Phenotypic characterization
I. lactea var. chinensis is a drought tolerant plant and can grow in desert steppe and saline lowland meadows. To validate this weed’s drought tolerance, we performed a native water deficiency stress and rehydration experiment. The results showed that the wilting rate of I. lactea var. chinensis plants was approximately 75% after 7 days of native water deficiency stress, and when rehydrated, all wilted plants recovered after 3 days, indicating that I. lactea var. chinensis possesses a strong phenotypic plasticity of drought stress (Fig. 1). The RWC of water deficiency-treated plants significantly decreased compared to the control (CK). Additionally, the plant height of water deficiency-treated plants was significantly shorter than the control. When water deficiency-treated plants were rehydrated, RWC and plant height partially recovered to some extent compared to the control (Fig. 2). To further investigate the phenotypes of plants treated and not treated with water deficiency stress, we measured the root and shoot length. The results showed that the root length and root/shoot ratio of the water deficiency stress plants significantly increased compared to the control, similar to the other plants under water deficiency treatment [26].
Primary transcriptome analysis
Three sample groups (CK, T, and R) with three biological repeats were employed to perform RNA-sequencing analysis using the Illumina HiSeq platform. A total of 57.85 Gb of data was obtained, and the three sample groups CK, T, and R, contained 42,979,607, 42,693,515, and 42,884,268 raw reads, respectively. Then, 42,905,695, 42,668,270, and 42,857,576 clean reads were obtained in the CK, T, and R groups after filtering low-quality reads, respectively. Following this, de novo transcriptome assembly with Trinity generated 126,979 unigenes, with an N50 of 1,176 bp. These unigenes contained an average length of 810 bp and an average GC content of 42.55% (Table S1), indicating that these unigenes were fine quality and suitable for further annotation. Annotations of the assembled unigenes were carried out according to six public databases including NR, SwissProt, KEGG, KOG, Pfam, and GO. A total of 41,360, 31,799, 18,512, 8,875, 28,920 and 28,385 unigenes were aligned, respectively (Table 1). Finally, 44,247 unigenes (34.85%) were successfully annotated in six databases.
Validity analysis of transcriptome data by qRT-PCR
To validate the reliability of the transcriptome data, eight candidate genes, including ERF053, HSPA1_8, ADC1, CDPK, IAA17, APRR5, MYBP, and ABCC2, were arbitrarily selected for qRT-PCR analysis. The results showed that the eight candidate genes’ expression levels in three samples with water deficiency stress, rehydration or well-watering treatment were similar to the RNA-Seq data (Fig. 3). Together, these data showed that the RNA-seq analysis was reliable in the present study.
Detection of DEGs
To evaluate I. lactea var. chinensis’s t drought tolerance, differentially expressed genes (DEGs) in three samples were shown in scatter plots (Fig. 4a). A total of 1,187, 275, and 865 DEGs according to |log2FC| >= 1 and P-adjust<0.05 were detected in comparison groups T/CK, R/T, and R/CK, respectively as shown in Fig. 4a and Table S2~S4. The up-regulated DEG numbers in the three comparison groups of T/CK, R/T, and R/CK, were 481, 185, and 402, respectively, and the corresponding down-regulated DEG numbers were 706, 90, and 463. In the T/CK comparison group, there were more down-regulated DEGs than up-regulated genes, indicating that water deficiency stress globally inhibits gene expression (Fig. 4a, Table S2).
Venn diagrams showing the number of up-regulated DEGs and down-regulated DEGs in each comparison (three ways) is shown in Fig. S1. Interestingly, 73 genes were found to be common between T/CK and R/T, but not R/CK, indicating that these DEGs in rehydration treatment were virtually same as those in the control group (P-adjust > 0.05). Among the 73 DEGs, 29 DEGs were up-regulated in T/CK and down-regulated in R/T (Fig. S1c, Table S5), while 44 DEGs were down-regulated in T/CK and up-regulated in R/T (Fig. S1g, Table S6). These data showed that the 73 DEGs may be have highly plasticity in plant water deficiency stress and rehydration treatment. For example, the gene related to primary-amine oxidase (PAO; TRINITY_DN60247_c0_g2) expression was up-regulated by 32.7 times in response to water deficiency stress and down-regulated to the control level after rehydration. Similarly, the gene related to germacrene D synthase (GDS; TRINITY_DN45420_c1_g3) expression increased 655.6 times after water deficiency treatment, then decreased after rewatering. By contrast, two UGT85A (TRINITY_DN44873_c0_g4 and TRINITY_DN53280_c0_g1) expression were down-regulated under the water deficiency stress, but up-regulated during the water recovery period. Additionally, two-way Venn diagrams illustrate the overlap between the DEGs identified following well water, drought, and water recovery treatments as shown in Fig. 4b, being similar to the report of Zhang et al. [27].
GO functional and pathway enrichment analysis of DEGs
To elucidate the potential mechanism of I. lactea var. chinensis’s t drought tolerance, DEGs were mapped against the GO database and subjected to enrichment analysis, which was classified into three major functional categories based on the criteria of P-adjust<0.05. In the comparison of water deficiency treatment with and the control, 42 GO terms were significantly enriched in biological process (BP) category. For example, the DEGs involved in the amino acid biosynthetic and metabolic process included the ‘glutamine family amino acid catabolic process’, ‘Proline catabolic process’, ‘amide biosynthetic process’, and ‘cellular amide metabolic process’; those associated with recognition included ‘cell recognition’ and the ‘recognition of pollen’; those concerning the organic substance metabolic process included the ‘nitrogen compound metabolic process’, ‘cellular aromatic compound metabolic process’, and ‘heterocyclic metabolic process’; and the DEGs related to the biosynthetic and metabolic process of nucleic acids and proteins included the ‘DNA metabolic process’, ‘RNA biosynthetic process’, ‘peptide biosynthetic process’, and ‘cellular protein metabolic process’. Additionally, the expected DEGs related to the response to various other types of abiotic stresses were detected in the ‘cellular response to osmotic stress’, ‘cellular response to salt stress’, and ‘cellular response to blue light’ (Fig. 5; Table S7~S9). For the cellular component (CC) category, only 15 GO terms were significantly enriched in the comparison of drought stress and the control, which were mainly related to the components of membrane cytoplasmic part and organelle part (Fig. 5). For the molecular function (MF) category, lots of proteinase activity types, including Proline dehydrogenase activity, protein kinase activity and transferase activity were highly enriched (Fig. 5). We found that only four and eight GO terms were significantly enriched in the R/T (Table S8) and R/CK group (Table S9), respectively.
To reveal the underlying mechanism of I. lactea var. chinensis’s tolerance to drought stress, pathway enrichment analyses of the DEGs based on the Kyoto Encyclopedia of Genes and Genomes (KEGG) database were performed. In the T/CK group, 64 DEGs were significantly enriched in six pathways (P-adjust < 0.05), including ‘Plant-pathogen interaction’, ‘alpha-Linolenic acid metabolism’, ‘Circadian rhythm - plant’, ‘ABC transporters’, ‘Arginine and Proline metabolism’, and so on (Table 2). Some of the 64 DEGs play an important role in plant drought tolerance [28-30]. For example, receptor-like kinases encoding genes BAK1/SERK1 (TRINITY_DN60727_c4_g2), hydroperoxide dehydratase encoding genes AOS (TRINITY_DN51813_c4_g1), and pseudo-response regulator encoding genes PRR5 (TRINITY_DN56996_c1_g3, TRINITY_DN62580_c5_g2, TRINITY_DN51806_c4_g2, and TRINITY_DN56996_c1_g1) were up-regulated in T/CK comparison group. The whole report for the DEGs in each comparison group, respectively, is shown in Table S10~S12.
TFs, transporter proteins (TPs), and ROS scavenging systems affect the plant response to drought stress
TFs In our database, at least 43 DEGs encoding TFs were found (Table S13), and were included in the heatmap (Fig. 6). The concerning TFs were divided into 14 subfamilies, including APETALA2/Ethylene-responsive transcription factor (AP2/ ERF), MYB, WRKY, Zinc finger proteins, NAC, growth-regulating factors (GRF), etc. As shown in Fig. 6, most DEGs coding TF unigenes exhibited a down-regulated expression in plants treated with water deficiency stress compared to the control, consistent with other DEGs aforementioned (Fig. 4 and Fig. 6). However, some DEGs coding TFs exhibited up-regulated expression in water deficiency treatment, e.g. two related genes encoding AP2/ERF (TRINITY_DN59365_c4_g7, TRINITY_DN48541_c0_g3), six related genes encoding MYB (TRINITY_DN58225_c1_g1, TRINITY_DN63921_c6_g1, TRINITY_DN56464_c2_g2, TRINITY_DN57679_c2_g2, TRINITY_DN43447_c3_g3, TRINITY_DN44333_c4_g3) and four related genes encoding Zinc finger (TRINITY_DN59156_c3_g4, TRINITY_DN64408_c4_g1, TRINITY_DN44819_c3_g3, TRINITY_DN57150_c6_g1) were up-regulated expression significantly in water deficiency treatment. In the R/CK group, most genes’ expression levels were similar to those in the T/CK group, suggesting that rehydration might affect a small amount of gene expression, which was validated by the gene expression level in R/T (Fig. 6).
TPs Forty-six DEGs encoding transporters were detected in the RNA-seq data, including ABC, Nitrate, Hexose, Polyamine, and so on (Fig. 7; Table S14). In T/CK and R/CK groups, most DEGs were down-regulated, similar to the DEGs related to TFs. While there were some up-regulated expression DEGs under water deficiency stress e.g. two genes related bidirectional sugar transporter (TRINITY_DN50598_c1_g2, TRINITY_DN44794_c7_g1), one gene related Potassium transporter (TRINITY_DN45507_c4_g1), one gene related Metal-nicotianamine transporter (TRINITY_DN46504_c5_g1), one gene related Polyol transporter (TRINITY_DN52400_c0_g1), etc. However, rehydration can affect the remarkable change of these DEGs in terms of their gene expression, as shown in R/T (Fig. 7).
ROS scavenging system In only the T/CK comparison group, there were 22 DEGs encoding ROS scavenging enzymes, categorized as the glutathione peroxidase (GPX) pathway, catalase (CAT) pathway, Peroxidase (POD) pathway, NADPH oxidoreductase pathway and water-water cycle (Fig. 8; Table S15). Among them, the GPX pathway contains the most DEGs in the ROS scavenging system, including glutathione S-transferases (GSTs) and glutaredoxin (GRX). Under water deficiency stress, one GRX (TRINITY_DN47416_c1_g1) and four GSTs (TRINITY_DN53475_c0_g3, TRINITY_DN52312_c2_g1, TRINITY_DN62418_c3_g3, and TRINITY_DN53475_c0_g2) had a significantly down-regulated expression, only two GSTs (TRINITY_DN67517_c0_g1 and TRINITY_DN47288_c0_g1) were significantly up-regulated. In the NADPH oxidoreductase pathway, most DEGs coding three key enzymes, viz., thioredoxin (TRX), NAD(P)H-quinine oxidoreductase, and peroxiredoxin (PrxR), were down-regulated (Fig. 8). Additionally, two SODs (TRINITY_DN58115_c0_g1, TRINITY_DN57370_c0_g2) were significantly down-regulated expression under water deficiency stress. In addition, these DEGs were associated with two PODs, two PEXs, one CAT, and one APX, which possibly participated in plant response to drought stress (Fig. 8). After rehydration, the expression level of 22 DEGs recovered to a certain extent compared to the control. To verify in situ the accumulation of H2O2 and O2·- under water deficiency treatment, histochemical with diaminobenzidine (DAB) and nitroblue tetrazolium (NBT) to quantization the ROS activity was performed. After water deficiency-treated, dark blue spots (stained with DAB, Fig. 9a) and brown spots (stained with NBT, Fig. 9b) were deposited the I. lactea var. chinensis leaves. And after rehydration-treated, the deposited of dark blue spots (stained with DAB, Fig. 9a) and brown spots (stained with NBT, Fig. 9b) turned to lighten. The histochemical experiment further demonstrate that ROS accumulation was involved in drought stress of I. lactea var. Chinensis.
Proline metabolism Proline dehydrogenase (ProDH), a mitochondrial enzyme, is a key enzyme in the first step of the Proline metabolic process (Fig. 10a). Under water deficiency stress, ProDH in I. lactea var. chinensis plants had a down-regulated expression. When rehydrated, the ProDH expression level increased compared to water deficiency treatment (Fig. 10a; Table S16). The genes located downstream of the Proline metabolism process, including ADC, SpeE, and APX, exhibited a down-regulated expression under drought stress (Fig. 10a; Table S16). The free Proline concentration determination result was as consistent as the transcriptome analysis. After water deficiency-treated, The I. lactea var. chinensis plants’ free Proline concentration sharply increased and would decrease after rehydration-treated (Fig. 10b). These data suggested that Proline can be accumulated under drought stress, consistent with previous reports of water deficiency-treated plants.