Altered gene expression changes in Arabidopsis leaf tissues and protoplasts in response to Plum pox virus infection
© Babu et al. 2008
Received: 24 December 2007
Accepted: 09 July 2008
Published: 09 July 2008
Skip to main content
© Babu et al. 2008
Received: 24 December 2007
Accepted: 09 July 2008
Published: 09 July 2008
Virus infection induces the activation and suppression of global gene expression in the host. Profiling gene expression changes in the host may provide insights into the molecular mechanisms that underlie host physiological and phenotypic responses to virus infection. In this study, the Arabidopsis Affymetrix ATH1 array was used to assess global gene expression changes in Arabidopsis thaliana plants infected with Plum pox virus (PPV). To identify early genes in response to PPV infection, an Arabidopsis synchronized single-cell transformation system was developed. Arabidopsis protoplasts were transfected with a PPV infectious clone and global gene expression changes in the transfected protoplasts were profiled.
Microarray analysis of PPV-infected Arabidopsis leaf tissues identified 2013 and 1457 genes that were significantly (Q ≤ 0.05) up- (≥ 2.5 fold) and downregulated (≤ -2.5 fold), respectively. Genes associated with soluble sugar, starch and amino acid, intracellular membrane/membrane-bound organelles, chloroplast, and protein fate were upregulated, while genes related to development/storage proteins, protein synthesis and translation, and cell wall-associated components were downregulated. These gene expression changes were associated with PPV infection and symptom development. Further transcriptional profiling of protoplasts transfected with a PPV infectious clone revealed the upregulation of defence and cellular signalling genes as early as 6 hours post transfection. A cross sequence comparison analysis of genes differentially regulated by PPV-infected Arabidopsis leaves against uniEST sequences derived from PPV-infected leaves of Prunus persica, a natural host of PPV, identified orthologs related to defence, metabolism and protein synthesis. The cross comparison of genes differentially regulated by PPV infection and by the infections of other positive sense RNA viruses revealed a common set of 416 genes. These identified genes, particularly the early responsive genes, may be critical in virus infection.
Gene expression changes in PPV-infected Arabidopsis are the molecular basis of stress and defence-like responses, PPV pathogenesis and symptom development. The differentially regulated genes, particularly the early responsive genes, and a common set of genes regulated by infections of PPV and other positive sense RNA viruses identified in this study are candidates suitable for further functional characterization to shed lights on molecular virus-host interactions.
Systemic virus infection in plants relies on complex molecular interactions between the invading virus and the host plant [1–4]. Through such interactions, the virus recruits diverse host gene products (host factors) and metabolites for the translation and replication of its genome, which disrupts the normal biological processes of the host [4–6]. On the other hand, the expression of resistance genes in the infected host is also triggered by specific molecular interactions to combat the infection. Thus, virus infection induces the activation and suppression of global gene expressions in the host [7–16]. These gene expression changes are the molecular basis of general stress and defence-like responses, viral pathogenesis and host symptom development.
Plum pox virus (PPV) is a member of potyviruses, and is considered the most devastating viral pathogen of many stone-fruit species such as Prunus persica, P. persica var. nucipersica, and P. domestica . Despite numerous extensive screenings for genetic resistance against the virus, no natural resistant germplams are currently available for conventional breeding programs. Very little is known about the effects of PPV on host cells at the molecular level. A better understanding of molecular PPV-plant interactions would help identify key host factors required for PPV infection and assist in the development of novel recessive resistance to PPV .
To date, a major problem in dissecting molecular virus-host interactions is the progressive and asynchronous nature of the infection in leaves or in whole plants [1, 16–18]. In systemically infected plants, some cells may be infected earlier, while others may be infected later. Some parts may have higher virus titres while others may have fewer viruses or even virus-free. In the whole plant, it may take from days to weeks for the invading virus to spread from the infection site and develop systemic symptoms in the intact plant. Thus, the infected tissues contain cells at various infection stages or with viral concentrations in a wide range. As a result, temporal and spatial expression changes of host genes are very difficult to investigate. In the absence of a synchronous virus infection system and a global gene expression profiling technology, Maule and colleagues elegantly assessed the expression of selected host genes in individual cells across an advancing infection front using a cell biological approach (immunocytochemistry and in situ hybridization) (see review by Maule et al.) . To date, microarray technology has been adapted to profile genome-wide gene expression changes in plants in response to infections of a number of positive sense RNA viruses [7–10, 12–16]. More recently, Yang et al.  creatively dissected infection foci away from the non-infected tissue into four distinct zones using an infectious clone of Turnip mosaic virus (TuMV) tagged with green fluorescence protein (GFP) (TuMV-GFP) and profiled global gene expression changes in each of these zones with microarray, making it possible to spatially and temporally correlate virus infection and accumulation with host gene expression. Collectively these works largely advance our understanding of virus-host interactions.
Considering the availability of whole genome microarrays, and increasing use of genetic and biochemical approaches in archetypal plants like Arabidopsis thaliana to explore the molecular interactions between viruses and their host plants , we used an Arabidopsis ATH1 Affymetrix array and analyzed global gene expression changes in Arabidopsis leaves infected with PPV at 17 days post inoculation (dpi). Further, we generated Arabidopsis synchronized single-cell protoplasts, transfected with a PPV infectious clone and assayed global gene expression changes in the protoplasts at 3, 6 and 12 hr post transfection (hpt). The aim of this study was to investigate general gene expression changes in infected leaf tissues and their association with PPV infection and symptom development, and to identify putative host gene factors differentially regulated upon PPV infection, particularly those at early time points involved in defence and viral pathogenesis.
At 17 dpi, newly expanded PPV-infected rosette leaves from the inoculated plants and the corresponding leaves from the mock-inoculated plants were sampled for microarray hybridizations. Consistency of the microarray data was assessed by comparing the mean normalized signal intensities obtained from the mock-inoculated control samples across three independent biological replicates. A high correlation (0.96 - 0.98) among the mock inoculated control samples was observed, indicating low biological variability between the replicates.
Gene expression data were analyzed by one-way ANOVA to identify differentially expressed genes in PPV-infected rosette leaves from the inoculated plants relative to leaves from the mock-inoculated plants. Using stringent selection criteria with the Benjamini and Hochberg false discovery rate (FDR) of 5% , corresponding to Q ≤ 0.05 , ~31.4% of the non-redundant (nr) genes (7,151 out of the 22,810 Arabidopsis genes printed on the chip) were identified as significantly differentially expressed (either induced or repressed) in response to PPV infection. These 7,151 genes were further filtered using a 2.5-fold increase or decrease in signal intensity. Using this fold change cut-off, we identified 2,013 and 1,457 Arabidopsis genes that were significantly induced and repressed by PPV in systemically infected tissues, respectively.
Assignment of upregulated and downregulated genes in each functional group identified in PPV-infected Arabidopsis leaves and protoplasts.
Number of genes on the array a
PPV-infected leaves Upregulated/Downregulated [Number (%)/Number (%)] b
pPPV-SK68-transfected protoplasts Upregulated/Downregulated [Number (%)/Number (%)] b
Defence and virulence
204 (13.7)/165 (11.1)
32 (2.15)/33 (2.22)
Cellular communication/Signal transduction mechanism/transmembrane signal transduction
94 (5.67)/74 (4.47)
16 (0.97)/18 (1.09)
Cell Cycle/DNA processing/chromatin regulation and cytoskeleton reorganization
87 (11.3)/30 (3.89)
10 (1.30)/9 (1.17)
138 (21)/100 (15.2)
20 (3.04)/20 (3.04)
548 (12.9)/226 (5.3)
35 (0.83)/41 (0.97)
Proteins with binding function or cofactor requirement
210 (22.8)/139 (15.1)
20 (2.17)/22 (2.39)
Protein synthesis and translation
37 (11.8)/39 (12.4)
1 (0.32)/1 (0.32)
77 (11.9)/69 (10.6)
9 (1.39)/12 (1.85)
175 (17.3)/102 (10.1)
32 (3.16)/29 (2.86)
78 (13.9)/19 (3.4)
7 (1.24)/7 (1.24)
Transposable elements, viral and plasmid proteins
22 (17.7)/44 (35.5)
8 (6.45)/21 (16.94)
343 (3.3)/450 (4.3)
73 (0.70)/91 (0.87)
The analysis showed that that the number of upregulated genes was more than that of downregulated genes in PPV-infected Arabidopsis leaves, probably reflecting more diverse genes induced by PPV infection. This notion is supported by a disproportionate allocation in the number of altered genes in the infected leaf tissues belonging to major functional groups of metabolism, transcription/splicing/RNA processing proteins, defence and development/storage proteins.
Virus infections often disturb the general metabolism in the host. This includes the accumulation of soluble sugars (e.g., sucrose synthase) and starch (e.g., starch synthase), an increase in the rate of respiration, a decrease in the rate of photosynthesis and elevated levels of amino acid synthesis (e.g., amino acid permease and aminotransferase) [25–30]. In this study, we found that in PPV-infected leaves, 23% (124/548) of genes associated with metabolism of starch, soluble sugars and amino acids and 2.7% (15/548) of genes related to glycolysis showed elevated expression levels [see Additional file 1]. Appearance of disease symptoms such as chlorosis on the Cucumber mosaic virus (CMV)-infected leaf tissues in cucumber was correlated with increased glycolysis, phosphoenol pyruvate phosphate, respiration and starch accumulation, decreased photosynthesis and reduction in total protein synthesis . The increased expression of these genes may result from the physical blockage of the transport path, the inactivation of sugar transport proteins or the induction of cell wall invertase [31–33]. Indeed, several cell wall invertase genes were also induced in PPV-infected leaves in this study. These data indicate that a series of intercrossing pathways involved in sugar metabolism was differentially regulated by PPV infection. It is possible that altered gene expression pathways involved in sugar metabolism might contribute to the mild symptom development observed in PPV-infected Arabidopsis leaves.
In addition to soluble sugars, starch and amino acids, PPV also induced 16% (87/548) of membrane-associated proteins, including proteins associated with integral membrane, plasma membrane, endoplasmic reticulum (ER), vacuole, mitochondria and peroxisomes [see Additional file 1]. Recently, membrane proteins have also been found to be upregulated in N. benthamiana plants infected with Tomato ringspot virus (ToRSV) and PPV . For example, genes encoding proteins such as plasma membrane polypeptide, putative integral membrane, COP-coated vesicle membrane, clathrin protein and mitochondrial malate dehydrogenase were induced in PPV-infected Arabidopsis leaves that have similar functional identity to those upregulated in N. benthamiana plants infected with ToRSV and PPV. It is believed that many positive-strand RNA viruses replicate their genome in association with cellular membranes. Viral proteins that are required for the formation of viral replication complexes (VRCs) are often shown to be located in ER [34–36]. It would be interesting to investigate whether these upregulated membrane proteins are co-localized to VRCs. Characterization of such host membrane factors will provide useful information regarding molecular mechanisms governing membrane specificity for viral RNA replication.
Of the upregulated genes coding for nuclear or cytosolic proteins in PPV-infected Arabidopsis leaves, two OGTs (O-linked N-acetyl glucosamine transferase), named SEC (SECRET AGENT) and SPY (SPINDLY, a gibberellin response protein), which can potentially posttranslationally modify the viral capsid protein (CP), are very interesting. Study by Chen et al.  has shown that the N-terminus of the CP of PPV is modified with O-linked N-acetyl glucosamine in N. clevelandii, N. benthamiana and Arabidopsis, indicating that the modifications are not host specific. In Arabidopsis, this modification is catalyzed by SEC but not by SPY. Such modification is not essential but plays a role in the infection process . SEC and SPY have partial functional redundancy in plants [37, 38]. It is not clear whether the upregulation of SPY in PPV-infected Arabidopsis is a compensation for the loss of SEC that is used for the modification of the PPV CP.
We also observed as many as 52 chloroplast genes that were upregulated in PPV-infected leaf tissues [see Additional file 1]. These represent approximately 60% of the chloroplast genes in the Arabidopsis genome . Several studies suggest that differential regulation of chloroplast gene expression is correlated to viral symptoms such as mosaic and chlorosis [13, 16]. Recently, Dardick  has reported that ToRSV or PPV infection in N. benthamiana leaves induces severe mosaic and chlorosis symptoms. However, the infection of this host with Prunus necrotic ringspot virus (PNRSV) hardly induces any symptoms. Using potato cDNA chips, it has been shown that infections of ToRSV and PPV result in the suppression of 34% (83/243) of chloroplast-related genes, while no common genes are among the induced genes in ToRSV- and PPV-infected N. benthamiana. In contrast to other two viruses, PNRSV infection downregulates 67% (10/15) and upregulates 33% (5/15) of the chloroplast-related genes . Analogous to PNRSV-infected N. benthamiana, PPV-infected Arabidopsis leaves did not show typical viral symptoms such as mosaic and chlorosis. The induction of such a large number of chloroplast-related genes in Arabidopsis may help maintain the normal function and morphology of the infected leaf and counteract possible detrimental effects resulting from PPV infection.
In virus infected plants, protein activities are often altered through posttranslational modifications such as ubiquitination . In the current study, about 9% (60 genes) of the upregulated protein fate genes encoding proteins associated with the 26S proteasome were very significant (Q ≤ 0.05) and over-represented (P < 3.5E-07). Of the seven upregulated 26S proteasome genes, two genes code for AAA-ATPase subunits including RPT3 that reside in the mature 26S proteasome  and five genes including Rpn1, 3 and 6 are involved in ubiquitination. Fundamentally, the 26S proteasome is known to control many biochemical processes by programmed degradation of regulatory protein targets . Upregulation of these genes may interfere with ubiquitin-dependent reactions in plants . It has been reported that, the 26S proteasome degrades the virus-encoded movement protein that facilitates the viral cell-to-cell spread . Therefore, the upregulation of the 26S proteasome genes may be a resistance response to PPV infection.
On the other hand, in N. benthamiana plants infected with ToRSV, about 72% (33/46) of the proteasome genes have been shown to be upregulated . A few archetype proteasome proteins such as subtilisin-like protease, aspartyl protease, ubiquitin conjugating enzyme, F-box protein and 26S proteasome regulatory subunits induced by PPV in Arabidopsis leaves are upregulated by ToRSV in N. benthamiana plants . Upregulation of such protein fate and ubiquitin genes has also been observed in other studies during viral infection [10, 16, 43, 44]. Study by Marathe et al.  has demonstrated that the induction of F-box proteins may mediate the degradation components of the CMV-Y resistance pathway. Though the above mentioned protease genes induced by PPV infection are known to function in the protein degradation, plant defence responses and hypersensitive reactions [45–47], their exact roles in PPV infection are to be elucidated.
Many virus-induced symptoms have been linked to alterations in plant hormone synthesis [4, 48, 49]. Of the 96 development/storage genes identified in this study, 10 auxin, 5 ethylene and 4 gibberellin-related genes were repressed in PPV-infected leaves [see Additional file 2]. Recent studies [48, 49] on TMV infection in Arabidopsis have shown that during the process of infection the TMV replicase protein interacts with a subset of Aux/IAA proteins, resulting in the activation of auxin responsive transcription factors (ARF). Such an alteration in the transcription levels of auxin-responsive genes and the disruption of host normal physiological process have shown to be associated with disease symptoms. In this study, PPV infection suppressed the expression of two such genes encoding auxin IAA proteins (At5g57420 and At1g52830). With respect to the downregulation of ethylene responsive genes, though there is no direct evidence suggesting interactions between virus and ethylene signalling components, specific viral proteins have been suggested to alter the ethylene signalling pathway . For instance, the P6 protein of Cauliflower mosaic virus (CaMV) is associated with the disruption of the ethylene response pathway and this may play a key role in the induction of symptoms such as mild vein chlorosis to severe chlorosis and stunting . In the case of downregulation of gibberellic acid (GA), a direct interaction between the Rice dwarf virus (RDV) P2 protein and the rice ent-kaurene oxidase, a key enzyme in the synthesis of GA in plants was found to be responsible for the significant reduction of GA1 in RDV-infected plants, and in the appearance of dwarf phenotype . Collectively these examples demonstrate that the suppression of hormone-related genes by viruses induces symptoms through altering the host hormone and developmental signalling pathway. It would be interesting to know whether any of the observed downregulated hormone-related genes is associated with the reduced growth and delayed development of PPV-infected Arabidopsis.
Another significantly over-represented subset of genes suppressed in PPV-infected tissues includes genes coding for the protein synthesis and translation machinery [see Additional file 2]. This functional category contains 33 ribosomal subunits: 21 large, 11 small and 1 plastid specific ribosomal proteins. In the Arabidopsis genome, there are ~227 ribosomal genes [13, 52]. About 15% of these ribosomal genes were printed on the array. Based on the analysis, we observed that approximately 81% (27/33) of the small and large ribosomal subunit genes were significantly repressed in PPV-infected leaves.
However, an opposite trend was observed in PPV-infected N. benthamiana plants , where 80% (105/131) of potato ribosomal genes were upregulated. Comparison of these potato upregulated ribosomal genes in Arabidopsis through BLAST (E-value < -10) sequence similarity searches identified 55 nuclear ribosomal subunits . Comparative analysis of 27 Arabidopsis nuclear ribosomal subunits that were downregulated by PPV in our study with those in Dardick's study (using Table S5 from Dardick's study for comparison) showed 8 (7 large and 1 small) Arabidopsis ribosomal subunits with an opposing trend. Since plant viruses lack protein synthesis ability, they rely on the host cell translation machinery to produce viral proteins. Thus, PPV might benefit from the upregulation of ribosomal genes in N. benthamiana which showed more susceptibility to PPV and displayed much more severe symptoms than Arabidopsis. However, at present there is no evidence to discern whether the observed downregulation of ribosomal genes is due to the stronger basal resistance of Arabidopsis to PPV infection or due to the natural intricacy of the different host responses to PPV that are diverse depending on the tissue type, time post infection, temperature, host and virus strain, or simply due to an artifact of hybridizations, given that two different platforms were used in the studies.
Cell wall-related genes are major determinants of cell morphogenesis in plants, encoding several hundreds of different structural proteins and cell wall-related enzymes [53, 54]. In this study, 75 cell wall-related genes encoding protein for structure (23 genes), reassembly (34 genes), expansins (9 genes) and matrix polymers (9 genes) were significantly downregulated in PPV-infected leaves [see Additional file 2]. This is consistent with results observed in Arabidopsis plants infected by TuMV  and in rice plants infected with RDV . In these studies, repression of cell wall related genes is correlated with symptom development. Of these downregulated genes, extensin is a cell wall-localized hydroxyproline-rich glycoprotein that usually forms a cross-linked network with pectin to create a highly impassable barrier against pathogens . Downregulation of extensin during PPV infection suggests a decreased extensin deposition . Conceivably, PPV exploits the plasmodesmata to move from initially infected cells to the neighbouring healthy cells. It is not likely that PPV cell-to-cell movement is related to the deposition of extensin. Nonetheless, the molecular mechanisms and the biological connotation underlying the downregulation of cell wall-related genes in the infected leaf remain to be understood.
In addition to the above described genes, many genes in the category of metal ion binding or metal binding were also downregulated. It is not clear if the downregulation of these genes can lead to physiological changes that are associated with PPV infection.
Resistance induced through signal transduction pathways in Arabidopsis may take place immediately after PPV infection. The Arabidopsis leaves infected with PPV at 17 dpi contained a mixture of uninfected, infected, and post-infected cells. Thus, the spatial and temporal information on gene expression was lost. Sampling at earlier time points such as 1 or 2 dpi can reduce the percentage of post-infected cells but will largely increase the percentage of non-infected cells. So, protoplasts have been suggested to overcome this barrier . There are several advantages in using Arabidopsis protoplasts to study plant-virus interactions. First, in contrary to unpredictable infections on the leaf, synchronized protoplast cells allows for a more precise determination of gene expression. Second, a much higher proportion of the cells are infected than that in infected leaf tissues . For example, we have shown that in Arabidopsis protoplasts transfected with plasmid pPPV-SK68, a PPV infectious clone, the transfection efficiency is approximately 35% . Hence, the protoplast system makes it possible to study molecular events as early as just a few hours post infection. Like other experimental systems, the protoplast system also has some limitations. Protoplasts lack the cell wall so the physiological integrity of the plant organ is disrupted. During isolation, transfection and culture of protoplasts, gene expression changes induced by harsh treatments such as osmotic stress, enzymatic digestion and centrifugation may mask those induced by virus infection . Physiological milieu that a virus encounters in protoplast cells might be quite different from that of an intact leaf cell. Sample bias may result from the minute amounts of RNA isolated from the protoplasts and from the subsequent amplification step. However, improved technologies and experimental designs may be used to minimize the negative impact of the system [59–62]. Recently, we reported a protocol describing the isolation and transfection of Arabidopsis protoplasts for studying PPV-host interactions . In order to identify early time points for profiling global gene expression changes in infected protoplast cells, a time course experiment was performed. Arabidopsis protoplasts were transfected with a PPV infectious cDNA clone, pPPV-SK68 derived from an M strain and with a non-infectious deletion mutant, pPPV-SK68Δ as control, respectively . In the mutant clone, the deletion abolishes the translation of viral proteins CI, 6K2 and NIa/VPg required for PPV replication. The quantitative results showed that the viral RNA in the protoplasts was detectable at 3 hpt, accumulated at 6 hpt and peaked at 12 hpt (Figure 1C), followed by a decline at 24 hpt . The time course progression of PPV infection in Arabidopsis protoplasts is similar to that of Cymbidium mosaic virus (CyMV) in orchid protoplasts . Thus, three time points, 3, 6 and 12 hpt were selected for this study.
As evident [see Additional file 4; Additional file 5; Additional file 6], comparison of the differentially expressed genes between PPV-infected Arabidopsis protoplasts (263 up- and 304 downregulated genes) and PPV-infected leaf tissues (2,013 induced genes and 1,457 suppressed genes) resulted in only 32 and 33 common genes that were significantly induced (≥ 2.5-fold) and repressed (≤ -2.5-fold), respectively. A few vital genes such as genes encoding heavy metal protein, disease resistance protein, FtsH protease, cytochrome P450 and β-1, 3-glucanase were significantly differentially regulated by PPV in the infected leaves and transfected protoplasts. The smaller percentage of significantly differentially regulated genes observed in the PPV-infected protoplasts than in the PPV-infected leaf suggests a remarkable gene expression differences between the two systems. However, as discussed above, the differences in the alteration of gene expression changes in the protoplasts induced by PPV infection could have been masked by harsh treatments during protoplast isolation, transfection and culture or might result from the two different PPV strains used. Though the PPV-D and -M strains used in this study share more than 95% sequence identity , these two strains may induce different symptoms on the same host [20, 65].
PPV infection in protoplasts induced the expression of 263 genes. In contrast to about 11% of the suppressed genes that were common at different time points, there were no induced genes shared at different time points, suggesting induction of host gene expression was much more transient than suppression. Genes related to defence, cellular signalling, primary and secondary metabolism, transcription and transporters were upregulated and over-represented. Within these five categories, ~52% (48/93) of genes were predominantly upregulated (≥ 2.5-fold) at 6 hpt [see Additional file 3]. Examples of the upregulated genes in the above functional categories include MADS box protein, zinc finger family protein, WRKY family transcription factor, avirulence (Avr)-induced protein, hydrophobic protein, disease resistance TIR-NBS-LRR (toll-interleukin-1-receptor/nucleotide-binding site/leucine-rich repeat), protein kinases and phosphatases.
Another interesting gene upregulated by PPV encodes a disease resistance protein, TIR-NBS-LRR. This gene was upregulated with 6.39-fold in PPV-infected protoplasts at 6 hpt and 2.12-fold at 12 hpt, and 10.84-fold in PPV-infected leaves [see Additional file 4]. In the Arabidopsis (Col-0 accession) genome, there are 149 NBS-LRR-like genes [66, 67]. The resistance gene NBS-LRR mostly confers resistance to viral diseases but not necessarily associates with cell death or tissue necrosis . Genes Sw5, N and Rx are homologous to the NBS-LRR gene. Rx resistance to Potato virus X (PVX; Potexvirus) is the most classical example, which shows an extreme resistance that inhibits virus replication without hypersensitive cell death . The Sw5 gene confers resistance to several Tospoviruses , and N and Rx are effective against many natural variants of TMV and PVX [70, 71]. Since the specificity determinants subsist in LRR domains , it would be interesting to find out if the NBS-LRR gene detected in the PPV-infected leaf plays a role in resistance responses to PPV infection.
The comparison of induced genes at different time points in PPV-infected protoplasts revealed that genes upregulated at 3 hpt might regulate their target genes at 12 hpt. For instance, two genes coding for cnd41 (At5g24820), a plastid specific DNA binding factor, and calmodulin (At3g15050, At3g16490), a signalling binding protein were induced at 3 hpt and their target genes encoding metabolic enzymes such as serine carboxypeptidase (At1g11080) were upregulated at 12 hpt [57, 58]. Similarly, transcription factors such as zinc finger proteins (At5g54020; At3g45480; At4g12140) was upregulated at 3 hpt and the genes encoding enzymes such as GDSL-motif lipase (At3g43550) and glycosyl hydrolases (At3g26140, At4g39010) were upregulated at 12 hpt . These results suggest that virus infection may induce expression of specific upstream regulator genes first which then activate the expression of downstream target genes.
Based on the expression analysis (Figure 3), the number of repressed genes in the transfected protoplasts was greater than that of the genes induced at all time points, suggesting that at early infection stages, suppression is more active than induction of host gene expression. Moreover, about 11% (33 out of 304) of the suppressed genes overlapped at different time points, indicating a relatively stable suppression of genes. Consistent with this, in pea cotyledons infected with Pea seed borne mosaic virus (PSbMV), suppression of diverse host genes including seed storage proteins (e.g., vicilin and convicilin), and starch synthase occurs only in a zone of six to eight cells immediately behind the infection front, which coincided with the onset of viral replication . In the present study, a gene coding for the seed storage protein cupin family (At2g18540) that shares sequence homology to the pea vicilin and convicilin was significantly repressed by PPV infection at 12 hpt in Arabidopsis protoplasts. Suppression of such host gene expression or "host shutoff" has been suggested to be due to stress responses, induced defence states or a mechanism employed by viruses to outcompete against host mRNAs for host resources for their translation and replication [16, 17, 43, 74]. Study by Rottier  has demonstrated that protein synthesis in protoplasts is blocked in cowpea protoplasts transfected with Cowpea mosaic virus. Despite this inhibition, viral mRNAs are preferentially translated. Since both viruses and hosts share the same translational apparatus, viruses must have a mechanism to avoid this inhibition. It is not clear whether the disappearance of host mRNAs is the consequence of mRNA destabilization triggered by stress response, or host transcription inhibition by the virus, or targeted degradation through a virus-induced ribonuclease activity.
We performed a cross sequence comparison analysis by blasting individual sequence of all the genes significantly differentially regulated by PPV in infected Arabidopsis leaves against 2,156 uniEST sequences derived from PPV-infected P. persica leaves . High-scoring pair (HSP; ≥ 100 score) and E-value (E-value ≤ 10–20) were used to identify orthologs. A total of 170 induced [see Additional file 9] and 18 repressed Arabidopsis genes showed high sequence similarity to P. persica uniESTs. Of these 170 induced genes, approximately ~39% (67/170) and ~16% (28/170) of the genes were involved in metabolism and defence, respectively. The defence-related genes include those coding for pathogenesis related (PR)-proteins, heat shock and germin-like proteins (Ger2) [see Additional file 9]. An increased expression of PR proteins requires salicylic acid (SA) signifying that compatible host-virus interactions induce defence-like responses through a SA dependent pathway . More recently, it has been shown that PPV infection in the apoplastic space of P. persica leaves induces a thaumatin-like protein (PR-5) . Induction of such protein is thought to be mediated by an increase in the level of peroxidase accumulation . A similar trend was also observed in PPV infected Arabidopsis leaves where genes encoding PR-5 and peroxidase were significantly induced and their counterparts were also found in the uniEST sequences derived from PPV-infected P. persica.
To determine whether the diverse RNA viruses have the ability to elicit common gene expression changes, we cross-compared the genes that were significantly differentially regulated by PPV infection in this study [Additional file 1] [Additional file 2] to those Arabidopsis genes that were differentially regulated by infections of other positive sense RNA viruses including CMV [9, 10]; TMV ; TuMV, and Oil seed rape mosaic virus (ORMV) , and to five diverse RNA viruses, i.e., TuMV, ORMV, Turnip vein clearing virus (TVCV), Potato virus X (PVX), and CMV . Collectively, 416 common genes [see Additional file 10] were identified from this analysis. These genes were further classified into 6 different categories. Group A contains 202 genes induced by infections of PPV and other positive sense RNA viruses. About ~40% (80/202) are metabolic genes and ~8% (17/202) are defence-related genes. Group B includes 41 Arabidopsis genes that were induced during PPV infection but repressed by infections of other positive sense RNA viruses. About ~27% (11/41) are involved in metabolism. Group C comprises of 119 Arabidopsis genes repressed during PPV infection but induced by infections of other positive sense RNA viruses. For example, a defence-related gene encoding glycosyl transferase (At1g70090), which was repressed in PPV-infected leaves at 17 dpi was observed to be induced more than 1.5-fold in TMV-infected leaves  at 14 dpi. Group D consists of 7 Arabidopsis genes repressed by infections of PPV and other positive sense RNA viruses. Group E comprises 25 genes that were induced during PPV infection but either induced or suppressed by infections of other positive sense RNA viruses. Group F contains 22 Arabidopsis genes repressed during PPV infection but induced during infections of more than one positive sense RNA virus. Previous studies have revealed that in virus-infected Arabidopsis and tomato leaves as many as one-third of the genes commonly induced by viruses are known or predicted to be involved in plant defence- and stress-related genes [5, 15]. In the current study, about 52% (217/416) of the functional proteins differentially regulated in response to infections of PPV and other positive sense RNA viruses are involved in defence, cellular signalling and metabolism. Twenty-nine cellular signalling genes involved in defence mechanisms including protein kinase (At3g08760, At4g32250), protein phosphatase 2C (At5g02760, At3g05640), pyruvate kinase (At2g36580), serine/threonine protein kinase (At3g17410) and calmodulin-binding family protein (At5g62390) [3, 78, 79] were induced by infections of PPV and other positive sense RNA viruses. Substantial induction of these defence, cellular signalling and metabolism genes by PPV and other positive sense RNA viruses in infected leaf tissues suggests that there is a shared pathway modulating their expression.
Microarray analyses conducted in this study allowed us to gain insights into global gene expression changes in Arabidopsis associated with PPV infection, symptom development and defence responses. At 17 dpi, PPV displayed mild growth retardation, slightly delayed flowering and bolting. PPV infection was associated with the downregulation of genes related to development/storage proteins, protein synthesis and translation and cell wall-related genes and the upregulation of genes involved in soluble sugars, starch, amino acid, intracellular membrane/membrane-bound organelles, chloroplast and protein fate. Transcriptional profiling of PPV-infected protoplasts revealed that at early infection stages, downregulation of host gene expression was more prominent than upregulation. Concomitant with this downregulation, several defence and cellular signalling genes were induced during PPV infection. These genes may be involved in signal transduction pathways to induce resistance to combat the virus. Further targeted functional studies of these early defence-related genes will help unravel their biological function during the course of PPV infection.
Arabidopsis thaliana accession Col-0 was grown in a growth chamber at 80% RH with a day/night regime of 16 h light at 22°C followed by 8 h dark at 18°C. Prior to the preparation of viral inoculum, leaf tissues collected from Prunus glandulosa infected with a Canadian PPV-D isolate were tested positive by ELISA and by RT-PCR. The inoculum was then prepared by homogenizing approximately 1 g of P. glandulosa leaf tissues infected with PPV-D in 10 ml inoculation buffer (PO4 buffer, pH 7.4 containing 2% PVP-40 (w/v) and 0.2% diethyl dithiocarbonic acid). Viral inoculum was mechanically rub-inoculated onto Arabidopsis leaves at 5–6 leaf stage. Control plants were mock inoculated with extracts from healthy P. glandulosa leaf tissues ground in inoculation buffer. All the mock- and PPV-infected Arabidopsis plants used in the microarray experiment were randomized and grown in different locations within a greenhouse under the same growth conditions. Fifty fully expanded rosette leaves above the inoculated leaves were harvested from 25 infected or mock inoculated control plants at 17 dpi. The leaves were immediately frozen with liquid nitrogen and stored at -80°C prior to RNA extraction. Six independent hybridizations were performed using total RNA isolated from three independent biological replicates of the virus-infected or mock-inoculated control samples.
The PPV infectious cDNA clone pPPV-SK68 derived from an M strain was a kind gift from Drs. Laszlo Palkovics and Ervin Balazs, Agricultural Biotechnology Center, Hungary. A control plasmid, pPPV-SK68Δ, was generated by deleting 1489 nt Sma I cDNA fragment from nt 4272 to 5760 (inclusive). This deletion resulted in the frame-shift of the portion of PPV genome after nt 5760 . The deletion region abolishes the translation of the viral protein domains CI-6k2-NIa/VPg-NIa/Pro-NIB-CP which is required for viral replication. Thus, clone pPPV-SK68Δ is non-infectious.
Protoplasts were isolated using fully expanded leaves of 3-week-old Arabidopsis accession Col-0 and transfected with the PPV infectious cDNA clone or the mutated PPV cDNA clone following the protocol we have developed recently . Transfected protoplasts were incubated at room temperature for 5 min, and then transferred to 25°C and incubated in the dark for 3, 6 and 12 h, respectively. At selected time points, transfected protoplast cells were harvested, immediately frozen in liquid nitrogen and stored at -80°C prior to RNA extractions. Gene expression differences were averaged out by pooling six RNA samples isolated independently from protoplasts transfected with pPPV-SK68 and pPPV-SK68Δ at each time point after transfection. For time course comparison, three independent biological replicates were maintained. Thus, for three different time points, 18 gene chips in total were used to hybridize with RNA isolated from protoplasts transfected with pPPV-SK68 and pPPV-SK68Δ.
Total RNA from virus-infected and mock-inoculated leaves of Arabidopsis plants and from protoplasts transfected with pPPV-SK68 and pPPV-SK68Δ were isolated using Trizol reagent (Invitrogen) following the manufacturer's suggestion. Total RNA was purified using RNeasy Mini columns (Qiagen) after DNase I treatment (Invitrogen). Integrity of total RNA was assessed using Bioanalyzer 2100 (Agilent Technologies, Boblingen, Germany). Primers OL-PPVDF3 (5'TCAATGGAATGTGGGTGATGA 3') and OL-PPVDR6 (5' GATTAGACTCTCACCCAGGTA 3'); and PPVMF18 (5' GGAAATTGAGAGATACCT 3') and PPVMR16 (5' TTTGTCTAAAAGTGGGTTT 3') were used to detect PPV in the virus-infected or mock-inoculated control samples and in pPPV-SK68- and pPPV-SK68Δ-transfected protoplasts, respectively.
Standard RNA processing and hybridization protocols were followed as recommended by the Affymetrix manual (Santa Clara, CA, USA). A double-amplification approach (Affymetrix) was used to generate cRNA for hybridization. Total RNA isolated from virus-infected or mock-inoculated leaves were subjected to double-amplification. Labelling, hybridization and detection of hybridization signals were performed at the McGill University and Genome Quebec Innovation Centre (please see Availability & requirements for more information).
Hybridization signals obtained from the scanned array images were quantified using MAS5.0 expression analysis algorithm, implemented in the Affymetrix GCOS (GeneChip® Operating) software version 1.2. Raw MAS5.0 data files obtained from scanned array images were imported into GeneSpring microarray analysis software version 7.3.1 (Silicon Genetics, Redwood City, CA, USA). Systemic variations across the chips were normalized using per gene and per chip normalization methods available in the GeneSpring software. Mean normalized value of the signal intensity for each gene in three biological replicate experiments was adopted as the expression value of the gene. The significant changes in gene expression during PPV infection in infected leaves and in protoplasts were determined using P values (P ≤ 0.05) derived from one-way ANOVA and the significant changes in the values of the identified genes were further adjusted using multiple testing correction of the Benjamini and Hochberg false discovery rate procedure  which applied a stringent cut off of 0.05 that corresponds to Q ≤ 0.05 . This multiple testing correction procedure is used to correct for the occurrence of false positives and to maximize the likelihood of finding significant gene sets. In order to identify genes that show significant changes of transcript levels during virus infection, we further searched for genes that were significantly up- (≥ 2.5-fold) or downregulated (≤ -2.5-fold) relative to the control. For clustering analysis, gene expression profiles from the protoplasts transfected with the PPV infectious clone at each time point were clustered based on their similarity in expression using hierarchical average linkage algorithm and Pearson correlation distance metric implemented in the Cluster 3.0. The clustering results were visualized by a gene tree heat map using the TreeView program . K-means clustering was employed to cluster the gene expression data from the PPV transfected protoplasts into several distinct expression profiles using Pearson correlation distance metric. The microarray data from this work was submitted to the NCBI database with Gene Expression Omnibus (GEO) accession number [GSE11217].
Each gene was assigned with a putative function by blasting the AGI probe identifier sequences against the Arabidopsis database deposited at TAIR (the Arabidopsis Information Resources) using the BLASTX program. Functional classification was performed by assigning the function of each gene based on protein sequence similarity using MIPS (Munich Information Centre for Protein Sequences; please see Availability & requirements for more information) Arabidopsis functional classification schema. Genes without a category name or with unclear classification were labeled as "unknowns". Over-or under-representation of genes involved in each functional group was evaluated using Fisher's exact test function [23, 24].
The PPV-infected P. persica leaf (PPL) EST library , consisting of 2156 uniEST sequences extracted from NCBI dbESTs [GenBank: DN552797, DN556565] were formatted so that they were searchable using the tBLASTx algorithm. Sequences of 2013 PPV induced and 1457 PPV repressed Arabidopsis transcripts in infected leaf tissues respectively were aligned to this collection to search for the presence of P. persica orthologs using the tBLASTx algorithm with default settings . The BLAST output for each transcript was then parsed for high-scoring pair (HSP) and the associated E-value. Any hits with HSP 100 and E-value 10-20 were taken as an indicative of significant similarity .
The Arabidopsis genes differentially regulated in PPV-infected leaf tissues were cross compared by searching against the Arabidopsis gene set that were differentially expressed in response to the infection by other positive sense RNA viruses [7–10, 16].
Primers used in semi quantitative RT-PCRa and in Northern hybridzationsb
Probe set IDs c
Forward Primer (5'---3')
Reverse Primer (5'---3')
A. Genes induced only in PPV-infected Arabidopsis protoplasts
Disease resistance protein, TIR-NBS-LRR class
Cation/hydrogen exchanger CHX10, putative
Avirulence induced gene (AIG) protein, related
Disease resistance protein, related
F-box family protein, related
Heat shock transcription factor
Hydrolase family protein
MAP kinase, putative
Protein kinase PK1- like
B. Genes induced in PPV-infected Arabidopsis protoplasts and leaf tissues
AMP-binding protein, putative
Cytochrome P450, putative
Cytochrome P450-like protein
Disease resistance protein, TIR-NBS-LRR
Calmodulin-binding family protein
We thank Drs. Laszlo Palkovics and Ervin Balazs (Agricultural Biotechnology Center, Hungary) for providing the PPV infectious clone; Drs. Hélène Sanfaçon, Brian Miki and Dan Brown (Agriculture and Agri-Food Canada) for very helpful discussion; Drs. Antonet Svircev and Lorne Stobbs (AAFC) for assistance in live virus work; and Dr. Daowen Wang (Chinese Academy of Sciences, China) and Dr. Hélène Sanfaçon for critical reading of the manuscript. This work was supported by grants to AW from AAFC and from the Natural Science and Engineering Research Council of Canada.
This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.