Tolerance of Triploid Hybrids of White Poplar ‘Beilinxiongzhu 1’ to Genetic Transformation Screening Agents In Vitro

Abstract

Genetic transformation of forest trees is essential for validating gene functions and breeding new varieties through molecular means. Appropriate selective pressure is critical for creating an effective screening system. ‘Beilinxiongzhu 1’ sensitivity testing showed that the critical tolerance concentrations for hygromycin (Hyg), kanamycin (Kan), and glyphosate (PPT) in leaf explants were 2.0 mg/L, 20 mg/L, and 1.0 mg/L, respectively. Among the physiological indicators, soluble sugar content, soluble protein content, and endogenous hormone levels were identified as key markers of the effects of the different antibiotic treatments. Transcriptome analysis showed that Hyg treatment resulted in a large number of differentially expressed genes (DEGs) involved in leaf cell wall synthesis and glucose metabolism. Under Kan treatment, the DEGs were associated with pathways such as ribosome biosynthesis and histone packaging in eukaryotes. Under PPT treatment, significant DEGs were related to ABC transporters. DEGs common to all three antibiotics were involved in glutathione metabolism pathways. A weighted gene co-expression network analysis identified TRXH2, H3.2, H2B, GST, U71K1, and CHS as key genes in response to antibiotic stress. By elucidating the physiological and molecular mechanisms by which different antibiotics affect leaf sprouting, our study serves as a reference for research into the genetic transformation of poplar leaves.

1. Introduction

Poplar is one of the most widely distributed tree species in the world, cultivated for its fast timbering and short rotation period. It is also an ideal model plant for research into gene function and genetic improvement in forest trees because of its simple nutrient requirements, small genome, and easy genetic transformation [1,2,3]. ‘Beilinxiongzhu 1’ poplar is a new hybrid triploid variety selected and bred by Beijing Forestry University. It combines the advantages of a hybrid genome and ploidy. It is characterized by non-flocculation of male plants, rapid growth, high basic wood density, long fibers, and high cellulose content. and has been broadly adopted in both ecologically protected and timber forests [4,5]. Genetic transformation constitutes the core of gene function verification, including in forest trees. Its use in breeding new tree species varieties by molecular means relies on the screening of resistant buds. In the genetic transformation of poplar, cell transformation is usually carried out using antibiotic marker genes conferring antibiotic resistance, which allows the identification of transformants based on their ability to grow in medium containing the corresponding antibiotic. In antibiotic-based screening, determination of the appropriate selection pressure is crucial for the establishment of an efficient screening system in which the growth of non-transformed cells is inhibited but the impact on transformed cells is minimized [6,7].

Commonly used screening antibiotics include hygromycin (Hyg), kanamycin (Kan), and glyphosate (PPT), which differ in their mechanisms of action. In poplar, different species, genotypes, tissues, and organs differ in their antibiotic sensitivity, such that in antibiotic screening the lethal concentration must be determined prior to the genetic transformation. In the transformation of American black 725 poplar leaves, a Hyg concentration of 2.5 mg/L was found to be suitable for the screening of transformants [8]. In Shanxin poplar and small black poplar, selection pressure for leaf differentiation was achieved with a Hyg concentration of 2.0 mg/L and 3.0 mg/L, respectively [9]. In giant poplar, leaf and stem segment differentiation were investigated using 4–6 mg Hyg/L [10]. A Hyg concentration of 1.5 mg/L was found to be suitable for transformation in Poplar 84K (Populus alba × P. glandulosa) [11], while a Kan concentration of 50 mg/L was applicable for transformation in Himalayan polar (Populus ciliata Wall.) [12]. In the leaves of Shanxin poplar, selection pressure was achieved using 30 mg Kan/L [13], whereas in the proposed Qing Shan Guan Yang hybrid poplar (Populus pseudo-cathayana × P. deltodides Bartr. cv. ‘Shan Hai’), adventitious shoots were induced in leaves using only 10 mg Kan/L [14]. Fewer studies have used PPT as a screening agent in poplar transgenics; instead, the activity of the herbicide has mostly been investigated in herbaceous plants. For example, PPT at a concentration of 1 mg/L was shown to inhibit completely both the growth of castor bean cotyledon nodes and bud differentiation [15]. In different genotypes of soybean, PPT induced indeterminate shoots, with the appropriate screening concentration determined to be 0.2–0.6 mg/L [16]. In 84K poplar shoots, the effective screening concentration of PPT was 5 mg/L [17].

Several studies have examined the appropriate critical concentration of different antibiotic species for poplar leaves, but little is known about the mechanism underlying the lethality of different antibiotics in poplar. Transcriptome sequencing has been used to investigate gene function, including the molecular composition of cells and tissues, in addition to specific biological processes and molecular mechanisms. With the continuous development of sequencing technology, transcriptome sequencing is increasingly being used in molecular research. Combined transcriptome and physiological studies have shed light on the multilevel regulation of gene expression processes and the molecular mechanisms behind many biological traits. In this study, the susceptibility to three antibiotics and the physiological response to antibiotic treatment were investigated to observe and record the status of leaves under different antibiotic gradients, and the key physiological indexes were measured using ‘Beilinxiongzhu 1’ leaf material. Based on transcriptome sequencing, we also analyzed differences in the annotation of differentially expressed genes (DEGs) in leaves under different antibiotic treatments, summarized the shared and specific genes under the three antibiotic treatments, analyzed differences in the expression of DEGs between the low- and high-concentration treatments, and identified hub-genes related to the screening of antibiotics using weighted gene co-expression network analysis (WGCNA) to provide insights into how antibiotics affect leaf sprouting in forest trees.

2. Materials and Methods

2.1. Test Materials

This study was conducted in the Plant Tissue Culture Laboratory of the College of Forestry, Hebei Agricultural University, using leaves of poplar ‘Beilinxiongzhu 1’ histocultured seedlings kept in the laboratory as experimental materials. Murashige and Skoog, 6-benzylaminopurine, naphthalene acetic acid, indole-3-butyric acid, hygromycin, kanamycin, and glyphosate were purchased from Beijing Coolaber Technology (Beijing, China). Sucrose was purchased from Tianjin Zhiyuan Chemical Reagent (Tianjin, China). Agar was purchased from Beijing Jintai Hongda Biotechnology (Beijing, China).

All three antibiotics were of analytical purity and stored in accordance with the correct practices. For use, the antibiotic powder was fully dissolved in sterile water to the appropriate concentration and filtrated through a sterile filter. It was sealed with a sealing film and stored at −20 °C.

Rooting medium consisted of 1/2 Murashige and Skoog (MS) + 0.4 mg indole-3-butyric acid (IBA)/L + 25 g sucrose/L + 6.5 g agar/L, and differentiation medium of MS + 0.4 mg/L 6-benzylaminopurine (6-BA) + 0.05 mg naphthaleneacetic acid (NAA)/L + 30 g sucrose/L + 7 g agar/L. The light cycle was 14 h light/10 h dark, and the temperature of the culture room was controlled at 25 °C.

2.2. Test Method

2.2.1. Determination of Hyg, Kan, and PPT Sensitivities

Leaves of rooting seedlings grown in rooting medium for 30 days were selected, choosing fully expanded 2nd–4th young leaves of the same size below the growth point of normal color; three or four incisions were made evenly along the main vein. The leaves were then inoculated with 0.4 mg 6BA/L +0.05 mg NAA/L +0.01 mg TDZ medium/L. After 9 days of culture in the absence of antibiotics, the TDZ was removed, and different concentrations of Hyg (0.5, 1.0, 1.5, 2.0, 2.5 mg/L), Kan (20, 40, 60, 80, 100 mg/L), or PPT (0.6, 0.8, 1.0, 1.2, 1.4 mg/L) were added to the cultures. Antibiotic-free cultures served as the control (CK). The antibiotic-containing and control media were changed every 10 days. Fifteen explants were included in each treatment, and the experiment was repeated three times. The growth status of each root seedling leaf was observed and recorded in detail. The germination rate was determined after 40 days of treatment.

According to differences in phenotypic traits, seven different antibiotic concentrations were further tested: CK, H1 (1.0 mg Hyg/L), H2 (2.5 mg Hyg/L), K1 (20 mg Kan/L), K2 (80 mg Kan/L), P1 (0.6 mg PPT/L), and P2 (1.2 mg PPT/L). The culture and leaf selection conditions were the same as described in Section 2.2.1. Physiological indexes were measured after 24 days of treatment. Each sample was weighed three times. Kits from Suzhou Greens Biotechnology were used to assay the soluble sugar (SS, anthrone colorimetry), soluble protein (SP, Coomassie brilliant blue method), and the contents of the endogenous hormones 6BA, zeatin (ZT), indoleacetic acid (IAA), and abscisic acid (ABA) in 0.5 g of leaves.

2.2.2. Transcriptome Sequencing

Leaves treated as described in Section 2.2.1 were weighted three times, and 0.3 g of each sample was wrapped in tin foil and frozen in liquid nitrogen. Transcriptome sequencing was completed by Beijing Nuohe Zhiyuan Technology (Beijing, China).

The significance of differences in gene expression was analyzed using DESeq2(version 1.10) software with the screening criteria |log2FC| ≥ 1 and p_adj ≤ 0. 05. The number of DEGs (up- and down-regulated) were counted for each comparison set, and GO function and KEGG pathway enrichment analyses were performed on the set of DEGs genes within each comparison set using clusterProfiler. UpSet plots were produced using the UpSet Plot module of TBtools to obtain the specific DEGs of different antibiotic types and concentrations in the treatment groups, and GO and KEGG enrichment analyses were performed on the relevant DEGs.

2.2.3. WGCNA Analysis

The DEGs generated by RNA-Seq were used as source data to construct a WGCNA based on the soft threshold value β = 9, and a clustering tree was constructed based on the correlation of the expression between genes. The modules were divided by the method of dynamic cut-tree (module size = 30, mergeCutHeight = 0.25), and the module eigenvalues were calculated using Epigengene; the correlations of trait data were illustrated using a correlation heatmap; and the correlation coefficients were calculated (gene significance) for all of the genes and traits within each module. To understand the details of the relevant modules, GO function and KEGG pathway enrichment analysis (p_adj < 0.05) were performed; the core genes were screened using relationships of the module network nodes, and the top 100 pairs of associated nodes with weight > 0.15 were screened. A co-expression network map among the genes was constructed using Cytoscape.

2.2.4. qRT-PCR Validation

The accuracy of the RNA-Seq data was verified by real-time fluorescent quantitative PCR (qRT-PCR). ACT7 served as the internal reference gene. Primer Premier 6.0 was used to design the primers, and the relative expression level was calculated using the 2−ΔΔCT method.

3. Results

3.1. Determination of the Screening Concentrations of the Three Antibiotics

Establishing the appropriate selection pressure is critical in the design of an efficient screening system to minimize damage to the transformed cells and to achieve a better screening effect by increasing the concentration of the screening agent in subsequent propagations. In this study, the leaves of the rooting seedlings were exposed to three different antibiotics (Hyg, Kan, and PPT), with non-antibiotic-exposed seedlings used as the control (CK).

After 40 days of culture (Figure 1 and Figure 2), the three antibiotics had significant inhibitory effects on the growth and development of the leaves. As seen in Figure 1, CK leaves were dark green, with adventitious buds, 0.8–1.5 cm long, forming clusters at the petioles and incision sites. In the Hyg-treated leaves, higher concentrations caused the browning of the leaves, a decrease in the regeneration rate of adventitious buds, and the eventual death of the leaves. At concentrations of 0.5–1.0 mg Hyg/L, the leaves were yellow–green, with a few thin, weak buds at the petiole and incision sites. Leaves treated with 1.5 mg Hyg/L lost their green color, and only a few buds developed; at Hyg concentrations ≥ 2.0 mg/L, no adventitious buds appeared at the incision sites, and the leaves turned brown and died. In the Kan treatments, the leaves gradually faded and rapidly lost their regeneration ability. At a concentration of 20 mg Kan/L, the leaves were unable to differentiate into adventitious buds. In response to PPT, the leaves turned brown to varying degrees and then died. Leaves treated with 0.6 mg PPT/L were etiolated and the petioles and incisions expanded, with the formation of a small number of yellow green adventitious buds with lengths of 0.3–0.8 cm. At a PPT concentration of 0.8 mg/L, the leaves were etiolated and died, the petioles and incisions expanded, and a few bud spots were seen. At ≥1.0 mg PPT/L, the regeneration of adventitious buds was completely inhibited and albino plants died.

3.2. Physiological Response of Poplar Leaves to Antibiotics

To explore the effects of the three antibiotics on osmolyte levels in the leaves, the soluble protein (SP) and soluble sugar (SS) contents were measured after 24 days of culture (Figure 3). The levels of the endogenous hormones (6BA and ZT), auxin (IAA), and abscisic acid (ABA), as important plant growth regulators, on day 24 were also measured. As seen in Figure 4, compared with CK, the leaf IAA content decreased significantly in each treatment (H1, H2, K1, K2, P1, P2). The decrease in the P1 treatment group was the smallest (35.8%). The 6-BA content in each treatment was lower than that in CK, with significant decreases in P1 and P2 by 47.1% and 76.7%, respectively. In the other treatments, the differences in the 6BA content were not significant. The ZT content was highest in CK leaves and was 2.4 times higher than in K1-treated leaves. The differences between the other Kan treatments and CK were not significant. The ABA content was significantly higher in P1 than in the other groups, among which the differences were not significant.

3.3. Transcriptome Analysis of the Antibiotic-Treated Leaves

3.3.1. Quality Control of the Transcriptome Data

Leaves of ‘Beilinxiongzhu 1’ poplar under the seven different antibiotic treatments (CK, H1, H2, K1, K2, P1, and P2) were subjected to transcriptome sequencing analysis, with three biological replicates in each group, resulting in the construction of 21 independent RNA-Seq libraries. The statistical results of the transcriptome sequencing data showed that the GC content in the raw data of the 21 samples ranged from 43.93% to 45.95%, with the GC contents of Q20 and Q30 > 96.84% and 92.01%, respectively, indicating the good quality of the sequencing data. The sequences of the samples were compared with the reference genome in the poplar 741 database; the comparison rate of each sample was >79%, such that the genome met the requirements allowing further analyses.

3.3.2. Intra- and Inter-Group Differentially Expressed Genes

The antibiotic treatments were compared with CK to identify differentially expressed genes (DEGs) using the DESeq2(version 1.10) software. The criterion for identifying DEGs was |log₂ (FoldChange)| ≥ 1 and p_adj ≤ 0.05. The results are shown in Figure 5A. In the H1-CK, H2-CK, and H1-H2 comparisons, 3811 DEGs (1751 up-regulated and 2060 down-regulated), 13,038 DEGs (5653 up-regulated and 7385 down-regulated), and 7001 DEGs (4279 up-regulated and 2722 down-regulated) were detected, respectively; in the K1-CK, K2-CK, and K1-K2 comparisons, 6688 DEGs (2415 up-regulated, 4273 down-regulated), 8180 DEGs (3258 up-regulated, 4922 down-regulated), and 666 DEGs (213 up-regulated, 453 down-regulated) were detected; and in the P1-CK, P2-CK, and P1-P2 comparisons, 5465 DEGs (3040 up-regulated, 2425 down-regulated), 8686 DEGs (682 up-regulated, 4004 down-regulated), and 787 DEGs (484 up-regulated, 303 down-regulated) were detected(Figure 5B,C).

3.3.3. GO Functional Enrichment Analysis of the DEGs in Group Comparisons

The GO database is used to identify the functions of genes and proteins. It is divided into three main categories: biological process (BP), cellular component (CC), and molecular function (MF), each of which describes the possible functions encoded by the genes, the cellular environment in which they are expressed, and the biological processes in which they participate. For the DEGs identified in the three antibiotic treatments, the top four pathways, ranked according to p_adj < 0.05, in each of the three GO categories in each treatment were determined (Supplementary Table S1).

As shown in Figure 5, in the low-concentration treatment groups (H1-CK, K1-CK, P1-CK), DEGs up-regulated under Hyg and Kan treatments were significantly enriched in cellular carbohydrate metabolic and disaccharide metabolic processes as well as flavin adenine dinucleotide binding. Endopeptidase inhibitor and peptidase inhibitor activities and related processes were significantly enriched under Hyg and PPT treatments. Common enriched pathways were also detected among the DEGs in the high-concentration Hyg and PPT groups (H2-CK, K2-CK, P2-CK), such as multi-organism processes, cell recognition, endopeptidase inhibitor and peptidase inhibitor activities, and related processes.

In the low-concentration treatment groups (H1-CK, K1-CK, P1-CK), the down-regulated DEGs of the Hyg and Kan treatments were annotated with defense response, DNA packaging complex, nucleosome, protein DNA complex, chromatin, chromosomal part, and chromosome, among others. In the Hyg and PPT groups, down-regulated DEGs were enriched in photosynthesis and thylakoid. In the Kan and PPT groups, only DEGs enriched in microtubule motor activity were noted.

In the high-concentration treatment groups (H2-CK, K2-CK, P2-CK), DEGs related to antioxidant activity were significantly enriched. In the Hyg and Kan groups, the DEGs were significantly enriched in microtubule-related processes, such as movement of cell or subcellular components, cytoskeletal protein binding, microtubule-based movement, and microtubule-based processes. DEGs under Hyg and PPT treatment were significantly enriched in photosynthesis, chloroplast part, thylakoid membrane, and other photosynthesis-related processes, similar to the low-concentration treatment group.

As shown in Figure 5, the antibiotic treatments also led to the enrichment of DEGs shared by the low- and high-concentration groups. Under Hyg and PPT treatment, the common pathways represented by the DEGs in the low- and high-concentration groups indicated similar, concentration-independent effects of the antibiotics on the leaves. In the Kan treatment, by contrast, there were almost no shared pathways. The GO pathway corresponding to antioxidant activity was significantly enriched only in the high-concentration Kan treatments.

3.3.4. KEGG Functional Enrichment Analysis of the DEGs in Group Comparisons

The annotation of the enriched DEGs in the GO analysis did not fully explain the molecular mechanisms of the different antibiotic treatments and the different concentrations. A KEGG enrichment was therefore performed; the significantly enriched pathways (p_adj < 0.05) are summarized in Supplementary Table S2.

As shown in Figure 6, in the low-concentration groups (H1-CK, K1-CK, P1-CK), up-regulated DEGs significantly enriched under Hyg and Kan treatments included those involved in valine, leucine, and isoleucine biosynthesis. Under Hyg and PPT treatments, a large number of amino acid metabolic processes were significantly enriched, such as alanine, aspartate, and glutamate metabolism, tyrosine metabolism, and amino sugar and nucleotide sugar metabolism. Tyrosine metabolism, MAPK signaling pathway-plant, ABC transporters, and other metabolic pathways were also enriched.

In the high-concentration treatment groups (H2-CK, K2-CK, P2-CK), glucosinolate biosynthesis was significantly enriched under the Hyg and Kan treatments, while phenylpropanoid biosynthesis, sesquiterpene and triterpene biosynthesis, and tryptophan metabolism were among the significantly enriched pathways in the Hyg and PPT treatments. As in the low-concentration treatments, the MAPK signaling pathway-plant and ABC transporters were also enriched in the high-concentration treatments.

In the low-concentration groups (H1-CK, K1-CK, P1-CK), down-regulated DEGs under Hyg and Kan treatments were significantly enriched in biosynthetic processes such as flavonoid, brassinosteroid, and brassinosteroid biosynthesis; motor proteins, phagosomes, and DNA replication were also enriched. DEGs significantly enriched in photosynthesis were detected only under Hyg and PPT treatments. In the Kan and PPT treatments, DEGs were significantly enriched in carbon fixation in photosynthetic organisms, photosynthesis-antenna proteins, and glyoxylate and dicarboxylate metabolism pathways.

In all three treatments in the high-concentration groups (H2-CK, K2-CK, P2-CK), photosynthesis-antenna proteins, carbon fixation in photosynthetic organisms, and brassinosteroid biosynthesis were significantly enriched. As in the low-concentration treatments, flavonoid biosynthesis, motor proteins, and phagosome pathways, but also starch and sucrose metabolism, fructose and mannose metabolism, and tropane, piperidine, and pyridine alkaloid biosynthesis were enriched under Hyg and Kan treatments. Photosynthesis and porphyrin metabolism pathways were significantly enriched under Hyg and PPT treatments, along with glyoxylate and dicarboxylic acid metabolism pathways, similar to the low-concentration groups.

3.3.5. Analysis of the DEGS across the Treatment Groups

Shared DEGs

The GO and KEGG enrichments of DEGs shared between groups are shown in Supplementary Tables S3 and S4. Ten BP pathways and five MF pathways were highly significantly enriched in the GO analysis of shared up-regulated DEGs under the three treatments (p_adj < 0.01). In the BP pathways, the DEGs were enriched in metabolic and biosynthetic processes such as disaccharide, glutamate, and dicarboxylic acid metabolic processes, the glutamate biosynthetic process, the dicarboxylic acid metabolic process, and the glutamate biosynthetic process. In the MF pathways, DEGs were enriched in enzyme activity-related pathways, such as sucrose synthase, glucosyltransferase, and glutamate synthase activities. Among the down-regulated DEGs shared among the six treatments, mangiferyl-3-dehydrogenase (NADP⁺) activity was the only identified MF pathway. In the KEGG enrichment, shared up-regulated DEGs were enriched in nitrogen metabolism, alanine, aspartate, and glutamate metabolism, and the pentose phosphate pathway. Shared down-regulated DEGs were enriched in glutathione metabolism, photosynthesis-antenna proteins, and brassinosteroid biosynthesis.

In the low-concentration treatment group, DEGs were enriched in fructose-bisphosphate aldolase, aldehyde-lyase, oxidoreductase, and other enzyme-activity-related GO pathways. In the high-concentration treatment group, among the up-regulated DEGs there were no co-enriched pathways, but the down-regulated DEGs were enriched in the polysaccharide catabolic process, beta-amylase activity, and amylase pathway. In the KEGG enrichment, up-regulated DEGs in the low-concentration treatment group were enriched in fructose and mannose metabolism pathways, the pentose phosphate pathway, and carbon fixation in photosynthetic organisms; the high-concentration treatment group was also enriched in fructose and mannose metabolism. Up-regulated DEGs in the high-concentration treatment group were enriched in the biosynthesis of various plant secondary metabolites and in exopolysaccharide biosynthesis. Down-regulated DEGs in the low- and high-concentration groups were not significantly enriched in any of the KEGG pathways.

Inter-Group Specific DEGS Analysis

The DEG enrichments determined in the GO pathways for specific groups are shown in Supplementary Table S3. A specific DEGs expression heat map showing the significant enrichment of KEGG pathways in each treatment group is presented in Figure 7. In the Hyg group, up-regulated specific DEGs were mainly enriched in endopeptidase and peptidase inhibitor activities, and down-regulated DEGs in pathways related to cell wall anabolism, such as cellulose metabolic and polysaccharide biosynthetic processes. Among the KEGG pathways, the up-regulated DEGs were enriched in nucleotide sugar biosynthesis and galactose metabolism, and thus in pathways that maintain cellular energy storage and support the cellular antioxidant response. Down-regulated DEGs were enriched in phagosomes, motor proteins, and the conversion of pentose and glucuronic acid, which accelerate cell movement and antagonize toxin accumulation in vivo.

Enrichment pathways were most abundant in the DEGs associated with Kan treatment. Up-regulated GO pathways were mainly chromatin assembly, regulation of kinase activity, and the cellular hormone metabolic process, pyrimidine nucleotide metabolic process, and other related processes. Down-regulated pathways mainly consisted of the cell wall macromolecule metabolic process, regulation of translation elongation, positive regulation of cellular protein metabolic processes, endopeptidase activity, and amino acid synthesis and metabolism processes. In the KEGG analysis, up-regulated DEGs were enriched in ribosome biosynthesis in eukaryotes, valine, leucine, and isoleucine biosynthesis, steroid biosynthesis, and several other amino acid metabolic pathways. Down-regulated DEGs were enriched in phenylalanine, tyrosine, and tryptophan biosynthesis, flavonoids biosynthesis, carbon fixation in photosynthetic organisms, and the biosynthesis of various plant secondary metabolites.

PPT up-regulated specific DEGs were enriched in aminoglycan metabolic, cell wall macromolecule catabolic, and amino acid metabolic processes. Down-regulated DEGs were significantly enriched in single-stranded DNA binding and iron sulfur cluster binding. In the KEGG enrichment, up-regulated specific DEGs in the PPT treatments were significantly enriched only in ABC transporters; there were no specific down-regulated DEGs significantly enriched in any pathway.

3.4. Gene Co-Expression Network Analysis

To investigate further the gene regulatory networks involved in the effects of the three antibiotics on the growth of ‘Beilinxiongzhu 1’ poplar leaves, a weighted gene co-expression network (WGCNA) of 21 samples was constructed. Using the DEGs generated by RNA-Seq as the source data, the threshold for the square of the correlation coefficient was set at 0.8. An optimal soft threshold β = 9 was used to construct the scale-free co-expression network (Figure 8A). A clustering tree was constructed based on the correlations of gene expression; modules were divided according to the dynamic cut tree method (module size = 30, mergeCutHeight = 0.25), based on a total of 31 modules (Figure 8B). A correlation analysis of module eigenvalues (Epigengene) and trait data revealed significant positive or negative correlations of the different modules with each physiological index (Figure 8C).

The correlation coefficients (gene significance) of all genes within each module with SS, SP, and IAA were calculated and box line plots were drawn (Figure 8D–F). The selection of modules with a mean GS ≥ 0.4 identified the modules red (2457 genes, r = 0.63, p < 0.01), dark orange (47 genes, r = 0.52, p < 0.05), blue (4255 genes, r = 0.7, p < 0.001), saddle brown (36 genes, r = 0.67, p < 0.001), and turquoise (8542 genes, r = −0.62, p < 0.01) as significantly associated with SS; the modules royal blue (104 genes, r = 0.75, p < 0.001), salmon (266 genes, r = 0.66, p < 0.01), and green (3242 genes, r = 0.63, p < 0.01) as significantly positively correlated with SP; and the modules royal blue (104 genes, r = 0.69, p < 0.001), red (2457 genes, r = 0.76, p < 0.001), magenta (805 genes, r = 0.64, p < 0.01), and salmon (266 genes, r = 0.78, p < 0.001) as significantly and positively correlated with the growth hormone IAA.

Analysis of Central Genes Related to Physiological Indexes

GO function and KEGG pathway enrichment analyses were performed on the modules. GO function enrichment of the magenta module revealed biological processes, cellular amino acid metabolism, steroid metabolism, and molecular functions involved in intramolecular lyase, steroid dehydrogenase, and amine hydrolase activities. The KEGG analysis showed that 11 pathways were significantly enriched, including flavonoid biosynthesis, phenylalanine, tyrosine, and tryptophan biosynthesis, pyruvate metabolism, and glutathione metabolism (Figure 9A).

Genes with high connectivity obtained in the magenta module included P_x_tomentosa39748 (bifunctional dihydroflavonol 4-reductase/flavanone 4-reductase), P_x_tomentosa51340 (anthocyanin reductase), P_x_tomentosa54276, P_x_tomentosa57673 (CHS1 chalcone synthase), the GO annotation involving catalytic activity, the KEGG annotation involving flavonoid biosynthesis, and P_x_tomentosa67474 (oxidase) in the GO annotation involving redox processes (Figure 9A).

In the magenta module, genes were positively correlated with CK, P1, and P2, and negatively correlated with H1, H2, K1, and K2 (Figure 9B). Module genes in the CK and PPT treatment groups were highly expressed, indicating their involvement in auxin synthesis.

Relevant central genes in SS and SP accumulation were analyzed (Supplementary Table S5). The red and blue modules correlated significantly with SS accumulation. Genes with high connectivity in the red module included P_x_tomentosa21724 (trxh2 thioredoxin), involved in cell redox balance in the GO annotations, P_x_tomentosa64914 (NADH-ubiquinone oxidoreductase complex I), P_x_tomentosa18833 (H3 histone 2), and P_x_tomentosa22071 (H2B histone) in the GO annotations related to cellular components, including nucleosomes, chromosomes, protein-DNA complexes, and organelles. Genes with high connectivity in the blue module included P_x_tomentosa06896 (GST glutathione S-transferase), P_x_tomentosa42907, P_x_tomentosa23019 (cadmium resistance in PCR plants), P_x_tomentosa40306 (RTL3 ribonuclease), and P_x_tomentosa36432 (GLIP1/GDSL esterase/lipase), related to the hydrolysis of ester bonds according to the GO annotations.

SP accumulation correlated positively with the royal blue, salmon, and green modules. Genes with high connectivity in the salmon module included P_x_tomentosa50106 (SILD, lignin biosynthesis), P_x_tomentosa00090 (CYC2 (S)-8-oxo limonenol synthase), related to catalytic activity and coenzyme binding in the GO annotations, P_x_tomentosa61781 and P_x_tomentosa63953 (C92C6 trimethyltridecatetraenyl synthase, cytochrome), related to ferrous iron binding, oxidoreductase activity, hematoxylin binding, tetrapyrrole binding, and others in the GO annotations. Genes with high connectivity in the green module included P_x_tomentosa31463 (ribosomal protein, chloroplast) and P_x_tomentosa49360 (ribosomal protein, chloroplast), involved in biological processes, such as translation, peptide metabolism processes, nitrogen and cytosolic amide metabolism processes, and organic matter biosynthesis processes in the GO annotations, and those related to molecular functions, structural components of ribosomes, P_x_tomentosa28030 (ribosomal protein, chloroplast), genes involved in nucleic acid and organic ring compound binding in the GO annotation, and the mRNA monitoring pathway in the KEGG annotation.

3.5. qRT-PCR Validation Results

To verify the accuracy of the RNA-Seq data, six DEGs were selected from the comparison groups and verified by qRT-PCR (Figure 10). Primer information is provided in Table S6. The qRT-PCR results showed that the expression trends of the six DEGs were consistent with the RNA-Seq data, indicating that the transcriptome sequencing results were reliable.

4. Discussion

The screening of transgenic plants is an important part of plant genetic transformation. The same plant can react differently to different screening agents, and different plants to the same screening agents. In this study, the antibiotics Hyg and Kan and the herbicide PPT were tested for their effects on the growth of ‘Beilinxiongzhu 1’ poplar leaves. The greatest damage to the leaves was induced by PPT, which caused obvious fading as well as whitish spots by day 7 (high concentration) or day 20 (low concentration). Buds appeared at the main vein incision at day 20 in the low-concentration group, whereas browning and death were observed in the high-concentration group. Hyg caused less damage to the leaves than PPT, with less yellowing during 14 days and a significant increase in the number of regenerated adventitious buds in response to a low concentration of the antibiotic. Kan had more moderate effects on the leaves than either PPT or Hyg, and it produced reliable screening results at low concentrations. Although the differences among Kan treatments were not obvious, the leaves were unable to form adventitious buds under low concentrations. These concentrations, therefore, effectively limited the growth of non-transformed cells, with fewer effects on the leaves, thus allowing the later emergence of buds. In genetic transformations, the Kan dose can therefore be reduced to 15 mg/L. In fact, low-toxicity screening agents are more effective than those with high toxicity in genetic transformation, as the rapid death of non-transformed cells has inhibitory effects on neighboring cells, thus potentially preventing the growth of transformants [18]. The much lower toxicity of Kan than of Hyg or PPT was consistent with the findings of Wang Dan [9] and Jiao Ziyuan [19]. In developing a protocol for cell transformation, low concentrations of screening agents should be used first, during the early stage of screening, followed by high concentrations during the induction of adventitious buds and to obtain regenerated plants [10]. In addition, multiple screenings should include succession and strengthening of the plants and then seedling formation, with molecular confirmation of the resistant plants at the later stage to ensure the reliability of the screening results [20].

Plant tolerance depends on the antibiotic type, exposure time, and concentration [21]. In this study, based on phenotypic trait differences, the physiological indexes of leaves with different exposure times were determined. The soluble sugar content of the Hyg-treated group was significantly higher than that of the CK group at 16 d of incubation, indicating damage to the cell membrane and a restoration of osmotic balance. At 24 days of culture, the SP content was significantly lower in the high thiomycin concentration group than in the CK group, suggesting that, beyond a certain threshold of antibiotic stress, the plant antioxidant system cannot effectively remove the drug, such that the cells become irreversibly damaged, consistent with the results of Zhang Lina [22]. The physiological indexes in the Kan-treated leaves were significantly increased by day 8 of exposure, which indicated that antibiotic stress, including membrane damage, occurred earlier than in response to Hyg. Low concentrations of PPT induced stress at day 16 of the exposure, and high concentrations at day 8.

Plants under stress release various types of endogenous hormones to activate the physiological and biochemical processes that allow them to adapt to the adverse environment [21,23]. In this study, the contents of 6-BA, ZT, and IAA in each antibiotic treatment were significantly lower than in the control, while the contents of IAA and ABA in the low concentration of PPT (P1) treatment were significantly higher than in any of the other treatment groups. These results indicated that antibiotic treatment inhibited poplar leaf growth and development, with the exception of a low concentration of PPT, in response to which plants were able to produce ABA and thus exhibit resistance to the drug.

Several genes related to the response to antibiotic stress were identified by transcriptome analysis of ‘Beilinxiongzhu 1’poplar leaves. Comparisons of each treatment with the control revealed many DEGs. Their further investigation in GO function and KEGG pathway enrichment analyses revealed significant differences in the biological functions induced by the different antibiotics.

Up-regulated DEGs in the Hyg-treated group were significantly enriched in processes related to fatty acid oxidation and endopeptidase activity. In Hyg-treated leaves, the chloroplasts were poisoned, producing toxic by-products, such as ROS, and fatty acid oxidation was induced, a sign of cellular stress [24], which in turn affected ATP production and thus cellular respiration [25,26]. Endopeptidase is an intracellular protein with hydrolase activity that breaks down long protein chains into shorter peptide fragments [27]. Hyg inhibits peptide chain lengthening by competing with lengthening factor EF-2 for ribosome binding sites in chloroplasts and mitochondria, thus disrupting translation and causing the browning and death of sensitive plant tissues [28]. Down-regulated DEGs were significantly enriched in cellulose-related processes, such as cellulose metabolism and cellulose synthase activity, such that the synthesis of cellulose, a basic component of the plant cell wall, was blocked by Hyg treatment, which was also responsible for the death of the leaves. Hyg-specific DEGs were significantly enriched in the interconversion of pentose and glucuronic acid, nucleotide sugar biosynthesis, purine metabolism, and linoleic acid metabolism. In addition, DEGs in the nucleotide sugar biosynthesis pathway were up-regulated, including GFPT2, HXK1, PGMC, MIOX2, GALE2, UXS3, and AGLUP, indicating disruptions of cellular energy stores and antioxidant responses.

Histones are crucial for chromatin assembly, the regulation of protein phosphorylation, and the regulation of kinase activity. Nuclear DNA is packaged with histone proteins to form the nucleosomes that make up chromatin. A reduced nucleosome occupancy of many defense-related gene promoters triggers a pattern of gene expression that may significantly improve plant fitness by promoting plant defenses [29,30]. The regulation of both protein phosphorylation and kinase activity is significantly associated with a plant‘s cellular stress response and leaf senescence [31,32]. In this study, the specifically down-regulated DEGs were significantly enriched in cellular amino acid biosynthesis processes, cell wall macromolecule catabolism metabolism, and the regulation of translational elongation, suggesting that protein synthesis in leaf cells under Kan treatment is blocked, leading to cell wall catabolism. The up-regulated DEGs were those involved in ribosome biosynthesis and purine metabolism processes in eukaryotes, while the down-regulated DEGs were related to phenylalanine, tyrosine, and tryptophan biosynthesis, flavonoids biosynthesis, and carbon fixation in photosynthesis organisms. These processes are significantly related to transcription and protein synthesis as well as the protection of plants against stress.

The PPT-specific up-regulated DEGs were enriched in amino acid synthesis processes, including cell wall macromolecule catabolism, drug catabolism, ATPase activity, and sulfur-containing amino acid biosynthesis. This indicated a positive response to the herbicide that included enhanced cell wall synthesis and drug catabolism as well as increased ATP and amino acid synthesis, thus stabilizing the leaf cells. In addition, among the significantly up-regulated DEGs were ABC transporters, which are essential for the plant immune system. ABC transporters are a class of transmembrane proteins that play dominant roles in cellular secretion and vesicles, which not only protect the plant from exogenous toxic substances but are also indispensable for normal plant development and environmental exposure [33,34,35]. With PPT treatment of the leaves, ABC transporters have an essential part in the plant immune system via the transmembrane transport of toxic substances. The ABC transporter is thought to be involved in a variety of physiological processes that enable plants to adapt to changing environments and cope with biotic and abiotic stresses. Down-regulated DEGs also included those involved in thiamine and vitamin metabolism processes. Thiamin (vitamin B1) is a cofactor for enzymes involved in major metabolic pathways and an environmental stress protector [36,37]. A reduction in vitamin B1 metabolism under PPT treatment can hinder normal cellular survival processes.

Shared DEGs detected in all antibiotic treatments were those involved in glutathione (GSH) metabolism, such as glutamate synthase, glucosyltransferase activity, and oxidoreductase activity. The amino acid derivative GSH, composed of glutamic acid, cysteine, and glycine, is an important antioxidant in plants, participating directly or indirectly in the scavenging of reactive oxygen species. It is also a metabolic substance involved in plant metabolic responses and the signaling processes required for plant defense [38,39]. Expression of the four genes encoding GST enzymes was down-regulated in all three antibiotic treatments, which would decrease leaf antioxidant capacity and thereby damage biomolecules in the cell.

In a gene co-expression network (WGCNA), genes in the same module may be related functionally. In this study, the physiological index data were correlated with the modules to identify the expression modules most strongly related to SS, SP, IAA, and ABA contents. Among them, the central genes of the red module (TRXH2, H3.2, and H2B) were highly expressed in CK. Previous studies have shown that overexpression of the AtTRXh2 gene increased cell tolerance to oxidative and salt stress in kale-type oilseed rape [40] and that the plant histone H3 gene responded to high salt stress [41,42].

Among the blue module hub genes (GSTX3, PCR2, RTL3, GLIP1), GST genes have detoxifying and antioxidant functions; AtGSTU17 (Arabidopsis thaliana) [43], ZmGST (maize) [44], OsGSTU3 (rice) [45], and ThGSTZ1 (willow Tamarix hispida) [46] genes play crucial roles in plant stress responses; CpGLIP1 overexpression has positive effects in Arabidopsis and waxberry (Chimonanthus praecox) against high temperature, drought, and low temperature stress, in addition to reducing the rate of water loss in transgenic poplar [47]. Central genes in the salmon (SILD, CYC2, C92C6, and U71K1) and magenta (DFRA, ANRPN, CHS1, and FL3H) modules were highly expressed in the CK and PPT treatments, and green module hub genes (RK17, RK34, and PSRP2) in the CK and Kan treatments. These modules correlated significantly and positively with SP and IAA, indicating their involvement in regulation and, together with IAA, in the response to antibiotic stress. Chalcone synthase (CHS), the rate-limiting enzyme of the flavonoid biosynthesis pathway, is widely present in higher plants [48] and catalyzes the generation of chalcone in the anthocyanin synthesis pathway; other proteins in this family catalyze the synthesis of a number of products that mediate plant stress tolerance. Changes in the expression of CHS genes in different plants alter floral color and affect stress tolerance [49]. ABA correlated positively with the purple module, and genes in this module were highly expressed in the P1 treatment, suggesting their role in the synthesis of ABA in improving plant resistance, consistent with the results of the endogenous hormone assay.

5. Conclusions

In antibiotic susceptibility tests of poplar leaves of ‘Beilinxiongzhu 1’, the tolerated concentrations of Hyg, Kan, and PPT were 2.0 mg/L, 20 mg/L, and 1.0 mg/L, respectively. In leaves in the Hyg group, osmotic balance was maintained by increasing the SS content at day 16 of the incubation, but the cells were irreversibly damaged after 24 days of culture. A response to the antibiotic stress induced by Kan treatment was observed on exposure day 8. Leaves exposed to high and low concentrations of PPT responded to the stress on days 8 and 16, respectively. Under the three antibiotic treatments, DEGs were co-enriched in GSH metabolism, which is involved in antioxidant activities and thus in plant defenses. In the low-concentration groups, the DEGs were those mainly responsible for fructose and mannose metabolism as well as carbon fixation in photosynthetic organisms, all of which would enhance cellular energy storage and cellular resistance. The DEGs expressed in the high-concentration groups were mainly involved in the biosynthesis of plant secondary metabolites, exopolysaccharide biosynthesis, and, thus, in defenses against external stress. DEGs in the Hyg-treated group maintained cellular energy storage and supported cellular antioxidant responses by participating in cell wall synthesis and sugar metabolism pathways. Kan initiated the expression of relevant defense genes by initiating the packaging of nuclear DNA with histones to form nucleosomes. DEGs in the PPT-treated group were enriched in ABC transporter proteins, which help plants to exclude drugs and toxins. The WGCNA analysis of specific gene modules showed significant correlations with SS, SP, and growth factors. Pivotal genes of the specific modules, such as TRXH2, H3.2, H2B, GST, U71K1, and CHS, mediated the antibiotic stress response by producing osmotic regulators, growth hormones, and abscisic acid.