JOURNAL OF VIROLOGY, July 1999, p. 5381–5387 0022-538X/99/$04.0010 Copyright © 1999, American Society for Microbiology. All Rights Reserved.
Vol. 73, No. 7
Combinatorial Screening and Intracellular Antiviral Activity of Hairpin Ribozymes Directed against Hepatitis B Virus JASPER
ZU
PUTLITZ,1† QIAO YU,2‡ JOHN M. BURKE,2
AND
JACK R. WANDS1*
Molecular Hepatology Laboratory, Massachusetts General Hospital Cancer Center, and Harvard Medical School, Boston, Massachusetts 02129,1 and Department of Microbiology and Molecular Genetics, Markey Center for Molecular Genetics, The University of Vermont, Burlington, Vermont 054052 Received 1 December 1998/Accepted 13 April 1999
A combinatorial screening method has been used to identify hairpin ribozymes that inhibit hepatitis B virus (HBV) replication in transfected human hepatocellular carcinoma (HCC) cells. A hairpin ribozyme library (5 3 105 variants) containing a randomized substrate-binding domain was used to identify accessible target sites within 3.3 kb of full-length in vitro-transcribed HBV pregenomic RNA. Forty potential target sites were found within the HBV pregenomic RNA, and 17 sites conserved in all four subtypes of HBV were chosen for intracellular inhibition experiments. Polymerase II and III promoter expression constructs for corresponding hairpin ribozymes were generated and cotransfected into HCC cells together with a replication-competent dimer of HBV DNA. Four ribozymes inhibited HBV replication by 80, 69, 66, and 49%, respectively, while catalytically inactive mutant forms of these ribozymes affected HBV replication by 36, 28, 0, and 0%. These findings indicate that the inhibitory effects on HBV replication were largely mediated by the catalytic activity of the ribozymes. In conclusion, we have identified catalytically active RNAs by combinatorial screening that mediate intracellular antiviral effects on HBV. have been shown to inhibit replication of human immunodeficiency virus type 1 in cell culture (23, 32, 33). An essential prerequisite for antiviral approaches involving ribozymes is the need to identify accessible target sites on substrate RNA. While substrate consensus sequences for cleavage by hammerhead and hairpin ribozymes have been identified, the secondary and tertiary structures of the substrate and potential interactions with cellular factors in vivo may reduce the accessibility of these sequences for the ribozyme. Therefore, the selection of potential target sites on substrate RNA (7) solely on the basis of the primary sequence (sequence selection) often yields poor results with respect to the intracellular activity of corresponding ribozymes. In contrast, activity selection of ribozymes from a large ribozyme library with randomized substrate-binding sequences has the potential to identify ribozyme species with in vitro and in vivo cleavage activity against any given substrate RNA (19, 20, 35). With respect to ribozyme-mediated inhibition of gene expression in intact cells, we believe that activity-selected ribozymes are likely to be more effective than sequence-selected ribozymes. Since previous attempts to use hammerhead ribozymes for the intracellular inhibition of HBV have been largely unsuccessful (3, 30), we wished to investigate activity-selected hairpin ribozymes for the potential to intracellularly inhibit HBV replication. In the present study, a library of modified hairpin ribozymes with randomized substrate-binding domains was incubated with full-length in vitro-transcribed HBV pregenomic RNA. Primer extension analysis was used to identify several ribozyme target sites within the HBV pregenomic RNA. Ribozymes targeting the identified cleavage sites were then tested for intracellular activity by cotransfection of eukaryotic expression vectors carrying ribozyme expression cassettes together with a DNA construct that encodes a replication-competent genome of HBV. Several ribozymes were identified that markedly inhibit HBV replication, demonstrating that activity-based selection is a useful approach for the identification of hairpin ribozymes that are capable of mediating intracellular antiviral effects on HBV.
Hepatitis B virus (HBV) is a major cause of human viral hepatitis, and exposure to the virus often leads to persistent viral infection of the liver, cirrhosis, and hepatocellular carcinoma (HCC). HBV is a DNA virus which replicates asymmetrically through reverse transcription of an RNA intermediate. HBV has a partially double-stranded 3.2-kb DNA genome from which four major classes of transcripts are synthesized. The 3.5-kb pregenomic RNA serves as a template for reverse transcription and also encodes the nucleocapsid protein and the reverse transcriptase. A subclass of this transcript with a 59-end extension codes for the precore protein which, after processing, is secreted as HBV e antigen (HBeAg). The 2.4-kb RNA encompasses the pre-S1 open reading frame (ORF), which encodes the large surface protein. The 2.1-kb RNA encompasses the pre-S2 and S ORFs, which encode the middle and small surface proteins, respectively. The smallest transcript (approximately 0.8 kb) codes for the X protein. Ribozymes are naturally occurring enzymes comprised of RNA that catalyze RNA cleavage and splicing reactions (10). Several different ribozyme motifs with RNA cleavage activity have been discovered (27). Of these, the hammerhead ribozyme (29) is one of the smallest and has the simplest minimal target sequence requirements. Hammerhead and hairpin ribozymes have been studied extensively as experimental tools for trans suppression of gene expression and possible therapeutic applications (2). Ribozyme modifications designed to enhance catalytic activity, nuclease resistance, and intracellular efficiency may be useful for increasing intracellular and therapeutic activity (4, 13, 16–18, 24). Engineered hairpin ribozymes
* Corresponding author. Mailing address: The Liver Research Center, 55 Claverick St., 4th Floor, Providence, RI 02903. Phone: (401) 444-2795. Fax: (401) 444-2939. E-mail: Jack Wands
[email protected]. † Present address: Department of Internal Medicine, University of Freiburg, 79106 Freiburg, Germany. ‡ Present address: Genetic Therapy, Inc., Gaithersburg, MD 20878. 5381
5382
ZU
PUTLITZ ET AL. MATERIALS AND METHODS
Preparation of the randomized ribozyme pool and HBV pregenomic RNA. The pool of ribozymes containing 10 randomized positions in the substrate-binding domain was transcribed from a synthetic DNA template pool using bacteriophage T7 RNA polymerase (pol) as previously described (12, 21). The sequence of the DNA template pool is 59-TACCAGGTAATGTACCACGACTTACGTC GTGTGTTTCTCTGGTNNRCTTCNNNNNNNCCCTATAGTGAGTCGTAT TA-39, where N is any base, R is A or G, and the underlined sequence represents the bottom strand of the T7 promoter. All DNA oligodeoxynucleotides were synthesized by using standard solid-phase phosphoramidite chemistry on an Applied Biosystems model 392 oligonucleotide synthesizer. Randomized sites were generated by mixing equimolar amounts of all four nucleotides during synthesis. After annealing of the primer T7-Top (59-TAATACGACTCACTAT A-39) to the DNA template pool, RNA transcripts were produced by transcription of the DNA templates with T7 RNA polymerase and purified as previously described (8). The HBV adw2 pregenomic RNA (3.3 kb) was transcribed from linearized plasmid pSP6-HBV (unpublished data; kind gift of Stefan Wieland) by using SP6 RNA polymerase (MEGAscript SP6 Kit; Ambion Inc., Austin, Tex.). Cleavage reactions using the randomized ribozyme pool. The 3.3-kb HBV pregenomic RNA substrate and the randomized pool of ribozymes were preincubated separately at 37°C for 10 min either in reaction buffer (12 mM MgCl2, 50 mM Tris HCl, pH 8.0) to permit folding (14) or in control buffer (1 mM EDTA, 50 mM Tris HCl, pH 8.0) in which ribozymes are inactive. Substrate RNA (0.1 mM) was then incubated with the randomized ribozyme pool (20 mM) in reaction buffer or control buffer. Under the assumption of truly random DNA synthesis, the ribozyme pool contained each specific ribozyme molecule at a concentration of 0.04 nM. Cleavage and control reaction mixtures were incubated for 2 h at 37°C (final volume, 5 ml), subsequently desalted with a CentriSep column (Princeton Separations, Adelphia, N.J.), and dried in preparation for primer extension reactions. Primer extension mapping of accessible cleavage sites. Cleavage sites were identified by primer extension of cleavage products alongside sequencing ladders generated with avian myeloblastosis virus reverse transcriptase and 59-end-labeled DNA primers as previously described (9). Each primer generated a readable sequence 200 to 250 nucleotides in length. Therefore, 14 different primers were needed to accurately map all of the cleavage sites within the 3.3-kb HBV pregenomic RNA. The nucleotide positions of the complementary primers with respect to the in vitro-transcribed pregenomic RNA sequence were P1 (229 to 238), P2 (451 to 468), P3 (680 to 700), P4 (923 to 945), P5 (1179 to 1196), P6 (1415 to 1434), P7 (1661 to 1680), P8 (1891 to 1910), P9 (2133 to 2154), P10 (2374 to 2395), P11 (2610 to 2627), P12 (2845 to 2862), P13 (3083 to 3103), and P14 (3318 to 3340). The first nucleotide of the in vitro-transcribed HBV pregenomic RNA is nucleotide 1816, taking into account the fact that nucleotide 1 is the first T of the unique EcoRI site. Construction of ribozyme expression vectors. Ribozymes targeting 17 of the selected sites were designed in accordance with Watson-Crick base pair rules in helix 1 and helix 2 (see Fig. 1). The double-stranded DNA fragments corresponding to both active and inactive ribozyme sequences were generated by PCR using the Sindbis virus-specific ribozyme construct pTZ-U6-8242 (26) as a template. PCR products were purified, digested with SalI and XbaI, and inserted into the SalI/XbaI-digested pTZ-U6 vector (15). In addition, the PCR products were also inserted into the XhoI/XbaI-digested pCI-neo vector (Promega, Madison, Wis.). pCI-neo contains a chimeric intron composed of the 59 donor site from the first intron of the human b-globin gene and the branch and 39 acceptor site from the intron of an immunoglobulin gene heavy-chain variable region. In addition, pCI-neo contains the simian virus 40 late polyadenylation signal. The sequences of DNA primers for the most active ribozyme constructs targeting HBV pregenomic RNA sites 1401, 1626, 1781, and 1976 are as follows: 1401-FP, 59-ATACTAGTCGACG AATTCTGAAGCATACCAGAGAAACAGATCTC; 1401-ina-FP, 59-ATACTA GTCGACGAATTCTaAAGCATACCAGAtAAACAGATCTC; 1626-FP, 59-AT ACTAGTCGACATTCTTAGAAACAAACCAGAGAAACAGATCTC; 1626ina-FP, 59-ATACTAGTCGACATTCTTAaAAACAAACCAGAtAAACAGATC TC; 1781-FP, 59-ATACTAGTCGACGACACACGAAGCGAACCAGAGAAAC AGATCTC; 1781-ina-FP, 59-ATACTAGTCGACGACACACaAAGCGAACCA GAtAAACAGATCTC; 1976-FP, 59-ATACTAGTCGACGTTTTGCGAAGCA AACCAGAGAAACAGATCTC; 1976-ina-FP, 59-ATACTAGTCGACGTTTTG CaAAGCAAACCAGAtAAACAGATCTC; RP, 59-CCGCTCTAGACCAGGTAA TG. FP indicates a forward primer, RP indicates a reverse primer, and ina indicates an inactive ribozyme. Underlined bases are restriction enzyme sites, boldface bases are the substrate-binding region of the hairpin ribozyme, and lowercase bases are mutations that abolish the catalytic activity of the hairpin ribozyme (6). Cells and transfections. The human HCC cell line HuH-7 (22) was grown in modified Eagle minimal essential medium (Cellgro Mediatech, Washington, D.C.) supplemented with 10% fetal calf serum, 1% nonessential amino acid solution (Life Technologies, Gaithersburg, Md.), and a 1% penicillin-streptomycin stock solution (5,000 U of penicillin G sodium per ml, 5,000 mg of streptomycin per ml; Cellgro Mediatech). Transfections were performed by using a modified calcium phosphate precipitation protocol (11), routinely using 10 mg each of DNA plasmids plus 1 mg of reporter plasmid pTKGH encoding cDNA for human growth hormone (hGH) (25) per 100-mm-diameter plate. DNA analysis. For the preparation of core-associated HBV DNA, transfected cells were lysed in 50 mM Tris-HCl (pH 8.0)–1 mM EDTA–1% Nonidet P-40.
J. VIROL. The lysate was centrifuged at 10,000 3 g for 5 min at room temperature. After the addition of CaCl2 and MgCl2 to a final concentration of 10 mM each, the supernatant was incubated with 20 U of DNase I (Boehringer Mannheim, Indianapolis, Ind.) per ml and micrococcal nuclease (final concentration, 150 U/ml; Pharmacia Biotech) for 2 h at 37°C. Next, EDTA (final concentration, 20 mM), 10% sodium dodecyl sulfate (SDS; final concentration, 1%), and proteinase K (final concentration, 1 mg/ml; Promega) were added and the mixture was incubated for 12 to 16 h at 37°C. Finally, the sample was extracted once with 1 volume of phenol-chloroform. After addition of 3 M sodium acetate, pH 5.5, the DNA was precipitated with 1 volume of isopropanol. The pellet was washed with 80% ethanol, vacuum dried, and resuspended in 20 ml of agarose gel loading buffer. DNA was fractionated by 1.2% agarose gel electrophoresis in Tris-acetate buffer (1). Nucleic acids were transferred to Hybond-N1 nylon membranes (Amersham Life Science, Arlington Heights, Ill.). After UV cross-linking of nucleic acids, membranes were prehybridized for 4 h at 42°C in 50% (vol/vol) formamide–53 SSPE (13 SSPE is 0.15 M NaCl, 0.01 M sodium dihydrogen phosphate, and 1 mM EDTA)–2.53 Denhardt’s solution (13 Denhardt’s solution is 0.02% polyvinylpyrrolidone, 0.02% Ficoll, and 0.02% bovine serum albumin)– 0.1% SDS–200 mg of denatured calf thymus DNA per ml. Hybridization with 32 P-labeled recombinant full-length HBV DNA was performed by incubation in the above-described buffer for 16 h at 42°C. After hybridization, membranes were washed once in 53 SSC–(13 SSC is 0.15 M NaCl plpus 0.015 M sodium citrate)–0.1% SDS for 5 min at 42°C and once in 13 SSC–0.1% SDS for 20 min at 65°C. Membranes were exposed to X-ray film at 280°C. RNA analysis. Total cellular RNA from cells cotransfected with the HBV head-to-tail dimer (HTD) and various ribozyme expression constructs was extracted with the Ultraspec RNA isolation system (Biotecx Laboratories, Houston, Tex.). Transcripts corresponding to U6 and cytomegalovirus (CMV) ribozymes were detected by primer extension with primers pA-U6 (59-ACCAGGT AATGTACGATC) and pB-CMV (59-GTCAGAAGCACTGACTGCG), respectively. HBV pregenomic RNA was detected by primer extension with the primer PREG-PE (59-AAACGAGAGTAACTCCACAG). Reactions also contained the primer pC-U6 (59-GGCCATGCTAATCTTCTCTG) that annealed to positions 42 to 61 of the endogenous U6 small nuclear snRNA and yielded signals corresponding to the U6 snRNA used as an internal control. HBV and ribozyme transcript copy numbers were calculated by radiodensitometry (Bio-Rad) and use of endogenous U6-specific signals as an internal reference. Enzymatic and immunological assays. For the determination of transfection efficiencies, a commercially available assay for secreted hGH (Tandem-R HGH; Hybritech, San Diego, Calif.) was used. HBV surface antigen (HBsAg) was determined by a commercially available radioimmunoassay (AUSRIA II; Abbott Laboratories, North Chicago, Ill.). HBV e antigen (HBeAg) was determined by a commercially available radioimmunoassay (Incstar Corp., Stillwater, Minn.) which does not exhibit cross-reactivity with HBV core antigen.
RESULTS Generation of a hairpin ribozyme library with randomized substrate-binding domains. A hairpin ribozyme library was generated by in vitro transcription of synthetic DNA templates. Nine positions within the substrate-binding domain were randomized, and one was fixed as a pyrimidine (Fig. 1A), while the bases which are important for catalytic activity (G8, A9, A10, and G11) remained fixed (Fig. 1A). The resulting library of hairpin ribozymes was expected to contain 5 3 105 different members. For cleavage assays, the quantity of ribozymes used (100 pmol; 6 3 1013 molecules) greatly exceeded the sequence complexity of the pool. Therefore, all potential sequences within the randomized pool were likely to be present in multiple copies in the assay. Cleavage of HBV pregenomic RNA by the ribozyme library in vitro. In vitro-transcribed HBV pregenomic RNA (0.1 mM) was incubated at 37°C for 2 h with the randomized pool of hairpin ribozymes (20 mM), either in the presence or in the absence of MgCl2. Because the hairpin ribozyme library contained 5 3 105 different members, the concentration of each individual ribozyme species was calculated to be 0.04 nM, assuming truly random DNA synthesis. Therefore, the HBV substrate RNA was calculated to be present in approximately 2,500-fold molar excess over each individual ribozyme species. Because most substrate RNA molecules remained uncleaved at the end of the digestion, most products resulted from a single cleavage event within target RNA molecules, not from multiple cleavage events within the same target.
VOL. 73, 1999
IN VITRO-SELECTED RIBOZYMES BLOCK HBV REPLICATION
5383
FIG. 1. Structure of the hairpin ribozyme (A) and in vitro site screening (B). (A) The modified hairpin ribozyme used in the present study consists of a 57-nucleotide RNA molecule which binds to and cleaves an RNA substrate (top strand). The catalytic RNA folds into a two-dimensional structure that resembles a hairpin consisting of helices 3 and 4 and internal loop B. In addition, helices 1 and 2 and internal loop A form between the ribozyme and its substrate. Recognition of the substrate by the ribozyme utilizes Watson-Crick base pairing within helices 1 and 2 (N, any nucleotide; R, A or G; Y, T or C). Cleavage occurs immediately 59 of the substrate G in loop A, as indicated by an arrow. The catalytic activity, but not substrate-binding activity, of the hairpin ribozyme is abolished by mutating G8 and G21 to A and U, respectively (6). (B) In vitro combinatorial activity-based screening. A large library (nine ribozyme positions are totally randomized [N], and one position is randomized between T and C [Y], yielding a complexity of approximately 5 3 105 members) of in vitro-generated hairpin ribozymes (Rz) was incubated with HBV pregenomic RNA as the substrate. Control reactions were performed in the absence of the ribozyme library and MgCl2. After cleavage occurred, primer extension assays using various primers (P1, P2, and P3) were performed to map the 59 ends of HBV pregenomic RNA cleavage products. Extension products were separated on sequencing gels together with RNA sequencing reactions. Criteria for the identification of extension products corresponding to authentic ribozyme cleavage sites were that the product be detectable only in the presence of both the ribozyme pool and MgCl2 and have a guanosine residue at its 59 end.
Mapping of hairpin ribozyme cleavage sites on HBV pregenomic RNA. Primer extension assays (Fig. 1B; see also Materials and Methods) were carried out to map 59 ends of HBV pregenomic RNA cleavage products that occurred after incubation with the hairpin ribozyme library. Extension products were separated on sequencing gels together with sequencing reaction mixtures (Fig. 2). With this method, sites on the HBV pregenomic RNA that were accessible for hairpin ribozymes and subsequent cleavage could be accurately mapped. As an example, Fig. 2, lane 3, illustrates in vitro cleavage of HBV pregenomic RNA target sites 1401, 1626, 1781, and 1976 by the randomized pool of hairpin ribozymes (the first nucleotide of the in vitro-transcribed HBV pregenomic RNA is nucleotide 1816, taking into account the fact that nucleotide 1 is the first T of the unique EcoRI site). Control reactions were performed in the absence of ribo-
zyme (lane 1) or MgCl2 (lane 2). Although specific cleavage products were readily detectable (compare lane 3 with lanes 1 and 2, respectively), it is important to note that the folded structure of the RNA template caused nonspecific terminations of the extension reaction that appeared as background bands. Therefore, criteria for the identification of extension products corresponding to real ribozyme cleavage sites were that the product be detectable only in the presence of both the ribozyme pool and MgCl2 and also have a guanosine residue at its 59 end. This guanosine residue represents the essential nucleotide located immediately 39 from the cleavage site within the consensus substrate cleavage sequence that has been described for hairpin ribozymes (12). Primer extension reactions covering the entire HBV pregenomic RNA were performed by using 14 different primers. In total, 40 potential cleavage sites within the HBV prege-
5384
ZU
PUTLITZ ET AL.
FIG. 2. In vitro cleavage of HBV pregenomic RNA target sites 1401, 1626, 1781, and 1976 by the randomized pool of hairpin ribozymes. Sites are numbered according to the position of the G located immediately 39 to the cleavage site in the in vitro-transcribed HBV pregenomic RNA sequence. Control reactions were performed in the absence of ribozyme (Rz) (lane 1) or MgCl2 (Mg11) (lane 2). Specific cleavage products (compare lane 3 with lanes 1 and 2, respectively) were readily detectable (arrows) and had a guanosine at the 59 terminus. Note that the folded structure of the RNA template caused nonspecific terminations of the extension reaction that appeared as background bands in lanes 1, 2, and 3. Lanes 4 to 7 contained parallel RNA sequencing reactions with the same primers used for primer extension that enabled the accurate mapping of cleavage sites.
nomic RNA were identified (data not shown). Sites were numbered according to the position of the G located immediately 39 to the cleavage site in the in vitro-transcribed HBV pregenomic RNA sequence. Design and construction of hairpin ribozymes for selected targets. Of the 40 cleavage sites on the HBV pregenomic RNA that were initially identified, 17 sites whose sequence was conserved among all HBV subtypes were chosen for the construction of corresponding hairpin ribozymes and viral inhibition assays. The locations of these 17 sites on the pregenomic RNA and their positions with respect to HBV subgenomic transcripts and ORFs are illustrated in Fig. 3. Cleavage sites were clustered between nucleotides 520 and 600, 1400 and 2100, and 3000 and 3200. Two single sites were localized at nucleotide positions 122 and 2595. Interestingly, no sites were identified between nucleotides 612 and 1256, suggesting that this region of the pregenomic RNA was not accessible for cleavage by hairpin ribozymes in vitro. To design hairpin ribozymes that cleaved the selected sites, the substrate-binding sequences of the ribozymes were in-
J. VIROL.
ferred from the HBV sequence surrounding the cleavage site. Sequences with Watson-Crick complementarity to the HBV substrate were introduced into helices 1 and 2 (compare Fig. 1A), with the exception of G11, which remained fixed (17). Catalytically inactive hairpin ribozyme variants were designed by replacing G8 with A and G21 with U. These base substitutions prevent ribozyme-catalyzed cleavage but still permit formation of the ribozyme substrate complex (6). All other hairpin ribozyme sequences were designed in accordance with established guidelines (34). PCR-generated sequences encoding the catalytically active hairpin ribozymes and their inactive counterparts were cloned into two different expression vectors. For pol III promoterdirected expression of hairpin ribozymes in transfected cells, ribozyme coding sequences were inserted into the U6 snRNA promoter-bearing expression vector pTZ-U6 (15). This vector provides short RNA stem-loop structures at both the 59 and 39 ends of the ribozyme that serve to stabilize correctly folded hairpin ribozymes and partially protect these molecules from exoribonuclease attack. Ribozymes synthesized from this expression construct were expected to be localized in the nucleus and cytoplasm of transfected cells. For pol II promoter-directed expression of hairpin ribozymes, the ribozyme coding sequences were inserted into the CMV immediate-early promoter-bearing expression vector pCI-neo. Transcripts containing hairpin ribozyme sequences that are generated from this vector were expected to be capped, spliced, polyadenylated, and translocated into the cytoplasm of transfected cells. Expression of ribozymes in cells. Ribozyme expression vectors were cotransfected into HuH-7 HCC cells together with a replication-competent HBV HTD. Two days after transfection, total cellular RNA was prepared. Endogenously synthesized U6 snRNA and ribozymes transcribed from U6 snRNA and CMV promoters, as well as HBV pregenomic RNA (data not shown), were detected by primer extension (Fig. 4). As demonstrated in Fig. 4, lanes 1 to 6, specific extension products corresponding to U6 ribozymes 1401, 1781, and 1976 and their catalytically inactive variants were detectable in transfected cells. The CMV ribozyme 1626 and its inactive variant were also readily detectable (Fig. 4, lanes 9 and 10). Inhibition of HBV replication in cells. Table 1 summarizes results from transient-cotransfection experiments with an HBV HTD and all 17 ribozyme expression constructs or the vector (pTZ-U6 or pCI-neo) in HuH-7 HCC cells. All cotransfections were performed at least three times, and plate-to-plate variations in transfection efficiency were determined by analyzing hGH levels in cell culture supernatants. Nucleocapsid-associated HBV replication products were detected on Southern blots with an HBV-specific probe and quantified by radiodensitometry. The inhibition percentage was calculated as follows: 1 2 ([SignalRz/Signalwt]/[hGHRz/hGHwt]) 3 100. No inhibition of HBV replication was observed when a series of ribozymes directed against an unrelated virus (Sindbis virus) were expressed from a U6 snRNA or CMV promoter (data not shown). Most of the U6 and CMV ribozymes inhibited HBV replication to some degree. Figure 5 summarizes the results obtained with ribozymes U6-1401, U6-1781, U6-1976, and CMV-1626 and their catalytically inactive controls (designated ina). Ribozyme U6-1401 inhibited HBV replication by 69% (lane 1), whereas U6-1401-ina affected HBV replication by 28% (lane 2). Inhibition values for the other ribozymes were as follows: U6-1781, 49% (lane 4); U6-1781-ina, 0% (lane 5); U6-1976, 80% (lane 7); U6-1976-ina, 36% (lane 8); CMV-1626, 66% (lane 10); CMV-1626-ina, 0% (lane 11). Thus, the ribozyme with the highest inhibitory potential against HBV replication was U6-1976, with 80% inhibition. To assess the specificity of
VOL. 73, 1999
IN VITRO-SELECTED RIBOZYMES BLOCK HBV REPLICATION
5385
FIG. 3. Locations of conserved cleavage sites for hairpin ribozymes within the HBV pregenomic (3.5-kb) RNA. The first nucleotide of the in vitro-transcribed HBV pregenomic RNA is located 5 nucleotides upstream from the naturally occurring transcriptional start point. The positions of the encapsidation signal (ε) and direct repeats DR1 and DR2 are shown. HBV subgenomic transcripts 2.4, 2.1, and 0.9 kb length are depicted by lines below the pregenomic RNA. The core (C); pol; large (preS1), middle (preS2), and small (S) surface antigen; and X ORFs are depicted as open arrows. The identified hairpin ribozyme cleavage sites were clustered between nucleotides 520 and 600, 1400 and 2100, and 3000 and 3200. Two single sites were localized at nucleotide positions 122 and 2595. pA, polyadenylation site.
the inhibitory effect of the four best, they were also tested for the ability to inhibit duck HBV in LMH chicken hepatoma cells. None of the four ribozymes affected duck HBV replication (data not shown). In conclusion, several of the in vitroselected ribozymes had inhibitory effects on HBV replication in transfected cells, and three ribozymes inhibited HBV replication by more than 50%. Effects of ribozymes on HBV antigen levels in cell culture supernatants. To assess the effects of the most effective ribozymes on HBsAg and HBeAg synthesis and secretion, cell culture supernatants from cotransfections with an HBV HTD and ribozymes U6-1401, U6-1781, U6-1976, and CMV-1626 were analyzed for HBsAg and HBeAg levels (Fig. 6). After correction for variations in transfection efficiency by the hGH
assay, the normalized antigen (Agn) levels were used to monitor the decrease in antigen expression relative to the control (1 2 [ribozyme Agn/control Agn] 3 100). Thus, variations in antigen levels reflecting antigen secretion were correlated to variations in transfection efficiency, and inhibition values are expressed in percent. As demonstrated in Fig. 6, all four ribozymes reduced extracellular HBeAg levels. The reductions were as follows: U6-1401, 62%; U6-1781, 38%; U6-1976, 64%; CMV-1626, 42%. Since all four ribozymes targeted sequences that were also present on HBV subgenomic transcripts coding for HBsAgs, extracellular HBsAg levels were determined. As demonstrated in Fig. 6, the effects of ribozymes on HBsAg levels were generally weaker. The reductions were as follows: U6-1401, 38%; U6-1781, 34%; U6-1976, 5%; CMV-1626, 12%. In conclusion, the ribozymes with the greatest potential to inhibit HBV replication also had significant effects on extraTABLE 1. Inhibition of HBV replication by selected ribozymes Cleavage site
FIG. 4. Expression of ribozymes and catalytically inactive variants in HuH-7 cells. Primer extension analysis of total cellular RNA from cells cotransfected with an HBV HTD and ribozyme expression constructs U6-1401, U6-1781, U6-1976, and CMV-1626, as well as catalytically inactive variants. HBV pregenomic RNA was detected by primer extension (data not shown). Transcripts corresponding to U6 and CMV ribozymes were detected by using primers pA-U6 (59-ACCAGGTAATGTACGATC) and pB-CMV (59-GTCAGAAGCACTGAC TGCG), respectively. Reaction mixtures also contained primer pC-U6 (59-GG CCATGCTAATCTTCTCTG), which yielded signals corresponding to endogenous U6 snRNA. Note that all of the ribozymes and their catalytically inactive variants were expressed (lanes 1 to 6 and 9 to 10). NC, negative control.
122 527 564 585 604 1401 1626 1650 1781 1824 1976 2113 2595 3026 3098 3123 3189 a
Inhibition (%)a U6 Rz
CMV Rz
23 13 39 45 19 69 35 43 49 5 80 18 0 18 44 44 45
21 18 22 21 25 19 66 19 5 0 10 0 0 0 0 0 36
Results of individual experiments varied by 6 10%. Rz, ribozyme.
5386
ZU
PUTLITZ ET AL.
J. VIROL.
that the in vitro secondary and tertiary structures of the HBV pregenomic RNA in this region prevent ribozymes from binding to their substrate RNA and subsequent cleavage. Hairpin ribozymes corresponding to the 17 conserved cleavage sites were designed and expressed in cells together with a replication-competent genome of HBV. While several ribozymes showed only modest inhibition of HBV replication, four ribozymes (U6-1401, U6-1781, U6-1976, and CMV-1626) markedly inhibited HBV replication and decreased extracellular HBeAg levels, whereas HBsAg levels were less affected. Catalytically inactive variants of these ribozymes had much weaker effects, suggesting that the majority of the inhibitory effects on HBV replication were mediated by the catalytic activity of the ribozymes rather than by an antisense mechanism. Therefore, we have identified hairpin ribozymes by in vitro selection that effectively inhibit HBV replication, presumably by cleavage of HBV pregenomic RNA under physiological conditions within the complex cellular environment. Another study has described hairpin ribozyme genes that were cotransfected into HuH-7 HCC cells together with small amounts of an HTD of HBV (31). Under these conditions, the virus particle-associated HBV levels, as determined by the endogenous pol assay, were reduced by up to 83%. However, this study did not correlate inhibition levels with plate-to-plate variations in transfection efficiency, and it is therefore unclear to which extent HBV replication was truly inhibited. Interestingly, the two hairpin ribozymes used in this study targeted cleavage sites on HBV pregenomic RNA at nucleotide positions 1703 and 2938 that were not identified in our experiments. However, in vitro cleavage was demonstrated by using FIG. 5. Intracellular inhibition of HBV replication by four hairpin ribozymes. Southern blot analysis of nucleocapsid-associated HBV DNA isolated from HuH-7 cells cotransfected with an HTD of HBV and ribozyme expression constructs U6-1401, U6-1781, U6-1976, and CMV-1626, as well as their catalytically inactive variants (designated ina). hGH levels for the assessment of plateto-plate variations in transfection efficiency and inhibition percentages calculated from radiodensitometry are depicted below each panel. Ribozyme U6-1401 inhibited HBV replication by 69% (lane 1), whereas U6-1401-ina affected HBV replication by 28% (lane 2). Inhibition values for the other ribozymes were as follows: U6-1781, 49% (lane 4); U6-1781-ina, 0% (lane 5); U6-1976, 80% (lane 7); U6-1976-ina, 36% (lane 8); CMV-1626, 66% (lane 10); CMV-1626-ina, 0% (lane 11). Results are representative of at least three independent experiments. DS, position of HBV double-stranded linear DNA.
cellular HBeAg levels. Extracellular HBsAg levels were affected to a lesser degree. This antigen is translated from the subgenomic 2.1-kb transcript, which was not used as target in the in vitro selection assay. It is therefore possible that the selected ribozymes did not efficiently anneal to this transcript, resulting in smaller reductions of HBsAg versus HBeAg levels. DISCUSSION In this study, we have used a combinatorial screening method, employing a complex library of hairpin ribozymes with all possible specificities, to identify conserved sites within the full-length HBV pregenomic RNA that could be effectively cleaved by corresponding hairpin ribozymes. A total of 40 cleavage sites were identified. Seventeen sites were conserved among all HBV subtypes and were subjected to further analysis in transfected HuH-7 HCC cells. Interestingly, most of the conserved cleavage sites were positioned in three distinct clusters on the HBV pregenomic RNA. However, a large segment of the pregenomic RNA, from nucleotide position 612 to position 1256, was entirely devoid of cleavage sites. It is possible
FIG. 6. Effects of four ribozymes on HBsAg and HBeAg levels in cell culture supernatants. Cell culture supernatants from cotransfections with an HBV HTD and ribozymes U6-1401, U6-1781, U6-1976, and CMV-1626 were analyzed for HBsAg and HBeAg levels. Variations in transfection efficiency were assessed by hGH assay. After correction for variations in transfection efficiency by the hGH assay, normalized antigen levels (Agn) were used to monitor the decrease of antigen expression relative to the control (1 2 [ribozyme Agn/control Agn] 3 100). Results were expressed as percent inhibition compared with the wild type (cotransfection of vector). All four ribozymes reduced extracellular HBeAg levels. The reductions were as follows: U6-1401, 62%; U6-1781, 38%; U6-1976, 64%; CMV-1626, 42%. Since all four ribozymes also targeted sequences that were present on HBV subgenomic transcripts coding for HBsAgs, extracellular HBsAg levels were determined. The effects of the four ribozymes on HBsAg levels were weaker. The reductions were as follows: U6-1401, 38%; U6-1781, 34%; U6-1976, 5%; CMV-1626, 12%.
VOL. 73, 1999
IN VITRO-SELECTED RIBOZYMES BLOCK HBV REPLICATION
short (ca. 40 nucleotides) substrates, not full-length HBV pregenomic RNA. Two studies have demonstrated that hammerhead ribozymes can specifically cleave HBV RNA in a cell-free system (30) and cell extracts (3). The first study used three hammerhead ribozymes transcribed from a single DNA template that were directed against three adjacent sites within the HBV pregenomic RNA (30). It was demonstrated that all three ribozymes cleaved HBV substrate RNA and that cleavage efficiency was similar to that of single-ribozyme constructs. The second study demonstrated that efficient hammerhead ribozyme-mediated cleavage of the viral encapsidation signal could be achieved in vitro and in cell extracts but not in intact cells (3). Cleavage was most efficient after treatment of cotransfected cells with proteinase K, extraction with phenol, and supplementation of the extract with Mg21 indicating that, in this case, cellular proteins and low Mg21 concentrations limited intracellular ribozyme activity. However, cellular proteins may also have positive effects on hammerhead ribozyme-mediated catalysis (5, 28). Because of its complex structure and its nature as a binding site for cellular and/or viral proteins, the viral encapsidation signal may not be the optimal target for ribozyme attack, and more accessible sites on the HBV pregenomic RNA might be identified that are susceptible to hammerhead ribozyme-mediated cleavage in intact cells. Our study demonstrates that in vitro selection yields hairpin ribozymes that can effectively reduce HBV replication in transfected cells. One of the major challenges of gene therapy for HBV is the delivery of therapeutic genes to infected cells. In the future, recombinant retroviruses, adenoviruses, or adenoassociated viruses containing therapeutic ribozyme genes will have to be generated to allow efficient delivery of the therapeutic transgenes into cells. Thus, the full potential of a molecular therapeutic approach involving hairpin ribozymes directed against HBV may be realized. ACKNOWLEDGMENTS J.Z. and Q.Y. contributed equally to the studies described in this publication. We thank Brennan Martin for help with the in vitro cleavage site screening work, Attila Seyhan for providing primers pA-U6 and pCU6, and Stefan Wieland for the construct pSP6-HBV. This work was supported by grants CA-35711 (J.R.W.), AI30534 (J.M.B.), and AA-02169 (J.R.W.) from the National Institutes of Health. J.Z. was on leave from the Department of Internal Medicine II, University of Freiburg, Freiburg, Germany, and is supported by the Stipendienprogramm “Infektionsforschung” of the German Cancer Research Center, Heidelberg, Germany. REFERENCES 1. Ausubel, F. M., R. Brent, R. E. Kingston, D. D. Moore, J. G. Seidman, J. A. Smith, and K. Struhl. 1987. Current protocols in molecular biology. John Wiley & Sons, Inc., New York, N.Y. 2. Bartolome, J., A. Madejon, and V. Carreno. 1995. Ribozymes: structure, characteristics and use as potential antiviral agents [see comments]. J. Hepatol. 22:57–64. 3. Beck, J., and M. Nassal. 1995. Efficient hammerhead ribozyme-mediated cleavage of the structured hepatitis B virus encapsidation signal in vitro and in cell extracts, but not in intact cells. Nucleic Acids Res. 23:4954–4962. 4. Beigelman, L., J. A. McSwiggen, K. G. Draper, C. Gonzalez, K. Jensen, A. M. Karpeisky, A. S. Modak, J. Matulic-Adamic, A. B. DiRenzo, P. Haeberli, et al. 1995. Chemical modification of hammerhead ribozymes. Catalytic activity and nuclease resistance. J. Biol. Chem. 270:25702–25708. 5. Bertrand, E. L., and J. J. Rossi. 1994. Facilitation of hammerhead ribozyme catalysis by the nucleocapsid protein of HIV-1 and the heterogeneous nuclear ribonucleoprotein A1. EMBO J. 13:2904–2912. 6. Berzal-Herranz, A., S. Joseph, B. M. Chowrira, S. E. Butcher, and J. M.
7. 8. 9. 10. 11. 12. 13. 14. 15. 16.
17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32.
33. 34. 35.
5387
Burke. 1993. Essential nucleotide sequences and secondary structure elements of the hairpin ribozyme. EMBO J. 12:2567–2573. Birikh, K. R., Y. A. Berlin, H. Soreq, and F. Eckstein. 1997. Probing accessible sites for ribozymes on human acetylcholinesterase RNA. RNA 3:429– 437. Butcher, S. E., and J. M. Burke. 1994. A photo-cross-linkable tertiary structure motif found in functionally distinct RNA molecules is essential for catalytic function of the hairpin ribozyme. Biochemistry 33:992–999. Butcher, S. E., and J. M. Burke. 1994. Structure-mapping of the hairpin ribozyme. Magnesium-dependent folding and evidence for tertiary interactions within the ribozyme-substrate complex. J. Mol. Biol. 244:52–63. Cech, T. R. 1987. The chemistry of self-splicing RNA and RNA enzymes. Science 236:1532–1539. Chen, C., and H. Okayama. 1987. High-efficiency transformation of mammalian cells by plasmid DNA. Mol. Cell. Biol. 7:2745–2752. Chowrira, B. M., A. Berzal-Herranz, and J. M. Burke. 1991. Novel guanosine requirement for catalysis by the hairpin ribozyme. Nature 354:320–322. Chowrira, B. M., and J. M. Burke. 1992. Extensive phosphorothioate substitution yields highly active and nuclease-resistant hairpin ribozymes. Nucleic Acids Res. 20:2835–2840. Esteban, J. A., A. R. Banerjee, and J. M. Burke. 1997. Kinetic mechanism of the hairpin ribozyme. Identification and characterization of two nonexchangeable conformations. J. Biol. Chem. 272:13629–13639. Good, P. D., A. J. Krikos, S. X. Li, E. Bertrand, N. S. Lee, L. Giver, A. Ellington, J. A. Zaia, J. J. Rossi, and D. R. Engelke. 1997. Expression of small, therapeutic RNAs in human cell nuclei. Gene Ther. 4:45–54. Jarvis, T. C., F. E. Wincott, L. J. Alby, J. A. McSwiggen, L. Beigelman, J. Gustofson, A. DiRenzo, K. Levy, M. Arthur, J. Matulic-Adamic, A. Karpeisky, C. Gonzalez, T. M. Woolf, N. Usman, and D. T. Stinchcomb. 1996. Optimizing the cell efficacy of synthetic ribozymes. Site selection and chemical modifications of ribozymes targeting the proto-oncogene c-myb. J. Biol. Chem. 271:29107–29112. Joseph, S., and J. M. Burke. 1993. Optimization of an anti-HIV hairpin ribozyme by in vitro selection. J. Biol. Chem. 268:24515–24518. Joyce, G. F. 1998. Nucleic acid enzymes: playing with a fuller deck. Proc. Natl. Acad. Sci. USA 95:5845–5847. Lieber, A., and M. A. Kay. 1996. Adenovirus-mediated expression of ribozymes in mice. J. Virol. 70:3153–3158. Lieber, A., and M. Strauss. 1995. Selection of efficient cleavage sites in target RNAs by using a ribozyme expression library. Mol. Cell. Biol. 15:540–551. Milligan, J. F., and O. C. Uhlenbeck. 1989. Synthesis of small RNAs using T7 RNA polymerase. Methods Enzymol. 180:51–62. Nakabayashi, H., K. Taketa, K. Miyano, T. Yamane, and J. Sato. 1982. Growth of human hepatoma cells lines with differentiated functions in chemically defined medium. Cancer Res. 42:3858–3863. Ojwang, J. O., A. Hampel, D. J. Looney, F. Wong-Staal, and J. Rappaport. 1992. Inhibition of human immunodeficiency virus type 1 expression by a hairpin ribozyme. Proc. Natl. Acad. Sci. USA 89:10802–10806. Sargueil, B., D. B. Pecchia, and J. M. Burke. 1995. An improved version of the hairpin ribozyme functions as a ribonucleoprotein complex. Biochemistry 34:7739–7748. Selden, R. F., K. B. Howie, M. E. Rowe, H. M. Goodman, and D. D. Moore. 1986. Human growth hormone as a reporter gene in regulation studies employing transient gene expression. Mol. Cell. Biol. 6:3173–3179. Seyhan, A. A., J. Amaral, and J. M. Burke. 1998. Intracellular RNA cleavage by the hairpin ribozyme. Nucleic Acids Res. 26:3494–3504. Symons, R. H. 1992. Small catalytic RNAs. Annu. Rev. Biochem. 61:641–71. Tsuchihashi, Z., M. Khosla, and D. Herschlag. 1993. Protein enhancement of hammerhead ribozyme catalysis. Science 262:99–102. Uhlenbeck, O. C. 1987. A small catalytic oligoribonucleotide. Nature 328: 596–600. von Weizsacker, F., H. E. Blum, and J. R. Wands. 1992. Cleavage of hepatitis B virus RNA by three ribozymes transcribed from a single DNA template. Biochem. Biophys. Res. Commun. 189:743–748. Welch, P. J., R. Tritz, S. Yei, J. Barber, and M. Yu. 1997. Intracellular application of hairpin ribozyme genes against hepatitis B virus. Gene Ther. 4:736–743. Yu, M., J. Ojwang, O. Yamada, A. Hampel, J. Rapapport, D. Looney, and F. Wong-Staal. 1993. A hairpin ribozyme inhibits expression of diverse strains of human immunodeficiency virus type 1. Proc. Natl. Acad. Sci. USA 90: 6340–6344. (Erratum, 90:8303.) Yu, M., E. Poeschla, and F. Wong-Staal. 1994. Progress towards gene therapy for HIV infection. Gene Ther. 1:13–26. Yu, Q., and J. M. Burke. 1997. Design of hairpin ribozymes for in vitro and cellular applications. Methods Mol. Biol. 74:161–169. Yu, Q., D. B. Pecchia, S. L. Kingsley, J. E. Heckman, and J. M. Burke. 1998. Cleavage of highly structured viral RNA molecules by combinatorial libraries of hairpin ribozymes. The most effective ribozymes are not predicted by substrate selection rules. J. Biol. Chem. 273:23524–23533.
