Coordinated control of proteasome subunit gene expression promotes stress resistance, proteostasis, and longevity

Abstract

The proteasome is essential for cellular protein homeostasis through selective destruction of damaged and misfolded proteins. Failure of proteasome-dependent turnover accompanied by accumulation and aggregation of aberrant proteins is a hallmark of aging and late-onset neurodegenerative diseases. SKN-1A/Nrf1, a member of the NFE2L/Nrf family of transcription factors, is a master regulator of proteasome biogenesis. Through transcriptional control of proteasome subunit gene expression, SKN-1A/Nrf1 controls homoeostatic and stress-responsive upregulation of proteasome levels in adaptation to proteasome dysfunction or protein misfolding. SKN-1A/Nrf1 acts in concert with another Nrf family transcription factor, SKN-1C/Nrf2, to regulate many aspects of physiology including stress responses, redox balance, immunity, and metabolism. Here, we demonstrate that a small deletion in the promoter of the pbs-5 gene, which encodes an essential proteasome subunit, uncouples its expression from transcriptional regulation by SKN-1A/Nrf1. This disruption leads to compensatory SKN-1A/Nrf1-dependent upregulation of other proteasome subunit genes, resulting in a homeostatic imbalance in proteasomal gene expression. This pbs-5 regulatory mutation phenocopies some, but not all, aspects of SKN-1A/Nrf1 inactivation, providing evidence that coordinated regulation of proteasomal subunit gene expression underlies a subset of SKN-1A/Nrf1’s physiological roles. In comparing the effects of the pbs-5 promoter deletion with isoform-specific inactivation of SKN-1A or SKN-1C, we show that the pbs-5 promoter mutation completely abrogates multiple lifespan extension paradigms. These results reveal that coordinated homeostatic regulation of proteasome subunit gene expression is critical for longevity and healthy aging.

Supplementary Information

The online version contains supplementary material available at https://doi.org/10.1007/s11357-026-02146-6.

Introduction

The ubiquitin proteasome system (UPS) is the major mechanism of selective protein degradation in eukaryotic cells [1–5]. The UPS plays a critical role in the maintenance of protein homeostasis through degradation of damaged and misfolded proteins in a timely and controlled manner. Degradation of ubiquitinated proteins is carried out by the 26S proteasome, an elaborately regulated multi-subunit protease complex [2–5]. Proteasome function is tightly regulated at multiple levels, including transcription, assembly, post-translational modification, and localization [2, 6]. Disruption of proteasome function impairs protein homeostasis, leading to accumulation and aggregation of abnormal proteins, a hallmark of aging and almost all age-dependent neurodegenerative diseases [7]. Understanding the cellular mechanisms that safeguard proteostasis through regulation of proteasome levels, assembly, localization, and activity may provide a basis for therapeutic strategies targeting age-related diseases.

In animal cells, the SKN-1A/Nrf1 transcription factor controls proteasome levels through transcriptional control of proteasome subunit genes [8, 9]. The levels and activity of SKN-1A/Nrf1 are regulated to ensure precise homeostatic and stress-responsive regulation of proteasome subunit expression. SKN-1A/Nrf1 levels and activity are increased under conditions of proteotoxic stress, such as proteasome inhibition or increased protein misfolding and aggregation. In these cases, SKN-1A/Nrf1 upregulates proteasome subunit gene expression to drive increased proteasome levels, thereby enabling cells to compensate for proteotoxic insults. As a result, inactivation of the SKN-1A/Nrf1 pathway in C. elegans causes sensitivity to the toxic effects of proteasome inhibition or widespread protein misfolding [8, 10]. In mice, inactivation of the SKN-1A/Nrf1 pathway causes progressive neurodegeneration [11–13]. Consistent with a protective role, genetic or pharmacological interventions that increase SKN-1A/Nrf1 activity ameliorate disease phenotypes in neurodegenerative disease models [14–17].

SKN-1A/Nrf1 belongs to the nuclear factor erythroid 2-like (SKN-1/NFE2L/Nrf) family of stress responsive Cap’n’Collar (CnC) basic leucine zipper (bZip) transcription factors. In C. elegans, this family consists of distinct protein isoforms encoded by the skn-1 gene [18, 19]. The two major isoforms are SKN-1A and SKN-1C, which are expressed in most or all tissues [20]. The third isoform, SKN-1B, is expressed in a pair of sensory neurons and mediates physiological and behavioral responses to food availability [21, 22]. In mammals, Nrf1 (also termed NFE2L1) and Nrf2 (also termed NFE2L2) are the functional counterparts of SKN-1A and SKN-1C, respectively [15, 20]. There is not a clear functional counterpart of SKN-1B in mammals. Although they recognize the same DNA sequence motif, SKN-1A/Nrf1 and SKN-1C/Nrf2 perform both distinct and overlapping functions in maintaining cellular homeostasis and responding to stress [23, 24]. Regulation of proteasome levels is solely a function of SKN-1A/Nrf1 [8, 10]. In contrast, SKN-1C/Nrf2 is a master regulator of oxidative stress responses and xenobiotic detoxification, a role that SKN-1A/Nrf1 can also support under certain conditions [20, 25–30]. skn-1 isoforms are also implicated in the control of lipid metabolism, regulation of feeding behaviors, innate immunity, and metabolic adaptation to diet [21, 31–35]. Analysis of isoform-specific alleles of the skn-1 gene in C. elegans provides a useful framework to dissect distinct and overlapping functions of Nrf family transcription factors in animals.

Precise control of SKN-1A/Nrf1 and SKN-1C/Nrf2 functions is critical for longevity. Moderate overexpression of either SKN-1A or SKN-1C increases lifespan in otherwise wild type C. elegans, indicating that both forms of SKN-1/Nrf are competent to extend lifespan [16, 36, 37]. Mutations that cause SKN-1 hyperactivation cause increased stress resistance and, in some cases, extend lifespan. However, in other cases, such mutations lead to reduced lifespan or prevent lifespan extension, despite increased stress resistance [32, 36–40]. Thus, although SKN-1/Nrf hyperactivity is protective under stress, it can be detrimental during aging. Notably, skn-1 inactivation abrogates the effects of genetic or dietary interventions that normally cause increased longevity [36, 41, 42], suggesting that tightly controlled SKN-1-mediated changes in gene expression are important drivers of lifespan extension. However, the underlying transcriptional programs are not fully understood, partly because many genes with diverse functions are responsive to changes in SKN-1/Nrf activity [20, 35, 43–45]. Additional complexity is created by the possibility for redundant or cooperative regulation of some genes by multiple SKN-1/Nrf family members. Identifying critical transcriptional targets of SKN-1/Nrf family transcription factors that are required for lifespan extension offers a basis to explore their therapeutic potential in promoting healthy aging or treating late-onset diseases.

Here, we characterize a deletion mutation in the promoter region upstream of the pbs-5 gene, which encodes the β5 subunit of the proteasome [10]. The β5 subunit is one of the three catalytic subunits of the 20S core particle of the proteasome, is required for proteasome assembly and function, and is the target of proteasome inhibitors used in cancer treatment [46–48]. We show that this mutation disrupts transcriptional regulation of pbs-5, uncoupling this gene from control by SKN-1A/Nrf1, which profoundly disrupts homeostatic control of proteasome levels and activity. We use this mutant as a genetic tool to evaluate the role of proteasome homeostasis in stress responses and lifespan extension. Our results demonstrate that coordinated control of proteasome subunit expression is essential for some, but not all, stress-responsive functions of SKN-1A. Strikingly, uncoupling pbs-5 from control by SKN-1A blocks multiple lifespan extension interventions. Our results reveal the physiological significance of precisely coordinated control of proteasome subunit gene expression in animals, including critical roles in protein homeostasis and longevity.

Results

The mg502 allele is a promoter mutation that uncouples the PBS-5/β5 proteasome subunit from SKN-1A/Nrf1-dependent transcriptional control

In a previous screen for misregulation of the SKN-1A/Nrf1 pathway, we identified the mg502 allele, which causes constitutive, skn-1a-dependent activation of rpt-3_p::gfp, a reporter of proteasome subunit gene expression [10]. We determined that mg502 is a 122 bp deletion in the promoter region of the CEOP1752 operon [10], which includes the genes K05C4.2 and pbs-5 (Fig. 1A). Since the mg502 lesion is in the promoter of the operon, the resulting defect in rpt-3_p::gfp expression likely stems from disrupted expression of one or both genes. pbs-5 encodes the β5 subunit of the 20S proteasome [49] and K05C4.2 encodes the C. elegans ortholog of JAGN1/jagunal, an integral membrane protein of the endoplasmic reticulum (ER) required for organization of the ER and regulation of protein trafficking [50–52]. SKN-1 binds to a short DNA motif termed the antioxidant response element (ARE) present in the promoters of many SKN-1-regulated genes, including proteasome subunits such as pbs-5 [24, 53, 54]. Interestingly, the mg502 deletion removes two AREs that are necessary for SKN-1-dependent upregulation of pbs-5 transcriptional reporters in response to stress [54] (Fig S1). The AREs coincide with a SKN-1 binding peak detected in ChIP-seq experiments, suggesting that SKN-1 binds to these elements [53]. Collectively, these findings suggest that the mg502 lesion, via removal of ARE elements in the COP1752 promoter, alters pbs-5 mRNA expression and disrupts its stress-responsive upregulation by SKN-1A. Accordingly, we will hereafter refer to animals harboring the mg502 lesion as pbs-5(∆ARE) mutants.

Fig. 1.

The pbs-5(mg502[∆ARE]) mutation disrupts regulation of proteasome subunit gene expression. A Schematic of CEOP1752 showing the location of the mg502 deletion. Antioxidant response elements (AREs; SKN-1 binding motifs) are indicated by red arrows. SKN-1 CHIP (MODENCODE) Data are from the Wormbase jbrowse2 track “SKN-1 Combined (GFP ChIP) recalled peaks w peaks”. B Quantitative real-time PCR showing that pbs-5 mRNA levels are reduced in pbs-5(∆ARE) mutant animals and that induction of pbs-5 mRNA following bortezomib exposure is lost. Error bars show mean ± SD. Ordinary two-way ANOVA. *** p < 0.001, * p < 0.05, ns p > 0.05. C Fluorescence micrographs showing expression of the rpt-3_p::gfp proteasome subunit reporter in control (DMSO treated) animals and following proteasome inhibition by treatment with 1 μM BTZ. L4 Animals were transferred to DMSO or drug-supplemented plates for 20 h prior to imaging. Constitutive activation of rpt-3_p::gfp expression in pbs-5(∆ARE) mutants is rescued by a construct that drives expression of PBS-5 under the CEOP1752 promoter (CEOP1752_p::PBS-5, see Fig S2). Scale bar shows 100 μm. D Quantification of rpt-3_p::gfp expression shown in (C). n > 20 animals were measured per genotype/condition. Error bars show mean ± SD. Ordinary two-way ANOVA with Šidák’s multiple comparisons test. **** p < 0.0001, ns p > 0.05. E Fluorescence micrographs showing expression of an mCherry::H2B transcriptional reporter knocked into the endogenous rpt-3 locus. L4 Animals were transferred to DMSO or drug-supplemented plates for 20 h prior to imaging. BTZ-responsive upregulation of rpt-3 is lost in skn-1a mutants, whereas rpt-3 is constitutively upregulated in pbs-5(∆ARE) mutants. Scale bar shows 100 μm. F Quantification of the mCherry::H2B rpt-3 reporter expression shown in (E). n > 25 animals were measured per genotype/condition. Error bars show mean ± SD. Ordinary two-way ANOVA with Bonferroni’s multiple comparisons test. **** p < 0.0001, *** p < 0.001, ** 0.01, ns p > 0.05. G Fluorescence micrographs showing expression of the gst-4_p::gfp oxidative stress response reporter in control (DMSO treated) animals and following proteasome inhibition by treatment with 2.4 μM BTZ. L4 Animals were transferred to DMSO or drug-supplemented plates for 20 h prior to imaging. gst-4_p::gfp expression is not constitutively upregulated in pbs-5(∆ARE) animals. Scale bar shows 100 μm. H Quantification of gst-4_p::gfp expression shown in (G). n = 20 animals were measured per genotype/condition. Error bars show mean ± SD. Ordinary two-way ANOVA. **** p < 0.0001, ** 0.01, ns p > 0.05

To examine whether the pbs-5(∆ARE) mutation interferes with regulation of pbs-5 expression, we compared pbs-5 mRNA levels in wild type and pbs-5(∆ARE) mutant animals. RNA was extracted from animals exposed to a DMSO control or to the proteasome inhibitor bortezomib (BTZ), which triggers SKN-1A-dependent upregulation of proteasome subunit genes, including pbs-5 [15, 55]. In pbs-5(∆ARE) animals, pbs-5 mRNA levels are moderately reduced compared to the wild type under control conditions (Fig. 1B). In the wild-type, exposure to BTZ leads to a ~ twofold induction of pbs-5 mRNA, but this response is completely lost in pbs-5(∆ARE) mutants (Fig. 1B). We therefore conclude that the pbs-5(∆ARE) mutation (1) reduces the basal expression of pbs-5 mRNA and (2) abrogates stress-responsive upregulation of the pbs-5 gene. Since transcriptomic and reporter transgene assays indicate that SKN-1A is necessary to promote both basal expression of proteasome subunit genes as well as their stress-responsive upregulation [10, 15, 16], these data are consistent with a model in which deletion of the AREs from the CEOP1752 promoter uncouples pbs-5 from transcriptional regulation by SKN-1A.

Uncoupling PBS-5/β5 from SKN-1A/Nrf1-mediated activation causes hyperactivation of proteasome subunit gene expression

Since K05C4.2 is co-expressed with pbs-5 in the CEOP1752 operon, we expect that K05C4.2 expression is also reduced in pbs-5(∆ARE) mutant animals. Because SKN-1A/Nrf1 function requires trafficking through the ER [10, 56, 57], it is possible that reduced K05C4.2 levels could lead to misregulation of SKN-1A. To address this possibility, we generated a plasmid construct in which only the PBS-5 proteasome subunit is driven from the wild type CEOP1752 promoter (CEOP1572_p::PBS-5; Fig S2). When introduced as a high-copy extrachromosomal array, the construct fully rescues rpt-3_p::gfp hyperactivation of pbs-5(∆ARE) mutants (Fig. 1C,D). Stress-responsive up-regulation of rpt-3_p::gfp is normal in the rescued animals, suggesting that misregulation of the SKN-1A/Nrf1 signaling axis in pbs-5(∆ARE) animals can be entirely attributed to reduced pbs-5 expression, whereas there is no effect stemming from reduced K05C4.2/Jagunal expression. Notably, although the rescuing array is high-copy, rescued animals show normal basal and stress-responsive regulation of the rpt-3_p::gfp reporter. This suggests that whereas reduced pbs-5 expression leads to SKN-1A hyperactivation, overexpression of pbs-5 does not lead to reduced SKN-1A activity. We conclude from these results that hyperactivation of the rpt-3_p::gfp proteasome subunit reporter in pbs-5(∆ARE) mutants is caused by reduced expression of the pbs-5 proteasome subunit gene, which in turn leads to constitutive SKN-1A-dependent hyperactivation of the rpt-3_p::gfp proteasome subunit reporter.

It is somewhat counterintuitive that reduced expression of pbs-5 triggers increased expression of a reporter for the rpt-3 proteasome subunit gene, since stoichiometrically balanced expression of proteasome subunits should be required for proper proteasome assembly [58]. We therefore sought to confirm that endogenous rpt-3 is upregulated in pbs-5(∆ARE) animals. We used CRISPR/Cas9 to insert an mcherry::H2B::SL2 cassette immediately upstream of the endogenous rpt-3 open reading frame, such that the expression of mcherry::H2B provides a fluorescent readout of endogenous rpt-3 gene transcription (Fig S3). Using this reporter, we confirmed that rpt-3 transcription is induced approximately 1.5-fold following exposure to BTZ (Fig. 1E,F). In animals lacking SKN-1A, the basal level of expression of rpt-3 is reduced, and there is no increase in expression following BTZ exposure, consistent with our previous analyses with rpt-3_p::gfp [10, 15]. These data are also consistent with transcriptomic studies that reveal that proteasome subunits are induced by 1.5 to twofold following exposure of C. elegans to BTZ or SKN-1A activation, and that basal expression of proteasome subunits is reduced in animals lacking SKN-1A [20, 35, 55, 59]. In pbs-5(∆ARE) animals, rpt-3 expression is elevated compared to wild type, reaching levels higher than those induced by BTZ in the wild type. This basally elevated expression is only slightly further increased when pbs-5(∆ARE) mutant animals are exposed to BTZ (Fig. 1E,F), again consistent with the rpt-3_p::gfp reporter (Fig. 1C,D). These data confirm that endogenous rpt-3 expression is controlled by SKN-1A similarly to the rpt-3_p::gfp reporter transgene. We conclude that rpt-3 is constitutively elevated by SKN-1A in pbs-5(∆ARE) mutants due to reduced expression of the PBS-5/β5 subunit of the proteasome.

A transcriptional reporter for the glutathione-S-transferase gene gst-4 (gst-4_p::gfp), is activated in a skn-1-dependent manner in response to various stressors [60–64]. In response to oxidative stress, gst-4_p::gfp is activated in a largely SKN-1C-dependent manner [20]. In contrast, gst-4_p::gfp is activated in a SKN-1A-dependent manner under conditions of proteasome impairment [15, 20]. We did not observe hyperactivation of gst-4_p::gfp in pbs-5(∆ARE) mutants (Fig. 1G,H), suggesting that the pbs-5(∆ARE) mutation does not cause hyperactivation of the SKN-1C-mediated oxidative stress response. Additionally, although SKN-1A is responsible for hyperactivation of rpt-3_p::gfp in pbs-5(∆ARE), SKN-1A-mediated hyperactivation of gst-4_p::gfp does not take place. gst-4_p::gfp levels are induced following proteasome inhibition by BTZ in pbs-5(∆ARE) mutants, suggesting that the SKN-1A-mediated activation of gst-4_p::gfp in response to proteasome inhibition is not compromised. Thus, the pbs-5(∆ARE) mutation appears to cause constitutive activation of a SKN-1A-mediated response involving activation of proteasome subunit genes, but not other SKN-1-mediated stress responses.

Uncoupling PBS-5/β5 from SKN-1A/Nrf1 causes increased levels and abnormal subcellular localization of proteasome subunits

Activation of SKN-1A/Nrf1 leads to increased proteasome levels, which is achieved through coordinated upregulation of all proteasome subunit genes [20, 56, 65]. Our data suggest that this coordinated control of proteasome subunit gene expression is disrupted in pbs-5(∆ARE) mutants, since transcription of pbs-5 is reduced, whereas transcription of rpt-3 is elevated. To better understand the effect of failure to correctly coordinate the expression of pbs-5 with that of rpt-3 (and presumably all other proteasome subunit genes that are transcriptionally controlled by SKN-1A), we analyzed the abundance of proteasome subunit proteins in pbs-5(∆ARE) mutants. We examined CRISPR/Cas9 engineered strains in which endogenous proteasome subunits are fused to GFP. We used two strains, one in which GFP is fused to PBS-4 (the β4 subunit of the 20S core particle; pbs-4::GFP), and one in which GFP is fused to RPN-8 (a subunit of the 19S regulatory particle; GFP::rpn-8) [66]. The abundance of both proteasome subunits is strikingly increased in pbs-5(∆ARE) mutant animals (Fig. 2A-D). Notably, in the skn-1a mutant background, the basal level of each proteasome subunit is reduced and the uncreased abundance caused by the pbs-5(∆ARE) mutation is lost. We conclude that reduced expression of the PBS-5/β5 proteasome subunit gene leads to increased levels of other proteasome subunits via the SKN-1A/Nrf1 pathway.

Fig. 2.

The pbs-5(ΔARE) mutation causes overproduction, mislocalization, and processing defects in proteasome subunit proteins. A, B Fluorescence micrographs showing abundance of endogenously tagged A PBS-4::GFP and B GFP::RPN-8 in L4 stage animals. Scale bar shows 100 μm. C, D Quantification of endogenously tagged proteasome subunit levels from (A) and (B). n ≥ 20 animals were measured per genotype. Proteasome subunit abundance is increased in pbs-5(∆ARE) animals in a skn-1a-dependent manner. Error bars show mean ± SD. Ordinary one-way ANOVA with Tukey’s multiple comparisons test. **** p < 0.0001, ns p > 0.05. E, F Higher magnification fluorescence micrographs showing differences in subcellular distribution of proteasomes within intestinal cells of L4 animals of each genotype. The brightness and contrast of each fluorescence image is adjusted to compensate for the increased fluorescence of the pbs-5(∆ARE) mutant animals, enabling comparison of localization patterns. Scale bar shows 10 μm. G, H Quantification of nuclear enrichment of endogenously tagged proteasome subunits from (E) and (F). Nuclear enrichment of proteasome subunits is diminished in pbs-5(∆ARE) mutants but are unchanged in either skn-1a mutants or pbs-5(∆ARE); skn-1a double mutants. Nuclear enrichment was measured in n > 25 intestinal cells per genotype. Error bars show mean ± SD. Ordinary one-way ANOVA with Tukey’s multiple comparisons test. **** p < 0.0001, ns p > 0.05. I Fluorescence micrographs showing subcellular distribution of proteasomes within intestinal cells of wild type L4 stage animals exposed to either DMSO control or 1 μM BTZ throughout life. Brightness and contrast of each fluorescence image is adjusted to compensate for the increased fluorescence of BTZ-exposed animals and allow comparison of localization patterns. Scale bar shows 10 μm. J, K Quantification of nuclear enrichment of endogenously tagged proteasome subunits from (I). Nuclear enrichment was measured in n > 25 intestinal cells per condition. Error bars show mean ± SD. Welch’s t test. ns p > 0.05. L Western blot showing processing of endogenously 3xFLAG-tagged PBS-2. Lysates for this experiment were prepared from stage-synchronized young adult animals. PBS-2 precursor and mature (proteolytically processed) PBS-2 are indicated by red and blue arrows respectively. Accumulation of PBS-2 precursor is dramatically increased in pbs-5(∆ARE) mutants, but this effect is lost in pbs-5(∆ARE); skn-1a double mutants

While carrying out these experiments we made the unexpected observation that the subcellular localization of proteasome subunit proteins is dramatically altered in the pbs-5(∆ARE) mutants. Proteasomes are enriched in the nuclei of most cell types in C. elegans and other eukaryotes including humans and yeast [66–71]. Consistently, in the wild type, both PBS-4::GFP and GFP::RPN-8 show nuclear enrichment in all cell types, which is particularly apparent in the large nuclei of the gut cells (Fig. 2E-H). This nuclear enrichment remains intact in skn-1a mutants (Fig. 2E-H). However, this enrichment is completely lost in the pbs-5(∆ARE) mutants, which show equal distribution between cytosol and nucleus of intestinal cells (Fig. 2E-H). Strikingly, this altered localization does not occur in in pbs-5(∆ARE); skn-1a double mutants. This result suggests that the correct subcellular localization of proteasome subunits depends on the stoichiometric expression of pbs-5 relative to other proteasome subunit genes. Alternatively, since SKN-1A is hyperactive in the pbs-5(∆ARE) mutants, it may indicate that chronic hyperactivation of SKN-1A/Nrf1 signaling drives altered subcellular localization of proteasomes. To test the latter hypothesis, we examined the subcellular localization of GFP-tagged proteasome subunits in animals in which the SKN-1A pathway is chronically activated by exposure to BTZ throughout development. Chronic BTZ exposure causes a similar ~ twofold increase in PBS-4::GFP and GFP::RPN-8 abundance as we observed in the pbs-5(∆ARE) mutants (Fig S4). However, there is no reduction in their nuclear enrichment (Fig. 2I-K). These data suggest two key points: (1) simply increasing the levels of proteasomal subunits through SKN-1A activation is not sufficient to diminish their nuclear enrichment and (2) the altered proteasome subunit localization in pbs-5(∆ARE) animals cannot be attributed solely to a reduction in PBS-5/β5 function, since BTZ, which inhibits proteasome function by blocking the PBS-5/β5 proteolytic active site [72, 73], does not cause loss of nuclear enrichment. We conclude that uncoupling pbs-5 gene expression from regulation by SKN-1A leads to dramatic imbalance of proteasome subunit gene expression and mislocalization of overproduced subunits.

Uncoupling PBS-5/β5 from SKN-1A/Nrf1 causes defective proteasome assembly

Proteasomes are assembled in the cytosol prior to transport into the nucleus [69–71]. The loss of nuclear enrichment of proteasomal components could indicate that either proteasome assembly and/or nuclear import of assembled proteasomes is compromised in pbs-5(∆ARE) mutant animals. The pbs-5/β5 subunit of the proteasome is required for assembly of the 20S proteasome subcomplex, so reduced pbs-5 expression is likely to limit 20S assembly [47, 48]. Indeed, reduced transcription of the β5 subunit of the proteasome (PSMB5) in humans was recently shown to disrupt proteasome assembly [74, 75]. The β1 and β2 subunits of the 20S proteasome are expressed as precursors that undergo maturation via auto-proteolytic removal of an N-terminal segment during proteasome assembly [47, 48]. Accumulation of β1 and β2 precursors is a characteristic defect in human cells with inadequate PSMB5/β5 transcription [74, 75]. To test whether pbs-5(∆ARE) mutant animals have a similar proteasome assembly defect, we generated CRISPR/Cas9 engineered strains in which endogenous PBS-2/β2 is fused to 3xFLAG at its C-terminus. By western blot, we detected a dramatic accumulation of PBS-2 precursors in pbs-5(∆ARE) mutants (Fig. 2L). Importantly, we did not detect this accumulation of PBS-2 precursors in pbs-5(∆ARE); skn-1a double mutants. We conclude that uncoupling pbs-5 gene expression from regulation by SKN-1A leads to defective proteasome assembly, likely driven by a stochiometric imbalance between reduced PBS-5/β5 expression and overproduction of other proteasome subunits.

Uncoupling PBS-5/Β5 from SKN-1A/Nrf1 disrupts proteasome function and the response to proteasome inhibition

Our findings so far suggest that the pbs-5(∆ARE) mutant presents an interesting case: although the SKN-1A pathway is hyperactive, expression of an essential proteasome subunit gene is reduced, resulting in overproduced subunits that show abnormalities in subcellular localization and assembly into proteasomes. However, it remains unclear whether -and how- this mutation impacts proteasome function. Inactivation of SKN-1A/Nrf1 per se causes impaired proteasome function as measured by degradation of an unstable Ub[G76V]::GFP reporter [15]. In contrast, hyperactivation of SKN-1A can enhance proteasome function [15, 16]. We find that accumulation of Ub[G76V]::GFP is increased in pbs-5(∆ARE) animals, indicating that proteasome function is impaired (Fig. 3A, B). This phenocopies the defect observed in skn-1a mutants (Fig. 3A,B), suggesting that uncoupling pbs-5 from regulation by SKN-1A is sufficient to induce a functional deficit equivalent to SKN-1A loss. These data support the model in which SKN-1A/Nrf1 promotes proteasome function through coordinated activation of all proteasome subunit genes [65]. The reduced expression of PBS-5/β5 is clearly not compensated by the strong upregulation of other proteasomal components, consistent with failure to incorporate upregulated proteasome subunits into functional proteasomes. Interestingly, accumulation of Ub[G76V]::GFP is further increased in pbs-5(∆ARE); skn-1a double mutants (Fig. 3A-D), suggesting that although pbs-5 is not coordinated with other proteasome subunit genes, SKN-1A retains some residual capacity to promote proteasome function.

Fig. 3.

Proteasome function and responses to proteasome inhibition are disrupted in pbs-5(∆ARE) mutants. A Fluorescence micrographs of animals expressing the proteasome substrate Ub[G76V]::GFP. Accumulation of the substrate is increased in skn-1a and pbs-5(∆ARE) mutant animals. Scale bar shows 100 μm. B Quantification of Ub[G76V] accumulation shown in (A). Ub[G76V]::GFP was quantified in n > 20 animals of each genotype. Error bars show mean ± SD. Ordinary one-way ANOVA with Tukey’s multiple comparisons test. **** p < 0.0001, *** p < 0.001. C Fluorescence micrographs showing accumulation of Ub[G76V]::GFP in animals exposed to the proteasome inhibitor BTZ. Accumulation of the substrate is increased in skn-1a and pbs-5(∆ARE) mutant animals. Scale bar shows 100 μm. D Quantification of Ub[G76V] accumulation shown in (C). Ub[G76V]::GFP was quantified in n ≥ 20 animals for each genotype/condition. Error bars show mean ± SD. Ordinary one-way ANOVA with Šidák’s multiple comparisons test. **** p < 0.0001, *** p < 0.001, ns p > 0.05. E Micrographs showing development of animals exposed to DMSO control or 1 μM BTZ throughout development. Scale bar shows 500 μm. F Survival of animals exposed to 1 μM BTZ showing that sensitivity of pbs-5(∆ARE) mutants is rescued by a construct that expresses PBS-5 under the CEOP1752 promoter. Late L4 stage animals were placed on plates supplemented with 1 μM BTZ and survival checked after 4 days. Results of n = 3 replicate experiments are shown, at least 20 animals’ survival was tested in each replicate. Error bars show mean ± SD. Ordinary one-way ANOVA with Šidák’s multiple comparisons test. **** p < 0.0001, ns p > 0.05. G Survival of animals exposed to 50 nM BTZ showing that sensitivity of pbs-5(∆ARE) mutants is further enhanced by skn-1a inactivation. Late L4 stage animals were placed on plates supplemented with 50 nM BTZ and survival was checked after 3 days. Results of n = 3 replicate experiments are shown, at least 20 animals’ survival was tested in each replicate. Error bars show mean ± SD. Ordinary one-way ANOVA with Tukey’s multiple comparisons test. **** p < 0.0001, *** p < 0.001, ns p > 0.05. H Survival of animals exposed to 1 μM BTZ showing that sensitivity of pbs-5(∆ARE) mutants is partially rescued by SKN-1A overexpression (o/e). Late L4 stage animals were placed on plates supplemented with 1 μM BTZ and survival was checked after 4 days. Results of n = 3 replicate experiments are shown, at least 20 animals’ survival was tested in each replicate. Error bars show mean ± SD. Ordinary one-way ANOVA with Šidák’s multiple comparisons test. **** p < 0.0001, ** p < 0.01

Induction of proteasome biogenesis by SKN-1A/Nrf1 is an essential cellular and organismal adaptation to proteotoxic stress [9]. Animals lacking SKN-1A are thus highly sensitive to developmental inhibition and killing by low concentrations of BTZ, whereas hyperactivation of SKN-1A increases BTZ resistance [10, 15]. pbs-5(∆ARE) animals phenocopy the severe proteasome inhibitor sensitivity caused by inactivation of skn-1a (Fig. 3E, F). This phenotype is fully rescued by extrachromosomal array transgenes that drive expression of pbs-5 under the wild type CEOP1752 promoter (Fig. 3E, F), confirming that reduced pbs-5 expression, but not altered expression of K05C4.2/Jagunal, causes proteasome inhibitor sensitivity. Notably, pbs-5(∆ARE) mutants are hypersensitive to BTZ even though SKN-1A/Nrf1 is hyperactive and expression of proteasome subunits (other than pbs-5) is elevated. Thus, concerted upregulation of pbs-5 alongside the other proteasome subunit genes is required for adaptation to proteasome inhibition. Interestingly, pbs-5(∆ARE); skn-1a double mutants are more sensitive to BTZ than either the pbs-5 or skn-1a single mutants (Fig. 3G). Thus, although the response is severely compromised, SKN-1A/Nrf1 retains some residual capacity to promote proteasome inhibitor resistance through a mechanism that does not involve upregulation of pbs-5.

To further address the possibility that SKN-1A can promote BTZ resistance through a mechanism that does not depend on upregulation of pbs-5, we measured the effect of SKN-1A overexpression in wild type and pbs-5(∆ARE) mutant animals. SKN-1A overexpression significantly increases the survival of pbs-5(∆ARE) mutant animals in the presence of a low concentration of BTZ (Fig. 3H), suggesting that SKN-1A can promote BTZ resistance in a pbs-5 upregulation-independent manner. However, the survival of pbs-5(∆ARE); skn-1a(o/e) animals remains reduced compared to wild type and is drastically lower than that of skn-1a(o/e) animals, which are normally more resistant than the wild type [15]. Thus, whilst SKN-1A retains a residual capacity to promote BTZ resistance without upregulation of pbs-5, the SKN-1A pathway primarily promotes inhibitor resistance through a mechanism that requires pbs-5 upregulation. These data support a model in which the SKN-1A/Nrf1 pathway drives proteasome inhibitor resistance through compensatory proteasome biogenesis, which requires activation of all proteasome subunit genes’ transcription [8, 10, 56, 65].

Uncoupling PBS-5/β5 from regulation by SKN-1A/Nrf1 minimally impairs oxidative stress resistance

Responses to oxidative stress are coordinately controlled by both the SKN-1A and SKN-1C isoforms of SKN-1 [20, 30]. SKN-1C serves as the primary mediator of these responses [15, 20]. Inactivation of SKN-1A per se does not cause hypersensitivity to oxidants, but animals lacking both SKN-1A and SKN-1C are much more sensitive than animals lacking SKN-1C only [20]. In this context, SKN-1A may enhance oxidative stress resistance by upregulating biogenesis of proteasomes, which are important for cellular clearance of oxidatively damaged proteins [76], and/or upregulating the same SKN-1C-regulated genes involved in redox balance and detoxification [20]. The relative contributions of these two regulatory outputs of SKN-1A to this response are not clear. To address this, we examined how the pbs-5(∆ARE) mutation affects oxidative stress resistance, since it severely disrupts SKN-1A-mediated regulation of proteasome biogenesis but does not interfere with SKN-1A-mediated regulation of gst-4_p::gfp, a reporter of the oxidative stress response. We found that pbs-5(∆ARE) mutant animals do not show increased sensitivity to arsenite (AS) or paraquat (PQ) compared to the wild type, similarly to animals lacking SKN-1A (Fig. 4A,B) [20]. Since SKN-1A promotes oxidative stress resistance in a manner that is redundant with SKN-1C, we measured whether the pbs-5(∆ARE) mutation enhances oxidative stress resistance in animals lacking SKN-1C. Following exposure to AS, the pbs-5(∆ARE); skn-1c double mutants show reduced survival compared to skn-1c single mutants (Fig. 4A), indicating that in the absence of SKN-1C function, uncoupling pbs-5 from transcriptional regulation by SKN-1A further impairs oxidative stress resistance. But importantly, the pbs-5(∆ARE); skn-1c double mutants still show a higher rate of survival than animals lacking both SKN-1A and SKN-1C, suggesting that the consequences of reduced pbs-5 expression are not as severe as complete inactivation of SKN-1A. Strikingly, whereas animals lacking both SKN-1A and SKN-1C are highly sensitive to PQ, the pbs-5(∆ARE); skn-1c double mutants do not show increased sensitivity compared to the wild type (Fig. 4B). Therefore, SKN-1A can promote PQ resistance by a mechanism that does not require regulation of pbs-5 gene expression. These data suggest that SKN-1A promotes oxidative stress resistance predominantly via redundant upregulation of antioxidant defense genes with SKN-1C.

Fig. 4.

Oxidative stress resistance is largely intact in pbs-5(∆ARE) mutants. A Survival of animals exposed to 3 mM arsenite showing that pbs-5(∆ARE) does not phenocopy loss of SKN-1A in the skn-1c mutant background. Late L4 stage animals from each genotype were placed on plates supplemented with 3 mM arsenite and assayed for survival after 24 h. Results of n = 3–6 replicate experiments are shown. Survival of at least 20 animals was tested for each replicate experiment. Error bars show mean ± SD. Ordinary one-way ANOVA with Šidák’s multiple comparisons test. **** p < 0.0001, *** p < 0.001, ns p > 0.05. B Development of animals exposed to 4 mM paraquat during development showing that pbs-5(∆ARE) does not phenocopy loss of SKN-1A in the skn-1c mutant background. Development was scored 7 days after synchronized egg lay. Results of n = 4 replicate experiments are shown, at least 100 animals’ development was measured in each replicate experiment. Ordinary one-way ANOVA with Tukey’s multiple comparisons test. **** p < 0.0001, ns p > 0.05, indicates p-values comparing each mutant to the wild type control

Uncoupling PBS-5/β5 from regulation by SKN-1A/Nrf1 enhances age-dependent decline in proteostasis

SKN-1A/Nrf1 boosts proteasome levels to allow cellular adaptation to protein misfolding and aggregation [16, 77]. In animals lacking SKN-1A, the age-dependent toxicity caused by the aggregation-prone amyloid beta (Aβ) peptide is enhanced, whereas SKN-1A over-expression is protective [15, 16]. To test whether coordinated regulation of proteasome subunit genes by SKN-1A/Nrf1 is required to avert age-related decline in protein homeostasis, we examined the toxicity of Aβ in pbs-5(∆ARE) animals. The age-dependent paralysis caused by Aβ expression in the muscle is significantly enhanced in pbs-5(∆ARE) animals, although this enhancement is not as severe as that observed in skn-1a mutants (Fig. 5A). We conclude that upregulation of pbs-5 by SKN-1A/Nrf1 is required to protect against the age-dependent toxicity of aggregation-prone Aβ. The more severe effect observed in skn-1a mutants suggests that SKN-1A, which is hyperactivated in the pbs-5(∆ARE) animals, is still competent to partially protect against Aβ toxicity despite the failure to regulate pbs-5 gene expression.

Fig. 5.

Age-dependent decline in proteostasis is accelerated in pbs-5(∆ARE) mutants. A Age-dependent paralysis caused by expression of amyloid beta (Aβ) peptide in the muscle is enhanced in skn-1a and pbs-5(∆ARE) mutants. The mean paralysis ± SD is shown for n = 3 replicate experiments. Paralysis of > 50 animals was scored for each replicate assay. Ordinary two-way ANOVA with Dunnett’s multiple comparisons test. **** p < 0.0001, *** p < 0.001, ns p > 0.05. P-values are indicated for the comparison between each mutant and the wild type control at each timepoint. B Fluorescence micrographs showing increased accumulation and aggregation of muscle-expressed KIN-19::tagRFP in png-1 and pbs-5(∆ARE) mutant day 1 adult animals. Scale bar shows 30 μm. C Quantification of KIN-19::tagRFP aggregates (fluorescent puncta) from (B). n > 15 animals were measured from each genotype. Error bars show mean ± SD. Ordinary one-way ANOVA with Tukey’s multiple comparisons test. **** p < 0.0001, ns p > 0.05. D Age-dependent loss of locomotor function caused by expression of KIN-19::tagRFP in the muscle is enhanced in png-1 and pbs-5(∆ARE) mutant animals. n = 10 day 1 adult animals were assayed from each genotype. Error bars show mean ± SD. Ordinary one-way ANOVA with Tukey’s multiple comparisons test. **** p < 0.0001, ** p < 0.01

We also analyzed the effects of SKN-1A and the pbs-5(∆ARE) mutation on a reporter of age-dependent protein aggregation, KIN-19::tagRFP. KIN-19 is an intrinsically aggregation prone protein, that accumulates in insoluble protein aggregates in aged animals [78]. When overexpressed in muscle, KIN-19::tagRFP forms aggregates that can be visualized as fluorescent puncta and that cause impairment of muscle function [78, 79]. We observed increased accumulation of KIN-19::tagRFP in the muscles of day 1 adult animals lacking PNG-1, an essential component of the SKN-1A pathway, and an equivalent increase in pbs-5(∆ARE) mutants (Fig. 5B,C). In wild type animals, muscle-specific overexpression of KIN-19::tagRFP causes age-dependent defects in muscle function [78, 79]. This effect is strongly enhanced in png-1 or pbs-5(∆ARE) mutant animals, indicating that SKN-1A/Nrf1-mediated regulation of the proteasome is required to limit toxicity of KIN-19::tagRFP aggregates. We conclude that PBS-5 misregulation phenocopies SKN-1A inactivation effects on age-dependent protein aggregation.

Uncoupling PBS-5/β5 from regulation by SKN-1A/Nrf1 abrogates SKN-1-mediated lifespan extension

Our data suggest that the pbs-5(ΔARE) mutation specifically interferes with concerted regulation of proteasome subunits by SKN-1A and is detrimental to proteostasis. The consequences of this loss of proteasome homeostasis are distinct from the effects of SKN-1A inactivation because proteasome homeostasis is perturbed without inactivation of other SKN-1A functions, such as in the oxidative stress response. These observations suggested that pbs-5(∆ARE) could be used to test the role of proteasome homeostasis in longevity. We were particularly motivated to test how impaired proteasome homeostasis affects genetic interventions that lead to SKN-1-dependent lifespan extension: reduced insulin signaling, germline ablation, and reduced mTOR signaling [36, 41, 42, 80]. Although lifespan extension is known to be skn-1-dependent in these contexts, skn-1 is implicated in the control of many cellular and physiological processes relevant to healthy aging, and the significance of these various processes downstream of SKN-1A/Nrf1 for aging are difficult to dissect [18]. Further, the contributions of the two major SKN-1 isoforms have not been systematically examined in these paradigms. We therefore included isoform-specific mutations that selectively inactivate either SKN-1A (analogous to inactivation of Nrf1 in mammals, master regulator of proteasome levels) or SKN-1C (analogous to inactivation of mammalian Nrf2, master regulator of redox homeostasis and detoxification), to further clarify the relevance of each Nrf in these modes of lifespan extension.

Hypomorphic alleles of the daf-2 insulin/IGF receptor gene cause increased lifespan [81, 82]. skn-1 contributes to this lifespan extension in a manner that is dependent on the cultivation temperature and the specific daf-2 mutation used [36, 80]. We chose to test the effect of the daf-2(e1370) allele in animals reared at 15 °C, which causes strictly skn-1-dependent lifespan extension [80]. We found that isoform-specific skn-1 alleles have strikingly distinct effects. In animals lacking SKN-1C, lifespan extension caused by daf-2(e1370) is completely unaffected, indicating that SKN-1C is dispensable (Fig. 6A,B). In contrast, the skn-1a mutants show reduced lifespan compared to the wild type and are completely refractory to lifespan extension caused by the daf-2 mutation (Fig. 6A,B). These data suggest that SKN-1A is the skn-1 isoform that mediates lifespan extension caused by reduced insulin signaling, consistent with the severe reduction in lifespan extension following DAF-2 depletion by RNAi in skn-1a mutant animals [34]. This effect is phenocopied in pbs-5(ΔARE) mutants, which show a reduced lifespan compared to the wild type and in which lifespan extension by the daf-2(e1370) mutation is completely blocked. Thus, SKN-1A-dependent proteasome homeostasis is required for lifespan extension by reduced insulin/IGF signaling.

Fig. 6.

The pbs-5(∆ARE) mutation abrogates lifespan extension. A, B Lifespan experiments showing that skn-1a and the pbs-5(∆ARE) mutations abrogate lifespan extension conferred by daf-2(e1370). Lifespan experiments were carried out with animals maintained at 15 ºC throughout life. In panel (A), lifespan of n > 90 animals was measured for each genotype. In panel (B), the percent increase in mean survival caused by the daf-2(e1370) mutation in each genetic background is plotted from three replicate experiments in which the survival of n > 90 animals was measured. C, D Lifespan experiments showing that the pbs-5(∆ARE) mutation abrogates lifespan extension conferred by rsks-1 inactivation. Lifespan experiments were carried out with animals maintained at 20 °C throughout life. In panel (C), lifespan of n > 80 animals was measured for each genotype. In panel (D), the percent increase in mean survival caused by the rsks-1(ok1255) mutation in each genetic background is plotted from three replicate experiments in which the survival of n > 80 animals was measured. E, F Lifespan experiments showing that the pbs-5(∆ARE) mutation abrogates lifespan extension conferred by germline ablation in glp-1ts animals. Lifespan experiments were carried out with animals maintained at 25 °C throughout life, which causes absence of germ cells in glp-1(bn18ts) mutant animals. In panel (C), lifespan of n > 70 animals was measured for each genotype. In panel (D), the percent increase in mean survival caused by the glp-1(bn18ts) mutation in each genetic background is plotted from three replicate experiments in which the survival of n > 70 animals was measured. Raw data and statistical analysis for all lifespan experiments are provided in supplementary tables S1-S3

Reducing mTOR signaling causes increased lifespan, in a partially skn-1-dependent manner [42, 83, 84]. The rsks-1 gene encodes ribosomal S6 kinase, a conserved downstream effector of mTOR signaling. To test control of longevity by mTOR, we examined the effect of skn-1a, skn-1c and pbs-5(ΔARE) on lifespan extension caused by inactivation of RSKS-1/S6K. These experiments were performed at 20 °C, since previous work had demonstrated lifespan extension by inactivation of rsks-1 and other mTOR components in animals cultivated at this temperature [42, 85]. When raised at 20 °C, the skn-1a, skn-1c and pbs-5(ΔARE) mutants all show a similar lifespan to the wild type (Fig. 6C,D). The lifespan extension caused by the rsks-1 mutation is blunted in both skn-1a and skn-1c mutants, suggesting that both isoforms contribute to lifespan extension in this context. Importantly, the lifespan extension of rsks-1 mutants was entirely abolished in pbs-5(ΔARE) mutant animals (Fig. 6C,D). These results indicate coordinated control of proteasome subunit gene expression is essential for this mode of lifespan extension.

Ablation of the germline causes increased lifespan in a skn-1-dependent manner [41, 86, 87]. To test this mode of lifespan extension, we used a temperature sensitive (ts) allele of glp-1/Notch. In glp-1(bn18ts) mutants raised at 25 °C, proliferation of germline stem cells is blocked resulting in significant lifespan extension [87–90]. As a result, these lifespan experiments were carried out using animals raised at 25 °C. Under this culture condition, the skn-1a, skn-1c and pbs-5(ΔARE) mutants all show a similar lifespan to the wild type (Fig. 6E,F). The lifespan extension caused by glp-1(ts) is significantly reduced in animals lacking SKN-1A or SKN-1C but is completely abrogated in the pbs-5(∆ARE) mutants (Fig. 6E,F). These data suggest that, although each SKN-1 isoform is required for full lifespan extension when the germline is ablated, lifespan extension is completely blocked without coordinated regulation of proteasome subunit gene expression. Collectively, these data indicate that concerted control of proteasome subunit gene expression is essential for multiple longevity paradigms.

Discussion

Here, we characterize the molecular, cellular and physiological consequences of a regulatory allele that disrupts SKN-1A-dependent transcriptional control of the PBS-5/β5 proteasome subunit. Our findings provide insight into the significance of coordinated regulation of proteasome subunits and SKN-1/Nrf signaling in proteostasis, stress responses, and longevity.

The promoter of the CEOP1752 operon, which controls expression of the PBS-5/β5 proteasome subunit, contains two SKN-1 binding motifs (AREs) that are required for regulation by SKN-1 [54]. In an unbiased mutagenesis screen, we identified a small deletion in the promoter of CEOP1752 that removes the two SKN-1/Nrf binding motifs and causes constitutive activation of rpt-3_p::gfp, a reporter of SKN-1A/Nrf1-dependent proteasome subunit gene transcription [10]. In this study, we show that loss of AREs from the CEOP1752 promoter causes reduced expression of PBS-5/β5, which in turn triggers activation of SKN-1A/Nrf1, leading to compensatory activation of other proteasomal subunit genes (e.g., rpt-3). In wild type animals, proteasomal degradation of SKN-1A/Nrf1 creates a negative feedback loop, ensuring that SKN-1A/Nrf1 drives increased expression of proteasome subunits only when necessitated by proteotoxic stress [10, 56, 57]. This regulatory feedback is evidently disrupted in pbs-5(∆ARE) mutants, resulting in constant activation of SKN-1A. Thus, our results provide molecular genetic support for tight regulatory feedback between proteasome subunit gene transcription and SKN-1A/Nrf1 activity.

The PBS-5/β5 subunit is essential for proteasome biogenesis and function [47]. Consistently, pbs-5(∆ARE) mutants are deficient in proteasome function despite constitutive SKN-1A activation, which normally drives enhanced proteasome function and proteostasis [15, 16]. The pbs-5(∆ARE) mutants’ defects in proteasome function and survival following BTZ exposure phenocopy those of animals lacking SKN-1A. Our data therefore reveal that loss of SKN-1A-mediated regulation of PBS-5/β5 is sufficient to disable physiological regulation of protein turnover by the SKN-1A/Nrf1 pathway. Thus, these data show that uncoupling regulation of pbs-5 from SKN-1A/Nrf1 creates a situation in which the SKN-1A/Nrf1 pathway cannot dynamically adapt proteasome levels to meet cellular needs and instead drives unbalanced overproduction of proteasome subunits that are not effectively utilized (Fig S4).

We found that the subcellular localization of proteasome subunits and the proteolytic processing of PBS-2/β2 is dramatically altered in pbs-5(ΔARE) mutants. These findings are consistent with a model in which insufficient PBS-5/β5 expression impairs proteasome assembly. Our results closely mirror recent findings in human cells showing that inactivation of the THAP1 transcription factor leads to reduced PSMB5/β5 gene expression [74, 75]. Thus, adequate expression of the β5 subunit is a conserved requirement for proteasome assembly and function. We find that proteasome subunit localization and PBS-2/β2 processing are restored in pbs-5(∆ARE); skn-1a double mutants. This result indicates that in the case of the pbs-5(∆ARE) mutant, defects in proteasome subunit localization and processing arise from the combined effects of insufficient pbs-5 expression and SKN-1A-dependent overproduction of other proteasome subunits, underscoring the importance of coordinated transcriptional control of proteasome subunit genes for proper proteasome biogenesis. In humans, THAP1 deficiency causes dystonia, a neurological disorder [91]. The similarities between THAP1 deficiency and pbs-5(∆ARE) mutant animals support the use of pbs-5(∆ARE) as a tractable whole-animal model to explore the cellular and organismal consequences of β5 insufficiency, providing insights relevant to dystonia pathogenesis.

At this stage, it is unclear whether the effects we observe in the pbs-5(∆ARE) mutant are unique or could also be caused by mutations in the promoters of other proteasome subunits. Interestingly, depletion of different individual proteasome subunits in human cells leads to distinct effects on cellular protein homeostasis and stress resistance [92–94]. These effects may reflect the distinct effects of different subunit depletions on proteasome assembly or function, distinct functions of proteasomal subcomplexes, or even extra-proteasomal functions of some subunits. Understanding the distinct molecular and physiological consequences of transcriptional misregulation of different proteasome subunit genes will be of great interest.

SKN-1A/Nrf1 promotes protein homeostasis in response to protein misfolding and aggregation [16, 77]. Animals lacking SKN-1A are more susceptible to age-dependent aggregation and toxicity of Aβ, which forms insoluble amyloid aggregates in Alzheimer’s disease [16]. We find that the toxicity of Aβ and another aggregation-prone protein KIN-19 is enhanced in pbs-5(∆ARE) mutants. Thus, these data indicate that homeostatic maintenance of an adequate pool of functional proteasomes is the primary mechanism by which SKN-1A provides protection against accumulation of protein aggregates during aging. In contrast, the pbs-5(∆ARE) mutants do not phenocopy skn-1a mutants in oxidative stress resistance. This is consistent with conserved redundant regulation of genes involved in redox homeostasis and oxidant detoxification by SKN-1A/Nrf1 and SKN-1C/Nrf2 [20, 25, 27, 28, 95]. Thus, the pbs-5(ARE) mutation enables dissection of proteasome-dependent and proteasome-independent functions of SKN-1A.

Our examination of lifespan extension in pbs-5(∆ARE) mutant animals suggests that proteasome homeostasis is a key requirement for healthy aging and longevity. In all three lifespan extension paradigms that we tested, lifespan extension is completely abrogated by the pbs-5(∆ARE) mutation. The defect caused by pbs-5(∆ARE) defect is significantly more severe than the effect of SKN-1A inactivation in some contexts. For example, lifespan extension caused by inactivation of RSKS-1/S6K or by germline ablation is blunted in skn-1a mutants but eliminated in pbs-5(∆ARE) mutants. This suggests that misregulation of SKN-1A/Nrf1 signaling in the pbs-5(∆ARE) mutant leads to a more severe outcome than simple inactivation of SKN-1A/Nrf1. Although the basis for this difference remains unknown, we favor a model in which improper coordination of proteasome subunit gene expression causes a more severe defect in aging animals than the loss of SKN-1A/Nrf1-mediated proteasome regulation itself. Alternatively, constitutive SKN-1A/Nrf1 hyperactivation occurring in pbs-5(∆ARE) mutant animals could be a driver of reduced lifespan. Indeed, inappropriate hyperactivation of SKN-1/Nrf is detrimental in many contexts [36, 37, 39, 40, 96]. These observations altogether suggest that whereas concerted activation of proteasome gene expression by SKN-1A/Nrf1 promotes longevity [16, 34], non-concerted activation of proteasome subunit genes and/or misregulation of SKN-1A/Nrf1 activity limits longevity. These findings support that precise homeostatic regulation of SKN-1/Nrf signaling is a critical requirement for organismal function in aging.

Our lifespan experiments were carried out using animals exposed to flurodeoxyuridine (FUdR) to inhibit progeny production and allow automated lifespan measurement with the WormBot platform [97]. Exposure to FUdR can cause changes in protein misfolding and responses to proteotoxic stress, which may have modified the lifespan phenotypes we observed [98–100]. In the future, it will be important to analyze the interplay between SKN-1A/Nrf1, proteasome biogenesis, and longevity in non-FUdR-treated animals to address the generality of our findings. We also note that pbs-5(∆ARE) lesion removes a regulatory element in the CEOP1752 promoter that is expected to cause reduced expression of both K05C4.2/Jagunal and pbs-5/β5. We support that pbs-5(∆ARE) abrogates longevity due to defects in proteasome regulation and function that ultimately stem from misregulation of pbs-5/β5. However, we cannot exclude the possibility that reduced expression of K05C4.2/Jagunal, which could cause changes in ER function or protein trafficking [50–52], contribute to changes in animal longevity. Future work to evaluate the potential contribution of K05C4.2/Jagunal to longevity will be needed to address this issue.

Given that maintaining balance in proteasome subunit levels is important, it is perhaps surprising that overexpression of PBS-5/β5 causes enhanced proteostasis and lifespan extension in C. elegans and Drosophila [101, 102]. Further, overexpression of PBS-5/β5 is protective in neurodegenerative disease models in C. elegans, Drosophila, and mice [103, 104]. Importantly, in these studies, overexpression of the β5 subunit of the proteasome was found to be accompanied by increased expression of other proteasome subunits. These effects are not limited to the β5 subunit of the proteasome, as overexpression of the rpn-6/Rpn6 subunit of the 19S proteasome also leads to increased proteasome activity and extended lifespan [86]. Thus, the adaptive capacity of cells to maintain balanced proteasome subunit expression when a single subunit is overexpressed appears to be a conserved and general phenomenon. Activation of SKN-1A/Nrf1 signaling, or other pathways that enhance proteasome homeostasis, is therefore a potential strategy for combating age-related proteostasis decline and neurodegenerative diseases.

Materials and Methods

C. elegans strains and maintenance

C. elegans were maintained on standard nematode growth media (NGM) at 20 °C and fed E. coli OP50. A list of strains used in this study is provided in Table S4.

Plasmid constructs and transgenesis

Cloning was carried out using the NEBuilder HiFi DNA Assembly kit (New England Biolabs #E2621). The pbs-5 rescue plasmid (pNL25) was assembled using a fragment containing the CEOP1752 promoter (848 bp immediately upstream of the K05C4.2 start codon), fused to a second fragment containing the pbs-5 coding sequence, predicted 3’UTR, and downstream intergenic sequence (1550 bp starting with the pbs-5 start codon), to generate a construct in which the expression of pbs-5 mRNA is driven by the CEOP175 promoter (Fig S2), and assembled into the pDONR221 backbone. Extrachromosomal arrays transgenes containing pNL25 (10 ng/μl), a myo-2::mcherry co-injection marker (2.5 ng/μl), and Invitrogen 1 kb plus ladder as carrier DNA (87.5 ng/μl).

RNA extraction and quantitative PCR

Drug treatments were carried out as described in [15]. For each replicate experiment, approximately 3000 synchronized L4 stage animals were incubated in the presence of BTZ for 7–9 h. drug treatments were carried out in liquid culture in 6-well tissue culture plates in a total volume of 700 μl. In addition to the desired concentration of BTZ, each plate contained E. coli OP50 (equivalent to E. coli from 4 mL saturated LB culture), 50 mg/mL Carbenicillin, and 0.01% DMSO. After the drug treatment worm pellets were washed with M9 to remove bacteria before resuspension in 1 mL TRIzol reagent (Invitrogen, #15,596). RNA extraction was performed according to the manufacturer’s instructions and all RNA samples were DNase-treated using the TURBO DNA-free kit (Invitrogen, #AM1907). Reverse transcription was carried out using the High-Capacity cDNA Reverse Transcription kit (ThermoFisher, #4,368,814). Real time PCR was performed using an Applied Biosystems 7500 Real Time PCR System instrument and Power SYBR Green PCR master mix (Thermo Fisher, #4,367,659). Expression levels were calculated using the ∆∆Ct method normalized to rps-23. The following oligonucleotide primers were used:

• pbs-5 qPCR F: CCACTTATCGGGATTCTGGATC

• pbs-5 qPCR R: GTTGTAGGTGATATCTCTGCCG

• rps-23 qPCR F: AAGGCTCACATTGGAACTCG

• rps-23 qPCR R: AGGCTGCTTAGCTTCGACAC

Genome Modification by CRISPR/Cas9

CRISPR/Cas9 genome editing was carried out through microinjection of guide RNA/Cas9 ribonucleoprotein (RNP) complexes (IDT #1,081,058, #1,072,532, and custom crRNA) along with single-stranded DNA oligos or PCR products as homology-directed repair (HDR) templates [105, 106].

rpt-3(nic913[rpt-3p::mcherry::H2B::SL2::rpt-3]) was generated using a guide RNA targeting the following sequence: ACCGATTTCCTCCAAAGCAG. The HDR template was a PCR product derived from the following primers:

• Forward: ttttttgcaacaacgttaaattttcatttctaggaacATGGTCTCAAAGGGTGAAGAAGA

• Reverse: TCCTGTAATGGAGCAATACCGATTTCCTCCAAAGCtgaagcactcatctttggagttcgattctgaaaat

The HDR PCR product was amplified from pNL47 [16] which contains the mcherry::H2B::SL2 cassette. This edit adds a mcherry::H2B::SL2 cassette immediately after the promoter of the rpt-3 gene (Fig S3).

pbs-4(nic1600[pbs-4::GFP]) was generated using a guide RNA targeting the following sequence: TAAGAAGGGATACCGCAAAT. The HDR template was a PCR product derived from the following primers:

• Forward: GTGATCATCGATAAGAAGGGATACCGCAAAgggagttcaagaccatctacaaggAGTACTAGCGGTGGCAGTGGAGG −3'

• Reverse: gccagaaatTTAGAAAAGAACATCATCCAACTTGTAGAGCTCGTCCATTCCGT

The HDR PCR product was amplified from pNL558, which encodes superfolder GFP (sfGFP) with C. elegans codon optimization and two short introns (introns from the rps-5 gene). The sfGFP insert was designed using the Wormbuilder C. elegans transgene adaptation tool (wormbuilder.org/transgenebuilder/) and synthesized by Twist Biosciences. This edit alters the pbs-4 gene to encode PBS-4 fused to sfGFP after the 196th residue, connected by the following flexible linker sequence: GSSRPSTRSTSGGSGGTGGSS. Six C-terminal amino acids (197–202) of PBS-4 are intact but are C-terminal to the linker::sfGFP insertion.

pbs-2(nic1799[pbs-2::3xFLAG]) was generated using a guide RNA targeting the following sequence: GGATATCACTCACTAAATTG. The HDR template was a single stranded DNA oligo with the following sequence: GACGTCGTCGAGTCCATGGATATCACTCACgactacaaagaccatgacggtgattataaagatcatgatatcgattacaaggatgacgatgacaagTAAattgcggtcttctgtcgatgtgtgtat.

All CRIPR edits were identified by diagnostic PCR and confirmed by sequencing.

Microscopy

Brightfield and fluorescence images were collected on a Leica M165FC equipped with a 910 Leica K5 sCMOS camera and using LAS X software, except images shown in Fig. 2, which were collected on a Zeiss AxioImager Z1 microscope equipped with an Axiocam 705 CMOS camera and using ZEN software. For fluorescence imaging, worms were immobilized using sodium azide and mounted on 2% agarose pads. All images were processed and analyzed using Fiji software. Images shown within the same figure panel were collected using the same exposure time and were then processed identically, except for panels of Fig. 2E,F,I in which images have been brightness/contrast adjusted to allow comparison of localization between animals with very different levels of fluorescence. To quantify rpt-3_p::gfp expression, the mean pixel intensity along a longitudinal section of each animal was measured. For rpt-3_p::mCherry, pbs-4::GFP, GFP::rpn-8, and rpl-28_p::ub[G76V]::GFP images, the mean pixel intensity was measured using a polygon selection around each animal. Nuclear enrichment of proteasome subunits (pbs-4::GFP and GFP::rpn-8) was measured by selecting an area around the nucleus using the polygon selection tool in ImageJ, measuring the mean fluorescent intensity within that region, and dividing it by the fluorescent intensity of an identically sized area in the adjacent cytoplasm. KIN-19::tagRFP aggregates were quantified using the Find Maxima tool in ImageJ. Fluorescence from vulval muscle (in which KIN-19::tagRFP is more highly abundant than the body wall muscle) was excluded from the analysis.

Bortezomib resistance assays

Bortezomib sensitivity was evaluated by assessing the ability of animals to grow during development or survive as adults on plates containing various concentrations of bortezomib (LC Laboratories, #B1408). The bortezomib solution was directly applied to NGM plates seeded with OP50 and left to dry for at least 2 h. Control plates supplemented with DMSO were prepared simultaneously. For development assays, five L4-stage animals were transferred to fresh plates supplemented with 1 μM bortezomib and incubated at 20 °C for three or four days. Progeny development was then assessed. Images were captured using a Leica M165FC microscope equipped with a Leica K5 sCMOS camera and LAS X software. For survival assays, 30 L4-stage animals were transferred to fresh plates supplemented with 50 nM or 1 μM bortezomib and incubated at 20 °C. Survival was scored after 3 or 4 days. For strains producing progeny that can develop in the presence of the tested bortezomib concentration, animals were transferred to fresh bortezomib-supplemented plates after 2–3 days to ensure that the animals being assayed for survival could be distinguished from their progeny.

Oxidative stress resistance assays

Arsenite survival assay

NGM plates containing sodium arsenite (RICCA Chemical; Cat # 7142–16) at final concentrations of 1 mM or 3 mM were prepared one day before the assay. Plates were left to dry for 2 h before being seeded with 250 μL of 10X concentrated OP50. The following day, approximately 30 L4-stage C. elegans were transferred to each plate and incubated at 20 °C for 24 h. After incubation, the number of live and dead animals was recorded.

Paraquat development assay

NGM plates containing paraquat were prepared one day before the experiment. A 1 M stock solution of methyl viologen hydrate (Thermo Scientific; Cat # 227,320,050) was freshly prepared in ddH₂O and added to the NGM media at a final concentration of 4 mM. Plates were dried for 2 h before being seeded with 250 μL of 10X concentrated OP50. The following day, 10–25 gravid C. elegans were transferred to each plate and incubated at 20 °C for 3–4 h. After incubation, the adults were removed, and the remaining progeny were allowed to develop for 7 days. The number of animals that matured into gravid adults was recorded, along with the total number of animals on the plate. For mutants balanced with tmC25, only homozygous mutant progeny (identified by the absence of the myo-2_p::GFP marker) were scored.

Preparation of 10X OP50 for oxidative stress assays

A single OP50 colony was inoculated in 5 mL of Luria–Bertani (LB) broth and grown overnight at 37 °C. The next day, 1 mL of the overnight culture was added to 1 L of Terrific Broth (TB) and incubated at 37 °C for 2 days. The culture was then centrifuged at 4,000 rpm for 15 min. After discarding the supernatant, the pellet was resuspended in 45 mL of M9 buffer and transferred to a 50 mL conical tube, creating a 20X stock solution, which was stored at 4 °C. To prepare the 10X working solution, the 20X stock was diluted 1:1 with M9.

Bortezomib treatment for imaging

NGM plates containing the desired concentration of bortezomib were prepared as described for bortezomib resistance assays. Control plates supplemented with DMSO were prepared in parallel. To image Ub[G76V]::GFP accumulation following proteasome inhibition, 25–30 L4-stage C. elegans were transferred to each plate and incubated at 20 °C for approximately 20 h before imaging. To image proteasome subunit levels and localization following chronic proteasome inhibition, 5–10 L4-stage C. elegans were transferred to each plate and incubated at 20 °C for 3–4 days. L4 stage animals in the progeny generation were then selected for imaging. Animals were imaged using the procedures described in the Microscopy section. Bortezomib-treated worms were compared to DMSO-treated control animals in all experiments.

Aβ-induced paralysis assay

For each assay, approximately 10 adult animals raised at 15 °C were transferred to 25 °C and allowed to lay eggs for a few hours to produce L1-stage-synchronized progeny. The L1 animals were then raised at 25 °C. Under these conditions, the animals reached adulthood after approximately 48 h. Starting at 48 h, animals were scored for paralysis every 24 h. An animal was considered paralyzed if it showed no movement after tapping the plate or gently prodding the body.

Locomotion assay (KIN-19::tagRFP)

First-day adult C. elegans, grown at 20 °C, were transferred to a fresh NGM plate seeded with OP50 and allowed to rest for approximately 30 s. Each worm was then gently prodded at the tail to initiate movement. Once the worm began to move, body bends were counted for 30 s. A body bend was defined as the movement of the worm’s body from the maximum to the minimum amplitude of its sinusoidal wave pattern. To minimize variability, all worms within a single experiment were assayed on the same plate.

Lifespan assays

Lifespan assays were performed by the Invertebrate Longevity and Healthspan Core at the University of Washington Nathan Shock Center of Excellence in the Basic Biology of Aging. Lifespan measurements of hermaphrodite C. elegans were made using the WormBot platform [97]. Briefly, the system works by quantifying the time of death (cessation of movement) of each animal in each well of a 12-well plate. Time of death was determined by manual analysis of videos captured throughout the lifespan experiment. The media used for the lifespan experiments was supplemented with 50 μM FUDR to stop progeny production. Raw data from all lifespan assays are shown in Tables S1-S3.

Western blot analysis

Animals were bleached-synchronized and then incubated at 20 °C for 3 days. Young adult animals were collected in 15 mL tubes, washed twice in M9 to remove bacteria, and then transferred to 1.5 mL tubes. 30 μl of the worm pellet was snap-frozen in liquid nitrogen and stored at − 80 °C. The worm pellet was briefly thawed at room temperature, mixed with 60 μl of M9, 60 μl of LDS Sample buffer (Invitrogen, #B0007) and 6 μl of Bond-Breaker™ TCEP solution, then heated to 95 °C for 10 min, followed by gentle vortexing for 10 min. Samples were centrifuged at 14,000 × g for 10 min to pellet debris. SDS-PAGE and western blotting was performed using Bolt™ 4–12%, Bis–Tris Plus gels (Invitrogen, #NW04122BOX) and iBlot™ 2 nitrocellulose mini stacks (Invitrogen, #IB23002), according to the manufacturer’s instructions. The following antibodies were used: mouse monoclonal anti-FLAG M2 antibody (SIGMA, #F1804; 1:2,000 dilution), mouse monoclonal anti-α-tubulin (SIGMA, #T6074; 1:10,000 dilution), and goat anti-mouse IgG antibody, HRP conjugate (Millipore, #12–349; 1:1000 dilution). The SuperSignal West Pico PLUS chemiluminescent substrate (Thermo Scientific, #34,577) was used for detection of the HRP signal.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

Conceptualization, NL; Funding Acquisition, NL; Investigation NL & IT; Methodology; NL & IT; Visualization, IT; Writing – Original Draft Preparation, NL & IT; Writing – Review and Editing, NL & IT.

Funding

This work was supported by the National Institutes of Health (NIH) National Institute of General Medical Sciences (grant R35GM142728 to NL) and funding from a Nathan Shock Center Pilot Award at the University of Washington Nathan Shock Center of Excellence in the Basic Biology of Aging (P30AG013280).

National Institute of General Medical Sciences,R35GM142728,Nicolas Lehrbach,National Institute on Aging,P30AG013280

Data Availability

All data are available in the main text or the supplementary materials.

Declarations

Competing interests

The authors declare no competing interests.

Ethics and consent to participate

Not applicable.