Plant pathogenic RNAs and RNA catalysisRobert H. Symons
Department of Plant Science, Waite Institute, University of Adelaide, Glen Osmond, SA 5064, Australia
Received March 18, 1997;Revised and Accepted May 28, 1997
ABSTRACT
The rolling circle replication of small circular plant pathogenic RNAs requires a processing step to convert multimeric intermediates to monomers which are then circularized. Eleven such RNAs are known so far, two are viroids, one is viroid-like and the remainder are satellite RNAs dependent on a helper virus for replication. The processing step is RNA-catalysed in all cases, at least in vitro. All plus forms of these RNAs self-cleave via the hammerhead structure whereas only eight of the minus RNAs self-cleave, five via the hammerhead structure and three via the hairpin structure. There are about 20 other viroids where the processing mechanim has yet to be determined but they are likely candidates for a new type of self-cleavage reaction which is predicted to be conserved in all these viroids. Hepatitis delta RNA is the only circular pathogenic RNA known to self-cleave in the animal kingdom. It is feasible that more single-stranded circular pathogenic RNAs are waiting to be discovered and these could be prospective for new types of self-cleavage reactions.
Small single-stranded circular plant pathogenic RNAs have been a rich source of self-cleaving RNAs and there is considerable promise of more to be found. All such self-cleaving RNAs identified so far range in size from 246 to >400 nucleotides (nt) and are replicated by a host RNA polymerase via a rolling circle mechanism (1 ). Two of these are viroids which are able to replicate in the plant independently of a helper virus. The remainder are satellite RNAs, so called because they require specific input from a helper RNA virus for their replication.
The aim of this article is to place as much emphasis on where the future lies as on a concise review of a number of aspects of RNA catalysis in small plant pathogenic RNAs. I refer the reader to review type articles which cover both broad and specific aspects relevant to RNA catalysis as considered here (2 -11 ).
Self-cleavage and rolling circle replication
Self-cleavage is a term used for the reaction whereby single-stranded RNAs are cleaved at a specific site in the absence of a cofactor apart from a divalent cation. In the reactions identified so far, cleavage occurs by nucleophilic attack by the 2'-hydroxyl at the cleavage site on the internucleotide phosphate such that the cleavage fragments contain a 2', 3'-cyclic phosphate and a 5'-hydroxyl (Fig. 1 ). The reaction is therefore distinguished from the self-splicing of Group I introns which requires a guanosine cofactor which is covalently incorporated at the 5'-end of the intron (Fig. 2 A). The 3'-hydroxyl of the 5'-exon then attacks the phosphate at the 5'-end of the 3'-exon to give the spliced exons and the release of the excised intron. In the case of the self-splicing of Group II introns, the initial step is the attack by the 2'-hydroxyl of an A residue internal to the intron on the phosphate at the exon-intron boundary (Fig. 2 B) followed by a similar second step to that of Group I splicing.
All evidence indicates that the circular pathogenic RNAs are replicated by a rolling circle mechanism (1 ,8 ) in vivo (Fig. 3 ). The replication cycle involves the copying of the dominant circular plus (+) strand by a host or viral-coded RNA polymerase to give a longer than unit length minus (-) strand. For most of the self-cleaving circular RNA pathogens so far identified, this long (-) strand can self-cleave in vitro to give monomeric products (Fig. 3 A). In vivo, these monomers are circularised and then copied to produce a linear (+) strand which self-cleaves to monomers which circularize to produce the dominant circular (+) progeny found in vivo. For those multimeric (-) RNAs which are not processed (Fig. 3 B), the RNA is copied to give a multimeric (+) strand which then undergoes cleavage to monomers which are then circularised.
Circular, self-cleaving plant pathogenic RNAs
The in vitro self-cleaving plant pathogenic RNAs identified so far all fall into three groups (Table 1 ); those where both the (+) and (-) RNAs self cleave via the hammerhead structure (Fig. 4 A), those where only the (+) RNA is cleaved via the hammerhead structure, and the remainder where the (+) RNA is cleaved by the hammerhead structure and the (-) RNA is cleaved via the hairpin structure (Fig. 4 B).
Of the 21 viroids identified so far (Fig. 5 ), only two have shown specific self-cleavage in vitro. Avocado sunblotch viroid (ASBV), a nuisance pathogen in the avocado industry, was the first viroid to show hammerhead self-cleavage in 1986 (12 ). This viroid at 246-251 nt, together with coconut cadang cadang viroid (CCCV) at 246 nt, are the two smallest viroids isolated so far; however the processing mechanism during replication of CCCV has yet to be determined. A most interesting difference between these two viroids as well as the 368-463 nt citrus exocortis viroid (CEV) and the 341-361 nt potato spindle tuber viroid (PSTV) is their intracellular localisation; ASBV is found on the chloroplast thylakoid membranes while CCCV, CEV and PSTV are localised in the nucleus of host plants (13 -15 ). The only other viroid isolated so far which has been shown to self-cleave is the 337-338 nt peach latent mosaic viroid (PLMV) and, like ASBV, both (+) and (-) RNAs self-cleave via the hammerhead structure (16 ; Fig. 4 A).
The hammerhead ribozyme structure is becoming well characterized
The small size of the hammerhead structure (Fig. 4 A), where fewer than 40-50 nt can form an active self-cleaving molecule, has led to an explosion of effort into its characterisation and manipulation both in vitro and in vivo. It was the discovery and characterization of the self-cleavage of ASBV and of the virusoid of lucerne transient streak virus (vLTSV) (12 ,21 -23 ), together with the predicted self-cleavage sites of two other virusoids (21 ) and the data from the self-cleavage of the (+) form of the satellite RNA of tobacco ringspot virus (sTRSV) (10 ,24 ), which defined the hammerhead structure and the naming of it (21 ).
The first in trans reactions with the hammerhead structure involved two short sequences derived from the sequence of the top and bottom strands of (-)ASBV (26 ) and also of a short and a long RNA transcript equivalent to the full length of the ASBV molecule (23 ). The core sequence of the hammerhead structure of (+)sTRSV was used by Haseloff and Gerlach (27 ) to develop their in vitro system for targeting the cleavage of foreign RNAs. The reader is referred to a number of reviews and papers which summarize much of this earlier work (5 -11 ,28 -31 ). Of considerable recent importance is the determination of the crystalline structure of the hammerhead, initially with modifications to prevent self-cleavage (32 -34 ), and more recently in an unmodified form (35 ).
The hairpin ribozyme structure
Although the hairpin self-cleavage reaction of (-) sTRSV was first identified in 1986 (36 ,37 ), the same year as the hammerhead reaction in (+)sTRSV and ASBV (12 ,24 ,36 ,37 ), it has received less attention than the hammerhead. However, there is a greatly increasing effort into the 50 nt hairpin structure (38 -42 ). So far only three satellite RNAs have been found where the (-) RNA self cleaves by the hairpin structure (Fig. 4 B) and the helper viruses all belong to the Nepovirus family (Table 1 ). In contrast to the hammerhead self-cleavage reaction, that of the hairpin ribozyme is readily reversible. As for the hammerhead structure, it lends itself readily to mutational analysis and reactions carried out both in cis and in trans (38 ,41 ,42 ) and such experiments together with phylogenetic comparisons have allowed the establishment of the secondary structure shown in Figure 4 B for (-)sTRSV. One can foresee the RNA crystallographers turning their attention to determining the three-dimensional structure of the active hairpin ribozyme and an increasing interest in exploring the potential for in trans cleavage of foreign RNAs both in vitro and in vivo (43 ).
Does the in vitro hammerhead self-cleavage reaction have its predicted role in vivo?
The two viroids and the four viroid-like satellite RNAs or virusoids associated with the four helper Sobemoviruses (Table 1 ) are isolated from infected plants in the circular form. Hence, the role of the self-cleavage reaction in the rolling circle replication of these RNAs is inferred from the absence or presence in infected plants of a high molecular weight minus strand. For example, in the case of vLTSV where both (+) and (-) strands self-cleave in vitro (21 -23 ), mostly monomeric (+) and (-) strands, together with some lower multimeric (+) forms, were found in infected plants and no high molecular weight (-) forms, evidence consistent with specific cleavage of the (-) strand in vivo (44 ). Likewise, the presence of high molecular weight (-) forms of vSNMV, vSCMoV and vVTMoV in infected plants and the absence of monomeric (-) forms (44 ,45 ) is consistent with the in vitro self-cleavage data (Table 1 ).
In order to provide more definitive evidence, three full length cDNA clones of vLTSV, mutated at sites that inactivate in vitro self-cleavage of the (-) RNA were inoculated as excised plasmid inserts, together with helper virus LTSV, on susceptible host plants (46 ). As predicted if hammerhead self-cleavage is involved in in vivo cleavage of (-) RNAs, high molecular weight (-)vLTSV RNAs were present in total RNA extracts of these mutant inoculated plants but not in wild-type inoculated plants. Surprisingly, the mutated virusoids also produced monomeric (-) RNAs in vivo; sequence analysis of these RNAs indicated reversions or pseudoreversions of the introduced mutations which most likely restored some in vivo self-cleavage activity in the (-) RNAs. In spite of the complications of these reversions, the results provide strong support for the proposed role of the hammerhead reaction in the processing in vivo of multimeric (-)vLTSV RNA.
The PSTV group of viroids-potential for a new ribozyme?
The PSTV (potato spindle tuber viroid) group of viroids contains about 20 different viroids (Fig. 5 ) and new ones are being described at regular intervals. Relevant to further discussion, a brief description of viroid classification is important. Viroids can be classified into two separate groups based on comparative sequence analysis and self-cleavage ability (5 ,48 -50 ). There are currently only two members in the ASBV group (Table 1 , Fig. 5 ) and both (+) and (-) RNAs self-cleave in vitro via the hammerhead structure. All other viroids identified so far fall into the PSTV group and none contain the sequence motifs required for hammerhead or hairpin self-cleavage nor have any been shown so far to be capable of a specific self-cleavage reaction in vitro (see further discussion below).
The domain model for the PSTV group of viroids was developed in 1985 (48 ) and has stood the test of time as more viroids were discovered and sequenced. Comparative pairwise sequence analysis of members of the PSTV subgroup of viroids (Fig. 5 ) indicated the presence of five domains (Fig. 6 ), the boundaries of which were defined by sharp changes in sequence homology, from high to low or vice versa. Different pairwise sequence comparisons were consistent in defining the exact position of the boundaries. The domains are: T1 and T2, the left and right terminal domains; P, the pathogenic domain; C, the central conserved domain; and V, the variable domain. By convention, residue numbering starts in the middle of the end loop of the T1 domain.
What is the UV-induced cross-linking between RNA strands telling us about local tertiary structure and RNA processing?
As mentioned above, UV-irradiation of PSTV cross-links G98 and U260 in the CCR of the native molecule (58 ) and in a region highly prospective for RNA processing during rolling circle replication. Two other RNAs also show similar UV-induced cross-linking. The (-) sTRSV RNA is cross-linked between G204 and U183 (Fig. 4 B) and close to the self-cleavage site in the hairpin structure (42 ,65 ). A similar situation also exists in the single-stranded, circular self-cleaving hepatitis delta RNA, a 1700 nt rod-like molecule where both the genomic and antigenomic RNAs show self-cleavage via a similar structure (66 ). This self-cleavage occurs at one end of the molecule which shows viroid-like features and near to the site where the genomic RNA is cross-linked by UV-irradiation (67 ).
The UV-induced cross-linking indicates a tertiary structure motif where juxtapositioning occurs of two bases from opposite strands. It is feasible that such a tertiary element may play a role in aiding the formation of an active self-cleavage structure in these molecules.
Future prospects
The single-stranded circular plant pathogenic RNAs and the more complex hepatitis delta RNA have provided us so far with most of the naturally occurring self-cleaving RNAs. As considered above, it is predicted that the PSTV group of viroids will provide another source of such RNAs. The carnation small viroid-like RNA (CarSV RNA) appears part of a more complex system (17 ,18 ) which is still to be resolved and it may provide some more surprises in RNA catalysis.
It is feasible that there are more single-stranded circular pathogenic RNAs waiting to be discovered and characterised in both the animal and plant kingdoms and these could be prospective for new types of self-cleavage reactions. The possibility should be considered that the transcription of circular single-stranded and double-stranded DNA may, in some cases, depend on self-cleavage reactions for processing of the transcripts. A good example here is the self-cleavage of a multimeric RNA transcript of the circular 881 bp Neurospora mitochondrial DNA plasmid to produce a linear 881 nt RNA which is then circularised (68 ). The self-cleaving domains in this RNA are different to those found for the hammerhead, hairpin and hepatitis delta RNA ribozymes (69 ).
And, finally, the transcript of satellite 2, a 300-350 bp repetitive DNA that is highly conserved in salamanders, contains a hammerhead self-cleavage structure that provides monomeric transcripts (70 ,71 ). This is the only example of a hammerhead-like RNA that has been found in an animal. This example, and that of the Neurospora mitochondrial DNA plasmid, indicate that non-pathogen sources may also provide a rich source of RNA catalysed cleavage reactions.
ACKNOWLEDGEMENTS
I gratefully acknowledge the financial support for our own work on RNA catalysis and plant pathogenic RNAs from the Australian Research Council Special Research Centre for Basic and Applied Plant Molecular Biology.
REFERENCES
1 Branch, A.D. and Robertson, H.D. (1984) Science223, 450-455.MEDLINE Abstract
3 Cech, T.R. and Herschlag, D. (1996) In Eckstein, F. and Lilley, D.M.J. (eds), Nucleic Acids and Molecular Biology, volume 10. Springer-Verlag, Berlin, pp. 1-17.
4 Pyle, A.M. (1996) In Eckstein F. and Lilley, D.M.J. (eds), Nucleic Acids and Molecular Biology, volume 10. Springer-Verlag, Berlin, pp. 75-107.
5 Keese P., Visvader, J.E. and Symons, R.H. (1988) In Domingo, E., Holland, J. and Ahlquist, P. (eds) RNA Genetics, volume 3; RNA Replication. CRC Press, Boca Raton, pp. 71-98.
27 Haseloff, J. and Gerlach, W.L. (1988) Nature334, 585-591.MEDLINE Abstract
28 Long, D.M. and Uhlenbeck, O.C. (1993) FASEBJ.7, 25-30.
29 Sheldon, C.C. Jeffries, A.C., Davies, C. and Symons, R.H. (1990) In Eckstein, F. and Lilley, D.M.J. (eds), Nucleic Acids and Molecular Biology, volume 4. Springer-Verlag, Berlin, pp. 227-242.
30 Tuschl, T., Thomson, J.B. and Eckstein, F. (1995) Curr. Opin. Struct. Biol. 5, 296-302.MEDLINE Abstract
31 Thomson, J.B., Tuschl, T. and Eckstein, F. (1996) In Eckstein, F. and Lilley, D.M.J. (eds), Nucleic Acids and Molecular Biology,volume 10. Springer-Verlag, Berlin, pp. 173-196.
39 Feldstein, P.A., Buzayan, J.M. and Bruening, G. (1989) Gene82, 53-61.MEDLINE Abstract
40 Feldstein, P.A., and Bruening, G. (1993) Nucleic Acids Res. 21, 1991-1998.MEDLINE Abstract
41 Hegg, L.A. and Fedor, M.J. (1995) Biochemistry34, 15813-15828.MEDLINE Abstract
42 Burke, J.M., Butcher, S.E. and Sarguiel, B. (1996) In Eckstein, F. and Lilley, D.M.J. (eds), Nucleic Acids and Molecular Biology, Springer- Verlag, Berlin, volume 10, pp. 129-143.
43 Couture, L.A. and Stinchcomb, D.T. (1996) Trends Genet.12, 510-515.
