High transgene expression is associated with systemic GFP silencing in Nicotiana benthamiana

Gene silencing in plants using topical dsRNA is a new approach that has the potential to be a sustainable component of the agricultural production systems of the future. However, more research is needed to enable this technology as an economical and efficacious supplement to current crop protection practices. Systemic gene silencing is one key enabling aspect. The objective of this research was to better understand systemic transgene silencing in Nicotiana benthamiana. Previous reports details sequencing of the integration site of the Green Fluorescent Protein (GFP) transgene in the well-known N. benthamiana GFP16C event revealed inadvertent co-integration of part of a bacterial transposase. To determine the effect of this transgene configuration on systemic silencing, new GFP transgenic lines with or without the transposase sequences were produced. GFP expression levels in the 19 single-copy events and three hemizygous 16C lines produced for this study ranged from 50-72% of the homozygous 16C line. GFP expression was equivalent to 16C in a two-copy event. Local GFP silencing was observed in all transgenic and 16C hemizygous lines after topical application of delivery formulations with a GFP targeting dsRNA. The 16C-like systemic silencing phenotype was only observed in the two-copy line. The partial transposase had no impact on transgene expression level, local GFP silencing, small RNA abundance and distribution, or systemic GFP silencing in the transgenic lines. We conclude that high transgene expression level is a key enabler of systemic transgene silencing in N. benthamiana.


48
RNA-based gene silencing is a sequence-specific, conserved mechanism in eukaryotes implicated in viral 49 defense, control of transposable elements, and gene regulation. Gene silencing using transgenic dsRNA-50 based approaches has been utilized to deploy a number of agriculturally important traits including virus 51 resistance in papaya (1), delayed fruit ripening in tomato (2), black-spot bruise resistance and lower 52 acrylamide levels post-cooking in potato (3), improved oil composition in soybeans (4) and insect control with a partial GFP coding sequence, abundant transitive small RNAs both 5' and 3' of the targeted 106 sequence were observed in N. benthamiana and Arabidopsis (28).

108
Transitivity is observed when targeting transgenes for silencing but reports of transitivity when targeting

116
Another factor contributing to transitivity is the RNAi effector (trigger) that is used. Transitivity 3' of the 117 target locus was observed using a 22nt amiRNA construct targeting chalcone synthase in Arabidopsis but

146
The seedlings were transplanted 9-10 days after seeding into 2.5in pots filled with Berger BM2 peat moss 147 potting mixture. Transplants were grown with the same conditions described above except for irrigation 148 frequency. Transplants were ebb-flow irrigated every other day for the first week and daily thereafter. The T-DNA inserts for each transformation construct were synthesized using a third-party vendor (Bio

DNA and RNA extraction and analysis
167 Leaf tissue was collected using a 4mm round biopsy punch. Eight to ten samples per leaf were collected 168 into 96-well plate preloaded with steel grinding balls. The plates were frozen prior to sampling and tissue 169 was collected on dry ice. Total RNA was extracted from leaf tissue using Trizol reagent (ThermoFisher).

170
1ml of Trizol was added to the frozen leaf discs. The plates were sealed, and the tissue was homogenized 171 at 1600rpm for 10 min using a Genogrinder. The manufacturer's instructions were followed for the 172 remainder of the procedure with exception of a 20 min centrifugation to precipitate total RNA. Glycol 173 blue (45µg) was added to aid in pellet recovery. The RNA was resuspended in 20µl of RNase free water, 174 and the concentration was measured using Quant-iT RNA BR assay kit (ThermoFisher). For qPCR 175 analysis, the samples were diluted to 5ng/µl and target gene expression was measured as described in DNA was extracted using Plant DNAzol (ThermoFisher) following the manufacturer's instructions. The

179
purified DNA was resuspended in water and the concentration was measured using UV spectroscopy. The 180 samples were diluted to 50 ng/µl and GFP copy number was estimated using qPCR. The qPCR reaction 181 mixtures comprised DNA (100ng total), and the reactions were assembled as referenced for the 182 expression analysis. Probes sets for NPTII and GFP coding region were utilized to estimate copy number  Instruments). Images were acquired using the Cannon EOS utility 2 software with tethered image 204 acquisition. For GFP images, 58mm Tiffen Green 11 and Yellow 12 filters were utilized to capture GFP 205 and chlorophyll fluorescence from ~480nm to ~600nm.

207
The images were processed using ImageJ with the software provider's guidance. Briefly, the program 208 operator utilized the threshold color panel to highlight a border around each leaf. A border image was 209 overlaid onto the leaf image and the pixel number within the leaf border was quantitated by the software.

210
The quantitated number of pixels represented the total leaf area. A similar thresholding process was used 211 to highlight a border around the visible leaf phenotype and to quantitate pixels within the phenotypic area.

212
GFP and CHL-H silenced areas were calculated by dividing the phenotypic area pixels by the total leaf 213 area pixels.  sequence without the partial transposase sequence (Fig 1A). Transgenic N. benthamiana plants were 240 created with each construct. Ten single-copy events were selected in the R0 generation using qPCR to 241 estimate copy number relative to the single-copy 16C line (15). Copy number was confirmed in the R1 242 generation using transgene segregation and an additional round of qPCR copy number quantification. All 243 events were confirmed as single copy in the R1 generation except event NT_W22241804 (pMON417670) 244 which was an unlinked, two-copy event. background when the leaves were excited with a blue light source (Fig 2A top). CHL-H silencing 271 appeared as yellow sectors on the application leaves (Fig 2B top). Tissue was collected from phenotypic 272 areas to measure gene expression and small RNA abundance. Reduced mRNA levels were observed for 273 both GFP (Fig 2A middle) and CHL-H (Fig 2B middle) when those genes were targeted by a specific 274 dsRNA. Transitive small RNAs were observed both 5' and 3' of the target region for the GFP gene ( Fig   275  2A bottom), but only 3' of the target region for CHL-H gene (Fig 2B bottom). The transitive small RNAs were predominantly 21nt in length but other biologically important size classes (e.g. 22nt and 24nt) were 277 also observed (Fig 2 inset bottom).

292
The GFP transgene was silenced in the 16C line and the 20 transgenic events produced for this study 293 using carbon dot delivery of a chemically synthesized 22nt dsRNA targeting the GFP transgene. Whole 294 plants were photographed 4 days after dsRNA application to qualitatively assess local GFP silencing (Fig   295  3A top). The plants were harvested by removing all the leaves 14 days after dsRNA application. The

296
leaves were arranged in developmental order and photographed under blue light (Fig 3A bottom). These 297 images were analyzed for local GFP silencing on the application leaf (Fig 3B top) and for systemic 298 silencing on younger leaves (Fig 3B bottom). Systemic GFP silencing covering 25% and 12% of the 299 total leaf area was observed in the 16C and NT_W22241804 lines, respectively. Weak systemic GFP 300 silencing was observed in many of the other events, but the silenced area was low, and did not continue to

374
Substantial variation spanning two orders of magnitude was observed for the total number of small RNAs 375 mapped in these samples. In systemic tissue, 5' and 3' transitive small RNAs were observed for the 16C 376 and NT_W22241804 line. Consistent with the visual phenotypic difference (Fig 3), the 16C line had 10-  without the partial transposase sequence (Fig 1 and 2s). We did not observe any enhancing effect on 414 systemic silencing as a result of including the transposase sequence in the transformation constructs (Fig   415  3). We also didn't observe any effect on the level of expression of the GFP transgene (Fig 1).

417
We observed both 5' and 3' transitivity for the GFP transgene after topical application of targeting 22 bp 418 dsRNAs but have only seen significant 3' transitivity for endogenous genes studied ((10), and we 419 considered that 5' transitivity could be unique to the 16C event and perhaps associated with systemic 420 silencing. We were able to replicate the 5' transitivity phenomenon for the GFP transgene in all of the 421 transgenic events produced for this study (Fig 5)

569
Untreated tissue from the first true leaf of each homozygous R2 transgenic events was sampled and the 570 small RNAs were sequenced. The experiment was arranged as a randomized complete block with 4 571 replications per treatment. The replicates for each treatment were pooled prior to small RNA sequencing.

572
The sequencing data are expressed as the sum of small RNA counts 19-25nt in length per 1x10 6 total 573 small RNA reads.