Arabidopsis thaliana wild-type ecotype Columbia (Col-0) and dcp5-1 (SALK_008881), xrn4-5 (SAIL_681_E01) and double lsm1a lsm1b (SALK_106536, SAIL_756_305) homozygous mutant lines were used in this study. Seeds were surface sterilized with 30% bleach/0.02% Triton-X100 solution and grown on Murashige and Skoog (MS) medium supplemented with 1% (w/v) sucrose and 0.3% phytagel, under a 16 h light/8 h dark (long-day) photoperiod, and 22°C/20°C. For sterile hydroponic culture about 100 seeds grown for 14 d in 300-mL Erlenmeyer flasks containing 100 mL of one-half Murashige and Skoog medium supplemented with 100 mg/L myo-inositol, 500 mg/L MES, 10 g/L sucrose, pH 5.7. Seedlings were treated with different ABA concentrations as indicated, harvested, quickly frozen in liquid nitrogen, and stored at −80°C.

For germination tests, 40 seeds were planted in MS medium containing various concentrations of ABA (0–0.5 μM, Sigma) in 4 replicas. The average number of seeds germinated each day was calculated and used to define the Pieper's index (average time required for germination of a single seed) (Jakubowski, 2015): Pieper′s index=(x1* s1+x2*s2+…+xn*sn)/(s1+s2+…+sn)

x - number of a day from seeds dissemination

For sensitivity to ABA at the early stages of vegetative growth 5-day-old seedlings were transferred from MS medium to MS medium with different concentration of ABA (0–10 μM). Root length was measured relative to control conditions 4 d after the transfer, for more than 30 roots for each data point. The ABA sensitivity results were subjected to a two-way analysis of ANOVA variance followed by t-tests using Microsoft Excel.

Total RNA was isolated from 2-week-old seedlings using Trizol reagent (Sigma) according to the manufacturer's instructions. For Northern blot analysis 15 μg of total RNA was separated in 1.1% agarose/6% formaldehyde gels and transferred to a Hybond N⁺ membrane by capillary elution. Northern blots were performed using random primed probes amplified from cDNA template with appropriate primers and radioactively labeled with DECAprimeTM II kit (Ambion) and [α-³²P]ATP (Hartmann Analytics), or oligoribonucleotide probes radioactively labeled with PNK (Thermo Scientific) and [γ-³²P]ATP (Hartmann Analytics). Membranes were hybridized overnight with radioactive probes in PerfectHyb buffer (Sigma), washed, analyzed with PhosphorImager Typhoon FLA 9000 (GE Healthcare) and quantified with ImageJ software (Molecular Dynamics). mRNA half-life measurement experiments were carried out as described (Souret et al., 2004). Two-week-old seedlings were transferred to flasks containing a buffer (1mM PIPES, pH 6.25, 1 mM sodium citrate, 1 mM KCl, 15 mM sucrose), and after a 30-min incubation, cordycepin (150 mg/L) was added. Total RNA samples at indicated time points were extracted using Trizol reagent and analyzed by Northern blot. For real-time RT-PCR analysis samples of total RNA (50 μg) were DNase-digested with TURBO DNA-free kit (Ambion), according to manufacturer's protocol. RNA quality was checked on Nanodrop 1000. RT of 5 μg of RNA was performed in 20-μl reaction using SuperScript III reverse transcriptase (LifeTech) and mix of random hexamers (Invitrogen) and oligo(dT) according to manufacturer's protocol. cDNA samples were diluted 9 times and used as a template in qPCR using the SYBR Green I Master (Roche) and the LightCycler 480 (Roche). The AT1G13320 gene, the expression of which does not alter after abiotic stress, was used for normalization (Czechowski et al., 2005). All presented data are derived from three biological replicas, each of which represents an average of three technical replicas. The results were subjected to a two-way analysis of ANOVA variance followed by t-tests. Oligonucleotides used for Northern hybridization and PCR reactions are listed in Supplementary Table 1.

Frozen seedlings were ground in liquid nitrogen with mortar and pestle and sonicated three times for 20 s in the extraction buffer (20 mM Tris, pH 7.5, 2 mM EDTA, 2 mM EGTA 50 mM β-glycerophosphate, 100 μM Na₃VO₄, 2 mM dithiothreitol [DTT], Complete protease inhibitor cocktail Roche) using approximately 0.5 mL of the extraction buffer for each 1 μg of plant material. After sonication, the extracts were centrifuged at 18,000 rcf for 30 min at 4°C, and the supernatants were used for further studies. In-gel kinase activity assays were performed using a method described previously (Wawer et al., 2010). Briefly, protein samples were separated in 12% SDS/PAGE gels with 0.5 mg/ml histone embedded in the separating gel as a kinase substrate. After electrophoresis, SDS was removed by washing in washing buffer (25 mM Tris/HCl, pH 7.5, 5 mM sodium fluoride, 0.5 mg/ml BSA, 0.1% Triton X-100, 0.5 mM DTT and 0.1 mM sodium orthovanadate) three times each for 30 min at room temperature. Proteins were renaturated overnight in renaturing buffer (25 mM Tris/HCl, pH 7.5, 5 mM sodium fluoride, 0.1% Triton X-100, 1 mM DTT and 0.1 mM sodium orthovanadate) at 4°C with three changes of buffer. The gel was incubated for 1.5 h at room temperature in 10 ml of reaction buffer (10 mM Tris/HCl, pH 7.5, 2 mM DTT, 0.1 mM EGTA, 15 mM MgCl2 and 20 μM ATP, supplemented with 50 μCi of [γ-³²P]ATP). Unincorporated [γ-32P]ATP was removed by extensive washing in 5% trichloroacetic acid with 1% sodium phosphate. The gels were stained with Coomassie Brilliant Blue R250, dried and exposed to autoradiography.

Western blot analysis was performed according to a standard procedure (Wawer et al., 2010) using polyclonal antibody anti-PYR1 (AS132634, 1:10,000) and anti-SnRK2.2/3/6 (AS142783, 1:2,000) from Agrisera. Anti-rabbit (Sigma Aldrich) horseradish peroxidase-conjugated antisera were used as secondary antibodies.

Chlorophyll (Chl) extraction and quantification were performed using 1 cm² leaf discs cut from 3-week-old plants, as described (Lichtenthaler, 1987). Briefly, Chl was extracted with 100% acetone and quantified spectrophotometrically at 662 and 645 nm. Chl content (mg per cm²) was calculated as sum of Chlorophyll a and b: Chlorophyll a:Chla=12.25 A662 - 2.79 A645 (μ g per ml solution)Chlorophyll b:Chlb=21.50 A645 - 5.10 A662 (μ g per ml solution)

To evaluate whether the 5′-3′ mRNA decay pathway is involved in ABA signaling we analyzed the impact of the depletion of major factors in decapping and exonucleolytic degradation, LSM1, DCP5 and XRN4, on typical ABA responses. First we investigated ABA sensitivity of the double lsm1a lsm1b, dcp5-1, and xrn4-5 loss-of-function mutants (Souret et al., 2004; Xu and Chua, 2009; Golisz et al., 2013) during germination by determining average seed germination time (Pieper's index). As shown previously, under normal growth condition, lsm1a lsm1b and dcp5-1 seeds germinated slightly later than the wild-type (the ratio of mutant vs. Col-0 Pieper's index > 1; Figure 1A) (Xu and Chua, 2009; Perea-Resa et al., 2012). For all concentrations of ABA, germination time of dcp5-1 seedlings was significantly longer (higher Pieper's index) that the dcp5-1 mutant is more susceptible to ABA. In turn, differences in germination time of lsm1a lsm1b plants were similar in the absence and in the presence of ABA and germination of xrn4-5 was not altered at any conditions. Next, to assess ABA sensitivity of the mutants during early stages of vegetative growth, we measured primary root length in the presence of ABA. To this end 5-day-old seedlings grown in hormone-free medium were transferred to vertical plates with various concentrations of ABA (5 and 10 μM). The roots of lsm1a lsm1b and xrn4-5 plants were clearly shorter than in the case of Col-0, whereas no effect was visible for the dcp5-1 line (Figures 1B–G). These observations strongly suggest that these two mutants are hypersensitive to ABA during later stages of development. Together, the data indicates that DCP5, LSM1 and XRN4 are involved in response to ABA.

ABA receptors, RCAR/PYR/PYL proteins constitute a 14-member family and all of them, except PYL13, are able to activate ABA signaling (Fujii et al., 2009). PYL5, has been found among several proteins important for growth and development that are encoded by unstable transcripts (Gutierrez et al., 2002; Narsai et al., 2007). To check whether decapping regulates the level of PYL transcripts in response to ABA we analyzed mRNAs of PYL5 and PYR1 in 2-week-old Col-0 and decapping mutants lsm1a lsm1b and dcp5-1 before and following ABA (50 μM) treatment. As expected, both transcripts are regulated by ABA and their level gradually decreases with time after the treatment (Figure 2). In lsm1a lsm1b and dcp5-1 mutant lines PYR1 level was slightly increased only in control conditions, while PYL5 was visibly up-regulated before and during ABA treatment in lsm1a lsm1b and to a lesser extent in dcp5-1. To check whether the impact of LSM1 on ABA receptor mRNAs is direct we tested the stability of PYL5 and PYR1 mRNAs in the lsm1a lsm1b mutant following transcriptional inhibition by cordycepin. Northern blot analysis showed that half-life of PYR1, but not of PYL5, was markedly increased in the absence of LSM1 (Figure 3A1), suggesting that PYR1 mRNA is a direct substrate of the LSM1-7 complex. This observation was confirmed for another mutant in the decapping complex, dcp5-1, where PYR1 mRNA was clearly, and PYL5 only marginally, stabilized (Figure 3A2). Surprisingly, half-lives of these transcripts were not altered in xrn4-5 plants. These results indicate that at least one of PYL mRNAs, PYR1, is a direct substrate of DCP- and LSM-mediated decapping, but its turnover is independent of the cytoplasmic 5'-3' exoribonuclease XRN4. This is consistent with a finding that in Arabidopsis this nuclease is involved in the decay of only a subset of cellular mRNAs (Souret et al., 2004; Rymarquis et al., 2011; Golisz et al., 2013).

DCP5 has been reported to repress translation of OLEO1 and OLEO2 mRNAs encoding seed storage proteins (Xu and Chua, 2009). Since PYR1 mRNA is a direct substrate of decapping complexes, we assumed that it might be translationally repressed, even if its level is not affected in plants lacking DCP5 and LSM1. To test this hypothesis we checked the level of PYR1 protein in the mutants using specific anti-PYR1 antibodies. We found that PYR1 strongly accumulated in dcp5-1 and to a lower extent in lsm1a lsm1b, but not in xrn4-5 plants (Figure 3B). These observation correlate well with extended half-life of PYR1 mRNA in these mutants (Figures 3A–C).

The elevated level of PYR1 receptor in lsm1a lsm1b and dcp5-1 mutants may increase ABA-dependent activation of SnRK2.2/3/6 (Fujii et al., 2009). We therefore measured the activity of SnRK2 kinases in lsm1a lsm1b, dcp5-1 and xrn4-5 mutants in vivo before and after exposure to ABA (50 μM) by an in-gel kinase assay using Arabidopsis crude protein extracts and histone H3 as a substrate (Yoshida et al., 2002). Extracts prepared from Col-0, snrk2.2/3 and ost1/snrk2.6 seedlings were used as controls. Consistent with previous reports, ABA treatment caused rapid activation of SnRK2.2/3/6 protein kinases (Figure 4A). Activity of SnRK2.6 and SnRK2.2/3 in lsm1a lsm1b mutant, but not in dcp5-1 or xrn4-5, was stronger than in Col-0 at early time points (15 and 30 min, depending on the experiment, see (Supplementary Figure 1), reaching later the same level as in control plants.

It appears that SnRK2 activity corresponds to changes in PYR1/PYL5 receptors only in lsm1a lsm1b plants, therefore we conclude that SnRK2 kinases are not generally regulated by the decapping/5′-3′ mRNA decay pathway. As increased activity of SnRK2.2/3/6 in lsm1a lsm1b might be caused by higher amount of kinases we tested their mRNA and protein level in the three mutants under study. Real-time quantitative PCR (RT-qPCR) analysis of SnRK2.2/3/6 mRNAs revealed no differences in the case of SnRK2.2 and SnRK2.3 between control and the mutants (Figure 4B), but as reported previously SnRK2.6 mRNA was strongly induced after ABA treatment in Col-0 plants (Chan, 2012), and only moderately in all mutants. In turn, using western blotting with anti- SnRK2.2/3/6 antibodies we observed a more intense signal corresponding to SnRK2.2/2.3 and SnRK2.6 kinases in lsm1a lsm1b, but not in dcp5-1 and xrn4-5 (Figure 4C). This result suggests that increased activity of ABA-dependent SnRK2s in the lsm1a lsm1b mutant is due to higher protein level of these kinases. Considering that LSM1-7 and DCP5 may be involved in translation repression we also tested the stability of SnRK2 mRNAs in lsm1a lsm1b and dcp5-1 mutants following transcriptional inhibition by cordycepin. Since SnRK2.2 and SnRK2.3 have particularly long half-lives (>120 min, data not shown, Narsai et al., 2007) we were able to determine only the half-life of SnRK2.6. This transcript was strongly stabilized in the lsm1a lsm1b mutant and to a lesser extent in the dcp5-1 line (Figure 4D). Higher stability of SnRK2.6 in lsm1a lsm1b plants correlates well with the increased level of the protein and suggests that at least SnRK2.6 mRNA is affected by decapping complex and its translation is repressed by the LSM1-7 complex.

To gain insight into the functional relationship between decapping/5′-3′ mRNA decay and the PYL/PYR/RCAR-PP2C-SnRK2 signaling pathway we examined the induction of ABA- and SnRK2-inducible transcripts, including RD29B, RAB18, RD20, LTP (At4g33550), RAP2.6, ABI1, and HAB1 (Fujita et al., 2009). As reported previously RAB18, LPT, RD20, ABI1, HAB1 were up-regulated in lsm1a lsm1b plants in control conditions (Golisz et al., 2013; Supplementary Figures 2, 3), while only LPT accumulated in dcp5-1 and HAB1 in xrn4-5 (Supplementary Figure 2). In contrast, RAB18 and RD20 mRNAs were significantly decreased in plants lacking XRN4. After 2 h exposure to ABA all examined transcripts were induced in Col-0 plants as expected, but this effect was much reduced in ABA-treated lsm1a lsm1b plants for almost all tested mRNAs, except RAP2.6 (Figure 5A). Similarly, the expression of RD29B, LTP, RAB18, and HAB1 after exposure to ABA was lower in dcp5-1 compared to Col-0, while only induction of RD29B and LTP was decreased in xrn4-5 plants.

Since higher ABA-induced activation of SnRK2 in lsm1a lsm1b seedlings occurs within 15–30 min and then reaches the same level as in Col-0, we checked whether the expression pattern of ABA-inducible genes exhibits any fluctuations in the mutant. Northern blot analysis of chosen transcripts at different time points after ABA treatment showed that all tested mRNAs are evenly induced, alike in control and mutant plants (Supplementary Figure 3). These results suggest that decapping-mediated accumulation of common ABA-responsive genes is independent of the activation of PYL receptors and SnRK2 kinases.

Recently it has been reported that SnRK2 kinases together with ABF2, ABF3, and ABF4 transcription factors may act as key regulators in mediating ABA-triggered chlorophyll (Chl) degradation and leaf senescence in Arabidopsis (Gao et al., 2016). To check whether the decapping complex contributes to the regulation of this pathway during subsequent vegetative growth stages we measured Chl in 3-week old plants after ABA treatment. Consistent with previous data chlorophyll level in plants treated with ABA was markedly reduced, but we did not observe differences in chlorophyll degradation between control and mutant plants (Supplementary Figure 4).

The PYL/PYR/RCAR-PP2C-SnRK2 pathway regulates gene expression through the phosphorylation of ABF transcription factors that cooperate with DREB factors (reviewed in Joshi et al., 2016). We therefore analyzed by RT-qPCR the mRNA levels of eight ABA-responsive transcription factors (four ABFs and four DREBs) involved in PYL/PYR/RCAR-PP2C-SnRK2 signaling during vegetative growth stage (Yoshida et al., 2015). Their level was either moderately altered or unaffected in lsm1a lsm1b, dcp5-1, and xrn4-5 plants in control conditions without ABA treatment (Supplementary Figure 2) with little similar tendencies between the mutants, except that mRNAs of all four tested DREBs were down-regulated in dcp5-1 and ABF1 was significantly decreased both in lsm1a lsm1b and dcp5-1. Interestingly, this analysis revealed induction of all tested TF mRNAs in response to ABA and showed that activation of DREB1A, DREB1B and DREB2A was markedly reduced in lsm1a lsm1b and dcp5-1 lines, whereas induction of ABF2 was lower only in lsm1a lsm1b plants (Figure 5B). These findings suggest that DCP5 and the LSM1-7 complex modulate, probably indirectly, the expression of TFs involved in the PYL/PYR/RCAR-PP2C-SnRK2 pathway.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.