Co-opted Cellular Sac1 Lipid Phosphatase and PI(4)P Phosphoinositide Are Key Host Factors during the Biogenesis of the Tombusvirus Replication Compartment

Replication of positive-strand RNA viruses depends on recruitment of host components into viral replication compartments or organelles. Using TBSV, we uncovered the critical roles of Sac1 PI(4)P phosphatase and its substrate, PI(4)P phosphoinositide, in promoting viral replication. Both Sac1 and PI(4)P are recruited to the site of viral replication to facilitate the assembly of the viral replicase complexes, which perform viral RNA replication. We found that Sac1 affects the recruitment of other host factors and enrichment of phosphatidylethanolamine and sterol lipids within the subverted host membranes to promote optimal viral replication. In summary, this work demonstrates the novel functions of Sac1 and PI(4)P in TBSV replication in the model host yeast and in plants.

ABSTRACT

Positive-strand RNA [(+)RNA] viruses assemble numerous membrane-bound viral replicase complexes (VRCs) with the help of viral replication proteins and co-opted host proteins within large viral replication compartments in the cytosol of infected cells. In this study, we found that deletion or depletion of Sac1 phosphatidylinositol 4-phosphate [PI(4)P] phosphatase reduced tomato bushy stunt virus (TBSV) replication in yeast (Saccharomyces cerevisiae) and plants. We demonstrate a critical role for Sac1 in TBSV replicase assembly in a cell-free replicase reconstitution assay. The effect of Sac1 seems to be direct, based on its interaction with the TBSV p33 replication protein, its copurification with the tombusvirus replicase, and its presence in the virus-induced membrane contact sites and within the TBSV replication compartment. The proviral functions of Sac1 include manipulation of lipid composition, sterol enrichment within the VRCs, and recruitment of additional host factors into VRCs. Depletion of Sac1 inhibited the recruitment of Rab5 GTPase-positive endosomes and enrichment of phosphatidylethanolamine in the viral replication compartment. We propose that Sac1 might be a component of the assembly hub for VRCs, likely in collaboration with the co-opted the syntaxin18-like Ufe1 SNARE protein within the TBSV replication compartments. This work also led to demonstration of the enrichment of PI(4)P phosphoinositide within the replication compartment. Reduction in the PI(4)P level due to chemical inhibition in plant protoplasts; depletion of two PI(4)P kinases, Stt4p and Pik1p; or sequestration of free PI(4)P via expression of a PI(4)P-binding protein in yeast strongly inhibited TBSV replication. Altogether, Sac1 and PI(4)P play important proviral roles during TBSV replication.

IMPORTANCE Replication of positive-strand RNA viruses depends on recruitment of host components into viral replication compartments or organelles. Using TBSV, we uncovered the critical roles of Sac1 PI(4)P phosphatase and its substrate, PI(4)P phosphoinositide, in promoting viral replication. Both Sac1 and PI(4)P are recruited to the site of viral replication to facilitate the assembly of the viral replicase complexes, which perform viral RNA replication. We found that Sac1 affects the recruitment of other host factors and enrichment of phosphatidylethanolamine and sterol lipids within the subverted host membranes to promote optimal viral replication. In summary, this work demonstrates the novel functions of Sac1 and PI(4)P in TBSV replication in the model host yeast and in plants.

INTRODUCTION

Positive-strand RNA [(+)RNA] viruses assemble numerous membrane-bound viral replicase complexes (VRCs) within large viral replication compartments in the cytosol of infected cells (1–10). The viral replication proteins with co-opted host proteins are involved in the biogenesis of replication compartments. In addition to membrane deformations and membrane proliferation, (+)RNA viruses also retarget cellular trafficking pathways and induce lipid biosynthesis to achieve robust replication (3, 6, 10–13). In spite of major recent advances in virus-host interactions, we still have not identified many of the subverted host factors in the biogenesis of VRCs and replication compartments, or their roles have not been defined.

Tomato bushy stunt virus (TBSV) and other tombusviruses code for two viral replication proteins, termed p33 and p92^pol, which are essential for virus replication (14–16). The TBSV p33 replication protein is the master regulator of VRC assembly and viral (+)RNA recruitment into VRCs. The auxiliary p33 has an RNA chaperone function. p92^pol is the viral RNA-dependent RNA polymerase (RdRp) that synthesizes the viral RNA progeny. The N-terminal segment of p92^pol is identical to p33 due to the translational readthrough of the stop codon in the p33 open reading frame (ORF). TBSV forms numerous VRCs, spherule-like structures that are vesicles with narrow openings toward the cytosol, utilizing peroxisomal membranes in plants and yeast (Saccharomyces cerevisiae), a model host (17–19).

Intensive research with TBSV in yeast has led to the identification of a long list of host factors co-opted for viral RNA replication (3, 5, 20, 21). Genomics- and lipidomics-based results also demonstrated a major role for global phospholipid and sterol biosynthesis in TBSV replication (22–24). Lipids, including phosphatidylethanolamine (PE), phosphatidylinositol 3-phosphate [PI(3)P] phosphoinositide, and sterols, play essential roles in the biogenesis of tombusviral replication compartments, the formation of VRCs, and activation of the virus-encoded p92 RdRp (25–30).

TBSV replication triggers characteristic alterations in cells, including stabilization of the actin network, recruitment of endosomal and COPII vesicles into the viral replication compartments, induction of subcellular membrane proliferation, aggregation of peroxisomes, and exploitation of the ATP-generating glycolytic and fermentation pathways (26, 30–36). Another feature of TBSV replication is the formation/stabilization of membrane contact sites (MCSs) between the endoplasmic reticulum (ER) and peroxisomes, which are required for sterol enrichment within the replication compartment to support efficient virus replication (5, 19, 37–39).

In this study, we dissected the proviral function of Sac1 (suppressor of actin mutations 1-like protein) lipid phosphatase, which was identified as a pro-TBSV host factor in an earlier high-throughput genome-wide screen in yeast (40). Sac1 is a phosphatidylinositol 4-phosphate [PI(4)P] lipid phosphatase that is conserved in eukaryotes (41–43). Sac1 controls several cellular processes, including phosphoinositide homeostasis, membrane trafficking, and actin dynamics (44, 45). The best-documented function of Sac1 is to facilitate the formation of MCSs, where nonvesicular lipid exchange could take place between various organelles (41–43, 46). The ER-resident Sac1 PI(4)P phosphatase is a critical cellular protein in MCS formation/function because it allows the directional transfer of sterols from the ER to the acceptor membranes by converting PI(4)P phosphoinositide to PI phosphatidylinositol. PI(4)P is needed by oxysterol binding proteins (OSBP in mammals, Osh proteins in yeast, and ORP proteins in plants) to exchange for sterols/oxysterols/ergosterols to transfer these lipids at the MCS (47–51). PI(4)P and Sac1 also affect phosphatidylserine transport at the MCS (52).

Sac1 is a transmembrane protein with both N- and C-terminal domains exposed to the cytosol. Deletion of SAC1 in yeast and mammalian cells leads to changes in Golgi apparatus morphology and function, and sac1 knockout in a mouse is embryonically lethal. It has been reported previously that Sac1 is localized to the Golgi membranes only when cells are starved for nutrients or growth factors but remains in the ER under normal growth conditions (45, 53, 54).

Mutations in Sac1 are involved in several human diseases, including the progression of breast cancer cells by regulating PI(4)P levels and the expression of CD44, a transmembrane receptor, which regulates cell adhesion and migration in breast cancer cells (45, 55–57).

There are nine Sac1 domain proteins in Arabidopsis thaliana (AtSac1 proteins), and three of them, namely, Sac1a, Sac1b, and Sac1c, could complement the sac1-null (sac1Δ) mutant of yeast; these proteins are also localized to the ER and have PI(4)P phosphoinositide phosphatase activity (58–60).

Identification of the critical proviral function of Sac1 opens up the possibility that PI(4)P phosphoinositide might affect tombusvirus replication. PI(4)P, similar to other PI phosphoinositides, is a minor, yet essential, component of the cytosolic leaflet of certain eukaryotic cell membranes (53). In general, phosphorylation or dephosphorylation of PI phosphoinositides, which are guided by responses to acute signaling inputs, provides spatial and temporal cues within the cell for phosphoinositide-binding proteins. These, in turn, promote the assembly of macromolecular complexes and initiate several physiological events, including membrane trafficking and actin dynamics (53).

In this paper, we find that deletion or depletion of Sac1 affected TBSV replication in yeast and plant cells. We document reduction in the size of viral replication compartments in sac1Δ yeast. Moreover, the effect of Sac1 seems to be direct, based on interaction with the p33 replication protein, copurification with the tombusvirus replicase, and its presence in the TBSV replication compartment in yeast cells. Altogether, we demonstrate a critical role for Sac1 in TBSV replication and propose that the proviral functions of Sac1 might include manipulation of lipid composition within VRCs. In addition, Sac1 might serve as an assembly platform for VRCs within the extensive TBSV replication compartments in cells.

RESULTS

The cellular Sac1 lipid phosphatase is critical for tombusvirus replication in yeast and plants.

To test the role of cellular Sac1p in tombusvirus replication, we launched TBSV replicon RNA (repRNA) replication in sac1Δ yeast from plasmids expressing the viral p33 and p92^pol replication proteins together with DI-72 repRNA. The TBSV repRNA, which contains four noncontiguous segments from the genomic RNA (gRNA), can replicate efficiently in yeast and plant cells expressing p33 and p92^pol (16, 61). Northern blot analysis revealed low-level TBSV repRNA replication (11 to 23% of that found in wild-type [WT] yeast) (Fig. 1A and B, compare lanes 4 to 6 to lanes 1 to 3). Western blot analysis showed that the tombusvirus p33 and p92^pol replication proteins were expressed at close to WT level (Fig. 1B). Plasmid-based expression of His₆-tagged Sac1p in sac1Δ yeast increased TBSV repRNA accumulation to ∼170% (Fig. 1C, lanes 10 to 12), indicating that the defect in sac1Δ yeast can be complemented. Overexpression of Sac1p in WT yeast resulted in an ∼3-fold increase in TBSV repRNA accumulation (Fig. 1C, lanes 7 to 9 versus lanes 1 to 3), suggesting that the amount of Sac1p available for TBSV replication is limited in WT yeast. Altogether, these data firmly established that cellular Sac1p is a critical proviral factor for TBSV replication in yeast.

FIG 1.

Sac1p PI(4)P phosphatase protein is an essential host factor for tombusvirus replication in yeast and plants. (A and B) Deletion of Sac1 inhibits TBSV replication in yeast. (Top gels) Northern blot analyses of TBSV repRNA using a 3′-end-specific probe demonstrate reduced accumulation of repRNA in the sac1Δ yeast strain (lanes 4 to 6) in comparison with the WT yeast strain (lanes 1 to 3). A plasmid-borne copper-inducible CUP1 promoter was used to express the viral proteins His₆-p33 and His₆-p92, while DI-72 (+)repRNA was expressed from the constitutive ADH1 promoter. Standard errors were calculated based on three separate repeats. TBSV repRNA accumulation was measured with a phosphorimager and normalized based on the 18S rRNA level. (Second gels) Northern blotting with 18S rRNA-specific probe was used as a loading control. (Bottom gels) Western blot analysis of the level of His₆-tagged proteins with anti-His antibodies. (C) Complementation of TBSV replication with His₆-Sac1 in the sac1Δ yeast strain. (Top gel) Northern blot analyses of repRNA using a 3′-end-specific probe. The viral proteins Flag-p33 and Flag-p92 were expressed from the CUP1 promoter, whereas DI-72 (+)repRNA was expressed from the galactose-inducible GAL1 promoter. (Second gel) Northern blot with 18S rRNA-specific probe used as a loading control. (Bottom gels) Western blot analysis of the levels of His₆-tagged Sac1 protein and Flag-p33 with anti-His and anti-Flag antibodies, respectively. Each experiment was repeated three times. (D and E) (Top gels) Accumulation of the TBSV gRNA in Sac1a-silenced or Sac1b-silenced N. benthamiana plants 2 days postinoculation (dpi) in the inoculated leaves was measured by Northern blot analysis. Inoculation of TBSV gRNA was done 8 days after starting VIGS. Agroinfiltration of the tobacco rattle virus (TRV) vector carrying NbSac1a, NbSac1b, or C-terminal GFP (as a control) sequence was used to induce VIGS. (Second gels) rRNA is shown as a loading control in an ethidium-bromide-stained agarose gel. (D) (Third gel) RT-PCR analysis of NbSac1a mRNA levels in the silenced and control plants. (Bottom) RT-PCR analysis of tubulin mRNA levels in the silenced and control plants. Each experiment was repeated three times. cGFP, C-terminal domain of GFP. (F) Reduced TBSV RNA replication by the reconstituted tombusvirus replicase in cell extracts (CFE) prepared from sac1Δ yeast in comparison to a similar CFE preparation obtained from WT yeast. Purified recombinant MBP-p33 and MBP-p92^pol replication proteins of TBSV and in vitro-transcribed TBSV DI-72 (+)repRNA were added to the CFEs. The mixed CFE in lanes 3 and 4 contained the membrane fraction from sac1Δ CFE and the soluble fraction obtained from the WT (BY4741) yeast CFE. (G) Nondenaturing PAGE analysis showing the ³²P-labeled TBSV repRNA products, including the (+)repRNA progeny and the dsRNA replication intermediate, made by the reconstituted TBSV replicase. The CFEs used were from either WT yeast (BY4741) or Sac1-depleted GAL::SAC1 yeast strains (cultured in a raffinose medium).

To test the relevance of Sac1 in TBSV replication in a plant system, we focused on Sac1a/b/c, which were shown to have PI(4)P phosphatase activities and could complement SAC1 deletion in yeast (59, 62). First, we used a virus-induced gene silencing (VIGS) approach to deplete the Sac1a level in Nicotiana benthamiana, which resulted in downregulation of Sac1a mRNA (Fig. 1D). Replication of the infectious TBSV gRNA was decreased by ∼3-fold in the Sac1a knockdown plants compared to the control nonsilenced plants (Fig. 1D, lanes 1 to 8 versus lanes 9 to 16). Second, VIGS-based silencing of Sac1b also resulted in an ∼3-fold reduction in TBSV replication in N. benthamiana (Fig. 1E). Altogether, these data suggest that cellular Sac1 is required for TBSV accumulation in plants. Thus, similar to the results obtained in yeast, the SAC1 genes are major proviral factors in tombusvirus replication in plants.

Reduced in vitro replication of TBSV in the absence of Sac1.

To obtain more direct evidence supporting the proviral role of Sac1p in TBSV replication, we prepared cell-free extracts (CFEs) from sac1Δ and WT yeasts, followed by in vitro reconstitution of the tombusvirus replicase using purified recombinant TBSV p33 and p92^pol replication proteins and T7-transcription-based repRNA template (63). CFEs from sac1Δ yeast supported ∼90% less efficient in vitro replication of TBSV repRNA than CFEs prepared from WT yeast (Fig. 1F, compare lanes 5 and 6 with lanes 1 and 2).

Since Sac1p is an ER-resident protein (45), we mixed the membrane fraction of CFEs from sac1Δ yeast with the soluble fraction from WT CFEs to assemble the replicase in vitro, followed by the replication assay. Interestingly, TBSV repRNA replication was supported at an ∼4-fold-reduced level (Fig. 1F, lanes 3 and 4). Thus, the above-mentioned results demonstrate that the cellular Sac1p present in the membrane fraction of yeast is important for TBSV replication in vitro.

To test if either negative-strand synthesis or positive-strand synthesis of TBSV is inhibited when Sac1p is depleted, we used the CFE-based replicase reconstitution assay with purified recombinant p33 and p92^pol replication proteins and (+)repRNA template (63). The double-stranded RNA (dsRNA) replication intermediate product produced during minus-strand synthesis (64) was reduced by ∼4-fold, whereas the (+)repRNA was decreased by ∼3-fold in yeast with depleted Sac1p levels (Fig. 1G). Thus, depletion of Sac1p interfered with the synthesis of both RNA strands, which is likely the consequence of inhibition of the in vitro assembly of TBSV replicase.

The tombusvirus replication protein interacts with the ER-localized Sac1 protein.

To explore if cellular Sac1p interacts with the TBSV p33 replication protein, we performed copurification experiments from yeast coexpressing Flag-tagged p33 and 3× hemagglutinin (3×HA)-tagged Sac1p. The yeast strain expressed 3×HA-tagged Sac1p under the native promoter from the original chromosomal location in the absence of WT Sac1p. After detergent solubilization of the membrane fraction and immobilization of Flag-p33 in the Flag-binding column, we eluted the bound proteins from the column, followed by Western blotting analysis. These assays revealed that Sac1p was copurified with p33 (Fig. 2A), suggesting that the proteins form a complex in yeast membranes. Similarly prepared yeast extract containing His₆-p33 excluded nonspecific binding by 3×HA-tagged Sac1p to the column. To further confirm the interaction, we conducted a split-ubiquitin-based membrane yeast two-hybrid (MYTH) assay (65). We found that p33 replication protein interacted with Sac1p in the MYTH assay (Fig. 2B), comparable to the interactions previously demonstrated between p33 and a select group of the yeast Osh proteins (oxysterol binding protein 1-like OSBP1/ORP proteins) (39).

FIG 2.

Interaction between tombusvirus replication proteins and Sac1. (A) Copurification of yeast 3×HA-Sac1 expressed under the natural promoter and from the original chromosomal location with TBSV Flag-p33 and Flag-p92^pol replication proteins from subcellular membranes. (Top two gels) Western blot analysis of copurified 3×HA-Sac1 (lanes 2) with Flag affinity-purified Flag-p33. HA-Sac1 and Flag-p33 were detected with anti-HA and anti-Flag antibodies, respectively. The negative control was from yeast expressing His₆-p33 and His₆-p92^pol purified in a Flag affinity column (lanes 1). (Bottom gel) Western blot of total 3×HA-Sac1 in the total yeast extracts. (B) The split ubiquitin (Ub)-based MYTH assay was used to test binding between TBSV p33 and the yeast Sac1 proteins in yeast. The bait p33 was coexpressed with the prey proteins shown. Ssa1p Hsp70 and the empty prey vector (NubG) were used as positive and negative controls, respectively. (Left) p33-Sac1 interactions. (Middle left) Demonstration that comparable amounts of yeasts were used for the experiments. (Middle right and right) Demonstration that the interaction of Sac1 with p33 is comparable to the interaction of the Osh proteins shown with p33. (C) Confocal microscopy images showing colocalization of TBSV RFP-p33 replication protein and GFP-Sac1 in sac1Δ yeast coexpressing p92^pol and repRNA. (D) Confocal microscopy images showing localization of GFP-Sac1 in sac1Δ yeast in the absence of TBSV components. DIC, differential inference contrast.

To examine if the cellular Sac1p and the tombusvirus p33 replication protein are colocalized in yeast, we coexpressed the red fluorescent protein (RFP)-tagged p33 and the green fluorescent protein (GFP)-Sac1p in sac1Δ yeast cells. Confocal laser microscopy showed the colocalization of a fraction of RFP-p33 with a portion of GFP-Sac1 (Fig. 2C). The presence of Sac1p in large punctate structures formed during p33 expression in yeast (32) indicates that a fraction of Sac1p molecules are recruited into the tombusvirus replication compartment. In contrast, separate expression of GFP-Sac1p showed the characteristic ER distribution of Sac1p, which was not present in large punctate structures in the absence of the tombusvirus replication protein (Fig. 2D).

To extend our findings to plants, we expressed blue fluorescent protein (BFP)-tagged TBSV p33 replication protein and RFP-tagged Arabidopsis Sac1 by agroinfiltration in N. benthamiana leaves and followed their localization with confocal microscopy. The formation of p33-BFP-decorated replication compartments, with or without TBSV infection, showed the efficient colocalization of AtSac1 within the large TBSV replication compartments consisting of the aggregated peroxisomes, which were marked with GFP-SKL (Fig. 3A). In the absence of viral components, AtSac1 did not colocalize with the peroxisomal or nuclear markers (RFP-H2B) (Fig. 3A) or with the mitochondria (Fig. 4A).

FIG 3.

Recruitment of Sac1 by the TBSV p33 replication protein into the viral replication compartment in N. benthamiana. (A) (Top two rows) Confocal microscopy images showing efficient colocalization of TBSV p33-BFP replication protein and RFP-AtSac1 within the viral replication compartment marked by GFP-SKL peroxisomal marker in N. benthamiana leaves. Expression of these proteins from the 35S promoter was done after coagroinfiltration into N. benthamiana leaves. The plant leaves were either mock inoculated or TBSV infected, as shown. Scale bars, 10 μm. (Bottom two rows) Confocal microscopy images showing different localization patterns for AtSac1 in comparison with GFP-SKL peroxisomal marker or RFP-H2B nuclear marker in N. benthamiana leaves. (B) (Top row) GFP-AtSac1 is recruited into the replication compartment, which supports viral RNA synthesis. The viral (+)repRNA carried six copies of the RNA hairpin (MS2hp) from the MS2 bacteriophage, which is specifically recognized by the MS2 coat protein (RFP-MS2-CP). The replication compartment was marked by the BFP-tagged p33 replication protein in N. benthamiana. (Bottom row) Images of mock-inoculated plants (no viral RNA replication). Note that RFP-MS2-CP contains a weak nuclear localization signal; therefore, it is present in the nucleus in the absence of replicating (+)repRNA-MS2hp in the cytosol. Expression of the above-mentioned proteins from the 35S promoter was done after coagroinfiltration into N. benthamiana leaves. The images were taken 3.5 days after agroinfiltration of plant leaves. Scale bars. 10 μm. Each experiment was repeated three times. (C) (Top row) Interactions between TBSV p33-cYFP replication protein and the nYFP-AtSac1 protein were detected by BiFC. The merged image shows the efficient colocalization of RFP-SKL with the BiFC signal, indicating that the interaction between the p33 replication protein and AtSac1 takes place at the aggregated peroxisomes. (Bottom row) Negative control for the BiFC experiment.

FIG 4.

Recruitment of Sac1 by the CIRV p36 replication protein into the mitochondrion-derived viral replication compartment in N. benthamiana. (A) Confocal microscopy images showing efficient colocalization of CIRV p36-BFP replication protein and RFP-AtSac1 within the viral replication compartment, marked by the GFP-AtTim21 mitochondrial marker in N. benthamiana leaves. Further details are shown in Fig. 3A. (B) Interaction between CIRV p36-cYFP replication protein and the nYFP-AtSac1 protein was detected by BiFC. The merged images show efficient colocalization of RFP-AtTim21 with the BiFC signal, indicating that the interaction between the CIRV p36 replication protein and AtSac1 occurs at the aggregated mitochondria.

To show that the recruitment of AtSac1 takes place in the active TBSV replication compartment, we monitored the accumulation of the replicating (+)repRNA with a biosensor detecting the newly generated (+)repRNA based on the incorporation of the MS2 bacteriophage RNA sequence, which is specifically recognized by the RFP-tagged MS2 coat protein (Fig. 3B) (66). These experiments confirmed the colocalization of co-opted AtSac1 with the replication compartment decorated with p33-BFP in N. benthamiana. Therefore, Sac1 likely plays a role in the formation of the tombusvirus replication compartments.

To provide additional evidence that the plant AtSac1 interacts with the TBSV p33 replication protein, we conducted bimolecular fluorescence complementation (BiFC) experiments with p33 and AtSac1 in N. benthamiana leaves. The BiFC signals revealed specific interactions between AtSac1 and p33 replication protein at the aggregated peroxisomes (Fig. 3C). Based on these data, we suggest that the cellular Sac1 lipid phosphatase is actively recruited into the large replication compartment induced by tombusviruses.

Similar confocal-microscopy-based experiments with the closely related carnation Italian ringspot virus (CIRV), which, unlike TBSV, builds replication compartments from aggregated mitochondria (67, 68), revealed the colocalization of AtSac1 with the CIRV p36-RFP replication protein and the GFP-AtTim21 mitochondrial membrane protein (Fig. 4A). We also conducted BiFC experiments, which revealed specific interactions between AtSac1 and the CIRV p36 replication protein at the aggregated mitochondria (Fig. 4B). Taken together, these data suggest that AtSac1 is recruited into the large mitochondrion-based CIRV replication compartment in plant cells.

Sac1 protein is recruited into tombusvirus-induced membrane contact sites.

One of the known cellular functions of the ER-resident Sac1p is performed at MCSs between the ER and the plasma membrane (PM) or Golgi membrane, which are needed to transport sterols to the PM and PI(4)P to the ER as a counterflow. By converting PI(4)P to PI in the ER, Sac1p makes the lipid flow directional (42, 43). TBSV p33 replication protein has been shown to stabilize MCSs, termed vMCSs, between the ER and peroxisomes through interacting with the ER-resident Scs2p VAP protein and the OSBP1-like ORP proteins, such as Osh3p, -5, -6, and -7 (37, 40). To test the subcellular location of Sac1p in yeast replicating TBSV and if Sac1p is also recruited to vMCSs, we performed confocal-microscopy-based colocalization and BiFC experiments between p33 and Scs2p VAP tethering protein (69) in yeast. These studies indicated that Sac1p is recruited to the subcellular sites of p33-Scs2p interaction (Fig. 5A), which has been shown to take place in dedicated subdomains in the ER membranes, as part of vMCSs (39).

FIG 5.

Effect of vMCS proteins on recruitment of Sac1 by the tombusvirus p33 replication protein in yeast. (A) Confocal microscopy images showing localization of yeast CFP-tagged Sac1p to vMCS marked by interaction between the p33 replication protein and the yeast Scs2p VAP protein, which was detected by BiFC. (B) Partial colocalization of ergosterols, RFP-p33, and GFP-Sac1 shown by confocal laser microscopy. Yeasts were stained with filipin dye to detect ergosterols. p92^pol and the repRNA were not expressed in these yeast cells. (C) SAC1 deletion prevents the formation of large sterol-enriched compartments in yeast replicating TBSV repRNA. Note the similar distributions of ergosterol in sac1Δ yeast with or without TBSV replication. Yeasts were stained with filipin dye to detect ergosterols. (D) Copurification of Sac1p and Osh7p proteins with the tombusvirus p33 replication protein from double-VAP-deletion (scs2Δ scs22Δ) (ΔΔ) yeast versus WT yeast. The Flag-tagged p33 was purified from the membrane fractions of yeast extracts using a Flag affinity column. Note that yeast coexpressed p92^pol to support the replication of repRNA. (Top gel) Western blot analysis of copurified His₆-tagged Sac1p or His₆-Osh7 using anti-His antibody. The lane 4 sample was from WT yeast overexpressing Scs2p from a plasmid. (Middle gel) Western blot of purified Flag-p33 replication protein detected with anti-Flag antibody. (Bottom gel) Western blot of His₆-tagged Sac1p or His₆-Osh7 in the total yeast extracts using anti-His antibody. (E) Copurification of Sac1p and Osh5, -6, and -7 proteins with the Scs2 VAP protein from WT yeast replicating TBSV repRNA (coexpression of p33, p92, and repRNA; marked with “+”) or without viral components (marked with “−”).Flag-tagged Scs2p was purified from the membrane fractions of yeast extracts using a Flag affinity column. (Top gel) Western blot analysis of copurified His₆-Osh5, His₆-Osh6, His₆-Osh7, or His₆-Sac1 using anti-His antibody. (Middle gel) Western blot of purified Flag-Scs2p detected with anti-Flag antibody. (Bottom gel) Western blot of His₆-Osh5, His₆-Osh6, His₆-Osh7, or His₆-Sac1 in total yeast extracts using anti-His antibody.

The vMCS-assisted enrichment of sterols at the replication sites is critical for TBSV replication (39). Accordingly, we colocalized RFP-p33, GFP-Sac1p, and ergosterol (stained with filipin) in yeast cells (Fig. 5B). The distribution of ergosterol changed in sac1Δ yeast, showing uneven, small, punctate-like structures, in contrast to the WT yeast, which showed the characteristic, mostly plasma membrane enrichment of ergosterol (Fig. 5C). Interestingly, the expression of the viral components in sac1Δ yeast also showed the uneven, punctate-like distribution of ergosterol, similar to the distribution in sac1Δ yeast not expressing viral components (Fig. 5C). As we demonstrated previously (39), the distribution of ergosterol is very different in WT yeast cells replicating TBSV RNA, showing large intracellular ergosterol-enriched punctate structures (Fig. 5C). Based on these data, we suggest that ergosterols are not efficiently redistributed to the sites of viral replication in sac1Δ yeast. Altogether, our data suggest that Sac1p is recruited into the tombusvirus replication compartment, including into vMCSs, to facilitate the redistribution of sterols to the sites of viral replication.

To test if recruitment of Sac1p to the replication compartment is affected by the ER-resident Scs2p VAP tethering protein (and its paralog, Scs22p) (69), which is required for vMCS formation (39), we purified the tombusvirus replicase from scs2Δ scs22Δ yeast, followed by measuring the copurified His₆-Sac1p by Western blotting. In comparison with the replicase preparation from WT yeast, the replicase preparation from scs2Δ scs22Δ yeast contained a comparable amount of Sac1p (Fig. 5D, compare lanes 6 and 7), suggesting that the recruitment of Sac1p to the viral replication compartment is not dependent on the Scs2p/Scs22p VAP proteins. Similarly, the amount of Osh7p ORP in the replication compartment is only slightly affected by deletion of Scs2p/Scs22p VAP genes or overexpression of Scs2p in WT yeast (Fig. 5D, compare lanes 2, 3, and 4). These data suggest that p33 replication protein might recruit Sac1p and Osh7p cellular factors independently of Scs2p/Scs22p VAP proteins to the viral replication compartment in yeast.

To examine if the amounts of the co-opted cellular protein components of the virus-induced vMCSs are increased by p33/p92 replication proteins, we purified Flag-Scs2p VAP protein from detergent-solubilized membrane fractions of yeast replicating TBSV repRNA or missing the viral components as a control. As expected based on prior work (43, 69), His₆-Osh6p, His₆-Osh7p, and His₆-Sac1p were copurified with Flag-Scs2p from yeast lacking viral components (Fig. 5E). Interestingly, the amounts of His₆-Osh5p, His₆-Osh6p, His₆-Osh7p, and His₆-Sac1p copurified with Flag-Scs2p were increased in the presence of viral components (Fig. 5E). These data suggest that the expressed p33/p92 replication proteins stabilize the complexes containing the cellular Scs2p VAP, ORP proteins, and Sac1p, likely within vMCSs.

Testing TBSV replication in osh3Δ, -6Δ, and -7Δ yeast with depleted Sac1p showed undetectable TBSV repRNA accumulation levels (Fig. 6A to C), suggesting that both the ORP and Sac1p proteins are critical for TBSV replication in yeast. The additional inhibitory effect of Sac1p depletion on TBSV repRNA accumulation in osh3Δ, -6Δ, and -7Δ yeast (Fig. 6C) in comparison with osh3Δ, -6Δ, and -7Δ yeast expressing high levels of Sac1p (Fig. 6B) indicates that Sac1p likely has other proviral functions in addition to its roles in vMCS and sterol enrichment (see below).

FIG 6.

Sac1p PI(4)P phosphatase and the oxysterol binding proteins do not have complementary functions in tombusvirus replication in yeast. (A) Scheme of the expression strategy for Sac1p in WT yeast versus triple-deletion (osh3,6,7Δ) yeast. The His₆-Sac1p was expressed using the galactose-inducible GAL1 promoter from its original chromosomal location in these haploid yeast strains. (B and C) Depletion of Sac1 inhibits TBSV replication in both WT and triple-deletion (osh3,6,7Δ) yeast strains. (Top gels) Northern blot analyses of repRNA using a 3′-end-specific probe demonstrating reduced accumulation of repRNA in either Sac1-induced (B) or Sac1-depleted (C) triple-deletion (osh3,6,7Δ) yeast strains. A plasmid-borne copper-inducible CUP1 promoter was used to express the viral proteins His₆-p33 and His₆-p92, while DI-72 (+)repRNA was expressed from the constitutive ADH1 promoter. (C) Note that we loaded twice as much RNA in lanes 4 to 6 than in lanes 1 to 3 to help visualize small amounts of repRNA. (Middle gels) Northern blotting with an 18S rRNA-specific probe was used as a loading control. (Bottom gel) Western blot analysis of the level of His₆-tagged p33 replication protein with anti-His antibodies. We loaded the same amount of total protein in each sample. (D) Western blot analysis of the levels of His₆-Sac1p in the indicated yeast strains cultured under repressive (glucose), inducing (galactose), or noninducing (raffinose) conditions.

PI(4)P phosphoinositide is required for tombusvirus replication.

Identification of the critical proviral function of Sac1p, which is a PI(4)P phosphatase (70), opens up the possibility that PI(4)P phosphoinositide might also affect tombusvirus replication. Therefore, we tested if depletion of PI(4)P kinases, which generate PI(4)P from PI, affects TBSV replication in yeast. Separate depletion of Stt4p and Pik1p PI(4)P kinases inhibited TBSV replication by ∼4-fold (Fig. 7A and B), whereas deletion of the LSB6 PI(4)P kinase gene in yeast had no effect (Fig. 7C). These data indicate that Stt4p and Pik1p PI(4)P kinases, which account for the vast majority of PI4 kinase (PI4K) activity in yeast cells (71), are critical host factors in TBSV replication, likely by providing PI(4)P.

FIG 7.

Depletion of PI4K enzymes reduces tombusvirus replication in yeast. (A and B) Depletion of Stt4p levels in TET::STT4 and Pik1p in TET::PIK1 yeast strains inhibits TBSV replication in yeast. Doxycycline (Dox) was used to downregulate the expression of Stt4p and Pik1p from the TET promoter. Replication of the TBSV repRNA in TET::STT4 and TET::PIK1 yeast strains, respectively, coexpressing the tombusvirus p33 and p92 replication proteins was measured by Northern blotting 16 h after initiation of TBSV replication. The accumulation level of His₆-p33 replication protein was tested by Western blotting and anti-His antibody. (C) Deletion of LSB6 PI4K does not inhibit TBSV replication in yeast. Northern and Western blot analyses were performed as for panel A. (D) Inhibition of PI4K enzymes by PAO decreases TBSV replication in N. benthamiana protoplasts. Three-week-old N. benthamiana leaves were used for protoplast preparations, which were transfected with TBSV RNAs using PEG4000. Incubation of protoplasts with PAO (50 μM) was for 16 h. Each experiment was repeated three times.

To confirm the role of PI(4)P in plants, we used phenylarsine oxide (PAO), a specific inhibitor of PI(4)P kinases (72, 73) in N. benthamiana protoplasts replicating TBSV gRNA. PAO treatment resulted in undetectable accumulation of TBSV RNAs in plant cells (Fig. 7D), suggesting that PI(4)P is also critically important for tombusviruses in plant cells.

To test if PI(4)P is localized within the tombusvirus replication compartment, we coexpressed the GFP-tagged FAPP-PH domain, which specifically binds to PI(4)P (11, 74) in yeast. The yeast cells also coexpressed RFP-p33 and p92^pol replication proteins and the TBSV repRNA to support the formation of the viral replication compartment and TBSV replication. Confocal microscopy revealed that GFP-FAPP-PH was partially colocalized with the RFP-p33 replication protein in yeast cells (Fig. 8A), indicating that PI(4)P was present within the viral replication compartment. The partial colocalization of p33 replication protein and PI(4)P phosphoinositide in yeast cells might be affected by the efficient utilization of PI(4)P by Sac1 at vMCSs, leading to reduction in PI(4)P levels and sterol enrichment within the viral replication compartment.

FIG 8.

Partial colocalization of PI(4)P phosphoinositide with the TBSV p33 replication protein in yeast and in N. benthamiana. (A) Confocal microscopy images showing partial colocalization of TBSV RFP-p33 with GFP-FAPP-PH, which specifically binds to PI(4)P in yeast. The arrows point to the areas where colocalization was detected. The yeast also coexpressed p92^pol and repRNA to support TBSV replication. (B) Expression of the Legionella DrrA effector, which binds strongly to PI(4)P through its C-terminal P4M domain, eliminates colocalization of TBSV RFP-p33 with GFP-FAPP-PH in yeast cells. Note that the p33 replication protein cannot form the large replication compartment in these cells. Further details are shown in panel A. (C) Confocal microscopy images showing partial colocalization of TBSV p33-RFP with GFP-FAPP-PH in N. benthamiana. The arrows point to the areas where colocalization was detected. The boxed images on the right are different from those on the left. Expression of these proteins from the 35S promoter was done after coagroinfiltration into N. benthamiana leaves replicating TBSV. (D) Expression of the Legionella DrrA effector interferes with the colocalization of TBSV p33-RFP with GFP-FAPP-PH in N. benthamiana cells. Note that p33 replication protein cannot form the large replication compartment in these cells. Further details are shown in panel C. (E) Confocal microscopy images showing partial colocalization of either p33-GFP or p33-BFP and PI(4)P in N. benthamiana protoplasts. N. benthamiana plants were agroinfiltrated to express either p33-GFP or p33-BFP. The next day, N. benthamiana protoplasts were prepared from the agroinfiltrated plants and transfected with TBSV gRNA. After 16 h of incubation, the protoplasts were fixed. Note that PI(4)P was detected by PI(4)P-specific antibody after the permeabilization of protoplasts. (F) Lack of colocalization of peroxisomes, marked by GFP-Pex13 and PI(4)P in N. benthamiana protoplasts lacking viral components. Note that PI(4)P was detected by PI(4)P-specific antibody. (G) Expression of the Legionella DrrA effector prevents the colocalization of TBSV p33-RFP or p33-BFP with PI(4)P in N. benthamiana protoplasts. Agroinfiltration was used to express DrrA, p33-GFP, and p19 suppressor of gene silencing for 1 day in N. benthamiana leaves, followed by protoplast isolation. Then, the protoplast preparations were transfected with TBSV RNA, followed by incubation for 16 h. PI(4)P was detected in cells after fixation and immunofluorescence detection with anti-PI(4)P antibody. Each experiment was repeated three times.

To further test the localization pattern of PI(4)P during virus replication, we expressed the DrrA effector protein of Legionella pneumophila, which through its C-terminal P4M domain binds exceptionally strongly to PI(4)P (75), thus possibly sequestering this minor lipid species away from the viral replication compartment inside the yeast cells. Expression of the DrrA effector changed the localization pattern of GFP-FAPP-PH (Fig. 8B). Moreover, RFP-p33 replication protein did not colocalize with GFP-FAPP-PH (Fig. 8B), suggesting that TBSV failed to recruit or enrich PI(4)P at the site of replication in the presence of the DrrA effector. We also noticed that DrrA expression inhibited the formation of the characteristic large replication compartments visible in WT yeast cells (compare Fig. 8A and B).

Using a similar approach in N. benthamiana plants through expression of GFP-FAPP-PH, we showed that PI(4)P accumulated within the large viral replication compartment detected via expression of p33-RFP in epidermal cells infected with TBSV (Fig. 8C). Expression of the DrrA effector in N. benthamiana leaves dramatically changed PI(4)P distribution and interfered with the formation of large replication compartments (Fig. 8D). Similar observations were made in N. benthamiana protoplasts replicating TBSV with PI(4)P-specific antibody. They included partial colocalization (enrichment) of PI(4)P with the p33 replication protein (Fig. 8E) and the absence of colocalization of PI(4)P with p33 when protoplasts were also coexpressing the DrrA effector (Fig. 8G). As expected based on the known localization pattern for PI(4)P, which includes the Golgi apparatus and the plasma membrane in plant cells (58, 60), PI(4)P was not colocalized with the peroxisomal marker in mock-transfected N. benthamiana protoplasts (Fig. 8F). Based on the above-mentioned data, we suggest that PI(4)P has a proviral function in tombusvirus replication and the DrrA effector, due to its strong binding to PI(4)P, sequesters PI(4)P, thus interfering with the enrichment of PI(4)P at the sites of tombusvirus replication.

Sac1 protein affects the recruitment of syntaxin18-like Ufe1p SNARE for tombusvirus replication.

The formation of the TBSV replication compartment is a complex process depending on the ER-resident SNARE proteins, including the syntaxin18-like Ufe1p (38, 76). We have previously shown that depletion of Ufe1p inhibits the formation of tombusvirus replication compartments via multiple mechanisms (38). To study if recruitment of the ER-resident Ufe1p into the viral replication sites was affected by Sac1p, we examined the subcellular distribution of RFP-Ufe1p in sac1Δ yeast via confocal microscopy. We observed that Ufe1p colocalized poorly with p33 replication protein within the small p33-containing punctate structures formed in sac1Δ yeast in contrast with the more extensive colocalization between Ufe1p and p33 in WT yeast at two different time points (Fig. 9C and D versus Fig. 9A and B).

FIG 9.

Recruitment of the syntaxin18-like Ufe1 SNARE by the TBSV p33 replication protein is inhibited in sac1Δ yeast. (A and B) Confocal microscopy images showing the colocalization of GFP-p33 with RFP-Ufe1 in WT yeast at 16 and 24 h of incubation. (C and D) Confocal microscopy images showing lack of colocalization of GFP-p33 with RFP-Ufe1 in sac1Δ yeast at 16 and 24 h of incubation. Note that the yeast cells did not express p92^pol and repRNA. Each experiment was repeated three times.

To test if Sac1p affects the recruitment of Ufe1p to the vMCSs, we visualized vMCSs via BiFC using VN-p33 and VC-Scs2 VAP protein (39). We found that Ufe1p was poorly colocalized with p33-Scs2p-decorated membranes in sac1Δ yeast in contrast with the more extensive colocalization in WT yeast (Fig. 10B versus Fig. 10A). These data support a model in which Sac1p is critical for the efficient recruitment of Ufe1p SNARE protein into the p33-containing replication compartment and to the vMCS.

FIG 10.

Ufe1 SNARE recruitment into vMCSs is inhibited in sac1Δ yeast. (A and B) Confocal microscopy images showing colocalization of RFP-Ufe1 with the BiFC signal from VN-p33 and VC-Scs2, which marks vMCSs, in WT and sac1Δ yeasts, respectively. Note that the yeast cells did not express p92^pol and repRNA. Each experiment was repeated three times.

Sac1p is required for the efficient recruitment of p33-driven Rab5-positive endosomes into the replication compartment in yeast.

The biogenesis of the large TBSV replication compartments, which consist of the Scs2p VAP-containing vMCSs and the Ufe1p-containing ER arrival site (ERAS) subdomains of the ER and numerous VRCs within aggregated peroxisomes, depends on retargeting of Rab5-positive endosomes, which are PE rich, to the sites of viral replication (19, 26, 30). To test if Sac1p affects the recruitment of Rab5-positive endosomes, we used confocal microscopy in combination with BiFC analysis. These experiments showed that the Rab5 (called Vps21 in yeast)-positive endosomes were not colocalized with the p33 replication protein-Scs2p VAP complex in sac1Δ yeast (Fig. 11B), whereas we observed partial colocalization in WT yeast (Fig. 11A).

FIG 11.

The yeast Rab5 protein is not recruited into vMCSs in sac1Δ yeast. (A and B) Confocal microscopy images showing colocalization of the Vps21-RFP (Rab5 homolog) with the BiFC signal from VN-p33 and VC-Scs2, which marks vMCSs, in WT and sac1Δ yeasts, respectively. Note that the yeast cells did not express p92^pol and repRNA. Each experiment was repeated twice. (C) Enrichment of PE phospholipid, visualized with duramycin, within the GFP-p33-decorated viral replication compartment in WT yeast detected by confocal microscopy. The yeast also coexpressed p92^pol and repRNA to support TBSV replication. (D) Lack of enrichment of PE phospholipid, visualized with duramycin, within the GFP-p33-decorated viral replication compartment in sac1Δ yeast detected by confocal microscopy. The yeast also coexpressed p92^pol and repRNA to support TBSV replication. Each experiment was repeated three times.

To confirm the biological significance of the lack of Rab5-positive endosome recruitment to the replication compartment in sac1Δ yeast, we visualized PE distribution in sac1Δ versus WT yeasts replicating TBSV repRNA. Unlike in WT yeast, in which PE is enriched within the large replication compartment decorated with GFP-p33 replication protein (Fig. 11C), the PE distribution was more dispersed in sac1Δ yeast, and PE was not enriched within the p33-decorated replication compartment (Fig. 11D). Based on these data, we propose that Sac1p plays a role in recruitment of Rab5-positive endosomes and thus facilitates PE enrichment at the sites of tombusvirus replication, which is required for efficient tombusvirus replication.

Sac1p is required for the efficient recruitment of several host factors into the replicase complex.

The CFE-based in vitro replicase assembly experiments showed low TBSV replicase activity when the CFE was obtained from sac1Δ yeast, which might be the consequence of poor replicase assembly. Therefore, we decided to test the protein composition of the purified tombusvirus replicase, which has many co-opted cellular cofactors. Western blot analyses were performed on the detergent-solubilized affinity-purified tombusvirus replicase to detect the amounts of known co-opted cellular proteins in preparations obtained from sac1Δ or WT yeasts. This proteomic approach revealed ∼4- to 5-fold-reduced amounts of co-opted Rpn11 deubiquitinase/metalloprotease and Cdc34 E2 ubiquitin-conjugating enzyme and ∼3 times less Rad6p and Ubc2 E2 ubiquitin-conjugating enzymes, whereas Ded1p DEAD box helicase and Vps4 AAA ATPase were present in the purified replicase from sac1Δ yeast only at ∼50 to 60% of the WT level (Table 1). The co-opted Cdc34p, Rad6p, Ubc2p, and Vps4p proteins are critical factors during VRC assembly (77–79), whereas Rpn11p and Ded1p affect the activities of VRCs, including viral (+)RNA synthesis and viral RNA recombination (80–82). Therefore, the reduced recruitment of these cellular factors into tombusvirus VRCs in sac1Δ versus WT yeasts could contribute to the serious defect in viral RNA accumulation in sac1Δ yeast. On the other hand, Sac1p does not seem to affect the recruitment of Vap27-1 VAP and Osh6p ORP proteins, which are components of virus-induced vMCSs (39), suggesting that these components associate directly with p33 replication protein to promote the formation of vMCSs. Altogether, our data suggest that Sac1p affects the assembly of the tombusvirus replicase by facilitating the recruitment of several host proteins into the VRCs.

TABLE 1.

Host proteins copurified with p33 from sac1Δ yeast

Proteina	% (±SE)b
Bro1, ESCRT associated	141
Ded1 helicase	58 ± 7
Cdc34 E2 enzyme	28 ± 11
Osh6 ORP	92 ± 25
Pbp2 RNA binding	130 ± 14
Rad6 E2	32
RH2 helicase (eIF4AIII)	113 ± 44
Rpn11 deubiquitinase	22 ± 6
Tdh2 GAPDH	121 ± 14
Tef1 eEF1A	104 ± 3
Ubc2 E2	36
Vap27-1 (Scs2-like)	144 + 37
Vps4 AAA+ ATPase	49 + 13
Vps23 ESCRT I	152 + 19

^aThe table lists the co-opted host proteins, which were expressed as well in sac1Δ yeast as in WT (BY4741) yeast based on the total protein samples (prior to purification). Co-opted host proteins copurified with p33 in lesser amounts from sac1Δ yeast are shown in boldface.

^bCopurification of the given host protein with Flag-tagged p33 replication protein from WT yeast was taken as 100%. The standard error (SE) was calculated based on the results of three separate experiments.

DISCUSSION

High-throughput screening of the yeast knockout library has led to the identification of Sac1p PI(4)P phosphatase as one of the proviral cellular factors in TBSV replication in yeast (40). In this paper, we have confirmed that the yeast Sac1p and the orthologous Sac1a and Sac1b isoforms in N. benthamiana plants are important proviral factors during TBSV replication. TBSV replicase reconstitution experiments in CFEs from Sac1p-depleted yeast revealed that both minus- and plus-strand syntheses were inhibited, pointing toward deficiency in the assembly of the TBSV replicase when Sac1p is limited. The role of Sac1p seems to be direct, based on the interaction with the p33 replication protein and the copurification of Sac1p with the tombusvirus replicase from yeast.

Sac1p is recruited into the virus-induced vMCSs to facilitate enrichment of sterols within the viral replication compartment.

Based on the known cellular function of Sac1p in cellular MCSs, where Sac1p facilitates the directional transport of sterols from the ER to the PM or Golgi membrane with the help of PI(4)P (47–51), we analyzed in detail if Sac1p might also be recruited to the virus-induced vMCSs. Accordingly, we found that Sac1p is recruited to vMCSs, based on confocal and BiFC analyses. In addition, the amounts of Sac1p and Osh6 and Osh7 proteins copurified with Scs2 VAP protein, a known ER-resident tethering protein in the MCS (43), are increased in yeast replicating TBSV repRNA versus yeast lacking viral components (Fig. 5E). Moreover, we did not observe the characteristic enrichment of ergosterols within the viral replication compartment in yeast lacking the SAC1 gene (Fig. 5C). These data support the model that TBSV stabilizes vMCSs, as indicated by the enhanced level of interactions among co-opted cellular vMCS components. The vMCSs then facilitate efficient sterol enrichment within the viral replication compartment (Fig. 12). Sterols are important cofactors and bind to the TBSV p33 replication protein (25), which is the master regulator of VRC assembly and viral (+)RNA recruitment into VRCs (25, 83). Enrichment of sterols within VRC membranes is also required to build tombusvirus VRCs resilient to the antiviral activity of the cellular RNA interference (RNAi) pathway (84).

FIG 12.

Model of the functional roles of the ER-resident Sac1 lipid phosphatase and PI(4)P in the biogenesis of the virus-induced replication compartment. (A) First, the co-opted Sac1, in collaboration with the ER-resident Scs2p VAP tethering protein and the cytosolic oxysterol binding proteins (ORP and Osh3, -5, -6, and -7 in yeast), participates in the formation of vMCSs between the ER and peroxisomes, which results in the enrichment of sterols and phospholipids in the peroxisomal membrane. Sac1 likely utilizes PI(4)P by converting it to PI in the ER to allow unidirectional transport, as shown in regular MCSs in healthy cells. Second, Sac1 promotes the preassembly of VRCs on the altered peroxisomal membrane by helping the abundant p33 replication protein of TBSV in the assembly process. The co-opted Sac1 facilitates the recruitment of Rab5 (Vps21)-positive endosomal structures and additional proviral host factors, such as the ER-resident syntaxin18-like Ufe1 SNARE, Vps4 AAA ATPase ESCRT protein, Rpn11 deubiquitinase, and Ded1 DEAD box RNA helicase. Sac1 might function in the second process in a complex with the viral replication protein and Ufe1 SNARE protein, which is part of the ERAS with Use1 SNARE protein, by constituting a VRC assembly hub. The possible organizing role of Sac1 in the actin network is not shown here. Note that the model is simplified here to clearly illustrate the roles of Sac1 and PI(4)P. Overall, by subversion of Sac1, PI(4)P and the additional host components for proviral functions allow tombusviruses to build nuclease- and protease-insensitive virus replication compartment (only one spherule is shown for simplicity). (B) Confocal microscopic images showing the small viral replication compartments decorated by GFP-p33 in yeast lacking Sac1 lipid phosphatase. The yeast also coexpressed p92^pol and repRNA to support TBSV replication.

Based on the retargeting of Sac1 into the CIRV-induced replication compartment, which consists of aggregated mitochondria, and the interaction between the cellular Sac1 and the CIRV p36 replication protein, we propose that CIRV also co-opts Sac1 functions to build its replication compartments.

PI(4)P phosphoinositide is a critical host factor during tombusvirus replication.

The known substrate of Sac1p is PI(4)P, which is a minor, yet essential, component of the cytosolic leaflet of Golgi and plasma membranes (53, 85). The presence of PI(4)P in membranes facilitates the recruitment of PI(4)P-binding proteins, including Sac1p. These in turn promote the assembly of macromolecular complexes and initiate membrane trafficking and actin dynamics. Accordingly, we found a major proviral function(s) for PI(4)P during tombusvirus replication. This conclusion is supported by the findings that (i) depletion of two PI(4)P kinases, namely, Stt4p and Pik1p, in yeast or chemical inhibition of PI(4)P kinases in plant protoplasts decreased tombusvirus replication in cells; (ii) PI(4)P is present within the viral replication compartment, based on colocalization studies; and (iii) depletion of free PI(4)P via expression of the DrrA effector (from Legionella) inhibited the enrichment of PI(4)P within the replication compartment in yeast and plant cells. However, Sac1p converts PI(4)P to PI in the ER membrane, suggesting that the role of PI(4)P is likely temporal during tombusvirus replication. Altogether, the data presented firmly establish that both Sac1 and its substrate, PI(4)P, are required for TBSV replication in yeast and plants.

Sac1p affects PE enrichment in VRCs.

The proviral role of Sac1p, however, does not seem to be restricted only to sterol enrichment in the viral replication compartment via vMCSs. We also showed that deletion of SAC1 leads to deficiency in recruitment of several host factors, including the p33-driven hijacking of Rab5-decorated endosomes that deliver PE-rich membranes for robust viral replication (Fig. 12) (26, 38). Indeed, we found that Rab5 (Vps21p in yeast) is not recruited into vMCSs and PE is not enriched at the sites of replication in sac1Δ yeast (Fig. 11). The p33-containing punctate structures, which represent the sites of TBSV replication, are small in sac1Δ yeast cells in comparison with those formed in WT yeast (Fig. 12B). This is likely due to the lack of membrane proliferation and the lack of recruitment of Rab5-positive endosomes into the viral replication compartment. Accordingly, in vitro TBSV replicase assembly based on yeasts with deletion or depletion of Sac1p showed poor efficiency (Fig. 2), decreasing the production of both dsRNA replication intermediate and the (+)RNA progeny. Additional experiments will be needed to determine if the role of Sac1 is direct in the above-mentioned proviral processes.

Sac1p might also function as a scaffold protein and assembly platform to facilitate rapid generation of VRCs.

Tombusviruses are successful pathogens of plants due to their ability to replicate rapidly and efficiently in infected cells. These viruses can quickly assemble VRCs for robust replication (84). The intensive assembly of large virus-induced membranous structures likely depends not only on protein and RNA chaperones, such as Hsp70 and eEF1A (63, 86), but also on scaffolding proteins, which bring several proteins and membranes together for rapid assembly. Based on the current work, it seems that Sac1p might act as a scaffolding protein, possibly for vMCS stabilization and during TBSV VRC assembly. Accordingly, the co-opted protein composition of the TBSV replicase is dramatically different in sac1Δ yeast (Table 1).

In collaboration with Ufe1 SNARE protein, Sac1 is a component of the TBSV replicase assembly hub.

Surprisingly, we found that, in the absence of Sac1p, TBSV cannot efficiently hijack the ER-resident Ufe1p SNARE protein. Ufe1p is a major component of the ERAS in the ER subdomain, which serves as an assembly hub for several tombusvirus VRC components and greatly affects the formation of the viral replication compartment (38). We propose that subversion of the ER-resident Sac1 helps TBSV to hijack Ufe1 and the ERAS subdomain of the ER (Fig. 12). Thus, on one hand, by co-opting Sac1, TBSV gains access to local manipulation and enrichment of membrane lipids in vMCSs and ultimately in VRCs. On the other hand, the subverted Sac1 serves as an assembly platform that accelerates the efficient generation of numerous VRCs in infected cells. Accordingly, confocal microscopic analysis of sac1Δ yeast revealed tiny p33 replication protein-containing foci (punctate structures) (Fig. 12B) instead of the large punctate structures that are representative of the tombusvirus replication compartments (32, 87). This suggests a lack of discrete viral replication compartment formation in sac1Δ yeast. The co-opted Sac1 is unlikely to perform all these proviral functions by itself, but instead, as a component of different protein complexes, such as with the syntaxin18-like Ufe1 SNARE protein-p33 replication protein complexes. Altogether, the data presented are in agreement with a model in which the co-opted Sac1 plays a significant role in the formation of the tombusvirus replication compartment by bringing together vMCSs to obtain optimal lipid composition in the subverted membranes and aggregated peroxisomes, where numerous spherules/VRCs could then efficiently form. Thus, the several proviral roles of Sac1 illustrate that vMCS formation and the assembly of VRCs are likely linked, as in the model shown in Fig. 12A. These “VRC assembly hub” functions of Sac1 and Ufe1 could facilitate the sequestering and concentration of the viral replication proteins and several co-opted host proteins and critical lipids within the limited ERAS subdomains within the ER membranes. We propose that by targeting Sac1 and Ufe1 SNARE protein (and other components of the ERAS), TBSV generates an efficient VRC assembly hub that includes the ERAS subdomain of the ER and vMCS with peroxisomes (Fig. 12A).

Since (+)RNA viruses remodel subcellular membranes to facilitate virus replication and avoid antiviral responses (4–6, 10, 88, 89), it is possible that Sac1 lipid phosphatase and exploitation of MCSs could be widespread among viruses. Indeed, several animal (+)RNA viruses, such as picornaviruses, are known to depend on PI(4)P and Sac1 for replication (6, 85). Poliovirus (PV) and coxsackievirus B3 hijack the Golgi and trans-Golgi network (TGN) membranes and cellular PI4K to their replication organelles to produce PI(4)P in situ (11, 85, 90, 91). The PI(4)P-enriched membranes then recruit cellular PI(4)P-binding proteins to aid viral replication. The 3C replication protein of picornaviruses binds directly to PI(4)P to support viral RNA synthesis (90, 92). PI(4)P is also required for sterol enrichment within the replication organelles of rhinovirus, Aichi virus, and PV (93–95). Co-opted cellular PI4K kinases and oxysterol-binding proteins are also important for HCV replication (91, 96–98). In addition to viruses, several intracellular bacterial pathogens also hijack Sac1, PI(4)P, and MCSs to support the pathogenesis process (99). Therefore, it will be interesting to learn if other viruses and pathogens could also exploit Sac1 lipid phosphatase for many functions, as tombusviruses do. Altogether, Sac1 seems to be a possible target for antiviral approaches.

MATERIALS AND METHODS

Yeast strains and expression plasmids.

Saccharomyces cerevisiae strains BY4741 (MATa his3Δ1 leu2Δ0 met15Δ0 ura3Δ0) and R1158 (100), BY4741 sac1Δ (BY4741 sac1Δ::kanMX4), BY4741 Lsb6Δ (BY4741 lsb6Δ::kanMX4), TET::PIK1, and TET::STT4 (yTHC library) were obtained from Open Biosystems. A yeast expression plasmid, pRS406-PHO5-GFP-hFAPP1-PH, was obtained from Addgene. Yeast strain NMY51 was obtained from Dualsystems. Yeast strains SEY6210 (MATa ura3-52 his3Δ200 lys2-801 leu2-3,112 trp1Δ901 suc2Δ9) and JRY6232 (SEY6210 osh5Δ::LEU2 osh6Δ::LEU2 osh7Δ::HIS3) were obtained from Christopher T. Beh (Simon Fraser University).

To generate yeast strain JRY6232 his3Δ, the KanMX4 gene was PCR amplified from plasmid pYM45 (European Saccharomyces cerevisiae Archive for Functional Analysis [EUROSCARF]) (101) using primers 5489 (ATGACAGAGCAGAAAGCCCTAGTAAAGCGTATTACAAATGAAACGTACGCTGCAGGTCGA) and 5490 (CTACATAAGAACACCTTTGGTGGAGGGAACATCGTTGGTACCATCGATGAATTCGAGCTC). The PCR product was used for transformation that created strain SEY6210 (osh5Δ::LEU2 osh6Δ::LEU2 osh7Δ::KanMX4 his3Δ). BY4741-based SAC1::GAL1-6×HA-Sac1-Nat (referred to here as GAL1::SAC1) and OSH7::GAL1-6×HA-Osh7-Nat (OSH7-6×HA) yeast strains expressing 6×HA-tagged Sac1p or 6×HA-tagged Osh7p, respectively, under the GAL1 promoter from their natural chromosomal locations, were created by using plasmid pYM-N32 from the EUROSCARF plasmid collection according to the authors’ recommendation (101). The scs2Δ scs22Δ double-mutant (BY4741 SCS2::kanMX SCS22::hphNT1) strain was created from BY4741 scs2Δ (BY4741 SCS2::kanMX) (Open Bioystems) by using plasmid pFA6a-hphNT from the EUROSCARF plasmid collection (101).

The following yeast expression plasmids have been previously described: pGAD-LEU-Cup1-His92 (102), pESC-URA-Cup1-His92 (31), pGBK-HIS-Cup1-His33/ADH1-DI72 (65), pGBK-HIS-Gal1-RFP33/Gal1-DI72 (26), pDest-565-DrrA (35), and RFP-MS2-CP and DI72-MS2CP (66).

Plant expression plasmids.

Plant expression construct pGD-hFAPP1-PH was created by cloning hFAPP1-PH fragments PCR amplified from pRS406-PHO5-GFP-hFAPP1-PH (Addgene) with the primer pairs GFP-FAPP1-PH/XbaI/F (TGCATCTAGAATTCGAATATAATGTCTAAAGGAGAAGAACTTTTCACTGG) and GFP-FAPP1-PH/PstI/R (TGCACTGCAGTCATGTATCAGTCAAACATGCTTTGGAGC) into pGD-2x35Spro-L using XbaI and PstI sites. The full-length sequence of AtSAC1 was PCR amplified with primer pairs 7644 (ACGCGGATCCATGGCGAAATCGGAAAACTCAACG) and 7645 (ACGCGTCGACTTACTGCAGAATGACCTTCGGGACCTTATCATCTC) using A. thaliana cDNA as the template. The PCR product was digested with BamHI and SalI and then ligated into pGD-35S-GFP, pGD-35S-RFP, and pGD-35S-nYFP (the N-terminal domain of yellow fluorescent protein [YFP]) vectors.

The following plant expression plasmids have been previously described: pGD-p33-RFP, pGD-p33-GFP, pGD-p33-BFP, and pGD-RFP-SKL (26, 30, 67) and pEarleyGate100-DrrA (35).

Wild-type and RFP-H2b transgenic N. benthamiana plants were potted in soil and placed in a growth room at 25°C under a 16-h light/8-h dark cycle.

Analysis of tombusvirus replication in yeast.

Tombusvirus replication was analyzed in BY4741 and BY4741 sac1Δ yeast strains. These yeast strains were transformed with LpGADCUP1-p92 and HpGBKCUP1-p33/ADH1-DI-72 to express the p33 and p92 viral proteins, along with the DI-72 repRNA. The yeasts were grown in synthetic complete medium lacking leucine and histidine (SC-LH⁻ medium) supplemented with 2% glucose containing 50 μM CuSO₄ to induce viral protein expression. Yeasts were harvested after 16 and 48 h, and total RNA was extracted. Northern blot analysis was performed with a 3′-end-specific repRNA probe to detect the repRNA accumulation level, and an rRNA-specific probe was used to observe the rRNA level as described previously (61). Viral protein p33 and p92 expression was detected by Western blotting using anti-His antibody (103). In the complementation assay, yeast strains were transformed with plasmids LpGADCUP1-p92 and HpGBKCUP1-p33/ADH1-DI-72, and UpYC-GAL1-HisSac1. Yeasts were grown in synthetic complete medium lacking uracil, leucine, and histidine (SC-ULH⁻ medium) containing 2% glucose and 100 μM bathocuproine disulfonate (BCS) overnight. BCS inhibits the expression of the viral proteins driven by the CUP1 promoter. Then, the yeasts were grown in SC-ULH⁻ medium containing 2% raffinose or galactose in the presence of BCS for 12 h. The BCS was washed out, and fresh SC-ULH⁻ medium supplemented with 2% raffinose or galactose containing 50 μM CuSO₄ was added to induce viral protein expression for 24 h. Northern blotting and Western blotting were done as described above.

The osh5Δ osh6Δ osh7Δ triple-mutant SEY6210 and WT SEY6210 yeast strains were transformed to express HA-tagged Sac1p under the GAL1 promoter from the chromosome using plasmid pYM-N32 from the EUROSCARF plasmid collection (101). Then, both strains were transformed with plasmids LpGADCUP1-p92 and HpGBKCUP1-p33/ADH1-DI-72. Cells were grown in SC-LH⁻ medium supplemented with 2% glucose in the presence of 100 μM BCS overnight. Then, the cells were grown in SC-LH⁻ medium supplemented with either 2% raffinose or 2% galactose in the presence of BCS for 12 h. Finally, the BCS was washed out, and fresh SC-LH⁻ medium supplemented with 2% raffinose or 2% galactose containing 50 μM CuSO₄ was provided for 24 h. Northern blotting and Western blotting were done as described above.

To test the effect of yeast PI(4)P kinases on TBSV replication, we transformed BY4741 and lsb6Δ yeasts with pGAD-LEU-Cup1-His92, pGBK-HIS-Cup1-His33/ADH1-DI-72, and pESC-URA-Empty plasmids. In the cases of TET::STT4 and TET::PIK1 yeast strains, the transformation was done with plasmids pGAD-LEU-Cup1-His92 and pGBK-HIS-Cup1-His33/ADH1-DI-72. The transformed BY4741, lsb6Δ, TET::STT4, and TET::PIK1 yeasts were cultured in SC-ULH⁻ medium supplemented with 2% glucose and 100 μM BCS at 29°C for 16 h, and then the yeasts were continuously cultured in SC-ULH⁻ medium supplemented with 2% galactose and 50 μM CuSO₄ at 29°C for 16 h. To deplete the targeted proteins in TET::STT4 and TET::PIK1 yeast strains, 10 μg/ml doxycycline was added to both the pregrowth medium and the incubation medium. After harvest, RNA and protein were analyzed as described above.

Analysis of TBSV replication in N. benthamiana protoplasts.

Isolation of protoplasts from N. benthamiana leaves were carried out as described previously (26). The isolated protoplasts were transfected with TBSV genomic RNA isolated from infected plants using polyethylene glycol (PEG) (104) and treated with 50 μM PAO, a PI4K inhibitor (72, 73). Sixteen hours after transfection, the protoplasts were harvested by centrifugation and analyzed for TBSV levels by Northern blotting analysis.

Confocal laser microscopy analysis of yeast cells.

To analyze the subcellular localization of PI(4)P, plasmid pRS406-PHO5-GFP-hFAPP1-PH, along with pESC-TRP-Cup1-His92, pGBK-HIS-Gal1-RFP33/ADH1-DI72, and pAG415-DrrA or pAG415-Empty, was transformed into yeast strain SEY6210. The transformed yeasts were grown under the conditions used in TBSV RNA analysis as described above. The confocal images were obtained sequentially with an Olympus FV1000 microscope (Olympus America) and merged using Olympus FLUOVIEW 1.5 software (22).

Colocalization and BiFC assay in yeast.

WT BY4741 and BY4741 sac1Δ strains were transformed with UpRS426-GFP-Sac1 (provided by Daniel Barajas), along with HpESC-GAL1-BFP-p33 (26). The transformed yeasts were grown in UH⁻ minimal medium supplemented with 2% glucose for 12 h, and then yeast cells were harvested and washed and the medium was changed to SD-UH⁻ minimal medium supplemented with 2% galactose. GFP-Sac1 expression was driven by its endogenous promoter present on the pRS426 plasmid. Cells were harvested and washed after 24 h. Then, 2 to 3 μl of cell suspension was dropped on poly-l-lysine-coated slides and visualized with a laser scanning microscope (39). Single frames were obtained.

A BiFC assay was carried out as described previously (38, 39). Briefly, BY4741 and sac1Δ strains were transformed with HpESC-nYFP-p33-DI72 and UpYC-cYFP-SCS2, along with the LpGAD-ADH1-CFP-Sac1, LpRS315-TEF1-RFP-Vps21, or LpGAD-ADH1-RFP-Ufe1 plasmid (26, 38). Interaction between TBSV nYFP-p33 replication protein and the Scs2-cYFP (the C-terminal domain of YFP) protein and colocalization with cyan fluorescent protein (CFP)-Sac1p was detected by growing transformed yeast colonies in liquid yeast minimal synthetic drop-out (SD) medium containing 2% galactose for 24 h at 29°C, followed by confocal laser microscopy.

Detection of PE distribution in yeasts.

The distribution of PE was studied in BY4741 WT and BY4741 sac1Δ yeast strains as described previously (30). GFP-p33 protein was expressed form the pRS425-pCUP1-GFP-p33 plasmid. Yeast cells were grown in SC-L⁻ medium supplemented with 2% glucose in the presence of 100 μM BCS overnight. Then, the BCS was washed out and fresh SC-L⁻ medium supplemented with 2% glucose containing 50 μM CuSO₄ was provided to induce GFP-p33 expression. After 16 h, yeast cells were harvested, and PE distribution was detected by using biotinylated duramycin peptide and streptavidin conjugated with Alexa Fluor 405 (30).

Ergosterol distribution in yeasts.

To examine ergosterol distribution, the BY4741 yeast strain was cotransformed with plasmids HpESC-GAL1-RFP-p33 and UpRS426-GFP-Sac1. The control strain was transformed with empty pESC and UpRS426-GFP-Sac1. The yeast strains were grown in SC minimal medium supplemented with 2% galactose for 24 h at 23°C. The yeast cultures were fixed with 3% formaldehyde for 1 h at room temperature. The formaldehyde was quenched by addition of glycine (0.1 M final concentration). The fixed cells were centrifuged and washed twice with distilled water. The washed cells were incubated with 5 mg/ml filipin complex (Sigma Chemicals) in the dark for 15 min at 23°C. Filipin-based fluorescence was observed by spotting 2 to 3 μl of the cell suspensions onto poly-l-lysine-coated microscope slides under a UV light microscope using a DAPI (4′,6-diamidino-2-phenylindole) filter (39).

Analysis of protein interactions in vivo using a yeast membrane two-hybrid assay.

The MYTH assay was done as previously described (39). Briefly, to determine the interaction between S. cerevisiae Sac1p (ScSac1p) and p33 replication protein, plasmid pGAD-BT2-N-His33 was transformed with one of the following plasmids: pPR-N-RE, pPR-N-RE-ssa1, pPR-N-RE-ScSac1, pPR-N-RE-ScOsh5, pPR-N-RE-ScOsh6, or pPR-N-RE-ScOsh7. Similar amounts of transformed yeast colonies were picked up and suspended in 100 μl water, 8 μl of which was cultured on TL⁻ plates (SD agar plates lacking threonine and leucine) as a loading control or on TLHA⁻ plates (SD agar plates lacking threonine, leucine, histidine, and adenine) to score protein interactions.

Virus-induced gene silencing of SAC1 genes in N. benthamiana.

Three SAC1-like genes were identified in A. thaliana: AtSAC1a (At3g51830), AtSAC1b (At5g66020), and AtSAC1c (At3g51460), which can rescue the cold-sensitive and inositol auxotroph yeast sac1-null mutant strain. These proteins are localized in the ER (59). The translation product of AtSAC1a showed 60% similarity to the S. cerevisiae Sac1p (59). To find N. benthamiana-specific SAC1 sequences, we performed a BLAST search in the TGI database (Gene Indices, Harvard University) using the sequences of AtSAC1a, AtSAC1b, and AtSAC1c against N. benthamiana TC sequences. The BLAST search of SAC1b and SAC1c resulted in the same hit, TC141032, with 72% identity with the Arabidopsis sequences. SAC1a resulted in a separate hit, TC148342. We designed primers based on the N. benthamiana sequences to PCR amplify a 474-bp (SAC1a) and a 398-bp (SAC1b) fragment to create the VIGS vectors, pTRV2-SAC1ANb and pTRV2-SAC1BNb. The primer pairs for PCR were as follows: 5960F (CGCCGGATCCGGGTGGATCAAGGTGATGAA)/5961R (CGCCTCGAGTGCCTTCCATTTGCTTTGA) and 5962F (CGCCGGATCCGGTGGATGGGTCAATGAAAC)/5963R (CGCCTCGAGGAGGGTCTGCTTGTCTCCAC), respectively. VIGS in N. benthamiana was performed as described previously (105). Briefly, Agrobacterium cultures were transformed with plasmid pTRV2-SAC1ANb, pTRV2-SAC1BNb, or pTRV2-cGFP as a control. N. benthamiana plants were agroinfitrated (optical density at 600 nm [OD₆₀₀], 0.5). On the 8th day of agroinfiltration, leaves were sap inoculated with TBSV. Then, total RNA was extracted on the 2nd day postinoculation from the infected leaves, followed by Northern blotting analysis. Silencing of the target genes was confirmed with reverse transcription (RT)-PCR using primers 5960F (CGCCGGATCCGGGTGGATCAAGGTGATGAA) and 6342 (GGCGTAGTTTCGTCCTCCTA) on total RNA extracts of pTRV2-SAC1ANb- and pTRV2-cGFP-agroinfiltrated plants. RT-PCR amplification of tubulin mRNA sequence was performed as a control (39).

Confocal microscopic analysis of plant cells.

To analyze the subcellular localization of AtSac1 in the presence or absence of TBSV components in N. benthamiana leaves, we introduced plasmids pGD-35S-RFP-AtSac1, pGD-35S-T33-BFP, and pGD-35S-GFP-SKL (as a peroxisome marker) via agroinfiltration. The effect of CIRV on AtSac1 distribution in N. benthamiana leaves was studied by agroinfiltration of a mixture of Agrobacterium transformants (strain C58C1) carrying plasmids pGD-35S-RFP-AtSac1, pGD-35S-C36-BFP, and pGD-35S-GFP-AtTim21 (as a mitochondrial marker). Then, TBSV or CIRV inoculation with sap was performed at 16 h postagroinfiltration. At 2.5 days postagroinfiltration, the agroinfiltrated leaves were subjected to confocal laser microscopy.

To demonstrate if AtSac1 is recruited into an active virus replication compartment, a modified repRNA carrying six repeats of a hairpin RNA from the MS2 bacteriophage, which is specifically bound by its coat protein, was utilized (66). N. benthamiana leaves were coinfiltrated with different combinations of agrobacteria carrying pGD-p33-BFP (OD₆₀₀, 0.3), pGD-GFP-AtSac1 (OD₆₀₀, 0.3), pGD-RFP-MS2-CP (OD₆₀₀, 0.5), pGD-p19 (OD₆₀₀, 0.3), pGD-(+)DI72-MS2hp (OD₆₀₀, 0.5), and pGD-CNV-20K (OD₆₀₀, 0.4). Then, the samples were subjected to confocal microscopic analysis at 3.5 days postagroinfiltration.

To detect interaction between AtSac1 and TBSV p33 or CIRV p36 replication protein, we performed BiFC assays. Plasmids pGD-T33-cYFP, pGD-C36-cYFP, pGD-C-cYFP (as a negative control), pGD-nYFP-AtSac1, pGD-nYFP-MBP (as a negative control), pGD-RFP-SKL (as a peroxisome marker), and pGD-RFP-AtTim21 (as a mitochondrial marker) were separately transformed into the Agrobacterium strain C58C1. The Agrobacterium transformants with different combinations were used to infiltrate N. benthamiana leaves, which were harvested and then subjected to confocal microscopic analysis at 2 days after agroinfiltration.

To observe the colocalization of p33 and PI(4)P in the presence or absence of the Legionella DrrA effector in N. benthamiana epidermal cells, pGD-p33-RFP, pGD-p19 pEarlygate-DrrA, and pGD-GFP-hFAPP1-PH were agroinfiltrated into N. benthamiana leaves, and the infiltrated leaves were inoculated with TBSV 1 day after the agroinfitration. Images were taken 2 days after infiltration. To visualize PI(4)P distribution in plant cells, mesophyll protoplasts were isolated from N. benthamiana leaves 1 day after agroinfiltration with pGD-p19, pGD-p33-GFP, or pGD-p33-BFP, and pEarlygate-DrrA as described above. The protoplasts were transfected with TBSV genomic RNA using PEG (104). Sixteen hours after transfection, protoplasts were harvested, fixed with 2% paraformaldehyde, and applied to a poly-l-lysine–coated slide. Using mouse anti-PI(4)P IgM (Echelon; Z-P004) and Alexa 568-conjugated anti-mouse secondary antibody (Invitrogen), PI(4)P distribution was analyzed with confocal microscopy (22).

Recombinant-protein purification from E. coli.

For the expression and purification of the recombinant MBP-tagged TBSV p33 and p92 replication proteins from E. coli, we followed a published protocol (106).

In vitro TBSV replication assay in cell-free yeast extract.

BY4741 and GAL1::Sac1 yeast strains were pregrown for 16 h at 29°C in yeast extract-peptone-glucose (YPG) medium supplemented with 2% raffinose. After the yeast cells were centrifuged at 3,000 rpm for 5 min and washed with yeast extract-peptone-dextrose (YPD) medium, the yeasts were grown for 8 h at 29°C in YPD medium. We followed a previously published protocol (63, 107) for the preparation of CFEs to reconstitute the TBSV replicase in vitro. The in vitro TBSV replication assays were performed in 20-μl total volumes containing 3 μl of CFE, 0.25 μg DI-72 (+)repRNA transcript, 200 ng purified MBP-p33, 200 ng purified MBP-p92^pol, and the CFE buffer. After incubation at 25°C for 3 h, the RNA products synthesized in the replication assays were separated by electrophoresis in 0.5× Tris-borate-EDTA buffer in a 5% polyacrylamide gel containing 8 M urea.

In the second TBSV replication assay, the membrane and soluble fractions of CFEs from BY4741 and sac1Δ yeasts were separated via centrifugation at 16,000 rpm. The BY4741 and sac1Δ strains were pregrown in YPD medium for 16 h, and then fresh YPD medium was supplied for an additional 6 h. After mixing the membrane and soluble fractions in different combinations, the in vitro TBSV replication assays were performed as described previously (63, 107).

Copurification of various host proteins and p33 from yeasts.

For the copurification of chromosomally expressed 6×HA-tagged Sac1p with tombusvirus Flag-p33 replication protein, the GAL1::SAC1 yeast strain was transformed with plasmids pGBK-His33-CUP1/DI72-GAL1 or pGBK-Flag33-CUP1/DI72-GAL1 and pGAD-CUP1-His92. For copurification of His₆-tagged Osh7 and His₆-Sac1 proteins with tombusvirus Flag-p33 replication protein, BY4741 and scs2Δ scs22Δ double-mutant yeast strains were cotransformed with plasmid pESC/DI72/His33 or pESC/DI72/FLAG33, pGAD-His92, and pYC2/NT-C plasmids expressing His₆-tagged Osh7p and His₆-Sac1p. For copurification of Flag-tagged Scs2p and His₆-tagged Osh5p, His₆-Osh6p, His₆-Osh7p, and His₆-Sac1p, BY4741 yeast was transformed with HpESC-Flag-Scs2, together with pYC2/NT-C plasmid expressing one of the His₆-tagged proteins. Control yeast samples were transformed with plasmid coding for YFP-tagged Scs2. Flag-based purification was carried out as described previously (39).

BY4741 and sac1Δ yeast strains were transformed with the following plasmids: pGBK-HIS-Cup-Flag33/Gal-DI-72 (108) expressing Flag-tagged p33 and the TBSV DI-72 repRNA; pGAD-LEU-Cup-Flag92 expressing Flag-tagged p92; and one of the UpESC-Cup-His plasmids expressing His₆-Bro1p, His₆-Ded1p, His₆-Cdc34p, His₆-Osh6p, His₆-Pbp2p, His₆-Rad6p, His₆-RH2, His₆-Rpn11p, His₆-Tdh2p, His₆-Tef1p, His₆-Ubc2p, His₆-Vap27-1, His₆-Vps4p, or His₆-Vps23p. The experiment was performed as described previously (38). Briefly, transformed yeasts were grown for 18 h in SC-ULH⁻ medium supplemented with 2% galactose and 100 μM BCS at 29°C. Then, the medium was changed to SC-ULH⁻ medium supplemented with 2% glucose and 100 μM BCS. After growing yeasts for 24 h at 29°C, the medium was changed to SC-ULH⁻ medium containing 2% glucose and 50 μM CuSO₄, and culturing continued for 10 h at 23°C. Then, Flag-tagged replicase was purified following a previously published protocol (82). Fractions eluted from anti-FLAG M2-agarose affinity resin were balanced based on the amount of purified Flag-p33. Both total and purified fractions were analyzed for the presence of His₆-tagged host proteins. Purified Flag-p33 was detected by Western blotting using anti-Flag antibody; copurified His₆-tagged host proteins were detected with anti-His antibody, followed by anti-mouse antibody conjugated to alkaline phosphatase. Colorimetric detection was performed with nitroblue tetrazolium (NBT) and 5-bromo-4-chloro-3-indolylphosphate (BCIP).