CUL3^BPM E3 ubiquitin ligases regulate MYC2, MYC3, and MYC4 stability and JA responses

Significance

Jasmonates (JAs) are essential phytohormones regulating a myriad of developmental and defensive responses in plants. MYC2, MYC3, and MYC4 are key transcriptional activators of JA-mediated gene expression and have become relevant signaling hubs. Repression of MYC activity is necessary for resetting JA signaling and to avoid harmful runaway responses, which contribute to plant fitness. Here, we identified a mechanism to reduce MYC protein levels by E3 ubiquitin ligases based on Cullin3 and BPM proteins as substrate adaptors (CUL3^BPM). BPM3 stability is enhanced by JA, establishing a negative feedback regulatory loop to control MYC levels and activity. Our results uncover a new layer of JA-pathway regulation that terminates MYC activity by CUL3^BPM-mediated degradation of MYC TFs.

Abstract

The jasmonate (JA)-pathway regulators MYC2, MYC3, and MYC4 are central nodes in plant signaling networks integrating environmental and developmental signals to fine-tune JA defenses and plant growth. Continuous activation of MYC activity is potentially lethal. Hence, MYCs need to be tightly regulated in order to optimize plant fitness. Among the increasing number of mechanisms regulating MYC activity, protein stability is arising as a major player. However, how the levels of MYC proteins are modulated is still poorly understood. Here, we report that MYC2, MYC3, and MYC4 are targets of BPM (BTB/POZ-MATH) proteins, which act as substrate adaptors of CUL3-based E3 ubiquitin ligases. Reduction of function of CUL3^BPM in amiR-bpm lines, bpm235 triple mutants, and cul3ab double mutants enhances MYC2 and MYC3 stability and accumulation and potentiates plant responses to JA such as root-growth inhibition and MYC-regulated gene expression. Moreover, MYC3 polyubiquitination levels are reduced in amiR-bpm lines. BPM3 protein is stabilized by JA, suggesting a negative feedback regulatory mechanism to control MYC activity, avoiding harmful runaway responses. Our results uncover a layer for JA-pathway regulation by CUL3^BPM-mediated degradation of MYC transcription factors.

Jasmonates (JAs) are oxygenated lipid derivatives (oxylipins) synthesized through the octadecanoid and hexadecatrienoic pathways and chemically similar to prostaglandins in animals (1, 2). JAs are essential phytohormones for plant development and environmental adaptation, since they 1) are key regulators of responses to biotic and abiotic stresses, 2) coordinate several developmental processes such as root growth and fertility, and 3) fine-tune competitive growth–defense tradeoffs that optimize plant fitness in response to resource limitations (1, 3).

Upon stress or endogenous stimuli, plants accumulate the bioactive JA (+)-7-iso-JA-Ile (JA-Ile) (4, 5), which is perceived by a coreceptor complex formed by the F-box protein CORONATINE INSENSITIVE 1 (COI1) and a JASMONATE-ZIM (JAZ) domain protein (6–8). COI1 is the F-box subunit of the SCF^COI1 (Skip1-Cullin1-F box) E3 ubiquitin ligase (9, 10). JAZs act as nuclear repressors of an increasing number of transcription factors (TFs) belonging to different families such as bHLH, MYB, YABBY, and EIN3/EIL (1, 11–15). The JAZ proteins exert their repressor activity by a double mechanism involving the recruitment of the general corepressor TOPLESS (TPL) by the adaptor protein NOVEL INTERACTOR OF JAZ (NINJA) (16) and through competition with the MEDIATOR25 (MED25) component of the general transcriptional activation machinery for interaction with MYCs (17). Hormone-triggered interaction of COI1 and JAZs leads to JAZ ubiquitination and degradation by the ubiquitin-proteasome system (UPS) (6, 7, 18, 19). Degradation of JAZs releases their TF targets that, in turn, induce a substantial transcriptional reprogramming (20–24).

Among JAZ targets, MYC2, MYC3, and MYC4 are key transcriptional regulators of JA-mediated gene expression that belong to the bHLH family of TFs (11, 25). JA-dependent transcriptional activation by MYCs requires recruitment of MED25, a subunit of the MEDIATOR transcriptional coactivator complex (26). Both MYC2 and MED25 regulate a dynamic chromatin looping between JA enhancers and their promoters (27). Continuous activation of JA responses by MYCs inhibits growth and is potentially harmful or even lethal for the cell. Thus, many mechanisms for resetting JA signaling and terminating MYC activity have been described, including the MYC-dependent expression of JAZ and bHLH repressors, alternative splicing forms of JAZ, etc. (28, 29). Recently, even a function of MYC2 in regulating the termination of JA signaling through activation of a small group of bHLH repressors that impair the formation of the MYC2–MED25 complex has been reported (30). Another repression mechanism includes JAV1 [JASMONATE-ASSOCIATED VQ-MOTIF GENE1 (14)], a repressor of JA-mediated defenses that is degraded after herbivory by JAV1-ASSOCIATED UBIQUITIN LIGASE1 [JUL1 (31)].

In addition to JA signaling and defense, MYCs have been involved in the regulation of other processes such as responses to abscisic acid (ABA), ethylene, or blue light (32–35). Therefore, MYCs are currently considered a node of convergence of several pathways, whose activity needs to be tightly regulated to optimize plant fitness. MYCs are short-lived proteins and their transcriptional activity requires phosphorylation, which also triggers MYC proteolysis (36, 37). Besides phosphorylation, MYC protein stability is regulated by JA, light quality, and the circadian clock, suggesting that modulation of protein levels is a major regulatory mechanism of MYC protein activity (36–39).

Several regulators of MYC protein stability have been described. For instance, the E3 ubiquitin ligase PUB10 targets MYC2 for proteasomal degradation (40). PUB10 affects JA-related gene expression and phenotypic responses, but its involvement in modulation of MYC2 protein levels by environmental cues has not been reported so far. Similarly, COP1 promotes MYC degradation in the dark, whereas photoreceptors counteract this activity in light (37). However, the importance of MYCs as nodes of convergence of signaling networks and activators of responses that are potentially harmful for the cell suggests that MYC activity has to be tightly regulated and that other regulatory mechanisms of MYC stability could be expected.

The BPM (BTB/POZ-MATH) proteins are adaptors of Cullin3-based E3 ubiquitin ligases in animals and plants, and form a small family of six members in Arabidopsis [AtBPM1 (At5g19000), AtBPM2 (At3g06190), AtBPM3 (At2g39760), AtBPM4 (At3g03740), AtBPM5 (At5g21010), and AtBPM6 (At3g43700) (41–45)]. They are characterized by the presence of two protein–protein interaction domains: BTB/POZ (broad complex, tramtrack, bric-a-brac/POX virus, and zinc finger domain) and MATH (Meprin and TRAF homology domain). The BTB/POZ domain mediates assembly of BTB/POZ proteins with CUL3a and CUL3b in plants and animals. The MATH domain determines the interaction of BTB/POZ-MATH proteins with their substrates (10, 41, 46–52).

CUL3^BPM E3 ligases are involved in the regulation of several physiological processes such as plant growth, fertility, stomatal dynamics, fatty acid metabolism, and ABA signaling, but their targets are still scarce (43, 44, 53).

Here, we identified MYC2, MYC3, and MYC4 as interactors of BPM proteins in vivo and show that CUL3^BPM E3 ligases target MYC2 and MYC3 for UPS-mediated degradation. Indeed, reduction of function of CUL3^BPM (amiR-bpm) reduces ubiquitination levels of MYC3, enhances MYC accumulation, and renders the plants more responsive to JA. Similarly, triple bpm235 mutants show JA-related phenotypes and a constitutive JA-regulated gene expression. Reduction of CUL3 function in the double mutant cul3ab also enhances JA responses, altogether indicating that CUL3^BPM regulates MYC activity. Interestingly, BPM3 protein is stabilized by JA, suggesting a negative feedback regulatory mechanism to control MYC levels and activity. Our results uncover a new tier of JA-pathway regulation by CUL3^BPM-mediated degradation of MYC TFs.

Results

MYC2, MYC3, and MYC4 Physically Interact with BPMs.

In order to identify new targets of BPMs, we performed a yeast two-hybrid (Y2H) screen using BPM3 as bait and an Arabidopsis complementary (c)DNA library prepared from Arabidopsis inflorescences. From over 2 million clones screened, 80 were subsequently confirmed by retransformation into yeast (43). Among these clones, we identified two independent prey clones corresponding to MYC2 (MYC2^28–292 and MYC2^187–509). These two clones delimit a small MYC2 region of interaction with BPM3 (MYC2^187–292), which is highly conserved within the MYC2 partners MYC3 and MYC4 (SI Appendix, Fig. S1). Y2H assays using all six BPMs as bait and the three MYCs as prey showed that MYC2 and MYC3 interact strongly with BPM1, BPM2, BPM3, and BPM4 and very weakly with BPM5 and BPM6 (Fig. 1A and SI Appendix, Fig. S2). In the case of MYC4, only a strong interaction with BPM2 could be observed. However, due to the toxicity of MYC expression in yeast, these interactions likely represent only a subset of all possible interactions in vivo.

Fig. 1.

MYC2, MYC3, and MYC4 interact with several BPM proteins. (A) Yeast cells cotransformed with MYC2, MYC3, and MYC4 fused to the GAL4 activation domain (prey) and with BPM1 to BPM6 fused to the GAL4–DNA-binding domain (bait) were selected and subsequently grown on yeast synthetic dropout lacking Leu and Trp (−2) as a transformation control (shown in SI Appendix, Fig. S2), or on selective media lacking Ade, His, Leu, and Trp (−4) to test protein interactions. One-third and 1/10 dilutions were included. pGADT7-MYC2, pGADT7-MYC3, and pGADT7-MYC4 cotransformations with pGBKT7 empty vectors were included as controls. (B and C) Immunoblots with anti-HA antibody of recovered HA-BPM3 (B) or with anti-myc antibody of recovered myc-BPM6 (C) after pull-down reactions using crude protein extracts from 35S:HA-BPM3 (+; B), 35S:myc-BPM6 (+; C), or Col-0 (−) Arabidopsis plants and resin-bound MBP or MBP-fused MYC and BPM3 proteins. Coomassie blue staining shows the amount of recombinant proteins used (Bottom). Asterisks show an unspecific band in MYC3 (B) and MYC4 (C). (D) Graphical representation obtained from the proteomic analysis, depicting the GFP-BPM6–enriched proteins. Proteins are ranked in a volcano plot according to their relative abundance ratio (BPM6/control) and the statistical analysis was performed by a negative binomial regression (adjusted P values). For each protein, the average number of spectra from the control samples (n = 3) was compared with the average number of spectra from the GFP-BPM6 samples (n = 3). The vertical red dashed lines display large-magnitude fold changes (x axis, FC < 0.5, Left; FC > 2, Right) whereas the horizontal red dashed lines display high statistical significance (y axis, adjusted P < 0.05 above the line). Green dots represent enriched BPM proteins exhibiting significant fold changes, whereas red dots represent enriched proteins of interest, among which are MYC proteins. Proteins associated with the volcano plot representation are listed in SI Appendix, Table S1 and Dataset S1.

To further test BPM–MYC interactions, we conducted pull-down experiments using maltose-binding protein (MBP)–MYC protein fusions bound to amylose resin and protein extracts from transgenic plants expressing hemagglutinin (HA)-BPM3 or myc-BPM6 from the 35S CaMV promoter (SI Appendix, Fig. S3). As shown in Fig. 1 B and C, similar to the positive control MBP-BPM3 that homodimerizes with HA-BPM3, all three MBP–MYC fusions pulled down both HA-BPM3 and myc-BPM6 from these extracts.

To find out if BPM–MYC interaction occurs in vivo, we attempted to identify proteins bound to BPM6 after immunoprecipitation of green fluorescent protein (GFP)-BPM6 from transgenic 35S:GFP-BPM6 lines (Materials and Methods) treated or not with MeJA by mass spectrometry analysis. As shown in Fig. 1D, SI Appendix, Table S1, and Dataset S1, several BPMs (BPM1, BPM4, and BPM5) were coimmunopurified by GFP-BPM6, both in the presence and absence of MeJA treatment, supporting the previous observation that BPMs can form heterodimers in vivo (41, 43). More strikingly, these assays identified MYC2, MYC3, MYC4, and JAZ1 both in untreated and JA-treated plants and NINJA in untreated plants, but not in 35S:GFP controls (Fig. 1D, SI Appendix, Table S1, and Dataset S1). These data suggested that MYCs are direct interactors of BPMs in vivo. Although MeJA might potentiate the interaction, particularly in the case of MYC2 and MYC4, it also occurs in the absence of MeJA treatment.

BPMs Regulate MYC Stability.

Protein–protein interaction results described above suggested that MYC2, MYC3, and MYC4 could be targets of CUL3^BPM E3 ubiquitin ligases. Therefore, we next introgressed MYC2-GFP from two independent 35S:MYC2-GFP lines into a previously described amiRNA knockdown (KD) BPM line (amiR-bpm), where transcripts for several BPM family members are significantly reduced (43). After confirmation that, similar to the original lines, the crossed lines maintained a lower expression level of BPM1, BPM4, BPM5, and BPM6 (SI Appendix, Fig. S4), we analyzed the effect of BPM reduction on MYC2 protein levels. As shown in Fig. 2A, MYC2 accumulated more in the mutant amiR-bpm than in wild-type (WT) plants, despite MYC2-GFP transgenic expression being similar or even slightly lower in amiR-bpm (Fig. 2A and SI Appendix, Fig. S5). Since MYC2 is constitutively expressed from the 35S promoter in the transgenic lines, we also analyzed its degradation rate after inhibition of protein synthesis by cycloheximide (CHX) (Fig. 2B). In WT plants, MYC2 is a short-lived protein with a rapid turnover that almost disappeared after 3 h. In contrast, MYC2 degradation was much slower in amiR-bpm, with higher levels remaining at all time points.

Fig. 2.

MYC2 and MYC3 accumulate in the amiR-bpm background. (A and B) Immunoblot analyses of MYC2-GFP and actin protein levels in 35S:MYC2-GFP lines in WT or amiR-bpm (ami) backgrounds. “#1” and “#2” indicate two independent 35S:MYC2-GFP transgenic lines. Loading of total protein extracts in A is indicated by 1 (18 μL) and 1/3 (6 μL). To monitor MYC2-GFP protein decay in B, seedlings were treated with 50 μM CHX and protein levels were analyzed at the indicated times (h). (C) Immunoblot analysis of MYC2-GFP and actin protein levels in wild-type and amiR-bpm backgrounds. To check MYC2-GFP protein decline, seedlings were treated with 50 μM CHX and protein levels were analyzed at the indicated times (h). Expression of MYC2-GFP under its natural promoter was assessed using the MYC2 C5025 recombineering line. (D) Immunoblot analysis of MYC3-HA and actin protein levels in 35S:MYC3-HA lines in WT and amiR-bpm backgrounds. #1 and #2 indicate two independent 35S:MYC3-HA transgenic lines. To check MYC3-HA protein levels, seedlings were treated with 50 μM CHX and protein levels were analyzed at the indicated times (h). Protein molecular mass is shown (Right).

To confirm these results, we analyzed MYC2-GFP levels in a “recombineering” line in which MYC2-GFP is expressed under its natural promoter in its natural chromosomal environment [recombineering line MYC2 C5025 (54, 55)]. These lines were obtained by introducing the GFP marker by homologous recombination in bacteria using big chromosomal pieces containing the gene of interest and transforming the recombineered chromosomal piece back into transgenic plants (54). Similar to the 35S:MYC2-GFP line, the recombineering MYC2-GFP showed an enhanced stability after CHX treatment (Fig. 2C).

Finally, we also analyzed the effect of BPM reduction on MYC3-HA protein levels using the amiR-bpm background. As shown in Fig. 2D, results of MYC3 were similar to those of MYC2, indicating that MYC3 is also a target of CUL3^BPMs.

We have previously shown that MYC2 protein stability is reduced in the dark through COP1 activity (37). To test if BPMs could also have a role in this process, we analyzed MYC2 levels in WT and amiR-bpm plants under white light or after transfer to darkness for 24 h. Consistent with the regulation of this process by COP1, destabilization of MYC2 was similar in both lines, indicating that BPMs are not involved in light/dark regulation of MYC2 (SI Appendix, Fig. S6).

CUL3^BPMs Ubiquitinate MYC3.

To further confirm that elevated MYC protein levels in the amiR-bpm background correlate with protein ubiquitination, we analyzed whether a representative member of the family (MYC3) is ubiquitinated in vivo and whether ubiquitination levels are dependent on BPMs. We chose MYC3, since this protein accumulates to higher levels than other MYCs and is easier to detect. We used p62 resin columns to purify ubiquitinated proteins in WT and amiR-bpm lines expressing MYC3-HA. As shown in Fig. 3A, the MYC3 lanes showed a higher-size smear of polyubiquitinated bands when the plants were incubated with proteasome inhibitors overnight, demonstrating that MYC3 is ubiquitinated in vivo. Remarkably, after purification of ubiquitinated proteins with the p62 resin, the polyubiquitinated-protein smear was lower in the amiR-bpm background compared with the WT. Quantification of ubiquitinated levels of MYC3-HA in WT and amiR-bpm lines in three independent experiments showed a consistent reduction of over 40% in all three experiments (Fig. 3B). Considering that amiR-bpm lines are only a reduction of BPM function, these results strongly support a major role of CUL3^BPMs in ubiquitination of MYC3.

Fig. 3.

MYC3 polyubiquitination levels are dependent on BPMs. (A) Affinity purification of polyubiquitinated MYC3-HA was carried out from total protein extracts obtained from MYC3-HA seedlings in WT (M3wt) and amiR-bpm (M3ami) backgrounds by incubation with Ub-binding p62 resin (+) or with empty agarose resin (negative control; −). Anti-Ub was used to detect total ubiquitinated proteins. Anti-HA allowed detection of MYC3-HA and its ubiquitinated forms (shown by brackets). To facilitate visualization and quantification of results, the amounts of M3ami and M3wt loaded into the gel were adjusted to obtain similar amounts of the main MYC3-HA band. PD, pull-down. (B) Protein-level analysis of samples described in A was carried out using ImageJ software. The ratio of polyubiquitinated MYC3-HA [Ub(n)-MYC3-HA] compared with unmodified MYC3-HA is shown for M3wt and M3ami samples. Results correspond to three independent experiments.

Knocking Down BPM Function Increases JA Responses.

Results described above suggest that BPMs could redundantly regulate MYC levels and, therefore, activity. To test this hypothesis, we analyzed typical JA responses in amiR-bpm and bpm mutants. We first attempted to obtain and characterize individual bpm mutants and combinations. We confirmed that available lines in stock centers for bpm2 (GK_391EO4), bpm4 (Salk_082761), and bpm5 (Salk_038471) were loss-of-function mutations. Line Salk_72848, however, seemed to be a reduction of function of bpm3 (SI Appendix, Fig. S7). Phenotypic analyses suggested functional redundancy among BPMs due to the lack of JA-related phenotypes in single mutants and some double combinations (bpm23, bpm25, bpm34, bpm35). However, triple bpm235 (SI Appendix, Fig. S8) showed a root-growth phenotype similar to that of amiR-bpm and 35S:MYC2 overexpressing lines, both in control conditions and after treatments with JA or COR (Fig. 4) (25). In control conditions, all three lines (35S:MYC2, bmp235, and amiR-bpm) have a slightly shorter root than the WT, indicative of a mild constitutive JA response. Quantification of basal JA levels in these plants showed that, similar to 35S:MYC2, amiR-bpm lines had increased levels of JA in control conditions, whereas bpm235 mutants had a higher mean JA level but not statistically different from the WT (SI Appendix, Fig. S9). This suggests that the root-growth phenotype in amiR-bpm may be the consequence of enhanced basal JA levels. However, this is unlikely the case with bpm235. Treatments with JA or coronatine (a JA-Ile mimic of bacterial origin) inhibited root growth in WT plants (Fig. 4 A–C) in a percentage similar to that of 35S:MYC2, bpm235, and amiR-bpm lines, whose final root length was shorter than WT plants. The shorter final root length in 35S:MYC2 lines is the result of a stronger response to JA due to elevated MYC2 levels. The similar behavior of bpm235 and amiR-bpm lines to that of 35S:MYC2 transgenics suggests a stronger JA response in these lines and is consistent with elevated levels of MYCs in bpm235 and amiR-bpm lines. Importantly, the effect of amiR-bpm on root-growth inhibition was partially suppressed in jin1-2, a MYC2 loss-of-function allele (25), which confirms that the shorter root in basal conditions is due to a constitutive response to JA of the mutant lines and is consistent with an effect of BPMs on the activity of several MYCs (Fig. 4D). Altogether, these results show that BPMs and MYC2 have opposite effects on root growth and suggest that MYCs are targets of CUL3^BPMs in vivo.

Fig. 4.

amiR-bpm and bpm235 show increased response to JA and COR. (A) Root-growth inhibition assay of 10-d-old Arabidopsis wild-type Col-0 seedlings, amiR-bpm line, bpm235 triple mutant, and 35S:MYC2 transgenic plants (OEMYC2) grown on Johnson’s media supplemented with JA or coronatine at the indicated concentrations. C, control. (Scale bar, 10 mm.) (B) Quantification of the root length of plants described in A treated with JA. (C) Quantification of the root length of plants described in A treated with COR. (D) Quantification of the root length of 8-d-old Arabidopsis wild-type Col-0 seedlings, amiR-bpm line, jin1-2, and two amiR-bpm lines (7.2 and 7.4) in the jin1-2 mutant background grown on Johnson’s media supplemented with 50 μM JA or 0.5 μM COR. Results shown in the graphs are the mean ± SD of measurements of 15 to 20 seedlings. Asterisks indicate statistically significant differences compared with WT (Student’s t test, *P < 0.01) and letters indicate statistically significant differences compared with WT and between different genotypes (Student’s t test, *P < 0.01).

To further support this hypothesis, we also analyzed JA responses in cul3 mutants. Since a complete loss of function of the two CUL3 genes (CUL3a and CUL3b) is lethal, we used a weak cul3ab allele (56). Additionally, CUL3 also participates in ethylene (ET) biosynthesis assembling CUL3^ETO1 E3 ligases that target ACS5 (48, 56, 57). Thus, cul3ab mutants accumulate ethylene and have a strong phenotype. To reduce this ET phenotype, we used a triple cul3ab,ein3 mutant and compared it with its ET-insensitive ein3 background. As shown in Fig. 5, the cul3ab,ein3 mutant had a stronger response to all JA or COR concentrations, compared with its ein3 background, indicating that the reduction of CUL3 activity in cul3 mutants also leads to enhanced JA response and behaves as the KD mutations in BPMs or overexpression of MYC2.

Fig. 5.

CUL3 knockdown increases responses to JA and COR. Root-growth inhibition assay of 10-d-old Arabidopsis wild-type Col-0 seedlings and ein3-1 and cul3ab ein3 mutants grown on Johnson’s media supplemented with JA (A) or COR (B) at the indicated concentrations. Results shown are the mean ± SD of measurements of 15 to 20 seedlings. Asterisks indicate statistically significant differences compared with ein3-1 (Student’s t test, *P < 0.01).

Knocking Down BPM Function Increases JA-Dependent Gene Expression.

Transcriptomic analyses of the amiR-bpm line and bpm235 mutants further supported the role of BPMs regulating MYC stability (Fig. 6). COR-treated amiR-bpm showed 121 genes differentially up-regulated and 77 genes differentially down-regulated compared with COR-treated wild-type plants (false discovery rate [FDR] < 0.01; fold change [FC] > 2; ref. 58). Among the differentially up-regulated genes, more than 27% (considering FDR < 0.01) or 43% (considering FDR < 0.001) are JA-regulated genes (up-regulated by MeJA after 0.5, 1, or 3 h according to the BAR database [http://bar.utoronto.ca]; Fig. 6 A and B). These observed values contrast with the expected number of JA-regulated genes by chance (1.3% considering FDR < 0.01 of BAR data; Fig. 6B). Moreover, gene ontology (GO) analysis of up-regulated genes in amiR-bpm showed a statistically significant overrepresentation of JA-related GO terms, such as JA response and wounding (Fig. 6C). These results indicate that reduction of BPM activity has a deep impact enhancing JA-regulated gene expression and support that amiR-bpm plants have a stronger response to JA. Under basal conditions, amiR-bpm plants differentially expressed 172 genes (112 up-regulated and 60 down-regulated) compared with WT (FDR < 0.01; FC > 2; SI Appendix, Fig. S10; ref. 58). In this case, the number of JA-regulated genes among those differentially regulated genes was lower but still significant (about 13.3% considering FDR < 0.01).

Fig. 6.

BPMs repress JA-mediated gene expression. (A) BAR data (FC ≥ 2 and FDR < 0.01) showing MeJA regulation of representative genes up-regulated in the transcriptomic profiling of amiR-bpm vs. Col-0 plants both treated with 0.5 μM COR for 6 h. (B) Percentage of expected (MeJA BAR) or observed (amiR-bpm) JA–up-regulated genes among those differentially expressed in the same transcriptomic analysis. MeJA BAR: % of genes expected to be induced by JA considering the whole-genome data of JA treatments (0.5, 1, and 3 h) in the BAR database using FC ≥2 and FDR (rank product) <0.01, <0.005, or 0.001. amiR-bpm: % of genes induced or repressed by MeJA treatment within a list of genes up-regulated by coronatine treatment. (C) GO enrichment analysis of up-regulated genes in amiR-bpm vs. WT using DAVID bioinformatics resources. P values were adjusted by the Benjamini–Hochberg method. JA-dependent biological processes are highlighted in red. (D) BAR data (FC ≥ 2 and FDR < 0.01) showing MeJA regulation of the 20% most up-regulated genes in the transcriptomic profiling of bpm235 vs. Columbia plants in basal conditions (untreated). (E) Percentage of expected (MeJA BAR) or observed (bpm235) JA–up-regulated genes among those differentially expressed in the same transcriptomic analysis. MeJA BAR: % of genes expected to be induced by JA considering the whole-genome data of JA treatments (0.5, 1, and 3 h) in the BAR database using FC ≥2 and FDR (rank product) <0.01, <0.005, or <0.001. bpm235: % of genes induced by MeJA treatment within the list of genes up-regulated in basal conditions in bpm235. (F) GO enrichment analysis of up-regulated genes in the triple mutant bpm235 vs. WT using DAVID bioinformatics resources. P values were adjusted by the Benjamini–Hochberg method. JA-dependent biological processes are highlighted in red. The columns in A and D correspond to the 0.5, 1, and 3 h time points of BAR data.

Results were even clearer in the case of the bpm235 mutant (Fig. 6 D–F). Under basal conditions, 531 genes were differentially expressed in bpm235 compared with the WT (500 up-regulated and 31 down-regulated; FDR < 0.01; FC > 2 or <−2; ref. 59). Among up-regulated genes, more than 44% (considering FDR < 0.01) or 50% (considering FDR < 0.001) are JA-response genes (induced by MeJA after 0.5, 1, or 3 h according to the BAR database; Fig. 6 D and E). Moreover, GO analysis of up-regulated genes in bpm235 showed a statistically significant overrepresentation of JA-related GO terms, such as JA response, JA signaling, wounding, and defense (Fig. 6F). Besides JA, other GO terms such as water transport, water deficit, or ABA were overrepresented in our GO analysis. This alteration in basal gene expression is consistent with BPMs regulating other pathways (in addition to JA), as previously reported (43, 45).

MYC-Dependent Gene Expression Is Up-Regulated in KD bpm Lines.

To further contrast if MYCs are targets of CUL3^BPMs, we next checked if genes up-regulated in amiR-bpm were targets of MYCs. Toward this, we obtained transcriptomic profiles of triple myc2myc3myc4 mutants compared with WT Col-0 plants after JA treatment for 6 h using Agilent microarrays. As shown in Fig. 7A, clustering analysis showed that genes differentially down-regulated in myc2myc3myc4 in response to JA are mostly up-regulated in the bmp235 and amiR-bpm backgrounds. Interestingly, the overlap in misregulated genes in bmp235 and amiR-bpm was relatively low (Fig. 7A and SI Appendix, Fig. S11), indicating that they affect complementary JA-regulated gene sets. This is consistent with the different genes knocked out (or down) in the two different mutants (bpm235 is a knockout mutant for BPM2 and 5 and a knockdown for BPM3, whereas amiR-bpm is a knockdown for BPM4 and 5 and a weak knockdown for BPM1 and 6). Therefore, these results suggest that different BPMs may have a complementary function in the regulation of particular JA responses.

Fig. 7.

MYC-dependent gene expression is up-regulated in KD bpm lines. (A) Hierarchical clustering analysis of genes down-regulated (log ratio <−1; FDR < 0.05) in the triple myc2,3,4 mutant in response to JA (myc234) and their expression in knockdown bpm lines (amiR-bpm) after COR treatment and in untreated triple bpm2,3,5 mutant plants. (B and C) The MYC binding sites (G box and/or G-box variants) are overrepresented in the promoters of genes up-regulated in amiR-bpm and bpm235. Percentage of expected (genome) or observed (amiR-bpm or bpm235) boxes (G box or variants) in differentially up-regulated genes in amiR-bpm or bpm235. Genome: % of genes expected to contain the G box or G-box variants (1 or 0.5 kb upstream of the TSS) considering the whole genome obtained with the dna-pattern tool in RSAT (http://www.rsat.eu). amiR-bpm and bpm235: % of genes containing the G box or G-box variants within the list of genes up-regulated in both genotypes. Asterisks indicate statistically significant differences compared with the genome (binomial, *P < 0.01).

MYC TFs recognize the G box (5′-CACGTG-3′) or G-box variants (5′-AACGTG-3′; 5′-CATGTG-3′) in their cognate promoters (60, 61). As shown in Fig. 7 B and C, analyses of the promoters (0.5 or 1 kb upstream of the transcriptional start site) of genes up-regulated in the amiR-bpm or bmp235 lines showed a significant overrepresentation of these motifs (G box and 5′-AACGTG-3′ in both lines and 5′-CATGTG-3′ only in bpm235). These results further support that up-regulation of these genes is mediated by an enhanced activity of MYCs in the KD bpm backgrounds.

JA Stabilizes BPM3.

Results described above suggest a mechanism to terminate MYC activity and reset JA signaling by CUL3^BPMs. Therefore, we next investigated if JA signaling regulates BPM activity. BPM genes do not seem transcriptionally regulated by JA (according to our own transcriptomic data, the BAR database, and the Genevestigator database, as shown in SI Appendix, Fig. S12). To check if JA could regulate BPMs posttranscriptionally, we analyzed myc-BPM3 and myc-BPM6 stability after JA treatment in time-course experiments. As shown in Fig. 8A, the amount of myc-BPM3 protein was very low in untreated conditions, despite the myc-BPM3 transgene being expressed constitutively under the strong 35S promoter. Interestingly, JA treatment greatly stabilized myc-BPM3 protein (Fig. 8A). Accumulation was visible after 1 h of treatment and increased to very high levels after 3 h. Subcellular fractionation of a different fusion protein (HA-BPM3) in the absence or presence of JA confirmed these results (JA stabilization) and showed that, consistent with previous reports (43), BPM3 is a nuclear protein (Fig. 8B). Simultaneous treatment of HA-BPM3 transgenic plants with JA and the COI1 specific inhibitor coronatine-O-methyloxime (COR-MO) (62) prevented the stabilizing effect of JA on HA-BPM3, indicating that this effect of JA is dependent on the COI1 receptor and, therefore, on the JA-signaling pathway (Fig. 8C).

Fig. 8.

JA induces BPM3 accumulation. (A) Immunoblot analysis (anti-myc antibody) of myc-BPM3 protein levels in 7-d-old 35S:myc-BPM3 transgenic plants untreated or treated with 100 μM JA at the indicated times. Ponceau staining was used as protein loading control. (B) Immunoblot analysis of HA-BPM3 and histone 2B protein levels in 9-d-old 35S:HA-BPM3 transgenic plants untreated (control; C), treated for 4 h (4 h), or grown for 9 d (9 d) in 50 μM JA. Cytoplasmic (S) and nuclear (N) proteins were isolated and tested with anti-HA and histone 2B antibodies, respectively. Histone 2B was used as a control of cytoplasmic and nuclear fraction isolation. (C) Immunoblot analysis of HA-BPM3 and histone 3 protein levels in 7-d-old 35S:HA-BPM3 transgenic plants untreated, treated for 3 h with 100 μM JA, or treated with 100 μM JA plus 10 μM COR-MO. Nuclear proteins were isolated and tested with anti-HA and histone 3 antibodies, respectively. (D) Immunoblot analysis of myc-BPM3 and actin protein levels in 35S:myc-BPM3 transgenic plants. Seven-day-old seedlings were treated overnight with 50 μM MG132 and then untreated or treated with 50 μM CHX, harvested after 0.5 or 1 h, and tested with anti-myc or anti-actin antibodies, respectively. Protein molecular mass is shown (Right).

Finally, inhibition of translation by cycloheximide or proteasome activity by MG132, bortezomib, or epoxomicin showed that BPM3 is a short-lived protein, degraded by the proteasome, as previously described (Fig. 8D and SI Appendix, Fig. S13A) (43).

In contrast to BPM3, BPM6 accumulation is not altered by JA (SI Appendix, Fig. S13B). Moreover, BPM6 seems to be a more stable protein than BPM3, whose levels do not decay after 4 h of CHX treatment (SI Appendix, Fig. S13C).

Discussion

MYC2, MYC3, and MYC4 are central nodes in plant signaling networks integrating environmental and developmental signals and balancing multiple physiological responses for plant fitness (28, 32–35). Their role in activating plant defenses has the tradeoff of inhibiting growth. Runaway activation of these TFs would be very harmful or even lethal for the cell and therefore needs to be tightly regulated spatially and temporally.

Many mechanisms to ensure a timely repression of these TFs have been uncovered. For instance, phosphorylation, which is required for MYC2 activation, also triggers MYC2 degradation, ensuring a rapid elimination of active protein (36). Additionally, competition for cis-regulatory elements with bHLH repressors balances MYC activity (15, 30, 63–65). Rapid transcriptional induction of their JAZ repressors, which are direct transcriptional targets of MYCs, and splicing variants (or atypical JAZs) resistant to degradation ensures a pulsed, narrow activation of MYCs and a quick rerepression (6, 7, 66–69). Daily variation in MYC protein levels fine-tunes MYC activity during the day/night cycles to optimize the growth–defense balance (38). This daily variation is regulated by circadian clock components [i.e., TIC2 (38) and light quality (37)]. Photoreceptors are required for MYC stability, and conditions that inactivate photoreceptors (dark or shade) reduce MYC protein levels by favoring COP1-mediated degradation (37).

Altogether, these examples underscore how different environmental cues influence plant physiology by modulating the activity of these TFs. They also show that regulation of protein levels is a major way of regulating MYC activity. So far, two ubiquitin ligases have been involved in the regulation of MYC stability, COP1 and PUB10 (37, 40). COP1 reduces MYC levels in the dark and canopy shade, and photoreceptors counteract its activity in the light (37). PUB10 regulates MYC accumulation and, therefore, JA responses, but has not been linked to environmental regulation so far.

Based on the importance of MYC TFs as signaling hubs, additional regulators of their stability should exist and remain to be uncovered. In this work, we identified a small family of E3 ubiquitin ligases that regulate MYC protein stability. This E3 family is based on CUL3 and uses BPM proteins as substrate adaptors (CUL3^BPM). Several members of the BPM family interact directly with MYC2, MYC3, and MYC4 in vivo and in vitro. Pull-down and mass spectrometry results of BPM6-coimmunoprecipitated proteins suggest that MYCs can interact with BPM3 and BPM6 directly and with similar affinity. Y2H assays confirmed a direct interaction. The differences observed in Y2H assays among different combinations of proteins are likely a consequence of the toxicity of BPM and MYC proteins in yeast. However, we cannot exclude that BPM6 may recognize MYCs through dimerization with other BPMs in vivo. This would be consistent with the identification of BPM1, BPM4, and BPM5 in the BPM6 immunoprecipitation purifications.

The JA-related phenotypes (root growth and gene expression) of the reduction-of-function mutants cul3ab, bpm235, and amiR-bpm confirmed that the CUL3^BPM complexes regulate MYC protein stability. This conclusion is also supported by the CUL3^BPM-dependent ubiquitination of MYC3, with the higher levels of MYC2 and MYC3 in amiR-bpm, and by the partial suppression of the root-growth phenotype of amiR-bpm in the jin1-2 background. This partial suppression is consistent with BPMs regulating several MYCs and not only MYC2, the only MYC mutated in jin1-2. It is intriguing, however, that BPMs do not seem to affect all JA-related MYC2-regulated phenotypes. For instance, amiR-bpm and bpm235 do not overaccumulate anthocyanins, in contrast to 35S:MYC2 transgenics. This suggests that BPMs may be more active in specific tissues (i.e., roots), or that different CUL3^BPM complexes may have functional differences in the regulation of particular JA responses. Gene expression results in amiR-bpm and bpm235 support this hypothesis by showing that individual BPMs have a complementary function regulating specific JA-regulated gene sets (as shown by the moderate overlap in misregulated genes between amiR-bpm and bpm235 shown in Fig. 7A and SI Appendix, Fig. S11). Alternatively, BPMs may activate additional pathways that counteract some of the MYC-dependent effects. Indeed, previous studies have identified other CUL3^BPM targets such as AtHB6 and PP2CA, a negative regulator of ABA signaling, and several ERF/AP2 TFs involved in fatty acid metabolism (43, 45, 70). Consistent with these reports, GO terms related to ABA physiology, such as ABA, water deficit, and water transport, are overrepresented in our transcriptomic analyses. Besides ABA, other hormone-related GO terms (i.e., salicylic acid [SA] and ET) are also overrepresented, indicating that, in addition to JA and ABA, BPMs may be involved in the regulation of multiple hormonal signaling pathways. Finally, the amiR-bpm plants might retain residual but sufficient accumulation of BPMs in the aerial part of the plants.

Results on BPM3 stability using the protein synthesis inhibitor cycloheximide and the proteasome inhibitors MG132, bortezomid, or epoxomicin indicate that this protein is unstable and subject to a quick turnover by the proteasome. It is noteworthy that BPM proteins are prone to autoubiquitination in the absence of their substrate, as seen for other Cullin-based ligases (71, 72). Stabilization of BPM3 by JA is dependent on COI1, which parallels the JA-mediated stabilization of MYC proteins (37). Therefore, BPM3 stabilization by JA may be a direct consequence of JA-mediated stabilization of its MYC substrates. This hypothesis also suggests a mechanism for activation of the CUL3^BPM3 complex and its involvement in a negative feedback regulatory loop to reduce protein levels after MYC activation (Fig. 9). This repressor mechanism of MYCs adds to the many described so far, such as the transcriptional induction of JAZ repressors, degradation-resistant JAZs, and splicing variants (6, 7, 66–69), and further underscores the importance of repressing MYC activity to avoid runaway harmful effects. In contrast to BPM3, JA does not regulate BPM6 stability. Stabilization of different BPMs by different signals would provide specificity among BPMs and deserves further research. In any case, BPMs may share overlapping but not fully redundant functions. This is particularly clear from gene expression results shown in Fig. 7, where knockdown lines amiR-bpm and bpm235 share many up-regulated genes but also show evident specificities. Thus, the clustering results suggest that many BPM genes regulate JA responses complementarily and with some specificity that may come from their differential patterns of tissue or temporal expression or relative effect on different MYCs.

Fig. 9.

Model of CUL3^BPM-dependent regulation of MYCs. Activation of the COI1-dependent pathway by JA-Ile activates JA-dependent gene expression through the transcription factors MYC2, MYC3, and MYC4. JA-pathway activation stabilizes BPM3, likely by repressing a negative regulator of BPM3 stability, therefore activating a negative feedback regulatory loop to reduce MYC protein levels and their activity. BPMs are substrate adaptors of Cullin3-based E3 ubiquitin ligases that target MYC proteins for proteasome degradation in response to JA (i.e., CUL3^BPM3) or other unknown signals (i.e., other CUL^BPMs).

Besides BPM3 stabilization, how JA regulates the activity of other BPMs is currently unknown. Our results using the COI1 specific inhibitor COR-MO (62) show that this JA effect is dependent on its COI1 receptor; however, additional signals may be required for BPM–MYC interaction. In this regard, phosphorylation of Thr328 in the transactivation domain is a prerequisite for MYC2 activation (36). Interestingly, phosphorylation of this residue also leads to MYC2 degradation, which ensures a short pulse of MYC2 activity. Although phosphorylation could be the obvious signal inducing BPM–MYC interaction, available 3D structures of MYC proteins do not expand the region containing Thr328 (17, 73). Therefore, the role of Thr328 phosphorylation in the interaction with BPMs remains to be addressed. Clarification of this issue or the identification of additional phosphorylated residues or other amino acid modifications seems essential to understand how BPMs recognize their targets and how specificity is achieved.

The fact that MYC function is regulated by several simultaneous mechanisms, including several E3 ligases (COP1, PUB10, and CUL3^BPM), indicates that MYCs have a broad function, which needs to be tightly controlled. Indeed, besides regulating stress responses, MYCs also inhibit growth by promoting photomorphogenesis, among other mechanisms (74). Thus, that several ubiquitin ligases regulate MYC protein stability is consistent with the idea that MYCs should be controlled not only in response to stress but also temporally and spatially.

In summary, we provide a new framework to understand the fine-tuned regulation of MYC activity. We uncovered a feedback regulatory loop of MYC protein levels mediated by the E3 ligase CUL3^BPM (Fig. 9) that facilitates termination of MYC activity and resetting of the JA pathway.

Materials and Methods

Plant Material and Growth Conditions.

Arabidopsis thaliana Columbia-0 ecotype is the genetic background of the wild-type, mutant, and transgenic lines used throughout the work. Seeds were vernalized at 4 °C for 3 d in the dark and grown in 0.5× Murashige and Skoog (MS) medium with 1% sucrose and 0.7% agar at 21 °C under a 16-h light/8-h dark cycle in a growth chamber.

Two 35S:MYC2-GFP and 1 35S:MYC3-HA transgenic lines previously reported (75) were crossed by Col-0 or introgressed by direct crossing into the amiR-bpm background (43), and F3 homozygous segregating plants were selected in 40 μM hygromycin plates for the wild-type background or in 40 μM hygromycin and 10 μM BASTA plates for the amiR-bpm background, respectively.

amiR-bpm was also introgressed by direct crossing into the MYC2 C5025 recombineering line in which MYC2-GFP expression is under its natural promoter (54, 55). F1 segregating progenies were analyzed by Western blot assays. The MYC2 C5025 line was also crossed by Col-0, and F1 segregating progenies were used as controls in the same experiment. Likewise, the amiR-bpm line (43) was introgressed by direct crossing into the jin1-2 mutant background (25), and F2 segregating progenies were selected in 10 μM BASTA and 50 μM JA plates.

T-DNA insertion lines for bpm2 (GK_391E04), bpm3 (Salk_72848), bpm4 (Salk_082761), and bpm5 (Salk_038471) were obtained from the Nottingham Arabidopsis Stock Centre. The bpm2,bpm3,bpm5 (bpm235) triple mutant was generated by crossing the corresponding double-mutant homozygous lines. F2 segregating progenies of these crosses were genotyped to obtain homozygous plants for each T-DNA insertion line.

Light/dark experiments were performed as described by Chico et al. (37).

Plasmid Constructs.

Full-length BPM1, BPM2, BPM3, BPM4, BPM5, and BPM6 coding sequences carrying a stop codon were PCR-amplified with Phusion polymerase (Thermo Fisher) and cloned by restriction into the pENTR1A (carrying a gentamicin resistance gene; modified from pENTRY1A from Invitrogen) vector for BPM3 and into pENTR3C (carrying a gentamicin resistance gene; modified from pENTRY3C from Invitrogen) for BPM1, BPM2, BPM4, BPM5, and BPM6. PCR primers and restriction enzymes used for amplification and cloning of the BPM genes into pENTRY vectors are listed in SI Appendix, Table S2.

Recombinant Proteins.

For pull-down assays, MBP–MYC proteins were generated as previously reported (6, 11). Full-length BPM3 and BPM6 coding sequences carrying a stop codon were PCR-amplified with Expand High Fidelity polymerase (Roche) or Phusion polymerase (Thermo Fisher) using Gateway-compatible primers (SI Appendix, Table S2). PCR products were cloned into pDONR207 using the Gateway BP II Kit (Invitrogen) to obtain pENTRY-BPM3 and pENTRY-BPM6 and sequence-verified. pENTRY-BPM3 was used in Gateway LR reactions (Invitrogen) and recombined in pDEST-H1 (76) and pEarley201 to obtain N-terminal MBP and HA fusions, respectively. Alternatively, pENTRY-BPM6 was used in Gateway LR reactions (Invitrogen) and recombined in pGWB21 to obtain N-terminal 10×Myc fusions, and in pMDC43 in order to obtain N-terminal GFP fusions.

Recombinant MBP fusions were expressed in Escherichia coli BL21 cells and purified on amylose resin columns (New England Biolabs) following the method previously described (6).

To generate transgenic plants expressing 35S:HA-BPM3 and 35S:myc-BPM6, respectively, the corresponding constructs were transferred to Agrobacterium tumefaciens GV3101 by heat shock and Arabidopsis Col-0 plants were then transformed by floral dipping (77).

Similar to generating plants expressing 35S:GFP-BPM6, transgenic Arabidopsis that already expressed an artificial microRNA directed to BPM2 and BPM3 were transformed by floral dipping using A. tumefaciens GV3101 pMP90 expressing the 35S:GFP-BPM6 construct.

Yeast Two-Hybrid Assays.

Full-length MYC2, MYC3, and MYC4 coding sequences cloned into pGADT7gateway (Gal4 AD) were used as previously described (6, 11). Likewise, full-length BPM1, BPM2, BPM3, and BPM6 constructs were cloned into pGBKT7gateway (GAL4 BD) and BPM4 and BPM5 into pGBT9 (GAL4 BD). To test protein interactions, the corresponding plasmids were cotransformed into Saccharomyces cerevisiae AH 109 cells following standard heat-shock protocols (75). Successfully transformed colonies were identified on yeast synthetic dropout lacking Leu and Trp. At 3 d after transformation, yeast colonies were suspended in water and cell density was adjusted to 3 × 10⁷ cells mL⁻¹ (OD₆₀₀ 1). A 10-μL sample of cell suspensions and 1/3 and 1/10 dilutions were plated out on yeast synthetic dropout lacking Ade, His, Leu, and Trp supplemented or not with 5 mM 3-aminotriazole to test protein interactions. Plates were incubated at 28 °C for 2 to 4 d. Yeast cotransformed with MYC2, MYC3, and MYC4 constructs into the pGADT7 vector together with a pGBKT7 empty vector were used as negative controls.

Protein Extracts, Pull-Down Assays, and Western Blots.

Ten-day-old Arabidopsis wild-type seedlings and lines expressing 35S:HA-BPM3 or 35S:myc-BPM6 were ground in liquid nitrogen and homogenized in extraction buffer containing 50 mM Tris⋅HCl (pH 7.4), 80 mM NaCl, 10% glycerol, 0.1% Tween 20, 1 mM DTT, 1 mM phenylmethylsulfonyl fluoride (PMSF), 50 μM MG132 (Sigma-Aldrich), and complete protease inhibitor (Roche). After two rounds of 15 min of centrifugation at 13,000 rpm at 4 °C, the supernatant was collected. For pull-down experiments, 6 mg of resin-bound MBP fusion protein was added to 1 mg of total protein extract and incubated for 2 h at 4 °C with rotation. After washing, samples were denatured, loaded on 8% sodium dodecyl sulfate (SDS) polyacrylamide gels, transferred to nitrocellulose membranes, and incubated with anti-HA (Roche) or anti–Myc-horseradish peroxidase antibody (Santa Cruz Biotechnology). An 8-μL aliquot of MBP-fused protein of each sample was run on SDS polyacrylamide gels and stained with Coomassie brilliant blue to confirm equal protein loading.

For immunoblotting analysis, 8 to 10 seedlings were harvested per sample, frozen in liquid nitrogen, and homogenized in 200 μL of 2× Laemmli SDS polyacrylamide gel electrophoresis (PAGE) protein loading buffer. The extracts were boiled at 95 °C for 5 min and kept on ice. After centrifugation (5 min, 13,000 rpm at room temperature) the supernatant was collected. A 20- to 40-μL volume of each sample was run on 8% SDS polyacrylamide gels, transferred to nitrocellulose membrane (Bio-Rad), and incubated with anti-GFP (Miltenyi Biotec), anti-HA, anti-Myc, or monoclonal anti-ACTIN (produced in mice; Sigma-Aldrich) antibodies. Blots were developed using ECL (Pierce).

Immunoprecipitation.

A. thaliana 35S:GFP-BPM6 seedlings were ground in liquid nitrogen. Seedling powder was transferred in a cold mortar and lysed with a ratio of weight/lysis (g/mL) buffer volume of 1/3 during 15 min (lysis buffer containing 50 mM Tris⋅HCl, pH 8, 50 mM NaCl, 1% Triton, and cOmplete EDTA-free protease inhibitors; Roche). The protein extracts were clarified by centrifugation and GFP-BPM6 complexes were immunoprecipitated using magnetic microparticles (MACS Purification System; Miltenyi Biotec) according to the manufacturer’s instructions and as previously described (78). µMACS magnetic microbeads were coated with a monoclonal anti-GFP antibody (Miltenyi Biotec). Beads coated with anti-HA antibodies were used for negative controls. Coimmunoprecipitation experiments were carried out in independent biological duplicates; each was divided into three affinity-purification replicates. Proteins were eluted out of the magnetic stand with SDS loading buffer from the kit.

Mass Spectrometry Analysis and Data Processing.

Eluted proteins were digested with sequencing-grade trypsin (Promega) and analyzed by nanoLC-MS/MS (liquid chromatography-tandem mass spectrometry) as described previously (79). Two instruments were used: a QExactive+ mass spectrometer coupled to an EASY-nanoLC 1000 (Thermo Fisher Scientific) and a TripleTOF 5600 mass spectrometer coupled to a nanoLC Ultra 2D Plus (AB Sciex). Data were searched against the TAIR 10 database with a decoy strategy. Peptides were identified with the Mascot algorithm (version 2.5; Matrix Science) and data were imported into Proline 1.4 software (http://proline.profiproteomics.fr/). Proteins were validated on Mascot pretty rank equal to 1, and 1% FDR on both peptide spectrum matches (PSM score) and protein sets (protein set score). The total number of MS/MS fragmentation spectra was used to quantify each protein. This spectral count was submitted to a negative-binomial test using an edgeR GLM regression through the R package (v3.5.0). For each identified protein, an adjusted P value corrected by Benjamini–Hochberg was calculated, as well as a protein fold change. The results are presented in a volcano plot using protein log2 fold changes and their corresponding adjusted log10 P values to highlight enriched proteins. Mass spectrometry data are deposited in the PRIDE database.

Affinity Purification of Ubiquitinated Proteins.

Soluble proteins of MYC3-HA seedlings in WT or amiR-bpm backgrounds were treated overnight with 50 µM bortezomib or alternatively with 50 µM MG132 and extracted in buffer BI (50 mM Tris⋅HCl, pH 7.5, 20 mM NaCl, 0.1% Nonidet P-40, 5 mM ATP, 1 mM PMSF, 50 mM MG132, 10 nM ubiquitin (Ub)-aldehyde, 10 mM N-ethylmaleimide, and plant protease inhibitor mixture; Sigma-Aldrich) before incubation with prewashed p62 agarose (Enzo Life Sciences) or the agarose alone at 4 °C for 4 h. Beads were washed twice in BI buffer and once with BI buffer supplemented with 200 mM NaCl. Proteins were eluted in SDS loading buffer at 100 °C for 5 min. The eluted proteins were separated by SDS/PAGE and analyzed by immunoblotting using anti-Ub (Enzo Life Sciences) or anti-HA antibodies (Roche).

Root Measurements.

For root-growth inhibition assays, the root length of 15 to 20 seedlings was measured 8 to 10 d after germination in the presence or absence of JA (Sigma-Aldrich) or coronatine (Sigma-Aldrich) at the concentrations indicated in each experiment. Three independent replicates (15 to 20 seedlings each) were measured for each sample. Values represent mean ± SD. Comparisons between different lines and the wild type (Col-0) were done by Student’s t test.

Microarray Assays.

Leaves of 3-wk-old wild-type, amiR-bpm, or bpm235 plants treated with 0.5 μM coronatine or mock were harvested and frozen. Three biological replicates were independently hybridized for each transcriptomic comparison. RNA was extracted using TRIzol reagent (Invitrogen) and amplification and labeling were performed basically as previously described (62) starting from 200 ng of total RNA. Genes with a reported FDR <0.05 and a fold change ≥2 and ≤−2 were selected for further investigation. Two different microarray designs were used, the Arabidopsis V4 Oligo Microarray 4 × 44 (Agilent; 021169) for the analysis of amiR-bpm plants and a custom oligo microarray 8 × 60 K (Agilent; GPL22511) in bpm235 that contains the same oligonucleotide probes as V4 and a new batch of probes covering a total of two probes per gene. Transcriptomic profiles of the triple myc234 mutant were previously obtained (74) and correspond to Agilent’s V4 design. For comparative analysis of the different experiments, common oligonucleotide probes were considered.

Chemical Treatments and Hormone Quantification.

Arabidopsis seedlings were germinated and grown on 0.5× MS plates. Seven-day-old seedlings were treated with sterile water containing 50 μM CHX, 50 μM MG132, and 50 or 100 μM JA and harvested at the indicated times. CHX was dissolved in 100% ethanol, MG132 in dimethyl sulfoxide, and JA in dimethylformamide. The chemicals were provided by Sigma-Aldrich.

JA levels were quantified as previously described in ref. 2.

Quantitative RT-PCR.

Quantitative RT-PCR experiments were performed with RNA extracted from 10-d-old seedlings expressing 35S:MYC2-GFP or 35S:MYC3-HA in wild-type and amiR-bpm backgrounds, respectively. RNA extraction and cleanup were done using TRIzol reagent (Invitrogen) followed by the High Pure RNA Isolation Kit (Roche) to remove genomic DNA contamination. cDNA was synthesized from 1 μg of total RNA with the High-Capacity cDNA Reverse Transcription Kit (Applied Biosystems). Four microliters from 1/10-diluted cDNA was used to amplify GFP and the housekeeping ACTIN8 gene using Power SYBR Green (Applied Biosystems). Primer sequences were GFP-F (forward) (5′-TATATCATGGCCGACAAGCA-3′), GFP-R (reverse) (5′-ACTGGGTGCTCAGGTAGTGG-3′), HA-F (5′-ACAAAGTGGTTGATAACAGC-3′), HA-R (5′-GAGCTCTAAGCGCTGCAC-3′), ACT8-F (5′-CCAGTGGTCGTACAAC-CGGTAT-3′), and ACT8-R (5′-TAGTTCTTTTCGATGGAGGAGCTG-3′).

Quantitative PCR was performed in 96-well optical plates in a 7300 Real-Time PCR System (Applied Biosystems). The PCR conditions were as follows: 2 min at 50 °C, 10 min at 95 °C, and 40 cycles of 15 s at 95 °C and 60 s at 60 °C. Data analysis shown was done using three technical replicates from one biological sample. Similar results were obtained with two other biological replicates.

Data Availability Statement.

Microarray datasets have been deposited in the Gene Expression Omnibus (GEO) database, https://www.ncbi.nlm.nih.gov/geo/ under accession nos. GSE131024 (“Transcriptomic profile of bpm2,3,5 triple mutant and Col-0”) and GSE131037 (“Transcriptomic profile of amiR-bpm and Col-0”). These data will be publicly available at the time of publication. Besides this, all materials, data, and associated protocols will be available to readers upon request from the corresponding author.