The overexpression of OsACBP5 protects transgenic rice against necrotrophic, hemibiotrophic and biotrophic pathogens

The most devastating diseases in rice (Oryza sativa) are sheath blight caused by the fungal necrotroph Rhizoctonia solani, rice blast by hemibiotrophic fungus Magnaporthe oryzae, and leaf blight by bacterial biotroph Xanthomonas oryzae (Xoo). It has been reported that the Class III acyl-CoA-binding proteins (ACBPs) such as those from dicots (Arabidopsis and grapevine) play a role in defence against biotrophic pathogens. Of the six Arabidopsis (Arabidopsis thaliana) ACBPs, AtACBP3 conferred protection in transgenic Arabidopsis against Pseudomonas syringae, but not the necrotrophic fungus, Botrytis cinerea. Similar to Arabidopsis, rice possesses six ACBPs, designated OsACBPs. The aims of this study were to test whether OsACBP5, the homologue of AtACBP3, can confer resistance against representative necrotrophic, hemibiotrophic and biotrophic phytopathogens and to understand the mechanisms in protection. Herein, when OsACBP5 was overexpressed in rice, the OsACBP5-overexpressing (OsACBP5-OE) lines exhibited enhanced disease resistance against representative necrotrophic (R. solani & Cercospora oryzae), hemibiotrophic (M. oryzae & Fusarium graminearum) and biotrophic (Xoo) phytopathogens. Progeny from a cross between OsACBP5-OE9 and the jasmonate (JA)-signalling deficient mutant were more susceptible than the wild type to infection by the necrotroph R. solani. In contrast, progeny from a cross between OsACBP5-OE9 and the salicylic acid (SA)-signalling deficient mutant was more susceptible to infection by the hemibiotroph M. oryzae and biotroph Xoo. Hence, enhanced resistance of OsACBP5-OEs against representative necrotrophs appears to be JA-dependent whilst that to (hemi)biotrophs is SA-mediated.

www.nature.com/scientificreports/ Transgenic rice OsACBP5-OEs conferred protection against biotrophic bacterial pathogen Xanthomonas oryzae. When transgenic rice OsACBP5-OEs were evaluated against the bacterial leaf blight disease caused by Xanthomonas oryzae pv. oryzae (Xoo) using the leaf-clipping method (Fig. 3A), the average lesion length on OsACBP5-OEs at 14 dpi was 0.7 cm, while the average lesion lengths in WT and vectortransformed plants were 5.7 cm and 5.5 cm, respectively, representing a fivefold reduction in disease development (Fig. 3B). These results indicate that OsACBP5-OEs displayed enhanced tolerance to Xoo.
SA and JA levels in transgenic rice OsACBP5-OEs were elevated. As SA and JA play important roles in regulating immune responses in rice 61 , possible relationship between SA and JA in enhanced pathogenresistance in OsACBP5-OEs, was investigated by SA and JA content measurements using gas chromatographymass spectrometry (GC-MS) on uninfected and R. solani-infected plant samples. A two-fold increase in endogenous SA was shown in OsACBP5-OEs in comparison to the controls (Fig. 4A,B). Similarly, the endogenous JA www.nature.com/scientificreports/ content in OsACBP5-OEs were two-fold higher than the controls (Fig. 4A, B). When qRT-PCR was performed, the expression of OsNPR1, an SA-signalling regulatory gene, and ALLENE OXIDE SYNTHASE1 (OsAOS1) encoding allene oxide synthase in JA biosynthesis, were upregulated in uninfected and R. solani-infected plant samples of OsACBP5-OEs in comparison to that of controls (Fig. 4C,D). These results indicate that both SA-and JA-signalling pathways are activated in OsACBP5-OEs.
Protection in rice OsACBP5-OEs are dependent on both SA-and JA-signalling pathways. To determine if SA-and JA-signalling pathways are involved in the enhanced resistance of rice OsACBP5-OEs against various plant pathogens, WT, OsACBP5-OE9 (OE-9), osnpr1 (SA-signalling-deficient mutant), oscoi1 (JA-signalling-deficient mutant) and OE-9 in the osnpr1 or oscoi1 backgrounds were infected with the necrotroph R. solani (Fig. 5A), the hemibiotroph M. oryzae (Fig. 5B) and the biotroph Xoo (Fig. 5C). The measurement of lesion length in R. solani-infected plants showed significantly higher susceptibility in OE-9oscoi1 and oscoi1 plants in comparison to the WT (Fig. 5D). No significant difference was observed between osnpr1 and the WT (Fig. 5D). The OE-9osnpr1 mimicked the response of OE-9 (Fig. 5D). These results indicated that the improved resistance of OsACBP5-OEs to the fungal necrotroph R. solani is JA-dependent. When the plants (WT, OE-9, osnpr1, oscoi1, OE-9osnpr1 and OE-9oscoi1) were infected with the hemibiotrophic fungal pathogen M. oryzae and the biotroph Xoo, OE-9osnpr1 plants were no longer resistant to the pathogen similar to osnpr1 (Fig. 5, E and F). No significant difference was observed in M. oryzae-or Xoo-infected WT and oscoi1 mutant (Fig. 5E,F). OE-9 and OE-9oscoi1 showed similar responses to representative hemibiotrophic (M. oryzae) and biotrophic (Xoo) pathogen infection (Fig. 5E,F). These findings illustrate that the SA-signalling pathway is responsible for the enhanced resistance of OsACBP5-OEs to infection caused by the representative hemibiotroph and biotroph.  www.nature.com/scientificreports/ Furthermore, GC-MS was performed to measure SA and JA content in R. solani-, M. grisea-and Xoo-infected WT, OE-9, osnpr1, oscoi1, OE-9osnpr1 and OE-9oscoi1. R. solani-infected oscoi1 and OE-9oscoi1 respectively showed 40-and three-fold lower JA content compared to the WT (Supplemental Fig. S2A). When qRT-PCR was performed, oscoi1 and OE-9oscoi1 respectively showed ~ ten-and three-fold lower expression of OsAOS1 in comparison to that of WT (Supplemental Fig. S3A). Similarly, M. grisea-and Xoo-infected osnpr1 and OE-9osnpr1 respectively showed ~ 20-and 2.5-fold lower SA content compared to the WT (Supplemental Fig. S2B,C). When qRT-PCR was performed, osnpr1 and OE-9osnpr1 respectively showed ~ ten-and three-fold lower expression of OsNPR1 compared to the WT (Supplemental Fig. S3B,C). These results further confirm that the improved resistance of OsACBP5-OEs to the necrotroph R. solani is JA-dependent and enhanced resistance of OsACBP5-OEs to hemibiotroph (M. grisea) and biotroph (Xoo) is SA-dependent. www.nature.com/scientificreports/ two W-boxes in the OsACBP5 5′-flanking region bind infected rice nuclear proteins.  (Fig. 6A). EMSAs using crude nuclear extracts from R. solaniinfected 5-week-old WT rice, C. oryzae-infected three-week-old WT rice, M. oryzae-infected three-week-old WT rice, Xoo-infected three-week-old WT rice and C. oryzae-infected three-week-old WT rice showed strong DNA-protein binding complexes with the W-boxes at − 1713/− 1708 and − 157/− 152 (Fig. 6B), indicating that two of the four putative W-boxes are essential in regulating OsACBP5 expression. In contrast, when the CGTCA and Skn-1 boxes were tested, they did not bind to nuclear extracts in EMSA (Supplemental Fig. S4). binding of (His) 6 -tagged OsACBP5 to 18:3-acyl-CoA esters 28 . As 18:3-FA is important for basal defence against fungal pathogens and is a precursor for JA biosynthesis 67 , the binding affinity of (His) 6 -OsACBP5 to 18:3-acyl-CoA ester was investigated by isothermal titration calorimetry (ITC) which provides a more precise method to measure protein-ligand binding than Lipidex assays 20 . Consistent with Lipidex assays, recombinant OsACBP5 (rOsACBP5) was shown to bind to 18:3-acyl-CoA with high affinities (Supplemental Fig. S5). ITC results (Supplemental Table S1) indicated that rOsACBP5 has a strong binding affinity to 18:3-acyl-CoA ester with a dissociation constant (K d ) value of 59.5 nM. When OsACBP5-OE leaves were further examined using GC-MS to test the level of the six major FA species  www.nature.com/scientificreports/   Increase in the cytosolic Ca 2+ concentration by the activation of CNGCs, CDPK and CaM/CML in PTI, is a regulator for production of ROS and NOS which results in the hypersensitive response 121,122 . Activation of FLS2 in PTI triggers the MAPK signalling pathway that induces known defence genes for the generation of antimicrobial compounds such as phytoalexins, camalexin and lignin 123,124 . Pathogen infection induces RIN4 in ETI, which activates the disease resistant proteins RPM1 and RPS2 92 . RPM1 and RPS2 then trigger a complex formed by HSP90, RAR1 and SGT1 leading to the hypersensitive response [96][97][98] . SGT1 also regulates early R gene-mediated plant defences upon pathogen infection 125

Discussion
OsACBP5 conferred broad-spectrum defence against phytopathogens. In this study, the function of OsACBP5 in plant defence was established by phenotypic analyses of five independent rice OsACBP5-OE lines in response to representative necrotrophic, hemibiotrophic and biotrophic pathogens. Previous work on transgenic Arabidopsis overexpressing its homologue, AtACBP3, had shown that AtACBP3 could confer NONEXPRESSOR OF PR GENES1 (NPR1)-dependent resistance to bacterial biotroph P. syringae, with increased susceptibility to the fungal necrotroph B. cinerea 17 . In contrast, this study revealed that transgenic rice OsACBP5-OEs displayed enhanced tolerance to necrotrophic fungal pathogens such as R. solani and C. oryzae (Fig. 1), hemibiotrophic fungal pathogens, M. oryzae and F. graminearum (Fig. 2) and a biotrophic bacterial pathogen, Xoo (Fig. 3). These findings demonstrated that OsACBP5 is more versatile against pathogens in transgenic rice. As Takato et al. had earlier illustrated that the overexpression of a Class III ACBP from grape (Vitis vinifera) could protect transgenic Arabidopsis against the biotroph P. syringae and a hemibiotroph C. higginsianum 30 , it appears that the Class III ACBPs are promising targets for disease prevention in both transgenic dicots and monocots. Similar to OsACBP5 in exhibiting broad-spectrum properties in defence, wide-range protection against R. solani, M. oryzae and Xoo have been reported in transgenic rice overexpressing a cysteinerich antimicrobial defensin from Allium cepa (Ace-AMP), but the molecular regulation on its action is less understood 68 . In comparison, defence-related proteins such as the rice wall-associated kinase (OsWAK25) and MoSM1, encoding a cerato-platanin protein from M. oryzae, when overexpressed in rice, conferred protection only against the hemibiotroph M. oryzae and the bacterial biotroph Xoo, but displayed increased susceptibility to the fungal necrotroph R. solani 57,69 . Correspondingly, the expression of OsWRKY13, encoding transcription factor WRKY, was induced by M. oryzae and Xoo infection 58,86 . The constitutive expression of OsWRKY13 displayed protection to M. oryzae and Xoo via the SA-signalling pathway 58 (GH3-8), whose expression is induced by auxin, enhanced resistance to M. oryzae and Xoo infection in rice by suppressing pathogen-induced IAA accumulation 58,141 . Other rice genes that promote similar pathogen resistance are summarised in Table 2.
Rice OsACBP5-OEs showed JA-mediated response against necrotrophs and SA-mediated response against (hemi)biotrophs. SA and JA, the two critical defence signalling hormones that play vital roles against necrotrophic, hemibiotrophic and biotrophic pathogens in rice 61,70 , were observed to accumulate in rice OsACBP5-OEs (Fig. 4A,B). The upregulated expression of OsNPR1, an SA-signalling regulatory gene, and OsAOS2 encoding allene oxide synthase in JA biosynthesis, in OsACBP5-OEs (Fig. 4C,D) likely stimulates SA-and JA-mediated defence responses. Furthermore, results from bioassays on transgenic rice OsACBP5-OE9 in oscoi1 and osnpr1 backgrounds suggest that necrotrophic resistance in rice OsACBP5-OEs is JA-dependent and (hemi)biotrophic resistance is SA-dependent (Fig. 5) 59,60 . These results from proteomic analysis indicated that the defence responses arising from OsACBP5 overexpression in transgenic Arabidopsis involved cell wall-mediated defence as well as salicylic acid (SA)-and jasmonic acid (JA)-mediated defence pathways 59,60 . Similarly, transgenic Arabidopsis overexpressing AtACBP3 displayed enhanced SA-mediated resistance to the biotrophic pathogen P. syringae 17 .
The current results on transgenic rice OsACBP5-OEs also demonstrated the cooperation between the SA and JA pathways in defence against representative necrotrophs (R. solani, C. oryzae), hemibiotrophs (M. oryzae, F. graminearum) and biotrophs (Xoo). In contrast, the SA and JA defence signalling pathways generally interact antagonistically in dicots 51,53,55,[71][72][73][74][75][76][77][78] , while such interaction is not well investigated in monocots such as rice 56,79 . Nonetheless, Tamaoki et al. reported that SA and JA can collaboratively stimulate a common defence signalling system in rice against pathogens 80 . Similar to the present study, transgenic MoSM1-OE rice displayed improved resistance to the hemibiotroph M. oryzae and the biotroph Xoo accompanied by elevated SA and JA content and upregulated expression of SA-and JA-signalling genes 57 . SA and JA accumulation in rice OsACBP5-OEs may have arisen from the ability of OsACBP5 in binding to 18:3-acyl-CoA ester because ITC data supported rOsACBP5 binding to 18:3-acyl-CoA ester with a K d value in the nanomolar range (Fig. S5). Interestingly, OsACBP5-OEs showed higher linolenic acid (18:3) content than the controls (Fig. S6) and 18:3-FA is a precursor for JA biosynthesis 67 . Previous reports have demonstrated that the Arabidopsis ssi2 mutant contains lower 18:3-FA content than the WT and was more susceptible to necrotroph B. cinerea infection 82 . These results resonate well with the current study which revealed that decreased susceptibility of OsACBP5-OEs to various pathogens, in comparison to the WT, is associated with increase in 18:3-FA content.
The physiological significance in the role of AtACBP3 in trafficking lipids such as acyl-CoA esters was evident in transgenic Arabidopsis overexpressing AtACBP3 which displayed accelerated leaf senescence, in contrast to an atacbp3 T-DNA insertional mutant and AtACBP3 RNA interference (RNAi) transgenic Arabidopsis lines which were delayed in dark-induced leaf senescence 15 . Subsequent acyl-CoA and lipid profiling revealed that AtACBP3 overexpression culminated in the accumulation of acyl-CoA and phosphatidylethanolamine (PE), while the downregulation of AtACBP3 reduced PE 15 . In dark-treated and premature senescing AtACBP3-OE plants, PC and phosphatidylinositol levels declined accompanied by increases in PA, lysophospholipids, and oxylipin-containing galactolipids (arabidopsides). It was concluded that the accumulation of PA and arabidopsides (A, B, D, E, and G) resulting from lipid peroxidation in AtACBP3-OEs likely caused leaf senescence 15 . In another study, it was reported that oxylipin-related FA (18:2-FA, 18:3-FA and MeJA) content was lower in atacbp3 and AtACBP3-RNAi than wild-type phloem exudates upon GC-MS analysis 28 . On ITC analysis, recombinant AtACBP3 was shown to bind medium-and long-chain acyl-CoA esters with K D values in the micromolar range 28 . Hu et al. concluded that the phloem-mobile AtACBP3 likely affected the FA pool and JA content in the phloem by its binding to acyl-CoA esters, ultimately influencing the level of oxylipins, which are crucial components of the plant wound responses mobilized via the vasculature 28 . www.nature.com/scientificreports/ Significance of W-boxes in regulating pathogen-inducible OsACBP5 expression. The WRKY family of TFs that regulate the transcription of plant defence genes through the W-box 84 , are crucial in protection against necrotrophic, hemibiotrophic and biotrophic pathogens [84][85][86][87][88][89][90][91][92][93][94] . For example, OsWRKY4 binds to the W-boxes in the 5′-flanking region in each of pathogenesis-related PR1b and PR5, and OsWRKY4 and OsWRKY80 were reported to be highly induced by R. solani infection 91,95 . Wang et al. also showed that transgenic rice overexpressing OsWRKY4 were protected against R. solani infection 95 . In this study, OsACBP5-OEs were proven tolerant to representative necrotrophs, hemibiotrophs and biotrophic phytopathogens and EMSAs revealed that only two of the four W-boxes (-1713/-1708 and -157/-152) in the 5′-flanking region of OsACBP5 regulate OsACBP5 expression during representative necrotrophic (Fig. 6B panels i, iv, v and viii), hemibiotrophic (Fig. 6B panels ii and vi) and biotrophic (Fig. 6B panels iii and vii) infection. These results correspond well with quantitative GUS assays on the pOS820 (2.2-kb OsACBP5pro::GUS; -1926/ + 304) transformants which displayed induced GUS expression after treatment with the pathogen-related phytohormones, SA and MeJA, in comparison to transformants from constructs that lacked either of these W-boxes, pOS891 (1.3-kb OsACBP5pro::GUS; − 1,281/ + 304) and pOS895 (0.6-kb OsACBP5pro::GUS; − 46/ + 304) (Fig. 6C,D). These results verified that the two W-boxes (− 1713/− 1708 and − 157/− 152) play a role in regulating OsACBP5 expression in response to necrotrophic, hemibiotrophic and biotrophic pathogens, as well as to the pathogen-related phytohormones, SA and MeJA. Previous results suggest that increased protection to R. solani and M. oryzae in transgenic rice overexpressing OsWRKY30 was associated with elevated levels of JA, as well as the stimulated expression of JA synthesisrelated genes (LOX and AOS2) and pathogenesis-related PR3 and PR10, following fungal pathogen infection 90 . Furthermore, Hiroyuki and Terauchi revealed that the W-boxes in the RICE THAUMATIN-LIKE PROTEIN1 (RTLP1) promoter function in response to M. oryzae infection 96 . Also, past investigations on the development of resistance in rice against the rice blast pathogen M. oryzae unveiled a critical role for WRKY TFs (OsWRKY45, OsWRKY13 and OsWRKY42) in plant defence 92 . OsWRKY45 has been assigned a vital role in SA-mediated signalling in rice against the hemibiotrophic pathogen M. oryzae 84 . Enhanced resistance of transgenic rice to the biotrophic pathogen Xoo and the hemibiotroph M. oryzae was achieved by OsWRKY13 overexpression that was related to activation of SA-signalling and suppression of the JA-dependent pathway 89 . Similar to OsACBP5, where W-boxes were observed to bind nuclear extracts from Xoo-infected rice, W-boxes in rice STRESS RESPON-SIVE NAC1 (SNAC1) interacted with Xoo-treated nuclear proteins and OsWRKY13 was subsequently identified to regulate SNAC1 expression during biotic stress 97 . Furthermore, the overexpression of OsWRKY13 or OsWRKY71 culminated in better tolerance to Xoo in transgenic rice 87,98 . Taken together these studies support a role for OsWRKY TFs in biotrophic, hemibiotrophic and necrotrophic fungal tolerance in rice via the SA-and JA-defence signalling pathways.
Several defence-related genes were upregulated in R. solani-infected OsACBP5-OEs. The role of SA and JA in hemi(bio)trophic and necrotrophic pathogen defence in transgenic rice OsACBP5-OEs was partially confirmed from pathogen assays, GC-MS and qRT-PCR. Transcriptomic and proteomic analyses further confirmed the mechanism of defence in transgenic rice OsACBP5-OEs. Although transcriptomic and proteomic assays were performed only on necrotrophic pathogen R. solani-infected transgenic rice OsACBP5-OEs, the upregulated genes and proteins from these assays were reported to be involved in defence against necrotrophic, hemibiotrophic and biotrophic phytopathogens, which are discussed in this section.
Transcriptomics and proteomics data provided an insight into the defence responses of transgenic rice OsACBP5-OEs to the necrotrophic pathogen R. solani infection. The innate immunity in plants appeared to be triggered through PTI followed by ETI, providing the first line of defence upon pathogen challenge 99,100 . Ten genes involved in PTI were up-regulated in OsACBP5-OEs upon R. solani infection (Fig. 7A). Cytoplasmic Ca 2+ concentration increases during PTI leading to the activation of CDPK in plant cells 101 . In this study, three genes involved in Ca 2+ signalling were up-regulated in R. solani-infected OsACBP5-OEs including CNGCs, CDPK, CaM/CML. Calcium signalling was reportedly accompanied by an increase of both ROS and NO leading to SAmediated defence 30,101 . Transcription factors WRKY22 and WRKY33 were activated by components of the MAPK cascade such as MEKK1, MKK1/2 and MKK4/5, resulting in induced expression of defence-related genes in R. solani-infected OsACBP5-OEs (Fig. 7A). Similar results were observed in Xoo-infected rice plants in which FLS2 perceived bacterial flagellin and activated the MAPK cascade which in turn activated WRKY22 and WRKY33 resulting in induced expression of defence-related genes 92,99,100 . Taken together, ROS production and activation of MAPKs and CDPKs cause an array of defences restricting pathogen progression.
In plants, a secondary immune response ETI is the basis for a second layer of defence 99,100 . The second signalling pathway consists of five genes encoding receptor proteins (RIN4, PBS1, RPM1, SGT1 and RAR1) to perceive pathogen infection. In this study, four such genes encoding receptor proteins including RIN4, RPM1, SGT1 and RAR1 displayed up-regulation in OsACBP5-OEs following R. solani infection (Fig. 7A). RPM1 recognizes modifications of RIN4 followed by P. syringae infection in Arabidopsis and RPM1 interacts with RIN4 triggering RPM1-mediated immunity 102 . RAR1 and SGT1 conferred resistance against Xoo and M. oryzae when overexpressed in rice 103,104 . RAR1 forms a complex with the molecular chaperones HSP90 and SGT1 to initiate a signalling cascade in diverse plant immune responses [105][106][107][108] . In this study, the upregulated expression of various components (HSP90, RAR1 and SGT1) of the complex likely caused a hypersensitive response in OsACBP5-OEs following R. solani infection (Fig. 7A). Previous studies have reported that hypersensitive responses are mostly accompanied by an increase in SA biosynthesis 109,110 .
Furthermore, several DEGs (NPR1, TGA, PR1, JAR1, COI1, JAZ and MYC2) involved in the SA-and JAsignalling pathways were enriched in R. solani-infected OsACBP5-OEs (Fig. 7B), suggesting that both pathways are involved (Fig. 7B) Quantification of SA and JA. SA and JA quantification was performed following Fina et al. 116 . Leaf tissue (300 mg) was homogenized and SA extracted in 80% methanol by shaking for 16 h at − 20 °C. The samples were then purified on a C18 cartridge (Bond Elut C18 6 cc, 500 mg, Agilent, CA, USA) in 80% methanol. Formic acid was added for the binding of SA to the cartridge. The SA was eluted with diethyl ether. The eluent was evaporated under nitrogen gas after removing the residual water. The sample was further methylated using diazomethane and dried under nitrogen gas. The sample was subsequently dissolved in 100% hexane for GC-MS analysis. The same protocol was followed for JA quantification. SA (10 µM) and JA (10 µM) were used as internal standards.
Expression and purification of OsACBP5. The (His) 6  transcriptome analysis. Total RNA was extracted from R. solani-infected WT, vector (pCAMBIA1304)transformed control and transgenic rice OsACBP5-OEs using the RNeasy Plant Mini Kit (Qiagen). RNAs samples were sequenced using BGISEQ-500 sequencer at Beijing Genomics Institute (BGI, Hong Kong). RNA concentration and quality were measured using Agilent 2100 Bio analyser (Agilent RNA 6000 Nano Kit). The BGISEQ-500 platform was used to sequence the cDNA libraries. SOAPnuke was used to filter reads and after filtering, the clean reads were stored in FASTQ format. The clean reads were mapped using Bowtie2 (https ://bowti e-bio.sourc eforg e.net/Bowti e2/index .shtml ) and the gene expression level was calculated using RSEM (https :// dewey lab.biost at.wisc.edu/RSEM). Differentially expressed genes (DEGs) were detected using DEGseq software based on the Poisson distribution. The KEGG database [148][149][150] was used for pathway analysis. The open reading frame (ORF) of each DEG was identified using GETORF database. To predict the transcription factor of each DEG, the ORF was aligned to transcription factor domains using the PLNTFdb database. DEGs were mapped to the PRGdb database using BLAST to detect plant disease resistance genes.
Sequential window acquisition of all theoretical mass spectra quantitative proteomic analysis. The trichloroacetic acid/acetone method was used for proteins extraction following Wu et al. 118 . The protein pellet was resuspended in 2 mL urea buffer (6 M urea and 4 mM calcium chloride in 200 mM 3-(N-morpholino) propanesulfonic acid (MOPS), pH 8.0) 120 . An equivalent amount of protein (100 μg) was reduced using 10 mM dithiothreitol (DTT) and alkylated in 40 mM iodoacetamide (IAA) in the dark. After alkylation, the concentration of urea in the mixture was reduced to less than 2 M by diluting with 4 mM CaCl 2 . The protein was digested with trypsin (1:20) followed by incubation at 37 °C overnight. Subsequently, the peptides were desalted utilising C18 SepPak reverse-phase cartridges and SWATH-MS analysis was performed 121 . The data was analysed from five biological repeats.
Statistical analysis. Significant differences in data between different samples were analyzed by the Student's t-test.