Allosterically Controlled Single-Chained Maxizymes with Extremely

Biomacromolecules 2001, 2, 788-799

788

Allosterically Controlled Single-Chained Maxizymes with Extremely High and Specific Activity Tomoko Kuwabara,†,‡,§ Makiko Hamada,†,§ Masaki Warashina,† and Kazunari Taira*,†,‡ Gene Discovery Research Center, National Institute of Advanced Intdustrial Science and Technology (AIST), 1-1-4 Higashi, Tsukuba Science City 305-8562, Japan; and Department of Chemistry and Biotechnology, Graduate School of Engineering, The University of Tokyo, Hongo, Tokyo 113-8656, Japan Received March 15, 2001

For the treatment of chronic myelogenous leukemia (CML), attempts have been made to design various ribozyme motifs that can specifically recognize and cleave BCR-ABL fusion mRNAs. In the case of L6 BCR-ABL b2a2 mRNA, it is difficult to cleave the abnormal mRNA specifically because the mRNA includes no sequences that can be cleaved efficiently by conventional hammerhead ribozymes near the BCR-ABL junction. We recently succeeded in designing a novel maxizyme, which specifically cleaves BCR-ABL fusion mRNA, as a result of the formation of a dimeric structure [Kuwabara, T.; et al. Mol. Cell 1998, 2, 617627; Tanabe, T.; et al. Nature 2000, 406, 473-474]. Specifically, we tailored the maxizyme with molecular switching function: the maxizyme splices a cleavable GUC site, but only when it appears within a strand of mRNA that possesses the abnormal splice junction. We demonstrated that this approach is generalizable [Tanabe, T.; et al. Biomacromolecules 2000, 1, 108-117]. All the maxizymes designed in the past functioned as a result of the formation of a dimeric structure. Questions have been asked whether a similar molecular switching might be possible within a single molecule when two monomer units of the maxizyme were connected via a linker sequence. We found that an analogous conformational change could not be induced within a single molecule when two maxizyme units were simply connected via a nonregulatable linker sequence. However, an active conformation was achieved by the introduction of an antisense modulator within the linker sequence that adjusted the overall structure to the correct form. Results of studies in cultured cells suggested that the desired conformational change could indeed be induced within the modified singlechained maxizyme and such a construct caused apoptosis only in leukemic cells with the Philadelphia chromosome. Introduction Chronic myelogenous leukemia (CML) is a clonal myeloproliferative disorder of hematopoietic stem cells associated with the Philadelphia chromosome,1 and it accounts for about 15-20% of all leukemias.2 The reciprocal chromosomal translocation t(9; 22) (q34; q11) can be subdivided into two types: K28 translocations and L6 translocations (Figure 1). They result in the formation of the BCR-ABL fusion gene which encodes two types of mRNA: K28 b3a2 (consisting of BCR exon 3 and ABL exon 2) and L6 b2a2 (consisting of the BCR exon 2 and ABL exon 2) (Figure 1).3-8 Both of these mRNAs are translated into a protein of 210 kDa (p210BCR-ABL) which has elevated tyrosine kinase activity and causes the malignant cell phenotype.9 Since BCR-ABL chimeric mRNAs are tumor-specific and pathogenetically important, they are obvious targets for nucleic acids therapeutics as a paradigm. The hammerhead ribozyme is one of the smallest RNA enzymes. It was first recognized as the sequence motif * Author to whom correspondence should be addressed at The University of Tokyo. Telephone: 81(Japan)-3-5841-8828 or 81(Japan)-298-61-3015. Fax: 81(Japan)-298-61-3019 or 81(Japan)-3-5841-8828. E-mail: taira@ chembio.t.u-tokyo.ac.jp. † National Institute of Advanced Intdustrial Science and Technology (AIST). ‡ The University of Tokyo. § The first two authors contributed equally to this work.

responsible for self-cleavage (cis action) in the satellite RNAs of certain viruses.10 The putative consensus sequence required for activity has three duplex stems and a conserved “core” of two nonhelical segments, plus an unpaired nucleotide at the cleavage site. The trans-acting hammerhead ribozyme11,12 consists of an antisense section (stems I and stem III) and a catalytic domain with a flanking stem-loop II section (Figure 2A, left). Such RNA motifs can cleave oligoribonucleotides at specific sites (most effectively at GUC).13-18 Because of its small size and potential utility as an antivirus agent, this ribozyme has been extensively investigated in terms of the mechanism of its action19-28 and possible applications in vivo.29-38 For such applications, it is clearly necessary to direct the ribozyme specifically to the cellular RNA target of interest. As mentioned above, CML has obvious tumor-specific and pathogenetically important mRNAs. Therefore, the gene therapy by using ribozymes that will disrupt chimeric BCRABL RNAs seems to be attractive. However, the design of ribozymes targeting against BCR-ABL RNAs has not been easy. This is because ribozymes surely have to target the junction sequence. In the case of the K28 BCR-ABL chimeric b3a2 RNA sequence, a potential ribozyme-cleavage site is located near the junction (Figure 1). However, there are no sequences potentially cleavable by hammerhead ribozymes

10.1021/bm010054g CCC: $20.00 © 2001 American Chemical Society Published on Web 07/26/2001

Allosterically Controlled Single-Chained Maxizymes

Biomacromolecules, Vol. 2, No. 3, 2001 789

Figure 1. Translocations and possible sites of cleavage by conventional ribozymes. GUC triplets are generally the triplets that are most susceptible to cleavage by a hammerhead ribozyme,13-18 and one such triplet is located 45 nucleotides from the junction (Figure 1, bottom). If this GUC triplet were targeted by a ribozyme, the normal ABL mRNA, which shares part of the sequence of the abnormal BCR-ABL mRNA, would also be cleaved by the ribozyme, with resultant damage to the host cells.39 In designing ribozymes and antisenses that might cleave b2a2 mRNA, we must be careful to avoid cleavage of normal ABL mRNA. Boxes in orange represent BCR exons and boxes in black represent ABL exon 2. Dotted lines connecting BCR and ABL exons indicate alternative splicing pathways.

within two or three nucleotides of the junction in the L6 BCR-ABL chimeric b2a2 RNA. In designing ribozymes that might cleave the b2a2 mRNA, we must be sure to avoid cleavage of the normal ABL mRNA itself (Figure 1). Recently, we discovered a novel motif of a dimeric maxizyme with sensor function.39 The shortened forms of hammerhead ribozymes (conventional minizymes), with low intrinsic activity, form very active dimers with a common stem II (Figure 2).40,41 Because of their dimeric structure, heterodimeric maxizymes are capable of recognizing two independent sequences,42 and one of the catalytic domains of the heterodimer can be converted to a sensor arm that can allosterically regulate the activity of the maxizyme (Figure 3A).39 Previous study demonstrated that the maxizyme did indeed cleave only the abnormal BCR-ABL mRNA.39 In general, heterodimeric maxizymes potentially generate a mixture of inactive (MzR‚MzR) dimers (consisting of two identical forms of maxizyme right), inactive (MzL‚MzL) dimers (consisting of two identical forms of maxizyme left), and the desired active (MzR‚MzL) dimers (consisting of maxizyme right and maxizyme left). In particular, in vitro, some heterodimeric maxizymes showed low activity due to such a mixed population of dimers.43,44 To avoid formation of such inactive homodimers and, simultaneously, to increase the amount of active (MzR‚MzL) dimers, we have connected

two types of maxizyme.45 However, the specificity has been destroyed by the simple connection of the two independent maxizyme units. Since the appropriate folding of catalytic RNA is a prerequisite for effective catalysis, we then tried to regulate the molecular switching of the connected maxizyme by the introduction of an antisense modulator within the linker sequence that adjusted the overall structure to the active conformation. We describe in this paper the catalytic activity of the allosterically controlled, single-chained, connected maxizyme (cMz) by two types of activation (as shown in Figure 3B), in the presence of the junction in BCR-ABL chimeric L6 b2a2 mRNA in cultured cells. Specific depletion of the p210BCR-ABL protein, as a result of cleavage of the mRNA by the controlled single-chained maxizyme, resulted in the enhanced cleavage of inactive procaspase-3 to yield active caspase-3, with resultant apoptosis, in BaF3/p210BCR-ABL cells. Such controlled single-chained maxizymes, whose activity can be controlled allosterically by sensor arms and also by an antisense modulator, should be powerful tools for the disruption of abnormal chimeric targets and might provide the basis for future gene therapy for the treatment of CML, in particular because the expression of cMz can be controlled by a single promoter and, thus, their delivery by viral vectors, for example, should be simplified.46

790

Biomacromolecules, Vol. 2, No. 3, 2001

Kuwabara et al.

Figure 2. Secondary structures of a hammerhead ribozyme and of maxizymes. (A) Secondary structure of the parental hammerhead ribozyme consisting of an antisense section (stems I and III) and a catalytic domain with a flanking stem-loop II section (left in Figure 2A). The secondary structure of a monomeric minizyme, namely, a conventional hammerhead ribozyme with a deleted stem-loop II region, is shown on the right. (B) Secondary structure of a maxizyme that is capable of forming a homodimer (top) and a heterodimer (bottom). Maxizyme left (MzL) and maxizyme right (MzR) form the heterodimeric maxizyme. The homodimer has two identical binding sites, while the heterodimer can generate two different binding sites, one of which is complementary to the sequence of one substrate and one of which is complementary to a second substrate.

Materials and Methods Construction of Plasmids for the Expression of Allosterically Controlled Single-Chained Maxizymes. DNA templates for allosterically controlled single-chained maxizymes (cMz) were synthesized chemically by a DNA/RNA synthesizer [model 394; Perkin-Elmer, Applied Biosystems (ABI), Foster City, CA]. Sequences for the templates were as follows: cMzA, 5′-GAA GGG CTT CTT TCA TCG AAA CCC TGA GGA ATA GAT CTA AAA GAA TGC TGT GAA AGA AAG CGA ATA CAC TCA CTG ATG AGA GTT ATT GAT GGT CAG-3′; and cMzB, 5′-GAA GGG CTT CTT TCA TCG AAA CCC TGA GGA AAA CTC AAA GTC AAA AAA AAC ACT CAC TAA TGA GAG TTA TTG ATG GTC AG-3′. The cMzA was designed to cleave the L6 BCR-ABL mRNA39 (at the posision after 3348), upon binding of its antisense mudulator suquence with the BCR region (3273 to 3281) on L6 BCR-ABL mRNA (Figure 4B). The cMzB was designed to cleave the same

position on the L6 BCR-ABL mRNA, upon binding of its antisense mudulator suquence with the ABL region (3328 to 3338) on L6 BCR-ABL mRNA (Figure 4B). Synthesized oligonucleotides encoding each ribozyme and the pol III termination sequence47 were converted to double-stranded sequences by PCR. After digestion with Csp 45 I and Sal I, each appropriate fragment was cloned downstream of the promoter of the gene for tRNAVal (Figure 5) of pV (in which the chemically synthesized promoter of a human gene for tRNAVal had been ligated between the EcoR I and Sal I sites of the pMX puro vector).39 The sequences of the constructs were confirmed by direct sequencing. Assays of Reporter Activity after Transient Transfection. The vectors shown in Figure 6A were used to transfect HeLa cells in combination with Lipofectin Reagent (GibcoBRL, Rockville, MD). Luciferase activity was measured with a Pica Gene kit (Toyo-inki, Tokyo, Japan) as described elsewhere.48 To normalize the efficiency of transfection by



Figure 3. Schematic representations of allosteric controls of the conformations of the heterodimeric maxizyme and the single-chained maxizymes. (A) Schematic representation of allosteric control by the maxizyme with sensor arms. To achieve high substrate-specificity, maxizymes were designed to retain an active conformation only in the presence of the abnormal BCR-ABL junction (upper panel), while their conformation should remain inactive in the presence of the normal ABL mRNA or in the absence of the BCR-ABL junction (lower panel). The catalytic action of Mg2+ ions is indicated by “Mg scissors”. (B) Schematic representation of allosteric control with two types of activation by a cMz. A conformational change should be induced within a single molecule when maxizymes (green and red) are connected via a linker sequence (blue). The active conformation is achieved by binding to the BCR-ABL junction. Then a part of the linker sequence acts as an antisense modulator within the complex to adjust the overall structure. The alternative sequence of events is also possible.

792


Kuwabara et al.

Figure 4. Secondary structures of an allosterically controlled single-chained maxizymes (cMz) and its target. (A) Generation of single-chained (dimeric) maxizyme motifs. To improve the cleavage specificty, an antisense modulator was introduced within the linker sequence so that the size of the population of correctly folded complexes would be increased by formation of basepairs with part of the target mRNA, with resultant correct base-pairing of the maxizyme with the cleavage site. (B) Secondary structures of two allosterically controlled single-chained maxizymes (cMzA and cMzB) with antisense modulator sequences (in blue).

reference to β-galactosidase activity, cells were cotransfected with the pSV-β-galactosidase control vector (Promega, Madison, WI) and then the chemiluminescent signal due to β-galactosidase activity was quantitated with a luminescent β-galactosidase genetic reporter system (Clontech, Palo Alto, CA) as described previously.48 Stable Transduction of BaF3/p210BCR-ABL Cells with Various Ribozymes/Maxizymes. Construction of stably transduced BaF3/p210BCR-ABL and H9 cells that harbored cMz or ribozyme construct was performed basically as described

previously.39,48 In a 10% WEHI-conditioned RPMI medium that contained interleukin-3, BaF3/p210BCR-ABL cells were transfected separately with plasmids pVasRz81, pV-cMzB, and pV-I-cMzB, all of which included a gene for resistance to puromycin.48 After 24 h, the medium was replaced with RPMI supplemented with 10% fetal calf serum and 3 µg/ mL puromycin. Stably transduced H9 cells that harbored cMz or ribozyme construct were generated using the respective plasmids (described above). The efficiency of transfection was very



Figure 5. Intracellular expressions of ribozymes and cMz. (A) Predicted secondary structures, based on calculations by the MulFold program (Biocomputing Office, Biology Department, Indiana University, IN) of tRNAVal-driven cMzA and cMzB. The sequence of the human gene for tRNAVal, including the binding sites of transcription factor TFIIIC (labeled A and B, corresponding to the A box and B box regions, respectively), is shown in uppercase letters, with numbering from 1 to 66. Extra sequences that were inserted artificially are indicated by lowercase letters. The sequences of MzR, MzL and the antisense modulator are indicated in green, red, and blue, respectively. (B) Secondary strustures of conventional ribozymes, wtRz and asRz81, and their cleavage sites. (C) Intracellular localization of tRNAVal-driven cMz and tRNAVal-driven conventional ribozymes (wtRz and asRz81) in BaF3/p210BCR-ABL cells that had been stably transduced with the respective maxizyme-coding or ribozyme-coding plasmids. Total RNA that isolated 3 days after removal of IL-3 from the medium and subjected to Northern blotting analysis. U6 snRNA, which remains in the nucleus, was included in the study as a control. The levels of transcripts from the expression vector pV-MzL/R, which produced two independent units of maxizymes MzR and MzL under the control of two independent promoters39 and that from the cMz expression vector under the control of a single promoter (pV-cMzB) were also compared (right). In the former case (pV-MzL/R), two independent units under the control of two independent promoters had to be delivered to each cell. In the latter case (pV-cMzB), only one unit had to be delivered to each cell. The levels of expression of MzL and MzR encoded by the pV-MzL/R vector were lower than those of MzL and MzR encoded by the pV-cMzB vector. Key: N, nuclear fraction; C, cytoplasmic fraction.

794


Kuwabara et al.

1077 (Sigma, St. Louis, MO). Apoptosis was determined as described previously,50 and cells were stained for 15 min with 10 µg/mL Hoechst 33342 (Nippon Gene Co., Toyama, Japan) for analysis of nuclear morphology. After washing and mounting in 90% glycerol/20 mM Tris (pH 8.0)/0.1% N-propyl gallate, samples were examined under a fluorescence microscope (Nikon, Tokyo). Western Blotting Analysis. Cell lysates were subjected to SDS-PAGE on a 15% PAGE. A rabbit polyclonal antibody, RCPP32, that recognized both procaspase-3 and processed p17 (caspase-3) was used to detect the activation of procaspase-3 in apoptotic BaF3/p210BCR-ABL and H9 cells. Blocking and detection were performed as described previously.51 Results

Figure 6. Intracellular activities and specificities of two forms of tRNAVal-driven cMz in HeLa cells. (A) Assay for measurement of the activities of tRNAVal-driven cMz and ribozymes in HeLa cells. (B) Intracellular activities and specificities of tRNAVal-enzymes with the chimeric BCR-ABL-luciferase gene or the ABL-luciferase gene as the reporter. Luciferase activity was normalized, in terms of the efficiency of transfection, by reference to the activity due to a cotransfected gene for β-galactosidase. The obtained results were avarage values at least in triplicate assays.

low, so we generated transduced cells using a line of retroviral producer cells (BOSC23 cells).48 Filtered supernatants of BOSC23 cells, which had been transfected with respective plasmid pV, pVwtRz, or pV-MzL/R, were added to H9 cells. The H9 cells were cultured for 72 h, and then puromycin was added for selection of resistant cells.48 The transduced cells were cultured for a further 60 h, and then dead cells were removed by the Ficol separation procedure and each transformant was obtained by selecting puromycinresistant cells. Northern Blotting Analysis. For assays of the expression of maxizymes in BaF3/p210BCR-ABL cells, total RNA was isolated with ISOGEN (Nippon Gene Co., Toyama). Cytoplasmic RNA and nuclear RNA were separated as described previously.39,42,49 Thirty micrograms of total RNA per lane were loaded on an agarose gel (FMC Inc., Rockland, ME), and then bands of RNA were transferred to a Hybond-N nylon membrane (Amersham Co., Buckinghamshire, U.K.). The membrane was probed with synthetic oligonucleotides, which were specific for each transcript and had been labeled with 32P-ATP by T4 polynucleotide kinase (Takara Shuzo Co., Kyoto, Japan). Prehybridization and hybridization were performed as described previously.49 Cell Viability and Apoptosis. Cell viability was determined by the trypan blue exclusion test. Dead cells were removed by the Ficol separation procedure with Histopaque-

Design of Allosterically Controlled Single-Chained Maxizymes (cMzA and cMzB) for the Specific Cleavage of BCR-ABL Chimeric L6 b2a2 mRNA. To test whether molecular switching might be possible within a single molecule, when two monomer units of the maxizyme were connected via a linker sequence, we designed and synthesized several single-chained maxizymes. When we connected two monomer maxizymes (MzL, shown in red in Figure 4A, and MzR, shown in green in Figure 4A), by a simple linker sequence at the sensor arms between the 5′-end of MzR and the 3′-end of MzL, the specificity of the single-chained, connected maxizyme decreased. Therefore, we connected each monomer maxizyme as shown in Figure 4A, in which MzR and MzL were connected at the cleavage site recognition arms, which correspond to stems I and III of the parental hammerhead ribozyme (Figure 2A, left), by a linker (in blue) that connected the 5′-end of MzL and the 3′-end of MzR. Moreover, since simply connected maxizymes lost specificity in vitro,45 we attempted to increase the specificity with respect to the cleavage of BCR-ABL L6 b2a2 mRNA by introducing an antisense modulator within the linker sequence. We hoped that the size of the population of correctly folded complexes would increase as a result of the formation of basepairs with part of the target mRNA, as shown schematically in the lower part in Figure 3B. Analysis of the activities of the two types of allosterically controlled single-chained maxizyme, cMzA and cMzB (Figure 4B), revealed that only cMzB catalyzed specific cleavage of the target mRNA in vitro.45 In designing the second generation of cMz, cMzA, and cMzB, we took care to ensure that the linker region would not form intramolecular basepairs within cMz. It was impossible to predict, at least by computer-folding analysis, which cMz would have greater specificity. Although we predicted that the antisense modulator region would not form intramolecular basepairs within cMz, we could not exclude the possibility that, in the case of cMzA, the linker region might form basepairs within the incorrectly folded complex and, thus, the effect of the antisense modulator would remain minimal. Allosterically Controlled Single-Chained Maxizymes under the Control of a Human tRNAVal Promoter. For application of allosterically controlled single-chained max-


izymes to gene therapy, the effective delivery into cells and the high-level expression of cMzs are obviously critically important. We use the promoter of a human gene for tRNAVal that is recognized by RNA polymerase III. The tRNAValexpression system results in high-level activity of ribozymes and maxizymes in cells.39,41,42,46,48 Therefore, we embedded cMzA and cMzB similarly in the 3′ portion of the human gene for tRNAVal (Figure 5A). A conventional antisensetype ribozyme (pVasRz81; Figure 5B) which had a parental hammerhead motif with a significantly longer antisense arm than cMz (Figure 5B) was also expressed under the control of the tRNAVal promoter. An expression vector for the wildtype hammerhead ribozyme, wt Rz, that was designed to cleave the L6 BCR-ABL mRNA at the same position as does cMz was also prepared. We used wt Rz as a negative control, because it cleaves the ABL mRNA as well as the BCR-ABL mRNA (Figure 5B). One factor that probably ensures the high-level activity of our tRNAVal-driven maxizymes is that the transcripts can be exported from the nucleus to the cytoplasm, thereby ensuring the co-localization of the tRNAVal-driven maxizyme with its target mRNA. Moreover, a significant portion of certain RNA-binding proteins, including hnRNP, which associate with (pre-)mRNA in the nucleus, dissociate from cytoplasmic mRNAs and, therefore, cytoplasmic mRNAs become more accessible to cytoplasmic tRNAVal-maxizymes. As shown by the results of the Northern blotting analysis in Figure 5C, all tRNAVal-driven transcripts were indeed exported to the cytoplasm with extremely high efficiency. We confirmed that Exportin-t (Xpo-t)53-56 is responsible for the export of tRNAVal-maxizymes to the cytoplasm [Kuwabara, T. et al. unpublished data; see below for the function of Xpo-t.]. In the case of cMzB, a single-chained maxizyme RNA of more than 100 nucleotides in length was attached to the 3′-modified end of human tRNAVal and the resultant tRNAVal-cMzB was efficiently exported to the cytoplasm (Figure 5C). It is also noteworthy that the level of tRNAValdriven transcripts of the cMz construct was significantly higher than those of independent tRNAVal-driven MzR and tRNAVal-driven MzL (Figure 5C, right). The steady-state level of a ribozyme is one of the important factors that govern the intracellular activity of a ribozyme.48 Efficacies of the Allosterically Controlled SingleChained Maxizymes in HeLa Cells. The intracellular activities and specificities of cMzs were examined in mammalian cells using a reporter gene in a transient expression system (Figure 6). We cotransfected HeLa S3 cells with expression plasmids that encoded cMzs under the control of the human tRNAVal promoter, as well as a target gene-expressing plasmid that encoded a chimeric target BCRABL sequence (or ABL alone) and a gene for luciferase (Figure 6A). The normal ABL expression vector encoded a sequence of 300 nucleotides (nt) that encompassed the target cleavage site and the junction between exon 1 and exon 2 of the normal ABL mRNA. The abnormal BCR-ABL mRNA expression vector encoded a sequence of 300 nt that encompassed the BCR-ABL junction and the same target cleavage site in the BCR-ABL mRNA.39 After transient expression of both genes, we measured the intracellular


activity of each maxizyme by estimating the luciferase activity in individual cell lysates. In the sections that follow, cMzA, cMzB, and all the other ribozymes/maxizymes examined in cultured cells refer to human tRNAVal-driven transcripts. The luciferase activity recorded when we used the target gene-expressing plasmid alone was taken as 100% (Figure 6B). Expression of the tRNAVal region (pV) by itself had no inhibitory effect. In accord with results obtained in vitro,45 cMzA did not exhibit specificity because it suppressed the expression not only of the BCR-ABL-luciferase gene but also of the ABL-luciferase gene, whereas cMzB (indicated by a star in Figure 6B) appeared to exhibit higher specificity since it suppressed expression of the abnormal BCR-ABL-luciferase gene more significantly than that of the normal ABLluciferase gene. A conventional antisense-type ribozyme (pVasRz81; Figure 5B), that we used as a control exhibited no specificity, as observed previously.39 It should be emphasized that since a mutant, crippled maxizyme (pV-IcMzB), which differed from the parental cMzB by a single G5-to-A5 mutation within the catalytic core, had no inhibitory effect, the inhibitory effect of the parental cMzB must have originated from the chemical cleavage effect and not from an antisense effect. Specific Inhibitory Effects of cMzs with an Endogenous BCR-ABL Cellular Target. Since the assay of luciferase activity in the transient expression system had indicted that cMzB rather than cMzA might be potentially useful in a future clinical setting, we examined the activity and specificity of cMzB against an endogenous BCR-ABL (L6 b2a2 mRNA) target. In previous experiments in our laboratory, we established the murine cell line BaF3/p210BCR-ABL, which expresses human L6 b2a2 mRNA constitutively, by integrating a plasmid construct that expressed p210BCR-ABL.39,52 The transformed BaF3/p210BCR-ABL cells can grow in the absence of interleukin 3 (IL-3) because of the tyrosine kinase activity of p210BCR-ABL (the product of BCR-ABL gene). The parental cell line BaF3 is, by contrast, an IL-3-dependent line of hematopoietic cells. We demonstrated that if the expression of p210BCR-ABL is inhibited, BaF3/p210BCR-ABL cells become IL-3-dependent and, in the absence of IL-3, they undergo apoptosis.39 Therefore, when we generated stable transformants of BaF3/p210BCR-ABL cells transduced with the tRNAVal-cMz-encoding plasmid or similar plasmids, we used 10% WEHI-conditioned RPMI medium as a source of IL-3. We then switched the medium to medium without IL-3 to monitor the occurrence of apoptosis. As indicated in Figure 7A, BaF3/p210BCR-ABL cells that expressed cMzB (red squares) died most rapidly among the transformants of BaF3/ p210BCR-ABL cells that had been generated by transduction with various plasmids, as listed in the right panel of Figure 7A. To estimate specificity, we also generated stable transformants of H9 cells, using the same sets of plasmids, to serve as controls since H9 cells are derived from human T cells and express normal ABL mRNA rather than abnormal BCRABL mRNA. Stably transduced H9 cells that harbored a maxizyme or ribozyme construct were generated as described in ref 39. As shown in the left panel of Figure 7A, H9 cells

796


Kuwabara et al.

Figure 7. Efficiency and specificity of the action of cMz against endogenous BCR-ABL mRNA. (A) Measurements of viability of tRNAVal-cMzBtransduced BaF3/p210BCR-ABL cells and H9 cells. (B) Morphology of tRNAVal-cMzB-transduced BaF3/p210BCR-ABL cells and H9 cells. (C) Cleavage of inactive procaspase-3 yielded active caspase-3 upon specific depletion of p210BCR-ABL protein by cMzB. Immunoblotting analysis was performed using the antibody RCPP32, which recognizes the 32-kDa precursor to caspase-3 (procaspase-3) and caspase-3 itself.

did not undergo apoptosis upon the constitutive expression of cMzB (red squares), demonstrating clearly that cMzB did not cleave normal ABL mRNA nonspecifically and that constitutive high-level expression of cMzB did not damage host cells. Supporting the results of the transient expression assay, neither H9 cells (purple squares in Figure 7A, left) nor BaF3/ p210BCR-ABL cells (purple squares in Figure 7A, right) underwent apoptosis in the presence of the inactivated cMz (I-cMzB). This result provides further confirmation that the observed effects originated from specific cleavage of the chimeric BCR-ABL gene. Furthermore, the conventional antisense-type ribozyme (asRz81) and the hammerhead ribozyme (wtRz) induced apoptosis not only in BaF3/ p210BCR-ABL cells but also in H9 cells since both catalyzed the nonspecific cleavage of normal ABL mRNA. Figure 7B shows the typical apoptotic morphology observed when cells ceased to proliferate. These cells were stained with the DNA-binding fluorochrome Hoechst 33342. It is clear from Figure 7B that only cMzB caused apoptotic cell death in BaF3/p210BCR-ABL cells specifically and that it had no similar effects on H9 cells. The Activation of Caspase-3 by the Allosterically Controlled Single-Chained Maxizymes. Transduction of a signal for apoptosis and subsequent apoptosis require the coordinated actions of several aspartate-specific cysteine proteases, known as caspases. In the BCR-ABL-mediated inhibition of apoptosis, the apoptotic pathway is interrupted upstream of activation of procaspase-3.50 Then, we tried to confirm that the cMzB-mediated apoptotic pathway did indeed involve the cleavage of inactive procaspase-3 to yield active caspase-3, with resultant apoptosis in BaF3/p210BCR-ABL

cells. Immunoblotting analysis using the antibody RCPP32, which recognizes both the 32-kDa inactive precursor of caspase-3 (procaspase-3) and the processed, active protease, caspase-3, allowed us to monitor the maturation of this cysteine protease (Figure 7C). We confirmed that the basal level of procaspase-3 was almost the same in both BaF3/p210BCR-ABL cells and control H9 cells (lane 1; Figure 7C). In cMzB-expressing BaF3/ p210BCR-ABL cells, the level of procaspase-3 decreased and the level of the p17 active subunit of caspase-3 increased significantly (Figure 7C, right panel, lane 3). By contrast, in stably transduced cMzB-expressing H9 cells, the level of procaspase-3 remained unchanged (Figure 7C, left panel, lane 3). These data again confirmed the highly specific action of cMz. By contrast, expression of the hammerhead ribozyme (lanes 4; wtRz) and the conventional antisense-type ribozyme (lanes 5; asRz81) was associated with the processing of procaspase-3 in both BaF3/p210BCR-ABL and H9 cells, demonstrating again the nonspecific actions of these RNAs. The extent of conversion of procaspase-3 to caspase-3 in stably transduced cMzB-expressing BaF3/p210BCR-ABL cells was slightly higher than that in asRz81- and wtRz-expressing BaF3/p210BCR-ABL cells. In our system, the apoptosis of cells originates from the depletion of p210BCR-ABL and/or p145 c-ABL proteins in the respective hematopoietic cells. (p145 c-ABL is a nuclear protein with low intrinsic tyrosine kinase activity, whereas p210BCR-ABL is a cytoplasmic, membraneassociated protein with a constitutively high level of tyrosine kinase activity that prolongs the survival of hematopoietic cells by inhibiting apoptosis.) These data strengthen our conclusion that the cMz induced apoptosis as a result of


specific depletion of p210BCR-ABL and the subsequent promotion of the activation of caspase-3 in leukemic cells. Taken together, all our data support the proposed change in conformation of cMzB (Figure 3B) in response to an encounter with the BCR-ABL junction and the base-pairing interaction with the antisense modulator in the linker sequence (the order of interactions might be reversed; namely, the base-pairing interaction with the antisense modulator might occur first, being followed by binding to the BCR-ABL junction). Most importantly, these two types of interaction guarantee that cMz undergoes a conformational change and that the specific cleavage of abnormal BCR-ABL mRNA occurs within the correctly folded complex, as shown schematically in Figure 3B. Discussion Expression and Export of tRNAVal-Driven Transcripts to the Cytoplasm. Ribozymes are potentially useful tools for suppression of the expression of specific genes since they can be engineered to act on other RNA molecules with high specificity.34,36,39,42,46 Although some trials have been successful,29,35,57,58 it remains difficult to design an effective ribozyme-expression system that can be used in vivo. One major challenge related to the use of ribozymes and antisense RNAs as therapeutic or genetic agents is the development of suitable expression vectors.48,59-62 Two kinds of expression system have been used to date, namely, the pol II system and the pol III system. In this and other recent studies, we used the pol III system and the promoter of a human gene for tRNAVal for expression of ribozymes/maxizymes.36,39,42,46,48 Not only is this promoter suitable for transcription of small RNAs, but also its use facilitates predictions of secondary structure by computer-folding analysis. More importantly, if properly designed, transcribed tRNAVal-ribozymes can be exported from the nucleus to the cytoplasm where they can find their mRNA targets [Kuwabara, T. et al. unpublished data].39,42,48 A recently identified protein, designated Exportin-t (Xpo-t), transports individual tRNAs from the nucleus to the cytoplasm.53-56 Xpo-t binds RanGTP in the absence of tRNAs, but it does not bind a tRNA in the absence of RanGTP. Therefore, a model for the transport of tRNAs was proposed wherein Xpo-t associates with RanGTP in the nucleus and then this complex binds a mature tRNA molecule. This entire complex is then translocated through a nuclear pore to the cytoplasm where the Ran-bound GTP is hydrolyzed, with release of the tRNA into the cytoplasm and recycling of Xpo-t to the nucleus.55 The minimal sequence or structure within a tRNA that can be recognized by Xpo-t was clarified.54 According to recent studies with Xenopus oocytes,53-56 the 3′ terminal CCA-region appears to be very important for recognition by Xpo-t. Studies in Xenopus oocytes led to the proposal that only correctly processed mature tRNAs are exported from nuclei in a RanGTP-dependent manner.53-56 However, the tRNAValdriven cMzB, shown in Figure 5A, was successfully exported to the cytoplasm (Figure 5C) despite the fact that the sequence at the 3′ end had been replaced by unnatural


sequences that included the sequence of the cMz and was more than 100 nucleotides in length (this is the longest RNA that we have ever attached to the 3′ end of tRNAVal). We determined that Xpo-t/RanGTP bound to immature tRNAs (several tRNAVal-driven transcripts such as the tRNAValdriven cMzB) in vitro and in somatic cells with recognition basically resembling the recognition of mature tRNAs [Kuwabara, T. et al. unpublished data]. The tRNAVal-vector may be useful for expression of functional RNAs other than ribozymes whose target molecules are localized in the cytoplasm. Although co-localization of ribozymes and their targets in the cytoplasm does not, by itself, guarantee effectiveness, it clearly increases the probability of success. In our hands, tRNAVal-driven ribozymes/maxizymes exhibit consistently high activities, at least in cultured cells36,39,42,46,48 and the activities of cytoplasmic ribozymes against specific target mRNAs are significantly greater than those of nuclear ribozymes.63-65 Specific Inhibition of Expression of the BCR-ABL Gene by a tRNAVal-Driven Allosterically Controlled SingleChained Maxizyme. For treatment of CML with ribozymes and, in particular, in the case of the L6 translocations on which we focused in this study, a conventional hammerhead ribozyme is unlikely to work effectively because appropriate cleavage sites are located far from the BCR-ABL junction and, thus, discrimination by the ribozyme between a correct target and an incorrect one remains difficult. Previous attempts at the cleavage of L6 BCR-ABL b2a2 mRNA involved a combination of a long antisense arm and a ribozyme sequence.66,67 Antisense sequences of about 1030 nucleotides, with the potential ability to bind to and cover the junction region for some distance beyond the cleavage site, were connected to one of the substrate-binding sites of hammerhead ribozymes. We demonstrated previously that an antisense-type ribozyme cleaved normal ABL mRNA nonspecifically in vitro.51 In the present study, the antisensetype ribozyme (asRz81) was used as a control with the endogenous BCR-ABL mRNA as target (Figure 7), and the results clearly demonstrated that nonspecific cleavage occurred within the normal ABL mRNA, most probably because hammerhead ribozymes are catalytically active even when the binding arms are as little as three nucleotides long.68,69 This result highlights once again the difficulties associated with the use of conventional hammerhead ribozymes, even after some rational modifications or engineering. We succeeded, in an earlier study, in designing an allosterically controllable maxizyme that forms a catalytically active conformation only in the presence of the BCR-ABL junction (Figure 3A). As a result, extremely specific and high-level activity was achieved in cells with, apparently, considerable potential utility in a medical setting.39,46 For application to clinical therapy, it would be useful if the same conformational change could be induced under the control of a single promoter, namely, if the dimeric maxizymes (Figure 3A) could be transcribed as a single-chained molecule (Figure 3B). In the latter case, only one unit would need to be delivered to cells. By contrast, in the former case, two independent units under the control of two independent promoters must be delivered to each cell. Another advantage

798


to the use of a single-chained maxizyme might be elimination of inactive homodimers, such as (MzR‚MzR) dimers (consisting of two identical forms of maxizyme right) and (MzL‚ MzL) dimers (consisting of two identical forms of maxizyme left).39,43,44 We designed several cMzs, and we explored this possibility in detail in vivo. As mentioned above, since simply connected maxizymes exhibited only limited specificity, we introduced an antisense modulator for additional allosteric control, as shown in Figures 3B and 4B. The cMzB had significantly elevated specificity in the transient expression assay with a reporter gene (Figure 6). The cMzB turned out to be an excellent suppressor of the expression of the endogenous BCR-ABL target (Figure 7), having extremely high specificity. All our data support the proposed conformational change, in living cells, of the allosterically controlled single-chained maxizyme, as depicted in Figure 3B. A major important finding in our study was that the introduction of an antisense modulator allowed further adjustment of the conformational change in the intracellular environment. It should be possible to optimize such antisense modulators to increase the activity of allosterically controlled single-chained maxizyme still further. The allosteric control depicted in Figure 3A can be used to suppress not only the L6 BCR-ABL b2a2 gene39 but also other genes,70-74 and thus, the novel motif depicted in Figure 3B should also be useful in many other systems. Last, creation of artificial allosteric enzymes is of great current interest. Discovery of dimeric short ribozymes with high activity40 and the fact that the activity of hammerhead ribozymes solely depends on the catalytic metal ions24-28,75 led us to design maxizymes.39,70-74 As demonstrated in this paper, the maxizyme can now be expressed as a singlechained molecule and the anticipated conformational change (Figure 3B) occurred in vivo in response to the presence of the abnormal splice junction. Clearly, the maxizyme, whose design is based on physical organic or biophysical chemistry, has evolved to the stage where it might significantly contribute to medical area. References and Notes (1) Nowell, P. C.; Hungerford, D. A. Science 1960, 132, 1497-1499. (2) Lichtman, M. A. Chronic Myelogenous Leukemia and Related Disorders in Hematology; McGraw-Hill: New York, 1990. (3) Rowley, J. D. Nature 1973, 243, 290-293. (4) Bartram, C. R.; de Klein, A.; Hagemeijer, A.; van Agthoven, T.; van Kessel, A. G.; Bootsma, D.; Grosveld, G.; Ferguson-Smith, M. A.; Davies, T.; Stone, M.; Heisterkamp, N.; Stephenson, J. R.; Groffen, J. Nature 1983, 306, 277-280. (5) Heisterkamp, N.; Stephenson, J. R.; Groffen, J.; Hansen, P. F.; de Klein, A.; Bartram, C. R.; Grosveld, G. Nature 1983, 306, 239242. (6) Groffen, J.; Stephenson, J. R.; Heisterkamp, N.; de Klein, A.; Bartram, C. R.; Grosveld, G. Cell 1984, 36, 93-99. (7) Shtivelman, E.; Lifshitz, B.; Gale, R. P.; Canaani, E. Nature 1985, 315, 550-553. (8) Shtivelman, E.; Lifschitz, B.; Gale, R. P.; Roe, B. A.; Canaani, J. Cell 1986, 47, 277-284. (9) Konopka, J. B.; Watanabe, S. M.; Witte, O. N. Cell 1984, 37, 10351042. (10) Symons, R. H. Trends Biochem. Sci. 1989, 14, 445-450. (11) Uhlenbeck, O. C. Nature 1987, 328, 596-600. (12) Haseloff, J.; Gerlach, W. L. Nature 1988, 334, 585-591. (13) Koizumi, M.; Iwai, S.; Ohtsuka, E. FEBS Lett. 1988, 228, 285288.

Kuwabara et al. (14) Sheldon, C. C.; Symons, R. H. Nucleic Acids Res. 1989, 17, 567985. (15) Ruffner, D. E.; Stormo, R. H.; Uhlenbeck, O. C. Biochemistry 1990, 29, 10695-10702. (16) Perriman, R.; Delves, A.; Gerlach, W. L. Gene 1992, 113, 157163. (17) Shimayama, T.; Nishikawa, S.; Taira, K. Biochemistry 1995, 34, 3649-3654. (18) Zoumadakis, M.; Tabler, M. Nucleic Acids Res. 1995, 23, 11921196. (19) Sawata, S.; Komiyama, M.; Taira, K. J. Am. Chem. Soc. 1995, 117, 2357-2358. (20) Zhou, D.-M.; Usman, N.; Wincott, F. E.; Matulic-Adamic, J.; Orita, M.; Zhang. L.-H.; Komiyama, M.; Kumar, P. K. R.; Taira, K. J. Am. Chem. Soc. 1996, 118, 5862-5866. (21) Zhou, D.-M.; Kumar, P. K. R.; Zhang. L.-H.; Taira, K. J. Am. Chem. Soc. 1996, 118, 8969-8970. (22) Warashina, M.; Takagi, Y.; Sawata, S.; Zhou, D.-M.; Kuwabara, T.; Taira, K. J. Org. Chem. 1997, 62, 9138-9147. (23) Kuimelis, R. G.; McLaughlin, L. W. Chem. ReV. 1998, 98, 10271044. (24) Zhou, D.-M.; Taira, K. Chem. ReV. 1998, 98, 991-1026. (25) Lilley, D. M. J. Curr. Opin. Struct. Biol. 1999, 9, 330-338. (26) Wang, S.; Karbstein, K.; Peracchi, A.; Beigelman, L.; Herschlag, D. Biochemistry 1999, 38, 14363-14378. (27) Warashina, M,; Takagi, Y.; Stec, W. J.; Taira, K. Curr. Opin. Biotechnol. 2000, 11, 354-362. (28) Takagi, Y.; Warashina, M,; Stec, W. J.; Yoshinara, K.; Taira, K. Nucleic Acids Res. 2001, 29, 1815-1834. (29) Erickson, R. P.; Izant, J. G. Gene Regulation: Biology of Antisense RNA and DNA; Raven Press: New York, 1992. (30) Ojwang, J. O.; Hampel, A.; Looney, D. J.; Wong-Staal, F.; Rappaport, J. Proc. Natl. Acad. Sci. U.S.A. 1992, 89, 10802-10806. (31) Rossi, J. J. Curr. Opin. Biotechnol. 1992, 3, 3-7. (32) Eckstein, F., Lilley, D. M. J., Eds. Catalytic RNA: Nucleic Acids and Molecular Biology; Springer-Verlag: Berlin, 1996; Vol. 10. (33) Leavitt, M. C.; Yu, M.; Yamada, O.; Kraus, G.; Looney, D.; Poeschla, E.; Wong-Staal, F. Hum. Gene Ther. 1994, 5, 1115-1120. (34) Kawasaki, H.; Ohkawa, J.; Tanishige, N.; Yoshinari, K.; Murata, T.; Yokoyama, K. K.; Taira, K. Nucleic Acids Res. 1996, 24, 30103016. (35) Turner, P. C., Ed. Ribozyme Protocols: Methods in Molecular Biology; Humana Press: Totowa, NJ, 1997; Vol. 74. (36) Kawasaki, H.; Ecker, R.; Yao, T.-P.; Taira, K.; Chiu, R.; Livingston, D. M.; Yokoyama, K. K. Nature 1998, 93, 284-289. (37) Scanlon, K. J., Ed. Therapeutic Applications of Ribozymes: Methods in Molecular Medicine; Humana Press: Totowa, NJ, 1998; Vol. 11. (38) Krupp, G.; Gaur, R. K., Eds. Ribozyme: Biochemistry and Biotechnology; Eaton Publishing: Natick, MA, 2000. (39) Kuwabara, T.; Warashina, M.; Tanabe, T.; Tani, K.; Asano, T.; Taira, K. Mol. Cell 1998, 2, 617-627. (40) Amontov, S.; Taira, K. J. Am. Chem. Soc. 1996, 118, 1624-1628. (41) Kuwabara, T.; Warashina, M.; Orita, M.; Koseki, S.; Ohkawa, J.; Taira, K. Nature Biotechnol. 1998, 16, 961-965. (42) Kuwabara, T.; Warashina, M.; Nakayama, A.; Ohkawa, J.; Taira, K. Proc. Natl. Acad. Sci. U.S.A. 1999, 96, 1886-1891. (43) Nakayama, A.; Kuwabara, T.; Warashina, M.; Taira, K. FEBS Lett. 1999, 448, 67-74. (44) Nakayama, A.; Warashina, M.; Kuwabara, T.; Taira, K. J. Inorg. Biochem. 2000, 78, 69-77. (45) Hamada, M.; Kuwabara, T.; Warasina, M.; Nakayama, A.; Taira, K. FEBS Lett. 1999, 461, 77-85 (46) Tanabe, T.; Kuwabara, T.; Warashina, M.; Tani, K.; Taira, K.; Asano, S. Nature 2000, 406, 473-474. (47) Geiduschek, E. P.; Tocchini-Valentini, G. P. Annu. ReV. Biochem. 1988, 57, 873-914. (48) Koseki, S.; Takebe, T.; Tani, K.; Asano, S.; Shioda, T.; Nagai, Y.; Shimada, T.; Ohkawa, J.; Taira, K. J. Virol. 1999, 73, 1868-1877. (49) Baier, G.; Coggeshall, K. M.; Baier-Bitterlich, G.; Giampa, L.; Telford, D.; Herbert, E.; Shih, W.; Altman, A. Mol. Immunol. 1994, 31, 923-932. (50) Dubrez, L.; Eymin, B.; Sordet, O.; Droin, N.; Turhan, A. G.; Solary, E. Blood 1998, 7, 2415-2422. (51) Kuwabara, T.; Warashina, M.; Tanabe, T.; Tani, K.; Asano, S.; Taira, K. Nucleic Acids Res. 1997, 25, 3074-3082. (52) Daley, G. Q.; Baltimore, D. Proc. Natl. Acad. Sci. U.S.A. 1988, 85, 9312-9316. (53) Arts, G.-J.; Fornerod, M.; Mattaj, I. W. Curr. Biol. 1998, 6, 305314.

Allosterically Controlled Single-Chained Maxizymes (54) Arts, G.-J.; Scott, K.; Pascale, R.; Bernard, E.; Mattaj, I. W. EMBO J. 1999, 17, 7430-7441. (55) Kutay, U.; Liposwky, G.; Izaurralde, E.; Bischoff, F. R.; Schwarzmaier, P.; Hartmann, E.; Go¨rlich. D. Mol. Cell. 1998, 1, 359-369. (56) Lund, E.; Dahlberg. J. E. Science 1999, 282, 2082-2085. (57) Fujita, S.; Koguma, K.; Ohkawa, J.; Mori, K.; Kohda, T.; Kise, H.; Nishikawa, S.; Iwakura, M.; Taira, K. Proc. Natl. Acad. Sci. U.S.A. 1997, 94, 391-396. (58) Rossi, J. J., Couture, L. A., Eds. Intracellular Ribozyme Applications: Principles and Protocols; Horizon Scientific Press: Norfolk, England, 1999. (59) Jennings, P. A.; Moloy, P. L. EMBO J. 1987, 6, 3043-3047. (60) Bertrand, E.; Pictet, R.; Grange, T. Nucleic Acids Res. 1994, 22, 293300. (61) Thompson, D. J.; Ayers, F. D.; Malmstrom, A. T.; Ganousis, L. M.; Chowrira, M. B.; Couture, L.; Stinchcomb, T. D. Nucleic Acids Res. 1995, 23, 2259-2268. (62) Yu, M.; Ojwang, J. O.; Yamada, O.; Hampel, A.; Rappaport, J.; Looney, D.; Wong-Staal, F. Proc. Natl. Acad. Sci. U.S.A. 1993, 90, 6340-6344. (63) Kato, Y.; Kuwabara, T.; Warashina, M.; Toda, H.; Taira, K. J. Biol. Chem. 2001, 276, 15378-15385. (64) Kuwabara, T.; Warashina, M.; Koseki, S.; Sano, M.; Ohkawa, J.; Nakayama, K.; Taira, K. Nucleic Acids Res. 2001, 29, 2780-2788.

Biomacromolecules, Vol. 2, No. 3, 2001 799 (65) Warashina, M.; Kuwabara, T.; Kato, Y.; Sano, M.; Taira, K. Proc. Natl. Acad. Sci. U.S.A. 2001, 98, 5572-5577. (66) Pachuk, C. J.; Yoon, K.; Moelling, K.; Coney, L. R. Nucleic Acids Res. 1994, 22, 301-307. (67) James, H.; Mills, K.; Gibson, I. Leukemia 1996, 10, 1054-1064. (68) Hertel, K. J.; Herschlag, D.; Uhlenbeck, O. C. EMBO J. 1996, 15, 3751-3757. (69) Birikh, K. R.; Heaton, P. A.; Eckstein, F. Eur. J. Biochem. 1997, 245, 1-16. (70) Tanabe, T.; Takata, I.; Kuwabara, T.; Warashina, M.; Kawasaki, H.; Tani, K.; Ohta, S.; Asano, S.; Taira, K. Biomacromol. 2000, 1, 108117. (71) Kuwabara, T.; Warashina, M.; Taira, K. Trends Biotechnol. 2000, 18, 462-468. (72) Kuwabara, T.; Warashina, M.; Taira, K. Curr. Opin. Chem. Biol. 2000, 4, 669-677. (73) Kurata, H.; Miyagishi, M.; Kuwabara, T.; Warashina, M.; Taira, K.; J. Biochem. Mol. Biol. 2000, 33, 359-365. (74) Warashina, M.; Kuwabara, T.; Taira, K. Structure 2000, 8, R207R212. (75) Yoshinari, K.; Taira, K. Nucleic Acids Res. 2000, 28, 1730-1742.

BM010054G

Allosterically Controlled Single-Chained Maxizymes with Extremely

Recommend Documents