The C4 protein from Curtovirus is known as a major symptom determinant, but the mode of action of the C4 protein remains unclear. To understand the mechanism of involvement of C4 protein in virus–plant interactions, we introduced the C4 gene from Beet severe curly top virus (BSCTV) into Arabidopsis under a conditional expression promoter; the resulting overexpression of BSCTV C4 led to abnormal host cell division. RKP, a RING finger protein, which is a homolog of the human cell cycle regulator KPC1, was discovered to be induced by BSCTV C4 protein. Mutation of RKP reduced the susceptibility to BSCTV in Arabidopsis and impaired BSCTV replication in plant cells. Callus formation is impaired in rkp mutants, indicating a role of RKP in the plant cell cycle. RKP was demonstrated to be a functional ubiquitin E3 ligase and is able to interact with cell-cycle inhibitor ICK/KRP proteins in vitro. Accumulation of the protein ICK2/KRP2 was found increased in the rkp mutant. The above results strengthen the possibility that RKP might regulate the degradation of ICK/KRP proteins. In addition, the protein level of ICK2/KRP2 was decreased upon BSCTV infection. Overexpression of ICK1/KRP1 in Arabidopsis could reduce the susceptibility to BSCTV. In conclusion, we found that RKP is induced by BSCTV C4 and may affect BSCTV infection by regulating the host cell cycle.
Geminiviruses are a group of single-stranded DNA viruses that infect a large range of plants and cause considerable agricultural losses. Their small genomes are amplified through a rolling-circle mechanism in plant cell nuclei (Gutierrez, 1999; Hanley-Bowdoin et al., 2000; Lazarowitz, 1992). The limited coding capacity of geminiviruses means that their replication is heavily dependent on host factors. However, many infected cells are differentiated cells that have exited the cell cycle and cannot support DNA replication. Therefore, geminiviruses must cause the host cells to re-enter the cell cycle to create an environment suitable for replication (Morra and Petty, 2000; Nagar et al., 1995, 2002).
There is evidence to support the proposition that amplification of geminiviruses is coupled with DNA replication of host cells. For instance, dsDNA replicative forms are much more abundant in S-phase nuclei of cultured cells (Accotto et al., 1993). Proliferating cell nuclear antigen (PCNA), an accessory protein of DNA polymerase δ, is induced in differentiated cells due to the presence of the Rep protein of Tomato golden mosaic virus (TGMV) (Nagar et al., 1995). Some viral replication-associated proteins (Begomovirus Rep or Mastrevirus RepA) may interfere with the host cell cycle by interacting with RBR, an activator of G/S-phase transition in plant cells (Ach et al., 1997; Grafi et al., 1996; Kong et al., 2000; Liu et al., 1999; Xie et al., 1995). The physical interaction between Tomato yellow leaf curl Sardinia virus (TYLCSV) REn/Rep and tomato PCNA may also contribute to the cell-cycle switch (Castillo et al., 2003). Recently, microarray analysis of the Arabidopsis transcriptome in response to Cabbage leaf curl virus (CaLCuV) infection has suggested that geminiviruses modulate the plant cell cycle by differential impacts on the CYCD/RBR/E2F regulatory network and promotion of progression into the endocycle (Ascencio-Ibanez et al., 2008).
Infection by many monopartite geminiviruses, such as Beet curly top virus (BCTV), leads to a vein-swelling phenotype associated with abnormal cell division in host plants (Stanley et al., 1986). BCTV C4 protein has been proposed as a major determinant of pathogenesis that contributes to the symptom of the host in BCTV infection (Stanley and Latham, 1992). In addition, expression of BCTV C4 protein in transgenic Nicotiana benthamiana results in ectopic cell division (Latham et al., 1997; Piroux et al., 2007). Previous studies also uncovered differential roles of C4 protein homologs encoded by some bipartite geminiviruses. For instance, the AC4 protein of TGMV may be a virus movement factor (Pooma and Petty, 1996), and the AC4 protein of African cassava mosaic virus (ACMV) may be a gene silencing suppressor (Chellappan et al., 2005; Vanitharani et al., 2004).
In eukaryotes, the mechanisms of cell-cycle regulation are conserved. Cell-cycle progression in plants is controlled by cyclin-dependent kinase (CDK)/cyclin complexes (De Veylder et al., 2007; Dewitte and Murray, 2003). The activities of CDK/cyclin complexes may be negatively regulated by CDK inhibitors (CKIs) (Sherr and Roberts, 1999; Verkest et al., 2005b). There are seven proteins in Arabidopsis related to the mammalian protein CKI p27Kip1, known as Kip-related proteins (KRPs) or interactors/inhibitors of Cdc2 kinases (ICKs) (De Veylder et al., 2001; Jakoby et al., 2006). Overexpression of ICK/KRPs in Arabidopsis blocks plant growth through reduction of cell numbers (De Veylder et al., 2001; Zhou et al., 2003). In mammalian cells, the degradation of CKI p27Kip1 is mediated by two pathways, the nuclear SCF ubiquitin ligase Skp2 pathway, and the cytoplasmic ubiquitin ligase KPC (Kip1 ubiquitination-promoting complex) pathway that includes KPC1 and KPC2 (Kamura et al., 2004; Kotoshiba et al., 2005). Recently, a similar redundant control pathway has been found in plants (Ren et al., 2008). However, the degradation mechanism of ICK/KRPs might be more complicated, for instance ICK4/KRP6 is degraded by RING-H2 group F1a (RHF1a) and RHF2a, two RING-type E3 ligases, during Arabidopsis gametogenesis (Liu et al., 2008).
This study focuses on C4 protein encoded by Beet severe curly top virus (BSCTV) (formerly named the BCTV CFH strain) (Fauquet et al., 2008; Park et al., 2002, 2004), a species of Curtovirus. In this report, we describe a geminivirus BSCTV C4-inducible Arabidopsis protein, similar to human KPC1, named AtKPC1. During the course of this study, this gene was reported by Ren et al. (2008) in their study of turnover the Arabidopsis cell-cycle inhibitor KRP1, and named RKP (related to KPC1). To avoid confusion, the name RKP is used throughout this paper. We found that mutation of RKP in Arabidopsis reduces susceptibility to BSCTV infection and impairs BSCTV replication in plant cells. Callus division is impaired in rkp mutants. In addition, RKP acts as a functional ubiquitin E3 ligase, and ICK/KRPs may be its targets. The protein level of ICK2/KRP2 was decreased in BSCTV infection. Overexpression of ICK1/KRP1 in Arabidopsis could reduce susceptibility to BSCTV. Thus, RKP, which is induced by C4 protein, could affect BSCTV infection by regulating the plant cell cycle.
Expression of BSCTV C4 in Arabidopsis induces abnormal cell division
The C4 protein from BCTV, a close species of BSCTV, is known to be a symptom determinant in BCTV infection, but the molecular function and the role of C4 in virus–plant interactions remains unclear. To reveal the molecular function of the C4 protein in plant cells, we cloned the C4 gene from BSCTV and expressed it in Arabidopsis. To achieve this, the 264 bp coding sequence (CDS) of the C4 gene, encoding a small protein containing 87 amino acids (Figure 1a), was introduced using the flower-dip transformation method into a transgenic vector under the control of the CaMV 35S promoter to produce transgenic plants constitutively overexpressing C4. However, no transgenic plants were obtained in several individual transformations. Instead, we used green callus-like tissues that were found on the selection plates under light growth conditions (Figure 1b). When viewed under a microscope, the callus cell type was similar to that of callus obtained by the normal tissue culture method. This indicates that overexpression of BSCTV C4 may induce abnormal cell division/differentiation, and that normal transgenic lines could not be generated. To resolve this problem, we cloned the C4 gene into the pER8 vector, in which expression of C4 protein is under the control of the lexA-VP16-ER (XVE)-inducible promoter, whose expression is induced by treatment of estrodiol (Zuo et al., 2000). We obtained transgenic plants in which expression of C4 was not detected under standard growth conditions but was detected on medium including the inducer estrodiol (Figure 1c). Interestingly, growth of pER8-C4 plants was normal on Murashige and Skoog (MS) plates but was blocked when C4 was induced. Curled cotyledons and short roots were observed when the seeds of pER8-C4 plants were germinated on MS plates including 2 μm estrodiol (Figure 1d). More surprisingly, after several weeks under estrodiol induction, the blocked seedlings turned into calli structures similar to those that constitutively express C4 (Figure 1e). These data indicate that C4 could induce abnormal cell division in transgenic plants.
RKP expression is induced by C4 protein and BSCTV infection
To dissect the molecular function of C4 protein in plants, we examined the effect of the C4 protein on expression of host genes. We performed genome-wide expression studies using ATH1 Arabidopsis whole-genome microarray chips. mRNA was isolated from 2-week-old seedlings of pER8-C4 and pER8-GFP (comprising the GFP gene cloned into the same vector) treated with estrodiol for 16 h, and subjected to microarray analysis. Compared with the pER8-GFP control, expression levels of various genes were increased or decreased in pER8-C4 plants. Among them, the expression of a few genes related to cell division was found to be altered. Several interesting genes were chosen for study, and this paper focuses on the function of one gene changed in microarray.
In the microarray data, the expression of one gene, which is identical to the putative gene At2g22010 in the Arabidopsis database, attracted our attention as its expression was approximately eight times greater than that in the control. BLAST analysis demonstrated that this gene might be the homolog of KPC1 in humans. Thus we named it AtKPC1. The amino acid sequence identity between KPC1 and AtKPC1 is 27% overall, but there are two highly conserved domains, including a SPRY domain, which acts as a protein interaction region, in the N-terminus, and a RING finger domain in the C-terminus. During our research, this gene was named RKP (related to KPC1) by Ren et al. (2008). In their study, the protein level of the cell-cycle negative regulator KRP1 was found to increase in the rkp mutant. To avoid confusion, we have adopted the gene name RKP here also.
To verify the array data, we performed semi-quantitative RT-PCR to check the expression of RKP. The expression of RKP was indeed increased in pER8-C4 plants treated with estrodiol for 16 h, under which conditions the C4 gene was expressed (Figure 2a). This result was further confirmed by a transient assay in which the promoter of the RKP was fused to the GUS reporter gene. This construct was co-transfected with either pCambia1300221-C4 or the control vector into Arabidopsis mesophyll protoplasts. Compared to the vector control, GUS activity in protein extracts from protoplasts was obviously enhanced when C4 was co-expressed (Figure 2b). Together, these results indicated that RKP expression is induced by C4 protein. To examine whether RKP expression is also induced during the infection of BSCTV carrying the C4 gene, RNA extracted from local leaves (rosette leaves) of Arabidopsis inoculated by BSCTV or the vector control was used for semi-quantitative RT-PCR analysis. Compared to the vector control, the RNA level of RKP increased in leaves infected by BSCTV (Figure 2c). Accordingly, BSCTV could induce RKP expression in Arabidopsis, and RKP may play a role in BSCTV infection.
Mutation of RKP reduces the susceptibility to BSCTV in Arabidopsis
BSCTV C4 expression, as well as BSCTV infection, induced expression of the endogenous gene RKP. To show whether RKP plays a role in BSCTV infection, the reverse genetics approach was used. Three independent T-DNA insertion lines, rkp-1 (SAIL_3_E03), rkp-2 (WiscDsLox466C1) and rkp-3 (SALK_121005), which showed loss of function of RKP, were obtained from the ABRC seed stock center. The T-DNA insertion positions are indicated in Figure 3(a) and homozygous mutants were verified by PCR using RKP gene-specific and T-DNA border primers (Figure S1). All alleles were confirmed by RT-PCR. Full-length RKP was not detected in any of the three mutants (Figure 3b). Because the RING finger domain is located at the end of the RKP C-terminus, the ligase activity of the protein should be affected in all the mutants.
Under standard growth conditions, we did not observe any apparent morphological phenotype in rkp mutants. To investigate whether loss of function of RKP affects BSCTV infection, the rkp mutant and control plants were agro-inoculated with BSCTV (Grimsley et al., 1986). Conventional agro-inoculation of Arabidopsis by BSCTV requires manual infiltration through wounds on individual plants (Lee et al., 1994; Park et al., 2002). In our study, the procedure was improved to be more efficient (see Experimental procedures). The number of plants with symptoms was counted every day post-inoculation. In independent repeated experiments, all three rkp alleles displayed delayed symptom appearance and a reduced ratio of symptomatic plants compared to control plants (Figure 3c). Total DNA from the whole plants (overground tissues of the inoculated plants) at various time points post-inoculation was extracted and subjected to a DNA gel-blot using the whole BSCTV genome DNA as the probe. As a result, the accumulation of viral DNA in rkp mutants was approximately 40–50% of that in control plants 15 days post infection in repeated experiments (Figure 3d). Together, these results indicate that mutation of RKP results in reduced susceptibility to BSCTV infection in Arabidopsis.
BSCTV DNA replication is decreased in rkp mutants
Virus resistance in plants has at least two possible reasons: impairment of virus DNA replication leading to reduction of virus accumulation, or restriction of virus movement from cell to cell to reduce the spread of virus. To discover the reason for reduced susceptibility in rkp mutants, several experiments were performed. First, we examined viral DNA accumulation in the inoculated rosette leaves of symptomatic plants after BSCTV infection. In this analysis, because leaves that expanded after infection were excluded, BSCTV DNA accumulation must primarily be a result of viral DNA replication. DNA gel results indicated that the level of viral DNA in the inoculated rosette leaves of rkp mutants was lower (approximately 60% of samples from leaves 16 days post infection in repeated experiments) than that in control plants (Figure 3e). This indicates that virus replication is affected in rkp mutants.
To confirm this result and to exclude an effect of short-distance movement of the virus in the inoculated rosette leaves, a transient replication assay was performed in the Arabidopsis mesophyll protoplasts. In this assay, protoplasts are free of cell walls and there is no viral movement among the cells. The mesophyll protoplasts from the rkp mutant and control plants were transfected by the BSCTV construct. Total DNA was extracted from the protoplasts at various times after transfection, and was subjected to DNA gel blots to examine the newly replicated viral DNA. The data showed that virus replication was reduced to approximately 40% in protoplasts of the rkp-1 mutant (Figure 3f). These two results prove that replication of BSCTV is impaired in the absence of RKP.
Callus formation is impaired in rkp mutants
Several reasons led us to propose that RKP is involved in plant cell-cycle control. First, replication of viral DNA is known to be associated with the host plant cell cycle. Second, RKP is the homolog of the cell-cycle regulator KPC1 that negatively regulates p27Kip1 in humans. Third, in this study, we have shown that the replication of BSCTV is impaired in rkp mutants. Thus, we would like to examine whether loss of function of RKP affects the plant cell cycle. Because no apparent developmental phenotypes were observed in rkp mutants, a callus formation assay was used to examine the cell-cycle process (Kim et al., 2006; Riou-Khamlichi et al., 1999). Hypocotyls of rkp mutants and wild-type plants were cut into segments and were cultured on MS medium supplemented with 2,4-dichlorophenoxyacetic acid (2,4-D) and 6-benzylaminopurine (6-BA) (Riou-Khamlichi et al., 1999). Calli were actively induced on both wild-type and mutant segments. However, callus growth was reduced on the rkp mutant segments compared to that on the control segment (Figure 4a). Similar results were obtained when root segments were used for callus induction (Figure 4b). These results demonstrate that cell division is impaired in rkp mutants under hormone treatments.
Interestingly, when the hypocotyls and roots of transgenic plants carrying an RKP promoter–GUS construct were cultured on callus formation mediums, the GUS activity increased significantly (Figure 4c,d). This suggests that expression of RKP is induced in dividing cells, and supports the hypothesis that RKP might be involved in plant cell-cycle regulation.
RKP is a functional E3 ligase
Previous research has shown that RING finger-containing proteins play roles as ubiquitin E3 ligases (Xie et al., 2002). The C-terminus of RKP contains a conserved C3HC4-type RING domain including conserved Cys and His residues. Therefore, we assume that RKP has ubiquitin E3 ligase activity. To confirm our hypothesis, RKP fused with the GST tag was expressed in Escherichia coli, and was purified using GST affinity beads from the soluble fraction. In the presence of wheat (Triticum aestivum) E1 and human E2 (UBCh5b), ubiquitination activity was observed in the presence of purified GST–RKP (Figure 5a). This assay showed that RKP is a functional ubiquitin E3 ligase that may regulate degradation of targeted proteins by the 26S proteasome pathway.
RKP interacts with ICK/KRPs
Human KPC1 is known to act as a ubiquitin ligase to regulate the degradation of p27Kip1 at G1 phase, therefore we analyzed the similarity of protein sequences between p27Kip1 and the Arabidopsis ICK/KRP family. Seven ICK/KRP proteins contained a conserved domain at their C-termini. Interestingly, this domain is similar to the p27Kip1 N-terminal domain that interacts with human KPC1. This suggests that RKP may interact with ICK/KRP proteins. An in vitro pull-down assay was performed to test our hypothesis. All members of the ICK/KRP family except ICK3/KRP5 could be produced as fusion proteins with GST tags in E. coli, and were purified using affinity beads. RKP labeled with 35S on its methionine residues was generated by in vitro transcription and translation using wheatgerm extracts. After binding and washing, we found that RKP was able to interact with all six ICK/KRP proteins in vitro but not with the GST protein itself as a control (Figure 5b).
Because it has been reported recently that degradation of KRP1 is mediated by RKP (Ren et al., 2008), we examined whether the degradation of other ICK/KRPs is also regulated by RKP. ICK2/KRP2, a cell-cycle inhibitor in G1/S phase and a regulator of leaf development (De Veylder et al., 2001; Verkest et al., 2005a), was chosen for a detailed study. There was no obvious difference in the RNA level of myc-ICK2/KRP2 between the rkp-1 and control plants. However, the protein level of myc-ICK2/KRP2 was higher in the mutant than in the wild-type (Figure 5c). This result is similar to that obtained for ICK1/KRP1 in the rkp mutant (Ren et al., 2008). Therefore, the degradation of at least two ICK/KRP family members is regulated by RKP.
Inter-play between BSCTV infection and the ICK/KRP protein level
Because ICK/KRP proteins may be the targets of RKP in certain situations and BSCTV can induce the expression of RKP, the protein levels of ICK/KRP proteins may be supposed to be decreased during BSCTV infection. We used ICK2/KRP2 to prove this hypothesis. 35S-myc-ICK2/KRP2 and 35S-GFP plasmids were co-transfected with either pCambia-BSCTV or the pCambia vector control into Arabidopsis mesophyll protoplasts. Total proteins were extracted from the protoplasts 1-day post-transfection and applied to protein gel blots. The level of myc-ICK2/KRP2 proteins was decreased during BSCTV infection compared to the vector control (Figure 5d), whereas the level of myc-ICK2/KRP2 mRNA was not affected by BSCTV in this assay. Therefore, the protein level of ICK2/KRP2 is affected by BSCTV infection.
An interesting question is whether the changes in ICK/KRP levels in plants affect the susceptibility to BSCTV. It has been reported that overexpression of ICK/KRP proteins in plants results in smaller size and serrated leaves by inhibition of cell-cycle progression (De Veylder et al., 2001; Zhou et al., 2002). To verify the effect of expression of ICK/KRP gene on BSCTV infection, the ICK1/KRP1 gene under the control of the 35S promoter was introduced into Arabidopsis, and several transgenic lines with various levels of leaf size reduction were obtained. To avoid experimental error, a transgenic line with serrated leaves but a similar leaf size to wild-type control was used for virus infection (Figure S2a,b). In independent repeated experiments, the 35S-ICK1/KRP1 plants displayed delayed symptom appearance and a reduced proportion of symptomatic plants compared to wild-type plants (Figure 5e). This result indicates that overexpression of ICK1/KRP1 in Arabidopsis could reduce susceptibility to BSCTV.
Much work has been performed to analyze the interaction between geminiviruses and host plants, but the mechanism remains unclear. Our results suggest that an Arabidopsis RING-type E3 ligase RKP, which is induced by the C4 protein of BSCTV, affects geminivirus replication by regulation of the cell cycle.
BSCTV C4 protein enhances host cell division
It has been suggested that geminiviral DNA replication is coupled to host cell-cycle transition due to the requirement for cellular factors (Gutierrez, 2000a). In our study, the callus-like tissues produced by overexpression of C4 in Arabidopsis suggest that abnormal host cell division is induced by the BSCTV C4 protein.
Previous research has shown that infection by most geminiviruses (bipartite Begomoviruses and Mastreviruses) does not lead to abnormal cell division and an increase in cell numbers. However, infection by curtoviruses, such as BCTV, is associated with morphological changes (for instance, vein swelling and leaf curling) due to interference with the host cell cycle (Gutierrez, 2000b). It is therefore interesting to determine why cell division is induced and whether it is necessary for virus amplification. Mutation analysis of BCTV genes suggests that the molecular basis for infection symptom is dependent on the viral protein BCTV C4 (Stanley and Latham, 1992). Abnormal cell division resulting from expression of BCTV C4 in Nicotiana benthamiana confirms that this gene alone can initiate host cell division (Latham et al., 1997). This evidence supports our hypothesis that cell division is enhanced in transgenic Arabidopsis expressing BSCTV C4.
Conversely, C4 protein from Tomato yellow leaf curl virus (TYLCV) was reported to be involved in virus movement, because the C4 mutant retained the ability to replicate in tomato protoplasts and was able to infect Nicotiana benthamiana plants systemically with a lower virus DNA level (Jupin et al., 1994; Rojas et al., 2001). It is interesting to determine whether the role of C4 in cell division and virus movement are compatible.
The molecular mechanism of action of BSCTV C4 is unclear, but previous studies have provided some insight on how C4 interferes with the host cell cycle. First, C4 might interact physically with some cell-cycle regulators in a manner similar to Rep/RepA proteins, which interact with RBR, but no cell-cycle factor has been reported. Second, C4 may affect the metabolism or distribution of hormones. For example, the brassinosteroid signaling pathway may be regulated by BSCTV C4 (Piroux et al., 2007). Third, C4 may regulate the expression of host genes directly or indirectly to control the cell cycle. For instance, the expression of RKP is upregulated by BSCTV C4. C4 may control transcription or act as a gene silencing suppressor to regulate the RNA level of some host genes (Vanitharani et al., 2004).
RKP acts as a cell-cycle regulator in Arabidopsis
Because C4 induces abnormal cell division, we suspect that expression of several important cell-cycle regulators must be altered in the presence of C4. After analysis of microarray data, RKP was chosen as a candidate due to its similarity to human KPC1, which acts as a regulator of p27kip1 during the cell cycle (Kamura et al., 2004). Therefore, we speculate that RKP might be involved in regulation of the plant cell cycle in Arabidopsis. There are no apparent morphological phenotypes for the rkp mutants, although there is only one copy for RKP in the Arabidopsis genome. This finding may be interpreted in two ways. First, there may be redundant pathways. In human cells, the cyclin-dependent kinase inhibitor p27Kip1 is degraded at the G0–G1 transition of the cell cycle by two ubiquitin/proteasome pathways, independently mediated by the nuclear ubiquitin ligase SCF (Skp2) and the cytoplasmic ubiquitin ligase KPC complex (Kamura et al., 2004). Because these two pathways appear to be redundant, the Skp2 homolog in Arabidopsis may work as a regulator by itself to retain cell-cycle balance even in the absence of RKP. The other reason may be that RKP plays a role in regulating the cell cycle under special conditions. RKP expression is low under normal conditions, but is induced dramatically by the BSCTV C4 protein. The callus formation experiment supports the hypothesis that cell division is impaired in rkp mutants under hormone treatments. At the same time, the GUS assay suggests that RKP expression is highly induced in dividing cells.
Human KPC1 is a functional ubiquitin ligase that regulates the degradation of p27kip1 directly (Kamura et al., 2004; Kotoshiba et al., 2005). Our data demonstrate that RKP also has ubiquitin ligase activity in vitro. The ICK/KRP protein family, known cell-cycle inhibitors, share a conserved domain with p27kip1. Because this domain could interact with cell cycle-dependent kinase as well as KPC1 in human cells, ICK/KRPs may be the targets of RKP. Recently, it has been reported that the degradation of ICK1/KRP1 is regulated by Skp2 and RKP in Arabidopsis (Ren et al., 2008). Our data suggest that ICK/KRPs may be the targets of RKP in vitro and in vivo. In plants, the members of the ICK/KRP family may play different roles at different cell-cycle checkpoints and in different tissues (De Veylder et al., 2001; Menges and Murray, 2002; Verkest et al., 2005b). Therefore, the fundamental function of RKP in plant cells needs to be studied in detail. Some other E3 ligases have been reported to regulate the stability of ICK/KRPs with the manner of member specificity (Liu et al., 2008), thus the abundance and specificity of this complexity needs to be further addressed.
RKP may affect BSCTV replication by cell-cycle regulation
When plants are infected by geminiviruses, the cell cycle may be altered in terminally differentiated cells. For example, PCNA, an accessory factor for DNA polymerase, was induced by TGMV in differentiated cells (Nagar et al., 1995). In our experiment, virus DNA replication was reduced in rkp mutants. Therefore, RKP might be involved in a cell-cycle switch affected by BSCTV infection in differentiated cells. RKP upregulation would trigger the degradation of ICK/KRPs and enhance the activity of cell cycle-dependent kinases to accelerate G1/S cell-cycle transitions, which can provide a suitable environment for virus replication. Interestingly, many components of the cell cycle, including cyclin-dependent kinase inhibitors, are the targets of viruses in mammalian cells. For instance, degradation of p27Kip1 is triggered by the Kaposi’s sarcoma virus cyclin–cdk6 complex (Ellis et al., 1999; Mann et al., 1999), and recruitment of SCF Skp2 activity to cyclin A complexes by Epstein–Barr virus protein EBNA3C results in ubiquitination and SCF Skp2-dependent degradation of p27 (Knight et al., 2005). In our experiment, the protein level of ICK2/KRP2 was also found to decrease in BSCTV infection. These results suggest that cyclin-dependent kinase inhibitors are critical factors in the cell cycle, and are always the targets of viruses in both plants and animals.
If the cell cycle is blocked at G1 phase, virus infection may be interrupted. There is no evidence for this hypothesis in plants, but animal virus research can help us to understand virus resistance mechanisms in rkp mutants. In animal cells, cell-cycle interference could affect virus infection. The replication of herpes simplex virus type 1 is inhibited in some temperature-sensitive cell-cycle mutant cells (Yanagi et al., 1978). Regulation of the activity of cyclin-dependent kinases, a key factor in cell-cycle progression, is an effective strategy for virus inhibition. For instance, roscovitine, a cyclin-dependent kinase inhibitor, prevents replication of varicella-zoster virus (Taylor et al., 2004). Human cytomegalovirus replication is inhibited by the expression of a CDK2 dominant negative mutant (Bresnahan et al., 1997). Modification of cyclin-dependent kinase inhibitor (CKI) p21 expression altered HIV-1 infection by regulation of the cell cycle (Zhang et al., 2007a). This evidence from mammalian cell research indicates that cell-cycle arrest inhibits virus replication. In rkp mutants, the C4 protein could not enhance expression of the RKP gene to enhance degradation of ICK/KRPs, even after BSCTV infection, so the activity of CDKs remains inhibited by ICK/KRPs. Delay in the G1/S cell-cycle transition may disturb the BSCTV replication environment, and thus could be a new strategy to reduce susceptibility to viruses in plants (Figure 5f). This model may be valid because expression of ICK1/KRP1 led to the reduction of BSCTV infectivity. There may be further cell-cycle genes related to virus replication and infection. Because the RKP and ICK/KRP genes are important in regulating G1/S transition, plants that show impaired cell-cycle G1/S transition may show reduced susceptibility. However, with respect to virus replication, the influence of cell division and endo-replication may be different and complicated. Thus more evidence is required to establish a precise model for the interaction between the cell cycle and virus infection. In rkp mutants, infection by BSCTV was not abolished but only impaired, which might be due to the redundant pathways for ICK/KRP degradation. If the SCF Skp2 pathway is also destroyed, susceptibility to the virus may decrease more drastically.
C4 protein could induce cell division but appeared not to be related to geminivirus replication (Jupin et al., 1994). The host cell cycle may be affected cooperatively by C4, Rep or other viral proteins. When C4 is mutated, other proteins may regulate the host cell cycle to establish a suitable environment for DNA replication. It is an interesting question whether the expression of RKP is regulated by other geminvirus proteins. The mechanism of RKP expression regulation requires further investigation. Thus, further functional dissection of RKP is necessary for a complete understanding of the interaction between the cell cycle and virus infection in plants.
Plant materials and growth conditions
Arabidopsis thaliana ecotype Columbia was used for this study. Seeds were surface-sterilized with 10% bleach for 20 min and rinsed three times with sterile water. Sterile seeds were suspended in 0.15% agarose and plated on MS medium plus 1.5% sucrose. Plates were stratified in darkness for 2–4 days at 4°C and then transferred to a tissue culture room at 22°C under a 16 h light/8 h dark photoperiod. After 2–3 weeks, seedlings were potted in soil and placed in a growth chamber at 22°C and 70% relative humidity under a 16 h light/8 h dark photoperiod.
Construction of the BSCTV plasmid
An infectious clone of BSCTV (previously named the BCTV CFH strain) (Stenger, 1994; Stenger et al., 1990), pCFH (ATCC number PVMC-6), was obtained from the American Type Culture Collection (ATCC, Manassas, VA, USA). An EcoRI–BamHI fragment (0.8 copy of genome) and an EcoRI–EcoRI fragment (full genome) were introduced from pCFH to binary vector pCambia1300 to generate pCambia-BSCTV, carrying 1.8 copies of the BSCTV genome as a tandem repeat.
Transformation vectors and construction of transgenic plants
To produce 35S-C4 plants, a 264 bp fragment containing BSCTV C4 cDNA was cloned into the vector pCambia1300-221 in which transgene expression is under the control of the CaMV 35S promoter. Inducible C4 plants were generated by cloning C4 cDNA into the vector pER8, in which transgene expression is under the control of the XVE-inducible promoter. For the RKP promoter/GUS fusion construct, a 5′ flanking sequence (2 kb promoter region just upstream of the ATG start codon of RKP) was amplified from genomic DNA by PCR and verified by sequencing. The PCR fragment was cloned into the PstI–XbaI site of binary vector pCambia1300-221 to obtain a transcriptional fusion of the RKP promoter and the GUS coding sequence. For the 35S-ICK1/KRP1 construct, the full-length coding sequence (CDS) of ICK1/KRP1 was cloned into the vector pCanG, modified from pCambia, in which transgene expression is under the control of the CaMV 35S promoter. Transformation of Arabidopsis was performed by the vacuum infiltration method (Bechtold and Pelletier, 1998), using Agrobacterium tumefaciens strain EHA105.
Verification of RKP T-DNA insertion mutants
Seeds of the T-DNA insertion lines rkp-1 (SAIL_3_E03), rkp-2 (WiscDsLox466C1) and rkp-3 (SALK_121005) were obtained from the Arabidopsis Biological Resource Center (Ohio State University, Columbus, OH, USA). Homozygous mutants were identified by PCR from genomic DNA using T-DNA left border primers (LB1, 5′-GCCTTTTCAGAAATGGATAAATAGCCTTGCTTCC-3′; p745, 5′-AACGTCCGCAATGTGTTATTAAGTTGTC-3′; LBb1, 5′-GCGTGGACCGCTTGCTGCAACT-3′) and RKP-specific primers (P3, 5′-CAAGTAATCAGTCTGACCCTG-3′; P4, 5′-TCATGTGCTTCTTTTGTGACC-3′).
Semi-quantitative RT-PCR amplification
To examine the expression of RKP by RT-PCR, DNase I-treated total RNA was denatured and subjected to reverse transcription using SuperScript II (200 units per reaction; Invitrogen, http://www.invitrogen.com/) at 37°C for 1 h followed by heat inactivation of the reverse transcriptase at 70°C for 15 min. To determine the changes in RKP expression in C4 transgenic plants and BSCTV infection, PCR amplification was performed using RKP forward (RT Fw, 5′-TTCGTAGTTACACACTTCAAC-3′) and reverse (RT Rev, 5′-TCATGTGCTTCTTTTGTGACC-3′) primers. To check RKP expression in rkp mutant plants, two pairs of primers were used (P1/P2, 5′-TGG CGCTGGCTTGTCATTTG-3′/5′-GACAAGAACCGAATTGCGTG-3′, and P3/P4). ACTIN1 expression levels were monitored using forward and reverse primers (F, 5′-CATCAGGAAGGACTTGTACGG-3′; R, 5′-GATGGACCTGACTCGTCATAC-3′ to serve as an internal control.
Plants carrying the RKP promoter fused with the GUS gene were treated under various conditions and used for histochemical detection of GUS expression. Materials were stained at 37°C overnight in 1 mg ml−1 5-bromo-4-chloro-3-indolyl-β-d-glucuronic acid (X-Gluc), 5 mm potassium ferricyanide, 5 mm potassium ferrocyanide, 0.03% Triton X-100 and 0.1 m sodium phosphate buffer, pH 7.0. To test the induction of GUS expression on callus formation medium, hypocotyls of 2-week-old transgenic seedlings were transferred onto medium containing 100 ng l−1 2,4-D and 100 ng l−1 6-BA for 4 days; roots of 8-day-old transgenic seedling were transferred onto medium containing 100 ng l−1 2,4-D and 300 ng l−1 kinetin for 6 days. For the protoplast GUS activity assay, the proteins were extracted and detected using a GUS reporter gene activity detection kit (MGT-M0877; FLUOstar OPTIMA, BMG Labtech, http://www.bmglabtech.com/), according to the manufacturer’s instructions.
Rosette leaves of 4-week-old plants were agro-inoculated with BSCTV (Briddon et al., 1989; Grimsley et al., 1986). A suspension of Agrobacterium tumefaciens strain EHA105 (Hood et al., 1993) carrying pCambia1300-BSCTV (carrying 1.8 copies of the BSCTV genome) at a dose of 0.02 (absorbance at 600 nm of 0.02) and mixed with 1% carborundum (320 grit, C192-500, Fisher Scientific, http://www.fishersci.com) was sprayed onto the leaves using an airbrush (SIL.AIR, http://www.silentaire.com) with an air pressure of 75 psi (1 psi = 6.89 kPa). The inoculated plants were covered for one night and grown in another greenhouse at a higher temperature (26°C). Symptoms appeared approximately 10 days after inoculation.
DNA gel blot
For the DNA gel blot, total DNA was extracted using CTAB buffer. Genomic DNA was stained using ethidium bromide as a loading control. After depurination and neutralization, total DNA was transferred to Hybond N+ nylon membranes (Amersham Pharmacia Biotech, http://www5.amershambiosciences.com/) by upward capillary transfer in 0.4 m NaOH solution. The membranes were hybridized at 65°C using the whole genome of BSCTV, labeled with [α-32P]dCTP, as a probe. Signal intensity was measured using ImageJ (National Institutes of Health, http://www.rsbweb.nih.gov/ij/).
Mesophyll protoplasts were isolated from rosette leaves of 4-week-old Arabidopsis in the soil and transfected with plasmid DNA based on a previously described protocol (Yoo et al., 2007), except that the leaves were sterilized in 70% ethanol for approximately 1 min; the entire operation was performed under sterile conditions. Transfected cells were kept in the dark at room temperature. For the BSCTV replication assay, approximately 3 × 105 cells were collected at 0, 2 and 4 days after transfection for DNA extraction. Total genomic DNA was extracted from the cells according to a previously described protocol (Fontes et al., 1994). Newly replicated viral DNA was identified by DNA gel blot. For the GUS activity assay, pCambia1300-221–promoterRKP::GUS was transformed together with pCambia1300 or pCambia1300-C4, and the protoplasts were collected 16 h later. For detection of the myc-ICK2/KRP2 protein level in the rkp mutant, 35S-myc-ICK2/KRP2 was prepared by inserting the PCR-amplified coding region of ICK2/KRP2 fused with MYC in the N-terminus of ICK2/KRP2 to the pBA002 vector under the control of the 35S promoter. The plasmid was transformed into protoplasts of wild-type and rkp-1 plants; the pGFP2 plasmid, in which the GFP gene was under the control of the 35S promoter, was transfected alongside for an internal control. For detection of myc-ICK2/KRP2 protein level during BSCTV infection, 35S-myc-ICK2/KRP2 and pGFP2 plasmids were co-transfected with either pCambia-BSCTV or vector control into wild-type protoplasts.
E3 ubiquitin ligase activity assay
The entire RKP ORF was cloned into the PGEX-6P-1 vector and expressed in Escherichia coli. Fusion proteins were prepared according to the manufacturer’s instructions (GE Healthcare, http://www.gehealthcare.com). The in vitro E3 ligase assays were performed as described previously (Zhang et al., 2007b).
In vitro pull-down assay
RKP labeled with 35S on its methionine residues was generated by in vitro transcription and translation using a T7/T3-coupled TnT kit (Promega, http://www.promega.com/). ORFs of ICK/KRP proteins were cloned into the PGEX-6P-1 vector and expressed in Escherichia coli. RKP (4 μl) labeled with 35S on its methionine residues was mixed with 1 μg GST proteins in 1 ml GST-binding buffer (GBB: 50 mm Tris at pH 8.0, 120 mm NaCl, 1 mm DTT, 0.5% NP-40, 1 mm PMSF [Amesco, http://www.amescodavao.com]), and incubated at room temperature for 60 min. The glutathione–Sepharose beads (GE Healthcare) were then rinsed five times in GBB containing 0.5 m NaCl and twice with GBB. Bound proteins were released by boiling in SDS sample buffer at 90°C for 4 min and subjected to SDS–PAGE. After running the gel, protein was fixed to the gel using isopropanol 25%/acetic 10% for 30 min and washing for 5 min with distilled water, after which the gel was soaked with 1 m Na-salicylate for 45 min, washed again with water, and vacuum-dried at 80°C. Autoradiography exposure was performed at −70°C.
Protein gel blot analysis
Protein extracts were prepared by grinding material in homogenization buffer (Yoo et al., 2007). Protein gel blotting was performed according to standard procedures using primary anti-c-myc antibody (9E10; Santa Cruz Biotechnology, http://www.scbt.com) and anti-GFP antibodies (JL-8; Clontech, http://www.clontech.com/), followed by secondary goat anti-mouse antibody conjugated to horseradish peroxidase, and visualized using chemiluminescence as instructed by the manufacturer (ECL; Amersham Pharmacia).
We would like to thank Dr Nam-Hai Chua from the Laboratory of Plant Molecular Biology, Rockefeller University for kindly providing us with the pER8 vector, the Arabidopsis Biological Resource Center at Ohio State University for providing the T-DNA insertion lines, and Mr Sanyuan Tang for technical assistance. This research was supported by grants CNSF30325030/30530400 from the Chinese Natural Science Foundation (CNSF). Q.X. is supported by grants KSCX2-YW-N-010 and CXTD-S2005-2 from the Chinese Academy of Science. H.G. is supported by CNSF grant 30525004.