Variation in Expression of the HECT E3 Ligase UPL3Modulates LEC2 Levels, Seed Size, and Crop Yields inBrassica napus[OPEN]

CharlotteMiller,a RachelWells,a NeilMcKenzie,aMartin Trick,a JoshuaBall,a AbdelhakFatihi,b BertrandDubreucq,b

Thierry Chardot,b Loic Lepiniec,b and Michael W. Bevana,1

a Department of Cell and Developmental Biology, John Innes Centre, Norwich NR4 7UH, United Kingdomb Institut Jean-Pierre Bourgin, Institut National de la Recherche Agronomique, AgroParisTech, Centre National de la RechercheScientifique, Université Paris-Saclay, Institut National de la Recherche Agronomique Versailles, route de Saint-Cyr, 78000 Versailles,France

ORCID IDs: 0000-0003-2228-1619 (C.M.); 0000-0002-1280-7472 (R.W.); 0000-0001-9818-9216 (N.M.); 0000-0001-8786-5012 (M.T.);0000-0003-4840-3768 (J.B.); 0000-0002-8216-6367 (A.F.); 0000-0003-2661-5479 (B.D.); 0000-0001-6689-6151 (T.C.); 0000-0002-5845-3323 (L.L.); 0000-0001-8264-2354 (M.W.B.).

Identifying genetic variation that increases crop yields is a primary objective in plant breeding. We used association analysesof oilseed rape/canola (Brassica napus) accessions to identify genetic variation that influences seed size, lipid content, andfinal crop yield. Variation in the promoter region of the HECT E3 ligase gene BnaUPL3.C03 made a major contribution tovariation in seed weight per pod, with accessions exhibiting high seed weight per pod having lower levels of BnaUPL3.C03expression. We defined a mechanism in which UPL3 mediated the proteasomal degradation of LEC2, a master transcriptionalregulator of seed maturation. Accessions with reduced UPL3 expression had increased LEC2 protein levels, larger seeds, andprolonged expression of lipid biosynthetic genes during seed maturation. Natural variation in BnaUPL3.C03 expressionappears not to have been exploited in current B. napus breeding lines and could therefore be used as a new approach tomaximize future yields in this important oil crop.

INTRODUCTION

In major oil-producing crops, such as oilseed rape (Brassica na-pus), the composition of seed storage lipids has been optimizedfor different end uses, from human nutrition to industrial appli-cations,by identifyingallelic variation inbiosynthetic enzymesandpathways (Napier and Graham, 2010). For example, elite oilseedrape varieties now have greatly enhanced nutritional value, withhigh linoleic acid, reduced erucic acid, and optimal linoleic tolinolenic acid ratios. However, increasing overall production ofstorage lipids tomeet projected future demands for both food andindustrial uses remains a key objective for achieving food securityand sustainable industrial production.

An interacting network of transcription factors establishesand maintains Arabidopsis (Arabidopsis thaliana) embryo de-velopment and promotes the accumulation of seed storage lipidsand proteins (Fatihi et al., 2016; Boulard et al., 2017). Loss-of-function mutations in four conserved regulatory genes leadto curtailed seed maturation, loss of dormancy, and ectopicvegetative development. FUSCA3 (FUS3), ABSCISIC ACIDINSENSITIVE3 (ABI3), and LEAFY COTYLEDON2 (LEC2) encodeAFL-B3-family transcription factors, while LEC1 encodes an

NFY-Y CCAAT binding transcription factor. These “LAFL” genes(Santos-Mendoza et al., 2008; Roscoe et al., 2015) induce seedmaturation and inhibit germination, and their expression is down-regulated at the initiation of seed dormancy and desiccationtolerance. LEC2 and FUS3 expression is repressed by miRNA–mediated mechanisms during early embryogenesis (Nodine andBartel, 2010; Willmann et al., 2011) to ensure the correct timing ofstorage reserve accumulation. At later stages of seed de-velopment, the B3-domain protein VAL3 recruits HISTONEDEACETYLASE19 to the promoters of LAFL genes and silencestheir expression by altering levels of histone methylation andacetylation (Zhou et al., 2013). The stability of LAFL proteins isalso controlled during seed maturation and dormancy. ABI3-INTERACTING PROTEIN2 is an E3 ligase that ubiquitylates bothABI3 (Zhang et al., 2005) andFUS3 (Duong et al., 2017), suggestingthat regulation of LAFL protein levels has an important role inseed maturation and the transition to dormancy. The SNF kinaseAKIN10 phosphorylates and stabilizes FUS3 (Chan et al., 2017;Tsai and Gazzarrini, 2012) and WRINKLED1 (WRI1; Zhai et al.,2017), a transcription factor regulated by LEC2 that promotesexpression of glycolytic and lipid biosynthesis genes. Improvedunderstanding of the control of these important seed maturationregulators may provide new ways to optimize seed compositionand yield.An important strategy in crop improvement aims to identify new

sourcesofgenetic variation fromdiversegermplasmresources forincreasing crop productivity (Bevan et al., 2017). Genome-wideassociation studies (GWAS) are increasingly used for identifyingvariation associatedwith traits in crops and theirwild relatives. For

1 Address correspondence to: michael.bevan@jic.ac.uk.The author responsible for distribution of materials integral to the findingspresented in this article in accordance with the policy described inthe Instructions for Authors (www.plantcell.org) is: Michael W. Bevan(michael.bevan@jic.ac.uk).[OPEN]Articles can be viewed without a subscription.www.plantcell.org/cgi/doi/10.1105/tpc.18.00577

The Plant Cell, Vol. 31: 2370–2385, October 2019, www.plantcell.org ã 2019 ASPB.

example, associations among sequence variation, gene expres-sion levels, andoil composition have been used to identify geneticvariation in knowngenes conferring oil quality traits in oilseed rape(Harper et al., 2012; Lu et al., 2017). GWAS also has potentialfor the discovery of new gene functions and, when utilized fully,can lead to a deeper understanding of mechanisms underlyingcomplex traits such as crop yield.

Here we use associative transcriptomics in B. napus to identifygenetic variation in the regulation of BnaUPL3.C03, encoding anortholog of the HECT E3 ubiquitin ligase UPL3 (Downes et al.,2003; ElRefy et al., 2003),which is associatedwith increasedseedsize and field yields. We establish a mechanism in which reducedexpression of BnaUPL3.C03 maintains higher levels of LEC2protein during seedmaturation by reducedUPL3-mediated LEC2ubiquitylation, leading to increased seed lipid levels and overallincreased seedyields. Analysis of elite oilseed rape varieties showsthat variation in the expression of BnaUPL3.C03 has not yet beenexploited in breeding programs and thus can be used to increasecrop yields.

RESULTS

Associative Transcriptomics Identifies a NegativeCorrelation between BnaUPL3.C03 Expression and SeedWeight per Pod in B. napus

A panel of 94 B. napus oilseed rape accessions with high geneticdiversity (Supplemental Table 1), for which leaf transcriptomedata from each accession has been mapped to a sequencedreference genome (Harper et al., 2012), was screened for yield-related phenotypic variation. High levels of trait variation wereobserved. Trait associations with both sequence variation, inthe form of hemi-single nucleotide polymorphisms (SNPs) in thepolyploid genome of B. napus, and gene expression levels, as-sessed as Reads Per Kilobase of unigene per Million (RPKM)mapped reads in a gene expression marker (GEM) analysis, werethen calculated. This identified significant associations betweenvariation in seed weight per pod (SWPP) and SNP variation inhomeologous regions of linkage groups A08 and C03 (Figure 1A;Supplemental Figure 1). SWPP phenotype data displayeda normal distribution appropriate for mixed linear model (MLM)analyses (Figure 1B).

Assessment of phenotypic variation segregating with alleles forthe most significant SNP marker, JCVI_5587:125, revealed thataccessions inheriting “T” at this locus had SWPP values ;20%lower than those accessions inheriting the hemi-SNP genotype,“Y” (C/T; Figure 1C). Figure 1A shows that variation in the hemi-SNP marker did not distinguish between the A08 and C03chromosomes. However, GEM analyses showed an associationof SWPP with varying expression of a single C genome-assignedunigenewithin this region,C_EX097784 (Figure 1D;SupplementalFigure 1). This unigene corresponded to an ortholog of the Ara-bidopsisUBIQUITINPROTEINLIGASE3 (UPL3) encoding aHECTE3 ligase (Downes et al., 2003; El Refy et al., 2003).

Two UPL3 homologs in B. napus, BnaUPL3.C03 and BnaU-PL3.A08, were identified based on protein sequence similarityand conserved synteny among Arabidopsis, Brassica rapa, and

Brassica oleracea. Supplemental Figure 2 illustrates the high proteinsequence similarity between the single Arabidopsis and the twoB. napus UPL3 orthologs. Associative transcriptomics analysesidentified significant differential expression of the BnaUPL3.C03gene betweenGWASaccessions displaying high variation in SWPP(Figure 1D; Supplemental Figure 1). Gene-specific RT-qPCR anal-yses confirmed this differential expression at the BnaUPL3.C03locus in seedlings of six lines selected for high- or low-SWPP(Figure 1E). Correlating BnaUPL3.C03 expression levels with SWPPrevealed a negative relationship, with accessions displaying lowBnaUPL3.C03 expression exhibiting high SWPP (Figure 1F). RT-qPCR analyses ofBnaUPL3.A08 expression in developing seeds ofsix lineswith high- or low-SWPP showed no significant difference inexpression levels and variation in SWPP (Supplemental Figure 3),whereasBnaUPL3.C03expressionwas lower indevelopingseedsofthesehighSWPPlinesascomparedwith lowSWPPlines (Figures2Band 6A). This is consistent with the absence of an association be-tween transcript levels at the BnaUPL3.A08 locus and SWPP de-termined by associative transcriptomics analyses (SupplementalFigure 1). This indicated that variation in the expressionofBnaUPL3.C03 contributed to variation in SWPP.A subset of 10 GWAS accessions (Supplemental Table 2) ex-

hibiting both differential expression of BnaUPL3.C03 and highvariation in SWPP was grown under field conditions in a replicatedyield trial. Mean plot yields across accessions showed a significantincrease in plot yields of high SWPP accessions (Figure 1G), in-dicating SWPP is an important measure of seed yield under fieldconditions.Previous studies in Arabidopsis, a close relative of B. napus, iden-

tified roles for UPL3 in mediating the proteasomal degradation ofGLABROUS3 (GL3) and ENHANCER OF GLABROUS3 (EGL3), bothknown regulators of trichome morphogenesis. Enhanced GL3/EGL3protein levels in an Arabidopsis loss-of-function upl3 mutant alteredleaf trichomemorphogenesis (Downesetal., 2003;ElRefyetal., 2003;Patraetal.,2013).Assessmentof leafhairsacrossasubsetofB.napusGWAS accessions with maximal variation in BnaUPL3.C03 expres-sion revealed segregation of a trichome phenotype (SupplementalFigure 4), suggesting differential BnaUPL3.C03 expression also in-fluenced trichome morphogenesis in oilseed rape. Although AtUPL3appears to have aconserved role in trichomemorphogenesis, there isno evidence that AtUPL3 has a role in seed development.In Arabidopsis,UPL3 transcript levels increased steadily during

seed development, with highest expression levels observedduring the seed maturation phase (Figure 2A), suggesting a po-tential role for UPL3 in seed maturation. BnaUPL3.C03 was dif-ferentially expressed in seedlings of B. napus accessionsdisplayinghighvariation inSWPP(Figure1F), thereforevariationofBnaUPL3.C03 expression was measured using RT-qPCR in de-veloping pods at 45 days post anthesis (DPA) in six accessionswith low- and high-SWPP. Figure 2B confirms that variation inBnaUPL3.C03 expression in pods was tightly correlated with thatin leaves measured using RNA sequencing.

Arabidopsis and B. rapa Mutants Lacking UPL3 FunctionExhibit Increased Seed Size

Thepotential influenceofAtUPL3 in seed formationwasassessedusing two Arabidopsis T-DNA insertion lines with essentially

Varying UPL3 Expression Modulates Crop Yield 2371

undetectable AtUPL3 expression levels (Supplemental Fig-ure 5). These lines were insertions in the 10th exon and werenamed upl3-4 (SALK_015334) and upl3-5 (SALK_151005) inreference to the upl3 alleles previously identified in Downeset al. (2003). The upl3-4 and upl3-5 mutant seeds were ;10%larger (Figure 2C) and displayed a 12% increase in seed lipidcontent (Figure 2D) relative to seeds of wild-type plants.Conversely, overexpression of 3HA-UPL3 from the 35S pro-moter reduced seed size and lipid content. Analysis of seedlipid composition revealed no significant changes in fatty acidcomposition (Supplemental Table 3). Two premature stop

codons in the single B. rapa UPL3 gene upstream of the pre-dicted catalytic Cys residue (Bra010737.1, SupplementalFigure 6) were identified using a Targeting Induced Local Le-sions IN Genomes (TILLING) resource for B. rapa (Stephensonet al., 2010). Supplemental Figure 7 shows that a line homo-zygous for the JI32517 BraUPL3 mutation also had largerseeds as compared with a line that segregated the BraUPL3mutation. These results established a potential role for theactivities of bothBnaUPL3.C03andBnaUPL3.A08 (descendedfrom BraUPL3) in regulating seed size in Arabidopsis andBrassica species.

Figure 1. Association of Variation in SWPP with SNPs and Differential Expression of BnaUPL3.C03 in a Panel of 94 B. napus Accessions.

(A)AssociationsbetweenSNPsonchromosomesA08/C03andSWPP.Thedottedgray linesoutline thegenomic locationofpeaksofSNPassociationswithSWPP.Markersareplotted inpseudomoleculeorder andassociationsas2log10Pvalues. Thecolored regionsunder thepseudochromosomerepresent theregions of sequence similarity to Arabidopsis chromosomes, as described in 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 MLM analyses.(C)Segregationof SWPP traitmeanswith themost highly associatingmarker (JCVI_5587:125) showamarker effect of;20%.Data are given asmean6 SE.P values were determined by Student’s t test.(D)Differential expression of a single C genome assigned unigeneC_EX097784 on chromosomeC03was associated with SWPP variation. This unigene isan ortholog of Arabidopsis UPL3, and is termed BnaUPL3.C03. The association exceeded the adjusted P value calculated by the Benjamini-Hochbergmethod (Q 5 0.026).(E)Correlation ofBnaUPL3.C03 expression in seedling leavesmeasured byRPKMandRT-qPCR in sixB. napusGWASaccessions (Supplemental Table 2)exhibiting maximal variation in C-EX097784 expression. Measurements of RT-qPCR of BnaUPL3.C03 expression in seedlings were expressed relative toBnaUBC10 expression levels. Lines with high SWPP are shown by orange data points, and lines with low SWPP are shown by gray data points.(F)CorrelationofBnaUPL3.C03 transcript abundance in seedlings,measuredusingRT-qPCR in sixB.napusGWASaccession exhibitingmaximal variationin SWPP, with variation in SWPP. Lines with high SWPP are shown by orange data points, and lines with low SWPP are shown by gray data points.(G) A subset of 10 GWAS accessions with maximal variation in SWPP (Supplemental Table 2) were grown in replicated plots in field conditions betweenMarch and August and mean plot yields measured after combining. Plot yields are shown as mean 6 SE. P values were determined by Student’s t test.

2372 The Plant Cell

Figure 2. A Loss-of-Function Mutation in Arabidopsis UPL3 Has Increased Seed Size and Altered Seed Storage and Seed Coat Phenotypes.

(A)UPL3expression increasesduringseeddevelopment inArabidopsis.AtUPL3expression,measuredbyRT-qPCR,duringseeddevelopment inwild-typeCol-0. Expression levels are relative to EF1ALPHA expression. Data are given as mean 6 SE; n 5 3 biological replicates.

Varying UPL3 Expression Modulates Crop Yield 2373

UPL3 Indirectly Influences Seed Mucilage Biosynthesisvia GL2

Arabidopsis upl3-4 mutant seeds exhibited altered mucilageextrusion upon imbibition (Figure 2E). UPL3-mediates the pro-teasomal degradation of the basic helix-loop-helix transcriptionfactors GL3 and EGL3 (Patra et al., 2013). These proteins functionas part of a complex to regulate the expression of GL2, encodinga homeodomain transcription factor that activates expressionof MUCILAGE-MODIFIED4 (MUM4), encoding a mucilage bio-synthetic enzyme (Shi et al., 2012). This EGL3/GL3 complex ac-tivatesGL2 expression in leaves, but repressesGL2 expression inseeds, depending on the type of MYB transcription factor in thecomplex (Song et al., 2011). The single repeat MYB GmMYB73,a possible homolog of Arabidopsis TRY and CPC, repressedGL2expression in seeds and interacted with EGL3 and GL3 (Liu et al.,2014). Thus, reducedUPL3-mediateddestabilization of EGL3andGL3 in the upl3-4 mutant may elevate GL3 and EGL3 levels,leading to increased repression of GL2 expression. A significantreduction in the expression of both GL2 (Figure 2F) and MUM4(Figure 2G)wasobservedat 5DPA indeveloping seedsof aupl3-4mutant and may explain the altered mucilage extrusion observedin the Arabidopsis upl3-4 mutant siliques.

GL2alsonegatively regulates seed lipid content bysuppressionof PHOSPHOLIPASE D ALPHA1 expression (Liu et al., 2014).However, no significant difference in the expression ofPLDɑ1wasobserved in wild-type and upl3 mutant seeds. Therefore, UPL3may target other proteins for degradation during seed maturationthat influence seed size and storage reserve accumulation.

upl3 Mutants Display Increased Expression of SeedMaturation Genes

Several genes influence both seed lipid content and seed size inArabidopsis, including APETALA2 (AP2; Ohto et al., 2009), LEC1,and LEC2 (SantosMendoza et al., 2005). The expression of thesegenes was assessed during the development of Arabidopsisupl3-4 andwild-type seeds using RT-qPCR. No differences in the

expression of AP2 (Figure 3A) or LEC2 (Figure 3B) were seenbetween upl3-4 and wild-type seeds. However, significant in-creases in LEC1 expression were observed in upl3-4 mutantseeds at 10 DPA as compared with the wild type (Figure 3C).Transcription of LEC1 is positively regulated by LEC2 (Santos-Mendoza et al., 2008; Baud et al., 2016). Given the observedincrease in LEC1, but not LEC2, expression, it was possible thataltered UPL3 expression may affect LEC2 protein levels, thusaltering LEC1 expression in upl3-4mutant seeds. This was testedby measuring expression of WRI1 and MYB118, which are alsoregulated by LEC2 (Barthole et al., 2014; Baud et al., 2009). In-creased expression of both geneswas observed in upl3-4mutantsiliques relative to the wild type from 10 DPA (Figures 3D and 3E),supporting the hypothesis that AtUPL3 may influence LEC2 proteinlevels and expression of target genes.

UPL3 Reduces LEC2-Mediated Transcription of SeedMaturation Genes

LEC1andLEC2bind to thepromoters andactivate the expressionof seed maturation genes, such as OLEOSIN1 (OLE1), a generequired for seed lipid accumulation (Santos-Mendoza et al.,2008b; Baud et al., 2016). To further assess the potential role ofAtUPL3 in LEC2-mediated gene expression, expression of OLE1in maturing seeds was measured in wild-type and upl3-4 mutantArabidopsis. OLE1 was expressed at higher levels in the upl3-4mutant (Figure 3F), consistent with the hypothesis that UPL3mayinfluence LEC2 protein levels. To assess if UPL3 directly affectsLEC2-mediated transcription ofOLE1, transient expression of theArabidopsisOLE1 promoter fused to firefly Luciferase (fLUC) wasperformed in Arabidopsis upl3-4 mutant mesophyll protoplasts.The low levels of endogenousOLE1 promoter activity (Figure 3G)were increased by cotransfection with 35S:3HA-LEC2. Co-transfection with both 35S:3HA-LEC2 and 35S:3FLAG-UPL3 signifi-cantly reduced LEC2-inducedOLE1promoter activity, suggestingthat UPL3 reduced LEC2-mediated transcriptional regulation ofseed lipid biosynthetic genes.

Figure 2. (continued).

(B)BnaUPL3.C03 expression in developing pods ofB. napus accessions described in Figure 1Ewas correlatedwith expression in leaves.BnaACTIN2wasused asan internal RT-qPCRcontrol. RT-qPCRanalysesused threebiological replicates. Lineswith highSWPPare representedbyorangedata points, andlines with low SWPP are shown as gray data points.(C) Loss-of-function mutants of AtUPL3 have enlarged seeds, while AtUPL3 expression from the 35S promoter in Col-0 plants reduced seed size. Areasof Arabidopsis seeds fromCol-0, a representative transformant, and two independent upl3 T-DNA loss-of-function mutants. Data are given asmean6 SE;n 5 six biological replicates per genotype using 100 seeds per replicate. P values were determined by Student’s t test. WT, wild type.(D) Loss-of-function mutants of AtUPL3 has elevated seed lipid content, while AtUPL3 expression from the 35S promoter in Col-0 plants reduced lipidcontent. Lipid content of Arabidopsis seeds fromCol-0, two upl3 T-DNAmutants, and a representative 35S:3HA-UPL3 transformant wasmeasured usingnear-infrared spectroscopy. Data are given asmean6 SE; n5 6 biological replicates per genotype using 100 seeds per replicate.P valueswere determinedby Student’s t test.(E) Altered mucilage extrusion of imbibed Arabidopsis seeds in the upl3-4 mutant, visualized by Ruthenium Red staining.(F)Altered expressionof the regulatory transcription factorGL2during early stagesof seeddevelopment inAtupl3-4mutant seeds.RNA fromwhole siliquesharvested at 5 to 10 DPA was used. Expression levels are relative to EF1ALPHA expression. Data are given as means 6 SE; n 5 3 biological replicates.P values were determined by Student’s t test.(G)Alteredexpressionof the regulatory transcription factorMUM4duringearlystagesof seeddevelopment inupl3-4mutantseeds.RNAfromwholesiliquesharvested at 5 to 10 DPA was used. Expression levels are relative to EF1ALPHA expression. Data are given as means 6 SE; n 5 3 biological replicates.P values were determined by Student’s t test. WT, wild type.

2374 The Plant Cell

UPL3 Mediates the Proteasomal Degradation of LEC2

HECT E3 ligases such as UPL3 mediate the proteasomaldegradation of substrate proteins by direct ubiquitylation oftheir substrates (Maspero et al., 2013). To test UPL3-mediated

degradation of LEC2, the predicted active site Cys of AtUPL3(Downes et al. 2003) was mutated to Gly (C1855G), and wild-typeand predicted inactive AtUPL3 formswere expressed inNicotianabenthamiana as C-terminal FLAG fusions and purified on FLAG-MA (magnetic) beads. UPL3-mediated degradation of LEC2-HIS

Figure 3. UPL3 Reduces LEC2-Mediated Gene Expression.

AlteredexpressionofArabidopsis seedmaturationgenes in theupl3-4mutant.RT-qPCRwasused tomeasuregeneexpressionduringseeddevelopment inCol-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 geneexpression. Data are given as means 6 SE n 5 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-qPCRmeasurementofOLE1 in10DPACol-0andupl3-4mutantArabidopsisplants.RNA fromwhole siliqueswasused.Expression levelsare relativeto EF1ALPHA gene expression. Data are given as means 6 SE; n 5 3 biological replicates. P values were determined by Student’s t test.(G) AtUPL3 reduces LEC2-mediated activation of theOLE1 promoter. Transient expression of the LEC2- regulatedOLE1 promoter:fLUC reporter gene inArabidopsis upl3-4 mutant leaf protoplasts. 35S:Renilla LUC vector was cotransfected in all treatments as a control, and the ratio of Firefly/Renilla LUCactivitywasused todetermineOLE1:fLUCgeneexpression levels. 35S:3HA-LEC2and35S:3FLAG-UPL3werecotransfectedasshown.Dataarepresentedas means 6 SE; n 5 3 independent transfections. P values were determined by Student’s t test.

Varying UPL3 Expression Modulates Crop Yield 2375

protein was performed in cell-free extracts of upl3-4 seedlings(Figure 4A). Wild-type UPL3-FLAG promoted a more rapid re-duction in LEC2-HIS levels than mutant UPL3-FLAG, and thisreduction was inhibited by the proteasome inhibitor MG132. Tosupport these observations of LEC2-HIS stability, LEC2 proteinlevels in Arabidopsis upl3-4 mutant plants, in wild-type plants,and in transgenic plants expressing 35S:3HA-UPL3 were as-sessed using LEC2-specific polyclonal antibodies. Increasedendogenous LEC2 protein levels were observed in mutantArabidopsis upl3-4 siliques as compared with wild-type plants(Figure 4B), while reduced LEC2 levels were seen in siliques of35S:3HA-UPL3 transgenic lines as compared with wild-typesiliques (Figure 4B). These observations indicated that AtUPL3activity influenced LEC2 protein stability, and this was reflectedby altered LEC2 protein levels in vivo in upl3-4 mutant andoverexpressing lines.

UPL3 Promotes the Formation of Higher MW Forms of LEC2

Immunoblot detection of endogenous LEC2 protein in de-veloping siliques of 35S:3HA-UPL3 transgenic plants revealedLEC2 forms with higher apparent molecular weight (MW) incomparison to those observed in wild-type protein extracts(Figure 4C), consistent with UPL3-mediated ubiquitylation ofLEC2. Co-expression of LEC2-HA with wild-type or mutantUPL3-3FLAG in N. benthamiana and pull-down of ubiquitylatedproteins using FLAG-TR-TUBE showed that higherMW forms ofLEC2-HAwere specifically formed by the activity of UPL3, as nohigher MW LEC2-HA products were detected by co-expressionwith predicted catalytically inactive UPL3-3FLAG (Figure 4D).These higher MW forms interacted with the ubiquitin bindingmotifs on FLAG-TR-TUBE protein, suggesting they may beubiquitylated forms LEC2. Taken together, these results provideevidence that UPL3 promotes the formation of higher MW formsof LEC2 and destabilizes LEC2 in vitro.

Variation in the Promoter and 59UTR Region ofBnaUPL3.C03 Is Sufficient for Differential Expression andInfluences Variation in Yield Traits

We observed significant differences in BnaUPL3.C03 ex-pression between GWAS accessions displaying variation inSWPP (Figure 1F; Supplemental Table 2). Alignment ofBnaUPL3.C03 promoter and 59 untranslated region (UTR)sequences (from 2 kb upstream of the ATG initiation codon)of a high- (Coriander) and a low-expressing (Dimension) ac-cession revealed multiple sequence differences, including34 SNPs and seven indels between 3 to 60 nucleotides(Supplemental Figure 8). The Dimension and CorianderBnaUPL3.C03 promoter and 59UTR regions were fused toa fLUC reporter gene (Figure 5A) and their activities assessedafter transfection of Arabidopsis mesophyll protoplasts.Figure 5B shows that theCoriander promoter and 59UTR regiondrove approximately three times more luciferase activity thanthe Dimension promoter in these cells. This difference inBnaUPL3.C03 expression levels was consistent with RNAsequencing and RT-qPCR data (Figures 1E and 1F) from

B. napus seedlings, showing that variation in promoter and59UTR activity is the primary source of variation in BnaUPL3.C03 expression between these two accessions.To assess the role of BnaUPL3.C03 promoter and 59UTR

sequence variation in variation in seed size, transgenic Ara-bidopsis lines expressing the Arabidopsis UPL3 coding regionfused to an N-terminal 3HA epitope driven by Corianderand Dimension BnaUPL3.C03 promoters were constructed.Transgenic linesweremade in theupl3-4mutant background toassess differential complementation of upl3-4 mutant pheno-types in response to transgene expression. UPL3 expressionwasmeasured in 10 DPA siliques of transgenic lines, the upl3-4mutant, and wild-type plants. The two B. napus promoter re-gions expressed the AtUPL3 coding region at the predictedlevels in developing siliques (Figure 5C). Seed sizes in the high-expressing Coriander promoter transgenic lines showed nearwild-type seed sizes, indicating complementation of the largeArabidopsis seed upl3-4 phenotype. Reduced levels of com-plementation of the large seed upl3-4 phenotype was observedin plants expressing AtUPL3 under the control of the low-expressing Dimension promoter (Figure 5D). Thus, variationin BnaUPL3.C03 promoter and 59UTR transcriptional activitycan influencefinal seedsizebydrivingdifferent levels ofAtUPL3coding region expression and 3HA-UPL3 protein accumulation(Figure 5D, bottom).

Differential Expression of BnaUPL3.C03 Leads to Variationin BnaLEC2 Protein Levels and Modulates Final SeedLipid Content

Associative transcriptomics identified a negative relation-ship between BnaUPL3.C03 expression levels and SWPP inB. napus accessions (Figures 2B and 6A). Consistent withseed size phenotypes seen in transgenic Arabidopsis (Figure5D), low BnaUPL3.C03 expressing lines had larger seeds(Figure 6B) and higher thousand seed weights (Figure 6C). Torelate these B. napus phenotypes to the proposed mechanismof UPL3-mediated destabilization of LEC2 in Arabidopsis(Figure 4), LEC2 antibody was used to assess LEC2 proteinlevels in the developing seeds of a subset of six B. napusaccessions varying in BnaUPL3.C03 expression (Figure 2B;Supplemental Table 2). Figure 6D shows that LEC2 proteinlevelswere higher in all threeB.napus accessionswith reducedBnaUPL3.C03 expression, as compared with accessions withhigher BnaUPL3.C03 expression in which no LEC2 was de-tected in seeds. These observations show that the mechanismofUPL3-mediated control of LEC2protein levels established inArabidopsis underlies variation in seed size observed inB. napus accessions. As predicted by this mechanism, significantdifferential expression ofBnaOLE1 (which is regulated by LEC2in Arabidopsis; Figure 3G) was observed across the subsetof B. napus accessions displaying differential expression ofBnaUPL3.C03 during seed maturation (Figure 6E). Finally,accessions displaying reducedBnaUPL3.C03 and consequentincreased BnaOLE1 expression levels exhibit significantlyhigher seed lipid levels relative to those displaying highBnaUPL3.C03 expression (Figure 6F).

2376 The Plant Cell

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. coliwas incubated at 22°C for the times indicated in total protein extractsfrom upl3-4 mutant seedlings, with either wild-type– or mutant–purified UPL3-3FLAG added (left), with and without 50 mM of MG132. Immunoblots ofreactionswereprobedwithanti-HISantibodies. “CBB”showsaportionof the reactionstainedwithColloidalCoomassieBlueasa loadingcontrol. Thegraphshows 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. wt,wild type; mut, mutant.(B) Immunoblots of protein samples from leaf or 10 to 15 DPA siliques of Col-0, lec2mutant, upl3-4mutant, and 35S:3HA-UPL3 plants electrophoresed onSDS-PAGE gels and probed with anti-LEC2, anti-HA, or anti-tubulin antibodies as a loading control. TUBULIN levels are to compare protein loading.(C)Expressionof3HA-UPL3 increaseshigherMWformsofLEC2 indevelopingsiliques. Immunoblotsofprotein samples frompooled5 to10DPAsiliquesofCol-0 or 35S:3HA-UPL3plantswere electrophoresed on4%to 20%SDS-PAGEgels andprobedwith anti-LEC2.HigherMW formsof LEC2protein seen inthe35S:3HA-UPL3 sampleare indicatedbyarrows.Thebottomshows thatbothwild-typeandmutantUPL3-3FLAG interactwithLEC2-HAduring transientco-expression in N. benthamiana leaves. wt, wild type; mut, mutant.

Varying UPL3 Expression Modulates Crop Yield 2377

Reduced BnUPL3.C03 Expression Has Not Yet BeenExploited in Current B. napus Breeding Material

To assess if variation inBnaUPL3.C03 expression levels has beenselected in the breeding of commercial oilseed rape varieties,expression was measured across a panel of seven current eliteB. napus lines. Expression levels were compared with those mea-sured in GWAS accessions exhibiting high differential expressionof BnaUPL3.C03 and high variation in yield traits (Figure 7). Therewas substantial variation in BnaUPL3.C03 expression in theelite lines, with relatively high levels of expression in several linescompared with the low-expressing line Licrown 3 Express. There-fore, there appears to be a significant potential for further yieldincreases in elite oilseed rape germplasm through reduction ofBnaUPL3.C03 expression.

DISCUSSION

Using associative transcriptomics (Harper et al., 2012), weidentified variation in the expression of a B. napus gene,BnaUPL3.C03, that modulates seed size, lipid content, and fieldyields in this important oilseed crop. BnaUPL3.C03 encodes anortholog of theArabidopsisHECTE3 ligaseUPL3.We show that itpromotes the formation of HMWproducts of LEC2with an affinityfor ubiquitin binding proteins and destabilizes this “hub” tran-scriptional regulator of seed storage processes (Boulard et al.,2017).ReducedBnaUPL3.C03expressionmaintains higher levelsof LEC2 during seed maturation, prolonging transcriptional acti-vation of storage lipid genes, leading to larger seedswith elevatedlipid levels in B. napus. Lines with relatively low BnaUPL3.C03expression had robust yield increases in field conditions ascompared with lines with higher BnaUPL3.C03 expression. Var-iation in UPL3 expression levels in elite oilseed rape cultivarsidentifies a promising approach for achieving further increases inoilseed yields by selecting for reduced UPL3 expression levels.

Comparison of assemblies of the BnaUPL3.C03 gene andflanking sequences from two B. napus lines with low and highBnaUPL3.C03 expression identified high levels of promoter se-quence variation that segregated with high SWPP and reducedBnaUPL3.C03 expression (Figures 5B and 5C). Transient assaysshowed that 2-kb 59 flanking regions and UTRs from high- andlow-expressing BnaUPL3.C03 accessions promoted the ex-pected expression differences (Figure 5B). Driving the expressionof theArabidopsisUPL3coding sequencewithB.napuspromotervariants led to differential complementation of the Arabidopsisupl3-4mutant seed phenotypes (Figure 5D). This established thatnatural variation in the activity of the BnaUPL3.C03 promoter and59UTR region was sufficient to cause variation in seed size and

yield traits in the accessions. Expression of BnaUPL3.A08 wasnot associated with SWPP (Figure 1D), nor did its expression varyin developing seeds of accessions with high or low SWPP(Supplemental Figure3). This suggested that theprimary influenceofUPL3 expression levels on SWPP is from variation inBnaUPL3.C03 expression. The multiple variants detected between twopromoters driving differential expression and the continuousprofile of BnaUPL3.C03 expression in the accessions (Figure 1E)suggest there may be multiple sequence variants that togetherreduce promoter activity. Although limited, variation in other re-gions of theBnaUPL3.C03gene, including the coding region,mayalso have the potential to contribute to variation in transcriptabundance and yield observed in B. napus varieties.Such genetic variation influencing gene regulation is an in-

creasingly important resource for trait improvement, includingincreasing seed yields. For example, natural variation in the copynumberof apromoter-silencingelementof theFZPgeneunderliesvariation in spikelet numbers in rice (Oryza sativa) panicles andis an important determinant of yield (Bai et al., 2017). Similarly,Quantitative Trait Locus analyses identified promoter variation inthe rice GW7 gene that, combined with variation that reducedexpression of an SPL16 transcriptional repressor of GW7, led to10% increases in grain yield and improved quality (Wang et al.,2015). A deletion in the promoter region of GW5 in rice cv Nip-ponbare lines reduced expression and increased grain width (Liuet al., 2017), and similarly, promoter deletions inGSE5 accountedfor wide grains in indica rice varieties (Duan et al., 2017). Moregenerally, genetic variation in chromatin accessibility (a mark ofpromoter activity) explained ;40% of heritable variation in manyquantitative traits in maize (Zea mays; Rodgers-Melnick et al.,2016). These reports, and the study described here, reveal theexceptional promise of accessing variation in promoter se-quences and altered transcriptional activity for identifying regu-latory mechanisms and for the quantitative manipulation ofcomplex traits such as yield in crop plants.UPL3 was first identified in Arabidopsis as a HECT E3 ligase

gene whose loss-of-function mutation causes increased leaf hairbranching (Downes et al., 2003; El Refy et al., 2003). Here, weshowed that UPL3 expression increased during seed maturationin Arabidopsis (Figure 2A), and an ortholog, BnaUPL3.C03, wasdifferentially expressed in the seeds of B. napus lines (Figures 2Band 6A) that varied in seed weight (Figure 6C) and seed lipidcontent (Figure 6F). LEC2, a transcriptional regulator of seedstorage processes, is more stable in siliques as a upl3-4 loss-of-function mutation in Arabidopsis (Figure 4B) and in B. napusaccessions with relatively lower BnaUPL3.C03 expression levels(Figure 6D). AtUPL3 physically interacted with LEC2 (Figure 4C),suggesting a direct functional relationship. UPL3-FLAG, but not

Figure 4. (continued).

(D) Purification of ubiquitylated forms of LEC2-HA from transiently expressed 35S:LEC2-HA and 35S:UPL3-3FLAG (wild type and mutant) vectors in N.benthamiana leaves. Protein samples were extracted and samples taken to assess protein expression levels using immunoblotting (loading panels). Theremainingsamplesweresplit in twoand;20mgofFLAG-TR-TUBEwasadded tooneset (lower loadingpanel). This fractionwaspurifiedonFLAG-MAbeadsand theother purifiedusingHA-MAbeads.Affinity-purifiedproteinswere subjected to immunoblotting usinganti-HA-HRPantibodies. Thebottom rightwasexposed for longer than the bottom left. The red arrow indicates the position of LEC2-HA protein. Higher MW forms of LEC2-HA detected in the FLAG TR-TUBE pull-down were dependent on the activity of UPL3-3FLAG. wt, wild type; mut, mutant.

2378 The Plant Cell

Figure 5. Variation inBnaUPL3.C03 Promoter Activities fromHigh- and Low-SWPPB. napus Accessions Is Sufficient To Generate Variation in Final SeedYield.

(A) The diagram shows fusions of 2-kb BnaUPL3.C03 promoter and 59UTR regions from the B. napus lines Dimension (DIM; High SWPP) and Coriander(COR; Low SWPP) to the Luciferase coding sequence. The diagrams are not to scale.(B) Differential expression of the Luciferase reporter gene by the DIM and COR promoter and 59UTR region. The Dimension (DIM) and Coriander (COR)promoters described in (A) fused to a fLUC reporter gene were transfected into Arabidopsis upl3-4mutant protoplasts. The activities of each promoter areshown relative to co-expressed 35S:Renilla luciferase. Data are given as means 6 SE, n 5 3. P values were determined by Student’s t test.(C)Differential expressionof theArabidopsisUPL3coding regionby theCoriander (COR) andDimension (DIM) promoter and59UTR regions.BnaUPL3.C03promoter and 59UTR regions from theB. napus lines DIM andCORwere used to express the coding region of ArabidopsisUPL3 fused to 3HA at its amino-terminus inupl3-4mutantArabidopsis.RT-qPCRofUPL3expression in leavesofwild-type and transgenicArabidopsis linesweremeasuredandare shown

Varying UPL3 Expression Modulates Crop Yield 2379

apredictedcatalytically inactive form,promoted theproteasomal-dependent instability of LEC2-HIS in vitro (Figure 4A), and ex-pression of 3HA-UPL3 in transgenic Arabidopsis promoted theformation of HMW forms of LEC2 in developing seeds (Figure 4C).Finally, co-expression in plants of UPL3-3FLAG promoted theformation of higher MW forms of LEC2HA that were specificallypurified by TR-TUBE, which has a high affinity for ubiquitylatedproteins (Figure 4D). Conversely, a predicted catalytically inactiveform of UPL3-3FLAG did not lead to the formation of these higherMW forms of LEC2-HA.

These observations suggested that UPL3 ubiquitin ligase ac-tivity directly modulated levels of LEC2 protein during seedmaturation. LEC1 is directly activated by LEC2, and together theyactivate the expression of genes involved in promoting seed lipidaccumulation, such as OLE1. Consistent with this, we observedelevated OLE1 expression levels in Arabidopsis upl3-4 mutantseeds relative to the wild type and showed that BnaOLE1 is dif-ferentially expressed between B. napus accessions displayingdifferential expression of BnaUPL3.C03 (Figure 6E). These ob-servations show that reduced levels of UPL3 maintain higherlevels of LEC2proteinduringseedmaturation, thusprolonging theduration of expression of seed maturation genes, leading to in-creased seed lipid in Arabidopsis upl3-4 mutants and B. napusaccessions with reduced BnaUPL3.C03 expression.

At earlier stages of seed development, the seed coat in Ara-bidopsis upl3-4 mutants has altered mucilage and reduced ex-pression of GL2, a transcription factor controlling epidermal celldifferentiation (Figure 2F; Lin et al., 2015). GL2 promotes ex-pression of the rhamnose biosynthesis gene MUM4 that is re-quired for seed mucilage production (Lin et al., 2015; Shi et al.,2012), and expression of MUM4 is also reduced during testadevelopment in upl3-4 (Figure 2G). This observation revealeda unifying role of UPL3 in regulating both testa and embryomaturation by modulating levels of transcription factors duringdifferent stagesof seeddevelopment. These transcription factors,GL3 and EGL3 (Patra et al., 2013), and LEC2 (this study) inturn modulate expression of other transcription factors and bio-synthetic genes involved in testa and embryo development.

Altered expression of LAFL genes has profound developmentalconsequencessuchasectopicembryogenesis (Stoneetal., 2001;Roscoeet al., 2015), but inducedexpressionofLEC2 in vegetativetissues does increase lipid accumulation (Andrianov et al., 2010;Santos Mendoza et al., 2005). These studies showed that theactivities of LEC2 expression in storage processes and embryodevelopment were difficult to separate, probably due to thetimings of expression, interdependence, and partial redundancyof LAFL gene function. By identifying a mechanism controllingLEC2 protein levels during seed maturation, we have shown thatit is possible to elevate lipid levels during normal Arabidopsis

embryodevelopment (Figure2D).Mis-expressionofWRI1permitsnormal seed development while increasing lipid content (Kanaiet al., 2016; van Erp et al., 2014) by extending seed maturation,consistent with our observations of elevated WRI1 expressionin upl3-4 (Figure 3E). Intensive breeding is optimizing oilseedlipid composition for different end-uses, but comparatively slowprogress is being made in increasing yields of oilseed crops suchas oilseed rape,with current rates of yield increase predicted to beinsufficient to meet future needs (Ray et al., 2013). Genetic vari-ation that reduces expression of BnaUPL3.C03 and increasesseed lipid content appears not to have been exploited in oilseedrape breeding (Figure 7), demonstrating how lipid content andseed yields could be increased without influencing composition.The presence of a potential LEC2 ortholog in soybean (Glycinemax; Manan et al., 2017), and the expression of LEC2 and otherB3-domain transcription factor homologs during seed lipid syn-thesis in sunflower (Helianthus; Badouin et al., 2017) and oil palm(Elaeis guineensis; Singh et al., 2013) reveals the potential ofUPL3-mediated regulation of LEC2 to increase seed lipid levelsand overall yields in other major oilseed crops.

METHODS

Plant Material and Growth Conditions

Phenotype data were collected from 94 accessions representing winter,spring, and Chinese oilseed rape (Brassica napus) from the OREGIN fixedfoundation diversity set (Harper et al., 2012). Plants were grown ina randomized, triplicated experimental design in a Keder greenhouse(https://www.kedergreenhouse.co.uk/) under natural light with no controlledheating. Before transplantation, plants were grown (18°C/15°C day/night,16-h light) for 4 weeks before 6weeks vernalization (4°C, 8-h light). Twentytypical pods were collected from each mature plant and digitally imaged.Pod length (“Podl”) was measured using the software ImageJ (Schneideret al., 2012). Podswereweighed (“PW”) before threshing to remove seeds.Seednumbers, averageseed length (“SL”),width (“SW”), area (“SA”), singleseed weight (“SSW”), and thousand grain weight (“TGW”) were measuredfor each sample using a Marvin device (GTA Sensorik). Numbers of seedsper pod (“SPP”), “SWPP,” and seed density (“SDen”) were calculated fromthese data. Field yields of selected accessions were grown in four repli-cated field plots (1.25 m 3 6 m, Church Farm, Bawburgh, Norfolk, UK) ina randomized design and harvested by combine. Total plot yield wasdetermined for each plot and an average plot yield taken.

All Arabidopsis (Arabidopsis thaliana) mutant and transgenic lines usedwere in an ecotype Columbia-0 (Col-0) background. Plants were grownon soil in a growth chamber (Sanyo-Panasonic Model MLR-351) usingcool white fluorescent lighting with 16-/8-h day/night at 22°C after 48-hstratification at 5°C. Two independent-sequence–indexed T-DNA in-sertion lines in the 10th exon of At4G38600 (UPL3), Salk_015334 (termedupl3-4) and Salk_151005 (termed upl3-5), were obtained from the Not-tingham Arabidopsis Stock Centre. Genotyping primers were designed

Figure 5. (continued).

relative to theAtEIF1ALPHA gene. Data are given asmeans6 SE; n5 at least three biological replicates of three independent transformants.P values weredetermined by Student’s t test.(D) The Coriander (COR) and Dimension (DIM) promoters show differential complementation of Arabidopsis upl3-4 seed size. Seed area was quantified inwild type, upl3-4 mutant, COR:3HA-AtUPL3, and DIM:3HA-AtUPL3 independent transgenic lines. Data shown are means 6 SE based on seed areameasurements across at least 100 seeds per genotype and with at least three biological replicates of three independent transformants. P values weredetermined by Student’s t test. (Lower) Immunoblots show 3HA-UPL3 protein levels in 10 to 15 DPA seeds and tubulin levels for comparison.

2380 The Plant Cell

using theprimer design tool http://signal.salk.edu/tdnaprimers.2.html. Theprimer sequences are in Supplemental Table 2. Genotyping used TAKARAEX taq (Takara Bio). Both SALK alleles abrogated gene expression(Supplemental Figure 5)

To identify loss-of-function mutations in UPL3 in Brassicas, a TILLINGpopulation (Stephenson et al., 2010) ofBrassica rapa (the A genome donortoB.napus)wasscreened formutations in thepredictedgeneBra010737.1on chromosome A08 that encodes an ortholog of Arabidopsis UPL3. Thepredicted B. rapa gene was defined by a full-length transcript. There wasonly one clear ortholog of UPL3 in B. rapa, consistent with two copiespresent in the amphidiploid B. napus on chromosomes A08 and C03.Two premature stop codon mutations were identified in Bra010737.1(Supplemental Figure 6) in lines JI32517 and JI30043. Primers were de-signed to amplify genomic DNA from the mutated region (SupplementalTable 4) and used to validate the TILLING mutations. Line JI32517-B wasused in further studies. This line was selfed and progeny screened forhomozygous (G>A) and wild-type (C>T) changes by sequencing of thelocus. Seeds were harvested from the wild type and the upl3 mutant andtheir area measured using the software ImageJ.

A subset of B. napus accessions and a panel of elite B. napus breedinglines were grown under glasshouse conditions after vernalization. Leafmaterial was harvested from the first true leaf and stored at270°C beforefurther processing. Developing pods were staged by tagging when petalswere beginning to emerge from the developing bud, taken as zero DPA.Samples for RNA isolation were collected at 45 DPA and stored at270°C.For expression analyses in Arabidopsis, wild-type Col-0 and upl3 mutantplants were grown as described, without vernalization, and floral budstagged when petals were beginning to emerge from the developing bud(0 DPA). Siliques were then harvested at 0, 5, 10, and 15 DPA, and tissuesamples were stored at 270°C.

Population Structure Analysis

Associative Transcriptomics analysis was performed as described inHarper et al. (2012). The population structure Q matrix was recalculatedusing 680 unlinked markers across the set of 94 lines. Run-length

Figure 6. Relationships between BnaUPL3.C03 Expression Levels inHigh- and Low-SWPP B. napus Accessions to Seed Size, Seed LEC2Protein Levels, and Seed Lipid Content.

(A) Comparison of BnaUPL3.C03 expression levels in 45-DPA seeds inB. napus Dimension (DIM) with high SWPP, and Coriander (COR) ac-cessions with low SWPP, measured by RT-qPCR. Expression levels arerelative to the BnaACTIN2 gene. Data are presented as means6 SE; n5 3for each genotype. P values were determined by Student’s t test.

(B) Seed sizes in the low-expressing BnaUPL3.C03 Line Dimension andthe high-expressing BnaUPL3.C03 Line Coriander.(C) Thousand seed weights of low-expressing BnaUPL3.C03 line Di-mension and the high-expressing BnaUPL3.C03 line Coriander. Data areshown as means 6 SE. Seeds were weighed in batches of 100 seed andthousand seedweight calculated based on these values. n5 3 batches foreach genotype assayed. P values were determined by Student’s t test.(D) LEC2 protein levels in 45-DPA seeds were detectable in threelow BnaUPL3.C03 expressing lines, and undetectable in three highBnaUPL3.C03 expressing accessions (Supplemental Table 2). Immuno-blotsof seedproteinextractwereprobedwithanti-LEC2 (top)andwithanti-tubulin (bottom) as a protein loading control.(E) Elevated expression of the LEC2-regulated gene BnaOLE1 in thelow-expressing BnaUPL3.C03 (Dimension) line. BnaOLE1 expressionwas quantified by RT-qPCR in low-expressing BnaUPL3.C03 (Di-mension) and high-expressing BnaUPL3.C03 (Coriander) accessions.Expression levels were relative to that of BnaACTIN2. n 5 3 for eachgenotype. Primers were designed to measure expression of bothBnaC01g17050D and BnaA01g14480D BnaOLE1. P values were de-termined 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-expressingBnaUPL3.C03 (Dimension) and high-expressing BnaUPL3.C03 (Co-riander) accessions. P values were determined by Student’s t test.

Varying UPL3 Expression Modulates Crop Yield 2381

comprised 10,000 burn-in followed by 10,000 steps. The ancestry modelwas admixture, and allele frequencies were independent between pop-ulations. Between 1 and 10 values of K were tested using three iterations.One SNP per 500-kb interval along pseudomolecules, excluding regionsless than 1,000kb fromcentromeres (Masonet al., 2017), were selected forBayesian population structure analysis via the program “STRUCTURE2.3.3” (Pritchard et al., 2000). This analysis incorporated a minor allelefrequency of 5%. The optimum number of K populations was selected asdescribed in Harper et al. (2012).

SNP Analysis

SNP data, STRUCTURE Q matrix, and phenotypes of the 94 accessionswere combined using the software TASSEL (V4.0, https://tassel.bit-bucket.io/). After the removal of minor alleles (frequency < 0.05);144,000SNPs were used to calculate a kinship (K) matrix to estimate the pairwiserelatedness between individuals. Data sets were analyzed using bothgeneralized LMs andMLMs. Goodness of fit of the model was determinedby a QQ plot (Supplemental Figure 9) of the observed versus the expected2log10P values. 2log10P values were plotted in chromosome order andvisualized using “R” programming scripts as described previously byHarper et al. (2012; http://www.R-project.org/). RPKM and trait data var-iation were analyzed by linear regression analysis using the software “R.”The2log10P values were plotted in pseudomolecule order and visualizedusing R scripts as described previously by Harper et al. (2012).

Arabidopsis and Brassica Seed Size Quantification

Arabidopsis seeds were harvested frommature plants and imaged at 103magnification. B. napus and B. rapa seeds were harvested from matureplants and scanned using a photocopier. Seed area was quantified usingImageJ particle analysis.

Ruthenium Red Staining

Seed mucilage phenotypes were assessed using methods described byMcFarlane et al. (2014). Stained seeds were imaged at 103magnification.

Seed Lipid Quantification and Profiling

Fatty acid profile analyses in Arabidopsis were performed using themethods described by Li et al. (2006). Lipid content ofB. napus seedsweremeasured using near-infrared spectroscopy (Wang et al., 2014).

PCR and Sequencing

All PCR reactions were performed using Phusion High Fidelity DNApolymerase (New England BioLabs) according to manufacturer’s in-structions. Capillary sequencing was performed by GATC Biotech.

cDNA Synthesis and RT-qPCR

RNA was extracted using the SPECTRUM Total Plant RNA kit (Sigma-Aldrich). RNA (1 mg) was treated with RQ1 RNase-Free DNase (Promega)and cDNA synthesis was performed with the GoScript Reverse Tran-scription system (Promega) using Oligo (dT). All protocols were performedusingmanufacturers’ guidelines. cDNA sampleswere diluted 1:10 inwaterbefore use. RT-qPCR was performed using SYBR Green Real-Time PCRmastermix (Thermo Fisher Scientific) and performed using Lightcycler 480(Roche). Primer sequencesused forRT-qPCRare inSupplemental Table 4.Primer efficiencies and relative expression calculations were performedaccording tomethodsdescribedbyPfaffl (2001). All RT-qPCRassayswererepeated at least twice.

DNA Constructs

The p35S:3HA-AtUPL3 transgenic line was generated by cloning AtUPL3cDNA into the pENTR TOPO-D vector (Thermo Fisher Scientific) using theprimers described in Supplemental Table 4. LR Clonase Mix II (ThermoFisherScientific)wasused to transfer theAtUPL3coding sequences (CDS)into the 35S PB7HA binary vector. The Arabidopsis UPL3 cDNA TOPOconstruct was cloned into the 33FLAG PW1266 vector. The active siteCys-1855 residue (Downes et al. 2003) was mutated to Gly using GeneArtkits and primers described in Supplemental Table 4. Full-length Arabi-dopsisLEC2cDNAwasamplifiedusingprimersdescribed inSupplementalTable 4. After cloning into pENTR TOPO-D, the LEC2CDSwas transferredby LR reaction to pEARLEY 103. Constructs were transformed intoAgrobacteriumtumefaciensstrainGV3101,andArabidopsisupl3-4mutantplants were transformed using the floral dip method (Clough and Bent,

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 seedsof seven current elite commercial cultivars of oilseed rapeweremeasured usingRT-qPCRandcomparedwith expression levels of BnaUPL3.C03 from GWAS accessions Samurai and Licrown 3 Express, which exhibit high and low BnaUPL3.C03 expressionlevels, and low and high SWPPphenotypes, respectively.BnaACTIN2 expressionwas used for comparison usingRT-qPCR control. Data are presented asmeans6 SE;n5 3 for each genotype.P values estimatedbyStudent’s t test show the significance ofBnaUPL3.C03 expression levels as comparedwith thelow-expressing accession Licrown 3 Express.

2382 The Plant Cell

1998). Promoter regions of theBnaUPL3.C03gene from the high- and low-expressing accessions Coriander and Dimension were amplified usingprimers described in Supplemental Table 4. BnaUPL3.C03 promoter PCRproducts, digested with Stu1 and Xho1, were ligated into pEarly 201-AtUPL3-3FLAG CDS plasmid. The resulting BnaUPL3.C03 promoter:AtUPL3-3FLAG CDS constructs was transformed into A. tumefaciensstrain GV3101, and Arabidopsis upl3-4 mutant plants were transformedusing the floral dip method (Clough and Bent, 1998).

Promoter Transactivation Assay

Promoters and full-length cDNAs of selected Arabidopsis Col-0 geneswere amplified by PCR using Phusion polymerase (Thermo Fisher Sci-entific) according to the manufacturer’s guidelines. Promoter primer se-quences are in Supplemental Table 4. PCR reactions were purified usingWizard SV Gel and the PCR Cleanup system (Promega) and inserted intopENTR D-TOPO vector (Thermo Fisher Scientific), and an LR reaction wasused to clone promoters into a fLUCreporter vector pUGW35. LEC2 andUPL3 CDS were transferred using LR Clonase into PB7HA and PW1266,respectively, to create p35S:3HA-LEC2 and p35S:3FLAG-UPL3. A 35S:Renilla luciferase construct was used to quantify relative promoter activ-ities. Plasmid preparations for transient assays were prepared using thePlasmid Maxi Kit (Qiagen) according to manufacturer’s instructions.

Promoter transactivation assays were performed using protoplastsisolated from upl3 leaves (Wu et al., 2009). Assays were carried in triplicateusing 5 mg of plasmid and 100 mL of purified protoplasts (;50,000 cells).After anovernight incubation at room temperature, transfectedprotoplastswere harvested and promoter activity assessed using the Dual LuciferaseReporter assay system (Promega). The ratio of fLUC to Renilla Luciferaseactivity was determined using the dual assay Promega protocol ona Glomax 20/20 luminometer (Promega). All transactivation assays wereconducted in triplicate and repeated at least twice.

Total Protein Extraction from B. napus Pods andArabidopsis Siliques

Material was ground to a fine powder in liquid nitrogen and resuspended inextraction buffer (1mL/g freshweight; 25mMof Tris-HCl at pH 8.0, 10mMofNaCl, 10mMofMgCl2, 4mMof4-benzenesulfonyl fluoride [AEBSF], and50 mM of MG132). After an incubation on ice for 30 min, samples werecentrifuged at 15,000 rpm for 5 min at 4°C. Supernatant was then addedto a fresh tube and centrifugation repeated for 10 min. Total proteincontent was assessed using Bradford reagent (Bio-Rad) according tomanufacturer’s instructions.

Immunoblot Analysis

LEC2proteinwasassesseddirectly usingaffinity-purified rabbit polyclonalantibodies raised against the antigenic peptide ARKDFYRFSSFDNKKLfromLEC2coupled tokeyhole limpet hemocyanin (NewEnglandPeptides).Antibodieswereusedat the followingdilutions: anti-HIS (cat. no.A7058, lot088M4865V; Sigma-Aldrich; 0.0005 dilution); anti-FLAG-HRP (cat. no.F1804, lot SLBW5142; Sigma-Aldrich; 0.001 dilution); anti-GFP-HRP (cat.no. 120-002-165, lot 5196017043; Miltenyi; 0.0002 dilution); anti-tubulin(cat. no. T9026, lot 086M4773V; Sigma-Aldrich; 0.0002 dilution); anti-HA-HRP (cat. no. 3F10 1,588,800, lot 12013819001; Roche; 0.001 dilution);and anti LEC2 (New England Peptides; 0.001 dilution). Secondary anti-bodies used for tubulin and LEC2 were anti-rabbit (cat. no. A0545, lot102M4823; Sigma-Aldrich) (0.0002 dilution); for tubulin anti-mouse (cat.no. A8924, lot SLBH4089; Sigma-Aldrich; 0.0002 dilution). Immunoblotswere developed with FemtoMax peroxidase substrate (Thermo FisherScientific).

Protein Expression in Escherichia coli BL21

The Arabidopsis LEC2 CDS was amplified from seedling cDNA usingprimers described in Supplemental Table 4. After purification, PCRproducts were TOPO-cloned into pET-24a to generate a C-terminal HISfusion. FLAG–TR–TUBE for expression in E. coli was performed as de-scribed in Dong et al. (2017). Plasmids were transformed into BL21 E. colicells, grown in liquid culture until the OD600 nm 5 ;1. Isopropyl b- D -1-thiogalactopyranoside was then added to 1mM and the culture incubatedat 28°C for 3 h to induce protein expression. Cultures were centrifuged at3,500 rpm for 10 min at 4°C and the cell pellet suspended in 7.5 mL ofsuspension buffer (50mMof HEPES at pH 7.5, 150mMof NaCl, 1%TritonX-100, 10% glycerol, and one Roche EDTA-FREE inhibitor cocktail tablet)plus 2.5 U/mL of Beconase. Cells were sonicated for 4 3 10 s with 20-sintervals on ice, and sonicates were centrifuged at 12,000g for 20 min at4°C. Protein purification was performed using Dynabeads His-tag mag-netic beads (Novex) or FLAG-MA beads (Sigma-Aldrich). Before use,beads were washed three times in 50 mM of HEPES at pH 7.5, 150 mM ofNaCl, and 10% glycerol. Sonicates were incubated with washed beads at4°C with rotation for at least 2 h. Beads were then washed three times withsuspension buffer and three times with wash buffer. Proteins were elutedwith elution buffer (13 PBS, 0.3 M of NaCl, 0.1% Tween-20, and 10%glycerol) containingeither300mMof Imidazole for LEC2-HISor200mg/mLof FLAG peptide (Sigma-Aldrich). Purified protein was quantified usingQubit reagents, buffer-exchanged with 50 mM of Tris-HCl at pH 8.0 and10% glycerol, and stored at 270°C in 20-mL aliquots.

Cell-Free Degradation Assay

One gram of 10 to 15 DPA seedlings of upl3-4 plants was extracted with1 mL of 50 mM of Tris-HCl at pH 7.5, 100 mM of NaCl, 10 mM of MgCl2,5 mM of DTT, and 5 mM of ATP, and centrifuged twice. Aliquots of 200 mLwere taken and 1mg of recombinant LEC2-HIS, and 5mL of purifiedUPL3-3FLAG or UPL3-mut-3FLAG from Nicotiana benthamiana were added.Reactions were performed at 22°C with or without 50 mM of MG132.Samples were taken at 0, 10, 30, 60, and 90 min and SDS sample bufferadded to stop reactions. Samples were denatured at 90°C for 10 min andelectrophoresed on 4% to 20% SDS-PAGE gels. Anti-HIS HRP antibodyvisualized LEC2-HIS protein levels.

Transient Expression in N. benthamiana

Constructs were transformed into A. tumefaciens GV3101 and 10-mLcultures grown at 28°C for 16 h before transfection into leaves of4-week–old N. benthamiana plants. MG132 (50 mM) in 10-mM MES-KOHat pH7.5 and 10mMofMgCl2was infiltrated into leaves 6 h before harvest.After 48h to60h, transfected leaveswereexcised, frozen in liquid nitrogen,and stored at270°C before extraction. Protein extracts were made using1 g of tissue/2 mL of extraction buffer. Extraction buffer was 10 mM of TrisHCl at pH 7.5, 150mMof NaCl, 0.5mMof EDTA, 0.5%Nonidet P-40, 10%glycerol, 1 mM of AESBF, and 13 EDTA-free protease cocktail (Roche),plus 50 mM of MG132. Samples were centrifuged at 4°C for 10 mins at15,000g, filtered through Miracloth (Merck Millipore) and used for affinitypurification. Where indicated, 20 mg of FLAG-purified FLAG-TR-TUBEproteinwasadded toeach5-mLextraction. Affinity purificationwascarriedat 4°C with rotation for 2 h. Pull-down experiments used co-expression ofUPL3-3FLAG with 3HA-LEC2, binding to FLAG-MA beads, washing withextraction buffer and elution with SDS-sample buffer. To detect LEC2 inplants total protein was extracted as described for total protein extractionfrom pooled 10 to 15 DPA siliques or leaves of 35SS:3HA-AtUPL3transgenic,Col-0,or lec2mutantplantsand immunoblottedwithanti-LEC2antibodies.

Varying UPL3 Expression Modulates Crop Yield 2383

Accession Numbers

Sequence data from this article can be found in the Arabidopsis GenomeInitiative or GenBank/European Molecular Biology Laboratory databasesunder the following accession numbers:

A. thaliana LEC2, At1G28300, accession OAP12706.1, UniProtQ1PFR7; A. thaliana UPL3, At4G38600, accession NP_001329354.1, UniProtQ6WWWW4; B. napus BnaUPL3.C03, BnaC03g60070D XP_013685717.1.BnaUPL3.A08, BnaA08g17000D-1, XP_022545849.1; and B. rapa BraUPL3Bra010737.1 (XM_009111293.2).

Supplemental Data

Supplemental Figure 1. Manhattan Plots for SWPP SNP and GEMAnalyses.

Supplemental Figure 2. Comparison of Predicted Protein Sequencesof the Arabidopsis UPL3 Gene and the Two B. napus Orthologs,BnaUPL3.C03 and BnaUPL3.A08.

Supplemental Figure 3. Expression of BnaUPL3.A08 in 45-DPASeeds in a Subset of B. napus GWAS Accessions Displaying HighVariation in SWPP.

Supplemental Figure 4. Segregation of a Leaf Hair Phenotype acrossB. napus Accessions Displaying Differential Expression of BnC03UPL3and High Variation in Seed Yield.

Supplemental Figure 5. RT-qPCR Analysis of T-DNA Insertion Allelesin AtUPL3.

Supplemental Figure 6. Predicted Sequence of B. rapa BraUPL3Protein.

Supplemental Figure 7. Seed Areas of B. rapa TILLING Lines.

Supplemental Figure 8. Promoter and 59UTR Variation Segregatingbetween B. napus GWAS Accessions Displaying Differential Expres-sion of BnUPL3.C03.

Supplemental Figure 9. Quantile–Quantile Plot of SWPP.

Supplemental Table 1. A List of B. napus Accessions and Trait DataUsed in GWAS and Associative Transcriptomics Analyses.

Supplemental Table 2. Segregation of SWPP and BnUPL3.C03Expression across a Subset of Accessions Used in GWAS.

Supplemental Table 3. Seed Fatty Acid Content and Composition inCol-0 and upl3-4 Mutant.

Supplemental Table 4. Primers Used.

ACKNOWLEDGMENTS

This work was supported by the Biotechnology and Biological SciencesCouncil (EuropeanResearchAreaNetwork forCoordinatingAction inPlantSciences, ERA-CAPS ABCEED grant, and strategic grants GRO BB/J004588/1 and GEN BB/P013511/1 to M.W.B.) and the French NationalResearchAgency (ANR-10-LABX-0040-SPStoL.L.).WethankJingkunMafor advice on transient expression and luciferase assays.

AUTHOR CONTRIBUTIONS

C.M., M.W.B., R.W., B.D., and L.L. designed the research; C.M., N.M.,R.W., J.B., M.W.B., T.C. and A.F. performed research; C.M. M.T. andM.W.B. analyzed data and wrote the article.

ReceivedAugust7,2018; revisedJuly22,2019;acceptedAugust12, 2019;published August 22, 2019.

REFERENCES

Andrianov, V., Borisjuk, N., Pogrebnyak, N., Brinker, A., Dixon, J.,Spitsin, S., Flynn, J., Matyszczuk, P., Andryszak, K., Laurelli, M.,Golovkin, M., and Koprowski, H. (2010). Tobacco as a productionplatform for biofuel: Overexpression of Arabidopsis DGAT and LEC2genes increases accumulation and shifts the composition of lipids ingreen biomass. Plant Biotechnol. J. 8: 277–287.

Badouin, H., et al. (2017) The sunflower genome provides insights into oilmetabolism, flowering and Asterid evolution. Nature 546: 148–152.

Bai, X., Huang, Y., Hu, Y., Liu, H., Zhang, B., Smaczniak, C., Hu, G.,Han, Z., and Xing, Y. (2017). Duplication of an upstream silencer of FZPincreases grain yield in rice. Nat. Plants 3: 885–893.

Barthole, G., To, A., Marchive, C., Brunaud, V., Soubigou-Taconnat, L., Berger, N., Dubreucq, B., Lepiniec, L., and Baud,S. (2014). MYB118 represses endosperm maturation in seeds of Arabi-dopsis. Plant Cell 26: 3519–3537.

Baud, S., et al. (2016) Deciphering the molecular mechanisms un-derpinning the transcriptional control of gene expression by mastertranscriptional regulators in Arabidopsis seed. Plant Physiol. 171:1099–1112.

Baud, S., Wuillème, S., To, A., Rochat, C., and Lepiniec, L. (2009).Role of WRINKLED1 in the transcriptional regulation of glycolytic andfatty acid biosynthetic genes in Arabidopsis. Plant J. 60: 933–947.

Bevan, M.W., Uauy, C., Wulff, B.B.H., Zhou, J., Krasileva, K., andClark, M.D. (2017). Genomic innovation for crop improvement. Nature543: 346–354.

Boulard, C., Fatihi, A., Lepiniec, L., and Dubreucq, B. (2017). Reg-ulation and evolution of the interaction of the seed B3 transcription fac-tors with NF-Y subunits. Biochim. Biophys. Acta. Gene Regul. Mech.1860: 1069–1078.

Chan, A., Carianopol, C., Tsai, A.Y.-L., Varatharajah, K., Chiu, R.S.,and Gazzarrini, S. (2017). SnRK1 phosphorylation of FUSCA3 posi-tively regulates embryogenesis, seed yield, and plant growth at hightemperature in Arabidopsis. J. Exp. Bot. 68: 4219–4231.

Clough, S.J., and Bent, A.F. (1998). Floral dip: A simplified method forAgrobacterium-mediated transformation of Arabidopsis thaliana. Plant J.16: 735–743.

Dong, H., et al. (2017) Ubiquitylation activates a peptidase that promotescleavage and destabilization of its activating E3 ligases and diversegrowth regulatory proteins to limit cell proliferation in Arabidopsis. GenesDev. 31: 197–208.

Downes, B.P., Stupar, R.M., Gingerich, D.J., and Vierstra, R.D.(2003). The HECT ubiquitin-protein ligase (UPL) family in Arabidopsis:UPL3 has a specific role in trichome development. Plant J. 35:729–742.

Duan, P., Xu, J., Zeng, D., Zhang, B., Geng, M., Zhang, G., Huang,K., Huang, L., Xu, R., Ge, S., Qian, Q., and Li, Y. (2017). Naturalvariation in the promoter of GSE5 contributes to grain size diversity inrice. Mol. Plant 10: 685–694.

Duong, S., Vonapartis, E., Li, C.-Y., Patel, S., and Gazzarrini, S.(2017). The E3 ligase ABI3-INTERACTING PROTEIN2 negatively regu-lates FUSCA3 and plays a role in cotyledon development in Arabidopsisthaliana. J. Exp. Bot. 68: 1555–1567.

El Refy, A., Perazza, D., Zekraoui, L., Valay, J.-G., Bechtold, N.,Brown, S., Hülskamp, M., Herzog, M., and Bonneville, J.-M.(2003). The Arabidopsis KAKTUS gene encodes a HECT protein andcontrols the number of endoreduplication cycles. Mol. Genet. Genomics270: 403–414.

2384 The Plant Cell

Fatihi, A., Boulard, C., Bouyer, D., Baud, S., Dubreucq, B., andLepiniec, L. (2016). Deciphering and modifying LAFL transcriptionalregulatory network in seed for improving yield and quality of storagecompounds. Plant Sci. 250: 198–204.

Harper, A.L., Trick, M., Higgins, J., Fraser, F., Clissold, L., Wells,R., Hattori, C., Werner, P., and Bancroft, I. (2012). Associativetranscriptomics of traits in the polyploid crop species Brassica napus.Nat. Biotechnol. 30: 798–802.

Kanai, M., Mano, S., Kondo, M., Hayashi, M., and Nishimura, M.(2016). Extension of oil biosynthesis during the mid-phase of seed de-velopment enhances oil content in Arabidopsis seeds. Plant Biotechnol.J. 14: 1241–1250.

Li, Y., Beisson, F., Pollard, M., and Ohlrogge, J. (2006). Oil content ofArabidopsis seeds: The influence of seed anatomy, light and plant-to-plant variation. Phytochemistry 67: 904–915.

Lin, Q., Ohashi, Y., Kato, M., Tsuge, T., Gu, H., Qu, L.-J., andAoyama, T. (2015). GLABRA2 directly suppresses basic helix-loop-helixtranscription factor genes with diverse functions in root hair development.Plant Cell 27: 2894–2906.

Liu, J., et al. (2017) GW5 acts in the brassinosteroid signalling pathway toregulate grain width and weight in rice. Nat. Plants 3: 17043.

Liu, Y.-F., et al. (2014) Soybean GmMYB73 promotes lipid accumulationin transgenic plants. BMC Plant Biol. 14: 73.

Lu, K., et al. (2017) Genome-wide association and transcriptome analysesreveal candidate genes underlying yield-determining traits in Brassicanapus. Front. Plant Sci. 8: 206.

Manan, S., Ahmad, M.Z., Zhang, G., Chen, B., Haq, B.U., Yang, J., andZhao, J. (2017). Soybean LEC2 regulates subsets of genes involved incontrolling the biosynthesis and catabolism of seed storage substancesand seed development. Front. Plant Sci. 8: 1604.

Mason, A.S., Higgins, E.E., Snowdon, R.J., Batley, J., Stein, A., Werner,C., and Parkin, I.A.P. (2017). A user guide to the Brassica 60K IlluminaInfiniumTM SNP genotyping array. Theor. Appl. Genet. 130: 621–633.

Maspero, E., Valentini, E., Mari, S., Cecatiello, V., Soffientini, P.,Pasqualato, S., and Polo, S. (2013). Structure of a ubiquitin-loadedHECT ligase reveals the molecular basis for catalytic priming. Nat. Struct.Mol. Biol. 20: 696–701.

McFarlane, H., Gendre, D., and Western, T. (2014). Seed coat ruthe-nium red staining assay. Bio Protoc. 4.

Napier, J.A., and Graham, I.A. (2010). Tailoring plant lipid composition:Designer oilseeds come of age. Curr. Opin. Plant Biol. 13: 330–337.

Nodine, M.D., and Bartel, D.P. (2010). MicroRNAs prevent precociousgene expression and enable pattern formation during plant embryogen-esis. Genes Dev. 24: 2678–2692.

Ohto, M.-A., Floyd, S.K., Fischer, R.L., Goldberg, R.B., and Harada,J.J. (2009). Effects of APETALA2 on embryo, endosperm, and seed coatdevelopment determine seed size in Arabidopsis. Sex. Plant Reprod. 22:277–289.

Patra, B., Pattanaik, S., and Yuan, L. (2013). Ubiquitin protein ligase 3mediates the proteasomal degradation of GLABROUS 3 and ENHANCEROF GLABROUS 3, regulators of trichome development and flavonoidbiosynthesis in Arabidopsis. Plant J. 74: 435–447.

Pfaffl, M.W. (2001). A newmathematical model for relative quantification inreal-time RT-PCR. Nucleic Acids Res. 29: e45.

Pritchard, J.K., Stephens, M., and Donnelly, P. (2000). Inference ofpopulation structure using multilocus genotype data. Genetics 155:945–959.

Ray, D.K., Mueller, N.D., West, P.C., and Foley, J.A. (2013). Yieldtrends are insufficient to double global crop production by 2050. PLoSOne 8: e66428.

Rodgers-Melnick, E., Vera, D.L., Bass, H.W., and Buckler, E.S.(2016). Open chromatin reveals the functional maize genome. Proc. Natl.Acad. Sci. USA 113: E3177–E3184.

Roscoe, T.T., Guilleminot, J., Bessoule, J.-J., Berger, F., andDevic, M. (2015). Complementation of seed maturation phenotypes byectopic expression of ABSCISIC ACID INSENSITIVE3, FUSCA3 andLEAFY COTYLEDON2 in Arabidopsis. Plant Cell Physiol. 56: 1215–1228.

Santos-Mendoza, M., Dubreucq, B., Baud, S., Parcy, F., Caboche,M., and Lepiniec, L. (2008). Deciphering gene regulatory networks thatcontrol seed development and maturation in Arabidopsis. Plant J. 54:608–620.

Santos Mendoza, M., Dubreucq, B., Miquel, M., Caboche, M., andLepiniec, L. (2005). LEAFY COTYLEDON 2 activation is sufficient totrigger the accumulation of oil and seed specific mRNAs in Arabidopsisleaves. FEBS Lett. 579: 4666–4670.

Schneider, C.A., Rasband, W.S., and Eliceiri, K.W. (2012). NIH Image toImageJ: 25 years of image analysis. Nat. Methods 9: 671–675.

Shi, L., Katavic, V., Yu, Y., Kunst, L., and Haughn, G. (2012). Arabi-dopsis glabra2 mutant seeds deficient in mucilage biosynthesis producemore oil. Plant J. 69: 37–46.

Singh, R., et al. (2013) Oil palm genome sequence reveals divergence ofinterfertile species in Old and New worlds. Nature 500: 335–339.

Song, S.-K., Ryu, K.H., Kang, Y.H., Song, J.H., Cho, Y.-H., Yoo, S.-D.,Schiefelbein, J., and Lee, M.M. (2011). Cell fate in the Arabidopsis rootepidermis is determined by competition between WEREWOLF and CA-PRICE. Plant Physiol. 157: 1196–1208.

Stephenson, P., Baker, D., Girin, T., Perez, A., Amoah, S., King,G.J., and Østergaard, L. (2010). A rich TILLING resource for studyinggene function in Brassica rapa. BMC Plant Biol. 10: 62.

Stone, S.L., Kwong, L.W., Yee, K.M., Pelletier, J., Lepiniec, L.,Fischer, R.L., Goldberg, R.B., and Harada, J.J. (2001). LEAFYCOTYLEDON2 encodes a B3 domain transcription factor that inducesembryo development. Proc. Natl. Acad. Sci. USA 98: 11806–11811.

Tsai, A.Y.-L., and Gazzarrini, S. (2012). AKIN10 and FUSCA3 interact tocontrol lateral organ development and phase transitions in Arabidopsis.Plant J. 69: 809–821.

van Erp, H., Kelly, A.A., Menard, G., and Eastmond, P.J. (2014).Multigene engineering of triacylglycerol metabolism boosts seed oilcontent in Arabidopsis. Plant Physiol. 165: 30–36.

Wang, J., Liu, H., and Ren, G. (2014). Near-infrared spectroscopy (NIRS)evaluation and regional analysis of Chinese faba bean (Vicia faba L.). CropJ. 2: 28–37.

Wang, S., et al. (2015) The OsSPL16-GW7 regulatory module determinesgrain shape and simultaneously improves rice yield and grain quality. Nat.Genet. 47: 949–954.

Willmann, M.R., Mehalick, A.J., Packer, R.L., and Jenik, P.D. (2011).MicroRNAs regulate the timing of embryo maturation in Arabidopsis.Plant Physiol. 155: 1871–1884.

Wu, F.-H., Shen, S.-C., Lee, L.-Y., Lee, S.-H., Chan, M.-T., and Lin,C.-S. (2009). Tape-Arabidopsis Sandwich—a simpler Arabidopsis pro-toplast isolation method. Plant Methods 5: 16.

Zhai, Z., Liu, H., and Shanklin, J. (2017). Phosphorylation of WRINKLED1by KIN10 results in its proteasomal degradation, providing a link betweenenergy homeostasis and lipid biosynthesis. Plant Cell 29: 871–889.

Zhang, X., Garreton, V., and Chua, N.-H. (2005). The AIP2 E3 ligaseacts as a novel negative regulator of ABA signaling by promoting ABI3degradation. Genes Dev. 19: 1532–1543.

Zhou, Y., et al. (2013) HISTONE DEACETYLASE19 interacts with HSL1and participates in the repression of seed maturation genes in Arabi-dopsis seedlings. Plant Cell 25: 134–148.

Varying UPL3 Expression Modulates Crop Yield 2385

DOI 10.1105/tpc.18.00577; originally published online August 22, 2019; 2019;31;2370-2385Plant Cell

Dubreucq, Thierry Chardot, Loic Lepiniec and Michael W. BevanCharlotte Miller, Rachel Wells, Neil McKenzie, Martin Trick, Joshua Ball, Abdelhak Fatihi, Bertrand

Brassica napusCrop Yields in Modulates LEC2 Levels, Seed Size, andUPL3Variation in Expression of the HECT E3 Ligase

 This information is current as of February 29, 2020

 

Supplemental Data /content/suppl/2019/08/22/tpc.18.00577.DC1.html

References /content/31/10/2370.full.html#ref-list-1

This article cites 52 articles, 14 of which can be accessed free at:

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