ABSTRACT
Salicylic acid (SA)-mediated immunity plays important roles in combating virus in plants. Two plant stress associated protein (SAPs) containing dual A20/AN1 zinc-finger domain were found to play important roles in SA-mediated immunity; however, detailed mechanisms remain elusive. In this study, another orchid homolog gene of Pha13, Pha21, was analyzed. Pha21 confers antiviral immunity in both transgenic orchid and Arabidopsis overexpressing Pha21. Expression of Pha21 is early-induced by SA treatment, and is involved in the expression of the orchid homolog of the master regulator NPR1. Pha21 but not Pha13 is involved in the expression of key RNAi-related genes, Dicer-like nuclease 4 (DCL4) and Argonaut 1 (AGO1) in orchids. The involvement of SAPs in expression of orchid DCL4 and AGO1 is not limited to orchid, as AtSAP5 also plays essential role in the expression of Arabidopsis DCL4 and AGO1. However, unlike Pha13 and AtSAP5, Pha21 does not play positive role in the expression of orchid homolog gene of RNA-dependent RNA polymerase 1 (RdR1), an important gene in RNAi pathway. Pha21 can be found localized in the nucleus, and confers self-E3 ligase and ubiquitin binding activities. Functional domain analysis revealed that both A20 and AN1 domains of Pha21 are required for decreasing virus accumulation, and the AN1 domain plays a more important role in the expression of orchid DCL4. Collectively, our data suggests SA regulated SAPs play important roles in antiviral immunity and is involved in delicate regulation of key genes in RNAi-mediated pathway.
IMPORTANCE Salicylic acid (SA)-mediated antiviral immunity plays an important role to protect plants from virus infection; however, the detailed mechanisms remain elusive. We previously demonstrated that two plant A20/AN1 proteins, orchid Pha13 and Arabidopsis AtSAP5, function similarly and serve as an important hub to regulate SA-mediated antiviral immunity. In this study, we identified another orchid A20/AN1 protein, Pha21, which is involved in SA-mediated antiviral immunity. In contrast to Pha13 and AtSAP5, Pha21 plays minor negative roles in the expression of PhaRdR1 (orchid homolog of RNA-dependent RNA polymerase 1). However, Pha21 and AtSAP5, but not Pha13, are involved in the expression of important players in RNAi pathway, Dicer-like nuclease 4 (DCL4) and Argonaut 1 (AGO1), in orchid and Arabidopsis. Our data demonstrates that plant A20/AN1 proteins are conserved players in SA-mediated antiviral resistance among plants, and provide links between the A20/AN1 proteins and the RNAi pathway.
INTRODUCTION
The plant hormone salicylic acid (SA) is important to trigger plant immunity especially against biotrophic pathogens such as viruses (1–4). In facing pathogen invasion, plants recognize conserved microbe-associated molecular patterns (MAMPs) by pattern-recognition receptors and trigger pathogen (or pattern)-triggered immunity (PTI) as the first line of defense (5, 6). Although PTI provides protection from invasion by most pathogens, some have evolved, and can utilize different effectors to suppress PTI for successful infection (7). Plants have evolved resistance (R) proteins capable of detecting these effectors to trigger effecter-triggered immunity (ETI) for a second line of plant defense (7). ETI is generally associated with programmed cell death leading to the formation of necrotic lesions, which is also known as hypersensitive response (HR) to prevent the further spreading of pathogen infection (1).
Current evidence suggests that the PTI is important to limit virus infection (8–10), and the viral double strand RNA (dsRNA) has been shown to serve as a conserved MAMP (10). In addition, ETI also plays an important role in combating virus infection in plants. For example, the N gene from Nicotiana glutinosa serves as an R protein to specifically recognize the effector from Tobacco mosaic virus (TMV) and trigger the ETI (11–14). SA plays important roles in triggering PTI and ETI (15). In addition, the TMV-infected tobacco plant showing local necrotic lesions was found to become more resistant against secondary virus infection in the distal leaves (16, 17). This systemic immune response is known as systemic acquired resistance, which is a common plant immune response and plays an important role in protecting plants from pathogen infection (4).
In addition, RNA silencing or RNA interference (RNAi) has also been demonstrated to play an important role in combating virus infection (18, 19). RNAi is activated through the appearance of the viral dsRNA upon virus infection. The dsRNA can be cleaved to short small-interfering dsRNA (siRNA) of 21–24 nucleotides (nt) in size by the Dicer-like (DCL) nuclease (20). DCL2 and DCL4 in Arabidopsis play an important role in generating the virus-derived siRNA (vsiRNA) (19, 21, 22). The vsiRNA is further unwound and one strand (guide-strand) is incorporated into the RNA-induced silencing complex (RISC) (19). This complex targets viral RNA with a sequence that is complementary to the guide-strand and degrades the target RNA by the catalytic component of the RISC, Argonaut (AGO) nucleases (19, 20, 23, 24). Among the Arabidopsis AGOs, AGO1 plays an important role in the antiviral immunity against RNA viruses. (24). More vsiRNA can be further generated de novo through the cellular RNA-dependent RNA polymerases (RDRs) to trigger the secondary RNA silencing against viruses (25–27). SA also plays important roles in the RNAi-pathway as SA treatment can induce genes involved in the RNAi pathway including RDRs in Arabidopsis, Nicotiana tabacum and Tomato, and DCLs and AGOs in Tomato (28–32).
SA induces the fluctuation of redox and serves as a signal to activate sets of defense genes (33, 34). The SA-induced redox change can modify the NPR1 (nonexpressor of pathogenesis-related protein 1) from the multimeric protein complex to a monomer via the oxidoreductases, Thioredoxin-3 and Thioredoxin-5 (35, 36). The NPR1 monomer moves into the nucleus and activates multiple defense-related genes including the PR (pathogenesis-related) gene in the SA-signaling pathway (37–39). In addition to the NPR1-dependent immune pathway, some data also suggest that the existence of a NPR1-independent pathway contributes to the virus resistance (40, 41). Despite the importance of the SA governed immune pathway in antiviral immunity, the regulation of this pathway remains largely elusive.
Proteins containing zinc-finger A20 and/or AN1 domains play an important role in plant response against various abiotic stresses, and are known as stress associated protein (SAPs). SAPs are conserved among different organisms (42, 43). Different numbers of SAP homologs (range from 1 to 19) have been identified in organisms including protists, fungi, animals, plants, and humans (42, 43). Compared to animals, more of these proteins are found in plants. So far, 18, 14 and 14 SAPs have been identified in rice, Arabidopsis, and Phalaenopsis orchid, respectively (42, 44). Biochemical studies of proteins containing A20 and/or AN1 domains revealed that the A20 domain confers E3 ligase and ubiquitin binding activity (43–49). The A20 domain of human A20, Rabex-5 (guanine nucleotide exchange factor), Arabidopsis AtSAP5 and orchid Pha13 have been reported to exhibit E3 ligase activity (44, 45, 49, 50). In addition, it has been shown that the A20 domain can also bind to various ubiquitin chains (44, 46–48). In contrast to the A20 domain, the biochemical function of the AN1 domain is not fully understood.
Our recent study indicates that plant SAPs play a pivotal role in SA governed antiviral immunity (44). Orchid Pha13 and Arabidopsis AtSAP5 are induced by SA treatment at the early stage, and are involved in expression of orchid or Arabidopsis NPR1 and NPR1-independent immune responsive genes including the induction of plant RdR1. In addition to Pha13, our previous virus-induced gene silencing (VIGS) assay also allowed us to identify the involvement of orchid PhaTF21 (designated here as Pha21) in SA–regulated immune response genes (51). Pha13 and Pha21 share 69.5% amino sequence identity, and are most closely related among the orchid SAPs (44). Whether Pha21 and Pha13 work in a cooperative manner in plant antiviral immunity remains to be explored.
Here, we performed a detailed study of Pha21, and compared its biochemical and physiological function to Pha13. Pha21 and Pha13 share similar biochemical properties including E3-ligase and ubiquitin binding activities. Our results also indicate that similar to Pha13, Pha21 is early-induced by SA, involved in the expression of the PhaNPR1 and PhaNPR1-independent genes, and plays an important role in antiviral immunity. Importantly, our studies show that Pha21 is involved in expression of orchid DCL4 and AGO1. Together with our previous data, these studies suggest that Pha13 and Pha21 coordinate the expression of genes important in the RNAi pathway including orchid RdR1, DCL4 and AGO1. In addition, our previous data and data presented in this study also showed that Arabidopsis AtSAP5 is involved in expression of RdR1, DCL4 and AGO1. Collectively, our data indicates that plant A20/AN1 proteins involved antiviral immunity are conserved among plants, and the antiviral immunity is partly through the RNAi pathway. Our findings suggest that A20/AN1 proteins may serve as a link between SA and the RNAi pathway at the early stage of SA induced-antiviral immunity.
RESULTS
Sequence and expression pattern analysis of Pha21
Previously, we reported that Pha21 (Orchidstra 2.0 database, http://orchidstra2.abrc.sinica.edu.tw, accession number PATC144963) is involved in the SA induced immune pathway in P. aphrodite (52). Pha21 contains dual zinc-finger domains, A20 and AN1, in the N-terminal and C-terminal, respectively (Fig. 1A-C), and belongs to the stress associated protein (SAP) family (43). Among the SAPs of P. aphrodite, Pha21 is most related to our previously reported Pha13 and shared 69.5% amino acid sequence identity (44). The A20 and AN1 zinc finger domains of Pha21 also shared high amino acid sequence identity among SAPs from different species including AtSAP5 from Arabidopsis, and OsSAP3 and OsSAP5 from Oryza sativa (Fig. 1B). (44). Unlike Pha13, no nuclear localization signal was identified in Pha21 (Fig. 1A).
The expression of Pha21 was analyzed in different tissues of P. aphrodite orchid including root, leaf, septal, petal, lip, and column. The results revealed that Pha21 expression was higher in the column, root and leaf (Fig. 1D). In addition, we also analyzed the absolute expression of Pha21 and Pha13 in the leaves of P. aphrodite by use of droplet digital PCR. The results indicated that Pha21 RNA expression is about 15 times lower than Pha13 in leaves of P. aphrodite (Fig. 1E)
Pha21 is involved in the expression of SA responsive genes, PhaNPR1 and PhaPR1
To further analyze the role of Pha21 in the SA-induced immune pathway, we transiently expressed two hairpin RNA (will generate 21 nt siRNA) of Pha21 (35S::hpPha21-1 and 35S::hpPha21-2) separately to knockdown Pha21 by agroinfiltration in P. aphrodite carrying hairpin RNA expression construct, phpPha21-1 and phpPha21-2 (Fig. 2A). Samples collected from the infiltrated site of orchid leaves were used to detect the RNA level of Pha21, PhaNPR1, and PhaPR1 by quantitative RT-PCR (qRT-PCR). As shown in Fig. 2A, the expression of PhaNPR1 and PhaPR1 was decreased in both Pha21-silenced plants. In addition to transient knockdown of Pha21, we also transiently overexpressed Pha21 (35S::FLAG-Pha21) through agroinfiltration in the leaves of P. aphrodite. However, transient overexpression of Pha21 in leaves did not significantly affect the RNA level of PhaNPR1 and PhaPR1 (Fig. 2B).
In addition, we also transiently silenced PhaNPR1 (phpPhaNPR1) in P. aphrodite to analyze the expression of Pha21. The results showed that the silenced PhaNPR1 showed decreased expression of PhaPR1; however, no obvious effect was observed on the expression of Pha21 (Fig. 2C). These results suggest that Pha21 is involved in the expression of PhaNPR1, but not vice versa.
Pha21 is induced by SA, jasmonic acid and ethylene
We also tested whether Pha21 was induced by defense-related plant hormones including SA, jasmonic acid (JA) and ethylene (ET). Orchid plants were treated with SA, JA and ET and samples were collected after treatment at different time points (up to 72 h) for analysis of the expression of Pha21 and the marker genes of each plant hormone by qRT-PCR. The induction of Pha21 by SA was observed at 1 h post-treatment (Fig. 2D). In addition, Pha21 was also induced by JA and ET at 72 h and 12 h, respectively (Fig. 2E and F). Our results showed that Pha21 can respond to multiple defense-related plant hormones, and SA induced Pha21 expression at the very early stage.
Pha21 plays a positive role in virus resistance
To analyze whether Pha21 plays a role in virus resistance, we first analyzed the expression of Pha21 in response to virus infection. We detected the RNA level of Pha21 in mock- or CymMV-inoculated P. aphrodite using qRT-PCR, and the results showed that the RNA of Pha21 is induced by CymMV infection (Fig. 3A). Furthermore, we transiently silenced or overexpressed Pha21 in CymMV-infected P. aphrodite to assay the effect on virus accumulation. Transient knockdown of Pha21 had no significant effect on CymMV accumulation (Fig. 3B); whereas transient overexpression of Pha21 decreased CymMV accumulation (Fig. 3C). In addition, we also generated transgenic P. equestris overexpressing Pha21 (35S::FLAG-Pha21). The expression of Pha21 is higher in the two individual asexually propagated progeny derived from two T0 lines of transgenic P. equestris (Pha21#8 and Pha21#9) as compared to the non-transgenic lines (WT) (Fig. 3D). We further inoculated CymMV into individual progenies derived from two transgenic lines, and the result showed that the accumulation level of CymMV is decreased to 21% and 34% in the two transgenic lines (Pha21#8 and Pha21#9) as compared to the non-transgenic lines (WT) (Fig. 3D). Our data suggest that Pha21 plays a positive role in virus resistance.
Transgenic Arabidopsis overexpressing Pha21 enhances antiviral resistance
To understand whether Pha21-mediated antiviral immunity is conserved in plants, we generated the transgenic Arabidopsis (Col-0) overexpressing Pha21 (35S::FLAG-Pha21). Homozygous T3 plants derived from 3 T1 transgenic lines, At-Pha21#4, #5 and #6, were selected for further antiviral assay. The expression level of Pha21 was confirmed by qRT-PCR on the homozygote progenies (Fig. 4A).
We mechanically inoculated Cucumber mosaic virus (CMV) to wild-type (WT) and Pha21 overexpressing Arabidopsis (At-Pha21#4, #5 and #6). Fourteen days post-inoculation, the accumulation of CMV in WT and Pha21 overexpressing Arabidopsis were detected by qRT-PCR. The results showed that Pha21 overexpressing Arabidopsis decreased the accumulation of CMV (Fig. 4A) compared to the WT, and severe disease symptoms were observed on the WT but not in the Pha21 overexpressing Arabidopsis (Fig. 4B).
Subcellular localization of Pha21
To better understand the role of Pha21 within cells, we first analyzed the subcellular localization of Pha21. We fused green fluorescent protein (GFP) in the N- or C-terminal of Pha21 driven by 35S promoter (pG-Pha21 and pPha21-G). The GFP-fused Pha21 was further transfected into protoplasts isolated from P. aphrodite. Twenty-four hours post-transfection, the localization of Pha21 was observed by using a confocal microscope. The results indicated that the C-terminal GFP fusion protein (Pha21-G) was observed in the nucleus in about 50% of cells; whereas the N-terminal GFP fusion protein (Pha21-G) showed a similar result to our GFP control vector, which had no nucleus-specific GFP (Fig. 5A). Our data suggested that Pha21 can move into the nucleus even without the predicated NLS signal.
Pha21 and Pha13 does not confer transcriptional activation ability in yeast two-hybrid assay
Since both Pha21 and Pha13 have the ability to move into the nucleus (44), we tried to understand whether Pha21 and Pha13 function as transcription factors. Therefore, we analyzed its transcriptional activation ability through yeast two-hybrid (Y2H) assay. We fused Pha21 and Pha13 to the Gal4 DNA binding domain to generate pGBKT7-Pha21 and pGBKT7-Pha13. We transformed the pGBKT7-Pha21 or pGBKT7-Pha13 in the yeast strain AH109 without any Gal4 transcriptional activation domain (AD) and further analyzed the transcriptional activation ability through the analysis of the activation of reporter genes in yeast. As shown in Fig. 5B, yeast with pGBKT7-Pha21 or pGBKT7-Pha13 cannot grow on the -Trp/Aureobasidin and -Trp/-Ade/-His medium, and the colonies did not turn blue on X-α-Gal containing medium.
Pha21 exhibits E3 ligase activity in which the A20 domain plays a major role
Several A20 and/or AN1 domain containing proteins confer E3 ligase activity (44–46, 53). To determine whether Pha21 functions as an E3 ligase, in vitro self-ubiquitination E3 ligase activity assay was performed using FLAG-Ubiquitin (FLAG-Ub), human E1 (hE1), human E2 (UBCH2, hE2), and purified His-tagged Pha21 (Pha21-His). The poly-ubiquitinated Pha21 can be detected in the presence of FLAG-Ub, hE1, and hE2 using anti-FLAG antibody to detect FLAG-Ubiquitin. The results revealed that Pha21 confers E3 ligase activity (Fig. 6A).
Furthermore, we also analyzed the self-ubiquitination E3 ligase activity of the A20 and/or AN1 domain of Pha21. We substituted the conserved cysteine and histidine to glycine on the A20 and/or AN1 domains of Pha21-His (Fig 1B and C) to generate A20 domain mutant (Pha21-A20m), AN1 domain mutant (Pha21-AN1m), and the double mutant (Pha21-A20mAN1m) for E3 ligase activity analysis. As shown in Fig 6A, the major E3 ligase activity was conferred by the A20 domain. A20 domain mutant and double mutant of Pha21 showed lower self-ubiquitination E3 ligase activity than wild-type Pha21 and Pha21-AN1 mutant (Fig. 6A).
A20 domain of Pha21 confers ubiquitin binding ability
A20 and/or AN1 proteins have also been shown to confer ubiquitin binding ability in animals and plants (44, 46, 47, 54, 55). To analyze whether Pha21 has ubiquitin binding ability, Y2H assay was performed to verify the interaction using Pha21 as bait and ubiquitin as prey. As shown in Fig 6B, Pha21 had a positive interaction with ubiquitin, suggesting that Pha21 conferred ubiquitin binding ability. In addition, to identify the ubiquitin binding domain of Pha21, a series of deletion mutants of Pha21 were generated, followed by Y2H assay. The result revealed that only truncated Pha21 fragments containing A20 domain (Pha211-136, Pha211-101, Pha211-40, Pha2116-158, Pha2116-136, Pha2116-101, Pha2116-40) showed a positive interaction with ubiquitin, suggesting that the A20 domain of Pha21 is responsible for the ubiquitin binding ability (Fig. 6B).
Interaction analysis between orchid SAPs
The A20/AN1 proteins may confer self-interaction and also interact with each other (56). Therefore, we also analyzed whether Pha21 and Pha13 confer self-interaction and also interact with each other. The Y2H assay was performed to verify the interaction between Pha21 and Pha13. Our Y2H results suggested that Pha21 and Pha13 did not interact with each other and no self-interaction of Pha21 or Pha13 was observed (Fig. 7).
Pha21 is involved in the expression of SA-induced PhaNPR1-dependent and - independent genes
To further understand how Pha21 is involved in the antiviral immunity, we performed transient silencing (35S::hpPha21-2) and overexpression assay (35S::FLAG-Pha21) of Pha21 in P. aphrodite, and analyzed the expression of the previously identified PhaNPR1-dependent and -independent antiviral genes, Phalaenopsis homolog of RNA dependent RNA polymerase 1 (PhaRdR1) and Glutaredoxin C9 (PhaGRXC9), respectively (44). Our results indicated that transient silencing of Pha21 RNA increased the expression of PhaRdR1, while PhaGRXC9 remained unchanged (Fig. 8A). Transient overexpression of Pha21 RNA decreased the expression of PhaGRXC9, but PhaRdR1 remain unchanged (Fig. 8B). Our data suggest that Pha21 affects the expression of PhaRdR1 and PhaGRXC9.
Pha21 is involved in the expression of PhaDCL4 and PhaAGO1
Because the RNAi pathway plays an important role in antiviral immunity (18, 19), we analyzed whether Pha21 is involved in the expression of core RNAi-related genes including the Phalaenopsis homolog genes of RdR2 (PhaRdR2, PATC124544), RdR6 (PhaRdR6, PATC131836), DCL2 (PhaDCL2, PATC143544), DCL4 (PhaDCL4, PATC150652), AGO1 (PhaAGO1, PATC157237), and AGO10 (PhaAGO10, PATC093469). Transient overexpression and silencing assay of Pha21 was performed in P. aphrodite. Overexpression of Pha21 increased the RNA expression of PhaDCL4 and PhaAGO1 (Fig. 9A); whereas transient silencing of Pha21 had no significant effect on the RNA expression of RNAi-related genes (Fig. 9B).
For comparison, we also analyzed the effect of Pha13 on the RNAi-related genes. The results showed that transient overexpression of Pha13 has no significant effect on the expression of PhaRdR2, PhaRdR6, PhaDCL2, PhaDCL4, PhaAGO1, and PhaAGO10 (Fig. 9C). Our results suggested that Pha21 is involved in the expression of PhaDCL4 and PhaAGO1.
To see whether orchid DCL4 (PhaDCL4) and AGO1 (PhaAGO1) also play a role in the antiviral immunity, we transiently silenced PhaDCL4 and PhaAGO1 through delivering the hairpin RNA into P. aphrodite by agroinfiltration carrying hairpin RNA expression constructs, phpPhaDCL4 and phpPhaAGO1. Our results showed that transient silencing of PhaDCL4 and PhaAGO1 increased the accumulation of CymMV (Fig. 9D and E).
The Arabidopsis homolog gene of Pha21, AtSAP5, is involved in the expression of DCL4 and AGO1
Our previous phylogenic analysis revealed that both Pha21 and Pha13 are closely related to Arabidopsis AtSAP5 (accession number: AT3G12630) (44). Therefore, we also analyzed whether AtSAP5 is involved in the expression of DCL4 and AGO1. We detected the expression of DCL4 and AGO1 in our previously generated transgenic Arabidopsis overexpressing AtSAP5 (AtSAP5-oe-4 and 11) and RNAi lines of AtSAP5 (AtSAP5-RNAi-3 and 7) by qRT-PCR (44). The expression level of AtSAP5 was confirmed by qRT-PCR (Fig. 10A). Our results showed that slightly increased expression of DCL4 and AGO1 were observed in overexpression transgenic lines, AtSAP5-oe-11 and decreased expression of DCL4 and AGO1 were observed in the RNAi lines, AtSAP5-RNAi-3 and AtSAP5-RNAi-7 (Fig. 10B and C). Our results suggest that AtSAP5 is important in the expression of DCL4 and AGO1.
A20 and AN1 domain of Pha21 play different roles in the expression of PhaDCL4, PhaAGO1, PhaGRXC9 and virus resistance
Our previous results indicated that Pha21 is involved in the expression of PhaDCL4, PhaAGO1, and PhaGRXC9 and virus accumulation (Figs 3C, 3D, 8B, and 9A). We further analyzed the roles of the A20 and/or AN1 domains of Pha21 in the expression of PhaDCL4, PhaAGO1, PhaGRXC9 and virus accumulation by overexpression of Pha21 and Pha21 with mutation in the A20 and/or AN1 domain. Therefore, we generated Pha21 with A20 domain mutant (Pha21A20m), AN1 domain mutant (Pha21AN1m), and the double mutant (Pha21A20mAN1m) through the substitution of the conserved cysteine and histidine to glycine on A20 and/or AN1 domain (Fig. 1B and C). We transiently overexpressed wild-type Pha21 or the mutants (A20 and/or AN1 domain mutant) in healthy or CymMV pre-infected P. aphrodite, and then detected PhaDCL4, PhaAGRO1, PhaGRXC9, and CymMV. The results indicated that overexpression of Pha21 A20 mutant (Pha21A20m) but not AN1 mutant (Pha21AN1m) or A20/AN1 mutant (Pha21A20mAN1m) increased the expression of PhaDCL4, which is similar to the wild-type Pha21 (Fig. 11A). Overexpression of any Pha21 A20 and/or AN1 mutant showed similar results to wild-type Pha21 with respect to increased expression of PhaAGO1 (Fig. 11A). Although overexpression of wild-type and any Pha21 A20 and/or AN1 mutant resulted in decreased expression of PhaGRXC9 and CymMV accumulation compared to the vector control, overexpression of wild-type Pha21 showed a greater effect on the expression of PhaGRXC9 and CymMV accumulation compared to any Pha21 A20 and/or AN1 mutant (Fig. 11A and B). Our data suggest that both the A20 and AN1 domains of Pha21 are required for expression of PhaGRXC9 and the decrease of CymMV accumulation, and the AN1 domain plays a major role in the expression of PhaDCL4. They also suggest that neither domain is required for the expression of PhaAGO1.
DISCUSSION
Plant A20/AN1 proteins, Pha13, Pha21 and AtSAP5 share similar biochemical properties and all involved in SA-mediated antiviral immunity
In this report, we performed detailed analysis of a previously identified orchid gene, Pha21, involved in SA and RNAi-mediated antiviral immunity. Pha21 shares high protein sequence identity to our previously identified Pha13 (Fig. S7 in Chang et al, 2018), and exhibits similar biochemical properties including self-E3 ligase, ubiquitin binding abilities (Fig. 6) and subcellular localization (Fig. 5A). Similar to Pha13, our data suggests that Pha21 is involved in the SA-signaling pathway downstream of SA and upstream of the expression of PhaNPR1 (Fig. 2A, C and D). In addition, Pha13 and Pha21 also positively mediate antiviral immunity (44). Transgenic Arabidopsis overexpressing Pha21 also enhanced resistance to virus infection (Fig 4) (44). In addition, our previous report and the data presented here indicate that Arabidopsis AtSAP5 is also induced by SA at the early stage, and is involved in expression of similar sets of SA mediated immune responsive genes, and positively regulates antiviral immunity (44).
Pha13-mediated expression of PhaGRXC9 and PhaRdR1 are dominant during SA treatment
Our previous study indicated that Pha13 and AtSAP5 plays a positive role in the expression of genes involved in the RNAi pathway and SA-governed oxidoreductases, RdR1 and GRXC9, respectively (44). We demonstrated that both PhaRdR1 and PhaGRXC9 play positive roles in antiviral immunity (44). In this study, we found that Pha21 has an opposite role to Pha13 in the expression of PhaRdR1 and PhaGRXC9 (Fig. 8). As our data indicates that the RNA level of Pha13 is approximately 15 times higher than Pha21 in leaves of orchids as measured by digital PCR (Fig. 1E), we suggest that the effect of Pha21 on the expression of PhaRdR1 and PhaGRXC9 is minor, and Pha13-mediated positive expression of PhaGRXC9 and PhaRdR1 are dominant during SA treatment.
Pha21 and AtSAP5 are all involved in expression of key genes in RNAi pathway, but regulation of PhaDCL4 and PhaAGO1 by Pha21 in orchid is different from regulation of DCL4 and AGO1 by AtSAP5 in Arabidopsis
Our data indicates that similar to Arabidopsis DCL4 and AGO1, orchid PhaDCL4 and PhaAGO1 also play a similar positive role in antiviral immunity in orchids as a slight decrease in the RNA of PhaDCL4 or PhaAGO1 (approximately 40% decrease) has a prominent effect on CymMV accumulation (approximately 100% increase) (Fig. 9D and E), (19). Interestingly, our data suggests that Pha21 regulates antiviral immunity partly through the effect on the RNAi pathway, as transient overexpression of Pha21 (1.6-folds) but not Pha13 increased the expression of two Phalaenopsis orchid homolog genes of DCL4 (PhaDCL4, 2-fold) and AGO1 (PhaAGO1, 1.6-fold) (Fig. 9A and C).
Although the effect of AtSAP5 on expression of DCL4 and AGO1 is not as prominent as Pha21, our data shown here also indicates that AtSAP5 is involved in the expression of DCL4 and AGO1. In both of our silencing lines, AtSAP5-RNAi-3 and −7 showed noticeable effect on the expression of DCL4 and AGO1 (Fig. 10), but that only one overexpression line, AtSAP5-oe-11 (expression increased about 7 times over AtSAP5) slightly increased the expression of DCL4 (40% increase) and AGO1 (20% increase). This suggests that AtSAP5 is still important for the maintenance of DCL4 and AGO1 expression.
Collectively, our data suggests that although Pha13 and Pha21 both participate in SA regulated immunity, they act differently in triggering downstream immune responsive genes.
Pha13 and Pha21 showed no transcriptional activation activity by Y2H assay
Although Pha21 and Pha13 are involved in expression of PhaGRXC9, PhaRdR1, PhaDCL4 and PhaAGO1, our Y2H analysis showed that Pha21 and Pha13 have no transcriptional activation activity (Fig. 5B).
A20 confers most E3 ligase and ubiquitin binding activity but the AN1 domain also plays important roles in expression of immune responsive genes
Similar to other analyzed A20/AN1 proteins, our biochemical assay indicated that both Pha21 and Pha13 confer E3 ligase activity and ubiquitin binding ability, and A20 domain of Pha21 and Pha13 is the major domain exhibiting self-E3 ligase and ubiquitin binding activity (Fig. 6) (44). It has been well demonstrated that human A20 proteins (without an AN1 domain) regulate the master immune transcription factor NF-kB through ubiquitin editing activity including E3 ligase and ubiquitin binding ability on different substrates (53). It is likely that Pha21 and Pha13 may function in a similar manner to human A20 indirectly, rather than directly function as transcription factors. Our previous domain functional analysis revealed that both the A20 and AN1 domains of Pha13 are required for expression of PhaRdR1, PhaGRXC9, and virus accumulation, and the AN1 domain of Pha13 is involved in the expression of PhaNPR1 (44). Here we also showed that A20 and AN1 domains of Pha21 are both required for expression of PhaGRXC9 and CymMV accumulation, and the AN1 domain plays a more important role in the expression of PhaDCL4 (Fig. 11). This indicates that although A20 confers most E3 ligase and ubiquitin binding activity, the AN1 domain also plays important roles in expression of immune responsive genes. Interestingly, our data suggests that neither domain of Pha21 is required for the expression of PhaAGO1, which suggests that protein regions other than A20 and AN1 domain in Pha21 are also involved in the expression of PhaAGO1 (Fig. 11A).
A20/AN1 proteins serve as important modulators in plant antiviral immunity
Our previous data and the findings presented here indicate that A20/AN1 protein-mediated antiviral immunity is conserved among plants, and A20/AN1 proteins may work alone (AtSAP5) or in a cooperative manner (Pha13 and Pha21) in antiviral immunity to induce the expression of the SA-mediated immune responsive genes, including NPR1, NPR1-independent oxidoreductases gene (GRXC9), and genes involved in the RNAi pathway (RdR1, AGO1 and DCL4) (Fig. 12).
MATERIALS AND METHODS
Plant materials and growth conditions
Orchid variety, Phalaenopsis aphrodite var. Formosa, was purchased from Taiwan Sugar Research Institute (Tainan, Taiwan). All orchid plants we used including
P. aphrodite, P. equestris and transgenic P. equestris (35S::FLAG-Pha21) were first analyzed for the infection with two prevalent orchid viruses, Odontoglossum ringspot virus (ORSV) and CymMV, as detected by RT-PCR with primer pairs, ORSV-F/ORSV-R and CymMV-F/CymMV-R (Table 1). Plants free from ORSV and CymMV were maintained in greenhouse conditions with a controlled 12-h photoperiod (200 μmol m-2s-2) at 25°C/25°C (day/night). The wild-type (WT) Arabidopsis WT (Col-0) and all transgenic Arabidopsis were maintained in a greenhouse with a controlled 12-h photoperiod (200 μmol m-2s-2) at 22°C/22°C (day/night) for three to four weeks before analysis. Cucumber mosaic virus (CMV) isolate 20 was maintained in the Arabidopsis (Col-0) as inoculation source in our study.
Sequence analysis
The amino acid sequence of Pha21 was analyzed by the PROSITE database, ExPASy Proteomics Server (http://ca.expasy.org/), and Conserved Domain Database of NCBI database (http://www.ncbi.nlm.nih.gov/).
Phytohormone treatment
Sodium salicylate (50 mM) (Sigma, St. Louis, MO, USA), methyl jasmonate (45 μM) (Sigma), and aminocyclopropanecarboxylic acid (660 μM) (Sigma) were directly rubbed on leaves of P. aphrodite by cotton swab. Leaf samples were collected at 0 h, 1 h, 3 h, 6 h, 12 h, 24 h, 48 h and 72 h after treatment.
RNA isolation and real-time quantitative RT-PCR (qRT-PCR) detection
Total RNA was extracted using the TOOLSmart RNA Extractor (BIOTOOLS, Taiwan) as described previously (44). The cDNA template for qPCR was synthesized from 500 ng of DNA-free RNA and oligo (dT) using PrimeScript RT Reagent Kit (Perfect Real Time), following the manufacturer’s instructions (Takara Bio, Shiga, Japan). qPCR was performed using SYBR Green PCR Master Mix (Applied Biosystems, Foster, CA, USA)) with ABI Prism 7500 sequence detection system (Applied Biosystems). Target gene PCR products were sequenced to validate the correct analysis of gene targets. PhaUbiquitin 10 or AtActin was used as an internal quantification control. Quantification of target gene expression was calculated according to the manufacturer’s instructions (Applied Biosystems). The primer pairs used in this study are listed in Table 1.
Droplet digital PCR
Total RNA was extracted using the TOOLSmart RNA Extractor (BIOTOOLS, Taiwan) as described previously (44). For droplet digital PCR (ddPCR), DNA-free RNA (1 μg) and oligodT primer were used for cDNA synthesis with PrimeScript RT Reagent Kit (Perfect Real Time) following the manufacturer’s instructions (Takara Bio). An amount of 5 μl of 25X dilute cDNA was used as a template for ddPCR reaction (total 20 μl) following the manufacturer’s instructions (Bio-Rad, Hercules, CA, USA). The sequence of FAM-labeled Taqman probes and primer pairs of Pha13 and Pha21 for ddPCR are listed in Table 1.
Construction of transient silencing vector
For construction of transient silencing vector of Pha21, the oligonucleotide pairs Pha21-hpRNA-1-F/Pha21-hpRNA-1-R and Pha21-hpRNA-2-F/Pha21-hpRNA-2-R (Table 1) were used to amplify the hairpin dsDNA fragments. The hairpin dsDNA fragments were cloned into the Gateway entry vector pENTR/D-TOPO (Thermo Fisher-Scientific, Waltham, MA, USA) following the manufacturer’s instructions to generate pENTR-Pha21-hpRNA-1 and pENTR-Pha21-hpRNA-2. Then, LR Gateway cloning reaction (Thermo Fisher-Scientific) was conducted to transfer the hairpin RNA fragments from pENTR-Pha21-hpRNA into 35S promoter driven pB7GWIWG2(I) (57) to obtain phpPha21-1 and phpPha21-2. The method for construction of the transient silencing vector of PhaNPR1, PhaDCL4, and PhaAGO1 (phpPhaNPR1, phpPhaDCL4, and phpPhaAGO1) were similar to that described above, except the oligonucleotide pairs PhaNPR1-hpRNA-F/PhaNPR1-hpRNA-R, PhaDCL4-hpRNA-F/PhaDCL4-hpRNA-R, and PhaAGO1-hpRNA-F/ PhaAGO1-hpRNA-R (Table 1) were used to generate the hairpin dsDNA fragments.
Construction of transient overexpression vector of Pha21, Pha21 A20 and/or AN1 domain mutant
To construct Pha21 transient overexpression vector, total P. aphrodite RNA was used as a template to amplify the N-terminus FLAG-tagged Pha21 by RT-PCR with the primer pairs FLAG-Pha21ORF-F/Pha21ORF-R (Table 1). The amplified fragment was cloned into Gateway entry vector pENTR/D-TOPO (Invitrogen) to generate pENTR-FLAG-Pha21 following the manufacturer’s protocol. Subsequently, LR Gateway cloning reaction (Invitrogen) was performed to transfer the FLAG-Pha21 fragment from pENTR-FLAG-Pha21 into the 35S promoter driven pK2GW7 (57), designated pPha21-oe. For generation of A20 and/or AN1 mutant on pPha21-oe (Fig. 1), site-directed mutagenesis was conducted by QuikChange Site-Directed Mutagenesis Kit (Agilent Technologies, Santa Clara, CA, USA) and pPha21-oe was used as a template. For A20 mutant, we substituted the conserved 3rd and 4th cysteine to glycine at A20 (C35G and C38G). For AN1 mutant, we substituted the conserved 3rd cysteine and 1st histidine to glycine at AN1 (C113G and H123G). The A20 mutated clones, AN1 mutated clones, and the A20 and AN1 mutated clones, were designed as pPha21A20m, pPha21AN1m, and pPha21A20mAN1m, respectively. Primer pairs used for site directed mutagenesis are listed in Table 1.
Transgenic Phalaenopsis orchid
For construction of overexpression vector to generate transgenic Phalaenopsis orchid, the FLAG-Pha21 fragment was transferred from pENTR-FLAG-Pha21 (described above) into 35S promoter driven binary vector, pH2GW7 (57), to obtain pHPha21. pHPha21 was used to generate transgenic P. equestris orchid using the method described by Hsing et al. (58).
Agroinfiltration
Agroinfiltration was conducted as previously described by (51) with some modification. Briefly, A. tumefaciens C58C1 (pTiB6S3ΔT)H competent cells were transformed with pCambia-CymMV, pB7GWIWG2, pK2GW7 and their derivatives using an electroporation system (Bio-Rad Laboratories, Hercules, CA, USA). Then, the A. tumefaciens strains were incubated at 28°C until the optical density, OD600, reached 0.8–1.0. After centrifugation, cells were resuspended in 20 ml AB-MES medium (17.2 mM K2HPO4, 8.3 mM NaH2PO4, 18.7 mM NH4Cl, 2 mM KCl, 1.25 mM MgSO4, 100 μM CaCl2, 10 μM FeSO4, 50 mM MES, 2% glucose (w/v), pH 5.5) with 200 μm acetosyringone (59), and cultured overnight. The overnight culture was centrifuged (3000 rpm, 10 min, in room temperature), supernatant was removed and the A. tumefaciens culture was resuspended in 2 ml of infiltration medium containing 50% MS medium (1/2 MS salt supplemented with 0.5% sucrose (w/v), pH 5.5), 50% AB-MES and 200 μm acetosyringone (59). The infiltration medium containing the transformed A. tumefaciens was applied for infiltration.
Cymbidium mosaic virus (CymMV) and Cucumber mosaic virus (CMV) inoculation and accumulation assay
To assay the effect of Pha21, PhaDCL4, or PhaAGO1 in CymMV accumulation, we first inoculated CymMV in P. aphrodite through the infiltration of A. tumefaciens carrying pCambia-CymMV (described above) in the leaf tip of P. aphrodite. The CymMV-infected P. aphrodite was maintained at least 14 days before further analysis. To assay the effect of transient silencing (Pha21, PhaDCL4, or PhaAGO1) or transient overexpression (Pha21, or the derived mutants), A. tumefaciens carrying the control vector pB7GWIWG2 (for silencing), pK2GW7 (for overexpression), silencing vectors or overexpression vectors (described above) were infiltrated into the leaves. After agroinfiltration, a pair of disks (6 mm diameter) were immediately (defined as 0 dpi) collected from both the control and assay vector infiltrated regions. After 5 dpi, another pair of disks were collected from the same infiltrated region. Total RNA extracted from the samples was used as a template to analyze the accumulation of CymMV by use of qRT-PCR. The fold change of CymMV accumulation at 0 dpi to 5 dpi was calculated for relative quantification.
For inoculation with CMV, CMV-infected Arabidopsis leaves were ground with 0.01 M potassium phosphate buffer by pestle and mortar for use as the inoculation source. Four-week-old Arabidopsis leaves were inoculated mechanically (pre-dusted with 300-mesh Carborundum) with the CMV inoculation source. After 14 dpi, the disease symptoms were observed, and three disks from three different distal leaves were collected for CMV accumulation analysis by use of qRT-PCR. Primers used for CMV detection are listed in Table 1.
Preparation and transfection of protoplasts
For construction of vectors used for subcellular localization analysis, the primer pairs, Pha21-ORF-F/Pha21-ORF-R and Pha21-ORF-F/Pha21-ORF-non-stop-R (Table 1) were used to amplify 2 sets of Pha21 ORF (with or without a stop codon). The amplified fragment was cloned into Gateway entry vector pENTR/D-TOPO (Invitrogen) to generate pENTR-Pha21 and pENTR-Pha21-non-stop following the manufacturer’s protocol. Subsequently, LR Gateway cloning reaction (Invitrogen) was performed to transfer ORF fragment of Pha21 from pENTR-Pha21 into p2FGW7 driven by 35S promoter (57) to obtain N-terminal GFP fused clones (pG-Pha21). To obtain C-terminal GFP-fused clones (pPha21-G), we transferred and pENTR-Pha21-non-stop into p2GWF7.
Protoplast isolation and transfection were as described by Lu et al. (51). Transformed protoplasts were detected for florescence signals by confocal microscopy (Zeiss LSM 780, plus ELYRA S.1) with excitation at 488 nm and emission at 500 to 587 nm for GFP, and excitation at 543 nm and emission at 600 to 630 nm for mCherry.
Transcriptional activation ability assay
The Pha13 and Pha21 ORF was amplified by PCR with gene specific primer pairs, NdeI-Pha13-F/EcoR1-Pha13-R and NdeI-Pha21-F/BamH1-Pha21-R, respectively (Table 1). The amplified ORF fragment was then cloned into the pGBKT7 vector (Takara Bio) to generate the pGBKT7-Pha13 and pGBKT7-Pha21 constructs, which fused to Gal4 BD sequence. These constructs were individually transformed into AH109 yeast strain (Takara Bio) and the self-activation ability was analyzed. The transformation and selection procedure was performed following the Yeast Protocols HandBook (Takara Bio). Transformants were selected on SD/-Trp (tryptophan) medium. The expression of reporter genes in yeast were tested on different selection media; SD/-Trp/Aureobasidin was used to test for activation of the AUR1-C (inositol phosphoryl ceramide synthase) reporter against Aureobasidin A, SD/-Trp/X-α-Gal medium was used to test for activation of α-galactosidase, and SD/-Trp/-Ade/-His medium was used to test for induction of the ADE2 and HIS3 reporters. The yeast containing Gal4 DNA-BD fused with murine p53 (pGBKT7-53) and the Gal4 AD fused with SV40 large T-antigen (pGADT7-T) from the Matchmaker Gold Yeast Two-Hybrid System kit (Takara Bio) were used as positive control. The yeast containing Gal4 DNA-BD fused with lamin (pGBKT7-Lam) and the pGADT7-T from the Matchmaker Gold Yeast Two-Hybrid System kit (Takara Bio) were used as a negative control.
Construction, expression, and purification of recombinant proteins
Full-length Pha21, or Pha21 A20 and/or AN1 domain mutant fragments were amplified by PCR with primer pairs NdeI-Pha21-F/NotI-Pha21-R (Table 1) using pPha21-oe, pPha21A20m, pPha21AN1m, and pPha21A20mAN1m as templates. The amplified fragments were cloned into the pET24b expression vector (Merck, Darmstadt, Germany) to produce fused C-terminal histidine tag (His-tag) expression plasmids, pETPha21, pETPha21A20m, pETPha21AN1m and pETPha21A20mAN1m. The constructed plasmids were transformed into Escherichia coli strain BL21 (DE3) for protein expression. Bacteria were cultured at 37°C to an OD600 of 0.5, then transferred to 25°C, for 1.5 h. Then, isopropylthio-β-galactoside (IPTG; Sigma) was added to a final concentration of 1 mM for protein induction. His-tagged recombinant protein was purified by TALON Superflow (GE Healthcare Life Sciences, Pittsburgh, PA, USA) according to the manufacturer’s description. His-tagged recombinant protein was eluded with 250 mM imidazole (Sigma).
E3 ubiquitin ligase activity assay
In vitro ubiquitination assays were performed as described by (45) with modification. An amount of 3 μg purified His-tagged recombinant proteins (Pha21 or derived mutants) was used for each ubiquitination reaction. Reactions were incubated at 30°C for 3 hours and analyzed by SDS-PAGE followed by immunoblot analysis. Blots were probed using anti-FLAG antibodies (Sigma) followed by HRP conjugated goat anti-mouse antibodies (GE Healthcare Life Sciences).
Yeast two-hybrid assay
For the ubiquitin binding ability assay, the pGBKT7-Pha21 (described above) was used as a bait vector. To map the region of Pha21 for ubiquitin binding, a deletion mutant of Pha21 was generated by PCR amplification with the primer pairs described in Table 1. The PCR-amplified fragment was individually cloned to pGBKT7 and used as a bait vector (Fig. 6B). Full length of orchid ubiquitin was amplified by RT-PCR with the primer pair, NdeI-PhaUBQ-F/BamHI-PhaUBQ-R, and cloned into pGADT7 vector (Takara Bio) to generate pGADT7-UBQ as a prey vector. pGBKT7-Pha21 and pGADT7-UBQ was co-transformed into AH109 yeast strain for yeast two-hybrid assay by using the Make Your Own “Mate & Plate” Library System (Takara Bio) following the user manual. Yeast strains containing the appropriate bait and prey plasmids were cultured in liquid 2-dropout medium (leucine- and tryptophan-) overnight. The overnight yeast culture was diluted to an OD600 of 0.06 and spotted on selection plates (containing histidine-, leucine-, tryptophan- for growth assay.
For the interaction analysis between Pha21 and Pha13, full-length Pha13 fragment was amplified by PCR with primer pair NdeI-Pha13-F/EcoR1-Pha13-R (Table 1). The Pha13 fragments were cloned into pGBKT7 and pGADT7 vector to generate pGBKT7-Pha13 as the bait vector and pGADT7-Pha13 as the prey vector. The full-length of Pha21 previously used to generate the pGBKT7-Pha21 (described above) was also cloned into pGADT7 to generate pGADT7-Pha21 as prey vector. The pGBKT7-Pha21, pGBKT7-Pha13, pGADT7-Pha21, and pGADT7-Pha13 were co-transformed into AH109 yeast strain for yeast two-hybrid assay as described above. The pGBKT7-53 and pGADT7-T from the Matchmaker Gold Yeast Two-Hybrid System kit were used as a positive control.
Accession numbers
Pha21 (PATC144963), Pha13 (PATC148746), PhaPR1 (PATC126136), PhaNPR1 (PATC135791), PhadR1 (PATC143146), PhaRdR2 (PATC124544), PhaRdR6 (PATC131836), PhaDCL2 (PATC143544), PhaDCL4 (PATC150652), PhaAGO1 (PATC157237), PhaAGO10 (PATC093469), PhaGRXC9 (PATC068819), PhaUBQ10 (PATC230548), PhaJAZ1 (PATC141437), PhaACO2 (PATC139319), AtActin (At3G18780), AtSAP5 (AT3G12630), DCL4 (AT5G20320), AGO1 (AT1G48410), OsSAP3 (LOC_Os01g56040.1), OsSAP5 (LOC_Os02g32840.1)
ACKNOWLEDGMENTS
We thank Shu-Chen Shen from the Confocal Microscope Core Facility at Academia Sinica for assistance in confocal microscopy images and Chii-Gong Tong from the Gene Transgenic Room at Academia Sinica for orchid transformation. This work was supported by grants from Academia Sinica, Taipei, Taiwan and the Ministry of Science and Technology (105-2313-B-001-004-MY3) of Taiwan. The funders had no role in study design, data collection and interpretation, or the decision to submit the work for publication