The essential role of the quasi-long terminal repeat sequence for replication and gene expression of an endogenous pararetrovirus, petunia vein clearing virus

Abstract

Petunia vein clearing virus (PVCV) is a type member of the genus Petuvirus within the Caulimoviridae family and is defined as one viral unit consisting of a single open reading frame (ORF) encoding a viral polyprotein and one quasi-long terminal repeat (QTR) sequence. Since some full-length PVCV sequences are found in the petunia genome and a vector for horizontal transmission of PVCV has not been identified yet, PVCV is referred to as an endogenous pararetrovirus. Molecular mechanisms of replication, gene expression and horizontal transmission of endogenous pararetroviruses in plants are elusive. In this study, agroinfiltration experiments using various PVCV infectious clones indicated that the replication (episomal DNA synthesis) and gene expression of PVCV were efficient when the QTR sequences are present on both sides of the ORF. Whereas replacement of the QTR with another promoter and/or terminator is possible for gene expression, it is essential for QTR sequences to be on both sides for viral replication. Although horizontal transmission of PVCV by grafting and biolistic inoculation was previously reported, agroinfiltration is a useful and convenient method for studying its replication and gene expression.

Introduction

True retroviruses are not known in plants, but pararetroviruses, which are viruses that are classified in the Caulimoviridae family, are common in most plant species. Pararetroviruses are non-enveloped reverse-transcribing viruses with non-covalently closed (open) circular double-stranded DNA (dsDNA) genomes of 7.1–9.8 kbp (Teycheney et al. 2020). They infect a wide range of monocotyledonous and dicotyledonous plants, and some pararetroviruses cause economically important diseases of tropical and subtropical crops (Bhat et al. 2016). Their horizontal transmission occurs through insect vectors and grafting (Teycheney et al. 2020). Animal retroviruses have genomic RNA in their particles, and plant pararetroviruses have genomic DNA in their particles. Since plant pararetroviruses share a common origin with animal retroviruses, they are both classified in the order Ortervirales (Krupovic et al. 2018; Teycheney et al. 2020).

Cauliflower mosaic virus (CaMV) is a type species of the genus Caulimovirus in the family Caulimoviridae and the most extensively studied pararetrovirus (Grimsley et al. 1986; Guilley et al. 1983; Hohn and Rothnie 2013; Sanfaçon and Hohn 1990; Teycheney et al. 2020). The replication cycle of CaMV is as follows. After CaMV infects a host cell, its open circular dsDNA genome is imported into the nucleus, where the discontinuities of the open circular genome are sealed to give closed circular (supercoiled) dsDNA by a putative host DNA repair enzyme that is associated with histone proteins to form a mini-chromosome. A greater-than-genome length transcript with terminal redundant sequences at both ends (35S RNA) is transcribed from this viral mini-chromosomal DNA by host DNA-dependent RNA polymerase II, is exported from the nucleus to the cytoplasm, and serves as a template (the pregenomic RNA) for reverse transcription. The resulting open circular dsDNA is associated with coat proteins to form the virus particle (Grimsley et al. 1986; Guilley et al. 1983; Hohn and Rothnie 2013; Pooggin and Ryabova 2018; Sanfaçon and Hohn 1990). In this study we focused on the replication of petunia vein clearing virus (PVCV), which is classified in the genus Petuvirus of the Caulimoviridae family. Since the molecular mechanism underlying the replication of PVCV remains elusive, the putative replication cycle of PVCV shown in Supplementary Figure S1 is inferred from that of CaMV (Hohn and Rothnie 2013).

Pararetroviruses have a quasi-long terminal repeat (quasi-LTR or QTR) sequence, corresponding to the LTR sequence of retroviruses, one for each copy of the viral sequence (Figure 1; Richert-Pöggeler et al. 2003; Sanfaçon and Hohn 1990). As the LTR sequence contains elements necessary for the replication of retroviruses (Coffin et al. 2021; Krupovic et al. 2018), the QTR sequence of pararetroviruses containing the promoter region and poly-A signal is also essential for their replication and transcription. Although pararetroviruses have only one QTR sequence, it is sufficient for functioning in the circular DNA genome (Noreen et al. 2007; Richert-Pöggeler et al. 2003; Teycheney et al. 2020). Since the LTR sequence of animal retroviruses is recognized when inserted into the host genome, one is found at each end of the retroviral sequence (Coffin et al. 2021). Most plant pararetroviruses including CaMV probably have no integrase and cannot integrate into the host genomic DNA (Krupovic et al. 2018; Richert-Pöggeler and Shepherd 1997).

Figure 1. PVCV constructs for agroinfiltration experiments. (A) Schematic drawings of PVCV constructs used in agroinfiltration experiments. The commercially available plasmid, pRI201-AN, which contains an Agrobacterium T-DNA sequence, was used for plasmid construction. P_35S and T_HSP indicate the CaMV 35S promoter and the HSP terminator sequence derived from A. thaliana, respectively. The green indicates the primer (tRNA)-binding site (PBS). (B) The nucleotide sequence of PVCV QTR (666 bp). It consists of an upstream sequence (50 bp from 1 to 50) and a downstream sequence (616 bp from 6,591 to 7,206) of a single ORF in the full-length PVCV sequence of 7,206 bp (NC_001839.2) deposited in the NCBI database. The partial QTR sequences in PHOH and EHOH are indicated by purple and blue arrows, respectively. The green, red, orange, and blue letters indicate the transcriptional activator binding site, the TATA-box, poly A signal, and PBS, respectively. Conserved protein domains in the ORF are indicated as viral movement protein (VMP), coat protein (CP), retropepsin (pepsin-like aspartic protease) (AP), reverse transcriptase (RT) and RNase H1 (RH1) (Teycheney et al. 2020). The target region of qPCR and the probe regions of Southern and Northern hybridizations are indicated by black arrows.

However, full-length pararetrovirus sequences and partial pararetrovirus-related sequences are usually found as endogenous viral elements (EVEs, viral DNA integrated into the host nuclear genome) in various plant species (Becher et al. 2014; Diop et al. 2018; Geering et al. 2014; Richert-Pöggeler et al. 2021), as it is also well known that endogenous retroviruses and retrotransposons are major components of animal genomes (Johnson 2019). The majority of pararetroviral EVEs, which are widespread in the genomes of vascular plants, are considered to have defective replication and do not induce any diseases in their host plants (Chabannes and Iskra-Caruana 2013; Staginnus and Richert-Pöggeler 2006; Takahashi et al. 2019). However, a few examples of the activation of endogenous pararetroviruses, including tobacco vein clearing virus (TVCV) in tobacco (Nicotiana edwardsonii), banana streak virus (BSV) in banana (Musa balbisiana), and PVCV in petunia (Petunia hybrida), have been reported (Chabannes et al. 2013; Lockhart et al. 2000; Ndowora et al. 1999; Richert-Pöggeler et al. 2003).

It was reported two decades ago that PVCV is activated by environmental stresses such as heat, drought and wounding (Richert-Pöggeler et al. 2003; Zeidan et al. 2000). Although PVCV cannot be mechanically inoculated, it can be transmitted horizontally to other petunia and tobacco plants (N. glutinosa) by grafting, and exhibits symptoms such as vein clearing, leaf yellowing and distortion (Richert-Pöggeler et al. 2003). Tandemly inserted multiple copies of endogenous PVCV sequences are likely to be activated, and the QTR sequences can function like the LTR sequences at both ends of the coding sequence of PVCV (Richert-Pöggeler et al. 2003). In endogenous pararetroviruses, the endogenous state in the host genome is referred to as a provirus, and the circular DNA reverse-transcribed from a pregenomic RNA transcript is referred to as the episomal state (episomal DNA) (Supplementary Figure S1).

Recently we observed blotched flowers and a vein clearing symptom in aged petunia plants (Kuriyama et al. 2020). To determine the cause of blotched flowers in star-type petunia cultivars, in which white areas of bicolor petals are caused by post-transcriptional gene silencing (PTGS) of the key enzyme of anthocyanin biosynthesis, we focused on endogenous PVCV sequences, because many plant viruses encode for a suppressor of RNA silencing (VSR) (Burgyán and Havelda 2011). DNA methylation of CG and CHG sites in the promoter region of the QTR sequence decreased and PVCV transcripts and episomal DNAs accumulated in aged plants, indicating that poor maintenance of DNA methylation activates PVCV. Moreover, the detection of VSR activity by PVCV indicated that the mechanism, namely the suppression of host PTGS by the VSR of PVCV, caused blotched flowers. In parallel, de novo DNA methylation of CHH sites in the promoter region of QTR by RNA-directed DNA methylation (RdDM), which is an RNA silencing mechanism, was also detected. Therefore, both activation and inactivation of an endogenous pararetrovirs were observed in petunia plants. The star-type petunia with PVCV is a unique pathosystem that allows visualization of the activation process of an endogenous pararetrovirus, via changes in petal colors, and the tug-of-war between a plant host (RNA silencing) and a pathogen (endogenous pararetrovirus).

Thus, we have reported on the activation mechanism of an endogenous pararetrovirus, but not on its replication and gene expression. In this study, we characterized the role of QTR sequences on the replication and gene expression of PVCV. The PVCV coding sequence with the QTR sequence derived from either proviral or episomal DNA was introduced into Nicotiana benthamiana leaves by agroinfiltration, and the activity of VSR. i.e., the inhibition of PTGS, was evaluated by the efficiency of gene expression as assessed by the intensity of green fluorescence. At the same time, the amount of episomal DNA detected by Southern hybridization was considered to reflect the replication efficiency of PVCV. These results shed light on the replication mechanism of endogenous pararetroviruses, which are components of the plant genome.

Materials and methods

Plant materials and growth conditions

Petunia (P. hybrida) and N. benthamiana plants were grown in pots in a room with a controlled environment under the following conditions: 40–50 µmol m⁻² s⁻¹, 16 h light and 8 h dark, 24°C. Seeds of N. benthamiana line 16c were kindly provided by Dr. David Baulcombe, the Sainsbury Laboratory, Norwich, UK.

Preparation of plasmids for agroinfiltration

Since endogenous proviral and episomal PVCV have slightly different nucleotide sequences relative to each other (Kuriyama et al. 2020; Richert-Pöggeler et al. 2003), DNA extracted from leaves of one-month-old (for endogenous provirus PVCV) and four-month-old (for episomal PVCV) petunia plants was used for cloning the PVCV sequence. PVCV-L (5′region), PVCV-S (middle region), and PVCV-R (3′region) fragments were amplified from these DNAs as templates by PrimeSTAR® Max DNA Polymerase (Takara, Kusatsu, Japan). The PVCV-S and PVCV-R fragments were mixed and cloned into the binary vector pRI201-AN (Takara), which contains the CaMV 35S promoter and the heat shock protein (HSP) terminator derived from Arabidopsis thaliana, and treated with HindIII and BamHI using the In-Fusion® HD Cloning kit (Takara). This plasmid that was obtained and the PVCV-L fragment were treated with SacI and NotI and cloned using DNA Ligation Kit Ver.2.1 (Takara). These were named PQOQ and EQOQ, respectively (Figure 1A). The PVCV-MinR fragment was amplified using the plasmid of PVCV-S+PVCV-R+pRI201-AN as a template, and cloned into pRI201-AN treated with HindIII and BamHI using the In-Fusion® HD Cloning kit. Then, the PVCV-MinL fragment was amplified by PCR using the DNA extracted from leaves of one- and four-month-old petunia plants as a template. The above two fragments were treated with SacI and NotI and cloned using ligase. These plasmids were named PHOH and EHOH, respectively (Figure 1A). The 5′QTR-PVCV fragment was amplified by PCR using PQOQ as a template, and cloned by the in-fusion method with pRI201-AN treated with HindIII and SalI. This plasmid was named QOT (Figure 1A). The PVCV-3′QTR fragment was amplified by PCR using PQOQ as a template, and cloned into pRI201-AN treated with NdeI and BamHI by the in-fusion method. This plasmid was named 35OQ (Figure 1A). The 5′QTR-PVCV and PVCV-3′QTR fragments, as well as the PVCV ORF fragment without QTR, were amplified by PCR using PQOQ as a template, and cloned into pRI201-AN treated with NdeI and SalI. These plasmids were named QO, OQ, and Only, respectively. The DNA fragment of the green fluorescent protein (GFP) gene was amplified from the plasmid pEGFP-1 (Takara), then cloned into pRI201-AN using the In-Fusion® HD Cloning kit. Nucleotide sequences of cloned PVCV fragments were verified. Structures and names of obtained plasmids are shown in Figure 1, and primers for PCR are listed in Supplementary Table S1.

Agroinfiltration

Binary plasmids were introduced into Rhizobium radiobacter (Agrobacterium tumefaciens) AGL1 (kindly provided by Dr. Yuri Munekage, Kanseigakuin University, Japan), which is a recA mutant strain, by the electroporation method using MicroPulser™ (Bio-Rad Laboratories, Hercules, CA, USA). RNA silencing-suppressor activities of proteins encoded by PVCV were investigated in wild type (WT) or GFP-expressing N. benthamiana line 16c by an Agrobacterium-mediated transient expression assay (agroinfiltration) (Yaegashi et al. 2007). Green fluorescence in leaves of N. benthamiana was measured by the Fluorecence Imaging System, FOBI (NeoScience, Seoul, South Korea).

Southern hybridization

Total genomic DNA was isolated from petunia and N. benthamiana leaves using the protocol previously described by Liu et al. (1995). Approximately 4 µg of purified DNA was separated by 1% agarose gel electrophoresis (AGE), and then stained with ethidium bromide. DNA fragments were transferred to a nylon membrane (Zeta-Probe, Bio-Rad Laboratories) by capillary transfer. DNA fragments of the PVCV ORF as probes were amplified by PCR, and then probes were made using the BcaBEST Labeling Kit (Takara) with [α-³²P] dCTP. PCR primers are listed in Supplementary Table S1. Hybridization was carried out in Perfect Hyb Plus hybridization buffer (Sigma-Aldrich, St. Louis, MO, USA) containing a ³²P-labeled DNA probe at 65°C for 6 to 16 h. Membranes were washed twice in 2×SSC (1×SSC, 0.15 M NaCl, 15 mM sodium citrate) with 0.1% SDS at 65°C for 15 min, washed twice again in 0.5×SSC with 0.1% SDS at 65°C for 15 min, and then analyzed with a Typhoon FLA 7000 image analyzer (GE Healthcare, Chicago, MI, USA) (Fukuhara et al. 2011).

Northern hybridization

Total RNA was isolated from agroinfiltrated spots of N. benthamiana leaves using the Trizol reagent following the manufacturer’s protocol (Thermo Fisher Scientific, Waltham, MA, USA). Approximately 5 µg of total RNA was electrophoresed by 1% denaturing AGE containing 5% formaldehyde with MOPS Buffer (20 mM MOPS, 5 mM sodium acetate, 1 mM EDTA, pH 7.0) and then stained with ethidium bromide. Separated RNAs were transferred to a nylon membrane (Zeta-Probe, Bio-Rad Laboratories) by capillary transfer. Preparation of ³²P-labeled DNA probe and hybridization were performed in a similar manner to Southern hybridization.

Quantitative real-time PCR

Total RNA was isolated from agroinfiltrated spots of N. benthamiana leaves using the Trizol reagent following the manufacturer’s protocol (Thermo Fisher Scientific), from which cDNAs were produced by the PrimeScript RT reagent Kit with gDNA Eraser (Takara). Quantitative real-time PCR (qPCR) was performed by the Thermal Cycler Dice Real Time System (Takara) with the GoTaq® qPCR Master Mix (Promega). Primers for qPCR were designed using the Primer3Plus program (http://www.bioinformatics.nl/cgi-bin/primer3plus/primer3plus.cgi/). Primers are listed in Supplementary Table S1, and the locations of primers for PVCV are indicated in Figure 1.

Results

The essential role of the QTR sequence for PVCV gene expression

We obtained two types of QTR sequences from the proviral PVCV sequence that exists in the petunia genome and the episomal PVCV sequence detected in PVCV-activated petunia plants (Kuriyama et al. 2020). These two QTR sequences have slightly different sequences, a state that is thought to be caused by a switch of the template during reverse transcription to produce episomal DNA from pregenomic RNA transcribed from a proviral sequence (Figure 1 and Supplementary Figure S1; Hohn and Rothnie 2013). The QTR sequence contains elements that are critical for transcription such as transcription activation, initiation and termination, and also for replication (reverse transcription) such as primer (tRNA) binding and template switch (Figure 1 and Supplementary Figure S1, Hohn and Rothnie 2013). We constructed two types of PVCV infectious clones containing a full-length coding region (a single ORF putatively encoding a polyprotein) with two kinds of complete (full-length) QTR sequences on both sides (proviral (PQOQ) and episomal (EQOQ) in Figure 1A). We also constructed two different PVCV infectious clones containing partial QTR sequences on both sides (PHOH and EHOH in Figure 1A), in which only the sequences necessary for transcription and replication remained, to investigate how these partial QTR sequences affect the replication and gene expression of PVCV. Furthermore, we created several PVCV plasmids with the 5′end QTR sequence replaced with an authentic promoter (the CaMV 35S promoter sequence; 35OQ in Figure 1A), the 3′end QTR sequence replaced with an authentic terminator (the Arabidopsis HSP terminator sequence; QOT in Figure 1A), and the ORF sequence with one QTR sequence on either side (QO and OQ in Figure 1A).

To characterize the role of the QTR sequence for PVCV gene expression, we analyzed the amount of GFP and PVCV transcripts (transcription) and VSR activity (transcription and translation) that is encoded in PVCV (Kuriyama et al. 2020). Using agroinfiltration, agrobacteria containing various PVCV plasmids (Figure 1A), together with the plasmid containing the constitutively-expressed GFP gene, were co-introduced into N. benthamiana 16c leaf cells over-expressing the GFP gene, and then the efficiency of PVCV gene expression was evaluated by VSR activity (green fluorescence intensity, Figure 2A, B) and the accumulation of GFP and PVCV transcripts (Figure 2C–E). Strong green fluorescence was observed in the PQOQ, EQOQ, QOT, and 35OQ regions and these intensities were comparable with that of the HC-Pro region, which is an authentic VSR encoded by potyviruses (Figure 2A, B). The GFP and PVCV transcripts highly accumulated in these four regions (Figure 2C–E). These results indicate that the complete QTR sequences from either proviral or episomal sequences on both sides of the ORF (PQOQ and EQOQ in Figure 2) are required to efficiently transcribe mRNAs and translate encoding proteins. Moreover, one of the QTR sequences can be exchanged by the promoter or terminator sequences derived from other organisms (QOT and 35OQ in Figure 2), indicating that the QTR sequence upstream of the ORF sequence functions as a strong promoter and that the QTR sequence downstream of the ORF functions as a functional terminator. Similar relative abundance of the PVCV transcript was detected in the EHOH and QO regions by qPCR (Figure 2D) but the faint signal of the full-length transcript was detected in the QO region by Northern hybridization (Figure 2E), indicating that the partial QTR sequence (H) functioned as the poly A signal to stabilize the full-length PVCV transcript.

Figure 2. Role of the QTR sequence for PVCV gene expression. (A) Photograph of a N. benthamiana leaf over-expressing the GFP gene (16c line), into which agrobacteria containing various plasmids were infiltrated. Green fluorescence was observed at 14 days post inoculation (dpi). The schematic drawing shows plasmids contained by agrobacteria. Two kinds of agrobacteria, one containing various PVCV plasmids indicated in the right drawing and another containing a GFP over-expressing plasmid, were co-infiltrated except for PQOQ (GFP-), which was one agrobacterium containing PQOQ that was infiltrated. HC-Pro was used as the positive control of VSR. EV indicates an empty vector. (B) The intensity of green fluorescence in each infiltration spot was calculated. The intensity of EV was set to 1 as the standard. Bars indicate ±standard errors (SE) of three biological replicates. Significant differences were analyzed by Tukey’s test, and are indicated by different letters (p<0.01). (C) and (D) Relative transcript abundance of GFP and PVCV genes in each infiltrated spot at 7 dpi was measured by qPCR. Data from mock and EV spots were obtained from six biological replicates (three biological replicates for the others). The actin gene was used as the internal standard. The amount of 35OT was set to 1 as the standard. Bars indicate ±SE, and significant differences (Tukey’s test) are indicated by different letters (p<0.05). (E) Detection of the PVCV full-length transcripts by Northern hybridization. Total RNA was isolated from infiltrated spots of leaves at 7 dpi, separated by 1% denaturing agarose gel electrophoresis (AGE) containing 5% formaldehyde, and then transferred to a nylon membrane. The PVCV transcripts were detected by a ³²P-labeled probe that is indicated in Figure 1B. Autoradiographies with long exposure (LE) or short exposure (SE) are shown. Total RNA isolated from leaves of PVCV-infected N. benthamiana plant by grafting was used as the positive control (PC). + and − indicate that transcripts may contain the terminal redundant sequence or not. 28S ribosomal RNA is shown as a loading control.

In order to more accurately evaluate the promoter activity and the terminator activity of PVCV QTR, four T-DNA plasmids in which PVCV QTR was linked with the GFP gene were prepared, and the expression levels of GFP in agroinfiltrated spots were evaluated (Figure 3). In this experiment, in order to maintain high GFP expression, co-infiltrations were performed by using agrobacteria carrying a T-DNA plasmid containing VSR (HC-Pro or PQOQ). The results indicated that the terminator activity of QTR was similar to that of the authentic terminator (compare 35GT with 35GQ in Figure 3B, C), although the promoter activity of QTR was weaker than that of the authentic strong promoter, CaMV 35S (compare 35GT with QGT in Figure 3B, C). This suggests that the promoter activity of PVCV QTR cannot be simply compared with the promoter activity of CaMV 35S, because it has been reported that various transcriptional cis-elements are present in the non-coding region of the pararetrovirus genome (Gupta et al. 2021).

Figure 3. Comparison of the promoter and terminator activities of QTR with those of an authentic promoter and terminator. (A) Schematic drawings of T-DNA regions of plasmids used in agroinfiltration experiments. The commercially available plasmid, pRI201-AN, was used for plasmid construction, and the GFP gene was used as the reporter gene. P_35S and T_HSP indicate the CaMV 35S promoter and the Arabidopsis HSP terminator, respectively, as an authentic promoter and terminator. (B) Photographs of a wild-type N. benthamiana leaf, into which two kinds of agrobacteria, one containing one of four GFP plasmids shown in (A) and another containing the HC-Pro expressing plasmid or the PVCV expressing plasmid (PQOQ), were co-infiltrated. Green fluorescence was observed at 7 days post inoculation (dpi). (C) Relative abundance of GFP transcript in each infiltrated spot at 7 dpi was measured by qPCR. The actin gene was used as the internal standard, and data were obtained from three biological replicates. The amount of GFP transcript in 35GT co-infiltrated with HC-Pro was set to 1 as the standard. Bars indicate ±SE, and significant differences (Tukey’s test) are indicated by different letters (p<0.01).

The essential role of the QTR sequence for PVCV replication

To evaluate the role of the QTR sequence during PVCV replication, we attempted to quantitatively detect episomal DNA from infiltrated regions and its outside regions in N. benthamiana leaves by Southern hybridization, because episomal DNA is a hallmark of PVCV replication, and is produced from pregenomic RNA via reverse-transcription by its own reverse transcriptase (Supplementary Figure S1; Hohn and Rothnie 2013). Large amounts of three putative topological isomers of PVCV episomal DNA (indicated by three black triangles in Figure 4A, B) were detected at the infiltrated sites of PQOQ and EQOQ. These two PVCV constructs have the complete QTR sequences on both sides of the ORF sequence (Figure 1). Small amounts of episomal DNA were also detected from the infiltrated sites of PHOH and EHOH (Figure 4A, B), which have partial QTR sequences on both sides of the ORF (Figure 1). Consistent with these results, the full-length transcripts of PVCV were detected from agroinfiltrated spots of PQOQ, EQOQ, PHOH and EHOH (lanes PQOQ, EQOQ, PHOH and EHOH in Figure 2E), and the signal intensities of full-length RNAs in these four lanes in Figure 2E are correlated with those of episomal DNAs in the equivalent four lanes in Figure 4A. However, no apparent signals derived from PVCV episomal DNA were detected from other infiltration sites (Figure 4A), although high molecular-weight signals, which were probably derived from N. bentamiana genomic DNAs containing integrated T-DNA with a PVCV sequence (white triangles in Figure 4A, B). These results indicate that the complete QTR sequences from either proviral or episomal sequence on both sides are essential for the replication of PVCV (i.e., synthesis of episomal DNA). The replacement of QTR into the other promoter or terminator sequences on either side of the PVCV coding region is possible for gene expression (Figures 2 and 3) but impossible for replication (Figure 4). Therefore, we successfully detected episomal PVCV DNA, which was probably reverse-transcribed from the full-length transcripts with terminal redundant sequences at both ends, by agroinfiltration.

Figure 4. Role of the QTR sequence for PVCV replication. (A) Replication of PVCV, namely, production of PVCV episomal DNA, in agroinfiltrated spots were analyzed by Southern hybridization. DNA was isolated from infiltrated spots of N. benthamiana leaves at 14 dpi. DNA isolated from leaves of PVCV-infected N. benthamiana plant by grafting was used as a positive control (PC), and the mixture of purified plasmid PQOQ with N. benthamiana genomic DNA (indicated as a plasmid) was used to determine the migration of plasmid PQOQ (17 kbp). PVCV DNA was detected by a ³²P-labeled DNA probe indicated in Figure 1B. Genomic DNA stained by ethidium bromide is shown as a loading control. Three putative topological isomers (linear, open and closed circular) of PVCV episomal DNA (7.2 kbp) are indicated by black triangles, and the T-DNA plasmids and genomic DNA are indicated by white triangles. (B) Migration of PVCV from agroinfiltrated spots was analyzed by Southern hybridization. Agroinfiltration was performed with one spot per N. benthamiana leaf. Lanes In, Out and Sys indicate an infiltrated spot (inside), inoculated leaf excluding an infiltrated spot (outside), and an upper (systemic) leaf, respectively. (C) Confirmation of episomal DNA production in agroinfiltrated spots were analyzed by Southern hybridization. To exclude the possibility that the signals of 5 and 7 kbp shown by black triangles in (A) and (B), which we expected as circular and linear PVCV episomal DNA, respectively, would be smaller plasmids caused by intramolecular homologous recombination of the T-DNA plasmids, we attempted to detect GFP DNA in agroinfiltrated spots, into which agrobacteria containing the T-DNA plasmid (QGQ, QTR-GFP-QTR) were infiltrated, by Southern hybridization. Agroinfiltration experiments using one Agrobacterium containing one of four GFP plasmids shown in Figure 3A or co-infiltration experiments using Agrobacterium containing one of four GFP plasmids with Agrobacterium containing the PVCV-expressing plasmid (PQOQ) were performed, and DNA was isolated from infiltrated spots at 14 dpi. The mixture of the purified plasmid QGQ with N. benthamiana genomic DNA (indicated as a plasmid) was used to determine the migration of the plasmid QGQ (11 kbp). GFP DNA was detected by a ³²P-labeled DNA probe. Genomic DNA stained by ethidium bromide is shown as a loading control. The putative GFP episomal DNA (approximately 1.4 kbp) is indicated by a black triangle, and the T-DNA plasmids including QGQ (11 kbp) and genomic DNA are indicated by white triangles. M indicates a molecular weight marker.

The PVCV clones with the complete QTR sequence on both sides are able to replicate (In lanes of PQOQ and EQOQ in Figure 4B) but are unable to move outside of the infiltrated regions in inoculated leaves and systemic leaves (Out and Sys lanes of PQOQ and EQOQ in Figure 4B). In other words, these artificial PVCV clones can efficiently propagate only in cells that contain the PVCV coding sequence with the complete QTR sequences on both sides in their genomic DNAs, which is a similar situation to the tandemly arranged endogenous PVCV sequences in petunia plants. However, their episomal DNAs (with putative viral coat proteins) cannot move to neighboring cells containing PVCV-free genomic DNAs.

Although we expected that the 5-kbp signals indicated by a black triangle in Figure 4A and 4B were closed circular episomal DNA of PVCV, we have to exclude the possibility that these signals are circular DNA molecules caused by an intramolecular homologous recombination of the T-DNA plasmid during T-DNA plasmid replication in agrobacterium cells or T-DNA integration in plant cells. To exclude this possibility and to confirm the production of PVCV episomal DNA in host plant cells in agroinfiltrated spots, we attempted to detect DNA molecules created by an intramolecular homologous recombination of the T-DNA plasmid (QGQ, QTR-GFP-QTR), which has the QTR sequence on both sides of the GFP gene and a similar structure to the plasmids PQOQ and EQOQ (QTR-ORF-QTR), in agroinfiltration spots, by Southern hybridization (Figure 4C). No signals around 5 kbp were detected, indicating that the possibility of intramolecular homologous recombination can be excluded. Since A. tumefaciens AGL1, which was used for agroinfiltration is a recA mutant strain (see Materials and methods), homologous recombination is probably rare in this strain. The approximately 1.5-kbp signal, namely putative episomal DNA, was detected only in the co-infiltrated spot of QGQ and PQOQ, indicating that functional PVCV proteins expressed from the T-DNA containing PQOQ likely reverse-transcribed DNA (synthesized episomal DNA containing the GFP sequence) from RNA transcribed from QTR-GFP-QTR as well as its own transcript from QTR-PVCV-QTR as a template. Namely, it is very interesting that PVCV proteins can recognize any transcripts transcribed from DNA fragments containing complete QTR sequences at both ends as active templates for reverse-transcription.

Collectively, we conclude that the 5-kbp signal shown in Figure 4A, B is a closed circular form of PVCV episomal DNA that was reverse-transcribed from the full-length transcripts with terminal redundant sequences at both ends as templates. We were able to detect the replication process of PVCV in N. benthamiana leaves by agroinfiltration. The result of agroinfiltration experiments, summarized in Figure 5, revealed the role of the QTR sequence in replication and gene expression of PVCV.

Figure 5. Schematic drawings of the results of agroinfiltration experiments. When the QTR sequences are present on both sides of the PVCV ORF (PQOQ and EQOQ), the replication (episomal DNA synthesis) and gene expression of PVCV (synthesis of PVCV proteins including VSR) were efficient. When partial QTR sequences are present on both sides of the PVCV ORF (PHOH and EHOH), replication was less efficient. Replication was not detected in the other constructs. Replacement of the QTR with another authentic promoter and/or terminator is possible for gene expression. Complete QTR sequences on both sides of the ORF is essential for the efficient replication and gene expression of PVCV.

Discussion

In this study, we established an experimental system based on the agroinfiltration of N. benthamiana leaves to evaluate the gene expression (transcription and translation) of PVCV as VSR activity by measuring the intensity of green fluorescence of GFP (Figures 2 and 3), and also to assess the replication efficiency of PVCV by detecting and quantifying PVCV episomal DNA by Southern hybridization (Figure 4). This system is similar to an experimental system using cDNA infectious clones of RNA viruses, which is a powerful tool for characterizing the replication of plant viruses. It has previously been reported that PVCV was horizontally transmitted by biolistic inoculation (particle bombardment) and grafting (Richert-Pöggeler et al. 2003). However, the biolistic inoculation method requires expensive equipment and grafting takes a long time to observe virus propagation. By using agroinfiltration it is easy to inoculate the PVCV infectious clone, and its replication can be evaluated about 1 to 2 weeks after infiltration (Figure 4).

By using this method, we characterized the essential role of the QTR sequence in gene expression and replication of the plant endogenous pararetrovirus PVCV. The results shown in Figures 2 and 3 indicate that two complete QTR sequences on both sides of the coding region of PVCV are necessary for both gene expression and replication. Consequently, when one PVCV unit, which consists of one QTR and the coding sequence, becomes a circular episomal DNA, one QTR sequence is sufficient for efficient replication and transcription of PVCV. Furthermore, these results indicate that the placement of two or more proviral units in tandem is necessary for the activation (transcription) of PVCV when PVCV is presented as a provirus in the host nuclear genome. In other words, in this study, we determined the essential conditions for the activation of an endogenous pararetrovirus which had previously been speculated (Richert-Pöggeler et al. 2003). Our results also showed that the sequences essential for virus propagation are compactly packed in the QTR sequence, which is 666 bp long (Figure 1).

Two decades ago, it was reported that there is a slight difference in sequences between the QTR sequences derived from the proviral PVCV and the episomal DNA due to a template switch during reverse transcription to produce episomal DNA from the pregenomic RNA transcribed from the proviral sequence (Richert-Pöggeler et al. 2003). Actually, we recently detected two types of QTR sequences in a PVCV-activated petunia plant (Kuriyama et al. 2020). In this study, we assumed that these two types of QTR sequences might have different replication and/or transcription abilities, but no difference was detected by this agroinfiltration method (Figures 1–4), suggesting that the propagation (replication) of PVCV from circular episomal DNA is as efficient as that from the activated proviral PVCV sequences in the petunia genome. This speculation is supported by our previous results in which PVCV transcripts derived from the proviral PVCV sequence were detected at a similar abundance as those derived from episomal DNA in aged (4-months old) petunia plants by RNA-seq analysis (Kuriyama et al. 2020).

Although horizontal transmission and systemic infection of PVCV was observed by grafting and biolistic inoculation (Richert-Pöggeler et al. 2003), here we found the replication of PVCV by agroinfiltration but could not detect its systemic infection (Figure 4B). Since it has been reported that PVCV systemic infection was observed several months after viral inoculation by these two methods, a long maintenance period for host plants after inoculation may be necessary for PVCV systemic infection although N. benthamiana leaves usually become necrotic 20 days after agroinfiltration. Furthermore, as shown in Supplementary Figure S2, since the PVCV replication was mainly activated in a midvein tissue in N. benthamiana as well as petunia plants, agroinfiltrated regions in N. benthamiana leaves might not be suitable for detecting the cell-to-cell movement of PVCV.

Figure 4 shows very interesting results when considering the replication process of PVCV. More specifically, it was shown that the PVCV proteins function not only in cis but also in trans, recognize transcripts from DNA fragments containing complete QTR sequences at both ends as a template for reverse transcription, and reverse-transcribe DNA from them. This is similar to the transposition mechanism of transposable elements, that is, a transposase encoded by autonomous transposons recognizes terminal repeat sequences of both autonomous and non-autonomous transposons (Feschotte et al. 2002). However, in addition to terminal redundant sequences transcribed from the PVCV ORF with complete QTR sequences at both ends, the polypurine sequence in the ORF is essential for PVCV replication via reverse transcription (Supplementary Figure S1, Richert-Pöggeler and Shepherd 1997). Since the polypurine sequence does not exist in the GFP sequence, it is inferred that the putative episomal DNA containing the GFP sequence (Figure 4C) was a single-stranded DNA and had accumulated in the cytoplasm. Therefore, no RNA must be transcribed from this episomal DNA. In fact, the low amount of GFP mRNA accumulated in a co-agroinfiltrated spot with PQOQ and QGQ (Figure 3) supports this speculation. Since various T-DNA plasmids can be easily prepared for agroinfiltration, it is a powerful tool for analyzing the molecular mechanism of PVCV replication.

Conflict of interest

The authors declare that they have no conflicts of interest.

Author contribution

KK, MT, HM, HT and TF designed the research. KK performed all experiments. KK and TF wrote the paper. All authors read and approved the final manuscript.

Funding

This work was supported by the Ministry of Education, Culture, Sports, Science and Technology (MEXT) of Japan ([No. 16H06435, No. 16H06429 and 16H21723] to H.T. and T.F.), the Japan Society for the Promotion of Science (JSPS) ([No. 19K22304] to T.F., [No. 21J12088] to K.K., and [the JSPS Core-to-Core Program entitled “Establishment of international agricultural immunology research-core for a quantum improvement in food safety”] to H.T.), and the Global Innovation Research (GIR) Organization of Tokyo University of Agriculture and Technology (to M.T. and T.F.).