The unfolded protein response (UPR) is required for protein homeostasis in the endoplasmic reticulum (ER) when plants are challenged by adverse environmental conditions. Inositol-requiring enzyme 1 (IRE1), the bifunctional protein kinase / ribonuclease, is an important UPR regulator in plants mediating cytoplasmic splicing of the mRNA encoding the transcription factor bZIP60. This activates the UPR signaling pathway and regulates canonical UPR genes. However, how the protein activity of IRE1 is controlled during plant growth and development is largely unknown. In the present study, we demonstrate that the nuclear and Golgi-localized protein BLISTER (BLI) negatively controls the activity of IRE1A/IRE1B under normal growth condition in Arabidopsis. Loss-of-function mutation of BLI results in chronic up-regulation of a set of both canonical UPR genes and non-canonical UPR downstream genes, leading to cell death and growth retardation. Genetic analysis indicates that BLI-regulated vegetative growth phenotype is dependent on IRE1A/IRE1B but not their canonical splicing target bZIP60. Genetic complementation with mutation analysis suggests that the D570/K572 residues in the ATP-binding pocket and N780 residue in the RNase domain of IRE1A are required for the activation of canonical UPR gene expression, in contrast, the D570/K572 residues and D590 residue in the protein kinase domain of IRE1A are important for the induction of non-canonical UPR downstream genes in the BLI mutant background, which correlates with the shoot growth phenotype. Hence, our results reveal the important role of IRE1A in plant growth and development, and BLI negatively controls IRE1A’s function under normal growth condition in plants.
When unfolded or misfolded proteins are accumulated in the ER, a much conserved response, called the unfolded protein response (UPR), is elicited to lighten the load of unfolded proteins in the ER by bringing the protein-folding and degradation capacities into alignment with the protein folding demands. However, over-activation of the UPR pathways under normal growth conditions affects plant growth and development. The bifunctional protein kinase / ribonuclease protein IRE1 is important for UPR gene regulation, but how IRE1’ protein activity is tightly controlled in plants is currently unknown. Here we report that BLISTER (BLI) negatively controls the IRE1’s function under normal growth condition in Arabidopsis. Through genetic analysis, our results also provide novel insights into how the protein kinase domain and ribonuclease domain contribute to the function of IRE1A in downstream gene expression.
Citation: Hong Z-H, Qing T, Schubert D, Kleinmanns JA, Liu J-X (2019) BLISTER-regulated vegetative growth is dependent on the protein kinase domain of ER stress modulator IRE1A in Arabidopsis thaliana. PLoS Genet 15(12): e1008563. https://doi.org/10.1371/journal.pgen.1008563
Editor: Li-Jia Qu, Peking University, CHINA
Received: June 12, 2019; Accepted: December 12, 2019; Published: December 23, 2019
Copyright: © 2019 Hong et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: All relevant data are within the manuscript and its Supporting Information files.
Funding: This project was financially supported by grants from the National Natural Science Foundation of China (31625004 and 31872653), the 111 Project (B14027), and the PLUS3 grant from the Boehringer Ingelheim Foundation. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscrip.
Competing interests: The authors have declared that no competing interests exist.
Protein folding in the ER is a fundamental process in eukaryotic cells. Protein folding demands on the secretory pathway escalate constantly during different developmental stages and environmental conditions while protein folding capacity is limited depending on ER chaperones, oxidoreductases, N-glycosylation etc. [1,2]. When protein folding demands exceed the protein folding capacity in the ER, unfolded or misfolded proteins accumulate, triggering the UPR pathway in order to recover protein homeostasis in the ER . Like yeast and mammalian cells, plant cells are also equipped with a set of membrane-associated transcription factors for sensing/transducing ER stress signals, although the sequence similarities among these factors are modest [4–14]. These membrane-associated transcription factors are activated in a similar way to those in yeast and mammals. In Arabidopsis, both bZIP28 and bZIP60 are ER-membrane-associated basic Leucine Zipper (bZIP) transcription factors. Upon ER stress induced by chemicals or abiotic stresses such as heat stress, bZIP28 relocates from ER to Golgi, where it is subjected to proteolysis by Golgi-resident Site-2 Protease (S2P). This cleavage releases its cytoplasmic N-terminal region that contains a DNA-binding domain, a nuclear localization signal and a transcriptional activation domain for downstream stress responsive gene expression [4,15–20]. The activation of bZIP60 is distinct and requires unconventional splicing of bZIP60 mRNA executed by the ER-membrane-associated protein IRE1 containing a protein kinase domain and a ribonuclease (RNase) domain in its C-terminus. Under ER stress conditions, the activated IRE1 recognizes the double stem-loop structure on bZIP60 mRNA and splices out a 23-base nucleotides, which results in a reading frame shift and elimination of the transmembrane domain of encoded bZIP60 [10,14,21]. When the soluble form of bZIP60 enters the nucleus, it activates downstream ER stress responsive genes [8,22]. Beside specifically targeting bZIP60 mRNA, IRE1 becomes more promiscuous under severe ER stress conditions, and attacks other mRNAs through a process called Regulated IRE1-Dependent Decay (RIDD), which is thought to be important for ER stress tolerance in Arabidopsis . The UPR has paradoxical outputs, cytoprotective effects for protein homeostasis and cytotoxic effects to induce programmed cell death, depending on the intensity and duration of the stimulus that the organism, tissues or cells are receiving . Over-activation of the UPR pathways affects plant growth and development. For example, constitutive over-expression of the activated form of bZIP28 in Arabidopsis induces UPR genes and results in delayed growth and development . Recently, it was reported that mutation of the rice transcription factor SQUAMOSA PROMOTER-BINDING PROTEIN-LIKE 6 (SPL6) causes up-regulation of IRE1 expression and persistent UPR, leading to cell death and apical panicle abortion in rice . However, how IRE1 activity is controlled under normal growth conditions in plants is currently unknown.
In the current study, we found that BLISTER (BLI) negatively regulates IRE1A function in Arabidopsis. Mutation of BLI results in the activation of IRE1A and up-regulation of both canonical and non-canonical UPR genes under normal growth conditions, however, the IRE1-bZIP60 pathway is not responsible for the vegetative growth retardation phenotype in the BLI mutant plants. Further studies demonstrated that the residues D570/K572 and D590 in the protein kinase domain of IRE1A are important for non-canonical UPR gene expression and growth inhibition, while the D570/K572 residues and N780 residue of IRE1A are required for its ribonuclease activity to induce canonical UPR downstream genes. Together, these results demonstrated the important role of IRE1A in growth and development, and revealed BLI as a new negative regulator of IRE1 in Arabidopsis.
Mutation of BLI results in up-regulation of UPR genes and growth retardation phenotype under normal growth conditions
Previously we identified a plant specific protein BLISTER (BLI), which interacts with the Polycomb-group (Pc-G) histone methyltransferase CURLY LEAF (CLF) and controls the expression of Pc-G target genes for cellular differentiation in Arabidopsis thaliana . Mutations of BLI in two loss-of-function Arabidopsis mutant alleles (bli-1 and bli-11) conferred growth retardation [26,27] and up-regulation of some canonical ER stress responsive genes under normal growth conditions, for examples, BiP3, NSF, ERDJ3A, ERDJ3B and SARA1A (S1A Fig). The spliced form of bZIP60 mRNA, encoding the activated nuclear form of bZIP60, was also dramatically up-regulated in both bli-1 and bli-11 mutant plants, while the unspliced bZIP60 mRNA was less up-regulated (S1A Fig), suggesting that bZIP60 is activated in the BLI mutant plants. However, the expression of neither IRE1A nor IRE1B is affected in the BLI mutant plants (S1A Fig). These results supports that UPR pathway is turned on when BLI is mutated in Arabidopsis, and BLI is a negative regulator of the ER stress response. Previously, we demonstrated that BLI-GFP was mainly observed in the nucleus with cytoplasmic signals in the transient expression system in tobacco (N. benthamiana) epidermal leaves . Our previous analyses revealed that ER stress modulators, such as bZIP28, bZIP60, NAC062 and NAC089, localize to either ER membrane or plasma membrane in plants when they are inactive [4–7], so we re-checked the subcellular localization of BLI in stably transformed Arabidopsis plants. To do so, BLI-GFP was co-expressed with the ER marker WAK2-RFP, Golgi marker SYP32-RFP or the trans-Golgi/early endosome marker VTI12-RFP, or the nucleus marker H2B-RFP  in Arabidopsis plants. Confocal microscopy revealed that BLI-GFP not only localized to nucleus but also co-localized with SYP32-RFP but not with VTI12-RFP or WAK2-RFP in root cells (Fig 1A–1R). We conclude that BLI localizes to the Golgi and nucleus, and negatively regulates the expression of ER stress responsive genes in plants under normal growth condition.
A-R, Subcellular localization of BLI in Arabidopsis bli-11 mutant plants. BLI:BLI-GFP was co-expressed with the Golgi marker SYP32-RFP, the trans-Golgi/early endosome marker VTI12-RFP, the ER marker WAK2-RFP, or the nucleus marker H2B-RFP. Root cells in the elongation zone (A-C, G-I, M-R) or meristematic zone (D-F, J-L) were observed under a confocal microscope and merged. Close-up pictures are shown in A-L. Scale bars are 20 μm.
IRE1 Has UPR-independent roles in BLI-regulated plant growth and development
To investigate whether the growth phenotype of BLI mutant plants depends on ER stress response modulators, we crossed bli (bli-1 hereafter) to either bzip28 or bzip60 and generated bli bzip28 and bli bzip60 double mutants. The bzip28 and bzip60 single mutants grew as normally as the WT (S2A and S2B Fig), and the expression of ER stress responsive genes such as BiP3, ERDJ3A and SARA1A in bzip28 and bzip60 single mutant seedlings was slightly changed (within one fold change) comparing to that in WT plants under normal growth conditions (S2C Fig). However, when crossed to the bli mutant, the bli bzip28 double mutant resembled the bli single mutant phenotype in terms of shoot growth (Fig 2A) and silique length (Fig 2B). The bli bzip60 double mutant also resembled the bli single mutant phenotype at the vegetative stage (Fig 2A), but the silique length in the bli bzip60 double mutant was further reduced comparing to that in the bli single mutant (Fig 2B). These results suggest that the growth retardation phenotype of bli at the seedling stage does not depend on bZIP28 or bZIP60 alone. Further gene expression analysis showed that the expression of ER stress marker genes was not up-regulated in the bli bzip60 double mutant seedlings, but still up-regulated in bli bzip28 double mutant seedlings comparable to the inducing level in bli single mutant (Fig 2C). In our previous study, mutation of bZIP28 reduces the up-regulation of BiP3, ERDJ3A, ERDJ3B and SARA1 under ER stress condition . ERDJ3B was considered as a bZIP28-specific downstream gene since knock-out bZIP28 almost abolished the up-regulation of ERDJ3B during UPR in Arabidopsis . However, the expression of ERDJ3B was only slightly up-regulated in bli mutant plants (Fig 2C and S1 Fig). The differences between bli and bli bzip28 mutants in terms of up-regulation of BiP3, NSF, ERDJ3A, SARA1 and NAC103 were also very modest. Nevertheless, our results strongly suggest that bZIP60 is activated in bli mutant. As the up-regulation of the analyzed ER stress responsive genes in the bli mutant depends on bZIP60, and bli bzip60 mutants are reminiscent of bli mutants at the seedling stage, it is likely that mis-expression of the analyzed ER stress genes does not contribute to the bli phenotype at seedling stage.
A-B, Phenotypic analysis. T-DNA mutant bli (bli-1) was crossed to either bZIP28 mutant (bzip28) or bZIP60 mutant (bzip60) to generate the respective double mutant plants. 2-week-old plant seedlings were photographed (A) and siliques lengths were measured later at reproductive stage (B). Error bars represent SD (n = 10). Letters above the bars in (B) indicate significant differences as determined by LSD test following ANOVA analysis (p<0.05). Bar = 5 mm. C, UPR gene expression analysis. Total RNA was exacted from 2-week-old plants for qRT-PCR. Fold change is the gene expression level in the mutants normalized to that in the WT, both of which were normalized to the expression of ACTIN. Asterisks indicate significance levels when comparing to the WT control in t-test. (*, p<0.05; **, p<0.01; n.s., not significant at p<0.05).
bZIP60 is an Arabidopsis ER-membrane-associated transcription factor which is activated by ER-localized IRE1A/IRE1B through unconventional splicing during the ER stress response [5,10,14]. IRE1A/IRE1B contain both protein kinase domains and RNase domains at the cytoplasm-facing side. The RNase activity of IRE1B is required for ER stress tolerance, whereas both protein kinase and RNase activity are required for normal root growth . To analyze the genetic interaction of BLI and IRE1A/IRE1B, we crossed bli to either ire1a or ire1b mutant to generate the bli ire1a and bli ire1b double mutants (S3 Fig). The ire1a and ire1b single mutants did not have obvious phenotypes when compared to the WT (S2A and S2B Fig), and mutation of neither IRE1A nor IRE1B altered the expression of ER stress responsive genes under normal growth conditions (S2C Fig). However, mutation of either IRE1A or IRE1B largely suppressed the bli mutant phenotype at seedling stage in terms of seedling size (Fig 3A), but not at reproductive stage in terms of silique length (Fig 3C and 3D). However, the leaves of bli ire1a and bli ire1b plants were thinner than that of the WT plants. Meanwhile, the root growth retardation phenotype in bli was not suppressed in bli ire1a and bli ire1b mutants under normal growth condition (S4 Fig). Therefore, the fresh weight of whole plants was only partially suppressed in the bli ire1a and bli ire1b mutant plants (Fig 3B). Loss of BLI function resulted in cell death, as revealed by Trypan Blue staining (Fig 3E), and ROS accumulation, as revealed by DAB staining (Fig 3F). Mutation of either IRE1A or IRE1B also suppressed cell death and ROS level in the bli mutant seedlings (Fig 3E and 3F). To investigate whether BLI is involved in the ER stress response, the sensitivity of WT plants, bli, ire1a, and ire1b single mutants, as well as bli ire1a and bli ire1b double mutants were checked. In the presence of tunicamycin (TM), an N-glycosylation inhibitor specifically inducing misfolded protein accumulation in ER, in the growth medium, the bli-1 mutant plants were very sensitive to ER stress in terms of shoot growth, leaf chlorosis and root growth, when compared to the WT, which could be partially rescued in the bli ire1a and bli ire1b double mutant plants (S4 Fig), although the expression of BLI is not affected by TM treatment (S2C Fig). Unfortunately, we could not obtain bli ire1a ire1b triple mutant plants. Arrested embryos from the ire1a ire1b bli+/- self-crossing siliques were observed (S5 Fig), suggesting that bli ire1a ire1b triple mutant is lethal. Taken together, these results demonstrate that BLI has both IRE1 dependent and independent roles in growth and development in Arabidopsis.
A-D, Phenotypic analysis. T-DNA mutant bli (bli-1) was crossed to either IRE1A mutant (ire1a) or IRE1B mutant (ire1b) to generate the respective double mutant plants. 2-weeks-old plants were photographed (A), weighted (B). Representative siliques were photographed (C) and siliques lengths were measurement at reproductive stage (D). Error bars represent SE (n = 3) in (B) and represent SD (n = 12) in (D). Letters above the bars indicate significant differences as determined by LSD test following ANOVA analysis (p< 0.05). Bar = 5 mm. E-F, Cell death and ROS accumulation analysis. 2-weeks-old plants were stained with trypan blue (E) or DAB (F). Arrows and arrow heads point to sites of cell death and ROS accumulation, respectively. Bar = 1 mm.
To understand how IRE1 is involved in the bli-regulated shoot growth, we performed RNA-Seq analysis with WT, bli, bli ire1a and bli ire1b seedlings grown under normal growth conditions. Compared to the WT, 867 and 1113 genes were significantly up-regulated and down-regulated in the bli mutant, respectively (Fig 4A and S1C Fig). Among them, the expression of 359 up-regulated genes (i.e. 41.4%) and 599 down-regulated genes (i.e. 53.8%) was also affected by ER stress inducer tunicamycin (5 μg/ml, 12 hours) in WT plants (S1C Fig), further demonstrated that the canonical UPR pathway is activated in bli mutant seedlings. Interestingly, 331 up-regulated genes and 251 down-regulated genes in the bli mutant were not mis-regulated in bli ire1a or bli ire1b seedlings (Fig 4A and S1 Dataset). Surprisingly, up-regulation of canonical UPR responsive genes in the bli mutant was suppressed in the bli ire1a double mutant but not in bli ire1b double mutants, whereas the up-regulation of non-canonical ER stress responsive genes such as MRN1 and LTP4 was suppressed in both the bli ire1a and bli ire1b mutant (Fig 4B and S1 Dataset). This suggests that the RNase activity of IRE1A is more activated than that of IRE1B in bli mutant plants, or the RNase activity of IRE1B is dependent on IRE1A in bli mutant plants under normal growth condition. Nevertheless, our results show that both IRE1A and IRE1B, but not bZIP60, are required for the shoot growth phenotype in bli mutant, and the bli shoot growth phenotype is probably attributed to the mis-expression of non-canonical UPR genes.
A-B, Gene expression analysis. T-DNA mutant bli (bli-1) was crossed to either IRE1A mutant (ire1a) or IRE1B mutant (ire1b) to generate the respective double mutant plants. Total RNA was exacted for RNA-Seq analysis (A) and qRT-PCR (B). Up-regulated or down-regulated genes in the mutants relative to the wild-type plants were used to draw the Venn diagrams. Fold change is the gene expression level in the mutants normalized to that in the WT, both of which were normalized to the expression of ACTIN. Error bars represent SE (n = 3). Asterisks indicate significance levels when comparing to the WT control in t-test. (*, p<0.05; **, p<0.01; n.s., not significant at p<0.05). bZIP60S, spliced bZIP60 transcript. Note that bZIP60-S, BiP3, NSF, DERD3A, SARA1A and NAC103 are up-regulated while MRN1 and LTP4 are not up-regulated by canonical ER stress inducer tunicamycin.
Residues in the protein kinase domain of IRE1A are important for BLI-regulated vegetative growth
It is known that IRE1B has a role in vegetative growth , we focused more on characterization of IRE1A in the current paper. Previously, it was shown that in yeast, D797N K799N (called 1KR32) mutations in the nucleotide-binding pocket abolish autophosphorylation and transautophosphorylation in in vitro kinase assays but retain the RNase activity in in vitro RNase assays . D828A mutation in yeast abolishes autophosphorylation in in vitro kinase assays, although it retains ATP binding activity . N1057A mutation of yeast IRE1 in the RNase domain only abolished its RNase activity . We made equivalent mutations of IRE1A (Fig 5A and S6 Fig) and analyzed the autophosphorylation of IRE1A in vitro. Calf intestinal alkaline phosphatase (CIAP) is widely used to eliminate the phosphate group from phosphorylated proteins, and molecular weight shift after CIAP treatment has been successfully used for detection the phosphorylation status of mammalian IRE1 . We have taken this advantage and detected IRE1A autophosphorylation on Phosphate-affinity polyacrylamide gel (Phos-tag) electrophoresis following CIAP treatment (Fig 5B). D590A (corresponding to yeast D828A) and D570N K572N (corresponding to yeast D797N K799N) mutations abolished the autophosphorylation of IRE1A, while N780A (corresponding to yeast N1057A) did not (Fig 5B). We then expressed these mutated forms of IRE1A or the non-mutated form of IRE1A driven by the IRE1A native promoter in the bli ire1a double mutant background (S7 Fig). Phenotypic analysis showed that expression of the non-mutated form of IRE1A (IRE1A) and the RNase dead form of IRE1A (N780A) reverted the bli ire1a phenotype to the bli phenotype (Fig 5C and 5D). However, expression of the kinase dead forms of IRE1A (D590A and D570N K572N) largely resembled the bli ire1a phenotype (Fig 5C and 5D). These results demonstrate that the protein kinase domain but not the RNase domain of IRE1A is required for BLI-regulated shoot growth phenotype. Further gene expression analysis showed that the up-regulation of canonical UPR marker genes was restored when IRE1A and D590A were expressed (Fig 6). In contrast, the up-regulation of UPR marker genes was not restored when N780A and D570N K572N were expressed (Fig 6). However, the up-regulation of non-canonical UPR marker genes (MRN1 and LTP4) was restored when IRE1A and N780A were expressed (Fig 4D). Thus, the protein kinase domain of IRE1A is required for the bli mutant phenotype and up-regulation of non-canonical UPR genes such as MRN1 in the bli mutant. D570N K572N mutation in the ATP-binding pocket and N780A mutation in the RNase domain affect IRE1A’s RNase activity to induce canonical UPR gene expression.
A-B, Site-specific mutagenesis analysis of the IRE1A auto-phosphorylation. Domain structure of IRE1A is shown in (A) with the mutated amino acids highlighted. The wild-type form and various mutated forms of IRE1A were expressed in E. coli and purified proteins were treated with CIAP for de-phosphorylation assays. Western blotting analysis was done to detect the molecular weight shifts. C-D, Genetic complementation analysis. T-DNA mutant bli (bli-1) was crossed to either IRE1A mutant (ire1a) or IRE1B mutant (ire1b) to generate the respective double mutant plants. The wild-type form and various mutated forms of IRE1A were expressed the bli ire1a double mutant background and 2-weeks-old T3 transgenic plants were photographed (C) and fresh weight were measured (D). Error bars represent SE (n = 3). There were at least 10 plants for each of the three biological replicates. Letters above the bars indicate significant differences as determined by LSD test following ANOVA analysis (p<0.05). Bar = 5 mm.
T-DNA mutant bli (bli-1) was crossed to either IRE1A mutant (ire1a) or IRE1B mutant (ire1b) to generate the respective double mutant plants. The wild-type form and various mutated forms of IRE1A were expressed in the bli ire1a double mutant background and 2-weeks-old T2 transgenic plants were collected for gene expression analysis. Fold change is the gene expression level in the mutants or the genetically complemented materials normalized to that in the WT, both of which were normalized to the expression of ACTIN. Error bars represent SE (n = 3). Asterisks indicate significance levels when comparing to the WT control in t-test. (*, p<0.05; **, p<0.01; n.s., not significant at p<0.05).
UPR signaling has protective functions in cells during ER stress response by increasing the ER protein folding capacity in the cell, enhancing the protein degradation of unfolded/misfolded proteins, and probably reducing the synthesis of new proteins in ER in plants. The Inositol-requiring enzyme 1 (IRE1) is one of the most important ER transmembrane sensors that activates the UPR for protein homeostasis under ER stress conditions through its specific splicing RNA target or non-specific RNA substrates in a process called RIDD . However, IRE1’s activity needs to be timely attenuated after ER stress, and even under normal physiological conditions, IRE1’s activity needs to be finely tuned. For example, mutation of the transcription factor SPL6 resulted in increased expression level of IRE1 therefore elevated level of IRE1 protein, leading to constitutive ER stress response and developmental arrests in rice panicles . However, in the current study, neither the expression of IRE1A nor IRE1B is affected in the bli mutant. We could not exclude the possibility that the protein level of IRE1A/IRE1B is affected in the bli mutant, since generation of antibody against Arabidopsis IRE1A/IRE1B is not successful and the rice IRE1 antibody  did not recognize the Arabidopsis IRE1A/IRE1B. However, our genetic complementation experiments demonstrated that the bli phenotype at seedling stage is dependent on the kinase domain of IRE1A, suggesting that the protein activity rather than the protein level of IRE1A is more probably affected by BLI. Little is understood on how IRE1’ activity is regulated in plants. In mammalian cells, it was proposed that, under normal conditions, the ER chaperone BiP/Grp78 binds to IRE1α to keep the protein in an inactive monomeric state, under ER-stressed conditions, releasing of BiP from IRE1α allows monomerization and autophosphorylation of its cytosolic domain, which triggers the activation of the RNase activity and downstream events . However, whether similar mechanisms are existed in plants to tune the IRE1’s activity has not yet been reported. In the current study, we found that BLI is a negative regulator of IRE1 in Arabidopsis. Mutation of BLI results in activation of IRE1 and bZIP60 splicing, leading to increased expression level of canonical UPR genes such as BiP3 and non-canonical UPR genes such as MRN1 in plants. It seems that high level of BiP in bli mutant does not inhibit but activate IRE1A/IRE1B in Arabidopsis. Both Arabidopsis IRE1A and IRE1B were observed in perinuclear ER in tobacco BY2 cells . Recently, IRE1B was also shown to be distributed in ER in Arabidopsis protoplasts under normal growth condition . Given that BLI was observed in the nucleus and Golgi in Arabidopsis root cells, we hypothesized that BLI may indirectly control the activity of IRE1A/IRE1B. Further studies are needed to understand how BLI regulates IRE1’s activity in plants. Suppressor screening of bli mutant is underway in the lab.
The N-terminal part of IRE1 is in the ER lumen while the C-terminus is facing the cytoplasm which contains a protein kinase domain and an RNase domain. Previously, D608N K610N mutations of Arabidopsis IRE1B (corresponding to D797N K799N of yeast IRE1, 1KR32) in the nucleotide-binding pocket abolished IRE1B’s autophosphorylation and reduced its RNase activity to splice bZIP60 mRNA, while the D628A mutation (corresponding to D828A mutation of yeast IRE1) within the conserved DFG kinase motif only abolished IRE1B’s autophosphorylation; the N820A mutation (corresponding to N1057A mutation of yeast IRE1) in the RNase domain only abolished IRE1B’s RNase activity in Arabidopsis during ER stress response . However, how these conserved amino acids are involved in IRE1A’s activity is not reported. In the current study, we carried out experiments to characterize these important residues of IRE1A in Arabidopsis bli mutant background under normal growth conditions. Our results showed that D570A K572A mutations and N780A mutation of IRE1A reduced its ribonuclease activity for canonical UPR downstream gene expression while D570A K572A mutations and D590A mutation in the protein kinase domain of IRE1A affected its protein kinase activity for non-canonical UPR gene expression. Thus, the regulatory effect of protein kinase activity on its RNase activity is common for these two IRE1 proteins in Arabidopsis plants, which is different from yeast cells in which 1KR32 mutation in the nucleotide-binding pocket retains the RNase activity . Autophosphorylation of IRE1B  and IRE1A (this study) has been demonstrated in vitro, however, detection of the autophosphorylation of these two proteins in vivo has not yet been successful, most probably because of the low level of these proteins in plants. Given that D590A mutation in IRE1A and D628A mutation in IRE1B do not affect the RNase activity of IRE1 in Arabidopsis, these results suggest that the ATP binding site of IRE1 rather than the protein kinase activity is more important for the activation of RNase activity in plants.
Besides the important roles of IRE1 in plant ER stress responses, the root-specific phenotype of Arabidopsis ire1a ire1b double mutant plants under unstressed conditions suggests that IRE1 also have other function in integrating physiological signals to maintain specific secretory activity . Mutant analysis showed that the root growth phenotype of ire1a ire1b double mutant in the absence of stress was dependent on IRE1B . Our results revealed further that the shoot growth retardation phenotype in bli mutant under normal growth conditions were dependent on IRE1A by analyzing the advance mutants in the bli mutant background. The root growth phenotype and reproductive development defects in bli mutant were not rescued by IRE1A or IRE1B mutation, suggesting that BLI also has IRE1-independent roles in plant growth and development, or IRE1A and IRE1B are too important for reproductive development so that the bli ire1a and bli ire1b double mutants could not recover. Nevertheless, our data demonstrated that the function of IRE1A in shoot growth regulation in bli mutant dependents on neither its RNase activity, nor its splicing target bZIP60. The protein kinase domain of IRE1A is required for the regulation of growth and development in bli mutant plants, probably through regulating of non-canonical UPR downstream genes.
In summary, we identified BLI as a negative regulator of IRE1A, constraining its function under normal growth conditions in Arabidopsis (Fig 7). In our working model, BLI may directly inhibit IRE1A’s function, or regulate the expression of an unknown protein to inhibit IRE1A’s function during vegetative growth. Once BLI is mutated, the activated IRE1A induces UPR genes and other non-canonical ER stress genes, leading to cell death and growth retardation in plants. D570/K572 in the ATP-binding pocket and N780 in the RNase domain are important for IRE1A to regulate canonical UPR genes while D570/K572 and D590 are important for its auto-phosphorylation and non-canonical UPR gene expression.
In the presence of BLISTER (BLI) in wildtype (WT) plants, BLI may directly or indirectly inhibit IRE1A activation and clustering. In the absence of BLI in bli mutant plants, the inhibitor is removed and IRE1A is activated. The RNase activity of IRE1A for bZIP60 mRNA splicing and downstream canonical UPR gene expression is dependent on N780 in the RNase domain and D570/K572 in the ATP-binding pocket of IRE1A. In contrast, the auto-phosphorylation activity of IRE1A is dependent on D570/K572 and D590 of IRE1A, which is important for non-canonical UPR gene expression and the shoot growth retardation phenotype.
Materials and methods
Plant material and growth conditions
All of the Arabidopsis thaliana seeds used in this study are in the Columbia (Col) background. Seeds of WT, bli-1 (SAIL_107_D04), bli-11 (GABI-Kat_663H12), bzip28 (SALK_132285), bzip60 (SALK_050203), ire1a (SALK_018112), ire1b (SAIL_238_F07), double mutants and transgenic plants were sterilized (10 min 70% Ethanol supplemented with 0.05% Triton X-100, 10 min 96% Ethanol) and sown on germination medium (MS; half-strength Murashige and Skoog medium supplemented with 1% sucrose, 0.05% MES, and 0.8% plant agar). Seeds were stratified for two days at 4°C and grown under long day (LD) conditions, (16/8 h light / dark cycle at 22°C) in the medium or in the soil. For ER stress treatment, plants were grown on MS medium for 6 days and then transferred to either MS medium or MS plus 0.3 μg/ml tunicamycin (TM, Sigma) for additional 6 days. For silique length measurement, two longest siliques from each of the five main inflorescences were selected and measured for each genotype. For statistical analysis of gene expression, student’s t test was used. For multiple comparisons of phenotypes, LSD (Least Significant Difference) test was carried out following the ANOVA (Analysis of Variance) analysis.
Trypan blue staining and DAB staining
Trypan blue staining and DAB staining were performed as described in previous literature . For trypan blue staining, 14-day-old leaves of wild-type, bli, bli ire1a and bli ire1b were boiled for 1 min in trypan blue staining buffer containing 12.5% phenol, 12.5% glycerol, 12.5% lactic acid, 48% ethanol, and 0.025% trypan blue (Sangon Biotech), incubated for 10 min at room temperature, followed by destaining five times in 70% chloral hydrate. For DAB staining, leaves of 14-day-old wild-type, bli, bli ire1a and bli ire1b leaves were incubated in 1 mg/ mL DAB (Sigma) solution (pH 3.8) for 5 h at room temperature in darkness and then boiled in 95% boiling ethanol for 10 min before photograph.
The IRE1A genomic sequences including 2.5 kb upstream promoter region were amplified from wild-type and inserted into pCambia 1300 to obtain the IRE1A:IRE1A construct. The site-specific-mutated PCR products were amplified by overlapping PCR based on the IRE1A:IRE1A construct. All the error-free constructs were introduced into the bli+/- ire1a-/- double mutant plants via Agrobacterium-mediated transformation. The T1 seeds were grown on selective MS medium plus 30 mg/L hygromycin (Roche) and genotyped to select transgenic plants in bli-/- ire1a-/- background. The T3 seeds were obtained and grown on MS medium for phenotypic analysis along with genotyping and gene expression analysis. At least five independent lines of each construct were selected and examined. All the primers used are listed in S1 Table.
RNA-Seq and qRT-PCR
For RNA-Seq analyses, 14-days-old seeding were collected and immediately frozen in liquid nitrogen. Total RNA were extracted with Trizol (Invitrogen) and sequenced on the Illumina HiSeq 4000 platform by commercial company (Major Bio) following the standard Illumina protocols . RNA-Seq reads were aligned to the reference genome of Arabidopsis thaliana (version TAIR10) using TopHat (version 2.0.13) (Kim et al., 2013) after filtering out low-quality (lowest base score < 20) reads using SeqPrep (https://github.com/jstjohn/SeqPrep) and Sickle (https://github.com/najoshi/sickle). Totally ~5.5–10 GB clean reads for each sample were obtained. Differential gene expression was assessed using Cuffdiff (http://cole-trapnell-lab.github.io/cufflinks/cuffdiff/index.html). The cutoff for significant differential expression was set as fold change (FC) ≥2 or FC≤0.5 and p-value p≤0.05. Parameters for sequencing quality control are listed in S2 Table. For qRT-PCR, RNA from 14-day-old seedlings were extracted using an RNA Prep Pure Plant kit (Tiangen). For reverse transcription, 2 mg of RNA and oligo (dT) primers were used to synthesize cDNA in a 20 μL reaction using M-MLV reverse transcriptase (TaKaRa). qRT-PCR was performed using SuperReal PreMix Color (Tiangen) in a CFX96 real-time system (Bio-Rad). There were three biological replicates for the RNA-Seq analysis. All the primers used are listed in S1 Table.
Genomic BLI, containing the BLI coding region and 1.7 kb upstream of the transcriptional start site, was amplified from genomic DNA, and cloned into pGKGWG. The BLI:BLI-GFP construct was introduced to the bli-11 mutant background and stable transformed plants were obtained. Confocal laser scanning microscopy was performed with root tissues using LSM 780 and LSM 510 microscopes (Zeiss). Image acquisition was carried out sequentially to prevent crosstalk between channels. GFP was excited at 488 nm, and emission was detected at 510–550 nm. RFP was excited at 561 nm and emission was detected at 575–630 nm. All the primers used are listed in S1 Table.
For de-phosphorylation assay, partial IRE1A sequences (AA345-841) were amplified from the wild-type Arabidopsis cDNA and cloned into pET28 to express the His-tag IRE1A truncation protein. Various mutated forms of IRE1A were amplified by overlapping PCR based on the wild-type form. All purified proteins (2 μg) were treated with or without CIAP (Takara) for 2 min at 37°C. The reaction were stopped by adding 4 μL 5X SDS buffer and boiling at 95°C for 5 min. All the proteins were resolved in Phos-tag gels and then immunoblotted with anti-His antibody (Abmart). All the primers used are listed in S1 Table.
Sequence data from this article can be found in the Arabidopsis Genome Initiative under the following accession numbers: At3g10800 (bZIP28), AT1G42990 (bZIP60), AT2G17520 (IRE1A), AT5G24360 (IRE1B), AT3G23980 (BLI), AT1G09080 (BiP3), AT4G21730 (NSF), AT3G08970, (ERDJ3A), AT3G62600, (ERDJ3B), AT1G09180 (SARA1), AT5G64060 (NAC103), AT5G42600 (MRN1) and AT5G59310 (LTP4). The RNA-Seq data in this article can be found in the Gene Expression Omnibus (GEO) under the accession number GSE124235.
S1 Fig. Mutation of BLI induces UPR gene expression.
A, UPR gene expression analysis. Total RNA was exacted from 2-week-old plants for qRT-PCR. Fold change is the gene expression level in the BLI mutants (bli-1 or bli-11) normalized to that in the wild-type plants (WT), both of which were normalized to the expression of ACTIN. Error bars represent SE (n = 3). bZIP60U, unspliced bZIP60; bZIP60S, spliced bZIP60. Asterisks indicate significance levels when comparing to the WT control in t-test. (*, p<0.05; **, p<0.01). B, Venn diagrams showing the overlapping regulated genes between BLI-dependent genes and canonical UPR genes. Canonical UPR genes were obtained by comparing gene expression profiles of WT plants treated with or without ER stress inducer tunicamycin (5 μg/ml) for 12 hours followed by RNA-Seq analysis. Up-regulation: fold change ≥2, p<0.05; down-regulation: fold change ≤0.5, p<0.05.
S2 Fig. Mutation of either IRE1A, IRE1B, bZIP28, or bZIP60 alone does not affect vegetative growth and reproductive development.
A-B, Phenotypic analysis. T-DNA mutants of IRE1A (ire1a), or IRE1B (ire1b) or bZIP28 (bzip28) or bZIP60 (bzip60) were grown together with wild-type plants (WT) in standard MS growth medium. 2-week-old plant seedlings were photographed (A) and siliques lengths were measured at reproductive stage (B). Error bars represent SD (n = 10). Letters above the bars indicate significant differences as determined by LSD test following ANOVA analysis (p<0.05). Bar = 5 mm. C, UPR gene expression analysis. Total RNA was exacted from 2-week-old plants for qRT-PCR analysis. Error bars represent SE (n = 3). Asterisks indicate significance levels when comparing to the WT control in t-test. (*, p<0.05; **, p<0.01). bZIP60U, unspliced bZIP60; bZIP60S, spliced bZIP60.
S3 Fig. Characterization of T-DNA mutants.
Total RNA was extracted from various plant materials grown under normal growth conditions and the expression of BLI, IRE1A or IRE1B was checked by RT-PCR. UBQ5 was used as a loading control.
S4 Fig. The BLI ER stress related phenotype is partially suppressed by IRE1A or IRE1B mutation.
(A-B) ER stress related phenotype of mutant plants. T-DNA mutant of BLI (bli-1) was crossed to either IRE1A mutant (ire1a) or IRE1B mutant (ire1b) to generate the respective double mutant plants. Wild-type (WT), single mutants and double mutants were vertically grown for 6 days on MS medium, transferred to either MS or MS plus 0.3 μg/ml tunicamycin (TM) plates and grown for additional 6 days, then photographed (A). Primary root length was measured (B). There were five plants in each of the three replicates. Error bars represent SE (n = 3). Letters above the bars indicate significant differences as determined by LSD test following ANOVA analysis (p<0.05). Bar = 10 mm.
S5 Fig. Analysis of siliques from self-crossing plants.
WT and ire1a ire1b bli+/- plants were grown in soils in well-controlled growth chamber and their siliques were examined after self-crossing. Arrow heads point to the aborted embryos.
S6 Fig. Multiple sequence alignment of the IRE1’ C-termini.
Protein sequences of the Arabidopsis IRE1s (AtIRE1A, At2g17520; AtIRE1B, At5g24360), rice IRE1 (OsIRE1, LOC_Os07g28820), yeast IRE1 (ScIRE1, NP_011946) and human IRE1 (HsIRE1, NP_001424) were subjected to sequence alignment using CLUSTAL 2.1 with the default setting. Only the C-terminal regions beyond the transmembrane domains are shown. The protein kinase domain is shaded in grey while the endonuclease domain is shaded in yellow. Amino acids predicted to be important for the protein kinase activity or endonuclease activity are highlighted in red.
S7 Fig. Validation of transgene expression.
The wild-type form and various mutated forms of IRE1A were expressed the bli ire1a double mutant background. Total RNA was extracted from various plant materials grown under normal growth conditions and the expression of total IRE1A was checked by RT-PCR. UBQ5 was used as a loading control.
S2 Table. Parameters for sequencing quality control of each sample.
- 1. Liu JX, Howell SH (2016) Managing the protein folding demands in the endoplasmic reticulum of plants. New Phytol 211: 418–428. pmid:26990454
- 2. Walter P, Ron D (2011) The unfolded protein response: From stress pathway to homeostatic regulation. Science 334: 1081–1086. pmid:22116877
- 3. Liu JX, Howell SH (2010) Endoplasmic reticulum protein quality control and its relationship to environmental stress responses in plants. Plant Cell 22: 2930–2942. pmid:20876830
- 4. Liu JX, Srivastava R, Che P, Howell SH (2007) An endoplasmic reticulum stress response in Arabidopsis is mediated by proteolytic processing and nuclear relocation of a membrane-associated transcription factor, bZIP28. Plant Cell 19: 4111–4119. pmid:18156219
- 5. Deng Y, Humbert S, Liu JX, Srivastava R, Rothstein SJ, et al. (2011) Heat induces the splicing by IRE1 of a mRNA encoding a transcription factor involved in the unfolded protein response in Arabidopsis. Proc Natl Acad Sci USA 108: 7247–7252. pmid:21482766
- 6. Yang ZT, Lu SJ, Wang MJ, Bi DL, Sun L, et al. (2014) A plasma membrane-tethered transcription factor, NAC062/ANAC062/NTL6, mediates the unfolded protein response in Arabidopsis. Plant J 79: 1033–1043. pmid:24961665
- 7. Yang ZT, Wang MJ, Sun L, Lu SJ, Bi DL, et al. (2014) The membrane-associated transcription factor NAC089 controls ER-stress-induced programmed cell death in plants. PLOS Genet 10: e1004243. pmid:24675811
- 8. Iwata Y, Koizumi N (2005) An Arabidopsis transcription factor, AtbZIP60, regulates the endoplasmic reticulum stress response in a manner unique to plants. Proc Natl Acad Sci USA 102: 5280–5285. pmid:15781873
- 9. Tajima H, Iwata Y, Iwano M, Takayama S, Koizumi N (2008) Identification of an Arabidopsis transmembrane bZIP transcription factor involved in the endoplasmic reticulum stress response. Biochem Biophy Res Commun 374: 242–247.
- 10. Moreno AA, Mukhtar MS, Blanco F, Boatwright JL, Moreno I, et al. (2012) IRE1/bZIP60-mediated unfolded protein response plays distinct roles in plant immunity and abiotic stress responses. PLOS One 7: e31944. pmid:22359644
- 11. Hayashi S, Wakasa Y, Takahashi H, Kawakatsu T, Takaiwa F (2012) Signal transduction by IRE1-mediated splicing of bZIP50 and other stress sensors in the endoplasmic reticulum stress response of rice. Plant J 69: 946–956. pmid:22050533
- 12. Lu SJ, Yang ZT, Sun L, Sun L, Song ZT, et al. (2012) Conservation of IRE1-regulated bZIP74 mRNA unconventional splicing in rice (Oryza sativa L.) involved in ER stress responses. Mol Plant 5: 504–514. pmid:22199238
- 13. Takahashi H, Kawakatsu T, Wakasa Y, Hayashi S, Takaiwa F (2012) A rice transmembrane bZIP transcription factor, OsbZIP39, regulates the endoplasmic reticulum stress response. Plant Cell Physiol 53: 144–153. pmid:22084314
- 14. Nagashima Y, Mishiba KI, Suzuki E, Shimada Y, Iwata Y, et al. (2011) Arabidopsis IRE1 catalyses unconventional splicing of bZIP60 mRNA to produce the active transcription factor. Sci Rep 1: pmid:22355548
- 15. Gao H, Brandizzi F, Benning C, Larkin RM (2008) A membrane-tethered transcription factor defines a branch of the heat stress response in Arabidopsis thaliana. Proc Natl Acad Sci USA 105: 16398–16403. pmid:18849477
- 16. Che P, Bussell JD, Zhou WX, Estavillo GM, Pogson BJ, et al. (2010) Signaling from the endoplasmic reticulum activates brassinosteroid signaling and promotes acclimation to stress in Arabidopsis. Sci Signal 3: ra69. pmid:20876872
- 17. Sun L, Lu SJ, Zhang SS, Zhou SF, Sun L, et al. (2013) The lumen-facing domain is important for the biological function and organelle-to-organelle movement of bZIP28 during ER stress in Arabidopsis. Mol Plant 6: 1605–1615. pmid:23558471
- 18. Sun L, Zhang SS, Lu SJ, Liu JX (2015) Site-1 protease cleavage site is important for the ER stress-induced activation of membrane-associated transcription factor bZIP28 in Arabidopsis. Sci China-Life Sci 58: 270–275. pmid:25634523
- 19. Tian L, Zhang Y, Ma EKH, Zhao H, Yuan M, et al. (2018) Basic-Leucine Zipper 17 and HMG-CoA reductase degradation 3A are involved in salt acclimation memory in Arabidopsis. J Integr Plant Biol pmid:30450762
- 20. Liu JX, Srivastava R, Che P, Howell SH (2007) Salt stress responses in Arabidopsis utilize a signal transduction pathway related to endoplasmic reticulum stress signaling. Plant J 51: 897–909. pmid:17662035
- 21. Deng Y, Srivastava R, Quilichini TD, Dong H, Bao Y, et al. (2016) IRE1, a component of the unfolded protein response signaling pathway, protects pollen development in Arabidopsis from heat stress. Plant J 88: 193–204. pmid:27304577
- 22. Sun L, Yang ZT, Song ZT, Wang MJ, Sun L, et al. (2013) The plant-specific transcription factor gene NAC103 is induced by bZIP60 through a new cis-regulatory element to modulate the unfolded protein response in Arabidopsis. Plant J 76: 274–286. pmid:23869562
- 23. Mishiba KI, Nagashima Y, Suzuki E, Hayashi N, Ogata Y, et al. (2013) Defects in IRE1 enhance cell death and fail to degrade mRNAs encoding secretory pathway proteins in the Arabidopsis unfolded protein response. Proc Natl Acad Sci USA 110: 5713–5718. pmid:23509268
- 24. Wang QL, Sun AZ, Chen ST, Chen LS, Guo FQ (2018) SPL6 represses signalling outputs of ER stress in control of panicle cell death in rice. Nat Plant 4: 280–288.
- 25. Schatlowski N, Stahl Y, Hohenstatt ML, Goodrich J, Schubert D (2010) The CURLY LEAF interacting protein BLISTER controls expression of polycomb-group target genes and cellular differentiation of Arabidopsis thaliana. Plant Cell 22: 2291–2305. pmid:20647345
- 26. Kleinmanns JA, Schatlowski N, Heckmann D, Schubert D (2017) BLISTER regulates polycomb-target genes, represses stress-regulated genes and promotes stress responses in Arabidopsis thaliana. Front Plant Sci 8: 1530. pmid:28955347
- 27. Song ZT, Sun L, Lu SJ, Tian Y, Ding Y, et al. (2015) Transcription factor interaction with COMPASS-like complex regulates histone H3K4 trimethylation for specific gene expression in plants. Proc Natl Acad Sci USA 112: 2900–2905. pmid:25730865
- 28. Geldner N, Dénervaud-Tendon V, Hyman DL, Mayer U, Stierhof YD, et al. (2009) Rapid, combinatorial analysis of membrane compartments in intact plants with a multicolor marker set. Plant J 59: 169–178. pmid:19309456
- 29. Ruberti C, Lai Y, Brandizzi F (2018) Recovery from temporary endoplasmic reticulum stress in plants relies on the tissue-specific and largely independent roles of bZIP28 and bZIP60, as well as an antagonizing function of BAX-Inhibitor1 upon the pro-adaptive signaling mediated by bZIP28. Plant J 93: 155–165. pmid:29124827
- 30. Deng Y, Srivastava R, Howell SH (2013) Protein kinase and ribonuclease domains of IRE1 confer stress tolerance, vegetative growth, and reproductive development in Arabidopsis. Proc Natl Acad Sci USA 110: 19633–19638. pmid:24145452
- 31. Rubio C, Pincus D, Korennykh A, Schuck S, El-Samad H, et al. (2011) Homeostatic adaptation to endoplasmic reticulum stress depends on Ire1 kinase activity. J Cell Biol 193: 171–184. pmid:21444684
- 32. Chawla A, Chakrabarti S, Ghosh G, Niwa M (2011) Attenuation of yeast UPR is essential for survival and is mediated by IRE1 kinase. J Cell Biol 193: 41–50. pmid:21444691
- 33. Korennykh AV, Korostelev AA, Egea PF, Finer-Moore J, Stroud RM, et al. (2011) Structural and functional basis for RNA cleavage by IRE1. BMC Biol 9: 15.
- 34. Bertolotti A, Zhang YH, Hendershot LM, Harding HP, Ron D (2000) Dynamic interaction of BiP and ER stress transducers in the unfolded-protein response. Nat Cell Biol 2: 326–332. pmid:10854322
- 35. Chen Y, Brandizzi F (2013) IRE1: ER stress sensor and cell fate executor. Trend Cell Biol 23: 547–555.
- 36. Hetz C, Glimcher LH (2009) Fine-tuning of the unfolded protein response: Assembling the IRE1 alpha interactome. Mol Cell 35: 551–561. pmid:19748352
- 37. Koizumi N, Ujino T, Sano H, Chrispeels MJ (1999) Overexpression of a gene that encodes the first enzyme in the biosynthesis of asparagine-linked glycans makes plants resistant to tunicamycin and obviates the tunicamycin-induced unfolded protein response. Plant Physiol 121: 353–361. pmid:10517826
- 38. Bao Y, Pu Y, Yu X, Gregory BD, Srivastava R, et al. (2018) IRE1B degrades RNAs encoding proteins that interfere with the induction of autophagy by ER stress in Arabidopsis thaliana. Autophagy 14: 1562–1573. pmid:29940799
- 39. Chen Y, Brandizzi F (2012) AtIRE1A/AtIRE1B and AGB1 independently control two essential unfolded protein response pathways in Arabidopsis. Plant J 69: 266–277. pmid:21914012
- 40. Qi H, Xia FN, Xie LJ, Yu LJ, Chen QF, et al. (2017) TRAF family proteins regulate autophagy dynamics by modulating AUTOPHAGY PROTEIN6 stability in Arabidopsis. Plant Cell 29: 890. pmid:28351989
- 41. Zhang SS, Yang H, Ding L, Song ZT, Ma H, et al. (2017) Tissue-specific transcriptomics reveals an important role of the unfolded protein response in maintaining fertility upon heat stress in Arabidopsis. Plant Cell 29: 1007–1023. pmid:28442596