DEVELOPMENTALLY REGULATED PLASMA MEMBRANE PROTEIN of Nicotiana benthamiana Contributes to Potyvirus Movement and Transports to Plasmodesmata via the Early Secretory Pathway and the Actomyosin System1[OPEN]

Virus movement in tobacco depends on interactions between tobacco and viral movement proteins and on their traffic to plasmodesmata. The intercellular movement of plant viruses requires both viral and host proteins. Previous studies have demonstrated that the frame-shift protein P3N-PIPO (for the protein encoded by the open reading frame [ORF] containing 5′-terminus of P3 and a +2 frame-shift ORF called Pretty Interesting Potyviridae ORF and embedded in the P3) and CYLINDRICAL INCLUSION (CI) proteins were required for potyvirus cell-to-cell movement. Here, we provide genetic evidence showing that a Tobacco vein banding mosaic virus (TVBMV; genus Potyvirus) mutant carrying a truncated PIPO domain of 58 amino acid residues could move between cells and induce systemic infection in Nicotiana benthamiana plants; mutants carrying a PIPO domain of seven, 20, or 43 amino acid residues failed to move between cells and cause systemic infection in this host plant. Interestingly, the movement-defective mutants produced progeny that eliminated the previously introduced stop codons and thus restored their systemic movement ability. We also present evidence showing that a developmentally regulated plasma membrane protein of N. benthamiana (referred to as NbDREPP) interacted with both P3N-PIPO and CI of the movement-competent TVBMV. The knockdown of NbDREPP gene expression in N. benthamiana impeded the cell-to-cell movement of TVBMV. NbDREPP was shown to colocalize with TVBMV P3N-PIPO and CI at plasmodesmata (PD) and traffic to PD via the early secretory pathway and the actomyosin motility system. We also show that myosin XI-2 is specially required for transporting NbDREPP to PD. In conclusion, NbDREPP is a key host protein within the early secretory pathway and the actomyosin motility system that interacts with two movement proteins and influences virus movement.

The movement of viruses in plants can be divided into three stages: intracellular, intercellular, and long-distance movement (Nelson and Citovsky, 2005;Benitez-Alfonso et al., 2010). Plasmodesmata (PD) are plasma membranemediated channels in cell walls that control the intercellular trafficking of micromolecules and macromolecules, including plant viruses (Boevink and Oparka, 2005;Lucas et al., 2009). Plant viruses encode movement proteins (MPs) that can regulate the size exclusion limit (SEL) of PD and mediate virus trafficking between cells (Lucas, 2006;Raffaele et al., 2009;Amari et al., 2010;Ueki et al., 2010). Based on the functions of MPs during virus movement, the viral MPs are divided into three major groups. The first group of MPs is represented by the 30-kD protein of Tobacco mosaic virus (TMV). The 30-kD proteins can interact with single-stranded RNAs and transport viral ribonucleoprotein complexes to cell walls, where they modify the SEL of PD to allow viruses to traverse the cell walls (Olesinski et al., 1996;Tzfira et al., 2000;Kawakami et al., 2004). The second group of MPs is known to form tubular structures that extend across the PD and allow virus to traverse. Viruses that encode this group of MPs include Cowpea mosaic virus, Grapevine fan leaf virus (GFLV), Cauliflower mosaic virus, and Tomato spotted wilt virus (Ritzenthaler and Hofmann, 2007;Amari et al., 2011). The third group of MPs is known as triple gene block proteins (TGBps), encoded by overlapping triple gene blocks. The three TGBps (TGBp1, TGBp2, and TGBp3) function coordinately to transport viral genomes to and through PD (Verchot-Lubicz, 2005;Jackson et al., 2009;Lim et al., 2009;Tilsner et al., 2013). Viruses that encode TGBps belong to the genera Potexvirus, Hordeivirus, and Pomovirus (Verchot-Lubicz et al., 2010). Potyviruses are different from the above viruses and lack a dedicated MP. To date, multiple potyviral proteins, including COAT PROTEIN, CYLINDRICAL INCLUSION (CI), HELPER COMPONENT PROTEINASE (HC-Pro), and VIRAL GENOME-LINKED PROTEIN, have been shown to function in the cell-to-cell movement of potyviruses (Nicolas et al., 1997;Rojas et al., 1997;Carrington et al., 1998;Wei et al., 2010).
Viruses of Potyvirus (family Potyviridae), the largest genus of plant-infecting viruses, cause great economic losses to world agriculture production (Fauquet et al., 2005). The potyviral genome is a positive sense, single-stranded RNA of approximately 10 kb in length. It contains a large open reading frame (ORF) encoding a polyprotein that is later processed into 10 mature proteins by three virus-encoded proteinases (Riechmann et al., 1992;Fauquet et al., 2005). A +2 frame-shift Pretty Interesting Potyviridae (PIPO) ORF that is embedded within the P3 ORF was recently identified and proposed to produce a P3N-PIPO (for the protein encoded by 59-terminus of P3 and frameshift PIPO) fusion (Chung et al., 2008;Vijayapalani et al., 2012). The P3N-PIPOs of Turnip mosaic virus (TuMV) and Tobacco etch virus were previously shown to localize at PD, interact with CI in planta, and transport CI to PD in a CI:P3N-PIPO ratio-dependent manner (Wei et al., 2010). Soybean mosaic virus with a mutant PIPO domain failed to cause systemic infection in its host plant (Wen and Hajimorad, 2010). Therefore, the potyvirus P3N-PIPO has been suggested as the classical MP (Tilsner and Oparka, 2012;Vijayapalani et al., 2012).
Viruses recruit host factors for their movement in plants Raffaele et al., 2009;Amari et al., 2010;Ueki et al., 2010). Compared with the progresses on viral MP characterization, identifications of MP-interacting host proteins are much behind Oparka, 2004;Raffaele et al., 2009;Amari et al., 2010). To date, about 20 host proteins have been identified to interact with specific viral MPs (Pallas and García, 2011). For example, the pectin methylesterase interacted with TMV MP, increased the SEL of PD, and facilitated TMV movement between cells ; an ankyrin repeatcontaining protein (ANK) interacted with TMV MP at PD, down-regulated callose formation, and aided viral movement (Ueki et al., 2010); the Arabidopsis (Arabidopsis thaliana) PLASMODESMATA-LOCALIZED PRO-TEIN1 (AtPDLP1) was reported to interact with GFLV MP and mediate tubule assembly during GFLV cell-tocell movement in plants (Amari et al., 2010(Amari et al., , 2011. TuMV P3N-PIPO was shown to interact with AtPCaP1, a plasma membrane cation-binding protein of Arabidopsis, and colocalize with this host protein at the PD. Knockout of AtPCaP1 expression resulted in a significant reduction of TuMV infection in Arabidopsis (Vijayapalani et al., 2012). Many viral MPs have been shown to traffic within plant cells via the early secretory pathway and/or along the actin filaments or microtubules. For example, the early secretory pathway and microtubules were required for GFLV MP trafficking to PD (Laporte et al., 2003). TuMV P3N-PIPO and CI were reported to utilize the early secretory pathway rather than the actomyosin motility system for their trafficking to PD (Wei et al., 2010). Several plant myosin motor proteins have been reported to participate in virus intracellular movement (Wei and Wang, 2008;Harries et al., 2010). Myosins VIII-1, VIII-2, and VIII-B were shown to transport a HEAT SHOCK PROTEIN70 homolog of Beet yellows virus to PD (Avisar et al., 2008a), but only myosin VIII-1 was needed for the nonstructural protein encoded by viral complementary strand of RNA4 (NSvc4) of Rice stripe virus traffic to PD (Yuan et al., 2011). A more recent study has indicated that both the secretory pathway and myosins XI-2 and XI-K were required for TuMV cell-to-cell movement (Agbeci et al., 2013). However, it remains largely unknown how the MP-interacting host factor(s) reach their target sites in cells.
Tobacco vein banding mosaic virus (TVBMV) is a distinct potyvirus mainly infecting solanaceous crops (Tian et al., 2007;Yu et al., 2007;Zhang et al., 2011). In this article, we provide evidence showing the length requirements of the PIPO domains for its function in mediating TVBMV movement and the restoration of the movement-defective TVBMV mutants. We also show the interactions between TVBMV P3N-PIPO and CI and NbDREPP, a developmentally regulated plasma membrane protein in Nicotiana benthamiana, and the route by which NbDREPP traffics to PD. Silencing of NbDREPP expression in N. benthamiana significantly impeded the cell-to-cell movement of TVBMV. Figure 1. Schematic diagram of pCamTVBMV-GFP-HISP3N and western-blot analysis of His-tagged P3 and P3N-PIPO. A, Genome diagram of pCamTVBMV-GFP-HISP3N. Amino acids at the cleavage sites between HC-Pro and P3 and between P3 and 6K1 are shown. The inserted six His residues are shown in boldface. Arrowheads indicate the positions of the three introns inserted in the TVBMV infectious cDNA clone. UTR, Untranslated region. B, Western-blot analysis of P3 and P3N-PIPO in the N. benthamiana (Nb) and N. tabacum (Nt) leaves infected with pCamTVBMV-GFP or pCamTVBMV-GFP-HISP3N. Total proteins were extracted from the systemically infected leaves at 14 d post agroinfiltration (dpai), separated on a 10% SDS-PAGE gel, blotted onto a nitrocellulose membrane, and then probed with an anti-His IgG. The predicted positions of His-tagged P3 and P3N-PIPO are shown. Molecular masses of the protein markers are also shown.

Expression of TVBMV P3N-PIPO Fusion in Plants
To determine the form of PIPO expressed in the TVBMV-infected plants, we modified pTVBMV-GFP, an infectious clone of the TVBMV HN39 isolate that carries a GFP gene and can express free GFP (Gao et al., 2012), to produce constructs pCamTVBMV-GFP, which can be inoculated via agroinfiltration, and pCamTVBMV-GFP-HISP3N, which expresses a His tag-labeled P3 (Fig.  1A). Both constructs were agroinfiltrated into leaves of N. benthamiana and Nicotiana tabacum. By 14 dpai, two unique proteins of approximately 40 and 24 kD, similar to the predicted size of TVBMV (His) 6 -P3 and (His) 6 -P3N-PIPO, were detected by western-blot assays using a His tag-specific antibody in TVBMV-GFP-HISP3N-infected but not TVBMV-GFP-infected N. benthamiana and N. tabacum plants (Fig. 1B). This result indicated that both P3 and the P3N-PIPO fusion were produced in the TVBMV-infected plants.

The Length of the PIPO Domain Determines the Ability of TVBMV Mutants to Move between Cells
The P3N-PIPO ORF of wild-type TVBMV encoded a PIPO domain of 60 amino acid residues (Zhang et al., 2011;Gao et al., 2012). To determine the length requirements of the PIPO domain during TVBMV infection, we relocated the stop codon within the P3N-PIPO ORF to produce four TVBMV mutants with truncated PIPO domains while maintaining the original amino acid sequence of P3. The resulting mutants PIPO7aaSTOP, PIPO20aaSTOP, PIPO43aaSTOP, and PIPO58aaSTOP encoded PIPO domains of 7, 20, 43, and 58 amino acid residues, respectively ( Fig. 2A). Because the PIPO domains in different potyviruses varied in length (Supplemental Fig. S1), we also produced three more TVBMV mutants with extended PIPO domains (PIPOPLUS8aa, PIPOPLUS22aa, and PIPOPLUS38aa). These three mutants encoded PIPO domains of 68, 82, and 98 amino acid residues, respectively (Supplemental Fig. S2A). The wild-type and mutant TVBMV were individually inoculated to N. benthamiana plants via agroinfiltration. By 7 dpai, plants agroinfiltrated with construct pCamTVBMV-GFP, pCamPIPO58aaSTOP, pCamPIPOPLUS8aa, pCamPIPOPLUS22aa, or pCamPIPOPLUS38aa developed mosaic symptoms in their upper young leaves (i.e. systemic leaves). Under UV illumination, these infected plants showed strong GFP fluorescence in their agroinfiltrated and systemically infected leaves ( Fig. 2B; Supplemental Fig. S2B). Plants agroinfiltrated with pCamPIPO7aaSTOP, pCamPIPO20aaSTOP, or pCamPIPO43aaSTOP failed to show any virus symptoms in their systemic leaves, and GFP fluorescence was observed only in the agroinfiltrated leaves (Fig. 2B). The reverse transcription (RT)-PCR results showed that TVBMV RNA had accumulated in the systemic leaves of pCamTVBMV-GFP-, pCamPIPO58aaSTOP-, pCamPIPOPLUS8aa-, pCamPIPOPLUS22aa-, or pCamPIPOPLUS38aa-agroinfiltrated plants but was not detected in the systemic leaves of plants agroinfiltrated with pCamPIPO7aaSTOP, pCamPIPO20aaSTOP, or pCamPIPO43aaSTOP by 7 dpai (Supplemental Fig. S3).
To determine whether the four truncated mutants could replicate in the agroinfiltrated N. benthamiana leaves as well as the parental TVBMV-GFP, we analyzed the accumulation levels of their negative-strand RNA by a northern-blot assay using an RNA probe specific for the TVBMV COAT PROTEIN ORF. The results showed that similar amounts of negative-strand RNA had accumulated in the leaves agroinfiltrated with pCamTVBMV-GFP or one of the four mutant constructs by 3 dpai ( Fig. 2C; Supplemental Fig. S4), indicating that the replication of these mutants was not affected by the mutations introduced into the PIPO ORF.
To further investigate the role of P3N-PIPO in TVBMV cell-to-cell movement, Agrobacterium tumefaciens cells harboring pCamTVBMV-GFP, pCamPIPO7aa-STOP, pCamPIPO20aaSTOP, pCamPIPO43aaTOP, pCamPIPO58aaSTOP, pCamPIPOPLUS8aa, pCamPI-POPLUS22aa, or pCamPIPOPLUS38aa were used to infiltrate N. benthamiana leaves. Prior to infiltration, the A. tumefaciens cells were resuspended and adjusted to optical density at 600 nm (OD 600 ) = 0.5 with induction buffer, which were further diluted at a ratio of 1:10,000 to ensure that initial transfection occurred in isolated foci of a single cell to allow cell-to-cell movement assessment. By 3 dpai, the leaves agroinfiltrated with pCamPIPO7aa-STOP, pCamPIPO20aaSTOP, or pCamPIPO43aaTOP showed GFP green fluorescence in single epidermal cells, while the leaves agroinfiltrated with pCamTVBMV-GFP, pCamPIPO58aaSTOP, pCamPIPOPLUS8aa, pCamPIPO-PLUS22aa, or pCamPIPOPLUS38aa showed GFP fluorescence in clusters of approximately eight epidermal cells (Fig. 2,D and E;Supplemental Fig. S2C). This result indicated that TVBMV mutants encoding a PIPO domain of 58 or more amino acid residues were competent in cell-tocell movement while those encoding a PIPO domain of less than 58 amino acid residues were defective in cell-tocell movement.

Movement-Defective Mutants Restored Their Ability to Cause Systemic Infection via Mutations
The above infectivity assays showed that mutant PIPO7aaSTOP, PIPO20aaSTOP, or PIPO43aaSTOP failed to cause systemic infection in N. benthamiana by 7 dpai ( Fig. 2B). However, by 14 dpai, some N. benthamiana plants agroinfiltrated with pCamPIPO7aaSTOP, pCamPIPO20aaSTOP, or pCamPIPO43aaSTOP developed systemic mosaic symptoms. To investigate the reason for this phenomenon, total RNAs were extracted from the systemically infected leaves for RT-PCR followed by sequencing of the progeny viruses. The results showed that the previously introduced stop codons (TGA, TAG, and TGA) in these mutants had mutated to CGA, TGG, or CGA, resulting in the production of the functional P3N-PIPO. Some progeny viruses were found to contain additional mutations in the other regions within the genomic RNA. Most of these mutations were synonymous mutations and a few were nonsynonymous. To determine the role of these nonsynonymous mutations in movement, we introduced them into constructs pCamPIPO7aaSTOP, pCamPIPO20aaSTOP, and pCamPIPO43aaSTOP, respectively. Agroinfiltration of these new mutants to N. benthamiana plants did not produce systemic infection in any assayed plant (data not shown).
To determine the frequency of spontaneous mutations at the introduced stop codon sites in the movementdefective mutants, N. benthamiana and N. tabacum plants were agroinfiltrated with individual movement-defective mutants and grown at 25°C or 30°C. By 14 dpai, systemic symptoms and GFP green fluorescence were observed in 65% to 77% of N. benthamiana and 17% to 35% of N. tabacum plants grown at 25°C. Only 8% to 9% of N. benthamiana and 2% to 10% of N. tabacum plants grown at 30°C showed systemic virus symptoms and GFP green fluorescence (Table I). This result indicated that the frequency of spontaneous mutations was host and temperature dependent. Of the three mutants, the highest mutation frequency was found in PIPO43aaSTOPinfiltrated host plants grown under both temperature conditions. The lowest mutation frequency was found in PIPO20aaSTOP-infiltrated N. tabacum plants (Table I).

TVBMV P3N-PIPO and CI Interact with DREPP of N. benthamiana
Candidates that interacted with TVBMV P3N-PIPO were identified from an N. benthamiana complementary DNA (cDNA) library using a yeast (Saccharomyces cerevisiae)-two hybrid assay. Sequence analysis showed that five of the candidates shared 87% amino acid sequence identity with NtDREPP, a developmentally regulated plasma membrane protein of N. tabacum (Supplemental Fig. S5A). NbDREPP was predicted to have a long intrinsically disordered (ID) region in its C terminus (Supplemental Fig. S5, B and C). The ID region contained four flexible, surface-exposed, and disorder-promoting residues (Ser, Pro, Glu, and Lys; Marín and Ott, 2014) that accounted for about 43% of the total amino acid in the NbDREPP. The full-length NbDREPP ORF was then cloned into the pGBKT7 vector, and the wild-type P3N-PIPO ORF or its mutants were individually cloned into the pGADT7 vector. Plasmids pGADT7-T-antigen and pGBKT7-laminin were used as a negative control treatment and pGADT7-T-antigen and pGBKT7-murine p53 as a positive control treatment during the assay. Like the yeast cells cotransformed with the positive control plasmids, the yeast cells cotransformed with pGBKT7-NbDREPP and pGADT7-P3N-PIPO or pGBKT7-NbDREPP and pGADT7-P3N-PIPO58aa gave blue colonies on the culture medium containing synthetic dextrose (SD)/2Trp/2Leu/+5-bromo-4-chloro-3-indolyla-D-galactopyranoside (X-a-Gal) or SD/2Trp/2Leu/ 2adenine/2His/+X-a-Gal (Fig. 3A). The yeast cells cotransformed with pGBKT7-NbDREPP and pGADT7-P3N-PIPO68aa or pGADT7-P3N-PIPO82aa and pGADT7-P3N-PIPO98aa also gave positive results (data not shown). These results demonstrated that both the wild-type and movement-competent TVBMV P3N-PIPO interacted with NbDREPP.
To further confirm this interaction, we fused NbDREPP or TVBMV P3N-PIPO to the C or N terminus of yellow fluorescent protein (YFP), respectively. The plasmids were then agroinfiltrated into N. benthamiana leaves. By 72 hpai, YFP yellow fluorescence was observed in NbDREPP-YC (for YFP C-terminus) and P3N-PIPO-YN (for YFP N-terminus)-or P3N-PIPO58aa-YN-coinfiltrated N. benthamiana cells using a confocal microscope (Fig. 3B). The N. benthamiana cells coinfiltrated with NbDREPP-YC and P3N-PIPO68aa-YN, P3N-PIPO82aa-YN, or P3N-PIPO98aa-YN also showed YFP yellow fluorescence  (data not shown). These results indicated that NbDREPP did interact with the wild-type or mutant P3N-PIPOs that support successful TVBMV cell-to-cell movement.
In this study, we also tested the interaction between TVBMV CI and NbDREPP via bimolecular fluorescence complementation (BiFC). Our results showed that by 72 hpai, YFP yellow fluorescence was observed in the leaf cells coexpressing NbDREPP-YC and CI-YN, NbDREPP-YN and CI-YC, or CI-YN and CI-YC (Fig. 3C). No YFP yellow fluorescence was observed in the leaf cells coexpressing the negative controls plasmids, NbDREPP-YC and YN, or CI-YN and YC (Fig. 3C). Therefore, the BiFC results indicated that TVBMV CI could interact with NbDREPP and CI itself.
The interaction between NbDREPP and CI was further determined through a coimmunoprecipitation assay. NbDREPP-c-myc and CI-HA (for HEMAGGLUTININ) were transiently coexpressed in N. benthamiana leaves via agroinfiltration followed by total protein extraction. Leaves coagroinfiltrated with pCamNbDREPP-c-myc and pCamHA or pCamCI-HA and pCamc-myc were used as negative controls. The extracted proteins were analyzed through immunoprecipitation followed by western-blot assays using antibodies specific for HA and c-myc. The results showed that NbDREPP was indeed coimmunoprecipitated with TVBMV CI (Fig. 3D).

NbDREPP Contributes to TVBMV Intercellular Movement
To determine the role of NbDREPP in TVBMV movement, we cloned a 564-bp NbDREPP fragment into a Tobacco rattle virus (TRV)-based vector and silenced the expression of NbDREPP in N. benthamiana through virusinduced gene silencing (VIGS; Fig. 4). An 819-bp fragment of NbANK, whose silencing reduced TMV movement (Ueki et al., 2010), was cloned into the TRV vector to silence NbANK expression. The empty TRV vector-infected (nonsilenced control), NbDREPP-silenced, and NbANKsilenced N. benthamiana plants were agroinfiltrated again with pCamTVBMV-GFP. Five days later, vein-clearing symptoms appeared in the systemic leaves of the TRV empty vector-infected and NbANK-silenced N. benthamiana plants. Under UV illumination, GFP green fluorescence was observed in the systemic leaves of these plants (Fig.  4A, left column). In contrast, GFP green fluorescence was observed in the infiltrated leaves but not the systemic leaves of NbDREPP-silenced plants by 5 dpai (Fig. 4A, left column). By 14 dpai with pCamTVBMV-GFP, the NbDREPP-silenced N. benthamiana plants also showed systemic mosaic and slight stunting symptoms. Under UV illumination, GFP green fluorescence, although weaker than that in the TRV empty vector-infected and NbANK-silenced plants, was observed in the systemic leaves of the NbDREPP-silenced plants (Fig. 4A, right column). Results of semiquantitative reverse transcription (RT)-PCR confirmed that the expression levels of NbDREPP in NbDREPP-silenced plants and NbANK in NbANK-silenced N. benthamiana plants were clearly knocked down through VIGS (Fig. 4B).
The results of fluorescence microscopy showed that by 2 dpai, TVBMV-GFP infection in the NbDREPPsilenced N. benthamiana leaves was predominantly restricted in single cells. However, its infection in the NbANK-silenced or TRV empty vector-infected N. benthamiana leaves expanded to clusters of about five cells (Fig. 4, C and D). By 5 dpai, TVBMV-GFP in the NbDREPP-silenced leaves expanded to clusters of about 20 cells, while its infection in the NbANK-silenced or TRV empty vector-infected N. benthamiana leaves had expanded to large foci with numerous cells (Fig. 4, C and D). These results clearly indicated that NbDREPP played an important role in TVBMV cell-to-cell movement in N. benthamiana.

NbDREPP Colocalizes with P3N-PIPO and CI at PD
To determine the subcellular localization pattern of NbDREPP in planta, we fused the NbDREPP gene to GFP, YFP, and DsRed to generate constructs pNbDREPP-GFP, pNbDREPP-YFP, and pNbDREPP-DsRed, respectively. After agroinfiltration of these constructs into N. benthamiana leaves, punctate fluorescent bodies were observed near the cell walls in all infiltrated leaves by 48 hpai (Fig. 5A). Under the higher magnifications, the fluorescent signal from the expressed NbDREPP-GFP, NbDREPP-YFP, or NbDREPP-DsRed fusion appeared as paired punctate bodies at both sides of the cell walls (Fig. 5A). When AtPDLP1-DsRed, a PD marker, was coexpressed with NbDREPP-GFP in the same cells, colocalized red and green fluorescence was observed at PD (Fig. 5B). In addition, NbDREPP was unable to interact with AtPDLP1 in a BiFC assay (Supplemental Fig.  S6). These results indicated that NbDREPP localized at PD by itself.
When NbDREPP-YC and P3N-PIPO-YN were coexpressed with AtPDLP1-DsRed in N. benthamiana leaf epidermal cells via agroinfiltration, YFP yellow fluorescence from the NbDREPP-YC/P3NPIPO-YN complex overlapped with the red fluorescence from AtPDLP1-DsRed (Fig. 5C), indicating that NbDREPP and P3N-PIPO interacted at PD. TVBMV CI by itself localized in the cytoplasm (Fig. 5D). In the presence of P3N-PIPO or during TVBMV infection, however, CI was found at PD and formed paired punctate bodies at the cell   (Fig. 5E). This finding suggested that TVBMV P3N-PIPO could interact with CI and transport CI to PD. In addition, in the presence of P3N-PIPO or during the infection of TVBMV, NbDREPP and CI accumulated predominantly as punctate bodies at the cell walls (Fig. 5, F and G).

NbDREPP Traffics to PD via the Early Secretory Pathway
To investigate the role of the early secretory pathway in NbDREPP intracellular trafficking, we conducted chemical treatment and dominant-negative inhibition assays. Brefeldin A (BFA) was shown to interfere with the intracellular transport of proteins, membrane materials, and soluble cargo between the endoplasmic reticulum (ER) and the Golgi apparatus within the endomembrane system (Nebenführ et al., 2002;Stefano et al., 2006;Robinson et al., 2007;Cheung and de Vries, 2008). Leaves of N. benthamiana were treated with BFA followed by confocal microscopy. The results showed that by 3 h post BFA treatment, the PD targeting of NbDREPP-GFP was disrupted, but treatment with dimethyl sulfoxide (DMSO) showed no such effect (Fig. 6, A and B).
The Actin Cytoskeleton and Myosin XI-2 Are Essential for Trafficking of NbDREPP to PD Chemical and protein inhibition assays were used to evaluate the role of the actomyosin system in the intracellular movement of NbDREPP. Treatment of N. benthamiana leaves with latrunculin B (LatB), a chemical inhibitor of actin polymerization (Morton et al., 2000), disrupted the accumulation of NbDREPP-GFP at PD in approximately 80% of the N. benthamiana cells examined (Fig. 7, A and C), indicating that the actin cytoskeleton was required for NbDREPP targeting to PD. We then transiently overexpressed myosin tails to disrupt the function of endogenous myosins (dominant-negative inhibition; Avisar et al., 2008aAvisar et al., , 2008bPeremyslov et al., 2008;Amari et al., 2011) and analyzed whether the trafficking of NbDREPP to PD was affected. NbDREPP-GFP was coexpressed with the tails of myosin XI-2, VIII-1, or VIII-B in N. benthamiana leaves via agroinfiltration. The results showed that the trafficking of NbDREPP to PD was disrupted in about 70% of the cells coexpressing NbDREPP-GFP and c-myc-tagged myosin XI-2 tail but not in the cells coexpressing NbDREPP-GFP and the tails of myosin VIII-1 or VIII-B (Fig. 7, B and C). Immunoblot analysis using a GFP-specific or a c-myc-specific antibody confirmed the accumulation of NbDREPP-GFP and various myosin tails in the coagroinfiltrated leaves (Fig. 7D).

NbDREPP Interacted with P3N-PIPO and CI to Facilitate TVMBV Intercellular Movement
Viral MPs need to interact with host factors to mediate virus movement in plants. Multiple host proteins have been identified to interact with viral MPs and to be involved in virus movement in plants (Chen et al., , 2005Raffaele et al., 2009;Amari et al., 2010Amari et al., , 2011Perraki et al., 2012). The potyviral CI known as one of the MPs of potyviruses (Carrington et al., 1998) was shown to target PD with the help of P3N-PIPO in a ratio-dependent manner (Wei et al., 2010). Recently, the plasma membrane-locating AtPCaP1 was shown to interact with TuMV P3N-PIPO during TuMV intercellular movement and to localize at the plasma membrane in Arabidopsis protoplasts (Ide et al., 2007;Vijayapalani et al., 2012). AtPCaP1 is a hydrophilic cationbinding protein without a predicted transmembrane domain and is associated with the plasma membrane via  N-myristoylation at the Gly residue at position 2 (Nagasaki et al., 2008). In this study, we identified that TVBMV P3N-PIPO could interact with NbDREPP, whose homolog was identified previously in N. tabacum and was predicted to be a developmentally regulated plasma membrane protein (Logan et al., 1997). NbDREPP shared 52% amino acid identity and a highly conserved N terminus with AtPCaP1 (Supplemental Fig.  S5A), suggesting that NbDREPP is a homolog of AtPCaP1. We further showed that NbDREPP localized at PD (Fig. 5,

A and B), interacted with both TVBMV P3N-PIPO and CI, and colocalized with P3N-PIPO and CI at PD (Figs. 3 and 5, F and G). It still remains unknown whether AtPCaP1 can localize at PD in epidermal cells and interact with CI.
The subcellular localization of a protein may be affected when it is coexpressed with interacting protein(s). TVBMV CI alone aggregated in the cytoplasm (Figs. 3C and 5D), while NbDREPP localized at PD (Fig. 5A). When coexpressed in N. benthamiana leaf epidermal cells, NbDREPP and CI colocalized at the cell periphery  (Supplemental Fig. S7). Furthermore, in the presence of P3N-PIPO, expressed from a transient expression vector (Fig. 5F) or from the virus (Fig. 5G), CI and NbDREPP were found to colocalize at PD. These results suggested that, in the infection process of TVBMV, P3N-PIPO, CI, and NbDREPP might form a complex at PD to comediate potyvirus movement between cells.
NbDREPP was predicted to be an ID protein with a long ID region in the C terminus (Supplemental Fig. S5, B and C). Most known ID proteins undergo folding upon interaction with their partners (Marín and Ott, 2014). Interaction of NbDREPP with P3N-PIPO and/or CI might induce NbDREPP to undergo disorder-to-order transitions prior to the modification of PD. Similar to the role of AtPCaP1 in TuMV movement, silencing of NbDREPP or SlPCaP1 (the homologous gene of AtPCaP1 in tomato [Solanum lycopersicum]) expression through VIGS impeded TVBMV infection ( Fig. 4; Supplemental  Fig. S8). However, knockdown of AtPCaP1 or NbDREPP expression in plants did not completely inhibit the intercellular movement of TuMV or TVBMV, suggesting that additional host factor(s) must participate in their intercellular movement.

NbDREPP Targeted to PD via the Early Secretory Pathway and Actomyosin Network
Although several host intracellular trafficking systems have been reported for virus intracellular trafficking (Lazarowitz and Beachy, 1999;Jackson, 2000;Fedorkin et al., 2001), the molecular mechanism by which viral MPs and their interacting host factors are delivered to the PD remained largely unknown (Harries et al., 2010). For example, targeting PD by Cowpea mosaic virus MP or Poa semilatent virus TGBp3 was not affected by disrupting the ER-Golgi transport pathway or the cytoskeleton network (Pouwels et al., 2002;Schepetilnikov et al., 2008). In contrast, the MPs of Beet yellows virus and Rice stripe virus required the participation of class VIII myosins (Avisar et al., 2008a;Yuan et al., 2011). Silencing the myosin XI-2 gene in N. benthamiana inhibited the movement of TMV but not Potato virus X, Tomato bushy stunt virus, or Turnip vein clearing virus (Harries et al., 2009). The routes by which viral MPs and their interacting host proteins move to PD may differ. For example, GFLV MP targeted PD through diffusion or trafficking along the microtubules (Laporte et al., 2003), while its interacting protein AtPDLP1 targeted PD via the early secretory pathway in a myosin XI-2-and XI-K-dependent manner (Thomas et al., 2008;Amari et al., 2011). Delivery of TuMV P3N-PIPO and CI to PD also required the early secretory pathway but not the actomyosin motility system (Wei et al., 2010). However, the cell-to-cell movement of TuMV required both a functional secretory pathway and actomyosin network, although the target proteins within these host intracellular transport machineries have not been determined (Agbeci et al., 2013). Here, we showed that the trafficking of NbDREPP to PD depended on the early secretory pathway in a myosin XI-2-dependent manner (Figs. 6 and 7). Like TuMV P3N-PIPO, TVBMV P3N-PIPO also was transported to PD via the early secretory pathway but not the actomyosin motility system (Supplemental Fig. S9). The results suggested that NbDREPP trafficked to PD by a route different from that of P3N-PIPO and formed movement complexes with P3N-PIPO and CI at PD to mediate the intercellular movement of TVBMV. Therefore, NbDREPP is a key host protein within the early secretory pathway and the actomyosin motility system that interacts with two MPs and influences virus movement.

PIPO Is Expressed as a P3N-PIPO Fusion in TVBMV-Infected Cells and Has Length Variability
A previous study suggested that the PIPO of TuMV was translated as a P3-PIPO fusion through a ribosomal frame-shifting or a transcriptional slippage strategy (Chung et al., 2008). Later, this fusion was confirmed in TuMV-infected plant tissues through western-blot assays using a P3-and PIPO-specific antibody (Vijayapalani et al., 2012). In this study, we continued this work by fusing a His tag to the N terminus of TVBMV P3. Our result showed that the TVBMV PIPO was indeed expressed as the P3N-PIPO fusion during TVBMV infection (Fig. 1B). Our findings, together with the previous reports, indicated that the P3N-PIPO fusion could be found in plant tissues infected with different potyviruses.
The length of the P3N-PIPO fusion was reported to vary among different potyviruses and even among different isolates of the same species (Chung et al., 2008;Cuevas et al., 2012). Our sequence alignments using the published sequences of Potato virus Y and Plum pox virus agreed with the above reports (Supplemental Fig. S2). More recently, Hillung et al. (2013) reported that the PIPO length variation was controlled by a host-driven selection mechanism. Our results presented here showed that the TVBMV mutants with a deletion of two amino acid residues or an extension of up to 38 amino acid residues in the PIPO domain were functional in mediating TVBMV movement ( Fig. 2; Supplemental Fig. S2). The longest movement-competent TVBMV PIPO domain tested in this study contained 40 more amino acid residues, which account for two-thirds of the wild-type PIPO domain, than the minimum movement-competent PIPO domain. This finding shed new light on the population diversity of potyviral P3N-PIPO.
RNA-dependent RNA polymerase (RdRp) of RNA viruses generates numerous mutations in viral genomes during error-prone RNA replication (Domingo and Holland, 1997;Moya et al., 2004;Lauring and Andino, 2010). Some mutations generated by viral RdRp are Figure 8. Model for NbDREPP intracellular trafficking and potyvirus cell-to-cell movement. NbDREPP targets to PD through a route within the early secretory pathway in a myosin XI-2-dependent manner. Potyvirus produces PIPO domains with different lengths during its low-fidelity replication. Only the PIPO domains that meet the length requirements are competent to facilitate potyvirus cell-to-cell movement. NbDREPP interacts with both P3N-PIPO and CI at PD. The three proteins are essential components in the movement complex mediating the intercellular movement of potyvirus.
lethal to viruses, while other mutations may increase the fitness of viruses and lead to viral quasispecies in nature (Domingo and Holland, 1997). Because the truncated mutations in the TVBMV PIPO domain had no influence on viral replication, mutant viruses might continue to produce mutations in this region. Although spontaneous mutations occurring precisely at sites of the introduced stop codons were rare, once a mutation led to an elimination of the introduced stop codon, the progeny viruses would restore systemic movement activity and cause systemic infection. It is noteworthy that temperature and host plants had profound effects on the frequency of spontaneous mutations (Table I). This finding agreed with a previous report showing that host plants and temperature could affect the fidelity of viral RdRp (Pita et al., 2007).

A New Model for NbDREPP and Potyvirus Movement
Combining our data with the results published previously (Wei et al., 2010;Amari et al., 2011;Vijayapalani et al., 2012), we propose a new model for NbDREPP intracellular trafficking and potyvirus cell-to-cell movement (Fig. 8). Unlike the two recently proposed models for potyvirus movement (Wei et al., 2010;Vijayapalani et al., 2012), our model emphasizes that (1) the length of P3N-PIPO is variable and only those that meet the length requirements can mediate potyvirus movement; (2) NbDREPP interacts with both P3N-PIPO and CI at PD, and they are all essential components of the potyviral movement complex; and (3) NbDREPP traffics to PD via the COPI-and COPII-dependent early secretory pathway and is empowered by the motor of myosin XI-2. The results presented here further improved our understanding of the mechanism of potyvirus movement and the discovery of new host resistance against the largest group of plant-infecting viruses.

Plasmid Construction
The full-length intron-containing TVBMV-GFP sequence was released from pTVBMV-GFP (Gao et al., 2012) using restriction enzymes SalI and SmaI (New England Biolabs) and inserted into pCambia0390 to generate pCamTVBMV-GFP for agroinfiltration into Nicotiana benthamiana plants. To construct pCamTVBMV-GFP-HISP3N, a TVBMV infectious clone with a His tag-labeled P3, 24 nucleotides encoding six His residues plus a Gly and a Leu were inserted into pCamTVBMV-GFP, two amino acids after the HC-Pro/P3 cleavage site (KHYRVG/GL), through PCR using the primers listed in Supplemental Table S1. Four constructs (i.e. pCamPIPO7aaSTOP, pCamPIPO20aaSTOP, pCamPIPO43aaSTOP, and pCamPIPO58aaSTOP) were obtained by introducing a stop codon in the pipo ORF to produce a truncated PIPO domain of seven, 20, 43, and 58 amino acid residues, respectively. Three constructs (pCamPIPOPLUS8aa, pCamPIPO-PLUS22aa, and pCamPIPOPLUS38aa), which produce extended PIPO domains, were obtained by disrupting the postulated stop codon of the TVBMV pipo ORF (GenBank accession no. EU734432) through PCR using specific primers as described previously (Liu and Naismith, 2008;Supplemental Table S1). The authenticity of the resulting plasmids was confirmed through sequencing before further use.

Plant Growth, Protein Transient Expression, and Virus Inoculation
Unless indicated otherwise, N. benthamiana and Nicotiana tabacum plants were grown inside a growth chamber at 25°C with a 16-h-light/8-h-dark cycle. Plasmid pCamTVBMV-GFP and its derivatives were introduced individually into Agrobacterium tumefaciens strain GV3101. The transformed A. tumefaciens cultures were grown overnight in Luria-Bertani medium containing appropriate antibiotics and then incubated in an induction buffer (10 mM MgCl 2 , 150 mM acetosyringone, and 10 mM MES) for 3 h at room temperature. Individual A. tumefaciens culture was adjusted to OD 600 = 0.3 for protein expression, OD 600 = 0.5 for virus inoculation, or as indicated otherwise. The diluted A. tumefaciens cultures were infiltrated individually into leaves of N. benthamiana or N. tabacum using needleless syringes.

Coimmunoprecipitation and Immunoblot Assays
For coimmunoprecipitation assays, agroinfiltrated N. benthamiana leaves were collected and ground in liquid nitrogen. Total proteins from the harvested N. benthamiana leaves were extracted using an extraction buffer containing 25 mM Tris, pH 7.5, 1 mM EDTA, 150 mM NaCl, 10 mM dithiothreitol, 1 mM phenylmethanesulfonyl fluoride, 10% (v/v) glycerol, 2% (w/v) polyvinylpolypyrrolidone, 0.15% (v/v) Nonidet P-40, and 13 protease inhibitor cocktail (Roche). The crude leaf extracts were centrifuged at 20,000g for 15 min, and the supernatants were incubated overnight at 4°C with 5 mg of rabbit anti-HA or anti-c-myc antibody as instructed (Sigma-Aldrich). Fifty-microliter protein A/G plus agarose beads (Santa Cruz Biotechnology) were equilibrated with the extraction buffer and then added to each assay sample. After 4 h of incubation at 4°C, the protein/bead complex was pelleted at 600g for 3 min followed by four washes in a washing buffer (25 mM Tris, pH 7.5, 1 mM EDTA, 150 mM NaCl, and 10% [v/v] glycerol). Proteins in the resulting samples were separated on 10% (w/v) SDS-PAGE gels. After transferring protein bands to nitrocellulose membranes, the membranes were probed with a mouse antic-Myc or anti-HA antibody (Santa Cruz Biotechnology). The detection signal was visualized using the Pierce ECL western-blot substrates (Thermo Fisher Scientific).

RT-PCR and Northern Blotting
Total RNAs were extracted from the harvested N. benthamiana leaves using Trizol reagent (TransGen Biotech) and then treated with RNase-free DNase I (Takara). RT was conducted with an oligo(dT) 18 primer and the EasyScript First-Strand cDNA Synthesis SuperMix kit (TransGen Biotech). PCR was performed using the specific primers listed in Supplemental Table S1 and a Phusion High-Fidelity DNA Polymerase (Thermo Fisher Scientific). The PCR products were visualized on agarose gels through electrophoresis.
Plant Physiol. Vol. 167, 2015 407 For northern-blot assays, the agroinfiltrated N. benthamiana leaves were harvested at 3 dpai. Total RNAs were extracted from the harvested leaf tissues, separated on 1% (w/v) denaturing agarose gels containing 2% (v/v) formaldehyde, and transferred to Hybond N + nylon membranes (GE Healthcare) by capillary action. The membranes were probed for viral RNA using a digoxigenin-labeled riboprobe specific for the TVBMV COAT PROTEIN ORF. The RNA band intensity was analyzed by Quantity One software (Bio-Rad).

Yeast Two-Hybrid and BiFC Assays
To identify host protein(s) that interact with TVBMV P3N-PIPO, an N. benthamiana cDNA library was constructed and screened using the GAL4 Y2H kit (Clontech) as instructed. Full-length P3N-PIPO ORF was PCR amplified and cloned into the bait vector pGBKT7. Briefly, the library Saccharomyces cerevisiae Y187 strain and the AH109 strain harboring pGBKT7-P3N-PIPO were mated, plated on the double dropout medium containing X-a-Gal (SD/2Leu/2Trp/+Xa-Gal), and then incubated at 30°C for 3 d. The positive cotransformants were then subjected to high-stringency screenings on the quadruple dropout medium with SD/2Leu/2Trp/2adenine/2His. The protein-protein interactions were then verified in yeast strain Y2HGold.
For BiFC assays, two fusion constructs were agroinfiltrated into N. benthamiana leaves using needless syringes. Cells expressing the fusion proteins were imaged using a Zeiss LSM510 META laser scanning confocal microscope.

Pharmacological Assays
The pharmacological assays were performed as described previously (Brandizzi et al., 2002). To disrupt the actin microfilament, leaves agroinfiltrated with various constructs were infiltrated with 5 mM LatB (Sigma-Aldrich) in 0.1% (v/v) DMSO at 40 hpai. The LatB-treated leaves were allowed to grow for 12 h before examining with the confocal microscope. To investigate the transport of NbDREPP between the ER and Golgi, the agroinfiltrated N. benthamiana leaves were infiltrated with BFA (50 mg mL 21 in 0.1% [v/v] DMSO) at 48 hpai. The BFAtreated leaves were allowed to grow for 3 h before confocal microscopy examination. Leaves infiltrated with 0.1% (v/v) DMSO were used as controls for these assays.

Fluorescence and Confocal Microscopy Observation
A. tumefaciens cultures (OD 600 = 0.5) harboring the parental or mutant TVBMV-GFP constructs were diluted with induction buffer at a ratio of 1:10,000 before agroinfiltration. The agroinfiltrated N. benthamiana areas were harvested at 3 dpai and examined with a Leica M165 FC fluorescence stereomicroscope. GFP was excited with light at 488 nm. All images were captured with a DFC 420C camera (Leica) and processed using the Leica Application Suite version 3.4.0.
For confocal microscopy analysis, epidermal cells of N. benthamiana leaves were examined with a Zeiss LSM510 META laser scanning confocal microscope equipped with a Plan-Neofluar 403/1.3 oil differential interference contrast lens or a Plan-Apochromat 633/1.4 oil differential interference contrast lens and a multitrack mode. GFP was excited at 488 nm, and the emitted signal was captured at 505 to 530 nm. YFP was excited at 514 nm and captured at 530 to 600 nm. DsRed was excited at 543 nm and captured at 615 nm. All images were captured digitally and processed using the Zeiss LSM Image Examiner version 4.0.

VIGS
A 564-bp fragment (nucleotide residues 145-708) of NbDREPP and an 819bp fragment (nucleotide residues 229-1,047) of NbANK were RT-PCR amplified from an N. benthamiana leaf cDNA using primers NbDREPP-EcoRIF and NbDREPP-BamHIR or NbANK-EcoRIF and NbANK-KpnIR (Supplemental Table S1). The PCR products were cloned individually into the pTRV2 vector (Liu et al., 2002) to generate pTRV2-NbDREPP and pTRV2-NbANK. A. tumefaciens cultures harboring pTRV1, pTRV2, pTRV2-NbDREPP, or pTRV2-NbANK were grown separately overnight in Luria-Bertani medium containing appropriate antibiotics. The cultures were pelleted individually and then incubated in an induction buffer (10 mM MgCl 2 , 150 mM acetosyringone, and 10 mM MES) for 3 h at room temperature. A. tumefaciens culture harboring pTRV1 was mixed with an equal volume of A. tumefaciens culture harboring pTRV2 (control vector), pTRV2-NbDREPP, or pTRV2-NbANK. The mixed A. tumefaciens cultures were infiltrated individually into N. benthamiana leaves as described (Liu et al., 2002). At 14 dpai, the upper nonagroinfiltrated N. benthamiana leaves were infiltrated again with A. tumefaciens culture harboring pCamTVBMV-GFP. The infiltrated plants were again grown in the growth chamber, and the first fully developed leaves above the pCamTVBMV-GFP-infiltrated leaves were harvested at 5 and 14 dpai. Silencing of the NbDREPP or NbANK gene in the agroinfiltrated N. benthamiana plants was analyzed by semiquantitative RT-PCR using primers NbDREPP-F and NbDREPP381-R or NbANK-F and NbANK392-R (Supplemental Table S1). The relative expression levels of the ELONGATION FACTOR1a gene were determined using primers EF1a-F and EF1a-R and used as internal controls for the assay (Supplemental Table S1).

Supplemental Data
The following supplemental materials are available.
Supplemental Figure S1. Amino acid alignment of the PIPO domains of different PVY and PPV isolates.
Supplemental Figure S2. Infectivity of the wild-type and mutant TVBMV with extended PIPO domains in N. benthamiana plants.
Supplemental Figure S3. Detection of the wild-type and mutant TVBMV RNA in the systemic leaves of N. benthamiana through RT-PCR at 7 dpai.
Supplemental Figure S4. Quantitative analysis of the Northern-blot results using Quantity One software (Bio-Rad).
Supplemental Figure S5. Amino acid alignment of NbDREPP with its orthologs and prediction of protein disorder.
Supplemental Figure S6. BiFC detection of the interaction between NbDREPP and AtPDPL1.
Supplemental Figure S9. Effect of the early secretory pathway and actomyosin system on the PD localization of TVBMV P3N-PIPO.
Supplemental Table S1. Sequences of the primers used in this study.