zhou et al., 1 research article · 30/03/2020 · 102 tissues, embryo, ghost (seed coat +...

Zhou et al.,

1

RESEARCH ARTICLE 1 2 Mechanism of auxin and GA crosstalk in diploid strawberry fruit initiation 3 4 Junhui Zhou, John Sittmann, Lei Guo, Xiaolong Huang, Anuhya Pulapaka and Zhongchi 5 Liua 6 7 Dept. of Cell Biology and Molecular Genetics 8 University of Maryland 9 College Park, MD 20742 10 USA 11 12 aCorresponding author: [email protected] 13 ORCID: 0000-0001-9969-9381 14 15

Short title: Mechanism of strawberry fruit initiation 16

One sentence summary: Genome editing and network analysis are applied to 17 investigating the mechanism of accessory fruit initiation in the wild strawberry, revealing 18 conserved as well as novel crosstalk mechanisms. 19

The author responsible for distribution of materials integral to the findings presented in 20 this article in accordance with the policy described in the Instructions for Authors 21 (www.plantcell.org) is: Zhongchi Liu ([email protected]) 22 23 24 25 Abstract 26 Strawberry, a high value fruit crop, has recently become amenable for genetic studies due 27 to genomic resources and CRISPR/CAS9 tools. Unlike ovary-derived botanical fruits, 28 strawberry is an accessory fruit derived from receptacle, the stem tip subtending floral 29 organs. Although both botanical and accessory fruits initiate development in response to 30 auxin and GA released from seeds, the downstream auxin and GA signaling mechanisms 31 underlying accessory fruit development remain unknown. Using wild strawberry, we 32 performed in depth molecular characterizations of accessory fruit development. We show 33 that auxin signaling proteins FveARF8/FveARF6 are bound and hence inhibited by 34 FveIAA4 and FveRGA1, repressors in auxin and GA signaling pathways. This inhibition 35 is relieved post-fertilization or by the application of GA or auxin. Mutants of FveRGA1 36 developed parthenocarpic fruit suggesting a conserved function of DELLA proteins in 37 fruit set. Further, FveARF8 was found to repress the expression of a GA receptor gene 38 GID1c to control fruit’s sensitivity to GA, revealing a novel crosstalk mechanism. We 39 demonstrate that consensus co-expression network provides a powerful tool for non-40 model species in the selection of interacting genes for functional studies. These findings 41 will facilitate the improvement of strawberry fruit productivity and quality by guiding 42 future production of parthenocarpic fruit. 43 44

was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (whichthis version posted April 1, 2020. ; https://doi.org/10.1101/2020.03.30.016188doi: bioRxiv preprint

https://doi.org/10.1101/2020.03.30.016188

Zhou et al.,

2

Introduction 45

Strawberry is a high-value fruit crop immensely important for agriculture and human 46

nutrition. Unlike most fruits developed from the ovary wall, strawberry fruit is derived 47

from receptacle, the stem tip supporting floral organs (Hollender et al., 2012). About 200 48

achenes, ovaries each with a single seed inside, dot the surface of the receptacle, making 49

them highly accessible for manipulation. Nitsch’s 1950 experiment showed that removal 50

of achenes from a strawberry receptacle prevented receptacle fruit enlargement. However, 51

exogenous application of auxin could substitute for the achenes and stimulated receptacle 52

fruit growth, indicating achenes as the source of auxin (Nitsch 1950). Later work showed 53

that emasculated wild strawberry flowers could still develop into fruit if auxin or GA, 54

alone or in combination, was supplied exogenously (Kang et al., 2013; Liao et al., 2018). 55

Transcriptome analyses and reporter gene expression confirm that fertilization-induced 56

biosynthesis of auxin and GA occurs in the embryo and endosperm (Kang et al., 2013; 57

Feng et al., 2019). The phytohormones are then transported to the receptacle to stimulate 58

fruit enlargement. Exactly how auxin and GA interact to coordinately regulate receptacle 59

fruit development is not known. 60

61

The fertilization-induced phytohormone synthesis for “fruit set”, the no return 62

commitment of the floral tissue to enter fruit pathway, is an evolutionarily conserved 63

strategy in angiosperms. The unique strawberry fruit structure as well as prior research 64

provide a strong foundation for further investigations into signal communication between 65

achenes and the receptacle and auxin/GA crosstalk during fruit development. This is 66

further aided by the rich genomic resources including genome sequencing of wild 67

strawberry Fragaria vesca and garden strawberry Fragaria x ananassa (Shulaev et al., 68

2011; Edger et al., 2018; Edger et al., 2019) as well as extensive transcriptome data and 69

network analyses during flower and fruit development (Kang et al., 2013; Hollender et al., 70

2014; Sanchez-Sevilla et al., 2017; Shahan et al., 2018). As garden strawberry is an allo-71

octoploid, wild strawberry F. vesca, a diploid species, is used in this study. 72

73

Prior research in model organisms illuminated auxin and GA signaling cascade. When 74

auxin is absent, IAAs bind to ARFs to prevent ARFs from activating or repressing 75


https://doi.org/10.1101/2020.03.30.016188

Zhou et al.,

3

downstream target genes. When auxin is synthesized or supplied, it binds its receptors 76

TIR/AFBs, which are F-box proteins that recruit and add ubiquitins to the Aux/IAA 77

proteins. This leads to the 26S proteasome-mediated degradation of Aux/IAAs and the 78

release of ARFs to execute their functions (Leyser, 2018). Similarly, in the absence of 79

GA, DELLA repressor proteins accumulate and bind transcription factors to prevent their 80

function. When GA is produced, GA binds to its receptor GID1 to facilitate binding of 81

GID1 to DELLA that leads to ubiquitination and proteasome-mediated degradation of 82

DELLA. Transcription factors previously bound by DELLA are now free to act upon 83

downstream genes (Daviere and Achard, 2013). 84

85

In model systems Arabidopsis and tomato (both with botanical fruits), fertilization-86

independent fruit formation (parthenocarpy) could be induced when repressors of auxin 87

or GA signaling response are mutated or reduced as demonstrated by arf8 mutants in 88

Arabidopsis and transgenic knockdowns of tomato ARF7, ARF8, or IAA9 (Wang et al., 89

2005; Goetz et al., 2006; Goetz et al., 2007; de Jong et al., 2009; Wang et al., 2009). 90

These results led to the proposal that the IAA9/ARF7/ARF8 repressor complex inhibits 91

fruit set in the ovary tissue of Arabidopsis and tomato. To investigate the relationship 92

between auxin and GA, several studies were carried out in Arabidopsis, pea, and tomato 93

which showed that auxin acts upstream of GA by activating the biosynthesis of GA 94

(Serrani et al., 2008; Dorcey et al., 2009; Ozga et al., 2009; Fuentes et al., 2012). 95

Recently, direct protein-protein interactions between SlDELLA and SlARF7 and SlIAA9 96

were shown to serve as another mechanism mediating crosstalk between auxin and GA in 97

tomato fruit (Hu et al., 2018). 98

99

In contrast, there has been little functional studies of fruit set in accessory fruits including 100

strawberry. Previously, we employed RNA-seq to profile finely dissected strawberry fruit 101

tissues, embryo, ghost (seed coat + endosperm), ovary wall, and receptacle pith and 102

receptacle cortex, at five early fruit developmental stages (Kang et al., 2013). Based on 103

the RNA-seq data and subsequent reporter GUS studies (Feng et al., 2019), the 104

endosperm within the achene emerges as the site of fertilization-induced auxin and GA 105

biosynthesis targeted for fruit development. However, auxin and GA perception and 106


https://doi.org/10.1101/2020.03.30.016188

Zhou et al.,

4

signaling genes, TIR/IAAs and GID1c/RGA1, are highly expressed in the receptacle, 107

indicating receptacle as the site of auxin and GA perception and response (Kang et al., 108

2013). Using the RNA-seq data described above, consensus co-expression network was 109

established (Shahan et al., 2018), which identifies robust co-expression modules in 110

receptacle fruit development. In addition, CRISPR/Cas9 genome editing has been 111

demonstrated in wild and garden strawberries (Zhou et al., 2018; Feng et al., 2019; 112

Martin-Pizarro et al., 2019), providing tools for dissecting gene function. 113

114

Here, we provide an in depth functional dissection of a few key auxin and GA signaling 115

components in the receptacle fruit development in diploid strawberry. We observed both 116

conserved and novel cross-talk mechanisms for GA and auxin signaling, identified 117

FveRGA1 as a key repressor of fruit set in strawberry, and validated the usefulness of 118

consensus co-expression network. The findings lay the foundation for future 119

improvement of strawberry fruit. 120

121

122


https://doi.org/10.1101/2020.03.30.016188

Zhou et al.,

5

Results 123

arf8 mutants develop larger and rounder fruits 124

Previously, we reported successful genome editing of FveARF8, and the arf8 mutants 125

showed faster seedling growth (Zhou et al., 2018). However, arf8 adult plants are similar 126

to WT plants in stature. To determine the defect of arf8 mutants in flower and fruit 127

development, we first characterized four independent CRISPR lines of arf8, two lines are 128

arf8 (-2, -2), one line is arf8 (-2+1, -2+1), and one line is arf8 (-8, -23+5) (Figure S1). 129

Second, we established that fruits derived from primary and secondary flowers are 130

similar in size (Figure S2, allowing us to collect data from both primary and secondary 131

fruits. Third, to remove impact of environment on plant vigor and hence fruit size, the 132

height and width of fruits were normalized respectively to the height and width of petals 133

of the same flower. 134

135

arf8 mutant fruit is larger and rounder than wild type (Figure 1). Significant increase in 136

the width and the height of arf8-1 (-2, -2) fruit is found when compared with wild type 137

(Figure 1A, B); the higher increase of fruit width results in rounder shape in developing 138

fruit. The other three CRISPR-induced arf8 lines exhibited similar phenotype as arf8-1 139

with increased fruit height and width (Figure 1C). Hence, arf8-1 is used for the 140

subsequent analyses. In ripe fruit, the width remains larger in arf8-1 mutants, but the 141

height becomes similar to WT (Figure 1A, D) probably due to compensatory mechanisms 142

at late stages of fruit development. arf8 mutants also exhibited a mild floral phenotype; 143

they showed an increase in petal number (Figure S2C). While 80% wild type flowers 144

have 5 petals, only 40% arf8 mutants produce 5 petals and the remaining flowers produce 145

6 or 7 petals (Figure S2D). The average petal size however appears the same between 146

WT and arf8-1 mutants (Figure S2E). 147

148

Using the Arabidopsis Ubiquitin 10 promoter, FveARF8 was over-expressed in F. vesca; 149

the resulting transgenic lines (#1, #5, and #26), which exhibit a significantly higher 150

FveARF8 transcript level than the non-transgenic control, showed retarded seedling and 151

plant growth (Figure S3). Transgenic lines #1 and #26 showed a reduction of fruit size 152

when normalized to petal (Figure S3C, D), while transgenic line #4 with a wild type 153


https://doi.org/10.1101/2020.03.30.016188

Zhou et al.,

6

expression level of FveARF8 exhibited normal plant statue (Figure S3A, B). Together, 154


https://doi.org/10.1101/2020.03.30.016188

Zhou et al.,

7

our data suggest that FveARF8 is a negative regulator of plant growth and fruit 155

enlargement. 156

157

Mechanism of auxin and GA signaling in fruit set 158

Fertilization-induced phytohormone synthesis and signaling are critical for “fruit set”, a 159

decision to proceed with fruit development upon successful fertilization (Gillaspy et al., 160

1993). In F. vesca, we previously showed that auxin and GA each substituted for 161

fertilization to stimulate fruit set, yet each hormone alone stimulated fruit development to 162

a size smaller than the pollinated control fruit (Kang et al., 2013). Only when both auxin 163

and GA were applied at the same time, did parthenocarpic strawberry develop to the 164

same size as the pollinated control, suggesting the requirement of both auxin and GA for 165

strawberry fruit set. Nevertheless, the molecular mechanism underlying this double 166

requirement is not known. 167

168

To test if FveARF8 could act to block fruit set pre-fertilization, we examined arf8-1 fruit 169

formation in the absence of fertilization. Wild type and arf8-1 mutant flowers were 170

emasculated to prevent fertilization. Fruit size at 0DPA (pre-fertilization) was compared 171

to fruit size at 7 DPA (post-emasculation) within the same genotype. No significant 172

change in fruit size was observed at 7DPA for arf8-1 as well as WT (Figure 2), 173

suggesting that fertilization-induced hormone production is still required for arf8 mutants 174

to develop fruit and that FveARF8 is unlikely involved in preventing fruit set. 175

176

During GA signaling, the GA-GID1 receptor complex targets the DELLA repressor 177

protein for proteasome-mediated degradation, resulting in the release of downstream 178

transcription factors (TFs) normally sequestered by the DELLA protein (Hartweck, 2008; 179

Sun, 2011). In F. vesca, gene06210 (FveRGA1) encodes one of the five annotated 180

probable DELLA proteins and is the only one with a full DELLA motif (Caruana et al., 181

2018). A loss-of-function mutation rga1-1 (also named srl-1) was identified in FveRGA1 182

during a chemical mutagenesis screen (Caruana et al., 2018) (Figure S1B). This mutant 183

offers an opportunity to determine if the DELLA repressor encoded by FveRGA1 184

regulates fruit set in strawberry. rga1-1 receptacle size was measured at Stage 1 (1DPA), 185


https://doi.org/10.1101/2020.03.30.016188

Zhou et al.,

8

stage 4 (9DPA), and stage 5 (11-12 DPA). While the receptacle size difference between 186


https://doi.org/10.1101/2020.03.30.016188

Zhou et al.,

9

WT and rga1-1 is not obvious at 1DPA, the receptacle becomes significantly larger and 187

taller at post-fertilization stages 4-5 (9DPA and 11-12 DPA) (Figure 2C, D), suggesting 188

that constitutive GA response in the rga1-1 receptacle may underlie the larger fruit 189

phenotype. To determine if rga1-1 mutant could develop parthenocarpic fruit, 190

emasculated WT and rga1-1 flowers were marked and the receptacle size measured. 191

While WT fruit remains unchanged when measured at stage 1 (0DPA) and stage 3 192

(7DPA), rga1-1 mutants develop significantly larger receptacle at stage 3 (7DPA) when 193

compared with stage 1 (0DPA) (Figure 2E, F). The data suggests that FveRGA1 encodes 194

a key negative regulator of fruit set, perhaps by binding and inhibiting the activity of 195

unknown TFs. 196

197

To investigate if the increased fruit size in arf8 or rga1-1 is due to increased cell division 198

or cell expansion, histological sections were conducted. Histological sections of wild type, 199

arf8-1 and rga1-1 fruits at 7DPA reveal an increased number of cell files across the width 200

of the pith in arf8 and rga1-1 receptacle (Figure S4); WT (YW5AF7) has an average of 201

28 cell files, arf8-1 has an average of 40 cell files, and rga1-1 has an average of 46.5 cell 202

files across the width of the pith (Figure S4J). In addition, arf8-1 and rga1-1 appear to 203

have dense cell files near the vasculatures (Figure S4F, I). Thus, increased cell division 204

might have contributed to the increased fruit width in both arf8 and rga1-1 mutants. 205

206

FveRGA1, a key crosstalk node and regulator of fruit set 207

To investigate the mechanism of how auxin and GA coordinate their effect on fruit set 208

and fruit growth and explain distinct effects of arf8-1 and rga1-1 on parthenocarpy, we 209

investigated if FveRGA1 interacts with FveARF8 and FveARF6. FveARF6 was included 210

as it was previously shown in Arabidopsis to interact with and inhibited by RGA1(Oh et 211

al., 2014). In BiFC, FveRGA1 interacts positively with FveARF6 and FveARF8 (Figure 212

3). In Y2H, full length FveRGA1 self-activates, thus a N-terminal 564 bp deletion was 213

made leading to FveRGA1-C (similar to the M5 version of AtRGA in Arabidopsis). 214

Consistent with the BiFC result, FveRGA1-C interacts positively with FveARF6 and 215

FveARF8 (Figure 3B), suggesting FveRGA1 as a key crosstalk node for hormone 216

signaling. 217


https://doi.org/10.1101/2020.03.30.016188

Zhou et al.,

10

218


https://doi.org/10.1101/2020.03.30.016188

Zhou et al.,

11

To determine which domain of FveARF8 interacts with FveRGA1, FveARF8 is divided 219

into three segments: DBD (DNA-binding), MR (Middle Region), and PB1 (Figure S5). 220

While each domain of FveARF8 fails to interact with FveRGA1-C, the combined DBD-221

MR domains of FveARF8 positively interact with FveRGA1-C (Figure 3C). The work 222

also indicates that the C-terminal portion of FveRGA1 is sufficient for the interaction 223

with these TFs. Since rga1-1 is a W601X mutation that deletes the last 10 amino acids 224

from the C-terminal SAW domain (Figure S1B), we tested if removing the SAW domain 225

in FveRGA1-C affect the interaction with FveARF8 or FveARF6 in Y2H. FveRGA1-C (-226

SAW) failed to interact with FveARF8 and FveARF6 (Figure 3D). rga1-1 may thus fail 227

to inhibit fruit set due to its inability to bind and inhibit FveARF6, ARF8, and perhaps 228

other unidentified transcription factors. 229

230

The emerging theme from above studies is that FveRGA1 acts to prevent fruit set 231

(transition from stage 1 to stage 2). Fertilization induced GA production may remove 232

FveRGA1 through GA-GID1-induced degradation, permitting fruit set to occur 233

accompanied by the activation of a number of TFs. 234

235

FveARF8 represses the expression of GA receptor FveGID1c 236

Since arf8-1 fruit could not develop in the absence of fertilization (Figure 2A), the larger 237

fruit of arf8-1 post-fertilization (Figure 1A) may reflect an increased sensitivity or 238

enhanced response to fertilization-induced hormonal signals, namely auxin and/or GA. 239

To test this hypothesis, emasculated flowers of arf8-1 were treated with 1mM NAA (1-240

naphthaleneacetic acid) or 1mM GA3. Mock treated fruits of the same genotype served as 241

the control. The ratio of fruit size between hormone treated vs. mock-treated fruit at 242

7DPA was compared; the ratio between treated and mock for arf8-1 mutants is 243

significantly higher than that of WT under GA treatment as well as under auxin (NAA) 244

treatment (Figure 4). Therefore, arf8-1 mutant may exhibit higher sensitivity or response 245

to GA and NAA than wild type, suggesting that one normal function of FveARF8 is to 246

dampen the receptacle’s sensitivity or response to GA and NAA perhaps in a feedback 247

regulatory loop. 248

249


https://doi.org/10.1101/2020.03.30.016188

Zhou et al.,

12

While it is understandable that arf8-1, a mutant gene in the auxin signaling pathway, 250


https://doi.org/10.1101/2020.03.30.016188

Zhou et al.,

13

affects sensitivity or response to auxin, why does arf8-1 also exhibit increased sensitivity 251

or response to GA? One possibility is that FveARF8 regulates the expression of GA 252

receptors to affect the sensitivity to GA. Three FveGID1 genes were identified from the F. 253

vesca genome (gene22353, gene20092, and gene27756) as the candidates of GA receptor 254

(Figure S6), only gene22353, named as either FveGID1a (Kang et al., 2013) or 255

FveGID1c (Li et al., 2019) and thereafter, is specifically expressed in the receptacle at 256

early fruit stages (Figure S6B-D). If FveGID1c encodes a GA receptor for fruit 257

development, FveGID1c should bind FveRGA1 when GA is present. Indeed, FveGID1c 258

directly interacts with FveRGA1 in yeast only when GA is added to the medium (Figure 259

4D) and this interaction requires the DELLA domain at the N-terminus as FveRGA1-C 260

(missing N-terminal 564 bp) fails to interact with FveGID1c even when GA is added 261

(Figure 4D). The interaction between FvGID1c and FveRGA1 is also demonstrated in 262

tobacco cells via BiFC (Figure 4E). CRISPR/Cas9 was used to knockout GID1c; a 263

biallelic mutant gid1c (-149; -166+15) was obtained (Figure S1C), which exhibited 264

strong dwarf but never flowered (Figure 4F). Although the GID1c sgRNA1 and sgRNA2 265

show 5 and 9 mismatches respectively to the other two GID1 genes (gene20092 and 266

gene27756) (Figure S6E), we still sequenced the other two GID1 genes (gene20092 and 267

gene27756) from the gid1c (-149; -166+15) plant and found no mutations in them. The 268

strong dwarf phenotype indicates that FveGID1c encodes an important GA receptor. 269

270

Next, we tested if FveGID1c expression is altered in arf8-1 at the early fruit stages 1-3. 271

FveGID1c expression level is similar in arf8-1 and WT at stage 1; at stages 2-3, 272

FveGID1c expression is reduced in WT but remains high in arf8-1 mutant receptacle, 273

significantly higher in arf8-1 when compared with WT (Figure 5), indicating that 274

FveARF8 may limit the sensitivity to GA by down-regulating FveGID1c at stages 2-3. 275

Given that FveARF8 is no longer inhibited by FveRGA1 at stages 2-3 due to fertilization-276

induced GA production and subsequent FveRGA degradation, FveARF8 may then be 277

free to repress FveGID1c expression at stage 2-3 to dampen sensitivity to GA in feedback 278

regulation. 279

280


https://doi.org/10.1101/2020.03.30.016188

Zhou et al.,

14

We then tested FveGID1c transcript level in rga1-1 mutant background. In rga1-1 mutant, 281


https://doi.org/10.1101/2020.03.30.016188

Zhou et al.,

15

a reduction of FveGID1c transcript is observed at stage 1 (Figure 5B), consistent with 282

that de-repressed FveARF8 in rga1-1 could act to repress FveGID1c transcription. Given 283

that FveARF8’s function may still be constrained by IAAs at stage 1, the repression of 284

FveGID1c is moderate (Figure 5B). At stage 2-3 (post-fertilization), freed from inhibition 285

by FveIAAs and FveRGA1, FveARF8 can repress FveGID1c expression in WT as well 286

as in rga1-1 (Figure 5B). The above qRT-PCR data are consistent with a negative 287

regulatory relationship of FveRGA1 to FveARF8 and to FveGID1c. 288

289

To test if FveARF8 directly represses FveGID1c, three segments of FveGID1c promoters, 290

P1, P2, and P3 (Figure S6F), were respectively fused to the AUR1-C reporter and 291

integrated into the yeast strain in a Y1H system. Self-activating activity of each segment 292

was examined at two Abi concentrations, 500 µM and 1000 µM (Figure 5C). Promoter 293

fragment 2 (P2), which does not self-activate and contains two potential AuxREs and two 294

potential G-boxes (Figure S6F), was used in the Y1H against transcription factor 295

FveARF6-AD or FveARF8-AD (Figure 5C). FveARF6-AD and FveARF8 (DBD)-AD 296

readily activate the FveGID1c (P2)::AUR1-C reporter, indicating direct binding of 297

FveARF8 and FveARF6 to the FveGID1c promoter. However, full length FveARF8 or a 298

truncated FveARF8 (DBD + MR) fail to activate the FveGID1c (P2) reporter (Figure 5C), 299

indicating that the MR domain of FveARF8 may possess strong repressor activity that 300

could counter Gal4-AD. 301

302

A transient expression reporter system in tobacco cells is also used to test direct effect of 303

FveARF8 and FveARF6 on the FveGID1c promoter. A 1kb promoter fragment of 304

FveGID1c was fused to the firefly luciferase (LUC) and integrated into vector PLAH-305

LARM (Taylor-Teeples et al., 2015), which also contains 35S::REN (Renilla luciferase) 306

as a control and 35S::FveARF8 or 35S::FveARF6 (Figure 5D). 1 µM GA was applied to 307

the tobacco cells to cause robust expression from pGID1c::LUC perhaps due to 308

endogenous transcription factors mediating GA responses. Compared with 35S::YFP 309

control; 35S::FveARF8 significantly reduces the expression from pGID1c::LUC (Figure 310

5D), suggesting that FveARF8 harbors strong repressor activity and binds directly to the 311

promoter of FveGID1c. FveARF6 appears to possess very weak repressor activity (Figure 312


https://doi.org/10.1101/2020.03.30.016188

Zhou et al.,

16

5D). Together, these data suggest that FveARF8, and perhaps FveARF6, may directly 313

bind and repress FveGID1c to limit sensitivity to GA at post-fertilization stage. 314

315

Consensus co-expression network identifies FveARF8, FveARF6, and FveIAA4 in 316

the same co-expression module 317

Our data revealed a previously unknown mechanism of GA-auxin crosstalk where auxin 318

represses the sensitivity to GA via activation of FveARF8. In auxin signaling pathway, 319

AUX/IAA proteins are known to bind and repress ARFs in the absence of auxin 320

(Guilfoyle, 2015; Salehin et al., 2015). Which FveIAAs could bind and inhibit FveARF8? 321

F. vesca genome V4.0a2 has about 34 FveARFs and 25 FveIAAs (Edger et al., 2018; Li et 322

al., 2019). To identify candidate FveIAAs, we mined the consensus co-expression 323

network generated previously (Shahan et al., 2018) (http://159.203.72.198:3838/fvesca/). 324

FveARF8 resides in module 80 of the Consensus_90 Fruit Network; this module 80 325

correlates positively in expression with early stage receptacle fruit tissues. Interestingly, 326

the two enriched GO terms for this module are “auxin activating signaling pathway” and 327

“cellular response to auxin stimulus”. FveARF6 and FveIAA4 are found in this module 328

with a high correlation co-efficient (>0.91) with FveARF8 (Figure 6; Data S1). Based on 329

RNA-seq data (Kang et al., 2013; Hollender et al., 2014; Hawkins et al., 2017), all three 330

genes are highly and specifically expressed in the receptacle, ovary wall and style during 331

early fruit development at stage 1-5 (Figure 6B). To validate their expression, qRT-PCR 332

was conducted using receptacle at fruit stages 1, 2, and 3; all three genes are expressed in 333

the receptacle at stage 1 (pre-fertilization) and show downward expression trend post-334

fertilization at stage 2 and 3 (Figure 6C). 335

336

To test if FveIAA4 in the same network module may interact and repress FveARF6/8 in 337

the absence of auxin, Y2H shows that FveIAA4 interacts with both FveARF8 and 338

FveARF6 (Figure 7). The C-terminal PB1 domain of FveARF8 is shown to be both 339

necessary and sufficient for the interaction with FveIAA4 (Figure 7B). BiFC confirms 340

these interactions (Figure 7C). The data suggest that fertilization-induced auxin may 341

cause degradation of FveIAA4, releasing FveARF8/6 to exert positive or negative 342

regulatory effects on downstream genes. 343


https://doi.org/10.1101/2020.03.30.016188

Zhou et al.,

17

344


https://doi.org/10.1101/2020.03.30.016188

Zhou et al.,

18

To test if FveARF6 and FveARF8 could work together as heterodimers, we first 345


https://doi.org/10.1101/2020.03.30.016188

Zhou et al.,

19

performed BiFC (Figure 7D) which showed that while FveARF6 and FveARF8 do not 346

interact with each other, FveARF6, but not FveARF8, interacts with itself (Figure 7D). 347

FveARF8’s inability to interact with FveARF6 is subsequently confirmed in co-348

immunoprecipitation assay (Figure 7E) using protein extracts from yeast containing 349

MYC-tagged Gal4BD-FveARF6 and HA-tagged Gal4AD-FveARF8 or Gal4AD-350

FveARF6. 351

352

The different MR sequence between FveARF8 and FveARF6 protein (Figure S5) 353

combined with their distinct ability to hormodimerize suggests that FveARF6 and 354

FveARF8 may not be functionally redundant. CRISPR construct was made for FveARF6 355

and transformed into WT F. vesca as well as arf8-1 (-2, -2) line# 156 that has segregated 356

away the original CRISPR-sgARF8 transgene. While we failed to obtain an arf6 single 357

mutant, arf6-1 (-1, -1) mutation was identified in arf8-1 (-2, -2) background in two 358

different transgenic lines (Figure S1D). These arf6-1; arf8-1 double mutant lines 359

developed fruits similar in size to WT and significantly smaller than arf8-1 single 360

mutants (Figure S7). The phenotype supports that FveARF6 and FveARF8 may play 361

different roles during fruit initiation and enlargement. 362

363

Discussion 364

Strawberry has long been a model to study fruit development, especially fruit set, when 365

Nitsch (1950) first observed the effect of auxin in stimulating receptacle fruit fleshy 366

formation when achenes were removed (Nitsch, 1950). However, the molecular 367

mechanisms underlying this elegant communication between seed and fruit flesh are not 368

well understood. Further, being the accessory fruit, where the fruit fleshy is derived from 369

receptacle instead of ovary wall, it poses additional questions concerning if the conserved 370

fertilization signals, auxin and GA, activate downstream fruit program in a similar or 371

distinct mechanism from botanical fruits. The work presented here provides basic 372

molecular frame work on how the GA and auxin signaling components interact and 373

coordinate fleshy fruit program in strawberry receptacle using both conserved and novel 374

mechanisms. The work laid the foundation for future engineering of parthenocarpic 375


https://doi.org/10.1101/2020.03.30.016188

Zhou et al.,

20

strawberry for enhancing fruit productivity when pollinators are scarce. 376

377

FveRGA1 is a key regulator of fruit set in diploid strawberry, but FveARF8 is 378

involved in downstream regulation of fruit development post-fertilization 379

rga1-1 is the only strawberry mutant so far reported to develop parthenocarpic fruit, 380

indicating FveRGA1 as a key regulator of fruit set in strawberry. FveRGA1 normally 381

prevents fruit sets by inhibition of a large number of TFs through protein-protein 382

interaction. In rga1-1 mutants, the inhibition is absent, and the TFs are free to activate 383

downstream genes for fruit development even in the absence of plant hormone, leading to 384

parthenocarpic fruit. This mutational effect of rga1-1 mimics the application of GA in 385

emasculated flowers, where GA-GID1c induced degradation of FveRGA1 lead to 386

parthenocarpic fruit (Kang et al., 2013). Our study identifies FveRGA1 as a target gene 387

for genetic manipulation to produce parthenocarpic strawberry. 388

389

The larger fruit size of arf8 mutants appears caused by different reasons from rga1-1. 390

Further, the effect of arf8 on receptacle size can be divided into pre- and post-fertilization 391

stages. At the pre-fertilization stage, the larger receptacle size is likely due to a defect on 392

flower development, which is also reflected by increased petal numbers in arf8 mutants 393

(Figure S2). It is important to note that the arf8-1 mutant receptacle size stays the same 394

from stage 1 to stage 3 when fertilization is prevented (Figure 2A-B). In contrast, at the 395

post-fertilization stages when auxin and GA are produced, the faster increase in fruit size 396

of arf8-1 mutants reflects an enhanced sensitivity to auxin and GA (Figure 4A-C). In 397

other words, same concentrations of phytohormones could mount a stronger response in 398

arf8-1 mutants. This is why arf8-1 mutants do not develop parthenocarpic fruit as GA 399

and auxin are still required for the arf8-1 mutants to develop fruit. 400

401

A model on the mechanism of GA-auxin crosstalk on fruit initiation 402

Figure 8 is a model that both summarizes previous understanding of GA and auxin 403

signaling pathways and incorporates our findings here. First, auxin and GA signaling 404

pathways act in parallel to promote fruit set and fruit development (thick green curved 405

arrows). At pre-fertilization (stage 1) receptacle (Figure 8B), RGA1 and IAA4, 406


https://doi.org/10.1101/2020.03.30.016188

Zhou et al.,

21

respectively bind and inhibits a large number of TFs including ARF8 (red bars), blocking 407


https://doi.org/10.1101/2020.03.30.016188

Zhou et al.,

22

fruit development. At post-fertilization (stage 2) receptacle (Figure 8C), GA and auxin 408

are produced from the fertilized seed and transported to the receptacle, where GA and 409

auxin interact with their respective receptors, FveGID1c and FveTIR, which subsequently 410

facilitate ubiquitin-mediated degradation of FveRGA1 and FveIAA4. Therefore, upon 411

fertilization, TFs in auxin and GA pathways are activated to promote fruit set and 412

subsequent fruit development (blue arrows). 413

414 We show that FveARF8’s DBD-MR and PB1 domains interact respectively with 415

FveRGA1 and FveIAA4 (Figure 3 and 7), indicating that FveARF8 activity could be 416

simultaneously inhibited by FveIAA4 and FveRGA1 prior to fertilization (Figure 8B). At 417

stage 2, when both auxin and GA are produced and FveIAA4 and FveRGA1 are both 418

degraded, FveARF8 and other TFs are now freed from both inhibitions to regulate the 419

transcription of downstream genes (blue lines) (Figure 8C). This explains why combined 420

treatment of auxin and GA is required to cause emasculated fruit to develop to the same 421

size as pollinated fruit (Kang et al., 2013). 422

423

One of the downstream targets of FveARF8 is the FveGID1c gene, and FveARF8 424

represses FveGID1c transcription to dampen sensitivity to GA in a feedback regulatory 425

loop (Figure 8C). This transcriptional down-regulation of FveGID1c by FveARF8 at 426

stage 2 is detected by qRT-PCR (Figure 5A) and appears to be direct (Figure 5C-D). As 427

ARFs normally have a large number of target genes in a genome (Oh et al., 2014), 428

FveARF8 may also promote fruit development indirectly through repressing other 429

negative regulators of fruit enlargement. This is indicated by a dotted blue arrow from 430

FveARF8 in Figure 8C. 431

432

FveGID1c (gene22353) encodes a GA receptor in F. vesca 433

Previous gene expression analysis (Kang et al., 2013), current Y2H assay in the presence 434

or absence of GA (Figure 4D), and CRISPR/CAS9-induced knockout of FveGID1c 435

(Figure 4F) all support that FveGID1c is an important GA receptor. While there are three 436

FveGID1 genes (Figure S6A), FveGID1c (gene22353) not only exhibits significantly 437

higher expression than the other two FveGID1b genes (gene20092 and gene27756) but 438


https://doi.org/10.1101/2020.03.30.016188

Zhou et al.,

23

also is more specifically expressed in the receptacle (Figure S6B-D). The severe 439

phenotype of gid1c (-149, -166+15) prevented the mutant plant to produce flowers. 440

Future work using inducible RNAi maybe necessary to allow the examination of gid1c 441

phenotype in fruit development. 442

443

Both conserved and novel GA-auxin crosstalk mechanisms exist in strawberry 444

While both auxin and GA act to promote fruit set in several organisms including 445

Arabidopsis, tomato, and strawberry, their interaction and crosstalk mechanism may vary. 446

In the botanical fruit of Arabidopsis and tomato, auxin was shown to act upstream of GA 447

by promoting GA biosynthesis (Serrani et al., 2008; Dorcey et al., 2009; Fuentes et al., 448

2012). Recent work in tomato revealed another crosstalk point through direct protein-449

protein interactions between SlARF7 and SlDELLA as well as SlIAA9 via SlARF7’s 450

MR2 and PB1 domains respectively (Hu et al., 2018). However, it is not known if 451

strawberry employs similar or novel crosstalk mechanisms to induce fruit formation 452

given the distinct floral tissue used by strawberry to form fruit flesh. 453

454

Our finding of direct interaction of FveIAA4 with FveARF8/FveARF6 and of RGA1 455

with FveARF8/FveARF6 illustrates conservation of crosstalk mechanism in a different 456

spatial context (ie. in accessary fruit). However, our data also reveal a previously 457

unknown crosstalk mechanism where FveARF8 represses FveGID1c transcription to 458

reduce sensitivity to GA in a feedback regulatory loop (Figure 8). The enhanced 459

sensitivity to auxin in arf8-1 (Figure 4A) also suggests that FveARF8 may act in a 460

feedback loop of auxin signaling. Proper regulation of FveARF8 protein activity through 461

FveIAA4 and FveRGA1 may serve to ensure that auxin and GA are limiting each other’s 462

signaling pathway to balance each other’s impact on fruit enlargement. One interesting 463

outcome of this balance is reflected in fruit shape. Previously, exogenous application of 464

GA to F. vesca fruit was shown to impact fruit shape differently from exogenous 465

application of NAA (Liao et al., 2018). Similarly, we showed that treatment of 466

emasculated WT or arf8-1 with NAA significantly increases fruit width and hence 467

rounder fruit shape, while treatment with GA significantly increases fruit length and 468


https://doi.org/10.1101/2020.03.30.016188

Zhou et al.,

24

hence longer fruit shape (Figure 4A). Hence, a balance between the two hormonal 469

pathways may impact fruit shape as well as size. 470

471

FveARF8 and FveARF6 exhibit differences in protein sequence, dimerization, 472

repression strength, and phenotype 473

Most ARF proteins can be divided into three domains, a N-terminal domain for DNA-474

binding and dimerization, a variable middle region (MR) that confers activation or 475

repression activity, and a C-terminal PB1 domain for interaction with ARFs as well as 476

Aux/IAA inhibitors (Tiwari et al., 2003; Boer et al., 2014; Guilfoyle, 2015; Powers et al., 477

2019). A recent report indicates that Arabidopsis activator ARF7 and ARF19 form 478

cytoplasmic assemblies as a mechanism to sequester them to modulate cellular 479

competence to respond to auxin (Powers et al., 2019). Sequence divergence in all three 480

domains contribute to distinct functional properties of ARFs within and between species. 481

482

Arabidopsis ARF6 and ARF8 genes are classified as activator ARFs that promote both 483

vegetative and reproductive development (Ulmasov et al., 1999; Tiwari et al., 2003). 484

Arabidopsis arf6; arf8 double mutants show more severe floral phenotypes than either 485

single mutants, suggesting partial redundancy between AtARF6 and AtARF8 (Nagpal et 486

al., 2005). In Arabidopsis and tomato, over-expression of miR167, which targets both 487

ARF6 and ARF8, also caused similar phenotypes as the arf6; arf8 double mutants (Wu et 488

al., 2006; Liu et al., 2014; Zheng et al., 2019). Most relevant to this work, Arabidopsis 489

arf8 mutants develop parthenocarpic fruit, and the expression of an “dominant negative” 490

AtARF8 causes parthenocarpy in Arabidopsis and tomato (Goetz et al., 2007). Thus, 491

ARF8 possesses a conserved role in preventing fruit set in Arabidopsis and tomato. 492

493

In this study, FveARF8 (gene31631) and FveARF6 (gene30394) are identified and named 494

based on their blast best hit with their Arabidopsis counterparts. A second F. vesca ARF6 495

genes (gene22728) is not studied here since it is not in the same co-expression module as 496

FveARF8. Sequence alignment among AtARF6/AtARF8/FveARF6/FveARF8 reveals 497

high levels of sequence conservation at the N-terminal DBD and C-terminal PB1 498

domains but poor sequence similarity at the MR domain (Figure S5). When compared 499


https://doi.org/10.1101/2020.03.30.016188

Zhou et al.,

25

with the MR of AtARF6, MR of FveARF6 has a large 40 amino acid deletion and a 500

reduced number of polyQ tracks (Figure S5), which may weaken the activation function 501

of FveARF6. Together, the poor sequence conservation in the MR domain, different 502

strength of repression on FveGID1c promoter, and different ability to homodimerize 503

suggest that FveARF6 and FveARF8 may function differently. This is supported by the 504

fruit phenotypes where a loss-of-FveARF6 led to fruit size reduction in arf8-1 505

background and a loss-of-FveARF8 caused fruit size increase (Figure S7). 506

507

Consensus co-expression network is a powerful tool 508

Previously, a consensus-clustering approach was applied to the standard WGCNA 509

(Weighted Gene Co-expression Network Analysis) to generate robust and reproducible 510

modules (Shahan et al., 2018). Since ARF and IAA genes are from large gene families, 511

the consensus co-expression network led us to a co-expression module consisting of 512

FveIAA4, FveARF6 and FveARF8 (Figure 6A). Our results support their protein-protein 513

interaction and functional roles in early fruit development and demonstrate the predictive 514

power of consensus co-expression network analysis. Such network is especially valuable 515

in guiding selection of genes belonging to large gene families as well as studies in non-516

model species where transformation is resource-intensive and challenging. 517

518

519


https://doi.org/10.1101/2020.03.30.016188

Zhou et al.,

26

Methods 520

Plant materials and growth condition 521

The Fragaria vesca cultivar Yellow Wonder 5AF7 (YW5AF7) was used in most of this 522

study. Cultivar Hawaii 4 (H4) was used in generating BZR1-RNAi lines. Plants were 523

grown in a growth chamber at a temperature of 25 °C/16 hour light followed by 22 °C/8 524

hour darkness with a relative humidity of 50%. The F. vesca genes studied here are 525

indicated by their gene IDs (genome version 2.0.a2; genome version 4.0.a2): FveARF8 526

(gene31631; FvH4_1g27650), FveARF6 (gene30394; FvH4_3g05010), FveIAA4 (gene 527

16569; FvH4_6g02870), FveGID1c (gene22353; FvH4_6g04960), and FveRGA1 (gene 528

06210; FvH4_4g34110). The names FveARF8, FveARF6, and FveIAA4 are assigned 529

based on their best blast hit with their Arabidopsis counterparts. 530

531

arf8 mutants were generated using CRISPR/CAS9 (Zhou et al., 2018). Four independent 532

arf8 transgenic lines #100 (-2, -2), #156 (-2, -2), #391-3 (-8, -23+5), and #391-31 (-2+1, -533

2+1) were detailed in Figure S1. Line #100 (-2, -2) and line #156 (-2, -2) are referred to 534

as arf8-1 in the text. T1 generation arf8 mutant plants were analyzed, and about 20% of 535

them have segregated away the CRISPR-sgARF8 transgene. 536

537

The rga1-1 (u230) mutant was previously generated by ENU chemical mutagenesis in a 538

ga20ox4 loss of function mutant background (Tenreira et al., 2017; Caruana et al., 2018). 539

This rga1-1; ga20ox4 was backcrossed with WT (H4) to yield rga1-1 single mutant, 540

which was used in all analyses. H4 was used as WT control whenever rga1-1 was used. 541

542

CRISPR/Cas9 genome editing and over-expression 543

To generate the arf8; arf6 double mutants, two seed gRNA sites (gRNA1: 544

TGCTTCAACCAACATGGAAG and gRNA2: GCCAGTGACACTAGTACCCA) 545

targeting an FveARF6 (gene30394) exon were inserted into the JH4 entry vector and then 546

incorporated into binary vector JH19 via gateway cloning (Zhou et al., 2018). Cotyledon 547

of arf8 mutant line #156 (-2, -2), which has segregated away its prior sgARF8-Cas9 548

construct, was used for transformation with the CRISPR-sgARF6 construct. Four 549

transgenic lines were generated. To detect mutation, genomic sequences spanning two 550


https://doi.org/10.1101/2020.03.30.016188

Zhou et al.,

27

target sites of FveARF6 were amplified by primers FvARF6-F5 and FvARF6-R5 (Data 551

S2) and analyzed by Sanger sequencing. 552

553

The FveGID1c CRISPR construct was created using JH4 and JH17 vectors (Zhou et al., 554

2018). Two seed guide RNA sequences (gRNA1: GCAACTGAGACACGCCTGG and 555

gRNA2: GTTCCCAAGGAGCCTTGTGG) targeting FveGID1c were inserted into the 556

entry vector JH4 and incorporated into binary vector JH17 via gateway cloning. In total, 557

3 transgenic lines were generated. The genomic region spanning the two target sites in 558

FveGID1c was amplified by primers FvGID1a-F1 and FvGID1a-R1 (Data S2) and 559

sequenced. Only one transgenic line, line #7 gid1c (-149, -166+15), has a clear bi-allelic 560

mutation (Figure 4F). 561

562

To test if the other two FveGiD1b genes (gene20092 and gene27756) were also mutated 563

in gid1c line #7, primer pairs gene20092-F&gene20092-R and gene27756-564

F&gene27756-R (Data S2) were used to respectively amplify the genome sequence of the 565

two FveGiD1b genes. The PCR products were then sequenced. 566

567

To generate the FveARF8 over-expression (OX) construct, full-length FveARF8 CDS 568

was PCR amplified from YW5AF7cDNA library with primers FvARF8-SalI-F and 569

FvARF8-Kpn-R (Data S2) and inserted into pENTR-2b vector via SalI/KpnI. The 570

pENTR-FveARF8 was incorporated into JH23 destination vector through gateway 571

cloning. In total, eight over-expression transgenic lines were generated and analyzed. 572

573

To construct the over-expression vector JH23 (Figure S6), the OCS terminator was 574

amplified from JH19 (Zhou et al., 2018) and inserted into pMDC99(Curtis and 575

Grossniklaus, 2003) at PacI and SacI sites; the AtUBQ10 promoter was amplified from 576

JH19 (Zhou et al., 2018) and then inserted into above PMDC99-OCS at HindIII and KpnI 577

sites to yield JH23. 578

579

Most above constructs were transformed in F. vesca (YW5AF7 or arf8-1 mutant in 580

YW5AF7 background) using previously published protocols (Kang et al., 2013; Caruana 581


https://doi.org/10.1101/2020.03.30.016188

Zhou et al.,

28

et al., 2018). 582

583

Receptacle size measurement 584

The receptacle of arf8 mutant and WT (YW5AF7 or Hawaii 4) were cut in half and 585

imaged under Stemi SV 6 dissecting microscope (Zeiss. Inc). The height and width were 586

then measured using the ZEN software. Flowers that just open with anthers not yet 587

dehiscent were considered stage 1 (0DPA). Flowers at 5-7 DPA were considered stage 3 588

usually with only 1 petal left. Receptacle measurements were normalized to petal size 589

(height and width of receptacle divided respectively by height and width of petal) to yield 590

relative receptacle size in Figure 1. To measure the receptacle size of rga1-1 (Figure 2) 591

the newly opened but non-dehiscent flowers were labeled as 1DPA (stage 1), the 592

receptacle length and width were measured at 9 DPA (stage 4) and 11-12 DPA (stage 5). 593

To determine parthenocarpy, arf8-1 or rga1-1 flowers were emasculated to prevent self-594

fertilization. Fruit size at 0DPA (pre-fertilization) was compared to fruit size at 7 DPA 595

(post-emasculation) within the same genotype. 596

597

For measuring the receptacle size of GA3- and NAA-treated fruits, WT (YW5AF7) or 598

arf8 flowers were emasculated at 0DPA under Stemi SV 6 dissecting microscope (Zeiss. 599

Inc), 1mM GA3 or 1mM NAA mixed with 0.1% Tween 20 was applied to the 600

emasculated receptacle. This treatment was repeated every two days until day 7 when the 601

receptacle size was measured. The relative GA- and NAA-treated receptacle height and 602

width were normalized to mock treated fruit of the same genotype. 603

604

For arf8-1, 8-19 fruits derived from 5 to 10 plants of the same genotype were measured; 605

and then combined to yield one bar in bar graph (or one box in box plot). For YW5AF7 606

(WT) control, fruit measurement was similarly made on fruits from 12 plants. 607

Measurement of other arf8 CRISPR lines were similarly conducted. For hormone 608

treatment experiments, the same groups of arf8-1 and YW5AF7 plants above were used 609

in fruit measurement. For rga1-1, twelve mutant plants were grown from the same 610

original mother plant through vegetative propagation via runner. Ten H4 (WT) control 611

plants were similarly propagated through an original H4 mother plant. The same 612


https://doi.org/10.1101/2020.03.30.016188

Zhou et al.,

29

measurement and analysis strategy described for arf8-1 was used. For arf8-1; arf6-1 613

double mutants, two transgenic lines were used, and each line has 3-4 plants derived from 614

the same calli. 10 YW5AF7 (WT) plants were grown together and used as controls. 615

616

Statistical analysis 617

Data from fruit size measurements, qRT-PCR, and LUC assay were all analyzed using 618

unpaired, unequal variance, two-tailed student's t test. Significant difference (P <0.01) 619

and difference (P <0.05) between treatments or genotypes are denoted by **and *, 620

respectively. Data are expressed as the means ± standard deviation for different 621

biological replicates. 622

623

BiFC and Y2H to test protein-protein interaction 624

For the BiFC, full-length CDS of FveARF8, FveARF6 were PCR amplified from cDNA 625

and ligated into PXY105 (cYFP) at BamHI and SalI sites, full-length CDS of FveRGA1 626

was ligated into PXY105 through Gibson cloning by using primers N-termless 627

Bamh1BIfc_FWP & Gib_Pxy105_06210_Xba1_R. Next, full-length CDS of FveARF8, 628

FveARF6, and FveIAA4, were amplified from cDNA and ligated into PXY106 (nYFP) at 629

BamHI and SalI sites. BiFC constructs were transformed into Agrobacterium EHA105. 630

Transformants were harvested when OD600 reached 1.0, and resuspended in 1/2 MS 631

medium (suppled with 50uM acetosyringone) to a final OD600 of 0.5. The agrobacterial 632

cells containing the cYFP fusion vector were mixed with agrobacterial cells containing 633

the nYFP fusion vector at 1:1 ratio. The mixture was infiltrated into young Nicotiana 634

benthamiana leave. After 48 hours, the YFP florescence signal was visualized and 635

photographed using Leica SP5X Confocal microscope (Leica Co. USA). 636

637

For the yeast two-hybrid (Y2H), full-length CDS of FveARF8, FveARF6, FveGID1c 638

were PCR amplified and ligated into GAL4 AD vector PGADT7 (Clontech Inc.) at the 639

BamHI/XhoI, BamHI/XhoI, EcoRI/XhoI double digestion sites, respectively. Full-length 640

CDS of FveIAA4 and FveGID1c were ligated into GAL4 BD vector PBGK T7 (Clontech 641

Inc.) through EcoRI/SalI double digestion sites. FveRGA1-C was cloned into the Y2H 642

vector pGBTK7 by Gibson assembly using primer pair GBTK7_M5_06210_F/ 643


https://doi.org/10.1101/2020.03.30.016188

Zhou et al.,

30

GBTK7_BamH1_06210_R (Data S2). To delete the SAW domain from FveRGA1-C, 644

primer pair RGA1-564-BglII-F & RGA1-564-SalI-R were used to amplify FveRGA1-C 645

(-SAW) fragment and inserted into pGBK T7 at BamHI-SalI sites. The auto activity of 646

each BD construct was examined by co-transformation with an empty AD vector PGAD 647

T7. 648

649

To make truncated FveARF8 for Y2H, FveARF8 (DBD), FveARF8 (MR), FveARF8 650

(PB1), and FveARF8 (MR+PB1) fragments were amplified respectively by primer pairs 651

ARF8-CDS-F1/ARF8-DBD-XhoI-R; FveARF8-MD-F/ARF8-MD-R; FveARF8-PBI-652

F/ARF8-CDS-R1, FveARF8-MD-F/ARF8-CDS-R1 (Data S2) and inserted into PGAD 653

T7 via EcoRI-XhoI/SalI sites. FveARF8 (DBD+MR) was amplified by ARF8-CDS-654

F1/ARF8-MD-R and inserted into PGAD T7 via BamHI/BglII-XhoI/SalI sites. 655

656

For Y2H essay, yeast constructs were transformed into PJ69-4A yeast strain using 657

LiAc/PEG method (Gietz and Schiestl, 2007). The transformed yeast cells were plated 658

onto -LW two drop-out or -HLWA four drop-out medium; -HLWA medium is often 659

supplemented with 5mM 3-AT to reduce background. Yeast cells were grown at 30 ℃ 660

for 3-4 days. 661

662

Each BiFC, Y1H, and Y2H experiment was performed three times. All repeats yielded 663

the same result. 664

665

Co-immunoprecipitation of proteins from yeast 666

The co-IP assay was performed following a published protocol (Gerace and Moazed, 667

2014) with minor modifications. The yeast strains containing PGBK T7-ARF6-Myc and 668

PGAD T7-ARF8-HA (or PGBK T7-ARF6-Myc and PGAD T7-ARF6-HA) were grown 669

to an OD600=1 at 28 0C. After centrifugation, the yeast cells were suspended in 1×lysis 670

buffer (100mM Na-HEPES, pH7.5, 400mM NaoAc, 2mM EDTA, 2mM EGTA, 10mM 671

MgOAc, and 10% Glycerol) and lysed by Mini-Beadeater (model: 2412PS-12W-B30), 672

Minebea Co., LTD. For each immunoprecipitation, 30μl Magnetic Dynabeads with 673

protein A (Invitrogen) were resuspend in 1×lysis buffer and incubated with 5μg c-Myc 674


https://doi.org/10.1101/2020.03.30.016188

Zhou et al.,

31

primary antibody (sc-40, Santa Cruz Biotechnology) for 10min at room temperature. The 675

protein A/c-Myc antibody slurry were dispensed into each cell lysate and rotate for 2h at 676

40C. Then, the Dynabeads- cMyc antibody-protein complex were washed 5 times with 677

wash buffer (100mM Na-HEPES, pH7.5, 400mM NaoAc, 2mM EDTA, 2mM EGTA, 678

10mM MgOAc, 0.25% NP-40, 3mM DTT, 1mMPMSF, and 10% Glycerol), followed by 679

elution with 4×SDS-PAGE Protein Loading Buffer (10% SDS, 500Mm DTT, 50% 680

Glycerol, 500mM Tris-HCL and 0.05% bromophenol blue dye) and loaded onto SDS-681

PAGE gel. 10% lysate was used as the “Input” loading. The western blots were 682

performed following General V3 Western Workflow Blotting Protocol (Bio-rad, 683

Laboratories, Inc). The anti-HA primary antibody (SKU: H6908, Sigma-Aldrich, Inc. 684

USA) was used to detect the pull-down products in western. The co-IP experiment was 685

repeated three times with the same result. 686

687

Yeast one-hybrid and the luciferase reporter assays 688

For yeast one hybrid assay, a 1077bp promoter fragment upstream of FveGiD1c 689

(gene22353) was amplified with primer FvGID1a-P-KpnI-F and FvGID1a-P-SalI-R 690

(TableS2) and ligated into pAbAi vector (Clontech Inc.) at KpnI and SalI sites. The 691

resulting GiD1c reporter vector was linearized by BstBI, and then introduced into 692

competent yeast cell Y1HGold to integrate at the ura3-52 locus. All procedures were 693

based on the protocol Matchmaker Gold Yeast One-Hybrid Library Screening System 694

User Manual (PT4087-1). The AD fusions in pGAD T7 with FveARF8 (and various 695

truncations) and FveARF6 are the same as those described for Y2H. 696

697

The pENTR vectors LEI02 and LEI04 (Figure S9) for LUC experiments are modified 698

from pLAH-R4R3-VP64Ter (Taylor-Teeples et al., 2015; Zhan et al., 2018). Specifically, 699

full length VP64-35S terminator was PCR-amplified from pLAH-R4R3-VP64Ter with 700

primers VP64-F and T35S-R (Data S2), and inserted into pCR™8/GW/TOPO™ vector 701

(Invitrogen™) via EcoRI, to generate LEI01 vector. The LEI01 vector was digested by 702

PacI and then self-ligated to drop off the VP64 sequence to produce LEI02. To construct 703

the negative control vector LEI04, YFP was amplified from pLAH-L1L4-Citrine and 704

then inserted into LEI02 vectors via EcoRI and AscI. Full length FveARF8 or FveARF6 705


https://doi.org/10.1101/2020.03.30.016188

Zhou et al.,

32

was PCR amplified and inserted into LEI02 at AscI and PacI to yield 35S::FveARF8 and 706

35S::FveARF6 respectively. The expression cassettes of 35S::FveARF8 (or 707

35S::FveARF6), pGiD1c::LUC, and 35S::REN were simultaneously incorporated into the 708

destination vector pLAH-LARm(Taylor-Teeples et al., 2015) through gateway cloning, 709

resulting in the final vectors 35S::FveARF8_pGiD1c::LUC_35S::REN and 710

35S::FveARF6_pGiD1c::LUC_35S::REN. The 35S::Citrine_pGiD1c::LUC_35S::REN 711

vector served as a control. 712

713

The luciferase reporter assay was performed following the instruction of Dual-714

Luciferase®Reporter Assay System (Promega, Inc, USA). Various constructs were 715

transformed into Agrobacterium strain EHA105 and infiltrated into tobacco Nicotiana 716

benthamiana leaves. 1µM GA was also applied during infiltration to increase the 717

autoactivity of GiD1c. After two days, a small leaf disc from infection site was collected 718

and finely ground in the passive lysis buffer (1×). 10 µl supernatant was mixed with LAR 719

solution (from the Promega kit), luciferase activities were measured using TD-20/20 720

Luminometer (Turner Designs, Inc, USA). The LUC readout was divided by the REN 721

readout to provide a normalized expression level. The LUC experiment was performed 722

three times with same results. The data shown is one of the three experiments. 723

724

RNA extraction and qRT-PCR 725

For testing FveGID1c expression (Figure 5), total RNA was extracted from YW5AF7, 726

arf8-1, Hawaii 4 (H4), and rga1-1 receptacles at stages indicated. Achenes were removed 727

from the receptacle, and 3 to 5 receptacles from several plants of the same genotype were 728

combined as one biological replicate. The CTAB RNA extraction method was 729

used(Gambino et al., 2008). The CTAB RNA extraction solution (2% CTAB, 2% PVPk-730

30, 10mM Tris-Hcl(pH8), 25mM EDTA, 2M Nacl, 0.5g/L Spermidine) was added to the 731

flesh tissues. The tissues were ground immediately following by the addition of 2/3 732

volume of 10M LiCl to precipitation RNA. In total, 1 ug total RNA was used for cDNA 733

synthesis with SuperScript IV VILO Master Mix (Invitrogen, USA). The cDNA product 734

was diluted 4 times and ~5 ul was used as a template in qRT-PCR using BioRad CFX96 735

Real-time system and SYBR Green PCR MasterMix (Applied Biosystems). The 736


https://doi.org/10.1101/2020.03.30.016188

Zhou et al.,

33

expression levels of four F.vesca genes FveARF8, FveARF6, FveIAA4, and FveGiD1c 737

were quantified via 4 primer pairs ARF8-F1-2/ARF8-R1-2, FveARF6-F1/FveARF6-R1, 738

FveIAA4-F1 /FveIAA4-XhoI-R and Fv GID1c-F2/ Fv GID1c-R2 (Data S2), respectively. 739

F. vesca gene03773 was used as the internal control with primers gene03773-740

F/gene03773-R (Caruana et al., 2018). The experiment was performed three times each 741

with RNA extracted from newly collected samples (ie. biological replicates). 742

743

Mining consensus co-expression network and other bioinformatics tools 744

The expression pattern of each F. vesca gene in wild type (YW5AF7) was visualized via 745

strawberry eFP browser (Hawkins et al., 2017) (http://mb3.towson.edu/efp/cgi-746

bin/efpWeb.cgi). The consensus co-expression network analysis of YW5AF7 floral and 747

fruit tissues(Shahan et al., 2018) was mined using the network website 748

(http://159.203.72.198:3838/fvesca/). The co-expression cluster/module was visualized 749

using Cytoscape (cytoscape.org). Protein alignments were performed using clustalw 750

(https://www.genome.jp/tools-bin/clustalw). Heatmap was generated with morpheus 751

(https://software.broadinstitute.org/morpheus/). 752

753

Tissue section and histological staining 754

Tissue section was performed following the protocol (Retamales and Scharaschkin, 2014) 755

with minor modification. YW5AF7, arf8-1, rga1-1 fruit (receptacle and achene) of each 756

stage was fixed in FAA (50% ethanol, 5% acetic acid, 3.7% Formaldehyde) for 3 days 757

and went through an ethanol series 50%, 60%, 70%, 80%, 90%, 95% (containing 0.1% 758

Eosin), and 100%. The tissues went through a series of changes in 25% xylenes/75% 759

ethanol, 50% xylenes/50% ethanol, 75% xylenes/25% ethanol, and finally 100% xylenes. 760

After two more changes of 100% xylene, paraffin chips were added into the xylene. 761

Afterwards, the tissue in xylene/paraffin mix is exchanged with 100% paraffin wax. After 762

4 changes of 100% paraffin in a duration of 4-5 days, a wax boat was poured. The tissues 763

were sectioned at 12 μm thickness on a microtome, and stained by Safranin-O/Fast Green 764

based on a protocol (http://microscopy.berkeley.edu/Resources/instruction/staining.htm). 765

Slides were mounted with Permount (Fisher) and dried overnight. 766

767


https://doi.org/10.1101/2020.03.30.016188

Zhou et al.,

34

Acknowledgements 768

This work has been supported by a grant from the National Science Foundation (IOS 769

1444987). We thank Dr. Ramin Yadegari at University of Arizona, Samuel Hazen at 770

UMass Amherst, Dr. Yanhai Yin from Iowa State University, and Dr. Kunsong Chen 771

from Zhejiang University for vectors. 772

773

Author Contributions 774

J. Z. and Z. L. designed experiments. J. Z., J. S., and L. G. performed experiments. X. H. 775

and A. P. assisted experiments. J. Z. and Z. L. analyzed and discussed the data, Z. L. and 776

J. Z. wrote the paper. 777

778

779


https://doi.org/10.1101/2020.03.30.016188

Zhou et al.,

35

Figure legend 780

781

Figure 1. arf8 mutants develop larger and rounder fruit 782

A. Images of bisected receptacle at stage 1 (0DPA, before fertilization), stage 3 (7DPA, 783

post-fertilization), stage 5 (15 DPA), and ripe stage in WT and arf8-1 (-2, -2). Note the 784

rounder fruit shape in arf8-1. Bars; 1 mm. 785

B. Average size of receptacle fruit (height and width, normalized to petal size) at stage 1 786

and stage 3. ** above bars represent significant difference (P < 0.01) between WT and 787

arf8-1 mutant. 788

C. Relative receptacle fruit size (height and width, normalized to petal size) in WT and 789

three additional arf8 mutant lines at stage 3 (7DPA). These arf8 mutant lines, #156 (-2, -790

2), #391-3 (-8, -23+5), and #391-31 (-2+1, -2+1) are illustrated in Figure S1. 791

D. WT and arf8 mutant receptacle fruits at ripe stage. Since petal has senesced, the 792

absolute fruit size was measured. arf8-1 mutants show a significant increase in width but 793

not in height at the ripe stage as shown in A. 794

795

796

Figure 2. rga1-1, but not arf8-1, develop parthenocarpic fruit 797

A. Photos of bisected receptacle of WT and arf8-1 at 0DPA and 7DPA. The flowers were 798

emasculated at the 0DPA. Bars; 1 mm 799

B. Quantitative measurement of receptacle size in emasculated fruit shown in A. No 800

significant difference in fruit size between 0DPA and 7DPA was observed in emasculated 801

WT and arf8 receptacle. 802

C. Images of bisected WT and rga1-1 receptacle at stage 1 (1DPA), Stage 4 (9DPA), and 803

stage 5 (11-12 DPA) without emasculation. rga1-1 receptacle fruit is larger and longer 804

than WT at stages 4-5. Bar: 1 mm. 805

D. Quantitative measurement of WT and rga1-1 receptacles at stage 1 and stage 4. rga1-1 806

develops significantly larger fruit, particularly in receptacle height. 807

E. Image of emasculated WT and rga1-1 receptacles at stage 1 (1DPA) and stage 3 808

(7DPA). Red arrows highlight the width of the receptacle. Fertilization-independent fruit 809

enlargement is evident in rga1-1 mutant. Bar, 1 mm. 810


https://doi.org/10.1101/2020.03.30.016188

Zhou et al.,

36

F. Quantitative measurements of emasculated WT and rga1-1 receptacles at stage 1 811

(0DPA) and stage 3 (7DPA). Emasculated WT receptacles at stage 1 and stage 3 are 812

similar in size. Emasculated rga1-1 receptacles are significantly larger and taller at stage 813

3 when compared with stage 1. 814

815

816

Figure 3. BIFC and Y2H revealed direct interaction of FveRGA1 with FveARF8, 817

and FveARF6 818

A. BiFC in tobacco leaf showing positive interactions between FveRGA1 and 819

respectively FveARF8 and FveARF6. 820

B. Y2H assay confirming the interactions between FveRGA1-C (fused to BD) and 821

respectively FveARF8 and FveARF6. Because of self-activating activity, a truncated 822

FveRGA1-C with its N-terminal DELLA domain removed is used in the Y2H. 823

C. Dissection of FveARF8 protein domains reveals the requirement of both DBD and MR 824

for the interaction with RGA1-C. 825

D. Removing the SAW domain from RGA1-C eliminates its ability to interact with 826

FveARF6 and FveARF8 in Y2H assay. 827

828

829

Figure 4. arf8-1 exhibits enhanced response to GA and auxin 830

A. Photos of bisected stage 3 receptacles of WT and arf8-1. The flowers were first 831

emasculated and then applied with NAA, GA3, or mock solution (see Methods) until 832

7DPA. Bars; 1 mm. 833

B. Relative receptacle size of WT and arf8-1 at stage 3 (7DPA) after GA treatment of 834

emasculated flowers. The height and width of hormone GA-treated receptacles are 835

divided respectively by the height and width of mock treated receptacles of the same 836

genotype. ** above bar represents significant difference (P < 0.01) between WT and 837

arf8-1. 838

C. Relative receptacle size of WT and arf8-1 at stage 3 (7DPA) after auxin (NAA) 839

treatment of emasculated flowers. Data is similarly measured and analyzed as B. 840


https://doi.org/10.1101/2020.03.30.016188

Zhou et al.,

37

D. Yeast two-hybrid assay showing the interaction between FveGID1c and FveRGA1 841

only in the presence of GA3 (200 M). The interaction is dependent on the DELLA 842

domain, which, when deleted in RGA1-C, could no longer support the interaction even in 843

the presence of GA. 844

E. BIFC assay in tobacco leaf cells confirming positive interaction between FveRGA1 845

and FveGID1c. The tobacco cells may produce GA to enable the interaction. 846

F. A CRISPR/CAS9-induced gid1c (-149, -166+15) biallelic mutant exhibiting severe 847

dwarf. 848

849

850

Figure 5. FveARF8 and FveARF6 directly represses the expression of FveGID1c 851

A. qRT-PCR showing relative expression of FveGID1c in WT (YW5AF7) and arf8-1 852

mutants at stage 1-3. 853

B. qRT-PCR showing relative expression of FveGID1c in WT (H4) and rga1-1 mutants 854

at stage 1-3. 855

C. Y1H assay showing activation of GID1c Promote fragment 2 (P2) when FveARF6 or 856

ARF8-DBD domain fused to GAL4 AD is introduced into the yeast. On the right, three 857

different promoter fragments (P1, P2, and P3) are tested for self-activation at two 858

different concentrations of antibiotics Abi500 and Abi1000. P2 does not have any self-859

activation activity. 860

D. Diagram of the LUC reporter system. The expression of LUC driven by the GID1c 861

promoter is repressed strongly by 35S::FveARF8 and weakly by 35S::FveARF6. Robust 862

self-activation in the absence of transacting factor (with YFP serving as a negative 863

control) is shown when 1 m GA3 is added in all essays. Y-axis shows relative LUC 864

expression normalized against 35S::REN expressed from the same vector. 865

866


https://doi.org/10.1101/2020.03.30.016188

Zhou et al.,

38

867

868

869

Figure 6. Network analysis identifies FveARF8, FveARF6 and FveIAA4 as potential 870

regulators of fruit initiation 871

A. Consensus Network analysis identifies FveARF8, FveARF6, and FveIAA4 in module 872

80 of Consensus 90 Fruit Network. The five-digit ID numbers for all other genes in the 873

same module are shown and also listed in Table S1. This module correlates with young 874

receptacle fruit. The consensus score and correlation score between genes are indicated 875

by color and thickness of lines respectively. 876

B. Heatmap of FveARF6, FveARF8 and FveIAA4 based on prior RNA-seq data, revealing 877

specific expression in young fruit and style. 878

C. qRT-PCR of FveARF8, FveARF6, and FveIAA4 in developing receptacle (pith and 879

cortex) at stages 1-3. Data represents mean ± SE of gene expression from three biological 880

replicates of YW receptacle tissues. The Y-axis indicates relative expression level 881

compared to reference gene PP2a (gene03773) derived using formula 2-ΔCt. Different 882

assaying techniques and tissue sampling methods may contribute to some differences 883

between qRT-PCR result and the RNA-seq data in (B). 884

885

886

Figure 7. Yeast two-hybrid, BiFC, and co-immunoprecipitation assays testing the 887

interaction between FveIAA4 and FveARF8/FveARF6 888

A. Yeast two-hybrid assay revealing the positive interaction between FveIAA4 and 889

FveARF6/FveARF8. 890

B. Domain dissection of FveARF8 showing that the PB1 domain is necessary and 891

sufficient for the interaction with FveIAA4. A diagram of ARF8 is drawn beneath, 892

showing DNA binding (DBD), Middle Region (MR), and PB1 domains. Numbers 893

beneath the diagram indicate amino acid residues. 894

C. FveARF8 and FveARF6 respectively interact with FveIAA4 shown by BiFC in 895

tobacco cells. 896


https://doi.org/10.1101/2020.03.30.016188

Zhou et al.,

39

D. BiFC in tobacco cells showing a lack of interaction between FveARF8 and FveARF6. 897

While FveARF8 does not homodimerize, FveARF6 does. 898

E. Co-immunoprecipitation using protein extracts from yeast cells containing c-Myc-899

tagged Gal4BD-FveARF6 and HA-tagged Gal4AD fusion constructs. INPUT lanes show 900

western blots of yeast protein extracts containing indicated constructs. Co-IP lanes are 901

western blots of immunoprecipitated proteins with anti-HA antibody. 902

903

904

Figure 8. A model that summarizes auxin/GA crosstalk in strawberry fruit initiation. 905

A. A diagram illustrating spatial separation of hormone synthesis in the achene from 906

hormonal signal perception and transduction in the receptacle. 907

B. In receptacle at stage 1 (pre-fertilization), when auxin and GA are absent, IAA4 and 908

RGA1 bind and repress ARF8 and other TFs in the auxin and GA signaling pathway. 909

Fruit set is blocked. 910

C. In the receptacle at stage 2 (post-fertilization), auxin and GA are transported to the 911

receptacle (from achenes) to induce signal transduction. Specifically, GA/GID1c and 912

auxin/TIR receptor complex caused degradation of FveRGA1 and FveIAA4, releasing 913

FveARF8 and TFs which promote fruit set and subsequent development. FveARF8 at the 914

same time represses the transcription of FveGID1c in a feedback regulation. 915

916

In B and C, the red lines describe negative regulation at post-translational level via 917

protein degradation or binding of repressors to transcription factors (TFs). Blue lines 918

describe positive (arrows) or negative (bar) regulation at transcription level. Light 919

shaded lines or names indicate “inactive” or “absent”. Dotted blue line from FveARF8 920

indicates FveARF8’s putative and indirect role in promoting fruit development. 921

922

923


https://doi.org/10.1101/2020.03.30.016188

Zhou et al.,

40

924


https://doi.org/10.1101/2020.03.30.016188

Parsed CitationsBoer, D.R., Freire-Rios, A., van den Berg, W.A., Saaki, T., Manfield, I.W., Kepinski, S., Lopez-Vidrieo, I., Franco-Zorrilla, J.M., de Vries,S.C., Solano, R., Weijers, D., and Coll, M. (2014). Structural basis for DNA binding specificity by the auxin-dependent ARF transcriptionfactors. Cell 156, 577-589.

Pubmed: Author and TitleGoogle Scholar: Author Only Title Only Author and Title

Caruana, J.C., Sittmann, J.W., Wang, W.P., and Liu, Z.C. (2018). Suppressor of Runnerless Encodes a DELLA Protein that ControlsRunner Formation for Asexual Reproduction in Strawberry. Molecular plant 11, 230-233.


Curtis, M.D., and Grossniklaus, U. (2003). A gateway cloning vector set for high-throughput functional analysis of genes in planta. PlantPhysiol 133, 462-469.


Daviere, J.M., and Achard, P. (2013). Gibberellin signaling in plants. Development 140, 1147-1151.Pubmed: Author and TitleGoogle Scholar: Author Only Title Only Author and Title

de Jong, M., Wolters-Arts, M., Feron, R., Mariani, C., and Vriezen, W.H. (2009). The Solanum lycopersicum auxin response factor 7(SlARF7) regulates auxin signaling during tomato fruit set and development. Plant Journal 57, 160-170.


Dorcey, E., Urbez, C., Blazquez, M.A., Carbonell, J., and Perez-Amador, M.A. (2009). Fertilization-dependent auxin response in ovulestriggers fruit development through the modulation of gibberellin metabolism in Arabidopsis. The Plant journal : for cell and molecularbiology 58, 318-332.


Edger, P.P., VanBuren, R., Colle, M., Poorten, T.J., Wai, C.M., Niederhuth, C.E., Alger, E.I., Ou, S., Acharya, C.B., Wang, J., Callow, P.,McKain, M.R., Shi, J., Collier, C., Xiong, Z., Mower, J.P., Slovin, J.P., Hytonen, T., Jiang, N., Childs, K.L., and Knapp, S.J. (2018). Single-molecule sequencing and optical mapping yields an improved genome of woodland strawberry (Fragaria vesca) with chromosome-scale contiguity. GigaScience 7, 1-7.


Edger, P.P., Poorten, T.J., VanBuren, R., Hardigan, M.A., Colle, M., McKain, M.R., Smith, R.D., Teresi, S.J., Nelson, A.D.L., Wai, C.M.,Alger, E.I., Bird, K.A., Yocca, A.E., Pumplin, N., Ou, S., Ben-Zvi, G., Brodt, A., Baruch, K., Swale, T., Shiue, L., Acharya, C.B., Cole, G.S.,Mower, J.P., Childs, K.L., Jiang, N., Lyons, E., Freeling, M., Puzey, J.R., and Knapp, S.J. (2019). Origin and evolution of the octoploidstrawberry genome. Nat Genet 51, 541-547.


Feng, J., Dai, C., Luo, H., Han, Y., Liu, Z., and Kang, C. (2019). Reporter gene expression reveals precise auxin synthesis sites duringfruit and root development in wild strawberry. Journal of experimental botany 70, 563-574.


Fuentes, S., Ljung, K., Sorefan, K., Alvey, E., Harberd, N.P., and Ostergaard, L. (2012). Fruit growth in Arabidopsis occurs via DELLA-dependent and DELLA-independent gibberellin responses. Plant Cell 24, 3982-3996.


Gambino, G., Perrone, I., and Gribaudo, I. (2008). A Rapid and effective method for RNA extraction from different tissues of grapevineand other woody plants. Phytochemical analysis : PCA 19, 520-525.


Gerace, E., and Moazed, D. (2014). Coimmunoprecipitation of proteins from yeast. Methods Enzymol 541, 13-26.Pubmed: Author and TitleGoogle Scholar: Author Only Title Only Author and Title

Gietz, R.D., and Schiestl, R.H. (2007). High-efficiency yeast transformation using the LiAc/SS carrier DNA/PEG method. Nat Protoc 2, 31-34.


Gillaspy, G., Ben-David, H., and Gruissem, W. (1993). Fruits: A Developmental Perspective. Plant Cell 5, 1439-1451.Pubmed: Author and TitleGoogle Scholar: Author Only Title Only Author and Title


https://doi.org/10.1101/2020.03.30.016188

Google Scholar: Author Only Title Only Author and Title

Goetz, M., Vivian-Smith, A., Johnson, S.D., and Koltunow, A.M. (2006). AUXIN RESPONSE FACTOR8 is a negative regulator of fruitinitiation in Arabidopsis. Plant Cell 18, 1873-1886.


Goetz, M., Hooper, L.C., Johnson, S.D., Rodrigues, J.C., Vivian-Smith, A., and Koltunow, A.M. (2007). Expression of aberrant forms ofAUXIN RESPONSE FACTOR8 stimulates parthenocarpy in Arabidopsis and tomato. Plant Physiol 145, 351-366.


Guilfoyle, T.J. (2015). The PB1 domain in auxin response factor and Aux/IAA proteins: a versatile protein interaction module in theauxin response. Plant Cell 27, 33-43.


Hartweck, L.M. (2008). Gibberellin signaling. Planta 229, 1-13.Pubmed: Author and TitleGoogle Scholar: Author Only Title Only Author and Title

Hawkins, C., Caruana, J., Li, J., Zawora, C., Darwish, O., Wu, J., Alkharouf, N., and Liu, Z. (2017). An eFP browser for visualizingstrawberry fruit and flower transcriptomes. Horticulture research 4, 17029.


Hollender, C.A., Geretz, A.C., Slovin, J.P., and Liu, Z. (2012). Flower and early fruit development in a diploid strawberry, Fragaria vesca.Planta 235, 1123-1139.


Hollender, C.A., Kang, C., Darwish, O., Geretz, A., Matthews, B.F., Slovin, J., Alkharouf, N., and Liu, Z. (2014). Floral Transcriptomes inWoodland Strawberry Uncover Developing Receptacle and Anther Gene Networks. Plant Physiol 165, 1062-1075.


Hu, J.H., Israeli, A., Ori, N., and Sun, T.P. (2018). The Interaction between DELLA and ARF/IAA Mediates Crosstalk between Gibberellinand Auxin Signaling to Control Fruit Initiation in Tomato. Plant Cell 30, 1710-1728.


Kang, C.Y., Darwish, O., Geretz, A., Shahan, R., Alkharouf, N., and Liu, Z.C. (2013). Genome-Scale Transcriptomic Insights into Early-Stage Fruit Development in Woodland Strawberry Fragaria vesca. Plant Cell 25, 1960-1978.


Leyser, O. (2018). Auxin Signaling. Plant Physiol 176, 465-479.Pubmed: Author and TitleGoogle Scholar: Author Only Title Only Author and Title

Li, Y., Pi, M., Gao, Q., Liu, Z., and Kang, C. (2019). Updated annotation of the wild strawberry Fragaria vesca V4 genome. Horticultureresearch 6, 61.


Liao, X., Li, M., Liu, B., Yan, M., Yu, X., Zi, H., Liu, R., and Yamamuro, C. (2018). Interlinked regulatory loops of ABA catabolism andbiosynthesis coordinate fruit growth and ripening in woodland strawberry. Proc Natl Acad Sci U S A 115, E11542-E11550.


Liu, N., Wu, S., Van Houten, J., Wang, Y., Ding, B., Fei, Z., Clarke, T.H., Reed, J.W., and van der Knaap, E. (2014). Down-regulation ofAUXIN RESPONSE FACTORS 6 and 8 by microRNA 167 leads to floral development defects and female sterility in tomato. Journal ofexperimental botany 65, 2507-2520.


Martin-Pizarro, C., Trivino, J.C., and Pose, D. (2019). Functional analysis of the TM6 MADS-box gene in the octoploid strawberry byCRISPR/Cas9-directed mutagenesis. Journal of experimental botany 70, 885-895.


Nagpal, P., Ellis, C.M., Weber, H., Ploense, S.E., Barkawi, L.S., Guilfoyle, T.J., Hagen, G., Alonso, J.M., Cohen, J.D., Farmer, E.E., Ecker,J.R., and Reed, J.W. (2005). Auxin response factors ARF6 and ARF8 promote jasmonic acid production and flower maturation.Development 132, 4107-4118.



https://doi.org/10.1101/2020.03.30.016188

Nitsch, J. (1950). Growth and morphogenesis of the strawberry as related to auxin. Am J Bot 37, 211-215.Pubmed: Author and TitleGoogle Scholar: Author Only Title Only Author and Title

Oh, E., Zhu, J.Y., Bai, M.Y., Arenhart, R.A., Sun, Y., and Wang, Z.Y. (2014). Cell elongation is regulated through a central circuit ofinteracting transcription factors in the Arabidopsis hypocotyl. eLife 3.


Ozga, J.A., Reinecke, D.M., Ayele, B.T., Ngo, P., Nadeau, C., and Wickramarathna, A.D. (2009). Developmental and hormonal regulationof gibberellin biosynthesis and catabolism in pea fruit. Plant Physiol 150, 448-462.


Powers, S.K., Holehouse, A.S., Korasick, D.A., Schreiber, K.H., Clark, N.M., Jing, H., Emenecker, R., Han, S., Tycksen, E., Hwang, I.,Sozzani, R., Jez, J.M., Pappu, R.V., and Strader, L.C. (2019). Nucleo-cytoplasmic Partitioning of ARF Proteins Controls AuxinResponses in Arabidopsis thaliana. Molecular cell 76, 177-190 e175.


Retamales, H.A., and Scharaschkin, T. (2014). A staining protocol for identifying secondary compounds in Myrtaceae. Applications inplant sciences 2.


Salehin, M., Bagchi, R., and Estelle, M. (2015). SCFTIR1/AFB-based auxin perception: mechanism and role in plant growth anddevelopment. Plant Cell 27, 9-19.


Sanchez-Sevilla, J.F., Vallarino, J.G., Osorio, S., Bombarely, A., Pose, D., Merchante, C., Botella, M.A., Amaya, I., and Valpuesta, V.(2017). Gene expression atlas of fruit ripening and transcriptome assembly from RNA-seq data in octoploid strawberry (Fragaria xananassa). Scientific reports 7, 13737.


Serrani, J.C., Ruiz-Rivero, O., Fos, M., and Garcia-Martinez, J.L. (2008). Auxin-induced fruit-set in tomato is mediated in part bygibberellins. Plant Journal 56, 922-934.


Shahan, R., Zawora, C., Wight, H., Sittmann, J., Wang, W., Mount, S.M., and Liu, Z. (2018). Consensus Coexpression Network AnalysisIdentifies Key Regulators of Flower and Fruit Development in Wild Strawberry. Plant Physiol 178, 202-216.


Shulaev, V., Sargent, D.J., Crowhurst, R.N., Mockler, T.C., Folkerts, O., Delcher, A.L., Jaiswal, P., Mockaitis, K., Liston, A., Mane, S.P.,Burns, P., Davis, T.M., Slovin, J.P., Bassil, N., Hellens, R.P., Evans, C., Harkins, T., Kodira, C., Desany, B., Crasta, O.R., Jensen, R.V.,Allan, A.C., Michael, T.P., Setubal, J.C., Celton, J.M., Rees, D.J., Williams, K.P., Holt, S.H., Ruiz Rojas, J.J., Chatterjee, M., Liu, B., Silva,H., Meisel, L., Adato, A., Filichkin, S.A., Troggio, M., Viola, R., Ashman, T.L., Wang, H., Dharmawardhana, P., Elser, J., Raja, R., Priest,H.D., Bryant, D.W., Jr., Fox, S.E., Givan, S.A., Wilhelm, L.J., Naithani, S., Christoffels, A., Salama, D.Y., Carter, J., Lopez Girona, E.,Zdepski, A., Wang, W., Kerstetter, R.A., Schwab, W., Korban, S.S., Davik, J., Monfort, A., Denoyes-Rothan, B., Arus, P., Mittler, R., Flinn,B., Aharoni, A., Bennetzen, J.L., Salzberg, S.L., Dickerman, A.W., Velasco, R., Borodovsky, M., Veilleux, R.E., and Folta, K.M. (2011). Thegenome of woodland strawberry (Fragaria vesca). Nat Genet 43, 109-116.


Sun, T.P. (2011). The molecular mechanism and evolution of the GA-GID1-DELLA signaling module in plants. Current biology : CB 21,R338-345.


Taylor-Teeples, M., Lin, L., de Lucas, M., Turco, G., Toal, T.W., Gaudinier, A., Young, N.F., Trabucco, G.M., Veling, M.T., Lamothe, R.,Handakumbura, P.P., Xiong, G., Wang, C., Corwin, J., Tsoukalas, A., Zhang, L., Ware, D., Pauly, M., Kliebenstein, D.J., Dehesh, K.,Tagkopoulos, I., Breton, G., Pruneda-Paz, J.L., Ahnert, S.E., Kay, S.A., Hazen, S.P., and Brady, S.M. (2015). An Arabidopsis generegulatory network for secondary cell wall synthesis. Nature 517, 571-575.


Tenreira, T., Lange, M.J.P., Lange, T., Bres, C., Labadie, M., Monfort, A., Hernould, M., Rothan, C., and Denoyes, B. (2017). A SpecificGibberellin 20-Oxidase Dictates the Flowering-Runnering Decision in Diploid Strawberry. Plant Cell 29, 2168-2182.



https://doi.org/10.1101/2020.03.30.016188

Tiwari, S.B., Hagen, G., and Guilfoyle, T. (2003). The roles of auxin response factor domains in auxin-responsive transcription. PlantCell 15, 533-543.


Ulmasov, T., Hagen, G., and Guilfoyle, T.J. (1999). Activation and repression of transcription by auxin-response factors. Proc Natl AcadSci U S A 96, 5844-5849.


Wang, H., Jones, B., Li, Z.G., Frasse, P., Delalande, C., Regad, F., Chaabouni, S., Latche, A., Pech, J.C., and Bouzayen, M. (2005). Thetomato Aux/IAA transcription factor IAA9 is involved in fruit development and leaf morphogenesis. Plant Cell 17, 2676-2692.


Wang, H., Schauer, N., Usadel, B., Frasse, P., Zouine, M., Hernould, M., Latche, A., Pech, J.C., Fernie, A.R., and Bouzayen, M. (2009).Regulatory features underlying pollination-dependent and -independent tomato fruit set revealed by transcript and primary metaboliteprofiling. Plant Cell 21, 1428-1452.


Wu, M.F., Tian, Q., and Reed, J.W. (2006). Arabidopsis microRNA167 controls patterns of ARF6 and ARF8 expression, and regulatesboth female and male reproduction. Development 133, 4211-4218.


Zhan, J., Li, G., Ryu, C.H., Ma, C., Zhang, S., Lloyd, A., Hunter, B.G., Larkins, B.A., Drews, G.N., Wang, X., and Yadegari, R. (2018).Opaque-2 Regulates a Complex Gene Network Associated with Cell Differentiation and Storage Functions of Maize Endosperm. PlantCell 30, 2425-2446.


Zheng, L.J., Nagpal, P., Villarino, G., Trinidad, B., Bird, L., Huang, Y.B., and Reed, J.W. (2019). miR167 limits anther growth to potentiateanther dehiscence. Development 146.


Zhou, J., Wang, G., and Liu, Z. (2018). Efficient genome editing of wild strawberry genes, vector development and validation. PlantBiotechnol J 16, 1868-1877.



https://doi.org/10.1101/2020.03.30.016188

zhou et al., 1 research article · 30/03/2020 · 102 tissues, embryo, ghost (seed coat +...

Documents