key: cord-266230-ia04jc9j
authors: Rott, M. E.; Tremaine, J. H.; Rochon, D. M.
title: Comparison of the 5′ and 3′ termini of tomato ringspot virus RNA1 and RNA2: Evidence for RNA recombination
date: 1991-11-30
journal: Virology
DOI: 10.1016/0042-6822(91)90801-h
sha: 
doc_id: 266230
cord_uid: ia04jc9j

Abstract The sequences of the 5′ terminal 1140 and 3′ terminal 1546 nt of tomato ringspot virus (TomRSV) RNA1 have been determined. These sequences share a high degree of nucleotide sequence similarity with the previously determined TomRSV RNA2 sequence. Eighty-eight percent of the 5' terminal 907 nt of TomRSV RNA1 and RNA2 contain identical nucleotide residues; the first 459 nt are identical at all positions, whereas the next 447 nt are identical at only 75.8% of the nucleotide positions. The region of similarity includes not only the 5' nontranslated leader but also sequence probably encoding polyproteins. The 3′ terminal 1533 nt of TomRSV RNA1 and RNA2 are identical and are noncoding. The sequences common to RNA1 and RNA2 account for almost 35% of the total genomic sequence. It is possible that the similar sequences at both ends of TomRSV RNA1 and RNA2 are a result of recombination between these two genomic RNA components.

o 1991 Academic press. inc.

involving the exchange of genetic information between genomic RNA molecules of RNA viruses, appears to play a number of important functions during the evolution and life cycle of RNA viruses (for reviews see 7, 2). The importance of nonhomologous RNA recombination in the generation of evolutionary diversity can be seen in the modularity of viralencoded genes. In addition, many RNA viruses produce defective-interfering RNAs which may be a result of recombination.

Homologous RNA recombination has been observed among the picornaviruses (3), coronaviruses (4), and bromoviruses (5) and is suspected to occur in the tobraviruses (6). It has been suggested that homologous recombination in the picornaviruses and coronaviruses is important for repairing defective genomes. In addition, coronaviruses undergo site-specific recombination to express downstream genes from leader-primed subgenomic transcripts (see 7).

Tomato ringspot virus (TomRSV) is a member of the nepovirus group (8) . Nepoviruses consist of 28-nm spherical particles composed of 60 copies of a single coat protein species and two separately encapsidated genomic RNAcomponents.

Nepoviruses share similarities in genomic structure and translational strategies with the plant coma-and potyviruses as well as the animal picornaviruses (9). Previously, we reported that the 3' termini of TomRSV RNA1 and RNA2 share an ' Sequence data from this article have been deposited with the EMBUGenbank Data Libraries under Accession Numbers M27935 and M73822.

2 To whom requests for reprints should be addressed. extended region of nucleotide sequence similarity, as determined by restriction enzyme cleavage maps and hybridization analysis (10). We report that extensive nucleotide sequence similarity also exists between the 5'termini of RNA1 and RNA2. The possibility that these repeated sequences may facilitate replication of TomRSV RNA, perhaps through recombination, will be discussed.

Two cDNA clones derived from TomRSV RNA1 (see Fig. 1 ) were used to determine the 5' and 3' terminal sequences of TomRSV RNAl. Clone J27, which has been previously described (IO), was used to sequence the 3' terminal 1546 nt of TomRSV RNA1 and clone 25P6 was used to sequence the 5' terminal 1108 nt. 25P6 was obtained in essentially the same manner as J27 except that random priming was used for firststrand cDNA synthesis. Subcloning, sequencing, sequence assembly, and analysis were essentially as described previously (1 I).

Clone 25P6 was found to hybridize to both TomRSV RNA1 and RNA2 in Northern hybridization studies (12) (data not shown). However, the restriction enzyme map of 25P6 matched that of the RNA1 -specific clone B54 (10) ( Fig. 1 ) but was distinct from that of the RNA2-specific clone 035 (see 7 7). To confirm that 25P6 was derived from RNAl, the region 5' to the HindIll site of B54 was partially sequenced in one direction and was found to be identical to the corresponding region obtained from 25P6 (data not shown). The 5' terminal sequence not encoded by 25P6 was determined by dideoxynucleotide sequence analysis using TomRSV RNA as a template and a specific oligonucleo-

"PI3 tide primer (S'GCCTICGATGGCAACC 3') complementary to nucleotide positions 115-l 30 as described previously (I I). The sequencing gel obtained from this experiment terminated with two strong stop points in all four lanes which likely correspond to the first and second nucleotides of TomRSV RNA1 (each denoted N in Fig. ZA) .

Nucleotide sequence comparison of the 5'termini of TomRSV RNA1 and RNA2 revealed that within the first 907 nt, 88.8% of the nucleotide positions are identical. The first 459 nt of TomRSV RNA1 and RNA2 are identical. This region includes the 5' noncoding regions of RNA1 and RNA2 as well as two potential in-frame translation initiation sites at AUGTB and AUG,,, . Beginning from the first potential in-frame initiation site at AUG,B, the N-terminal regions of the TomRSV RNA1 and RNA2 polyproteins are identical for the first 132 amino acids, and of the next 145 amino acid residues 75.3% of the positions are identical (Fig. 3A ). It is perhaps significant that the second in-frame initiation site at AUG,, occurs shortly after the point where the homology between the RNA1 and the RNA2 polyproteins becomes less than perfect. As described in a previous paper (1 I), AUG,,, is in a better Kozak context for the initiation of translation than AUG,, (13). Since it is unknown whether AUG,, and/or AUG,,, act as initiation sites for translation, it cannot be said whether it is the conservation of amino acid or nucleotide sequence between AUG,, and AUG,,, on RNA1 and RNA2 which is most significant. The deduced amino acid sequence of the regions beginning shortly after AUG,,, on RNA1 and RNA2 could be aligned with the deduced amino acid sequences encoded by the 5' terminal region of RNA1 of the nepoviruses tomato blackring (TBRV) and grapevine chrome mosaic (GCMV) (14, 15) ( Fig. 3B ). The fact that these regions of similarity are present only at the N-termini of the TBRV and GCMV RNAl-encoded polyproteins but are present at the N-termini of both TomRSV RNA1 -and RNA2encoded pofyproteins suggests that a large portion of coding and noncoding sequences at the 5' terminus of TomRSV RNA1 have been duplicated and are now present at the 5' terrnini of both RNA1 and RNA2. The function of this coding region in RNA1 is unknown, however, it has been suggested that it may have a role in proteofytic prmessing of the viral-encoded polyprotein (75). It is interesting that in vitro translation studies of cherry feafroll virus (CLRV) RNA2 (16) another nepovirus with a large RNA2 component ( 17) resulted in proteolytic processing of the RNA2 polyprotein in the absence of RNAlencoded protease. It is possible that a subgroup of nepoviruses with large RNA2 components, which would include TomRSV and CLRV (181, may encode another protease on RNA2 which is involved in proteolytic processing. However, it is possibtethirt proteotytic processing of the CLRV RNA2 polyprotein may not be due to a specific viral-encoded protease.

The 3' noncoding regions of TomfZ@V RNA1 and RNA2 are almost identical for 1533 nt (excluding the 3 poly(A) tail sequence) with only three nucleotide differences at positions 703, 720, and 770 as shown in Fig.  2 B. These sequences are preceded by 13 and 17 nt of noncoding sequence which are unique to RNA1 and RNA2, respectively (UAAAUCCUCWUUG and tJ&JG-UUGGCUUCCUGAA, underlined nuclsotides indicate stop codons for the large ORFs of RNA1 and RNA2, respectively). When we first reportedthesequence similarity at the 3'termini of TomRSV RPVAl and RNA2 ( ICI), we proposed that extensive 3' terminel identity between RNA1 and RNA2 may be characteristic for other nepoviruses with large RNA2 components. This has been confirmed for the nepovirus CLRV (19) ** ****** * * **** ******** * ** ******** *********** * x** AALTAPWLLCSYKSGVSSPPpPMTQRQQFAAIKRRLVQKGQQ~IRELIRARKAAKYAlbFAARKKA AAITAPWLLRPCK-GEAPPPPPLTQRQQFAALKKRLAVKGQQIIREHIRARKAAKYAAIAKAKKA ** ** ** ************ ****** *******X******R****** ** x**** AAVAAQKARAEAPRLAAQKAAIAKILRDRQLVSLPPPPPPS~RL~~A~SKSASLQRL~FH AALfLAVKAAQEAPRLAAQKISKILRDRDRDVAALPPPPPPS~RL~EABLASKAESLRRLKAFK **** ** *** ** RANRVRPVLNNSFPSPP TFSRVRPALNTSFPPPP shows 3' noncoding sequence identity between RNA1

In summary, the total amount of duplicated seand RNA2 for a length which is similar to that found in quences between TomRSV RNA1 and RNA2 as well as TomRSV. However, very little sequence similarity is de-within RNA2 (see 11) accounts for almost 35% of the tectabJe between the 3' noncoding regions of TomRSV total genomic sequence. The extensive amount of nuand CLRV (19). The 3' noncoding regions of the two cleotide sequence identity at the 5' and 3' termini of RNA components for several other nepoviruses are TomRSV RNA1 and RNA2 may be required for recognialso identical but much shorter (less than 300 nt) (74, tion by a highly selective replicsee. It is slso possible 15). Extensive sequence similarity at the 3'termini has that RNA recombination is responsible for maintaining also been reported for members of the tobravirus nucleotide sequence identity at the 5' and 3'termini of group (20, 21) and includes both potential coding se-RNA1 and RNA2. RNA recombination has been postuquences and noncoding sequences. lated to explain the duplication of 820 nt at the 3'ter- FIG. 2. Nucleotide sequence and deduced amino acid sequence of the 5'(A) and 3'(B) regions of TomRSV RNA1 (A) The first two nuolaotides which could not be determined from the sequencing gel are each represented by an N. The second in-frame AUG which is a potential initiation site for translation is underlined. Nucleotides are numbered on the left beginning at the first N. Amino acids are numbered on the right snd begin at the first M. Numbering in (B) begins at the termination codon (WA) for the RNAl-encoded long ORF as determined from clone J.27. Three nucleotide substitutions in the 3' noncoding region of RNA2 (I 7) compared to RNA1 are shown below the sequence of RNA1 mini of RNA1 and RNA2 of the tobravirus tobacco rattle virus (TRV) strain PLB (6). However, after several passages of a pseudorecombinant consisting of RNA1 from TRV strain TCM and RNA2 from strain PLB, which differ at their 3'termini in 39 of 820 nt, a recombinant RNA molecule was not detected. Not only was RNA recombination in this system not detected experimentally, but the viability of a pseudorecombinant consisting of heterologous 3' termini in RNA1 and RNA2 suggests that in this system precise 3'terminal sequences in RNA1 and RNA2 are not essential for replication. It has been suggested that the high frequency of homologous recombination in the animal picornaviruses is important for removing deleterious mutations introduced by the poor fidelity of the RNA replicase (22) . Since it has been suggested that the picornavirus replicative machinery may not function efficiently in trans, nondefective genes, located on different RNA molecules which possess errors in genes involved in replication, can only be utilized after genetic recombination with an RNA molecule which encodes functional replicative genes (23, 24) . It is possible that, in TomRSV, replication begins in cis with RNA1 and that vans replication of RNA2 occurs only following disassociation and reassociation of the initial negative-strand transcript with the corresponding region in RNA2. A similar mechanism involving recombination could account for the sequence conservation observed between the 5' termini of RNA1 and RNA2. Such a mechanism has recently been proposed for leader-primed generation of subgenomic RNAs in coronaviruses (see 7). The size of the duplicated sequences in TomRSV may be the minimum required to facilitate efficient replication through RNA recombination between RNA1 and RNA2. Alternatively, the entire length may not be required for recombination but may serve other important functions in addition to a postulated role in recombination. We are planning further experiments to determine the biological significance of the repeated sequences between TomRSV RNA1 and RNA2 and whether they are in fact involved in RNA recombination during replication.

The technical assistance of L. Lee is gratefully acknowledged. This work was partially supported by an NSERC operating grant awarded to J. H. Tremaine.

/n "The Molecular Biology of the Positive Strand RNA Viruses

Cold Spring Harbor Symp. Quart Biol

Descriptions of Plant Viruses

Nematode Vectors of Plant Viruses

V///rh hf