3710
J. Org. Chem. 1988, 53, 3710-3722
er/petroleum ether) to give two fractions: (1) dimer 6b (210 mg, 34%) and (2) a mixture (3:2) of 6b-A and 6b-B (350 mg, 36%). The mixture was heated in benzene to 80 OC for 2 h [2.2.1]Bicyclic 6b-A was converted quantitatively into 6b-B. 6b-A: 90-MHz ‘H NMR (CDCI,) 6 6.22, 5.98 (dd, J = 6 Hz, J = 3 Hz, 2 H, ring CH=CH), 5.34 (m, 2 H, CH=CH), 3.72 (s,3H,0CH3),3.27, 2.81 (m, 2 H, 2 bridgehead H), 1.20-3.00 (m, 4 H, 2 CH2), 1.60 (m, 3 H, CH,). cis-Hydrindane6b-B: IR (CHCl,) 3060 (w), 2980 (m), 2940 (m), 2880 (m), 2835 (m), 1710 (vs), 1655 (w), 1635 (w), 1400 (s),1380 (w), 1375 (w), 1345 (w), 1310 (m), 1260 (vs), 1120 (m), 1085 (m), 1035 (w), 995 (w); 90-MHz ‘H NMR (CDCl,) 6 6.86 (m, 1H,CH=C),5.44 (m, 2 H,CH=CH), 3.73 (s, 3 H, OCHJ, 1.8-2.2 (m, 7 H, 2 CH2, 3 CH), 1.20 (d, J = 7.5 Hz, 3 H, CH3); 20-MHz 13CNMR (CDC1,) b 167.4 (s, C=O), 147.3 (d, =CH), 132.3, 130.6
(d, CH=CH), 131.0 (9, C=CH), 51.5 (9, OCHJ, 50.1, 35.3, 33.1 (d, 3 CH), 41.1, 27.9 (t, 2 CH2), 17.7 (9, CH,); mass spectrum (70 eV, room temperature), m / z (relative intensity) 192 (30, M’), 160 (13), 133 (a), 127 (33), 126 ( l l ) , 121 (lo), 115 (la), 91 (13), 67 (23), 66 (100); exact mass calcd for Cl2H1602 192.1150304, found 192.1149426.
Acknowledgment. We are indebted to Dr. L. Ernst of the University of Braunschweig for his NMR work (ref lo), to C. Sabieraj for the FVT experiments, and to Ulrike Eggert for experimental contributions. Our work was supported by the Deutsche Forschungsgemeinschaft and the Fonds der Chemischen Industrie.
Synthesis and Spectroscopic Analysis of Branched RNA Fragments: Messenger RNA Splicing Intermediates Masad J. Damhat and Kelvin K. Ogilvie*J Department of Chemistry, McGill University, Montreal, Quebec, H3A 2K6 Canada Received March 3, 1988 RNA splicing is now established as a major RNA processing reaction in eukaryotic cells. Splicing of messenger RNA precursors generates a lariat RNA structure containing a branched RNA core (Wallace J. C.; Edmons M. Proc. Natl. Acad. Sci. U.S.A. 1983, 80, 950-954) of the type A(2’p5’G)3’p5’X (AGX),where X is a pyrimidine residue. Understanding the mechanism which generates these molecules as well as the role they play in the splicing reaction are central issues in cell biology. Chemically synthesized branched RNA fragments of defined sequence and structure are likely to play a key role in understanding the full biological role of these molecules. Herein we describe the chemical synthesis of a series of trinucleotides (AUu,A G ~ACc, , AGU,AUG,AGc, A C ~ATT, , GUu, aAuu) and a tetranucleotide (UpA’U) with branched linkages. These molecules were fully characterized by UV, NMR (‘H, 13C, ‘lP), HPLC analyses of enzymatic digests, and gel electrophoresis.
Introduction’ In vivo and in vitro studies of the biosynthesis of messenger RNA precursors (pre-mRNA) have established that splicing of these primary transcripts proceeds via a novel RNA form, a lariat RNA.2 This intermediate is an RNA branch with RNA chains linked to an adenosine residue by both 2‘-5‘ and 3’-5’ vicinal phosphodiester linkages. Analysis of branch structures in yeast and higher eukaryotes has indicated that the adenosine-branched nucleoside is usually linked to guanosine and a pyrimidine through the vicinal 2‘-5’ and 3‘-5‘ phosphodiester linkages, respectively. The structure of the branch is A(2’p5’G)3’p5’U (AGu) in the case of adenovirus 2 transcripts2band A(2’p5’G)3’p5’C (AGc)in both @-globin2” and yeast2dactin RNA precursors. In spite of the remarkable advances made in the elucidation of the splicing process, many aspects of this reaction are still poorly understood. The exact mechanism of splice site and branch point selection is unknown as are the role and conformational properties of the lariat molecules. Some authors have speculated that the branch point sequences of the lariat molecules may serve as a recognition signal for achieving accurate splicing.2c,3 Alternatively, the primary sequence and/or the three-dimensional structure of the branch may play a key role (e.g. may possess catalytic activity) in directing and t Present address: J. Tuzo Wilson Research Laboratories, Department of Chemistry, University of Toronto, Erindale College, Mississauga, Ontario, Canada L5L 1C6. f Present address: Vice-president (Academic),Acadia University, Wolfville, Nova Scotia, Canada BOP 1x0.
0022-3263/88/1953-3710$01.50/0
regulating the splicing process. In order to gain a better understanding of the role of lariat RNA, the development of chemical synthesis and the study (e.g. conformational analysis) of branched nucleotide fragments is a prerequisite. In this report we describe the chemical synthesis and structural characterization of a series of trinucleotides and a tetranucleotide with branched linkages. A preliminary report of some of these results has a ~ p e a r e d . ~
Results and Discussion Synthesis of Branched RNA. Branched RNA fragments were rapidly prepared by introducing both 2’-5’ and 3’-5‘ vicinal phosphate linkages simultaneously. The basis of this strategy, as presently applied to the phosphite triester synthesis of oligonucleotides using nucleoside phosphoramidites, is summarized in Scheme 11. The key (1) Abbreviations:
the branched nucleotide sequence X(2’- 5‘Y)3’-
p-5‘Z where X is the branched nucleoside is abbreviated as Xqz. The linear sequences X3’-p-5’Y3’-p-5’Zand X2’-~-5’y2’-~-5’Z are written as XpYpZ and XpypZ, respectively. The branched tetramer U3’-p-5’A(2’-~-5’U)3’-~-5’U is abbreviated as UpA”“ (2) (a) Wallace, J. C.; Edmons, M. R o c . Natl. Acad. Sei. (U.S.A.)1983, 80, 950-954. (b) Padgett, R. A.; Konarska, M. M.; Grabowski, P. J.; Hardy, S. F.; Sharp, P. A. Science (Washington, D.C.)1984,225,898-903. (e) Ruskin, B.; Krainer, A. R.; Maniatis, T.; Green, M. R. Cell 1984,38, 317-338. (d) Domdey, H.; Apostol, B.; Lin, R. J.; Newman, A.; Brody, E.; Abelson, J. Cell 1984,39,611-621. (e) Sharp, P. Science (Washington,
D.C.) 1987, 235, 766-771. (3) Hornig, H.; Aebi, A.; Weissmann, C. Nature (London) 1987, 324,
589-590. (4) Damha, M. J.; Pon, R. T.; Ogilvie, K. K. Tetrahedron Lett. 1985,
26. 4839-4842.
0 1988 American Chemical Societv
J. Org. Chem., Vol. 53, No. 16, 1988 3711
Messenger RNA Splicing Intermediates S c h e m e I"
S c h e m e 11"
-
I. tetrazole 2. I 2 I H f l
(/'-Pr)ZN
k
SI6 R
l a OSI b OSi c OS1 d OS1 e H
B
SI0
(,-PrIpN
2..
b.
U
R .Me R = CE
R R' 3 a O S i Me b OSI CE c OS1 Me
GEL
GBz,NPE CBr T
B
U U
GBz
d OSI Me G B z . N P E e OSI CE f O S I CE CBr g H MeT ~~~a~~~
-
MMTO-~.+A"
M
HO OH
M
T
O
RO-P-0
4a (I
~
A
"
0-P-OR
1 - P r)pN
I
8. R = M M T
N ( / -Pr&
R'
5 a , R = Me b. R =CE M
M
T
o
~
A
B MMTO r ~T
HO
8
A
T
-
AB2
O
ROPO
R"
Bi
a O S i Me AB' b O S I C E AB' c O S I Me A B r
R
d O S i Me e OS1 Me +
N(I-P~)~
osI osI
Me
CE h OSI CE i osI CE
I (/-Pr)2N
j
78, R = Me
AB'
S.R=H
Bz U U U
U U GBZ
AB'
GBz.NPE U GBz GBz
ABZ
GBz.NPE GBz.NPE
ABZ
GBz.NPE GBz.NPE
AB'
CBL
ABZ
GBZ.NPE c B Z
cBz
O S I CE
Me A B ' T I O S I Me GBr.NPE U
k H
b,R=CE
B3
GBz.NPE
cBZ T U
OSi = (tert-butyldimethylsily1)oxy.
reagents used were the nucleoside 5'-phosphoramidites 3 and the 2',3'-diphosphoramidite derivatives 5 and 7. These derivatives were prepared by the reaction of nucleosides 1 with either methyl (2a) or /3-cyanoethyl N,N-diisopropylphosphoramidochloridite (2b) in 60-95 % yields by using established procedures (Scheme I).5-7 The branched trinucleotides AUu,AGG,ACc, AGU,AUG, AGC,ACG,ATT, and GUu (8a-1) were prepared via one-flask procedures using either route A or B (Scheme 11). Thus, 5'-0-(monomethoxytrityl)-N6-benz0yladen0~ine (4a),the corresponding 5'-methyl or cyanoethyl phosphoramidites 3a-g (3.0 equiv) and tetrazole (12.0 equiv) were stirred in anhydrous tetrahydrofuran (THF) for 1to 2 h (route A). Similarly, the adenosine 2',3'-diamidites 5a and 5b, uridine la (3.0 equiv), and tetrazole (8.0 equiv) were reacted under identical conditions (route B). After completion of reactions, collidine and an excess of an aqueous iodine solution were added to oxidize the phosphite triester intermediates. The organic layer was washed with an aqueous solution of sodium bisulfite and evaporated, and the crude products were detritylated with a 0.1 M benzenesulfonic acid (BSA)/acetonitrile solution (0 " C , 10 min). After chromatography on silica gel, the detritylated methyl- and cyanoethyl-protected nucleotides 9 were isolated in 15-3090 and 50-75% yields, respectively. The use of an excess of 5'-phosphoramidite is essential for obtaining good yields of the trinucleotides (route A). When the ratio of uridine amidite 3a to adenosine 4a was changed from 3:l to 2:l the yield of trinucleotide AUU(9a) decreased by 20% (Table I). We have found the ratio of 3:1:12 of reactants 5'-methyl phosphoramidite 3:nucleoside (5) Beaucage, S. L.; Caruthers, M. H. Tetrahedron Lett. 1981, 22, 1859-1862. (6) (a) McBride, L. J.; Caruthers, M. H. Tetrahedron Lett. 1983,24, 245-248. (b) Adams, S. P.; Kavka, K. S.; Wykes, E. J.; Holder, S. B.; Galluppi, G. R. J . Am. Chem. SOC.1983,105,661-663. (7) (a) Usman, N.; Pon, R. T.; Ogilvie, K. K. Tetrahedron Lett. 1985, 26, 4567-4570. (b) Usman, N.; et al. J. Am. Chem. SOC.,in press.
ROP6 b P O R
la
I
I
(/-Pr)2N
N(/-Pr)2
5 a
OSi = (tert-butyldimethylsily1)oxy. S c h e m e 111"
a
OSi = (tert-butyldimethylsily1)oxy.
4a:tetrazole gives the best results for branched RNA synthesis using route A. The experimental conditions employed in these condensations are summarized in Table I. Similarly, the arabinoadenosine-diuridine trimer 10 (aAUU)was prepared by the reaction of the arabinonucleoside bisamidite 7 with the nucleoside la (route B, Scheme 111). A better yield of this nucleotide was obtained when the condensation time was increased from 3 to 16 h (Table I). This result reflects, in part, the marked difference in steric environment of the 2'-position in arabinonucleosides.8 In addition, we found that activation
Damha and Ogilvie
3712 J. Org. Chem., Vol. 53, No. 16,1988 Table I. Preparation of Branched RNA Fragments entry
la
lb 2
3 4 5
nucleoside (mmol)
4a 4a la 4a 4a 4a
(0.25) (0.25) (0.75) (0.25) (0.25) (0.25)
11 12
4b (0.25) 4a (0.25) l a (0.83) la (2.00) 4a (0.25) 4a (0.25) la (0.75) 4a (0.25)
13
9a (0.14)
6 7 8a
8b 9 10
a Estimated
yield by
amidite (mmol)
tetrazole (mmol)
THF
3a (0.55) 3a (0.75) 5a (0.25) 3d (0.75) 3c (0.75) 3a (0.38) 3d (0.38)
1.7 3.0 2.0 3.0 3.0 3.0
2.0
3a (0.75) 3g (0.75) 7 (0.20) 7 (0.50) 3f (0.75) 3b (0.75) 5b (0.25) 3e (0.38) 3f (0.38) 15 (0.30)
3.0 2.8 1.0 4.2 3.0 3.0 2.0 3.0 1.2
BSA
(mL)
time (h)
(mmol)
2.5 0.8 2.0 2.0 1.0 2.5
1.5 1.5 1.5 1.5 1.5
1.5 1.5
2.0
1.2
3.0 1.5 1.5 1.5 1.5
3.5 160 1.0 0.8 2.0 2.0
0.9
2.0
product
38 38 38 38
38 38
2.0
76 38 38 38
%"
AUu 9a AUU9a AUU9a AGG 9f AGG8e AUU8a AUG8c AGu 8d A G 8f~ GUU91 ATT 9k aAuu 10 aAUu 10 ACc 9j AUu9b A U U 9b 8g-j
41 65 68 76
U,AUU 16a
53
0 16
39 24 61 42 27 61 21
13 20 32
31PNMR.
of the arabinoadenosine diamidite 7 with tetrazole was slower than that of its rib0 counterparts 5a and 5b (31P NMR).g The adenosine-diguanosine trinucleotide MMT-AGG(8e) could not be prepared by using guanosine units lacking O6 protection. TLC of the crude reaction of adenosine 4a, guanosine amidite 3c, and tetrazole showed, after the oxidation step, mainly unreacted 4a, fast-moving non-tritylated products, and base-line material. However, when the guanosine amidite 3d bearing 06-p-nitrophenylethyl (NPE) protectionlo was used instead, the corresponding trinucleotide 9f was isolated in 76% yield after the detritylation step. The failure of the former reaction may be ascribed to the phosphitylation of the 06-position of guanine under the phosphoramidite condensation conditions.'l The 31PNMR spectrum of a mixture of 06-unprotected amidite 3c and tetrazole (1:4) in acetonitrile showed complete conversion of 3c to products with peaks a t 128.6 ppm (tetrazoyl intermediate,l* ca. 60%) and a t 139.0-135.2 ppm (06adducts," ca. 40%), confirming the reaction of the tetrazolyl phosphoramidite moiety with the 06-position of guanine. As expected, only the tetrazolylamidite product (127.8 ppm) was observed when the 06-NPE protected amidite 3d was treated under the same conditions conditions. The necessity of the protection of the 06-position of guanine residues has also been reported by Imbach et al.I3and Nielsen et al.I4 in analogous phosphoramidite condensations. The 06-p-nitrophenylethy1 protected guanosine phosphoramidites 3d and 3e were prepared by the route shown in Scheme IV. The 06-protected derivative 12 was prepared from 5',3',2'-tris-O-(tert-butyldimethylsilyl)-N2(8) Damha, M. J.; Usman, N.; Ogilvie, K. K. Tetrahedron Lett. 1987, 28, 1633-1636. (9) Damha, M. J. Ph.D. Thesis, McGill University, Quebec, Canada,
1987. (10) Himmelsbach, F.; Schultz, B. S.; Trichtinger, T.; Charubala, R.; Pfleiderer, W. Tetrahedron 1984, 40, 59-72. (11) (a) Pon, R. T.; Damha, M. J.; Ogilvie, K. K. Nucl. Acids Res. 1985, 13, 6447-6465. (b) Pon, R. T.; Usman, N.; Damha, M. J.; Ogilvie, K. K. Nucl. Acids Res. 1986, 14, 6453-6470. (12) (a) Matteucci, M. D.; Caruthers, M. H. J. Am. Chem. SOC.1981, 103,3185-3191. (b) Froehler, B. C.; Matteucci, M. D. Tetrahedron Lett. 1983, 24, 3171-3174. (13) Huss, S.; Gosselin, G.; Imbach, J. L. Tetrahedron Lett. 1987,28, 415-418. (14) Nielsen, J.; Marugg, J. E.; Taagaard, M.; van Boom, J. H.; Dahl, 0. Recl. Trau. Chim. Pays-Bas 1986, 105, 33-34.
Scheme IV"
SI0
os,
11
HO*
SiO*
i;P
">p,-N/
'l-pr
CI
2a,b SI0
os1
SI0
os1
12
I C
-
3d,e OSi = (tert-butyldimethylsily1)oxy.
benzoylg~anosinel~ (11) in 87% yield via a Mitsunobu reaction according to Pfleiderer et al.1° Selective removal of the 5'-TBDMS group of 12 with an aqueous solution of zinc bromide16gave nucleoside IC in 58% yield. Guanosine (15) Ogilvie, K. K.; Schifman, A. L.; Penney, C. L. Can. J.Chem. 1979,
57, 2230-2238.
(16) Seela, F.; Ott, J.; Potter, D. V. L. J. Am. Chem. SOC.1983, 105, 5879-5886.
Messenger RNA Splicing Intermediates
J . Org. Chem., Vol. 53, No. 16, 1988 3713 Scheme VIa
Scheme Va
""'"VU
1 tetrazole 2 I~r'Hz0-
3 BSA
6 bsi I
/ - P r 2 N/
MeO-P=O
P
OMe
I
\ OMe
u
I
0Sl
SI0
15
OY? si0
' 0 R
Ro*
OSi = (tert-butyldimethylsily1)oxy.
amidites 3d and 3e were prepared by the reaction of IC with either phosphoramidochloridite 2a or 2b, in 78% and 70% yields, respectively. Accordingly, the 06-protected guanosine 4b'0J1bwas used in the synthesis of the guanosine-diuridine trimer 91. The isomers of MMT-A'G (8c)and MTT-AGU(8d) were prepared simultaneously, via route A, by condensing adenosine 4a with the uridine and guanosine methyl phosphoramidites 3a and 3d. The nucleotide isomer mixture of MMT-AUG(8c)and MMT-AGU(8d) was first separated (39%, combined yield) from MMT-AGG(8f, 24%) and MMT-AUU (8a, 16%) by silica gel column chromatography and subsequently detritylated with BSA to give the isomer mixture of A"G (9c) and AGU (9d) in 72% yield. These products could not be separated at this stage. They were, however, fully separated by high pressure liquid chromatography (HPLC) after complete deprotection (vide infra). The isomers MMT-A'G (8h) and MMT-AGc (8i) were obtained from the reaction of adenosine 4a with the cytidine and guanosine cyanoethyl phosphoramidites 3e and 3f. The crude mixture of MMT-ACG(8h), MMT-AGc(8i), MMT-AGG(8g),and MMT-ACc (8j) was detritylated with BSA to give nucleotides 9g, 9h, 9i, and 9j in 32% overall yield. These trinucleotides were cleanly separated after complete deprotection (vide infra). The significant difference obtained in the coupling yields between the methyl and cyanomethyl phosphoramidites (compare entries l b and 10, Table I) appears to be related, a t least in part, to the relative ease of activation of these derivatives with tetrazole. Activation studies of the uridine methyl (3a) and cyanoethyl phosphoramidite (3b) with tetrazole (4 equiv) in acetonitrile indicated that 3a was 98% converted to the tetrazolyl intermediate12 (130.4 ppm) while only 80% conversion of 3b (128.5 ppm) was observed under these conditions (31P NMR). In addition, significant amounts of triesters 13 and 14 were detected in the condensations of adenosine 4a with the nucleoside 5'-(cyanoethylphosphoramidites). For instance, triesters pU (13a) and the 5'-5' symmetrically linked dimer UpU (14a) were obtained (30-40% combined yield) in the condensation of adenosine 4a and the uridine amidite 3b (entry 10, Table I). Similarly, the cytidine byproducts pC (13b) and CpC (14b) were obtained in the condensation of adenosine 4a and cytidine amidite 3f (entry 9, Table I). These products are most likely formed via 2'- or 3'-hydroxy-assisted cleavage of the vicinal phosphite linkage of the 2'-5'- or 3'-5'-linked dimer intermediates (Scheme V). The released terminal nucleoside (ROH) and/or 2-cyanoethanol (R'OH) can then react with the activated nucleoside phosphoramidite present in excess
0I
0Sl
os1
MeO-P=O
I
I
II
MeO-P=O
I
O
I P SI0
"
I OSI
os1
16r R = M M T b,R=H (I
OSi = (tert-butyldimethylsily1)oxy.
in the condensation reaction to yield, after the oxidation step, compounds 13 and 14. 0
Si?
13r,B=U, R=CE
b, B = C B Z ; ~ = c ~ C. B=U. R =
Me
?Si
SIC, b S I
14r, B = C B z , R = C E b.B=U:R=CE c . B = U; R = Me
These side products were also obtained by using the methyl phosphoramidites [e.g. pU (13c) and UpU (14c)l. They formed, however, in only ca. 5% combined yield (31P NMR) and were easily separated during the purification step. As expected, side products did not form (31PNMR, TLC) in the preparation of the trinucleotides AUU(9a and 9b) from the bis(methy1 and cyanoethyl amidites) 5a and 5b (route B, Scheme 11). The branched tetranucleotide MMT-UpAUU(16a) was prepared in 53% yield by the reaction of uridine 3'phosphoramidite 15, Auu (9a), and tetrazole (Table I) and was subsequently detritylated with BSA to yield UpAUu (16b) in 92% yield (Scheme VI). Deprotection and Purification of Branched RNA Fragments. The nucleotides were completely deprotected by consecutive treatment with (i) tert-butylamine to remove the methyl phosphate protecting groups where present, (ii) aqueous ammonia/ethanol to remove the benzoyl protecting groups, and (iii) 1 M tetra-n-butylammonium fluoride/THF to simultaneously remove silyl and (p-nitropheny1)ethyl groups. Nucleotides containing guanosine residues had to be treated with ammonium hydroxide for about 24 h to insure complete removal of the W-benzoyl groups. When the debenzoylation time of AGG was decreased to 6 h, considerable amounts of di- and monobenzoylated products were isolated ('H NMR). After complete deprotection the crude material obtained was
3714 J. Org. Chem., Vol. 53,No. 16, 1988
Damha and Ogilvie
Table 11. Physical Propertien of Free Branched and Linear R N A Fragments
uva nucleotide AUu A41
R/
NMRb
max
min 234 228
261 256 251 254 264 257 258 257 264 261 260
P-3' -0.14 -0.53 4.61 4.69
234 227 238 228 228 230 234 232 232
260
232
258
230
-0.66 -0.66
-0.69 -0.63 4.94 -1.21
4.79 4.85
P-2' -1.68 -1.27 -1.53 -1.31 -1.63 -1.38 -1.73 -1.10 -1.50 -1.56 -1.22
-0.66 -0.66
-0.75
A
F
G
EMdR.
0.01
0.05
0.37 0.31 0.32 0.29 0.43 0.35 0.37 0.34 0.49 0.40 0.34
0.81 0.72 0.72 0.14 0.76 0.73 0.73 0.78 0.81 0.18 0.19
0.78 0.71 0.11 0.61 0.77 0.71 0.71 0.11 0.48 0.85 0.72
HPLC tn,' 12.7 12.4 13.2 12.8 12.1 11.8 13.3 10.4 15.1 12.8 11.0
0.09
0.41
0.19
0.67
13.3
0.07
0.34
0.76
0.66
10.2
0.07
0.37
0.11
0.66
10.2
0.05 0.05 0.04 0.11 0.06 0.06 0.06 0.14 0.08
4 74
UPAS
258
'
229
-0.72
4.19
@HPO.bSample conditions described in the Experimental Section. 'Cellulose TLC,solvents, A, F. G. dElectmphoretic mobility (24% polyacrylamide/8 M urea gel) relative to xylene cyanol (XC)EM R. = 0.00 and bromophenol blue (BPB) EM R. = 1.00. 'HPLC condition
A.
4 (¶,.In
R l Inin1
Figure 1. HPLC analysis of the crude mixture of AuG and AGu after comnlete denrotection and desaltine on e Senhadex G-25 column. insert: HPLC analysis of fractions a, b. and c of a mixture of AUc and AGu eluting, in this order, from a Sephadex G-25 column (Experimental Section). HPLC condition A (Experimental Section). ~
~
~~~.~ ~
~~
~~~~
~~~~~~
~~~~~~~~~~~
~
~~~
~~
~~~~~~~~~~
~
~~
passed through a column of Dowex Na+ ion exchange resin and subsequently desalted by size exclusion column
chromatography (Sephadex G-25) .and cellulose TLC. Milligram quantities (5-35 mg) of the fully deproteded nucleotides were obtained. The isomeric mixture ACcand AGCwas separated from AGOand ACc by chromatography on cellulose TLC. Each of the isomer mixtures Auc, Acu and ACc, AGCwas separated into their respective components by HPLC using a reverse-phase column (Figure 1). The purity of all fully deprotected nucleotides was checked by analytical polyacrylamide gel electrophoresis, cellulose TLC, HPLC, and NMR (lH, I3C, 31P)spectroscopy. All these methods indicated that the nucleotides isolated were homogeneous. The 31PNMR, HPLC, UV, TLC, and gel electrophoresis data of the branched and some linear nucleotides are collected in Table 111. It is worth noting the following. (i) In the series AUU,AUc, Acu, and ACc the HPLC t g values follow the order Acu < Auu < ACc < A"c, Similarly, in the series Acc, Acc, Acc, and AGOthe HPLC tR values follow the order Acc < Acc < AGO< Ace. Therefore, under our HPLC conditions, adenosine-branched nucleotides containing 3'-5'-linked guanosine residues elute more slowly than those containing 3'-5'-linked pyrimidines. (ii) The branched trinucleotides exhibit base4e.g. ATn Auu) and structure(e.g. aA", Auu) specific mobility under
xc
(.*a-
-
- 0 . .
-
.*
*
e
- e-
BPB
1
2
3
4
5
h A
2
0
CL
3
4
5
WQYQ
6
B
Figure 2. Electrophoresisof branched RNA fragments on a 24% polyacrylamide/8M urea gel. A lane 1, ADc;lane 2, AcG i ADc; lane 3, Ace; lane 4, ACc;lane 5, ADG;lane 6, A'u; lane 7. ACu, lane 8. Auc + AGu; lane 9, AuG; lane 10, UPAPG; lane 11, U P A S . B lane 1, UpApU; lane 2, AUu;lane 3, AT^; lane 4, Act; lane 5, AD&lane 6, aAuu; lane 7, GUu;lane 8, UpAUu;lane 9, ApGpApUpC. All samples were loaded with xylene cyanol (XC)and bromophenol blue (BPB).
Messenger RNA Splicing Intermediates
J. Org. Chem., Vol. 53,No. 16, 1988 3715
Table 111. IH N M R (400 M H z ) Spectral Data of Free Branched RNA Fragments"
AUU AUG
AGU AGG ACG
'4% ACC ATT
GUU A U U
UPAUu -PAUu
5'U-
H 1' 6.21 7.3 6.07 7.5 6.19 5.7 6.10 6.4 6.15 7.1 6.16 6.1 6.16 7.3 6.19 7.4 6.06 7.0 6.50 3.9 6.25 5.9
A H2 8.00
H8 8.31
8.03
8.30
7.86
8.23
7.79
8.21
7.95
8.30
7.79
8.22
7.94
8.28
8.00
8.33 7.99
8.16
8.30
8.16
8.44
H1' 5.96 5.5 5.90 5.4 5.92 4.4 5.88 5.5 5.90 6.2 5.91 4.4 5.94 4.5 6.30 6.9 6.9 5.97 4.4 5.92 4.2 5.97 4.6 5.72 5.4
3'p5'B H5(2) 5.88
H6(8) 7.88 8.02
5.84
7.87 8.01 8.02
5.96
7.84
6.01
7.88
1.85
7.69
5.92
7.93
5.88
7.89
5.92
7.92
17.0
Figure 3. 2D 'H-RELAYED-COSY (300 MHz) spectrum of Auu, with the 'H NMR spectrum shown along the top axis. Solid and broken lines indicate the COSY and RELAYED-COSY off-diagonal resonances, respectively. our electrophoretic conditions as shown in Figure 2. The trinucleotides A U G , A G ~ACG, , and AGChad, however, the same mobility. The linear trimers UpApU, UPAPG, and UpApG had slower electrophoretic mobility than the cor-
H6(8) 7.36
5.66
7.37
7.65
5.61 4.4 5.54 0.0 6.00 6.5 5.3 5.71 2.4 4.69 4.2 5.63
5.74 5.74
7.34 7.34 7.61
5.74
5.34
1.78
7.23
5.79
7.69
5.71
7.44
5.57
7.41
3.0
5.81
7.70
01)
-6.0
2'p5'C H5(2) 5.65
7.72
1
"Sample conditions described in the Experimental Section; 3JH1,-~2. (and 3&1-H2!,, chemical shifts.
- 5.0
H1' 5.60 2.8 5.60 2.8 5.63 4.9 5.60 4.8 5.55
in the case of ATT) values are shown below the HI'
EO
3
1
i o
PPm
Figure 4. Pre-steady-state NOE measurements on AUUat 37 O C : (a) 'H NMR (200 MHz, with line broadening) spectrum between 8.5 and 3.0 ppm; (b-d) difference spectra (on-resonance minus off-resonancepreirradiation) after presaturation for 5.0 s of (b) the proton resonance at 8.31 ppm (peak AH&, (c) the proton resonance at 7.88 ppm (doublet BHB),(d) the proton resonance at 7.36 ppm (doublet The assignments of the proton resonances are given in Table V. The experimental conditions are described in the Experimental Section. responding AUu, AUG, and A G U branched trimers. ATT exhibited lower electrophoretic mobility than the pentanucleotide ApGpApUpC. (iii) AUG and AGuhave slightly different elution behavior on a Sephadex G-25 column; on this column AUG was eluted more slowly than AGU(see insert, Figure 1). Structure Determination of Branched RNA Fragments. Nuclear Magnetic Resonance. The general procedure can be illustrated by the structure determination of AUU. The first step consisted of the identification of the three ribose spin systems A, B, and C with 2D-J correlated spectroscopy ('H-RELAYED-COSY," Figure 3). Since the chemical shift dispersion of some proton reso(17) (a) Bolton, P. H.; Bodenhausen, G. Chem. Phys. Lett. 1982,89,
139-144. (b) Eich, G.; Bodenhausen, G.; Emst, R. R. J. Am. Chem. SOC. 1982, 104, 3731-3732.
Damha and Ogilvie
3716 J. Org. Chem., Vol. 53, No. 16, 1988
nucleotide AUU
AGG
ACc ATT
GUU aAUu UPAU”
Table IV. lSC N M R 1‘ A 90.2d 91.7 -2’pU 91.4 -3pu A 90.0d -2’pG 89.9 -3’pG 89.7 A 90.ld -2’pC 92.4 -3’pC 92.1 A 90.I d -2‘pT 87.0 87.9 -3’pT G 89.7d -2’pU 91.7 -3’pU 91.2 aA 86.4d -2’pU 91.4 91.6 -3’pU 91.5 5‘pU A 88.1d -2’pU 91.6 -3’pU 91.1
(75.4 MHz) Spectral Datan of Free Branched R N A 2‘ 3‘ 4’ 5’ 5 78.3q 78.0s 88.8d 64.4 70.8 84.1d 66.0d 104.8 77.0 76.4 72.6 86.0d 105.4 68.0d 78.29 77.49 87Bd 64.0 76.6 72.2 85.1d 66.8d 76.3 73.1 86.3d 68.2d 78.19 77.89 88.6d 64.3 98.4 77.3 70.3 83.5d 65.6d 76.8 72.2 85.5d 67.7d 99.2 78.3q 77.99 88.5d 64.3 14.3 71.7 87.1d 66.5d 41.1 73.6 88.2d 68.4d 42.0 14.5 77.6q 77.59 88.2d 64.0 76.9 71.0 84.3d 65.gd 104.8 76.3 72.4 86.0d 67.7d 105.3 80.09 81.1s 86.gd 64.0 76.6 71.9 85.2d 67.1d 105.0 105.2 72.3 85.7d 67.8d 76.4 76.2d 86.1d 63.2 105.0 75.5d 78.7, 76.7q 85.59 67.7d 76.8 71.0 84.6d 66.2d 104.6 76.4 72.4 86.0d 67.7d 105.3
Fragments 6
2 155.1
8
143.9
143.2 144.3 154.8 155.0
143.5 139.3 139.7 143.8
155.0
b
143.0 144.2 139.3 140.1 141.1 143.1 144.1 155.6
144.2
155.8
142.6
143.6 144.3 144.1 143.1 144.1
nSample conditions described in the Experimental Section. bSignal not observed due to exchange of H8 signal with deuterium. dDoublet. branched residue consisted of two sets of doublets due to geminal and vicinal coupling to 31P2‘and 31P3’.
q Quartet; C2’ and C3’ of central
nances was limited (300 MHz) some proton-proton connectivities could not be established from the ‘H-RELAYED-COSY spectra alone and it was therefore necessary to make use of ‘H-13C heteronuclear 2D NMR18 (spectrum not shown). Secondly, the bases of each sugar residue were assigned via one-dimensional ‘H-lH nuclear Overhauser enhancement (NOE) measurements (Figure 4). Therefore, the ‘H spin system A belongs to adenosine, and B and C to the uridine residues. In the third step, the scalar coupling connectivities between the ‘H spin systems and the 31PNMR resonances were identified using 1H-31P heterodec~upling’~ (Figure 5). The 31PNMR spectrum of AUUshowed two well-resolved proton-decoupled 31P signals (I and 11,see insert Figure 5). Figure 5a shows the proton NMR spectrum (400 MHz, sugar proton region) of AUU. Figure 5b shows that the spectral pattern of A(H3’,H2’) and B(H4’,H5’,H5”) has been perturbed by irradiation a t peak I of the 31Presonance. The perturbation of the AH2’ signal is due to a four-bond phosphorusproton coupling with 4JH21-p3r = 2.2 Hz (vide infra). Therefore, the 31Presonance I and ‘H spin system B are assigned to the 3’-~-5’-uridineresidue. When peak I1 was irradiated AH2’ and C(H4’,H5’,H5”) was perturbed (Figure 512). Therefore, the 31Presonance I1 and ‘H spin system C are assigned to the zr-p-5’-uridine residue. Clearly, the combination of these NMR techniques allowed for the unambiguous characterization of this nucleotide, i.e., presence of vicinal 2’-5’ and 3’-5‘ phosphodiester linkages and assignments of lH,13C, and 31P NMR resonances (Tables 11-IV). All branched trinucleotides were characterized in this fashion and the following general characteristics were observed. (i) Without exception, the nucleotide residue attached to the central nucleoside via the 2‘4‘ linkage (2’-residue) exhibits ‘H, 31P,and 13C resonances a t higher field than that attached through the 3’-5’ linkage (3’-residue). (ii) The base and anomeric proton chemcical shift (6) values of the 2’ residues are primarily influenced by the (18) (a) Maudsley, A. A.; Muller, L.; Ernst, R. R. J.Magn. Reson. 1977, 28, 463. (b) Bodenhausen, G.; Freeman, R. J. Ibid. 1977, 28, 471. (19) Cheng, D. M.; Kan, L. S.; Iuorno, V. L.; Ts’o, P.0. P. Biopolymers
1984,23, 575-592.
t
*
/I
12
LO
I S
40
86
w-
Figure 5. Insert: proton-decoupled 31Presonances of AUU(I = -0.74 ppm, I1 = -1.68 ppm) obtained at 121.5 MHz: (a) lH-NMR (400 MHz) spectrum of AUUbetween 5.4 and 3.3 ppm (H2,,H3,, H4,, H5,,H5”,proton region); (b) same as (a) except an irradiation power was applied at peak I; (c) same as (a) except an irradiation power was applied at peak 11; (d) same as (a) except that an irradiation power was applied at both peaks I and I1 (broad band decoupling). The arrows indicate the re ion of changes in the spectral pattern due to decoupling. I, fP resonance of 3’p5’uridine residue; 11, 31Presonance of 2’p5’-uridine residue. The sugar spin systems A, B, and C correspond to the adenosine,
3’p5’-uridine, and 2’p5’-uridine residues, respectively.
nature of the central nucleoside (i.e. adenosine or guanosine), and vice versa. For instance, the 6 values of both the adenine and 2’-guanine in AGC,AGU,and AGGare very similar. In contrast, the 6 values of the 2’-uracil protons in A U U and GUUdiffer substantially. (iii) The base and anomeric proton 6 values of a given 3’ residue are nearly independent of the nature of both the central nucleoside and 2’ residue. For example, the 3’guanosines in ACG,AUG,and AGGhave nearly the same 6 values. Similarly, the 6 values of the 3‘-uridines in GUU, ACTU, aAUU,and AGUare virtually the same. (iv) The 3J~1t-H2t values of adenosine residues containing 2’-5’-linked guanosines are 1-1.5 Hz smaller than those containing 2’-5’-linked pyrimidines (e.g. 3JH1t-H2t of AGU, AGc, A G ~
