Proc. Natd. Acad. Sci. USAVol. 89, pp. 5710-5714, July 1992Biochemistry

Minimal sequence requirements for ribozyme activity(minimized ribozymes/RNA endonucleases/RNA cleavage/DNA-containing ribozymes)

MAXINE J. MCCALL, PHILIP HENDRY, AND PHILIP A. JENNINGSCommonwealth Scientific and Industrial Research Organization, Division of Biomolecular Engineering, Laboratory for Molecular Biology, P.O. Box 184,North Ryde, NSW 2113, Australia

Communicated by W. J. Peacock, March 12, 1992 (receivedfor review August 28, 1991)

ABSTRACT The hammerhead ribozyme, as engineered byJ. Haseloff and W. L. Gerlach [(1988) Nature (London) 334,585-591], is an RNA molecule containing two regions ofconserved nucleotides, a double helix, called helix H, whichconnects the two conserved regions, and flanking arms ofvariable sequence, which hybridize the ribozyme to its specifictarget. Here we show that this ribozyme may be reduced in sizeand still retain cleavage activity by replacing helix II with justa few nucleotides that cannot form Watson-Crick base pairsbetween themselves. Furthermore, the nucleotides replacingheli II and the nucleotides in the flanking arms may besubstituted with DNA, and this small, DNA-containing ri-bozyme is fully as active as the original, full-size ribozyme.Cleavage activity of the minimized ribozyme depends on thenumber and sequence of the few nucleotides that replace helixII; optimal activity, thus far, is achieved by four or fivedeoxyribopyrimidines. The minimized ribozyme, or "mimn-zyme," is active as a monomer, as shown by its nearly constantactivity over a concentration range varying 25,000-fold, by themobility of the minizyme-substrate complex in nondenaturingpolyacrylamide gels as compared with other nucleic acid mol-ecules of known size, and by other observations. Theseminizymes provide an excellent model system for studying thestructure and mechanism of catalytic RNA; they might also beuseful in a variety of biological applications.

Ribozymes are RNA molecules that can cut or ligate otherRNA or DNA molecules in a catalytic fashion (1, 2). Thehammerhead ribozyme, one of the best-known ribozymes,has been studied extensively in isolated chemical systems(3-7) and used in gene-control studies in living cells (8-11).The essential features of a hammerhead ribozyme weredetermined initially by comparing the base sequences (3, 12)of a number of plant virusoids, linear satellite RNAs, and aviroid that can all cut themselves at unique locations (12-14);later they were refined in a comprehensive study of basemutations (15). These essential features are (i) three doublehelices ofany base sequence, which meet at the cleavage site,(ii) the site ofcleavage that has a sequence UX, with cleavageoccurring on the 3' side of X, and (iii) two stretches ofnucleotides of sequence 5'-CUGAZGA and 5'-GAAA, whereX and Z are any base except guanine. It is possible to separatethese essential features into two parts, a ribozyme and asubstrate, in several ways (4, 5). A hammerhead ribozyme asdefined by Haseloff and Gerlach (5) is shown in Fig. 1 at thetop. This ribozyme contains the two stretches of conservednucleotides and one of the three double helices (helix II),whereas the substrate forms double helices I and III incombination with the ribozyme. Because the base sequencesof helices I and III are not conserved, the ribozyme may bedesigned to cut a substrate of virtually any sequence.

Our primary aim in this work was to minimize the size ofa Haseloff-Gerlach ribozyme. Our motivation was 2-fold: (i)we are interested in determining the structure ofthe ribozymein complex with its substrate, and (ii) we are interested intesting ribozymes as pharmaceuticals. For structural studiesby NMR or x-ray crystallography and for testing syntheticribozymes in vivo, milligrams of material are required. Byreducing its size and by replacing many RNA nucleotideswith DNA, the ribozyme could be made more rapidly, morecheaply, and in greater yield by the step-wise, solid-phase,synthetic methods currently in use. For the structural deter-mination, a smaller ribozyme would simplify the work. Inbiological applications, a smaller ribozyme may be taken upby cells more efficiently, and, if it consists mainly of DNA,it might be more resistant to ribonucleases and, thus, survivelonger in the cell.

MATERIALS AND METHODSSolid-Phase Synthesis of Oligonucleotides. All oligonucleo-

tides were synthesized on an Applied Biosystems 380 and/or391 synthesizer by using 2-cyanoethylphosphoramiditechemistry. DNA monomers were from Applied Biosystems;RNA monomers, protected at the 2' position with a tert-butyldimethylsilyl group, were from Peninsula Laboratoriesor MilliGen (Bedford, MA). The 3' nucleotide in all moleculesis a deoxyribonucleotide. All RNA-containing oligonucleo-tides, with 5'-trityl groups removed, were processed asfollows. The oligonucleotide was cleaved from the column inNH4OH/ethanol, 3:1, and heated overnight at 550C. Thesolutions were evaporated to near dryness, coevaporatedseveral times with H20/ethanol, 3:1, and then the amount ofmaterial was estimated by measuring UV absorbance. The2'-group of the sugar was deprotected by treatment overnightwith 1 M tetrabutylammonium fluoride in tetrahydrofuran (10Al per OD260 unit). The tetrabutylammonium ions wereremoved by passing the oligonucleotide solution twicethrough a Dowex 50X8-200 cation-exchange column in theNa+ form; eluate volume was reduced by extraction with2-butanol, and the oligonucleotide was precipitated withsodium acetate and ethanol. The oligonucleotide was thenpurified by electrophoresis on a 10-20%o (depending onlength) polyacrylamide gel containing 7 M urea. The band ofinterest was visualized by UV shadowing or ethidium bro-mide staining, excised, and soaked in several changes ofwater over 24 hr. The supernatant was removed from the gelslices, concentrated with 2-butanol, and extracted with phe-nol/chloroform and with ether. The oligonucleotide was thenprecipitated with sodium acetate and ethanol, washed withcold 80% ethanol, redissolved in 10 mM Tris'HCI, pH 8.0/0.2mM EDTA, quantified by UV spectroscopy, and frozen. Theoligonucleotides were phosphorylated on their 5'-ends byusing [y-32P]ATP and T4 polynucleotide kinase (Bresatec,

Abbreviations: CAT, chloramphenicol acetyltransferase; nt, nucle-otide(s).

5710

The publication costs of this article were defrayed in part by page chargepayment. This article must therefore be hereby marked "advertisement"in accordance with 18 U.S.C. §1734 solely to indicate this fact.

Dow

nloa

ded

by g

uest

on

Feb

ruar

y 4,

202

0

Proc. Natl. Acad. Sci. USA 89 (1992) 5711

4, Helix III Helix I Substrate5' ACCUGCGGG[C AUGAAGUGUC 3' . . . . . . .-3' tGGACGCCC A UACUUCACAG 5'

A c Conserved -wV R4U,RNAA A ~~~~~~basesAGU

A-U Helix IIG-CG-CUUl ConnectorUUI

3' tGGACGCCC A UACUUCACAG 5'ACUK MORNA

3' tGGACGCCC A UACUUCACAG5'

A A M4URNA

UU3' tggacgccC tacttcacagN'

AJA M4UDNA

3 tggacgccC tacttcacagN ------

ttAGU M4t,DNA

tttt

5'GGUUUACCUGCGGG AUGAAGUGUC UU3'UGGACGCCC A UACUUCACAGGC

A

G [AU UUU

MS4URNA

FIG. 1. Base sequence, schematic representation, and names ofribozyme, minizymes, and their common substrate. Conserved ri-

bonucleotides are drawn in boxes and depicted schematically bythick lines, other ribonucleotides are depicted by uppercase lettersand thin lines, and deoxyribonucleotides are depicted by lowercaseletters and wavy lines. R, ribozyme containing helix II; M, minizymenot containing helix II; MS, a minizyme and substrate in the same

molecule; the first subscript indicates the bases in the connector, andthe second subscript indicates whetherRNA orDNA nucleotides arein the flanking arms that form helices I and III with the substrate. Theribozyme and minizymes cleave the substrate after the cytidineresidue marked by a downward arrow.

Adelaide, Australia) under standard conditions, except thatseveral units of ribonuclease inhibitor (RNasin, Promega)were added to the reaction mixture. The phosphorylationreaction was stopped with EDTA; the oligonucleotide solu-tion was extracted with phenol/chloroform and ether andconcentrated with 2-butanol. The 5'-end-labeled oligonucle-otide was precipitated twice, firstly from 2.5 M ammoniumacetate and ethanol and secondly from 0.3 M sodium acetateand ethanol. The pellet was washed with cold 80%o ethanoland then dissolved in 10 mM Tris HCl, pH 8/0.2 mM EDTAto make a stock solution with nominal concentration of 1 tLM.The exact concentration ofthe stock solution was determinedin the following way. All supernatants from the two precip-itation steps and the final ethanol wash were pooled, con-

centrated under vacuum, and loaded on a denaturing poly-acrylamide gel alongside a known, small fraction of the stocksolution. After electrophoresis, gel slices containing thestock and the lost oligonucleotide were excised, and theamount of material they contained was quantified by Ceren-kov counting. These data were used to determine the exactconcentration of the 5'-end-labeled oligonucleotide solution.

Transcription of Oligonucleotides. One of the oligoribonu-cleotides used in this study (MS4URNA) was transcribed froma DNA oligonucleotide containing a truncated T3 promoter.The DNA duplex was generated by extension with T7 DNApolymerase (Sequenase, United States Biochemical) afterhybridizing two oligonucleotides with sequences as follows:T3 primer: 5' AAT TAA CCC TCA CTA and MS template:5'-ACC TGC GGG TTT CAA AAT CAT CAG ATG AAG

TGT CCG AAG ACA CTT CAT GAC CCG CAG GTA AACCTT TAG TGA GGG TTA ATT.Another RNA molecule, the 173-nucleotide (nt) chloram-

phenicol acetyltransferase (CAT) substrate, was transcribedwith T7 RNA polymerase from a Pvu II digest of pGEM3containing the BamHI CAT fragment from pCM4 as de-scribed in ref. 5. Transcripts were made with the Stratagenetranscription kit: a typical reaction in a 30-gl vol contained 5pmol of template; 2 units of T3 RNA polymerase; 500 ,uMeach of rATP, rGTP, and rCTP; 100 ,uM rUTP; 10 pmol of[a-32P]rUTP (17 ,uCi; 1 Ci = 37 GBq); 10 mM dithiothreitol;and 10 units of RNasin (Promega). The reaction to constructCAT RNA was stopped after 30 min, but that to makeMS4U,RNA was stopped after only 5 min to minimize self-cleavage of the transcribed molecule. The mixtures wereextracted with phenol/chloroform; RNAs were precipitatedwith sodium acetate and ethanol, and the pellets were dis-solved in sterile water. Concentration of the solution wasdetermined by knowing the number of uridines in the se-quence, the specific activity of the [a-32P]rUTP used, and theactivity of an aliquot of the transcript solution.

Cleavage Reactions. Standard conditions for the cleavagereactions are as follows. Reactions were at 37°C in a 30-,lI volof 10 mM MgCl2/50 mM Tris HCI, pH 8.0; the ribo/minizymes and substrate (labeled on the 5' end with 32p) wereheated separately at 70°C for 3 min in Tris buffer, then cooledquickly on ice before mixing, adding MgCl2, and incubating.At appropriate times, a 3-ILI aliquot was removed from thereaction and added to 6 ,ul of quenching solution containing80% formamide, 20 mM EDTA, and dye. The quenchedsamples were then analyzed by electrophoresis on 15%polyacrylamide gels containing 7M urea as a denaturant. Thesubstrate and product of cleavage were visualized by auto-radiography, and gel slices corresponding to their positionswere excised and quantified by Cerenkov counting. The ratioof [product] to ([substrate] plus [product]) was plotted versustime as shown. Rate constants were calculated by fitting, bya Newton-Raphson iterative procedure, data for percentageof product formed (%oP) as a function oftime () to an equationof the form %P = %6P. - C-exp[-kt], where %oP. is per-centage of product at infinite time, C is the differencebetween percentage of product at t = oo and t = 0, and k is thefirst-order rate constant. A listing of the program, which runson MS-DOS computers, is available on request.Nondenaturing Gels. The molecular species in a mixture

that initially contained equimolar amounts (0.1 uM) of 5'-end-labeled substrate and ribozyme or minizyme and inwhich conditions for the cleavage reaction were as describedabove were analyzed under nondenaturing conditions byelectrophoresis on 10% polyacrylamide gels kept at 4°C.After a 30-min reaction at 37°C, samples were mixed with anequal volume of 30% sucrose/0.1% dye and stored at -20°C.The polyacrylamide gels were preelectrophoresed overnightat constant voltage (100 V) in a buffer of 50 mM Tris borate,pH 8/10mM MgCI2 and then preelectrophoresed for 2 hr nextday after replacing the buffer with fresh solution (16, 17).Then samples were loaded and electrophoresed for 8 hr atconstant voltage (400 V). In addition, the same samples wereelectrophoresed for 6 hr at 400 V on another 10% polyacryl-amide gel, which was buffered by 90 mM Tris borate, pH 8/2mM EDTA containing no Mg2+. The components of severalbands in the nondenaturing gels were determined by excisingthe bands, eluting the molecules in formamide, and separat-ing them by electrophoresis on a denaturing, polyacrylamidegel; the 3'-product of cleavage of a 5'-end-labeled substrateis unobservable by this method.

RESULTS AND DISCUSSIONThe hammerhead ribozyme (5) used is shown in Fig. 1 at top;its sequence (the two conserved regions are drawn in boxes)

Biochemistry: McCall et al.

Dow

nloa

ded

by g

uest

on

Feb

ruar

y 4,

202

0

Proc. Natl. Acad. Sci. USA 89 (1992)

is at left, and a schematic drawing is at right. We call thisribozyme R4U,RNA and use it as a reference in assessingcleavage activities ofother molecules. TheRNA substrate forthis ribozyme is also shown in Fig. 1 (top), where the cut sitefollows the cytidine residue marked by a downward arrow.The ribozyme and substrate together are designed to formtwo double helices each containing 10 base pairs (bp) oneither side of the central cytidine. When the substrate isradioactively labeled at its 5' end, the cleavage reaction canbe monitored easily by electrophoresing the reaction mixtureon a denaturing polyacrylamide gel.To see how much a ribozyme could be reduced in size

without complete loss of activity, we synthesized chemically,in small amounts, a series ofRNA molecules that containedthe conserved bases but omitted as many of the noncon-served bases as possible. A drastic reduction of the lowerdomain from 22 to 10 nt, shown in Fig. 1 as MORNA, wasunsuccessful in that it eliminated all RNA-cutting activity(Fig. 2a). But a less drastic reduction from 22 to 14 nt, shownin Fig. 1 as M4U,RNA, preserved much cutting activity whilereducing the size (Fig. 2a). The diminished ribozyme,M4U,RNA, which lacks a helix II, is less active than thefull-size ribozyme R4U,RNA (Fig. 2a, Table 1), but its mea-surable activity indicates that helix II (previously thoughtessential) is dispensable to the cleavage reaction. We definea ribozyme that does not contain helix II as a "minizyme."

80a

R4U,RNA

40-

20 M4U, RNA

MO, RNA

:0

0 20 4 0 80

ff 8080 -

b

M4t, DNA60 -

= R4U, RNA O

M4U, DNA.

40-

20 1 / / -M 4U, RNA

20

0 20 40 60 80

time (min)

FIG. 2. Plots of percentage of product versus time for cutting ofsubstrate by ribozyme and minizymes, as labeled. Reactions were in50 mM Tris-HCI, pH 8/10 mM MgCl2 at 37°C; substrate concentra-tion was 0.1 ,uM, and ribozyme and minizyme concentrations were0.6 AtM in all reactions. (a) Effect of reducing size of an all-RNAribozyme. (b) Effect of replacing RNA nucleotides by DNA in a

minizyme ofconstant length. (Data in a for the ribozyme R4U,RNA areincluded in b for reference.)

Table 1. Kinetic data for cleavage by various minizymesRibozyme/minizyme Rate constant, min-1 %P.

R4U,RNA 0.15 59M4U.RNA 0.05 31M4U,DNA (4U) 0.05 54M4tDNA (4T) 0.16 (0.03) 62 (5)M2t,DNA (2T) <0.001M3t,DNA (3T) 0.014 65M5tDNA (5T) 0.23 71M4a,DNA (4A) 0.02 64MttCtDNA (TTCT) 0.21 76M4tKrp-DNA + Kruippel substrate 0.23 76M4tCAT-DNA + short CAT substrate 0.03 70M4tCAT-DNA + long CAT substrate 0.02 94

Reactions were in 50 mM Tris HCI, pH 8/10 mM MgCl2 at 370C;substrate concentration was 0.1 AM, and ribozyme and minizymeconcentrations were 0.6 AiM. Except where indicated otherwise, thesubstrate is that shown at top of Fig. 1. Minizyme M4tDNA wassynthesized several times, and its kinetics of cleavage was deter-mined independently on four separate occasions; SDs of the meanrate constant and mean percentage of product at infinite time (%P.)are in parentheses.

To see how many deoxyribonucleotides could be intro-duced into the minizyme (M4U,RNA) without activity loss, wesynthesized a series of molecules in which some or all of thenonconserved RNA nucleotides were replaced by DNAnucleotides. First, most of the ribonucleotides in theminizyme that contribute to forming double helices I and IIIwere replaced with DNA (wavy lines in M4U,DNA, Fig. 1).Unexpectedly, this minizyme with DNA arms cut the sub-strate to a greater extent than the all-RNA minizyme (Fig. 2b,Table 1). To continue the trend, all nucleotides connectingthe two regions of conserved bases were replaced with DNA(wavy lines in M41,DNA, Fig. 1), and the RNA-cutting activityof this minizyme equaled that of the reference (Fig. 2b, Table1). Thus, we have made a diminutive ribozyme that containsjust 12 RNA nt, with 4 DNA nt connecting the two regions ofconserved RNA bases, and all DNA elsewhere in the helix-forming arms; and this small, DNA-rich molecule cleaves thesubstrate as well as the full-size, all-RNA ribozyme, at leastfor the particular substrate studied here. We have not at-tempted to replace any of the conserved ribonucleotides withDNA in the minizymes because such replacements in anall-RNA hammerhead ribozyme decrease its activity signif-icantly (18, 19). The increase in cleavage activity seen uponreplacing ribonucleotides with deoxyribonucleotides in theminizyme probably arises from a slight alteration to thestructure of the catalytic core of the minizyme-substratecomplex; a different structure must be adopted by a ri-bozyme-substrate complex in which ribonucleotides are re-placed by deoxyribonucleotides in the substrate, and forwhich an activity decrease is seen (20).To make the minizyme even smaller, the number of nu-

cleotides in the helix-forming arms may be reduced. Weexperimented with minizyme M4t,DNA and found that thenumber of base pairs in helices I and III can be reduced from10 to 4, in both double helices, with maintenance of cleavageactivity (unpublished data). Thus a minizyme can be at leastas small as 22 nt.

Effect of Length and Sequence of the Connector. Within thesmall DNA-rich minizyme, the only nucleotides that can bevaried now are those connecting the two regions ofconservedRNA bases; all other nonconserved nucleotides are in thehelix-forming arms and, hence, are determined by the basesequence of the target. Therefore, to find the optimal com-position of the connector, new minizymes were designedbased on the most active minizyme, M4t,DNA. To find theoptimal length, we compared the cleavage activities of four

5712 Biochemistry: McCall et al.

Dow

nloa

ded

by g

uest

on

Feb

ruar

y 4,

202

0

Proc. Natl. Acad. Sci. USA 89 (1992) 5713

minizymes containing the DNA nucleotides 2T, 3T, 4T, or 5Tin their connectors. Observed rate constants and %oP. (Table1) show that the 5T minizyme is marginally better than the 4Tminizyme and that both are substantially better thanminizymes with shorter connectors. Presumably, a connec-tor with <4T restricts the minizyme in adopting the activeconformation for cleavage. We did not investigate further theeffect of increasing the connector's length because our aimhere was to construct a minizyme, reasonably balancedbetween activity and size. To look at the effect of the 4-ntsequence, we compared cleavage activities of fourminizymes containing the RNA nucleotides 4U, or the DNAnucleotides 4A, 4T, or 5'-TTCT in their connectors. Ob-served rate constants and %oP. (Table 1) indicate that, of thisseries, connectors of 4T and TTCT provide for the greatestactivity. The connectors 4T and TTCT probably are suffi-ciently flexible to allow the conserved RNA nucleotides toadopt the active conformation for cleavage, whereas theconnectors 4A and 4U are more rigid in structure, withpossible stacking of adenine bases in 4A and interactionsinvolving 2'-hydroxyl groups in 4U (21). Goodchild and Kohli(22) also have reduced the number of nucleotides joining thetwo stretches of conserved RNA bases in a hammerheadribozyme, although they kept bases that potentially couldform Watson-Crick pairs (and, potentially, a ribozyme di-mer). One of their all-RNA molecules, which has the se-quence 5'-GGCC in the equivalent of our connector, cleavesits substrate very slowly. A second molecule, which has5'-GGCGCC in the connector, cleaves the same substratewith higher initial rates. Because the structure of these twoconnectors would be relatively rigid, due to stacking ofguanine bases (23) and the possible formation of a GC basepair, a connector >4 nt is probably necessary to allow theirenzyme to adopt the active conformation. These initial re-sults suggest that further improvement of the cleavage ac-tivity of minizymes may be possible.Minizymes are Active asMonomers. We considered whether

some form ofhelix II might arise from the association ofbasesin the connectors of two minizyme molecules and whethersuch a minizyme dimer was the active species. A precedentexists for this: Forster et al. (24) have proposed that self-cleavage ofRNA occurs for avocado sunblotch viroid and inRNA from the newt by the combination of two differenthammerhead domains to form active dimers. However, re-

sults from the following three experiments indicate that theminizyme acts as a monomer. (i) For M4tDNA, the putativehelix II would be formed by T*T base pairs, which arerelatively unstable. A more stable helix II, consisting of T-Abase pairs, might be formed by mixing M4tDNA with M4aDNA;and, were dimerization required for cleavage, the mixedminizyme-dimer should be more active than the dimersformed by either component alone. Cleavage of substrate (0.1.M) by M4tDNA (0.4 AM), by M4aDNA (0.4 ,uM), and byM4tDNA mixed with M4aDNA (0.2 ,uM each) was followedunder standard conditions. The observed rate constants forthe reactions (0.14 min-1 for M4tDNA, 0.02 min-' forM4aDNA, and 0.04 min-1 for M4tDNA mixed with M4aDNA)indicate no synergy between the two minizymes in cleavingsubstrate. (ii) The transcribed oligonucleotide MS4U.RNA,which has a minizyme and attached substrate (see Fig. 1,bottom), shows little change in the rate constant for cleavage(-0.1 min-') over a concentration range varying 25,000-foldfrom 1.3 ,M to 55 pM (data points in Fig. 3). Were theconcentration of MS4URNA in the cleavage reactions signif-icantly less than the association constant for the putativedimerization, the rate of cleavage by the dimer would beexpected to depend on the square of the concentration (solidline in Fig. 3); the data do not follow this relationship. (iii)

b-p. markers A B C

112 3 4 5 6 F78 910 11 I1213141516190 - ~ S M C S SMC C' S S' R C C'

147 - _111 _110 -

67- _52- *

0M34 -

- It.21

:,

.:fpt

-2-

-3

-11 -10 -9 -8 -7 -6 -5

lOg [MS4U,RNA] (M)

FIG. 3. Observed rate constants (ED) for the self-cleavage of theunimolecular minizyme-substrate MS4U,RNA at various concentra-tions. Reactions were at 370C in 50 mM Tris HCl, pH 8/10 mMMgCl2. Solid line represents the expected variation of rate constantwith concentration (k o concentration2), were the reaction bimolec-ular and dimer-formation rate limiting.

FIG. 4. Mobilities of various substrates, minizymes, and ri-bozymes, and their complexes and products of cleavage in a non-denaturing, 10o polyacrylamide gel at 4°C. The electrophoresisbuffer was 50 mM Tris borate, pH 8/10 mM MgCl2. Species in eachof lanes 4-16 represent the reaction products for 30 min at 37°C understandard conditions. S, substrate only in reaction; S', noncleavablesubstrate; M, minizyme; R, ribozyme; C, equimolar mix of substrateand minizyme (or ribozyme); and C', equimolar mix of noncleavablesubstrate and minizyme (or ribozyme). The molecules in set A (lanes4-6) are an all-RNA, 21-base synthetic substrate of sequence 5'-AUUUGCGAGUCCACACUGGAG (from the Kruppel gene ofDro-sophila melanogaster) and its cognate, 34-base minizyme based onM4tDNA; those in set B (lanes 7-11) are the all-RNA, 21-basesubstrate described in Fig. 1, a noncleavable substrate of identicalbase sequence in which the central ribocytidine has been replaced bya deoxyribocytidine, and the minizyme M4tDNA; those in set C (lanes12-16) are the same cleavable and noncleavable substrates of set Band ribozyme R4U.RNA- Size markers are labeled in base pairs.Lanes: 1, DNA double helices formed by a Hpa II restriction digestof pUC19; 2, 52-bp DNA double helix; and 3, 21-bp double helixconsisting ofan all-RNA strand identical to the molecule in lane 4 andan all-DNA strand of complementary sequence. To clearly see thebands of interest, the gel above the 190-bp mark is not shown; therewere no bands above this mark in lanes 2-16.

Biochemistry: McCall et al.

1

Dow

nloa

ded

by g

uest

on

Feb

ruar

y 4,

202

0

Proc. Natl. Acad. Sci. USA 89 (1992)

Complexes of a ribozyme and two different minizymes withvarious cleavable and noncleavable substrates were analyzedby electrophoresis on a 10%o polyacrylamide gel under non-denaturing conditions (Fig. 4). A single complex is formedbetween the minizyme M4t,DNA and its noncleavable sub-strate, as revealed by the strong band in lane 11, set B, of Fig.4. This complex moves in the gel as a species of =31 bp,slightly ahead of the 34-bp DNA marker in lane 1. [Thisestimate of size in bp of a DNARNA complex probablyrepresents an upper limit because mobilities of double-stranded RNA molecules and, probably, of double-strandedRNADNA molecules in nondenaturing gels are 10-20%olower than the corresponding mobilities ofduplex DNA (25)].A complex formed by one minizyme (34 nt) and one substratemolecule (21 nt) would contain 55 nt or, formally, 27.5 bp.Therefore, in this complex, there is just one minizymemolecule. The adjacent lane 10 contains the same minizymeM4t,DNA with its cleavable substrate. The strong band at '=31bp in lane 10 was shown to consist of minizyme, uncleavedsubstrate, and the 5'-product of cleavage (see Materials andMethods). Presumably, the minizyme complexed with un-cleaved substrate comigrates in the gel with the minizymeand its undissociated 5'- and 3'-products of'cleavage. Otherbands lower down lane 10 were shown to contain theminizyme plus the 5'-product, the minizyme plus (presum-ably) the 3'-product and, at the gel bottom, the 5'-productalone. There are no species of molecular mass > -31 bp ineither lane 10 or 11. Similar results were obtained for adifferent, 21-nt, cleavable substrate and its respective, 34-ntminizyme (compare lane 6, set A, with lane 10, set B, of Fig.4), indicating that these observations probably are general forminizymes. Therefore, all available evidence is consistentwith minizymes acting as monomers.

Other pertinent information can be drawn from Fig. 4. (i)A single complex, which moves in the gel as an -42-bpspecies, is formed by the ribozyme R4U,RNA (42 nt) and itsnoncleavable substrate (21 nt), as shown by the strong bandin lane 16, set C, of Fig. 4. A 1:1 complex of these moleculeswould contain 63 nt or, formally, 31.5 bp. The same "42-bpspecies is present in lane 15 of Fig. 4; it was shown to consistof ribozyme, uncleaved substrate, and the 5'-product ofcleavage, analogous to the minizyme-substrate complexes inlanes 6 and 10. In addition, lane 15 reveals that somehigher-polymeric species are present in very small amountsin the reaction mixture. All these species, which appearreproducibly, were shown to contain ribozyme, uncleavedsubstrate, and 5'-product of cleavage. These higher molec-ular-mass species depend on Mg2+ for stability because theyare absent in 10% polyacrylamide gels buffered by 90 mMTris borate/2 mM EDTA. Therefore, from the experimentson gel mobility, we conclude that this full-size ribozyme actspredominantly as a monomer in forming a complex with itssubstrate. (ii) The presence of uncleaved substrate com-plexed with minizyme and ribozyme in the reaction mixturessuggests that a proportion of the complexes might not be inthe active conformation for cleavage; this would account forthe observed extents of cleavage being <100o in Fig. 2 andTable 1. (iii) The minizyme-substrate complex migrates inthe gel as an =31-bp species, which is only slightly slowerthan would be expected were it a regular double helix of 27.5bp. Because the mobility of a molecule in a gel is determinedby its convex volume (the volume occupied by an imaginary,convex hull enveloping the molecule) (17), this suggests thathelices I and III in the complex are approximately paralleland that the conserved nucleotides protrude very little be-yond the surface of these helices.

Cleavage of Other Substrates. To see whether minizymeswere active generally against other RNA molecules, wesynthesized two more minizymes based on M4tDNA but with

appropriate changes to the sequence of the flanking arms.The new substrates were synthetic oligoribonucleotides of 21bases; one of these is the Krfippel substrate described in setA of Fig. 4, and the other, of sequence 5'-GCAUUUCAGU-CAGUUGCUCAA, is part of the gene for CAT. These shortsubstrates were cleaved by their respective minizymes at theexpected sites, as judged by counting bands on a denaturingpolyacrylamide gel that had the products of the cleavagereactions directly beside the products of alkaline hydrolysisofthe respective substrates. To test the activity ofminizymesagainst a longer RNA substrate, we made a 173-nt CATtranscript (5) containing the sequence of the shorter sub-strate, in which the expected cleavage site was 139 nt fromthe 5' end. The minizyme cleaved the transcript at theexpected site, as judged by sizes of the cleavage products.Rate constants and %6P. measured for these reactions understandard conditions are given in Table 1.

In summary, cleavage activity of a ribozyme of the Has-eloff-Gerlach type can be maintained despite size reductionand replacement of many RNA nucleotides by DNA. One ofthe three conserved double helices, helix II, is dispensable toformation of the active structure. The minimized ribozyme,or minizyme, is active as a monomer. We think that theseminizymes will be useful in future structural and functionalstudies of catalytic RNA. They should also prove useful ingene-control studies in living cells (8-11), where the DNAcomponent should make them more resistant to RNases.

We thank G. Both and T. Brown for help in oligonucleotidesyntheses, F. Santiago for technical assistance, and H. Drew forsuggestions and comments.

1. Cech, T. R. & Bass, B. L. (1986)Annu. Rev. Biochem. 55,599-629.2. Altman, S., Baer, M., Gold, H., Guerrier-Takada, C., Kirsebom,

L., Lawrence, N., Lumelsky, N. & Vioque, A. (1987) in MolecularBiology of RNA: New Perspectives, eds. Inouye, M. & Dudock,B. S. (Academic, London), pp. 3-15.

3. Forster, A. C. & Symons, R. H. (1987) Cell 5, 9-16.4. Uhlenbeck, 0. C. (1987) Nature (London) 328, 596-600.5. Haseloff, J. & Gerlach, W. L. (1988) Nature (London) 334,585-591.6. Jeffries, A. C. & Symons, R. H. (1989) Nucleic Acids Res. 17,

1371-1377.7. Koizumi, M., Iwai, S. & Ohtsuka, E. (1988) FEBS Lett. 239,

285-288.8. Cotten, M. & Birnstiel, M. L. (1989) EMBO J. 8, 3861-3866.9. Cameron, F. H. & Jennings, P. A. (1989) Proc. Natl. Acad. Sci.

USA 86, 9139-9143.10. Sarver, N., Cantin, E., Chang, P., Zaia, J., Ladna, P., Stephens, D.

& Rossi, J. (1990) Science 247, 1222-1225.11. Saxena, S. K. & Ackerman, E. J. (1990) J. Biol. Chem. 265,

17106-17109.12. Forster, A. C. & Symons, R. H. (1987) Cell 49, 211-220.13. Buzayan, J. M., Gerlach, W. L. & Bruening, G. (1986) Nature

(London) 323, 349-353.14. Prody, G. A., Bakos, J. T., Buzayan, J. M., Schneider, I. R. &

Bruening, G. (1986) Science 231, 1577-1580.15. Ruffner, D. E., Stormo, G. D. & Uhlenbeck, 0. C. (1990) Biochem-

istry 29, 10695-10702.16. Diekmann, S. & Wang, J. C. (1985) J. Mol. Biol. 186, 1-11.17. Calladine, C. R., Collis, C. M., Drew, H. R. & Mott, M. (1991) J.

Mol. Biol. 221, 981-1005.18. Perreault, J.-P., Wu, T., Cousineau, B., Ogilvie, K. K. & Ceder-

gren, R. (1990) Nature (London) 344, 565-567.19. Perreault, J.-P., Labuda, D., Usman, N., Yang, J.-H. & Cedergren,

R. (1991) Biochemistry 30, 4020-4025.20. Yang, J.-H., Perreault, J.-P., Labuda, D., Usman, N. & Cedergren,

R. (1990) Biochemistry 29, 11156-11160.21. Saenger, W. (1984) Principles of Nucleic Acid Structure (Springer,

New York).22. Goodchild, J. & Kohli, V. (1991) Arch. Biochem. Biophys. 284,

386-391.23. McCall, M. J., Brown, T. & Kennard, 0. (1985) J. Mol. Biol. 183,

385-3%.24. Forster, A. C., Davies, C., Sheldon, C. C., Jeifries, A. C. &

Symons, R. H. (1988) Nature (London) 334, 265-267.25. Gast, F.-U. & Hagerman, P. J. (1991) Biochemistry 30, 4268-4277.

5714 Biochemistry: McCall et al.

Dow

nloa

ded

by g

uest

on

Feb

ruar

y 4,

202

0