Cell death resulted from loss of fumarylacetoacetate hydrolase in Arabidopsis is related to phytohormone jasmonate but not salicylic acid

Fumarylacetoacetate hydrolase (FAH) catalyzes the final step in Tyr degradation pathway essential to animals but not well understood in plants. Previously, we found that mutation of SSCD1 encoding Arabidopsis FAH causes cell death under short day, which uncovered an important role of Tyr degradation pathway in plants. Since phytohormones salicylic acid (SA) and jasmonate (JA) are involved in programmed cell death, in this study, we investigated whether sscd1 cell death is related to SA and JA, and found that (1) it is accompanied by up-regulation of JA- and SA-inducible genes as well as accumulation of JA but not SA; (2) it is repressed by breakdown of JA signaling but not SA signaling; (3) the up-regulation of reactive oxygen species marker genes in sscd1 is repressed by breakdown of JA signaling; (4) treatment of wild-type Arabidopsis with succinylacetone, an abnormal metabolite caused by loss of FAH, induces expression of JA-inducible genes whereas treatment with JA induces expression of some Tyr degradation genes with dependence of JA signaling. These results demonstrated that cell death resulted from loss of FAH in Arabidopsis is related to JA but not SA, and suggested that JA signaling positively regulates sscd1 cell death by up-regulating Tyr degradation.


Results
cell death in sscd1 is uncorrelated to SA signaling although it is accompanied by up-regulation of SA-inducible PR1. The sscd1 mutant grows normally under long day (LD), but displays obvious cell death symptoms after transferred to SD for 3 days 41 . To investigate whether cell death in sscd1 is related to SA, we first analyzed expression of PR1, one of SA-inducible genes, in wild-type and sscd1 seedlings transferred from LD to SD for 1, 2 and 3 days by quantitative real-time polymerase chain reaction (RT-qPCR). As shown in Fig. 1a, no significant difference in the expression level of PR1 between wild type and sscd1 was observed before seedlings were transferred to SD or after they were transferred to SD for 1 day, however, the expression level of PR1 was significantly increased in sscd1 compared to wild type when seedlings were transferred to SD for 2 days and that this increase was much more obvious after seedlings were transferred to SD for 3 days.
Since SA-inducible gene PR1 was significantly up-regulated in the sscd1 mutant compared to wild type when seedlings were transferred from LD to SD for 2-3 days (Fig. 1a), we next measured the content of SA to investigate whether up-regulation of PR1 is resulted from accumulation of SA in the sscd1 mutant. Unexpected, the content of SA was not significantly increased in sscd1 compared to wild type before seedlings were transferred to SD or after they were transferred to SD for 2 or 3 days (Fig. 1b). Therefore, the up-regulation of PR1 in sscd1 was not related to SA.
NPR1 is a SA receptor in SA signaling 46 and expression of SA-inducible PR1 is abolished in the npr1 mutant 19 . To investigate whether loss of NPR1 influences the up-regulation of PR1 as well as the cell death in sscd1, a double mutant of sscd1 and npr1-1 was generated and then expression of PR1 was analyzed as well as the seedlings phenotype was observed. As shown in Fig. 1c, expression of PR1 was almost undetected in the npr1-1 mutant whereas it was significantly induced in the sscd1npr1 double mutant although its level was much lower than that in the sscd1 single mutant after seedlings were transferred to SD for 3 days. Furthermore, the rate of seedlings death in sscd1npr1 was similar to that in sscd1 (Fig. 1d). These results demonstrated that both up-regulation of PR1 and cell death in sscd1 are independent of NPR1 and that cell death in sscd1 is uncorrelated to SA signaling although it is accompanied by the up-regulation of SA-inducible PR1. cell death in sscd1 is accompanied by up-regulation of JA-inducible genes and accumulation of jasmonic acid. Since cell death in sscd1 is uncorrelated to SA signaling ( Fig. 1), we next investigated whether it is related to JA signaling. We first analyzed the expression of JA-inducible genes including VSP2, PDF1.2, and THI2.1 in wild-type and sscd1 seedlings which were transferred from LD to SD for 1, 2 and 3 days. The results Scientific RepoRtS | (2020) 10:13714 | https://doi.org/10.1038/s41598-020-70567-0 www.nature.com/scientificreports/ showed that the expression level of these genes was similar in wild type and sscd1 before seedlings were transferred to SD or after they were transferred to SD for 1 day, however, it was significantly increased in sscd1 compared to wild type after seedlings were transferred to SD for 2 days and that this increase was much more obvious after seedlings were transferred to SD for 3 days (Fig. 2a-c). Then, we measured the content of jasmonic acid in wild type and sscd1 before seedlings were transferred to SD or after they were transferred to SD for 2 and 3 days to investigate whether the up-regulation of these genes is resulted from the accumulation of jasmonic acid. The result showed that there was no significant difference in the content of jasmonic acid between wild type and sscd1 before seedlings were transferred to SD, however, the content of jasmonic acid was significantly increased in sscd1 compared to that in wild type after seedlings were transferred to SD for 2 days and that this increase was much more distinct after seedlings were transferred to SD for 3 days (Fig. 2d). These results indicated that the cell death in sscd1 is accompanied by both up-regulation of JA-inducible genes and accumulation of jasmonic acid and suggested that the up-regulation of JA-inducible genes is caused by the accumulation of jasmonic acid.
cell death in sscd1 is repressed by mutation of COI1. COI1 is a JA receptor in JA signaling 47 . To investigate whether JA signaling mediates the sscd1 cell death, we generated the sscd1coi1 double mutant through a cross of sscd1 with coi1-2 28 to break the JA signaling, and then observed the phenotype of seedlings. It was interesting that the phenotype of seedlings death was obviously rescued in sscd1coi1 compared to sscd1 (Fig. 3a). For example, 65% of 7-old sscd1 seedlings grown under SD were dead whereas the rate of sscd1coi1 seedlings death was only 43% (Fig. 3b). This result suggested that the cell death in sscd1 is repressed by breakdown of JA signaling through mutation of COI1 and that JA signaling positively regulates the sscd1 cell death. The content of SA in wild-type (WT) and sscd1 seedlings that were grown under LD for 3 weeks, and then transferred to SD for 0, 2 and 3 days. (c) Relative expression level of SA-inducible genes PR1 in wild-type (WT), npr1-1, sscd1 and sscd1npr1 seedlings that were grown under LD for 3 weeks, and then transferred to SD for 3 days. (d) The rate of seedlings death in sscd1 and sscd1npr1 seedlings grown on MS under SD for 6-9 days. LD, long day; SD, short day. The expression of gene was analyzed by RT-qPCR, relative expression level was normalized to those of ACTIN2 and the control (in wild type) was set to 1. Mean ± SE from three biological replicates. Asterisk ** represents the significance of differences (two-tailed Student's t-test) at the level of P < 0.01.
Scientific RepoRtS | (2020) 10:13714 | https://doi.org/10.1038/s41598-020-70567-0 www.nature.com/scientificreports/ wild-type (WT) and sscd1 seedlings that were grown under LD for 3 weeks, and then transferred to SD for 0, 1, 2 and 3 days. (d) The content of jasmonic acid in wild-type (WT) and sscd1 seedlings that were grown under LD for 3 weeks, and then transferred to SD for 0, 2 and 3 days. LD, long day; SD, short day. The expression of genes was analyzed by RT-qPCR, relative expression level was normalized to those of ACTIN2 and the control (in wild type) was set to 1. Mean ± SE from three biological replicates. Asterisk * and ** represent the significance of differences (two-tailed Student's t-test) at the levels of P < 0.05 and P < 0.01, respectively. found that ROS marker genes such as APX2, OXI1, BAP1 and ZP were up-regulated before an occurrence of cell death in the sscd1 mutant 48 , so, we next investigated whether the repression of cell death in sscd1 by mutation of COI1 is correlated with the expression of these genes. Since the cell death phenotype of sscd1 seedlings that were grown under SD appeared on the 6th day 41,49 , therefore, we tested the expression of APX2, OXI1, BAP1 and ZP in seedlings grown under SD for 5 days. As shown in Fig. 4, the expression pattern of APX2, OXI1, BAP1 and ZP was similar in both WT and coi1-2, however, the up-regulation of these genes in sscd1 was significantly suppressed in sscd1coi1 (Fig. 4), which indicated that the up-regulation of ROS marker genes in sscd1 could be suppressed by the mutation of COI1.
SUAC treatment activates the expression of JA-inducible genes. Previously, we speculated that the cell death in sscd1 is resulted from the accumulation of SUAC and also found that treatment of Arabidopsis wild-type seedlings with SUAC mimicked the cell death phenotype of sscd1 41 . We next investigated whether SUAC treatment activates the expression of JA-inducible genes. To this end, we analyzed the expression of VSP2 and THI2.1 in wild-type seedlings treated with SUAC, in which some leaves started wilting. The result showed the expression of both VSP2 and THI2.1 was significantly increased upon SUAC treatment (Fig. 5), indicating that SUAC treatment could activate the expression of JA-inducible genes.
Treatment with MeJA causes the COI1-dependent up-regulation of some Tyr degradation pathway genes. Since the cell death in sscd1 is accompanied by the accumulation of jasmonic acid (Fig. 2d) and could be repressed by breakdown of JA signaling through mutation of COI1 (Fig. 3), we next investigated whether treatment of Arabidopsis wild-type and coi1-2 seedlings with MeJA influences the Tyr degradation pathway by analyzing the expression of Tyr degradation pathway genes including TAT3, HGO, MAAI, and SSCD1.
The results showed that the expression level of TAT3, HGO, and MAAI except SSCD1 was significantly increased in wild type upon MeJA treatment (Fig. 6), especially, an increase of TAT3 expression level in wild type treated with MeJA was much more significant compared with HGO and MAAI (Fig. 6a-c). However, it was interesting that the expression level of these genes was not significantly increased in the coi1-2 mutant upon MeJA treatment (Fig. 6). These results suggested that MeJA up-regulates the expression of some Tyr degradation pathway genes, which would promote Tyr degradation, however, the breakdown of JA signaling through mutation of COI1 could eliminate an effect of JA on Tyr degradation pathway.

Discussion
Tyr degradation pathway is essential to animals 42 but it is not well understood in plants. Previously, we found that mutation of SSCD1 encoding Arabidopsis FAH, an enzyme catalyzing the final step of Tyr degradation pathway, results in spontaneous cell death under SD, which uncovered an important role of Tyr degradation pathway in plants 41 . Afterwards, we found that sugar suppresses cell death caused by disruption of FAH in Arabidopsis, indicating that Tyr degradation is regulated by sugar in plants 49 . Recently, we found that cell death resulted from loss of FAH in sscd1 is related to chlorophyll (Chl) biosynthesis, suggesting a crosstalk between Tyr degradation and Chl biosynthetic pathways in mediating the sscd1 cell death 48 . Phytohormones such as SA and JA are involved in PCD [13][14][15][16]24,25,50 . In this study, the investigation whether cell death resulted from loss of FAH in Arabidopsis is related to SA and JA would expand our understandings on the regulation of Tyr degradation pathway in plants.
Through testing expression of SA-inducible PR1 and content of SA, we found that cell death in sscd1 was accompanied by the up-regulation of SA-inducible PR1 (Fig. 1a), however, the content of SA was not significantly altered between in sscd1 and wild type (Fig. 1b), which indicated that the up-regulation of PR1 in sscd1 is independent of SA. Similarly, an increase of PR1 expression in the loh1 mutant displaying spontaneous cell death phenotype is also independent of SA 51 . Breakdown of SA signaling by mutation of NPR1 that encodes a receptor of SA 46 represses expression of PR1 19 . In our study, the expression of PR1 was also repressed in sscd1npr1 compared to sscd1 (Fig. 1c), however, the rate of seedlings death was similar in sscd1npr1 and sscd1 (Fig. 1d), suggesting that the cell death in sscd1 is uncorrelated to both SA signaling and the up-regulation of PR1. In addition, we also generated the sscd1nahG double mutant by crossing sscd1 with nahG harboring a bacterial gene encoding salicylate hydroxylase that catalyzes the decarboxylation of SA 52,53 and found that the degree of cell death was similar between sscd1nahG and sscd1 (data not shown), indicating that the degradation of SA would not affect the cell death in sscd1, which further confirmed the sscd1 cell death is not related to SA.  THI2.1 (b) in wild-type seedlings that were grown under LD for 3 weeks, and then transferred to SD and treated with ddH 2 0 or 1,280 μg/mL SUAC for 3 days. SUAC, succinylacetone; LD, long day; SD, short day. The expression of genes was analyzed by RT-qPCR, relative expression level was normalized to those of ACTIN2 and the control (without SUAC treatment) was set to 1. Mean ± SE from three biological replicates. Asterisk * and ** represent the significance of differences (two-tailed Student's t-test) at the levels of P < 0.05 and P < 0.01, respectively. Scientific RepoRtS | (2020) 10:13714 | https://doi.org/10.1038/s41598-020-70567-0 www.nature.com/scientificreports/ However, cell death in sscd1 was accompanied by the up-regulation of JA-inducible genes as well as the accumulation of jasmonic acid (Fig. 2). The up-regulation of JA-inducible genes in sscd1 should be resulted from the accumulation of jasmonic acid, but why the cell death of sscd1 is accompanied by the accumulation of jasmonic acid? In animals, loss of FAH results in the accumulation of Tyr degradation pathway's abnormal metabolite SUAC that is toxic to cells and tissues resulting in severe metabolic disorder diseases 42 . In plants, we have found that treatment of Arabidopsis wild-type seedlings with SUAC mimicked the sscd1 cell death phenotype 41 and demonstrated that the cell death of sscd1 seedlings correlates with the accumulation of SUAC 54 . Recently, we found that SUAC affects Chl biosynthesis, resulting in the generation of ROS and then inducing cell death 48 . Some researcher's work has shown that JA could be synthesized in response to singlet oxygen that is one form of ROS 25,55 . Singlet oxygen is very unstable and difficult to detect within a cell 55 , however, some genes were specifically induced by singlet oxygen 40 . Recently, we found that the genes induced specifically by singlet oxygen 40 were up-regulated in sscd1 48 , suggesting that an effect of SUAC on Chl biosynthesis results in the generation of singlet oxygen in the sscd1 mutant. Furthermore, we found that treatment of Arabidopsis wild-type seedlings with SUAC activated the expression of JA-inducible genes (Fig. 5). Taken together, we concluded that cell death in sscd1 was accompanied by the accumulation of JA (Fig. 2d) is due to the synthesis of JA in response to singlet oxygen.
TAT catalyzes the first step in Tyr degradation pathway 56 . For the first time, Titarenko et al. 57 reported that TAT could be induced by wounding as well as by JA. The gene for the F-box protein COI1 was identified for its irreplaceable role in JA signal transduction [26][27][28] . Mutations in the COI1 gene result in plants compromised in all known JA responses: defense against biotic and abiotic stresses, growth inhibition, and fertility [26][27][28] . Titarenko et al. 57 reported that wounding induced TAT in wild type but not in the coi1 mutant, suggesting that woundinduced TAT is dependent on JA signaling. Brosché and Kangasjärvi 58 reported that expression of TAT3 encoding Arabidopsis putative TAT 45 was induced by JA. In this study, we not only confirmed that expression of TAT3 was induced by JA (Fig. 6a) but also found that expression of some of Tyr degradation pathway's genes including HGO and MAAI was also induced by JA (Fig. 6b,c), however, the expression of these genes in the coi1-2 mutant was not significantly induced by JA (Fig. 6a-c), which suggested that JA signaling up-regulates Tyr degradation in plants.
JA plays an important role in cell death regulation. Singlet oxygen-and JA-mediated cell death in irradiated flu plants is likely to be a form of PCD 59 . Inactivation of the EXECUTER1 protein abrogates not only singlet  HGO (b), MAAI (c) and SSCD1 (d) in wild-type (WT) and coi1-2 seedlings that were grown under LD for 7 days, and then transferred to SD and treated with ddH 2 0 or 100 μM MeJA for 3 days. MeJA, methyl jasmonate; LD, long day; SD, short day. The expression of genes was analyzed by RT-qPCR, relative expression level was normalized to those of ACTIN2 and the control (in wild type without MeJA treatment) was set to 1. Mean ± SE from three biological replicates. Asterisk * and ** represent the significance of differences (two-tailed Student's t-test) at the levels of P < 0.05 and P < 0.01, respectively. Scientific RepoRtS | (2020) 10:13714 | https://doi.org/10.1038/s41598-020-70567-0 www.nature.com/scientificreports/ oxygen-mediated cell death of flu plants but also accumulation of JA, however, inactivation of JA biosynthesis in the aos/flu double mutant does not affect singlet oxygen-mediated cell death 55 , hence, JA does not act as second messengers during singlet oxygen-mediated cell death but forms an integral part of a stress-related signaling cascade activated by singlet oxygen that encompasses several signaling pathways known to be activated by abiotic and biotic stressors 55 . In our study, the cell death of sscd1 seedlings was repressed by mutation of COI1 (Fig. 3). Accordingly, the up-regulation of ROS-inducible genes APX2 and OXI1, as well as singlet oxygen specifically induced genes BAP1 and ZP was also repressed by mutation of COI1 (Fig. 4), suggesting that the breakdown of JA signaling reduces the generation of ROS in the sscd1 mutant. We have just discussed above that JA signaling upregulates Tyr degradation. Therefore, the accumulation of JA in sscd1 would promote cell death by up-regulating Tyr degradation producing more SUAC. However, blockage of JA signaling by mutation of COI1 breaks the action of JA in Tyr degradation in sscd1, resulting in repression of cell death. Taken all above together, we concluded that cell death resulted from loss of FAH in Arabidopsis is related to JA but not SA, and proposed a model for the relationship between JA and Tyr degradation pathway in mediating the sscd1 cell death. In the sscd1 mutant, the accumulation of SUAC results in the generation of singlet oxygen, which induces cell death as well as JA synthesis. The accumulation of JA in sscd1 accelerates Tyr degradation by up-regulating Tyr degradation pathway, producing more SUAC, which promotes cell death. Once JA signaling is broken by mutation of COI1, the up-regulation of Tyr degradation by JA in sscd1 is eliminated, reducing production of SUAC, as a result, the sscd1 cell death is repressed.

Methods
Plant material and growth conditions. The sscd1 mutant was isolated previously in our laboratory 41 . For RT-qPCR analysis and determination of SA and jasmonic acid in Figs. 1 and 2, the seeds were germinated on MS medium and grown under LD for 1 week and then the seedlings were transplanted to a new MS medium for additional 2 weeks' growth under LD, and then transferred to SD.
Construction of double mutants. The sscd1coi1 double mutant was created by first selecting F 2 individuals from a cross between sscd1 and coi1-2 on plates containing 25 mM MeJA by screening for decreased sensitivity to JA 28 , and then F 3 lines were selected by sequencing the SSCD1 gene 41 . The primers for sequencing the SSCD1 gene are as follows: forward primer is 5′-CCT CGT CCT GCC GTC GCT AT-3′ and reverse primer is 5′-CTT GTG GAT GGC CCT GAC CT-3′.
The sscd1npr1 double mutant was created by selecting F 2 individuals from a cross between sscd1 and npr1-1 (a recessive mutation with a single base mutation in NPR1 19 ) by sequencing SSCD1 and NPR1, respectively. The primers for sequencing the NPR1 gene are as follows: forward primer is 5′-GTG TGC TCT TCA TTT CGC TGTTG-3′ and reverse primer is 5′-ACC CGG TGA TGT TCT CTT CGTA-3′.
RT-qPCR was performed in 96-well blocks using a SYBR qPCR mix (ROCHE, https ://lifes cienc e.roche .com/) with a BIO-RAD CFX CONNECT Real-Time PCR detection system (https ://www.biora d.com/) following the manufacturer's instructions. The RT-qPCR amplifications were performed under the following conditions: initial denaturation at 95 °C for 10 min, followed by 40 cycles of 95 °C for 15 s and 60 °C for 60 s. The primers of genes tested by RT-qPCR are listed in Table 1, and ACTIN2 was used as an internal control. The gene expression for each sample was calculated on three analytical replicates, and the relative expression was quantified using the 2 −ΔΔCt method. The experiment was performed in three independent biological repeats. The significance of differences between datasets was evaluated using the two-tailed Student' t-test.
Determination of the dead seedlings. Seedlings of sscd1 and sscd1npr1 were grown under SD and the number of dead seedling (all leaves were completely bleached) was counted from day 6 to 9. Seedlings of sscd1 and sscd1coi1 were grown under SD for 7 days and the number of dead seedlings was counted. The rate of seedling death was calculated as the percentage of dead seedlings from 250 to 300 seedlings. At least three independent biological repeats were performed.
Detection of jasmonic acid and SA. 0.5 g of leaves from WT and sscd1 seedlings that were grown under LD for 3 weeks and then transferred to SD for 0, 2 and 3 days was harvested for jasmonic acid and SA extraction. The harvested tissues were immediately ground to a fine powder in liquid N 2 , and then exposed to extraction buffer (1.0 mL of 80% methanol) at 4 °C overnight. The samples were centrifuged at 10,000g for 5 min, and the Table 1. Primers of genes tested by RT-qPCR.