Manipulation of Auxin Response Factor 19 affects seed size in the woody perennial Jatropha curcas

Seed size is a major determinant of seed yield but few is known about the genetics controlling of seed size in plants. Phytohormones cytokinin and brassinosteroid were known to be involved in the regulation of herbaceous plant seed development. Here we identified a homolog of Auxin Response Factor 19 (JcARF19) from a woody plant Jatropha curcas and genetically demonstrated its functions in controlling seed size and seed yield. Through Virus Induced Gene Silencing (VIGS), we found that JcARF19 was a positive upstream modulator in auxin signaling and may control plant organ size in J. curcas. Importantly, transgenic overexpression of JcARF19 significantly increased seed size and seed yield in plants Arabidopsis thaliana and J. curcas, indicating the importance of auxin pathway in seed yield controlling in dicot plants. Transcripts analysis indicated that ectopic expression of JcARF19 in J. curcas upregulated auxin responsive genes encoding essential regulators in cell differentiation and cytoskeletal dynamics of seed development. Our data suggested the potential of improving seed traits by precisely engineering auxin signaling in woody perennial plants.


Results
Down regulation of JcARF19 affected auxin signalling transduction in Jatropha. According to the whole-genome scan results, we hypothesized that JcARF19 might participate in controlling seed size. To test the hypothesis, firstly, we investigated potential tissue-specific roles of JcARF19 in Jatropha by profiling its expression patterns in roots, stems, leaves, fruits and seeds using quantitative real-time PCR. JcARF19 showed highly similar expression profiles to ubiquitous expression in all organs and highest in endosperm (Fig. 1b), providing evidence that JcARF19 might be related to seed traits in Jatropha. To test the function of JcARF19 in auxin signalling transduction and therefore plant development in Jatropha, we used synthetic tobacco rattle virus (sTRV) based virus-induced gene silencing (VIGS) method, which we developed previously and allowed us to rapidly identify gene function in various plants 40,41 . At 27 days after inoculation of agrobacterium containing VIGS vectors of JcARF19 and positive control Jatropha Chlorata 42 (JcCH42) or empty vector control (EV), we observed distinct smaller newly expanding leaves in JcARF19-silenced plants than EV-treated plants (Fig. 1c) and the difference between them was remarkably as shown in Fig. 1d 34 . Smaller leaf size was also observed in the positive control treatment containing the silenced marker gene, JcCH42, which encodes a subunit of Magnesium (Mg) chelatase, involved in photosynthesis (Fig. 1c) 42 . Recent research showed that auxin treatment enhances ARF19 binding to its target gene promoters, which correlates with the enhancement of transcriptional activity of the ARF19 in Arabidoposis thaliana (Arabidopsis) 43 .We next tested the gene expression of downstream transcription factors esp. Lateral Organ Boundaries-domain (LBDs). Upon auxin treatment, two genes JcLBD18 and JcLBD29 Scientific RepoRts | 7:40844 | DOI: 10.1038/srep40844 which encode putative LBDs of Jatropha were up-regulated differentially (Fig. 1e), especially JcLBD18 which had 7-fold higher expression in EV-treated plants. But in JcARF19-silenced plants, there was no obvious induction of JcLBD18 or JcLBD29 upon auxin treatment, indicating that JcARF19 was essential for proper auxin-mediated signalling transduction process in leaf cells of Jatropha.
Transgenic overexpression of JcARF19 increased seed size and dry seed weight in Arabidopsis. To further investigate the function of JcARF19, at first, we got partial cDNA sequence from a database of sequenced cDNA library prepared from Jatropha seeds 11 . By integrating known genomic sequence of JcARF19 and cloned sequenced information from 5′RACE and 3′RACE, we finally got the full-length coding sequence of JcARF19 (Genbank accession NO. KX988008, detailed sequence information could be found in Supplementary file). JcARF19 protein encoded 1133 amino acids and consists of major functional domains, an amine-terminus B3 DNA binding domain and a carboxyl-terminus (CT) Phox and Bem1p (PB1) domain. Amino acid sequence alignment showed that ARF19s from various plants had high sequence similarity in the three conserved domains (Fig. S1), suggesting functional conservation of ARF19 family proteins. To further investigated the function of JcARF19, we generated thirty transgenic lines for CaMV35S:JcARF19 in Arabidoposis 44 . An obvious increase in plant size and seed size was observed in the 35S:JcARF19 plants compared to wild-type (WT) Col-0 plants ( Fig. 2a-d). Seed length and seed dry weight of 35S:JcARF19 lines were increased remarkably compared with those of WT Col-0 plants as well (Fig. 2e,f). We further conducted oil traits analysis to check the effects of JcARF19 on oil yield or oil composition. There was no significant change in either oil content per dry seed weight or oil composition in transgenic line of 35S:JcARF19 compared with those of WT Col-0 plants (Fig. S2). Since seed size and weight were significantly increased in transgenic line of 35S:JcARF19, the lipid . Samples were collected after treatment of IAA (10 nM) and each treatment had three biological replicates. For qRT-PCR, each biological replicate was replicated three times. Numbers represent mean relative values from three independent experiments with standard deviation. The relative expression level of JcIAA9 in sTRV-silenced plants was normalized as 1.
content per seed was obviously increased. These results indicated that overexpression of JcARF19 increased seed size, seed dry weight and oil yield, but had no effect on oil composition of seeds in Arabidopsis.
Increased seed size and seed numbers by overexpression of JcARF19 in Jatropha. Jatropha was also transformed with 35S:JcARF19 vector 31 . The ectopic expression of JcARF19 affected calluses formation during transformation procedure (Fig. S3), producing bigger calluses than that of empty vector (EV). After transferring to soil, the ectopic expression of JcARF19 also affected flowering time (Fig. 3a) and other characteristics ( Fig. 3c-f). Under normal growth conditions in a greenhouse, WT Jatropha Jc-MD required around 8 months to produce the first inflorescence whereas 10 primary independent JcARF19 overexpression lines formed their first inflorescences after only 5 months, with a 3 months reduction (Figs 3a and 4a-c). Quantitative reverse transcriptase PCR (qRT-PCR) analysis verified the presence of the transgene and ectopic expression of JcARF19 in transgenic Jatropha plants (Fig. 3b). JcARF19 overexpression lines also had greater branching compared with WT Jc-MD control plants (Figs 3c and 4a). A four-fold increase in seed set from JcARF19 ectopic expression plants was collected within one year of transplanting. Furthermore, a 17.2% increase of single seed weight and 17.1% increase in fruit size were found in JcARF19 ectopic expression plants (Figs 3d-f and 4d-f). The longer cell length (18.4%) and higher cell number (16.5%) in JcARF19 ectopic expression plants explained the increased seed weights and lengths (Figs 3g,h and 4g,h). These results indicated that overexpression of JcARF19 increased length, number and weight of seeds and also oil yield by increasing seed cell number and length in Jatropha.
We further germinated T1 JcARF19-overexpression Jatropha seeds and found that the germination percentage of T1 JcARF19-overexpression Jatropha seeds was higher than these of the wild type control (Fig. S4).
Expression of Auxin responsible genes upon the ectopic expression of JcARF19 in developing seeds. In VIGS assay, we have demonstrated that JcARF19 is essential for auxin signaling transduction on downstream transcription factors JcLBD18 and JcLBD29 in Jatropha leaves (Fig. 1e). To further understand its role in auxin pathway during seed developing, we analyzed the expression of JcLBD18 and JcLBD29 in JcARF19 overexpression seeds and found that both of two LBD genes were higher expressed either at early stage or middle stage compared with those of WT control. We further checked the expression of other downstream genes in auxin pathway such as cell cycle and cell number controlling. We found that the expression level of Expansin1 (JcEXP1) and AUXIN-REGULATED GENE INVOLVED IN ORGAN SIZE (JcARGOS) 45 , together with cell cycle regulators JcCDKA1, JcCYCD2 and JcCYCD5, are induced in JcARF19-ectopic expression Jatropha seeds endosperm, at either early or middle stage of seed development (Fig. 5a), providing molecular explanations for the increased cell number in JcARF19 overexpression Jatropha seeds. We found that the expression of several genes encoding important regulators in cell differentiation and cytoskeletal dynamics have been enhanced including ARGOS 46 , small GTPases auxin-Rho of Plants (ROP), ROP-interactive protein RIC and receptor-like auxin-(Transmembrane Kinase) TMK 47,48 , either in early or middle stages of seed development of J. curcas (Fig. 5b), which also explain the cell size expansion in JcARF19 ectopic expression Jatropha seeds.  JcIAA9 we identified previously, we hypothesized that ARF19 may function via direct protein-protein interaction with IAA9. The secondary structure of JcARF19 and JcIAA9 were predicted to have 4 α-helix and 5 β-sheet folding module (Fig. 6a). We found that JcARF19 interacts with JcIAA9 physically in vitro pull-down assays (Fig. 6b). We used glutathione S-transferase (GST) fused JcIAA9 as bait and JcARF19 as prey. We found that the GST-fused JcIAA9 COOH-terminal protein (JcIAA9-CT) could interact strongly with 6*Histidine-tagged ARF19 COOH-terminal protein (JcARF19-CT), in contrast to much weaker interaction found on the protein pair of J. integerrima (JiARF19 and JcIAA9). This difference was surprising because the ARF19-PB1 and IAA9-PB1 polypeptides of the two species differ by only one amino acid. The S → G mutation which located in interaction interface β5 of ARF19 proteins affects ARF19-IAA9 binding ability. We also provided evidence to show the physical interaction between JcARF19 and JcIAA9 in vivo and the vital role of key amino acid in heterodimer formation by Bimolecular Fluorescence Complementation (BiFC) assays (Fig. 6c). However, although we identified a putative protein-protein interaction pair of IAA9-ARF19 here (Fig. 6), it is still unclear what the significance of this interaction is so far and how IAA9 affects the function of ARF19 in auxin signaling pathway.

Discussion
In herbaceous plants, species with small seeds sometimes have larger seed set than larger-seed species, assuming a limited total amount of energy. This energetic trade-off has been observed in genetic mutants such as APETALA2 [49][50][51] and CURLY LEAF28 52 . In this study we showed that manipulation of the auxin pathway in J.curcas not only increased seed size but also enhanced total seed yield. Previous knowledge on the molecular mechanisms of seed size was mainly limited to model herbaceous plants, particularly Arabidopsis and rice 53 . Our work suggests that manipulation of auxin is an alternative approach to increase seed size in woody plants.
Several lines of evidence support the involvement of JcARF19 in seed size determination. First, JcARF19 was mapped in the major quantitative trait locus (QTL) region and was significantly associated with seed size 11 . Second, by using expression QTL (eQTL) analysis to link variants with functional candidate genes, we provided evidences that seed traits were affected by the genetic interaction of JcARF19 and JcIAA9 11 . Third, the C-terminal   of JcARF19, which is essential for protein-protein interaction among ARF proteins, was identical between ARF19 homologs of bigger seed size in Jatropha. Point mutations on the single-nucleotide polymorphisms (SNPs) of JcARF19 affected their direct physical interactions (Fig. 6). Fourth, overexpression of JcARF19 increased seed size in both Arabidopsis and Jatropha by increasing both cell numbers and cell length. Fifth, ectopic expression of JcARF19 upregulated auxin responsive genes encoding important regulators involving in cell differentiation and cytoskeletal dynamics.
Woody plants such as Jatropha had longer life cycle than herbaceous ones. It took as long as two years for Jatropha to get the first flower blooming under our current lab condition in Beijing North of China, on contrast of 4-5 months in tropical countries such as Singapore. We tried to plant Jatropha in South of China and so far they do not flower yet. For this reason we used a few independent T0 Jatropha plants and we also performed genetic analyses on laboratory model plant Arabidopsis. The increased seeds yield of T3 JcARF19 ectopic Arabidopsis was also consist with the results of overexpression in Jatropha, confirming our claims of improved seed traits with JcARF19 ectopic expression strategy. Furthermore, we found that the germination percentage of T1 JcARF19-overexpression Jatropha seeds was obviously higher than that of the wild type (Fig. S4). Early research had reported that the auxin has close relationship with the seeds germination rate 54 . It indicated that the traits we observed at T0 generation can be inherited into the next generation and our improved seed traits by JcARF19 ectopic expression are reliable as well 55 . Nevertheless, the observed impact on plant architecture, seed size and yield by manipulation of ARF19 need not only more observations on T1 and T2 generations plants under greenhouse condition, more researches under field condition are also necessary for the feasibility of a big scale commercialization of this strategy. Auxin is a multifunctional hormone that regulates pattern formation in plants 56 . The location and timing of auxin accumulation and signal transduction play critical roles in various aspects of plant development 57,58 . In the future, to avoid growth abnormalities in auxin signaling pathway transgenic plants, it is advisable to use a weaker or an organ and developmental specific promoter rather than a stronger promoter like CaMV 35S promoter because the amount of the hormone produced by the transgene and the response should be confined to the target tissue at an appropriate level as did in cotton fiber cell and other success reports 59 . We present a common sharing genetic framework for the control of cell division, differentiation and size for various plant organs, e.g. seed and root. Given the early working model of ARF19 in auxin signaling transduction 53,55,60 , the seed size controlling results in this study can be best integrated as the working model presented in Fig. S5. In this seed size controlling model, auxin activates the transcription of JcARF19 via RETINOBLASTOMA-RELATED (RBR) protein and cytokinin-dependent transcription factor ARABIDOPSIS RESPONSE REGULATOR12 (ARR12) 60 . JcARF19 is involved in promoting cell differentiation and thus cell number increasing in early stage of seed development by regulating the transcription of LBD18 and LBD29 55 . ARGOS gene family is auxin-induced and involved in the regulation of cell number for the duration of organ growth 45 . Ectopic expression of ARGOS prolongs the expression of AINTEGUMENTA (ANT) and cell cycle regulator CycD3; as well as the neoplastic activity of leaf cells 45 . Overexpression of ARGOS genes modifies plant sensitivity to Ethylene, leading to improved drought tolerance in both Arabidopsis and maize 46 . The auxin-(Transmembrane Kinase) TMK sensing and auxin-Rho of Plants (ROP) signaling networks have been demonstrated to control auxin signaling pathway 47,48 . JcARF19 might be also involved in enlarged cell size by TMK Auxin-Sensing and ROP GTPase signaling complex in middle stage of seed development. Considering that orthologs of JcARF19 exist in many other plant species, including castor bean, alfalfa (Medicago sativa), soybean and apple, the manipulation of ARF19 may provide a broad application to increase plant biomass and seed productivity in many other species.This study provides evidence that an auxin signaling integrator ARF19 plays vital roles in determining seed size. ARF19 is conserved in higher plants and involved in auxin pathway signal transduction 55 . Nevertheless, it is necessary to test the ARF19 ectopic expression Jatropha lines under field conditions to get conclusive statements of its commercial viability. Meanwhile besides of plant genome, plant rhizospheric or leaf-residing microbiomes via plant endogenous auxin signalling pathway have been successfully to improve crop yield dramatically including Jatropha 61,62 . ARF19 transcription factors mediated auxin pathway is essential for growth and yield promoting effect by beneficial microbes 63 . Manipulation of the auxin signaling pathway can result in larger seed sizes and improved seed yield in J. curcas, this ARF19-ectopic expression plant may become a more attractive commercial plant.

Materials and Methods
Plant materials and growth condition. Three species of plants were used in this study. For Jatropha, seeds were obtained from the Jatropha curcas (Jc-MD) elite plants which were pre-selected by Drs. Yan Hong and Chenxin Yi 64 . The seeds were germinated on ½ Murashige and Skoog salt medium at 25 °C under a 16 h light/8 h dark photoperiod with a light intensity of 100 μmol·m −2 s −1 . When two or three true leaves were grown 64 , the seedlings were transplanted into pots filled with soil and grown at 25 °C in greenhouse under natural light condition. Plant management, including pesticide spraying, watering and artificial fertilization, was carried out according to normal practice 31 . For Arabidopsis, seeds of ecotype Col-0 were vernalized on ½ Murashige and Skoog salt medium at 4 °C in darkness for about three days and transferred into a growth chamber at 22 °C under16 h dark/8 h light photoperiod. The seedling with four true leaves were transplanted into soil and grown in the same condition. Arabidopsis transformation was performed according to the floral-dipping method previously described 65 . For Nicotiana benthamiana, seeds were sown in soil to germinate. The seedlings were transplanted in separated pots and grown in a greenhouse at 25 °C with 14 h dark/10 h light photoperiod.
RNA extraction and analysis. RNA was isolated and analyzed according to previously described methods 23 . 100 mg samples from different organs were harvested and extracted with plant RNA purification reagent (Invitrogen, Carlsbad, CA, USA). Nanodrop (Thermo Scientific, Wilmington, DE, USA) was used to quantify RNA concentration. M-MLV reverse transcriptase (Promega, Madison, USA) was used for reverse transcription. For quantitative PCR analysis, THUNDERBIRD SYBR qPCR Mix (TOYOBO) was used and run in Bio-Rad CFX96 qPCR machine. Each treatment was repeated with three biological replicates, and with three technical replicates for each biological sample. The Jatropha UBQ transcript was served as an internal control for RNA samples. The primers for target genes are listed in Table S1. Standard deviation was calculated based on the three biological replicates.
Virus induced gene silencing. We used the sTRV method described by Ye et al. 40,41 , using psTRV1 and psTRV2. PCR-based cloning was used to clone partial cDNAs of JcARF19 to psTRV2 to generate psTRV2 derivatives. Another psTRV2 clone with insertion of Jatropha Chlorata 42 (JcCH42) was served as a positive control 34 . psTRV1, psTRV2 and psTRV2 derivatives were electroporated into Agrobacterium strain AGL1. Vacuum agroinfiltration was used to inoculate those Agrobacterium into Jatropha seedling with two or three true leaves. At least 5 Jatropha seedlings were agroinfiltrated with psTRV1 and psTRV2-JcARF19, psTRV2-JcCH42 or psTRV2 vector only accordingly. After infiltration, plants were grown in a growth chamber at 25 °C with a 16 h light ⁄8 h dark photoperiod 40 . Phenotypes of Jatropha plants at 27 days post-infiltration (dpi) with various sTRV constructs were recorded and leaves in same leaf position were picked and leaf width were measured. Values (n = 5) were shown as mean ± SD and statistic analysis with Student T-test. **Indicates P < 0.01, *indicates P < 0.05. Transgenic plant plasmid construction. JcARF19 gene was identified from a database of sequenced cDNA library prepared from Jatropha seeds (detailed sequence information could be found in Supplementary file). The full-length cDNA fragment of JcARF19 was PCR-amplified with primers (Table S1). The PCR fragment was inserted in the sense orientation into suitable sites of pCABMIA1300-3HA vector 34 .

Scanning electronic microscopy (SEM) and light microscopy. For observation of Arabidopsis seeds
with the scanning electron microscope (SEM), collected seeds from WT Col-0 and 35S:JcARF19 overexpression plants respectively were fixed with a tape inside a sample chamber, following freezing in liquid N2. Images were collected using a SEM (JSM-6360LV, JEOL, USA) with an acceleration voltage of 20 kV. For observation of Jatropha seed endosperm with light microscopy, endosperm discs from WT Jc-MD and 35S:JcARF19 overexpression plants respectively were excised from mature Jatropha endosperm (7 WAF, weeks after fertilization) and fixed overnight in 2.5% glutaraldehyde in 0.1 M phosphate buffer, pH 7.2 as described previously 40 . Endosperm discs were rinsed three times in 0.1 M phosphate buffer for 15 min each, and were then post-fixed in 1% (w/v) aqueous OsO 4 for 1 h. Tissues were dehydrated in an ethanol series and embedded in Spurr's resin. Semi-thin sections with thickness of 500 nm were stained in 0.1% toluidine blue and photographed with a Zeiss Axioplan 2 microscope (Carl Zeiss, Germany). Cell size and cell number per disc were analyzed with ImageJ and calculated, followed with statistic analysis with Student T-test. **Indicates P < 0.01. Values are mean ± SEM (n = 10).
Arabidopsis seed size and weight measurement. Mature seeds were harvested from WT Col-0 and 35S:JcARF19 overexpression plants grown under the same conditions. 100 seeds from ten independent transgenic lines and WT Col-0 were weighted and recorded with three technical replicates. Values (n = 10) are given as mean ± SD. A DM5000B microscope (Leica) and ImageJ analysis software were used to measure seed sizes. Values (n = 10) are given as mean ± SD. Statistic analysis with Student T-test. **Indicates P < 0.01.

Fatty acid analysis.
Total lipid was extracted and transmethylated from 100 dry Arabidopsis seeds as described previously 66 . The resulting FAMEs were separated and detected by GCMS-QP2010 Ultra (Shimadzu, Kyoto, Japan). The GC analysis was performed under conditions described before 31 . The data were presented based on three biological replicates and each biological replicate had three technical replicates. Values (n = 3) are given as means ± standard deviation.
Explant material preparation and Jatropha transformation method. Cotyledons were harvested from WT plants Jc-MD sterilized seedlings that were 7-9 days old and were cut into small pieces (5 mm × 5 mm) used as explants. After co-cultivation, shoot regeneration, shoot elongation and rooting, we got the JcARF19 overexpression line. Detailed protocol can be found in Qu et al. 31 .

JcARF19-overexpressing Jatropha agronomic traits measurement and statistical analysis.
Wild-type Jc-MD and JcARF19 transgenic overexpression Jatropha plants were grown in the same condition. Values are mean ± SEM (n = 10). Single seed weight and seed length for each seeds were measured for three of WT plants Jc-MD and T1 JcARF19 overexpression seeds of three lines, JcARF19OE #1, #10 and #13. Values are mean ± SEM (n = 50). Germination percentage were measured for five of WT plants Jc-MD and T1 JcARF19 overexpression seeds of JcARF19 OE #10 and #13. Values are mean ± SEM (n = 5). Student T-test was used for statistical analyses for all agronomic traits. **Indicates P < 0.01, *indicates P < 0.05.
In vitro GST pull-down assay. The C-terminal sequence of JcIAA9 and JcARF19 were amplified by PCR using Phusion High-Fidelity DNA Polymerase (Thermo-Fisher, Finnzymes, Espoo, Finnland) and subcloned into pGEX6P-1 or pET28-SUMO vectors to generate GST fusion or 6*His fusion constructs. Point mutations were performed to generate vector of JiARF19 fusion with 6*His tag by QuikChange Site-Directed Mutagenesis Kit (Stratagene, Agilent, Wilmington, DE, USA). In vitro pull-down assays were performed with 2 μg of GST fusion proteins and 2 μg of His-tagged proteins. GST fusion proteins were incubated in a binding buffer (50 mM Tris-HCl at pH 7.5, 100 mM NaCl, 0.25% Triton X-100, 35 mM β-mercaptoethanol) with 25 μL of glutathione sepharose 4B (GE Healthcare, Uppsala, Sweden) for 3 h at 4 °C and GST beads were washed six times with binding buffer. His-tagged JcARF19-CT and JiARF19-CT proteins were added into GST beads and the mixture was incubated overnight at 4 °C. After washing again with binding buffer six times, pulled-down proteins were separated on 12% SDS-polyacrylamide gel and detected by Western blotting using anti-His or anti-GST antibody as previously described 67-69 . Bimolecular fluorescence complementation (BiFC). BiFC was carried using previously described vectors and methods 67,68 . The C-terminal sequence of JcARF19 and JcIAA9 were cloned in corresponding restrict enzyme sites of BiFC vectors. Point mutations were performed to generate JiARF19 by QuikChange Site-Directed Mutagenesis Kit (Stratagene, Agilent, Wilmington, DE, USA). The resulting cassettes including fusion proteins and constitutive promoters were cloned into pGreen binary vector HY105 and transformed into Agrobacterium. For BiFC experiments, 3-week-old Nicotiana benthamiana leaves were co-infiltrated with Agrobacterium as previously described. Two days after incubation, fluorescence and DAPI staining were analyzed by confocal microscopy 68,70 . The confocal laser scanning microscope technique we used was referred to the Leica SP8 microscope instruction.