The entire genome sequence and annotation files of Cannabis sativa (GCA_900626175.1) were downloaded from the National Center for Biotechnology Information (NCBI) database. The CRBRx female strain of Cannabis sativa was used [32]. Arabidopsis thaliana (GCA_000001735.4),Malus pumila(GCF_002114115.1) were downloaded from NCBI database. Arabidopsis thaliana TPS gene data were obtained from the TAIR database (https://www.arabidopsis.org) （Supplementary Table 1）.The transcriptomics data were retrieved from NCBI (PRJNA498707) [33].

The known TPS protein sequence from Arabidopsis thaliana was initially used as a query sequence to identify members of the Cannabis sativa TPS gene family, and basic local alignment search tool in the TBtools [34]software package was used for sequence alignment. The E-value was set to 1e-5 to recover sequences of candidate Cannabis sativa TPS genes, which were then compared again in the NCBI database using the protein explosion function and were submitted to the CD Search and Pfam databases for further screening of the conserved domain. To isolate the Cannabis sativa TPS gene family members, sequences that were incomplete and superfluous were eliminated from the domain. ExPASy software (http://web. expand. The org/protparam/) tool was employed to forecast the physicochemical characteristics of the Cannabis sativa TPS protein. Research into the subcellular positioning of CsTPS proteins was conducted via the WoLFPSORT website (www.genscript.com/wolf-psort.html) [33].

The amino acid sequence of Cannabis sativa TPS was mapped to the Cannabis sativa chromosomes by analyzing the Cannabis sativa genome annotation file, and chromosome location information for each TPS gene was determined. A chart depicting the respective physical positions of chromosomes was created, assigning names to genes based on their chromosomal locations. MCscanX [35] was used to calculate the repeat events of the Cannabis sativa TPS genes. In addition, to visualize the collinearity between the Cannabis sativa and Arabidopsis thaliana genomes, we used the Dual Systematic Plot function in the TBtools [34] software package and highlighted the collinearity relationship of the TPS genes.

Utilizing the STRING website available online (https://string-db.org/cgi/Input. pl) forecasts the interplay among Cannabis sativa TPS proteins. For reference purposes, the standard settings of the Arabidopsis thaliana protein library are utilized. Data storage is in TSV format, with visualization achieved through the Cytoscape 3.8.0 software (www.cytoscape. org).

Phylogenetic analysis and classification of the TPS gene family and analysis of TPS gene structure and motif.

The MEME Suite web (https://meme-suite.org/mem e/tools/meme) [36] server was used to analyze the TPS motifs. Settings were adjusted to ZOOPS for the distribution of sites and 10 for counting motifs. Utilizing the 2000 base pair sequence preceding each TPS Cannabis sativa gene as the promoter area, this sequence was extracted and uploaded to the PlantCare (http://www.plantcare.co.uk/) site for a statistical examination of its cis component's makeup [37]. TBtools software was used to analyze and visualize the composition of cis-expressed elements in the conserved domains, exons, introns, motifs, and promoter regions of TPS genes. The neighbor-joining method was applied using the MEGA 7 software (www.megasoftware.net) package to build a phylogenetic tree of TPS proteins in Cannabis sativa and other reported species [38]. Alignment of the amino acid sequences was conducted through the FUSTALW algorithm, and the bootstrap was set to 1000 times. All other parameters were set to their standard settings.

Purchase of Plant Total RNA Kit (510105) from Hangzhou Xinjing Biochemical Reagent Development Co., Ltd. (Hangzhou, China). Reverse Transcription Kit (article number QP057) was purchased from Guangzhou Yijin Biotechnology Co., Ltd. (Guangzhou China). The RNA extraction process adhered to the methods and steps by Hangzhou Xinjing Biochemical Reagent Development Co., Ltd. Post RNA extraction, the sample was preserved at a temperature of −80 °C. In the analysis, an appropriate quantity of RNA was obtained by melting the reverse transcription reagent while it was maintained on ice. The procedure for the reverse transcription program proceeded in this manner: 25 °C for 5 min, 42 °C for 15 min, 85 °C for 5 min, and 4 °C for the retention phase. An ultramicro ultraviolet spectrophotometer was used to gauge the levels of cDNA products, which were then preserved in the fridge at −20 °C for subsequent application. The above operation steps are the same as the operation methods of other personnel in this research group [39].

Analysis of Cannabis sativa TPS gene expression profile and fluorescence quantitative real-time polymerase chain reaction (qRT- PCR)

Extracting TPS fragments per kilobase of transcript per million mapped reads (FPKM) value from RNA sequencing (RNA seq) data of Cannabis sativa [33]. The CsTPS gene expression was visually analyzed using TBtools software, and all FPKM values were processed using row scaling（Supplementary Table 2）. Transcriptome data were verified by via qRT-PCR. The cDNA was used as a template for qRT-PCR experiments using the Hieff UNICON® Universal Blue qPCR SYBR Green Master Mix kit (Shanghai, China). For qRT-PCR, the AriaMx Real-Time PCR System was employed. Program: 95 °C for 2 min, followed by 40 cycles of 95 °C for 10 s, 60 °C for 30 s [40]. Each experiment was replicated thrice. The creation of qPCR primers utilized Premier software (version 5.0), with their synthesis carried out by Beijing Ruibo Xingke Biotechnology Co., Ltd.

To identify TPS family members in Cannabis sativa genomes, we performed a whole genome scan using the Blastp method and Hidden Markov Model and then checked the conserved domain using CD-Search and Pfam databases(PF02365), ultimately identifying 41 CsTPS genes in Cannabis sativa. The amino acids encoded by CsTPS range from 180 to 891, with the molecular weight ranging from 20705.90 to 102324.64 and the isoelectric point ranging from 4.96 to 7.03. The majority of genes were found in the cytoplasm and chloroplasts, while the rest were situated in the peroxisome, nucleus, plasma membrane, and mitochondria (Table 1). Prediction of subcellular localization revealed 19 proteins in the chloroplasts, 2 in the nucleus, however, CsTPS-8 was also identified in the peroxisomes, CsTPS4, CsTPS5, CsTPS6, CsTPS15, CsTPS19, CsTPS24, CsTPS25, CsTPS26, CsTPS28, CsTPS29, CsTPS32, CsTPS34, CsTPS36, CsTPS37 and CsTPS41 in the cytoskeleton, CsTPS-30 in the plastid, and CsTPS-38 in mitochondria（Supplementary Table 3）. The results of this study may lay the groundwork for additional investigations into TPS genes within Cannabis sativa.

Cannabis sativa's genomic annotation data facilitated the examination of CsTPS gene locations on chromosomes. 41 CsTPS genes were scattered randomly across eight chromosomes and a scaffold fragment of Cannabis sativa (Fig. 1A). Predominantly, the CsTPS gene was found at the extremities of each chromosome, exhibiting reduced presence in the central region. Chromosomes 4 and 6 housed the majority of the CsTPS genes. Chromosome 1 showed the lowest distribution, with only one genes. One genes (CsTPS) was distributed on one scaffold fragment. that was not attached to chromosomes (CsTPS41).（Supplementary Table 4）.

Events of gene replication are crucial in creating diversity within gene families and in comprehending how species evolve. For the purpose of aiding the study of Cannabis sativa gene replication occurrences, we removed every scaffold segment from the Cannabis sativa genome, keeping only the genetic data of eight chromosomes. (Fig. 1B). One pairs of segmental gene replication were found on chromosomes NC_044371.1and NC_044379.1, and no gene pairs were found on the remaining chromosomes. The duplication of the CsTPS gene might have been a significant factor in the evolutionary process of Cannabis sativa.

Three-dimensional structure analysis of Cannabis sativa TPS showed that the subfamily structure was similar, but there were significant differences in the structure of genes from different subfamilies, such as different proteins with different structures (Fig. 2).

To elucidate the evolutionary relationship between Cannabis sativa TPS proteins, a phylogenetic tree based on neighbor-joining was constructed using the sequences of TPS proteins from Arabidopsis thaliana and a few other species, the other species are Oryza sativa, Arabidopsis thaliana, Nicotiana tabacum, Vitis vinifera, Citrus limon, Populus trichocarpa, Brassica napus, Brassica rapa, Cucurbita maxima, Chocolope, Ziziphus jujuba, Ricinus communis, Perilla frutescens, Salvia rosmarinus, Thymus caespititius, Origanum vulgare, Lavandula latifolia, Salvia officinalis, Salvia stenophylla, Solanum lycopersicum, Humulus lupulus, Antirrhinum majus, Physcomitrella patens, Abies grandis, Picea abies, Fragaria ananassa, Mentha x piperita, Cichorium intybus, Clarkia breweri, Gossypium arboretum, Solanum tuberosum, and Lycopersicum esculentum. (Fig. 3). All the CsTPS proteins clustered with AtTPS. Consistent with the classification of AtTPSs, CsTPSs were divided into five subfamilies: a (14 members), b (16 members), c (7 members), d (not present), and e/f (4 members). Interestingly, the e/f subgroup classification was relatively close, which has also been observed in other species (Supplementary Table 5).

Generally, plant TPS gene family members contain two conserved domains: the conserved motif in the N-terminal domain is R(R)X8W (R: arginine, W: tryptophan, and X: alternative amino acids), and two highly conserved aspartic acid motifs are present in the C-terminal domain. The DDXXD motif is associated with the coordination of divalent ions and water molecules and the stability of the active sites. The NSE/DTE motif stands as the secondmotif. These motifs are located on the side of the active site entrance and bind the magnesium trinuclear clusters. Most TPSs belong to monoterpene, sesquiterpene, and diterpene synthases, each containing DDXXD and DXDD motifs. We analyzed all protein sequences of the CsTPS gene family. MEME was used to analyze conserved TPS protein motifs, and 10 conserved motifs designated motifs 1–10 were identified. As shown in Fig. 4 [41],the number of conserved motifs in proteins TPS01–TPS41 varied from 3 to 10, with multiple groups having the same motif. For example, TPS17 and TPS18 had the same conservative cardinality with motifs 1–10. We found that motif 1 was the most common motif present in 41 TPS proteins, followed by motifs 7 and 8. These results indicate that motifs 1, 7, and 8 are conserved in the TPS and are crucial for the function of Cannabis sativa TPS domain proteins. Further analysis was conducted on the multi sequence alignment of CsTPS protein （Supplementary Fig. 1）.The N-terminus and C-terminus domains were found in almost all proteins, but the conserved motif of CsTPS2 was missing, possibly due to changes in the arginine tryptophan motif R (R) X8W and DDXXD motifs in TPS-a, or even the absence of a protein. Generally, the arrangement of genes and amino acids aligns with the findings of phylogenetic studies. The role of the CsTPS protein within this category can be deduced from the evolutionary connections of established TPS proteins.

For a deeper insight into Cannabis sativa's TPS gene structure, we observed the arrangement of its coding regions (CDS), noncoding regions (UTR), and introns (Fig. 4). Every CsTPS gene possesses introns varying between one and fourteen, and within the same subfamily, the count of introns is comparable. As an illustration, members of the TPS-b subfamily possess seven introns, in contrast to the TPS-e/f subfamily's eleven introns, and occasionally, there might be unique deviations.

We detected motifs A–L in TPS 17, TPS18, TPS21, TPS22, and so on, while TPS 2 and TPS 15 lacked many conserved motifs, including motifs 1, 6, and 7. However, within the structural domain, significant changes in the positions and quantities of proteins were observed to a certain extent (Fig. 5). According to the results of Motif (Supplementary Fig. 2) and conserved domains, it can be concluded that motifs are conserved within a subfamily, which may be related to their evolutionary conservatism. These results indicate that the CsTPS identified in this study are accurate（Supplementary Table 6）. Findings from extensive phylogenetic analyses, motifs, conserved domains, and gene structures revealed the TPS gene's significant conservation throughout an extended evolutionary journey [40].

Cis-regulatory sequences are noncoding sequences in the gene promoter region that play an significant role in regulating the transcription of related genes. The study extracted the 2000 bp upstream of each CsTPS gene to analyze the promoter region. The promoter region of the CsTPS gene contains a diverse array and quantity of cis-acting elements. The cis-acting elements related to hormone regulation include auxin, gibberellin, and abscisic acid response elements. The abiotic stress encompass cis elements for defense and stress reactions, along with elements responding to low temperatures. Plant growth and development are influenced by factors such as light, salicylic acid, and seed-specific regulatory responsive elements（Supplementary Table 7）. Further analysis revealed that the CsTPS gene has cis-elements that bind to myeloblastosis (MYB) to participate in light and drought responses. One possibility is that CsTPS interacts with CsMYB to form a regulatory network (Fig. 6).

To search for TPS homologous genes in Cannabis sativa and Arabidopsis thaliana, we conducted an association analysis of the entire genomes of Cannabis sativa and Arabidopsis thaliana, Malus pumila Mill and highlighted the TPS gene（Supplementary Table 8）. Collinearity analysis (Fig. 7) showed that Cannabis sativa had a large number of homologous genes with other species, including 3 homologous gene pairs with Arabidopsis thaliana,6with Malus pumila Mil. Collinearity analysis detected CsTPS1, CsTPS4, CsTPS5, and CsTPS6 in the three plants, suggesting that these genes may be highly conserved. According to the predicted results (Fig. 7), Cannabis sativa has homologous genes in each species; however, Cannabis sativa is a dicot with few genes homologous to monocot plants such as Oryza sativa L [33]. In contrast, it shares a greater number of genes with dicotyledonous plants, such as Arabidopsis thaliana, which is consistent with the evolutionary relationships. Homologous genes may have similar functions, which warrants further study in subsequent functional analyses [29].

The STRING database was employed to forecast the traits of the protein protein interaction (PPI) within the Cannabis sativa TPS gene family (Fig. 8). In the diagram, the nodes represent the names of proteins. A node's degree value increases with the number of nodes it connects to. The degree value is indicated by the node shape's size and the color's depth. In summary, our findings reveal that the PPI network comprises 10 nodes and 16 edges. It is worth noting that we noticed that the degree values of these proteins decreased in sequence. The CsTPS37 protein has the highest degree value and is considered to have stronger interactions with other proteins, thus playing an important role in regulating plant growth.

Global changes were also observed across multiple samples and genes. To determine the expression mode of the CsTPS gene, we extracted the FPKM value of the CsTPS gene from RNA-sequencing data of Cannabis sativa flower, bracts, leaves, stems, seeds, and roots, compared their gene expression modes, and constructed a heat map. The results are shown in Fig. 9. Twelve CsTPS genes have tissue-specific expression, which suggests that these genes promote plant growth and development.

Using qRT-PCR, we validated the expression of TPS17, TPS18, TPS20, TPS21, TPS22, TPS26, TPS27, TPS40, TPS41, TPS12, TPS13, and TPS16 in different tissues (Fig. 10). The genes TPS13, TPS16, TPS17, TPS18, TPS26, TPS27, and TPS41 showed high expression levels in bracts, while TPS12, TPS20, and TPS21 genes were predominantly found in roots, and genes TPS40 and TPS22 were notably expressed in leaves. These findings indicate that these genes are specifically expressed in different tissues（Supplementary Table 9）.

The entire genome sequence and annotation files of Cannabis sativa (GCA_900626175.1) and the genome and annotation files of species, such as Arabidopsis thaliana (GCA_000001735.4), were downloaded from the National Center for Biotechnology Information (NCBI) database. The transcriptomic data used in this study were completed by our group and are publicly available and can be found in the NCBI database (PRJNA498707).