The N-terminal fragment of the tomato torrado virus RNA1-encoded polyprotein induces a hypersensitive response (HR)-like reaction in Nicotiana benthamiana

N. benthamiana plants (wild type and the 16c line, expressing the green fluorescent protein [GFP]) were grown under greenhouse conditions at 24-25 °C. An infectious clone of the tomato torrado virus Kra isolate [16] was used as the virus source for plant inoculation.

All primers used in this study are listed in Table 1. Plants were inoculated with ToTV, and from those showing disease symptoms, total RNA was isolated using TRI Reagent (Life Technologies) as described previously [19]. Up to 2 µg of the RNA was reverse transcribed using 100 ng of random hexamers and 200 U of RevertAid Reverse Transcriptase (Thermo Scientific). The cDNA was used for polymerase chain reaction (PCR) with primers flanking the sequence between the first ATG codon from ToTV RNA1 (the 5UTR/PrCo_F primer) and the first codon of the protease cofactor motif (the 5UTR/PrCo_R primer) within the same RNA1. After agarose gel electrophoresis (1 % agarose in TBE buffer), DNA fragments were extracted from the gel and cloned into two expression vectors: pBIN61 and pgR107 [20]. Briefly, the vectors were digested with SmaI and treated with alkaline phosphatase (Thermo Scientific). One hundred nanograms of the re-purified SmaI-digested pBIN61 was used for recombination-based cloning with 300 ng of PCR-amplified 11K sequence using an In-Fusion cloning Kit (Clontech).

The SmaI-digested pgR107 vector was ligated to the PCR-amplified 11K using T4 DNA ligase (Thermo Scientific). Both the recombination mixture and the ligation mixture were used to transform competent cells of E. coli strain TOP10 (Life Technologies). The recombined plasmid was extracted, its sequence was verified, and it was used for transformation of A. tumefaciens strains C58C1 (pBIN61-based constructs) or GV3101 (pgR107-based constructs).

A suspension of A. tumefaciens carrying either the pBIN61- or pgR107-based expression vector were grown at 28 °C for 24-48 h in LB medium with kanamycin and tetracycline. The bacteria were harvested and suspended in infiltration buffer (10 mM MES, pH 5.8, 0.5 μM acetosyringone, and 10 mM MgCl₂) and kept at room temperature for at least 2 h. The mixture was adjusted to an OD₆₀₀ of 1.0 and used to infiltrate two leaves of 6-week-old N. benthamiana seedlings, which were subsequently maintained under greenhouse conditions at 24-25 °C. Local protein expression was done using pBIN61-derived constructs, whereas both local and systemic expression of 11K was carried out using pgR107-11K (the PVX-based expression vector). For each infiltration, at least five plants were used, and the entire experiment was performed three times.

For the GFP-based silencing patch co-infiltration assay, the suspension of A. tumefaciens transformed with pBIN61-GFP (OD₆₀₀ 0.1) was mixed with Agrobacterium (OD₆₀₀ = 1.0) transformed with either pBIN61-11K, pBIN61-p19 or pBIN61 (empty vector) and co-infiltrated into leaves of N. benthamiana. Plants were maintained under greenhouse conditions at 24-25 °C. GFP fluorescence was excited using a hand-held UV lamp (UV Tech) and monitored 72 h after infiltration. Photographs were taken using a Nikon D5100 camera.

Total RNA was extracted as described previously [19]. For reverse transcription (RT) quantitative real-time PCR (RT-qPCR), 1 μg of total RNA was first treated with DNase (Thermo Scientific) and then subjected to cDNA synthesis. The resulting cDNA was used for RT-qPCR with the appropriate primers (Table 1) and iTaq™ universal SYBR® Green Supermix (Bio-Rad). The raw Ct data were normalised against actin or EF1α mRNA.

Total soluble proteins were isolated from infiltrated leaf patches using extraction buffer (1 M Tris-HCl, pH 7.5, and 20 % glycerol). The homogenate was centrifuged (16,000 × g, 30 min, 4 °C) to remove cell debris, and the protein concentration was measured using a Bio-Rad Protein Assay (Bio-Rad). An equal amount (50 μg) of the proteins was subjected to sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE). After electrophoresis, proteins were electroblotted onto a PVDF filter (Roche), and GFP was probed using primary monoclonal antibody (anti-GFP monoclonal antibody, Thermo Scientific). The GFP protein was visualised using Western Blue^® Stabilized Substrate for Alkaline Phosphatase (Promega).

Protein secondary structure was predicted using the GeneSilico Metaserver [21], and amino acid sequence alignments were done using ClustalW [22]. Sequence identities and similarities were calculated using BioEdit [23]. Protein sequence variability was assessed using the Protein Variability Server [24].

To investigate the biological function of the 11K domain, which is composed of 105 amino acid residues upstream of the protease cofactor (ProCo) motif encoded on ToTV RNA1 (Fig. 1B), the coding region between the 5’UTR and the ProCo motif was amplified by RT-PCR and was inserted into the pBIN61 vector. As shown in Fig. 2A, when this construct was used to infiltrate N. benthamiana plants, an HR-like reaction was induced 72 hours post-infiltration (hpi) to produce pBIN61-11K. In comparison, patches infiltrated with the empty pBIN61 vector did not develop an HR-like phenotype. To determine whether the 11K RNA was transcribed in leaf patches infiltrated with pBIN61-11K, total RNA was extracted from the infiltrated leaves, treated with DNase and used for RT-PCR. As expected, a ca. 350-nt 11K-specific amplicon was obtained only when RNA template extracted from HR expressing leaves was used. Neither the RNA isolated from non-infiltrated plants nor RNA extracted from patches infiltrated with pBIN61 gave the expected amplification product (Fig. 2B). This demonstrates that the 11K domain ectopically expressed in N. benthamiana induces a cell-death-like reaction in the host.

PVX expressed from the pgR107 plasmid induced mild disease symptoms in N. benthamiana, and in emerging leaves of the plants infected with pgR107, only a mild mosaic was observed (Fig. 3A). However, leaves infiltrated with pgR107-11K developed a cell-death phenotype that was already visible at 3 dpi. Symptoms were stronger than those displayed by plants expressing the 11K domain from pBIN61-11K. Such a phenotype was not observed in leaves infiltrated with either pgR107 or pgR107-11Kas (Fig. 3A).

In the course of time, PVX infection expanded toward new leaves. PVX alone induced a mild disease phenotype in upper non-infiltrated leaves. Similar symptoms were observed in plants infected with PVX-11Kas. To assess whether the plants were actually infected with PVX or PVX-11Kas, total RNA was extracted from upper non-infiltrated leaves and subjected to RT-PCR with 11K-specific primers. As expected, the ca. 350-nt amplicon was obtained from RNA isolated from PVX-11K- and PVX-11Kas-infected plants (Fig. 3B). However, only in plants infected with PVX-11K was strong systemic necrosis observed. Necrosis started to develop around the veins and proceeded to the entire leaf. The reaction was so strong that plants could not develop new apical leaves and thus remained stunted (Fig. 3A).

Next, we investigated whether the severe symptoms induced in PVX-11K-infected N. benthamiana were due to enhanced replication and therefore to an increase in accumulation of PVX-11K in infected tissues. For this purpose, total RNA was extracted from upper non-infiltrated leaves and used for RT-qPCR. The expression of three PVX genes (RdRP, 25K, and CP) was measured and normalized against the expression level of actin or EF1 α genes. The results indicated that the levels of accumulated genomic (the RdRP gene) and subgenomic (25K and CP genes) PVX RNAs were comparable in plants infected with PVX, PVX-11K and PVX-11Kas (Fig. 4). In detail, the level of RdRP RNA decreased ca. 1.6-fold in plants infected with PVX-11K, whereas it increased slightly (ca. 1.2-fold) in those infected with PVX-11Kas. The expression level of 25K did not change considerably in PVX-11K-infected plants, but CP gene expression increased 1.4-fold in plants infected with PVX-11Kas, and it did not change in plants infected with PVX-11K (Fig. 4). This suggests that the severe disease symptoms observed in PVX-11K infection were not caused by a dramatic increase in PVX RNA expression but instead were induced by the biologically active 11K domain.

Thomas et al. [25] reported previously that the P38 protein (coat protein) from turnip crinkle virus induces severe necrosis when ectopically expressed from a PVX vector. In the same study, the authors showed a possible role of P38 as a suppressor of RNA silencing. To verify the hypothesis that the disease symptoms induced in N. benthamiana upon 11K expression were a consequence of suppression of PTGS pathways, a co-infiltration silencing assay was performed. Transient expression of GFP in N. benthamiana proceeded until 3 dpi and then decreased as a result of PTGS. However, in the presence of a PTGS suppressor, expression of the reporter gene was maintained at a high level. Indeed, at 3 dpi, strong GFP fluorescence was observed in leaf patches infiltrated with pBIN61-GFP together with pBIN61-p19 (a silencing suppressor protein of tomato bushy stunt virus [26]) (Fig. 5A). In comparison, GFP fluorescence was hardly observed in leaves infiltrated with pBIN61. Interestingly, in leaves co-expressing GFP and 11K, a brightening was observed (Fig. 5A). To determine whether the effect was caused by enhanced expression of GFP in infiltrated leaves, total soluble protein extracts were prepared from the GFP-expressing patches and were subjected to SDS-PAGE followed by western blot analysis with GFP-specific monoclonal antibodies. High GFP expression was confirmed in leaves co-infiltrated with pBIN61-GFP and pBIN61-p19 (Fig. 5B). However, in leaves expressing GFP together with pBIN61 or pBIN61-11K, the level of GFP was substantially lower. This suggests that the brightening in leaves infiltrated with pBIN61-11K was not due to increased GFP expression. This was also confirmed by comparing GFP mRNA levels. Relative RT-qPCR analysis showed that GFP mRNA accumulated to the highest level only when co-expressed with the p19 silencing suppressor (Fig. 5C).

The consensus prediction of putative secondary structure motifs in the 11K domain (based on eighteen predictors implemented in the GeneSilico Metaserver) indicated the potential to form two α-helical motifs. Importantly, no β-sheets were predicted to be formed within the 11K domain. Moreover, the predicted helical motifs consisted of two separate domains: helix 1 between amino acids 24 and 56, and helix 2 between amino acids 63 and 105 (Fig. 6A). The same analysis was performed for amino acid sequences of the corresponding regions of tomato marchitez virus, tomato necrotic dwarf virus, tomato chocolate virus and tomato chocolate spot virus. The two predicted helical motifs were found in all of the aforementioned viruses, at almost the same positions in their corresponding 11K sequences (Fig. 6A). Interestingly, those two predicted helices were composed of amino acids with similar properties (Fig. 6B), and although the RNA sequence differs substantially between the analogous regions of other torradoviruses, the predicted secondary structures in their 11K domains are highly conserved (Fig. 6B and C). Moreover, the proposed helix-building amino acids of the 11K domain had the lowest variability score (Fig. 6D). This confirmed that although the two helices in the 11K domain are sequence-defined, the domains have a conserved structure rather than a conserved sequence.