Cloning and expression characterisation of LiDXS2 gene in Oriental Lily (Lilium 'Sorbonne')

The 'Sorbonne' of Lilium orientalis used in this study is planted in the Horticultural Engineering Center of Northeast Agricultural University, as shown in Figure 1. A. thaliana wild type (WT) (Col-0) was provided by the Garden Plant Breeding Laboratory of Northeast Agricultural University. Light conditions were as follows: vegetative growth period of 12 hr, reproductive growth period of 14 hr, humidity of 40–60% and temperature of 22°C–24°C. The entire flower development process of Lily 'Sorbonne' was divided into five stages: green bud stage, pink bud stage, half-opening stage, blooming stage and decay stage. The inner wheel petals of each stage were selected for the experiment. Different tissues (stigma, filaments, anthers, petals and leaves) were collected for the qRT-PCR detection. All samples of the three bioreplicates were collected and immediately frozen at -80°C for further analysis.

Figure 1.

L. orientalis’Sorbonne’.

An amount of 100 mg was taken from the same part of the lower petal of the wheel of 'Sorbonne' Lily during the five different stages of development. After all the samples collected were removed, they were wrapped with a tin foil, quickly transferred to the liquid nitrogen for freezing and then placed at –80°C until use. The equipment, working platform, glassware and pipette used were cleaned according to the solid-phase RNase scavenger instructions, and the deionised water was treated. The RNase-free water obtained was used to prepare the electrophoresis solution and 75% ethanol using Axygen RNase-free centrifuge tube and gun tip. RNA extraction was performed according to the instructions of TransZol Plant kit. During the operation, a mask should be worn, gloves should be replaced in time and attention should be paid to the cleanliness of the operating environment to avoid RNase pollution. The concentration and purity of the RNA samples were detected by ultra-micro-amount of ultraviolet optical meter, and the integrity of the RNA extracted by 1.2% agarose gel electrophoresis was detected.

The synthesis of cDNA was conducted according to the instructions of the TranScript^® One-Step gDNA Removal and cDNA Synthesis SuperMix. After the sample was gently mixed, the product was used for PCR: incubation at 42°C for 30 min and for qPCR: incubation at 42°C for 15 min by heating at 85°C for 5 s. The samples were preserved at -20°C in a refrigerator.

The partial cDNA sequence of LiDXS2 was obtained from the transcriptome data of Lilium 'Sorbonne' petals, which was amplified from cDNA using the PrimeSTAR^® HS DNA Polymerase Kit. The pairs of primers (LiDXS2-F: 3′-ATGGCTTTCTCAGGCTCTCTC-5′; LiDXS2-R: 3′-CTAGCTCAGATGCATGGCCT-5′) were used for the amplification. The PCR reaction was performed in a sample of 50 μL volume (28.5 μL dd H₂O, 10 μL 5× PrimeSTAR buffer (Mg²⁺ Plus), 4 μL dNTP mixture (2.5 mM each), 1 μL LiDXS2-F, 1 μL LiDXS2-R, 5 μL cDNA, 0.5 μL PrimeSTAR HS DNA Polymerase) using the following conditions: denaturation at 98°C for 8 min, 30 thermal cycles (98°C/10 s, 56°C/5 s, 72°C/2 min 15 s) and a final extension at 72°C for 10 min. Gene expression analysis was performed by qRT-PCR with three independent biological replicates. The ChamQ Universal SYBR qPCR Master Mix was used to measure the expression levels according to the manufacturer’s instructions. Design qPCR-specific primers using the Primer 5.0 software are shown in Table 1. The Actin gene (GenBank: AB438963) was selected as the internal reference gene and calculated using the 2^-ΔΔCT method.

Using DNAMAN 9.0 to compare Oriental LiDXS2 for amino acid sequences, the phylogenetic analysis of LiDXS2 of Oriental Lily was conducted through MEGA-X, the developmental tree was constructed by using the neighbour-joining method with the Bootstrap method and the number of inspections is 1000. WoLF PSORT, ChloroP 1.1 Server and TargetP-2.0 Server were used for LiDXS2 subcellular localisation prediction.

The coding sequences of LiDXS2 with termination codons removed were subcloned into the pCAMBIA1300-GFP vector by AarI and Eco31I restriction enzymes using T4 DNA ligase. The constructs and empty plasmids were transformed into the leaves of Nicotiana benthamiana, which were then visualised by a laser scanning confocal microscope (Olympus FV3000, Japan).

To gain a preliminary understanding of the function and the location where the LiDXS2 gene plays a role in the formation of lily fragrance, the relative expression of LiDXS2 gene in different organs (stigma, filament, anther, petal and leaf) during the blooming stage of Lily 'Sorbonne' was determined by qRT-PCR (Figure 2), as well as in the green bud stage, pink bud stage, halfopening stage, full blooming stage and the decay stage (Figure 3). RNA was extracted in different organisations, reversed it into cDNA and designed as qPCR-specific primers. The Actin gene (GenBank: AB438963) was selected as the internal reference gene.

Figure 2.

Parts of the flower of L. orientalis’Sorbonne’: (A) petal, (B) leaf, (C) filament, (D) stigma and (E) anther.

Figure 3.

L. orientalis’Sorbonne’ flower development stages: (S1) green bud stage, (S2) pink bud stage, (S3) halfopening stage, (S4) full blooming stage and (S5) decay stage.

Restriction enzymes BamHI and SacI were used to cut pBI121 vector. Primers containing homologous sequences were designed, and PCR amplification was performed using pBZ-LiDXS2 plasmid as template. The reaction conditions were 98°C for 8 min, 98°C for 10 s, 60°C for 5 s, 72°C for 2 min and 15 s, 30 cycles; 72°C for 10 min. The linear vector and the inserted fragment were connected by the ClonExpress^® II One Step Cloning Kit. The linear vector was mixed with the inserted fragment in a molar ratio of 1:2. The reorganisation reaction products were transformed into Escherichia coli, identified by PCR and verified by reorganised vector sequencing. The reorganised vector was named as pBI121-LiDXS2.

The plant expression vector pBI121-LiDXS2 was transformed into Agrobacterium GV3101 by freezethaw method. To verify the function of LiDXS2 gene in model plants, inflorescence dipping method was used to transform Arabidopsis T0 plants by Agrobacterium, and the infected transgenic Arabidopsis T0 seeds were purified, vernalized and sterilised. They were seeded in 0.5 ms medium containing 50 mg ⋅L⁻¹ kanamycin for antibiotic screening. RNA was extracted from the Naman Mustang T3 generation and WT and reverse-transcribed into cDNA for phenotypic analysis.

A total of four flowering-related genes (AtDXR, AtMCT, AtHDR, AtGPPS) were chosen, and RNA was extracted from the LiDXS2 transgenic A. thaliana and Col-0, respectively. The expression of flowering-related genes in transgenic plants and WT plants was detected. The expression of LiDXS2 and the genes related to flower formation were analysed.

The length of the LiDXS2 gene ORF area is 2142 bp, with a total of 713 amino acids (Figure 4). The protein molecular formula is C₃₃₇₄H₅₃₈₁N₉₄₅O₁₀₁₄S₃₂; the total number of atoms is 10746; the protein molecular weight is 76.43 kDa and the theoretical pI is 6.77. The unstable coefficient is 35.69 (<40), which is a stable protein with a fat coefficient of 88.13. The hydrophilicity and hydrophobicity of LiDXS2 protein are predicted. The total average hydrophobic index (GRAVY) is –0.086. Therefore, it is presumed that LiDXS2 protein is hydrophilic. LiDXS2 protein does not have a cross-membrane area, and hence it does not belong to cross-membrane protein. LiDXS2 protein contains an N-terminal DXP synthetic structure (DXP_ synthase_N), a TPP binding area (TPP_DXS) and a pyromylase, pyrimidine binding area (Transket_pyr), and hence LiDXS2 protein belongs to the DXS super family.

Figure 4.

Agarose gel electrophoresis of the polymerase chain reaction product. The amplification product of LiDXS2. (M) DL2000 DNA marker. (1) Negative control, (2) and (3) LiDXS2.

Through the online analysis of NCBI website Blastp function, the amino acid sequence encoded by LiDXS2 gene was compared with Artemisia annua, Bixa orellana, Stevia rebaudiana, Crataegus pinnatifida var. major and Catharanthus roseus. The results showed that the MEGA-X software was used to construct the phylogenetic tree of Oriental Lily LiDXS gene and other plant DXS gene (Figure 5). The similarity of DXS protein between LiDXS2 of Lily 'Sorbonne' and other five plants was more than 77%, and both contained two highly conserved regions of WDVGHQ and IAEQHA (Figure 6). The results of the system development tree show that LiDXS2 is close to the relationship between Ginkgo biloba DXS2 protein and C. roseus DXS protein, which is far-related to Populus trichocarpa DXS2 and A. thaliana DXS3.

Figure 5.

Phylogenetic analysis of LiDXS2 and DXS from other plants: GbDXS1 (Ginkgo biloba DXS1, AAS89341.1); GbDXS2 (Ginkgo biloba DXS2, AAR95699.1); AtDXS1 (Arabidopsis thaliana DXS1, NP_193291.1); AtDXS2 (Arabidopsis thaliana DXS2, NP_850620.2); AtDXS3 (Arabidopsis thaliana DXS3, NP_196699.1); CmDXS (Chrysanthemum morifolium DXS, BAE79547.1); HbDXS1 (Hevea brasiliensi DXS1, AAS94123.1); CrDXS (Catharanthus roseus DXS, CAA09804.2); OsDXS1 (Oryza sativa DXS1, XP_015640505.1); OsDXS2 (Oryza sativa DXS2, XP_015642490.1); PtDXS1 (Populus trichocarpa DXS1, XP_006381844.1); PtDXS2 (Populus trichocarpa DXS2, XP_024460342.1); NtDXS (Nicotiana tabacum DXS, CBA12009.1); SrDXS (Stevia rebaudiana DXS, CAD22155.2) and AaDXS (Artemisia annua DXS, PWA87995.1).

Figure 6.

Amino acid sequence homology comparison alignment of LiDXS2 and DXS from other plants: AaDXS1 (Artemisia annua DXS1, PWA87995.1); BoDXS2a (Bixa orellana DXS2a, AMJ39460.1); SrDXS (Stevia rebaudiana DXS4, ALJ30089.1); CpDXS2 (Crataegus pinnatifida var. major DXS2, ALL29183.1) and CrDXS (Catharanthus roseus DXS, CAA09804.2).

Through the laser co-focusing microscope, the fluorescent signal is observed under the excitation light at 488 nm wavelength. The air carrier of pCAMBIA1300-GFP was used as the control, and the positioning results are shown in Figure 7. The green fluorescent signal of the PCAMBIA1300-LiDXS2-GFP carrier appears only in the chloroplast, and the green fluorescent channel, chloroplast fluorescence channel and open field are opened at the same time. The GFP fluorescence signal overlaps well with the chloroplast spontaneous fluorescence signal. The results of the subcell positioning show that the 'Sorbonne' Lily LiDXS2 gene is positioned in chloroplasts.

Figure 7.

Subcellular localisation of LiDXS2. The fusion proteins PCAMBIA1300-LiDXS2-GFP and PCAMBIA1300-GFP as control proteins were detected using a confocal laser scanning microscope. Bar = 10 μm.

The expression patterns of LiDXS2 gene in different organs of lily during the blooming period of 'Sorbonne' were analysed by real-time fluorescence quantitative PCR. The expression of LiDXS2 gene was tissuespecific in the 'Sorbonne' Lily, the expression level was the highest in petals and the relative expression level was significantly higher than that in other organisations. The expression level was about 30 times higher than that in the filaments, almost no expression was observed in anthers and there were extremely significant differences among different organs (Figure 8). The results of qRT-PCR showed that the LiDXS2 gene was mainly expressed in petals, but the expression level was low in stigma, filaments, anthers and leaves.

Figure 8.

Relative expression levels of LiDXS2 in different flower parts at the first full-blooming day. ***p < 0.001.

The expression level of LiDXS2 gene was also different at different flowering stages, which was significantly lower in the green bud stage than in the other four periods, and the expression volume began to increase significantly (Figure 9). The expression of the LiDXS2 gene in the blooming stage was 1.4 times higher than that of pink bud stage, 1.6 times higher than that of half-opening stage and 4 times higher than that of decay stage. The results showed that the LiDXS2 gene may play an important role in the formation of 'Sorbonne' Lily.

Figure 9.

Relative expression analysis of flower developmental stage of the LiDXS2 gene in the Lilium Oriental hybrids. (S1) Green bud stage, (S2) pink bud stage, (S3) half-opening stage, (S4) full blooming stage and (S5) decay stage. ***p < 0.001.

After receiving infected Arabidopsis seeds, transgenic Arabidopsis T0 seeds were screened on the medium containing kanamycin for two generations, and genomic DNA of wild Arabidopsis and transgenic Arabidopsis T3 were extracted and amplified with LiDXS2-specific primers. A genetically modified strain for good growth, namely, OE1, OE2 and OE3 (Figure 10A), was chosen and compared with WT. To further judge the expression level of LiDXS2 in the genetically modified plant system, RNA from the leaves of transgenic strains was further extracted for qPCR identification. The results of fluorescence quantitative PCR showed that the expression level of LiDXS2 was different in different transgenic strains, and the highest expression level was found in OE2. However, the expression levels of the three plants were different from the expression levels of AtDXS. Therefore, OE1, OE2 and OE3 were all selected for the next experiment (Figure 10B).

Figure 10.

Identification of transgenic A. thaliana: (A) agarose gel electrophoresis of transgenic plants; (M) DL2000 DNA marker; (OE1, OE2, OE3) expression level of LiDXS2 in transgenic plants, (B) expression level of LiDXS2 in transgenic plants.

The T3-generation transfer gene was compared with the growth phenotype and the WT of the same conditions, mainly to observe the difference between the flowering time and the growth tendency. Compared with the WT, transgenic A. thaliana showed increased plant height and more flowers. Therefore, under the same conditions, LiDXS2 will affect the growth and development process of the plant, which will result in different plant phenotypes (Figure 11).

Figure 11.

Phenotype observation of T3-generation positive plants in transgenic Arabidopsis; (OE1, OE2, OE3) expression level of LiDXS2 in transgenic plants, (WT) wild type.

Since DXS is the first rate-limiting enzyme in the MEP pathway, overexpression of LiDXS2 in Arabidopsis must have an impact on the expression of other component genes in the MEP pathway. To explore its influence, the three genetic strains were selected and the expression patterns of key synthase genes downstream of DXS in the MEP pathway were analysed. The results showed that the expression levels of selected four key enzyme genes (AtDXR, AtMCT, AtHDR and AtGPPS) were raised to average, but the differences in upward increases differently, indicating that overexpression of LiDXS2 would affect the expression levels of other genes downstream of MEP. The expression level of AtHDR in the three strains was significantly different, and the highest expression level is about 2.2 times higher than that of WTs. The expression level of AtGPPS in OE1 is also about 2.3 times higher than that of WTs, but the expressions of OE2 and OE3 are not significant. AtDXR is the first key synthetase gene of AtDXS in the middle and downstream MEP pathway. The expression level of AtDXR is about 1.8 times higher than that of the WT, in which OE2 and OE1 show significant differences. The expression trend of AtMCT in transgenic lines was similar to that of AtDXR, both of which showed upregulated expression levels (Figure 12).

Figure 12.

Gene expression levels in transgenic A. thaliana. (A) AtDXR; (B) AtMCT; (C) AtHDR and (D) AtGPPS. The expression levels of MEP pathway-related genes in different transgenic lines were determined by qRT-PCR. A. thaliana Actin (AtActin) was used as the internal reference. (OE1, OE2, OE3) expression level of LiDXS2 in transgenic plants, (WT) wild type. The columns represent average expression values for each line and the error bars show the standard deviation of three biological replicates. The GraphPad statistical analysis was used for testing significant differences in expression levels. *p < 0.05, **p < 0.01.

In previous studies, the content of GPPS in mint young leaves was higher, so we speculated that the reason for the high content of GPPS in OE1 was that the leaves of OE1 were younger at the time of sampling (Yu et al., 2023). In addition, other experiments verified that AKGGPPS3 was induced by abscisic acid (ABA) and its expression was significantly upregulated, while the ABA content of OE2 was higher than that of OE1 and overexpression 3 (OE3). Therefore, the content of exogenous hormone ABA may also be the reason for affecting the expression level of GPPS in different transgenic A. thaliana.

Compared with the WT, A. thaliana lines transformed by LiDXS2 showed early flowering and increased plant height, and the flowering time in A. thaliana was affected by the gibberellin pathway. Therefore, the contents of gibberellins (GA) and ABA in the leaves of transgenic plants were measured. The results showed that the GA content in OE2 of LiDXS2 transgenic strain was significantly increased, while ABA content was slightly decreased (Figure 13). The changes in GA and ABA contents were consistent with the expression of transgenic A. thaliana.

Figure 13.

Determination of ABA and GA contents in LiDXS2 transgenic Arabidopsis and WT Arabidopsis. (A) GA content of transgenic and WT of A. thaliana leaves; (B) ABA content of transgenic and WT of A. thaliana leaves. (OE1, OE2, OE3) expression level of LiDXS2 in transgenic plants, (WT) wild type. “a”: significant difference at p < 0.05.