Structures of Viroids Virusoids and Satellites

)3 tz 8+

STRUCTURES OF

VIROIDS

VIRUSOIDS

AND

SATE LLI TES

Jim Hasel-of f B. Sc. (Hons )

for Gene Technology Adelaide University Centre ' South Aus traÌ i a.

Thesis submltLed to the university of Adelaide in fulfil-lment of the requirements for the degree of Doctor of PhilosoPhY.

May, 1 983. CON TENTS

S TATEMEN T

A C KNOV\¡LED GEMEN TS

SUMMARY

CHA PTE R 1 TNTRODUCTION

A. Viroids 1 B. Vl rusoids 4 C. Aims 5

CHAPTER 2 RNA SESUENCE DETERMINATION

IN TROD U C TI ON I o MATERIALS

ME THOD S A. Isol-ation and sequence determination of linear viroid or virusoid fragments

A-1 AnalYtical RNase digests 9 A-2 PreParatlve RNase digests 11 A-3 5'-32P fabell-ing of RNA fragments 11 A-4 3'-32P l-abel1ing of RNA fragments i) Synthesis of Ir'-3'rldpcp 12 ri ) 3'-32P l-abeì,1i-ng 12

h-)AE Polyacrylamide geI fractlonation 13 A-6 Sequence determination of RNA fnagments using the partial enzymic cfeavage technique 14 A-7 Sequence determination of RNA fragments using the dideoxynucleotide chaln termj-nation technique i) PhosPhatase treatment 16 ii ) PolYadenYì-ation 16 ij-i) Reverse transcriPtion 1T A-B Bolyacnylamide gel elecLrophoresis 18 B. Sequence determination using cloned viroid or virusoid sequences B-1 Synthesls and cl-oning of viroid and

virusoid ds cDNA i) LinearizaLion and polyadenyl-ation 1g ii ) First strand cDNA sYnthesis 20 iii ) Second strand cDNA sYnthesis 20 iv ) Restrj-ctlon enzyme cleavage and fractionation 20

v) M1 3 cloning 21 B-2 Sequence determination of recombinant

phage M1 3 ¿t B-3 Sequence determination of RNAs using cfoned DNA Prlmers i) PneParation of Primer ¿¿ ii ) RNA-DNA hYbnidizatlon ¿J iii ) Reverse transcriPtion 23

RESULTS and DISCUSSION A. Techniques 24 B Partial- enzymic cleavage of viroid and virusold RNAs 24

c Partial enzymic cleavage of radiol abel-l-ed RNA fnagments 25

D Dideoxynucl-eotide chain termination sequencing of RNA fragments 25 ¿õ E Cloning of vinoid and virusoid sequences

F Sequence determination using cfoned viroid

or vinusoid sequences - 27

G Detenmination of complete vinoid or virusoid primanY structunes 2B

CHAPTER 3 CHRYSANTHEMUM STUNT V]ROID

INTRODUCTI ON 2g

MATER]ALS 30

ME THOD S A. Primary structure determination 30

B. Secondary structure determination 3',I

RESULTS A. Sequence determination 31 B. Primary sequence and secondary sLructure 33

DISCUSS]ON A. Homology between CSV and PSTV 33 B. Replication of CSV and PSTV 34 C. Rel-ationship of this isolate of CSV to other viroid isol-ates 36

CHAPTER 4 COCONUT CADANG-CADANG VIROID

INTRODUCT] ON 39 M ETHOD S A. Isolation of the ccRNAs 41 B. Sizing of the ccRNAs 42 C. Fingerprinting of the ccRNAs 43 D. Sequence and structure deLermination of

the ccRNAs 44

RESULTS ANd DTSCUSSION A. Sizing of the ccRNAs 46 B. Fingerprinting of the ccRNAs 47 C. Sequence and structure determination of

the ccRNAs 48 D. ccRNAs differ in size but not sequence complexitY 50 E. Varlation in sequence between different ccRNA isoÌates 52 F. Structural simitarities between ccRNAs and viroids 53

G . Re plication of ccRNAs 55

H . ccR NA s l-ow variants and the time course of infection 56

I. Origin of cadang-cadang disease 5B

CHAPTER 5 VELVET TOBACCO MOTTLE V]RUS AND SOLANUM

NODIFLORUM MOTTLE V]RUS

INTRODUCTION 60

MATERIALS and METHODS

A. Viruses and RNA 61 B. RNase Fingerprinting 62 C. BNA sequence analysis i-) Partial- enzYmic digestion 62 fj- ) Di deoxynucleotide chain termlnation 62 D. Synlhesis and cloning of double-strand cDNA 63

RESULTS

A. RNase fingerprints of VTMoV and SNMV RNA 2 63

B. Primary structures of VTMoV and SNMV RNA 2 64 C. Secondary structures of VTMoV and

SNMV RNA 2 66 D. Possibl-e polypeptide translation products from RNA 2 species and their compfements o/

DI SCUSS I ON 6B

CHAFTER 6 SUBTERRANEAN CLOVER MOTTLE VIRUS

INTRODUCTION 72

MATER]ALS 73

METHODS A. Synthesls and restriction endonucl-ease

cleavage of ds cDNA 73

B. Fingerprlnting of SCMoV RNAs (¿1 C. Sequence determlnation of SCMoV RNA 2 and

RNA 2I 74

RESU LT S

A Analysis of SCMoV RNA 1 nucleotide

sequences 75

B RNase fingerprinting of SCMoV RNAs 2 and

FNAs 2t 76

C Sequence determination of SCMoV-A RNA 2 and RNA 2I 7B

DI SCUSSI ON A. RelationshiPs between the v'arious isolates

of SCMoV 79 B. Sequence homology between SCMoV RNA 2 and RNA 2I 79 C. Sequence homologY between SCMoV, VTMOV,

SNMV and LTSV RNAs 2 BO D. Satellite RNA of TobRV B2 E.'sequence homoì-ogY between TobRV s atellite RNA and vlrusoids B4

CHAPTER 7 VIROIDS VIRUSOIDS AND SATELLITES

INTRODUCTION B6

ME THOD S

A. tsolation of RNA B7 B. Blot hybridization B7

RESULTS A. Analysls of VTMoV and SNMV RNA 2 sequences present in vinus and infected tissues 8B

DI SCUSS I ON

A. Multimers of VTMoV and SNMV RNA 2 8,9 B. A possible site for RNA processing 90 C. Vlroid, vlrusoid and satetlite RNAs 92 STAT EMEN T

This thesis has not previously been submibted for an academic award at this or any other University, and is the original work of the author, except where due reference is made in the tex t .

JIM HASELOFF AC KN Ol/tr LED GEMENTS

I wish to thank Prof . VÙ. H. El1iot for permission to underLake these studies in the Department. I aÌso wlsh to thank my supervisor, Bob Syrnons for the advice and support provlded to me during the course of this wonk. In addition, I wish to express my appreciation to the following people: Dr . Pe ter Palukaitis , for purified CSV and arouslng my interest in viroids; Drs. Richard Fnanckl and John RandIes, for heJ-pf ul- discussions and interesting vinuses; Dr. Nizar Mohamed, Julita Imperial and Judith Rodrigue z, for providing ample amounts of ccRNAs; Dr. George BrueninS, f or unpubl-ished resul-ts and stimulaLing discussion 1n the J-ab; Dr. Detl-ev Riesner' f or providing unpubl-ished resufts and eintoPf; Dr. Al,l-an Gould, f or his help and dangerous sense of humour; Karf Gordon, for his red wine and socj-o-politics (more fike rosè), and scientific discussion; My fel-1ow viroid/virusoj-d infected co-workers, Dn. Peter Murphy, Jane Visvader and Paul- Keese, as weIf, as the other numerous faboratory and departmenlal lnmates; Jenny Rosey, Sharon Freund and Lisa Waters, for exceflent technical assj-stance and preparation of the figures for thls thesis; Mrs. To, for her care in typing this thesis; and finally my family, for putting up with this student. SUMMARY The work described in this thesis coticerrls the establishment and application of technj-ques for the rapid sequence determination of small circular RNAs such as those of viroids and virusoids. The determined sequences of chrysanthemum stunt viroid, the vari-ant RNAs of coconut cadang-cadang viroid and the virusoids of velvet tobacco mottle virus, solanum nodiflorum mottl-e virus and subterranean clover mottle virus are presented. The overall conclusions from the work are outlined briefly below. 1. Viroids contain highly conserved sequences central to their rod-like native structures. 2. Virusoids also contain highly conserved. sequences central to their rod-like native structures and share the pentanucleotid.e sequence GAAAC with that of viroids. 3. In addition the conserved. seguences of vírusoids are shared by the linear Rr\A of tobacco ringspot virus. Presumably the coInmon sequences of each cl-ass of

RIIAs reflect common function, and perhaps suggest some functional similarity between viroids and virusoids.

Sequence homology between virusoids and the satellite RNA of tobacco ringspot virus al-Iows pred.iction of sites for processing of these RNAs from multimeric RIJA intermediates of replication. CHAPTER 1

TNTRODUCTION A. Viroids Viroids constitute a unique class of infectious plant pathogens, and as such are a fairly recent dicovery. The viroid concept vJas first recognized when the infectlous agents of the spindJ-e tuber disease of potato (Diener: and Raymer, 1967 ) and the exocortis disease of citrus (Semancik and lltleathers,

1968 ) , which vrere thought to be viruses, I^i ere shown to possess unusual propertles : ( 1 ) phenol or other organic sofvents had no effect upon the lnfectivity of buffered extracts from lnfected plants; (2) no virus particles coul-d be isolated or dernonstrated by el-ectron microscopy; ( 3 ) the infectious agents displayed a resistance to nucfeases and an elutlon profile off cel-Ìufose columns simil-ar to double-stranded RNA; and (4 ) the inf ectious agents I^/ere always present in high speed supernatants, possessing sedimentation coefficients of 1 0- 1 5S. As a more detailed knowl-edge of the sizes and physlco-chemicaf properties of these two disease agents became availabl-e (Raymer and Diener' 1969; Diener and Raymer, 1969; Semancik and Weathers 1972a) , it became apparent that these two agents were the first of a new cfass of infectious nucÌeic acids (Diener, 1971b; Semanclk and Vrleathers, 1972b; Sånger, 1972); the Lerm virold I^Ias proposed (Diener, 1971b), and L the causal agents were renamed potato spindle tuber viroid ( pSfV ) and citrus exocortls viroid ( CEV ) . Since that t1me, viroids have been shown Lo be the causative agents of a further eight plant diseases, âhd are listed in Table 1 -1 . These viroids consj-st of infectious l-ow mol-ecul-ar weight RNA species which are unencapsidated, sì-ngJ-e sLranded, coval-ent1y-closed circul-ar moleçul-es with a hì-gh degree of intramol-ecul-ar base-pairing (Diener, 1972; Semancik et â1., 1975; Sången et âf., 1976). Physj-co-chemica] sludies of several- viroids (Henco et â1 . , 1977; KJ-ump et âf., 197B; Langowski et â1., 1978; Domdey et âf., 1978) cul-minated in a model fon viroid structure in which the clrcufar RNA mofecuÌes exist as extended rod-l-ike structures, characterized by series of base-paired sections interspersed with single-stranded loop sections. Determination of the compl-ete nucl-eotide sequence of PSTV ( Gross et â1. , 1 978) , together with

dye-binding expeniments ( Riesner et â1 . , 1 979 ) and

tRNA-binding experiments (v'lild eL â1. , 1 980 ) established the validity of this model. Apart from their unique structures ' the singJ-e feature which distinguishes viroids flrom vlruses is their apparent Ìack of encoded polypeptlde products. Viroids RNAs appear to be naked with no associated protein (Diener, 1971a; Semancik and V'leathers, 1972a), Table 1 -1 Viroids that are presentlv knownl

Viroid References

1 potalo spindle tuber viroid ( PSTV ) (Diener, 1977 )

¿ citrus exocortis ( Semancik and viroid (CEV) Íieathers, 1972; Sänger, 1972)

3 chrysanthemum stunt ( Di ener and Lawson , vinoid (CSV) 1973)

4 chrysanthemum chlorotic ( Romaine and Horst, mottle virold ( ChCMV ) 1975)

5 cucumber pale frult (Van Dorst and vi roid ( CPFV ) Peters, 1974)

6 coconut cadang-cadang viroid ( cccv ) (Randl-es, 1975) 7. hop stunt viroid (HSV) (Sasaki and Shikata, 1977 )

ö avocado sunbl-otch ( Thomas and viroid ( ASBV ) Mohamed, 1979)

9 tomato bunchy top vi roi d (TBTV) (Vlalter, 1981 )

'l 0. Lomato planta macho ( Galindo et âf . , vi-roid ( TPMV ) 1 982)

Recent studies indicate that the causative agent of burdock stunt disease may possess propertles atypical of those of other viroids (Chen Weì and 'T j-en Po, personal communication), and has been tenatively omltted from the list. 3 and no viroid-coded transfation products have been found either in vivo ( Conjero and Semancik, 1977; FIores et â1., 1978; Conjero et af. , 197g; Camacho and Sången,

1 982a, b ) or ln vitro ( Davies et â1 . , 1 97 4; Semancik et

âf . , 1 977 \ . If viroids do not encode functional polypeptide translation products, they must rely entirely on plant host cefl- components for their replicati-on. Vlnold repl-icalion has been shown to be inhibited by actinomycin D, inhibiting DNA-dependent RNA synthesis (Diener and smith, 1975; Takahashi and Diener, 1975; Uünlbach and Sänger, 197g), and o'-amanitin aL concentrations which inhibit RNA poJ-ymerase II, and thus mRNA synthesis (Mühlbach and Sånger, 197g)- Whil-e the effects of these drugs on vlroid replication may be indirect, due to general effects on host cell metaboJ_ism, some evldence suggests that DNA-dependent RNA polymerase If, the target for o-amanitinr mâY play a direct rofe in viroid replicatlon. Rackwitz et al.

( 1 981 ) have shown that RNA pol_ymerase If purified from healthy plant tissues is capable of the o'-amanitln sensitive transcription of full,-length llnear complementany viroid RNAs from viroid template in vltro. However, the in vitro transcription of viroid templates has also been shown for RNA-dependent RNA polymerases isolated from healthy plant tissue ( Boege et a1. ,1 982 ) 4 and from cucumber mosaic virus infected plant tissue

( D. S. Gllt and R. H Symons, unpubl-ished resutts ) ' So, whil-e the exact naLure ofl the enzymes involved in viroid replication in vivo remains uncfear, the f oll_owing detail-s are known. (t ) virolds appear to repJ-icate through compfementary RNA lntermediates ( Grill and Semancik, 197B; Zaitl-in et âl ., 1980; Hadidi et â1.,

1 981; ZeIcer et â1., 1982). (2) Longer than unit-length complenentary viroid RNA intermediates have been detected in viroid infected tissue extracts ( Bnanch et I Owens and Diener â1. , 1 981 ; Rohde anO Sånger, 981 ; ' 1gB2; Bruening et â1. , 1982 ) (3 ) 0tigomeric series of RNAs of avocado sunbl0tch vlroid (ASBV) have been detected in infected avocado tissue (Bruening et al., 1g82). Various workers have therefore postulated roll-ing circle mechanisms for the transcription of larger than unit length viroid intermediates (Branch et â1., 1gB2; Owens and Dlener, 1982; Bruening a! âI',

1 982 ) . Such mechanisms require that unit-length Iinear viroid, produced by either specific tnanscniption or cleavage of ol-igomeric RNAs, b€ ligated to produce the final circular Product.

B. Virusoids Four members of a new and unique group of plant viruses have been recently described ( Randles et 5

âI., 1981; Gould and Hatta, 1981; Tien-Po et â1., 1981 ) '

These vlruses, velvet tobacco mottle virus ( VTMoV ) , sofanum nodiflorum motLte virus ( SNMV ) , lucerne transient streak virus (LTSV) and subterranean clover mottl-e virus (SCMoV ) were isolated in AusLral-asia (see Figure 1-1 ) and each consist of 30 nm polyhedral- capsids containing two maior slngle-strand RNA species. RNA 1 is a l-inear molecule of about 4,500 res j-dues, whereas RNA 2 is a circufar covalently closed mol-ecul-e of 300-400 residues with a high degree of internal- base-pairlng. The RNA 2 molecules therefore possess physical properties simll-ar to those of vlrolds and have been termed virusoids.

Gould et aÌ. ( 1 981 ) have shown that both RNA 'l and RNA 2 of VTMoV and SNMV are required for viral infection and Lhat therefore these vi-rusoid molecules contribute some function essential- for replication. In

contras t , Jones et al-. (1983) have shown that the virusoid of LTSV is apparently not required for viral- lnfection and that it therefore behaves as a satell-ite

RNA. The nature of the refatj-onship between the RNA

components of SCMoV is unknown.

C. AIMS Three unresofved questions stand behind the work described in this thesis.

ORIGIN OF VIRUSES CONTAIN¡NG VIRUSOIDS

Northern

Territory Queensland

Western Austalia Brisbane South Australia

New South Wales Perth e Sydney

Adelaide Vict

Melbourne New Zealand o WMoV SNMV Tasmania LTSV o SCMOV 6

(1) How do viroids nepl-icate ?

(2) Vlhat function ( if any ) do virusoids contribute to

vi rus replication ?

(3 ) Does the replication of viroids and vinusoids share

common features ? In order to address these Seneral- questions of virold and virusoid function, the primary and secondary structures of à number of these molecul-es v'iere determined. TL was hoped that the location of regions essential for function within these molecules woufd be mirroned by the presence of conserved nucfeotide sequences an d/ or structures. The various techniques required for viro id/v lnusoid sequence determinatlon are descrlbed in lhe following chapten. Chapters 3, 4,5 and 6 describe varlous apptications of these techniques, and the final- chapter deals with the overall concÌusions from the

I^IOf k. CHAPTER 2

RNA SEOUENCE DETERMINATION 7

] N TRODU CT I ON The techniques availabte for lhe rapid sequence determination of RNA rely on the presence of a fixedreferencepointwithintheRNA,eg'auniquesite for primed synthesls of transcnipts (Zlmmern and Kaesberg, 1978 Symons' 1978 ) or a 5t or 3' radiolabeÌIed terminus ( Donis-Kel-fer et âI' , 1977 ) ' and therefore these techniques ane i-deaJ-1y applied to ri-near RNAs.ModifiedapproacheSmuStbeusedfortheSequenCe determination of circul-ar RNAs, and the approaches used in this work fall into two classes' First, by exploitlng the base-paired nature ofviroidsandvirusoids,sPecificllnearRNAfragments may be produced by partial RNase cleavage of the circufarmolecu]-es.Thelinearvj.roidorvirusoid fragmentsmayberadlo]-abelled,purifledandSequenced using the partlal enzymic dlgestion ( Donis-KefIer eL àI ., 1977) or dideoxynucleotide chain Lermination (Zimmern and Kaesberg, 19781. Symons, 1978 ) sequencing techniques. Second, double-strand cDNA can be transcribedfnomfinearizedviroidorvirusoidRNA' inserted and propagated in a bacteriophage M1 3 vector' and the inserted DNA sequenced directJ-y (Sanger et â1., to prime 1 980 ) o r excised or transcrlbed and used dideoXynucfeotldechaintermj.nationSequencingofthe B ci-rcular viroi-d or virusoid RNA (Zimmern and K.aesbeng, lgTB; Symons, 1978\ . The details of these techniques are given below.

MATERIALS RNases A and T1, cal-f lntestinal alkaline phosphatase, E. col j- tRNA, deoxynucl-eoside triphosphates and isopropylthiogaì-actoside ( IPTG ) were obtained from the Sigma Chemj-caf Co. T,, DNA ligase and T + pol-ynucleotlde kinase, Kl-enow f ragment of E.col-i DNA poJ-ymerase were obtained from Boehringer. T, RNA ligase, dj-deoxynucfeoside 4

trlphosphates and d ( TBC ) idere ob tained f rom P L Biochemicals. M13 sPecif ic 17-mer primer and a1l restrictlon endonucleases htere obtalned from New Engl-and Biolabs. RNas" U2I^Ias obtained from Sankyo.

5-bromo -4-chloro-3 - indoyl-gal- ac tos i de ( BCIG ) bras obtained from Bethesda Research Laboratories. Avian myelobfastosis virus reverse

transcriptase v,/as obtained f rom Lif e Science Inc. , Petersburg, Fl-orida. lo-3zplocrp and lo-32lloRtt, at specific

activities of 1 0OO Ci /mmol , and I v -32p] nrP at a 9 speciflc activitY of 2000 Cilmmol- v,/ere prepared bY Dr. R.H. Symons as PreviouslY described (Symons, 1977; 1981 ). E.coli pol-y (A ) polymerase was purif ied according to Sippel (1973). Phy M RNase !\¡as prepared

( Donis-Ke1Ìer, 1 980 ) from culture supernatants of

Phv s ar um oJ-yce halum, the inocul-um of which was kindJ-y provided by the school- of Biological Sciences, Flinders University of South Australia. The extracelful-ar RNase of Bacill-us cereus r^¡as prepared as descrlbed by Lockard et al. ( 1 978 ) from a cufture supplied by Dr ' G Brownfee.

METHODS A. Isolation and seq uence determination of linear viroid or virusoid fragments A-'l AnaÌytical RNase diges ts Circular RNA vras digested with a range of RNase concentrations to obtain optimal conditions for the production of specific finear fragments. Flve atiquots, each of 0.1 ug to 0.5 ug purified circufar RNA were resuspended in 9 u1 of the appropriate high salt RNase digestion buffer (600rnM NaCl, 10rnM MgCl, wiLh

either 2OmM sodium citnate pH 3.5 for RNas. U2, otr 2OmM

Tris-HCl- pH 7.5 for RNase T or RNase A digestions) and

pl-aced on ice. 1 U1 of 100,000 units/m1 RNas" Tt , 10

I mglm1 RNase A or 1OO unlts/ml- RNas. U 2 v{as added to one of the tubes, which was to contain the highest concentration of RNase. The tube conLents \^/ene mixed and 1 u1 removed to a second tube; this v¡as f oll-owed subsequently by another two similar 1 0-fo1d RNase dilutions. No RNase was added to the fifth tube. For example, if RNase T., I^/as used, the f ive tubes would

1 'l O and 0 units/ml RNase T', contain 10, OOO, ,000 ' 1OO, respectively . After incubation at 0 o c for 60 minutes,

the digestions vrere terminated by extractlon with 1 00 ¡r 1 water saturated phenol : chl-oroform (l : 1 ) and 100 pI 0.2M NaCl, 0.1mM NaTEDTA; the aqueous phases were removed, washed twice with ether and precipated with 3 vol_umes of ethanol. After 20 minutes at -B0oc, the tubes \¡,rere centrifuged at lo,ooog for 15 minutes aL 4oc, the Supernatants I^Iere removed, and the precip j-tates v¡ene

dried in vacuo with 1 O P C1 lr-"Pl nrP. The dried pellets conlainlng RNA fragments ')a and Iv -"pl RrP were resuspended in 5 u L,o, heated aL

BOo C f or 1 rninute, snap cooled on ice, and 1 U I of 5x pol-ynucleotide kinase buf f er ('l 25mM Tris-HCl pH 9'0,

5OmM MgC1r, 5OmM DTT) added. O.25 units of polynucleotide kinase was added, and the reaction was incubated aL 37"C for 30 minutes. 5 Ut of formamide loading buffer (95% deionj-zed fonmamlde, 'l 0mM EDTA, O.O2% xylene cyanol FF, O.O2% bromophenol blue) was 11 added, the tubes v\¡ere heated aL B0o c f or 1 minute, snap cool-ed on ice, and l-oaded onLo a 20x40x0.05 cm 6% polyacnyl_amide geI containing TBE (9Oml¡ Tris-borate pH 8.3, 2mM NaTEDTA) and 7M urea (Sanger and Coulson,

1978 ) . Af ter el-ectrophoresis, the gel was autoradiographed.

A-2 Preparative RNase digests

Purlfied circular RNA (2 to 20 Ug) v, as resuspended in 50 p1 of RNase T., and or RNase U Z high sal-t digestion buffer (see above) and cooled on 1ce. The appropriate amount of RNase Tt, RNase A or RNase U 2' as determined by analytical ribonucl-ease dlgesLs, h/as added and incubation continued for 60 minutes at 0oC.

Digestions hlere terminated by extraction with 1 50 pl water saturated phenol : chloroform (l : 1) and 100 pJ-

H^0. The aqueous phase was removed, washed twice wlth 1 ¿ mf ether, and 450 pl ethanol I^ras added: Af ter 20 minutes aL -BOo C the sample I^ras centrlf uged at 1 0,000g for 15 minutes at 4"C and the supernatant discarded. The precipitated RNA fragments coul-d be either 5' or 3 I radiol-abell-ed.

J¿ A-3 5 ' - P-labe11ing of RNA fragments The pellet, with 20O uCi of added ty-32PlATP

(2OOO Cilmmol), b/as dried in vacuo' resuspended in B U1 12

1.5mM spermidine' heated aL B0oC for 1 minute and snaP

on ice. 2 pl of 5x 1, pol-ynucf eotide kinase cool-ed ¿+

( (4 T, polYnucleotide buffer see above ) and 1 Ul units) 4 kinase v,rere added and the reaction incubated al 37oC for 30 minutes.

32 A-4 3 r - P-Iabel-l-ing of RNA fragments i) Synthesis of Ir' -32PTopcp 5oo uci Iv-32P -]nrP (2000 Ci /mmol ) vras dnied in vacuo, resuspended in 5 Ul ,aO, and 2 VI'l 0 mg/m\ 3r dcMP, 2 Ul 5x Tq poÌYnucleotide kinase buffer and 1 Ul

r¡i ( ) T, poJ-ynucleotide klnase ere added. The 4 unlts 4 reaction r^ras incubated aL 37oC for 30 minutes, heaLed at 90oC for -l minute and stored aL -15" C bef ore use.

32 ii ) 3 ' - P-labelJ.ing The preclpitated RNA fragments were dried in vacuo, Fe spended in 20 Ul i0mM Tris-HCl pH 9.0 I E containing 0.01 units cal-f intes tinaf al-kali-ne phosphatase, and incubated aL 500c for 20 minutes. The

reacti-ons I^Iere then extracted wi th 1 00 Ul water

saturated phenol- : chl-oroform (r : 1 ) and 1 00 Ul 0.2M NaCI, 0.1 mM NaTEDTA. The aqueous phase was removed, washed twice with'1 mf ether and the RNA precipitated with 450 Ul ethanof at -80oC for 20 minutes. The reaction tubes vüere centrifuged at 1 0,0009 for 15 13 minutes aL 4"C and the supernatant hlas discarded. The precipitated r phosphatase treated RNA fragments were dried in vacuo, tr€suspended in 5 Ul ,rO, heated at B0oC f or one minute and snap cooled on j-ce. 1 U1 of l¡'-ttt,ldpCp (50 uCi), 6 ¡r1 of 2x T,t RNA ligase buffer

('1 OOmM HEPES pH 7 .5, 6.6mM DTT, 3OmM MgCl, , 20% (v/v ) redistlll-ed DMSO, 100 UM ATP) and 1 U1 Tq RNA J-igase

( 4 .6 units, 1 .5 Ug ) vüere added, and the reaction v{as incubated aL 4"C for 16 hours.

A-5 Polyacrvl-amide se1 frac tionation

1 O pl of formamide loading buffer (95% deionized formamlde, 10mM NaTEDTA, O.O2% bromophenol blue, O,O2% xyl,ene cyanol FF) was added to each reactlon mixture containing 5t - or 3 I - radiofabelfed RNA fragments. The reaction mixtures vùere heated at B0oC for 3O seconds, snap cool-ed on ice and Ioaded onto an

80x20x0.05 cm 6% polyacrylamide gel containing 90mM Tris-borate pH 8.3, 2nY1 EDTA and 7M urea. After el-ectrophoresis for 6 hours aL 25 fiA, the gel was autoradlographed aL room temperature for 5-30 mlnutes and the resultant autoradiograph used as a template to focate and excise the 32p-l-abel-led fragmenLs. Excised bands v,rere eluted by soaking overnight aL room

tempenature in 500 pì- of 500mM ammonium acetate, 1mM Na^EDTA, 0.1% SDS, whlch contained 60 E.coli tRNA as ¿ ug 14 canrier if the fragments \^Iere to be Sequenced using the partial_ enzymic cleavage technique . After soaking, the efutlon buffer i.Jas removed and the RNA fra6Ements were precipitated by the addltion of 1 mI ethanol and storage aL -80oC for at f east 30 minutes. After centrlfugation aL 10,0oog for 15 minutes at 4oc, the pelleted fragments v,rere resuspended in 'l OO pJ- 0.2M NaCt, 0.1mM NaTEDTA and re-ethanol precipitated with 300 pI ethanol-. After centrif ugation the pellets i^¡ere dried in vacuo '

Punlf ied 5t o r 3 | radiof abel-led RNA f ragments \^Iere used for sequence determinatlon by elther the partial enzymic cleavage on dideoxynucfeotide chain termination technì-ques.

A-6 Sequence determinalion of RNA fragments using the partial enzvmic cleavage technique 32 5r- or )l P-labelfed RNA fnagments, with 60 UC E.coli LRNA, were resuspended in 12 pf HZO and six

aliquots of 'l u I dispensed . The f ollowing buf f ers b¡ere added to the six tubes. Tube N (No enzyme) 9 Ul 20mM Na citrate pH 5.0

1 mM NaTEDTA

7M urea Tube T (RNase T") 9 ul 20mM Na citrate pH 5.0

1 mM NaTEDTA

7M urea 15

( 20mM Na cltrate pH 3.5 Tube U RNase 'z) 9 Ul 1mM NaTEDTA

714 urea

Tube L (alka]i ladder) 5 ut 50mM NarCO, / NaHCO, pH 9.0 Tube D (RNase PhyM) 9 yl 20mM Na citrate pH 5.0 'lmM NaTEDTA

7M urea

Tube B (BaciÌ1us cereus RNase ) 5 ill 20mM Na citrate pH 5.0

1mM Na ED TA a Tubes N, T, U, P, and B were heated at B0oC for I minute, snap cool-ed on ice, and the ribonucleases added.

Tube N

Tube T 1 U1 10,000 units/mL RNase T.,

U 'l RNase ïl Tube Ul 5 units/ml- ¿

Tube L 'l Tube P UÌ RNase PhyM extract

Tube B pl B ceneus RNase extract

Tube L hlas heated at 1 00oC for 90 seconds while the remalning tubes \^Iere incubated at 5 0o C f or 20 minutes . At compfetion of the seQBencing reactions, the tubes v\,ere stored at -B0o C or were kept on ice while being prepared lmmedlatel-y f or ef ectrophores j-s. Bef ore polyacryl-amide gef electrophoresis, formamide l-oading buf f er (95% de j-on ized formamide, 1OmM EDTA , O.02% (w/v ) 16 bromophenol blue, O.02% (w/v ) xylene cyanol FF) vlas added to the samples to a final volume of 12 pl. Samples were heated aL B0oC for 1 minute and snap cool-ed on ice before e l- e c t n o P h o r e s i s .

A-7 Sequence de termination of RNA fragments us ing the dideoxynucleotide chain termination technique i ) Phosphatase treatment Purif ied 5'-32p-l,abell-ed RNA f ragments, obtalned by RNase T.'| digestion, were each resuspended in 'l 20 uI 5OmM Tris-Hcl pH 9.0, heated at B0oc f or minute and snap cool-ed on ice; 1 Ul ( 0.0'l units ) cal-f intestinal afkal-ine phosphatase was added, and the reactions incubated at 50oC for 2O minutes. The reactlons were extracted with 1 00 ul water saturated phenol : chforoform (l : 'l ) and 100 Ul 0.2M NaCl, 0'imM EDTA; the aqueous phases were washed twice wlth 1 ml ether and the fragments precipitated wlth 300 Ul ethanol aL -BOoC for 20 minutes. After centrlfugation at 10'000g for 15 minutes at 4oC, the pellets b¡ene dried in vacuo. f i ) PoJ-yadenYlation Phosphratase treated RNA fragments were resuspended in 1o uI water, heated at B0oc for 1 minute

2mM ATP 4 u l- of and snap coof ed on ice. 2 VI of ' 5x E.col-i poly(A) polymerase reaction rnixture, comprising 105 U1 H.O, 50 U1 1M Tnis-HCI pH 7.9, 25 pI 17

O.1M MnClr, 'l O Ul 1M MgCJ-Z and 1O pI O'1M DTT, and 6 p1 of E. col-i poly (A ) poJ-ymerase extract were added, and the reaction incubated aL 37"C for 60 minutes' The reactions vüere extracted with 1oo u1 water saturated phenoJ- : chloroform (l : 1 ) and 100 u] 0.2M NaCl, 0.1mM EDTA' washed twice with 1 ml ether and precipitated with 300 Ul ethanol at -BOoC for 2O minutes' After centrifugation at lO,OOOg for 15 minutes aL 4oC, the pell-els were dried in vacuo. fii) Reverse transcriPtion Polyadeny]-atedRNAfragmentswereresuspended 1O H,O with '1 0.25 mglml dT8C, heated at B0oC in ilr ¿ ul and al-f owed to cool- to room ternperature over 20 minutes. 2.5 U1 aliquots \^Iere dlspensed into f our tubes to give reaction mixtures of 5 Ul, contalning 5OmM Tris-HCl pH 8.3, 5OmM KCI-, BmM MgClr, 1OmM DTT, 2 unlts of avian myelobl-astos j-s virus reverse transcriptase, 2 yCi I o-"p_loRtp or dcTP, 5o irM of the remaining deoxynucl-eoside triphosphates and a single dideoxynucl-eoside triphosphate species essentially as described by Symons ( 1 978 ) . Re actions blere incubated at 3ZoC for 30 minutes; 5 Ul of formamlde loading buffer (95% deionized formamide, 'l OmM NaTEDTA, O-02% (w/v) bromophenol b1ue, O.O2%(w/v) xylene cyanoÌ FF) hlas added to each reaction, âhd the tubes vJere heated at 1 00oC for 2 ninutes and snap cool-ed on ice before 1B

e 1 e c t r o p h o r e s i s .

A-B Pol-y acnyÌamide gel efectrophoresis Radio]-abel]-edproductsofthepartiaJ-enzymic cleavage and dideoxynucfeotide chain terrnination technlques vJere fractionated by el-ectrophoresis in BOxZOx0.O5 cm B% polyacrylamide gels or 4Ox20x0'05 cm

20% polyacrylamide gels containing 90mM Tris-borate pH 8.3, 2mM NaTEDTA and 7M urea (Sanger and Coulson' 1978). The sequence determination of some RNA fragments\dascompl-icatedbythepresenceofband compression artif acts (Kramer and MiJ-l-s, 1978 ) arising from incompJ-ete denaluration of RNA or cDNA fragments during electrophoresis. In order to el-iminate band compression, f ragmenLs \^lere f ractionated in

potyacryl-amide geJ-s conLaining 98% f ormamlde ' Sequenclng reaction mÌxtures were

preclpitated by adding 'l 0O p I O ' 2M NaCl, 0 ' 1mM NaTEDTA and 300 Ul ethanol, and the sampl-es I^Iere kept at -B0oC for 20 minutes, centrifuged at 1 0,000g for 15 minutes aL 40c, and the peJ-lets dried 1n vacuo. The samples were resuspended 1n 5 ul formamide loading buffer, heated at SOoC for 1 minute, cooled on ice and loaded onto a 40x20x0.05 cm 20% polyacryfamide 8el contaì-ning 98%

1 6mM NarHPOO AmM NaH dò formamide buffered with ' zPo,-r 19 described by ManiaLis and Efstratiadis ( 1 980 ) ' Following electrophoresis, gels b¡ere autoradiographed aL

-BOo c, using cal_cium tungstate inLensifying screens.

B. Sequence determination using cl-oned viroid or vírusoid sequences

B- 1 Syn thes is and cloning of vinoid and vinusoid

ds cDNA i) Llnearization and polyadenylation Purified cj-rcul-ar RNA (2-10 Ug) in 10 U1

distilled water v,¡as heated at 1 00o C f or 30 minutes in a seal-ed capil-Iary to Élenerate randomfy cleaved full-length linean molecules. As described above for radiol-abeffed RNA fragments in A-7 ( i ) , terminaÌ 2, (3 t )-phosphate groups vüere removed from the cleaved molecules by the addition of Tnis-HCl pH 9.0 to 50mM and

O. O1 units calf intestinaf alkaline phosphatase followed by incubation aL 50oc for 20 minutes. Reactions I^Iene phenol-chl-oroform extracted, twice ether washed and ethanol precipitated. Phosphatase treated RNA mofecufes I^/ere resuspended in 50 U1 water, heated aL B0oC for 1 minut,e and snap coofed on ice; 10 ul of 2mM ATP, 20 pJ- of 5x E.col-i poly(A) pol-ymerase neaction mixLure (see A-7 (i j- ) ) , âhd 30 U] E. coÌi poly (A ) pol-ymerase extract r^rere added and the reaction incubated at 37"C for 60 20 minutes. Reactions vJene extracted with 100 pJ- water satunated phenol- : chloroform, twlce ether washed and precipitated with 3 volumes of ethanol-. ii ) First strand cDNA sYnthesis

Polyadenylated RNA vüas resuspended in a 5O p1 reaction mixture containing 50mM Tris-HC1 pH 8.3, 50mM

KC1, 1OmM MgCJ-.,,'l OmM DTT, lrnM each of dATP, dGTP and drrP, zoo uM Io-32r-locrr (60 uCi), 0.6 ug (dT)to and 30 units avian myel-oblastosis virus reverse transcriptase, and incubated at 42oC fon 60 minutes. The reaction mixture r^ras heated aL 100oC for 1 minute, snap cool-ed on ice, 'l Ul (10 Ug) heat-treated RNase A added and incubation continued at 50oC fon 20 minutes. Reactlons hrene extracted with 'l 00 Ul 0.2M NaCl, 0.1mM NaTEDTA and

1 00 U1 water saturated phenol : chl-oroform (t : 1), twice ether washed and precipitat,ed with 3 vol-umes of ethanol. i j-i ) Second strand cDNA synthesis Si-ng1e strand cDNA v¡as resuspended in 10 Ul H^0, boiled for 1 minute and snap cool-ed on ice, added ¿ to make up a 25 Ul reaction mixture containing 5OmM Tris-HCl pH 7.5, 1OmM MgClr, 10mM DTT, 1mM dATP, dCTP, dGTP and dTTP and 2 units Klenow fragment of E.coli DNA polymerase 1, incubated for 4 hours at 37oC and kept aL

-2Oo C b efore use . j-v ) Restriction enzyme cl-eavage and f ractionation 21

Doubl-e strand cDNA I^/as subj ected to digestion by various restriction endonucleases under conditions as specif ied by t,he suppl-iers of the enzymes. The cleaved ds cDNA bras fractionated by electnophoresis in a 20x40x0.05 cm 6% polyacrylamide gel (Sanger and Coulson,

1977 \ containing TBE buffen ( see above ) and 2M urea. Following electrophoresis, the gel vlas autoradiognaphed, the ds cDNA fragments ütere exclsed and each eÌuted in 400 ut 0.5M ammonium acetate, 0.1% SDS, 0.1mM NaTEDTA aL room temperature overnight, âhd ethanol precipiated tw1ce. v) M1 3 cloning Purified ds cDNA fnagrnents r^/ere llgated into an appropriate restriction site of the repllcative form of the phage M13 mp7 using phage Tr* DNA J-igase (Goodman and MacDonald, 1979; Messing, Crea and Seeburg, 1 981 ). Further speciflc details are provided in fol-lowing chapters. Ligated M13 RF and ds cDNA uras used to transform competent E. coli JMl 01 and the celfs vùere plated on agar media with BCIG ( 5-bromo-4-chl-oro

-3- indoy I-P-D - gal ac tos ide ) and I PTG ( Is opropyJ- -p

-D-thlogalactopynanoside ) . Recornbinant M1 3 phage v\rere screened by sequencing as described bel-ow.

B-2 Sequence deterrnination of recombinant phage M13 M13 phage containing cl-oned ds cDNA inserts 22 r^iere selected, âs judged by insertional- inactivation of ß-galactosidase activity (Messing, Crea and SeebüPBr

1 981 ), âtrd the inserted sequences hrere determined using the dideoxynucleotide chain termination technique as described by Sanger et al-. (1980 ) with a 17-mer M'l 3 specific oligonucl-eotide primer ( GTA, CGACGZC2AGT ) .

B-3 Sequence determination of RNAs usln cl-oned DNA pr]-mers i ) Preparation of primer The DNA primers used in this work were of two types, being either fragments restricted from recombinant phage M1 3 RF DNA or transcribed using recombinant phage M13 ss DNA as a tempfate. In the former case, recombinant M1 3 RF \^/as isol-ated using the method of Birnboim and DoJ-y (1979) , restricted with the appropriate enzyme, âhd the fragment ( s ) containing vlrold or virusoid sequences purifed by efectrophoresis through a 6% polyacrylamide gel containing 7M urea (Sanger and Coul-son, 1978). In the l-atter case, recombinant M1 3 ss DNA containing viroid or virusoid sequences of the same polarity as the RNA sequence r^ras transcrlbed uslng an M1 3 specific ol-igonucfeotide primen and the Klenow fragment of E.coli DNA polymerase 1 with lo-ttplocrp and l-o-32plonrp (specific activities of

1 000 Ci /mmol ) essentially as described by Bruening et 23 a1 . (1982). The resulting partially double-stranded DNA mol-ecules v,rene sub j ected lo restnictlon enzyme digestion, âtrd the l-abelled fragments fractionated by polyacrylamide gel el-ectnophoresis. In both cases the purlfled DNA primers were eluted from polyacrylamide gels by soaklng (Maxam and Gil-bert, 1980 ) . ii) RNA-DNA hYbnidization RNA-DNA hybnids were prepared as f ol-lo\^IS. The purified restriction fragments and 1 to 2 ve of the appnopniate RNA were resuspended in 25 ¡rI of 0.18M NaCl, lOmM Tris-HCl pH 7.0, 1mM EDTA, 0.05% SDS, heated aL 1000c for 2 minutes, and incubated aL 600c for 2 hours. The RNA-DNA hybrids I^Iere twice ethanol precipitated and dried ln vacuo.

1]-IJ Reverse transcriPtion The RNA-DNA hybrids hiere reverse transcribed in the presence of dideoxynucleoside triphosphates essentiaJ-1y as descnibed above in A-7 (ii j- ) . However, if 1t "P-radiof abef f ed transcripts of recombinant M1 3 ss DNA \^rere used as primens, Io-32p]ONTPs I,\,ere omitted and replaced by the unl-abefled dNTP species. Revense transcripts v/ere fractionated by polyacrylamide gel electrophoresis as described above in A-8. Al l experimen ts invol-ving the use of recombinant DNA molecul-es were performed within safety guidelines set out by the Austrafian Academy of Science 24

Committee on Recombinant DNA mofecul-es (ASCORD).

RESULTS AND DISCUSSION A. Techniques The various rapid geI sequencing techniques used for the determination of viroid and vÍnusoid sequences are outlined in Figure 2-1. These techniques share the advantage of requiring onl-y smafl amounts of purified RNA for thelr use and, while there are disadvantages inherent 1n the use of each procedure, the combined use of different techniques alfows rapld and reLiabfe RNA s equence anal-ysis. The various approaches are revlewed bnief lY bel-ow.

B. Partial- enzymi-c cl-eavage of viroid and virusoid RNAs Under conditions of high saÌt concentration

( ( 60OmM NaCl, 1 OmM MgCl, ) and Iow temperature 0oC ), the highJ-y base-paired natj-ve structures of viroids and virusoids are stabil ized and so cJ-eavage by the single strand specific RNases Tl uz and A 1s initialJ-y limited to relatively few accessibfe sites on the molecules. Thus, pârtial- ribonucfease digestion of the native cincul-ar RNA moÌecuf es glves rise to rel-atively f ew, speciflc linear RNA fragments which may then be radiol-abelled and fractionated by polyacryl-amide gel ef ectrophoresis (Figs 2-2ß1. The main disadvantage in Figure 2-La

purified. circular RNA

A-I, A-2 Partial RNase , A or U2 cleavage lt

A-3 5r- 32 P-labelling A-4 3r- 32P'1abe11ing

A-5 Polyacrylamide gel fractionation

A-7 r) Phosp.hatase tre atment

A-6 Partial enzymic fI) Polyadenylation cleavage sequencing I 111) RTase dideoxy- sequencing

A-8 Polyacrylamide gel electrophoresis Figure 2-Lb

purified circular RNA

B-t t) Linearization, phosphatase treatment, polyadenylation

11) 10 strand cDNA synthesis, tu\ase treatment

lrr ) 20 strand. cDNA synthesis

fV) Restrict.ion, gel fractionation

V) Ligation into M13 vector, transformation

B-3 I) excision or transcription of primer I B-2 Klenow dideoxy- 11) R.irrA - DNA hybridization sequencing III) RTase d.ideoxy-sequencing

A-8 Polyacrylamide gel electrophoresis Figure 2-2 Analytical RNase A digestions of SNMV RNA 2.

As described 1n the text, 0.5 pg purified SNMV RNA 2 v,/as variously digested with 0, 0.1, 1, 10 or

100 u g/nI RNase A unden conditions of high sal-t and

1ow temperature, and lhe resulting l-inear RNA f nagments 5'-32p-,-abelted and f ractionated by denaturing polyacrylamide gel electrophoresis. The presence of a band in the Lrack containing SNMV RNA 2 untreated with RNase A corresponds to a smal-l amount of full-length l-inear breakdown product (377 residues in size ) present with the i-ntact circufar

RNA. Concentrations of between 0 'l and 'l p g/ml RNase A hrere used to obtain fnagments suitable for sequencing. RNa'se A 0¡g/ml) o o.1 1 10 100 I

SNMV RNA 2 LINEAR I -¡p

-.,-

O7*

ii .ri 'ç

6% TBE

-i'' 7M UREA Flgure 2-3 Preparative RNase A and RNase T

digestlons of VTMoV RNA 2 5 Ue amounts of purified VTMoV RNA 2 were digested wlth 0.2r 0.1 or 0.05 Ug/m1 RNase A and 150 or 75 units/ml RNase T., under condi tions of high salt and 1ow temperature. 5r-radlolabelled products are shown fractionated on an B0 cn long 6% poly -acrylamide gel containing 7Vl urea. OnIy the bottom portion of the gel vras autoradiographed and the band corresponding to full-Iength 1j-near VTMoV RNA 2 (365/366 resldues) mlgrated about 30 cm from the origin. Followlng a 5 mlnute autoradiographic exposUr€ ¡ bands r^rere excised and eluted for sequence detenmination.

I I

1 RNase A RNase T1 50 100200 75 150 ng/ml U/ml .Èf LINEAR VTMOV ifl RNA 2 \

\ \ I Ë {

.& þ t ì Ç

_1. È. Ë a

6% TBE ü 7M Urea 80 cm

:. ,.k 20b + Iri -1 * 25 using this technique is that' if the native circular RNA molecul-e possesses an exposed singl-e-stnand negion (such as a termlnal hairpin l-oop) containing accessible sites f or aIl- nibonucl-eases, 1t is dif f icult t'o obtain RNA fragments spanning such an exposed region. This disadvantage is not shared by those sequencing methods which rely on cloned viroid or virusoid fragments.

C. Partial- enzymic cleavage of nadiof abel-l-ed RNA fnagments

The purlfied 3r- or 5r- radiofabelled RNA fragments obtained after partial RNase cleavage of intact vlnoid or virusoid molecules were sequenced uslng the partial enzymic cleavage method ( Donis-Ke11er , 1977; Lockard et â1., 1978; Krupp and Gnoss, 1979; Donis-KefÌer, 1980). An example of one sequencing gel 1s given in Figure 2-+. In particuJ-ar, the use of fragments l-abelled separately aL either the 5r on 3r terminj-i alfowed the sequence determination of long RNA fragments and, with shorter fragments, resol-ved gel compression artefacts ( Kramer and Mi11s, 1978 ) when the relevant nucfeotide sequences v\rere determined from both directions.

D. Dideoxynucl-eotide chain terminalion sequencinA of RNA fragments Figure 2-4 Partial enzymi-c dlgestlon sequencing technlque.' A 5'-32p-labelred RN.A f ragment, obtalned by partial RNase T, digestio n of SNMV RNA 2 under non-denaturlng co ndltlons, v,ras sub j ected to partial d iges t ion un d.e r d enaturing conditions with various

RNases as describ ed in the text (N, no enzyme; T, RNase T,, ; U, RNas e U17 L, aIkaI1 ladder; P, RNase PhyM; B, Bacil-1us cereus RNase). The resulting f ragments brere separated by I0 cn, B% polyacryJ-amide geI elecLrophoresis and the resulting autoradiograph is shown with residues 132 to 175 of SNMV RNA 2. NTUL PB 3 - t¡ UC --t 1r. AG t a GU -- a U A G A-17O U G A u-160G c - q - t) "' aa - a 150 t a - o a Jt

: It 140 o 2 G o t G G

G o t c

U A 5 26

' All- 5'-32p-radiol-abel-l-ed fragments pnoduced by RNase T.'| cleavage of viroid or virusoid RNA possess a 3r proximal- guanine residue, together with a 2t (3 ' )-phosphate group. Treatment of such fragments with calf intestinaÌ phosphatase ( Efs LraLiadis et âf . , 1977 ) removed both the radlofabefled 5' phosphate and 2, (3t)-phosphate groups, and the RNA fragments couÌd then be used as templ-ates for polyadenylation usinpg

E. col-i poly (A ) poJ-ymerase (Sippel, 1973; GouId et âI' ,

1978 ) . The synthetic ol-igonucleotide d ( TBc ) blas then used as a specific primer for reverse transcrlption in the presence of dideoxynucfeoside triphosphates. The polyacrylamide gel fractionated products of such a transcription neaction are shown in Figure 2-5. The dideoxynucl-eotide chain termlnation and partial enzymic digestion techniques are complementary, âflowing further confirmation of sequences and the resofution of occasional band compressions.

E. Cfoning of viroid and virusoid sequences Purified circufar RNAs I^/ere hydrolysed by

prolonged heating aL 1 00oC to generate nandomly nicked

full-length finear mofecufes. Terrninal- 2' (3' ) -phosphates were removed by treatment with caLf intestinaf phosphatase and the RNA mol-ecules

polyadenyl-ated with E. coli poly ( A ) polymerase. First Flgure 2-5 Dldeoxynucleotide chain terminatlon s eq uen c 1ng of RNA f.ragmen ts An RNase T,, cleaved fragment of VTMoV RNA 2 was phosphatase- treated , polyadenyl-ated arid reverse transcrlbed in the presence of dideoxynucl-eoside triphosphates as described in the text. Two Ioadings of the resultant radiolabelled franscripts are shown -fractionated on an B0 cm B% polyacrylamide ge1 containing 714 urea. Residues 355 to 190 (3';5') of VTMoV RNA 2 are shown.

I I

1 ACG T 280 tt t - c( O 190 - t ö - 290 { Ò 200 - Ê.4 ì Ç. Õ{ 300 210 - I) ao - \ -: 220 - \ ì ra I o -310 230 - \ -t I lç (, 240 - rrÛ æ - 320 -: , 250 - \ : * I t e I- - - 330 260 - r - Q \ (I 210 \ - ì (Tr¡- 340 \ -

280- \ -ìr¡ It IT C3 I J a 8% TBE \ I t C 350 290 - 7M UREA - !i ç'r..] ìl rD I SOcm IT t

ì---a \ 355 300 - ìr - ¿l strand 32 P-cDNA r^las synthesized using reverse

transcrj-ptase and oligo- ( dT ) as pnimer ( Gould and 1 O Symons, 1982). The RNA.DNA hybrids hrere heat denatured and the tempfate was removed by RNase A treatment.

Immediately prion to the synthesis of the second DNA s trand , the cDNA hras heat-denatured; this step I^Ias f ound necessary to avoid the formation of anomalous ds cDNAs with Iarge apparent mol-ecuf ar weights. The sel-f -prlmed second DNA strand was transcribed using the Kfenow fragment of E.coIi DNA poJ-ymerase 1, and the resulting

population of circularly penmuted ds cDNA molecules \^ras

then digested with an appropriate restriction enzyme ( s ) to pnoduce specific DNA fragments for cloning (see Figure 2-6\. The restriction fragments were then purlfied by non-denaturlng polyacrylamide ge1 electrophoresis and ligated into a bacteriophage M1 3 vector, tnansformed and propagated in an E.co1i host (MessÍng, Crea and SeeburB, 1981 ; Sanger et âf ., 1980).

F. Sequence determination using cfoned viroid or virusoid sequences Viroid or virusoid sequences vüere determined using recombinant M13 phage in either of two b/ays. First, recombinant phage ss DNA !ùas sequenced using the dideoxynucl-eotide chain termination technique with an

M1 3 specific primer ( Sanger et â1. , 1 980 ) ( see Figure Figure 2-6 Synthesis and nestriction endonuclease cleavage of SCMoV RNA 2 ds cDNA. As described 1n the textr PUrified SCMoV RNA 2 was finearized, phosphatase treated, polyadenyl-ated and used as a tempfate for first strand cDNA synthesis

( 1 o ). The ss cDNA vüas denatured and treated with RNase (1o + RNase) and second strand cDNA was synthesised by seff priming- (2o). Doubl-e-strand cDNA was digested with Hae III, Hha I, Sau96 I, Hpa II + Sau3A I or Hpa II + Taq I. Samples taken aL each step r^rere fractj-onated by 6% polyacrylamide gel efectrophoresis in the prêsence of 2M urea. SampJ-es 1o, 1o + RNase and 20 were loaded wlth (+H) and without (-) heating aL'l 00oC for 2 minutes. Specific bands produced by restriction endonuclease cleavage v.tere suitabl-e f or cf oning. H ãH fO Hsi F ItHr¡l<Ð< 3iå

FULL.LENGTH ? LINEAR SCMoV RNA2 sr DNA

6% TBE 2M Urea 2B

2-7 ) . In this wâV, cl-oned sequences corresponding to either viroid/v irusoid or its compÌement, depending on orientation of the cloned fragmenLs, coufd be determined. Second, cl-oned sequences were excised or transcribed for use as primers for reverse transcrlbed dideoxynucl-eotide chain termination sequencing of intact vinoid or vlnusoid RNAs (Zimmern and Kaesberg, 1978; Symons, 1978). The primed transcripts obtalned in this way are representative of the whole popul-ation of RNA templates, and thus this sequencing technique has been used to estimate the rel-ative proportions of RNA sub-species within heterogeneous populations ( see Figure 2-8).

G. Determination of complete vlroid or virusoid primary stnuctures Construction of the complete prlmary structures of each cincular mofecul-e depended on the obtaining of numerous overlapping sequences by one or more of the techniques described above. Once the entire base sequence of a molecule had been determined by merglng the various overlapping sequences, secondary structure model-s b¡er?e constructed uslng the methods of

Tinoco et al. (1971 ) . The following chapters pnesent the determi-ned structures of several viroid and virusoid RNAs and outl-ine some of the interesting features of these unique mol-ecules. Figure 2-7 M1 3 dideoxynucl eotide chain termination sequencing.

Single-strand DNAs isolated from recombinant bacteriophage M1 3 hrere sequenced by the dideoxynucleotide chain termination technique us 1ng a 17-mer specific primer. The sequence of d. cl-oned

Sau3A I fragment of VTMoV RNA 1 ds cDNA is s hown determined from both indivldually cloned DNA s trands. (+) (-) ACGT ACGT

Sau 3A I + Sau 3A I .+ SITE - 160 S !TE -01 - 150 - 101

q 140 20' ,t - - 130 30¡ o - - ö .l -120 - 40r I - 110 - 501 a ) ô a I -100 - 60r (,-a t ¿ o a -e0 - 701 - t I a o ¡a a, t rt g0l . -80 : - a i 'a ¡ La g0l 'tt -70 t -

i a a -60 -100r \ 't Ò' a T I.ö I - 1101 Ê -50 a a

-40 -1201

31 -1291 Figure 2-B Dideoxynucleotlde chain termination seq uencfnA of intact RNA using cloned DNA pnimers.

Recombinant M1 3 ssDNA containing sequences corresponding to SNMV RNA 2 residues 131 to 216 vüas transcribed to produce a 1abelled complementary strand as described in the text. The cÌoned insert v{as excised, punified and the l-abelÌed transcript hybri_dized to intact VTIt4oV and SNMV RNA 2; this prlmer was elongated uslng reverse tnanscriptase in the presence of dideoxynucleotides. The transcripts r^rene fnactionated by B0 cm 6% potyacryl-amide gel el-ectophoresis and the tracks A, C, G and T correspond to the dideoxynucleotj-de specles present.

Note the sequence heterogeneity evident in VTMoV RNA

2 aL residue 108 which results in band doubling, and in the termination of reverse transcription aL residue 49. T G CA TGCA

19 6tAl 6u A6 *E 6U hh. ouo' I blr¡ I I ,oo" t "..': I 16 77 tA .¡it eit ooÉ I ì G AG - ¡lD 6c A =lr ; -, oo ì oou Ë i¡ -, ä !r i3 uu õ \¡ A ;;- ( r_c ì l¡ì- _6' tE) \*l -r" t u' l}t il *-.i! U ¡a{ - I ú 3o - Èr_.j-_ì t c,A ì= _,bf¡ Ò- t ç!u _ __ O oo' r¡- c , \._ G (I t A ea t U - - G {rt ú 6 \ o"' A 6 rll t - 125 t A 126 SNMV RNA 2 VTMoV RNA 2 CHAPTER 3

CHRYSANTHEMUM STUNT VIROID 29

INTRODUCTTON Chrysanthemum stunt disease bras first

described by Dimock (1947 ) and vras epidemic among cul-tivated Chrysanthemum morifol-ium varieties during 1945-1947 in the U.S.A. (Srierly and Smith, 1949). At that time the production and distribution of chrysanthemum varieties in the U. S. A. \^/ere highly centrallzed, allowing the rapld spread of the disease

( KelÌer, 1 953 ) . Transmission of the chrysanthemum stunt disease by both grafting and sap inoculation (Brierly

and Smith, 1949; 0lson, 1949; Brierl-y and Smith, lt951 ) demonstrated its infectious nature and al-l_owed the development of control measures to prevent accldental_ spread of disease (fefler, 1953). Dlener and Lawson (l9lZ ) demonstnated LhaL the causative agent of

chrysan themum stunt disease r¡ras a viroid, a l_ow mofeculan weight RNA species with properties slmil_an to,

but distinct from potato spindle tuber viroid ( pSTV ) .

Chrysanthemum stunt viroid ( CSV ) was shown to possess a greater electnophoretic mobility Lhan PSTV, and the host

range of CSV hras shown to be mostJ_y confined to some plant species in t,he compositae family whereas pSTV witl replicate in species from a number of plant famil_ies (Diener, 1979; Hollings and Stone, 1973). Ribonuclease fingerpninting has al_so shown that the primary sequence of CSV differs significantly from that of PSTV (Gross, 30

Domdey and Sanger, 1977 \ . However, cDNA-RNA hybridization techniques have indicated that aL Ieast 20% of the PSTV primary sequence is common to that of CSV (Owens, Smith and Diener, 1978). RNA sequence debermination studies vlere undertaken to obtain primary and predicted secondary structures of CSV and to compare these with the structures previously determined for PSTV (Gross et âl ., 1978).

MATERIALS The CSV isol-ate used in this work I^ras originalJ-y obtalned from infected Chrysanthemum cuftivars, Charm type, âhd was kindly provided by T. C. Lee, Adelaide Botanic Gardens, via Dr. R. I. B. Francki, University of Adelaide. CSV, purified from infected chnysanthemuns as described by Pal-ukaitis and Symons (1 980 ) , vras kindly provlded by Dn. P. PaIukaitls.

ME THOD S A. Primary structure determination The primary sequence of CSV was determined using techniques described in Chapter 2. Linear viroid fragments vrrere obtained by partlaJ- digestion of purified CSV unden conditions of high saÌt and l-ow temperature using 3750 units/rnÌ RNas" T.l , 2 pg/n)' RNase A or j-near 2 units/ml- RNase U Z. The I RNA f ragments \,rere 31

32 5 P-radiolabelled using |- l nrp and T, "-32p + polynucleotide kinase, f?actionaLed by denaturing polyacrylamide gel el-ectrophonesis and sequenced using onJ-y the partial enzymic cleavage method, as descrlbed in Chapter 2. Some pantial enzymic digests of virold fragments r^iene fnactionated in polyacnyÌamide gels containing 9B% formamide, as descnibed in Chapter 2 A-8, in order to etiminate band compression artifacts (Kramer and Mil-l-s, 1978 ) arising f rom incomplete denaturation of the fragments during electrophoresis. The compÌete base sequence of CSV üras assembl-ed from sequence data of a l-arge number of RNA fragments obtained by partiaJ- RNase diges t ion .

B. Secondary stnucture detenmination

A possible secondary structure model- Ì,ùas constructed from the compfete CSV sequence using the matrix method of Tinoco et al-. (1971, 1973) and the predicted thermodynamic stabil-ity of the modef was calcul-ated using vafues provided by Dr. D. Reisner (Steger, Gross, Randles, Sanger and Reisner, in preparation ) .

RESU LTS A. Sequence determination Purified circufar CSV vías subjected to partlal 32 dj-gestj-on with RNase T., UZ or A under conditions of high

'l sal-t concenLration (60OmM NaCl, 0mM MgCJ- Z and at 0oC in order to limit cleavage by the single-strand specific RNases to rel-ativeJ-y f ew accessible si tes on the highty base-paired RNA mol-ecul-e. The resulting viroid f ragments hrene 5'-32p-labetl-ed in vitro using T q poJ-ynucleotide kinase ano Iv -32p] nre and f ractionated by size on a denaturÍng polyacrylamlde gel. Figure 3-1 a shows the gel patterns obtained fon partial digestions of CSV with RNa.ses T1 , UZ and A. Digestion with either of the sì-ngì-e base specif ic RNases Tl (G specif ic ) or U^¿' ( A s pecif ic ) gave ri-se to f ewer f ragments than digestion with the C and U specific RNase A.

The gel fractionated 5 I - fabeffed fragments obtalned by pantial- RNase digestì-on of CSV wene excised, efuted and sequenced using the partial enzymic cl-eavage method as described in Chapter 2 A-6. An example of one sequencing gel is given in Figure 3-1b. For some regions of the viroid molecul-e, sequencing u/as complicated by band compression ( Kramer and MiJ-1s, 1978) due to the presence of stabfe base-paired halrpin structures. However, these band compressions coufd be el-iminated by the use of sequencing gels which contained 98% formamide rather than 7VI urea in order to ensune complete denaturation of the RNA fragments. Figure 3- 1 Purification and nucleotlde sequence determination of CSV fragments. (a) Autoradiogram of the trt)- 32 P-labell-ed products of

the partial digestion of CSV by RNases A , U2 and Tl after fractionation by electrophoresis ona

polyacrylamide slab gef as descnibed in Chapter 2 A-5. The largest radiolabell-ed fragment is full

I ength l-inear C SV ( CSVL ) which migrated about 30 cm fnom the onigln. XC is the posltion of the xylene cyanol FF dye marker which corresponds to fragments

about B0 residues long. A number of the shorter CSV

f nagments (including band X ) 'were excised and ef uted for sequencing by pantial enzymic digestion.

(b) Autoradiogram of part of a sequencing gel (g% polyacnylamide ) contalning the various partial

enzymic digests of fragment X. Digestions, as described in Chapter 2 A-6, r^¡ere with RNase T.,' (G), RNas. UZ (A) alkal-i (N) to produce the reference l-adder, RNase Phy M (A+U) and BacilÌus cereus RNase

(C+U). Part of the nucleotide sequence of fragment X from residues 207 to 265 is given. c) o U) >< c) r:

I I

RNose É' ,Jliffft t I "¡ ffiffi t t I RNose Ll2 {ù¡Þüt ¡r'Þ, \ RNose 1-1 I \ \ \ \l ô X

N) ú l. Ll.l r.) L¡r È G) I Þ z_ c+> c +c) \\

( C'ì C Cì'rrì'ìC )''6ì 6l (- Gì Cì Ì- ,' r- 6.ì 61cr-ìC ( r-ì I l--. rì ì I sl r O F..) LJ O 33

B. Primary Sequence and Secondary Structure of CSV The compl-ete base sequence of CSV I¡¡as assembled from sequence data of a large number of RNA fragments obtained using the pantiat RNase digestion technique (Figune 3-2lt. The 356 residues are numbered according to the scheme of Gross et a1. (1929) ror PSTV and the main overtapping sequences used for the pnimary structure determination are shown in Figure 3-2. A possible secondary stnucture modef h/as constructed from the CSV sequence using the methods of

Tinoco et a1 . (1 97 1 ; 1 973 ) and 1s compared with the published structure for PSTV ( Gross et â1. , 197e\ (Figure 3-3). The refative number of G.C base pairs in the predicted CSV stnucture (64 G.C, 44 A.U, 16 G.U) is l-ower than that of PSTV (7 3 G. C, 37 A. U , 16 G. U ) and,

1 using val-ues kindly provided by Dr. D. n/diÀ"er (Steger, ill Gnoss, Randles, Sanger and R1eìi\sner, in prep araLion ) , the l,) thermodynamic stabitities of the proposed models for CSV and PSTV brere calculated to be AG (25oC, 1M NaCl) = -540 KJlmol- and -610 KJlmol respectively.

DTSCUSSION

A. Homology between CSV and PSTV The striking feature of both the primary and postul-ated secondary structures of CSV is the extent of homology with the previousJ-y sequenced PSTV molecule Flgure 3-2 Primany sequence of CSV. The sequence of the 356 residues of CSV is given and the residues numbered according to the published s equence of PSTV ( Gross et âI. , 1978l,. The 247 residues homologous wÍth PSTV are boxed. !{i thin the circular sequence are given the l-ocations of overlapping sequences obtained from RNase fragments of CSV; these sequences do not represent the entire J-ength of these fragments. Each sequence is labelled with the RNase (4, T1, or Ur) which gave the fragment from which that sequence vüas derived.

Figure 3-3 The predicted secondary Structures of CSV and PSTV (Gross et a1. 1978). The boxed areas contain residues homologous between the two viroids. 160 1 PSTV 50 100 c c c uooc ouo o cucoo ÂAOO ouco UUUCC cAc oGc ffifou^""" co c uu U 359 350 ""3oo 250 200 150 CSV 60 100 c u^ c co UU cA coA cô@l ôûcc $o, ôo^cc CO OCU' UCUO uucu oo qFã4 cup c uc uu^F-o¡ uC ^ 356 350 - 3oo 250 200 34

(Gross et â1., 1978). Of the 356 residues of csv, 247 residues (69% ) are homologous with those of PSTV, and occur in two main areas in the primary structure (Figure 3-2) extending f rom residues 247 to 'l 10 and 148 to 206. These areas are separated by two regions of about 40 residues each containing only two small areas of homology. The postulated secondary structure model for CSV (Flgure 3-3) shows that the two main areas of homology each correspond to one base-paired end of the native mofecuÌe. These are separated by the two regions of lesser homology which are positioned almost exactly opposite each other in the native mofecule and are predominantl-y base-paired. Thus, Lhe conservative arrangement and base-palring of such non-conserved regions in the pnimary sequence al-lows the CSV mol-ecufe to form a stabl-e secondary strucLure simll-ar to that of

PSTV.

B. Replicatlon of CSV and PSTV Although the host ranges of PSTV and our isolate of CSV differ significantly ( see Introduction ) , they do overlap in such plant hosts as the composlte Gynura aurantiaca (llenerr1979; Palukaitis and Symons, 1980; Niblett et âf., 1980). It is feasible that replication of the two viroids in these plants will occur by simil-ar mechanlsms in view of theln similarities in size and 35 sequence. Assuming the existence of translatable l-inear forms of the RNAs (Kozak, 1979; Konarska eL âl ., 1981 ), the possibl-e polypeptide products of both the virold and putative complementary RNA stnands of csv and PSTV can be predicted from their known primary structures. Maj or diffenences are found between the possible polypeptide tnanslation products of CSV and PSTV (see Chapter 7), suggestlng that neither viroid codes for proteins involved in their replication. This is consistent with the l-ack of evidence fon any viroid transl-ation 1n vivo

( conj ero and Semancik, 1 977 ) and in vitro ( Davies et âf . , 197 4; Semancik et al- . , 1977 ) . In contrast, the overall secondary structures of CSV and PSTV are conserved despite diffenences in sequence. Given the fack of evidence for functlonal viroid-coded translation pnoducts, the replication of CSV and PSTV rnay invofve necognition by host enzymes which are capabl-e of RNA-dependent RNA synthesis. Thus, the sequence and structural- features common to both CSV and PSTV may play a role in such recognition processes. An example of such a conserved feature 1s si-tuated aL the centre of the native mol-ecules (Fig. 3-3) (CSV residues 74-110, 247-284; PSTV residues 76-112, 247-284) and incÌudes two relatively Ìarge single-stranded reglons which are compl-etely conserved between CSV and

PSTV . 36

C. Rel-ationshiP of thls isol-ate of CSV to othen viroid isol-ates Both Owens et al-. ( 1978 ) and Gross et al-.

(1 977 ; 1 982 ) have used isolates of CSV, of different origins from ours, for comparative studies with a PSTV lsol-ate obtained from Dn. T. 0. Dienen. The sequence of this PSTV i sol-ate has been determined (Gross et al,. , 1978). Owens et al. (1978) used DNA compfementary to PSTV to show that bY hYbridization anal-ysis, their isofate of CSV contained about 2O% sequence homology with PSTV, whereas a viroid isolated from Columnea ery throphae contained about 40% sequence homology. In addition, thein CSV isofate and the Columnea viroid had electrophonetic mobilitles in a non-denaturing polyacnyl-amide gel which wene rnarkedly faster than that of PSTV, indicating appreciably different sizes and/or secondary structures. In contrast, our isol-ate of CSV

shares 69% sequence homology wlth PSTV, is only 3 residues shorter, and possesses a similar secondary structure. The data suggest that, even allowing for some errors in estimates of sequence homology determined

by cDNA-RNA hybnidization analysis, our isolate of CSV may be,more closely reÌated to PSTV in size and sequence than are the Owens et af. (1978) isolates of CSV and 37

Columnea viroid.

Gross et al . ( 1 977 ) have compared the oligonucleotide fingenprints of an isol-ate of CSV' obtained from Dr. M. Hol-lings, with those of PSTV and obtained distinctl-y dif f erent patterns. I t v'Ias concfuded that CSV and PSTV differ significantly 1n sequence. Subsequent sequence determination of that isol-ate of CSV, white confirming its non-identity with PSTV, showed 73% sequence homology with PSTV (Gross et al-., 1982). Flgure 3-4 shows the predicted structures of the two isolates of CSV, together with that of PSTV. It can be seen that both isolates of CSV are closeJ-y rel-ated, sharing 346 residues in common, whi-le the CSV isolate of Gross et aI. (1978) i- s only 2 residues smal-1er in s1ze. The f ew seq uenc e differences between the two isofates are mainly located in the left hand sides (as they are drawn) of the nat ive mol-ecu1es. I t is unknown whether the differences in structure between these CSV isofates corresponds to differences in biol-ogical properties.

Three independent isol-ates of CSV have been characterized and it seems likeJ-y that a1l differ. These three CSV isolates, together with the Cofumnea vinoid, are closeJ-y related to PSTV. It therefore appears tikel-y that there exists a group of viroids, Figure 3-4 Comparison of the secondary structures of two sequenced isolates of CSV.

The secondary structune model-s for the Australian CSV isolate (CSV-R; this work) and English CSV isolate (CSV-E; Gross et ãL., 1982) are presented. The two viroid isolates share over 97% base sequence homologyr âhd the few residues not shaned by both isol-ates are shown boxed.- 150 PST V 50 100 *-. 1 u AA^^ I A Gg I AC c c u u cc A A A oc o A G C AC A u c oc u ^ ôougooo ôuo cc^ ocoo co^c Aooô ccco GA^^ AOGGU AOAAOOCGO CUCOOô OÔC UUCAO ucc.ccooo cuoGôcoA uooc AÂog ôÂCU^cu AACU OUOOUUCC oouu Ac^ccu cucc 0^GcA0 ^^6^ \ u AOg qoccc oocuuc0cu'cuco uuucc ccocucc cAc oo cocc oc uccuu oooc cuuu uccc^ u CCG^ USUGOO OOOg UUCGUU UUCU uuuuucocc oAocc'cuco GUUGG'UUCA.CGCCAÀOO Âôuc c cG c c c^ A C OC C U IJU C uc c c c^ o\ c ^c ûAAU U rc o UU u 200 c 300 cc A 250 359 350

100 150 A 50 U CSV- ^ c o I ¡r O^ u^ cc u o)o^ o c co c u UC c u uc o^ c ^c AOGGU U U U U OC O C C uþl^ cucc oô Aooô cuA U U ^ ^ ^ôÂo c uucô ucc.ccoog cuooôGÂOU cO^ o@þôocsc ^ ooô^c uacu oouucc ocu Acuccu ccuG uoc t@l'Ã'þ.ero 6Aoocs Ôûcc ^GO^CC ô^^^ U UUCUOO o^Go cuc.uccuu oûuu ucccA cuuc . o^ôu c AOO OOCCC oouuuccuuco ocu.ucuooc'Gcoûccco cco uoAgoo oooc uuc uuuuu'uuccoc c cu ccuuoñuo^ ccÂoo c UC u la c^côû'uu U c uc u U uc fttc ^ I f ^fil ^c Â u o 200 u ï" " ^c 300 ^^Cç 250 356 350 otr

@ @ @ @ O CSV-E @ l c0 c .l ll C @ oô¡ 0 ¡u c c UU t AU c ucû u I uor l^ u r cu ¡l l ¡00ôu 0 ¡ UCU c o0^00â0u côr.0ûc0^ 0 00 00t ¡00^cc cuct GAGA ¡001G c U u UU A 00c 0 0^10 c UUC UcC'cC000 c u 00 UU cc 0u00 0c ¡ U ctu c cc 0cu0 0rl 001 ' ^Âl0A^ Ô ll il I : I lill ^C I ill lll illt ilill I. I lil il ill ilil1| lllril rïI ill 0^00 cucu ucl uu 0r uu uc(c¡c I I I lil|l || lllll ccuuc0 0 cu tu00c c UC cco lil u00 rl I lt il ill lllllilll CUUC. 0 a^0 UC ¡00 ll lll U il ulll UUUUUU U. Cc0 c uc tc lJr il u l 0Gc 0 o0 c0û cuu illll uc c, uu ccu u0 l oo.Ùokc.c 0 A AU u rC UU I c l^ uu CA ^ ^l u@ -t € ET c c U I ouu ^ ^ I l^^ I I I ^t @ O @ cÞ @ @ @ -)0 lncluding PSTV, whì-ch share common sequences and possibly secondary structures, and which may be derived fnom a common ancestnal viroid. CHAPTER 4

COCONUT CADANG-CADANG VTROID 39

]NTRODUCTION Cadang-cadang is a serious and economicalJ-y important disease of coconut palms which vras first neported in 1927 on San MigueI IsIand in the Phillppines (ocfemia, 1g37). In the fol-lowing years, incidence of cadang-cadang disease I^Ias reported in surrounding areaS of the Phil-ippines and nor^r, 56 years af ter its f irst known occurrence, the disease j-s found widespread over the south-east part of Luzon and many neighbouring islands ( Figure 4- 1 ) . By 1 96?, only 1 00 of the 250, 000 coconut palms on san Miguef Island had survived the disease (Bigornia,1977). It is estimated that cadang-cadang dlsease is still responsj-bl-e for the death of 5OO, OOO palms each year in the Philippines ( e ' ZeLazny, personal communication). Recent work has indicated that tinangaja disease of coconuts on Guam, âD island 1 ,500 miles east of the Philippines' has the same aetiology as cadang-cadang disease (Boccardo et â1., 1981 ). Hohrever, cadang-cadang dlsease has not been found in any other coconut growing area. The first symptom of the cadang-cadang disease is the devel-opment on the affected palm of smal-1, rounded coconuts whlch are distinctively scarified. Laten, the fronds develop characteristic yellow spots and as the disease advances the cro\^In of Figure 4-1 Incidence of cadang-cadang disease of coconuts. Regions of the Philippine islands whene bhe disease is widespnead are shown cross-hatched, while isolated incidences of the disease occur in surrounding areas (Zetazny, 1979). The different ccRNA i solates used in thi-s work hrere each obtained from separate diseased coconut palms in one of the following locations : 1, San Miguel Island; 2, Sorsogon; 3, Llgao; 4, Lake Baao; 5, Tinambac; 6, Guinayangan; and 7, San Nasciso. 123 E Southern Luzon

anes 5

7 1 a 50 I km t3N o o Burtas U $¡ Samar

o ol " Masbate

Leyte ß 40 the palm is reduced to a tuff of short yelÌow fronds. The course of the disease invariabl-y ends with the death of the infected palm, which occurs from 3 years to more than 15 years ( usual-1y about 1 0 years ) after the appearance of fhe first symptoms (Zelazny and Niven,

1980 ) . Li ttle r^ras known of the nature of the pathogenic agent responsible for the cadang-cadang disease until Randl-es (1975) showed the existence of two

RNA s pecies that r^rere present only in diseased paIms. The RNA species ürere named ccRNA 1 and ccRNA 2 in order of increasing size, âhd wene shown to share thermal- denaturationr nucl-ease sensitivity, centrifugation and electrophoretlc properties with viroids ( Randfes , 1975; Randl-es et âf . , 197 6) . The ccRNAs hrere also subsequently shown to be cincul-ar (Randles and Hatta

( 1979) , like viroids. Randfes and Pafukaitis 1 979 ) , using cDNA-RNA hybridization techni-ques, demonstrated that ccRNA'1 and ccRNA 2 shared common sequences and that the sequences of the ccRNAs \^Iere not found in healthy palms.

Recently it has been shown that ccRNA 1 and ccRNA 2 occun as fast and slow efectrophoretic variants and that the occurrence of the variants is nel-ated to the stage of disease development in the coconut palms

( Imperial et ãI., 1 981 ) . The fast electrophoretj-c 41 varlants, ccRNA 'l f ast and ccRNA 2 f asL are present 1n infected palms aL early stages of the disease, and as the disease progresses over a period of years the ccRNA 1 slow and ccRNA 2 slow variants first appean and then predominate ( Imperial et âf., 1 981 ; Mohamed et âf' , 1982). All four ccRNA species have been recenbly shown to be infectious (Mohamed and Imperial' and as the ccRNAs are unpubl-ished results ) ' slngle-strand covalentl-y closed circul-ar RNAs with high degrees of secondary stnucture ( Randles et âf., 1976; Randles and Hatta, 1979), they possess both biological and physical properties simil-ar to those of viroids. In order to further investigate the intrigulng rel-ationships between the cadang-cadang disease, the vaniant ccRNAs and viroids, the sequences and structures of the dif f erent ccRNAs I^¡ere determined and compared.

METHOD S A. Isol-ation of the ccRNAs Purif ied ccRNAs \^Iere kindly provided by Dr. Nizan A. Mohamed. Fnonds were harvested from naturally infected coconut palms from a number of sites in the Philippines. Nucleic acids I^iere extracted f rom the l-eaf

tiÀsue as described by Imperial et a1. ( 1 981 ) using their Method 1. Tndividual- ccRNAs LIere purified by 3 42 cycfes of polyacrylamide gel efectrophoresis ( Imperial- et âf ., 1981 ).

B. Sizing of the ccRNAs

Sizes of the ccRNAs I^rene estimated by el-ectrophoresis in 6% polyacrylamide gels ( 4OxZO x0.05 cm) containing 9B% formamide (Maniat is and Efstr adlatis,

1980 ) fne f olJ-owing vrere used as mo lecufar weig ht

( mankens - solanum nodlfl-orum mottle virus SNMV ) RNA 2,

377 residues ( Haseloff and Symons, 1 982); vefvet tobacco mottl-e virus (VTMoV ) ntln 2, 365 resi dues ( Haselo ff and

Symons, 1982); chrysanthemum stunt v iroid (CSV ) , 356 residues (Hasel-of f and Symons, 1981 ) ; I strain o f cucumber mosaic virus ( CMV ) RNA 4, 1 027 residues (Goutd and Symons,l982); CMV satel-lite RNA, 336 resldue e

( Gordon and Symons, 1 983 ) ; chicken 1 BS rRNA, 180 0 residues ( Spohr et â1., 1976); alfal- fa mosaic vi rus (AMV ) nl¡R 4, BBl residues (Brederode et âf., 1980); yeast 5 . BS RNA, 1 5B nesidues ( Rubin, 1 97 3) ; yeas r 5s RNA, 121 residues (Vtiyazaki, 1974); Escherichia col-i phenylalanine tRNA , 7 6 nucleotides ( Banrefl- and Sanger, 1969). The circular RNAs (SNMV RNA 2, VTMoV RNA 2, CSV and the ccRNAs'ürere boil-ed for 15 minutes in distilled water before electrophoresis to produce the l-inear f orms. Af ter ef ectrophoresis, gels I^Iere stained 43 r^rith 0.01 % tol-uidine bfue and destained with water.

C. Fingerprinting of the ccRNAs Purified ccRNAs (0.5 Ug) vüere dried down, resuspended in 5 UI 5mM Tris-HCI pH 7.5, and digested with 0.1 UC RNase A at 37oC for t hour or with 20 unlts RNase T.', aL 56oC for 30 minutes. The resultant oligonucleotide fragments urere transferred to another tube contaì-ning 5 uci dried down Iv-32p] arP (2ooo

Cilmnol ) and 1 .5 U1 of 5x Tr* polynucfeotide kinase buffer ( 25OmM Tris-HC1 pH 9.0, 50mM MgCIr, 5OmM dithiothreitol-), and 0.5 pI (0.5 units) Tq polynucl-eotlde kinase added. The reaction Idas incubated aL 37oC for 20 mj-nutes and 5 Ul formamide l-oading buffer (95% (v/v ) Oeionized formamide, 10mM NaTEDTA, 0.02% (w/v) bromophenol bIue, O.02% (w/v) xylene cyanol FF) added. Radiol-abel-Ied ollgonucl-eotides h¡ere fractionated by two dimensional- polyacryl-amide gel el-ectrophonesis. For the first dimension, preparatlons !,rere ef ectrophoresed in 40x20x0.05 cm 10% polyacrylamide gels containing 25nM sodium citrate pH 3.5 (leWachter and Fiers, 1972; Frisby , 1977 ) . After the xyl ene cyanol

FF dye marker had migrated 14 cfl, electrophoresis \^ras stopped r and gel strips r^¡ere excised and embedded aL the bottom of 40x20x0.05 cm 25% polyacrylamide gels 44 contalning B9mM Tris-borate pH 8.3, ZmM NaTEDTA (Frisby,

1977). Polymerization of the second dimension gels r^ras catalysed by the addition of 300 Ul 10% (w/v) ammonium persulphate,30 Ul TEMED, 50 Ul 10% (w/v) ascorbic acid and 70 llf 30% (w/v ) HZOZ per 50 ml of gel soÌution to ensune compfete polymerization in the region of the f irst dimension gel strip. Samples v\rere electrophoresed upwards until the bromophenol bl-ue dye marken had migrated 1B cn, and the fractionated 32 P-labeÌed oJ-igonucfeotides were detected by autoradiography.

D. Sequence and structure determination of the ccRNAs

Punified ccRNAs \^rere sequenced essentially as described in chapter 2. Purified ccRNAs ( 5 ug) isorated f nom sì-n91e inf ected coconut paJ-ms i^Iere sub j ected to l-imited digestlon by 3 units/ml_ RNase UZ, 5 ng/nl RNase A or 2000 units/mI RNas. T.l under non-denaturing conditions ( 600mM NaCf, 1 OmM MgCl Z aL 0oC ) . The res.ulting l-inear RNA f ragments brere 5'-radiof abef led . r 32-- using I V---P_IATP and Tq polynucfeotide kinase or, after treatment wi-th calf intestinal phosphatase, 3r-radiolabelled using [¡'-3'rldpCp and Tq RNA t-igase, and fractionated by denaturing polyacryl_amide gel el-ectrophoresls as detailed in Chapter 2 A-3 r4 r5. Radiolabel-l-ed f ragments b¡ere located by autoradiography, excised, el-uted, and sequenced using partial enzymic 45 cleavage methods. The sequences of numerous overlapping fragments vüere assembled to give the compl-ete primary structure of each circul-ar mol-ecuf e. Secondary structures of the ccRNAs vüere mappgd using St nuclease ( Wurst et âf . , 1 978 ) . Futl J-ength 5t - on 3 | -radiol-abelled ccRNAs, obtained as descrj-bed above by RNas" T1 digesLion, were suspended in 20 Ul 200mM NaCl-, 0.05mM ZnSOO, 50mM sodiurn acetate pH 4.6, containing 5 Ug Escherichia coli tRNA c arrier, and incubated aL 37"C for 10 minutes with 0.1, 'l or 10 units of S" nuclease ( Boehringer ) . The reaction I mixtures r^i ere extnacted wlth phenol, precipitated with ethanol- and fractionated by el-ectrophoresis in a polyacrylamide gel containing BM urea and TBE buffer

( 9OmU Tris-borate pH 8.3, 2ml4 NaTEDTA ) . Products of partial enzymic sequencing reactions of the same ccRNA species vüere run as markers, thus allowing sifes of S1 nucl-ease sensitivity to be l-ocated. Data so derlved r^rere used during consLructj-on of the secondary structure models of ccRNA 'l f ast and ccRNA 1 slow. Secondary structure modeÌs for the native ccRNAs hrere constructed using the base pairing matrix procedure of Tinoco et al-. (1971; 1 973 ) and the thermodynamic stabilities of the pnedlcted RNA s tructures r^rere calcul-ated using values kindl-y provided by Dr. D. Reisner (C. SLe65er, H.J . Gross, J.W. Randles, 46

H.L. Sanger and D. Reisnen, unpublished results).

RESULTS and DISCUSSION A. Sizing of the ccRNAs The sizes of the finear fasl and slow variants of ccRNA 1 and ccRNA 2 were estimated by electrophoresis in polyacrylamide geJ-s containlng 98% formamide, using Iinear RNAs of known mol-ecular weight as markens (Figure 4-2). No differences in mobiJ-ity were observed between the different ccRNA isolates. The sizes of the RNAs were estimated to be : ccRNA 'l f ast, 250 residues; ccRNA 1 slow, 300 residues; ccRNA 2 fast, 500 residues; ccRNA 2 slow, 600 resldues. Therefore the ccRNA 2 fast and ccRNA 2 sÌow forms are approximately twj-ce the size of the correspondlng ccRNA'l forms and the sl-ow forms of both ccRNA 1 and ccRNA 2 are 20% larger than the corresponding fast forms. These estimates of size differ from those of ccRNA 1 (31013 resldues) and ccRNA 2 (438t5 residues) as obtained by length measurements made after el-ectron microscopy (Randl-es and Hatta, 1979). However, Randles and Hatta (1979) Oid not differentiate between the fast and sfow ccRNA variants which mâV, together with perhaps under-estimated experimentaf error, have resulted in the differences observed. Figure 4-2 Slze estimation of the fast and sfow forms of ccRNA 1 and ccRNA 2 by eleclnophonesis in a 6% polyacrylamide ge1 containing 98% formamide. 148, 620C and X2 vlere different isol-ates of ccRNA obtained from Ligao, PhiJ-ippines. Sizes of the ccRNAs wene determined frorn a standard curve of mobillties of the RNA markers (described in Methods) pl-otted against their known sizes on a logarithmic sca1e. 0nJ-y the Iinear RNAs, hot circuf ar RNAs vüene used for size estimation. Fon SNMV RNA 2, VTMoV RNA 2, CSV, and the ccRNAs, the linear forms are indicated by arrov\rs. Fon CMV p lus satellite (sat) RNA, the bands are (in order from the top) RNAs 1,

2, and 3 runnlng as a broad band, RNA 4 and sateflite RNA. Marker RNAs, in order from the top, are chicken 185, AMV RNA 4, yeast 5.8S RNA, yeast 5S RNA and E.coIi phenylalanine TRNA. SNMV RNA 2 It VTMov RNA 2 csv CMV + sat *i* 'f,+ MARKERS ¡i r+a I t-t"", e zocl """*o Il, r¿e'l o2oGl ccRNA t-slow rl x2 I I, 62OC ccRilA 2-fast ^l X2 ccRNA 2-slow I o 2 ez

SIZE (no. ot bases)

c ê

o ê o =o c o ! g / Ð z I z N o I = o u I ø_ o z Þ o a ¡2 É 3 I ¿ ¿ N q ê I o ê o o o Þ f o o ú d I I oN o o o t o 47

B. Fingerprinting of the ccRNAs The sequence rel-ationships between the four ccRNAs \^rere further investigated using RNase A and RNase T, fingerprints. The individual purified cincular I ccRNAs isolated from a single infected palm vùere digested to compl-etion with RNase A or RNase T., , 5'-32p radj-ol-abelled, and the resulLanb oligonucfeotides f ractionated by two-dimens j-onaI gel el-ectrophoresis. The RNase A fingerprints of the four forms of ccRNA extnacted f rom the same tree (isol-ate Ligao t.,, ) show essentiaì-1y identicaf patterns of l-abelled oligonucleotides ( Figune 4-3 ) . This indicates that the three larger ccRNAs contaln the same sequences as the smalf est ccRNA 1 f ast. SimiJ-arJ-y, essentiaJ-ly identical patterns of radiolabell-ed oligonucfeotides vüere obtained af ter digestion with RNase T.,' ( dat'a not shown ) , however three extra oligonucl-eotides \^Iere found in the RNase T., f ingenprints of the f ast ccRNA f orms, which I^iere not in the fingerprints of the sfow ccRNA forms; the significance of this is not known, but may be rel-ated either to sequence heterogeneity obsenved in ccRNA fast forms (see below) or to difficulty experienced in ensuring the RNas" T1, in contrast to RNase A digestions, always go to compl-etion. The ccRNA 2 fast and slow forms are estimated to be twice the sizes of their respective ccRNA 1 forms Figure 4-3 RNase A fingerprints of ccRNAs. The fast and sl-ow forms of ccRNA l and ccRNA 2 were digested wi th RNasê A, 5r- 32 P-label-1ed and fnactionated by two dimensional gel e I e c t r o p h o r e s i s . Migration in the first dimension is from left to right and in the second dimension from bottom bo top. CCRNA 1 FAST CCRNA 1 SLOW CCRNA 2 FAST CCRNA 2 SLOW

1a D o ID qt a b a t) t o t a a. a .D D- o a - .D o J - o o 1o a a a a -(o:oo-. a a a DO oa lo O a a a (a a e a e a oo at: ra la O ¡ .D (D- ¡ r .D rD - 48

( Figure 4-2l, and, since the RNase fingerprints show that ccRNA 1 and ccRNA 2 possess similar sequence conplexity, it is possible that the ccRNA 2 fast and slow forms a"e dimers of the ccRNA 1 f ast and sl-ow forms respectively.

Furthermore, it is J-ikely that each ccRNA s l-ow speices contains only repeated sequences of the respective ccRNA fast species. In order to extend these observaLions, the sequences of the ccRNAs lrere determined.

C. Sequences and structure deterrnination of the ccRNAs Native circufar ccRNAs vüere subjected to limited digestion either by RNas" T.l , which catalysed cleavage of ccRNA'l species at singl-e sites and ccRNA 2 species aL elther or both of two sites to pnoduce specific full- length Iinean ccRNAs (Flgure 4-4), or by ribonucleases A or U2, which produced smafl-en overl-apping RNA fragments. These l1near RNA molecules vüere 5 r - or 3 ' - radiol-abelled and then purif ied by polyacryl-amide gel electnophonesis. The sequences of these 5'- or 3 r - l-abelled f ragments vüere determined by the partial enzymic digestion technique. The use of f ragments f abel-1ed separatel-y at both the 5 r - and 3 r - ends al-l-owed the sequence determination of long RNA fragments up to 574 nesidues long and, with shorter fragments, resol-ved gef compression artefacts ( Kramer and Mills, 1978) when the rel-evant nucLeotide sequences Figure 4-4 Partial RNase T . digest of ccRNAs. Purif led ccRNAs (lsoIate Llgao T., ) hrere digested with 2000 units/ml RNase T., under condltions of hlgh salt and 1ow lemperature as descnibed in the text.

Resulting 11near RNAs were 5,-32p-tabell-ed and fractionated on a 5% polyacrylamlde geI containing 7M urea. Di-gestion of ccRNA 2 species gave rlse to linear RNAs corresponding in sizes to those of the ful-1-length ccRNAs 2 and ccRNAs 1. ccRNAl ccRNAl ccRNA2 ccRNA2

f ast slow fast slow Origin

ccRNA2 slow linear tpr+.ff - ccRNA2 f ast linear

##r-..t+r ccRNA 1 slow linear

ccRNAl faEt llnear

è. 5% TBE 7M Urea 49

b/ere determined from both directions. The sequences of overlapping fragments r^rere assembl-ed to construct the

complete primary structunes of the circul-ar RNA molecules. Secondary structure models for the native ccRNAs were constructed using the base-pairing matrix pnocedure of Tinoco et aÌ. (1971), val_ues for the thermodynamic stability of the predicted RNA structunes

(C. Steger, H.J. Gross, J.!ù. RandJ_es, H.L. Sanger and D. ,,1 Rþli)sner, unpubrished results), and experimental evidence ltJ for the l-ocation of ribonucfease sensitive single stranded regions on the native mol-ecul-es. rn addition, specific fulf-rength linear ccRNAs, ppoduced by rimited ribonuclease T,, cf eavâge, vJere eithe r 5, - or 3r-32p fabeÌ1ed and the susceptible singJ-e strand regions in the native structures rocated by the s1 nucl-ease mapping procedure ( Vùurst et â1. , 1978) . The sltes of cleavage were detenmined by co-el_ectrophoresis of the radiol-abel-led fnagments of the s1 nuclease digest with products of sequencing reactions using the partial enzymic digestlon procedure (Figure 4-5). Thus the possibility of speclf ica1ly l-1ne arj zing the circul_ar ccRNAs by limited cleavage with RNase T., both facilitated sequence determination of the mofecul-es and afl-owed s1 nucl-ease mapping of the renatured l_inearized mol-ecuf es. Figure 4-5 Partial enzymì-c digestion sequencing and S1 nuclease mapp ing of the ccRNAs Full--ì-ength l-inear lsolate Baao 54 ccRNA 1 fast was pnoduced by partlal RNase T., digestion under non-denaturing condiLì-ons, 5'-32p-rabelled and purified by polyacryl-amide gel efetrophoresis as descnibed in the text. Punified radiolabell-ed RNA vras,subjected to treatment by no enzyme (N), RNase (T), RNase (U), (L), RNase (P), T.t¿ U. alkali PhyM Bacillus ceneus RNase ( B ) and 1 0 units nucl-ease S1

(S). The products were fractionated by B0 cm 6% polyacryl-amide gef electroporesis, and cleavage sites for nucl-ease S1 are shown anrovJed on the predicted secondary structure modef of ccRNA 1 f ast. 2A 1() I '\^'^î,.r" g7 sl l^il""^ cc ACU (¡ U J\ AGA GCCGC 2.+ GGGG AU AGGG.CACC AC AC GCAGG. c U (¡ U CC UCU.CGGCG 1à cccc ;; UCCC GUGG U U A AU "r'íÌf"i ì A @ I ltcâÞ câ -l^ zta ôl 226 i ¡iaR I A , "3i , 224 R 218

ú'tfe?ar¡O co¡-(o 5?89È g ÂlôtôlñÂtN dÈi æb

D. ccRNAs differ in size but not sequence complexity. TheSequencesandpredictedstructuresofthe ccRNA 1 fast and sl-ow forms isolated fnom a slngJ-e infected coconut palm (isol-ate Baao 54, Figure 4-1) are shown in Figure 4-6 together with the known structures of four viroids, PSTV ( Gross et âf. , 1978) , CSV (HaseIoff and symons, 1981 ), CEV (Visvader et al ., 1982) and ASBV (Symons, 1981 ). The two native ccRNAs 1 possess extensive regions of intramolecul-an base pairing and can form rod-tike native structures similar to other viroids. The ccRNA 1 fast and ccRNA 1 slow possess 246 and 287 nesidues respectively, âDd have cafculated thermodynamic stabililies of -320 and -360 KJmol- -,1 respectively. The ccRNA 1sl-ow contains the entire sequence and structure of the smal-ler ccRNA'l fast but differs by an additionaf duplicated sequence and 'l struc ture of 41 residues (residues 03-143 in ccRNA 1 fast ) which is added at the right-hand end of the nati-ve mofecul_e between resldues 123 and 124 0f ccRNA',l

fast ( Figune 4-6) . Thus , the rod-1ike, base-paired nat j-ve structure is maintained in the larger molecul-e. The nucl-eotide sequences of ccRNA 2 fast and ccRNA 2 slow, consisting of 492 and 574 residues respectiveJ-y, are perf ect dimers of the respect j-ve ccRNA

'1 f orms. A schematic summary of the rel-ationships Figure 4-6 Sequences and redicted secondary structures of the Baao 54 isol-ate of ccRNA 1 f as t and ccRNA 1 sl-ow ane shown with those of PSTV, CSV, CEV and ASBV. The structures are aligned under the central conserved negions of these viroids (boxed). Cadang-cadang RNA 1 fast i20\ u¡ I r." \,\ oooo ¡u

t^ 2a€ 2ao

Cadang-cadang RNA I slow 80 t@ !ao 20 ao l"t" ^l

cccc uc uccc ouúo

r60 2A7 2AO PSTV

r20 tao rto ! \. "t^ êc^l¡ t" I I 2ao 220 lAO 35e "ìo

CSV 80 r80 \, I

¡uo. I c E¿LUo tuI uu I u 3ao "_32o ^.._ gæ cEv t m 40 l0 a0 tæ r20 tao tto

I t" oo¡ucu Jcu aouucc oou uc¡c

I l7 ü ilo 320

s u8P!æræ :J I u ! I ASBV ^ ^/

, J¡ ".1' I t t lct rao c r0o r¡o 51 between the pnimary structures of all foun ccRNAs and their predicted secondary structures is glven in Figure 4-7. V,¡hile each of the monomeric ccRNA 1 forms can base pair intramoÌecul-arJ-y to forrn a single rod-like conformer, the ccRNA 2 forms, due to their dimeric nature, can each form either of two rod-like conformers (A or B, Figure 4-7) and a large number of intermediate cruciform-shaped structures, one of which is given 1n

Figure 4-7 . The ccRNA 1 fast and ccRNA 1 sfow mol-ecufes each possess, under experiemental- conditions, a highly accessibl-e site for cleavage by RNase T.,, at the right-hand termlnal hairpin loop of the predicted native structure (between residues 124 and 125 1n Baao 54 ccRNA 1 fast and between residues 145 and 146 in Baao 54 ccRNA

1 slow ) . Limited RNase T., digestion of each ccRNA 2 species produced specific finear RNA fragments corresponding to both the respective full-length linear ccRNA 2 and ccRNA 'l mofecufes. Sequence determination of these fragments showed that cleavage of the ccRNA 2 molecules occunred aL two sites located at the same sequences as for the two ccRNA 1 molecul-es. This suggests that either the predicted confonmer A of the two ccRNA 2 mol-ecufes on possible cruciform intermediates exist in sol-utlon, whereby the appropniate terminal hairpin loops are exposed. However, the Figune 4-7 Schernatic representatlon of the sequences and pnedicted structure relationships between the ccRNAs.

The circular sequences of the four ccRNAs are shown wibh bIack, âhd cross-hatched boxed negions representing the sequences highly conserved between ccRNAs , PSTV, CSV and CEV. The white, and stippled boxed regi-ons represent those sequences duplicated within the ccRNA 'l sl-ow species. Positions cornesponding to residue'1 of ccRNA fast or ccRNA 1 slow are indicated by black dots. Both ccRNAs 2 are dimers of the respective ccRNA 1 forms and can potentially form either of two rod-like conformers A or B as wel-l- as a Iarge number of cruciform-shaped

intermedlates, of which one is shown. Each ccRNA 1 species possesses a single, highly accessibl_e site for RNase T. cleavage Iocated on a terminal hairpin I loop; these sites are indicated by arror^rs. Each ccRNA 2 species possesses two such accessible sites for RNase T. cleavage and arrows indlcate where I these sites al-so occur on hairpin loops in the different ccRNA 2 conformers. ccRNA I last

5 I

ccRNA 2 last

-Hr

J e

B ccRNA I slow

ccRNA 2 sl A -l

j a-

B 52 existence of type B conformens cannot be precluded.

E. Variation Ín sequence between different ccRNA isol-ates Diffenent lsol-ates of cadang-cadang RNAs hlere each obtained from single infected coconut palms from different locafities in the Philippines ( figure 4-1 ) . Sequence differences between the isotates consist of two types. First, the sequences of the ccRNA 1 sl-ow forms can differ. While afl ccRNA 1 fast forms are essential_Iy identical ( see below ) , ccRNA ',l sl-ow f orms can diffen in the length of the repeated sequence inserted between resi-dues 123 and 124 of ccRNA 1 fast. Three different repeated sequences found in nj-ne sequenced isol-ates of ccRNA 1 slow are given in Figure 4-B; these vary in length from 41 to 55 residues but ane al-l internalty base-paired to produce dupJ-icated structures as well- as Sequences al the right-hand ends of the natlve molecul-es. Interestingly, the right-hand ends of the mofecules of PSTV, CSV, CEV and ASBV (Figure 4-6) are similarJ-y distanced from the centraf conserved regions of these molecules. In contrast, the right-hand side of the ccRNA'l fast mol-ecuIe 1s shorten while those of the elongated ccRNA 1 sfow mol-ecufes are cfoser in size to those of PSTV, CSV' CEV and ASBV.

Second, four of the six isoÌates of ccRNA 1 Figure 4-B Sequence variatlon between ccRNA 1 sl-ow of three ccRNA isol-ates. The sequences and structures of various ccRNA'l fast and ccRNA 1 slow isol-ates were determined AS described in the text. As essentially all sequence variation occurred aL the right hand end of the ccRNA 1 slow molecul-es, only this region is shown. Boxed regions represent those sequences which are duplicated in the ccRNA 1 slow mol-ecul-es and which are 41 (isolate Baao 54), 50 (isol-ate Ligao 148) or 55 residues ( isolate Ligao T1 ) long. AIl sequenced ccRNA 1 slow isol-ates correspond to one of these forms ( Table 4-1 ) . ccRNA I last '120

GCG

cGc

ccRNA I slow Isolatc Ligao T,

GCG

cGc

'160 ccRNA I slow lsolate Ligao l4B 100 120

u ccG

cGc c

I 200 180 160 ccRNA I slow lsolate Baäo 54

100 120 140 uc GCG CUGGG

coc GGCCC 53 fast sequenced each consist of two populations of mofecules, one of 246 nesj-dues and the other of 247 residues, which differ in the presence or absence of a C aL residue 1 9B ( Figure 4-9, Table 4-1 ) ' Simil-ar sequence heterogeneities have al-so been reported fon CEV (Gnoss et â1., 1982 ) and the viroid-like RNA of vTMoV

( Hasel_off and symons, 1982) . The nefative proporfions of the two ccRNA 1 fast subspecies vary between different lsol-ales as Iisted in Table 4-1. For the two ccRNA 2 fast isol-ates sequenced, the relative proportions of the Lwo forms are the same as those of the corresponding ccRNA 1 fast. In contrast to the 'ccRNA 1 and 2 fast species, no similar sequence heterogeneity has been observed in nine isol-ates of ccRNA 1 slow and the one sequenced lsolate of ccRNA 2 slow (TabIe 4-1), Each isolate of the ccRNA slow species thus consists entirely of either one subspecies or the other; in all except one case, the C at the position corresponding to ccRNA 1 f ast residue 'l 9B was absent. The - various sequence differences between the ccRNA isolates do not seem to correlate with differences in geographic focation.

F. Structural- simil_arities between ccRNAs and viroids The ccRNAs shane two regions of sequence

homology, each of abouL 2O nucfeotides, with the viroids Figure 4-9 Sequence hetenoEeneit y r^ri thin the ccRNAs.

Purified ccRNAs were found to consist of either one

or a mixture of two RNA species which differed in

the presence or absence of an extra C residue at the

position corresponding to ccRNA 1 fast residue 197 or 198. Port j,ons of the predicted native stnuctures of these two species are shown with stars to indicate the sequence differences. Boxed regions

indicate those sequences common to PSTV, CSV and

CEV. 60

GCCGC U

CG CG A *c 200

U GCCGC U UCC CC G c

CGGCG A **cc U

2 o r ccRllÀ isolates were purified from nucleic-acid extracts of infecteci coconut single, Palms. Ì t ReI¡tive proportions of sequence variants were

Re l ati ve ProPorti ons Totaì length Length of sequence ccRNA lsoìate* of sequence variantst of ccRNA dupl i catl on

ccc (resi dues ) ( res I dues )

ccRNA I fast

Baao 54 1.0 0 ?46

Tinambac 1;0 0 246

Li qao l4tÌ 0.8 0.2 246 1247+

Li gao 620C 0.6 0.4 246 /247

Li 9ao 1910 0.4 0.6 246 /247

Li gao Tl o.2 - 0.8 246 /247 ccRNA 2 fast

Baao 54 1.0 0 492

Lì gao Tl 0.2 0.8 A92-494** ccRNA 1 slow

Baao 54 1.0 0 287 4t

Ligao'l48 1.0 0 296 50

Li gao 620C 1.0 d 296 50

Li gao 19lD 1.0 0 296 50

Lì gao Tl 1.0 0 301 55

Li gao 5 1.0 0 ?96 50

Gui nayangan 1.0 0 ?96 50

San Mi grreì 1.0 0 296 50

San Nasci so 0 I 0 297 50

ccRNA 2 slow

Baao 54 1.0 0 574 4t 54

PSTV, CSV and CEV (Figure 4-6). The latter three viroids are closely rel-ated, sharlng abouL 50% sequence homology. The two conserved regions are base-pained in the predicted structures of the native molecul-es to form highJ-y conserved secondary structures. The conserved regions shared by the ccRNAs correspond to regions of PSTVTCSV and CEV postul-ated to be involved in base-pairing of viroid compl-ementary RNAs with a plant smal-l- nuclean RNA (snRNA ) in a manner anal-ogous to that proposed for the interaction of mRNA intron-exon splice junctions with mammal-ian Ula snRNA (Gross et âf ., 1982; Diener, 1981), and in the formation of a stabilizing stem-1oop structure in the viroid complements ( Gross et â1., 1982) . The proposed lnteraction between viroid compl-ements and snRNA is postulated to reflect the origin of vlroids from an lntnon ancestor (Diener, 1981) or as a basis for pathogenesis ( Gross et âf. , 1982; Diener, 1 981 ) , but has not been proposed to be directJ-y invol-ved in viroid neplication. However, the RNA or compl-ementary RNA of aL least one viroid, ASBV, is incapabl-e of base-pairing wi-th a UIa-like snRNA or of formatlon of the conserved stern-loop structune, despite up to 1B% sequence homology betwen ASBV and other vlroids (Symons, 1981 ). It is possible that the central- conserved regions of viroids, including ASBV, tr€f lect f unctional- simll-arities rel-ated 55 to viroid replication rathen than to the postulated snRNA b inding.

G. Replication of ccRNAs As PSTV,CSV,CEV and the ccRNAs are capable of autonomous repJ-ication, the enzymes involved are of Ho\^Iever no viroid-encoded considerabl-e interest. ' translation products have been found in vitro (DavÍes et âl . , 197 4; Semancik et aJ . , 1977 ) or in vivo ( Conj eno and Semancik, 1977). Although PSTV, CSV and CEV share around 50% sequence homology' none of these vlroids nor thein putative complementary RNAs can theoretically encode similar transl-ation products (Hasel-of f and Symons, 1981; Vlsvader et al-., 19BZ), even assumlng the existence of LransLatabfe l-inear viroid RNAs in vivo (Kozak, 1979; Konarska et â1 . , 1981 ) . Possibl-e protein-coding regions simifan to those of other viroids are not found in the ccRNAs on their complements nor are there any AUG initiation codons present. It therefore seems highly unlikely that the ccRNAs can code for any functional- polypeptide product. All evidence indicates that ccRNAs and other viroids must rely entlreJ-y on host components for their replication.

Larger that unit-length compfementary (-) RNA intermediates have been detected in PSTV, CEV and A SBV

It infected tissues ( Branch et âf . , 1 981 ; Rohde and Sanger, 56

1981; Owens and Diener, 1982; Bruening et â1., 1982) ' In addition, âtr oligomeric series of RNAs of ASBV (+) have been detected in infected avocado tissue; the dimer of ASBV has been purifled and characterized as a single-stnand, circul-ar molecule slmifar to the ccRNA 2 molecules. Rolling circl-e mechanisms have been postul-ated for the synthesis of oligomeric compl-ementary RNAs from circular viroid templ-ates (Branch et a1', 1981; Owens and Diener,19B2; Bruening et â1., 1982), and oligomeric viroid (+) sequences could be simply generated by transcniption of multimeric ( - ) s trand templ-ates. Unit length viroid pnoduced by either specific transcription or cleavage of oligomeric viroid RNAs must be ligated to produce the finaÌ circufar product. Such a model for viroid replication, invoÌving oJ-igomeric RNA intermediates, could readily account for the formatÍon of the dimeric forms of both ccRNA 1 fast and ccRNA 1 slow; that is, ccRNA 2 fast and ccRNA 2 sl-ow nespectively. Rate-Iimiting steps during the transcription or the possible processing of viroid transcrJ-pts would al-1ow the dirneric ccRNAs 2 to accumul-ate over the monomeric ccRNA 1 species.

H ccRNA slow variants and the tlme course of infection In the initial stages of cadang-cadang disease, only the fast forms of ccRNA 1 and ccRNA 2 are 57 present in infected Pa1ms and it is onJ-y after a furthen 24-30 months that the sfow variants of ccRNA 1 and ccRNA 2 first appear and in the foll-owing yeans predominate (Mohamed et al ., 1982]}. These data, pl-us prelJ-minary evidence that the ccRNA fast species are more infectious (ImperlaI 1981 ) are that the ccRNA slow species et âf., ' consistent with the de novo generation of the ccRNA sfow varlants durlng each cadang-cadang disease Ìnfection. This proposition is supported by the following sequence data. 1) The ccRNA 1 sf ow forms dif f er f rom ccRNA 'l fast by the insertion of a single repeated sequence (Figure 4-6) and coul-d be simply generated from the ccRNA 1 fast by processing and/ or transcription mechanisms ' 2) The ccRNA 1 slow isolates can differ in the size of their inserted sequence repeats (Figure 4-8, Table 4-1 ) suggesting separate orlgins for these ccRNA slow variants. 3) !,Ihi1e most ccRNA fast isolates contain a sequence

heLerogeneity aL residue 1 98, and consist of varying ratlos of the 246 and 247 residue species, each of the nine sequenced ccRNA slow isolates consists of a singl-e

homogeneous poputation, either with or without a C residue at the position homologous to ccRNA 1 fast residue ,l 98, and with only one slze of repeated Sequence 5B

(Figur" 4-9, Table 4-1 ). These data are consistent with the generation of ccRNA sl-ow forms from ccRNA fast by single, rare sequence duplication events occurring separatefy in each cadang-cadang inf ected paIm. AII ccRNA sl-ow mol-ecul-es wou1d, therefore, origlnate from single panent moÌecul-es and may accumulate in pneference to ccRNA fast species due to a competetive advantage in repJ-ication.

I. 0rigln of cadang-cadang disease

The ccRNAs share biological- properties and sequence and structuraf homology with viroids so that application of the term coconut cadang-cadang viroid

( CCCV ) i-s fully justified. However, whereas other viroids consist of a single predominant infectious RNA species, CCCV consists of several variant RNA species.

It is feasible that CCCV may have arisen from a pre-existing viroid and that mutation or infection of ner^/ hosts, such as the coconut palm and rel-ated host specles ( Randles et â1 . , 1 980 ) , resulted in the production of the variant ccRNAs by abberant transcrlption and/or pnocessing mechanisms whlch nonmally occur faithfully 1n the replicaiton of other vinoids. The outbreak and subsequent apparent rapid spread of the cadang-cadang disease in the Philipplnes this century (ZeIazny, 1979) is consistent with such an 59 origln of the ccRNAs. As viroids do not appear to encode functional polypeptide products, it seems llkel-y that these pathogens rely entirel-y on the interaction of the viroid RNA with host celf components for replicatlon. If so, the homology between cccv, which replicates in several species of the monocotyl-edonous plant family Pal-maceae (Randl-es et al- . , 1980 ) , and other viroids, which replicate in dicotyledonous plant hosts ( Dienen, 1979) , may mj-rnor simifar homology between cell-uIar components responsibl-e for viroid repl-ication in these different hosL plants. The exact nature and function of these possibly conserved host cel-l- components is as yet unknoü/n. CHAPTER 5

VELVET TOBACCO MOTTLE VIRUS

AND

SOLANUM NOD]FLORUM MOTTLE VIRUS 60

INTRODUCT]ON As outlined in Chapter 1, a nev'r unique group of plant viruses has been reponted in Austral-asia (Randfes et â1., 1981; Gould and HaLLa, 1981; Tlen Po et â1., 1981; Francki et â1., i983). The viruses consist of 30 nm diameter pofyhedral capsids containing two major single-strand RNA specles; the RNA 1 species ane l-inear rnolecules of abou L 4 ,5OO residues (Mr 1 .5 x 1 O6 ) wheneas the RNA 2 spec j-es are circular, covalentl-y closed mol-ecules of 300-4OO residues ( Mr 1 .25 x I O5 ) with a high degnee of internal base-pairing and have been tenmed virusoids as they share physical- propertj-es simil-ar to those of viroids.

So f ar, there are f our members of this ner^/ group of plant viruses; vel-vet tobacco mottl-e virus

( VTMoV ), solanum nodiflorum mottle vinus ( SNMV ) , l-ucerne transient streak virus (LtSV) and subterranean cl-over mottl-e virus ( SCMoV ). The best characterized of these ane VTMoV and SNMV which possess a bipartite genome since both RNA 1 and the virusoid RNA 2 are necessary for infection. The genetic functions provided by the two vi ral- RNAs have not been determined except for that of coding for the coat protein ( Gould et âf., 1 981 ).

VTMoV and SNMV are serofogicall-y reÌated wh1le cDNA-RNA hybnidization analysis gave estimates of sequence homology fon the vlral- RNAs 1 of between 20% and 50% 61 depending on the stringency of the assay conditons (Gould and HaLLa, 1981). 0n the other hand, hybridization anal-ysis indicated that the complete sequence of VTMoV RNA 2 (Mn 1.2 x lO5) is contained in

SNMV RNA 2 (l4r 1.3x 105) (GouId and Hatta, 1981). Despite the cfose sequence simil-arities between the RNAs 2 of VTMoV and SNMV, helther RNA will suppont the replication of the heterologous RNA 1 (Goufd et â1., 1981 ) which indicates a hi-ghJ-y specif ic relationship between the RNA 1 and RNA 2 of each vlrus. Al-though viroids and virusoids appear to share similar physicaÌ charactenistics, viroids are not encapsidated and repl-icate autonomousl-y (Diener, 1979; Gross and Riesner' 1980). In order to further lnvestigate t,he intriguing nel-ationshlps between the RNAs of VTMoV and SNMV, we have sequenced the RNA 2 species of each vlrus and compared their structunes with those of viroids.

MATERIALS and METHODS

A. Viruses and RNA VTMoV and SNMV I"Iere kindJ-Y pnovided by Drs.

R.I.B. Francki, J.V,l Randles and A.R Goul-d. Viruses b¡ere purified from lnfected Nicotiana cLevelandii and viraf RNAs isolated and purifled essentiai-1y as descnibed by Randles et al-. (1981). 62

B. RNase FingenP rinting

The RNase A and RNase T.,, fingerPrints of

VTMoV and SNMV RNA 2 were determined as descrlbed for

CCCV in Chapter 4

C. RNA sequence determínation

1) Partial- enzvmic di estion Specific Iinear RNA fragments vrere obtained from clrcular RNA 2 molecules by partiaJ- RNase digestion under non-denaturi-ng conditions as descnibed 1n Chapter 2, except that 150 units/mI of RNase T.' and 0-25 unlts/ml- RNas" U2 were required for VTMoV RNA 2 and 300 units/ml of RNase Tl and 0.25 units/ml RNas" UZ fon SNMV RNA z. The resul-tant RNA fragrnents were 5'-32p-labell-ed in vitro , f?actionated by polyacrylamide geI el-ectrophoresis and sequenced by the partial enzymic digestion technique as described previously. 2) Dideoxynucleotide chain termination As described in Chapter 2, RNA fragments produced by RNas" T1 digestion \^rere al-so sequenced using the dideoxynucleotide chain terrnination technique. Specific purified 5t - '-P-1 abetled fragments vJere dephosphorylated with cal-f intestinal phosphatase and polyadenylated, using E.cot1 poly(A) pol-ymerase.

out using d ( T C) as Sequencing reactlons were carried B 63 the speciflc Primer-

D. Synthesis and cJ-oning of double-strand cDNA Double-strandcDNAwaSsynthesizedfromSNMV

RNA 2 as described in chapter 2, and dlgested with the restriction endonucfease Sau3A I. The DNA fragment corresponding to residues 'l 31 to 216 of SNMV RNA 2 was purified and ligated into the BamH f site of the repticative form of phage M13 mp7 using Tr* DNA Iigase as described in Chapter 2, Recombinant phage vüere scneened by sequence detenmination using a specific M1 3 primer

( cTA, CGACG^C^AGT ) and the dideoxynucleotide chain 4¿¿ termination sequencing technique. Recombinant M1 3 (Birnboim and Dof 1979) repJ-icative f orm was isol-ated Y, ' digested.with Sau3A I and the cfoned insert purified on a 6% polyacrylamide gel (Sanger and Coulson, 1978; Maxam and Gilbert, 19BO) and used as a primen fon the sequencing of RNA 2 of VTMoV and of SNMV by the dideoxynucleotlde chain termination technique (Zlmmern and Kaesberg, 1978; Symons, 1978; 1 981 ).

RESU LTS A. RNase fingerPnints of VTMoV and SNMV RNA 2

Figure 5-1 shows the RNase A and RNase T.,' flngerpnints of both the VTMoV and SNMV RNA 2 molecules. The fingerPrints of the two RNAs share many sPots in Figure 5-1 RNase fingerprints of SNMV and VTMoV RNA

Purified circular sNMV and vrMov RNA 2 were digested

with RNase A or RNas" T1, 5'-32p-tabetled and separated by two dimenslonar ger er-ectrophoresis. The direction of finst dimension erectrophoresis is left to right, and the direction of second dimension el-_ectrophoresls 1s bottom to top. A, RNase A dlgested vrMov RNA 2; B, RNase A digested sNMv RNA

2; C, RNase T., digested VTMoV RNA 2; D RNase T., digested sNMV RNA z. origonucreotides unique to the vrMov or sNMV fingerprlnts are indicated by arrohrs. A B i¡ a a o'o t (t ,oa 'i , O {î\ a Ioì I .o\ I a +o I no-a O a a o a o t_ t* c D O O I t ID I O I (D I ID o oo a O \ o ì ì' r'\r¡ a O - O I z\ \ .1 '¡ o oo a o o î': ..(\ oo L t* 64 common, which confirms the exlsLence of sequence homology between VTMoV and SNMV RNA 2 suggested by cDNA-RNA hybridization studies (Gould and Hatta, 1981). However, the fingerprj-nts of each RNA species contain unique oligonucleotides showing that the smalfer VTMoV RNA 2 is not wholly contained within SNMV RNA 2 and that the RNAs are related but distinct species.

B. Primary stnuctures of VTMoV and SNMV RNA 2

The base sequences of VTMoV and SNMV RNA 2 r^rene determlned by using both the partial- enzymic digestion and dideoxynucleotide chain termination techniques with linear RNA fragments derived from partial RNase cl-eavage of the native circular RNAs 2. The sequence determination of RNA fnagments from both 5t-Lerminii, using the partial enzymic digestion technique, âod 3 r - terminii-, using the dideoxynucl-eotide chain terminat j-on technique, âf l-owed conf irmation of sequences and the resolution of occasional band compressions ( Kramer and Mi11s, 1978 ) which i^Iere seen on sequencing in one direction but not in the other. The complete sequences of the two RNAs wene evenbualì-y obtained from the sequences of numerous ovenlapping RNA fragments.

The complete base sequences of the two RNAs are given in Figure 5-2. Al- though the RNAs are Figure 5-2 The primary sequences of SNMV RNA 2 and the 365 residue form of VTMoV RNA 2 are shown in l1near form and aligned for maximum sequence homology. The 366 residue form of VTMoV RNA 2 has an extra UMP residue aL position 108 (arrowed). The sequence differences between VTMoV RNA 2 and SNMV RNA 2 are boxed. R.esidue 1 1n each case corresponds to the left-hand end of the secondary structure model of Figure 5-3. SNMVRNA2 . zo . ¡to . 60 GUUccUGcccuUGGGGAcUGAUuuUUGGuucGccuGGuccGUGUccGUAGUGGAUGUGUA GUUccuGcccuUGGGGAcUGAUUuUUGGuucGccuGGUccGUGUccGUAGUGGAUGUGUÂ VTMoV RNA 2 80 . 100 . 120 UCGACCUCGAC UCCACUCUGAUGAGUC AAGGACGAAACGGAUGUACCGCU UCUUG AAGGACG AAACGGAUGU ACCGC UUC UUG CUCGACCUCGAC UC.CACUCUGAUGAGUC :B: . lle 80 100 I

. 160 179 U C ACGCCC GC U GIAIGU AGA UGU AGU clqc AU A c u GG ac u AG; G au c G rccc lcGc u c a c tt uGlqGU aG AUGU AGU .ld. AU A CUGGACUAGUGAUCGAGGG AGGCUC c u c AcGc c cGc loo t?8 ' 110 ' 239 . 200 220 UAC clõ]ccauåucrtõlc GCG C UCC AAUGACUUGGGGUC ACUGUGUAA ollo UACUACAG UAC cl4lccAUGUGAluaic GCG CUC C AAUGAC UUGGGGUC ACUGUGUAA AIgG UACUACAG 2?0 238 t80 200 299 ?60 . AC G G u a o, o, u G A a co rïo.¡e1o GGGAGUCAAGGACGC GGGAGCUGGAC C CUCUCACC lt C GGGAGUCAAGGACGC GGGAGCUGG ACCCUCUC ACCAC G G U A G u GU U G A AGG U G C[gjA 280 ' 298 210 260 . 357 320 310 6çuF-¡T CCGGCAUCAGAG AUUGCAC ACCACCGGU AUC ACG UA ccc[E | 66c,o¡culccu[v_.1 CCGGCAUCAGAG GAUUGCACAC CACCGGUAUCACG - 310 320 300

360 377 ET.cAGGcUGGcAGGUAAc UCCÀGGCUGGCAGGUAAC ' 300 365 65 covalentl-y closed circular molecuÌes' the sequences are presented in linear forrn fon convenience and ease of comparj-son. sNMV RNA 2 consists of 377 nesidues while vTMoV RNA 2 consists of two approximately equimoÌar species, one of 366 residues which, like SNMV RNA, has a u at residue 108, and another species of 365 residues where this residue is defeted. This sequence heterogeneity within RNA 2 of VTMoV was determined by sequence anafysis of individual purified fragments which differed in size by one residue and which I^Iere denlved from either the 366 or 365 residue species. Further confirmation of this sequence hetenogeneity and an estimate of the nefative proportlons of the two species vrere obtained using a cloned DNA fragment (derived from residues 131 to 216 of SMNV RNA 2) as a primer on the mixture of the intact vTMov RNA 2 species (see Figure 2-B\. This 365 residue species is arbitraril-y presented in Figure 5-2 and numbering of the VTMoV RNA 2 sequence witl refer to this sPecies. The extensj-ve sequence homology beLween VTMoV and SNMV RNA 2, originalJ-y reported on the basis of hybridization anal-ysis with cDNA ( Gould and Hatta, 1 981 ) and show by RNase fingerprinting, is confirmed by the sequence data. As suggested by RNase fingerprinting data, each RNA 2 species contains unique sequences, thus

95% of VTMoV RNA 2 is homologous with SNMV RNA 2 and 92% 66 ofSNMVRNA2ishomologouswithVTMoVRNA2.The sequence diffenences are unevenly scattered throughout the two RNAs with a cluster of base differencés around residues 339-359 0f SNMV RNA 2 and residues 333-347 0f vTMoV RNA 2, while there is. al-most compÌete sequence homology between nesidues 360-146 of SNMV RNA 2 and residues 348-1 45 of VTMoV RN A 2.

C . Secondary s truc tures of VTMoV and SNMV RNA 2 Secondary structure models for the two RNAs ( were constructed as described by Tinoco et al- . 1 97 1 ) and are shown in Figure 5-3. Both RNAs form extensively base-paired rod-like structures which are simifar to those described for viroids (Sangen et â1 ., 1976; Gross et af. , 1gB2; Hasel_off et â1., 1982) . The structures are consistent with the known sites of high sensitivity to ribonuclease under the conditions of high sal-t concentration used to generate specific RNA fragments from the circular RNAs for sequencing ' Thus, the terminaÌ singl-e-strand hairpin loops and the centraf singJ-e-strand regions of both RNAs ( residues 70-1 00 and 285-305 ) \^rere especial-1y susceptible to RNase cf eavage. The properties of the proposed structures are summarized (Table 5-1) and are compared lo those of the pubJ_ished structures of four viroids. The vTMoV and sNMV RNA 2 mofecules possess proportions of G:C base Tab I e 5-1 Properties of proposed secondary structures for Rl'lA 2 of VTMoV and SNMV compared with those of several viroids

No. of base G:C Res i dues No.of pai rs base pai rs base (K¡/mol^G* at RNA resi- as% pa'ired 25"C in dues A:U G:C G:U of total lo lM NaCl )

VTMoV RNA 2 365 3B 72 l3 59 67 - 350

VTMoV RNA 2 366 38 71 l4 58 67 - 345

SNMV RNA 2 377 4l 76 20 55 73 -455

ASBV 247 43 28 12 34 67 -280

PSTV 359 37 73 l6 58 70 - 610

CSV 356 44 64 l6 52 70 -540

CEV 371 34 72 l8 58 67 -590

*Parameters for calculation províded by Dr. D. Riesner (Steger, Gross, Randles, Sänger and R'iesner, personaì communicatjon). Figure 5-3 Pnedicted secondary structures of SNMV RNA 2 and the 365 residue fonm of VTMoV RNA 2 plus the segment of the 366 residue form of VTMoV RNA 2 contaì-ning the extra UMP nesidue (residue 108). The sequence differences between the RNAs are boxed. SNMV RNA 2 c 20 40 60 ,ao AO 100 120 140 160 't80 U luu I AG G c I u c cc GA Ac \r" UU ec \r ru[]e r g I \uEl ^ . "GCU GUAGA ucuecu.c@"^ ,^t uo^" U I ccuGcc CUUGGGG UG UUGGUUCGCCUGGUC GUG CCG GUGG UGUGU.AUC.CACUCUGAUG CC GG CG UG C UUCUU G ccAccu.ccAc.cuc cu.AGuG G^GGG GGcuc ccEluc^ cGc "^ c o ou o cq^cco corccu@ e$cs{þFrucrþ .cAc Goc cacc.acAcG u rc Æ-uìe¡c¡curc GG çC.GC o c crþþ o¡ l-ülc¡c cuccc uco^G.qqo scEl suc cþþ'c eucB rcruce oþ ruo - ^c.uo ^gug uc UU cl o é)l"u l" \à^ u c o u AA u/, ^u \9 I I g ^o l^ I \"^\"" I G "l¡c o l^ 377 360 - 340 .l; 300 260 260 240 220 20,0

VTMoV RNA 2 'r20 160 't80 20u 60 .go " u .1oo 140 G ug u G GG U c c I lu c A cA uc cu G cc U UA A c AU A U CC U U c rI U c I ^ u E ^l oueo ucurou'{ u^c c^ u3^c U CCUGCC GG cuG UGGUU GCCU UôCGUG CCG GUGG UGUGU.AUCC UCUGAUG C c GG CG UG AC UUCU GC G^CCU.CG^C'CUG CU^G.UG GAGOG uc cc u.c coc.ueþ1. ::: o c G s GGACGG. UC aee.[þccec GGC. C^CC'ACACG U^GG AG^CU^C GG CC GC AC.UG GGGA CG CUGGA SUUG GAU r@l cucc c ucg^G.gcc ous ;;; :i;; ;;;;;; lä ;;; :; l;;; ^" cG A u-^^ ,1" ^^r G UU AUG ccu , Eu Et G lE^n I I ^s l SG @ 2OO 365 360 320 300 280 260 ^c 240 22O

100 G c U c uucu fl erccu

GGG CG CUGGA El 67 pairing which are l_ower than that of cccv and higher than that of ASBV, but which are similar to those of the similarly slzed PSTV, CSV and CEV. The thermodynamic stabilities of the proposed model-s were cal-culated using vafues kindly provided by Dr D. Riesner (steger, Gross, Randles, Sanger and Riesner' unpublished data) ' The val-ues of -455 KJlmol for SNMV RNA 2 and of -345 and (Table -455 KJlmo1 for the Lwo forms of vTMov RNA 2 5-1 ) are consistent with their thermaL denaturation properties; thus VTMoV RNA 2 gave a Tm of 57oC in 0.15M NaCl, O.O'l 5M sodium citrate, pH 7, while SNMV RNA 2 gave a highen Tm of 64"C under the same conditions (Goul-d and

Hatta, 1 981 ; Goul_d, 1 981 ) . The pnedicted stabilities of the RNA 2 0f VTMoV and of SNMV are lower than those of the simitarl-y sized viroids PSTV, CSV and CEV but highen than those of the smaller CCCV and ASBV (TabIe 5-1 ).

D. Possible PolY peptide translation Products from RNA 2 species and theÍr conPl-ements Since the RNA 2 species of VTMoV and SNMV are nequÌred with the homologous RNA'l for viraÌ infection (Gould et â1 ., 1981 ) , the RNA 2 mol-ecules must code for some protein product ( s ) and/or contain structuraÌ information essential for viraf replication. Evidence suggests that an AUG codon, hot necessaril-y that neanest the 5t terninus of the mRNA, functions as the initiation 6B signal- f or eukaryotic mRNA translation (Banal-1e and

Brownlee , 1 97 B; Ko zak, 1 982; Lomedlco and McAndrew, 1982) and that eukaryotic ribosomes do not interact with cincular RNAs (Kozak, 1979; Konarska et âf ., 1981). Therefore, Lranslation of the RNA 2 species woul-d require the existence of specific Ìinear RNA fonms. Assuming that this condition is met, the extensive sequence homology between SNMV RNA 2 and the 366 residue form of VTMoV RNA 2, and their putative complementary RNA sequences, aIl-ows them to code for several- slmil-ar polypeptide products ( figure 5-4). However, only one small polypeptide product is shared between SNMV RNA 2 and the 365 resi-due form of VTMoV RNA 2 and their compl-ements. AIl possible transl-ation products are l-ess than 100 amino acids in length and therefore the genes coding for the viral- coat protelns (approximately 300 amino acids (Randles et â1 ., 'l 981 ; Hollings et âl ., 1979) ) must reside in the RNA 1 species.

DISCUSSION

In overal-1 structune, VTMoV and SNMV RNA 2 resembLe viroids in being smal-1 single-strand covaJ-ent1y cfosed circufar RNA mol-ecules which form rod-like native structures with extensive base-pained regions interspensed with single-strand regions. Howeven, in contrast with viroids which replicate autonomously and Figure 5-4 Possible polypeptide products of SNMV RNA

2 and the 365 and 366 residue forms of VTMoV RNA 2

(A), and their putative comlementary RNAs ( B ) are s hown in schematic form. Each possible transfatlon product is given with the residue number of the flrst base of the initiation codon plus the termination codon ( s ) in parentheses. For the compl-ementary sequences, the same residue numb e rs ane retained and therefore run in the 3 ' to 5t direction. The cl_ear aneas represent regions of amino acids sequence homoJ-ogy and the bl ack areas regions of non-homology in the RNAs shown. The cross-hatched areas are regions of sequence homology between different products of the two VTMoV RNAs which correspond to reglons of non-homology in SNMV RNA 2. l¡'lhere an identical product is obtained fnom the two forms of VTMoV RNA 2, only one is shown. (amino acids) LENGTH OF POLYPEPTIOE PFOOUCT lst nosrdu" ol 60 t0 !o 90 ro0 FNA 2 to 20 30 ao so Spec'as Coóon (u^G) A VIMoV (365) 54 (uGA) VIMoV (366) 5a (uG^t SNMV 51

(u^G) vIMov (365) 70 (UAG,UGA} vlMoV (366) ?o (UAG UGA) SNMV ?o

(UAG) vIMov (365) 93 (uGA) VIMoV (366, 93 (UGA, 93

(365) (U^G) 184 (366) (UAG) SNMV 185

(365) 303 (uA G) (366) sNvv 30¡ (UGA)

(365) 226 (UAG) (f,66) SNMV (UAG)

(u^c) r75 (UGA) vtMov (366) r76 (U^G) SNMV 117 69 are not encapsidated ( Diener, 1 979; Gross and Riesner,

.l gBO ) , the RNA 2 species are essential components of a bipartite genome and are encapsidated (Gould et â1.,

1981 ) . !,¡hi1e viroids do not appear to code f or functional_ protein pnoducts ( conj ero and semancik, 1 977 ; Davies et â1., 1974; Semancik et â1., 1977; Haseloff et â1., 1g82]t, thene 1s no information available for the RNA 2 speci-es. The fack of conservation of possible translation pnoducts between SNMV RNA 2 and the 365 and 366 residue forms of vTMov RNA 2, desplte greater than go% base sequence homology, suggests either that the 365 residue form of vTMoV RNA 2 may be non-functional or that RNA 2 coded translation products may have no f unction in viral repl-ication. Although the invol-vement of RNA 2 coded transl-ation products in vlral replication cannot be excluded, 1t seems likeIy that the unique viroid-like structunes of the RNA 2 molecul-es encode some function besides that of a templ-ate, and that this

function is required for the replication of both RNA 1 and RNA 2 species. Since neither VTMoV RNA 2 nor SNMV RNA 2

supports the replications of the heterologous RNA 1 species (Gould et âI ., 1981 ), the bioJ-ogical- speclficities of the RNA 2 species must be determined by differences in primary and/or secondary structunes. The 70 only extenslve region of sequence differences between

VTMoV and SNMV RNA 2l-ies around VTMoV RNA 2 residues 333-347 and SNMV RNA 2 resdues 339-359. Hence, this reglon may be involved in determi-ning the specificity of the relationship between the RNA 1 and RNA 2 species, although the involvement of othen structural differences cannot be excluded. An unexpected complication during the sequence determination of the RNA 2 mol-ecul-es v{as the occurence of sequence heterogeneity in VTMoV RNA 2 whÍch consisted of two RNA species differing in the presence or absence of a U residue at posltion 108 (Figune 5-2, 5-3 ) and existing in approximatety equlmolar amounts. The two RNA species may have arisen either by a single rnutation in a panent molecule fol-l-owed by independent repllcation of the resultant two RNA species, or a mixture of the two species may be produced during each cycì-e of repJ-ication by transcriptional and/ or processing events. In each case, iL is possible that one of the two RNA species may be non-functional. Sequence heterogeneity within RNA populations has been reported for RNA phage Oß (Domingo et a1., 19781,,

vesicular stomatltis virus ( Holland et â1. , 1979) ,

satellite tobacco necnosis vinus (Donis-Kel-l-er et âl . , 1981), citrus exocortis viroid (Gross et âl ., 1982) and coconut cadang-cadang vinoid ( Hasel-off et â1., 1982). 71

In the secondary structures of the viroids PSTV, CSV, CEV and CCCV there is a central reglon of the native rod-like structures which is highl-y conserved in both sequence and structure ( Hasefoff et âf. , 1982) - ASBV (Symons, 1981 ) does not share this common structure except for the residues GAAACC (ASBV nesidues 45-50) which, as in the other vlroids are present on a single strand l-oop in the central- region of the native molecule. Interestingì-y, VTMoV and SNMV RNA 2 also contain the sequence GAAAC (nesidues B6-90 in both mol-ecul-es ) which is al-so present in a singJ-e-strand region in the centre of the proposed secondary structures. Howeven, there 1s no extensive base sequence homology or compl-ementarity between the vinoids

PSTV, CSV, CEV, ASBV and CCCV and t,he VTMoV and SNMV RNA 2 mol-ecuf es. It, will be of considerable interest to determine the exact mechanisrns by which VTMoV and SNMV support the replication of the homologous RNA 1 species as wel-1 as the molecular basis for the speclficity between the RNA 1 and RNA 2 species, and to determine whether the common structunal- features of viroids and virusolds mirnor some common function. Chapter 6 outÌines work with subterranean cl-over mottle virus which al-l-ows some definition of the virusoid sequences involved in such relationshiPs. CHAPTER 6

SUBTERRANEAN CLOVER MOTTLE VIRUS 72

INTRODUCTION Several lsolates of subterranean cfover mottle virus ( SCMoV ) from Western Austral-ia have been

described ( Francki et âf. , 1 983 ) . As judged by electron microscopy, pFeparations of SCMoV consist of honogeneous populations of polyhedral- virus panticles about 30 nm in diameLer, and serological tests using antiserum to punified SCMoV failed to reveal- any antigenic diffenences between the isol-ates. However, the various

SCMoV isolates appear to contain differi-ng encapsidated RNA components, i¡ühile al-1 isol-ates contained a J-inear, singJ-e-stranded RNA species of approxlmateJ-y 4500 residues in size (RNA 1, Mr 1.5 x 106), each isol-ate contained either one or both of Lwo viroid-like clrcul-an RNA species, RNA 2 (approximately 400 resides) and RNA

2t ( approximately 300 residues ) . The SCMoV isol-ates used in this wonk v,/ere the isolates A, B, D and E of Francki et aI. (1983). Some time after i-ts isolation SCMoV-A Idas found to contain

both RNA 2 and RNA 2t with the RNA 1 species. In

contrast, SCMoV-E contains onJ-y RNA 1 and RNA 2 whil-e

SCMoV-B and SCMoV-D contain onJ-y RNA 'l and RNA 2t. The fractionated RNA components of these isol-ates are shown

in Figure 6-1. Interestingly, while isol-ates A, D and E urere obtained as f ield isoJ-ates, SCMoV-B was produced from SCMoV-A by passage through single l-esions on pea l-eaves. Figure 6-1 RNA components of SCMoV isolates. RNAs extracted from SCMoV isolates E, A, BandD r^rere fractionated by electrophoresis 1n a4% polyacryl-amide geI containing 7M urea. R NAs $Ie re detected by slaining with tol-uidine blue. RNA 1 and the circular forms of RNAs 2 and 2t are shown. The linear forms of RNAs 2 and 2t.migrated from the gel. SCMoV ieolates

EABD

-fl-{--li-it--.'-

*s- t

4S TBE +7M uroâ 73

The finding of SCMoV isolates which differed in their RNA 2 and/or RNA 2t components suggested either

( 1 ) the existence of two serologically indistinguishabfe viruses which each possessed different RNA components or (2) the existence of a single virus whlch could support the repì-ication of either or both RNA 2 and RNA 2t. The fol-lowing work describes efforts to distinguish between these possibllities and to determine

the structures of SCMoV RNAs 2 and 2t .

MATERI A LS Isol-ates of SCMoV were kindJ-y provided by Dn. Richard Francki. Viral propagation, purlfication and extraction of viral RNAs I^¡ere perf ormed by Mr. Chris

Davies as described ( Francki et âI. , 1 983 ) . SCMoV RNA species vüere purified by polyacryl-amide gel el-ectrophonesis (Gould, 1981 ) , frostly by Mr. Dav j-es.

ME THODS A. Synthesis and restriction endonucl-ease cl-eavage of

ds cDNA

Random-primed first strand cDNA was transcribed from SCMoV RNA 1 essentially as descrlbed by TayJ-or et al-. (1976l, . Purified SCMoV RNA 1 (2 Ue) bras resuspended in 25 p1 containing 50mM Trls-HC1 pH 8.3,

5OmM KC1, 'l OmM MgClr, l OmM DTT, 1O0pM Io-3tr] acrr 74

(fO ¡rCi), 5OOUM dATP, dGTP and dTTP' 2 mg/mI DNase 1 treated salmon sperm DNA and 20 units avj-an myelobl-astosis virus reverse transcriptase. After incubation aL 37oC for 2 hours, the reaction I^¡as boiled and treated wlth RNase A befone synthesis of second-strand cDNA as described in Chapter 2, Me thods B-1. Synthesized doubfe-strand cDNA was digested with varlous restriction endonucfeases and fractionated by polyacrylamide gel el-ectrophoresis as descrlbed al-so in Chapter 2, MeLhods B-1,

B. Fingerp rinting of SCMoV RNAs Purified circul-an SCMoV RNAs 2 and 2t were RNase A fingerprinted uslng techniques outlined in Chapter 4.

C. Sequence determination of SCMoV RNA 2 and RNA 2l Purified SCMoV RNA 2 and RNA 2t were each subjected to partial- digestion under non-denaturing conditions, âS described in Chapter 2, using 150 T1 0.1 RNase A or 0.25 units/ml- units/ml RNas" ' vc/nl- RNas. UZ. The nesulting linear RNA fragments were eithe r 5' - or 3'-32P radiolabelled, fractionated by polyacryl-amide gel el-ectrophoresis and sequenced using the partial enzymic digestion technique essentiaJ-Iy as descrlbed in Chapter 2, except that sequencj-ng gels 75 contained TBE buffer, 714 urea and 25% (v/v ) Oeionized formamide. The sequences of overlapping RNA f n agmen ts wene used to obtain the primary stnuctures of the circular RNAs and the methods of Tinoco et a1- (1971) used to predict the native secondary stnuctures of the mol-ecules.

RESU LTS A. Analysis of SCMoV RNA l nucl-eotide sequences In order to analYse the sequence rel-ationships between the RNAs 1 of the SCMoV isolates, double-strand cDNAs I^Iere transcribed from purified RNAs 1 and then digested with the various sequence-specific restriction endonucfease, Alu I' Hae III, Hha I, Hpa II (see Figure 6-2) and Acc 1 (data not shown). PoJ-yacrylamide get electrophoresis of digested ds cDNAs results in gef pattenns which reflect the anrangement of nestrlction enzyme recognition sequences on the original 6-2 all- four RNA 1 molecules. As shown in Figure ' isol-ated RNA 1 species gave rise to indistinguishable patterns of ds cDNA restriction fragments for all

enzymes used. llhi)-e, ( 1 ) the synthesized ds cDNAs may

not be wholJ-y repnesentative of the respecti-ve RNA 1 species, and/or (2) small nucleotide sequence differences between the RNAs 1 may exist which do not affect the size or number of observed ds cDNA Figure 6-2 Restriction endonucl-ease digestion of SCMoV RNA 1 ds cDNA.

Double-strand cDNA was synthesized from RNA 1 purifled from four dlfferent isol-ates of SCMoV. Isolates 1, 2, 3 and 4 correspond to SCMoV-8, SCMoV-4, SCMoV-B and SCMoV-D, F€spectlvely. The ds cDNAs vüere untreated or dlgesfed with AIu I, Hae

III, Hha I or Hpa II and fractionated on a 6% polyacrylamide gel containing 2M urea.

3,-32p-tabell-ed Hpa Ir digested M13 mp7 RF vras incfuded as size markers, with sizes of 1596, 829, B1B, 652, 545, 543, 472, 454, 357 t 183, 176, 156, 129, 123, 79 and 60 base pairs. RESTRICTION DIGESTS 0F SCMoV RNA 1 ds cDNA

UNCUT Alu I Hoe III Hho I Hpo II

SCMoV IS0LATES 123 4123t+W|TTTTÑ --axa- -!rI

i I I I I i

a I ? I I t ì I I \ t lí I ¡ I å

F .Ç I I .1 i. \ I a { f' F ì F t ta t I t I \ L t- tÉ -¡iÒ p \ -lr -lr .l' 76 nestriction fragments, the results do suggest that the four isotated RNA 1 sPecies are certainÌy closely rel-ated in nucf eotide sequence t and may be identical.

B. RNase fing erprinting of SCMoV RNAs 2 and RNAs 2t In addition, the sequence reÌationships between the RNAs 2 of the SCMoV isol-ates were investigated using the RNase fingerpninting technique. Purified cincul-ar RNA 2 and RNA 2t species vüene digested 2P-radiolabelled to compl-etlon by RNase A, 5'-3 and fractionated by two-dimensional polyacryÌamj-de gel electrophonesis as descnibed in Chapter 4. The resulting oligonucleotide patterns are glven in Figure 6-3 and show the f ollowing. ( I ) f rre ol-igonucleotide patterns obtained from the RNAs 2 of SCMoV-E and SCMoV-A are essentially identical . (Z ) fne ofigonucleotide patterns obtained from the RNAs 2t of SCMoV-4, SCMoV-B and SCMoV-D are af so essentialJ-y identical-. (3 ) Sorne ol-igonucleotides appear to be common to both the RNA 2 and RNA 2t specì-es (see Figure 6-3). Although RNase A (C and U specific) fingerprinting may not reveal- minor sequence differences present in pyrimidine-rich regions of the molecules, the data suggest that, l-ike the isol-ated SCMoV RNAs 1 , each RNA 2 and RNA 2t species is eithen closely rel-ated or identlcal to similan species from different SCMoV Flgure 6-3 RNase A fingerprj_nts of SCMoV RNAs 2 and 2t.

Circular RNAs 2 were purifled from SCMoV-E and

SCMoV-A r âhd circular RNAs 2t r^rere purif 1ed f rom

SCMoV-4, SCMoV-B and SCMoV-D. AlI RNAs r¡¡ere digested with RNase A, 5,-32p-tabell-ed and fractlonated by 2-dlmensional polyacrylamide ge1 el-ectrophoresis. The resulting oligonucleotide flngerprints are shown with the dinections of el-ectrophoresis in the first dimenslon being left to right, and, irt the second dimension, bot,tom to top. ,ZYNU ,ZYNU ,ZYNU ZYNU ZYNU o-ôncs 8-ôncs Y-ônc8 Y-ônc8 3-ônc8

o o

a o

o O a o a o o a a a o a t J o a o a o I |¡ O a I o t ,J.Or 0. JI I o t ¡ ¡O O O o 0 , , 4 I o o Ç 77

1soÌates. The possible identity of the different RNA 2 or RNA 2t species is also supported by the identical fragments obtained after partÍaI RNase T uz orA cJ-eavage under non-denaturing conditions (see Methods, this chapter) of RNA 2/2t fnom different isolates, and by preì-iminary sequence data (not shown). In contnast, there are considerable differences between the oJ-igonucleotide fingenprints, and thus primary sequences, of the RNA 2 and RNA 2t. So while some degree of sequence homol-ogy 1s indicated by lhe number of shared oligonucfeotides, RNA 2 and RNA 2t each contain unique sequences and do not differ simply in the possession of repeated sequences. From the evidence presented, it seems 1ikely that the different isol-ates of SCMoV, which are serologically indistinguishabl-e ( Francki et âf., 1 983 ), contain essentially identical RNA 1 species and differ only in containlng either or both RNA 2 and RNA 2t specles. Th is is al-most certainlSr the case f or SCMoV-4, containing both RNA 2 and RNA 2t and SCMoV-B (which hlas derived directly from SCMoV-A) containing only RNA 2t. Furthermore, although RNA 2 and RNA 2t appean to diffen significantly in nucleotide sequence, no sequence differences v\rere observed between simil-ar RNAs 2/21 obtalned from diffenent SCMoV isolates. 7B

C. Sequence determination of SCMoV-A RNA 2 and RNA 2t Linear RNA fragments v¡ere obtained from the RNA 2 and RNA 2t of the SCMoV-A isolate by partiaì- ribonuclease digestion under non-denaturing conditions. These fragments v¡ere radiolabell-ed and sequenced using the partial- enzymic cJ-eavage method, and the sequences of overlapping fragments vlere assembled to glve the compl-ete primary stnuctures of the circul-ar molecul-es. The RNA 2 and RNA 2t specles each consists of 388 and

327 nesidues respectively. In addition, the two RNA species share a singJ-e common region of abouL 220 residues of aÌmost complete sequence homoÌogy. The pr imary structures of SCMoV-A RNA 2 and RNA 2t are presented in Fi6Sure 6-4 in a convenient linear form with the shared sequences shown boxed. Secondary structure models for these mo.l-ecufes were constnucted using the methods of Tinoco et at. (1971) and are shown in Figure

6-5. Both R NAs may base-pair intramolecularl-y to fo rm hel-ical rod-like structures similar to those of VTMoV and SNMV RNA 2 and viroids. Strikingly, the nucfeotide sequences conserved between SCMoV-A RNA 2 and RNA 2t are located so as to form the entire base-paired left-hand sides (as drawn) of the native mol-ecul-es. Consequently, it is the differing lengths of the unconserved right-hand sides of the natj-ve molecul-es which account f or their dif f erence in si-ze. Figure 6-4 Primary structures of SCMoV RNA 2 and

RNA 2I .

Nu cfeotlde sequences common to the two RNAs are s hown boxed . ItO SCMoV R NA 2 A6A.66CAU A C CC UC C UC6 C6 6AUUUU6AA66 UG UU UC AGCUACCCAA A l6 SCMoV RNA 2' A6A66CAU UU C C C UC C UC6 C66AUUUU6 AA6G U6 UU CU A6CUACCCAA t

UAUUC CACG C U6 UC U6UACUU A UAU C A6 UACAC UGACGA6 U CC C UAAA66

UAUU C C A C6 C U6 UC U6 UAC UU 6 UAUC A6 UA C A C U6 A C.6 A6 U C C C UAAA66 JO 1ú t+0 AC6AAACA6C6CACC6CAA C U U 6 G C C'A 6 A C C U c6ccAAu c A cccccAcAc. AC6AAACA6CGCACC6CAA UCUAC6UAUACCC C6AUUC6 AC UU6CUU66A tú ,la ta, tr, C AA6 C CAAAAACC 66 U CC C C AAC6 C A6 UUUAG UAU C AA6 UC6 U CG C AUC C 6CUA6C6UUC6ACA6A6U6C tLo

zØ 1¡0 40 AC6 CUC C CGA6G6 466 AA6 UUUG C6 C C UU6A66 U UCU6CAC66UCGU66U cGc6r 6GAC6CGGUUCU6GUC lro

2aD lío AACA66 AAAA6 U6 UU66AAU6 UUU6 AAG6 U CU U6 C 6 6UUGUCAA66ACC A

A C A C U C A C C C 6.G.6 A G 6 C C A U C G 6 6 C A 6 A U U A U A C U A 6UUGUCAA66ACC 2oo 1ú tt A 6UC6UUAGU 6U UATUAUA U AUUACUAC c TUAC6U6UUAC U UU6UUA66U

U 6UCGUUA6U UC UACUAUA c AUUACUAC A CUAC6U6UUAC UU6UUA66U úo l¡a llo ¡o¡ 66CCCCA(CUTACUUUCG c 6AAG6CUA6 6A AAC6UCCAU

66CCCCACCUCACUUUC6 U 6AA66CUA6 A6 AAC6UCCAU )æ !¿o tz7 Figure 6-5 Secondarv structures of SCMoV RNA 2 and

RNA 2I . NucÌeotide sequences common to the two RNAs are shown boxed. RNA 2 SCMoV 200 120 t1 o, 1ó0 180 ó0 80 100 20 10 CA cc rt cc oouur* LUL!c- c AAA AA. cc u'., ( CCAAU cc Gcca ACC0 OIJ A ^ u rr^ CA6AC ^r8 ôor'nu"o' - OG UOGU cGu ;;;;;. 0c A IJC ;; .;;i;;^*oîiiîå.. 'o7'*o***ou^ôouuol ^ CA c6 I 280 260 2r{, 220 380 3ó0 3t0 )20

SCMoV RNA 2' 60 80 120 1¿0 20 rc ó0 A CG corou^ c .^/ c u \ c accoo,, uc cuuG uun^Gau^ocoru U

¡c GCG c W û "7u^;ooo"-r9oôoJ\î'*"'**T.

2 2lr0 220 200 -100 321 300 280 79

DISCUSSION

A. Rel-atlonships between the various isolates of SCMoV The four SCMoV isolates used in this wonk have nov\r been shown to be indistinguishabfe serologicaì-l-y (Francki et âl . , 1983 ) , to contain indistinguishable RNA 1 species and are likeJ-y to dlffer only in containing either one or both RNA 2 and RNA 2t. Glven this, then irrespective of whether SCMoV RNAs 2/21 are required for viral infection, âs appears to be the

case with the RNAs 2 of VTMoV and SNMV ( Gould et âf. , 1981 ), or are satellite RNAs, as appears to be the case

for LTSV RNA 2 (Jones et âf., 1983), it seems that SCMoV RNA 2 and RNA 2t must be functionally equivafent, LhaL is, interchangeable.

B. Se uence homolo between SCMoV RNA 2 and RNA 2t

The native structunes of SCMoV RNA 2 and RNA 2t show remarkable conservation of the left-hand sides of the mofecufes (Figure 6-5). It therefore seems

reasonable to propose that these common sequences and structures mirror the apparent interchangeabl-e functions

of the mofecules. Thus, if RNA 2 and RNA 2t ane satell-ite RNAs, the conserved left-hand sides of the molecules may contain recognition signals required fon replication by viraf and/or host componen ts If the RNAs are functionally simil-ar to VTMoV and SNMV RNA 2, BO

the left-hand sides of the rnolecul-es may al_so contaj_n some undefined function required for viral replication.

InterestingJ-y, the conserved sequences of SCMoV RNA 2 and RNA 2t share homotogy wlLh VTMoV and SNMV RNAs ?.

C. Sequence homology between SCMoV, VTMoV, SNMV and LTSV

RNAs 2 The determined sequences and pnedicted

secondary structures of the viroid-like RNAs of SCMoV,

VTMoV and SNMV and two isolates of LTSV ( Keese and Symons, unpublished resutts) are shown in Figure 6-6, and common sequences indicated. First, the sequence

GAUUUU 1s present on al_l_ RNAs in a simitar position on the native structures (startlng at approximately resj-due number 20 in atI cases). The conservation of this sequence suggests that it may play some rol_e which is common to the repl-ication of aÌ1 six RNA.s, including the

LTSV RNAs 2, Second, the vj_rusoids of SCMoV, VTMoV, and SNMV all contain conserved sequences whlch are centnar to their native structures. These consist of two regions, one of 24 resldues and the other of 9 residues, which are postioned on opposite sides of the rod-like molecul-es. SimiJ_arJ_y, viroids af so contain highly conserved sequences which are centraÌ to their rod-like native sturctures (chapter 4), and the pentanucreotide sequence GAAAC is present, predorninantly Figune 6-6 Seouence and structural homology between

SCMoV RNA 2 and RNA 2t and SNMV, VTMoV and LTSV RNAs

The pnoposed secondary stnucture model-s fon the RNAs ane shown wiLh conserved sequences indicaLed in gneen.

LTSV-A RNA 2 40 60 120 140 60 I I I GU t_ G cc c G A CG c U I U c G € ^u ^ "û"": ^ GGAU GU cc cu uccA ccu cc^ .,d:.?iù,."t"-r"I! o :i.l',:+ CAAIJUGAG ACOUUUCGGCC'UGCC GGCC.UC UCAOUG^G CGG ucAo 1-c :i ^t:ii"l"liìr+ "" rccrculuc U cu cpþ UA cue uc U ccu^c GA ACGU. GGA.GOU CAGU ccc r¿:;i¡ GUUAA CU C UGCAA^GCCGG AUGG ccGG AG CA AGU AGUC u þþcu.[þco j ^CUCACUC.GCC u I c U C uu c ñì--' AG cu^ l, cc c c q ^G ^GG ^ I ^wg';'^ I 324 u ls I I I l" 300 2ao' 260 240 220 200 80

LTSV-N RNA 2 40 60 80 1O0 12O '140 160 2l I c cl c l_l^t I cflt ÀU C G A o I ^+-,¡ñ?¡-s:{"" ^ jo^ j. c "l ô'^ cG cu ucc^ ccu ccA I u o u,,. o o u o ucAs " +:,uþFlì'q"^"" ! l-c : ::::: : " "díïìt:" "Fr:+-s :i"$;'¡. " ^,, ^ " " " "^ @"^ rê u cu uc ucu ccu^c c c u. c s GG u c c u c o c ð¡-c-uf, c fu?t'dcl cu u^ c uc UGCAAAGCCGG UGS ccGG AGUC^CUC cc U "Fj3;å:"--"-"1 Êcu^[crftcco ^GC ^ ^ I ^. ^ !---J ^ ^G ^ ,, c \J c u u r cc A I G c l. G o / " | '-' uo ", I ^c ô c ^tto/or'^ ^ c 324 I I ./t ^l"l\þ"1 I 320 300 280 260 ', 240 220 200 180 StMoV RNA 2 ,180 2 00 80 100 120 140 160 20 40 ó0 lca,L lc c t, /r/ c--c clc A AA otoo /uou co oc,n"t^ct 3u cÂGac ccAAU Aa cc 0c CA Àcc0 OU AÁc- JLUao- cAAGg uc c CG coo oc io* .;;i Yo^ uc cc 'i' GA Â C /// A^;;;, A^;;,.; / AAôo^ ca ûoo,o. 280 260 24 220 ô 340 I60 3r0 120

SCMoV RNA 2 120 110 tó0 20 t0 60 80 \- .t ccu¡u^ c c u c ,ccoou u c ^l uc o^oau^otnru ccu oc

u' o o oo c^cu Go *"ufc ¡oococ U u^;oo E ^ "u^, ^ 7 J\ ^crê 2oo 1oo )27 300 280 ?60 260 220 "

180 SNMV RNA 2 100 120 140 160 1 20 40 I Ac c" \o[ \,. ^ cc I "^u^c c^ uo^c I Bu cGcuc ccPlucr ccc ccu GUAGA uourou.cþl uucuuo c c^qGo ' o G CCUGCC CUUGGGG ^c c oo[\ ouo cþþ c¡ucft ¡c¡uc¡ cEì ruo ou Acuo o c o cr[þ OCUOOA GUUO GAU CA c cuccc ucc^q'c6o o GGACGC GGAC c uF¡-îl cfã¡lc o s Gf,ihF¡-tã¡c c^cc.^c^co u¡c,6ü1c¡c¡c UAC . GG Gq lro u l^ .L\ U Ug UO cu lc \g^ vc ^a ul' u uc UU lu d"f cl Nlt,8,ß, üi ^co I - a I ^ I I 200 I I ^c 220 I \.^\,.! l- 280 260 240 377 360 340 320 300

180 140 c 160 G 100 120 VTMoV RNA 2 I I 40 cc cA uc a c Ac rI A U uql A U CC U UC ^cl uoueou'{ u^c c^ uoAc u cu c c GS I G GCUC C c coc.uGlg cueo UG AC UUCU OC GAGGO U ccuqcc s6 ^c oquu Gccu uçccuc o q qUO c^uc c@ r.uo ou ecuo o ouuo oru oe@l rc CUCCC UCg^C'3qg o ccc.[þccec C^CC.ACACG UAGG.^GACU^C'GG CC'G ccG^ cc. cucG^ ^cC o lfãlo^c^uc^ u -^^ ul t C C U AG l-^l u U Et I AC G U llluc 240 2zO 2OO 2AO 260 365 360 32O 30O B1

in the central conserved sequences of singÌe-stranded ' SCMoV, VTMoV, SNMV virusoids and all sequenced virolds. However the virusoid mol-ecules do nol appear to contain the conserved stem-J-oop structure ( Gross et â1., 1982) or postulated Ula snRNA homology ( Diener, 1 981 ; Gross et âf., 19BZ ) which are present in all- sequenced vinoids except ASBV. The GAUUUU and central conserved sequences which are shaned by the SCMoV RNAs 2/2t and VTMoV and sNMV RNAs 2 are focated within the left-hand sides of the native mol-ecules which, in SCMoV RNAs 2/21, are highJ_y conserved. In contrast to the SCMoV virusoids, VTMoV and SNMV RNA 2 will only replicate in conjunction wlth the RNA 1 species from the same virus, and the RNA 2 species are thenefore not ínterchangeabl-e (Gould et

â1 ., 1981) desplte greater than 90% sequence homoJ-ogy. It is therefore interesting to note that the main sequence differences between VTMoV and SNMV RNA 2 1ie clustered opposite the GAUUUU sequence and that the remainder of the left-hand side of the native structure is al-most completely conserved ( Chapter 5 ) . Thus, these obsenvations are consistent with both the functionaÌ simitarity of the SCMoV, VTMoV and SNMV virusoids and the importance of the left-hand sides of these mol-ecul-es in replication. It is tempting to speculate that as lhe SCMoV 82 virusoids share common structures with VTMoV and SNMV

RNA 2, they may afso share common biological properties.

Th at is, SCMoV RNA 2 and/or RNA 2t may be required for viraf infection in a manner simil-ar to that shown for the RNAs 2 of VTMoV and SNMV; ând the central conserved sequences of these mol-ecul-es may play some nole in fulfil-ling this requirement. However, this proposition must be viewed with some scepticj-sm; fon while LTSV RNA 2 (which has been shown to behave as a satellite RNA (Jones et âf., 1983) and is not required for viraÌ infection) does not share the central- conserved sequences of SCMoV, VTMoV and SNMV RNAs, another satel I i te RNA, that of tobacco ringspot virus (TobRV), does.

D. SateIlite RNA of TobRV TobRV belongs to the nepovirus group and consists of 28 nm isometric particles containing two slngle-strand RNAs of molecular weighL 2.7 x lO6 (RNA 1 ) and 1.3 x 10"^ (RNA 2) which comprise the entire genome of the virus. RNA 1 and RNA 2 each possess a 3r-polyA tnact (Mayo et â1., 1979a) and a 5'-covalently l-inked protein (Vpg) (Mayo et al-. , 1979b) . In 1969, a novel- RNA species hras found in cultures of TobRV that had pneviously been apparently fnee of it (Schneider, 1969l'. This RNA species, which was dependent on TobRV for B3 reptication (Schneider, 1971), shaned almost no nucleotide squences in common with the supponting RNA (Schneiden, 1977) and hlas cl-assified as a satellite RNA. Since that time, ãL leasL 24 distinct isolates of TobRV have produced a satellite RNA of unknown origin during l-aboratory propagation (Kiefen et ãI ., 1982). The satellite RNA consists of a singl-e linear species

( Olener et âf . , 1 97 4; Schneiden, 1 977 ) of approximately 350 residues (Sogo et al., 1974), and does not share the

3'-poJ-yadenyJ-ate and 5 | -l-inked protein that are characteristic of the TobRV genomic RNAs ( Kiefer et âf. , 1982) , but lnstead bear 5'-hydroxyl and 3r-phosphate groups ( Kiefer et â1. , 1982; G. Bruening, personaf communication ) . Kiefer et a1 . (1 982) have shown that TobRV also encapsidates a muÌtimenic series of larger than unit length TobRV satetlite RNA sequences and that doubfe-strand RNA fractions isolated from infected plants can be denatured to pnoduce simil-ar mul-timeric series of both satellite and compl-ementary RNA sequences. Schneider and Thompson (1 977 ) have shown that the doubl-e-strand RNAs purffied from infected tissue are infective only after denaturation and addition of TobRV, and Sogo and Schnieder (1982) demonstrated that while the double-strand RNA preparaLions conbained predominantly l-inear mol-ecuIes, circular and I racket' shaped molecul-es vJere also B4 detected. Thus Kiefer et a1. ( 1 982) have proposed a rolling circle type mechanism for the replication of the satelllte RNA of TobRV which would account for the production of circul-ar and longer than unit length RNA sequences of the sateltite RNA and its compl emen t .

E. Sequence homoJ-ogy between TobRV satellite RNA and virusoids

The compì-ete sequence of a satell-ite RNA associated with the budblight st,rain of TobRV has noh¡ been determined ( Bruening, unpubl-ished results ) and was kindly pnovided by Dr. George Bruening. The sequence which consists of 357 nesidues is shown in Figure 6-7 with the pnedicted secondary structure of the molecul-e (Bruening, unpublished results). The pnedicted native mol-ecul-e possesses an overal-1 rod-11ke structure with four prominent stem-1oop structures and a slngle-strand 5t-proximal region. AIso indicated in Figune 6-7 is the remankabl-e extent of sequence homology between TobRV satellite RNA and the virusoid RNAs of VTMoV, SNMV and SCMoV. The homol-ogous sequences correspond to the central- conserved regions of the virusoids, and are al-so positioned in the rod-Iike centre of the TobRV mol-ecuÌe. Furthermore, the GAAAC sequence which is comnon to the central- conserved regions of viroids and virusoids

(except that of LTSV) is also present 1n the TobRV Figure 6-7 Sequence and p noposed secondary structure of TobRV satel-lite RNA. This data vüas kindly provided by Dr. George Bruening

( unpubl-ished results ) . Sequences which are conserved between TobRV satellite, SCMoV RNAs 2 and 2' , SNMV and VTMoV RNA 2 are shown indicated in green. TobRV satelllte RNA

I ,. tsic-a-l U U u U a u'c G G AÂ i. : ü--t'o U 50'C 6'cu G ^ U G'c c.c ç. c'c G.C ¡51 U U.A G U l0 l0l c C G'C ¡20 G.C ¡ u 20 u G.C UU aG I u9 o! c G.c U G \ C CU OGçC GCUAC U u\ G6 ACGU ACUAGlJ ' c 5 Ga G C G U :Y9 (, q:9191c C'CAUACCCUGUe cc u i cuca ^ I .CG CGCG 6uGu cuuc ccÂuG c,^G c UC uuu 6u cu C AGU UGCG UCAUCG U G GC cccGci s auo0gGca6cGCC GAUUÀC a' a G Ua CÈ t u ^GGC c clc IJ U cc a G c u c I a0 Â.rJ I cc U I U U ç ^c AA U CU ¡c0 G.C ¡50 u 770 I a c c I u ta0 ¡a G.C c ^ u t, c E-tto GA u u u .l u ?00-u 'a G^ c c'u G G u'a Ul G.U AC u'^ c.c 2t0 c ç AA AC B5 satellite RNA, b¡hile tha GAUUUU sequence found in all virusoids, incl-uding that of LTSV, iS absent ' The presence of sequences and structures whlch are shared by the TobRV satellite RNA and VTMoV,

SNMV and SCMoV virusoids suggests the nesidence of common functions and/or signals wlthin lhese RNAs. These RNAs replicate with viruses of quite different properties ( nepovirus group versus vTMoV/sobemovirus group), and this may indicate that the possible conserved functions and/or signals shared by the RNAs are invol-ved in interaction with host cel-l components rather than components of the different viruses. Klefer et a1 . (1982) concl-uded from evidence outlined above that TobRV satellite RNA may replicate via circul-ar and rnultimeric RNA intermediates thnough a roll-ing-circle type mechanism similar to that proposed for viroids (Branch e! âf., 1981; Owens and Diener, 1982; Kiefer et âf., 1982; Bruening et al., 1982) . The work presented in the final chapter is the nesult of preliminary attempts to determine whether virusoids also replicate via a rol_l_ing-circle type mechanism, with the attendant possibility of involvement of the consenved sequences. CHAPTER 7 vrRorDS, VIRUSOIDS AND SATELLITES B6

]NTRODUCTION

Larger than unit-length complementary ( - ) RNA

intermediates have been detected in PSTV- and CEV-

infected plant tissues ( Branch et âf. , 1 981 ; Rohde and il Sanger,19B1; Owens and Dienerr l9B2), and appean to exist malnly 1n extensively double-stranded RNA. 0wens

and Cress ( 1 980 ) and Bnanch et al. ( 1 981 ) have shown that RNase trealment of doubl-e-strand RNA intermedlates from PSTV infected pì-ants results in the production of complemenLary PSTV RNA of unit length or slightly tanger whil-e 0wens and Diener (1982) demonstrated that, denaturation of the double-strand RNAs released monomeric PSTV strands that had been complexed with multimeric compl-ementary RNAs. In addition, multimeric series of both ASBV and its complement have been found

in viroid-infected tissue (Bruening et â1., 1982), and dirneri-c RNAs have been shown f or CCCV. Various workers have postul-ated rolling-circl-e. type mechanisms f or the repl-ication of viroids ( Branch et âf. , 1 981 ; Owens and Dlener, 1982; Brueni-ng et âl ., 1982). As outlined in the previous chapter, the replication of the linear satellite RNA of TobRV shares features common to viroid replication. Compl-ementary RNA intermediates have been detected in high mol_ecular weight double-strand RNAs, âhd RNase treatment of these duplex RNA intermediates reduced the molecules to a size B7 slightj-y but significantly larger than unit-length ( Sogo and Schneider, 1982). Double-strand RNAs vüere shown to produce detectable satellite activlty only if denatured

( Schneider and Thompson, 1 977 ) , and to consist of multimeric series of apparently concatenate forms of RNAs of both pol-aritles ( Kiefer et â1., 1982). Thus Kief er et al- . ( 1982l, have also postul-ated a rol-ling-circfe type mechanism for the replication of TobRV satellite RNA. It seemed feasible that virusoids may al-so replicate via a rol-11n9-circle type mechanj-sm and the remainder of this chapter 1s devoted to descniption of the experimental- support for thls notion and to its possible ramlfications.

ME THODS

A. Isolation of RNA VTMoV and SNMV $rere kindly provided by Drs.

R.I.B. Francki, J.Vt. Randles and A. R. GouId. RNAs v\iene

extracted from purified VITUS (GouJ-d, 1981 ) , and from vi rus infected Nicotiana clevelandii ( Randles et â1.,

1 981 ) using phenol-SDS extractions as previously described.

B. Blot hybridization Nucleic acid samples hlere denatured by treatment with 1M gl-yoxal and 50% (v/v) dimethyl BB sufphoxide ( McMaster and Carmichael, 1977 ) and electrophoresed on 2.0% agarose slab gels (l5xl 4*.0.15' cm) in 1OmM sodium phosphate pH 6.5 aL 30 mA. Nucleic acids were transferred to nitrocel-l-ulose by blotting and baked in vacuo at B0oC (Thomas, 1 980). Nltnoceflul-ose sheets h¡ere prehybridized, hybridized and washed essentialJ-y as described by Thomas (1980 ). Complementary 32p-o¡la hybridization probe r^ras prepared using recombinant M13 ss DNA, containing sequences corresponding to SNMV RNA 2 residues 13'1 to 216 (see Chapter 5), essentiatJ-y as described by Bruening et a1. (1982).

RESU LTS

A. Analysis of VTMoV and SNMV RNA 2 se uences resent in vinus and infected tissues Using recombinant M13 ss DNA containing 32P-cDNA SNMV RNA 2 sequences, bras syntheslsed cl-oned ^ and isolated as a probe specific for VTMoV and SNMV RNA

¿ Fi gure 7-1 shows the pattern obtained when bhis 32 P-probe hras used to detect VTMoV and SNMV RNA 2 sequences present in RNAs extracted from virions and infected plants. It can be seen that for both viruses, mul-timeric series of (+) RNAs are found both encapsidated and in tlssue extracts. In addition, the dimeric forms of VTMoV and SNMV are detectabl-e by Figure 7 -1 Multimeric RNAs containing VTMoV and

SNMV RNA 2 sequences. Nucleic acids r¡rere extracted from either purified virus (V) or virus infected plant tissue (E) for both VTMoV and SNMV. The nucleic acids hrere glyoxaJ- treated, electrophoresed on a 2% agarose gel and subsequenbly transferred to nitrocellulose. VTMoV and SNMV RNA 2 sequences were detected using_ a ')a "P- cDNA p robe prepared f rom a cloned SNMV RNA 2 sequence as described in the text. 3r- 32 P-labelled Hpa II cut M13 mp7 RF vJas contransferred to provide si-ze markers. PLUS OLIGOMERS OF VIRUSOIDS

tr. - 1596

{;^-lf - 819

2% Agarose gel - 454 a - 357 Glyoxal RNA s

VE V E M VTMOV SNMV B9 toluidine blue staining of viri-on RNAs fractionated bY denaturing polyacrYl-amlde gel electrophoresis ( results not shown).

DISCUSS]ON

A. Multimers of VTMoV and SNMV RNA 2 Fnom btot hybridization experiments, such as that shown in Figune 7-1, it is apparent that multlmeric series of RNAs containing VTMoV and SNMV RNA 2 sequences are found both packaged in virions and in infected plant tissues, as j-s the case f or TobRV satel-lite RNA. On this evidence, it seems IikeIy that the vinusoids of

VTMoV and SNMV replicate via a roll-ing-circfe type mechanj-sm similar to that- proposed for TobRV satel-lite RNA (Kiefer et â1., 1982) atthough the existence and propertles of mul-timeric complementary RNAs and ds RNAs have yet to be determined. RoIling-circl-e mechanisms require a circufar

templ-ate to al-l-ow tnanscription of multimeric RNA intermediates. Therefore, such a model woul-d require that aL some stage during replication, the linear TobRV satellite RNA be ligated to produce a circulan template mol-ecul-e. As one of the f inal steps in replication, unit-length finear virusoid or satellite RNA must be produced by either specific transcription or cJ-eavage of

mul-timerlc RNAs . Th e 5' - hydroxyl and 3 t -phosphate 90 groups present on TobRV sateltite RNAs (Kiefer et âf., 1982; G. Bruening, personal communication) suggest that these molecufes are produced by specÍfic cleavage rathen than being primary transcripts. fn the case of virusoids, the unit-l-ength l-inear mof ecuf es must also be ligated to produce the final- circulan product. Thus, it is feasible that the TobRV satellite RNAs are simply defective 1n l-igation and correspond to l-inear RNA intermediates in virusoid replication which, in contrast, a?e capable of cj-rcuf arizatlon.

B. A possible site for RNA pnocessing The 5t terrninus of TobRV satellite RNA is adjacent in the molecule to the centnal- conserved sequences shared with the vinusoids of VTMoV, SNMV and SCMoV. When the sequences of these molecufes are aJ-igned as in Figure 7-2, extensive sequence homology

be tween TobRV satellite and VTMoV and SNMV RNA becomes apparent. Homologous sequences extend from the central- conserved regions to residues corresponding to the 5l

terminus of TobRV satel-lite RNA (VTMoV and SNMV RNAs 2 residues 49) and include several- residues correspondÍng to the 3r terminus. Therefore, 1t is proposed that the TobRV satetlite and virusoid RNAs are pnoduced by cleavage of multimeric RNA pnecursors at sites corresponding to between residues 357 and 1 for TobRV Figure 7-Z A possible site for RNA processing.

A) The proposed structures of TobRV satellite RNA, SNMV and VTMoV RNAs 2 are shown with the nucl-eotide sequences conserved between these 3 RNAs indicated in colour. These sequences consist of those shown in Figunes 6-6 and 6-T , and include addi tional sequences not shared by SCMoV RNAs 2 and 2t .

B) Comparison of the 3r and 5t proximal-, and central conserved reglons of TobRV with the cornesponding regions in SNMV and VTMoV RNA 2 and SCMoV RNAs 2 and 2t . Regions of the RNAs are shown in linear form, and proposed sites for RNA'processing of SNMV, VTMoV and SCMoV virusoids. Conserved nesidues are lndicated in col-our. A

I To bRV ¡ lio {-r AAA U c u c u.c i l!-t'o G'cu t.Ga^ 1o - c.c c'c'o ^ G.C ¡ tì c'c G u'a ¡! u :¡0 G.C ¡ u C G'C r l3! c c u G.c G c'c a^ U u UU Gv ccccoGc c.Gûrcclul¡ û \ c¡ .cG ^Gl ..1 99 içgY oC GCGU 99Y , î,cc G OCUAC.CUC .Gûuac 9Y:t î:Y1fY.+ç o CGCGCC u AucGcccAGc6cç u u u c u c u c AE U ucc G UCAUCC U GC.CG cccc Gue v cuuc cgûo Gô ç u .U cla CA lc c I cu U. G c cc UA CL c .c cc cu ¡le ¡l' u ¡¡0 u G.C Û IJ A .c ¡14 c'c ^^ G u'a tlc l.g UtJ c.G u cu'c GA u U u uu. . G^ c ¡ ôc-u G G c u U l U l A c u u A ¡ c c c c A A A c SNMV

I 20. 40 ô0 100 r20 r40 160 rô0 [u, I co' I I cl^ u c u A U CC g gA c \ohj \, e^v g I ^c ^C "^ !c^c uo uuoouucoccuqcuc ouo ¡!c c t/0 t¡uc9u g Cq^CCu.CO^G.CUg êu.ôUO sAO00 oocuc c{}ucr. coc ocu 0 u^g¡. rse^se.cg) u^t c¡ c ^c 0 co^coo oo^c cuffiõì olõiìcoo qqc q30 0 cpþ ceu@ rcevcr of| euc o9 c {þ@$,crc . cac c.^c^cc.ueo-116} AOA C ooo{þ ocugoA 0uuc 9û c cuccc uc0ao op\ ouo ^cue c UC o \cl^ u c u/" @r c | ô ' ^.^ | 'û lü I "rrü¡ o l' 3tt .1:$,"- 340 320 280 260 ^c ?40 220 200 VTMoV

lô0 r00 2ou 40 0 / 100 120 140 0 U c lu U c g u c c I tr0 u g0 c cA uc I lr c I u cc u uc 0^ t_ ^ U ug Ê u C c o ouô urc cA ueAc ccuocc o0 c uo0uu sccv uccou0 u 0 uucg oc 0Accu,coAc,cu0 cu^0 UO oôoo oocuG c c.uclgl. uourcu.{J c c :: ^c gOACgO. g OUO AgC C^ uc rcruce ruo cu ¡cue 0 uC c cFõî¡Ic ooo.[þocrc 0OC¡CÁ€9.^€ACC uACg.A0 C .cc 909^ cg cuogÀ ouuq sû o@ cuccc !CoAO'OoC { ' ^c ^ccct o co A u ,1. ^^r c UU !lItr a90 Io lq" â l I c 0 I I 2@ 365 360 'g' glo 320 300 2øO 260 240 E 220

100 o c u cc uucu @ orccu gOO^ slCO ' CUggA 280 a

TcbRV scielTite

UACC UU 66 U6 ¡.F

ViMcV ond SIiMV RNA 2

¡Ê¡. 2 Gu6ulr 6 A A( AA + ¡O i*rÞr"¡y'a

SiMcV RlilAs 2 c¡¡d 2'

S;-Prt UÀ L cAc6 u A t u'yåglj",, G u A t A CUA st llo ^

I 91 satellite RNA and between nesidues 48 and 49 for VTMoV and SNMV RNAs 2. Unit length linean VTMoV and SNMV RNAs 2 woul-d then be l-igated to produce the mature circular forms of the RNAs. Implicit in this proposal j-s the assumption that the conserved sequences surrounding these putative sites for RNA pnocessing are in some I^Iay functional, perhaps in detenmining the specific sites of cleavage. The conservation of these sequences in RNAs from different viruses may suggest their interaction with host, rather than viral, components (TobRV, VTMoV and SNMV share common host plants, such as Nicotiana cf evel-andii ). Interestingly, during sequence determinatj-on of VTMoV and SNMV RNAs 2, essentiaJ-J-y compl-ete termlnation of reverse transcription I4Ias observed at positions corresponding to residues 49 of the RNA templates. Thls I¡Ias seen whether intact RNA 2 (see Figure 2-B) or puri-fied finear RNA fragments u¡ere used as templates, and hras pnesumabl-y due to the presence of sequences and/or secondary structunes capable of causlng reverse transcniptase to chain terminate. For example, an B-base-pair stem 3-base-pair loop structure can be formed in VTMoV and SNMV RNA 2 aL residues 40 to 58, however the same structure cannot be formed at the corresponding sequences of TobRV satellite RNA (or SCMoV). It is unknown whethen the precì-se coincidence 92 of sites for termination of neverse transcrlption with predicted pnocessing sites in vTMov and sNMV RNA 2 is a product of chance of, perhaps indirectl-y, of function. In contrast, SCMoV RNA 2 and RNA 2t share Iittle sequence homology with the other RNAs (Flgure 7-2\ outslde the central- conserved region, with the exception of several residues approximately cornesponding in location to the 5t terminus of TobRV satellj_te RNA. Based so1ely on thls limited sequence and structural- homotogy, Possíb1e processing sites fon

SCMoV RNA 2 and RNA 2t are between residues 62 and 63 in each mol-ecuf e. The l-essen extent of sequence homology between the SCMoV virusolds and the other small viral- RNAs (Figure 7-2) may' be rel-ated to the limited and excl-usive host nange of scMov, which is not known to share host pJ-ant species with TobRV, VTMoV or SNMV

( Franckl et âf . , 1 983 ) . Thus functional nucleotide sequences might vary to acconodate the different requirements of host components in different species.

C. Viroid, virusoid and satel-Iite RNAs I t noI^I appears that there exists in pl-ants a range of replicating RNA species which overal-1 share many common features. Thus virolds, virusoids and TobRV satellite RNA a1l consist of ss ú*O speci-es of between L- ß 240 and 4OO residues which, except that of TobRV o? satellite, are rod-like base-paired circufar molecules ' These RNA species do not appear to code for functional polypeplide translation products, but repÌicäte lhrough multimeric RNA intermedj-ates whj-ch are probably transcribed by a roll-ing-circle type mechanism. The unit-Iength pnogeny species must be produced by specific transcription or processing events and thenr eXcept in the case of TobRV satel-Iite, J-igated' Given these simitarities, the RNAs f all- j-nto one of two cl-asses. The first contains viroids which a?e chanactenisticalty naked and capabl-e of independent of encapsidated RNAs replication. The second conslsts ' Iike those of VTMoV and SNMV which appear to contribute

some f unction to a vinal- Senome I otr l-ike those of TobRV and LTSV which are satel-1ites. The members of each of these groups share at l-east some conserved sequences with others of the same group (see chapter 4 and 6), and overalf share remarkable conservation of regions central to their native stnuctures ( with the notabÌe exception of LTSV RNA 2) . Furthermone the pentanucleotide sequence GAAAC is present on all sequenced viroids, virusoids (except LTSV RNA 2) and TobRV satellite RNA, and is l-ocated within the central conserved regions of, these molecul-es.

A s ingle questlon looms f nom th j-s tangl-e of observatons. Do viroids shane common functions, and 94 perhaps origins, with the second group of encapsidated RNAs? The possible invol-vement of conserved vinusoid and satellite sequences in interaction with plant host components v,IaS inferred from data presented j-n Chapters 6 and T, and viroids appear to nely entireJ-y on host components for reptication. It therefore seems reasonable to suggest that these biologicalJ-y disparaLe RNA species may shane some common mechanisms in neptication whlch involve functionally similar, if not identical, host components. The two groups of RNA specles may be derived from a common ancestral- species or alternatively be products of convergent evol-ution. These suggestions are of course based on inference rather than direct evidence as the appnoach taken in this work aIlows onJ-y a glimpse of the functions and origins of these mol-ecules as reflected in their comparative structures. Confirmation on denial- of and ultimately, ansvlering of Lhe these possibilities ' three questions originally posed in Chapter'l wlll rely upon studies of the host and viral compon"ìt" involved in the repl-ication of these RNAs, Tãther than the RNAs themselves. RE FEREN CE S

( FEBS Lett Barrell , B. G. and Sanger, F. 1 969) ' 3 ,

27 5-27 8 . Bigorni-a, A. E. (1977) Philipp' J' coconut stud' 2, 5-33. Bi rnboim, H. C. and Do1Y, J. (1979) Nuc. Acids Res. 9,

6527 -6537 .

Boccardo, G., Beaverr-R. G.r RandJ-es, J. W and

ImperiaJ-, J. S. (1981)Pngtopath. 71 ' 1104-1107. Boege, F. , Rohde , V,l. and Sánger, H. L. (1982) niosci ReP' 2, 185-194'

Branch, A. D., Robertson, H. D. and Dickson, E. (1981 ) Proc. Natl-. Acad. 5c j. USA 78, 6381-6385.

Brederode, F. TH., Koper-Zwarthoff, E . C. and Bol-, J F

(1980) Nuc. Ac ids Res. B 221 3-?223 .

Br ierly , P . and Smith, F F (1949 ) Phgtopath 39, 501 .

Br ierlV , P . and Smith, F F (1951 ) Pf ant D is Reptr. 35 , 524-526.

Bruening, G., Gouì-d, A R Murphy, P. J. and Symons, R

H. (1982) FEBS Lett - 148, 71-78

Camacho, H. A. and Sanger, H (1982a\ Arch Virol-. in press.

Camacho, H. A. and Sange n , H. (1982b) Arch. Viro7. 1y¡ press. Conjero, V. and Semancik, J. S. (1977) V i rologg 77 , 221 -232. Conjero, V., Picazo, I. and Segado, P. (1979) viroToss 97 , 454-456.

Davies, J. I,'l ., Kaesberg, P. and Diener, r. 0. (1974) Virologg 61 , 282-286. De Wachter, R. and Fiers, Vü. (1972) AnaL. Biochen. 49, 184-197.

Dienen, T O. (1971a) virolosg 43, 75-89.

Diener, T O. (1971b) virofoss 45, 41 1 -428 .

Diener, T O. (1972) Virofoss 50, 606-609.

Diener, T 0. (1979]' f'Vi-roids and Viroid Diseasesrr Wiley In terscience , New York. Diener, T. O. (1981) Proc. Natl. Acad.Sci. UsA 78,

501 4-501 5 .

Dlener, T. 0. and Lawson, R. H. (1973) virol-ogg 51 , 94-100. Dienen, T. 0. and Raymer, Vü. B. (1967) science 'l 58, 378-381.

Diener, T. 0. and Raymer, V,I. B. ( 1 969 ) V irof ogg 37 ,

351 -366 . Diener, T. 0. and Smith, D. R. (1975) viroloss 63

421 -427 . Diener, T. 0., Schneider, I. R. and Smith, D. R. (1974)

Virol-osv 57, 577-581 . Dimock, A. W. (1947) N.Y. st. FfowerGrower'sBuLf. 26, 2. Domdey, H., Jank, P., Sanger, H. L and Gross, H' J' (1978) Nuc. Acids Res' 5, 1221 -1236' Domingo, E., Sabo, D., Taniguchi, T and V'Jelssman' C' (1978) ceft 13, 735-744'

Donis-Keller, H. ( 1 9BO ) Nuc. Ac ids Res. B, 31 33 -31 42. Donis-KeIIer, H., Maxam, A. M. and Gilberl, V'¡' (1977)

Nuc. Ac ids Res. 4 , 2527 -2538 ' Donis-Kel-l-er, H. , BrowninS, K. S. and Clark, J ' M ' (1981) ViroTosg 110' 43-54' N. Donis-KelIer', Efstratiadis, A. , Vournakis, J. ' Chaconas, G., Dougall, D. K. and Kafatos, F' C. (1977) Nuc. Acids Res. 4, 4165-4174' ( Fl.ores, R. , Chroboczek, J. and Semancik, J. S. 1 978 )

Phgs io7. Pfant Pathol. 1 3 , 193-201 ' Davies, Fnancki, R. I. B., Rand1es, J. !'¡. r Hatta, T. ' , C., Chu, P. !ü. G. and Mcl-ean, G. D. (1983) in press.

Frisby, D . (1977) Nuc. Acid Res. t 2975-2995. (1979) Go odman , H. M. and MacDonal-d; R. J tniethods EnzgnoT. 68, 499-560.

Gordon, K . H. J. and Symons, R. H. ( 1 983 ) Nuc. Ac ids Res. 11, 947-960.

Gou1d, A R. (1981) virofoss 108, 1 23-1 33

GouId, A R. and Hatta, T. (1981 ) Virologg 109,

137 -1 47 .

GouId, A. R. and SYmons, R H. (1982) nur. J. Biochen ' 126, 217-226. Palukaitis, P Symons, R' H' and Mossop' Goul-d, A. R. ' ', D. W. (1978) virofoss 84, 443-455. GouJ-d, A. R., Francki, R. I' B' and Randles, J' W' (1981) ViroTosg 110, 420'426' Grilt, L. K. and Semancik, J. s. (1978 ) Proc. Natl-. Acad. 5cj. USA 75, 896-900' Gross, H. J. and Riesner, D' (1980) Angew' Chem' Int' Ed. Eng7. 19, 231-243. --Gross, H. J., Domdey, H., Sånger, H' L' (1977) Nuc'

Ac ids Res . 4 , 2021 -201 I - Gross, H. J., Domdey, H.r Lossow, C', Jank, P', Raba' M., AlbertV, H. and Sånger, H' L' (1978)

N ature 27 3 , 203- 208 ' Raba, M', Alberty' Gross, H. J., Krupp, G. r Domdey, H' ' Lossow, C. H., Ramm, K' and Sånger H' L' H. ' (jgBZ) Eur. J. Biochen. 121 , 249-257. Hadidi, A., Cress, D. E. and Diener, T' O' (1981)Proc' Natl- . Acad: 5c j. USA 78, 6932-6935.

Haseloff, J. and Symons, R. H. ( 1 981 ) Nuc. Ac ids Res. 9

27 41 -27 52. Haseloff , J. and SYmons' R. H. (1 982 ) ruuc. Ac ids Res.

1 0, 3681 -3691. ( Haseloff, J., Mohamed, N A. and Symons, R. H. 1 982 ) Nature 299 , 31 6-322.

Henco, K., Riesner, D and SJnger, H. L (1977) Nuc Ac ids Res. 4, 177-194.

HolIand, J. J., Gnabau, E. A. , Jones, C L. and Semler, B L. (1979) cett 1 6 , 495-504. ( Holl-ings, M and Stone, 0. M 1 973 ) Ann. Appl Bio1.

7 4 , 333-348.

R Hollings, M , Stone, 0. M. , Barton, R. J. and Greber, q (1979) Rep. Gfasshouse Crops Res ' rnst.

1978. 150-151 , Imperlal, J. S.' Rodriguez, J. B' and Randles, J' W' (1981) J. Gen. Virol. 56,77-85' Jones, A. T., Mayo, M. A. and Duncan, G' H' (1983) J

Gen. V itof . 64 , 1n Press '

Ke11er, J. R. ( 1 953 ) Corneff univ. Agtic. Exp. 5tan. l,Iemoir 324,40. D. Schneider, I' R' and Klefen, M. C., DauberL, S. ' BrueninS, G. (1982) vitofogg 121 ,262-273' KJ-ump, H. , Riesner, D. and Sänger, H. L ' (1978 ) Ùuc. Ac ids Res. 5 ' 1 581 -1587 ' H and Gross, H Konarska, M., Filipowicz, I,{. ' Domdey, J. (1981) Eur. J. Biochem. 114 , 221 -227 .

Kozak, M. (1978) eet-t- 15, 1109-1123.

Kozak, M. (19'(9) Nature 2BO, B2-85.

Kramer, F R. and Milts, D. R. (1978) Proc. Natl. Acad. 5ci. UsA 75, 5334-5338.

Krupp, G and Gross, H. J. (1979)¡vuc Ac ids Res . 6, 3481-3490. J. Henco, K. Riesner' D. and Sánger, H. L Langowski, ' ' (197 8 ) ¡vuc . Ac ids .Res. 5 , 1589-1610 '

J' E Lockand, R. E., Ai-znen-Deweerd, B. ' Heckman, MacGee, J., Tabor, M' I/t and RaJ BhandâFY r U

L. ( 1 9?B ) Nuc. Ac ids Res 5, 37-55.

Lomedico, P. T; and McAndrew' S J ( 1 98 2) Nature 299 ,

221 -226 - Maniatis, T. and Efstrâtiadis, A. (1980)¡lethods EnzgnoT. 65, 299-305 Maxam, A. M. and GiIbert, tü. (1980) Methods Enzgnol. 65, 499-560. (1979a) J Mayo, M 4., Barkerr-JI . and Harrison, B. D. 603-610. Gen,. Virol. 43 ' (1979b) Mayo, M A., Banker, H. and Harrison, B. D. J Gen. virol- ' 43, 735-740' Messing, J., Crea, R. and Seeburg, P. H. (1981 ) IVuc. Acids Res. 9, 309-321. Miyazaki, M. (1974) J. Biochem. 75, 1407-1410'

Mohamed, N. 4., Haseloff, J -, ImPerial, J. S. and

Symons, R. H. (1982)t.'cen. vitof . 63, 181-188. Mulbach, H. P. and Sänger, H. L. (1979lr Nature 278,

1 B5-1 BB. Niblett, C. L,, Dickson' E.' Horst, R. K. and Romaine, C. P. (1980) Phstopath. 70,610-615. Ocfemia, c. p. (1937) Philipp. Asric. 26,338-340. 17 OJ- son, J C. (1949ll Bul-1 . ChrVsanthemum Soc. Amer. , 2-9.

Owens, R A. and Diener, T. O. (1982) Proc. Natl-' Acad' Scj. UsA 79, 113-117. ( Owens, R. 4., Smith. D. R' and Di ener, T. 0. 1 978 ) Virolosg 89, 3BB-394'

PaIukaltis, P. and SYmons, R' H (1980) J. Gen. Virol-. 46, 477-489, l¡ (1981 RacixwiLz, H. R-, Rohde, V¡' and Sanger, H. L. )

Nature 291 , 297-301 .

J . ü¡ . ( 1 975) pnstopath. 65 , 1 63-1 67 ' RandJ-es ' Randl-es, J. ü1. and Hatta, T. (1979) vitoTosg 96 47 -53. Randles J. lü. and Pafukaitis, P. (1979) J. Gen. Virol. ' 43 , 649-662. Randles, J. lirl., RiJ-l-o, E. P. and Diener, T. O (1976) v¡rot oss 74, 128-139. ( Randles, J. W., Boccardo, G. and ImPerial, J. s. 1 980 )

Phstoþath. 70, 1 B5-1 89.

C. HaLLa, T. Gou1d, A R. and Randles t J . Vü. , Davies, , ' Francki, R . I. B. (1981) ViroToss 10 ai

1',1 1-122. Raymer, l,ù. B. and Diener, T. 0. (1969) v¡rologg 37, 343-350.

D. Henco, K. Rokohl, U. , KIoLz, G Riesner, , ' Kteinschmidt, A. K., Gross, H. J , Domdey, H

H L. (1979) I4of . Biol. 133 and Sänger, J. '

85-115 .

Rohde, ll,l . and Sänger, H L. (1981) aiosci. Rep. 1, 327 -336. jol- Rubin, G . M. (1973) r. B . Chen . 248 , 3860-3875.

A (1977) Sange n , F., NickJ-en, S. and Coulson, R. Proc. Natl . Acad. Scj. UsA 7 4 , 5463-5467 ' Sanger, F. and Coul-son, A. R. (1978) FEBs Lett.87, 107-110. Sanger, F., Coul-sonr A. R.r Barrell, B. G', Smith, A' J' H. and Roe, B. A. (1980) J. Mol. Biof. 143,

161 - 1 78. Sånger, H. L. (197,2\ aav. Biosci. B, 103-116' Sánger, H. L., KIoLz, G., Riesner, D.' Gnoss, H' J' and Kleinschmidt, A. K. (1976) Ptoc. Natl-. Acad. 5cj. USA 73, 3852-3856. Schnelder, I. R. (1969) science 166, 16?7-1629. Schneider, I. R. (1 971 ) vito)ogg 45, 1OB-1?2, Sch,rr¡eider, I. R. (1977) Defective pl-ant viruses. In rrBel-tsvill-e Symposia in Agricul-tunett (.1- A. Romb€r8, J. D. Anderson and R. L. PoweJ-1, eds.) 1, 201-219.

Schneider, I. R. and Thompson, S. M. (1977 ) v iroTogt¿ 78, 453-462. Semancik, J. S. and !,leathers, L. G. (1968) virofogv 36, 326-328. Semanclk , J . S. and lrleathers, L. G. ( 197 0 ) p¡ gtopath. 60, 732-736. Semancik, J. S. and V'leathers, L. G. (1972a) v¡rologg 47, 456-466.

Semancik, J . S. and fieathers, L G. ( 197 2b) IVa ture New BioTogg 237 ¡ 242-244.

Semancik, J. S. Morris, T. J., lrleathers, L. G Rodorf, B. F. and Keanns, D. R. (1975) Virolosg 63, 160-167. (1977) Semancik, J. S., Conjero' V and Genhart, J.

Virofogg 80, 218-221 .

J 37 31 Si ppel , A. E. (1 97 3 ) Eu¡. . Biochen. , -40. 117 Sogo, J M. and Schneider, I. R. (1982) v irol-ogg ¡ 401 - 415. (1974) Sogo, J. M Schneider, I. R. and KoIler, Th. ViroTogg 57 , 459-466. M. Imaizumi, T. and Schrerren, K Spohr, G. , Mirault, ' (1976) Eur. J. Biochen. 62, 3'l 3-322.

Symons, R. H. (1971) Nuc. Acids Res. 4, 4347-4355. Symons, R. H. (1978).Aust. J. Biol. Scj. 31 , 25-37.

Symons, R. H . ( 197 9 ) /Vuc. Ac jds Res. 7 , 825- 837 . Symons, R. H. (1981) lVuc. Acids Res. 9, 6527-6537. Takahashi, T. and Diener, T. O. (1975)viroTosU 64, 106-114. Taylor, J. M. r Illmensee, R. and Summers, J. (1976) Biochem. BioPhgs. Acta. 442 , 324-330. Tien-Po, Davj-es, C., Hatta, T. and Francki, R. I. B

( 1 981 ) FEBS Lett. 132 , 353-356.

Tinoco, I. JY., Uhlenbeck, o t, , and Levi-ne, M D

( 1 97 1 ) N atu re 230 , 362'-367 . Tinoco, I. JF., Borer, P- N., DengJ-en, B., Levi-ne, M D., Uhlenbeck' O. C., Crothers, D' M' and

( 1 Nature 246 40-41 GralÌa, J. 973 ) ' ' Visvader, J. E., GouJ-d, A. R.' Bruening, G' E' and Symons, R H. (1982) FEBs Lett. 137, 2BB-292. ( V,¡i1d, U Ramm, K Sanger, H. L. and Riesner, D. 1 980 ) Eur. J. Bjochen. 103, 227-235.

J. N. and Maxam, A. M. ( 1 978 ) l'l,lurst, R. M. , Vournakis ' BiochemistrU 11 , 4493-4499 ZaLLl-in, M., Niblett' C. L., Dickson' E. and Goldberg, R. B. (1980) YrroToss 104, 1-9. Zelazny, B. (1979) ecta. Phgtopath. Acad. ^scj. Hung. 14, 115-126.

Zelazny , B. and Niven, B. S. ( 1 980) plant Disease 64 , 841 -842. M. Robertson, H. D. and Dickson, ZeLcer, 4., ZaiLlin, ' E. (1982\ ¡. Gen. virol. 59', 139-148'

Zimmern, D. and Kaesberg, P. ( 1 978 ) Proc. Natl . Acad. Scj. usA 75,4257-4261'