The necessity for ensuring quality-assurance measures in RT-qPCR analysis of gene expression is well recognized and a set of guidelines have been outlined for appropriate normalization strategy to control for non-specific variation between samples . Although a range of endogenous reference genes have been listed as good candidates for normalization of gene expression, identification of the most suitable reference genes for the given experimental conditions, rather than using reference genes published in the literature, is extremely important in functional genomics studies [47, 48]. In addition, certain reference genes may be stably expressed in one plant species but are not be well suited for use in other species . Apart from other fields of research, this knowledge is highly relevant to studies in plant host-virus interactions, as viruses are known to modulate key cellular processes in plants which may involve changes in the expression of endogenous host genes normally used as reference genes in RT-qPCR [51, 52]. Moreover, viruses manipulate different host cellular transcription pathways and the extent to which these pathways are affected will be dependent on the specific virus-host combination [53, 54]. Consequently, we evaluated geometric averaging of multiple reference genes as a means to avoid experimental bias in gene expression data.
In this study, we analyzed a set of six putative reference genes (Ubiquitin, Actin, GAPDH, EF1-a, SAND and NAD5) for their expression stability in leaf samples collected from a red-fruited wine grape cultivar (cv. Merlot) grown under field-conditions. Since expression stability of reference genes is known to vary with environmental conditions under which plants are grown, the type of plant tissue used and under a diverse set of biotic and abiotic stress conditions, we validated the expression stability of these six genes under our experimental conditions using the geNorm software and selected Actin and NAD5 to normalize RT-qPCR data obtained for the flavonoid biosynthetic pathway genes in virus-infected and virus-free grapevine leaves . In a previous study, GAPDH was ranked as one of the top three reference genes (GAPDH <Actin <EF1-a/SAND) for gene expression studies in grape berry development . However, we found that GAPDH is the least reliable in the context of our investigations on relative expression of the flavonoid biosynthetic pathway genes in grapevine leaf samples (Figure 4). These results clearly highlight the importance of validating reference genes as the most invariant internal controls for a particular experimental condition prior to investigating the relative expression of target genes by RT-qPCR.
By using gene-specific RT-qPCR, we present evidence in this study, the first example to our knowledge, that overall up-regulation of PAL, an enzyme that commits the flux of primary metabolism into the flavonoid biosynthetic pathway, and both "early" (CHS, CHI, F3'H, F3'5'H, F3H and FLS) and "late" genes (DFR, LDOX, UFGT and LAR) of the pathway occurred in GLRaV-3-infected symptomatic grapevine leaves (Figure 5). In red-fruited cultivars of wine grapes, anthocyanin pigments accumulate predominantly in berry skins displaying various shades of colors ranging from brick red to dark blue and their biosynthesis is developmentally triggered at the onset of véraison via the activation of flavonoid biosynthetic pathway genes . Under normal circumstances, these cultivars do not exhibit such coloration in their foliage during the growing season. Thus, changes in leaf color (Figure 2) and accumulation of specific classes of anthocyanins (Figure 6 and 7) only in GLRaV-3-infected symptomatic leaves supported our hypothesis that expression of the flavonoid biosynthetic pathway genes was activated in virus-infected leaves. Although this study was based on the expression analysis of flavonoid biosynthetic pathway genes and qualitative and quantitative variation of anthocyanins, flavonols and proanthocyanidins, it should be noted that mRNA expression is only one aspect of functional gene regulation of the pathway that result in changes in color of leaves in virus-infected plants. Since changes in leaf coloration begins to occur soon after véraison, even though GLRaV-3 can be detected in leaves of infected grapevines during the entire season including pre-véraison, it remains to be studied if the specific induction of anthocyanins in virus-infected leaves during post-véraison is tightly coupled with a cascade of physiological and/or molecular events triggered as a consequence of virus-host interactions during véraison.
In plants, delphinidin- and cyanidin-based anthocyanins exhibit blue and reddish color, respectively, under the acidic conditions of plant vacuoles . HPLC profiling of total anthocyanins showed that both cyanidin-3-glucoside and malvidin-3-glucoside accumulated in virus-infected symptomatic leaves and they are virtually undetected in virus-free green leaves (Figure 6a and Figure 7a&7b). We believe that presence of these two classes of anthocyanins, although cyanidin-3-glucoside is slightly but not significantly higher than malvidin-3-glucoside in virus-infected leaves, contributes to red and reddish-purple discoloration of virus-infected leaves. Since F3'5'H regulates the synthesis of delphinidin-based anthocyanins and F3'H regulates the synthesis of cyanidin-based anthocyanins, expression profiles of these two genes in concert with increased expression of anthocyanin-specific gene UFGT and its transcription factor gene MybA1 would ensure the flux of flavonoid intermediates towards the synthesis of these two classes of anthocyanins in virus-infected leaves. The levels of F3'H and F3'5'H gene transcripts observed in virus-free green leaves is in agreement with recent reports that F3'H gene was only slightly detectable and F3'5'H gene was expressed at non-detectable levels in green, fully expanded grapevine leaves [15, 57]. Our results also showed significantly higher levels of flavonols in virus-infected leaves than in virus-free leaves, and the predominant flavonols were quercetin followed by myricetin (Figure 6b and 7c&7d). Bogs et al. showed that total amounts of proanthocyanidins decline with leaf maturity and the two LAR isogenes have different patterns of expression with LAR1 showing seed-specific expression and insignificant levels in mature leaves and LAR2 readily present in different tissues, including leaves . Hummer and Schreier reported that proanthocyanidins as condensed tannins can precipitate proteins and several methods using protein precipitation have been used to estimate proanthocyanidins in various agricultural products . Using this approach, we showed that higher amounts of proanthocyanidins are present in virus-infected leaves than in virus-free leaves (Figure 6c) and the data correlate with strong induction of proanthocyanidin-specific genes; namely, LAR1, LAR2 and ANR. Since LAR and ANR genes provide two separate pathways for the synthesis of the terminal units of proanthocyanidin polymers, specific induction of LAR1 in virus-infected leaves (Figure 5a) would suggest that this gene may be contributing towards higher amounts of proanthocyanidins . Overall, these results are compatible with our hypothesis that activation of the flavonoid biosynthetic pathway genes occurred in GLRaV-3-infected symptomatic leaves during post-véraison period resulting in de novo synthesis of specific flavonoid classes and leading to phenotypic expression of GLRD symptoms. It is also likely that these flavonoid compounds confer protection from oxidative damage and/or against attack by opportunistic pathogens due to their antioxidant and free radical scavenging properties [8, 9, 60].
The use of more sensitive and gene-specific RT-qPCR technique enabled us to study relative abundance of the three highly homologous CHS gene family transcripts in virus-infected grapevine leaves. The results showed that three members of the CHS family (CHS1, CHS2 and CHS3) identified to date in grapevine, accumulated to varying levels, with CHS3 expression being significantly higher than the other two isogenes indicating its important role in color development in virus-infected leaves. This result is consistent with previous studies that CHS3, which is phylogenetically divergent from a cluster formed together by CHS1 and CHS2, was predominant in grape berry skins of red-fruited cultivars during coloration [61, 62]. The exact role of CHS1 and CHS2 in the biosynthesis of flavonoids may be insignificant, although their expression was implicated in the production of proanthocyanidins in unpigmented tissues of both red- and white-fruited grapevine cultivars [61, 62]. Among the two flavanone-3-hydroxylase isogenes, F3H1 showed higher expression levels than F3H2, and LAR1 of the two LAR isogenes of leucoanthocyanidin reductase was expressed at higher levels in virus-infected leaves. No such differential expression was observed in CHI isogenes. Thus, members of multigenic families appear to be induced differentially during the biosynthesis of flavonoids in virus-infected leaves of cv. Merlot showing GLRD symptoms.
Induced accumulation of anthocyanins and development of reddish-purple coloration in GLRaV-3 infected grapevine leaves appears to be analogous in some ways with stimulation of pigmentation in other plant species infected with taxonomically disparate viruses . It has been shown that mottling symptoms present on the seed coats of virus-infected soybean plants or induction of floral anthocyanin pigmentation in petunias can be caused by suppression of CHS posttranscriptional gene silencing (PTGS) via the expression of a virus-encoded silencing suppressor protein and that the reversion to pigmentation in virus-infected tissues is correlated with an increase in the CHS mRNA level [64–66]. Since CHS is the first committed enzyme in the flavonoid biosynthetic pathway, it is tempting to speculate that modulation of PTGS suppression of CHS isogenes by GLRaV-3-encoded silencing suppressor protein(s) occurs during post-véraison in virus-infected grapevine leaves leading to a cascade of molecular events resulting in up-regulation of CHS3 and the ensuing production of secondary metabolites conferring color to otherwise green leaves. However, identification of silencing suppressors of GLRaV-3 awaits further validation of this possibility.
An alternative explanation would be that, since grapevine leaves begin to show GLRD symptoms only during post-véraison even though GLRaV-3 can be detected in infected plants throughout the season (i.e. both during pre- and post-véraison) and the virus is phloem-limited, appearance of reddish-purple coloration in symptomatic leaves could be due to a consequence of changes occurring in host metabolism and altered phloem translocation during véraison. In this context, up-regulation of the flavonoid biosynthetic pathway genes in GLRaV-3-infected Merlot leaves may not entirely represent a host defense response to pathogen infection and, therefore, our results differ somewhat from other compatible plant-pathogen interactions in grapevine leaves and hybrid poplar, where genes encoding key enzymes of the flavonoid biosynthetic pathway were strongly induced after infection with phytoplasma or fungal pathogens [67–70]. Nevertheless, the present study contributes towards a better understanding of virus-host interactions leading to the development of GLRD symptoms in red-fruited wine grape cultivars.
In the present study, we observed higher transcript levels of MybA1 gene that encodes a MYB transcription factor in virus-infected leaves (Figure 5b). Although other MYB transcription factors have recently been reported in grapevines, our rationale for analyzing only MybA1 was because of its main role in the regulation of anthocyanin biosynthesis via expression of the UFGT gene [20, 71]. However, further research is necessary to determine whether fine regulation of the flavonoid biosynthetic pathway genes in virus-infected leaves involves a combinatorial action(s) of different R2R3-MYB transcription factors, including basic helix-loop-helix (bHLH) and WD40 factors expressed in a spatially and temporally controlled manner [3, 72].
It has been documented that the flavonoid biosynthetic pathway in fruits and vegetative tissues of plants is up-regulated by different environmental stress factors and in response to nutritional status . It has also been suggested that in woody perennials like red-osier dogwood, anthocyanins accumulate during senescence to provide optical masking of chlorophyll in order to reduce the risk of photo-oxidative damage to leaf cells . However, reduced levels of chlorophylls and carotenoids and higher amounts of specific classes of anthocyanins and the resulting changes in coloration of GLRaV-3-infected grapevine leaves during post-véraison may represent specific host-virus interactions as discussed above rather than a generalized abiotic stress response to environmental and/or nutritional imbalances. An integrated approach involving proteomic and metabolomic analyses combined with studies on modulation of cellular transcriptome would provide additional data for a comprehensive understanding of events that underlie changing colors of virus-infected grapevine leaves in red-fruited cultivars during post-véraison stage of berry development. Such information would also help to delineate grapevine's response to compatible virus infections from generic stress responses stimulated by a variety of abiotic and environmental factors.
Since berries in many red-fruited wine grape cultivars infected with GLRD show uneven ripening with reduced levels of extractable anthocyanins from berry skins (Naidu et al., unpublished results), the methodologies and results described in this study is providing leads for a deeper exploration of impacts of GLRD on berry skin pigments at the molecular level. In addition, there are several outstanding questions in GLRD-grapevine interactions that need to be addressed. They include: Do other red-fruited wine grape cultivars exhibit similar responses in the expression of flavonoid biosynthetic pathway genes and the profile of flavonoids to infection with GLRaV-3? Do genetically different GLRaVs trigger homologous responses in different red-fruited wine grape cultivars? Is the absence of dramatic symptoms in white-fruited wine grape cultivars an indication of non-responsiveness of the flavonoid biosynthetic pathway to virus infection? Indeed, GLRD-grapevine offers an excellent model system to address these questions.