TY - JOUR
AU - Lomonossoff, George
AB - Abstract We previously reported the novel partitioning of regional GFP-silencing on leaves of 35S–GFP transgenic plants, coining the term “partitioned silencing”. We set out to delineate the mechanism of partitioned silencing. Here, we report that the partitioned plants were hemizygous for the transgene, possessing two direct-repeat copies of 35S–GFP. The detection of both siRNA expression (21 and 24 nt) and DNA methylation enrichment specifically at silenced regions indicated that both post-transcriptional gene silencing (PTGS) and transcriptional gene silencing (TGS) were involved in the silencing mechanism. Using in vivo agroinfiltration of 35S–GFP/GUS and inoculation of TMV–GFP RNA, we demonstrate that PTGS, not TGS, plays a dominant role in the partitioned silencing, concluding that the underlying mechanism of partitioned silencing is analogous to RNA-directed DNA methylation (RdDM). The initial pattern of partitioned silencing was tightly maintained in a cell-autonomous manner, although partitioned-silenced regions possess a potential for systemic spread. Surprisingly, transcriptome profiling through next-generation sequencing demonstrated that expression levels of most genes involved in the silencing pathway were similar in both GFP-expressing and silenced regions although a diverse set of region-specific transcripts were detected.This suggests that partitioned silencing can be triggered and regulated by genes other than the genes involved in the silencing pathway. Nicotiana benthamiana, GFP-partitioned silencing, cell-autonomous-like silencing, post-transcriptional gene silencing, systemic silencing. Introduction RNA silencing, a sequence-specific process of RNA degradation, is initiated by the formation of complementary double-stranded RNAs (dsRNAs; Goto et al., 2007; Vaistij et al., 2002). dsRNAs can be directly generated from transcripts of inverted repeat (IR) transgenes (Broderson and Voinnet, 2006; Dunoyer et al., 2005), imperfectly matched fold-back RNAs (Hutvagner et al., 2001; Park et al., 2002), viral dsRNAs (Deleris et al., 2006), and the coordinated action of the plant-specific DNA-dependent RNA polymerase IV (POLIV) and RDR2 proteins (Pontes et al., 2006; Xie et al., 2004). In each case, dsRNAs are subsequently processed into small interfering RNAs (siRNAs) of 21–24 nt by Dicer-like proteins (DCLs). siRNAs participate in the cleavage or translation inhibition of sequence-specific target mRNAs through the action of RNA-induced silencing complex (RISC)-containing argonaute complexes (AGOs; Molnar et al., 2011), and in some cases through epigenetic modification of DNA and histone proteins (Ghildiyal and Zamore, 2009; Jones et al., 1999). Aberrant RNAs (aRNAs) are defined as improperly metabolized RNAs, canonically produced from transgenes, and include incorrectly spliced mRNAs (Cheng et al., 2003), mRNAs whose termination or polyadenylation is impaired (Herr et al., 2006; Luo and Chen, 2007), and collateral products of normal transcripts from the high expression of direct-repeat (DR) transgenes (Luo and Chen, 2007). aRNAs result in gene silencing through a mechanism that requires RDR6 (Allen et al., 2005; Wassenegger and Krczal, 2006; Yoshikawa et al., 2005). aRNAs are copied to form complementary dsRNAs, which feed into the siRNA pathway to subsequently degrade target mRNAs. In Neurospora, a DNA/RNA-dependent RNA polymerase, QDE-1, which interacts with a replication protein A, was shown to be essential for generating aRNA (Lee et al., 2010). Nevertheless, the pathway by which aRNAs are sensed inside the plant cell to initiate active silencing is still to be delineated (Kalantidis et al., 2008; Molnar et al., 2011). After the initiation of silencing, mobile signal molecules have been shown to spread from the initial sites of silencing to other parts of the plant through both short-distance (cell-to-cell movement) and long-distance (source-to-sink phloem flow) systemic transmission. Silencing is continually spread to the rest of the plant non-cell autonomously. For short-distance transmission, silencing initially spans the adjacent 10–15 cells via plasmodesmata, and then advances through another 10–15 cells (Frizzi and Huang, 2010; Himber et al., 2003). For long-distance transmission, silencing has been shown to systematically spread to recipient root-stocks or scions through grafting experiments (Palauqui et al., 1997; Tournier et al., 2006; Voinnet and Baulcombe, 1997; Voinnet et al., 1998). The spread of silencing via the phloem was more efficient in the shoot-to-root direction in Arabidopsis (Molnar et al., 2010). The signal molecules responsible for communicating gene silencing are 21–24 nt siRNAs. siRNAs have been detected in recipient cells such as the phloem and its neighbouring cells. These mobile siRNAs trigger the activation of silencing pathways, resulting in the degradation of target mRNAs in a sequence-dependent manner (Bai et al., 2011; Dunoyer et al., 2005; Kasai et al., 2011; Melnyk et al., 2011). Secondarily produced siRNAs move farther to neighbouring cells to continually degrade target mRNAs and to guide protein factors that confer epigenetic modifications such as DNA and histone methylation (Molnar et al., 2010). In rice, siRNAs targeted to promoter regions were shown to induce de novo DNA methylation and histone modifications (Miki and Shimamoto, 2008; Okano et al., 2008; Vaucheret, 2005). Methylated DNA is thought to act as a template for aRNA transcription, which repeatedly forms dsRNAs to create a self-perpetuating RNA silencing loop (Vaucheret, 2005). Such methylated DNA and histone modifications (Cheung and Lau, 2005), known as the direct causes of TGS, were reported to mediate epigenetic regulation by changing the transcriptome profile of gene expression in plants at a specific developmental stage (He et al., 2011). Epigenetic regulation is important for many aspects of plant physiology, including plant development, defence against pathogens, and enhanced adaptation to the natural environment (Molnar et al., 2010; Xie et al., 2004). Small RNA is one of the factors known to influence epigenetic modifications in plants. To investigate the dynamics and causes of spontaneous gene silencing, we isolated a specific transgenic plant defined as having a ‘GFP-partitioned’ phenotype, isolated from the diverse phenotypes of 35S–GFP transgenic N. benthamiana generated in a previous study (Sohn et al., 2011b). The transgenic plant possessed a characteristic pattern of GFP-expressing and silenced regions, even within a single leaf, despite having identical genetic and environmental backgrounds. Our study uses this unique phenotype to delineate the features of this specific silencing, carrying out comprehensive analyses of the phenotypic characteristics of silencing, transgene structure, the transcriptome profile including silencing pathway-involved genes (SPIGs) by NGS, and the competence of systemic silencing. Finally, possible factors to regulate the partitioned silencing are discussed. Materials and methods Production of GFP-partitioned transgenic N. benthamiana The GFP-partitioned plant was previously isolated from hundreds of transgenic plants transformed with Agrobacterium tumefaciens harbouring pGreen0229-35S:GFP (35S–GFP) via leaf-disc transformation. The GFP-partitioned plants had been selected over several generations to immobilize their phenotype up to the T5 generation. The pedigree numbers of progeny plants were designated as T1-T2-T3-Tn according to the progress of the generation (Sohn et al., 2011b). GFP phenotypes and expression patterns were photographed with an Olympus C-5050Z digital camera under UV illumination (BIB-150P; Spectronics Corp. USA), and microscopic images of GFP fluorescence were taken by a stereomicroscope (Olympus SZX12; Japan). Conditions for agroinfiltration and virus inoculation To determine the dominant pathway of silencing, Agrobacterium cells harbouring pGreen0229-35S:GUS (35S–GUS) or 35S–GFP were co-infiltrated with the presence of viral suppressors, pGreen0229-35S:2b (35S–2b; Choi et al., 2008) and pGreen0229-35S:NS3 (35S–NS3; Sohn et al., 2011a). The 2b protein has the ability to enhance the expression of a co-infiltrated foreign gene by suppressing the onset of PTGS in plant cells (Choi et al., 2008). Non-structural protein 3 (NS3) of a rice stripe virus (RSV) also displays powerful enhancement of transient expression of foreign genes (Sohn et al., 2011a). Our detailed agroinfiltration procedures are as previously described (Choi et al., 2008). A TMV-based vector expressing GFP (TMV–GFP; Rabindran et al., 2001) was prepared and inoculated onto partitioned plants as previously described (Choi et al., 2009). Southern and northern blot analysis, and siRNA detection Genomic DNA (gDNA) was digested with HindIII, which cleaves at a single site localized between the 35S terminator (T35S) and nos promoter (Pnos) in the T-DNA region of pGreen0229. The digested gDNA was then electrophoresed on a 0.6% agarose gel. The copy numbers derived from the Southern blot were compared with the segregation ratios of Basta®-resistant or susceptible plants, to confirm the accuracy of the transgene numbers. To investigate transgene methylation, gDNA was fully digested with Sau96I. The analysis by Sau96I digestion was a simple method to approximately examine the extent of DNA methylation in the 35S–GFP cassette (Jones et al., 1999; Vaistij et al., 2002). The steady-state level of GFP RNA transcripts and detection of siRNAs were performed as previously described (Sohn et al., 2011a). For hybridization and visualization of bands, GFP or bar genes labelled with 32P–dCTPs were used as the probes, as previously reported (Sohn et al., 2011a). Structure of transgenes in the plant genome To determine the structure of transgenes in transgenic N. benthamiana, gDNA from leaf tissues of GFP-expressing and silenced regions separately was fully digested with EcoRI. The T-DNA has a single cleavage site of EcoRI located ahead of the 35S promoter (P35S). The digested DNAs were then size-fractionated on a 0.6% agarose gel. The agarose gel containing the 2–10kb fragments was cut into 15 blocks, and the embedded gDNAs extracted and confirmed by Southern blotting to localize to the gDNA fragment fused to the T-DNA (gDNA:T-DNA). The gDNA:T-DNA fragment (~5.8kb) was inserted into the pUC18 vector and selected using a 32P–dCTP-labelled GFP gene. Subsequently, both the 5’- and 3’-flanking sequences were cloned using the GeneFishing™ PCR kit (Seegene K1021, USA) and sequenced. The nucleotide sequence of the transgene was registered at EMBL under the accession number HF675000. NGS transcriptome analysis To investigate the differentially expressed genes (DEGs) between GFP-expressing and silenced regions, RNA-seq data (four replicates) was obtained using the HiSeq 2000 platform (Illumina, USA). Processing of the paired-end sequences (101bp read length) was conducted using CLC Genomics Workbench (CLC Bio, ver. 5.1.5, Denmark), and sequenced bases with a quality lower than Q20 or equivalent were discarded. The total numbers of reads for the GFP-expressing and -silenced regions were 87 122 428 and 88 841 280, respectively. The 100 480 contigs were assembled (Supplementary Data) with the RNA-seq data (ERP002109, EMBL) at the guidance of ESTs (56 102 ESTs of N. benthamiana, NCBI) and registered to EMBL under the accession CBMM010000001. To analyse DEGs, all the sequence reads were mapped to the contigs using the CLC Genomics Workbench, and then their read counts were analysed by DESeq package R-3.0.2 (http://www.R-project.org/). DEGs, which were differentially expressed by over 2-fold between the two regions, were obtained using the cutoffs of upper-quartile normalization 40% and false discovery rate (FDR, P≤0.01, Benjamini and Hochberg corrected). The annotation and function of DEGs were analysed using BLASTx (E<10–5, NCBI) and Blast2GO (Conesa et al., 2005). The validation for expressional difference between two regions was analysed with representative SPIGs and region-specific genes by quantitative RT-PCR (qRT-PCR, Lim and Ha, 2013). Transgenic plant grafting The steady-green plant (266-3-1) was introduced as the recipient plant, which is a single-copied transgenic plant strongly expressing GFP (Sohn et al., 2011b). The donor, a partitioned plant, was used as either the stock or scion to track the direction of systemic spreading. These two transgenic plants were grown for about 6–8 weeks post sowing, and a wedge-grafting method was performed. The grafted plants were maintained under humid conditions for 2 weeks to prevent the scions from severely dehydrating. Then, they were transferred and grown further until the seeds had set for almost 3 months. Results The phenotypic characteristics of partitioned silencing Partitioned-silenced plants typically showed a phenotypic transition on the leaf that does not alter as the plant grows. At early stages following seed germination, approximately five fully green fluorescent leaves were generated, which were then followed by approximately three partitioned-silenced leaves. After this point, almost fully silenced leaves dominated until the flowering stage (Figs 1A and 2A–D). Thus, the partitioned silencing was observed between lower GFP-expressing and upper GFP-silenced leaves. Similar to our stereoscopic observations, the GFP-expressing and silenced regions (outer and inner blades, respectively) of the partitioned leaves were separated with a very clear border (Fig. 1B–D). This GFP expression pattern was also observed on otherwise fully silenced leaves, specifically at the distal ends (tips) of the leaves (Fig. 1A), and on limbs and flower sepals (Fig. 1E–H). The shapes of the partitioned borders (localized on the leaf blade and main stem) and the area of silenced regions or spots (red drops on the leaves) did not enlarge or change, despite the plants’ dynamic growth (white- and yellow-dashed squares in Fig. 2A–D). The lack of change in partitioned regions was confirmed by trichome location markers that were adjacent to the silenced regions (arrowheads; Fig. 2E–H). Partitioned leaves also appeared on lateral shoots developing from the main stem following flowering, where the GFP-expressing and silenced regions were adjacent to each other (PL1 and 2; Fig. 2I, J). Fig. 1. View largeDownload slide Representative phenotypes of partitioned-silenced plant. (A) GFP expression at the mature partitioned plant (257-9-8-12, T3). The regions indicated by arrows of B, C, and D are enlarged in B, C, and D. (B) GFP-expressing region at the outer-blade of the leaf. TC indicates trichomes. (C) Partitioned borderline (Pb) between the GFP-expressing and silenced regions. (D) Silenced region at the inner-blade of the leaf. (E, F) Specific GFP expression in limbs and sepals of a flower, photographed under visible and UV light, respectively. (G, H) Maintenance of specific GFP expression in sepals until seed maturation. Fig. 1. View largeDownload slide Representative phenotypes of partitioned-silenced plant. (A) GFP expression at the mature partitioned plant (257-9-8-12, T3). The regions indicated by arrows of B, C, and D are enlarged in B, C, and D. (B) GFP-expressing region at the outer-blade of the leaf. TC indicates trichomes. (C) Partitioned borderline (Pb) between the GFP-expressing and silenced regions. (D) Silenced region at the inner-blade of the leaf. (E, F) Specific GFP expression in limbs and sepals of a flower, photographed under visible and UV light, respectively. (G, H) Maintenance of specific GFP expression in sepals until seed maturation. Fig. 2. View largeDownload slide The time-course observations of the change in partitioned-silenced patterns during the active growing stage, from 45–56 days post sowing. (A–D) The time-course development of the fully-green, partitioned, and fully silenced leaves in sequence. The dashed squares indicated by E to J are enlarged in E to J. (E–H) The time-course enlarged images of two silenced spots and a partitioned borderline (Pb). The maintenance of the initial shapes (e.g. sizes and pattern) was demonstrated by the locations of trichomes, indicated by arrows. (I, J) Another pattern of partitioned silencing on the main stem and lateral leaves. The location and pattern of the Pb on the stem, and PL 1 and 2 were maintained for 12 days. Fig. 2. View largeDownload slide The time-course observations of the change in partitioned-silenced patterns during the active growing stage, from 45–56 days post sowing. (A–D) The time-course development of the fully-green, partitioned, and fully silenced leaves in sequence. The dashed squares indicated by E to J are enlarged in E to J. (E–H) The time-course enlarged images of two silenced spots and a partitioned borderline (Pb). The maintenance of the initial shapes (e.g. sizes and pattern) was demonstrated by the locations of trichomes, indicated by arrows. (I, J) Another pattern of partitioned silencing on the main stem and lateral leaves. The location and pattern of the Pb on the stem, and PL 1 and 2 were maintained for 12 days. In general, silencing has been shown to spread systemically in a non-cell autonomous manner (Frizzi and Huang, 2010; Himber et al., 2003). However, in our transgenic plants, the initial pattern of partitioned silencing remained unchanged (Fig. 2), persisting until flowering, by which point the partitioned-silenced leaves were senescent (decolourization in Fig. 1A).Maintenance of the initial silencing pattern suggests that, in a manner different from the well-known systemic spreading of silencing, silencing in the partitioned plant is determinate once it has been established. This also suggests the presence of a region-specific and self-governing mechanism of silencing. If this is so, it is possible that unique regulation systems exist between two regions of the same cell-type and developmental stage, and therefore this partitioned-silenced plant is a unique phenotype that enables exploration of specific genes or mechanisms involved in determinate or cell-autonomous silencing. Under high-temperature growth conditions, especially in the summer season, a breakdown of partitioned silencing was observed (Supplementary Data). For example, when grown in temperatures over 30 °C, the phenotype of partitioned plants transformed into a variegated pattern (red arrows in Supplementary Data and Supplementary Data, Supplementary Data), or the initiation of silencing seemed to be activated earlier, whereby fully silenced leaves emerged in the absence of partitioned-silenced leaves (yellow arrows in Supplementary Data). However, grown below 30 °C, the partitioned-silenced phenotype re-appeared stably (Supplementary Data). This phenomenon indicates that the active initiation and establishment of silencing is affected by temperature, and that high temperature may accelerate silencing initiation, affect a wider region, or bypass the intermediate partitioned stage. This effect is consistent with previous results showing an increase in RNA silencing and high-level accumulation of siRNAs under conditions of increased temperature (Chellappan et al., 2005; Szittya et al., 2003). Partitioned silencing appears only in hemizygous transgenic plants To immobilize the partitioned-silenced phenotype, we analysed GFP expression from partitioned plants down to the T5 generation and found repeated segregation into four different phenotypes; (i) fully silenced, (ii) steady-green, (iii) partitioned, and (iv) non-transgenic plants (Table 1). The fully silenced plants generated a fully silenced phenotype in all progeny, whereas the steady-green and partitioned plants always segregated into an approximately 1:2:1 ratio of fully silenced: steady-green plus partitioned: non-transgenic plants. This suggests that the fully silenced plants were probably homozygous for the transgene. In contrast, the steady-green and partitioned plants might be hemizygous plants. However, the phenotypic development of the hemizygous plants into either the steady-green or partitioned phenotype was spontaneous. Unless the partitioned-silenced leaf emerged until the growth stage with more or less the tenth leaf, the plant would remain steady-green for the rest of its life. We hypothesized that epigenetic modification of the GFP transgene was responsible for the alteration of GFP transcription in segregating progeny, and sought to examine the cause of the spontaneous partitioned silencing. Table 1. The phenotypic segregations of T5 progeny plants derived from the transgenic plant 257 (A) The phenotype and pedigree number of progenitor plants (T4). (B) The number of plants classified by phenotypes and herbicide resistance. (C, D) The phenotypic and herbicide-resistant segregation ratios calculated by summing up those from three progenitor plants excluding fully silenced phenotype (257-9-5-7-12). Progenitor plants (T4) A  No. of herbicide (Basta®) resistance and phenotype segregation (T5) B  Phenotypes  Pedigree number  Resistant  Susceptible          Silenced  Steady-green  Partitioned  Non-transgenic  Fully silenced  257-9-5-7-12  24  0  0  0  Steady-green  257-9-5-7-7  6  12  1  5  Partitioned  257-9-5-7-9  10  5  2  6  Partitioned  257-9-8-3-10  3  11  5  5  Phenotype segregationC  19  36  16  Herbicide resistanceD  55    16  Progenitor plants (T4) A  No. of herbicide (Basta®) resistance and phenotype segregation (T5) B  Phenotypes  Pedigree number  Resistant  Susceptible          Silenced  Steady-green  Partitioned  Non-transgenic  Fully silenced  257-9-5-7-12  24  0  0  0  Steady-green  257-9-5-7-7  6  12  1  5  Partitioned  257-9-5-7-9  10  5  2  6  Partitioned  257-9-8-3-10  3  11  5  5  Phenotype segregationC  19  36  16  Herbicide resistanceD  55    16  View Large The DR structure of two transgenes Seven progeny plants derived from a partitioned progenitor plant, including two fully silenced, two steady-green, and three partitioned plants (Fig. 3A–G), were analysed for their number, structure, and expression of the transgenes. For the partitioned plants particularly, DNA and RNA were extracted excluding the neighbouring regions of the border to prevent contamination from opposing regions. Southern blot analysis showed that all seven plants contained two copies of the transgene (Fig. 3H). Consistent with this, all progeny from fully silenced plants survived Basta® spray treatment (the ‘Fully silenced’column in Table 1). In contrast, the steady-green (Fig. 3A, B) and partitioned plants (Fig. 3E–G) demonstrated a 3:1 segregation of the herbicide resistance trait in their progeny (Table 1), indicating that these plants were hemizygous. The segregation of 3:1 is highly suggestive of two transgenes having been incorporated into the same locus, thereby behaving like a single dominant gene in phenotypic segregation. In fact, two transgenes were confirmed to be incorporated as a DR. The two GFP genes were interrupted with the backbone sequences of pGreen0229 (Fig. 3I). Fig. 3. View largeDownload slide Molecular analysis of copy number, RNA transcription, and methylation of the 35S–GFP transgene in seven progeny plants from the partitioned progenitor plant. (A–G) The progeny plants segregated into the steady-green (S;, A, B), silenced (S; C, D), and partitioned (P; E–G) phenotypes. PG-n and PS-n indicate the partitioned-green and -silenced regions, respectively. (H) The incorporation of two GFP transgenes. WT indicates the wild-type, non-transgenic plant. (I) Transgene structure incorporated into the plant genome. The vector backbone was inserted between the two GFP transgenes. The flanking genomic sequence has high similarity with Niben-scf929-contig 3 (http://www.solgenomics.net). (J) The detection of GFP RNA transcripts by northern blot analysis. (K) The evidence of DNA methylation on the 35S–GFP cassette with the methylation-sensitive restriction enzyme Sau96I. (L) The map of schematic digestion by Sau96I of the 35S–GFP cassette. The red or black lines below indicate the expected DNA fragments fully or incomplete digested. (M) The evidence of unmethylation of the nos-bar cassette, the selection marker of pGreen0229-35:GFP, with Sau96I. Fig. 3. View largeDownload slide Molecular analysis of copy number, RNA transcription, and methylation of the 35S–GFP transgene in seven progeny plants from the partitioned progenitor plant. (A–G) The progeny plants segregated into the steady-green (S;, A, B), silenced (S; C, D), and partitioned (P; E–G) phenotypes. PG-n and PS-n indicate the partitioned-green and -silenced regions, respectively. (H) The incorporation of two GFP transgenes. WT indicates the wild-type, non-transgenic plant. (I) Transgene structure incorporated into the plant genome. The vector backbone was inserted between the two GFP transgenes. The flanking genomic sequence has high similarity with Niben-scf929-contig 3 (http://www.solgenomics.net). (J) The detection of GFP RNA transcripts by northern blot analysis. (K) The evidence of DNA methylation on the 35S–GFP cassette with the methylation-sensitive restriction enzyme Sau96I. (L) The map of schematic digestion by Sau96I of the 35S–GFP cassette. The red or black lines below indicate the expected DNA fragments fully or incomplete digested. (M) The evidence of unmethylation of the nos-bar cassette, the selection marker of pGreen0229-35:GFP, with Sau96I. Sequence analysis of the 5’-flanking region of the transgenes revealed a loss of the majority of right-border (RB) sequences. As previously reported, during the process of Agrobacterium-mediated transformation, vector backbones are often co-integrated with a frequency of 45% owing to the failure of T-strand termination of the left-border (LB) sequence, and DR structures were also observed with a frequency of 33% in rice (Kim et al., 2003). The loss of RB sequences is also a general phenomenon during the process of T-DNA incorporation (Kim et al., 2003; Kumar and Fladung, 2002). The 5’-flanking genomic sequences were matched to the scaffold #929-contig 3 of N. benthamiana (the draft genome ver. 0.4.2, sol genomics network). However, owing to the lack of detailed gene annotation of that scaffold, it was impossible to identify the number and location of the chromosomes where the transgenes were integrated. The potential genes in the flanking sequences that allow transcription of the anti-sense RNA of the transgenes were not found based on BLASTN analysis. Accordingly, no anti-sense GFP RNA for the production of dsRNA was detected in a northern blot analysis using anti-sense-specific oligo probes (data not shown). Our inability to detect dsRNA leads us to implicate aRNA as the trigger for silencing. The dominant mechanism leading to partitioned silencing Consistent with the GFP expression phenotypes observed using stereoscopy, GFP mRNAs were strongly expressed in steady-green plants (SG-1 and SG-2) but not in fully silenced plants (Fig. 3J). As expected, we also found GFP mRNA expression in the green fluorescent regions (PG-1 to PG-3) but not in the partitioned-silenced regions (PS-1 to PS-3) of the same partitioned plants (Fig. 3J). To examine DNA methylation of the 35S–GFP transgenes, genomic DNAs were digested with the CpG methylation-sensitive enzyme, Sau96I, and Southern blotting with a 35S–GFP probe was performed (Jones et al., 1999; Vaistij et al., 2002). DNA fragments of less than 0.5kb (red lines in Fig. 3L) were detected in all green regions (SG-1, SG-2 and PG-1 to PG-3), indicating that the 35S–GFP DNAs were not methylated in GFP-expressing leaves (Fig. 3K). In contrast, larger fragments ranging from 0.7 to over 1.0kb (black lines in Fig. 3L) were detected in all silenced regions (S-1, S-2 and PS-1 to PS-3) (Fig. 3K). These results suggest that 35S–GFP DNAs were at least partially methylated in the GFP-silenced or partitioned-silenced regions. Much stronger bands were observed in the fully silenced plants (S-1 and S-2) compared to that of the partitioned-silenced regions. This is probably due to the presence of increased CpG methylation at the GFP locus in the fully silenced plants. We also checked methylation of the bar gene which is located close-by the 35S–GFP transgene. Consistent with exhibiting Basta® resistance in all transgenic plants, the same band patterns of the bar gene on the Southern blot were observed in all transgenic plants (Fig. 3M). This indicated that the bar gene was not methylated and that the GFP locus was specifically methylated. Small RNAs are well-known to silence their target genes either transcriptionally or post-transcriptionally (Miki and Shimamoto, 2008; Molnar et al., 2010). We found that the 21 and 24 nt siRNAs of the GFP transgene were clearly detected in the fully silenced and partitioned-silenced regions only (PS, S in Fig. 4A, C). Taken together with our evidence regarding DNA methylation and siRNA detection, the partitioned silencing seems to be associated with both PTGS and TGS pathways. Fig. 4. View largeDownload slide The detection of siRNAs in the different regions of a partitioned plant. (A) The four different regions assayed are fully green (G), partitioned green (PG), partitioned silenced (PS), and fully silenced (S). (B) The GFP RNA transcripts detected only in the green regions (G and PG). WT indicates the wild-type, non-transgenic plant. (C) The 21 and 24 nt siRNAs detected only in the silenced regions (S1 and S2). Fig. 4. View largeDownload slide The detection of siRNAs in the different regions of a partitioned plant. (A) The four different regions assayed are fully green (G), partitioned green (PG), partitioned silenced (PS), and fully silenced (S). (B) The GFP RNA transcripts detected only in the green regions (G and PG). WT indicates the wild-type, non-transgenic plant. (C) The 21 and 24 nt siRNAs detected only in the silenced regions (S1 and S2). To delineate which pathway may play a more major role in partitioned silencing, agroinfiltrations with 35S–GFP/GUS binary vectors (Fig. 5A, B) and inoculation with RNA transcripts of the recombinant TMV-harbouring GFP gene (TMV–GFP, Fig. 5E) were applied to the partitioned-silenced leaves. Viral suppressors 2b or NS3 were co-infiltrated (Fig. 5C, D) to enhance and clarify the transient expression of the 35S–GFP/GUS vectors (Choi et al., 2008; Sohn et al., 2011a). However, such an enhanced expression by 2b was always limited to GFP-expressing regions (Fig. 5G, H). Neither a reversion nor an alteration of the silencing pattern was observed, consistent with our previous data (Sohn et al., 2011b). This suggests that although 2b prevents the initiation of PTGS through direct binding to siRNAs (Goto et al., 2007), it does not affect the already established gene silencing, as previously reported (Brigneti et al., 1998; Li et al., 1999). For TMV–GFP inoculation, we first confirmed the green hotspots in WT plants, indicating that TMV–GFP had successfully infected the plant and been amplified, and that the plants expressed viral GFP (Fig. 5I). Then, we observed green spots only in the GFP-expressing regions but not in the partitioned-silenced regions (Fig. 5J). Even in the enlarged image (Fig. 5K), green hotspots were seen adjacent to the silencing region but unable to penetrate it. This finding suggests that TMV–GFP RNAs lost their infectivity in the silencing region, possibly from co-degradation of the TMV–GFP RNA by the established GFP-specific RNA silencing. Fig. 5. View largeDownload slide Evidence of PTGS as the leading machinery in partitioned silencing. (A–E) The schematic gene maps of 35S–GFP, 35S–GUS, 35S–2b, 35S–NS3, and TMV–GFP. (F) The location of agroinfiltration for 35S–GFP in the absence or presence of 35S–2b. (G) The partitioned leaf prior to agroinfiltration. (H) The enhanced GFP-expression after 3 days post-infiltration (dpi) on G. (I, J) The GFP spot on wild-type (WT) and partitioned leaf at 4 dpi of TMV–GFP inoculation. (K) The enlarged image of the dashed-square in J under a stereoscopic microscope. (L–O) Before (L, N) and after GUS-staining (M, O) following infiltration with 35S–GUS in the presence of 35S–2b and 35S–NS3, respectively, at 3 dpi under visible or UV light. Fig. 5. View largeDownload slide Evidence of PTGS as the leading machinery in partitioned silencing. (A–E) The schematic gene maps of 35S–GFP, 35S–GUS, 35S–2b, 35S–NS3, and TMV–GFP. (F) The location of agroinfiltration for 35S–GFP in the absence or presence of 35S–2b. (G) The partitioned leaf prior to agroinfiltration. (H) The enhanced GFP-expression after 3 days post-infiltration (dpi) on G. (I, J) The GFP spot on wild-type (WT) and partitioned leaf at 4 dpi of TMV–GFP inoculation. (K) The enlarged image of the dashed-square in J under a stereoscopic microscope. (L–O) Before (L, N) and after GUS-staining (M, O) following infiltration with 35S–GUS in the presence of 35S–2b and 35S–NS3, respectively, at 3 dpi under visible or UV light. In contrast, for 35S–GUS infiltration with 2b or NS3, GUS was entirely expressed on the partitioned-silenced (Fig. 5L–O) and variegated leaves (Supplementary Data). This result suggests that the potential of instant methylation on the 35S promoter might be ineffective or too weak to quickly silence the transcription of GUS expression. These results strongly suggest that the PTGS plays a primary role in silencing and is therefore at least partially responsible for the partitioned silencing. According to previous reports (Frizzi and Huang, 2010; Vaucheret, 2005), 24 nt siRNAs also induce DNA and histone modifications with the involvement of the RISC. Then, the methylated DNA is thought to become inactivated in RNA transcription, or to act as a template for transcribing aRNAs. Hence, the central pathway of partitioned silencing is analogous to the RdDM pathway (Eamens et al., 2008; Molnar et al., 2011). NGS transcriptome analysis on silencing pathway-involved genes (SPIGs) The comparison of overall transcriptome profiles between GFP-expressing and silenced region was performed. Region-specific genes (contigs) were analysed by comparing the expressional differences between the two regions. Numbers of GFP-expressing and silenced region-specific genes using DESeq (P≤0.01, Benjamini and Hochberg-corrected), a BLASTX analysis (E<10–5), and Blast2GO were 281 and 81 respectively (Supplementary Data, Supplementary Dataand Supplementary Data). Over three times more numbers of region-specific genes were identified in GFP-expressing regions, and among them it is notable that diverse kinds of nucleic acid-related and responsive genes to various stimuli (e.g., stress, (a)biotic, chemical, ethylene, etc.) were characteristically expressed there. Expressional differences of 11 randomly selected region-specific genes, which showed over 2-fold expressions between the two regions (Supplementary Data), were verified by qRT-PCR between inner and outer blades of leaves of WT plants (Fig. 6). The enhanced expression of region-specific genes resulted from the different expression profile in GFP-expressing and silenced regions, rather than that in leaf sectors (i.e. outer versus inner blades of leaves). This suggests that the two regions have fundamentally different transcriptional profiles, despite being adjacent and phenotypically identical, apart from their GFP-expressing status. To assess whether this difference involved gene silencing pathways, as could be expected, we compared transcriptome profiles of various SPIGs between the two regions. The 18 representative SPIGs were compared based on a previous classification into four groups carried out in Arabidopsis; namely, common, PTGS-, miRNA-, and TGS-specific proteins (Frizzi and Huang, 2010) (Table 2). The reliability of our transcriptome analysis of the 18 SPIGs was confirmed by qRT-PCR with six key genes (GFP, RDR6, AGO4, HYL1, MET1, and CMT3). We found that qRT-PCR resulted in the same expression pattern as the fold enrichments observed in transcriptome analysis (Supplementary Data and Supplementary Data). Although a corrected P-value of MET1 (0.140), its expressional difference was identical to the fold enrichment derived from transcriptome analysis. These data show that there are no striking differences in gene expression profiles between the two regions when analysing genes involved in RNA-silencing pathways. We demonstrate that most of the components commonly related to RNA degradation are already transcribed in plant cells of both regions; however, the genes associated with silencing initiation are not activated in the GFP-expressing cells at the time of our analysis. These findings also suggest that plant cells are always prepared for gene silencing and can degrade target RNA once dsRNAs or aRNAs are produced. Taken together with overall transcriptome profiles of region-specific genes (Supplementary Data, Supplementary Data, and Supplementary Dataand Fig. 6), it seems the initiation or arrest of silencing might be dependent on region-specific genes, but not SPIGs. Notably, the finding of many more specific genes being expressed in the GFP-expressing regions suggests that stable transgene expression might be a more complicated and aggressive process than transgene silencing. Table 2. The expressional fold changes of putatively silencing pathway-involved genes, such as PTGS, miRNA, and TGS, between GFP-expressing and silenced regions (A) The gene expression levels were verified through quantitative RT-PCR using specific primers (Supplementary Data, Supplementary Data). (B) The SUVH6 contig was only assembled. The other SUVHs were not able to be identified owing to the lack of information at sol genomics. (C) The range of base mean read counts mapped to the genes were indicated by <100, <1000, <10 000 and >10 000. (D) The expressional fold change in the silenced to expressing region was calculated by read counts. The significance of difference (P-value, Benjamini and Hochberg-corrected) was analysed using DESeq. Function (putative)  Gene  Type  Description  Contig No.  Length (bp)  Read countsC  Fold changeD  P-value  Photo-synthesis related  RbcS  Ribulose bisphosphate carboxylase, small chain  783  698  >10 000  1.4  0.399  CAB1  Light harvesting chlorophyll a/b-binding protein  284  608  <10 000  0.9  0.915  FBPase  Fructose-1,6- bisphosphatase precursor  5290  1594  <10 000  0.7  0.890  Transgene  GFPA  Green fluorescent protein  13  1001  >10 000  0.01  <0.001  Common  RDR  RDR2  RNA-dependent RNA polymerase RdRP2  37681  3561  <1000  1.0  0.999  RDR6A  RNA-dependent RNA polymerase SDE1  10447  4215  <1000  1.8  0.510  DCL  DCL1  Endoribonuclease, Dicer- like 1 protein  3745  5839  <1000  1.8  0.596  DCL3  Endoribonuclease, Dicer- like 3 protein  8494  4410  <1000  1.2  0.927  DCL4  Endoribonuclease, Dicer- like 4 protein  7266  1832  <100  0.4  0.775  AGO  AGO1  Argonaute 1–2  24205  1013  <1000  1.3  0.750  AGO4A  Argonaute 4-2  10368  3204  <10 000  1.4  0.700  HEN1  RNA methylase (HUA ENHANCER 1)  43652  519  <100  0.9  0.950  PTGS  SGS3  Suppressor of gene silencing 3  5111  1513  <1000  2.1  0.330  SDE3  RNA helicase SDE3  6867  2237  <10 000  1.2  0.860  miRNA  HYL1A  dsRNA/miRNA binding domain (DsRBD)  38405  2220  <1000  1.2  0.827  TGS  DNA methylase  MET1A  DNA (cytosine-5) methyltransferase  41253  5109  <1000  3.6  0.140  CMT3A  Chromomethylase 3b  22233  3170  <1000  3.8  0.008  DRM2  DNA methyltransferase  15472  1287  <1000  1.2  0.870  RNA pol IV  DNA-dependent RNA polymerase IV  4995  1912  <1000  1.4  0.760  DRD1  Defective in RNA-directed DNA methylation 1(CHR38)  20768  3937  <1000  1.8  0.544  HDA2  Histone deacetylase 2a/b  11571  1074  <1000  1.0  1.00  SUVH6B  H3-K9 methyltransferase 6  19464  1835  <1000  1.3  0.750  Function (putative)  Gene  Type  Description  Contig No.  Length (bp)  Read countsC  Fold changeD  P-value  Photo-synthesis related  RbcS  Ribulose bisphosphate carboxylase, small chain  783  698  >10 000  1.4  0.399  CAB1  Light harvesting chlorophyll a/b-binding protein  284  608  <10 000  0.9  0.915  FBPase  Fructose-1,6- bisphosphatase precursor  5290  1594  <10 000  0.7  0.890  Transgene  GFPA  Green fluorescent protein  13  1001  >10 000  0.01  <0.001  Common  RDR  RDR2  RNA-dependent RNA polymerase RdRP2  37681  3561  <1000  1.0  0.999  RDR6A  RNA-dependent RNA polymerase SDE1  10447  4215  <1000  1.8  0.510  DCL  DCL1  Endoribonuclease, Dicer- like 1 protein  3745  5839  <1000  1.8  0.596  DCL3  Endoribonuclease, Dicer- like 3 protein  8494  4410  <1000  1.2  0.927  DCL4  Endoribonuclease, Dicer- like 4 protein  7266  1832  <100  0.4  0.775  AGO  AGO1  Argonaute 1–2  24205  1013  <1000  1.3  0.750  AGO4A  Argonaute 4-2  10368  3204  <10 000  1.4  0.700  HEN1  RNA methylase (HUA ENHANCER 1)  43652  519  <100  0.9  0.950  PTGS  SGS3  Suppressor of gene silencing 3  5111  1513  <1000  2.1  0.330  SDE3  RNA helicase SDE3  6867  2237  <10 000  1.2  0.860  miRNA  HYL1A  dsRNA/miRNA binding domain (DsRBD)  38405  2220  <1000  1.2  0.827  TGS  DNA methylase  MET1A  DNA (cytosine-5) methyltransferase  41253  5109  <1000  3.6  0.140  CMT3A  Chromomethylase 3b  22233  3170  <1000  3.8  0.008  DRM2  DNA methyltransferase  15472  1287  <1000  1.2  0.870  RNA pol IV  DNA-dependent RNA polymerase IV  4995  1912  <1000  1.4  0.760  DRD1  Defective in RNA-directed DNA methylation 1(CHR38)  20768  3937  <1000  1.8  0.544  HDA2  Histone deacetylase 2a/b  11571  1074  <1000  1.0  1.00  SUVH6B  H3-K9 methyltransferase 6  19464  1835  <1000  1.3  0.750  View Large Fig. 6. View large Download slide Expressional differences of randomly selected 11 regional-specific genes of Supplementary Data including wild-type (WT) plant through qRT-PCR. The GFP-expressing and silenced regions respectively corresponds to the outer and inner-blade of leaves in WT plant. The expression levels were represented by means ± SD and analysed by independent two sample t-test for differential expression. (A) The relative expression levels of silenced region-specific (DnaJ, H4 and RCOM) and expressing region-specific (HYP, ERP, BBD2, ARF14, RL1, UBE2 and SPS1) genes in partitioned plant. (B) The relative expression of the specific genes in the outer and inner blade of leaf in WT plant. Fig. 6. View large Download slide Expressional differences of randomly selected 11 regional-specific genes of Supplementary Data including wild-type (WT) plant through qRT-PCR. The GFP-expressing and silenced regions respectively corresponds to the outer and inner-blade of leaves in WT plant. The expression levels were represented by means ± SD and analysed by independent two sample t-test for differential expression. (A) The relative expression levels of silenced region-specific (DnaJ, H4 and RCOM) and expressing region-specific (HYP, ERP, BBD2, ARF14, RL1, UBE2 and SPS1) genes in partitioned plant. (B) The relative expression of the specific genes in the outer and inner blade of leaf in WT plant. Downward systemic transmission of silencing Our finding that the silencing pattern of partitioned regions is not altered during tobacco development (Figs 1 and 2) is in contrast to the general concept of systemic silencing. To provide crucial insights into the mechanism of partitioned silencing, the existence of a competent systemic silencing system was analysed using a grafting experiment. The steady-green plant (266-3-1) was used as a recipient plant, either as a root-stock or scion. One dozen partitioned and steady-green plants, grown for eight weeks post sowing, were grafted. At 42 days post-grafting (dpg), root-ward transmission of silencing was clearly observed in the form of silencing strips (SS) along the main stem of the steady-green root-stocks from the fully silenced scions of the partitioned plants, and the SS reached to almost the bottom of the root-stocks (Supplementary Data). In the reverse direction, the shoot-ward transmission to the steady-green scions was hardly detected, even at 42 (Supplementary Data), and 83 dpg (Supplementary Data). These data indicate that the mobile silencing signals, presumably 21 and 24 nt siRNAs, exist in the silenced region, but could move only in the root-ward direction, probably through the phloem. However, the transmission rate and intensity were too slow and weak to co-silence the existing leaves and the newly emerging lateral shoots adjacent to the SS (Supplementary Data). This one-directional movement is contrary to some previous reports (Bai et al., 2011; Kasai et al., 2011; Palauqui et al., 1997; Voinnet et al., 1998). To confirm the direction of transmission at the active growing stage, younger plants grown for six weeks were used. Similar to the results as before, the direction of transmission was only root-ward (Fig. 7A, B). However, the transmission rate of silencing was higher and faster, leading to the co-silencing of existing leaves of the root-stock. This systemic silencing pattern on the leaves of root stocks at 33, 63, and 70 dpg showed an outward radial wave-like movement from the mid-veins to the lateral veins and finally to the neighbouring mesophyll cells (Fig. 7A, B and red circles 1 and 2 in Fig. 7D–F). This radial wave of silencing suggested that mobile signals initially reached the mid and lateral veins through the phloem flow, and then spread to neighbouring cells through cell-to-cell movement, with the signals amplified by a self-perpetuating silencing loop. The grafting experiment indicated that partitioned-silenced regions definitely possess mobile signals for systemic transmission of silencing. However, through some unknown mechanism, transmission within partitioned plants is repressed, so that the silencing pattern appears to be cell-autonomous or determined. Fig. 7. View largeDownload slide The evidence for the progressed systemic silencing in the stock grafted at the active growing stage. The steady-green (G) and partitioned-silenced (PS) plants grown for six weeks post-sowing were grafted to examine the strength of systemic transmission. (A, B) The progressed transmission of root-ward systemic silencing to the leaves of the stock (PS/G-1). The grafted plant of PS-scion to G-stock was photographed in the front (A) and on the top (B) at 33 dpg. (C) Non-detection of shoot-ward transmission in G/PS-2 grafted with G-scions to PS-stocks at 33 dpg. (D–F) The time-course transmission of systemic silencing on a leaf of PS/G-1. Circles 1 and 2 showed the gradual enlargement of the silencing regions with the progress of time from 33–70 dpg. Circle 3 showed the gradual re-expression of GFP as the silencing moved outward of the leaf blade. Fig. 7. View largeDownload slide The evidence for the progressed systemic silencing in the stock grafted at the active growing stage. The steady-green (G) and partitioned-silenced (PS) plants grown for six weeks post-sowing were grafted to examine the strength of systemic transmission. (A, B) The progressed transmission of root-ward systemic silencing to the leaves of the stock (PS/G-1). The grafted plant of PS-scion to G-stock was photographed in the front (A) and on the top (B) at 33 dpg. (C) Non-detection of shoot-ward transmission in G/PS-2 grafted with G-scions to PS-stocks at 33 dpg. (D–F) The time-course transmission of systemic silencing on a leaf of PS/G-1. Circles 1 and 2 showed the gradual enlargement of the silencing regions with the progress of time from 33–70 dpg. Circle 3 showed the gradual re-expression of GFP as the silencing moved outward of the leaf blade. Discussion A de novo model for silencing research The partitioned phenotype provides a useful model in which to investigate the fine-tuned regulation between stable gene expression and specific silencing. Recently, a similar phenotype, called a “bi-zonal pattern”, was discovered from a seedling graft system in Arabidopsis (Liang et al., 2012), whereby GFP-expressing apical meristems grafted onto GFP-silenced stocks developed into bi-zonally silenced leaves (GFP-silenced only on the lower parts of the leaf). This phenotype suggested that small RNAs (sRNAs) were transported from cell to cell via plasmodesmata, rather than the phloem, for the upward transmission in apical meristem, owing to incompleteness of the vascular system, and that the bi-zonal pattern was caused by a lagging of cell-to-cell movement of sRNAs into the whole GFP-expressing apical meristem (Liang et al., 2012). Similar to the bi-zonal patterning, we initially hypothesized that the cell-autonomous-like nature of the partitioned silencing occurred owing to movement of mobile signal molecules only to the inner blade, limited by a lack of vein networks at the stage of leaf primordial-like tissue. However, GFP expression in limbs, sepals, and leaf-tips, continued development of lateral partitioned leaves, and region-specific transcriptome data, indicate that a limited or lagged movement of mobile signals does not explain our data. Regulating factors responsible for partitioned silencing siRNAs ranging from 20–24 nt are generated by the coordinated actions of different DCLs and AGOs (Molnar et al., 2011), and a variety of RNA-silencing pathways have been proposed according to the involvement of different DCL family members. For this study, a simplified pathway model in terms of PTGS, TGS, and miRNA (Frizzi and Huang, 2010) was applied to elucidate the underlying process of partitioned silencing. Through NGS transcriptome analysis, most of the SPIGs were demonstrated to be intrinsically expressed and ready for RNA silencing, regardless of their GFP-transgene expression status. This “ready-to-start” concept has been indirectly demonstrated previously, whereby specific target genes become silenced simply by the introduction of anti-sense RNAs or short dsRNAs using RNAi technology (Frizzi and Huang, 2010; Brodersen and Voinnet, 2006). These RNAs are substrates for siRNA production for the PTGS pathway. In partitioned-silenced regions, the amplified siRNAs are suggestive of a potential boost of target RNA degradation. We also observed up-regulation of the TGS-related genes (e.g. MET1 and CMT3). The up-regulated TGS-related genes might contribute to the production of more aRNAs to strengthen the PTGS pathway. As demonstrated in vivo by agroinfiltration and inoculation assays, the TGS pathway seems to play a subsidiary role for maintaining the active state of PTGS. Furthermore, the continued random occurrence of the partitioned phenotype from hemizygous progenitors combined with our DNA methylation data shows that TGS is not a major silencing mechanism here. Thus, we conclude that the partitioned silencing is predominately maintained by RdDM pathway. In the partitioned plants, spontaneous dsRNA production seems unlikely owing to the DR structure of GFP transgenes, with the intervention of the vector backbone (~3.0kb) between them. Therefore, the formation of aRNAs is probably responsible for silencing initiation. siRNAs guide the up-regulation of DNA/histone modification genes and a PTGS-amplifying gene, resulting in the strong establishment of an RNA silencing self-perpetuating loop. In our transcriptome data, region-specific genes may play critical roles in the initiation or arrest of gene silencing. Particularly, further study into GFP-expressing region-specific genes (Supplementary Data) may give new insights into methods of promoting the stable expression of transgenes. Therefore, region-specific genes will be an important factor of understanding determined-like silencing, and also help to develop a target-oriented gene expression or silencing techniques for the plant biotechnology field. Existence of silencing signals in partitioned silencing Over a decade ago, various studies reported systemic-acquired silencing, such as the transmission of silencing to upper leaves (Voinnet and Baulcombe, 1997) and the unidirectional transmission of silencing (Palauqui et al., 1997). In contrast, signals were also confirmed to be transmitted from silenced scions to non-silenced stocks by the direction of phloem flow (Tournier et al., 2006). siRNAs, such as 21 nt miR399 (Pant et al., 2008) and 24 nt sRNAs (Molnar et al., 2010), were identified as mobile signals for systemic spreading, and the movement from shoot to stock was confirmed to be more efficient than that in the opposite direction, which was consistent with previous findings of the source-to-sink movement through the phloem flow. Recently, long-range root-to-shoot silencing was shown to be spread largely by a series of cell-to-cell short-range mobile silencing events, rather than by transport through the phloem (Liang et al., 2012). This means that cell-to-cell movement is the major route to communicate with shoot-ward cells. Taken together, these observations show that silencing can move bi-directionally, and that siRNAs are key molecules, responsible for conveying the signal to other parts of the plant. However, the silencing spreading inside our partitioned plants has not been shown directly and even looks to be cell-autonomous-like silencing. The grafting experiments confirmed that partitioned plants also had the potential for systemic transmission, though only in a downwards direction. Silencing was not spread shoot-ward across the borderline of grafted junctions until 83 dpg. Fundamental questions arise as to why the silencing spread inside partitioned plants is repressed, what factors force the maintenance of an initial pattern of silencing, and whether the shoot-ward transmission is real or artefactual. Under high-temperature conditions, the partitioned phenotype was no longer maintained, leading to the early initiation of silencing, a decrease in number of partitioned leaves, increases in the size of silenced regions, and a shift of leaf phenotype from partitioned to variegated. Such differences demonstrate that the silencing pattern is affected by temperature, whereby silencing is exacerbated by high temperatures. This exacerbation can be explained by an increased amount of siRNAs (~21–26 nt) resulting from high temperatures (Szittya et al., 2003), possibly through enhanced RDR6 transcription (Qu et al., 2005). The characteristics of partitioned silencing are therefore highly reminiscent of a cell-autonomous, rather than a non-cell-autonomous mechanism. As such, the partitioned-silenced phenotype will provide an important opportunity to unveil the underlying mechanisms utilised by silencing initiation factors controlling cell-autonomous-like and determined silencing. Abbreviations: Abbreviations: 35S CaMV 35S promoter AGO argonaute ARF14 AP2/ERF and B3 domain-containing transcription factor RAV1-like BBD2 bifunctional nuclease 2-like CMT chromomethylase DnaJ chaperone DnaJ ERP elicitor responsible GFP green fluorescent protein H4 histone H4-like HDA histone deacetylase HYL1 dsRNA/miRNA binding domain 1 (DsRBD) HYP hypothetical protein MET DNA (cytosine-5) methyltransferase miRNA microRNA NGS next-generation sequencing nos nopaline synthase RCOM CASP-like RDR RNA-dependent RNA polymerase SDE RNA helicase SPS1 solanesyl diphosphate synthase 1-like TMV tobacco mosaic virus UBE2 ubiquitin-conjugating enzyme E2-17 kDa-like. Acknowledgements The authors gratefully acknowledge the efforts of Min Sue Choi and Yeon-Jung Cho for selecting and analysing the partitioned plants and Ana L. Toribio for cooperation in the contig registration into EMBL. This study was carried out cooperatively with the John Innes Centre (JIC), and basically funded by the National Academy of Agricultural Science (NAAS, PJ008673) of Rural Development Administration (RDA), the Republic of Korea. The work at JIC was supported by the United Kingdom Biotechnological and Biological Sciences Research Council (BBSRC) Institute Strategic Programme Grant “Understanding and Exploiting Plant and Microbial Metabolism” BB/J004561/1 and by the John Innes Foundation. References Allen E Xie Z Gustafson AM Carrington JC . 2005. microRNA-directed phasing during trans-acting siRNA biogenesis in plants. Cell  121, 207– 221. Google Scholar CrossRef Search ADS PubMed  Bai S Kasai A Yamada K Li T Harada T . 2011. A mobile signal transported over a long distance induces systemic transcriptional gene silencing in a grafted partner. Journal of Experimental Botany  62, 4561– 70. Google Scholar CrossRef Search ADS PubMed  Brigneti G Voinnet O Li WX Ji LH Ding SW Baulcombe DC . 1998. Viral pathogenicity determinants are suppressors of transgene silencing in Nicotiana benthamiana. EMBO Journal  17, 6739– 6746. Google Scholar CrossRef Search ADS PubMed  Broderson P Voinnet O . 2006. The diversity of RNA silencing pathways in plants. Trends in Genetics  22, 268– 280. Google Scholar CrossRef Search ADS PubMed  Chellappan P Vanitharani R Ogbe F Fauquet CM . 2005. Effect of temperature on Geminivirus-induced RNA silencing in plants. Plant Physiology  138, 1828– 1841. Google Scholar CrossRef Search ADS PubMed  Cheung P Lau P . 2005. Epigenetic regulation by histone methylation and histone variants. Molecular Endocrinology  19, 563– 573. Google Scholar CrossRef Search ADS PubMed  Cheng Y Kato N Wang W Li J Chen X . 2003. Two RNA binding proteins, HEN4 and HUA1, act in the processing of AGAMOUS pre-mRNA in Arabidopsis thaliana. Developmental Cell  4, 53– 66. Google Scholar CrossRef Search ADS PubMed  Choi SM Yoon JY Chung BN . 2009. Genome analysis and characterization of a tobacco mosaic virus isolate infecting balsam (Impatiens balsamina). Archives of Virology  154, 881– 885. Google Scholar CrossRef Search ADS PubMed  Choi MS Yoon I-S Rhee Y et al.   2008. The effect of cucumber mosaic virus 2b protein to transient expression and transgene silencing mediated by Agro-infiltration. The Plant Pathology Journal  24, 296– 304. Google Scholar CrossRef Search ADS   Conesa A Gotz S Garcia-Gomez JM Terol J Talon M Robles M . 2005. Blast2GO: a universal tool for annotation, visualization and analysis in functional genomics research. Bioinformatics  21, 3674– 3676. Google Scholar CrossRef Search ADS PubMed  Deleris A Gallego-Bartolome J Bao J Kasschau KD Carrington JC Voinnet O . 2006. Hierarchical action and inhibition of plant dicer-like proteins in antiviral defense. Science  313, 68– 71. Google Scholar CrossRef Search ADS PubMed  Dunoyer P Himber C Voinnet O . 2005. DICER-LIKE 4 is required for RNA interference and produces the 21-nucleotide small interfering RNA component of the plant cell-to-cell silencing signal. Nature Genetics  37, 1356– 1360. Google Scholar CrossRef Search ADS PubMed  Eamens A Wang M-B Smith NA Waterhouse PM . 2008. RNA silencing in plants: yesterday, today, and tomorrow. Plant Physiology  147, 456– 468. Google Scholar CrossRef Search ADS PubMed  Frizzi A Huang S . 2010. Tapping RNA silencing pathways for plant Biotechnology. Plant Biotechnology Journal  8, 655– 677. Google Scholar CrossRef Search ADS PubMed  Ghildiyal M Zamore PD . 2009. Small silencing RNAs: an expanding universe. Nature Reviews Genetics  10, 94– 108. Google Scholar CrossRef Search ADS PubMed  Goto K Kobori T Kosaka Y Natsuaki T Masuta C . 2007. Characterization of silencing suppressor 2b of cucumber mosaic virus based on examination of its small RNA-binding abilities. Plant and Cell Physiology  48, 1050– 1060. Google Scholar CrossRef Search ADS PubMed  He G Elling AA Deng XW . 2011. The epigenome and plant development. Annual Review of Plant Biology  62, 411– 435. Google Scholar CrossRef Search ADS PubMed  Herr AJ Molnár A Jones A Baulcombe DC . 2006. Defective RNA processing enhances RNA silencing and influences flowering of Arabidopsis. Proceedings of the National Academy of Sciences, USA  103, 14994– 15001. Google Scholar CrossRef Search ADS   Himber C Dunoyer P Moissiard G Ritzenthaler C Voinnet O . 2003. Transitivity-dependent and -independent cell-to-cell movement of RNA silencing. EMBO Journal  22, 4523– 4533. Google Scholar CrossRef Search ADS PubMed  Hutvagner G McLachlan J Pasquinelli AE Balint E Tuschl T Zamore PD . 2001. A cellular function for the RNA-interference enzyme Dicer in the maturation of the let-7 small temporal RNA. Science  293, 834– 838. Google Scholar CrossRef Search ADS PubMed  Jones L Hamilton AJ Voinnet O Thomas CL Maule AJ Baulcombe DC . 1999. RNA–DNA interactions and DNA methylation in post-transcriptional gene silencing. Plant Cell  11, 2291– 2301. Google Scholar CrossRef Search ADS PubMed  Kalantidis K Schumacher HT Alexiadis T Hem JM . 2008. RNA silencing movement in plants. Biology of the Cell  100, 13– 26. Google Scholar CrossRef Search ADS PubMed  Kasai A Bai S Li T Harada T . 2011. Graft-transmitted siRNA signal from the root induces visual manifestation of endogenous post-transcriptional gene silencing in the scion. PLoS One  6, e16895. Google Scholar CrossRef Search ADS PubMed  Kim S-R Lee J Jun S-H Park S Kang H-G Kwon S An G . 2003. Transgene structures in T-DNA-inserted rice plants. Plant Molecular Biology  52, 761– 773. Google Scholar CrossRef Search ADS PubMed  Kumar S Fladung M . 2002. Transgene integration in aspen: structures of integration sites and mechanism of T-DNA integration. The Plant Journal  31, 543– 551. Google Scholar CrossRef Search ADS PubMed  Lee H-C Aalto AP Yang Q Chang S-S Huang G Fisher D Cha J Poranen MM Bamford DH Liu Y . 2010. The DNA/RNA-dependent RNA polymerase QDE-1 generates aberrant RNA and dsRNA for RNAi in a process requiring replication protein A and a DNA helicase. PLOS Biology  8, e1000496. Google Scholar CrossRef Search ADS PubMed  Li HW Lucy AP Guo H-S Li WX Ji LH Wong SM Ding SW . 1999. Strong host resistance targeted against a viral suppressor of the plant gene silencing defense mechanism. EMBO Journal  18, 2683– 2691. Google Scholar CrossRef Search ADS PubMed  Liang D White RG Waterhouse PM . 2012. Gene silencing in Arabidopsis spreads from the root to the shoot, through a gating barrier, by template-dependent, non-vascular, cell-to-cell movement. Plant Physiology  159, 984– 1000. Google Scholar CrossRef Search ADS PubMed  Lim S-H Ha S-H . 2013. Marker development for the identification of rice seed color. Plant Biotechnology Reports  7, 391– 398. Google Scholar CrossRef Search ADS   Luo Z Chen Z . 2007. Improperly terminated, unpolyadenylated mRNA of sense transgenes is targeted by RDR6-mediated RNA silencing in Arabidopsis. Plant Cell  19, 943– 958. Google Scholar CrossRef Search ADS PubMed  Melnyk CW Molnar A Baulcombe DC . 2011. Intercellular and systemic movement of RNA silencing signals. EMBO Journal  30, 3553– 3563. Google Scholar CrossRef Search ADS PubMed  Miki D Shimamoto K . 2008. De novo DNA methylation induced by siRNA targeted to endogenous transcribed sequences is gene-specific and OsMet1-independent in rice. The Plant Journal  56, 539– 49. Google Scholar CrossRef Search ADS PubMed  Molnar A Melnyk CW Bassett A Hardcastle TJ Dunn R Baulcombe DC . 2010. Small silencing RNAs in plants are mobile and direct epigenetic modification in recipient cells. Science  328, 872– 875. Google Scholar CrossRef Search ADS PubMed  Molnar A Melnyk C Baulcombe DC . 2011. Silencing signals in plants: a long journey for small RNAs. Genome Biology  12, 215. Google Scholar CrossRef Search ADS PubMed  Okano Y Miki D Shimamoto K . 2008. Small interfering RNA (siRNA) targeting of endogenous promoters induces DNA methylation but not necessarily gene silencing in rice. The Plant Journal  53, 65– 77. Google Scholar CrossRef Search ADS PubMed  Palauqui J-C Elmayan T Pollien J-M Vaucheret H . 1997. Systemic acquired silencing: transgene-specific post-transcriptional silencing is transmitted by grafting from silenced stocks to non-silenced scions. EMBO Journal  16, 4738– 4745. Google Scholar CrossRef Search ADS PubMed  Pant BD Buhtz A Kehr J Scheible WR . 2008. MicroRNA399 is a long-distance signal for the regulation of plant phosphate homeostasis. The Plant Journal  53, 731– 738. Google Scholar CrossRef Search ADS PubMed  Park W Li J Song R Messing J Chen X . 2002. CARPEL FACTORY, a Dicer homolog, and HEN1, a novel protein, act in microRNA metabolism in Arabidopsis thaliana. Current Biology  12, 1484– 1495. Google Scholar CrossRef Search ADS PubMed  Pontes O Li CF Nunes PC Haag JR Ream T Vitins A Jacobsen SE Pikaard CS . 2006. The Arabidopsis chromatin-modifying nuclear siRNA pathway involves a nucleolar RNA processing center. Cell  126, 79– 92. Google Scholar CrossRef Search ADS PubMed  Qu F Ye X Hou G Sato S Clemente TE Morris TJ . 2005. RDR6 has a broad-spectrum but temperature-dependent antiviral defense role in Nicotiana benthamiana. Journal of Virology  79, 15209– 15217. Google Scholar CrossRef Search ADS PubMed  Rabindran S Dawson WO . 2001. Assessment of recombinants that arise from the use of a TMV-based transient expression vector. Virology  284, 182– 189. Google Scholar CrossRef Search ADS PubMed  Sohn S-H Huh SM Kim K-H Park JW Lomonossoff G . 2011a. Effect of rice stripe virus NS3 on transient gene expression and transgene co-silencing. The Plant Pathology Journal  27, 310– 314. Google Scholar CrossRef Search ADS   Sohn S-H Choi MS Kim K-H Lomonossoff G . 2011b. The epigenetic phenotypes in transgenic Nicotiana benthamiana for CaMV 35S–GFP are mediated by spontaneous transgene silencing. Plant Biotechnology Reports  5, 273– 281. Google Scholar CrossRef Search ADS   Szittya G Silhavy D Molnár A Havelda Z Lovas A Lakatos A Bánfalvi Z Burgyán J . 2003. Low temperature inhibits RNA silencing-mediated defense by the control of siRNA generation. EMBO Journal  22, 633– 640. Google Scholar CrossRef Search ADS PubMed  Tournier B Tabler M Kalantidis K . 2006. Phloem flow strongly influences the systemic spread of silencing in GFP Nicotiana benthamiana plants. The Plant Journal  47, 383– 394. Google Scholar CrossRef Search ADS PubMed  Vaistij FE Jones L Baulcombe DC . 2002. Spreading of RNA targeting and DNA methylation in RNA silencing requires transcription of the target gene and a putative RNA-dependent RNA polymerase. Plant Cell  14, 857– 867. Google Scholar CrossRef Search ADS PubMed  Vaucheret H . 2005. RNA polymerase IV and transcriptional silencing. Nature Genetics  37, 659– 60. Google Scholar CrossRef Search ADS PubMed  Voinnet O Baulcombe DC . 1997. Systemic signalling in gene silencing. Nature  389, 553. Google Scholar CrossRef Search ADS PubMed  Voinnet O Vain P Angell S Baulcombe DC . 1998. Systemic spread of sequence-specific transgene RNA degradation is initiated by localised introduction of ectopic promoterless DNA. Cell  95, 177– 187. Google Scholar CrossRef Search ADS PubMed  Wassenegger M Krczal G . 2006. Nomenclature and functions of RNA-directed RNA polymerases. Trends in Plant Science  11, 142– 151. Google Scholar CrossRef Search ADS PubMed  Xie Z Johansen LK Gustafson AM Kasschau KD Lellis AD Zilberman D Jacobsen SE Carrington JC . 2004. Genetic and functional diversification of small RNA pathways in plants. PLOS Biology  2, 642– 652. Google Scholar CrossRef Search ADS   Yoshikawa M Peragine A Park MY Poethig RS . 2005. A pathway for the biogenesis of trans-acting siRNAs in Arabidopsis. Genes and Development  19, 2164– 2175. Google Scholar CrossRef Search ADS PubMed  © The Author 2014. Published by Oxford University Press on behalf of the Society for Experimental Biology. All rights reserved. For permissions, please email: journals.permissions@oup.com
TI - Cell-autonomous-like silencing of GFP-partitioned transgenic Nicotiana benthamiana
JF - Journal of Experimental Botany
DO - 10.1093/jxb/eru200
DA - 2014-05-27
UR - https://www.deepdyve.com/lp/oxford-university-press/cell-autonomous-like-silencing-of-gfp-partitioned-transgenic-nicotiana-kxGnAeUv0v
SP - 4271
EP - 4283
VL - 65
IS - 15
DP - DeepDyve
ER -