MOLECULAR PLANT-MICROBE INTERACTIONS Vol. 4, No. 2, pp. 111-121, 1991, Current Review The Intriguing Viroids and Virusoid What Is Their Information Content and How Did They Evolve? Robert H. Symons Department of Biochemistry, University of Adelaide, Adelaide, S.A. S000, Australia Received 29 October 1990, Accepted 10 December 1990. ‘The low molecular weight, circular, single-stranded path- ogenic RNAs of plants are creating considerable interest because of their small size of 246 to 388 nucleotides, their unique structural features, and their ability to often produce dramatic symptoms as well as no symptoms on infected plants. In addition, many are important pathogens of agri- cultural crops. Their apparent inability to code for any proteins indicates that all interactions with cellular com- ponents must be by their sequence and their secondary and tertiary structures. Characterization of such interac- tions is likely to contribute to our understanding of the relation of structure to function of RNA in all cells. For general reviews, see Diener (1979, 1983, 1987), Riesner and Gross (1985), Keese and Symons (1987), and Keese er al. (1988). ‘These RNAs can be divided into two groups: the viroids,, which replicate independently because they do not require a helper virus, and the encapsidated viroidlike satellite RNAS or virusoids, which require a helper virus. The 16 ids that have been characterized so far (Table 1) can be divided into two main groups on the basis of comparative sequence homology. Avocado sunblotch viroid (ASBY) is the only member so far of one group, whereas the remainder of the viroids can be divided into two subgroups based on related sequences in the central part of the rodlike viroid molecules (see below; Keese and Symons 1985, 1987; Koltunow and Rezaian 1989). Only four virusoids have been discovered so far, all in Australasia (Table 2; Francki 1985, 1987; Keese and ‘Symons 1987; Symons, in press), although adifferent isolate of lucerne transient streak virusoid (VLTSV) has been found in Canada (Abouhaidar and Paliwal 1988). The virusoids in infected plants exist almost solely as circular molecules, either free or encapsidated within virions of the helper virus. Current address of R. H. Symons: Department of Plant Science, Waite Institute, Glen Osmond, S.A. 5064, Australia, This article is based on a lecture presented at a symposium marking the appointment of T. O, Diener as University of Maryland Distinguished Professor. The symposium, entitled “Subvieal Pathogens: Their Impact fon Molecular Biology, Plant and Human Medicine, and Evolutionary “Thought,” was held on March 5, 1990, at the University of Maryland twas cosponsored by the Center for Agricultural Biotechnology, Mary- Jand Biotechnology Institute, and the Department of Botany, University of Maryland, College Park. © 1901 The American Phytopathological Society This is in contrast to the similarly sized satellite RNA of tobacco ringspot virus (sTRSV) that is found as both linear and circular single-stranded molecules in vivo, but only the linear form is encapsidated (Kiefer et al. 1982; Bruening 1990). Although it is generally considered that viroids and virus- oids do not code for any polypeptides, the evidence is only circumstantial and definitive proof is lacking (Singer 1987; Keese and Symons 1987). Despite the inability of circular RNAs to be translated by eukaryotic ribosomes (Kozak 1979), the possibility cannot be dismissed that subgenomic fragments of viroids and virusoids may act as messenger RNAS in vivo and that initiation of peptide synthesis may occur at codons other than AUG (Dasso and Jackson 1989; Kozak 1989), The major problem in relating sequence and structure of viroids and virusoids to function is the scarcity of marker functions. When using mutants for this type of work, these can be obtained in two ways, either by isolating naturally occurring mutants of a viroid or virusoid (sequence vari- ants) or by preparing them by site-directed mutagenesis, of CDNA clones (Owens and Hammond 1988, 1990). How- ever, assay of the variation of the pathogenic effect on inoculated plants is one of the few ways available that have yielded definitive data (Visvader and Symons 1986). More recently, preliminary data that indicate an effect on ‘movement from cell to cell of potato spindle tuber viroid (PSTV) by mutations in the right-hand terminal loop of the viroid molecule should help to identify structural fea- tures that determine viroid host range (Owens and Ham- mond 1990). In the area of replication, in vitro analysis of the proc- essing of oligomeric precursors of ASBV and the four virusoids has identified a specific self-cleavage structure known as the hammerhead and has shown that about 20-30% of each molecule is required for the formation of these structures. This area is discussed below, together with sequence domains in viroids in terms of function and providing evidence for the evolution of viroids, and possibly virusoids, by RNA recombination. Sequence domains in viroids and virusoids indicate evolution by RNA rearrangement. Comparative pairwise sequence analysis of members of PSTV subgroup BI (Table 1) indicated the presence of five domains (Fig. 1), the bound- aries of which were defined by very sharp changes in se- uence homology, from high to low or vice versa (Keese and Symons 1985, 1987; Keese et al. 1988). Different pair- wise comparisons were always consistent in defining the Vol. 4, No.2, 1991 111‘exact position of the boundaries. The domain model was developed using viroid sequences only from PSTV subgroup BI (Table 1); these were the only ones available at that time, However, the viroids of apple scar skin viroid (ASSV) subgroup B2 have the same domains as shown by pairwise comparison of sequences (Koltunow and Re- zaian 1989). ‘The PSTV group of Table | is divided into two subgroups ‘on the basis of the sequences in the central conserved (C) domain, Within the C domain of each subgroup, there are sequences of about 30 nucleotides that are highly con- served among all members of each subgroup but which are quite different from those of the other subgroup (Vis- vader et al. 1985; Koltunow and Rezaian 1989). These central conserved sequences make up about one third of the C domain in each case. However, a common feature of both subgroups is the presence of a short inverted repeat sequence within the conserved nucleotides, which is shown by arrows in Figure 1 This domain model led to the proposal that the evolution of viroids involved the rearrangements of domains between viroids infecting the same cell followed by further evolution (Keese and Symons 1985). Experimental support of such a model is difficult to obtain, but new viroid sequences ‘Table 1. Classification of voids” according to groups and subgroups Length Designation Abbreviation __(aueleotides) A, ASBY group 'ASBV subgroup ‘Avocado sunblotch viroid ABV 246-251 B, PSTY group BI, PSTV subgroup Chrysanthemum stunt viroid csv 354 and 356 Citrus exocorts viroid cev 370-375 Coconut cadang-cadang viroid cccy 246 and 247 Coconut tinangajaviroid civ 254 CColumnea latent vieoid cy 370 Cucumber pale fruit veoid™ cPEV 403 Hop latent viroid HLY 256 Hop stunt viroid HSV 297-303 Potato spindle tuber viroid PSTV 359 ‘Tomato apical stunt void TASV. 360) ‘Tomato planta macho viroid TRMV 360 1B2, ASSV subgroup, ‘Apple scar skin viroid ASS 330 Australian grapevine viroid AGV 369 Grapevine yellow speckle viroid GS 367 Grapevine viroid 1B Gyis 363 Classification scheme of Keltunow and Rezaian (1989) for voids that "Cucumber pale {uit vitoid is really a sequence variant of hop stunt viroid. Virwsold Length abbreviation (nucleotides) Lucerne transient streak virus VETS 34 Solanum nodiflorum motte virus ySNMV. an Subterranean clover motte virus YSCMoV 332 and 388 Velvet tobacco motte virus YVTMoV 365 and 366 "Eneapsidated,virodlike, satelite RNAS. 112 Molecular Plant-Microbe Interactions that continue to appear provide strong indirect evidence. For example, in the recently sequenced columnea latent viroid (Hammond et al. 1989), the left-hand terminal (T1) and right-hand terminal (T2) domains show high sequence homology to the same domains in PSTV and tomato apical stunt viroid, respectively, with boundaries sharply defined and consistent with other pairwise comparisons (Fig. 2). ‘The presence of subdomain lengths of sequences of tomato planta macho viroid in the pathogenic (P) domain and of PSTV and of hop stunt viroid (HSV) in the C domain (Fig. 2) indicate that rearrangements can occur within a domain as well as at the boundaries. An even more scrambled viroid is Australian grapevine viroid (AGV) of ASSV subgroup B2 (Table 1) in which nearly all the 370 nucleotides appear to be derived from segments of citrus exocortis viroid (CEV), PSTV, ASSV, and grapevine yellow speckle viroid (Rezaian 1990). The sequence similarity of the segments in AGV varies from 52 to 100% of the corresponding sequences in the putative parent viroids (Koltunow and Rezaian 1989). Only one example has been found so far in the virusoids of such RNA rearrangements. Four separate isolates of subterranean clover mottle virus (SCMoV) contained two different-sized virusoids, which could occur together or separately (Francki 1985, 1987). Sequence analysis of the Virusoids showed the remarkable situation where the left- me c vom ce coao te fe =) Lett mand Pathogenic Variable Right nang Fig. 1. Model of void domains forthe potato spindle tuber vitoid (PSTV) group of vroids. The five domains, TI, P,C, V, and T2, were determined from sequence homologies between the viroids, The arrows depict inverted repeat sequence that can form a stem loop. Sequence elements ate those of PSTY subgroup BI (see Table I). R and Y indicate a short, ligopurine-oligopyrimidine helix, The same domains exist in apple sar skin viroid subgroup B2 (Koltunow and Rezaian 1989), This ‘modified from Keese and Symons (J98), (Reprinted with permission {rom Viroids and Viroi-like Pathogens, pages I-47. Copyright CRC Press, Ine, Boca Raton, FL.) Fig. 2. Schemati diagram of columnea latent viroid showing the sequence homologies to other viroids. Domains (abbreviations as given in Fig. 1) areshown together with residue numbers tthe boundaries ofeach domain, ‘The central conserved region (CCR) contains the inverted repeat sequence ‘This figure is adapted from Hammond et al. (1989). (Reprinted. by permission of Oxford University Press)hand part of each molecule was almost identical, whereas the right-hand parts differed in size and were completely different except for some short sequences (Fig. 3; Davies et al, 1990). Presumably, the two virusoids arose from recombination between three parents that have yet to be discovered. The mechanism of these RNA rearrangements is un- known. One feasible mechanism is discontinuous transcrip- tion where an RNA polymerase copying one viroid or Virusoid template changes over to copy a juxtapositioned second template at some point, most probably determined by the tertiary structures of the two templates (Keese and ‘Symons 1985, 1987). Examples of recombination between plus strand RNA. viruses are accumulating (King 1988). In the case of plant viruses, perhaps the most definitive example is that pro- vided by Allison et al. (1990) for the tripartite cowpea v(388)SCMoV 988 a, 360 chlorotic mottle virus where full infection requires the simultaneous inoculation of the three genomic RNAs RNAs | and 2 are each monocistronic whereas RNA 3 is dicistronie, coding for protein 3a (probably responsible for virus movement in the plant) and the coat protein Inoculation of plants with normal RNAs I and 2 and two RNAs 3, one with a deletion in the 3a cistron and the other with a deletion in the coat protein cistron, led to normal infection and the recovery of wild-type RNA 3. No systemic infection resulted when each of the mutant RNAs 3 was inoculated separately with RNAs I and 2. Hence, the RNA recombination event must have occurred with high frequency in many of the inoculated cells. Such. a result should encourage workers in the viroid and virusoid field to conduct similar experiments to test viroid and virusoid evolution by RNA rearrangements P domain of CEV plays. role in pathogenicity. Naturally ds 300 200 Fig. 3. Proposed secondary structures for v(388)SCMoV and v(332)SCMoV RNAS (9SCMoV, subterranean clover motte virusoid). Nucleotides conserved between both RNAS at similar positions inthe secondary structures fe boxed, An I8-nucleotide sequence in which Id of the nucleotides are common in both RNAS is indicated by a thick lin. The site of selfcleavage in each RNA is shown by an arrow. This iguee is reproduced from Davies eta. (1990) Vol.4,No.2, 1991 113elon ie balan CORY ‘otal number of nucteotides 96 a fences Between any two” 2 Maximum variation () between 0% 258 e117 20-24 68 4 " ° a8 Te ow Fig. 4. Summary of sequence analysis of 17 sequence variants of citrus exocorts viroids (CEVs) (Visvader and Symons 1985). Most nucleotide changes oceur in the Py Keese er al. (1988). (Re Inc, Boce Raton, FL.) occurring isolates of viroids, prepared from a single plant, often contain more than one sequence variant of a viroid, and these can be separated and sequenced by the prepa- ration of full-length cDNA clones from the viroid mixture (Visvader and Symons 1985; Rakowski and Symons 1989). Since DNA clones of viroids and their RNA transcripts that are infectious when inoculated onto susceptible plants (Cress et al. 1983) can be prepared, this offers a unique Opportunity to relate sequence, and hence structure, with pathogenicity. Sequence analysis of PSTV, CEV, and HSV variants showed that almost all sequence differences for each viroid are located within the P and variable (V) domains (Schnol- zer et al. 1985; Visvader and Symons 1985; Shikata 1990), This indicates that variation in severity of symptom ex- pression of these sequence variants is most likely determined by variation of sequence in one or both of these domains. Most data are available for CEV, and the results showed that, in 17 sequence variants, essentially all sequence varia tion occurred within the P and V domains in regions called P, and Pp (Fig. 4, Visvader and Symons 1985; Keese et al. 1988), To define whether the P or V domain or both were responsible for modulating symptom expression, chimeric infectious DNA clones were prepared in which approxi- mately one half of one sequence variant was joined through the C domain with the other half of another variant (Fig. 5). Two sequence variants were used, one which induced severe symptoms on tomato seedlings and the other which induced very mild symptoms. DNA constructs were con- veniently prepared using the BamHI and HindIII sites in the C domain. Infectivity results clearly indicated that the P domain determined symptom expression, while the V domain may have had an effect on the level of viroid that develops in infected plants (Visvader and Symons 1986). It was confirmed that the sequence of the progeny viroids was always the same as that of the clone used for inocula- 114 Molecular Plant-Microbe Interactions ind Py domains within the normal pathogenic (P) and va ined with permission from RNA Genetics, Vol. II, Variability of RNA Genomes, pages 71-98, Copyright CRC Press, lc (V) domains of CEV. This figure is reproduced from Fig. §. Schematic diagram of two parental citrus exocortsviroids (CEVS) and two ehimerieviroids i circular form, Parent CEV-A(2) induces severe symptoms on tomato seedings, whereas the other parent, CEV-DE30(a), induces very mild symptoms. Py and Py domains (Fig, 4) and nucleotide Tare indicated, B indicates BamHI site and H indicates HindIIt site in eDNA clones; these were used to construct the two chimeras (CEV- S./Mg and CEV-M,/S,). Ths figure is reproduced from Visvader and ‘Shmons (1986). (Reprinted by permission of Oxford University Press.) tion, However, the results provide no information on the molecular basis of the induction of the pathogenic response. Periodicity of sequences in viroids indicates origin by duplication. Most viroids exhibit a structural periodicity characterized by repeat units of 11 or 12 nucleotides for the PSTV subgroup, 60 nucleotides for ASSV, and 80 for ASBV (Juhasz et al. 1988). Although the repeat sequence was found in only roughly one third of the total sequence of members of the PSTV subgroup, the longer repeats of ASSV and ASBV occupied the full length of each molecule (Table 3). The periodicities of Table 3 appear to be specific for viroids, since none were found in other small RNAs, such as small nuclear RNAs and the virusoids, or in random sequences generated when the same base composition as that of some of the viroids was used (Juhasz et al. 1988). However, a conspicuous periodicity was not found for HSV or its sequence variant, cucumber pale fruit viroid, both members of the PSTV subgroup, which remains a puzzle, It was suggested by Juhasz et al. (1988) that the observed periodicity of viroids may play a role in the interactionof the DNA-like viroid RNA with proteins, since one of the characteristics of the protein-binding ability of DNA is structural periodicity (Travers 1987). An additional possibility is that this structural periodicity of most viroids reflects their evolutionary origin where se- quence duplication followed by mutation allowed an increase in size and the development of infectious molecules (Diener 1989). An excellent example of such partial dup! cation of a viroid molecule is shown during the infec- tion of coconut palms by coconut cadang-cadang viroid (CCCV). The 246-nucleotide viroid appears early in infec- Table 3. Examples of structural perio Repeat unit (at) 2 Consensus (at 69-172) 2 ‘CNGRRGRRAYCN (nt 244-129) Repeated 4.5 times in 330 nt Repeated 3 times in 247 nt Assv 0 AsBV 80 {rom Juhasz ea. (1988), R, purine (A or G) ¥, pyrimidine (Cor U);N,@ nonconserved nucleotide (A.C, G, or U}; and at, nucleotide ‘Abbreviations are given in Table | tion but, as symptoms develop, new higher molecular weight forms arise and eventually dominate the viroid population as the disease progresses (Imperial ef al. 1981; Mohamed et al, 1982). Duplication of the whole T2 domain occurs and is initiated at three separate sites within the V domain to give extra sequences of 41, 50, 55, or 2X 50 nucleotides (Fig. 6; Haseloff ef al. 1982; Keese and ‘Symons 1987; Keese et al. 1988). The functional implications of the development of these higher forms of CCCV during disease progression are not known, Obviously, they must provide some replication advantage, for example, by providing increased compe- tition for binding to these repeated sequences of some host component important for replication but which isin limited supply (Keese et al. 1988). It has been suggested that this duplication of the T2 domain of CCCV occurs by discon- tinuous transcription by an RNA polymerase switching or jumping from one template to another (Keese and Symons 1985), a model analogous to that proposed for the gen- eration of defective interfering RNAs of influenza virus (Jennings er al, 1983). Presumably, the sequence duplication in the other members of the PSTV group of viroids could have occurred in a similar way to eventually build up the full-length molecule. Replication pathway for viroids and virusoids occurs via ‘rolling circle mechanism. It is generally agreed that these covalently closed circular RNAS are replicated by a rolling 50 56 2x50 a coy ex ace! wut ice Meaty C240) ia didi, itty tite —> an : 2 le ’ eo ee BI Ge hide i m6 ! GOON Hae th dal, 2 el ae Ga, ae? 2g ui we cooy bs 848) eo . 29 v" x ash gb EE iyi ik gy iS hil ite? vy 6 Y Fig. 6. Partial sequence duplications in coconut eadang-eadang viroid (CCV). The two sequences, X and Y, with a total of 41, $0, 55, oF 2 X50 nucleotides are duplicated as indicated. The arrows pointing to the circles mark the boundaries of the duplicated sequences. Circied nucleot ight indicate the boundaries of the X and Y sequences, while the filled are sites of mutation in sequence variants. This figure is reproduced from Keese etal, (1988), (Reprinted with permission from RNA Genetics, Vol. II, Variability of RNA Genomes, pages 71-98, Copyright CRC Press, Inc, Boca Raton, FL.) Vol. 4, No.2, 1991. 115circle mechanism (Fig. 7; Branch and Robertson 1984; ‘Symons 1989), although many aspects of their replication are uncharacterized at the molecular level. In one of the two variations of the rolling circle mechanism (Fig. 7A), the infectious circular plus RNA is copied continuously by an unidentified RNA-dependent RNA polymerase to form aconcatameric minus strand (step 1). Specific cleavage of this strand produces monomers (step 2) that are circu- larized by a host RNA ligase and then copied by the same or a different RNA polymerase (step 4). Specific cleavage of the long linear plus strand produces plus monomers (step 5) that are circularized (step 6) to give the progeny RNA. ASBY and vLTSV most likely follow this pathway (Hutchins et al, 1986; Forster and Symons 1987a, 19876). In the other variation (Fig. 7B), the linear minus strand (step 1) is not cleaved but copied directly to give a linear plus strand (step 3) that is cleaved to give plus monomers and, finally, circular progeny. Most viroids and three of the four virusoids most likely follow this route (Branch et al. 1988; Hutchins et al. 1985; Davies et al. 1990). ‘There are many unanswered questions about these two related pathways. What RNA polymerase is involved in the replication? Does the same enzyme copy both the plus and minus strands? In the case of viroids, the enzyme must be host encoded, but there is no direct evidence to indicate whether it is RNA polymerase I, II, or IIT or some variant of their multicomponent complexes. In the case of the satellite virusoids, it is generally assumed that the helper viral-coded RNA polymerase, with or without host com- ponents, is responsible but no direct proof is available. ‘The possibility that a host-derived enzyme similar to that required for viroid replication is needed cannot be dis- missed. There is no information for a specific promoter sequence on the circular RNAS that defines the site of initiation of transcription, The sites of viroid and virusoid synthesis in the cell are also unknown. In the case of PSTV and CEV, Riesner and his colleagues have clearly shown that the nucleolus is a major site of viroid accumulation (Riesner 1987; Hard- ers et al. 1989), but nothing is known about the actual site of synthesis. A comprehensive analysis by in situ hy- bridization is required to provide more definitive data on these aspects for both viroids and virusoids. ‘Mechanism for processing precursors of the viroids of the PSTV group is still to be resolved. A major aspect of the rolling circle mechanismis the highly specific cleavage reaction required to produce monomers from coneatameric RNA. As considered below, a unique self-cleavage reaction that is mediated by RNA jin the absence of protein and demonstrated in vitro most likely operates in vivo for ASBV and the four virusoids. However, for all viroids of the PSTV group, no such definitive data exist, and resolution of the processing mechanism remains a major challenge in the viroid field So far, attempts to detect a nonenzymatic processing event for oligomeric PSTV have shown only very low (Robertson er al. 1985) or negligible (Tsagris et al. 1987a) activity. However, Tsagris er al. (1987) appear to have achieved reasonably specific cleavage of PSTV precursors and the production of circular monomers in low yield when greater-than-unit-length precursors were incubated with 116 Molecular Plant-Microbe Interactions extracts of plant nuclei, The results were taken to indicate that in vivo processing of PSTV precursors requires the action of specific plant RNases to give linear monomers which are then ligated to the circular form. As the authors, state, the possibility that proteins in the nuclear extract may have a structural rather than a catalytic role in the processing of the precursor RNA cannot be eliminated. The results are not quite as convincing as may first appear, because most RNases cleave single-stranded regions to give S'-hydroxyl and 2’,’-phosphate termini, which are those required for covalent coupling of RNA by plant RNA ligases (Branch et al. 1982). Furthermore, given that the precursor oligomeric RNA consistently folds into one or ‘more configurations, a nonspecific RNase could nick ex- posed single-stranded regions at approximately monomeric lengths with the subsequent RNA ligase coupling of juxta- positioned 5’-and 3-termini. Obviously much more specific data are needed to provide further support for or against a specific nuclease cleavage of precursor RNAS. It seems feasible that the in vivo mechanism for the viroids of the PSTV group could turn out to be some type of self-cleavage. Already three types of self-cleavage have been identified in only 12 self-cleaving RNAs that have been characterized so far: nine via the hammerhead reaction (see below), a different and still to be fully characterized structure for the minus sTRSV (Bruening et al. 1988; Bruening 1990), and an unknown structure for the plus and minus RNAs of hepatitis delta virus (Taylor 1990; Branch et al. 1990). It is eminently feasible that more types of self-cleavage will be found, one of which could account for the processing of precursors of all members of the PSTV Brow Site-directed mutagenesis defines a potential site for CE processing. Monomeric cDNA clones of PSTV and CEV are infectious on tomato seedlings when cloned at the BamHI site (see Fig. 8A for CEV) in certain BamHi-cut plasmid or phage vectors, as are the RNA transcripts and excised monomers (Tabler and Sanger 1984; Visvader and Symons 1985). However, monomeric cDNA clones pre- pared at other sites in the viroid molecule, for example the other site in Figure 8A, are not infectious unless they ‘can be excised as monomers with sticky ends. Presumably, the excised monomers are ligated to monomeric circles in vivo to initiate the rolling circle transcription from the circular DNA. Fig. 7. Rolling circle models for the replication of and the satellite RNA of tobacco ringspot virus. A, Pathway where both plus and minus RNAs cleave. B, Pathway where only the plus RNA, leaves, Sites of cleavage are indicated by arrows. Ths figure is reproduced from Forster and Symons (1987a). (Reprinted with permission of Cell, Press)By remarkable coincidence, infectivity of the intact mon- omeric DNA clones is correlated with the presence of an I I-nucleotide sequence 3’to the BamHI site in the viroid sequence that is repeated after the BamHI cloning site in the vector (black bar in Fig. 8B, wild-type sequence contains a G at the site of circled A and U) (Tabler and Sanger 1984; Visvader and Symons 1986). Since monomeric clones without a repeat sequence or with only a six-nucleotide repeat are not infectious, this indicates that the processing, B Lecco, see vector 8 ‘of RNA transcripts occurs within the I -nucleotide repeat sequence. Site-directed mutagenesis of G97 of the CEV sequence to U or A (Fig. 8B) resulted in an infectious clone, but 4 noninfectious excised insert (Visvader et al. 1986). 'Sur- prisingly, the infectious CEV progeny from the infectious mutated DNA clone contained the wild-type sequence. ‘The result was explained by the in vivo processing of the RNA transcripts, generated in vitro or in vivo, at one of Mi tue , ne BEeVEtER /goncuucey, 32 eke ee 1 ‘Summary of data indicating ste of cleavage of plus citrus exocorts void (CEV) and other potato spindle tuber viroid (PSTV) subgroup ids during rolling circle replication. A, Sites of the CEV molecule used in the construction of full-length cDNA clones. B, Potential in vivo processing sites of RNA transcripts derived from two point mutant cDNA clones of CEY; the circled nucleotides represent the point mutations introduced at G97 to give either A97 or U97. The wild-type cDNA clone a well as the excised insert were infectious when inoculated on tomato seedlings. In contrat, only the intact mutant clones were infectious. The IT-nucleotide repeat sequence correlates with infectivity and the proposed sites of processing at positions I, 2, or 3, which are indicated. C, Palindromie sequences and structures that may form at the proposed 5 and ‘processing sites of the RNA transcripts of the U9T mutant cDNA clone. The II-nucleotide repeat sequence is boxed. D, Conserved nucleotides land potential structures in the conserved central domain of all PSTV subgroup BI viroids. Nis a nonconserved nucleotide, The II-nucleotide sequence is boxed, and the proposed processing sites are indicated by arrows. In hop stunt viroid, a U is inserted at the positon of the filled circle. This U could pair with an A that is represented by N, leaving an unpaired C which is marked C. Potential A:G base pairs are joined by broken lines. Tis figure is adapted from figures in Visvader eal. (1988). (Reprinted by permission of Oxford Univesity Press.) Vol. 4, No.2, 1991 117the three phosphodiester linkages on the 3’ side of the point mutation (identified as 1, 2, or 3 in Fig. 8B). It is approxi- mately the same site identified by Tsagris er al. (1987) as one of two possible processing sites in PSTV from their experiments using nuclear extracts as described above. Given that processing occurs at one of the three proposed sites in CEV and that some other processing mechanism not operating, then one possible two-dimensional proc- essing structure is possible and is given in Figure 8C (Riesner et al. 1979; Visvader et al. 1985; Diener 1986). The boxed nucleotides are the I1-nucleotide repeats of Figure 8B, while the seven nucleotides (CGAGCUC) near the 3” end are vector sequences that provide limited base pairing. Processing would require excision of monomeric RNA by cleavage at two positions corresponding to sites 1,2, or 3 (indicated by arrows) and its subsequent enzymatic ligation to a circular form. Alternatively, the excision mech- anism may directly produce a circular monomer. During continuous rolling circle transcription of a circu- lar minus template, the structure analogous to that of Figure 8C is given in Figure 8D. All nucleotides are derived from the top strand of the C domain (Fig. 1), and those shown are conserved in all viroids of the PSTV subgroup. It is possible that the double cleavage considered in Figure 8C is replaced by a more complex sequence of events for oligo- meric precursors considered in Figure 8D. For example, if the palindromic structure is the actual structure required for cleavage, then cleavage may only occur on the top strand to release a linear monomer followed by the rearrangement of the structure to place the next cleavage site in an active structure. In this way, there would be sequential release of linear monomers. There is no real evidence for any of these models. Diener (1986) has considered models analogous to those in Figure 8D, while Riesner (1990) has described structural models with similar hybridization in the central conserved region for an oligomer of nearly four units. Despite the limited data, considerable speculation, and two-dimensional model building, we really cannot be confident about any aspect, t 9 ° | @ 0 on 0 on t t Fig. 9. The self-cleavage reaction that involves a phosphoryl transfer reaction from the S-hydroxyl ofthe ¥ nucleotide residue the Y-bydroxy! of the 5’ residue. This figure is reproduced from Symons (1983), 118 Molecular Plant-Microbe Interactions of the processing of oligomeric precursors of the PSTV ‘group of viroids. Hammerhead self-cleavage reaction in ASBV and the four virusoids. This area has been extensively reviewed recently (Symons 1989, 1990, in press; Sheldon eal. 1990; Forster et al. 1990; Davies et al, 1990), and only key aspects will be considered here. Plus and minus RNA transcripts prepared from cDNA clones of ASBV and vLTSV self-cleave in a highly specific manner both during transcription and after isolation in the complete absence of protein. The reaction involves a Mg*-catalyzed phosphoryl transfer reaction that cleaves the RNA to give termini with a 2’,3-cyclic phosphate and a S-hydroxyl (Fig. 9). On the basis of the specific sites of cleavage, a conserved two-dimensional hammerhead structure was developed (Fig. 10) that could be applied to the self-cleaving RNAs of plus and minus ASBV, plus and minus vLTSV, plus solanum nodiflorum mottie vi- plus velvet tobacco mottle virusoid, and plus vSCMoV, as well as the plus sTRSV. ‘The hammerhead structure of Figure 10 consists of three base-paired stems (I, II, and III) around an open single stranded region with 13 nucleotides (boxed) that are con- served in nine self-cleaving RNAS identified so far which cleave via this structure, Although definitive evidence has yet to be obtained, it seems likely that the self-cleavage reaction characterized in vitro is also responsible for proc essing of oligomeric precursors in vivo, The two-dimensional hammerhead structure of Figure 10 obviously does not explain why there is a nonenzymatic cleavage of RNA at the specific site. It is considered that, in the presence of Mg"*, the hammerhead assumes an active tertiary structure which lowers the activation energy suffi- ciently and specifically at the internucleotide bond of the self-cleavage site to allow the phosphoryl transfer of the self-cleavage reaction. When cleaved, the structure is con- sidered to relax and thus prevent the reverse reaction from occurring, since the overall reaction is theoretically re- versible. Inthe case of ASBV and the four virusoids, the sequences involved in the hammerhead reaction provide information W g:8 wal waNWN 5! only Fig. 10. Consensus hammerhead structure around the sef-leavage site (indicated by arrows) of avocado sunblotch viroi, the four virusoids, find the plus satellite RNA of tobacco ringspot virus (¥LTSV, lucerne transient streak virusoid) The 13 conserved nucleotides are boxed, and the nonconserved nucleotides are given as N.for a key step in the replication cycle. They are also the only sequences in these RNA pathogens for which we have information indicating their functional role. Thus, in the case of ASBV, the plus and minus self-cleavage sites are 14 nucleotides apart, and the sequences of the two hammer- head structures involve approximately the central one third of the total molecule (Hutchins et al. 1986; Forster et al. 1987). Whether this part of the ASBV molecule has any other functional role is unknown. The plus and minus self-cleavage sites of vLTSV are only six nucleotides apart, and the sequences that make up the plus and minus self-cleavage structures overlap and constitute the right-hand one third of the VLTSV molecule (Forster and Symons 1987a, 1987b). Regarding ASBV, we do not know if this region has any other function, and ‘we do not have any data on the information content of the rest of the molecule. Only the plus RNAS of the other three virusoids self- cleave. About 60 nucleotides are required for the hammer head structures, and these are derived from the top strand ‘on the left-hand side of the rodlike molecules (Keese and ‘Symons 1987; Davies et af. 1990). Taken together with the opposite, partially base-paired strand of each virusoid, the total accounts for about one third of each molecule No other functional information is known for these RNAS, except for specificity in relation to their helper viruses (Francki 1985, 1987; Symons, in press). Elements of tertiary structure are present ids and virusoids. It seems feasible that in the rodlike structures of viroids and virusoids additional elements of tertiary structure occur both in the native state as well as during various stages of their replication. Such structural elements would most likely play a functional role in the life cycle ‘of these pathogens as well as in the initiation of pathogenic effects One well-defined example of local tertiary structure in PSTV is the UV-induced covalent cross-linking between G98 and U260 in the highly conserved region of the C domain (Fig. 11; Branch et al, 1985). Presumably this region of the viroid is in an activated state that allows specific and efficient cross-linking to occur when irradiated with UY light. Similar cross-linking was also found in the conserved bulged helix of HeLa cell $S RNA (Fig. 11; Branch ef al. 1985) as well as in a viroidlike domain of hepatitis delta virus RNA (Branch er al. 1989) and 7SL RNA, which is a component of the signal recognition particle (Zwieb and Schiller 1988). The site of cross-linking in HeLa 5S RNA is in a region of the molecule analogous to that of the protein TFIIIA. binding site on Xenopus 5S RNA and near the binding " oy tas “cosee"06 1,6, Fig. 11. Schematic diagram of UV-induced crosslinked sites in HeLa ‘SS ¢RNA and potato spindle tuber viroid (PSTV), Double-headed arrows indicate positions of covalent attachment. This figue is reproduced from, Branch era. (1985) site for ribosomal protein LS, while 7SL RNA binds a number of specific proteins. Thus, the UV-sensitive element in PSTV could form a binding site for one or more proteins involved in replication and transport (Branch et al. 1990). ‘The numerous structural forms possible in PSTV and re- lated viroids, as has been well-documented by Riesner and his colleagues (Riesner 1990), could also provide multiple ‘opportunities in vivo for the binding of host protein and for the interaction with other cellular components. Another type of tertiary element is the hammerhead structure implicated in the in vivo processing of oligomeric precursors of ASBV, the four virusoids, and the plus STRSV, as already considered above. Likewise, processing in the PSTV group of viroids may eventually be charac~ terized as another activated tertiary structure. Putative conserved elements of introns in viroids and virusoids most likely have nothing to do with common and RNA processing. During the past nine years, it has become popular to speculate on the possibility that iroids and virusoids contain the conserved elements of group I introns, that these indicate a common origin, and, by strong implication, are involved in the processing of the multimeric precursors during rolling circle replication (€-g., Diener 1981; Dinter-Gottlieb 1986, 1987). Features that are usually absent from such sequence comparisons are a critical statistical analysis of the probability of the observed partial sequence homologies occurring by chance and the presence of such homologies in other RNAs which would be considered to be completely unrelated. It is a sobering thought to consider that the comparison of the sequence of PSTV with 25 computer-generated random sequences of the same size and base composition gave sequence homologies of 29.5 + 4.2%; (Fig. 2 in Keese et al 1988). In the case of the virusoids, processing of oligomeric precursors most likely occurs in vivo via a self-cleavage reaction involving the hammerhead structure, a reaction extensively characterized in vitro (Fig. 10; Forster and Symons 1987a, 1987b; Forster ef a/. 1987; Sheldon and Symons 1989; Davies ef al. 1990). As Diener (1989) points out, the absence of these group I-type sequence elements in the hammerhead structure effectively destroys any consideration that they play a role in RNA processing in these pathogenic RNAS. As already considered above, if a similar type of self-cleavage reaction is involved in the processing of precursors of the PSTV group of viroids, then this would provide further evidence against any role for these various group I intronlike related sequences. iven that these group T intronlike sequences in viroids and virusoids do have some functional role, it is feasible that they could be involved in some aspect of tertiary structure, for example in a way similar to proline that usually involved at bends in protein molecules. Our under- standing of the role of various sequence motifs in the tertiary structure of RNA is increasing; one recent example would be the structural and functional roles played by pseudo- knots (Pleij 1990). ‘ACKNOWLEDGMENTS I wish to acknowledge my great appreciation to the Center for Agricul tural Biotechnology, University of Maryland, fo the invitation to present Vol. 4, No.2, 1981 119the material described inthis manuscript at a symposium held to honor the appointment of T. O. Diener as a Distinguished Professor of the University of Maryland. His early work on viroids provided me with the inspiration to enter the viroid field in 1975, and, since that time, Thave valued greatly our friendship. LITERATURE CITED. [Abouhaidar, M. G., and Paliwal, Y.C. 1988. Comparison ofthe nucleotide sequences of virod-ike satelite RNA of the Canadian and Australasian strains of lucerne transient streak virus, J. Gen, Virol. 69:2369-2573. Allison, R., Thompson, C., and Ahlquist, P. 1990. Regeneration of a functional RNA virus by recombination between deletion mutants and ‘equitement for cowpea chlorotic motl virus 3a and coat protein genes for systemic infection, Proc. Natl. Acad. Sei, USA 87-1820-1824 Branch, A. D,, and Robertson, H. D. 1984, A replication eycle for vroids ‘nd other stall infectious RNAs. Science 22%:450-455 Branch, A. D., Robertson, H. D., Greer, C., Gegenheimer, P., Peebles, C., ‘and Abelson, J. 1982, Cell-reecireularization of viroid progeny RNA bby an RNA ligase from wheat germ, Science 217:1147-1149. Branch, A. D., Benenfeld, , J., and Robertsoa, H. D. 1985. Ultraviolet, Tight induced crosslinking reveals a unique region of local tertiary structure in potato spindle tuber viroid and HeL.a 58 RNA. Proc. Nat ‘Acad, Sei, USA 82:6590-6594 Branch, A. D., Benenfeld, B. J, and Robertson, H. D. 1988. Evidence Tor a single rolling circle in the replication of potato spindle tuber viroid. Proc. Natl, Acad. Sc. USA 85:9128-9132. Branch, A. D., Benenfeld, B. J., Baroudy, B. M., Wells, F. V., Gerin, “JL, and Robertson, HD. 1989. An ultravioletsensitive RNA struct tural clement ina vroid-like domain of the hepatitis deta viru, Science 245:649-652, Branch, A. D, Levin, B. J.,and Robertson, H. D, 1990, The brotherhood of citcular RNA pathogens: Virois, circuiar satellites and the delta agent. Semin. Virol. 1143-152. Bruening, G. 1990, Replication of satelite RNA of tobacco ringspot ‘Semin. Virol. £127-134 Bruening, G., Buzayan, J. M., Hampel, A., and Gerlach, W. L. 1988 Replication of smal satelite RNAS and vroids: Possible participation ‘of on-enzymic reactions, Pages 127-145 in: RNA Genetics, Vol. IL E, Domingo, J. . Holland, and P. Ahlquist, eds. CRC Press, Boca Raton, FL Cress, D. E., Kiefer, M. C., and Owens, R. A. 1983. Construction of infectious potato spindle tuber viroid cDNA clones. Nucleic Acids Res. T1:6821-6835, asso, C,, and Jackson, R. J. 1989. Efficient initiation of mammalian ‘translation at a CUG codon. Nucleic Acids Res, 17:6885.6497, Davies, C., Haseloff, J, and Symons, R. H, 1990. Structure, slf-cleavage ‘and replication of two viroid-like satelite RNAS (virusoids) of sub- {erranean clover motile virus, Virology 177-216-224 Diener, T. 0. 1979, Viroids and Viroid Diseases. Wiley and Sons, New York. 252 pp. Diener, T. 0. 1981. Are virois escaped introns? Proc. Natl. Acad. Se. USA 78:3014-5018, Diener, T. 0. 1983 Vitoids. Adv. Virus Res, 28:241-283 Diener, T. 0, 1986, Viroid processing: A mode! involving the central ‘conserved region and hairpin I, Proc, Natl. Acad. Sci. USA 83:58-62 Diener, T. 0. 1987, Biological properties. Pages 9.35 in: The Viroids. TO, Diener ed. Plenum Press, New York. Diener, ¥.0. 1989. Circular RNAs: Relic of precellular evolution, Proc Natl) Acad, Si, USA 86:9370-9374, Dinter-Gotlieb, G. 1986. Viroids and virusoids are related to group 1 introns, Proc, Natl, Acad, Sci, USA 83:6250-6254, Dinter-Gotlieb, G, 1987. Possible viroid origin. Viroids, virusoids and group Lintrons. Pages 189-203 in: The Viroid. T.O. Diener, ed. Plenum Press, New York, Forster, A. C., and Symons, R. H.1987a, Selfcleavage of plus and minus TRNAS of a virusoid and a structural model for the ative sites. Cell 49:211-220, Forster, A. C., and Symons, R. H. 1987. Self-cleavage of a virusoid| is performed by the proposed $S-nuceotide active site. Cell 509-16. Forster, A.C, Jelfies, A. C., Sheldon, C. C., and Symons, R. H. 1987 ‘Structural and ionic requirements for self-leavage of virusoid RNAS and trans self-cleavage of viroid RNA. Cold Spring Harbor Symp. ‘Quant, Biol, $2:249-289 120 Molecular Plant-Microbe Interactions Forster, A. C., Davies, C., Hutchins, C., and Symons, R. H. 1990 ‘Characterization of sei-cleavage in vtoid and virusoid RNAS. Methods Enzymol. 181-583-607. rancki, RIB. 1985, Plant virus satellites, Annu, Rev, Microbiol. 39:151- 1, Francki, R, 1. B. 1987, Possible void origin: Encapsidated viroidtike RNA, Pages 205-208 i: The Viroids. T. O, Diener, ed, Plenum Press, New York, Hammond, R., Smith, D. R.,and Diener, TO. 1989. Nucleotide sequence ‘and proposed secondary sructure of columnea latent vroid: A natural ‘mosaic of viroid sequences. Nucleic Acids Res. 17:10083-10094, Harders, J, Lukaes, N Robert-Nicoud, M., Jovin, J. M., and Riesner, 'D. 1989. Imaging of viroids in nuclei from tomato leaf tissue by in situ hybridization and confocal laser seanning microscopy. EMBO J. 5:3941.3949, Haseloff, J, Mohamed, N. ., and Symons, R. H. 1982. Viroid RNAS ofthe Gadang-eadang dies of coconuts, Nature (London 29:16. 321 Hutchins, C. J, Keese, P., Visvader, J. E., Rathjen, PL D., Melnnes, J.L, and Syinons, R. Hi. 1985. Comparison of multimeric plus and ‘minus forms of viroids and virusoids, Plant Mol. Biol. 4:293-204, Hutchins, C. J., Rathjen, P. D., Forster, A. C., and Symons, R. H. 1986, Selfcleavage of plus and minus RNA transeripts of avocado snbiotch viroid, Nuclelc Acids Res, 14:3627-3680, Imperial, J.S., Rodriguez, M.J.B., and Randles, J. W. 1981. Variation inthe vroid like RNA associated with cadang-eadang disease: Evidence for an inerease in molecular weight with disease progress. J. Gen. Virol 36.77.85 Jennings, P. A., Finch, J. T, Winter, G., and Robertson, J. S. 1983, Does the higher order structure ofthe influenza virus ribonucieoprotein sguide sequence rearrangements in influenza viral RNA? Cell 38:619- 6, Juhasr, A., Hegyi, H., and Solymosy, F. 1988, A novel aspect of the ‘information content of viroids. Biochim. Biophys. Acta 950:4S5-458, Keese, P., and Symons, R. H. 1985. Domains in viroids: Evidence of intermolecular RNA rearrangements and their contribution to vitoid evolution. Proc, Natl. Acad, Sei, USA 82:4582-4586, Keese, P, and Symons, R. H. i987. The structure of vroids and virusoids Pages 1-47 in: Viroids and Viroid-like Pathogens. J. S. Semancik, ed CRC Press, Boca Raton, FL, Keese,P., Visvader, J. E., and Symons, R. H. 1988. Sequence variability in plant viroid RNAS. Pages 71-98 in: RNA Genetics, Variability of RNA Genomes, Vol. Ill. E. Domingo, J.J. Holland, and P. Ahlguis, eds. CRC Press, Boca Raton, FL. Kiefer, M. C., Daubert, 8. D., Schneider, 1. R., and Bruening, G. 1982. Multimeric forms of satelite tobacco ringspot virus RNA. Virology TITS King, A. M. Q. 1988. Genetic recombination in positive strand RNA. viruses. Pages 149-165 in: RNA Genetics, Vol. H. E. Domingo, J.J Holland, and P. Ahlguist, eds. CRC Press, Boca Raton, FL. Koltunow, A. M., and Rezaian, M.A. 1989. A scheme for viroid ‘classification, Intervirology 30:194-201, Kozak, M. 1979. Inability of circular RNA to attach to eukaryotic ribosomes. Nature (London) 280:82-85, Kozak, M. 1989, Context effects and inefficient initiation at non-AUG, ‘codons in eucaryotic exl-free translation systems. Mol, Cell Bio 93507-5080. Mohamed, N.A., Haseloff,J., Imperial, J.S.,and Symons, R. H. 1982. Characterization of the different electrophoretic forms of the eadang- ceadang viroid. J, Gen, Vitol. 63:181-188, ‘Owens, R. A., and Hammond, R, W. 1988, Structure and function Felationships in plant viroid RNAs. Pages 107-125 in: RNA Genetics, Vol. Il, E. Domingo, J.J. Holland, and P. Ahiqust, eds. CRC Press, Boca Raton, FL. ‘Owens, R.A., and Hammond, R. W. 1990. Mutational analysis of viroids, ‘Semin, Virol, 1101-106, Picj, C. W. A. 1990, Pscudoknots: A new motif in the RNA game Trends Biochem, Sci. 15:143-147 Rakowski, A. G., and Symons, R.H. 1989, Comparative sequence studies ‘of variants of avocado sunbiotch virod, Virology 173:382-356, Reraian, M.A. 1990, Australian grapevine viroid~ Evidence for extensive recomibination between vitoids. Nucleic Acids Res, 18:1813-1818, Riesner, D. 1987. Subcellular location and in situ association with cellularcomponents. Pages 99-116 in: The Vitoids. T. 0. Diener, ed. Plenum Press, New York. Riesner,D. 1990. Suture of vieoids and thei replication intermediates, ‘Are thermodynamic domains also functional domains? Semin. Virol, 183559, Riesner, B., and Gross, H. J 1985 Viroids, Annu, Rev. Biochem. $4:531- 564 Riesner, D., Henco, K., Rokohl, U., Klot, G., Kleinschmidt, A. K., Domdey, H.,Jank, P., Gross, H. J.” and Stnget, H. L. 1979, Steucture and funetion of viroids. J. Mol. Biol, 133385-115, Robertson, H. D., Rosen, D. L., and Branch, A. D. 1985, CelLfree synthesis and processing of an infectious dimeric transcript of potato spindle tuber void RNA. Virology 142:441-487 Singer, H. L. 1987. Viroid function. Viroid replication. Pages 117-166 in: The Viroids. T. O. Diener, d, Plenum Press, New York. Schnolzer, M., Haas, B., Ramm, K., Hofmann, H., and Singer, H. L. 1985, Correlation between structure and pathogenicity of potato spindle tuber void (PSTV). EMBO J. 4:2181-2190, Sheldon, C.-C, and Symons, R. H. 1989, RNA stem stability in the formation ofa sel-cleaving hammerhead structure, Nucleie Acids Res 17:5679-5685, Sheldon, C. C. Jeffries, A. C., Davies, C., and Symons, R. H. 1990 RNA 'self-cleavage by the hammerhead Structure, Pages 227-282 in: Nucleic Acids and Molecular Biology, Vol. 4. F. Eckstein and D, M. 4. Lilley, eds. Springer-Verlag, Berlin. Shikata, E1990. New viroids from Japan, Semin, Viol, 107-115, Symons, R. H. 1989, Self-cleavage of RNA in the replication of small, pathogens of plants and animals, Trends Biochem, Sei. 14-445-450. Symons, RH. 1990. Selfcleavage of RNA in the replication of viroids and virusoids. Semin. Virol. 1:117-126 ‘Symons, R, H. Encapsidated circular satelite RNAS (virusoids) of plants, Curr. Top. Microbiol, Immunol. In press ‘Tabler, M., and Singer, H. L, 1984, Cloned single- and double-stranded DNA copies of potato spindle tuber viroid (PSTV) RNA and. co- Inoculated subgenomic DNA fragments ae infectious. EMBO J.3:3055- 3062, ‘Taylor, J. 1990, Structure and replication of hepatitis delta virus. Semin, Virol 1135-141 Travers, A.A. 1987. DNA bending and nucleosome positioning, Trends ‘Biochem, Sei 12:i08-112. ‘Tsagris, M., Tabler, M., and Sanger, H. L, 1987a, Oligomeric potato spindle tuber viroid (PSTV) RNA does not process autocatalytically under conditions where other RNAs do. Virology 157:227-231 ‘Tsagris, M., Tabler, M., MUblbaeh, H.-P., and Stinger, H. L. 19876, Linear oligomeric potato spindle tuber viroid (PSTV) RNAS are accurately processed in vitro to the monomeric circular viroid proper when incubated with a nuclear extract from healthy potato cells, EMBO 5621732183, Visvader, J. E., and Symons, R. H. 1985, Eleven new sequence variants of citrus exocortis viroid and the correlation of sequence with patho- sgenicity. Nucleic Acids Res. 13:2907-2920, Visvader, J. F., and Symons, R. H. 1986. Replication of in vitro con structed viroid mutants: Location of the pathogenicity-modulating domain of citrus exocortis vitoid. EMBO J. $:2051-2055, Visvader, J. E., Forster, A. C., and Symons, R. H, 1985, Infectivity and ‘in vitro mutagenesis of monomeric cDNA clones of citrus exocorts ‘roid indicate the ste of processing of void precursors. Nucleic Acids Res, 13:5843-5856, 2owieb,C. and Sehr, D. 1989, Low resolution three-dimensional models ‘of the 7SL RNA of the signal recognition particle, based on an intra- molecular crosslink introduced by mild irradiation with ultraviolet light, Biochem, Cell Biol, 67:434-492, Vol. 4, No. 2, 1991 121