N-glycosylation of SnRK2s controls NADPH maintenance in peroxisomes during prolonged ABA signalling

doi:10.21203/rs.3.rs-2714625/v1

Download PDF

Article

N-glycosylation of SnRK2s controls NADPH maintenance in peroxisomes during prolonged ABA signalling

https://doi.org/10.21203/rs.3.rs-2714625/v1

This work is licensed under a CC BY 4.0 License

Journal Publication

published 05 Aug, 2024

Read the published version in Nature Communications →

Version 1

posted

You are reading this latest preprint version

Unfavourable conditions threaten the survival and agricultural yield of plants and are often maintained in the long term, such as drought season and high salinity field. The phytohormone ABA plays key roles in the regulation of plant stress adaptation and is often sustained at high level for long periods of time^1-3. Much is known about ABA signal perception and activation/sensitization in the early signalling stage^4-6. However, the molecular mechanism underlying desensitization of ABA signalling largely unknown. Here, we revealed that, in the endoplasmic reticulum (ER)-Golgi network, SnRK2.2/2.3, the key regulators of ABA signalling, are unexpectedly N-glycosylated, which promotes their gradual redistribution from the nucleus to the peroxisomes in Arabidopsis roots, leading to a desensitization of the transcriptional response in the nucleus during prolonged ABA signalling. On the peroxisomal membrane, SnRK2s directly interact with glucose-6-phosphate (G6P)/phosphate translocator 1 (GPT1) to maintain NADPH homeostasis through increased activity of the peroxisomal oxidative pentose phosphate pathway (OPPP). The resulting maintenance in NADPH is required for modulation of hydrogen peroxide (H₂O₂) accumulation, relieving ABA-induced root growth inhibition. The subcellular dynamics of SnRK2s achieved by N-glycosylation seems to reflect a switch to control ABA responses from transcriptional regulation in the nucleus to metabolic control in the peroxisomes to help plants adapt to long-term environmental stress.

Biological sciences/Plant sciences/Plant hormones

Biological sciences/Plant sciences/Plant stress responses/Abiotic

The phytohormone abscisic acid (ABA) controls stress tolerance and development in plants. After ABA perception by the soluble receptors pyrabactin resistance 1 (PYR1)/PYR1-like (PYL)/regulatory components of ABA receptors (RCAR)^7,8, sucrose nonfermenting-1 (Snf1)-related kinase 2 (SnRK2) members (mainly SnRK2.2, SnRK2.3 and SnRK2.6) are released from the protein phosphatase type 2C (PP2C)-SnRK2 inactivation complex^9,10. Subsequently, activated SnRK2s phosphorylate and activate downstream effectors, such as transcription factors in the nucleus¹¹, ion channel and NADPH oxidases (NOXs)/respiratory burst oxidase homologs (RBOHs) on the plasma membrane^12-14. The SnRK2 kinases function as the central hub in this well established early ABA signalling^15-17. A common feature of signal-response regulatory modules is that the continuous presence of stimulation often results in attenuated responsiveness to subsequent challenges in the long term. This ability to adapt is referred to as signalling desensitization^18,19. Pretreatment with salt stress or water-deficit conditions significantly dampens ABA signalling, a phenomenon that has been observed for almost 40 years^20,21. However, the underlying molecular mechanism of desensitization in long-term ABA signalling remains largely unknown.

Asparagine (Asn/N)-linked glycosylation (N-glycosylation) catalysed by a group of N-glycan processing enzymes (NPEs) is one of the most prominent and abundant co/posttranslational modifications for secretory and membrane proteins within the ER-Golgi network in organisms across all domains^22-24. When a nascent glycoprotein is present in the ER lumen, the oligosaccharyltransferase complex (OST) attaches a N-glycan from dolichol-pyrophosphate-linked lipid anchor to a specific N residue in the NX(S/T) motif (the third one may also be cystine occasionally). Subsequently, N-glucosidase I (GI) and GIIα/β heterodimer remove the outmost two glucose residues. GII-trimmed glycoprotein is captured by calnexin/calreticulin (CNX/CRT) for quality control. Successfully folded glycoprotein can be further processed to form a mature N-glycan structure by NPEs sequentially, such as mannose-removing enzymes α-mannosidases (MNSs), β1,2-N-acetylglucosaminyl-transferase I (GnTI), Golgi-α-mannosidase II (GMII), and fucosyltransferases (FUC) ^24,25. Based on genetic analyses, several NPEs ware reported to be required for plant tolerance to abiotic stresses. Defective alleles of GIIα, display sensitivity to high-temperature stress²⁶. The loss-of-function mutants of NPEs are hypersensitive to high salt stress, such as mutants of the catalytic subunit of OST, STT3A (staurosporine and temperature sensitivity 3 A isoform)²⁷, Complex glycan 1(CGL1/GnTI)^28,29, GMII^28,30, MNS1 and MNS2³¹, β1,2-xylosyltransferase (XYLT) ²⁸, and α1,3-Fucosyltransferase 11 and 12 (FUT11/FUT12)^28,32. Here, we unexpectedly found that the nonsecretory proteins SnRK2.2/2.3 are modified by NPEs with a N-glycan addition at its C-terminal domain II, which positively regulates desensitization of ABA signalling in Arabidopsis.

NPEs promote desensitization of the transcriptional ABA response

Through screening of Arabidopsis mutants in response to exogenous ABA, an allele embedding a T-DNA insertion in β-subunit of glucosidase II (gIIβ-2, Extended Data Fig.1a) exhibited hypersensitivity to ABA-mediated repression of primary root growth and ABA-induced detached leaf senescence (Fig. 1a, b and Extended Data Fig. 2a, b). Consistently, two more loss-of-function alleles, gIIβ-3 and gIIβ-4, and their F1 hybrids all showed the hypersensitivity to ABA treatment (Extended Data Fig. 2c, d). A genomic DNA fragment harbouring GIIβ and its native promoter fully rescued the defective morphology of gIIβ-2, suggesting that the mutations in the GIIβ are responsible for the enhanced ABA sensitivity (Fig. 1a, b).

As GIIβ is involved in the N-glycosylation pathway, a complex process employing sequential actions of many enzymes that encompasses the biosynthesis, transfer and modification of sugar moieties in ER and Golgi apparatus^24,33. The enhanced sensitivity to ABA of gIIβ alleles prompted us to check the effects of other NPEs in response to ABA. Several previously reported high salt-sensitive mutants were tested. Loss-of-function mutant of STT3A, stt3a-2 exhibited hypersensitivity to ABA treatment (Fig. 1c, d). In contrast, STT3A overexpression (STT3A-OE) obviously reduced sensitivity to ABA treatment compared with that of the WT. Moreover, mutants of GIIα and three NPEs, CGL1, GMII, and FUT11, residing in the Golgi apparatus, namely, rsw3, cgl1-T, hgl1-2 and fut11²⁴, all exhibited hypersensitivity to ABA (Fig. 1e, f). The results of the morphological analyses suggest that the NPEs involved in the N-glycosylation pathway negatively regulate the ABA responses.

GIIβ was induced by prolonged ABA treatment, and the GIIβ protein also accumulated in an ABA dose-dependent manner (Extended Data Fig. 1c, f, g). The temporal expression pattern of the STT3A gene was similar to that of GIIβ (Extended Data Fig. 1d). Tissue expression analysis showed that GIIβ was significantly expressed in the roots (Extended Data Fig. 1e). To further determine the biological roles of NPEs in the ABA signalling pathway, we investigated the transcription of ABA-responsive marker genes in stt3a-2 roots, as the transcriptional response is the most important and conserved ABA response present from semiterrestrial algae to angiosperms³⁴. NCED3, P5CS1, and RAB18 all displayed a biphasic expression pattern (rapid activation followed by a gradual decline from 4 h to 72 h) in the WT roots. Intriguingly, in the stt3a-2, the biphasic inflection point was delayed from 3 h to 4 h for P5CS1 and to 6 h for NCED3 and RAB18. In contrast, this biphasic inflection point shifted forward to 2 h for P5CS1 and NCED3 in the STT3A-OE line, suggesting that NPEs positively regulate desensitization of the ABA transcriptional response during prolonged ABA signalling (Fig. 1g-i).

To investigate the genome-wide effect of NPEs on ABA transcription dynamics, we performed a time course RNA sequencing (RNA-seq) analysis with Col-0/WT and stt3a-2 harvested at 0 h to 72 h after application of ABA (see the Methods). The transcriptomics datasets of ABA-treated Col-0 were first subjected to k-means clustering analysis to determine the temporal expression patterns with 0 h as control³⁵. The results showed that 8 out of 20 clusters containing 12,337 genes exhibit ABA-induced biphasic expression patterns over time. Among them, 5,271 ABA-upregulated genes composing 4 clusters (clusters 1, 8, 13, and 18) displayed rapid induction in response to short-term ABA application, followed by a gradual attenuation during prolonged ABA treatment (Fig. 1j, k and Extended Data Fig. 3a). The expression patterns of NCED3, P5CS1, and RAB18 from the transcriptome were similar to those from RT-PCR (Extended Data Fig. 3b). The expression profiles from stt3a-2 were further retrieved to compare with the WT datasets, and the results showed that the biphasic switch to desensitization of 1,632 out of 5,271 genes was delayed in stt3a-2 (Fig. 1j, k and Extended Data Fig. 3a). For example, ABA-responsive genes whose associated Gene Ontology (GO) terms were most highly enriched in responses to vesicle-mediated transport in cluster 1 ware rapidly induced and peaked at 6 h in the WT, and the desensitization inflection point of 689 out of 1,288 genes shifted from 6 h to 24 h in stt3a-2 (Fig. 1j, k). A total of 7066 genes in the remaining 4 clusters (clusters 2, 6, 7, and 14) were rapidly downregulated in response to short-term ABA treatment (1-3 h), followed by gradual increase in the WT background. 840 out of 1,499 genes in cluster 6, showed a delayed biphasic inflection point from 1 h in WT to 2 h in stt3a-2 after ABA application (Extended Data Fig. 3c). With respect to the control, a biphasic expression pattern of ABA-responsive genes was not found in the no-ABA treatment datasets (Col-0, ck; Fig. 1k, Extended Data Fig. 3a, c). Moreover, our RNA-seq datasets significantly overlap with the ABA-induced transcriptomes previously reported (Extended Data Fig. 4)³⁶. In summary, the clustering analysis results reveal that the NPE STT3A promotes genome-wide desensitization of the ABA transcriptional response during prolonged ABA signalling.

SnRK2.2 and SnRK2.3 are N-glycosylated

To further decipher the molecular mechanisms of NPEs acting on ABA transcriptional desensitization, we examined the pair-wise interaction between NPEs and key components of the ABA-signalling pathway. Yeast two-hybrid (Y2H) assays showed that GIIβ interacts with SnRK2.2/2.3/2.6 but not with 14 PYLs, 2 PP2Cs (ABA insensitive 1 (ABI1) and ABI2), 2 downstream transcription factors (ABI5 and ABA responsive element binding factor 2 (ABF2)), and SnRK2.10, an ABA-unresponsive SnRK2 kinase³⁷ (Fig. 2a). Coimmunoprecipitation (Co-IP) and Pull-down assays further validated that SnRK2.2/2.3/2.6 can directly interact with GIIβ (Fig. 2b and Extended Data Fig. 5a). Firefly luciferase complementation imaging (LCI) assays showed that GIIβ, STT3A, GIIα, CGL1 and FUT11 all exclusively bind to SnRK2s, and the binding activities ware strongly enhanced under prolonged ABA treatment (Fig. 2c and Extended Data Fig. 5b-f). In general, N-glycosylation is thought to be processed in the ER and Golgi apparatus²⁴. Bimolecular fluorescence complementation (BiFC) assays showed that the interactions between SnRK2.2/2.3/2.6-YFP^C and GIIβ-YFP^N took place in the ER (Extended Data Fig. 5g). Collectively, our findings suggest that NPEs directly interact with SnRK2.2/2.3/2.6 in the ER-Golgi network.

The interactions between NPEs and SnRK2s suggest the occurrence of two possibilities: NPEs are phosphorylated by SnRK2 kinases, and/or SnRK2s are modified by N-glycosylation. Phos-tag-based assays showed that GIIβ cannot be phosphorylated by GST-SnRK2.2 in vitro (Extended Data Fig. 6a). To verify the second possibility, we transiently co-expressed SnRK2.3 and STT3A in tobacco leaves, as STT3A could induce a larger molecular weight difference of SnRK2.2/2.3 than other NPEs. Western blots revealed that STT3A coexpression produced a larger band shift of approximately 2-3 kDa (about one N-glycan³⁸, Fig. 2d). Subsequently, PNGase F and Endo H digestions confirmed that the band shift was due to N-glycan modification (Fig. 2e)³⁹. Like SnRK2.3, SnRK2.2 can also be N-glycosylated, but SnRK2.6 cannot in our tests (Fig. 2f, g). As the apparent molecular masses of SnRK2.6 and N-glycosylated SnRK2.2/2.3 are very similar to each other, we verified the N-glycosylation of endogenous SnRK2.2/2.3 in the snrk2.6 background. Upon ABA treatment, glycosylated SnRK2.2/2.3 gradually increased in abundance over time (12 h to 2 d, Fig. 2h). However, only a portion of the N-glycan was removed by PNGase F and Endo H digestions, indicating that SnRK2.2/2.3 N-glycans vary in structure in Arabidopsis in vivo (Fig. 2i). Taken together, these findings suggest that SnRK2.2/2.3 are modified with N-glycan in a prolonged ABA-dependent manner.

As N-glycosylation plays crucial roles in many biological processes, including protein folding, protein stability, and protein targeting²⁴, we checked whether the subcellular localization of SnRK2.2/2.3 was affected by N-glycosylation. SnRK2.2/2.3-GFP was mainly located in the nucleus to activate transcriptional responses, as previously reported. However, when coexpressed with GIIβ, SnRK2.2/2.3-GFP unexpectedly underwent nuclear export (or suppressed nuclear import) in tobacco leaves (Extended Data Fig. 6b and e)^9,34. Quantitative analysis of the SnRK2.2-GFP fluorescence signal showed that the localization change of SnRK2.2-GFP was tightly associated with GIIβ coexpression (Extended Data Fig.6c and d). Similarly, co-expression of STT3A-mCherry also led to SnRK2.2/2.3-GFP subcellular change and their co-localization in the ER, suggesting that N-glycosylation of SnRK2.2/2.3 induces its subcellar change in transiently expressed Nicotiana benthamiana leaves (Fig. 2j and Extended Data Fig.6f-i).

Based on these observations, we hypothesized that the N-glycosylation site mutation would disrupt the subcellular change of SnRK2.2/2.3. A total of 11 N-site mutations of SnRK2.3 were coexpressed with STT3A-mCherry in N. benthamiana leaves, and only the SnRK2.3^N323A still was retained in the nucleus (without change) upon STT3A-mCherry coexpression (Fig. 2j and Extended Data Fig. 6g-i). Western blots showed that the SnRK2.3^N323A mutation led to abolition of the N-glycosylation of SnRK2.3, suggesting that the N³²³ was the bona fide N-glycosylation site (Fig. 2k). The identified SnRK2.3^N323 site resides in a noncanonical N-glycosylation motif NRC⁴⁰, which is also present in SnRK2.2 but not in SnRK2.6. This motif resides in the C-terminal domain II of SnRK2s, an extended conformation protruding out of the three-dimensional structure required for ABA activation (Extended Data Fig. 7a)⁴¹. Moreover, the identified N-glycosylation site of SnRK2.2/2.3 is highly conserved across mammal AMP-activated protein kinases (AMPKs) and yeast Snf1 (Extended Data Fig. 7b).

SnRK2s are epistatic to NPEs in ABA signalling.

We generated the quadruple mutant gIIβ-2/snrk2.2/2.3/2.6, and the morphological analysis revealed that loss of SnRK2.2/2.3/2.6 fully suppressed the sensitivity phenotype of gIIβ-2 to ABA in terms of both primary root length and ABA-induced leaf senescence (Extended Data Fig. 8a-d). Moreover, stt3a-2, rsw3, cgl1-T, hgl1-2, and fut11 were introduced into the snrk2.2/2.3 double mutant background to generate triple mutants, respectively. As expected, snrk2.2/2.3 fully repressed the ABA-hypersensitive phenotype of stt3a-2, rsw3, cgl1-T, hgl1-2, and fut11 respectively (Fig. 2l, m and Extended Data Fig. 8e, f). These genetic assays suggest that SnRK2s are epistatic to NPEs in the ABA signalling pathway.

To investigate the biological role of the N-glycosylation of SnRK2 kinases, we generated transgenic lines harbouring SnRK2.3^N323A or SnRK2.3^WT in snrk2.2/2.3 background, respectively. Upon ABA application, SnRK2.3^WT and SnRK2.3^N323A all recovered the hyposensitivity of the snrk2.2/2.3 double mutant (Extended Data Fig. 8g, h). Moreover, the eight SnRK2.3^N323A lines were more sensitive to ABA than the SnRK2.3^WT lines, resembling the morphological differences observed between the NPE mutants and the WT (Fig. 1a-f, Extended Data Fig. 8g, h). Taken together, these findings suggest that the N-glycosylation of SnRK2.2/2.3 desensitizes the ABA responses.

SnRK2.2/2.3 redistribute to the peroxisome membrane

According to the subcellular change of SnRK2s in N. benthamiana leaves, we investigated its dynamics in Arabidopsis. Both SnRK2.2-GFP and SnRK2.3-GFP reporter lines complemented the ABA hyposensitivity of the snrk2.2/2.3 double mutant, and exhibited protein accumulation patterns similar to that of endogenous SnRK2s, representing rapid induction in response to short-term ABA and subsequent degradation within 6-72 h after ABA treatment (Extended Data Fig. 9a-e)⁴². Microscopy observation showed that strong nuclear localization of SnRK2.2/2.3-GFP (little diffusion in the cytosol) was detected at 2 h of ABA treatment in root tip cortical cells⁹. However, accompanied by a decrease in total fluorescence in the cells over time, a gradual translocation from the nucleus to vesicle/organelle-like structures in the cytoplasm occurred under prolonged ABA treatment for both SnRK2.2-GFP and SnRK2.3-GFP (Fig. 3a, b). With the aid of red fluorescent protein (RFP)-tagged organelle marker genes and organelle staining (see the Method), SnRK2.2/2.3-GFP were found to translocate onto the peroxisomal membrane under prolonged ABA signalling in Arabidopsis roots (Fig. 3c and Extended Data Fig.10a-e). However, this process was not blocked by nuclear export inhibitor leptomycin B (LMB), suggesting that the translocation of SnRK2s might consist of two steps: nuclear SnRK2 decrease and peroxisomal membrane redistribution (Extended Data Fig.10f).

For comparative observations, the SnRK2.2-GFP reporter was introduced into WT, stt3a-2 and STT3A-OE backgrounds, respectively. Intriguingly, SnRK2.2-GFP’s redistribution was delayed in stt3a-2. Stronger fluorescent signal was present in the nucleus and less in the peroxisomes in stt3a-2 compared with the WT at the same time points after ABA treatment, such as at 6 h and 48 h (Fig. 3d, e). In contrast, the SnRK2.2-GFP fluorescent signal was detected more rapidly/earlier in the peroxisomes, while weaker nuclear fluorescent signal was observed in the STT3A-OE background. For example, at 3 h, obvious peroxisomal signal was induced in STT3A-OE, whereas it was present at the 6 h after ABA treatment in the WT (Fig. 3d, e). Moreover, stronger nuclear fluorescence and less peroxisomal localization were detected for SnRK2.3^N323A-GFP compared with SnRK2.3^WT-GFP, such as at 48 h and 72 h after ABA treatment (Fig. 3f, g and h). On the other hand, the expression patterns of RAB18 and NCED3 showed that the biphasic switch of the ABA transcriptional response from activation to desensitization was also delayed in SnRK2.3^N323A plants (Fig. 3i, j). The findings suggest that desensitization of the ABA transcriptional response is orchestrated by SnRK2 redistribution from the nucleus to the peroxisomes.

Furthermore, we sought to determine whether the molecular mechanisms regulated by the N-glycosylation of SnRK2s are required for plant adaptation to salt stress. First, gIIβ mutants exhibited enhanced sensitivity to 125 mM NaCl, as did mutants deficient in STT3A, GIIα/RSW3, CGL1 and FUT11 (Extended Data Fig. 9f-i)^24,27,31. In contrast, overexpression of GIIβ and STT3A relieved the growth inhibition induced by high salt (Extended Data Fig. 9j, k). SnRK2.3^N323A plant also resulted in a hypersensitivity to high-salinity stress (Extended Data Fig. 9l, m). These morphological assays suggested that N-glycosylation of SnRK2 kinases plays positive biological roles in the response to high-salinity stress. Second, we investigated the subcellular dynamics of SnRK2s under high-salt and osmotic stresses. The results showed that SnRK2.2-GFP also was redistributed to the peroxisomes during prolonged salinity treatment, as well as during prolonged osmotic stress treatment (Extended Data Fig. 10g, h). These findings indicate that plants may exploit SnRK2 subcellular dynamics to better adapt to high-salt and osmotic stresses.

N-glycosylation of SnRK2s modulates H₂O₂ accumulation

The desensitization of the transcriptional response appears to be mainly related to the decrease of nuclear SnRK2s, prompting an interesting question concerning the biological role of peroxisome-redistributed SnRK2.2/2.3 achieved by N-glycosylation during prolonged ABA signalling. It is well known that peroxisomes play key roles in the regulation of reactive oxygen species (ROS) homeostasis⁴³, and H₂O₂ is essential for ABA signal transduction^13,44. To elucidate the biological role of peroxisomal SnRK2s, we investigated the H₂O₂ accumulations during prolonged ABA application in the WT and N-glycosylation-defective mutants. H₂O₂was induced in the WT roots after ABA application and peaked at 48 h, and then declined until 168 h (Fig. 4a). The biphasic pattern of H₂O₂ level was further confirmed with 3,3’-diaminobenzidine (DAB) staining and the H₂O₂ fluorescent indicators 3’-(p-hydroxyphenyl) fluorescein (HPF) and BES-H₂O₂-Ac (see the Methods) in both Col-0 and SnRK2.3^WT root tip (Extended Data Fig. 11a-f). Intriguingly, further investigations showed that the switch to the desensitization phase of ABA-induced H₂O₂ was delayed from 48 h to 96 h in stt3a-2 and SnRK2.3^N323A, each of which exhibited significantly higher H₂O₂ levels during 96-168 h after ABA treatment, suggesting that N-glycosylation of SnRK2s inhibits H₂O₂accumulation under prolonged ABA signalling in root tip (Fig. 4a, Extended Data Fig. 11a-f). As a control, the superoxide O₂^•- level did not show significant divergence at most time points of prolonged ABA signalling in different genotypes (stt3a-2 vs. Col-0; SnRK2.3^N323A vs. SnRK2.3^WT; Extended Data Fig. 12a-d).

H₂O₂ is required for ABA modulated stomatal closure^13,44, senescence of leaves⁴⁵ and primary root growth inhibition¹³. To verify the biological role of H₂O₂ scavenging mediated by N-glycosylation of SnRK2s, we carried out morphological analyses with H₂O₂ scavengers. Applications of the antioxidant reduced glutathione (GSH) significantly repressed the ABA-induced root growth inhibition and reduced root meristem cell numbers both in Col-0 and the SnRK2.3^WT line, supporting the notion that H₂O₂-mediated oxidative signalling plays key roles in ABA-induced plant growth inhibition⁴⁶. Moreover, the recovery of root growth was more pronounced in stt3a-2 and SnRK2.3^N323A plants after treatment with GSH or H₂O₂-scavenger KI (Extended Data Fig. 13a-f). For example, the primary root length of stt3a-2 is 62% that of Col-0 upon ABA treatment, whereas GSH application ultimately restored this percentage to 90% (Extended Data Fig. 13a, b). These findings suggest that glycosylated SnRK2s relieve ABA-induced growth inhibition by reducing H₂O₂ content during prolonged ABA signalling, which helps plants to better adapt to long-term unfavourable conditions.

The SnRK2-GPT1 module regulates NADPH maintenance

Plant peroxisomes mainly counter H₂O₂ overload with catalases (CATs) in the matrix and ascorbate peroxidases (APXs) involved in the ascorbate-glutathione cycle with reducing power from NADPH on the membrane⁴³. To identify the H₂O₂ scavenger regulated by N-glycosylation of SnRK2.2/2.3, we investigated the potential SnRK2-interacting partners among 13 published peroxisomal isozymes involved in H₂O₂-scavenging and NADPH generation (Extended Data Fig. 14a). The results of LCI assays showed that only glucose-6-phosphate (G6P)/phosphate translocator 1 (GPT1) exhibited reliable interaction with SnRK2.2/2.3 (APX3 binds to SnRK2.2/2.3 in an ABA-independent manner, and pull-down assay failed to show a direct interaction between them. Extended Data Fig. 14a, b). GPT1 was previously reported to be translocated to the peroxisomal membrane under stress conditions⁴⁷. Time-course Co-IP analyses showed that short-term ABA induced very weak interactions between GPT1 and SnRK2s before 6 h, however, stronger binding activity was observed at 24-72 h after ABA treatment (Fig. 4b). BiFC assays showed that the interactions between SnRK2.2/2.3 and GPT1 occur on the peroxisomal membrane (Fig. 4c, and Extended Data Fig. 14c). Moreover, GPT1 protein level was found to be induced by prolonged ABA treatment (Extended Data Fig. 14d). These findings suggest that SnRK2s bind to GPT1 on the peroxisomal membrane during prolonged ABA signalling.

As GPT1 is thought to be involved in peroxisomal oxidative pentose phosphate pathway (OPPP) for NADPH generation (Extended Data Fig. 14a)⁴⁷, we investigated the dynamics of NADPH accumulation. In the Col-0 background, ABA induces an increase of NADPH/NADP⁺ before 24 h which is followed by a decline between in the middle, then a re-accumulation from 48 h to 168 h, indicating two NADPH-producing phases induced by short-term and long-term ABA respectively (Fig. 4d). Intriguingly, compared with the WT (Col-0 and SnRK2.3^WT), the deficiency of N-glycosylation in stt3a-2 and SnRK2.3^N323A delayed the phasic switch of NADPH/NADP⁺ and apparently repressed the later accumulation (96 h later) of NADPH/NADP⁺, suggesting that N-glycosylated SnRK2s function in NADPH maintenance during prolonged ABA signalling when they are redistributed to the peroxisomes (Fig. 4d). As controls, the endogenous GSH/GSSG and NADH/ NAD⁺ levels did not significantly differ in stt3a-2 or SnRK2.3^N323A compared with the Col-0 and SnRK2.3^WT, respectively, although their accumulation was indeed affected by ABA application (Extended Data Fig. 15a, b). Furthermore, applications of OPPP inhibitors, polydatin and glucosamine-6-phosphate (GN6P), dramatically decreased the prolonged ABA-induced NADPH/NADP⁺ accumulation and negated the significant difference between N-glycosylation-deficient mutants and the WT (stt3a-2 vs. Col-0 and SnRK2.3^N323A vs. SnRK2.3^WT, Extended Data Fig. 15c). Correspondingly, the significant difference in H₂O₂ accumulation between N-glycosylation-deficient mutants and the WT no longer occurred upon inhibitor applications (Extended Data Fig. 15d-f). The findings suggest that N-glycosylated SnRK2.2/2.3 regulate NADPH maintenance relying on OPPP activity, sequentially leading to H₂O₂ modulation during prolonged ABA signalling.

To verify whether GPT1 regulates NADPH concentrations during prolonged ABA treatment, we attempted to measure the NADPH/NADP⁺ level in a mutant lacking GPT1. As the gpt1(-/-) homozygous mutant exhibits an embryo-lethal phenotype⁴⁸, artificial miRNAs of GPT1 (amiR-GPT1) was employed. The expression of amiR-EV (an artificial miRNA expression vector) or amiR-GPT2 led to significant decreases in NADPH/NADP⁺ levels and corresponding increases in H₂O₂ levels after 96 h ABA treatment in stt3a-2 and SnRK2.3^N323A compared with Col-0 and SnRK2.3^WT, respectively (Extended Data Fig. 16a-f). However, the significant differences in both NADPH/NADP⁺ and H₂O₂ levels between N-glycosylation-deficient mutants and the WT (stt3a-2 vs. Col-0, SnRK2.3^N323A vs. SnRK2.3^WT) were negated when amiR-GPT1s were expressed (amiR-GPT1-1 and amiR-GPT1-2 in Extended Data Fig. 16a-f, and i). Taken together, these findings suggest that the SnRK2-GPT1 module orchestrates NADPH homeostasis and H₂O₂ scavenging during prolonged ABA signalling on the peroxisomes.

Using GST-SnRK2.2/2.3 and GPT1 N-terminus (GPT1^N) purified from Escherichia coli, we carried out in vitro phosphorylation assays. The results showed that GPT1^N was phosphorylated by SnRK2.2/2.3 but not at a previously identified phosphorylation site, GPT1Ser27 (Extended Data Fig. 14g)⁴⁷. Subsequent LC‒MS/MS and site mutation analyses validated that the phosphorylation occurred on the Ser32 site of GPT1 (Fig. 4e, and Extended Data Fig. 14e, f, h). Furthermore, in the amiR-GPT1-1 background, GPT1^WT, GPT1^S32A, and GPT1^S32D with the anti-amiR-GPT1-1 sequence (in which the amiR-GPT1-1 binding site was replaced with synonymous codons, as shown in Fig. 4f) were expressed in Arabidopsis roots, respectively. During ABA treatments, the nonphosphorylation-mimicking mutant GPT1^S32A exhibited a significantly decreased NADPH/NADP⁺ ratio and increased H₂O₂ levels compared with those of GPT1^WT, mainly in the prolonged ABA signalling stage (Fig. 4g, h, Extended Data Fig. 16g, h). In contrast, expression of the phosphorylation-mimicking mutant GPT1^S32D gave rise to significantly increased NADPH accumulations and decreased H₂O₂ levels accordingly during prolonged ABA signalling in both amiR-EV and amiR-GPT1-1 backgrounds (Fig. 4g, h). Taken together, our findings suggest that phosphorylation of GPT1 at the Ser32 site catalysed by SnRK2s controls NADPH homeostasis on the peroxisomal membrane, which alleviates H₂O₂-mediated oxidative stress during prolonged ABA signalling.

Finally, to verify whether the NADPH maintenance regulated by N-glycosylation is due to the peroxisome targeting of SnRK2.2/2.3 during prolonged ABA signalling, we constructed two fixed SnRK2.3s specifically located on the peroxisomal membrane (SnRK2.3-mPTS^PEX26) and in the nucleus (SnRK2.3-NLS) respectively, and they possessed normal kinase activity but different protein expression patterns (Extended Data Fig. 17a-f). When expressed in snrk2.2/2.3 double mutant background, SnRK2.3-mPTS^PEX26 only exhibited one NADPH-accumulation phase induced by long-term ABA treatment, while SnRK2.3^WT displayed two NADPH-producing phases (Fig. 4i). Moreover, SnRK2.3-mPTS^PEX26 possessed significantly higher NADPH level than SnRK2.3^WT, which repressed H₂O₂ to very low level during prolonged ABA signalling (Fig. 4j). SnRK2.3-NLS also showed one NADPH-accumulation phase but induced by short-term ABA (before 24 h), lacking the NADPH maintenance phase although its NADPH level was not lower than SnRK2.3^WT during prolonged ABA signalling (before 168 h, Fig. 4i). The accumulation level of H₂O₂ was significantly higher in SnRK2.3-NLS than that in SnRK2.3^WT (Fig. 4j). Moreover, the transcriptional expression of ABA-induced genes RAB18 and NCED3 was enhanced by both 2 h and 48 h ABA treatments in SnRK2.3-NLS, whereas almost no any obvious expression of the two genes were tested in SnRK2.3-mPTS^PEX26 (Extended Data Fig.17g, h). Totally, the findings support the notions that SnRK2s activate ABA-transcriptional response in the nucleus, and are redistributed to the peroxisomes to maintain the NADPH accumulation for relieving H₂O₂ stress during prolonged ABA signalling.

Here, we revealed a novel molecular mechanism by which N-glycosylated SnRK2.2/2.3 are subcellularly redistributed the peroxisomes and control prolonged ABA-signalling for long-term stress acclimation in plants. As master regulators of ABA signalling, SnRK2.2/2.3/2.6 were previously found to locate in the nucleus to activate the transcriptional response, and SnRK2.6 also has been shown to interact with RBOHD/F to activate H₂O₂ signalling on the plasma membrane under short-term ABA treatment^11,12,14. According to our results, SnRK2.2/2.3 were unexpectedly found to be N-glycosylated in the ER-Golgi network during prolonged ABA signalling, which ultimately orchestrates the gradual redistribution of these proteins from the nucleus onto the peroxisomal membrane, leading to desensitization of the transcriptional response in the nucleus. Peroxisomal SnRK2.2/2.3 kinases subsequently directly interact with the G6P transporter GPT1 to enhance peroxisomal OPPP activity in controlling NADPH maintenance, which is critical for redox rebalancing under long-term stress conditions (Extended Data Fig. 18). Our findings suggest that N-glycosylation of SnRK2.2/2.3 regulates desensitization of the ABA response, which does not involve just deactivation of ABA signalling, but actually appears to reflect a transition state from transcriptional reprogramming gradually switching to metabolic regulation for NADPH maintenance in the peroxisomes.

N-glycosylation of SnRK2s is proposed here to be a basic principle promoting the cellular state transition during prolonged ABA signalling. The N-glycosylation is specific for secretory and membrane proteins processed in the ER-Golgi network^22,23. How SnRK2s are associated with ER-Golgi needs further investigations. The subcellular dynamics of SnRK2s function as a direct cellular effector to regulate prolonged ABA signalling. SnRK2s’ redistributions from the nucleus to the peroxisomes was mainly detected in heterotrophic root tissue of Arabidopsis. Nuclear export inhibitor LMB did not block this process (Extended Data Fig. 10f). indicating that it is not directly translocation from nucleus to peroxisomes. Combined with the results from CHX and MG132 treatments, this process may involve nuclear decrease and peroxisomal membrane re-sorting, both of which are regulated by N-glycosylation of SnRK2s (Fig.3a-h, Extend Data Fig. 10f).

NADPH accumulation underwent two distinct accumulation phases, indicating that different molecular mechanisms ware involved in NADPH generation during prolonged ABA signalling. Our observations suggested that SnRK2-GPT1 module mainly regulated the latter generation phase (Fig. 4d, g, i). In plants, the photosynthetic electron transfer chain is a major resource of NADPH in leaves, however at night, the OPPP located in chloroplast and cytosol provides NADPH, and also in heterotrophic tissues^49,50. Both phases of NADPH generation induced by ABA is thought to be a result of OPPP activity in Arabidopsis roots, which is supported by OPPP inhibitor assays (Extended Data Fig. 15c). Considering H₂O₂dynamics, the initial increase in NADPH abundance is proposed to be provided by cytosolic OPPP for ABA-induced H₂O₂ production via activity of RBOHD/F^13,50, whereas the latter NADPH accumulation regulated by SnRK2-GPT1 comes from peroxisomal OPPP pathway, which is involved in relieving H₂O₂-mediated growth inhibition during prolonged ABA signalling (Fig. 4a, d, g-j, and Extended Data Fig. 13a-f). The temporal expression pattern of GPT1 protein also supports this hypothesis, as the decrease in GPT1 before 24 h-ABA application indicates that more G6P is retained in the cytosol, whereas the subsequent GPT1 induction implies increased peroxisomal OPPP activity during prolonged ABA signalling (Extended Data Fig. 14d).

To summarize, peroxisomes are highly dynamic in terms of their size, shape, abundance and metabolic processing in response to environmental changes and are thus thought to be cellular sensors that respond to metabolic changes and abiotic stresses^43,51,52. as plant-specific homologues of yeast Snf1 and mammalian AMPK, which are key regulators of the limited energy status in response to stresses, SnRK2 kinases were found to directly orchestrate NADPH homeostasis in the peroxisomes during prolonged ABA signalling. Our findings would shed light on the molecular mechanisms involved in plant acclimation under long-term stress conditions.

Zhu, J.-K. Abiotic Stress Signaling and Responses in Plants. Cell 167, 313-324 (2016).
Gupta, A., Rico-Medina, A. & Cano-Delgado, A. I. The physiology of plant responses to drought. Science 368, 266-269, doi:10.1126/science.aaz7614 (2020).
Weng, J. K., Ye, M., Li, B. & Noel, J. P. Co-evolution of Hormone Metabolism and Signaling Networks Expands Plant Adaptive Plasticity. Cell 166, 881-893, doi:10.1016/j.cell.2016.06.027 (2016).
Raghavendra, A. S., Gonugunta, V. K., Christmann, A. & Grill, E. ABA perception and signalling. Trends Plant Sci 15, 395-401, doi:10.1016/j.tplants.2010.04.006 (2010).
Hauser, F., Li, Z., Waadt, R. & Schroeder, J. I. SnapShot: Abscisic Acid Signaling. Cell 171, 1708-1708 e1700, doi:10.1016/j.cell.2017.11.045 (2017).
Lin, Z. et al. Initiation and amplification of SnRK2 activation in abscisic acid signaling. Nat Commun 12, 2456, doi:10.1038/s41467-021-22812-x (2021).
Park, S. Y. et al. Abscisic acid inhibits type 2C protein phosphatases via the PYR/PYL family of START proteins. Science 324, 1068-1071, doi:10.1126/science.1173041 (2009).
Ma, Y. et al. Regulators of PP2C phosphatase activity function as abscisic acid sensors. Science 324, 1064-1068, doi:10.1126/science.1172408 (2009).
Fujita, Y. et al. Three SnRK2 protein kinases are the main positive regulators of abscisic acid signaling in response to water stress in Arabidopsis. Plant Cell Physiol 50, 2123-2132, doi:10.1093/pcp/pcp147 (2009).
Fujii, H. & Zhu, J. K. Arabidopsis mutant deficient in 3 abscisic acid-activated protein kinases reveals critical roles in growth, reproduction, and stress. Proc Natl Acad Sci U S A 106, 8380-8385, doi:10.1073/pnas.0903144106 (2009).
Yoshida, T. et al. AREB1, AREB2, and ABF3 are master transcription factors that cooperatively regulate ABRE-dependent ABA signaling involved in drought stress tolerance and require ABA for full activation. Plant J 61, 672-685, doi:10.1111/j.1365-313X.2009.04092.x (2010).
Geiger, D. et al. Activity of guard cell anion channel SLAC1 is controlled by drought-stress signaling kinase-phosphatase pair. Proc Natl Acad Sci U S A 106, 21425-21430, doi:10.1073/pnas.0912021106 (2009).
Kwak, J. M. et al. NADPH oxidase AtrbohD and AtrbohF genes function in ROS-dependent ABA signaling in Arabidopsis. EMBO J 22, 2623-2633, doi:10.1093/emboj/cdg277 (2003).
Sirichandra, C. et al. Phosphorylation of the Arabidopsis AtrbohF NADPH oxidase by OST1 protein kinase. FEBS Lett 583, 2982-2986, doi:10.1016/j.febslet.2009.08.033 (2009).
Miyazono, K. et al. Structural basis of abscisic acid signalling. Nature 462, 609-614, doi:10.1038/nature08583 (2009).
Fujii, H. et al. In vitro reconstitution of an abscisic acid signalling pathway. Nature 462, 660-664, doi:10.1038/nature08599 (2009).
Yang, G., Yu, Z., Gao, L. & Zheng, C. SnRK2s at the Crossroads of Growth and Stress Responses. Trends Plant Sci 24, 672-676, doi:10.1016/j.tplants.2019.05.010 (2019).
Sibley, D. R. & Lefkowitz, R. J. Molecular mechanisms of receptor desensitization using the beta-adrenergic receptor-coupled adenylate cyclase system as a model. Nature 317, 124-129, doi:10.1038/317124a0 (1985).
Lemmon, M. A., Freed, D. M., Schlessinger, J. & Kiyatkin, A. The Dark Side of Cell Signaling: Positive Roles for Negative Regulators. Cell 164, 1172-1184, doi:10.1016/j.cell.2016.02.047 (2016).
Larosa, P. C., Handa, A. K., Hasegawa, P. M. & Bressan, R. A. Abscisic Acid accelerates adaptation of cultured tobacco cells to salt. Plant Physiol 79, 138-142, doi:10.1104/pp.79.1.138 (1985).
Peng, Z.-Y. & Weyers, J. D. B. Stomatal sensitivity to abscisic acid following water deficit stress. Journal of Experimental Botany 45, 835-845, doi:10.1093/jxb/45.6.835 (1994).
Hebert, D. N., Lamriben, L., Powers, E. T. & Kelly, J. W. The intrinsic and extrinsic effects of N-linked glycans on glycoproteostasis. Nat Chem Biol 10, 902-910, doi:10.1038/nchembio.1651 (2014).
Kowarik, M. et al. N-linked glycosylation of folded proteins by the bacterial oligosaccharyltransferase. Science 314, 1148-1150, doi:10.1126/science.1134351 (2006).
Strasser, R. Plant protein glycosylation. Glycobiology 26, 926-939, doi:10.1093/glycob/cww023 (2016).
Nagashima, Y., von Schaewen, A. & Koiwa, H. Function of N-glycosylation in plants. Plant Sci 274, 70-79, doi:10.1016/j.plantsci.2018.05.007 (2018).
Burn, J. E. et al. The cellulose-deficient Arabidopsis mutant rsw3 is defective in a gene encoding a putative glucosidase II, an enzyme processing N-glycans during ER quality control. Plant J 32, 949-960, doi:10.1046/j.1365-313x.2002.01483.x (2002).
Koiwa, H. et al. The STT3a subunit isoform of the Arabidopsis oligosaccharyltransferase controls adaptive responses to salt/osmotic stress. Plant Cell 15, 2273-2284, doi:10.1105/tpc.013862 (2003).
Kang, J. S. et al. Salt tolerance of Arabidopsis thaliana requires maturation of N-glycosylated proteins in the Golgi apparatus. Proc Natl Acad Sci U S A 105, 5933-5938, doi:10.1073/pnas.0800237105 (2008).
Frank, J., Kaulfurst-Soboll, H., Rips, S., Koiwa, H. & von Schaewen, A. Comparative analyses of Arabidopsis complex glycan1 mutants and genetic interaction with staurosporin and temperature sensitive3a. Plant Physiol 148, 1354-1367, doi:10.1104/pp.108.127027 (2008).
Strasser, R. et al. Molecular cloning and characterization of Arabidopsis thaliana Golgi alpha-mannosidase II, a key enzyme in the formation of complex N-glycans in plants. Plant J 45, 789-803, doi:10.1111/j.1365-313X.2005.02648.x (2006).
Liu, C. et al. Trimming of N-Glycans by the Golgi-Localized alpha-1,2-Mannosidases, MNS1 and MNS2, Is Crucial for Maintaining RSW2 Protein Abundance during Salt Stress in Arabidopsis. Mol Plant 11, 678-690, doi:10.1016/j.molp.2018.01.006 (2018).
Strasser, R., Altmann, F., Mach, L., Glossl, J. & Steinkellner, H. Generation of Arabidopsis thaliana plants with complex N-glycans lacking beta1,2-linked xylose and core alpha1,3-linked fucose. FEBS Lett 561, 132-136, doi:10.1016/S0014-5793(04)00150-4 (2004).
Zielinska, D. F., Gnad, F., Schropp, K., Wisniewski, J. R. & Mann, M. Mapping N-glycosylation sites across seven evolutionarily distant species reveals a divergent substrate proteome despite a common core machinery. Mol Cell 46, 542-548, doi:10.1016/j.molcel.2012.04.031 (2012).
Shinozawa, A. et al. SnRK2 protein kinases represent an ancient system in plants for adaptation to a terrestrial environment. Commun Biol 2, 30, doi:10.1038/s42003-019-0281-1 (2019).
Ikeuchi, M. et al. Wounding Triggers Callus Formation via Dynamic Hormonal and Transcriptional Changes. Plant Physiol 175, 1158-1174, doi:10.1104/pp.17.01035 (2017).
Sun, H. X., Li, Y., Niu, Q. W. & Chua, N. H. Dehydration stress extends mRNA 3' untranslated regions with noncoding RNA functions in Arabidopsis. Genome Res 27, 1427-1436, doi:10.1101/gr.218669.116 (2017).
Soma, F. et al. ABA-unresponsive SnRK2 protein kinases regulate mRNA decay under osmotic stress in plants. Nat Plants 3, 16204, doi:10.1038/nplants.2016.204 (2017).
Shen, J., Ding, Y., Gao, C., Rojo, E. & Jiang, L. N-linked glycosylation of AtVSR1 is important for vacuolar protein sorting in Arabidopsis. Plant J 80, 977-992, doi:10.1111/tpj.12696 (2014).
Maley, F., Trimble, R. B., Tarentino, A. L. & Plummer, T. H., Jr. Characterization of glycoproteins and their associated oligosaccharides through the use of endoglycosidases. Anal Biochem 180, 195-204, doi:10.1016/0003-2697(89)90115-2 (1989).
Matsui, T. et al. N-glycosylation at noncanonical Asn-X-Cys sequences in plant cells. Glycobiology 21, 994-999, doi:10.1093/glycob/cwq198 (2011).
Yoshida, R. et al. The regulatory domain of SRK2E/OST1/SnRK2.6 interacts with ABI1 and integrates abscisic acid (ABA) and osmotic stress signals controlling stomatal closure in Arabidopsis. J Biol Chem 281, 5310-5318, doi:10.1074/jbc.M509820200 (2006).
Ali, A. et al. Rheostatic Control of ABA Signaling through HOS15-Mediated OST1 Degradation. Mol Plant 12, 1447-1462, doi:10.1016/j.molp.2019.08.005 (2019).
Pan, R., Liu, J., Wang, S. & Hu, J. Peroxisomes: versatile organelles with diverse roles in plants. New Phytol 225, 1410-1427, doi:10.1111/nph.16134 (2020).
Pei, Z. M. et al. Calcium channels activated by hydrogen peroxide mediate abscisic acid signalling in guard cells. Nature 406, 731-734, doi:10.1038/35021067 (2000).
Hung, K. T. & Kao, C. H. Hydrogen peroxide is necessary for abscisic acid-induced senescence of rice leaves. J Plant Physiol 161, 1347-1357, doi:10.1016/j.jplph.2004.05.011 (2004).
Yang, L. et al. ABA-mediated ROS in mitochondria regulate root meristem activity by controlling PLETHORA expression in Arabidopsis. PLoS Genet 10, e1004791, doi:10.1371/journal.pgen.1004791 (2014).
Baune, M. C. et al. The Arabidopsis Plastidial Glucose-6-Phosphate Transporter GPT1 is Dually Targeted to Peroxisomes via the Endoplasmic Reticulum. Plant Cell 32, 1703-1726, doi:10.1105/tpc.19.00959 (2020).
Niewiadomski, P. et al. The Arabidopsis plastidic glucose 6-phosphate/phosphate translocator GPT1 is essential for pollen maturation and embryo sac development. Plant Cell 17, 760-775, doi:10.1105/tpc.104.029124 (2005).
Kruger, N. J. & von Schaewen, A. The oxidative pentose phosphate pathway: structure and organisation. Curr Opin Plant Biol 6, 236-246, doi:10.1016/s1369-5266(03)00039-6 (2003).
Baxter, C. J. et al. The metabolic response of heterotrophic Arabidopsis cells to oxidative stress. Plant Physiol 143, 312-325, doi:10.1104/pp.106.090431 (2007).
Hu, J. et al. Plant peroxisomes: biogenesis and function. Plant Cell 24, 2279-2303, doi:10.1105/tpc.112.096586 (2012).
He, A., Dean, J. M. & Lodhi, I. J. Peroxisomes as cellular adaptors to metabolic and environmental stress. Trends Cell Biol 31, 656-670, doi:10.1016/j.tcb.2021.02.005 (2021).

There is NO Competing Interest.

Download PDF

Journal Publication

published 05 Aug, 2024

Read the published version in Nature Communications →

Version 1

posted

You are reading this latest preprint version

N-glycosylation of SnRK2s controls NADPH maintenance in peroxisomes during prolonged ABA signalling

Status:

Journal Publication

Version 1

Abstract

Figures

Full Text

Discussion

References

Additional Declarations

Supplementary Files

Status:

Journal Publication

Version 1