The MoClo system, based on Golden Gate cloning technology, has been used for viral gene cloning and binary vector construction for plant transformation (Weber et al., 2011; Engler et al., 2014; Marillonnet and Grützner, 2020). Briefly, restriction-ligations were set up in one tube containing approximately 20 fmol of each plasmid DNA, 10 U of the required type IIS restriction enzymes, such as BsaI and BpiI (New England BioLabs, Frankfurt, Germany), 10 U T4 DNA ligase, and 1×T4 DNA ligase buffer (Fermentas, St. Leon-Rot, Germany) in a final reaction volume of 20 μL. The reaction was incubated in a thermocycler for 5 hours at 37°C, 5 min at 50°C and 10 min at 80°C. The mixture was transformed into competent DH5α cells using heat shock. Colonies were selected on LB plates containing appropriate antibiotics and extrapolated for the entire transformation.

Level 0 (L0) fragments were synthesized as linear fragments and cloned into the pBHK cloning vector (Bioneer, Daejeon, Republic of Korea). Adaptor sequences were introduced at both the 5′- and 3′- ends of the recognition site of BpiI or BsaI type IIS restriction enzyme as design changes prior to synthesis during initial assembly. For geminiviral-based deconstructed vector construction, viral sequences for L0 parts were based on the genome of the TYLCV Korean Isolate (2,774 bp, GenBank accession no. KF225312), the HYVV Korean Isolate (2,763 bp, GenBank accession no. GQ477135), and the BMCTV Worland isolate (2,930 bp, GenBank accession no. U56975). Overlapping complementary sense ORFs and the C1, C2, and C3 genes were modified and synthesized. For the replication-related gene expression vectors, the C3 start codon harbored in C12 (C1 and C2), and the C2 start codon harbored in C1 were modified to prevent spontaneous C3 and C2 truncation, respectively. The complete sequences are shown in Supplementary Figures S2 - S5 . ORFs of each synthesized replication-related genes in pBHK (L0), s35SP (L0; pICH41388) and NosT (L0; pICH41421) were inserted into Level 2 destination vector (pICH86969) using Golden Gate cloning method with BsaI to make binary vectors as follows: TYLCV-based replication-related gene vectors, pTC1, pTC12, pTC123, HYVV-based replication-related gene vectors, pHC1, pHC12, pHC123, and BMCTV-based replication-related gene vectors, pBC1, pBC12, and pBC123, respectively.

Plant codon-optimized Pontellina plumata-derived turbo GFP (tGFP, level 0; pICSL80005) was used as a reporter gene in this study. tGFP (L0; pICSL80005), d35SP, tobacco mosaic virus (TMV) 5′-leader sequence (Ω) (L0; pICH51288), and PinIIT (L0; pBHK) were inserted into Level 1-2 destination vector (L1-2; pICH47751) using Golden Gate cloning method with BsaI to construct expression vector pSPtGFP, i.e., the GFP expression vector. Each IR was cloned into L1-1 and L1-3 using PCR and specific primers ( Supplementary Table S1 ). Each IR (in L1-1 and L1-3), pSPtGFP (L1-2), and end linker L3E were inserted into the Level M1 destination vector (pAGM8031) using the Golden Gate cloning method with BpiI to construct IR-carrying tGFP expression vectors, i.e., TYLCV-based IR vector, pTIRtGFP; HYVV-based IR vector, pHIRtGFP; and BMCTV-based IR vector, pBIRtGFP. The pSPtGFP expression vector served as a nonreplicating control. A schematic diagram of the geminivirus-based deconstructed vectors is shown in Figure 1 .

Binary vectors were separately introduced into A. tumefaciens GV3101 by using 50–500 ng of recombinant plasmid in the freeze-thaw method, as described by Jyothishwaran et al. (2007). Recombinant strains were grown overnight at 28°C with agitation in LB medium supplemented with appropriate antibiotics for infiltration. Cells were pelleted using centrifugation 2,000 × g for 5 min, and then suspended in infiltration buffer [10 mM 2-(N-morpholino) ethanesulfonic acid (MES) (pH 5.6) and 10 mM MgCl₂ supplemented with 100 μM acetosyringone] and diluted in infiltration buffer to OD₆₀₀ = 0.5, unless otherwise described. When mixing the two constructs, each Agrobacterium concentration was set to OD₆₀₀ = 1.0, and mixed 1:1. For co-infiltration with tomato bushy stunt virus (TBSV) p19 (kindly provided by Prof. Inhwan Hwang, POSTECH, Republic of Korea), each was set to OD₆₀₀ = 1.0, and mixed 1:1:0.5. After incubation at 25°C for 3 h, cells were infiltrated into the leaves of 4–5 weeks old N. benthamiana plants grown in a hydroponic growing media at 23°C under a 16 h light and 8 h dark photoperiod at a light intensity of 100 mol m^-2 s^-1. The resulting bacterial suspensions were infiltrated using syringe without needles into fully expanded leaves through a small puncture (Huang et al., 2006).

Genomic DNA was isolated from plant tissue using the NucleoSpin Plant II (Machery-Nagel GmbH & Co, Düren, Germany) as per the manufacturer’s instructions. Genomic DNA was quantified on a NanoDrop spectrophotometer (Thermo Scientific, Waltham, MA, USA) and diluted to 50 ng/μL. PCR was performed in reaction volumes of 20 μL using 100 ng of genomic DNA with Ex-Taq DNA polymerase (TaKaRa, Shiga, Japan). The primers ( Supplementary Table S1 ) used in the PCR analysis were designed to amplify each replicon recirculation and C1, C12, and C123 in the co-infiltrated N. benthamiana genome. The PCR products were electrophoresed on 1.0% (w/v) agarose gels, stained with StaySafe Nucleic Acid Gel Stain (Real Biotech Corporation, Taipei, Taiwan), and visualized under ultraviolet light.

Quantitative PCR was performed using a CFX96 Touch Real-Time PCR Detection System (Bio-Rad, Hercules, CA, USA) real-time PCR machine together with AccuPower 2×GreenStar qPCR Master Mix (Bioneer). Triplicate reactions were performed, and tGFP gene copy numbers were normalized using β-actin of N. benthamiana (GenBank accession no. JQ256516) as the reference. PCR primer sequences are listed in Supplementary Table S1 .

Total RNAs was purified from infiltrated N. benthamiana leaves at 1, 3, 5, and 7 day post infiltration (DPI) using the Spectrum™ Plant Total RNA Kit (Sigma-Aldrich, St. Louis, MO, USA), and the residual DNA was removed with RNase-free DNase I (Takara, Shiga, Japan). First-strand cDNA was synthesized from 5 μg of total RNA and oligo dT primers using RNA-to-cDNA EcoDry Premix (Takara, Shiga, Japan) according to the manufacturer’s protocol. Quantitative real-time PCR analysis was performed with 100 ng of cDNA in a 20 μL reaction volume using AccuPower 2×GreenStar™ qPCR Master Mix (Bioneer, Daejeon, Republic of Koread). Gene-specific primers used are listed in Supplementary Table S1 . qRT-PCR was performed with an initial step at 95°C for 10 min followed by 40 cycles of 95°C for 20 s, 58°C for 20 s, and 72°C for 25 s. Fluorescence was recorded after the last step of every cycle. Three replicates were performed for each sample. Amplification, data processing, and detection were performed using the CFX96 Real-Time PCR Detection System (Bio-Rad, Hercules, CA, USA). Quantification cycle (Cq) values were examined using the 2^-ΔCT method to determine changes in gene expression.

Total soluble proteins were extracted by homogenizing agroinfiltrated leaf samples harvested 1, 3, 5, and 7 DPI with 1:2 (w/v) ice-cold extraction buffer [200 mM Tris-Cl (pH 7.0), 100 mM NaCl, 10 mM EDTA, 0.5% Triton X-100, and a protease inhibitor cocktail (Roche Diagnostics GmbH, Mannheim, Germany)]. Cleared supernatants were obtained by centrifugation at 15,000 × g for 15 min at 4°C. The protein concentration in the leaf samples was determined using the Bradford Protein Assay Reagent kit (Bio-Rad, Hercules, CA, USA) (Bradford, 1976) with bovine serum albumin as the reference standard.

The tGFP samples were prepared by dilution in extraction buffer and 100 μL of each sample was added to a 96-well plate (SPL, Republic of Korea) in duplicate. The tGFP fluorescence intensity was examined using a Victor³ microplate reader (PerkinElmer, Waltham, MA, USA). The excitation and emission wavelengths were 485 nm and 538 nm, respectively. All measurements were performed at 25°C and the negative control (extract of un-infiltrated plant leaves) was subtracted before graphing. E. coli-derived tGFP protein (Evrogen, Moscow, Russia) was used to generate a standard curve. Leaves expressing tGFP harvested at 1, 3, 5, and 7 DPI were viewed under UV illumination generated by a LAS4000 (Fujifilm, Tokyo, Japan).

Total soluble protein (25 μg) was mixed with sample buffer [10% glycerol, 60 mM Tris–HCl (pH 6.8), 2% SDS, 0.5M dithiothreitol, 0.01% bromophenol blue] at 25°C and were separated using 12% SDS-polyacrylamide gel electrophoresis and then electrophoretically transferred to a iBlot 2 polyvinylidene fluoride (PVDF) Regular Stacks (Invitrogen, Carlsbad, CA, USA) following the manufacturer’s instructions. The protein-transferred membranes were blocked with 5% non-fat skim milk in Tris-buffered saline containing 0.05% Tween-20 (TBST, pH 7.4) for 2 h at 25°C. tGFP protein was detected with mouse anti-tGFP monoclonal antibody (Origene, Rockville, MD, USA) at 1:5,000 dilutions. The secondary antibody used was alkaline phosphatase-conjugated goat anti-mouse IgG (Sigma-Aldrich, St. Louis, MO, USA) at 1:10,000 dilutions. The membranes were developed using nitro blue tetrazolium chloride and 5-bromo-4-chloro-3-indolyl phosphate (Sigma-Aldrich, St. Louis, MO, USA). Purified E. coli expressing tGFP protein (Evrogen, Moscow, Russia) was used as a positive control.

The amount of tGFP protein extracted from the agroinfiltrated leaf tissues was determined using indirect ELISA. Briefly, a 96-well Maxisorp microtiter plate (NUNC, Roskilde, Sjelland, Denmark) was coated with tGFP plant extracts in coating buffer (15 mM Na₂CO₃, 35 mM NaHCO₃, pH 9.6) at 4°C overnight, and washed with 200 μL of PBST (PBS with 0.1% Tween-20, pH 7.4) four times. The wells were then blocked with 1% BSA in PBS at 37°C for 2 h. After washing the wells three times, 100 μL of diluted anti-mouse tGFP monoclonal antibody (1:5,000) (Origene, Rockville, MD, USA) were added to the wells and incubated at 37°C for 2 h. After washing four times, anti-mouse IgG-conjugated horseradish peroxidase (1:10,000) (GenDepot, Baker, TX, USA) was added to each well and incubated at 37°C for 2 h. After five times washing, 100 μL of TMB (3,3′,5,5′-tetramethylbenzidine) containing H₂O₂ solution were added and the plate was incubated for 15 min at 25°C. The enzyme reaction was stopped by quickly pipetting 50 μL of 2 N H₃PO₄ into each well. The absorbance at 450 nm was measured using a microplate reader. To calculate the relative amount of tGFP in the plant sample, the OD value from each sample was subtracted from the untransformed plant OD value before converting by reference to an ELISA standard curve constructed with purified bacterial tGFP (Evrogen, Moscow, Russia). Plant samples were analyzed by diluting from 1:1,000 to 1:4,000 with a coating buffer.

We deconstructed TYLCV, HYVV, and BMCTV strains belonging to the Geminiviridae family to develop viral vector systems for the high expression of target proteins. These vectors were based on the replication machinery related to IR and Rep to induce viral replicons and the transcription unit of the target recombinant protein gene instead of the sequence-coding coat and movement proteins (Gleba et al., 2004). This system used two vectors containing different portions of the TYLCV, HYVV, or BMCTV genomes. The first vector carried the transcription unit of tGFP gene and the IRs of the TYLCV, HYVV, or BMCTV genome, which are essential for each episomal replication, incorporating at its 5′- and 3′-ends, and the second vector contained each replication-related gene.

We modified and constructed vectors using MoClo systems to investigate the efficient combination of each IR and replication-related gene for the amplification of episomal replicons and high protein expression. Each IR-carrying tGFP vector, i.e., pTIRtGFP, pHIRtGFP, and pBIRtGFP, was controlled by d35SP, Ω, and PinIIT, respectively. PinIIT is a strong terminator (An et al., 1989) and reports have shown that the use of PinIIT results in 10–50 times greater HBsAg accumulation and 8.5−fold higher GFP fluorescence than NosT (Richter et al., 2000; Diamos and Mason, 2018b). pSPtGFP was used as the negative control ( Figure 1 ). Each Rep-supplying vector, i.e., C1ΔC2 (named pTC1, pHC1, and pBC1), C1/C2/ΔC3 (named pTC12, pHC12, and pBC12), and C1/C2/C3 (named pTC123, pHC123, and pBC123), was controlled by s35SP and NosT. Agrobacterium strains carrying these constructs were co-infiltrated at the same ratio, and N. benthamiana leaves were then sampled in a time-dependent manner.

Following a previous report, genomic DNA was extracted from 3 DPI leaf samples (Regnard et al., 2010) to confirm the formation of episomal replicons containing three different IRs, respectively, in co-infiltrated leaves. The formation of episomal replicon from T-DNA is essential for the amplification using the primers oriented for both ends ( Figure 2A ). Approximately 1.2 kb PCR products were amplified an only from circularized unit-length replicons ( Figure 2B ). The results showed that replicon formation occurred in the co-infiltrated N. benthamiana leaves. We also tested whether the amplification of episomal DNA was associated with an increase in the tGFP copy number. Real-time PCR was used to determine the tGFP gene copy number in co-infiltrated N. benthamiana leaves with each replicon vector combination. In combination with TIRtGFP and replication-related genes, tGFP gene amplification levels were high, ranging from 433- to 701-times compared to the negative control ( Figure 2C ). In addition, in combination with HIRtGFP and replication-related genes, tGFP gene amplification levels increased 201- to 731-times compared with the negative control ( Figure 2D ). The data indicated that the highest tGFP amplification was achieved with the combination of TIRtGFP+TC123 or HIRtGFP+HC123, which was 701- to 731-times increase compared to that in the negative control. However, in the case of the combination of BIRtGFP and replication-related genes, the tGFP gene amplification level showed a 2- to 11-times increase compared to the negative control ( Figure 2E ). Co-infiltration with BIRtGFP+BC1 resulted in the highest copy number (11-times) of tGFP gene among the combinations of BIRtGFP and BMCTV replication-related genes, which was different from the other cases ( Figures 2C–E ).

The impact of tGFP gene amplification on transient expression was determined by comparing tGFP expression of the three geminivirus-based deconstructed vectors. We investigated the potential of various geminivirus-derived IR- and replication-related gene combinations to enhance tGFP mRNA expression and fluorescence. Plant extracts prepared from samples taken from agroinfiltrated N. benthamiana leaves over 7 d were analyzed using qRT-PCR and tGFP fluorescence.

The relative mRNA expression of tGFP in each vector combination peaked at 1 DPI, after which there was a steady decline. The mRNA expression patterns at 3 DPI varied depending on the presence or absence of geminivirus-derived replication-related factors. In the case of tGFP only, the relative mRNA expression decreased to 87% at 3 DPI, and IR-carrying tGFP only, TIRtGFP, HIRtGFP, and BIRtGFP also sharply declined to 96.3%, 94.5%, and 88.2%, respectively, at 3 DPI ( Figures 3A–C ). However, when co-expressed with each IR and replication-related gene, relative mRNA expression was maintained at 47.8, 38.8, and 50% in TIRtGFP+TC123, TC12, or TC1 ( Figure 3A ) and 40.5, 51.6, and 63.1% in HIRtGFP+HC123, HC12, and HC1 ( Figure 3B ), respectively, compared to that of the negative control at 3 DPI. In contrast, BMCTV-related combinations showed lower mRNA expression in all combinations compared to the negative control at 1 DPI and dropped quickly as in the controls, with the exception of BIRtGFP+BC1, which was maintained at 99.8% at 3 DPI ( Figure 3C ).

The expression levels among vector combinations were further compared by determining the fluorescence intensity of the crude extract of agroinfiltrated N. benthamiana leaf tissues transiently expressing tGFP with various vector combinations. Unlike the maximum mRNA expression at 1 DPI, no tGFP fluorescence was detected, and the highest fluorescence was quantified at 3 DPI for all combinations, after which it decreased. Leaf tissues co-infiltrated with TIRtGFP+TC123, HIRtGFP+HC123, or BIRtGFP+BC1 showed the highest fluorescence intensity of vector combinations with high mRNA expression ( Figure 3 ). The relative fluorescence intensities of TIRtGFP+TC123, HIRtGFP+HC123, and BIRtGFP+BC1 at 3 DPI increased 31-, 22-, and 11-fold, respectively, compared to the negative control on 7 DPI ( Figures 3D–F ). When tGFP was expressed from TIRtGFP+TC123 at 3 DPI, higher mRNA expression and fluorescence intensity were observed compared to those of the other combinations and other DPIs ( Figures 3A, D ). These results demonstrated a similar correlation between replicon amplification and tGFP expression in various vector combinations, strongly suggesting that higher gene expression is due to the high copy number of replicons.

Each single vector or vector combination was infiltrated into a single leaf to determine the fluorescence intensity and phenotypic changes in tGFP in the infiltrated N. benthamiana leaf tissues. An illustration of the approach used to directly compare tGFP expression between different vector combinations are presented in Figure 3G . Images are representative of tGFP fluorescence intensity and photographs of infiltrated leaves in TYLCV, HYVV, and BMCTV at 3, 5, and 7 DPI under UV and visible light ( Figures 3H–J ). The green fluorescence of the infiltrated areas at 3–7 DPI was brighter than that of the negative control under UV light ( Figures 3H–J ). Severe tissue necrosis was not observed under visible light in agroinfiltrated N. benthamiana leaves expressing tGFP with a combination of various geminivirus-derived IR and replication-related genes; however, inspection of leaves under visible light revealed that fluorescing leaf regions infiltrated with vector combinations were gradually brighter and curlier than those of the negative control as DPI increased ( Figures 3H–J ).

We chose three deconstructed geminiviral vector combinations (TIRtGFP+TC123, HIRtGFP+HC123, and BIRtGFP+BC1) for high production of tGFP. Bottlenecks in gene silencing exist during the transcription in transient expression systems. TBSV p19 co-infiltrated to enhance transcript levels and protein accumulation in N. benthamiana leaf tissues by suppressing gene silencing (Qiu et al., 2002). We co-infiltrated a p19 vector with the three most effective combinations based on our data to test whether the expression of p19 in our replicon system can further elevate tGFP mRNA levels and fluorescence intensity. We found that co-expression with p19 resulted in a shift in maximal mRNA expression from 1 to 3 DPI. The highest tGFP relative mRNA expression reached 529-, 749-, 714-, and 726-times (increased 19.6-, 27.7-, 26.4-, and 26.9-times compared to that of tGFP at 3 DPI, respectively) in the tGFP+p19, TIRtGFP+TC123+p19, HIRtGFP+HC123+p19, and BIRtGFP+BC1+p19 combinations compared to the negative control at 7 DPI, respectively, and then gradually decreased over time in all three combinations ( Figure 4A ).

The fluorescence intensity in tGFP+p19, TIRtGFP+TC123+p19, and HIRtGFP+HC123+p19 reached the highest levels (2-, 7.1-, and 6.4-times, respectively) at 3 DPI, whereas that of BIRtGFP+BC1+p19 reached the highest level (18.5-times) at 5 DPI compared to the negative control of each DPI ( Figure 4B ). It gradually decreased, as did the mRNA expression for all combinations except tGFP+p19, which increased over time ( Figure 4B ). The fluorescence intensity under UV light was much stronger in the leaf spots infiltrated with p19 than in the leaf spots infiltrated without p19 ( Figures 4C–F ).

The impact of transient protein production was determined using SDS-PAGE, western blot analysis, and enzyme-linked immunosorbent assay (ELISA). Plant extracts were prepared from agro-infiltrated N. benthamiana leaves at 3 DPI for determining tGFP protein production. The E. coli-derived tGFP as a positive control was observed at 27 kDa and β-actin was used as an internal loading control. In total, 25 μg of leaf total soluble protein (TSPs) from N. benthamiana leaf tissue samples at 3 DPI were separated using SDS-PAGE and stained with Coomassie Brilliant Blue. As expected, SDS-PAGE gels showed the three major protein bands in the mass range of 14–45 kDa, i.e., ribulose 1,5-bisphosphate carboxylase oxygenase small (rbcS) and large (rbcL) subunits (Ellis, 1979). In addition, a 27 kDa recombinant tGFP protein were weakly and strongly expressed in agroinfiltrated N. benthamiana leaves in all three geminiviral vector combinations without p19 and with p19, respectively ( Figures 5A, C ). A faint 27 kDa tGFP band was detected in tGFP-only and HIRtGFP+HC123 combination, and distinct in TIRtGFP+TC123 combination. However, this was not observed in the BIRtGFP+BC1 combination ( Figure 5A ). In the case of p19 co-expression, TIRtGFP+TC123 and HIRtGFP+HC123 showed stronger tGFP bands compared to that of tGFP-only and BIRtGFP+BC1 ( Figure 5C ). Western blot analysis using a mouse monoclonal tGFP antibody at 3 DPI revealed that tGFP expression with p19 was higher than that without p19 in the tGFP-only, TIRtGFP+TC123, HIRtGFP+HC123, and BIRtGFP+BC1 combinations under reducing conditions ( Figures 5B, D ). The production of tGFP in co-infiltrated N. benthamiana leaves was estimated using ELISA ( Figure 5E ). When co-expressed with p19, tGFP protein production increased 4.8-, 3.4-, 3.9-, and 11.5-times compared to tGFP alone, TIRtGFP+TC123, HIRtGFP+HC123, and BIRtGFP+BC1 combinations, respectively, reaching the highest levels of 0.1, 1.1, 1.0 and 0.7 mg/g fresh weight (FW), corresponding to 1.9, 12.1, 10.6, and 6.9% TSP, respectively ( Figures 5E, F ). These results indicated that p19 can increase tGFP mRNA and protein accumulation, most likely by suppressing post-transcriptional silencing of the transgene in agroinfiltrated N. benthamiana leaves.