The p23 of Citrus Tristeza Virus Interacts with Host FKBP-Type Peptidyl-Prolylcis-Trans Isomerase 17-2 and Is Involved in the Intracellular Movement of the Viral Coat Protein

Citrus tristeza virus is a member of the genus Closterovirus in the family Closteroviridae. The p23 of citrus tristeza virus (CTV) is a multifunctional protein and RNA silencing suppressor. In this study, we identified a p23 interacting partner, FK506-binding protein (FKBP) 17-2, from Citrus aurantifolia (CaFKBP17-2), a susceptible host, and Nicotiana benthamiana (NbFKBP17-2), an experimental host for CTV. The interaction of p23 with CaFKBP17-2 and NbFKBP17-2 were individually confirmed by yeast two-hybrid (Y2H) and bimolecular fluorescence complementation (BiFC) assays. Subcellular localization tests showed that the viral p23 translocated FKBP17-2 from chloroplasts to the plasmodesmata of epidermal cells of N. benthamiana leaves. The knocked-down expression level of NbFKBP17-2 mRNA resulted in a decreased CTV titer in N. benthamiana plants. Further, BiFC and Y2H assays showed that NbFKBP17-2 also interacted with the coat protein (CP) of CTV, and the complexes of CP/NbFKBP17-2 rapidly moved in the cytoplasm. Moreover, p23 guided the CP/NbFKBP17-2 complexes to move along the cell wall. To the best of our knowledge, this is the first report of viral proteins interacting with FKBP17-2 encoded by plants. Our results provide insights for further revealing the mechanism of the CTV CP protein movement.


Gene Cloning
Primers used for the amplification of genes coding for proteins p23 and CP of CTV were designed based on multiple alignments of the CTV sequences available in GenBank. Primers used for the amplification of FKBP17-2 coding genes were designed based on cDNA sequences of CaFKBP17-2 from C. aurantifolia and NbFKBP17-2 from N. benthamiana, respectively. Total RNA was extracted using TRIzol reagent (Invitrogen, Carlsbad, CA, USA) according to the manufacturer's instructions. The isolated RNA was used for firststrand cDNA synthesis via reverse transcription with the Maloney murine leukemia virus reverse transcriptase (M-MLV, Promega, Madison, WI, USA) and p(N)6 random primer (TaKaRa, Dalian, China) at 37 • C for 2 h. The PCR products were purified and ligated into the pMD18-T vector (TaKaRa, Dalian, China). Clones were identified via PCR and at least three positive clones of each product were sequenced at Sangon Biotech (Shanghai) Co., Ltd., Shanghai, China. One clone of each gene with a consensus sequence was selected for further study. The primers are listed in Table S1. GenBank accession numbers of genes analyzed in this study are as follows: CaFKBP17-2 (MW600634), NbFKBP17-2 (MW600635), CTV p23 (MW600636), and CTV CP (EF063109.1).

DUALMembrane Yeast Two-Hybrid Assay
The DUALmembrane yeast two-hybrid (mYTH) system was used for testing proteinprotein interactions following the manufacturer's protocols (Dualsystems Biotech AG, Switzerland). The cDNAs of CaFKBP17-2 and NbFKBP17-2 were individually cloned into the library vector pPR3-N by fusing to the C-terminal half of Ubiquitin (Cub), and p23 was fused to the N-terminal half of Ubiquitin (NubG) of the bait vector pDHB1. The recombined pPR3-N and pDHB1 plasmids were co-transformed into Saccharomyces cerevisiae strain NMY51. The transformed yeast cells were grown on plates containing selection medium lacking tryptophan and leucine (SD-Trp,-Leu) (Clontech, Mountain View, CA, USA) at 30 • C for 2-4 days. The surviving clones were transferred to quadruple dropout medium lacking tryptophan, leucine, histidine and adenine (SD-Trp,-Leu,-His,-Ade) (Clontech, Mountain View, CA, USA) in the presence of 50 mM 3-amino-1,2,4-triazole (3-AT) (Sigma-Aldrich, St. Louis MO, USA). The self-interaction of CTV p20 was used as a positive control.

BiFC and Subcellular Localization Assays
Gateway Technology (Invitrogen, Carlsbad, CA, USA) was used for the construction of BiFC vectors. To facilitate subsequent Gateway cloning, primer sequences were flanked with attB recombination sites at their 5 ends (Table S1). PCR products containing attB sites were gel-purified and recombined into the pDONR/Zeo vector using the Gateway™ BP Clonase™ Enzyme Mix (Invitrogen, Carlsbad, CA, USA). The genes in the pDONR/Zeo vectors were recombined into the modified pEarlygate201-YN or pEarlygate202-YC vectors [39] using the Gateway™ LR Clonase™ II Enzyme mix (Invitrogen, Carlsbad, CA, USA).
For subcellular localization assays, the genes coding for enhanced yellow fluorescent protein (eYFP) and enhanced cyan fluorescent protein (eCFP) were ligated into vector pCNF3 at the KpnI and SmaI sites [40] to generate plant expression constructs 2× 35S-MCS-eYFP (pCNY) and 2× 35S-MCS-eCFP (pCNC), respectively. Genes were amplified from the corresponding clones by using primers with an XbaI or BamHI digestion site (Table S1), digested with enzymes XbaI and BamHI and inserted into vector pCNY or pCNC. The mCherry-HDEL vector was used as an endoplasmic reticulum (ER) marker [41]. The plant expression constructs were individually transformed into cells of Agrobacterium tumefaciens strain GV3101 by using a heat shock method. The GV3101 cells harboring the expression constructs were resuspended in the infiltration buffer (10 mM MgCl 2 , 10 mM 2-(4-morpholino) ethanesulfonic acid, and 150 mM acetosyringone) and infiltrated into N. benthamiana leaves, as previously described [42]. Callose deposition at plasmodesmata of infiltrated N. benthamiana samples were stained with a water solution of 0.05% Cells 2021, 10, 934 4 of 14 (wt/vol) aniline blue (Coolaber, Beijing, China) at 30 min before image taking as previously described [43].
Videos were taken using the time series program in TCS-SP8 CLSM. Collected images were analyzed with the Leica Application Suite for Advanced Fluorescence (LAS AF Lite, version 3.3) software.

Virus-Induced Gene Silencing (VIGS) and CTV Inoculation
VIGS vectors pTRV1 and pTRV2 were used for the RNA silencing analyses [44]. A fragment of NbFKBP17-2 cDNA (position 80-296 nt of the gene) was amplified from N. benthamiana and then cloned into pTRV2. The resulting plasmid pTRV2-FKBP17-2 was electroporated into A. tumefaciens cells and co-infiltrated with an A. tumefaciens culture harboring pTRV1 (final optical density of 0.7 at OD600) into leaves of 4-week old N. benthamiana plants. At 7 dpi, leaves of NbFKBP17-2 silenced plants and control plants were agroinfiltrated with CTV infectious clone.

Real-Time RT-PCR
Total RNA was extracted from N. benthamiana samples using TRIzol reagent (Invitrogen, Carlsbad, CA, USA). The cDNA was prepared using 5× All-In-One RT MasterMix (with AccuRT Genomic DNA Removal Kit) (Abm, VAN, Canada).
The relative expression levels of CTV genome and NbFKBP17-2 mRNA were tested using RT-qPCR. RT-qPCR was performed in a CFX96™ Real-time system (Bio-Rad, Hercules, CA, USA) using the iTaq TM Universal SYBR ® Green Supermix (Bio-Rad, Hercules, CA, USA). CTV-specific primers amplifying the intergenic region between the RNA-dependent RNA polymerase and p33 ORFs were used for RT-qPCR assays of the CTV genome quantity [45]. The N. benthamiana actin gene (accession number AY179605) was used as an internal control. Values for relative FKBP17-2 expression was calculated by the 2 -∆∆C T method [46] and were normalized to the transcriptional value of an actin gene [38,47]. Results were obtained from three independent assays. The primers for RT-qPCR are listed in Table S1.
showed that these spots colocalized with the blue signals of aniline blue stained callose at PD ( Figure 1C), but did not move in the cytoplasm under time-lapse microscopy observation ( Figure S1). The results further confirmed that the interactions between the CTV protein p23 and FKBP17-2 proteins from C. aurantifolia and N. benthamiana, and their interaction signals mainly happened at PD. There was no interaction signal observed for the control combinations of p23-YC with unfused YN and FKBP17-2-YN with unfused YC ( Figure  S2). The yeast cell of co-transformed with p20-Cub and NubG-p20 was used as a positive control, and the yeast cell with p23-Cub and NubG was used as a negative control, respectively. A series of dilutions (1, 10 -1 , 10 -2 , and 10 -3 ) were shown at the top of the figure. (C) Bimolecular-fluorescence complementation assays. The p23 fused to YFP C-terminal (p23-YC) was separately co-expressed with CaFKBP17-2 and NbFKBP17-2 fused to YFP N-terminal (CaFKBP17-2-YN and NbFKBP17- Shaded letters indicate the conserved amino acid residues. The FK506 binding domain (FKBd) was indicated by a red box. (B) Split-ubiquitin yeast two-hybrid assays. p23 was fused to the C-terminal half of ubiquitin (Cub). CaFKBP17-2 and NbFKBP17-2 were respectively fused to the N-terminal half of ubiquitin (NubG). The yeast cell of co-transformed with p20-Cub and NubG-p20 was used as a positive control, and the yeast cell with p23-Cub and NubG was used as a negative control, respectively. A series of dilutions (1, 10 −1 , 10 −2 , and 10 −3 ) were shown at the top of the figure. (C) Bimolecular-fluorescence complementation assays. The p23 fused to YFP C-terminal (p23-YC) was separately co-expressed with CaFKBP17-2 and NbFKBP17-2 fused to YFP N-terminal (CaFKBP17-2-YN and NbFKBP17-2-YN) by agroinfiltration on N. benthamiana leaves. Aniline blue stained callose at plasmodesmata (PD) was used as a marker. Arrow heads pointed to PD. Confocal images were taken at 48 hpi. Bars, 20 µm.
The interactions of p23 with CaFKBP17-2 and NbFKBP17-2 were further investigated using the BiFC assay in N. benthamiana leaves. In this assay, p23 was fused to the C-terminal fragment of YFP (p23-YC), and CaFKBP17-2 and NbFKBP17-2 were respectively fused to the N-terminal fragment of YFP (CaFKBP17-2-YN and NbFKBP17-2-YN). The co-expression of p23-YC and CaFKBP17-2-YN, and p23-YC and NbFKBP17-2-YN via agroinfiltration resulted in extensive YFP fluorescence signals in the cytoplasm of agroinfiltrated cells at 48 h post infiltration (hpi). The YFP fluorescence signals of p23 interacting with CaFKBP17-2 and NbFKBP17-2 emerged as discontinuous spots along the cell membranes and granular structures in the cytoplasm, further staining of the infiltration leaf sections showed that these spots colocalized with the blue signals of aniline blue stained callose at PD ( Figure 1C), but did not move in the cytoplasm under time-lapse microscopy observation ( Figure S1). The results further confirmed that the interactions between the CTV protein p23 and FKBP17-2 proteins from C. aurantifolia and N. benthamiana, and their interaction signals mainly happened at PD. There was no interaction signal observed for the control combinations of p23-YC with unfused YN and FKBP17-2-YN with unfused YC ( Figure S2).

Subcellular Localization of p23 and NbFKBP17-2 in N. benthamiana Leaf Cells
The subcellular localization of p23 in the epidermal cells of N. benthamiana leaves was determined using the transient expression vectors pCNC and pCNY. The fluorescence signals in infiltrated leaf cells were visualized under confocal microscopy at 2 dpi. When eCFP was expressed alone, the fluorescence signals visualized for free eCFP located in the cytoplasm and nucleus almost uniformly, as well as in actin filaments close to the nucleus ( Figure S3). The fluorescence signal of p23-eCFP accumulated preferentially as parallel spots at plasmodesmata (Figure 2A), similar to what was previously reported [17]. The fluorescent signals of diffused NbFKBP17-2-eYFP appeared as bodies with diameter about 2.5-4.5 µm in chloroplasts, and most of these bodies colocalized with chlorophyll autofluorescent signals ( Figure 2B). Further overlapping the fluorescence spectra of p23 with chlorophyll and NbFKBP17-2 with chlorophyll showed that NbFKBP17-2, but not p23, localized to chloroplasts ( Figure 2C).

Subcellular Localization of p23 and NbFKBP17-2 in N. benthamiana Leaf Cells
The subcellular localization of p23 in the epidermal cells of N. benthamiana leaves was determined using the transient expression vectors pCNC and pCNY. The fluorescence signals in infiltrated leaf cells were visualized under confocal microscopy at 2 dpi. When eCFP was expressed alone, the fluorescence signals visualized for free eCFP located in the cytoplasm and nucleus almost uniformly, as well as in actin filaments close to the nucleus ( Figure S3). The fluorescence signal of p23-eCFP accumulated preferentially as parallel spots at plasmodesmata (Figure 2A), similar to what was previously reported [17]. The fluorescent signals of diffused NbFKBP17-2-eYFP appeared as bodies with diameter about 2.5-4.5 μm in chloroplasts, and most of these bodies colocalized with chlorophyll autofluorescent signals ( Figure 2B). Further overlapping the fluorescence spectra of p23 with chlorophyll and NbFKBP17-2 with chlorophyll showed that NbFKBP17-2, but not p23, localized to chloroplasts ( Figure 2C).  cence signals of NbFKBP17-2-eYFP were translocated from chloroplasts to plasmodesmata and cytoplasm ( Figure 3A). The co-localization feature of p23-eCFP and NbFKBP17-2-eYFP was similar to their interaction signals in BiFC assays ( Figure 1C). Occasionally, we found the fluorescence signal of p23-eCFP as small dots in a few chloroplasts where NbFKBP17-2-eYFP also presented ( Figure 3B). The overlapping fluorescence spectra also showed that p23 translocated NbFKBP17-2 partially from chloroplasts to PD ( Figure 3C,D). The co-localization features of p23 and NbFKBP17-2 were examined by co-expressing the two proteins in the epidermal cells of N. benthamiana leaves. A profound change of NbFKBP17-2-eYFP localization was observed in the presence of p23-eCFP. Most YFP fluorescence signals of NbFKBP17-2-eYFP were translocated from chloroplasts to plasmodesmata and cytoplasm ( Figure 3A). The co-localization feature of p23-eCFP and NbFKBP17-2-eYFP was similar to their interaction signals in BiFC assays ( Figure 1C). Occasionally, we found the fluorescence signal of p23-eCFP as small dots in a few chloroplasts where NbFKBP17-2-eYFP also presented ( Figure 3B). The overlapping fluorescence spectra also showed that p23 translocated NbFKBP17-2 partially from chloroplasts to PD ( Figure 3C,D).

NbFKBP17-2 and p23 Were Involved in CTV CP Movement
Since p23 alone and p23 together with NbFKBP17-2 accumulated preferentially at PD as tested above, we suspected that the co-localization features of these proteins might be involved in the virus movement. Considering that many plant viruses move as nucleocapsid, we tested the accumulation type of CTV CP in N. benthamiana cells. Results showed that CTV CP alone located at PD, and in nucleus and cytoplasm ( Figure S4).
BiFC assay showed that CP-YC and NbFKBP17-2-YN interacted in N. benthamiana leaf cells and YFP fluorescence signals appeared as numerous bodies with variable sizes in the cytoplasm of agroinfiltrated N. benthamiana cells at 48 hpi, but did not colocalize with chloroplasts. There was no interaction signal observed for the control combination

NbFKBP17-2 and p23 Were Involved in CTV CP Movement
Since p23 alone and p23 together with NbFKBP17-2 accumulated preferentially at PD as tested above, we suspected that the co-localization features of these proteins might be involved in the virus movement. Considering that many plant viruses move as nucleocapsid, we tested the accumulation type of CTV CP in N. benthamiana cells. Results showed that CTV CP alone located at PD, and in nucleus and cytoplasm ( Figure S4).
BiFC assay showed that CP-YC and NbFKBP17-2-YN interacted in N. benthamiana leaf cells and YFP fluorescence signals appeared as numerous bodies with variable sizes in the cytoplasm of agroinfiltrated N. benthamiana cells at 48 hpi, but did not colocalize with chloroplasts. There was no interaction signal observed for the control combination of CP-YC with unfused YN ( Figure 4A). The interaction of CP with NbFKBP17-2 was further investigated by Y2H assay. The yeast cells transformed with paired vectors p20-Cub and NubG-p20 (a positive control), and CP-Cub and NubG-NbFKBP17-2 were able to grow on selective medium (SD-Trp,-Leu,-His,-Ade) complemented with 50 mM 3-AT, while yeast transformants carrying CP-Cub+NubG, and Cub+NubG-NbFKBP17-2 did not grow, indicating that the CP interacted with NbFKBP17-2 ( Figure 4B). Surprisingly, time-lapse microscopy observation found that the CP/NbFKBP17-2 bodies with different sizes could rapidly and irregularly move in the cytoplasm ( Figure 4C, upper panel), while the self-interaction of CP appeared as uniform distribution in nucleus and plasmodesmata ( Figure S5), but did not form bodies in the cytoplasm and chloroplasts ( Figure 4C, low panel). Further assays revealed that the interaction complexes moved on the ER ( Figure 4D).
of CP-YC with unfused YN ( Figure 4A). The interaction of CP with NbFKBP17-2 was further investigated by Y2H assay. The yeast cells transformed with paired vectors p20-Cub and NubG-p20 (a positive control), and CP-Cub and NubG-NbFKBP17-2 were able to grow on selective medium (SD-Trp,-Leu,-His,-Ade) complemented with 50 mM 3-AT, while yeast transformants carrying CP-Cub+NubG, and Cub+NubG-NbFKBP17-2 did not grow, indicating that the CP interacted with NbFKBP17-2 ( Figure 4B). Surprisingly, timelapse microscopy observation found that the CP/NbFKBP17-2 bodies with different sizes could rapidly and irregularly move in the cytoplasm ( Figure 4C, upper panel), while the self-interaction of CP appeared as uniform distribution in nucleus and plasmodesmata ( Figure S5), but did not form bodies in the cytoplasm and chloroplasts ( Figure 4C, low panel). Further assays revealed that the interaction complexes moved on the ER ( Figure  4D). Given the fact that NbFKBP17-2 interacts with p23, we suspected that p23 might take part in the movement. The expression vector p23-eCFP and BiFC vectors NbFKBP17-2-YN and CP-YC were co-infiltrated into epidemic cells of N. benthamiana leaves. It was Given the fact that NbFKBP17-2 interacts with p23, we suspected that p23 might take part in the movement. The expression vector p23-eCFP and BiFC vectors NbFKBP17-2-YN and CP-YC were co-infiltrated into epidemic cells of N. benthamiana leaves. It was found that p23-eCFP and CP/NbFKBP17-2 complexes indeed co-localized as round particles in the cytoplasm and near cell membranes ( Figure 5A). Time-lapse observation showed that these particles also moved quickly in the cytoplasm ( Figure 5B and Figure S6). found that p23-eCFP and CP/NbFKBP17-2 complexes indeed co-localized as round particles in the cytoplasm and near cell membranes ( Figure 5A). Time-lapse observation showed that these particles also moved quickly in the cytoplasm (Figures 5B and S6).

Knock-Down Expression of NbFKBP17-2 in N. benthamiana Led to Decreased CTV Accumulation
To examine whether the interaction between p23 and NbFKBP17-2 could influence the accumulation of CTV genome, we constructed a tobacco rattle virus (TRV)-based RNA silencing vector pTRV2-NbFKBP17-2 to knock down NbFKBP17-2 mRNA via virus induced gene silencing (VIGS). The VIGS vector pTRV2-NbFKBP17-2 was co-infiltrated with plasmid pTRV1 into N. benthamiana leaf cells (TRV: NbFKBP17-2). Plants co-infiltrated with the plasmids pTRV2 and pTRV1 were used as negative controls (TRV: 00). At 7 dpi, CTV infectious cDNA clones were infiltrated into NbFKBP17-2-silenced and control plants. Mild chlorosis phenotype was observed on Mock-inoculated NbFKBP17-2-silenced and control plants at 14 dpi ( Figure 6A, panels 1 and 2). At 14 days post CTV inoculation, although the CTV-infected control plants did not show visible disease symptom ( Figure  6A, panel 3), the new leaves of CTV-infected NbFKBP17-2-silenced plants showed local chlorosis and white blotches ( Figure 6A, panel 4). RT-qPCR assays showed that actin mRNA accumulated stably in N. benthamiana plants during CTV infection, which was consistent with the results from CTV-infected citrus plants [38,46]. The relative expression level of endogenous NbFKBP17-2 in N. benthamiana was knocked down to about 0.33-, 0.44-and 0.68-fold ( Figure 6B), and the accumulation levels of CTV genome in NbFKBP17-2-silenced plants were significantly lower than those in control plants ( Figure 6C). The

Knock-Down Expression of NbFKBP17-2 in N. benthamiana Led to Decreased CTV Accumulation
To examine whether the interaction between p23 and NbFKBP17-2 could influence the accumulation of CTV genome, we constructed a tobacco rattle virus (TRV)-based RNA silencing vector pTRV2-NbFKBP17-2 to knock down NbFKBP17-2 mRNA via virus induced gene silencing (VIGS). The VIGS vector pTRV2-NbFKBP17-2 was co-infiltrated with plasmid pTRV1 into N. benthamiana leaf cells (TRV: NbFKBP17-2). Plants co-infiltrated with the plasmids pTRV2 and pTRV1 were used as negative controls (TRV: 00). At 7 dpi, CTV infectious cDNA clones were infiltrated into NbFKBP17-2-silenced and control plants. Mild chlorosis phenotype was observed on Mock-inoculated NbFKBP17-2-silenced and control plants at 14 dpi ( Figure 6A, panels 1 and 2). At 14 days post CTV inoculation, although the CTV-infected control plants did not show visible disease symptom ( Figure 6A, panel 3), the new leaves of CTV-infected NbFKBP17-2-silenced plants showed local chlorosis and white blotches ( Figure 6A, panel 4). RT-qPCR assays showed that actin mRNA accumulated stably in N. benthamiana plants during CTV infection, which was consistent with the results from CTV-infected citrus plants [38,46]. The relative expression level of endogenous NbFKBP17-2 in N. benthamiana was knocked down to about 0.33-, 0.44-and 0.68-fold ( Figure 6B), and the accumulation levels of CTV genome in NbFKBP17-2-silenced plants were significantly lower than those in control plants ( Figure 6C). The results indicated that the decreased expression of endogenous NbFKBP17-2 reduced CTV accumulation in N. benthamiana plants.  Asterisks denoted significant differences in Student's t-test between the two treatments (two-sided, ** p < 0.01).

Discussion
CTV p23 protein has multiple functions on the viral replication, movement, suppression of RNA silencing, and pathogenesis [16][17][18][19]. Accordingly, CTV p23 shows specific subcellular localizations in the nucleolus, Cajal bodies, and plasmodesmata. The p23 is a unique protein of CTV. It is suggested that CTV p23 might have evolved for regulating specific interactions of CTV with its citrus hosts [14]. Considering the unique properties of the viral protein, we examined a yeast library of C. aurantifolia by using CTV p23 as a bait and identified a CTV p23 interacting protein CaFKBP17-2. Furthermore, we confirmed the interaction between CTV p23 and NbFKBP17-2, a homolog of CaFKBP17-2 from an experimental host N. benthamiana for CTV [26]. It was reported that some FKBPs, such as FKBP13 and FKBP16, localized in plant chloroplasts, could regulate photosynthetic membrane assembly [33]. Our results showed that NbFKBP17-2 alone localized in chloroplasts. However, CTV p23/NbFKBP17-2 complexes located at PD as tested in both BiFC and co-localization analyses, suggesting that p23 might change the NbFKBP17-2 subcellular location. CTV p23 is involved in eliciting CTV-like symptoms (vein clearing and chlorotic spots) when it is expressed ectopically as a transgene in several Citrus species [18,19]. It is also the CTV determinant of the seedling yellow syndrome in sour orange and grapefruit (C. paradisi Macf.) [49]. The ectopic expression of CTV p23 as a transgene does not induce similar symptoms in N. benthamiana [19], but expression CTV p23 from a

Discussion
CTV p23 protein has multiple functions on the viral replication, movement, suppression of RNA silencing, and pathogenesis [16][17][18][19]. Accordingly, CTV p23 shows specific subcellular localizations in the nucleolus, Cajal bodies, and plasmodesmata. The p23 is a unique protein of CTV. It is suggested that CTV p23 might have evolved for regulating specific interactions of CTV with its citrus hosts [14]. Considering the unique properties of the viral protein, we examined a yeast library of C. aurantifolia by using CTV p23 as a bait and identified a CTV p23 interacting protein CaFKBP17-2. Furthermore, we confirmed the interaction between CTV p23 and NbFKBP17-2, a homolog of CaFKBP17-2 from an experimental host N. benthamiana for CTV [26]. It was reported that some FKBPs, such as FKBP13 and FKBP16, localized in plant chloroplasts, could regulate photosynthetic membrane assembly [33]. Our results showed that NbFKBP17-2 alone localized in chloroplasts. However, CTV p23/NbFKBP17-2 complexes located at PD as tested in both BiFC and co-localization analyses, suggesting that p23 might change the NbFKBP17-2 subcellular location. CTV p23 is involved in eliciting CTV-like symptoms (vein clearing and chlorotic spots) when it is expressed ectopically as a transgene in several Citrus species [18,19]. It is also the CTV determinant of the seedling yellow syndrome in sour orange and grapefruit (C. paradisi Macf.) [49]. The ectopic expression of CTV p23 as a transgene does not induce similar symptoms in N. benthamiana [19], but expression CTV p23 from a PVX sgRNA is able to incite symptoms in N. benthamiana [50]. Thus, the interaction between p23 and NbFKBP17-2 resulting in the changed subcellular localization of NbFKBP17-2 might affect the leaf cell photosynthesis, then induce disease like symptoms [14]. [51,52]. Except for movement proteins to be responsible for viral movement [53][54][55], other proteins are also known to be involved in the intracellular movement of some viruses. The proteins of some viruses, such as the 126-kDa replication protein of tobacco mosaic virus [56], the membrane-associated P3 of tobacco etch virus, P3N-PIPO of turnip mosaic virus [57,58], the P6 of CaMV [59], and the N protein of tomato spotted wilt tospovirus [60], can form mobile inclusion bodies involved in the virus intracellular movement. The viral coat proteins have multiple functions and are involved in almost every stage of the viral infection cycle [20]. The interaction of host CHUP1 with CaMV P6 can help the viral inclusion bodies moving inside plant cells along microfilament to reach to plasmodesmata [60]. In this study, we found that the interaction complex CP/NbFKBP17-2 could rapidly move in the cytoplasm of agroinfiltrated N. benthamiana cells without CTV infection. Since CTV-CP and NbFKBP17-2 alone located in cytoplasm and chloroplasts, respectively, the interaction might occur in cytoplasm before NbFKBP17-2 transportation into chloroplasts and be necessary for CP transportation in cytoplasm of plant cells. In addition, although we found the moveable CP/NbFKBP17-2 bodies on the ER membrane, the exact trace of the bodies was difficult to be captured since they moved at different layers of the cell. The association of ER structures with the movement action needs an extensive study. The movement proteins encoded by some plant RNA viruses may bind to the viral replication complex and facilitate their association with plasmodesmata and the subsequent viral spread [61]. CPs from a number of plant RNA viruses are required for effective cell-to-cell and long distance movement [54,62]. Viruses in the genera Cucumovirus, Potyvirus, Potexvirus, Comovirus, and some closteroviruses require their MPs and CPs for both cell-to-cell and systemic movement of the viral genomes [62][63][64]. CTV p23 accumulates preferentially at plasmodesmata and facilitates CTV cell-to-cell movement [65]. Our results showed that the p23 could translocate the CP/NbFKBP17-2 complexes from cytoplasm to cell membranes, which provide a new insight into a host protein involved in intracellular trafficking of the CP and the unique protein p23 of CTV. The p33 is another unique non-conserved and PD localized protein of CTV [61], recent study demonstrated interactions of p33 with three proteins CP, p20, and p23, and the four proteins colocalized in the CTV replication factories. In addition, p33 and p23 colocalized as a number of spots on plasma membrane, as well as in a few vesicle-like structures. Then, the p23/NbFKBP17-2/CP motile complexes may function on facilitating the movement of a ribonucleoprotein replication complex to the PD. Further understanding the roles of p33 in the movement of the p23/NbFKBP17-2/CP complexes will help reveal the CTV movement mechanism.

The intracellular and intercellular movement of a plant virus is necessary for the virus systemic infection
The knocked-down expression of NbFKBP17-2 in N. benthamiana does not result in a visible phenotype. However, the decreased expression of NbFKBP17-2 could reduce CTV titer in N. benthamiana. In some N. benthamiana plants with decreased NbFKBP17-2 expression, CTV infection causes severe chlorosis of new leaves, suggesting that the hijack of NbFKBP17-2 by CTV p23 could greatly decreased NbFKBP17-2 accumulation in chloroplasts, which might result in the reduction of chlorophyll. Thus, like a recently reported p23-interacting host protein GAPDH [30], the interaction of NbFKBP17-2 with CTV p23 might facilitate CTV infectious cycle.
Collectively, our findings provide the first evidence that the p23 of CTV has the ability of intracellular trafficking with the viral coat protein by interaction with a host protein.