This paper is available on line at
REVIEW ARTICLE
Plant virus gene expression strategies
Pedro I. Bustamante 1
Laboratorio de Biotecnología, Universidad Mayor, Campus Huechuraba, Santiago-Chile,
E-mail: pbustama@risc.umayor.cl
Roger Hull
John Innes Centre, Norfolk Rearch Park, Norwich, NR4 7UH, U.K.
E-mail: roger.hull@bbsrc.ac.uk
1
Corresponding author
Plant virus can cau rious loss to most, if not all, major crops upon which depend for food. Many virus are endemic,causing moderate loss each year. Others, such as tho causing rice tungro, give periodic vere epidemics. There are no fully collated figures for world-wide loss due to virus but some examples has been listed, i.e., rice tungro in SE ASIA and african cassava mosaic in Africa with 1,500 and 2,000millions dollars per year in loss respectively.
However, in recent years the understanding of the genome organisation of plant virus has incread in parallel with development of molecular biological techniques. The ability to obtain nucleotide quences of complete viral genomes has also permitted the elucidation and understanding of expression strategies ud by many different plant virus.This review is aimed to summari some aspects of the main strategies ud by plant virus to express their genomes.To date the Virus Identification Data Exchange (VIDE)databa (plant virus databa operated at the Australian National University in Canberra, Australia) contains 569characters for more than 890 plant virus species in 55 genera,according Gibbs (1994), and cited by Murphy et al. (1995).The VIDE databa is accessible through the Internet from the BioWeb rver biology.anu.edu.au/Groups/MES/vide/ (Brunt et al., 1996). Plant virus can cau rious loss to most, if not all, major crops upon which we depend for food. Many virus are endemic, ca
using moderate loss each year.Others, such as tho causing rice tungro, give periodic vere epidemics. There are no fully collated figures for world-wide loss due to virus but some examples has been listed by Hull (1994), i.e. rice tungro in SE Asia, african casava mosaic in Africa and potato virus in UK with 1,500,2,000 and 30-50 millions dollars per year in loss respectively.
In recent years the understanding of the genome organisation of plant virus has incread rapidly in parallel with the development of molecular biological techniques.The ability to obtain nucleotide quences of complete viral genomes has also permitted the elucidation and understanding of expression strategies ud for many different plant virus.
For many years, the only nucleic acid found in plant virus was RNA, but it is now clear that virus infecting plants may contain any one of the four types of genetic material:single-stranded RNA (ssRNA, about 75% of plant virus),double-stranded RNA (dsRNA, reovirus), single-stranded DNA (ssDNA, geminivirus) or double-stranded DNA (dsDNA, caulimo- and badnavirus). Of tho for which the genome is known or can be extrapolated by being in the
same group as a known virus, the vast majority have ssRNA of the (+) or mesnger polarity (termed (+) RNA). The (+)strand plant virus are classified into more than 25 distinct taxonomic groups (
Murphy et al., 1995) and show a wide variation in capsid morphology ranging from the rod shaped tobravirus, the filamentous potyvirus, to the icosahedral virus (e.g. bromovirus, sobemovirus, comovirus,tombusvirus, nepovirus, tymovirus ).
There are also some economically important virus with minus-strand and ambin genomic RNA species (rhabdovirus, tospovirus and tenuivirus ). Tospovirus is the only genus of plant virus in the Bunyaviridae family (German et al., 1992). However, Toriyama (1995) has also propod to include the Tenuivirus group as a new genus of plant-related virus of the Bunyaviridae family.
This review is aimed to summari some aspects of the main strategies ud by plant virus to express their genomes.
Genome organisation
Although the majority of known plant virus have RNA genomes, it is the smaller division of plant DNA virus which are better known. In the following ctions, mainly information related to plant RNA virus will be discusd and for information about plant DNA virus or some viroids, there are veral reviews that cover in depth the aspects (Symons, 1991;Lazarowitz, 1992,Timmermans et
al.,1994,Rothnie et al., 1994).
Plant RNA virus show a wide variation in their genome structure and organisation and may have different terminal structures such as cap structures or genome-linked proteins (VPg) at the 5' end, and a poly(A)-tail or tRNA-like structure at the 3' end of their RNA (reviewed by Goldbach et al.,1991). For some virus the genome needed for infection is divided between two or more gments which may be encapsidated in the same particle or in parate particles (multicomponent) and even like the tobacco necrosis virus (TNV), have associated satellite RNAs (Hull, 1990; Hull and Davies, 1992, Matthews, 1991). Most, if not all, plant virus genomes encode four or more proteins with functions that operate at various stages in the infection cycle.
Information on the genome organisation and quence similarities of the non-structural proteins, in particular of their RNA-dependent RNA polymeras (RdRps) and helicas, show that most plant RNA virus are genetically related and appear to have possible evolutionary links with some animal RNA virus (Ishihama and Barbier,1994,Strauss et al., 1996).
Table 1. Characteristic of RNA virus superfamilies
Group Lineage Virus groups Helica
type Common features Morphology Hosts and
vectors
Super 1 (POL
绿萝简笔画
1)Picorna-like
Poty-like
Sobemo-like
Arteri-like
Picornaviridae
Comovirus
Nepovirus
Calicivirus
Potyvirus
Bymovirus
Sobemovirus
Luteovirus
Nodavirus
Corona virus
Arterivirus
Torovirus
III
II
None
5’-VPG
3’-poli (A)
No subgenomic RNAs
Polyprotein processing
No overlapping ORFs
5’-cap
3’-poly (A)
Nested t of mRNAs
Enveloped
Icosahedral
注射用替加环素
Separate
encapsidation
Rod-shaped
Isometric
Mammals
Plants
Mammals
Plants
Plants
Inct
Mammals
Super 2 (POL
2)
Phage
Flavi-like
二十四节气图片Pesti-like
Carmo-like
RNA coliphages
Flavivirus
Pestivirus
Carmovirus
Tombusvirus
None
II
II
None
One ORF
No 3’-poly (A)
Enveloped
Icosahedral
Bacteria
Humans
Mammals
Plants
Super 3 (POL
3)
Tymo-like
Rubi-like
新语录
Tobamo-like
Tymovirus
Carlavirus
Potexvirus
Capillovirus
Rubella
Hepatitis E
Alphavirus
Tobamovirus
Tricornavirus
Hordeivirus
Tobravirus
Closterovirus
I
I
I
5’-caps
Subgenomics mRNAs
No Overlapping ORFs
表扬英语
Readthrough (most)
Icosahedral
Filamentous
Enveloped
Rod-shaped
Plants
Minus strand
RNA Paramyxovirida
e Rhabdoviridae Orthomyxovirid事业编怎么报名
ae Arenaviridae Filoviridae
Enveloped
Self-complementary
termini
Helical capsid
Overlapping ORFs
Some
gmented
genomes
Some with M
protein
Pleiomorphic
Enveloped rod
Mammals
Birds
Fish
Incts
Plants
Double-strand
RNA Reoviridae Bimaviridae
Segmented genome
5’-cap
3’-OH
ss RNA intermediates
Vertebrates
Plants
Arthropods
以色列耶路撒冷Mollusks Vpg, genome-linked protein; ORF, open reading frame
Adapted from Straus et al., 1996
The analogous modular arrangement of the coding quences also suggests that the virus may employ similar RNA replication strategies (Dolja and Carrington, 1992;Koonin and Dolja, 1993). This has led to the proposal (Goldbach, 1986;Koonin, 1991a;Koonin et al., 1991;Dolja and Carrington, 1992;Koonin and Dolja, 1993), bad on the three different types of quence motifs in the RdRps, of the division of the positive-strand RNA virus into three 'Supergroups'.
Supergroup I, which includes, the Picorna-like, Poty-like, Sobemo-like and Arteri-like. They have common features as, a VPg protein covalently linked to the 5' end of the RNA, 3'-poly (A), no subgenomic RNAs, polyprotein processing, no overlapping ORFs (e Table 1 for more detail). Supergroup II, which includes the coliphages, Flavi-like, Pesti-like and Carmo-like virus. They shares such features as, enveloped virions and no 3' -poly (A) (e Table 1).Supergroup III, which includes the Tymo-like, Rubi-like and Tobamo-like virus. They have common features as, 5' caps, subgenomics mRNAs, no overlapping ORFs and read-through expression strategy (most of them) (e Table 1).
The grouping is bad on quence homology of three similarly organid non-structural proteins of Sindbis virus, including the RNA capping enzyme, RNA helica and RdRp (Koonin, 1991a;Koonin, 1991b). Subgrouping can also be bad on conrved quence motifs of helicas (helica superfamilies 1, 2 and 3), proteas and the prence of capping enzymes (reviewed in Koonin and Dolja, 1993). RdRp however is the only domain of positive-strand RNA virus allowing an all-inclusive phylogenetic analysis.
Replication of plant RNA virus
Bustamante P.I. and Hull R.
Most virus encode proteins that are involved in viral nucleic acid replication. The discovery of the RdRps marked a major breakthrough in understanding the replication of progeny RNA from genomic viral RNA (reviewed in David et al., 1992; Ishihama and Barbier, 1994).
Potential RdRps have been described for many plant RNA virus including brome mosaic virus (BMV) (Hardy et al., 1979), cowpea chlorotic mosaic virus (CCMV) (Miller and Hall, 1984), turnip yellow mosaic virus (TYMV) (Mouches et al., 1984), alfalfa mosaic virus (AlMV) (Houwing and Jaspers, 1986), cucumber mosaic virus (CMV) (Hayes and Buck, 1990), TMV (Young et al.., 1987), turnip crinkle virus (TCV) (Song and Simon, 1994), red clover necrotic mosaic dianthovirus (RCNMV) (Bates et al.., 1995), tomato spotted wilt virus (TSWV) (Adkins et al.., 1995). While it is accepted that the role of RdRp in replication of RNA virus is esntial, the mechanism of its function is unclear and may differ for different virus groups. In in vitro studies on BMV, CCMV, AlMV, and TYMV the enzyme has only been shown to synthesi minus-strand RNA while complete replication of both minus-strand and new progeny plus-strand RNA has been demonstrated for CMV (Hayes and Buck, l990).
In addition, host factors have also been implicated in the replication complexes of TYMV (Mouches et al.., 1984), TMV (Meshi et al.., 1988), cowpea mosaic virus (CPMV) (Dersrs et al.., 1984), BMV
(Quadt and Jaspars, 1990;Quadt et al., 1993), CMV (Hayes and Buck, l990). The requirement for host-factors goes some way in explaining the inability of some extracted viral RdRps to fully complete a replication cycle. Propod mechanisms for the preci mode of action of veral viral RdRps as well as their structure and organisation have been reviewed extensively (for comprehensive reviews e Marsh et al., 1989,David et al., 1992,Ishihama and Barbier, 1994).
In general however, the viral RdRps are complex moieties, acting as RNA replicas or transcriptas, synthesising both (-) and (+) strands. Moreover, RdRps not only cataly RNA polymerisation but, in many virus, also effect RNA modifications (e.g. RNA methyltransfera activity).
Replication of plus-strand RNA virus
Replication of plant positive-strand RNA virus takes place in the cytoplasm of infected cells. RNA polymeras appear to be membrane-bound, and some proteins implicated in replication have membrane-binding domains, e.g. P58 encoded by RNA1 of CPMV. However the preci sites where RNA replication takes place have not been clearly defined and probably differ for different virus. Granular inclusion bodies have been invoked as the sites for TMV-RNA replication (Saito et al., 1987;
Okamoto et al., 1988). Replication of (+) strand RNA virus can be parated into four overlapping steps:(i) The uncoating of the virus, which expos the nucleic acid to the replication process. (ii) Translation, during which the viral RNA rves as a mesnger RNA and produces structural and non-structural proteins. This process is further divided into the primary or early translation of proteins required for replication, e.g. the RdRp, and condary or late translation of proteins with late functions, e.g. the coat protein. (iii) Replication of the genome which yields progeny RNA molecules takes place in two stages, both catalyd by an RdRp:(1) Synthesis of a full-length complementary (negative) RNA strand using the genomic (positive) RNA strand as a template. (2) Synthesis of progeny genomic RNA and subgenomics RNAs using the negative-strand RNA as a template. And finally (iv) the progeny genomic strands are encapsidated.The virus-encoded proteins required for RNA replication have been deduced from the composition of purified polymeras capable of copying genomic RNA to produce a negative strand, from the u of mutants, for divided genome virus from the minimum number of RNA gments needed to infect protoplasts and from the prence of conrved quence motifs found in polymeras in other systems (Quadt and Jaspars, 1990).
Initiation of the synthesis of a negative-strand on a positive-strand RNA template requires binding of
the polymera to a recognition site at the 3' end of the template. The 3' end of the RNA of many virus can be folded into a characteristic condary or tertiary structure which includes the RNA polymera binding site. Sequences at the 5' end of the genomic RNA are also required for RNA infectivity (French and Ahlquist, 1987) and presumably reflect the requirement for binding of the polymera at the 3' end of negative-strand RNA.
One system currently being ud to study positive-strand RNA virus replication is the plant bromovirus group (Ahlquist, 1992). The bromovirus are icosahedral, positive-strand, tripartite RNA virus in the alphavirus-like superfamily. The two bromovirus proteins required for RNA replication, 1a and 2a, are translated from genomic RNA1 and RNA2, respectively, while proteins required for infection spread are translated from genomic RNA3 and a subgenomic mRNA, RNA4, transcribed from negative-strand RNA3. Protein 1a (109 K) contains an N-terminal m7G methyltransfera-like domain thought to be involved in capping viral RNA (Rozanov et al., 1992) and a C-terminal helica-like domain (Gorbalenya et al., 1988). Protein 2a (94 K) contains a central polymera-like domain (Kamer and Argos, 1984). Site-specific mutagenesis studies showed that all three conrved domains in 1a and 2a are required for RNA synthesis (Kroner et al., 1990;Traynor et al., 1991).
Bromovirus RNA synthesis can be divided into three distinct steps:negative-strand synthesis, positiv
e-strand synthesis, and subgenomic mRNA transcription. Each of the steps is differentially regulated. For example, negative-strand RNA accumulation plateaus by 8 h post-inoculation, while positive-strand genomic RNA and subgenomic mRNA continue to accumulate until or beyond 20 h post-inoculation (Kroner et al., 1990). French and Ahlquist (1987) described that BMV-directed replication of RNA3 in vivo depends on cis-acting quences in three regions of RNA3:the 3' and 5' noncoding regions and the intercistronic noncoding region. Later, Janda and Ahlquist (1993) demonstrated that BMV RNA3 derivates can be replicated and direct subgenomic mRNA transcription in yeast expressing BMV proteins 1a and 2a from DNA plasmids.
Recently, it has been shown that yeast expressing 1a and 2a and replicating RNA3 derivatives can be extracted to yield BMV-specific template-dependent RdRp activity (Quadt et al., 1995). Moreover, even though RdRp activity was asssayed on in vitro-supplied BMV-RNA templates, it was found that RdRp can only be isolated from cells expressing certain BMV RNA template quences as well as 1a and 2a. Strong correlation between extracted RdRp activity and BMV (-)-strand RNA accumulation in vivo was found for all RNA3 derivatives tested. Thus, extractable in vitro RdRp activity paralleled formation of a complex capable of viral RNA synthesis in vivo. The results suggest that asmbly of active RdRp requires not only viral proteins but also viral RNA, either to dir
ect contribute some nontemplate function or to recruit esntial host factors in the RdRp complex (Quadt et al., 1995)
Plant virus gene expression strategies
Zaccomer et al. (1995) have reviewed recently other elements believed to be involved in virus replication:
1)tRNA-like structures. It has long been known that the
RNA genomes of certain positive-strand plant virus have tRNA-related properties (reviewed in Mans et al., 1992). The 3' tRNA-like structures have been shown to be involved in minus-strand synthesis in the ca of TMV (Dawson et al., 1988), BMV (Miller et al., 1985) and TYMV (Tsai and Dreher, 1991).
2)Pudoknots. In addition to the pudoknots in the
tRNA-like structures, a few virus have pudoknots upstream of the structures which appear to participate in RNA replication. In TMV the most downstream of the six double-helical structures that compo the three pudoknots already mentioned located just upstream of the tRNA-like structure,
is required for replication (Takamatsu et al., 1990). Also Leathers et al. (1993) have reported that this region probably is involved in translation. However, pudoknots prent either in BMV-RNA3 (Lahr et al., 1993) or TYMV (Tsai and Dreher, 1992) are only involved in RNA replication.
3)Poly(A) structures. In CPMV, both M-RNA and B-RNA
contain the quence UUUUAUU immediately followed by the poly(A) tail. This heptanucleotide quence together with the first four A residues immediately downstream can adopt a hairpin structure. A similar structure can also be formed by the M-RNA of RCMCV (Shanks et al., 1986). In CPMV B-RNA, deletions from the 3' end of the RNA can prevent formation of the hairpin and dramatically interfere with RNA replication (Eggen et al., 1989).
4)Internal control region (ICR)-like quences. Similarities
exist between viral RNA quences (bromovirus, cucumovirus, tobamovirus, tobravirus and tymovirus) and the ICR2 of the RNA polymera III promoter of eukaryotes (Marsh et al., 1989). A role for the quences in replication has been demonstrated for BMV RNA (Pogue et al., 1992) and is also propod for the RNA of CMV (Boccard and Baulcombe, 1993) and AlMV (van der Vosn et al., 1993). The prence of ICR-like quences suggests that a host RNA polymera III s
ubunit and/or one of its cofactors could participate in viral RNA replication.
Replication of minus-strand RNA virus
Negative-strand RNA virus are a large and diver group of enveloped virus. They are found in hosts from the plant and animal kingdoms, and have a wide range of morphologies, biological properties and genome organisations (Conzelmann, 1996). A major distinction is made between virus who genome consists of a single RNA molecule (order Mononegavirales), including the families Rhabdoviridae, Paramyxoviridae and Filoviridae, and tho possing multipartite (gmented) genomes, comprising the families Orthomyxoviridae (six to nine gments), Bunyaviridae (three gments) and Arenaviridae (two gments).
Characteristically, the genetic information of negative-strand RNA virus is exclusively found in the form of a ribonucleoprotein complex (RNP) in which the genomic or antigenomic ssRNA is tightly encapsidated in a nucleoprotein (N or NP) and associated with the virus RdRp. In the ca of non-gmented virus, the latter consists of a catalytic subunit (L) and a non-catalytic cofactor, a phosphoprotein (P). After infection of a cell, the RNP rves as a template for two distinct RNA synthesis functions, transcription of subgenomic, usually non-overlapping mRNAs and the replication
of full-length RNAs (for detailed reviews e Galinski, 1991). The RNP genomes appear to poss only one promoter, at the 3' end of the RNA where the virus RdRp enters for both mRNA transcription and genome replication (Conzelman, 1996).
For virus in the family Bunyaviridae, the polymera protein, either acting alone or in concert with undefined viral or cellular factors, must first functions as a cap-dependent endonuclea to generate a primer for transcriptions of a non-encapsidated transcript of subgenomic length. At some point, the polymera must switch to a process of independently initiating transcription at the preci-3' end of the template and producing an encapsidated, full length transcript (Schmaljohn, 1996). Presumably, some viral or host factor is required to signal a suppression of the transcription termination signal responsible for generation of truncated mRNA and also to prevent the addition of the capped and methylated structures to the 5' termini of the cRNAs (Schmaljohn, 1996). For the rhabdovirus vesicular stomatitis virus, the switch to antigenome synthesis appears to be controlled by the N protein (Banerjee, 1987).
In animal virus such Influenza A (Orthomyxoviridae), which has a genome consisting of eight ssRNA gments of negative polarity, the replication and transcription of the virus genome are catalyd by a virus-encoded RdRp (Kobashagi et al., 1992, Huang et al., 1990). The RdRp is comp
od of three subunits, PB1, PB2 and PA, which are tightly associated at the double-stranded stem region of the panhandle formed by the 5' and 3' termini of each RNA gment (Huang et al., 1990, Hsu et al., 1987). RdRp plays an esntial role in both replication and transcription but little is known about the molecular mechanism of replication. However, some evidence suggests that PB1, PA and the nucleoprotein can support the replication of the influenza virus genome as well as the transcription to yield uncapped poly (A)+RNA but PB2 is specifically required for the synthesis of capped RNA (Nakagawa et al., 1995).
Virion-associated RdRp polymera activity has been also found in plant rhabdovirus. In the ca of wheat rotte stunt virus, both detergent-treated virions and isolated nucleocapsids exhibit RNA polymera activity. Like animal rhabdovirus, the enzyme activity can be regained upon mixing of L and NS proteins and using N-associated RNA template. Products synthesid in vitro by the virion-associated RNA polymera of plant rhabdovirus contain genome-length and single-strand virus complementary RNA (vcRNA) indicating that the RdRp acts not only as transcripta but also as replica.
Replication of ambin virus
昆明旅RdRp activity has been detected in detergent-disrupted virions of animal-infecting members of the Bunyaviridae (Vialat and Bouloy, 1992) and has been directly linked to the L-protein of bunyamwera virus, the type member of the family (Jin and Elliot, 1991). An RdRp activity has been found associated with virions of TSWV, a plant- and inct-infecting member of the family Bunyaviridae. Radiolabelled nucleoside triphosphate was incorporated into trichloroacetic acid-precipitable products by detergent-disrupted, purified TSWV virions. The predominantly double-stranded RNA products were RNa-resistant at high but not low salt concentrations. Discrete products of approximately 3.0 kb were synthesid that hybridid to purified TSWV RNA and transcripts of cDNA clones encompassing parts of each of the three genomic RNAs. The predominant products were viral n although significant amounts of viral complementary n S RNA products were also synthesid (Adkins et al., 1995).
Bustamante P.I. and Hull R.
Barbier et al. (1992) working with the Tenuivirus rice stripe virus (RSV) a virus with some genome organisation features in common with TSWV, isolated an RNA polymera activity by CsCl centrifugation from purified RSV ribonucleoproteins (RNPs). The active fraction contained two viral structural proteins, a 30 K nucleocapsid (N) protein and a 230 K putative polymera protein. An in vi
tro RNA synthesis system was reconstituted using this RNA-free protein fraction and short model templates carrying the conrved 5' and 3' terminal quences. This showed that, as in the ca of influenza virus, a minimum promoter function resides in the panhandle condary structure formed by the complementary termini or in the 3' terminal quence of 11-14 nucleotides in length.
Modes of gene expression
Another major problem facing RNA virus with limited genome size is their obvious dependence on the host eukaryotic protein-synthesising system. The small genomes are also expected to encode a range of virus proteins. The strategies of expression that have emerged from recent studies suggest that the viral genomes appear to have evolved to overcome the obvious constraints of the plant host system.
The eukaryotic 80S ribosome is usually able only to translate the first ORF in the 5' region of an mRNA, according the "scanning ribosome model" propod by Kozak (1991). The model states that the 40S ribosomal subunit (carrying Met-tRNA imet and various initiation factors) binds initially at the 5’end of mRNA. The ubiquitous m7G cap and the associated cap-binding protein(s) explain the predilection of eukaryotic ribosomes to engage mRNA at the 5’-end. Then the migrating 40S ribosom
al subunit stalls at the first AUG codon, which is recognid in large part by ba pairing with the anticodon in Met-tRNA imet. However, the stop-scanning step and hence lection of the initiator codon, is susceptible to modulation, by context, at least in vertebrates and lection of more distal AUG is permitted under certain defined circumstances (Kozak, 1991).
The possibility that might be cas of internal translation initiation has been shown. Pelletier and Sonenberg (1988) have propod that there is efficient internal initiation on poliovirus RNA. The evidence comes from experiments exploiting the fact that translation of a dicistronic mRNA with two non-overlapping ORFs (A and B) generally gives a low yield of B protein (located downstream) compared with A.
Pelletier and Sonenberg (1988) ud a construct in which the entire 5' untranslated region (736 nt) of type 2 poliovirus was placed in the intercistronic region of a capped dicistronic mRNA. When the cells expressing the dicistronic mRNA are infected with the poliovirus, the synthesis of protein A (upstream) was inhibited and protein B enhanced, demonstrating that downstream cistron translation is independent of upstream. In addition, cell-free extracts from poliovirus-infected cells translated cistron B but not A. Similar results have been published also by Jang et al. (1988) for encephalomyocarditis virus RNA and more recently for bovine viral diarrhea virus by Poole et al. (19
95).
The main strategies ud by plant virus to allow protein synthesis in a eukaryotic system from positive n RNA genome containing more than one gene are discusd below (e Figure 1 for some illustrations).
Subgenomic RNAs
The expression of internal genes such coat protein (CP) of the positive RNA virus is frequently mediated via subgenomic RNAs, considered in this study as mRNAs (e Figure 1a). The mRNAs are encapsidated in some virus, but not in others. Among plant RNA virus, the mechanism of synthesis of the subgenomic RNA encoding the CP has been examined in veral virus, i.e., TMV (Palukaitis et al., 1983), CMV (Jaspars et al., 1985). From the studies, two mechanisms have been propod to explain the synthesis of subgenomic RNA species:(1) During (-) RNA strand synthesis by the RdRp, premature termination could lead to the formation of (-) RNA strands of subgenomic length that could rve as template to generate the subgenomic (+) RNA; alternatively (2) the subgenomic (+) RNA could be synthesid via internal initiation on (-) RNA strands of genomic length.
The evidence from in vivo and in vitro experiments with various RNA virus clearly tends to favour the cond mechanism. Since subgenomic RNAs contain at their 3' end the elements required for the production of complementary subgenomic RNA chains, various explanations have been put forward to account for the lack of autonomous replication of subgenomic RNA. The are that (1) the quence contained within the subgenomic RNA is insufficient for replication of the subgenomic RNA; (2) the subgenomic RNA, which is frequently a highly efficient mRNA, may not be available for replication; and (3) the subgenomic RNA would be produced late in infection or at time when negative-strand synthesis has cead. From different experiments, the first explanation is certainly the most likely.
Miller et al. (1985), studying the mechanism of BMV subgenomic RNA4 formation from genomic RNA3 by using the in vitro RdRp system provided, the first unequivocal evidence that the subgenomic RNA of a positive-strand RNA virus is synthesid (at least in vitro) by internal initiation of positive-strand RNA synthesis on a negative-strand template.
