Enhanced Transgene Expression in Sugarcane by Co-Expression of Virus-Encoded RNA Silencing Suppressors

Post-transcriptional gene silencing is commonly observed in polyploid species and often poses a major limitation to plant improvement via biotechnology. Five plant viral suppressors of RNA silencing were evaluated for their ability to counteract gene silencing and enhance the expression of the Enhanced Yellow Fluorescent Protein (EYFP) or the β-glucuronidase (GUS) reporter gene in sugarcane, a major sugar and biomass producing polyploid. Functionality of these suppressors was first verified in Nicotiana benthamiana and onion epidermal cells, and later tested by transient expression in sugarcane young leaf segments and protoplasts. In young leaf segments co-expressing a suppressor, EYFP reached its maximum expression at 48–96 h post-DNA introduction and maintained its peak expression for a longer time compared with that in the absence of a suppressor. Among the five suppressors, Tomato bushy stunt virus-encoded P19 and Barley stripe mosaic virus-encoded γb were the most efficient. Co-expression with P19 and γb enhanced EYFP expression 4.6-fold and 3.6-fold in young leaf segments, and GUS activity 2.3-fold and 2.4-fold in protoplasts compared with those in the absence of a suppressor, respectively. In transgenic sugarcane, co-expression of GUS and P19 suppressor showed the highest accumulation of GUS levels with an average of 2.7-fold more than when GUS was expressed alone, with no detrimental phenotypic effects. The two established transient expression assays, based on young leaf segments and protoplasts, and confirmed by stable transgene expression, offer a rapid versatile system to verify the efficiency of RNA silencing suppressors that proved to be valuable in enhancing and stabilizing transgene expression in sugarcane.


Introduction
RNA silencing is an ancient pathway shared by eukaryotic organisms to regulate gene expression. It particularly operates as an adaptive defense mechanism, which is initiated by the formation of double stranded RNAs (dsRNAs) to destroy aberrant RNAs in the cell [1][2][3][4]. The silencing pathway is very complex in higher eukaryotes, but some of its distinct steps and key components are well characterized. The dsRNA trigger is first cleaved by the RNase III-type DICER-LIKE proteins into small RNA species of 21-26 nucleotide duplexes named short-interfering RNAs (siRNAs) or microRNAs (miRNAs) [5,6], which are denatured and incorporated into the multi-component RNA-induced silencing complex (RISC) with an Argonaute (AGO) protein at its catalytic core [7]. The RISC complex then binds complementary mRNAs guided by single-stranded siRNAs, thereby mediating processes such as translational inhibition, RNA degradation or chromosome modification [8][9][10]. Unlike the miRNAs produced by the miRNA precursors [5,6,10], the siRNAs can also be amplified from the target RNA by cellular host RNA-dependent RNA polymerases (RdRPs) to produce additional dsRNAs that will be processed into secondary siRNAs [11]. The Suppressor of Gene Silencing 3, a dsRNA binding protein, is also required for post-transcriptional gene silencing (PTGS) in plants [12,13].
Plant viruses have evolved several counter-defensive strategies to efficiently suppress their host RNA silencing mechanism. The production of virus-encoded suppressors of RNA silencing is one of the strategies used to counteract host antiviral defense [14][15][16]. So far, several suppressors of RNA silencing have been identified from different types of viruses, and they show a high diversity in primary sequence and protein structure, though sharing certain mechanistic features [16][17][18]. Viral suppressors seem to interfere with the RNA silencing pathway at distinct steps, since they potentially have different molecular targets in the host and operate differently in widely used silencing inhibition assays [19,20]. For instance, the Tomato bushy stunt virus (TBSV)-encoded P19 [21,22], one of the most studied suppressors, sequesters 21-nt siRNAs in a non-specific manner, preventing their incorporation into the RISC complex to act as guides; it also inhibits the spread of the ds siRNA duplex identified as the signal of RNA silencing [18,23] as well as the translational efficiency of AGO1 mRNA by modulating the endogenous miR168 level [24]. The suppression activity of the Barley stripe mosaic virus (BSMV)-encoded cb was demonstrated in Agrobacterium-mediated transient assays [25,26], and the molecular basis of its silencing suppression is similar to that of P19 [20].
The P1/HC-Pro or HC-Pro from Potato virus Y [27] or Tobacco etch virus (TEV) [28] was the first identified suppressor serving as a model to study the mechanism of silencing suppression. HC-Pro is proposed to act on the RISC complex [20,29] or downstream of an RdRP by interfering with the DICER protein [30,31], or by sequestrating the 21-nt siRNA duplexes [19,20] and inhibiting the 39 modification of the si/miRNAs [29,32].
Transgenic and transient expression via Agrobacterium coinfiltration into Nicotiana benthamiana and Arabidopsis thaliana have been extensively used to probe the phenomenon of RNA silencing and the function of viral suppressors at the whole plant level [17]. Protoplasts of N. benthamiana and A. thaliana proved useful to investigate transient gene expression [33], RNAi-mediated silencing of gene expression [34,35] and the RNA silencing suppressor function [36][37][38] at the cellular level. Viral suppressors of silencing were quantitatively evaluated by transient co-expression with the Green Fluorescent Protein (GFP) in germinating lima bean (Phaseolus lunatus L.) cotyledons via particle bombardment [39].
Sugarcane (Saccharum spp. hybrid) is an economically important sugar and bioenergy producing polyploid crop, which is amenable for improvement through genetic engineering [40][41][42]. Transgene silencing is currently one of the major limiting factors to produce improved transgenic varieties, and to achieve commercially useful expression levels of transgenes in this crop [43][44][45]. In the present study, the strategy of using viral RNA silencing suppressors to counteract RNA silencing was adopted in sugarcane in an attempt to enhance transgene expression and stability. Four RNA silencing suppressors were evaluated for their silencing suppression efficiency by their transient and stable transgenic co-expression with the Enhanced Yellow Fluorescent Protein (EYFP) or the bglucuronidase (GUS) reporter gene. These include the TEVencoded P1/HC-Pro, the BSMV-encoded cb, the TBSV-encoded P19, and one putative suppressor, the Sugarcane bacilliform virus (SCBV)-encoded OrfI. In addition, a P19 suppressor mutant, P19/R43W, whose overexpression did not induce developmental defects, was also evaluated [46]. The suppressor-reporter gene constructs were tested in the dicot Nicotiana benthamiana plants and monocot onion epidermal cells, to determine that they are expressing functional suppressors. An efficient transient suppressor-reporter gene co-expression system, based on young leaf segments and protoplasts of sugarcane, was first established and it was subsequently used to demonstrate that several silencing suppressors enhanced EYFP and GUS expression to a significant level. That the transient expression system provided a rapid analysis of viral RNA silencing suppressor efficiencies was further supported by the generation of stable transgenics. Combined, these results show the usefulness of the system to probe the activity of these suppressors, while these proved to be valuable in enhancing and stabilizing transgene expression in sugarcane.

Assaying the Activity of Viral RNA Silencing Suppressors in Model Plant Systems
To verify that the genetic constructs were expressing functional suppressors (P1/HC-Pro, P19, and cb), we first tested these in N. benthamiana using the standard suppressor activity assay [21] (Legend, Figure S1). A construct expressing GFP was coagroinfiltrated with a suppressor-expressing construct into N. benthamiana leaves, and GFP expression was compared to the treatment with GFP construct alone. These experiments verified that expression of the suppressors led to the expected enhanced and prolonged GFP expression in the dicot N. benthamiana ( Figure  S1).
To determine whether the suppressors were active in a monocot system, onion epidermal cells were co-bombarded with constructs expressing the EYFP gene and those expressing the suppressor genes ( Figure S2a). In this case, the SCVB OrfI suppressor was also included. It has been established that an increase in the number of fluorescent cells correlates with the effectiveness of suppressor activity [47], thus we monitored the effect of cobombardment with a suppressor on the number of EYFPexpressing cells ( Figure S2b). The results of these comparisons indicated that, unlike in N. benthamiana, for unknown reasons not all suppressors performed optimally, even though the suppression effect was often more evident when two suppressors were combined ( Figure S2b). For instance, the SCVB OrfI and cb suppressors exhibited the most prominent effect ( Figure S2b), but whether this is related to their origin of being a monocot-infecting virus remains to be determined. Due to the relative small number of cells that expressed EYFP ( Figure S2a), quantification with western analysis was technically not feasible, so any suppressor effect at the cellular level could not be quantified. Therefore, either the onion cells yielded unexpected results in not responding to certain suppressors in an expected manner to be explored in future experiments, and/or the system itself was insufficiently quantifiable. Thus, we felt that a more robust transient system for monocot expression, preferably sugarcane itself, needed to be established.

Development of Transient Expression Systems for the Rapid Testing of viral RNA Silencing Suppressor Efficiencies in Sugarcane
To evaluate the ability of viral RNA silencing suppressors in enhancing transgene expression in sugarcane, we first established two rapid and efficient transient systems for the co-expression of the suppressor and the reporter gene in the same tissues or cells.

Optimization of transient expression in vivo in young leaf
segments. To investigate the optimal parameters of imaging EYFP-expressing cells in sugarcane young leaf segments following bombardment, we used multicolor fluorescence imaging at low and high optical magnifications combined with ImageJ data analysis. Images of young leaf segment cells were acquired at 48 h after bombardment with pUbi:EYFP:Tnos (EYFP under the control of the maize ubiquitin 1 (Ubi) promoter and the Agrobacterium nopaline synthase terminator (Tnos); Figure S3) (0.5 mg) using filter sets for different fluorophores such as rhodamine, eGFP and YFP with different detection spectra, at 15x and 150x magnifications. Imaging at a high magnification (150x) allowed the distinction of EYFP-expressing cells from autofluorescence derived from damaged cells or any object present on the tissue surface, but the field of vision was restricted to few cells only; however, imaging at a low magnification (15x) provided a significantly larger area of cells for image capture and analysis ( Figures S4 and S5). The overlaid image generated from merging the YFP and bright field images clearly showed that the EYFP-expressing cells were intact ( Figure S5). Furthermore, images taken with the rhodamine filter failed to show any prominent fluorescent spots, indicating that there is no visible autofluorescence from damaged/dead cells in leaf segments expressing pUbi:EYFP:Tnos+pUbi:Tnos or from those bombarded with pUbi:Tnos ( Figure S3) or water (negative controls) at 15x and 150x magnifications ( Figures S4 and S5). Quantitative assessment of EYFP expression in leaf segments of negative controls using ImageJ indicated that the level of autofluorescence from damaged/dead cells or any object present on the tissue surface is minor ( Figure S6). Although the EYFP foci count number of both negative controls (12.1060.30) was about 9.4% of the total foci count number of pUbi:EYFP:Tnos+pUbi:Tnos (129.40615.70), their EYFP expression level (gray value6pixels) (0.01610 5 ) was only 1% of that of pUbi:EYFP:T-nos+pUbi:Tnos (0.95610 5 60.15610 5 ) ( Figure S6). Since the degree of cell damage during bombardment and the chance of having foreign objects on the tissue surface will be the same for each treatment, the level of autofluorescence derived from damaged/dead cells or any foreign object present on the tissue surface can be assumed to be uniform for all treatments. Therefore, we are confident that the sugarcane leaf segmentbased transient gene expression system and the data analysis by ImageJ coupled with fluorescent microscopy at 15x magnification, adopted in the current study, provide a reliable approach for the quantitative analysis of EYFP expression in order to investigate the effect of RNA silencing suppressors in sugarcane.
To optimize the conditions of transient expression of EYFP in sugarcane young leaf segments, we investigated the duration of 0, 3 and 5 days of pre-culture of leaf segments on media prior to bombardment and the dosage of the introduced gene. Data collected at 48 h after bombardment with EYFP (pUbi:EYFP:Tnos; Figure S3) showed that leaf segments bombarded without preculture (0 day) had the lowest EYFP expression (expression level: 3.65610 5 60.26610 5 ; foci count: 228.00617.00). However, leaf segments pre-cultured for 3 days and 5 days displayed significantly (p,0.05) higher EYFP expression levels (8.12610 5 60.62610 5 and 7.41610 5 60.56610 5 , respectively) and foci counts (438.00616.00 and 443.00627.00, respectively) than those without pre-culture at 48 h after bombardment. These results show that pre-culture of sugarcane young leaf segments for 3-5 days prior to bombardment enhanced EYFP expression (as measured by EYFP foci count and expression level).
To determine a suitable amount of DNA to be used for a higher EYFP expression, six EYFP plasmid (pUbi:EYFP:Tnos) concentrations (from 0.125 mg to 4.0 mg DNA per bombardment) were tested in 3 day-pre-cultured sugarcane young leaf segments. Data collected at 48 h after bombardment showed that EYFP expression increased in a linear manner with increasing amounts of EYFP DNA from 0.125 mg (foci count: 237.00618.00; expression level: 1.19610 5 5 60.40610 5 ), until it reached a plateau with 1.0 mg, 2.0 mg and 4.0 mg. The highest EYFP foci count of 369.00 (629.00) and expression level of 6.70610 5 (60.70610 5 ) were obtained with 4 mg of EYFP DNA per bombardment, but this increase was not significantly different from the one obtained with 0.5 mg, 1 mg or 2 mg of EYFP DNA per bombardment. However, we opted to use the 0.25 mg dose in the transient experiments with young leaf segments, to decrease the fluorescence background of EYFP expression (from the EYFP plasmid alone) and avoid interference with image data collection (data not shown).

Establishment of a cellular transient expression system
based on protoplasts. A homogenous cell suspension was obtained from compact globular white-yellow embryogenic sugarcane callus, originating from leaf rolls ( Figures S7a-b), in liquid MS medium with 2,4-dichlorophenoxyacetic acid (2,4-D) (3 mg/L). Subsequently, protoplasts were successfully isolated from this cell suspension after culturing for 2-3 days, with an average yield of about 2610 6 protoplasts per 1 mL of suspension ( Figure S7c). The polyethylene glycol (PEG)-calcium chloride (CaCl 2 ) transfection efficiency of the isolated protoplasts (1610 5 protoplasts; 100 mL) was assessed by using 10 mg of EYFP plasmid DNA (pUbi:EYFP:Tnos) and three transfection (protoplast incubation with DNA) periods of 5 min, 10 min and 15 min. Image analysis data collected at 24 h after co-transfection of the protoplasts showed that around 30% of the transfected protoplasts expressed EYFP, i.e. 30% of protoplasts were successfully transfected ( Figures S7d-f). This transfection rate was maintained to the same level during the 15 min time period (data not shown), indicating that this is a less critical factor to be considered.
Dosage effect of the viral RNA silencing suppressors on transgene expression. Because P19 enhanced transgene expression in N. benthamiana and in our preliminary experiments using sugarcane young leaf segments, we determined the optimal amount of the P19 suppressor to increase EYFP expression and GUS activity in sugarcane young leaf segments and protoplasts, respectively. When EYFP (pUbi:EYFP:Tnos) (0.25 mg) was introduced with no suppressor in 3 day-pre-cultured young leaf segments, EYFP expression reached its peak at 24 h (EYFP foci count) or 48 h (EYFP expression level) post-introduction and then declined (Figures 1a and b). However, co-bombardment of EYFP (0.25 mg) with increasing doses of P19 (pUbi:P19:Tnos; Figure S3) extended the EYFP expression peak to at least 72 h and enhanced EYFP expression (foci count and expression level) (Figures 1a and  b). The highest increases in EYFP expression were obtained with 0.125 mg (335.00614.00; 2.90610 5 60.30610 5 ) and 0.25 mg (335.00616.00; 2.10610 5 60.20610 5 ) of P19. The 0.125 mg and 0.25 mg P19 concentrations resulted in 1.6-fold increase in EYFP foci counts, and 3.2-fold and 2.3-fold increase in EYFP expression levels, respectively, compared to those in the absence of P19 (216.00612.00; 0.90610 5 60.10610 5 ) at the 120 h time point (Figures 1a and 1b). Co-introduction of EYFP (0.25 mg) with higher doses of P19, such as 0.5 mg and 1.0 mg, still enhanced the EYFP foci count (356.00623.00 and 343.00617.00 at 120 h, respectively) by 1.7-fold and 1.6-fold, respectively, but a significant decrease was observed with the 2-mg dose (241.00622.00 at 120 h) ( Figure 1a).

Enhancement of Transient Gene Expression by Coexpression of viral RNA Silencing Suppressors in Sugarcane Young Leaf Segments and Protoplasts
After establishing the two transient expression systems for sugarcane, we tested the five suppressors, P19, P19/R43W, cb, P1/HC-Pro and SCBV OrfI (under the control of the Ubi promoter and nos terminator; Figure S3), in young leaf segments and in protoplasts. In the leaf segment system, when EYFP  Figure S8). However, when EYFP was co-introduced with one of the five suppressors, EYFP expression reached its maximum within 48-96 h post-bombardment and maintained its peak for a longer time (Figures 2 and S8). P19 and cb induced the highest peaks of EYFP expression levels (3.30610 5 60.30610 5 and 2.60610 5 60.20610 5 , respectively) and foci counts (406.00622.00 and 334.00617.00, respectively) at 192 h postbombardment (Figures 2a and 2b). Each of these two suppressors resulted in a highly significant improvement in the transient  Figure S3), and EYFP expression as measured by foci count (a) and expression level (b) was monitored for 120 h postbombardment of sugarcane young leaf segments (3 day-pre-cultured). Vector with no suppressor (pUbi:Tnos; Figure S3) was used as a negative control. Values represent two independent experiments and ten technical repeats, and are reported with the standard error. Quantitation of EYFP foci counts and expression levels is provided in Materials and Methods.

Enhancement of Stable Transgene Expression by Coexpression of the Viral RNA Silencing Suppressor P19 in Sugarcane
In order to study the long term protection of transgene expression by the RNA silencing suppressor P19, a total of 41 transgenic plants, representing seven independent stably transformed lines were generated, from leaf roll disc explants, by cobombardment of the GUS reporter (pUbi:GUS:Tnos) and the P19 (pUbi:P19:Tnos) suppressor genes. Successful gene co-integration was confirmed by Southern blot analysis (data not shown). Lines co-expressing P19 and GUS (P19-GUS) developed normally and showed a significant (p,0.05) enhancement in GUS activity, i.e. an average of 53.2% increase when compared to lines expressing GUS alone (Table 2; Figure 4a). Significant (p,0.05) increases of 1.9-fold (in 39% of transgenics) to 3.5-fold (in 20% of transgenics) in GUS activity were observed in P19-GUS transgenics when compared to those expressing GUS alone (Table 2). No significant increase in GUS activity was noted in the remaining of the P19-GUS transgenics (16 plants) (data not shown).
P19 expression and protein accumulation in the high GUS expressing P19-GUS plants (belonging to 6 lines) were below detectable levels (Northern and qRT-PCR, Figure 4a; western data not shown). This phenomenon has been previously reported in different species transformed with the wild-type P19 gene [31,[48][49][50][51]. However, P19 was highly expressed in the low GUS expressing P19-GUS plants (one line) (Figure 4a).
Even though seven independent transgenic lines were generated, a detailed molecular analysis is presented here for the representative highest GUS expressing line 1 and the lowest GUS expressing line 3. For instance, P19 was expressed in P19-GUS line 1 (three representative plants) by an average of only 1.9% relative to the P19 highest expressing plant (100%) of P19-GUS line 3 ( Figure 4a). GUS levels, on the other hand, accumulated in plants of P19-GUS line 1 by an average of 21.2-fold more than in those of P19-GUS line 3 (Figure 4a). At the DNA level, P19-GUS line 1 displayed a simpler profile than P19-GUS line 3, which showed a multiple loci integration pattern of the P19 and GUS genes (data not shown).
To further investigate the contrasting results observed with the P19-GUS transgenic line 3 (expressing high P19, but low GUS levels), we checked for the presence of any possible mutation in the P19 gene as well as for the methylation status of the coding region of P19 and GUS and of the Ubi promoter driving the expression of both genes. The P19 gene, derived from cDNA synthesized from RNA of each of the three plants of P19-GUS line 3, did not have any mutations (data not shown). The methylation of the GUS gene (driven by the Ubi promoter) and the Ubi promoter (driving P19 or GUS) was assessed by Southern blot hybridization, using genomic DNA from two representative plants of P19-GUS line 3 digested with methylation-sensitive HpaII (H), and methylationinsensitive MspI (M), restriction endonucleases, and probes for the GUS coding region and the Ubi promoter, respectively. As shown in Figure 4b, the P19-GUS line 3 plants exhibited a high level of methylation of the GUS gene and the Ubi promoter, compared to their unmethylation status in the P19-GUS line 1 plants (expressing high GUS and low P19 levels); the majority of the hybridizing HpaII fragments in line 3 were of higher molecular weight. Methylation of the P19 gene was also revealed in the P19-GUS line 3 plants when Southern blot hybridization was performed using the P19 gene as a probe (data not shown).

Reproducible Transient Expression Systems for the Rapid Screening for Functional RNA Silencing Suppressors in Sugarcane
Transient gene expression is influenced by several factors including species and physiological status of the explant [52], transformation parameters [53,54], timing of gene integration [55], cell death and loss of DNA [56,57], and gene structure [58]. PTGS also plays an important role in the post-introduction gene expression decline, so-called transient expression [39,59].
In the present study, we have established two simple and reproducible transient expression systems for screening for functional viral RNA silencing suppressors in sugarcane. The first transient system is based on the co-expression of the suppressor and the target gene in vivo in young leaf segments via particle bombardment. It provides an easy and rapid evaluation of the suppressor activity due to the simplicity of the transformation method and the type of target tissue that offers a large and homogeneous surface for detection of the EYFP reporter gene. The use of image analysis was important to asses the EYFP expression level over time by determining indexes such as foci count, EYFP  expression (gray values6pixel number) and total expression levels (combination of gray values6pixel number in green and red channels) (ImageJ software), and to correlate them with the levels of transgene protection by viral suppressors. This assay system is non-destructive and has been previously shown to be useful in sugarcane for studying efficiencies of terminators [58], and in lima bean for analyzing the effect of RNA silencing suppressors [39,59] and the activity of a soybean promoter [60].
Our establishment of a second transient expression system, based on sugarcane protoplasts transfected via PEG and using the GUS reporter gene, has allowed us to study the efficiency of RNA silencing suppressors in a single cell and over time. Although the isolation of sugarcane protoplasts was first achieved in the 1970s [61], and several scientists have reported their subsequent use in plant regeneration [62,63] and stable transformation by electroporation or by PEG [64][65][66], this is considered to be the first established sugarcane protoplast-based RNA silencing assay. Quantitative analysis of GUS made it possible to correlate the suppressor activity with the efficiency of silencing suppression at the cellular level over time. The protoplast cells are considered to be more homogeneous than the cells within a plant and allow the collection of consistent data for a detailed time course analysis of the suppressor activity. The high yield of viable isolated protoplasts as well as the improved transfection efficiency played key roles in the establishment of the sugarcane protoplast system for transient monitoring of the efficiency of silencing suppressors. The 30% transfection efficiency of sugarcane protoplasts achieved in this study is very acceptable as compared to the 50-70% protoplast transfection efficiency in rice [67].
The protoplast-based RNA silencing assay offers unique advantages to study the mechanisms of RNA silencing in protoplast cells. However, some of the differing results between protoplasts and in planta studies, which are caused by protoplasts as single cells or by the cells within a plant body, may have biological differences that affect the RNA silencing pathway [36]. Therefore, it is necessary to have a combination of single cell and in planta studies to generate information on the kinetic features of the RNA silencing suppressors.

Enhancement of Transgene Expression in Sugarcane by the use of the P19 Viral RNA Silencing Suppressor
In the present study, five viral RNA silencing suppressors, P19, P19/R43W, cb, P1/HC-Pro and SCBV OrfI, were co-expressed transiently with the EYFP or GUS reporter gene in sugarcane leaf segment tissues and protoplasts in order to investigate their activity as suppressors of post-transcriptional transgene silencing. Although these suppressors have been well studied previously, the primary transient expression assay system used relied on Agrobacteriuminfiltration of GFP into model plant systems like N. benthamiana leaves [16], as verified in the present study ( Figure S1). However, several monocot species, including sugarcane, are not amenable to Agrobacterium-infiltration, thus requiring alternative approaches for transient studies. We demonstrated that the viral RNA silencing suppressors operate in sugarcane leaf segments and protoplasts by co-bombardment or co-transfection with the target reporter genes, EYFP and GUS. Our findings indicated that P19, P19/R43W and cb suppressors contributed to increased and extended EYFP or GUS expression, and this increase was highly dependent on the system used. In young leaf segments, P19 and cb significantly enhanced EYFP expression and extended its peak to over 240 h  post-bombardment ( Figure 2). In protoplasts, P19, P19/R43W and cb resulted in a significant increase in GUS activity until 24 h post-transfection ( Figure 3). The fact that P19 and cb worked well in our systems indicate that siRNA sequestration is probably more effective for suppression than inhibiting the RISC complex, as is the case with P1/HC-Pro that did not have any effect [1,20,68]. The ability of P19, P19/R43W and cb to enhance transgene expression in transient assays has been well documented in N. tabacum and N. benthamiana [20,36,46], but using Agrobacteriummediated delivery of these suppressors. The effect of a viral suppressor on the expression of the cointroduced EYFP or GUS was found to be dependent on the dose of the co-delivered suppressor in our sugarcane transient expression system. Co-expression of EYFP with increasing doses of the P19 suppressor was noted to enhance EYFP expression and prolong the EYFP expression peak by at least 120 h in young leaf segments (Figure 1). Similarly, an increase in GUS activity was achieved in protoplasts by increasing the amount of co-transfected P19 (Table 1). These results confirm a previous report where the suppression activity of P19 was dosage dependent in N. benthamiana [69].
The sugarcane transgenic independent lines co-expressing the GUS reporter and the suppressor P19 generated in the present study showed a significant (p,0.05) enhancement in their GUS activity levels by 1.9-fold to 3.5-fold more than those expressing GUS alone ( Table 2). The P19-GUS transgenic plants were noted to develop normally, with no detrimental phenotypic effects, indicating that P19 was tolerated within the stably transformed sugarcane plants. This is an improvement to the transgenic expression of P19 in A. thaliana, N. tabacum and N. benthamiana that often yielded plants with deformed phenotypes [31,46,51]. While we can not rule out other possibilities, the normal development and growth in the P19-GUS transgenic lines is consistent with the low expression levels of P19 (Figure 4a).
The majority of the P19-GUS transgenic sugarcane plants exhibited high GUS levels with a low detectable P19 expression level (Figure 4a) to avoid plant toxicity, possibly due to the use of the strong constitutive Ubi promoter. The methylation status of the Ubi promoter as well as that of GUS and P19 genes in one silenced line certainly adds to the understanding of this phenomenon. Among the seven P19-GUS transgenic lines characterized in this study, line 3 (silenced line), represented by three plants, exhibited low GUS transcript level and activity (Figure 4a). We envision at least two possible reasons for the lower GUS expression levels. One is that P19 is defective or less efficient. Alternatively, GUS expression is silenced at the step(s) that can not be overcome by P19. We cloned the P19 cDNAs derived from RNA of each of the three plants of line 3. Among 5 of the cDNA clones, we did not find any mutations, indicating that the P19 transcript was not altered. Conversely, the P19-GUS line 3 exhibited a high level of methylation in both the GUS gene and the Ubi promoter, compared to their unmethylation status in the non-silenced P19-GUS line 1 expressing high GUS levels (Figure 4b).
Two types of transgene-induced gene silencing are known to exist in plants [70,71]. One type acts at the transcriptional level (TGS), through repression of transcription, where the transgene possesses sequence homology to the promoter of the silenced gene, and the other type works at the PTGS level, through mRNA degradation, requiring homology in the transcript. The TGS is often associated with increased DNA methylation, while PTGS involves sequence-specific RNA degradation, although methylation in either coding or transcribed regions of silenced transgenes has been detected in many cases of PTGS [72][73][74]. The methylation of the coding region of the GUS gene in the P19-GUS line 3 (Figure 4b) is more likely to be associated with PTGS, while the Ubi promoter methylation (Figure 4b) is related to TGS. The P19 suppressor is probably not functional due to its silencing at the PTGS level, while GUS is silenced at the TGS and PTGS levels. This data is consistent with the methylation of the zein gene and the phaseolin promoter in silenced transgenic soybean lines [75].

Conclusion
Two transient transgene expression assays based on young leaf segments and protoplasts, and confirmed by stable transgene expression, were successfully established in the present study to provide a rapid and reproducible versatile system to screen for functional RNA silencing suppressors in sugarcane and other plant species. This system is the first to be developed in sugarcane, and it combines in vivo, single cell and in planta studies to generate more information on the kinetics of the suppressor activities.
The use of RNA silencing suppressors, specifically the TBSVencoded P19 suppressor, proved to be an efficient strategy in allowing for high levels of foreign protein production, whether by transient or stable transgene expression, to counteract the deleterious effects of RNA silencing in sugarcane. This approach provides a suitable platform for the cost-effective production of high-value recombinant proteins for the exploitation of a variety of biotechnologically attractive plant species, such as sugarcane and other high biomass producers, as biofactories. Researchers have previously used the Artichoke mottled crinkle virus-encoded P19 suppressor in Agrobacterium infiltration transient gene expression systems to produce high yields of biopharmaceuticals, namely a human antibody against the tumour-associated antigen tenascin-C [76] and the HIV-1 Nef protein [77] in N. tabacum and N. benthamiana, respectively.

Target Tissue, DNA Particle Bombardment and Transgenic Plants
Stalk tops of field-grown sugarcane (Saccharum spp. hybrid, commercial variety CP72-1210) were collected during the growing season at the Texas A&M AgriLife Research Annex Farm (Weslaco, Texas). No specific permits were required for the described field study and the location, and the location is not privately owned or protected in any way. The field study did not involve endangered or protected species. Young leaf segments were prepared from the sugarcane stalk tops for transient gene expression assays as described by Beyene et al [58]. Briefly, leaf blades and sheaths were removed down to leaf 1 (the top visible dewlap leaf), and the upper 20-30 cm portion of shoot (leaf roll stalk) was surface sterilized in 70% (v/v) ethanol for about 20 min. The two outermost leaf sheathes were discarded, and leaves at position -3 were unfolded, their midribs removed and their blade cut into about 2.5 cm 2 leaf segments. For transient gene expression, segments were pre-cultured adaxial side down onto MS medium [79] with 2,4-dichlorophenoxyacetic acid (2,4-D) at 0.6 mg/L (MS0.6 medium) [80] and casein hydrolysate (500 mg/ L) for 3 days in the dark at 28uC before DNA particle bombardment.
For stable transgene expression, sugarcane leaf roll discs were obtained from the stalk tops as described for leaf segments, and used as explants for DNA particle bombardment. Immature leaf rolls close to the apical meristem were sliced transversely into 1 mm thick sections and cultured on MS0.6 medium for 10-28 days. Leaf roll discs were preconditioned on MS0.6 osmoticum (MS0.6 with 36.44 g/L of D-mannitol and 36.44 g/L of Dsorbitol) for 4 h prior to and after DNA particle bombardment. Bombarded leaf roll discs were maintained on MS0.6 medium for 7 days in the dark at 28uC for recovery. They were later broken into small pieces and incubated in the dark at 28uC on callus induction medium, MS0.6 with Bialaphos (4 mg/L) selection, for a total of 4 weeks, with sub-culturing every two weeks. For shoot regeneration, calli from leaf roll discs were grown on MS supplemented with kinetin (2 mg/L), naphthalene acetic acid (2 mg/L) and Bialaphos (4 mg/L) for 6-8 weeks under a light (16 h)/dark (8 h) photoperiod. Green shoots of approximately 2 cm in height were transferred into MS rooting medium containing indole-3-butyric acid (4 mg/L) and Bialaphos (4 mg/ L). Rooted plantlets were transferred to potting soil (Metromix, Scotts, Hope, AR) in pots and maintained in the greenhouse.
DNA coating for particle gun bombardment was performed according to Beyene et al [58]. Briefly, tungsten particles (M17, 1.1 micron; Bio-Rad Laboratories, Hercules, CA) were sterilized in absolute ethanol and resuspended in nuclease-free water to a final concentration of 60 ng/mL following the manufacturer's instructions. Plasmid DNA was precipitated onto tungsten particles at a concentration of 2.0 mg (GUS; for stable expression) or 0.5 mg (EYFP; for transient expression) DNA per mg of tungsten using calcium chloride (CaCl 2 ) (2.5 M) and spermidine (0.1 M). The molecular ratio of the GUS or EYFP plasmid to the suppressor plasmid was 1:2. The DNA-coated tungsten particles were resuspended in 40 mL of absolute ethanol, and 4 mL of this suspension was used per bombardment of the target tissue (leaf segment or leaf roll). For DNA particle discharging, a modified particle inflow gun [81] with helium gas (110 psi) was used, as described previously [52].

Protoplast Isolation and Transfection
Protoplasts were isolated from suspension cell cultures of callus originating from sugarcane immature leaf roll discs using the modified methods of Chen et al [64] and Yoo et al [33]. Briefly Transfection of protoplasts with plasmid DNA was performed according to Yoo et al [33]. Briefly, protoplasts (1610 5 ; 100 mL) were transferred into a 2-mL microcentrifuge tube (round-bottom) and mixed gently with plasmid DNA (5 mg of GUS reporter plasmid and 10 mg of suppressor plasmid in 10 mL volume). Equivalent volumes of sterile water (mock-transfection) and empty vector (pUbi-Tnos) were used as controls for transfection. Protoplasts were mixed gently with a PEG-calcium solution [40% polyethylene glycol-4000 (PEG), 0.2 M D-mannitol and 100 mM CaCl 2 ] (110 mL) and incubated for 10 min at room temperature. Transfection was terminated by the dilution of the mixture in W5 solution (440 mL). Transfected protoplasts were collected by centrifugation for 2 min at 1006g and suspended in W5 solution (250 mL). GUS expression was analyzed after incubation of the protoplasts in the dark for 24 h at room temperature. The number of protoplasts expressing EYFP was determined manually using a SZX7 fluorescence stereomicroscope with a DP71 cooled CCD camera (Olympus, Center Valley, PA) and a YFP filter (85.5x magnification).

Southern Blot, Northern Blot and Quantitative RT-PCR Analyses
Genomic DNA and total RNA were isolated from liquid nitrogen-ground leaf tissues (0.5-1 g fresh weight) collected from young leaves of 3-4 month-old sugarcane transgenic plants according to Tai and Tanksley [82] and Damaj et al [83], respectively.
Pre-hybridization, hybridization, washing and detection of DNA and RNA gel blots were performed as described by Sambrook et al [86] and Mangwende et al [85], using Church's buffer. The GUS-specific probe was obtained from pUbi:GUS ( Figure S3) by BbsI and SacI digestion, and all of the five RNA silencing suppressor probes were prepared from their respective constructs ( Figure S3) after digestion with PstI. For methylation analysis, DNA probes were obtained by further digesting the GUSspecific probe with MspI into 8 fragments, and by releasing the Ubi promoter from pUbi:EYFP:Tnos ( Figure S3) with HindIII and NcoI. Probes were labeled with [a-32 P] dCTP using the Random Primers DNA Labeling kit (Invitrogen, Carlsbad, CA) [85].

Analysis of b-glucuronidase Activity
Quantitative b-glucuronidase (GUS) assays were performed on transfected protoplasts and leaf tissue of transgenic plants using 4methylumbelliferyl-b-D-glucuronide (MUG) as a fluorescent substrate [87].
Transfected protoplasts were harvested by centrifugation at 1006g for 2 min, and stored at -80uC until analysis. Frozen protoplasts were ruptured in GUS extraction buffer (50 mM sodium phosphate buffer pH 7.0, 10 mM 2-mercaptoethanol, 10 mM EDTA pH 8.0, and 0.1% [v/v] Triton X-100) (100 mL) by vortexing for 2 s, and incubated on ice for 5 min. Total soluble protein extracts were collected by centrifugation at 10006g for 2 min at 4uC.
Leaf tissue (500 mg) of transgenic plants (3-4 month-old), ground into powder in liquid nitrogen, was suspended in GUS extraction buffer (750 mL) by brief vortexing and incubated on ice for 1 h. Total soluble protein extracts were collected by centrifugation at 12,0006g for 10 min at 4uC.
Fluorometric GUS assay was carried out on total soluble protein extracts from protoplasts (25 mL) and leaf samples (10 mL of extract and 15 mL of GUS extraction buffer) in 4 mM MUG assay buffer (25 mL) by incubation for 60 min at 37uC. The reaction was stopped by the addition of 0.2 M sodium carbonate (950 mL). Fluorescence was measured at 455 nm (emission) and 365 nm (excitation) using a VersaFluor TM Fluorometer (Bio-Rad Laboratories). Protein concentrations were determined by the Lowry assay method using the DC protein assay kit (Bio-Rad Laboratories). Protein extracts from protoplasts transfected with empty vector and from leaves of non-transgenic plants were used as a negative control.

EYFP Imaging and Analysis
Images (4,08063,072 pixels) of sugarcane young leaf segments expressing EYFP were collected every 6 h post-bombardment for at least 240 h by a SZX7 fluorescence stereomicroscope with a DP71 digital camera (Olympus) fitted with YFPHQ filters (excitation of 490-500 nm and emission of 515-560 nm) under 15x magnification. EYFP expression was quantified using the ImageJ version 1.42u software according to the revised method of Chiera et al [59,60] and as described by Beyene et al [58]. Briefly, a series of images taken over time from each sample were saved in separate folders and imported into AdobeH ImageReady TM as frames. The sequence of images was then resized to 8006600 pixels and exported as ''mov'' files. For image analysis, ''mov'' files were imported into the ImageJ software, and a representative 4006300 pixel area was selected and cropped. This 4006300 pixel area was used as the original image for calculating the EYFP foci count and EYFP expression level. Subsequently, all image series were separated into red, green and blue channels and their background was corrected. For further analysis, only the green channel was used, since the contribution of the red channel was found to be lower than 5% of total EYFP expression. Plugins for quantification of EYFP expression and foci count were kindly provided by J. Chiera and C. Hernandez-Garcia (Department of Horticulture and Crop Sciences, The Ohio State University, Columbus, Ohio). EYFP expression was calculated by multiplying the mean grayscale value per pixel by the number of EYFP expressing pixels with the resulting EYFP expression values being unitless. The number of EYFP expressing foci was determined by counting the number of spots bigger than ten pixels.

Statistical Analysis
Data were collected from 2-3 independent experiments, with 6-10 replicates per experiment, and subjected to an analysis of variance (ANOVA) using the General Linear Model procedure of the Statistical Analysis System 8.1 (SAS Institute Inc., Cary, NC). Mean separation was performed using the Student-Newman-Keuls (SNK) test. Figure S1 Comparison of suppressors in Nicotiana benthamiana. The plasmids carrying the silencing suppressors P1/HC-Pro (HcPro) and cb are pGD binary vectors specifically generated to be used with Agrobacterium, and modified from the binary vector pCAMBIA 1303, with a multiple cloning site downstream of a Cauliflower mosaic virus 35S promoter, and upstream of an Agrobacterium nopaline synthase poly(A) signal [88,89]. P19 is expressed from the binary vector pCass4N, a derivative of a pBin19 binary vector [46]. The plasmid carrying the Green Fluorescent Protein (GFP) gene is 35S-GFP (provided by David C. Baulcombe, University of Cambridge, Cambridge, UK) [90]. All of the silencing suppressors were infiltrated into Nicotiana benthamiana using the Agrobacterium strain EHA [91], which exhibited a less virulent host response than others (data not shown). Three week-old N. benthamiana plants were infiltrated with 35S-GFP and silencing suppressors at an optical density of 0.8, mixed as indicated. The plants were then photographed at different days post-infiltration (dpi) under a 488 nm wavelength UV light with a 4 s exposure and no flash, to monitor the levels of the fluorescent GFP expressed. The plants in the first column, labeled EHA, are those infiltrated with untransformed Agrobacterium, as a negative control. When necessary, supplementation with Agrobacterium EHA was done to ensure that each leaf was inoculated with 0.5 mL of bacterial culture. For each suppressor, the expression was verified by immuno-blotting (data not shown). Genes encoding the suppressors were under control of the maize ubiquitin 1 promoter, as described in Materials and Methods. Plasmid DNA of the appropriate construct was introduced into onion cells using a PDS-1000/He particle delivery system. Bombardment was performed at 9 cm from targets using gold particles (1.0 micron; Bio-Rad Laboratories) coated with plasmids expressing EYFP or viral suppressors under 27 inch Hg and 1100 psi helium pressure. Plasmid DNA was precipitated onto the gold particles using calcium chloride (2.5 M) and spermidine (0.1 M).

Supporting Information
For co-introduction of two and three different plasmids, 4.5 mg and 3.0 mg of each plasmid was used, respectively. The bombarded epidermal peels were incubated on MS osmoticum for 48-72 h at 25uC in the dark. Fluorescence was monitored using a fluorescence binocular microscope Olympus SZX10 with an excitation wavelength of 490 nm. (b) Comparison of different suppressors. Similarly sized onion peel sections were bombarded as described for (a), and the number of fluorescent cells was counted. (TIF) Figure S3 Map of suppressor and reporter gene constructs for stable sugarcane transformation. For genetic construct assembly, refer to Materials and Methods. P1/HC-Pro is derived from Tobacco etch virus, cb from Barley stripe mosaic virus, P19 and P19/R43W, a mutant of P19, from Tomato bushy stunt virus, and SCBV OrfI from Sugarcane bacilliform virus; GUS: b-glucuronidase; EYFP: Enhanced Yellow Fluorescent Protein; pUbi: Maize ubiquitin 1 promoter; Tnos: Agrobacterium tumefaciens nopaline synthase terminator. Boxes are not drawn to scale. (TIF) Figure S4 Multicolor fluorescence imaging of cells in sugarcane young leaf segments expressing the EYFP reporter gene, under low optical magification. Images of EYFP expression were collected with a SteReo Lumar.V12 fluorescence stereomicroscope and an AxioCam ICc3 digital camera (15x magnification) (Carl Zeiss) from young leaf segments at 48 h after bombardment with EYFP (pUbi:EYFP:Tnos; Figure  S3) (0.5 mg per shot) (bar = 0.5 mm). Vector with no EYFP (pUbi:Tnos; Figure S3) and water were used as negative controls. Images were taken under bright light as well as with filters for rhodamine (filter model FS20) (excitation: 546/12 nm, emission: 575-640 nm), eGFP (filter model FS38) (excitation: 470/40 nm, emission: 525/50 nm) and YFP (filter model FS46 HE) (excitation: 500/25 nm, emission: 535/30 nm). eGFP and YFP images were taken under 700 ms exposure, and rhodamine images were taken under autoexposure. (TIF) Figure S5 Multicolor fluorescence imaging of cells in sugarcane young leaf segments expressing the EYFP reporter gene, under high optical magnification. Images of EYFP expression were collected with a SteReo Lumar.V12 fluorescence stereomicroscope and an AxioCam ICc3 digital camera (150x magnification) (Carl Zeiss) from young leaf segments at 48 h after bombardment with EYFP (pUbi:EYFP:Tnos; Figure  S3) (0.5 mg per shot) (bar = 0.05 mm). Vector with no EYFP (pUbi:Tnos; Figure S3) and water were used as negative controls. Images were taken under bright light as well as with filters for rhodamine (filter model FS20) (excitation: 546/12 nm, emission: 575-640 nm), eGFP (filter model FS38) (excitation: 470/40 nm, emission: 525/50 nm) and YFP (filter model FS46 HE) (excitation: 500/25 nm, emission: 535/30 nm). Overlaid images were generated from merged images of bright light and YFP filter. eGFP and YFP images were taken under 700 ms exposure, and rhodamine images were taken under autoexposure. (TIF) Figure S6 Quantitative assessment of transient expression of the EYFP reporter gene in sugarcane young leaf segments. EYFP expression as measured by foci count (a) and expression level (b) was monitored in 3 day-pre-cultured young leaf segments at 48 h after bombardment with EYFP (pUbi:EYFP:Tnos; Figure S3) (0.5 mg per shot). Vector with no EYFP (pUbi:Tnos; Figure S3) and water were used as negative controls.

Enhanced Gene Expression and Silencing Suppressors
Values represent means with standard error from three independent experiments and 10 replicates per experiment. Means with the same letter are not significantly different (p.0.05). Quantitation of EYFP foci counts and expression levels is provided in Materials and Methods. (TIF) Figure S7 Transient expression of the EYFP reporter gene in sugarcane protoplasts. (a) A seven day-old leaf roll disc growing on MS shoot regeneration medium (9.5x magnification, bar = 1.0 mm); (b) Callus regenerated from leaf roll disc after subculture on MS medium for 4-6 weeks and grown in MS liquid medium to obtain suspension cells (12x magnification, bar = 1.0 mm); (c) Protoplasts isolated from suspension cells under bright light (400x, bar = 20 mm); (d) and (e) Protoplasts transfected with the pUbi:EYFP:Tnos vector expressing EYFP under bright light (100x magnification) and EYFP filter (100x magnification), respectively (bar = 100 mm); (f) Overlaid image of (d) and (e) showing transfection efficiency; a transfected protoplast is indicated by an arrow (bar = 100 mm). Microphotographs of (a) and (b) were collected using an Olympus SZX7 fluorescence microscope with a DP71 camera. Microphotographs of (c), (d), (e) and (f) were obtained with an Olympus BX51 fluorescence microscope with a DP72 camera. (TIF) Figure S8 Effect of viral RNA silencing suppressors on transient expression of the EYFP reporter gene in sugarcane young leaf segments. Images of EYFP expression were collected with a SZX7 fluorescence stereomicroscope and a DP71 digital camera (15x magnification) (Olympus) from the same young leaf segments at 24-240 h after co-bombardment with EYFP (pUbi:EYFP:Tnos; Figure S3) (0.25 mg) and each of the RNA silencing suppressors (driven by the Ubi promoter, Figure  S3), HC-Pro, cb, P19, P19/R43W and SCBV OrfI (0.5 mg) (bar = 0.5 mm). Vector with no suppressor (pUbi:Tnos; Figure  S3) was used as a negative control. (TIF)