Identification and VIGS-based characterization of Bx1 ortholog in rye (Secale cereale L.)

The first step of the benzoxazinoid (BX) synthesis pathway is catalyzed by an enzyme with indole-3-glycerol phosphate lyase activity encoded by 3 genes, Bx1, TSA and Igl. A gene highly homologous to maize and wheat Bx1 has been identified in rye. The goal of the study was to analyze the gene and to experimentally verify its role in the rye BX biosynthesis pathway as a rye ortholog of the Bx1 gene. Expression of the gene showed peak values 3 days after imbibition (dai) and at 21 dai it was undetectable. Changes of the BX content in leaves were highly correlated with the expression pattern until 21 dai. In plants older than 21 dai despite the undetectable expression of the analyzed gene there was still low accumulation of BXs. Function of the gene was verified by correlating its native expression and virus-induced silencing with BX accumulation. Barley stripe mosaic virus (BSMV)-based vectors were used to induce transcriptional (TGS) and posttranscriptional (PTGS) silencing of the analyzed gene. Both strategies (PTGS and TGS) significantly reduced the transcript level of the analyzed gene, and this was highly correlated with lowered BX content. Inoculation with virus-based vectors specifically induced expression of the analyzed gene, indicating up-regulation by biotic stressors. This is the first report of using the BSMV-based system for functional analysis of rye gene. The findings prove that the analyzed gene is a rye ortholog of the Bx1 gene. Its expression is developmentally regulated and is strongly induced by biotic stress. Stable accumulation of BXs in plants older than 21 dai associated with undetectable expression of ScBx1 indicates that the function of the ScBx1 in the BX biosynthesis is redundant with another gene. We anticipate that the unknown gene is a putative ortholog of the Igl, which still remains to be identified in rye.

Introduction Benzoxazinoids (BXs) are a group of defense-related molecules accumulated as glycosylated precursors in grasses and sporadically in a few dicot species [1]. The first reports published in the 1950s identified the BX as components of rye resistance against Fusarium [2] and maize resistance to the European corn borer Ostrinia nubilalis [3]. Further studies confirmed the diverse roles of BX in innate immunity against pests and fungal diseases in maize [4,5], wheat [6] and rye [7]. On the other hand, degradation (i.e. detoxification) of the benzoxazinoids was found to be a pathogen strategy to overcome the chemical defense of the host [8]. Studies by Kettle et al. (2015) and Schulz et al. (2013) [9,10] indicated that BX were responsible for rye allelopathy. Nutritional investigations revealed that the BXs themselves or the BX-derived compounds found in rye or wheat food products show potentially health-promoting activities [11]. The diverse spectrum of biological functions as well as the possible nutritional values caused that the BXs became a widely studied group of secondary metabolites in plants.
Characterization loss-of-function phenotypes is one of the main strategies to study biological function of a gene known only from its nucleotide sequence. Virus induced gene silencing (VIGS) is a method of choice to induce silencing of a target gene and it is particularly suitable for species with no collections of knock-out mutants and no methods of genetic transformation. The technique is based on targeted post-transcriptional gene silencing (PTGS) in host species activated by introduction a viral vector bearing a fragment of the target gene. It was first developed in dicots [25] and by constructing vectors derived from Barley Stripe Mosaic Virus (BSMV) the method was adapted for barley [26] and later for wheat [27], Brachypodium distachyon [28] and oat [29]. The method can be used to silence a single gene or simultaneously two different genes [30] and can be applied to genes expressed in root, leaves [31], spikes and grains [32]. VIGS approach has been used to verify gene function in biochemical pathway of morphine biosynthesis [33]. For a recent comprehensive review on VIGS see [34]. Designing the VIGS experiments and adapting the method for a new species the use of internal silencing standard is required. This is achieved by using a vector bearing a fragment of gene encoding phytoene desaturase (PDS), which is a key enzyme of carotenoid synthesis. Its silencing inhibits synthesis of carotenoids and leads, as a consequence, to photobleaching of chlorophyll. This is easily visible as white leaf sectors of the experimental plants. Silencing of the PDS and the presence of white leaf sectors serves as a visual marker confirming correctly constructed vectors and successful inoculation. Rye belongs to so called recalcitrant species for which transformation methods are practically not available, there are no mutant collections, rye genome has not been sequenced and there are very limited resources on genomic sequences of rye genes. In such species VIGS is a method of choice to characterize biological function of a specific gene. Rye is an important cereal used as grain, forage and a green manure crop. Less demanding growth conditions combined with frost and drought resistance make it suitable for the colder regions and poor soils of Northern and Eastern Europe, North America and Northern Asia. Rye is among the first species where benzoxazinoids have been identified; however, the genes involved in the BX biosynthesis pathway remain unknown until recently. The first two, ScBx1 and ScBx2, were proposed by La Hovary in his PhD thesis [35]; however, the structure and the nucleotide sequences of the subsequent five genes-ScBx1, ScBx2, ScBx3, ScBx4 and ScBx5 -were published by Bakera et al. [23]. The two genes Bx6 and Bx7 encoding 7-hydroxylase and 7-O-methylotranferase, and converting 2,4-dihydroxy-1,4-benzoxazin-3-one (DIBOA-Glc) to (DIMBOA-Glc) in maize, have not been found in wheat or rye, although the same reactions are expected to occur in these species. However, the Bx6-like gene encoding DIBOA-Glc dioxygenase has been described in rye (GenBank: HG380520.1).
Conversion of indole-3-glycerol phosphate (IGP) to indole, catalyzed by indole-3-glycerol phosphate lyase (IGL), is the branching point between two pathways: BX synthesis and tryptophan synthesis. Proteins with IGL activity are encoded by three different genes: Bx1, TSA or Igl [36]. Although the genes are evolutionarily related and show high sequence homology, they differ in their expression patterns and, more importantly, they have different enzyme characteristics. BX1 (encoded by Bx1) acts as a monomer and delivers free indole, which is a substrate in the BX biosynthesis pathway [36]. The tryptophan synthase alpha (TSA) subunit is active only in a complex with the tryptophan synthase beta (TSB) subunit, and the catalyzed reaction provides indole directly to the tryptophan synthesis pathway [36,37]. The third protein with IGL activity is encoded by the Igl gene. The enzyme is active as a monomer and in this respect can replace the BX1 enzyme in providing free indole to the BX pathway [14,18,22,36]. The two genes involved in BX synthesis, i.e. Bx1 and Igl, have different expression patterns. Bx1 is developmentally regulated, while Igl is mostly herbivory-induced [4,36]. The cDNA clones identified in rye were annotated as Bx biosynthesis genes mostly based on the sequence homology to already known genes [23].
Considering high homology of Bx1 to TSA, to TSA-like as well as to Igl, the correct annotation of the identified putative Bx1 gene in rye required experimental verification. This was the aim of the present study. The strategy of functional analysis of the gene was based on correlation of its native expression pattern and experimental VIGS-induced silencing with the levels of benzoxazinoids in rye leaves.

A. Post-transcriptional silencing of ScBx1 correlates with lower BXs content
Inoculation of rye seedlings with BSMV-based control vectors showed that expression of the analyzed gene and the BX content was significantly enhanced compared with non-treated plants. This indicated that inoculation with the BSMV-based vectors activated two contradictory processes: up-regulation of the analyzed gene, which was the reaction for virus infection (Fig 3A), and gene silencing, which was induced by the BSMV-based VIGS vector. On average

B. Transcriptional gene silencing
Another series of VIGS-BSMV experiments was designed to induce transcriptional silencing (TGS) of the analyzed gene. The experimental vector BSMV:α,β (pScBx1-fragment I) ,γ (pScBx1-fragment II) marked as BSMV:pScBx1 with gDNA fragments derived from the promoter region of ScBx1 and the control vector BSMV:α,β (-) ,γ (-) marked as BSMV:00 were used for seedling inoculation  Table 2). The 3 plants were analyzed for the ScBx1 expression, cytosine methylation within the ScBx1 promoter fragments and the content of hydroxamic acids. The leaf samples for the analysis were taken 14, 21 and 99 dpi.  Relative expression of ScBx1 was very low in non-inoculated plants and in plants abraded with silicon carbide without or with FES buffer. Inoculation with BSMV:00 vector led to over 40-fold induction of ScBx1 relative expression, which reached the levels 0.032 (14 dpi), 0.024 (21 dpi) and 0.015 (99 dpi) (Fig 6A).

ScBx1 promoter fragments in plants inoculated with BSMV:pScBx1 vector show elevated cytosine methylation
Genomic DNA corresponding to fragment I of the ScBx1 promoter contains 3 sites recognized by MSRE: single MluI site and two HhaI. Genomic DNA corresponding to fragment II of the ScBx1 promoter contains two HpyCH4IV sites and a single HhaI site (Fig 5). The rate of cytosine methylation in all these sites has been analyzed. The rates of cytosine methylation in fragment I MluI site in non-inoculated plants were 1.5% (14 dpi), 1.6% (21 dpi) and 1.0% (99 dpi). The methylation rates in plants inoculated with the BSMV:00 vector were 2.5% (14 dpi), 1.4% (21 dpi) and 1.2% (99 dpi). The rates of methylation in the MluI site in plants inoculated with  the experimental vector BSMV:pScBx1 varied from 1.3% to 2.0% (14 dpi), from 4.8% to 35.0% (21 dpi) and from 2.0% to 12.2% (99 dpi) (Fig 7A, MluI, pScBx1-fragment I).

fragment II).
Elevated cytosine methylation in ScBx1 promoter correlates with lower ScBx1 transcript level and lower BX content On average normalized expression of ScBx1 in plants inoculated with the experimental vector (BSMV:pScBx1) compared to the control (BSMV:00) was 20 x lower at 14 dpi, 1.5 x lower at 21 dpi and 1.7 x lower at 99 dpi (Fig 8A). The strongest impact of inoculation with BSMV:pScBx1 on relative expression of ScBx1 was at 14 dpi ( Fig 8A).

Discussion
Bx1, the first gene of the BX biosynthesis pathway, encoding indole-3-glycerol phosphate lyase (IGL), was cloned for the first time in maize using the recessive bx1 mutation [14]. This VIGS-based analysis of Bx1 ortholog in rye opened the possibility to identify Bx1 orthologs in other cereal species: hexaploid wheat [20], wild barley [18], and diploid wheat species [19]. Putative Bx orthologs in rye were identified by La Hovary [35] and recently by Bakera et al. [23]. It is worthwhile to point that although the benzoxazineless1 (bx1) mutation was originally used to identify and to clone Bx1, the gene itself and the encoded protein (BX1) are not the only ones responsible for synthesis of indole, which is the substrate for the BX biosynthesis pathway. Conversion of indole-3-glycerol phosphate to indole is catalyzed by enzymes encoded by at least three different genes: Bx1, TSA and Igl. All of the genes as well as the encoded proteins show a high level of sequence similarity, which is the result of the common evolution history [36,38,39]. The two Bx1 and Igl genes encode enzymes that are active as monomers and can deliver free indole to the BX biosynthesis pathway [36]. The third one encodes the tryptophan synthase alpha subunit (TSA), which forms a complex with the tryptophan synthase beta subunit (TSB) and only in this configuration catalyzes indole synthesis [36,37]. The IGL enzyme (encoded by the Igl gene) is active as a monomer and in this respect it can replace BX1 (encoded by the Bx1 gene) in providing free indole for BX biosynthesis [18,36]. The possible presence of two functionally redundant genes, which show a high level of nucleotide similarity, indicates that the final assignment of the putative Bx1 gene should be the result of both in silico analysis and experimental verification.
A rye candidate of the Bx1 ortholog has been isolated from the BAC library and assigned as the putative ScBx1 gene based on its nucleotide sequence analysis [23]. Here the experimental characterization of the gene in rye is presented. The ScBx1 transcript was detected only in young rye seedlings, indicating a strong, developmentally regulated expression pattern. The highest level, found in leaves 3 dai, was followed by a sharp decline; thus in leaves collected 21 dai the transcript was already undetectable. The expression pattern correlated with BX content. In non-induced seedlings the highest level of BXs matched the highest transcript level found 3 dai. In older leaves the content of benzoxazinoids declined, although it never reached the level of zero as it did for the Bx1 transcript. In contrast to the undetectable transcript level, the content of BXs in rye leaves remained at low but still detectable levels throughout the whole analyzed period, i.e. until 99 dai. The seedling-specific expression of the analyzed gene corresponds well to expression patterns of Bx1 orthologs found in rye [35], maize [12,14], and wheat [22]. Based on these results we assume the presence of another gene encoding protein with IGL-enzyme activity, which acts redundantly with the ScBx1. We expect that the gene, which remains to be identified in rye, will be expressed at least in plants older than 21 dai.
Further results indicated strong activation of the putative ScBx1 gene upon inoculation with the BSMV-based vector. The expression remained elevated during the whole analyzed period, i.e. 99 days after imbibition. The changes of ScBx1 expression induced by the BSMV-based vector are similar to changes of Bx1 orthologs caused by biotic stresses in other species: in maize upon herbivory infestation [4,40,41] or after treatment with stress-dependent plant elicitor peptide ZmPep1 [42], after treatment with jasmonic acid (JA) or after pathogen and herbivory infection [43]. The experimentally determined expression pattern is in agreement with the presence of stress-specific motifs detected in the ScBx1 promoter sequence [23]. The seedlingspecific and stress-dependent expression of the gene correlates well with the content of BXs, as it has been confirmed by a high correlation coefficient between seedling-specific and BSMVinduced expression of Bx1 with the accumulation of BXs.
Biological function of the putative ScBx1 gene was analyzed using the virus-induced gene silencing (VIGS) strategy. The chosen method proved to be very effective in a number of plant species. The method is particularly useful in species where transformation-based tools are difficult to use or are not available (for a comprehensive review see [44]. Rye belongs to this group of recalcitrant plants. Although the species is a nonhost for Barley stripe mosaic virus the presented results proved that the BSMV-based VIGS vector could be successfully used for gene silencing experiments. The inoculation rates were efficient and ranged from 4.7% to 12.5% as estimated from inoculations of 569 seedlings. Experimental inoculations were done with the Bx1 silencing vectors containing fragments of the PDS marker gene and of the analyzed Bx1 cloned alternatively to the β or γ subunit of the BSMV-based VIGS vector [30]. Both constructs resulted in similar inoculation efficiencies (10.4% and 9.3%) indicating that both β and γ subunits of the BSMV vector were equally suitable for the purpose. It is worth noting that the system of simultaneous posttranscriptional silencing of the two genes allowed for easy visual screening of photobleached plants, which presumably showed decreased expression of the analyzed gene. Only the plants showing successful inoculation with the BSMV-based vector were used for further molecular analysis. The experiments proved that PDS fragment conserved between barley and wheat was effective as a visual screening marker of VIGS in rye. The analysis of 9 plants confirmed that the ScBx1 expression and the BX content were significantly lowered compared with control plants inoculated with the BSMV:PDS vector. Lowered BX content was highly correlated with decreased transcript level of the analyzed gene indicating strong functional association between ScBx1 expression and BX biosynthesis. It is consistent with other examples of using VIGS as a reverse genetics tool. A similar approach of genes' silencing was used to investigate regulation of morphine biosynthesis in opium poppy [33]. The results of BSMV-based silencing of CYP96B22 encoding P450 monooxygenase was the basis to conclude about a role of the enzyme during basal resistance of barley against Magnaporthe [45]. Similar strategy was also applied to dissect contribution of glycerol-3-phosphate dehydrogenase and GLI1-encoded glycerol kinase in wheat systemic acquired resistance [46] and to establish function of zinc finger protein TaLSD1 in wheat resistance against stripe rust [47].
Opposite to the VIGS-based posttranscriptional silencing, there are relatively few reports on viral vector-induced transcriptional silencing (TGS) of endogenes in plants. It is partly because the endogenes were thought to be recalcitrant to the RNA-directed transcriptional silencing and most of the first attempts were focused on transgenes. Plant RNA viruses such as Potato virus X (PVX), Tobacco rattle virus (TRV) and Cucumber mosaic virus (CMV) vectors were used as tools to induce TGS of transgenes [48][49][50]. More recent studies, however, proved that endogenes can also be successfully targeted with virus-based vectors and consequently transcriptionally silenced. A high frequency of cytosine methylation at the targeted gene promoters was found in petunia and tomato plants inoculated with a Cucumber mosaic virus (CMV)-based vector [51]. Four Arabidopsis endogenes were silenced using a Tobacco rattle virus RNA 2 (TRV2)-based vector [52]. Similar approach was used to confirm the role of cytosine methylation in promoter of PcMYB10 in anthocyanin biosynthesis in pear fruits [53]. Novel insight on RNA-directed-DNA-methylation, the process which is the basis of virusinduced TGS is presented recently by Bond and Baulcombe (2015) [54].
The VIGS-BSMV vector reported here was constructed to accommodate the two different fragments derived from the ScBx1 promoter region. Four sites recognized by three MSRE (MluI, HhaI and HpyCH4IV) and located within the cloned regions served for analysis of cytosine methylation. Native cytosine methylation in the two target regions differed significantly. The methylation was low (0.5% to 1.6%) in both MSRE sites of the pScBx1-fragment I and it was much higher (67.3% to 100%) in HpyCH4IV and HhaI sites of the pScBx1-fragment II. The level of native methylation was not affected in plants inoculated with the BSMV:00 vector. Methylation in plants inoculated with the experimental vector changed compared with the controls, although the degree was different depending on the analyzed region. Cytosine methylation in the MluI site of pScBx1-fragment I remained unchanged in plants 14 dpi and subsequently (21 and 99 dpi) significantly increased. In the HhaI site the methylation remained at a low level for the whole analyzed period. Cytosine methylation in the HpyCH4IV site of the pScBx1-fragment II was high in the control plants and did not show any significant changes in experimental plants. Methylation of cytosine in the HhaI site was almost 100% in control and in experimental plants.
The significant increase of methylation in the MluI site of pScBx1-fragment I was the most probable cause of ScBx1 silencing and consequently decreased levels of benzoxazinoids in rye leaves. The MluI site with induced cytosine hypermethylation is located in the 5 0 -UTR region of analyzed ScBx1 80 bp upstream of the ATG codon and the pScBx1-fragment I contains two TATA-box domains (Fig 5). According to [55] and [56] the close proximity (about 100 bp) of the induced methylation to transcription start site was crucial for RNA-directed DNA methylation and for transcriptional silencing. The native hypermethylation of the pScBx1-fragment II and no apparent effect on Bx1 gene transcription in control plants agrees with the finding that methylation of more downstream regions of the promoter is unassociated with gene expression [55]. Similarly, the TGS approach to analyze function of the gene was used by Wang et al. (2013). The hypermethylation of the PcMYB10 promoter found in green-skin pear was experimentally induced in corresponding hypomethylated DNA of red-skin pear. The result of Agrobacterium tumefaciens infiltration assay was methylation of the originally hypomethylated PcMYB10 promoter transcriptional silencing of the PcMYB10 gene and anthocyanin-free green-skin phenotype [53]. The results confirmed the role of PcMYB10 promoter methylation in anthocyanin synthesis. It is worth to note that selected method allowed for a detection of methylation changes only within sites recognized by MSRE and the possible methylation in the neighboring regions could not be excluded. Bisulfite sequencing, the alternative to MSRE method, could not be applied because we expected that virus induced methylation pattern was not homogenous in all cells what might lead to ambiguous sequencing results.
The results presented here show that the analyzed ScBx1 gene is functionally involved in benzoxazinoid biosynthesis in rye and its expression is developmentally regulated. Both traits, which also characterize Bx1 orthologs in other monocots, confirm that the analyzed gene is a rye ortholog of the Bx1 gene. Our results prove the applicability of the BSMV-based virus induced transcriptional and posttranscriptional silencing for functional analysis of rye genes. Moreover, based on the obtained results, we anticipate the presence of the Igl ortholog in rye, which contributes along with the ScBx1 to BX biosynthesis in rye. We expect that the gene will be constitutively expressed at least in plants older than 21 dai. The gene remains to be identified.

Plant material and growth conditions
Two rye (Secale cereale L.) hybrid cultivars, Stach F1 and Konto F1, were used. Kernels after 48 hours of imbibition in water-saturated filter paper were cultivated in pots (5 cm diameter) filled with Aura peat substrate (Lasland, Poland) mixed with sand in a 10:3 v/v ratio. The seedlings were cultivated in a growth chamber with a 16 h photoperiod, at 22˚C in the day and 20˚C at night. The relative humidity was in the range 60-80%, and the light intensity was 130-230 μMÁs -1 m -2 . Hypocotyl and leaf samples were collected for transcript quantification and for hydroxamic acid analysis. The first leaves of 5-day old seedlings were used for inoculation with VIGS-BSMV vectors.

Construction of VIGS-BSMV vectors
For VIGS studies the plasmids (pBSMV-T7-α, pBSMV-T7-γMCS and pBSMV-T7-γPDS275) carrying full-length cDNA clones of α and γ molecules of BSMV with the promoter of T7 RNA polymerase [57,58] were kindly provided by Dr. Merete Albrechtsen (University of Aarhus, Frederiksberg, Denmark). The pT7-BSMV-βBamHI plasmid carrying cDNA of the β molecule contained a unique BamHI cloning site engineered between the open reading frame of βc (ORF βc) and poly(A) [30]. The plasmid pBSMV-T7-γMCS contained a multiple cloning site (MCS) with unique PacI, XmaI, BamHI, enabling cloning of DNA fragments upstream of the γb open reading frame. The plasmid pBSMV-T7-γPDS275 contained a 275-bp fragment of barley phytoene desaturase (PDS) encoding gene, accession number AY062039 [58]. The 'empty' BSMV:βBamHI and BSMV:γMCS molecules were described as β (-) and γ (-) respectively. The β and γ molecules of pT7-BSMV-βBamHI and pBSMV-T7-γMCS were modified by insertion of a cDNA fragment for posttranscriptional silencing and gDNA fragments for transcriptional silencing (S1 and S2 Tables, Figs 9 and 5). The fragments were amplified using primers (S1 Table) designed for cDNA (JQ716987) or gDNA (unpublished). The putative regulatory sequences of the ScBx1 promoter were predicted using plant-dedicated SofBerry http:// linux1.softberry.com/all.htm software. Based on the analysis the two fragments were selected. Both contained putative promoter specific motifs as well as the sites recognized by methylation-sensitive restriction enzymes (MSRE) needed for methylation analysis of the region. Amplicons after digestion with BamHI and gel purification were ligated to the BamHI site of the pBSMV-T7-βBamHI and pBSMV-T7-γMCS plasmids. To confirm the presence of the insert the constructs were sequenced using primers designed for the insert-flanking region of pBSMV-T7 plasmids (S1 Table). Fragments of gDNA of the putative ScBx1 promoter selected based on the presence of TATA-box motif and CG islands (Fig 5) (unpublished results) were used for construction of the BSMV:β and BSMV:γ subunits designed for transcriptional gene silencing. Genomic fragment I (324 bp), amplified with primers 2701_Fw and 2991_Re and genomic fragment II (313 bp) amplified with primers 1448_Fw and 1760_Re were cloned into BSMV:βBamHI and BSMV:γMCS respectively (S2 Table, Fig 5).

Nucleic acid isolation, reverse transcription and transcript quantification
Genomic DNA was isolated from young leaves according to the modified CTAB procedure [59]. The RNA was extracted from leaf samples using RNA 3-zone reagent (Novazym, Poland) according to the manufacturer's instructions. RNA concentration and 260/A280 ratio were measured using a NanoDrop spectrophotometer (NanoDrop Technologies, USA). The 260/ A280 ratio was always higher than 1.8 and agarose gel was used to further determine the quality of obtained RNA. Isolated total RNA was treated with 10 U of DNase (DNase I recombinant RNase-free, Roche, USA) and 10 U Protector RNase inhibitor [60] per 10 μg of RNA for 40 min at 37˚C followed by DNase inactivation according to the manufacturer's protocol. Complete digestion was confirmed by PCR with primers specific for actin (AY145451) using 100 ng of DNase digested RNA as a template. No product on agarose gel was detected after 36 cycles of amplification. Four micrograms of RNA (non-degraded, DNase-treated with undetectable gDNA impurities) were used as a template for the reverse transcription reaction with oligo d(T) primers and the RevertAid First Strand cDNA Synthesis Kit (Thermo Scientific, USA). The obtained cDNA was diluted twofold and used directly as a template for quantitative PCR (qPCR). The standard qPCR reaction mix was composed of: 2 μL master mix 5x HOT FIREPol EvaGreen qPCR Mix Plus (Solis Biodyne), 0.2 μL primer F (10 μM), 0.2 μL primer R (10 μM), cDNA 1 μL and water to 10 μL. The reaction was performed in a Rotor-Gene 6000 model 5-plex thermocycler (Corbett). The efficiencies of amplification E for all primer pairs were in the range 0.9 to 1.00 and R 2 0.99875-1.00000. The specificity of amplification was verified by melting curve analysis. Template concentrations ranging from 10 2 to 10 6 copies of analyzed amplicon and 10 3 -10 7 copies of the reference gene HvAct (AY145451.1) per reaction were used as the standards for qPCR. The resulting data were normalized to actin (AY145451.1) using the two standard curves method. Threshold line, Ct values, standard curves and relative quantifications were determined using the proprietary Rotor-Gene 6000 software v 1.7 supplied by the thermocycler manufacturer. The qPCR experiments were performed using cDNA from each analyzed plant with three technical repetitions each. The whole procedure of RNA isolation, reverse transcription, qPCR conditions and data analysis met the MIQE criteria outlined by Bustin et al. [61].

In vitro transcription and plant inoculation with VIGS-BSMV-derived RNA
The plasmids used as templates for in vitro transcription were isolated using the standard alkaline lysis method, linearized with MluI or SpeI (S2 Table) and purified according to the protocol by Kawalek et al. [30]. Resultant DNA, suspended in RNase free water, was used as template for in vitro transcription. Infectious RNA was synthesized using the in vitro transcription kit AmpliCap-Max T7 HighYield Message Maker Kit (CellScript, USA) according to the manufacturer's protocol. A single in vitro reaction with 1 μg of plasmid-derived template yielded 30-40 μg of RNA. One μl of reaction mixture with approximately 2 μg of RNA of each subunit transcript (α, β and γ) was mixed with 18 μl of FES buffer (0.1 M glycine, 0.06 M K 2 HPO 4 , 1% Na 4 P 2 O 7 (w/v), 1% Bentonite (w/v), 1% Celite (w/v)) [62]. The resulting 21 μl was rubbed into the leaf surface of 5-day-old rye seedlings dusted with silicon carbide (Silicon carbide 400 mesh, Sigma-Aldrich). The leaves of the control plants were powdered with silicon carbide and rubbed without or with FES buffer. Directly after the inoculation, the plants were transferred to a high humidity growth chamber (26˚C, 100% relative humidity, 16 h photoperiod and 130-230 μM photons s -1 m -2 light intensity). After 2 days, the plants were transferred to standard growth conditions. From several VIGS-BSMV inoculated plants only plants showing virus infection and/or PDS silencing (i.e. leaf photo bleaching) symptoms were selected for further studies. The leaves were collected 14 and 21 or 14, 21 and 99 days post inoculation (dpi) for further analysis of posttranscriptional or transcriptional silencing respectively (Fig 10). At least three plants for each vector / time point combination from two independent VIGS experiments were selected and analyzed. In order to show the silencing of the analyzed Bx1 gene in the experimental plants the relative expression of the gene in plants inoculated with the BSMV vector (designated as BSMV:PDS or BSMV:00) was assumed as 1.00. This value was subsequently used as a basis to calculate the normalized expression of ScBx1 in the experimental plants inoculated with the ScBx1 silencing VIGS vectors. The outcome was presented on the logarithmic scale where values below the base line 1.00 indicated that the expression of ScBx1 in experimental plants was lower than expression in plants inoculated with control BSMV:PDS or BSMV:00.
Detection and quantification of cytosine methylation in the selected region of rye gDNA Cytosine methylations in target regions of the genomic DNA fragment I and fragment II ( Fig  5) was quantified using methylation sensitive restriction enzymes (MSRE) and the qPCR approach according to the procedure of OneStep qMethyl Kit (Zymo Research, USA). Briefly: 0.9 μg of gDNA isolated from control and experimental plants (i.e. inoculated with BSMV:α, β (-) ,γ (-) or BSMV:α,β (pScBx1-fragment I) ,γ (pScBx1-fragment II) respectively) were suspended in a restriction buffer. Each of them was split into a test and a reference sample and only the test sample was digested with MSRE. DNA from the test and reference samples served as qPCR templates for amplification with primers (2811_Fw/2994_Re and 1448_Fw/1760_Re, S1 Table) flanking the analyzed region. The difference between cycle threshold (Ct) values for test and reference DNA samples (i.e. ΔCt) was used to calculate the ratio of cytosine methylation according to the equation below [63]: Cytosine methylation ratio ð%Þ ¼ 100 x 2 À DCt Sample preparation and analysis of selected benzoxazinoids 100 mg of lyophilized leaf samples were mixed with diatomaceous earth (ASE Prep DE, Dionex, Sunnyvale, CA) and extracted with 70% methanol in stainless steel extraction cells of an accelerated solvent extraction system (ASE 200, Dionex, Sunnyvale, CA). Extractions were carried out at 10 MPa operating pressure and 40˚C. After evaporation to dryness under reduced pressure, extracts were reconstituted in 1 ml of methanol containing 0.1% (v/v) acetic acid and stored at -20˚C. Before the analyses, extracts were centrifuged for 20 min at 23 000 x g at 4˚C. Quantitative analyses were performed on a Waters Acquity UPLC system (Waters, Milford, MA) equipped with a triple quadrupole mass spectrometer (Waters TQD). BXs and their metabolites were separated on a Waters BEH C18 column (50 x 2.1 mm, 1.7 μm) with a linear, 7 min long gradient from 3 to 10% of acetonitrile containing 0.1% (v/v) formic acid (solvent B) in 0.1% formic acid (solvent A). All separations were carried out at 50˚C at the flow rate of 700 μlÁmin -1 . After each gradient elution, the column was washed with 90% solvent B for 1 minute and then re-equilibrated with 3% solvent B in solvent A for 3 minutes prior to the next injection. Injections were done in the "partial loop needle overfill" mode of a Waters Acquity autosampler. 2.5 μl was injected from each sample and the analysis was repeated twice. The column's effluent was introduced into the ion source of the mass spectrometer, which operated in the negative ion mode with the following parameters: capillary voltage -2.8 kV, extractor 3 V, RF lens 100 mV, source temperature 130˚C, desolvation temperature 400˚C, desolvation VIGS-based analysis of Bx1 ortholog in rye gas flow 1000 lÁh -1 , cone gas flow 100 lÁh -1 . The collision cell entrance and exit were set to -2 and 0.5, respectively. Parameters of quadrupoles 1 and 3 were set to achieve unit-mass resolution. Cone voltage and collision energy were optimized for each compound to attain a maximal response ( Table 3).
The quantitation method was calibrated from the standard solutions of DIBOA, DIMBOA, HBOA, DIMBOA-Glc and MBOA. DIMBOA-Glc was used as a reference standard for DIBOA-Glc quantitation. Calibration was performed between 2.5 and 30 ngÁμl -1 , and it was found to be linear only within this range. Samples with concentrations of analytes higher than 30 ngÁμl -1 were appropriately diluted (typically between 2 and 5 times) using 0.1% formic acid and re-analyzed. Waters MassLynx 4.1 SCN 849 software was used for data acquisition and processing.

Statistical analysis
The statistical significance of the change (Student's t-test) was calculated as Ã p < 0.05 and ÃÃ p < 0.005. Pearson's coefficient was used to determine whether there was a relationship in the expression of ScBx1 and benzoxazinoid content. The statistical analyses were performed using the software package STATISTICA (Data Analysis Software System, StatSoft Inc.). All quantitative results of cytosine methylation, transcripts and BXs levels represented medium value from three technical repetitions of single individual plant inoculated with BSMV-based vector.
Supporting information S1 Table. List of primers and reaction conditions. Shading indicates BamHI recognition site used for amplicon cloning.