The ABA receptor PYL9 together with PYL8 plays an important role in regulating lateral root growth

Abscisic acid is a phytohormone regulating plant growth, development and stress responses. PYR1/PYL/RCAR proteins are ABA receptors that function by inhibiting PP2Cs to activate SnRK2s, resulting in phosphorylation of ABFs and other effectors of ABA response pathways. Exogenous ABA induces growth quiescence of lateral roots, which is prolonged by knockout of the ABA receptor PYL8. Among the 14 members of PYR1/PYL/RCAR protein family, PYL9 is a close relative of PYL8. Here we show that knockout of both PYL9 and PYL8 resulted in a longer ABA-induced quiescence on lateral root growth and a reduced sensitivity to ABA on primary root growth and lateral root formation compared to knockout of PYL8 alone. Induced overexpression of PYL9 promoted the lateral root elongation in the presence of ABA. The prolonged quiescent phase of the pyl8-1pyl9 double mutant was reversed by exogenous IAA. PYL9 may regulate auxin-responsive genes in vivo through direct interaction with MYB77 and MYB44. Thus, PYL9 and PYL8 are both responsible for recovery of lateral root from ABA inhibition via MYB transcription factors.

kinases) 11,13 . SnRK2 kinases are activated by activation loop autophosphorylation 14 and are the key nodes in ABA signaling pathway.
High concentrations of ABA inhibit both primary and lateral root growth 15,16 , while low concentrations of ABA are known to promote root growth 17 . Studies with ABA-deficient mutants indicate that ABA is also crucial for maintenance of root growth under water-stressed and normal growth conditions 18,19 . This is partially due to the promotion of stem cell maintenance, which is vital for root growth, by nanomolar concentrations of ABA in the root meristem 20 .
Previous studies found that ABA and auxin are integrated into each other's pathways in many developmental processes. ABA regulates auxin responses in many aspects. Loss of function of ABA-insensitive 3 leads to a reduction of auxin-induced lateral root initiation 21 . Overexpression of ABA-insensitive 4 (ABI4) impairs lateral root development by reducing the expression of the auxin-efflux transporter PINFORMED 1 (PIN1) 22 . Furthermore, Auxin Response Factor 2 (ARF2) is a transcriptional repressor involved in plant growth and directly regulates the homeodomain gene HB33. Altered auxin distribution in arf2-101 and two HB33 overexpressing lines in response to ABA treatment indicates that ABA and auxin might act synergistically in inhibiting root growth 23 . Furthermore, the pyl8 mutant shows a reduced sensitivity to ABA-mediated primary root growth inhibition 24 . Besides its role in primary root, PYL8 was found to interact with MYB DOMAIN PROTEIN 77 (MYB77) and functions in auxin-mediated lateral root growth 25 .
Auxin, in turn, also regulates ABA responses in many ways. During seed germination, the expression of ABI3 is controlled by auxin through the auxin response factors AUXIN RESPONSE FACTOR 10 and AUXIN RESPONSE FACTOR 16 26 . Some auxin responsive genes are also regulators of stress responses. Among them is MYB77 that acts as a coactivator with ARFs 27 . Auxin-responsive gene expression was greatly attenuated in myb77 knockout mutants and MYB77 overexpression lines mimicked the wild type treated with exogenous IAA 27 . In the subgroup 22 where MYB77 belongs to, there are three other members: AtMYB44/AtMYBR1, AtMYB70 and AtMYB73 28 . Previous research has suggested that MYB44 represses ABA signaling during drought and senescence by interacting with PYL8 and PYL9 29,30 . Thus, these MYB proteins are associated with both auxin and stress responses 31 .
The members of PYR1/PYL/RCAR family have distinct properties. For example, two orthologs AtPYL13 and OsPYL12 inhibit several clade A PP2Cs in an ABA-independent manner, and most importantly, OsPYL12 is unable to bind to ABA [32][33][34] . PYR1 and PYL1-2 are dimers in solution, while PYL4-10 are monomers 35 . The diversity of PYLs may contribute to the versatility of ABA signaling.
Here, we report a function of PYL9 in regulating lateral root growth upon ABA treatment, in a manner similar to that of PYL8. Previous studies found that PYL9/RCAR1-mediated ABA signaling pathway could be modulated by ROP11 during in vitro reconstitution of ABA signaling pathway in Arabidopsis protoplasts 36 and PYL9 could interact with an R2R3 MYB transcription factor, MYB44 29 . We found that PYL9 acts together with PYL8 in regulating the recovery of lateral root growth from inhibition after ABA treatment. Our results suggest that both PYL8 and PYL9 are nodes of crosstalk between ABA and auxin, and their signal is transduced by a group of MYB transcription factors to regulate the lateral roots.

Results
The pyl8-1pyl9 double mutant shows a reduced inhibition of primary root growth and lateral root formation to exogenous ABA compared to pyl8. ABA is a key factor in regulating root architecture under stress 16,20 . To dissect its role, we analyzed Arabidopsis mutants impaired in ABA signaling pathway, especially null mutants in ABA receptors. Previous studies demonstrated that PYL8 plays an important role in ABA signaling in roots and the work was focused on primary roots 24 . We found that in addition to regulating the growth of primary roots, PYL8 could also promote lateral root recovery from ABA inhibition 25 . Since PYL9 has a very high sequence identity with PYL8 and both of them have a relative high expression level in roots 37 , we obtained a T-DNA knock-out mutant of pyl9 (SALK_083621) (Fig. 1a) 24 and tested the root growth of pyl9 under ABA treatment. pyl9 mutant did not have an obvious lateral root growth phenotype on ABA medium (see Supplementary Fig. S1). We also quantified the primary root length and the lateral root number as well as average lateral root length both on the control medium and on ABA medium, still found no significant difference between pyl9 and wild type (WT) (see Supplementary Fig. S1). PYL8 and PYL9 are in the same clade of the PYL phylogenetic tree 11 . We therefore analyzed the root growth of the pyl8-1pyl9 double mutant 38 (see Supplementary Fig. S2). Consistent with its visible phenotypes (Fig. 1b), the pyl8-1pyl9 double mutant showed a faster primary root growth than the py8-1 single mutant after treatment with ABA. On the 9 th day post transfer (dpt) to plates with ABA, the pyl8-1pyl9 double mutant showed a longer primary root than pyl8-1 single mutant at all concentrations of ABA tested (Fig. 1c). We also quantified the lateral root number of both mutants (Fig. 1d). On the 7 dpt to different concentrations of ABA, pyl8-1pyl9 had more lateral roots than pyl8-1 on 1, 5 and 10 μ M ABA medium, and both of them differ significantly from WT at all concentrations except 1 μ M ABA medium. Thus pyl8-1pyl9 double mutant was less sensitive to ABA than pyl8-1 single mutant in primary root growth and lateral root formation.
Lateral root growth stays in a longer quiescence in pyl8-1pyl9 double mutant. Besides primary root growth phenotypes, we have also observed that the pyl8-1pyl9 double mutant had shorter lateral roots (Fig. 1b). Consistent with its visible phenotypes, the pyl8-1pyl9 double mutant showed a slower lateral root growth than the pyl8-1 single mutant after treatment with ABA. Lateral root growth was severely suppressed in both mutants exposed to ABA especially in the double mutant. Average lateral root length and total lateral root length were both decreased (Fig. 2a). The pyl8 single mutant was reported to have a longer quiescent phase on ABA plates before recovery 25 . Our results here suggest that PYL9 may promote growth recovery of lateral root under ABA treatment together with PYL8. Lateral root with a length shorter than 0.5 mm is considered to be quiescent 39 . We found that the two mutants and the wild type had similar quiescent days on control medium. However both mutants have a prolonged quiescent phase than Col-0 under ABA treatment. Moreover, the pyl8-1pyl9 double mutant had an even longer quiescent phase than the py8-1 single mutant (Fig. 2b). This suggests that PYL9 functions together with PYL8 in promoting growth recovery of lateral roots under ABA treatment.

IAA suppresses ABA-induced growth inhibition of lateral roots in pyl8-1pyl9. Exogenous IAA
can release growth inhibition of lateral roots under ABA treatment 25 . To determine whether this happens in the pyl8-1pyl9 double mutant, we analyzed the lateral root growth of pyl8-1pyl9 seedlings grown on ABA-containing medium supplemented with IAA (Fig. 3a). We found that the ABA-dependent lateral root growth suppression of pyl8-1pyl9 was partially overcome by 10 nM IAA, and was fully rescued by 100 nM IAA (Fig. 3b). We also determined lateral root growth over the course of time. We found that both pyl8-1 and pyl8-1pyl9 mutants had a slower Asterisks indicate comparison between pyl8-1 and pyl8-1pyl9. *P < 0.05, **P < 0.01, ***P < 0.001, Student's t test. P-values were adjusted for multiple comparisons by "Benjamini & Hochberg" method. Primary root length and lateral root number of pyl8-1 (at 5 μ M, 10 μ M, 15 μ M, 20 μ M) and pyl8-1pyl9 (at all ABA concentrations) were significantly larger than Col-0, P < 0.05. lateral root elongation than wild type under ABA treatment but not under control conditions. Exogenous application of 10 nM IAA rescued the ABA-dependent growth defects of lateral root in pyl8-1 but not in pyl8-1pyl9. In contrast, in the presence of 100 nM IAA, the lateral root growth defects of both mutants were fully rescued (Fig. 3c). This means that IAA could also rescue the ABA-dependent inhibition of lateral root elongation caused by loss of PYL8 and PYL9. However, the pyl8-1pyl9 double mutant requires a higher concentration of IAA than pyl8-1 single mutant. These results suggest that the lateral root growth defect of pyl8-1pyl9 mutant may be caused by auxin deficiency.

ABA-induced overexpression of PYL9 alters both primary and lateral root growth.
Since the function of PYL9 on root growth was relatively weaker than that of PYL8, we analyzed the root architecture of PYL9 overexpression lines. The PYL9 gene, under the control of RD29A promoter, was transformed into Arabidopsis. This RD29A gene was first discovered as a cold-induced gene. Later it was found to be responsive to not only cold but also dehydration and ABA 40 . Two homozygous pRD29A:PYL9 transgenic lines with high PYL9 expression were used for further analysis 38 . Both transgenic lines had a dramatic increase in PYL9 expression compared to Col-0 wild type under ABA treatment (see Supplementary Fig. S3).  We transferred 4-day-old seedlings to both control medium and medium with 1 μ M ABA (Fig. 4a). We monitored the lateral root length over time (Fig. 4b). The two transgenic lines did not have any obvious difference from wild type on control medium since the PYL9 gene was not induced. When 1 μ M ABA was applied, the transgenic lines produced longer lateral root than the wild type after 3 days and the difference was more significant over time (Fig. 4b). As loss of PYL9 caused a prolonged quiescent phase, we wondered whether this increase in average lateral root length was due to a shortened quiescent phase. Under 1 μ M ABA treatment, more than half of the lateral roots had a quiescent phase shorter than 3 days in both transgenic lines (Fig. 4c). Thus, overexpression of PYL9 promotes lateral roots to escape from ABA-dependent inhibition.
To test whether overexpression of ABA receptor PYL9 conferred hypersensitivity to ABA in primary root growth, we analyzed root growth of pRD29A:PYL9 transgenic lines with treatment of different concentrations of ABA. After transferring to medium supplemented with ABA and growing for a week, the primary root of transgenic lines showed an obvious reduction (Fig. 5a). After ABA treatment, both lines showed a significant reduction of lateral root number compared to the wild type (Fig. 5b). This suggests that PYL9 functions in the promotion of lateral root escaping from quiescence as well as the inhibition of primary root growth and lateral root formation under ABA treatment.

PYL9 interacts with MYB transcription factors. Recently, interactions between PYLs and MYBs
have been reported 25,29,30 . In our study, several different clones in each combination were tested on yeast growth medium deprived of Trp, His and Leu (Fig. 6a). All of these clones with MYB44/PYL8 had visible colonies, suggesting that PYL8 strongly interacts with MYB44. However, only one of these clones with MYB44/PYL9 had visible colonies (Fig. 6a), suggesting that PYL9 may weakly interact with MYB44. Other MYB proteins were also tested using multiple independent colonies. PYL6 was used as a negative control and had no interactions with any of the three tested MYBs (Fig. 6b, bottom panel). PYL8, as a positive control, had interactions with MYB77 and MYB73 (Fig. 6b, upper panel). Similar to clones with MYB44/PYL9, the minority of these independent clones with MYB77/PYL9 had visible colonies. However, we did not detect any interaction between MYB73 and PYL9 in the yeast two-hybrid assay (Fig. 6b,  We used pGADT7-MYBs and pBD-GAL4-PYLs. Our results and the previous findings suggest that PYL8 strongly interacts with MYBs, while PYL9 has a relatively weak interaction with MYBs. To further analyze the interactions between PYL9 and MYBs in plant cells, we used the firefly luciferase (LUC) complementation assay in Arabidopsis protoplasts. PYL9 was fused to the N-terminal domain of firefly luciferase (LUC) and MYB proteins were fused to the C-terminal domain of LUC. We co-transformed PYL9-nLUC with the MYB-cLUC into Arabidopsis protoplasts (Fig. 6c). Coexpression of PYL9-nLUC with MYB44/73/77-cLUC, but not MYB61-cLUC produced measurable luciferase activity. These results suggest that PYL9 also interacts with some MYBs in vivo.
PYL9 enhances the activity of MYBs in the presence of ABA and IAA. We asked whether the interaction with PYL9 might affect the regulation of downstream genes by MYBs. Previous studies showed that MYB77 could recognize cis-elements MBSI (CNGTTR) and MBSII (GTTAGTTA) and preferentially binds to the MBSI motif 25 . IAA19 is one of these downstream genes regulated by MYB77 27 . pIAA19:LUC was used as a reporter in the Arabidopsis protoplasts transient expression assay 25 . We co-expressed PYL9, MYB77 with pIAA19:LUC in the protoplasts (Fig. 7a). As expected, the luciferase signal was induced by IAA and was enhanced by MYB77. Similar to PYL8, PYL9 enhanced the activity of MYB77 to increase the expression of pIAA19:LUC in the presence of IAA (Fig. 7a).
Previous studies demonstrated that PYL8 enhances the interaction between MYB77 and MBSI motifs in vitro 25 . To further understand whether PYL9 and MYB77 regulate their target genes as a protein complex, we used a transgenic line with tagged PYL9 driven by PYL9 native promoter 38 . pPYL9:PYL9-YFP-HA was introduced into the quadruple mutant pyr1pyl1pyl2pyl4 and this tagged line was used in chromatin immunoprecipitation assay (see Supplementary Fig. S4). We pre-incubated the seedlings with 5 μ M ABA and 1 μ M IAA for 5 hours, and the assay was performed as described 41 . We added extra EGS or ethylene glycol-bis (succinic acid N-hydroxysuccinimide ester) to strengthen the PYL9-MYBs interaction during cross-linking. Because MYB77 preferentially binds to MBSI, we filtered candidate genes using this criterion. We checked several DNA fragments that harbored this cis-element, including the promoter region of IAA1, IAA7, IAA17, IAA19 and HAT2. Different primer pairs were designed flanking this motif. The ChIP-PCR result showed that the promoter region of IAA7 was enriched with PYL9-YFP-HA after treatment of IAA and ABA (Fig. 7b). The ChIP-quantitative-PCR result showed nearly 10-fold enrichment in pPYL9:PYL9-YFP-HA transgenic lines than in the wild type negative control after the treatment of ABA and IAA. While in the control, there was no such enrichment (Fig. 7b,c). The motif in this promoter fragment was CTGTTG, belonging to MBSI. We also found another CTGTTG in the promoter of HAT2 but this one did not show enrichment in ChIP-quantitative-PCR. This might be due to the position of the motif. The one in IAA7 is located around 500 bp upstream from the start codon, however, the one in HAT2 is located within 100 bp from its start codon. This suggests that the transcription complex of PYL9 and MYB proteins may bind to promoters around 500 bp from where transcription starts. This in vivo evidence strongly supports the hypothesis that PYL9 regulates the transcriptional activity of MYB77 or other MYB proteins directly in vivo and participates in the auxin signal pathway. Discussion ABA and auxin are two important phytohormones; one is well known for its function under stress, and the other one is well known for its growth-inducing activity. Our study here reveals a crosstalk node between them in regulating root growth.
Roots provide a tight connection between plants and the soil environment, and are often the first to sense the soil environment. Root systems of ABA-deficient mutants like aba2-1 and aba3-1 are less affected than wild type seedlings upon osmotic stress 42 . ABA at low concentrations has been reported to promote root growth 17 . The endogenous ABA controls root architecture both in the presence and absence of osmotic stress through maintaining meristem in dormancy and inhibiting QC division and suppressing the differentiation of stem cells in the primary root 20 . High concentrations of ABA are well known to inhibit the growth of both primary and lateral root (Fig. 1c,d). ABA has a much stronger effect on lateral root than on the primary root, suggesting that different signaling mechanisms in the two types of roots 39 . In our study, the lateral root number in all the tested ABA concentrations showed a decrease compared to control medium (Fig. 1d,b). What's more, the lateral root elongation is more sensitive to ABA than the inhibition of seed germination and this reversible growth arrest occurs at a specific developmental stage, that is right after the lateral root emergence with a length less than 0.5 mm 16 . As shown in Fig. 2b, longer quiescence was observed on ABA-containing medium and this might be due to a dormant state of lateral roots. Although ABA is well known for its function in seed dormancy and growth inhibition, different ABA sensitivities suggest different pathways in lateral roots compared with other tissues.
Salt stress induces a quiescence phase in post emergence lateral root growth and then recovery takes place several days later, which involves genes of ABA biosynthesis, signaling and transcription regulation 39 . PYL8 is responsible for this quiescence upon ABA treatment 25 . PYL9 has a 77% amino acid sequence identity with PYL8 and particularly high GUS activity can be detected in stele cells of pPYL9:GUS transgenic plants 24 . Although pyl9 single mutant did not show obvious difference from wild type on ABA medium (see Supplementary Fig. S1),  pyl8-1pyl9 double mutant had an even longer primary root and more lateral roots than pyl8-1 on ABA medium (Fig. 1c,d). The induced overexpression of PYL9 upon ABA treatment led to a shorter primary root and fewer lateral roots (Fig. 5a,b). These results suggest that PYL9 functions in the repression of lateral root formation and primary root elongation by ABA. PYL9 is a functional ABA receptor 11 and the ABA signal perceived by it and other PYLs is passed down eventually to ABA-RESPONSIVE ELEMENT (ABRE)-BINDING FACTOR (ABF) proteins. And overexpression of ABF2 confers hypersensitivity to ABA in primary root 43 . Other mutants impaired in ABA signaling pathway have also been tested including abi1-1, pyr1/pyl1/2/4 and snrk2.2/3/6 25 . All these ABA-resistant mutants showed a reduced repression of primary root growth and lateral root formation upon ABA treatment 25 and this correlates with our results in pyl8-1pyl9 double mutant. However, these ABA-resistant mutants have shorter quiescence compared to WT on ABA medium, which is opposite to our pyl8-1pyl9 double mutant. This suggests a different pathway involving PYL9. Recently, interactions between PYLs and MYBs have been reported 25,29,30 . PYL8 strongly interacts with the MYBs 25,30 . Further studies found that PYL9 and probably PYL7 also interact with these MYBs 29 , which is different from the previous published result 30 . Our results also showed that PYL9 interacted with MYB77 and MYB44 (Fig. 6), however in our Y2H assay there was no obvious interaction between PYL9 and MYB73 (Fig. 6b), which is different from previous reports 29 . This is probably due to different yeast systems and plasmids. MYB77 is a positive regulator of lateral root growth and its interaction with ARF7 suggests that MYB77 may function through auxin signal transduction pathway 27 . The lateral root growth of myb77 mutants is more sensitive upon ABA treatment and exogenous IAA could reverse this like it could reverse in pyl8-1 and pyl8-1pyl9 25 . So the interaction between PYL9 and MYB77 suggests they may function together. MYB77 preferentially recognizes cis-elements MBSI (CNGTTR) and previous EMSA data showed that PYL8 protein enhances MYB77 binding to MBSI in an ABA-dependent manner 25,27 . MYB44 has a clear preference for MBSII (GTTAGTTA) type but still binds to MBSI 44 . Our chromatin immunoprecipitation assay indicated that anti-HA antibody pulls down the protein complex of PYL9-MYB proteins (Fig. 7b,c). The exogenous IAA and ABA enhanced their interaction and the effect on the downstream gene promoters 25 (Fig. 7). Quantitative PCR showed that one fragment of IAA7 promoter, containing MBSI motif, has a huge enrichment (Fig. 7b,c). IAA7 is a member of the Aux/IAA protein family, which is induced by auxin 45 . Besides PYL8 and PYL9, PYL7 is also an interacting protein of MYB44 in Y2H assays 29 . This implies that more PYLs might be involved in the PYL-MYB pathway.
Auxin and ABA are widely accepted as regulators of root growth. These two phytohormones have a lot of interactions and previous studies have revealed that auxin could potentiate ABA response in roots 46 . Unlike in p35S:VP1 where ABA fully inhibits auxin induced lateral root initiation 47 , the prolonged quiescence phase of pyl8-1 on ABA medium was shortened by application of IAA 25 . Osmotic stress represses lateral root development which could be overcome by auxin 42 , however exogenous ABA represses lateral root formation, that is the initiation of lateral root primordia, which could not be rescued by auxin 16 . In our study, the quiescence in pyl8-1 and pyl8-1pyl9 was overcome by application of IAA (Fig. 3). What's more, the severe defects in the double mutant require more IAA to be rescued (Fig. 3b,c). Whether auxin rescues or not might be dependent on the state of lateral roots and these activities might be mediated by different mechanisms. Lateral root primordia initiation is highly dependent on auxin 48 , however in our study, we only considered the lateral roots that are already formed when we transferred seedlings to ABA medium.
The function of PYL9 and MYB77 complex is dependent on IAA treatment. In Arabidopsis protoplasts, the reporter (pIAA19:LUC) showed a strong signal when PYL9 was co-transformed with MYB77 in the presence of IAA (Fig. 7a). In transgenic seedlings, PYL9 can pull down auxin responsive promoter fragments in the presence of IAA and ABA (Fig. 7b,c). Besides, PYL9 itself is a functional ABA receptor. Taken together, this suggests that PYL9 connects ABA and auxin signaling pathway. MYB77 modulates auxin response by forming a heterodimer with ARF7 and their double mutant myb77-2arf7 (myb77-2nph4-1) has even smaller lateral root density than nph4-1, which is already smaller than wild type and myb77-2 27 . In myb77 mutants, auxin-responsive genes are attenuated while they are increased in over-expression lines 27 . Similar results were obtained in p35S:MYB44 transgenic lines where auxin-responsive genes are increased, suggesting that salt response and auxin signaling cross-talk at the transcriptional level 31 . This suggests that MYB77 and MYB44 might have some overlapping functions in regulating auxin-responsive genes. IAA7, which might be the target of PYL9-MYB77 complex, is also a key component regulating ABA and auxin-dependent post-embryonic growth 49 . It is suppressed when ABA represses embryonic axis elongation by potentiating auxin signaling in the elongation zone 49 .
PYL9 binds type 2C protein phosphatases and was found to interact with R2R3 MYB proteins recently. Our results show that PYL9 not just binds to MYB transcription factors, but also regulates their transcriptional activity, and that this is a point of convergence between ABA and auxin pathways.
Seeds were surface-sterilized for 8 min in 10% bleach and then rinsed in sterile deionized water for four times. Sterilized seeds were grown on 0.6% Phytagel (Sigma) or 0.8% agar media containing 1/2 MS nutrients (catalog no. M524, PhytoTechnology Laboratories), 1% sucrose adjusted to pH 5.7 (control media), and kept at 4 °C for 2 days. Seedlings were grown vertically before transfer to control media or media supplemented with the indicated concentrations of ABA (Sigma, A1049) or IAA (Sigma, I2886). Seedlings were grown at 22 °C under 16-h light/8-h dark cycles.
For protoplast analysis, seedlings were grown on Jiffy 7 peat soil (42 mm Pellets) in a Percival chamber with a relatively short photoperiod (12 hours of light at 23 °C, 12 hours of dark at 20 °C) under low light (about 100 μ E m −2 s −1 ) and 50 to 70% relative humidity under well-watered conditions. Phenotype analysis. The primary root length was measured from the first day after transfer. The results of lateral root were observed using HIROX KH-7700 digital microscope with MX-5040RZ lens. The total and average lateral root lengths of individual plants were quantified by summing or averaging the lengths of all the lateral roots on each plant. Lateral roots shorter than 0.5 mm were categorized as quiescent 39 .
Scientific RepoRts | 6:27177 | DOI: 10.1038/srep27177 Plasmid construction. The pIAA19-LUC and pGADT7-MYBs constructs were generated as described 25 . The pBD-GAL4-PYLs constructs were generated as described 11 . Yeast two-hybrid assay. Yeast two-hybrid assays were performed as described 25 . PYLs fused to the GAL4 DNA binding domain were used as baits. MYB44, MYB73 and MYB77 fused to the GAL4-activating domain were used as preys. Interaction was determined by growth assay on media lacking His or His and Ade with or without 5 μ M ABA. Dilutions (10 −1 , 10 −2 and 10 −3 ) of saturated cultures were spotted onto the plates and photographed after 5 days.
Transient expression assay in Arabidopsis. Assays for transient expression in protoplasts were performed as described 25 . All steps were at room temperature. pIAA19::LUC was used as the auxin-responsive reporter. ZmUBQ-GUS was used as the internal control. After transforming, protoplasts were incubated in washing and incubation solution (0.5 M mannitol, 20 mM KCl, 4 mM MES, pH 5.7) with or without ABA and IAA at the indicated concentrations for 12 hours of light and LUC luminescence was measured with a plate reader (Wallac VICTOR2 plate reader). LUC complementation assays were performed as described 25 .
Chromatin immuoprecipitation assay. Ten-day-old transgenic seedlings and anti-HA antibodies (1:100 for ChIP assay, HA-Tag, 26D11, Mouse mAb, M20003, Abmart, Shanghai, China) were used for ChIP experiments 50 . Briefly, the transgenic seedlings were ground into a fine powder with liquid nitrogen and resuspended in nuclei isolation buffer. 1.5 mM EGS was added into the nuclei isolation buffer for protein-protein cross-linking. The nucleus were then collected by centrifugation and resuspended with nuclei lysis buffer. The resuspended chromatin was sonicated to fragments with various sizes (250 bp-1 kb) subsequently. PYL9-HA-YFP-MYBs protein complex was precipitated from input DNA with or without anti-HA antibodies. Protein A agarose beads (Millipore, USA) were added into the incubation mixture for additional 2 h at 4 °C. The immune complexes were eluted from the washed protein A beads. The DNA was purified with phenol/chloroform (1:1, v/v) and precipitated. The purified DNA and input DNA were used as templates. The enrichment of DNA fragments was determined by quantitative PCR with specific primers.