Non‐canonical AUX/IAA protein IAA33 competes with canonical AUX/IAA repressor IAA5 to negatively regulate auxin signaling

Abstract

The phytohormone auxin controls plant growth and development via TIR1‐dependent protein degradation of canonical AUX/IAA proteins, which normally repress the activity of auxin response transcription factors (ARFs). IAA33 is a non‐canonical AUX/IAA protein lacking a TIR1‐binding domain, and its role in auxin signaling and plant development is not well understood. Here, we show that IAA33 maintains root distal stem cell identity and negatively regulates auxin signaling by interacting with ARF10 and ARF16. IAA33 competes with the canonical AUX/IAA repressor IAA5 for binding to ARF10/16 to protect them from IAA5‐mediated inhibition. In contrast to auxin‐dependent degradation of canonical AUX/IAA proteins, auxin stabilizes IAA33 protein via MITOGEN‐ACTIVATED PROTEIN KINASE 14 (MPK14) and does not affect IAA33 gene expression. Taken together, this study provides insight into the molecular functions of non‐canonical AUX/IAA proteins in auxin signaling transduction.

In contrast to other AUX/IAA‐family repressor, IAA33 is stabilised by auxin and competes with canonical IAA5 for binding to downstream auxin response factors ARF10/16.

Introduction

The phytohormone auxin regulates almost every aspect of plant growth and development (Mockaitis & Estelle, 2008; Lavy & Estelle, 2016; Leyser, 2018). Since the identification of the TIR1 auxin receptor, the auxin signaling pathways have been well investigated (Dharmasiri et al, 2005; Kepinski & Leyser, 2005). A canonical auxin signaling pathway starts from auxin perception by a co‐receptor complex, comprised of TIR1/AFB receptors and AUX/IAA proteins, followed by AUX/IAA protein ubiquitination and degradation, and eventually derepression of AUXIN RESPONSE FACTORs (ARFs) and the transcriptional activation of auxin‐induced gene expression (Peer, 2013). The domain II of AUX/IAA proteins mediates the interaction with TIR1/AFB receptors, a process which is promoted by auxin (Ramos et al, 2001; Kepinski & Leyser, 2004; Tan et al, 2007). There are 23 ARFs in Arabidopsis (Hagen & Guilfoyle, 2002). Structural studies revealed that there are three conserved regions of homology in these ARFs: an N‐terminal B3‐type DNA‐binding domain (DBD) and two C‐terminal regions, which share homology with domains III and IV of AUX/IAA proteins (Guilfoyle & Hagen, 2007; Korasick et al, 2014). Domains III and IV are responsible for homo‐ and heterodimerization of AUX/IAA proteins or ARFs (Tiwari et al, 2003). Between the N‐terminal DBD and C‐terminal domains, ARFs contain a variable middle domain, which has been proposed to confer either activation or repression properties of these transcription factors. Based on the amino acid composition of the middle regions, ARFs can be classified as activators or repressors (Tiwari et al, 2003). ARF10 and ARF16, which have been shown to regulate root stem cell identities (Ding & Friml, 2010), were characterized as transcriptional repressors (Wang et al, 2005; Bennett et al, 2014). Several recent studies suggest a high complexity of the auxin signaling pathway (Jing et al, 2015; Wang et al, 2015; Yu et al, 2015; Dezfulian et al, 2016), which can also explain how auxin modulates diverse aspects of plant growth and development.

Although canonical auxin signaling has been well studied, whether the non‐canonical AUX/IAA proteins, which lack the conserved domain II, take part in the auxin signaling or the mechanism of non‐canonical auxin signaling still remains elusive. A recent study showed that the non‐canonical AUX/IAA proteins, IAA20 and IAA30, were required for the proper vascular patterning. The double mutant iaa20/30 formed ectopic protoxylem, while overexpression of IAA30 caused discontinuous protoxylem and occasional ectopic metaxylem (Muller et al, 2016). Chen et al (2018) reported that a RING finger E3 ubiquitin ligase (SOR1) controlled root‐specific ethylene responses by modulating a non‐canonical AUX/IAA protein (OsIAA26) stability. Recently, Cao et al reported that auxin‐mediated C‐terminal cleavage of the TRANSMEMBRANE KINASE 1 (TMK1) leads to phosphorylation of two non‐canonical AUX/IAA proteins (IAA32 and IAA34) and their subsequent stabilization to regulate differential growth of the apical hook (Cao et al, 2019).

In this study, we investigated the role of IAA33, a non‐canonical AUX/IAA protein without typical domains I and II, which are essential components to mediate the canonical auxin signaling through the TIR1‐dependent pathway. The provided evidence shows that IAA33 is involved in auxin signaling through interacting with ARF10 and ARF16, which have been reported to control root distal stem cell (DSC) identity (Ding & Friml, 2010), and consequently regulates root DSC identity. IAA33 negatively regulates auxin response through the competition with IAA5, a canonical AUX/IAA protein, and thus releases the repression of ARF10/16 in this process. Furthermore, different from the up‐regulation of the transcription and the destruction of canonical AUX/IAA proteins such as IAA5, auxin stabilizes the IAA33 protein through interaction with MPK14 without influencing its transcription.

Results

IAA33 negatively regulates auxin responses

The canonical AUX/IAA proteins act as transcriptional repressors and mediate auxin signaling through interaction with TIR1 receptors (Villalobos et al, 2012; Weijers & Wagner, 2016). IAA33 has no domains I and II, which represses ARF‐mediated transcription and mediates the interaction between AUX/IAA protein and TIR1, respectively (Appendix Fig S1A). Consistent with the absence of domain II, IAA33 could not interact with TIR1, while the canonical domain II‐containing IAA5 could interact with TIR1 in yeast (Appendix Fig S1B). To address whether the non‐canonical IAA33 could regulate auxin signaling, we examined auxin responses using DR5rev::GFP as a marker in iaa33 and IAA33 ^OE lines. A highly increased transcriptional auxin response was observed in iaa33, which is reflected by the increased DR5rev::GFP signal (Fig 1A). Consistently, when IAA33 is overexpressed, the auxin response was repressed, which is shown by the reduced DR5rev::GFP signal (Fig 1A). Similarly, co‐expression of DR5::LUC with the 35S::IAA33 construct in protoplast cells isolated from Arabidopsis leaves led to an obvious reduction in luminescence intensity (Fig 1B), further suggesting that overexpression of IAA33 represses auxin response. Our qRT–PCR results also showed that the relative expression levels of auxin‐induced genes such as IAA3/4/5/11/13/18/28 were increased in iaa33 compared with Col‐0 (Fig 1C). Taken together, these results suggest that IAA33 negatively regulates auxin response.

Figure 1. IAA33 negatively regulates auxin responses.

1. Expression of the auxin reporter DR5rev::GFP in Col‐0, iaa33, and IAA33 ^OE roots. DR5rev::GFP crossed with iaa33 or IAA33 ^OE seedlings, respectively. Scale bars, 50 μm.
2. Transient expression analysis of overexpression of IAA33 on DR5::LUC activity in A. thaliana protoplasts. IAA33 was co‐transfected with DR5::LUC. The LUC‐to‐REN ratio was shown to indicate the expression level of the DR5::LUC. LUC: firefly luciferase activity, REN: Renilla luciferase activity. Data shown are means ± standard errors (n = 9); *: means significant difference compared to control (P < 0.05) based on Duncan's test.
3. The relative expression of IAA3/4/5/11/13/18/28 in Col‐0 and iaa33. RNA was isolated from the roots of 6‐day‐old Col‐0 and iaa33 seedlings using an RNeasy Plant Mini Kit (Qiagen, Hilden, Germany) following the manufacturer's protocol. Data shown are means ± standard errors (n = 3); *: means significant difference compared to Col‐0 (P < 0.05) based on Duncan's test.

IAA33 controls root stem cell identity

Analysis of an IAA33p::GUS‐GFP line showed that IAA33 is expressed in root tips (Fig 2E). Since auxin plays an essential role in the maintenance of root stem cell identity (Blilou et al, 2005; Ding & Friml, 2010), we also examined the role of IAA33, which negatively regulates auxin response, in this developmental context. The results showed that the iaa33 mutant exhibited reduced root distal stem cell (DSC) differentiation, shown by a higher rate of seedling roots with 2 layers of DSCs (Fig 2A–C, Appendix Fig S2). In contrast, the IAA33 overexpression (IAA33 ^OE) lines displayed enhanced root DSC differentiation reflected by a higher rate of seedling roots without DSCs (Fig 2A–C, Appendix Fig S2). The allele test showed that the F1 of iaa33‐cas9 crossed with iaa33 also had a higher rate of 2 layers of DSCs, which is similar to iaa33 mutant (Appendix Fig S2C). These observations are in line with IAA33 negatively regulating auxin response.

Figure 2. IAA33 controls root distal stem cell (DSC) identity.

1. In Lugol‐stained 5‐day‐old Col‐0, iaa33, iaa5, IAA33 ^OE, and IAA5 ^OE seedling roots, root distal stem cell (DSC) differentiation is reduced in iaa33 and IAA5 ^OE seedlings and enhanced in IAA33 ^OE and iaa5 seedlings (red asterisk indicates QC cells, and yellow arrowheads indicate distal stem cell). Scale bars, 20 μm.
2. Quantitative evaluation of root DSC layers in 5‐day‐old seedlings of Col‐0, iaa33, and IAA33 ^OE, which are grown on MS medium (n = 50).
3. The relative expression of IAA33 in Col‐0, iaa33, and IAA33 ^OE. Data shown are means ± standard errors (n = 3), **: means of iaa33, IAA33 ^OE differ significantly from mean of Col‐0 (P < 0.01) based on Duncan's test.
4. Quantitative evaluation of root DSC layers in 5‐day‐old seedlings of Col‐0, iaa5, and IAA5 ^OE, which are grown on MS medium (n = 50).
5. IAA33 is expressed in root. Yellow asterisk indicates QC cells. Black scale bars, 1 cm. White scale bars, 50 μm.

IAA33 interacts with ARF10 and ARF16

It was previously reported that among the non‐canonical AUX/IAA proteins, ARF10/16 could interact with IAA32, IAA33, and IAA34 (Piya et al, 2014). Both ARF10 and ARF16 are well known to regulate auxin signaling and root stem cell identity (Wang et al, 2005; Ding & Friml, 2010; Bennett et al, 2014). The expression of IAA32p::GUS and IAA34p::GUS, which are expressed in the apical hook (Cao et al, 2019), is absent from the root tip (Appendix Fig S3). This suggested that IAA32/34 are likely not involved in the regulation of root development. Therefore, we next only focused on examining whether IAA33 could interact with root‐expressed ARF10 and ARF16 (Rademacher et al, 2011). A yeast two‐hybrid analysis confirmed that IAA33 could interact with both ARF10 and ARF16 (Appendix Fig S5A), which was consistent with the previous report (Piya et al, 2014). Furthermore, bimolecular fluorescence complementation (BiFC) assays also showed strong fluorescence signal in tobacco epidermal cells co‐expressing IAA33 fused to the N‐terminal half of YFP (^NYFP) and ARF10 or ARF16 fused to the C‐terminal half of YFP (^CYFP), whereas no signal was observed in the empty vector control (Fig 3A). To confirm this, we performed in vitro pull‐down assays with GST‐tagged IAA33 in combination with His‐tagged ARF10 or ARF16. The results showed that IAA33 could interact with ARF10 or ARF16 in vitro (Fig 3B). Finally, we further confirmed the interaction between IAA33 and ARF10 or ARF16, using co‐immunoprecipitation assays in tobacco leaves. IAA33‐GFP protein was immunoprecipitated by anti‐MYC antibody from tobacco leaf cells co‐expressing ARF10‐MYC or ARF16‐MYC (Fig 3C and D). Together, all these results indicate that IAA33 interacts with ARF10 or ARF16.

Figure 3. IAA33 interacts with ARF10 and ARF16.

• A
  BiFC analysis of interaction between IAA33 and ARF10/16. The split YFP system was used in BiFC assays. ^CYFP and ^NYFP are empty vectors. The different combinations of plasmids were transformed into tobacco epidermal cells, and the YFP signals were detected with a confocal microscope. Scale bars, 50 μm.
• B
  Western blotting results of the GST pull‐down assay of IAA33‐GST and ARF10‐His or ARF16‐His. MBP‐His was used as the negative control.
• C, D
  In vivo Co‐IP assays of IAA33 with ARF10 (C) or ARF16 (D). ARF10‐MYC or ARF16‐MYC was co‐expressed with IAA33‐GFP in tobacco leaves. Protein extracts (Input) were immunoprecipitated with anti‐MYC antibody (IP). Immunoblots were developed with anti‐GFP antibody to detect IAA33 and with anti‐MYC to detect ARF10 and ARF16.

Source data are available online for this figure.

Furthermore, genetic analysis showed that the reduced DSC differentiation in iaa33 and the higher frequency of DSC differentiation in IAA33 ^OE were both repressed by overexpression of or mutations in ARF10 or ARF16, respectively (Appendix Fig S4), indicating that IAA33 controls root stem cell identity through ARF10 and ARF16.

The canonical IAA5 has also been reported to interact with ARF10 or ARF16 in yeast (Piya et al, 2014). We confirmed this interaction through BiFC assays, pull‐down assays, and Co‐IP analysis (Appendix Fig S5). Consistently, the overexpression of IAA5 (IAA5 ^OE) lines exhibited reduced root DSC differentiation, shown by a higher rate of seedling roots with 2 layers of DSCs. In contrast, the iaa5 mutant displayed enhanced root DSC differentiation reflected by a higher rate of seedling roots without DSCs (Fig 2A and D).

We also tested the interaction between IAA33 and other repressor ARFs or activator ARFs by yeast two‐hybrid assays and found that in addition to ARF10/16, IAA33 also strongly interacted with ARF1 and ARF18 (Appendix Fig S6). In agreement with previously published data (Piya et al, 2014), this indicated that IAA33 preferentially binds to repressor ARFs.

IAA33 competes with IAA5 for ARF10/16 binding

IAA5 is a canonical AUX/IAA protein, which could interact with TIR1 (Shimizu‐Mitao & Kakimoto, 2014) (Appendix Fig S1) and negatively regulated transcriptional auxin response (Appendix Fig S7). Furthermore, both IAA33 and IAA5 strongly interacted with ARF10 or ARF16 (Fig 3 and Appendix Fig S5), while the interaction between IAA33 and IAA5 was weak (Appendix Fig S8). This indicated that IAA33 prefers to interact with ARF10/16 rather than heterodimerize with IAA5. Therefore, we examined whether IAA33 could compete with IAA5 to interact with ARF10 or ARF16. In vitro pull‐down assays indicated that increasing the level of GST‐tagged IAA33 clearly reduced the interaction between IAA5 with ARF10 or ARF16, suggesting that IAA33 could compete with IAA5 to interact with ARF10 or ARF16 (Fig 4A). This result was also confirmed by yeast three‐hybrid assays. In the presence of co‐expressed IAA33, the interaction between IAA5 and ARF10 or ARF16 was blocked, which was shown through the inactivation of the reporter (Fig 4B). However, in the presence of co‐expressed IAA5, we still observed the activation of the reporter, which showed the interaction between IAA33 and ARF10 or ARF16 (Appendix Fig S9). In addition, IAA5, IAA33, ARF10, and ARF16 all repressed transcriptional auxin response, which was shown by decreased DR5::LUC activity in transient expression assays in Arabidopsis protoplasts (Fig 4C and D). Furthermore, repression of transcriptional auxin response by ARF10 and ARF16 could be partially alleviated through co‐expression of IAA5. However, this negative regulation of IAA5 on ARF10 and ARF16 is removed by IAA33 (Fig 4C and D). All these results indicate that IAA33 regulates auxin signaling through competition with IAA5.

Figure 4. IAA33 interacts with ARF10 and ARF16 through the competition with IAA5.

• A
  Immunoblot of an immunoprecipitation (IP)/protein competition assay co‐expressing IAA5‐MBP and ARF10‐His or ARF16‐His with increasing amounts of IAA33‐GST. The amounts of IAA5‐MBP and 1X IAA33‐GST were 20 μg, respectively. The anti‐MBP antibody used for IP is indicated at the top panel, while anti‐GST for IP is shown at the bottom panel.
• B
  Yeast three‐hybrid assay analyzing the IAA5‐ARF10/16 interaction in the presence or absence of co‐expressed IAA33. Co‐transformants were spotted on SD‐Leu‐Trp medium to check for viability and on SD‐Met‐His‐Leu‐Trp medium to test the interaction and competition, respectively.
• C, D
  IAA33 removes the repression of IAA5 on ARF10‐ and ARF16‐mediated auxin signaling. Values are the means ± standard errors of three biological replicates. Different letters above bars indicate a significant difference (P < 0.05) based on Duncan's test.

Source data are available online for this figure.

Auxin can induce the accumulation of IAA33

The canonical AUX/IAA proteins act as transcriptional repressors that are degraded through the 26S proteasome pathway in the presence of auxin and thus release the repression of ARF transcription factors and activate auxin signaling (Gray et al, 2001; Ramos et al, 2001). On the other hand, the activated auxin signaling could induce the expression of these canonical AUX/IAAs (Walker & Key, 1982). IAA5 is a canonical IAA with 4 typical domains (Appendix Fig S1). Therefore, it was not surprising that we observed the clear up‐regulation of IAA5 expression with auxin treatment as shown by qRT–PCR and GUS staining of an IAA5p::GUS line (Appendix Fig S10A and B). As expected, the IAA5 protein was degraded in the presence of auxin (Appendix Fig S10C and D), a process which is dependent on the 26S proteasome pathway since MG132 co‐treatment could slow down the degradation of IAA5 (Appendix Fig S10E). However, both qRT–PCR analysis and GUS staining assays in IAA33p::GUS‐GFP lines suggested that the transcription of the non‐canonical IAA33 is not affected by auxin (Fig 5A and B). Furthermore, though IAA33 is degraded through the 26S proteasome pathway, which is inhibited by MG132 (Fig 5D), we did not detect the auxin‐induced degradation of IAA33 (Fig 5C and E, Appendix Fig S11). On the contrary, we observed that auxin stabilized the IAA33 protein, which was shown by the accumulated IAA33 upon NAA treatment (Fig 5C and E, Appendix Fig S11). All these results suggest that, compared to canonical AUX/IAA proteins such as IAA5, the non‐canonical IAA33 is regulated differently in response to auxin at both transcriptional and protein levels.

Figure 5. Auxin induces IAA33 protein accumulations.

1. GUS staining analysis of IAA33p::GUS‐GFP seedlings when treated with or without NAA (10 μM) for 4 h. Scale bars, 50 μm.
2. Transcript abundance of IAA33 when treated with or without NAA. Five‐day‐old seedlings were treated with or without NAA (10 μM) for 4 h, and used for RNA extraction and subjected to qRT–PCR analysis. Data shown are means ± standard errors (n = 3).
3. Confocal images of 35S::GFP‐IAA33 seedling roots when treated with or without NAA (10 μM) for 4 h. Scale bars, 50 μm.
4. IAA33 protein level was examined by Western blot at different time points after CHX only or CHX and MG132 co‐treatment in 5‐day‐old 35S::GFP‐IAA33 seedlings. The relative intensity of band detected by anti‐GFP antibody to that by anti‐actin antibody without treatment was set to 1.0.
5. IAA33 protein level was examined by Western blot when 5‐day‐old 35S::GFP‐IAA33 seedlings were treated with auxin (10 μM NAA) at indicated time points. The relative intensity of band detected by anti‐GFP antibody to that by anti‐actin antibody without treatment was set to 1.0.

Source data are available online for this figure.

IAA33 interacts with MPK14

To elucidate the molecular mechanism of auxin‐induced IAA33 protein accumulation, we performed a yeast two‐hybrid assay to identify interacting proteins of IAA33. Through this assay, MITOGEN‐ACTIVATED PROTEIN KINASE 14 (MPK14) was identified to interact with IAA33 (Fig 6B). The interaction between MPK14 and IAA33 was also confirmed by co‐immunoprecipitation (Co‐IP) assays using proteins extracted from Arabidopsis mesophyll protoplast transiently expressing different construct combinations (IAA33‐MYC/MPK14‐YFP, MPK14‐YFP, IAA33‐MYC). IAA33‐MYC was successfully detected in the anti‐GFP immunoprecipitates of cells co‐expressing IAA33‐MYC and MPK14‐YFP (Fig 6A). Additional evidence that IAA33 interacts with MPK14 came from a bimolecular fluorescence complementation (BiFC) assay. A strong fluorescence signal was observed in tobacco epidermal cells co‐expressing IAA33 fused to the N‐terminal half of YFP (^NYFP) and MPK14 fused to the C‐terminal half of YFP (^CYFP), whereas no signal was observed in the empty vector control (Fig 6C). Last, in vitro pull‐down assays also confirmed the interaction between IAA33 and MPK14, since MPK14‐MBP was pulled down with IAA33‐GST (Fig 6D). All these results demonstrate that the IAA33 protein interacts with MPK14 in vitro and in vivo.

Figure 6. IAA33 interacts with MPK14.

1. In vivo Co‐IP assays of IAA33 with MPK14. IAA33‐MYC was co‐expressed with MPK14‐YFP in Arabidopsis mesophyll protoplast. Protein extracts (Input) were immunoprecipitated with anti‐GFP antibody (IP). Immunoblots were developed with anti‐GFP antibody to detect MPK14 and with anti‐MYC to detect IAA33.
2. IAA33 specifically interacts with MPK14 in yeast. IAA33 was used as bait, and MPK14 was used as prey. Empty‐AD was co‐transformed as negative control.
3. BiFC analysis of interaction between IAA33 and MPK14. The split YFP system was used in BiFC assays. ^CYFP and ^NYFP are empty vectors. The different combinations of plasmids were transformed into tobacco epidermal cells, and the YFP signals were detected with a confocal microscope. Scale bars, 50 μm.
4. Western blotting results of the GST pull‐down assay of IAA33‐GST and MPK14‐MBP.

Source data are available online for this figure.

Auxin‐induced IAA33 accumulation is mediated through MPK14

The interaction between MPK14 and IAA33 suggested that auxin‐induced IAA33 accumulation might be associated with its phosphorylation status. Indeed, we detected an increased phosphorylation of IAA33 with auxin treatment through the in vivo kinase assay using an anti‐phospho‐threonine antibody (Fig 7A). Consistently, the phosphorylation activity of MPK14 was also found to be activated, which was shown via a time‐course treatment with auxin (Fig 7B). To further study whether auxin‐activated MPK14 kinase activity is involved in IAA33 protein phosphorylation and stabilization, we generated kinase‐dead (MPK14^AA‐YFP) and kinase‐activated (MPK14^DD‐YFP) versions of MPK14 to examine their effect on IAA33 when both were transiently expressed in Arabidopsis mesophyll protoplast (Fig 7C). Our results showed that the phosphorylation level of IAA33 was reduced upon co‐expression with MPK14^AA, but increased upon co‐expression with MPK14^DD compared to the WT control (Fig 7D). To further examine whether MPK14‐regulated IAA33 phosphorylation could influence its protein stability, we expressed different variants of MPK14 in mesophyll protoplast cells extracted from 35S::GFP‐IAA33 seedlings and examined the effect on IAA33 protein stability through Western blot analysis. Our results showed that the IAA33 protein levels were increased upon expression of MPK14^DD compared to the WT control (Fig 7E). Accordingly, auxin‐induced IAA33 protein accumulation was reduced in an mpk14 mutant compared with that in Col‐0 (Fig 7F), indicating that the accumulation of IAA33 induced by auxin was regulated by MPK14. In order to test whether MPK14 regulates auxin signaling, we performed a transient expression assay in Arabidopsis protoplasts and found that expression of MPK14 had a promoting effect on transcriptional auxin response (DR5::LUC) in wild type. Consistent with the results that MPK14 stabilized IAA33, which interacts with repressor ARFs such as ARF10/16, this promoting effect was enhanced in the iaa33 mutant (Appendix Fig S12). In addition, we also checked the root stem cell identity phenotype and found that, similar to iaa33, the mpk1/14 double mutant exhibited a reduced DSC differentiation phenotype (Appendix Fig S13). Taken together, auxin‐induced IAA33 accumulation is mediated through MPK14.

Figure 7. Auxin‐mediated IAA33 protein accumulation is regulated by MPK14.

1. Phosphorylation assay of IAA33 in planta under auxin treatment. Total proteins were extracted from the IAA33‐MYC transient overexpressing in Arabidopsis mesophyll protoplast treated with or without NAA (10 μM) for 12 h and were immunoprecipitated with anti‐MYC agarose beads and separated on SDS–PAGE. IAA33‐MYC as equal loading controls was detected with anti‐MYC antibody. The relative intensity of band detected by anti‐phospho‐threonine antibody to that anti‐MYC antibody without NAA treatment was set to 1.0.
2. Phosphorylation assay of MPK14 in planta under auxin treatment. Total proteins were extracted from the 35S::MPK14‐YFP seedlings treated with NAA (10 μM) at indicated time points and separated on SDS–PAGE. The relative intensity of band detected by anti‐phospho‐MAPK antibody to that anti‐actin antibody without NAA treatment was set to 1.0.
3. Diagram of MPK14 protein. Mutations to generate the kinase‐dead (MPK14^AA) and constitutively active (MPK14^DD) MPK14 variants are indicated.
4. Phosphorylation assay of IAA33 in planta by MPK14. Total proteins were extracted from Arabidopsis mesophyll protoplast cells, which were transiently expressed IAA33‐MYC and different variants of MPK14‐YFP. The proteins were immunoprecipitated with anti‐MYC agarose beads and separated on SDS–PAGE. The relative intensity of band detected by anti‐phospho‐threonine antibody without MPK14 translation was set to 1.0.
5. Immunoblot analysis of IAA33 treated with MPK14. Different variants of MPK14‐YFP were transiently overexpressed in 35S::GFP‐IAA33 mesophyll protoplast cells. The relative intensity of IAA33 band to that anti‐Rubisco antibody without MPK14 translation was set to 1.0.
6. Immunoblot analysis of IAA33 when treated with or without NAA in Col‐0 or mpk14 mutant background. Total proteins were extracted from the IAA33‐MYC transient overexpressing in Arabidopsis mesophyll protoplast treated with or without NAA (10 μM) for 12 h and separated on SDS–PAGE. The relative intensity of IAA33 band to that anti‐actin antibody without NAA treatment was set to 1.0.

Source data are available online for this figure.

Discussion

The AUX/IAA protein family plays a central role in auxin signaling through interactions with AUXIN RESPONSE FACTORs (ARFs) (Parry & Estelle, 2006; Mockaitis & Estelle, 2008; Vernoux et al, 2011; Boer et al, 2014; Korasick et al, 2014; Guilfoyle, 2015). The canonical AUX/IAA proteins have four highly conserved domains (domains I‐IV), which underlie the functional properties of these proteins (Tiwari et al, 2004; Guilfoyle, 2015). Domain III/IV in AUX/IAA proteins mediates protein interactions between AUX/IAA and ARFs or other AUX/IAA proteins (Korasick et al, 2014; Nanao et al, 2014). In the absence of auxin, AUX/IAA proteins interact with ARFs through domain III/IV and repress ARF transcriptional activities through domain I. However, in the presence of auxin, the canonical AUX/IAA proteins are recognized by TIR1 through domain II and ubiquitinated and degraded via the 26S proteasome, leading to derepression of AUXIN RESPONSE FACTORs (ARFs) and transduction of auxin signaling (Dreher et al, 2006; Chapman & Estelle, 2009; Korasick et al, 2015; Salehin et al, 2015). However, in addition to canonical AUX/IAA proteins, there are also six non‐canonical Arabidopsis AUX/IAA proteins, such as IAA20, IAA30, IAA31, IAA32, IAA33, and IAA34, which have no typical domains I and II (Appendix Fig S1). Whether these non‐canonical AUX/IAA proteins regulate auxin signaling has long been an outstanding question. In this study, we found that, though IAA33 has no typical domains I and II, it negatively regulates auxin signaling. Recently, Cao et al reported that the non‐canonical IAA32/34 can also repress the transcription of genes, which contained auxin response elements although IAA32/34 also contained no typical domains I and II (Cao et al, 2019). This may occur through a similar mechanism as IAA33, since IAA32/34 also preferred to bind to repressor ARFs other than activator ARFs and protected the activity of these repressor ARFs (Piya et al, 2014). However, whether IAA32/34 regulate auxin response through competing with canonical AUX/IAA proteins for ARF binding requires further studies in the future.

Gilkerson et al (2015) mutated all 16 lysines in Arabidopsis IAA1, which eliminated all possible canonical ubiquitin acceptor sites, and found that the protein of IAA1 was still unstable and ubiquitinated. They showed that auxin‐dependent IAA1 ubiquitination may involve non‐canonical oxyester bonds to serines, threonines, or cysteines, which can be changed by posttranslational modifications (phosphorylation or disulfide bridges). To support this, in vitro phosphorylation of AUX/IAA proteins by phytochrome was detected (Colon‐Carmona et al, 2000). Recently, Xu et al, reported that light inhibited auxin signaling through stabilization of AUX/IAA proteins by interactions of cryptochrome 1 and phytochrome B with AUX/IAA proteins, respectively (Xu et al, 2018). Another study also reported that shade‐accumulated phytochrome A competed with TIR1 by directly binding and stabilizing AUX/IAA proteins to weaken auxin signaling and negatively regulate shade response (Yang et al, 2018). Several studies suggest that MAP kinases play important roles in auxin signaling. For example, auxin has been shown to induce the activity of MAP kinase in root of Arabidopsis seedlings (Mockaitis & Howell, 2000), and the MAPKK inhibitor PD98059 changes the transcription of auxin‐responsive genes in rice (Zhao et al, 2013), suggesting a potential role for MAP kinases in mediating auxin signaling. The Arabidopsis Group C MAP kinases consist of MPK1, MPK2, MPK7, and MPK14 (Ichimura et al, 2002), and the roles in Arabidopsis have not been well delineated. A previous study showed that MPK2 displayed increased kinase activity in response to the synthetic auxin 2,4‐D (Mizoguchi et al, 1994). Together with our observation that auxin could activate MPK14 kinase activity (Fig 7B), this suggests that this group of MAP kinase might play an important role in auxin signaling pathway.

Based on our results, we propose the following model (Fig 8). In the absence of auxin, both IAA33 and IAA5 could interact with ARFs, with IAA33 mainly binding to repressor ARFs. However, auxin induces the degradation of canonical AUX/IAA proteins such as IAA5 through the 26S proteasome‐dependent ubiquitin and the accumulation of IAA33 through phosphorylation by MPK14. The auxin‐induced IAA33, which has no repression activity on ARFs because of the absence of domain I, could interact with repressor ARFs through competition with a canonical IAA5 and thus enhance the activity of repressor ARFs and regulate auxin signaling and control root stem cell identity. The auxin‐TIR1/AFB‐dependent transcriptional regulation of auxin‐induced genes still operates in parallel with the action of the auxin–MPK pathway, which stabilizes IAA33 proteins. Although auxin can reduce the transcriptional auxin response through the IAA33‐ARF10/16 module, the canonical IAAs and activator ARFs still play a major role in this situation. The overall auxin effect on auxin‐mediated gene expression is the combinatorial result of multiple levels of regulation (Fig 8). Non‐canonical AUX/IAA proteins, such as IAA33, contribute to this, likely fine‐tuning the auxin response in a cell‐specific manner. Especially in root tips, where auxin levels are higher, the non‐canonical IAA33 might prevent an excessive auxin response, which will affect root stem cell identity (Ding & Friml, 2010), through interaction with repressor ARFs such as ARF10/16.

Figure 8. A proposed model of IAA33‐mediated auxin signaling to control root stem cell identity.

When the auxin level is low, the IAA33 protein is degraded by a 26S proteasome‐dependent mechanism, which leads to a low level of IAA33 protein, and most of the ARFs are inhibited by IAA5 and other canonical AUX/IAA proteins. When the auxin level increases, auxin induces the phosphorylation activity of MPK14, which then phosphorylated IAA33 to lead to accumulation of IAA33; meanwhile, canonical AUX/IAA proteins are degraded through 26S proteasome‐dependent pathway. The interaction of IAA5 with repressor ARFs is competed by increased IAA33 proteins; meanwhile, activator ARFs were also de‐repressed because of the degradation of canonical AUX/IAA proteins. Therefore, the overall auxin effect on auxin‐mediated gene expression is the combinatorial result of multiple levels of regulation. The pink ARFs mean repressor ARFs, and brown ARFs represent activator ARFs. The blue IAAs represent canonical AUX/IAA proteins.

Furthermore, different from the up‐regulation of canonical expression of AUX/IAAs in response to auxin (Vernoux et al, 2011; Weijers & Wagner, 2016; Winkler et al, 2017), the transcription of IAA33 was not affected by auxin treatment (Fig 5A and B), indicating a distinct transcriptional regulation of non‐canonical AUX/IAAs in response to auxin. This study suggests that, different from the dual action of auxin on the transcription and protein stability of canonical AUX/IAAs, auxin displays a distinct effect on the non‐canonical IAA33. Though IAA20, IAA30, IAA31, IAA32, IAA33, and IAA34 all belong to non‐canonical AUX/IAA proteins in Arabidopsis, it seems that the lack of domain II in IAA20, IAA30, IAA31, IAA32, and IAA34 might be an independent loss in their respective lineages in the core angiosperms (Mutte et al, 2018). It will be interesting to determine whether auxin has a similar effect on other non‐canonical AUX/IAA proteins. In addition, it was recently shown that TMK1 phosphorylates and stabilizes IAA32/34, which is similar to the here proposed MPK14‐IAA33 regulation, but TMK1 did not interact with IAA33 (Cao et al, 2019). It will thus be interesting to investigate the likely interaction between TMK1 and MPK14 and further study the molecular mechanism of auxin‐activated MPK14. Furthermore, though different kinases such as phyA, phyB, CRY1, TMK1, and MPK14 could phosphorylate AUX/IAA proteins and regulate the protein stability and auxin signaling (Xu et al, 2018; Yang et al, 2018; Cao et al, 2019; Fig 8), whether these different regulation modules are tissue‐ or cellular‐dependent processes is still an open question for future studies.

Materials and Methods

Plant materials and growing conditions

All of the A. thaliana mutants and/or transgenic lines utilized are in a Col‐0 background; the following have been described elsewhere: arf10, arf16, arf10/16 (Wang et al, 2005), DR5rev::GFP (Friml et al, 2003), IAA32p::GUS, IAA34p::GUS (Cao et al, 2019), iaa33 (CS876160), iaa5 (CS9578) (Overvoorde et al, 2005), and mpk1 (SALK_063847), mpk14 (SALK_022928) (Kohorn et al, 2014), which were obtained from the Arabidopsis Biological Resource Center (ABRC, Columbus, OH, USA). The mpk1/14 double mutant was obtained by crossing of mpk1 and mpk14. The 2411‐bp upstream region from the IAA33 start codon was amplified using I‐5™ 2× High‐Fidelity Master Mix (Beijing TsingKe Biotech Co., Ltd, catalog number TP001) and linked to the GUS‐GFP reporter in a gateway vector pKGWFS7.1 to obtain IAA33p::GUS‐GFP reporter construct. The 906‐bp upstream region from the IAA5 start codon was amplified using I‐5™ 2× High‐Fidelity Master Mix (Beijing TsingKe Biotech Co., Ltd, catalog number TP001) and linked to the GUS reported in gateway vector pKGWFS7.1 to obtain IAA5p::GUS reporter construct. The IAA33 ^OE and IAA5 ^OE constructs were constructed by introducing the IAA33 and IAA5 coding region (CDS), respectively, into the binary vector pK7WG2.0 and subjected to the plant transformations. The 35S::GFP‐IAA33 and 35S::GFP‐IAA5 constructs were constructed by introducing the IAA33 and IAA5 coding region (CDS), respectively, into the binary vector pB7WGF2.0 and subjected to the plant transformations. The ARF10 ^OE and ARF16 ^OE constructs were constructed by introducing the ARF10 and ARF16 CDS, respectively, into the binary vector pGWB18 and subjected to the plant transformations. Prior to germination, seeds were surface‐sterilized by fumigation in chlorine gas, held for 2 days at 4°C on solidified half‐strength Murashige and Skoog (MS) medium, and then transferred to a growth room providing a 16‐h photoperiod and a constant temperature of 20°C. All primers used for the generation of the constructs are listed in Appendix Table S1.

Microscopy and histochemical GUS staining

Root tips were imaged by an LSM‐700 laser‐scanning confocal microscopy (Zeiss). The process was performed according to the method described by Ding et al (Ding & Friml, 2010). Histochemical GUS staining was performed according to the method described by Gonzalez‐Garcia et al (2011).

Yeast two‐hybrid assay

Yeast two‐hybrid assay was performed according to the Matchmaker GAL4 Two‐Hybrid System 3 manufacturer's manual (Clontech). The coding sequences of ARF10 and ARF16 were inserted separately into the EcoRI‐XhoI cloning site of prey plasmid pGADT7 (Takara, USA), while the coding sequences of IAA5 and IAA33 were inserted into the cloning site of the bait plasmid pGBKT7, separately. Each of the constructs (including an empty vector for control purposes) was transferred separately into yeast Y2HGold using the PEG/LiAc method. After culturing on synthetic medium plates (SD medium) lacking Trp and Leu (SD‐Trp‐Leu) for 2 days, the transformants were transferred onto SD‐Trp‐Leu‐His or SD‐Trp‐Leu‐His‐Ade containing AbA (Aureobasidin A) and X‐α‐gal (5‐bromo‐4‐chloro‐3‐indolyl‐b‐d‐galactopyranoside) for blue color development. Primers used for yeast two‐hybrid assays are listed in Appendix Table S1.

Yeast three‐hybrid assay

Yeast three‐hybrid analysis was performed according to the previously described methods (Licitra & Liu, 1996). The constructs expressing ARF10 or ARF16 and bridge protein IAA5 or IAA33 were generated. The yeast strain Y2HGold was transformed with a pair of plasmids, pBridge‐ARF10‐IAA33 or pBridge‐ARF16‐IAA33 and IAA5‐pGADT7, pBridge‐ARF10‐IAA5 or pBridge‐ARF16‐IAA5 and IAA33‐pGADT7. pBridge‐ARF10 or pBridge‐ARF16 and IAA5‐pGADT7 or IAA33‐pGADT7 were used as positive control. The pBridge‐ARF10 or pBridge‐ARF16 and pGADT7, pBridge and IAA5‐pGADT7 or IAA33‐pGADT7 were used as negative control. Double transformants were selected on dropout media (SD‐Leu‐Trp). Protein interactions were confirmed by growth on selective media (SD‐Met‐His‐Leu‐Trp).

BiFC assays

To generate the constructs for the BiFC assays, full‐length cDNA fragments of IAA5, IAA33, ARF10, and ARF16 were amplified and cloned into the pDONRZeo (Invitrogen) vector for fusion with the N‐terminus of YFP or the C‐terminus of YFP by LR reaction. The recombinant constructs were introduced into the Agrobacterium tumefaciens strain EHA105, which were then injected into 4‐week‐old Nicotiana benthamiana leaves. The transfected leaves were grown at 25°C for 3 days before observation. The tobacco epidermal cells were then imaged under Zeiss LSM 700 confocal microscope. About forty cells of three biological repetitions were analyzed in each combination. The combination with the empty vector was the negative control.

Pull‐down assays

The purified GST‐fused protein (IAA33‐GST) or MBP‐fused protein (IAA5‐MBP) and purified His‐fused proteins (ARF10‐His and ARF16‐His) and Glutathione Sepharose™ 4 Fast Flow (GE Healthcare, catalog number 17‐5132‐01) or Amylose Resin (Biolabs, catalog number E8021S) were simultaneously incubated overnight at 4°C. His‐fused free protein was used as the control. After washing the beads three times with 25 mM HEPES, pH 7.5, 150 mM NaCl, 1 mM DTT, the beads were boiled in SDS–PAGE gel loading buffer at 99°C for 10 min, and subsequently, Western blotting was performed using anti‐GST or anti‐MBP antibody and anti‐His antibody. Primers used for pull‐down are listed in Appendix Table S1.

Co‐immunoprecipitation and Western blotting

Cells were harvested and lysed in cell lysis buffer (10 mM Tris–HCl, pH 7.5, 150 mM NaCl, 0.5 mM EDTA, 0.5% NP‐40, and 1 mM PMSF) on ice for 30 min with pipetting every 10 min. Cell lysates were centrifuged, and the supernatant was incubated with MYC‐Trap magnetic agarose beads (ChromoTek, catalog number ytma‐20) at 4°C for 2 h. The beads were washed three times with dilution buffer (10 mM Tris–HCl, pH 7.5, 150 mM NaCl, 0.5 mM EDTA) and resuspended the beads in SDS loading buffer. The resuspended beads were boiled for 10 min at 99°C, and Western blotting was followed using anti‐MYC (ABclonal, catalog number AE010) or anti‐GFP (TransGen Biotech, catalog number HT801‐02) antibody. For Western blotting, proteins from cell lysates were denatured and subjected to SDS–polyacrylamide gel electrophoresis (Bio‐Rad) and transferred to PVDF membranes (Millipore, catalog number IPVH00010). The membranes were blocked in 1× TBST with 5% milk for 2 h, immunoblotted with indicated antibodies at 4°C overnight, followed by incubation for 2 h with horseradish peroxidase‐conjugated secondary antibodies at room temperature. Blots were visualized by SuperSignal West Pico Luminol/Enhancer Solution (Thermo Fisher Scientific).

Quantitative real‐time PCR and transient expression assays

For qRT–PCR, the RNA template required was isolated using an RNeasy Plant Mini Kit (Qiagen, Hilden, Germany) following the manufacturer's protocol. Using DNase I to remove contaminating genomic DNA, a 2 μg aliquot was reverse‐transcribed using a Transcriptor First Strand cDNA Synthesis Kit (Roche, Basel, Switzerland), following the manufacturer's protocol. The subsequent qRT–PCRs were run on a MyiQTM Real‐time PCR Detection System (Bio‐Rad, Hercules, CA, USA) using FastStart Universal SYBR Green Master Mix (Roche, Basel, Switzerland). Each sample was represented by three biological replicates, and each biological replicate by three technical replicates. The reference sequence was AtACTIN2 (At3 g18780). Primers used for qRT–PCR are listed in Appendix Table S1.

For transient expression assays, the IAA5 or IAA33 coding sequences were amplified and the resulting sequences introduced into pBI221 to place them under the control of the CaMV 35S promoter. The promoter sequence of DR5 was amplified and introduced into the pGreenII0800‐LUC reporter vector. Both recombinant plasmids were then transferred into the protoplasts of A. thaliana. Firefly luciferase (LUC) and Renilla luciferase (REN) activities were measured using the Dual‐Luciferase Reporter Assay System (http://www.promega.com). LUC activity was normalized against REN activity (Hellens et al, 2005).

Phosphorylation assays

To examine the phosphorylation levels of IAA33 during auxin treatment or different variants of MPK14, extracted proteins from these treated protoplasts were mixed with MYC‐Trap magnetic agarose beads (ChromoTek, catalog number ytma‐20) at 4°C for 2 h. The beads were washed three times with dilution buffer (10 mM Tris–HCl, pH 7.5, 150 mM NaCl, 0.5 mM EDTA) and resuspended the beads in SDS loading buffer. The resuspended beads were boiled for 10 min at 99°C, and Western blotting was followed using anti‐phospho‐threonine (anti‐Phospho‐Thr) (Cell Signaling, catalog number 9381) and anti‐MYC (Abclonal, catalog number AE010) antibodies. To examine the phosphorylation activity of MPK14 during auxin treatment, total proteins from 35S::MPK14‐YFP seedlings treated with auxin for different time points were isolated with the extraction buffer and separated on SDS–PAGE. Protein levels and phosphorylation of MPK14 were detected with anti‐GFP (TransGen Biotech, catalog number HT801‐02) or anti‐phospho‐MAPK (Cell Signaling, catalog number 9101) antibody, respectively.

Statistical analysis

Statistical analysis was performed by one‐way ANOVA using the statistical software SPSS 17.0 (SPSS, Chicago, IL). Based on the ANOVA results, Duncan's test for mean comparison was performed. All data were represented by means and standard error.

Author contributions

ZD designed the study. BL, QY, and JL performed the experiments. BL, QY, JL, XW, ZY, KH, HL, XK, CL, HT, IDS, X‐SZ, and ZD analyzed the data. BL and ZD wrote the paper.

Conflict of interest

The authors declare that they have no conflict of interest.

Supporting information

Appendix

Source Data for Appendix

Review Process File

Source Data for Figure 3

Source Data for Figure 4

Source Data for Figure 5

Source Data for Figure 6

Source Data for Figure 7

The EMBO Journal (2020) 39: e101515