Agrobacterium-Mediated Gene Transient Overexpression and Tobacco Rattle Virus (TRV)-Based Gene Silencing in Cassava

Agrobacterium-mediated transient expression and virus-induced gene silencing (VIGS) are very useful in functional genomics in plants. However, whether these methods are effective in cassava (Manihot esculenta), one of the most important tropical crops, remains elusive. In this study, we used green fluorescent protein (GFP) and β-glucuronidase (GUS) as reporter genes in a transient expression assay. GFP or GUS could be detected in the infiltrated leaves at 2 days postinfiltration (dpi) and were evidenced by visual GFP and GUS assays, reverse-transcription PCR, and Western blot. In addition, phytoene desaturase (PDS) was used to show the silencing effect in a VIGS system. Both Agrobacterium GV3101 and AGL-1 with tobacco rattle virus (TRV)-MePDS-infiltrated distal leaves showed an albino phenotype at 20 dpi; in particular, the AGL-1-infiltrated plants showed an obvious albino area in the most distal leaves. Moreover, the silencing effect was validated by molecular identification. Notably, compared with the obvious cassava mosaic disease symptom infiltrated by African-cassava-mosaic-virus-based VIGS systems in previous studies, TRV-based VIGS-system-infiltrated cassava plants did not show obvious virus-induced disease symptoms, suggesting a significant advantage. Taken together, these methods could promote functional genomics in cassava.


Introduction
Aiming at analyzing genomic sequences and functions, functional genomics has rapidly developed as a result of sequencing projects of different species, especially of important crops. Cassava (Manihot esculenta), a kind of tuber crop, is widely cultivated in the tropics and some subtropical areas. Because of its low planting cost and high efficiency, it has become one of the most important industrial crops in the world [1]. Although cassava sequencing and resequencing were accomplished years ago, cassava functional genomics has developed slowly due to the time-consuming nature and low transformation rate of obtaining stable transgenic cassava plants [2]. Therefore, it is necessary to find another way to promote rapid gene function studies of cassava.
Agrobacterium-mediated transient overexpression is widely used in gene function studies in plants [3]. The most important advantages of this method are its rapid and simple process and that it does not require complex equipment, unlike other overexpression assays [4]. The simple protocol is accomplished by creating an overexpression vector, transforming it into Agrobacterium, and infiltrating the Agrobacterium into plant culture [5]. Green fluorescent protein (GFP) and β-glucuronidase (GUS) are two convenient reporter genes in biology which can be observed by fluorescent microscopy or direct staining [6,7]. Normally, they can continue to express, reaching their peak at 3 days postinfiltration (dpi), and exert obvious expression for at least one week [4,8]. In addition, they can be efficiently expressed in a wide range of plant species. So far, transient overexpression assays have been successfully used in many species, such as Arabidopsis spp. [4], tobacco [5], Mimulus lewisii [6], Piper colubrinum [7], rose [9], soybean [10], Theobroma cacao L. [11], cotton [12], Brassica juncea [13], potato [14], tomato [15], and lettuce [15]. However, its effectiveness in cassava requires further investigation.
In this study, Agrobacterium-mediated transient overexpression and VIGS were established in cassava. The phenotype and examination of these assays illustrated the feasibility of the protocols, which should help promote the rapid analysis of functional genomics in cassava.

The Effect of the Transient Expression Assay in Cassava Leaves
To investigate the effect of GFP transient expression, the 35S::GFP vector [33] was transformed into Agrobacterium AGL-1 and GV3101, and the transformed strains were infiltrated into cassava leaves, respectively (Figure 1). At 2 dpi, the infiltrated leaves were harvested and examined by confocal laser-scanning microscopy. Agrobacterium with 35S::GFP-infiltrated leaves showed clear green fluorescence in both the cytoplasm and nucleus, while the empty Agrobacterium-infiltrated leaves showed no fluorescence ( Figure 2A). Reverse-transcription PCR was performed to clarify the effect of transient expression. Agrobacterium with the 35S::GFP-infiltrated sample showed a bright band by PCR, indicating that GFP was expressed in the cassava leaves ( Figure 2B). Moreover, GFP was also detected by Western blot, which was consistent with confocal observation ( Figure 2C).     To further confirm the effect of transient expression, GUS was used as the second marker. After GUS staining, Agrobacterium with the 35S::GUS-infiltrated area became light blue ( Figure 3A), indicating that GUS was expressed at the protein level. In addition, reverse-transcription PCR showed that the expression of 35S::GUS could be examined at the transcript level ( Figure 3B), with significant GUS activity ( Figure 3C). These results suggested that GFP and GUS were expressed with biological activity in the transient expression assay.
Based on the above results of GFP and GUS overexpression, the effects of transient overexpression showed no significant difference in GFP expression and GUS activity between AGL-1 and GV3101 (Figures 2 and 3, Table 1).  Briefly, the Kolmogorov-Smirnov test and Levene's test were performed to check the normality of the data distribution and the homogeneity of variance of the data, respectively. The statistical analysis was performed using Student's t-test and the Tukey-Kramer test. Asterisk symbol (*) indicates significant difference at p < 0.05. n = 10 per group. All data are expressed as mean ± SD of three independent experiments.

Silencing of MePDS in Cassava
To reveal the effect of TRV-based VIGS on cassava, phytoene desaturase (PDS) was selected as a reporter gene. A 370 bp sequence of MePDS was amplified and inserted into the multiple cloning site of the pTRV2 vector. The pTRV1, pTRV2, and pTRV2-MePDS vectors were transformed into Agrobacterium, respectively. The cultivated Agrobacterium solution was infiltrated into cassava local Agrobacterium with no plasmid was used as the mock, and Agrobacterium containing 35S::GUS was used as the sample. The leaves were stained using GUS staining solution. Bar = 1 cm. (B) The relative transcript level of GUS is shown by reverse-transcription PCR, and EF1 was used as a reference gene. (C) The GUS activities of infiltrated leaves. Statistical tests were performed using IBM SPSS (v21). Briefly, the Kolmogorov-Smirnov test and Levene's test were performed to check the normality of the data distribution and the homogeneity of variance of the data, respectively. The statistical analysis was performed using Student's t-test and the Tukey-Kramer test. Asterisk symbol (*) indicates significant difference at p < 0.05. n = 10 per group. All data are expressed as mean ± SD of three independent experiments.

Silencing of MePDS in Cassava
To reveal the effect of TRV-based VIGS on cassava, phytoene desaturase (PDS) was selected as a reporter gene. A 370 bp sequence of MePDS was amplified and inserted into the multiple cloning site of the pTRV2 vector. The pTRV1, pTRV2, and pTRV2-MePDS vectors were transformed into Agrobacterium, respectively. The cultivated Agrobacterium solution was infiltrated into cassava local leaves and axillary buds (Figure 1). At 20 dpi, AGL-1-infiltrated cassava showed an obvious albino area in the distal leaves, especially for the area around the main vein ( Figure 4). leaves and axillary buds (Figure 1). At 20 dpi, AGL-1-infiltrated cassava showed an obvious albino area in the distal leaves, especially for the area around the main vein ( Figure 4). To further investigate the level of MePDS silencing, DNA was extracted from distal leaves and sequences of pTRV1, pTRV2, or pTRV2-MePDS were detected by PCR. The result indicated that the sequences of pTRV vectors could be detected in the distal leaves of all AGL-1-infitrated cassava plants but not in the wild-type (WT) cassava ( Figure 5A). In addition, the transcript levels of pTRV1 and pTRV2 sequences in the distal leaves were analyzed. Both the transcripts of pTRV1 and pTRV2 could be obviously examined in the pTRV and pTRV-MePDS cassava plants but showed no PCR band in WT cassava ( Figure 5B). Notably, the pTRV-MePDS cassava leaves showed an obvious albino area around the main vein. The transcript level of MePDS in the distal leaves of MePDS-VIGS plants was 37.9% and 53.1% of that in the mock ( Figure 5C), displaying significantly lower chlorophyll content than the mock ( Figure 5D). Based on these results, we could conclude that the gene silencing of MePDS in the distal leaves of MePDS-VIGS plants was caused by VIGS, and the albino phenotype in the distal leaves was derived by VIGS but not the infection-induced effects, which resulted in no significant difference in local leaves.
Compared with the phenotype of Agrobacterium AGL-1-infiltrated distal leaves, GV3101infiltrated distal leaves showed the albino phenotype in the part around the petiole (Supplementary Figure S1), with slightly weaker PCR detection bands, a lower MePDS transcript level, and lower chlorophyll content than the mock (Supplementary Figure S2). For AGL-1, 75.00% of cassava plants showed less than 60% transcript level of MePDS in the distal leaves, and only 37.50% of the plants exhibited the albino phenotype (Table 2). For GV3101, 62.50% of cassava plants showed a lower transcript level of MePDS in the distal leaves, and only 12.50% of plants showed the albino phenotype To further investigate the level of MePDS silencing, DNA was extracted from distal leaves and sequences of pTRV1, pTRV2, or pTRV2-MePDS were detected by PCR. The result indicated that the sequences of pTRV vectors could be detected in the distal leaves of all AGL-1-infitrated cassava plants but not in the wild-type (WT) cassava ( Figure 5A). In addition, the transcript levels of pTRV1 and pTRV2 sequences in the distal leaves were analyzed. Both the transcripts of pTRV1 and pTRV2 could be obviously examined in the pTRV and pTRV-MePDS cassava plants but showed no PCR band in WT cassava ( Figure 5B). Notably, the pTRV-MePDS cassava leaves showed an obvious albino area around the main vein. The transcript level of MePDS in the distal leaves of MePDS-VIGS plants was 37.9% and 53.1% of that in the mock ( Figure 5C), displaying significantly lower chlorophyll content than the mock ( Figure 5D). Based on these results, we could conclude that the gene silencing of MePDS in the distal leaves of MePDS-VIGS plants was caused by VIGS, and the albino phenotype in the distal leaves was derived by VIGS but not the infection-induced effects, which resulted in no significant difference in local leaves.
Compared with the phenotype of Agrobacterium AGL-1-infiltrated distal leaves, GV3101-infiltrated distal leaves showed the albino phenotype in the part around the petiole (Supplementary Figure S1), with slightly weaker PCR detection bands, a lower MePDS transcript level, and lower chlorophyll content than the mock (Supplementary Figure S2). For AGL-1, 75.00% of cassava plants showed less than 60% transcript level of MePDS in the distal leaves, and only 37.50% of the plants exhibited the albino phenotype (Table 2). For GV3101, 62.50% of cassava plants showed a lower transcript level of MePDS in the distal leaves, and only 12.50% of plants showed the albino phenotype (Table 2).
These results suggested that both AGL-1 and GV3101 with TRV could induce MePDS silencing and regulate the color of leaves in cassava, with better effects in AGL-1-transformed TRV-based VIGS.  (Table 2). These results suggested that both AGL-1 and GV3101 with TRV could induce MePDS silencing and regulate the color of leaves in cassava, with better effects in AGL-1-transformed TRVbased VIGS.

Discussion
Although genome sequencing of cassava has been completed for several years now, little progress has been made in determining the functional genomics of cassava [34]. In order to address this question and promote functional genomics in cassava, this study investigated Agrobacterium-mediated gene transient overexpression and TRV-based VIGS.
The Agrobacterium-mediated transient overexpression assay is a powerful tool for analyzing gene function in vivo [35]. In this study, Agrobacterium AGL-1 and GV3101 with GFP or GUS overexpressing plasmids were used for transient overexpression. In a previous study, GV3101 containing the 35S::GUS vector was used in MCOL2215 and 60,444 cassava varieties to test the transient expression effect, but no GUS was detected. However, AGL with the 35S::GUS vector showed a positive result [36]. In this study, both GFP and GUS were expressed by the GV3101 strain in SC124 and could be detected at the transcript (Figures 2 and 3) and protein (Figures 2 and 3) levels, indicating that they could be efficiently expressed in cassava leaves. Generally, the GV3101-and AGL-1-based transient overexpression assays had relatively high success rates (Table 1), while the negative results might have been due to the RNA cosilence process [37,38], the morphology, or the structure of the plants [39]. Therefore, whether the target genes are overexpressed in the transient overexpression assay should be analyzed first.
In addition, Agrobacterium-mediated VIGS is another powerful tool to investigate gene function in vivo. Different Agrobacterium strains (GV3101 and AGL-1) with pTRV vectors were used to silence the MePDS gene. AGL-1 with the transformation of pTRV1 and pTRV2-MePDS vector-infiltrated distal leaves showed obvious albino phenotypes at 20 dpi ( Figure 4). Interestingly, the albino phenotype only showed in distal leaves and the albino area was mainly located around the leaf vein, similar to the results in California poppy [40] and Solanum pseudocapsicum [41]. From the albino leaf blade, we found that the area near the petiole became white, while the area distant from the petiole was still green (Figure 4), indicating that TRV-based MePDS silence might spread from the bottom to the tip of the blade in cassava. Additionally, TRV-based Agrobacterium-mediated VIGS was examined by genome DNA or cDNA to ensure the Agrobacterium-mediated transformation and expression of pTRV1 and pTRV2 or pTRV2-X [41,42]. PCR results indicated that Agrobacterium-mediated pTRV1 and pTRV2 or pTRV2-MePDS could be successfully transformed and expressed in the distal leaves of cassava plants, with a 37.9%-53.1% transcript level of MePDS and 58%-76.5% of chlorophyll content in the distal plant leaves ( Figure 5). Because the relative transcript level of MePDS of the albino cassava plants was 71% of the mock and the detected bands were weak, GV3101 with pTRV1 and pTRV2-MePDS infiltrated cassava plants showed less obvious change, and only the part around the petiole became albino ( Supplementary Figures S1 and S2). VIGS is a post-transcriptional gene silencing method [24,25]. In a VIGS system, the expressed levels of silenced marker genes were 30%-40% of control in Solanum pseudocapsicum L. [41], 37% of control in columbine [43], 41%-60% of control in barley [44], 28%-38% of control in tomato [45], and 30%-70% of control in rose [27]. Although these genes could not silenced completely, the albino phenotype was obvious. Consistently, the transcript level of MePDS in MePDS-VIGS plants were 37.9%-53.1% of that in the mock, with an obvious albino phenotype and a 58%-76.5% level of chlorophyll. Compared with GV3101, Agrobacterium AGL-1 showed better effects in the VIGS system (Table 2), which might have been due to the specific hypervirulence of AGL-1 [46]. AGL-1 is derived from pTiBo542, which has a high induction of the vir gene [47], which is necessary for T-DNA transfer [48].
In this study, an Agrobacterium-mediated transient overexpression system was established in cassava. The successful examination of the transcription, translation, and biological activity of expressed proteins suggests its application in subcellular localization experiments, bimolecular fluorescence complementation (BIFC), co-immunoprecipitation (Co-IP), enzyme activity detection, and so on. In addition, a TRV-based Agrobacterium-mediated VIGS system was also successfully verified in cassava ( Figure 1B), as shown by the expressions of pTRV and the corresponding gene, as well as the albino phenotype. Previous studies have found that cassava could be silenced by an African-cassava-mosaic-virus-based VIGS system; however, this could lead to serious malformation of newly grown leaves, thereby resulting in interference in the assay of plant disease resistance [49][50][51]. On the contrary, TRV-based VIGS-system-infiltrated cassava plants only showed mild virus-induced disease symptoms and might be more appropriate for gene function analysis. In conclusion, this study provided two valid and efficient methods for the characterization of gene function, so as to promote functional genomics in cassava.

Plant Materials
The cassava plants South China 124 (SC124) were kindly provided by Dr. Wei Hu (Institute of Tropical Bioscience and Biotechnology, Haikou, China). Two-week-old tissue culture cassava plants (SC124) were transferred to small pots and grown in a chamber for two weeks until the experiment at 25°C under a 16 h light/8 h dark cycle.

Vectors and Vector Construction
The pEGAD and pBI121 vectors were used to express GFP and GUS, respectively. The full-length sequences of GFP and GUS were driven by the 35S promoter. The VIGS assay used pTRV1 and pTRV2 vectors, which have been described previously [52]. The partial sequence of the MePDS gene was cloned into the multiple cloning site of the pTRV2 vector, and the primers are listed in Supplementary  Table S1.

Agrobacterium Infiltration of Cassava
Agrobacterium strains GV3101 and AGL-1 were transformed by different plasmids: pEGAD, pBI121, pTRV1, pTRV2, and pTRV2-MePDS, respectively. GV3101 was cultivated in liquid LB medium containing 50 mg/L kanamycin, 20 mg/L rifampicin, and 50 mg/L gentamycin, while AGL-1 was cultivated in LB medium with 50 mg/L kanamycin, 20 mg/L rifampicin, and 50 mg/L carbenicillin. Both were then shook at 28°C at 200 rpm for 2 days. The bacteria was centrifuged at 4000 rpm for 10 min, then the supernatant was discarded. The remaining bacteria was washed with double-distilled water, then centrifuged, and the supernatant was discarded again. The bacterial sediment was resuspended in MMA solution (10 mM MgCl, 10 mM MES, and 150 µM acetosyringone), the OD 600 was adjusted to 1, and then the resuspended bacterial solution was placed in the dark for 3 h. For the transient expression assay, the standing bacterial solution was infiltrated into the second and third leaves from the top of the cassava by a 1 mL needle. For the VIGS assay, the bacterial solutions containing pTRV1 or pTRV2-MePDS were mixed with the same volume, and the mixture of pTRV1 or pTRV2 was used as the mock. The mixed solution was infiltrated into both the leaves and axillary buds of cassava plants to keep the silencing effect [48]. After infiltration, the cassava plants were removed to the chamber at the same light and temperature.

Reverse-Transcription PCR and Quantitative Real-Time PCR
The first strand cDNA was synthesized by the RevertAid First Strand cDNA Synthesis Kit (K1621, Thermo, Waltham, MA, USA). Then, the cDNA was adjusted to an equal concentration. Reverse-transcription PCR was performed by EasyTaq PCR SuperMix (AS111, TRANS, Beijing, China) with a PCR program of (1) 94°C for 3 min; (2)

Confocal Microscopy Scanning
At 2 dpi, the infiltrated area of cassava leaves was harvested, cut into small pieces, and then observed by confocal laser-scanning microscopy (TCS SP8, Leica, Heidelberg, Germany).

Protein Extraction and Western Blot
Cassava leaves were harvested and ground by liquid nitrogen and phosphate buffer solution (pH 7.4). The extracted solution was centrifuged at 12,000 rpm and 4°C for 10 min. The supernatant was boiled with SDS-PAGE sample loading buffer (P0015, Biyotime, Shanghai, China) for 5 min. After centrifugation for 10 min, the supernatant could be used in Western blot. The Western blot assay was performed according to a previous study [53]. Briefly, the protein samples were loaded into 12% polyacrylamide gel and separated by electrophoresis. The polyacrylamide gel was transferred to a PVDF membrane (475855-1R, Millpore, Massachusetts, America) by a Trans-Blot SD Semi-Dry Electrophoretic Transfer Cell (1703940, Bio-rad, Hercules, America). Then, the PVDF membrane was blocked and incubated in 5% skim milk with anti-GFP antibody (AG281, Biyotime, Haimen, China).

GUS Staining and Activity Detection
GUS staining and GUS activity assay were carried out according to a previous study with slight modifications [48]. The infiltrated cassava leaves were harvested and immersed into GUS staining solution (50 mM NaH 2 PO 4 , 50 mM Na 2 HPO 4 , 10 mM EDTA-Na 2 , 0.5 mM K 4 [Fe(CN) 6 ], 0.5 mM K 3 [Fe(CN) 6 ], 0.1% Triton-X100, and 2 mM X-Gluc; pH 7.0) with vacuum infiltration for 0.5 h in the dark, and then the plant leaves were incubated in the dark at 37°C for at least 12 h. After staining, the leaves were immersed into 70% alcohol to remove chlorophyll.
The infiltrated leaves were harvested and ground by liquid nitrogen and phosphate buffer solution (pH 7.4). The extracted solution was centrifuged (12,000 rpm, 10 min, 4°C) and the supernatant was used for further assay. The extraction was incubated in 1 mM 4-Methylumbelliferyl-b-D-glucuronide (4-MUG) at 37°C, the reaction mixture was taken out per 5 min, and Na 2 CO 3 was added to stop the reaction. The samples were detected by a microplate system (Infinite M200 Pro, TECAN, Hombrechtikon, Switzerland) with 365 nm excitation and 455 nm emission. The GUS activity was calculated by the standard curve made by different concentrations of 4-methylumbelliferone (4-MU).

PCR Detection of Cassava Plants
The transient expressed cassava plants were detected by PCR using genomic DNA and cDNA, with a PCR program of (1) 94°C for 3 min; (2) 29 cycles of 94°C for 30 s, 55°C for 30 s, and 72°C for 2 min; and (3) 72°C for 10 min and 16°C for storage. The primers are listed in Supplementary Table S4.

Chlorophyll Content of Cassava Leaves
The chlorophyll content of the cassava leaves was quantified as previously described [54]. The harvested leaves were ground in 80% acetone. After centrifugation, the absorbance of OD 647 and OD 665 of the supernatant solution was detected by a microplate system (Infinite M200 Pro, TECAN, Hombrechtikon, Switzerland). The total chlorophyll content was 17.90 × OD 647 + 8.08 × OD 665 .

Statistical Analysis
All data are expressed as mean ± SD of three independent experiments. Statistical tests were performed using IBM SPSS (v21). Briefly, the Kolmogorov-Smirnov test and Levene's test were used to check the normality of the data distribution and the homogeneity of variance of the data, respectively. Then, Student's t-test and the Tukey-Kramer test were used for statistical analysis. The asterisk symbol (*) indicates significant difference at p < 0.05.