Biochemical Society Transactions
764
likely efficacy of antisense oligodeoxynucleotidesas chemotherapeutic agents. Two major problems remain. The first of these, as mentioned earlier, is the stability of the ODNs in cell systems. Our present experiments suggest that the half life of these is between 1 and 3 h, which is unlikely to be sufficient for a clinical preparation. The second concerns the efficiency of uptake into the cells which, while we have not yet quantified it, we believe is unlikely to be greater than 10% of the total amount of ODN administered. It will therefore be essential by suitable means (e.g. liposome encapsulation) [8] to increase this uptake. This work was supported by the SERC and by Stiefel International.
1. 2. 3. 4.
5. 6.
7. 8.
Editorial, (1985) The Lancet 325, 1045-1046 Editorial, (1992) The Lancet 339,963-964 Beral. V. & Booth, M. (1986) The Lancet, 327,495 Scheffner, W., Werness, B. A., Huibregtse, J. M., Levine, A. J. & Howley, P. M. (1990) Cell 63, 1129-1 136 Dyso, N., Howley, P. M., Munger, K. & Harlow, E. (1989) Science 243,934-937 Porter, C. D. & Archard, 1., (1987) J. Gen. Virol. 68, 673-682 Archard, L. C., Johnson, K. K. & Malcolm, A. D. B. (1985) FEBS Lett. 192,53-56 Leonetti, J.-P., Machy, P., Degols, G., Lebleu, B. & Leseran, L. (1990) Proc. Natl. Acad. Sci. U.S.A. 87, 2448-245 1
Received 24 July 1992
Ribonuclease H-mediated inhibition of translation and reverse transcription by antisense oligodeoxynucleotides Claudine Boiziau, Beatrice Larrouy, Serge Moreau, Christian Cazenave, David Shire* and Jean-JacquesToulmet Laboratoire de Biophysique Moleculaire, INSERM CJF90- 13, Universite de Bordeaux II, 33076 Bordeaux Cedex, France, 5anofi Elf BioRecherches, 3 I676 Labege, France
Introduction The association of an oligodeoxynucleotide (oligonucleotide) with single-stranded RNA can interfere with the expression of the encoded information. The so-called antisense strategy has been used to control several steps of gene expression. Both splicing of pre-mRNA and translation of mRNA has been impaired by antisense oligonucleotides, i.e. by oligomers complementary to a portion of the target RNA [l, 21. Recently we extended this approach to the inhibition of reverse transcription: an oligonucleotide complementary to the template RNA can prevent the synthesis of cDNA [3,4]. From in vitro studies with purified components (RNA, enzymes) or in cell-free extracts it was shown that the target RNA did not remain intact during the biological process (translation or reverse transcription) in the presence of oligo-
Abbreviations used: AMV, avian myeloblastosis virus; PS, phosphorothioate; PO, phosphodiester; MP, methylphosphonate; HIV, human immunodeficiency virus. tcorrespondence to Dr. Jean-JacquesToulmC, Laboratoire de Biophysique MolCculaire, Universitk de Bordeaux 11, Bit. 3a, 146 rue Lko Saignat, 33076 Bordeaux Cedex, France.
Volume 20
nucleotides. The RNA was cleaved at the oligonucleotide binding site due to the recognition of the antisense DNA/RNA hybrid by RNase H [5,6]. Of course, once the target is cleaved it can no longer direct cDNA or protein synthesis. The reversible binding of the antisense oligonucleotide leads to a permanent damage of the RNA. RNase H-induced cleavage is therefore responsible for an improved efficiency of antisense effects. The specificity of the effect induced by antisense oligonucleotides which makes this approach so attractive resides in the selective recognition of the complementary sequence through the formation of AT (AU) and GC pairs. Mismatches strongly destabilize imperfect hybrids. If the antisense sequence is short enough, weak mismatched duplexes resulting from the binding of the oligonucleotide to partially complementary sequences are not expected to give rise to non-specific effects on gene expression. However, RNase H can recognize a mismatched DNA/FWA duplex as a substrate [6]. Therefore, this enzyme might cleave FWA at non-target sites exhibiting only a partial with the target sequence Of the antisense oligonucleotide. Consequently, W a s e H might be responsible for the inhibition of the expression of non-target genes.
Anti-sense Oligonucleotides
Numerous chemical modifications have been introduced into oligonucleotides that aimed at improving the antisense efficiency. In particular, nuclease resistant oligomers have been obtained by modification of the phosphodiester backbone, inversion of the anomeric configuration of the nucleoside unit, or substitution on the 2' position [7]. Most of these molecules do not elicit RNase H. The use of sandwich molecules made of a stretch of unmodified units between two blocks of modified segments which do not allow RNase H activity might lead to compounds of improved specificity. We will describe here the molecular mechanisms by which antisense oligonucleotides block protein and cDNA synthesis and discuss the consequences of the use of nuclease resistant oligomers on these mechanisms.
Inhibition of translation and reverse transcription by unmodified oligonucleotides Oligomers complementary to the coding region of mRNA are very poorly efficient at inhibiting translation. In most cases the antisense oligonucleotide/ mRNA hybrid is melted out by the unwinding activity associated with the elongating ribosome. Even sequences several hundred nucleotides long do not reduce the extent of translation [8]. Indeed, an oligonucleotide targeted to the coding sequence does not prevent translation in rabbit reticulocyte lysate unless the cell-free extract is complemented with Escherichziz coli FWase H activity [9]. Such an addition is not required when translation is performed in wheat germ extract, which contains RNase H; it is not necessary in Xenopus oocytes either. In these latter cases the endogenous RNase H activity mediates the cleavage of the target RNA allowing antisense oligomers complementary to the coding region to inhibit protein synthesis [6]. In contrast, this enzyme is not required when the 5' leader region of the message, upstream of the initiator AUG, is targeted. Oligonucleotides complementary to this part of mRNA can outcompete the initiation complex and thereby arrest translation [ 10, 113. The inhibition of reverse transcription by synthetic complementary oligonucleotides, has also been shown to involve the RNase H activity borne by the reverse transcriptase. An antisense oligonucleotide induces the cleavage of the template at its binding site thus preventing cDNA synthesis beyond this region. As a consequence, reverse transcription is not inhibited when the polymerization is performed with a recombinant enzyme
devoid of RNase H activity. The inhibitory efficiency is correlated with the affinity of the oligomer for its complementary sequence, i.e. with the length of the oligonucleotide. A 50% reduction of cDNA synthesis was achieved by 0.3 and 20 p~ of a 17mer and a 12-mer, respectively, using rabbit Bglobin mRNA as a template for the avian myeloblastosis virus (AMV) enzyme [ 51.
Inhibition of translation by nuclease resistant oligomers: Effect of alpha and phosphorothioate analogues Chemical modifications have been introduced into synthetic oligonucleotides to provide them with nuclease resistance. W e investigated the antisense properties of some nuclease resistant analogues namely alpha, phosphorothiate and methylphosphonate oligomers. In wheat germ extract a 17 mer phosphorothioate (PS) sequence complementary to the coding region was shown to block protein synthesis. PS oligonucleotide/mRNA hybrids are recognized and cleaved by RNase H. This is of particular interest as the oligomer which remains intact in this process is available for binding to a second RNA molecule. The combination of nuclease resistance and RNase H-mediated cleavage of the target leads to a catalytic process: microinjection of rabbit /?-globin mRNA and antisense PS-17 mer leads to a selective inhibition of B-globin synthesis. We have shown that 50% reduction was induced by 0.1 ,UMoligonucleotide, i.e. at a concentration 4 times lower than that of the target RNA [ 121. The hybrids formed by alpha-oligonucleotides and RNA are not attacked by RNase H. Consequently, a 17 mer a complementary to the coding region does not prevent polypeptide synthesis. In contrast, a 17 mer a-sequence targeted to the 5' leader region (nt 3-19) of the /?-globin mRNA was able to arrest the initiation of translation. In rabbit reticulocyte lysate, in which the RNase H contribution is low, this a-17 mer is a more efficient inhibitor of B-globin synthesis than the unmodified sequence [ 111.
Inhibition of translation by sandwich molecules In addition to its complementary sequence an oligonucleotide can bind to sequences which exhibit partial complementarity and give rise to mismatched duplexes. Due to co-operative interactions in double-stranded nucleic acids the stability of unperfect hybrids is low compared with fully paired duplexes, at least if mismatches occur in the central
I992
765
Biochemical Society Transactions
766
part of the duplex. Therefore, if the oligomer is short enough (15 to 20 nucleotides long) no antisense effect resulting from the formation of partial duplexes is expected. However, we observed that an unmodified 15-mer targeted to the AUG initiation codon of the rabbit a-globin mRNA prevented the in vitro synthesis of both B- and a-globin with a similar efficiency when translation was performed in wheat germ extract. Northern blot analysis of the 8-globin mRNA after translation revealed that the message was cleaved at three different positions corresponding to sites at which the 15-mer could form up to 13 base pairs. These partial duplexes were recognized by RNase H, which attacked the RNA moiety in such hybrids. Therefore, RNase H-mediated cleavage was the source of undesired effects; as previously mentioned the RNase H activity is very low in reticulocyte lysate compared to wheat germ extract. We synthesized a ‘sandwich‘ analogue of the 15-mer in which a stretch of five unmodified phosphodiester (PO) residues was inserted between two blocks of methylphosphonate (MP) nucleosides. It was demonstrated that MP oligonucleotides do not elicit RNase H activity. We reasoned that a MP/PO/ MP chimera will recognize the complementary sequence as specifically as the parent oligomer and that it would retain the capacity to induce the cleavage of the target RNA in the PO region. But the RNase H activity will be restricted to the central part of the molecule. Therefore we expected an improved specificity from these composite antisense oligonucleotides. Indeed, we demonstrated that the anti-a-globin mRNA MP/PO/MP 15-mer restored some selectivity: the addition of 1 ,DMoligomer decreased a- and /3-globin synthesis by 60% and 15%, respectively. Northern blot analysis of B-globin mRNA after translation in vitro in the presence of the sandwich antisense oligonucleotide revealed a reduced number of breakdown fragments compared with the situation obtained with the unmodified oligonucleotide. The analysis of the mismatched duplexes indicated that the improved specificity of the sandwich oligonucleotide was probably due to the reduced stability of the complex formed with the complementary sequence compared with the one obtained with the parent phosphodiester 15-mer [ 131.
Inhibition of reverse transcription by modified oligonucleotides Effect of alpha-derivatives Alpha-oligonucleotides do not allow RNase H activity when bound to an RNA sequence. As a
Volume 20
consequence, an a-17-mer complementary to the 5’ end of the rabbit B-globin mRNA chosen as a template, did not prevent cDNA synthesis primed 130 nucleotides away. This is in agreement with the RNase H-dependent inhibition of reverse transcription by antisense oligonucleotides described above. The use of a-derivatives allowed us to demonstrate that inhibition of cDNA synthesis could occur through a second mechanism: an antisense a-17 mer bound contiguously to the primer impedes its use by either the Moloney murine leukaemia virus or the avian (AMV) reverse transcriptase 151. Of course, the template RNA remains intact during the process. It should be noted that a-derivatives, which bind in a parallel orientation to their complementary sequence, cannot be lengthened by reverse transcriptase. A co-operative interaction between the primer and the antisense oligonucleotides is strictly required: a gap of a single nucleotide between the two sequences releases the inhibition. Use of phosphorothioate analogues
PS oligonucleotides have been described as inhibitors of the development of the Human Immunodeficiency Virus (HIV) in cultured cells [14]. The antiviral property of these oligomers was in part ascribed to a direct interaction with the HIV reverse transcriptase [IS]. This inhibition was sequenceindependent. We investigated the effect of a PS-17mer (17PS cap), complementary to the 5’ end (nucleotides 3-19) of the rabbit /3-globin mRNA, on the cDNA synthesis by the AMV reverse transcriptase. As shown in Fig. 1, the addition of the PS17-mer led to the decrease of the full length cDNA fragment and to the concomitant appearance of shortened fragments whose size was consistent with a synthesis aborted at the antisense oligomer binding site. At high concentration ( > 30 nM) the overall cDNA synthesis decreased suggesting a nonspecific inhibition of reverse transcription by these modified oligomers.
Conclusion W e investigated the effects of unmodified and nuclease resistant oligomers on in vitro translation and reverse transcription. W e have demonstrated that both steps can be inhibited through two different mechanisms. Firstly, a direct competition between the antisense oligomer and the translating, or transcribing, entity is responsible for the hybridarrested synthesis of protein, or cDNA. This mechanism accounts for the results obtained with oligomers complementary to the 5‘ leader sequence
Anti-sense Oligonucleotides
Fig. I Effect of antisense phosphorothioate oligonucleotide on reverse transcription Rabbit b-globin mRNA was used as a template by AMV reverse transcriptase either in the absence (lane C) or in the presence of the chemically modified oligonucleotide I7PScap, complementary t o nucleotides 3- I 9 of the B-globin message, at the ) at the top of different lanes. The concentration ( p ~indicated cDNA synthesis was primed by an Unmodified phosphodiester oligonucleotide complementary t o the I 13- I29 sequence [32P]LabelledcDNA fragments were analysed on a 20% polyacrylarnide sequencing gel (See [5] for experimental conditions) Size markers are indicated on the left
M
C
17PScap 0.01 0.03 0.05 0.1
123
-
110 -
of the message (in the case of translation) or to the region contiguous to the primer binding site (for reverse transcription). Secondly, antisense oligodeoxynucleotides can mediate the degradation of the target RNA by RNase H. This is the major - if not the only - way to interfere with the elongation step of polypeptide synthesis or with the cDNA polymerization, i.e. when oligomers are targeted to the mRNA coding region or to a site remote from the primer. W e have demonstrated that RNase H allows antisense oligonucleotides to act via a catalytic process. This is of particular interest in the case of long-lived oligomers such as phosphorothioate analogues. However, the RNase H activity is also responsible for non-specific inhibition of non-target genes due to its ability to cleave the RNA part of mismatched hybrids. Sandwich molecules made of a stretch of unmodified residues between two segments which do not elicit RNase H led to antisense compounds of improved specificity. It has also been demonstrated that methylphosphonate/ phosphodiester chimeras are resistant to exonucleases and are efficiently taken up by cells [ 161 (D. M. Tidd, personal communication).
Phosphorothioates have been recognized as sequence-independent inhibitors of reverse transcription. We have demonstrated here that these analogues can induce a sequence-specific inhibition of reverse transcription. W e have not analysed the template RNA after the reaction yet, but it has been demonstrated that phosphorothioate derivatives induce RNase H activity. Therefore, these analogues very probably inhibit reverse transcription via the RNase H pathway. This work was supported by the Conseil Regional dAquitaine, by a grant from the Direction des Recherches, des Etudes et des Techniques, and by a grant from the Association de la Kecherche contre le Cancer. C.B. is a recipient of the Ligue contre le Cancer and B.L. a DRET fellowship. 1. Uhlmann, E. & Peyman, A. (1990) Chern. Rev. 90, 544-579 2. Hkltne, C. & ToulmC, J. J. (1990) Biochim. Biophys. Acta 1049,99-125 3. Loreau, N., Boiziau, C., Verspieren, P., Shire, D. & Toulrnk, J. J. (1990) FERS Lett. 252, 53-56 4. Boiziau, C., Blonski, C., Thuong, N. T., Shire, D. & Toulrne, J. J. (1991) Nucleic Acids Res., Symp. Ser. 24,121-125 5. Boiziau, C., Thuong, N. T. & ToulmC, J. J. (1992) Proc. Natl. Acad. Sci. U.S.A. 89,768-772 6. Cazenave, C., Loreau, N., Thuong, N. T., Toulrne, J. J. & Helene, C. (1987) Nucleic Acids Res. 15, 47 17-4736 7. Goodchild, J. (1990) Bioconjugate Chern. 1, 165-187 8. Shakin, S. N. & Liebhaber, S. A. (1986) J. Biol. Chern. 261, 16018- 16025 9. Minshull, J. & Hunt, T. (1986) Nucleic Acids Res. 14, 6433-6451 10. Bertrand, J. K., Irnbach, J. I,., Paoletti, C. & Malvy, C. (1989) Biochern. Biophys. Res. Commun. 164, 311-318 11. Boiziau, C., Kurfurst, R., Cazenave, C., Roig, V., Thuong, N. T. & Toulmk, J. J. (1991) Nucleic Acids Res. 19,1113-1 119 12. Cazenave, C., Stein, C. A., Loreau, N., Thuong, N.T., Neckers, L. M., Subasinghe, C., Helkne, C., Cohen, J. S. & Toulrnk, J. J. (1989) Nucleic Acids Res. 17, 425 5-4273 13. Larrouy, B., Blonski, C., Boiziau, C., Stuer, M., Moreau, S., Shire, D. & Toulrnk, J. J. (1992) Gene, in the press 14. Matsukura, M., Shinokuza, K., Zon, G. Mitsuya, H., Reitz, M., Cohen, J. S. & Broder, S. (1987) Proc. Natl. Acad. Sci. U.S.A. 84,7706-77 10 15. Majumdar, C., Stein, C. A., Cohen, J. S., Broder, S. & Wilson, S. H. (1989) Biochemistry 28, 1340-1346 16. Giles, R. V. & Tidd, D. M. (1992) Nucleic Acids Res. 20,763-770
Received 22 July 1992
I992
767
764
likely efficacy of antisense oligodeoxynucleotidesas chemotherapeutic agents. Two major problems remain. The first of these, as mentioned earlier, is the stability of the ODNs in cell systems. Our present experiments suggest that the half life of these is between 1 and 3 h, which is unlikely to be sufficient for a clinical preparation. The second concerns the efficiency of uptake into the cells which, while we have not yet quantified it, we believe is unlikely to be greater than 10% of the total amount of ODN administered. It will therefore be essential by suitable means (e.g. liposome encapsulation) [8] to increase this uptake. This work was supported by the SERC and by Stiefel International.
1. 2. 3. 4.
5. 6.
7. 8.
Editorial, (1985) The Lancet 325, 1045-1046 Editorial, (1992) The Lancet 339,963-964 Beral. V. & Booth, M. (1986) The Lancet, 327,495 Scheffner, W., Werness, B. A., Huibregtse, J. M., Levine, A. J. & Howley, P. M. (1990) Cell 63, 1129-1 136 Dyso, N., Howley, P. M., Munger, K. & Harlow, E. (1989) Science 243,934-937 Porter, C. D. & Archard, 1., (1987) J. Gen. Virol. 68, 673-682 Archard, L. C., Johnson, K. K. & Malcolm, A. D. B. (1985) FEBS Lett. 192,53-56 Leonetti, J.-P., Machy, P., Degols, G., Lebleu, B. & Leseran, L. (1990) Proc. Natl. Acad. Sci. U.S.A. 87, 2448-245 1
Received 24 July 1992
Ribonuclease H-mediated inhibition of translation and reverse transcription by antisense oligodeoxynucleotides Claudine Boiziau, Beatrice Larrouy, Serge Moreau, Christian Cazenave, David Shire* and Jean-JacquesToulmet Laboratoire de Biophysique Moleculaire, INSERM CJF90- 13, Universite de Bordeaux II, 33076 Bordeaux Cedex, France, 5anofi Elf BioRecherches, 3 I676 Labege, France
Introduction The association of an oligodeoxynucleotide (oligonucleotide) with single-stranded RNA can interfere with the expression of the encoded information. The so-called antisense strategy has been used to control several steps of gene expression. Both splicing of pre-mRNA and translation of mRNA has been impaired by antisense oligonucleotides, i.e. by oligomers complementary to a portion of the target RNA [l, 21. Recently we extended this approach to the inhibition of reverse transcription: an oligonucleotide complementary to the template RNA can prevent the synthesis of cDNA [3,4]. From in vitro studies with purified components (RNA, enzymes) or in cell-free extracts it was shown that the target RNA did not remain intact during the biological process (translation or reverse transcription) in the presence of oligo-
Abbreviations used: AMV, avian myeloblastosis virus; PS, phosphorothioate; PO, phosphodiester; MP, methylphosphonate; HIV, human immunodeficiency virus. tcorrespondence to Dr. Jean-JacquesToulmC, Laboratoire de Biophysique MolCculaire, Universitk de Bordeaux 11, Bit. 3a, 146 rue Lko Saignat, 33076 Bordeaux Cedex, France.
Volume 20
nucleotides. The RNA was cleaved at the oligonucleotide binding site due to the recognition of the antisense DNA/RNA hybrid by RNase H [5,6]. Of course, once the target is cleaved it can no longer direct cDNA or protein synthesis. The reversible binding of the antisense oligonucleotide leads to a permanent damage of the RNA. RNase H-induced cleavage is therefore responsible for an improved efficiency of antisense effects. The specificity of the effect induced by antisense oligonucleotides which makes this approach so attractive resides in the selective recognition of the complementary sequence through the formation of AT (AU) and GC pairs. Mismatches strongly destabilize imperfect hybrids. If the antisense sequence is short enough, weak mismatched duplexes resulting from the binding of the oligonucleotide to partially complementary sequences are not expected to give rise to non-specific effects on gene expression. However, RNase H can recognize a mismatched DNA/FWA duplex as a substrate [6]. Therefore, this enzyme might cleave FWA at non-target sites exhibiting only a partial with the target sequence Of the antisense oligonucleotide. Consequently, W a s e H might be responsible for the inhibition of the expression of non-target genes.
Anti-sense Oligonucleotides
Numerous chemical modifications have been introduced into oligonucleotides that aimed at improving the antisense efficiency. In particular, nuclease resistant oligomers have been obtained by modification of the phosphodiester backbone, inversion of the anomeric configuration of the nucleoside unit, or substitution on the 2' position [7]. Most of these molecules do not elicit RNase H. The use of sandwich molecules made of a stretch of unmodified units between two blocks of modified segments which do not allow RNase H activity might lead to compounds of improved specificity. We will describe here the molecular mechanisms by which antisense oligonucleotides block protein and cDNA synthesis and discuss the consequences of the use of nuclease resistant oligomers on these mechanisms.
Inhibition of translation and reverse transcription by unmodified oligonucleotides Oligomers complementary to the coding region of mRNA are very poorly efficient at inhibiting translation. In most cases the antisense oligonucleotide/ mRNA hybrid is melted out by the unwinding activity associated with the elongating ribosome. Even sequences several hundred nucleotides long do not reduce the extent of translation [8]. Indeed, an oligonucleotide targeted to the coding sequence does not prevent translation in rabbit reticulocyte lysate unless the cell-free extract is complemented with Escherichziz coli FWase H activity [9]. Such an addition is not required when translation is performed in wheat germ extract, which contains RNase H; it is not necessary in Xenopus oocytes either. In these latter cases the endogenous RNase H activity mediates the cleavage of the target RNA allowing antisense oligomers complementary to the coding region to inhibit protein synthesis [6]. In contrast, this enzyme is not required when the 5' leader region of the message, upstream of the initiator AUG, is targeted. Oligonucleotides complementary to this part of mRNA can outcompete the initiation complex and thereby arrest translation [ 10, 113. The inhibition of reverse transcription by synthetic complementary oligonucleotides, has also been shown to involve the RNase H activity borne by the reverse transcriptase. An antisense oligonucleotide induces the cleavage of the template at its binding site thus preventing cDNA synthesis beyond this region. As a consequence, reverse transcription is not inhibited when the polymerization is performed with a recombinant enzyme
devoid of RNase H activity. The inhibitory efficiency is correlated with the affinity of the oligomer for its complementary sequence, i.e. with the length of the oligonucleotide. A 50% reduction of cDNA synthesis was achieved by 0.3 and 20 p~ of a 17mer and a 12-mer, respectively, using rabbit Bglobin mRNA as a template for the avian myeloblastosis virus (AMV) enzyme [ 51.
Inhibition of translation by nuclease resistant oligomers: Effect of alpha and phosphorothioate analogues Chemical modifications have been introduced into synthetic oligonucleotides to provide them with nuclease resistance. W e investigated the antisense properties of some nuclease resistant analogues namely alpha, phosphorothiate and methylphosphonate oligomers. In wheat germ extract a 17 mer phosphorothioate (PS) sequence complementary to the coding region was shown to block protein synthesis. PS oligonucleotide/mRNA hybrids are recognized and cleaved by RNase H. This is of particular interest as the oligomer which remains intact in this process is available for binding to a second RNA molecule. The combination of nuclease resistance and RNase H-mediated cleavage of the target leads to a catalytic process: microinjection of rabbit /?-globin mRNA and antisense PS-17 mer leads to a selective inhibition of B-globin synthesis. We have shown that 50% reduction was induced by 0.1 ,UMoligonucleotide, i.e. at a concentration 4 times lower than that of the target RNA [ 121. The hybrids formed by alpha-oligonucleotides and RNA are not attacked by RNase H. Consequently, a 17 mer a complementary to the coding region does not prevent polypeptide synthesis. In contrast, a 17 mer a-sequence targeted to the 5' leader region (nt 3-19) of the /?-globin mRNA was able to arrest the initiation of translation. In rabbit reticulocyte lysate, in which the RNase H contribution is low, this a-17 mer is a more efficient inhibitor of B-globin synthesis than the unmodified sequence [ 111.
Inhibition of translation by sandwich molecules In addition to its complementary sequence an oligonucleotide can bind to sequences which exhibit partial complementarity and give rise to mismatched duplexes. Due to co-operative interactions in double-stranded nucleic acids the stability of unperfect hybrids is low compared with fully paired duplexes, at least if mismatches occur in the central
I992
765
Biochemical Society Transactions
766
part of the duplex. Therefore, if the oligomer is short enough (15 to 20 nucleotides long) no antisense effect resulting from the formation of partial duplexes is expected. However, we observed that an unmodified 15-mer targeted to the AUG initiation codon of the rabbit a-globin mRNA prevented the in vitro synthesis of both B- and a-globin with a similar efficiency when translation was performed in wheat germ extract. Northern blot analysis of the 8-globin mRNA after translation revealed that the message was cleaved at three different positions corresponding to sites at which the 15-mer could form up to 13 base pairs. These partial duplexes were recognized by RNase H, which attacked the RNA moiety in such hybrids. Therefore, RNase H-mediated cleavage was the source of undesired effects; as previously mentioned the RNase H activity is very low in reticulocyte lysate compared to wheat germ extract. We synthesized a ‘sandwich‘ analogue of the 15-mer in which a stretch of five unmodified phosphodiester (PO) residues was inserted between two blocks of methylphosphonate (MP) nucleosides. It was demonstrated that MP oligonucleotides do not elicit RNase H activity. We reasoned that a MP/PO/ MP chimera will recognize the complementary sequence as specifically as the parent oligomer and that it would retain the capacity to induce the cleavage of the target RNA in the PO region. But the RNase H activity will be restricted to the central part of the molecule. Therefore we expected an improved specificity from these composite antisense oligonucleotides. Indeed, we demonstrated that the anti-a-globin mRNA MP/PO/MP 15-mer restored some selectivity: the addition of 1 ,DMoligomer decreased a- and /3-globin synthesis by 60% and 15%, respectively. Northern blot analysis of B-globin mRNA after translation in vitro in the presence of the sandwich antisense oligonucleotide revealed a reduced number of breakdown fragments compared with the situation obtained with the unmodified oligonucleotide. The analysis of the mismatched duplexes indicated that the improved specificity of the sandwich oligonucleotide was probably due to the reduced stability of the complex formed with the complementary sequence compared with the one obtained with the parent phosphodiester 15-mer [ 131.
Inhibition of reverse transcription by modified oligonucleotides Effect of alpha-derivatives Alpha-oligonucleotides do not allow RNase H activity when bound to an RNA sequence. As a
Volume 20
consequence, an a-17-mer complementary to the 5’ end of the rabbit B-globin mRNA chosen as a template, did not prevent cDNA synthesis primed 130 nucleotides away. This is in agreement with the RNase H-dependent inhibition of reverse transcription by antisense oligonucleotides described above. The use of a-derivatives allowed us to demonstrate that inhibition of cDNA synthesis could occur through a second mechanism: an antisense a-17 mer bound contiguously to the primer impedes its use by either the Moloney murine leukaemia virus or the avian (AMV) reverse transcriptase 151. Of course, the template RNA remains intact during the process. It should be noted that a-derivatives, which bind in a parallel orientation to their complementary sequence, cannot be lengthened by reverse transcriptase. A co-operative interaction between the primer and the antisense oligonucleotides is strictly required: a gap of a single nucleotide between the two sequences releases the inhibition. Use of phosphorothioate analogues
PS oligonucleotides have been described as inhibitors of the development of the Human Immunodeficiency Virus (HIV) in cultured cells [14]. The antiviral property of these oligomers was in part ascribed to a direct interaction with the HIV reverse transcriptase [IS]. This inhibition was sequenceindependent. We investigated the effect of a PS-17mer (17PS cap), complementary to the 5’ end (nucleotides 3-19) of the rabbit /3-globin mRNA, on the cDNA synthesis by the AMV reverse transcriptase. As shown in Fig. 1, the addition of the PS17-mer led to the decrease of the full length cDNA fragment and to the concomitant appearance of shortened fragments whose size was consistent with a synthesis aborted at the antisense oligomer binding site. At high concentration ( > 30 nM) the overall cDNA synthesis decreased suggesting a nonspecific inhibition of reverse transcription by these modified oligomers.
Conclusion W e investigated the effects of unmodified and nuclease resistant oligomers on in vitro translation and reverse transcription. W e have demonstrated that both steps can be inhibited through two different mechanisms. Firstly, a direct competition between the antisense oligomer and the translating, or transcribing, entity is responsible for the hybridarrested synthesis of protein, or cDNA. This mechanism accounts for the results obtained with oligomers complementary to the 5‘ leader sequence
Anti-sense Oligonucleotides
Fig. I Effect of antisense phosphorothioate oligonucleotide on reverse transcription Rabbit b-globin mRNA was used as a template by AMV reverse transcriptase either in the absence (lane C) or in the presence of the chemically modified oligonucleotide I7PScap, complementary t o nucleotides 3- I 9 of the B-globin message, at the ) at the top of different lanes. The concentration ( p ~indicated cDNA synthesis was primed by an Unmodified phosphodiester oligonucleotide complementary t o the I 13- I29 sequence [32P]LabelledcDNA fragments were analysed on a 20% polyacrylarnide sequencing gel (See [5] for experimental conditions) Size markers are indicated on the left
M
C
17PScap 0.01 0.03 0.05 0.1
123
-
110 -
of the message (in the case of translation) or to the region contiguous to the primer binding site (for reverse transcription). Secondly, antisense oligodeoxynucleotides can mediate the degradation of the target RNA by RNase H. This is the major - if not the only - way to interfere with the elongation step of polypeptide synthesis or with the cDNA polymerization, i.e. when oligomers are targeted to the mRNA coding region or to a site remote from the primer. W e have demonstrated that RNase H allows antisense oligonucleotides to act via a catalytic process. This is of particular interest in the case of long-lived oligomers such as phosphorothioate analogues. However, the RNase H activity is also responsible for non-specific inhibition of non-target genes due to its ability to cleave the RNA part of mismatched hybrids. Sandwich molecules made of a stretch of unmodified residues between two segments which do not elicit RNase H led to antisense compounds of improved specificity. It has also been demonstrated that methylphosphonate/ phosphodiester chimeras are resistant to exonucleases and are efficiently taken up by cells [ 161 (D. M. Tidd, personal communication).
Phosphorothioates have been recognized as sequence-independent inhibitors of reverse transcription. We have demonstrated here that these analogues can induce a sequence-specific inhibition of reverse transcription. W e have not analysed the template RNA after the reaction yet, but it has been demonstrated that phosphorothioate derivatives induce RNase H activity. Therefore, these analogues very probably inhibit reverse transcription via the RNase H pathway. This work was supported by the Conseil Regional dAquitaine, by a grant from the Direction des Recherches, des Etudes et des Techniques, and by a grant from the Association de la Kecherche contre le Cancer. C.B. is a recipient of the Ligue contre le Cancer and B.L. a DRET fellowship. 1. Uhlmann, E. & Peyman, A. (1990) Chern. Rev. 90, 544-579 2. Hkltne, C. & ToulmC, J. J. (1990) Biochim. Biophys. Acta 1049,99-125 3. Loreau, N., Boiziau, C., Verspieren, P., Shire, D. & Toulrnk, J. J. (1990) FERS Lett. 252, 53-56 4. Boiziau, C., Blonski, C., Thuong, N. T., Shire, D. & Toulrne, J. J. (1991) Nucleic Acids Res., Symp. Ser. 24,121-125 5. Boiziau, C., Thuong, N. T. & ToulmC, J. J. (1992) Proc. Natl. Acad. Sci. U.S.A. 89,768-772 6. Cazenave, C., Loreau, N., Thuong, N. T., Toulrne, J. J. & Helene, C. (1987) Nucleic Acids Res. 15, 47 17-4736 7. Goodchild, J. (1990) Bioconjugate Chern. 1, 165-187 8. Shakin, S. N. & Liebhaber, S. A. (1986) J. Biol. Chern. 261, 16018- 16025 9. Minshull, J. & Hunt, T. (1986) Nucleic Acids Res. 14, 6433-6451 10. Bertrand, J. K., Irnbach, J. I,., Paoletti, C. & Malvy, C. (1989) Biochern. Biophys. Res. Commun. 164, 311-318 11. Boiziau, C., Kurfurst, R., Cazenave, C., Roig, V., Thuong, N. T. & Toulmk, J. J. (1991) Nucleic Acids Res. 19,1113-1 119 12. Cazenave, C., Stein, C. A., Loreau, N., Thuong, N.T., Neckers, L. M., Subasinghe, C., Helkne, C., Cohen, J. S. & Toulrnk, J. J. (1989) Nucleic Acids Res. 17, 425 5-4273 13. Larrouy, B., Blonski, C., Boiziau, C., Stuer, M., Moreau, S., Shire, D. & Toulrnk, J. J. (1992) Gene, in the press 14. Matsukura, M., Shinokuza, K., Zon, G. Mitsuya, H., Reitz, M., Cohen, J. S. & Broder, S. (1987) Proc. Natl. Acad. Sci. U.S.A. 84,7706-77 10 15. Majumdar, C., Stein, C. A., Cohen, J. S., Broder, S. & Wilson, S. H. (1989) Biochemistry 28, 1340-1346 16. Giles, R. V. & Tidd, D. M. (1992) Nucleic Acids Res. 20,763-770
Received 22 July 1992
I992
767
