IAA3-mediated repression of PIF proteins coordinates light and auxin signaling in Arabidopsis

The exogenous light signal and endogenous auxin are two critical factors that antagonistically regulate hypocotyl growth. However, the regulatory mechanisms integrating light and auxin signaling pathways need further investigation. In this study, we identified a direct link between the light and auxin signaling pathways mediated by the auxin transcriptional repressor IAA3 and light-controlled PIF transcription factors in Arabidopsis. The gain-of-function mutation in IAA3 caused hyposensitivity to light, whereas disruption of IAA3 led to an elongated hypocotyl under different light intensity conditions, indicating that IAA3 is required in light regulated hypocotyl growth. Genetic studies showed that the function of IAA3 in hypocotyl elongation is dependent on PIFs. Our data further demonstrated that IAA3 interacts with PIFs in vitro and in vivo, and it attenuates the DNA binding activities of PIFs to the target genes. Moreover, IAA3 negatively regulates the expression of PIFs-dependent genes. Collectively, our study reveals an interplay mechanism of light and auxin on the regulation of hypocotyl growth, coordinated by the IAA3 and PIFs transcriptional regulatory module.

Introduction Sessile plants adjust their growth and development in response to the changing environment through a process in which environmental cues and plant intrinsic pathways are tightly integrated [1]. As a vital external signal for plant growth, light provides not only energy but also an informational signal that critically impacts plant development from seed germination to flowering [2][3][4]. Growing in the soil under natural conditions, Arabidopsis seedlings undergo skotomorphogenesis and produce elongated hypocotyls as one of the characteristic features of this process. When the seedlings emerge from the soil, light rapidly stimulates photomorphogenic development, and hypocotyl elongation is consequently inhibited [2,5].
Light signals are sensed by different types of photoreceptors, including phytochromes, cryptochromes, phototropins, and UV Resistance Locus 8 (UVR8) [2], which perceive different wavelengths and intensities of incoming light [6]. These photoreceptors convey the signal to downstream PHYTOCHROME-INTERACTING FACTORs (PIFs, including PIF1, PIF3, PIF4, and PIF5). PIFs are an important subfamily of basic helix-loop-helix (bHLH) transcription factors and act as negative regulators of photomorphogenesis [7]. Among them, PIF3 is the first identified member, and it directly interacts with phyB [8]. Several PIF3 homologs, including PIF1, PIF4, and PIF5, are also found to regulate photomorphogenesis [9]. The quadruple PIF mutant, pif1 pif3 pif4 pif5 (pifq), exhibits a constitutively photomorphogenic phenotype under dark conditions, indicating that these PIFs function redundantly to repress photomorphogenesis [10]. PIF proteins accumulate in the dark to maintain skotomorphogenesis. After light irradiation, they are phosphorylated and degraded, leading to photomorphogenetic seedling growth [11]. Several studies have shown how PIF proteins are regulated at the post-transcriptional level. HECATE (HEC) proteins heterodimerize with PIF1 and affect the transcriptional activation activity of PIF1 by preventing DNA binding [12]. Cryptochromes directly interact with PIF4 and PIF5 and modulate their activities in limiting blue light responses [13]. PIF proteins are also involved in diverse internal and external signaling pathways, such as the gibberellin (GA), brassinosteroid (BR), and high temperature pathways [14][15][16][17][18].
Auxin is a core endogenous phytohormone that regulates plant growth and development. In contrast to the light signal, auxin promotes hypocotyl elongation [5,19]. Several auxin biosynthesis deficient or excessive mutants, such as sav3/taa1 and yucca, display defects in hypocotyl elongation [20,21]. It is well established that auxin triggers interaction between the TRANSPORT INHIBITOR RESPONSE 1/AUXIN SIGNALING F-BOX PROTEIN (TIR1/ AFB) co-receptor and the AUXIN/INDOLE-3-ACETIC ACID (AUX/IAA) transcriptional repressors, resulting in the degradation of AUX/IAA proteins through the 26S proteasomemediated pathway. As a result, the inhibition of downstream auxin response factor (ARF) transcription factors are released, and a subset of auxin-responsive genes are activated [22]. There are 29 AUX/IAA proteins in Arabidopsis. The typical AUX/IAA proteins contain four conserved domains, I, II, III, and IV [23]. Domain I functions as a repression domain and recruits the TOPLESS (TPL) corepressor for its function. Domain II is required to interact with TIR1/ AFB and thus determines the stability of AUX/IAA proteins. Domains III and IV mediate the binding of AUX/IAAs and ARFs [24]. Several mutations in AUX/IAA genes have been identified through genetic analysis in Arabidopsis [25,26]. Among them, the semidominant shy2 (short hypocotyl 2)/iaa3 mutations cause shortened hypocotyl and inhibit lateral root development [26][27][28][29]. The increasing evidence has demonstrated that IAA3 acts as a node in the regulation of root meristem development by other internal signals, such as cytokinin, brassinosteroids and strigolactones [30][31][32]. However, the precise role of IAA3 in hypocotyl elongation awaits further investigation.
The interplay between light and auxin contributes to the regulation of a wide range of developmental processes. It was reported that light and auxin co-regulate a number of downstream genes, including the SMALL AUXIN UPREGULATED (SAUR) genes [33]. Furthermore, several reports have indicated that the light and auxin signaling pathways are integrated by the interaction between the light receptors and several auxin signaling components. For example, one study shows that phyA from oat could interact with several IAA proteins from Arabidopsis and pea in an in vitro recombinant system [34]. More recently, photoreceptors are shown to interact with IAA proteins to prevent the degradation of IAA proteins mediated by auxin receptors [35,36]. In addition, photoreceptors also target ARF6 and ARF8 to inhibit their DNA-binding activities [37]. Beyond these examples of photoreceptors interacting with IAA or ARF proteins, we wish to explore whether there are additional regulatory modules that directly mediate the crosstalk between the light and auxin signals.
Here, we report that the auxin transcriptional repressor IAA3 regulates hypocotyl elongation by directly modulating PIF protein activities. An IAA3 gain-of-function mutation led to insensitivity to light, whereas disruption of IAA3 caused elongated hypocotyl under different light intensity conditions. Genetically, the regulation of IAA3 in hypocotyl elongation is dependent on PIFs. Notably, IAA3 negatively regulates the expression of PIF target genes, by directly interacting with PIFs to interfere with PIFs binding to DNA. Our findings uncover a novel molecular mechanism in which IAA proteins coordinate with PIFs to repress the transcription of target genes, thereby ensuring precise control of hypocotyl growth in response to light and auxin cues.

Gain-of-function mutation in IAA3 causes hyposensitivity to light
Previous studies showed that a gain-of-function mutation in IAA3, shy2-3, caused shortened hypocotyl growth. Interestingly, the shy2-3 mutation could rescue the long hypocotyl phenotype of several light-signaling mutants such as phyB [26,29], indicating a possible role of IAA3 in light signaling to control seedling photomorphogenesis. To assess this putative IAA3 function, we first examined the light response of the shy2-3 gain-of-function mutant by treating the seedlings with different light fluence rates. Wild-type Ler seedlings responded to light normally, with an elongated hypocotyl under low light fluence conditions. In comparison, the hypocotyl of the shy2-3 mutant displayed a weaker response to light (Fig 1A). Analysis of hypocotyl elongation kinetics revealed that when the light intensity was low, shy2-3 had reduced hypocotyl elongation compared to wild-type plants (Fig 1B). The hypocotyl response to light in shy2-3 was similar to that of the pifq mutant, which also exhibited insensitivity to light (S1 Fig). Analysis of the shy2-3 phenotype under dark conditions revealed further similarities to the pifq mutant [7,38], as etiolated shy2-3 seedlings also exhibited a shortened hypocotyl and reduced hook curvature (S2 Fig). As IAA proteins are crucial components in auxin signaling pathway, we further analyzed the effect of auxin on wild type and shy2-3 mutant. We treated the seedlings with auxin biosynthesis inhibitors Kyn and PPBo to reduce the endogenous auxin level [39,40]. The results showed that the hypocotyl response of wild type plants to light became weakened when auxin was reduced, whereas the shy2-3 mutant showed no obvious change (Fig 1A and 1B). Together, these results suggest that IAA3 is required for auxin-regulated hypocotyl elongation in light conditions.

Disruption of IAA3 leads to elongated hypocotyl
IAA3 belongs to the auxin-induced AUX/IAA gene family, whose members often have high functional redundancy. As reported previously, several other IAAs, such as IAA6 and IAA7, may also contribute to hypocotyl elongation [24,25]. To elucidate the function of IAA3 and overcome the possible genetic redundancy, we used a dominant-negative strategy to disrupt the function of the related AUX/IAA genes. We generated a 35S-IAA3ΔEAR-VP16 construct by fusing the VP16 transcriptional activation domain to a mutated version of IAA3 lacking the EAR (ERF-associated Amphiphilic Repression) domain to effectively transform the IAA3 transcriptional repressor into an activator. This approach has successfully been used in previous studies to disrupt the redundant functions of multiple repressors [41,42]. We analyzed the hypocotyl elongation of the 35S-IAA3ΔEAR-VP16 transgenic plants under dark and different light intensities. The 35S-IAA3ΔEAR-VP16 transgenic plants displayed similar hypocotyl length with wild type plants in the dark (Fig 2A and 2B). However, compared with wild-type plants, the 35S-IAA3ΔEAR-VP16 transgenic plants showed constitutive longer hypocotyls under different light fluence rates (Fig 2C and 2D), further indicating the important role of IAA3 in light-regulated hypocotyl elongation.

Genetic interactions between IAA3 and PIFs
The similar hypocotyl phenotypes of shy2-3 and pifq under both light and dark conditions indicated a functional link between IAA3 and PIFs during hypocotyl development. To further reveal the possible functional link between IAA3 and PIFs, we first examined the tissue-specific  We then analyzed the genetic interactions between IAA3 and PIFs. We first crossed shy2-3 with each of the following 35S-PIF-MYC overexpression lines: PIF1-MYC, PIF3-MYC, PIF4--MYC, and PIF5-MYC. The seedlings of all of these PIF-MYC lines had elongated hypocotyls compared to wild type when grown under light, but the longest hypocotyls were observed in PIF4-MYC. The genetic analysis showed that overexpression of individual PIFs significantly rescued the shortened hypocotyl phenotype of the shy2-3 mutant (Figs 3A and S4). Furthermore, we transformed the IAA3ΔEAR-VP16 into the pifq mutant and analyzed two IAA3-ΔEAR-VP16/pifq independent lines (S5 Fig). The hypocotyl lengths of IAA3ΔEAR-VP16/pifq lines were comparable to those of pifq mutants (Fig 3B and 3C), suggesting that the function of IAA3 in hypocotyl regulation is dependent on PIFs.

IAA3 physically interacts with PIFs
The coordination of IAA3 and PIFs in hypocotyl elongation prompted us to investigate whether IAA3 and PIFs physically interact. For in vitro pull-down analysis, we co-expressed recombinant His-PIF1 and His-PIF4 with MBP or MBP-IAA3. We detected specific binding of His-tagged PIF1 and PIF4 with MBP-IAA3, with no band detected for the control MBP (Fig 4A), indicating that IAA3 interacted with these PIFs in vitro. We then conducted firefly luciferase complementation imaging (LCI) assays to investigate the in vivo interactions of IAA3 and PIFs in N. benthamiana leaves and Arabidopsis protoplasts. Luciferase activity signals were detected between IAA3 and PIFs (PIF1, PIF3, PIF4, and PIF5) in both N. benthamiana leaves and Arabidopsis protoplasts (Fig 4B and 4C). Finally, co-immunoprecipitation (Co-IP) experiments using tobacco plants coexpressing IAA3-GFP and PIF4-HA further confirmed that IAA3 interacts with PIF4 in vivo. Notably, we found that only the unphosphorylated form of PIF4 protein was immunoprecipitated by IAA3 ( Fig 4D). We therefore used LCI analysis to compare the affinity of IAA3 with the unphosphorylated form (designated PIF3 A20) and phosphorylated form of PIF3 (designated PIF3 D19), in which the light-induced phosphorylation sites were mutated to Ala or the phosphomimic mutations were introduced as previously described [43]. Consistent with the findings for PIF4, the LCI analysis indicated preferential binding of IAA3 to the unphosphorylated form of PIF3 (S6 Fig). Collectively, these in vitro and in vivo studies demonstrated that IAA3 physically interacts with PIFs at the protein level and preferentially binds to unphosphorylated PIFs.
PIF genes encode several important domains including an APB domain for phytochrome binding in the N-terminal and a bHLH domain for DNA binding in the C-terminal [44]. We generated two truncated forms of PIF4 to identify the regions of the protein required for interaction with IAA3 ( Fig 4E). We found that both the PIF4ΔC without the bHLH domain and PIF4ΔN without the APB domain are responsible for interaction with IAA3 ( Fig 4G). We also made truncated forms of IAA3 with only the N-terminal (IAA3ΔC) and C-terminal (IAA3ΔN) regions (Fig 4F). The results showed that only IAA3ΔN (lacking domains I and II) could interact with PIF4ΔN or PIF4ΔC (Fig 4H and 4I), indicating that the IAA3 C-terminal region containing domains III and IV is crucial for the interaction with PIF4.
We also tested the ability of PIF4 to interact with several other members of the AUX/IAA family. The analysis showed that PIF4 also interacted with IAA7, IAA10, IAA15, IAA26, IAA32, and IAA33 (S7 Fig), indicating that multiple AUX/IAA family members may function redundantly to cooperate with PIFs to control plant growth and development.

IAA3 inhibits PIFs binding to DNA
Based on the finding that IAA3 directly interacts with PIFs, we investigated whether changes in IAA3 protein abundance affected PIF accumulation in plants. Western blot analysis showed that PIF4 protein accumulation was not discernibly altered in the shy2-3 background (S8A  IAA3 was originally identified as a transcriptional repressor in the auxin signaling pathway [27]. To test whether IAA3 regulates the transcriptional activities of PIFs, we performed a transient expression assay using the reporter construct ProPIL1::LUC, in which the LUCIFERASE (LUC) gene was fused with the promoter region of PIL1, a direct target gene of PIF [45,46]. LUC activity analysis indicated that PIF4 alone directly activated the expression of PIL1, whereas IAA3 alone had no significant effect on its expression (Fig 5A). Remarkably, coexpression of IAA3 and PIF4 significantly repressed the activation executed by PIF4 (Fig 5A). The IAA3 and PIF4 protein levels were monitored by immunoblotting to ensure the comparable expression in each combination (Fig 5B). These results indicate that IAA3 interferes with the PIF transcriptional activity.
Together with our finding in LCI assay that IAA3 interacts with the DNA binding domain of PIF4, we then explored whether IAA3 interacts with PIFs to directly inhibit their DNAbinding ability. The electrophoretic mobility shift assay (EMSA) was conducted with purified recombinant His-PIF4 and MBP-IAA3 proteins. The promoter sequence of a PIF4 target gene, PIL1, was used as the binding probe ( Fig 5C). As expected, His-PIF4 specifically bound to the promoter of PIL1 (Fig 5D). Notably, such binding was progressively diminished when increasing amount of IAA3 protein was added (Fig 5D), supporting that IAA3 directly inhibits PIFs binding to their target genes.

IAA3 negatively regulates the expression of PIFs' target genes
To further investigate the gene regulatory networks of IAA3 and PIFs, we conducted RNA-seq transcriptome analysis of both shy2-3 and pifq seedlings. We found that 2088 genes and 1050 genes were differentially expressed in the shy2-3 and pifq mutant, respectively (q value<0.05; fold change�2 and �2) (Fig 5E and S1 Table). Of the1050 PIFs-regulated genes, 653 genes (62%) were oppositely affected by IAA3 (Fig 5E). The Heat map analysis also showed those genes display the similar pattern of FPKM value in shy2-3 and pifq mutants compared with that of in wild type (Fig 5F). Based on the repressive effect of IAA3 to PIFs, we then focused on the 262 genes that were activated by PIFs but repressed by IAA3. Gene ontology (GO) analysis of these genes showed that many auxin-responsive genes, such as SAURs, are enriched (S9A and S9B Fig). In addition, we also found several PIFs direct target genes were downregulated in shy2-3 mutant, such as PIL1, PRE1, LTP3 and EXP8 [45][46][47][48]. Quantitative RT-PCR analysis confirmed that the expression levels of these PIFs' target genes were significantly lower in the shy2-3 mutant than in wild-type controls (Fig 5G). Consistently, these genes were significantly upregulated in 35S-IAA3ΔEAR-VP16 lines (S10 Fig). Interestingly, we found those PIFs' target genes could be significantly upregulated by auxin (Fig 5H). Taken together, these results indicate that IAA3 negatively modulates the PIFs-dependent gene expression.

PIFs are required for auxin-mediated hypocotyl elongation
The activation of transcriptional auxin signaling pathways is essential for the auxin-mediated hypocotyl elongation response [49]. Based on our findings that PIFs also regulate genes that are responsive to auxin (S9 Fig), we then determined whether PIFs are required for auxin-for the combinations listed in the rectangles (PIF4ΔC/IAA3 and PIF4ΔC/IAA3ΔN in (H) and PIF4ΔN/IAA3 and PIF4ΔN/IAA3ΔN in (I); negative controls and other combinations that exhibited no interaction are indicated below the graphs. https://doi.org/10.1371/journal.pgen.1009384.g004

Fig 5. IAA3 impairs PIFs binding to DNA and negatively regulates PIFs' target genes. (A) The transient dual-LUC reporter assay
showing the effects of IAA3 on PIF transcriptional regulation of PIL1. The ATG18-GFP protein is as the negative control in the assay. Data are means ± SD (n = 3 experiments). Significant differences are indicated with asterisks ( ��� P < 0.001, two-tailed Student's t-test). (B) Western blotting for the detection of expressed PIF4-MYC, IAA3-GFP and ATG18-GFP in transformed tobacco leaves from (A) using an anti-MYC or anti-GFP antibody, respectively. (C) A fragment for PIL1 promoter containing Gbox motif (CACGTG) was selected to used as probe in the EMSA experiment. (D) EMSA experiment showing IAA3 inhibition of the DNA-binding activity of PIFs. The cold probe was added as a competitor. MBP was used as a negative control. (E) Venn diagrams showing significant overlap between IAA3-regulated genes and PIFs-dependent genes. Detailed gene information is given in S1 Table. (F) Heat map of 625 DEGs coregulated by IAA3 and PIFs. (G) and (H) qRT-PCR analysis of known PIFs' direct target genes. The transcription levels of PIFs' direct target genes in Ler and shy2-3 mutant (G). Auxin induction of PIFs-regulated genes. Five-day Col-0 seedlings grown in light condition were treated with 5 μM IAA for 24 hours (H). The expression was normalized to the ACTIN2 (ACT2) expression level, and was relative to wild-type levels. Data represents mean ±SD from three biological replicates. Significant differences are indicated by ��� P < 0.001(two-tailed Student's t-test).
https://doi.org/10.1371/journal.pgen.1009384.g005 mediated hypocotyl elongation. Similar to other AUX/IAA gain-of-function mutants, the shy2-3 hypocotyl was insensitive to auxin treatment, which otherwise enhanced the hypocotyl elongation in wild-type plants (Fig 6A and 6B). Remarkably, as in the shy2-3 mutant, the hypocotyl of the pifq mutant also exhibited pronounced resistance to auxin (Fig 6A and 6C). These results support that PIFs are required for auxin-mediated hypocotyl elongation, consistent with the previous report that PIF4 and PIF5 affect hypocotyl sensitivity to auxin [50].

Discussion
The growth responses of sessile plants to fluctuating environmental conditions must be precise and adaptive. Hypocotyl growth is a developmental event with high plasticity, and it is antagonistically modulated by environmental light and endogenous auxin signals [5]. However, how light and auxin are integrated to regulate hypocotyl growth remains largely unknown. IAAs and PIFs are transcriptional regulators that negatively regulate the auxin and light signaling pathways, respectively [10,23]. In this study, we discovered that IAA3 represses the DNAbinding activity of PIFs through physical interaction, thereby coordinating the light and auxin signals to control hypocotyl elongation.
Several previous reports suggested a close relationship between IAA3 and PIFs. One study indicated that IAA3 is involved in PIF4-mediated high temperature-induced hypocotyl growth [51]. Moreover, IAA3 and PIFs co-regulated a great amount of genes [19]. However, a direct connection between IAA3 and PIFs has not previously been established. In this study, we found that IAA3 directly interacts with multiple PIFs and impairs the binding of PIFs to their target gene promoters (Figs 4 and 5). Notably, we observed preferential binding of IAA3 to unphosphorylated forms of PIFs (Figs 4D and S6). This suggests that even if PIFs are not sufficiently phosphorylated to be degraded under specific conditions, they may nevertheless be subject to repressive regulation by other factors such as IAAs. Under natural conditions, different plant organs may sense and respond to environmental signals such as light in different ways [33]. Thus, the alternative regulation of phosphorylated or unphosphorylated PIFs may be an adaptation allowing specific responses in different organs to the natural environment: compared to the hypocotyl, young leaves may expand to perceive more light, resulting in a greater abundance of the phosphorylated form of PIFs in the leaves and abundant unphosphorylated PIFs in the hypocotyl. In this model, the preferential binding of IAA3 to unphosphorylated PIFs would allow for precise regulation of hypocotyl growth.
It has been reported that IAA proteins could inhibit the DNA-binding activities of ARFs [52,53]. Our finding showed that IAA3 interacts with the DNA binding domain of PIF4 ( Fig  4G) and inhibits PIF activity by interfering with their binding to target DNA (Fig 5D). The previous studies also reported that IAA proteins regulate ARF activities by recruiting the TPL/ TPR co-repressors [54]. Accordingly, IAA3 may also repress the transcriptional activities of PIFs by recruiting TPL. It is possible that these two mechanisms cooperate together to precisely regulate PIF activities and hypocotyl elongation in response to light and auxin signals.
Although the four PIFs function redundantly to regulate the hypocotyl elongation, they may have distinct contributions for hypocotyl regulation. PIF4 showed a relatively higher expression in hypocotyl than other PIFs (S3G Fig), and the overexpression of PIF4 lines showed the longest hypocotyls compared with that in other PIFs overexpression lines (Fig 3A). These results imply that PIF4 might play a dominant role in hypocotyl regulation, thus resulting in the highest degree of shy2-3 hypocotyl phenotypic rescue by PIF4 overexpression. Furthermore, IAA3 may display distinct binding ability with the different PIF members, probably a stronger binding ability with PIF1 compared with other PIFs (Fig 4C). As reported previously, PIF1 plays a predominant role in repressing seed germination [55]. It will be interesting to explore whether and how seed germination is regulated by IAA3 and PIF1 transcriptional regulatory module.
ARFs are typical transcription factors that act downstream of IAAs to respond to the auxin signal [56]. Previous reports have indicated that IAA3 associates with ARF7 and ARF19 to regulate root growth [28]. Several ARFs also function redundantly to regulate hypocotyl elongation, and probably through interaction with PIFs or BZR1 [19,57]. Interestingly, the analysis of tissue specific expression of IAA3 and PIFs indicated that IAA3 and PIFs may coexist primarily in cotyledon and hypocotyl (S3 Fig). This also indicates a specificity of signal integration achieved through IAA3 interaction with different types of transcription factors in specific tissues.
Auxin synthesis is considered to occur in the cotyledon, followed by transport from the shoot to the root [5]. Light is a key factor affecting cotyledon auxin levels, as it impacts PIFs accumulation as well as YUCCA transcript levels [58]. On the basis of previous reports and the results of the present study, we propose a model that illustrates how light and auxin signals be integrated in specific tissues (Fig 7). Low light signals trigger auxin synthesis in the cotyledon, which is mediated by a transcriptional regulatory module consisting of PIFs and YUCCA. Auxin is subsequently transported to promote hypocotyl elongation, which is mediated by the interaction between IAAs and different types of transcription factors, including PIFs, ARFs, and BZR1.
Our data showed that PIF4 could also interact with several non-canonical IAA proteins (S7 Fig), such as IAA32 and IAA33, which lack the conserved domains I and II [24]. Mechanistically, these non-canonical IAAs may function distinctly from canonical IAAs. IAA32 was recently identified as a target of transmembrane kinase 1 (TMK1) in the regulation of apical hook development [59]. As previously reported, PIFs also affect apical hook development [38]. Thus, the interaction between IAA32 and PIF4 observed in the present study is consistent with their involvement in apical hook development, most likely as downstream components in the TMK1-mediated signaling pathway. Alternatively, since different AUX/IAA family members function redundantly during hypocotyl elongation [25,26], these PIF-interacting non-canonical IAAs may also regulate hypocotyl elongation. The functions of these non-canonical IAAs in hypocotyl elongation warrant further study, along with the connection between non-canonical and canonical IAAs.

Plant materials and growth conditions
All Arabidopsis thaliana mutants and transgenic lines used in this study except shy2-3 are in the Col-0 background. shy2-3 is in the Ler background [27]. pif1-1 pif3-3 pif4-2 pif5-3 (pifq)  The different responses of specific seedling tissues to changing light conditions and auxin reflect different patterns of IAA and PIF activity in these tissues. In young leaves, low light is perceived by photoreceptors and then triggers the accumulation of PIFs, leading to increased YUCCA expression and auxin abundance. The auxin in the leaves is transported to the hypocotyl, in which IAA3 is degraded through the TIR1-mediated proteasome pathway, releasing the inhibition of several transcription factors including PIFs, ARFs, and BZR1. The target genes of these transcription factors are activated to promote hypocotyl elongation. When the light conditions are sufficiently high, PIF accumulation in the leaves is reduced, leading to decreased YUCCA expression, and then reduced auxin levels in the leaves. Thus, the auxin in the hypocotyl which is transported from leaves is also reduced. Consequently, IAA3 represses PIFs, ARFs, and BZR1 in the hypocotyl. The blue arrows in the diagram represent auxin flow and the thickness of the arrows represent auxin level.
Surface-sterilized seeds were sown on Murashige and Skoog (MS) medium (4.4 g/L MS salt, 1% sucrose, pH 5.7, and 0.8% agar) and imbibed for 3 d at 4˚C in the dark. Light-grown seedlings were grown under white fluorescent light (about 130 μmol m -2 s -1 ) with a 16 h light/8 h dark photoperiod unless otherwise indicated. Dark-grown seedlings were exposed to white light for 4 h then transferred to dark conditions. Hypocotyl lengths were measured from the intersection of the cotyledons to the root using ImageJ software, and more than 15 seedlings were measured for each set of experiments.

RNA extraction, reverse transcription, and real-time PCR
Total RNA was extracted from seedlings using Trizol reagent (Invitrogen). Total RNA was subjected to reverse transcription at 42˚C for 1 h with M-MLV reverse transcriptase (Promega). Real-time PCR was performed on a Light Cycler 480 system (Roche) with SYBR Premix Ex Taq reagent (TaKaRa). Each experiment was performed at least three times using different pools of seedlings under the indicated conditions.

Firefly luciferase complementation assay
The coding sequences of PIF1/3/4/5, PIF4ΔN, and PIF4ΔC were individually inserted into the multiple cloning site of pCAMBIA1300-nLUC, and the coding sequences of IAA3, IAA3ΔN, and IAA3ΔC were inserted into the multiple cloning site of pCAMBIA1300-cLUC [62]. Different vector combinations were transformed into Agrobacterium strain GV3101, which was subsequently infiltrated into tobacco leaves. After infiltration, the tobacco leaves were grown under weak light for 12 h then transferred to light conditions for 48 h. The abaxial sides of leaves were sprayed with 1 mM luciferin, followed by a 10 min incubation in the dark. Luciferase activity was recorded on a Xenogen IVIS Spectrum instrument.
For transformation into Arabidopsis protoplasts, plasmids were extracted and purified using Plasmid Maxiprep kits (Vigorous). Arabidopsis protoplasts were prepared from 10-dayold green seedlings or 4-week-old adult seedlings following a standard protocol. Ten micrograms of both the nluc-tagged and cluc-tagged constructs were co-transformed into protoplasts and incubated for 12 to 16 h in darkness. After the incubation period, luciferin (Gold Biotechnology) was added, and luciferase activity was recorded on an LB 985 NightSHADE system (Berthold Technologies).

Transient transcription activation assay
The ProPIL1::LUC reporter was constructed by amplifying 1.8 kb of the Arabidopsis PIL1 promoter and inserting the fragment into the pGreen II 0800-LUC vector. The coding sequences of PIF4, IAA3, and ATG18 were amplified and inserted into the PQG110 vector to generate effector constructs. The DLR system (Promega) was used for transient expression analysis in leaves of tobacco [63]. The activities of firefly (Photinus pyralis) luciferase and Renilla reniformis luciferase (LUC and REN, respectively) were sequentially measured from a single sample on a GLO-MAX 20/20 luminometer (Promega). The ratio of LUC/REN was calculated as an indicator of the final transcriptional activity.

Protein extraction and immunoblotting
Whole seedlings were finely ground in liquid nitrogen and suspended thoroughly in protein extraction buffer (60 mM Tris-HCl, pH 6.8, 25% glycerol, 2% SDS, 14.4 M β-mercaptoethanol, 0.1% bromophenol, and 1 M DTT) by vortexing. Proteins were extracted by boiling for 10 min at 100˚C then roughly separated via centrifugation (13,000 rpm for 15 min at 4˚C). Supernatants were loaded onto 10% SDS-PAGE gels, and the proteins were transferred to a nitrocellulose filter membrane (Millipore) according to a standard protocol. Anti-MYC (ABclonal) was diluted 5000-fold for incubation with the membranes. Goat anti-mouse HRP-conjugated secondary antibody (Promega) was diluted 7500-fold.

Co-immunoprecipitation assay
For Co-IP experiments, the 35S-IAA3-GFP and 35S-PIF4-HA constructs were transformed into Agrobacterium strain GV3101, which was subsequently infiltrated into tobacco leaves. After infiltration, the tobacco leaves were grown under weak light for 12 h then transferred to light conditions for 48 h. Soluble and nuclear proteins from tobacco leaves were isolated using IP buffer (50 mM Tris-HCl, pH 8.0, 150 mM NaCl, 10% glycerol, 0.5 mM EDTA, 0.1% Triton X-100, 5mM β-mercaptoethanol, and 1× cocktail). The soluble proteins from tobacco leaves co-expressing 35S-GFP and 35S-PIF4-HA or 35S-IAA3-GFP and 35S-PIF4-HA were immunoprecipitated with GFP-binding protein (GBP, a gift from Professor Junyu Xiao of Peking University) that binds Ni-NTA Agarose (Qiagen). The mixtures were maintained at 4˚C with gentle shaking for 3 h then washed with cold IP buffer three times. The samples were detected by anti-HA (Qiagen) and anti-GFP antibodies (Abmart).

Protein expression and purification
The coding sequences of PIF1 and PIF4 were cloned into the pCold-TF vector (GE Healthcare) and transformed into Transette (DE3) competent cells (TransGen). Protein expression was induced by 0.1 mM IPTG at 16˚C, and proteins were purified with Ni-NTA Agarose (Qiagen) following the manufacturer's instructions. The coding sequence of IAA3 was cloned into pMAL-c2X vector (NEB) then transformed into Transette (DE3) competent cells (TransGen). Protein expression was induced by 0.4 mM IPTG at 16˚C, and proteins were purified with amylose agarose (NEB) following the manufacturer's instructions.

Pull-down assay
The in vitro-purified MBP and MBP-IAA3 were incubated with Amylose Resin (NEB) in pulldown buffer (50 mM Tris-Cl, pH 8.0, 150 mM NaCl, 10% glycerol, 0.5 mM EDTA, 0.1% Triton X-100, 5 mM β-mercaptoethanol, and 1× protease inhibitor cocktail) for 4 h at 4˚C. The fusion protein His-PIF1 and His-PIF4 was added and incubated for another 3 h at 4˚C. After five washes with pull-down buffer, the precipitated resins were collected by brief centrifugation then treated with protein extraction buffer. Proteins were separated by SDS-PAGE and detected with anti-MBP (NEB) or anti-His (Tiangen).

EMSA assay
The oligonucleotide probes for PIL1 were synthesized with the sequence information described previously [46]. The PIL1 probe was labeled with digoxigenin (DIG) by company. The LightShift Chemiluminescent EMSA Kit (Pierce) was used in the assay. The biotin-labeled probe was incubated with the proteins in 1× binding buffer for 60 min. A 500-fold excess of cold probe was added to the reaction system for competition. The 6% polyacrylamide gel was used for electrophoresis.

RNA sequencing
The five-day-old seedlings of Ler, shy2-3, Col-0 and pifq plants grown in light were collected to extract RNA for RNA sequencing using the Illumina HiSeq X Ten system in biotechnology company. The bioinformatic and statistical analysis of the RNA-seq data was performed as described previously [64]. Genes with changes of more than two-fold (Q-value � 0.05) were defined as differentially expressed genes.