Potential Role of Photosynthesis in the Regulation of Reactive Oxygen Species and Defence Responses to Blumeria graminis f. sp. tritici in Wheat

Photosynthesis is not only a primary generator of reactive oxygen species (ROS) but also a component of plant defence. To determine the relationships among photosynthesis, ROS, and defence responses to powdery mildew in wheat, we compared the responses of the Pm40-expressing wheat line L658 and its susceptible sister line L958 at 0, 6, 12, 24, 48, and 72 h post-inoculation (hpi) with powdery mildew via analyses of transcriptomes, cytology, antioxidant activities, photosynthesis, and chlorophyll fluorescence parameters. The results showed that H2O2 accumulation in L658 was significantly greater than that in L958 at 6 and 48 hpi, and the enzymes activity and transcripts expression of peroxidase and catalase were suppressed in L658 compared with L958. In addition, the inhibition of photosynthesis in L658 paralleled the global downregulation of photosynthesis-related genes. Furthermore, the expression of the salicylic acid-related genes non-expressor of pathogenesis related genes 1 (NPR1), pathogenesis-related 1 (PR1), and pathogenesis-related 5 (PR5) was upregulated, while the expression of jasmonic acid- and ethylene-related genes was inhibited in L658 compared with L958. In conclusion, the downregulation of photosynthesis-related genes likely led to a decline in photosynthesis, which may be combined with the inhibition of peroxidase (POD) and catalase (CAT) to generate two stages of H2O2 accumulation. The high level of H2O2, salicylic acid and PR1 and PR5 in L658 possible initiated the hypersensitive response.


Introduction
Plants are challenged by various biotic stresses, including viral, bacterial, and fungal pathogenic stresses, which have substantial economic and ecological impacts [1,2]. Plants depend on their innate immune system to perceive and respond to biotic stimuli [3]. The first line of the immune response is called pathogen-associated molecular pattern (PAMP)-triggered immunity (PTI), which is triggered by various PAMPs [4]. During this period, the physiological and biochemical statuses of plants correspondingly change, which include phenomena such as cytoskeletal reorganization, cell wall reinforcement, reactive oxygen species (ROS) bursts, and stomatal closure [5][6][7]. To defend against pathogens further, resistance (R) proteins of plants recognize specific effectors generated by pathogens in the infection process to trigger the second line of defence referred to as effector-triggered immunity

Difference in H2O2 Accumulation and Related DEG Expression between L658 and L958
H2O2 accumulation is a typical host defence response to biotrophic fungi [4]. In this study, the results of 3'-diaminobenzidine (DAB) staining showed that H2O2 similarly accumulated under primary germ tube (PGT) at 6 and 12 hpi and then accumulated under penetration peg (PP) at 24 hpi in both L658 and L958 (Figure 2A,B). Subsequently, a stronger H2O2 response was generated in the whole infected and adjacent cells in L658 at 48 hpi, while H2O2 faded gradually in L958 at 48 and 72 hpi (Figure 2A,B). In addition, the percentage of interaction sites stained by DAB had no significant difference between L658 and L958 at 12 and 24 hpi, while that in L658 was significantly (p < 0.05) higher at both 6 and 48 hpi and lower at 72 hpi than that in L958 ( Figure 2C).

Difference in H 2 O 2 Accumulation and Related DEG Expression between L658 and L958
H 2 O 2 accumulation is a typical host defence response to biotrophic fungi [4]. In this study, the results of 3 -diaminobenzidine (DAB) staining showed that H 2 O 2 similarly accumulated under primary germ tube (PGT) at 6 and 12 hpi and then accumulated under penetration peg (PP) at 24 hpi in both L658 and L958 (Figure 2A,B). Subsequently, a stronger H 2 O 2 response was generated in the whole infected and adjacent cells in L658 at 48 hpi, while H 2 O 2 faded gradually in L958 at 48 and 72 hpi (Figure 2A,B). In addition, the percentage of interaction sites stained by DAB had no significant difference between L658 and L958 at 12 and 24 hpi, while that in L658 was significantly (p < 0.05) higher at both 6 and 48 hpi and lower at 72 hpi than that in L958 ( Figure 2C). the solid and dotted lines represent L658 and L958, respectively. The vertical bars represent the means ± SEs. The asterisks represent significant differences as follows: ** p < 0.01 and * p < 0.05. The asterisk at the top of the line chart represents the difference between L658 and L958 at each time point. The asterisk on the trend line represents the difference between two adjacent time points for the same genotype.
To further determine the difference of H2O2 at the transcript level, we identified some DEGs related to H2O2 generation and H2O2 scavenging (Table 1). In L658, the expression of H2O2 generationrelated DEGs, such as oxalate oxidase (OXO) was upregulated, especially at 12 hpi, while the expression of many H2O2-scavenging DEGs, such as POD and CAT isozymes, was downregulated in L658 compared with L958 (Table 1). In L958, the expression of other H2O2 generation-related DEGs, including respiratory burst oxidase homologue protein (RBOH), polyamine oxidase (POX), and amine oxidase (AOX), were usually upregulated, while DEGs encoding POD and CAT were commonly upregulated in L958 compared with L658 (Table 1). This result indicated that although different DEGs regulated H2O2 generation in L658 and L958, the capacity to scavenge H2O2 was more seriously inhibited in L658 than L958. Each point represents at least 100 infection sites of each of three leaf pieces, and the lowercase letter at the top of the bar chart represents statistically significant differences at p < 0.05. Activities of SOD (D), POD (E), and CAT (F) at different inoculation time points in the two genotypes; the solid and dotted lines represent L658 and L958, respectively. The vertical bars represent the means ± SEs. The asterisks represent significant differences as follows: ** p < 0.01 and * p < 0.05. The asterisk at the top of the line chart represents the difference between L658 and L958 at each time point. The asterisk on the trend line represents the difference between two adjacent time points for the same genotype.
To further determine the difference of H 2 O 2 at the transcript level, we identified some DEGs related to H 2 O 2 generation and H 2 O 2 scavenging (Table 1). In L658, the expression of H 2 O 2 generation-related DEGs, such as oxalate oxidase (OXO) was upregulated, especially at 12 hpi, while the expression of many H 2 O 2 -scavenging DEGs, such as POD and CAT isozymes, was downregulated in L658 compared with L958 (Table 1). In L958, the expression of other H 2 O 2 generation-related DEGs, including respiratory burst oxidase homologue protein (RBOH), polyamine oxidase (POX), and amine oxidase (AOX), were usually upregulated, while DEGs encoding POD and CAT were commonly upregulated in L958 compared with L658 (Table 1). This result indicated that although different DEGs regulated H 2 O 2 generation in L658 and L958, the capacity to scavenge H 2 O 2 was more seriously inhibited in L658 than L958. The antioxidant enzyme activities further confirmed the transcriptome data. The results showed that although the activity of SOD and POD had no significant change in L658 ( Figure 2D,E), CAT activity was sharply decreased from 6 to 12 hpi (p < 0.01) and from 24 to 48 hpi (p < 0.05) ( Figure 2F). In L958, the activity of SOD significantly increased from 6 to 12 hpi while the CAT activity also significantly increased from 12 to 24 hpi ( Figure 2F). In contrast, the CAT activity in L958 was significantly higher (p < 0.01) than that in L658 at 12 and 48 hpi ( Figure 2F). Therefore, the inhibition of H 2 O 2 scavenging-related genes and enzyme activities in L658 likely caused a more abundant accumulation of H 2 O 2 than in L958.

Changes in Photosynthesis and Photosynthesis-Related Genes in Response to Bgt
Photosynthesis is not only the main generator of H 2 O 2 , but also regarded as a plant defence component [13]. To investigate the influence of Bgt in photosynthesis, we identified a large number of photosynthesis-related DEGs, most of which encoded chlorophyll a/b-binding proteins (Cab), proteins in reaction centres, ATP synthase (ATPase), ribulose bisphosphate carboxylase small chain/large chain (RbcS/L) and Rubisco activase (RCA) ( Figure 3A). Interestingly, we found that almost all of the photosynthesis-related DEGs were downregulated in L658 compared with L958, and 54 (98%) out of 55 and 75 (96%) out of 78 photosynthesis-related DEGs were downregulated in L658 at 12 and 48 hpi, respectively ( Figure 3B, Table S3). Furthermore, 41 and 98 DEGs were also downregulated in L658 at 12 hpi and 48 hpi compared with 0 hpi, respectively, and 45 DEGs were also downregulated at 12 hpi compared with 6 hpi in L658 ( Figure 3C, Table S4). On the contrary, only eight photosynthesis-related DEGs were downregulated in L958 at both 24 and 72 hpi compared with 0 hpi, and 25 DEGs were downregulated at 24 hpi compared with 12 hpi ( Figure 3C, Table S4). This result indicated that the expression of photosynthesis-related genes in L658 was more seriously inhibited compared with that in L958, which could lead to the disruption of photosynthesis. To further confirm the photosynthesis-related transcriptome data, we detected the photosynthetic parameters. As shown in Figure 4, the net photosynthetic rate (Pn) of L658 decreased from 0 hour (without inoculation) to 12 hpi and from 48 to 72 hpi ( Figure 4A) and stomatal conductance (Gs) exhibited a reduction from 0 to 12 hpi and an increase from 12 to 72 hpi ( Figure 4B). Compared with Pn, the intercellular CO 2 concentration (Ci) of L658 increased significantly (p < 0.01) from 6 to 12 hpi and from 24 to 72 hpi ( Figure 4C). The opposite change trend of Pn and Ci in L658 indicated that non-stomatal limitation, such as the inhibition of photosynthesis-related gene expression, was the main reason for the decline of Pn. However, the Pn, Gs, and Ci of L958 showed similar change trends, with a decrease from 0 to 12 hpi and from 48 to 72 hpi ( Figure 4A-C), which indicated that the Pn of L958 decreased under stomatal limitation.
L958 at various inoculation time points compared with 0 h (left) and between two adjacent time points (right).
To further confirm the photosynthesis-related transcriptome data, we detected the photosynthetic parameters. As shown in Figure 4, the net photosynthetic rate (Pn) of L658 decreased from 0 hour (without inoculation) to 12 hpi and from 48 to 72 hpi ( Figure 4A) and stomatal conductance (Gs) exhibited a reduction from 0 to 12 hpi and an increase from 12 to 72 hpi ( Figure 4B). Compared with Pn, the intercellular CO2 concentration (Ci) of L658 increased significantly (p < 0.01) from 6 to 12 hpi and from 24 to 72 hpi ( Figure 4C). The opposite change trend of Pn and Ci in L658 indicated that non-stomatal limitation, such as the inhibition of photosynthesis-related gene expression, was the main reason for the decline of Pn. However, the Pn, Gs, and Ci of L958 showed similar change trends, with a decrease from 0 to 12 hpi and from 48 to 72 hpi ( Figure 4A-C), which indicated that the Pn of L958 decreased under stomatal limitation.  Figure 1. The asterisks represent significant differences as follows: ** p < 0.01 and * p < 0.05. The asterisk at the top of the line chart represents the difference between L658 and L958 at each time point. The asterisk on the trend line represents the difference between two adjacent time points for the same genotype.
Furthermore, we used chlorophyll fluorescence to evaluate changes in PSII. In L658, the efficiency of excitation capture by open PSII reaction centres (Fv'/Fm'), actual photochemical efficiency of PSII (ФPSII), electron transport rate (ETR), and coefficient of photochemical chlorophyll fluorescence quenching (qP) decreased from 0 to 12 hpi and from 24 to 72 hpi ( Figure 4D-G), and the  Figure 1. The asterisks represent significant differences as follows: ** p < 0.01 and * p < 0.05. The asterisk at the top of the line chart represents the difference between L658 and L958 at each time point. The asterisk on the trend line represents the difference between two adjacent time points for the same genotype.
Furthermore, we used chlorophyll fluorescence to evaluate changes in PSII. In L658, the efficiency of excitation capture by open PSII reaction centres (Fv'/Fm'), actual photochemical efficiency of PSII (Φ PSII ), electron transport rate (ETR), and coefficient of photochemical chlorophyll fluorescence quenching (qP) decreased from 0 to 12 hpi and from 24 to 72 hpi ( Figure 4D-G), and the maximal photochemical efficiency of PSII in dark-adapted leaves (Fv/Fm) increased from 0 to 6 hpi and then decreased to 72 hpi ( Figure 4H). Although the change tendency of Fv'/Fm', Φ PSII , ETR and qP of L958 was similar to that of L658, the Φ PSII , ETR and qP of L958 were inhibited more strongly than those of L658 ( Figure 4D-G). Interestingly, non-photochemical quenching (NPQ) significantly (p < 0.05) increased from 0 to 12 hpi and decreased from 12 to 72 hpi in L658, while it decreased from 0 to 6 hpi and increased from 48 to 72 hpi in L958 ( Figure 4I). These data indicated that although the electron transport chain was blocked in both L658 and L958, the different changes of NPQ from 48 to 72 hpi may cause the differences in H 2 O 2 accumulation at 48 and 72 hpi between L658 and L958.

Identification of DEGs Related to Phytohormones and PR Proteins
To understand the difference in hormone regulation between L658 and L958, we identified some phytohormone DEGs. The DEGs involved in SA synthesis, such as salicylic acid-binding protein (SABP) and isochorismate synthase (ICS), were downregulated in L658. However, NPR1, which is a downstream key regulator of SA, was upregulated in L658 at all time points (Table 2, Figure S1). In contrast, the DEGs involved in JA/ET pathways, such as the JA biosynthesis-related DEGs encoding lipoxygenase (LOX) and allene oxide synthase (AOS), and nearly all ET-related DEGs, including ethylene-responsive transcription factor (ERF), ethylene insensitive (EIN), 1-aminocyclopropane-1-carboxylate synthase (ACS), and 1-aminocyclopropane-1-carboxylate oxidase (ACO), were globally downregulated in L658, especially at 72 hpi (Table 2, Figure S1). On the contrary, the expression of NPR1 was inhibited in L958, whereas JA/ET-related genes were highly expressed (Table 2, Figure S1). The results of the qRT-PCR analysis further confirmed that the expression of NPR1 and EBF1 in L658 was greater than that in L958 ( Figure 5A,B), and the expression of LOX was lower in L658 than L958 ( Figure 5C). These results showed that the SA pathway was activated and the JA/ET pathways were supressed in L658, while the JA pathway was overactivated in L958.     Figure S1). On the contrary, the expression of NPR1 was inhibited in L958, whereas JA/ET-related genes were highly expressed (Table 2, Figure S1). The results of the qRT-PCR analysis further confirmed that the expression of NPR1 and EBF1 in L658 was greater than that in L958 ( Figure 5A,B), and the expression of LOX was lower in L658 than L958 ( Figure 5C). These results showed that the SA pathway was activated and the JA/ET pathways were supressed in L658, while the JA pathway was overactivated in L958.  The dotted line means that the relative expression in L958 was set as 1.
The dynamic expression patterns revealed that most of the DEGs encoding pathogenesis-related 1 (PR1) and pathogenesis-related 5 (PR5) were expressed at greater levels in L658, especially at 24 and 48 hpi, while nearly all DEGs encoding pathogenesis-related 10 (PR10) and pathogenesis-related 14 (PR14) were downregulated at all times in L658 (Table 3, Figure S2). However, both PR10 and PR14 exhibited opposite expression tendencies, with higher expression in L958 compared with L658 (Table 3, Figure S2). Furthermore, the results of the qRT-PCR analysis also confirmed that the relative expression of PR14 was lower in L658 than in L958 at 12 and 48 hpi ( Figure 5D). Overall, the high expression of NPR1 in L658 activated the expression of PR1 and PR5, which further induced host cell death ( Figure  S3). However, the high expression of JA/ET-related genes and PR10 and PR14 in L958 may have suppressed the formation of HR ( Figure S3).

Discussion
Powdery mildew is a leaf disease that causes substantial injuries to host photosynthesis, which results in the reduction of photosynthesis-related protein abundance and gene expression [52,55,56]. Previously, plant defence and photosynthesis were studied separately. However, a growing number of studies suggest that photosynthesis is highly related to plant immunity to pathogens [13,14,19,42,43]. Therefore, it is meaningful to investigate the cross-talk between photosynthesis and plant immunity. Dual RNA-seq is widely used to detect the gene expression in various pathosystems, thus providing an effective way to explain the molecular mechanism of plant-fungal interactions [57].
In addition, the measurement of photosynthesis-related parameters, including the net photosynthetic rate and chlorophyll fluorescence, is a very powerful method to monitor and quantify changes in photosynthesis [58,59]. In the present study, we characterized the expression profile of photosynthesis-related genes in resistant and susceptible wheat lines by comparing the wheat transcriptome after inoculation with Bgt, at the same time we measured the photosynthesis-related parameters. In addition, we screened some H 2 O 2 -and defence-related genes and determined cytological reactions to reveal the relationship between photosynthesis and immunity response.

Initial Stronger H 2 O 2 Burst and Secondary Lasting H 2 O 2 Burst Were Regulated by Antioxidant Enzyme Activities and H 2 O 2 -Related Genes and Occurred in Incompatible Reactions
ROS burst is one of the earliest immunity responses after the successful recognition of a pathogen [60]. In general, there are two distinct peaks of ROS burst induced during incompatible reactions: the first phase is often a low-amplitude and transient peak that is followed by a sustained burst of comparatively greater intensity that correlates with disease resistance; however, only the first transient burst of ROS is induced in compatible reactions [60][61][62][63]. In the present study, the first localized H 2 O 2 burst was observed at 6 hpi, and the strength was stronger in L658 than L958 (Figure 2A-C). Subsequently, a stronger H 2 O 2 accumulation generated around the whole infected and adjacent cells at 48 hpi in L658 while H 2 O 2 faded gradually in L958 (Figure 2A-C [16,64,65]. However, powdery mildew does not secrete oxalic acid, thus, the primary role of OXO in cereals would seem to be the production of H 2 O 2 to lignify cell walls against further fungal invasion [65,66]. Hence, DEGs encoding OXO in L658 may be involved in the first H 2 O 2 burst to reinforce the cell wall or transmit defence signals. In addition to OXO, the expression of H 2 O 2 -scavenging DEGs in L658, including those encoding POD and CAT isozymes, emerged as being universally downregulated (Table 1), which corresponded with the inhibition of the POD and CAT enzyme activities in L658 ( Figure 2E,F). The suppression of CAT enzyme activity promotes the increase of H 2 O 2 concentration [29], and transgenic lines that overexpress CAT are more sensitive to pathogens because the capacity to remove H 2 O 2 is over activated [67]. Furthermore, although the correlation between POD genes and plant defence is equivocal, PODs are generally considered to be H 2 O 2 -detoxifying enzymes that may indirectly modulate resistance by regulating H 2 O 2 concentrations [68]. Taken together, the downregulation of POD and CAT gene expression in L658 was likely related to the suppression of enzyme activity, which further led to the increase of H 2 O 2 accumulation. Moreover, the first H 2 O 2 accumulation during the stage from 6 to 12 hpi was possibly involved in reinforcing the cell wall while the second lasting and stronger H 2 O 2 accumulation from 48 to 72 hpi further induced the HR ( Figure S3).
As for L958, three DEGs encoding RBOH were upregulated (Table 1). The RBOH family is reported to mediate the production of apoplastic ROS during the defence responses, however, various RBOH proteins might have different functions in disease resistance [69,70]. For example, silencing Rboh of N. benthamiana led to a suppression of HR and increased the susceptibility to Phytophthora infestans [71]. However, the Arabidopsis atrbohF mutant exhibited more resistance to Peronospora parasitica and increased the HR response [72]. Therefore, whether RBOH is involved in positive or negative resistance to Bgt in L958 needs to be further investigated, but DEGs encoding POD and CAT were clearly upregulated in L958 compared with L658, and the enzyme activities of SOD, POD, and CAT were also significantly higher in L958 compared with L658 (Table 1, Figure 2E,F). Therefore, although other DEGs likely regulated H 2 O 2 generation, the capacity to scavenge H 2 O 2 was enhanced in L958 compared with L658, which caused the first weaker H 2 O 2 burst and failure of the second H 2 O 2 burst.

Global Downregulation of Photosynthesis-Related Genes Likely Disrupted Photosynthesis and Promoted the Generation of H 2 O 2
Generally, virulent or avirulent pathogen invasion often results in a decrease in photosynthesis to different extents, and the reduction in photosynthesis capability may represent a "hidden cost" of defence [29,73]. The plant transcriptome revealed that many photosynthesis-related genes are downregulated under biotic stress [29]. For example, photosynthesis and the expression of the photosynthesis-related transcripts, such as Cab and RbcS, were reduced in barley-Blumeria graminis f. sp. hordei (Bgh) incompatible interaction [74], and the downregulation of Rubisco and ATPase also occurred under other biotic stresses [10,29,75]. However, in some cases, such as in the incompatible interaction between Arabidopsis and P. syringae, the repression of photosynthesis was not correlated with the downregulation of RbcS and Cab transcripts [76], which may be related to the lack of participation of a majority of plant cells in the defence reaction, thus, the repression of RbcS and Cab cannot be detected in the whole leaves' RNA. In the present study, the downregulation of Cab, RbcS, and ATPase transcripts paralleled the decline in photosynthetic rate in L658, especially from 6 to 12 hpi and from 48 to 72 hpi ( Figures 3A and 4A). Meanwhile, the reduction of Pn in L658 was accompanied by the decrease of Gs and the increase of Ci ( Figure 4B,C), which indicated that non-stomatal limitation led to the decrease of Pn in L658 [77,78], which was consistent with the above results. Taken together, the downregulation of photosynthesis-related DEGs in L658 likely switched off photosynthesis activity to strengthen the defence reaction. Inversely, only a small part of photosynthesis-related genes in L958 were downregulated ( Figure 3C). Furthermore, the Pn, Gs, and Ci decreased consistently in L958 at 12 and 72 hpi ( Figure 4A-C), which suggested that the decrease of Pn in L958 was primally controlled by stomatal limitation rather than to regulate immune response.
In addition, the expression of a DEG encoding Fd (TraesCS4A01G356200) was upregulated in L658 compared with L958 at 24 and 48 hpi ( Figure 3A,B). In photosynthesis, Fd is a major element of photosynthesis-associated proteins that accept electrons from PSI and reduce NADP + by ferredoxin: NADP + oxidoreductase (FNR). In some plant-pathogen interactions, Fd has been suggested to participate in pathogen defence directly or to induce the production of ROS to mount the HR [29,34,38]. Therefore, it would be reasonable to assume that the upregulation of Fd in L658 may be involved in plant defence.
To further reveal the defence mechanism of photosynthesis, we monitored the status of PSII by detecting chlorophyll fluorescence indexes. In a previous study, Bgh infection led to a reduction of Φ PSII , an increase of NPQ, and a reduction of Fv/Fm in barley [79]. In general, the increase of NPQ is a typical host response in plant-fungus interactions to protect plants when light energy absorption exceeds the capacity for light utilization, but the decrease in NPQ predisposes chloroplasts to the production of ROS, which might be a priming mechanism for chloroplast ROS signaling during later immune responses [14,80,81]. In our study, Fv'/Fm', Fv/Fm, Φ PSII , ETR, and qP significantly decreased in L658 at 12 and 72 hpi ( Figure 4D-H). This response implied that photosynthetic electron transport was blocked, which could result in electron leakage in L658 for the generation of H 2 O 2 . Furthermore, the increase in NPQ in L658 at 12 hpi and the continuous decrease after 12 hpi may indicate that L658 needed photoprotection before 12 hpi but tended to generate ROS at the late stage to induce the HR ( Figure 4I). In L958, the Φ PSII , ETR, and qP were supressed more seriously than in L658 ( Figure 4E-G), which may be due to damage to the chloroplasts after Bgt invasion. Interestingly, the NPQ of L958 increased at 72 hpi ( Figure 4I), indicating that L958 did not activate the second ROS burst and thus was insufficient to induce the HR.

Mutually Antagonistic Interactions between SA and JA/ET Are Involved in the Regulation of Defence Against Bgt by Regulating Different PR Genes Expression
The difference of H 2 O 2 accumulation and photosynthesis might influence the different regulation of phytohormones, including SA, JA, and ET, after inoculation with Bgt. Generally, SA and JA interact antagonistically under different pathogen attacks and the SA bone fide receptor NPR1 acts as an important regulator to mediate cross-talk between SA and JA [82]. In addition, ET has positive effects on JA but may have positive or negative impacts on SA depending on the pathosystem [83,84]. Regarding the defences against biotrophic fungi, the SA signaling pathway is commonly activated and then directly regulates the conformation of NPR1 to induce the formation of SAR, the activation of PR1 and PR5 and to suppress the JA/ET signaling pathways [82,85,86]. In previous studies, the overexpression of NPR1 in tobacco and apple enhanced resistance to Botrytis cinereal and apple powdery mildew, respectively [87,88]. In a recent study, the repression of NPR1 in Brassica napus L. by RNAi significantly reduced resistance to Sclerotinia sclerotiorum, and silencing NPR1 inhibited SA defence response but enhanced the JA/ET defence response [89]. In the present study, although ICS expression was downregulated at 12 hpi, the transcript TraesCS7A01G021800 encoding NPR1 was upregulated in L658 (Table 2, Figure S1). Compared with SA, the expression of JA biosynthesis genes, including LOX and AOS, ET biosynthesis genes, including ACS and ACO, and ET-mediated signaling genes, such as ERF and EIN, were downregulated in L658, especially at 72 hpi (Table 2, Figure S1). Furthermore, the expression of PR1 and PR5 was upregulated in L658 (Table 3, Figure S2). Taken together, the downregulated expression of JA/ET-related genes indicated that JA/ET levels were also decreased, which could be attributed to increased SA levels in L658. The increased SA and the upregulated expression of NPR1, PR1, and PR5 likely induced the HR in L658.
However, in L958, JA/ET-related genes were strongly expressed (Table 2, Figure S1), which suggested that JA/ET levels increased in L958. Interestingly, the expression of the other two clusters of genes encoding PR10 and PR14 was largely upregulated in L958 as well (Table 3, Figure S2). It is reported that the expression of the PR10 gene is induced by JA and ABA [90]. Furthermore, silencing of PR10-like proteins in Medicago truncatula resulted in the reduced colonization of the oomycete Aphanomyces euteiches and induced the expression of PR5 [91]. In another research, a wheat orthologue of LTP, which is a member of the PR14 family, acted as a negative regulator to resist Puccinia striiformis f. sp. tritici (Pst), and the knockdown of this gene increased wheat resistance to Pst, which was accompanied by the accumulation of H 2 O 2 and SA [92,93]. Therefore, the increase in JA/ET level and the high expression of PR10 and PR14 in L958 may act as negative regulators in the defence response to promote the establishment of susceptibility.

Plant and Pathogen Materials
The resistant wheat line L658 carrying Pm40 and its susceptible sister line L958 without Pm40 were used for comparative analyses by observations of cytological changes and determinations of physiological and biochemical indices. In addition, RNA-seq was performed for wheat plants after they were inoculated with Bgt, as described previously [54]. The fungus was a single-spore isolate collected from a field at Wenjiang Farm, Chengdu, Sichuan Province, and propagated weekly on the highly susceptible wheat line CY20. The wheat plants were cultivated in a growth chamber (Microclima MC1750E, Snijders Scientific, Tilburg, Holland) at 18 • C and 80% relative humidity under a 16-h light/8-h dark photoperiod. Seven-day-old seedlings of L658 and L958 were inoculated with fresh Bgt spores at a density of 100-200 conidia/mm 2 by shaking infected leaf segments.

Cytological Observations of H 2 O 2 and Cell Death
H 2 O 2 accumulation was detected by the DAB (Sigma-Aldrich, Shanghai, China) staining method as described by Wang et al. [94]. The first leaves of L658 and L958 were harvested at 6, 12, 24, 48, and 72 hpi. The inoculated leaves were cut off and the end of cut leaves were immersed in 0.1% (w/v) DAB solution (pH 3.8) for 12 h to take up and react with the DAB fully. Cell death was detected by the trypan blue (Sigma-Aldrich, Shanghai, China) [53]. Leaf segments were immersed in 1% (w/v) TPB solution and then boiled for 2 min to stain the necrotic cells. Afterward, the leaf segments were decoloured in an ethanol-acetic acid (1:1, v/v) mixture until the chlorophyll was removed and then cleared in saturated chloral hydrate. The leaf segments were subsequently stored in 50% glycerol solution for examination and image collection via bright-field microscopy (Nikon Eclipse 80i, Nikon Corporation, Tokyo, Japan).

Determination of Antioxidant Enzyme Activity
A total of 0.2 g of fresh leaves was harvested at 6, 12, 24, 48, and 72 hpi for preparation of the crude enzyme, and samples at 0 h (without inoculation) were used as controls. Each sample was detected with three biological replicates. The fresh leaves were immediately frozen in liquid nitrogen and then homogenized in 2 mL of 50 mM pre-chilled phosphate buffer (pH 7.8) containing 1% (w/v) polyvinylpyrrolidone. The homogenate was then centrifuged at 12,000× g for 20 min at 4 • C, after which the supernatant was used for determination of enzyme activities. The activities of SOD, CAT, and POD were measured at 560, 240, and 470 nm, respectively, as described previously [51].

Determination of Photosynthesis Indices
The photosynthesis index data, including the Pn, Gs, and Ci, were collected with a portable photosynthesis system (LI-6400-02B, LI-COR, Lincoln, NE, USA) as described previously [51], and the measurement conditions included a vapour pressure deficit (VPD) of 0.6 kPa under an actinic light intensity of 1000 µmol/m 2 /s and an air temperature of 22 • C. The first measurements were performed before inoculation, and it was set as the controls (0 h). Subsequently, wheat leaves were inoculated with 100-200 conidia/mm 2 fresh spores at 9:00 am. After 6 h (15:00 p.m.), 12 h (21:00 p.m.), 24 h (9:00 a.m.), 48 h (9:00 a.m.), and 72 h (9:00 a.m.) post inoculation, the photosynthesis indices were collected. The measurements at six time points were performed under the light period of the growth chamber. The data of six independent replications of each genotype were collected by measuring the centre of the first leaf of the wheat plants. The means of these six replications were then analysed statistically.

Measurements of Chlorophyll Fluorescence and Quantum Yields of PSII
Chlorophyll fluorescence was measured with a portable photosynthesis system (LI-COR 6400XT, Lincoln, NE, USA) equipped with a leaf chamber fluorometer (LI-6400-40; LI-COR, Lincoln, NE, USA) as described previously [51]. Measurements were taken from the centre of the first leaves of L658 and L958 at 0 (without inoculation), 6, 12, 24, 48, and 72 hpi. Six plants of each genotype were checked independently at the different timepoints, and the average represented the phenotypic value for subsequent statistical analysis. The data of the qP, maximum fluorescence in the light (Fm'), variable chlorophyll fluorescence yield in the light (Fv'), and Φ PSII were collected in the light. Furthermore, the maximum fluorescence (Fm) and variable chlorophyll fluorescence yield (Fv) were measured with a fluorescence meter after the plants were dark adapted for 30 min.

RNA Extraction, cDNA Library Construction, and RNA-Seq
First leaves were randomly collected from L658 and L958 plants at 0, 6, 12, 24, 48, and 72 hpi, with each sample comprising three biological replicates. A total of 36 leaf samples were frozen in liquid nitrogen and then sent to a service company (Beijing Biomarker Technology Co, Ltd., Beijing, China) for RNA-seq. Briefly, the total RNA in wheat was extracted according to the TRIzol reagent method (Invitrogen Life Technologies, CA, USA). After the integrity and purity of the RNA was checked via an Agilent 2100 Bioanalyzer (Agilent Technologies, CA, USA) and a NanoDrop ND-1000 spectrophotometer (Thermo Scientific, DE, USA), a cDNA library was constructed by the use of a NEBNext Ultra RNA Library Prep Kit for Illumina (NEB, USA). Last, RNA-seq was performed on an Illumina HiSeq™ 2500 platform in accordance with the 150-bp paired-end sequencing strategy. All the raw sequencing data have been submitted to the NCBI Sequence Read Archive (SRA), and the accession number is SRP117269 [54].

Alignment of RNA-Seq Reads and Differential Gene Expression Analysis
Reads that contained an adapter or a poly-N tail or those that were of low quality were removed from the raw reads. The filtered high-quality reads were then mapped to the IWGSC RefSeq v1.0 wheat genome reference (https://wheat-urgi.versailles.inra.fr/) by HISAT2 [95], the process of which allowed up to 2 base mismatches. All identified genes were functionally annotated via BLASTX of the non-redundant (NR) database and the Swiss-Prot protein database [96]. The expression level of the transcripts was subsequently calculated and normalized to an RPKM value by StringTie [97]. Furthermore, EdgeR [98] was applied to detect DEGs in pairwise comparisons with a threshold of a false discovery rate (FDR) ≤ 0.001 and |log2(fold change) | ≥ 2.

Reliable Analysis by qRT-PCR
To confirm the reliability of the RNA-seq data, we randomly chose 12 expressed genes for qRT-PCR experiments. Transcript One-Step gDNA Removal and a cDNA Synthesis Supermix Kit (Transgen Biotech, Beijing, China) were used to synthesize cDNA for qRT-PCR. The primer sequences used were designed by NCBI Primer-BLAST and are listed in Additional Table S2. The reactions were conducted on a Thermal Cycler CFX96 Real-Time System (Bio-Rad Laboratories, Hercules, CA, USA) together with SYBR Green qPCR Master Mix (Omega, Beijing, China). The relative expression of the abovementioned 12 genes was calculated according to the 2 −∆∆Ct method [99], with tubulin [100] and GAPDH [101] used as reference genes. The experiments were performed for three biological replicates, and each biological replicate involved three technical repeats.

Statistical Analysis
Significant differences in the mean physiological parameters, photosynthesis parameters and relative gene expression were determined by independent sample t-tests or Duncan's multiple range test via IBM Statistical Package for Social Science (SPSS) 19.0 software version 19.0 (SPSS Inc., Chicago, IL, USA). The data are presented as the mean values ± standard error (SEs).

Conclusions
We investigated the differences in the physiological and biochemical status and in gene expression between L658 and L958 to reveal the mechanisms of Pm40-regulated resistance against Bgt, as shown in Figure 6. During the pre-penetration stage, the expression of genes encoding OXO was upregulated in L658 to generate H 2 O 2 . At the same time, the inhibition of photosynthesis caused by the downregulated expression of genes encoding Cab, ATPase, and RbcS/L generated O 2-, which was dismutated into H 2 O 2 by SOD. The suppression of POD and CAT further promoted the accumulation of H 2 O 2 in L658. The first stage of H 2 O 2 accumulation was at 6-12 hpi, and the H 2 O 2 generated during this stage likely acted as a signaling molecule to provoke the second and relatively more intense burst of H 2 O 2 at 48-72 hpi. The second H 2 O 2 accumulation possibly enhanced SA levels by directly promoting the expression of SA-related genes or by indirectly inhibiting the expression of JA-/ET-related genes. Subsequently, SA likely activated the expression of NPR1, PR1, and PR5 and ultimately induced HR in L658.

Figure 6.
Presumptive schematic of the resistance mechanism synergistically regulated by photosynthesis, ROS and plant hormones in the Pm40-expressing wheat line L658. The green colour represents low expression levels, and the red colour represents high expression levels.
Supplementary Materials: Supplementary materials can be found at www.mdpi.com/xxx/s1. Figure S1. Expression patterns of DEGs involved in the SA, JA, and ET pathways in L658 and L958. The green colour represents low expression levels, and the red colour represents high expression levels. Figure S2. Expression patterns of DEGs encoding different PR proteins in L658 and L958. The green colour represents low expression levels, and the red colour represents high expression levels. Figure S3. Localization of hypersensitive cell death at interaction sites in L658 and L958 after inoculation with Bgt and the changes in main photosynthesis parameters in the two genotypes. Necrosis (dark blue) staining by trypan blue at interaction sites in L658 (A) and L958 (B) at various time points. Table S1. Summary of RNA-seq data and read mapping. Table S2. List of forward and reverse primers designed for use in qRT-PCR.

Acknowledgments:
We are grateful to Biomarker Technology Co., Ltd. (Beijing, China) for the technical assistance.

Conflicts of Interest:
The authors declare no conflict of interest. The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript, or in the decision to publish the results.  Figure S1. Expression patterns of DEGs involved in the SA, JA, and ET pathways in L658 and L958. The green colour represents low expression levels, and the red colour represents high expression levels. Figure S2. Expression patterns of DEGs encoding different PR proteins in L658 and L958. The green colour represents low expression levels, and the red colour represents high expression levels. Figure S3. Localization of hypersensitive cell death at interaction sites in L658 and L958 after inoculation with Bgt and the changes in main photosynthesis parameters in the two genotypes. Necrosis (dark blue) staining by trypan blue at interaction sites in L658 (A) and L958 (B) at various time points. Table S1. Summary of RNA-seq data and read mapping. Table S2. List of forward and reverse primers designed for use in qRT-PCR.