2.1. The Olfactory Attraction of the Whitefly Vector Is Conserved in All Betasatellite SubgroupsThe betasatellites are essential pathogenicity determinants for begomoviruses with a typical feature of a single ORF βC1 encoded in the complementary-sense [27]. In the last five years, the number of betasatellites global isolates has increased from ~1300 to ~3100, and species have increased from 61 to 119. Field surveys have revealed that the outbreak of begomovirus-betasatellite diseases is often associated with the expansion of whitefly distribution [28,29,30]. Therefore, we speculate that the previous discovery of begomovirus-whitefly mutualism may be a conserved trait. Begomovirus-associated betasatellites are believed to have evolved independently and encode single conserved βC1 proteins [18,27,31], indicating that no suitable roots can be found in the phylogenetic tree for betasatellite species. Thus, to test this hypothesis, we performed a phylogenetic analysis of all available 119 βC1 proteins encoded by these distinct begomovirus-associated betasatellite species, using an unrooted tree. The phylogenetic analysis showed that betasatellite-encoded βC1 proteins could be divided into three subgroups (I, IIA, and IIB) (Figure 1A). Interestingly, the βC1 from a Japan-derived betasatellite (EpYVB, eβ) associated with a begomovirus Eupatorium yellow-vein virus (EpYVV), regarded as a causing pathogen of the earliest known record of a plant virus disease in 752 AD [32], and the recent China-derived Siegesbeckia yellow vein betasatellite (SiYVB, sβ) belonged to subgroup I. There was another African subclade in subgroup I of βC1 proteins, including African species such as cotton leaf curl Gezira betasatellite (CLCuGeB) and tomato leaf curl Ghana betasatellite 2 (ToLCuGHB2). Cotton leaf curl Multan betasatellite (CLCuMuB, cβ) with the most records in NCBI of 585 accessions, belonged to subgroup IIA. Compared to subgroups I and IIA, subgroup IIB contained more than half of the total numbers of betasatellite species, including our previously described Tomato yellow leaf curl China betasatellite (TYLCCNB, tβ).One open question is how betasatellites and their encoded βC1 proteins have existed in nature and evolved with their hosts for such long time. To investigate whether functional evolution of betasatellites and βC1 proteins are related to the host preference of its whitefly vector, we performed insect preference assays for four betasatellites (including eβ, sβ, cβ, and tβ) belonging to three subgroups in Nicotiana benthamiana (Nb) plants. Given that betasatellites primarily rely on their helper viruses to spread, multiply, and encapsulate, and generally have a loose specificity by different helper components [33,34]. We chose Tomato yellow leaf curl China virus TYLCCNV DNA-A (TA) as a helper virus to explore the function of other betasatellites. As shown in Figure S1A, plants inoculated with TA + tβ developed disease symptoms with severe curling of leaves and shot twisting, while only TA infection failed to cause symptoms. Plants inoculated with TA + cβ, TA + sβ, or TA + eβ displayed moderate leaf curling without shoot twisting. We further determined DNA accumulation of betasatellites in systemically infected leaves of plants infected with these four betasatellites in combination with the helper virus TA. The accumulation of tβ was higher in plants than in other betasatellites, which is consistent with the observed disease symptoms (Figure S1B–E). Therefore, the helper virus TA can trans-replicate these four betasatellites in Nicotiana benthamiana (Nb), and some specificity exists for betasatellites trans-replication by begomoviruses. Similar to our previous observation on TA + tβ infection, TA + cβ complex-infected Nb attracted more whiteflies than vector control healthy Nb plants (Figure 1B). Intriguingly, two betasatellites from the subgroup I showed different attractions to the whitefly vectors. The TA + eβ complex-infected Nb could efficiently attract more whitefly, but the TA + sβ complex-infected Nb failed to attract more whitefly than those of vector control healthy Nb plants. Considering some specificity between betasatellites and the helper begomoviruses, we employed EpYVV DNA-A (EA) as the original helper virus of its associated betasatellite eβ to infect Nb plants. The eβ-mediated whitefly attraction was confirmed again. These results suggest that the begomovirus-whitefly mutualism might evolve a millennium ago when subgroup I betasatellites diverged.Because plant terpenoid-related defenses play critical roles in the host preference, we performed volatile metabolite analysis to explore if the terpenoid-based defense still involves in the betasatellites-mediated whitefly attraction. A pivotal plant volatile sesquiterpene α-bergamotene, its emission significantly decreased on TA + tβ, TA + cβ, TA + eβ, and EA + eβ-infected Nb than that of vector control healthy Nb plants (Figure 1C,D). Consistent with TA + sβ infection displayed no changes in whitefly attraction, the level of α-bergamotene emission in TA + sβ-infected Nb plants showed similar to that of vector control healthy Nb plants. As expected, the transcription levels of sesquiterpene synthase gene NbTPS1 were significantly reduced in plants infected by begomovirus-betasatellite complexes, except for TA + sβ infection, which did not affect NbTPS1 expression (Figure S2). To further explore whether the attraction by betasatellites is specific to the whitefly vector, we performed insect choice assays with begomovirus non-vector green peach aphid (Myzus persicae). More non-vector aphids were preferred on uninfected plants compared to TA + tβ infected plants (Figure S3). These results are consistent with the fact that α-bergamotene has been reported as an attractant to aphids [35]. These results suggest that whitefly attraction is specifically conserved in betasatellites from all three subgroups I, IIA, and IIB, by regulating plant terpenoid-based defense. 2.2. The Evolutionarily Conserved Serine-33 Is Essential for βC1-Mediated Whitefly AttractionThe above results show that siegesbeckia yellow vein betasatellite sβ belonging to betasatellite subgroups I do not possess traits of whitefly attraction caused our interest in understanding the behind mechanism. Mutations are considered one crucial factor in genetic diversity among plant viruses [17]. All functions of betasatellites are accredited to a single complementary protein βC1, and the phosphorylation modification on βC1 protein is a major antiviral plant response and pathogenesis [36,37,38]. We thus wanted to determine if some mutations on phosphorylation sites of βC1 protein make sβ not change the host preference. To test this, we analyzed the primary amino acid sequence of βC1 proteins encoded by 119 betasatellite molecules and predicted potential phosphorylation sites using the NetPhos 3.1 server (http://www.cbs.dtu.dk/services/NetPhos; accessed on 22 September 2021). Each βC1 protein has several evolutionarily conserved phosphorylation sites, among which serine-33 is the most conserved residue in 105 of 119 Betasatellite species in the three subgroups (Figures S4 and S5). Intriguingly, most of βC1 proteins with non-Serine-33 in 14 Betasatellite species were found in subgroup I, esp. the African subclade and sβ isolated from Asia. The phosphorylation mimics impair the functions of βC1 as a viral suppressor of RNA silencing and a symptom determinant [36,37]. To identify whether the serine-33 of βC1 functions has a critical role in the olfactory attraction of whitefly, we mutated the serine-33 to aspartate (S33D) to mimic the phosphorylation of βC1 proteins in tβ and eβ and named as tβS33D and eβS33D. When the begomovirus-betasatellite complexes TA + tβS33D and EA + eβS33D infected Nb plants, we observed that TA + tβS33D and EA + eβS33D abolished betasatellite-mediated whitefly vector attraction (Figure 1B). The emission of α-bergamotene and the expression of NbTPS1 on TA + tβS33D and EA + eβS33D infected Nb was also comparable to healthy plants, which was significantly higher than that of TA + tβ and EA + eβ-infected Nb plants (Figure 1D and Figure S2A,D). These results suggest that an evolutionarily conserved phosphorylation site serine-33 of βC1 protein is essential for betasatellite-mediated whitefly attraction.To further confirm the conserved phosphorylation site serine-33 of βC1 protein-mediated vector choice, we used the heterologous Potato virus X (PVX) model system for systemic ectopic expression of betasatellited-encoded βC1 proteins and βC1 mutations through agro-inoculation of Nb plants. Plants inoculated with PVX-GFP served as a control. There were no obvious morphological differences between these recombinant PVX vector-infected plants (Figure S6). We performed the two-choice assay to determine whether the expression of a single βC1 protein is sufficient to affect the whitefly attraction. Similar to the whitefly attraction toward the host infected with begomovirus-betasatellite complex, more whiteflies were attracted to Nb plants infected with recombinant PVX-eβC1, PVX-tβC1 compared to PVX-GFP control plants (Figure 2A). Whereas PVX-eβC1S33D and PVX-tβC1S33D-infected Nb plants significantly decreased whitefly attraction compared with PVX-eβC1 and PVX-tβC1-infected Nb plants. Consistent with what we have observed on the begomovirus TA + sβ complex-infected plants, PVX-sβC1-infected plants could not change the whitefly preference. If the βC1 proteins with serine-33 available for phosphorylation are vital for whitefly attraction, we considered that the substitution of serine-33 with cysteine-31 in the sβC1 protein could compromise the ability to attract whitefly and suppress plant terpene-related defense. Thus, we postulated that mutation of cysteine-31 to a serine residue in sβC1 protein (sβCC31S) could restore its function on whitefly attraction. Indeed, we observed that PVX-sβC1C31S-infected plants attracted more whiteflies than those of PVX-sβC1 and PVX-GFP-infected plants (Figure 2A). Consistent with the effect on changes in the whitefly olfactory behavior, changes in α-bergamotene emission and NbTPS1 expression in infected plants are relevant to the specific variant of serine-33 in βC1 proteins. Similar to PVX-eβC1 and PVX-tβC1 infection, PVX-sβCC31S infection displayed a decrease of α-bergamotene emission and downregulation of NbTPS1 expression compared to PVX-GFP infection (Figure 2B–D). PVX-eβC1S33D and PVX-tβC1S33D infection exhibited much fewer effects on α-bergamotene emission and NbTPS1 expression than those of PVX-GFP infected plants. These results further demonstrate that the dynamic phosphorylation state of serine-33 in βC1 proteins confers the evolutionarily conserved whitefly attraction by the begomovirus-betasatellite complex. 2.3. The Conserved Serine-33 in βC1 Proteins Is Responsible for Suppressing MYC2 DimerizationWe previously demonstrated that βC1 interacts with MYC2 to suppress plant terpenoid-based defense and promotes whitefly preference [13]. The phosphorylation mimics of βC1 abolish its interaction with ASYMMETRIC LEAVES 1 (AS1) to attenuate virus disease symptoms in N. benthamiana leaves [37]. We asked whether the phosphorylation mimics of βC1 proteins at the conserved serine-33 also alter its interaction with MYC2 to affect plant terpenoid-based defense and whitefly preference. To investigate this, we determined the interaction of tβC1 and its phosphorylation mimics with AtMYC2 by a bimolecular fluorescence complementation (BiFC) assay. Agrobacterium tumefaciens strains containing expression vectors for tβC1-cEYFP or tβC1S33D-cEYFP with nEYFP-MYC2 were co-infiltrated into N. benthamiana leaf cells. The strong interaction (represented by fluorescence) between tβC1-cEYFP and nEYFP-MYC2 or tβC1S33D-cEYFP and nEYFP-MYC2 was observed in nuclei (Figure S7), which has been confirmed by our previous study with DAPI staining [13]. As a negative control, no fluorescence was observed when tβC1-cEYFP or tβC1S33D-cEYFP was coexpressed with the nEYFP-GUS vector. This conserved serine-33 residue of βC1 proteins does not play an essential role in its interaction with MYC2.MYC2 dimerizations contribute to the binding to the G-boxes in the promoter of its downstream target genes and regulate multiple signaling events [39,40]. Given that the TYLCCNB-encoded tβC1 suppresses MYC2 activity by interfering with MYC2 dimerization [13], we hope that the conserved serine-33 in βC1 proteins are responsible for interfering with MYC2 dimerization. To examine this hypothesis, we performed a modified BiFC assay following our previous report [4]. In the presence of coexpressing tβC1, cβC1, and eβC1, the interaction signal strength of MYC2-MYC2 decreased to approximately half of its original intensity (Figure 3A,B). These results confirmed the conserved function of betasatellite-encoded βC1 proteins on the interference with MYC2 dimerization.Additionally, the phosphorylation mimics had a minor influence on the suppression of MYC2 dimerization compared with its wild-type βC1 proteins. sβC1 did not affect the formation of MYC2 dimerization, but the gain of function sβC1C31S harbored a similar effect on suppression of MYC2 dimerization to the other three βC1 proteins. Furthermore, an in vitro competitive pull-down assay further confirmed the above results. The amount of GST-MYC2 pulled down by MBP-MYC2 was reduced by increasing the amount of His-tβC1 in the mix reaction (Figure 3C). By contrast, increasing the amount of His-tβC1S33D protein showed less effect on MYC2 self-association than adding the same content of His-tβC1 in the reaction. Together, these results suggest that the conserved serine-33 is a key residue in βC1 proteins for interfering with MYC2 dimerization.Next, we asked whether the conserved residue of βC1 proteins at serine-33 influences the MYC2 trans-activation activity. Using TPS10 promoter: Luciferase (LUC) as a reporter, and MYC2, tβC1, and tβC1S33D as effectors, we transiently expressed with the indicated effector and reporter constructs in N. benthamiana leaf cells. GUS was used as the control effector. LUC activities were subsequently quantified. tβC1coexpressing with MYC2 significantly decreased the LUC activity, whereas tβC1S33D coexpression did not show significant differences in MYC2-induced LUC activity compared to GUS coexpression (Figure 3D,E). These results indicate that the conserved serine-33 also plays a key role in βC1 suppression of MYC2 trans-activity. 2.4. The Conserved Serine-33 in βC1 Proteins Contributes to Its Compacity to Affect Whitefly PerformanceTo further confirm the importance of serine-33 sites of βC1 on whitefly performance, we conducted whitefly two-choice assays and bioassays using stable βC1-transgenic Arabidopsis thaliana. Consistent with these above results on virus-infected N. benthamiana (Figure 1B and Figure 2A), eβC1 transgenic Arabidopsis was more attractive to whitefly compared to wild type Col-0, while eβC1S33D transgenic Arabidopsis attracted fewer whiteflies compared to Col-0 and eβC1 (Figure 4A), indicating that the phosphorylation of βC1 at serine-33 disrupts βC1-mediated whitefly preference. We note that sβC1 transgenic Arabidopsis failed to attract more whiteflies, but sβC1C31S transgenic Arabidopsis enhanced whitefly attraction compared to Col-0 and sβC1. These results indicate that the dynamic phosphorylation of serine-33 in βC1 proteins commonly renders the host more vector attraction through the evolution of betasatellites in nature.Because βC1 suppresses the activity of MYC2, which is involved in regulating multiple secondary metabolites against herbivore insects [40], to investigate whether the conserved serine-33 of βC1 is involved in plant resistance to whiteflies, we performed whitefly bioassays using Col-0 and βC1-overexpressing Arabidopsis plants. Whiteflies laid more eggs on tβC1, eβC1, and sβC1C31S transgenic plants than on Col-0 (Figure 4B). Conversely, similar eggs were laid by whiteflies on tβC1S33D, eβC1S33D, and sβC1 transgenic plants compared to on Col-0 plants. We thus considered that whitefly development could be affected by expressing βC1 proteins in plants. As expected, whitefly developed significantly faster on tβC1, eβC1, and sβC1C31S transgenic plants than on Col-0 plants, as indicated by more pupae (Figure 4C). Whitefly development on tβC1S33D, eβC1S33D, and sβC1 transgenic plants progressed similarly to Col-0. These data suggest that the dynamic phosphorylation of serine-33 in βC1 proteins contributes to whitefly performance in plants.