The transcription factor ANAC017 links mitochondrial retrograde signaling with the ubiquitin–proteasome system to control mitochondrial function in Arabidopsis

Abstract

Mitochondrial biogenesis requires the import of more than a thousand proteins encoded by nuclear DNA. The translocase of the outer mitochondrial membrane (TOM) complex serves as the primary gateway for specific recognition of precursor proteins, which are synthesized in the cytosol. Little is known about the regulation of the abundance of the TOM complex. Using forward genetics, we identified key 26S proteasome subunits, including REGULATORY PARTICLE NON-ATPASE1A (RPN1A), that affect the abundance of TOM-complex subunits through the ubiquitin–proteasome pathway. Loss of proteasome function through rpn1a mutation or MG132 treatment increased the abundance of TOM20 isoforms and induced mitochondrial stress marker genes. By contrast, overexpression of ANAC017, an endoplasmic reticulum–anchored transcription factor that activates mitochondrial retrograde signaling under stress, lowered TOM20 abundance and reduced mitochondrial protein import. The rates of mitochondrial protein import and respiratory activity were also altered. Genetic analyses placed the proteasome downstream of ANAC017, since the reduction in TOM20 required the RPN1a subunit. Transcriptome profiling after antimycin A treatment showed broad ANAC017-dependent reprogramming of ubiquitin–proteasome system genes. A second tier formed by ANAC053- and ANAC078-bound promoters of proteasome subunits, including RPN1a, is required to restrain TOM20 accumulation. These findings establish a two-step transcriptional circuit that engages the ubiquitin–proteasome system to tune TOM abundance and coordinate protein import with organelle function.

This study reports a NAC transcription factor regulatory cascade consisting of ANAC017, ANAC053, and ANAC078 controls mitochondrial biogenesis and function via the ubiquitin–proteasome system in Arabidopsis thaliana, thereby linking transcriptional and post-transcriptional layers of mitochondrial retrograde regulation.

Introduction

Mitochondria are defining structures within eukaryotic cells, crucial for energy production and a myriad of metabolic processes. Although the mitochondrion has its own genome, the majority of mitochondrial proteins are encoded by the nuclear genome. As a result, mitochondrial homeostasis relies on coordinated nuclear gene expression and feedback signaling between the mitochondria and the nucleus. When mitochondria become dysfunctional, they send signals to the nucleus to trigger an adaptive response, leading to changes in nuclear gene expression and cell physiology in a process known as mitochondrial retrograde signaling (MRS) (Ng et al., 2014; Kleine and Leister, 2016). In Arabidopsis thaliana, MRS is regulated predominantly by the endoplasmic reticulum–localized transcription factor ANAC017. During stress conditions, such as oxidative stress, ANAC017 translocates to the nucleus, where it initiates MRS to reprogram mitochondrial function (Ng et al., 2013). One of the key downstream targets of ANAC017 is ALTERNATIVE OXIDASE 1A (AOX1a), which plays a pivotal role in helping plants to manage oxidative stress (Selinski et al., 2018). ANAC017-mediated MRS also regulates several other stress responses, including tolerance to cellulose biosynthesis inhibitors such as C17 and isoxaben (Hu et al., 2016; Broad et al., 2024).

Nuclear-encoded mitochondrial proteins are synthesized in the cytosol and then imported into the mitochondria. These proteins carry targeting signals recognized by mitochondrial surface receptors at the translocase of the outer membrane (TOM) complex, which acts as the main entry point (Murcha et al., 2014b). The core mechanisms and components involved in this process are thought to have arisen early in eukaryotic evolution, as evidenced by the structural and functional conservation of key elements of the TOM complex across diverse eukaryotic lineages (Duncan et al., 2013). The TOM complex consists of the core translocation pore TOM40, together with membrane-anchored receptor proteins that face the cytosol, i.e., TOM20, TOM9, and OUTER MEMBRANE 64 (Murcha et al., 2014a). The small accessory proteins TOM5, TOM6, and TOM7 are also associated with the TOM complex and contribute to its stability (Murcha et al., 2014a). TOM20, a peripheral subunit of the TOM40 complex, functions as a general import receptor by recognizing mitochondrial targeting signals in protein presequences (Yamamoto et al., 2011). Similar to those in yeast, mutations in Arabidopsis TOM20 proteins result in significantly reduced protein import rate and impaired growth (Ramage et al., 1993; Lister et al., 2007). Although the TOM complex has been functionally characterized, the mechanisms that govern its abundance in plants are poorly understood.

In both yeast and mammals, the ubiquitin–proteasome system (UPS) plays a crucial role in the clearance of damaged or defective proteins at the mitochondrial outer membrane (MOM), a process termed mitochondrial-associated degradation (MAD) (Heo et al., 2010; Ruan et al., 2020). MAD targets some yeast MOM proteins, including Fzo1 (fusion protein 1), Mdm12 (mitochondrial distribution and morphology protein 12), Mdm34, and Tom70, for ubiquitination and degradation to maintain mitochondrial protein homeostasis (Cohen et al., 2008; Ota et al., 2008; Wu et al., 2016; Belgareh-Touze et al., 2017). Of these, only Tom70 is a member of the TOM complex and is regulated by the E3 ligase Reverse Spt-Phenotype 5 (RSP5) (Wu et al., 2016). It remains unclear whether TOM-complex proteins are similarly regulated by the UPS in plants, and the existence of a plant MAD pathway has yet to be demonstrated. However, the protein import complexes of other plant cell organelles, including chloroplasts and peroxisomes, are regulated by the UPS (Akhter et al., 2023; Sun and Jarvis, 2023). The molecular mechanisms that regulate the UPS pathway to control the organelle protein import machinery have also not been identified in any species, revealing a critical gap in our understanding of organelle development and function.

In this study, we demonstrate the existence of the MAD pathway in plants and show that MOM proteins undergo ubiquitination and subsequent degradation via the UPS. This pathway is regulated by a two-tier transcriptional cascade in which ANAC017 controls the expression of ANAC053, which in turn regulates genes involved in the UPS pathway to modulate the abundance of the TOM complex.

Results

A forward genetic screen reveals the role of the ubiquitin–proteasome system (UPS) in activating mitochondrial retrograde signaling

Previously, we demonstrated that plants with constitutively induced MRS exhibit remarkable tolerance to the cellulose biosynthesis inhibitor isoxaben (Hu et al., 2016; Broad et al., 2024). To further investigate this phenomenon, we applied antimycin A (AA), a well-characterized inhibitor of mitochondrial complex III, and isoxaben to wild-type plants. Consistent with previous findings, AA suppressed isoxaben-induced growth inhibition, reinforcing the idea that isoxaben tolerance can serve as an effective proxy for constitutive MRS activation (Figure 1A and 1B). To identify additional components involved in MRS, we used isoxaben tolerance as a screening tool. Approximately 5000 homozygous Arabidopsis thaliana transfer DNA (T-DNA) insertion mutants (Alonso et al., 2003) were grown vertically on medium containing 2.5 nM isoxaben, alongside wild-type controls. Mutants exhibiting longer roots were identified as isoxaben-tolerant candidates. Interestingly, we identified a set of mutants in components of the UPS pathway that exhibited different levels of isoxaben tolerance compared with Col-0, namely, 26S PROTEASOME REGULATORY SUBUNIT 1A (RPN1A), RPN2A, REGULATORY PARTICLE AAA-ATPase 2A (RPT2A), RPT5A, and 20S PROTEASOME BETA SUBUNIT E1 (PBE1) (Figure 1C and 1D). These mutants were particularly intriguing, as they suggested a potential link between the UPS pathway and mitochondrial protein degradation. As controls, we used the ANAC017 overexpression line ANAC017OE2, which constitutively activates the MRS and exhibits strong tolerance to isoxaben, and the ANAC017 loss-of-function mutant rao2-1, which shows sensitivity to isoxaben treatment (Figure 1C). Isoxaben-induced growth inhibition is most obvious in roots (Ogden et al., 2024), but we also noticed pronounced leaf chlorosis in rao2-1. Visual scoring confirmed that rao2-1 developed significantly paler shoots, whereas ANAC017OE2 and rpn1a mutants displayed less severe stress. To probe the stress physiology underlying these phenotypes, we stained seedlings with 3,3′-diaminobenzidine and nitroblue tetrazolium. Isoxaben triggered a burst of reactive oxygen species (ROS) that was evident in wild-type tissues and further intensified in rao2-1 (Supplemental Figure 1A–1C). These results suggest that rao2-1 mutants show increased sensitivity to isoxaben in both roots and shoots.

Figure 1.

A forward genetic screen identified UPS mutants that exhibit isoxaben tolerance and induced expression of MRS marker genes.

(A) Col-0 grown vertically on 1/2 MS medium supplemented with 1% sucrose and 0.8% agar (control) with 1 μM AA, 2.5 nM isoxaben (ISX), or both (AA + ISX). Representative images were taken from 10-day-old seedlings. Scale bar, 1 cm.

(B) The primary root lengths in (A) were measured, and the data are shown as mean ± SE, N = 16–23. Different letters indicate significant differences at p < 0.05 determined by one-way ANOVA with Tukey’s honestly significant difference (HSD) test. The exact p values are provided in Supplemental Data 6.

(C) Col-0, an ANAC017-overexpressing line (ANAC017OE2), and T-DNA insertion mutants of RPN1A (rpn1a-5 and rpn1a-2), PBE1 (pbe1-2 and pbe1-3), RPT2A (rpt2a-3), RPN2A (rpn2a-1), and RPT5A (rpt5a-1) were grown on 1/2 MS medium supplemented with 1% sucrose and 0.8% agar, with (ISX) or without (control) 2.5 nM isoxaben. Representative images were taken from 10-day-old seedlings. Scale bar, 1 cm.

(D) The ratio of primary root length in (C) of 10-day-old seedlings under 2.5 nM isoxaben or control conditions was calculated by averaging the control and dividing each isoxaben value by the control average; data are shown as mean ± SE, N = 18–22. Different letters indicate significant differences at p < 0.05 determined by one-way ANOVA with Tukey’s HSD test. The exact p values are provided in Supplemental Data 6.

(E) UPS mutants show induced expression of MRS marker genes under control conditions. The bar chart shows the relative expression of UP-REGULATED BY OXIDATIVE STRESS (UPOX1), ALTERNATIVE OXIDASE 1A (AOX1a), OUTER MITOCHONDRIAL MEMBRANE PROTEIN OF 66 kDa (OM66), and NAD(P)H DEHYDROGENASE B2 (NDB2). Data are shown as means of 40-ΔCt values ± SE of three biological replicates, and asterisks indicate significant differences in expression between mutants and Col-0. Significant differences were identified by one-way ANOVA followed by Tukey’s HSD test (∗p < 0.05). The exact p values are provided in Supplemental Data 6.

To determine whether the UPS pathway mutants displayed constitutive MRS activation, we performed quantitative reverse transcription PCR (RT–qPCR) of several MRS marker genes, including UP-REGULATED BY OXIDATIVE STRESS 1 (UPOX1), AOX1a, OUTER MITOCHONDRIAL MEMBRANE PROTEIN OF 66 kDa (OM66), and NAD(P)H DEHYDROGENASE B2 (NDB2). These marker genes were significantly induced in ANAC017OE-2. Similarly, mutant alleles of rpn1a and pbe1 exhibited significant induction of all tested marker genes, whereas rpt2a-3 and rpt5a-1 showed induction of only UPOX1 and AOX1a (Figure 1E). Although ANAC013 and ANAC017 are key regulators of the MRS, their transcript levels were largely unchanged in the UPS mutants (Supplemental Figure 1D). Phenotypically, rpn1a mutants exhibited no significant developmental abnormalities, displaying only a slight reduction in rosette growth compared with Col-0 (Supplemental Figure 2A). These data demonstrate that mutations in proteasome subunits result in mitochondrial dysfunction, evidenced by the constitutive activation of MRS. This suggests that a fully functional proteasome interacts with mitochondrial signaling and/or function.

Mutants of genes encoding components of the UPS exhibit altered abundance of mitochondrial outer membrane proteins

Given that rpn1a mutants showed strong isoxaben tolerance and MRS marker gene induction, we investigated the effect of RPN1a on mitochondrial protein abundance in Arabidopsis. Western blot analysis was performed on mitochondria isolated from untreated Col-0, Col-0 plants pre-treated with the proteasome inhibitor MG132, and rpn1a mutants (Supplemental Figure 2B and 2C). Interestingly, we observed increased abundance of the TOM20 isoforms in rpn1a mutants (Figure 2A and Supplemental Figure 3A) and in Col-0 plants treated with MG132 four days before mitochondrial isolation (Figure 2A and Supplemental Figure 3A).

Figure 2.

ANAC017 and UPS-component mutants affect mitochondrial proteins.

(A) Immunoblot analysis of purified mitochondria extracted from 10-day-old liquid-cultured seedlings: Col-0 vs. rao2-1 and ANAC017OE2, Col-0 vs. rpn1a-5 and rpn1a-2, and control (DMSO-treated Col-0) vs. MG132 (50 μM MG132–treated Col-0). Antibodies used are shown on the left and targeted Sam50, Porin, TOM40, TOM20-2, TOM20-3, and TOM20-4 (outer membrane proteins) and MIC60, RISP (the complex III subunit), and AOX (inner membrane proteins). Ten micrograms of mitochondrial proteins were loaded for each set. Protein loading controls for the results presented here are provided in Supplemental Figure 4.

(B) The [³⁵S]-radiolabeled precursor protein MGE2 was incubated with mitochondria isolated from Col-0, rao2-1, and ANAC017OE2 lines and rpn1a-5 lines under conditions that supported protein import. Aliquots were removed at 5-, 10-, and 20-min time points and treated with Proteinase K. m, mature protein; p, precursor protein.

(C) The activity of mitochondrial respiratory chain complexes from Col-0, rao2-1, and ANAC017OE2 lines and Col-0 vs. rpn1a-5 and rpn1a-2 lines measured using a Clark-type oxygen electrode and shown as mean ± SE of three biological replicates. Asterisks indicate significant differences (∗p < 0.05) between the oxygen consumption of wild-type mitochondria (Col-0) and mitochondria from different mutants as determined by Student’s t test.

These results indicate that the UPS component RPN1A is involved in TOM20 degradation. Western blot analyses of an ANAC017 overexpression line (OE2), which constitutively activates MRS, and the ANAC017 loss-of-function line rao2-1 (Ng et al., 2013; Meng et al., 2019) revealed that the rao2-1 mutants largely mimicked the rpn1a mutants (Figure 2A and Supplemental Figure 3A). By contrast, the abundance of the three TOM20 isoforms was significantly reduced in ANAC017OE2 (Figure 2A and Supplemental Figure 3A). Two additional independent lines, ANAC017OE3 and ANAC017 knockout line 1 (anac017 KO1), were used to confirm the changes in abundance of TOM20 isoforms (Supplemental Figure 3B). Abundance of the other outer membrane proteins, SAM50 and Porin, and the inner membrane proteins RISP (Rieske FeS protein of the cytochrome bc₁ complex) and MIC60 (mitochondrial contact site and cristae organizing system) was largely unchanged in all lines (Figure 2A and Supplemental Figure 4).

As the abundance of TOM20 isoforms was significantly reduced in ANAC017OE overexpression lines, we measured the rate of protein import into the mitochondria in vitro using a plant mitochondrial precursor protein, mitochondrial GrpE protein homolog 2 (MGE2, At4g26780) (Gill-Hille et al., 2022) (Figure 2B). The import rate of MGE2 was unaffected in rao2-1 and rpn1a-5 mutants (Figure 2B and Supplemental Figure 3C); however, significant differences in the rate of protein import were observed in ANAC017OE2. MGE2 import into ANAC017OE2 mitochondria was reduced to 72% compared with that into wild-type mitochondria (Supplemental Figure 3C). These results suggest that lower TOM20 abundance in ANAC017OE2 leads to reduced mitochondrial protein import. By contrast, rpn1a mutants, which showed increased TOM20 levels, did not display enhanced import. This is consistent with previous findings that TIM components, not TOM20, are rate limiting for import (Neupert and Brunner, 2002; Wang et al., 2012; Murcha et al., 2014b). Consistent with the altered abundance of mitochondrial proteins, a similar trend was observed in respiratory capacity (Figure 2C). Activity of the alternative oxidase pathway was increased in both rpn1a mutants and ANAC017OE2, and a slight increase was also observed in rao2-1 mutants. In addition, respiration through the Cyt c oxidase pathway showed a general increase in both rpn1a mutants and ANAC017OE2 but was unchanged in rao2-1. These results indicate that the mutants show similar changes in mitochondrial function.

To rule out the possibility that the observed changes in protein abundance could be attributed to alterations at the transcriptional level, we performed RT–qPCR analysis of AOX1a and several MOM genes, including TOM20-2, TOM20-3, TOM20-4, and TOM40. There were no significant changes in transcript abundance of the tested MOM genes in the rpn1a, rao2-1, and ANAC017OE2 lines (Supplemental Figure 5). However, the matrix-facing inner membrane protein AOX did show changes in abundance in all mutants (Figure 2A). Consistent with the well-described role of ANAC017 as a positive transcriptional regulator of AOX1a in Arabidopsis (Ng et al., 2013), AOX1a protein abundance was reduced in the loss-of-function mutant (rao2-1) but significantly increased in the ANAC017 overexpression line ANAC017OE2. AOX levels were also elevated in rpn1a mutants, as well as in plants treated with MG132 (Figure 2A). The increase in AOX abundance in rpn1a lines and in plants pre-treated with MG132 revealed a previously unreported role for the UPS in determining AOX abundance (Figure 2A).

The mitochondrial outer membrane protein TOM20-3 is ubiquitinated

Our findings that TOM20 protein abundance was altered in rpn1a and upon treatment with MG132 led us to investigate whether these proteins are ubiquitinated, using immunoblot analysis. To enhance the sensitivity of the analysis, we generated transgenic plants expressing 6×MYC-tagged ubiquitin (35S:6×MYC-UBQ) (Sun et al., 2022). The 35S: TOM20-3:3×FLAG construct was co-infiltrated with the 35S:6×MYC-UBQ construct into tobacco leaves, with the 3×FLAG empty vector as a control. Proteins extracted from the infiltrated leaves were subjected to immunoprecipitation with FLAG-Trap agarose and blotted with anti-FLAG and anti-MYC antibodies. High-molecular-weight smears were detected with both antibodies, supporting TOM20-3 ubiquitination (Figure 3A). To further confirm TOM20-3 ubiquitination, we performed additional ubiquitination studies in transiently transformed tobacco leaves expressing TOM20-3:3×FLAG or a 3×FLAG empty vector control. Proteins extracted from infiltrated leaves were subjected to immunoprecipitation with FLAG-Trap agarose (Chromotek) and blotted with anti-TOM20-3 and anti-UBQ11 antibodies, respectively, using purified Col-0 mitochondrial protein as a native control. Consistent with earlier findings in our study, TOM20-3 displayed higher molecular weight modifications, and blots with anti-UBQ11 showed a higher-molecular-weight smear, indicating TOM20-3 ubiquitination (Figure 3B).

Figure 3.

Proteins in the TOM complex are ubiquitinated.

(A) Transient expression assay for TOM20-3 ubiquitination. The 35S:TOM20-3:3×FLAG and 35S:3×FLAG plasmids were transiently transfected into N. benthamiana leaves with 35S:6×MYC-UBQ11 and co-transfected for 2 days. Total proteins were extracted, incubated with FLAG-Trap agarose, and analyzed by immunoblotting using α-FLAG and α-MYC antibodies. poly-Ub, poly-ubiquitinated.

(B) Transient expression assay for TOM20-3 ubiquitination. The 35S:TOM20-3:3×FLAG and 35S:3×FLAG plasmids were transiently expressed in N. benthamiana leaves for 2 days; total proteins were extracted and incubated with FLAG-Trap agarose, then analyzed by immunoblotting using α-TOM20-3 and α-UBQ11 antibodies. The first lane, Col-0 Mito (Col-0 mitochondria), was extracted from Arabidopsis thaliana and used as a native marker.

UPS-mediated regulation of MOM protein abundance is under the control of ANAC017

We next investigated whether the changes in mitochondrial protein abundance in rpn1a mutants were due to a mechanism similar to that observed in rao2-1 and ANAC017OE2. We therefore performed genetic crosses of rpn1a-5 with rao2-1 and rpn1a-5 with ANAC017OE2. For this assay, we assessed genetic interactions by testing the ability of mutants to grow on AA, which requires AOX1a. Whereas aox1a plants grew like wild-type plants on control medium, their growth was suppressed in the presence of AA (Strodtkötter et al., 2009). Under control conditions, rao2-1, rpn1a-5, rao2-1/rpn1a-5, ANAC017OE2, and rpn1a-5/ANAC017OE2 showed only slight growth alterations, whereas aox1a grew like Col-0 (Figure 4A–4D). As expected, rao2-1 plants exhibited increased sensitivity to AA in both shoots (Figure 4A) and roots (Figure 4B and D), which is due to their inability to induce AOX1a upon AA treatment (Strodtkötter et al., 2009). In line with this observation, ANAC017OE2 exhibited increased tolerance to AA (Meng et al., 2019) (Figure 4A–4D).

Figure 4.

ANAC017 acts upstream of the UPS to regulate mitochondrial protein abundance.

(A) The indicated seedlings were grown horizontally on 1/2 MS plates (control) or 50 μM AA plates for 10 days for shoot observation (two upper panels). (Lower) Seedlings were grown on 1/2 MS plates for 10 days and then sprayed with 50 μM AA and imaged 72 h later.

(B) The indicated seedlings were grown vertically on 1/2 MS plates (control) or 50 μM AA plates for 10 days for observation of primary root length.

(C and D) Quantified primary root lengths are shown in the boxplot: first and third quartiles, with median and Tukey’s whiskers. Different letters indicate significant differences at p < 0.05 determined by one-way ANOVA with Tukey’s HSD test. The exact p values are provided in Supplemental Data 6.

(E) Immunoblotting analysis of purified mitochondria extracted from 10-day-old seedlings grown on 1/2 MS plates, sprayed with 50 μM AA or water (control) as indicated and harvested 6 h post treatment; antibodies used are shown at left. Ten micrograms of mitochondrial proteins were loaded for each set. Protein loading controls for the results presented here are provided in Supplemental Figure 6.

Interestingly, the AA sensitivity of rao2-1 was suppressed by rpn1a-5 (Figure 4A–4D), likely owing to increased AOX1a abundance in these mutants. Through western blot analysis of purified mitochondria, we confirmed that AOX1a abundance was increased in rpn1a-5/rao2-1, both under control conditions and after AA treatment (Figure 4E and Supplemental Figure 6). These findings suggest that the UPS-pathway gene RPN1a can alter the abundance of AOX1a independently of ANAC017, thereby suppressing the sensitivity of rao2-1 to AA. In addition, we showed that the isoxaben sensitivity of rao2-1 was rescued by the RPN1A mutation, but not through AOX1a, as the aox1a mutant shows no sensitivity to isoxaben (Supplemental Figure 7).

To understand the molecular interaction between ANAC017 and the UPS-pathway-mediated regulation of TOM complex abundance, we performed western blot analysis to determine the levels of TOM proteins in the double mutants. This was carried out using purified mitochondria from plants treated with or without AA for 6 h. TOM20 isoforms exhibited increased protein abundance in rao2-1, as observed previously (Figure 2A). A similar increase was also seen in rpn1a-5/rao2-1 (Figure 4E and Supplemental Figure 6). These results suggest that both ANAC017 and RPN1a negatively regulate TOM20 abundance and act within the same pathway, as no additive effect was observed. Interestingly, after AA treatment, there was a decrease in TOM20 isoforms in Col-0, and this decrease was partially suppressed in rao2-1, rpn1a-5, and rpn1a-5/rao2-1 mutants. This suggests that AA-mediated degradation of TOM20 isoforms requires ANAC017 and RPN1a (Figure 4E and Supplemental Figure 6). By contrast, the ANAC017OE2-mediated decrease in abundance of TOM20 isoforms was strongly suppressed in rpn1a-5/ANAC017OE2 (Figure 4E). These results demonstrate that ANAC017 mediates the changes in TOM20 abundance via the UPS pathway and requires the RPN1a subunit of the proteasome complex.

Genes encoding components of the UPS respond to antimycin A treatment, that is modulated by ANAC017

The evidence that ANAC017 regulates MOM protein abundance via the UPS pathway prompted us to further investigate the molecular link between ANAC017 and the UPS pathway. First, to characterize the connection between ANAC017 and regulation of the UPS, we examined changes in the expression of UPS-related genes (Supplemental Data 1) using previously published transcriptional data for various ANAC017 overexpression and mutant lines (Meng et al., 2019). Notably, nearly 17% of all UPS-related genes were differentially regulated in ANAC017 overexpression or mutant lines (Supplemental Figure 8A and Supplemental Data 2), suggesting that ANAC017 plays a significant role in the transcriptional regulation of UPS genes.

To explore the potential link, we performed RNA sequencing (RNA-seq) analysis of Col-0, rao2-1, anac017KO1, ANAC017OE2, and ANAC017OE3 after 3 h of AA treatment. The greatest number of significant differentially expressed genes (DEGs log₂[fold change] > 1, false discovery rate < 0.05; Supplemental Data 3 and 4) occurred in the ANAC017 loss-of-function lines rao2-1 and anac017KO1, each of which had > 6500 affected genes, followed by the ANAC017 overexpression lines (> 4000 genes; Figure 5A). To confirm that AA treatment induced mitochondrial dysfunction, expression of established marker genes was assessed. All showed increased transcript abundance in an ANAC017-dependent manner (Figure 5B), consistent with prior reports.

Figure 5.

Genes encoding components involved in the UPS are responsive to mitochondrial dysfunction, and this is affected by ANAC017.

(A) Numbers of genes that were up- or downregulated upon AA treatment of Col-0 and in ANAC017 mutant lines. Three sets of comparisons were performed: a genotype comparison under control conditions (lower), a treatment comparison with the response in each genotype determined (middle), and a genotype by treatment comparison (upper).

(B) Responses of marker genes for mitochondrial dysfunction, to ensure that the treatment and genotypes responded as documented previously. Expression is shown as mean ± SE of transcripts per million.

(C) Hierarchical clustering of 533 genes encoding components of the UPS, divided into 5 clusters according to their expression patterns, with word clouds for Cluster 2 and Cluster 3. Expression is shown as the Z score of the normalized counts.

Among a curated list of 1749 UPS-related genes (Supplemental Data 1), 533 displayed significant transcript changes (Supplemental Data 5), and 184 of these genes changed significantly in wild-type plants upon AA treatment (Supplemental Data 5). Hierarchical clustering revealed that AA triggered broad transcriptional reprogramming of UPS genes (Figure 5C). Under control conditions, expression patterns were largely similar across lines, consistent with the limited DEG counts in mock-treated samples and the latent nature of ANAC017. Upon treatment, ∼70% of UPS genes were altered, forming 5 clusters that distinguished Col-0 from ANAC017 mutants. Cluster 2 (268 genes) increased in rao2-1 and anac017KO1 but decreased in overexpression lines, indicating that ANAC017 acts as a negative regulator. Cluster 3 (91 genes) showed the opposite trend, consistent with positive regulation by ANAC017. Together, these clusters encompassed 359 of the 533 responsive genes.

Cluster 2 included UBOX, RING finger, and ubiquitin-conjugating genes linked to mitochondrial or organelle function, including CHIP (At3g07370), a chaperone-interacting E3 ligase for chloroplast preprotein degradation (Lee et al., 2009); SPL2 (At1g54150), a chloroplast envelope E3 ligase (Mohd Ali et al., 2023); NLA (At1g02860), a negative regulator of phosphate limitation responses (Park et al., 2023); and BRUTUS-like protein (At3g18290), involved in iron and chilling responses (Pullin et al., 2025; Xing et al., 2025), that links to mitochondrial signaling via ANAC044 and PAP signaling (Yan et al., 2024). A total of 11 ubiquitin-conjugating enzymes and 14 PUB genes also changed significantly (Supplemental Data 5).

Cluster 3 genes, positively regulated by ANAC017 and upregulated in Col-0 after treatment, included Cullin2, a chloroplast envelope protein (Froehlich et al., 2003); ATG8 family members involved in mitochondrial autophagy (Li et al., 2022); and the E3 ligases AIP2 and AIRP2 that mediate ABA signaling (Zhang et al., 2005; Giraud et al., 2009; Cho et al., 2011). SGR9 (At5g02750), which interacts with F-actin filaments associated with the endoplasmic reticulum and mitochondria (Nakamura et al., 2011; Ng et al., 2013), was also upregulated, highlighting the integration of ANAC017 with organelle dynamics and stress signaling.

ANAC017 regulates proteasome-mediated degradation of mitochondrial proteins via ANAC053 and ANAC078

To further understand the molecular mechanisms that underpin ANAC017-mediated UPS regulation, we investigated the roles of the transcription factors ANAC078 and ANAC053. These transcription factors are known to function as homo- and heterodimers, directly binding to the promoters of genes encoding 26S proteasome subunits to regulate their expression (Figure 6A) (Nguyen et al., 2013; Gladman et al., 2016). We previously showed that ANAC017 directly binds to the promoter of ANAC053 to control its expression (Figure 6A) (He et al., 2022; Zhu et al., 2023). ANAC053 and ANAC078 recognize Proteasome-Related cis-Elements (PRCEs), which are present in the promoters of numerous genes encoding 26S proteasome components, including RPN1a (Figure 6B) (Nguyen et al., 2013; Gladman et al., 2016). Genome-wide DNA affinity purification sequencing (DAP-seq) showed binding of ANAC053 and ANAC078 to a variety of promoters encoding 26S components (Supplemental Figure 9) (O'Malley et al., 2016) and, most notably, the components identified in the forward genetic screen in this study (Figure 6C). We performed a dual-luciferase assay in Nicotiana benthamiana to confirm that both ANAC053 and ANAC078 strongly induce the expression of RPN1A by binding to its promoter. Although ANAC017 was also able to activate RPN1A expression, it did so to a much lesser extent. These results suggest that ANAC053 and ANAC078 are able to directly activate RPN1A expression (Figure 6D and Supplemental Figure 10).

Figure 6.

ANAC017 regulates UPS-mediated regulation of mitochondrial proteins via ANAC053/ANAC078.

(A) A sketch showing the two-tier transcriptional network.

(B) Position weight matrix for the PRCE (Gladman et al., 2016) and schematic diagram of identified PRCE binding sites in the RPN1a promoter. TSS, transcription start site.

(C) AnnoJ browser snapshots of ANAC053, ANAC078, and ANAC017 DAP-seq peaks (O'Malley et al., 2016) in the promoters of the screened UPS genes in Figure 1C.

(D) Transcriptional activation assays performed by cotransfecting tobacco protoplasts with the constructs shown in the upper panel. The luciferase activity was normalized to that of cells coexpressing the empty effector vector. Error bars represent the mean ± SE of three biological replicates. Statistical significance was calculated by Student’s t test, and p values are indicated.

(E) Immunoblotting analysis of purified mitochondrial extracts from 10-day-old liquid-cultured seedlings, including Col-0 vs. anac053-1/anac078-1, anac053-1, and anac078-1, with the antibodies used shown at left (SAM50, TOM40, TOM20-2, TOM20-3, TOM20-4, and AOX). Ten micrograms of mitochondrial proteins were loaded for each set. Protein loading controls for the results presented here are provided in Supplemental Figure 11.

(F) Protein abundance was normalized to total protein loading and relative to Col-0 with Image Lab software for quantification. The bars shown are the mean ± SE (n = 3 biological replicates). The asterisks indicate statistically significant differences (Student’s t test, ∗p < 0.05), and the exact p values are provided in Supplemental Data 6.

(G) Relative expression of SAM50, TOM40, TOM20-2, TOM20-3, TOM20-4, AOX1a, ANAC053, and ANAC078 in Col-0, anac053-1/anac078-1, anac053-1, and anac078-1. Ten-day-old seedlings were harvested, and transcript abundance was quantified by RT–qPCR. Data shown are the means of 40-ΔCt values ± SE of three biological replicates, and asterisks indicate statistically significant differences (Student’s t test, ∗p < 0.05); the exact p values are provided in Supplemental Data 6.

We performed western blot analysis of mitochondria isolated from Col-0, anac053-1, anac078-1, and anac053-1/anac078-1 plants to investigate the abundance of TOM-complex proteins. We found that TOM-complex proteins in the MOM, including TOM20-2, TOM20-3, and to a lesser extent TOM20-4, accumulated in all three mutant lines, whereas there was no change in SAM50 and TOM40 (Figure 6E and 6F and Supplemental Figure 11). Interestingly, AOX proteins showed a slight decrease in all mutants (Figure 6E and 6F). We further confirmed that there were no significant changes in the analyzed MOM proteins at the transcriptional level in anac053-1, anac078-1, and anac053-1/anac078-1 (Figure 6G). These results show that ANAC053 and ANAC078 regulate the accumulation of MOM proteins, with ANAC053 acting downstream of ANAC017.

Discussion

We conclude that the MAD pathway operates in plants, on the basis of altered mitochondrial function in mutants of genes encoding UPS components. Changes observed include altered abundance of all three TOM20 isoforms, impaired respiratory activity, constitutive expression of genes that are markers of mitochondrial stress, and enhanced resistance to AA in an anac017 mutant, revealing the effect of rpn1a mutants on mitochondrial function. Our characterization focused on one UPS component, RPN1a, identified through a forward genetic screen; however, several other UPS-related genes were also recovered. Given that only ∼5000 T-DNA lines were screened, it is likely that additional UPS components influencing mitochondrial function were not captured in this screen. This is supported by the RNA-seq data, which showed that numerous genes encoding UPS components exhibited altered transcript abundance in Col-0 upon AA treatment and that some of these were altered in the anac017 mutant lines. It is also important to note that the observed changes in protein abundance of the TOM20 isoforms and AOX1a likely occur through distinct or partially overlapping pathways. Whereas rpn1a mutants did not show altered transcript abundance of genes encoding the TOM20 isoforms, AOX1a transcripts were significantly upregulated in these mutants (Supplemental Figure 5). Furthermore, activation of a constitutive mitochondrial stress response involves changes in auxin and ethylene, both of which act as regulators of AOX1a (He et al., 2022). Thus, the current evidence does not establish that components of the UPS directly regulate AOX1a protein abundance. Our study identifies a two-tier transcriptional regulatory network that regulates the MAD pathway responsible for turnover of mitochondrial protein-import machinery in Arabidopsis (Figure 7). ANAC017, a master regulator of MRS, directly binds to the ANAC053 promoter (He et al., 2022), which, together with its closest homolog, ANAC078, regulates the expression of UPS pathway components (Gladman et al., 2016). Transcriptome analysis revealed that over 500 genes encoding UPS components changed significantly in transcript abundance upon perturbation of mitochondrial function and that ANAC017 affects these changes. Although ANAC017 was originally identified in a screen to identify regulators of MRR (Ng et al., 2013), subsequent studies have shown its diverse roles in many other biological processes, including senescence (Meng et al., 2019; Broda et al., 2021). In many of these processes, protein turnover via the UPS has also been described (Zhang et al., 2021), and the molecular components are just beginning to be uncovered.

Figure 7.

Model illustrating how ANAC017 links mitochondrial retrograde signaling (MRS) to ubiquitin–proteasome system (UPS)-mediated control of the mitochondrial import machinery.

During oxidative stress (e.g., treatment with AA), the endoplasmic reticulum–anchored transcription factor ANAC017 is cleaved and relocates to the nucleus, where it activates MRS. This activation induces the expression of genes such as Alternative Oxidase 1a (AOX1a), which sustain electron flow through the mitochondrial electron transport chain and reduce the accumulation of ROS. In parallel, ANAC017 reprograms the UPS through both positive and negative regulation. This occurs directly and indirectly through a second regulatory tier involving ANAC053 and ANAC078, which together fine-tune proteasome capacity by modulating the expression of subunits such as Regulatory Particle Non-ATPase 1a (RPN1a). These changes in proteasome function influence the turnover of TOM20 isoforms, adjusting mitochondrial protein import during stress. Genetic analyses place RPN1a downstream of ANAC017, as the reduction in TOM20 observed in the ANAC017 overexpression line (ANAC017OE2) was suppressed in the rpn1a-5 mutant background, confirming their functional relationship. Together, these results support a model in which ANAC017 integrates transcriptional regulation of mitochondrial retrograde signaling and the UPS with post-translational control of the mitochondrial protein-import system. Through this coordination, ANAC017 helps to maintain mitochondrial biogenesis and energy homeostasis under stress conditions. Positive and negative regulatory interactions are indicated by arrows and flat-ended lines, respectively, and unidentified components are shown in gray with dotted connectors. IMS, intermembrane space; I–V, respiratory chain complexes I to V; NDex and NDin, NAD(P)H dehydrogenases in outside and inside orientations in the inner membrane; Cyt c, cytochrome c; Ub, ubiquitin; E2 and E3, ubiquitin-conjugating enzymes and ubiquitin ligases, respectively.

We demonstrated that TOM proteins accumulated in an anac017 mutant but were depleted in an ANAC017-overexpressing line (Figure 2A). Similarly, TOM-complex proteins accumulated in anac053 and anac078 single and double mutants (Figure 6). These results reveal a novel transcriptional regulatory network that governs the UPS pathway for mitochondrial protein degradation. Given that UPS components participate in the degradation of plastid and peroxisome protein-import machinery (Pan et al., 2018; Ling et al., 2019), it is possible that ANAC017 also regulates the degradation of chloroplast and peroxisome outer-membrane proteins. Indeed, ANAC017 has been linked to chloroplast phenotypes, as it regulates the mitochondrial alternative respiratory pathway, which is essential for efficient photosynthesis (Zhang et al., 2017; Zhu et al., 2023). Moreover, ANAC017 regulates ethylene signaling (He et al., 2022), which is known to regulate chloroplast retrograde signaling (Jiang et al., 2020; Gommers et al., 2021). This suggests that ANAC017 is a crucial player in coordinating the biogenesis and function of key metabolically linked organelles, including chloroplasts, mitochondria, and likely peroxisomes, through both transcriptional and post-transcriptional regulation.

Our data suggest that TOM20 isoforms are ubiquitinated and subsequently degraded by the UPS. Nevertheless, the E3 ligase involved in ubiquitination of the TOM complex remains to be identified (Figure 7). A path forward is to integrate candidate prioritization from our RNA-seq analysis with localization- and domain-based filters typical of plant E3 families to focus on ligases likely resident at, or acting on, the MOM; genetic perturbation of these candidates can then be used to test their effects on TOM20 stability and ubiquitination. Because MRS marker genes are induced in rpn1a mutants (Figure 1E), this suggests that any disruption to proteasome-mediated degradation of TOM-complex proteins leads to mitochondrial dysfunction. This is expected, as changes in TOM-complex abundance may affect their stoichiometry. This imbalance could disrupt transmembrane protein transport in mitochondria, thus affecting biogenesis.

The mechanism by which MRS induction confers tolerance to cellulose biosynthesis inhibitors remains to be determined. Isoxaben treatment does not induce a mitochondrial stress response (Broad et al., 2024), but rather, constitutively induced MRS confers tolerance (Figure 1C and 1D). This tolerance is unlikely to result from reduced TOM20 levels alone. The enhanced tolerance is more likely due to the broader transcriptional reprogramming driven by ANAC017, which activates a wide range of stress-responsive genes in addition to TOM20. One possible explanation is that a detoxification response induced by constitutively active MRS (Meng et al., 2019) counteracts the effects of isoxaben. For instance, genes in the multidrug and toxic compound extrusion (MATE) efflux transporter family are induced by constitutively active MRS (Meng et al., 2019) and could be involved in exporting isoxaben from the cytosol, thereby conferring tolerance. A similar mechanism has previously been shown to confer tolerance to the herbicide paraquat (Xia et al., 2021).

Although a complete MAD pathway has not yet been delineated in plants, parallels from budding yeast provide a useful framework. In yeast, the degradation of TOM70, a TOM receptor with no plant ortholog, occurs through ubiquitination by the cytosolic E3 ligase RSP5. After ubiquitination, TOM70 binds to the ubiquitin-binding domain of DOA1 (Degradation of Alpha 1). DOA1 then interacts with CDC48 and facilitates the extraction of ubiquitinated client proteins from membranes (Wu et al., 2016; Liao et al., 2020). Interestingly, this pathway is specific to TOM70, as the other yeast TOM receptor, TOM22, is not targeted for degradation through this pathway (Wu et al., 2016). This specificity implies that individual MOM proteins may be regulated by different components of the UPS.

Material and methods

Plant materials

The Arabidopsis thaliana Columbia-0 (Col-0; CS70000) accession was used as the wild-type control for all experiments in this study. T-DNA insertion lines for RPN1A (AT2G20580), rpn1a-5 (SALK_127430C) (Wang et al., 2009) and rpn1a-2 (SALK_129604) (Huang et al., 2006); PBE1 (AT1G13060), pbe1-2 (SALK_092686C) (Han et al., 2019) and pbe1-3 (SALK_066170); RPT2A (AT4G29040), rpt2a-3 (SALK_130019C) (Yao et al., 2024); RPN2A (AT2G32730), rpn2a-1 (SALK_088636C); and RPT5A (AT3G05530), rpt5a-1 (SALK_203008C) were obtained from the Arabidopsis Biological Resource Center (ABRC). rao2-1, anac017KO1, ANAC017OE2, and ANAC017OE3 were described previously in Ng et al. (2013) and Meng et al. (2019); anac053-1 (SALK_009578C) and anac078-1 (SALK_025098) were described previously in Gladman et al. (2016) and Han et al. (2019); and aox1a was described previously in Giraud et al. (2008). Primers for T-DNA screening and RT–PCR are listed in Supplemental Data 7.

Plant growth and treatments

In all experiments, Arabidopsis seeds were stratified in the dark for 48 h, then transferred to growth chambers and grown at 22°C under 120 μmol m⁻² s⁻¹ photosynthetic photon flux density with a 16-h light/8-h dark photoperiod. All seedlings were grown vertically in 10-cm petri plates positioned in custom-built acrylic racks to maintain consistent spacing between plates with vertical orientation at a 5° angle for root analysis or grown horizontally for shoot analysis. Seeds were sown on half-strength MS (Murashige and Skoog Basal Salt Mixture) medium (M524, Phytotech Labs) supplemented with 0.1% (v/v) Gamborg’s B5 vitamin solution (G1019, Sigma-Aldrich), 1% (w/v) sucrose (Sigma-Aldrich), and 0.8% (w/v) agar (A7921, Sigma-Aldrich) with the pH adjusted to 5.7. For chemical treatments on plates, 50 μmol AA (A8674, Sigma-Aldrich) was used, with equal amounts of ethanol added to the mock plates to control for a potential solvent effect. Plants for mitochondrial preparations were grown in a liquid culture of half-strength MS in vented transparent containers. For MG132 (#13697 Cayman Chemical [Sapphire Biosciences]) treatment, seeds were grown in liquid culture with shaking at 200 rpm for 6 days, 50 μM MG132 was added to the culture with 0.01% Tween 20 (P9416, Sigma-Aldrich), equal amounts of DMSO (276855, Sigma-Aldrich) were added to the mock control, and seedlings were incubated for another 4 days.

RNA extraction and RT–qPCR

Total plant RNA was extracted using the Spectrum Plant Total RNA Kit (Sigma-Aldrich). DNA contamination was eliminated with the On-Column DNase I (Sigma-Aldrich) digestion kit according to the manufacturer’s instructions. For RT–qPCR, first-strand cDNA was synthesized with the Tetro cDNA Synthesis Kit (Bioline); qPCR was performed using the SensiFAST SYBR & Fluorescein Kit (Bioline) on a QuantStudio 12K Flex Real-Time PCR system (Applied Biosystems) according to the manufacturer’s instructions. The PCR conditions were 95°C for 2 min, followed by 40 cycles of 95°C for 20 s and 60°C for 30 s, using primers listed in Supplemental Data 7. Data were analyzed using QuantStudio 12K Flex software (Applied Biosystems). All experiments were performed with at least three biological replicates consisting of three to four pooled seedlings from independent plates.

Mitochondrial isolation and immunoblot analysis

Mitochondria were isolated from 10-day-old liquid-cultured seedlings as described previously (Lister et al., 2007). Fractions were stored at −80°C and maintained on ice when in use. Protein concentration was determined using the Pierce Coomassie Plus (Bradford) Assay Reagent (#23236, Thermo Scientific). Samples were boiled in SDS–PAGE sample buffer (10% [m/v] SDS, 1% [v/v] β-mercaptoethanol, 18.75% [v/v] glycerol, 0.1% [m/v] bromophenol blue, and 150 mM Tris–HCl [pH 6.8]) for 10 min and resolved on TGX Stain-Free protein gels (Bio-Rad) with Precision Plus Protein WesternC Blotting Standards (#1310376, Bio-Rad) added, and transferred to a Hybond-C Extra nitrocellulose membrane; immunodetection was performed as described previously (Wang et al., 2012). The antibodies used are listed in Supplemental Data 8.

RNA-seq

For RNA-seq, 10-day-old plants were pooled, with four to five seedlings per replicate. Total RNA was extracted as described above from four biological replicates of each genotype. RNA was eluted in molecular-grade DNase- and RNase-free water, and RNA integrity was confirmed using a TapeStation 2200 system (Agilent). RNA-seq libraries were prepared using the TruSeq Stranded mRNA Library Prep Kit (Illumina) according to the manufacturer’s instructions and sequenced on the HiSeq 1500 system (Illumina) as 60-bp reads or on the NextSeq 550 system as 75-bp reads, with an average quality score (Q30) of above 97%. The raw reads (on average, 32 million per sample) were quality controlled using FastQC (https://www.bioinformatics.babraham.ac.uk/projects/fastqc/) and trimmed using TrimGalore (https://github.com/FelixKrueger/TrimGalore). Transcript abundance was quantified at the gene level by pseudo-aligning reads against a k-mer index (k-mer length, 31) built from transcript models for the Araport 11 annotation (Cheng et al., 2017) using the kallisto program with 100 bootstraps, yielding transcripts per million (TPM) values and estimated counts (Bray et al., 2016). These transcript counts were imported and aggregated to the gene level using tximport (Soneson et al., 2015), producing a count matrix for differential expression analysis. DESeq2 was then used (Love et al., 2014), filtering out low-count genes and fitting a negative binomial model across contrasts of interest. Normalized counts were obtained for downstream analyses. Variance-stabilizing transformation was used for clustering and heatmap generation, with genes grouped by hierarchical clustering and visualized in Z-score space. For clusters of differentially expressed genes, Gene Ontology enrichment (biological process) was performed using clusterProfiler (Yu, 2024) with the Arabidopsis TAIR database, and the results were visualized by dot plots and enrichment networks. For UPS gene clusters, functional descriptions were cleaned and used to generate word clouds summarizing key themes. The raw sequence data reported in this paper have been deposited in the Genome Sequence Archive at the National Genomics Data Center, China National Center for Bioinformation/Beijing Institute of Genomics, Chinese Academy of Sciences (GSA: CRA030868) and are publicly accessible at https://ngdc.cncb.ac.cn/gsa.

Total protein extraction and immunoprecipitation

Arabidopsis seedlings or N. benthamiana leaves were ground to a fine powder in liquid nitrogen, and total proteins were extracted in extraction buffer (25 mM Tris–HCl [pH 7.5], 150 mM NaCl, 1 mM EDTA, 10% [v/v] glycerol, 0.5% [v/v] Triton X-100, 1 mM phenylmethylsulfonyl fluoride, and 1× cOmplete protease inhibitor cocktail [Sigma-Aldrich]) on ice before centrifugation at 12 000 × g for 10 min at 4°C twice.

The immunoprecipitation was carried out using FLAG-Trap agarose beads (ffa-20 from ChromoTek) according to the manufacturer’s instructions. In brief, total proteins were lysed in lysis buffer (10 mM Tris–HCl [pH 7.5], 150 mM NaCl, 0.5 mM EDTA, and 0.5% NP-40) with 1× cOmplete protease inhibitor cocktail (Sigma-Aldrich) and 1 mM phenylmethylsulfonyl fluoride and centrifuged at 12 000 × g at 4°C to separate the insoluble pellet from the soluble fraction. The soluble lysates were used as input for further immunoprecipitation. A 30-μl agarose bead slurry was washed in dilution buffer (10 mM Tris–HCl [pH 7.5], 150 mM NaCl, and 0.5 mM EDTA). Soluble lysates were incubated by rotation at 4°C with beads for 1 h, and beads were sedimented by centrifugation at 2500 × g for 5 min at 4°C. The supernatant was saved as the flowthrough/non-bound fraction. The beads were resuspended using wash buffer (10 mM Tris–HCl [pH 7.5], 150 mM NaCl, 0.5 mM EDTA, and 0.05% NP-40), sedimented by centrifugation at 2500 × g for 5 min at 4°C, and washed three times with ice-cold wash buffer. The supernatant was removed, and 80 μl SDS–PAGE sample buffer was added. After boiling for 5 min at 95°C, the immunoprecipitated samples were analyzed by immunoblotting.

In vitro mitochondrial protein import

[³⁵S]-Met-labeled precursor proteins were synthesized using the rabbit reticulocyte TNT in vitro transcription/translation lysate system (Promega) and imported into freshly isolated mitochondria with samples taken at 5, 10, 15, and 20 min, when the import rate had not reached saturation. Equal quantities (50 μg) of mitochondria from different genotypes were used for each import reaction. Following import, mitochondria were precipitated at 20 000 RCF for 5 min and subjected to SDS–PAGE. Gels were stained with Coomassie Brilliant Blue, dried, and exposed to a BAS TR2040 phosphor imaging plate (Fuji) for at least 48 h. The exposed plate was visualized using the BAS 2500 Bio-Imaging Analyzer (Fuji). Mature radiolabeled imported proteins were quantified at each time point and normalized to the highest time point of the wild type, which was set to 1, using Quantity One software (Bio-Rad). Three biological replicates were performed for each mutant and precursor protein. Average import amounts and significant differences relative to that of Col-0 were determined.

Measurements of respiratory parameters of plant mitochondria

Oxygen consumption by isolated mitochondria was measured as described previously (Li et al., 2019) using a Clark-type O₂ electrode (Hansatech, United Kingdom). Oxygen consumption driven by NADH oxidation was measured by adding malate (5 mM) and NADH (1 mM). Rotenone (5 μM) was added to specifically inhibit complex I to determine complex I NADH-driven oxygen consumption; the activity remaining after rotenone addition represented oxygen consumption driven by internal and external NADH dehydrogenases. Complex II–driven oxygen consumption was measured by adding succinate (5 mM) and assessing the malonate (5 mM)-sensitive oxygen consumption rate, with malonate acting as a complex II inhibitor. Activity of the cytochrome c oxidase pathway was measured in a 1-ml reaction volume of aerated respiration medium containing 0.3 M sucrose, 5 mM KH₂PO₄, 10 mM TES, 10 mM NaCl, 2 mM MgSO₄, and 0.1% (w/v) BSA (pH 7.2) in the presence of saturating concentrations of succinate (5 mM) and NADH (1 mM), ATP (0.5 mM), ADP (0.3 mM), malate (10 mM), pyruvate (10 mM), coenzyme A (12 μM), thiamine pyrophosphate (0.2 mM), and NAD⁺ (2 mM). Potassium cyanide (KCN, 1 mM) and n-propyl gallate (0.5 mM) were used to inhibit complex IV and AOX, respectively. Data were analyzed using three biological replicates. Significant differences were determined using Student’s t test (p ≤ 0.05).

Plasmid construction

The ubiquitin coding sequence was amplified from the AtUBQ11 gene (At4g05050), cloned into the NotI–BamHI sites of the pE3n vector (Dubin et al., 2008), and subcloned into the pK7WG2 vector (Karimi et al., 2002) for stable plant transformation (generating the 6×MYC-UBQ construct). The coding sequence of TOM20-3 was amplified with gene-specific primers (Supplemental Data 7) and cloned into the SacI–BamHI sites of pCAMBIA1300-FLAG (Yue et al., 2017) to fuse the 3×FLAG tag, generating TOM20-3-3×FLAG.

To construct vectors for the Dual-LUC reporter assay, the pGreenII 0800-LUC reporter vector and pGreenII 62-SK effector vector were both cut at the BsaI and Eco31I sites (Hellens et al., 2005). The coding sequences of ANAC053, ANAC078, and ANAC017 were amplified using gene-specific primers (Supplemental Data 7) and cloned into the pGreenII 62-SK effector vector, generating 62SK-ANAC053, 62SK-ANAC078, and 62SK-ANAC017. The 1454-bp upstream region of the translational start site of RPN1a (proRPN1a) was amplified by PCR from Arabidopsis Col-0 genomic DNA with specific primers (Supplemental Data 7) and cloned into the pGreenII 0800-LUC reporter vector. All clones were verified by sequencing.

Transient transcription dual-luciferase assay

Transactivation analysis of the RPN1A promoter was performed using the dual-luciferase reporter system in protoplasts extracted from 4-week-old N. benthamiana leaves (Shi et al., 2019). After incubating the transfected protoplasts overnight, the cells were lysed, and luciferase (LUC) activity was measured according to the instructions of the Dual-Luciferase Reporter Assay System (Promega, E1960). The Renilla reniformis luciferase (REN) gene driven by CaMV 35S was used as an internal control. Promoter activity was expressed as the ratio of LUC to REN, and the analysis was repeated three times.

For observation of LUC activity, an Agrobacterium strain carrying a transcription factor or promoter was resuspended in infiltration medium (10 mM MgCl₂ and 200 μM acetosyringone) and mixed when OD₆₀₀ reached 0.8. The mixtures were then injected into tobacco leaves using a needless syringe. After 3 days, the transformed tobacco leaves were sprayed with D-luciferin sodium salt and observed using a Tanon 5200 imaging apparatus. Three independent measurements were performed.

ROS measurements

Ten-day-old seedlings were stained for H₂O₂ and O₂^·− as described previously (Ivanova et al., 2014).

UPS gene list

A list of UPS-related genes in Arabidopsis was compiled from genome annotations and the literature by combining 26S proteasome genes (Book et al., 2010) encoding DUBs (de-ubiquitinating enzymes, ubiquitin-specific proteases), E1s (ubiquitin-activating enzymes, UBAs), E2s (ubiquitin-conjugating enzymes, UBCs), and E3 ligases from the RING (Really Interesting New Gene) family (Stone et al., 2005), HECT (Homologous to E6-AP C-Terminus) family (Stone et al., 2005), U-box family (Wiborg et al., 2008), and CRL (Cullin–RING ligase) families (including the SCF–F-box [Skp1–Cullin1–F-box] family (Hua et al., 2011), CUL3–BTB [Cullin3–Bric-a-brac/Tramtrack/Broad complex] type (Gingerich et al., 2005), CUL4–DDB1–DWD [Cullin4–DNA Damage-Binding1–DDB1-binding WD40] type (Lee et al., 2008), and APC/C [Anaphase-Promoting Complex/Cyclosome] type (Capron et al., 2003)). Genes encoding members of the SUMO (Small ubiquitin-related modifier) family and ubiquitin-like (UBL) family were retrieved from the MapMan (Thimm et al., 2004) annotation set labeled “UBQ” (Supplemental Data 1).

Funding

This work was supported by the Zhejiang Provincial Government Kun Peng Fellowship and the Yangtze River Scholar Fellowship awarded to J.W. G.A.K. was supported by a DECRA Fellowship from the Australian Research Council (DE210101200). S.P. was supported by a Villum Foundation grant, two Novo Nordisk Foundation grants, and a Danish National Research Foundation grant (25915, 19OC0056076, 20OC0060564, and DNRF155, respectively).

Acknowledgments

We thank Prof. Jianxiang Liu (College of Life Science, Zhejiang University) for providing the ANAC053 and ANAC078 mutant lines. No conflict of interest declared.

Author contributions

J.W., G.A.K., Y.W., and Y.Z. conceived, designed, and supervised the study. Y.Z. carried out most of the experiments. M.O., S.P., and G.A.K. performed the initial forward genetic screen to identify the UPS components analyzed in this study. M.M. and Y.W. contributed to the analysis of mitochondrial protein abundance. G.A.K., J.W., and Y.Z. wrote the manuscript. All authors contributed to data interpretation and to the design of subsequent experiments.

Published: December 10, 2025