Introduction

Pest is one of the major enemies for China’s forestry production. And the occurrence frequency and damage degree of forest pests in China have increased year by year. Due to the influence of climate warming and human factors, especially the single tree species of plantation, the harm of forest pests has become more and more serious. China is one of the countries who has the most serious problem of forest pests in the world [1, 2]. During the past few years of forestry development in China, the infestation of H. cunea has been more pronounced among the associated pests and diseases. H. cunea is widely distributed in many areas of China and is a serious foreign forestry pest [3, 4]. It is a globally growing Lepidoptera forest pest species. which has the characteristics of strong reproductive capacity, strong adaptability to the environment, long damage cycle, and wide transmission routes. It is extremely harmful to forest resources [5], which can feed on about 600 kinds of deciduous trees in the world. In the event of an outbreak of this pest, the leaves of the entire plant will be eaten, seriously affecting the photosynthesis of the tree, thus affecting its normal growth and development, and even leading to the death of the tree [2, 6].

Omics technology can explore the molecular mechanisms of plant development and inheritance by analyzing the changes of DNA, proteins, metabolites and mineral nutrients in plants [7, 8]. At present, multi-omics joint analysis has been widely used in the analysis of complex biological mechanisms of plants. Transcriptomics and metabolomics are relatively mature, in-depth and simple techniques [9, 10]. Researchers have used transcriptomics technology to study the molecular mechanism of Arabidopsis, rice, Camellia sinensis and other plants in response to biotic stresses such as silverleaf whitefly, brown planthopper and tea geometrid [11,12,13]. Under thrips feeding, a variety of signaling pathways in alfalfa were activated and the expression of a large number of functional genes and TFs was induced, and the expression levels of transcription factor families such as WRKY, NAC and MYB were significantly up-regulated/down-regulated [14]. Zhao et al. showed that the feeding of tea green leafhopper could significantly affect the production of metabolites in the phenylpropanoid and flavonoid pathways in tea leaf, and the genes encoding key metabolic enzymes related to phenylpropanoid and flavonoid biosynthesis, such as phenylalanine ammonia lyase (PAL) and flavonol synthase (FLS), were significantly up-regulated [15]. It was found that after Apolygus lucorµm fed on Gossypium, the genes encoding JA biosynthesis were significantly up-regulated, and the contents of JA and JA-Ile were significantly increased [16].

Secondary metabolites are the results of adaptation to ecological environment and various stresses in the long-term evolution of plants, among which terpenes, flavonoids, polyphenols and other substances, as the main physiological and biochemical substances of plant insect resistance, play a vital role in the defense response of host plants to insects [17]. The analysis of poplar leaves with or without exposure to leaf beetles found that the content of phenolic substances in the leaves decreased after being damaged by beetles [18]. After being damaged by Macrosiphum rosae, the total phenolic content of Rosa hybrid was higher than that before stress, and the flavonoid content decreased first and then increased [19]. Studies have shown that flavonoids isoquercitrin and quercetin in cotton tissues at some specific period have a favorable effect on the occurrence and damage of Apolygus lucorum, while the increase of condensed tannins and terpenes can inhibit the damage of Apolygus lucorum [20]. Furthermore, others studied that the high content of total alkaloids in the bark of the insect-resistant Japanese pagodatree is one of the important factors for its insect resistance [21].

Poplar belongs to Salicaceae, Populus, which has the characteristics of fast growth, short rotation period, strong adaptability and wide application. It is widely used in greening, afforestation and other aspects [22]. As a woody model plant, poplar is planted all over the world, providing a large number of raw materials for the production of pulp, wood and biofuel [23]. ‘P.‘xin lin 1’ belongs to Salicaceae, Populus, which is an excellent forest variety cultivated in 2005. It grows fast with straight trunks, round crowns and dense branches and is widely planted in the cities of Shenyang, Jinzhou, Huludao and other places in China, as well as the good promotion effect is achieved [24]. However, in the actual planting process, poplar is easy to be attacked by various pests during its growth and development period due to the single planting variety, unreasonable tree age structure and large afforestation area [25]. Forestry pests cause great damage to poplar. Through genetic engineering breeding, insect-resistant genes can be efficiently expressed and have high lethality to pests, so as to control pests [26]. Current studies have shown that transgenic poplars with the same or different insect-resistant genes have a superposition effect, showing spectral insect resistance. However, the stability and effectiveness of insect resistance of transgenic poplars still need time to be further verified [27].

In this study, transcriptomics and broad-target metabolomics techniques were used to jointly analyze the effects of H. cunea on the gene expression level and metabolite content in ‘P.‘xin lin 1’, to explore the key genes and key metabolites, and to analyze the biological process of poplar in response to insect stress. Revealing the regulatory network of poplar trees in response to insect stress is of great importance and provides information for future screening of insect resistant genes and breeding of new insect-resistant varieties.

Results

Phenotype and physiological indexes

There were all significant differences (p < 0.01) in phenylalanine ammonia lyase (PAL) activity, phenolic, and flavonoid contents at different feeding times. The PAL activity, phenolic and flavonoid contents all showed a trend of decreaseing, increaseing, decreaseing and then increasing with the increase of feeding time (Fig. 1H and J). Meanwhile, the highest phenolic and flavonoid contents were both observed at 24h, whereas the lowest both at 16h.

Fig. 1
figure 1

Changes of phenotypes and physiological indexes after H. cunea larvae feeding at different time points. A Potted seedlings used in the experiment. B Not to be gnawed (CK). C To be gnawed for two hours (2h). D To be gnawed for four hours (4h). E To be gnawed for eight hours (8h). F To be gnawed for 16h (16h). G To be gnawed for 24h (24h). PAL activity (H), total phenolic (I), and flavonoid (J) contents of CK, 2h, 4h, 8h, 16h, and 24h

Full size image

Transcriptome-wide identification of DEGs after infestation by H. cunea

In order to evaluate the transcriptional profiles of poplar (‘P.‘xin lin 1’) leaves at different time points, a total of 18 cDNA libraries with three biological replicates for six samples were used to construct and which were also used for RNA-seq. As a result, a total of 824,571,664 raw reads were obtained from 18 samples and each sample gained more than 6 Gb of clean data. The average fragments scoring Q30 was all over 93.13% and the average GC content was 44.91%, indicating a high quality of sequencing (Table S1). Besides, the result of PCA indicated that all arrays using the FPKM value was shown in Fig. 2A and separated excellently. Moreover, to investigate the expression patterns of the DEGs, a comparative transcriptomic analysis was performed using the DEseq2 software. A total of 12132 significant DEGs were identified in 18 groups (Table S2). Five comparison groups were selected for analysis (2h vs. CK, 4h vs. CK, 8h vs. CK, 16h vs. CK, 24h vs. CK), which identified 8925 DEGs Table S3), wherein 6179 DEGs in 2h vs. CK, 3353 DEGs in 4h vs. CK, and 1840 DEGs in 8h vs. CK, 4185 DEGs in 16h vs. CK, 3746 DEGs in 24h vs. CK. They all showed p-adjust value < 0.05 and |log2 (Fold Change) |≥1; and there were 298 common DEGs among these five groups. In 2h vs. CK comparison group, 3231 DEGs were found to be up-regulated, 2948 were down-regulated, which indicated that many DEGs occurred in the first two hours stages. In 4h vs. CK comparison group, 1926 DEGs were up-regulated, while 1427 were down-regulated. In 8h vs. CK comparison group, 1155 DEGs were up-regulated, while 685 were down-regulated. In 16h vs. CK comparison group, 1850 DEGs were up-regulated, while 2335 were down-regulated. In 24h vs. CK comparison group, 2218 DEGs were up-regulated, while 1528 were down-regulated (Fig. 2B, D).

Fig. 2
figure 2

Functional annotation of DEGs from ‘P.‘xin lin 1’ after H. cunea larvae feeding at different time points. A PCA score plot of expression profiles from different samples. B Venn diagram of DEGs among different samples. C The top 20 KEGG enrichment pathways of DEGs. D The statistics of up-regulated and down-regulated DEGs in different comparison groups. E level-2 GO functional classifications of DEGs. F Statistical analysis of the top 20 KEGG enrichment pathways of DEGs

Full size image

Analysis of transcription factors after infestation by H. cunea

TFs play an important role in the regulation of plant growth and development [28]. These proteins that can bind specific cis-acting elements in promoter regions to determine gene expression levels, thereby controlling plant growth and development [29]. In this study, a total of 825 DEGs encoded 56 TF families: including AP2/ERF (12%), MYB (12%), C2C2(8%), bHLH (7%), NAC (7%), WRKY (6%) and so on (Fig. 3A and Table S4). Besides, transcription factor families that are not listed in the table or graph (the number of transcription factors is less than 1%) were classified as others. Figure 3B-E showed a heatmap of the top four TFs and all genes of the AP2/ERF, MYB, C2C2 and bHLH families were significantly expressed in different treatments.

Fig. 3
figure 3

Changes in the expression levels of DEGs encoding transcription factors. A Statistical analysis of transcription factor families. Heatmap of DEGs involved in (B) AP2/ERF; (C) MYB; (D) C2H2; (E) bHLH transcription factor family

Full size image

GO and pathway enrichment analysis of DEGs after infestation by H. cunea

All DEGs of the five comparison groups were divided into 32 level-two functional classification terms including 15 biological processes, two cellular components, and 15 molecular functions. Among the biological processes, the majority of genes were classified into cellular process and metabolic process, which responded to stimulus. Interestingly, the result showed that there were only two terms in cellular component including cellular anatomical entity and protein-containing complex. Besides, binding, catalytic activity and transcription regulator activity were the first three terms corresponding to molecular function (Fig. 2E and Table S5). In addition, KEGG pathway analysis of DEGs in the five comparison groups were performed and 139 pathways were annotated. Furthermore, the top 20 KEGG pathways enriched of DEGs were shown in Fig. 2C and divided into five categories-metabolism, cellular processes, environmental information processing, genetic information processing, organismal systems (Fig. 2F and Table S6). In these pathways, metabolic, biosynthesis of secondary metabolites, plant hormone signal transduction, MAPK signaling and flavonoid biosynthesis pathway were considered as the most common pathways in metabolism. These results suggested that there were many changes in genes involved in plant hormone signal transduction pathway, MAPK signaling pathway and flavonoid biosynthesis pathway.

Metabolome identification of DAMs after infestation by H. cunea

We study metabolite species and content through of metabolites by metabolomics. To get insight into the differences in metabolites under different time treatments after feeding of H. cunea larvae, a metabolome profiling via untargeted analysis used ultra-performance liquid chromatography-tandem mass spectrometry (UPLC-MS/MS) was performed. The PCA of CK, 2h, 4h, 8h, 16h, 24h samples were analyzed and which revealed 27.39% and 19.13% variations in PC1 and PC2, respectively (Fig. 4A). In addition, the venn diagram revealed that there were 9 common compounds in five comparison groups (Fig. 4B). Besides, a total of 842 DAMs were divided into 12 categories (Fig. 4C and Table S7), and most of these compounds belong to the flavonoid, followed by phenolic acids. A total of 598 DAMs (138 up-regulated and 460 down-regulated) were obtained in 2h vs. CK, which was the highest number of DAMs. A total of 377 DAMs (70 up-regulated and 307 down-regulated) were obtained in 4h vs. CK. In 8h vs. CK groups, 42 DAMs were found to be up-regulated, while 82 were down-regulated. In 16h vs. CK groups, 58 DAMs were found to be up-regulated, while 298 were down-regulated and a total of 146 DAMs were obtained in 24h vs. CK (85 up-regulated and 61 down-regulated) (Fig. 4D). The top 20 enriched KEGG pathways were shown in Fig. 4E (Table S8). As we can see, these DAMs were remarkably enriched in the pathways of flavone and flavonol biosynthesis and the biosynthesis of flavonol glycosides, which suggested that compounds associated with these pathways were more induced in response to infestation by H. cunea.

Fig. 4
figure 4

Identification of differential metabolites of H. cunea larvae after feeding on ‘P.‘xin lin 1’ at different time points. A PCA of different samples. B Venn diagram showing the number of metabolites at CK, 2h, 4h, 8h, 16h and 24h. C Component analysis of the identified DAMs from leaves. D Statistical analysis of up-regulated and down-regulated DAMs in different comparison groups. E Top 20 KEGG enrichment pathways of DAMs

Full size image

Validation of RNA-Seq data by RT-qPCR

In order to verify the availability and accuracy of the transcriptome data, the expression of 8 genes selected from six different periods (CK, 2h, 6h, 8h, 16h, 24h) was detected in ‘P.‘xin lin 1’ after feeding of H. cunea larvae by RT-qPCR (Fig. 5). The primer sequences of candidate genes were listed in Table S9. The results showed that the relative expression of these candidate genes exhibited similar patterns to be detected by RNA-seq.

Fig. 5
figure 5

RT-qPCR validation of the expression levels of eight DEGs identified by RNA-seq. The left Y-axis indicates the relative gene expression levels analyzed by RT-qPCR, and the right Y-axis represents the FPKM values obtained by RNA-seq. The X-axis represents different samples

Full size image

DEGs related to plant hormone signal transduction pathway after infestation by H. cunea

In transcriptome analysis, ‘P.‘xin lin 1’ after infestation by H. cunea larvae confirmed that plant hormone signal transduction pathway was significantly enriched pathway. The results showed that a total of 328 DEGs were identified, encoding 33 enzymes in eight hormone signal transduction pathways in our study. The FPKM value of each gene was used to plot the heatmap (Fig. 6A). These genes were assigned to multiple signals, including auxin, cytokinine, gibberellin and so on. In auxin signal transduction pathway, the expression of two TIR1 genes (POPTR_007G048200v3, POPTR005G141800v3) were found to be significantly higher at 8h and then declined. After infestation by H. cunea, the majority of SAUR genes had high expression levels at CK, 2h and 4h, but showed low expression levels at 8h, 16h and 24h. In cytokinine signal transduction pathway, in which there were 30 DEGs encoding four enzymes, the expression of CRE1 genes POPTR_010G102900v3 was highest when infestation happened by H. cunea at 2h. All the genes in AHP, B-ARR, A-ARR were differentially expressed under different time treatments. The majority of GID1 genes expression level in gibberellin signal transduction pathway showed no obvious trend. Besides, GID2 gene (POPTR_014G165700v3) expression decreased after infestation. There were 21 DEGs encoding five enzymes in abscisic acid signal transduction pathway, and all PYR/PYL genes were expressed the highest in CK period (except POPTR_010G000600v3). Besides, two genes encoding ABF (POPTR_009G101200v3, POPTR_002G018400v3) showed a high level of expression in response to infestation by H. cunea at 16h. All genes encoding CYCD3 also exhibited the highest expression level in brassinosteroid signal transduction pathway in CK period. There were a total of 24 genes encoding two enzymes in jasmonic acid signal transduction pathway, the expression of five JAZ genes (POPTR_001G166200v3, POPTR_003G068900v3, POPTR_006G139400v3, POPTR_003G165000v3, POPTR_006G217200v3) and two MYC2 genes (POPTR_002G176900v3, POPTR_014G099700v3) gradually increased after infestation. Besides, there were a total of three compounds enriched in this pathway and they involved two classes (organic acids, amino acids and derivatives). Figure 6B showed that three compounds accumulated differently in six stages. The results of this study showed that the expression of jasmonoyl-L-Isoleucine increased with the extension of the infestation time by H. cunea.

Fig. 6
figure 6

Analysis of the plant hormone signal transduction pathway. A The color scale from Min (green) to Max(red) refers to the expression value from low to high. Uppercase letters indicate genes which encode enzymes. Solid arrows represent established biosynthesis steps and broken arrows indicate the involvement of multiple enzymatic reactions. Two solid gray lines respresent cell membrane. The grey dashed line respresents the nuclear membrane. B Histogram of the accumulation of compounds in six samples

Full size image

DEGs related to MAPK signaling pathway after infestation by H. cunea

The MAPK cascade plays an important role in coping with various stresses in different plants and at different stages of their growth and development, Fig. 7 illustrated the MAPK signaling pathway. GO, KEGG enrichment, and clustering analyses indicated that the MAPK signaling pathway was differentially affected by feeding of H. cunea, which was complex. Therefore, a total of 211 structural genes in different periods were identified. The result showed that most of them were assigned to flg22 signals and showed high expression level after 8 h of infestation. There were a total of 36 DEGs encoding 8 enzymes in the H2O2 signal transduction pathway, the expression of ANP1 genes was dramatically down-regulated after infestation. Besides, in ethylene branch, the expression levels of many genes were significantly up-regulated by H. cunea infestation. In the jasmonic acid signal transduction pathway, it also found that two MYC2 genes (POPTR_002G176900v3, POPTR_014G099700v3) were significantly up-regulated after infestation. The results implied that these genes played an important role in response to stress after H. cunea larvae feeding at different time periods.

Fig. 7
figure 7

Analysis of the MAPK signaling pathway. The color scale from Min (green) to Max(red) refers to the expression value from low to high. Uppercase letters indicate genes that encode enzymes. Solid arrows represent established biosynthesis steps and broken arrows indicate the involvement of multiple enzymatic reactions

Full size image

DEGs related to flavonoid, flavone and flavonol, anthocyanin biosynthesis pathway after infestation by H. cunea

Flavonoid is a significant class of secondary metabolites and which is widely found in plants [30]. In order to have a better understanding about the relationship between metabolites and genes in flavonoid, flavone and flavonol and anthocyanin biosynthesis pathways, 45 DEGs encoding 13 enzymes were analyzed aiming to show the relationship between gene expression and metabolite accumulation intuitively (Fig. 8), including PAL (two DEGs), CH4 (one DEG), 4CL (nine DEGs), CHS (six DEGs), CYP75A (one DEG), CYP75B1 (four DEGs), DFR (six DEGs), F3H (two DEG), FG2 (five DEGs), FLS (six DEGs), PAL (two DEGs). There were 39 DEGs in the flavonoid biosynthesis pathway, 11 DEGs in the flavone and flavonol biosynthesis pathway and one DEG enriched in the anthocyanin biosynthesis pathway. In biosynthetic step, from p-Coumaric acid to p-Coumaroyl-CoA and naringenin to apiforol, the expression of two 4CL genes (POPTR_006G169700v3, POPTR_006G169700v3) and one DFR gene (POPTR_006G087200v3) were significantly up-regulated compared with CK in the process of H. cunea infestation, respectively. For FH3 genes, all genes were activated at 2h. Besides, the majority of CYP75B1 genes had no significant changes. One FG2 gene (POPTR_008G024900v3) was significantly up-regulated in the flavone and flavonol biosynthesis pathway. Furthermore, in the anthocyanin biosynthesis pathway, one BZ1 gene also was activated at 2h. The DAMs in these three pathways were shown in Fig. 8, including cyanidin-3-O-glucoside, delphinidin-3-O-glucoside, petunidin-3-O-glucoside, cosmosin, trifolin, nictoflorin, phloretin, apigenin, 3’-O-Methylluteolin, 3-O-Methylquercetin, taxifolin, isoquercitrin, kaempferol and so on.

Fig. 8
figure 8

Analysis of the flavonoid, flavone and flavonol, anthocyanin biosynthesis pathways. A Overview of the flavonoid, flavone and flavonol, anthocyanin biosynthesis pathways. B Heatmap of the expression of genes associated with the flavonoid, flavone and flavonol, anthocyanin biosynthesis pathways. The expression profiles of genes at the six stages are shown in separate columns within the colored boxes, with each row representing a distinct gene. The color gradient indicates the expression levels, with red indicating up-regulated and green indicating down-regulated. C Histogram of the accumulation of compounds in the six samples

Full size image

Discussion

With the development of high-throughput experimental methods of omics, including transcriptomics, metabolomics profiling, proteomics and so on, the ability to characterize plant responses to biological (such as insect attacks) and abiotic stresses has been enhanced [31]. These methods normally are used to assess the interactions between plants and herbivorous insects. For example, transcriptomics has been used to identify gene expression at the RNA level. Metabolomics makes it possible to identify the types, quantities and changes of metabolites in disturbed organisms [32]. In the last few years, the combined analysis of transcriptomics and metabolomics has been used for a variety of plants species to identify gene-to-metabolites networks associated with insect infestations, such as rice [33], maize [34], cotton [35], arabidopsis [36] and wheat [37]. This study explored the dynamic response mechanism of physiology, transcriptomics and metabolomics of poplar (‘P.‘xin lin 1’) at different feeding time points of H. cunea.

TFs genes

Infestation by H. cunea induced substantial overall changes in ‘P.‘xin lin 1’ leaf transcriptome. Compared with the control group, the DEGs of five comparison groups were analyzed. The results increased our understanding of mechanisms underlying the dynamic responses of ‘P.‘xin lin 1’ plants to H. cunea feeding. Gene expression analyses showed that 8925 genes in H. cunea-infested plants were differentially expressed. The previous studies displayed more up-regulated DEGs than down-regulated ones in response to feeding by H. cunea larvae. The result was consistent with the previous studies which concerning cotton was infested by chewing insects Anthonomus grandis or Helicoverpa armigera [38]. Similarly, there were more up-regulated DEGs than down-regulated ones when rice plants were attacked byrice stem borer [39] and arabidopsis plants were damaged by Myzus persicae, Brevicoryne brassicae or Pieris rapae [40]. Besides, this was the same phenomenon as Mythimna separata insects specifically induced defence responses in maize [41]. Nevertheless, there was also a study indicated that more DEGs were down-regulated not up-regulated or the number of DEGs in up-regulated and down-regulated were equivalent when rice [42] or cotton [43] were damaged by brown planthopper N. lugens or aphid Aphis gossypii. The reason for this difference would be that plant species, infected tissues, infection duration, and herbivore species caused different gene expressions [40]. In this study, 826 TFs were identified among ‘P.‘xin lin 1’ genes which responded to H. cunea feeding, suggesting that plant defense against insect herbivore attack was a multidimensional dynamic procedure. The top four families with the highest number of TFs after H. cunea infestation were AP2/ERF, MYB, C2C2, bHLH. Previous research has widely present that AP2/ERF involved in the regulation of growth and development, metabolism [31, 44], and responded to biotic and abiotic stresses. For example, plants defend against herbivorous insects and necrotic pathogens by modulating different branches of the jasmonic acid signaling pathway [44]. Furthermore, members of MYB were widely involved in plant secondary metabolism regulation, cell morphogenesis, stress response. Evidence proved that NbMYB42, NbMYB107, NbMYB163, and NbMYB423 were resistant to whitefly [45]. The role of TFs in ‘P.‘xin lin 1’ insect resistance remains to be further researched.

Plant hormone signal transduction pathway

A number of studies have shown that the plant hormone signal transduction pathway plays an important role in plants’ resistance to herbivores [39]. In this study, GO and KEGG enrichment indicated that ‘P.‘xin lin 1’ phytohormone signaling pathways were differentially affected after infestation by H. cunea. In the auxin signal transduction pathway, POPTR_007G048200v3 and POPTR005G141800v3 encoding the TIR1, was found to be significantly high in 8h period and then declined which demonstrated that different induced defense responses occurred in different time periods. Nevertheless, the expression of the majority of PYR/PYL genes in the abscisic acid signaling pathway had no significant changes after H. cunea infestation, indicating that these genes showed a weak response to H. cunea infestation. The expression levels of several genes in the jasmonic acid signaling pathway, including the JAZ genes (POPTR_003G068900v3, POPTR_006G139400v3) and MYC2 gene (POPTR_002G176900v3), were significantly up-regulated after H. cunea infestation. In addition, data in this study showed that a large number of JA-related genes were up-regulated by larval feeding and which was consistent with rice plants’ responses to rice stem borer Chilo suppressalis infestation [39]. Furthermore, it was show that genes associated with the JA pathway were most induced when rice plants were infested by Cnaphalorocis medinalis [32]. A previous study indicated that silencing OsHI-LOX and OsPLDa4/5 genes involved in JA biosynthesis in rice could lead to a decrease in JA levels and trypsin protein inhibitor content, thereby promoting the growth and development of rice striped stem borer Chilo suppressalis [32]. The results in this research, along with previous studies, suggested that jasmonic acid signaling pathway played a central role in plant defense against insect herbivore attack and these genes performed an important function in ‘P.‘xin lin 1’ s response to insect H. cunea. Both JA and JA-Ile metabolism accumulation were detected. Studies showed that TrypPI, polyphenoloxidase and plant volatiles were used by the application of MeJA and JA on rice plants activated various defense metabolite [46,47,48]. JA and JA-Ile emerged as an important signal compounds activated defense responses to herbivores [49]. Together with those of previous reports, it can speculate that these two compounds were possible involved in poplar‘s defense against H. cunea.

MAPK signaling pathway

After attacking by herbivores, plants activated MAPK signaling, reshaping the transcriptome, and thus the proteome, in preparation for defense against attack [49, 50]. In this study, two MYC2 genes including POPTR_002G176900v3 and POPTR_014G099700v3 were also found to significantly increase from CK to 24h in jasmonic acid branch of MAPK pathways. Meanwhile, this pathway also responded to wounding, which further proved the role of jasmonic acid pathway in insect resistance. In addition, in ethylene branch, the expression levels of many genes, such as one CTR1 gene (POPTR_009G139400v3) and one MKK9 gene (POPTR_015G030700v3), were significantly up-regulated after H. cunea infestation and this pathway also activated the defense response, which were likely transcriptional defense regulation. The expression of MPK3/6 genes POPTR_009G066100v3 at 4h had no evident change. After 8h of feeding, the expression of this gene significantly increased. These results suggested that the defense response of ‘P.‘xin lin 1’ was tightly controlled, regulating the expression of different genes at various time points to initiate optimal defense processes. Previous studies demonstrated the role of MAPK pathway in insect resistance. For example, the expression of several MAPKs, including OsMPK3 and MAPKKs was significantly up-regulated after C. medinalis feeding, which proved that MAPK signaling played a central role in regulating the induced defense responses of rice to this pest and which is consistent with this research [51]. This suggested that ‘P.‘xin lin 1’ also responded to the infestation of H. cunea by regulating the up-regulated genes of CTRI (POPTR_009G139400v3) and MKK9 (POPTR_015G030700v3). In rice plants, OsWRKY70 interacted with OsMPK3 and OsMPK6 and was regulated by them. And under attack by C. suppressalis, priority was given to defense over growth by positively regulating JA and negatively regulating gibberellin biosynthesis [52]. All the results indicated that MAPK pathway was crucial for ‘P.‘xin lin 1’ defense against H. cunea insect.

Flavonoid, flavone and flavonol, anthocyanin biosynthesis pathway

Many defensive secondary metabolites present in plants, such as flavonoids, terpenes, and alkaloids, play crucial roles in the plants against pests [53,54,55]. In this study, a total of 842 DAMs were identified associated with regulating stress responses. The results showed that most DAMs belonged to flavonoids. The KEGG analysis results showed that most of the DAMs were significantly enriched in the flavone and flavonol biosynthesis pathway. Similarly, the transcriptome data also showed that a large number of DEGs enriched in the flavonoid pathway. Flavonoids were the unique secondary metabolites in plants to resist biotic and abiotic stresses [56]. Flavonoid biosynthesis pathway contained several branches which lead to the biosynthesis of flavanones, flavone and flavonol, isoflavones, anthocyanidins and so on [57]. In these pathway map, it was observed that the infestation of H. cunea activated the expression of some genes (PAL, CH4, 4CL, CHS, etc.) and led to the accumulation of metabolites (such as kaempferol). This confirmed the plant hormone signal transduction pathway. In this study, the transcript expression of several flavonoid pathway-related genes was significantly up-regulated after infection by H. cunea and the genes were 4CL (POPTR_006G169700v3, POPTR_006G169700v3), DFR (POPTR_006G087200v3) and FG2(POPTR_008G024900v3). FH3 genes were activated at 2h and then the expression level decreased gradually, and reached to the lowest at 16h, which indicated that plant defense against insect feeding changed dynamically. Besides, in the measurement of physiological indexes, the PAL activity, phenolic and flavonoid contents were higher at 8h. This pattern of secondary metabolites accumulative has been mentioned in a number of studies. After Castanopsis eyrei infestation by herbivorous pests, there was a significant increase in the flavonoid content. Plants could prevent herbivores from oviposition, feeding and developing through the accumulation of flavonoids [32]. Furthermore, aphid attack led to changes in the metabolism of phenylalanine and tyrosine in wheat that possibly because wheat produced phenolic acids to protect itself from aphids when it was attacked infested [58]. In addition, the previous study indicated that feeding an artificial diet containing flavonoid naringenin significantly reduced the fertility and intrinsic natural growth rate of Acyrthosiphon pisum Harris [59]. And artificial application of flavonoids could enhance resistance to N. lugens [60]. Combined with previous studies, a significant increase in substances such as flavonoids and phenolic acids was observed 8 h after the attack by H. cunea, which initiated an optimal defense process. Studies have shown that PAL were key genes in the flavonoid and flavonoid and flavonol biosynthesis pathways [61]. In this study, the increase of the PAL activity indicated the activation of flavonoid pathway. At the same time, the expression change trends of cyanidin-3-O-glucoside, delphinidin-3-O-glucoside and petunidin-3-O-glucoside expression were down-regulated at the N-4 h locus, up-regulated at the 4–8 h locus and down-regulated at the 8–16 h locus, which was consistent with the trend of PAL. We speculated ‘P.‘xin lin 1’ affected the accumulation of flavonoids by regulating PAL transcription factors to respond to the damage of H. cunea. All in all, flavonoids were the key defense substances in ‘P.‘xin lin 1’ after infestation by H. cunea.

Methods

Plant materials and treatments

This study was carried out in the teaching and research base of Jilin Agricultural University (43°81΄N;125°41΄E). The three-month-old seedlings of ‘P.‘xin lin 1’ were used as experimental targets. The source of H. cunea was provided by Chinese Academy of Forestry Sciences and which was bred in the alien insect pest research base of Jilin Agricultural University. The feeding test began in August 2023 and the specific operation methods are as follows. The third instar larvae of H. cunea were starved for 24 h before the initiation of the formal experiment. Subsequently, the insects were placed on the leaves of ‘P.‘xin lin 1’ seedlings with the same growth state, three leaves were selected from each seeding, and five third instars were placed on each leaf. The leaves without insect feeding test were taken as control. There were five replicates per treatment group. Leaves inoculated for 2h, 4h, 8h, 16h and 24h were collected. All leaves were mixed and frozen in liquid nitrogen quickly.

Physiological index measurement

The leaves were carefully grounded with liquid nitrogen, and then 0.1 g sample was taken in a centrifuge tube for the determination of relevant physiological indices. Assay kits of PAL activity, plant flavonoid content, plant total phenol content were purchased from Beijing Box Production Technology Co. Ltd. (Beijing, China). The specific operation steps and calculation formulas followed the instructions and formulas provided by the kit. Each sample has three technical replicates.

RNA isolation, cDNA library construction, and transcriptome sequencing

Plant total RNA was extracted from samples using the Plant Total RNA isolation kit (Takara, Beijing, China) in three biological replicates. RNA concentration and purity were determined by using a Nanophotometer spectrophotomete, and RNA integrity was assessed by agarose gel electrophoresis. Most mRNA in eukaryotes have the structural feature of polyA tails. PolyA-mRNA was enriched by Oligo(dT) magnetic beads. The fragment lysis buffer was used to lyse RNA into fragments, which were then used as templates to synthesize the first strand of cDNA. Double-stranded cDNA was synthesized by adding dNTPs (buffer dTTP, dATP, dGTP, dCTP) and DNA polymerase I, and then purified by DNA purification magnetic beads. Final PCR enrichment was performed to obtain the final cDNA library. Initial quantification was performed using the Qubit dye method to detect the library insert size by a fragment analyzer. Q-PCR method accurately quantified the library effective concentration (library effective concentration > 2 nM) and completed the library detection. They were sequenced using the Illumina platform (HiSeq™ 2500, United States). The raw data of 125–150 bp paired-end reads were generated. After filtering and removing the low-quality raw reads, the clean reads were mapped to the Populus trichocarpa reference genome using HISAT2 [62].

Differential gene screening and functional enrichment

Differential expression analysis among sample groups was performed using DESeq2 R software package (1.20.0). Multiple hypothesis testing correction for hypothesis testing probability (P-value) using Benjamini-Hochberg method yielded the False Discovery Rate (FDR). DEGs were screened using the criteria of |log2 Fold Change| ≥ 1, and p value < 0.05. GO analysis of DEGs was implemented using GOseq R package based on the Wallenius non-central hypergeometric distribution. The enrichment of DEGs in the KEGG pathway was detected using cluster-Profiler3.4.4 software.

Sample preparation and metabolite measurements

The leaf samples were vacuum freeze-dried in a lyophilizer (Scientz-100 F, Scientz Biotechnology, Zhejiang, China), and then grounded with a grinder (MM 400, Retsch, Nordrhein-Westfalen, Germany) (30 Hz, 1.5 min). 1200 µL pre-cooled 70% methanol water was added, vortex centrifuged (12000 rpm, 3 min), and the supernatant was collected, a filtered through a microporous filter membrane (0.22 μm pore size) and stored in sample bottle. Leaves extracts were analyzed using UPLC-MS/MS.

Metabolite analysis

Overall metabolic differences between samples and groups were obtained by PCA. Orthogonal partial least squares discriminant analysis (OPLS-DA) was used to filter orthogonal variables independent to categorical variables in metabolites. And non-orthogonal and orthogonal variables were filtered separately. Hierarchical clustering analysis (HCA) was performed on the metabolite accumulation patterns of different samples. DAMs met the criteria of VIP ≥ 1, fold change ≥ 2 or ≤ 0.5 and P-value < 0.05, and the enrichment pathways of DAMs were analyzed using KEGG database.

Real-time quantitative PCR detection

To validate the transcriptome data, eight DEGs for RT-qPCR were selected randomly. RNA Easy Fast Plant Tissue Kit (TianGen, Beijing, China) was used to extract RNA from leaves. All cDNAs were obtained using Prime Script RT reagent Kit and gDNA Eraser kit (TaKaRa, Beijing, China), and qRT-PCR detection was performed using TaKaRa SYBR Green Mix kit (TaKaRa, Beijing, China), as well as using Gel Doc Go imaging system and Rapid Real-time detection system. All primers were designed using Primer Premier 5.0 and listed in Table S9. Ptactin was used as housekeeping gene. The relative expression levels of DEGs were calculated according to 2−ΔΔCT method. We used three biological replications and three technical replications.

Statistical analysis of data

Statistical analysis was performed using version IBM statistics SPSS 26.0. Analysis of variance (ANOVA) and F test were used to test for the significance of the sample [63]. One-way ANOVA was used to examine the effect of H. cunea larvae eating leaves on ‘P.‘xin lin 1’.

$$X_{ij}=\mu+T_i+e_{ij}$$

Where Xij is the performance of an individual tree j in treatment i; µ is the overall mean; Ti is the fixed effect of treatment, and eij is the random error.