We cloned the coding sequences (CDS) of BpMYB123 whose gene identifier is BPChr03G03353 in recently published B. platyphylla genome (Chen et al., 2021b). Primers with adaptors containing specific restriction sites were designed for further cloning. Using cDNA of B. platyphylla as template, PCR products were cloned into PMD19-T vector for cDNA sequence validation. The CDS of BpMYB123 was inserted into the binary vector pROK2 by using BamHI and KpnI double-enzyme digestion. Using the similar method, the CDS of BpMYB123 (XbaI and SalI) was cloned into binary vector of pBI121-GFP. The antisense fragments of BpMYB123 were amplified by PCR, and then cloned into the binary vector pROK2-RNAi. The primer sequences used are shown in Supplementary Table 1.

The three binary vectors, pROK2-BpMYB123, pBI121-BpMYB123-GFP and pROK2-RNAi-BpMYB123, were transformed into Agrobacterium strain EHA105 by freeze-thaw method (Weigel and Glazebrook, 2006). The B. platyphylla transgenic lines were generated by the leaf disc method (Lv et al., 2020a). To identify true transgenic lines, DNA was extracted from transgenic plants using a DNA extraction kit (TIANGEN, Beijing, China) and was used for PCR. RNA was extracted from transgenic lines using the CTAB method (Chang et al., 1993) and used for qRT-PCR to calculate the expression levels of BpMYB123 in transgenic lines. The primer sequences for PCR and qPCR are listed in Supplementary Table 1.

Three BpMYB123 overexpression lines were used for RNA-seq experiments. The RNA-Seq data was submitted to NCBI SRA database, and the BioProject ID was PRJNA776425. The clean reads were mapped to the B. platyphylla genome by using HISAT2 (Kim et al., 2019). The threshold of p-adjusted (padj) value <0.05 and fold change (FC) ≥2 were employed to identify the differentially expressed genes (DEGs). DEGs were annotated by eggNOG-mapper v2 (Cantalapiedra et al., 2021) and that were further analyzed by Gene Ontology (GO) enrichment analysis (Chen et al., 2020) to identify biological processes that were over-represented in DEGs.

The TF-centered Y1H was used to identify the binding cis-elements of BpMYB123 following the previously described method (Ji et al., 2014). Positive colonies were chosen and sequenced to identify candidate cis-element sequences.

To do Y1H, the CDS of BpMYB123 was first cloned into the pGADT7 vector (Clontech), while the promoter of BpLEA14 was ligated into the pAbAi vector (Clontech). The construct, pAbAi-BpLEA14, was transformed into Y1HGold competent cells, and then cultured on SD/-Ura and SD/-Ura + AbA medium. Finally, pGADT7-BpMYB123 plasmids were transformed into pAbAi-BpLEA14 yeast competent cells, and cultured on SD/-Leu and SD/-Leu + AbA medium. The Y1H analysis was conducted using the Matchmaker Gold Yeast One-Hybrid Library Screening System Kit (Clontech). The primer sequences are listed in Supplementary Table 1.

pBI121-BpMYB123-GFP transgenic plants were subjected to ChIP experiments following a method as described previously (Li et al., 2014) using anti-GFP antibodies. The precipitated DNAs were used as the templates for ChIP-PCR. There were six MYB1AT cis-elements in the promoter of BpLEA14, we designed the primers that span at the six sites. The primer sequences used are shown in Supplementary Table 1.

Birch seedlings were planted in soil in the greenhouse (16/8 h light/dark, 24°C) for two months. These seedlings were then used for drought treatment, the plants were placed in the greenhouse without watering for ten days. After that, the plants were irrigated for recovering. The phenotypes were the photographed.

To measure the physiological changes, the plants under drought treatment for seven days were harvested for measuring the activities of SOD and POD, the contents of H₂O₂, superoxide radicals (O₂^–), MDA and electrolyte leakage (EL). MDA, SOD and POD were measured following the methods as described previously (Wang et al., 2010), while EL was measured as described in our early publication (Lv et al., 2020b). H₂O₂ and O₂^– were measured using the kit (Hydrogen peroxide content kit, H2O2-2-Y, Superoxide anion kit, SA-2-G, Suzhou Comin Biotechnology).

B. platyphylla seedlings were cultured in bottles pre-filled with 1/2 MS + 0.02 mg/L NAA + 2% (w/v) sucrose media, and then placed in a tissue culture room pre-set to 16/8-h light/dark cycles and an average temperature of 25°C for two months. They were then subjected to drought stress, and the leaves of these stressed plants were harvested for nitroblue tetrazolium (NBT) and 3, 3′-diaminobenzidine (DAB) stainings, as described previously (Zhang et al., 2011).

In order to localize BpMYB123 proteins in the cells, the pBI121-BpMYB123-GFP plasmids were transferred into Agrobacterium strain EHA105, and then the agrobacteria were delivered into Nicotiana benthamiana epidermal cells using the injection method as described earlier (Pečenková et al., 2017). After the infection, the materials were placed at room temperature for 48 h. Fluorescence was observed and photographed under the confocal laser scanning microscope (LSM 800, Zeiss, Germany).

The student’s t-test was employed to test the statistically significant differences of various measures between BpMYB123 transgenic lines and wild-type. The threshold for statistically significant differences was set to p-value <0.05.

Based on Sanger sequencing, CDS of BpMYB123 is 915 base pairs (bp), which is 81 bp longer than 834bp of BPChr03G03353 from Betula platyphylla reference genome (Supplementary File 1). The subcellular localization experiment showed that BpMYB123 transcription factor was localized in the nucleus (Figure 1).

Nine BpMYB123 overexpression (BpMYB123-OE) transgenic lines, and eight BpMYB123 repression (BpMYB123-RE) transgenic lines were obtained using leaf disc transformation method (Supplementary Figure 1). These transgenic lines and wild-type plants had been growing in the greenhouse for two months. After the drought treatment (without watering for 10-day) and a three-day watering recovery. Several leaves of BpMYB123-OE transgenic lines near the root shriveled and wilted, and half of them survived. In the wild-type, one of the third of the leaves survived, while all the leaves in the BpMYB123-RE transgenic lines withered (Figure 2A). We conclude that BpMYB123 improved drought tolerance of B. platyphylla. We harvested the leaves after drought treatment for about a week and measured the contents of proline, MDA and EL (Figures 2B–D). The content of proline in BpMYB123-OE was significantly higher than wild-type (Figure 2B), whereas BpMYB123-RE transgenic lines were lower than wild-type. Plant cells subjected to stress can lead to membrane lipid peroxidation, damage of the structure of biofilm and changes in the permeability of cell plasma membrane. MDA is one of the final products of membrane lipid peroxidation. The integrity of the plasma membranes of the cells can be determined by the results of EL. The results of MDA content and EL in transgenic plants and wild-type cells before and after stress were shown in Figures 2C,D. Before drought stress, the content of MDA and EL in cells of all transgenic lines were largely the same without a significant difference. After drought stress treatment, the content of MDA and EL in cells increased, indicating that the cell membranes were oxidized. MDA was produced and the permeability of cell membrane was enhanced under drought stress. While the contents of MDA and EL in BpMYB123-OE transgenic lines were significantly decreased, the contents of MDA and EL in BpMYB123-RE transgenic lines were significantly increased in comparison with wild-type.

Biochemical and physiological traits were measured to determine if the gain- and loss-of-function in BpMYB123 transgenic lines under drought treatment. The DAB and NBT staining were used to measure H₂O₂ and O₂^– levels; and thus represented the activities of POD and SOD. This is because POD and SOD are the enzymes responsive for removing H₂O₂ and O₂^–. After drought treatment, all three BpMYB123-OE lines showed significantly reduced H₂O₂ and O₂^– levels compared to wild-type (Figures 3A,B). POD and SOD enzyme activities in the BpMYB123-OE transgenic lines were significantly higher than those in the wild-type (Figures 3C,D). Correspondingly, the contents of H₂O₂ and O₂^– were significantly lower than those in the wild-type (Figures 3E,F). Thus, overexpression transgenic plants could eliminate reactive oxygen species (ROS) and diminish the toxicity of ROS.

Using a TF-centered Y1H method, we obtained multiple yeast colonies containing pHIS2 element library plasmids with candidate sequence motifs, which could potentially be bound by BpMYB123 protein. We used the sequencing results to search the PLACE database¹ and identify cis-elements in the inserted sequences. Some motifs were identified. There are six MYB1AT motifs in the promoter of BpLEA14. Because MYB1AT (WAACCA) is an element closely related to abiotic stress, and we considered it was the candidate motif (Table 1).

1480 DEGs were identified in BpMYB123-OE lines as shown in Supplementary Table 2. Gene ontology (GO) enrichment analysis was performed using the DEGs form a pairwise comparison of BpMYB123-OE vs wild-type. The result showed a large number of DEGs responded to abiotic stress (Supplementary Table 3). There were 132 genes belonged to biological process of response to hormones (GO:0009725); 61 out of these 132 genes belong to biological process of respond to abscisic acid (GO:0009737), and many DEGs were responsive to metabolic process, such as cellular glucan metabolic process (GO:0006073); glycogen metabolic process (GO:0005977); energy reserve metabolic process (GO:0006112). According to the adjusted p-value, the top 20 GO terms were shown (Supplementary Figure 2).

BPChr06G29277 encodes a late embryogenesis abundant protein (LEA) and its protein sequence is highly homologous to AtLEA14 in A. thaliana; it was thus named BpLEA14. BpLEA14 was significantly up-regulated (Log2 based fold change = 4.35, adjusted p-value = 6.04E-51) in BpMYB123-OE lines. We believe that BpLEA14 plays an important role in enhancement of drought tolerance in BpMYB123 transgenic plants.

To confirm that BpMYB123 could bind to BpLEA14, we performed Y1H (Figure 4B). The result confirmed the binding of BpMYB123 to the promoter of BpLEA14.

ChIP-PCR was used to identify putative target gene of BpMYB123. There were six MYB1AT (WAACCA) cis-elements in the promoter of BpLEA14 (Figure 4A). To verify that BpMYB123 could bind to BpLEA14 by MYB1AT (WAACCA) motifs, we performed ChIP-PCR. After ChIP was done, the regions spanning the six MYB1AT motifs in BpLEA14 promoter were amplified using PCR and the results were shown in Figure 4C. These results substantiated the binding of BpMYB123 to three MYB1AT (WAACCA) motifs in the promoter of BpLEA14.

The CDS of BpLEA14 (456bp) was cloned from B. platyphylla, 1611bp shorter than the reference genomic region of BPChr06G29277 (2067bp). The sequence was shown in the Supplementary File 1. The CDS of BpLEA14 was inserted into the binary vector pROK2. Seven transgenic lines were obtained by the leaf disc method, and three lines with the highest gene expression were selected for subsequent experiments (Supplementary Figure 3). Two-month-old B. platyphylla transgenic lines were grown in a greenhouse at 25°C under 16 h light/8 h dark period. Before the drought experiment was performed, all plants were fully irrigated. After 12 days without water, the plants were subjected to dehydration. The photos were taken three days later after the rehydration we initiated (Figure 5A). The content of MDA and EL in BpLEA14 overexpression (BpLEA14-OE) transgenic lines were significantly decreased compared with wild-type (Figures 5B,C). That indicates BpLEA14 improves drought tolerancein B. platyphylla.