![[Feedback]](/icons/banner/feedbackACT.gif){kind=icon}
![[For Subscribers]](/icons/banner/subscriberACT.gif){kind=icon}
![[Archive]](/icons/banner/archiveACT.gif){kind=icon}
![[Search]](/icons/banner/searchACT.gif){kind=icon}
![[Contents]](/icons/banner/tocACT.gif){kind=icon}
Previous Article | Next Article 
Journal of Virology, October 2006, p. 10045-10054, Vol. 80, No. 20
0022-538X/06/$08.00+0 doi:10.1128/JVI.00991-06
Copyright © 2006, American Society for Microbiology. All Rights Reserved.
,
Aurélie M. Rakotondrafara,2,3, and
W. Allen Miller1,2,3,4*
Interdepartmental Genetics Program,1 Molecular, Cellular and Developmental Biology Program,2 Plant Pathology Department,3 Biochemistry, Biophysics and Molecular Biology Department, Iowa State University, Ames, Iowa 500114
Received 13 May 2006/ Accepted 24 July 2006
 |
ABSTRACT |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
INTRODUCTION |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Many families of positive-strand RNA viruses produce nested subgenomic mRNAs during infection (36, 43). These sgRNAs have the same 3' ends as genomic RNA but have 5' truncations or deletions relative to the genomic RNA. This places open reading frames (ORFs) that are 5' distal on genomic RNA near the 5' ends of sgRNAs, allowing the sgRNAs to serve as mRNAs for translation of these downstream ORFs. Examples of viral pathogens that produce sgRNAs are severe acute respiratory syndrome coronavirus (24), equine arteritis virus (53), Sindbis virus (17), rubella virus (40), tobacco mosaic virus (20), citrus tristeza virus (19), and barley yellow dwarf virus (BYDV) (28, 29). While the control and mechanism of synthesis of the subgenomic RNAs of these and related viruses have been studied (36, 43), little is known about how translation of the viral genomic and subgenomic RNAs is coordinated as they accumulate in the cell (17, 40). The competition among viral RNAs for the host translation machinery and regulation of their translation are likely key control points in viral-gene expression necessary for a successful infection. Here, we provide evidence that translation of BYDV genomic RNA and a subgenomic RNA is regulated in trans by a second, specialized viral sgRNA.
BYDV, a major pathogen of wheat and other cereal crops (32), has a positive-sense RNA genome of 5,677 nucleotides (nt) that encodes six ORFs (Fig. 1) (37). Three 3'-coterminal sgRNAs are generated in infected cells. They are not encapsidated and thus are absent at the initial stage of infection. sgRNA1 is the mRNA for the coat protein (CP) (ORF 3), a putative movement protein (ORF 4), and a C-terminal extension of the coat protein required for aphid transmission (ORF 5). Highly abundant sgRNA2 harbors a small ORF (ORF 6) that encodes a predicted polypeptide of 4.3 to 7.2 kDa, depending on the isolate. The sequence of ORF 6 is poorly conserved among BYDV isolates (9), and it is absent in other members of the genus Luteovirus (13, 49). The product of ORF 6 has not been detected in infected cells, although sgRNA2 can be translated in vitro (55). sgRNA3 comprises the 3'-terminal 332 nt of the BYDV genome (26). It accumulates sporadically, encodes no ORF, and has no known function.
|
In addition to conferring cap-independent translation in cis, the BTE inhibits translation of viral genes and nonviral reporter genes in trans (54, 55). In wheat germ extract, sgRNA2, which harbors the BTE at its 5' end, trans inhibits translation of gRNA more than that of sgRNA1 (55). The inhibition does not require translation of ORF 6, but it requires a functional BTE in sgRNA2 (55). Premature addition of sgRNA2 at the initial moment of infection strongly inhibits BYDV RNA accumulation (50). Based on these data, the following trans regulation model of gene expression was proposed (Fig. 1). Early in BYDV infection, only the replicase genes, ORF 1 and ORF 2 (translated as a fusion with ORF 1 by ribosomal frameshifting) (4), are translated via BTE-mediated cap-independent translation in cis. Once replicase is produced, viral RNA is replicated, and gRNA and sgRNAs accumulate. The highly abundant sgRNA2 would selectively inhibit translation of gRNA relative to sgRNA1 in trans. Structural and movement proteins would then be preferentially translated from sgRNA1, and replicase expression would be shut off. Thus, the BYDV viral life cycle would switch from an early to a late stage of gene expression (Fig. 1) (55). Here, we provide evidence that strongly supports the notion that these events occur in virus-infected cells. Moreover, we show that the different translation efficiencies of the BYDV mRNAs in the presence of sgRNA2 are determined by the proximity to the 5' end of the mRNA of the stem-loop structure that base pairs to the 3' translation element to allow cap-independent translation. These observations reveal a novel translational control mechanism by a trans-regulatory RNA and a new function for a viral subgenomic RNA.
|
MATERIALS AND METHODS |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
SG2, was described previously
(28) as SG2G/C. It
differs from pPAV6 by a G-to-C mutation at position 4810, which
prevents sgRNA2 synthesis. pSG2 and pSG2BF allow T7 transcription of
sgRNA2 RNA and its mutant sgRNA2BF, respectively
(55). sgRNA2BF contains a
GAUC duplication at the BamHI site (BF) of sgRNA2 that destroys the in
vitro trans inhibition function of sgRNA2
(55). Clone pGfLUC was described previously (22) as p5'UTR-LUC-TE869-(A)60. GfLUC is the gRNA reporter transcript derived from SmaI-linearized pGfLUC by in vitro T7 transcription. It encodes the firefly luciferase ORF flanked by the UTRs of BYDV. pRenilla-CP393 was cloned by replacing nt 2843 to 4565 of pPAV6 with the Renilla ORF of pRluc (Promega, Madison, WI). pSG1rLUC was cloned by ligating the Bst1107I-BsmI fragment of pRenilla-CP393 into Bst1107I/BsmI-cut pSG1, which was described previously (29). SG1rLUC is the sgRNA1 T7 transcript from SmaI-cut pSG1rLUC. It has the same 5' UTR as sgRNA1, except for the omission of 14 bases at the extreme 3' end. pSG1fLuc was constructed by replacing the 5' UTR of pGfLUC with the 5'UTR of pSG1rLUC.
PAV6-FLAG and sgRNA2-FLAG were constructed by inserting a FLAG tag (amino acid sequence, DYKDDDDK) at the 3' end of ORF 6 in pPAV6 and pSG2, respectively. The FLAG tag insert was created by a three-step PCR approach by using primers bearing a FLAG tag and pSG1 as templates, as described previously (28). The final PCR fragment was cloned into Acc65I-SmaI-digested pPAV6 or NotI-SmaI-digested pSG2. sgRNA2-LIII CS RNA was in vitro transcribed from a PCR-generated template corresponding to the sgRNA2 sequence. The PCR fragment was amplified using LUC 869-LIII-CS as a template (48) and a forward primer bearing a T7 promoter sequence and a reverse primer corresponding to the 3' end of sgRNA2.
pA-GfLUC contained a C15A mutation in the loop of stem-loop A (SL-A)and a GAC-to-CUG mutation within the loop of stem loop D (bases 105 to 107) (see Fig. 5A). First, a BYDV 5'-UTR PCR fragment with a GAC-to-CUG mutation within the loop of stem-loop D was generated by three-step PCR and cloned into NotI-BssHI-digested pGfLUC. The clone obtained was then used as a PCR template to generate the additional C15A mutation. pA-GfLUC was then cloned by ligating the NotI-BssHI-digested PCR fragment into NotI-BssHI-cut pGfLUC. pD-SG1fLUC contained an A10C mutation in the loop of stem-loop A and an AGUUA-to-CUGACAA mutation of the loop of stem-loop D (bases 111 to 115) (see Fig. 5A). pD-SG1fLUC was generated by PCR base mutagenesis of SG1fLUC cDNA using the Gene Tailor Site-Directed Mutagenesis System (Invitrogen). All constructs were verified by automated sequencing at the Nucleic Acid Facility of Iowa State University.
|
Western immunoblotting. Western blotting was performed according to the Amersham Pharmacia Biotech protocol. Total protein from inoculated protoplasts was separated on a 10% polyacrylamide gel and transferred to a polyvinylidene difluoride membrane. After being blocked overnight with phosphate-buffered saline-Tween buffer (PTB) containing 5% dried low-fat milk, the membrane was incubated with the primary anti-FLAG antibodies (Sigma, St. Louis, MO) in a 1:25,000 dilution in PTB for 2 h, with a 1:600 dilution of fluorescein-conjugated secondary antibodies for 1 h, and then with a 1:2,500 dilution of alkaline phosphatase-conjugated anti-fluorescein antibodies for another hour. The membrane was washed three times with PTB after each incubation. After the final incubation with attophose substrate with a volume of 24 µl/cm2 for less than 20 min, the membrane was air dried and scanned on a STORM 840 chemiluminescence imager (Molecular Dynamics).
In vivo translation. Oat protoplasts were prepared and electroporated with RNA as described previously (12). For the two-step electroporation method, the voltage was reduced to 280 V. For the in vivo translation assay of the SG2fLUC reporter construct, we included a capped, polyadenylated Renilla luciferase reporter RNA as an internal control to normalize electroporation variation. The Renilla luciferase ORF was flanked by the 5' and 3' UTRs of the firefly luciferase gene from pGEMLUC (Promega). Luciferase activity was measured 4 h after electroporation.
In the two-step electroporation, oat protoplasts
were inoculated with 1 pmol of infectious BYDV PAV6 or PAV6GS2
RNA and incubated for 24 h at room temperature prior to the
second electroporation to allow viral replication and sgRNA
accumulation. When transfected directly with the nonreplicative sgRNA2
or sgRNA2BF RNA transcripts, the cells were incubated for 4 h
prior to the second electroporation. In the second step, protoplasts
were inoculated again with 1 to 2 pmol of GfLUC (A-GfLUC), SG1rLUC, or
both, as indicated. In all cases, firefly luciferase and
Renilla luciferase were analyzed 4 h after the
second electroporation, as described previously
(51), and the Promega
(Madison, WI) Stop-N-Glo system was used to assay both luciferase
activities.
Northern blot analysis. Total RNAs were extracted from oat protoplasts 24 h postinoculation (p.i.) by using Trizol reagent (Invitrogen, Carlsbad, CA) according to the manufacturer's instructions. The RNAs were then analyzed by Northern blotting as described previously (29). A 32P-labeled probe complementary to the 1.5-kb 3' end of BYDV-PAV genome RNA was used to detect BYDV gRNA and sgRNAs (29).
|
RESULTS |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
To more closely mimic natural infection, gRNA reporter GfLUC, sgRNA1 reporter SG1rLUC, and sgRNA2 were mixed simultaneously in the same wheat germ translation reaction mixture at different ratios. Again, sgRNA2 inhibited translation of GfLUC much more than that of SG1rLUC (Fig. 2D). The difference in translation inhibition by sgRNA2 of GfLUC and SG1rLUC reporters was greater when GfLUC, SG1rLUC, and sgRNA2 were added together in the same translation reaction (compare Fig. 2B and D). The GfLUC/SG1rLUC (gRNA/sgRNA1) expression ratio decreased as the sgRNA2 concentration was increased (Fig. 2E). These results resemble previous observations when gRNA, sgRNA1, and sgRNA2 were all added to a wheat germ translation extract at ratios approximating those in infected cells. In that experiment, gRNA was inhibited by 99% while sgRNA1 was inhibited by only 65%, allowing it to remain a relatively efficient message (55). Thus, the reporter RNAs provide valid representations of the translation of the actual viral RNAs in the presence of sgRNA2. Importantly, these results also reveal that no ORFs or gene products of gRNA or sgRNA1 are necessary for the differential inhibition of translation.
Differential inhibition of GfLUC and SG1rLUC translation in virus-infected cells.
Having validated that reporter
constructs GfLUC and SG1rLUC translate like gRNA and sgRNA1 in the
presence or absence of sgRNA2 in vitro, we tested the trans
regulation model in oat protoplasts by measuring the translation
efficiencies of the two reporter constructs in the presence of
replicating BYDV RNA. We employed a two-step electroporation method
(Fig.
3A) (50). First, oat
protoplasts were transfected by electroporation with the infectious
transcript of the BYDV genome, PAV6, or with transcript
PAV6SG2. PAV6SG2 RNA has a G4810C point mutation that
prevents sgRNA2 synthesis but still permits genomic-RNA replication
(28). After a 24-hour
incubation to allow viral replication and sgRNA accumulation (Fig.
3B), the protoplasts were
electroporated again, this time with GfLUC or SG1rLUC reporter RNA.
Then after another 4-hour incubation to allow translation of these
RNAs, firefly luciferase and Renilla luciferase activities
were measured (Fig. 3C and
D).
|
SG2 (Fig.
3B). After the second
electroporation to introduce reporter mRNA, we observed that the
presence of replicating PAV6 RNA caused nearly an 80% drop in the
translation of the gRNA reporter GfLUC compared to uninfected cells but
caused only a 20% reduction in the translation of the sgRNA1 reporter
SG1rLUC (Fig. 3C, PAV6).
In cells transfected with PAV6SG2 RNA, translations of
GfLUC and SG1rLUC (Fig.
3C, PAV6SG2) were
reduced by only 40% and less than 5%, respectively. Thus, infection
with PAV6SG2 (which makes no sgRNA2) inhibited the translation
of GfLUC less than did wild-type PAV6 RNA.
When GfLUC
and SG1rLUC were coelectroporated in the second step in cells
previously transfected with PAV6 RNA, GfLUC translation dropped by 88%,
while translation of SG1rLUC actually increased slightly, compared to
that of uninfected cells (Fig.
3D, PAV6). SG1rLUC
translated at least 10 times more efficiently than GfLUC relative to
uninfected controls. Thus, the differential inhibition effects of the
replicating PAV6 on translation of GfLUC and SG1rLUC were
greater when both reporter RNAs were present simultaneously.
PAV6SG2 had a less inhibitory effect on translation of GfLUC
than did wild-type PAV6 infection (Fig.
3D, PAV6SG2),
strengthening the concept that sgRNA2 is the major influence on the
differential inhibition of GfLUC and SG1rLUC translation. The 40%
inhibition of GfLUC translation in the presence of
replicating PAV6SG2 likely results from the presence of the
BTE at the 3' ends of gRNA and sgRNA1, which, being much less
abundant than sgRNA2, would be expected to inhibit reporter gene
expression to a lesser extent, as was
observed.
BYDV sgRNA2 alone selectively trans inhibits translation in vivo. It is possible that the differential inhibition of reporter gene expression in infected cells is controlled by viral proteins or by host proteins whose expression or function is affected by viral infection, as well as by sgRNA2. To avoid the complicating effects of a viral infection, we observed the effects of sgRNA2 or sgRNA2BF RNA alone on GfLUC or SG1rLUC translation in cells in another two-step electroporation assay. Oat protoplasts were first electroporated with either sgRNA2 or sgRNA2BF transcripts or mock transfected. Four hours later, the same protoplasts were electroporated with GfLUC and SG1rLUC RNAs. Both luciferase activities were measured after another 4-hour incubation to allow translation of the reporter RNAs. In cells preelectroporated with sgRNA2, GfLUC translation dropped by 70% compared to that in cells initially mock transfected or preelectroporated with sgRNA2BF (Fig. 4). In contrast, the presence of sgRNA2 caused only a 10% reduction in translation of SG1rLUC RNA. These data show that sgRNA2 alone is sufficient to mediate selective trans inhibition of gRNA. It requires a functional BTE but does not require viral replication or infection in vivo.
|
The 5' UTRs of gRNA and sgRNA1 determine the differential trans inhibition effects. The previous results led us to wonder what property of the viral RNA determines differential inhibition. Because neither the ORFs nor the 3' UTRs affect inhibition by sgRNA2 (Fig. 2C), we conclude that features of the different 5' UTRs determine the ability to be inhibited in trans. A striking difference between the two 5' UTRs is that the BCL is in the 5'-proximal stem-loop in sgRNA1 but in the fourth stem-loop (SL-D) from the 5' end in the gRNA (Fig. 6A). The loop of SL-D is located 104 to 109 nt from the 5' end of the gRNA, whereas the sgRNA1 BCL is just 10 nt from the 5' end of its RNA (Fig. 6A, sg1SL-A). To determine the effect of the distance of the BCL from the 5' end on cap-independent translation in the presence of sgRNA2, we engineered GfLUC so that the 5'-proximal stem-loop of its 5' UTR (SL-A) was able to kiss (base pair to) the BTE, and the natural kissing bases of SL-D were mutated so that it could no longer interact with the 3' BTE (Fig. 6B). This construct, A-GfLUC, translates at an efficiency similar to that of GfLUC (48). To determine whether this relocation of the BCL to the 5'-proximal stem-loop of the gRNA influences the selectivity of trans inhibition by sgRNA2, we tested the translation efficiency of A-GfLUC relative to GfLUC and SG1fLUC in the presence of sgRNA2 in wheat germ extract (Fig. 6C). A-GfLUC behaved similarly to SG1fLUC in the presence of various ratios of excess sgRNA2 and thus was inhibited less by sgRNA2 than was GfLUC (Fig. 6C).
|
To further analyze the effect of the position of the BCL on selective trans inhibition by sgRNA2, we moved the BCL of SG1fLUC 100 nt downstream to the fourth stem-loop (sg1SL-D) (Fig. 6A) from the 5' end (construct D-SG1fLUC) (Fig. 6B). We tested the translation efficiency of D-SG1fLUC relative to SG1fLUC and GfLUC in the presence of 20-fold-excess sgRNA2 and sgRNA2BF in wheat germ extract (Fig. 6E). While the 5'-distal position of the BCL in D-SG1fLUC did not affect cap-independent translation in the absence of sgRNA2, D-SG1fLUC was inhibited to a level similar to that of GfLUC in the presence of sgRNA2. As expected, sgRNA2BF did not inhibit the translation of either reporter RNA (Fig. 6E). Taken together, the effects of moving the BCL in either gRNA or sgRNA1 reporters demonstrate that the proximity of the kissing loop to the 5' end of the RNA is the major determinant of susceptibility to inhibition of translation by sgRNA2.
Lack of a role for sgRNA2 as an mRNA. We next determined whether sgRNA2 functions as an mRNA in cells. sgRNA2 encodes a small, poorly conserved ORF (ORF 6) that is translatable in vitro (55). Unlike gRNA and sgRNA1, sgRNA2 harbors the BTE in its 5' UTR. No significant sequence complementarity between the 3' and 5' UTRs is predicted in sgRNA2. Thus, sgRNA2 is unlikely to be circularized by 5'-3' base pairing. Because circularization of eukaryotic mRNAs is generally required for efficient translation in vivo but not in vitro (35, 45), we speculated that sgRNA2 may not be translatable in vivo (in protoplasts).
To test the translatability of sgRNA2 in vivo, we fused a FLAG tag to ORF 6 and attempted to detect the P6-FLAG fusion using anti-FLAG antibodies (Fig. 7A). In immunoblots of total protein from cells inoculated with the infectious transcript of the BYDV genome bearing the FLAG tag fusion (PAV6-FLAG), no protein corresponding to P6-FLAG was detected (Fig. 7B), despite the production of abundant sgRNA2 by the PAV6-FLAG virus (Fig. 7C). To determine whether ORF 6 could be translated directly from sgRNA2 in the absence of viral replication, transcript corresponding to sgRNA2 with a FLAG tag (sgRNA2-FLAG) was electroporated into oat protoplasts. As a positive control, this RNA was translated in vitro and P6-FLAG was detected by Western blotting (Fig. 7B). The immunoblot of total protein from protoplasts electroporated with sgRNA2-FLAG revealed no detectable P6-FLAG, even after varied times of sample collection, long exposures, or heavy gel loading (Fig. 7B and data not shown). Thus, sgRNA2 is either not translated or its product (P6) is highly unstable in cells. To test the latter possibility, the in vitro translation product of sgRNA2-FLAG was spiked in an extract of uninfected cells prepared by the same method that was used for attempted detection of P6-FLAG from electroporated cells. The cell extract affected mobility of in vitro-translated P6-FLAG, but P6-FLAG was clearly stable enough to be detected (Fig. 7B). Thus, the cell extract does not degrade FLAG-tagged P6 significantly.
|
|
DISCUSSION |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Interestingly, the difference in inhibition by sgRNA2 of translation of gRNA and sgRNA1 reporters was greater in vivo than in vitro. Moreover, these differences were augmented when GfLUC, SG1rLUC, and sgRNA2 were all present simultaneously in vitro and in vivo (compare Fig. 2B and C, Fig. 3C and D). This mixture most closely mimics natural infection and reveals a level of gene regulation in which BYDV RNAs are well coordinated.
A new mechanism of subgenomic-mRNA gene expression control. The data strongly support the trans regulation model in Fig. 1. Early in BYDV infection, when sgRNAs are absent, only ORF 1 and ORF 2 (replicase genes) are translated via BTE-mediated, cap-independent translation to express the RNA-dependent RNA polymerase. This in turn replicates viral genomic RNA and generates sgRNA1 and particularly large amounts of sgRNA2 (Fig. 3B). The abundant sgRNA2 trans inhibits translation of gRNA only. Unlike in wheat germ extract, where sgRNA1 was inhibited somewhat but to a lesser extent than gRNA, sgRNA1 translation was inhibited only very slightly or not at all in vivo. Thus, the expected large quantity of coat protein can be translated from sgRNA1 as the viral life cycle enters a later stage. This model predicts that the presence of abundant sgRNA2 at the moment of inoculation will block virus replication by prematurely preventing translation of the polymerase, and indeed that was observed previously (50).
We now conclude that there are at least two levels of temporal control of viral-gene expression via subgenomic mRNAs. First and foremost is synthesis of the subgenomic RNAs. They are absent initially; thus, only ORFs 1 and 2 can be translated early in infection from gRNA. sgRNA1 synthesis positively controls the expression of ORFs 3, 4, and 5. The trans regulation of translation by sgRNA2 provides the second level of control. It acts negatively to turn off translation of gRNA, favoring translation of sgRNA1 only. Not only does this control the level of RNA-dependent RNA polymerase (RdRp) produced, we propose that this selective trans inhibition could also free the gRNA of ribosomes, making it available for replication and encapsidation.
Note that shutoff of gRNA translation by sgRNA2 is
not absolutely required for RNA replication in protoplasts, as
indicated by the accumulation (at 24 h p.i.) of viral RNAs in
isolated protoplasts inoculated with mutant BYDV RNA
(PAV6SG2). Thus, sgRNA2 may serve as a fine-tuning device to
maximize replication and/or its effects may be seen more clearly later
in infection or in whole plants, where CP and movement proteins are
necessary.
The proximity of the kissing stem-loop in the 5' UTR to the 5' end determines sensitivity to translation inhibition by sgRNA2. The abilities of gRNA and sgRNA1 to be differentially inhibited by sgRNA2 are attributable to their different 5' UTRs (Fig. 2C). Thus, neither the reporter gene nor the distance of the BTE from the 5' UTR affected the efficiency of translation in the presence of sgRNA2. In contrast, the mutations that alter the location of the BCL relative to the 5' end did change the response of the reporter RNA to sgRNA2 (Fig. 6).
The favored translation of sgRNA1 over gRNA in the presence of sgRNA2 may be explained by different structures of the 5' UTRs per se or by differential requirement for a host factor(s). The positive correlation of proximity of the 5' UTR kissing stem-loop to the 5' end and competitiveness of the mRNA in the presence of sgRNA2 (Fig. 6) supports a simple scanning efficiency mechanism. Given evidence that BTE-mediated translation requires 5'-end-dependent ribosome scanning (21), a longer tract, with significant secondary structure, between the 5' terminus and the BCL may require more translation factors to facilitate ribosome scanning (48). In the absence of sgRNA2, the factors may be in sufficient supply to allow efficient translation of gRNA. Only in the presence of competing sgRNA2 would the factors be reduced enough to hinder translation of gRNA, whereas RNA with the 5'-proximal BCL (sgRNA1) would have a lower factor dependence and hence be less inhibited by competing sgRNA2 (48).
Selective translational control by viral 5' UTRs in cis has been observed in other viruses. For example, subgenomic RNA 4 of brome mosaic virus has a translational competitive advantage over the other three viral RNAs (46). The 5' UTR of the coat protein-encoding sgRNA of turnip crinkle virus mediates translation more efficiently than the gRNA 5' UTR (47). This is not surprising, because the coat protein is needed and expressed at orders of magnitude greater levels than the replication proteins translated from the gRNA. The 5' UTRs of influenza virus mRNAs (18, 41) and the 5' end of the capsid ORF of Sindbis virus subgenomic mRNA (17) mediate selective translation of viral mRNAs when translation of host mRNAs is shut off. In contrast to the above-mentioned RNAs, BYDV sgRNA translation is regulated in trans.
A role for viral proteins in selective translation of BYDV sgRNA1 in the presence of sgRNA2 has been ruled out, but host proteins may participate, as is the case for other viruses that use different mechanisms to control translation. The 5' UTR of brome mosaic virus RNA 2 confers a specific requirement for the host translation factor DED1 in Saccharomyces cerevisiae(39). The cellular protein GRSF-1 participates in selective translation of influenza virus mRNAs (25, 42). With the exception of turnip crinkle virus, all of the above-mentioned translation regulation involves capped viral mRNAs. In contrast, the selective translation mediated by BYDV 5' UTRs is between two uncapped mRNAs dependent on a 3' cap-independent translation element and is mediated by a third viral RNA (sgRNA2).
Potential mechanism(s) of trans inhibition of translation of gRNA by sgRNA2. Regulatory RNAs inhibit gene expression by at least two mechanisms. One mechanism is by base pairing of the regulatory RNA to the target RNA(s) to block translation or to recruit an inhibitory protein(s) to the target RNA(s). Examples include the microRNAs, small interfering RNAs, and bacterial small RNAs (8, 33, 38). This mechanism is unlikely for trans inhibition by BYDV sgRNA2 because the sequence within the BTE of sgRNA2 that is complementary to the 5' UTR of the gRNA is not necessary for its trans inhibition activity (Fig. 5). Moreover, it stimulates translation in cis, and also it trans inhibits translation of nonviral mRNAs to which it has no sequence homology (54, 55).
The more likely mechanism is that BYDV sgRNA2 is a molecular decoy that competes for translation initiation factors. In support of this, addition of eukaryotic translation initiation factor 4F (eIF4F) restored translation of mRNA in extracts inhibited by the addition of BTE RNA (54). Indeed, recently eIF4F has been found to interact directly with the BTE and not with the nonfunctional mutant with the filled BamHI site (E. P. Kneller, K. Treder, E. Allen, and W. A. Miller, unpublished data). Thus, eIF4F binding correlates with the trans inhibition function. It is highly unlikely that translation of sgRNA2 is necessary for its function. Not only is no translation product detectable (Fig. 7), but mutant sgRNA2 containing a frameshift mutation that disrupts ORF 6 still selectively inhibited translation of gRNA in vitro (55).
Other trans-regulatory RNAs from viruses have very different functions. Red clover necrotic mosaic virus RNA2 has a 34-nt trans activator sequence, which is required for transcription of sgRNA from RNA1 (52) and for encapsidation (6). This trans activator base pairs to RNA1 to facilitate sgRNA synthesis. Flock house virus sgRNA3, but not its translation product, trans activates replication of viral genomic RNA2 (14, 15), which then down-regulates the synthesis of sgRNA3 from genomic RNA1 (57). Epstein-Barr virus EBER RNAs may function similarly to VA RNAs in blocking the host protein kinase RNA-activated antiviral response, because they can rescue replication of adenovirus lacking VA RNAs (7, 10). The herpesvirus microRNAs downregulate both viral and host gene expression at different stages of the viral infection (38, 44). Among the above-mentioned viral regulatory RNAs, VA RNAs, EBER RNAs, and most likely BYDV sgRNA2 are not mRNAs. Others function as both a coding RNA and a noncoding regulatory RNA.
It remains to be investigated whether other viruses employ this type of selective, negative regulation of translation. Competitive inhibition of translation among RNAs of other multi-RNA viruses has been observed primarily in vitro (30, 31, 46). The RNA elements that confer translational competitiveness of viral subgenomic RNAs have been mapped to both the 5' (17, 31, 40) and 3' (23) UTRs. However, (i) selective trans inhibition of translation of one viral RNA but not another, (ii) trans inhibition of cap-independent translation, and (iii) inhibition by an apparently untranslated RNA all are properties that so far are known only in BYDV infection. It will be particularly interesting to know whether other viruses that generate multiple subgenomic mRNAs, such as severe acute respiratory syndrome coronavirus and other members of the Nidovirales, or the Closteroviridae, are also controlled by a subgenomic RNA in trans. By extension, it is possible that a host mRNA with particularly high affinity for translation factors could negatively regulate translation of other host mRNAs by this mechanism.
|
ACKNOWLEDGMENTS |
|---|
We thank Sang Ik Song for construction of pRenilla-CP393 and Sondra Schlesinger and Steve Whitham for critical reading of the manuscript and for providing insightful comments.
|
FOOTNOTES |
|---|
R.S.
and A.M.R. contributed equally to the study.
Present address: Department of Medicine/Gastroenterology, University of Alabama at Birmingham, Birmingham, AL 35294.
|
REFERENCES |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
