key: cord-0008819-czmv6itq
authors: Strauss, Ellen G.; Strauss, James H.
title: RNA viruses: genome structure and evolution
date: 2005-06-09
journal: Curr Opin Genet Dev
DOI: 10.1016/s0959-437x(05)80196-0
sha: 73821617665e87d15a49e94479d262fec3eb01b0
doc_id: 8819
cord_uid: czmv6itq

The explosive pace of sequencing of RNA viruses is leading to rapid advances in our understanding of the evolution of these viruses and of the ways in which their genomes are organized and expressed. New insights are coming not only from genomic nucleotide sequence comparisons, but also from direct sequencing of transcribed mRNAs and of RNAs that serve as intermediates in replication.

During the past decade, advances in the technology of cloning and sequencing have made possible the very rapid acquisition of information about the organization of RNA virus genomes. At the current time, complete genomic sequences exist for at least one representative of almost all the known RNA virus groups; this information has led to the elucidation of evolutionary relationships between many superficially diverse groups. It has often been found that overall strategies of replication, such as the relative location within a genome of genes of similar function, the presence or lack of subgenomic mRNAs, readthrough of termination codons for downstream products, or use of an 'ambisense' transcription strategy, show that certain viruses are distantly related, even when no sequence homology remains [1] . In addition, a number of short amino-acid-sequence elements have been recognized as indicators of the function of particular proteins and it has been suggested that proteins sharing such motifs are also related by descent from common ancestors. These indicators include the 'GDD motif, which is characteristic of RNA polymerases [2,3,4-], the 'DEAD motif [5] and the 'G-x-x-GKS/T motif, which are characteristic of RNA helicases [5] [6] [7] , and the sequence elements surrounding the amino acids in the active sites of viral-encoded proteases [8] .

Even among well studied viruses known to be closely related, sequencing often clarifies their relationships.

From sequence analysis, the genera of the Picornaviridae have been realigned [9] and the taxonomy of the Paramyxoviridae clarified (see discussion below). Sequence comparisons have also been useful in epidemiological studies, both longitudinal studies to determine rates of change of a particular virus in nature, and com-parisons of geographic isolates to pinpoint the origins of epidemic viruses [10-12,13-] .

One of the most exciting and unanticipated results of genome structure comparisons has been the discovery that plant virus counterparts exist for almost every major group of RNA-contalning animal viruses. In the case of the bunyaviruses, the plant and animal representatives are so closely related as to be placed within the same family, and an argument could be made that the plant tenuiviruses should be considered as belonging to the same genus as the animal uukuviruses and phleboviruses. In other cases, the plant and animal viruses are counterparts in the sense that they share genome organization and transcription strategies, and may share sequence homology in a number of proteins (even though each may possess unique genes required for replication in their respective hosts), but the plant and animal virions may be quite different in morphology. The existence of such viruses indicates that plant and animal viruses have radiated from a small number of ancestral prototypic viruses, and that the repertoire of successful replication modes may be limited. These studies have also made it clear that recombination has played an important role in the evolution of RNA viruses, and that viruses can acquire the ability to jump across wide phylogenetic barriers, whether by recombination or adaptation, rather more easily than would have been suspected a decade ago.

Because it is impossible to discuss all of the significant advances of the past year in a short review, we have selected three areas of particular interest. These include: a newly described mechanism used by the paramyxoviruses to shift the translation frame that is useful for the classification of these viruses; updated information on the replication strategy of coronaviruses; and additional insights into plant virus counterparts of animal viruses.

Coding strategy of the V/P genes of the Paramyxoviridae: a new mechanism for translational frame-shifting Many viruses are known to increase the information content of their genomes by translating their RNA in more than one reading frame. Differential splicing (for viruses that replicate in the nucleus), translation initiation at more than one start codon, and ribosomal frame shifting have been described. In the V/P gene of the paramyxovirus simian virus 5 (SV5), two non-templated nucleotides are added during transcription of mRNAs to shift the reading frame in some, but not all, transcripts [14] . Within the past year, reports of similar mechanisms used by several other members of the paramyxovirus family have been described. Not only is the addition of non-templated G residues to shift the reading frame found in many, but not all, of these viruses, but the details differ among the different viruses and appear to be useful in the classification of the members of this family.

All members of the family Paramyxoviridae, in the order Mononegavirales, harbor a genome comprising single segment of negative polarity RNA (approximately 10--12kb). The family contains three currently recognized genera: pneumovirus, paramyxovirus, and morbillivirus. The genus paramyxovirus includes Newcastle disease virus (NDV), the type virus, Sendai virus, human parainfluenza virus (PIV)-I, PIV-2, PIV-3, and PIV-4, mumps virus, and SV5, all of which contain a neuraminidase activity. The morbillivirus genus encompasses measles virus, rinderpest virus, and canine distemper virus, all of which are quite similar to paramyxoviruses but lack the neuraminidase activity. Pneumoviruses (respiratory syncytial virus and pneumonia virus of mice) are distinct, and contain a number of extra genes in addition to the nucleocapsid (N), phosphoprotein (P), membrane protein (M), glycoprotein (G), fusion protein (F), and large polymerase protein (L) genes common to all members of the family. All of the genes are monocistronic with the exception of the P (or V/P) genes.

Paramyxoviruses and morbiUiviruses, but not pneumoviruses, increase the coding capacity of the V/P gene by translating products from more than one reading frame. The details of the V/P gene strategy are illustrated in figure 1. Two mechanisms are used: initiation of translation at two different AUGs, and the ad dition of non-templated G residues to shift the reading frame. In some cases all three reading frames are utilized and up to four protein products are produced from the V/P gene. In measles [15] , PIV-1 [16. ], PW-3 [17..] , and Sendai viruses [18], a P protein of approximately 600 amino acids is translated from an mRNA that is a faithful complement of the genome. In addition, a smaller protein (C) of about 200 amino acids is translated from the same mRNA by internal initiation at a methionine in a second reading frame. The V protein, which is amino-coterminal with P but contains a carboxyterminal domain that is highly conserved and rich in cysteine residues, is encoded by an mRNA that is formed when a single non-templated G residue is inserted into the mRNA during transcription. This shifts the frame to the third possible frame, resulting in a V protein of about 400 amino acids. Note that for PIV-3, an mRNA containing two non-templated G residues is also produced and is translated into a fourth protein, the D protein.

In PIV-2 [19o], Pry-4 [20,], sv5 [14] , mumps [21" ], and probably NDV, there is no single long open reading frame (ORF). Instead there are two significant ORFs which overlap in the middle of the gene. Translation of the mRNA resulting from faithful copying of the genome gives rise to the V protein. In the case of PIV-2 and SV5, exacdy two non-templated G residues are inserted to shift the frame to the P protein frame; in PIV-4 the number of G's inserted is more variable, although the specific insertion of two gives the P protein. For mumps, insertion of two G's results in the mRNA for the P protein, whereas insertion of four G's gives the message for another product, the I protein ( Fig. 1) . All of the P proteins of this latter group are 391-399 amino acids long.

have not yet been established, nevertheless it is fascinating that in viruses that are so closely related and similar in many aspects of their replication, some should use one gene to translate only one protein, whereas up to four proteins are produced from a single gene in others. The disparity in use of the P protein among the Paramyxoviridae also suggests either that evolution to use multiple reading frames within a gene proceeds more rapidly than might have been predicted some time ago, or that the ancestral paramyxovirus used multiple reading frames, and that as the function of some of these translation products became non-essential during evolutionary divergence some paramyxoviruses lost the ability to produce them.

Based on this information, there has been a suggestion to reclassify the Paramyxoviridae [20" ]. The morbilliviruses and pneumoviruses would remain unchanged, but the paramyxovirus genus would be split into two genera. One genus would contain PIV-1, PIV-3, and Sendal virus, and the second would contain PIV-2, PIV-4, SV5, mumps, and NDV. Notably, a recent analysis of the aligned sequences of the L protein of these various viruses led to precisely the same assignment [22--] .

Many RNA viruses are known to produce subgenomic mRNAs. The plus-stranded togaviruses as well as a number of plus-stranded plant viruses transcribe a subgenornic mRNA for the structural protein(s), and all of the minus-stranded viruses produce mRNAs that are less than genome length. These mRNAs lack a cL~-acting terminal element (either a sequence or a structure) essential for initiation of replication by the viral replicase, and thus cannot replicate in infected cells. The coronaviruses also produce up to seven different subgenomic mRNAs of genomic polarity. Sethna et al. [23] reported the presence of minus-strand copies of these mRNAs in coronavirus-infected cells and suggested that the subgenomic RNAs could replicate. As discussed below, during the past year a number of groups have reported that coronavirus mRNAs are indeed actively replicated in infected cells. The Coronaviridae are plus-stranded RNA viruses with the largest known RNA genomes, more than 27 kb. The genome organization of a typical coronavirus is illustrated in figure 2a (reviewed in [24] ). Almost 20 kb at the 5' end is devoted to two long ORFs encoding the RNA replicase. The replicase is translated from the genomic RNA as two large potyproteins, with the larger, presumably produced by ribosomal frame-shifting, cormsponding to the entire 20 kb region. These polyproteins are believed to be posttranslationally cleaved by two cysteine proteases, which are encoded within them, into an unknown number of final products [25, 26] . -/q'a atreaqtuatu papoaua-gru> o~ ptr¢ 'p!sdeaoapnu i-ca uoBnlo^a pue uo!lez!ue~o ~u~D lost from some coronavirus-like viruses but not from others. All of the shared characteristics suggest that the toroviruses and coronaviruses have diverged from a common ancestor, but this argument will be considerably strengthened ff it is found that toroviruses, like coronaviruses, have independently replicating mRNAs.

One of the more interesting discoveries of the past decade has been the finding that many plant viruses have animal virus counterparts (Table 1) to which they are related to varying degrees. This is a topic that has received a great deal of attention in the last year, as more complete sequences of both plant and animal viruses have been determined. A recent volume of Seminars in Virology was devoted exclusively to this topic [33"].

The relationships between the animal picornaviruses and the plant comoviruses and between the animal togaviruses and the plant tobamoviruses, bromoviruses, and alfalfa mosaic virus were first described several years ago [34] [35] [36] [37] . These relationships involve similarities in genome organization and clear amino-acid-sequence homology in some, but not all, of the encoded proteins, suggesting quite strongly that the 'picorna-like plant viruses' and Picornaviridae are evolutionarily related to one another as are 'Sindbis-like plant viruses' and Togaviridae.

The genomes of Sindbis virus, the type alphavirus, family Togaviridae, and tobacco mosaic virus (TMV), are compared in figure 3 . The relationship between Sindbis virus and TMV is straightforward. In each case the replicase genes are translated from the genomic RNA, readthrough of a termination codon is required to translate the RNA polymerase, and there are long stretches of clear amino acid sequence homology in three genes. The structural proteins are translated from a subgenomic mRNA and appear to be unrelated to one another. TMV is a rodshaped virus, and Sindbis virus is enveloped. The recent demonstration that the nucleocapsid protein of Sindbis virus is structurally related to chymotrypsin is an exciting development and suggests that this capsid may have been obtained by recombination from a cellular protease (Rossman, personal communication). The RNA replication signals in the 3' non-translated region (NTR) are also different in the two viruses. The Sindbis virus 3' NTR contains a number of sequence elements, including repeated sequences and a 19 nucleotide conserved element, that are believed to function as linear elements, and which terminate in a poly(A) tract. TMV contains a 3'terminal nucleotide sequence capable of forming a tertiary structure similar to that of tRNA that is recognized by an aminoacyl tRNA synthetase. These tRNA-like sequences in TMV and a number of other plant virus RNAs are thought to be important for RNA replication and are known to be essential for infectivity, because certain point mutations within the sequence are lethal [38] . Upstream of this structure are a number of stem and loop elements in which certain bases in the loops can hydrogen bond with sequences adjacent to the stems, forming 'pseudo- knots' [39] . Recently, it has been shown that the 3' NTR of TMV (containing both the tRNA and pseudoknot domains) can substitute functionally for a poly(A) tail in the expression of heterologous mRNAs in both plant and animal cells [40.°] . Some synergistic interaction with other viral elements is suggested by the fact that maximal expression occurs in certain constructs in which the 5' NTR and 3' NTR of TMV flank the heterologous reporter gene [40".] .

Members of at least five other groups of plant viruses belong to the 'Sindbis-like superfamily', although in some cases the RNA genomes are divided into multiple inde~ pendent RNA segments, as in the case of bromoviruses and hordeiviruses (Fig. 3) . Barley stripe mosaic virus (BSMV), a hordeivirus, has now been sequenced and found to have a number of interesting properties that illustrate the evolution of these viruses [41. ]. In BSMV, the three domains of sequence similarity to Sindbis virus are encoded on separate gene segments. Remarkably, homologs to the helicase domain found in the Sindbis virus non-structural protein, nsP2, are present on BSMV RNA segments, although only one is required for RNA replication [41-] . Two notable characteristics of hordeiviruses are: that the 3' NT~ combine several motifs from 3' to 5', that is, a tRNA structure, a number of psuedoknots and finally a poly (A) tract adjacent to the end of the ORF (reviewed in [42] ); and that a certain amount of plasticity is apparent in the polymerase, encoded on RNA 3.

Three different forms of RNA 3 were originally identified in different strains of BSMV, although it has now been shown that all three can occur together in certain strains. Form IV contains a deletion in an essential polymerase domain, and is therefore defective and rapidly removed from the population. However, both Form III and Form II, which contain a tandem duplication of 350--370 nucleotides at the amino-temlinus of the ORF for the polymerase, often occur [41. ]. Comparable variation is not seen within the second ORF of RNA 3 which encodes a cysteine-rich polypeptide necessary for systemic infection of plants. Thus, a remarkable series of recombination events were involved in the production of the BSMV genome, and the virus may still be evolving toward an optimal organization. translate two polypeptides from overlapping ORFs in the small segment by initiation at two different AUG codons; and phlebovirus and uukuvirus small segments have two non-overlapping ORFs that are read in opposite orientations (i.e. one is translated from a genome sense mRNA and one from an antigenome sense mRNA) [43] . This strategy, termed 'ambisense', is also found in the bipartite Arenaviridae. Virions of Bunyaviridae with ambisense small segments contain small segment RNA of both polarities (although not in equal concentrations), but their medium segment RNA is all negative strand. At the 3' termini of the three Bunyaviridae RNA segments there are short nucleotide sequences (10-12 residues in length) that are conserved within a given genus and that are complementary to conserved sequences at the 5' termini such that the RNAs can form panhandles. It has been suggested that these sequences are promoter elements for the initiation of RNA synthesis, and that in a mixed infection only Bunyaviridae with identical or nearly identical termini would be replicated by the same polymerase. This would imply that only members of the same genus can exchange genome segments. Nine out of 10 of the terminal nucleotides are identical for phleboviruses and uukuviruses. In addition, sequence similarities have been found in the amino acid sequences of the N proteins and GI and G2 glycoproteins of these two genera. Surprisingly, the terminal sequence for the tenuiviruses is 90% identical to uukuviruses, suggesting that these two groups are very closely related, but the termini of tospoviruses bear no relationship to the termini of any other bunyavirus. All of these characteristics lead us to suggest that, despite their ambisense strategy, the tospoviruses have diverged significantly from other Bunyaviridae, in contrast to tenuiviruses, which may be very closely related to Bunyaviridae, despite their distinct morphology. Furthermore, we propose that the viral phleboviruses and uukuviruses belong to a single genus.

During the past year, a number of interesting insights into the interrelationships among RNA viruses have been obtained. There appear to be many similarities among the replication strategies of seemingly diverse viruses; representatives of plant virus groups and animal virus families often share certain genes and features of genome organization while differing in other aspects of their replication.

As more viral genomic sequences are obtained, a clearer picture of the RNA virus phylogenetic tree emerges, and it appears that RNA viruses extant today have evolved from a small number of protoviruses. In addition to divergent evolution, RNA viruses also evolve by recombination and the acquisition of new genes either from other viruses or from their hosts. The net result of such a reshuffling of entire viral genes is a form of modular evolution, where segments of the genome are transferred as a module.

In recent years, a number of new strategies of viral gene expression have been discovered. Because most RNA virus genomes are small, perhaps limited by the inherent error frequency caused by RNA replication without proofreading, RNA viruses are quite efficient, and have evolved a range of mechanisms by which to expand the available coding capacity and differentially regulate their gene products. These include translating more than one reading frame starting at different initiation sites, differential splicing of mRNAs during transcription, and production of multiple mRNAs in which the reading frame has been shifted by the insertion of non-templated nucleotides. Furthermore, in at least one system subgenomic mRNAs can replicate independently, and this replication may be an additional important mechanism for regulating the amounts of individual gene products.

Sequences of 50 RNA-dependent RNA polymerases from 43 plus-strand viruses and seven double-strand viruses have been aligned, and phylogenetic trees constructed

Two Related Superfamifies of Putative Helicases Involved in Replication, Repair, and Expression of DNA and RNA Genomes

Evolution of RNA Viruses

A Conserved nTP-Binding Motif in Putative Helicases

Processing the Nonstructural Proteins of Sindbis Virus: Nonstructural Proteinase is in the C-Terminal Half of nsP2 and Functions Both in Cts and Tranx

Sequence Alignments of Picornaviral Capsid Proteins

Biochemical Identification of Viruses Causing the 1981 Outbreaks of Foot-and-Mouth Disease in the UK. Nature

Evolution of Enterovirus Type 70: Ollgonucleotide Mapping Analysis of RNA Genome

Genetic Stability of Ross River Virus During Epidemic Spread in Nonimmune Humans

Structure of the Ockelbo Virus Genome and its Relationship to other Sindbis Viruses. Virologl

Complete sequence of the genome of Ockelbo virus, a strain of Sindhis virus which causes epidemic disease in northern Europe. From sequence comparisons it was found that Ockelbo virus must have been transferred to northern

Two mRNAs That Differ by Two Nontemplated Nucleotides Encode the Amino Coterminal Proteins P and V of the Paramyxovirus SV5

Measles Virus Editing Provides an Additional Cysteine-rich

The P gene of Human Parainfluenza Virus Type 1 Encodes P and C Proteins But not a Cysteine-rich V Protein

Demonstration that PIV-1 does not need a cysteine-rich V protein, and that the addition of non-templated G residues is not essential for all paramyxo-viruses

The P Gene of Bovine • • Parainfluenza Virus 3 Expresses all Three Reading Frames From a Single mRNA Editing Site

A comprehensive look at all of the mRNA transcripts and translated protein products of the P gene of PIV-3, illustrating that all three reading frames are used over a stretch of more than 300 nucleotides. This may be the first gene sequenced in which all three frames are used

Two Nontemplated • Nucleotide Additions are Required to Generate the P mRNA of Parainfluenza Virus Type 2 Since the

To study the structure of the mRNAs from the PV gene, both genomic and mRNA sequences were amplified by PCR and sequenced

Sequence Analysis of the Phosphoprotein (P) Genes of Human Parainlluenza Type 4A and 4B Viruses and RNA Editing at Transcript of the P Genes: the Number of G Residues Added is Imprecise

RNA Editing by G-Nucleotide In-. sertion in Mumps Virus P-Gene mRNA Transcripts

Demonstration that the unedited transcript encodes the cysteine.rich V protein and that variable numbers of inserted G's give transcripts for two other products

of Human Respiratory Syncytial Virus and Predicted Phylogeny of Nonsegmented Negative-strand Viruses. Virolog)

The complete sequence of the polymerase gene of RSV is presented and compared ,~ith the L genes of Rhabdoviridae and Paramvxoviridae. Comprehensive phylogenetic trees are shown

Coronavirus Subgenomic Minus-strand RNAs and the Potential for mRNA Replicons

Coronavirus: Organization, Replication and Expression of Genome

Coronavirus Genome: Prediction of Putative Functional Domains in the Non.structural Polyprotein by Comparative Amino Acid Sequence Analysis

Identification of a Domain Required for Autoproteolyric Cleavage of Murine Coronavirus Gene A Polyprotein

The E3 Protein of Bovine Coronavirus is a Receptor-destroying Enzyme With Acetyl-esterase Activity

Coronavirus Leader-RNA-primed Transcription: an Alternative Mechanism to RNA Splicing

SAWlCKI DIz Coronavirus Transcription: Subgenomic Mouse Hepatitis Virus Replicative Intermediates Function in RNA Synthesis

Bovine Coronavirus • mRNA Replication Continues Throughout Persistent Infection in Cell Culture

Minus-strand Copies of • . Replicating Coronavirus mRNAs Contain Antileaders

Direct sequencing of minus strands corresponding to coronavirus mRNAs to show that they are exactly complementa D, to the mRNAs

• • Comparison of the Genome Organization of Toro-and Coronaviruses: Evidence for Two Nonhomologous RNA Recombination Events During Berne Virus Evolution

Sequence analysis to suggest that coronaviruses and toroviruses are related and that recombination has played a role in their evolution. 33. MARCY BWJ (ED): Related Viruses of Plants and Animals

A collection of brief reviews that summarizes the know facts about related plant and animal viruses

Striking Similarities in Amino Acid Sequence Among Nonstructural Proteins Encoded by RNA Viruses that have Dissimilar Genomic Organization

Sindbis Virus Proteins nsPl and nsP2 Contain Homology to Nonstructural Proteins from Several RNA Plant Viruses

Homologous Sequences in Non-structural Proteins from Cowpea Mosaic Virus and Picornavirus

Similarity in Gene Organization and Homology Between Protein of Animal Picornaviruses and a Plant Comovirus Suggest Common Ancestry of these Virus Families

Mutational Analysis of the Sequence and Structural Requirements in Brome Mosaic Virus RNA for Minus Strand Promoter Activity

BOSCH 12 The Spatial Folding of the 3' Noncoding Region of Aminoacylatable Plant Viral RNAs

RNA Pseudoknot Domain of Tobacco • • Mosaic Virus can Functionally Substitute for a Poly(A) Tail in Plant and Animal Cells

An extremely interesting dissection of the functions of various domains in the 3'-NTR of a plant virus

Identifica-• tion of Barley Stripe Mosaic Virus Genes Involved in Viral RNA Replication and Systemic Movement

An excellent synthesis of all that is known about BSMV, the functions of the various genes, and the heterogeneity of RNA 3 within a population of get-tomes

Hordeivirus Relationships and Genome Organization

Uukuniemi Virus Segment S: Ambisense Coding Strategy, Packaging of Complementary Strands into Virions and Homology to Members of the Genus Phlebovirus

The S RNA • • Segment of Tomato Spotted Wilt Virus has an Ambisense Character

Ambisense Seg-• merit 4 of Rice Stripe Virus: Possible Evolutionary Relationship with Phleboviruses and Uukuviruses (Bunyavtridae)

Sequence analysis to show that the terminal sequences are self-complementary and very similar to those of phleboviruses and uukuviruses

First description of a virus with two ambisense segments

Division of Biology

Work of the authors is supported by grants A110793 and A120612 from the National Institutes of HeAth and by grant DMB 9104054 from the National Science Foundation.