WO1997046672A2

WO1997046672A2 - Antisense nucleic acids and hammerhead ribozymes

Info

Publication number: WO1997046672A2
Application number: PCT/EP1997/002923
Authority: WO
Inventors: Georg Sczakiel; Rainer Haas; Ralf Kronenwett
Original assignee: Deutsches Krebsforschungszentrum Stiftung des öffentlichen Rechts
Priority date: 1996-06-05
Filing date: 1997-06-05
Publication date: 1997-12-11
Also published as: EP0951540A2; WO1997046672A3; JP2000511428A

Abstract

The present invention relates to antisense constructs, particularly antisense nucleic acids and hammerhead ribozymes, which are specific for transcripts encoded by chromosomal translocations, such as the bcr-abl fusion as well as to pharmaceutical compositions containing said antisense constructs.

Description

"ANTISENSE NUCLEIC ACIDS AND HAMMERHEAD RIBOZYMES"

Description

The present invention relates to antisense constructs, particularly antisense nucleic acids and hammerhead ribozymes, which are specific for transcripts encoded by chromosomal translocations, as well as to pharmaceutical compositions containing said antisense constructs.

Antisense nucleic acids and ribozymes have been shown to be potent inhibitors of gene expression and viral functions (Helene & Toulme, 1990; Stein & Chang, 1 993; Marschall et al., 1 994, James & Al-Shamkhani, 1995; Sczakiel & Nedbal, 1 995). Sequence selectivity is particularly important in case of cellular target sequences that differ only slightly from non-target sequences. In case of shortchain antisense oligonucleotides ( < 30 nt), selective binding to the target can be monitored in vitro by comparing the melting temperatures of the duplexes formed and sequence selectivity has been demonstrated convincingly in living cells (Schwab et al. , 1994) Conversely, for long-chain duplexes ( > 30 to 40 nucleotides) discrimination between a completely matching target and a sequence that differs by only one or a few bases seems to be unlikely. Hence, the melting temperature is not appropriate to investigate selectivity in vitro .

For long-chain antisense RNA the association rate with their target RNA in vitro correlates with their effectiveness in vivo This was found for naturally occuπng antisense RNA in procaryotes (Wagner & Simons, 1 994) as well as for antisense RNA-mediated inhibition of the replication of the human immunodeficiency virus type 1 (HIV-1 ) in human cells (Rittner et al. , 1 993). In the latter case, a substantial difference in the association rates of antisense strands of almost the same length was observed . Consequently, the search for fast-hybridizing antisense species supports the identification of improved inhibitors and sequence selectiv ity is not necessarily a matter of duplex stability but, in case of long-chain complementary RNA sequences, a matter of the association kinetics. Thus, the search for selective antisense inhibitors has to be based on the investigation of the association kinetics of complementary nucleic acids.

In case of ribozymes, both, the binding to the target as well as its cleavage are thought to contribute to specific destruction of the target. Hammerhead ribozymes can be designed such that they consist of only one antisense arm, for example Helix Ill-forming sequences, and the catalytic domain but lack the second antisense arm (Helix I) except a number of at least three base pairs

(Tabler et al. , 1 994) . This asymmetric design results in two functional domains, the antisense arm for binding and the catalytic domain for cleavage. Thereby, association kinetics can be performed systematically. The binding domain as well as the catalytic domain can be placed such that only in case of a perfectly matching target sequence both, binding and cleavage of the target sequence can occur.

Biologically relevant model systems for the study of sequence selectivity include transcriptionally active sequences that result from genomic aberrations. Chromosomal translocations that are associated with proliferation of malignant cells are of particular interest. For example, the t(9;22) translocation results in the bcr-abl fusion gene which consists of bcr sequences at its 5' portion and abl sequences at the 3' portion (reviewed in Kurzrock et al. , 1 988) . This translocation is thought to play an important role in the development of chronic myelogenous leukemia (CML) in man. This view is supported by studies in mice showing that the expression of the bcr-abl gene causes malignancies resembling the pathology of CML (Daley et al., 1 990; Kelliher et al. , 1990) . The abl wild type gene is apparently involved in the control of cell cycle in normal cells (Konopka & Witte, 1985; Sawyers et al. , 1994) . The function of the bcr gene product is still not clear, though it is regarded to act as a multifunctional signal- transducing protein (Maru & Witte, 1 991 ; Maru et al. , 1 995; Voncken et al. , 1995) . Thus, the selective inhibition of the bcr-abl gene is required to spare the wild type genes in normal hematopoietic cells. Accordingly, the technical problem underlying the present invention is to provide a new system for both specific and selective inhibition of expression of fusion genes formed by chromosomal translocating which might cause severe disorders such as cancer.

The solution to the above technical problem is achieved by providing a nucleic acid comprising a nucleotide sequence containing

(a) a portion complementary to a first chromosomal DNA sequence, and

(b) a portion complementary to a second chromosomal DNA sequence, wherein the first and second chromosomal DNA sequences form at least a part of a chromosomal translocation resulting in a fusion gene, said part containing the translocation point.

The terms "nucleic acid" and "nucleotide sequence" refer to endogenously expressed, semisynthetic, synthetic or chemically modified nucleic acid molecules of deoxyribonucleotides and/or ribonucleotides.

The expression "a portion complementary to" refers to the region of a nucleic acid that can form base pairs with a given nucleic acid, e.g. a biologically relevant target nucleic acid and, thus, can form double-stranded nucleic acids with the target.

The term "chromosomal translocation" refers to the combination of DNA sequences of chromosomla loci that are not linked together in a normal cell. In a preferred embodiment of the present invention the chromosomal translocation is t(9;22) .

The term "fusion gene" refers to a gene that is the result of chromosoml rearangements and consists of combined DNA sequences of different chromosomal loci. Fusion genes can express genetic information that does not exist as such in a normal cell.

In a preferred embodiment of the present invention the nucleotide sequence consists substantially of ribonucleotides. Preferred examples of said nucleotide sequences are listed below:

In a further embodiment of the present invention the portion (a) and/or the portion (b) of the above defined nucleotide sequence contain(s) the catalytic domain of a hammerhead ribozym. The nomenclature and numbering system for hammerhead ribozymes used in the present invention correspond to Hertel et al. ( 1992) .

The term "catalytic domain" refers to a nucleic acid sequence that is able to catalyze the site-specific cleavage of another nucleic acid in trans. The cleavage competent complex between the target nucleic acid and the nucleic acid containing the catalytic domain is formed and facilitated via a portion of the latter strand that is complementary to the target. In a preferred embodiment the above defined nucleic acid comprises the sequence of an asymmetric hammerhead ribozyme; i.e. one of the regions flanking the catalytic domain is smaller in length than the other. The portion (a) and/or portion (b) form(s) at least part(s) of the Helix I- and/or Helix Ill-forming region(s) of a hammerhead ribozyme.

The expression "Helix I- and/or Helix Ill-forming region(s)" refers to the region(s) of the ribozyme sequence that flanks the catalytic domain at the 3'-side and/or 5'-side, respectively. Preferred examples of said asymmetric hammerhead ribozyme sequences are listed below.

Another embodiment of the present invention relates to a DNA sequence that contains the nucleic acid which upon transcription to RNA, corresponds to the nucleic acid as defined above. The term "upon transcription" refers to an enzymatic conversion from DNA sequences into RNA sequences by an appropriate RNA polymerase.

Another embodiment of the present invention relates to a DNA and/or RNA vector comprising the nucleic acid or the DNA sequence, as defined above. The term "vector" refers to a DNA and/or RNA replicon that can be used for the amplification of a foreign nucleotide sequence. In particular useful in the context of asymmetric hammerhead ribozymes are DNA or RNA vectors which allow the expression of the hammerhead ribozyme. An example is a DNA vector containing the nucleic acid encoding for an asymmetric hammerhead ribozyme downstream of an appropriate promotor so that transcription from said promotor generates the functional and catalytically active asymmetric hammerhead ribozyme.

Another embodiment of the present invention relates to a host organism containing the nucleic acid, the DNA sequence or the vector, as defined above. The term "host organism" relates to a virus, bacterium, fungus, plant or mammal.

A preferred embodiment of the present invention relates to a host organism carrying an episomal replication system that is able to amplify the nucleic acid encoding the above defined nucleotide sequence or the asymmetric hammerhead ribozyme. The term "episomal replication system" refers to a replication system that is independant of the genome of the host; i.e. a plasmid DNA or a virus are episomal replication systems.

A further preferred embodiment of the present invention relates to a genetically engineered host organism containing at least one of the above defined nucleic acids in its genome. The term "genome" refers to the entire genetic material that is inheritable to the next generation. In particular, the present invention relates to an organism which inherit the property to synthesize a functional, catalytically active asymmetric hammerhead ribozyme.

A further embodiment of the present invention relates to a method for the production of the above defined nucleic acid, the DNA sequence or the vector, comprising cultivating the above defined host organism under suitable conditions and isolating the desired products from the culture (cells and/or culture medium) according to procedures known in the art.

In a further embodiment the present invention relates to a pharmaceutical composition containing the above defined nucleic acid, the DNA sequence or the vector, optionally in association with a pharmaceutically acceptable carrier and/or diluent. The pharmaceutical composition may be used for the treatment of disorders based on chromosomal translocations such as chronic myelogenous leukemia, acute lymphoblastic leukemias, acute myelogenous leukemias, and low-grade Non-Hodgkin lymphomas. In particular, the above defined nucleic acid, the DNA sequence or the vector may be used for eliminating a target RNA in a human patient by parenteral or other means of administration including purging of hematopoetic stem cells from malignant cells in the treatment of patients. The bcr-abl-directed nucleic acids described here serve for both, effective inhibition in treated bcr-abl-positive cells as well as for unaffected expression of the bcr and abl wild type genes by sequence selectivity. The latter property is crucial since normal expression of the bcr and abl wild type genes is essential for normal cell proliferation. The experimental testing of sequence selectivity as described here is novel.

The Figures show: Fig.1 : Depiction of the bcr-abl-derived constructs used in this work. (A) The top shows the schematic map of the Bluescript-based plasmids for in vitro transcription of target RNA. Filled arrows indicate the orientation of the T7 and T3 promotor. Additionally, the three target RNA abl1 b, bcr and bcr-abl are shown. Numbers within open bars indicate exons of the bcr-abl, abl and bcr genes, respectively. Numbers given on top indicate the length of RNA. (B) Parental antisense RNA and ribozymes. Numbers on top indicate RNA length. (C) Sequences of the complex formed between ribozymes and bcr-abl substrate RNA. Cleavage positions are indicated by an arrow and bcr-abl fusion points by interrupted line.

Fig.2: Schematic depiction of the kinetic in vitro selection and identification of fast-hybridizing antisense RNA and ribozyme species. Fig.3: Annealing of αBA62-derived antisense RNA species with bcr-abl, abl1 b or bcr target RNA as a function of the chain length of the antisense strands. (A) Time course of the composition of the single strand fraction and hybrid fractions of the above listed hybridization reactions by denaturing polyacrylamide gelelectrophoresis. The numbers on top of the lanes indicate the time points in minutes. Arrowheads indicate RNA length in nucleotides. Note that due to Phosphorlmager signals in the range of 20 to 30 nt in the hybrid fraction of the annealing reaction with abl1 b RNA it was possible to calculate rate constants. These signals, however, are not clearly visible in this figure. (B) Plot of the association rate constants ( k_obs) versus the length of the antisense strand.

Fig.4: Annealing of αARz72-derived antisense RNA species with bcr-abl, abl1 b or bcr target RNA as a function of the chain length of the antisense strands (for part A and B of this figure see notes in legend to Fig.3) . Fig.5: Association rate constants (k_obs) versus the length of the antisense strand for the annealing of αBA72-, αBA80-, and αBRz57-derived antisense RNA species with bcr-abl, abl1 b or bcr target RNA, respectively. Fig.6: Selectivity of RNaseH-mediated cleavage of bcr, abl1 b, and bcr-abl RNA sequences in the presence of the antisense oligodeoxyribonucleotides αBA23 and αBA28 respectively. Numbers on top of the lanes indicate the incubation times in minutes. The bcr-abl cleavage products P1 and P2 are indicated by solid arrow heads. The open arrowhead indicates an additional cleavage product.

Fig.7: Selectivity of RNaseH-mediated cleavage of bcr, abl1 b, and bcr-abl RNA sequences by αBA23 in the presence of varying amounts of nuclear extracts from human cells. The portion of nuclear extracts in the total reaction volume is indicated in % . Numbers on top of the lanes represent the incubation time in minutes. The bcr-abl cleavage products P1 and P2 are indicated by solid arrow heads.

Fig.8: Folding potential of the sequences 1 20 nt upstream and downstream of the bcr-abl fusion point. The calculation was performed as described recently

(Sczakiel et al., 1 993) at a window size of 50 nt and a step width of 1 nt. Note that the local minimum indicated at pos. 9 represents the ΔG value for the structure from pos. 9 to pos.58 which contains the structural element shown in the figure ( + 1 1 to + 58).

The following Examples illustrate the invention: RNA strands representing the abl1 b, bcr, and bcr-abl (b3/a2) sequences

The human Philadelphia-positive (Ph⁺ ) cell line K562 (Lozzio & Lozzio, 1 975) was used to isolate mRNA and to perform reverse transcription followed by PCR amplification of cDNA containing the abl1b, bcr, and bcr-abl sequences. The primers for PCR were chosen such that the resulting constructs included approximately 300 nucleotides upstream and 300 nucleotides downstream of the bcr- abl fusion point b3/a2 or the equivalent regions of the abl1 b RNA and of the bcr RNA, respectively (Fig. 1 A) . All three target transcripts were comprised of approximately 600 nucleotides for two reasons, (i) The authentic local structure in the vicinity of the fusion point sequence should be preserved by this RNA length. This view is based on evidence suggesting that RNA can possess a domain-like structure that is formed in the course of RNA transcription leading to independent folding units (Zarrinkar & Williamson, 1994). A stretch of 300 nt on either side of the fusion point presumably permits the formation of the naturally occuring three dimensional local RNA structure, (ii) A chain length of 600 nucleotides is at least four-fold greater than the length of the antisense species used here. Thus, in the course of the in vitro selection assay (Rittner et al. , 1993) the single strand fraction can be separated from the duplex fraction which is a methodological prerequisite.

For in vitro transcription of RNA, the amplified bcr-abl, abl1 b and bcr cDNA fragments were ligated into the plasmid 'bluescript' and the sequences of the resulting constructs were examined by sequence determination. The wild type sequences abl1 b and bcr, respectively, as well as a transcript representing the b3/a2 fusion were obtained . However, a nucleotide sequence comparison showed that the sequences isolated in this work differed from the original sequences. The abl1b cDNA sequence contained an additional ATG at the 3' end of exon l b (Shtivelman et al. , 1 986) . A similar observation was made by Bernards et al. ( 1 987) . The bcr cDN A showed an 'A' to 'G' exchange at positions 552 and 579 of the published bcr sequence (Heisterkamp et al. , 1985) . However, these differences are not relevant for the data presented here since they are located outside of the target region of the antisense and ribozyme constructs used here (Fig.1 ) .

Antisense RNA and asymmetric ribozymes directed against the bcr-abl fusion point sequence The cloned bcr-abl sequence pBSbcr/abl600 was used to generate plasmids for the in vitro transcription of three parental b cr-a bl-directed antisense RNA species and two asymmetric hammerhead ribozymes (Fig.1 B,.) . The three antisense species differ in the length of the portion which is complementary to abl sequences (1 2, 22, or 30 nt, respectively) and share the same bcr-directed sequence of 50 nt. Asymmetric hammerhead ribozymes bind via one antisense arm instead of two. They were shown to perform equally effective as their symmetric equivalents in vitro and in living cells (Tabler et al., 1 994). One bcr- ab l-directed ribozyme (αBRz57) used in this study was designed to bind via abl sequences and to cleave within the bcr portion. A second bcr-abl-directed ribozyme (αARz72) predominantly binds via bcr sequences and cleaves within the abl portion (Fig.1 C). Identification of selectively binding complementary RNA: association kinetics

The parental antisense RNA and ribozymes were synthesized by in vitro transcription and ³²P-labelled at one end. Limited alkaline hydrolysis resulted in pools of RNA species that have one end in common and are successively shortened at the opposite end. The asscociation rate constants for such related species were measured as schematically shown in Fig.2 at 37°C and physiological ionic strength (100 mM NaCl).

The association rate constants in relation to the chain length of related species derived from either of the three parental antisense constructs αBA62, αBA72, or αBA80 and the two parental asymmetric ribozymes αARz72 or αBRz57, respectively, are shown in Figs.3-5. The maximal association rate constants reached values of 1 to 3 x 10⁴ M^-1s^-1. For the 25mers to 31 mers derived from αBA62, association was almost fastest with the bcr-abl target but more than one order of magnitude slower with either of the bcr or abl1 b targets (Fig.3). Similarly, the even longer αBA72-derived species ranging between 34 and 39 nucleotides (Fig.5) as well as the αARz72- derived species ranging between 58 and 62 nucleotides in length showed kinetically selective annealing (Fig.4) . However, none of the αBRz57-derived species performed selectively and in the range of 53 nt to 62 nt, annealing with bcr-abl sequences was even slower than annealing with abl1b sequences (iig.5). Hammerhead ribozymes that cleave specifically bcr-abl fusion point sequences

For the two selected bcr-a bl-directed ribozymes αARz33 (59mer derived from αARz72) and αBRz42 (72mer derived from αBRz57), selectivity was measured under experimental conditions at which both, the association as well as the cleavage step contribute to specific destruction of either of the three targets bcr- abl, bcr, or abl1 b, respectively. Measurements were first performed at a temperature of 37°C and physiological ionic strength. However, to monitor more sensitively undesired cleavage of abl1 b or bcr sequences, the reactions were also performed at 50°C at which the cleavage reaction is faster. The half life of the bcr-abl target was in the range of 10 to 1 2 hours at 37 °C and approximately 2.5 hours at 50°C RNA (Tab.1 ) . Conversely, the half life of the bcr sequence was greater than 1 50 hours for both ribozymes. For αARz33 which does not only cleave the abl1 b sequence but also binds via a portion of 19 consecutive 3/b/1 b-directed bases (Fig.1 C), selectivity was temperature-dependent but not as pronounced as in case of αBRz42. At 37°C, using αARz33 the half life of the bcr-abl sequence was approximately 7.5-fold shorter than the half life of the abl1 b sequence and at least ten-fold shorter when compared with the half life of the bcr sequence (Tab.1 ). At 50°C, the difference between the half lives of the bcr-abl and the abl1 b sequence was reduced to a factor of only 2 (Tab.1 ) . No cleavage occured in the use of in vitro inactive derivatives of both ribozymes.

Tab.1 : Half lives of the cleavage reaction of the ribozymes αARz33 and αBRz42 with the bcr, abl1 b, and bcr-abl target strands.

Selective cleavage of bcr-abl RNA by antisense oligodeoxyribonucleotides and

RNaseH

The assay for identifying fast-annealing antisense species as shown in Figs.3-5 was performed with RNA. Deoxyribonucleotides (DNA) were used to monitor the selective annealing of antisense species derived from this experiments. The use of antisense DNA instead of synthetic RNA allows to follow the annealing reaction by RNaseH, which recognizes RNA-DNA hybrid strands of a minimal length of 6 to 8 base pairs. Under the experimental conditions given, the association step of the complementary nucleic acids was rate-limiting and not the RNaseH-mediated cleavage of the DNA-RNA heteroduplexes. All of the three targets (bcr-abl, bcr, and abl1 b) were incubated in the presence of the chemically synthesized oligodeoxyribonucleotides αBA23, αBA25, αBA28, or αBA30, as well as in the presence of RNaseH, respectively. Oligonucleotides αBA25 and αBA30 correspond exactly to the αBA62-derived RNA-25mer and -30mer (Fig.3). All αBA62-derived sequences have two additional non-complementary 'G'- nucleotides at their 5'-ends. To investigate their role two antisense oligonucleotides lacking the two 'G' at the 5'-terminus (αBA23 and αBA28) were synthesized. For the species αBA23 and αBA28,the results are shown in Fig.6.

All constructs strongly cleaved the bcr-abl target at conditions that resulted in poor or no significant degradation of bcr or abl1 b wild type sequences indicating that selective binding of the antisense species occured in a mixture of all potential target strands. When using oligonucleotides αBA28 and αBA30 with 1 6 nucleotides complementary to bcr, an additional cleavage product appeared (open triangle in Fig.6) that cannot be assigned to cleavage of the bcr RNA at or near the break point.

To investigate sequence selectivity under conditions that are closer to a living cell, the RNaseH assay with αBA23 and αBA25 was performed in the presence of nuclear extracts isolated from human cells. When nuclear extracts formed 3% to 30% of the total reaction volume, the bcr-abl RNA disappeared significantly faster than the bcr and abl RNA indicating selective degradation of the bcr-abl strand. In the absence of antisense-oligonucleotides and RNase H, degradation of all three RNAs occurred at an undistinguishable rate which increased at higher amounts of nuclear extracts. This result excludes an endogenous nuclease activity that preferentially degrades the bcr-abl RNA. Conversely, at increasing amounts of nuclear extract the bcr-abl RNA was degraded even faster relative to the abl and bcr RNA (Fig. 7). A quantification of band intesities by PhosphorImager (data not shown) indicated that at 10% nuclear extract, sequence selectivity of αBA23 was approximately 2-fold greater than measured with 3% nuclear extract (Fig. 7) . It became even greater at higher amounts of nuclear extract (30%, data not shown) which strongly suggests that sequence selectivity is at similar levels or even stronger in vivo.

Association rate constants of antisense oligodeoxyribonucleotides

To shed some light on the role of the 2'-OH group for association, the annealing kinetics of complementary DNA and RNA for the set of αBA62-derived antisense oligonucleotides with bcr-abl target RNA were compared (Tab.2). All DNA species annealed remarkably slow (10³ - 10⁴ M^{- 1}s^-1) with the bcr-abl target reflecting the limited accessibility of the fusion point sequences and 3- to 5-fold slower than the homologous complementary RNA in vitro under similar experimental conditions.

Due to prerequisites for efficient transcription in vitro, all antisense species contained two 'G' residues at their 5' end (αBA25 and αBA30) . In addition, the association kinetics were performed with αBA23 and αBA28, which were derivates lacking the two non-complementary 'G' residues at the 5'-end. The association rate constants of the two longer species αBA28 and αBA30 were similar regardless of the 5'-GG residues (ca. 3.6 x 10³ M^-1s^{- 1}, Tab.2) . For the shorter oligonucleotides αBA23 and αBA25, an increased association rate constant was measured for αBA23 which lacks the additional two 'G' residues

(6.8 x 10³ versus 4.4 x 10³ M^-1s^-1 ) indicating that in this case, the addition of the two non-complementary 'G' residues at the 5' end is of significant disadvantage for duplex formation in vitro.

Tab.2: Second order rate constants ( k_obs) of the association between antisense oligonucleotides and the bcr, abl1 b, and bcr-abl target strands.

Selectivity of antisense nucleic acids and hammerhead ribozymes

Selective binding in vitro was observed for short complementary sequences ( < 30 nt) as well as for longer complementary RNA strands ( > 30 nt) . This could not be explained by the equillibrium of the annealing reaction, i.e. the stability of the resulting duplex (T_m value) . Conversely, the association rate is a suitable parameter for assessing selectivity. It should be noted, however, that the oligodeoxyribonucleotides annealed slower in comparison with the analogous RNA. For instance the oligodeoxyribonucleotides αBA25 and αBA30 (Tab.2) annealed 3- to 5-fold slower than the respective RNA strands (Fig.3) . However, the comparable selectivity between tested oligomeric DNA and RNA indicates that th is parameter is independent of the type of nucleic acids used . Thus, the strategy used in this work is also valid for the large variety of chemically modified antisense nucleic acids and ribozymes. As confirmed by selective cleavage of bcr-abl RNA in the presence of RNaseH (data not shown), the two 'G' residues present at the 5' end of the constructs αBA25 and αBA30 had no significant effect on the selectivity. Nevertheless a slight decrease of annealing was observed in comparison with the derivatives lacking the two 'G' residues (Tab.2) .

In many cases, the transition from selective to non-selective complementary RNA species is sharp, i.e. the addition or deletion of only one to three nucleotides can significantly influence the kinetic selectivity as well as the association rate constant which is consistent with earlier studies in the use of short-chain antisense species (Southern et al., 1 992) as well as long-chain antisense RNA (Rittner et al. , 1 993).

Annealing of complementary RNA can be enhanced by facilitators such as hnRNP proteins (e.g. hnRNP A1 protein; ref. : Pontius & Berg, 1 990) or the compound cetyltrimethylammonium bromide in vitro (for review see: Pontius, 1 993) . Cellular factors could also lead to faster annealing in vivo. Assuming that the facilitator-mediated increase of annealing is influenced by the extent of sequence complementarity, selectivity could be increased as well. This view is consistent with the increased selectivity of the antisense species αBA23 in the presence of nuclear extracts (Fig. 7) .

For technical reasons, it was difficult to assess the selectivity of the shortened derivatives of the two asymmetric ribozymes. In an experiment with the αARz72-derived ribozyme αARz33, ribozyme-mediated cleavage of the abl1 b

RNA was observed. This observation might reflect the ability of αARz33 to form a 19 bp double strand with the abl1 b RNA in solution resulting in the cleavage of abl RNA. Annealing of αARz33 with either of the target sequences was not observed under the conditions used for the in vitro selection assay shown in Fig.4. Following the association reaction the samples were incubated under semi-denaturing conditions ( 1 % SDS, urea) prior to gelelectrophoresis which could explain the lack of binding . This finding implies that the in vitro selection method cannot be used for complementary species that are shorter than appro ximately 20 nt since the double strands formed are not stable. Conversely, for a chain length of greater than 25 bp the results are reliable and, in case of the antisense species, association was even measured for 20mers as for example for the annealing of the αBA72-derived 20mer with abl1 b (Fig.5).

The selectivity of αBRz42 in the cleavage experiment (Tab. 1 ) was greater than found in the annealing experiment (Fig. 7). This observation can be explained by the fact that the half life of intact target RNA in the cleavage experiment (Tab. 1 ) is a result of both, efficient annealing as well as cleavage. Cleavage by αBRz42 can only occur with either bcr-abl or bcr but not with abl sequences.

Thus, cleavage of the abl target cannot be detected and cleavage of the bcr target does not occur due to the fact that this RNA does not anneal (Fig. 7) .

Reduced accessibility of the bcr-abl b3/a2 fusion point sequence

The maximal association rate constants determined in this work( 1 to 3 x 10⁴ M- ¹s^-1) were not as great as the rate constants of naturally occuring complementary RNA which usually are in the range of 1 0⁵ to 10⁶ M^-1s^-1 (Wagner & Simons, 1994). The association rate constants that have been determined for artificial antisense RNA and ribozymes directed against HIV- 1 sequences were found to reach values between 10⁴ and 1 0⁵ M^-1s^-1 (Homann et al., 1 993b; summarized in: Sczakiel, in the press) . This observed decreased kinetic accessibility of the bcr- abl b3/a2 fusion sequence suggests that there is a similar inaccessibility in structural terms that could be caused by a relatively stable local structure. A stable local RNA structure could negatively influence the level of the annealing rate and, thus, the efficacy of complementary RNA in living cells. However, a reduced accessibility does not necessarily affect selectivity as can be derived from the results of this work. For example, the annealing pattern of the set of αBRz57-derived constructs with the three target strands (Fig.5) indicates that the accessibility (local structure) is different in the bcr-abl versus the abl1 b sequence (pos.53-63 in Fig.5) and strongly suggests the existence of structural differences in the relevant parts. The extent of RNA folding, i.e. the extent of intramolecular interactions can be monitored by the local folding potential. This computer-calculated parameter is a measure of the lowest possible free energy of a structure into which a given stretch of sequence can fold and gives some information on the probability of effective interactions between complementary RNA as well as the efficacy of antisense RNA in living cells (Sczakiel et al., 1993) . Thus, it seems to be reasonable to assume that the local folding potential can monitor the accessibility of a given sequence stretch. The folding potential for the bcr-abl sequences in general as well as in the vicinity of the fusion point is low (ΔG_{rnean(-300 to + 300)} = - 7.6 kcal/mol; ΔG_{mean(-120 to + 120} = -8.8 kcal/mol) when compared with other sequences indicating relatively extensive intramolecular interactions (Fig.8) : These results are consistent with the weak inhibition that was reported so far. The local minimum of the folding potential 9 nt downstream of the fusion point (ΔG = -1 6.8 kcal/mol, Fig.8) indicates that a stable secondary structure element can be formed in this region. In fact, the sequences between pos. + 1 1 and

+ 58 (Fig.8) downstream of the fusion point can form into a predicted low energy (stable) stem-bulge-stem-loop element that is also found as an individual structural element when the secondary structures are predicted for longer stretches of the bcr-abl sequence of up to the total length of 600 nt. All fast- hybridizing antisense species with a longer abl portion than 1 2 nt (αBA72, αBA80, αARz72 and αBRz57) anneal approximately two- to three-fold slower than the species that contain only 1 2 nt of abl sequences (αBA62j indicating that complementary sequences which extend into the above described element ( + 1 1 to + 58) do not support binding and might even be hindering.

The folding potential of the bcr-abl sequence shows higher values upstream of the fusion point at various window sizes. This means that the sequences in the upstream bcr portion do not seem to be involved as extensively in intramolecular folding and seem to be more available for intermolecular interactions, i.e. for initiation/elongation of double strand formation with antisense species as is indicated by the fast-hybridizing species αBA25 and αBA30 (Fig.3, Tab.2) . Implications for the design of b cr-ab l-directed complementary nucleic acids

The pairing reaction between two complementary strands consists of at least two critical steps. Firstly, the sequence-specific recognition of both strands via Watson-Crick base pairing and, subsequently, the initiation and elongation of duplex formation. Both steps may occur via the same bases which seems to be likely in case of short antisense sequences or, in specific cases, even with longer antisense RNA (Homann et at. , 1993b). However, this is not an absolute requirement and in case of naturally-occuring complementary RNA, initial inter- actions take place via reversible loop-loop interactions ('kissing') followed by initiation of double strand formation at another site, often involving the free 3'- end of the antisense strand and an accessible complementary portion of the target strand (for review see: Wagner & Simons, 1 994). Ideally, antisense species as well as ribozymes should be designed such that accessible sequences of the target strand meet accessible complementary sequences.

In living cells, however, duplex formation alone does not seem to be sufficient for efficacy, i.e. the destruction of the target. In case of antisense DNA (oligodeoxyribonucleotides), presumably RNaseH is responsible for degradation of the target RNA. In case of antisense RNA, it is speculated that an RNaselll-like activity recognizes perfectly formed duplex RNA that exceeds a certain minimal length in the range of 25 to 30 base pairs (Nellen & Sczakiel, in the press). For this reason, short-chain ( < 30 nt) antisense species that had been identified by kinetic selection have to be synthesized as deoxyribonucleotides whereas long- chain species ( > 30 nt) may also be synthesized and applied as RNA.

The limited kinetic and structural accessibility of the sequences in the vicinity of the bcr-abl fusion point fit with the weak or almost absent antisense-mediated inhibition of bcr-abl gene expression observed in human cells. This suggests a closer look at the local structures of the fusion region sequences. The local minimum at pos + 9 is clearly disadvantageous as a point for antisense recognition and sequences further downstream are unlikely to be suitable for selectivity. For this reason, selective and efficient antisense sequences should be directed against the bcr portion neighbouring the fusion point as well as the first 8 nucleotides of the abl sequence.

Construction of templates for in vitro synthesis of target RNA, antisense RNA and ribozymes.

The cDNA for synthesis of the target RNA bcr-abl, abl1 b and bcr were derived from the CML cell line K562 (Lozzio & Lozzio, 1 975) by RT-PCR. Cellular RNA of K562 was extracted as described (Chomczynski & Sacchi, 1 987) . Standard methods for reverse transcription and PCR were applied. For synthesis of the bcr-abl cDNA primers 93/25 and 93/21 , for abl1 b cDNA primers 93/21 and 93/22 and for bcr cDNA primers 93/25 and 93/26 were used . The sequences of the primers were: 93/21 : 5'-AGGAGTGTTTCTCCAGACTG-3' , 93/22: 5'-TGC- TTCCTTTTGTTATGGAA-3', 93/25: 5'-ATGTCTCCCAGCATGGCCTT-3', 93/26: 5'-TTACTTCGATCCCATTCATG-3'. The resulting PCR products were blunt ended by mung bean nuclease and cloned into the Smal-cleaved Bluescript M 1 3 (Stratagene ) , yielding the plasmids pBSbcr/abl600, pBSabl 1 b603 and pBSbcr600. The bcr-abl sense RNA could be transcribed in vitro by T7 RNA polymerase, the abl1b and the bcr sense RNA by T3 RNA polymerase (Fig. 1 A) . Proceeding from pBSbcr/abl600, we constructed the cDNA of three bcr-abl- directed antisense RNA species by PCR. An unique bcr-directed primer (bcr50) hybridizing 50 nucleotides upstream of the fusionpoint and containing a Ppu10I restriction site at the 5'-end was used for PCR amplification of the cDNA. The abl-directed primers hybridized 12 (abl12), 22 (abl22) and 30 (abl30) nucleotides downstream of the fusion point, respectively, and contained the sequence of the

T7 promotor and a Xbal site. Additionally, using primer bcr50 and primer rzabl23, which contained the ribozyme sequences as well as T7 promotor sequences, the cDNA for a hammerhead ribozyme was designed by PCR. This ribozyme binds predominantly via the bcr portion of the bcr-abl RNA and cleaves within the abl portion. The cDNA for a second ribozyme, which cleaved within the -bcr-portion and used a bl-directed antisense sequences for binding, was constructed in the similar manner using primers all50 (containing the T7 sequence) and rzbcr7 (containing the ribozyme sequence; Fig.1 B,C). The

The PCR products were cloned into pUC 1 31 using Xbal and Ppu 10I sites resulting the plasmids pBA62, pBA72, pBA80, pARz72 and pBRz57. All plasmids were controlled experimentally by sequence analysis and the antisense RNA and ribozymes αBA62, αBA72, αBA80, αARz72 and αBRz57 were transcribed in vitro by T7 RNA polymerase.

For in vitro transcription of αARz33 and αBRz42 as well as their corresponding in vitro inactive derivatives, PCR products containing a T7 promotor were used . The PCR fragment for αARz33 was amplified from the plasmid pARz72 using primer bcrl 1 (5'-GCCAAGCTTGCAGAGTTCAAAAGCCCTTC-3') and rzabl23_a (-5'-GCCTCTAGATAATACGACTCACT-3') or rzabl23_i (in vitro inactive; 5'-GCC-

TCTAGATAATACGACTCACTATAGGGCTGCCTAATAAGGCCT-3') . PCR fragment for αBRz42 was synthesized proceeding from the plasmid pBRz57 using primer abl36 (5'-GCCTCTAGATAATACGACTCACTATAGGGCGCTCAAAGT- CAGATG

CTACT-3') and rzbcr7_a (5'-GCCAAGCTTAGTTTCGGCCTCGAGGCCTC-3') or rzbcr7, (in vitro inactive; 5'-GCCAAGCTTAGTTTCGGCCTCGAGGCC7TATTAGC- AAA-3'). The sequences of both in vitro active ribozymes are: (i) αARz33: 5'- GGGCUGCCUGAUGAGGCCUCGAGGCCGAAACUGGCCGCUGAAGGGCUUUU- GAACUCUGCAAGCU-3', (ii) αBRZ42: 5'-GGGCGCUCAAAGUCAGAUGCUACUG- GCCGCUGAAGGGCUUUUGCUGAUGAGGCCUCGAGGCCGAAACUAAGCU-3'. Helix 2 sequences are indicated in bold letters. The underlined G-residues are exchanged to A-residues for in vitro inactive ribozymes.

In vitro transcription of RNA

Plasmid DNA was linearized by the following enzymes prior to in vitro transcription: pBSbcr/abl600 by EcoRl, pBSbcr600 by BamHl and pBSabl603 by BamHl or Sstl followed by filling in of the 3'-protruding ends of the Sstl-site by the

Klenow enzyme. The plasmids pBA62, pBA72, pBA80, pARz72 and pBRz57 were linearized by Hindlll. PCR products were cleaved with Hindlll before in vitro transcription. T7 RNA polymerase (Boehringer Mannheim) was used for in vitro transcription of PCR products and plasmid DNA except pBSabl1 b603 and pBSbcr600, which were transcribed by T3 RNA polymerase (Boehringer Mannheim) . Five μg of linearized template DNA were transcribed in vitro as described (Rittner et al., 1 993) . After adding of 200 μl 20 mM MgSO₄ the mixtures were treated with 20 U DNasel (Boehringer Mannheim) and the transcripts were purified by gel filtration (Sephadex G-50, Pharmacia; TE-buffer ( 10 mM Tris/HCI pH 8, 1 mM EDTA)) . For in vitro transcription of radiolabelled RNA 0.5 μg of template DNA were used under same conditions as described above. Unlabelled nucleoside triphosphates ATP, CTP, GTP at final concentrations of 1 .25 mM, unlabeled UTP at a final concentration of 0.04 mM and 20 μCi of [α-³²P]UTP (3000 Ci/mmol, Amersham, Braunschweig) were included in 20 μl reactions. RNA was purified by gel filtration.

³²P-end labelling of RNA The 5'-ends of in vitro transcribed RNA ( 10 pmol) were ³²P-labelled by dephos- phorylation with calf intestinal phosphatase and subsequent rephosphorylation with 50 μCi of [γ-³²P]ATP (3000 Ci/mmol) and polynucleotide kinase (Boehringer Mannheim) as described (Sambrook et al., 1 989) .

3'-labelling of RNA was performed as described by Barrio et al. (1 978). Briefly, 5 pmol of RNA were incubated for 14 hours at 1 5 °C with 330 μM ATP, 50 mM Hepes-buffer pH 8.3, 20 mM MgCI₂, 10 units T4 RNA ligase (Boehringer Mannheim) and 50μCi of [³²P]pCp (3000 Ci/mmol; Amersham, Braunschweig) in a final volume of 20 μl. RNA was purified by gel filtration.

Alkaline hydrolysis of antisense RNA and ribozymes

In order to generate a random mixture of bcr-ab/-d\rec\ed antisense RNA and ribozymes the ³²P-end-labelled parental RNA and ribozymes were successively shortened by alkaline hydrolysis as described (Rittner et al. , 1 993). Briefly, 2.5 pmol of end labelled RNA in TE-buffer ( 10 mM Tris/HCI pH 8.0, 1 mM EDTA) were heated with 1 .5 volumes of 0.5 M NaHCO₃ to 96°C for 1 2 to 14 minutes, then chilled on ice and desalted by gel filtration. After ethanol precipitation, the RNA was dissolved in TE-buffer. Subsequently, the mixtures of RNA species were heated to 75 °C for 10 minutes and cooled slowly to 37°C before using in hybridization assays.

In vitro selection method for identification of fast-hybridizing antisense RNA species

The experimental procedure was described recently (Rittner et al. , 1 993) and is schematically depicted in Fig. 2. Briefly, 5'- or 3'-labelled hydrolysed RNA was incubated with 1 .6 pmol, 3.2 pmol or 6.4 pmol of bcr-abl, abl1 b or bcr target RNA, respectively, in a final volume of 20 μl of a solution containing 100 mM NaCl, 20 mM Tris/HCI pH 7.4 and 10 mM MgCI₂ at 37 °C. At different time points of incubation, 3 μl aliquots were transferred into 30 μl stop buffer (50 mM Tris/HCI pH 8; 1 5 mM EDTA, 0.2% SDS, 8 M urea, 0.04% bromphenolblue, 0.04% xylenecyanol) precooled on ice. Single stranded antisense RNA and target RNA/antisense RNA duplexes were separated by native agarose gelelectrophoresis ( 1 .2%, 89 mM Tris/HCI pH 8,3, 89 mM boric acid, 2 mM EDTA). Single stranded and duplex RNA was excised from the gel and recovered by centrifugation of the excised gel slices which had been frozen in liquid nitrogen. After precipitation with ethanol, RNA was redissolved with stop buffer and analysed by polyacrylamide gelelectrophoresis under denaturing conditions (12% polyacrylamide gels containing 7 M urea in 89 mM Tris-borate buffer pH 8.3). Gels were dried and exposed to X-ray film. Determination of hybridization rates for individual antisense RNA species

For quantitative analysis of band intensities, dried polyacrylamide or agarose gels were scanned by a Phosphorlmager (Molecular Dynamics) . Using the 'IMAGE QUANT' software (Molecular Dynamics), band intensities were measured . The data were transferred to the programme 'EXCEL' (Microsoft) . Band intensities of hybridization or ribozyme cleavage reactions of individual antisense RNA or ribozyme species were plotted against the time axis. A curve for an exponential decay was fitted by non linear regression using the programme 'GRAFIT' (Erithacus Software, London, UK).

RNaseH cleavage reactions

The three ³²P-labelled substrate RNA species bcr-abl, abl1 b and bcr (100 pM) were incubated at 37°C with the different phosphorothioate antisense oligonucleotides αBA23 (5'-GCTGAAGGGCTTTTGAACTCT-3'), αBA25 (5'-GGGCTG-

AAGGGCTTTT

GAACTCTGC-3'), αBA28 (5'-GCTGAAGGGCTTTTGAACTCTGCTTAAA-3') or αBA30 (5'-GGGCTGAAGGGCTTTTGAACTCTGCTTAAA-3'), respectively (final concentration 1 μM) and 1 U Escherichia coli RNaseH (Boehringer Mannheim) in a- final volume of 20 μl containing 100 mM NaCl, 20 mM Tris/HCI pH 7.4 and 10 mM MgCI₂ At different time points of incubation, 3 μl aliquots were transferred into 30 μl stop buffer (50 mM Tris/HCI pH 8; 1 5 mM EDTA, 0.2% SDS, 8 M urea, 0.04% bromphenolblue, 0.04% xylenecyanol) precooled on ice. After heating at 95 °C for 5 minutes the reaction products were analysed by polyacrylamide gelelectrophoresis (4%) under denaturing conditions and subsequent autoradiography of the dried gel. The experiments with nuclear extracts were performed such that the reaction mixtures contained different amounts (3%, 10% or 30%) of nuclear extracts from human SW480 cells (3.5 mg/ml protein) . In order to remove the proteins from samples withdrawn at different time points of the RNaseH reaction, an extraction with phenol/chlorophorm/isoamylalcohol (25:24: 1 ) was performed before adding to 30 μl stop buffer. Nuclear extracts were prepared as described

(Schreiber et al., 1 989).

Ribozyme activity in vitro Ribozyme activity was measured under single turnover conditions. The preparation of RNA, the cleavage reaction, and the separation of products by polyacrylamide gelelectrophoresis was performed as described recently (Homann et al., 1 993a). Briefly, 30 fmol (1 .5 nM) of radioactively labelled target RNA and an at least ten-fold excess of unlabelled ribozyme RNA ( > 300 fmol; > 15 nM) were incubated at 37 °C or at 50°C in a final volume of 20 μl containing 100 mM

NaCl, 10 mM MgCI₂, and 20 mM Tris/HCI pH 7.4. Aliquots of the reaction mixture were withdrawn at different time points and the reaction was stopped by adding 30 μl of a solution containing 50 mM Tris/HCI pH 8; 1 5 mM EDTA, 0.2% SDS, 8M urea, 0.04% bromphenolblue and 0.04% xylenecyanol. Samples were stored at 0°C and analysed by electrophoresis under denaturing conditions

(8M urea) with 5% polyacrylamide gels (0, 1 25% bisacrylamide) . Gels were dried and scanned by a Phosphorlmager.

Determination of association rate constants of antisense oligodeoxyribonucleo- tides

Radioactively labelled phosphorothioate antisense oligodeoxynucleotides ( 1 nM final concentration) were incubated with either of the three target RNA bcr-abl, ab/1 b or bcr (300 nM or 450 nM final concentration) in a total volume of 20 μl of hybridization buffer (see above). After different incubation times, 3 μl aliquots were withdrawn, transferred into 25 μl of pre-cooled stop buffer (see above), and analysed by native agarose gelelectrophoresis. Gels were dried, exposed to X-ray film and scanned by a Phosphorlmager.

Ex vivo tests with human cells

Patients and cells

Human peripheral blood mononuclear cells (MNC) were obtained from patients with chronic myelogneous leukemia in chronic phase, in accelerated phase or in blast crisis. Red blood cells and cell debris were removed by density centrifugation using the lymphocyte separation medium Lymphoprep (Nycomed Pharma, Oslo, Norway) . Separation of CD34 + cells from the MNC fraction was performed using the miniMACS immunomagnetic separation system (Miltenyi Biotec, Bergisch Gladbach, Germany) according to the manufacturer's instructions. For transfection experiments cells were cultured in RPMI- 1 640-Medium supplemented with10% heat-inactivated fetal calf serum, 100 lU/ml penicillin, 100 μg/ml streptomycin, 2 mmol/L L-glutamine.

Oliqodeoxyribonucleotodes (ODN)

The ODN used for transfection studies had the following sequences:

b3a2-directed: αBA23: 5'-gctgaagggcttttgaactctgc-3'

αBA28: 5'-gctgaagggcttttgaactctgcttaaa-3'

b2a2-directed: αb2a2-26: 5'-cgctgaagggcttcttccttattgat-3'

scrambled controls: nsBA23: 5'-ttattgagggtgatccgctagcc-3'

nsBA28: 5'-agaggtcacgcttttagagattgcttca-3' nsb2a2-26: 5'-tggtcatacaggcctatttcgtcttg-3'

The two internucleotide linkages at the 3' and 5' ends of the ODN were phosphorothioates, the internal deoxyribonucleotides were connected by phosphodie sters. The ODN were purified by reverse phase high performance liquid chromatography and lyophilized after synthesis.

Treatment of CML cells by ODN

Before transfection 2.5 μl ODN (200 μM, corresponding to a final concentration of the ODN in 500 μl cell suspension of 1 μM) were mixed with 1 5 μl DOTAP (N- [1 -(2,3-Dioleoyloxy)propyl]-N,N,N-trimethylammonium methyosulfate, Boehringer Mannheim, Mannheim, Germany) and Hepes-buffer (20 mM, pH 7.4) to a final volume of 75 μl. The mixture was incubated for 1 5 minutes at room temperature for formation of ODN/cationic lipid complexes and added dropwise to the cell suspension. The final volume of MNC as well as CD34 + cells was 500 μl at a final denisty of 1 x 10⁶ cells/ml. Cells were incubated with the ODN for 6 hours at 37°C. Cells were pelleted by centrifugation and resuspended in fresh medium. After 1 8 hours a second transfection was performed using the half amount of

ODN/DOTAP complexes. Then the cells were added to MethoCult H4433 "Complete" Methylcellulose Medium (StemCell Technologies Inc., Vancouver, Canada) and plated in duplicates. Cell densities in methylcellulose medium: MNC (1 x 10⁵/ml), CD34 + (1 x 10³/ml) . Colonies were counted after 14 days.

Reverse transcription and polymerase chain reaction

For analysis of the bcr-abl fusion point RNA from 1 x 10⁷ cells was extracted . The fusion point was determined by reverse transcription, followed by polymerase chain reaction and analysis by agarose gel electrophoresis.

Tab. 3: Specific inhibition of clonogenic growth of primary CML cells by treatment with bcr-ab l-directed antisense oligodeoxyribonucleotides.

Abbreviations: CP: chronic phase

AP: accelerated phase

BC: blast crisis

MNC: mononuclear cells (peripheral blood)

CD34 + : CD34-enriched hematopoietic cells from peripheral blood

REFERENCES Barrio, J.R. , Barrio, M.C., Leonard, N.J., England, T.E. & Uhlenbeck O.C.

( 1978) . Synthesis of modified nucleoside 3', 5'-biphosphates and their incorporation into oligoribonucleotides with T4 RNA ligase. Biochemistry 17, 2077- 2081 . Bernards, A., Rubin , CM., Westbrook, C.A., Paskind, M. & Baltimore, D.

(1987) . The first intron in the human c-abl gene is at least 200 kilobases long and is a target for translocations in chronic myelogenous leukemia. Mol. Cell. Biol. 1, 3231 -3236. Chomczynski, P. & Sacchi, N. (1987). Single-step method of RNA isolation by acid guanidinium thiocyanate-phenol-chloroform extraction. Anal. Biochem. 162, 156-159. Daley, G.Q., Van Etten, R.A. & Baltimore D. (1990). Induction of chronic myelogenous leukemia in mice by the P210 bcr/abl gene of the Philadelphia chromosome. Science 247, 824-830.

Heisterkamp, N., Stam, K., Groffen, J., de Klein, A. & Grosveld, G. (1985). Structural organization of the bcr gene and its role in the Ph' translocation.

Nature 315, 758-761.

Helene, C. & Toulme, J.J. (1990) Specific regulation of gene expression by antisense, sense and antigene nucleic acids. Biochim. Biophys. Acta 1049, 99- 125.

Hertel, et al., (1992), Numbering system for the hammerhead. Nucl. Acids Res. 20, 3252. Homann M, Tzortzakaki S, Rittner K, Sczakiel,G. & Tabler,M. (1993a) Incorporation of the catalytic domain of a hammerhead ribozyme enhances the inhibitory effect on the replication of the human immunodeficiency virus type 1. Nucl. Acids Res.21, 2809-2814. Homann, M., Rittner, K. & Sczakiel,G. (1993b) Complementary large loops determine the rate of duplex formation in vitro in case of an effective antisense RNA directed against the human immunodeficiency virus type 1. J. Mol. Biol. 233, 7-15. James, W. & Al-Shamkhani, A. (1995). RNA enzymes as tools for gene ablation.

Curr. Opin. Biotechnol.6, 44-49.

Kaebisch, A., Perenyi, L., Seay, U., Lohmeyer, J. & Pralle, H. (1994). Unmod ified phosphodiester antisense oligodeoxynucleotides to the BCR-ABL junction do not suppress Philadelphia-positive clonogenic cells. Acta Haematol. 92, 1 90- 1 96. Kearney, P., Wright, L.A., Milliken, S. & Biggs, J.C. ( 1 995) . Improved specificity of ribozyme-mediated cleavage of bcr-abl mRNA. Exp. Hematol. 21 , 986-989.

Kelliher, M.A. , McLaughlin, J. , Witte, O.N. & Rosenberg, N. ( 1 990). Induction of a chronic myelogenous leukemia-like syndrome in mice with v-abl and BCR/ABL. Proc. Natl. Acad. Sci. USA 87, 6649-6653.

Konopka, J.B. & Witte, O.N. ( 1 985) . Detection of c-abl tyrosine kinase activity in vitro permits direct comparison of normal and altered abl gene products. Mol. Cell Biol. 5, 31 1 6-31 23.

Kurzrock, R., Gutterman, J.U. & Talpaz M. ( 1 988). The molecular genetics of Philadelphia chromosome-positive leukemias. N. Engl. J. Med. 319, 990-998.

Lozzio, C.B. & Lozzio, B.B. ( 1 975). Human chronic myelogenous leukemia cell line with positive Philadelphia chromosome. Blood AS, 321 -334.

Maekawa, T., Kimura, S., Hirakawa, K., Murakami, A., Zon, G. & Abe, T. (1995) . Sequence specificity on the growth suppression and induction of apop- tosis of chronic myeloid leukemia cells by bcr-abl anti-sense oligodeoxynucleo- side phosphorothioates. Int. J. Cancer 62, 63-69.

Marshall, P., Thompson, J.B. & Eckstein, F. ( 1 994) . Inhibition of gene expression with ribozymes. Cell. Mol. Neυrobiol. 14, 5523-5538. Martiat, P. , Lewalle, P. , Taj, A.S., Philippe, M., Larondelle, Y., Vaerman, J.L. ,

Wildmann, C , Goldman, J.M. & Michaux, J.L. ( 1 993). Retrovirally transduced antisense sequences stably suppress p210 bcr-abl expression and inhibit the proliferation of BCR/ABL-containing cell lines. Blood 81 , 502-509. Maru, Y. & Witte, O.N. ( 1 991 ). The BCR gene encodes a novel serine/threonine kinase activity within a single exon. Cell 67, 459-468.

Maru, Y., Peters, K.L. , Afar, D.E.H., Shibuya, M., Witte, O.N. & Smithgall, T.E. (1995) . Tyrosine phosphorylation of BCR by FPS/FES protein-tyrosine kinases induces association of BCR with GRB-2/SOS. Mol. Cell. Biol. 15, 835-842.

Nellen,W. & Sczakiel,G. Mechanisms of antisense RNA-mediated gene silencing, Molecular Biotech. , (in press).

Nichols, G.L. ( 1995) . Antisense oligonucleotides as therapeutic agents for chronic myelogenous leukemia. Antisense Res. Dev. 5, 67-69.

O'Brien, S.G. , Kirkland, M.A., Melo, J.V., Rao, M.H. , Davidson, R.J. , McDonald, C. & Goldman, J.M. ( 1 994). Antisense bcr-abl oligomers cause non-specific inhibition of chronic myeloid leukemia cell lines. Leukemia 8, 21 56-21 62.

Pachuk, C.J., Yoon, K., Moelling, K. & Coney, L.R. ( 1994). Selective cleavage of bcr-abl chimeric RNAs by a ribozyme targeted to non-contiguous sequences. Nucl. Acids Res. 22, 301 -307.

Pontius, B.W. & Berg, P. (1990). Renaturation of complementary DNA strands mediated by purified mammalian heterogenous nuclear ribonucleoprotein A1 protein: Implications for a mechanism for rapid molecular assembly. Proc. Natl. Acad. Sci. USA 87, 8403-8407.

Pontius, B.W. (1 993). Close encounters: Why unstructured, polymeric domains can increase rates of specific macromolecular association. Trends Biochem. Sci. 18, 1 81 -1 86.

Rittner, K. , Burmester, C. & Sczakiel, G. (1 993) . In vitro selection of fast-hybridizing and effective antisense RNAs directed against the human immunodeficiency virus type 1 . Nucl. Acids Res. 21 , 1 381 -1 387. Sambrook, J., Fritsch, E.F. & Maniatis F. (1 989) . Molecular cloning: a laboratory manual. Cold Spring Harbor Laboratory Press, New York.

Sawyers, C.L., McLaughlin, J.,Goga, A., Havlik, M. & Witte, O.N. ( 1994) . The nuclear tyrosine kinase c-abl negatively regulates cell growth. Cell 77, 1 21 -1 31 .

Schreiber, E., Matthias, P., Mϋller, M.M. & Schaffner, W (1989). Rapid detection of octamer binding proteins with 'mini-extracts', prepared from a small number of cells. Nucl. Acids Res. 17, 6419.

Schwab, G., Chavany, C, Duroux, I., Goubin, G., Lebeau, J., Helene, C. & Saison-Behmoaras, T. (1994) . Antisense oligonucleotides adsorbed to polyalkyl- cyanoacrylate nanoparticles specifically inhibit mutated Ha-ras-mediated cell proliferation in nude mice and tumorigeneity in nude mice. Proc. Natl. Acad. Sci. USA 91 , 10460-10464.

Sczakiel, G., Homann, M. & Rittner, K. (1993). Computer-aided search for effective antisense RNA target sequences of the human immunodeficiency virus type 1 . Antisense Res. Dev. 3,45-52.

Sczakiel, G. & Nedbal, W. ( 1995) . The potential of ribozymes as antiviral agents. Trends Microbiol. 3, 21 3-21 7.

Sczakiel, G. Hammerhead ribozymes with long antisense flanks: domain structure and binding kinetics. In: Eckstein, F. & Lilley,D.M.J. (eds. ) : Nucl. Acids and Mol.

Biol., Springer-Verlag, Heidelberg (in press).

Shtivelman, E. , Lifshitz, B., Gale, R.P., Roe, B.A. & Canaani, E. ( 1 986) . Alternative splicing of RNAs transcribed from the human abl gene and from the bcr-abl fused gene. Cell 47, 277-284.

Skorski, T., Nieborowska-Skorska, M., Barletta, C, Malaguarnera, L., Szczylik, C , Chen, S.-T., Lange, B. & Calabretta, B. ( 1 993) . Highly efficient elimination of Philadelphia¹ leukemic cells by exposure to bcr/abl antisense oligodeoxynucleotides combined with mafosfamide. J. Clin. Invest. 92, 1 94-202.

Skorski, T., Nieborowska-Skorska, M., Nicolaides, N.C., Szczylik, C, Iversen, P., lozzo, R.V., Zon, G. & Calabretta, B. ( 1994). Suppression of Philadelphia¹ leukemia cell growth in mice by BCR/ABL antisense oligodeoxynucleotide. Proc. Natl. Acad. Sci. USA 91 , 4504-4508.

Smetsers, T.F.C.M., Skorski, T., van de Locht, L.T.F., Wessels, H.M.C., Pennings, A.H.M., de Witte, T., Calabretta, B. & Mensink, E.J.B.M. (1994). Anti- sense BCR-ABL oligonucleotides induce apoptosis in the Philadelphia chromosome-positive cell line BV1 73. Leukemia 8, 1 29-140.

Smetsers, T.F.C.M., van de Locht, L.T.F. , Pennings, A.H.M. , Wessels, H.M.C., de Witte, T.M. & Mensink, E.J.B.M . ( 1 995) . Phosphorothioate BCR-ABL anti- sense oligonucleotides induce cell death, but fail to reduce cellular bcr-abl protein levels. Leukemia 9, 1 1 8- 1 30.

Snyder, D.S. , Wu, Y., Wang, J.L. , Rossi, J.J., Swiderski, P. , Kaplan, B.E. & Forman, S.J. ( 1 993). Ribozyme-mediated inhibition of bcr-abl gene expression in a Philadelphia chromosome-positive cell line. Blood 82, 600-605.

Southern, E. M., Maskos, U. & Elder, K. ( 1 992) . Analyzing and comparing nucleic acid sequences by hybridization to arrays of oligonucleotides: evaluation using experimental models. Genomics 13, 1008-101 7.

Stein, C. A. & Chang, Y.-C. (1993) . Antisense oligonucleotides as therapeutic agents - is the bullet really magical? Science 261 , 1 004-101 2. Szczylik, C , Skorski, T., Nicolaides, N.C. , Manzella, L. , Malaguarnera, L.,

Venturelli, D. , Gewirtz, A.M. & Calabretta, B. ( 1991 ) . Selective inhibition of leukemia cell proliferation by BCR-ABL antisense oligodeoxynucleotides. Science 253, 562-565. Tabler, M., Homann, M., Tzortzakaki, S. & Sczakiel, G. (1994). A three nucleotide helix I is sufficient for full activity of a hammerhead ribozyme: advantages of an asymmetric design. Nucl. Acids. Res. 22, 3958-3965. Tari, A.M., Tucker, S.D., Deisseroth, A. & Lopez-Berestein, G. (1994). Liposomal delivery of methylphosphonate antisense oligodeoxynucleotides in chronic myelogenous leukemia. Blood 84, 601-607.

Thomas, M., Kosciolek, B., Wang, N. & Rowley, P. (1994). Capping of bcr-abl antisense oligonucleotides enhances antiproliferative activity against chronic myeloid leukemia cell lines. Leuk. Res. 18, 401-408.

Vaerman, J.L., Lammineur, C, Moureau, P., Lewalle, P., Deldime, F., Blumenfeld, M. & Martiat, P. (1995). BCR-ABL antisense oligodeoxyribonucleotides suppress the growth of leukemic and normal hematopoietic cells by a sequence- specific but nonantisense mechanism. Blood, 86, 3891-3896.

Voncken, J.W., van Schaick, H., Kaartinen, V., Deemer, K., Coates, T., Landing, B., Pattengale, P., Dorseuil, O., Bokoch, G.M., Groffen, J & Heisterkamp, N. (1995). Increased neutrophil respiratory burst in bcr-Nu\\ Mutants. Cell SO, 719-

728.

Wagner, E.G.H. & Simons, R.W. (1994). Antisense RNA control in bacteria, phage and plasmids. Annu. Rev. Microbiol.48, 713-740.

Wright, L., Wilson, S.B., Milliken, S., Biggs, J. & Kearney, P. (1993). Ribozyme- mediated cleavage of the bcr/abl transcript expressed in chronic myeloid leukemia. Exp. Hematol. 21, 1714-1718.

Zarrinkar, P.P. & Williamson, J.R. (1994). Kinetic intermediates in RNA folding. Science 265, 918-924.

Claims

1 . A nucleic acid comprising a nucleotide sequence containing (a) a portion complementary to a first chromosomal DNA sequence, and

2. The nucleic acid according to claim 1 , wherein the first and second chromosomal DNA sequences derive from the same or different chromosome(s).

3. The nucleic acid according to claim 1 or 2, wherein the chromosomal translocation is t(9;22) .

4. The nucleic acid according to anyone of claims 1 to 3, wherein the fusion gene is the bcr-abl fusion gene.

5. The nucleic acid according to anyone of claims 1 to 4, wherein the nucleotide sequence consists substantially of ribonucleotides.

6. The nucleic acid according to claim 5, wherein the nucleotide sequence contains at least one of the following sequences:

7. The nucleic acid according to claim 5, wherein the portion (a) and/or the portion (b) contain(s) the catalytic domain of a hammerhead ribozyme.

8. The nucleic acid according to claim 7, wherein the portion (a) forms at least part of the Helix I- and/or Helix Ill-forming region of a hammerhead ribozyme.

9. The nucleic acid according to claim 7 or 8, wherein the portion (b) forms at least part of the Helix I- and/or Helix Ill-forming region of a hammerhead ribozyme.

10. The nucleic acid according to anyone of claims 7 to 9, wherein the nucleotide sequence contains at least one of the following sequences:

1 1 . A DNA sequence which upon transcription corresponds to the nucleic acid of anyone of claims 5 to 10.

12. A vector comprising the nucleic acid of anyone of claims 1 to 10 or a DNA sequence of claim 1 1 .

1 3. A host organism containing the nucleic acid of anyone of claims 1 to 1 0, the DNA sequence of claim 1 1 or the vector of claim 1 2.

14. A method for the production of the nucleic acid of anyone of claims 1 to 10, a DNA sequence of claim 1 1 or the vector of claim 1 2, comprising cultivating the host organism of claim 13, under suitable conditions and isolating the desired product from the culture.

15. A pharmaceutical composition containing the nucleic acid of anyone of claims 1 to 10, the DNA sequence of claim 1 1 or the vector of claim 1 2, optionally in association with a pharmaceutically acceptable carrier and/or diluent.

16. The pharmaceutical composition according to claim 1 5 for the treatment of disorders based on chromosomal translocations.

17. The pharmaceutical composition according to claim 1 6, wherein the disorder is chronic myelogenous leukemia.