Comparative analysis of bHLH transcription factors in ve Rosaceae species, and expression analysis of PbbHLHs in response to drought stress in pear

The bHLH (basic helix-loop-helix) transcription factor family plays important roles in regulating plant growth and development. However, informations about bHLH in rosaceous fruit species are still limited. In this study, a total of 198 PbbHLH genes were identied in the pear, with 188 bHLH genes in apple, 129 bHLH genes in peach, 112 in strawberry and 122 in Chinese plum. These Rosaceae bHLH genes, plus the 150 Arabidopsis thaliana bHLH genes, were divided into 34 groups, which included one Rosaceae-specic group. Evolutionary pattern analysis showed that whole-genome duplication (WGD) and segmental duplication played critical roles in expansion of the PbbHLH gene family. Ks value indicated that the two WGD duplication events (a recent WGD and an ancient WGD event) lead to the expansion of bHLH gene family. Tissue expression analysis shows that PbbHLHs may have diversity functions in different tissues. Furthermore, eight up-regulated and seven down-regulated PbbHLH genes were identied as the candidate genes in response to drought stress.

duplication played critical roles in expansion of the PbbHLH gene family. Ks value indicated that the two WGD duplication events (a recent WGD and an ancient WGD event) lead to the expansion of bHLH gene family. Tissue expression analysis shows that PbbHLHs may have diversity functions in different tissues. Furthermore, eight up-regulated and seven down-regulated PbbHLH genes were identi ed as the candidate genes in response to drought stress.

Conclusion
A comprehensive analysis of bHLH genes was performed in ve Rosaceae species. The phylogenetic, evolution and expression analyses of the bHLH gene family in pear will be meaningful for investigating the biological roles of PbbHLH genes.

Background
The basic helix-loop-helix (bHLH) proteins are an important superfamily of transcription factors, which are important regulators of plant development and physiological processes [1][2][3]. Each bHLH transcription factor contains a bHLH domain which is involved in binding DNA, including two functional regions, the basic region and the HLH region [4,5]. The basic region is located at the N terminus of the bHLH domain, and it consists of approximately 15 amino acids, which typically include six basic residues [6,7]. The HLH region contains two amphipathic α helices at the C-terminal end, with a loop region of variable sequence and length; the amphipathic α helices of two bHLH proteins can interact to form homodimeric or heterodimeric complexes, to promote protein-protein interactions [4,[8][9][10]. Certain conserved amino acids in the basic region determine recognition. The E-box (5´-CANNTG-3´) is a consensus hexanucleotide sequence. There are different types of E-boxes which depend on the identity of the two central bases, of which the G-box (5´-CACGTG-3´) is the most common type [1,11]. In addition, anking nucleotides outside the core have also been shown to play a role in determining binding speci city [6,12].
Identi cation of the bHLH gene family in various species could facilitate our understanding of their evolution and functions. Compared to animals, only a small number of plant bHLH proteins have been identi ed, while systematic studies on bHLH genes has been performed in only a few plants, such as Arabidopsis (Arabidopsis thaliana) and rice (Oryza sativa) [13,14]. In the Arabidopsis and rice genomes, which are two major model species for owering plants, 162 and 131 bHLH genes have been identi ed, respectively [13,14]. Previous reports have demonstrated that some plant bHLH proteins play important roles in plant growth and tolerance to drought. For example, Arabidopsis AtMYC2 (bHLH) and AtMYB2 (MYB) function as transcriptional activators in abscisic acid signaling [15]. The Arabidopsis myc/bHLH gene ALCATRAZ controls cell separation in fruit dehiscence [16]. OsbHLH148, a basic helix-loop-helix protein, interacts with OsJAZ proteins in a jasmonate signaling pathway, leading to drought tolerance in rice [17].
The Pear (Pyrus × bretschneideri Rehder) is one of the most important fruits in the world [18]. Sequencing of the pear genome was recently completed and published, and the genomes of other rosaceous species, apple (Malus × domestica Borkh.), peach (Prunus persica (L.) Batsch), wild strawberry (Fragaria vesca L.) and Chinese plum (Prunus mume Siebold & Zucc.) have also been sequenced [19][20][21][22], but little bioinformatic analysis on the bHLH gene family in the Rosaceae had been reported till now. The availability of the genome sequences of ve Rosaceae fruit species provides us with an unprecedented opportunity to carry out genome-wide comparative analyses of bHLH gene family in the Rosaceae phylogenetic trees.
In this study, we rst identi ed the bHLH gene family in the pear genome before executing a detailed evolutionary and expression pattern analysis, and providing an exhaustive genome analysis of the bHLH genes in the Rosaceae, as well as analyzing the expression pro les of PbbHLHs in pear different tissues and response to drought stress.

Identi cation and classi cation of bHLH genes in the Rosaceae
To define the bHLH gene families in the Rosaceae, we used the HMMER3 software pro le and BLASTP program to identify the members of bHLH genes in Pear and the three other Rosaceae species. SMART analysis was used to further discriminate and con rm the candidate bHLH proteins. Finally, a total of 198 bHLH genes were identi ed in Pear (Table S1), 129 in peach, 112 in strawberry and 122 in Chinese plum (Table S2), while 188 MdbHLHs of apple had been identi ed in a previous report [23]. The bHLH genes were named according to the method proposed for Arabidopsis [24]. Among these 198 candidate PbbHLH genes, 169 were mapped onto all but chromosomes 9 and 11 of the pear chromosomes, while 29 PbbHLH genes were located on the scaffold (Table S1 and Fig. S1). Six PbbHLHs were located on chromosomes 1, 4 and 12, twelve on chromosomes 10 and 17, 18; fourteen located in chromosomes 6 and 14; ve on chromosomes 7 and 16,13 on chromosome 2, ten on chromosome 3, 21on chromosome 5, nine on chromosome 13, 25 on chromosome 15, but only three genes on chromosome 8.
Multiple sequence alignments, conserved residues in the bHLH domains and predicted DNA-binding ability To examine sequence features of these Rosaceae bHLH domains, we performed multiple sequence alignments of the Rosaceae bHLH amino acid sequences and checked each alignment by hand. In the alignment, we identi ed 11 and 12 residues with at least 50% conservation in pear and apple bHLH sequences, respectively, and 13, 11 and 12 in peach, strawberry and Chinese plum, respectively. Among these conserved residues, six residues (Glu-13, Arg-16, Leu-27, Lys-36, Tyr-49 and Leu-53) ( Fig. 1 and Table S3) were present in more than 75% of sequences, suggesting that these residues are extremely important for the function of bHLH proteins.
It is generally acknowledged that bHLH domains with at least five basic amino acids in the basic region are considered to bind DNA. A total of 167 DNA-binding proteins and 31 DNA non-binding proteins were identi ed in Pear, with 164, 109, 98 and 104 DNA-binding proteins identi ed in apple, peach, strawberry and Chinese plum, respectively, and 24, 20, 14 and 18 DNA non-binding proteins identi ed in apple, peach, strawberry and Chinese plum, respectively (Table 1). According to the presence or absence of residues Glu-13 and Arg-16 in the basic region, the DNA-binding bHLH proteins were further divided into two groups: E-box-binding proteins or non-E-box-binding proteins. Twenty-nine, 22, 12, 15 and 16 non-E-boxbinding proteins were found in Pear, apple, peach, strawberry and Chinese plum, respectively. According to the presence or absence of the His-9 residue, the E-box-binding proteins can be further divided into two subgroups, G-box-binding proteins and non-G-box-binding proteins. A total of 113, 116, 73, 61 and 68 Gbox-binding proteins were found in Pear, apple, peach, strawberry and Chinese plum, respectively (Table  1).

Intron/exon structure in the Rosaceae bHLH domains
To better analyze the intron distribution within the bHLH domain of all of the Rosaceae bHLH genes, we performed a multiple alignment analysis between all the Rosaceae bHLH coding sequences and genome sequences, and eleven different distribution patterns (designated I to XI), with intron numbers ranging from 0 to 3, were found within the bHLH domain. Our results showed that approximately 80% of Rosaceae bHLH genes (159 PbbHLHs (from Pear), 157 MdbHLHs (from apple), 101 PpbHLHs (from peach), 90 FvbHLHs (from wild strawberry) and 96 PmbHLHs (from Chinese plum)) have introns in their bHLH domains (Fig. 2). The most common pattern differed among the ve species. The most common pattern in pear and peach was pattern VI (including one intron), while pattern I (including three introns) was the most common pattern in apple, strawberry and Chinese plum. The most common intron pattern in pear and peach (pattern VI) was the same as in Arabidopsis. Patterns I and VI were found to be the most common ones, present in most bHLH genes.

Phylogenetic analysis of the bHLH proteins
To examine the evolutionary relationships among Rosaceae bHLH proteins, and further predict the function of PbbHLHs, a neighbor-joining phylogenetic tree was conducted, using the alignment sequences of 749 Rosaceae bHLH proteins and 150 Arabidopsis bHLH proteins. The phylogenetic tree showed that these bHLH proteins were categorized into 34 groups (Fig. 3). The smallest group was group 19, having only six members, while the largest group was group 30, containing 65 members. Interestingly, the phylogenetic tree showed that group 20 contained 33 bHLH proteins which were from only Rosaceae species, indicating the Rosaceae-speci city of those bHLH proteins from group 20. To investigate the potential evolutionary mechanisms of the bHLH gene family, we performed the paralogous relationships across the entire pear genome, using a local synteny-based method. We observed that most PbbHLH genes were distributed on 16 chromosomes with an uneven distribution, and 63 paired homologous relationships from segmental duplications appeared in pear ( Fig. 5A and Table   S4). Interestingly, Chromosomes 15 has the most duplicated genes (Fig. 4A), indicating that it plays important roles in duplication events of the PbbHLH gene family. Furthermore, take two homologous gene pairs duplicated in chromosomes 15, we performed an all-vs.-all local BLASTP based on a method similar to the one used for PGDD across the whole pear genome to identify synteny blocks. Highly Conserved synteny was observed in the regions, several of which contained over 100 syntenic gene pairs ( Fig. 4B).
In addition, 158 paired orthologous relationships were found in pear and apple, 173 in pear and peach, 135 in pear and strawberry, and 130 in pear and Chinese plum (Fig S2.). The numbers of the orthologous relationships among pear, apple and peach were greater than those in pear, strawberry and Chinese plum, suggesting that bHLH genes in pear, apple and peach may have originated from a common ancestor.
Ka/Ks ratio and Ks value drive the selective pressure on the evolution of bHLH gene family Ka/Ks ratio is used to explore the selective pressure on duplicated genes, the ratio being an indication of negative selection (where Ka/Ks<1) or positive selection (where Ka/Ks>1) [26,27]. Positive selection is associated with functional divergence. Our result showed that the Ka/Ks ratio of most bHLH gene pairs in Rosaceae was less than one (Fig 6A), suggesting these bHLH genes mainly experienced negative selection. However, several bHLH gene pairs exhibited Ka/Ks>1, including three gene pairs in pear, 12 gene pairs in apple, one gene pair in peach and four gene pairs in strawberry (Fig 6A), suggesting positive selection possibly playing an important role in the functional divergence of these genes.
The synonymous (Ks) value is used to estimate the stage of evolution for the WGD or segmental duplication events. In our study, the paralogous gene pairs were used to calculate the Ks values for dating the evolutionary time of bHLH genes in Rosaceae. As shown in Fig. 6A, the Ks value of most bHLH duplicated genes in pear and apple ranged from 0.1-0.3, this consistent with a recent WGD duplication event in pear and apple. Interesting, the Ks values for duplicated gene pairs which the ratio of Ka/Ks was more than one almost ranged from 0.1 to 0.3 (Fig 6A), suggesting that the recent WGD event in pear and apple is an active stage of evolution for bHLH gene family. In peach, strawberry and Chinese plum, the Ks value of bHLH genes was singly ranged from 0.5-3, suggesting they might have duplicated from a more ancient duplication event. In addition, we characterized the mean Ks value of the PbbHLH gene pairs, showing that these PbbHLH gene pairs were distributed at the two Ks value peaks ( Fig. 6B and Table S4), further suggesting that the recent WGD event and the ancient WGD event led to the expansion of PbbHLH gene family.

Expression pro les of PbbHLH genes in pear different tissues
To further understand the potential functions of the PbbHLH genes, we analyzed the expression pro les of PbbHLH genes in different tissues, including stem, leaf, bud, sepal, petal, ovary and fruit, based on the public RNA-seq data (Fig 7 and Table S5). The result of heatmap showed that the expression pattern of each PbbHLH gene varied greatly in the different tissues. However, we found that 93.9% of PbbHLHs were expressed in at least one tissue (stage) of pear. Among these PbbHLH genes, 48.5% of the genes (96 genes) were expressed in all tissues. Especially, PbbHLH195 was highly expressed in all tissues, suggesting its comprehensive functions in pear development. Furthermore, some PbbHLH genes are highly expressed in a speci c tissue of pear, with low exoression in other tissues. For example, PbbHLH15, 18, 70, 84, 89, 93, 113 and 131 are accumulated mainly in leaf, suggesting these genes may play roles in leaf development or immunity. In addition, PbbHLH138, 167 and 176 are strongly expressed in fruit, with low expression in other tissues, suggesting the potential function in pear fruit development.
On the other hand, several PbbHLH genes have high level of transcripts in reproductive organs. For example, PbbHLH13, 25 and 100 show mainly higher expression in ovary, suggesting these PbbHLH genes may be necessary for reproductive growth. These above results demonstrated that PbbHLH genes may have different functions in tissue-dependent manner.

Expression pro les of PbbHLHs under drought stress
Plants have formed a set of mechanisms to response for stresses during long-term evolution. Previous reports have shown that several bHLH genes play important roles in response to abiotic stress. However, information is limited on PbbHLH genes response to drought stress in pear. To investigate PbbHLHs response to drought stress in pear, the expression pro les of PbbHLH genes which under drought treatment for 0, 1h, 3h, 6h and re-watering 24h were obtained from the transcriptome data sets previously reported ( Fig. 8 and Table S6). Most PbbHLH genes (143) showed differential gene expression. Among them, the expression of 13 PbbHLH genes were signi cantly up-regulated, while 23 PbbHLH genes were signi cantly down-regulated. Furthermore, we respectively selected 8 genes (PbbHLH26, 30, 77, 84, 113, 131, 152 and 166), 7 genes (PbbHLH1, 12, 15, 52, 58, 79, and 92) that were up-regulated and downregulated by at least 10-fold after drought treatment for 6h and restore original expression after rewatering 24h, suggesting these genes may participate in drought stress response.
To verify the reliability of transcriptome data, quantitative real-time PCR (qRT-PCR) was further used to analyze the relative transcript abundance of six selected genes (PbbHLH22, 30, 39, 52, 58 and 131) (

Discussion
The bHLH transcription factors play important roles in plant growth and defense [15,[28][29][30]. In previous reports, systematic and comprehensive whole-genome analyses of bHLH family proteins have been studied in a number of species. For example, 150 members of the bHLH family were identi ed in Arabidopsis [31], with 230 BrabHLH genes identi ed in the Chinese cabbage genome [32], 159 bHLH genes in tomato [33] and 167 bHLH genes in rice [24]. Although the Pear genome sequence project has been completed [18], no detailed analysis of the bHLH gene family has been published, and most functions of the bHLH genes in pear are unclear. In this study, a total of 198 bHLH genes were identi ed in Pear (Table S1). In three other members of the Rosaceae, 129 bHLH genes were identi ed in peach, 112 in strawberry and 122 in Chinese plum (Table S2). Previous reports had shown that 188 bHLH genes were present in apple [23]. Notably, the number of bHLH genes in pear and apple is largely more than that in peach, strawberry and Chinese plum, this may be due to that pear and apple belong to the subfamily Maloideae, peach and Chinese plum belong to the Prunoideae, and strawberry belongs to Rosoideae [10], and the recent WGD duplication events occurred in pear and apple, suggesting that WGD duplication contributed to the expansion of bHLH genes in pear and apple. Additionally, all the bHLHs in Arabidopsis and ve Rosaceae species were unevenly divided into 34 subfamilies (Fig. 3). Moreover, the group 20 contained bHLH proteins from only rosaceous species, indicating that the PbbHLHs in group 20 may play a species-speci c role in the Rosaceae.
The basic region of the bHLH domain has been associated with DNA-binding activity [34,35]. The basic region of the Rosaceae bHLH protein, which functions in DNA binding, contains 17 residues. According to our analysis, the Rosaceae bHLH proteins share similar DNA-binding mechanisms (Fig. 1). According to the presence or absence of residues Glu-13 and Arg-16 in the basic region, the DNA-binding bHLH proteins were further divided into two groups, the E-box-binding proteins and the non-E-box-binding proteins, and, based on the presence or absence of the His-9 residue, the E-box-binding proteins could be further divided into two subgroups, the G-box-binding proteins and the non-G-box-binding proteins. A total of 167, 164, 109, 98 and 104 DNA-binding proteins (including 113, 116, 73, 61 and 68, respectively, G-boxbinding proteins) were found in the pear, apple, peach, strawberry and Chinese plum, respectively ( Table 1). These results were similar to those obtained for the Arabidopsis bHLH proteins [31]. In the further, we need to study the functions of the basic region in the bHLH domain, this will be bene cial to better understand bHLH transcription factors.
To analyze intron distribution within the coding sequence of the bHLH domain in all the Rosaceae bHLH genes, we did alignment analysis with the coding and genome sequences using BLAST software, and found 11 different intron distribution patterns (designated I to XI) (Fig. 2). The intron number ranged from 0 to 3 within the coding sequence of the bHLH domain. In our study, 80.5% of the Rosaceae bHLH genes contained introns in their coding sequence of the bHLH domain (Fig. 2). Pattern I (containing three introns) was the most common pattern in apple, strawberry and the Chinese plum, and was similar to that in the Arabidopsis bHLHs [31], but it was only the second most common one in pear and peach; this nding was consistent with that of tomato bHLHs [33]. These results showed that intron sequences and distribution in different species were different, even though their bHLH domains were conserved.
Gene duplication events are thought to have occurred frequently during plant genome evolution, and gene families have evolved via several different duplication modes, including WGD or segmental duplication, tandem duplication or dispersed duplication [36,37]. Gene duplications can lead to expansion of gene families and can generate new functions [38,39]. In the present work, we found that segmental or WGD duplications were found in a majority of the bHLH genes in pear and apple, while dispersed duplications played important roles in the expansion of bHLH genes in peach, strawberry and the Chinese plum (Fig. 4). A recent WGD duplication events (30-45 MYA) occurred in pear and apple, while peach, strawberry, and Chinese plum did not experience such an event, indicating that bHLH genes were divergent in different plant species, and the recent WGD event may have caused the different numbers of bHLH genes in the members of the Rosaceae investigated (Table 2). This nding was consistent with the result that the number of bHLH genes in pear and apple was larger than in peach, strawberry and the Chinese plum.
Ks, as a proxy for evolution time, has been widely used for estimating the evolutionary dates of segmental duplication events [40,41]. Based on the result of Ka/Ks ratio and Ks value, the duplicated bHLH genes mainly focus on two Ks value peaks in pear and apple, Ks ~ 0.15-0.3 and Ks ~ 1.5-1.8 ( Fig. 5A and 5B), this consistent with a recent WGD duplication event in pear (Ks ~ 0.15-0.3) and apple (Ks ~ 0.2) occurred at 30-45 MYA, and the ancient WGD in pear (Ks ~ 1.5-1.8) and apple (Ks ~ 2) resulted from a paleohexaploidization (γ) event that took place ~ 140 MYA. Interestingly, the Ka/Ks ratio of several bHLH genes was more than one during the recent WGD duplication event (Fig. 5A), suggesting that these bHLH genes may have undergone positive selection and generated new functions during evolution.
Several previous reports have shown that bHLH transcription factors play important roles in plant growth and development. For example, the bHLH transcription factor SPATULA controls Arabidopsis root growth by controlling the size of the root meristem [42]. The rice bHLH transcription factor OsBLR1 regulates leaf angle via brassinosteroid signaling [43]. However, there is limited information available on the role of bHLH transcription factors in terms of controlling growth in pear. The expression patterns are able to provide essential information for studying gene function. Therefore, in the present study, we analyzed the expression level of PbbHLH genes in stem, leaf, bud, sepal, petal, ovary and fruit using transcriptome data ( Fig. 6 and Table S5). 96 PbbHLH genes were expressed in all the tissues of pear. Among these genes, PbbHLH195 showed the highest expression level, suggesting PbbHLH195 may play important roles in growth and development of pear. Furthermore, the PbbHLH genes showed the different expression in different tissues and different stages of development, suggesting the diverse functions of PbbHLH genes in pear. For example, PbbHLH167 and PbbHLH138 both highly expressed in in all fruit developmental stages of pear ( Fig. S3 and table S7), and, with low expression in other tissues, suggesting that PbbHLH167 and PbbHLH138 may involve in pear fruit development. In the further, based on the expression pro les of PbbHLHs in different tissues of pear, we can further study the diverse functions of PbbHLH genes in pear.
To date, several reports have investigated the important roles of bHLHs in response to drought stress. For instance, the bHLH family member ZmPTF1 regulates drought tolerance in maize by promoting root development and abscisic acid synthesis [44]. AtbHLH68 transcription factor contributes to the regulation of ABA homeostasis and drought stress tolerance in Arabidopsis thaliana [45]. A novel bHLH transcription factor PebHLH35 from Populus euphratica confers drought tolerance through regulating stomatal development, photosynthesis and growth [46]. In the present study, we found that 72.2% PbbHLH genes showed the different expression in reponse to drought stress based on the transcriptome data and qRT-PCR results in pear ( Fig. 8; Fig. 9 and Table S6). Among them, 15 PbbHLH genes signi cantly up-regulated or down-regulated by at least 10-fold after drought treatment for 6 h and then restored original expression after re-watering 24 h, suggesting that these candidate genes may play essential roles in pear responses to drought stress. However, the molecular mechanism of PbbHLH genes tolerance to drought still needs further investigation.

Conclusions
In the present study, we conducted a comparative analysis of bHLH gene family in ve Rosaceae species. 749 bHLH genes were identi ed in ve Rosaceae species, among them, 198 belonged to the pear. We divided these bHLH genes into 34 subfamilies. Evolutionary analysis suggested that the recent WGD (30-45 MYA) may drive the expansion of bHLH genes in pear and apple. The transcriptome data and qRT-PCR analysis suggested that bHLH genes might play important roles in pear development and drought stress response. Our work provides useful information for further potential functional studies in revealing their regulation mechanism.

Identi cation of bHLH genes in the Rosaceae
The Pear (Pyrus × bretschneideri) genome sequence was downloaded from the pear genome project (http://peargenome.njau.edu.cn/). The genome sequences of peach and strawberry were downloaded from Phytozome (http://phytozome.jgi.doe.gov/pz/portal.html#), and the Chinese plum genome sequence was downloaded from the Prunusmume Genome Project (http://prunusmumegenome.bjfu.edu.cn/index.jsp). The sequences of Arabidopsis (A. thaliana) bHLH genes were retrieved from The Arabidopsis Information Resource (TAIR; http://www.arabidopsis.org/) [47]. To identify the members of the bHLH gene family in the Rosaceae, two approaches were used in this research. Firstly, the HLH domain (http://pfam.xfam.org, PF00010) was downloaded from the Pfam website (http://pfam.xfam.org/family) [48], and it was used as a query to perform BLAST searches in HMMER3 software package [49] against the genome databases of four rosaceous species (pear, peach, strawberry and Chinese plum) [18][19][20][21][22]. Additionally, using the sequence of Arabidopsis bHLH as a query [31], the pear, peach, strawberry and Chinese plum bHLH sequences were blasted. Moreover, to further con rm all candidate bHLH genes, the predicted bHLH domain was detected by the Simple Modular Architecture Research Tool (SMART) program (http://smart.embl-heidelberg).

Multiple sequence alignments
Multiple sequence alignments were aligned using ClustalX [50] and SMS programs (http://www.bioinformatics.org/sms2/) [51], and domains were identi ed using the InterProScan tool [52,53]. To obtain information on the intron/exon structure, the cDNA alignment of bHLH domain sequences was obtained, according to the amino acid sequence alignment, and the information on the intron distribution pattern and intron splicing phase were derived from the aligned cDNA sequences.

Phylogenetic analysis
Phylogenetic trees were constructed using the neighbor-joining (NJ) method with MEGA 7 [54]. The reliability of the trees obtained was tested using the bootstrap method with 1,000 replicates, and branches with fewer than 50% bootstrap values were rejected. NJ analysis was performed with the Poisson model.

Chromosomal location and duplication pattern analyses of bHLH genes in pear and four other Rosaceae species
The chromosomal location information on the bHLH genes was obtained from genome annotation documents. The data were then plotted using Circos software [55]. Analysis of synteny among the ve Rosaceae genomes was conducted according to duplication gene pairs, using a method similar to that developed for the Plant Genome Duplication Database (PGDD) (http://chibba.agtec.uga.edu/duplication/) [56]. The MCScanX program was used to identify the type of gene duplication in each bHLH gene family [57].

Calculating Ka and Ks of the bHLH gene family in pear
To explore whether positive selection drove the evolution of the bHLH gene family, the rates of nonsynonymous (Ka) and synonymous (Ks) substitutions in all paralog pairs were analyzed using the coding sequence (CDS) of bHLH gene paralogs in pear. The ratio of these two values (Ka/Ks) is an indication of purifying (Ka/Ks<1) or diversifying selection (Ka/Ks>1) [58]. MCScanX downstream analysis tools were used to annotate the Ka and Ks substitution rates of syntenic gene pairs. To date segmental duplication events, 63 consecutive bHLH gene pairs in pear were chosen to calculate the mean Ks. Calculating the dates of duplication events was carried out according to the previous reports [18,59].

Genome-wide expression analysis of bHLH gene family in pear
To investigate the expression pro le of bHLH gene family members in pear different tissues, we utilized the transcriptome sequencing data presented in several earlier studies, while the expression data obtained from pollen was used to measure the expression level of bHLH genes. The RNA-seq data was downloaded from our center website (http://peargenome.njau.edu.cn/). The RPKM (reads per kilobase per million mapped reads) values were used to measure the gene expression levels. The log2 transformed RPKM values were utilized to measure the expression level of the PbbHLH genes and to generate the heatmaps.
qRT-PCR Total RNA was extracted using a Plant Total RNA Isolation Kit (FOREGENE, Chengdu, China) and genomic DNA contamination was removed by DNase I. Puri ed total RNA (3 μg) was used for reverse transcription with RevertAid First Strand cDNA Synthesis Kit (Thermo Scienti c, USA). Subsequently, quantitative PCR (qPCR) was performed on 20-μL samples using SYBR Green (TaKaRa, Dalian, China) on an IQ5 real-time PCR machine (Bio-Rad, Hercules, CA, USA) according to the manufacturer's speci cations. The thermal cycle was performed at 95°C for 5 min followed by 45 cycles of 95°C for 15 s each and 60°C for 15 s. The pear TUB gene was used as an internal control and relative expression levels were determined using the 2 −△△CT method as described previously [60]. Primers were designed using Primer 5 software [61] based on the target genes and are shown in Table S8.      Expression pattern of PbbHLH genes in pear different tissues. The heat map was generated by transcriptome data of pear seven different tissues, including stem, leaf, bud, sepal, petal, ovary and fruit.

Figure 9
Relative expression levels of six PbbHLH genes in response to drought treatment. Quantitative RT-PCR was used to analyze the expression level of six PbbHLH genes in response to drought stress. 0, 1, 3, 6 and D24, represents for 0, 1, 3 and 6 h under drought treatment and recovered for 24 h in water, respectively.

Supplementary Files
This is a list of supplementary les associated with this preprint. Click to download.