L-DNAs as potential antimessenger oligonucleotides: a reassessment

Share Embed


Descrição do Produto

Nucleic Acids Research, 1993, Vol. 21, No. 18 4159-4165

L-DNAs as potential antimessenger oligonucleotides: a reassessment Anna Garbesi, Massimo L.Capobianco, Francesco P.Colonna, Luisa Tondelli, Federico Arcamone1, Giorgio Manzini2, Cornelis W.Hilbers3, Jan M.E.Aelen3 and Marcel J.J.Blommers4 Istituto ICoCEA-CNR, Via della Chimica 8, 1-40064 Ozzano Emilia (BO), 'Menarini Ricerche Sud, via Sette Santi 3, 1-50131 Firenze, 2Dipartimento di Biochimica, Biofisica e Chimica Macromolecolare, Universita di Trieste, via Giorgieri 1, 1-34127 Trieste, Italy, 3Laboratory for Biophysical Chemistry, Toernooiveld, 6525 ED Nijmegen, The Netherlands and 4CIBA-GEIGY AG, Physics, PO Box, CH-4002 Basel, Switzerland Received July 16, 1993; Revised and Accepted August ',0, 1993

ABSTRACT Unnatural L-2'-deoxyribonucleosides L-T, L-dC, L-dA and L-dG were prepared from L-arabinose and assembled, by solution or solid phase synthesis, to give L-oligonucleotides (L-DNAs), which contain all four natural bases. The affinity of these modified oligomers for complementary D-ribo- and D-deoxyribo-oligomers was studied with NMR, UV and CD spectroscopies and mobility shift assay on native PAGE. All experimental results indicate that L-DNAs do not, in general, recognize single-stranded, natural DNA and RNA. Hence, contrary to previous suggestions, it is not possible to envisage their use as wide scope antimessenger agents in the selective control of gene expression.

INTRODUCTION Much effort has been devoted in recent years toward the use of oligonucleotides as valuable tools in molecular and cell biology and as potential therapeutic agents (antisense methodology)('). Short, synthetic nucleotide chains, complementary to a segment of double-stranded DNA (anti gene agents) or single-stranded RNA (anti messenger agents), can block selectively the expression of a gene, by specific binding to the target nucleic acid(2). One of the many challenges to be met for the development of oligonucleotides as pharmacologically interesting molecules is the sensitivity of the native phosphodiester bond to nucleases, which leads to fast degradation inside cells. This problem has been addressed mainly by chemical synthesis of oligonucleotide analogs(3), having modified internucleosidic linkages (i.e. phosphorothioate and methylphosphonate) or modified sugars (i.e. 1 '-a-anomers and 2'-substituted riboses), which are more resistant to enzymatic digestion. Little attention has been paid to L-oligonucleotides, where the natural D-ribose is replaced by its enantiomer L-ribose. When this work was started, only four

papers(4'5'6'7) had appeared in the literature concerning the nuclease sensitivity and/or the affinity of L-oligonucleotides for complementary DNA or RNA. While an extremely high stability to spleen and venom phosphodiesterases was uniformnly described for L-ribo- and L-deoxyribo-oligomers, a less clear picture emerged as to their interactions with natural sequences. Tazawa et al.(4), on the basis of U.V. and C.D. measurements, reported the formation of a complex between L-(ApA) and poly U, having a lA:2U stoichiometry and a melting temperature slightly lower than that of the analogous complex formed by D-(ApA). Anderson et al.(6) failed to see, by U.V. spectroscopy, any temperature dependent interaction between poly dA and either L- or D-(dU)18. Van Boeckel et al.(7) presented evidence, from IH-NMR spectroscopy, that L-r(CAAGG) binds to D-ribopentamers having a complementary sequence in both parallel and antiparallel orientations. Since the L/D hybrid observed in both cases, which does not show Watson-Crick base pairing, is more stable than the corresponding D/D antiparallel duplex, the authors suggested that 'the association of nucleotide strands of opposite handedness provides a mechanism for chiral amplification' of a small initial excess of D-ribonucleotides, ultimately leading to the evolution of this enantiomeric form. Given the interest of this hypothesis for prebiotic evolution, we thought it worth to test if a comparably stable L/D hybrid could be formed also by short complementary L-DNA and D-DNA. This experiment would also contribute information expedient to our main purpose i.e. to enquire about the ability of L-oligonucleotides, derived from either L-ribose or L-2'-deoxyribose, to function as effective antisense compounds. In this respect, the result of van Boeckel et al.(7) is rather discouraging, indeed, since it indicates that the binding of a L-oligomer to a natural sequence can occur by a mechanism which is less selective than canonical Watson -Crick base pairing. While the present work was in progress, four reports(8'9' 10. 1) appeared in the literature on the affinity of Loligonucleotides for natural sequences; in all cases, however, only

4160 Nucleic Acids Research, 1993, Vol. 21, No. 18 the behaviour of L-homo-oligomers was studied. Nevertheless, the experimental results led some of the authors to suggest that 'enantio-DNA recognizes complementary RNA specifically but not complementary DNA'(10) and 'L-RNA effectively recognizes the natural-configuration RNA sequence'("1). The result of our study demonstrates that, at least, the first of these statements is, as a general one, wrong, since mixed-sequence L-2'-deoxyoligonucleotides do not bind complementary D-RNA.

RESULTS Modeling calculation At the beginning of the present investigation, a molecular modeling study was undertaken with the aim to compare potential energies and other relevant features of putative L/D duplexes with those of natural D/D ones. To this end, we combined L-d(CGTTCC) with complementary DNA and RNA sequences and created three families of artificial heterochiral duplexes: B-type DNA/DNA, A-type DNA/RNA and ladder DNA/DNA. It was further assumed that 1) the bases of the interacting strands would pair according to Watson -Crick scheme in both antiparallel and parallel arrangements and 2) that in helical complexes one of the coils adopts (as initial guess) an A- or B-type structure. Ladder models, in which both strands loose their helical property, as proposed by van Boeckel et al., on the basis of modeling and energy minimisation 7), tend to have calculated potential energies significantly higher than helical duplexes. One of the reasons is obviously connected to the reduced intra strand phosphatephosphate distances. Using a distance-geometry approach (see Materials and Methods), a set of 40 different duplexes which fulfil the starting prerequisites was generated. The most common feature of the four classes of helical hybrids (see data in Table 1) is that the adoption of a right handed structure by the L strand does not induce severe strain. When the duplexes are subjected to molecular mechanics/dynamic simulations they do not dissociate and, surprisingly, the potential energy of the A-type antiparallel complex is the same as that of the corresponding natural duplex. This does not happen for the other L/D hybrids. Thus, the calculation suggests that L-DNA has a preference to bind natural RNA over DNA and, furthermore, that the resulting duplex should be formed under ordinary conditions. After hybridization studies had disproved this forecast, we went back to the molecular modeling calculation to check if, despite its inherent, strong limitations, it contained clues to factors amenable to the experimental finding. One such clue may possibly be found in the conformational heterogeneity of the docked L strand, as

manifested by the average pairwise root mean square deviation (RMSD) of each of the central L-T residues(12) under the assumption that structures with low RMSD values have low, unfavorable, conformational entropies. This parameter was calculated for all available structures. The Sconf values in Table 1 are the average RMSD values which result from 90 pairwise superpositions of the L-T residues. The lower Sconf value found for B-type versus A-type natural duplex may, in fact, indicate that the conformational space is more restricted in the first case, in agreement with experiment. The lowest Sconf value is found for the antiparallel L/D A-type duplex, precisely the one with the most favourable calculated potential energy. That this is in fact one of the reasons why the hybrid duplex is not formed remains to be proved, since the above reported finding may be accidental.

NH2 u 0 N

0

HN H2N "la~,

OH

0

H

L-dC

N N

O

H L-dG (Y-1 /%)

(Y-62X)

L-2'-deoxy-5'-3'-di-O-benzoyluridine (DIBU)

NH2 N

N

fH

\ KNNO

OH

H L-dA (Y-25%)

L-T

(Y-49X)

Figure 1. Yields of the four L-nucleosides from DIBU.

Table 1.

Etot B-type helix anti parallel D-D anti parallel D-L parallel D-L A-type helix anti parallel D-D anti-parallel D-L parallel D-L

Ebon

Eang

Edih

Ehbo

EvdW Ecoul Scont

-105 4 -96 4 -90 4

43 50 52

101 108 104

-541 -132 420 -568 -134 443 -568 -134 443

1.02 0.92 0.96

-104 4 - 105 5 -99 4

44 51 51

111 121

-628 -134 500

-804 -136 659

122

-771 ---133 628

1.17 0.85 1.07

Energy of the more stable models in Kcal/mol. E,0,t = total energy, Ebon= bonding energy, Ed'h = dihedral energy, Ehb() = H-bond energy, E,,dw = van der Waals energy, EcQUI = Coulomb energy, SC,,l = RMSD of L-T residues (see text).

a B=T b B=CDPA c B=GDPA

d B=ABz

1 a-d Figure 2. The four L-phosphorarnidites.

Nucleic Acids Research, 1993, Vol.

21,

No. 18 4161

3' end. Following deprotection, all 2'-deoxy-oligomers were Synthesis by ion exchange column chromatography and converted purified The L-2'-deoxynucleosides L-T, L-dC, L-dA and L-dG were salts. The ribo-oligonucleotide 7 was a certified sodium the into 1 prepared, as reported elsewhere( 3), from the common (Paris). GENSET from product intermediate 3',5'-di-O-benzoyl-2'-deoxy-L-uridine (DIBU), which was obtained, on a large scale, from commercial LHybridization studies described bv fcllnwinc nrnhince, five-qten svnthesis IkIU 111oWthe Xit auI al -.v_7W sosv___ 3.11s_%-3L As pointed out in the introduction, the starting aim of the present Holy1(14) The synthetic procedure(13) is summarized in Figure j investigation was to verify if a short L-DNA strand interacts (yiel ds are in respect to DIBU). strongly with complementary D-DNA sequences, as reported to Thie protected L phosphoramidites la-d were prepared by occur with 'random' sequence L-RNA/D-RNA combination(7). standlard procedures(15-17). The diphenylacetyl protecting group 1:1 mixtures of L-d(CGTTCC) with either 4 (antiparallel (DP A) for the amino functions of cytosine and guanosine was or 5 (parallel complement) were prepared in 0.1 complement) in less than removal chos'en after having verified its complete M Tris buffer (pH 6.7) containing 10 mM MgCl2 and the two hours under the usual deblocking conditions (30% aq. change of UV absorbance with temperature of the resulting NH4IOH at 50°C for 16 hours). The compounds 1 a-d (see solutions was monitored between 5 and 70°C. No evidence of FiguIre 2) were more than 96% pure by 31P-NMR spectroscopy. melting behaviour was found, since absorption increased a study: ie following sequences (5'- 3') were used in the present linearly with an overall change lower than 4 % (data not shown). Under the same conditions, the duplex formed by D-d(CGTT2):L-d(CGTTCC) CC) with its antiparallel complement 4 gave a quite sharp melting 3):D-d(CGTTCC) profile, with Tm = 25°C (data not shown). The possible 4):D-d(GGAACG) interaction between the L-hexamer and the two complementary 5) D-d(GCAAGG) natural sequences was further investigated by 1H-NMR in 6) D-d(AGTAGTCAAGCTCGGACC) D-r(AGUAGUCAAGCUCGGACC) H20/D20 9/1 at low temperature, the experiment that had 7) shown the formation of a stable hybrid between RNA strands 8) D-d(GGTCCGAGCTTGACTACT) (D-ap) of opposite handedness, as guessed from the protection of imino L-d(GGTCCGAGCTTGACTACTA) (L-ap) protons from exchange with water(7). In our case, however, the 10) L-d(TCATCAGTTCGAGCCTGGA) (L-p) proton spectra (not shown) of the 1:1 mixtures of 2 and Hex;amers 2-5 were prepared by the hydroxybenzotriazole phoscomplementary 4 or 5 in 0.2 M KCl at pH 6.8 and 5°C did not phot iester method in solution and deprotected accordingly(18). any evidence of association between L- and D- oligomers. present by automated obtained were 8-s10 6, TI he longer oligomers These findings prove that short DNA oligomers of solidI-phase synthesis, using the phosphoramidite chemistry, and complementary sequences but opposite chirality do not form depr*otected by the standard proceduresh9r. Sincecohemmercial hybrids stabilized by base stacking and/or hydrogen bonding. deri vatized supports were used in all cases, the L sequences 9 the The next experiments were done with mixed-sequence L at and 10 carry an irrelevant extra D-nucleoside (underlined) oligomers with length in the range normally required for inhibiting mRNA expression with antisense oligonucleotides. As 'sense' strand, we chose a segment of the junction region of t(14; 18) translocation, which gives origin to an oncogene, whose A/A(max) expression can be selectively inhibited by the natural antisense ..................... oligomer 8(20). The hybridization ability of the L strands 9 and -- -_---- -10 toward complementary natural DNA 6 and RNA 7 was probed | by UV and CD spectroscopies and by non denaturing polyacrylamide gel electrophoresis. For UV melting experiments, I equimolar concentrations of 6 (DNA) or 7 (RNA) and complementary 8 (D-ap), 9 (L-ap) and 10 (L-p) were combined t , in 0.1 M Tris buffer (pH 7) and 0.1 M NaCl. The absorbance gf , ---------------------profiles versus temperature of the resulting solutions are shown : 3 * in Figure 3 (DNA/DNA) and 4 (RNA/DNA). Sharp hyperchromic transitions are observed for the natural DNA/DNA ,* -----------------.andRNA/DNA duplexes, with Tm values of 67 and 69°C, respectively. On the contrary, when the natural strands are *f combined with their antiparallel (L-ap) or parallel (L-p) modified a steady increase of absorbance is observed, without I complements 3 detectable cooperative transition. Hence, if complexes any 80 70 60 0 50 40 30 10 20 involving base stacking are formed between mixed-sequence complementary strands of opposite handedness, their stability T (Celsius) lower than that of the corresponding must be more than + ~ DNA/Lap DNA/Lp DNA/L-ap D~~ NA/L-p natural This picture is not changed when melting experiments are done at lower buffer (0.01 M Tris) and increasing NaCl (0.1, 1 and L1s1%

9)

-

--------

duplexes.- 70°C

DNA/D-p

Figuire

3. Normalized absorbance profile (at 260 nm) vs temperature of the with the complementary DNA in Tris 0.1M,

mixtures of natural and L-sequences NaC 'IO.V. IILIVII M. DH i,%a%-,i PJLX 7.0. /.V.

M) concentrations with L-DNA/D-DNA combinations and at 0. 1 M buffer, 1 and 4 M salt concentrations with

4

L-DNA/D-RNA.

4162 Nucleic Acids Research, 1993, Vol. 21, No. 18

A/A(max)

7. OOOE+0 S

1 k

- k- -

....

----

0,95 /

0,85 0,8

-----

0

10

20

/ 30

40

50

60

70

iy

80

T (Celsius) + RNA/D-p : RNA/L-p

ORNA/L-ap

Figure 4. Normalized absorbance profile (at 260nm) vs temperature of the mixtures of natural and the L-sequences with complementary RNA in Tris 0. IM, NaCI O.IM, pH 7.0; 0.O1M spermine was present in the solutions containing the modified sequences.

7. OOOE+0 1

-5 .OOOE+O1 ** ... *.*.,, 220.0 WL i: D-DNA/0-ONA T-5C 2: O-ONA/0-ONA T-75C 3: ----D-DNA/L-ONA ap T-SC 4: ------------ O-DNA/L-ONA p T-5C

[nmJ

3±0.0

Figure 5. Circular dichroism spectra of the mixtures of natural and L-sequences with the complementary DNA in Tris 0. IM, NaCl 0. 1M, pH 7.0. Signal after melting is also reported for the natural duplex.

The lack of association between D and L strands was confirmea by circular dichroism, a technique that, being very sensitive to stereochemical variations, has been widely employed to detect and study unusual nucleic acid structures like Z-DNA(21) and

-5.OOOE+0i 5: 6: 7: 8: -----

220 .0 D-RNA/D-ONA D-RNA/0-ONA

WL tnm] T-5C T-75C

D-RNA/L-DNA ap

T-SC

D-RNA/L-ONA p

T-5C

3S0 .0

Figure 6. Circular dichroism spectra of the mixtures of natural and L-sequences with complementary RNA in Tris 0. I M, NaCl 0. I M, pH 7.0; 0.01 M spermine was present in the solutions containing the modified sequencies. Signal after melting is also reported for the natural duplex.

quadruple helices(22). As shown in Figure 5 and 6, a large CD is found when 6 or 7 are mixed with their natural antiparallel complement 8, whereas in the spectra obtained upon addition of the L sequences 9 and 10 no such effect is observed. Furthermore, while a large thermal effect is observed with the natural duplexes (due to separation into single strands), the spectra of the L/D mixtures are practically unaffected by temperature. Thus, judging from the results of UV and CD experiments, natural, single-stranded DNA and RNA with a random composition do not interact with complementary L-DNA. This finding was definitively proved by mobility shift experiments on native PAGE, at 5°C. When the natural DNA 6 (Figure 7) is mixed with an equimolar amount of either L-ap 9 or L-p 10 (lanes 7 and 8 respectively) the component strands migrate separately with unchanged mobility; on the contrary, when 6 is combined with the natural complement D-ap 8, only one band is detected (lane 6) with a lower mobility than those of the separate strands. Nothing changes when the target is the natural RNA 7 (Figure 8, lanes 6, 5 and 2 respectively), showing again no duplex formation between L and D strands. DISCUSSION The present experimental results show that mixed-base L-DNA do not bind complementary natural DNA and RNA. In agreement with this finding, other authors(23) failed to observe any UV detectable association between mixed-base L-DNA and D-DNA dodecamers. The experimental behaviour of L-homo-oligomers toward complementary natural strands is varied. Lack of association has been reported for several sequences: L-d(U18)(6), L oa- and 3-

Nucleic Acids Research, 1993, Vol. 21, No. 18 4163

Figure 7. Non denaturating PAGE at 5°C of DNA/DNA. lane 1: Bromophenol blue. lane 2: D-ap 8. lane 3: DNA 6. lane 4: L-ap 9. lane 5: L-p 10. lane 6: mixture of D-ap 8 and DNA 6. lane 7: mixture of L-ap 9 and DNA 6. lane 8: mixture of L-p 10 and DNA 6.

Figure 8. Non denaturating PAGE at 5°C of DNA/RNA. lane 1: mixture of Bromophenol blue and D-ap 8. lane 2: mixture of D-ap 8 and RNA 7 (present in a small excess). lane 3: RNA 7. lane 4: L-p 10. lane 5: mixture of L-p 10 and RNA 7. lane 6: mixture of L-ap 9 and RNA 7. lane 7: L-p 9.

d(T8)pO(CH9)30H(8), L ce- and /-(T8) and L d(U8), all carrying an acridine derivative at the 3' end (10). Instead, triple helix formation was observed between i) poly U and L-d(A6) (9) ii) poly T and acridine-derivatized L ce- and 3-d(A4)(10) iii) poly rA and L-(U 12)(1). The complexes are consistently less stable than natural ones. The finding that base composition may play a major role in the hybridization ability of sugar modified oligonucleotides is not new. For example, hybrids between oa-DNAs with an even number of purines and pyrimidines, in random arrangement, and $-DNA and -RNA have stabilities close to those of the corresponding natural complexes(3b), while ao-d(GGAAGG):3d(CCTTCC) melts 7°C below(24) and ca-d(G2T12G1):O-r(A12) 25°C above(25) the corresponding natural complexes. The contradiction we have found for L/D hybrids, between the experimental results and the predictions based upon modeling

calculations, is not novel. The most stable heterochiral RNA duplex calculated by van Boeckel et al.(7) had a ladder-like structure, with the two strands running parallel, and an energy content significantly higher than that of the corresponding natural A-type duplex. Instead, 'H-NMR experiments showed that a complex, devoid of Watson -Crick pairing, was formed by the L strand and both parallel and antiparallel complementary D sequences and that, moreover, the L/D hybrid was more stable than the corresponding natural duplex. The AshleyO' 1) modeling study suggested two stable heterochiral combinations: an A-type parallel duplex and a triple helix formed by two L homopyrimidine strands and a D homo-purine one. UV melting experiments displayed a sharp hyperchromic transition only for the 2L-r(UI2):poly rA mixture, with a Tm slightly lower than that of the natural counterpart. However, the reported IH-NMR spectra at 5°C of the imino proton resonances of the two complexes are different and the hyperchromicity of the DDD hybrid is about twice that of the LLD one. In this connection, it is interesting to note that the triplexes formed by poly U and L- or D-d(A6) have identical hyperchromicity, but the DDL complex Tm melts 25°C below the DDD triplex(9). Considering these examples and our own finding, we are now inclined to think that potential energy calculations do not allow sound prediction on the structure and stability of complexes involving oligonucleotides of opposite chirality. Thus, when a qualitative agreement between calculations and experimental results is found, like in one of the examples discussed above, it is most likely a fortuitous finding, devoid of any cognitive value. A situation somewhat similar to that described here is the selfassociation of 2',5'-linked oligonucleotides, for which conflicting predictions had been made by theoretical calculations, in spite of available reference data from X-ray crystal structures of free dimers. The synthesis of several 2',5'-oligonucleotides has very recently allowed to probe the real hybridization properties of this natural but uncommon DNA(27). Altogether, it seems that modeling calculations do not lead to sound predictions about complexes involving oligomers whose behaviour cannot be derived, in an easily predictable way, from that of already known systems. This should not come as a surprise, since empirically based force field development has been obtained by optimizing the agreement between existing experimental data and potential energy predictions. The obvious implication is that any force field is inherently 'biased' by the molecules taken as reference standard and may lead to unreliable predictions if applied to systems that greatly differ from those already known. Furthermore, while the behaviour of real molecules depends upon free energy, the predictions of force field are based on potential energy calculations. As clearly stated in a recent article(26), which thoroughly evaluates the present state of molecular modeling, 'in any case where there is a non negligible entropic contribution to free energy, it is quite possible that prediction based on the CPFF alone will significantly differ from those that would be made if free energy surface were known'. Finally, in complex systems like nucleic acids, the solvation effects play a role that may be too poorly accounted for by a pseudovacuum model. However, because of the computation cost, free energy computations with explicit inclusion of solvent are not yet practicable for large molecules, as the ones at hand. In spite of the present limitations, modeling calculations may be of value in the field of antisense research, once some of the aspects discussed above have been further improved

4164 Nucleic Acids Research, 1993, Vol. 21, No. 18

CONCLUSIONS The hybridization of L-oligodeoxynucleotides with complementary DNA and RNA strands has been studied with UV, IHNMR and CD spectroscopies and gel elecrophoresis The experimental results make up a coherent picture, showing that L-DNA which contains all four natural bases in a random distribution does not bind single-stranded DNA and RNA in either parallel or antiparallel orientations. The fact that some L-homo-oligomers form triple helices with complementary natural polynucleotides, though interesting from a structural point of view, is not a sound criterion to evaluate the ability of L oligomers to selectively control gene expression, in general. Since in the antimessenger strategy the synthetic oligomer must bind to random sequence single-stranded RNA and in the antigene approach to homopurine/homopyrimidine double-stranded DNA2, these two type of sequences should be used first in appropriate control experiments(3h). In fact, the present result demonstrates that L-DNAs have no future as wide scope antimessenger compounds, contrary to a previous suggestion based upon hybridization experiments with the homosequence L-d(A6)9. As mentioned before, ci-DNAs are another class of sugarmodified oligonucleotides where the results obtained with 'biased' sequences have no general validity. Further studies, in progress, will show if homo-purine and/or homo-pyrimidine L-DNAs can function as specific anti-gene tools and establish the potential of L-RNAs(I1) in the control of gene expression.

MATERIALS AND METHODS Modeling calculations Molecular models of duplexes with mixed chirality were generated using the program InsightIl (BIOSYM). The structures were created on the basis of the geometrical properties which were translated into inter atom distances and given as an input for distance-geometry calculations using the program DGII. To this end, the distances which define the idealised helix structure of the D-strand as well as the Watson-Crick base pairs, were calculated and used as input for distance-geometry calculations. The embedding was done in four dimensions. After embedding, the structures were refined with 10000 steps simulated annealing to improve the sampling of the conformational space. The Lduplex was thus docked on the D-strand by embedding of the distance matrix. The resulting structures were used as starting structures for molecular mechanics and dynamics calculations using the AMBER force field. All structures were energy minimised with 100 steps steepest descent minimisation followed by 500-1000 steps conjugated gradient minimisation until the maximum derivative is less than 0. 1 kcal/A. Typically molecular dynamics calculations were carried out for 100 ps at 300 K. In order to diminish the effects introduced by the exclusion of solvent molecules, the electrostatic interactions were scaled by a factor of 4/R. Materials D-amidites were purchased from Chem Gene and Pharmiiacia. Solid phase syntheses were performed on a Pharmacia Gene Assembler II plus (1.3 ,umol scale) using Pharmacia cartridges. Triethylammonium bicarbonate (TEAB) solution was prepared by passing CO2 gas through a 0°C cooled solution of 2M triethylamine (Fluka) in HPLC grade water until pH 7.4.

DEAE Sephacell was purchased from Pharmacia, Dowex 50Wx8 from Fluka. HPLC were performed with a Varian 5000 and a Waters 600E instruments equipped with Hypersil 5-ODS and Spherisorb Sl0 columns. NMR spectra were measured on a Varian VXR 200, UV measurements were done with a Perkin Elmer spectrophotometer model 554 equipped with a MGW Lauda RC5 thermostat and a MGW Lauda R40/2 digital thermometer. CD measurements were recorded on a JASCO J500A spectropolarimeter. PAGE was done at 3°C on 16% polyacrylamide using Stains-all as dying agent.

Purifications of the oligomers The crude oligomers were purified on anion exchange chromatography columns (DEAE Sephacell) using a gradient of TEAB from 0.05M to 2.OM. The fractions were analyzed by reversed phase HPLC, detecting at 260nm. Fractions with purity higher than 96% were pooled and coevaporated with water several times to decompose the TEAB excess. The purified oligomers were converted to the sodium salts by passing them through a small column of Dowex (Na), analyzed by HPLC on the anion exchange column and eventually lyophilized. The purity of the RNA sequence was checked by HPLC (SAX column) and found to be 92% pure. UV experiments The concentration of the aqueous solution of each oligomer (> 10 OD/ml) was measured at 260 nm, using the Borer's method'28 for calculating molar extinction coefficients. Equimolar amounts of each strand were then diluted in the buffer solution 0. IM TRIS and 0. l M NaCl corrected with lN HCl to pH 7.0; the final concentration was approximately 2.8 x 10-6M per strand. The cells were heated to 80°C for 15 minutes, then slowly cooled down. The gradient of temperature never exceeded 0.5°C/min starting from 4°C (moisture condensation on the walls of the cells was prevented by flushing nitrogen inside the cell holder).

CD experiments CD spectra were recorded with a Jasco J500A spectropolarimeter equipped with a DP 100 data processor and with a thermostatted water jacket cell holder. The solutions were prepared as previously described for the UV experiments. Gel electrophoresis The electrophoretic mobilities of the oligomers in 100mM sodium phosphate buffer, pH 7.2, have been obtained using 18% polyacrylamide gels, run at 5°C in a thermostatted slab gel unit, at 10 V/cm. Gel staining was performed by soaking with 0.01 % Stains-all dye in 1/ 1 water/formamide.

ACKNOWLEDGEMENTS This work was supported by CNR 'Progetto Finalizzato Chimica Fine II'.

REFERENCES Cohen. J. S. Ed. (1989) Oligodeoxynucleotides. Antisense inhibitors of gene expression: CRC Press: Boca Raton, Fl. 2. Heene, C. and Toudmr. J.-J. (1990) Biochin. Biophys. Acta 1049, 99-105. 3. a: Uhimrlanfl. E. aind Pevmzzan, A. (1990) Chemn. Rev. 90, 543-584. b: Morvan, F.: PorLlllb, H.: Decols, G.; Lefebvre, I.; Pompon, A.; Sproat, 1.

Nucleic Acids Research, 1993, Vol. 21, No. 18 4165

4. 5.

6. 7.

8.

9. 10. 11. 12.

13. 14. 15.

16. 17. 18.

19. 20. 21. 22. 23. 24.

25. 26. 27.

28.

B. S.; Rayner, B.; Malvy, C.; Lebleu, B. and Imbach, J.-L. (1993) J. Med. Chem.36, 280-287. Tazawa, 1. M.; Tazawa, S.; Stempel, L. M. and Ts'o, P. 0. P. (1970) Biochemistry 9, 3499-3514. Holy, A. and Sorm, F. (1971) Collect. Czech. Chem. Commun. 36 3282 -3299. Anderson, D. J.; Reischer, R. J.; Taylor, A. J. and Wecher, W. J. (1984) Nucleosides Nucleotides 3, 499-512. van Boeckel, C. A. A.; Visser, G. M.; Hegstrom, R. A. and van Boom, J. H. (1987) J. Mol. Evol. 25, 100-105. Morvan, F.; Genu, C.; Rayner, B.; Gosselin, G. and Imbach, J.-L. (1990) Biochem. Biophys. Res. Commun. 172, 537-543. Fujimori, S.; Shudo, K. and Hashimoto, Y. (1990) J. Am. Chem. Soc. 112, 7436 -7438. Asseline, U.; Hau, J.-F.; Czernecki, S.; Le Diguarher, T.; Perlat, M.-C.; Valery, J.-M. and Thuong, N. T. (1991) Nucleic Acids Res. 19, 4067-4074. Ashley, G. W. (1992) J. Am. Chem. Soc. 114, 9731-9736. Blommers, M. J. J.,(1990) Ph. D. Thesis, Katholieke Universiteit, Nijmegen, The Netherlands. Spadari, S.; Maga, G.; Focher F.; Ciarrocchi, G.; Manservigi, R.; Arcamone, F.; Capobianco, M; Carcuro, A.; lotti, S. and Garbesi, A. (1992) J. Med. Chem. 35,4214-4220. Holy, A. (1972) Collect. Czech. Chem. Commun. 37, 4072-4087. Ti, G. S.; Gaffney, B.L. and Jones, R. A. (1982) J. Am. Chem. Soc. 104, 1316-1319. Sinha, N. D.; Biernat, J.; McManus, J.; Koster, H. (1984) Nucleic Acids Res. 12, 4539-4547. Gait, M. J. (1985) 'Oligonucleotide synthesis' IRL Press. Marugg, J. E.; Tromp, M.; Jhurani, P.; Hoyng, C. F.; van der Marel, G. A. and van Boom, J. H. (1984) Tetrahedron 40, 73-78 Pharmacia Gene Assembler Manual Morelli, S.; Delia, D.; Ryte, A.; Mazzei, M. and Nicolin, A. (1992) Conference on Antisense Oligonucleotides and Ribonucleases H, September 27-October 2, Arcachon, France. Quadrifoglio, F.; Manzini, G.; Vasser, M.; Dinkelspiel, K. and Crea, R. (1981) Nucleic Acids Res. 9, 2195-2206. Bonazzi,S.; Capobianco, M.; De Morais, M., M.; Garbesi, A.; Gottarelli, G.; Mariani, P.; Ponzi Bossi, M., G.; Spada, G., P. and Tondelli, L. (1991) J. Am. Chem. Soc. 113, 5809-5816. Rayner, B. and Imbach, J.-L. (1992) private communication. Paoletti, J.; Bazile, D.; Morvan, F.; Imbach, J.-L. and Paoletti, C. (1989) Nucleic Acids Res. 17, 2693 -2704. Gagnor, C.; Bertrand, J.-R,; Thenet, S.; Lemaitre, M.; Morvan, F.; Rayner, B.; Malvy, C.; Lebleu, B.; Imbach, J.-L. and Paoletti, C. (1987) Nucleic Acids Res. 15, 10419-10436. Pearlman, D. A. and Kollman P.A. (1991) J. Am. Chem. Soc. 113, 7167-7177. a: Kierzek, R.; He, L. and Turner, D. H. (1992) Nucleic Acids Res. 20, 1685-1690. b: Dougherty, J. P.; Rizzo, C. J. and Breslow R. (1992) J. Am. Chem. Soc. 114, 6254-6255. c: Hashimoto, H. and Switzer, C. (1992) J. Am. Chem. Soc. 114, 6255-6256. Borer, P. N., (1975) CRC Handbook of Biochemistry and Molecular Biology 1, 589.

Lihat lebih banyak...

Comentários

Copyright © 2017 DADOSPDF Inc.