Age groups including nine of 5-, 100-, and 700-year-old P. orientalis donors were selected in Beijing Botanical Garden to collect strong and pest-free branches that were 3–5 m above the ground. They were cut into 10–15-cm cuttings at the base 1.0–1.5 cm from buds. The cuttings were dipped in sodium lignosulfonate:NAA : IBA = 1:1:1 at the concentration of 1,000 ppm for 1 min, with 25 cuttings in each treatment and three repeats for each treatment. The cutting experiment was carried out under full sunshine and automatic spray at the Chinese Academy of Forestry.

Stage 1 (S1) was the P. orientalis cuttings have just been taken. Stage 2 (S2) was when the white or red callus began to grow at the cut site of the cuttings 45 days after cutting. Stage 3 (S3) was when the emergence of AR tips from the callus was evident 90 days after cutting. A scalpel was used to peel off the xylem at a distance of 0.5 cm from the base of the cuttings of S1, S2, and S3, and the remaining parts were then put into liquid nitrogen as metabolome and proteome materials.

After the culture of the cuttings propagated from 100-year-old P. orientalis for 30 days, the callus was incised, and the non-incised callus was used as control. The wounded cuttings were dipped in sodium lignosulfonate:NAA : IBA = 1:1:1 at the concentration of 1,000 ppm for 1 min, with 25 cuttings in each treatment and three repeats for each treatment. After culturing for 40 days, the rooting rates and root numbers were counted, and 0.5-cm branches (without xylem) at the wounded (ck, w-s1, and w-s2) and non-wounded (ck, nw-s1, and nw-s2) cutting base were collected at 0, 20, and 40 days and then put into liquid nitrogen as metabolome and proteome materials.

A powder sample (50 mg) was collected in a 2-ml centrifuge tube containing grinding beads. Metabolites were extracted using 400 μl of extracting solution containing 0.02 mg/ml internal standard (Vmethanol : Vwater = 4:1, with 0.02 mg/ml of l-2-chlorophenylalanine as the internal standard). The sample solution was ground at −10°C and 50 Hz for 6 min in a frozen tissue grinder, and then extraction was performed by low-temperature ultrasound at 5°C and 40 kHz for 30 min. Afterward, the sample was placed at −20°C for 30 min, centrifuged at 13,000 g for 15 min at 4°C, and transferred into a tube for liquid chromatography-mass spectrometry (LC-MS) detection and analysis. All metabolites of the same volume were collected and mixed to prepare quality control (QC) samples. One QC was inserted into every 10 samples to investigate the repeatability of the whole analysis.

LC-MS analyses were performed using a UHPLC-Q Exactive HF-X system (1290, Agilent Technologies, Shanghai, China) with a UPLC HSS T3 column (100 mm × 2.1 mm i.d., 1.8 µm) coupled to Q Extractive unit (Orbitrap MS, Thermo, Waltham, MA, USA). Mobile phase A was 95% water + 5% acetonitrile (containing 0.1% formic acid), and mobile phase B was 47.5% acetonitrile + 47.5% isopropanol + 5% water (containing 0.1% formic acid). The gradient elution procedure of the mobile phases is according to Xie et al. (2019). The mass spectrum (MS) signals of the sample were scanned in positive and negative ion modes, with a mass range of 70–1,050 m/z. The sheath gas flow rate was set at 50 psi, the auxiliary gas flow rate at 13 psi, the auxiliary gas heating temperature at 425°C, the ion spray voltage in the positive mode at 3,500 V, the ion spray voltage in the negative mode at −3,500 V, the ion-transfer tube temperature at 325°C, and the normalized collision energy at 20–40–60 V cyclic collision energy. The resolution of primary MS was 60,000, and that of secondary MS was 7,500. The data were collected using data-dependent acquisition (DDA) mode.

After analysis, LC-MS raw data were imported into metabonomic processing software progenesis for baseline filtering, peak recognition, integration, retention time correction, and peak alignment. Finally, a data matrix of retention time, mass-to-charge ratio, and peak intensity was obtained. In addition, the MS and tandem MS data were matched with the public metabonomic databases HMDB (http://www.hmdb.ca/) and Metlin (https://metlin.scripps.edu/), and the self-built database of Majorbio to obtain the metabolite information.

The matrix data after the database search were uploaded to the Majorbio Biological Cloud Platform (https://cloud.majorbio.com) for data analysis. First, data preprocessing was conducted. The missing values of the data matrix were removed using the 80% rule; that is, more than 80% of the non-zero variables in at least one group of samples were retained, and then vacant values were filled in (with the minimum value of the original matrix). To reduce the errors caused by sample preparation and instrument instability, the response intensity of the MS peaks was normalized using the sumNormalizer to obtain the normalized data matrix. Additionally, the variables with relative standard deviation (RSD) of QC samples >30% were deleted, and logarithmic processing with log10 was performed to obtain the final data matrix for subsequent analysis.The differences in the preprocessed matrix file were analyzed. The R package ropls performed principal component analysis (PCA) and orthogonal least partial squares discriminant analysis (OPLS-DA). The stability of the model was evaluated by seven cycles of interactive verification. Moreover, Student’s t-test and the fold-change analysis were carried out. Differential metabolites were selected based on the variable weight value (variable importance in projection (VIP)) obtained from the OPLS-DA model and the p-value from Student’s t-test. The metabolites with fold change (FC) >1.5 or <0.67, VIP > 1, and p < 0.05 were significant differential metabolites. The differential metabolites were annotated by the Kyoto Encyclopedia of Genes and Genomes (KEGG) database (https://www.kegg.jp/kegg/pathway.html) to obtain the pathways involved in the differential metabolites. Pathway enrichment analysis was conducted using the Python software package scipy.stats, and the biological pathway most relevant to the experimental treatment was obtained by Fisher’s exact test.

For proteome analysis, iTRAQ technology was used to obtain raw data, Uniprot database was used for searching, t.test function in R language was used to calculate the p-value of the significant difference between samples, and at the same time, the fold of difference between groups was calculated. The screening criteria for significantly differentially expressed proteins (DEPs) were as follows: upregulated proteins with p < 0.05 and FC > 1.2, and downregulated proteins with p < 0.05 and FC < 0.83. Then, KEGG pathway analysis was performed for these DEPs.

The known proteins related to the biosynthetic pathways of phenolics and flavonoids were found in proteomics data during AR formation of the cuttings propagated from P. orientalis donors at different ages. Meanwhile, relative contents of phenolics and flavonoids were obtained from metabonomics data during AR formation of the cuttings propagated from P. orientalis donors at different ages. The correlation between proteins and metabolites was calculated as Pearson’s correlation coefficients. Correlation pairs were deemed statistically significant when the absolute value of Pearson’s correlation coeffificient (PCC) >0.9 and p-value<0.01.

In order to investigate physiological changes, endogenous phenolics, flavonoids, total soluble sugar, soluble protein, and lignin contents were quantified by using an assay kit (Cominbio, Suzhou, China) according to the methods of Qian et al. and Du et al. (Du et al., 2019; Wei et al., 2020).

For lignin staining, transverse sections were dipped in 1% phloroglucinol and 12% HCl for 10 min, shifted in an acidic solution (10% H₂SO₄ and 75% glycerol), and observed under bright-field Leica M205 FA microscope (Leica Microsystems, Wetzlar, Germany). Indole-3-acetic acid (IAA) and zeatin were extracted using isopropanol/water/hydrochloric acid and quantified by using a UPLC-MS/MS platform (Agilent 1290 Infinity II UPLC system equipped with the AB Sciex QTrap 6500+ mass spectrometer) and calibrated by using inositol as an internal control.

Observation of the AR formation process of cuttings propagated from P. orientalis donors at different ages revealed that the rooting rates and root numbers of cuttings propagated from 5-year-old P. orientalis donors (control) were 78.14% and 7.17, respectively, which were significantly higher than those of 100- and 700-year-old P. orientalis donors (p < 0.05) ( Figure 1 ). However, no significant difference was observed in the rooting rates and root numbers of the cuttings propagated from 100- and 700-year-old P. orientalis donors (p > 0.05). These results suggest that the rooting rates of cuttings may decrease significantly in the age of ancient trees.

Ultraperformance liquid chromatography–tandem mass spectrometry sequencing was performed to uncover relevant metabolites during the AR formation of P. orientalis cuttings. The relatively clustered distribution within each group implied good repeatability, and the evident difference in metabolites among groups suggested biological differences among samples. A total of 8,623 anionic and 9,231 cationic metabolites were detected in this experiment. Further comparison of the numbers of differentially accumulated metabolites (DAMs) among groups revealed the obvious differences among S1, S2, and S3 of the cuttings propagated from 5-, 100-, and 700-year-old P. orientalis donors ( Figure 2A ). The numbers of DAMs at S3/S1 and S2/S1 were less than those at S3/S2, which indicated that the numbers of DAMs in cuttings increased with the AR formation process. In addition, the numbers of DAMs in cuttings propagated from 100- and 700-year-old P. orientalis donors were the least at S2 and S3, respectively, which denoted a similarity in their AR formation process.

To explore the types of DAMs in different stages of AR formation, we performed KEGG pathway enrichment analysis of significantly upregulated DAMs in these groups. The upregulated metabolic pathways (flavone and flavonol biosynthesis pathway and phenylalanine metabolism biosynthesis pathway) were significantly enriched at S2/S1 and S3/S1 of cuttings propagated from 100- and 700-year-old P. orientalis donors (p < 0.05) ( Figure 2B ). The flavonoid biosynthesis pathway and fructose and mannose metabolism pathway were enriched at S3/S1 of cuttings propagated from 5- and 100-year-old P. orientalis donors, which suggested that in addition to defense response, energy synthesis is required for AR formation. Notably, a significant accumulation of phenolics occurred in the phenylalanine metabolism biosynthesis pathway at S2 and S3 of cuttings propagated from 100-year-old P. orientalis donors, which indicated that the syntheses of secondary metabolites and stress-resistant substances were possible to inhibit the AR formation of ancient trees.

To discover the key DAMs involved in the phenylpropanoid and flavonoid biosynthesis pathway that inhibit the AR formation of cuttings propagated from ancient trees ( Tables S1, S2 ), PCC analysis was conducted on DAMs and DEPs ( Tables S3, S4 ). As for the phenylpropanoid biosynthesis pathway, 450 pairs of correlations, including 216 and 234 pairs of positive and negative correlations, were observed ( Figure 3A ; Tables S5, S6 ). Meanwhile, for the flavonoid biosynthesis pathway, 420 pairs of correlations, including 222 pairs of positive correlations and 198 pairs of negative correlations, were generated. With |PCC| > 0.8 and p < 0.01 as the criteria, 13 DAMs (such as coniferyl alcohol, cinnamoyl-CoA, and caffeic acid) and 16 DAMs (such as cinnamic acid, caffeic aldehyde, and 4-coumarate) were screened with strong positive and strong negative correlations in the phenylpropanoid biosynthesis pathway, respectively ( Figure 3B ). Meanwhile, nine DAMs such as phloretin, cinnamoyl-CoA, and isoliquiritigenin had strong negative correlations, and eight DAMs such as quercetin, naringenin, and chlorogenic acid had strong negative correlations in the flavonoid biosynthesis pathway. In addition, the upregulated expression of peroxidase (PER) related to the phenylpropanoid biosynthesis pathway promoted the accumulation of lignin synthesis metabolites ( Figure 4A ). Moreover, the synthesis pathways of phenylpropane, phenolics, and flavonoids were the main pathways that formed lignin.

In view of the aforementioned research, the ancient P. orientalis produced a large number of secondary metabolites during the cutting process. The contents of phenolics, flavonoids, and lignin increased with the increase in tree age of P. orientalis and during AR formation. The cuttings propagated from 5-year-old P. orientalis donors contained significantly higher auxin content than the 100- and 700-year-old P. orientalis donors (p < 0.05) ( Figure 4B ). A negative correlation existed between secondary metabolites. In addition, no significant difference was observed at the same stage of cuttings propagated from 100- and 700-year-old P. orientalis donors, which suggested small differences in ancient P. orientalis cuttings.

To promote AR formation, we employed wounding to break down the mechanical barrier of lignification of callus in cuttings propagated from ancient P. orientalis donors. After wounding, the rooting rates and rooting numbers of cuttings propagated from 100-year-old P. orientalis donors were 11.1% and 2.31, respectively, which were significantly higher than those of cuttings without wounding, accompanied by a significantly shortened rooting time (p < 0.05) ( Figure 5 ). These data suggest that wounding significantly improved the rooting outcome of ancient P. orientalis cuttings.

Our study further identified the key metabolites of cuttings in the process of wounding-promoted AR formation. The screening of significant DAMs that were upregulated in w-s2 than in nw-s1 and nw-s2 was mainly enriched during the biosynthesis of zeatin, aminoacyl-tRNA, and phenylpropanoid at p < 0.05 ( Figure 6 ). This result indicated the promoted cell division and energy accumulation in AR formation after wounding. Meanwhile, DAMs that were upregulated in w-s1 compared with nw-s1 were enriched significantly in aminoacyl-tRNA biosynthesis and arachidonic acid metabolism pathway and thus served as a material basis for AR formation.

Our subsequent analysis clarified changes in the contents of key metabolites after wounding. The contents of 5′-methylthioadenosine, cis-zeatin-O-glucoside, and adenine in zeatin biosynthesis gradually increased after wounding ( Figure 7A ). Thus, wounding reactivated and promoted cell division, which regulated the differentiation and growth of various cells (Antoniadi et al., 2022; Sun et al., 2023). Metabolites, such as l-glutamine, l-histidine, l-isoleucine, l-leucine, and l-arginine, are involved in aminoacyl-tRNA biosynthesis and are sources of carbon skeleton for basic metabolite synthesis on wounding response ( Figure 7B ) (Guan et al., 2021). Meanwhile, after the first wounding, secondary metabolic activities were enhanced, with increased contents of phenylalanine, phenolics, flavonoids, and other secondary metabolites. Therefore, the stress resistance of the callus improved after wounding.

Finally, this study verified relevant DAMs during wounding-promoted AR formation. The contents of zeatin and soluble proteins increased significantly after wounding and reached the peak at w-s2 (p < 0.05) ( Figure 7C ). Hence, zeatin may promote cell redivision around the wound, and soluble proteins regulate osmotic stress and provide a carbon source for cell growth. In addition, the contents of phenolics and flavonoids increased after wounding, which indicated improved free radical scavenging activity and damage resistance.