J. Mol. Biol. (1992) 223, 233-245

Sequence and Structure of the Catalytic RNA of Hepatitis Delta Virus Genomic RNA Huey-Nan

Wul, Yueh-Ju Wang2, Chien-Fu Hung’, Han-Jung Lee’? 2 and Michael M. C. Lai2

‘Institute

of Molecular

“Howard

Hughes University

(Received

Biology,

Academia

Sinica,

Nankang,

Taiwan

11529

Medical Institute and Department of Microbiology of Southern California School of Medicine Los Angeles, CA 90033, U.S.A. 15 March

1991; accepted 4 September

1991)

Human hepatitis delta virus (HDV) RNA has been shown to contain a self-catalyzed cleavage activity. The sequence requirement for its catalytic activity appears to be different from that of other known ribozymes. In this paper, we define the minimum contiguous sequence and secondary structure of the HDV genomic RKA required for the catalytic activity. By using nested-set deletion mutants, we have determined that the essential sequence for the catalytic activity is contained within no more than 85 nucleotides of HDV RNA. These results are in close agreement with the previous determinations and confirmed the relative insignificance of the sequence at the 5’ side of the cleavage site. The smallest catalytic RNA, representing HDV genomic RNA nucleotide positions 683 to 770, was used as the basis for studying the secondary structure requirements for catalytic activity. Analysis of the RNA structure, using RNase Vi, nuclease S, and diethylpyrocarbonate treatments showed that this RNA contains at least two stem-and-loop structures. Other larger HDV RNA subfragments containing the catalytic activity also have a very similar secondary structure. By performing site-specific mutagenesis studies, it was shown that one of the stem-and-loop structures could be deleted to half of its original size without affecting the catalytic activity. In addition, the other stem-and-loop contained a six base-pair helix, and the structure, rather than the sequence, of this helix was required for the catalytic activity. However, the structure of a portion of the stem-and-loop remains uncertain. We also report that this RNA can be divided into two separate molecules, which alone did not have cleavage activity but, when mixed, one of the RNAs could be cleaved in trans. This study thus reveals some features of the secondary structure of the HDV genomic RNA involved in self-catalyzed cleavage. A model of this RNA structure is presented. Keywords:

hepatitis

delta virus RNA; ribozyme; secondary site-specific mutagenesis

1. Introduction

structure;

trans-cleavage:

separate enzyme and substrate oligonucleotides, with the enzyme RNA possessing the ability to cleave the substrate RNA in trans (Uhlenbeck, 1987). A transcript of newt satellite DNA also possesses catalytic activity of this type (Epstein & Gall, 1987). The second type is represented by the negative-stranded RNA of satellite tobacco ringspot virus (Feldstein et al., 1989; Hampel & Tritz, 1989). The catalytic activity of this RNA resides within a stretch of 50 nucleotides and can cleave a substrate RNA in trans (Hampel et aE., 1990). This catalytic domain differs in sequence from the hammerhead RNAs. The third type is represented by hepatitis

Several types of naturally occurring RNA, excluding transcripts of cellular genes, have been shown to possess self-catalyzed cleavage activities, and each of them displays different sequence and structural requirements. The first type is represented by plant viroid, virusoids and satellite RNAs, each of which contains a similar “hammerhead”-like self-cleaving domain (Forster & Symons, 1987). This domain consists of three branched helices and two single-stranded regions of highly conserved nucleotides (Forster & Symons, 1987). Furthermore, this domain could be divided into

233 0022-2836/Q2/010233-13

$03.00/O

(0

1992

Academic

Press

Limited

delta

virus

(HDV?)

RNA.

which

also

has

a distincat

st,ructural requirement for its catalytic activity (LVu rt al.. 1989). More recently, a ~Veurosporn mit,ochondrial RNA has been shown to possess catalytic activit’y and may represent, yet another class of catalyt,ic RNA (Saville $ Collins, 1990). HDV contains a circular single-stranded R&A of I.7 kb with a high degree of intramolecular complrmentarity (Wang et al., 1986; Makino et al., 1987), a characteristic of the catalytic virusoid RNAs. However, unlike virusoid RNAs, HDV RNA has an open reading frame on its antigenomic strand that encodes a virus-specific protein, the hepatitis delta antigen (Wang rt al.. 1986). This combination of R,NA catalytic activity and protein-encoding cspability differs from most other catalytic RNAs. The catalytic domain of the HI)V genomlc RXA has been shown to consist of no more tha.n 85 nucleotides, and contains neither t,he consensus selfcleaving sequencenor the hammerhead struct,ure of t,hc catalytic virusoid RNAs (Wu et al., 1989; Wu 02 Lai. 1990: Perrotta & Been, 1990). The cleavage o(*curs between nucleotides 688 and 689, yielding a 3’ cleavage product with a 5’-hydroxyl guanosine residue and a 5’ cleavage product wit’h a 2’,3’-(*y(alic phosphoryl uridine residue (Wu et al., 1989). To st,udy further the catalytic act’ivity of HDV genomic RNA, we have begun to determine the precise sequence and secondary st’ructure involved in HDV genomic RKA catalysis. In this paper we further define the minimum RNA sequence constituting the HDV genomic catalytic RKA. This RNA forms at, least, two stem structures which are required for the catalytic activity. We have also shown that) these RNA fragment’s can self-cleave in trans.

2. Materials

and Methods

(a) Plasmid

DNA

The construction of plasmids pD(654-786). pD(654-801) and pD(654-732) was described by Wu et al. (1989) under the names pHN3-24. pHN54 and pHNY-7. respectively. All other plasmids containing truncated lengths of HDV genomic RNA sequence were derived from pD(654-786), and the scheme for their construction is outlined in Fig. 1. Plasmids pD(654-770), pD(654-769) and pD(654&764) are 3’.deletion mutants of pD(654-786) with respect to the HDV sequence. Plasmids pD(654-X) (x = 767, 765, 763. 760, 758 and 756) and plasmids pD( y-770) ( y = 663,667 and 683) were derived by processive deletions from the EcoRI site and Hind111 site of pD(654-770), respectively. Plasmid pD(683-770) was further deleted from the EcoRI site to generat)e clones pD(683-z) (z = 764, 749, 746, 744, 743, 741 and 735). Plasmid pD(683-727) was derived from pD(683--735). For all of the constructions, each parental plasmid was linearized by an appropriate restriction enzyme, digested from the restriction site by exonuclease III and mung bean nuclease treatment, and then ligated to a linker according to standard procedures (Maniatis et al., 1982). t Abbreviations used: HDV, hepatitis delta virus; kb, lo3 bases or base-pairs; DEP. diethylpyrocarbonate: AMV, avian myeloblastosis virus.

Mutants listed in Tablt~ I itr~ Ilttk(Lr-s~.atI ~IIIC~ itltc~rn;~l deletion/suhstit,ution mutants clerivrtl tiorti phllill pD(683 7iO). Each plasmid \vas 1aon&ruc%rd IJ~ ~r~srrtin~ a pair of appropriate synthetic DNA fragments into ck restriction enzyme sit,? of an appropriatr pli~srnici construct. downstream from the HI)\’ seqrrrnc~ IJIW rxample. pI)(683-770) -I,1 u-asc~mst ruc+Ati hx insr,rt ing tht, follo\ving fragment. 5’p(:AT(‘(‘TAAT(:OC(:rZATUG(:UAATT(’(~ 3’ (:~\TTA(I(‘(:CTTA(‘(‘(‘(‘TT,\~\(:(’~’T.\(:F,:~’.

3’

into the RamHI sit,e of plasmid pD(683 749). Mutant I to mutant 6 in Fig. 4 arp substit,utiorl mutant’s of plasmid pD(683-770)-1)4. each c~ontaininp nucleotide substitution(s) within HI)\’ genomi~ RNA nucleotides 690 to 695 and/or nucleotides 719 to 724. Each mut)ant was constructed by replacing thr wild-typt, sequence with synthetic oligotlrrcleotidrs t)hat c.ontairr mutated sequences. Plasmid pD(654A774) (Fig. 5) was construc+rtl 1)~ inserting a pair of synthetic DNA fragments into t)he EcoRT site of p1)(654732). Plasmid pD(683-~728) (Fig. 5) was ronstructed by inserting a HanlHI linker into the &oRI site of pD(683-727). The sequence of each plasmid was confirmed by the dideoxyribonucleotide chain-termination sequencing method (Sanger Ptal., 1977).

(b)

RSA transcription

and cloaoage reactions

Plasmids were linearized with a restriction enzyme and then transcribed using phage T7 RNA polymarasr according to methods described (Wu et al.. 1989). Briefly. transcription was performed in 40 mM-Tris HCI (pH 8.1). 5 tnM-dithiottireitoi. 12 mM-MgU,. I mM-spermidinr. 0.1 mg bovine serum albumin/ml, 0.9 mM each of ribonucaleoside triphosphates, @OS mCi of [cx-~*P]GTP. 15 units of bacteriophage T7 RNA polymerase at 37°C for 1 h. The transcription products were incubated with 3..5 M-urea al 95°C for 90 s. and then analyzed by electrophorrxis on 6o, or 8”, (w/v) polyacrylamide gels containing 7 >f-urea. Tn general. R,NA molecules that, contained the catalytic, activity would self-cleave during the process of the transcription reaction followed by denaturation. To further confirm the catalytic activity of each RNA tnoleculr. we also performed cleavage reactions using t,he purified primary transcripts. Briefly. t’hr 32P-labeled full-length transcr:ption products were purified from gel and incubat,ed with 40 m&t-Tris. HCI (pH 8.1) and 12 mM-Mg(ll, at 37°C” or 60°C‘ for 30 min. a,nd then analyzed l)y electrophoresis. Some RNA transcripts were heat-denatured and renatured for several cycles to enhance the cleavage activity as described (Wu & Lai. 1990). The HD(761-774) and HD(76&770) RNAs were synthesized from synthetic DN4 templates c,ontaining a T7 promoter and using T7 RNA polymerase as described by Milligan ef aZ. (1987).

(c) RN,4 secondary structure analysis The RNA products transcribed from pD(654-801) and pD(683-770) were treated with RNase-free DNase, extracted with phenol/chloroform and recovered by precipitation with ethanol. The purified RNA transcripts were then resuspended in water and partially digested under non-denaturing conditions with diethylpyrocarbonate (DEP), nuclease S, or RNase V, according to published methods (Rerkhout et al., 1989). For DEP

Hepatitis

Delta

Virus

RNA

Summary

of stem-and-loop

235

Structure

Table 1 Nucleotide

126

sequence

II

mutants

differencest

(: G G (’ .2 A (: A 1’ 11 (‘ (’ G A U G G G A C C G 11 (I c‘ C C U C’ G G tl A A V G G (’

pD(683-770)

\-es

G G G (1 .4 A C B VU

(’ (’ G A G G G G B C C G U C C g g a u c c V A A L’ G G C

pD(683-770)-L]

so

: a u C C C U (’ G G I’ ~~___ A A Y G G C (:(:G(‘r\A(‘Ar.U(‘CGAGGGGBCg( -(: G (4 (’ A A (: A 1. Ll (‘ (’ G A c G G a u C C G IT (1 C C C I: (’ G G IT A A U G G (’

pD(683-770)-I,2

Yes

pD(683-770)-1~3

Yes

(: G G (’ h A (’ A IT Ii (: g G A u c c G A C C G V C C C C U C G G IT A A li G G (’

pD(683-770)-L4

NO

(: G (: g A A II u IT IT (‘ (’ G B G G G G A C (’ G U C C C C I’ (’ G G 11 A A - 1T (: (: (’

pD(683-770)-1~5

Yes

(: (: (: (‘ A A (: A I’ V (‘ gQA

pD(683%770)-L4a

\‘PS

pD(683-770)-D]

l-es Yes

G (‘

pD(683-770)-D2 pD(683-770)-D3

G G (‘

pD(683-770)-D4

Yes

A A l’ G G (’

pD(SS3--770)-I,5

No

u c c G A C C G IT C g g a C (’ c G I’ .I A I’ G G (’

(: U G (’ A A C A 1. U (’ g G A u c c - - - - G L’ C 4 g a V (’ c G I. ~~__ A A 11 G G C (: (4 G (’ A A C A 1’ U C g G A u c c ~ - - - a -__ LT (’ c G IT A~--A CTG G (’ G (: G g iz A u u I’ U (1 (’ G $ G G G - -

-

GGGgAAuucggauc(~GG--~~

- ~~

G (: (: g 9 $ u u U V (’ (’ GAG

(’ IT (‘ G G

G G - -

-mCgauc -

-

A B [IQ ‘1 g1:

:(:

l-es

t Only the sequence from nucleotides 726 to 763 is shown for each mutant. Sequences from nurleotidrs 6X3 to i25. and Tti3 to 770 of each mutant arc identical with pD(683-770) and are not shown in this Table. Complementary sequences within the nucleotide 726-763 region are underlined. The location of nucleotides which have been delet,ed are indicated by dashes. Sequences which differ from the wild-type pD(683&770) are designated by lower-case letters. $ Yes indicates cleavage of at least 50?/, of the RNA during transcription reactions. No indicat,es that the RXA failed to (.Ieave under any of the conditions trstrd.

digestion, 20 ~1 of UEP was added to RKA in 200 ~1 of 5 mM-sodium cacodylate (pH 7.5), 10 mM-&$I,, and incubated at 37°C for 30 min. For S, digestion, 0.05 unit

of nuchlease S, was mixed with mixture

containing

280 mM-NaCl,

RNA

in 4 ~1 of reaction

50 mM-sodium

acetate

(pH 4.5). 4.5 rnM-zinc sulfate and incubated at 37°C for 5 min. For RNase V, digestion, @02 unit of RNase V, was incubated with RN’A in 100 ~1 of 20 mi\l-T’ris. HCI (pH 7.5). 2 rnM-MgCl,, 100 mM-NaC1 at 30°C for 8 min. After incubation. the RNA was purified by extraction with with

phenol/chloroform

and

recovered

by

precipitation

ethanol. The cleavage sites were detected by primer extension. The primer oligonucleotide was end-labeled with [y-32P]ATT’ and phage T4 polynucleotide kinase. The partially digested RNA was dissolved in 4 ~1 of annealing buffer (40 mi\l-Tris. HCl (pH 7.5), 25 mM-MgCl,. 50 m&l-NaCl), mixed with 1 ~1 of 32P-labeled primer and heated at 7,5”C for 2 min. After the reaction mixture was cooled to room temperature, 1 unit of AMV reverse transcriptase in 10 ~1 of RT buffer (50 mM-Tris. HCI (pH 7.5). 75 mM-KCI. 10 mM-dithiothreitol, 100 pg bovine serum albumin/ml, &5 mm-dNTPs) was added and incubated at 37 “C for 15 min. Reaction products were mixed with 15 ~1 of formamidr sample buffer (95% formamide, 20 rnnl-EDTA). heat-denatured at 100°C for 1 min, and separated by elrctrophoresis in 6% polyacrylamide gels containing X M-urea.

3. Results (a) Determination of the minimum qenomic RNA sequence containing cleavage activity

contiguous HD V the autocatalytic

We have shown that HDV subgenomic fragments can undergo self-cleavage between nucleotides 688 and 689, and that the autocleavage activity resides

within a stretch of no more than 133 nucleotides encompassing the cleavage site (Wu et al., 1989). Subsequently, several smaller HDV genomic RNA subfragments, the smallest one being 85 nucleotides, have been demonstrated to contain the self-cleavage activity (Perrotta & Been. 1990; Wu & Lai, 1990). To define precisely the minimum contiguous sequence required for self-cleavage, we constructed a series of nested-set deletion mutants at both the 5’ and 3’ t’ermini of an HDV genomic RNA subfragment that displays the self-cleaving activity. The cDNA clone pD(654-786) (previously named pHN3-24: Wu et al., 1989), which contains HDV genomic RNA from nucleotides 654 to 786 plus several nucleotides derived from the vector, was used to construct a series of mutants containing different lengths of HDV-specific sequence (Fig. 1). The HDV RNA was transcribed from the T7 promoter using T7 RNA polymerase, and examined for self-cleavage activity as described in Materials and Methods. RNAs that did not undergo selfcleavage during transcription reactions in ,vitro were further incubated at higher temperatures or subjected to heat denaturation-renaturation cycles as described by Wu & Lai (1990) to ascertain whether cleavage could be detected under more sensitive conditions. Of nine 3’-terminal deletion mutants that were tested (Fig. 2(a)), the RNAs from pD(654-770), pD(654-769) and pD(654-767) were cleaved into two fragments, while none of the smaller RNA transcripts was cleaved under any of the incubation conditions used (Fig. 2(a)). Since the smallest RNA that showed cleavage activity was pD(654-767),

H.-S.

Wu et al.

1 1. 2. 3.

1. 2. 3.

HmdIII digestion Exonuclease III/Mung nuc1ease treatment Ligation

;;;:::I;;;;

bean r-

EcoRI digestion Exonuclease EcoRI linker iI;

7

C-1

oOf654-764)

III/Mung ligation ::

I

3

bean

nuclease

tredtnent

EcoRI digestlorl Exonuclease 111muncl nuclease treatment BarnHI linker liqation

bean .

t pD(663-770) pD(667-770) pD(683-770)

1. 2. 3.

,+) (i, (+1

FmRI digestion Exonuclease III/Mung nuc1ease treatment BumHI linker ligation

bean

b

pDl654-767) pDf654+765) pD(654-763) pD(654-760) pD(654-758) pD(654-756)

(+) C-1 j-1 t-1 t-1 (-1

pD(683-764) pD(683-749) pD(683-746) pD(683-744) pD(683-743) pD(683-7411 pD,(683-735)

(-) (-) (-) (-) (-) (-1 (-1

1. 2. 3.

BarnHI diyestiorl Exonuclease III/Mung nuc1ease treatment EcoRI linker ligation

bean

4

Figure 1. Schematic diagram of the construction of plasmids used in this study. All clones constructed for this study were derived from pD(6544786) (previously named pHN3324; Wu et aZ., 1989). The HDV sequence (hatched box) was deleted to various extents by treatment with exonuclease III and mung bean nuclease after linearizing the plasmids (pD(654-786), pD(654-770) and pD(683-770)) with appropriate restriction endonucleases. After ligation in the presence of a linker, plasmids containing different lengths of HDV sequence downstream from the T7 promoter (T7 P) were obtained. Plasmids were linearized using the appropriate restriction endonucleases and transcribed by T7 RNA polymerase. The sequences derived from the vector and linkers in each plasmid were identical with those shown for the and pD(6677770) had GGGA upstream parental plasmids, except that pD(6633770) had GGGAGCTCCGGATCCGGG from the HDV sequence. pD(683-770) was previously named pHJ12L (Wu & Lai, 1990). The presence (+ ) or absence (- ) of self-cleavage of each RNA also is indicated.

and the largest one that failed t,o cleave was pD(654-765), the 3’ boundary of the minimum HDV catalytic sequence was localized between nucleotides 765 and 767. In addition, the smaller of the two RNA cleavage products, which represent the 5’ cleavage fragments (Wu et al., 1989), were similar in size for all of the RNAs tested, suggesting that their cleavage sites were identical. Three 5’ end nested-set deletion mutants derived from pD(654-770) were tested. The RNAs made from each of these clones, pD(663-770): pD(667-770) and pD(683-770), underwent cleavage to differing extents (Fig. 2(b)). When transcription was performed in the presence of [JJ-~‘P]GTP, only the smaller cleavage product and the full-length transcript were labeled (Fig. 2(c)), establishing that the smaller RNA was the 5’ cleavage product. The 3’ cleavage products of each RNA were of the same size, indicating that their cleavage sites were identical. Since the smallest self-cleaving RNA was pD(683-770), and it cleaved to about 80% under the optimum incubating conditions (Wu & Lai, 1990), we conclude that the 5’ boundary of the

minimum HDV catalytic sequence was no more than six nucleotides 5’ from the cleavage site. To further establish the minimum contiguous HDV sequence containing self-cleavage activity, we generated additional nested-set deletions on the 3’ side of the HDV sequence in the pD(683-770) clones. Eight clones were tested, and none of these RNAs was cleaved under any of the conditions used (Fig. 2(c)). These results confirmed that’ the minimum contiguous HDV genomic RNA sequence required for self-cleavage is 85 nucleotides (from nucleotides 683 to 767). (b) Determination

of the secondary structure self-cleaving HD V RNAs

of the

On the basis of Zuker’s program (Zuker & Stiegler, 1981), we suggested that the HDV genomic RNA subfragments t’hat underwent self-cleavage could assume a clover-leaf conformation with several stem-and-loop structures (Wu et al., 1989). It was not known whether this was the conformation assumed by the self-cleaving HDV genomic

Hepatitis

Delta virus

(a)

RNA

237

Structure

(b)

(cl

Figure 2. Analysis of transcription products of various plasmids. The EcoRI or BarnHI-linearized plasmidswere transcribed by T7 RNA polymerase in the presence of [cI-~‘P]GTP or [y-‘*P]GTP (rG). Clones pD(654-786), pD(654-770), pD(6544769) pD(654-764) pD(6633770), pD(667-770), pD(6833770) and pD(6833727) were linearized by EcoRI, and the remaining clones were linearized by BumHI. The products of the transcription reactions were separated on an 8% (a) or 12 o/0 ((b) and (c)) polyacrylamide gel containing 7 M-urea. Arrows indicate full-length transcripts, filled arrowheads indicate 3’ cleavage products, and open arrowheads indicate 5’ cleavage products. The minor bands in most lanes of (c) were non-specific degradation products

RNA and whether all of the self-cleaving HDV RNAs had a similar secondary structure. To understand the conformation of the self-cleaving HDV RNAs, we used the purified run-off transcripts of the EcoRI-linearized pD(654-801) for secondary structure analysis. Since pD(654-801) RNA underwent self-cleavage almost to completion during transcription reaction (Wu et al., 1989), the RNA structure probed was essentially that of the 3 cleavage

product

(nucleotides

689

to 801)

rather

than the full-length transcript (nucleotides 654 to 801). RNA transcribed in vitro was subjected to limited treatment with single-stranded RNA-specific nuclease S, or DEP, or doublestranded RNA-specific RNase V,. The partially digested RNA then was examined by primer extension using an oligonucleotide primer complementary to the 3’ end of this RNA (nucleotides 789 to 801).

The results shown in Figure 3(b) demonstrate a cleavage pattern more consistent with the model of RNA conformation shown in Figure 3(a), than with the more stable RNA conformation predicted previously (Wu et al., 1989). In particular. a singlestranded loop region that is unique to this conformation

was

identified

at nucleotides

696 to 701.

These structural studies also suggested the presence of at least two stem-and-loop structures in the selfcleaving RNA (Fig. 3(a)). A few overlaps between the DEP-sensitive sites and RNase VI-sensitive sites suggested

that

the conformation

of some regions

of

RNA may be flexible. These overlaps were particularly evident at the junctions between the singlestranded and double-stranded RNA regions. Since the combined cleavage sites of these nucleases and chemical treatments did not include all of the nucleotide linkages (Fig. 3(a)), some portions of this

I 770

G



A A

CG*

'GC.

G,UA CG f -GC-

I

CG

720

(cl

G

C

‘4OlCG-+

GUCCG,

Stem-and-loop

GUAA ’ u 160 750

G

II

740 AC CCG&%GG c GGCUCCYC u

Stem-and-loop

UGCA

id)

725

Figure 3. Secondary structure of the HDV RKA subfragments. (a) The schematic drawing of a model of the secondary structure of pD(6544801) R?u’A as predicted by computer analysis and nuclease and chemical probing. The sites cut by either nuclease 8, or DEP are indicated by arrows; those cut by ribonuclease Vi are indicated by dots. The major cmting sites are marked by larger symbols. The vector-derived nucleotides are shown in lower case. (1)) Primer extension with pD(654~801) RNA after treatment with DEP. nuclease S, or ribonuclease Vi. After each treatment (see Materials and Methods). a 32P-labeled primer oligonuclrotidr (complementary to nucleotides 789 to 801) was annealed to t,he RNA and extended with reverse transcriptase. Lane 1. without treatments: lane 2, with treatments. The extra bands and enhanced bands after treatments are indicated by arrows. Sequencing ladders were run in parallel to serx-e as size markers. Lane zY!in the S, nuclease panel shows a 5 min digestion, while lane 3 shows an 8 min digestion. The major cutting sites are marked by larger symbols. (c) A schematic. drawing of the secondary structure of pD(683 770) RKA as predicted from the computer analysis and nuclease and chemical probing. The symbols are the same ax for (a). (d) Analysis of pII(683 770) RNA by thr same methods as in (b). The primer oli~onu~leoticIr used was c~omplrmrntar~ to nuclrotides 759 to 770 of HIIV ernomic RSA. Suc~lrc)titlrs adjxc,riit to the l)rirner c~ould root he probed unarnhi~uously.

3’-uuaag

pD(683-770)

DEP

240

H-S.

structural model, particularly the upper loop region of stem-and-loop I, were only tentative. Similar studies were performed on the purified run-off transcripts of EcoRI-digested pD(683-770), which is t,he smallest self-cleaving HDV RNA. This pI)(683-770) RNA cleaved t,o more than 501’,, during transcription reaction (Fig. 2(b)). Thus, the RNase and DEP treatments should give the strutt,ural information of both the full-length transcript (nucleotides 683 to 770) and the 3’ cleavage product (nucleotides 689 to 770; Fig. 3(d)). The RNA cleavage patterns were consistent wit,h the strucatural model presented in Figure 3(c). which has a core structure similar to that of pl)(654-801). Similarly, the structure of the upper half of stemand-loop I was less certain since this region was susceptible to both V1 nuclease and DEP t’reat>ments, suggesting that alternative structures were present. The structure of stem-and-loop TT of this RNA could not be determined with certainty since the primer used in these experiments was complementary t,o nucleotides 759 to 770. (c) Site-specijk

mutagenesis

of stem-and-loop

Wu et al

tional three-nucleotide deletion in thth sttbm region. pD(683-770)-D5, completely inact,ivatetl t ht~ &If‘ cleavage activity. These results are in agn~rmrtlt with da,ta obtained from the linker-scan mut.ant,s. Considering the dat,a from both thr lin ktar-scaat1 and deletion/substitut’ion mutants. it appears I hat within nucleotides 732 t’o 761 (st,c~m-anti-lool) TT), only a seven-base-pair st)em of variahlr st~qurn(~~~ is necessary for the autocatalytic cleavagt> activit? of HDV genomic RNA. We also examined whether this stem st)ructure was essential for thca selfcleavage activity of larger HI)\’ R?r’A srrbfragments. We constructed plasmid pl)(654A774). which consisted of the HDV sequence from rrucleotides 654 to 774 with a substitution of 13 unrelated nucleotides (GAAUU(:GGAAr!li(‘) for H 1)V nucleotides 733 to 760 (Fig. 5(a)). These 13 nuclrotides formed a six-base-pair stem and a threenucleotide loop, replacing the original H l)\,Y-spe(aifit stem-and-loop II. This RNA still cleaved it,self. reducing the requirement of t,hr stern to six basrpairs and confirming that the sequence rrquirrment was not’ specific (Fig. 5(b), lane 4. and ((b) lane 5).

II

To determine whether the stem-and-loop structures found in the HDV genomic RNAs are required for the catalytic cleavage activity, we first generated site-specific mutations within stem-and-loop II and examined their effects on self-cleavage activity. Five linker-scan mutants of clone pD(683-770) were constructed (Table l), in which HamHT or EcoRI linker sequences were substituted for various HDV-specific nucleotides in the stem-and-loop IT region (nucleotides 732 to 761). The mutations in the loop region (pD(683S770)-L2) and at’ each end of the proposed stem (pD(683-770)-L3 and -1,5) did not affect the self-cleavage activity (Table 1). However, mutations in the middle part of the stem (pD(683-770)-Ll and -L4) completely inactivated the self-cleavage activity. These results suggested that the middle part of the proposed stem-and-loop structure IT was essential for the catalytic activity. The self-cleavage activity of pD(683%770)-L4 RSA was restored when compensatory mutations were introduced into the stem region (pD(683-770).-TAa). Thus, structure rather than sequence of the stem region was required for the catalytic activity. To confirm further the findings that’ the loop and either side of the stem II could tolerate base changes, we constructed two deletion mutants of pD(683-770)-L4a, pD(683S770)-Dl and -D2. These deletion mutants lacked four and nine nucleotides of loop IT and its adjacent nucleotides, respectively (Table 1). Both mutant RNAs retained the self-cleavage activity. We also examined the importance of the bulged uridine residue (nucleotide 757) within this stem region. Two mutant RNAs, pD(683-770)-D3 and -D4, which have stems consisting of seven perfect base-pairs of different nucleotides, retained selfcleavage activity (Table l), indicating that the bulged uridine residue was not essential. An addi-

(d) Site-speci$c

mutagenesis

of helix

I

To establish the importance of t,he six base-pairs formed between nucleotides 690 to 695 and nucleotides 719 to 724 in the predicted stem-and-loop I (Fig. 3), we constructed mutants within this region and examined their self-cleavage activity. Six substitut’ion mutants of the clone pD(683- 77O-J)4 (Table 1) were generated (Fig. 4(a)). Mutations that dest,royed eit,her one (mutants 3. 4 and 5) or two (mutant 6) base-pairs of’ the stem completely self-cleavage activit’y (Fig. 4(b)). abolished Continued incubation at TiOY’, conditions which resulted in improved cleavage of the wild-type RNA (Fig. 4(h) and (c)), did not result in the cleavage of mutant’s 3, 4. 5 and 6 (Fig. 4(c)). However. t,hr selfcleavage activity was restored when compensatory mutat)ions were int’roduced. restoring t,hr basepaired structure (mutants 1 and 2. Fig. 4(b) and ((1)). Thus, as demon&rated for stem-and-loop Il. structure, rather than sequence of the base-paired region of the proposed stem-and-loop 1. was required for the catalytics activit,y.

(e) The self-cleaving KNA

can be separated two domains

into

The HDV RNA self-cleavage characterized so far involved a catalytic domain and a substrate domain that were contained within the same RNA molecule. We sought to determine whether the catalytic and substrate domains of HDV RNA could be separated and undergo cleavage in trans. pD(6546774) RNA was chosen (Fig. 5(a)) for this study. The pD(654-732) RNA, which represents the 5’ side of pD(6546774) and includes the catalytic cleavage site (between nucleotides 688 and 689), did not show cleavage activity in the in vitro transcription reaction or as a purified RNA (Fig. 5(b), lanes 1 and

Hepatitis Stem-and-loop

241

Structure

following 2). However, HD(761-774) RNA, which

I

pStem-and-loop

(b)

Delta Virus RNA

II

(c)

Figure 4. Self-cleavage activity of helix I mutants. (a) Sequences of the mutant RNAs in the helix I region. All 6 mutants were derived from pD(683-770)-D4 RNA (wild-type) (Table l), whose entire sequence is shown in the conformation similar to that in Fig. 3. For mutants, only the sequences of nucleotides 690 to 695 and 719 to 724 (helix I) are shown. The presence (+ ) or absence (- ) of the cleavage activity is indicated on the diagram. The nucleotides absent’ from HDV genomic RNA are shown in lower case. (b) Ueavage reactions of different mutant RNAs. REAs were synthesized by T7 RNA polymerase at 37°C for 1 h. The reaction products were heat,-denatured in the presence of 3.5 M-urea and analyzed on a 100/b polyacrylamide gel containing 7 M-urea. The top bands are the primary transcripts, and the lower bands are the 3 cleavage products. The 5’ cleavage products were run off the gel. (c) Same as (b) except that the transcription products were further incubated at 50°C for 10 min before denaturation.

the addition of an represents the 3’ side of pD(654A774), two new RNA products were detected (Fig. 5(b), lane 3). These two RNA products were characterized further by combining RNA transcripts different conditions. When under pD(654-732) RNA transcribed in the presence of [Y-~*P]GTP was used, only the large cleavage product was labeled (Fig. 5(c), lane 2), indicating that this product was the 5’ cleavage product. The large cleavage product also had the same electrophoretic mobility as the 5’ cleavage product of pD(654A774) (Fig. 5(c), lane 5). In contrast, the small cleavage product was labeled with [32P]pCp by RNA ligase (Fig. 5(c), lanes 3 and 4). consistent w’ith its being the 3’ cleavage product. I’nder these conditions, the 5’ cleavage product was not labeled due to its 2’,3’-cyclic phosphoryl end (Wu et al., 1989). The end-nucleotide analysis of the cleavage products (Uhlenbeck. 1987) further established that these t,wo RNA products were cleaved between nucleotides 688 and 689 (data not shown). These results indicated that the HD(761-774) RNA restored the catalytic cleavage of pD(654-732) RNA. probably because the stem-and-loop IT structure was restored, and that the two RXAs cleaved in trans. Kinetic st,udies of the tram-cleavage reaction of these two RNAs showed that the initial reaction rate was a function of the enzyme concrntration (Fig. 6(a)), and the extent of cleavage reached more than 800, when t,he enzyme-tosubstrate ratio was more than tenfold at 50°C (Fig. 6(a)). Similar results were obtained when pD(683-770)-D4 RNA (Table 1). which contains a different stem-and-loop II sequence (Fig. 5(a)), was divided into two separate molecules (pD(683-728) and HD(760-770)). The combination of these two R,NAs restored the catalytic cleavage (data not shown). reactions with pD(654-732) Tn the trans-cleavage and HD(761-774) RNAs, the proportion of RNA cleaved was low. One possible explanation could be that the two separate RNA molecules had intramolecular secondary structure that prevented the intermolecular interactions. Previously. we have shown that the extent of &s-cleavage of HDV RNA subfragments could be increased by repeated heat denaturation and renaturation, probably as a result of unfolding and refolding of the RNA structure (Wu & Lai, 1990). Thus, we sought to determine whether this treatment could increase the amount of trans.cleavage as well. Figure 6(b) shows that the trans-cleavage of pD(654-732) RNA by HD(761-774) RNA increased from loo, to 300/b after five cycles of heat denatura~t~ion and renaturation.

4. Discussion In this paper, we have demonstrated that t’he minimum contiguous HDV genomic RNA sequence required for autocleavage activity is from nucleotides 683 to 767. Thus, the actual sequence containing

A774)

pD(654

u

7’0 cc / u

‘CGC GC AU

pD(654-732)

WA

C CG GC GC

-

720 pD(683-770)-D4

CG CG

688,689

‘c

GC

UG

CCUGA

U

u

732

G

A

G GC AACr ,John Jaeger for help in t.he (~omputt~r~ analysis of RNA a.nd Yi-Shuian Huang and Tung-(iuang Hsueh for help in the construction and sequencing of some of the mutant,s. We also thank Drs John Polo and Tom MacNaughton for critical reading and editorial assistance. and Daphne Shimoda for typing the manuscript. This work was supported by grant NSC-80-0203-800 I 14 from the National Science Council of the Republic of (‘hina (to H.-N.W.) and U.S. Public Heatt.11 Service research grant AI 26741 from the National Tnst,itutes of’ Health (to M.M.C.L.). .M.,M.(‘.L. is an investigator of the Howard Hughes Medical Institute. References Berkhout’. R., Silverman, R. H. & Jeang, K.-T. (1989). Tat trans.activates the human immunodeficiency virus through a nascent RNA t.arget. (‘~11. 59. 273 282. Chao, Y.-C’.. Chang, M-F., Gust, I. & Lai, M. M. (‘. (1990). Sequence conservation and divergence of hepatitis fi virus RNA. Virology, 178, 3844392. Chao, Y.-C’.. J,ee. fl. M.. Tang, H. S.. (:ovindarajati. S. h Lai. M. M. t’. (1991). Molecular c*loning and characterization of an isolate of hepatitis delta virus from Taiwan. Hepatology. 13, 34.5~ 352. Epstein. 1,. M. & Gall, *J. G. (1987). Self-cleaving t.ranscripts of satellite I)NA from the newt. (‘P/I, 48. 5355543. Feldstein. I’. A., Buzayan, CJ.M. & Breuning, G. (1989). Two sequences participating in the autolytic processing of satellite t.obacco ringspot virus complrmentary RNA. Gene, 82. 53361. Forster, A. (1. $ Symons, R. H. (1987). Self-cleavage of plus and minus RNAs of a virusoid and a structural model of the active sites. Cell, 49: 21 I-220. Hampel. A. & Tritz, R. (1989). RNA catalytic properties of the minimum ( -) STRSV sequence. Riochmnistry, 28, 4929-4933. Hampel, A., Tritz, R.. Hicks, M. & Cruz, 1’. (1990). “Hairpin” catalytic RNA model: evidence for helices and sequence requirement for substrat’e RNA. Nucl. Acids Res. 18, 299-304. Makino, S.. Chang. M.-F.. Shieh, C.-K., Kamahora. T., Vannier, D. M., Govindarajan, S. & Lai, M. M. (‘. (1987). Molecular cloning and sequencing of a human hepatitis delta (6) virus RNA. Nature (London), 329, 3433346. Maniatis, T., Fritsch, E. F. Kr Sambrook. .I. (lY82). Molecular

Ching:

A

i,ctboratory

Menurcl

(Ml

Press. t ‘01~1 Spring Spring Harbor T,abora.tory Harbor, NY. Milligan. ,J. F.. Groebe, I). R.. Witherell, t:. W. K: Uhlenbeck, 0. (1. (1987). Oligoribonucleotide synthesis using T7 RNA polymerase and synthet,ic DNA templates. Nucl. Acids Res. 15. 8783-8798. Perrotta, A. T. & Been, M. D. (1990). The self-cleaving domain from the genomic RNA of hepat,itis delta virus: sequence requirements and the effects of denaturant. I%‘ucl. Acids Res. 18, 6821-6827. Perrotta. A. T. & Been, M. D. (1991). A pseudoknot structure required for efficient self-cleavage of hepatitis delta virus RNA. Nature (London), 350, 434-436. Sanger, F.. Nicklen, S. C Co&on. A. R. (1977). DNA sequencing with chain-terminating inhibitors. Prof. Nat. dead. Ski., l’.fJ. A 74, 5463.-5467.

Hepatitis

Delta Virus RNA

Saville. B. J. & Collins, R. A. (1990). A site-specific selfcleavage reaction performed by a novel RNA in Neurospora mitochondria. Cell, 61, 685-696. I’hlenbeck, 0. C. (1987). A small catalytic oligoribonucleotide. Nature (London), 328, 596-600. Wang, K.-S.. Choo. O.-L., Weiner, A. J., Ou, J.-H., Najarian. R. C.. Thayer, R. M., Mullenbach, G. T., Denniston, K. J., Gerin, J. L. 6 Houghton, M. (1986). Structure, sequence and expression of the hepatitis delt’a (6) viral genome. Naturr (London), 323. 508-514.

Structure

245

Wu. H.-N. & Lai, M. M. C. (1990). RNA conformational requirements of self-cleavage of hepatitis delta virus RNA. Mol. Cell. Biol. 10, 5575-5579. Wu, H.-N., Lin, Y.-J., Lin, F.-P., Makino, S., Chang, M.-F. & Lai, M. M. C. (1989). Human hepatitis 6 virus RNA subfragments contain an autocleavage activity. Proc. Nat. Acad. Sci., [‘..~.A. 86, 1831-1835.

Zuker, M. & Stiegler, P. (1981). Optimal computer folding of large RNA sequences using thermodynamics and auxiliary information. Nucl. Acid,? Res. 9. 133-148.

Edited by F. E.

Cohen

Sequence and structure of the catalytic RNA of hepatitis delta virus genomic RNA.

Human hepatitis delta virus (HDV) RNA has been shown to contain a self-catalyzed cleavage activity. The sequence requirement for its catalytic activit...
6MB Sizes 0 Downloads 0 Views