Plants of N. tabacum cv. Samsun NN, N. benthamiana, cucumber, and tomato were grown in an air-conditioned greenhouse under natural light at 25°C/20°C (day/night). Inocula of CMV-Y virions and in vitro transcripts of D3Yα were prepared as reported by Takeshita et al. (2008). Absence of D3Yα in the purified CMV-Y virions was certified using 3a-ORF-specific reverse transcription polymerase chain reaction (RT-PCR) described by Takeshita et al. (2008).

To examine the adaptability of D3Yα on different host species, we used cucumber and tomato for inoculation assay (Table 1). For cucumber and tomato plants, when the fourth true leaf stage to expand, the second true leaf was inoculated with purified virus particles of CMV-Y (1 mg/ml) or with those of CMV-Y (1 mg/ml) plus in vitro transcripts of D3Yα. We inoculated three cucumber and 9 to 12 tomato plants in the first assay, and repeated the second assay by using different three cucumber and 7 to 8 tomato plants with the same inocula as described above (Table 1). At 28 days post inoculation (dpi), the upper, non-inoculated leaves were collected from cucumber and tomato plants and tested by RT-PCR as described by Takeshita et al. (2008).

To monitor the D3Yα generation through serial inoculation on N. benthamiana or N. tabacum, we mechanically inoculated the plants with in vitro transcripts of cDNA clones of CMV-Y RNAs1, 2, 3 as reported by Suzuki et al. (1991). As Kaplan et al. (2004) reported that N. tabacum cv. Samsun NN which was inoculated with in vitro transcripts of cDNA of CMV RNAs does not generate D RNA3, we used in vitro transcripts for the first inoculation. Those inoculated plants were designated as the first passage in Fig. 1A, and we collected upper, non-inoculated leaves at 10 dpi for sap inoculation in the second passage, and at 10 and 40 dpi for 3a-ORF-specific RT-PCR. For sap inoculation in the third passage, we further collected upper, non-inoculated leaves at 14 dpi in the second passage. For 3a-ORF-specific RT-PCR, we sampled upper, non-inoculated leaves at 14, 30, 90, and 180 dpi in the second and 14 and 30 dpi in the third passage (Fig. 1B and C).

To investigate the involvement of temperature in generation or retention of D3Yα, N. benthamiana plants were inoculated with D3Yα-free purified CMV-Y virions or CMV-Y virions containing D3Yα, then those plants were kept at 15°C or 36°C till sampling at 7 and 28 dpi. The inoculum was prepared using tobacco plants which were inoculated with D3Yα-free CMV-Y virions (in the first passage as described above) at 7 to 10 dpi for checking the generation of D3Yα (Fig. 2A). Also, we prepared sap from N. tabacum plants that were inoculated with CMV-Y virions (1 mg/ml) plus in vitro transcripts of D3Yα for the retention of D3Yα at 15°C or 36°C (designated a positive control in Fig. 2A). We inoculated the second passage plants with sap from upper, non-inoculated leaves of the infected N. tabacum plants in the first passage. For 3a-ORF-specific RT-PCR analysis, upper, non-inoculated leaves of the second passage kept at 15°C were sampled at 28 dpi while we collected upper, non-inoculated leaves of the plants kept at 36°C at 7 dpi because the plants could not survive beyond 10 dpi.

To examine whether the 2b protein is necessary for accumulating D3Yα, N. benthamiana plants grown at 25°C/20°C (day/night) were inoculated with CMV-Y or two CMV-Y deletion mutants (C2-A1 and C2-H1) as described by Takeshita et al. (2012). C2-A1 vector (ΔY2b-C in Fig. 2B) truncated C terminus of the 2b protein, and C2-H1 (ΔY2b in Fig. 2B) has the entire region of the 2b protein deleted including ORF region which partially overlaps with C terminus of the 2a protein (Supplementary Fig. 1). For the first passage, each in vitro transcript using cDNA clone of CMV-Y RNA1,3 and C2-A1/C2-H1 vector were prepared as described by Takeshita et al. (2012). Then, upper, non-inoculated leaf tissues in the first passage were collected at 10 dpi for sap inoculation in the second passage. We sampled upper, non-inoculated leaves in the second passage at 40 dpi for 3a-ORF-specific RT-PCR analysis.

Total RNA was extracted from leaves using RNAiso Plus (TaKaRa Bio, Kusatsu, Japan) according to the manufacturer’s protocol. For cDNA synthesis, total RNA was diluted to 50-300 ng/μl and stored at −80°C until use. cDNA was synthesized using ReverTra Ace qPCR RT Master Mix with gDNA Remover (TOYOBO, Osaka, Japan) according to the manufacturer’s protocol. Two microliters of the diluted total RNA was added to 4.0 μl of deionized distilled water, then incubated at 65°C for 5 min. After 2.0 μl of 4× DN Master Mix with gDNA Remover was added, tubes were incubated at 37°C for 5 min. Finally, 2.0 μl of 5× RT Master Mix II was added, and samples were incubated at 37°C for 15 min for reverse transcription, then at 95°C for 5 min to inactivate the reverse transcriptase. The synthesized cDNA was stored at −30°C until use. For the PCR, 2 μl of cDNA mixture was used in a PCR mixture containing 5.0 μl of 2× PCR buffer, 2.0 μl of 2 mM dNTPs, 0.3 μl of 10 μM forward and reverse primers, 0.2 μl of deionized distilled water, and 0.2 μl of KOD Fx Neo (TOYOBO) polymerase. The 3a-ORF-specific RT-PCR was used to detect D3Yα as described by Takeshita et al. (2008) and yield amplicons of 857 bp for RNA3 and 695 bp for D3Yα. The thermalcycling conditions were 94°C for 2 min; 30 cycles of 94°C for 30 s, 58°C for 30 s, and 68°C for 2.5 min; 72°C for 2 min. The PCR products were resolved using 1.0% agarose gel electrophoresis.

The nucleotide sequences of 3a-ORF-specific RT-PCR products were determined using the Big Dye Terminator DNA Sequencing Kit v. 3.1 (Applied Biosystems, Waltham, MA, USA) and the ABI Prism 310 Genetic Analyzer. The sequence was analyzed using the program GENETYX-Win v. 15 (Genetyx Corp., Tokyo, Japan).

To quantify CMV-Y, ΔY2b-C, and ΔY2b, the CP-gene-targeting one-step quantitative RT-PCR (RT-qPCR) analysis was performed as previously reported by Ota et al. (2021). Two primer sets were used to amplify the CP gene and SAMD as a reference gene in N. benthamiana (Khaing et al., 2020; Liu et al., 2012). Obtained data was analyzed by ΔΔCt method and statistically processed by t-test (P < 0.05 or P < 0.01) in Microsoft Excel software.

Cowpea develops local lesions against CMV-Y infection according to Kim and Palukaitis (1997). We grew cowpea at 25°C/20°C during the experiment period. Non-inoculated, upper leaves of three to four N. benthamiana that were inoculated with CMV-Y, ΔY2b-C, or ΔY2b during the second passage were collected at 7 and 28 dpi, then we used these tissues for sap inoculation onto the primary leaves of cowpea to observe the development of local lesions at 7 dpi. Also, we examined the active replication of CMV-Y from N. benthamiana held at 15°C at 7 or 28 dpi or held at 36°C at 7 dpi. The number of local lesions per leaf was counted and the average number of local lesions and standard errors were calculated, then were statistically processed by t-test (P < 0.05 or P < 0.01).

To examine host adaptation of D3Yα in other plant species, we inoculated cucumber and tomato plants with purified CMV-Y virions or with a mixture of CMV-Y virions and D3Yα as described earlier (Table 1). RT-PCR using 3a-ORF-specific primers showed that when mixture of CMV-Y and D3Yα was used as the inoculum, the upper, non-inoculated leaves of cucumber plants had accumulated D3Yα by 28 dpi, but only two out of twenty tested tomato plants retained D3Yα throughout experiment 1 and 2 (Table 1). These results suggested that D3Yα varies in its adaptability to different plant host species. Interestingly, D3Yα was not detected in cucumber and tomato plants at 28 days after inoculation with only CMV-Y, suggesting that the generation of D3Yα depends on the host (Table 1).

In our previous study, no subviral RNAs other than D3Yα were detected in total RNA extracted from CMV-Y-infected N. benthamiana and N. tabacum plants, which were inoculated with in vitro transcripts of CMV-Y RNAs mixed with D3Yα, for at least 5 passages at 7-to-10-day intervals (Takeshita et al., 2008). Using a mixture of CMV-Y RNA transcripts as inoculum to establish that D3Yα has been generated is a key step to elucidate the mechanisms involved in D3Yα generation during virus infection. To test for the generation of D3Yα experimentally, we first examined N. tabacum that were inoculated with CMV-Y RNA1 to RNA3 transcripts in the first passage and did not detect D3Yα using 3a-ORF-specific RT-PCR throughout the experiment (10 and 40 dpi) (data not shown).

We then serially inoculated tobacco plants (the second passage) with sap from N. tabacum in the first passage to examine generation and maintenance of D3Yα in the plants over a longer period after inoculation (Fig. 1A). In the second passage, although D3Yα was not detected at 14 dpi in N. tabacum plants, it was detected at 30 dpi (Fig. 1B). We detected D3Yα at 90 and 180 dpi in the second passage; however, more plants had no D3Yα over time (Fig. 1B). These results indicated that by several weeks after inoculation, D3Yα levels start to decline for the next several months (Fig. 1B). Like during the second passage at 14 and 30 dpi, the same phenomenon occurred in the third passage (Fig. 1C). D3Yα was similarly generated in N. benthamiana (data not shown). All of the D3Yα RT-PCR products from the second and the third passage had the same nucleotide sequence that was reported by Takeshita et al. (2008) (Supplementary Figs. 2 and 3).

To test the influence of temperature other than 25°C/20°C (day/night) on the generation and retention of D3Yα in plants, we first tested the effect of 15°C using N. benthamiana, which were inoculated with sap from the first passage plants infected with CMV-Y or CMV-Y plus D3Yα, but only incubating the plants at 15°C during the second passage (Fig. 2A). CMV-Y-infected N. benthamiana kept at 15°C showed stunting but not yellow mosaic symptom (Supplementary Fig. 4). None of the plants inoculated with CMV-Y had D3Yα at 7 and 28 dpi (Fig. 2A). On the hand, the plants, which were inoculated with CMV-Y plus D3Yα and held at 15°C, had retained D3Yα by 28 dpi (Fig. 2A). In subsequent test of the effects of high temperature on generation and maintenance of D3Yα, N. benthamiana plants were inoculated with sap from the first passage plants infected with CMV-Y or CMV-Y containing D3Yα and kept at 36°C during the second passage. By 7 dpi, none of the CMV-Y-inoculated plants generated D3Yα (Fig. 2A), and none survived beyond 10 dpi (Supplementary Fig. 4). However, the plants, which were inoculated with CMV-Y containing D3Yα, retained D3Yα at 36°C for at least 7 dpi (Fig. 2A). Once D3Yα was present in the virus populations, then D3Yα seemed to be retained at a wide range of temperatures. The temperature conditions that are suitable for retaining D3Yα do not always seem to correspond with the temperature at which D3Yα is generated by CMV-Y in N. benthamiana.

The 2b protein of CMV counteracts the host RNA silencing and the SA-mediated antiviral resistances (Nakahara and Masuta, 2014), drives selection of interviral recombinants between CMV RNA1/2 and tomato aspermy virus RNA3 (Shi et al., 2008), and promotes unloading of CMV from vasculature into nonvascular cells (Takeshita et al., 2012). To assess the potential effects of the 2b protein on D3Yα generation, we inoculated N. benthamiana with the vector CMV2-A1 (Otagaki et al., 2006) or CMV-H1 (Matsuo et al., 2007). The CMV2-A1 vector (ΔY2b-C) has a functional 2b protein but the 3′ portion of its ORF was deleted; the CMV-H1 vector (ΔY2b) is a deletion mutant of CMV-Y that lacks the entire gene for the 2b protein (Supplementary Fig. 1). For achieving long-lasting systemic infection after inoculation with ΔY2b, we selected N. benthamiana as a host to compare infection after inoculation with CMV-Y, ΔY2b-C, and ΔY2b (Fig. 2B). The CMV-Y- and ΔY2b-C-infected plants showed severe yellow mosaic and green patches in upper, non-inoculated leaves at 7 and 28 dpi, respectively, while ΔY2b mutant induced a slight vein clearing throughout the experiment (Supplementary Fig. 5). The 3a-ORF-specific RT-PCR amplified DNA fragments from the entire region of the 3a ORF but did not produce any smaller ones from the ΔY2b-C or ΔY2b-infected plants even at 40 dpi during the second passage (Fig. 2B). To assess replication and accumulation levels of CMV-Y, ΔY2b-C, and ΔY2b, we mechanically inoculated cowpea leaves using sap from N. benthamiana held at 25°C/20°C and used RT-qPCR targeting CP gene of CMV-Y to measure relative viral levels in the upper, non-inoculated leaves from the infected N. benthamiana at 7 and 28 dpi (Fig. 3). The number of local lesions on primary leaves of cowpea plants, which were inoculated with sap from CMV-Y- or ΔY2b-C-infected N. benthamiana at 7 or 28 dpi, did not differ significantly (Fig. 3A). In addition, RT-qPCR targeting the CP gene of CMV-Y showed a significant difference in levels between CMV-Y and ΔY2b-C and between CMV-Y and ΔY2b at 7 dpi but not at 28 dpi (Fig. 3B). The number of local lesions on cowpea was positively correlated with viral levels in N. benthamiana (Fig. 3A and B). Also, we inoculated the primary leaves of cowpea with CMV-Y-infected N. benthamiana kept at 15°C or 36°C to exam CMV replication activity (Fig. 3C). The number of local lesions on the primary leaves of cowpea inoculated with sap from N. benthamiana held at 15°C indicated that CMV-Y was replicating without generating D3Yα at 7 and 28 dpi (Fig. 3C). Interestingly, at 7 and 28 dpi with CMV-Y, sap from N. benthamiana held at 15°C after inoculation induced more local lesions than those that were kept at 25°C/20°C (day/night) or 36°C (Fig. 3A and C). The results suggested that the intact 2b protein rather than a high level of virus, facilitates the generation of D3Yα.