TheO-Fucosyltransferase SPINDLY Attenuates Auxin-Induced Fruit Growth by Inhibiting ARF6 and ARF8 binding to Coactivator Mediator Complex inArabidopsis

30 31 The phytohormone auxin plays a pivotal role in promoting fruit initiation and growth upon 32 fertilization in flowering plants. Upregulation of auxin signaling by genetic mutations or 33 exogenous auxin treatment can induce seedless fruit formation from unpollinated ovaries, termed 34 parthenocarpy. Recent studies suggested that the class A AUXIN RESPONSE FACTOR6 35 (ARF6) and ARF8 in Arabidopsis play dual functions by first inhibiting fruit initiation when 36 complexed with unidentified corepressor IAA protein(s) before pollination, and later promoting 37 fruit growth after fertilization as ARF dimers. However, whether and how posttranslational 38 modification(s) regulate ARF6- and ARF8-mediated fruit growth were unknown. In this study, 39 we reveal that both ARF6 and ARF8 are O-fucosylated in their middle region (MR) by 40 SPINDLY (SPY), a novel nucleocytoplasmic protein O-fucosyltransferase, which catalyzes the 41 addition of a fucose moiety to specific Ser/Thr residues of target proteins. Epistasis, biochemical 42 and transcriptome analyses indicated that ARF6 and ARF8 are downstream of SPY, but ARF8 43 plays a more predominant role in parthenocarpic fruit growth. Intriguingly, two ARF6/8-44 interacting proteins, the co-repressor IAA9 and MED8, a subunit of the coactivator Mediator 45 complex, were also O-fucosylated by SPY. Biochemical assays demonstrated that SPY-mediated 46 O-fucosylation of these proteins reduced ARF-MED8 interaction, which led to enhanced 47 transcription repression activity of the ARF6/8-IAA9 complex but impaired transactivation 48 activities of ARF6/8. Our study unveils the role of protein O-fucosylation by SPY in attenuating 49 auxin-triggered fruit growth through modulation of activities of key transcription factors, a co-50 repressor and the coactivator MED complex. 51 52 53 54 .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.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 The copyright holder for this preprintthis version posted June 27, 2024. ; https://doi.org/10.1101/2024.06.26.599170doi: bioRxiv preprint 3

55 56 In flowering plants, the developmental transition from ovary to fruit, termed fruit 57 initiation or fruit set, usually occurs after fertilization. Phytohormones play key roles in 58 controlling fruit set, growth and maturation1-4. Among them, auxin synthesized in developing 59 seeds promote fruit set and subsequent growth. Treatment of the unfertilized ovary with auxin 60 can induce parthenocarpy, i.e., seedless fruit formation without fertilization, indicating that 61 activation of auxin signaling is essential for fruit initiation5-7. Parthenocarpy is often a beneficial 62 trait in crops. Seedless fruits are preferred by the consumers, and this trait is also associated with 63 flavor improvement and longer shelf life8. Moreover, it offers more consistent fruit yield in 64 variable environmental conditions such as elevated temperatures that can severely reduce pollen 65 viability and limit pollinator availability both of which can limit fruit production9. Elucidation of 66 the regulatory mechanism of auxin-induced parthenocarpy will contribute to improving yield 67 stability under stressful climate conditions. 68 69 The nuclear auxin signaling pathway consists of three families of major components, (1) 70 auxin coreceptors TRANSPORT INHIBITOR 1/AUXIN SIGNALING F-BOX (TIR1/AFB), (2) 71 transcription co-repressors Auxin/INDOLE-3-ACETIC ACID (Aux/IAA), and (3) AUXIN 72 RESPONSE FACTOR (ARF) transcription factors10-14. Aux/IAA proteins (thereafter abbreviated 73 as IAAs) function as negative regulators of auxin signaling by binding to ARFs at their target 74 promoters. Formation of the IAA-ARF complexes leads to the recruitment of co-repressor 75 TOPLESS (TPL), which represses transcription by preventing ARF binding to the coactivator 76 Mediator complex15. TPL may also recruit the CDK8 kinase module (CKM), a repressive 77 module of the Mediator complex, to block transcription of ARF target genes16. When auxin 78 levels increase, it promotes TIR1/AFB-IAA interactions, thereby triggering degradation of IAA 79 proteins to release ARFs to activate the downstream auxin response pathway. AFBs, IAAs and 80 ARFs all belong to multi-gene families in angiosperms, whereas much reduced gene redundancy 81 of these auxin signaling components was found in early-emerging land plants, e.g., 82 Physcomitrella patens (moss) and Marchantia polymorpha (liverwort)17. IAA proteins contain 83 three domains: Domain I includes a transcriptional repression EAR motif that recruits TPL, 84 domain II contains an interaction motif for TIR1/AFB, and the conserved C-terminal Phox and 85 .CC-BY-NC-ND 4.0 International licenseavailable under a (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 The copyright holder for this preprintthis version posted June 27, 2024. ; https://doi.org/10.1101/2024.06.26.599170doi: bioRxiv preprint 4 Bem1 (PB1) domain is required for multimerization between IAAs and ARFs10. ARFs contain a 86 conserved N-terminal DNA binding domain (DBD) for binding auxin response elements 87 (AuxREs) and ARF dimerization, a middle region (MR) and the PB1 domain. ARFs can be 88 divided into two functional groups based on their amino acid composition in the middle region 89 (MR) and their activities as transcription activators or repressors in transient expression assays. 90 The MRs of class A ARFs (also called activator ARFs) have a high glutamine content while 91 classes B and C ARFs (also called repressor ARFs) are rich in serine, threonine and proline 92 residues18,19. The canonical AFB/IAA/ARF signaling cascade described above is mainly based 93 on class A ARFs whose members include ARF5, ARF6, ARF7, ARF8 and ARF19 in 94 Arabidopsis thaliana. Recent studies in moss and liverwort suggest that class B ARFs may 95 function as transcription repressors by competing with class A ARFs for binding to auxin-96 responsive promoters20,21. Class C ARFs are also transcription repressors, although they may 97 regulate auxin-independent development21. 98 99 The mechanism of auxin-induced fruit set has been studied by genetic analyses. In 100 multiple species, mutations or silencing of class A ARF(s) leads to parthenocarpic fruits from 101 emasculated flowers, supporting the repressive role of class A ARFs in ovary-derived fruit 102 set1,7,22-25. However, by characterizing higher order mutant combinations of four class A SlARFs 103 (SlARF5, SlARF7, SlARF8A, SlARF8B) in tomato (Solanum lycopersicum), we recently 104 demonstrated that these class A SlARFs display dual functions during fruit development26. 105 Before pollination, all four SlARFs act as inhibitors of fruit set when associated with SlIAA9. 106 After fertilization, the elevated auxin levels in the ovary result in SlIAA9 degradation and free 107 the class A SlARFs to function as activators in subsequent fruit growth. The positive role of 108 SlARFs in fruit growth is reflected by the biphasic bell-shape curve of parthenocarpic fruit size 109 in response to varying doses of these SlARFs. The maximum parthenocarpic fruit size was 110 reached by removal of SlARF8A and SlARF8B, while knocking out all four SlARFs abolished 111 fruit growth completely. Consistent with this idea, expression of truncated SlARF8A or 8B 112 lacking the SlIAA9-interacting PB1 domain in transgenic tomato resulted in the production of 113 large seedless fruits26. Similarly, AtARF6 and AtARF8 in Arabidopsis also showed dual role in 114 fruit initiation/growth, although the specific IAA(s) regulating this process have not been 115 identified27. The arf6 and arf8 single mutants produced longer and wider fruits from emasculated 116 .CC-BY-NC-ND 4.0 International licenseavailable under a (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 The copyright holder for this preprintthis version posted June 27, 2024. ; https://doi.org/10.1101/2024.06.26.599170doi: bioRxiv preprint 5 unfertilized flowers22,27. In contrast, the arf6 arf8 double mutant produced short pistils with 117 similar length as that of WT without fertilization27. These observations suggest that although 118 both ARF6 and ARF8 inhibit fruit set before anthesis, they then promote fruit elongation after 119 fertilization. 120 121 Class A ARFs in Arabidopsis have been shown to be dynamically regulated by post-122 translational modifications (PTMs), including phosphorylation, Small Ubiquitin-Like Modifier 123 (SUMO)-modification (SUMOylation), and ubiquitination18,28. Phosphorylation of ARF7/19 124 MRs, which is induced by a peptide hormone TRACHEARY ELEMENT DIFFERENTIATION 125 INHIBITORY FACTOR (TDIF), enhances ARF DNA binding affinity and reduces interaction 126 with IAA proteins to promote lateral root formation29. SUMOylation of ARF7 contributes to 127 hydropatterning of the Arabidopsis root in response to moisture by promoting ARF7-IAA3 128 interaction to inhibit lateral root initiation specifically in dry environments30. ARF6 129 ubiquitination promotes its degradation in response to abscisic acid (ABA) treatment in 130 Arabidopsis seedlings31. However, whether and how PTMs regulate ARF6- and ARF8-mediated 131 fruit set/growth were unknown. In this study, we reveal that both ARF6 and ARF8 are O-132 fucosylated by SPINDLY (SPY), a novel nucleocytoplasmic protein O-fucosyltransferase 133 (POFUT), which catalyzes the addition of a fucose moiety to the hydroxyl oxygen of specific 134 Ser/Thr residues of target proteins32. Importantly, O-fucosylation of MRs of ARF6/8 reduced 135 their binding to MED8, a subunit of the coactivator Mediator complex. The pleiotropic 136 phenotypes of the spy mutants and recent discovery of hundreds of SPY target proteins by 137 proteomic studies all point to the important roles of SPY in regulating diverse cellular 138 processes33-35. But the molecular mechanism of SPY regulation has only been defined for a 139 handful of its targets32,34,36-39. Here, we demonstrated the role of SPY in attenuating auxin-140 induced fruit set and growth by O-fucosylating ARF6 and ARF8, and their interacting proteins, 141 co-repressor IAA9 and MED8 of the coactivator Mediator complex. 142 143 144

145 146 spy mutants displayed parthenocarpic fruit growth 147 .CC-BY-NC-ND 4.0 International licenseavailable under a (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 The copyright holder for this preprintthis version posted June 27, 2024. ; https://doi.org/10.1101/2024.06.26.599170doi: bioRxiv preprint 6 148 Genetic and proteomic analyses indicate that protein O-fucosylation by the nucleocytoplasmic-149 localized SPY regulates diverse developmental processes in Arabidopsis, although most of the 150 molecular mechanisms are unknown33-35,40. Two previous mutant studies investigated the effect 151 of spy mutations on parthenocarpy with contradicting results. One study reported that spy-2 and 152 spy-3 in the Col-0 ecotype displayed parthenocarpy after emasculation 41, while another study 153 did not observe any parthenocarpy phenotype in spy-3 (Col-0 ecotype) or spy-4 (Ws ecotype)42. 154 To clarify whether SPY plays a role in parthenocarpy, we examined the pistil phenotypes of four 155 spy alleles, including spy-8 and spy-19 in the Ler background and spy-3 and spy-23 in the Col-0 156 background. All four spy mutants showed significantly longer pistils from emasculated flowers 157 comparing to those of wild-type controls (Fig. 1a-1d). Furthermore, the pistil phenotype of spy-158 3 was partially rescued by introduction of PSPY:GFP-SPY or PSPY:GFP-SPY-NLS (nuclear 159 localization sequence), but not by PSPY:GFP-SPY-NES (nuclear export sequence) (Fig. 1e-1f, 160 Supplementary Fig. 1a-1b). These results indicate that nuclear-localized SPY inhibits fruit 161 initiation and elongation before pollination. Consistent with this notion, we found that SPY 162 protein levels were reduced after anthesis as detected using a PSPY:FLAG-SPY spy-3 transgenic 163 line (Supplementary Fig. 1c). We also examined overall protein O-fucosylation before and 164 after anthesis by protein blot analysis using a biotinylated Aleuria aurantia lectin (AAL), a 165 terminal fucose-specific lectin. The PSPY:FLAG-SPY spy-3 line showed reduced protein O-166 fucosylation at 3 DAA and 5 DAA compared to that at –2 DAA and 0 DAA (Supplementary 167 Fig. 1d). 168 169 arf6 showed additive interaction with spy, whereas arf8 was epistatic to spy in 170 parthenocarpic growth 171 172 The arf6 and arf8 mutants were shown previously to display longer and wider fruit after 173 emasculation comparing to that of WT 27. Consistent with the previous report, we found that both 174 arf6-2 and arf8-3 showed increased fruit length and width after emasculation, although arf8 175 displayed stronger parthenocarpic growth than arf6 (Fig. 2a, 2b, and Supplementary Fig. 2a). 176 These pistil phenotypes were partially rescued by PARF6/8:FLAG-ARF6/8 (Supplementary Fig. 177 3a, 3b). Another class A ARF, ARF7 (also known as NPH4), is also expressed in the developing 178 .CC-BY-NC-ND 4.0 International licenseavailable under a (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 The copyright holder for this preprintthis version posted June 27, 2024. ; https://doi.org/10.1101/2024.06.26.599170doi: bioRxiv preprint 7 pistil (Arabidopsis eFP Browser43, http://bar.utoronto.ca/efp_arabidopsis/cgi-bin/efpWeb.cgi). 179 However, the nph4-1 mutation alone or in combination with arf6 and/or arf8 did not increase 180 fruit growth after emasculation (Supplementary Fig. 2b, 2c), suggesting that ARF7 is not a 181 major regulate of this process. 182 To examine the genetic interaction between SPY and ARF6/8, epistasis analysis was 183 performed among spy-3, arf6-2 and arf8-3 using single, double, triple mutants. Most of the 184 mutant alleles are homozygous, except that mutants containing both arf6 and arf8 mutations 185 were sesquimutants (arf6 -/- arf8 +/- and arf8 -/- arf6 +/-) because the arf6 arf8 double 186 homozygote is sterile. The spy, arf6 and arf8 single homozygous mutants produced longer pistils 187 than that of WT with arf8 displayed the strongest phenotype (Fig. 2a-2b). Importantly, we found 188 that spy and arf6 additively promoted parthenocarpic fruit elongation in the spy arf6 double 189 homozygous mutant. Similarly, in the arf6 arf8+/- sesquimutant background, the spy mutation 190 further increased fruit length. However, the homozygous double mutant spy arf8 showed similar 191 fruit length as that of arf8. In the arf8 arf6+/- sesquimutant background, the spy mutation did not 192 further enhance fruit growth either. These results indicated that arf8 is epistatic to spy, 193 suggesting that ARF8 is downstream of SPY in regulating parthenocarpy. The additive 194 interaction between spy and arf6 in parthenocarpic growth is likely because both ARF6 and 195 ARF8 are downstream of SPY, but ARF8 plays a more predominant role in this process. This 196 notion is further supported by our results in the next section showing that both ARF6 and ARF8 197 are O-fucosylated by SPY. 198 199 Besides auxin, previous studies show that another phytohormone gibberellin (GA) also 200 promotes fruit initiation and growth in Arabidopsis. The quadruple or global della mutant 201 (knockout four or all 5 DELLA genes) displays strong parthenocarpy6,44. Because SPY represses 202 GA responses in hypocotyl elongation by enhancing DELLA activity via O-fucosylation, we 203 tested whether SPY’s regulation of parthenocarpy is through its repression of both GA and auxin 204 responses by comparing GA vs auxin responses in parthenocarpic fruits of WT and different 205 mutant backgrounds. Application of GA3 or the auxin analog picloram45 to pistils of emasculated 206 WT flowers promoted parthenocarpic fruit growth (Fig. 2c-2d). Treatment of GA or auxin also 207 increased parthenocarpic fruit growth in spy, arf6, and spy arf6. However, only GA treatment, 208 but not auxin, increased fruit growth in arf8 or spy arf8. These results indicate that SPY 209 .CC-BY-NC-ND 4.0 International licenseavailable under a (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 The copyright holder for this preprintthis version posted June 27, 2024. ; https://doi.org/10.1101/2024.06.26.599170doi: bioRxiv preprint 8 represses parthenocarpy by regulating both GA and auxin responses, and that ARF8 plays a more 210 predominant role than ARF6 in mediating auxin-induced parthenocarpic growth downstream of 211 SPY. 212 213 ARF6 and ARF8 are O-fucosylated by SPY 214 Based on the genetic interaction between spy and arf6/8, we tested whether ARF6 and 215 ARF8 are direct targets of SPY. By transient co-expression of FLAG-ARF6/ARF8 with SPY in 216 Nicotiana benthamiana, we found that both affinity-purified FLAG-ARF6 and -ARF8 were O-217 fucosylated as detected by biotinylated fucose-specific lectin, AAL (Supplementary Fig. 4a-218 4b). To confirm that ARF6 and ARF8 are O-fucosylated in Arabidopsis, we generated transgenic 219 Arabidopsis lines carrying either PUBQ10:FLAG-ARF6 or PUBQ10:FLAG-ARF8, and then 220 introduced these transgenes into the spy-8 background. Affinity purified FLAG-ARF6 and -221 ARF8 proteins from the WT background, but not those from the spy-8 background, were O-222 fucosylated as detected by AAL-biotin (Fig. 3a-3b). To identify the O-fucosylation (Fuc) sites in 223 these proteins, affinity-purified FLAG-ARF6 and -ARF8 from N. benthamiana and transgenic 224 Arabidopsis were analyzed by liquid chromatography (LC)-electrospray ionization (ESI)-mass 225 spectrometry (MS). One O-fucosylated ARF6 peptide and three O-fucosylated ARF8 peptides 226 were identified (Fig. 3c-3d, Supplementary Table 1, Supplementary Data Sets 1-4). Notably, 227 all identified O-Fuc sites are in the MR of ARF6/8. We also performed an AAL pulldown assay 228 using N. benthamiana that transiently expressed full-length or truncated ARF6 proteins in the 229 presence or absence of SPY. Besides the full-length ARF6, only the MR fragment (amino acid 230 residues 376-778) but not the N-terminal DBD-containing fragment (amino acid residues 1-375) 231 or C-terminal PB1 domain (amino acid residues 779-935) was O-fucosylated (Fig. 3e). These 232

support that the major O-Fuc sites in ARF6 and ARF8 reside within their MR sequence. 233 234 In addition to O-fucosylation, we found that both ARF6 and ARF8 contained two other 235 types of PTMs, phosphorylation and O-link-N-acetylglucosamine (GlcNAc) modification 236 (Supplementary Table 1 and Supplementary Fig. 5). Intracellular protein O-GlcNAcylation is 237 catalyzed by an O-GlcNAc transferase (OGT), SECRET AGENT (SEC) that is a paralog of SPY 238 in Arabidopsis46. Recent proteomic studies showed that SPY and SEC have unique targets as 239 well as common targets in Arabidopsis34,35,47,48. The interplay between SPY and SEC is complex 240 .CC-BY-NC-ND 4.0 International licenseavailable under a (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 The copyright holder for this preprintthis version posted June 27, 2024. ; https://doi.org/10.1101/2024.06.26.599170doi: bioRxiv preprint 9 as these two types of glycosylation can interact antagonistically or additively, depending on the 241 target proteins32,38,46,49. We found that the sec mutants produced short pistils after emasculation 242 with similar length as that of WT (Supplementary Fig. 4c, 4d), suggesting that O-243 GlcNAcylation does not alter the function of ARF6/8 during fruit set significantly. Therefore, we 244 focused on the regulatory mechanism of SPY on ARF6/ARF8 in the rest of this study. 245 246 SPY and ARF6/8 regulate many common target genes during fruit set 247 248 Our genetic and biochemical analyses indicated that ARF6 and ARF8 are downstream of 249 SPY in regulating parthenocarpy. To identify ARF6-, ARF8- and SPY-responsive genes that are 250 involved in fruit set and growth, transcriptome analysis was performed by RNA-seq using the 251 following samples: (1) unpollinated pistils at 2 days before anthesis (−2 DAA, from stage 10 252 flowers) of arf6-2, arf8-3, spy-3 and WT (Col-0); and (2) 0 DAA WT pistils (from stage 14 253 flowers) that were already self-pollinated. Two biological repeats were included in each set of 254 samples. The differentially expressed gene (DEG) lists for ARF6, ARF8 or SPY-responsive 255 genes (108, 1143, 271 DEGs, respectively) were identified by comparing each mutant vs WT (– 256 2 DAA) dataset using the criteria of fold change > 1.5 and p < 0.05 (Fig. 4a, Supplementary 257 Fig. 6a, Supplementary Table 2). The DEG list for fertilization-responsive genes (2528 total) 258 was generated by comparing 0 DAA WT vs –2 DAA WT dataset with fold change > 1.5 and p < 259 0.05 (Supplementary Fig. 6a, Supplementary Table 2). Consistent with the stronger 260 parthenocarpic phenotype of arf8, the arf8 mutation resulted in altered transcript levels of many 261 more genes than arf6, and almost all ARF6 DEGs (99 out of 108 total) were included in ARF8 262 DEG list (Fig. 4a). Comparison of DEG lists for SPY, ARF6 and ARF8 revealed that 67% of 263 SPY DEGs (181 DEGs out of 271 total) were co-regulated by ARF6 and/or ARF8 (Fig. 4a, 264 Supplementary Table 3). Among these co-regulated DEGs, 67 genes were up-regulated by both 265 spy-3 and arf8 (Fig. 4b, 4d, Supplementary Table 3), and 103 genes were down-regulated by 266 both spy-3 and arf8 (Fig. 4c, 4d, Supplementary Table 3). Comparison of SPY- and ARF8-267 responsive genes with fertilization-responsive genes (0 DAA WT vs –2 DAA WT) further 268 showed that 62% of SPY- and ARF- coregulated genes were also fertilization-responsive genes 269 (113 DEGs out of 181 total, Supplementary Fig. 6b-6e). The significant overlap among the 270 SPY-, ARF6/ARF8-, and fertilization-responsive genes provide strong support that SPY 271 .CC-BY-NC-ND 4.0 International licenseavailable under a (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 The copyright holder for this preprintthis version posted June 27, 2024. ; https://doi.org/10.1101/2024.06.26.599170doi: bioRxiv preprint 10 regulates fruit set at least in part mediated by ARF6 and ARF8. 272 273 Gene Ontology (GO) analysis of SPY- and ARF6/8-coregulated DEGs identified 48 274 enriched GO terms (Supplementary Table 4). Although the coregulated DEGs belong to a 275 variety of biological processes, certain groups were more represented, including response to 276 sugar, fatty acid, nutrient and hormone, and biogenesis related genes for carbohydrate/amino 277 acid metabolic processes (Fig. 4e). Seven SPY- and ARF-coregulated genes were selected to 278 verify their expression in –2 DAA WT, spy-3 and arf8, and 0 DAA WT by RT–qPCR 279 (Supplementary Table 5). Among them, three genes encode transcription repressor or 280 transcription factor: IAA19 in auxin signaling50,51, ETHYLENE RESPONSE FACTORs (ERF107 281 and ERF023) that are responsive to oxylipins52. In addition, ERF107 regulates nitrate 282 assimilation53, and ERF023 is responsive to cellular nitrogen status54. Two genes are involved in 283 sugar response/metabolic process, and their expression is repressed by sugar: BASIC LEUCINE-284 ZIPPER 1 (bZIP1) encoding a S1 subgroup bZIP transcription factor55,56 and GNTL encoding a 285 beta-1,6-N-acetylglucosaminyl transferase-like enzyme for the synthesis of glycan and/or 286 glycosylation of proteins57-59. ARABINOGALACTAN PROTEIN 12 (AGP12) was implicated in 287 nutrient uptake in developing seeds60, and SHORT LIFE (SHL) encodes a plant-homeodomain 288 (PHD) protein that functions as a histone reader for chromatin remodeling in repressing floral 289 induction and regulating fertility61-63. The RT–qPCR assays confirmed that AGP12, ERF023 and 290 IAA19 were upregulated, whereas GNTL, ERF107, bZIP1 and SHL were downregulated in spy-3, 291 arf8 or by fertilization (WT 0 DAA) in comparison to that in WT –2 DAA (Fig. 5a, 5b). Up-292 regulation of IAA19 expression by arf8 and spy reflects elevated auxin response in these mutant 293 pistils. AGPs are known to function in cell wall reorganization57. Induction of AGP12 by arf8 294 and spy is consistent with the proposed role of this gene in nutrient uptake of ovule based on its 295 expression at the chalaza of the ovule and funiculus60. Auxin has been shown to promote sugar 296 transport and metabolism in ovaries after fertilization in several species64-66, which in turn 297 inhibits fruit abortion caused by programmed cell death and promotes fruit growth64. In 298 Arabidopsis pistils at –2 DAA, we found that two sugar-repressed genes bZIP1 and GNTL were 299 downregulated by arf8 and spy, suggesting that these two mutations led to an increased sugar 300 content/signaling and that bZIP1 and GNTL may play a negative role in fruit set/growth in WT 301 .CC-BY-NC-ND 4.0 International licenseavailable under a (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 The copyright holder for this preprintthis version posted June 27, 2024. ; https://doi.org/10.1101/2024.06.26.599170doi: bioRxiv preprint 11 before anthesis. Consistent with this idea, two bZIP1 knockout mutants displayed longer pistils 302 after emasculation comparing to WT (Supplementary Fig. 7a, 7b). 303 304 To identify direct target genes of ARF6/ARF8 that are coregulated by SPY, we compared 305 our SPY- and ARF6/ARF8-responsive DEG list (181 total) with a published ARF6 ChIP-seq 306 dataset generated using seedling samples51. This comparison identified 51 overlapping genes as 307 direct targets of ARF6/ARF8, including six of the selected seven genes (except bZIP1) verified 308 by RT-qPCR (Supplementary Table 6, Supplementary Fig. 6f). Although bZIP1 is not present 309 in the published ARF6 ChIP-seq dataset, its promoter region contains tandem AuxREs. It is 310 possible that bZIP1 is a direct ARF target in the pistil, but not at the seedling stage. ChIP-qPCR 311 assay was performed to verify direct binding of ARF6/8 to the promoters of these genes. 312 Transgenic Arabidopsis seedlings carrying either PUBQ10:FLAG-ARF6 or PUBQ10:FLAG-ARF8 in 313 the WT background were used for chromatin crosslinking and immunoprecipitation using anti-314 FLAG beads. The non-transgenic WT was included as the negative control. qPCR was 315 performed using primers that span the ARF6 binding peaks in the promoter of each target gene, 316 except for bZIP1 where primers span two tandem AuxREs in its promoter. The ChIP-qPCR 317 analysis showed a significant enrichment in all seven target promoters in the PUBQ10:FLAG-ARF6 318 and PUBQ10:FLAG-ARF8 lines compared to the WT control (Fig. 5c, 5d), supporting that these 319 genes are direct targets of ARF6 and ARF8. 320 321 SPY did not affect ARF6 or ARF8 DNA binding, protein stability or nuclear localization 322 Because SPY and ARF6/ARF8 coregulate many target genes during fruit set, we 323 hypothesized that O-fucosylation of ARF6 and ARF8 by SPY may enhance ARF6/8 function to 324 inhibit fruit set. To investigate the molecular mechanism involved, we first examined whether 325 this posttranslational modification increases binding of ARF6/ARF8 to their target promoters by 326 ChIP-qPCR using transgenic PUBQ10:FLAG-ARF6 or PUBQ10:FLAG-ARF8 lines in WT or spy 327 background. We did not observe significant reduction in enrichment of target promoter 328 sequences by the spy mutation (Fig. 5c, 5d), suggesting that SPY does not alter the DNA binding 329 activities of ARF6/8. Recent studies have reported that the protein stability and 330 nucleocytoplasmic partitioning of two class A ARFs, ARF7 and ARF19, in Arabidopsis are 331 differentially regulated during root development67,68. By immunoblot analysis and confocal 332 .CC-BY-NC-ND 4.0 International licenseavailable under a (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 The copyright holder for this preprintthis version posted June 27, 2024. ; https://doi.org/10.1101/2024.06.26.599170doi: bioRxiv preprint 12 fluorescence microscopy, we showed that FLAG-or GFP-tagged ARF6 and ARF8 proteins 333 accumulated to similar levels in the WT and spy backgrounds (Supplementary Fig. 8a-8d). 334 Unlike the nucleocytoplasmic partitioning reported for ARF7 and ARF1967, ARF6 was only 335 detected in the nucleus in roots and pistils. 336 337 SPY enhanced IAA9-ARF6/8 transcription repression, but reduced ARF6/8 transactivation 338 activity 339 As described in the Introduction, class A ARFs play dual function in fruit development in 340 tomato and in Arabidopsis. In tomato, four class A-SlARFs together with SlIAA9 inhibit fruit set 341 but promote subsequent fruit growth after fertilization when elevated auxin levels trigger SlIAA9 342 degradation 26. ARF6 and ARF8 in Arabidopsis also showed similar dual function in fruit 343 development 27. We reasoned that AtIAA9 is likely the functional ortholog in this process because 344 its mRNA levels are elevated in ovaries (pistils) before anthesis69 (Arabidopsis eFP Browser43, 345 http://bar.utoronto.ca/efp_arabidopsis/cgi-bin/efpWeb.cgi) and IAA9 has been reported to 346 interact with ARF6 and ARF8 to repress auxin-induced adventitious root initiation70. Consistent 347 with this idea, the iaa9-1 mutant produced slightly longer pistils after emasculation compared to 348 WT (Supplementary 7c, 7d). The subtle parthenocarpic phenotype of iaa9-1 is likely because 349 the T-DNA insertion site in this allele is located in its fourth intron71, which is predicted to cause 350 C-terminal truncation at the end of the PB1 domain. To examine the effects of IAA9 and SPY on 351 the transcription activities of ARF6 and ARF8, dual luciferase (LUC) assays72 were performed 352 using the transient expression system in N. benthamiana. A synthetic auxin-responsive promoter 353 P3(2x) was fused to the firefly LUC (fLUC) as the reporter for this assay because P3(2x) was 354 shown to be responsive to AtARFs73. 35S:Renilla LUC (rLUC) was used as the internal control 355 to normalize variations in transformation efficiency. Five effectors, 35S:FLAG-ARF6, 356 35S:FLAG-ARF8, 35S:Myc-IAA9, 35S:HA-SPY and 35S:HA-(spy-19) were included in the 357 assays. A catalytic-domain mutant allele, spy-1932, served as a negative control. To avoid 358 variations in the IAA9 protein levels, the IAA9 construct used in this assay encodes a dominant 359 stabilized mutant protein with a P188S substitution in the conserved degron to prevent auxin-360 induced degradation74-76. As expected, expression of ARF6 or ARF8 alone induced P3(2x):fLUC 361 transcription. Importantly, co-expression of SPY, but not spy-19, attenuated transactivation 362 activities of both ARFs (Fig. 6a, 6b, and Supplementary Fig. 9). Co-expression of ARF6 or 363 .CC-BY-NC-ND 4.0 International licenseavailable under a (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 The copyright holder for this preprintthis version posted June 27, 2024. ; https://doi.org/10.1101/2024.06.26.599170doi: bioRxiv preprint 13 ARF8 with IAA9 repressed P3(2x):fLUC. Notably, co-expression of ARF, IAA9 and SPY 364 further repressed P3(2x):fLUC. These results suggested that O-fucosylation by SPY reduced 365 ARF6/8 transactivation activity, while enhanced the transcription repression activity of the 366 IAA9-ARF6/8 complexes. 367 368 SPY reduced ARF6 interaction with Mediator 8 369 370 Considering that SPY enhanced the transcription repression activity of the IAA9-ARF6/8 371 complexes, we tested whether SPY promotes ARF6-IAA9 interaction by co-IP assay. HA-IAA9 372 was expressed alone or co-expressed with FLAG-ARF6 and/or Myc-SPY or FLAG-GFP (as a 373 negative control) in N. benthamiana. After immunoprecipitated with anti-HA beads, we found 374 that IAA9 binding to ARF6 was not affected by co-expression of SPY (Fig. 6c). Notably, IAA9 375 also interacted with SPY (Fig. 6c), and its N-terminal region (amino acid residues 1-208) was O-376 fucosylated by SPY (Supplementary Fig. 10). Because ARFs function as homo- and hetero-377 dimers77, we examined whether SPY affects ARF6-ARF8 dimerization by co-IP assays. FLAG-378 ARF6 was expressed alone or co-expressed with Myc-ARF8 and/or Myc-SPY or Myc-GFP in N. 379 benthamiana. After immunoprecipitated with anti-FLAG beads, we found that SPY did not alter 380 ARF6-ARF8 binding affinity (Fig. 6d). Recent studies indicate that the Mediator complex is 381 required for class A ARF7/19-mediated transcription activation of auxin-induced genes for 382 lateral root initiation. ARF7 and ARF19 interact directly with MED8 of the head Mediator 383 module and MED25 of the tail Mediator module16. By yeast two-hybrid (Y2H) assay, we showed 384 that both ARF6 and ARF8 interacted with MED8 (Fig. 7a), but not with MED25 385 (Supplementary Fig. 11a). Notably, MED8 binds to the MR fragments of ARF6 and ARF8 386 (Fig. 7b), which contain O-Fuc sites modified by SPY (Fig. 3). Moreover, AAL pulldown assay 387 showed that MED8 was also O-fucosylated by SPY when transiently co-expressed in N. 388 benthamiana (Supplementary Fig. 11b), which is consistent with the detection of several O-389 Fuc-MED8 peptides from Arabidopsis extracts in a recent proteomic study35. To test whether 390 SPY modulates ARF-MED8 interaction, a co-IP assay was performed using N. benthamiana that 391 expressed FLAG-ARF6 alone or co-expressed with Myc-MED8 and/or HA-SPY, HA-spy-19 or 392 Myc-GFP. After IP’ed with anti-Myc beads, we found that co-expression of SPY, but not spy-19, 393 reduced MED8 binding to ARF6 (Fig. 7c). To confirm this observation in Arabidopsis, we 394 .CC-BY-NC-ND 4.0 International licenseavailable under a (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 The copyright holder for this preprintthis version posted June 27, 2024. ; https://doi.org/10.1101/2024.06.26.599170doi: bioRxiv preprint 14 generated 35S:MED8-GFP transgenic line and showed that MED8-GFP was functional in planta 395 to rescue the med8-2 mutant phenotype (Supplementary Fig. 11c, 11d). We also generated 396 transgenic Arabidopsis carrying both PUBQ10:FLAG-ARF6 and 35S:MED8-GFP in WT or spy 397 background. Co-IP assays were performed using protein extracts from transgenic lines carrying 398 either PUBQ10:FLAG-ARF6 alone or both PUBQ10:FLAG-ARF6 and 35S:MED8-GFP in WT or spy 399

(Fig. 7d). Importantly, the spy mutation enhanced MED8-ARF6 interaction without 400 altering MED8 protein levels, supporting that O-fucosylation of ARF6 and MED8 reduced their 401 binding affinity in planta. 402 403

404 405 In this study, we demonstrated that SPY inhibits auxin-induced fruit growth in 406 Arabidopsis by O-fucosylating ARF6/8, IAA9 and MED8. Co-IP assays showed that SPY-407 mediated O-fucosylation of these proteins reduced ARF-MED8 interaction, which leads to 408 enhanced transcription repression activity of the ARF6/8-IAA9 complex (Fig. 8), presumably 409 before pollination when auxin levels are low in the pistil. After pollination, elevated auxin levels 410 in the pistil trigger IAA9 degradation and release ARF6 and ARF8 to activate fruit growth-411 related genes by interacting with MED8 of the coactivator Mediator complex (Fig. 8). This 412 recruitment of the Mediator complex is known to promote the assembly of the RNA polymerase 413 II preinitiation complex (PIC)78,79. SPY-mediated O-fucosylation of ARF6/8 and MED8 also 414 attenuates transactivation activities of ARF6/8 by reducing ARF-MED8 interaction. We found 415 that SPY protein levels and its POFUT activity were reduced after pollination, although the 416 mechanism is unknown. This further enhances ARF6/8 transactivation activities by promoting 417 ARF-MED8 interaction. Importantly, we mapped the MED8 interaction domain in ARF6/8 to 418 their MRs, which also contain SPY target sites for O-fucosylation. This adds a novel regulatory 419 mechanism via PTM to modulate ARF activity, in addition to altered binding affinity to IAA 420 proteins by phosphorylation of DBD and/or MR of ARF5 and ARF729,80, reduced DNA binding 421 by SUMOylation in the DBD of ARF7/1930, and decreased protein stability by 422 ubiquitination31,68. Besides O-fucosylation, ARF6 and ARF8 are highly phosphorylated and O-423 GlcNAcylated. However, the pistils of the sec mutants lacked any phenotype comparing to WT, 424 suggesting that O-GlcNAcylation does not play a significant role in regulating ARF6/8 during 425 .CC-BY-NC-ND 4.0 International licenseavailable under a (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 The copyright holder for this preprintthis version posted June 27, 2024. ; https://doi.org/10.1101/2024.06.26.599170doi: bioRxiv preprint 15 fruit set. This could be due to a lower expression of SEC in the pistil than other tissues 426 (Arabidopsis eFP Browser43). 427 428 Although recent O-fucosylome studies have identified hundreds of SPY target 429 proteins34,35, the molecular mechanism of SPY regulation has only been elucidated for a handful 430 of its targets. These include enhanced protein-protein interaction for GA signaling repressors 431 DELLAs32, increased or decreased protein stability for bHLH transcription factors TEOSINTE 432 BRANCHED 1, CYCLOIDEA, PCFs (TCP14/15) in cytokinin response, and transcription 433 repressor PSEUDO RESPONSE REGULATOR 5 in circadian clock, respectively34,37,39, DNA 434 binding and/or transcription repression activity of bHLH transcription factor SPATULA for style 435 development 38, and protein translocation of co-chaperonin CPN20 in abscisic acid signaling36. 436 Here, we found that SPY not only O-fucosylates ARF6 and ARF8, but also two ARF6/8-437 interacting proteins, IAA9 and MED8, resulting in reduced interaction between ARF and MED8. 438 Thus, SPY-mediated O-fucosylation of multiple components in nuclear protein complexes may 439 contribute to another layer of transcription regulation. 440 441 The MRs of class A ARFs are enriched in glutamine residues and contain intrinsically 442 disordered regions (IDRs)18,19,67. The MR, together with the PB1 domain, of ARF7 and ARF19, 443 are required for cytoplasmic condensate formation, which blocks their function by preventing 444 nuclear localization67. Intriguingly, unlike ARF7 and ARF19, we did not observe cytoplasmic 445 condensate formation of ARF6 in the root cells of the maturation zone. The ARF6-GFP fusion 446 protein was only detected in the nucleus, suggesting that nucleocytoplasmic partitioning may not 447 be a universal regulatory mechanism for all class A ARFs. Previous studies also showed that the 448 MR of different class A ARFs specifies the distinct function of individual ARF in transcription 449 regulation by interactions with chromatin remodelers81 and/or transcription factors18. Here we 450 show that MR of ARF6/ARF8 interacts with MED8 of the coactivator Mediator complex, and 451 that O-fucosylation by SPY plays an important role in modulating this interaction. Further 452 studies will determine whether this is a general regulatory mechanism for all class A ARFs. 453 Besides MED8, three other MEDs (MED12, 13,14) in the Mediator complex are present in the 454 recently reported O-fucosylome in Arabidopsis35. Our study lays the foundation for elucidating 455 the molecular mechanism of SPY-mediated transcription regulation through its modulation of the 456 .CC-BY-NC-ND 4.0 International licenseavailable under a (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 The copyright holder for this preprintthis version posted June 27, 2024. ; https://doi.org/10.1101/2024.06.26.599170doi: bioRxiv preprint 16 coactivator MED complex. Finally, understanding the regulatory mechanism of class A-ARFs-457 controlled parthenocarpy has significant implication in developing climate-resilient crops as 458 production of seedless fruit without fertilization can ensure consistent fruit yield under stressful 459 environmental conditions. 460 461 462

463 Plant materials, growth conditions, plant transformation, and statistical analysis 464 Arabidopsis thaliana Columbia-0 (Col-0) was the wild-type control for most of the mutants used 465 in this study. The exceptions are spy-8 and spy-19, which are in the Landsberg erecta (Ler) 466 background. Plants were grown under long-day (22 °C, 16 h light/8 h dark) conditions. All 467 genotyping primers used in this study are listed in Supplementary Table 7. The genotyping 468 primers for spy-3 were designed using dCAPS Finder 2.0 (http://helix.wustl.edu/dcaps/)82. The 469 following mutants and transgenic lines have been described previously: (1) the spy mutants, spy-470 3, spy-8, spy-19, and spy-23 (WiscDsLox241C03)32,35,83, and the sec mutants, sec-284 and sec-471 585; (2) The arf mutants, arf6-2, nph4-1 (arf7), arf8-3, arf6-2 nph4-1, nph4-1 arf8-3, arf6-2 -/- 472 arf8-3 +/- 86-89; (3) The bZIP1 mutants, bzip1-1 (SALK_059343) and bzip1-2 (SALK_069489)90; 473 (4) PSPY:GFP-SPY spy-3, PSPY:GFP-SPY-NES spy-3, and PSPY:GFP-SPY-NLS spy-337; (5) 474 PARF6:ARF6-GFP and PARF7:ARF7-YFP lines in the Col-0 background67,91; (6) iaa9-171 and 475 med8-2 (CS16505)92 that were obtained from the Arabidopsis Biological Resource Center. For 476 the genetic interaction study, arf6-2, arf8-3, arf6-2 -/- arf8-3 +/- were crossed to spy-3, and the 477 F2 plants were genotyped by PCR to identify homozygous arf spy double mutants and arf 478 sesquimants (arf6 -/- arf8 +/- and arf8 -/- arf6 +/-) in the spy-3 background. Because the 479 homozygous arf6 arf8 double mutant is sterile87, the arf sesquimutants (arf6 -/- arf8 +/- and arf8 480 -/- arf6 +/-) in SPY or spy-3 background were identified in the segregating F3 population by 481 genotyping for phenotype analysis. 482 PARF6:FLAG-ARF6 and PARF8:FLAG-ARF8 constructs were introduced into arf6-2 or 483 arf8-3 separately by Agrobacterium-mediated transformation. Homozygous transgenic lines 484 containing a single insertion site were obtained as described previously93. A representative 485 PARF6:FLAG-ARF6 line (#1-11-7) and a PARF8:FLAG-ARF8 line (#2-1-4) were used for the 486 complementation test. PUBQ10:FLAG-ARF6 and PUBQ10:FLAG-ARF8 constructs were introduced 487 .CC-BY-NC-ND 4.0 International licenseavailable under a (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 The copyright holder for this preprintthis version posted June 27, 2024. ; https://doi.org/10.1101/2024.06.26.599170doi: bioRxiv preprint 17 into WT (Ler) separately by Agrobacterium-mediated transformation. The transgenes, 488 PUBQ10:FLAG-ARF6 (in line #4-2-1) and PUBQ10:FLAG-ARF8 (in line #4-1-1), were then 489 introduced into the spy-8 background by genetic crosses. The resulting homozygous 490 PUBQ10:FLAG-ARF6 spy-8 line and PUBQ10:FLAG-ARF8 spy-8 line were obtained by genotyping. 491 To test the activity of 35S:MED8-GFP in planta, the pGWB405:MED8-GFP construct was 492 introduced into med8-2 by Agrobacterium-mediated transformation. Homozygous transgenic 493 lines containing a single insertion site were obtained, and a representative 35S:MED8-GFP 494 med8-2 line (#1-32-1) was used for phenotype analysis. To generate the double transgenic lines 495 for co-IP assays, PUBQ10:FLAG-ARF6 spy-8 was transformed with the 35S:MED8-GFP construct 496 (pGWB405:MED8-GFP). A representative double transgenic line FLAG-ARF6 MED8-GFP spy-497 8 (#1-3-1) was crossed with PUBQ10:FLAG-ARF6 in the WT (Ler) background to generate FLAG-498 ARF6 MED8-GFP in WT (#1-4-52). The PARF6:ARF6-GFP line in the Col-0 background was 499 crossed with the spy-3 mutant to generate PARF6:ARF6-GFP spy-3 line. The PSPY:FLAG-SPY 500 construct was introduced into spy-3 by Agrobacterium-mediated transformation to obtain the 501 PSPY:FLAG-SPY spy-3 lines, and a representative homozygous line (#1-2-17) was used for 502 further analysis. 503 For agroinfiltration, 3 or 4-week-old plants of Nicotiana benthamiana were used. 504 Statistical analyses were performed by Tukey's honestly significant difference (HSD) mean 505 separation tests with SPSS Statistics 17.0 software. 506 507 Plasmid construction 508 The following plasmids were described previously: 35S:Myc-SPY and 35S:Myc-SEC for 509 transient expression46; P35S:FLAG-GFP-NLS for the negative control of transient expression34; 510 pEarleyGate201 and 203 vectors94, pDONR207:SPY95, pEG3F-GW34, pDEST32-HA96 and 511 pDEST22-FLAG96 for cloning new constructs. Primers and plasmid constructs are listed in 512 Supplementary Tables 7 and 8, respectively. All DNA constructs generated from PCR 513 amplification were sequenced to ensure that no mutations were introduced. 514 515 Flower emasculation and hormone treatments 516 Flower emasculation and hormone treatments were conducted following the methods described 517 previously42. Flowers were emasculated at stage 1097, approximately 2 days before anthesis (–2 518 .CC-BY-NC-ND 4.0 International licenseavailable under a (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 The copyright holder for this preprintthis version posted June 27, 2024. ; https://doi.org/10.1101/2024.06.26.599170doi: bioRxiv preprint 18 DAA), to remove sepals, petals, and anthers. The pistils of emasculated flowers were immersed 519 uniformly in 100 µM GA3 or 50 µM picloram (Sigma-Aldrich, CAS-1918-02-1) solution that 520 also contained 0.01% (v/v) Triton X-100, or in 0.01% (v/v) Triton X-100 alone for mock 521 treatment. Seven days after emasculation, the final length of the pistils was measured using 522 ImageJ. 523 524 Transient expression and dual luciferase assay in Nicotiana benthamiana 525 For dual luciferase assays and pulldown assays, transient expression of different epitope-526 tagged proteins in N. benthamiana was performed as described with slight modifications98. The 527 N. benthamiana leaves were harvest after 48 hr of agroinfiltration99 and luciferase activity was 528 quantified using the dual-luciferase reporter assay system (Promega). To determine the relative 529 promoter activity, the ratio of fLUC to rLUC activity was calculated for each sample. Three 530 biological repeats were conducted for each effector combination. 531 532 AAL-agarose pull-down, co-IP assays and protein blot analyses 533 For AAL pulldown assays, FLAG-GFP/ARF6/ARF6-N/ARF6-MR/ARF6-C or IAA9/IAA9-534 N/IAA9-C were individually expressed or co-expressed with Myc-SPY in N. benthamiana by 535 agroinfiltration, following the established protocol95. Protein extracts were incubated with AAL-536 agarose beads (Vector Labs, AL-1393-2, 2 mg lectin/mL) to enrich for O-fucosylated proteins, 537 and analyzed by immunoblot analysis as described34. 538 To investigate the effect of SPY on ARF6-IAA9 interaction, HA-IAA9P188S was 539 expressed alone or co-expressed with FLAG-ARF6 and/or Myc-SPY or FLAG-GFP in N. 540 benthamiana. Co-immunoprecipitation (co-IP) assays were performed using anti-HA beads 541 (Sigma-Aldrich, A2095), following the described procedure15. To examine the effect of SPY on 542 ARF6-MED8 interaction, FLAG-ARF6 was expressed alone or co-expressed with Myc-MED8 543 and/or HA-SPY, or Myc-GFP in N. benthamiana. Co-immunoprecipitation (co-IP) assays were 544 performed using anti-Myc beads (Sigma-Aldrich, A7470), following the described procedure15. 545 To verify the effect of spy on ARF6-MED8 interaction in planta, co-IP assays were performed 546 using ChromoTek GFP-Trap® Magnetic Agarose (Proteintech, GTMA-400) and protein extracts 547 from transgenic Arabidopsis lines carrying either FLAG-ARF6 or both FLAG-ARF6 and MED8-548 GFP in WT (Ler) or spy-8 background. 549 .CC-BY-NC-ND 4.0 International licenseavailable under a (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 The copyright holder for this preprintthis version posted June 27, 2024. ; https://doi.org/10.1101/2024.06.26.599170doi: bioRxiv preprint 19 Immunoblot analyses were performed using horseradish peroxidase (HRP)-conjugated 550 anti-FLAG M2 mouse monoclonal (Sigma-Aldrich A8592, 1:10,000 dilution) and mouse HRP-551 anti-MYC monoclonal antibodies (BioLegend #626803, 1:5,000 dilution), mouse HRP-anti-HA 552 (6E2) monoclonal antibody (Cell Signaling Technology #2999S, 1:5,000 dilution), mouse anti-553 GFP (Roche #11814460001, 1:2,000 dilution), mouse anti-Tubulin (Sigma T5168, dilution 554 1:500,000). HRP-conjugated donkey anti-mouse IgG (Jackson ImmunoResearch #715-035-150) 555 was used to detect anti-GFP at 1:5,000 dilution. Biotinylated-Aleuria aurantia lectin (AAL-556 biotin, Vector Labs B-1395, 1:30,000) followed by Streptavidin-HRP (Jackson ImmunoResearch 557 #016-030-084, 1: 100,000x dilution,) were used for AAL blot. To detect O-GlcNAcylated 558 proteins, the anti-O-GlcNAc monoclonal antibody CTD110.6 (5,000× dilution, Cell Signaling 559 Technology, Cat. No. 9875) followed by a goat anti-mouse IgM-HRP secondary antibody 560 (30,000× dilution, Thermo-Fisher Scientific, Cat. No. 31440) were used. 561 562 Y2H assays 563 The ProQuest Two-Hybrid system (Invitrogen) and yeast strain pJ69-4A were used for the Y2H 564 assays. Yeast transformation and 3-amino-1,2,4-triazole (3-AT) tests were performed according 565 to the manufacturer's protocol as described previously96 with slight modifications: 3-AT 566 concentrations in the plates were 0, 2, 5, 10, 25, 50, 75, and 100 mM in most cases, except that 0, 567 2, 10, 25 mM were used for the ARF6 or ARF8 domain mapping. 568 569 Confocal microscopy 570 GFP signals in pistils were analyzed using PSPY:GFP-SPY spy-3, PSPY:GFP-SPY-NES spy-3, and 571 PSPY:GFP-SPY-NLS spy-337. The pistils at stage 10 were imaged using a Zeiss 880 equipped with 572 a 20x objective. The pistils from stage 10 to stage 14 were imaged using a Zeiss Axio 573 Zoom.V16. Identical image settings were used for direct comparison. 574 For detecting the protein localization of ARF6-GFP and ARF7-YFP in the WT or spy-3 575 background, the primary root cells in 3d-old seedlings were stained with 10 μM propidium 576 iodide (Sigma, P4170) and imaged using a Zeiss 880 equipped with a 20x objective. The pistils 577 at stage 10 were also analyzed. Excitation and detection were set as follows: GFP or YFP, 578 excitation at 488 nm and detection at 493–558 nm; PI staining, excitation at 561 nm and 579 .CC-BY-NC-ND 4.0 International licenseavailable under a (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 The copyright holder for this preprintthis version posted June 27, 2024. ; https://doi.org/10.1101/2024.06.26.599170doi: bioRxiv preprint 20 detection at 605–695 nm. Confocal images were processed using the Fiji package of ImageJ. 580 Identical image settings were used for direct comparison. 581 582 RNA-seq analysis 583 Total RNAs were purified from unpollinated pistils from stage 10 flowers (–2 DAA) of arf6-2, 584 arf8-3, spy-3 and WT (Col-0), and pollinated pistils from WT stage 14 flowers (0 DAA). RNA-585 Seq cDNA libraries (two biological repeats) were prepared with the QuantSeq 3’mRNA-Seq 586 library prep kit FWD for Illumina (Lexogen). DNA sequencing was performed with Illumina 587 Next-Seq500 High-Output 75bp SR. Sequence alignment and DE (differential expression) 588 analysis were conducted on a commercial server with pre-established computational pipelines (8 589 omics Gene Technology Co. Ltd., Beijing, China)100. Co-regulated genes among spy-3, arf6-2, 590 arf8-3 and fertilization-responsive gene lists were then identified (fold change > 1.5; p < 0.05). 591 Venn diagrams were made using online tool at InteractiVenn.net101. GO analysis was performed 592 using Panther v.18.0102. Heatmap analysis was made by using online tool at MetaboAnalyst 593 5.0103. 594 595 RT-qPCR and ChIP-qPCR analyses 596 Total RNAs from pistils stage 10 or stage14 were isolated with RNeasy Plant Mini Kit (Qiagen). 597 First-strand cDNA was then synthesized using a Transcriptor First Strand cDNA Synthesis kit 598 (Roche). qPCR analyses were performed using FastStart Essential DNA Green Master mix 599 (Roche) and LightCycler 96 instrument (Roche). The PCR program was performed as described 600 before96. Relative transcript levels were determined by normalizing with PP2A (At1g13320). 601 Mean values of fold change were calculated from three biological replicates. For ChIP-qPCR 602 analysis, seedlings of the PUBQ10:FLAG-ARF6 and PUBQ10:FLAG-ARF8 transgenic lines in the 603 WT or spy-8 background were grown in liquid 0.5x MS and 1% sucrose for 10 days in 604 continuous light. The ChIP procedure was performed as described104. Primers for all qPCR 605 analyses are listed in Supplementary Table 7. 606 607 Tandem affinity purification of FLAG-ARF6/8 from N. benthamiana or Arabidopsis for 608 immunoblot or MS analysis 609 .CC-BY-NC-ND 4.0 International licenseavailable under a (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 The copyright holder for this preprintthis version posted June 27, 2024. ; https://doi.org/10.1101/2024.06.26.599170doi: bioRxiv preprint 21 The FLAG-ARF6/8 proteins (containing a 6xHis-3xFLAG-tag) that were transiently expressed 610 in N. benthamiana were tandem affinity-purified using a His-Bind resin followed by anti-Flag-611 M2-agarose beads (Sigma-Aldrich) as described32 with slight modifications. 3 g of starting tissue 612 was used for MS analysis and a smaller scale with 0.7 g of starting tissue was used for AAL blot 613 analysis. The extraction and purification buffer included 50 mM fucose and 2x protease 614 inhibitors. His-FLAG-ARF6/8 proteins were also purified from PUBQ10:FLAG-ARF6/8 transgenic 615 Arabidopsis lines in WT or spy-8 backgrounds. The tandem affinity purification procedures were 616 as previously described32 for both AAL blot and MS analyses, with the following modifications: 617 10 g of starting tissues was used, and 50 mM fucose and 2x protease inhibitors were included 618 during purification. The cleared extract was incubated with 0.4 mL of His-Bind resin for 1.5 h at 619 4 °C and was loaded onto a disposable plastic column. The second purification step was carried 620 out with 10 μL of anti-Flag-M2-agarose beads (Sigma-Aldrich). The purified protein was 621 digested by trypsin on-beads for MS analysis as described32. 622 623 Identification of O-fucosylation sites by liquid chromatography (LC)–electrospray 624 ionization (ESI)–mass spectrometry (MS) 625 Trypsin-digested 6His-3xFLAG-ARF6 and 6His-3xFLAG-ARF8 proteins, purified from N. 626 benthamiana and from A. thaliana, were separated by reverse phase nano-HPLC, and analyzed 627 by electrospray ionization (ESI)-MS using a Thermo ScientificTM Orbitrap FusionTM TribridTM 628 mass spectrometer equipped with electron transfer dissociation (ETD)105. Nano-HPLC was 629 performed as described previously32. MS1 spectra were acquired in the Orbitrap at a resolution 630 of 60,000 followed by data-dependent, 3 second Top-N, MS/MS experiments. Precursors were 631 isolated by resolving quadrupole with a 3 m/z window. An MS/MS decision tree was made for 632 each sample that included collision-activated dissociation CAD, ETD, and higher-energy 633 collisional dissociation (HCD). Precursors with a charge state less than 5 were selected for low 634 resolution MS/MS, fragmented with CAD and ETD, and scanned out of the linear ion trap at a 635 normal scan rate. Calibrated reaction times were used for ETD events. Precursors with a charge 636 state above 4 were selected for high resolution MS/MS, fragmented with HCD and ETD, and 637 scanned out of the Orbitrap at a resolution of 30,000. Electron transfer-higher energy collisional 638 dissociation (EThcD) was later added to the decision tree for precursors that had a charge state 639 above 4 and m/z above 900, and scanned out of the Orbitrap at a resolution of 30,000. For some 640 .CC-BY-NC-ND 4.0 International licenseavailable under a (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 The copyright holder for this preprintthis version posted June 27, 2024. ; https://doi.org/10.1101/2024.06.26.599170doi: bioRxiv preprint 22 samples, peptides were targeted for fragmentation based on intact mass and retention time 641 observed in previous experiments. 642 To detect additional ARF6 peptides within the MR sequence, trypsin-digested ARF6 643 samples from N. benthamiana underwent a second digestion with chymotrypsin overnight at 644 room temperature. ARF6 and ARF8 samples from N. benthamiana were subsequently cleaned up 645 using hydrophilic interaction liquid chromatography (HILIC) prior to a second MS analysis, a 646

developed by Keira Mahoney (unpublished data) and previously described34. Cleaned up 647 samples were analyzed immediately by MS or stored at -35°C. 648 Data files were searched using Byonic Version 3.8.13 (Protein Metrics)106. ARF6 and 649 ARF8 data files were searched against a database containing the sequence of 6His-3xFLAG-650 ARF6 and 6His-3xFLAG-ARF8, respectively. Search parameters included specific cleavage C-651 terminal to R and K residues with up to five allowed missed cleavages, 10 ppm tolerance for 652 precursor mass, 15 ppm tolerance for high-resolution MS/MS, and 0.35 Da mass tolerance for 653 low-resolution MS/MS. After the chymotrypsin digest, cleavages C-terminal to F, L, Y, and W 654 residues were added to the search parameters. Variable modifications selected included oxidation 655 of M residues, phosphorylation of S, T, and Y residues, alkylation of C residues, and O-656 GlcNAcylation, O-fucosylation, and O-hexosylation of S and T residues. No protein false 657 discovery rate cutoff or score cutoff was applied prior to the output of search results. Peptides 658 were manually validated, and modification sites were determined manually using ETD spectra. 659 660 Accession Numbers 661 Arabidopsis Genome Initiative locus identifiers for the genes mentioned in this article are as 662 follows: SPY (AT3G11540), ARF6 (AT1G30330), ARF7 (NPH4, AT5G20730), ARF8 663 (AT5G37020), IAA9 (AT5G65670), SEC (AT3G04240), MED8 (AT2G03070), MED25 664 (AT1G25540), PP2A (AT1G13320), ERF023 (AT1G01250), ERF107 (AT5G61590), bZIP1 665 (AT5G49450), AGP12 (AT3G13520), SHL (AT4G39100), GNTL (AT3G52060), IAA19 666 (AT3G15540). 667 668 Data Availability 669 Raw and processed RNA-Seq data have been deposited at in the NCBI Sequence Read Archive 670 under BioProject PRJNA1095421. The mass spectrometry proteomics data have been deposited 671 .CC-BY-NC-ND 4.0 International licenseavailable under a (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 The copyright holder for this preprintthis version posted June 27, 2024. ; https://doi.org/10.1101/2024.06.26.599170doi: bioRxiv preprint 23 to the ProteomeXchange Consortium via the PRIDE107 partner repository with the dataset 672 identifier PXD051232 (Project DOI: 10.6019/PXD051232). Source Data files will be provided 673 with this paper before publication. 674 675

676 We thank Jason Reed and Lucia Strader for helpful discussions, and for sharing Arabidopsis arf 677 mutants and reporter lines. We also thank Jyan-Chyun Jang for providing the bzip1 mutants. We 678 are grateful to Mingyuan Zhu for helping with confocal microscopy analysis. This work was 679 supported by the National Institutes of Health (GM100051 and GM150029 to TPS, GM037537 680 to DFH), and the United State Department of Agriculture (2023-67013-39532 to TPS). A special 681 thank you to Protein Metrics for providing Byonic. 682 683 Author Contributions 684 T.P.S. and Y.W. conceived and designed the research project. Y.W., R.Z. and J.H. performed 685 experiments, and T.P.S., Y.W., R.Z. and J.H. analyzed the data and generated figures. L.W. and 686 H.W. helped with RNA-seq data analysis. S.K. performed LC-ESI-MS analysis, and S.K., J.S., 687 and D.F.H. analyzed MS data. T.P.S. wrote the manuscript with input from all co-authors. 688 689 Competing Financial Interests Statements 690 The authors declare no competing financial interests. 691 692 693 .CC-BY-NC-ND 4.0 International licenseavailable under a (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 The copyright holder for this preprintthis version posted June 27, 2024. ; https://doi.org/10.1101/2024.06.26.599170doi: bioRxiv preprint 24

694 695 1 Fenn, M. A. & Giovannoni, J. J. Phytohormones in fruit development and maturation. 696 Plant J 105, 446-458 (2021). 697 2 Srivastava, A. & Handa, A. K. Hormonal regulation of tomato fruit development: a 698 molecular perspective. J Plant Growth Regul 24, 67-82 (2005). 699 3 Seymour, G. B., Ostergaard, L., Chapman, N. H., Knapp, S. & Martin, C. Fruit 700 development and ripening. Annu Rev Plant Biol 64, 219-241 (2013). 701 4 Ruan, Y. L., Patrick, J. W., Bouzayen, M., Osorio, S. & Fernie, A. R. Molecular 702 regulation of seed and fruit set. Trends Plant Sci 17, 656-665 (2012). 703 5 Gorguet, B., van Heusden, A. W. & Lindhout, P. Parthenocarpic fruit development in 704 tomato. Plant Biol 7, 131-139 (2005). 705 6 Dorcey, E., Urbez, C., Blazquez, M. A., Carbonell, J. & Perez-Amador, M. A. 706 Fertilization-dependent auxin response in ovules triggers fruit development through the 707 modulation of gibberellin metabolism in Arabidopsis. Plant J 58, 318-332 (2009). 708 7 Sharif, R., Su, L., Chen, X. & Qi, X. Hormonal interactions underlying parthenocarpic 709 fruit formation in horticultural crops. Hortic Res 9 (2022). 710 8 Varoquaux, F., Blanvillain, R., Delseny, M. & Gallois, P. Less is better: new approaches 711 for seedless fruit production. Trends Biotechnol 18, 233-242 (2000). 712 9 Resentini, F., Orozco-Arroyo, G., Cucinotta, M. & Mendes, M. A. The impact of heat 713 stress in plant reproduction. Frontiers in plant science 14, 1271644 (2023). 714 10 Guilfoyle, T. J. The PB1 domain in auxin response factor and Aux/IAA proteins: a 715 versatile protein interaction module in the auxin response. Plant Cell 27, 33-43 (2015). 716 11 Lavy, M. & Estelle, M. Mechanisms of auxin signaling. Development 143, 3226-3229 717 (2016). 718 12 Morffy, N. & Strader, L. C. Structural Aspects of Auxin Signaling. Cold Spring Harb 719 Perspect Biol 14 (2022). 720 13 Weijers, D. & Wagner, D. Transcriptional responses to the auxin hormone. Annu Rev 721 Plant Biol 67, 539-574 (2016). 722 14 Leyser, O. Auxin Signaling. Plant Physiol 176, 465-479 (2018). 723 15 Leydon, A. R. et al. Repression by the Arabidopsis TOPLESS corepressor requires 724 association with the core mediator complex. eLife 10 (2021). 725 16 Ito, J. et al. Auxin-dependent compositional change in Mediator in ARF7- and ARF19-726 mediated transcription. Proc Natl Acad Sci U S A 113, 6562-6567 (2016). 727 17 Mutte, S. K. et al. Origin and evolution of the nuclear auxin response system. eLife 7, 728 10.7554/eLife.33399 (2018). 729 18 Cancé, C., Martin-Arevalillo, R., Boubekeur, K. & Dumas, R. Auxin response factors are 730 keys to the many auxin doors. The New phytologist 235, 402-419 (2022). 731 19 Guilfoyle, T. J. & Hagen, G. Auxin response factors. Curr Opin Plant Biol 10, 453-460 732 (2007). 733 20 Lavy, M. et al. Constitutive auxin response in Physcomitrella reveals complex 734 interactions between Aux/IAA and ARF proteins. eLife 5, e13325 (2016). 735 21 Kato, H. et al. Design principles of a minimal auxin response system. Nat Plants 6, 473-736 482 (2020). 737 .CC-BY-NC-ND 4.0 International licenseavailable under a (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 The copyright holder for this preprintthis version posted June 27, 2024. ; https://doi.org/10.1101/2024.06.26.599170doi: bioRxiv preprint 25 22 Goetz, M., Vivian-Smith, A., Johnson, S. D. & Koltunow, A. M. AUXIN RESPONSE 738 FACTOR8 is a negative regulator of fruit initiation in Arabidopsis. Plant Cell 18, 1873-739 1886 (2006). 740 23 Liu, S. et al. Tomato AUXIN RESPONSE FACTOR 5 regulates fruit set and 741 development via the mediation of auxin and gibberellin signaling. Sci Rep 8, 2971 742 (2018). 743 24 Du, L. et al. SmARF8, a transcription factor involved in parthenocarpy in eggplant. Mol 744 Genet Genomics 291, 93-105 (2016). 745 25 de Jong, M., Wolters-Arts, M., Feron, R., Mariani, C. & Vriezen, W. H. The Solanum 746 lycopersicum auxin response factor 7 (SlARF7) regulates auxin signaling during tomato 747 fruit set and development. Plant J 57, 160-170 (2009). 748 26 Hu, J., Li, X. & Sun, T. P. Four class A AUXIN RESPONSE FACTORs promote tomato 749 fruit growth despite suppressing fruit set. Nat Plants 9, 706–719 (2023). 750 27 Israeli, A. et al. Modulating auxin response stabilizes tomato fruit set. Plant Physiol 192, 751 2336-2355 (2023). 752 28 Han, S. & Hwang, I. Integration of multiple signaling pathways shapes the auxin 753 response. J Exp Bot 69, 189-200 (2018). 754 29 Cho, H. et al. A secreted peptide acts on BIN2-mediated phosphorylation of ARFs to 755 potentiate auxin response during lateral root development. Nat Cell Biol 16, 66-76 756 (2014). 757 30 Orosa-Puente, B. et al. Root branching toward water involves posttranslational 758 modification of transcription factor ARF7. Science 362, 1407-1410 (2018). 759 31 Li, K., Wang, S., Wu, H. & Wang, H. Protein levels of several arabidopsis auxin 760 response factors are regulated by multiple factors and aba promotes ARF6 protein 761 ubiquitination. Int J Mol Sci 21 (2020). 762 32 Zentella, R. et al. The Arabidopsis O-fucosyltransferase SPINDLY activates nuclear 763 growth repressor DELLA. Nat Chem Biol 13, 479-485 (2017). 764 33 Sun, T. P. Novel nucleocytoplasmic protein O-fucosylation by SPINDLY regulates 765 diverse developmental processes in plants. Curr Opin Struct Biol 68, 113-121 (2021). 766 34 Zentella, R. et al. SPINDLY O-fucosylates nuclear and cytoplasmic proteins involved in 767 diverse cellular processes in plants. Plant Physiol 191, 1546-1560 (2023). 768 35 Bi, Y. et al. SPINDLY mediates O-fucosylation of hundreds of proteins and sugar-769 dependent growth in Arabidopsis. Plant Cell 35, 1318-1333 (2023). 770 36 Liang, L. et al. O-fucosylation of CPN20 by SPINDLY Derepresses Abscisic Acid 771 Signaling During Seed Germination and Seedling Development. Frontiers in plant 772 science 12 (2021). 773 37 Wang, Y. et al. Nuclear localized O-fucosyltransferase SPY facilitates PRR5 proteolysis 774 to fine-tune the pace of Arabidopsis circadian clock. Mol Plant 13, 446-458 (2020). 775 38 Jiang, Y. et al. O-glycosylation of the transcription factor SPATULA promotes style 776 development in Arabidopsis. Nat Plants 10, 283-299 (2024). 777 39 Steiner, E. et al. The Putative O-Linked N-Acetylglucosamine Transferase SPINDLY 778 Inhibits Class I TCP Proteolysis to Promote Sensitivity to Cytokinin. Plant Physiol 171, 779 1485-1494 (2016). 780 40 Mutanwad, K. V. & Lucyshyn, D. Balancing O-GlcNAc and O-fucose in plants. FEBS J 781 289, 3086-3092 (2022). 782 .CC-BY-NC-ND 4.0 International licenseavailable under a (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 The copyright holder for this preprintthis version posted June 27, 2024. ; https://doi.org/10.1101/2024.06.26.599170doi: bioRxiv preprint 26 41 Jacobsen, S. E. & Olszewski, N. E. Mutations at the SPINDLY locus of Arabidopsis alter 783 gibberellin signal transduction. Plant Cell 5, 887-896 (1993). 784 42 Vivian-Smith, A. & Koltunow, A. M. Genetic analysis of growth-regulator-induced 785 parthenocarpy in Arabidopsis. Plant Physiol 121, 437-451 (1999). 786 43 Winter, D. et al. An "Electronic Fluorescent Pictograph" browser for exploring and 787 analyzing large-scale biological data sets. PLoS One 2, e718 (2007). 788 44 Fuentes, S. et al. Fruit growth in Arabidopsis occurs via DELLA-dependent and DELLA-789 independent gibberellin responses. Plant Cell 24, 3982-3996 (2012). 790 45 Savaldi-Goldstein, S. et al. New auxin analogs with growth-promoting effects in intact 791 plants reveal a chemical strategy to improve hormone delivery. Proc Natl Acad Sci U S A 792 105, 15190-15195 (2008). 793 46 Zentella, R. et al. O-GlcNAcylation of master growth repressor DELLA by SECRET 794 AGENT modulates multiple signaling pathways in Arabidopsis. Genes Dev 30, 164-176 795 (2016). 796 47 Xu, S. L. et al. Proteomic analysis reveals O-GlcNAc modification on proteins with key 797 regulatory functions in Arabidopsis. Proc Natl Acad Sci U S A 114, E1536-E1543 (2017). 798 48 Shrestha, R., Karunadasa, S., Grismer, T., Reyes, A. V. & Xu, S. L. SECRET AGENT O-799 GlcNAcylates hundreds of proteins involved in diverse cellular processes in Arabidopsis. 800 Mol Cell Proteomics 23, 100732 (2024). 801 49 Bi, Y. et al. Arabidopsis ACINUS is O-glycosylated and regulates transcription and 802 alternative splicing of regulators of reproductive transitions. Nat Commun 12, 945 (2021). 803 50 Liscum, E. & Reed, J. W. Genetics of Aux/IAA and ARF action in plant growth and 804 development. Plant Mol Biol 49, 387-400 (2002). 805 51 Oh, E. et al. Cell elongation is regulated through a central circuit of interacting 806 transcription factors in the Arabidopsis hypocotyl. eLife 3, e03031 (2014). 807 52 Walper, E., Weiste, C., Mueller, M. J., Hamberg, M. & Droge-Laser, W. Screen 808 Identifying Arabidopsis Transcription Factors Involved in the Response to 9-809 Lipoxygenase-Derived Oxylipins. PLoS One 11, e0153216 (2016). 810 53 Gaudinier, A. et al. Transcriptional regulation of nitrogen-associated metabolism and 811 growth. Nature 563, 259-264 (2018). 812 54 Peng, M., Bi, Y. M., Zhu, T. & Rothstein, S. J. Genome-wide analysis of Arabidopsis 813 responsive transcriptome to nitrogen limitation and its regulation by the ubiquitin ligase 814 gene NLA. Plant Mol Biol 65, 775-797 (2007). 815 55 Kang, S. G., Price, J., Lin, P. C., Hong, J. C. & Jang, J. C. The arabidopsis bZIP1 816 transcription factor is involved in sugar signaling, protein networking, and DNA binding. 817 Mol Plant 3, 361-373 (2010). 818 56 Dröge-Laser, W. & Weiste, C. The C/S(1) bZIP network: A regulatory hub orchestrating 819 plant energy homeostasis. Trends Plant Sci 23, 422-433 (2018). 820 57 De Coninck, T., Gistelinck, K., Janse van Rensburg, H. C., Van den Ende, W. & Van 821 Damme, E. J. M. Sweet modifications modulate plant development. Biomolecules 11 822 (2021). 823 58 Zalepa-King, L. & Citovsky, V. A plasmodesmal glycosyltransferase-like protein. PLoS 824 One 8, e58025 (2013). 825 59 Usadel, B. et al. Global transcript levels respond to small changes of the carbon status 826 during progressive exhaustion of carbohydrates in Arabidopsis rosettes. Plant Physiol 827 146, 1834-1861 (2008). 828 .CC-BY-NC-ND 4.0 International licenseavailable under a (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 The copyright holder for this preprintthis version posted June 27, 2024. ; https://doi.org/10.1101/2024.06.26.599170doi: bioRxiv preprint 27 60 Pereira, A. M. et al. Differential expression patterns of arabinogalactan proteins in 829 Arabidopsis thaliana reproductive tissues. J Exp Bot 65, 5459-5471 (2014). 830 61 Lopez-Gonzalez, L. et al. Chromatin-dependent repression of the Arabidopsis floral 831 integrator genes involves plant specific PHD-containing proteins. Plant Cell 26, 3922-832 3938 (2014). 833 62 Müssig, C., Kauschmann, A., Clouse, S. D. & Altmann, T. The Arabidopsis PHD-finger 834 protein SHL is required for proper development and fertility. Mol Gen Genet 264, 363-835 370 (2000). 836 63 Qian, S. et al. Dual recognition of H3K4me3 and H3K27me3 by a plant histone reader 837 SHL. Nat Commun 9, 2425 (2018). 838 64 Kanayama, Y. Sugar metabolism and fruit development in the tomato. Horticulture J 86, 839 417-425 (2017). 840 65 Tang, N., Deng, W., Hu, G., Hu, N. & Li, Z. Transcriptome profiling reveals the 841 regulatory mechanism underlying pollination dependent and parthenocarpic fruit set 842 mainly mediated by auxin and gibberellin. PLoS One 10, e0125355 (2015). 843 66 Wang, H. et al. Regulatory features underlying pollination-dependent and -independent 844 tomato fruit set revealed by transcript and primary metabolite profiling. Plant Cell 21, 845 1428-1452 (2009). 846 67 Powers, S. K. et al. Nucleo-cytoplasmic Partitioning of ARF Proteins Controls Auxin 847 Responses in Arabidopsis thaliana. Mol Cell 76, 177-190 (2019). 848 68 Jing, H. et al. Regulation of AUXIN RESPONSE FACTOR condensation and nucleo-849 cytoplasmic partitioning. Nat Commun 13, 4015 (2022). 850 69 Swanson, R., Clark, T. & Preuss, D. Expression profiling of Arabidopsis stigma tissue 851 identifies stigma-specific genes. Sex Plant Reprod 18, 163-171 (2005). 852 70 Lakehal, A. et al. A Molecular Framework for the Control of Adventitious Rooting by 853 TIR1/AFB2-Aux/IAA-Dependent Auxin Signaling in Arabidopsis. Mol Plant 12, 1499-854 1514 (2019). 855 71 Overvoorde, P. J. et al. Functional genomic analysis of the AUXIN/INDOLE-3-ACETIC 856 ACID gene family members in Arabidopsis thaliana. Plant Cell 17, 3282-3300 (2005). 857 72 Matsuo, N., Minami, M., Maeda, T. & Hiratsuka, K. Dual luciferase assay for monitoring 858 transient gene expression in higher plants. Plant Biotechnol 18, 71-75 (2001). 859 73 Pierre-Jerome, E., Moss, B. L., Lanctot, A., Hageman, A. & Nemhauser, J. L. Functional 860 analysis of molecular interactions in synthetic auxin response circuits. Proc Natl Acad Sci 861 U S A 113, 11354-11359 (2016). 862 74 Shimizu-Mitao, Y. & Kakimoto, T. Auxin sensitivities of all Arabidopsis Aux/IAAs for 863 degradation in the presence of every TIR1/AFB. Plant Cell Physiol 55, 1450-1459 864 (2014). 865 75 Tian, Q. & Reed, J. W. Control of auxin-regulated root development by the Arabidopsis 866 thaliana SHY2/IAA3 gene. Development 126, 711-721 (1999). 867 76 Uehara, T., Okushima, Y., Mimura, T., Tasaka, M. & Fukaki, H. Domain II mutations in 868 CRANE/IAA18 suppress lateral root formation and affect shoot development in 869 Arabidopsis thaliana. Plant Cell Physiol 49, 1025-1038 (2008). 870 77 Rienstra, J., Hernandez-Garcia, J. & Weijers, D. To bind or not to bind: how AUXIN 871 RESPONSE FACTORs select their target genes. J Exp Bot 74, 6922-6932 (2023). 872 .CC-BY-NC-ND 4.0 International licenseavailable under a (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 The copyright holder for this preprintthis version posted June 27, 2024. ; https://doi.org/10.1101/2024.06.26.599170doi: bioRxiv preprint 28 78 Richter, W. F., Nayak, S., Iwasa, J. & Taatjes, D. J. The Mediator complex as a master 873 regulator of transcription by RNA polymerase II. Nat Rev Mol Cell Biol 23, 732-749 874 (2022). 875 79 Soutourina, J. Transcription regulation by the Mediator complex. Nature Rev Mol Cell 876 Biol 19, 262-274 (2018). 877 80 Han, S. et al. BIL1-mediated MP phosphorylation integrates PXY and cytokinin 878 signalling in secondary growth. Nat Plants 4, 605-614 (2018). 879 81 Wu, M. F. et al. Auxin-regulated chromatin switch directs acquisition of flower 880 primordium founder fate. eLife 4, e09269 (2015). 881 82 Neff, M. M., Turk, E. & Kalishman, M. Web-based primer design for single nucleotide 882 polymorphism analysis. Trends Genetics 18, 613-615 (2002). 883 83 Silverstone, A. L. et al. Functional analysis of SPINDLY in gibberellin signaling in 884 Arabidopsis. Plant Physiol 143, 987-1000 (2007). 885 84 Hartweck, L. M., Scott, C. L. & Olszewski, N. E. Two O-Linked N-acetylglucosamine 886 transferase genes of Arabidopsis thaliana L. Heynh. have overlapping functions 887 necessary for gamete and seed development. Genetics 161, 1279-1291 (2002). 888 85 Xing, L. et al. Arabidopsis O-GlcNAc transferase SEC activates histone 889 methyltransferase ATX1 to regulate flowering. EMBO J 37 (2018). 890 86 Reed, J. W. et al. Three auxin response factors promote hypocotyl elongation. Plant 891 Physiol 178, 864-875 (2018). 892 87 Nagpal, P. et al. Auxin response factors ARF6 and ARF8 promote jasmonic acid 893 production and flower maturation. Development 132, 4107-4118 (2005). 894 88 Harper, R. M. et al. The NPH4 locus encodes the auxin response factor ARF7, a 895 conditional regulator of differential growth in aerial Arabidopsis tissue. Plant Cell 12, 896 757-770 (2000). 897 89 Wilmoth, J. C. et al. NPH4/ARF7 and ARF19 promote leaf expansion and auxin-induced 898 lateral root formation. Plant J 43, 118-130 (2005). 899 90 Dietrich, K. et al. Heterodimers of the Arabidopsis transcription factors bZIP1 and 900 bZIP53 reprogram amino acid metabolism during low energy stress. Plant Cell 23, 381-901 395 (2011). 902 91 Rademacher, E. H. et al. A cellular expression map of the Arabidopsis AUXIN 903 RESPONSE FACTOR gene family. Plant J 68, 597-606 (2011). 904 92 Kidd, B. N. et al. The mediator complex subunit PFT1 is a key regulator of jasmonate-905 dependent defense in Arabidopsis. Plant Cell 21, 2237-2252 (2009). 906 93 Hu, J. et al. Potential sites of bioactive gibberellin production during reproductive growth 907 in Arabidopsis Plant Cell 20, 320-336 (2008). 908 94 Earley, K. W. et al. Gateway-compatible vectors for plant functional genomics and 909 proteomics. Plant J 45, 616-629 (2006). 910 95 Kumar, S. et al. Structure and dynamics of the Arabidopsis O-fucosyltransferase 911 SPINDLY. Nat Commun 14, 1538 (2023). 912 96 Hu, J., Israeli, A., Ori, N. & Sun, T. P. The Interaction between DELLA and ARF/IAA 913 mediates crosstalk between gibberellin and auxin signaling to control fruit initiation in 914 tomato. Plant Cell 30, 1710-1728 (2018). 915 97 Bowman, J. Arabidopsis: An Atlas of Morphology and Development. ISBN : 978-1-4612-916 7600-5 (Springer-Verlag, 1994). 917 .CC-BY-NC-ND 4.0 International licenseavailable under a (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 The copyright holder for this preprintthis version posted June 27, 2024. ; https://doi.org/10.1101/2024.06.26.599170doi: bioRxiv preprint 29 98 Zhang, Z. L. et al. SCARECROW-LIKE 3 promotes gibberellin signaling by 918 antagonizing DELLA in Arabidopsis. Proc Natl Acad Sci USA 108, 2160-2165 (2011). 919 99 Zhou, X. et al. The ERF11 Transcription Factor Promotes Internode Elongation by 920 Activating Gibberellin Biosynthesis and Signaling. Plant Physiol 171, 2760-2770 (2016). 921 100 Tian, W. et al. SDC mediates DNA methylation-controlled clock pace by interacting with 922 ZTL in Arabidopsis. Nucleic Acids Res 49, 3764-3780 (2021). 923 101 Heberle, H., Meirelles, G. V., da Silva, F. R., Telles, G. P. & Minghim, R. InteractiVenn: 924 a web-based tool for the analysis of sets through Venn diagrams. BMC Bioinformatics 16, 925 169 (2015). 926 102 Mi, H. et al. Protocol Update for large-scale genome and gene function analysis with the 927 PANTHER classification system (v.14.0). Nature Protocols 14, 703-721 (2019). 928 103 Pang, Z. et al. MetaboAnalyst 5.0: narrowing the gap between raw spectra and functional 929 insights. Nucleic Acids Res 49, W388-W396 (2021). 930 104 Huang, X. et al. The master growth regulator DELLA binding to histone H2A is essential 931 for DELLA-mediated global transcription regulation. Nat Plants 9, 1291-1305 (2023). 932 105 Udeshi, N. D., Compton, P. D., Shabanowitz, J., Hunt, D. F. & Rose, K. L. Methods for 933 analyzing peptides and proteins on a chromatographic timescale by electron-transfer 934 dissociation mass spectrometry. Nature Protocols 3, 1709-1717 (2008). 935 106 Bern, M., Kil, Y. J. & Becker, C. Byonic: advanced peptide and protein identification 936 software. Curr Protoc Bioinformatics 40, 13.20.11-13.20.14 (2012). 937 107 Perez-Riverol, Y. et al. The PRIDE database resources in 2022: a hub for mass 938 spectrometry-based proteomics evidences. Nucleic Acids Res 50, D543-D552 (2022). 939 940 941 .CC-BY-NC-ND 4.0 International licenseavailable under a (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 The copyright holder for this preprintthis version posted June 27, 2024. ; https://doi.org/10.1101/2024.06.26.599170doi: bioRxiv preprint 30 FIGURE LEGENDS 942 943 Figure 1. spy mutations promoted parthenocarpic fruit growth. a-d, spy mutations in both 944 Ler and Col-0 backgrounds promoted parthenocarpic fruit growth. In a and c, photo showing 945 representative pistils of different genotypes 7 days after emasculation. Bar = 2 mm. In b and d, 946 pistil lengths. n>15. e-f, the nuclear-localized SPY partially rescued the pistil phenotype of spy-947 3. In e, Bar = 2 mm. In f, pistil lengths. n>15. GFP-SPY, GFP-SPY-NLS and GFP-SPY-NES 948 were accumulated at similar levels in these lines (Supplementary Fig. 1b). In boxplots b, d and 949 f, center lines and box edges are medians and the lower/upper quartiles, respectively. Whiskers 950 extend to the lowest and highest data points within 1.5x interquartile range (IQR) below and 951 above the lower and upper quartiles, respectively. Different letters above the boxes represent 952 significant differences (p < 0.05) as determined Tukey's honestly significant difference (HSD) 953 mean separation test. Two biological repeats showed similar results. 954 955 Figure 2. ARF6 and ARF8 mediated auxin-induced parthenocarpic growth downstream of 956 SPY. a-b, epistasis analysis of spy-3, arf6-2 and arf8-3 mutations. spy-3 and arf6 additively 957 promoted parthenocarpic fruit elongation, whereas spy-3 arf8 showed similar phenotype as arf8. 958 Mutant alleles are homozygous unless specified as heterozygous, including arf6 +/- and arf8 +/-. 959 In a, photo showing representative pistils of different genotypes 7 days after emasculation. Bar = 960 5 mm. In b, pistil lengths. n>15. c-d, the spy arf6 double mutant was responsive to picloram 961 treatment, whereas spy arf8 was not. All mutants were responsive to GA. In c, photo showing 962 representative pistils of different genotypes 7 days after emasculation and treatment with mock 963 solvent (M) or 50 µM picloram (A, for auxin analog) or 100 µM GA3 (G). Pistils from self-964 pollinated flowers (P) were included for comparison. Bar = 5 mm. In d, pistil lengths. n>15. In 965 boxplots b and d, center lines and box edges are medians and the lower/upper quartiles, 966 respectively. Whiskers extend to the lowest and highest data points within 1.5x IQR below and 967 above the lower and upper quartiles, respectively. Different letters above the boxes represent 968 significant differences (p < 0.05) as determined by Tukey's HSD mean separation test. Two 969 biological repeats showed similar results. 970 971 .CC-BY-NC-ND 4.0 International licenseavailable under a (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 The copyright holder for this preprintthis version posted June 27, 2024. ; https://doi.org/10.1101/2024.06.26.599170doi: bioRxiv preprint 31 Figure 3. Identification of O-fucosylation sites in ARF6 and ARF8. a, FLAG-ARF6 was O-972 fucosylated by SPY in Arabidopsis. b, FLAG-ARF8 was O-fucosylated by SPY in Arabidopsis. 973 In a-b, FLAG-ARF6/8 proteins were affinity-purified from transgenic Arabidopsis carrying 974 PUBQ10:FLAG-ARF6 or -ARF8 in WT or spy-8 background, and protein blots were probed with 975 either AAL-biotin or anti-FLAG antibody as labeled. In a-b, arrow in the top panel indicates 976 FLAG-ARF6 or -ARF8, and arrow in the bottom panel indicates O-fucosylated FLAG-ARF6 or 977 -ARF8. In b, * indicates a non-specific background band. c-d, ARF6 and ARF8 O-fucosylation 978 sites identified by MS analysis. The schematic shows the ARF6 (c) or ARF8 protein (d); The 979 marked S/T residues are confirmed O-Fuc sites. The sequence within square brackets contains 980 undetermined O-Fuc sites. *, also identified in a recent proteomic study35. e, AAL pulldown 981 assay confirmed that MR-ARF6 contains major O-Fuc site(s). FLAG-tagged full-length (FL) or 982 truncated ARF6 proteins were expressed alone (–) or co-expressed (+) with Myc-SPY in N. 983 benthamiana. FLAG-GFP, a negative control. O-fucosylated proteins were pull-downed by 984 AAL-agarose. Immunoblot containing input (top panel) or AAL-agarose pull-down samples 985 (bottom panel) was probed with anti-FLAG and anti-Myc antibodies as labeled. PS, Ponceau S-986 stained blot showing even loading. N, N-terminal DBD domain; MR, middle region; C, C-987 terminal PB1 domain of ARF6. Two biological repeats showed similar results. 988 989 Figure 4. Identification of ARF6-, ARF8- and SPY-responsive genes in pistils by RNA-seq 990 analysis. RNA-seq analysis was performed using −2 DAA pistils (stage 10) of arf6, arf8, spy-3 991 and WT. The differentially expressed gene (DEG) lists for ARF6-, ARF8- and SPY-responsive 992 genes are in Supplementary Table 2. a-c, Venn diagrams of coregulated DEGs by ARF6, ARF8 993 and SPY. Total DEGs in a, Up-regulated DEGs in b, Down-regulated DEGs in c. d, Heat map of 994 SPY and ARF8 coregulated 181 DEGs. e, Enrichment of selected biological processes in ARF8 995 and SPY co-regulated 181 DEGs by GO term analysis. 996 997 Figure 5. Confirmation of ARF6/ARF8 target genes by RT-qPCR and ChIP-qPCR. a-b, 998 RT-qPCR analysis confirming selected genes that were upregulated (in a) or downregulated (in 999 b) in -2 DAA pistils of arf8 and spy-3 mutants in comparison to WT. These genes were also 1000 upregulated or downregulated, respectively, after pollination (0 DAA WT vs –2 DAA WT 1001 pistils). For all RT-qPCR analyses, the housekeeping gene PP2A was used to normalize different 1002 .CC-BY-NC-ND 4.0 International licenseavailable under a (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 The copyright holder for this preprintthis version posted June 27, 2024. ; https://doi.org/10.1101/2024.06.26.599170doi: bioRxiv preprint 32 samples. Means ± SE of 3 biological replicas are shown. Expression level in –2 DAA WT pistil 1003 was set to 1. c-d, ChIP-qPCR analysis showed ARF6 (in c) and ARF8 (in d) binding to promoter 1004 regions of selected ARF-responsive genes, although spy mutation did not affect ARF binding. -2 1005 DAA pistils of the PUBQ10:FLAG-ARF6/ARF8 lines in WT or spy mutant backgrounds and anti-1006 FLAG beads were used for the ChIP experiment. The relative enrichment was calculated by 1007 normalizing against ChIP-qPCR of non-transgenic WT samples using PP2A as control. Means ± 1008 SE of 3 biological replicas are shown. In a-d, Different letters above the bars represent 1009 significant differences (p < 0.05) as determined by Tukey's HSD mean separation test. 1010 1011 Figure 6. ARF6/8-IAA9 transcription repression activities were enhanced by SPY, while 1012 ARF6/8 transactivation activities were reduced by SPY. a-b, Dual luciferase assay in the N. 1013 benthamiana transient expression system showing the opposing effect of SPY on ARF vs 1014 ARF+IAA9. 35S:Renilla LUC (rLUC) was the internal control for transformation efficiency. The 1015 reporter construct contained P3(2x):fLUC. Effector constructs included 35S:FLAG-ARF6, 1016 35S:FLAG-ARF8, 35S:Myc-IAA9P188S, or 35S:HA-SPY as labeled. In a, relative fLUC activity 1017 was calculated by normalizing with rLUC activity in each sample. Means ± SE of 3 biological 1018 replicas are shown. Different letters above the bars represent significant differences (p < 0.05) as 1019 determined by Tukey's HSD mean separation test. *** p = 0.0002. In b, each effector protein 1020 was expressed at similar levels in different samples. Effector proteins in N. benthamiana extracts 1021 were detected by immunoblot using anti-FLAG, anti-Myc and anti-HA antibodies as labeled. c, 1022 Co-IP assay showing that SPY did not affect ARF6-IAA9 interaction in N. benthamiana. HA-1023 IAA9P188S was expressed alone or co-expressed with FLAG-ARF6, Myc-SPY or FLAG-1024 ARF6+Myc-SPY or Myc-GFP (a negative control). Anti-HA beads were used for IP, and input 1025 and IP’ed samples were detected with anti-HA, anti-Myc and anti-FLAG antibodies, separately. 1026 FLAG-ARF6 in the IP eluate from HA-IAA9P188S+FLAG-ARF6 sample was set as 1.0. d, Co-IP 1027 assay showing that SPY did not affect ARF6-ARF8 dimerization in N. benthamiana. FLAG-1028 ARF6 was expressed alone or co-expressed with Myc-ARF8 +/- HA-SPY. Myc-GFP was 1029 included as a negative control. Anti-FLAG beads were used for IP, and input and IP’ed samples 1030 were analyzed by immunoblotting with different antibodies as labeled. Myc-ARF8 in the IP 1031 eluate from FLAG-ARF6+Myc-ARF8 sample was set as 1.0. In b-d, PS-stained blot images 1032 showing even loading. In a-d, two biological repeats showed similar results. 1033 .CC-BY-NC-ND 4.0 International licenseavailable under a (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 The copyright holder for this preprintthis version posted June 27, 2024. ; https://doi.org/10.1101/2024.06.26.599170doi: bioRxiv preprint 33 1034 Figure 7. SPY reduced ARF6-MED8 interaction. a-b, ARF6 and ARF8 interacted with MED8 1035 and the MR of ARFs contain the MED8 binding sequence in Y2H assay. The strength of 1036 interaction was indicated by the ability of cells to grow on –His plates with 0-100 mM 3-AT as 1037 labeled. c, Co-IP assay showing SPY but not spy-19 reduced ARF6-MED8 interaction in N. 1038 benthamiana. FLAG-ARF6 was expressed alone or co-expressed with Myc-MED8 –/+ HA-SPY, 1039 HA-spy-19 or Myc-GFP (a negative control). Anti-Myc beads were used for IP. Input and IP’ed 1040 samples were detected by immunoblot analysis as labeled. The FLAG-ARF6 protein levels in the 1041 IP eluate from FLAG-ARF6+Myc-MED8 sample was set as 1.0. d, Co-IP assay in Arabidopsis 1042 showing spy mutation enhanced ARF6-MED8 interaction. Transgenic lines carrying either 1043 PUBQ10:FLAG-ARF6 or both PUBQ10:FLAG-ARF6 and 35S:MED8-GFP in WT or spy-8 1044

were used for IP with anti-GFP beads. The input and IP’ed samples were detected by 1045 immunoblot analysis as labeled. The FLAG-ARF6 protein levels in the IP eluate from FLAG-1046 ARF6+GFP-MED8 sample was set as 1.0. In c-d, PS-stained blot images showing even loading. 1047 In a-d, two biological repeats showed similar results. 1048 1049 Figure 8. Model of regulatory mechanism of ARF activity by SPY-mediated O-fucosylation 1050 in fruit growth. Before pollination, the IAA9-ARF6/8 complexes function as transcription 1051 repressors to inhibit fruit set. SPY O-fucosylates ARF6/8, IAA9 and MED8, a subunit of the core 1052 Mediator complex (cMED), which reduces ARF-MED8 interaction to enhance transcription 1053 repression activities of the ARF-IAA9 complexes. The co-repressor TPL, recruited by IAA9, 1054 also interferes with ARF binding to cMED15 and may recruit the CKM repressive module (not 1055 shown) to block transcription of ARF target genes16. After pollination, elevated auxin levels in 1056 the pistil trigger IAA9 degradation and release ARF6 and ARF8 homo and/or hetero-dimers to 1057 activate fruit growth-related genes by recruiting the coactivator Mediator complex and 1058 promoting the assembly of RNA Pol II preinitiation complex (PIC). In addition, SPY protein 1059 level and/or activity are reduced after pollination via an unknown mechanism. This further 1060 enhances ARF6/8 transactivation activities by promoting ARF-MED8 interaction. 1061 1062 Supplementary Figure 1. SPY expression patterns and levels in pistils. a, Confocal 1063 microscopy showing localization of GFP-SPY in both cytoplasm and nucleus, GFP-SPY-NES in 1064 .CC-BY-NC-ND 4.0 International licenseavailable under a (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 The copyright holder for this preprintthis version posted June 27, 2024. ; https://doi.org/10.1101/2024.06.26.599170doi: bioRxiv preprint 34 the cytoplasm, and GFP-SPY-NLS in the nucleus. Images of pistils at 2 days before anthesis (–2 1065 DAA). Bar = 20 µm. b, GFP-SPY, GFP-SPY-NLS and GFP-SPY-NES were accumulated at 1066 similar levels in these transgenic lines. Protein blot was probed with an anti-GFP antibody. c, 1067 FLAG-SPY was reduced in the PSPY:FLAG-SPY spy-3 line after anthesis. d, PSPY:FLAG-SPY 1068 spy-3 line showed reduced protein O-fucosylation at 3 DAA and 5 DAA compared to that at –2 1069 DAA and 0 DAA. O-fucosylated proteins in total proteins extracted from the PSPY:FLAG-SPY 1070 spy-3 line and spy-3 (a negative control) were detected by protein blot analysis using AAL-1071 biotin. * indicates reduced O-fucosylated proteins. In b-d, Ponceau S (PS)-stained blot showing 1072 protein loading. In a-d, two biological repeats showed similar results. 1073 1074 Supplementary Figure 2. SPY-regulated fruit growth is mediated by ARF8 together with 1075 ARF6, but not ARF7 (NPH4). a, Pistil width of spy-3, arf6-2 and arf8-3 mutants. Mutant 1076 alleles are homozygous unless specified as heterozygous, including arf6+/- and arf8+/-. b-c, 1077 Epistasis analysis of arf6, nph4-1 and arf8. The nph4-1 mutation did not display parthenocarpy 1078 after emasculation. In b, photo showing representative pistils of different genotypes 7 days after 1079 emasculation. Bar = 2 mm. In c, pistil lengths, n>15. In boxplots a and c, center lines and box 1080 edges are medians and the lower/upper quartiles, respectively. Whiskers extend to the lowest and 1081 highest data points within 1.5x interquartile range (IQR) below and above the lower and upper 1082 quartiles, respectively. Different letters above the boxes represent significant differences (p < 1083 0.05) as determined by Tukey's HSD mean separation test. In a-c, two biological repeats showed 1084 similar results. 1085 1086 Supplementary Figure 3. FLAG-ARF6/8 partially rescued the arf6 and arf8 parthenocarpic 1087 phenotype. a, photo showing representative pistils of different lines 7 days after emasculation. 1088 Bar = 2 mm. b, pistil lengths. n>15. The center lines and box edges in the box plot are medians 1089 and the lower/upper quartiles, respectively. Whiskers extend to the lowest and highest data 1090 points within 1.5x IQR below and above the lower and upper quartiles, respectively. Different 1091 letters above the boxes represent significant differences (p < 0.05) as determined by Tukey's 1092 HSD mean separation test. Two biological repeats showed similar results. 1093 1094 .CC-BY-NC-ND 4.0 International licenseavailable under a (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 The copyright holder for this preprintthis version posted June 27, 2024. ; https://doi.org/10.1101/2024.06.26.599170doi: bioRxiv preprint 35 Supplementary Figure 4. ARF6 and ARF8 are O-fucosylated and O-GlcNAcylated. a-b, 1095 FLAG-ARF6 and -ARF8 were O-fucosylated by SPY and O-GlcNAcylated by SEC, 1096 respectively. FLAG-ARF6/8 proteins were expressed alone or co-expressed with SPY or SEC in 1097 N. benthamiana. Immunoblots containing affinity-purified FLAG-ARF6/8 were probed with 1098 anti-FLAG, AAL-biotin or anti-O-GlcNAc antibody as labeled. Arrow in the top panel indicates 1099 FLAG-ARF6 or -ARF8, arrow in the middle panel indicates O-fucosylated FLAG-ARF6 or -1100 ARF8, and arrow in the bottom panel indicates O-GlcNAcylated FLAG-ARF6 or -ARF8. c-d, 1101 sec mutants did not alter growth of unpollinated pistils. c, photo showing representative pistils of 1102 different genotypes 7 days after emasculation. Bar = 1 mm. d, pistil lengths. n>15. The center 1103 lines and box edges in the box plot are medians and the lower/upper quartiles, respectively. 1104 Whiskers extend to the lowest and highest data points within 1.5x IQR below and above the 1105 lower and upper quartiles, respectively. Different letters above the boxes represent significant 1106 differences (p < 0.05) as determined by Tukey's HSD mean separation test. Two biological 1107 repeats showed similar results. 1108 1109 Supplementary Figure 5. PTM sites in ARF6 and ARF8. a-b, O-Fuc, O-GlcNAc and 1110 phosphorylation sites in ARF6 and ARF8 identified by MS analysis. The schematic shows the 1111 ARF6 (a) or ARF8 protein (b); The marked S/T residues are confirmed PTM sites. The sequence 1112 within brackets contains PTM(s) for which the specific residue(s) could not be determined. * 1113 indicates PTM that was reported previously35,47. 1114 1115 Supplementary Figure 6. Identification of coregulated genes among fertilization-responsive 1116 vs ARF6-, ARF8- and SPY-responsive genes in pistils by RNA-seq analysis. RNA-seq 1117 analysis was performed using −2 DAA pistils of arf6, arf8, spy-3 and WT, and 0 DAA WT 1118 pistils. The differentially expressed gene (DEG) lists for ARF6-, ARF8- and SPY-responsive 1119 genes, and for fertilization-responsive genes (WT 0 DAA vs WT –2 DAA) are in 1120 Supplementary Table 2. a, Total, up- or down-regulated DEGs in WT 0 DAA or in each mutant 1121 (vs WT –2 DAA). b-d. Venn diagrams of coregulated DEGs among fertilization (WT 0 DAA vs 1122 WT –2 DAA), arf6, arf8 and spy-3. e, Heat map of coregulated genes among fertilization-, 1123 ARF8- and SPY-responsive DEGs. f, Venn diagram of overlapping genes between total 1124 SPY/ARF8 coregulated genes (181 DEGs) and the ARF6 ChIP-seq gene list51. 1125 .CC-BY-NC-ND 4.0 International licenseavailable under a (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 The copyright holder for this preprintthis version posted June 27, 2024. ; https://doi.org/10.1101/2024.06.26.599170doi: bioRxiv preprint 36 1126 Supplementary Figure 7. The bzip1 and iaa9 mutants displayed longer pistils after 1127 emasculation. In a and c, photos showing representative pistils of different genotypes 7 days 1128 after emasculation. Bar = 2 mm. In b and d, pistil lengths were shown in boxplots. n>15. The 1129 center lines and box edges are medians and the lower/upper quartiles, respectively. Whiskers 1130 extend to the lowest and highest data points within 1.5x IQR below and above the lower and 1131 upper quartiles, respectively. Different letters (in b) or the asterisk (in d) above the boxes 1132 represent significant differences (p < 0.05) as determined by Tukey's HSD mean separation test. 1133 Two biological repeats showed similar results. 1134 1135 Supplementary Figure 8. ARF6 and ARF8 protein accumulation or nuclear localization 1136 were not affected by spy. a, PUBQ10:FLAG-ARF6 and PUBQ10:FLAG-ARF8 in WT vs spy 1137 background. Immunoblots containing total proteins extracted from seedlings were probed with 1138 anti-FLAG or anti-tubulin antibody. b-d, PARF6:ARF6-GFP in WT vs spy background. In c, 1139 PARF7:ARF7-YFP in WT background was included as a control. GFP/YFP signals detected by 1140 confocal microscopy showing root tips (b) or upper roots (c) of 3d-old seedlings or stage-10 1141 pistils (d). In c, ARF6-GFP was only detected in the nuclei of root cells in the maturation zone, 1142 whereas ARF7-YFP localized in cytoplasmic condensates. In b-c, roots were stained with 1143 propidium iodide before imaging. In b-d, bar = 10 µm. 1144 1145 Supplementary Figure 9. ARF6 transactivation activities were reduced by SPY, but not by 1146 spy-19. a-b, Dual luciferase assay in the N. benthamiana transient expression system. 1147 35S:Renilla LUC (rLUC) was the internal control for transformation efficiency. The reporter 1148 construct contained P3(2x):fLUC. Effector constructs included 35S:FLAG-ARF6 and/or 35S:HA-1149 SPY, and/or 35S:HA-spy-19 as labeled. In a, relative fLUC activity was calculated by 1150 normalizing with rLUC activity in each sample. Means ± SE of 3 biological replicas are shown. 1151 Different letters above the bars represent significant differences (p < 0.05) as determined by 1152 Tukey's HSD mean separation test. In b, each effector protein was expressed at similar levels in 1153 different samples. Effector proteins in N. benthamiana extracts were detected by immunoblot 1154 using anti-FLAG, anti-Myc and anti-HA antibodies as labeled. Ponceau S (PS)-stained gel 1155 images showing similar sample loading. Two biological repeats showed similar results. 1156 .CC-BY-NC-ND 4.0 International licenseavailable under a (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 The copyright holder for this preprintthis version posted June 27, 2024. ; https://doi.org/10.1101/2024.06.26.599170doi: bioRxiv preprint 37 1157 Supplementary Figure 10. AAL pulldown assay showed that N-terminal domain of IAA9 1158 was O-fucosylated by SPY. FLAG-tagged full-length (FL) or truncated IAA9 proteins were 1159 expressed alone (–) or co-expressed (+) with Myc-SPY in N. benthamiana. FLAG-GFP, a 1160 negative control. O-fucosylated proteins were pull-downed by AAL-agarose. Immunoblot 1161 containing input (top panel) or AAL-agarose pull-down samples (bottom panel) was probed with 1162 anti-FLAG and anti-Myc antibodies as labeled. PS, Ponceau S-stained blot showing even 1163 loading. N, N-terminal domain of IAA9 (amino acid residues 1-208); C, C-terminal PB1 domain 1164 of IAA9 (amino acid residues 209-326). Two biological repeats showed similar results. 1165 1166 Supplementary Figure 11. SPY O-fucosylates MED8. a, ARF6/8 did not interact with MED25 1167 in Y2H assay. b, MED8 was O-fucosylated by SPY. FLAG-MED8 was expressed alone (–) or 1168 co-expressed (+) with SPY in N. benthamiana. Immunoblots containing affinity-purified FLAG-1169 MED8 proteins were probed with AAL-biotin or anti-FLAG antibody as labeled. c-d, The 1170 35S:MED8-GFP transgene rescued the late flowering phenotype of med8-2. In c, photo was 1171 taken at 30d-old, and bar = 2 cm. In d, n=10. The center lines and box edges in the box plot are 1172 medians and the lower/upper quartiles, respectively. Whiskers extend to the lowest and highest 1173 data points within 1.5x IQR below and above the lower and upper quartiles, respectively. 1174 Different letters above the boxes represent significant differences (p < 0.05) as determined by 1175 Tukey's HSD mean separation test. Two biological repeats showed similar results. 1176 1177 Supplementary Table 1. Summary of ARF6 and ARF8 PTMs by MS analysis 1178 Supplementary Table 2. RNA-seq DEGs 1179 Supplementary Table 3. Co-regulated gene lists 1180 Supplementary Table 4. GO term analysis of SPY/ARF8 coregulated genes 1181 Supplementary Table 5. Selected genes for RT-qPCR and ChIP-qPCR 1182 Supplementary Table 6. Overlap between SPY/ARF8-DEGs and ARF6 ChIP-seq dataset 1183 Supplementary Table 7. List of primers 1184 Supplementary Table 8. List of constructs 1185 1186 .CC-BY-NC-ND 4.0 International licenseavailable under a (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 The copyright holder for this preprintthis version posted June 27, 2024. ; https://doi.org/10.1101/2024.06.26.599170doi: bioRxiv preprint Ler spy-8 spy-19a Fig. 1 b d SPY-NLS spy-3 GFP-SPY spy-3 SPY-NES spy-3 Col-0 spy-3 ec spy-23 Col-0 spy-3 f L er sp y-8 sp y-19 0 . 0 0 . 2 0 . 4 0 . 6 0 . 8 Pistil length (mm) 0 2 4 6 8 Ler spy-8 spy-19 a b c C ol-0 spy-3 spy-21 0 . 0 0 . 2 0 . 4 0 . 6 0 . 8 1 . 0 Pistil length (mm) 0 2 4 6 8 10 Col-0 spy-3 spy-23 a bc Col-0 spy-3 SPYGFPSPY spy-3 SPYGFPSPYNLS spy-3 SPYGFPSPYNES spy-3 0 2 4 6 8 Col-0 spy-3 Pistil length (mm) 0 8 6 2 4 c a a b bSPY-NLS spy-3 GFP-SPY spy-3 SPY-NES spy-3 Figure 1. spy mutations promoted parthenocarpic fruit growth. a-d, spy mutations in both Ler and Col-0 backgrounds promoted parthenocarpic fruit growth. In a and c, photo showing representative pistils of different genotypes 7 days after emasculation. Bar = 2 mm. In b and d, pistil lengths. n>15. e-f, the nuclear-localized SPY partially rescued the pistil phenotype of spy-3. In e, Bar = 2 mm. In f, pistil lengths. n>15. GFP-SPY, GFP-SPY-NLS and GFP-SPY-NES were accumulated at similar levels in these lines (Supplementary Fig. 1b). In boxplots b, d and f, center lines and box edges are medians and the lower/upper quartiles, respectively. Whiskers extend to the lowest and highest data points within 1.5x interquartile range (IQR) below and above the lower and upper quartiles, respectively. Different letters above the boxes represent significant differences (p < 0.05) as determined Tukey's honestly significant difference (HSD) mean separation test. Two biological repeats showed similar results. .CC-BY-NC-ND 4.0 International licenseavailable under a (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 The copyright holder for this preprintthis version posted June 27, 2024. ; https://doi.org/10.1101/2024.06.26.599170doi: bioRxiv preprint Fig. 2 a c arf6 spy arf6 spy arf8 spy arf8 arf6 arf8 +/- spy arf6 arf8 +/- WT arf8 arf6 +/- spy arf8 arf6 +/- b d P: Pollinated; M: Mock; A: 50 μM picloram; G: 100 μM GA3. WT spy arf6 spy arf6 arf8 spy arf8 P M A G P M A G P M A G P M A G P M A G P M A G P M Auxin GA Col-0 spy-3 arf6 spy-3 arf6 arf8 spy-3 arf8 arf6 arf8+/- spy-3 arf6 arf8+/- arf86+/- spy-3 arf86+/- 0.0 0.2 0.4 0.6 0.8 1.0 Pistil length (cm)Pistil length (mm) 0 4 8 6 2 10 barf6 spy arf6 spy arf8 spy arf8 arf6 arf8 +/- spy arf6 arf8 +/- WT arf8 arf6 +/- spy arf8 arf6 +/- b c a a a b a a a Col-0 #10 Col-0 Mock Col-0 50 uM picloram Col-0 GA spy-3 #10 spy-3 Mock spy-3 50 uM picloram spy-3 GA arf6 #10 arf6 M arf6 50 uM picloram arf6 GA spy-3 arf6 #10 spy3arf6 M spy3arf6 50 uM picloram spy3arf6 GA arf8 #10 arf8 M arf8 50 uM picloram arf8 GA spy-3 arf8 #10 spy3arf8 M spy3arf8 50 uM picloram spy3arf8 GA 0.0 0.5 1.0 1.5 Pistil length (cm) a Pistil length (mm) 0 10 15 5 WT spy arf6 spy arf6 arf8 spy arf8 b c d a b c d a b c d a b c d a b cc ab c d Figure 2. ARF6 and ARF8 mediated auxin-induced parthenocarpic growth downstream of SPY. a-b, epistasis analysis of spy-3, arf6-2 and arf8-3 mutations. spy-3 and arf6 additively promoted parthenocarpic fruit elongation, whereas spy-3 arf8 showed similar phenotype as arf8. Mutant alleles are homozygous unless specified as heterozygous, including arf6 +/- and arf8 +/-. In a, photo showing representative pistils of different genotypes 7 days after emasculation. Bar = 5 mm. In b, pistil lengths. n>15. c-d, the spy arf6 double mutant was responsive to picloram treatment, whereas spy arf8 was not. All mutants were responsive to GA. In c, photo showing representative pistils of different genotypes 7 days after emasculation and treatment with mock solvent (M) or 50 µM picloram (A, for auxin analog) or 100 µM GA3 (G). Pistils from self-pollinated flowers (P) were included for comparison. Bar = 5 mm. In d, pistil lengths. n>15. In boxplots b and d, center lines and box edges are medians and the lower/upper quartiles, respectively. Whiskers extend to the lowest and highest data points within 1.5x IQR below and above the lower and upper quartiles, respectively. Different letters above the boxes represent significant differences (p < 0.05) as determined by Tukey's HSD mean separation test. Two biological repeats showed similar results. .CC-BY-NC-ND 4.0 International licenseavailable under a (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 The copyright holder for this preprintthis version posted June 27, 2024. ; https://doi.org/10.1101/2024.06.26.599170doi: bioRxiv preprint Fig. 3 50 150 a e c d 75 α-FLAG 50 37 25 Input 75 50 37 25 150 AAL Pulldown Myc-SPY + + + + + FL – – –– – kD GFP CN MRFLAG fusion b ARF6 ARF6 T643* 780375 PB1DBD 935MR ARF8 T380* S473[T380 -S382] 690372 [S428 -T432] MR 811PB1DBD PS kD spy FLAG-ARF8 WT 100 75 75 100 [T643 -S654] 150 kD 150 WT spy FLAG-ARF6 α-FLAG AAL α-FLAG AAL * ARF6 FL ARF6 N ARF6 MR ARF6 C GFP O-Fuc- ARF6 FL O-Fuc- ARF6 MR 100 α-Myc (SPY) α-FLAG Figure 3. Identification of O-fucosylation sites in ARF6 and ARF8. a, FLAG-ARF6 was O- fucosylated by SPY in Arabidopsis. b, FLAG-ARF8 was O-fucosylated by SPY in Arabidopsis. In a-b, FLAG-ARF6/8 proteins were affinity-purified from transgenic Arabidopsis carrying PUBQ10:FLAG- ARF6 or -ARF8 in WT or spy-8 background, and protein blots were probed with either AAL-biotin or anti-FLAG antibody as labeled. In a-b, arrow in the top panel indicates FLAG-ARF6 or -ARF8, and arrow in the bottom panel indicates O-fucosylated FLAG-ARF6 or -ARF8. In b, * indicates a non- specific background band. c-d, ARF6 and ARF8 O-fucosylation sites identified by MS analysis. The schematic shows the ARF6 (c) or ARF8 protein (d); The marked S/T residues are confirmed O-Fuc sites. The sequence within square brackets contains undetermined O-Fuc sites. *, also identified in a recent proteomic study35. e, AAL pulldown assay confirmed that MR-ARF6 contains major O-Fuc site(s). FLAG-tagged full-length (FL) or truncated ARF6 proteins were expressed alone (–) or co-expressed (+) with Myc-SPY in N. benthamiana. FLAG-GFP, a negative control. O-fucosylated proteins were pull- downed by AAL-agarose. Immunoblot containing input (top panel) or AAL-agarose pull-down samples (bottom panel) was probed with anti-FLAG and anti-Myc antibodies as labeled. PS, Ponceau S-stained blot showing even loading. N, N-terminal DBD domain; MR, middle region; C, C-terminal PB1 domain of ARF6. Two biological repeats showed similar results. .CC-BY-NC-ND 4.0 International licenseavailable under a (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 The copyright holder for this preprintthis version posted June 27, 2024. ; https://doi.org/10.1101/2024.06.26.599170doi: bioRxiv preprint Fig. 4 a b d c Total DEGs (vs WT) Down-regulated DEGs (vs WT)Up-regulated DEGs (vs WT) 90 137 907 arf6 (108) spy-3 (271) arf8 (1143) 55 440 9 45 71 453 arf6 (80) spy-3 (148) arf8 (597) 41 32 0 7 56 56 464 spy-3 (123) arf8 (546) 15 110 2 arf6 (28) Response to hexose Response to fatty acid Response to nutrient levels Response to oxygen levels Response to carbohydrate Amino acid metabolic process Organic acid metabolic process Response to light stimulus Response to lipid Response to chemical Response to hormone Response to abiotic stimulus Small molecule metabolic process Catabolic process Response to stimulus 2 4 6 8 10 20 40 60 80 12 9 6 -log10 (p value) count Fold Enrichment e 0 -1 -2 -3 3 2 1 spy-3 arf8 (vs WT) log2 (FC) Figure 4. Identification of ARF6-, ARF8- and SPY-responsive genes in pistils by RNA-seq analysis. RNA-seq analysis was performed using −2 DAA pistils (stage 10) of arf6, arf8, spy-3 and WT. The differentially expressed gene (DEG) lists for ARF6-, ARF8- and SPY-responsive genes are in Supplementary Table 2. a-c, Venn diagrams of coregulated DEGs by ARF6, ARF8 and SPY. Total DEGs in a, Up-regulated DEGs in b, Down-regulated DEGs in c. d, Heat map of SPY and ARF8 coregulated 181 DEGs. e, Enrichment of selected biological processes in ARF8 and SPY co- regulated 181 DEGs by GO term analysis. .CC-BY-NC-ND 4.0 International licenseavailable under a (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 The copyright holder for this preprintthis version posted June 27, 2024. ; https://doi.org/10.1101/2024.06.26.599170doi: bioRxiv preprint 0 0.5 1 1.5 GNTL ERF107 bZIP1 SHL Relative mRNA level/PP2A a b bcc a b bc c a b c d a b b b Fig. 5 a b c d 0 2 4 6 8 10 AGP12 ERF023 IAA19 Relative mRNA level/PP2A a b c d aa a b a b b c 0 1 2 3 4 5 AGP12 ERF023 IAA19 GNTL ERF107 bZIP1 SHL Fold enrichment/PP2A a b a a a b a a b b a a ab a b a a b a a b 0 1 2 3 4 5 6 7 AGP12 ERF023 IAA19 GNTL ERF107 bZIP1 SHL Fold enrichment/PP2A a b a a b a aa b b a a b a a b a a b a aWT –2D WT 0D spy arf8 WT spy FLAG-ARF6 FLAG-ARF6 WT spy FLAG-ARF8 FLAG-ARF8 WT –2D WT 0D spy arf8 Figure 5. Confirmation of ARF6/ARF8 target genes by RT-qPCR and ChIP-qPCR. a-b, RT-qPCR analysis confirming selected genes that were upregulated (in a) or downregulated (in b) in -2 DAA pistils of arf8 and spy-3 mutants in comparison to WT. These genes were also upregulated or downregulated, respectively, after pollination (0 DAA WT vs –2 DAA WT pistils). For all RT-qPCR analyses, the housekeeping gene PP2A was used to normalize different samples. Means ± SE of 3 biological replicas are shown. Expression level in –2 DAA WT pistil was set to 1. c-d, ChIP-qPCR analysis showed ARF6 (in c) and ARF8 (in d) binding to promoter regions of selected ARF-responsive genes, although spy mutation did not affect ARF binding. -2 DAA pistils of the PUBQ10:FLAG-ARF6/ARF8 lines in WT or spy mutant backgrounds and anti-FLAG beads were used for the ChIP experiment. The relative enrichment was calculated by normalizing against ChIP-qPCR of non-transgenic WT samples using PP2A as control. Means ± SE of 3 biological replicas are shown. In a-d, Different letters above the bars represent significant differences (p < 0.05) as determined by Tukey's HSD mean separation test. .CC-BY-NC-ND 4.0 International licenseavailable under a (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 The copyright holder for this preprintthis version posted June 27, 2024. ; https://doi.org/10.1101/2024.06.26.599170doi: bioRxiv preprint 50 50 a Fig. 6 b c GFP 150 100 75 50 37 50 1.0 0.8 α-HA α-Myc α-FLAG 150 100 37 50 100 37 α-HA α-Myc α-FLAG PS InputIP (α-HA) HA-IAA9 – ARF6 SPY ARF6+SPY GFP 100 kD ARF6 ARF6 100 50 37 100 150 kD ARF6 ARF8 IAA9 SPY α-HA α-Myc α-FLAG PS 0 1 2 3 4 5 6 7 8 - ARF6 SPY ARF6+SPY IAA9 ARF6+IAA9 IAA9+SPY ARF6+IAA9+SPY ARF8 ARF8+SPY ARF8+IAA9 ARF8+IAA9+SPY Relative fLUC activity *** ef a e b f ef gf b c d e d 100 50 α-Myc α-FLAG α-Myc α-FLAG FLAG-ARF6 – ARF8 ARF8+SPY GFPkD 100 150 37 150 75 50 100 Input 100 75 37 IP (α-FLAG) ARF8 GFP ARF8 GFP PS α-HA (SPY)100 1.0 1.1 _ Figure 6. ARF6/8-IAA9 transcription repression activities were enhanced by SPY, while ARF6/8 transactivation activities were reduced by SPY. a-b, Dual luciferase assay in the N. benthamiana transient expression system showing the opposing effect of SPY on ARF vs ARF+IAA9. 35S:Renilla LUC (rLUC) was the internal control for transformation efficiency. The reporter construct contained P3(2x):fLUC. Effector constructs included 35S:FLAG-ARF6, 35S:FLAG-ARF8, 35S:Myc-IAA9P188S, or 35S:HA-SPY as labeled. In a, relative fLUC activity was calculated by normalizing with rLUC activity in each sample. Means ± SE of 3 biological replicas are shown. Different letters above the bars represent significant differences (p < 0.05) as determined by Tukey's HSD mean separation test. *** p = 0.0002. In b, each effector protein was expressed at similar levels in different samples. Effector proteins in N. benthamiana extracts were detected by immunoblot using anti-FLAG, anti-Myc and anti-HA antibodies as labeled. c, Co-IP assay showing that SPY did not affect ARF6-IAA9 interaction in N. benthamiana. HA-IAA9P188S was expressed alone or co-expressed with FLAG- ARF6, Myc-SPY or FLAG-ARF6+Myc-SPY or Myc-GFP (a negative control). Anti-HA beads were used for IP, and input and IP’ed samples were detected with anti-HA, anti-Myc and anti-FLAG antibodies, separately. FLAG-ARF6 in the IP eluate from HA-IAA9P188S+FLAG-ARF6 sample was set as 1.0. d, Co-IP assay showing that SPY did not affect ARF6-ARF8 dimerization in N. benthamiana. FLAG-ARF6 was expressed alone or co- expressed with Myc-ARF8 +/- HA-SPY. Myc-GFP was included as a negative control. Anti-FLAG beads were used for IP, and input and IP’ed samples were analyzed by immunoblotting with different antibodies as labeled. Myc-ARF8 in the IP eluate from FLAG-ARF6+Myc-ARF8 sample was set as 1.0. In b-d, PS-stained blot images showing even loading. In a-d, two biological repeats showed similar results. .CC-BY-NC-ND 4.0 International licenseavailable under a (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 The copyright holder for this preprintthis version posted June 27, 2024. ; https://doi.org/10.1101/2024.06.26.599170doi: bioRxiv preprint Fig. 7 50 a b MED8 Empty MED8 Empty MED8 MED8 ARF6-NT ARF6-MR ARF6-PB1 ARF8-NT ARF8-MR ARF8-PB1 Prey Empty +HisEmpty 3-AT (mM) 0 10 2 25 –His Bait Empty c d Bait Prey EmptyEmpty –His+His ARF6Empty Empty ARF8 MED8Empty ARF6 ARF8 MED8 MED8 3-AT (mM) 0 0 2 5 10 25 50 75 100 InputIP: GFP – MED8-GFP α-GFP α-FLAG α-GFP α-FLAG150 75 75 150 MED8-GFP spy-8 FLAG-ARF6 2.01.0 100 50 100 PS 37 75 1.0 0.4 37 50 150 75 α-Myc α-FLAG (ARF6) α-Myc α-FLAG (ARF6) InputIP (α –Myc) FLAG-ARF6 – MED8 MED8+SPY GFP 100 α-HA (SPY) PS 150 50 kD MED8 GFP MED8 GFP 0.9 MED8+spy-19 kD Figure 7. SPY reduced ARF6-MED8 interaction. a-b, ARF6 and ARF8 interacted with MED8 and the MR of ARFs contain the MED8 binding sequence in Y2H assay. The strength of interaction was indicated by the ability of cells to grow on –His plates with 0-100 mM 3-AT as labeled. c, Co- IP assay showing SPY but not spy-19 reduced ARF6-MED8 interaction in N. benthamiana. FLAG- ARF6 was expressed alone or co-expressed with Myc-MED8 –/+ HA-SPY, HA-spy-19 or Myc- GFP (a negative control). Anti-Myc beads were used for IP. Input and IP’ed samples were detected by immunoblot analysis as labeled. The FLAG-ARF6 protein levels in the IP eluate from FLAG- ARF6+Myc-MED8 sample was set as 1.0. d, Co-IP assay in Arabidopsis showing spy mutation enhanced ARF6-MED8 interaction. Transgenic lines carrying either PUBQ10:FLAG-ARF6 or both PUBQ10:FLAG-ARF6 and 35S:MED8-GFP in WT or spy-8 background were used for IP with anti- GFP beads. The input and IP’ed samples were detected by immunoblot analysis as labeled. The FLAG-ARF6 protein levels in the IP eluate from FLAG-ARF6+GFP-MED8 sample was set as 1.0. In c-d, PS-stained blot images showing even loading. In a-d, two biological repeats showed similar results. .CC-BY-NC-ND 4.0 International licenseavailable under a (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 The copyright holder for this preprintthis version posted June 27, 2024. ; https://doi.org/10.1101/2024.06.26.599170doi: bioRxiv preprint Fig. 8 Before Pollination (Low Auxin) ARF-IAA9 Complex Transcription Repressor F F DBD DBD MR cMED MED8 PICSPY PB1 PB1 IAA9 TPL MR F F PB1 ARF6/8 MR After Pollination (High Auxin) PICcMEDMED8 ARF Dimer Transcription Activator ARF6/8 TPLIAA9 PB1 PB1 DBD DBD MR SPY SPY Fruit Set & Fruit GrowthFruit Set Figure 8. Model of regulatory mechanism of ARF activity by SPY-mediated O-fucosylation in fruit growth. Before pollination, the IAA9-ARF6/8 complexes function as transcription repressors to inhibit fruit set. SPY O-fucosylates ARF6/8, IAA9 and MED8, a subunit of the core Mediator complex (cMED), which reduces ARF-MED8 interaction to enhance transcription repression activities of the ARF-IAA9 complexes. The co-repressor TPL, recruited by IAA9, also interferes with ARF binding to cMED15 and may recruit the CKM repressive module (not shown) to block transcription of ARF target genes16. After pollination, elevated auxin levels in the pistil trigger IAA9 degradation and release ARF6 and ARF8 homo and/or hetero-dimers to activate fruit growth-related genes by recruiting the coactivator Mediator complex and promoting the assembly of RNA Pol II preinitiation complex (PIC). In addition, SPY protein level and/or activity are reduced after pollination via an unknown mechanism. This further enhances ARF6/8 transactivation activities by promoting ARF-MED8 interaction. .CC-BY-NC-ND 4.0 International licenseavailable under a (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 The copyright holder for this preprintthis version posted June 27, 2024. ; https://doi.org/10.1101/2024.06.26.599170doi: bioRxiv preprint