|Home | About | Journals | Submit | Contact Us | Français|
Positive-strand RNA viruses depend on recruited host factors to control critical replication steps. Previously, it was shown that replication of evolutionarily diverse positive-strand RNA viruses, such as hepatitis C virus and brome mosaic virus, depends on host decapping activators LSm1-7, Pat1, and Dhh1 (J. Diez et al., Proc. Natl. Acad. Sci. U. S. A. 97:3913–3918, 2000; A. Mas et al., J. Virol. 80:246 –251, 2006; N. Scheller et al., Proc. Natl. Acad. Sci. U. S. A. 106:13517–13522, 2009). By using a system that allows the replication of the insect Flock House virus (FHV) in yeast, here we show that LSm1-7, Pat1, and Dhh1 control the ratio of subgenomic RNA3 to genomic RNA1 production, a key feature in the FHV life cycle mediated by a long-distance base pairing within RNA1. Depletion of LSM1, PAT1, or DHH1 dramatically increased RNA3 accumulation during replication. This was not caused by differences between RNA1 and RNA3 steady-state levels in the absence of replication. Importantly, coimmunoprecipitation assays indicated that LSm1-7, Pat1, and Dhh1 interact with the FHV RNA genome and the viral polymerase. By using a strategy that allows dissecting different stages of the replication process, we found that LSm1-7, Pat1, and Dhh1 did not affect the early replication steps of RNA1 recruitment to the replication complex or RNA1 synthesis. Furthermore, their function on RNA3/RNA1 ratios was independent of the membrane compartment, where replication occurs and requires ATPase activity of the Dhh1 helicase. Together, these results support that LSm1-7, Pat1, and Dhh1 control RNA3 synthesis. Their described function in mediating cellular mRNP rearrangements suggests a parallel role in mediating key viral RNP transitions, such as the one required to maintain the balance between the alternative FHV RNA1 conformations that control RNA3 synthesis.
The group of positive-strand RNA [(+)RNA] viruses includes many plant, animal, and human pathogens of economical and medical importance. Despite their great diversity, all positive-strand RNA viruses share some fundamental common features in their replication processes. First, upon infection, their genomes act as mRNAs to express the viral proteins and then, at later steps, as templates for replication and encapsidation. Since these functions are mutually exclusive, profound rearrangements in the viral genome RNA-protein (RNP) composition are required to regulate such transitions. The nature and control of these rearrangements are not well defined. Second, replication of (+)RNA viruses universally occurs on virus-induced rearranged host intracellular membranes (1). Third, because (+)RNA viruses have genomes with limited coding capacity, they greatly depend on the cellular machinery to multiply. Host factors, therefore, play crucial roles in regulating all steps of their life cycles (2, 3). The identification of common required host factors will expand our knowledge of key steps of (+)RNA life cycles and might provide novel and stable targets for the generation of broad-spectrum antiviral drugs.
The cellular decapping activators LSm1-7 heptameric ring, PatL1, and DDX6 (also designated Rck/p54) are examples of host proteins required for the replication of a wide range of (+)RNA viruses. These proteins are highly conserved from yeast (designated LSm1-7, Pat1, and Dhh1) to humans and can be found in cytoplasmic foci, named P-bodies. These foci are highly dynamic granules, with unclear functions, that contain translationally repressed mRNAs together with multiple proteins from the mRNA decay and microRNA machineries (4, 5). All P-body components rapidly cycle in and out of these granules, indicating that there is a constant exchange of molecules with the cytoplasm, where all P-body components are also diffusely present (6–8). In noninfected cells, LSm1-7, PatL1, and DDX6 form a complex and promote cellular mRNA decay by accelerating decapping in the 5′-3′deadenylation-dependent mRNA decay pathway. Although their precise way of functioning is not fully understood, they have been suggested to facilitate RNP rearrangements required for the transition of cellular mRNAs from an active translatable state to a translationally repressed state that allows the assembly of the decapping complex (9, 10). The LSm1-7 ring is constituted by seven LSm proteins that belong to the conserved Sm family of proteins, which are characterized by the presence of the Sm fold. It has been described that proteins containing Sm folds act as chaperones facilitating a variety of RNA-RNA and RNA-protein interactions (11). PatL1 has emerged as a pivotal protein in mRNA decay. Recent evidence supports its function as a scaffold protein, allowing the sequential binding of repression and decay factors on mRNPs that eventually leads to degradation (12). Finally, DDX6 belongs to the family of DEAD box helicases. These highly conserved enzymes accelerate structural transitions of RNAs and RNPs in an ATP-dependent manner that resembles the activities of certain groups of protein chaperones (13, 14). Interestingly, in contrast to their decay function in cellular mRNAs, PatL1, LSm1-7, and DDX6 have all or in part been shown to be required for replication of diverse (+)RNA viruses, including human hepatitis C virus (HCV), West Nile virus (WNV), and Dengue virus (DNV), as well as the plant brome mosaic virus (BMV) (15–22). Remarkably, Hfq, a homolog of LSm1 in bacteria, is also required for the replication of the (+)RNA phage Qβ (23, 24). The direct interaction of PatL1, LSm1-7, and DDX6 with essential regulatory signals in the viral genomes and with viral proteins support that this positive regulation is mediated by a direct and specific effect (18, 21, 22, 25). The conserved dependence on LSm1-7, PatL1, and DDX6 of viruses that infect different kingdoms of life underlines a remarkable robustness in the use of proteins from the decapping pathway to regulate (+)RNA viral life cycles; however, the underlying mechanisms involved remain to be elucidated.
To further explore how wide spread the use of LSm1-7, PatL1, and DDX6 in (+)RNA virus life cycle is and to gain further insights into their mechanism of action, here we tested their function on the replication of Flock House virus (FHV), a natural insect pathogen and well-studied member of the Nodaviridae family (reviewed in reference 26). The simplicity of its genome makes FHV a highly tractable system to study basic aspects of (+)RNA biology. The FHV bipartite genome consists of two capped but nonpolyadenylated RNA segments (Fig. 1A). RNA1 (3.1 kb) encodes protein A, the only FHV protein required for replication (27), which occurs in outer mitochondrial membranes (28). RNA2 (1.4 kb) encodes the capsid precursor α, which is required for virion formation but is dispensable for virus replication (29–31). Hence, RNA1 is capable of autonomous replication in cells (32). Furthermore, RNA1 templates that do not express a functional protein A can be complemented in trans (33, 34). During replication, RNA1 also produces subgenomic RNA3 (387 nucleotides). This RNA3 corresponds to the 3′ end of RNA1, and its synthesis requires a long-distance base pairing between two cis-acting elements in RNA1 (35). Thus, different RNP rearrangements within genomic RNA1 are predicted to lead to the synthesis of either negative-stranded RNA1 [(−)RNA1), when the polymerase synthesizes a complete cRNA, or (−)RNA3, when the polymerase prematurely stops due to the secondary or tertiary structures formed by the long-distance base pairing. RNA3 encodes protein B2, required for suppression of RNA silencing in infected hosts (36), and B1, a protein of unknown function. In addition, RNA3 fulfills an important role in the regulation of FHV gene expression by coordinating the production of proper RNA1 and RNA2 levels. RNA3 transactivates RNA2 replication, while high levels of RNA2 inhibit RNA3 synthesis (35, 37, 38). The downregulation of RNA3 most probably disallows disproportionately high RNA2/RNA1 ratios within FHV-infected cells. This feedback mechanism is of fundamental importance for viruses with segmented genomes, because it would ensure the timely and optimal expression of different viral proteins for the different stages of the viral life cycle, giving, for example, to RNA1 replication an advantage over RNA2 replication during the initial stages of infection, when RNA1 translation products are required for the establishment of viral replication.
A remarkable feature of FHV replication is its ability to replicate in a wide variety of eukaryotic cell types, including Drosophila, Caenorhabditis elegans, plants, mammals, and the budding yeast Saccharomyces cerevisiae (39–43). Thus, host factors required for FHV replication must be highly conserved. Because S. cerevisiae is easy to manipulate genetically, the FHV/yeast system is an excellent model system to study the effect of host factors in (+)RNA virus replication. By using this system, we show here that depletion of the LSM1, PAT1, or DHH1 gene dramatically increases RNA3 accumulation, altering RNA3/RNA1 ratios. This effect was not explained by differences between RNA1 or RNA3 steady-state levels in the absence of replication. Importantly, coimmunoprecipitation studies indicated that LSm1-7, Pat1, and Dhh1 interact with FHV protein A and the viral genome. Together, these results are consistent with a replication-dependent way of action of LSm1-7, Pat1, and Dhh1 to control proper RNA3/RNA1 ratios. Since no effects were observed on the early replication steps of RNA1 recruitment to the replication complex and RNA1 synthesis, a direct role on RNA3 synthesis was suggested. Moreover, RNA3/RNA1 ratio alterations were independent of the membrane compartment, where replication occurs, and require the ATPase activity of Dhh1, an essential feature for Dhh1-dependent RNP rearrangements. These findings highlight the role of LSm1-7, Pat1, and Dhh1 as master regulators of (+)RNA virus replication. Their described function on cellular mRNP rearrangements would support a parallel role in mediating key viral RNP transitions, such as the one required to control RNA3 synthesis.
Deletion yeast strains lsm1Δ, pat1Δ, and dhh1Δ were derived from Saccharomyces cerevisiae strain YPH500 (MATα ura3-52 lys2-801 ade2-101 trp1Δ63 his3-Δ200 leu2-Δ1) (17, 19). Yeast cells were grown at 30°C in medium selective for the desired plasmids, with 2% glucose, until mid-log phase. Yeast cells were then washed and induced in selective minimal medium with 2% galactose at 26°C for 24 h, unless stated otherwise in Results. Yeast containing plasmid pCupF1 with a frame shift (pCupF1fs) was grown without adding copper to use basal CUP1 promoter activity.
Standard molecular cloning techniques were used throughout (44). All products generated by PCR were verified by sequencing, and detailed methods and primer sequences are available from the authors upon request. Plasmid pCupF1fs is a pF1fs derivative (34, 42) in which the GAL1 promoter was replaced by the CUP1 promoter. Plasmids pF1fsmut1, pFA, pFAD692E (35, 45), and pFA-HCV (46) were described previously. Plasmid PV167 expresses RNA3 under the GAL1 promoter and a 3′-flanking antigenomic hepatitis delta virus ribozyme (δRz). This plasmid can be transcribed but does not support RNA3 replication when expressed with protein A in yeast. Plasmid expressing wild-type (WT) Dhh1 was described previously (15). The QuikChange II site-directed mutagenesis kit (Agilent Technologies) was used to introduce mutations into the WT Dhh1 sequence by following the protocol provided by the manufacturer. Mutant E196A, containing a point mutation in motif II, was generated using the primer 5′GTTCATTATTCATCATGGACGCAGCCGATAAAATGTTATCTCG-3′, and mutant S226A T228A, containing two point mutations in motif III, was generated using the primer 5′CTCCAACTCACCAATCTTTATTATTTGCCGCTGCTTTCCCACTAACGG-3′.
Total RNA was isolated from yeast cells using a hot-phenol method and analyzed by Northern blotting as previously described (47). Unless otherwise noted, 3-μg aliquots of total yeast RNA were analyzed by formaldehyde-agarose gel. RNAs were detected with 32P-labeled probes specific for positive- and negative-strand FHV RNA1 and RNA3 and for 18S rRNA as previously described (15, 34). Radiolabeled probes were generated using the MAXIscript in vitro transcription kit (Ambion), Northern blots were imaged on a Typhoon 8600 instrument (Amersham), and band intensities were quantified using ImageQuant software (Molecular Dynamics).
Total protein was extracted from equivalent numbers of yeast cells as previously described (48), separated in 10% SDS-polyacrylamide gels, and immunoblotted. The following antibodies were used throughout the experiments. Primary antibodies, to specifically detect FHV protein A (28), phophoglycerate kinase (PGK) (Molecular Probes), yeast Dhh1 (Santa Cruz Biotechnology), yeast Pat1 (generous gift from F. Wyers) (49), yeast LSm1 (generous gift from R. Parker) (50), and FLAG (Sigma), as well as fluorescently labeled secondary antibodies to be detected in the Odyssey infrared imaging system (LI-COR), were used. Quantification of proteins using the Odyssey infrared imaging system (LI-COR) was done according to the instructions supplied by the manufacturer (LI-COR).
For immunoprecipitation experiments, 50- to 100-ml yeast cultures were grown in synthetic medium with the appropriate amino acids with 2% glucose at 30°C to mid-log phase. Yeast cells then were washed and induced in selective medium at 26°C to mid-log phase with 2% galactose for 24 h. All subsequent steps were carried out at 4°C. The cultures were centrifuged at 4,000 rpm, washed once in cold medium, and lysed by vortexing with glass beads in 1× NET buffer (50 mM Tris, pH 7.5, 150 mM NaCl, 0.1% NP-40, or IGEPAL CA-630 [from Sigma]) containing a protease inhibitor cocktail (Promega), 5 μg/ml of pepstatin, leupeptin, and aprotinin (Sigma), and 40 U/ml RNAsin RNase (Promega). Lysates were quantified, and equal amounts of lysates were incubated with anti-Flag matrix (Sigma) for 3 h. Where indicated, samples were treated with 10 mg/ml RNase A (Sigma) for 30 min prior to incubation with matrix. Following incubation, samples were precipitated and washed 3 to 5 times in 1× NET buffer. The immunoprecipitates were eluted by boiling samples in SDS loading buffer for 5 min at 95°C for Western analysis. For RNA analysis, pellets were eluted in RNA elution buffer (50 mM Tris-HCl, pH 7.4, 5 mM EDTA, 10 mM dithiothreitol, and 1% SDS), treated with 20 μg proteinase K (Ambion) for 30 min at 42°C, purified with phenol-chloroform, and ethanol precipitated.
RNAs from immunoprecipitation assays were treated with a Turbo DNase I DNA-free kit according to the manufacturer's instructions (Ambion) and were directly used for reverse transcription-PCR (RT-PCR) assays. Twenty nanograms of RNA was reverse transcribed using random primers and Superscript III (Invitrogen) according to the manufacturer's instructions. The cDNA was amplified with specific primers for FHV RNA1 (5′-AACTCGCGCTAATCCAGGAA-3′ and 5′-TCCCTGCGCACGTTGTT-3′) and a SybrGreen master mix reagent kit in an ABI Prism 900HT (Applied Biosystems, Life Technologies). Primer sequences are available upon request.
To explore the impact of LSm1-7, Pat1, and Dhh1 on FHV replication, we used the previously established FHV trans-replication system for yeast cells. In this system, separate plasmids express template genomic RNA1 and protein A, disconnecting the cis-encoded template function of RNA1 from its coding potential (34, 35) (Fig. 1B). Plasmid pCupF1fs expresses, from the CUP1 promoter, a genomic RNA1 derivative carrying a frameshift mutation that blocks production of protein A; thus, it can only be used as a template for replication. In turn, protein A is provided in trans by plasmid pFA, which expresses a GAL1-driven RNA1 derivative in which the 5′ and 3′ FHV untranslated regions (UTRs) are replaced by cellular sequences. Thus, this RNA1 derivative can be used as mRNA for protein A expression but not as a template for replication. The corresponding pCupF1fs and pFA plasmids were transformed into WT and lsm1Δ, pat1Δ, and dhh1Δ yeast strains, and cells were grown in media without copper to exploit the basal CUP1 promoter activity. These growing conditions prevent RNA1 overproduction. The expression of protein A then was induced, and accumulation of FHV positive-strand and negative-strand RNA1 and RNA3 [referred to here as (+)RNA1, (−)RNA1, (+)RNA3, and (−)RNA3] was examined by Northern blotting. This was done at 24 h after galactose induction to examine early events in RNA3 and RNA1 production and at 72 h after galactose induction, a time point at which a balance in RNA3 and RNA1 production is already achieved (Fig. 2A). Interestingly, depletion of LSm1, Pat1, or Dhh1 protein dramatically increased subgenomic (+)RNA3 and (−)RNA3 levels both at 24 and 72 h after galactose induction. In contrast, genomic (+)RNA1 and (−)RNA1 accumulation was either unchanged or only moderately higher (Fig. 2B and andC).C). These altered FHV RNA1 and RNA3 accumulation patterns in the deletion strains resulted in an increase of 7- to 20-fold in the (+)RNA3/(+)RNA1 ratio and 4- to 5-fold in the (−)RNA3/(−)RNA1 ratio over those in WT yeast cells (Fig. 2D). Similar results were obtained at 24 and 72 h after galactose induction except for the pat1Δ strain, in which the (+)RNA3/(+)RNA1 ratio related to WT cells was 3-fold higher at 24 h than at 72 h. Together, these results indicate that LSm1-7, Pat1, and Dhh1 play a key role in maintaining proper FHV subgenomic/genomic ratios.
Since LSm1-7, Pat1, and Dhh1 promote degradation of cellular mRNAs (9), the altered FHV RNA levels observed in lsm1Δ, pat1Δ, and dhh1Δ yeast might result from increased stability of FHV RNA1 and RNA3. To test this possibility, we analyzed the steady-state levels of RNA1 and RNA3 by expressing these RNAs in the absence of protein A and, thus, the absence of replication (Fig. 3A). WT, lsm1Δ, pat1Δ, and dhh1Δ strains were transformed with plasmid pCupF1fs to transcribe RNA1 or with plasmid PV167 to transcribe RNA3, and RNA1 and RNA3 accumulation levels were examined by Northern blotting (Fig. 3A). As previously observed, under these conditions, in addition to the bands corresponding to RNA1 and RNA3, unspecific lower and higher comigrating bands are detected (45, 51). In lsm1Δ and pat1Δ strains, an increase of 2- to 2.5-fold over the WT level was observed for RNA1 and RNA3 accumulation, while in the dhh1Δ strain, RNA1 and RNA3 steady-state levels were not altered. These variations among deletion strains might explain the slightly higher RNA1 and RNA3 accumulation observed during replication in lsm1Δ and pat1Δ mutant strains compared to that in the dhh1Δ mutant strain, and they strongly suggest that differences in basal FHV RNA steady-state levels are not the primary reason for the dramatic increase in FHV RNA3 levels observed in the deletion strains.
We next examined whether depletion of LSm1-7, Pat1, or Dhh1 influences protein A expression levels. To explore this possibility, we transformed WT, lsm1Δ, pat1Δ, and dhh1Δ yeast strains with plasmid pFA and analyzed protein A expression levels by Western blotting (Fig. 3B). Similar protein A expression levels were detected in the WT and the lsm1Δ deletion mutant, while these were reduced by 55 and 44% in pat1Δ and dhh1Δ mutant strains. Given that LSm1-7, Pat1, and Dhh1 form a complex and have similar effects on FHV RNA3/RNA1 ratios, the protein A expression differences observed in pat1Δ and dhh1Δ are very unlikely to account for the RNA3/RNA1 ratio alterations.
Together, these results are consistent with a LSm1-7, Pat1, and Dhh1 replication-dependent control of FHV RNA3/RNA1 ratios.
If LSm1, Pat1, and Dhh1 regulate RNA3 subgenomic accumulation during the replication step, then they would be expected to interact with protein A, the only FHV protein involved in replication, and the FHV RNA. To test this possibility, we made use of a plasmid expressing Dhh1 tagged with the FLAG epitope that was previously shown to efficiently coimmunoprecipitate LSm1 and Pat1 (52). WT yeast cells expressing Dhh1 protein with or without a FLAG tag were transformed with plasmids pCupF1fs and pFA, and immunoprecipitation analyses were carried out using antibodies against the FLAG epitope (Fig. 4). Pat1 and LSm1 coimmunoprecipitated with Dhh1, verifying this interaction in the presence of FHV replication. Interestingly, Dhh1 also coimmunoprecipitated protein A. None of these interactions were sensitive to treatment with RNase A, suggesting that they associate as a complex rather than cooccupy an RNA molecule. In addition, FHV (+)RNA1 was detected by quantitative PCR at significant levels only in the immunoprecipitates where the FLAG epitope was expressed, whereas when these samples were treated with RNase A or in cells expressing nontagged Dhh1, only background levels of FHV (+)RNA1 were found. Altogether, these observations demonstrated that Dhh1, LSm1-7, and Pat1 can physically interact in vivo either directly or through additional components with protein A and FHV RNA1.
Protein A is the sole FHV protein required for replication. It recruits FHV genomic RNA1 to a mitochondrial membrane-associated state where it assembles the replication complex and initiates synthesis and amplification of the viral genome (45). We have previously shown that BMV RNA recruitment to the site of replication depends on LSm1-7, Pat1, and Dhh1. To examine whether they are playing a similar role in FHV replication, the WT and lsm1Δ, pat1Δ, and dhh1Δ mutant yeast strains were transformed with plasmid pCupF1fs that expresses RNA1, alone or together with pFAD692E-YFP, a derivative of plasmid pFA. This derivative expresses a mutated version of protein A that abolishes its replication activity while maintaining its capacity to specifically recruit the FHV genome to the mitochondrial membrane. In addition, an insertion of the yellow fluorescent protein (YFP) open reading frame (ORF) downstream of the protein A ORF increases the size of the pFAD692E-YFP-derived transcript to distinguish it from RNA1 (45). As previously described, in WT yeast, expression of pCupF1fs alone yielded very low levels of (+)RNA1 because its steady state only depends on its transcription from the CUP1 promoter and its degradation from the cellular decay machinery. However, when pCupF1fs and pFAD692E-YFP were coexpressed, a 6-fold higher accumulation of RNA1 was observed as a consequence of its protein-A-induced membrane association and subsequent stabilization (45) (Fig. 5). Neither of the deletion mutants analyzed affected FHV RNA1 recruitment. As expected because of the effect of LSm1-7 and Pat1 on RNA1 steady-state levels (Fig. 3A), in lsm1Δ and pat1Δ yeast strains the accumulation of RNA1 in the absence of protein A was higher than that in the WT. Importantly, this increase correlated with a parallel one in RNA1 accumulation in the presence of protein A. In the dhh1Δ yeast strain, FHV RNA1 levels paralleled those in WT cells, both in the absence and presence of protein A. Thus, the ratios of protein A-stimulated RNA1 accumulation relative to RNA1 accumulation in the absence of protein A were similar in all cases (Fig. 5C). Altogether, these results indicate that FHV RNA1 recruitment does not depend on LSm1-7, Pat1, and Dhh1.
Since RNA3 is produced during the course of RNA1 replication, we next sought to uncouple these two processes to investigate whether LSM1, PAT1, and DHH1 deletions could also perturb RNA1 accumulation independently of RNA3 synthesis. For this, we used pF1fsmut1, in which genomic RNA1 is replicated in the absence of RNA3 production (35). It is established that a long-distance base pairing between two cis-acting elements within RNA1 is required for RNA3 production (35). Since plasmid pF1fsmut1 transcribes an RNA1 derivative carrying multiple mutations in one of the cis-acting sequences, this base pairing is disrupted and consequently RNA3 synthesis is inhibited. WT, lsm1Δ, pat1Δ, and dhh1Δ yeast strains were transformed with plasmids pF1fsmut1 and pFA to express protein A, and accumulation of (+)RNA1 and (−)RNA1 was analyzed by Northern blotting (Fig. 6). In lsm1Δ and pat1Δ strains, (+)RNA1 and (−)RNA1 accumulation was increased by approximately 1.5- and 3-fold, respectively. This increase was consistent with the enhanced (+)RNA1 steady-state levels found in these strains in the absence of replication (Fig. 3A) and would support that the moderately higher RNA1 accumulation levels observed (Fig. 6) are not related to replication effects. In compelling agreement with this argument, no difference in RNA1 replication levels were observed between the WT and the dhh1Δ mutant strain (Fig. 6), a strain that had no effect on RNA1 steady-state levels (Fig. 3A). Together, these results suggest that LSm1, Pat1, and Dhh1 regulate later replication steps that specifically involve RNA3 synthesis. Although RNA3 constructs exist that allow the study of RNA1-independent replication of RNA3 in some higher eukaryotic cell lines (37), we did not succeed in establishing such a system in S. cerevisiae; thus, we were not able to address the effect of the deletion mutants on RNA3 replication in the absence of RNA1 replication.
Replication of plus-strand viruses is invariably associated with membranes (53). However, the preference of different (+)RNA viruses for different membrane compartments is not well understood. It has been proposed that the presence of specific host factors in close proximity to different membrane compartments contributes to these preferences. In FHV, replication occurs in the outer membrane of mitochondria, where the viral RNA is recruited by protein A via an N-proximal sequence that contains a transmembrane domain. Since LSm1-7, Pat1, and Dhh1 have been shown to localize in P-bodies in close proximity to mitochondria (54), we questioned whether these factors would also affect FHV replication when associated with other membrane compartments. It has been previously shown that FHV RNA replication complexes can be retargeted to the endoplasmic reticulum (ER) through the replacement of the N-proximal sequence by ER-targeting sequences from the HCV NS5B protein (46). These alternative replication complexes were also shown to be fully functional, although FHV replication kinetics were enhanced (46). In these experiments, we focused on protein Dhh1, since it has an effect on the RNA3/RNA1 ratios similar to those of the other two components of the complex without any additional effect on FHV RNA steady-state levels; thus, it shows a clearer phenotype (Fig. 2 to to3).3). WT and dhh1Δ yeast strains were cotransformed with pFA-HCV, which expresses the HCV-protein A chimera, and pCupF1fs, which provides RNA1 as a template for replication. RNA1 and RNA3 replication levels were analyzed as described before, except that Northern blotting was performed 12 h after galactose induction instead of 24 h. This was done because the enhanced RNA ER-targeted FHV replication results in a saturation of FHV replication levels at 24 h that hampers comparisons among different conditions. DHH1 deletion resulted in a dramatic increase in (+)RNA3 and (−)RNA3 accumulation, while (+)RNA1 and (−)RNA1 levels were not or were moderately affected (Fig. 7A and andB),B), altering the RNA3/RNA1 ratios (Fig. 7C). This effect was comparable to the one achieved when mitochondrion-targeted protein A was expressed (Fig. 2).
Dhh1 protein belongs to the highly conserved family of DEAD box helicases. Members of this family are involved in all aspects of RNA metabolism and carry out structural RNP transitions by unwinding local short duplex RNA and by providing nucleation centers to establish larger RNA-protein complexes (13). This unwinding activity requires ATP binding but not necessarily ATP hydrolysis, while ATP hydrolysis is key for the release and recycling of the DEAD box protein. Two characteristic motifs in these proteins are motifs II and III. Motif II, also called the DEAD motif, is critical for ATP binding and hydrolysis. Motif III is defined by the SAT sequence that mediates the communication between ATP binding and RNA binding sites to create a high-affinity RNA binding site (55). Motif II is essential for the function of Dhh1 in the cell; however, the role of motif III has not been explored yet (56, 57). To further investigate the role of Dhh1 in FHV replication, we generated two Dhh1 mutants, the single E196A mutant in motif II and the double mutant S226A/T228A in motif III, referred to here as mutant DAAD and AAA, respectively (Fig. 8A). We cotransformed a dhh1Δ yeast strain with plasmids pCupF1fs and pFA together with either an empty plasmid, a plasmid expressing WT Dhh1, or plasmids expressing the corresponding Dhh1 mutants. Importantly, the introduced mutations did not affect protein stability, since parallel expression levels were observed for wild-type and mutant Dhh1 proteins (Fig. 8B). FHV RNA replication then was analyzed as described before (Fig. 8C). The accumulation of (+)RNA1 and (−)RNA1 levels was similar in all cases. This was expected, since DHH1 deletion did not significantly affect RNA1 replication (Fig. 2). However, interesting differences were observed in RNA3 accumulation. As anticipated, expression of WT Dhh1 restored the normal levels of (+)RNA3 and (−)RNA3 accumulation. While expression of mutant AAA restored RNA3 levels to those found when WT Dhh1 protein was expressed, expression of mutant DAAD yielded levels similar to those reached with the empty plasmid. This indicates that motif II, but not motif III, is required for the function of Dhh1 on FHV replication. Thus, the ATPase activity of Dhh1 is required for maintaining proper RNA3 levels.
Viral (+)RNA genomes, as cellular mRNAs, are in constant interaction with proteins. Changes in the composition of these RNP structures are essential to direct viral genomes through different steps of the infection cycle. By using the FHV/yeast model system, we show here that host decapping activators that mediate key cellular mRNP transitions regulate the synthesis of subgenomic FHV RNA3, a replication step that requires formation of a particular RNP structure in RNA1. Depletion of LSM1, PAT1, or DHH1 altered subgenomic RNA3/genomic RNA1 ratios by dramatically increasing RNA3 accumulation. Two main observations support that this effect is mediated by a direct and specific replication-dependent mechanism affecting RNA3 synthesis. First, LSm1, Pat1, and Dhh1 coimmunoprecipitated with the viral genome and protein A, the viral polymerase. Second, LSm1, Pat1, and Dhh1 did not affect the early replication steps of RNA1 recruitment to the replication complex or RNA1 synthesis. The slightly higher RNA1 levels under replication conditions observed in lsm1Δ and pat1Δ strains compared to the WT correlated with parallel increases in RNA1 steady-state levels in the absence of replication. Given the known roles of LSm1-7 and Pat1 in cellular mRNA degradation, these strongly suggest that independently of their effect on RNA3 synthesis, LSm1-7 and Pat1, but not Dhh1, have an additional role in RNA1 and RNA3 stability. Together, these data support that LSm1-7, Pat1, and Dhh1 assist to maintain proper RNA3/RNA1 ratios by repressing RNA3 synthesis. This is predicted to be of fundamental importance for the FHV life cycle, given the role of RNA3 in coordinating RNA1 and RNA2 levels and consequently the synthesis of different FHV proteins at optimal ratios for different stages of the FHV life cycle.
Two mechanisms were proposed for synthesis of RNA3 from RNA1: internal initiation and premature termination (38). According to the internal initiation model, protein A synthesizes (+)RNA3 from an internal site on (−)RNA1. In contrast, the premature termination model predicts that protein A uses (+)RNA1 to begin synthesizing (−)RNA1 but terminates prematurely, thus generating (−)RNA3. (−)RNA3 then acts as the template to generate (+)RNA3. Current evidence strongly supports the latter mechanism. First, the accumulation of FHV (−)RNA3 synthesis does not depend on (+)RNA3 (38). Second, FHV (−)RNA3 serves as the template for the synthesis of (+)RNA3 (58). Third, a long-distance intramolecular base pairing interaction in (+)RNA1 is required for RNA3 synthesis (35). This base pairing occurs between two cis elements in (+)RNA1, one located just upstream of the RNA3 start site and the other 1.5 kb upstream (35, 58). It is proposed that this interaction promotes a highly structured conformation, resulting in premature termination of (−)RNA1 synthesis and generating (−)RNA3. Since (+)RNA1 would serve as the template for both (−)RNA1 and (−)RNA3, a proper balance between alternative FHV RNA1 conformations is required to control proper RNA3 levels. In this scenario, the decapping complex LSm1-7/Pat1/Dhh1 might provide the required chaperone and helicase activities. They might be recruited to the replication complex to stabilize the local RNA structures that repress excess RNA production. Alternatively, their interaction with protein A-containing FHV RNP1 complexes might be transient and highly dynamic. In line with this, it has been described that the Dhh1 DEAD motif, which is critical for ATP hydrolysis and for maintaining proper RNA3 levels, triggers the exit of DEAD box helicases from cellular mRNP complexes (59, 60). Thus, a parallel dynamic interaction with the viral RNA might govern genomic and subgenomic FHV production.
The use of LSm1-7/Pat1/Dhh1 chaperone and helicase activities by FHV is in agreement with the proposal that (+)RNA viruses with short genomes, such as FHV, do not code for proteins with helicase function; thus, they would need to subvert cellular helicases and chaperones to assist replication and other viral processes that require changes in RNP composition (61). In agreement with this idea, it has been recently shown that FHV also hijacks Ded1, another cellular DEAD box helicase, to promote synthesis of (+)RNA1, likely by unwinding local structures at the 3′ end of (−)RNA1 (61). Furthermore, as for Ded1, the ATPase activity of Dhh1 was essential for this function.
The proposed roles of LSm1-7, Pat1, and Dhh1 in FHV replication have features in common with the ones already described for other (+)RNA viruses. Using a BMV/yeast replication system, it was shown that these proteins are required for both translation and exit from the translation machinery to the viral replication complex of genomic BMV (+)RNA, a step that requires profound RNP rearrangements (17, 19, 64). Remarkably, this function is conserved in humans, since DDX6 can replace its yeast counterpart Dhh1 in BMV RNA translation and replication (15), and human LSm1-7, PatL1, and DDX6 promote translation and replication of HCV RNA (21). Furthermore, DDX6 has also been shown to promote replication of dengue virus (DENV) (22) and LSm1-7 replication of WNV (16). Current data support that, as shown here for FHV, these effects occur through interactions with viral components. The LSm1-7 ring interacts directly with cis sequences located in the UTRs of HCV and BMV that regulate translation and replication (21, 25), and DDX6 interacts with the DENV 3′UTR (22). Furthermore, LSm1 coimmunoprecipitates and colocalizes with the BMV polymerase (62), and DDX6 colocalizes with the DENV and WNV replication complexes (16, 22). Interestingly, Hfq, the bacterial homolog of Lsm1, supports replication of (+)RNA phage Qß by mediating rearrangements in secondary structures at the 3′ end of the Qß RNA genome that allows the association of the viral polymerase and subsequent initiation of replication (63).
Taken together, our results highlight the conserved use of LSm1-7, Pat1, and Dhh1 proteins by evolutionarily diverse (+)RNA viruses and strongly suggest that these proteins promote RNP rearrangements in (+)RNA viral genomes required for key replication steps. This points at LSm1-7, Pat1, and Dhh1 as promising targets for developing novel broad-spectrum antiviral strategies.
We thank Roy Parker and F. Wyers for reagents. We are also grateful to A. Meyerhans for critically reading the manuscript.
This work was supported by a grant from the Spanish Ministerio de Ciencia e Innovación BFU2010-20803. I.A.-R. was supported by Fundação para a Ciência (SFRH/BD/630/2002) through the Gabba program from the University of Porto, Portugal. P.M.V.W. was supported by a Howard Hughes Medical Institute predoctoral fellowship and an NIH Predoctoral Training grant in molecular biosciences (T32 GM07215).
Published ahead of print 27 March 2013