The reference genome and protein sequences of B. napus were downloaded from http://yanglab.hzau.edu.cn/BnIR/ while the reference genome of B. rapa, B. oleracea and Arabidopsis was obtained from http://plants.ensembl.org/index.html. The Hidden Markov Model (HMM) file of the LOR domain was acquired from Pfam protein family database (http://pfam.xfam.org/) (El-Gebali et al., 2019). LOR genes were searched from B. napus reference genome using HMMER 3.3 (Potter et al., 2018). The preliminarily identified LOR family members from B. napus (BnLOR) were validated by blasting with LOR gene family from A. thaliana as queries and all candidate genes were further validated at CDD (https://www.ncbi.nlm.nih.gov/cdd/), Pfam and SMART (http://smart.embl.de/) to verify the existence of LOR domain. The length of the BnLOR protein sequences as well as the molecular weight (Mw) and the isoelectric points (pI) were predicted by the pI/Mw tool in ExPASy (http://www.expasy.ch/tools/pi_tool.html). Besides, the subcellular location of BnLOR genes were predicted by the Cell-PLoc web-server (http://www.csbio.sjtu.edu.cn/bioinf/Cell-PLoc/).

To identify the evolutionary relationship among LOR gene family from different species, the protein sequences of LORs from Arabidopsis, B. rapa, B. oleracea, and B. napus were aligned by the Clustal W algorithm with default parameters (Larkin et al., 2007). The ALN file was converted into MEGA 6.0 to build a neighbor-joining (NJ) phylogenetic tree with following parameters: Poisson model, and pairwise deletion for the reliability of interior branches (Tamura et al., 2013). The phylogenetic tree was modified and displayed using the iTol tool (https://itol.embl.de/).

The gene structures and protein conserved motifs of the BnLOR family members were analyzed in this study. The exon-intron structures of each BnLOR gene were extracted using gene structure view (advanced) in TBtools (Chen et al., 2020), which provided a graphical representation of the gene structures. The conserved motifs within the BnLOR protein sequences were identified using the Multiple Em for Motif Elicitation program (MEME, https://meme-suite.org/meme/) with default parameters set for maximum number of motifs, distribution of motif occurrences, and optimum motif width (Bailey et al., 2009). The identified exon-intron structures and conserved motifs were then visualized by TBtools (Chen et al., 2020).

The chromosomal positions and relative distances of the BnLOR family members were drafted using TBtools (Chen et al., 2020). The WGD landscape within B. napus was generated by MCScanX (Wang et al., 2012). To confirm WGD events, the shorter aligned sequence was required to cover over 70% of the longer sequence and the similarity of the aligned regions had to be over 70% (Gu et al., 2002; Yang et al., 2008). WGD events includes segmental duplication and tandem repeats. Tandem repeats are defined as two genes located in the same chromosomal fragment with a length of less than 100 kb (Wang et al., 2010b). Segmental duplications refer to two genes experienced polyploidization followed by chromosome rearrangements (Yu et al., 2005). The synteny map was created using Circos (Krzywinski et al., 2009). Additionally, the MCScanX was used to show the collinearity relationship between B. napus, B. rapa and B. oleracea. Finally, the KaKs_Calculator 2.0 program was utilized to calculate the non-synonymous (ka) and synonymous (ks) substitutions of each duplicated LOR gene (Wang et al., 2010a).

The promoter sequence (about 2000 bp upstream of the translation start codon) of BnLOR family members were extracted from the B. napus reference genome. PlantCARE program (http://bioinformatics.psb.ugent.be/webtools/plantcare/html/) was used to predict the putative CRE responsive to transcription binding sites, hormone treatments, biotic and abiotic stresses. PlantRegMap program (http://plantregmap.gao-lab.org/) was used to predict the potential TFs using B. napus as the target species (Jin et al., 2017).

To identify orthologous gene clusters among multiple species, the reference genomes of eight commonly studied species including Arabidopsis, B. oleracea, B. rapa, B. napus, Glycine max (G. max), Oryza sativa (O. sativa), Triticum aestivum (T. aestivum) and Zea mays (Z. mays) were analyzed using the Ortho Venn 2.0 tool (https://orthovenn2.bioinfotoolkits.net/home). The e-value threshold and the inflation value were set at 1e-5 and 1.5, respectively. The protein sequences of LORs in these species were uploaded, analyzed, and visualized. (Wang et al., 2022; Wang et al., 2023).

To determine the gene expression pattern of BnLOR gene family, transcriptome file of various tissues including blossomy pistil, flower, leaf, ovule, pericarp, pistil, root, silique, sepal, stamen, stem, and wilting pistil were obtained from a previous study by Sun et al. (2017) under the project ID of PRJNA394926 in the NCBI. Moreover, transcriptome data of B. napus under dehydration, salt stress, ABA treatment and cold stress conditions were obtained from Zhang et al. (2019) under the project ID of CRA001775. Differential expression analysis of BnLOR genes was performed using the DSEeq2 R package and the heatmaps were created by TBtools software (Chen et al., 2020).

Homology modeling was used to create the three-dimensional (3D) structures of BnLOR proteins from their amino acid sequences. The amino acid sequences of BnLORs were submitted to SWISS-MODEL program (https://swissmodel.expasy.org/) for model generating. The quality of the models was assessed using ERRAT and PROCHECK tests through SAVES program (https://saves.mbi.ucla.edu/). The resulting 3D structures of BnLOR proteins were visualized using the Pymol software (Karnes and Shaikh, 2021).

One leading cultivar of B. napus (ZD622) was chosen in this research to validate the relative expression levels of BnLORs under salinity and ABA stress (Ma et al., 2022). Mature and healthy seeds of ZD622 underwent sterilization by submersion in 1% NaClO for 15 minutes. The seeds were thoroughly rinsed with distilled water before being germinated in a petri dish for 24 hours with a damp filter paper placed underneath. After germination, the seedlings were transferred to a 12 cm² growth box containing full-strength Hoagland’s solution, along with a sponge placed inside. Germinated plants were shifted to growth chamber with 16-hr photoperiod, 24/16 day/night temperature, 300 μmol/m²/s light intensity and 60% to 70% humidity. After one week, uniformly developed seedlings were subjected to salt stress (200 mM) and ABA treatment (25 μM). Leaf samples were immediately frozen using liquid nitrogen and stored in a -80°C refrigerator in preparation for RNA isolation at 0 h, 4 h, and 24 h after treatment. Each sample had at least three replicates.

The MiniBEST Plant RNA Extraction Kit (Takara, Japan) was employed to isolate total RNA from leaf samples. The PrimeScriptTM RT reagent Kit (Takara, Japan) was used to carry out reverse transcription reactions with 1 μg RNA. The synthesized cDNAs were used as templates for qRT-PCR by employing the SuperReal PreMix Plus/SYBR Green (Tiangen, China) on a Bio-Rad CFX96 (BIO RAD, USA). The qRT-PCR conditions were elucidated in our recently published research (Wang et al., 2022). B. napus actin-7 was used as the internal control gene. All qRT-PCR primers used were listed in Table S1 . The relative expression levels were calculated using the 2^–ΔΔCt method with three replicates each (Livak and Schmittgen, 2001). Statistical analysis of the data was performed using the SPSS 20.0 statistical package (SPSS, Chicago, IL, USA). One-way variance analysis (ANOVA) was conducted followed by Duncan’s multiple range test (p < 0.05) to determine significant differences. Figures were generated using GraphPad Prism 7.0 (Yang et al., 2022).

According to the HMM file (PF04525) of LOR gene family, 72 candidate BnLOR genes were identified through model search of the B. napus reference genome. After filtering out the redundant genes and alternative splicing, the existence of LOR domain in candidate genes was further verified by CDD, Pfam and SMART program. As a result, 56 LOR genes in B. napus with complete open reading frame and LOR domain were identified. Their details, along with their encoded proteins, were listed in Tables S2 , S3 . The predicted length of BnLOR proteins range from 90 amino acids (BnLOR15/48) to 257 amino acids (BnLOR17) ( Tables S2 , S3 ). The MWs of BnLOR proteins ranged from 10.49 kDa (BnLOR48) to 29.28 kDa (BnLOR17), suggesting structural variance and functional diversity among members of this gene family ( Table S2 ). The pIs of 15 BnLOR proteins were less than 7, indicating that they were acidic proteins ( Table S2 ). The rest 41 proteins were alkalinity proteins with pIs higher than 7 ( Table S2 ). The largely variance of pIs indicated the possibility of the proteins existing in different parts of cells. The grand average of hydropathicity (GRAVY) prediction results showed that all proteins were hydrophilic ( Table S2 ). The instability index of 25 proteins were lower than 40, indicating that these proteins were relatively stable, while the rest 31 proteins were relatively unstable ( Table S2 ). The aliphatic index of 56 BnLOR proteins fluctuated from 71.48 (BnLOR11) to 100.88 (BnLOR28). Aliphatic index is positively correlated with the thermal stability of protein, suggesting that the thermal stability of the BnLOR proteins was largely variant. Besides, the prediction of subcellular localization of BnLOR proteins exhibited significant variation. Among the 56 identified BnLOR proteins, 25 were exclusively localized in a single compartment such as nucleus, cytoplasm, or chloroplast, while the remaining 31 proteins were found in at least two different subcellular locations. Notably, chloroplast was the most frequent localization site for BnLOR proteins (75.00%), followed by nucleus (48.21%), cytoplasm (33.93%), cell membrane (14.29%), mitochondrion (14.29%), peroxisome (12.50%), cell wall (10.71%), and vacuole (3.57%).

The phylogenetic tree among B. napus (56 BnLORs), model plants Arabidopsis (20 AtLORs), ancestral species B. rapa (33 BrLORs) and B. oleracea (30 BoLORs) was constructed by NJ method. According to its topological structures, BnLORs could be divided into three subgroups ( Figure 1 ). Subgroup I was the largest group with 27 BnLOR genes and could be further divided into four clades (Clade A to D) ( Figure 1 ). Subgroup II (Clade E) was the smallest group with only 3 BnLOR genes ( Figure 1 ). Subgroup III had 26 BnLOR genes and could be further divided into three clades (Clade F to H) ( Figure 1 ). Clade A, C and H were relatively large, each contains 10 or more family members. Clade D and E had only 2 or 3 members each while Clade B, F and G had 4 to 8 family members, respectively. Both Subgroup I and III had plenty of members and branches, indicating that the structures and function of these two subgroups were relatively diverse. Each clade had LOR family members from four species (B. napus, Arabidopsis, B. rapa and B. oleracea), indicating that the phylogenic relationship among these four species were relatively close ( Figure 1 ). At the end of the phylogenetic tree branch, there were 46 homologous gene pairs, among which 22 were Bn-Bo homologous gene pairs, 17 were Bn-Br homologous gene pairs and 2 were Br-Bo homologous gene pairs ( Figure 1 ). Bn-Bo and Bn-Br homologous gene pairs could be found in every clade, but their distribution rates were inconsistent. It is noticed that all homologous gene pairs in Clade B, C, D and F belonged to Bn-Bo or Bn-Br homologous gene pairs ( Figure 1 ).

The reference genome sequences of BnLORs were submitted to GSDS server to display their gene structures. Results showed that the exon-intron composition from Clade A to Clade D was similar ( Figure 2 ). Each of them owned 3 exons separated by 2 introns apart from BnLOR48/15/53, which had one intro inside two exons ( Figure 2 ). Meanwhile, the motif structures from Clade E to H were also similar ( Figure 2 ). Most of them contained two exons and one intron with a few exceptions ( Figure 2 ). BnLOR3 and BnLOR54 had three exons separated by two introns; BnLOR47/16/17 only had one exon; BnLOR20 had two exons and two introns ( Figure 2 ). Among all the 56 BnLORs, 19 BnLOR family members had no upstream or downstream non-coding regions regulating the transcription of coding regions ( Figure 2 ).

The reference protein sequences of BnLORs were uploaded to MEME Suite to identify their conserved motif. As a result, the MEME program identified 15 conserved motifs (p<0.05) with the length varied from 11 to 36 amino acids ( Figure 2 ; Supplementary Figure S1 ; Table S4 ). The range of conserved motifs in BnLOR proteins was from 4 to 52 ( Figure 2 ). Most of the BnLOR proteins contained motif 3 while only 4 proteins (BnLOR31/1/11/36) contained motif 14 ( Figure 2 ). It was noticed that the motif structures of these BnLOR proteins showed high similarities. Most proteins had motif 1 at the head and motif 2 at the tail ( Figure 2 ). Besides, BnLOR proteins on the same clade or adjacent clade had similar motif structures. For example, all proteins in Clade A have the same motif composition except for BnLOR50 ( Figure 2 ). Each protein in Clade F shared the same motif composition ( Figure 2 ). In Clade G, every protein had the same motif distribution except BnLOR14 and 42, which lacked motif 12 ( Figure 2 ). Most proteins from Clade A to G had motif 6 ( Figure 2 ). At the same time, we also noticed that the distribution of motifs in BnLOR proteins showed specificity. For instance, motif 11 only existed in Clade C while motif 10 only existed in Clade G and H ( Figure 2 ). Motif 15 only appeared in Clade F while motif 13 only appeared in Clade C ( Figure 2 ). Motif 14 only existed in Clade H and there was no motif 12 in Clade C or H ( Figure 2 ).

The position of 56 BnLOR gene were retrieved from the B. napus reference genome and their position was visualized and mapped on chromosomes with TBtools (Chen et al., 2020). To facilitate identification, these genes were named as BnLOR1 to BnLOR56 based on their chromosomal location ( Figure 3 ). Our analysis revealed that the distribution of BnLORs across the A and C genomes of B. napus was uneven, with 29 genes located on the A genome and 27 genes located on the C genome ( Figure 3 ). Notably, chromosomes A3 and C3 exhibited the highest number of BnLOR genes, with 7 and 8 genes, respectively ( Figure 3 ). There were 4 genes on chromosome C6 and C8 and 3 genes on chromosome A6, A7, A9, C4 ( Figure 3 ). Two genes were located on chromosome A1, A2, A4, A5, C1, C5 and C9, one BnLOR gene was located on chromosome A10 and C7 while no BnLOR gene was found on chromosome A8 and C2 ( Figure 3 ). For random chromosomes, there were 0 to 3 BnLOR genes, respectively ( Figure 3 ).

WGD has been widely recognized as a crucial impetus underlying the evolution of species, providing the foundation for morphological and physiological changes in plants. WGD events include tandem repeats and segmental duplication. B. napus was originated from the spontaneous hybridization between B. rapa and B. oleracea followed by allopolyploidization, leading to the prevalence of paralogous genes with multiple copies in rapeseed. Thus, in this study, we aimed to gain insights into the expansion of BnLOR genes by analyzing the WGD. As a result, 42 pairs of gene duplication events occurred in 28 BnLOR family members on 19 chromosomes ( Figure 4 ; Table 1 ). Among them, there were five tandem repeats (BnLOR9/8, BnLOR16/17, BnLOR17/18, BnLOR26/27, BnLOR50/51) ( Table 1 ). The other 37 were all segmental duplication ( Table 1 ). Our findings suggest that BnLOR genes have undergone a considerable number of duplication events throughout the evolutionary process, thereby contributing to the rapid expansion of the BnLOR gene family.

To elucidate the phylogenetic relationship among LOR family members, we employed MCScanX to conduct collinearity analysis across B. napus, B. rapa, and B. oleracea genomes. Our analysis identified 148 sets of collinearity relationships, including 78 pairs of collinear genes between B. napus and B. rapa and 70 pairs of collinear genes between B. napus and B. oleracea ( Figure 5 and Table S5 ). The collinear genes on each chromosome were distributed on two to several chromosomes. For example, the collinear genes of B. napus on Chromosome A2 were located on Chromosome A2, A7 of B. rapa and on Chromosome C2 and C8 of B. oleracea ( Figure 5 ). Notably, chromosomes A3 and C3 exhibited the largest number of collinear gene pairs.

To further explore the diversity of duplicated gene pairs during evolution, we calculated the Ka/Ks value. Except for one gene (BnLOR15) that could not be calculated in B. napus, the Ka/Ks value of all the segmental duplications and tandem repeats of BnLOR genes were less than 1, indicating that BnLOR genes were under the pressure of purifying selection during the evolutionary process of B. napus ( Table 1 ).

After uploading the promoter sequence of BnLORs onto PlantCARE web server, a total of 1333 CREs with annotations were discovered. Nearly half (41.34%) of the CREs were associated with light responsiveness, such as AE-box, Box 4, ATCT-motif, GATA-motif, G-Box, GT1-motif, ACE, MRE, GA-motif, TCT-motif, I-box, AT1-motif, etc. 32.11% of them were involved in hormone responsiveness, such as ABRE, P-box, CGTCA-motif, TCA-element, TGA-element, TATC-box and TGACG-motif. Among them, ABRE was an ABA responsive CRE. 49 out of the 56 BnLOR promoters had ABRE element, indicating that most of BnLORs might be inducted by ABA treatment. 18.38% of the CREs were related to stress responsiveness, such as LTR, ARE, TC-rich repeats, GC-motif, MBS, and WUN-motif. About 80% of the CREs in BnLOR promoter regions were associated with anaerobic induction, implying that BnLOR gene family might confer the ability to survive under inundation stress or high-altitude environments. CREs related to circadian control were also detected in the BnLOR promoter regions, suggesting that this gene family might be regulated by light induction and circadian rhythms. Besides, CREs involved in metabolic regulation and meristem expression were also discovered.

This study focused on CREs in response to hormone treatments such as auxin responsive elements (TGA element and AuxRR core), ethylene responsive element (ERE), gibberellin (GA) responsive elements (GARE motif and TATC box), SA responsive elements (SARE), ABA responsive elements (ABRE) and methyl jasmonate (MeJA) related motif (TGACG motif and CGTCA motif) (Song et al., 2020). Besides, CREs associated with abiotic stress were also our focus, such as low temperature responsive element (LTR), heat stress responsive element (HSE), anaerobic induced elements (ARE), drought-induced elements (MBS), etc. (Song et al., 2020). Results showed that ABA and MeJA responsive CREs were the most abundant in the promoter regions of BnLOR genes, with 135 and 154 elements respectively, followed by auxin (47), GA (41) and SA (38) responsive elements ( Figure 6A ; Table S6 ). Among the stress responsive elements, the anaerobic induced elements were the most (123). Moreover, a relatively high number of low-temperature responsive elements, defense responsive elements and stress responsive elements were also detected, with 40, 39 and 37 CREs respectively. Generally speaking, BnLOR genes should be able to respond to various abiotic stresses and hormone treatments. However, we noticed that some genes contained very limited CREs such as BnLOR17/14/31, indicating that their expression might weakly or couldn’t respond to abiotic stresses and hormone treatments.

To further understand the regulation network of BnLORs, the putative TFs regulating BnLOR family members were also predicted. As a result, 56 BnLORs were regulated by 347 TFs ( Figure 6B ; Table S7 ). Except for BnLOR56 which couldn’t be retrieved, the rest 55 BnLOR genes were regulated by 1 (BnLOR38) to 15 (BnLOR49) TFs. Among the 347 TFs, NAC had the largest amount (31), followed by C2H2 (30), Dof (25), ERF (25), AP2 (19), B3(18), bHLH (17) and MIKC_MADS (17). These eight types of TFs occupied more than 50% of the total numbers ( Figure 6B ; Table S7 ). Most of these TFs are related to abiotic stresses and hormone treatments. For instance, The NAC Transcription factor family is unique to plants and plays a crucial role in plant growth and development as well as biotic and abiotic stress resistance (Wu et al., 2020). Dof proteins are involved in various biological processes, including but not limited to endosperm development, plant defense, seed germination, gibberellin response and auxin response (Yu et al., 2019). AP2/ERF proteins play vital roles in regulating plant growth and development and various abiotic stress response (Wu et al., 2021). It has noticed that BnLORs were mainly regulated by these TFs, further supporting the prediction that BnLORs perform crucial functions in the regulation of abiotic stresses and hormone treatments.

To deepen our understanding about the evolutionary history and the orthologous relationship among LOR genes, orthologous gene clusters in different species were compared and annotated. The quantity of orthologous proteins in each cluster exhibited variability, with the number ranging from 5 to 37 ( Figure 7 ; Table S8 ). Twenty-nine orthologous proteins in B. oleracea and 28 orthologous proteins in B. rapa were predicted to have similar conserved domains to BnLOR proteins. Only 6 orthologous proteins were discovered in O. sativa, showing a large difference, this might be due to their early evolutionary differentiation. For the other six species (Arabidopsis, B. oleracea, B. rapa, G. max, T. aestivum and Z. mays), the number of orthologous proteins ranged from 8 (Z. mays) to 14 (Arabidopsis). Ten BnLORs proteins only had orthologous proteins in B. oleracea while another 10 BnLOR proteins only had orthologous proteins in B. rapa. The orthologous proteins of BnLOR5 could only be found in G. max. None of the BnLORs had orthologous proteins in every species. BnLOR49 had orthologous proteins in up to 6 species except B. rapa. Additionally, BnLOR4/29/44 proteins had orthologous proteins in 5 different kinds of species. Two orthologous clusters, comprising a total of 31 proteins, were found to be overlapping in each species ( Figure 7 ; Table S8 ). Furthermore, 9 singleton proteins (BnLOR8/18/22/23/30/45/46/50/56) didn’t have any orthologous proteins in any of the studied species.

The expression profiles of 56 BnLOR genes among 12 distinct tissues including blossomy pistil, flower, leaf, ovule, pericarp, pistil, root, silique, sepal, stamen, stem, wilting pistil were obtained by Sun et al. (2017) under the project ID of PRJNA394926. Results indicated that some members of the BnLOR gene family exhibited closely resembling expression patterns across diverse tissues of B. napus ( Figure 8A ; Table S9 ). For example, BnLOR28/29/35 had high expression levels in almost every tissue while BnLOR4/5/33/39/43 had relatively high expression levels in each tissue. However, the expression levels of BnLOR1/16/18/22/46/54/56 were relatively low in every tissue. BnLOR22 were found not expressed in any tissue. These 8 genes with low relative expression levels were all belonged to Clade H, suggesting that genes in this clade might be pseudogenes or act in response to biotic/abiotic stresses or hormone treatments. Certain genes in the BnLOR family displayed tissue-specific expression patterns. For instance, BnLOR13/21/40/45 were mainly expressed in stamen. These four genes also had a small amount of expression in blossomy pistil, wilting pistil and flower, but they didn’t express in other tissues, suggesting that these genes might have a potential association with reproductive growth. BnLOR6 and BnLO32 were also highly expressed in reproductive organs, but their distribution was different. Both of them were largely expressed in stamen and flower, followed by wilting pistil and blossomy pistil. BnLOR10 and BnLOR12 had high expression levels in ovule and moderate expression levels in other tissues. BnLOR14 and BnLOR42 demonstrated elevated expression levels in blossomy pistil with moderate expression levels in other tissues, indicating that these two genes might involve in plant reproductive growth, but might regulate reproductive growth through different pathways. BnLOR19 was highly expressed in pericarp, suggesting that it might take part in the vegetative growth. The remaining genes were slightly expressed in different tissues.

To examine the initial changes in gene expression, we obtained the transcriptional profile of B. napus under temperature, salinity and ABA stress under the project ID of CRA001775 (Zhang et al., 2019). The stress treatments consisted of drought stress, salt stress (200 mM), ABA treatment (25 μM), and low temperature stress (4 °C). The differentially expressed genes (DEGs) with a false rate less than 0.5 and an absolute fold change over 2 were chosen and displayed in Figure 8B . In general, the relative expression levels of most BnLOR gene at 4 h was higher than that at 24 h under salt stress, ABA treatment and cold stress while and the situation was similar under dehydration stress, indicating that the response speed of mRNA was very fast in BnLOR gene family ( Figure 8B ; Table S10 ). The relative expression levels of BnLOR5/29/33/35/39/42/43 genes showed a significant up-regulation while BnLOR1/6/13/16/21/22/32/40/45/50/54 were significantly down-regulated under four different treatments. Among them, BnLOR33 exhibited significant upregulation in response to cold stress, indicating its potential significance in facilitating adaptation to such stress. For the remaining 19 BnLOR genes, their relative expression levels varied under different stress. Almost all of these 19 genes could increase their expression levels under cold stress, but the degree was different, and the relative expression levels at 4 h were larger than 24 h. Meanwhile, the expression of BnLOR4/8/9/14/34/41 was partially up regulated at 4 h and decreased at 24 h under dehydration stress. Several other genes demonstrated slight up or down regulation in response to dehydration, salt stress, and ABA treatment.

To further verify the expression pattern under salinity and ABA stress, twelve BnLOR family members were chosen at random for qRT-PCR analysis. The relative expression levels of these genes had altered in varying degrees under salt stress or ABA treatment, indicating that these genes were capable of being induced by salinity and ABA stress ( Figure 8C ; Table S10 ). Under salt stress, the relative expression levels of BnLOR4/29/39/44/51 increased dramatically at 4 h and went down a little bit at 24 h, but still showed an overall upward trend. The remaining seven genes increased at the initial stage of salt stress but decreased dramatically compared to control after 24 h, indicating an exceptionally swift response under salt stress. Under ABA treatment, the relative expression levels of BnLOR4/39/44/51 exhibited an upward trend, with a larger increase observed at 4 h compared to 24 h. Conversely, the relative expression levels of BnLOR3 and BnLOR28 decreased at both 4 h and 24 h after treatment. The relative expression levels of the remaining six genes increased at the initial stage and then decreased over time, as depicted in Figure 8C and Table S10 . Overall, the expression pattern trends of qRT-PCR and transcriptome file were relatively consistent, but there were some differences in the specific values.

To deepen our understanding about the structures of BnLOR proteins, homology modeling was carried out to predict the 3D structures of BnLOR proteins. The predicted models were generated based on reported templates and optimized to maximize alignment coverage, percentage identity, global model quality estimate (GMQE), and overall quality factor. Models were considered reliable if the sequence identity between the template and the BnLOR protein was greater than 30%. Results showed that the 3D structures of 31 BnLOR proteins could be predicted by SWISS-MOLDEL ( Figure 9 ; Table S11 ). The LOR protein family was discovered to exhibit structural similarities to the mammalian PLSCR (Phospholipid scramlase) protein family (Baig, 2018). X-ray crystallography technique was employed to determine the 3D structures of AtLOR1 and found that this protein comprised a 12-stranded β-barrel structure that encompassed a central C-terminal α-helix (Bateman et al., 2009). According to this structural model, sequence conservation was found to be mainly located within the secondary structures and β-hairpin turns between strands 2 to 3 and 4 to 5 (Bateman et al., 2009). All 31 proteins shared the same template 2q4m.1.A. The tertiary structures of the BnLOR proteins were relatively similar and complex with a few exceptions. For the 31 BnLOR proteins with predicted 3D structures, BnLOR15/48/50/53 were significantly different from the other 27 members, with no central C-terminal α-helix and less than 12 stranded β-barrels. Among these four genes, the expression pattern of BnLOR15/48/53 couldn’t be detected by RNA-seq, indicating that they were either low expressive genes or pseudogene. The relative expression level of BnLOR50 could be induced by abiotic stress suggests that it may be involved in facilitating adaptation to such stress.