Identification of AaSHI members in the A. annua genome
To obtain sequence information of AaSHI, nine AtSHI family proteins of A. thaliana were used to blast with four sets of haplotype chromosome genomes of A. annua (Supplementary Table 1). Multiple sequence alignments and protein integrity alignment analysis found that unctg_3838g01590241 (LQ-9_phase0) was half the length of the unctg_3207g01555681 and lacked the zinc finger domains, while chr5g00243221 (LQ-9_phase1) also lacked the zinc finger domains (Supplementary Figue 1). As a result, five AaSHI genes in the HAN1_phase0 genome were selected for subsequent analyses, named AaSHI1 (chr1g04025621), AaSHI2 (chr3g03131551), AaSHI3 (chr5g03332101), AaSHI4 (chr6g00096641), AaSHI5 (chr6g00640341), respectively. Then we analyzed the physical and chemical properties of these identified genes. These genes encoded 276–390 amino acids with the molecular weight (MW) ranged from 68.19 kDa to 95.67 kDa, and their isoelectric points ranged from 5.05 to 5.13.
Table 1
Detailed information for five AaSHI members in the A. annua genome.
NAME | Gene ID | ORF length (bp) | Protein length (aa) | Mw (kDa) | pI |
AaSHI1 | chr1g04025621 | 1,005 | 334 | 83.54 | 5.08 |
AaSHI2 | chr3g03131551 | 1,041 | 346 | 85.71 | 5.07 |
AaSHI3 | chr5g03332101 | 972 | 323 | 79.64 | 5.11 |
AaSHI4 | chr6g00096641 | 1,173 | 390 | 95.67 | 5.05 |
AaSHI5 | chr6g00640341 | 831 | 276 | 68.19 | 5.13 |
Phylogenetic tree construction, multiple sequence alignment and gene structure analysis
To further investigate the evolutionary relationships and their evolutionary conservation among individual members of the AaSHI family, a phylogenetic tree of 42 SHI members in A. annua and 5 other species was constructed using MEGAX software (Fig. 2A). Phylogenetic tree showed that these amino acid sequences formed three branches, and the members of AaSHI were distributed in two branches, and were most closely related to the dicotyledonous plant A. thaliana, followed by V. vinifera, while they were further related to Z. mays, O. sativa. Amino acid sequence alignment and conserved domain analysis indicated that all five AaSHI proteins contained the circular zinc finger domains and IGGH domains (Fig. 2B). Meanwhile, MEME online software was used to predict conservative structural domains (Fig. 2C, Supplementary Table 3), and 5 motifs were identified. Among them, motif4 and motif5 were distributed in five AaSHI proteins, which just corresponded to the circular zinc finger domain and IGGH domain. Gene structure analysis shows that they all have two exons and one intron, proving that the five AaSHI sequences share the same structure.
Chromosome localization and synteny analysis
Based on the genome information of A. annua, the chromosome location analysis of AaSHI genes showed that AaSHI1, AaSHI2 and AaSHI3 were distributed on chr1, chr3 and chr5 respectively, while AaSHI4 and AaSHI5 were distributed on chromosome 6, and there were no tandem duplication events among members of A. annua SHI gene family during the evolution (Fig. 3A). Synteny analysis detected that two AaSHI genes, AaSHI3 and AaSHI4, participated in a segmental duplication event. Large-scale comparative synteny maps of AaSHI and AtSHI, SlSHI, VvSHI genes showed that A. annua and V. vinifera had the highest synteny with 8 pairs of genes, but only 6 pairs of synteny genes with S. lycopersicum (Fig. 3B, Supplementary Table 4).
Analysis of cis-acting elements of AaSHI genes promoter
To analyze the potential biological functions of the AaSHI gene, cis-acting element analysis of the promoter region located 3.0 kb upstream of the start codon of the AaSHIs was performed via the PlantCARE online website (Supplementary Table 5). As shown in Fig. 4A, the promoter regions of AaSHIs mainly contained four major types of cis-acting elements: Plant growth and development, phytohormone responsive, light responsive, abiotic and biotic stress. MYB binding sites were the most common in the AaSHI genes promoter, followed by MYC binding elements were also more distributed except for AaSHI5, which indicated that AaSHI may interact with MYB and MYC TFs to regulate the growth of A. annua. Interestingly, the ABRE element and W-box on the AaSHI5 promoter were significantly more than the other four AaSHI genes, suggesting that its function may differ from the others. In addition, some jasmonic acid response elements could also be found in the all AaSHI genes promoter. To intuitively illustrate the distribution of cis-acting elements, we counted the number of elements of each type and presented them in a bar graph (Fig. 4B). These data indicated that the AaSHI genes potentially have diverse roles in the phytohormone and environmental response of A. annua.
The expression profile of AaSHI gene family by RNA-seq database
Using different tissues, light and dark treatment and phyllotaxy RNA-seq database to analyze the gene expression profile [21]. Evaluate the expression level of AaSHI gene family with artemisinin biosynthetic genes and GST-specific TFs by TPM, visual mapping was performed after normalization using the TBtools software. The results showed that AaSHI1, AaSHI2, AaSHI3 and AaSHI4 were all specifically expressed in the trichome, which was consistent with the expression pattern of artemisinin, while AaSHI5 was specifically expressed in the stem (Fig. 5A). The results of gene expression in distinct leaf positions indicated that except for AaCYP71AV1, other genes showed significant leaf order, that is, they were highly expressed in tender leaves and showed a decreasing trend with the order of leaf arrangement (Fig. 5B). Meanwhile, AaSHI genes with artemisinin biosynthetic genes and GST-specific TFs were all induced by light, AaSHI5 was not detected due to its low transcript level (Fig. 5C).
The expression pattern of AaSHIs
The expressions of AaSHI genes were evaluated by qRT-PCR. As shown in Fig. 6A-6D, AaSHI1-4 were highly expressed in the young leaves but weakly expressed in root, which is in line with the artemisinin biosynthetic genes. AaSHI5 showed the highest expression in the stem (Fig. 6E). In addition, the expression levels of artemisinin biosynthetic pathway genes are highest in the tender leaves and gradually decreases as leaf age. Similar expression pattern was observed in AaSHI1, AaSHI2, AaSHI4, and AaSHI5 (Fig. 6F-6J). Considering that AaSHI5 was preferentially expressed in stem rather than young leaves, we speculated that the AaSHI1, AaSHI2, and AaSHI4 are related to the biosynthesis of artemisinin.
Correlation analysis with artemisinin biosynthetic genes
Next, we conducted co-expression analysis of AaSHIs, artemisinin biosynthetic genes, and two key regulators of artemisinin biosynthetic pathway AaGSW1 and AaORA, based on their tissue/organ and light-treated transcriptome data. The Pearson coefficient was used to calculate the correlation between the AaSHI TFs and artemisinin biosynthetic genes. Pairs of genes that meet the criteria will be screened and then visualized using Cytoscape software. TPM of AaSHIs in GST was indicated by a color gradient. Correlation analysis showed that three genes (AaSHI1, AaSHI2 and AaSHI4) had a positive correlation of all four artemisinin biosynthetic genes as well as GST-specific TFs. However, AaSHI3 and AaSHI5 were not correlated with these key genes (Fig. 7, Supplementary Table 6). Therefore, we chose AaSHI1, AaSHI2 and AaSHI4 for further functional study.
Subcellular localization of the AaSHI proteins
The subcellular localization prediction was conducted on the Plant mPLoc website, and the all three TFs were predicted to be localized to the nucleus. Subsequently, the C-ternimal of AaSHIs were fused with YFP and expressed in N. benthamiana leaves to experimentally verify the subcellular localization. As shown in Fig. 8, pHB-AaSHI1/2/4-YFP were detected in the nucleus exclusively, while the control YFP (Yellow Fluorescent Protein) displayed in both nucleus and cytoplasm. This is consistent with website predictions and their potential functional localization as a TF.
AaSHIs transactivated the expression of artemisinin biosynthetic genes
To verify the activation effect of AaSHIs on four artemisinin biosynthetic genes, we used tobacco leaves for transient transformation. The reporter vector was obtained by inserting the successfully cloned promoter into the pGreenⅡ 0800-LUC vector with homologous recombination method and transferred into A. tumefaciens GV3101 (pSoup) for dual-LUC assay. Meanwhile, AaSHIs were inserted in pHB vector driven by the 35S promoter (Fig. 9A). The results indicated that AaSHI1 has significant activation effect on the promoters of AaADS and AaCYP71AV1, while AaSHI2 could solely activate the expression of AaADS gene (Fig. 9B-9E). In contrast, AaSHI4 had no activation effect on these artemisinin pathway genes. In order to explore the binding sites of AaSHI members on four specific artemisinin biosynthetic genes, we analyzed the SHI-binding sites in the 3,000 bp upstream promoter regions of these four genes. It was reported that the binding sites for SHI proteins were ACTCTAC, ACTCCAT, ACTCAAC and ACTCTAA [15, 17, 18, 20], so we concluded that the possible binding site for SHI TFs was ACTCnAn. The analysis results showed that there were six potential binding sites on the promoters of AaADS and three on AaCYP71AV1 (Supplementary Fig. 2). Yeast one-hybrid assays were performed to further examine the binding ability. The ORF of AaSHI1 and AaSHI2 were inserted into the pB42AD effector vector. Each predicted binding site motif along with the four nucleotide sequences on both sides was artificially synthesized into three repeat fragment and inserted into the pLacZ reporter vector. As shown in Fig. 9F and 9G, AaSHI1 could directly bind to AaADS and AaCYP71AV1 promotors while AaSHI2 could only bind to AaADS promotor. These data suggested AaSHI1 and AaSHI2 served as direct positive regulators of artemisinin biosynthetic genes
AaSHI1 promote the expression of artemisinin biosynthetic genes
Considering that AaSHI1 had better transcriptional activation effects on artemisinin biosynthetic genes as compared with AaSHI2, transient transformation was conducted to verify the function of AaSHI1 in regulating artemisinin biosynthesis in vivo. pHB-AaSHI1-YFP constructs in Agrobacterium strain cell GV3101 was injected into the back of the first pair of true leaves of A. annua (Fig. 10A). After 48 hours of cultivation, samples were collected and qRT-PCR was used to detect the expression level of artemisinin biosynthetic genes. The results showed that AaSHI1 significantly promoted AaADS expression in A. annua, about 8.6-fold of that in the empty control (Fig. 10B). Meanwhile, AaSHI1 also significantly activated the expression of AaCYP71AV1, with about 2.5-fold that of the control. These observations were consistent with previous dual-LUC assays. In general, these results indicated that AaSHI1 had positive roles in regulating artemisinin biosynthesis.