A BIN2-like1 Protein Confers Dwarf by Interacting with BZR1 of Brassinosteroid Signaling in Allotetraploid Brassica Napus

Brassinosteroids (BRs) are steroid hormones that play essential roles in plant growth and development. In this study, we identied a new dwarf mutant in Brassica napus. By map-based cloning, BnaC04.BIL1 (BnaC04g41660D) gene, a BIN2-like1 (BIL1) encoding a GLYCOGEN SYNTHASE KINASE 3 (GSK3-like) protein kinase, was isolated. To date, how BIL1 involves in BR signal transduction remains uncovered. Genetic transformation experiments conrmed that the BnaC04.BIL1 is responsible for the plant dwarf phenotype in the Bndwarf2 mutants. Overexpression of BnaC04.BIL1 not only reduced plant height, but also resulted in compact plant architecture. Using CRISPR/Cas9, two sgRNAs were designed to target BnaC04.BIL1 gene. The gene editing experiments generated mutations of BnaC04.BIL1 sequence, which were stably transmitted to successive generations, and lead to restoration of plant height and plant architecture. The molecular mechanism of Bndwarf2 dwarng was further veried by Y2H and BiFC assays. Results shown that a Thr187Ser amino acid substitution residing in the conserved region promotes the interaction between BnaC04.BIL1 Mut with BnaBZR1, thus enhances the negative regulation of plant growth. The genetic and molecular evidence claries rst the BnaC04.BIL1 can sharply change plant architecture in natural plant accessions in allotetraploid, and provides new insights into the molecular mechanisms of BR signaling.


Introduction
Oilseed rape (Brassica napus L.) is one of the most important oil crops worldwide, and provides highquality edible oil for human diets, protein-rich feed for animals, and raw materials for industrial processes. Breeding its cultivars with dwarf or semi-dwarf phenotype is a major objective in the genetic improvement because dwar ng architecture can be helpful to increase harvest index and enhance lodging resistance (Hedden 2003). To nd available germplasm or genes associated with dwar ng plant type for B. napus breeding, some efforts have been carried out. For example, the dwarfness-associated genes in B. napus, including DS-1 (Liu et al. 2010), ndf-1 ), DS-3 (Zhao et al. 2017, DS-4 (Zhao et al. 2019), G7 (Cheng et al. 2019), BnaDwf.C9 (Wang et al. 2020), have been positioned or identi ed. Additionally, the Bndwf1 was ne-mapped on the A9 chromosome to a 152-kb interval (Wang et al. 2016a). However, the molecular mechanism(s) underlying the development of the dwarf phenotype in B. napus remain elusive for ideal plant type breeding due to absence of successfully applied cultivar in vast oilseed rape production region.
Among these genes, glycogen synthase kinase-3 (GSK3)-like kinase BIN2 is a key suppressor that regulates plant growth and development by determining the phosphorylation status of BES1 and BZR1 (Choe et al. 2002;Li and Nam 2002;Wang et al. 2002;Yin et al. 2002). GSK3-like kinases are a highly conserved Ser/Thr kinases that are implicated in a wide range of cellular and developmental processes (Woodgett 2001). In Arabidopsis, the GSK3/SHAGGY-like gene family has 10 gene members that can be classi ed into four subgroups (Jonak and Hirt 2002). In this family, the Arabidopsis GSK3-like kinase (AT4G18710, BIN2/UCU1/DWF12/AtSK21) which belongs to the group II, has activity to negatively regulate the BR signal transduction by phosphorylating BZR1/BES1 (Choe et al. 2002;He et al. 2002;Li and Nam 2002). The gain-of-function bin2 mutant was discovered to be insensitive to BRs in Arabidopsis, and has the shaggy phenotypic characteristic of dwar ng architecture. It also confers curved leaves, and an impaired cell elongation (Perez-Perez et al. 2002). The coding sequence of the BIN2 gene, substitutes consecutive glutamate residues in the highly conserved TREE domain, which results in the negatively regulating growth by phosphorylating the BES1 and BZR1 proteins, that result in the degradation of BZR1 to reduce its activity . Based on sequence similarity of BIN2 with its two closest group II Arabidopsis homologs, BIN2-Like1 (BIL1) and BIN2-Like2 (BIL2), which belong to the AtSKs group (Jonak and Hirt 2002). It was further suggested that BIL1 and BIL2 may also be involved in BR signaling.
Overexpression of BIL1 or BIL2 gene driven by their native promoters in wild-type Arabidopsis plants exhibits the dwarf phenotype (Yan et al. 2009). However, the evidence of BIL1 and BIL2 genes involved in BR signal transduction is still insu cient, and the mechanism of plant dwarf phenotype caused by overexpression of BIL1 and BIL2 genes remains to be elucidated. Therefore, it is urgent to further explore their participation and even related mechanism.
In this study, a pure dwarf mutant, Bndwarf2, was found in advanced sel ng generation in a nearly pure line CB1501-1 in B. napus. To expedite this study, the dwarf gene BnaC04.BIL1 was isolated using mapbased cloning. The BnaC04.BIL1 gene encoding a GSK3-like kinase, belongs to GSK II subfamily. Genetic transformation experiments con rmed that the BnaC04.BIL1 was responsible for the plant dwarf phenotype in the Bndwarf2 mutants. The molecular mechanism of Bndwarf2 dwar ng was veri ed by Y2H and BiFC assays demonstrating that a Thr187Ser amino acid substitution residing in the highly conserved region of BnaC04.BIL1 promoted the interaction between BnaC04.BIL1 Mut with BnaBZR1.
Overall, this study clari es the role of BnaC04.BIL1 in the regulation of plant height, which may help to improve lodging resistance in oilseed rape, and therefore provides new insights into the molecular mechanisms of BR signaling in allotetraploid.

Plant materials
A pure dwarf mutant, Bndwarf2 was found in advanced sel ng generation in a nearly pure line CB1501-1 in B. napus by our lab. The populations for mapping the BnDWARF2 locus, were generated from the crosses between Bndwarf2 and the canola variety Zhpngshuang 11 (ZS11). All oilseed rape materials were grown in growth chamber and the elds of the Jiangpu Agricultural Experimental Station at Nanjing Agricultural University.
Tobacco was grown in growth chamber. The illumination period was 14 hours with temperature at 26 o C and 10 hours with temperature at 20 o C. When tobacco leaves at 5-leaf stage were used for the gene subcellular localization and bimolecular uorescence complementary assay (BiFC).
Map-based cloning SNP and SSR markers were used to map the dwarf gene. 70 dwarf plants, 24 tall plants and parents from F 2 population were genotyped using a Brassica 60 K SNP Bead Chip Array (Illumina, Inc), which have a total of 52,157 SNP markers. The SNP genetic map was constructed by JoinMap 4.1 mapping software (Ooijen et al. 2006), then the BnDWARF2 locus was primarily mapped onto physical and genetic map. The mapping interval sequence was downloaded from the Brassica napus Genome Browser (http://www.genoscope.cns.fr/brassicanapus/cgi-bin/gbrowse/ colza/). Using this genomic sequence, SSR marker primers were designed by aid of SSR Hunter 1.3 (Li et al. 2005), and Primer Premier 5.0 (Singh et al. 1998). A total of 318 polymorphic SSR markers were obtained. These SSR markers helped to ne-map the BnDWARF2 locus using a size-enlarged populations comprised of F 2:3 plants.
To identify genes associated to the dwarf trait, sequence of the ne mapping interval was obtained from the Brassica napus Genome Browser for reference to next-step experiments. Then, all of the genes in the ne mapping interval were cloned from Bndwarf2 and parent ZS11. And, the resulting sequences were aligned using ClustalX 1.83 and GeneDoc software. The speci c primers of the genes are listed in Table   S6.

Sequence analysis
The B. napus BIL1 genes were obtained by screening the B. napus Genome Browser (http://www.genoscope.cns.fr/brassicanapus/) with known A. thaliana BIL1 gene as a query. The Conserved Domain Database was used to search the protein functional in the National Center for Biotechnology Information (NCBI) (http://www.ncbi.nlm.nih.gov). Predicted A. thaliana BIL1 amino acid sequences were obtained from the TAIR website (http://www.arabidopsis.org/Blast). Moreover, the protein sequences of other species were obtained from the NCBI using the A. thaliana BIL1 protein sequence as a query. All obtained protein sequences were aligned using ClustalX 1.83 (Crooks et al. 2004). Additionally, a phylogenetic tree was constructed using MEGA 7.0 (Kumar et al. 2016) with maximum likelihood method, and the bootstrap values were estimated with 1000 replicates. The functional protein association network was analyzed by STRING (https://string-db.org/cgi/).

RNA extraction and qRT-PCR
Total RNA was extracted from various samples using TRIzol reagent (Sigma; http://www.sigmaaldrich.com/). First-strand cDNA synthesis was carried out using a Reverse Transcription System (Takara, Tokyo, Japan). The cDNA was used as the template for qRT-PCR analysis with speci c primers (Table S6). The qRT-PCRs were carried out with SYBR Green Real-time PCR Master mix using a CFX96-2 PCR machine (BIO-RAD, USA). Relative expression levels were calculated using the 2 −ΔΔCt method with Actin as an internal control.

Subcellular localization
The coding sequence of BnaC04.BIL1 (BnaC04g41660D) and BnaBZR1 (BnaCnng34980D) were ampli ed using two pairs of primers GFP-BnaC04.BIL1-F/R and GFP-BnaBZR1-F/R (Table S6) and inserted into pA7-GFP entry vector. The recombinant plasmid BnaC04.BIL1-GFP, BnaBZR1-GFP, and the empty vector pA7-GFP were introduced into tobacco leaf cells by the particle bombardment method. The tobacco leaf cells were then bombarded by PDS-1000/He (Bio-Rad, USA), with an 1100 psi split membrane and gold particles coated with the plasmid DNA. After bombardment, the tobacco leaf cells were incubated on MS medium in a dark chamber at 28 o C for 16 h. Fluorescence was observed using a LSM780 confocal microscopy imaging system (Zeiss, Germany).

Plant transformation
The 1223-bp BnaC04.BIL1 open reading frame was ampli ed from Bndwarf2 using the primers BnaC04.BIL1-F/R (Table S6) and cloned into the Xba I-BamH I sites of the overexpression pBI121 vector with CaMV35S promotor to construct the 35S::BnaC04.BIL1-pBI121 plasmid. The plasmid was introduced into Agrobacterium tumefaciens strain EHA105 by a heat shock method. The positive A. tumefaciens were transformed into ZS11 as previously described (Tan et al. 2011). Seeds were collected, followed by screening of the transgenic plants.

CRISPR/Cas9 target locus selection and construct assembly
In order to target BnaC04.BIL1, the sequence-speci c sgRNAs were designed using the web-based tool CRISPR-P 2.0 (http://crispr.hzau.edu.cn/CRISPR2/). The S1 target sites in ORF 5' terminal and the S2 in the conserved region for BnaC04.BIL1 were selected, and the speci c sgRNA primers are listed in Table  S6. The two targets were assessed using PCR and Sanger sequencing of Bndwarf2 to ensure that no polymorphisms existed between the sgRNAs and the corresponding target sequences.
The binary pYLCRIPSR/Cas9 multiplex genome targeting vector system (pYLCRISPR/Cas9P 35S-H), which was provided by Prof. Yaoguang Liu (South China Agriculture University), in which Cas9p is driven by the cauli ower mosaic virus 35S promoter (P35S), and four plasmids with sgRNA cassettes driven by the promoters of AtU3b, AtU3d, AtU6-1, and AtU6-29. This system was used for construct assembly according to a method previously described by Ma et al. (2015). The resulting constructs contained a Cas9p expression cassette, sgRNA expression cassettes with target sequences.

Bimolecular Fluorescence Complementarity (BiFC) assay
For the BiFC analysis, the cDNA of BnaBZR1 was cloned into BamH I site of pSPYNE vector, and BnaC04.BIL1 Mut and BnaC04.BIL1 WT were cloned into BamH I site of pSPYCE vector to generate the BnaBZR1-YFPn, BnaC04.BIL1 Mut -YFPc, and BnaC04.BIL1 WT -YFPc fusion proteins, respectively (Waddt et al. 2008). The plasmids were transformed into Agrobacterium tumefaciens strain GV3101. A. tumefaciens cells containing each construct were prepared and mixed to an OD 600 of 0.6:0.6 and transformed into tobacco leaf cells. YFP uorescence was observed using the LSM780 confocal microscopy imaging system (Zeiss, German) 40 to 48 h after in ltration. The PCR primers used for the BiFC assays are listed in Table S6.

RNA-sequencing analysis
The Bndwarf2 mutants were evaluated by RNA sequencing, with the ZS11 plants serving as the control. Seedlings were grown in pots containing peat: vermiculite (1:1 v/v). Because the Bndwarf2 showed wrinkled leaves, shorter petioles, and shorter hypocotyls, the whole plants of 6-week-old were harvested, and immediately frozen in liquid nitrogen, then stored at -80 °C until analyzed. Total RNA was extracted using Trizol reagent kit (Invitrogen, Carlsbad, CA, USA) according to the manufacturer's protocol. The RNA was used to construct sequencing libraries that were analyzed with the HiSeq 2500 platform (Illumina, San Diego, CA). Paired-end clean reads were aligned to the B. napus "Darmor-bzh" reference genome (Chalhoub et al. 2014) using HISAT2.2.4 (Kim et al. 2015). For each transcription region, a FPKM (fragment per kilobase of transcript per million mapped reads) value was calculated with StringTie v1.3.1 (Pertea et al. 2015). Differential expression analysis of Bndwarf2 and ZS11 (three biological replicates per sample) were performed using DESeq2 software (Love et al. 2014). Genes with a false discovery rate < 0.005 and absolute fold change ≥2 were considered DEGs.
The GO annotation of DEGs was performed using the Gene Ontology database (http://www.geneontology.org/), and GO terms with corrected P < 0.05 were considered to be signi cantly enriched. In addition, DEGs were analyzed using the KEGG database (Kanehisa et al. 2000), and KEGG enrichment pathways of DEGs were determined using edge R package (Robinson et al. 2010).

Results
Characterization of the Bndwarf2 mutant A pure dwarf mutant, Bndwarf2 was obtained in advanced sel ng generation in a nearly pure line CB1501-1 in B. napus. The Bndwarf2 mutant showed an obvious dwarf and etiolated phenotype after the germination in the dark at 24 o C for 6 d in comparison with the cultivar Zhongshuang 11 (ZS11, a conventional B. napus cultivar) that was used as a parent for map-based cloning gene responsible for the dwar sm, or as wild type (Fig. 1a). At seedling stage, the Bndwarf2 mutant plants had shorter hypocotyls and shorter petioles (Figs. 1b, c). The leaves of Bndwarf2 mutants showed darker green, thickened, and wrinkled leaves, and had signi cant higher Chl a, Chl b, and Chl contents than those of ZS11 (Table S1). At owering stage, the Bndwarf2 mutant showed signi cant difference in plant height from ZS11 (Fig.  S1). While at maturity stage, the Bndwarf2 mutant showed dwarf stature (33.62 ± 1.12 cm) with no apical dominance, that was signi cantly lower than that for ZS11 (193.54 ± 4.80 cm) (Fig. 1d). The siliques of Bndwarf2 mutants were signi cantly shorter compared to that of ZS11 (Fig. 1e). In addition, the Bndwarf2 mutants had smaller seeds, lower 1000-seeds weight and compact plant architecture (Fig. 1f, Table S2). The F 1 plants (105.30 ± 5.16 cm) generated by cross of ZS11 with Bndwarf2 were in-between that of ZS11 and Bndwarf2. The observed multiple morphological abnormal of Bndwarf2 mutants showed similar characteristics in appearance to the other BR-related mutants, such as bri1 (Clouse et al. 1996), dwf12 (Choe et al. 2002), bin2 , and ucu1 (Perez-Perez et al. 2002).

Map-based cloning
To investigate the genetic regulation mechanism for Bndwarf2, the F 1 (ZS11 × Bndwarf2) and RF 1 (Bndwarf2 × ZS11) plants were obtained by crossing Bndwarf2 with ZS11, all had the dwarf trait, indicating that dwarf trait was controlled by dominant genes. The phenotypic segregation ratio of dwarf plants to tall plants in the F 2 population was in a Mendelian model of 3:1 (209 dwarf plants vs. 78 tall plants, < 0.05 ). Among 289 BC 1 individuals, 139 as dwarf types and 150 as tall types, also approximately tted an expected Mendelian inheritance ratio of 1:1 (dwarf plants vs. tall plants). In subsequent segregating F 2:3 populations, the genetic regulation was con rmed (Table S3). These results indicated that the dwarf trait was controlled by a dominant nuclear gene, which was named as BnDWARF2 in the subsequent study.
To map BnDWARF2, 94 plants (70 dwarf plants and 24 tall plants) from the F 2 population were used for single nucleotide polymorphism (SNP) marker genotyping. Although the chip (Illumina, Inc) has 52,157 SNP markers, only 7457 polymorphic markers were used to construct the SNP genetic linkage map after removing the invalid markers. The BnDWARF2 locus was located on C04 chromosome between the SNP marker M33367 and M35244 (Fig. 2a). To ne map the BnDWARF2 locus, 318 primer pairs of simple sequence repeat (SSR) markers were designed to uniformly cover the preliminary mapping interval. A further 889 individuals from the F 2:3 populations, nally narrowed down the BnDWARF2 locus to a 34.62kb region between SSR markers S3 and S4 (Fig. 2b). No other markers to further narrow the mapping interval were found for this mapping population and its parents. A total of 5 putative genes were localized in the 34.62-kb region according to the gene annotation of the B. napus reference genome (Fig. 2c). Sequence cloning was performed for the mapping interval, and the results showed that only BnaC04g41660D (BnaC04.BIL1) gene had 10 SNPs differences between ZS11 and Bndwarf2. The BnaC04.BIL1 had two amino acid residues substitutions at aa-187 (Thr-to-Ser mutation, named Thr187Ser) and aa-399 (Gln-to-His mutation, named Gln399His) (Fig. 2e).
Sequence analysis of BnaC04.BIL1 BnaC04.BIL1 contains a 1233-bp open reading frame (ORF) with 11 introns and has 3 copy genes in B. napus (Fig. 2d, Fig. S2). BnaC04.BIL1 is a homologous gene of the Arabidopsis AT2G30980 gene, which encodes a GSK3-like protein kinase (Charrier et al. 2002). The conservative domain analysis showed that the amino acid sequence 65-357 was the conserved domain of STKc_GSK3, and Thr187Ser is in the conserved domain (Fig. 2e). The amino acid multiple sequence analysis showed that BnaC04.BIL1 gene had a series of amino acid residues conserved in GSK3 kinase, such as GSK3 domain signature SYICSR and plant-speci c TREE motif (Fig. S2). It was perfectly aligned with the genes for GSK3/Shaggy kinases with regarding to a series of amino acid residues such as the GSK3 signature SYICSR within domain VIII that was absent from MAP kinase sequences (Dornelas et al. 1999). The E-K mutation in the highly conserved TREE motif is thought to preventing the BR-mediated BIN2 inhibition (Peng and Li 2003), thus resulting in the increased BIN2 stability (Peng et al. 2008;Vert and Chory 2006). The phylogenetic tree clustering and construction were analyzed by MEGA 7.0 selection Neighbor-joining method. The results showed that BnaC04.BIL1 gene and ArabidopsisBIN2 gene were homologous, belonging to GSK3 II subfamily (Fig. S2). These indicated that BnaC04.BIL1 protein might also be interacting with BnaBZR1 to negatively regulate BR signaling transduction. Further bioinformatics analysis suggested that the protein interaction network of BnaC04.BIL1 showed that the BnaC04.BIL1 has binding domains with other proteins between aa-20 and aa-369 (Fig. S2), which was consistent with previous reports (Kim et al. 2009). The Gln399His mutation is not in the conserved domain, but in C-terminal. Accordingly, it suggested that the Thr187Ser mutation may be responsible for affecting the interaction with BnaBZR1.
Expression patterns of BnaC04.BIL1 and the subcellular localization To explore the possible function of BnaC04.BIL1 gene from Bndwarf2 mutant in different tissues, the transcription levels of BnaC04.BIL1 in leaves, roots, hypocotyl, stems, buds, owers, siliques, and seeds were analyzed. The qRT-PCR analysis showed that the BnaC04.BIL1 gene was expressed in all tissues, which indicated that BnaC04.BIL1 expressed constitutively (Fig. 3a). The expression level of BnaC04.BIL1 was higher in leaves, hypocotyls, siliques, and seeds, while its level in buds and stems were lower.
Previous research showed that Arabidopsis BIN2 (Ryu et al. 2010) and rice BZR1 (Bai et al. 2007) were localized in the nucleus. To de ne the subcellular location of expression, pA7-GFP, BnaC04.BIL1-GFP, and BnaBZR1-GFP constructs were then introduced into the tobacco leaf cells by the particle bombardment method. The merged image of BnaC04.BIL1-GFP and nuclear localization signal (NLS)-mCherry signals showed that BnaC04.BIL1 was localized to the nucleus, with the merged image of BnaBZR1-GFP showing that BnaBZR1 was localized in the nucleus and cell membrane (Fig. 3b) Table S4). At the seedling stage, the OE-BIL1 transgenic lines displayed darker green and wrinkled leaves compared to those of ZS11 (Fig. 4b). These results suggest that the BnaC04.BIL1 gene not only controls the plant height, but also regulates the seed size. It follows that, the yield of per OE-BIL1 transgenic plants showed a signi cantly reduction compared to that of ZS11 (Table S4). The T 2 progeny plants were examined from six T 1 transgenic lines in growth chamber, which showed the expected Mendelian inheritance ratio of 3:1 in T 2 progeny (dwarf vs. tall plants, < 0.05 , 1 = 3.84; P > 0.05; Table S5). The T 2 progeny plants displayed perfect co-segregation between the transgene and the dwarf phenotype. Consistently, the expressions of BnaC04.BIL1 gene in homozygous T 3 lines (OE-BnaC04.BIL1 transgenic genes) were signi cantly higher than those of ZS11 plants (Fig. 4c). These results con rmed that the BnaC04.BIL1 is the causal mutation for the dwar sm and controls smaller seeds, which were also observed in Bndwarf2.

Knockout of BnaC04.BIL1 restored plant height
An CRISPR-BnaC04.BIL1 (CR-BIL1) construct for BnaC04.BIL1 was transformed into Bndwarf2 plants by using Agrobacterium-mediated transformation. According to a PCR examination using construct-speci c primers (Table S6), some T 1 -positive transgenic lines were selected for veri cation by Sanger DNA sequencing of the target sites. The T 1 -positive plants displayed the expected tall phenotype within apical dominance at maturity (Fig. 5a), suggesting that the plant height of Bndwarf2 were restored (Table S4).
The transgenic CR-BIL1 plants showed dramatically larger seeds than the Bndwarf2 mutants (Fig. 5b). These results further veri ed that the BnaC04.BIL1 was responsible for the plant dwarf phenotype.
To obtain stable mutant lines, nine independent T 1 editing lines of BnaC04.BIL1 were self-pollinated to produce T 2 and T 3 progeny. Genetic analysis of the T 2 progeny plants from four positive plants was also performed, showing the expected Mendelian inheritance ratio of 1:3 in T 2 progeny (dwarf vs. tall plants; Table S5). The targeted mutations of progeny from these T 1 lines were further veri ed with a direct sequence analysis of the PCR products of the target sites. Sequencing of the mutated region revealed that various mutations, including mutation and deletion of different nucleotides, were produced in proximity to target sites in all selected lines. A total of nine T 3 lines with homozygous mutations in BnaC04.BIL1 were detected, including three S1 target (in ORF 5' terminal) mutants and six S2 target (in the conserved region) mutants (Fig. 5c). All of these homozygous mutations were detected at the target sites within BnC04.BIL1 as predicted, and caused frameshifts that resulted in non-functional proteins. These results suggested that the loss-of-function of BnaC04.BIL1 leads to the restored plant height of Bndwarf2.

BnaC04.BIL1 interacts with BnaBZR1
In order to investigate whether the mutation(s) in conserved domain of BnaC04.BIL1 promotes or attenuates the interaction with BnaBZR1 protein, we tested the effect of the Thr187Ser substitution on the interaction between BnaC04.BIL1 and BnaBZR1 using yeast two-hybrid (Y2H) assays. The Y2H assays showed that BnaC04.BIL1 Mut could interact with BnaBZR1, but BnaC04.BIL1 WT cannot interact with BnaBZR1 (Fig. 6a). As expected, these results clearly suggested that the Thr187Ser amino acid substitution residing in the conserved region affected the interaction between BnaC04.BIL1 with BnaBZR1.
For further veri cation, bimolecular uorescence complementation (BiFC) analysis was used to determine whether BnaC04.BIL1 interacts with BnaBZR1 in vivo. In Nicotiana benthamiana leaf epidermal cells coexpressing the C-terminal half of yellow uorescent protein (YFP) fused to BnaC04.BIL1 Mut or BnaC04.BIL1 WT , and the N-terminal half of YFP fused to BnaBZR1. We observed that the strong YFP uorescence appeared in the cell membrane and nucleus when BnaC04.BIL1 Mut interacts with BnaBZR1 (Fig. 6b), whereas other combinations had none YFP uorescence in cells. These results further con rmed that unlike the response of BnaC04.BIL1 WT , BnaC04.BIL1 Mut could interact with BnaBZR1.

Transcriptome analysis
To further explore the possible mechanism underlying dwarf and determine whether other metabolic pathways were also in uenced in the Bndwarf2 mutant, differentially expressed genes (DEGs) between Bndwarf2 plants and ZS11 plants were detected using RNA sequencing technology. A total of 365,670,572 clean reads from six samples (three duplications for Bndwarf2 and ZS11) were obtained (Table S7). Ultimately, a total of 7,270 DEGs were isolated from the Bndwarf2 and ZS11 plants.
Compared with the ZS11 lines, 3,625 genes were up-regulated and 3,645 genes were down-regulated in the Bndwarf2 mutants (Table S8). GO analysis showed that the biological process category contained 21 terms, and molecular function category contained 12 terms and cellular component contained 12 (Table  S9). In the biological process category, the top three terms were in the metabolic, cellular, and singleorganism processes. In the molecular function category, the top three terms were binding, catalytic activity, and transporter activity. In the cellular component category, the top three terms were cell, cell part, and membrane (Fig. S3). These results indicated that Bndwarf2 mutation seriously affected the plants cell elongation and development.
Furthermore, Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analysis showed that a total of 129 pathways were affected in the Bndwarf2, and particularly, the biosynthesis and signal transduction of BR seemed to be destroyed in dwarf plants (Table S10). The related DEGs included 15 genes related to the BR biosynthesis pathway (Table S11) and 6 genes related to the BR signal transduction (Table S12), all of which exhibited up-regulated expression in Bndwarf2 plants, compared to ZS11. For example, three BnaBZR1 genes were increased (Table S12), which can affect plant growth and development in various aspects through the regulating downstream genes expression.

Discussion
Plant type improvement is a major goal for crop breeding, where plant height is an important growth habit that is a fundamental yield determining trait in crops. In the 1960s and 1970s, the dwarf trait genes (Rht1 and sd1) were introduced into wheat and rice that were crucial to the rst "Green Revolution" (Hedden 2003;Khush 2001). The semi-dwarf genes result in a shortened culm with improved lodging resistance and a greater harvest index (Islam and Evans 1994). However, there are few studies with respect to dwarf oilseed rape. Because of the lower mechanization level of oilseed rape production and few varieties suitable for mechanization harvest, oilseed rape production faces severe challenge.
Most of our knowledge about BIN2 functions came mostly from gain-of-function results. For example, genetic screening in Arabidopsis for BR-insensitive dwarf mutants resulted in the isolation of eight gainof-function bin2 alleles (Choe et al. 2002;Li and Nam 2002;Perez-Perez et al. 2002). However, no single gain-of-function mutation for BIL1 or BIL2, two closest homologs of BIN2 encoding group II AtSKs, has been isolated. Although previous report found that transgenic Arabidopsis plants overexpressing BIL1 or BIL2 gene confer the dwarf phenotype (Yan et al. 2009), it remains unknown whether BIL1 or BIL2 might also participate in BR signaling, similar to the function of BIN2 (Choe et al. 2002;Perez-Perez et al. 2002). In our study, a gain-of-function mutation for BIL1 in oilseed rape has been discovered, and most importantly, it exhibits the BR-insensitive dwarf phenotype. For example, the Bndwarf2 mutant displayed the BR signaling phenotypes: shorter hypocotyls, shorter petioles, wrinkled leaves, smaller seeds, and obvious dwarf compared with the ZS11 (Fig. 1, Fig. S1, Tables S1, S2). These characteristics were similar to the phenotypes of BR-insensitive mutants such as bri1 (Clouse et al. 1996), dwf12 (Choe et al. 2002), anducu1 (Perez-Perez et al. 2002). Through map-based cloning, the BnaC04.BIL1 was identi ed to be a BIN2-Like1 (BIL1), showing a Thr187Ser amino acid substitution residing in the conserved region (Fig. 2,  Fig. S2). Genetic transformation experiments con rmed that the BnaC04.BIL1 was responsible for the plant dwarf phenotype in the Bndwarf2 mutants. Overexpression of BnaC04.BIL1 under the background of ZS11 reduced plant height compared with ZS11 (Fig. 4, Table S4). This result was consistent with previous reports, showing that overexpressing BIL1 gene confers the dwarf phenotype in Arabidopsis (Yan et al. 2009). Meanwhile, a CRISPR/Cas9 vector knockdown of BnaC04.BIL1 showed that the plant height of Bndwarf2 was restored (Fig. 5, Table S4). The genetic evidence clari es the BnaC04.BIL1 can sharply change plant architecture in natural plant accessions in allotetraploid.
Furthermore, our study provides evidence that BnaC04.BIL1 can interact with BnaBZR1, thus negatively regulating BR signaling in allotetraploid Brassica napus, similar to the function of BIN2 on BR signaling in Arabidopsis (Choe et al. 2002;Li and Nam 2002). First, the protein interaction network of BnaC04.BIL1 showed that the BnaC04.BIL1 has binding domains with other proteins between aa-20 and aa-369 (Fig.   S2), which was consistent with previous reports, showing the BSU1 interacts with BIN2 on Tyr200 in Arabidopsis (Kim et al. 2009;Kim et al. 2017). Second, the subcellular localization analysis demonstrated that BnaC04.BIL1 exists in the nucleus (Fig. 3). Consistently, the Arabidopsis BIN2 functioned in nucleus to negatively regulate BR signaling (Ryu et al. 2010). Third, the biosynthesis and signal transduction of BR were destroyed in dwarf plants (Table S10, Table S11, Table S12). And, since the BnaBZR1 genes were increased in Bndwarf2 mutants (Table S12), we further speculated that these genes can affect plant growth and development in various aspects through the regulating downstream genes expression Wang et al. 2012;Nolan et al. 2018). In fact, previous results revealed that many genes regulated by BZR1 and/or BES1, and some proteins interacting with BZR1/BES1, were closely associated with the BR signaling (Guo et al. 2013;Li 2010). Fourth, Y2H and BiFC assays con rmed that the Thr187Ser amino acid substitution residing in the conserved region affected the interaction between BnaC04.BIL1 with BnaBZR1 (Fig. 6), thus negatively regulating the BR signal transduction Perez-Perez et al. 2002). The genetic and molecular evidence clari es rst the BnaC04.BIL1 can sharply change plant architecture in natural plant accessions in allotetraploid, and provides new insights into the molecular mechanisms of BR signaling.
Further study has identi ed Bndwarf2, a dwarf and compact mutant in B. napus, and the dwarf trait is controlled by a semi-dominant nuclear gene (Table S3). The plant height of F 1 derived from the cross of Bndwarf2 with the tall parent, decreased by about 50% compared to that of tall plant (Table S2). Particularly, the Bndwarf2 displayed an extreme reduction in height at maturity (33.62 ± 1.12 cm), which is different from the previously reported dwarf mutants in B. napus (Foisset et al. 1995;Wang et al. 2016a;Wang et al. 2016b;Zeng et al. 2011;Zhao et al. 2017). For example, the dwarf mutant BnC.dwf only had 95 cm, and the dwarf locus was controlled by a recessive gene (Zeng et al. 2011). And, the dwarf trait of Bndwf1 mutant (80 cm) was controlled by a dominant gene (Wang et al. 2016a). The F 1 plants have compact properties such as shortened branch, shortened gap between siliques, shortened gap between branches and dwar ng plant height by BnDWARF2 gene (Fig. 1, Fig. S1, Table S2). This nding implicates that the plant architecture of homozygous or heterozygous individuals derived Bndwarf2 mutant is compact (Fig. 1). This kind of compact architecture can be undoubtedly helpful to increase planting density, enhance lodging resistance and increase planting density, therefore the compact plant architecture is ideal for machinery production of oilseed rape.
The germplasm Bndwarf2 that is sparse lacks strong vigorous growth habit. However, the compact plant architecture can be used in hybrid cultivar development in which the compact type and hybrid vigor can be combined well. This is helpful to breeding of variety breeding with the objectives such as high-yield, good quality and suitable for machinery. On the other hand, the growth vigor in pure line or cultivar may be improved in some genetic background. Some reports have demonstrated that the genes in BIN2 regulation network can also interact with BIN2, leading to improvement of the growth inhibition caused by BIN2 gene overexpression caused by natural biological accession state or by transgenics (He et al. 2019;He et al. 2020;Ling et al. 2017;Sun et al. 2018). We speculate that some gene may interact with BnaC04.BIL1 to attenuate its role in limit growth vigor as that the Arabidopsis homolog BIN2 crosstalk experiments have shown. Furthermore, expressions of some regulator genes may probably alter the expression level of BnaC04.BIL1 that is constitutively expressed in the various organs, and reduced expression level may improve the growth vigor. Map-based cloning of BnDWARF2. a The BnDWARF2 locus was mapped primarily on C04 chromosome between the SNP markers M33367 and M35244. b The BnDWARF2 locus was ne-mapped in the 34.62 kb region between SSR markers S3 and S4. The numerals indicate the number of recombinants. c The genes in the mapping interval. d The gene structure and the mutation sites in BnaC04.BIL1. e The protein structure and the mutation sites of the BnaC04.BIL1 protein, and the STKc_GSK3 superfamily domain was predicted. Solid lines show the position of the amino acid transition Figure 3 Expression pattern of BnaC04.BIL1 and subcellular localization of its encoding protein. a Expression pattern of BnaC04.BIL1 detected by qRT-PCR in bud, stem, ower, silique, seed, root, hypocotyl, and leaf from Bndwarf2. The BnActin gene was used as a reference gene and the expression level of bud was set to 1. The bud, stem, and ower samples are from owering stage. The silique samples are from podding stage. The seed samples are from maturity stage. The root and hypocotyl samples are from 7-day-old seedlings grown on medium, and the leaf samples are from seedling stage. b Subcellular localization of BnaC04.BIL1 and BnaBZR1 proteins in tobacco leaf cells. Plasmids pA7-GFP, BnaC04.BIL1-GFP, and BnaBZR1-GFP were introduced into tobacco leaf cells by particle bombardment, respectively. Bars = 20 generation. The PAM is underlined, and nucleotide mutations and deletions are marked with red, with details labeled at right. S1 indicates the sgRNA primer in ORF 5' terminal, and S2 indicates the sgRNA primer in the conserved region