Relationships between the Activities in Vitro and in Vivo ofVarious Kinds of Ribozyme and Their Intracellular Localizationin Mammalian Cells*
Received for publication, November 22, 2000, and in revised form, January 26, 2001Published, JBC Papers in Press, January 30, 2001, DOI 10.1074/jbc.M010570200
From the ‡Department of Chemistry and Biotechnology, Graduate School of Engineering, University of Tokyo,7-3-1 Hongo, Tokyo 113-8656 and the §Gene Discovery Research Center, National Institute of Advanced IndustrialScience and Technology, 1-1-4 Higashi, Tsukuba Science City 305-8562, Japan
Nineteen different functional RNAs were synthesizedfor an investigation of the actions of ribozymes, in vitroand in vivo, under the control of two different promot-ers, tRNA or U6, which localize transcripts either in thecytoplasm or in the nucleus. No relationships werefound between the activities of these RNAs in culturedcells and the kinetic parameters of their respectivechemical cleavage reactions in vitro, indicating that inno case was chemical cleavage the rate-limiting step invivo. For example, a hepatitis delta virus (HDV) ri-bozyme, whose activity in vitro was almost 3 orders ofmagnitude lower than that of a hammerhead ribozyme,still exhibited similar activity in cells when an appro-priate expression system was used. As expected, exter-nal guide sequences, the actions of which depend onnuclear RNase P, were more active in the nucleus. Anal-ysis of data obtained with cultured cells clearly demon-strated that the cytoplasmic ribozymes were signifi-cantly more active than the nuclear ribozymes,suggesting that mature mRNAs in the cytoplasm mightbe more accessible to antisense molecules than arepre-mRNAs in the nucleus. Our findings should be use-ful for the future design of intracellularly active func-tional molecules.
Since the discovery of the first two ribozymes (1, 2), severalnew types of ribozyme with self-cleavage activity have beenfound in nature (3–8). Small ribozymes that can be designed tocleave RNA strands intermolecularly include hammerhead, hair-pin, and HDV1 ribozymes. These trans-acting ribozymes recog-nize their RNA substrates via formation of Watson-Crick basepairs, and they cleave these RNAs in a sequence-specific manner.Because of their specificity, trans-acting ribozymes show promise
as tools for the dysfunction of target RNAs (9–21).The constitutive expression of a ribozyme in vivo, under the
control of a strong promoter, represents an attractive strategyfor the application of trans-acting ribozymes to gene therapy.As described in our previous reports (22, 23), we have suc-ceeded in establishing an effective ribozyme expression system,with subsequent efficient transport of transcripts to the cyto-plasm, which is based on a promoter that is recognized by RNApolymerase III (pol III). High levels of expression under thecontrol of the pol III promoter are advantageous for the exploi-tation of ribozymes in vivo. Therefore, we chose an expressionsystem with the promoter of a human gene for tRNAVal. Manyribozymes, such as hammerheads and hairpins, have been ef-fectively expressed under the control of promoters of gene fortRNAs (9, 11–15, 20, 21, 24).
A major advantage of our tRNAVal-directed expression sys-tem is that, with appropriate modification of the tRNAVal por-tion, it is possible to colocalize the expressed ribozyme in thecytoplasm with its target mRNA (14, 15, 22, 23, 25). Ribozymesexpressed under the control of the tRNAVal promoter are ex-ported to the cytoplasm as effectively as natural tRNAs via theaction of Xpo(t),2 a tRNA-binding protein (26, 27) that func-tions with Ran GTPase, which regulates the transport bycatalyzing the hydrolysis of GTP. Mature mRNAs are exportedto the cytoplasm for translation. Thus, both ribozymes andtheir target mRNAs can be co-localized in the same cellularcompartment.
By contrast, an external guide sequence (EGS), which isadded in trans and is able to bind to its target RNA, appears tofunction in the nucleus because its effect depends on the activ-ity of ribonuclease P (RNase P) (17, 28–30). The EGS RNAbinds to the target RNA, yielding a structure that resemblesthe pre-tRNA that is recognized as a substrate by RNase P.RNase P normally cleaves precursors to tRNAs to generate the59 termini of mature tRNAs. Because RNase P is expressedconstitutively in cells and accumulates, in particular in thenucleus, the use of an EGS as a gene-inactivating agent doesnot require expression of additional RNase P from exogenouslyintroduced genes. Although an EGS does not have intrinsiccleavage activity, when it acts in cooperation with endogenousRNase P, it can effectively inactivate its target mRNA.
Although there have been many studies both in vitro and invivo of the activities of the ribozymes mentioned above, furtherdetailed information on the parameters that determine theiractivities as gene-inactivating agents in vivo is necessary sothat we will be able to optimize their effects by optimizing the
* This work was supported in part by grants from the Ministry ofEconomy, Trade and Industry of Japan and also by a grant-in-aid forscientific research from the Ministry of Education, Culture, Sports,Science and Techology, Japan. The costs of publication of this articlewere defrayed in part by the payment of page charges. This article musttherefore be hereby marked “advertisement” in accordance with 18U.S.C. Section 1734 solely to indicate this fact.
¶ These authors contributed equally to this work.i Recipient of a research fellowship for young scientists from the
Japan Society for the Promotion of Science.** To whom correspondence should be addressed: Dept. of Chemistry
and Biotechnology, Graduate School of Engineering, University ofTokyo, Hongo, Tokyo 113-8656, Japan. Tel.: 81-3-5841-8828 or 81-298-61-3015; Fax: 81-298-61-3019; E-mail: [email protected].
1 The abbreviations used are: HDV, hepatitis delta virus; pol, polym-erase; EGS, external guide sequence; CML, chronic myelogenous leu-kemia; nt, nucleotide(s); Wt, wild-type; Rz, ribozyme. 2 T. Kuwabara and K. Taira, unpublished data.
requisite parameters. In addition, although it has been claimedfor each individual ribozyme that it has potential utility as aneffective gene-inactivating agent, there has been no systematicanalysis in which the activities of various ribozymes have beencompared under similar conditions in vivo. In this study, wedesigned several types of functional RNA targeted to the junc-tion site of the BCR-ABL chimeric mRNA that causes chronicmyelogenous leukemia (CML). Using this system, we haveaccumulated data that might allow correlations to be madebetween ribozyme activities in cultured cells and the efficaciesof the same ribozymes in vivo, namely, in mice (15, 21). CMLoccurs as a result of reciprocal chromosomal translocations thatresult in the formation of the BCR-ABL fusion gene. One of thechimeric mRNAs transcribed from an abnormal BCR-ABL(B2A2) gene (consisting of exon 2 of BCR and exon 2 of ABL;Refs. 31 and 32) provides a suitable substrate for comparisonsof ribozymes. We used six kinds of functional RNA, includinghammerhead, hairpin, and HDV ribozymes; our maxizyme andin vitro selected minizymes; and EGSs to examine parametersthat determined activities in vitro and in vivo.
Our goal was to determine whether activity in vitro mightreflect activity in mammalian cells. Moreover, since we haveevidence that suggests that tRNAVal-driven ribozymes withhigh level of activities are efficiently exported to the cytoplasm,whereas similarly expressed tRNA ribozymes with poor activ-ities are accumulated in the nucleus (22), we decided to exam-ine the correlation between nuclear localization and/or trans-port of functional RNAs and the activity in vivo. For thispurpose, we used two kinds of promoter. One promoter was thepromoter of the gene for tRNAVal described above, and theother was a U6 promoter (33, 34). Transcripts expressed underthe control of these promoters are located in the cytoplasm andthe nucleus, respectively.
We found that the intrinsic cleavage activity of a ribozyme isnot the sole determinant of activity in cultured cells and that itis the cytoplasmic localization and the association of the ri-bozyme with its substrate that regulate activity.
MATERIALS AND METHODS
Construction of Vectors for Expression of Ribozymes and EGSs—Theconstruction of vectors for expression of ribozymes from the tRNAVal
promoter using pUC-dt (a plasmid that contains the chemically synthe-sized promoter for a human gene for tRNAVal between the EcoRI andSalI sites of pUC 19) was described previously (22, 23). pUC-dt wasdouble-digested with Csp45I and SalI, and a fragment having a linkersequence with 59 Csp45I site and the restriction sites for KpnI andEcoRV and the terminator sequence TTTTT at the 39 end with 39 SalIsite was cloned into the double-digested plasmid to yield pUC-tRNA/KE. The KpnI and EcoRV sites were used for subsequent insertion ofthe each ribozyme sequence. The construction of vectors for ribozymeexpression from the U6 promoter has been described elsewhere (17).The EcoRI and XhoI sites were used for insertion of each ribozymesequence.
Analysis of the Cleavage Activity of Individual Ribozymes in Vitro—Each ribozyme and two substrates, namely BCR-ABL and ABL RNAs,were prepared in vitro using T7 RNA polymerase. Assays of ribozymeactivity in vitro were performed, in 25 mM MgCl2 and 50 mM Tris-HCl(pH 8.0) at 37 °C, under enzyme-saturating (single-turnover) condi-tions, as described elsewhere (14). Each ribozyme (50 mM) was incu-bated with 2 nM 59-32P-labeled substrate. The substrate and the prod-ucts of each reaction were separated by electrophoresis on an 8% polya-crylamide, 7 M urea denaturing gel and detected by autoradiography.
In Situ Hybridization—HeLa S3 cells on a coverslip, which had beentransfected in advance with an appropriate plasmid, were washed infresh phosphate-buffered saline and fixed in fix/permeabilization buffer(50 mM HEPES/KOH, pH 7.5, 50 mM potassium acetate, 8 mM MgCl2, 2mM EGTA, 2% paraformaldehyde, 0.1% Nonidet P-40, 0.02% SDS) for15 min at room temperature. Cells were rinsed three times in phos-phate-buffered saline for 10 min each. Seventy micrograms of Cy3-labeled oligodeoxynucleotide probe with a sequence complementary tothe ribozyme and 20 mg of tRNA from Escherichia coli MRE 600 (RocheMolecular Biochemicals, Mannheim, Germany), dissolved in 10 ml of
deionized formamide, were denatured by heating for 10 min at 70 °C.The mixture was then chilled immediately on ice, and 10 ml of hybrid-ization buffer, containing 20% dextran sulfate and 2% BSA in 43 SSC,were added. Twenty microliters of the hybridization solution containingthe probe were placed on the coverslip, and the coverslip was invertedon a glass slide, sealed with rubber cement, and incubated for 16 h at37 °C. Cells were rinsed in 23 SSC, 50% formamide and in 23 SSC atroom temperature for 20 min each. The coverslip was mounted withVectashield (Vector Laboratories, Burlingame, CA) on a glass slide, andcells were analyzed with a confocal laser scanning microscope (LSM510; Carl Zeiss, Jena, Germany).
Northern Blotting Analysis—Cells were grown to ;80% confluence(1 3 107 cells) and were transfected with a tRNAVal-Rz expressionvector with the Lipofectiny reagent (Life Technologies, Inc.). Thirty-sixhours after transfection, cells were harvested. For the preparation ofthe cytoplasmic fraction, collected cells were washed twice with phos-phate-buffered saline and then resuspended in digitonin lysis buffer (50mM HEPES/KOH, pH 7.5, 50 mM potassium acetate, 8 mM MgCl2, 2 mM
EGTA, and 50 mg/ml digitonin) on ice for 10 min. The lysate wascentrifuged at 1,000 3 g, and the supernatant was collected as thecytoplasmic fraction. The pellet was resuspended in Nonidet P-40 lysisbuffer (20 mM Tris-HCl, pH 7.5, 50 mM KCl, 10 mM NaCl, 1 mM EDTA,and 0.5% Nonidet P-40) and held on ice for 10 min, and the resultantlysate was used as the nuclear fraction. Cytoplasmic RNA and nuclearRNA were extracted and purified from the cytoplasmic fraction and thenuclear fraction, respectively, with ISOGEN reagent (Wako, Osaka,Japan). Thirty micrograms of total RNA per lane were loaded on a 3.0%NuSievey (3:1) agarose gel (FMC Inc., Rockland, ME). After electro-phoresis, bands of RNA were transferred to a Hybond-Ny nylon mem-brane (Amersham Pharmacia Biotech, Buckinghamshire, United King-dom). The membrane was probed with a synthetic oligonucleotide thatwas complementary to the sequence of the relevant ribozyme. Eachprobe was labeled with 32P by T4 polynucleotide kinase (Takara ShuzoCo., Kyoto, Japan).
Measurement of Luciferase Activity—Luciferase activity was meas-ured with a PicaGene® kit (Toyo-inki, Tokyo, Japan) as described else-where (15). In order to normalize the efficiency of transfection byreference to b-galactosidase activity, cells were cotransfected with thepSV-b-galactosidase control vector (Promega, Madison, WI), and thenthe chemiluminescent signal due to b-galactosidase was quantitatedwith a luminescent b-galactosidase genetic reporter system (CLON-TECH, Palo Alto, CA) as described previously (15).
RESULTS
Design of Ribozymes and EGSs—In order to express ri-bozymes in vivo, we used two kinds of pol III promoter (Fig.1A). Transcripts with the promoter of the gene for tRNAVal canbe efficiently transported to the cytoplasm when the appropri-ate choice of and combination of linker and ribozyme sequenceis made (14, 15, 22, 23, 25). We used the mouse U6 promoter,which controls expression of U6 RNA that is localized in thenucleus, for expression and accumulation of transcripts in thenucleus. Ten functional RNAs (Fig. 1B) directed against siteswithin a limited region (,100 nt) of B2A2 and ABL mRNAs(Fig. 1C; target sites are underlined; the identical cleavage sitecould not be chosen because of different cleavable sequence foreach different ribozyme) and expression vectors that encodedeach respective functional RNA were designed such that eachfunctional RNA was produced under the control both of thetRNAVal promoter and of the U6 promoter. The product trans-lated from B2A2 chimeric mRNA causes CML. We demon-strated previously that the RNA maxizyme that functions as adimer cleaves B2A2 chimeric mRNA in vitro and in vivo withextremely high specificity without any damage to normal ABLmRNA (15, 21). Since it seemed possible that a hairpin ri-bozyme might also distinguish abnormal B2A2 mRNA fromnormal ABL mRNA (even though conventional hammerheadribozymes fail to do so), we decided to use these two substratesin this study. Using two different substrates, we hoped to gainmore insight into the differences among the activities of thevarious ribozymes in vivo in more general terms.
The hammerhead ribozyme is one of the smallest trans-acting ribozymes (6, 7, 19, 35–37). The stem II region of this
Comparison of Ribozymes in Vitro and in Vivo 15379
ribozyme can be varied, and many derivatives with variousmodifications have been studied (38–44). The maxizyme is onesuch derivative and acts as a dimer (Fig. 1B), and, in general,maxizymes have high level activity in vivo (14, 15, 21, 23, 25).The term “maxizymes (minimized, active, X-shaped (functionsas a dimer), and intelligent (allosterically controllable) ri-bozymes)” was the name given to the minimized, allostericallycontrollable dimeric ribozymes with high level activity in vivo(15, 21, 25, 45–47). The minizymes shown in Fig. 1B are min-imized hammerhead ribozymes with stem II deletions and rel-atively high activity, and each of them was identified recentlyby in vitro selection (43, 44). These minizymes function in vitroeven at low concentrations of Mg21 ions. Therefore, they mayhave advantage at low concentrations of Mg21 ions in vivo;thus, we included them in our study. The maxizyme is aneffector-inducible trans-activated ribozyme, and the maxizymeshown in Fig. 1B recognizes the junction region of B2A2; itcleaves B2A2 mRNA but not normal ABL mRNA and, there-fore, we used this well characterized tRNAVal-driven maxizymeas a positive control in studies in cultured cells (15, 21, 25,45–47). Hairpin ribozymes, consisting of four helical regionsinterrupted by two internal bulges, have been used successfullyas gene-inactivating agents (9, 11, 48). The two bulges interactwith each other and hairpin ribozymes do not require Mg21
ions for catalysis, an observation that suggests that a base(s) inthis region might be essential for catalysis (49–52). In thisstudy, we designed two such hairpin ribozymes targeted to twodifferent sites. Although studies of the mechanism of action ofHDV ribozymes with a pseudoknot structure indicate that acytosine base distal to the cleavage site acts as a general acidcatalyst (51–53), very little information is available about theactivity of HDV ribozymes in vivo. Both genomic and antig-enomic versions of the HDV ribozyme can be generated, andthe latter type was made to act in trans (18, 54, 55). Weprepared two such trans-acting HDV ribozymes targeted to twodifferent sites.
Ribonuclease P (RNase P) cleaves tRNA precursors (pre-tRNAs) to generate the 59 termini of mature tRNAs (28–30,56–58). Studies of RNase P resulted in the design of EGSRNAs. An EGS is designed to bind to a target RNA to generatea structure that mimics that of a pre-tRNA structure and isrecognized as a substrate for RNase P. Upon formation of thisstructure, the target RNA can be cleaved by RNase P (17, 28,30). RNase P is synthesized constitutively in cells and, thus, foruse of an EGS as a gene-inactivating agent, it is not necessaryto engineer the expression of additional RNase P. An EGS itselfdoes not have cleavage ability. However, in cooperation withendogenous RNase P, it can bring about the cleavage of itstarget mRNA.
We constructed a total of 19 plasmids for expression of eachribozyme with the exception of the maxizyme under control ofthe tRNAVal or the U6 promoter. The sequences of ribozymesand the EGS were inserted as shown in Fig. 1A. All sequencesin plasmids were confirmed by sequencing. The integrity ofeach construct was also confirmed by examination of the cleav-age activity in vitro of each respective transcript, as describedbelow.
Cleavage Activities of Ribozymes in Vitro—We first deter-
mined activities of the various ribozymes in vitro. In order tocompare chemical cleavage activities rather than associationand/or dissociation kinetics, we measured the activities of thefunctional RNAs in vitro in the presence of a saturating excessof each ribozyme (single-turnover conditions). Ribozymes andsubstrate RNAs were transcribed in vitro by T7 RNA polymer-ase. As substrates, we used RNAs of 92 and 121 nt, whichcorresponded to regions that spanned the junctions of ABLRNA and B2A2 RNA, respectively (Fig. 1C). Because the tRNA-Val promoter is an internal promoter, in other words the DNAsequence that corresponds to tRNAVal contains the promoterregion, transcripts from this promoter are always linked to aportion of the tRNA, which might or might not interfere withthe ribozyme’s activity. Therefore, ribozymes tested in vitroincluded a modified tRNA promoter region (about 90 nt) orabout 20 nt of the U6 promoter (Fig. 1A).
As shown in Fig. 2, wild-type hammerhead ribozymes (WtRz) had the highest activity in the case of both the tRNAVal andU6 promoter-driven transcripts and against both the B2A2 andABL substrates. Since the majority of the target sites arelocated in the exon 2 region of ABL mRNA and since thecomputer-predicted secondary structure of this region is almostthe same for both substrates (Fig. 1C, dark blue), with theexception of the region upstream of the junction, no significantdifferences between the rates of cleavage by the ribozymes ofB2A2 and ABL substrates were expected or observed. In gen-eral, rates of cleavage by tRNAVal-driven ribozymes wereslightly higher than those by U6-driven ribozymes, demon-strating that the tRNA portion did not hinder the activity of theribozyme to any great extent. In terms of the rate of chemicalcleavage, no other ribozyme approached the efficiency of thehammerhead ribozyme; the activity of the majority of ri-bozymes against the relatively long substrates (92 and 121 nt)was 2 or more orders of magnitude lower than that of Wt Rz invitro.
Localization of tRNAVal- and U6-driven Transcripts—Colo-calization of a ribozyme with its substrate is an importantdeterminant of the activity of the ribozyme (10, 15, 22, 23). Atranscribed ribozyme might be expected to cleave pre-mRNAsin the nucleus or to be exported to the cytoplasm to cleavemature mRNAs (19). Our earlier data indicate that tRNAVal-driven ribozymes with high level of activities are efficientlyexported to the cytoplasm, while similarly expressed tRNA-ribozymes with low level of activities are accumulated in thenucleus (22). However, there has been no systematic attempt toidentify the cellular compartment in which a ribozyme actsmost effectively. Given that it should be necessary for a ri-bozyme to be transported to the cytoplasm in mammalian cellsfor colocalization with its target mRNA, we developed ourexpression system for cytoplasmic expression of ribozymes,because mature mRNAs are exported to the cytoplasm fortranslation. By contrast, an EGS is likely to be operative in thenucleus because cleavage depends on RNase P, which is activeonly in the nucleus. In order to confirm this hypothesis (seenext section), ribozymes were expressed from both types ofpromoter and their localization was determined both by in situhybridization and by Northern blotting analysis of fractionatedcell lysates.
Fig. 1. Design of the functional RNAs directed against BCR-ABL (B2A2) mRNA and ABL mRNA. A, the two types of expression system.The tRNAVal portion was attached to each ribozyme because the promoters of the gene are internal (A and B boxes). No extra sequence wasattached to the U6-driven transcripts. However, 17 nt were attached to ribozymes transcribed by T7 polymerase in vitro for kinetic analysis usebecause of the design of the primer used. B, secondary structures of ribozymes and EGSs. The Wt Rz, minizyme, and maxizyme had the identicaltarget site. Minizymes A and B were of the same length but had different sequences. The cleavage sites attached by hairpin ribozymes, HDVribozymes, and EGSs A and B were different, as indicated in C. C, secondary structures of the ABL and B2A2 mRNA substrates. The regions nearthe splicing junction are indicated by different colors for each exon. The substrate-recognition site of each ribozyme is indicated by underlining indifferent colors. The arrows show sites of cleavage.
Comparison of Ribozymes in Vitro and in Vivo 15381
HeLa cells were transfected with plasmids with a tRNAVal orU6 promoter. After 36 h, we examined the localization of eachexpressed ribozyme by in situ hybridization and Northern blot-ting analysis of fractionated cells. For in situ hybridization,Cy3-labeled probes were incubated with fixed and permeabi-lized cells and then the fluorescence of Cy3 was detected byconfocal microscopy. In this way, we confirmed the expectedlocations of all 19 different transcripts. Typical examples of ourresults are shown in Fig. 3, in which the red signals indicatethe presence of a hairpin ribozyme (top) or an EGS (bottom). Inall cases examined, without exception, tRNAVal-driven tran-scripts were transported to the cytoplasm and U6-driven tran-scripts localized in the nucleus.
For Northern blotting analysis, HeLa cells were fractionatedto yield nuclear and cytoplasmic fractions and RNA was ex-tracted from each fraction. This RNA was allowed to hybridizewith an appropriate 32P-labeled probe after electrophoresis(Fig. 4). Without exception, tRNAVal-driven ribozymes andEGSs expressed under control of the tRNAVal promoter werefound in the cytoplasmic fraction and U6-driven transcriptswere found in the nuclear fraction. The levels of all transcriptswere very similar (they differed by less than 20%), irrespectiveof the type of ribozyme expressed and the expression system(tRNAVal or U6 promoter). Since the steady-state level of thetranscript (reflecting its stability in cells) is a major determi-
nant of ribozyme activity in vivo, if levels of transcripts had notbeen similar, our comparison of activities in vivo would havebeen more difficult (see the next section).
The Activities of Various Functional RNAs in CulturedCells—We cotransfected HeLa cells with an expression plasmidthat encoded an appropriate ribozyme unit(s) under the controlof the tRNAVal or U6 promoter, and a plasmid that encoded thetarget BCR-ABL or ABL sequence fused with a gene for lucif-erase (luc), to evaluate the intracellular activity of ribozymes.The plasmid pB2A2-luc contained a sequence of B2A2 mRNA,while pABL-luc contained a sequence of 300 nt that encom-passed the same target cleavage site and the junction betweenexon 1 and exon 2 of normal ABL mRNA. After transientexpression of the ribozyme, substrate, and luciferase in indi-vidual cell lysates, we estimated the intracellular activity ofeach ribozyme by measuring luciferase activity.
Our results are shown in Fig. 5. The luciferase activity re-corded when we used each target gene-expressing plasmid(pB2A2-luc or pABL-luc) was taken as 100%. The data pre-sented are the results of three to six independent experiments.However, the various sets of experiments were performed ondifferent days and transfection efficiencies varied, dependingon the conditions of cells on each specific day. As a consequence,standard errors (error bars) were relatively large. However,when experiments were carried out on the same day, standard
FIG. 2. The cleavage activities of ribozymes in vitro. The cleavage of the B2A2 substrate by tRNAVal-driven ribozymes (A) and U6-drivenribozymes (B) was examined. The 59-32P-labeled substrate (112 nt) and the cleavage products were detected by autoradiography. Reactions wereperformed under single-turnover conditions. The rates of cleavage of the B2A2 substrate and the ABL substrate, kobs, by the tRNAVal-drivenribozymes and U6-driven ribozymes were measured and the results are summarized in C and D, respectively. Mini, minizyme; Hair, hairpin.
errors were in the range of 10–20%. Nonetheless, the rankorder of the activities of ribozymes always remained the same;thus, the data presented in Fig. 5 can be compared at leastqualitatively.
Expression of the tRNAVal portion by itself had no inhibitoryeffect. In all the cases when tRNAVal-ribozymes were directedagainst B2A2 target (results indicated by purple colors in Fig.5) and ABL target (indicated by blue colors), the luciferaseactivity decreased (the U6-driven maxizyme was not con-structed in this study). As expected and in accord with previousfindings (15, 20, 21, 25, 45–47), the tRNAVal-driven maxizymeshowed high level specificity, cleaving only B2A2 mRNA with-out damaging ABL mRNA (Fig. 5A). No other ribozyme was
FIG. 3. Confocal microscopic im-ages showing the detection by in situhybridization of ribozymes and EGSexpressed in mammalian cells. Cy3-labeled probes were used for detection ofthe tRNAVal-driven and U6-driven hair-pin ribozymes. Similar images were ob-tained for all 19 different constructs, i.e.tRNA-driven ribozymes and EGSs weretransported to the cytoplasm, and U6-driven ribozymes and EGSs were local-ized in the nucleus.
FIG. 4. Nuclear localization of functional RNAs. The steady-state levels of tRNAVal-driven ribozymes (A) and U6-driven ribozymes(B) and their localization are shown. Approximately the same levels ofexpression of functional RNAs from both promoters were observed. N,nuclear fraction; C, cytoplasm fraction. Without exception, tRNAVal-driven ribozymes and EGSs were localized in the cytoplasm and U6-driven ribozymes and EGSs were localized in the nucleus.
FIG. 5. Inhibitory effects in cultured cells of tRNAVal-drivenribozymes (A) and U6-driven ribozymes (B) on the expression ofchimeric BCR-ABL-luciferase and ABL-luciferase genes. A plas-mid that encoded a ribozyme or EGS and a plasmid that encoded thetarget gene were used to cotransfect HeLa cells. Decreases in luciferaseactivity (61) indicate the cleavage of transcripts by ribozymes in cells.The effects on the two different substrates are indicated by differentcolors. Mini, minizyme; Hair, hairpin.
Comparison of Ribozymes in Vitro and in Vivo 15383
able to distinguish between these two substrates. The extent ofthe decrease in luciferase activity was almost the same when invitro selected minizymes were tested; they were slightly lesseffective than other ribozymes in cultured cells. As expected,tRNAVal-driven EGSs were ineffective since they were exportedto the cytoplasm and their intracellular actions are known todepend on nuclear RNase P (Fig. 5A). We had expected that invitro selected minizymes might be more active than their pa-rental ribozymes because the minizymes were selected for theability to act at low concentrations of Mg21 ions (43). However,our expectations were not confirmed in cultured cells. Impor-tantly, our data in cells demonstrate that those ribozymes,whose activity in vitro was almost 3 orders of magnitude lowerthan that of a hammerhead ribozyme, still exhibited significantactivity in cells when an appropriate, high level expressionsystem, which allows transport of ribozyme transcripts to thecytoplasm, was used.
When EGSs were expressed under control of the U6 pro-moter, they did demonstrate intracellular activity (Fig. 5B).The various other U6-driven ribozymes did not have any sig-nificant negative effects on expression of the luciferase gene.These results demonstrated clearly that only exported ri-bozymes (EGSs are not ribozymes per se) have significantcleavage activity because of the requirement for their colocal-ization with the target mRNA in the cytoplasm. Our datademonstrate that the target mRNAs in the cytoplasm are sig-nificantly more accessible to ribozymes than are the corre-sponding nuclear pre-mRNAs.
DISCUSSION
Kinetics of Reactions in Vitro Reveal the Superior CleavageActivity of a Hammerhead Ribozyme against Relatively LongSubstrates—To investigate the actions of various ribozymes,we first determined kinetic parameters in vitro under single-turnover conditions. The rate constants, kobs, which are sum-marized in Fig. 2 (C and D) and were obtained in the presenceof excess ribozyme, reflected the rate of the chemical cleavagestep because almost all of each substrate had been captured byeach ribozyme when reactions were started by the addition ofMg21 ions to the pre-heated and cooled ribozyme-substratemixtures. Thus, the rate of the association step, which is asecond-order reaction and was low under these conditions,could be ignored. In our studies, we used two relatively longsubstrates (92 and 121 nt) because kinetic parameters for shortsubstrates have been well established and our purpose was tocompare the activities of various ribozymes in vitro and in vivoagainst structured RNA substrates.
Fig. 2 shows that the activity of the hammerhead ribozyme(Wt Rz) was significantly higher than that of all the otherribozymes. This was true for both substrates and for two dif-ferent transcripts (tRNAVal- and U6-driven). The activityagainst the structured long substrate of the hammerhead ri-bozyme was about 2 orders of magnitude higher than that ofthe most of the other ribozymes. Nevertheless, the absoluteactivity, with the rate constants of 0.01–0.02 min21 for thecleavage by the hammerhead ribozyme of the long substrate,was 2 orders of magnitude lower than the absolute activityagainst a short substrate. Under similar conditions, short RNAsubstrates can be cleaved by hammerhead ribozymes with rateconstants of 1–2 min21. The difference reflects the fact thatlonger RNA substrates tend to form structures that limit accessby ribozymes (60). With our long substrates we showed that thehammerhead ribozyme was the best ribozyme for cleavage ofsuch structured RNAs despite the fact that the hairpin ri-bozyme can cleave short substrates as efficiently as hammer-head ribozymes, with rate constants of 1–2 min21 (24).
Absence of Any Correlation between the Activities of Ri-
bozymes in Vitro and in Cultured Cells—Each functional RNAexhibited cleavage activity both in vitro and in cultured cellsbut with varying efficiency. As expected, the hammerhead ri-bozyme had significant activity in cultured cells (Fig. 5). We didnot expect 100% inhibition in our transient expression assaysbecause not all cells would have been transfected by the variousrespective plasmids. As seen from Fig. 5, there was no correla-tion between the trends in the ribozyme activity in vitro (Fig. 2)and in those in cultured cells (Fig. 5). The kinetic parametersobtained in vitro indicated that the hammerhead ribozyme wassuperior to other ribozymes, but the activities of other tRNAVal-driven ribozymes in cultured cells were significantly improvedrelative to that of the hammerhead ribozyme (Fig. 5A).
It has been suggested that the rate-limiting step in vivo of areaction mediated by a catalytic RNA, such as a ribozyme, isthe substrate-binding step (59, 60). Our analysis is clearlyconsistent with this suggestion because the efficacies of ri-bozymes in vivo depend more strongly on the expression systemand the localization within cells than on cleavage activities invitro. It is now apparent that the rate-limiting step in ri-bozyme-mediated reactions in vivo is not the cleavage step.Thus, hairpin and other ribozymes with limited activities invitro can have significant inhibitory effects in vivo, as demon-strated previously (9, 11). Even HDV ribozymes, which are veryinefficient in vitro, had clear inhibitory effects in cultured cells.
Cytoplasmic mRNAs Were Cleaved Significantly More Effec-tively than Nuclear Pre-mRNAs—Ribozymes expressed underthe control of the tRNAVal promoter and the U6 promoter werefound, as anticipated, in the cytoplasm and in the nucleus,respectively. The tRNAVal-driven ribozymes that had been ex-ported to the cytoplasm had higher activities than the corre-sponding tRNAVal-driven ribozymes that retained in the nuc-leus (Fig. 5). Nuclear pre-mRNAs might be less accessible toribozymes than cytoplasmic mature mRNAs becausepre-mRNAs form complexes with heterogeneous nuclear prot-eins and small nuclear ribonuclear proteins and interact withvarious RNA-binding proteins, for example, proteins involvedin splicing and in the export of processed mRNAs. It is alsolikely that higher ordered structures of mRNAs are disruptedmore effectively in the cytoplasm by various RNA helicases(60). Thus, a ribozyme can attack its target site during thebreathing of the cytoplasmic target mRNA.
The present analysis confirmed that cleavage by variousribozymes occurs more efficiently in the cytoplasm than in thenucleus. Without exception, the tRNAVal-driven ribozymes thathad been exported to the cytoplasm (Figs. 3 and 4) had inhib-itory effects (Fig. 5A), whereas U6-driven ribozymes that hadremained in the nucleus were completely ineffective (Fig. 5B),despite the fact that both types of ribozyme were targeted tothe identical site (Fig. 1C) and both had similar activity in vitro(Fig. 2, C and D). By contrast, the EGS in the nucleus mediatedcleavage more effectively than the EGS that had been exportedto the cytoplasm because the action of the EGS requires RNaseP, which is localized in the nucleus. It should be noted also thatRNase P might have an RNA unwinding activity.
We confirmed unambiguously that mature mRNAs in thecytoplasm were more accessible to ribozymes than pre-mRNAsin the nucleus. Thus, if we are to exploit ribozyme activity incells, ribozymes must be concentrated in the cytoplasm whileEGSs must remain in the nucleus. Our findings should beuseful for the selection of expression systems and the futuredesign of intracellularly active ribozymes.
Acknowledgments—We thank Professor Sidney Altman andDr. Cecilia Guerrier-Takada (Department of Molecular, Cellular andDevelopmental Biology, Yale University, New Haven, CT) for helpfulcomments and for the gifts of EGS expression vectors.
1. Cech, T. R., Zaug, A. J., and Grabowski, P. J. (1981) Cell 27, 487–4962. Guerrier-Takada, C., Gardiner, K., Marsh, T., Pace, N., and Altman, S. (1983)
Cell 35, 849–8573. Buzayan, J. M., Gerlach, W. L., and Bruening, G. (1986) Nature 323, 349–3534. Sharmeen, L., Kuo, M. Y., Dinter-Gottlieb, G., and Taylor, J. (1988) J. Virol.
62, 2674–26795. Wu, H.-N., and Lai, M. M. (1989) Science 243, 652–6546. Symons, R. H. (1992) Annu. Rev. Biochem. 61, 641–6717. Birikh, K. R., Heaton, P. A., and Eckstein, F. (1997) Eur. J. Biochem. 245, 1–168. Gesteland, R. F., Cech, T. R., and Atkins, J. F. (eds) (1999) The RNA World,
2nd Ed., Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY9. Ojwang, J. O., Hampel, A., Looney, D. J., Wong-Staal, F., and Rappaport, J.
(1992) Proc. Natl. Acad. Sci. U. S. A. 89, 10802–1080610. Sullenger, B. A., and Cech, T. R. (1993) Science 262, 1566–156911. Yu, M., Leavitt, M. C., Maruyama, M., Yamada, O., Young, D., Ho, A. D., and
Wong-Staal, F. (1995) Proc. Natl. Acad. Sci. U. S. A. 92, 699–70312. Bertrand, E., Castanotto, D., Zhou, C., Carbonnelle, C., Lee, N. S., Good, P.,
Chatterjee, S., Grange, T., Pictet, R., Kohn, D., Engelke, D., and Rossi, J. J.(1997) RNA 3, 75–88
13. Kawasaki, H., Eckner, R., Yao, T.-P., Taira, K., Chiu, R., Livingston, D. M.,and Yokoyama, K. K. (1998) Nature 393, 284–289
14. Kuwabara, T., Warashina, M., Orita, M., Koseki, S., Ohkawa, J., and Taira, K.(1998) Nat. Biotechnol. 16, 961–965
15. Kuwabara, T., Warashina, M., Tanabe, T., Tani, K., Asano, S., and Taira, K.(1998) Mol. Cell 2, 617–627
16. Lan, N., Howrey, R. P., Lee, S. W., Smith, C. A., and Sullenger, B. A. (1998)Science 280, 1593–1596
17. Plehn-Dujowich, D., and Altman, S. (1998) Proc. Natl. Acad. Sci. U. S. A. 95,7327–7332
19. Krupp, G., and Gaur, R. K. (eds) (2000) Ribozyme, Biochemistry and Bio/Technology, Eaton Publishing, Natick, MA
20. Tanabe, T., Takata, I., Kuwabara, T., Warashina, M., Kawasaki, H., Tani, K.,Ohta, S., Asano, S., and Taira, K. (2000) Biomacromolecules 1, 108–117
21. Tanabe, T., Kuwabara, T., Warashina, M., Tani, K., Taira, K., and Asano, S.(2000) Nature 406, 473–474
22. Koseki, S., Tanabe, T., Tani, K., Asano, S., Shioda, T., Nagai, Y., Shimada, T.,Ohkawa, J., and Taira, K. (1999) J. Virol. 73, 1868–1877
23. Kuwabara, T., Warashina, M., Nakayama, A., Ohkawa, J., and Taira, K.(1999) Proc. Natl. Acad. Sci. U. S. A. 96, 1886–1891
24. Hampel, A. (1998) Prog. Nucleic Acids Res. Mol. Biol. 58, 1–3925. Kuwabara, T., Warashina, M., and Taira, K. (2000) Trends Biotechnol. 18,
462–46826. Arts, G. J., Kuersten, S., Romby, P., Ehresmann, B., and Mattaj, I. W. (1998)
EMBO J. 17, 7430–744127. Arts, G. J., Fornerod, M., and Mattaj, I. W. (1998) Curr. Biol. 8, 305–31428. Forster, A. C., and Altman, S. (1990) Science 249, 783–78629. Yuan, Y., and Altman, S. (1994) Science 263, 1269–127330. Guerrier-Takada, C., Salavati, R., and Altman, S. (1997) Proc. Natl. Acad. Sci.
U. S. A. 94, 8468–847231. Muller, A. J., Young, J. C., Pendergast, A. M., Pondel, M., Landau, N. R.,
Littman, D. R., and Witte, O. N. (1991) Mol. Cell. Biol. 11, 1785–1792
32. Nowell, P. C., and Hungerford, D. A. (1960) Science 132, 1497–149933. Das, G., Henning, D., Wright, D., and Reddy, R. (1988) EMBO J. 7, 503–51234. Ohkawa, J., and Taira, K. (2000) Hum. Gene Ther. 11, 577–58535. Hertel, K. J., Herschlag, D., and Uhlenbeck, O. C. (1996) EMBO J. 15,
3751–375736. Zhou, D.-M., and Taira, K. (1998) Chem. Rev. 98, 991–102637. Yoshinari, K., and Taira, K. (2000) Nucleic Acids Res. 28, 1730–174238. McCall, M. J., Hendry, P., and Jennings, P. A. (1992) Proc. Natl. Acad. Sci.
U. S. A. 89, 5710–571439. Tuschl, T., and Eckstein, F. (1993) Proc. Natl. Acad. Sci. U. S. A. 90,
6991–699440. Fu, D. J., Benseler, F., and McLaughlin, L. W. (1994) J. Am. Chem. Soc. 116,
4591–459841. Long, D. M., and Uhlenbeck, O. C. (1994) Proc. Natl. Acad. Sci. U. S. A. 91,
6977–698142. Amontov, S. V., and Taira, K. (1996) J. Am. Chem. Soc. 118, 1624–162843. Conaty, J., Hendry, P., and Lockett, T. (1999) Nucleic Acids Res. 27,
2400–240744. McCall, M. J., Hendry, P., Mir, A. A., Conaty, J., Brown, G., and Lockett, T. J.
(2000) Mol. Biotechnol. 14, 5–1745. Kuwabara, T., Warashina, M., and Taira, K. (2000) Curr. Opin. Chem. Biol. 4,
669–67746. Kurata, H., Miyagishi, M., Kuwabara, T., Warashina, M., and Taira, K. (2000)
J. Biochem. Mol. Biol. 33, 359–36547. Warashina, M., Kuwabara, T., and Taira, K. (2000) Structure 8, R207-R21248. Kruger, M., Beger, C., Li, Q.-X., Welch, P. J., Tritz, R., Leavitt, M., Barber,
J. R., and Wong-Staal, F. (2000) Proc. Natl. Acad. Sci. U. S. A. 97,8566–8571
49. Earnshaw, D. J., Hamm, M. L., Piccirilli, J. A., Karpeisky, A., Beigelman, L.,Ross, B. S., Manoharan, M., and Gait, M. J. (2000) Biochemistry 39,6410–6421
50. Fedor, M. J. (2000) J. Mol. Biol. 297, 269–29151. Warashina, M., Takagi, Y., Stec, W. J., and Taira, K. (2000) Curr. Opin.
Biotechnol. 11, 354–36252. Takagi, Y., Warashina, M., Stec, W. J., Yoshinari, K., and Taira, K. (2001)
Nucleic Acids Res. 29, in press53. Nakano, S., Chadalavada, D. M., and Bevilacqua, P. C. (2000) Science 287,
1493–149754. Anavoranich, S., and Perreault, J.-P. (1998) J. Biol. Chem. 273, 13182–1318855. Nishikawa, F., Roy, M., Fauzi, H., and Nishikawa, S. (1999) Nucleic Acids Res.
27, 403–41056. Smith, D., and Pace, N. R. (1993) Biochemistry 32, 5273–528157. Warnecke, J. M., Held, R., Busch, S., and Hartmann, R. K. (1999) J. Mol. Biol.
290, 433–44558. Thomas, B. C., Li, X., and Gegenheimer, P. (2000) RNA 6, 545–55359. Castanotto, D., Scherr, M., Lee, N.-S., and Rossi, J. J. (2000) in Ribozyme,
Biochemistry and Bio/Technology (Krupp, G., and Gaur, R. K., eds) pp.421–440, Eaton Publishing, Natick, MA
60. Warashina, M., Kuwabara, T., Kato, Y., Sano, M., and Taira, K. (2001) Proc.Natl. Acad. Sci. U. S. A. 98, in press
61. Krummel, D. A. P., Knet, O., MacMillan, A. M., and Altman, S. (2001) J. Mol.Biol. 295, 1113–1118
62. Koseki, S., Ohkawa, J., Yamamoto, R., Takebe, Y., and Taira, K. (1998)J. Controlled Release 53, 159–173
Comparison of Ribozymes in Vitro and in Vivo 15385
