The whole genome data of C. paliurus were available from the Genome Sequence Archive (GSA) database (https://ngdc.cncb.ac.cn/gsa) provided by our research group. The Hidden Markov Model (HMM) profile of the HLH domain (PF00010) was obtained from the Pfam database (version 30.0) (Finn et al., 2014), and was used as a query to search for all protein sequences with default E-values in the whole genome and to identify genes with specific conserved domains by HMMER software (version 3.3; http://hmmer.org/) (Johnson et al., 2010). All screened sequences were aligned and checked with the online tools Batch CD-search (https://www.ncbi.nlm.nih.gov/Structure/bwrpsb/bwrpsb.cgi) (Marchler-Bauer and Bryant, 2004), Pfam, and SMART (http://smart.embl-heidelberg.de) (Letunic et al., 2021) to verify the existence of the conserved bHLH domain. The ExPASy software (https://web.expasy.org/protparam/) was used to obtain basic physical and chemical characteristics of these bHLH genes respectively.

The A. thaliana MYB sequences data were download from PlantTFDB database (http://planttfdb.cbi.pku.edu.cn/index). The construction of a phylogenetic tree consisted of proteins from A. thaliana and C. paliurus was performed with MEGA X (version 6.0) (Kumar et al., 2018) software using the neighbor-joining (NJ) method with 1000 bootstrap replicates. Multiple sequence alignment (MSA) of C. paliurus and A. thaliana bHLH proteins was performed using ClustalX 2.11 software (Thompson et al., 1997), and Weblogo3 (http://weblogo.threeplusone.com/create.cgi), while Jalview software (http://www.jalview.org/) was used to visualize and analyze the sequences of conserved domains in CpbHLH proteins. The GFF3 (Generic Feature Format Version 3) file, containing the positional and gene structure information of genes on the chromosomes, was obtained from whole genome data of C. paliurus. The TBtools software (version 1.098774) (Chen et al., 2020) was adopted to map the CpbHLH genes onto specific chromosomes.

The exon/intron structures of CpbHLH genes were visualized by TBtools software (version 1.098774) (Chen et al., 2018), whereas fifteen conserved motifs were obtained using the online software MEME (http://MEME-suite.org/) (upper limit of the recognition motif was 20, minimum motif width was 6, and maximum motif width was 50, zoops) (Bailey et al., 2009). The online tool PLACE (Higo et al., 1999) was used to analyze the cis-acting elements of CpbHLH genes.

Raw data were obtained via RNA sequencing of leaves treated with different salt concentration in hydroponic experiment (Zhang et al., 2021). CpbHLHs with reads per kilobase of transcript per million mapped reads or fragments per kilobase of transcript per million mapped reads (RPKM and FPKM, respectively) > 1 were collected for further analyses of all of the transcriptome data. TBtools was performed to generate the heatmap (Chen et al., 2020). Gene ontology (GO) analysis was carried out by the Blast2GO program (Conesa et al., 2005), with selecting the NCBI database as the reference database. The results were divided into three categories, namely molecular function, biological process, and cellular component. The NCBI database (https://www.ncbi.nlm.nih.gov/) were used to search the functions of AtbHLHs, which were predicted to be orthologous genes of CpbHLHs. STRING (https://string-db.org/) (Szklarczyk et al., 2019) was performed to predict the functional interaction network of candidate genes with option value>0.7.

The experiment was carried out at Baima Experimental Base of Nanjing Forestry University (31°35′ N, 119°09′ E). C. paliurus seeds were collected from Jinzhongshan county (24° 58′ N latitude, 110° 09′ E longitude), Guangxi province, China, in October 2018. After treated by exogenous GA3 (gibberellin A3) and stratification method (Fang et al., 2006), the germinated seeds were sown in nonwoven containers (10.0 cm height, 8.0 cm diameter) in April 2019.

Hydroponic experiment: After three months, uniform size seedlings (height: 40 ± 2.79 cm) were selected and transplanted to polypropylene containers (50L) with 1/2-strength Hoagland’s nutrient solution (pH 6.0 ± 0.2). Two weeks after hydroponic transplanting, four salt concentration (0%, 0.15%, 0.3%, and 0.45% NaCl) regimes were implemented in completely randomized design with three biological replicates for each treatment. The detailed information has been described in our previous study (Zhang et al., 2022a).

Pot experiment: After one-year growth in the nonwoven containers, the seedlings were transplanted into the big nonwoven containers (25 cm height, 20 cm diameter) and cut into 3-5 cm height in early spring in 2020. In February 2022, saplings with similar size were selected and all their stems were cut to 120 cm height, whereas in early April 2022, the selected saplings were transplanted from the nonwoven containers into plastic pots (26 cm height, 26 cm top diameter and 20 cm bottom diameter) containing peat: substrates of perlite: rotten bird dung: soil =5: 2:2:1 (v/v/v/v). The plastic pots were placed in plastic trays to prevent NaCl leaching. The substrate was a loam with pH 6.4, and the contents of total N, total P, and total K in the soil were 79.7, 66.5, 2.40, and 9.7 g kg⁻¹, respectively.

Salt treatments were conducted in early May 2022, and a completely randomized design was adopted with three replications per treatment and six plants per replication. Based on previous research (Zhang et al., 2022b), three levels of NaCl concentration were set up: CK (control, distilled water), T1 (0.2% NaCl) and T2 (0.4% NaCl). 1L solution were gradually add to the soil every three days (Chen et al., 2021), and electrical conductivity in the substrate was also monitored to keep the soil salt concentration relatively stable. Six complete and mature leaves were respectively collected from the upper, middle and lower positions of each sampled tree at the 45 days after the treatments (obvious differences were observed) ( Figure 1 ) and were immediately frozen in liquid nitrogen and stored at −80°C until needed for further analysis.

Plant materials were ground under RNase-free conditions. Trizol reagent kit (Invitrogen, Carlsbad, CA, USA) was used to extract RNA from 9 samples of the 3 treatments (CK, 0.2% NaCl and 0.4% NaCl); subsequently, MonScript RTIII All-in-One Mix with dsDNase kits (Monad, Nanjing, China) was used to acquire cDNA, following the manufacturer’s instructions. The qRTPCRs were performed on BiosystemsTM 7500 Real-Time PCR Systems (Monad, China). Primer Premier 6.0 (Premier Biosoft International, Palo Alto CA, USA) was used to design qRT-PCR primers for 12 genes ( Supplementary Table S1 ). SYBR Premix Ex Taq kit (Takara Biotechnology, Dalian, China) was applied to conduct qRT-PCR analysis. The cDNA diluted 20 times and an 18sRNA gene (Chen et al., 2019) were selected as the template and the internal standard, respectively. PCR reaction conditions were as 95 °C for 3 min; denaturation 5 s at 95 °C; 60 °C for 30 s; 40 cycles. Three technical and three biological replicates were used for each sample. After reaction, the relative expression levels of target gene and internal reference gene were calculated with the 2^−ΔΔCT method (Penfield, 2001).

One-way analysis of variance (ANOVA) was conducted to identify significant differences in the related gene expression among the treatments, followed by Duncan’s test for multiple comparisons. All statistical analyses were performed using IBM SPSS Statistics Version 22 software package (SPSS Inc., IBM Company Headquarters, Chicago, IL, USA). Data were presented as means ± standard deviation (SD).

Based on the Genome-wide data of C. paliurus, a total of 174 supposed CpbHLH proteins were discovered by using the HMMER software with default parameters. Subsequently, SMART and CD-Search were performed to confirm the existence of the conserved bHLH domain. After removing redundant sequences, 159 bHLH protein sequences of C. paliurus with typical complete bHLH domain were obtained and they were named CpbHLH1 to CpbHLH159 according to their location on chromosomes ( Figure 2 ; Supplementary Table S2 ). Sequence analysis showed that the average length of the CpbHLH proteins was 354 amino acids. The relative molecular weight (Nw) ranged from 10454.72 Da (CpbHLH39) to 175494.7 Da (CpbHLH86), whereas the isoelectric point (pI) ranged from 4.65 (CpbHLH48) to 9.66 (CpbHLH66) ( Supplementary Table S2 ).

To gain in-depth knowledge of the function of CpbHLH family, the bHLH domains of the CpbHLH proteins were searched and the presence of the conserved amino acid residues were analyzed based on multiple sequence alignment. The alignment results ( Figure 3 ) showed that the CpbHLH domains were composed of four conserved regions, namely one basic region, two helix regions and a loop region. Consistent with previous studies (Heim et al., 2003), the conservation of basic region and helix region is higher than that of the loop region. The bHLH domains of C. paliurus were made up of 79 amino acid residues, of which 24 were highly conserved (> 50% consensus ratio) and 8 were extremely conservative (> 75% consensus ratio). Among the 24 highly conserved amino acid residues, six conserved residues were found in the basic region (His-9, Ala-12, Glu-13, Arg-14, Arg-16, Arg-17), seven conserved residues were found in the first helix region (Ile-20, Asn-21, Arg-23, Leu-27, Leu-30, Val-31, Pro-32), one conserved residues were found in the loop region (Asp-64), and ten conserved residues were found in the second helix region (Lys-65, Ala-66, Ser-67, Leu-69, Ala-72, Ile-73, Tyr-75, Val-76, Lys-77, Leu-79).

It is generally believed that the basic region performs DNA binding functions, and is critical for the bHLH family to achieve its biological function (Carretero-Paulet et al., 2010). Therefore, the DNA-binding ability of the 159 CpbHLH proteins were predicted based on the conserved amino acid residues in the basic region ( Supplementary Table S3 ). The remaining 159 CpbHLH members were classified into three categories: G-box (His/Lys-9, Glu-13 and Arg-17), E-box (Glu-13 and Arg-16) and non-E-box (Glu-13 and Arg-16 do not appear together) in accordance with the classification method reported previously (Katiyar et al., 2012). The predicted results revealed there were 93 G-box-binding proteins, 43 non-G-box-binding proteins and 23 non-E-box-binding proteins in 159 CpbHLHs ( Supplementary Table S3 ).

In order to explore the evolutionary relationship among the CpbHLH members, the 159 CpbHLH proteins were aligned with 140 bHLH proteins from Arabidopsis, afterwards the phylogenetic tree was constructed using total 299 bHLH proteins based on the alignment ( Figure 4 ). In accordance with the classification of bHLH proteins from Arabidopsis and other plants (Heim et al., 2003; Li et al., 2020b; Li et al., 2021), 299 bHLH protein sequences were classified into 26 subfamilies, and were named from Ia to XV on the basis of the nomenclature of AtbHLHs proposed by Heim et al. (Heim et al., 2003). Figure 4 showed that the XII subfamily was the largest (contained 35 CpbHLH proteins), while the smallest subfamily (VI) contained only one CpbHLH protein. According to results from Heim et al. (Heim et al., 2003), CpbHLH proteins in the same subfamily would have similar functions, consequently, the clustering results of phylogenetic tree could contribute to predict the function of CpbHLH proteins.

Diversity of exon-intron structures, which could cause divergences in coding regions, is significant to the evolution of multiple gene families (Xu et al., 2012b). Hence, the gene structural characteristics of the CpbHLH family were investigated. The number of exons in the 159 CpbHLH genes varied from 1 to 13 ( Figure 5C ). In addition, 20 (12.6%) genes were intronless and distributed across subfamilies IIId, IIIe, VIIIa, VIIIb and VIIIc(2), while 13 (8.2%) genes contained one intron, and certainly the remaining genes had two or more introns. The 159 genes in different families varied widely in structure, including the number and relative location of introns and exons ( Figure 5C ). On the contrary, the intron/exon patterns of genes in the same subfamily had highly similarity, such as in subfamilies Ib(1) (five three-exon genes), Ib(2) (five three-exon genes), III(d+e) (nine one-exon genes), IVc (eight five-exon genes), and VIIIb (seven one-exon genes) ( Figure 5C ).

It is generally accepted that motifs figure prominently in interaction and signal transduction between different modules of the gene transcription process (Toledo-Ortiz et al., 2003). To further understand the evolutionary relationships among these CpbHLH proteins, the conserved motifs were analyzed by using MEME. Twenty motifs were identified and their sequences and length were counted ( Figure 5B , Supplementary Table S4 ). In addition, eight of twenty motifs were annotated by Pfam and CD-search ( Supplementary Table S4 ). Obviously, the composition patterns tended to be consistent with the results from our phylogenetic tree and gene structures, being resemble among genes within the same group, but varying greatly between groups ( Figure 5 ). The number of motifs in 159 CpbHLHs ranged from one (CpbHLH66) to nine (CpbHLH50). All 159 CpbHLH genes contained motif 1 and motif 2, except CpbHLH66, only containing motif 1 ( Figure 5B ). Interestingly, some conserved motifs were nested in specific groups. For example, motif 13 only existed in group Ia, motif 16 in group VIIIb, motif 18 in group XII, and motif 19 in group IX respectively ( Figure 5B ). This phenomenon might be the reason why functions for CpbHLH proteins tend to be specific to a particular group.

The highly differentiated sequences outside the conserved bHLH domain suggest that CpbHLH proteins may have a variety of biological functions. GO annotation of these 159 proteins was performed to understand the biological processes associated with CpbHLH genes. The results are shown in ( Figure 6 ; Supplementary Table S5 ). The identified CpbHLH proteins were classified into three main Gene ontology (GO) terms, which were CC (cellular component), MF (molecular function), and BP (biological process). Within MF category, the majority of CpbHLH proteins were annotated for “molecular function” (139/159), “nucleic acid binding” and “DNA binding”, respectively. These functions were closely related to the primary roles that TFs have. As for CC category, most of the CpbHLH proteins were assigned to cellular components and the nucleus (139/159). However, there were also a small number of CpbHLH proteins distributed in cytoplasm (8/159), organelle part (7/159), cytosol (4/159), symplast (CpbHLH37/117/132) and chloroplast (CpbHLH68/109) ( Figure 6 ; Supplementary Table S5 ). Furthermore, the BP aspect showed that CpbHLH proteins participated in various biological processes. Proteins annotated to be related to multiple biosynthetic and metabolic possessed the largest number of CpbHLHs (141/159). Besides, CpbHLH proteins may function in regulating biological processed, such as regulation of cellular process (111/159), transcription (109/159), DNA-templated (109/159) and gene expression (109/159). The BP analysis also showed that many CpbHLHs could respond to stimuli (46/159), including different types of biotic and abiotic stressors, while CpbHLH38/68/109 were predicted to be involved in respond to salt stress ( Figure 6 ; Supplementary Table S5 ).

Conserved motifs located in gene promoter regions are recognition and binding sites for proteins. In this study, a large number of cis-regulatory elements (CREs) of CpbHLH genes were identified, and they were classified into three main categories (plant growth and development, phytohormone responsive, as well as abiotic and biotic stresses) according to their roles ( Figure 7 ). Our result showed that CAT-box (105) and O2-site (86), which were involved in the meristem expression and zein metabolism regulation respectively were most frequently found motifs related to plant growth and development. On the contrary, the number of HD-Zip 1 (the differentiation of the palisade mesophyll cells), AACA-motif (involved in endosperm-specific negative expression) and MSA-like (cell cycle regulation) elements were 8, 3 and 2 respectively. Additionally, RY-element (seed-specific regulation) and GCN4_motif (endosperm expression) were also identified in the promoters of the CpbHLH genes ( Figure 7 ). The most common elements in phytohormone responsive category were ABRE (the abscisic acid-responsive element), CGTCA-motif and TGACG-motif (elements involved in MeJA responsiveness) and the TCA element (SA-responsive element) ( Figure 7 ). In the last category, a lot of important CREs related to plant abiotic stress were detected. Most abundant of these were the ABRE (drought response element), ARE (anaerobic induced response element), MBS (drought induced response element) and LTR (low temperature response element). Other stress response CREs, such as GC-motif (anoxic specific inducibility element), TC-rich (defense and stress response element) and ERE elements (oxidative stress responsive elements were also identified ( Figure 7 ).

Analysis of gene expression profiles is an effective way to determine gene functions. Hence, the leaves of C. paliurus treated with different salt concentrations (0%, 0.15%, 0.3%, and 0.45% NaCl) for 30 days in hydroponic experiment were sequenced and analyzed (Zhang et al., 2021). The raw sequencing data were submitted to the NCBI BioProject database under project number PRJNA700136. The RPKM (Reads Per Kilobase per Million mapped reads) values of 159 CpbHLH genes were obtained from the transcriptome data to estimate the expression levels of bHLH family members. However, CpbHLH119/121/138/151 were not analyzed because of the absence or low level of expression in the transcriptome data. Figure 8 showed that 155 of these genes were expressed in all concentrations of NaCl treatments with different expression patterns, providing evidence that CpbHLH genes are significantly affected by salt stress.

Based on the similarity of expression patterns, the 155 CpbHLH genes were clustered into 8 clusters, named A1-A8 ( Figure 8 ). The genes in cluster A1 were mainly expressed in the middle (0.30% NaCl) or high (0.45% NaCl) salinity condition and did not change significantly under low (0.15% NaCl) salinity condition. In contrast, CpbHLHs in cluster A6, A7 and A8 was strongly and preferentially expressed under low salt concentrations and down-regulated under high salt concentration. In cluster A2, the expression of CpbHLH genes did not change significantly under low and middle salt stress, but reached its highest value at high salinity treatment. However, expressions of most genes in cluster A4 varied with salt concentration treatments, and expression of these genes were all down-regulated under salt treatments and reached its lowest value at 0.45% NaCl treatment. However, very low expression levels of these genes in cluster A3 and A5 were observed at middle and high salt concentrations, respectively ( Figure 8 ). In particular, among these 155 genes, the expression of some genes were strongly induced or inhibited under salt stress. For example, compared with the CK, the expressions of CpbHLH36/74/75 in cluster A4 were down regulated by nearly folds of 3 in the low salinity treatment (0.15% NaCl), especially CpbHLH74 down regulated by nearly folds of 9 in the high salinity treatment (0.45% NaCl). Similarly, seven differentially expressed genes (DEGs) (CpbHLH68/69/71/108/146/152/158) were identified in the A5, A6 and A7, indicating a response to salt stress ( Figure 8 ).

Combining the results from both GO annotation and expression profiles analysis in hydroponic experiment, twelve salt-induced candidate genes (CpbHLH36/38/68/69/71/74/75/108/109/146/152/158) were selected for further qRT-PCR analysis using templates from pot experiment with three salt concentrations (0% NaCl, 0.2% NaCl and 0.4% NaCl) ( Figure 9 ). Notably, eight candidate genes (CpbHLH36/68/71/75/109/146/152/158) were up regulated or decreased dramatically under different salt treatments, indicating that the expression of these genes was significantly induced or inhibited under salt stress ( Figure 9 ). Among the eight genes, four genes (CpbHLH36/146/152/158) were down-regulated under salt stress, with three of these genes (CpbHLH146/152/158) being lowest expressed at 0.4% NaCl and one gene (CpbHLH36) being lowest expressed at 0.2% NaCl. On the contrary, three genes (CpbHLH68/71/109) were significantly induced by salt stress ( Figure 9 ). In particular, three genes (CpbHLH36/68/146) responded strongly to salt treatments. Compared to the control, the variation trend of their expression in the pot experiment was highly consistent with that in the hydroponic experiment ( Figure 8 ; Figure 9 ), indicating their vital functions in response to salt stress. For example, the expression level of CpbHLH36 in both experiments was strongly inhibited under salt stress, whereas the inhibition degree was greater in low salt concentration than in high salt concentration.

It was reported that bHLH proteins exert regulatory effects by forming homodimers or heterodimers between bHLH proteins or between bHLH and non-bHLH proteins (Herold et al., 2002; Hernandez et al., 2007). Thus, the interaction network of three candidate genes was predicted by STRING ( Figure 10 ), based on the CpbHLH homologous genes in A. thaliana. The investigation of CpbHLH146 (MYC2 ortholog) showed that it was involved in light, abscisic acid (ABA), and jasmonic acid (JA) signaling pathways and controlled additively subsets of JA-dependent responses with MYC3 and MYC4 ( Figure 10B , Supplementary Table S6 ). Among the proteins interacting with MYC2, those related to JA signaling pathway accounted for the majority, including JAZ1, JAZ3, JAZ5, JAZ8, JAZ10, JAZ12 and TILY7. Besides, PFT1 was determined as phytochrome and flowering time regulatory protein and the EIN3 probablely acted as a positive regulator in the ethylene response pathway ( Figure 10A , Supplementary Table S6 ). The predicted network for CpbHLH36 (NIG ortholog) showed that it plays central roles in regulating various proteins, and coincidently several of which were also involved in the jasmonic acid signaling pathway (JAZ1 and JAZ10) ( Figure 10A , Supplementary Table S6 ). Other proteins, GSTU1 and GSTU2, could be involved in the conjugation of reduced glutathione to a wide number of exogenous and endogenous hydrophobic electrophiles and have a detoxification role against certain herbicides, whereas bHLH11 and TRFL8 both function in DNA binding ( Figure 10A , Supplementary Table S6 ). Finally, the results of predicted network ( Figure 10C , Supplementary Table S6 ) also indicated that CpbHLH68 (ortholog of bHLH106) has crucial roles in DNA binding, whose function is the same as most of the proteins that interact with it. In addition, several interacting genes possibly regulate light responses, for example CRY1 and CPY2 are cryptochromes, and UVR2 and UVR3 involved in repair of UV radiation-induced DNA damage. However, PRMT4B has been identified as a positive regulator of oxidative stress tolerance that promotes the expression of antioxidant enzymes such as APX1 and GPX1 ( Figure 10A , Supplementary Table S6 ). Overall, the results of the protein interaction network analysis indicated that the three candidate genes interact with proteins of various functions, making them crucial players in regulating plant growth and stress responses.

Based on the whole genome of C. paliurus, 159 bHLH genes were systematically identified in the present study ( Supplementary Table S2 ). The number of CpbHLH genes was the same as that identified in tomato (Sun et al., 2015), but smaller than that in Arabidopsis (162 genes) (Toledo-Ortiz et al., 2003) and apple (175 genes) (Yang et al., 2017), whereas greater than that in grape (94 genes) (Wang et al., 2018a), strawberry (113 genes) (Zhao et al., 2018) and jujube (92 genes) (Li et al., 2019). Overall, 159 CpbHLH proteins were further categorized into 26 subfamilies ( Figure 4 ), according to the phylogenetic tree with the nomenclature protocol of bHLH proteins in C. paliurus and Arabidopsis (Heim et al., 2003), in agreement with results from previous studies (Pires and Dolan, 2010; Sun et al., 2015; Chu et al., 2018). However, the CpbHLHs were distributed almost evenly across 20 subfamilies, similar to Camellia sinensis (Sun et al., 2015) and O. fragrans (Li et al., 2020b). Moreover, our result indicated that no CpbHLHs were found in subfamily X, whereas the most CpbHLH members were detected in subfamily XII ( Figure 4 ), with the number of members in this family increasing from 17 in Arabidopsis to 22 in C. paliurus. Differences in the numbers of bHLH genes among plant species may be due to gene replication events or genome size or gene loss during evolution (Flagel and Wendel, 2009; Li et al., 2020b).

Based on the analysis of the conserved motif and intron/exon ( Figure 5B, C ), the results showed that CpbHLHs in the same subfamily of the phylogenetic tree were similar in genetic and motif structures, further confirming the accuracy of subgroup classification of phylogenetic tree ( Figure 4 ; Figure 5 ). Totally twenty motifs were identified in 159 CpbHLH proteins ( Figure 5B ). However, among them, motifs 1 and 2 existed in almost every CpbHLH protein and represented main components of the bHLH domain with high capability of conserved DNA binding, suggesting that the two motifs had very important implications about the functioning of bHLH genes (Zhang et al., 2020b). Nonetheless, the remaining 18 conserved non-bHLH domains can also feature separately in CpbHLHs in their respective subfamilies, similar to the other plant species (Chu et al., 2018; Li et al., 2020b). For example, most bHLH genes of subfamily III(d+e) in Panax ginseng (Chu et al., 2018) contained MYC-N structures (bHLH-MYC_N domain, Pfam : PF14215), which have been proved functioning in regulating the biosynthesis of phenylpropane. In this study, all CpbHLHs of III(d+e) also contained MYC-N structures (motif 5, 8, 10) ( Figure 5B ; Supplementary Table S4 ), implying that CpbHLHs of the same subgroup may have the similar roles. It was reported that gain/loss of exons and introns may result in the functional diversification of gene families (Xu et al., 2012a), whereas introns are related to gene evolution, and especieally the genes with few or no introns are more highly expressed in plants (Chung et al., 2006; Ren et al., 2006). In the present study, the intron-less CpbHLHs were distributed across subfamilies III (d+e) and VIIIb ( Figure 5C ), in accordance with the phenomenon in P. ginseng (Chu et al., 2018), apple (Yang et al., 2017) and Osmanthus (Li et al., 2020b), suggesting CpbHLHs of these subgroups could facilitates rapid and timely response to various stresses (Jeffares et al., 2008).

Transcriptional regulation is a basic process of gene regulation in response to stress signals and a mass of TFs are involved in regulating plant responses to a given stress (Riechmann et al., 2000). The results of GO annotation in this study showed the functions of the CpbHLH genes are diverse ( Figure 6 ; Supplementary Table S5 ), supporting that the bHLH TFs plays a crucial role in regulating plant growth, development and stress response (Shen et al., 2021). Several lines of evidence showed that salt stress had adverse effects on photosynthesis and the accumulation of secondary metabolites in C. paliurus (Zhang et al., 2021; Zhang et al., 2022a). Therefore, the detection of salt stress response genes from CpbHLHs will be helpful to achieve salt-tolerant breeding of C. paliurus.

The transcriptome sequencing analysis of salt treatments in the hydroponics provided specific expression data for the CpbHLHs, which makes it possible to further study the function of these genes. The RPKM values from our hydroponics showed that a large number of CpbHLH genes were induced/repressed under NaCl stress ( Figure 8 ). According to the RPKM data, ten significantly differentially expressed genes (CpbHLH36/68/69/71/74/75/108/146/152/158) were predicted to function in responding to salt stress ( Figure 8 ). Moreover, the molecular function annotations of 159 CpbHLHs indicated that three genes (CpbHLH38/68/109) strongly responded to salt stress ( Supplementary Table S5 ). Thus, the 12 genes mentioned above were predicted to be candidate genes in response to salt stress and were selected for further qRT-PCR analysis, using salt-treated templates collected from our pot experiment. The qRT-PCR results showed that the expression of three genes (CpbHLH36/68/146) strongly responded to the salt treatments ( Figure 9 ), and the variation trend of their expression levels was highly similar in the two salt stress experiments ( Figure 8 ; Figure 9 ), indicating that these genes were specific for the regulation of salt tolerance in C. paliurus.

Phylogenetic analysis can be used to derive orthogonal relationships based on sequence similarity and protein structure, while the most closely related bHLH genes in the phylogenetic tree may share a similar function (Wang et al., 2021). The existed research indicated that AtbHLH106 could enhance salt tolerance of plant by directly interacting with the G-box of salt tolerant genes (Ahmad et al., 2015), whereas the CpbHLH68 was clustered in the same clade that possess high bootstrap value with AtbHLH106 ( Figure 4 ), suggesting CpbHLH68 may be involved in response to salt stress. In addition, DNA sequences are decisive factors of the binding specificity between transcription factors and their genomic targets (Gordân et al., 2013), and our results from the DNA-binding ability of 159 CpbHLHs showed that CpbHLH68 was G-box-binding protein ( Supplementary Table S3 ), which further suggests that CpbHLH68, similar to AtbHLH106, may respond to salt stress by binding to G-box of target genes. Moreover, AtbHLH6 (ATMYC2) has been reported to exhibit a significant response to salt and drought stresses (Abe et al., 1997; Aleman et al., 2016), while AtNIG1 (a salt stress-responsive gene) was the first known TF participating in salt stress signal by binding calcium ions and bound to the E-box sequence (CANNTG) (Kim and Kim, 2006). Our study showed that CpbHLH36 and CpbHLH146 were clustered in the same clade with AtbHLH6(MYC2) and AtbHLH28(AtNIG1) ( Figure 4 ), suggesting that CpbHLH36 and CpbHLH146 are also E-box proteins ( Supplementary Table S3 ), and very likely to be involved in the regulation of salt stress signaling pathways.

In general, the function of a given gene can be inferred from its homologous genes (Yue et al., 2016; Qu et al., 2022). Therefore, Arabidopsis orthologs were used to predict the regulatory network of these three candidate genes (CpbHLH36/68/146) in this study. Some previous researches showed that AtMYC2 was involved in the regulation of ABA-inducible genes under drought stress conditions (Gordân et al., 2013) and could provide a possible mechanistic link between ABA signaling and JA signaling (Abe et al., 1997; Zhang et al., 2021). The predicted interaction genes of CpbHLH146 (MYC2 ortholog) were mainly involved in the regulation of JA signaling ( Figure 10A , Supplementary Table S6 ). The interaction of plant hormone ABA and JA played a major role in abiotic stress tolerance (Xiong et al., 2002; Zhang et al., 2012b) and ABA-dependent pathways the was one of important abiotic stress response signaling transduction pathways (Zhang et al., 2012a). The promoter region of most ABA regulatory genes contains many ABA responsive elements (Leonhardt et al., 2004; Yamaguchi-Shinozaki and Shinozaki, 2005; Fujita et al., 2011). In this study, a high occurrence of ABRE (ABA-responsive element) and CGTCA-motif (MeJA-responsive element) cis-acting elements was detected in the promoters of CpbHLH146 ( Figure 7 ). Thus, it can be inferred that this gene may have an important role in regulating stress resistance by regulating the expression of key genes in the ABA signaling pathway. Furthermore, most interaction genes of CpbHLH36 (AtNIG1 ortholog) and CpbHLH68 (bHLH106 ortholog) were mainly involved in DNA binding ( Figure 10A , Supplementary Table S6 ), which further supports our hypothesis that these two genes regulate plant salt stress mainly via recognizing G-box of target genes. Moreover, CpbHLH36 (AtNIG1 ortholog) was also interacted with some JA signaling pathway proteins ( Figure 10A , Supplementary Table S6 ), and it was in the same cluster of the phylogenetic tree with CpbHLH146 (MYC2 ortholog) ( Figure 4 ). Besides, the similar expression trend of CpbHLH36 was observed between pot experiment and hydroponic experiment, the same as to CpbHLH146 ( Figure 8 ; Figure 9 ). Therefore, it could be concluded that there is an indirect interaction between CpbHLH36 and CpbHLH146 at the protein level and these two genes coordinately control the expression of downstream genes, whereas the plant salt tolerance may depend upon the co-expression of these two genes.

In short, combined with the above results, CpbHLH36/68/146 could be the key putative candidates in response to salt stress in C. paliurus. However, characterizations of these three genes involved in the regulation of salt tolerance varied. CpbHLH36/68/146 are all G-box proteins, and may respond to salt stress by binding to G-box of target genes. Secondly, CpbHLH36 may participate in salt stress signal by binding calcium ions and regulating the expression of key genes in the JA signaling pathway. Thirdly, CpbHLH146 was very likely to be involved in the regulation of salt stress in ABA signaling pathways. Moreover, it is noted that there exists an indirect interaction between CpbHLH36 and CpbHLH146 at the protein level, thus we guess the salt tolerance of C. paliurus may depend upon the co-expression of these two genes.

In conclusion, it is the first report to identify the TF family based on the whole genome of C. paliurus. A total of 159 CpbHLH genes were detected and divided into 26 subfamilies, according to their evolutionary characteristics. In addition to investigating their structures and DNA-binding abilities, expression analysis from both the pot and hydroponic experiments and the regulatory network were also performed to determine which genes are most active for salt stress responses in this species. A total of 12 candidate genes were selected in response to salt stress, whereas the 3 genes (CpbHLH36/68/146) were further verified to be involved in regulating the salt tolerance of C. paliurus based on a pot experiment and protein interaction network analysis. Our findings would not only provide a basis for further understanding regulatory mechanisms of bHLH proteins TFs, but also drive progress in genetic improvement for the salt tolerance of C. paliurus.