For transcriptome analyses, one-year-old T. grandis seedlings were transplanted to pots with 300 g of a soil mixture and grown in a shady plastic greenhouse. After two weeks, the surviving seedlings were irrigated for 12 h, and then irrigation was withdrawn to start drought treatment.

For drought resistance tests, plants were grown in a potting soil mixture in growth chambers at 22°C with 16 h/day illumination. The relative humidity was approximately 70% ( ± 5%). After growing for one week, the plants were irrigated with 1 L of water per tray for 6 h, and then drought treatment was imposed by withdrawing irrigation for half of the plants until most died. The other half were grown under a standard irrigation regime as a control. After 20 days of drought treatment, representative pots were placed to take photos.

For DAB and NBT staining, Arabidopsis seeds were surface sterilized and cold treated at 4°C for three days. Then, the seeds were plated on an MS medium containing 3% (w/v) sucrose and 0.8% (w/v) agar and grown at 22°C with a 16-h daily light period. The 10-day-old seedlings were transferred to the MS (Murashige and Skoog) medium with 10% PEG (with controls) for three days before staining. For DAB staining, the samples were immersed in a DAB solution (SL1805, Coolaber) for 4 h and then in 95% ethanol for decoloring. For NBT staining, the samples were first immersed in NBT staining solution (S19048, Yuanye) until a dark blue color appeared (approximately 1 h) and then in 95% ethanol for decoloring. Photographs were taken using a stereo microscope (SZX16, Olympus).

Phylogenetic analysis generated a bootstrap neighbor-joining evolutionary tree using MEGA 7.0 with 1,000 bootstrap replicates. Amino acid sequence alignment was conducted using DNAMAN software.

To obtain TgSQS overexpression in transgenic plants, the coding sequence of TgSQS was PCR-amplified with the primers listed in Table S1 and introduced into the Super1300 vector driven by a Super promoter (multiple CaMV35S are connected in series). The construct was transformed into Agrobacterium strain GV3101 for transfection of Arabidopsis using the floral-dip method (Devarenne et al, 2002). Seeds were screened on an MS agar medium containing 50 mg/L of hygromycin. The selected T3-generation transgenic plants with 100% resistance to hygromycin were considered homozygous lines and harvested for further analysis.

The coding sequence of TgSQS was cloned into a modified pCAMBIA1300, which had GFP fused to the N-terminal, and introduced into Agrobacterium tumefaciens strain GV3101. The transient gene expression analysis of TgSQS and an endoplasmic reticulum marker in Nicotiana benthamiana was performed according to a previous study (Nakashima et al., 1995). Agrobacteria were grown overnight in LB (Luria-Bertani) media and brought to an OD600 of 0.8 in the injection solution. After three days of injection, green fluorescent protein (GFP) and red fluorescent protein (RFP) fluorescence was observed via confocal laser scanning microscopy (LSM510, Carl Zeiss).

The coding sequence of TgSQS was introduced into the pET-32a vector (provided by Zuying Zhang) and transformed into competent Rosetta (DE3) to produce a recombinant protein containing a Trx-His-tag at the N-terminus. The cloning sites were BamHI and SacI. The transformed bacteria were grown to an OD600 of 0.6 at 37°C in an LB medium, and then 0.4 mM isopropyl β-D-1-thiogalactopyranoside (IPTG) was added to induce TgSQS expression. After culturing for 20 h at 16°C, the cells were collected and lysed by sonication. The crude enzyme solution was subjected to Trx-His-Tag purification using a Ni Sepharose 6 Fast Flow gravity column to obtain purified protein. Imidazole was removed from the purified protein using centrifugal ultrafiltration (Amicon Ultra 30 kDa, Millipore, Burlington, MA, USA). After BCA protein quantification, 10-μl samples were used to run a 10% SDS-PAGE gel.

The enzyme reaction of the recombinant protein TgSQS was conducted as follows: 50 µg of purified protein was incubated at 30°C with 20 µg of FPP triammonium salt (Sigma–Aldrich, St. Louis, MO, USA), 50 mM Tris–HCl (pH7.5), 25 mM MgCl₂, 1 mM DTT, 2% glycine, and 3 mM NADPH in a total reaction volume of 500 µl. After 8 h, the reaction product was extracted three times with 500 µl of n-hexane, concentrated to 100 µl, and analyzed using GC–MS. GC–MS detection was performed using Trace GC Ultra-ISQ (Thermo Scientific) with a DB-5MS column (30 m × 0.25 mm × 0.25 μm; Agilent). The GC temperature program was as follows: 80°C, raised to 300°C at a rate of 15°C/min, and held at this temperature for 18 min. The carrier gas was ultrahigh-purity helium at a flow rate of 1.0 ml/min. The GC interface temperature was 290°C, and the sample injection volume used was 1 μl.

Dry A. thaliana seeds were accurately weighed to 0.039 g and ground with liquid nitrogen. Ultrasonically extracted samples were extracted for 20 min three times, and the solvent was removed under low temperature and pressure and dissolved in 1 ml of n-hexane. A total of 1.5 ml of a 2 mol/L KOH-ethanol solution was added to 0.4 ml of the extraction solution. Samples were saponified ultrasonically for 10 min in a 60°C water bath and placed in a 60°C oven for 60 min after vortexing. They were shaken for 1 min, and 1 ml of water and 1 ml of n-hexane were added after cooling. Samples were extracted for 2 min, and then the supernatants were removed. We added 0.25 g of anhydrous sodium sulfate to purify the supernatants, dried the supernatants with liquid nitrogen, and then dissolved the precipitates in 0.1 ml of n-hexane to start GC–MS analysis. The GC–MS detection was performed using a Trace GC Ultra-ISQ (Thermo Scientific) with a DB-5MS column (30 m × 0.25 mm × 0.25 μm: Agilent). The GC temperature program was as follows: 80°C, raised to 300°C at a rate of 15°C/min, and held at this temperature for 18 min. The carrier gas was ultrahigh-purity helium at a flow rate of 1.0 ml/min. The GC interface temperature was 290° C, and the sample injection volume used was 1 μl.

Total RNA was isolated from the leaves of T. grandis seedlings treated with drought stress using the RNAprep Pure Plant Kit (DP441, Tiangen); mRNA was purified from total RNA using poly-T oligo-attached magnetic beads. First-strand cDNA was synthesized using a random hexamer primer and M-MuLV reverse transcriptase (RNase H-). The library fragments were purified with the AMPure XP system (Beckman Coulter, Beverly, USA) to select cDNA fragments that were preferentially 370–420 bp in length. Library quality was assessed using a Qubit2.0 Fluorometer, Agilent Bioanalyzer 2100 system, and qRT-PCR. The clustering of the index-coded samples was performed on a cBot Cluster Generation System using the TruSeq PE Cluster Kit v3-cBot-HS (Illumia) according to the manufacturer’s instructions. After cluster generation, the library preparations were sequenced on an Illumina Novaseq platform, and 150-bp paired-end reads were generated.

Raw reads of the fastq format were first processed using in-house perl scripts. At the same time, the Q20, Q30, and GC content of the clean data were calculated. Differential expression analysis of two conditions/groups (two biological replicates per condition) was performed using the DESeq2 R package (1.20.0). The obtained P-values were adjusted using the Benjamini–Hochberg method, which is designed to control the false discovery rate. Genes with an adjusted P-value <0.05 found by DESeq2 were assigned as differentially expressed. Gene Ontology (GO) enrichment analysis and KEGG pathways of differentially expressed genes were implemented using the cluster Profiler R package, in which gene length bias was corrected.

qRT-PCR was performed with a C1000 Touch™ Thermal Cycler system (Bio-Rad) and the ChamQ SYBR qPCR Master Mix kit (Vazyme). The relative expression level was calculated according to the 2^−△△Ct method. The actin gene was used as a reference gene. Ct represents the PCR cycle number at which the amount of target reaches a fixed threshold. The corresponding primers are listed in Table S2.

We accurately weighed 0.039 g of dry A. thaliana seeds and 0.5 g of fresh leaves (from T. grandis seedlings treated with drought stress) and ground them in liquid nitrogen. Samples were extracted with ultrasound for 20 min three times, and the solvent was removed under low temperature and pressure and dissolved in 1 ml of n-hexane. A volume of 1.5 ml of a 2 mol/L KOH-ethanol solution was added to a 0.4 ml extraction solution. Samples were ultrasonically saponified for 10 min in a 60°C water bath and placed in a 60°C oven for 60 min after vortexing. They were shaken for 1 min, and 1 ml of water and 1 ml of n-hexane were added after cooling. Samples were extracted for 2 min, and the supernatants were removed. We added 0.25 g of anhydrous sodium sulfate to purify the supernatants, dried the supernatants with liquid nitrogen, and dissolved the precipitates with 0.1 ml of n-hexane for silylation to form trimethylsilyl ether derivatives of sterols, which were used for GC–MS.

Samples (1 µl) were injected into a GCMS-QP201PLUS apparatus (Shimadzu), which consisted of an automated sampler injection system, a split/splitless injector, and a DB-5MS column (30 m × 0.25 mm × 0.25 μm: Agilent). The GC temperature program was as follows: 80°C, raised to 290°C at a rate of 20°C/min, and held at this temperature for 18 min. The carrier gas was ultrahigh-purity helium at a flow rate of 1.0 ml/min and a 1:10 split injection ratio. Authentic campesterol, stigmasterol, β-sitosterol, and cycloartenol were purchased from Sigma-Aldrich. For quantification, a standard curve for individual authentic standards was generated.

The coding sequence of TgWRKYs was obtained from RNA sequencing and cloned into the pCAMBIA 1300-GFP vector to express the effector. The 1870-bp promoter fragments of TgSQS were sub-cloned into the pGreenII 0800-LUC vector as a reporter. The corresponding primers are listed in Table S1 . Individual effector vectors and recombinant reporter vectors were transferred to A. tumefaciens GV3101. The reporter was mixed with each kind of effector or the empty pCAMBIA 1300-GFP vector (control) at a 1:1 (v:v) ratio and then injected into tobacco leaves as described previously (Nakashima et al., 1995). A dual-luciferase assay was carried out using the Dual-Lumi™ Luciferase Reporter Gene Assay Kit (Beyotime, RG089). The firefly luciferase (FLUC) and Renilla luciferase (RLUC) values were measured using a Promega GloMax20/20 Luminescence Detector. The relative LUC activity was calculated as the ratio between FLUC and RLUC activities, and three biological replicates were applied.

Yeast one-hybrid (Y1H) assays were carried out with the Matchmaker Gold Yeast One-Hybrid System Kit (Clontech) according to the manufacturer’s protocol (PT4087-1, Clontech, USA). The corresponding primers are listed in Table S1 . Based on the distribution of predicted W-box binding sites in the TgSQS promoter region, three short fragments of the TgSQS promoter (the adenine residue of the translational start codon ATG) were assigned position +1, and the numbers flanking the sequences of the TgSQS promoter fragments were counted based on this number. The positions of three 200-bp sequences containing a W-box were as follows: W1, 1775 to 1575 bp; W2, 1275 to 1075 bp; and W3, 375–175 bp). They were subcloned into the pAbAi vector to obtain pAbAi-W1/W2/W3. Figure S2 provides the sequence of the TgSQS promoter. The coding sequence of TgWRKY3/6/7 was inserted into the pGADT7 vector to construct the prey-AD vector. Then, the linearized pAbAi-W1/W2/W3 vector was transformed into Y1HGold. After determining the minimal inhibitory concentration of Aureobasidin A (AbA) for positive transformants, the prey-AD vector was transformed into the bait yeast strain. Successfully transformed yeast strains were grown on the corresponding SD medium (SD/−Leu, SD/−Leu + AbA) for 3–5 days to take photos.

Since the genome of T. grandis has not yet been sequenced, we searched the transcriptome data (NCBI accession Nos. SRX9417001–SRX9417009) and found a unigene annotated as squalene synthase. This sequence was used as a query in the blastx of NCBI, and the results showed that this unigene had high identity (>80%) with squalene synthase from Taxus cuspidate, Ginkgo biloba L., and Pinus massoniana, indicating that this unigene likely encodes squalene synthase in T. grandis; thus, it was named TgSQS. The 1230-bp coding length of TgSQS was determined using sequence alignment in blastx. Phylogenetic tree analysis of TgSQS with SQSs from angiosperms, gymnosperms, algae, and animals showed that TgSQS was most closely related to gymnosperms, such as T. cuspidate, and furthest from animals ( Figure 1A ). Multiple alignments of TgSQS with squalene synthase from A. thaliana, T. cuspidate, and P. massoniana showed that there were six highly conserved regions, I–VI, in its amino acid sequence, among which II and IV contained aspartic acid (DXXXD) active sites ( Figure 1B ), which were reported to bind to isopentene phosphate groups (Rushton et al., 2010), and VI was reported to perform the function of membrane targeting and anchoring (Ishihama et al., 2011). The DeepTMHMM-based predictions of transmembrane domains revealed that TgSQS had one C-terminal transmembrane domain (amino acid residues 389–407), which was highly like those predicted for A. thaliana ( Figure S1B, C ) and Taraxacum koksaghyz SQS (Valitova et al, 2016). We further investigated the subcellular localization of TgSQS through transient expression in N. benthamiana leaves. As can be seen in Figure S1A , the TgSQS protein fused with GFP at the N-terminal colocalized with the endoplasmic reticulum marker, which was consistent with the literature (Valitova et al., 2016). These results suggest that TgSQS encodes functional squalene synthase.

To understand the catalytic function of TgSQS, the recombinant expression plasmid pET-32a-TgSQS was constructed and transformed into E. coli Rosetta (DE3) to produce a recombinant protein containing a Trx-His-tag at the N-terminus. The pET-32a empty vector served as a control. After induced expression, extraction, and purification, the purified recombinant proteins were analyzed using SDS-PAGE. TgSQS was successfully extracted from the supernatant of lysates with a molecular weight of around 70 kDa ( Figure 2A ). The enzyme activity of TgSQS was assayed using the purified protein with the cofactors NADPH and Mg²⁺, while Trx-His-tag was used as a control. The catalytic products were analyzed using GC–MS; the peak at 13.02 min in the profile of TgSQS reaction products ( Figure 2B ) corresponded to that observed in the standard squalene sample ( Figure 2C ), but no such peak was detected in the control ( Figure 2D ). Moreover, the characteristics of peaks in the mass spectrum of TgSQS ( Figure 2E ) were consistent with those of standard squalene ( Figure 2F ). These results proved that TgSQS has the catalytic activity of squalene synthase in vitro.

To further verify the function of TgSQS, the overexpression plasmid super1300-TgSQS was transferred into A. thaliana, and three constitutive overexpression lines with significantly higher expression levels of TgSQS were obtained ( Figure 3A ). The content of squalene in mature seeds was detected, and the results showed that the content of squalene in the three overexpression lines was significantly higher than that of the wild type ( Figure 3B ), indicating that TgSQS also has the catalytic activity of squalene synthase in vivo.

In plants, the dominant sterols comprise β-sitosterol, campesterol, and stigmasterol, while β-sitosterol is the most abundant. It has been evidenced in many in vitro and in vivo studies that β-sitosterol possesses various biological actions, such as immunomodulatory, antimicrobial, anticancer, anti-inflammatory, lipid-lowering, antioxidant, and anti-diabetic activities (Babu and Jayaraman, 2020). Therefore, increasing the β-sitosterol content in T. grandis seeds is very important to improve their nutritional value. As it is difficult to achieve genetic transformation in T. grandis, to verify the role of TgSQS in the sterol biosynthetic pathway, we used the super promoter to express TgSQS in A. thaliana ( Figure 3A ) and tested the β-sitosterol content in mature seeds, as can be seen from Figure 3C , which indicates that the β-sitosterol content in the mature seeds of three TgSQS-overexpressing lines increased by 15%, 15%, and 30% compared with the wild type, indicating that constitutively increasing TgSQS expression in Arabidopsis significantly increased β-sitosterol content in mature seeds. At the same time, the campesterol and stigmasterol content were also detected in these materials ( Figure 3D ); the results showed that the campesterol content in three TgSQS-overexpressing lines were also significantly higher than that of the wild type. However, the stigmasterol content showed no difference between the three TgSQS-overexpressing lines and the wild type.

Due to the important role of TgSQS in sterol biosynthesis, we studied its transcriptional regulation mechanism. Through cloning and analysis of its promoter (1,860 bp before ATG), three W-box cis-acting elements were found in the entire promoter region. Since the W-box element was the DNA binding site of WRKY family transcription factors, WRKY TFs were regarded as candidates to regulate TgSQS expression. In total, 23 TgWRKYs with full-length coding sequences were found by searching the transcriptome data of T. grandis (NCBI accession Nos. SRX9417001–SRX9417009). All of them were constructed in the overexpression vector pCAMBIA1300-GFP, and the dual luciferase assay was carried out with the empty vector as the control. TgWRKY3, TgWRKY6, and TgWRKY7 significantly increased the expression level of the reporter gene ( Figures 4B, C ), indicating that they may regulate the expression of TgSQS.

To verify whether TgWRKY3, TgWRKY6, and TgWRKY7 could directly bind to the promoter of TgSQS, a yeast one-hybrid experiment was carried out. In the Y1H assay, the pAbAi vector carrying W1, W2, and W3 (200-bp sequences containing W-box from the TgSQS promoter, Figures 4D , S2 ) and the pGADT7-TgWRKY3/6/7 recombinant vector served as the reporter and the effector, respectively. Only Y1HGold yeast co-transformed with pGADT7-TgWRKY3 and pAbAi-W1 grew normally on the SD/−Leu medium supplied with 200 ng/ml AbA, whereas the yeast carrying the pGADT7-TgWRKY6/7 and pGADT7 empty vector did not ( Figure 4E ), indicating that TgWRKY3 directly bound the W-box-containing region in the TgSQS promoter.

Preliminary experimental results showed that T. grandis seedlings had strong drought tolerance. Therefore, we carried out drought treatment on T. grandis seedlings and sampled the leaves at the D40 stage (drought treatment for 40 days when soil moisture was zero but seedlings stayed green) and D60 stage (drought treatment for 60 days when seedlings became wilted) to perform transcriptome sequencing with normally watered seedlings (CK, sampled at the D60 stage) as a control. PCA showed low variability among biological repeats, which indicated that there was a high correlation between repetitions ( Figure 5A ). The nine libraries produced over 6G of clean bases, with Q30 percentages (percentage of sequences with sequencing error rates <0.1%) ranging from 94% to 95% ( Table S3 ). All bases were assembled into 66,798 unigenes with a mean length of 1,241 bp ( Table S4 ). The GO, NT, NR, SwissProt, PFAM, KO, and KOG databases were used to annotate the predicted protein sequence, and 33,318 unigenes were annotated by at least one database ( Table S5 ). These data showed that RNA-seq was of high quality and could be used for further analysis.

A total of 2,013 DEGs, including 1,505 up and 508 downregulated genes, were identified in the D40 stage compared with CK ( Figure 5B ). In the D60 stage, compared with CK, 9,156 genes were upregulated and 8,880 genes were downregulated ( Figure 5C ), indicating that there were many more DEGs in the latter. Moreover, the DEGs of D40vsCK and D60vsCK did not completely overlap. As shown in Figure S6 , a total of 1,336 DEGs overlapped, of which 461 were jointly upregulated and 266 were jointly downregulated. Brassinosteroid biosynthesis and terpenoid backbone biosynthesis were found in the KEGG path enrichment scatter plot at the D60 stage compared with CK ( Figures S4, S5 ), indicating that genes related to terpenoid and sterol biosynthesis were involved in the response of T. grandis to drought stress, considering that campesterol is a precursor of brassinosteroids. Consequently, we combined the sterol biosynthesis pathway with the T. grandis transcriptome data in this study to draw a simplified map and mark the expression changes of related genes in CK, D40, and D60 stages in the form of a heat map ( Figure 5D ). The expression levels of HMGS, HMGR, MK, DXS, IPPI, FPPS, SQS, and DWF1 increased significantly at the D60 stage ( Table S6 ). At the same time, the sterol content in the leaves of the CK, D40, and D60 stages was detected, and the results showed that the content of sitosterol, campesterol, and stigmasterol in the D60 stage was significantly higher than in the CK, while there was no significant difference between the D40 stage and the CK ( Figure 6 ). This result was consistent with the fact that the differential genes of the D40 stage in the transcriptome were far less than those in the D60 stage, which might be because the soil moisture at the D40 stage had just dropped to zero and the T. grandis seedlings were still in the early stages of drought stress. Together, these results indicate that the sterol biosynthesis pathway is involved in the response of T. grandis to drought stress.

To verify the function of TgSQS under drought stress, three homozygous overexpressing A. thaliana lines and the wild type were used for drought treatment. After 20 days of drought, the wild-type leaves began to wilt, while the leaves of TgSQS-overexpressing lines stretched normally, indicating that they were more drought-tolerant than the wild type ( Figure 7C ). To deeply understand the drought tolerance mechanism, the accumulation of H₂O₂ and superoxide in the overexpressing lines and the wild type after simulated drought stress was observed by DAB and NBT staining. In this study, 10-day-old seedlings were treated with 10% PEG for 72 h, followed by 3,3-diaminobenzidine (DAB) staining to determine the presence of H₂O₂ and nitro blue tetrazolium (NBT) staining to show the presence of the superoxide anion. As shown in Figures 7A, B , under control conditions, wild-type and transgenic plants showed similar basal levels of H₂O₂ and superoxide, but under 10% PEG treatment, TgSQS-OE leaves and roots showed less H₂O₂ and superoxide accumulation than wild-type leaves and roots. These results suggest that TgSQS improves drought tolerance by reducing ROS accumulation.