# Source paper archive - Source title: Large seed size accompanying GmSDP1 suppression is linked to cellular remodeling and phytohormone regulation - Author(s): Thiya Mukherjee, Sophia Crowley, Maneesh Lingwan, Kevin Chu, Katherine M Murphy, Katia Gutierrez, Anastasiya Klebanovych, Russell Williams, Anuradha Dhingra, Kirk J Czymmek, Ben Mansfeld, Timothy P Durrett, Douglas Allen - Source landing URL: https://www.biorxiv.org/content/10.64898/2026.11.08.765164v1 - PDF URL: https://www.biorxiv.org/content/10.64898/2026.11.08.765164v1.full.pdf - PDF SHA-256: 8a6d97255886f41565b2e1cae9836af9853905c5ef630f4096e11afbec50c20b - Converter: pymupdf4llm 0.0.27 - Limits: Automated PDF text extraction; reading order, equations, tables and formatting may be incomplete or incorrect. Images are omitted, and no OCR was performed. Markup and links are neutralized for safe downloading. Consult the original PDF for the authoritative version. --- [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.08.765164; this version posted November 10, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.08.765164) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY 4.0 International license.](http://creativecommons.org/licenses/by/4.0/) 1 **Large seed size accompanying** _**GmSDP1**_ **suppression is linked to cellular** 2 **remodeling and phytohormone regulation** 3 _*Thiya Mukherjee_ _[1,2]_ _, Sophia Crowley_ _[1,6]_ _, Maneesh Lingwan_ _[1]_ _, Kevin L. Chu_ _[1,5]_ _, Katherine M._ 4 _Murphy_ _[1]_ _, Katia Gutierrez_ _[1]_ _, Anastasiya Klebanovych_ _[1]_ _, Russell Williams_ _[1]_ _, Anuradha Dhingra_ _[3]_ _, Kirk_ 5 _J. Czymmek_ _[1]_ _, *Ben Mansfeld_ _[3]_ _, *Timothy P. Durrett_ _[4]_ _, *Doug K Allen_ _[1]_ 6 7 1 Donald Danforth Plant Science Center, St. Louis, MO 63132 8 2 Department of Natural Sciences, Texas A&M University- San Antonio, San Antonio, TX 78224 9 3 Department of Biology, Washington University in Saint Louis, St. Louis, MO 63130 10 4 Department of Biochemistry and Molecular Biophysics, Kansas State University, Manhattan, 11 KS 66506 12 13 Present Address: 14 5 Bayer Crop Science, Chesterfield, MO 63017 15 6 Ann &Robert H. Lurie Children’s Hospital of Chicago, Chicago, IL 60611 16 17 _*Corresponding_ _author:_ _tmukherjee@tamusa.edu;_ _dallen@danforthcenter.org;_ 18 _tdurrett@ksu.edu; bmansfeld@wustl.edu_ 19 20 The author responsible for distribution of materials integral to the findings presented in this 21 article 22 In accordance with the policy described in the author guidelines 23 (https://academic.oup.com/plphys/pages/author-guidelines) is Thiya Mukherjee 24 25 **Abbreviations** 26 GmSDP1- Glycine max Sugar-Dependent 1 27 RFO- Raffinose Family Oligosaccharide 28 TAG- Triacylglycerol 29 RNAi- RNA interference 30 WT- Wild-type 31 32 **Abstract** 33 Soybean ( _Glycine max_ ) seeds contain approximately 20% oil and 40% protein and are 34 important for food, feed, and biofuel. Improving soybean productivity and seed quality requires 1 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.08.765164; this version posted November 10, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.08.765164) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY 4.0 International license.](http://creativecommons.org/licenses/by/4.0/) 35 optimizing carbon partitioning during seed development. Previously, seed-specific suppression 36 of the lipase-encoding gene _Sugar-Dependent 1_ ( _SDP1_ ) increased seed oil content by 37 approximately 5% while slightly reducing raffinose family oligosaccharides without 38 compromising protein content. Seed-specific _SDP1_ suppression also increased seed size by 39 18–25% without affecting seed number, resulting in greater seed biomass per plant. To 40 determine the developmental basis of the increased seed size, we investigated the cellular and 41 molecular mechanisms underlying this phenotype. Microscopy and biochemical analyses 42 showed that larger seeds were associated with increased seed cell size and enhanced 43 lignocellulosic cell wall accumulation. Temporal biochemical and transcriptomic analyses further 44 revealed coordinated changes in hormone homeostasis, including altered levels of abscisic 45 acid, 12-oxo-phytodienoic acid, and jasmonoyl-isoleucine, together with differential expression 46 of F-box and RING/U-box family genes. The results indicate that the enhanced seed size 47 resulting from _SDP1_ suppression is associated with coordinated changes in cellular 48 development, cell wall accumulation, hormone homeostasis, and regulation of gene expression 49 during seed development. Developmental and metabolic changes associated with the increased 50 seed size phenotype support SDP1 manipulation as a potential strategy for improving soybean 51 seed yield and quality. 52 53 **Introduction** 54 Global food demand is increasing with population growth and poses a major challenge to 55 sustainable agriculture and food security (Singh, 2024). The world population is projected to 56 reach 8.5 billion by 2030, resulting in market pressure for increased food, feed, and renewable 57 bioresources (The Sustainable Development Goal report, 2022). Meeting this need will require 58 developing crops with improved productivity and enhanced nutritional value. Because seeds 59 serve as a major source of nutrients and agricultural products, optimizing seed composition in 60 economically important crop species remains a critical objective for crop improvement and 61 agricultural research (Mukherjee et al., 2025). 62 Soybean ( _Glycine max_ L. Merr.) is an oilseed crop of high economic importance worldwide. As a 63 legume, soybeans provide a major source of plant-derived protein for human consumption and 64 livestock feed (Allen et al., 2009; Allen and Young, 2013; Yang et al., 2019). As an oilseed, 65 soybean contributes substantially to global plant-based oil production used for food and 66 industrial applications (Allen and Young, 2013). During the 2024/2025 marketing year, soybean 67 represented approximately 62% of global oil seed production, and the United States remained 68 the second-largest producer with a crop value of $43.9 billion (Soystats, 2026). Continued 2 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.08.765164; this version posted November 10, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.08.765164) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY 4.0 International license.](http://creativecommons.org/licenses/by/4.0/) 69 improvements in seed quality and yield are necessary to support agricultural sustainability and 70 economic competitiveness. For the 87 million acres in US soy production, an oil increase by 1% 71 at the expense of undervalued carbohydrate would translate to $1to $2 billion added revenue in 72 agriculture industry (Mukherjee et al., 2025). 73 Mature soybean seeds typically contain approximately 20% oil and 40% protein, although these 74 values vary depending on genotype and environmental conditions (Litterer et al., 2006; 75 Kambhampati et al., 2021; Aznar-Moreno et al., 2022; Morley et al., 2023). Improving protein 76 and oil simultaneously remains difficult because their levels in seeds are inversely correlated at 77 maturity (Johnson et al., 1955; Hymowitz et al., 1972; Wilcox and Shibles, 2001; Kambhampati 78 et al., 2020). This negative association has long constrained breeding efforts aimed at 79 maximizing seed value. In addition to oil and protein, less valued carbohydrates (30–35%) 80 account for most of the remaining seed biomass (Middelbos and Fahey, 2008). A substantial 81 portion of these includes raffinose family oligosaccharides (RFOs), which are poorly digested by 82 monogastric animals and are considered anti-nutritional components that reduce feed efficiency 83 (Coon et al., 1990; Suarez et al, 1999; Valentine et al., 2017; Aznar-Moreno et al., 2022). 84 Breeding efforts prioritizing yield improvement often negatively affect seed value by decreasing 85 protein and enhancing carbohydrates (Tamagno et al., 2022). These tradeoffs emphasize the 86 importance of understanding carbon allocation among competing storage pathways during seed 87 development. Carbon partitioning and seed metabolism dictates the balance between lipid, 88 protein, and carbohydrate accumulation and ultimately determines seed composition and value 89 (Kambhampati et al., 2026) . Developing strategies to redirect carbon flux toward high-value 90 seed reserves will provide opportunities to improve soybean nutritional quality while maintaining 91 productivity. 92 Seed development in soybean follows an ordered progression through three distinct phases of 93 embryogenesis: early development (morphogenesis), mid development (cell enlargement and 94 seed filling), and late development leading to desiccation and maturation (Thien Nguyen et al., 95 2016; Mukherjee et al., 2025) Final seed composition is restricted by the resources the seed 96 receives throughout development and the metabolic flux through pathways that convert these 97 resources to seed reserves (Allen and Young, 2013).During the seed-filling phase, maternal 98 carbon inputs (sugars and amino acids, (Allen and Young, 2013; Pegler et al., 2023; Mukherjee 99 et al., 2025)), together with carbon released from starch turnover support the continued 100 synthesis of lipids, proteins, and carbohydrates, including RFOs and cell wall polysaccharides. 101 By late-seed development, the supply of vegetative carbon through the phloem has ceased and 102 developing seeds partially remobilize stored lipids and proteins to sustain carbohydrate or other 3 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.08.765164; this version posted November 10, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.08.765164) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY 4.0 International license.](http://creativecommons.org/licenses/by/4.0/) 103 biosynthetic demands (Chia et al., 2005; Baud et al., 2008; Baud and Lepiniec, 2009; 104 Kambhampati et al., 2021; Mukherjee et al., 2025). 105 One strategy to improve seed oil accumulation is to minimize lipid degradation during late seed 106 development. Seed-specific suppression of the triacylglycerol (TAG) lipase Sugar-Dependent 1 107 (SDP1) through RNA interference (RNAi) has been shown to increase total seed oil content by 108 limiting TAG turnover (Kelly et al., 2013; Kim et al., 2014; Kanai et al., 2019; Azeez et al., 2022; 109 Aznar-Moreno et al., 2022). Prior work demonstrated that suppression of _SDP1_ in soybean 110 seeds increases fatty acid accumulation at maturity; primarily through reduction of RFO 111 production while maintaining total protein content (Aznar-Moreno et al., 2022). Consistent with 112 earlier observations, enhanced oil accumulation resulting from _SDP1_ suppression did not 113 negatively affect seed maturation or germination (Kelly et al., 2013; Kim et al., 2014; Azeez et 114 al., 2022; Aznar-Moreno et al., 2022). 115 In addition to restricting late-stage seed TAG turnover and enhancing seed oil accumulation, 116 suppression of _SDP1_ has consistently been associated with increased seed weight across 117 multiple species. Previous studies reported an increase in seed weight ranging from 13–30% in 118 _Camelina sativa_ (Dhankher, 2021), 16% in the _Arabidopsis thaliana sdp1_ mutant (Kim et al., 119 2014), 12–16% in _Physaria fendleri_ (Azeez et al., 2022), and 4–14% in _Glycine max_ (Kanai et 120 al., 2019). Our prior work demonstrated a 14–23% increase in seed weight in soybean RNAi 121 _SDP1_ -suppressed lines, which, together with elevated fatty acid accumulation, resulted in an 122 overall 20% increase in oil yield relative to wild type (Aznar-Moreno et al., 2022). Despite these 123 consistent phenotypic gains, the biochemical and cellular mechanisms underlying enhanced 124 seed weight following _SDP1_ suppression remain unknown. 125 In this study, we investigated the developmental basis of the increased seed weight phenotype 126 resulting from _SDP1_ suppression. Increased seed weight became evident during mid seed filling 127 and persisted through maturation. This increase was associated with greater seed size, 128 enhanced cell expansion, and increased cell wall accumulation, without changes in seed 129 number. Developmental transcriptomic and biochemical analyses further implicated changes in 130 phytohormone signaling and the ubiquitin–proteasome system during seed development. 131 Together, these findings provide insight into the developmental and metabolic changes 132 associated with suppression of seed lipid turnover and support _SDP1_ as a potential target for 133 increasing seed oil content and yield while maintaining seed quality in soybean and other 134 species. 135 136 **Material and Methods** 4 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.08.765164; this version posted November 10, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.08.765164) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY 4.0 International license.](http://creativecommons.org/licenses/by/4.0/) 137 138 _Plant growth conditions and sample preparation_ 139 Soybean seeds ( _Glycine max_ cv Williams 82) were planted in one-gallon pots with Berger BM7 140 35% bark soil (10121500; Hummert, Earth City, MO, USA) and grown in a greenhouse 141 supplemented with 14h day: 10h night photoperiod with 25 ° C/23 ° C Day/night temperature and 142 50% relative humidity. Seeds were sampled across development as described before (Aznar 143 Moreno et al., 2022). Briefly, pods were harvested at three reproductive stages (R5.5, R6, and 144 R7) and at maturity (R8). Seeds were immediately harvested, flash-frozen in liquid nitrogen, and 145 stored at –80 °C until further use. Prior to analysis, seed coats were removed, and cotyledons 146 were ground using an in-house stainless-steel hammer-crushing pestle and mortar. The ground, 147 frozen tissue was aliquoted, with other fractions kept fresh or lyophilized for subsequent 148 analyses. 149 150 _Phenotypic and agronomic measurements_ 151 Four to six plants for wild-type (WT) and each of the threeT4 homozygous transgenic 152 genotypes; A ( _sdp1-1 RNAi A_ ), B ( _sdp1-3, RNAi D_ ) and C ( _sdp1-3, RNAi C_ ) (Aznar-Moreno et 153 al., 2022) were grown in one-gallon pots, randomized within the greenhouse. Phenotypic traits, 154 including node and pod number were recorded when approximately 95% of the pods had 155 reached maturity, as indicated by dry yellow color and the absence of new pod growth. Seeds 156 were hand-harvested, and a seed count per plant was determined. Average seed weight was 157 determined by weighing 100 seeds individually from each of four to six plants per genotype. 158 Average individual seed area was obtained from ten seeds each from four to six plants per 159 genotype. Red-green-blue (RGB) images were captured using a standard digital camera (Nikon 160 Coolpix L830) connected to Raspberry Pi computer for capturing, labeling, and transferring the 161 images, as previously described (Duenwald et al., 2025). Bright, indirect lighting was used, and 162 a color card was used for color correction and size standardization. Images were processed 163 using open-source software PlantCV (Gehan et al., 2017). First, images were color corrected, 164 and then seeds were masked (segmented) from the background. Individual seeds were 165 measured using the morphology package for seed size and shape **(Supplementary figure** 166 **S1D)** . To capture an image of all seeds for visual comparison **(Supplementary figure S1C)** the 167 same protocol was used, except a Nikon D7200 was used instead of a Coolpix. 168 169 _Microscopy analysis_ 170 5 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.08.765164; this version posted November 10, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.08.765164) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY 4.0 International license.](http://creativecommons.org/licenses/by/4.0/) 171 _Sample Preparation_ 172 Three to five mature soybean seeds were fixed in 4% paraformaldehyde (15710; Electron 173 Microscopy Sciences, Hatfield, PA, USA and Leica Microsystems, Wetzlar, Germany) in 1X PBS 174 buffer (phosphate buffered saline pH 7.4, 137 mM NaCl, 2.7 mM KCl, 10 mM Na 2 HPO 4, 1.8 mM 175 KH 2 PO 4 ), (19342-10; Electron Microscopy Sciences, Hatfield, PA, USA) for 48 hours at 4°C. 176 After fixation, samples were thoroughly rinsed in 1XPBS buffer, briefly dried and embedded in 177 O.C.T. compound cryo-embedding medium (62550-0; Electron Microscopy Sciences, Hatfield, 178 PA, USA). Each sample was sectioned using Leica CryoJane CM1950 cryostat (Leica 179 Biosystems, Buffalo Grove, IL, USA) to expose the tissue structure. This step was done to avoid 180 hand-free razorblade sectioning and ensure the exposed tissue surface was even and flat. 181 Sectioned seeds were incubated in 25 μ M Calcofluor White fluorescent dye (29067; Biotium, 182 Fremont, CA, USA) to stain cell walls. After rinsing, samples were mounted in distilled water on 183 the No.1.5H glass-bottom WillCo-dish® (HBST-5040; Willco Wells, Amsterdam, The 184 Netherlands) designed for inverted microscope confocal imaging. 185 186 _Imaging_ 187 Each biological replicate was imaged using the same acquisition conditions as described below. 188 4-10 fields of view (FOV) were captured per replicate using Leica TCS SP8-X laser confocal 189 microscope (Leica Microsystems, Wetzlar, Germany) equipped with 405-nm UV laser, HC PL 190 APO CS 20X/NA 0.7 dry objective, pinhole 1 AU, and PMT detector. Each FOV was acquired as 191 a 12-bit, either 2D 1024X1024 pixels image with the pixel size of 0.568 X 0.568 μ m, or in some 192 cases as a 5- μ m z-stack with a z-step size of 2 μ m and a voxel size of 0.568 X 0.568 X 2 μ m. 193 To detect a cell wall stained with the Calcofluor White fluorescent dye (Excitation /Emission 194 360/430 nm), a UV laser (405 nm) was used with Acousto-Optical Beam Splitter prism set to the 195 emission detection of 415-470 nm. 196 197 _Image processing and data analysis_ 198 Cell area and cell count were analyzed using the Fiji open-source software (Schindelin et al., 199 2012). All images were processed as 2-dimentional (2D) or Maximum Intensity Projections 200 (MIPs) of 1024 x 1024-pixel FOV using the Trainable Weka Segmentation plugin (Arganda 201 Carreras et al., 2017). This approach enabled segmentation and selection of the inner regions 202 delineated by cell wall staining. A binary mask was then generated, followed by the application 203 of “Fill holes” and “Watershed” filters. Subsequently, the Analyze Particles plugin was used to 204 calculate the total area of each cell in μ m [2] and determine the total number of cells per 6 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.08.765164; this version posted November 10, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.08.765164) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY 4.0 International license.](http://creativecommons.org/licenses/by/4.0/) 205 1024x1024 pixel FOV. Cells that were incomplete or in contact with the frame borders were 206 excluded from the analysis. Density distribution curves of cell area were plotted using OriginPro 207 software (2024; OriginLab, Northampton, Massachusetts, USA). 208 209 _Fresh weight and dry weight measurements of developing seeds_ 210 Seeds were harvested from four individual plants per genotype for each of the three _sdp1_ 211 suppressed lines and the WT control at three developmental stages: R5.5, R6, and R7. Fresh 212 weight (FW) and dry weight (DW) were determined using an analytical balance (AG204 Delta 213 Range, Mettler Toledo, Columbus, Ohio, USA). For FW measurement, the weight of individual 2 214 mL microcentrifuge tubes was first recorded. A single seed was then placed in each tube, and 215 the combined weight was measured. Seed FW was calculated by subtracting the pre-recorded 216 empty tube weight from the total weight. Immediately following FW determination, seeds were 217 placed in a drying oven set at 60°C. Dry weight was measured at 3-day intervals over a 20-day 218 period, until no moisture remained as indicated by a stable DW value. 219 220 _Quantification of metabolites across development_ 221 Total lipids, starch, and oligosaccharides were quantified from three to four biological replicates 222 per reproductive stage. To minimize the effects of canopy position, seeds were collected from 223 multiple positions within the canopy of each plant replicate, pooled, and ground together. For 224 each plant replicate, two to four technical replicates were quantified. 225 226 _Fatty Acid Methyl Ester (FAME) analysis for lipid content_ 227 10-15mg of lyophilized tissue was weighed into 8ml glass tubes and analyzed using 1 mg of 228 C17:0 TAG internal standard (T2151; Millipore Sigma, Saint Louis, MO, USA). To each sample, 229 2ml of freshly prepared 5% (v/v) sulfuric acid: methanol was added. To prevent oxidation of 230 lipids, 25ul of 0.2% butylated hydroxytoluene (101162; MP Biomedicals, Irvine, CA, USA) in 231 methanol was added. Reactions were incubated on a heat block at 110°C for three hours with 232 intermittent vortexing, initially after 20 mins and then after each hour interval. After three hours, 233 tubes were cooled at room temperature, and the reaction was quenched using 3ml of 0.9% 234 sodium chloride. FAMEs were extracted using 1ml hexane and quantified by gas 235 chromatography-flame ionization detection (GC-FID) using a DB23 column (30m,0.25mm 236 i.d.,0.25 um film; J and W Scientific, Folsom, CA, USA). 20:1 split mode was used in the GC. 237 Flame detector temperature was set to 250°C with the oven temperature ramped from 170°C to 7 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.08.765164; this version posted November 10, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.08.765164) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY 4.0 International license.](http://creativecommons.org/licenses/by/4.0/) 238 250°C at a rate of 10°C /min with a program hold time of 6 min respectively. Quantification was 239 performed using peak area comparison with internal standard. 240 241 _Starch quantification_ 242 Starch extraction and quantification was performed using the Megazyme starch assay kit 243 (Megazyme International, Ireland), AOAC Official Method 996.11 (Mccleary et al., 1997; 244 McCleary et al., 2019) with assay volumes adjusted for 96-well plate use as described 245 previously (Kambhampati et al., 2021). Briefly, 15-25mg of lyophilized tissue was washed twice 246 with 80% (v/v) ethanol at 85°C, incubated with DMSO at 110°C for 10 min, digested with α 247 amylase at 110°C for 12 min with intermittent vortexing, and digested with amyloglucosidase at 248 50°C for 1 hr. Samples were then centrifuged and the aliquots of the supernatant were 249 incubated with the GOPOD reagent mixture at 50°C for 20 mins before absorbance at 510 nm 250 was measured. Starch concentrations were determined using standard curves of serially diluted 251 regular maize starch (700004337/K-RSTCL; Megazyme International, Ireland) standard 252 extractions performed alongside biological samples. 253 254 _Raffinose Family Oligosaccharide (RFO) quantification_ 255 Oligosaccharide extraction was performed as described previously (Dierking and Bilyeu, 2008; 256 Aznar-Moreno et al., 2022). Briefly, sugars were extracted from approximately 15–25 mg of 257 lyophilized tissue using 10 µg/ml ribitol (A5502; Millipore Sigma, Saint Louis, MO, USA) as an 258 internal standard and 1 ml of 50% ethanol. For late developmental and mature seed samples, 259 additional pulverization was performed using a ball mill at 30 Hz for 5 mins or until fully ground. 260 Samples were then heated at 70 °C for 30 mins with intermittent shaking and vortexing. 261 Following heat treatment, samples were centrifuged at 15,000 × g for 15 mins to pellet debris. 262 The supernatant was transferred to a new tube and completely dried using a centrivap. Sugars 263 were reconstituted in 50% methanol and 3 ul sample was separated chromatographically using 264 an Agilent InfinityLab Poroshell 120 HILIC-Z (2.7-µm particle size, 100 × 2.1 mm) column 265 (685775-924; Agilent, Santa Clara, CA, USA). LC-MS/MS data were acquired using a Thermo 266 TSQ Altis mass spectrometer coupled to a Thermo Dionex UltiMate 3000 UHPLC system 267 (Thermo Fisher Scientific, Waltham, MA, USA) and a CTC Analytics HTC-PAL autosampler 268 (CTC Analytics, Zwingen, Switzerland). The LC instrument was operated with a flow of 0.2 269 ml/min and sugars were eluted using mobile phase buffers 10mM Ammonium formate in water, 270 pH 6.8 (buffer A) and 10mM Ammonium formate (1:9) water: ACN (buffer B) with an oven 271 temperature of 40°C, adapted from prior methods (Czajka et al., 2020). The gradient conditions 8 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.08.765164; this version posted November 10, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.08.765164) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY 4.0 International license.](http://creativecommons.org/licenses/by/4.0/) 272 were as follows (all percentages in v/v): Start (0 min) at 95% B, 3.5 min at 60% B, 6 min at 30% 273 B, 7.5 min at 30% B, 9 min at 95% B, 14.70 min at 95% B. The total run time was 15 min. The 274 TSQ Altis mass spectrometer was operated with a Heated Electrospray ionization (H-ESI) ion 275 source type in negative ionization mode with a static ion spray voltage of -3700 V and a transfer 276 tube temperature of 325°C, sheath/aux/sweep gas settings of 55/7/1 (Arb), and a vaporizer 277 temperature of 350 °C. Direct infusion of standards was performed to determine the diagnostic 278 fragments and collision energies for each sugar to generate the selected reaction monitoring 279 (SRM) approach of targeted sugar determination **(Supplementary materials and methods,** 280 **Table 1)** . All peaks were integrated using the latest version of Skyline software (22.2-64-bit 281 version; MacCossLab Software, University of Washington, USA)(Pino et al.2020). Quantification 282 was done using a calibration curve. 283 284 _Transcriptomic analysis of developing soybean seeds_ 285 Transcriptomic analysis was performed with seeds from three to four plants per each of _SDP1_ 286 RNAi suppressed line B and WT across early (R5.5), mid (R6) and late (R7) developmental 287 stages. Seeds were ground using a GenoGrinder (Cole-Parmer HG-600). Approximately 50 mg 288 of ground tissue was used for total RNA extraction employing the MagMax Plant RNA Isolation 289 Kit in the KingFisher Flex (ThermoFisher Scientific, Waltham, MA, USA) automated system. 290 RNA was quantified using a Nanodrop spectrophotometer and an Agilent TapeStation 4200. 291 Samples with an RNA Integrity Number (RIN) of 8 or higher were deemed of high quality. Library 292 preparation and sequencing were conducted by Novogene using the Illumina NovaSeq X Plus. 293 Briefly, mRNA libraries were prepared with poly(A) enrichment and sequenced using a paired 294 end 150 bp (PE150) approach. The resulting demultiplexed raw reads were provided in FASTQ 295 format. These were then pseudo-aligned to the WT transcriptome using Salmon v1.10.2 (Patro 296 et al., 2017) default settings. Read count data were imported into R using the tximport package 297 (Soneson et al., 2015). Samples were then defined by their developmental stage (R5.5, R6, R7) 298 and genotype. Differential expression on the integer count values was performed using DESeq2 299 (Love et al., 2014). Genes with a sum of less than 10 reads or a mean of 20 reads across all 300 samples were excluded from analysis. Principal component analyses (PCA) were performed 301 using the variance stabilized transformed read counts DESeq2. Differential expression analysis 302 was performed using “apeglm”(Zhu et al., 2019) as the Log Fold Change Shrinkage method. 303 Genes were defined as being significantly differentially expressed if they had an adjusted p 304 value of less than 0.05. Normalized counts were plotted using ggplot and tidyverse (Wickham et 305 al., 2019) functions in R. Gene Ontology (GO) term enrichment analysis was performed using 9 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.08.765164; this version posted November 10, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.08.765164) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY 4.0 International license.](http://creativecommons.org/licenses/by/4.0/) 306 the TopGO R package with the entire set of expressed genes set as background (Alexa et al., 307 2006). Terms were considered enriched if they passed a p-value of 0.05 on the Fisher test with 308 the “weight01” algorithm and a minimum node size of 100 genes. 309 310 _Phosphoenolpyruvate carboxykinase_ ( _PEPCK) enzyme activity assay_ 311 PEPCK activity assay was performed with developing seeds from four plants, each of WT and 312 three transgenic genotypes. Enzyme extraction and assay procedures were adapted from 313 (Osorio et al., 2014) and (Sharwood et al., 2016). Approximately 100 mg of fresh ground tissue 314 was homogenized in 1 ml of cold extraction buffer containing 50 mM HEPES (pH 7.0), 5 mM 315 dithiothreitol (DTT; freshly prepared), 1% (w/v) polyvinylpolypyrrolidone (PVPP), 2 mM 316 Ethylenediaminetetraacetic acid (EDTA), 2 mM Manganese(II) chloride (MnCl ₂), and 0.05% 317 Triton X-100. Samples were vortexed for 30 s, homogenized for 2 min using a bead beater, and 318 incubated on ice for 10 min. The homogenate was centrifuged at 13,000 × g for 30 s at 4 °C, 319 and the clarified supernatant was used for enzyme assays. PEPCK activity was measured in a 320 flat-bottom 96-well microplate. Each reaction contained 10 µL of protein extract and 150 µL of 321 assay buffer comprising 50 mM HEPES (pH 7.0), 4% (w/v) β -mercaptoethanol, 100 mM KCl, 90 322 mM NaHCO ₃, 1 mM ADP, 2 mM MnCl ₂, 0.14 mM NADH, 15 mM phosphoenolpyruvate (PEP), 323 and 6 units of malate dehydrogenase. Absorbance at 340 nm was monitored at 25 °C using a 324 Tecan Microplate Reader M 200PRO, with readings taken every 5–10 s for up to 5 min with 325 intermittent shaking. Control reactions lacking PEP were included to correct for background 326 NADH oxidation. PEPCK activity was calculated from the linear decrease in absorbance at 340 327 nm using the molar extinction coefficient of NADH and expressed as enzyme units (U), defined 328 as µmol NADH oxidized per minute. Protein concentration was determined using the Bradford 329 assay with bovine serum albumin standards, and enzyme activity was normalized accordingly. 330 331 _Hormone analysis_ 332 333 _Phytohormone extraction_ 334 Phytohormones including cis-zeatin (cZ), trans-zeatin (tZ), DL-dihyrdrozeatin (DHZ), trans 335 zeatin riboside (tZR), abscisic acid (ABA), indole-3-acetic acid (IAA), IAA-aspartic acid (IAA 336 Asp), salicylic acid (SA), 12-oxo-phytodienoic acid (OPDA), jasmonoyl-isoleucine (JA-Ile), and 337 jasmonic acid (JA) were extracted from seeds collected at mid (R6) and late (R7) reproductive 338 stages. Samples were harvested from three to four plants per transgenic genotype and 339 corresponding WT controls. Hormone extraction was performed in triplicate at a tissue 10 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.08.765164; this version posted November 10, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.08.765164) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY 4.0 International license.](http://creativecommons.org/licenses/by/4.0/) 340 concentration of 50 mg ml ⁻ ¹ using ice-cold 1:1 methanol:acetonitrile (MeOH:ACN). For 341 extraction, a stock solvent of 1:1 MeOH:ACN was prepared and supplemented with a mixed 342 stable isotope-labeled internal standard (IS) solution containing 0.4 μ M d5-tZ and d5-tZR; 1.0 343 μ M d4-SA, d6-ABA, d5-JA, d5-IAA, ¹³C ₆ ¹ � N-IAA-Asp, and d5-dinor-OPDA; and 10.0 μ M 344 ¹³C ₆ ¹ � N-JA-Ile. The total extraction solvent volume was calculated as (number of samples + 2) 345 × (975 μ l MeOH:ACN + 25 μ l IS mix). For the experiment described here, 17.55 ml MeOH:ACN 346 and 450 μ l IS mix were prepared. Each sample received 1 ml of extraction solvent containing 347 internal standards. Samples were homogenized using a TissueLyser II (Qiagen) for 5 min at 15 348 Hz, followed by centrifugation at 16,000 × g for 5 min at 4 °C. Supernatants were transferred to 349 fresh 2 ml tubes, and the remaining pellets were re-extracted twice with 800 μ l ice-cold 1:1 350 MeOH:ACN (without IS). All supernatants were pooled and dried using a vacuum centrifuge. 351 Dried extracts were reconstituted in 100 μ l of 30% methanol, centrifuged to remove particulates, 352 and filtered through a 0.8 μ m polyethersulfone spin filter (Sartorius, Stonehouse, UK). The 353 filtrates were transferred to HPLC vials for subsequent LC–MS/MS analysis. Details of 354 chemicals used for the assay, including reference standards, are included as supplemental 355 material. 356 357 _LC-MS/MS Analysis_ 358 Phytohormones including cis-/trans-zeatin (c/tZ), DHZ, tZR, SA, ABA, IAA, IAA-Asp, JA, JA-Ile, 359 and OPDA were quantified using a targeted selected reaction monitoring (SRM) isotope dilution 360 based LC–MS/MS method **(Supplementary materials and methods, Table 2)** . Quantitative 361 analysis was performed using an Ultimate 3000 UPLC system coupled to a TSQ Altis triple 362 quadrupole mass spectrometer equipped with a heated electrospray ionization (HESI) source 363 (Thermo Fisher Scientific, Waltham, MA, USA). A 5 μ l aliquot of each reconstituted sample was 364 injected onto a 3.0 × 100 mm, 1.8 μ m ZORBAX Eclipse XDB-C18 column (Agilent Technologies, 365 Santa Clara, CA, USA). Chromatographic separation was achieved over 22 min using a binary 366 gradient consisting of mobile phase A (0.1% acetic acid in water) and mobile phase B (0.1% 367 acetic acid in 3:1 acetonitrile:methanol). The gradient program was as follows: 10% B from 0 to 368 2.0 min, increased to 25% B at 7.0 min, ramped to 65% B at 15.0 min, increased to 95% B at 369 18.0 min, and held at 95% B until 24.0 min. The flow rate was maintained at 0.3 ml min ⁻ ¹.Mass 370 spectrometry source parameters were set as follows: sheath gas, 60 arbitrary units; auxiliary 371 gas, 5; sweep gas, 2; ion transfer tube temperature, 325 °C; vaporizer temperature, 350 °C; and 372 spray voltage of +3000 V in positive ion mode and −3300 V in negative ion mode. MS conditions 373 included a collision-induced dissociation (CID) gas pressure of 1.5 mTorr, with Q1 and Q3 11 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.08.765164; this version posted November 10, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.08.765164) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY 4.0 International license.](http://creativecommons.org/licenses/by/4.0/) 374 resolutions set to 0.7 full width at half maximum (FWHM) and a cycle time of 0.6 s. Analytes and 375 corresponding internal standards were monitored using optimized SRM transitions in a polarity 376 switching acquisition method, with cytokinins and auxins detected in positive ion mode and 377 other analytes detected in negative ion mode. Data acquisition was performed using Xcalibur 378 software (version 4.2.47; Thermo Fisher Scientific), and quantitative analysis was conducted 379 using QuanBrowser (Thermo Fisher Scientific). Phytohormone concentrations were determined 380 by comparing analyte-to-internal standard peak area ratios to calibration curves generated using 381 matched standards and internal standards. Internal standard concentrations in calibration 382 solutions were 2.5 μ M for ¹³C ₆ ¹ � N-JA-Ile, 0.10 μ M for d5-tZ and d5-tZR, and 0.25 μ M for all 383 other internal standards. Mixed calibration standards were prepared over a range of 0.5 fmol to 384 250 pmol on-column, and the final calibration range for each analyte was adjusted based on its 385 abundance in the samples. 386 387 _Lignocellulosic wall isolation_ 388 Lignocellulosic wall content was extracted from plants each of WT and the three RNAi lipase 389 suppressed genotypes A, B and C. Alcohol Insoluble Residue (AIR), as described earlier (Caffall 390 et al., 2009; Foster et al., 2010; Pettolino et al., 2012; Kataya et al., 2025) was used for the 391 extraction of lignocellulosic wall material. Briefly, ~100 mg of powdered lyophilized seed 392 samples was sequentially extracted with 80% ethanol (v/v), 1:1 (v/v) chloroform: methanol, 393 100% methanol and 100% acetone. The pellet was resuspended in each solvent and incubated 394 for 10 mins at room temperature on a rotating mixer. Between each extraction, samples were 395 pelleted by centrifugation at 10,000 g for 10 mins using a table-top microcentrifuge. After solvent 396 extraction, the pellet was washed twice with water and subsequently used for starch removal. 397 Mature soybean seeds contain minimal starch (Kambhampati et al., 2021). However, to prevent 398 interference from trace amount of starch in the sample, alpha-amylase starch digestion was 399 performed. The resulting pellet was washed with sterile water and dried overnight at 40°C using 400 a centrifugal vaccum concentrator. The dried AIR material represented the lignocellulosic (cell 401 wall) enriched fraction. 402 403 _Statistical Analysis_ 404 To test for a significant difference of the means of different parameters between the WT and the 405 _SDP1_ -suppressed genotypes, two-tailed Student’s t-test of unequal variance was performed. 406 Means were considered significantly different at p ≤ 0.05. Pearson correlation analysis was 12 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.08.765164; this version posted November 10, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.08.765164) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY 4.0 International license.](http://creativecommons.org/licenses/by/4.0/) 407 performed, and p-value was determined using the corresponding t-distribution of the Pearson 408 correlation coefficient (PCC) to determine significance of the correlation analysis. 409 410 **Results** 411 412 _Seed-specific suppression of SDP1 enhances seed weight and seed area_ 413 Previously, we found that suppressing the _SDP1_ gene increases seed weight (Aznar-Moreno et 414 al., 2022), consistent with earlier observations in Arabidopsis (Kim et al., 2014) and soybean 415 (Kanai et al., 2019). In this study, we conducted a detailed greenhouse experiment to 416 characterize the increase in seed size during development. Four to six plants were grown from 417 three independent homozygous _SDP1_ suppressed lines A, B, and C, with the WT controls. As 418 vegetative growth influences the canopy and thus the productivity from a plant; node number, 419 pod number, total seeds were measured from each of the individual T4 plants per line and WT 420 controls. No significant difference was obtained in node, pod, or seed number relative to the 421 control plants **(Figure 1A-C).** Next, the 100-seed weight was assessed in each plant. The 422 results indicated that seeds from the lipase suppressed transgenic soybean genotypes had 423 significantly higher mass than the WT (p ≤ 0.001), resulting in an 18-25% increased seed weight 424 **(Figure 1D and Supplementary figure S1C)** . Pod and node number were modestly correlated 425 (r=0.52, p=0.06; **Supplementary figure S1A)** . Seed and pod number were strongly related 426 (r=0.88, p ≤ 0.001; **Supplementary figure S1B** ); however, seed weight did not correlate with 427 seed number (r=-0.25, p=0.647) **(Figure 1E)** . Seed weight was also compared to seed area. 428 Seed area increased by 13-21% (p ≤ 0.001), compared to the control **(Figure 1F and** 429 **Supplementary figure S1D)** . The results indicated that _SDP1_ suppression in soybean seeds 430 did not affect plant architecture and branching apart from the change in seed size. As the 431 change in seed weight was not associated with seed number, total seed biomass was increased 432 on a per plant basis. 433 434 _Enhanced seed size is due to increased cell size with no significant change in cell number_ 435 We sought to understand if the significant increase in seed size was attributable to changes in 436 cell number or morphology. We examined seed cross sections for each of the three RNAi 437 genotypes and compared them to the WT. Cells of all _SDP1-_ suppressed mature seeds showed 438 a significant increase in cell area relative to the WT, with cells of lines A, B and C showing 11, 12 13 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.08.765164; this version posted November 10, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.08.765164) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY 4.0 International license.](http://creativecommons.org/licenses/by/4.0/) 439 and 4% increased area, respectively (p ≤ 0.001; **Figure 2A-E)** . In contrast, no significant 440 change in cell number was observed **(Figure 2F)** . 441 442 _RNAi sdp1 seed weight is increased by mid-development_ 443 The _sdp1_ mutation enhanced soybean seed size at maturity. WT and SDP1 altered seeds 444 matured at the same time; however, it was unclear when larger seeds emerged during seed 445 filling. Fresh weight (FW) and dry weight (DW) were analyzed from developing WT and lipase 446 altered genotypes, harvesting seeds together. During the early stages of filling (R5.5), lines A, B 447 and C did not differ relative to WT **(Figure 3)** ; however, by mid development (R6) and through 448 later stages (R7), differences could be observed. A, B, and C were 33%, 48% and 38% greater 449 in FW and 43%, 61% and 45% increase in DW compared to WT at R6. At R7, the same lines 450 had 33%, 33% and 51% greater FW and 35%, 32% and 57% higher DW respectively. 451 452 _Developmental composition of soybean seed biomass indicated elevated oil levels_ 453 The contributions to altered single seed weight **(Figure 1 D)** were further evaluated by 454 assessing seed composition over development. T4 seeds were collected at R5.5, R6, R7 455 development stages, and at maturity (R8). Biomass was analyzed from three to four plants, with 456 two to four technical replicates per plant, for each transgenic line and WT. Biomass analysis on 457 a per seed basis was obtained from percent biomass storage reserve content **(Supplementary** 458 **table S1)** and average individual seed weight **(Figure 1 D and Figure 3).** 459 FAME analysis indicated that at R5.5, total lipid content (mg/seed) was reduced in all _SDP1_ 460 suppressed lines compared to WT. This difference was significant for B (p ≤ 0.01) and C (p ≤ 461 0.001). From R6 through R8, oil content was elevated in all three transgenic lines compared to 462 WT (p ≤ 0.001) **(Table 1).** The fatty acid profile of the transgenics changed over development, 463 specifically in the levels of palmitic (C16:0) and linoleic (C18:2) acid. Stearic acid (C18:0) was 464 increased at R5.5 for B (p ≤ 0.01) and C (p ≤ 0.001) and the difference decreased until maturity. 465 Oleic acid (C18:1) was elevated for A (p ≤ 0.05, p ≤ 0.01 and p ≤ 0.001). Linolenic acid (C18:3) 466 on the other hand significantly reduced for selected transgenic genotype (p ≤ 0.05 and p ≤ 467 0.001) across development respectively **(Supplementary table S2)** . 468 Starch content in soybeans is transient and peaks at R6 and declines through R7 and R8 469 (Kambhampati et al., 2021). This trend was observed in both WT and _SDP1_ -suppressed seeds 470 **(Table 1);** however, at R5.5, transgenic seeds had significantly lower starch than WT (p ≤ 471 0.001). The starch levels in transgenics increased at R6, matched WT levels by R7 in most lines 472 **(Table 1)** and were similar at maturity. 14 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.08.765164; this version posted November 10, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.08.765164) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY 4.0 International license.](http://creativecommons.org/licenses/by/4.0/) 473 Mature soybeans contain appreciable quantities of RFOs (Kambhampati et al., 2021) including 474 raffinose and stachyose that cannot be digested by monogastric animals and detract from the 475 overall value of soybean. Prior studies indicated that _GmSDP1_ suppression reduces the level of 476 RFOs, particularly stachyose, on a percent biomass basis at maturity (Aznar-Moreno et al., 477 2022). Here, the RFOs in _SDP1_ -suppressed soybean genotypes were compared to the WT over 478 development. Raffinose and stachyose were undetected in R5.5 seeds **(Table 1)** . Mid 479 development (R6) seeds had elevated raffinose in B and C (p ≤ 0.01) **(Table 1)** without changes 480 in stachyose. R7 seeds were elevated in raffinose and stachyose levels in A, B, and C seeds (p 481 ≤ 0.001) **(Table 1)** . Consistent with prior observations at maturity (Aznar-Moreno et al., 2022), 482 our results indicated a decline in raffinose and stachyose levels on a percent biomass basis in 483 the T4 mature seeds of _SDP1_ -suppressed lines **(Supplementary table S1)** ; however, due to 484 enhanced seed size, total raffinose content (mg/seed) of mature _SDP1_ -suppressed seeds was 485 higher than WT controls, with the difference being statistically significant for A and B (p ≤ 0.01) 486 **(Table 1)** . Stachyose levels in lines B and C (p ≤ 0.001) were reduced relative to the WT, except 487 for A, which showed an increase in stachyose content (p ≤ 0.001) **(Table 1)** . The metabolic 488 consequences of these changes were evident in the mature seeds. The absolute amount of oil 489 produced was increased by 27% to 34% per seed, because of the larger seed size and the 490 proportional increase in oil ( **Supplementary table S3)** . Total raffinose was elevated 1% to 15% 491 ( **Supplementary table S3)** ; however **,** starch and stachyose changes were genotype-specific 492 ( **Supplementary table S3).** 493 494 _Increased number of differentially expressed genes during late development_ 495 To better understand the mechanism by which silencing _SDP1_ enhances cell size **(Figure 2A-E)** 496 and increases seed size **(Figure 1D and 1F),** we performed RNAseq across seed development. 497 Principal Component Analysis (PCA) of the transcriptomic data from the WT and _SDP1_ 498 suppressed RNAi samples from line B showed consistent clustering between developmental 499 timepoints, with minimal differences between WT and RNAi samples. PCA indicated that the first 500 principal component (PC1) accounted for 88% of variance separating samples from early (R5.5 501 and R6) and late (R7) seed development. The second principal component (PC2) with 5% 502 variance reflected smaller changes associated with changes between early (R5.5) and mid 503 stage (R6) development **(Figure 4A)** . 504 When comparing gene expression between RNAi lines and the WT control, R7 stage contained 505 the highest number of differentially expressed genes (DEGs) with 283 genes and 572 genes 506 being up and down regulated respectively. In contrast, very few DEGs were identified in R5.5 15 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.08.765164; this version posted November 10, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.08.765164) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY 4.0 International license.](http://creativecommons.org/licenses/by/4.0/) 507 and R6 **(Figure 4B and Supplemental file S1)** . As the R7 stage showed the highest number of 508 DEGs, we performed a Gene Ontology (GO) enrichment analysis at this stage. Downregulated 509 genes were significantly enriched for various metabolic processes including hormone 510 associated metabolic processes, acetyl-CoA metabolic processes, oligosaccharide metabolic 511 processes, monosaccharide biosynthetic process, gluconeogenesis, glycolytic processes, 512 proteasome core complex assembly processes and metabolic processes associated with cell 513 division and proliferation **(Figure 4C).** Selected GO term enrichment of upregulated genes was 514 associated with cell wall organization including glucan metabolic processes, beta-glucan 515 biosynthetic processes, genes associated with JA metabolic processes and plant ovule 516 development **(Figure 4C).** A detailed list of all DEGs and GO-term enrichment are included in 517 **Supplemental file S1.** 518 519 _Seed-specific SDP1 suppression indicated altered expression of candidate genes from signaling_ 520 _pathways to regulate seed size_ 521 A closer inspection of the developmental dataset identified candidate genes of the F-Box family 522 proteins and Really Interesting New Gene (RING)/U-Box super family proteins as differentially 523 expressed between genotypes. F-box and RING/U-box superfamily proteins contain members 524 that are essential components of the ubiquitin-proteasome system and are known to regulate 525 seed size through interactions with phytohormones like JA signaling pathways (Koo et al., 2014; 526 Durand et al., 2016; Hu et al., 2021). At the late developmental stage (R7), expression of five F 527 box family genes and RING/U-box genes was significantly altered **(Supplementary figure S2,** 528 **S3).** Further, two members of the Protein Phosphatase 2C (PP2C) family proteins were also 529 differentially regulated over development **(Supplementary figure S4)** . 530 531 _Reduced PEPCK activity supports suppressed gluconeogenesis in RNAi genotypes_ 532 Isotopic labeling studies suggest that gluconeogenesis may occur late in seed maturation when 533 no phloem-based source of sugars is available (Kambhampati et al., 2021). The _SDP1_ 534 suppressed transgenic seeds indicated downregulated genes associated with gluconeogenesis 535 **(Figure 4C and Supplemental file S1).** Although changes in transcript levels are not indicative 536 of altered flux, a reduction in gluconeogenesis would be consistent with reduced lipid turnover 537 and fatty acid breakdown. 538 To further evaluate this, the activity of PEPCK, which is considered a rate-limiting enzyme for 539 gluconeogenesis, was analyzed over the course of soybean seed development in the three 16 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.08.765164; this version posted November 10, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.08.765164) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY 4.0 International license.](http://creativecommons.org/licenses/by/4.0/) 540 transgenic genotypes and WT **(Figure 5)** . R5.5 seeds of _sdp1_ RNAi line A, B and C had 16 541 41% of PEPCK activity measured in WT. Similarly, R6 and R7 seeds were reduced relative to 542 WT. PEPCK activity was highest in the R7 stage. The differences in PEPCK activity between the 543 transgenic soybean seeds and the WT seeds were statistically significant at all developmental 544 stages except for B at R6 stage **(Figure 5)** and were consistent with a reduction in 545 gluconeogenic capacity that may be important when phloem sugars are not available. 546 Transcripts for several related genes that would be consistent with a role for gluconeogenesis, 547 including malate dehydrogenase, pyruvate kinase, and pyruvate dehydrogenase, were reduced 548 at R7 in the SDP1-suppressed seeds ( **Supplemental file S1)** consistent with a reduction in this 549 pathway late in development. 550 551 _SDP1 suppression results in altered phytohormone levels during mid- and late-stage seed_ 552 _development_ 553 Gene expression values suggested that seed-specific suppression of _SDP1_ might play a role in 554 altering the levels of genes associated with hormone regulation and hormone metabolic 555 processes during late seed development **(Figure 4C and Supplemental file S1)** . 556 Phytohormones, including cytokinins, auxin, ABA, and JA, are major regulators of seed size and 557 morphology (Cao et al., 2020; Hu et al., 2021; Sharma et al., 2022; Zhang et al., 2025a; Zhang 558 et al., 2025b). Some of them control the turnover of key transcription factors through interactions 559 with F-box protein-mediated ubiquitination, thereby integrating hormonal regulation with the 560 balance between plant growth and defense (Santner and Estelle, 2010; Durand et al., 2016). To 561 evaluate the impact of _SDP1_ suppression on phytohormone homeostasis, targeted LC–MS/MS 562 analysis was performed to quantify total cytokinins (sum of tZ, cZ, DHZ, and tZR), total auxins ( 563 IAA, IAA-Asp), ABA, SA, OPDA, JA, and JA-Ile at two stages of seed development: R6 and R7. 564 JA was not detected at either developmental stage. Relative levels of hormones for the three 565 _sdp1_ RNAi lines were compared to WT ( **Table 2)** with absolute levels included in **Supplemental** 566 **file S2.** 567 Notably, OPDA was reduced in all _sdp1_ RNAi lines at both developmental stages, with more 568 pronounced decreases at R7. At R6, OPDA levels declined by 34%, 14%, and 28% in A, B, and 569 C, respectively. At R7, reductions were 22-51% of WT **(Table 2)** . OPDA is the precursor of JA 570 biosynthesis and plays an important role in balancing the growth-defense tradeoff in crops (Zhu 571 et al., 2026). Other components of the JA signaling pathway provided further supporting 572 evidence of the role of JA in regulating seed yield. JA-Ile levels were decreased in the 573 transgenic lines, at R6 and in two out of the three transgenic genotypes at R7 **(Table 2)** . 17 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.08.765164; this version posted November 10, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.08.765164) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY 4.0 International license.](http://creativecommons.org/licenses/by/4.0/) 574 Suppression of JA-Ile can positively regulate seed size (Hu et al., 2021; Jiang et al., 2025; 575 Zhang et al., 2025a). Thus, changes in OPDA and JA-Ile support an SDP1 mechanism that 576 affects F-box and RING-based ubiquitination of growth-related transcription factors, including 577 those associated with seed size. 578 Other hormones were altered by the _SDP1_ suppression. ABA is a major regulator of seed 579 development, controlling seed maturation, storage reserve accumulation, dormancy, and 580 germination. In Arabidopsis, ABA levels peak at distinct times during early and late seed 581 development (Zhang et al., 2025b). Consistent with this dynamic regulation, ABA content in 582 _SDP1_ -suppressed lines differed from WT in a stage-dependent manner. During R6, ABA content 583 was significantly reduced in lines B and C by 58% or 44%, respectively, relative to WT. The 584 reduction in line A was moderate (20% relative to WT). By R7, ABA levels were increased in 585 lines A and C **(Table 2)** . 586 Total cytokinin levels were lower in all transgenic lines relative to WT at all stages (Table 2). 587 Cytokinins are well known for their roles in regulating cell division and seed development, 588 reviewed in(Sharma et al., 2022), but their effects on plant growth and development are 589 complex and context-dependent, with contrasting effects reported during root and shoot 590 meristem development (Lindsay et al., 2006; Werner et al., 2010; Bianco et al., 2013). The 591 larger-seed phenotype observed in the _SDP1_ -suppressed genotype, despite the absence of 592 detectable changes in seed number **(Figure 1E)** or cell number **(Figures 2F and 4C)** and the 593 reduction in cytokinin levels, suggests that the relationship between cytokinin homeostasis and 594 seed size is not straightforward 595 Auxin and SA effects were variable and not easily interpreted **(Table 2)** . Auxin regulates seed 596 development through its effects on silique, embryo, endosperm, and seed coat development 597 (reviewed in (Cao et al., 2020). Previous studies have shown that auxin accumulation and 598 distribution vary throughout seed development (Lur’ and Setter, 1993; Bernardi et al., 2012; 599 Locascio et al., 2014). Similarly, we found that total auxin levels exhibited genotype- and stage 600 specific variation that was more difficult to clearly assess **(Table 2).** SA levels showed an 601 apparent stage-dependent response to _SDP1_ suppression ( **Table 2)** but the relationship with 602 seed size remains unclear. 603 604 _Seed-specific suppression of SDP1 enhances lignocellulosic cell wall content_ 605 Suppression of _SDP1_ resulted in a significant increase in soybean seed size **(Figure 1D and** 606 **1F)** . Microscopy analysis revealed that this phenotype was primarily driven by an increase in 607 cell area **(Figure 2E)**, with no detectable change in cell number **(Figure 2F)** . Based on these 18 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.08.765164; this version posted November 10, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.08.765164) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY 4.0 International license.](http://creativecommons.org/licenses/by/4.0/) 608 observations, we hypothesized that enlarged cell size would be accompanied by alterations in 609 cell wall composition in the RNAi lipase lines. Consistent with this hypothesis, transcriptomic 610 analysis of late-developing (R7 stage) soybean seeds showed upregulated genes involved in 611 the β -glucan biosynthetic process **(Figure 4C and Supplemental file S1)** . To further investigate 612 this, we quantified total lignocellulosic content in mature (R8 stage) seeds from the three 613 independent transgenic lines—A, B, and C—alongside WT controls. All transgenic lines 614 exhibited a significant increase in total lignocellulosic content per seed compared to the WT 615 based on per seed **(Figure 6)** . Specifically, transgenic line C increased by 23% (p ≤ 0.05) 616 relative to WT, A was elevated. by 30% (p ≤ 0.01), and B by 32% (p ≤ 0.001) **(Figure 6;** 617 **Supplemental file S3).** 618 619 **Discussion** 620 621 _SDP1 suppression enhances soybean seed size through increased cell size_ 622 Seed size is a key agronomic trait that directly influences crop fitness and yield (Moles et al., 623 2005; Fan et al., 2006; Song et al., 2007; Linkies et al., 2010; Li and Li, 2014; Zhang et al., 624 2024a; Zhang et al., 2024 b; Wang et al., 2025; Zhang et al., 2025b) Mature seed size is 625 determined by the coordinated development of three genetically distinct compartments—the 626 maternally derived seed coat, the diploid embryo, and the triploid endosperm (Lopes and 627 Larkins’, 1993; Reiser and Fischer, 1993; Chaudhury and Berger, 2001; Haughn and 628 Chaudhury, 2005; Zhang et al., 2025a). Interconnected networks of gene expression, 629 phytohormone signaling, nutrient availability, and other signaling pathways collectively perform 630 developmental regulation and control embryo and endosperm growth and ultimately influence 631 seed yield traits (Garcia et al., 2005; Zhang et al., 2025a). Relatively few studies have identified 632 genes that simultaneously improve both seed oil accumulation and yield. Our results support a 633 conserved role for _SDP1_ suppression in promoting increased seed weight that is sustained 634 through seed filling to maturity **(Figure 1D, F and Figure 3)** . These findings further strengthen 635 the potential of seed-specific _SDP1_ suppression as a biotechnological strategy to improve both 636 seed yield and oil content (Kanai et al., 2019; Aznar-Moreno et al., 2022), without compromising 637 protein accumulation (Aznar-Moreno et al., 2022). By combining of microscopy, transcriptomic, 638 and biochemical analyses, we provide mechanistic insight into the phenotype. Changes in seed 639 size were the result of increased cell area **(Figure 2E)**, indicating that cell expansion, rather 640 than enhanced cell proliferation, underlies the observed phenotype. This interpretation is 641 supported by our transcriptomic analyses, which revealed downregulated genes associated with 19 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.08.765164; this version posted November 10, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.08.765164) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY 4.0 International license.](http://creativecommons.org/licenses/by/4.0/) 642 cell division and upregulated genes involved in cell wall organization and β -glucan biosynthetic 643 processes **(Figure 4C; Supplemental file S1)** . Furthermore, the significant increase in 644 lignocellulosic cell wall content **(Figure 6)** provided additional physiological evidence supporting 645 enhanced cell wall remodeling during seed development. 646 647 _Enhanced seed size is associated with altered JA signaling_ 648 The seed-specific suppression of _SDP1_ was associated with altered JA-related hormone 649 profiles during soybean seed development. Both OPDA and JA-Ile were significantly reduced in 650 _SDP1_ -suppressed seeds relative to WT **(Table 2)** . OPDA is a precursor in the JA biosynthetic 651 pathway (Tien et al., 2013), and JA signaling has established roles in regulating plant growth 652 and development (Tien et al., 2013; Hu et al., 2021; Jiang et al., 2025; Kamble, 2025). Although 653 the effects of JA on growth can vary among species and developmental context (Hyun et al., 654 2008; Singh et al., 2011), the concurrent reduction in OPDA and JA-Ile is consistent with altered 655 JA metabolism in _SDP1_ -suppressed seeds. Because JA-Ile is the bioactive conjugate that 656 promotes JA signaling, reduced accumulation raises the possibility that JA signaling is 657 attenuated in _SDP1_ -suppressed seeds. 658 Transcriptomic analysis showed upregulated genes associated with JA metabolic processes and 659 plant ovule development **(Figure 4C; Supplemental file S1)** . These transcriptional changes, 660 together with the reduced accumulation of OPDA and JA-Ile, provide independent evidence of 661 altered JA-associated processes during seed development. Previous genetic studies provide 662 support for a relationship between reduced JA signaling and increased seed or reproductive 663 organ size (Hu et al., 2021; Wang et al., 2023; Zhang et al., 2025a). Mutant analyses of 664 positive and negative regulators of JA signaling have demonstrated effects on seed 665 development and size (Hu et al., 2021; Wang et al., 2023; Zhang et al., 2025a). Thus, a 666 reduction in JA-Ile, along with altered expression of genes associated with plant ovule 667 development and JA metabolism support the possibility that altered JA signaling contributes to 668 the enhanced seed size phenotype observed following _SDP1_ suppression. 669 The transcriptomic analysis also identified changes in components of the ubiquitin-mediated 670 protein degradation system. GO annotation analysis revealed downregulation of genes 671 associated with proteasome core complex assembly **(Figure 4C; Supplemental file S1)**, and 672 several F-box and RING/U-box genes were differentially expressed **(Supplemental Figures S2** 673 **and S3)** . The ubiquitin–proteasome system is an important regulator of plant growth and 674 development, and F-box and RING/U-box proteins participate in diverse regulatory processes 675 (Sadanandom et al., 2012; Morreale and Walden, 2016; Shu and Yang, 2017; Zheng and 20 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.08.765164; this version posted November 10, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.08.765164) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY 4.0 International license.](http://creativecommons.org/licenses/by/4.0/) 676 Shabek, 2017; Ban and Estelle, 2021; Hina et al., 2024). However, the specific F-box and 677 RING/U-box genes identified in this study have not been established as regulators of JA 678 signaling in soybean. Moreover, the differential expression of RING/U-box genes was not 679 uniform, with both up- and downregulated members identified. Therefore, these transcriptional 680 changes are best considered as evidence of altered ubiquitin-related regulatory processes 681 rather than direct evidence of altered JA signaling. 682 The potential relationship between JA signaling and ubiquitin-mediated proteolysis is well 683 established through the SCF [COI1] complex, in which the F-box protein COI1 functions acts as the 684 receptor component required for JA-Ile-dependent degradation of JAZ repressors (Katsir et al., 685 2008; Santner and Estelle, 2010; Durand et al., 2016). Here, the F-box genes identified as 686 differentially expressed in _SDP1_ -suppressed seeds were not identified as _COI1_ itself, and their 687 functional relationship to JA signaling in soybean remains unknown. Similarly, changes in 688 RING/U-box gene expression cannot be assumed to indicate altered JA signaling because 689 these E3 ligases regulate numerous cellular and developmental pathways. Thus, although the 690 concurrent changes in JA-related metabolites and ubiquitin-related genes are notable, additional 691 functional studies will be required to determine whether any of these ubiquitin pathway 692 components contribute to JA regulation in _SDP1_ -suppressed seeds. 693 694 _Late-stage suppression of gluconeogenesis prevents lipid turnover_ 695 Recent studies indicate that seeds undergo beta-oxidation during development (Koley et al., 696 2025; Kambhampati et al., 2026) and more broadly other systems regularly recycle 697 macromolecules as a part of cellular homeostasis in different conditions (Koley et al., 2026). 698 Some levels of lipid turnover in seeds were suggested decades ago (Goepfert and Poirier, 2007; 699 Graham and Eastmond 2002). Prior work in soybean suggested that gluconeogenesis may 700 increase during late seed development, when sugar is no longer supplied by the phloem, and is 701 tightly coordinated with lipid turnover (Kambhampati et al., 2021). Based on those observations, 702 we proposed that fatty acids released through TAG hydrolysis could directly or indirectly help 703 sustain carbohydrate production when a phloem source of carbon was absent (Aznar-Moreno et 704 al., 2022; Mukherjee et al., 2025). Suppression of _SDP1_ could therefore maintain the lipid sink 705 and limit available carbon for respiration, or carbohydrate biosynthesis, including RFO 706 production. In this study, GO-annotation of transcriptomes revealed that downregulated genes 707 were enriched for acetyl-CoA metabolic processes and gluconeogenesis during late seed 708 development **(Figure 4C and Supplemental file S1)** . A reduction in PEPCK activity, a crucial 709 step in gluconeogenesis, was observed across development in the RNAi genotypes **(Figure 5)** 21 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.08.765164; this version posted November 10, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.08.765164) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY 4.0 International license.](http://creativecommons.org/licenses/by/4.0/) 710 suggesting that suppression of _SDP1_ may restrict gluconeogenic carbon flux during seed 711 development. 712 Collectively, our findings demonstrate that seed-specific suppression of _SDP1_ enhances seed 713 size, increases oil accumulation, and improves seed yield-related traits through changes in 714 developmental and metabolic processes. Beyond its established role in TAG degradation, _SDP1_ 715 suppression alters carbon partitioning, cellular remodeling, phytohormone homeostasis, and 716 multiple regulatory pathways during seed development. These integrated changes are 717 associated with greater oil content and larger seeds without compromising seed number per 718 plant. Therefore, seed-specific suppression of _SDP1_ represents a promising biotechnological 719 strategy for simultaneously improving seed oil content and agronomic performance in soybean. 720 721 722 **Author Contributions** 723 The experimental plan was conceived by TM, BM, TPD, and DKA. Phenotyping experiments 724 were performed by TM, SC, KG and KMM. AK and KC performed microscopy imaging. Analysis 725 of microscopic data was performed by TM and KG. TM, KLC, ML, RW, AD performed biomass, 726 enzyme and hormone analysis with mass spectrometry. AD, BM, and TM performed RNAseq 727 analysis. The initial draft was composed by TM and DKA with subsequent help from TPD and 728 BM and additional comments provided, and changes incorporated from all others. All authors 729 approved the submission of the final version of the manuscript. 730 731 **Data Availability** 732 All data presented in this manuscript is provided through the manuscript and supplements. Raw 733 files for mass spectrometry and microscopy have been deposited in Zenodo 734 (10.5281/zenodo.23378262). Raw RNAseq FASTQ files are uploaded to NCBI SRA under 735 PRJNA1556110. 736 737 **Conflict of Interest** 738 The authors declare no competing interests. 739 740 **Acknowledgements** 741 We wish to thank the Phenotyping Facility (RRID:SCR_019049) and the Advanced Bioimaging 742 Facility (RRID:SCR_018951) at the Donald Danforth Plant Science Center for support with 743 imaging and image analysis of seeds, the Bioanalytical Chemistry Facility (RRID:SCR_001047) 22 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.08.765164; this version posted November 10, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.08.765164) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY 4.0 International license.](http://creativecommons.org/licenses/by/4.0/) 744 that helped with hormone analysis and provided mass spectrometry instrumentation for 745 metabolite analyses, and the Plant Growth Facility (RRID:SCR_024902) where soybeans were 746 grown for all experiments. We also like to thank Ms. Jiahong Zhou and Ms. Courtney Harmon 747 for her valuable insight related to correlation analysis and cell counting. Funding was provided 748 by the United Soybean Board, grant no. 24-203-S-B-1-A,26-203-S-B-1-A and 27-203-A-1-B. In 749 addition, funding from the National Science Foundation (DBI-2242822), the United States 750 Department of Agriculture National Institute of Food and Agriculture grant no. (2023-67017 751 39419) and the Department of Energy Office of Science, Office of Biological and Environmental 752 Research (DE SC0023142). 753 754 **References** 755 Alexa A, Rahnenführer J, Lengauer T. 2006. Improved scoring of functional groups from gene 756 expression data by decorrelating GO graph structure. _Bioinformatics._ 22: 1600–1607. 757 Allen DK, Ohlrogge JB, Shachar-Hill Y.2009. The role of light in soybean seed filling metabolism. 758 _Plant J._ 58: 220–234 759 Allen DK, Young JD. 2013. Carbon and nitrogen provisions alter the metabolic flux in developing 760 soybean embryos. _Plant Physiol._ 161: 1458–1475 761 Arganda-Carreras I, Kaynig V, Rueden C, Eliceiri KW, Schindelin J, Cardona A, Seung HS. 762 2017. Trainable Weka Segmentation: A machine learning tool for microscopy pixel classification. 763 _Bioinformatics._ 33: 2424–2426 764 Reiser L, Fischer RL. 1993. The Ovule and the Embryo Sac. _Plant Cell._ 5:1291-1301 765 Azeez A, Parchuri P, Bates PD. 2022. Suppression of Physaria fendleri SDP1 Increased Seed 766 Oil and Hydroxy Fatty Acid Content While Maintaining Oil Biosynthesis Through Triacylglycerol 767 Remodeling. _Front Plant Sci._ 13 768 Aznar-Moreno JA, Mukherjee T, Morley SA, Duressa D, Kambhampati S, Chu KL, Koley S, Allen 769 DK, Durrett TP. 2022. Suppression of SDP1 Improves Soybean Seed Composition by 770 Increasing Oil and Reducing Undigestible Oligosaccharides. _Front Plant Sci. 2022_ :13. 771 Ban Z, Estelle M. 2021. CUL3 E3 ligases in plant development and environmental response. 772 _Nat Plants._ 7: 6–16 773 Baud S, Dubreucq B, Miquel M, Rochat C, Lepiniec L. 2008. Storage Reserve Accumulation in 774 Arabidopsis: Metabolic and Developmental Control of Seed Filling. _Arabidopsis Book._ 6: e0113 775 Baud S, Lepiniec L. 2009. Regulation of de novo fatty acid synthesis in maturing oilseeds of 776 Arabidopsis. _Plant Physiol. Biochem._ 47: 448–455 23 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.08.765164; this version posted November 10, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.08.765164) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY 4.0 International license.](http://creativecommons.org/licenses/by/4.0/) 777 Bernardi J, Lanubile A, Li QB, Kumar D, Kladnik A, Cook SD, Ross JJ, Marocco A, Chourey PS. 778 2012. Impaired auxin biosynthesis in the defective endosperm18 mutant is due to mutational 779 loss of expression in the ZmYuc1 gene encoding endosperm-specific YUCCA1 Protein in Maize. 780 _Plant Physiol_ **.** 160: 1318–1328 781 Bianco MD, Giustini L, Sabatini S. 2013. Spatiotemporal changes in the role of cytokinin during 782 root development. _New Phytol_ .199:324-338 783 Caffall KH, Pattathil S, Phillips SE, Hahn MG, Mohnen D.2009.Arabidopsis thalian T-DNA 784 mutants implicate GAUT genes in the biosynthesis of pectin and xylan in cell walls and seed 785 testa. Mol.Plant.2:1000-1014 786 Cao J, Li G, Qu D, Li X, Wang Y.2020. Into the seed: Auxin controls seed development and 787 grain yield. _Int J Mol Sci._ 21(5):1662 788 Chaudhury AM, Berger F. 2001. Maternal control of seed development. _Semin Cell Dev Biol_ **.** 12: 789 381–386 790 Chia TYP, Pike MJ, Rawsthorne S. 2005. Storage oil breakdown during embryo development of 791 Brassica napus (L.). _J Exp Bot_ 56: 1285–1296 792 Coon CN, Leske KL, Akavanichan O, Cheng TK. 1990. Effect of Oligosaccharide-Free Soybean 793 Meal on True Metabolizable Energy and Fiber Digestion in Adult Roosters. _Poult. Sci._ 69:787 794 793 795 Czajka JJ, Kambhampati S, Tang YJ, Wang Y, Allen DK.2020. Application of stable isotope 796 tracing to elucidate metabolic dynamics during Yarrowia lipolytica α -lonone fermentation. 797 _iScience_ 23:100854 798 Dhankher 2021. * Genbank (NCBI, Sequence Accession No. XM_010489120.1 � ; Pub lished. 799 PNAS 800 Dierking EC, Bilyeu KD. 2008. Association of a Soybean Raffinose Synthase Gene with Low 801 Raffinose and Stachyose Seed Phenotype. Plant Genome. 1:(2) 802 Duenwald JG, Srivastava D, Murphy KM. 2025. Imaging and Analysis for Quantifying Maize 803 (Zea mays) Abiotic Stress Phenotypes. _J.Vis.Exp_ . 217 804 Durand AN, Pauwels L, Goossens A. 2016. The ubiquitin system and jasmonate signaling. 805 _Plants_ . 5: 183–205 806 Fan C, Xing Y, Mao H, Lu T, Han B, Xu C, Li X, Zhang Q. 2006. GS3, a major QTL for grain 807 length and weight and minor QTL for grain width and thickness in rice, encodes a putative 808 transmembrane protein. _Theor.Appl.Genet._ 112: 1164–1171 809 Foster CE, Martin TM, Pauly M.2010. Comprehensive compositional analysis of plant cell walls 810 (lignocellulosic biomass) Part 1: Lignin. Jove 37 24 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.08.765164; this version posted November 10, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.08.765164) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY 4.0 International license.](http://creativecommons.org/licenses/by/4.0/) 811 Garcia D, Gerald JNF, Berger F. 2005. Maternal control of integument cell elongation and 812 zygotic control of endosperm growth are coordinated to determine seed size in Arabidopsis. 813 Plant Cell 17: 52–60 814 Gehan MA, Fahlgren N, Abbasi A, Berry JC, Callen ST, Chavez L, Doust AN, Feldman MJ, 815 Gilbert KB, Hodge JG, _et al._ 2017. PlantCV v2: Image analysis software for high-throughput 816 plant phenotyping. _PeerJ_ . 5: e4088 817 Goepfert S, Poirier Y. 2007. β -Oxidation in fatty acid degradation and beyond. _Curr Opin Plant_ 818 _Biol_ **.** 10: 245–251 819 Graham IA, Eastmond PJ. 2002. Pathways of straight and branched chain fatty acid catabolism 820 in higher plants. _Prog Lipid Res._ 41(2):156-81 821 Haughn G, Chaudhury A. 2005. Genetic analysis of seed coat development in Arabidopsis. 822 _Trends Plant Sci_ . 10: 472–477 823 Hina A, Khan N, Kong K, Lv W, Karikari B, Abbasi A, Zhao T. 2024. Exploring the role of FBXL 824 fbxl gene family in Soybean: Implications for plant height and seed size regulation. Physiol 825 Plant. 176(1):e14191 826 Hu S, Yang H, Gao H, Yan J, Xie D. 2021. Control of seed size by jasmonate. _Sci China Life_ 827 _Sci._ 64: 1215–1226 828 Hymowitz T, Collins FI, Panczner J, Walker WM. 1972. Relationship Between the Content of Oil, 829 Protein, and Sugar in Soybean Seed. _Agron.J._ 64:5:613-616 830 Hyun Y, Choi S, Hwang HJ, Yu J, Nam SJ, Ko J, Park JY, Seo YS, Kim EY, Ryu SB, _et al_ . 2008. 831 Cooperation and Functional Diversification of Two Closely Related Galactolipase Genes 832 for Jasmonate Biosynthesis. _Dev. Cell._ 14: 183–192 833 Jiang S, Wu L, Huang L, Li Y. 2025. Genetic and molecular mechanisms of phytohormone 834 mediated seed size control in crops. _J.Genet.Genom._ 52: 1460–1474 835 Johnson HW, Robinson HF, Comstock RE. 1955. Genotypic and Phenotypic Correlations in 836 Soybeans and Their Implications in Selection. _Agron.J._ 47:477-483 837 Kambhampati S, Aznar-Moreno JA, Bailey SR, Arp JJ, Chu KL, Bilyeu KD, Durrett TP, Allen DK. 838 2021. Temporal changes in metabolism late in seed development affect biomass composition. 839 _Plant Physiol_ . 186: 874–890 840 Kambhampati S, Aznar-Moreno JA, Hostetler C, Caso T, Bailey SR, Hubbard AH, Durrett TP, 841 Allen DK. 2020. On the inverse correlation of protein and oil: Examining the effects of altered 842 central carbon metabolism on seed composition using soybean fast neutron mutants. 843 _Metabolites_ . 10(1):18 25 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.08.765164; this version posted November 10, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.08.765164) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY 4.0 International license.](http://creativecommons.org/licenses/by/4.0/) 844 Kambhampati S, Combs-Giroir R, Demiwal P, Koley S, Allen DK. 2026. Connecting the dots in 845 plant metabolism: Isotopic labeling and metabolic flux analysis. _Curr Opin Plant Biol._ 92:102924 846 Kamble NU. 2025. Tiny seeds, big decisions: Jasmonate-mediated regulation of seed size in 847 Arabidopsis via the SOD7-KLU module. Plant Cell. 37:(7) koaf160 848 Kanai M, Yamada T, Hayashi M, Mano S, Nishimura M. 2019. Soybean (Glycine max L.) 849 triacylglycerol lipase GmSDP1 regulates the quality and quantity of seed oil. _Sci Rep._ 9:8924 850 Kataya A, Nascimento JRS, Xu C, Garneau MG, Koley S, Kimberlin A, Mukherjee T, Mooney 851 BP, Xu D, Bates PD, Allen DK, Koo AJ, Thelen JJ. 2025. Comparative omics reveals 852 unanticipated metabolic rearrangements in a high-oil mutant of plastid acetyl-CoA carboxylase. 853 J. Proteome Res.24:2675-2688 854 Katsir L, Schilmiller AL, Staswick PE, Yang He S, Howe GA. 2008. COI1 is a critical component 855 of a receptor for jasmonate and the bacterial virulence factor coronatine. 856 _Proc.Natl.Acad.Sci.U.S.A._ 105: (19) 7100-7105 857 Kelly AA, Shaw E, Powers SJ, Kurup S, Eastmond PJ.2013. Suppression of the SUGAR 858 DEPENDENT1 triacylglycerol lipase family during seed development enhances oil yield in 859 oilseed rape (Brassica napus L.). _Plant Biotechnol J._ 11: 355–361 860 Kim MJ, Yang SW, Mao HZ, Veena SP, Yin JL, Chua NH. 2014. Gene silencing of Sugar 861 dependent 1 (JcSDP1), encoding a patatin-domain triacylglycerol lipase, enhances seed oil 862 accumulation in Jatropha curcas. _Biotechnol Biofuels_ . 7(1):36 863 Koley S, Foley K, Perrine Z, Morley SA, Kambhampati S, Gomez O, Chu KL, Chou YH, Wei M, 864 Tzeng SC, _et al._ 2026. Metabolic rewiring and biomass redistribution enable optimized 865 mixotrophic growth in Chlamydomonas. _Proc.Natl.Acad.Sci.U.S.A_ 123:(4):e2522572123 866 Koley S, Jyoti P, Lingwan M, Wei M, Xu C, Chu KL, Williams RB, Koo AJ, Thelen JJ, Xu D, _et al._ 867 2025. Persistent fatty acid catabolism during plant oil synthesis. _Cell Rep._ 44(4):115492 868 Koo AJ, Thireault C, Zemelis S, Poudel AN, Zhang T, Kitaoka N, Brandizzi F, Matsuura H, Howe 869 GA. 2014. Endoplasmic reticulum-associated inactivation of the hormone jasmonoyl-L 870 isoleucine by multiple members of the cytochrome P450 94 family in Arabidopsis. J. Biol. 871 Chem.289: 29728–29738 872 Lindsay DL, Sawhney VK, Peta CBS.2006. Cytokinin-induced changes in _CLAVATA1_ and 873 _WUSCHEL_ expression temporally coincide with altered floral development in Arabidopsis. Plant 874 Sci. 170:1111-1117 875 Li N, Li Y. 2014. Ubiquitin-mediated control of seed size in plants. Front Plant Sci. 5 876 Linkies A, Graeber K, Knight C, Leubner-Metzger G. 2010. The evolution of seeds. _New Phytol_ **.** 877 186: 817–831 26 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.08.765164; this version posted November 10, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.08.765164) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY 4.0 International license.](http://creativecommons.org/licenses/by/4.0/) 878 Litterer LA, Plaisance KL, Schnurr JA, Storey KK, Jung HJG, Gronwald JW, Somers DA. 2006. 879 Biosynthesis of UDP-glucuronic acid in developing soybean embryos: Possible role of UDP 880 sugar pyrophosphorylase. _Physiol Plant_ **.** 128: 200–211 881 Locascio A, Roig-Villanova I, Bernardi J, Varotto S. 2014. Current perspectives on the hormonal 882 control of seed development in Arabidopsis and maize: A focus on auxin. _Front Plant Sci._ 25 883 Lopes MA, Larkins’ BA. 1993. Endosperm Origin, Development, and Function. _Plant Cell._ 884 5(10):1383-99 885 Love MI, Huber W, Anders S. 2014. Moderated estimation of fold change and dispersion for 886 RNA-seq data with DESeq2. _Genome Biol._ 15(12):550 887 Lur’ H-S, Setter TL. 1993. Role of Auxin in Maize Endosperm Development’ Timing of Nuclear 888 DNA Endoreduplication, Zein Expression, and Cytokinin. _Plant Physiol._ 103(1): 273-280 889 McCleary B V., Charmier LMJ, McKie VA. 2019. Measurement of Starch: Critical Evaluation of 890 Current Methodology. _Starch._ 72:(1-2) 891 Mccleary B V, Gibson TS, Mugford DC, Lukow O, Jackson, D S, Rabe, E, Patel, N, Williams, P 892 C, Gelroth, J, Camire, M E, _et al_ . 1997. Measurement of Total Starch in Cereal Products by 893 Amyloglucosidase-oc-Amylase Method: Collaborative Study. _Food Composition and Additives._ 894 _J.AOAC Int._ 80:(3) 895 Middelbos IS, Fahey GC. 2008. Soybean Carbohydrates. _Soybeans_ . 269-296 896 Moles AT, Ackerly DD, Webb CO, Tweddle JC, Dickie JB, Pitman AJ, Westoby M. 2005. Factors 897 that shape seed mass evolution. _Proc.Natl.Acad.Sci._ 102(30):10540-10544 898 Morley SA, Ma F, Alazem M, Frankfater C, Yi H, Burch-Smith T, Clemente TE, Veena V, Nguyen 899 H, Allen DK. 2023. Expression of malic enzyme reveals subcellular carbon partitioning for 900 storage reserve production in soybeans. _New Phytol._ 239: 1834–1851 901 Morreale FE, Walden H. 2016. Types of Ubiquitin Ligases. _Cell._ 165: 248-248.e1 902 Mukherjee T, Kambhampati S, Morley SA, Durrett TP, Allen DK. 2025. Metabolic flux analysis to 903 increase oil in seeds. _Plant Physiol._ 197 (2): kiae 595 904 Osorio, S., Vallarino, J.G., Szecowka, M., Ufaz, S., Tzin, V., Angelovici, R., Galili, G., and 905 Fernie, A.R. 2014. Extraction and Measurement the Activities of Cytosolic Phosphoenolpyruvate 906 Carboxykinase (PEPCK) and Plastidic NADP-dependent Malic Enzyme (ME) on Tomato 907 ( _Solanum lycopersicum_ ). _Bio-protocol_, 4(9): e1122. 908 Patro R, Duggal G, Love MI, Irizarry RA, Kingsford C. 2017. Salmon provides fast and bias 909 aware quantification of transcript expression. _Nat Methods_ 14: 417–419 910 Pegler JL, Grof CP, Patrick JW. 2023. Sugar loading of crop seeds – a partnership of phloem, 911 plasmodesmal and membrane transport. _New Phytol._ 239: 1584–1602 27 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.08.765164; this version posted November 10, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.08.765164) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY 4.0 International license.](http://creativecommons.org/licenses/by/4.0/) 912 Pettolino FA, Walsh C, Fincher GB, Bacic A. 2012. Determining the polysaccharide composition 913 of plant cell walls. Nature protocols. 7:9: 1590-1607 914 Pino LK, Searle BC, Bollinger JG, Nunn B, Maclean B, Maccoss MJ. 2020. The Skyline 915 Ecosystem: Informatics for Quantitative Mass Spectrometry Proteomics. _Mass Spectrom_ 916 _Rev._ 39(3):229-244 917 Sadanandom A, Bailey M, Ewan R, Lee J, Nelis S. 2012. The ubiquitin-proteasome system: 918 Central modifier of plant signalling. _New Phytol._ 196: 13–28 919 Santner A, Estelle M. 2010. The ubiquitin-proteasome system regulates plant hormone 920 signaling. _Plant J._ 61: 1029–1040 921 Sharwood RE, Sonawane BV, Ghannoum O, Whitney SM.2016. Improved analysis of C4 and 922 C3 photosynthesis via refined in vitro assays of their carbon fixation biochemistry. _J.Exp.Bot_ 923 67(10):3137-48. 924 Schindelin J, Arganda-Carreras I, Frise E, Kaynig V, Longair M, Pietzsch T, Preibisch S, Rueden 925 C, Saalfeld S, Schmid B, _et al._ 2012. Fiji: An open-source platform for biological-image analysis. 926 _Nat Methods._ 9: 676–682 927 Sharma S, Kaur P, Gaikwad K. 2022. Role of cytokinins in seed development in pulses and 928 oilseed crops: Current status and future perspective. _Front Genet._ (13) 929 Shu K, Yang W. 2017. E3 ubiquitin ligases: Ubiquitous actors in plant development and abiotic 930 stress responses. _Plant Cell Physiol._ 58: 1461–1476 931 Singh AK, Fu D-Q, El-Habbak M, Navarre D, Ghabrial S, Kachroo A. 2011. Silencing Genes 932 Encoding Omega-3 Fatty Acid Desaturase Alters Seed Size and Accumulation of Bean pod 933 mottle virus in Soybean. _Mol Plant Microbe Interact_ 506: 506–515 934 Singh Y. 2024. Prospective of agriculture biotechnology to minimize hunger and malnutrition. 935 _IJABR_ . 8: 676–684 936 Soneson C, Love MI, Robinson MD. 2015. Differential analyses for RNA-seq: transcript-level 937 estimates improve gene-level inferences. F1000Res. 4: 1521 938 Song XJ, Huang W, Shi M, Zhu MZ, Lin HX. 2007. A QTL for rice grain width and weight 939 encodes a previously unknown RING-type E3 ubiquitin ligase. _Nat Genet_ 39: 623–630. 940 Stephens M.2016. False discovery rates: a new deal. Biostatistics. 18(2): 275-294 941 Suarez FL, Springfield J, Furne JK, Lohrmann TT, Kerr PS, Levitt MD. 1999. Gas production in 942 human ingesting a soybean flour derived from beans naturally low in oligosaccharides. _Am J_ 943 _Clin Nutr_ . 69(1):135-9 28 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.08.765164; this version posted November 10, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.08.765164) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY 4.0 International license.](http://creativecommons.org/licenses/by/4.0/) 944 Tamagno S, Sadras VO, Aznar-Moreno JA, Durrett TP, Ciampitti IA. 2022. Selection for yield 945 shifted the proportion of oil and protein in favor of low-energy seed fractions in soybean. _Field_ 946 _Crops Res._ 279 947 Thien Nguyen Q, Kisiala A, Andreas P, Neil Emery R, and Narine S.2016.Soybean Seed 948 Development: Fatty Acid and Phytohormone Metabolism and Their Interactions. Curr 949 Genomics.2016:17(3):241-60. 950 Tien DN, Oo MM, Soh M-S, Park SK. 2013. Bioengineering of Male Sterility in Rice (Oryza 951 sativa L.) _Plant Breed Biotechnol._ 1: 218–235 952 Valentine MF, De Tar JR, Mookkan M, Firman JD, Zhang ZJ. 2017. Silencing of soybean 953 raffinose synthase gene reduced raffinose family oligosaccharides and increased true 954 metabolizable energy of poultry feed. _Front Plant Sci._ 8 955 Wang X, Chen Y, Liu S, Fu W, Zhuang Y, Xu J, Lou Y, Baldwin IT, Li R. 2023 Functional 956 dissection of rice jasmonate receptors involved in development and defense. _New Phytol_ 238: 957 2144–2158 958 Wang X, Sun J, Yi Z, Dong S. 2025. Effects of seed size on soybean performance: germination, 959 growth, stress resistance, photosynthesis, and yield. _BMC Plant Biol._ 25: 219 960 Werner T, Nehnevajova E, Kollmer I, Novak O, Strnad M, Kramer U, Schmulling T. 2010. Root 961 specific reduction of cytokinin causes enhanced root growth, drought tolerance and leaf mineral 962 enrichment in Arabidopsis and Tobacco. _Plant Cell_ 22:3905-3920 963 Wickham H, Averick M, Bryan J, Chang W, McGowan L, François R, Grolemund G, Hayes A, 964 Henry L, Hester J, _et al._ 2019. Welcome to the Tidyverse. _J Open Source Softw._ 4: 1686 965 Wilcox JR, Shibles RM. 2001. Interrelationships among seed quality attributes in soybean. _Crop_ 966 _Sci_ 41: 11–14 967 Yang S, Miao L, He J, Zhang K, Li Y, Gai J. 2019. Dynamic transcriptome changes related to oil 968 accumulation in developing soybean seeds. _Int J Mol Sci._ 20 (9):2202 969 Zhang J, Yao J, He K, Yu C, Du J, Du J, Fu Q, Yao R, Howe GA, Han X, _et al._ 2025a Jasmonate 970 signaling coordinates with the SOD7–KLU pathway to regulate seed size in Arabidopsis 971 thaliana. _Plant Cell._ 37(8): koaf178 972 Zhang P, Yang Z, Jia S, Chen G, Li N, Karikari B, Cao Y. 2024a. Genome-Wide Association 973 Study and Candidate Gene Mining of Seed Size Traits in Soybean. _Agronomy._ 14:6 974 Zhang Y, Anfang M, Rowe JH, Rizza A, Li Z, Su N, Bar H, Charrier L, Geisler M, Jones AM, _et_ 975 _al_ . (2025b) ABA importers ABCG17 and ABCG18 redundantly regulate seed size in Arabidopsis. 976 _Plant J._ 121 (6):e70096 29 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.08.765164; this version posted November 10, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.08.765164) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY 4.0 International license.](http://creativecommons.org/licenses/by/4.0/) 977 Zhang Y, Bhat JA, Zhang Y, Yang S. 2024b. Understanding the Molecular Regulatory Networks 978 of Seed Size in Soybean. _Int J Mol Sci._ 25:3 979 Zheng N, Shabek N. 2017. Ubiquitin Ligases: Structure, Function, and Regulation. _Annu Rev_ 980 _Biochem._ 86:129-157 981 Zhu A, Ibrahim JG, Love MI. 2019. Heavy-Tailed prior distributions for sequence count data: 982 Removing the noise and preserving large differences. _Bioinformatics_ . 35: 2084–2092 983 Zhu J, Hu S, Cai Y, Wang C, Xie D, Shan X. 2026 Transcending the jasmonate-mediated 984 growth–defense trade-off: Pathways to unlocking plant potential. Plant Commun. 7:(6)101828 985 The Sustainable Development Goals Report. 2022. United Nations 986 Soystats 2026. A Reference Guide to Important Soybean Facts and Figures. American Soybean 987 Association. 988 989 **Figure Legends** 990 991 Figure 1 992 Phenotypic and agronomic measurements of WT and _SDP1-_ suppressed soybean genotypes 993 The total number of nodes (A), pods (B), and seeds (C) were determined from four-six plants 994 each for WT, A ( _sdp1-1 RNAi A_ ), B ( _sdp1-3, RNAi D_ ) and C ( _sdp1-3, RNAi C_ ), labeled here as 995 _sdp1._ Average individual seed weight was calculated from 100 seeds per each of four-six plants 996 per genotype (D). The correlation between average individual seed weight and seed number (E) 997 was determined using the Pearson correlation coefficient (PCC) and the p-value was 998 determined using the corresponding t-distribution of the PCC. Average individual seed area was 999 measured using RGB images of 10 seeds from each of four-six plants per genotype, processed 1000 with PlantCV software (F). Box-plot center lines indicate the median, ‘x’ represent the mean; box 1001 limits, 25 [th] and 75 [th] percentiles; whiskers, 1.5 times the interquartile range from the 25 [th] and 75 [th] 1002 percentiles. ***p ≤ 0.001 indicates a statistically significant difference based on a two-tailed 1003 Student’s t-test. 1004 1005 Figure 2 1006 Cellular morphology of WT and _SDP1-_ suppressed soybean genotypes 1007 Confocal microscopy imaging of three to five matured soybean seeds from each of WT and 1008 three _sdp1_ suppressed genotypes A ( _sdp1-1 RNAi A_ ), B ( _sdp1-3, RNAi D_ ) and C ( _sdp1-3, RNAi_ 1009 _C_ ), labeled here as _sdp1_, bar=50.01 um (A) Density histogram using 4-10 images per seed from 1010 three to five seeds per each of WT and A (B) WT and B (C) and WT and C (D) indicated a shift 30 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.08.765164; this version posted November 10, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.08.765164) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY 4.0 International license.](http://creativecommons.org/licenses/by/4.0/) 1011 towards bigger cell area in the transgenic seeds as compared to WT controls. Cell area (E) and 1012 cell number (F) of _SDP1_ -suppressed and WT genotypes (F). Data represents mean ± standard 1013 error. ***p ≤ 0.001 indicates a statistically significant difference based on a two-tailed Student’s 1014 t-test. 1015 1016 Figure 3 1017 Fresh weight and dry weight measurement of soybean seeds 1018 Fresh weight (FW) and dry weights (DW) measurements of seeds from four individual plants of 1019 the three _SDP1-_ suppressed genotypes A ( _sdp1-1 RNAi A_ ), B ( _sdp1-3, RNAi D_ ) and C ( _sdp1-3,_ 1020 _RNAi C_ ), labeled here as _sdp1_ and WT. Data represents mean ± standard error. **p ≤ 0.01 ***p 1021 ≤ 0.001 indicates a statistically significant difference based on a two-tailed Student’s t-test. 1022 1023 Figure 4 1024 Stage specific differential transcriptomic responses to _SDP1_ suppression 1025 Soybean seeds from three to four plants per three reproductive stages (R5.5, R6 and R7) from 1026 WT and the transgenic genotype B ( _sdp1-3, RNAi D_ ), was used to determine transcriptomic 1027 responses to _SDP1_ suppression. Principal component analysis (PCA) (A) Number of 1028 differentially expressed genes (DEGs) (B) and Selected Gene Ontology (GO) term enrichment 1029 analysis of R7 seeds (C). The colors of the lines represent the –log10(Fisher.weight01) p-value 1030 as calculated by the topGO R package. 1031 1032 Figure 5 1033 PEPCK activity across soybean seed development 1034 PEPCK activity, as a signature of gluconeogenesis, over the course of seed development from 1035 four individual plants of the three _SDP1-_ suppressed genotypes A ( _sdp1-1 RNAi A_ ), B ( _sdp1-3,_ 1036 _RNAi D_ ) and C ( _sdp1-3, RNAi C_ ), labeled here as _sdp1_ and WT. Box-plot center lines indicate 1037 the median, ‘x’ represent the mean; box limits, 25 [th] and 75 [th] percentiles; whiskers, 1.5 times the 1038 interquartile range from the 25 [th] and 75 [th] percentiles. *p ≤ 0.05, **p ≤ 0.01 indicates a 1039 statistically significant difference based on a two-tailed Student’s t-test. 1040 1041 Figure 6 1042 Lignocellulose content 1043 Lignocellulose wall content from four plants each of the three _SDP1-_ suppressed genotypes A 1044 ( _sdp1-1 RNAi A_ ), B ( _sdp1-3, RNAi D_ ) and C ( _sdp1-3, RNAi C_ ), labeled here as _sdp1_ and WT. 31 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.08.765164; this version posted November 10, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.08.765164) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY 4.0 International license.](http://creativecommons.org/licenses/by/4.0/) 1045 Data represents mean ± standard error. *p ≤ 0.05, **p ≤ 0.01, ***p ≤ 0.001 indicates a 1046 statistically significant difference based on a two-tailed Student’s t-test. 1047 1048 Supplementary figure S1 1049 Phenotypic and agronomic measurements of WT and _SDP1-_ suppressed soybean genotypes. 1050 The correlation between total number of pods and nodes (A) and total numbers of seeds and 1051 pods (B) was determined using the Pearson correlation coefficient. P-value was obtained using 1052 the corresponding t-distribution of the Pearson correlation coefficient. 1053 Mature seed images were taken using Nikon D7200 camera and a color checker card is 1054 introduced as reference for color correction and white balance (C) 1055 Seed area calculation using the morphology package in PlantCV after masking the seed from 1056 the background (D) 1057 1058 Supplementary figure S2 1059 Across development expression of F-box family protein 1060 Soybean seeds from three to four plants per three reproductive stages (R5.5, R6 and R7) from 1061 WT and the transgenic genotype B ( _sdp1-3, RNAi D_ ), was used to determine transcriptomic 1062 responses to _SDP1_ suppression. Expression of five family members of the F-box family protein 1063 Glyma.02G018450 (A), Glyma.04G042900 (B), Glyma.08G252300 (C), Glyma.17G237800 (D) 1064 and Glyma.19G196400 (E). Each point represents one sample in the RNA-seq dataset; 1065 boxplots show the interquartile range, and whiskers extend to the most extreme non-outlier 1066 values. Log2 fold changes were shrunk using the adaptive shrinkage estimator (ashr; Stephens, 1067 2016) implemented in lfcShrink(). *p~adj~ ≤ 0.05, **p~adj~ ≤ 0.01, ***p~adj~ ≤ 0.001, based on 1068 DESeq2 Wald test p-values adjusted for multiple testing (Benjamini-Hochberg). 1069 1070 Supplementary figure S3 1071 Across development expression of RING/U-box superfamily protein 1072 Soybean seeds from three to four plants per three reproductive stages (R5.5, R6 and R7) from 1073 WT and the transgenic genotype B ( _sdp1-3, RNAi D_ ), was used to determine transcriptomic 1074 responses to _SDP1_ suppression. Expression of ten family members of the RING/U-box 1075 superfamily protein Glyma.03G179300 (A), Glyma.06G150400 (B), Glyma.07G230600 (C), 1076 Glyma.09G216900 (D), Glyma.12G047300 (E), Glyma.12G173300 (F), Glyma.13G139400 (G), 1077 Glyma.16G076000 (H), Glyma.17G009900 (I) and Glyma.19G1800000 (J). Each point 1078 represents one sample in the RNA-seq dataset; boxplots show the interquartile range, and 32 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.08.765164; this version posted November 10, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.08.765164) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY 4.0 International license.](http://creativecommons.org/licenses/by/4.0/) 1079 whiskers extend to the most extreme non-outlier values. Log2 fold changes were shrunk using 1080 the adaptive shrinkage estimator (ashr; Stephens, 2016) implemented in lfcShrink(). *p~adj~ ≤ 1081 0.05, **p~adj~ ≤ 0.01, based on DESeq2 Wald test p-values adjusted for multiple testing 1082 (Benjamini-Hochberg). 1083 1084 Supplementary figure S4 1085 Across development expression of two Protein phosphatase 2C family protein 1086 Soybean seeds from three to four plants per three reproductive stages (R5.5, R6 and R7) from 1087 WT and the transgenic genotype B ( _sdp1-3, RNAi D_ ), was used to determine transcriptomic 1088 responses to _SDP1_ suppression. Expression of two family members of the Protein phosphatase 1089 2C family protein Glyma.14G082200 (A) and Glyma.U032205 (B). Each point represents one 1090 sample in the RNA-seq dataset; boxplots show the interquartile range, and whiskers extend to 1091 the most extreme non-outlier values. Log2 fold changes were shrunk using the adaptive 1092 shrinkage estimator (ashr; Stephens, 2016) implemented in lfcShrink(). *p~adj~ ≤ 0.05, 1093 ***p~adj~ ≤ 0.001, based on DESeq2 Wald test p-values adjusted for multiple testing 1094 (Benjamini-Hochberg). 1095 1096 **Table Legends** 1097 1098 Table 1 1099 Biomass composition of _SDP1-_ suppressed soybean seeds 1100 Soybean seed storage biomass was quantified from three to four plants per three reproductive 1101 stages and maturity (n=3-4 with 2-4 technical replicates) from WT (Williams 82) and three 1102 transgenic genotypes A ( _sdp1-1 RNAi A_ ), B ( _sdp1-3, RNAi D_ ) and C ( _sdp1-3, RNAi C_ ). Total 1103 lipid, starch and oligosaccharides (raffinose and stachyose) were determined on a mg per seed 1104 basis. Data represents mean ± Standard error. **p ≤ 0.01, and ***p ≤ 0.001, indicate statistically 1105 significant difference based on a two-tailed Student’s t-test. 1106 1107 Table 2 1108 Relative levels of total cytokinins, total auxins, abscisic acid (ABA), 12-oxo-phytodecanoic acid 1109 (OPDA), Jasmonoyl-isoleucine (Ja-Ile), and salicylic acid (SA), in the three _sdp1_ suppressed 1110 transgenic genotypes A ( _sdp1-1 RNAi A_ ), B ( _sdp1-3, RNAi D_ ) and C ( _sdp1-3, RNAi C_ ) as 1111 compared to WT during R6 (mid) and R7 (late) development stages. Data represents mean ± 33 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.08.765164; this version posted November 10, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.08.765164) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY 4.0 International license.](http://creativecommons.org/licenses/by/4.0/) 1112 standard error (n=3-4 plants/genotype). *p ≤ 0.05, **p ≤ 0.01, ***p ≤ 0.001 indicates a 1113 statistically significant difference based on a two-tailed t-test. 1114 1115 Supplementary table S1 1116 Biomass composition of _SDP1-_ suppressed soybean seeds 1117 Soybean seed storage biomass was quantified from three to four plants per three reproductive 1118 stages and maturity (n=3-4 with 2-4 technical replicates) from WT and three transgenic 1119 genotypes A ( _sdp1-1 RNAi A_ ), B ( _sdp1-3, RNAi D_ ) and C ( _sdp1-3, RNAi C_ ). Percent Biomass 1120 lipid, starch, and oligosaccharides (raffinose and stachyose) were determined. Data represents 1121 mean ± Standard error. *p ≤ 0.05, **p ≤ 0.01, and ***p ≤ 0.001, indicate statistically significant 1122 difference based on a two-tailed Student’s t-test. 1123 1124 Supplementary table S2 1125 Fatty acid composition of WT and _SDP1-_ suppressed seeds 1126 Fatty acid composition—including palmitic acid (C16:0), stearic acid (C18:0), oleic acid (C18:1), 1127 linoleic acid (C18:2), and linolenic acid (C18:3)—was analyzed from three to four plants at each 1128 of three reproductive stages plus maturity (n = 3–4, with 2–4 technical replicates each) for WT 1129 and three transgenic genotypes: A ( _sdp1-1 RNAi A_ ), B ( _sdp1-3, RNAi D_ ) and C ( _sdp1-3, RNAi_ 1130 _C_ ). Results are expressed as mol %. Data represent mean ± standard error. *p ≤ 0.05, **p ≤ 1131 0.01, and **p ≤ 0.001 indicate statistically significant differences based on a two-tailed Student’s 1132 t-test. 1133 1134 Supplementary table S3 1135 Percent storage reserve content change (mg/seed basis) in the seeds of three _SDP1-_ 1136 suppressed genotypes: A ( _sdp1-1 RNAi A_ ), B ( _sdp1-3, RNAi D_ ) and C ( _sdp1-3, RNAi C_ ) as 1137 compared to WT at maturity (R8) indicating the overall oil yield gain due to enhanced seed size. 1138 1139 Supplemental file_S1 1140 GO annotation and DEGs of transcriptomes across developing soybean seeds from genotype B 1141 versus WT 1142 1143 Supplemental file_S2 34 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.08.765164; this version posted November 10, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.08.765164) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY 4.0 International license.](http://creativecommons.org/licenses/by/4.0/) 1144 Absolute quantification of hormone (ng/g sample) of R6 and R7 seeds from three _SDP1-_ 1145 suppressed genotypes: A ( _sdp1-1 RNAi A_ ), B ( _sdp1-3, RNAi D_ ) and C ( _sdp1-3, RNAi C_ ) as 1146 compared to WT 1147 1148 Supplemental file_S3 1149 Percent lignocellulose content change in the seeds of three _SDP1-_ suppressed genotypes: A 1150 ( _sdp1-1 RNAi A_ ), B ( _sdp1-3, RNAi D_ ) and C ( _sdp1-3, RNAi C_ ) as compared to WT at maturity 1151 (R8) 1152 1153 1154 1155 1156 1157 1158 1159 1160 1161 1162 1163 1164 1165 1166 1167 1168 **Table 1** 1169 Biomass composition of _SDP1-_ suppressed soybean seeds 1170 Soybean seed storage biomass was quantified from three to four plants per three reproductive 1171 stages and maturity (n=3-4 with 2-4 technical replicates) from WT and three transgenic 1172 genotypes A ( _sdp1-1 RNAi A_ ), B ( _sdp1-3, RNAi D_ ) and C ( _sdp1-3, RNAi C_ ). Total lipid, starch 1173 and oligosaccharides (raffinose and stachyose) were determined on a mg per seed basis. Data 1174 represents mean ± Standard error. **p ≤ 0.01, and ***p ≤ 0.001, indicate statistically significant 1175 difference based on a two-tailed Student’s t-test. 35 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.08.765164; this version posted November 10, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.08.765164) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY 4.0 International license.](http://creativecommons.org/licenses/by/4.0/) 1176 1177 1178 1179 |Developmental<br>Stages|Genotype|Storage reserves (mg/seed)|Col4|Col5|Col6| |---|---|---|---|---|---| |||Lipid|Starch|Raffinose|Stachyose| |R5.5|WT|3.91±0.21|3.70±0.11|0.00±0.00|0.00±0.00| ||A|3.82±0.05|2.95±0.08 *******|0.00±0.00|0.00±0.00| ||B|3.40±0.22******|2.97±0.10 *******|0.00±0.00|0.00±0.00| ||C|2.95±0.08 *******|2.39±0.05 *******|0.00±0.00|0.00±0.00| ||||||| |R6|WT|19.61±0.37|8.72±0.34|0.04±0.01|0.02±0.01| ||A|27.68±0.64<br>*******|12.26±0.51<br>*******|0.05±0.02|0.01±0.01| ||B|33.71±0.36<br>*******|13.70±0.83<br>*******|0.09±0.03******|0.04±0.03| ||C|28.24±0.89<br>*******|12.43±0.33<br>*******|0.09±0.02******|0.02±0.02| ||||||| |R7|WT|29.97±0.42|3.61±0.27|0.87±0.04|3.85±0.31| ||A|41.20±2.96<br>*******|3.96±0.76|1.11±0.09*******|5.90±0.69*******| ||B|44.36±1.66<br>*******|2.32±0.19*******|1.50±0.08*******|6.85±0.24*******| ||C|51.80±0.94<br>*******|4.03±0.75|1.81±0.25*******|7.54±0.82*******| ||||||| |R8|WT|39.05±0.29|0.52±0.05|1.48±0.09|6.51±0.22| ||A|52.18±0.65<br>*******|0.56±0.02|1.70±0.10******|7.41±0.42*******| ||B|50.78±1.62<br>*******|0.47±0.06|1.68±0.03******|4.98±0.22*******| ||C|49.49±0.31<br>*******|0.51±0.05|1.49±0.04|4.22±0.12*******| 1180 **Table 2** 1181 Relative levels of total cytokinins, total auxins, abscisic acid (ABA), 12-oxo-phytodecanoic acid 1182 (OPDA), Jasmonoyl-isoleucine (Ja-Ile), and salicylic acid (SA), in the three _SDP1-_ suppressed 1183 transgenic genotypes A ( _sdp1-1 RNAi A_ ), B ( _sdp1-3, RNAi D_ ) and C ( _sdp1-3, RNAi C_ ) as 1184 compared to WT during R6 (mid) and R7 (late) development stages. Data represents mean ± 1185 standard error (n=3-4 plants/genotype). *p ≤ 0.05, **p ≤ 0.01, ***p ≤ 0.001 indicates a 1186 statistically significant difference based on a two-tailed t-test. 1187 36 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.08.765164; this version posted November 10, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.08.765164) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY 4.0 International license.](http://creativecommons.org/licenses/by/4.0/) |Hormones|DevelopmentalStages|Col3|Col4|DevelopmentalStages|Col6|Col7| |---|---|---|---|---|---|---| ||R6|R6|R6|R7|R7|R7| ||A|B|C|A|B|C| |Total cytokinins|0.98±0.10|0.91±0.09|0.75±0.02|0.59±0.06******|0.67±0.11|0.71±0.15| |Total auxins|0.99±0.13|0.92±0.09|1.18±0.22|1.54±0.24|0.95±0.24|1.90±0.48| |ABA|0.80±0.12|0.42±0.09 ******|0.56±0.07 ******|1.96±0.32|0.98±0.32|1.69±0.37| |OPDA|0.66±0.13|0.86±0.13|0.72±0.13|0.22±0.02 *******|0.51±0.10 *****|0.33±0.05 *******| |Ja-Ile|0.64±0.08*****|0.53±0.12*****|0.27±0.02******|0.96±0.01 *****|1.05±0.07|0.91±0.03 *****| |SA|1.27±0.15|1.32±0.45|1.28±0.39|0.60±0.04 ******|0.86±0.09|0.64±0.02******| 1188 1189 37 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.08.765164; this version posted November 10, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.08.765164) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY 4.0 International license.](http://creativecommons.org/licenses/by/4.0/) Figure 1 Phenotypic and agronomic measurements of WT and _SDP1-_ suppressed soybean genotypes The total number of nodes (A), pods (B), and seeds (C) were determined from four-six plants each for WT, A ( _sdp1-1 RNAi A_ ), B ( _sdp1-3, RNAi D_ ) and C ( _sdp1-3, RNAi C_ ), labeled here as _sdp1._ Average individual seed weight was calculated from 100 seeds per each of four-sixplants per genotype (D). The correlation between average individual seed weight and seed number (E) was determined using the Pearson correlation coefficient (PCC) and the p-value was determined using the corresponding t-distribution of the PCC. Average individual seed area was measured using RGB images of 10 seeds from each of four-six plants per genotype, processed with PlantCV software (F). Box-plot center lines indicate the median, ‘x’ represent the mean; box limits, 25 [th] and 75 [th] percentiles; whiskers, 1.5 times the interquartile range from the 25 [th] and 75 [th] percentiles. ***p ≤ 0.001 indicates a statistically significant difference based on a two-tailed Student’s t-test. 1 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.08.765164; this version posted November 10, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.08.765164) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY 4.0 International license.](http://creativecommons.org/licenses/by/4.0/) Figure 2 Cellular morphology of WT and _SDP1-_ suppressed soybean genotypes Confocal microscopy imaging of three to five matured soybean seeds from each of WT and three _SDP1-_ suppressed genotypes A ( _sdp1-1 RNAi A_ ), B ( _sdp1-3, RNAi D_ ) and C ( _sdp1-3,_ _RNAi C_ ), labeled here as _sdp1_, bar=50.01 um (A) Density histogram using 4-10 images per seed from three to five seeds per each of WT and A (B) WT and B (C) and WT and C (D) indicated a shift towards bigger cell area in the transgenic seeds as compared to WT controls. Cell area (E) and cell number (F) of _SDP1-_ suppressed and WT genotypes (F). Data represents mean ± standard error. ***p ≤ 0.001 indicates a statistically significant difference based on a two-tailed Student’s t-test. 2 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.08.765164; this version posted November 10, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.08.765164) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY 4.0 International license.](http://creativecommons.org/licenses/by/4.0/) Figure 3 Fresh weight and dry weight measurement of soybean seeds Fresh weight (FW) and dry weights (DW) measurements of seeds from four individual plants of the three _SDP1-_ suppressed genotypes A ( _sdp1-1 RNAi A_ ), B ( _sdp1-3, RNAi D_ ) and C ( _sdp1-_ _3, RNAi C_ ), labeled here as _sdp1_ and WT. Data represents mean ± standard error. **p ≤ 0.01 ***p ≤ 0.001 indicates a statistically significant difference based on a two-tailed Student’s t-test. 3 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.08.765164; this version posted November 10, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.08.765164) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY 4.0 International license.](http://creativecommons.org/licenses/by/4.0/) Figure 4 Stage specific differential transcriptomic responses to _SDP1_ suppression Soybean seeds from three to four plants per three reproductive stages (R5.5, R6 and R7) from WT and the transgenic genotype B ( _sdp1-3, RNAi D_ ), was used to determine transcriptomic responses to _SDP1_ suppression. Principal component analysis (PCA) (A) Number of differentially expressed genes (DEGs) (B) and Selected Gene Ontology (GO) term enrichment analysis of R7 seeds (C). The colors of the lines represent the –log10(Fisher.weight01) pvalue as calculated by the topGO R package. 4 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.08.765164; this version posted November 10, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.08.765164) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY 4.0 International license.](http://creativecommons.org/licenses/by/4.0/) Figure 5 PEPCK activity across soybean seed development PEPCK activity, as a signature of gluconeogenesis, over the course of seed development from four individual plants of the three _SDP1-_ suppressed genotypes A ( _sdp1-1 RNAi A_ ), B ( _sdp1-3,_ _RNAi D_ ) and C ( _sdp1-3, RNAi C_ ), labeled here as _sdp1_ and WT. Box-plot center lines indicate the median, ‘x’ represent the mean; box limits, 25 [th] and 75 [th] percentiles; whiskers, 1.5 times the interquartile range from the 25 [th ] and 75 [th ] percentiles. *p ≤ 0.05, **p ≤ 0.01 indicates a statistically significant difference based on a two-tailed Student’s t-test. 5 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.08.765164; this version posted November 10, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.08.765164) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY 4.0 International license.](http://creativecommons.org/licenses/by/4.0/) Figure 6 Lignocellulose content Lignocellulose wall content from four plants each of the three _SDP1-_ suppressed genotypes A ( _sdp1-1 RNAi A_ ), B ( _sdp1-3, RNAi D_ ) and C ( _sdp1-3, RNAi C_ ), labeled here as _sdp1_ and WT. Data represents mean ± standard error. *p ≤ 0.05, **p ≤ 0.01, ***p ≤ 0.001 indicates a statistically significant difference based on a two-tailed Student’s t-test. 6 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.08.765164; this version posted November 10, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.08.765164) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY 4.0 International license.](http://creativecommons.org/licenses/by/4.0/) Supplementary figure S1 Phenotypic and agronomic measurements of WT and _SDP1-_ suppressed soybean genotypes. The correlation between total number of pods and nodes (A) and total numbers of seeds and pods (B) was determined using the Pearson correlation coefficient. P-value was obtained using the corresponding t-distribution of the Pearson correlation coefficient. Mature seed images were taken using Nikon D7200 camera and a color checker card is introduced as reference for color correction and white balance (C). Seed area calculation using the morphology package in PlantCV after masking the seed from the background (D). 7 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.08.765164; this version posted November 10, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.08.765164) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY 4.0 International license.](http://creativecommons.org/licenses/by/4.0/) 8 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.08.765164; this version posted November 10, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.08.765164) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY 4.0 International license.](http://creativecommons.org/licenses/by/4.0/) Supplementary figure S2 Across development expression of F-box family protein Soybean seeds from three to four plants per three reproductive stages (R5.5, R6 and R7) from WT and the transgenic genotype B ( _sdp1-3, RNAi D_ ), was used to determine transcriptomic responses to _SDP1_ suppression. Expression of five family members of the F-box family protein Glyma.02G018450 (A), Glyma.04G042900 (B), Glyma.08G252300 (C), Glyma.17G237800 (D) and Glyma.19G196400 (E). Each point represents one sample in the RNA-seq dataset;boxplots show the interquartile range, and whiskers extend to the most extreme non-outlier values. Log 2 fold changes were shrunk using the adaptive shrinkage estimator (ashr; Stephens 2016) implemented in lfcShrink(). *p~adj~ ≤ 0.05, **p~adj~ ≤ 0.01, ***p~adj~ ≤ 0.001, based on DESeq2 Wald test p-values adjusted for multiple testing (Benjamini-Hochberg). 9 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.08.765164; this version posted November 10, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.08.765164) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY 4.0 International license.](http://creativecommons.org/licenses/by/4.0/) 10 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.08.765164; this version posted November 10, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.08.765164) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY 4.0 International license.](http://creativecommons.org/licenses/by/4.0/) Supplementary figure S3 Across development expression of RING/U-box superfamily protein Soybean seeds from three to four plants per three reproductive stages (R5.5, R6 and R7) from WT and the transgenic genotype B ( _sdp1-3, RNAi D_ ), was used to determine transcriptomic responses to _SDP1_ suppression. Expression of ten family members of the RING/U-box superfamily protein Glyma.03G179300 (A), Glyma.06G150400 (B), Glyma.07G230600 (C), Glyma.09G216900 (D), Glyma.12G047300 (E), Glyma.12G173300 (F), Glyma.13G139400 (G), Glyma.16G076000 (H), Glyma.17G009900 (I) and Glyma.19G1800000 (J).Each point represents one sample in the RNA-seq dataset; boxplots show the interquartile range, and whiskers extend to the most extreme non-outlier values. Log 2 fold changes were shrunk using the adaptive shrinkage estimator (ashr; Stephens 2016) implemented in lfcShrink(). *p~adj~ ≤ 0.05, **p~adj~ ≤ 0.01, based on DESeq2 Wald test p-values adjusted for multiple testing (Benjamini-Hochberg). 11 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.08.765164; this version posted November 10, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.08.765164) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY 4.0 International license.](http://creativecommons.org/licenses/by/4.0/) Supplementary figure S4 Across development expression of two Protein phosphatase 2C family protein Soybean seeds from three to four plants per three reproductive stages (R5.5, R6 and R7) from WT and the transgenic genotype B ( _sdp1-3, RNAi D_ ), was used to determine transcriptomic responses to _SDP1_ suppression. Expression of two family members of the Protein phosphatase 2C family protein Glyma.14G082200 (A) and Glyma.U032205 (B). Each point represents one sample in the RNA-seq dataset; boxplots show the interquartile range, and whiskers extend to the most extreme non-outlier values. Log 2 fold changes were shrunk using the adaptive shrinkage estimator (ashr; Stephens 2016) implemented in lfcShrink(). *p~adj~ ≤ 0.05, **p~adj~ ≤ 0.01, ***p~adj~ ≤ 0.001, based on DESeq2 Wald test p-values adjusted for multiple testing (BenjaminiHochberg). 12