The Ribosomal Shunt Translation Strategy of Caulif
http://www.100md.com
病菌学杂志 2006年第8期
Division of Plant Sciences, 108 Waters Hall
Department of Computer Science, 201 Engineering Bldg. West, University of Missouri, Columbia, Missouri 65211
Plant Biology Dept., The Samuel Roberts Noble Foundation, Ardmore, Oklahoma 73402
Dept. of Biochemistry, Dakota Wesleyan University, Mitchell, South Dakota 57301
Dept. of Microbiology, Miami University, Oxford, Ohio 45056
ABSTRACT
We have screened portions of the large intergenic region of the Cauliflower mosaic virus (CaMV) genome for promoter activity in baker's yeast (Saccharomyces cerevisiae) and have identified an element that contributes to promoter activity in yeast but has negligible activity in plant cells when expressed in an agroinfiltration assay. A search of the yeast genome sequence revealed that the CaMV element had sequence similarity with the R region of the long terminal repeat (LTR) of the yeast Ty1 retrotransposon, with significant statistical confidence. In plants, the same CaMV sequence has been shown to have an essential role in the ribosomal shunt mechanism of translation, as it forms the base of the right arm of the stem-loop structure that is required for the ribosomal shunt. Since the left arm of the stem-loop structure must represent an imperfect reverse copy of the right arm, we propose that the ribosomal shunt has evolved from a pair of LTRs that have become incorporated end to end into the CaMV genome.
INTRODUCTION
Cauliflower mosaic virus (CaMV), the type member of the caulimovirus group, is a pararetrovirus; it replicates by reverse transcription, but replication does not require the integration of the CaMV genome into host chromosomes. The lack of an integration step is one feature that distinguishes the replication strategy of CaMV from that of other plant retrotransposons or the animal retroviruses, and this is reflected in the structure of the CaMV genome (50). The CaMV genome is composed of circular, double-stranded DNA that is approximately 8,000 bp in size (Fig. 1A). The replication intermediate that serves as the template for reverse transcription is the CaMV 35S RNA (Fig. 1A) (47). The 35S RNA is formed when the host RNA polymerase II initiates transcription at nucleotide (nt) position 7435. A transcript termination signal at nucleotide 7613 is initially bypassed such that the RNA Pol II synthesizes viral RNA over the full length of the viral genome until it reaches the transcript termination signal for a second time, at which point transcription stops. Consequently, the first 179 bp of the 5' end of the 35S RNA are reiterated on the 3' end of the 35S RNA (50).
The terminal redundancy in the 35S RNA is required for the template switch that occurs during the synthesis of the first strand of viral DNA by the CaMV reverse transcriptase. In this capacity, the terminal redundancies of the 35S RNA fulfill the function of the R region present in the long terminal repeats (LTRs) of retrotransposons and retroviruses. The structure and function of LTRs are most easily visualized in the retrotransposon forms that are integrated into host chromosomes (Fig. 1B). For example, the integrated form of Ty1 consists of two complete LTRs that flank the open reading frames (ORFs), and each of the LTRs is composed of three distinct regions: U3, R, and U5 (4, 5, 13). Transcription is initiated within the leftward LTR at the junction of U3 and R and is terminated at the junction of R and U5 in the rightward LTR (Fig. 1B). The resultant mRNA has a copy of R followed by U5 at its 5' end, whereas the 3' end consists of U3 ending with the R region (4, 13). Consequently, in comparing the terminal redundancies of CaMV to the LTRs of the retrotransposons, both groups have clearly defined R regions, whereas the U3 and U5 regions can be only identified in the LTRs of the retrotransposons.
A second feature that is a hallmark of the caulimoviruses is the ribosomal shunt mechanism of translation (51). The ribosomal shunt has undoubtedly evolved over time to compensate for the complexity and length of the leader sequence of the 35S RNA. This leader sequence is approximately 604 bp in length, and it contains up to nine short ORFs (sORFs), depending on the strain, that differ in size from 9 to 102 nucleotides (Fig. 1C). Such a leader could be expected to be a formidable impediment to translation in any eukaryotic organism, as eukaryotic ribosomes generally initiate translation at the first AUG that is found in a favorable context and encountered nearest the 5' end of the transcript (29). In the majority of eukaryotic transcripts, ribosomes are unable to reinitiate the transcription of a second cistron. However, extensive studies have shown that the complexity of the 35S RNA leader sequence is bypassed through the formation of a large stem-loop structure (15, 20), which brings into close proximity the essential elements of the ribosomal shunt (the shunt donor and acceptor sites). In translation of the 35S RNA, host ribosomes enter the 5' end of the 35S RNA and scan a short distance until they reach the shunt donor site (between sORFs A and B) (Fig. 1C). At this point in the sequence, ribosomes bypass the central region of the leader and land at the shunt acceptor site within sORF F (10, 16, 48). The shunt mechanism allows ribosomes to reach the start codon of gene VII. The subsequent expression of genes I through V on the 35S RNA involves alternative mechanisms, including transactivation of the translation of genes I through V by the CaMV gene VI product (18, 44) and splicing (14, 28).
In the present study, we developed a system using Saccharomyces cerevisiae to probe the expression strategy of CaMV. S. cerevisiae has proven useful for studying the transcription and replication of positive-sense plant viruses, such as Brome mosaic virus and Tomato bushy stunt virus (39, 42), and genome-wide screens have revealed yeast (Saccharomyces cerevisiae) proteins that are involved in the replication of these viruses (30, 43). Previous studies have also shown that there is a degree of conservation between the transcriptional machinery of yeasts and plants. For example, the CaMV 35S promoter is active in yeast species (21) and the CaMV 35S terminator sequence also can function as a terminator in yeast, although the sequence elements that are responsible for termination are different from those that are active in plants (23). Furthermore, the CaMV translational transactivator protein has limited activity in yeast (56).
Our study has revealed the presence of a cryptic yeast promoter within the CaMV genomic region that corresponds to a portion of the 35S RNA leader sequence. A BLAST search of the yeast genome showed that this CaMV promoter element was similar to the R region of the LTR of the yeast Ty-1 retrotransposon. In plants, the same CaMV sequence has been shown to have an essential role in the ribosomal shunt mechanism of translation, as it forms the base of the right arm of the stem-loop. Since the left arm of the stem-loop must represent an imperfect reverse copy of the right arm, we propose that the ribosomal shunt has evolved from a pair of LTRs that have become incorporated end to end into the CaMV genome.
MATERIALS AND METHODS
Construction of CaMV-CAT fusion constructs for expression in yeast. The plasmid pMono-CAT was constructed by inserting the chloramphenicol acetyltransferase (CAT) gene from pCM-1 (Pharmacia, Piscataway, NJ) into the yeast shuttle vector pYES2 (Invitrogen, Carlsbad, CA). To make CaMV-CAT fusions, the DNA sequence from the 35S RNA transcriptional start site (nucleotide 7433) to an XhoI site within gene II (nucleotide 1642) of CaMV strain D4 (55) was amplified by PCR to create a DNA segment with flanking KpnI and XhoI sites for insertion into pYES2. The CAT gene was inserted in frame into the CaMV sequence through the introduction of flanking restriction enzyme sites, and CaMV sequences downstream from CAT were eliminated. Point mutations were introduced into the CaMV and CAT sequences by overlapping PCR. Nucleotide sequences of plasmid inserts were determined to ensure that the CAT gene had been inserted in frame. The bicistronic -glucuronidase (GUS)/CAT plasmids were all based on plasmid pJB79, which was constructed by inserting a Hind III-BamHI fragment containing the GUS gene into pJS163 (a plasmid in which the CAT gene had been inserted into the XhoI site of the yeast vector, pYES2) (56).
Construction of plasmids for gene expression studies in whole leaves. To evaluate the strength of the cryptic yeast promoter in whole leaves, GUS reporter constructs were inserted into an Agrobacterium tumefaciens binary vector, pKYLX7 (53). To create pKYGUS, the GUS gene was modified by PCR to be flanked by KpnI and XhoI sites, which facilitated its cloning into KpnI/XhoI sites engineered into pKYLX7. pCMS129 was created by replacing the 35S promoter, which was present on an EcoRI/HindIII DNA fragment, with a CaMV segment from nucleotides 7672 to 131. pKYMdg was created by digesting pCMS129 with EcoRI and HindIII, filling in the sticky ends, and religating the plasmid. Triparental matings and agroinfiltration were performed as described previously by Palanichelvam and Schoelz (41). Leaf samples were evaluated for GUS expression 4 days after infiltration.
Recombinant DNA techniques and yeast transformation. Restriction enzymes and T4 DNA ligase were purchased from Promega Corp. (Madison, WI) or New England Biolabs (Beverly, MA). DNA ligation, plasmid transformation, and plasmid purification were performed according to the procedure of Maniatis et al. (33). Plasmids were propagated in Escherichia coli strain JM101 (38) grown in double yeast tryptone broth containing 100 μg of ampicillin per milliliter. Purified plasmid DNA from E. coli was transformed into S. cerevisiae strain JC746Dip (MATa/MATa ura3/ura3 his3/his3 leu2/leu2 trp/trp) by the lithium acetate method (24). Yeast transformants were grown in synthetic dextrose medium (0.67% yeast nitrogen base without amino acids, supplemented with auxotrophic requirements and 2% raffinose) lacking uracil.
Detection of the CAT mRNA in yeast by Northern blot analysis. Yeast was grown in minimal medium to an optical density at 600 nm of 1.6 to 1.9, and total RNA was isolated by the glass bead method (2) (type IV glass beads, 250 to 300 microns in diameter; Sigma, St. Louis, MO). Equal concentrations of RNA were denatured using glyoxal, sodium phosphate (pH 7.0), and dimethyl sulfoxide and run on a 1% agarose gel in the presence of 0.01 M sodium phosphate (57). Northern hybridization was performed according to the procedure of Maniatis et al. (33) by using a 32P-labeled CAT DNA probe.
Measurement of CAT protein levels in yeast extracts by CAT ELISA. The induction of the GAL1 promoter has been described previously by Schneider and Guarente (54). Yeast cells (40 ml at an optical density at 600 nm of 1.6 to 1.9) were pelleted by centrifugation, and the pellet was washed twice with distilled water. The pellet was resuspended in 200 μl of 0.25 M Tris-HCl (pH 7.8), the cells were sonicated for a total of 120 s (four times for 30 s each, with incubation on ice for 1 min between sonications), and cell debris were removed by centrifugation. The concentration of total protein was determined by using a modified Lowry protein assay (Bio-Rad, Hercules, CA), and the amount of CAT protein was measured using a CAT enzyme-linked immunosorbent assay (ELISA) kit (Boehringer Mannheim, Indianapolis, IN). CAT and GUS levels were measured by at least three tests for each plasmid, with three replicates per test.
Nucleotide sequence analysis. Nucleotide sequences of CaMV strain D4 present in clone pCMS119 (from nucleotides 7949 to 131) were determined at the DNA Sequencing Core Facility of the University of Missouri, Columbia. The genome size of 8,034 bp for D4 was determined through a comparison of the partial D4 nucleotide sequence to the complete CM1841 nucleotide sequence (17). RNA secondary structure for the CaMV strain D4 leader sequence was generated by using the mFOLD program (63). tRNA primer binding sites for the putative left LTR were found in the tRNAscan-SE genomic tRNA database (32). Multiple sequence alignments were performed using Clustal W (58).
To assess the statistical significance for the homologous relationship between the R region in yeast LTR and CaMV sequences, we performed a sequence-profile analysis using HMMER (11), which applies profile-hidden Markov models for the statistical descriptions of a sequence family's consensus (60). We downloaded 297 Ty1 and Ty2 LTRs from the yeast genome database (http://www.yeastgenome.org) (27), from which we obtained 40 full LTRs. To avoid the bias from data redundancy, we kept only one copy of LTR R in the profile data set when there were multiple identical sequences. As a result, 20 LTR-R regions were used to construct the multiple sequence alignment profile LTR-R using Clustal W (58). Eight complete CaMV genome sequences were obtained from GenBank (BBC, GenBank ID M90542.1; NY8153, M90541.1; CM1841, V00140.1; CMV-1, M90543.1; B29, X79465.1; Strasbourg, V00141.1; D/H, M10376.1; and Xinjiang, AF14064.1) as well as a CaMV D4 partial genomic sequence (DQ355155). As with the yeast LTR sequences, only one copy of a CaMV sequence was retained in the instance of multiple identical sequences. Consequently, we searched the LTR-R profile against seven CaMV sequences. The E value from HMMER was used to evaluate the significance of the homologous relationship between yeast LTR-R and CaMV R regions. We utilized the E value of 0.05 suggested by HMMER as the significance threshold (12).
Mapping the 5' end of the CAT mRNA derived from the CaMV sequence. Total RNA was extracted from yeast cells transformed with pCMS119 and used for 5'-end mapping of the CAT mRNA by using the GeneRacer Kit (Invitrogen, Carlsbad, CA). Following the dephosphorylation and decapping of the full-length mRNA, the GeneRacer RNA oligonucleotide was ligated to the 5' end and used for reverse transcription. The resulting cDNA was PCR amplified by using a GeneRacer-specific 5' primer and a gene-specific primer corresponding to the 3' end of the CAT gene. Two fragments, which were approximately the size of the CAT gene, were amplified. The two fragments were separated from each other by gel isolation and subsequently cloned by using a one-shot Topo cloning kit (Invitrogen, Carlsbad, CA). The resultant clones were submitted for sequencing at the DNA Sequencing Core Facility of the University of Missouri, Columbia to identify the transcript initiation sites.
Measurement of GUS expression in yeast cellular extracts and agroinfiltrated plant disks. After protein concentrations from yeast and plants were determined, GUS expression was measured by using a chemiluminescence GUS assay with the GUS-light kit (Tropix, Bedford, MA). GUS concentrations in experimental samples were determined by comparison with a standard curve made with purified GUS supplied by the kit.
Nucleotide sequence accession number. The sequence for CaMV strain D4 was submitted to GenBank and assigned accession number DQ355155.
RESULTS
A survey of expression competence from different points within the CaMV large intergenic region in yeast. To assess the expression of CaMV genes that were present on the 35S RNA in yeast, a CAT reporter gene was inserted into two positions within the 35S RNA leader sequence and in frame with genes VII and I (Fig. 2). All reporter-fusion constructs were placed under the control of the yeast GAL1 promoter and CYC1 terminator sequences present in the yeast plasmid pYES2. Yeast cells were transformed with each plasmid, and a CAT ELISA was used to quantitatively measure CAT expression in yeast cellular extracts. The expression of each plasmid was then compared to pMono-CAT, in which the CAT gene was expressed from the GAL1 promoter.
The expression of CAT decreased as it was inserted further within the 35S leader sequence, which likely reflected a diminishing ability of ribosomes to reinitiate translation as they migrated through the sORFs of the leader sequence. For example, CAT protein was expressed from the construct pABC-CAT to a level that was approximately 11% of that of the pMono-CAT control plasmid (Fig. 2). The upstream sORFs A, B, and C did have an inhibitory effect on expression of CAT in yeast, but their presence upstream from the CAT gene was not sufficient to completely block expression of the CAT protein. In contrast, the expression of CAT protein in the construct pF-CAT was negligible (0.4% of the pMono-CAT control). Since upstream open reading frames and extensive secondary structure are reported to inhibit the translation of mRNAs in yeast (35), we hypothesized that CAT would not be expressed from points within genes VII and I. As expected, the CAT gene was expressed to only a very low level from within gene I. Surprisingly, the highest level of CAT expression was observed with pVII-CAT. Whereas sORF F was expressed at a level of only 0.4% of the pMono-CAT control in yeast, pVII-CAT was expressed at 2.5 times the level of the monocistronic CAT control, a nearly 700-fold difference in expression between sORF F and gene VII (Fig. 2).
In the construct pVII-CAT, the CAT gene was inserted in frame into the coding sequence and so we anticipated that a CAT fusion product would be produced in these cells. To determine whether the CAT protein was in fact fused to CaMV sequences within ORF VII, we altered the CaMV and CAT start codons through PCR-directed mutagenesis. A mutation of the gene VII start codon had no effect on the level of CAT expressed from pCMS100, whereas a mutation of the CAT start codon in pCMS93 almost completely abolished CAT expression (Fig. 2). We also directly compared the size and amount of CAT protein produced in yeast that contained pVII-CAT to that produced in the pMono-CAT cells. In agreement with the CAT ELISA results, a Western blot assay indicated a slightly higher level of the CAT protein in pVII-CAT compared to that in pMono-CAT. In addition, the CAT protein produced from the pVII-CAT construct had the same molecular weight as the CAT protein produced in cells containing the pMono-CAT plasmid (data not shown). These results demonstrated that the CAT start codon is used to express CAT rather than CAT being initiated from the upstream gene VII start codon.
Sequences within sORF F and gene VII form part of a cryptic yeast promoter. The high level of expression of CAT protein from pVII-CAT could be due to a ribosomal shunt mechanism, an internal ribosomal entry site (IRES), or the presence of a cryptic yeast promoter. To investigate whether the high level of CAT expression in yeast was due to a ribosomal shunt mechanism, we deleted 239 nucleotides, from the beginning of the 35S RNA to 25 nucleotides upstream of sORF D (nucleotides 7435 to 7672), to form pCMS95. The effect of this deletion would be to abolish a significant portion of one of the arms of the mRNA stem-loop structure (10, 48). However, this deletion had no effect on CAT expression (Fig. 3), and smaller deletions of sORF A and sORF B also had no effect on CAT expression (data not shown). These results indicated that ribosomal shunting was not responsible for expression of CAT in the construct pVII-CAT, and this was confirmed in the experiments described below.
To further delineate the CaMV sequences that were responsible for CAT expression and to test the possibility of internal initiation in yeast, we created a dicistronic plasmid (Fig. 3, pJB79) in which the GUS gene was placed immediately upstream from the CAT gene. As expected, GUS was expressed to a high level, whereas the CAT protein levels expressed from pJB79 were low (4.4% of that of the monocistronic CAT control). To determine whether the CaMV sequences could act as an IRES in yeast, the CaMV sequence present in pCMS95 was inserted between the GUS and CAT genes to form pCMS99 (Fig. 3). CAT expression from pCMS99 was approximately twice the level of that from pMono-CAT and nearly 42-fold higher than that of pJB79 (Fig. 3). To further delineate the CaMV regions that were responsible for CAT expression, we deleted the CaMV sequences comprising sORFs D and E from pCMS99 to make the construct pCMS101. The CAT expression level of pCMS101 was comparable to that of pMono-CAT and almost 50% of that of pCMS99 (Fig. 3). Therefore, CaMV sequences that were responsible for CAT expression from the dicistronic construct mapped primarily to nucleotides 7949 to 131. The expression of GUS from pCMS99 and pCMS101 was equivalent to GUS expression of pJB79, which showed that the enhancement in CAT expression in these constructs was not due to an increase in transcription from the GAL1 promoter (Fig. 3).
To distinguish whether the CaMV sequences in pCMS101 functioned as a promoter or as an IRES in yeast, we deleted the yeast GAL1 promoter sequence and examined the effect of this deletion on the expression of CAT and GUS in pCMS119. Interestingly, the level of CAT protein expressed from pCMS119 was comparable to that of pCMS101 (Fig. 3), whereas GUS expression was abolished. We concluded that CAT expression was not dependent upon the GAL1 promoter and that sequences within CaMV sORF F and gene VII were part of a cryptic promoter that functioned in yeast. A further deletion of CaMV sequences from nucleotides 7949 to 7995 essentially abolished the activity of the putative CaMV promoter, as CAT protein expression from pCMS120 was comparable to that of the control plasmid pJB79 (Fig. 3). Furthermore, the addition of the GAL1 promoter to pCMS120 had no effect on the CAT protein level (Fig. 3, compare pCMS120 and pCMS121), confirming that CAT expression was driven by a cryptic promoter present in the CaMV sequences.
A comparison of the CAT transcripts produced in yeast cells provided further evidence that CaMV sequences centered around sORF F and gene VII functioned as an active promoter in yeast. Two prominent transcripts that contain CAT sequences can be seen in yeast cells transformed with pCMS101 (Fig. 4). The larger transcript, expressed from the GAL1 promoter, consisted of GUS and CAT sequences, whereas a second transcript was smaller than the pMono-CAT transcript (Fig. 4). The larger size of the CAT mRNA that was derived from the pMono-CAT plasmid could be attributed to sequences from the pYES2 polylinker within its untranslated leader sequence. Removal of the GAL1 promoter, as in constructs pCMS119 and pCMS120, abolished the larger GUS/CAT transcript (Fig. 4), demonstrating that this transcript was not responsible for the expression of CAT. In contrast, the elimination of CaMV sequences from nucleotides 7949 to 7995 abolished the smaller transcript (Fig. 4, lanes pCMS120 and pCMS121) and CAT protein expression (Fig. 3).
With the deletion of the GAL1 promoter, a new band appeared that was larger than the transcript associated with GAL1 (Fig. 4, lanes pCMS119 and pCMS120). We considered the possibility that this RNA band might originate from a cryptic promoter present in the pYES2 plasmid that was activated in the absence of the GAL1 promoter. However, the presence of this new larger transcript was not correlated with the expression of CAT. The CAT protein expression level was high in yeast that contained pCMS119 and low in yeast that contained pCMS120 (Fig. 3), yet the larger CAT RNA band was present in both of these yeast cells (Fig. 4). Taken together, the CAT protein expression studies and RNA gel blots showed that the CaMV region that comprises sORF F and gene VII drives the transcription of the CAT mRNA in yeast.
Identification of two transcript initiation sites within gene VII. The size of the CAT mRNA derived from the CaMV yeast promoter, which was detected in the Northern blot analysis of pCMS101 and pCMS119, indicated that the initiation of transcription occurs within gene VII near the CAT start codon. To precisely determine the transcript initiation site for the CAT mRNA in pCMS119, we used a GeneRacer kit. Two PCR fragments, corresponding to the 5' end of the CAT mRNA, were cloned, and their nucleotide sequence was determined. Based on the intensity of the PCR-amplified fragments and nucleotide sequence, the most likely transcript initiation site was at the AG dinucleotides 35 nt upstream from the CAT start codon (Fig. 5). A second site was located 56 nucleotides upstream of the CAT start codon at CT.
The yeast promoter located in sORF F and gene VII does not have activity in leaf mesophyll cells. To determine whether the CaMV region we identified as a yeast promoter functioned as a plant promoter, we inserted CaMV sequences from nucleotides 7672 to 131 upstream from a GUS reporter gene to create pCMS129. The promoter activity of pCMS129 was compared to that of the CaMV 35S promoter and a promoterless GUS construct (Fig. 6) in an agroinfiltration assay of Nicotiana edwardsonii leaves. GUS activity from pCMS129 was 285-fold less than the GUS activity from the 35S promoter and only slightly higher than the promoterless control plasmid (Fig. 6). These results showed that the CaMV cryptic yeast promoter had negligible activity in plant leaf tissue in an agroinfiltration assay. However, this test could not eliminate the possibility that the cryptic promoter might function in plants in the presence of one or more CaMV proteins.
The CaMV cryptic yeast promoter has sequence homology to the R region within the long terminal repeat of the yeast retrotransposon Ty1. The constructs in Fig. 3 revealed that a 213-bp CaMV sequence functioned as a promoter in yeast and that the CaMV sequence from nucleotides 7949 to 7995 was essential for promoter activity. To identify sequences in the yeast genome that have homology to this CaMV region, we searched the Saccharomyces genome database (3) with the BLASTN 2.0 program (1). This search yielded a single match between CaMV sequences from nucleotides 7952 to 7993 and a S. cerevisiae delta element (Fig. 7A, YDRCdelta3) located on chromosome IV at nucleotide coordinates 513338 to 513294 (25). YDRCdelta3 is a single copy of the LTR of Ty1. The CaMV sequence was 76% identical to the YDRCdelta3 sequence and 70% identical to the sequence of Ty912 (8), the first member of the Ty1 family to be sequenced (Fig. 7A).
In addition to the direct alignment of nucleotide sequences between sORF F and the R region of the Ty1 LTR, the position of the methionine (Met) tRNA binding site was also conserved. This sequence serves as the primer binding site for first-strand DNA synthesis in both Ty1 and the CaMV 35S RNA. In the case of the Ty1 LTR, the R region is located 41 nucleotides upstream from the Met tRNA binding site, whereas in CaMV the distance is 39 nucleotides (Fig. 7A). The Ty1 U5 region, the sequence that intervenes between the R region and the Met tRNA binding site, has diverged significantly from the corresponding CaMV sequence. However, the positioning of the Met tRNA binding sites in relation to either CaMV sORF F or the R region of the Ty1 LTR provides a second structural landmark that extends the relationship between the large intergenic region of CaMV and the LTR of Ty1. Consequently, most of the R region of the Ty1 LTR is represented in the CaMV large intergenic region, including the junction between the R and U5 regions, and this putative R region is located within the CaMV sequence the expected distance from its first-strand primer binding site.
To assess the statistical significance for the homologous relationship between the R region in the Ty1 LTR and CaMV sequences, we performed a sequence-profile analysis. We searched the profile of yeast LTR R regions against each of the six CaMV genome sequences and the CaMV D4 partial genome sequence by using HMMER (Fig. 8) (11). Our analysis demonstrated that the E values between the profiles of LTR R regions in yeast and the genomic sequence of CaMV range from 0.026 to 0.00043, which is much lower than the E value significance threshold of 0.05 suggested by HMMER (12). Among the seven CaMV sequences, the D4 strain has the most significant E value of 0.00043. Thus, the sequence profile analyses suggest that the yeast LTR R region and the R region of CaMV are homologous with significant confidence.
Within the context of the CaMV replication cycle, the CaMV sequence from nucleotides 7952 to 7993 has a function that is entirely unrelated to an LTR, as it has been identified as the ribosome shunt landing site (Fig. 7B) (48). This stable stem structure has been previously shown by mutational analysis to be essential for the ribosomal shunt (10). Figure 7B illustrates that the CaMV sequence that corresponds to the R region in the Ty1 LTR has previously been predicted to form the base of the right arm of the large stem-loop structure within the leader sequence of the 35S RNA (48).
DISCUSSION
We have identified a 42-nucleotide CaMV sequence in the large intergenic region that contributes to three distinct functions in two different organisms. The sequence had previously been shown to have an essential role in the ribosomal shunt mechanism of translation of the CaMV 35S RNA in plants; it corresponds to the base of the right arm of a large stem-loop structure that forms in the leader of the 35S RNA (15, 20). This stem-loop structure is required for the formation of the ribosomal shunt landing site (10, 16, 48) to initiate the translation from ORF VII or ORF I. In yeast, the same sequence forms part of a cryptic promoter/enhancer element. This sequence directed a high level of expression of a CAT reporter gene in yeast, but it had negligible activity in plant cells in an agroinfiltration assay. A BLAST search of the yeast genome revealed a third function for this sequence element, as it shares homology to the R region of the LTRs of the yeast retrotransposon Ty1. A search using the sequence profile built from the LTRs in yeast against the various CaMV genomic sequences indicates that the homology is statistically significant. In the discussion below, we evaluate each of these functions and present our arguments for an evolutionary relationship between the ribosomal shunt and LTRs.
A sequence element in sORF F functions as a cryptic yeast promoter/enhancer and has homology to the R region present in the LTR of the yeast retrotransposon Ty1. In our study, we considered three different mechanisms that might explain the high level of expression of CAT from the construct pVII-CAT in baker's yeast: a ribosomal shunt, an IRES, and a cryptic promoter. We eliminated the ribosomal shunt and the IRES as possible mechanisms because both would have been dependent on the yeast GAL1 promoter, and elimination of the GAL1 promoter abolished CAT expression. Instead, our mutagenesis studies and Northern analysis revealed that the CaMV sequence had promoter or enhancer activity in yeast.
A BLAST analysis of the yeast genome revealed that the CaMV promoter/enhancer element had sequence homology to the delta element YRDCdelta3, which is located on yeast chromosome IV. Delta elements are solo copies of the LTR of Ty1; they result from a recombination event that occurs between the LTRs of a full-length Ty1 element. As such, delta elements represent a site in the genome where a full-length element had once been inserted. Within the yeast genome, there are 32 full-length copies of Ty1 and 185 solo Ty1 LTRs, either complete versions of the LTR or fragments (27). The BLAST search did not reveal a match between the CaMV sequence and other Ty1 LTRs, but once the initial match was made between the CaMV sORF F sequence and YRDCdelta3, it was possible to align the CaMV sequence with the original sequence of Ty1 (8). This alignment could be extended to the primer binding site for first-strand DNA synthesis for both CaMV and Ty1. In the case of the Ty1 LTR, the R region is located 41 nucleotides upstream from the Met tRNA binding site, whereas in CaMV, the distance is 39 nucleotides.
The portion of CaMV that contributes to the cryptic promoter activity corresponds to the untranslated leader region (ULR) of the Ty1 mRNA (Fig. 7A). Transcription of Ty1 begins at nucleotide position 240 within the LTR (13), and the start codon for the gag ORF is found at nucleotide position 294; the CaMV promoter/enhancer corresponds to approximately 83% of this sequence (Fig. 7A). This portion of the Ty1 LTR has not been reported previously to have enhancer activity. However, it is well established that transcriptional enhancers can be located within ULRs. For example, the copia retroelement of Drosophila contains seven transcriptional enhancer elements within a ULR that is adjacent to its 5' LTR (34, 37). The enhancer elements in the copia ULR have been shown to significantly increase the expression of a minimal heat shock (hsp) promoter element and to bind to the Drosophila CCAAT/enhancer binding protein (DmC/EBP) (37, 62). The copia 5' LTR also enhances the expression of the minimal hsp promoter, but the highest level of expression was attained when both the LTR and ULR were placed upstream from the minimal hsp promoter (37, 62).
In addition, a transcriptional enhancer was recently found within the ULR of CaMV at a point between the 35S RNA transcriptional start site and sORF A (45). It is worthwhile to note that the CaMV 35S promoter was first characterized in 1985 (40) and it is considered to be one of the most extensively studied plant promoters (45), yet the existence of this enhancer element is just coming to light now, perhaps because its presence was masked by other more powerful enhancers upstream from the 35S promoter (45). The discovery of this enhancer in the 35S leader sequence underscores how much there is to be learned about even well-characterized regulatory elements. At a minimum, our work illustrates that sequence alterations in the Ty1 ULR have the capacity to function as a promoter/enhancer element in yeast.
A model for the evolution of the ribosomal shunt mechanism of translation in CaMV from LTRs. So why might a cryptic yeast promoter/enhancer be located within the CaMV large intergenic region It is unlikely that the promoter/enhancer function is responsible for its conservation from yeasts to plants. This sequence lacked promoter/enhancer activity in an agroinfiltration assay (Fig. 6), although we cannot eliminate the possibility that the cryptic promoter might function in plants in the presence of one or more CaMV proteins. Furthermore, the homology of the 42-nucleotide CaMV sequence to the R region in the Ty1 LTR has no discernible function as an R region for CaMV. The comparable region for CaMV, its terminal redundancy in the 35S RNA, is located 338 nucleotides upstream from the 42-nucleotide stretch that is homologous to the Ty1 R region (Fig. 7A). Nonetheless, the 42-nucleotide CaMV sequence does have an essential role in the translation of the CaMV 35S RNA in plants; it encompasses the shunt acceptor site (Fig. 7B) (10, 16, 48). Since neither of the functions attributed to the 42-nucleotide CaMV sequence in yeast are relevant to how CaMV functions in plants, we propose that the sequence has evolved from the R region of an ancestral LTR into the stem-loop structure required for the ribosomal shunt.
Once the base of the right arm of the 35S RNA stem-loop is recognized to be a remnant of an LTR, then other elements of the LTR come into focus to provide an alternate picture of the structure of the 35S RNA leader sequence (Fig. 9). Most importantly, the left arm of the stem-loop can also be recognized as the remnant of the R region of Ty1, although in an opposite orientation to the R region that is represented in the right arm (Fig. 9B). In essence, the CaMV large intergenic region can be redrawn as two LTRs in opposite orientations (Fig. 9C). The R region remnant in the left LTR extends from nucleotide 7510 to 7563. It occupies more sequence than the R remnant in the right LTR because of insertions of foreign sequences over time (Fig. 9B); these insertions contribute a larger bulge loop in the large stem-loop structure (Fig. 7B). The boundaries between U5 and R correspond to the shunt donor and acceptor sites (Fig. 9A and C).
The extreme boundaries of the U3 regions might be predicted to occur within the large stem-loop structure at the top of stem section 3 (15, 48) at nucleotides 7724 and 7790, with 66 nucleotides separating the two LTRs (Fig. 1C and 9C). Although sequences of the two LTRs might change over time, mutations in one arm must be compensated for in the other arm to retain the stem-loop structure for the ribosomal shunt mechanism (47). Indeed, Pooggin et al. (48) concluded that mutations occurred within the 35S RNA leader sequence at "a surprisingly low frequency." The requirement for a functional ribosome shunt could provide one constraint that could limit change within the 35S RNA leader sequence.
The boundary of U5 on the right LTR is defined by the methionine tRNA binding site, which is the primer binding site for first-strand CaMV DNA synthesis during reverse transcription (Fig. 9C) (50). An inspection of the nucleotide sequence which is upstream from the left LTR reveals one attractive candidate for an additional tRNA primer binding site, as the 3' end of a valine tRNA from S. cerevisiae provides a reasonably good fit. Interestingly, the 3' end of this primer binding site is adjacent to the transcriptional start site of the 35S RNA at nucleotide 7435. The U5 region of the left LTR is a little larger than its right counterpart because of the inclusion of two short sequence elements that enhance the transcription of the 35S RNA (45).
The alternate views of the 35S RNA leader sequence presented in Fig. 9A and C both fit the model for the formation of the ribosomal shunt (10, 16, 48). The model in Fig. 9A illustrates the essential elements of the shunt, shunt donor, and acceptor sites as well as the sORFs in the leader sequence. Although the sORFs are a prominent feature of the leader sequence, the elimination of start codons in the leader only delays infectivity rather than abolishing it (48). Consequently, none are essential. Pooggin et al. (48) suggested that the sORFs may play a structural role in maintenance of the stem-loop structure or in positioning ribosomes for the shunt. In contrast, the model in Fig. 9C addresses the derivation of the ribosomal shunt mechanism; it suggests that the shunt is derived from two LTRs positioned in opposite directions. This model is consistent with the concept that sORFs might contribute to the stem-loop structure rather than encoding peptides that contribute to CaMV infections.
The structure of the CaMV large intergenic region illustrated in Fig. 9C implies that a CaMV progenitor may have served as a landing pad for other retrotransposons such that two LTRs would be juxtaposed next to the 35S promoter. There are, in fact, precedents for this arrangement of sequences. It is well established that retrotransposons of all types can be found in clusters (5, 52), and Ty1 is no exception. In a survey of transposable elements in the yeast genome, Kim et al. (27) documented 16 compound insertions in which retrotransposons had inserted within other retrotransposons. In addition, they found other cases in which Ty elements were inserted adjacent to each other and in opposite orientations. In one case, the retroelements were separated by only 10 base pairs (27).
It cannot be determined whether the tandem LTRs present in the CaMV large intergenic region are derived from a fungal or plant progenitor, although either scenario is possible. Ty1-copia-type retroelements are found in all plants (46, 59, 61), and they could have served as a source of the Ty1 LTR remnants in CaMV. Although pararetroviruses do not become integrated into their hosts as a part of their infection cycle, integrated copies of some pararetroviruses have been found in plant genomes (19, 26, 31, 49). Consequently, the LTR remnants in the CaMV large intergenic region might have arisen through recombination with a plant retroelement, at a time when a CaMV progenitor may have existed as an integrated copy in its host genome. However, attempts to align the CaMV LTR remnant with Ty1/copia-like and Ty3/Gypsy retroelements present in Arabidopsis (a total of 1,447 retroelements) failed to reveal any conservation (data not shown). Consequently, it does not appear that the CaMV LTR remnant is derived from either of these sources. In contrast, since the CaMV LTR remnants have homology with a yeast retroelement, they might have arisen from a mycorrhizal or endophytic fungus. Some fungi are capable of vectoring viruses into plants (22), although pararetroviruses have not been documented to be spread in this manner (22).
Mobile elements and segments of repetitive DNA have been documented to play a role in the evolution of their hosts through their insertion into and adjacent to host genes (6, 36, 61). Britten (6, 7) has compiled several examples that have been found in animals and insects and has suggested three criteria for inclusion in this group: (i) that the insertion occurred far in the past and that it is not a transient mutation; (ii) that the element is derived from recognizable group of similar sequences; and (iii) that the element serves a useful function. We propose that the LTR remnants that we have identified in the CaMV large intergenic region should be included in this group. Although we cannot establish a date for when the LTRs were inserted into the CaMV genome, they must be as old as CaMV itself, as they are an integral component of the CaMV genome. The elements themselves are related to the Ty1-copia class of retrotransposons, a class that is ubiquitous in plants (59, 61). Finally, the LTR remnants form the base of the large stem-loop structure that forms within the CaMV 35S RNA leader sequence; they are a structural requirement for the ribosomal shunt (10, 16, 48). The insertion of elements adjacent to each other and in opposite orientations may represent an economical way in evolution to form a sizable RNA local secondary structure, which could be much more efficient than the coevolution of two independent sequences to form the arms of a local secondary structure. This mechanism might appear in other genomes and for other types of RNA local structural motifs.
ACKNOWLEDGMENTS
We thank Barbara Wiggins for technical assistance and John Cannon for help with yeast strains and techniques.
This research was supported by the Missouri Agricultural Experiment Station, the Food for the 21st Century program at the University of Missouri, and the Research Board at the University of Missouri. The research of X.-F.W. and D.X. was supported by the U.S. Department of Energy's Genomes to Life program (www.doegenomestolife.org) under the project "Carbon Sequestration in Synechococcus sp.: from Molecular Machines to Hierarchical Modeling" (www.genomes-to-life.org).
REFERENCES
Altschul, S. F., W. Gish, W. Miller, E. W. Myers, and D. J. Lipman. 1990. Basic local alignment search tool. J. Mol. Biol. 215:403-410.
Ausubel, F., R. Brent, R. Kingstone, D. Moore, D. Seidman, J. G. Smith, and K. Struhl. 1995. Short protocols in molecular biology, 3rd ed. John Wiley and Sons, New York, N.Y.
Balakrishnan, R., K. R. Christie, M. C. Costanzo, K. Dolinski, S. S. Dwight, S. R. Engel, D. G. Fisk, J. E. Hirschman, E. L. Hong, R. Nash, R. Oughtred, M. Skrzypek, C. L. Theesfeld, G. Binkley, C. Lane, M. Schroeder, A. Sethuraman, S. Dong, S. Weng, S. Miyasato, R. Andrada, D. Botstein, and J. M. Cherry. Saccharomyces genome database. http://www.yeastgenome.org/. [Online.] Accessed 7 March 2005.
Boeke, J. D., and S. B. Sandmeyer. 1991. Yeast transposable elements, p. 193-261. In J. Broach, E. Jones, and J. Pringle (ed.), The molecular and cellular biology of the yeast Saccharomyces, vol. 1. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, N.Y.
Brandes, A., J. S. Heslop-Harrison, A. Kamm, S. Kubis, R. L. Doudrick, and T. Schmidt. 1997. Comparative analysis of the chromosomal and genomic organization of Ty1-copia-like retrotransposons in pteridophytes, gymnosperms and angiosperms. Plant Mol. Biol. 33:11-21.
Britten, R. J. 1996. DNA sequence insertion and evolutionary variation in gene regulation. Proc. Natl. Acad. Sci. USA 93:9374-9377.
Britten, R. J. 1997. Mobile elements inserted in the distant past have taken on important functions. Gene 205:177-182.
Clare, J., and P. Farabaugh. 1995. Nucleotide sequence of a yeast Ty element: evidence for an unusual mechanisms of gene expression. Proc. Natl. Acad. Sci. USA 82:2829-2833.
Daubert, S. D., and G. Routh. 1990. Point mutations in cauliflower mosaic virus gene VI confer host-specific symptom changes. Mol. Plant-Microbe Interact. 3:341-345.
Dominguez, D. I., L. A. Ryabova, M. M. Poogin, W. Schmidt-Puchta, J. Fütterer, and T. Hohn. 1998. Ribosome shunting in cauliflower mosaic virus. Identification of an essential and sufficient structural element. J. Biol. Chem. 273:3669-3676.
Eddy, E. R. 1998. Profile hidden Markov models. Bioinformatics 14:755-763.
Eddy, E. R. 1998. Multiple alignment and multiple sequence based searches. Cold Spring Harbor computational genomics course handout. [Online.] ftp://ftp.genetics.wustl.edu/pub/eddy/papers/tigs-9808.pdf. Cold Spring Harbor Laboratory, Cold Spring Harbor, N.Y.
Elder, R. T., E. Y. Loh, and R. W. Davis. 1983. RNA from the yeast transposable element Ty1 has both ends in the direct repeats, a structure similar to retrovirus RNA. Proc. Natl. Acad. Sci. USA 80:2432-2436.
Froissart, R., M. Uzest, V. Ruiz-Ferrer, M. Drucker, E. Hebrard, T. Hohn, and S. Blanc. 2004. Splicing of Cauliflower mosaic virus 35S RNA serves to downregulate a toxic gene product. J. Gen. Virol. 85:2719-2726.
Fütterer, J., K. Gordon, J. M. Bonneville, H. Sanfaon, B. Pisan, J. R. Penswick, and T. Hohn. 1988. The leading sequence of cauliflower mosaic virus large RNA can be folded into a large stem-loop structure. Nucleic Acids Res. 16:8377-8390.
Fütterer, J., Z. Kiss-Laszlo, and T. Hohn. 1993. Nonlinear ribosome migration on cauliflower mosaic virus 35S RNA. Cell 73:789-802.
Gardner, R. C., A. J. Howarth, P. Hahn, M. Brown-Leudi, R. J. Shepherd, and J. Messing. 1981. The complete nucleotide sequence of an infectious clone of cauliflower mosaic virus by M13mp7 shotgun sequencing. Nucleic Acids Res. 9:2871-2888.
Haas, M., A. Geldreich, M. Bureau, L. Dupuis, V. Leh, G. Vetter, K. Kobayashi, T. Hohn, L. Ryabova, P. Yot, and M. Keller. 2005. The open reading frame VI product of Cauliflower mosaic virus is a nucleocytoplasmic protein: its N terminus mediates its nuclear export and formation of electron-dense viroplasms. Plant Cell 17:927-943.
Harper, G., J. O. Osuji, J. S. Heslop-Harrison, and R. Hull. 1999. Integration of banana streak badnavirus into the Musa genome: molecular and cytogenetic evidence. Virology 255:207-213.
Hemmings-Mieszczak, M., G. Steger, and T. Hohn. 1997. Alternative structures of the cauliflower mosaic virus 35S RNA leader: implications for viral expression and replication. J. Mol. Biol. 267:1075-1088.
Hirt, H., M. Kgl, T. Murbacher, and E. Heberle-Bors. 1990. Evolutionary conservation of transcriptional machinery between yeast and plants as shown by the efficient expression from the CaMV 35S promoter and 35S terminator. Curr. Genet. 17:473-479.
Hull, R. 2002. Matthews' plant virology, 4th ed. Academic Press, San Diego, Calif.
Irniger, S., H. Sanfacon, C. H. Egli, and G. H. Braus. 1992. Different sequence elements are required for function of the cauliflower mosaic virus polyadenylation site in Saccharomyces cerevisiae compared with in plants. Mol. Cell. Biol. 12:2322-2330.
Ito, H., Y. Fukuda, K. Murata, and A. Kimura. 1983. Transformation of intact yeast cells treated with alkali cations. J. Bacteriol. 153:163-168.
Jacq, C., J. Alt-Mrbe, B. Andre, W. Arnold, A. Bahr, et al. 1997. The nucleotide sequence of Saccharomyces cerevisiae chromosome IV. Nature 387:75-78.
Jakowitsch, J., M. F. Mette, J. van der Winden, M. A. Matzke, and A. J. M. Matzke. 1999. Integrated pararetroviral sequences define a unique class of dispersed repetitive DNA in plants. Proc. Natl. Acad. Sci. USA 96:13241-13246.
Kim, J. M., S. Vanguri, J. D. Boeke, A. Gabriel, and D. F. Voytas. 1998. Transposable elements and genome organization: a comprehensive survey of retrotransposons revealed by the complete Saccharomyces cerevisiae genome sequence. Genome Res. 8:464-478.
Kiss-László, Z., J. Fütterer, and T. Hohn. 1995. Splicing of cauliflower mosaic virus 35S RNA is essential for viral infectivity. EMBO J. 14:3552-3562.
Kozak, M. 1999. Initiation of translation in prokaryotes and eukaryotes. Gene 234:187-208.
Kushner, D. B., B. D. Lindenbach, V. Z. Grdzelishvili, A. O. Noueiry, S. M. Paul, and P. Ahlquist. 2003. Systematic, genome-wide identification of host genes affecting replication of a positive-strand RNA virus. Proc. Natl. Acad. Sci. USA 100:15764-15769.
Lockhart, B. E., J. Menke, G. Dahal, and N. E. Olszewski. 2000. Characterization and genomic analysis of tobacco vein clearing virus, a plant pararetrovirus that is transmitted vertically and related to sequences integrated in the host genome. J. Gen. Virol. 81:1579-1585.
Lowe, T. M., and S. R. Eddy. 1997. tRNAscan-SE: A program for improved detection of transfer RNA genes in genomic sequence. Nucleic Acids Res. 25:955-964.
Maniatis, T., E. F. Fritsch, and J. Sambrook. 1982. Molecular cloning: a laboratory manual. Cold Spring Harbor Laboratory, Cold Spring Harbor, N.Y.
Matyunina, L. V., I. K. Jordan, and J. F. MacDonald. 1997. Naturally occurring variation in copia expression is due to both element (cis) and host (trans) regulatory variation. Proc. Natl. Acad. Sci. USA 93:7097-7102.
McCarthy, J. E. G. 1998. Posttranscriptional control of gene expression in yeast. Microbiol. Mol. Biol. Rev. 62:1492-1553.
McDonald, J. F. 1993. Evolution and consequences of transposable elements. Curr. Opin. Genet. Dev. 3:855-864.
McDonald, J. F., L. V. Matyunina, S. Wilson, I. K. Jordan, N. J. Bowen, and W. J. Miller. 1997. LTR retrotransposons and the evolution of eukaryotic enhancers. Genetica 100:3-13.
Messing, J. 1979. A multiple cloning system based on single-stranded DNA bacteriophage M13. Recomb. DNA Tech. Bull. 2:43-48.
Noueiry, A. O., and P. Ahlquist. 2003. Brome mosaic virus RNA replication: revealing the role of the host in RNA virus replication. Annu. Rev. Phytopathol. 41:77-98.
Odell, J. T., F. Nagy, and N.-H. Chua. 1985. Identification of DNA sequences required for activity of the cauliflower mosaic virus 35S promoter. Nature 313:810-812.
Palanichelvam, K., and J. E. Schoelz. 2002. A comparative analysis of the avirulence and translational transactivator functions of gene VI of cauliflower mosaic virus. Virology 293:225-233.
Panavas, T., and P. D. Nagy. 2003. Yeast as a model host to study replication and recombination of defective interfering RNA of Tomato bushy stunt virus. Virology 314:315-325.
Panavas, T., E. Serviene, J. Brahser, and P. D. Nagy. 2005. Yeast genome-wide screen reveals dissimilar sets of host genes affecting replication of RNA viruses. Proc. Natl. Acad. Sci. USA 102:7326-7331.
Park, H. S., A. Himmelbach, K. S. Browning, T. Hohn, and L. A. Ryabova. 2001. A plant viral "reinitiation" factor interacts with the host translational machinery. Cell 106:723-733.
Pauli, S., H. M. Rothnie, G. Chen, X. He, and T. Hohn. 2004. The cauliflower mosaic virus 35S promoter extends into the transcribed region. J. Virol. 78:12120-12128.
Peterson-Burch, B. D., and D. F. Voytas. 2002. Genes of the Pseudoviridae (Ty1/copia retrotransposons). Mol. Biol. Evol. 19:1832-1845.
Pfeiffer, P., and T. Hohn. 1983. Involvement of reverse transcription in the replication of cauliflower mosaic virus: a detailed model and test of some aspects. Cell 33:781-789.
Pooggin, M. M., T. Hohn, and J. Fütterer. 1998. Forced evolution reveals the importance of short open reading frame A and secondary structure in the cauliflower mosaic virus 35S RNA leader. J. Virol. 72:4157-4169.
Richert-Pggeler, K. R., and R. J. Shepherd. 1997. Petunia vein clearing virus: a plant pararetrovirus with the core sequences for an integrase function. Virology 236:137-146.
Rothnie, H., Y. Chapdelaine, and T. Hohn. 1994. Pararetroviruses and retroviruses: a comparative review of viral structure and gene expression strategies. Adv. Virus Res. 44:1-67.
Ryabova, L. A., M. M. Pooggin, and T. Hohn. 2002. Viral strategies of translation initiation: ribosomal shunt and reinitiation. Prog. Nucleic Acid Res. Mol. Biol. 72:1-39.
SanMiguel, P., A. Tikhonov, Y.-K. Jin, N. Motchoulskaia, D. Zakharov, A. Melake-Berhan, P. S. Springer, K. J. Edwards, M. Lee, Z. Avramova, and J. L. Bennetzen. 1996. Nested retrotransposons in the intergenic regions of the maize genome. Science 274:765-768.
Schardl, C. L., A. D. Byrd, G. Benzion, M. A. Altschuler, D. F. Hildebrand, and A. G. Hunt. 1987. Design and construction of a versatile system for the expression of foreign genes in plants. Gene 61:1-11.
Schneider, J. G., and L. Guarente. 1991. Vectors for expression of cloned genes in yeast. Methods Enzymol. 194:373-388.
Schoelz, J. E., R. J. Shepherd, and R. D. Richins. 1986. Properties of an unusual strain of cauliflower mosaic virus. Phytopathology 76:451-454.
Sha, Y., E. P. Broglio, J. F. Cannon, and J. E. Schoelz. 1995. Expression of a plant viral polycistronic mRNA in yeast, Saccharomyces cerevisiae, mediated by a plant virus translational transactivator. Proc. Natl. Acad. Sci. USA 92:8911-8915.
Thomas, P. S. 1980. Hybridization of denatured RNA and small DNA fragments transferred to nitrocellulose. Proc. Natl. Acad. Sci. USA 77:5201-5205.
Thompson, J. D., D. G. Higgins, and T. J. Gibson. 1994. CLUSTAL W: improving the sensitivity of progressive multiple sequence alignment through sequence weighting, position-specific gap penalties and weight matrix choice. Nucleic Acids Res. 22:4673-4680.
Voytas, D. F., M. P. Cummings, A. Konieczny, F. M. Ausubel, and S. R. Rodermel. 1992. Copia-like retroelements are ubiquitous among plants. Proc. Natl. Acad. Sci. USA 89:7124-7128.
Wan, X.-F., and D. Xu. Computational methods for remote homolog identification. Curr. Protein. Pept. Sci., in press.
White, S. E., L. F. Habera, and S. R. Wessler. 1994. Retrotransposons in the flanking regions of normal plant genes: a role for copia-like elements in the evolution of gene structure and expression. Proc. Natl. Acad. Sci. USA 91:11792-11796.
Wilson, S., L. V. Matyunina, and J. F. MacDonald. 1998. An enhancer region within the copia untranslated leader contains binding sites for Drosophila regulatory proteins. Gene 209:239-246.
Zuker, M. 2003. Mfold web server for nucleic acid folding and hybridization prediction. Nucleic Acids Res. 31:3406-3415.(Monir Shababi, June Bourq)
Department of Computer Science, 201 Engineering Bldg. West, University of Missouri, Columbia, Missouri 65211
Plant Biology Dept., The Samuel Roberts Noble Foundation, Ardmore, Oklahoma 73402
Dept. of Biochemistry, Dakota Wesleyan University, Mitchell, South Dakota 57301
Dept. of Microbiology, Miami University, Oxford, Ohio 45056
ABSTRACT
We have screened portions of the large intergenic region of the Cauliflower mosaic virus (CaMV) genome for promoter activity in baker's yeast (Saccharomyces cerevisiae) and have identified an element that contributes to promoter activity in yeast but has negligible activity in plant cells when expressed in an agroinfiltration assay. A search of the yeast genome sequence revealed that the CaMV element had sequence similarity with the R region of the long terminal repeat (LTR) of the yeast Ty1 retrotransposon, with significant statistical confidence. In plants, the same CaMV sequence has been shown to have an essential role in the ribosomal shunt mechanism of translation, as it forms the base of the right arm of the stem-loop structure that is required for the ribosomal shunt. Since the left arm of the stem-loop structure must represent an imperfect reverse copy of the right arm, we propose that the ribosomal shunt has evolved from a pair of LTRs that have become incorporated end to end into the CaMV genome.
INTRODUCTION
Cauliflower mosaic virus (CaMV), the type member of the caulimovirus group, is a pararetrovirus; it replicates by reverse transcription, but replication does not require the integration of the CaMV genome into host chromosomes. The lack of an integration step is one feature that distinguishes the replication strategy of CaMV from that of other plant retrotransposons or the animal retroviruses, and this is reflected in the structure of the CaMV genome (50). The CaMV genome is composed of circular, double-stranded DNA that is approximately 8,000 bp in size (Fig. 1A). The replication intermediate that serves as the template for reverse transcription is the CaMV 35S RNA (Fig. 1A) (47). The 35S RNA is formed when the host RNA polymerase II initiates transcription at nucleotide (nt) position 7435. A transcript termination signal at nucleotide 7613 is initially bypassed such that the RNA Pol II synthesizes viral RNA over the full length of the viral genome until it reaches the transcript termination signal for a second time, at which point transcription stops. Consequently, the first 179 bp of the 5' end of the 35S RNA are reiterated on the 3' end of the 35S RNA (50).
The terminal redundancy in the 35S RNA is required for the template switch that occurs during the synthesis of the first strand of viral DNA by the CaMV reverse transcriptase. In this capacity, the terminal redundancies of the 35S RNA fulfill the function of the R region present in the long terminal repeats (LTRs) of retrotransposons and retroviruses. The structure and function of LTRs are most easily visualized in the retrotransposon forms that are integrated into host chromosomes (Fig. 1B). For example, the integrated form of Ty1 consists of two complete LTRs that flank the open reading frames (ORFs), and each of the LTRs is composed of three distinct regions: U3, R, and U5 (4, 5, 13). Transcription is initiated within the leftward LTR at the junction of U3 and R and is terminated at the junction of R and U5 in the rightward LTR (Fig. 1B). The resultant mRNA has a copy of R followed by U5 at its 5' end, whereas the 3' end consists of U3 ending with the R region (4, 13). Consequently, in comparing the terminal redundancies of CaMV to the LTRs of the retrotransposons, both groups have clearly defined R regions, whereas the U3 and U5 regions can be only identified in the LTRs of the retrotransposons.
A second feature that is a hallmark of the caulimoviruses is the ribosomal shunt mechanism of translation (51). The ribosomal shunt has undoubtedly evolved over time to compensate for the complexity and length of the leader sequence of the 35S RNA. This leader sequence is approximately 604 bp in length, and it contains up to nine short ORFs (sORFs), depending on the strain, that differ in size from 9 to 102 nucleotides (Fig. 1C). Such a leader could be expected to be a formidable impediment to translation in any eukaryotic organism, as eukaryotic ribosomes generally initiate translation at the first AUG that is found in a favorable context and encountered nearest the 5' end of the transcript (29). In the majority of eukaryotic transcripts, ribosomes are unable to reinitiate the transcription of a second cistron. However, extensive studies have shown that the complexity of the 35S RNA leader sequence is bypassed through the formation of a large stem-loop structure (15, 20), which brings into close proximity the essential elements of the ribosomal shunt (the shunt donor and acceptor sites). In translation of the 35S RNA, host ribosomes enter the 5' end of the 35S RNA and scan a short distance until they reach the shunt donor site (between sORFs A and B) (Fig. 1C). At this point in the sequence, ribosomes bypass the central region of the leader and land at the shunt acceptor site within sORF F (10, 16, 48). The shunt mechanism allows ribosomes to reach the start codon of gene VII. The subsequent expression of genes I through V on the 35S RNA involves alternative mechanisms, including transactivation of the translation of genes I through V by the CaMV gene VI product (18, 44) and splicing (14, 28).
In the present study, we developed a system using Saccharomyces cerevisiae to probe the expression strategy of CaMV. S. cerevisiae has proven useful for studying the transcription and replication of positive-sense plant viruses, such as Brome mosaic virus and Tomato bushy stunt virus (39, 42), and genome-wide screens have revealed yeast (Saccharomyces cerevisiae) proteins that are involved in the replication of these viruses (30, 43). Previous studies have also shown that there is a degree of conservation between the transcriptional machinery of yeasts and plants. For example, the CaMV 35S promoter is active in yeast species (21) and the CaMV 35S terminator sequence also can function as a terminator in yeast, although the sequence elements that are responsible for termination are different from those that are active in plants (23). Furthermore, the CaMV translational transactivator protein has limited activity in yeast (56).
Our study has revealed the presence of a cryptic yeast promoter within the CaMV genomic region that corresponds to a portion of the 35S RNA leader sequence. A BLAST search of the yeast genome showed that this CaMV promoter element was similar to the R region of the LTR of the yeast Ty-1 retrotransposon. In plants, the same CaMV sequence has been shown to have an essential role in the ribosomal shunt mechanism of translation, as it forms the base of the right arm of the stem-loop. Since the left arm of the stem-loop must represent an imperfect reverse copy of the right arm, we propose that the ribosomal shunt has evolved from a pair of LTRs that have become incorporated end to end into the CaMV genome.
MATERIALS AND METHODS
Construction of CaMV-CAT fusion constructs for expression in yeast. The plasmid pMono-CAT was constructed by inserting the chloramphenicol acetyltransferase (CAT) gene from pCM-1 (Pharmacia, Piscataway, NJ) into the yeast shuttle vector pYES2 (Invitrogen, Carlsbad, CA). To make CaMV-CAT fusions, the DNA sequence from the 35S RNA transcriptional start site (nucleotide 7433) to an XhoI site within gene II (nucleotide 1642) of CaMV strain D4 (55) was amplified by PCR to create a DNA segment with flanking KpnI and XhoI sites for insertion into pYES2. The CAT gene was inserted in frame into the CaMV sequence through the introduction of flanking restriction enzyme sites, and CaMV sequences downstream from CAT were eliminated. Point mutations were introduced into the CaMV and CAT sequences by overlapping PCR. Nucleotide sequences of plasmid inserts were determined to ensure that the CAT gene had been inserted in frame. The bicistronic -glucuronidase (GUS)/CAT plasmids were all based on plasmid pJB79, which was constructed by inserting a Hind III-BamHI fragment containing the GUS gene into pJS163 (a plasmid in which the CAT gene had been inserted into the XhoI site of the yeast vector, pYES2) (56).
Construction of plasmids for gene expression studies in whole leaves. To evaluate the strength of the cryptic yeast promoter in whole leaves, GUS reporter constructs were inserted into an Agrobacterium tumefaciens binary vector, pKYLX7 (53). To create pKYGUS, the GUS gene was modified by PCR to be flanked by KpnI and XhoI sites, which facilitated its cloning into KpnI/XhoI sites engineered into pKYLX7. pCMS129 was created by replacing the 35S promoter, which was present on an EcoRI/HindIII DNA fragment, with a CaMV segment from nucleotides 7672 to 131. pKYMdg was created by digesting pCMS129 with EcoRI and HindIII, filling in the sticky ends, and religating the plasmid. Triparental matings and agroinfiltration were performed as described previously by Palanichelvam and Schoelz (41). Leaf samples were evaluated for GUS expression 4 days after infiltration.
Recombinant DNA techniques and yeast transformation. Restriction enzymes and T4 DNA ligase were purchased from Promega Corp. (Madison, WI) or New England Biolabs (Beverly, MA). DNA ligation, plasmid transformation, and plasmid purification were performed according to the procedure of Maniatis et al. (33). Plasmids were propagated in Escherichia coli strain JM101 (38) grown in double yeast tryptone broth containing 100 μg of ampicillin per milliliter. Purified plasmid DNA from E. coli was transformed into S. cerevisiae strain JC746Dip (MATa/MATa ura3/ura3 his3/his3 leu2/leu2 trp/trp) by the lithium acetate method (24). Yeast transformants were grown in synthetic dextrose medium (0.67% yeast nitrogen base without amino acids, supplemented with auxotrophic requirements and 2% raffinose) lacking uracil.
Detection of the CAT mRNA in yeast by Northern blot analysis. Yeast was grown in minimal medium to an optical density at 600 nm of 1.6 to 1.9, and total RNA was isolated by the glass bead method (2) (type IV glass beads, 250 to 300 microns in diameter; Sigma, St. Louis, MO). Equal concentrations of RNA were denatured using glyoxal, sodium phosphate (pH 7.0), and dimethyl sulfoxide and run on a 1% agarose gel in the presence of 0.01 M sodium phosphate (57). Northern hybridization was performed according to the procedure of Maniatis et al. (33) by using a 32P-labeled CAT DNA probe.
Measurement of CAT protein levels in yeast extracts by CAT ELISA. The induction of the GAL1 promoter has been described previously by Schneider and Guarente (54). Yeast cells (40 ml at an optical density at 600 nm of 1.6 to 1.9) were pelleted by centrifugation, and the pellet was washed twice with distilled water. The pellet was resuspended in 200 μl of 0.25 M Tris-HCl (pH 7.8), the cells were sonicated for a total of 120 s (four times for 30 s each, with incubation on ice for 1 min between sonications), and cell debris were removed by centrifugation. The concentration of total protein was determined by using a modified Lowry protein assay (Bio-Rad, Hercules, CA), and the amount of CAT protein was measured using a CAT enzyme-linked immunosorbent assay (ELISA) kit (Boehringer Mannheim, Indianapolis, IN). CAT and GUS levels were measured by at least three tests for each plasmid, with three replicates per test.
Nucleotide sequence analysis. Nucleotide sequences of CaMV strain D4 present in clone pCMS119 (from nucleotides 7949 to 131) were determined at the DNA Sequencing Core Facility of the University of Missouri, Columbia. The genome size of 8,034 bp for D4 was determined through a comparison of the partial D4 nucleotide sequence to the complete CM1841 nucleotide sequence (17). RNA secondary structure for the CaMV strain D4 leader sequence was generated by using the mFOLD program (63). tRNA primer binding sites for the putative left LTR were found in the tRNAscan-SE genomic tRNA database (32). Multiple sequence alignments were performed using Clustal W (58).
To assess the statistical significance for the homologous relationship between the R region in yeast LTR and CaMV sequences, we performed a sequence-profile analysis using HMMER (11), which applies profile-hidden Markov models for the statistical descriptions of a sequence family's consensus (60). We downloaded 297 Ty1 and Ty2 LTRs from the yeast genome database (http://www.yeastgenome.org) (27), from which we obtained 40 full LTRs. To avoid the bias from data redundancy, we kept only one copy of LTR R in the profile data set when there were multiple identical sequences. As a result, 20 LTR-R regions were used to construct the multiple sequence alignment profile LTR-R using Clustal W (58). Eight complete CaMV genome sequences were obtained from GenBank (BBC, GenBank ID M90542.1; NY8153, M90541.1; CM1841, V00140.1; CMV-1, M90543.1; B29, X79465.1; Strasbourg, V00141.1; D/H, M10376.1; and Xinjiang, AF14064.1) as well as a CaMV D4 partial genomic sequence (DQ355155). As with the yeast LTR sequences, only one copy of a CaMV sequence was retained in the instance of multiple identical sequences. Consequently, we searched the LTR-R profile against seven CaMV sequences. The E value from HMMER was used to evaluate the significance of the homologous relationship between yeast LTR-R and CaMV R regions. We utilized the E value of 0.05 suggested by HMMER as the significance threshold (12).
Mapping the 5' end of the CAT mRNA derived from the CaMV sequence. Total RNA was extracted from yeast cells transformed with pCMS119 and used for 5'-end mapping of the CAT mRNA by using the GeneRacer Kit (Invitrogen, Carlsbad, CA). Following the dephosphorylation and decapping of the full-length mRNA, the GeneRacer RNA oligonucleotide was ligated to the 5' end and used for reverse transcription. The resulting cDNA was PCR amplified by using a GeneRacer-specific 5' primer and a gene-specific primer corresponding to the 3' end of the CAT gene. Two fragments, which were approximately the size of the CAT gene, were amplified. The two fragments were separated from each other by gel isolation and subsequently cloned by using a one-shot Topo cloning kit (Invitrogen, Carlsbad, CA). The resultant clones were submitted for sequencing at the DNA Sequencing Core Facility of the University of Missouri, Columbia to identify the transcript initiation sites.
Measurement of GUS expression in yeast cellular extracts and agroinfiltrated plant disks. After protein concentrations from yeast and plants were determined, GUS expression was measured by using a chemiluminescence GUS assay with the GUS-light kit (Tropix, Bedford, MA). GUS concentrations in experimental samples were determined by comparison with a standard curve made with purified GUS supplied by the kit.
Nucleotide sequence accession number. The sequence for CaMV strain D4 was submitted to GenBank and assigned accession number DQ355155.
RESULTS
A survey of expression competence from different points within the CaMV large intergenic region in yeast. To assess the expression of CaMV genes that were present on the 35S RNA in yeast, a CAT reporter gene was inserted into two positions within the 35S RNA leader sequence and in frame with genes VII and I (Fig. 2). All reporter-fusion constructs were placed under the control of the yeast GAL1 promoter and CYC1 terminator sequences present in the yeast plasmid pYES2. Yeast cells were transformed with each plasmid, and a CAT ELISA was used to quantitatively measure CAT expression in yeast cellular extracts. The expression of each plasmid was then compared to pMono-CAT, in which the CAT gene was expressed from the GAL1 promoter.
The expression of CAT decreased as it was inserted further within the 35S leader sequence, which likely reflected a diminishing ability of ribosomes to reinitiate translation as they migrated through the sORFs of the leader sequence. For example, CAT protein was expressed from the construct pABC-CAT to a level that was approximately 11% of that of the pMono-CAT control plasmid (Fig. 2). The upstream sORFs A, B, and C did have an inhibitory effect on expression of CAT in yeast, but their presence upstream from the CAT gene was not sufficient to completely block expression of the CAT protein. In contrast, the expression of CAT protein in the construct pF-CAT was negligible (0.4% of the pMono-CAT control). Since upstream open reading frames and extensive secondary structure are reported to inhibit the translation of mRNAs in yeast (35), we hypothesized that CAT would not be expressed from points within genes VII and I. As expected, the CAT gene was expressed to only a very low level from within gene I. Surprisingly, the highest level of CAT expression was observed with pVII-CAT. Whereas sORF F was expressed at a level of only 0.4% of the pMono-CAT control in yeast, pVII-CAT was expressed at 2.5 times the level of the monocistronic CAT control, a nearly 700-fold difference in expression between sORF F and gene VII (Fig. 2).
In the construct pVII-CAT, the CAT gene was inserted in frame into the coding sequence and so we anticipated that a CAT fusion product would be produced in these cells. To determine whether the CAT protein was in fact fused to CaMV sequences within ORF VII, we altered the CaMV and CAT start codons through PCR-directed mutagenesis. A mutation of the gene VII start codon had no effect on the level of CAT expressed from pCMS100, whereas a mutation of the CAT start codon in pCMS93 almost completely abolished CAT expression (Fig. 2). We also directly compared the size and amount of CAT protein produced in yeast that contained pVII-CAT to that produced in the pMono-CAT cells. In agreement with the CAT ELISA results, a Western blot assay indicated a slightly higher level of the CAT protein in pVII-CAT compared to that in pMono-CAT. In addition, the CAT protein produced from the pVII-CAT construct had the same molecular weight as the CAT protein produced in cells containing the pMono-CAT plasmid (data not shown). These results demonstrated that the CAT start codon is used to express CAT rather than CAT being initiated from the upstream gene VII start codon.
Sequences within sORF F and gene VII form part of a cryptic yeast promoter. The high level of expression of CAT protein from pVII-CAT could be due to a ribosomal shunt mechanism, an internal ribosomal entry site (IRES), or the presence of a cryptic yeast promoter. To investigate whether the high level of CAT expression in yeast was due to a ribosomal shunt mechanism, we deleted 239 nucleotides, from the beginning of the 35S RNA to 25 nucleotides upstream of sORF D (nucleotides 7435 to 7672), to form pCMS95. The effect of this deletion would be to abolish a significant portion of one of the arms of the mRNA stem-loop structure (10, 48). However, this deletion had no effect on CAT expression (Fig. 3), and smaller deletions of sORF A and sORF B also had no effect on CAT expression (data not shown). These results indicated that ribosomal shunting was not responsible for expression of CAT in the construct pVII-CAT, and this was confirmed in the experiments described below.
To further delineate the CaMV sequences that were responsible for CAT expression and to test the possibility of internal initiation in yeast, we created a dicistronic plasmid (Fig. 3, pJB79) in which the GUS gene was placed immediately upstream from the CAT gene. As expected, GUS was expressed to a high level, whereas the CAT protein levels expressed from pJB79 were low (4.4% of that of the monocistronic CAT control). To determine whether the CaMV sequences could act as an IRES in yeast, the CaMV sequence present in pCMS95 was inserted between the GUS and CAT genes to form pCMS99 (Fig. 3). CAT expression from pCMS99 was approximately twice the level of that from pMono-CAT and nearly 42-fold higher than that of pJB79 (Fig. 3). To further delineate the CaMV regions that were responsible for CAT expression, we deleted the CaMV sequences comprising sORFs D and E from pCMS99 to make the construct pCMS101. The CAT expression level of pCMS101 was comparable to that of pMono-CAT and almost 50% of that of pCMS99 (Fig. 3). Therefore, CaMV sequences that were responsible for CAT expression from the dicistronic construct mapped primarily to nucleotides 7949 to 131. The expression of GUS from pCMS99 and pCMS101 was equivalent to GUS expression of pJB79, which showed that the enhancement in CAT expression in these constructs was not due to an increase in transcription from the GAL1 promoter (Fig. 3).
To distinguish whether the CaMV sequences in pCMS101 functioned as a promoter or as an IRES in yeast, we deleted the yeast GAL1 promoter sequence and examined the effect of this deletion on the expression of CAT and GUS in pCMS119. Interestingly, the level of CAT protein expressed from pCMS119 was comparable to that of pCMS101 (Fig. 3), whereas GUS expression was abolished. We concluded that CAT expression was not dependent upon the GAL1 promoter and that sequences within CaMV sORF F and gene VII were part of a cryptic promoter that functioned in yeast. A further deletion of CaMV sequences from nucleotides 7949 to 7995 essentially abolished the activity of the putative CaMV promoter, as CAT protein expression from pCMS120 was comparable to that of the control plasmid pJB79 (Fig. 3). Furthermore, the addition of the GAL1 promoter to pCMS120 had no effect on the CAT protein level (Fig. 3, compare pCMS120 and pCMS121), confirming that CAT expression was driven by a cryptic promoter present in the CaMV sequences.
A comparison of the CAT transcripts produced in yeast cells provided further evidence that CaMV sequences centered around sORF F and gene VII functioned as an active promoter in yeast. Two prominent transcripts that contain CAT sequences can be seen in yeast cells transformed with pCMS101 (Fig. 4). The larger transcript, expressed from the GAL1 promoter, consisted of GUS and CAT sequences, whereas a second transcript was smaller than the pMono-CAT transcript (Fig. 4). The larger size of the CAT mRNA that was derived from the pMono-CAT plasmid could be attributed to sequences from the pYES2 polylinker within its untranslated leader sequence. Removal of the GAL1 promoter, as in constructs pCMS119 and pCMS120, abolished the larger GUS/CAT transcript (Fig. 4), demonstrating that this transcript was not responsible for the expression of CAT. In contrast, the elimination of CaMV sequences from nucleotides 7949 to 7995 abolished the smaller transcript (Fig. 4, lanes pCMS120 and pCMS121) and CAT protein expression (Fig. 3).
With the deletion of the GAL1 promoter, a new band appeared that was larger than the transcript associated with GAL1 (Fig. 4, lanes pCMS119 and pCMS120). We considered the possibility that this RNA band might originate from a cryptic promoter present in the pYES2 plasmid that was activated in the absence of the GAL1 promoter. However, the presence of this new larger transcript was not correlated with the expression of CAT. The CAT protein expression level was high in yeast that contained pCMS119 and low in yeast that contained pCMS120 (Fig. 3), yet the larger CAT RNA band was present in both of these yeast cells (Fig. 4). Taken together, the CAT protein expression studies and RNA gel blots showed that the CaMV region that comprises sORF F and gene VII drives the transcription of the CAT mRNA in yeast.
Identification of two transcript initiation sites within gene VII. The size of the CAT mRNA derived from the CaMV yeast promoter, which was detected in the Northern blot analysis of pCMS101 and pCMS119, indicated that the initiation of transcription occurs within gene VII near the CAT start codon. To precisely determine the transcript initiation site for the CAT mRNA in pCMS119, we used a GeneRacer kit. Two PCR fragments, corresponding to the 5' end of the CAT mRNA, were cloned, and their nucleotide sequence was determined. Based on the intensity of the PCR-amplified fragments and nucleotide sequence, the most likely transcript initiation site was at the AG dinucleotides 35 nt upstream from the CAT start codon (Fig. 5). A second site was located 56 nucleotides upstream of the CAT start codon at CT.
The yeast promoter located in sORF F and gene VII does not have activity in leaf mesophyll cells. To determine whether the CaMV region we identified as a yeast promoter functioned as a plant promoter, we inserted CaMV sequences from nucleotides 7672 to 131 upstream from a GUS reporter gene to create pCMS129. The promoter activity of pCMS129 was compared to that of the CaMV 35S promoter and a promoterless GUS construct (Fig. 6) in an agroinfiltration assay of Nicotiana edwardsonii leaves. GUS activity from pCMS129 was 285-fold less than the GUS activity from the 35S promoter and only slightly higher than the promoterless control plasmid (Fig. 6). These results showed that the CaMV cryptic yeast promoter had negligible activity in plant leaf tissue in an agroinfiltration assay. However, this test could not eliminate the possibility that the cryptic promoter might function in plants in the presence of one or more CaMV proteins.
The CaMV cryptic yeast promoter has sequence homology to the R region within the long terminal repeat of the yeast retrotransposon Ty1. The constructs in Fig. 3 revealed that a 213-bp CaMV sequence functioned as a promoter in yeast and that the CaMV sequence from nucleotides 7949 to 7995 was essential for promoter activity. To identify sequences in the yeast genome that have homology to this CaMV region, we searched the Saccharomyces genome database (3) with the BLASTN 2.0 program (1). This search yielded a single match between CaMV sequences from nucleotides 7952 to 7993 and a S. cerevisiae delta element (Fig. 7A, YDRCdelta3) located on chromosome IV at nucleotide coordinates 513338 to 513294 (25). YDRCdelta3 is a single copy of the LTR of Ty1. The CaMV sequence was 76% identical to the YDRCdelta3 sequence and 70% identical to the sequence of Ty912 (8), the first member of the Ty1 family to be sequenced (Fig. 7A).
In addition to the direct alignment of nucleotide sequences between sORF F and the R region of the Ty1 LTR, the position of the methionine (Met) tRNA binding site was also conserved. This sequence serves as the primer binding site for first-strand DNA synthesis in both Ty1 and the CaMV 35S RNA. In the case of the Ty1 LTR, the R region is located 41 nucleotides upstream from the Met tRNA binding site, whereas in CaMV the distance is 39 nucleotides (Fig. 7A). The Ty1 U5 region, the sequence that intervenes between the R region and the Met tRNA binding site, has diverged significantly from the corresponding CaMV sequence. However, the positioning of the Met tRNA binding sites in relation to either CaMV sORF F or the R region of the Ty1 LTR provides a second structural landmark that extends the relationship between the large intergenic region of CaMV and the LTR of Ty1. Consequently, most of the R region of the Ty1 LTR is represented in the CaMV large intergenic region, including the junction between the R and U5 regions, and this putative R region is located within the CaMV sequence the expected distance from its first-strand primer binding site.
To assess the statistical significance for the homologous relationship between the R region in the Ty1 LTR and CaMV sequences, we performed a sequence-profile analysis. We searched the profile of yeast LTR R regions against each of the six CaMV genome sequences and the CaMV D4 partial genome sequence by using HMMER (Fig. 8) (11). Our analysis demonstrated that the E values between the profiles of LTR R regions in yeast and the genomic sequence of CaMV range from 0.026 to 0.00043, which is much lower than the E value significance threshold of 0.05 suggested by HMMER (12). Among the seven CaMV sequences, the D4 strain has the most significant E value of 0.00043. Thus, the sequence profile analyses suggest that the yeast LTR R region and the R region of CaMV are homologous with significant confidence.
Within the context of the CaMV replication cycle, the CaMV sequence from nucleotides 7952 to 7993 has a function that is entirely unrelated to an LTR, as it has been identified as the ribosome shunt landing site (Fig. 7B) (48). This stable stem structure has been previously shown by mutational analysis to be essential for the ribosomal shunt (10). Figure 7B illustrates that the CaMV sequence that corresponds to the R region in the Ty1 LTR has previously been predicted to form the base of the right arm of the large stem-loop structure within the leader sequence of the 35S RNA (48).
DISCUSSION
We have identified a 42-nucleotide CaMV sequence in the large intergenic region that contributes to three distinct functions in two different organisms. The sequence had previously been shown to have an essential role in the ribosomal shunt mechanism of translation of the CaMV 35S RNA in plants; it corresponds to the base of the right arm of a large stem-loop structure that forms in the leader of the 35S RNA (15, 20). This stem-loop structure is required for the formation of the ribosomal shunt landing site (10, 16, 48) to initiate the translation from ORF VII or ORF I. In yeast, the same sequence forms part of a cryptic promoter/enhancer element. This sequence directed a high level of expression of a CAT reporter gene in yeast, but it had negligible activity in plant cells in an agroinfiltration assay. A BLAST search of the yeast genome revealed a third function for this sequence element, as it shares homology to the R region of the LTRs of the yeast retrotransposon Ty1. A search using the sequence profile built from the LTRs in yeast against the various CaMV genomic sequences indicates that the homology is statistically significant. In the discussion below, we evaluate each of these functions and present our arguments for an evolutionary relationship between the ribosomal shunt and LTRs.
A sequence element in sORF F functions as a cryptic yeast promoter/enhancer and has homology to the R region present in the LTR of the yeast retrotransposon Ty1. In our study, we considered three different mechanisms that might explain the high level of expression of CAT from the construct pVII-CAT in baker's yeast: a ribosomal shunt, an IRES, and a cryptic promoter. We eliminated the ribosomal shunt and the IRES as possible mechanisms because both would have been dependent on the yeast GAL1 promoter, and elimination of the GAL1 promoter abolished CAT expression. Instead, our mutagenesis studies and Northern analysis revealed that the CaMV sequence had promoter or enhancer activity in yeast.
A BLAST analysis of the yeast genome revealed that the CaMV promoter/enhancer element had sequence homology to the delta element YRDCdelta3, which is located on yeast chromosome IV. Delta elements are solo copies of the LTR of Ty1; they result from a recombination event that occurs between the LTRs of a full-length Ty1 element. As such, delta elements represent a site in the genome where a full-length element had once been inserted. Within the yeast genome, there are 32 full-length copies of Ty1 and 185 solo Ty1 LTRs, either complete versions of the LTR or fragments (27). The BLAST search did not reveal a match between the CaMV sequence and other Ty1 LTRs, but once the initial match was made between the CaMV sORF F sequence and YRDCdelta3, it was possible to align the CaMV sequence with the original sequence of Ty1 (8). This alignment could be extended to the primer binding site for first-strand DNA synthesis for both CaMV and Ty1. In the case of the Ty1 LTR, the R region is located 41 nucleotides upstream from the Met tRNA binding site, whereas in CaMV, the distance is 39 nucleotides.
The portion of CaMV that contributes to the cryptic promoter activity corresponds to the untranslated leader region (ULR) of the Ty1 mRNA (Fig. 7A). Transcription of Ty1 begins at nucleotide position 240 within the LTR (13), and the start codon for the gag ORF is found at nucleotide position 294; the CaMV promoter/enhancer corresponds to approximately 83% of this sequence (Fig. 7A). This portion of the Ty1 LTR has not been reported previously to have enhancer activity. However, it is well established that transcriptional enhancers can be located within ULRs. For example, the copia retroelement of Drosophila contains seven transcriptional enhancer elements within a ULR that is adjacent to its 5' LTR (34, 37). The enhancer elements in the copia ULR have been shown to significantly increase the expression of a minimal heat shock (hsp) promoter element and to bind to the Drosophila CCAAT/enhancer binding protein (DmC/EBP) (37, 62). The copia 5' LTR also enhances the expression of the minimal hsp promoter, but the highest level of expression was attained when both the LTR and ULR were placed upstream from the minimal hsp promoter (37, 62).
In addition, a transcriptional enhancer was recently found within the ULR of CaMV at a point between the 35S RNA transcriptional start site and sORF A (45). It is worthwhile to note that the CaMV 35S promoter was first characterized in 1985 (40) and it is considered to be one of the most extensively studied plant promoters (45), yet the existence of this enhancer element is just coming to light now, perhaps because its presence was masked by other more powerful enhancers upstream from the 35S promoter (45). The discovery of this enhancer in the 35S leader sequence underscores how much there is to be learned about even well-characterized regulatory elements. At a minimum, our work illustrates that sequence alterations in the Ty1 ULR have the capacity to function as a promoter/enhancer element in yeast.
A model for the evolution of the ribosomal shunt mechanism of translation in CaMV from LTRs. So why might a cryptic yeast promoter/enhancer be located within the CaMV large intergenic region It is unlikely that the promoter/enhancer function is responsible for its conservation from yeasts to plants. This sequence lacked promoter/enhancer activity in an agroinfiltration assay (Fig. 6), although we cannot eliminate the possibility that the cryptic promoter might function in plants in the presence of one or more CaMV proteins. Furthermore, the homology of the 42-nucleotide CaMV sequence to the R region in the Ty1 LTR has no discernible function as an R region for CaMV. The comparable region for CaMV, its terminal redundancy in the 35S RNA, is located 338 nucleotides upstream from the 42-nucleotide stretch that is homologous to the Ty1 R region (Fig. 7A). Nonetheless, the 42-nucleotide CaMV sequence does have an essential role in the translation of the CaMV 35S RNA in plants; it encompasses the shunt acceptor site (Fig. 7B) (10, 16, 48). Since neither of the functions attributed to the 42-nucleotide CaMV sequence in yeast are relevant to how CaMV functions in plants, we propose that the sequence has evolved from the R region of an ancestral LTR into the stem-loop structure required for the ribosomal shunt.
Once the base of the right arm of the 35S RNA stem-loop is recognized to be a remnant of an LTR, then other elements of the LTR come into focus to provide an alternate picture of the structure of the 35S RNA leader sequence (Fig. 9). Most importantly, the left arm of the stem-loop can also be recognized as the remnant of the R region of Ty1, although in an opposite orientation to the R region that is represented in the right arm (Fig. 9B). In essence, the CaMV large intergenic region can be redrawn as two LTRs in opposite orientations (Fig. 9C). The R region remnant in the left LTR extends from nucleotide 7510 to 7563. It occupies more sequence than the R remnant in the right LTR because of insertions of foreign sequences over time (Fig. 9B); these insertions contribute a larger bulge loop in the large stem-loop structure (Fig. 7B). The boundaries between U5 and R correspond to the shunt donor and acceptor sites (Fig. 9A and C).
The extreme boundaries of the U3 regions might be predicted to occur within the large stem-loop structure at the top of stem section 3 (15, 48) at nucleotides 7724 and 7790, with 66 nucleotides separating the two LTRs (Fig. 1C and 9C). Although sequences of the two LTRs might change over time, mutations in one arm must be compensated for in the other arm to retain the stem-loop structure for the ribosomal shunt mechanism (47). Indeed, Pooggin et al. (48) concluded that mutations occurred within the 35S RNA leader sequence at "a surprisingly low frequency." The requirement for a functional ribosome shunt could provide one constraint that could limit change within the 35S RNA leader sequence.
The boundary of U5 on the right LTR is defined by the methionine tRNA binding site, which is the primer binding site for first-strand CaMV DNA synthesis during reverse transcription (Fig. 9C) (50). An inspection of the nucleotide sequence which is upstream from the left LTR reveals one attractive candidate for an additional tRNA primer binding site, as the 3' end of a valine tRNA from S. cerevisiae provides a reasonably good fit. Interestingly, the 3' end of this primer binding site is adjacent to the transcriptional start site of the 35S RNA at nucleotide 7435. The U5 region of the left LTR is a little larger than its right counterpart because of the inclusion of two short sequence elements that enhance the transcription of the 35S RNA (45).
The alternate views of the 35S RNA leader sequence presented in Fig. 9A and C both fit the model for the formation of the ribosomal shunt (10, 16, 48). The model in Fig. 9A illustrates the essential elements of the shunt, shunt donor, and acceptor sites as well as the sORFs in the leader sequence. Although the sORFs are a prominent feature of the leader sequence, the elimination of start codons in the leader only delays infectivity rather than abolishing it (48). Consequently, none are essential. Pooggin et al. (48) suggested that the sORFs may play a structural role in maintenance of the stem-loop structure or in positioning ribosomes for the shunt. In contrast, the model in Fig. 9C addresses the derivation of the ribosomal shunt mechanism; it suggests that the shunt is derived from two LTRs positioned in opposite directions. This model is consistent with the concept that sORFs might contribute to the stem-loop structure rather than encoding peptides that contribute to CaMV infections.
The structure of the CaMV large intergenic region illustrated in Fig. 9C implies that a CaMV progenitor may have served as a landing pad for other retrotransposons such that two LTRs would be juxtaposed next to the 35S promoter. There are, in fact, precedents for this arrangement of sequences. It is well established that retrotransposons of all types can be found in clusters (5, 52), and Ty1 is no exception. In a survey of transposable elements in the yeast genome, Kim et al. (27) documented 16 compound insertions in which retrotransposons had inserted within other retrotransposons. In addition, they found other cases in which Ty elements were inserted adjacent to each other and in opposite orientations. In one case, the retroelements were separated by only 10 base pairs (27).
It cannot be determined whether the tandem LTRs present in the CaMV large intergenic region are derived from a fungal or plant progenitor, although either scenario is possible. Ty1-copia-type retroelements are found in all plants (46, 59, 61), and they could have served as a source of the Ty1 LTR remnants in CaMV. Although pararetroviruses do not become integrated into their hosts as a part of their infection cycle, integrated copies of some pararetroviruses have been found in plant genomes (19, 26, 31, 49). Consequently, the LTR remnants in the CaMV large intergenic region might have arisen through recombination with a plant retroelement, at a time when a CaMV progenitor may have existed as an integrated copy in its host genome. However, attempts to align the CaMV LTR remnant with Ty1/copia-like and Ty3/Gypsy retroelements present in Arabidopsis (a total of 1,447 retroelements) failed to reveal any conservation (data not shown). Consequently, it does not appear that the CaMV LTR remnant is derived from either of these sources. In contrast, since the CaMV LTR remnants have homology with a yeast retroelement, they might have arisen from a mycorrhizal or endophytic fungus. Some fungi are capable of vectoring viruses into plants (22), although pararetroviruses have not been documented to be spread in this manner (22).
Mobile elements and segments of repetitive DNA have been documented to play a role in the evolution of their hosts through their insertion into and adjacent to host genes (6, 36, 61). Britten (6, 7) has compiled several examples that have been found in animals and insects and has suggested three criteria for inclusion in this group: (i) that the insertion occurred far in the past and that it is not a transient mutation; (ii) that the element is derived from recognizable group of similar sequences; and (iii) that the element serves a useful function. We propose that the LTR remnants that we have identified in the CaMV large intergenic region should be included in this group. Although we cannot establish a date for when the LTRs were inserted into the CaMV genome, they must be as old as CaMV itself, as they are an integral component of the CaMV genome. The elements themselves are related to the Ty1-copia class of retrotransposons, a class that is ubiquitous in plants (59, 61). Finally, the LTR remnants form the base of the large stem-loop structure that forms within the CaMV 35S RNA leader sequence; they are a structural requirement for the ribosomal shunt (10, 16, 48). The insertion of elements adjacent to each other and in opposite orientations may represent an economical way in evolution to form a sizable RNA local secondary structure, which could be much more efficient than the coevolution of two independent sequences to form the arms of a local secondary structure. This mechanism might appear in other genomes and for other types of RNA local structural motifs.
ACKNOWLEDGMENTS
We thank Barbara Wiggins for technical assistance and John Cannon for help with yeast strains and techniques.
This research was supported by the Missouri Agricultural Experiment Station, the Food for the 21st Century program at the University of Missouri, and the Research Board at the University of Missouri. The research of X.-F.W. and D.X. was supported by the U.S. Department of Energy's Genomes to Life program (www.doegenomestolife.org) under the project "Carbon Sequestration in Synechococcus sp.: from Molecular Machines to Hierarchical Modeling" (www.genomes-to-life.org).
REFERENCES
Altschul, S. F., W. Gish, W. Miller, E. W. Myers, and D. J. Lipman. 1990. Basic local alignment search tool. J. Mol. Biol. 215:403-410.
Ausubel, F., R. Brent, R. Kingstone, D. Moore, D. Seidman, J. G. Smith, and K. Struhl. 1995. Short protocols in molecular biology, 3rd ed. John Wiley and Sons, New York, N.Y.
Balakrishnan, R., K. R. Christie, M. C. Costanzo, K. Dolinski, S. S. Dwight, S. R. Engel, D. G. Fisk, J. E. Hirschman, E. L. Hong, R. Nash, R. Oughtred, M. Skrzypek, C. L. Theesfeld, G. Binkley, C. Lane, M. Schroeder, A. Sethuraman, S. Dong, S. Weng, S. Miyasato, R. Andrada, D. Botstein, and J. M. Cherry. Saccharomyces genome database. http://www.yeastgenome.org/. [Online.] Accessed 7 March 2005.
Boeke, J. D., and S. B. Sandmeyer. 1991. Yeast transposable elements, p. 193-261. In J. Broach, E. Jones, and J. Pringle (ed.), The molecular and cellular biology of the yeast Saccharomyces, vol. 1. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, N.Y.
Brandes, A., J. S. Heslop-Harrison, A. Kamm, S. Kubis, R. L. Doudrick, and T. Schmidt. 1997. Comparative analysis of the chromosomal and genomic organization of Ty1-copia-like retrotransposons in pteridophytes, gymnosperms and angiosperms. Plant Mol. Biol. 33:11-21.
Britten, R. J. 1996. DNA sequence insertion and evolutionary variation in gene regulation. Proc. Natl. Acad. Sci. USA 93:9374-9377.
Britten, R. J. 1997. Mobile elements inserted in the distant past have taken on important functions. Gene 205:177-182.
Clare, J., and P. Farabaugh. 1995. Nucleotide sequence of a yeast Ty element: evidence for an unusual mechanisms of gene expression. Proc. Natl. Acad. Sci. USA 82:2829-2833.
Daubert, S. D., and G. Routh. 1990. Point mutations in cauliflower mosaic virus gene VI confer host-specific symptom changes. Mol. Plant-Microbe Interact. 3:341-345.
Dominguez, D. I., L. A. Ryabova, M. M. Poogin, W. Schmidt-Puchta, J. Fütterer, and T. Hohn. 1998. Ribosome shunting in cauliflower mosaic virus. Identification of an essential and sufficient structural element. J. Biol. Chem. 273:3669-3676.
Eddy, E. R. 1998. Profile hidden Markov models. Bioinformatics 14:755-763.
Eddy, E. R. 1998. Multiple alignment and multiple sequence based searches. Cold Spring Harbor computational genomics course handout. [Online.] ftp://ftp.genetics.wustl.edu/pub/eddy/papers/tigs-9808.pdf. Cold Spring Harbor Laboratory, Cold Spring Harbor, N.Y.
Elder, R. T., E. Y. Loh, and R. W. Davis. 1983. RNA from the yeast transposable element Ty1 has both ends in the direct repeats, a structure similar to retrovirus RNA. Proc. Natl. Acad. Sci. USA 80:2432-2436.
Froissart, R., M. Uzest, V. Ruiz-Ferrer, M. Drucker, E. Hebrard, T. Hohn, and S. Blanc. 2004. Splicing of Cauliflower mosaic virus 35S RNA serves to downregulate a toxic gene product. J. Gen. Virol. 85:2719-2726.
Fütterer, J., K. Gordon, J. M. Bonneville, H. Sanfaon, B. Pisan, J. R. Penswick, and T. Hohn. 1988. The leading sequence of cauliflower mosaic virus large RNA can be folded into a large stem-loop structure. Nucleic Acids Res. 16:8377-8390.
Fütterer, J., Z. Kiss-Laszlo, and T. Hohn. 1993. Nonlinear ribosome migration on cauliflower mosaic virus 35S RNA. Cell 73:789-802.
Gardner, R. C., A. J. Howarth, P. Hahn, M. Brown-Leudi, R. J. Shepherd, and J. Messing. 1981. The complete nucleotide sequence of an infectious clone of cauliflower mosaic virus by M13mp7 shotgun sequencing. Nucleic Acids Res. 9:2871-2888.
Haas, M., A. Geldreich, M. Bureau, L. Dupuis, V. Leh, G. Vetter, K. Kobayashi, T. Hohn, L. Ryabova, P. Yot, and M. Keller. 2005. The open reading frame VI product of Cauliflower mosaic virus is a nucleocytoplasmic protein: its N terminus mediates its nuclear export and formation of electron-dense viroplasms. Plant Cell 17:927-943.
Harper, G., J. O. Osuji, J. S. Heslop-Harrison, and R. Hull. 1999. Integration of banana streak badnavirus into the Musa genome: molecular and cytogenetic evidence. Virology 255:207-213.
Hemmings-Mieszczak, M., G. Steger, and T. Hohn. 1997. Alternative structures of the cauliflower mosaic virus 35S RNA leader: implications for viral expression and replication. J. Mol. Biol. 267:1075-1088.
Hirt, H., M. Kgl, T. Murbacher, and E. Heberle-Bors. 1990. Evolutionary conservation of transcriptional machinery between yeast and plants as shown by the efficient expression from the CaMV 35S promoter and 35S terminator. Curr. Genet. 17:473-479.
Hull, R. 2002. Matthews' plant virology, 4th ed. Academic Press, San Diego, Calif.
Irniger, S., H. Sanfacon, C. H. Egli, and G. H. Braus. 1992. Different sequence elements are required for function of the cauliflower mosaic virus polyadenylation site in Saccharomyces cerevisiae compared with in plants. Mol. Cell. Biol. 12:2322-2330.
Ito, H., Y. Fukuda, K. Murata, and A. Kimura. 1983. Transformation of intact yeast cells treated with alkali cations. J. Bacteriol. 153:163-168.
Jacq, C., J. Alt-Mrbe, B. Andre, W. Arnold, A. Bahr, et al. 1997. The nucleotide sequence of Saccharomyces cerevisiae chromosome IV. Nature 387:75-78.
Jakowitsch, J., M. F. Mette, J. van der Winden, M. A. Matzke, and A. J. M. Matzke. 1999. Integrated pararetroviral sequences define a unique class of dispersed repetitive DNA in plants. Proc. Natl. Acad. Sci. USA 96:13241-13246.
Kim, J. M., S. Vanguri, J. D. Boeke, A. Gabriel, and D. F. Voytas. 1998. Transposable elements and genome organization: a comprehensive survey of retrotransposons revealed by the complete Saccharomyces cerevisiae genome sequence. Genome Res. 8:464-478.
Kiss-László, Z., J. Fütterer, and T. Hohn. 1995. Splicing of cauliflower mosaic virus 35S RNA is essential for viral infectivity. EMBO J. 14:3552-3562.
Kozak, M. 1999. Initiation of translation in prokaryotes and eukaryotes. Gene 234:187-208.
Kushner, D. B., B. D. Lindenbach, V. Z. Grdzelishvili, A. O. Noueiry, S. M. Paul, and P. Ahlquist. 2003. Systematic, genome-wide identification of host genes affecting replication of a positive-strand RNA virus. Proc. Natl. Acad. Sci. USA 100:15764-15769.
Lockhart, B. E., J. Menke, G. Dahal, and N. E. Olszewski. 2000. Characterization and genomic analysis of tobacco vein clearing virus, a plant pararetrovirus that is transmitted vertically and related to sequences integrated in the host genome. J. Gen. Virol. 81:1579-1585.
Lowe, T. M., and S. R. Eddy. 1997. tRNAscan-SE: A program for improved detection of transfer RNA genes in genomic sequence. Nucleic Acids Res. 25:955-964.
Maniatis, T., E. F. Fritsch, and J. Sambrook. 1982. Molecular cloning: a laboratory manual. Cold Spring Harbor Laboratory, Cold Spring Harbor, N.Y.
Matyunina, L. V., I. K. Jordan, and J. F. MacDonald. 1997. Naturally occurring variation in copia expression is due to both element (cis) and host (trans) regulatory variation. Proc. Natl. Acad. Sci. USA 93:7097-7102.
McCarthy, J. E. G. 1998. Posttranscriptional control of gene expression in yeast. Microbiol. Mol. Biol. Rev. 62:1492-1553.
McDonald, J. F. 1993. Evolution and consequences of transposable elements. Curr. Opin. Genet. Dev. 3:855-864.
McDonald, J. F., L. V. Matyunina, S. Wilson, I. K. Jordan, N. J. Bowen, and W. J. Miller. 1997. LTR retrotransposons and the evolution of eukaryotic enhancers. Genetica 100:3-13.
Messing, J. 1979. A multiple cloning system based on single-stranded DNA bacteriophage M13. Recomb. DNA Tech. Bull. 2:43-48.
Noueiry, A. O., and P. Ahlquist. 2003. Brome mosaic virus RNA replication: revealing the role of the host in RNA virus replication. Annu. Rev. Phytopathol. 41:77-98.
Odell, J. T., F. Nagy, and N.-H. Chua. 1985. Identification of DNA sequences required for activity of the cauliflower mosaic virus 35S promoter. Nature 313:810-812.
Palanichelvam, K., and J. E. Schoelz. 2002. A comparative analysis of the avirulence and translational transactivator functions of gene VI of cauliflower mosaic virus. Virology 293:225-233.
Panavas, T., and P. D. Nagy. 2003. Yeast as a model host to study replication and recombination of defective interfering RNA of Tomato bushy stunt virus. Virology 314:315-325.
Panavas, T., E. Serviene, J. Brahser, and P. D. Nagy. 2005. Yeast genome-wide screen reveals dissimilar sets of host genes affecting replication of RNA viruses. Proc. Natl. Acad. Sci. USA 102:7326-7331.
Park, H. S., A. Himmelbach, K. S. Browning, T. Hohn, and L. A. Ryabova. 2001. A plant viral "reinitiation" factor interacts with the host translational machinery. Cell 106:723-733.
Pauli, S., H. M. Rothnie, G. Chen, X. He, and T. Hohn. 2004. The cauliflower mosaic virus 35S promoter extends into the transcribed region. J. Virol. 78:12120-12128.
Peterson-Burch, B. D., and D. F. Voytas. 2002. Genes of the Pseudoviridae (Ty1/copia retrotransposons). Mol. Biol. Evol. 19:1832-1845.
Pfeiffer, P., and T. Hohn. 1983. Involvement of reverse transcription in the replication of cauliflower mosaic virus: a detailed model and test of some aspects. Cell 33:781-789.
Pooggin, M. M., T. Hohn, and J. Fütterer. 1998. Forced evolution reveals the importance of short open reading frame A and secondary structure in the cauliflower mosaic virus 35S RNA leader. J. Virol. 72:4157-4169.
Richert-Pggeler, K. R., and R. J. Shepherd. 1997. Petunia vein clearing virus: a plant pararetrovirus with the core sequences for an integrase function. Virology 236:137-146.
Rothnie, H., Y. Chapdelaine, and T. Hohn. 1994. Pararetroviruses and retroviruses: a comparative review of viral structure and gene expression strategies. Adv. Virus Res. 44:1-67.
Ryabova, L. A., M. M. Pooggin, and T. Hohn. 2002. Viral strategies of translation initiation: ribosomal shunt and reinitiation. Prog. Nucleic Acid Res. Mol. Biol. 72:1-39.
SanMiguel, P., A. Tikhonov, Y.-K. Jin, N. Motchoulskaia, D. Zakharov, A. Melake-Berhan, P. S. Springer, K. J. Edwards, M. Lee, Z. Avramova, and J. L. Bennetzen. 1996. Nested retrotransposons in the intergenic regions of the maize genome. Science 274:765-768.
Schardl, C. L., A. D. Byrd, G. Benzion, M. A. Altschuler, D. F. Hildebrand, and A. G. Hunt. 1987. Design and construction of a versatile system for the expression of foreign genes in plants. Gene 61:1-11.
Schneider, J. G., and L. Guarente. 1991. Vectors for expression of cloned genes in yeast. Methods Enzymol. 194:373-388.
Schoelz, J. E., R. J. Shepherd, and R. D. Richins. 1986. Properties of an unusual strain of cauliflower mosaic virus. Phytopathology 76:451-454.
Sha, Y., E. P. Broglio, J. F. Cannon, and J. E. Schoelz. 1995. Expression of a plant viral polycistronic mRNA in yeast, Saccharomyces cerevisiae, mediated by a plant virus translational transactivator. Proc. Natl. Acad. Sci. USA 92:8911-8915.
Thomas, P. S. 1980. Hybridization of denatured RNA and small DNA fragments transferred to nitrocellulose. Proc. Natl. Acad. Sci. USA 77:5201-5205.
Thompson, J. D., D. G. Higgins, and T. J. Gibson. 1994. CLUSTAL W: improving the sensitivity of progressive multiple sequence alignment through sequence weighting, position-specific gap penalties and weight matrix choice. Nucleic Acids Res. 22:4673-4680.
Voytas, D. F., M. P. Cummings, A. Konieczny, F. M. Ausubel, and S. R. Rodermel. 1992. Copia-like retroelements are ubiquitous among plants. Proc. Natl. Acad. Sci. USA 89:7124-7128.
Wan, X.-F., and D. Xu. Computational methods for remote homolog identification. Curr. Protein. Pept. Sci., in press.
White, S. E., L. F. Habera, and S. R. Wessler. 1994. Retrotransposons in the flanking regions of normal plant genes: a role for copia-like elements in the evolution of gene structure and expression. Proc. Natl. Acad. Sci. USA 91:11792-11796.
Wilson, S., L. V. Matyunina, and J. F. MacDonald. 1998. An enhancer region within the copia untranslated leader contains binding sites for Drosophila regulatory proteins. Gene 209:239-246.
Zuker, M. 2003. Mfold web server for nucleic acid folding and hybridization prediction. Nucleic Acids Res. 31:3406-3415.(Monir Shababi, June Bourq)