S
K
eq
. (N/S)
G
UUUC
51
49
1.0
G
U
UUC
42
58
0.7
GU
U
UC
41
59
0.7
GUU
U
C
41
59
0.7
GUUU
C
59
41
1.4
G
s
2
UUUC
62
38
1.6
G
s
2
U
UUC
75
25
3.0
Gs
2
U
U
UC
77
23
3.3
Gs
2
UU
U
C
75
25
3.0
Gs
2
UUU
C
61
39
1.5
The percentage of 3"-endo (N) conformers for the underlined nucleotide was calculated
from %S = 100 * J
1',2'
/(J
1',2'
+J
3',4'
), %N = 100 - %S (56).
2D NMR spectra of GUUUC and Gs
2
UUUC
Although the effect of s
2
U in stabilizing RNA structure can be seen from the sugar conformation
determination, the helical structure of the RNA can be directly determined from
the observation of certain NOE cross-peaks measured in the 2D ROESY NMR experiment. Comparison of the relative
intensities of the cross-peaks from H8/H6 to H1' and from H6/H8 to H2' and H3' reveals the preferred conformation around the
glycosidic bond (
43
,
44
). Comparison of the ROESY spectra (Fig.
3
) revealed that while the residues of Gs
2
UUUC exist predominantly in the anti conformation about the glycosidic bond, a distribution of syn and
anti forms is seen for the residues of GUUUC, as evidenced from the stronger
cross-peak intensities from H8/H6 to H1'. In the case of Gs
2
UUUC, the observation of a strong NOE from H6 to H2' and the absence of an NOE to H1' for the s
2
U residue clearly shows that s
2
U is strictly in the anti conformation. However, for other residues, although
NOEs are observed for both H1' and H2' from H8/H6, the intensities of the cross-peaks to H1' protons are weak compared to H2' protons thereby suggesting the predominance
of the
anti
conformation over the
syn
conformation.
Duplex formation
Pentamer RNA duplexes, containing either three uridines, a single 2-thiouridine and two uridines, or a single 4-thiouridine and two uridines were prepared to assess the intrinsic
effect of the uridine thiolation on base pair formation. The duplexes were
generated by incremental addition of the complementary purine strand to each of
the three pyrimidine strands, with observation of marker signals from both
strands in the
1
H NMR spectrum recorded at 40oC. Although the exactly 1:1 complexes formed Watson-Crick base pairing as indicated by the imino proton resonances at
low temperature, significant broadening in the aromatic region was observed
(spectra not shown); this is possibly because of the tendency of the pyrimidine
strands to aggregate at low temperature and hence a competition between duplex
formation and self-aggregation resulted in the signal broadening. In order to overcome this
aggregation problem, a slight excess of the purine strand was added to all the
three duplexes which significantly improved the quality of the spectra. It has
been reported earlier that in situations similar to the present study
(aggregation of one of the strands in the duplex formation), addition of slight
excess of the non-aggregating strand also improved the quality of the spectrum (
45
). The final concentration of the duplexes used for the NMR studies was 2 mM, in 50 mM phosphate buffer containing 200 mM NaCl, 0.1 mM EDTA, pH 7.0.
Temperature dependence of the imino proton NMR resonances
The downfield imino proton region of the three duplexes as a function of
temperature is shown in Figure
4
. While the s
2
U and s
4
U duplexes exhibited all five resonances corresponding to the five Watson-Crick base pairs, the unmodified duplex control showed only four
resonances due to the overlap of two of the uridine imino resonances. The
resonances for the s
2
U-duplex and uridine duplex were assigned by 1D NOE difference experiments
(spectra not shown). For the s
4
U-duplex, imino resonances from the three Us were nearly overlapped,
therefore the precise irradiation of the individual resonances necessary to
make the NOE assignment was not possible and the assignment was made purely
based on the comparison with s
2
U and U duplexes and from the sequential disappearance of resonances due to
melting effects. The imino resonances from the two terminal G-C base pairs were easily identified from their chemical shift positions
compared to the A-U base pairs and also from the broad nature of the signals. The imino
resonance of the s
2
U-A base pair appeared downfield and the s
4
U-A base pair appeared upfield compared to the corresponding resonance in
the unmodified U-duplex indicating that s
2
U forms a stronger H-bond whereas a weaker H-bond is observed for s
4
U.
UV thermal denaturation studies of duplexes and thermo- dynamic parameters
The UV melting studies were carried out with two different salt concentrations,
100 mM NaCl and 1.0 M NaCl, in 25 mM phosphate buffer, 0.05 mM EDTA at pH 7.0.
The duplex concentration used for melting studies was ~8 * 10
-4
M. The use of such a high concentration was required for these short pentamer
duplexes which exhibit very low
T
m
s at normal UV concentrations of 3-10 [mu]M. Studies on such small duplexes using high concentrations for
T
m
measurements have been reported earlier for pentamer length RNA and DNA homo and heteroduplexes (
46
). These concentrated solutions required the use of 1 mm UV cells, as well as measurements off the RNA [lambda]
max
of 260 nm. Hence, for the s
2
U-duplex, the observation was made at 280 nm, while 270 nm was used for the
U-duplex. For the s
4
U duplex, the [lambda]
max
at 330 nm was in the measurable range and hence was used for the
T
m
measurement. The
T
m
data shown in Figure
5
and the thermodynamic parameters (Table
3
) derived from the
T
m
data allow a quantitative measurement of duplex stability indicating that
compared to the unmodified control RNA duplex (
T
m
= 19.0oC, [Delta]Go
0
= 2.8 kcal/mol), the duplex containing 2-thiouridine stabilizes the RNA duplex structure (
T
m
= 30.7oC, [Delta]Go
0
= 4.8 kcal/mol), whereas destabilization of the RNA duplex occurs upon substitution with 4-thiouridine (
T
m
= 14.5oC, [Delta]Go
0
= 2.2 kcal/mol).
CD spectroscopy
The CD spectra of the three single strands, GUUUC, Gs
2
UUUC and Gs
4
UUUC were recorded at two different temperatures, 10 and 30oC , in 25 mM phosphate buffer, pH 7.0, containing 100 mM NaCl and 0.5 mM EDTA, and sample concentration of 35 [mu]M. The CD spectra shown in Figure
6
for Gs
2
UUUC exhibit a response characteristic for A-form RNA at both temperatures with a large positive band at 265 nm and a negative band near 220 nm. The negative band at 330 nm is characteristic of a thiocarbonyl group at
the C2 position of uridines (
47
-
49
) which also indicates stacking in the single stranded oligonucleotide (
48
). The CD spectra of Gs
4
UUUC and the unmodified control GUUUC have a similar profile with a positive
band at 270 nm and a transition from negative to positive at 255 nm that is
more typical of a B-form than A-form (
46
,
50
). For both GUUUC and Gs
4
UUUC, the amplitude of the positive band at 270 nm is decreased slightly at 30oC, whereas for Gs
2
UUUC the amplitude of the positive band at 265 nm did not change when the
temperature was raised from 10 to 30oC. However, the amplitude of the negative band at 330 nm for Gs
2
UUUC was decreased slightly on raising the temperature suggesting a partial
disruption of stacking at the higher temperature.
DISCUSSION
Uridine thiolation is among the most ubiquitous and conserved modifications
found in tRNAs. 4-Thiouridine is highly conserved at position 8 in eubacterial tRNA, s
2
U and its five modified derivatives occupy position 34, the so-called `wobble base' at the first position of the anticodon and s
2
T is found at position 55 (
1
). In tRNAs having consecutive pyrimidines in the anticodon loop, the wobble
base is modified to either the 2-thio or 2'-O-methyl derivatives (
3
), presumably to provide conformational rigidity. Previous studies to understand the role of nucleoside modification were carried out on the nucleoside and dinucleotide levels and established that
s
2
U preferred a 3'-endo sugar characteristic of A-form of RNA (
7
,
9
). A rigid 3'-endo sugar pucker was associated with wobble restriction of the
codon-anticodon interaction, leading to the `Modified Wobble Hypothesis' for
the s
2
U derivatives found at position 34 (
51
).
ACKNOWLEDGEMENTS
We wish to thank Jim McCloskey for helpful discussions and his group for
assistance with mass spectrometry. This work was supported by NSF grant MCB-9317196 and ACS grant JFRA-405. We acknowledge the NIH for partial support of NMR and the
oligonucleotide synthesis facility (RR0626, CA42014), and the NSF for mass
spectrometry support (CHE-902690).
REFERENCES
1 Steinberg, S., Misch, A., Sprinzl, M. (1993) Nucleic Acids Res., 21, 3011-3015.MEDLINE Abstract
2 Limbach, P. A., Crain, P.F., McCloskey, J.A. (1994) Nucleic Acids Res., 22, 2183-2196.
3 Yokoyama, S., Nishimura, S. (1995) In Soll, D., RajBhandary, U.L. (ed.), tRNA. Structure, Biosynthesis, and Function. ASM Press, Washington, pp. 207-224.
4 Agris, P. F. (1996) In Cohn, W. M., K. (ed.), Progress in Nucleic Acid Research and Molecular Biology. Academic Press, pp. 74-129.
5 Bjork, G. R. (1992) In Hatfield, D. L., Lee, B.J., Pirtle, R.M. (ed.), Transfer RNA in Protein Synthesis. CRC Press, Ann Arbor, pp. 23-85.
6 Yokoyama, S., Watanabe, T., Murao, K., Ishikura, H., Yamaizumi, Z., Nishimura, S. and Miyazawa, T. (1985) Proc. Natl. Acad. Sci. USA, 82, 4905-4909.MEDLINE Abstract
7 Agris, P. F., Sierzputowska-Gracz, H., Smith, W., Malkiewicz, A., Sochacka, E., Nawrot, B. (1992) J. Am. Chem. Soc., 114, 2652-2656.
8 Sierzputowska-Gracz, H., Sochaka, E., Malkiewicz, A., Kuo, K., Gehrke, C.W. and Agris, P.F. (1987) J. Am. Chem. Soc., 109, 7171-7177.
9 Smith, W. S., Sierzputowska-Gracz, H., Sochacka, E., Malkiewicz,A., Agris, P.F. (1992) J. Am. Chem. Soc., 114, 7989-7997.
10 Watanabe, K., Yokoyama, S., Hansske, F., Kasai, H., Miyazawa, T. (1979) Biochem. Biophys. Res. Commun., 91, 671-677.MEDLINE Abstract
11 Davanloo, P., Sprinzl, M., Watanabe, K., Albani, M., Kersten, H. (1979) Nucleic Acids Res., 6, 1571-1581.MEDLINE Abstract
12 Yamamoto, Y., Yokoyama, S., Miyazawa, T.,Watanabe, K., and Higuchi, S. (1983) FEBS, 157, 95-99.
13 Agris, P. F., Malkiewicz, A., Kraszewski, A., Everett, K., Nawrot, B., Sochacka, E., Jankowska, J. and Guenther, R. (1995) Biochimie, 77, 125-134.
33 Gait, M. J. (1990) Oligonucleotide Synthesis-A Practical Approach. Oxford University Press, London.
34 Gasparutto, D., Livache, T., Bazin, H., Duplaa., Guy, A., Khordin, A., Molko, D., Roget, A., and Teoule, R. (1992) Nucleic Acids Res., 20, 5159-5166.MEDLINE Abstract
35 Westman, E., Stromberg, R. (1994) Nucleic Acids Res., 22, 2430-2431.MEDLINE Abstract
36 Wincott, F., Direnzo, A., Shaffer, C., Grimm, S., Tracz, D., Workman, C., Sweedler, D., Gonzalez, C., Scaringe, S. and Usman, N. (1995) Nucleic Acids Res., 23, 2677-2684.MEDLINE Abstract
