1

RESEARCH ARTICLE 1

2

Variation in expression of the HECT E3 ligase UPL3 modulates LEC2 3

levels, seed size and crop yields in Brassica napus 4

5

Charlotte Miller1, Rachel Wells1, Neil McKenzie1, Martin Trick1, Joshua Ball1, 6

Abdelhak Fatihi2, Bertrand Dubreucq2, Thierry Chardot2, Loic Lepiniec2, Michael 7

W Bevan1* 8 1 Department of Cell and Developmental Biology, John Innes Centre, Norwich UK 9 2 Institut Jean-Pierre Bourgin, INRA, AgroParisTech, CNRS, Université Paris-Saclay, 10

INRA Versailles, route de Saint-Cyr, FR 11

12

*Corresponding author13

Michael W. Bevan. Orcid ID 0000-0001-8264-235414

Cell and Developmental Biology Dept15

John Innes Centre16

Norwich Research Park17

Norwich NR4 7UH, UK18

Tel +44 1603 45052019

michael.bevan@jic.ac.uk20

21

Short title: UPL3 modulates LEC2 levels and seed yield in Brassica napus. 22

One-sentence summary: 23

24

The author responsible for distribution of materials integral to the findings presented in 25

this article in accordance with the policy described in the Instructions for Authors 26

(www.plantcell.org) is: Michael W. Bevan (michael.bevan@jic.ac.uk). 27

28

ABSTRACT 29

Identifying genetic variation that increases crop yields is a primary objective in plant 30

breeding. We used association analyses of Brassica napus (oilseed rape/canola) 31

accessions to identify genetic variation that influences seed size, lipid content and final 32

crop yield. Variation in the promoter region of the HECT E3 ligase gene BnaUPL3.C03 33

made a major contribution to variation in seed weight per pod, with accessions 34

exhibiting high seed weight per pod having lower levels of BnaUPL3.C03 expression. 35

We defined a mechanism in which UPL3 mediated the proteasomal degradation of 36

LEC2, a master transcriptional regulator of seed maturation. Accessions with reduced 37

UPL3 expression had increased LEC2 protein levels, larger seeds and prolonged 38

expression of lipid biosynthetic genes during seed maturation. Natural variation in 39

BnaUPL3.C03 expression appears not to have been exploited in current Brassica napus 40

breeding lines and could therefore be used as a new approach to maximize future yields 41

in this important oil crop. 42

43

Plant Cell Advance Publication. Published on August 22, 2019, doi:10.1105/tpc.18.00577

©2019 American Society of Plant Biologists. All Rights Reserved

2

44

INTRODUCTION 45

In major oil-producing crops, such as Brassica napus (oilseed rape), the composition of 46

seed storage lipids has been optimized for different end uses, from human nutrition to 47

industrial applications, by identifying allelic variation in biosynthetic enzymes and 48

pathways (Napier and Graham, 2010). For example, elite oilseed rape varieties now 49

have greatly enhanced nutritional value, with high linoleic acid, reduced erucic acid and 50

optimal linoleic to linolenic acid ratios. However, increasing overall production of storage 51

lipids to meet projected future demands for both food and industrial uses remains a key 52

objective for achieving food security and sustainable industrial production. 53

An interacting network of transcription factors establishes and maintains Arabidopsis 54

embryo development and promotes the accumulation of seed storage lipids and 55

proteins (Fatihi et al., 2016; Boulard et al., 2017). Loss-of-function mutations in four 56

conserved regulatory genes lead to curtailed seed maturation, loss of dormancy, and 57

ectopic vegetative development. FUSCA3 (FUS3), ABSCISIC ACID INSENSITIVE 3 58

(ABI3) and LEAFY COTYLEDON 2 (LEC2) encode AFL-B3-family transcription factors, 59

while LEAFY COTYLEDON 1 (LEC1) encodes an NFY-Y CCAAT binding transcription 60

factor. These “LAFL” genes (Santos-Mendoza et al., 2008a; Roscoe et al., 2015) induce 61

seed maturation and inhibit germination, and their expression is down-regulated at the 62

initiation of seed dormancy and desiccation tolerance. LEC2 and FUS3 expression is 63

repressed by miRNA-mediated mechanisms during early embryogenesis (Nodine and 64

Bartel, 2010; Willmann et al., 2011) to ensure the correct timing of storage reserve 65

accumulation. At later stages of seed development, the B3-domain protein VAL3 66

recruits HISTONE DEACETYLASE19 (HDAC19) to the promoters of LAFL genes and 67

silences their expression by altering levels of histone methylation and acetylation (Zhou 68

et al., 2013). The stability of LAFL proteins is also controlled during seed maturation and 69

dormancy. ABI3-INTERACTING PROTEIN 2 (AIP2) is an E3 ligase that ubiquitylates 70

both ABI3 (Zhang et al., 2005) and FUS3 (Duong et al., 2017), suggesting that 71

regulation of LAFL protein levels has an important role in seed maturation and the 72

transition to dormancy. The SNF kinase AKIN10 phosphorylates and stabilizes FUS3 73

3

(Chan et al., 2017; Tsai and Gazzarrini, 2012) and WRI1 (Zhai et al., 2017), a 74

transcription factor regulated by LEC2 that promotes expression of glycolytic and lipid 75

biosynthesis genes. Improved understanding of the control of these important seed 76

maturation regulators may provide new ways to optimize seed composition and yield. 77

An important strategy in crop improvement aims to identify new sources of genetic 78

variation from diverse germplasm resources for increasing crop productivity (Bevan et 79

al., 2017). Genome-Wide Association Studies (GWAS) are increasingly used for 80

identifying variation associated with traits in crops and their wild relatives. For example, 81

associations between sequence variation, gene expression levels and oil composition 82

have been used to identify genetic variation in known genes conferring oil quality traits 83

in oilseed rape (Harper et al., 2012; Lu et al., 2017). GWAS also has potential for the 84

discovery of new gene functions and, when utilized fully, can lead to a deeper 85

understanding of mechanisms underlying complex traits such as crop yield. 86

Here we use Associative Transcriptomics in Brassica napus to identify genetic variation 87

in the regulation of BnaUPL3.C03, encoding an ortholog of the HECT E3 ubiquitin ligase 88

UPL3 (Downes et al., 2003; El Refy et al., 2003), that is associated with increased seed 89

size and field yields. We establish a mechanism in which reduced expression of 90

BnaUPL3.C03 maintains higher levels of LEC2 protein during seed maturation by 91

reduced UPL3-mediated LEC2 ubiquitylation, leading to increased seed lipid levels and 92

overall increased seed yields. Analysis of elite oilseed rape varieties shows that 93

variation in the expression of BnaUPL3.C03 has not yet been exploited in breeding 94

programmes and thus can be used to increase crop yields. 95

RESULTS 96

Associative Transcriptomics identifies a negative correlation between 97

BnaUPL3.C03 expression and seed weight per pod in B. napus 98

A panel of 94 B. napus oilseed rape accessions with high genetic diversity 99

(Supplemental Table 1), for which leaf transcriptome data from each accession has 100

been mapped to a sequenced reference genome (Harper et al., 2012), was screened 101

for yield-related phenotypic variation. High levels of trait variation were observed. Trait 102

4

associations with both sequence variation, in the form of hemi-SNPs in the polyploid 103

genome of B. napus, and gene expression levels, assessed as Reads Per Kilobase of 104

unigene per Million (RPKM) mapped reads in a Gene Expression Marker (GEM) 105

analysis, were then calculated. This identified significant associations between variation 106

in seed weight per pod (SWPP) and SNP variation in homoeologous regions of linkage 107

groups A08 and C03 (Figure 1A and Supplemental Figure 1). SWPP phenotype data 108

displayed a normal distribution appropriate for Mixed Linear Model analyses (Figure 109

1B). 110

Assessment of phenotypic variation segregating with alleles for the most significant 111

SNP marker, JCVI_5587:125, revealed that accessions inheriting “T” at this locus had 112

SWPP values ~20% lower than those accessions inheriting the hemi-SNP genotype, “Y” 113

(C/T) (Figure 1C). Figure 1A shows that variation in the hemi-SNP marker did not 114

distinguish between the A08 and C03 chromosomes. However, GEM analyses showed 115

an association of SWPP with varying expression of a single C genome-assigned 116

unigene within this region, C_EX097784. (Figure 1D and Supplemental Figure 1). This 117

unigene corresponded to an ortholog of the Arabidopsis UBIQUITIN PROTEIN LIGASE 118

3 (UPL3) encoding a HECT E3 ligase (Downes et al., 2003; El Refy et al., 2003). 119

Two UPL3 homologs in B. napus, BnaUPL3.C03 and BnaUPL3.A08, were identified 120

based on protein sequence similarity and conserved synteny between Arabidopsis, B. 121

rapa and B. oleracea. Supplemental Figure 2 illustrates the high protein sequence 122

similarity between the single Arabidopsis and the two B. napus UPL3 orthologs. 123

Associative Transcriptomics analyses identified significant differential expression of the 124

BnaUPL3.C03 gene between GWAS accessions displaying high variation in SWPP 125

(Figure 1D and Supplemental Figure 1). Gene-specific RT-qPCR analyses confirmed 126

this differential expression at the BnaUPL3.C03 locus in seedlings of six lines selected 127

for high- or low- SWPP (Figure 1E). Correlating BnaUPL3.C03 expression levels with 128

SWPP revealed a negative relationship, with accessions displaying low BnaUPL3.C03 129

expression exhibiting high SWPP (Figure 1F). RT-qPCR analyses of BnaUPL3.A08 130

expression in developing seeds of six lines with high- or low- SWPP showed no 131

significant difference in expression levels and variation in SWPP (Supplemental Figure 132

5

3), whereas BnaUPL3.C03 expression was lower in developing seeds of these high 133

SWPP lines compared to low SWPP lines (Figure 2B and see below). This is consistent 134

with the absence of an association between transcript levels at the BnaUPL3.A08 locus 135

and SWPP determined by Associative Transcriptomics analyses (Supplemental Figure 136

1). This indicated that variation in the expression of BnaUPL3.C03 contributed to 137

variation in SWPP. 138

A subset of 10 GWAS accessions (Supplemental Table 2) exhibiting both differential 139

expression of BnaUPL3.C03 and high variation in SWPP were grown under field 140

conditions in a replicated yield trial. Mean plot yields across accessions showed a 141

significant increase in plot yields of high SWPP accessions (Figure 1G), indicating 142

SWPP is an important measure of seed yield under field conditions. 143

Previous studies in Arabidopsis, a close relative of B. napus, identified roles for UPL3 in 144

mediating the proteasomal degradation of GLABROUS 3 (GL3) and ENHANCER OF 145

GLABROUS 3 (EGL3), both known regulators of trichome morphogenesis. Enhanced 146

GL3/EGL3 protein levels in an Arabidopsis loss-of-function upl3 mutant altered leaf 147

trichome morphogenesis (Downes et al., 2003; El Refy et al., 2003; Patra et al., 2013). 148

Assessment of leaf hairs across a subset of B. napus GWAS accessions with maximal 149

variation in BnaUPL3.C03 expression revealed segregation of a trichome phenotype 150

(Supplemental Figure 4), suggesting differential BnaUPL3.C03 expression also 151

influenced trichome morphogenesis in oilseed rape. Although AtUPL3 appears to have 152

a conserved role in trichome morphogenesis, there is no evidence that AtUPL3 has a 153

role in seed development. 154

In Arabidopsis, UPL3 transcript levels increased steadily during seed development, with 155

highest expression levels observed during the seed maturation phase (Figure 2A), 156

suggesting a potential role for UPL3 in seed maturation. BnaUPL3.C03 was 157

differentially expressed in seedlings of B. napus accessions displaying high variation in 158

SWPP (Figure 1F), therefore variation of BnaUPL3.C03 expression was measured 159

using RT-qPCR in developing pods at 45 Days Post Anthesis (DPA) in six accessions 160

with low- and high- SWPP. Figure 2B confirms that variation in BnaUPL3.C03 161

expression in pods was tightly correlated with that in leaves measured using RNAseq. 162

6

Arabidopsis and B. rapa mutants lacking UPL3 function exhibit increased seed 163

size 164

The potential influence of AtUPL3 in seed formation was assessed using two 165

Arabidopsis T-DNA insertion lines with essentially undetectable AtUPL3 expression 166

levels (Supplemental Figure 5). These lines were insertions in the 10th exon and were 167

named upl3-4 (SALK_015334) and upl3-5 (SALK_151005) in reference to the upl3 168

alleles previously identified (Downes et al 2003). upl3-4 and upl3-5 mutant seeds were 169

approximately 10% larger (Figure 2C) and displayed a 12% increase in seed lipid 170

content (Figure 2D) relative to seeds of WT plants. Conversely, over-expression of 3HA-171

UPL3 from the 35S promoter (see Figure 4B for protein expression levels) reduced 172

seed size and lipid content. Analysis of seed lipid composition revealed no significant 173

changes in fatty acid composition (Supplemental Table 3). Two premature stop codons 174

in the single B. rapa UPL3 gene upstream of the predicted catalytic cysteine residue 175

(Bra010737.1, Supplemental Figure 6) were identified using a TILLING (Targeting 176

Induced Local Lesions IN Genomes) resource for B. rapa (Stephenson et al., 2010). 177

Supplemental Figure 7 shows that a line homozygous for the JI32517 BraUPL3 178

mutation also had larger seeds compared to a line that segregated the BraUPL3 179

mutation. These results established a potential role for the activities of both 180

BnaUPL3.C03 and BnaUPL3.A08 (descended from BraUPL3) in regulating seed size in 181

Arabidopsis and Brassica species. 182

UPL3 indirectly influences seed mucilage biosynthesis via GL2 183

Arabidopsis upl3-4 mutant seeds exhibited altered mucilage extrusion upon imbibition 184

(Figure 2E). UPL3-mediates the proteasomal degradation of the bHLH transcription 185

factors GL3 and EGL3 (Patra et al., 2013). These proteins function as part of a complex 186

to regulate the expression of GLABROUS 2 (GL2), encoding a homeodomain 187

transcription factor that activates expression of MUCILAGE-MODIFIED 4 (MUM4), 188

encoding a mucilage biosynthetic enzyme (Shi et al., 2012). This EGL3/GL3 complex 189

activates GL2 expression in leaves, but represses GL2 expression in seeds, depending 190

on the type of MYB transcription factor in the complex (Song et al., 2011). The single 191

repeat MYB GmMYB73, a possible homolog of Arabidopsis TRY and CPC, repressed 192

7

GL2 expression in seeds and interacted with EGL3 and GL3 (Liu et al., 2014). Thus, 193

reduced UPL3-mediated destabilization of EGL3 and GL3 in the upl3-4 mutant may 194

elevate GL3 and EGL3 levels, leading to increased repression of GL2 expression. A 195

significant reduction in the expression of both GL2 (Figure 2F) and MUM4 (Figure 2G) 196

was observed at 5 DPA in developing seeds of a upl3-4 mutant and may explain the 197

altered mucilage extrusion observed in the Arabidopsis upl3-4 mutant siliques. 198

GL2 also negatively regulates seed lipid content by suppression of PHOSPHOLIPASE 199

D ALPHA 1 (PLDα1) expression (Liu et al., 2014). However, no significant difference in 200

the expression of PLDɑ1 was observed in WT and upl3 mutant seeds. Therefore, UPL3 201

may target other proteins for degradation during seed maturation that influence seed 202

size and storage reserve accumulation. 203

upl3 mutants display increased expression of seed maturation genes 204

Several genes influence both seed lipid content and seed size in Arabidopsis, including 205

APETALA 2 (AP2) (Ohto et al., 2009), LEC1 and LEC2 (Santos Mendoza et al., 2005). 206

The expression of these genes was assessed during the development of Arabidopsis 207

upl3-4 and WT seeds using RT-qPCR. No differences in the expression of AP2 (Figure 208

3A) or LEC2 (Figure 3B) were seen between upl3-4 and WT seeds. However, 209

significant increases in LEC1 expression were observed in upl3-4 mutant seeds at 10 210

DPA compared to WT (Figure 3C). Transcription of LEC1 is positively regulated by 211

LEC2 (Santos-Mendoza et al., 2008b; Baud et al., 2016). Given the observed increase 212

in LEC1, but not LEC2, expression, it was possible that altered UPL3 expression may 213

affect LEC2 protein levels, thus altering LEC1 expression in upl3-4 mutant seeds. This 214

was tested by measuring expression of WRINKLED 1 (WRI1) and MYB118, which are 215

also regulated by LEC2 (Barthole et al., 2014; Baud et al., 2009). Increased expression 216

of both genes was observed in upl3-4 mutant siliques relative to WT from 10 DPA 217

(Figures 3D and 3E), supporting the hypothesis that AtUPL3 may influence LEC2 218

protein levels and expression of target genes. 219

220

221

8

UPL3 reduces LEC2-mediated transcription of seed maturation genes 222

LEC1 and LEC2 bind to the promoters and activate the expression of seed maturation 223

genes, such as OLEOSIN 1 (OLE1), a gene required for seed lipid accumulation 224

(Santos-Mendoza et al., 2008b; Baud et al., 2016). To further assess the potential role 225

of AtUPL3 in LEC2-mediated gene expression, expression of OLE1 in maturing seeds 226

was measured in WT and upl3-4 mutant Arabidopsis. OLE1 was expressed at higher 227

levels in the upl3-4 mutant (Figure 3F), consistent with the hypothesis that UPL3 may 228

influence LEC2 protein levels. To assess if UPL3 directly affects LEC2-mediated 229

transcription of OLE1, transient expression of the Arabidopsis OLE1 promoter fused to 230

firefly luciferase was carried out in Arabidopsis upl3-4 mutant mesophyll protoplasts. 231

The low levels of endogenous OLE1 promoter activity (Figures 3G) were increased by 232

co-transfection with 35S:3HA-LEC2. Co-transfection with both 35S:3HA-LEC2 and 233

35S:3FLAG-UPL3 significantly reduced LEC2-induced OLE1 promoter activity, 234

suggesting that UPL3 reduced LEC2-mediated transcriptional regulation of seed lipid 235

biosynthetic genes. 236

UPL3 mediates the proteasomal degradation of LEC2. 237

HECT E3 ligases such as UPL3 mediate the proteasomal degradation of substrate 238

proteins by direct ubiquitylation of their substrates (Maspero et al., 2013). To test UPL3-239

mediated degradation of LEC2, the predicted active site cysteine of AtUPL3 (Downes et 240

al 2003) was mutated to glycine (C1855G) and wild-type and predicted inactive AtUPL3 241

forms were expressed in N. benthamiana as C-terminal FLAG fusions and purified on 242

FLAG-MA beads. UPL3-mediated degradation of LEC2-HIS protein was carried out in 243

cell free extracts of upl3-4 seedlings (Figure 4A). Wt UPL3-FLAG promoted a more 244

rapid reduction in LEC2-HIS levels than mutant UPL3-FLAG, and this reduction was 245

inhibited by the proteasome inhibitor MG132. To support these observations of LEC2-246

HIS stability, LEC2 protein levels in Arabidopsis upl3-4 mutant plants, in WT plants, and 247

in transgenic plants expressing 35S:3HA-UPL3 were assessed using LEC2-specific 248

polyclonal antibodies. Increased endogenous LEC2 protein levels were observed in 249

mutant Arabidopsis upl3-4 siliques compared to WT plants (Figure 4B), while reduced 250

LEC2 levels were seen in siliques of 35S:3HA-UPL3 transgenic lines compared to WT 251

9

siliques (Figure 4B). These observations indicated that AtUPL3 activity influenced LEC2 252

protein stability, and this was reflected by altered LEC2 protein levels in vivo in upl3-4 253

mutant and over-expressing lines. 254

UPL3 promotes the formation of higher MW forms of LEC2 255

Immunoblot detection of endogenous LEC2 protein in developing siliques of 35S:3HA-256

UPL3 transgenic plants revealed LEC2 forms with higher apparent MW in comparison 257

to those observed in WT protein extracts (Figure 4C), consistent with UPL3-mediated 258

ubiquitylation of LEC2. Co-expression of LEC2-HA with wt or mutant UPL3-3FLAG in 259

Nicotiana benthamiana and pull-down of ubiquitylated proteins using FLAG-TR-TUBE 260

showed that higher MW forms of LEC2-HA were specifically formed by the activity of 261

UPL3, as no HMW LEC2-HA products were detected by co-expression with predicted 262

catalytically inactive UPL3-3FLAG (Figure 4D). These HMW forms interacted with the 263

ubiquitin binding motifs on FLAG-TR-TUBE protein, suggesting they may be 264

ubiquitylated forms LEC2. Taken together, these results provide evidence that UPL3 265

promotes the formation of higher MW forms of LEC2 and destabilizes LEC2 in vitro. 266

Variation in the promoter and 5’UTR region of BnaUPL3.C03 is sufficient for 267

differential expression and influences variation in yield traits 268

We observed significant differences in BnaUPL3.C03 expression between GWAS 269

accessions displaying variation in SWPP (Figure 1F and Supplemental Table 2). 270

Alignment of BnaUPL3.C03 promoter and 5’UTR sequences (from 2 kb upstream of the 271

ATG initiation codon) of a high- (Coriander) and a low-expressing (Dimension) 272

accession revealed multiple sequence differences, including 34 Single Nucleotide 273

Polymorphisms (SNPs) and 7 indels between 3-60nt (Supplemental Figure 8). The 274

Dimension and Coriander BnaUPL3.C03 promoter and 5’UTR regions were fused to a 275

firefly luciferase reporter gene (Figure 5A) and their activities assessed following 276

transfection of Arabidopsis mesophyll protoplasts. Figure 5B shows that the Coriander 277

promoter and 5’UTR region drove approximately three times more luciferase activity 278

than the Dimension promoter in these cells. This difference in BnaUPL3.C03 expression 279

levels was consistent with RNAseq and RT-qPCR data (Figures 1E and 1F) from B. 280

10

napus seedlings, showing that variation in promoter and 5’UTR activity is the primary 281

source of variation in BnaUPL3.C03 expression between these two accessions. 282

To assess the role of BnaUPL3.C03 promoter and 5’UTR sequence variation in 283

variation in seed size, transgenic Arabidopsis lines expressing the Arabidopsis UPL3 284

coding region fused to an N-terminal 3HA epitope driven by Coriander and Dimension 285

BnaUPL3.C03 promoters were constructed. Transgenic lines were made in the upl3-4 286

mutant background to assess differential complementation of upl3-4 mutant phenotypes 287

in response to transgene expression. UPL3 expression was measured in 10 DPA 288

siliques of transgenic lines, the upl3-4 mutant, and WT plants. The two B. napus 289

promoter regions expressed the AtUPL3 coding region at the predicted levels in 290

developing siliques (Figure 5C). Seed sizes in the high-expressing Coriander promoter 291

transgenic lines showed near wild-type seed sizes, indicating complementation of the 292

large Arabidopsis seed upl3-4 phenotype. Reduced levels of complementation of the 293

large seed upl3-4 phenotype was observed in plants expressing AtUPL3 under the 294

control of the low-expressing Dimension promoter (Figure 5D). Thus, variation in 295

BnaUPL3.C03 promoter and 5’UTR transcriptional activity can influence final seed size 296

by driving different levels of AtUPL3 coding region expression and 3HA-UPL3 protein 297

accumulation (Figure 5D lower panel). 298

Differential expression of BnaUPL3.C03 leads to variation in BnaLEC2 protein 299

levels and modulates final seed lipid content. 300

Associative transcriptomics identified a negative relationship between BnaUPL3.C03 301

expression levels and SWPP in B. napus accessions (Figures 2B and 6A). Consistent 302

with seed size phenotypes seen in transgenic Arabidopsis (Figure 5D), low 303

BnaUPL3.C03 expressing lines had larger seeds (Figure 6B) and higher thousand seed 304

weights (Figure 6C). To relate these B. napus phenotypes to the proposed mechanism 305

of UPL3-mediated destabilization of LEC2 in Arabidopsis (Figure 4), LEC2 antibody was 306

used to assess LEC2 protein levels in the developing seeds of a subset of six B. napus 307

accessions varying in BnaUPL3.C03 expression (Supplemental Table 2 and Figure 2B). 308

Figure 6D shows that LEC2 protein levels were higher in all three B. napus accessions 309

with reduced BnaUPL3.C03 expression, compared to accessions with higher 310

11

BnaUPL3.C03 expression in which no LEC2 was detected in seeds. These 311

observations show that the mechanism of UPL3-mediated control of LEC2 protein levels 312

established in Arabidopsis underlies variation in seed size observed in B. napus 313

accessions. As predicted by this mechanism, significant differential expression of 314

BnaOLE1 (which is regulated by LEC2 in Arabidopsis (Figure 3G)) was observed 315

across the subset of B. napus accessions displaying differential expression of 316

BnaUPL3.C03 during seed maturation (Figure 6E). Finally, accessions displaying 317

reduced BnaUPL3.C03 and consequent increased BnaOLE1 expression levels exhibit 318

significantly higher seed lipid levels relative to those displaying high BnaUPL3.C03 319

expression (Figure 6F). 320

321

Reduced BnUPL3.C03 expression has not yet been exploited in current B. napus 322

breeding material 323

To assess if variation in BnaUPL3.C03 expression levels has been selected in the 324

breeding of commercial oilseed rape varieties, expression was measured across a 325

panel of seven current elite B. napus lines. Expression levels were compared to those 326

measured in GWAS accessions exhibiting high differential expression of BnaUPL3.C03 327

and high variation in yield traits (Figure 7). There was substantial variation in 328

BnaUPL3.C03 expression in the elite lines, with relatively high levels of expression in 329

several lines compared to the low expressing line Licrown x Express. Therefore, there 330

appears to be a significant potential for further yield increases in elite oilseed rape 331

germplasm through reduction of BnaUPL3.C03 expression. 332

333

DISCUSSION 334

Using Associative Transcriptomics (Harper et al., 2012), we identified variation in the 335

expression of a B. napus gene, BnaUPL3.C03, that modulates seed size, lipid content 336

and field yields in this important oilseed crop. BnaUPL3.C03 encodes an orthologue of 337

the Arabidopsis HECT E3 ligase UPL3. We show that it promotes the formation of HMW 338

12

products of LEC2 with an affinity for ubiquitin-binding proteins and destabilizes this 339

“hub” transcriptional regulator of seed storage processes (Boulard et al., 2017). 340

Reduced BnaUPL3.C03 expression maintains higher levels of LEC2 during seed 341

maturation, prolonging transcriptional activation of storage lipid genes, leading to larger 342

seeds with elevated lipid levels in B. napus. Lines with relatively low BnaUPL3.C03 343

expression had robust yield increases in field conditions compared to lines with higher 344

BnaUPL3.C03 expression. Variation in UPL3 expression levels in elite oilseed rape 345

cultivars identifies a promising approach for achieving further increases in oilseed yields 346

by selecting for reduced UPL3 expression levels. 347

Comparison of assemblies of the BnaUPL3.C03 gene and flanking sequences from two 348

B. napus lines with low and high BnaUPL3.C03 expression identified high levels of 349

promoter sequence variation that segregated with high SWPP and reduced 350

BnaUPL3.C03 expression (Figures 5B and 5C). Transient assays showed that 2kb 5’ 351

flanking regions and UTRs from high- and low-expressing BnaUPL3.C03 accessions 352

promoted the expected expression differences (Figure 5B). Driving the expression of 353

the Arabidopsis UPL3 coding sequence with B. napus promoter variants led to 354

differential complementation of the Arabidopsis upl3-4 mutant seed phenotypes (Figure 355

5D). This established that natural variation in the activity of the BnaUPL3.C03 promoter 356

and 5’UTR region was sufficient to cause variation in seed size and yield traits in the 357

accessions. Expression of BnaUPL3.A08 was not associated with SWPP (Figure 1D), 358

nor did its expression vary in developing seeds of accessions with high or low SWPP 359

(Supplemental Figure 3). This suggested that the primary influence of UPL3 expression 360

levels on SWPP is from variation in BnaUPL3.C03 expression. The multiple variants 361

detected between two promoters driving differential expression and the continuous 362

profile of BnaUPL3.C03 expression in the accessions (Figure 1E) suggest there may be 363

multiple sequence variants that together reduce promoter activity. Although limited, 364

variation in other regions of the BnaUPL3.C03 gene, including the coding region, may 365

also have the potential to contribute to variation in transcript abundance and yield 366

observed in B. napus varieties. 367

13

Such genetic variation influencing gene regulation is an increasingly important resource 368

for trait improvement, including increasing seed yields. For example, natural variation in 369

the copy number of a promoter silencing element of the FZP gene underlies variation in 370

spikelet numbers in rice panicles and is an important determinant of yield (Bai et al., 371

2017). Similarly, QTL analyses identified promoter variation in the rice GW7 gene that, 372

combined with variation that reduced expression of an SPL16 transcriptional repressor 373

of GW7, led to 10% increases in grain yield and improved quality (Wang et al., 2015). A 374

deletion in the promoter region of GW5 in Nipponbare rice lines reduced expression and 375

increased grain width (Liu et al., 2017), and similarly, promoter deletions in GSE5 376

accounted for wide grains in indica rice varieties (Duan et al., 2017). More generally, 377

genetic variation in chromatin accessibility (a mark of promoter activity) explained about 378

40% of heritable variation in many quantitative traits in maize (Rodgers-Melnick et al., 379

2016). These reports, and the study described here, reveal the exceptional promise of 380

accessing variation in promoter sequences and altered transcriptional activity for 381

identifying regulatory mechanisms and for the quantitative manipulation of complex 382

traits such as yield in crop plants. 383

UPL3 was first identified in Arabidopsis as a HECT E3 ligase gene whose loss-of-384

function mutation causes increased leaf hair branching (Downes et al., 2003; El Refy et 385

al., 2003). Here, we showed that UPL3 expression increased during seed maturation in 386

Arabidopsis (Figure 2A), and an ortholog, BnaUPL3.C03, was differentially expressed in 387

the seeds of B. napus lines (Figures 2B and 6A) that varied in seed weight (Figure 6C) 388

and seed lipid content (Figure 6F). LEC2, a transcriptional regulator of seed storage 389

processes, is more stable in siliques a upl3-4 loss-of-function mutation in Arabidopsis 390

(Figure 4B) and in B. napus accessions with relatively lower BnaUPL3.C03 expression 391

levels (Figure 6D). AtUPL3 physically interacted with LEC2 (Figure 4C), suggesting a 392

direct functional relationship. UPL3-FLAG, but not a predicted catalytically inactive form, 393

promoted the proteasomal- dependent instability of LEC2-HIS in vitro (Figure 4A), and 394

expression of 3HA-UPL3 in transgenic Arabidopsis promoted the formation of HMW 395

forms of LEC2 in developing seeds (Figure 4C). Finally, co-expression in plants of 396

UPL3-3FLAG promoted the formation of higher MW forms of LEC2HA that were 397

specifically purified by TR-TUBE, which has a high affinity for ubiquitylated proteins 398

14

(Figure 4D). Conversely, a predicted catalytically inactive form of UPL3-3FLAG did not 399

lead to the formation of these higher MW forms of LEC2-HA. 400

These observations suggested that UPL3 ubiquitin ligase activity directly modulated 401

levels of LEC2 protein during seed maturation. LEC1 is directly activated by LEC2, and 402

together they activate the expression of genes involved in promoting seed lipid 403

accumulation, such as OLE1. Consistent with this, we observed elevated OLE1 404

expression levels in Arabidopsis upl3-4 mutant seeds relative to WT and showed that 405

BnaOLE1 is differentially expressed between B. napus accessions displaying differential 406

expression of BnaUPL3.C03 (Figure 6E). These observations show that reduced levels 407

of UPL3 maintain higher levels of LEC2 protein during seed maturation, thus prolonging 408

the duration of expression of seed maturation genes, leading to increased seed lipid in 409

Arabidopsis upl3-4 mutants and B. napus accessions with reduced BnaUPL3.C03 410

expression. 411

At earlier stages of seed development, the seed coat in Arabidopsis upl3-4 mutants has 412

altered mucilage and reduced expression of GL2, a transcription factor controlling 413

epidermal cell differentiation (Figure 2F) (Lin et al., 2015). GL2 promotes expression of 414

the rhamnose biosynthesis gene MUM4 that is required for seed mucilage production 415

(Lin et al., 2015; Shi et al., 2012), and expression of MUM4 is also reduced during testa 416

development in upl3-4 (Figure 2G). This observation revealed a unifying role of UPL3 in 417

regulating both testa and embryo maturation by modulating levels of transcription 418

factors during different stages of seed development. These transcription factors, GL3 419

and EGL3 (Patra et al., 2013), and LEC2 (this study) in turn modulate expression of 420

other transcription factors and biosynthetic genes involved in testa and embryo 421

development. 422

Altered expression of LAFL genes has profound developmental consequences such as 423

ectopic embryogenesis (Stone et al., 2001; Roscoe et al., 2015), but induced 424

expression of LEC2 in vegetative tissues does increase lipid accumulation (Andrianov et 425

al., 2010; Santos Mendoza et al., 2005). These studies showed that the activities of 426

LEC2 expression in storage processes and embryo development were difficult to 427

separate, probably due to the timings of expression, interdependence and partial 428

15

redundancy of LAFL gene function. By identifying a mechanism controlling LEC2 protein 429

levels during seed maturation, we have shown that it is possible to elevate lipid levels 430

during normal Arabidopsis embryo development (Figure 2D). Mis-expression of WRI1 431

permits normal seed development while increasing lipid content (Kanai et al., 2016; van 432

Erp et al., 2014) by extending seed maturation, consistent with our observations of 433

elevated WRI1 expression in upl3-4 (Figure 3E). Intensive breeding is optimizing 434

oilseed lipid composition for different end-uses, but comparatively slow progress is 435

being made in increasing yields of oilseed crops such as oilseed rape, with current rates 436

of yield increase predicted to be insufficient to meet future needs (Ray et al., 2013). 437

Genetic variation that reduces expression of BnaUPL3.C03 and increases seed lipid 438

content appears not to have been exploited in oilseed rape breeding (Figure 7), 439

demonstrating how lipid content and seed yields could be increased without influencing 440

composition. The presence of a potential LEC2 ortholog in soybean (Manan et al., 441

2017), and the expression of LEC2 and other B3-domain transcription factor homologs 442

during seed lipid synthesis in sunflower (Badouin et al., 2017) and oil palm (Singh et al., 443

2013) reveals the potential of UPL3-mediated regulation of LEC2 to increase seed lipid 444

levels and overall yields in other major oilseed crops. 445

METHODS 446

Plant material and growth conditions 447

Phenotype data were collected from 94 accessions representing winter, spring and 448

Chinese oilseed rape from the OREGIN Brassica napus fixed foundation diversity set 449

(Harper et al., 2012). Plants were grown in a randomised, triplicated experimental 450

design in a Kedar greenhouse under natural light with no controlled heating. Prior to 451

transplantation, plants were grown (18/15°C day/night, 16hr light) for 4 weeks before 6 452

weeks vernalisation (4°C, 8hr light). 20 typical pods were collected from each mature 453

plant and digitally imaged. Pod length (Podl) was measured using ImageJ (Schneider et 454

al., 2012). Pods were weighed (PW) before threshing to remove seeds. Seed numbers, 455

average seed length (SL), width (SW), area (SA), single seed weight (SSW) and 456

thousand grain weight (TGW) were measured for each sample using a Marvin device 457

(GTA Sensorik GmbH, Germany). Numbers of seeds per pod (SPP), seed weight per 458

16

pod (SWPP) and seed density (SDen) were calculated from these data. Field yields of 459

selected accessions were grown in four replicated field plots (1.25m x 6m, Church 460

Farm, Bawburgh, Norfolk, UK) in a randomized design and harvested by combine. Total 461

plot yield was determined for each plot and an average plot yield taken. 462

All Arabidopsis thaliana mutant and transgenic lines used were in Col-0 background. 463

Plants were grown on soil in a Sanyo growth chamber using cool white fluorescent 464

lighting with 16/8 hr day/night at 22°C after 48 hours stratification at 5°C. Two 465

independent sequence indexed T-DNA insertion lines in the 10th exon of At4G38600 466

(UPL3), Salk_015334 (termed upl3-4) and Salk_151005 (termed upl3-5), were obtained 467

from The Nottingham Arabidopsis Stock Centre (NASC). Genotyping primers were 468

designed using the primer design tool http://signal.salk.edu/tdnaprimers.2.html: The 469

primer sequences are in Supplemental Table 2. Genotyping used TAKARA EX taq 470

(Takara Bio, USA). Both SALK alleles abrogated gene expression (Supplemental Figure 471

5) 472

To identify loss-of-function mutations in UPL3 in Brassicas, a TILLING population 473

(Stephenson et al., 2010) of B. rapa (the A genome donor to B. napus) was screened 474

for mutations in the predicted gene Bra010737.1 on chromosome A08 that encodes an 475

ortholog of Arabidopsis thaliana UPL3. The predicted B. rapa gene was defined by a 476

full- length transcript. There was only one clear ortholog of UPL3 in B.rapa, consistent 477

with two copies present in the amphidiploid B. napus on chromosomes A08 and C03. 478

Two premature stop codon mutations were identified in Bra010737.1 (Supplemental 479

Figure 6) in lines JI32517 and JI30043. Primers were designed to amplify genomic DNA 480

from the mutated region (Supplemental Table 4) and used to validate the TILLING 481

mutations. Line JI32517-B was used in further studies. This line was selfed and progeny 482

screened for homozygous (G>A) and “WT” (C>T) changes by sequencing of the locus. 483

Seeds were harvested from the “WT” and the upl3 mutant and their area measured 484

using ImageJ. 485

486

A subset of B. napus accessions and a panel of elite B. napus breeding lines were 487

grown under glasshouse conditions after vernalisation. Leaf material was harvested 488

17

from the first true leaf and stored at -70℃ prior to further processing. Developing pods 489

were staged by tagging when petals were beginning to emerge from the developing 490

bud, taken as zero days post anthesis (DPA). Samples for RNA isolation were collected 491

at 45 DPA and stored at -70℃. For expression analyses in Arabidopsis, WT Col 0 and 492

upl3 mutant plants were grown as described, without vernalization, and floral buds 493

tagged when petals were beginning to emerge from the developing bud (0 DPA). 494

Siliques were then harvested at 0, 5, 10 and 15 DPA and tissue samples stored at -495

70℃. 496

Population structure analysis 497

Associative Transcriptomics analysis was carried out as described (Harper et al., 2012). 498

The population structure Q matrix was re-calculated using 680 unlinked markers across 499

the set of 94 lines. Run-length comprised 10,000 burn-in followed by 10,000 steps. The 500

ancestry model was admixture, and allele frequencies independent between 501

populations. Between 1 and 10 values of K were tested using 3 iterations. One SNP per 502

500 kb interval along pseudomolecules, excluding regions less than 1000 kb from 503

centromeres (Mason et al., 2017), were selected for Bayesian population structure 504

analysis via STRUCTURE 2.3.3 (Pritchard et al., 2000). This analysis incorporated a 505

minor allele frequency of 5%. The optimum number of K populations was selected as 506

described (Harper et al., 2012). 507

SNP analysis 508

SNP data, STRUCTURE Q matrix and phenotypes of the 94 accessions were combined 509

using TASSEL (V4.0). Following the removal of minor alleles (frequency <0.05) 510

~144,000 SNPs were used to calculate a kinship (K) matrix to estimate the pairwise 511

relatedness between individuals. Data sets were analyzed using both Generalised and 512

Mixed Linear Models (GLM and MLM). Goodness of fit of the model was determined by 513

a QQ plot (Supplemental Figure 9) of the observed versus the expected -log10P values. 514

-log10P values were plotted in chromosome order and visualized using R programming 515

scripts as previously described (Harper et al 2012) ( http://www.R-project.org/.) . 516

Relationships between transcript levels (RPKM) and trait data variation were analysed 517

18

by Linear regression analysis using R. -log10P values were plotted in pseudomolecule 518

order and visualized using R scripts as described previously (Harper et al 2012). 519

Arabidopsis and Brassica seed size quantification 520

Arabidopsis seeds were harvested from mature plants and imaged at 10x magnification. 521

B. napus and B.rapa seeds were harvested from mature plants and scanned using a 522

photocopier. Seed area was quantified using ImageJ particle analysis. 523

Ruthenium red staining 524

Seed mucilage phenotypes were assessed using methods described by (McFarlane et 525

al., 2014). Stained seeds were imaged at 10x magnification. 526

Seed lipid quantification and profiling 527

Fatty acid profile analyses in Arabidopsis were carried out using the methods described 528

by (Li et al., 2006). Lipid content of B. napus seeds were measured using Near Infrared 529

Spectroscopy (Wang et al., 2014). 530

PCR and sequencing 531

All PCR reactions were carried out using Phusion High Fidelity DNA polymerase (New 532

England BioLabs) according to manufacturer's instructions. Capillary sequencing was 533

carried out by GATC Biotech (Germany). 534

cDNA synthesis and RT-qPCR 535

RNA was extracted using the SPECTRUM Total Plant RNA kit (Sigma, UK). 1ug of RNA 536

was treated with RQ1 RNase-Free DNase (Promega, USA) and cDNA synthesis was 537

carried out using the GoScript Reverse Transcription system (Promega, USA) using 538

OligoDT. All protocols were carried out using manufacturers’ guidelines. cDNA samples 539

were diluted 1:10 in water before use. RT-qPCR was carried out using SYBR green real 540

time PCR mastermix (Thermofisher) and performed using Lightcycler 480 (Roche, 541

Switzerland). Primer sequences used for RT-qPCR are in Supplemental Table 4. Primer 542

19

efficiencies and relative expression calculations were performed according to methods 543

described (Pfaffl, 2001). All RT-qPCR assays were repeated at least twice. 544

DNA constructs 545

The p35S:3HA-AtUPL3 transgenic line was generated by cloning AtUPL3 cDNA into the 546

pENTR TOPO-D vector (Thermofisher, UK) using the primers described in 547

Supplemental Table 4. LR Clonase mix II (Thermofisher, UK) was used to transfer the 548

AtUPL3 CDS into the 35S PB7HA binary vector. The Arabidopsis UPL3 cDNA TOPO 549

construct was cloned into the 3xFLAG PW1266 vector. The active site cysteine 1855 550

residue (Downes et al 2003) was mutated to glycine using GeneArt kits and primers 551

described in Supplemental Table 4. Full-length Arabidopsis LEC2 cDNA was amplified 552

using primers described in Supplemental Table 4. Following cloning into pENTR TOPO-553

D, the LEC2 CDS was transferred by LR reaction to pEARLEY 103. Constructs were 554

transformed into Agrobacterium tumefaciens strain GV3101, and Arabidopsis upl3-4 555

mutant plants were transformed using the floral dip method (Clough and Bent, 1998). 556

Promoter regions of the BnaUPL3.C03 gene from the high- and low- expressing 557

accessions Coriander and Dimension were amplified using primers described in 558

Supplemental Table 4. BnaUPL3.C03 promoter PCR products, digested with Stu1 and 559

Xho1, were ligated into pEarly 201-AtUPL3-3FLAG CDS plasmid. The resulting 560

BnaUPL3.C03 promoter:AtUPL3-3FLAG CDS constructs was transformed into 561

Agrobacterium tumefaciens strain GV3101, and Arabidopsis upl3-4 mutant plants were 562

transformed using the floral dip method (Clough and Bent, 1998). 563

Promoter transactivation assay 564

Promoters and full-length cDNAs of selected Arabidopsis Col-0 genes were amplified by 565

PCR using Phusion polymerase (Thermo Scientific) according to the manufacturer's 566

guidelines. Promoter primer sequences are in Supplemental Table 4. PCR reactions 567

were purified using Wizard SV Gel and PCR Clean up system (Promega, UK) and 568

inserted into pENTR D-TOPO vector (Thermofisher, UK), and an LR reaction was used 569

to clone promoters into a Firefly Luciferase reporter vector fLUC (pUGW35). LEC2 and 570

UPL3 CDS were transferred using LR Clonase into PB7HA and PW1266 respectively to 571

20

create p35S:3HA-LEC2 and p35S:3FLAG-UPL3. A 35S:Renilla luciferase construct was 572

used to quantify relative promoter activities. Plasmid preparations for transient assays 573

were prepared using the Qiagen Plasmid Maxi Kit according to manufacturer’s 574

instructions. 575

Promoter transactivation assays were carried out using protoplasts isolated from upl3 576

leaves (Wu et al., 2009). Assays were carried in triplicate using 5g of plasmid and 577

100l of purified protoplasts (approximately 50,000 cells). After an overnight incubation 578

at room temperature, transfected protoplasts were harvested and promoter activity 579

assessed using the Dual Luciferase Reporter assay system (Promega, USA). The ratio 580

of Firefly Luciferase to Renilla Luciferase activity was determined using the dual assay 581

Promega protocol on a Glomax 20/20 luminometer (Promega, USA). All transactivation 582

assays were conducted in triplicate and repeated at least twice. 583

Total protein extraction from B. napus pods and Arabidopsis siliques 584

Material was ground to a fine powder in liquid nitrogen and resuspended in extraction 585

buffer (1ml/g fresh weight) (25mM Tris-HCl pH 8.0, 10mM NaCl, 10mM MgCl2, 4mM 586

AEBSF and 50 µM MG132). Following an incubation on ice for 30 minutes, samples 587

were centrifuged at 15,000rpm for 5 minutes at 4°C. Supernatant was then added to a 588

fresh tube and centrifugation repeated for 10 minutes. Total protein content was 589

assessed using Bradford reagent (Bio-Labs) according to manufacturer's instructions. 590

Immunoblot analysis 591

LEC2 protein was assessed directly using affinity-purified rabbit polyclonal antibodies 592

raised against the antigenic peptide ARKDFYRFSSFDNKKL from LEC2 coupled to 593

keyhole limpet haemocyanin (New England Peptides, USA). Antibodies were used at 594

the following dilutions: anti-HIS (Sigma A7058 lot 088M4865V) 1/2000; anti-FLAG-HRP 595

(Sigma F1804 lot SLBW5142) 1/1000; anti-GFP-HRP (Miltenyi 120-002-165 lot 596

5196017043) 1/5000; anti-tubulin (Sigma T9026 lot 086M4773V) 1/5000; anti-HA-HRP 597

(Roche 3F10 1588800 lot 12013819001) 1/1000; anti LEC2 (New England Peptides) 598

1/1000. Secondary antibodies used for tubulin and LEC2 were anti-rabbit (Sigma A0545 599

21

lot 102M4823) 1/5000; for tubulin anti-mouse (Sigma A8924 lot SLBH4089) 1/5000. 600

Immunoblots were developed with FEMTO Max peroxidase substrate (Fisher Scientific). 601

Protein expression in E. coli BL21 602

The Arabidopsis LEC2 CDS was amplified from seedling cDNA using primers described 603

in Supplemental Table 4. Following purification, PCR products were TOPO cloned into 604

pET-24a to generate a C-terminal HIS fusion. FLAG-TR-TUBE for expression in E. coli 605

was as described (Dong et al 2017). Plasmids were transformed into BL21 E. coli cells, 606

grown in liquid culture until the OD 600nm was approximately 1. IPTG was then added 607

to 1mM and the culture incubated at 28°C for 3 hours to induce protein expression. 608

Cultures were centrifuged at 3,500rpm for 10 minutes at 4°C and the cell pellet 609

suspended in 7.5ml of suspension buffer (50mM HEPES pH 7.5, 150mM NaCl, 1% 610

TritonX-100; 10% glycerol; 1 Roche EDTA-FREE inhibitor cocktail tablet) plus 2.5 U/ml 611

Beconase. Cells were sonicated for 4 x 10 seconds with 20 sec intervals on ice, and 612

sonicates were centrifuged at 12,000g for 20 minutes at 4°C. Protein purification was 613

carried out using Dynabeads ® His-tag magnetic beads (Novex) or FLAG-MA beads 614

(Sigma). Prior to use, beads were washed 3 times in 50 mM HEPES pH7.5; 150 mM 615

NaCl; 10% glycerol. Sonicates were incubated with washed beads at 4°C with rotation 616

for at least 2hrs. Beads were then washed 3 times with suspension buffer and 3 times 617

with wash buffer. Proteins were eluted with elution buffer (1x PBS, 0.3M NaCl, 0.1% 618

Tween-20, 10% glycerol) containing either 300mM Imidazole for LEC2-HIS or 200 g/ml 619

FLAG peptide (Sigma). Purified protein was quantified using Qubit reagents, buffer 620

exchanged with 50mM Tris-HCl pH 8.0, 10% glycerol and stored at -70°C in 20 l 621

aliquots. 622

Cell-free degradation assay 623

One g of 10-15 DPA seedlings of upl3-4 plants was extracted with 1 ml of 50mM Tris-624

HCl pH 7.5, 100mM NaCl, 10mM MgCl2, 5mM DTT and 5mM ATP, and centrifuged 625

twice. Aliquots of 200 l were taken and 1 g of recombinant LEC2-HIS, and 5 l of 626

purified UPL3-3FLAG or UPL3-mut-3FLAG from N. benthamiana were added. 627

Reactions were carried out at 22oC with or without 50 M MG132. Samples were taken 628

22

at 0, 10, 30, 60 and 90 mins and SDS sample buffer added to stop reactions. Samples 629

were denatured at 90oC for 10 mins and electrophoresed on 4-20% SDS-PAGE gels. 630

Anti-HIS HRP antibody visualized LEC2-HIS protein levels. 631

Transient expression in Nicotiana benthamiana 632

Constructs were transformed into Agrobacterium tumefaciens GV3101 and 10 ml 633

cultures grown at 28°C for 16 hours before transfection into leaves of 4 week- old 634

Nicotiana benthamiana plants. 50 M MG132 in 10 mM MES-KOH, pH 7.5, 10 mM 635

MgCl2 was infiltrated into leaves 6 hr before harvest. After 48-60 hours, transfected 636

leaves were excised, frozen in liquid nitrogen and stored at -70oC before extraction. 637

Protein extracts were made using 1 gm tissue/ 2ml extraction buffer. Extraction buffer 638

was 10 mM Tris HCl, pH 7.5, 150 mM NaCl, 0.5 mM EDTA, 0.5% NP40, 10% glycerol, 639

1mM AESBF, 1x EDTA-free protease cocktail (Roche), plus 50 M MG132. Samples 640

were centrifuges at 4oC for 10mins at 15,000 g, filtered through Miracloth and used for 641

affinity purification. Where indicated, 20 g FLAG-purified FLAG-TR-TUBE protein was 642

added to each 5 ml extraction. Affinity purification was carried at 4oC with rotation for 2 643

hrs. Pull-down experiments used co-expression of UPL3-3FLAG with 3HA-LEC2, 644

binding to FLAG-MA beads, washing with extraction buffer and elution with SDS-sample 645

buffer. To detect LEC2 in plants total protein was extracted as described above from 646

pooled 10-15 DPA siliques or leaves of 35SS:3HA-AtUPL3 transgenic, Col 0 or lec2 647

mutant plants and immunoblotted with anti-LEC2 antibodies. 648

649

Accession numbers 650

Sequence data from this article can be found in the Arabidopsis Genome Initiative or 651

GenBank/EMBL databases under the following accession numbers: 652

Arabidopsis thaliana LEC2, At1G28300, accession OAP12706.1, UniProt Q1PFR7. 653

Arabidopsis thaliana UPL3, At4G38600, accession NP_001329354.1, UniProt 654

Q6WWWW4. Brassica napus BnaUPL3.C03, BnaC03g60070D XP_013685717.1. 655

23

BnaUPL3.A08, BnaA08g17000D-1, XP_022545849.1. Brassica rapa BraUPL3 656

Bra010737.1 (XM_009111293.2). 657

658

Supplemental Data 659

660

Supplemental Figure 1. Manhattan plots for SWPP SNP and GEM analyses. 661

662

Supplemental Figure 2. Comparison of predicted protein sequences of the Arabidopsis 663

UPL3 gene and the two B. napus orthologs, BnaUPL3.C03 and BnaUPL3.A08. 664

665

Supplemental Figure 3. Expression of BnaUPL3.A08 in 45 DPA seeds in a subset of 666

B.napus GWAS accessions displaying high variation in SWPP.667

668

Supplemental Figure 4. Segregation of a leaf hair phenotype across Brassica napus 669

accessions displaying differential expression of BnC03UPL3 and high variation in seed 670

yield. 671

672

Supplemental Figure 5. RT-qPCR analysis of T-DNA insertion alleles in AtUPL3. 673

674

Supplemental Figure 6. Predicted sequence of Brassica rapa BraUPL3 protein. 675

676

Supplemental Figure 7. Seed areas of B. rapa TILLING lines. 677

678

Supplemental Figure 8. Promoter and 5ʹUTR variation segregating between Brassica 679

napus GWAS accessions displaying differential expression of BnUPL3.C03. 680

681

Supplemental Figure 9. Quantile-Quantile Plot of Seed Weight Per Pod. 682

683

Supplemental Table 1. A list of B. napus accessions and trait data 684

used in GWAS and Associative Transcriptomics analyses. SWPP is 685

seed weight per pod. 686

Supplemental Table 2. Segregation of SWPP and BnUPL3.C03 expression across a 687

subset of accessions using in GWAS. 688

24

Supplemental Table 3. Seed fatty acid content and composition in Col 0 and upl3-4 689

mutant. 690

691

Supplemental Table 4. Primers used. 692

693

694

695

Acknowledgements 696

697

This work was supported by an ERA-CAPS grant (“ABCEED”) to MWB and LL. MWB 698

was also supported by a Biological and Biotechnological Sciences Research Council 699

(BBSRC) Institute Strategic Grants GRO (BB/J004588/1) and GEN (BB/P013511/1). LL 700

is also supported by Labex Saclay Plant Sciences-SPS (ANR-10-LABX-0040-SPS). We 701

thank Jingkun Ma for advice on transient expression and luciferase assays. 702

703

Author Contributions 704

CM, MWB, RW, BD and LL designed the research; CM, NMcK, RW, JB, MWB and AF 705

performed research, CM and MWB analyzed data and wrote the paper. 706

707

Competing Interests 708

709

The authors declare no competing interests. 710

Figure 1. Association of variation in Seed Weight Per Pod (SWPP) with SNPs and 711

differential expression of BnaUPL3.C03 in a panel of 94 Brassica napus accessions. 712

A. Associations between single-nucleotide polymorphism (SNP) on chromosomes A08/C03713

and SWPP. The dotted grey lines outline the genomic location of peaks of SNP714

associations with SWPP. Markers are plotted in pseudomolecule order and associations715

as -log10 P values. The colored regions under the pseudochromosome represent the716

regions of sequence similarity to A. thaliana chromosomes, as described previously717

(Harper et al. 2012). The dashed horizontal red line indicates the Bonferroni-corrected718

significance threshold.719

B. Normal distribution of the SWPP trait in the set of B. napus accessions, showing the720

data were suitable for Mixed Linear Model analyses.721

25

C. Segregation of SWPP trait means with the most highly associating marker 722

(JCVI_5587:125) show a marker effect of ~20%. Data are given as mean ± SE. P values 723

were determined by Student’s t-test. 724

D. Differential expression of a single C genome assigned unigene C_EX097784 on725

chromosome C03 was associated with SWPP variation. This unigene is an ortholog of726

Arabidopsis UBIQUITIN PROTEIN LIGASE 3 (UPL3), and is termed BnaUPL3.C03.727

The association exceeded the adjusted P value calculated by the Benjamini-Hochberg728

method (Q= 0.026).729

E. Correlation of BnaUPL3.C03 expression in seedling leaves measured by RPKM and RT-730

qPCR in six B. napus GWAS accessions (Supplemental Table 2) exhibiting maximal731

variation in C-EX097784 expression. Measurements of RT-qPCR of BnaUPL3.C03732

expression in seedlings were expressed relative to BnaUBC10 expression levels. Lines733

with high SWPP are shown by orange data points, and lines with low SWPP are shown734

by gray data points.735

F. Correlation of BnaUPL3.C03 transcript abundance in seedlings, measured using RT-736

qPCR in six B. napus GWAS accession exhibiting maximal variation in SWPP, with737

variation in SWPP. Lines with high SWPP are shown by orange data points, and lines738

with low SWPP are shown by gray data points.739

G. A subset of ten GWAS accessions with maximal variation in SWPP (Supplemental Table740

2) were grown in replicated plots in field conditions between March-August and mean741

plot yields measured after combining. Plot yields are shown as mean ± SE. P values 742

were determined by Student’s t-test. 743

744

Figure 2. A loss-of-function mutation in Arabidopsis UPL3 has increased seed size and 745

altered seed storage and seed coat phenotypes. 746

A. UPL3 expression increases during seed development in Arabidopsis. AtUPL3747

expression, measured by RT-qPCR, during seed development in wt Col. Expression748

levels are relative to EF1ALPHA expression. DPA is Days Post Anthesis. Data are given749

as mean + SE of three biological replicates.750

B. BnaUPL3.C03 expression in developing pods of Brassica napus accessions described in751

Figure 1E was correlated with expression in leaves. BnaACTIN2 was used as an internal752

RT-qPCR control. RT-qPCR analyses used three biological replicates. Lines with high753

SWPP are represented by orange data points, and lines with low SWPP are shown as754

26

gray data points. 755

C. Loss-of-function mutants of AtUPL3 have enlarged seeds, while AtUPL3 expression756

from the 35S promoter in Col 0 plants reduced seed size. Areas of Arabidopsis seeds757

from Col 0, a representative transformant, and two independent upl3 T-DNA loss-of-758

function mutants. Data are given as mean + SE (means calculated across 6 biological759

replicates per genotype using 100 seeds per replicate). P values were determined by760

Student’s t-test.761

D. Loss-of-function mutants of AtUPL3 has elevated seed lipid content, while AtUPL3762

expression from the 35S promoter in Col 0 plants reduced lipid content. Lipid content of763

Arabidopsis seeds from Col 0, two upl3 T-DNA mutants, and a representative 35S:3HA-764

UPL3 transformant was measured using Near Infrared Spectroscopy. Data are given as765

mean + SE (means calculated across 6 biological replicates per genotype using 100766

seeds per replicate). P values were determined by Student’s t-test.767

E. Altered mucilage extrusion of imbibed Arabidopsis seeds in the upl3-4 mutant, visualized768

by Ruthenium Red staining.769

F. Altered expression of the regulatory transcription factor GL2 during early stages of seed770

development in Atupl3-4 mutant seeds. RNA from whole siliques harvested at 5-10 DPA771

was used. Expression levels are relative to EF1ALPHA expression. Data are given as772

means + SE (n=3 biological replicates). P values were determined by Student’s t-test.773

G. Altered expression of the regulatory transcription factor MUM4 during early stages of774

seed development in upl3-4 mutant seeds. RNA from whole siliques harvested at 5-10775

DPA was used. Expression levels are relative to EF1ALPHA expression. Data are given776

as means + SE (n=3 biological replicates). P values were determined by Student’s t-test.777

778

Figure 3. UPL3 reduces LEC2-mediated gene expression. 779

Altered expression of Arabidopsis seed maturation genes in the upl3-4 mutant. RT-qPCR was 780

used to measure gene expression during seed development in Col 0 and upl3-4. RNA from 781

whole siliques was used, harvested at the times indicated on the x axis. Expression levels are 782

relative to EF1ALPHA gene expression. Data are given as means + SE (n=3 biological 783

replicates). P values were determined by Student’s t-test. 784

A, AP2 expression; B, LEC2 expression; C, LEC1 expression; D, MYB118 expression; E, WRI1 785

expression; F. RT-qPCR measurement of OLE1 in 10 DPA Col 0 and upl3-4 mutant Arabidopsis 786

plants. RNA from whole siliques was used. Expression levels are relative to EF1ALPHA gene 787

expression. Data are given as means + SE (n=3 biological replicates). P values were 788

27

determined by Student’s t-test. 789

G. AtUPL3 reduces LEC2-mediated activation of the OLE1 promoter. Transient expression of 790

the LEC2- regulated OLE1 promoter:Firefly Luciferase (fLUC) reporter gene in Arabidopsis upl3-791

4 mutant leaf protoplasts. 35S:Renilla LUC vector was co-transfected in all treatments as a 792

control, and the ratio of Firefly/Renilla LUC activity was used to determine OLE1:fLUC gene 793

expression levels. 35S:3HA-LEC2 and 35S:3FLAG-UPL3 were co-transfected as shown. Data 794

are presented as means + SE. (n=3 independent transfections). P values were determined by 795

Student’s t-test. 796

797

Figure 4. UPL3 mediates the proteasomal degradation of LEC2-HIS and mediates the 798

formation of higher MW forms of LEC2 in plants. 799

A. Cell-free degradation of LEC2-HIS protein. Purified LEC2-HIS expressed in E. coli was 800

incubated at 22oC for the times indicated in total protein extracts from upl3-4 mutant 801

seedlings, with either wt- or mutant purified UPL3-3FLAG added (see side panel), with 802

and without 50 μM MG132. Immunoblots of reactions were probed with anti-HIS 803

antibodies. CBB shows a portion of the reaction stained with Colloidal Coomassie Blue 804

as a loading control. The graph shows results of four independent cell-free reactions 805

using the same batch of purified LEC2-HIS protein. P values were determined by 806

Student’s t test. 807

B. Immunoblots of protein samples from leaf or 10-15 DPA siliques of Col 0, lec2 mutant, 808

upl3-4 mutant and 35S:3HA-UPL3 plants electrophoresed on SDS-PAGE gels and 809

probed with anti-LEC2, anti-HA or anti-tubulin antibodies as a loading control. TUBULIN 810

levels are to compare protein loading. 811

C. Expression of 3HA-UPL3 increases higher MW forms of LEC2 in developing siliques. 812

Immunoblots of protein samples from pooled 5-10 DPA siliques of Col 0 or 35S:3HA-813

UPL3 plants were electrophoresed on 4-20% SDS-PAGE gels and probed with anti-814

LEC2. Higher MW forms of LEC2 protein seen in the 35S:3HA-UPL3 sample are 815

indicated by arrows. The lower panel shows that both wild-type and mutant UPL3-816

3FLAG interact with LEC2-HA during transient co-expression in N. benthamiana leaves. 817

D. Purification of ubiquitylated forms of LEC2-HA from transiently expressed 35S:LEC2-HA 818

and 35S:UPL3-3FLAG (wild type and mutant) vectors in Nicotiana benthamiana leaves. 819

Protein samples were extracted and samples taken to assess protein expression levels 820

using immunoblotting (loading panels). The remaining samples were split in two and 821

approximately 20 g FLAG-TR-TUBE was added to one set (lower loading panel). This 822

28

fraction was purified on FLAG-MA beads and the other purified using HA-MA beads. 823

Affinity purified proteins were subjected to immunoblotting using anti-HA-HRP 824

antibodies. The lower right panel was exposed for longer than the lower left panel. The 825

red arrow indicates the position of LEC2-HA protein. Higher MW forms of LEC2-HA 826

detected in the FLAG TR-TUBE pull down were dependent on the activity of UPL3-827

3FLAG. 828

829

Figure 5. Variation in BnaUPL3.C03 promoter activities from high- and low- SWPP B. 830

napus accessions is sufficient to generate variation in final seed yield. 831

A. The diagram shows fusions of 2 kb BnaUPL3.C03 promoter and 5’UTR regions from the832

B. napus lines Dimension (DIM) (High SWPP) and Coriander (COR) (Low SWPP) to the833

Luciferase coding sequence. The diagrams are not to scale. 834

B. Differential expression of the Luciferase reporter gene by the DIM and COR promoter835

and 5’UTR region. The DIM and COR promoters described in panel A fused to a Firefly836

Luciferase reporter gene were transfected into Arabidopsis upl3-4 mutant protoplasts.837

The activities of each promoter are shown relative to co-expressed 35S:Renilla838

luciferase. Data are given as means + SE (n=3). P values were determined by Student’s839

t-test.840

C. Differential expression of the Arabidopsis UPL3 coding region by the COR and DIM841

promoter and 5’UTR regions. BnaUPL3.C03 promoter and 5’UTR regions from the B.842

napus lines Dimension (DIM) and Coriander (COR) were used to express the coding843

region of Arabidopsis UPL3 fused to 3HA at its amino-terminus in upl3-4 mutant844

Arabidopsis. RT-qPCR of UPL3 expression in leaves of wild-type and transgenic845

Arabidopsis lines were measured and are shown relative to the AtEIF1ALPHA gene.846

Data are given as means + SE (at least 3 biological replicates of three independent847

transformants). P values were determined by Student’s t-test.848

D. The COR and DIM promoters show differential complementation of Arabidopsis upl3-4849

seed size. Seed area was quantified in WT, upl3-4 mutant, COR:3HA-AtUPL3 and850

DIM:3HA-AtUPL3 independent transgenic lines. Data shown are means + SE based on851

seed area measurements across at least 100 seeds per genotype and with at least 3852

biological replicates of three independent transformants. P values were determined by853

Student’s t-test. The lower panels are immunoblots showing 3HA-UPL3 protein levels in854

10-15 DPA seeds and tubulin levels for comparison.855

856

29

Figure 6. Relationships between BnaUPL3.C03 expression levels in high- and low- SWPP 857

B. napus accessions to seed size, seed LEC2 protein levels, and seed lipid content. 858

A. Comparison of BnaUPL3.C03 expression levels in 45 DPA seeds in B. napus Dimension 859

(DIM) with high SWPP, and Coriander (COR) accessions with low SWPP, measured by 860

RT-qPCR. Expression levels are relative to the BnaACTIN2 gene. Data are presented as 861

means + SE where n=3 for each genotype. P values were determined by Student’s t-862

test. 863

B. Seed sizes in the low-expressing BnaUPL3.C03 line Dimension and the high expressing 864

BnaUPL3.C03 line Coriander. 865

C. Thousand seed weights of low-expressing BnaUPL3.C03 line Dimension and the high- 866

expressing BnaUPL3.C03 line Coriander. Data are shown as means + SE. Seeds were 867

weighed in batches of 100 seed and thousand seed weight calculated based on these 868

values. n=3 batches for each genotype assayed. P values were determined by Student’s 869

t-test. 870

D. LEC2 protein levels in 45 DPA seeds were detectable in three low BnaUPL3.C03 871

expressing lines, and undetectable in three high BnaUPL3.C03 expressing accessions 872

(Supplemental Table 2). Immunoblots of seed protein extract were probed with anti-873

LEC2 (top panel) and with anti-tubulin (lower panel) as a protein loading control. 874

E. Elevated expression of the LEC2-regulated gene BnaOLE1 in the low expressing 875

BnaUPL3.C03 (Dimension) line. BnaOLE1 expression was quantified by RT-qPCR in 876

low expressing BnaUPL3.C03 (Dimension) and high expressing BnaUPL3.C03 877

(Coriander) accessions. Expression levels were relative to that of BnaACTIN2. n=3 for 878

each genotype. Primers were designed to measure expression of both BnaC01g17050D 879

and BnaA01g14480D BnaOLE1. P values were determined by Student’s t-test. 880

F. Increased seed lipid content in the low expressing BnaUPL3.C03 (Dimension) line. Lipid 881

content of mature seeds of low expressing BnaUPL3.C03 (Dimension) and high 882

expressing BnaUPL3.C03 (Coriander) accessions. P values were determined by 883

Student’s t-test. 884

885

Figure 7. Selection for low BnaUPL3.C03 expression levels has not yet been exploited in 886

elite breeding lines. 887

BnaUPL3.C03 expression levels in 45 DPA seeds of seven current elite commercial cultivars of 888

oilseed rape were measured using RT-qPCR and compared with expression levels of 889

BnaUPL3.C03 from GWAS accessions Samurai and Licrown x Express, which exhibit high and 890

30

low BnaUPL3.C03 expression levels, and low and high SWPP phenotypes, respectively. 891

BnaACTIN2 expression was used for comparison using RT-qPCR control. Data are presented 892

as means +SE (n=3 for each genotype). P values estimated by Student’s t-test show the 893

significance of BnaUPL3.C03 expression levels compared to the low-expressing accession 894

Licrown x Express. 895

896

897

898

REFERENCES 899

Andrianov, V., Borisjuk, N., Pogrebnyak, N., Brinker, A., Dixon, J., Spitsin, S., 900

Flynn, J., Matyszczuk, P., Andryszak, K., Laurelli, M., Golovkin, M., and 901

Koprowski, H. (2010). Tobacco as a production platform for biofuel: 902

overexpression of Arabidopsis DGAT and LEC2 genes increases accumulation and 903

shifts the composition of lipids in green biomass. Plant Biotechnol. J. 8: 277–287. 904

Badouin, H. et al. (2017). The sunflower genome provides insights into oil metabolism, 905

flowering and Asterid evolution. Nature 546: 148–152. 906

Bai, X., Huang, Y., Hu, Y., Liu, H., Zhang, B., Smaczniak, C., Hu, G., Han, Z., and 907

Xing, Y. (2017). Duplication of an upstream silencer of FZP increases grain yield in 908

rice. Nat Plants 3: 885–893. 909

Barthole, G., To, A., Marchive, C., Brunaud, V., Soubigou-Taconnat, L., Berger, N., 910

Dubreucq, B., Lepiniec, L., and Baud, S. (2014). MYB118 represses endosperm 911

maturation in seeds of Arabidopsis. Plant Cell 26: 3519–3537. 912

Baud, S. et al. (2016). Deciphering the Molecular Mechanisms Underpinning the 913

Transcriptional Control of Gene Expression by Master Transcriptional Regulators in 914

Arabidopsis Seed. Plant Physiol. 171: 1099–1112. 915

Baud, S., Wuillème, S., To, A., Rochat, C., and Lepiniec, L. (2009). Role of 916

WRINKLED1 in the transcriptional regulation of glycolytic and fatty acid biosynthetic 917

genes in Arabidopsis. Plant J. 60: 933–947. 918

Bevan, M.W., Uauy, C., Wulff, B.B.H., Zhou, J., Krasileva, K., and Clark, M.D. 919

(2017). Genomic innovation for crop improvement. Nature 543: 346–354. 920

Boulard, C., Fatihi, A., Lepiniec, L., and Dubreucq, B. (2017). Regulation and 921

evolution of the interaction of the seed B3 transcription factors with NF-Y subunits. 922

Biochim. Biophys. Acta 1860: 1069–1078. 923

Chan, A., Carianopol, C., Tsai, A.Y.-L., Varathanajah, K., Chiu, R.S., and 924

Gazzarrini, S. (2017). SnRK1 phosphorylation of FUSCA3 positively regulates 925

31

embryogenesis, seed yield, and plant growth at high temperature in Arabidopsis. J. 926

Exp. Bot. 68: 4219–4231. 927

Clough, S.J. and Bent, A.F. (1998). Floral dip: a simplified method for Agrobacterium-928

mediated transformation of Arabidopsis thaliana. Plant J. 16: 735–743. 929

Dong, H., Dumenil, J., Lu, F-H., Na, L.,Vanhaeren, H., Naumann, C., Klecker, M., 930

Prior, R., Smith, C., McKenzie, N., Saalbach, G., Chen, Y., Xia., T., Gonzalez, 931

N., Seguela, M., Inze, D., Dissmeyer, N., Li, Y., and Bevan, MW. (2017). 932

Ubiquitylation activates a peptidase that promoted cleavage and destablization of 933

its activating E3 ligases and diverse growth regulatory proteins to limit cell 934

proliferation in Arabidopsis. Genes and Development 31: 197-208. 935

Downes, B.P., Stupar, R.M., Gingerich, D.J., and Vierstra, R.D. (2003). The HECT 936

ubiquitin-protein ligase (UPL) family in Arabidopsis: UPL3 has a specific role in 937

trichome development. Plant J. 35: 729–742. 938

Duan, P., Xu, J., Zeng, D., Zhang, B., Geng, M., Zhang, G., Huang, K., Huang, L., 939

Xu, R., Ge, S., Qian, Q., and Li, Y. (2017). Natural Variation in the Promoter of 940

GSE5 Contributes to Grain Size Diversity in Rice. Mol. Plant 10: 685–694. 941

Duong, S., Vonapartis, E., Li, C.-Y., Patel, S., and Gazzarrini, S. (2017). The E3 942

ligase ABI3-INTERACTING PROTEIN2 negatively regulates FUSCA3 and plays a 943

role in cotyledon development in Arabidopsis thaliana. J. Exp. Bot. 68: 1555–1567. 944

El Refy, A., Perazza, D., Zekraoui, L., Valay, J.-G., Bechtold, N., Brown, S., 945

Hülskamp, M., Herzog, M., and Bonneville, J.-M. (2003). The Arabidopsis 946

KAKTUS gene encodes a HECT protein and controls the number of 947

endoreduplication cycles. Mol. Genet. Genomics 270: 403–414. 948

van Erp, H., Kelly, A.A., Menard, G., and Eastmond, P.J. (2014). Multigene 949

engineering of triacylglycerol metabolism boosts seed oil content in Arabidopsis. 950

Plant Physiol. 165: 30–36. 951

Fatihi, A., Boulard, C., Bouyer, D., Baud, S., Dubreucq, B., and Lepiniec, L. (2016). 952

Deciphering and modifying LAFL transcriptional regulatory network in seed for 953

improving yield and quality of storage compounds. Plant Sci. 250: 198–204. 954

Harper, A.L., Trick, M., Higgins, J., Fraser, F., Clissold, L., Wells, R., Hattori, C., 955

Werner, P., and Bancroft, I. (2012). Associative transcriptomics of traits in the 956

polyploid crop species Brassica napus. Nat. Biotechnol. 30: 798–802. 957

Kanai, M., Mano, S., Kondo, M., Hayashi, M., and Nishimura, M. (2016). Extension of 958

oil biosynthesis during the mid-phase of seed development enhances oil content in 959

Arabidopsis seeds. Plant Biotechnol. J. 14: 1241–1250. 960

Lin, Q., Ohashi, Y., Kato, M., Tsuge, T., Gu, H., Qu, L.-J., and Aoyama, T. (2015). 961

GLABRA2 Directly Suppresses Basic Helix-Loop-Helix Transcription Factor Genes 962

32

with Diverse Functions in Root Hair Development. Plant Cell 27: 2894–2906. 963

Liu, J. et al. (2017). GW5 acts in the brassinosteroid signalling pathway to regulate 964

grain width and weight in rice. Nat Plants 3: 17043. 965

Liu, Y.-F. et al. (2014). Soybean GmMYB73 promotes lipid accumulation in transgenic 966

plants. BMC Plant Biol. 14: 73. 967

Li, Y., Beisson, F., Pollard, M., and Ohlrogge, J. (2006). Oil content of Arabidopsis 968

seeds: the influence of seed anatomy, light and plant-to-plant variation. 969

Phytochemistry 67: 904–915. 970

Lu, K. et al. (2017). Genome-Wide Association and Transcriptome Analyses Reveal 971

Candidate Genes Underlying Yield-determining Traits in Brassica napus. Front. 972

Plant Sci. 8: 206. 973

Manan, S., Ahmad, M.Z., Zhang, G., Chen, B., Haq, B.U., Yang, J., and Zhao, J. 974

(2017). Soybean LEC2 Regulates Subsets of Genes Involved in Controlling the 975

Biosynthesis and Catabolism of Seed Storage Substances and Seed Development. 976

Front. Plant Sci. 8: 1604. 977

Mason, A.S., Higgins, E.E., Snowdon, R.J., Batley, J., Stein, A., Werner, C., and 978

Parkin, I.A.P. (2017). A user guide to the Brassica 60K Illumina InfiniumTM SNP 979

genotyping array. Theor. Appl. Genet. 130: 621–633. 980

Maspero, E., Valentini, E., Mari, S., Cecatiello, V., Soffientini, P., Pasqualato, S., 981

and Polo, S. (2013). Structure of a ubiquitin-loaded HECT ligase reveals the 982

molecular basis for catalytic priming. Nat. Struct. Mol. Biol. 20: 696–701. 983

McFarlane, H., Gendre, D., and Western, T. (2014). Seed Coat Ruthenium Red 984

Staining Assay. BIO-PROTOCOL 4. 985

Napier, J.A. and Graham, I.A. (2010). Tailoring plant lipid composition: designer 986

oilseeds come of age. Curr. Opin. Plant Biol. 13: 330–337. 987

Nodine, M.D. and Bartel, D.P. (2010). MicroRNAs prevent precocious gene expression 988

and enable pattern formation during plant embryogenesis. Genes Dev. 24: 2678–989

2692. 990

Ohto, M.-A., Floyd, S.K., Fischer, R.L., Goldberg, R.B., and Harada, J.J. (2009). 991

Effects of APETALA2 on embryo, endosperm, and seed coat development 992

determine seed size in Arabidopsis. Sex. Plant Reprod. 22: 277–289. 993

Patra, B., Pattanaik, S., and Yuan, L. (2013). Ubiquitin protein ligase 3 mediates the 994

proteasomal degradation of GLABROUS 3 and ENHANCER OF GLABROUS 3, 995

regulators of trichome development and flavonoid biosynthesis in Arabidopsis. 996

Plant J. 74: 435–447. 997

33

Pfaffl, M.W. (2001). A new mathematical model for relative quantification in real-time 998

RT–PCR. Nucleic Acids Res. 29: e45–e45. 999

Pritchard, J.K., Stephens, M., and Donnelly, P. (2000). Inference of population 1000

structure using multilocus genotype data. Genetics 155: 945–959. 1001

Ray, D.K., Mueller, N.D., West, P.C., and Foley, J.A. (2013). Yield Trends Are 1002

Insufficient to Double Global Crop Production by 2050. PLoS One 8: e66428. 1003

Rodgers-Melnick, E., Vera, D.L., Bass, H.W., and Buckler, E.S. (2016). Open 1004

chromatin reveals the functional maize genome. Proc. Natl. Acad. Sci. U. S. A. 113: 1005

E3177–84. 1006

Roscoe, T.T., Guilleminot, J., Bessoule, J.-J., Berger, F., and Devic, M. (2015). 1007

Complementation of Seed Maturation Phenotypes by Ectopic Expression of 1008

ABSCISIC ACID INSENSITIVE3, FUSCA3 and LEAFY COTYLEDON2 in 1009

Arabidopsis. Plant Cell Physiol. 56: 1215–1228. 1010

Santos-Mendoza, M., Dubreucq, B., Baud, S., Parcy, F., Caboche, M., and 1011

Lepiniec, L. (2008a). Deciphering gene regulatory networks that control seed 1012

development and maturation in Arabidopsis. Plant J. 54: 608–620. 1013

Santos-Mendoza, M., Dubreucq, B., Baud, S., Parcy, F., Caboche, M., and 1014

Lepiniec, L. (2008b). Deciphering gene regulatory networks that control seed 1015

development and maturation in Arabidopsis. Plant J. 54: 608–620. 1016

Santos Mendoza, M., Dubreucq, B., Miquel, M., Caboche, M., and Lepiniec, L. 1017

(2005). LEAFY COTYLEDON 2 activation is sufficient to trigger the accumulation of 1018

oil and seed specific mRNAs in Arabidopsis leaves. FEBS Lett. 579: 4666–4670. 1019

Schneider, C.A., Rasband, W.S., and Eliceiri, K.W. (2012). NIH Image to ImageJ: 25 1020

years of image analysis. Nat. Methods 9: 671–675. 1021

Shi, L., Katavic, V., Yu, Y., Kunst, L., and Haughn, G. (2012). Arabidopsis glabra2 1022

mutant seeds deficient in mucilage biosynthesis produce more oil. Plant J. 69: 37–1023

46. 1024

Singh, R. et al. (2013). Oil palm genome sequence reveals divergence of interfertile 1025

species in Old and New worlds. Nature 500: 335–339. 1026

Song, S.-K., Ryu, K.H., Kang, Y.H., Song, J.H., Cho, Y.-H., Yoo, S.-D., Schiefelbein, 1027

J., and Lee, M.M. (2011). Cell fate in the Arabidopsis root epidermis is determined 1028

by competition between WEREWOLF and CAPRICE. Plant Physiol. 157: 1196–1029

1208. 1030

Stephenson, P., Baker, D., Girin, T., Perez, A., Amoah, S., King, G.J., and 1031

Østergaard, L. (2010). A rich TILLING resource for studying gene function in 1032

Brassica rapa. BMC Plant Biol. 10: 62. 1033

34

Stone, S.L., Kwong, L.W., Yee, K.M., Pelletier, J., Lepiniec, L., Fischer, R.L., 1034

Goldberg, R.B., and Harada, J.J. (2001). LEAFY COTYLEDON2 encodes a B3 1035

domain transcription factor that induces embryo development. Proc. Natl. Acad. 1036

Sci. U. S. A. 98: 11806–11811. 1037

Tsai, A.Y.-L. and Gazzarrini, S. (2012). AKIN10 and FUSCA3 interact to control lateral 1038

organ development and phase transitions in Arabidopsis. Plant J. 69: 809–821. 1039

Wang, J., Liu, H., and Ren, G. (2014). Near-infrared spectroscopy (NIRS) evaluation 1040

and regional analysis of Chinese faba bean (Vicia faba L.). The Crop Journal 2: 28–1041

37. 1042

Wang, S. et al. (2015). The OsSPL16-GW7 regulatory module determines grain shape 1043

and simultaneously improves rice yield and grain quality. Nat. Genet. 47: 949–954. 1044

Willmann, M.R., Mehalick, A.J., Packer, R.L., and Jenik, P.D. (2011). MicroRNAs 1045

regulate the timing of embryo maturation in Arabidopsis. Plant Physiol. 155: 1871–1046

1884. 1047

Wu, F.-H., Shen, S.-C., Lee, L.-Y., Lee, S.-H., Chan, M.-T., and Lin, C.-S. (2009). 1048

Tape-Arabidopsis Sandwich - a simpler Arabidopsis protoplast isolation method. 1049

Plant Methods 5: 16. 1050

Yau, R. and Rape, M. (2016). The increasing complexity of the ubiquitin code. Nat. Cell 1051

Biol. 18: 579–586. 1052

Zhai, Z., Liu, H., and Shanklin, J. (2017). Phosphorylation of WRINKLED1 by KIN10 1053

Results in Its Proteasomal Degradation, Providing a Link between Energy 1054

Homeostasis and Lipid Biosynthesis. Plant Cell 29: 871–889. 1055

Zhang, X., Garreton, V., and Chua, N.-H. (2005). The AIP2 E3 ligase acts as a novel 1056

negative regulator of ABA signaling by promoting ABI3 degradation. Genes Dev. 1057

19: 1532–1543. 1058

Zhou, Y. et al. (2013). HISTONE DEACETYLASE19 interacts with HSL1 and 1059

participates in the repression of seed maturation genes in Arabidopsis seedlings. 1060

Plant Cell 25: 134–148. 1061

1062

Figure 1. Association of variation in Seed Weight Per Pod (SWPP) with SNPs and differential expression of BnaUPL3.C03 in a panel of 94 Brassica napus accessions.

A. Associations between single-nucleotide polymorphism (SNP) on chromosomes A08/C03 and SWPP. The dotted grey lines outline the genomic location of peaks of SNP associations with SWPP. Markers are plotted in pseudomolecule order and associations as -log10 P values. The colored regions under the pseudochromosome represent the regions of sequence similarity to A. thaliana chromosomes, as described previously (Harper et al. 2012). The dashed horizontal red line indicates the Bonferroni-corrected significance threshold.

B. Normal distribution of the SWPP trait in the set of B. napus accessions, showing the data were suitable for Mixed Linear Model analyses.

C. Segregation of SWPP trait means with the most highly associating marker (JCVI_5587:125) show a marker effect of ~20%. Data are given as mean ± SE. P values were determined by Student’s t-test.

D. Differential expression of a single C genome assigned unigene C_EX097784 on chromosome C03 was associated with SWPP variation. This unigene is an ortholog of Arabidopsis UBIQUITIN PROTEIN LIGASE 3 (UPL3), and is termed BnaUPL3.C03. The association exceeded the adjusted P value calculated by the Benjamini-Hochberg method (Q= 0.026).

E. Correlation of BnaUPL3.C03 expression in seedling leaves measured by RPKM and RT-qPCR in six B. napus GWAS accessions (Supplemental Table 2) exhibiting maximal variation in C-EX097784 expression. Measurements of RT-qPCR of BnaUPL3.C03 expression in seedlings were expressed relative to BnaUBC10 expression levels. Lines with high SWPP are shown by orange data points, and lines with low SWPP are shown by gray data points.

F. Correlation of BnaUPL3.C03 transcript abundance in seedlings, measured using RT-qPCR insix B. napus GWAS accession exhibiting maximal variation in SWPP, with variation in SWPP.Lines with high SWPP are shown by orange data points, and lines with low SWPP are shownby gray data points.

G. A subset of ten GWAS accessions with maximal variation in SWPP (Supplemental Table 2)were grown in replicated plots in field conditions between March-August and mean plot yieldsmeasured after combining. Plot yields are shown as mean ± SE. P values were determinedby Student’s t-test.

Figure 2. A loss-of-function mutation in Arabidopsis UPL3 has increased seed size and altered seed storage and seed coat phenotypes.

A. UPL3 expression increases during seed development in Arabidopsis. AtUPL3 expression,measured by RT-qPCR, during seed development in wt Col. Expression levels are relative toEF1ALPHA expression. DPA is Days Post Anthesis. Data are given as mean + SE of threebiological replicates.

B. BnaUPL3.C03 expression in developing pods of Brassica napus accessions described inFigure 1E was correlated with expression in leaves. BnaACTIN2 was used as an internal RT-qPCR control. RT-qPCR analyses used three biological replicates. Lines with high SWPP arerepresented by orange data points, and lines with low SWPP are shown as gray data points.

C. Loss-of-function mutants of AtUPL3 have enlarged seeds, while AtUPL3 expression from the35S promoter in Col 0 plants reduced seed size. Areas of Arabidopsis seeds from Col 0, arepresentative transformant, and two independent upl3 T-DNA loss-of-function mutants. Dataare given as mean + SE (means calculated across 6 biological replicates per genotype using100 seeds per replicate). P values were determined by Student’s t-test.

D. Loss-of-function mutants of AtUPL3 has elevated seed lipid content, while AtUPL3 expressionfrom the 35S promoter in Col 0 plants reduced lipid content. Lipid content of Arabidopsisseeds from Col 0, two upl3 T-DNA mutants, and a representative 35S:3HA-UPL3transformant was measured using Near Infrared Spectroscopy. Data are given as mean + SE(means calculated across 6 biological replicates per genotype using 100 seeds per replicate).P values were determined by Student’s t-test.

E. Altered mucilage extrusion of imbibed Arabidopsis seeds in the upl3-4 mutant, visualized byRuthenium Red staining.

F. Altered expression of the regulatory transcription factor GL2 during early stages of seeddevelopment in Atupl3-4 mutant seeds. RNA from whole siliques harvested at 5-10 DPA wasused. Expression levels are relative to EF1ALPHA expression. Data are given as means + SE(n=3 biological replicates). P values were determined by Student’s t-test.

G. Altered expression of the regulatory transcription factor MUM4 during early stages of seeddevelopment in upl3-4 mutant seeds. RNA from whole siliques harvested at 5-10 DPA wasused. Expression levels are relative to EF1ALPHA expression. Data are given as means + SE(n=3 biological replicates). P values were determined by Student’s t-test.

Figure 3. UPL3 reduces LEC2-mediated gene expression. Altered expression of Arabidopsis seed maturation genes in the upl3-4 mutant. RT-qPCR was used to measure gene expression during seed development in Col 0 and upl3-4. RNA from whole siliques was used, harvested at the times indicated on the x axis. Expression levels are relative to EF1ALPHA gene expression. Data are given as means + SE (n=3 biological replicates). P values were determined by Student’s t-test. A, AP2 expression; B, LEC2 expression; C, LEC1 expression; D, MYB118 expression; E, WRI1 expression; F. RT-qPCR measurement of OLE1 in 10 DPA Col 0 and upl3-4 mutant Arabidopsis plants. RNA from whole siliques was used. Expression levels are relative to EF1ALPHA gene expression. Data are given as means + SE (n=3 biological replicates). P values were determined by Student’s t-test. G. AtUPL3 reduces LEC2-mediated activation of the OLE1 promoter. Transient expression of theLEC2- regulated OLE1 promoter:Firefly Luciferase (fLUC) reporter gene in Arabidopsis upl3-4 mutantleaf protoplasts. 35S:Renilla LUC vector was co-transfected in all treatments as a control, and theratio of Firefly/Renilla LUC activity was used to determine OLE1:fLUC gene expression levels.35S:3HA-LEC2 and 35S:3FLAG-UPL3 were co-transfected as shown. Data are presented as means

+ SE. (n=3 independent transfections). P values were determined by Student’s t-test.

Figure 4. UPL3 mediates the proteasomal degradation of LEC2-HIS and mediates the formation of higher MW forms of LEC2 in plants.

A. Cell-free degradation of LEC2-HIS protein. Purified LEC2-HIS expressed in E. coli was

incubated at 22oC for the times indicated in total protein extracts from upl3-4 mutant seedlings, with either wt- or mutant purified UPL3-3FLAG added (see side panel), with and without 50 μM MG132. Immunoblots of reactions were probed with anti-HIS antibodies. CBB shows a portion of the reaction stained with Colloidal Coomassie Blue as a loading control. The graph shows results of four independent cell-free reactions using the same batch of purified LEC2-HIS protein. P values were determined by Student’s t test.

B. Immunoblots of protein samples from leaf or 10-15 DPA siliques of Col 0, lec2 mutant, upl3-4mutant and 35S:3HA-UPL3 plants electrophoresed on SDS-PAGE gels and probed with anti-LEC2, anti-HA or anti-tubulin antibodies as a loading control. TUBULIN levels are to compareprotein loading.

C. Expression of 3HA-UPL3 increases higher MW forms of LEC2 in developing siliques.Immunoblots of protein samples from pooled 5-10 DPA siliques of Col 0 or 35S:3HA-UPL3plants were electrophoresed on 4-20% SDS-PAGE gels and probed with anti-LEC2. HigherMW forms of LEC2 protein seen in the 35S:3HA-UPL3 sample are indicated by arrows. Thelower panel shows that both wild-type and mutant UPL3-3FLAG interact with LEC2-HA duringtransient co-expression in N. benthamiana leaves.

D. Purification of ubiquitylated forms of LEC2-HA from transiently expressed 35S:LEC2-HA and35S:UPL3-3FLAG (wild type and mutant) vectors in Nicotiana benthamiana leaves. Proteinsamples were extracted and samples taken to assess protein expression levels usingimmunoblotting (loading panels). The remaining samples were split in two and approximately20 µg FLAG-TR-TUBE was added to one set (lower loading panel). This fraction was purifiedon FLAG-MA beads and the other purified using HA-MA beads. Affinity purified proteins weresubjected to immunoblotting using anti-HA-HRP antibodies. The lower right panel wasexposed for longer than the lower left panel. The red arrow indicates the position of LEC2-HAprotein. Higher MW forms of LEC2-HA detected in the FLAG TR-TUBE pull down weredependent on the activity of UPL3-3FLAG.

Figure 5. Variation in BnaUPL3.C03 promoter activities from high- and low- SWPP B. napus accessions is sufficient to generate variation in final seed yield.

A. The diagram shows fusions of 2 kb BnaUPL3.C03 promoter and 5’UTR regions from the B.napus lines Dimension (DIM) (High SWPP) and Coriander (COR) (Low SWPP) to theLuciferase coding sequence. The diagrams are not to scale.

B. Differential expression of the Luciferase reporter gene by the DIM and COR promoter and5’UTR region. The DIM and COR promoters described in panel A fused to a Firefly Luciferasereporter gene were transfected into Arabidopsis upl3-4 mutant protoplasts. The activities ofeach promoter are shown relative to co-expressed 35S:Renilla luciferase. Data are given asmeans + SE (n=3). P values were determined by Student’s t-test.

C. Differential expression of the Arabidopsis UPL3 coding region by the COR and DIM promoterand 5’UTR regions. BnaUPL3.C03 promoter and 5’UTR regions from the B. napus linesDimension (DIM) and Coriander (COR) were used to express the coding region ofArabidopsis UPL3 fused to 3HA at its amino-terminus in upl3-4 mutant Arabidopsis. RT-qPCRof UPL3 expression in leaves of wild-type and transgenic Arabidopsis lines were measuredand are shown relative to the AtEIF1ALPHA gene. Data are given as means + SE (at least 3biological replicates of three independent transformants). P values were determined byStudent’s t-test.

D. The COR and DIM promoters show differential complementation of Arabidopsis upl3-4 seedsize. Seed area was quantified in WT, upl3-4 mutant, COR:3HA-AtUPL3 and DIM:3HA-AtUPL3 independent transgenic lines. Data shown are means + SE based on seed areameasurements across at least 100 seeds per genotype and with at least 3 biologicalreplicates of three independent transformants. P values were determined by Student’s t-test.The lower panels are immunoblots showing 3HA-UPL3 protein levels in 10-15 DPA seedsand tubulin levels for comparison.

Figure 6. Relationships between BnaUPL3.C03 expression levels in high- and low- SWPP B. napus accessions to seed size, seed LEC2 protein levels, and seed lipid content.

A. Comparison of BnaUPL3.C03 expression levels in 45 DPA seeds in B. napus Dimension (DIM) with high SWPP, and Coriander (COR) accessions with low SWPP, measured by RT-qPCR. Expression levels are relative to the BnaACTIN2 gene. Data are presented as means + SE where n=3 for each genotype. P values were determined by Student’s t-test.

B. Seed sizes in the low-expressing BnaUPL3.C03 line Dimension and the high expressing BnaUPL3.C03 line Coriander.

C. Thousand seed weights of low-expressing BnaUPL3.C03 line Dimension and the high- expressing BnaUPL3.C03 line Coriander. Data are shown as means + SE. Seeds were weighed in batches of 100 seed and thousand seed weight calculated based on these values. n=3 batches for each genotype assayed. P values were determined by Student’s t-test.

D. LEC2 protein levels in 45 DPA seeds were detectable in three low BnaUPL3.C03 expressing lines, and undetectable in three high BnaUPL3.C03 expressing accessions (Supplemental Table 2). Immunoblots of seed protein extract were probed with anti-LEC2 (top panel) and with anti-tubulin (lower panel) as a protein loading control.

E. Elevated expression of the LEC2-regulated gene BnaOLE1 in the low expressing BnaUPL3.C03 (Dimension) line. BnaOLE1 expression was quantified by RT-qPCR in low expressing BnaUPL3.C03 (Dimension) and high expressing BnaUPL3.C03 (Coriander) accessions. Expression levels were relative to that of BnaACTIN2. n=3 for each genotype. Primers were designed to measure expression of both BnaC01g17050D and BnaA01g14480D BnaOLE1. P values were determined by Student’s t-test.

F. Increased seed lipid content in the low expressing BnaUPL3.C03 (Dimension) line. Lipid content of mature seeds of low expressing BnaUPL3.C03 (Dimension) and high expressing BnaUPL3.C03 (Coriander) accessions. P values were determined by Student’s t-test.

Figure 7. Selection for low BnaUPL3.C03 expression levels has not yet been exploited in elite breeding lines. BnaUPL3.C03 expression levels in 45 DPA seeds of seven current elite commercial cultivars of oilseed rape were measured using RT-qPCR and compared with expression levels of BnaUPL3.C03 from GWAS accessions Samurai and Licrown x Express, which exhibit high and low BnaUPL3.C03 expression levels, and low and high SWPP phenotypes, respectively. BnaACTIN2 expression was used for comparison using RT-qPCR control. Data are presented as means +SE (n=3 for each genotype). P values estimated by Student’s t-test show the significance of BnaUPL3.C03 expression levels compared to the low-expressing accession Licrown x Express.

DOI 10.1105/tpc.18.00577; originally published online August 22, 2019;Plant Cell

Dubreucq, Thierry Chardot, Loic Lepiniec and Michael W BevanCharlotte Miller, Rachel Wells, Neil McKenzie, Martin Trick, Joshua Ball, Abdelhak Fatihi, Bertrand

yield in Brassica napus.Variation in expression of the HECT E3 ligase UPL3 modulates LEC2 levels, seed size and crop

 This information is current as of July 2, 2020

 

Supplemental Data /content/suppl/2019/08/22/tpc.18.00577.DC1.html

Permissions https://www.copyright.com/ccc/openurl.do?sid=pd_hw1532298X&issn=1532298X&WT.mc_id=pd_hw1532298X

eTOCs http://www.plantcell.org/cgi/alerts/ctmain

Sign up for eTOCs at:

CiteTrack Alerts http://www.plantcell.org/cgi/alerts/ctmain

Sign up for CiteTrack Alerts at:

Subscription Information http://www.aspb.org/publications/subscriptions.cfm

is available at:Plant Physiology and The Plant CellSubscription Information for

ADVANCING THE SCIENCE OF PLANT BIOLOGY © American Society of Plant Biologists