Alternative Splicing Analysis Revealed the Role of Alpha-Linolenic Acid and Carotenoids in Fruit Development of Osmanthus fragrans

Alternative splicing refers to the process of producing different splicing isoforms from the same pre-mRNA through different alternative splicing events, which almost participates in all stages of plant growth and development. In order to understand its role in the fruit development of Osmanthus fragrans, transcriptome sequencing and alternative splicing analysis was carried out on three stages of O. fragrans fruit (O. fragrans “Zi Yingui”). The results showed that the proportion of skipping exon events was the highest in all three periods, followed by a retained intron, and the proportion of mutually exclusive exon events was the lowest and most of the alternative splicing events occurred in the first two periods. The results of enrichment analysis of differentially expressed genes and differentially expressed isoforms showed that alpha-Linolenic acid metabolism, flavonoid biosynthesis, carotenoid biosynthesis, photosynthesis, and photosynthetic-antenna protein pathways were significantly enriched, which may play an important role in the fruit development of O. fragrans. The results of this study lay the foundation for further study of the development and maturation of O. fragrans fruit and further ideas for controlling fruit color and improving fruit quality and appearance.


Introduction
Alternative splicing (AS) refers to the process in which pre-mRNA finally produces a variety of different transcription products by changing splicing sites during gene expression, which is an important regulatory mechanism during gene expression [1]. Alternative splicing was first described in the study of the adenovirus hexon gene, which showed that a single gene could produce multiple mRNAs with different functions [2]. The discovery overturned the theory of molecular biology at the time that "one gene corresponds to one protein." Since then, alternative splicing, as a post-transcriptional processing mechanism of genes, has been found to exist widely in eukaryotes and is the main form of regulating gene expression in eukaryotes [3]. Alternative splicing of pre-mRNA was believed to be one of the important causes of protein functional diversity, which allows a gene to encode multiple different transcriptional and protein products, greatly increasing protein diversity and the complexity of gene expression, and was an important factor in regulating gene expression and generating proteomic diversity [4].
There are five main types of alternative splicing events, including skipped exon, mutually exclusive exon, alternative 5 splice site, alternative 3 splice site, and retained intron [5]. The proportion of variable splicing events in genes varies among species, and species in the same community have the same types of major variable splicing events. The main type of alternative splicing events in fungi and protists is intron retention. The frequency of skipped exon events was also higher in plants than in fungi and protozoa [6]. events, the number of SE was the highest, 15,516, accounting for 52.43% of all the alter-native splicing events, and 13,074, accounting for 15.24% of all the alternative splicing events at A3SS. A total of 7936 (9.25%) sites were variable at A5SS. There were 2524 MXEs in total, accounting for 2.94%. The number of RI was 17,274 (20.14%) ( Figure 1C, Table  S3). Most of the alternative splicing events took place in the first two stages ( Figure 1D).

Analysis of Differentially Expressed Alternative Splicing Events
Differentially expressed genes (DEGs) and differentially expressed alternative splicing (DAS) events at the different development stages were identified using Log2FoldChage > 2 and q-value < 0.05. There were 10,807 differentially alternatively spliced genes and 14,912 genes belonging to differentially alternatively spliced events in the first stage compared with the second stage and 125 differentially alternatively spliced genes and 675 differentially spliced genes belonging to differentially alternatively spliced events in the second stage compared with the third stage. There were a total of 69 genes that were both differentially alternative splicing genes and that had differential alternative splicing events occurring in all three stages (Figure 2A). The corresponding gene families were matched according to the existing gene annotation files to obtain the gene families to which the differentially expressed genes belonged. It finds that the transcription factors (TF) were mainly concentrated in the bHLH family, AP2-EREBP family, MYB family, and MADS family ( Figure 2B; Table S4). By comparing the differentially expressed genes with the existing splicing factor family data of Arabidopsis thaliana, 17 splicing factor (SF) families were obtained, in which glycine-rich protein, 17S U2 snRNP, related to spliceosome, 35S U5-associated proteins, and SR protein played a crucial part ( Figure 2B; Table S5). Five alternative splicing events were detected, including skipped exon (SE), the alternative 5 splicing site (A5SS), the alternative 3 splicing site (A3SS), mutually exclusive exon (MXE), and retained intron (RI) ( Figure 1B). In this study, we analyzed the alternative splicing events of O. fragrans fruit transcription genes and identified 85,781 alternative splicing events in O. fragrans fruit. According to the types of alternative splicing events, the number of SE was the highest, 15,516, accounting for 52.43% of all the alternative splicing events, and 13,074, accounting for 15.24% of all the alternative splicing events at A3SS. A total of 7936 (9.25%) sites were variable at A5SS. There were 2524 MXEs in total, accounting for 2.94%. The number of RI was 17,274 (20.14%) ( Figure 1C, Table S3). Most of the alternative splicing events took place in the first two stages ( Figure 1D).

Analysis of Differentially Expressed Alternative Splicing Events
Differentially expressed genes (DEGs) and differentially expressed alternative splicing (DAS) events at the different development stages were identified using Log2FoldChage > 2 and q-value < 0.05. There were 10,807 differentially alternatively spliced genes and 14,912 genes belonging to differentially alternatively spliced events in the first stage compared with the second stage and 125 differentially alternatively spliced genes and 675 differentially spliced genes belonging to differentially alternatively spliced events in the second stage compared with the third stage. There were a total of 69 genes that were both differentially alternative splicing genes and that had differential alternative splicing events occurring in all three stages (Figure 2A). The corresponding gene families were matched according to the existing gene annotation files to obtain the gene families to which the differentially expressed genes belonged. It finds that the transcription factors (TF) were mainly concentrated in the bHLH family, AP2-EREBP family, MYB family, and MADS family ( Figure 2B; Table S4). By comparing the differentially expressed genes with the existing splicing factor family data of Arabidopsis thaliana, 17 splicing factor (SF) families were obtained, in which glycine-rich protein, 17S U2 snRNP, related to spliceosome, 35S U5-associated proteins, and SR protein played a crucial part ( Figure 2B; Table S5).

Gene Ontology and Kyoto Encyclopedia of Genes and Genomes Enrichment Analysis
In order to clarify the functional categories of differentially expressed genes du the development of O. fragrans fruit, we performed GO enrichment analysis on the ferentially expressed gene sets of comparison between two adjacent periods (S1-S2 S2-S3), and the top 10 GO items significantly enriched were listed. Compared with first and second periods, DEGs mainly included lipid metabolic process, small mole metabolic process, chloroplast, plastid, phosphotransferase activity, alcohol group a ceptor, kinase activity, and oxidoreductase activity ( Figure 3A), and DAS events en ment in small molecule metabolic process, chloroplast, plastid, phosphotransferas tivity, alcohol group as acceptor, kinase activity, and oxidoreductase activity (Figure Following consideration of the first and second segments, both differentially alternat spliced genes and differentially alternatively spliced events were enriched in small ecule metabolic process, chloroplast, plastid, phosphotransferase activity, alcohol g as acceptor, kinase activity, and oxidoreductase activity ( Figure 3A). The third period contrasted with the second period, with a significant concentration of DEGs in the m metabolic process, alpha-amino acid metabolic process, mRNA process cell perip and oxidoreductase activity ( Figure 3B), and DAS events concentration on mRNA abolic process, alpha-amino acid metabolic process, mRNA process, cell periphery, oxidoreductase activity ( Figure 3B). DEGs and DAS between the second and third st were accompanied by the mRNA metabolic process, alpha-amino acid metabolic pro mRNA process, cell periphery, and oxidoreductase activity ( Figure 3B). Distinction tween these three periods collectively enrich the oxidoreductase activity.
KEGG enrichment analysis of genes can predict the key metabolic pathway volved in genes. In order to understand the key metabolic pathways related to th

Gene Ontology and Kyoto Encyclopedia of Genes and Genomes Enrichment Analysis
In order to clarify the functional categories of differentially expressed genes during the development of O. fragrans fruit, we performed GO enrichment analysis on the differentially expressed gene sets of comparison between two adjacent periods (S1-S2 and S2-S3), and the top 10 GO items significantly enriched were listed. Compared with the first and second periods, DEGs mainly included lipid metabolic process, small molecule metabolic process, chloroplast, plastid, phosphotransferase activity, alcohol group as acceptor, kinase activity, and oxidoreductase activity ( Figure 3A), and DAS events enrichment in small molecule metabolic process, chloroplast, plastid, phosphotransferase activity, alcohol group as acceptor, kinase activity, and oxidoreductase activity ( Figure 3A). Following consideration of the first and second segments, both differentially alternatively spliced genes and differentially alternatively spliced events were enriched in small molecule metabolic process, chloroplast, plastid, phosphotransferase activity, alcohol group as acceptor, kinase activity, and oxidoreductase activity ( Figure 3A). The third period was contrasted with the second period, with a significant concentration of DEGs in the mRNA metabolic process, alpha-amino acid metabolic process, mRNA process cell periphery and oxidoreductase activity ( Figure 3B), and DAS events concentration on mRNA metabolic process, alphaamino acid metabolic process, mRNA process, cell periphery, and oxidoreductase activity ( Figure 3B). DEGs and DAS between the second and third stages were accompanied by the mRNA metabolic process, alpha-amino acid metabolic process, mRNA process, cell periphery, and oxidoreductase activity ( Figure 3B). Distinctions between these three periods collectively enrich the oxidoreductase activity.
KEGG enrichment analysis of genes can predict the key metabolic pathways involved in genes. In order to understand the key metabolic pathways related to the development of O. fragrans fruit, we, respectively, carried out KEGG pathway enrichment analysis on the differential gene sets of three groups of controls. Contrasted to stage one (S1) and stage two (S2), DEGs augmented in porphyrin and chlorophyll metabolism, photosynthesis-antenna proteins, photosynthesis, carotenoid biosynthesis, and alpha-linoleic acid metabolism ( Figure 4A). DAS events were enriched in MAPK signaling pathway-plant, peroxisome, and cyanoamino acid metabolism pathways ( Figure 4A). Observing the second (S2) and third periods (S3), DEGs were enriched in porphyrin and chlorophyll metabolism, photosynthesisantenna proteins, alpha-linoleic acid metabolism, and carotenoid biosynthesis ( Figure 4B); DAS events were enlisted in alpha-linolenic acid metabolism, phagosome, and carotenoid biosynthesis ( Figure 4B). KEGG analysis of the DEG and DAS genes exposed that most of the genes were enriched in porphyrin and chlorophyll metabolism, alpha-linoleic acid metabolism, and carotenoid biosynthesis pathways.  (DAS) events. Dark blue, light blue, and orange represent the first 10 (A) GO analysis of differentially expressed genes in period one and period two; (B) GO analyses of differentially expressed genes in period two and period three. Biological processes (BP), cellular components (CC), and molecular functions (MF) down-regulated and up-regulated proteins in order by p-value.

RT-PCR Validation of Differentially Expressed Alternative Splicing Genes
In order to verify the AS events detected by RNA-seq, we extracted RNA samples from O. fragrans fruit at three different stages and designed primers for four genes that produce AS events for RT-PCR analysis (Table S6). For example, ofr.gene4802 has two corresponding transcripts, in which the expression level of the transcript gene4802-mRNA-1 is increasing in three periods, which is consistent with the transcript data we have measured ( Figure 5C). Comparing the expression levels of the two transcripts shows that the transcript ofr.gene4802-mRNA-1 exerts a dominating feature in the development of O. fragrans fruit ( Figure 5C). The expression level of the two corresponding transcripts of ofr.gene55649 was increased in all three periods ( Figure 5D), and the expression level obtained from the experimental results is highly consistent with the data we measured. In comparison, the transcript BGI_novel_T008614 exerts a dominant role ( Figure 5D). gene4802-mRNA-1 is increasing in three periods, which is consistent with the transcript data we have measured ( Figure 5C). Comparing the expression levels of the two transcripts shows that the transcript ofr.gene4802-mRNA-1 exerts a dominating feature in the development of O. fragrans fruit ( Figure 5C). The expression level of the two corresponding transcripts of ofr.gene55649 was increased in all three periods ( Figure 5D), and the expression level obtained from the experimental results is highly consistent with the data we measured. In comparison, the transcript BGI_novel_T008614 exerts a dominant role ( Figure 5D).

Up-Regulation of Alpha-Linolenic Acid Metabolism in O. fragrans Fruit Development
Alpha-linolenic acid content is continuously accumulated in plants, which is the substrate of fatty acid oxidation. It participates in the synthesis of volatile compounds and is an indispensable substance in the human body but can only be obtained from plants [17,18]. The more mature the fruit, the higher the content of alpha-linolenic acid, which has also been verified in tomatoes and Plukenetia volubilis [18,19]. The alpha-linolenic acid pathway was significantly enriched in our enrichment outcome (Figure 4). A total of 37 genes were enriched in the alpha-linolenic acid synthesis pathway in the transcriptome data of O. fragrans fruit development, which is composed of seven gene families (Table S5). Lipoxygenase (LOX), hydroperoxide dehydratase (AOS), 12-oxophytodienoic acid reductase (OPR), and other gene families have a trend of increasing with fruit development.
Lipoxygenase (LOX) gene is the key enzyme gene in the first step of the transformation of α-linolenic acid into other substances [20]. In O. fragrans fruit, the LOX gene family showed an overall up-regulation trend, consisting of 11 genes (29.73%), among

Up-Regulation of Alpha-Linolenic Acid Metabolism in O. fragrans Fruit Development
Alpha-linolenic acid content is continuously accumulated in plants, which is the substrate of fatty acid oxidation. It participates in the synthesis of volatile compounds and is an indispensable substance in the human body but can only be obtained from plants [17,18]. The more mature the fruit, the higher the content of alpha-linolenic acid, which has also been verified in tomatoes and Plukenetia volubilis [18,19]. The alpha-linolenic acid pathway was significantly enriched in our enrichment outcome (Figure 4). A total of 37 genes were enriched in the alpha-linolenic acid synthesis pathway in the transcriptome data of O. fragrans fruit development, which is composed of seven gene families (Table S5). Lipoxygenase (LOX), hydroperoxide dehydratase (AOS), 12-oxophytodienoic acid reductase (OPR), and other gene families have a trend of increasing with fruit development.
Lipoxygenase (LOX) gene is the key enzyme gene in the first step of the transformation of α-linolenic acid into other substances [20]. In O. fragrans fruit, the LOX gene family showed an overall up-regulation trend, consisting of 11 genes (29.73%), among which the transcript BGI_novel_T021035 (Table 1) corresponding to ofr.gene 24,865 genes could be up to 10 times higher in the third-period stage (S3) compared to the first-period stage (S1) ( Figure 5C). Addition of LOX expression can accelerate the softening process of fruit after ripening [21]. During the ripening process of O. fragrans fruit, the fruit pericarp becomes soft continuously ( Figure 1A). Therefore, we hypothesized that the expression of the LOX gene was significantly up-regulated, which may result in the fruit pericarp keeping softening. The overall gene expression level of the alpha-linolenic acid metabolism pathway increased with fruit ripening ( Figure 5C), which was consistent with the change in the alpha-linolenic acid metabolism pathway in peaches [22]. The increase in LOX gene expression promoted the increase in downstream reaction substrate content, and alpha-linolenic acid accumulated continuously in O. fragrans fruit, which accelerated the softening of the fruit pericarp.

Carotenoid Metabolism Affects the Color Formation of O. fragrans Fruit
The carotenoid metabolism pathway is an important pathway that affects fruit color. During the ripening of Mangifera indica and Vitis vinifera, the carotenoid content gradually decreases and the change in carotenoid content is closely related to the change in fruit color [23,24]. The overall carotenoid content was decreased continuously from 0.032 mg/g in the first period to 0.021 mg/g in the third period, which was consistent with the overall carotenoid expression changes in transcription data ( Figure 6B), where a total of 78 genes, including 12 gene families, were enriched in the carotenoid pathway (Table S7). The differential regulation of gene expression levels was more obvious in the middle and downstream genes of the carotenoid metabolism pathway. In Mangifera indica, the difference in α-carotene and β-carotene content was responsible for the different colors of the flesh [25]. Our study found that the gene expression level difference is more obvious after the production of α-carotene and β-carotene in the carotenoid biosynthesis pathway of fruit at different stages ( Figure 6B); the expression of genes of α-carotene and β-carotene are highest in the first period of fruit and then gradually decrease, which may be the reason for the gradual purple color of the fruit.   The gene expressions of zeta-carotene isomerase (Z-ISO), lycopene epsilon-cyclase (LCYE), lycopene beta-cyclase (LCYB), zeaxanthin epoxidase (ZEP), and 9-cis-epoxycarotenoid dioxygenase (NCED) in the carotenoid metabolism pathway also showed a downward trend, in which low expression impedes carotenoid synthesis ( Figure 6A, Table 2). As a key enzyme, Z-ISO determines the conversion of carotenoid synthesis, thus regulating fruit coloring [26]. The transcriptional regulation of Z-ISO expression has changed during tomato evolution, which may result in differences in fruit color [27,28]. In our study, the transcript expression level of Z-ISO in the second phase decreased by 1/7 compared with that in the first phase, and the expression level in the third phase was roughly the same as that in the second phase ( Figure 6B). LCYE and LCYB may be the limiting step in carotenoid accumulation. In Citrus sinensis, the low carotenoid content was consistent with the low expression of LCYE in the green stage. LCYB controls the accumulation of carotenoids in the middle and lower reaches of Citrus reticulata [29,30] and has also been verified in Cucurbita moschata, M. indica, and V. vinifera [28,31,32]. In our discussion, the expression levels of LCYE and LCYB were consistent, the expression levels decreased to 1/6 or 1/10 from the first stage to the second stage, and the expression levels of the second and third stages were consistent, with lower transcription levels. The carotenoid content decreased continuously from the first stage to the third stage ( Figure 6B), and the low carotenoid content coincided with the low expression of LCYE and LCYB; that low expression of LCYE and LCYB inhibited carotenoid accumulation in O. fragrans. ZEP is a key action point in the carotenoid pathway and an increase in ZEP gene expression during fruit ripening inhibits carotenoid accumulation, which has been demonstrated in Vaccinium myrtillus and Prunus armeniaca [33][34][35][36][37]. In O. fragrans, the expression level of ZEP can be reduced to 1/10, and the reduction is most obvious in the second stage of O. fragrans fruit development, which indicates that reduced carotenoid content due to low expression of Z-ISO, LCYE, LCYB, and ZEP genes may be an essential reason for the O. fragrans color transitions.

Sample Selection and Preparation
The fruit of the O. fragrans "Zi Yingui" was collected about 150 days after flowering in spring, and the samples were collected according to the color of the fruit skin, with the first period being full green, the second period being half green and half purple, and the third period being full purple. Three O. fragrans "Zi Yingui" trees of equal lengths and healthy growth were selected for collection on the campus of Nanjing Forestry University in close proximity to each other, and 10 g were taken from each of the three trees in each period as three replicates. After picking, we peeled off the skin and put it into liquid nitrogen to keep it fresh and took pictures of the samples under the body view mirror.
The determination method of carotenoid content in the pericarp of "Ziyingui" was to take 0.5 g of pericarp powder ground with liquid nitrogen into a 10 mL centrifuge tube, add 6 mL of 95% ethanol solution precooled at 4 • C, and extract it in darkness at 4 • C for 24 h, shaking many times during the period. The centrifuge was set at 4 degrees, 5000 rpm, centrifuged for 5 min, and the absorbance values of 1 mL of supernatant solution were measured at the wavelengths of 665, 649, and 470 nm, respectively. The formula for calculating carotenoid content is as follows [38]:

Transcriptome Sequencing and Differential Alternative Splicing Analysis
Total RNA was treated by mRNA enrichment method; that obtained RNA was fragmented by interruption buffer, reverse transcription was conducted by random N6 primer, and cDNA double strand was synthesized to form double-stranded DNA. The resulting double-stranded DNA was blunt-ended and phosphorylated at the 5 end, forming a sticky end with an "A" protruding from the 3 end, and lighted with a bubble-like linker with a "T" protruding from the 3 end. The ligation product was amplified by PCR with specific primers. The PCR product was thermally denatured into a single strand, and then that single-strand DNA was circularized by a section of bridge prim to obtain a singlestrand circular DNA library, and the single-strand circular DNA library was sequenced on DNBSEQ.
Clean reads were obtained by filtering out reads with low quality, adapter contamination, and high content of unknown base N. Clean reads were then aligned to the reference genome for new transcript prediction, SNP, InDel, and differential splicing gene detection. The novel transcript with protein code potential was in addition to that reference gene sequence to form a complete reference sequence after the novel transcript was obtained, and the alternative splicing condition of the sample was detected by using rMATs [39]. DESeq2 R package was used to screen a differentially expressed gene and differentially expressed alternative splicing transcripts and to perform in-depth cluster analysis and functional enrichment analysis [40]. GO functional classification and KEGG enrichment analysis of alternatively spliced transcripts were performed using the pHYPER package [41] in R software. The p-value was FDR-corrected and, in general, a function with a Qvalue ≤ 0.05 is considered significantly enriched. The ORF of unigene was detected using getorf [42], then aligned to the transcription factor protein domain with hmmsearch [43] (data from TF), and then the unigene was characterized according to the transcription factor family characteristics described by PlantTFDB [44]. According to the existing splicing factor (SF) family information of A. thaliana, we compared the homologous genes of O. fragrans to identify the splicing factor family of O. fragrans.

RT-PCR Validation of AS Events
The total RNA of fruits was extracted with RNAprep Pure Plant Plus Kit (DP441, TIANGEN, Beijing, China) according to the manufacturer's instructions. Then, 5 µg RNA was reversed transcribed by Evo M-MLV RT Premix for RT-PCR (AG11728, Accurate Biology, Changsha, China) for cDNA synthesis. The reaction steps of RT-PCR were as follows: 94 • C for 30 s, followed by 29 cycles of 98 • C for 10 s and 55 • C for the 30 s and 72 • C for 1 min, and 72 • C for 5 min. Three replicates were created for each sample, with OfACT as the internal reference gene, primers were designed using the prime3 website (https://primer3.ut.ee/ (accessed on 13 April 2023)), and all primers used in this experiment were listed in Table S6. RT-PCR products were identified by 1.5% agarose gel electrophoresis.

Conclusions
The fruit color of O. fragrans gradually changed from green to purple during the development process and the fruit texture gradually softened. Based on this, transcription and alternative splicing analysis were conducted for the three developmental stages of O. fragrans. According to the type of alternative splicing events, the content of skipped exon type was the highest and most of the alternative splicing events occurred in the first two periods. Transcription factors were primarily concentrated in the bHLH family, AP2-EREBP family, MYB family, and MADS family. Splicing factors mainly center on glycine-rich proteins, 17S U2 snRNP, related to spliceosome, 35S U5 associated proteins, and SR proteins.
This study found that alpha-linolenic acid metabolism and carotenoid metabolism pathways may be important pathways during fruit development of O. fragrans, among which the up-regulation of LOX, AOS, and OPR may be the reason for the accumulation of alpha-linolenic acid. The down-regulation of Z-ISO, LCYE, LCYB, and ZEP may be the reason for the decrease in carotenoid content, resulting in the color of fruit changing from green to purple gradually. This study analyzed the fruit development process of O. fragrans from the perspective of alternative splicing, which provided new evidence for further study on the fruit development of O. fragrans.