The leafy stem cuttings of M. xiaojinensis (Mx) was used as the materials, because Mx has a high apomictic rate to ensure the juvenile materials stability (Xu et al., 2017). Semi-lignified leafy cuttings (8-10 cm in length) were excised from basal suckers (juvenile phase, Mx-J) and shoots from the canopy of reproductively mature trees (adult phase, Mx-A). The bases of leafy cuttings were immersed 1~2 cm in depth into a 3.0 g L^-1 indole butyric acid (IBA, Sigma-Aldrich, St. Louis, MO, USA) solution for 1 min (Xu et al., 2017). Cutting dipped in IBA-free water was used as a control. After plugging the cuttings into 50 cell trays containing fine sand, they were incubated in a solar greenhouse. The rooting ability was evaluated at 35 days after treatment. Three biological replicates, each with at least 50 leafy cuttings, were used for the experiment to manage experiment errors.

Tissue-cultured ‘M9T337’ plantlets were sub-cultured in Murashige and Skoog (MS) medium containing 7.5 g L^-1 agar and 30 g L^-1 sugar (pH 5.8) with 0.5 mg L^-1 IBA and 0.2 mg L^-1 6-benzylaminopurine (6-BA) (Cheng et al., 2020). After 30 days, stem cuttings were transferred into 1/2 medium containing rooting 7.5 g L^-1 agar and 30 g L^-1 sugar (pH 5.8) with 0.5 mg L^-1 IBA and 0.1 mg L^-1 1-naphthalene acetic acid (NAA) for rooting. The tobacco (Nicotiana tabacum) plants were sub-cultured in MS medium without hormone. The plantlets were grown under a 16 h light/8 h dark photoperiod with day/night temperatures of 25 ± 1°C and 20 ± 1°C.

Paraffin sections of stem bases were prepared as previously described (Xu et al., 2017; Cheng et al., 2020; Jin et al., 2022), with some modifications. The bases of ‘M9T337’ stem cuttings were collected at 6, 9, and 12 days after transplanting on 1/2 MS medium with 0.5 mg L-1 IBA and 0.1 mg L-1 NAA. The bases of tobacco stem cuttings were excised at 2, 4 and 6 days after subculture on hormone-free MS medium. The samples were fixed in FAA solution (70% ethanol: formaldehyde: acetic acid, 95:5:5 [v/v/v]) for 2 days at room temperature, and store at 4 °C. Then samples were dehydrated with a graded ethanal series (50%, 70%, 85%, 95%, and 100%), infiltrated with xylene, and embedded in paraffin. Cross sections with a 10 μm in thickness were cut with a Leica RM2245 (Leica Microsystems, Wetzlar, Germany) rotary microtome, transferred onto glass slides, deparaffined with xylene, and re-hydrated through an ethanol series, and stained with toluidine blue. Slides were observed using an optical microscope DM2500 (Leica Microsystems, Wetzlar, Germany) and photos were obtained using an attached digital camera DFC420 (Leica Microsystems, Wetzlar, Germany).

Total RNA was extracted from approximately 0.5 g of frozen sample using the TIANGEN Plant RNA Kit (TIANGEN biotech CO., LTD, Beijing, China, DP305). For each sample, 1 μg DNase-treated RNA was used to synthesize first-strand cDNA with oligo d(T) or random primer and HiScript^® II Q RT SuperMix (Cat. R223-01, Vazyme, China). A LightCycler 480 instrument (Roche, Basel, Switzerland) and ChamQ SYBR Color qPCR Master Mix (Vazyme, Nanjing, China) were used for qRT-PCR. The relative expression levels of genes were normalized to the reference gene EF1α and calculated using the 2 ^-ΔΔCt method (Livak and Schmittgen, 2001). All reactions were performed with at least three biological replicates. The primers are listed in Supplemental Table S1 .

The MdWRKY87 promoter fragment (-2000 bp to -0 bp from the MdWRKY87 ATG start codon) was inserted into pCambia1391 vector generating the proMdWRKY87: GUS construct. Tobacco leaves were transformed with the Agrobacterium tumefaciens strain GV3101 cells harboring a proMdWRKY87: GUS or DR5: GUS construct. Agrobacterium cells were re-suspended in buffer with (10 mM MgCl₂, 10 mM MES-KOH, pH 5.6; adding 200 μM acetosyringone immediately prior to use) to an OD₆₀₀ of 0.8~1.0. After injected with 1 ml needleless syringes, the leaves were treated with 50μM IBA and collected 6 hours later. The leaves were submerged in the GUS staining solution for 24 h at 37°C. After staining, tissues were cleared by immersing in 70% ethanol. All primers used are listed in Supplemental Table S1 .

The ORF fragment (stop codon removed) of MdWRKY87 containing SmaI and XbaI sites were inserted into the Super1300-GFP vector to generate the pSuper : MdWRKY87-GFP construct. Subcellular localization was conducted as previously described (Cheng et al., 2021). A. tumefaciens cells (GV3101) expressing pSuper : MdWRKY87-GFP and a Cherry-labelled nuclear marker (NF-YA4-mCherry) was re-suspended using the buffer (10 mM MgCl2, 10 mM MES-KOH, pH 5.6; 200 μM acetosyringone). pSuper : NF-YA4-mCherry was used as a nuclear marker. The tobacco leaves were injected with the re-suspended A. tumefaciens cells using a 1 ml needleless syringe. Three days after infiltration, fluorescence signals of the infiltrated leaves were detected using a laser scanning confocal microscope (Leica TCS SP5 II, Wetzlar, Germany). The primers used for construction are listed in Supplemental Table S1 .

The coding fragment of MdWRKY87 were fused to the GAL4-BD in pBD-GAL4 vector. The transcriptional activation analysis was conducted as previously described (Cheng et al., 2021). The transactivation activity was verified by the growth of yeast AH109 harboring full-length of MdWRKY87 on SD/-Trp and SD/-Trp-His plates and was confirmed by a X-α-Gal staining assay. All primers used are listed in Supplemental Table S1 .

The MdWRKY87 ORF sequence without stop codon was cloned into the pBD-VP16 vector (Han et al., 2016). The reporter vector contained a GAL4-luciferase (LUC) containing five copies of the GAL4-binding element and a minimal CaMV35S promoter at the 5’ end of the LUC gene (Han et al., 2016).The effector vectors or reporter vectors were introduced into A. tumefaciens strain GV3101. The A. tumefaciens cells was re-suspended to an OD₆₀₀ of 1.0 using the buffer (10 mM MgCl₂, 10 mM MES-KOH, pH 5.6; 200 μM acetosyringone). A. tumefaciens cells harboring effector vector and reporter vector were mixed 1:1, then injected into the tobacco (N. benthamiana) leaves by using a 1 mL needleless syringe. After spraying 1 mM luciferin onto the leaves, luciferase imaging was performed using NEWTON 7.0 (VILBER LOURMAT, Paris, France). An assessment of LUC and REN activities was conducted using the Duo-Lite Luciferase Assay System (DD1205–01, Vazyme, Nanjing, China) and BioStack Ready (BioTek Instruments Inc., Winooski, Vermont, USA). LUC/REN ratio was used to calculate the results. The primers used for construction are listed in Supplemental Table S1 .

The MdWRKY87 ORF sequence without stop codon was cloned into pRI101 vector to generate 35S:MdWRKY87-OE construct. The construct was introduced into wild type tobacco (N. tabacum) leaves by A. tumefaciens-mediated transformation as previously described (Zhu et al., 2022). The infected leaves were selected on MS medium containing 100 mg L^-1 kanamycin and 300 mg L^-1 cefotaxime sodium to generate MdWRKY87-overexpressing (MdWRKY87-OE) transgenic lines. Transgenic plants were propagated by subculture on hormone-free MS medium. The primers used for construction are listed in Supplemental Table S1 .

Total RNA was extracted from the stem bases of MdWRKY87-OE and wild type plants using the TIANGEN Plant RNA Kit (TIANGEN biotech CO., LTD, Beijing, China, DP305). A total amount of 3 µg RNA per sample was used in RNA-seq library construction. An Illumina Hiseq (Illumina, CA, USA) system was used for RNA sequencing by Novogene (Novogene, Tianjin, China). A quality assessment was performed on raw data using FastQC. Following Trimmomatic filtering out adapters and unpaired reads, the remaining clean reads were used to calculate the expression of gene by using Kallisto, an RNA-seq quantification program (Bray et al., 2016). The N. tabacum genome was used as the reference genome (Edwards et al., 2017). The count of reads was normalized to Transcripts Per kilobase of exon model per Million mapped reads (TPM). The log2TPM values were subjected to generate the heat map by TBtools software (Chen et al., 2020). The RNA-seq data were deposited in the NCBI Sequence Read Archive (accession number PRJNA917351).

According to our previous research, semi-lignified leafy cuttings from Mx-J and shoots from the canopy of Mx-A were used in this study (Xu et al., 2017; Li et al., 2021). As the previous results (Xu et al., 2017; Li et al., 2021), Mx-J cuttings exhibited a high adventitious rooting ability ( Figures 1A–C ). After IBA treatment, the rooting percentage of Mx-J cuttings (85.14%) was significantly higher than that of Mx-A cuttings (3.57%, Figure 1B ). Neither the cuttings of Mx-A nor Mx-J exhibited the ability of adventitious rooting ( Figure 1B ). Moreover, the adventitious root number per cutting of Mx-J was significantly more than that of Mx-A ( Figure 1C ). We previously identified the expression of the WRKY transcription factor family genes in the cutting stems of Mx-A and Mx-J after IBA treatment (Che et al., 2021). The expression of MdWRKY87 gene was significantly induced in the Mx-J cutting treated with IBA ( Figure 1C ). However, there was no difference in the mRNA levels of MdWRKY87 in the Mx-A cuttings treated with IBA or untreated control ( Figure 1C ).

To further define the relationship between MdWRKY87 expression levels and adventitious rooting formation, we next examined the expression pattern of MdWRKY87 during adventitious rooting formation of tissue culture plantlets in apple. According to the paraffin sections of stem bases of apple plantlets, we found that primordia with dome-shaped adventitious structures were clearly visible at 6~9 days after transplanting ( Figure 1E ) as our previous results (Cheng et al., 2020). After 12 days of transplantation, adventitious roots began to appear. As expected, the gene expression levels of MdWRKY87 increased significantly with the emergence of adventitious root from stem base ( Figure 1F ).

To further validate MdWRKY87 expression in response to IBA treatment, Agrobacterium tumefaciens cells (GV3101) harboring the proMdWRKY87:GUS or DR5:GUS construct (auxin-responsive reporter) were transiently transformed into tobacco leaves. After IBA treatment, the levels of GUS proteins obviously increased in both of leaves transformed with proMdWRKY87:GUS and DR5:GUS construct ( Figure 1G ), suggesting that the promoter activity of MdWRKY87 responds to auxin. These results indicate high MdWRKY87 expression correlates positively with adventitious rooting and may regulate auxin-mediated adventitious root development.

Multiple sequence alignments of WRKY22 homologs from apple indicated that MdWRKY87 harbored a conserved WRKYGQ domain and a C2H2 (C-X5-C-X23-H-X1-H) zinc-finger motif at its C terminus ( Figure 2A ) and belonged to WRKY group IIe (Eulgem et al., 2000). To confirm whether MdWRKY87 functions as a transcription factor, we expressed MdWRKY87 fused to green fluorescence protein in tobacco leaves and observed that the fusion protein localized to the nucleus ( Figure 2B ). A transactivation assay in yeast indicated that the MdWRKY87 protein has transcriptional activation activity ( Figure 2C ). We also performed a dual-luciferase transactivation assay in tobacco leaves. The results showed that the luciferase activity of co-expression of the reporter with pBD- WRKY87 -VP16 was significantly lower than that in the pBD-VP16 control ( Figures 2D, E ). These results suggested that MdWRKY87 is indeed a transcription repressor of WRKY group IIe subfamily.

To characterize the role of MdWRKY87 in adventitious rooting, we generated overexpression lines of MdWRKY87 (MdWRKY87-OE) in tobacco plants by Agrobacterium-mediated transformation ( Supplemental Figure S1 ). We tested the effects of MdWRKY87 overexpression on adventitious root formation in MdWRKY87-OE lines #9, #5 and #6 ( Figures 3A , B ). During the adventitious rooting process, there were obvious morphological differences between MdWRKY87-OE transgenic plants and non-transformed wild type (WT) ( Figure 3A ). The adventitious rooting rate, root number and root length per stem were significantly higher in MdWRKY87-OE lines than that in WT plants ( Figures 3C–E and Supplemental Movie S1 ). These results supported the notion that high MdWRKY87 expression correlates positively with adventitious rooting.

To check whether MdWRKY87 affects the initiation of adventitious root primordia, we conducted the cross sections of the stems of WT and MdWRKY87-OE transgenics lines during adventitious rooting. The initiation of adventitious root primordia was accelerated in MdWRKY87-OE transgenics lines from 2 days after subculture on MS ( Figure 4 ). Moreover, the adventitious root primordium in MdWRKY87-OE transgenics lines were well-developed compared with the WT plantlets at 4 days after subculture, suggesting that high MdWRKY87 expression accelerates the initiation and development of adventitious root primordia.

To determine whether WRKY87 regulates adventitious root formation through modulating auxin polar transport, we examined adventitious rooting capacity in wild-type, MdWRKY87-OE tobacco plants stem cuttings grown on MS medium supplemented with 1-N-naphthylphthalamic acid (NPA). Adventitious rooting was almost absolutely inhibited in both wild-type and MdWRKY87-OE tobacco plants treated with 20µM NPA ( Figure 5A ). To check whether NPA affects the initiation of adventitious root primordia, we conducted the cross sections of the stems of WT and MdWRKY87-OE lines during adventitious rooting. The initiation of adventitious root primordia was both inhibited in wild-type and MdWRKY87-OE transgenic lines under NPA treatment ( Figure 5B ). Dense aerial roots developed on the stems of transgenic and wild-type plants under NPA treatment ( Figure 5C ).

To understand how MdWRKY87 regulates adventitious root development, we conducted a comparative transcriptome analysis of the stem of WT and MdWRKY87-OE tobacco plants. A total 6690 differentially expressed genes (DEGs) in tobacco were identified, including 2000 downregulated and 4690 upregulated genes ( Supplemental Figure S2 ). These DEGs were then subjected to Gene Ontology (GO) functional classification. Within the three GO categories identified, a greatest number of DEGs was significantly enriched in GO categories ‘biological process’ ( Supplemental Figure S3 ). Within the ‘biological process’ GO categories, the top 4 GO terms were “regulation of cellular process”, “RNA biosynthetic process”, “nucleic acid-templated transcription” and “transcription, DNA-templated”, respectively ( Figure 6A ). To further dig out the key genes regulated by MdWRKY87, we analyzed these four GO categories. Venn diagram analysis showed that 272 overlapping DEGs were identified in these four GO categories ( Figure 6B ). Most of these overlapping genes were annotated as transcription factor (TF) genes and transcriptional regulator (TR) genes ( Figure 6C ). Among these overlapping DEGs, there were 242 TF genes, 17 TR genes, and 13 other genes ( Figure 6C ). Within these TF genes, AP-ERFBP, NAC, and HB type TF genes account for a large proportion ( Figure 6C , Supplemental Figure S4 ). Most of these TR genes are of the types Orphans and AUX/IAA ( Figure 6C , Supplemental Figure S4 ).

Further analysis indicated these TF and TR genes mainly enriched in auxin and cytokinin signaling pathway. It is well known auxin and cytokinin appear to play antagonistic roles in the adventitious rooting process. Among these TF and TR genes family, previous study has demonstrated that AUX/IAAs, AINTEGUMENTA (ANT), AINTEGUMENTA LIKE1 (AIL), WUSCHEL-RELATED HOMEOBOX (WOX), and type-B cytokinin Response Regulator (RR) family genes involved in adventitious rooting process (Ramírez-Carvajal et al., 2009; Rigal et al., 2012; Lakehal et al., 2019; Geng et al., 2023). In genes up-regulated by MdWRKY87, WUSCHEL-RELATED HOMEOBOX11 (WOX11), and AINTEGUMENTA LIKE1 (AIL) are positive regulators of adventitious root formation. For genes down-regulated by MdWRKY87, AUX/IAAs and type-B Response Regulator genes are negative regulator of adventitious root formation. Hence, negative regulation of AUX/IAAs and type-B Response Regulator genes and positive regulation of WOX11 and AIL by MdWRKY87 contributes to the enhanced adventitious rooting ability in transgenic tobacco plants. The expression level of these genes was confirmed by quantitative real-time (qRT)-PCR, thus supporting the RNA-seq results ( Figure 7 ).

For rooting recalcitrant woody plants, juvenility is necessary for efficient adventitious rooting. In general, rooting rates in adult plants are usually lower than in juvenile plants (Xu et al., 2017). Recent studies have provided a paradigm for the molecular basis of age-dependent adventitious rooting ability (Sun and Zhu, 2021). According to previous reports, we can understand the molecular basis from at least two independent signaling pathways: (i) via the accumulation of EIN3 protein in adult plants, which directly suppresses expression of WUSCHEL RELATED HOMEOBOX (WOX) genes to inhibit rooting (Ma et al., 2020); (ii) the miR156-SPLs pathway, which modulates root regeneration by crosstalk with auxin signaling pathway (Xu et al., 2016; Xu et al., 2017; Ye et al., 2020a; Ye et al., 2020b; Li et al., 2021). Whether there is other regulation mechanism controlling age-dependent adventitious rooting ability? Here we demonstrated that higher expression level of MdWRKY87 contributing to adventitious rooting-competent in juvenile apple rootstock independent of miR156/SPL pathway. As revealed by qRT-PCR, the expression level of MdWRKY87 was positively correlated with adventitious rooting ability ( Figures 1D , F ). Over-expressing of MdWRKY87 in tobacco leads to enhanced adventitious rooting ability, more adventitious root number and accelerated adventitious rooting process ( Figures 3 , 4 and Supplemental Movie S1 ). In addition, SPL family genes was not found in the transcriptome data of MdWRKY87-OE transgenic plants. It has been widely reported that WRKY transcription factors participate in the regulation of plant growth and development, abiotic stress responses, and disease response. However, the function of WRKY involved in adventitious rooting remains largely unknown. Here, we found WRKY87 transcription factors play essential role in adventitious root formation. In agreement to this result, PuWRKY75 was identified to control the low phosphorus driven adventitious root formation in Populus ussuriensis (Pu) (Wang et al., 2022). In addition, it was interesting that PuWRKY75 act as a transcriptional enhancer, but MdWRKY87 act as a transcriptional inhibitor ( Figures 2C–E ). There will be more WRKY genes, which be identified involving in adventitious rooting process in the future research.

It was well known that auxin play a dominant role in regulation of adventitious root formation (Lakehal et al., 2019). To further identify how MdWRKY87 regulate adventitious root formation, we analyzed the interaction between MdWRKY87 and auxin. MdWRKY87 has significantly enhanced expression levels in response to exogenous IBA treatment ( Figures 1D , F ). In addition, pro MdWRKY87: GUS transiently transformed tobacco leaves exhibited induction of GUS activity after spraying with IBA and the promoter region of MdWRKY87 has auxin response element ( Figure 1G ). These suggest that MdWRKY87 acts downstream of auxin to regulate adventitious root formation. However, the adventitious rooting ability was inhibited in MdWRKY87-OE transgenic tobacco lines upon treatment with the polar auxin transport inhibitor NPA ( Figure 5 ). Taken together, we conclude MdWRKY87 act not only downstream of auxin, but also feedback regulation during adventitious rooting process. Consistent with this, previous data also demonstrate that WRKY71/EXB1 play pivotal roles in shoot branching by regulating auxin pathways (Guo et al., 2015). However, specific mechanism underlying the crosstalk between MdWRKY87 and auxin signaling needs to be further demonstrated.

MdWRKY87 was known to control adventitious root formation ( Figure 3 ), but the regulatory role of MdWRKY87 was unknown. Comparative transcriptome profiling between the wild type and MdWRKY87-OE transgenic plants was conducted in this study ( Figure 6 ). A set of 272 overlapping DEGs were identified through GO enrichment analysis, implying their potential importance for MdWRKY87-dependent adventitious rooting formation ( Figure 6B ). Among these DEGs, we found MdWRKY87 overexpression can activate the expression of adventitious rooting-induced genes, such as WOX11 and AIL, and the counterparts of WOX11 and AIL promote adventitious root formation in A.thaliana and poplar (Rigal et al., 2012; Hu and Xu, 2016; Geng et al., 2023). In addition, MdWRKY87 overexpression can inhibit the transcription of adventitious rooting-repressed genes, such as AUX/IAAs and RRs, and the counterparts of AUX/IAAs and type-B RRs inhibit adventitious root formation in Arabidopsis, apple, and poplar (Ramírez-Carvajal et al., 2009; Lakehal et al., 2019; Zhao et al., 2020). Similarly, it was demonstrated that PuWRKY75 interacted with PuMYB40 and directly co-regulate PuLRP1 and PuERF003 to promote adventitious root formation in P. ussuriensis (Wang et al., 2022).

In conclusion, a relatively high expression level of MdWKRY87 contribute to improving adventitious rooting ability. Based on this potential mechanisms, artificial methods for adventitious rooting ability improving will be created via manipulating MdWRKY87 gene expression, especially for rooting recalcitrant woody perennial species.