Understanding physiological and molecular mechanisms of Populus deltoides ‘DanHongYang’ tolerance to waterlogging by comparative transcriptome analysis

Populus deltoides ‘DanHongYang’ (DHY) was identified as a waterlogging-resistant cultivar in our previous study. Here, the phenotype, physiological features and transcriptome profiling of P. deltoides ‘DHY’ between the treatments of waterlogging and control were compared. Waterlogging treatment led to distinctly formation of adventitious roots from P. deltoides ‘DHY’ stems. The activities of ascorbate peroxidase and glutathione reductase significantly increased in the leaves of P. deltoides ‘DHY’ by waterlogging treatment. Comparative transcriptomic analysis showed that 2,447 and 9,465 differentially expressed genes (DEGs) were screened between the leaves and roots of P. deltoides ‘DHY’ under waterlogging and control, respectively. The KEGG analysis showed the most significantly up-regulated DEGs in the leaves and roots were enriched to the pathways of glycolyis and proline synthesis. Some genes involved in stress response, endogenous hormones, antioxidant system and adventitious root development in the waterlogged were identified to contribute to the waterlogging tolerance of P. deltoides ‘DHY’. In addition, some candidate transcription factors such as RAP, NAC, WRKY, and bHLH were also found to be associated with the waterlogging tolerance of P. deltoides ‘DHY’. These findings provided the insights into the physiological and molecular mechanisms underlying the tolerance of P. deltoides ‘DHY’ to waterlogging stresses.


Introduction
Waterlogging stress has a significant negative effect on plant growth. Excessive soil moisture can disrupt the water balance of plants and significantly affect the morphology and metabolism of plants, restricting their growth (Irfan et al., 2010). Waterlogging puts severe stress on plant growth because excess water in the DPLC3, waterlogging treat: DPLF1, DPLF2, DPLF3)and roots (control: DPRC1, DPRC2, DPRC3, waterlogging treat: DPRF1, DPRF2, DPRF3)samples used for sequencing with three biological replicates per treatment.

Physiological index measurement
After 65 days of submergence treatment, the roots, stems, and leaves of the plants were separately harvested and their fresh biomass were measured. Chlorophyll content were determined according to the method of Wellburn (1994). Soluble sugar content was determined using anthrone colorimetric method (Moustakas et al., 2011). The MDA content was extracted using thiobarbituric acid (TBA) reagent and boiled at 100 °C for 20 min as previously described by . Proline content was estimated using described method with L-proline as standard (Yang et al., 2011). The activities of antioxidant enzymes (T-SOD, CuZn-SOD, POD, CAT, APX, soluble-Protein and GR) were determined according to the instructions of corresponding kits (NanjingJiancheng Biotech Co., Nanjing, China).
The photosynthesis parameters of poplar leaves were determined from 09:00 to 12:00 hr (solar time) using a portable open system infrared gas analyzer (LI-6400, LI-COR Inc., Lincoln, NE, USA). The chlorophyll fluorescence parameters were determined using a fluorometer with a modulated pulse (MINI-PAM) (Walz, Effeltrich, Germany) on the same leaflets used for the leaf gas exchange parameters measurement. Determination parameters include the maximum emitted fluorescence (Fm). The maximum fluorescence (Fm'), the maximum photosystem II (PSII) photochemical efficiency (Fv/Fm), the electron transport rate, the quantum yield of electron transport (Yield) and the photochemical (qP). Before measuring, the leaf darkadapted for 30 min, Specific operating steps are completed according to the method of Nascimento (Nascimento et al., 2018).

Library construction and transcriptome sequencing
The total RNA of leaves and roots was extracted using a TAKARA MIniBEST Plant RNA Extraction Kit in accordance with the manufacturer s instructions. RNA samples were tested for degradation and impurities by using 1% agarose electrophoresis. Sample purity was measured using a NanoDrop 2000 microspectrophotometer (IMPLEN, CA, USA), and the integrity and concentration of the RNA sample were detected using an Agilent 2100 RNA Nano 6000 Assay Kit (Agilent Technologies, CA, USA). Sequencing libraries were obtained using the NEBNext1Ultra RNA Library Prep Kit for Illumina* (NEB, USA) in accordance with the manufacturer s instructions and the construction of 12 cDNA libraries were completed. The concentration and quality of these libraries (12 cDNA) were tested using Agilent 2100 and Qubit 2.0, respectively. The Illumina HiSeq 2500 sequencing platform (Baimaike, Beijing) was used to sequence 12 cDNA libraries from the control / waterlogging leaves and roots. The original sequence data was obtained by Illumina HiSeq sequencing. High-quality reads (clean reads) were obtained by removing low-quality reads, joint-pollution reads and reads with N ratios greater than 5%. HISAT v2.0.4 software with default parameters was used to map clean reads filtered from original Reads to the Populus trichocarpa reference genome database (Tuskan Grabherr et al., 2006). All clean reads of the 12 libraries were obtained by assembling full-length transcriptional sequences using Trinity software and the DE Bruijn method (Grabherr et al., 2011).
Transcriptome sequences are annotated with trinotate, and functional annotations are mainly performed using databases, including PFAM, Nr, Swissprot, GO, COG and KEGG annotations. DEG analysis RPKM value were used to represent the expression abundance of reads corresponding to Genes. In this study, we used TopHat2 to compare the treatment group and the control group with the reference group and selected | log 2 Ratio | > 1 and Q < 0.05 genes as DEGs (Kim et al., 2013). Enrichment analysis of DEGs was analyzed using the GO and KEGG databases to obtain a detailed description of the DEGs during seed germination. The DEGs were clustered with p < 0.05, which indicated that the cluster distribution was significant under the comparison between waterlogging and control.
Statistical analysis of data Data were analyzed using Excel and SPSS (SPSS，Chicago，IL，USA) by ANOVA followed by Turkey's significant difference test at p < 0.05. All data had three biological repeats.

Results
Effects of waterlogging on morphology, physiology in P. deltoides 'DHY' The growth of poplar was significantly inhibited after waterlogging ( Figure 1A-D and Figure 6B). A large number of roots died with some new lateral roots ( Figure 1E), and a part of adventitious roots were generated little by little from the stems compared with the control (Figure 1F and 1G). Simultaneously, the ratio of chlorophyll a, chlorophyll b, carotenoid, and chlorophyll a/b in the leaves of P. deltoides 'DHY' were significantly lower than those of the control under waterlogging ( Figure 1H). The activities of antioxidant enzymes including ascorbate peroxidase (APX), glutathione reductase (GR), catalase (CAT), peroxidase (POD), total superoxide dismutase (T-SOD), CuZn superoxide dismutase (CuZn-SOD) in the leaves of P. deltoides 'DHY' changed to varying degrees after 65 d of waterlogging (Table   1). The activities of CAT and GR were significantly higher in waterlogged poplars than in the control ( Table  1). The activities of APX, POD, T-SOD, CuZn-SOD were significantly lower in the waterlogged poplars than in the control. In addition, malondialdehyde (MDA) and soluble protein content were significantly increased compared with the control. Sequence alignment and mapping to the reference genome A total of 12 cDNA libraries were sequenced using Illumina HiSeq, and the sequencing data were stored in NCBI (SRR10828697-SRR10828686). The Q30 of the 12 samples exceeded 94% (Supplementary Table  1). A total of 42,569 genes were obtained by using Trinity software for de novo assembly. The average lengths of genes and N50 were 2,458 and 1,732 bp, respectively. The total length of 4,435 genes was more than 400 bp, and 13,146 genes were over 2,000 bp (Supplementary Table 2). The gene sequences of the samples were compared with commonly used functional databases using Trinotate (Supplementary Table 3). The comparison with the NR database obtained 42,017 annotated genes. The comparison with KOG, Swiss-Prot, PFAM, COG, GO,KEGG,and eggNOG databases obtained 20,756,29,154,398,14,329,25,089,12,982,and 33,486 annotated genes, respectively. Moreover, 14,329 and 25,089 genes were annotated into the 25 COG and 50 GO pathways, respectively, in the transcriptome database of P. deltoides 'DHY' (Supplementary Figure   1). These above results indicate that the assembly quality of sequencing data meets the requirements of transcriptome analysis.
Comparative analysis of DEGs The total number of genes upregulated in the leaves and roots of P. deltoides 'DHY' under waterlogging conditions were 920 and 4,283, whereas 1,527 and 5,182 were downregulated, respectively ( Figure 2A). The leaves and roots of P. deltoides 'DHY' under waterlogging were 1,426 and 8,444 genes specifically expressed, and 1021 commonly expressed in the leaves and roots of waterlogging P. deltoides 'DHY', respectively, compared with the control ( Figure 2B  DEGs of KEGG analysis KEGG enrichment analysis of DEGs was performed in this study to assess the response mechanism of P. deltoides 'DHY' under waterlogging stress. As shown in the Figure 3A, a total of 518 DEGs were enriched to 105 pathways in the leaves of control (DPLC) versus the leaves of waterlogging treatment (DPLF) group, mainly including photosynthesis, photosynthesis-antenna protein, glycolysis, amino acid biosynthesis, starch and sucrose metabolism, peroxisome, pentose phosphate pathway, and pyruvate metabolism. A total of 1,936 DEGs were enriched to 126 pathways, including amino acid biosynthesis, starch and sucrose metabolism, ascorbic acid and uronic acid metabolism, glycolysis, and fatty acid degradation in the roots of control (DPRC) versus the roots of waterlogging treatment (DPRF) group ( Figure 3B). Furthermore, 305 and 716 DEGs were screened from leaves and roots, respectively, by further transcriptome analysis which may be related to waterlogging tolerance ( Figure 3C and D). The cluster analysis of the expression amount of DEGs showed that 117 and 271 were upregulated and 188 and 445 were downregulated DEGs in the leaves and roots, respectively. The X-axis shows the enrichment factor. A high q-value is represented by green and a low q-value is represented by red (q < 0.05). Bubble size indicates DEG number (increases with DEG number) DEGs related to adventitious root primordia initiation under waterlogging stress A large number of adventitious roots appeared on the stems of the surface of the waterlogged roots of P. deltoides 'DHY' after 65 d of waterlogging treatment ( Figure 1E). Simultaneously, 14 DEGs were found to be related to adventitious root germination in waterlogged roots ( Figure 4A). The expression of Protein Root Initiation Defective 3 (RID3), Aberrant root formation protein 4 (ALF4), Protein Lateral root primordium 1.1 (LRP1.1) and Protein Short-root 1 (SHR1) genes were significantly upregulated, suggesting that these genes may be involved in the formation of adventitious roots of poplars under waterlogging stress. The sample names are showed at the bottom. The change in expression level is represented by the change in color, blue indicates a lower expression level and red indicates a higher expression level DEGs related to stress response under waterlogging stress A total of 14 and 30 DEGs were related to stress response in roots and leaves, respectively ( Figure 4B and C). The DEGs upregulated in the leaves of P. deltoides 'DHY' under waterlogging stress included CBLinteracting serine / threonine-protein kinase 7 (CIPK7), LOB domain-containing protein (LBD1), calmodulin-interacting protein 111 (CIP111), 3 calmodulin-like protein (CML) and mitogen-activated protein kinase kinase kinase 18 (MAPKKK18) genes. Moreover, 2 CIPK, 2 LBD, 3 Ankyrin repeat domaincontaining protein (AKR), Calcium-dependent protein kinase SK5 (CDPK), 2 Mitogen-activated protein kinase (MPK), and 5 Calcium-dependent protein kinase (CPK) genes were upregulated in the roots under waterlogging conditions. In addition, the expression of aquaporin genes in the leaves and roots of P. deltoides 'DHY' was downregulated under waterlogging stress.
The DEGs related to sugar metabolism under waterlogging stress The soluble sugar content in the leaves of P. deltoides 'DHY' was significantly higher than that of the control after 65 d of waterlogging ( Figure 6A). The RNA-seq results showed 22 and 38 DEGs in the leaves and roots of the control and waterlogged poplar, respectively ( Figure 6C and D). Furthermore, 14 and 25 DEGs significantly upregulated in the leaves and roots, respectively ( Figure 7A and B). The expression of ENO, glyceraldehyde 3-phosphate dehydrogenase (GAPDH), phosphoglycerate kinase (PGK), pyruvate kinase (PK), triosephosphate isomerase (TPI), fructose-bisphosphate aldolase (ALDO), hexokinase (HK), aldose 1epimerase (GALM), 6-phosphofructokinase 1 (PFK), diphosphate-dependent phosphofructokinase (PFP) and 2,3-bisphosphoglycerate-dependent phosphoglycerate mutase (PGMA) were significantly higher expressed in roots and leaves, whereas PGM was only significantly upregulated in roots. Moreover, 9 and 21 DEGs of genes were involved in glucose metabolism pathways in the leaves and roots, respectively. The expression levels of isocitrate dehydrogenase (IDH)s, malate dehydrogenase (CMDH), pyruvate decarboxylase (PDC) and alcohol dehydrogenase (ADH1) were significantly upregulated in the leaves of P. deltoides 'DHY', and the expression levels of malate dehydrogenase (MDH1), 3 PDC, 3 aldehyde dehydrogenase (ALDH), 5 ADH, phosphoenolpyruvate carboxylase 4 (PPC4) and sucrose synthase (SUS2) were significantly upregulated in the roots.   The net photosynthetic rate, stomatal conductance, intercellular CO2 concentration, and transpiration rate in the leaves of P. deltoides 'DHY' were significantly lower than those in the control under waterlogging stress ( Figure 8A). The maximum photosynthetic efficiency, the maximum fluorescence under light, the maximum fluorescence after dark adaptation, the apparent electron transport rate, photochemical quenching coefficient, and actual photochemical quantum yields also decreased significantly compared with the control ( Figure 8B). RNA-seq analysis showed that 50 DEGs related to photosynthesis and light response were involved in this pathway ( Figure 8C and D), including genes such as photosystem I (PSI), photosystem II (PSII), lightharvesting complex I chlorophyll a/b binding protein 1 (LHCA), and light-harvesting complex II chlorophyll a/b binding protein 3 (LHCB). The expression levels of most of the DEGs involved in photosynthesis and light response were significantly downregulated compared with the control.

DEGs related to antioxidant and defense systems under waterlogging stress
A total of 2 catalases (CAT), peroxidase (PER), respiratory burst oxidase homolog (ROBH), and 3 superoxide dismutases (SOD) were significantly downregulated and the gene expression levels of 3 glutathione s-transferases (GST) and 3 GST23s were significantly increased in the leaves of P. deltoides 'DHY' under waterlogging conditions ( Figure 10A). Moreover, 3 GSTs, CAT1 and 3 SODs were significantly upregulated in the waterlogged roots of P. deltoides 'DHY'. However, peroxidase (POD) and 3 RBOHs were significantly downregulated ( Figure 10B). In addition, GSTU17 was found to be significantly downregulated in both leaves and roots under waterlogged conditions. DEGs related to TFs under waterlogging stress A total of 700 transcription factors were annotated in the leaves and roots of poplar under waterlogging stress and control and were classified into 62 families ( Figure 11A), of which AP2/ERF-ERF family was the largest, including 75, followed by NAC, MYB, WRKY, C2H2, bHLH, and AUX/IAA. A total of 27 and 39 transcription factors of stress and defense response in the leaves and roots of P. deltoides 'DHY' were observed through calorimetry analysis of DEG expression ( Figure 11B and C). Significantly upregulated TFs included 2, 1, 2, and 8 members of the NAC, MYB, bHLH, and AP2/ERF family in flood-stressed leaves, whereas the downregulated transcription factor gene had one NACs. The significantly upregulated ones in the waterlogged roots included 5 members of WRKY family, 2 NAC022 of 2 NAC family, 1 MYB17 of MYB family, and 2 TFs of bHLH family; whereas the significantly downregulated ones included 1 WRKY22 of WRKY family and 8 TFs of AP2/ERF family. In addition, MYB86 of MYB family was found to be significantly downregulated in both leaves and roots under waterlogged conditions.

Discussion
Initiation adventitious root primordia paly key role in the waterlogging tolerance of P. deltoides 'DHY' P. deltoides 'DHY' under waterlogging treatment had fewer roots below the ground, and the leaves turned yellow but not abscission compared with the control form figure one. Simultaneously, the chlorophyll content showed a significant downward trend, indicating that P. deltoides 'DHY' was inhibited to some extent under flood stress. Studies have shown that the decrease in chlorophyll content is thought to be caused by the acceleration of chlorophyll degradation or obstruction of chlorophyll synthesis (Zhu et al., 2018). In addition, a large number of adventitious roots were produced under waterlogging stress compared with the control. The formation of adventitious roots is an active adaptation mode of plants to waterlogging (Colmer et al., 2009), which can rapidly replace the primary roots suffocating and dying due to hypoxia. The porosity of the internal tissue of roots is greatly increased, and the energy of oxygen uptake and transportation is significantly improved, which is conducive to reducing the long-distance oxygen transportation (Rich et al., 2012；Xu et al., 2017. This feature has been reported in corn (Mano et al., 2009), cucumber (Xu et al., 2017), and soybean (Kim et al., 2015) under waterlogging stress.
Four important genes related to adventitious root formation were also excavated from the transcriptome data and significantly upregulated in the roots of P. deltoides 'DHY' under waterlogging stress. Studies have reported that the expression of ALF4 gene mutants in Arabidopsis thaliana thwarts the formation of lateral roots, which will also change the structure of the root system (DiDonato et al., 2014). SHR gene was identified in Arabidopsis to regulate the growth of adventitious roots. SHR is the regulator of adventitious roots in Arabidopsis thaliana, also necessary for the initiation and formation of lateral root primordia to maintain the infinite growth of lateral and anchored roots, and regulates the root-related development process (Lucas et al., 2011). Besides, the RID3 gene germinated in hypocotyls in Arabidopsis thaliana, which encoded an evolutionarily conserved methyltransferase-like protein localized in the nucleus and most prominently accumulated in the nucleoli (Ohbayashi et al., 2011). The AR primordium in Arabidopsis thaliana originates from the hypocotyl formation layer, forming lateral roots similar to the surrounding tissues of the native root (Verstraeten et al., 2013;Agulló-Antón et al., 2014), and LRP1.1 is highly expressed in P. deltoides 'DHY'.
Therefore, results of this study indicate that waterlogged treatment induced the upregulated expression of some adventitious roots in P. deltoides 'DHY', which in turn produced adventitious roots to improve the unfavorable factors caused by hypoxia in waterlogged conditions and gradually restore normal metabolic pathways in plants.
Candidate genes related to stress response contribute to the waterlogging tolerance of P. deltoides 'DHY' The response mechanisms of plants under waterlogging stress include changes in signal transduction, antioxidant system, and stress response (Ruperti et al., 2019). Therefore, screening stress response genes under waterlogging is crucial for understanding the molecular mechanism of poplar tolerance to waterlogging (Fukao et al., 2019). CML, CBL and CDPK in plants can monitor dynamic levels of calcium, and these proteins transmit signals and trigger responses from downstream genes (Valmonte et al., 2014). For example, the increased expression of NtCDPK4 and OsCPK13 enhanced the tolerance to drought and high salt (Wan et al., 2007), and AtCPK6 overexpression in Arabidopsis thaliana increased the expression of stress-related genes and proline content (Xu et al., 2010). These results were verified in this study. The expression of LBD1 and CDPK genes in the leaves and roots of P. deltoides 'DHY' were significantly upregulated under waterlogging stress in this study.
Three LBD genes were found to be upregulated in the leaves and roots of P. deltoides 'DHY' under waterlogging stress in this experiment. MtLBD1 can control the root structure of Medicago truncatula under salt stress (Ariel et al., 2010). Therefore, LBD gene was speculated to play a positive role in regulating waterlogging stress. A total of 3 Aquaporin TIPs (TIP) and aquaporin PIP (PIP)in leaves, and 4 TIPs in the roots of P. deltoides 'DHY' were downregulated under the waterlogging stress in this study. The plasma membrane of the aquaporin gene plays an important role in maintaining the water balance of plants under the condition of excessive water content of protein gene and inner protein of vacuole membrane to reduce the damage of osmotic stress between cells caused by excessive water content (Galmés et al., 2007). Therefore, the physiological interference caused by waterlogging to plants could be reduced by downregulating the expression of these aquaporin genes in P. deltoides 'DHY' under the stress of waterlogging is reasonable (Kadam et al., 2017).

Function of Hormonal regulation of AR emergence in the waterlogging tolerance of P. deltoides 'DHY'
The formation of adventitious roots and plant ethylene-induced adventitious roots under waterlogging stress depended on the directional transport of polar auxins (Vidoz et al., 2010). RNA-seq results showed that ARF17 and GH3 genes were upregulated in the leaves and roots of P. deltoides 'DHY' under waterlogging conditions. ARF17 positively regulated GH3 genes, and members of the GH3 protein family participated in the homeostasis of auxin (Kramer et al., 2015). In addition, the PIN protein gene upregulated in the roots of P. deltoides 'DHY' is involved in the transport of intracellular auxin (Adamowski et al., 2015). Hence, auxin may play an important role in regulating the growth of adventitious roots. This study found that NCED1, a key gene involved in ABA biosynthesis, was upregulated in the leaves of P. deltoides 'DHY', but NCED1 and NCED3 were downregulated in roots, and ABA receptor-related genes were downregulated in the leaves and roots under waterlogged conditions. This result is consistent with the decrease in ABA content and expression of related genes in hypocotyls of cucumber after 48h of waterlogging and can promote the synthesis of adventitious roots (Dawood et al., 2016;Xu et al., 2017).
Studies have shown that ethylene signals can be detected within a few hours after waterlogging stress, play a key role in the plants' response to waterlogging, and activate downstream gene expression to promote the growth of adventitious roots (Phukan et al., 2016). However, the long-term increase of ethylene content can do harm to plant root growth (De and Hanecakova, 2008). Accumulation of ethylene in plant tissues increased the rate of active oxygen generation, leading to the destruction of macromolecules and ultimately cell death (Ahmed et al., 2006). RNA-seq results showed that 28 DEGs involved in the ethylene biosynthetic pathway were screened out from waterlogged and control P. deltoides 'DHY', and the expression levels of most DGEs were significantly downregulated and low-expressed. This finding means that the roots of P. deltoides 'DHY' can reduce the release of ethylene and increase the plants' tolerance to waterlogging stress by downregulating the expression of the above genes in waterlogged conditions. Role of sugar in the waterlogging tolerance of P. deltoides 'DHY' Many studies have found that the content of soluble sugar in plants is significantly increased under waterlogging conditions, which is an adaptation and protection mechanism for adversity and provides energy under hypoxic conditions (Peng et al., 2018). This study also detected that the soluble sugar content in P. deltoides 'DHY' under waterlogged conditions was significantly higher than that of the control. RNA-seq analysis further excavated 22 and 38 DEGs involved in sugar metabolism pathways. IDH, CMDH, and PDC genes in the leaves were upregulated, which promoted the normal function of the tricarboxylic acid cycle metabolic pathway (Rocha et al., 2009). One MDH1 was also upregulated in the roots, indicating that the energy metabolism pathways in the leaves and roots of P. deltoides 'DHY' may be relieved by hypoxia from adventitious roots, allowing plants to perform normal mitochondrial respiration under aerobic conditions. In addition, the PDC in the root downregulated the expression, which inhibited the tricarboxylic acid cycle (Rocha et al., 2009). A significant upregulated SUS2 gene was also found in the roots of P. deltoides 'DHY' under waterlogging treatment. The protease has been proved to catalyze the transformation of sucrose into glucose, which led to the inhibition of mitochondrial respiration (Zabalza et al., 2009), thereby increasing the tolerance of red poplar to waterlogging. ADH was upregulated in the roots and leaves of P. deltoides 'DHY' under waterlogging stress, which promoted the ethanol fermentation pathway and accelerated the degradation of ethanol (Arora et al., 2017). This result may also be one of the reasons why P. deltoides 'DHY' resisted waterlogging stress by activating ethanol metabolism.
Candidate genes associated with photosynthesis and chlorophyll fluorescence participate in the waterlogging tolerance of P. deltoides 'DHY' The gas photosynthetic parameters and photosynthetic fluorescence parameters of P. deltoides 'DHY' leaves were significantly decreased under waterlogging conditions compared with the control, and similar findings were also found in P. deltoides and Jerusalem artichoke (Miao et al., 2017;Yan et al., 2018). Moreover, most genes related to photosynthesis, including PSII, PSI, cytochrome b6/f complex, ATP synthase, photosynthetic electron transport chain, and photosynthetic antenna protein, were also significantly reduced in the leaves of P. deltoides 'DHY'. A total of 13 transcripts were significantly differentially expressed in PSII, but only 11 transcripts were differentially expressed in PSI, indicating that PSII is more sensitive than PSI under waterlogging stress. This result is consistent with previous reports that PSI is generally more stable than PSII under stress (Doupis et al., 2016;Zhu et al., 2018). In summary, transcriptome comparison and physiological test results showed that long-term waterlogging stress of P. deltoides 'DHY' was closely related to the primary photosynthetic metabolic process, and the decrease in photosynthetic activity was also related to the blocked chlorophyll synthesis.
Candidate genes involved in antioxidant and defense systems requirements for the waterlogging tolerance of P. deltoides 'DHY' Plants have evolved a series of enzymatic ROS clearance mechanisms to maintain intracellular ROS balance. The activities of APX and GR under waterlogging conditions in the leaves of P. deltoides 'DHY' were significantly increased compared with the control, indicating that these two enzymes participate in the removal of excessive ROS in the plant, thereby reducing the damage to the plants. ROBH is a rate-limiting enzyme produced by reactive oxygen species. ROBH forces mitochondrial respiratory pathway to reach the saturation of redox chain under hypoxia stress, accumulating NAD(P)H and producing ROS (Yamauchi et al., 2014;He et al., 2015), which is the main source of ROS outbreak under waterlogging stress (Shabala et al., 2014). ROBH expression levels in the leaves and roots of P. deltoides 'DHY' were significantly downregulated under waterlogging stress. Hence, the rate of ROS production in the leaves and roots of P. deltoides 'DHY' was declining, and the oxidative damage suffered gradually. In addition, the expression of GSTF and GST in the leaves and APX and SOD antioxidant enzyme genes in roots were upregulated under waterlogging conditions, indicating that these two genes are involved in the important role of tolerance to waterlogging (Waszczak et al., 2018).
Activation of proline biosynthesis in the waterlogging tolerance of P. deltoides 'DHY' Proline has excellent ROS scavenging ability, and abiotic stress leads to an increase in plant proline content (Alia and Sardhi, 1991). The proline content in P. deltoides 'DHY' leaves also showed a significant upward trend under the waterlogging stress, which is consist with waterlogging resistance in maize (Kravi et al., 2013). Results of KEGG analysis showed that five genes (ALDO, GAPDH, ENO, PK, GOT) related to proline synthesis pathway were significantly upregulated in the leaves and roots of P. deltoides 'DHY'. Simultaneously, this finding is consistent with the results of DEGs in the leaves and roots of our transcriptome. The above results indicate that P. deltoides 'DHY' can tolerate waterlogging stress by inducing the proline synthesis pathway gene expression under waterlogging stress to promote proline synthesis.
Candidate TFs associated with the waterlogging tolerance of P. deltoides 'DHY' Transcription factors are key regulators of responses to various biotic or abiotic stresses (Singh et al., 2002). Transcription factors known to cope with waterlogging stress include bZIP , NAC (Wang et al., 2019), WRKY (Jiang et al., 2009), MYB , AP2/ERF (Sasidharan et al., 2015) and bHLH family (Jing et al., 2017). AP2/ERF family members, including RAP2.1.1, 3 RAP2.3, RAP2.4, and 3 DREB transcription factors were significantly upregulated in response to hypoxia in P. deltoides 'DHY' leaves, but 2 RAP 2.12, RAP 2.4, and DREB2A in roots were significantly downregulated. The lack of oxygen in the submerged roots can be alleviated due to the germination of adventitious roots. However, the oxygen content in the leaves was low, which promoted hypoxia in response to transcription factors. This result is consistent with the stability of the transcription factor RAP2.12 and results of dynamic regulation of nuclear localization by oxygen concentration (Monika et al., 2015). A total of 2 NAC, 2 WRKY, and bHLH transcription factor genes were upregulated in the leaves and roots under waterlogging stress. These transcription factor genes were inferred to play an important role in the waterlogging tolerance of P. deltoides 'DHY', and their functions in waterlogging tolerance need further verification.

Conclusions
This study provided a new understanding of the mechanism of waterlogging tolerance and the formation of adventitious roots of poplar under waterlogging conditions through the comparative analysis of RNA-seq and physiological indexes. Waterlogging can make a series of changes in plant morphology, physiology, photosynthesis, antioxidant enzyme system, energy metabolism, and the germination of the adventitious roots to promote the adaptation and long-term growth of P. deltoides 'DHY'. The mechanism of tolerance to waterlogging was elucidated from morphology and physiology aspects to molecule to provide an important basis for molecular breeding of poplar.
Authors' Contributions G.L., M.H., Z.L., F.X and Z.Z conceived and designed the experiment. G.L., J.Y., W.Z. and F.X. analyzed the data. Y.L., participated validation and investigation. G.L. and F.X. wrote the paper. All authors prepared, read, and approved the final manuscript.