Genome-wide Survey of bHLH Super Gene Family in Brassica napus and Their Roles in Roots

Background: Basic helix-loop-helix (bHLH) gene family is one of the largest transcription factors in plants and are functionally characterized in diverse species. However, less is known about their functions in the economically important allopolyploid oil crop, Brassica napus . Results : We identified 602 potential bHLHs in B. napus genome ( BnbHLHs ) and categorized them into 36 subfamilies, including seven newly separated subfamilies, based on phylogeny, protein structure and exon-intron organization analysis. The intron insertion patterns of this gene family were corrected and a total of eight types were identified in the bHLH regions of BnbHLHs . Chromosome distribution and synteny analyses revealed that hybridization between Brassica rapa and Brassica oleracea was the main expansion mechanism for BnbHLHs . Expression analyses showed that BnbHLHs were wide and formed six main patterns, suggesting they may participate in various aspects in B. napus during the development. The expression profiles under five hormone treatments (IAA, ABA, ACC, GA3, 6-BA) in roots further revealed the active response of BnbHLHs with a large proportion of which being induced. qRT-PCR analysis confirmed the expression profiles of five candidate BnbHLHs under five hormone inductions. Up to 246 BnbHLHs from nine subfamilies were predicted to have potential roles relating to root development by joint analysis of expression profile and homolog function. Further, the MYB/bHLH/WD40 (MBW) protein complex regulating root hair development were verified in B. napus by yeast two-hybrid experiment. Conclusion: The 602 BnbHLHs identified from B. napus could be classed into 36 subfamilies, and those members from the same subfamily generally have similar sequence motifs. BnbHLHs may widely involve in root biological process in B. napus . Overall, this study provides important insights into the characterization and potential functions of B. napus bHLH super gene family and thus will be useful in future gene function research.


Background
Transcription factors (TF) are an important kind of genes that were widely distributed in eukaryote kingdom and usually contain two different functional domains involved in DNA binding and transcript activities [1,2]. The bHLH transcription factor was characterized by an approximately 50-60aa conserved DNA binding domain consisted of two main regions, namely the basic and the helix-loop-helix (HLH) regions. The basic region is a 10-15aa region at the N-terminal, which functions as DNA recognition and allows the binding of HLH region [3]. The HLH region, composing of two relatively conserved amphipathic helices linked by a divergent loop, is approximately 40aa with more hydrophobic amino acids, which contributes to its function in the dimerization between HLH regions [4,5].
The bHLH gene family existed in land plants over 400 million years ago and were highly conserved during plant evolution [6]. As one of the largest transcription factor gene families in plants, the number of bHLH genes (bHLHs) were largely increased during the evolution. For example, there was only one bHLH gene in Cyanidioschyzon Merolae [6], 98 bHLHs in moss [7], 208 in Zea mays [8], 167 in Arabidopsis [7], 159 in tomato [9] while 230 in Chinese cabbage [10]. The substantial increase of bHLHs resulted in their important roles in diverse physiological and developmental processes, with majority are involved in metabolic and developmental processes. For instance, AtbHLH045/MUTE controls sequential cell fate; SlbHLH22 in tomato promotes early flowering and accelerates fruit ripening; SPT is required either for carpel development or is involved in endocarp margin development in Arabidopsis or Prunus persica [10,11]; MYC2 regulates sesquiterpene and artemisinin biosynthesis in various species like Arabidopsis, Aquilaria sinensis and Artemisia annua [12,13]. Meanwhile, some bHLHs are related to abiotic stress response, including cold, drought, and salt stresses etc, e.g. PIF4 in Arabidopsis mediates plant architecture to response to high temperatures [14]; StbHLH1 in potato also responses to high temperature by regulating anthocyanin biosynthesis [15]; while MdbHLH3 in Malus domestica responds to low temperature [16]. Moreover, bHLHs also respond to various hormone inductions. For instance, Arabidopsis MYC2 is well known for its conserved roles in abscisic acid (ABA), jasmonic acid (JA) and light signaling pathways [17][18][19]; AP2/ERF is a jasmonate-responsive transcription factor involved in secondary metabolism [20].
In this study, we identified 602 bHLHs in the important economy crop, Brassica napus Yeast two-hybrid experiment further verified the interaction relationships of BnbHLH544 (BnGL3) protein in the MBW complex, providing strong support to their roles in hair root development.

Sequence retrieval
A preliminary search for B. napus bHLH proteins in Genoscope (http://www.genoscope.cns.fr/ brassicanapus/) is performed using BLASTP with at least one representative sequence of the bHLH domain for each bHLH subfamily. The redundant sequences are discarded to ensure the candidate genes mapped to unique loci in their respective genome. We then confirm the putative non-redundant sequences to ensure that the candidates contain the bHLH domain using ExPASy (http://expasy.org/prosite/) [
Mapchart software is used to draw the chromosome map of candidate BnbHLHs. Locations of bHLHs in Arabidopsis, B. rapa, and Brassica oleracea are also determined using the same method. We then use CoGe software (https://genomevolution.org/coge/) to conduct a gene synteny analysis of bHLHs in Arabidopsis, B. napus, B. rapa, and B. oleracea.

Intron/exon structure analysis
To find the intron distribution and splicing phase in the candidate BnbHLHs, we compare and view the coding DNA (CDS) and DNA sequences of candidates using GSDS software (http://gsds.cbi.pku.edu.cn/). Then manually locates the intron insertion sites in the corresponding protein sequences.

Identification of conserved motifs
Full-length protein sequences of candidate BnbHLHs are analysed using MEME software (http://meme-suite.org/tools/meme) [29], with the following parameters: optimum motif width ≥ 6 and ≤ 250 and maximum number of motifs to identify = 30.

Expression analysis of BnbHLHs
The B. napus expression datasets are downloaded from the BioProject (NCBI database: PRJNA358784). The data were obtained from various tissues at different B. napus developmental stages and under five hormones induction (IAA, GA3, 6-BA, ABA, and ACC) were used to analyse the expression profiles of candidate BnbHLHs respectively. The expression profiles are analysed using Cluster 3.0 software [30] and the heatmaps are drawn using the R package. All genes with FPKM <1 are excluded from the heatmap, as they may be pseudogene or may be expressed only under specific stresses or inductions.
For qRT-PCR analysis, seeds of B. napus variety Zhongshuang 11 (ZS11) are obtained from the College of Agriculture and Biotechnology, Southwest University, and germinated on petri dishes. At the five-leaf stage, seedlings are treated in Hoagland liquid medium containing five phytohormone (50 µM ABA, 120 µM GA, 75 µM 6-BA, 60 µM ACC, and 10 µM IAA) respectively. The seedlings are grown in an artificial climatic chamber at 25°C with a 16 /8 h photoperiod (day/night). The root tissues are then harvested at 0, 1, 3, 6, 12 and 24 h after the treatments and immediately frozen in liquid nitrogen and stored at -80 °C for RNA isolation.
Total RNA is extracted using EASYspin total RNA Extraction kit (Biomed, Beijing). The total RNA sample is treated with DNase I (Promega, USA) before use. First-strand cDNA synthesis is performed using the M-MuLV RT kit (Takara Biotechnology, Japan). The fluorescence is measured after the extension step by using the CFX Connect™ Real-Time System (Bio-Rad, Chongqing, China) and the SYBR-Green PrimeScript RT-PCR Kit (Takara Biotechnology, Japan). The B. napus Actin7 ( BnActin7) (GenBank accession no. AF024716) is used as the reference gene. The primers used in this analysis are listed in Additional file 9: Table S9. Three biological replicates are included for each treatment, and each consists of three technical replicates. The reaction conditions for real-time PCR are as follows: initial denaturation at 95℃ for 3 min, followed by 40 cycles of denaturation at 95℃ for 10 s and annealing at 58℃ for 20 s. The relative expression levels are determined using the 2

Yeast two-hybrid assays
The full-length ORF of BnGL3 (BnaC09g12820D, BnbHLH544), BnWER (BnaA02g02300D), BnTTG (XP_013643414) and BnCPC (BnaA05g01400D) genes are amplified from ZS11 seedling root cDNA. For yeast two-hybrid assay, the coding cDNA sequences of BnGL3, BnWER, BnTTG and BnCPC are recombined into pGADT7 vector, while their BD domains (BnGL3 1231-1818 , BnWER 1-360 , BnTTG  and BnCPC  ) are recombined into pGBKT7 vector respectivley. The primers used in this analysis are listed in Additional file 9: Table   S9. Empty vectors of pGADT7 and pGBKT7 are co-transformed as a negative control. The corresponding constructs are co-transformed into yeast strain AH109 and then test for protein-protein interaction relationship respectively, following the manufacturer's protocols (Takara Biosciences, Clontech, Japan).

A large number of bHLHs were identified in Brassica napus
To identify bHLH encoding genes in B. napus genome (Darmor-bzh), a preliminary repeated BLASTP search (e values of < 1.0) is performed using the representative sequences of Arabidopsis bHLH proteins as queries (Additional file 1: Table S1). To ensure the integrity of the bHLH protein data in B. napus, we refer to the method of Guo et al.
Originally, a large number of deduced amino acid sequences containing bHLH domains are obtained. Then, redundant sequences are discarded, and the bHLH domains are verified in the remaining sequences by ExPASy. Consequently, 11 genes are excluded from our dataset as no bHLH domain is identified by ExPASy analysis. Meanwhile, the sequence information of 65 BnbHLHs from Darmor-bzh are corrected by the data from ZS11 genome (Additional file 1: Table S1). Finally, a total of 602 BnbHLHs with relatively complete open reading frame (ORF) are obtained in this study, account for approximately 0.60% of the B. napus protein-coding genes. The corresponding proportion in wheat, rice, maize and Arabidopsis are 0.55, 0.47, 0.59, and 0.61% respectively [13]. The candidate BnbHLHs are then named according to their chromosomal distribution orders (Additional file 1: Table   S1). Physicochemical property analysis showed that the BnbHLH proteins (BnbHLHs) varied in length from 63 to 1440 amino acids (aa); their molecular weight ranged from 6.9 (BnbHLH105) to 165 kDa (BnbHLH381); and the isoelectric points are from 4.36 (BnbHLH562) to 11.79 (BnbHLH023). Subcellular localization analysis demonstrates that all BnbHLHs are located in nucleus (Additional file 1: Table S1).
For further comparative analysis across different species, we identified 255 bHLHs in B.

Sequences characterics of the bHLH domains of BnbHLH proteins
To investigate the sequence features, we perform multiple alignment analysis of the 602 bHLH domains of candidate BnbHLHs. The result is visualized using Weblogo online software.
Our result shows that the length of the bHLH domains of BnHLHs is approximately 55 amino acids, ranging from 39-57 aa. The bHLH domain is generally conserved in this gene family in B. napus, in which two typical conserved regions are included, namely the basic and HLH regions. Ten residues are identified with conservation of more than 70% consensus ratio in the bHLH domains (Fig. 1a), inclduing four are located in the basic region, five in the two helix regions and one in the loop region. Consistent with other studies [7][8][9][10]33], Leu-25 is the most conserved residue, with a conservation of almost 100% (Additional file 3: Table S3), indicating its essential role for bHLH proteins.
Interestingly, Phe-30 is partly substituted by Ser in Arabidopsis, rice, tomato etc, however, no such situation is observed in B. napus, B. oleracea and B. rapa, (Additional file 3: Table   S3), suggesting a higher conservation and/or close relationship in these three species.
To further characterize the BnbHLHs sequence features, the criterion given by Massari and Murre [34] is used (Fig. 1). Our result shows that the 602 BnbHLHs are separated into two major categories according to the sequence profiles of the bHLH domains: 132 (21.93%) atypical BnbHLHs (non-DNA-binding proteins ), and 470 (78.07%) typical BnbHLHs (Fig. 1).
And the latter is further consisted of three categories, including 300 (49.83%) G-box binding proteins, 103 (17.11%) E-box binding and 67 (11.13%) non-E-box binding proteins ( Fig. 1). The sequences of the 132 atypical BnbHLHs are divergent in the basic region but are relatively conserved in the HLH region, especailly in the loop region (Fig. 1d). Similar situation is found for the non-E-box binding proteins (Fig. 1c), suggesting its close relationship to the atypical BnbHLHs. In contrary, the residues in the basic region of the Ebox/G-box DNA binding types are more conserved than the HLH region (Fig. 1b, c).

Protein structures of BnbHLHs were conserved in each subfamily
To determine the evolutionary relationship of BnbHLHs among Brassicease species, we construct a NJ phylogenetic tree on the basis of the alignment of 1243 bHLH domains from B. napus (602), B. rapa (230), B. oleracea (244) and Arabidopsis (170).
The 1243 bHLH proteins are divided into 35 subfamilies, comprising the largest number to date (Fig. 2a). Among which, 28 subfamilies are consistent with the previous research thus their names keep the same [7]; seven subfamilies (S33-S39) containing bHLHs from these four species with a higher bootstrap value are newly identified in this study; while two previous reported subfamilies (S6 and S8) are not existed in this study as they were only found in lower plants (moss and algae) [7]. Compared with the division of AtbHLHs, S5 subfamily in B. napus is divided into S5 and S33; S17 is divided into S17 and S34; S21 is divided into S21 and S35; S24 is divided into S24, S36 and S37; S30 is divided into S30 and S38; whereas the orthologs of S39 in Arabidopsis are previously defined as orphan genes [7] which are defined as a new subfamily in this study. The distributions of BnbHLHs in the 36 subfamilies are biased, varying from two (S22 and S38) to 62 genes (S25). In addition, the BnbHLHs of different DNA binding types have a biased distribution tendency among different subfamilies as well, but the BnbHLHs in a given subfamily usually share the same DNA binding type (Fig. 2b). A total of 11 subfamilies (S2, S3, S5, S7, S10, S11, S13, S14, S24, S25 and S26) contain G-box binding proteins; five subfamilies (S1, S9, S17, S27 and S37) contained E-box-binding proteins; three subfamilies (S20, S23 and S39) contained non-E-box-binding proteins; while seven subfamilies (S16, S21, S22, S33, S34, S35 and S38) contained non-DNA-binding proteins (Fig. 2b).
To further discovery the non-bHLH domains and explore their distribution patterns within each subfamily, the MEME tool is applied. A total of 27 conserved motifs with variable length (8-103 amino acids) are obtained ( Fig. 2c Additional file 4: Table S4). Among which, motif 1 and 2 are distributed in all BnbHLHs, and made up the basic and the two helix regions of the bHLH domains, respectively. The loop region is located between motif 1 and 2, indicating that this region is variable than the basic and helix regions. Outside the bHLH domain, members of the same subfamily generally share several same motifs. For example, all BnbHLHs in subfamily S9 contain motif 22; proteins in subfamily S26 all contain motif 5 (Fig2b, Additional file 4: Table S4). Moreover, some motifs have been characterized and were defined as additional functional properties, e.g. motifs 4, 8 and 20 were detected in many proteins in subfamilies S2 and S5 in various species, such as TabHLH239, AtMYC2, TabHLH184 and ZmbHLH103, which were found to be significantly matched with an ACT domain that contributed to the recruitment of the C1 R2R3-MYB factor to the C1 binding sites located in the promoters of flavonoid biosynthetic genes [35]. Meanwhile, motif 6 in these two subfamilies are also conserved, which overlapped with the MIR and MYC_N domains that can interact with JAZs (jasmonate ZIM-domain) [36].
Besides, some motifs are demonstrated to be subfamily-specific, yet their functions are still unclear (Fig. 2c).

Intron insertion patterns of BnbHLHs were conserved within each subfamily
The intron and exon structure is an important clue to understand the gene evolutionary relationship and functional diversification within a gene family [37]. The intron and exon patterns of candidate BnbHLHs are determined by comparing their full length CDS and DNA sequences using GSDS web server.
A total of eight intron insertion patterns (pattern a to k) are observed in the bHLH domains in B. napus, containing 0 to 2 intron insertion sites (Fig. 3). The nomenclature of the intron insertion patterns of BnbHLHs is referred to that of Carretero-Paulet et al. 2010 [7]. In this study, the previous defined patterns. The intron insertion sites in the basic and helix regions are located at three highly conserved residues, Arg-11 (the E-box recognition site), Phe-21 and Lys-33 (Fig. 3).
Further, the intron insertion sites of most patterns are conserved, excepting pattern j (Fig.   3). Among them, pattern a, c, e and f are similar, thus are likely to be homologous, where pattern f lacks the first intron as compared with pattern a, pattern e lacks the second intron while pattern c has the second intron inserted at L-50 as compares with pattern e. Similar situation is observed in pattern h and i. Meanwhile, phylogenetic analyses show a close relationship within intron insertion pattern a, c, e and f, further confirms their close relationship as well. Moreover, pattern k and i are indicated as the ancestral types because they are existed in members from algae [7]. Patterns f, a and k are the three types accounting for the majority of BnbHLHs (41.2%, 26.2% and 20.1% respectively). This trend is similar to the results in other species, such as Arabidopsis, rice, potato, poplar and tomato [8][9][10][11]13]. Accordingly, these three patterns are obtained by many subfamilies while the remainings (pattern c, e, h, i and j) are existed in only one or two subfamilies, indicating a different expansion trend.
The distributions of intron insertion patterns are generally conserved within most subfamilies. For example, members in subfamilies S12, S10, S11, S7, S9, S5, S33, S2, S1, S13, S23 and S38 contain pattern f, excepting several genes that may attribute to incomplete genomic annotation information (Fig. 2b). The conservation of intron insertion pattern of BnbHLHs within each subfamily provides an independent criterion for the reliability of our phylogenetic analyses (Fig. 2b). Meanwhile, the intron insertion patterns of BnbHLHs is almost the same with their orthologs in Arabidopsis. The only exception is S27 subfamily which contains pattern a for B. napus members while their homologs in Arabidopsis is pattern f [7]. To confirm this result, we further compare the corresponding results in other species, e.g. rice [33]. And we confirm that it should be pattern a for the homologs in this subfamily, including Arabidopsis homologs (At080, At081, At122, At128, At129 and At130).
Overall, intron insertion patterns of BnbHLHs are conserved within most subfamilies and coordinated with these of AtbHLH orthologs as well. And the intron insertion sites in the basic and loop regions are more conserved than those in the helix regions.

BnbHLHs
In this study, up to 602 BnbHLHs are identified, while the gene number in lower plants are much lower, like Volvox carteri has only three [6]. This indicates a large scale expansion for this gene family has occurred during the evolution. To explore the expansion mechanism of this gene family in B. napus, the chromosomal locations and syntenic relationships of BnbHLHs are analysed based on the genome information from Genoscope and CoGe databases.
Chromosomal location analysis showed that there are 294 and 306 BnbHLHs in A n -and C nsubgenomes, respectively, indicating no biased tendency between these two sub-genomes (Fig. 4a). The BnbHLHs are distributed on all of the 19 chromosomes. The A n -subgenome has an average of 27.4 BnbHLHs on its 10 chromosomes with A10 has a minimum number of 14 and A09 has a maximum of 41 genes. The average number of BnbHLHs in C nsubgenome is 29.9, with C06 contains a minimum of 19 genes and the C03 has as many as 48 genes. Thus, the genes on each chromosome is uneven within both subgenomes.
Among the 602 BnbHLHs, 475 genes have syntenic relationships, and 382 of which are inherited from B. rapa or B. oleracea genomes (Additional file 5: Table S5). In contrast, only 79 genes (7.8%) of 79 syntenic pairs are originated from segmental duplication in B. napus genome, and 34 genes (5.8%) from 28 syntenic pairs are from tandem duplication (Fig. 4b). These results provide an outstanding example of genome-wide allopolyploidization between B. rapa and B. oleracea that mainly contributes to the large BnbHLHs expansion in B. napus (63.3%). Moreover, we find that the genome-wide duplicated genes take the largest number of BnbHLHs in majority subfamilies, while the segmental duplicated genes take the biggest amount in subfamilies S10 and S25 (nine genes each subfamily), and the tandem duplicated genes in S12 is the most (11 genes).
Furthremore, the BnbHLHs with intron pattern f expands most in B. napus (31 duplicats), contributing to the largest proportion of BnbHLHs (Additional file 1: Table S1).
Taken together, the main expansion mechanism of BnbHLHs is whole-genome duplication (allopolyploidization), while segmental and tandem duplication events preferentially occurred in certain subfamilies with relative specific intron patterns. It is common that members of a given subfamily generally exhibit the same/similar expression profile. For example, members in S2 subfamily are mainly expressed in root and leaf at seedling, budding, and flowering stages (Additional file 6: Table S6). Moreover, the bHLHs in the same subfamily probably process the same or similar expression profiles across different species as well, thus may share conserved functions during evolution. For example, AtbHLH155/CPU and AtbHLH156/LHW in subfamily S23 play an essential role in establishing vascular cells and the size of vascular initial population in root meristem [38].
The corresponding homologs in B. napus are also expressed in roots (Additional file 6: Table S6

Plenty of BnbHLHs are induced by hormones treatments in root
As discussed above, plenty of BnbHLHs are highly expressed in roots, implying possible roles in root related biological processes. To further explore their functional characteristics in roots, a comprehensive expression analysis of candidate BnbHLHs in roots under five hormone treatments (auxin indole-3-acetic acid, IAA; abscisic acid, ABA; cytokinin 6-benzyladenine, 6-BA; ethylene precursor 1-aminocyclopropane-1-carboxylic acid, ACC; and gibberellic acid, GA) is performed, based on the RNA-seq data (BioProject ID PRJNA358784).
To further verify the results by RNA-Seq analysis, five BnbHLHs that have high expression levels in roots (Additional file 6: Table S6) and obviously respond to hormone inductions (Additional file 7: Table S7)  subfamily show similar expression pattern under the five hormone treatments (Fig. 6). In addition, cis-acting elements analysis revealed that the promoter regions of these five BnbHLHs contain more than one cis-acting elements that are related to hormone response (Additional file 8: Table S8). This further supports our above results. and BnbHLH126 which provides a valuable foundation for future functional research.

Phylogenetic tree and subfamily division
As an important plant transcription factor super gene family, genome-wide analysis of bHLHs has already been widely performed in a mass of species [7][8][9][10]. However, we find Generally, the division of gene family is performed on the basis of the topology and bootstrap value of the phylogenetic tree [40,41]. However, we find that subfamilies VII (a+b), IX and IIIf in the result of Pires and Dolan did not consist of a consensus node, but across different branches/clades instead [6]. By contrast, the situation was well defined in the results of Carretero-Pault, indicating it is relatively more credible. Furthermore, most of the orphan genes (62 genes) identified by Pires and Dolan were classified into different subfamilies by Carretero-Paulet et al. Consequently, an obvious decrease of orphan gene proportions (total 15 genes, 2%) was observed, suggesting a relative better solution for the classfication of orpan genes. To date, this gene family have been genome-wide characterized in various plants, such as Z. mays [8], tomato [9], and B. rapa [10], etc., and the classification in most of those studies referred to that of Pires and Dolan. As a result, there did exist several inadequacies for the subfamily classification in those researches. For instance, in Z. mays , subfamilies VII, VIII, IX, XVI and XVIII did not clustered in a consensus node respectively [8]; the same situation were observed in B.
In this study, the 602 candidate BnbHLHs are classified into 43 subfamilies, as refered to the method of Carretero-Pault et al [7] (Fig. 2). Of which, onesubfamily (S39) is newly identified (AtbHLH151 homolgs), while 11 ones are separated from four former subfamilies that had a relatively low bootstrap value support in previous study [7], including three subfamilies are separated into two new subfamilies (S5 and S33; S17 and S34; S21 and S35; S30 and S38, respectively) and one subfamily is divided into three new subfamilies (S24, S36 and S37) in our study. The difference may be attributed to more sequences from Brassicaceae species applied that have close evolutionary relationship. It is reported that many other characteristics of candidates generally supply important clues to support the subfamily classification, including highly conserved intron pattern and motif distribution within each subfamily [40,41]. Similarly, the sequence characteristics within each subfamily are also highly conserved in our results, which provides an independent support to our phylogeny analysis and classification as well. For example, subfamily S39 had a relatively high bootstrap value (Fig. 2a), share conserved non-E-box and k intron pattern (Fig. 2b, c); S24 subfamily defined by Carretero-Pault et al. [7] is separated into subfamilies S24, S36 and S37 in this study (Fig. 2a), and these three new subfamilies all contained subfamily-specific DNA binding type, intron insertion patterns and motifs (Fig.   2b, c); the former S17 subfamily is separated into S17 and S34 subfamilies in present study (Fig. 2a), and the BnbHLHs in the new S17 subfamily had a E-box in DNA binding domain while these in S33 areee all non-DNA binding types; Similarly, the former S5 subfamily in Carretero-Paulet's study [7] is divided into two new subfamilies in our study as well.
Taken together, here we provide a more systematic classification of bHLH proteins in plants, which lays a good foundation for exploring the evolutionary characteristics of bHLH gene family.  (Table 1).
As mentioned above, the bHLHs are widely expressed in root tissues in B. napus.
Together, these results suggested possible wide roles of bHLHs hormone signaling pathways in plant roots.
To date, the mechanisms of bHLHs function in plant development processes have been extensively studied in numerous plants at biochemical, genetic, and molecular levels [7].
And it is revealed that bHLH proteins usually functioned in homo-or heterodimeric forms proteins bind to the bHLH protein [35,47,48]. Moreover, this complex is proved to be conserved in plants during evolution [73,76,77], which regulates five processes, namely the production of anthocyanidin, proanthocyanidin, seed-coat mucilage, and trichomes and root hairs development [70]. For example, the roles of this complex in controlling the development of trichomes and root hairs are verified in many species, ranging from monocots (Z. mays ) to dicots (Arabidopsis, Arabisalpina, Gossypium hirsutum and Petunia hybrida) [23,25,78]. And four residues of GL3 and EGL3 proteins are identified to be vital for the protein interaction, including Phe-177 of AtGL3, Pro-377 of AtGL3, Asp-477 of AtGL3 and Ser-589 of AtGL3 [70]. Similarly, these key residues are observed in BnbHLH orthologs as well (Additional file 1: Table S1), suggesting they may have the same/or similar roles in B. napus. To further verify the interaction relationship among these three types of proteins as well as the existance of the MBW complex in B. napus, the yeast-twohybrid (Y2H) experiment is applied in present study. Our result shows that BnbHLH544 (ortholog of EGL3 ) can interact with BnCPC (BnaA05g01400D, ortholog of CPC)/BnWER (BnaA02g02300D, ortholog of WER), and BnTTG (XP_013643414, ortholog of TTG) (Fig. 7), indicating this complex is existed in B. napus as well, which probably functions through the same mechanism.

Availability of data and material
The datasets supporting the conclusions of this article are included within the article and its Additional files.

Ethics approval and consent to participate
Not applicable.