ABSTRACT
In contrast to desiccation-tolerant orthodox seeds, recalcitrant seeds are desiccation sensitive and are unable to survive for a prolonged time. Here, our analyses of Oryza species with contrasting seed desiccation tolerance reveals that PROTEIN L-ISOASPARTYL METHYLTRANSFERASE (PIMT), an enzyme that repairs abnormal isoaspartyl (isoAsp) residues in proteins, acts as a key player that governs seed desiccation tolerance to orthodox seeds but is ineffective in recalcitrant seeds. We observe that, unlike the orthodox seed of Oryza sativa, desiccation intolerance of the recalcitrant seeds of Oryza coarctata are linked to reduced PIMT activity and increased isoAsp accumulation due to the lack of coordinated action of ABA and ABI transcription factors to upregulate PIMT during maturation. We show that suppression of PIMT reduces, and its overexpression increases, seed desiccation tolerance and seed longevity in O. sativa. Our analyses further reveal that the ABI transcription factors undergo isoAsp formation that affect their functional competence; however, PIMT interacts with and repairs isoAsp residues and facilitates their functions. Our results thus illustrate a new insight into the mechanisms of acquisition of seed desiccation tolerance and longevity by ABI transcription factors and the PIMT module.
INTRODUCTION
Most flowering plants produce orthodox seeds, which are desiccation tolerant and can survive in a dry quiescent state for a prolonged period (longevity). However, a few flowering plant species (8%) produce recalcitrant seeds, which are desiccation sensitive and are unable to survive in a dry, quiescent state (Roberts, 1973; Finch-Savage et al., 1992). Previous studies indicate that acquisition of desiccation tolerance of orthodox seeds is a highly coordinated molecular event that is largely mediated by the hormone abscisic acid (ABA) and coordinated participation of different master regulators – ABSCISIC ACID INSENSITIVE 3 (ABI3), FUSCA3 (FUS3), LEAFY COTYLEDON 1 (LEC1) and LEC2 – and their downstream gene products and processes (Parcy et al., 1994; Sugliani et al., 2009; Zinsmeister et al., 2016; Leprince et al., 2017). These processes include the accumulation of several diverse protective molecules, such as late embryogenesis abundant (LEA) proteins, heat shock proteins (HSPs), and mono-, di- oligosaccharides (soluble sugars) during the maturation phase in orthodox seeds (Blackman et al., 1995; Wehmeyer and Vierling, 2000; Berjak and Pammenter, 2008). Various studies suggest that such mechanisms are either absent or might be present but are ineffective in recalcitrant seeds (Finch-Savage et al., 1990; Still et al., 1994; Berjak and Pammenter, 2008). Therefore, it is reasonable to presume that various mechanisms or processes that govern orthodox seeds to be desiccation tolerant and recalcitrant seeds to be sensitive are yet to be revealed. Recently, an evolutionarily conserved protein repairing enzyme, PROTEIN L-ISOASPARTYL METHYLTRANSFERASE (PIMT; EC 2.1.1.77), was shown to be a key factor governing a significant role in preserving seed germination vigor and longevity of orthodox seeds (Ogé et al., 2008; Verma et al., 2013; Petla et al., 2016). PIMT repairs abnormal isoaspartyl (isoAsp) residues, which arise from the spontaneous modification of asparaginyl (Asn) or aspartyl (Asp) residues of a protein, through S-adenosyl-l-methionine (AdoMet)-dependent methylation (Johnson et al., 1987; McFadden and Clarke, 1987; Kamble and Majee, 2020). PIMT is ubiquitously distributed in almost all living organisms, including bacteria, nematodes, mammals, plants, etc. During aging or storage, a number of proteins that play a key role in seed germination vigor and longevity were shown to be susceptible to isoAsp modification, resulting in poor seed vigor and longevity. However, during germination upon rehydration, PIMT, which is highly abundant in orthodox seeds, repairs the isoAsp residues in these proteins and restores their biological functions (Nayak et al., 2013; Petla et al., 2016). Intriguingly, during seed maturation when moisture content sharply declines, proteins, including those that are necessary for seed maturation and the acquisition of seed desiccation tolerance, are likely to undergo detrimental protein modifications, such as isoAsp formation, that can detrimentally affect their functions. Therefore, a seed has to avoid or protect cellular components, including proteins, both during water loss when undergoing maturation and upon imbibition during germination. However, the consequence of isoAsp formation and the necessity and role of PIMT during seed maturation and successful acquisition of desiccation tolerance remains largely unknown. Furthermore, the occurrence, regulation and possible role, if at all, of PIMT in recalcitrant seeds are yet unknown. To address these questions, we studied two closely related rice species, Oryza sativa and Oryza coarctata (also known as Porteresia coarctata), which are characterized by contrasting responses to seed desiccation tolerance.
In this study, we elucidate how PIMT is differentially regulated by ABA levels and ABA INSENSITIVE (ABI) transcription factors (ABI-TFs) during seed maturation and how it contributes to desiccation tolerance, and consequent longevity, in orthodox seeds of O. sativa, but not in recalcitrant seeds of O. coarctata.
RESULTS
Desiccation intolerance of recalcitrant rice seeds is associated with low PIMT activity, and PIMT activity is upregulated during desiccation re-establishment in germinating rice seeds
First, morpho-physiological and recalcitrant behavior of the fully grown plant, spikelet and seeds of O. coarctata and O. sativa were analyzed as shown in Fig. 1A,B. In agreement with previous studies (Aldridge and Probert, 1993), we observed that O. coarctata caryopses were relatively larger with a bigger embryo (Fig. 1C,D) and exhibited desiccation intolerance even at 19 days after pollination (DAP) and lost complete viability at maturity, in contrast to O. sativa orthodox seeds (Fig. 1E). Furthermore, O. coarctata mature seeds failed to germinate even after gibberellic acid (GA) treatment (Fig. 1D). Altogether, we confirmed that O. coarctata seeds are indeed desiccation-intolerant, recalcitrant seeds in accordance with the previous report (Aldridge and Probert, 1993). To investigate whether desiccation intolerance of O. coarctata seeds is linked to PIMT activity, we initially measured PIMT activity in mature seeds of O. coarctata, and compared it with that of O. sativa. Our results showed that in comparison with O. sativa seeds, a significantly lower level of PIMT activity was observed in O. coarctata seeds (Fig. 2A). Our analysis further revealed that PIMT activity in the embryo and endosperm of O. coarctata seeds was significantly reduced compared with that of O. sativa (Fig. S2A). Interestingly, except for seed, PIMT activity was found to be similar in root, stem, leaf and flower in these species (Fig. 2B). PIMT activity analysis during the course of seed development (Fig. S1) revealed that, in contrast to O. sativa seeds, in which PIMT activity was increased during the maturation phase (stages S4 and S5), PIMT activity remained the same until stage S3 and declined thereafter in the case of O. coarctata (Fig. 2C). Subsequently, isoAsp accumulation was also analyzed in these species. Our data revealed that O. coarctata mature seeds accumulated a significantly greater amount of isoAsp than did O. sativa seeds (Fig. 2D). Markedly increased isoAsp content was noticed in the embryo as well as endosperm of O. coarctata compared with that of O. sativa (Fig. S2B). isoAsp content in root, stem, leaf and flower showed no significant difference in these species (Fig. 2E). isoAsp accumulation was also analyzed during seed development and significantly greater isoAsp content was observed particularly at the S4 and S5 stages in O. coarctata compared with O. sativa (Fig. 2F). To confirm that reduced PIMT activity is not due to drying damage or loss of viability of O. coarctata seeds, we also analyzed the activity of some antioxidant proteins in mature seeds (De Gara et al., 2003). Mature seeds of O. coarctata showed significantly higher ascorbate peroxidase and superoxide dismutase activity but lesser catalase activity compared with O. sativa seeds (Fig. S3). This suggests that low PIMT activity in O. coarctata is due to differential regulation of PIMT in O. sativa and O. coarctata seeds and not due to drying damage or loss of viability of O. coarctata seeds (Fig. 2A-F). Furthermore, to analyze the function of PIMT during seed desiccation in O. sativa, we performed seed-desiccation treatment as described in the Materials and Methods (Downie et al., 2003; Maia et al., 2011). We observed the induction of PIMT activity when seeds were maintained in the optimized concentration of polyethylene glycol (PEG) or desiccated directly (Fig. 2G). Subsequently, we also estimated isoAsp levels associated with molecular injuries and desiccation tolerance in terms of germination following these treatments. It is important to note that direct desiccation treatment caused more desiccation injuries and also caused more induction of PIMT activity (Fig. 2H, Fig. S4). Additionally, we also analyzed the effect of controlled deterioration treatment (CDT) on seed germination in O. sativa and O. coarctata. The results suggest that after 10 days of CDT seed germination was significantly affected in O. sativa. However, seeds from O. coarctata failed to germinate without CDT (Fig. S5, 0 day). Altogether, these results strongly corroborate that the desiccation intolerance of the recalcitrant seeds of O. coarctata is associated with reduced PIMT activity and increased isoAsp accumulations compared with O. sativa (Fig. 2A-H).
Phenotypic analysis. (A) Image of full-grown O. sativa and O. coarctata plants. Scale bar: 5 cm. (B) Mature seeds of each species, with and without husk. Scale bar: 0.2 cm. (C) Electron microscopic image of mature seeds showing embryo (EM) and endosperm (EN) of O. sativa and O. coarctata. Scale bars: 200 µm. (D) Germination percentage of O. sativa and O. coarctata with or without gibberellic acid (GA). Data are mean±s.d. of three biological repeats. Significant differences among means are denoted by the different letters using Duncan's multiple range test (α=0.05). ND, not detected. (E) O. sativa and O. coarctata embryo viability was analyzed using tetrazolium staining; dark red staining indicates viable seeds. Scale bars: 5 mm.
Phenotypic analysis. (A) Image of full-grown O. sativa and O. coarctata plants. Scale bar: 5 cm. (B) Mature seeds of each species, with and without husk. Scale bar: 0.2 cm. (C) Electron microscopic image of mature seeds showing embryo (EM) and endosperm (EN) of O. sativa and O. coarctata. Scale bars: 200 µm. (D) Germination percentage of O. sativa and O. coarctata with or without gibberellic acid (GA). Data are mean±s.d. of three biological repeats. Significant differences among means are denoted by the different letters using Duncan's multiple range test (α=0.05). ND, not detected. (E) O. sativa and O. coarctata embryo viability was analyzed using tetrazolium staining; dark red staining indicates viable seeds. Scale bars: 5 mm.
Comparative analysis of PIMT activity and isoAsp accumulation in O. sativa and O. coarctata. (A-H) PIMT activity (A,B,C,G) and isoAsp accumulation (D,E,F,H) were determined in mature seeds (A,D), in different organs (B,E), during seed development in O. sativa and O. coarctata (C,F) and during desiccation re-establishment in O. sativa (G,H). Data are mean±s.d. of three biological repeats. (A,D) **P<0.05 (Student's t-test). (B,C,E,F,G,H) Significant differences among means (α=0.05) are denoted by the different letters.
Comparative analysis of PIMT activity and isoAsp accumulation in O. sativa and O. coarctata. (A-H) PIMT activity (A,B,C,G) and isoAsp accumulation (D,E,F,H) were determined in mature seeds (A,D), in different organs (B,E), during seed development in O. sativa and O. coarctata (C,F) and during desiccation re-establishment in O. sativa (G,H). Data are mean±s.d. of three biological repeats. (A,D) **P<0.05 (Student's t-test). (B,C,E,F,G,H) Significant differences among means (α=0.05) are denoted by the different letters.
O. coarctata possesses multiple PIMT transcript variants encoding biochemically active and inactive PIMT isoforms
To compare the biochemical properties of PIMT isoforms of O. coarctata with O. sativa, we cloned PIMT cDNAs from O. coarctata (Petla et al., 2016). Sequence analysis revealed four different transcript variants of PIMT, of which two are OcPIMT1 (OcPIMT1-1 and OcPIMT1-2) variants and two others are OcPIMT2 (OcPIMT2-1 and OcPIMT2-2) variants. OcPIMT1-1 encodes a 231-aa PIMT protein with all five conserved domains similar to OsPIMT1. However, OcPIMT1-2 encodes a relatively shorter PIMT1 protein (213 aa) that lacks the region II domain. OcPIMT2-1 encodes a 277-aa protein, fairly similar to OsPIMT2 protein, which contains all five domains and an N-terminal extension, and OcPIMT2-2 encodes a 252-aa protein that lacks few amino acid residues in pre-region I (TISAP----) and region I (ALDVGS---) domains along with few deletions and replacement of adjacent amino acid residues (Fig. 3A, Fig. S6). Multiple sequence alignment of OsPIMTs and OcPIMTs revealed more than 80% sequence identities between respective PIMT isoforms of these two species (Fig. 3A). Phylogenetic analyses showed that OcPIMTs clustered with OsPIMTs and other monocots (Fig. S7). For comparative enzymatic studies of recombinant PIMT isoforms between O. coarctata and O. sativa, including ΔOsPIMT2, ΔOcPIMT2-1 (PIMT2 lacking the N-terminal extension), were purified and assessed for enzyme activity (Fig. 3B). Only OcPIMT1-1, OcPIMT2-1 and ΔOcPIMT2-1 showed significant enzyme activity, whereas OcPIMT1-2 and OcPIMT2-2 had barely any activity because not all five domains are present (Fig. 3C). Our results also showed that specific activity of OcPIMT1-1, OcPIMT2-1 and ΔOcPIMT2-1 was fairly similar to the corresponding OsPIMT isoforms. Detailed biochemical properties of these enzymatically active OcPIMT isoforms and OsPIMT isoforms revealed comparable optimum pH and temperature, Km and Vmax values for isoAsp-containing peptide substrate and AdoMet (Table 1, Fig. S8).
Biochemical characterization of O. coarctata OcPIMTs. (A) Multiple sequence alignments of PIMT protein sequences of O. sativa and O. coarctata using Clustal W 2.1. Region-I, Region-II and Region-III, which are conserved in methyltransferase proteins utilizing AdoMet, are marked in blue. Pre-region-I and Post-Region III, which are unique to PIMT, are marked in red. Asterisks, full colons (:) and periods (.) represent identical amino acid residues, conserved substitutions and semi-conserved substitutions, respectively. (B) SDS-PAGE analysis of recombinant OsPIMT and OcPIMT proteins. M, molecular weight markers. (C) Comparison of PIMT enzyme activity among recombinant OsPIMTs and OcPIMTs. Data are mean±s.d. of three biological repeats. Significant differences among means are denoted by the different letters using Duncan's multiple range test (α=0.01).
Biochemical characterization of O. coarctata OcPIMTs. (A) Multiple sequence alignments of PIMT protein sequences of O. sativa and O. coarctata using Clustal W 2.1. Region-I, Region-II and Region-III, which are conserved in methyltransferase proteins utilizing AdoMet, are marked in blue. Pre-region-I and Post-Region III, which are unique to PIMT, are marked in red. Asterisks, full colons (:) and periods (.) represent identical amino acid residues, conserved substitutions and semi-conserved substitutions, respectively. (B) SDS-PAGE analysis of recombinant OsPIMT and OcPIMT proteins. M, molecular weight markers. (C) Comparison of PIMT enzyme activity among recombinant OsPIMTs and OcPIMTs. Data are mean±s.d. of three biological repeats. Significant differences among means are denoted by the different letters using Duncan's multiple range test (α=0.01).
PIMT isoforms are differentially expressed during seed maturation in O. sativa and O. coarctata
To investigate whether the reduced PIMT activity in O. coarctata seeds is due to the decreased expression of PIMT in seeds, accumulation of PIMT transcripts and protein isoforms were analyzed in O. coarctata and compared with O. sativa. Transcript abundance of PIMT1 and PIMT2 were comparable in root, stem, leaf and flower in both species, but significantly lower in O. coarctata seeds compared with O. sativa seeds (Fig. 4A). To investigate this further, transcript accumulation was also compared during seed development, and a similar pattern of transcript accumulation was observed until the S3 stage, but significantly differed at the later stages of seed development between these species. In contrast to O. sativa seeds, in which OsPIMT transcript accumulation sharply rose at the S4 and S5 stages, OcPIMT transcripts abundance did not show such rises during stages S4 and S5 of seed development (Fig. 4B). To confirm further, immunoblot analysis was carried out in mature seeds (Fig. 4C) and during seed development (Fig. 4D) using an anti-OsPIMT antibody. Cross-reactivity of the anti-OsPIMT antibody to OcPIMT isoforms was first confirmed using recombinant OcPIMT isoforms (Fig. S9). Subsequent data revealed that in O. coarctata, unlike O. sativa, PIMT protein accumulation was observed until the S3 stage, but very little PIMT was detected at S4, and it became undetectable at S5 and in mature seeds (Fig. 4D). We also analyzed the induction of PIMT during seed desiccation re-establishment using immunoblot analysis in germinating O. sativa seeds after desiccation treatment (Fig. 4E). The results indicated that desiccation does indeed induce PIMT during the re-establishment of desiccation tolerance in germinating O. sativa seeds by both direct and PEG-mediated desiccation (Fig. 4E). The differential accumulation pattern of PIMT proteins in mature seeds of these species was also confirmed by immunolocalization study. In accordance with a previous report (Petla et al., 2016), PIMT protein was found to be concentrated in the embryo (shoot apical meristem, radicle, coleoptile, epiblast) and in the aleurone layer of the endosperm of O. sativa seeds, whereas PIMT was undetectable in any of these tissues in O. coarctata seeds (Fig. 4F, Fig. S10).
PIMTs are differentially expressed during seed maturation in O. sativa and O. coarctata. (A,B) qRT-PCR analysis of OcPIMT and OsPIMT genes in different organs (A) and during the course of seed development (B). The relative expression value of each gene was normalized with the endogenous control 17S rRNA and calculated using the ΔΔCT method (Applied Biosystems). Data are mean±s.d. of three biological repeats. Significant differences among means are denoted by the different letters using Duncan's multiple range test (α=0.05). (C-E) Western blot analysis of PIMTs in mature seeds of O. sativa and O. coarctata 100 µg (in lanes 1 and 2) and 50 µg (in lanes 3 and 4) protein along with recombinant OcPIMTs as positive controls and anti-PIMT antibody as probe (C), during seed development of O. coarctata (D) and during re-establishment of desiccation in O. sativa (E). The membrane was stripped and then re-probed with anti-tubulin antibody. (F) Immunolocalization of PIMT proteins in mature seeds of O. sativa and O. coarctata. Seed sections were probed with anti-PIMT antibody (left) or with secondary antibody only (right). CO, coleoptile; EN, endosperm; EP, epiblast; RA, radicle; SAM, shoot apical meristem; SC, scutellum. Scale bars: 0.2 mm.
PIMTs are differentially expressed during seed maturation in O. sativa and O. coarctata. (A,B) qRT-PCR analysis of OcPIMT and OsPIMT genes in different organs (A) and during the course of seed development (B). The relative expression value of each gene was normalized with the endogenous control 17S rRNA and calculated using the ΔΔCT method (Applied Biosystems). Data are mean±s.d. of three biological repeats. Significant differences among means are denoted by the different letters using Duncan's multiple range test (α=0.05). (C-E) Western blot analysis of PIMTs in mature seeds of O. sativa and O. coarctata 100 µg (in lanes 1 and 2) and 50 µg (in lanes 3 and 4) protein along with recombinant OcPIMTs as positive controls and anti-PIMT antibody as probe (C), during seed development of O. coarctata (D) and during re-establishment of desiccation in O. sativa (E). The membrane was stripped and then re-probed with anti-tubulin antibody. (F) Immunolocalization of PIMT proteins in mature seeds of O. sativa and O. coarctata. Seed sections were probed with anti-PIMT antibody (left) or with secondary antibody only (right). CO, coleoptile; EN, endosperm; EP, epiblast; RA, radicle; SAM, shoot apical meristem; SC, scutellum. Scale bars: 0.2 mm.
ABI-TF-mediated ABA signaling positively regulates PIMT expression in O. sativa orthodox seeds, but is ineffective in O. coarctata recalcitrant seeds
To investigate the mechanism of differential expression of PIMT1 and PIMT2 genes during seed development of O. sativa and O. coarctata seeds, we compared the upstream regulatory region, particularly ABA-responsive cis regulatory elements, of PIMT genes of both species based on the observation that ABA treatment induces PIMT transcription and activity in orthodox seeds (Verma et al., 2013; Wei et al., 2015; Petla et al., 2016). Even though considerable differences were observed among the promoter sequences of PIMT genes between these two species, several ABA-responsive cis elements, such as CATG core (GBL), ACGT core (RY element), CACCG motifs, etc., which are known to bind ABI-TFs ABI3, ABI4 and ABI5, are present in both the PIMT promoters of O. sativa and O. coarctata (Choi et al., 2000; Niu et al., 2002; Monke et al., 2004; Koussevitzky et al., 2007). However, the number and position of these cis elements vary between these two species (Fig. S11). We then assessed and compared the promoter activity of PIMT genes of both species in response to ABA in seeds by generating stable promoter-GUS transgenic lines in the O. sativa background (Fig. S12). These promoter-GUS fusion studies indicate that OcPIMT promoters, like OsPIMT promoters, were able to express the GUS reporter gene in seeds and could induce GUS upon ABA treatment in the O. sativa background (Fig. 5A, Fig. S13). This suggests that reduction of PIMT expression in O. coarctata may not be due to differences in the OcPIMT promoter regions, but may instead be due to inadequate ABA content or ineffectiveness of ABA signaling in O. coarctata seeds. To investigate such possibilities, we analyzed the ABA content in developing and mature seeds of O. coarctata and O. sativa. As shown in Fig. 5B, O. coarctata had significantly higher ABA content in developing seeds (19 DAP) and almost 6-fold more in mature seeds (29 DAP) compared with O. sativa seeds, indicating that the reduced expression of PIMT during seed maturation in O. coarctata is not due to the insufficient ABA content but could be due to ineffectiveness of the ABA signaling pathway. Although PIMT expression is known to be regulated by the ABA signaling pathway, the detailed mechanisms are still not known. In Arabidopsis, it has been indicated that PIMT2 is directly induced by ABI3 (Tian et al., 2020). Therefore, to examine whether the ABA-responsive master regulators ABI3, ABI4 and ABI5 are involved in the regulation of PIMT gene expression in O. sativa, we performed a yeast one-hybrid (Y1H) assay and a promoter-driven GUS assay using proPIMT:GUS in a tobacco leaf transient expression system. In the Y1H assay, it was evident that ABI3, ABI4 and ABI5 were able to bind to both OsPIMT1 and OsPIMT2 promoters with a strength that was comparable to the positive control interaction (Fig. S14). To confirm ABI-TF-mediated regulation of PIMT1 and PIMT2 expression in planta, we carried out a promoter-driven GUS assay in Nicotiana benthamiana leaf by co-transforming proOsPIMT1:GUS or proOsPIMT2:GUS constructs with 35S:OsABI3/4/5:YFP effector constructs. Expression and localization of these transcription factors were initially confirmed by transient expression in N. benthamiana leaves (Fig. S15). As shown in Fig. 5C, ABI3, ABI4 and ABI5 were able to induce proOsPIMT1:GUS or proOsPIMT2:GUS in tobacco leaf. To confirm and quantify these results, we employed a dual luciferase assay in N. benthamiana leaves. Like GUS, luciferase was induced by these ABI-TFs in tobacco leaf. Quantitative analysis revealed that the luciferase activity was significantly higher when tobacco leaves co-transformed proOsPIMT1:LUC with OsABI4 than with OsABI3 or OsABI5, and proOsPIMT2:LUC with either OsABI3 or OsABI5 than with OsABI4 (Fig. 5D,E).
ABI-TF-mediated ABA signaling positively regulates PIMT expression in O. sativa, but is ineffective in O. coarctata. (A) Histochemical localization of GUS activity in vertically dissected seeds treated with or without ABA of OsPIMT and OcPIMT proPIMT(s)-GUS transgenic lines. ‘1’ and ‘2’ represent independent transgenic lines used for GUS staining. (B) Quantification of ABA levels in O. sativa and O. coarctata seeds. (C) Transactivation of proOsPIMT(s)-GUS by ABI-TFs. Histochemical localization of GUS activity analyzed in N. benthamiana leaf agroinfiltrated with a promoter reporter construct (proOsPIMT1/2:GUS) and an effector construct (OsABI-TF:YFP). Different combinations of constructs were infiltrated in four parts of the leaf as indicated. (D,E) Relative quantification of luciferase/Renilla (LUC/REN) ratio in N. benthamiana leaves co-infiltrated with reporter constructs proOsPIMT1:LUC (D) and proOsPIMT2:LUC (E) with effector construct OsABI-TF:YFP as indicated. Data are mean±s.d. of three biological repeats. Significant differences among means are denoted by the different letters using Duncan's multiple range test (α=0.05).
ABI-TF-mediated ABA signaling positively regulates PIMT expression in O. sativa, but is ineffective in O. coarctata. (A) Histochemical localization of GUS activity in vertically dissected seeds treated with or without ABA of OsPIMT and OcPIMT proPIMT(s)-GUS transgenic lines. ‘1’ and ‘2’ represent independent transgenic lines used for GUS staining. (B) Quantification of ABA levels in O. sativa and O. coarctata seeds. (C) Transactivation of proOsPIMT(s)-GUS by ABI-TFs. Histochemical localization of GUS activity analyzed in N. benthamiana leaf agroinfiltrated with a promoter reporter construct (proOsPIMT1/2:GUS) and an effector construct (OsABI-TF:YFP). Different combinations of constructs were infiltrated in four parts of the leaf as indicated. (D,E) Relative quantification of luciferase/Renilla (LUC/REN) ratio in N. benthamiana leaves co-infiltrated with reporter constructs proOsPIMT1:LUC (D) and proOsPIMT2:LUC (E) with effector construct OsABI-TF:YFP as indicated. Data are mean±s.d. of three biological repeats. Significant differences among means are denoted by the different letters using Duncan's multiple range test (α=0.05).
Altogether, these analyses conclusively revealed that the PIMT1 promoter is mainly regulated by OsABI4 whereas the PIMT2 promoter is primarily regulated by OsABI3, OsABI4 and OsABI5 in a differential manner in O. sativa seeds (Fig. 5D,E). To investigate whether the low expression of PIMT genes in O. coarctata seeds is due to the lack of abundance of these ABI-TFs, we compared transcript accumulation of these transcription factors in O. coarctata and O. sativa seeds. We observed that orthodox rice seeds showed high expression of ABI3, ABI4 and ABI5 during seed maturation, whereas expression of these transcription factors was severely reduced in the case of O. coarctata (Fig. S16A-C). Furthermore, we examined chlorophyll and carotenoid content in the seeds of these two species as chlorophyll degradation in maturing seeds is intimately connected to ABA signaling (Koornneef et al., 1984; Parcy et al., 1994; Monma et al., 1994; Nambara et al., 1995; Gonzalez-Jorge et al., 2013). Our data suggest that O. coarctata seeds are unable to follow the process of chlorophyll degradation in seeds, possibly due to severely compromised ABI-TF mediated ABA signaling (Fig. S16D,E). Altogether, ABI-TF expression is severely compromised in O. coarctata seeds, playing a significant role in the expression of PIMT1 and PIMT2 during seed development.
Suppression of PIMT results in reduced seed desiccation tolerance and consequent longevity whereas seed-specific overexpression results in increased seed desiccation tolerance and longevity in O. sativa
In our previous section, we showed that desiccation intolerance and longevity of O. coarctata are linked to reduced PIMT activity. Therefore, we were interested to know whether suppression of PIMT activity alters seed desiccation tolerance and longevity in O. sativa. For this purpose, we generated a maize ubiquitin (ZmUBQ1) promoter-driven intron-spliced hairpin RNA interference (RNAi) construct that targets both OsPIMT1 and OsPIMT2 simultaneously (Fig. S17). Decreased expression of both OsPIMT1 and OsPIMT2 in these lines was confirmed by qRT-PCR, western blot and activity analysis (Fig. S18). Subsequently, to examine whether PIMT overexpression in seeds results in improved seed desiccation tolerance and longevity in recalcitrant rice seeds, we attempted to overexpress PIMT in O. coarctata; however, because of the recalcitrant nature of O. coarctata and lack of established tissue culture methods, as reported in previous studies, genetic transformation of O. coarctata was unsuccessful (Ramanan et al., 1996; Jena, 1994; Finch et al., 1990; Jelodar et al., 1999; Latha et al., 1998). Subsequently, to examine whether PIMT overexpression in seeds results in improved seed desiccation tolerance and longevity in rice seeds, OsPIMT1 and OsPIMT2 were overexpressed in seeds using a rice embryo globulin gene (REG2) promoter. PIMT overexpression lines were confirmed through PIMT transcript, PIMT activity, and protein abundance (Fig. S19). Initially, we analyzed seed desiccation tolerance in terms of germination percentage in transgenic lines. As shown in Fig. 6A, OsPIMT RNAi lines exhibited significantly reduced desiccation tolerance compared with wild-type seeds. Significantly improved seed desiccation tolerance was observed in both OsPIMT1 and OsPIMT2 overexpressed seeds compared with wild-type or segregating null seeds (Fig. 6A). To ascertain whether the reduction of desiccation tolerance of OsPIMT RNAi lines and increased seed desiccation tolerance of the PIMT overexpressing lines are correlated with the PIMT and isoAsp content, the accumulation of isoAsp and PIMT activity were quantified in these lines. Upon imposing seed desiccation, PIMT RNAi lines exhibited significantly increased isoAsp content, whereas PIMT overexpression lines had significantly reduced isoAsp (Fig. S20A,B). Subsequently, we analyzed the effect of CDT on seed germination and seed viability in O. sativa wild-type, OsPIMT1 and OsPIMT2 seed-specific expression lines and OsPIMT RNAi lines before and after CDT, because better seed desiccation tolerance allows seeds to be viable for prolong period of time (longevity). The data suggest that after 10 days of CDT, seed germination and viability was more affected in OsPIMT RNAi than in wild-type seeds, whereas, by contrast, OsPIMT1 and OsPIMT2 seed-specific expression lines showed significantly increased seed germination (Fig. 6B) and viability (Fig. 6C) compared with wild-type seeds. Altogether, these data support the conclusion that desiccation intolerance is strongly associated with reduced PIMT activity and increased isoAsp accumulation, and our study demonstrates that PIMT plays a significant role in seed desiccation tolerance and consequent seed longevity in O. sativa.
Suppression of PIMT results in reduced seed desiccation tolerance and consequent longevity whereas seed-specific overexpression results in increased seed desiccation tolerance and longevity in O. sativa. (A,B) Germination percentage before (mature seeds) and after seed desiccation treatments (A) and before and after CDT (B). (C) Seed viability analysis among seeds of wild type (WT), OsPIMT1 and OsPIMT2 seed-specific overexpression, and OsPIMT RNAi lines. For germination analysis before and after seed desiccation treatments and CDT; data from three independent transgenic lines for each construct and their corresponding null segregant lines (indicated by prime symbol) is shown (WT, wild-type O. sativa; OsPIMT1 1: L2 and null segregant of L2; OsPIMT1 2: L4 and null segregant of L4; OsPIMT1 3: L6 and null segregant of L6; OsPIMT2 1: L2 and null segregant of L2; OsPIMT2 2: L28 and null segregant of L28; OsPIMT2 3: L7 and null segregant of L7; OsPIMT: L75 and null segregant of L75; OsPIMT: L82 and null segregant of L82; OsPIMT 3: L95 and null segregant of L95). Germination was scored after 4 days. Data are mean±s.d. of three repetitions with 50 seeds each. (D-F) qRT-PCR analysis of OsABI3 (D), OsABI4 (E) and OsABI5 (F). The relative expression value of each gene was normalized with endogenous controls 17S rRNA and calculated using the ΔΔCT method (Applied Biosystems). Data are mean±s.d. of three biological repeats. Significant differences among means are denoted by the different letters using Duncan's multiple range test (α=0.05). 1, OsPIMT1 L2; 2, OsPIMT2 L28; 3, OsPIMT L95.
Suppression of PIMT results in reduced seed desiccation tolerance and consequent longevity whereas seed-specific overexpression results in increased seed desiccation tolerance and longevity in O. sativa. (A,B) Germination percentage before (mature seeds) and after seed desiccation treatments (A) and before and after CDT (B). (C) Seed viability analysis among seeds of wild type (WT), OsPIMT1 and OsPIMT2 seed-specific overexpression, and OsPIMT RNAi lines. For germination analysis before and after seed desiccation treatments and CDT; data from three independent transgenic lines for each construct and their corresponding null segregant lines (indicated by prime symbol) is shown (WT, wild-type O. sativa; OsPIMT1 1: L2 and null segregant of L2; OsPIMT1 2: L4 and null segregant of L4; OsPIMT1 3: L6 and null segregant of L6; OsPIMT2 1: L2 and null segregant of L2; OsPIMT2 2: L28 and null segregant of L28; OsPIMT2 3: L7 and null segregant of L7; OsPIMT: L75 and null segregant of L75; OsPIMT: L82 and null segregant of L82; OsPIMT 3: L95 and null segregant of L95). Germination was scored after 4 days. Data are mean±s.d. of three repetitions with 50 seeds each. (D-F) qRT-PCR analysis of OsABI3 (D), OsABI4 (E) and OsABI5 (F). The relative expression value of each gene was normalized with endogenous controls 17S rRNA and calculated using the ΔΔCT method (Applied Biosystems). Data are mean±s.d. of three biological repeats. Significant differences among means are denoted by the different letters using Duncan's multiple range test (α=0.05). 1, OsPIMT1 L2; 2, OsPIMT2 L28; 3, OsPIMT L95.
To gain a deeper insight into PIMT-mediated seed desiccation tolerance, we investigated whether PIMT affected the expression or function of genes/proteins, particularly ABA- and desiccation-induced genes, which play a key role in seed desiccation tolerance. To investigate this, we analyzed the expression of ABI-TFs and their known downstream target genes, such as OsHSF (Kotak et al., 2007), galactinol synthase (OsGolS1) (Zhang et al., 2019) and RESPONSIVE TO ABA GENE 16C (OsRAB16c) (Rabbani et al., 2003) in mature seeds (Fig. 6D-F, Fig. 7A-C) and PEG-treated seeds (Fig. S21) of OsPIMT RNAi and overexpression lines. Overexpression lines had increased expression of these genes as well as ABI-TFs, whereas the RNAi line exhibited similar or reduced expression of these genes compared with wild type. Galactinol synthase (OsGolS1) synthesizes galactinol in raffinose family oligosaccharide (RFO) biosynthesis. Therefore, we compared galactinol accumulation and its precursor molecule Myo-inositol (Fig. 7D,E) in mature seeds of PIMT overexpression, RNAi and wild-type plants. PIMT-overexpressing seeds exhibited significant accumulation of galactinol, whereas RNAi seeds exhibited similar or compromised accumulation of such metabolites compared with wild type. Furthermore, Myo-inositol content was significantly lower in PIMT-overexpressing seeds compared with RNAi and wild-type plants because it acts as a precursor molecule for galactinol synthesis. Accumulation of mannitol in rice seeds during natural aging has also been reported (Yan et al., 2018). Therefore, we analyzed accumulation of mannitol and observed that PIMT-overexpressing seeds exhibited significant accumulation of mannitol, whereas RNAi seeds exhibited compromised accumulation of mannitol compared with wild type (Fig. 7F). We further quantified galactinol accumulation in O. sativa and O. coarctata mature seeds and observed significantly less accumulation of galactinol in O. coarctata than in O. sativa seeds (Fig. 7G). These data confirm that PIMT does indeed facilitate the expression and function of other genes or proteins required for seed desiccation tolerance.
Suppression of PIMT results in reduced expression of downstream genes of ABI-TFs and metabolite accumulation. (A-G). qRT-PCR analysis of OsHSF (A), OsGolS1 (B) and OsRAB16c (C), and GC-MS-based quantification of galactinol (D), Myo-inositol (E) and mannitol (F) in mature seeds of selected lines of PIMT overexpression and RNAi lines. (G) GC-MS-based quantification of galactinol from mature seeds of O. sativa and O. coarctata. The relative expression value of each gene was normalized with endogenous controls 17S rRNA and calculated using the ΔΔCT method (Applied Biosystems). Data are mean±s.d. of three biological repeats. Significant differences among means are denoted by the different letters using Duncan's multiple range test (α=0.05).
Suppression of PIMT results in reduced expression of downstream genes of ABI-TFs and metabolite accumulation. (A-G). qRT-PCR analysis of OsHSF (A), OsGolS1 (B) and OsRAB16c (C), and GC-MS-based quantification of galactinol (D), Myo-inositol (E) and mannitol (F) in mature seeds of selected lines of PIMT overexpression and RNAi lines. (G) GC-MS-based quantification of galactinol from mature seeds of O. sativa and O. coarctata. The relative expression value of each gene was normalized with endogenous controls 17S rRNA and calculated using the ΔΔCT method (Applied Biosystems). Data are mean±s.d. of three biological repeats. Significant differences among means are denoted by the different letters using Duncan's multiple range test (α=0.05).
ABI-TFs are susceptible to isoAsp accumulation upon thermal insult
Next, we investigated whether these ABI-TFs are susceptible to isoAsp formation during seed desiccation. We first checked the sequence of the OsABI-TFs for probable isoAsp modification sites and several highly susceptible sites for isoAsp formation (Asp and Asn residues preceding glycine, serine and histidine) were identified (Clarke, 2003) (Fig. S22). To examine whether these potential Asp and Asn residues do undergo isoAsp formation under stressful environments, recombinant OsABI4 and OsABI5 were purified to near-homogeneity by nickel-charged affinity chromatography (Fig. 8A,B). These purified proteins were subsequently subjected to thermal insult at 37°C for increasing time periods, as described previously (Ghosh et al., 2020a). Afterwards, isoAsp accumulation was quantified in treated and untreated proteins. Our results showed that isoAsp content in OsABI4 and OsABI5 was notably increased upon thermal insult compared with untreated proteins (Fig. 8C,D). Both ABI4 and ABI5 showed significant accumulation of isoAsp as early as 2 h of treatment at 37°C (Fig. 8C,D). To further substantiate whether deamidation of Asn residues really occurs upon stress to form isoAsp, MS/MS analysis of these treated and untreated OsABI4 and OsABI5 proteins was carried out. The Asn residues susceptible to isoAsp formation were identified directly by comparing peptides from a tryptic digest of control and thermally insulted OsABI4 and OsABI5 enzymes by the AB Sciex QTRAP 6600 system. Those sequences that produced more than 50% coverage with 99% confidence were taken into consideration for further analysis. We found that Asn55, Asn200, Asn275, Asn279 and Asn285 in OsABI5 were deamidated and were indeed susceptible to isoAsp formation during stressful environments (Fig. 8E-G, Fig. S23, Table S1). Sequence analysis for probable isoAsp modification showed only Asn45 in OsABI4 as highly susceptible sites for isoAsp formation; however, we could not identify Asn45 in our ESI peptide data for OsABI4 (Table S1).
ABI-TFs are susceptible to isoAsp accumulation upon thermal insult. (A,B) SDS-PAGE analysis of purified recombinant proteins OsABI4 (A) and OsABI5 (B). M, marker. (C,D) isoAsp accumulation in OsABI4 (C) and OsABI5 (D) upon thermal insult. Data are mean±s.d. of three biological repeats. Significant differences among means are denoted by the different letters using Duncan's multiple range test (α=0.05). (E-G) ESI mass spectra of tryptic peptide resulting from treated OsABI5. Site(s) of deamidation are represented by square brackets enclosing residue(s) of the peptide sequence (e.g. [Dea]). Analysis of y and b ion fragmentation patterns with Protein Pilot version 5.0.1 are also indicated.
ABI-TFs are susceptible to isoAsp accumulation upon thermal insult. (A,B) SDS-PAGE analysis of purified recombinant proteins OsABI4 (A) and OsABI5 (B). M, marker. (C,D) isoAsp accumulation in OsABI4 (C) and OsABI5 (D) upon thermal insult. Data are mean±s.d. of three biological repeats. Significant differences among means are denoted by the different letters using Duncan's multiple range test (α=0.05). (E-G) ESI mass spectra of tryptic peptide resulting from treated OsABI5. Site(s) of deamidation are represented by square brackets enclosing residue(s) of the peptide sequence (e.g. [Dea]). Analysis of y and b ion fragmentation patterns with Protein Pilot version 5.0.1 are also indicated.
PIMT physically interacts with ABI-TFs and repairs isoAsp residues, and positively influences expression of their target gene(s)
In our previous section, we observed that the OsABI4 and OsABI5 are susceptible to isoAsp formation, leading to reduced transcript accumulation of several ABA-responsive genes along with ABI transcription factors in PIMT RNAi lines, and markedly increased accumulation in overexpressing lines when compared with the wild-type plants. These observations raised an intriguing possibility of the involvement of PIMT in facilitating the activity of these ABI-TFs (ABI3, ABI4 and ABI5). To investigate such possibilities, we initially determined whether PIMT1 and ΔPIMT2 (the major PIMT2 isoform in seed) physically interact with these ABI-TFs in yeast cells through yeast two-hybrid (Y2H) assays (Petla et al., 2016). For this purpose, OsPIMT(s) coding sequences were subcloned into the pGBKT7 BD vector, whereas ABI-TFs (ABI3, ABI4 and ABI5) were cloned in the pGADT7 AD vector. As shown in Fig. 9A, both PIMT1 and ΔPIMT2 showed positive interactions with these transcription factors. To confirm these interactions in plant cells, we performed a bimolecular fluorescence complementation (BiFC) assay in N. benthamiana leaves. For this purpose, PIMT (PIMT1 and PIMT2) was transiently expressed as a YFP N-terminal fusion protein, whereas ABI-TFs were expressed as a YFP C terminal-fused protein in N. benthamiana leaves. We also used catalase and ascorbate peroxidase and Arabidopsis RNA helicase, which are known substrates of PIMT, as positive controls (Nayak et al., 2013; Petla et al., 2016; Ghosh et al., 2020a). Reconstituted fluorescence was detected in the nucleus when PIMT and any of these transcription factors were expressed together (Fig. 9B, Fig. S24). Collectively, our results suggest that PIMT proteins physically interact with ABI3, ABI4 and ABI5. To check whether PIMT is able to restrict isoAsp accumulation in vitro, we conducted PIMT repair assays with thermally insulted OsABI4 and OsABI5 proteins, and then isoAsp content was determined. To perform the PIMT repair assay, either catalytically active PIMT (ΔΟsPIMT2) or catalytically inactive PIMT (OcPIMT1-2) and AdoMet were used (Fig. 3B) with heat-treated OsABI4 and OsABI5 as substrates. Following the PIMT repair assay, isoAsp accumulation was measured. Results showed that, following thermal insult, the level of isoAsp accumulation in these proteins was significantly reduced when both catalytically active PIMT and AdoMet were present in the PIMT repair reaction mix (Fig. 9C,D). To confirm whether the decrease of isoAsp content in these ABI-TF proteins was only because of PIMT activity, a catalytically inactive PIMT isoform (OcPIMT1-2) was used in the reaction mix (Fig. 3B). As opposed to catalytically active PIMT, isoAsp content in OsABI4 and OsABI5 was found to be unaltered in the presence of catalytically inactive PIMT isoforms (OcPIMT1-2) and AdoMet (Fig. 9C,D). These results strongly suggest that PIMT repairs isoAsp accumulation, which may consequently restore the activity of OsABI4 and OsABI5. Subsequently, to validate this observation in planta, OsABI4-10xMYC and OsABI5-10xMYC were infiltrated in tobacco alone or co-infiltrated with ΔOsPIMT2: YFP in N. benthamiana leaves (Fig. 9E). Infiltrated plants were then exposed to high temperature for 24 h and then proteins were collected from infiltrated leaf samples. ABI4 and ABI5 proteins were pulled with a myc antibody and used for isoAsp quantification. Results revealed that the isoAsp content in OsABI4 and OsABI5 was significantly lower when co-infiltrated with ΔOsPIMT2 compared with plants infiltrated with only OsABI4 and OsABI5 (Fig. 9F). Similar results were obtained when plants were infiltrated with OsABI4 or OsABI5 alone or with ΔOsPIMT2 and treated with dehydration stress instead of heat stress (Fig. S25). These results indicated that OsABI4 and OsABI5 are indeed susceptible to isoAsp formation in planta. Next, we analyzed the transcriptional activity of these ABI-TFs (ABI3, ABI4 and ABI5) using promoters of their respective target genes through a dual luciferase assay system in the presence or absence of ΔOsPIMT2 before and after heat treatments. The result of these analyses revealed that luciferase activity is significantly higher when proOsHSF:LUC were co-infiltrated with OsABI3:YFP and ΔOsPIMT2:YFP than with only OsABI3:YFP (Fig. 10A). We also analyzed the activity of OsABI4:YFP and OsABI5:YFP using proOsPIMT1:LUC and proOsGolS1:LUC, respectively, in the presence and absence of ΔOsPIMT2:YFP before and after heat treatments (Fig. 10B,C). Collectively, our results strongly suggest that ABI-TFs acquire isoAsp residues, which compromise the transcriptional activity of these ABI-TFs, eventually negatively influencing the function of downstream genes required for desiccation tolerance; however, PIMT repairs that damage to keep these ABI-TFs functionally competent (Figs 9-11).
PIMT physically interacts with and repairs isoAsp modification in ABI-transcription factors. (A,B) Interaction analyses of OsPIMTs and ABI-TFs by Y2H (A) and BiFC (B). (A) Y2H interactions between OsPIMTs and OsABIs [positive (PC) and negative (NC) controls for interaction provided by manufacturer]. (B) BiFC analysis in N. benthamiana. OsPIMTs were fused to the N-terminal fragment of YFP (nYFP) and OsABI-TFs were fused to the C-terminal fragment of YFP (cYFP). Images were taken under confocal microscopy. YFP, YFP fluorescence channel; BF, brightfield channel; merge, merged image of YFP and BF channel. Scale bars: 10 μm. (C,D) Relative isoAsp content of thermally insulted OsABI4 (C) and OsABI5 (D) with and without PIMT repair. Following heat treatment, PIMT repair assay was carried out with heat-treated OsABI4 (C) and OsABI5 (D). (E) Immunoblot analyses of ΔOsPIMT2:YFP, OsABI4:Myc and OsABI5:Myc expressed in N. benthamiana with anti-YFP or anti-Myc antibodies. (F) Analysis of in vivo isoAsp accumulation in immunoprecipitated OsABI4 and OsABI5 protein co-infiltrated with ΔOsPIMT2:YFP or alone.
PIMT physically interacts with and repairs isoAsp modification in ABI-transcription factors. (A,B) Interaction analyses of OsPIMTs and ABI-TFs by Y2H (A) and BiFC (B). (A) Y2H interactions between OsPIMTs and OsABIs [positive (PC) and negative (NC) controls for interaction provided by manufacturer]. (B) BiFC analysis in N. benthamiana. OsPIMTs were fused to the N-terminal fragment of YFP (nYFP) and OsABI-TFs were fused to the C-terminal fragment of YFP (cYFP). Images were taken under confocal microscopy. YFP, YFP fluorescence channel; BF, brightfield channel; merge, merged image of YFP and BF channel. Scale bars: 10 μm. (C,D) Relative isoAsp content of thermally insulted OsABI4 (C) and OsABI5 (D) with and without PIMT repair. Following heat treatment, PIMT repair assay was carried out with heat-treated OsABI4 (C) and OsABI5 (D). (E) Immunoblot analyses of ΔOsPIMT2:YFP, OsABI4:Myc and OsABI5:Myc expressed in N. benthamiana with anti-YFP or anti-Myc antibodies. (F) Analysis of in vivo isoAsp accumulation in immunoprecipitated OsABI4 and OsABI5 protein co-infiltrated with ΔOsPIMT2:YFP or alone.
PIMT positively influences ABI transcription factors target gene expression. (A-C) Relative quantification of luciferase/Renilla (LUC/REN) ratio in N. benthamiana leaves co-infiltrated with different reporter and effector constructs. Schematics of different reporter and effector constructs used are shown on the left, with or without ΔOsPIMT2:YFP before (Control) and after heat treatment (Heat), respectively. Data are mean±s.d. of three biological repeats. Significant differences among means are denoted by the different letters using Duncan's multiple range test (α=0.05).
PIMT positively influences ABI transcription factors target gene expression. (A-C) Relative quantification of luciferase/Renilla (LUC/REN) ratio in N. benthamiana leaves co-infiltrated with different reporter and effector constructs. Schematics of different reporter and effector constructs used are shown on the left, with or without ΔOsPIMT2:YFP before (Control) and after heat treatment (Heat), respectively. Data are mean±s.d. of three biological repeats. Significant differences among means are denoted by the different letters using Duncan's multiple range test (α=0.05).
A working model depicting the role of ABI-TFs and the PIMT module in O. sativa seed desiccation tolerance and subsequent longevity. Seed maturation, including seed desiccation tolerance, is mediated by the accumulation of ABA and its downstream signaling components, ABI-TFs. Coordinated action of ABA and ABI-TFs subsequently induces PIMT in orthodox O. sativa seeds. ABI-TFs undergo isoAsp modification during stressful environments, seed desiccation and storage, possibly due to reactive oxygen species accumulation. PIMT interacts with and repairs isoAsp residues accumulated in ABI-TFs to maintain the functional competence of ABI-TFs to impart seed desiccation tolerance and subsequent longevity and germination vigor in O. sativa seeds.
A working model depicting the role of ABI-TFs and the PIMT module in O. sativa seed desiccation tolerance and subsequent longevity. Seed maturation, including seed desiccation tolerance, is mediated by the accumulation of ABA and its downstream signaling components, ABI-TFs. Coordinated action of ABA and ABI-TFs subsequently induces PIMT in orthodox O. sativa seeds. ABI-TFs undergo isoAsp modification during stressful environments, seed desiccation and storage, possibly due to reactive oxygen species accumulation. PIMT interacts with and repairs isoAsp residues accumulated in ABI-TFs to maintain the functional competence of ABI-TFs to impart seed desiccation tolerance and subsequent longevity and germination vigor in O. sativa seeds.
DISCUSSION
During seed maturation, orthodox seeds acquire seed desiccation tolerance, in which ABI-mediated regulatory networks play an indispensable role as ABI transcription factors (ABI3, ABI4 and ABI5) activate a large set of genes that are involved in several mechanisms that contribute to seed desiccation tolerance. Some of these transcription factors were also shown to play important role in regulating the germination process (Koornneef et al., 1984; Parcy et al., 1994; Monma et al., 1994; Nambara et al., 1995; Gonzalez-Jorge et al., 2013; Lopez-Molina et al., 2001). Therefore, the functionality of these transcription factors must be protected during seed development, particularly during maturation when cellular environments become favorable for protein modifications and damage due to the gradual loss of water and increased reactive oxygen species. However, how these proteins are protected from such damage to maintain their unabated function throughout the drying process during seed maturation remains an intriguing question. The present study demonstrates that protein-repairing enzyme PIMT plays a key role in the acquisition of seed desiccation tolerance by protecting isoAsp-mediated protein damage and functions during seed maturation. In contrast to orthodox O. sativa seeds, recalcitrant O. coarctata seeds exhibit significantly reduced PIMT activity, resulting in over-accumulation of isoAsp in proteins in maturing seeds, which may contribute to desiccation intolerance upon maturity (Fig. 2). Our comparative expression and biochemical analyses demonstrate that O. coarctata encodes two biochemically active PIMT isoforms that share more than 90% sequence identities and similar enzymatic properties with their respective OsPIMT isoforms. O. coarctata encodes two biochemically active PIMT isoforms but lacks coordinated action of ABA and ABI-TFs to upregulate PIMT during the late maturation phase whereas, in contrast, PIMT is markedly induced during seed maturation in O. sativa seeds (Figs 2, 4 and 5). These analyses during rice seed development reveal that the PIMT-mediated repair process, which seems to be crucial for seed desiccation tolerance and longevity of orthodox seeds, is highly compromised in desiccation-sensitive recalcitrant seeds of O. coarctata. The positive relationship between PIMT activity and seed longevity and desiccation tolerance was further confirmed by genetic studies using PIMT overexpression and RNAi lines (Figs 6, 7). We show that suppression of PIMT activity significantly reduces seed longevity and desiccation tolerance, and that seed-specific expression enhances seed longevity and desiccation tolerance to O. sativa seeds. Our findings are consistent with previous studies that have shown that several mechanisms or processes that confer desiccation tolerance are either absent or ineffective in many desiccation-intolerant recalcitrant seeds. For example, several specific LEA proteins, which are required for desiccation tolerance in orthodox seeds of Medicago trancatula, are highly reduced or absent in desiccation-sensitive seeds of the phylogenetically close species Castanospermum australe (Delahaie et al., 2013). Furthermore, similar results were also observed in the desiccation-tolerant midge Polypedilum vanderplanki: the Anhydrobiosis-Related gene Island (ARId), which includes PIMT along with heatshock proteins, LEA proteins, core components of antioxidants, aquaporins, and trehalose metabolic genes, is present in Polypedilum vanderplanki, but absent in the desiccation-sensitive midge P. nubifer (Gusev et al., 2014). It has been well established by previous studies that induction of seed-desiccation tolerance was shown to be dependent on ABA signaling, which is achieved by the accumulation of ABA and subsequent coordinated expression and action of various master regulators of ABA signaling components during the seed maturation phase (Koornneef et al., 1984; Parcy et al., 1994). Interestingly, comparatively greater accumulation of ABA in O. coarctata recalcitrant seeds than the O. sativa orthodox seeds suggests that the lack of desiccation tolerance in O. coarctata is not linked to ABA accumulation, but rather is due to the ABA signaling pathway in O. coarctata failing or being compromised in some way. Our subsequent analyses reveal that PIMT1 and PIMT2 genes in O. sativa seeds are primarily regulated by ABI3, ABI4 and ABI5 in a differential manner. In Arabidopsis, it has been shown that PIMT2 is directly induced by ABI3 (Tian et al., 2020). However, in O. coarctata seeds, ABI3, ABI4 and ABI5 expression is significantly reduced, and hence ABA-induced genes, including PIMT, exhibit reduced expression in recalcitrant seeds (Fig. S14). We also show that PIMT physically interacts with ABI3, ABI4 and ABI5 and modulates their target gene expression, possibly by repairing isoAsp residues (Figs 8-10). It is evident in our study that transcript accumulation of several ABA-regulated genes, which play a significant role in seed desiccation tolerance, are highly compromised in PIMT RNAi lines but increased in PIMT overexpression lines (Figs 6, 7). Our data also suggest that ABI3, ABI4 and ABI5 are likely to undergo isoAsp modifications (Figs 8-10), which not only affect expression of their target genes but also their own expression, as these transcription factors are known to cross-regulate their own expressions and act combinatorially during seed maturation (Koornneef et al., 1989; Bossi et al., 2009). Previous studies have also identified several proteins, particularly those which reside in seeds under low moisture content, including PRH75, catalase, superoxide dismutase, for which functions are compromised by isoAsp formation but subsequently repaired by PIMT (Nayak et al., 2013; Ghosh et al., 2020a).
In summary, our present study offers a new insight into the mechanisms of desiccation tolerance in orthodox seeds, whereby the PIMT-mediated protein repair system plays a decisive role by protecting the functionality of regulatory proteins, which are crucial for the acquisition of seed-desiccation tolerance, by repairing isoAsp induced damage; however, such mechanisms of desiccation tolerance were nonfunctional in recalcitrant seeds (Fig. 11).
MATERIALS AND METHODS
Plant materials and growth conditions
Oryza coarctata (also known as Porteresia coarctata) Roxb Tateoka plants were collected from riverbank of the Sunderbans in the coastal regions of the Bay of Bengal, India, and were maintained in a growth chamber. Oryza sativa (indica) seeds were obtained from the Indian Agricultural Research Institute (IARI), New Delhi, India, and grown for seed collection. Transgenic plants were grown in a growth chamber. For Oryza sativa and Oryza coarctata, the growth chamber was maintained in a 16 h light (250 µmol m−2 s−1 light intensity)/8 h dark cycle at 28/23°C day/night and 55% humidity. For collection of seed development samples, flowers were tagged on the day of pollination (Agarwal et al., 2011; Petla et al., 2016) and collected on different days after pollination and flash-frozen in liquid nitrogen and stored at −80°C until used. Just before analysis, seeds were pulled according to Fig. S1 as S1 to S5 stages. N. benthamiana were used in this study and were grown and maintained in a plant growth facility maintained at 22°C±2°C, relative humidity of 60% with a 16 h light (200 µmol m−2 s−1 light intensity)/8 h dark cycle (Ghosh et al., 2020a).
Preparation of protein extracts, measurement of PIMT activity and isoAsp quantification
Total proteins were extracted from different organs and tissues of O. sativa and O. coarctata as described by Petla et al. (2016). Protein concentration was estimated according to the methods of Bradford (1976). PIMT activity was measured using the vapor diffusion method described by Mudgett and Clarke (1993) with minor modifications (Verma et al., 2010). The assay mixture (total of 30 µl) contained 10 µM [3H] AdoMet (370 GBq/mmol; Perkin Elmer) and 625 µM of Val-Tyr-Pro-(L-iso Asp)-His-Ala [VYP-(L-iso D)-HA] (Peptron, Korea) and HEPES buffer pH 7.5, unless otherwise mentioned. The 30 µl reaction was allowed to continue at 37°C for 1 h and then stopped by quenching with 30 µl of freshly prepared 0.2 N NaOH and 1% SDS, which results in hydrolysis of the methyl ester to methanol. Radioactivity was measured using a liquid scintillation counter (Verma et al., 2010). Fifty micrograms of crude protein was used for the PIMT assay or isoAsp estimation. PIMT activity was calculated by subtracting isoAsp peptide-independent (endogenous) activity from the isoAsp peptide-dependent activity.
ISOQUANT isoAsp detection kit (Promega) was used for the estimation of isoAsp according to the kit's protocol with minor alterations as described by Ghosh et al. (2020a).
Activity of the antioxidant proteins ascorbate peroxidase (APX), superoxide dismutase (SOD) and catalase (CAT) were analyzed as described by Ghosh et al. (2020a).
Extraction of RNA and quantitative real-time PCR
Total RNA was extracted from developing and mature seeds of O. sativa and O. coarctata as described by Singh et al. (2003). cDNA was synthesized using 2 µg of DNaseI-treated total RNA following the manufacturer's instructions (Verso cDNA synthesis kit, Thermo Scientific). qRT-PCR was carried out as described by Ghosh et al. (2020b) on an ABI StepOne Real-Time PCR system using suitable primer pairs (Table S2). The expression of PIMT genes was normalized to the expression of UBQ5, EF1α, 17S rRNA or UBQ10.
Isolation and cloning of PIMT gene and promoter sequences
Five prime and 3′ RACE was performed to determine the untranslated region of OcPIMT genes using a RACE kit (Invitrogen) following the manufacturer's protocol. The full-length PIMT cDNA sequences from O. coarctata were amplified using gene-specific primers based on PIMT sequences of O. sativa. All cDNAs were cloned in the pJET1.2 vector and subsequently sequenced. Full-length genomic sequences of OcPIMT genes were also amplified using genomic DNA and cloned in the pJET1.2 vector and subsequently sequenced.
Genomic DNA containing the promoter regions of OcPIMT genes were amplified from total genomic DNA by genome walking using a GenomeWalker kit (BD Biosciences).
Bacterial overexpression and purification of recombinant PIMT(s)
OcPIMT cDNA was subcloned into the NdeI/XhoI sites of the expression vector pET-23b (Novagen). For expression of recombinant OsABI4 and OsABI5, respective amplicons were cloned, via pENTR/D-TOPO, into pDEST17 to provide a His-tagged recombinant OsABI4 and OsABI5 using Gateway technology (Invitrogen). The resulting plasmids were introduced into the host strain Escherichia coli BL21 (DE3) or Lemo21 cells (Sambrook and Russell, 2001). Transformed E. coli cells were grown in Luria–Bertani medium (10 g Tryptone, 10 g NaCl and 5 g yeast extract per l) at 37°C up to A600 of 0.5 absorbance units and induced by Isopropyl β-d-1-thiogalactopyranoside for 8 h at 25±1°C. The bacterial cells were collected by centrifugation (at 6000 g for 10 min) and lysed by sonication in a buffer containing 20 mM phosphate buffer pH 7.5, 10 mM β-mercaptoethanol, supplemented with bacterial protease inhibitor cocktail (Sigma-Aldrich). The extracts were analyzed by 12% SDS-PAGE as described previously by Salvi et al. (2016). Proteins expressed in pellet fraction were solubilized as described previously by Majee et al. (2004). The solubilized proteins from the pellet, as well as proteins expressed in soluble fractions, were purified to homogeneity using nickel-charged affinity columns (GE Healthcare). Five micrograms of purified protein were used for the PIMT activity. Proteins were visualized by SDS-PAGE by staining with CBB R-250 and then images were captured using the Bio-Rad Gel Doc system XR+.
Thermal stress treatment to OsABI4 and OsABI5 and restoration by PIMT
For stress treatments, purified recombinant OsABI4 and OsABI5 proteins (50 µg) were separately incubated at 37°C for 0, 2, 4 or 8 h with an untreated set as control. Samples were collected at each time point and were analyzed for isoAsp accumulation as described above. PIMT repair reactions were performed for 1 h at 37°C using heat-treated OsABI4 and OsABI5 as the substrate with enzymatically active (ΔOsPIMT2) and inactive (OcPIMT1-2) PIMT. The repair reaction consisted of 30 µM PIMT, 10 µM AdoMet in 0.1 M HEPES buffer (pH 7.5). After PIMT repair reactions, isoAsp content was measured as described above.
Sample preparation and LC-MS/MS analysis
Thermally treated and untreated purified recombinant OsABI4 and OsABI5 protein samples were run on 12% SDS-PAGE. In-gel digestion was performed using trypsin (Promega) following reduction with DTT and alkylation with iodoacetamide as described by Ghosh et al. (2020). Eluted proteins were collected, vacuum-dried and reconstituted in 0.1% (v/v) formic acid and used for LC-MS/MS on AB Sciex 250 TripleTOF 6600 with Ekspert nano-LC 425 system with C18 column (0.075×150 mm, 3 µm) using DuoSpray Ion drive electrospray ionization (ESI) source. The peptide mixture was fractionated using a linear gradient of 2% acetonitrile (0.1% formic acid) to 90% acetonitrile (0.1% formic acid) at a flow rate of 300 µl/min for 60 min. Spectra were acquired at 3 Hz with a mass range of m/z 400-1200 for mass spectrum 1 (MS1). The additional top three intense peaks were selected for MS/MS analysis with a scan rate of 10,000 Da/s in positive mode. LC-MS/MS data were processed with Analyst 1.7 (AB Sciex) as described by Ghosh et al. (2020a) and peak lists were searched against 435,086 proteins of the UniProt rice proteome database.
Western blot and immunolocalization analysis
Western blot analysis was carried out using suitable antibodies as described by Petla et al. (2016). Protein samples were separated by 12% SDS PAGE and then electroblotted onto a PVDF membrane. The membranes were probed with suitable antibodies and subsequently developed using Luminata Forte western detection reagents (Bio-Rad). the antigenic peptide QNDASVRYVPLTSRSAQ, which corresponds to any PIMT protein with Post region III (VRYPLTS), was custom synthesized. See Table S2 for a list of antibodies along with the dilution used in this study.
Immunolocalization of PIMT proteins in mature and developing seeds of O. sativa and O. coarctata was carried out using an anti-PIMT antibody as described by Petla et al. (2016).
Plasmid construction and generation of transgenic plants
To make the OsPIMT1 and OsPIMT2 seed-specific overexpression construct, OsPIMT1 and OsPIMT2 were subcloned into the binary vector pCAMBIA1301 under the control of the OsREG2 promoter (Qu and Takaiwa, 2004). To generate knockdown lines of OsPIMT1 and OsPIMT2, intron-spliced hairpin RNAi strategies were used. For RNAi constructs, unique sequences of OsPIMT1 and OsPIMT2 were chosen as the targeted interference region. Subsequently, the ihpRNAi cassette was developed using Gateway technology in the pANDA vector under the control of the maize ubiquitin1 promoter following the manufacturer's protocol (Invitrogen).
For promoter GUS analysis, individual OcPIMT and OsPIMT promoters were fused with the β-glucuronidase (GUS) reporter gene via pENTR/D-TOPO into promoter-GUS fusion vector using Gateway technology (Invitrogen).
Agrobacterium-mediated transformation of rice was carried out as described by Petla et al. (2016). Transgenic plants were selected based on suitable antibiotic resistance-specific gene integration, with transcript analysis followed by western blot and PIMT activity analysis.
ABA quantification and GC-MS analysis
ABA was quantified from developing and mature seeds of O. sativa and O. coarctata following a method described by Vadassery et al. (2012) with minor modifications. Seeds were ground in liquid nitrogen, and 250 mg frozen powder was homogenized and extracted in methanol containing an internal standard D6-ABA (OIChemIm, http://www.olchemim.cz/). The extracted supernatant sample was lyophilized, reconstituted and finally injected into an AB Sciex QTRAP 6500 Exion AD LC system (AB Sciex) with an Acquity UPLC BEH C18 column (2.1 100 mm, 1.7_m, Waters) using turbo spray ion drive electrospray ionization (ESI) source.
For metabolite extraction of O. sativa and O. coarctata seeds, the method described by Kundu et al. (2018) was followed with minor modifications. Seeds were homogenized using liquid nitrogen in 1 ml chilled methanol (HPLC grade) and then adonitol (0.2 mg/ml) as an internal standard was added to this extraction. The samples were mixed, vortexed and incubated at 70°C for 15 min. Subsequently, 800 μl of HPLC grade water and 600 μl of chloroform were added. After centrifugation at 10,000 g for 15 min, the supernatant was collected and lyophilized until the complete solvent evaporated. For derivatization, dried lyophilized samples were suspended in 80 μl of methoxyamine hydrochloride in 20 mg/ml pyridine and then incubated at 30°C for 90 min. Trimethylsilylation was carried out using 80 μl MSTFA [N-methyl-N-(trimethylsilyl) trifluoroacetamide] followed by incubation at 37°C for 30 min. Samples were then centrifuged at 14,000 g for 5 min and supernatants were used for further analysis.
The GC-MS analysis was performed using a Shimadzu QP2010 Ultra using Rtx-5Sil-MS column (0.25 mm×30 m×0.25 μm, Restek Corporation). The temperature program consisted of 60°C isothermal heating for 2 min, followed by a ramp rate of 5°C/min to 250°C with hold time 5 min and, subsequently, a final ramp rate of 10°C/min to 280°C with hold time 15 min. The chromatograms were examined using GC-MS solution software. The NIST12 library was used for peak identification.
Y1H and Y2H assays
Interactions between OsPIMT promoters and ABI transcription factors were determined through Y1H assay using the Matchmaker® Gold Yeast One-Hybrid Library Screening System (Clontech) according to the manufacturer's instructions. Promoter regions or oligos consisting cis-acting elements of OsPIMTs were cloned upstream of the Aureobasidin A (AUR1-C) gene into the pAbAi vector. The resulting constructs were linearized and integrated into the genome of yeast Y1H Gold cells. Coding sequences of ABI transcription factors were cloned into the pDEST-GADT7 plasmid using Gateway technology (Invitrogen). Yeast cells containing the intergrated promoter sequence were then transformed with pDEST-GADT7:ABI constructs. DNA–protein interaction was determined based on the ability of the co-transformants to grow on SD/-Ura medium in the presence of Aureobasidin A.
For Y2H one-to-one protein-protein interaction, OsPIMT coding sequences were sub-cloned into the pDEST-GBKT7BD vector using Gateway technology (Invitrogen), and ABI-TFs (ABI3, ABI4 and ABI5) were cloned into the pDEST-GADT7AD vector. Plasmids were co-transformed into the Y2H Gold strain (Clontech) and then transformed cells were selected on dropout (-Ade, -His, -Leu, -Trp) with X-α-Gal and Aureobasidin A agar media as described by Rao et al. (2018).
Transactivation of promoter analysis through transient reporter GUS expression in N. benthamiana leaf
For in vivo transient analysis of promoter activity promoter regions for O. coarctata and O. sativa PIMTs were sub-cloned in a promoter-GUS fusion vector using Gateway technology (Invitrogen). Transcription factors that showed interaction in the Y1H analysis were sub-cloned in pEG101 to provide a C-terminal YFP tag and resulting plasmids were transformed to Agrobacterium strain EHA109 and then co-infiltrated into N. benthamiana leaf. Histochemical staining for β-glucuronidase (GUS) was performed using GUS staining solution [100 mM sodium phosphate buffer (pH 7.0), 0.1% Triton X-100, 0.5 mM potassium ferricyanide (K3FeCN6), 0.5 mM potassium ferrocyanide (K4FeCN6), and 1 mM 5-bromo-4-chloro-3-indolyl-β-D-glucuronide (X-gluc) (Gold Biotechnology)] (Jefferson, 1987). Leaf samples were bleached in acetic acid:glycerol:ethanol (1:1:3, v/v) and photographed (Salvi et al., 2020). To perform GUS staining in seeds, we imbibed rice seed of the two independent homozygous transgenic lines of proPIMT1: GUS and proPIMT2: GUS from O. coarctata and O. sativa in 100 µM ABA (Sigma-Aldrich) or only water for 24 h. Seeds were then vertically sectioned in half and placed in the GUS buffer and incubated at 28°C until stain was visible.
Dual luciferase reporter assay
For the quantitative luciferase (LUC) assay, the promoter regions were fused with LUC to generate proOsPIMT1: LUC, proOsPIMT2: LUC; proOsHSF: LUC and proOsGols1: LUC constructs into pGreen Luc II vector with Renilla as an internal control. The resulting plasmids were transformed to Agrobacterium strain EHA109 and co-infiltrated into N. benthamiana leaf with ABI-TFs fused to C-terminal YFP tag. For the dual luciferase assay, N. benthamiana leaf were similarly co-infiltrated with promoter:LUC reporter and transcription factors fused to C-terminal YFP tag constructs with or without the ΔOsPIMT2:YFP construct as described previously. Infiltrated plants were incubated overnight at 24°C in the dark followed by incubation in a 16 h light/8 h dark cycle for 24 h. Following heat treatment to half of the plants, leaf samples were collected from heat treated and normally grown (infiltrated control) plants. Dual LUC assays were performed using a dual luciferase reporter assay system (Promega) according to the manufacturer's instructions (Ju et al., 2019).
Seed desiccation tolerance
To evaluate the desiccation tolerance in O. sativa a previously published protocol (Downie et al., 2003; Maia et al., 2011) was followed. Both ‘direct desiccation’ and ‘PEG-induced desiccation’ were carried out. Seeds were imbibed on two layers of wet filter paper for 30-36 h. Just before radical emergence, seeds were placed in 50% polyethylene glycol (PEG 8000) (PEG-induced desiccation). For direct desiccation, seeds were incubated for 3 days at 25°C at 8% relative humidity achieved using saturated LiCl solution in a closed chamber. During direct desiccation and PEG-induced desiccation, samples were taken at different time intervals to measure PIMT activity and isoAsp content. After 3 days of incubation, seeds were humidified in humid air (100% relative humidity) for 24 h at 28°C in the dark, and then seeds were thoroughly washed with water; rehydrated in H2O at 28°C in a 16 h light/8 h dark regime and then germination was analyzed.
BiFC
OsPIMTs were subcloned into pSITE BiFC nEYFP-C1 vector and OsABI transcription factors were subcloned into pSITE BiFC-cEYFP-N1 vectors via pENTR/D-TOPO (Martin et al., 2009) using Gateway technology (Invitrogen). The resulting plasmids were transformed to Agrobacterium strain EHA109 (Sambrook and Russell, 2001) and were co-infiltrated into 1-month-old N. benthamiana leaves as described previously. Reconstitution of YFP fluorescence in leaf samples was observed after 48 h.
isoAsp content in immunoprecipitated protein samples
OsABI4/OsABI5 transcription factors were cloned into the plant expression vector pGWB420 to obtain a C-terminal 10X Myc tag using Gateway technology (Invitrogen). The resulting plasmid was transformed to Agrobacterium strain EHA109 (Sambrook and Russell, 2001) and then infiltrated alone or co-infiltrated with ΔOsPIMT2:YFP into 1-month-old N. benthamiana leaves as described previously. Infiltrated plants were incubated overnight at 24°C in the dark followed by incubation with 16 h light/8 h dark cycle for 24 h. Following heat or dehydration treatment, leaf samples were collected from heat-treated and normally grown infiltrated control plants (Nayak et al., 2013; Ghosh et al., 2020a). Total crude proteins were extracted in IP buffer (50 mM Tris pH 7.4,150 mM NaCl, 1% Triton X-100, 1 mM EDTA, 1 mM MgCl2, and protease inhibitor cocktail). Subsequently, myc-tagged OsABI4/OsABI5 was immunoprecipitated with an anti-Myc antibody coupled to the protein A sepharose resin by incubating overnight at 4°C with gentle shaking. The beads were collected by centrifugation (at 12,000 g for 30 s) and washed several times. The protein was then eluted and isoAsp content in immunoprecipitated protein samples was analyzed as described previously.
CDT, tetrazolium assay and germination analysis
For germination analysis of mature and developing seeds, flowers were tagged on the day of pollination and collected at 19 DAP and were used for the germination analysis and tetrazolium staining (Berridge et al., 1996; Verma and Majee, 2013). For germination, seeds were imbibed either in water or 50 µM gibberellic acid (GA3). CDT and seed viability analysis using tetrazolium staining were carried out as described by Petla et al. (2016).
Estimation of total chlorophyll and carotenoid content
Chlorophyll A, chlorophyll B and total chlorophyll content in mature seed were estimated as described by Ghosh et al. (2020b). Carotenoid content was measured using the method of Saxena et al. (2020).
Phylogenetic analysis of PIMT proteins
PIMT protein sequences from various organisms of different taxonomic groups were derived through BLAST searches from the National Centre for Biotechnology Information (NCBI) and phytozome from the Joint Genome Institute (JGI) database. Multiple sequence alignment for PIMT proteins was generated using Clustal W incorporating local pairwise alignment. Maximum likelihood analysis was conducted for the construction of a phylogenetic tree with 1000 replicates and the final tree was created and the tree was visualized using the Tree explorer tool of MEGA 6.
Statistical analysis
All data presented in this study are expressed as mean±s.d. Standard deviations were calculated from three biological replicates. The statistical analysis was conducted by Student's t–test (**P<0.0001 or *P<0.05) or one-way analysis of variance (ANOVA) to authenticate the validity of results. Duncan's multiple range test (α=0.01/0.05) was performed using the SPSS program (SPSS Statistics, IBM), to test the statistical significance. Letters on the graphs show the result of Duncan's multiple range test (α=0.05/0.01).
Acknowledgements
We gratefully thank the NIPGR metabolome facility and proteomic facility (DBT grant BT/INF/22/SP28268/2018), confocal microscopy facility and central instrumentation facility. We are thankful to DBT-eLibrary Consortium (DeLCON) for providing access to e-resources.
Footnotes
Author contributions
Conceptualization: N.U.K., M.M.; Methodology: N.U.K.; Validation: M.M.; Formal analysis: N.U.K.; Investigation: N.U.K.; Resources: N.U.K.; Data curation: N.U.K.; Writing - original draft: N.U.K.; Writing - review & editing: M.M.; Visualization: N.U.K.; Supervision: M.M.; Project administration: M.M.; Funding acquisition: M.M.
Funding
This work was supported by a grant from the Department of Biotechnology, Ministry of Science and Technology, India (BT/HRD/NBA/39/05/2018-19), the Government of India and a core grant from the National Institute of Plant Genome Research. N.U.K. gratefully thanks the University Grants Commission, Government of India, and the National Institute of Plant Genome Research for research fellowships.
Peer review history
The peer review history is available online at https://journals.biologists.com/dev/article-lookup/doi/10.1242/dev.200600.
References
Competing interests
The authors declare no competing or financial interests.