AKIR-1 regulates proteasome subcellular function in Caenorhabditis elegans

Summary Polyubiquitinated proteins are primarily degraded by the ubiquitin-proteasome system (UPS). Proteasomes are present both in the cytoplasm and nucleus. Here, we investigated mechanisms coordinating proteasome subcellular localization and activity in a multicellular organism. We identified the nuclear protein-encoding gene akir-1 as a proteasome regulator in a genome-wide Caenorhabditis elegans RNAi screen. We demonstrate that depletion of akir-1 causes nuclear accumulation of endogenous polyubiquitinated proteins in intestinal cells, concomitant with slower in vivo proteasomal degradation in this subcellular compartment. Remarkably, akir-1 is essential for nuclear localization of proteasomes both in oocytes and intestinal cells but affects differentially the subcellular distribution of polyubiquitinated proteins. We further reveal that importin ima-3 genetically interacts with akir-1 and influences nuclear localization of a polyubiquitin-binding reporter. Our study shows that the conserved AKIR-1 is an important regulator of the subcellular function of proteasomes in a multicellular organism, suggesting a role for AKIR-1 in proteostasis maintenance.


INTRODUCTION
Protein degradation is one of the essential mechanisms that the cell uses to maintain its homeostasis and protein balance.The major machinery for degrading individual proteins is the ubiquitin-proteasome system (UPS), by which a cascade of enzymes for ubiquitin activation (E1), conjugation (E2), and ligation (E3) marks selected proteins with polyubiquitin chains as a signal for degradation. 1Hydrolysis of the tagged proteins into small peptide fragments is performed by the proteasome, a large 2.5-megadalton (MDa) protein complex composed of two subcomplexes: a catalytic core particle (CP or 20S) and a regulatory particle (RP or 19S).The catalytic activities, i.e., trypsin-, chymotrypsin-and caspase-like activity of the barrel-shaped 20S particle, are contained within the two seven-subunit b-rings, which are stacked between two seven-subunit a-rings.The CP can be capped at either one (26S) or both ends by the multisubunit regulatory particle, which functions in recognition, binding, deubiquitination, unfolding, and translocation of the substrates into the lumen of the CP. 2 The proteasomal degradation capabilities are broadened with the ability of the free CP to degrade some proteins independently of ubiquitination 3 and the existence of proteasome variants with alternative subunits of the CP or the regulatory particle, or activating complexes. 4uring the last decades, studies based on biochemical and immunological methods and live-cell imaging have reported proteasomes to be localized both in the cytoplasm and in the nucleus in various cell types. 5,6Despite vast accumulating evidence on variation in nuclear and cytoplasmic distribution of proteasomes in different cell lines and physiological conditions, the regulatory mechanisms governing nuclear localization and function of the proteasomes are only beginning to emerge. 6][9] However, there are also reports both in yeast and human cells showing that nuclei may lack proteolytic activity, suggesting that proteasomes localize mainly on the cytoplasmic side of the nuclear membrane. 10,11In addition, in some circumstances ubiquitinated proteins seem to be exported from the nucleus to the cytoplasm for degradation. 12][15][16] For nuclear transport, receptors of the karyopherin beta family are generally required. 17Some of these receptors have been shown to be involved in the import of proteasomes: the importin a family member SRP1 in yeast, 18

Knockdown of akir-1 alters proteasome activity
Accumulation of polyubiquitinated proteins is commonly caused by altered proteasome levels or activity.We first analyzed the total amount of endogenous proteasomes in whole animal lysates by western blot analysis using an antibody against 20S alpha subunits.No clear change in 20S levels was detected in lysates of either akir-1 mutants (Figure 2A) or akir-1 RNAi-treated animals (Figure 2B) compared to control animals.Next, we performed an in-gel proteasome activity assay on whole animal lysates and detected a slight increase in total proteasome activity upon akir-1 RNAi (Figure 2C, left graph).Interestingly, the 20S CP appears to be the main contributor of this increased activity (Figure 2C).To test whether the increased CP activity after akir-1 RNAi is caused by changes in the amount of CP complexes, we analyzed the CP levels by immunoblotting native gels with the antibody against 20S alpha subunits.No clear change in the levels of the 20S CPs was detected in akir-1 RNAi-treated animals compared to control RNAi-treated animals (Figure 2D).To specifically measure proteasome activity in intestinal cells of C. elegans, we employed our previously established UPS reporter animals expressing the photoconvertible proteasomal substrate UbG76V-Dendra2 in the intestine. 37,38In these animals, a decrease in the amount of photoconverted UbG76V-Dendra2 reporter reflects in vivo proteasome activity.We first measured the fluorescence of the photoconverted UbG76V-Dendra2 in the whole intestine and detected a similar rate of degradation in control and akir-1 RNAi-treated animals (Figure 2E, extended in Figure S2).As akir-1 depletion leads to nuclear accumulation of polyubiquitinated proteins, we next monitored photoconverted UbG76V-Dendra2 fluorescence intensity at single-cell level, specifically in the nucleus and cytoplasm, in living animals.The fluorescence of UbG76V-Dendra2 at 18 h after photoconversion was more intense in the intestinal nuclei than in the cytosol upon akir-1 RNAi (Figure 2F), demonstrating that akir-1 knockdown results in slower proteasomal degradation in intestinal nuclei in vivo.Taken together, our results suggest that AKIR-1 influences subcellular proteasome activity.

Intestinal nuclei display reduced proteasome levels upon akir-1 depletion
As the accumulation of polyubiquitinated proteins and the altered in vivo proteasome activity in the intestine of akir-1 mutant and RNAitreated animals occur specifically in the nucleus, we investigated subcellular distribution of the proteasome by immunostaining dissected intestines with the anti-20S antibody.We have previously shown using this antibody that the proteasome is concentrated in nuclei in the intestine of wild-type C. elegans. 33Consistently, dissected intestines of wild-type animals displayed high immunofluorescence in the nuclei (Figures 3A and S3).Compared to the controls, both akir-1 mutants and akir-1 RNAi-treated animals showed a consistent reduction in 20S immunofluorescence intensity in intestinal nuclei, when we measured immunofluorescence profiles along a line intersecting the cytoplasm and the nucleus (Figures 3A and S3).As a complementary approach, we investigated the proteasome subcellular localization using transgenic C. elegans expressing extrachromosomal arrays of GFP-tagged RPT-5, a subunit of the 19S RP particle of the proteasome, under its endogenous rpt-5 promoter.The transgenic animals showed a mosaic fluorescence pattern of the ubiquitously expressed GFP::RPT-5.GFP::RPT-5 animals displayed a stronger nuclear fluorescence compared to the cytoplasmic fluorescence in intestinal cells, and upon akir-1 RNAi a clear reduction in nuclear fluorescence was also observed (Figure 3B), indicating nuclear decrease of 26S proteasome in the intestine.In addition, we investigated the localization of the proteasome-associated deubiquitinase (DUB) UBH-4.UBH-4, and its human homolog UCHL5, have previously been shown to interact with the 19S subunit RPN-13 34,[39][40][41] and to be broadly expressed in C. elegans tissues, including the intestine. 34,42Here, we used a CRISPR-engineered UBH-4:GFP strain 42,43 and observed that these animals displayed strong nuclear fluorescence in intestinal nuclei similar to the pattern detected with the anti-20S antibody (Figures 3C and 3A).In accordance with the 20S immunofluorescence and GFP::RPT-5 results, UBH-4:GFP fluorescence decreased in the intestinal nuclei of animals exposed to akir-1 RNAi when compared to control animals (Figure 3).Together, these results demonstrate that proteasome levels decrease in intestinal nuclei upon the loss of AKIR-1.

Depletion of akir-1 affects nuclear polyubiquitinated proteins differently in oocytes and body-wall muscle cells compared to intestinal cells
While performing 20S immunostaining with dissected akir-1 mutants, we noticed that loss of akir-1 altered proteasome subcellular distribution not only in the intestinal cells but also in the oocytes.Oocytes of wild-type animals displayed cytoplasmic and strong nuclear 20S immunofluorescence (Figures 4A and S4).Compared to the wild-type animals, nuclear proteasome localization is clearly reduced in oocytes of akir-1 mutants (Figure 4A).A similar decrease in intensity of nuclear 20S immunostaining was also detected in oocytes of animals exposed to akir-1 RNAi (Figure S4).Further, analysis of UBH-4::GFP animals treated with akir-1 RNAi confirmed reduced levels of nuclear proteasomes in oocytes (Figure 4B).Taken together, these results show that the depletion or downregulation of akir-1 causes a decrease in nuclear proteasomes in oocytes, similarly to our observation in intestinal cells.However, despite the clear reduction in nuclear proteasomes, we did not   (A and B) Western blot analysis with antibody against proteasomal 20S alpha subunits using lysates of akir-1(gk528) mutants and control (N2) animals (A), and of control and akir-1 RNAi-treated wild-type animals (B).Anti-alpha-tubulin antibody was used as a normalization control.The quantification graphs show the average fold change in akir-1(gk528) mutants compared to control (N2) animals (set to 1; n = 6 independent experiments) (A), and in akir-1 RNAi-treated wild-type animals compared to control RNAi-treated animals (set to 1; n = 8 independent experiments) (B).
observe accumulation of polyubiquitinated proteins in the oocyte nuclei of akir-1 mutants (Figure 5A).Most oocytes showed a relatively uniform cytoplasmic polyubiquitin immunofluorescence pattern, with weaker signal in the nuclei of both control animals and akir-1 mutants (Figure 5A).Occasionally, some polyubiquitin-positive staining was detected at the rim of the nuclear membrane in akir-1-depleted animals (Figure 5A, lower panel).The polyubiquitin staining pattern in oocytes of akir-1 RNAi-treated animals resembled the results of akir-1 mutants (Figures 5A and S5A).Importantly, our results revealed that polyubiquitinated proteins do not accumulate to a similar degree in oocyte nuclei, as in intestinal nuclei, upon loss of akir-1.
AKIR-1 is required for C. elegans muscle development, integrity, and function. 31Therefore, we tested the impact of the loss of akir-1 on polyubiquitinated proteins in muscles using our previously established transgenic strain expressing the polyubiquitin-binding reporter in the body-wall muscle cells. 35,36Quantification of the mean fluorescent intensity of the polyubiquitin-binding reporter in live animals showed no difference between control and akir-1 RNAi treatments either in wild-type (N2) background (mean fold change 1.15 G SD 0.19, n = 3 independent experiments) or in the RNAi-sensitive rrf-3 background (mean fold change 0.94 G SD 0.09, n = 2 independent experiments), and no difference in the fluorescence pattern itself was observed (Figure S5B).We also assessed proteasome activity using the photoconvertible UbG76V-Dendra2 reporter expressed in body-wall muscle cells of C. elegans. 37No difference in the degradation of photoconverted UbG76V-Dendra2 was detected in akir-1 RNAi-treated animals compared to control RNAi-treated animals (Figure S5C).
As the RNAi experiments were performed by placing stage 1 larvae (L1) on RNAi plates and assessing the phenotype at day 1 of adulthood, we investigated whether remaining maternal AKIR-1 contribution might influence the lack of polyubiquitin accumulation in body-wall muscle cells.To this end, polyubiquitin-binding reporter animals (in rrf-3 background) were continuously exposed to akir-1 RNAi, and their F1 offspring treated in similar manner were monitored at day 1 of adulthood.We assigned reporter fluorescence signals either as nuclear or non-nuclear, based on colocalization with Hoechst nuclear staining.A slight increase in nuclear localization of the reporter was detected in akir-1 RNAi-treated F1 animals (control 14%, akir-1 RNAi 39% nuclear signal) (Figure 5B).Taken together, our results suggest that bodywall muscle cells respond differently to akir-1 depletion in terms of nuclear accumulation of polyubiquitinated proteins than oocytes or intestinal cells.

Perturbed nuclear transport mimics the akir-1 RNAi-induced polyubiquitin phenotype
Akirin proteins have been implicated in several physiological processes. 280][31] To investigate phenotypes induced by akir-1 downregulation under our experimental conditions, we compared the progeny number and lifespan of rrf-3 animals treated with akir-1 or control RNAi.We observed a 40% reduction in the number of progeny (Figure 6A) and a 2-day decrease in mean lifespan (Figures 6B and 6C) upon akir-1 downregulation, which are in agreement with previous reports on the akir-1 mutant and an epidermis-specific RNAi strain. 29,30To exclude that our detected effects on polyubiquitinated proteins and the proteasome are not linked to developmental defects, or to the completion of intestinal cell divisions in L1 larval stage, 44 we also performed akir-1 RNAi treatment starting from the L4 larval stage.These intestinal polyubiquitin-binding reporter animals, examined at day 3 of adulthood, displayed the same distinctive converged fluorescent pattern (Figure 6D), as detected when RNAi treatment was started at L1 larval stage (Figure 1A).
Previously, Akirins and AKIR-1 have been shown to act in the regulation of transcription via chromatin remodeling complexes. 28Therefore, we examined whether transcriptional regulation is required for the ability of AKIR-1 to affect proteasome function.In C. elegans AKIR-1 interacts specifically with the NuRD (NuRD I, II and MEC) chromatin remodeling complexes, 30,31 which have been reported to display some subunit overlap (Figure S6A). 30,45These chromatin remodeling complex genes were not recognized as hits in our genome-wide RNAi screen, which could have been due to differences in experimental setup and fluorescence detection between the genome-wide scale and select gene RNAi, or due to the developmental criteria followed in our genome-wide RNAi screen.Thus, we next performed downregulation by select gene RNAi for each member of the NuRD complexes and MEC complex on the intestinal polyubiquitin-binding reporter animals but detected In-gel proteasome activity assay with whole animal lysates of wild-type animals exposed to control or akir-1 RNAi treatment (upper gel).Coomassie staining of the same gel (lower gel).The quantifications show the average fold change in chymotrypsin activity in akir-1 RNAi-treated wild-type animals compared to control RNAi-treated animals (set to 1; n = 5 independent experiments).Proteasome activity is indicated as total (Total; CP + RP-CP + RP2-CP), as core particle (CP), and as CP activity subtracted from total activity (Total À CP).RP, Regulatory particle.(D) Immunoblot analysis of whole animal lysates separated under non-denaturing condition using the antibody against proteasomal 20S alpha subunits.Ponceau S staining was used for total protein normalization.The quantification graph shows the average fold change of core particles (CP) in akir-1 RNAi-treated wild-type animals compared to control RNAi-treated animals (set to 1; n = 3 independent experiments).(E) Representative fluorescence micrographs of control and akir-1 RNAi-treated transgenic C. elegans expressing photoconvertible UbG76V-Dendra2 reporter (vha-6p::UbG76V::Dendra2) in intestinal cells.The 0 h (left panels) and 18 h (right panels) indicate time after photoconversion.Scale bar, 500 mm.The graph shows the mean percentages of fluorescence intensity of the photoconverted UbG76V-Dendra2 18 h after the photoconversion relative to the fluorescence at the point of photoconversion (0 h, set as 100%); n = 6 independent experiments with triplicate images of 6-7 animals per image (total number of animals is 108 per treatment).(F) Representative confocal fluorescence micrographs of intestinal cells with photoconverted UbG76V-Dendra2 (18 h after conversion) in transgenic C. elegans treated with control or akir-1 RNAi.Scale bar, 10 mm.The graph shows the ratio between nuclear and cytoplasmic mean fluorescence per cell.n = 2 independent experiments (total number of nuclei is 23 in control RNAi and 27 in akir-1 RNAi treatment).Welch's t-test (two-tailed distribution and unequal variance) was used for statistical analyses.Error bars, SD; ns, not significant; *p < 0,05; **p < 0,01; ***p < 0.001.See also Figure S2.no phenotypes mimicking the akir-1 RNAi-induced effect (Figure S6B).To directly affect transcription, we also performed RNAi against ama-1, a subunit of the RNA polymerase II, and while we observed a growth phenotype, the localization of polyubiquitin-binding reporter was not affected (Figure S6C).As the human homolog of AKIR-1, AKIRIN2, has recently been shown to mediate the nuclear import of proteasomes in conjunction with the cargo receptor, importin 9, 19 and as interactions between AKIR-1 and importins have been reported, 30,32 we tested whether C. elegans importins are similarly involved in regulating nuclear proteasomes.In C. elegans, 13 members of the karyopherin family have been identified (Figure 7A). 46We performed RNAi against all these karyopherin family members using our intestinal polyubiquitin-binding reporter strain (Figures 7B and S7).Of these, downregulation of importin a3, ima-3, showed clearly a similar polyubiquitin phenotype as akir-1 RNAi (Figure 7B), although the intestinal fluorescent puncta formed with weaker intensity.DNA staining with Hoechst confirmed that the polyubiquitin-binding reporter concentrated to the intestinal nuclei upon ima-3 RNAi treatment (Figure 7C).In addition, downregulation of importin b1, imb-1, partially mimicked the akir-1 RNAi-induced fluorescence phenotype, but to a lesser extent than ima-3 RNAi (Figure S7B), and with an apparent sick phenotype.We next tested for genetic interactions between akir-1 and ima-3 by performing ima-3 RNAi on akir-1 mutants (Figure 7D).Compared to either akir-1 mutants or ima-3 RNAi-treated wild-type animals, the akir-1 mutant animals exposed to ima-3 RNAi showed severely delayed growth and reduced body size (Figure 7D), suggesting that AKIR-1 and IMA-3 may cooperate in some cellular processes by acting either in the same complex genetic pathway or in parallel pathways.While it has been suggested that individual importins act as receptors for several different types of protein cargo, 47 we reasoned that AKIR-1 is likely involved in a more limited role.We therefore used a nuclear localized GFP transgenic reporter strain (sur-5p::NLS-GFP) 33,48 to investigate whether AKIR-1 affects nuclear import in general.As expected, RNAi against ima-3 markedly reduced the GFP signal in intestinal nuclei, but akir-1 RNAi had no effect on the nuclear localization of the GFP reporter suggesting a more specific role for akir-1 in nuclear import than for importins (Figure 7E).Taken together, our genetic experiments suggest that AKIR-1 acts in the nuclear import of proteasomes cooperating with importins a3 and b1.

DISCUSSION
Our results reveal a conserved role of the Akirin protein family in regulation of nuclear transport of proteasome and elaborate the knowledge on their complex action in a multicellular organism by presenting a role for C. elegans AKIR-1 as a proteasome regulator.Recently, a screen for regulators of the levels of the transcription factor MYC by de Almeida and colleagues identified an unexpected function for human AKIRIN2 as a regulator of the turnover of a subset of nuclear proteins in human cancer cells. 19As downregulation of AKIRIN2 led to reduced nuclear fluorescence signal of tagged proteasome subunits, and as AKIRIN2 was shown with cryoelectron microscopy (cryo-EM) to bind to the 20S proteasome, they suggested that AKIRIN2 acts as a mediator of proteasome nuclear import.We have identified the AKIRIN2 homolog akir-1, in an unbiased genome-wide RNAi screen for novel proteasome regulators in C. elegans, and show that AKIR-1 functions in nuclear localization of endogenous proteasomes.More precisely, we demonstrate that animals lacking akir-1 have reduced levels of nuclear proteasomes, both in intestinal cells and oocytes.Thus, our results reveal the significance of Akirins in regulating nuclear proteasome localization at an organismal level and that this function is conserved between human cells and invertebrates.The molecular mechanism behind many of the physiological functions of Akirins is believed to be transcriptional regulation due to their interaction with chromatin remodeling complexes and transcription factors, 28,49,50 and as such the effect of AKIR-1 on proteasome localization could be indirect.However, de Almeida and co-workers have shown that the transcript levels of MYC or other AKIRIN2 regulated proteins are not altered upon AKIRIN2 knockout. 19Similarly, we observed no phenotype resembling the akir-1 Thus, our results are in agreement with the presented role of AKIRIN2 and suggest that also AKIR-1 could regulate nuclear transport of proteasomes in C. elegans.However, additional mechanisms regulating the subcellular distribution of proteasomes likely exist in C. elegans, as we show a reduction, but not a complete loss, of nuclear proteasomes upon both downregulation and loss of akir-1 in intestinal cells and oocytes.In addition, the akir-1 null mutants display a developmental defective but viable phenotype, whereas the AKIRIN2 knockout results in apoptosis of human cancer cells. 19In human cells, AKIRIN2 is regulating re-importing of proteasomes into the nucleus upon reformation of the nuclear envelope in late mitosis, 19 but our results reveal that AKIR-1 is also involved in regulating nuclear transport of proteasomes in non-dividing cells, as downregulation of akir-1 results in reduction of nuclear proteasomes and increase in nuclear accumulation of polyubiquitinated proteins in post-mitotic intestinal cells.Thus, our study suggests a more complex role of AKIR-1 in proteasomal degradation in a multicellular organism.1][32] Nuclear transport of proteins is mediated by the importins and exportins of the karyopherin family, 17 and we show that knockdown of two importin family members, ima-3 and imb-1, mimics the akir-1 RNAi-induced nuclear accumulation of the polyubiquitin-binding reporter in intestinal cells.The small fluctuation in observed phenotypes between akir-1 and importins knockdowns might derive from variations in protein stability, receptor redundancy, or the existence of more complex proteasome transport mechanisms.We further show that akir-1 genetically interacts with ima-3, either within the same complex pathway or via a parallel pathway.Previously, it has been reported that GFP-tagged AKIR-1 interacts with IMA-3 30 and that AKIR-1 binds to IMA-2 in a yeast two-hybrid screen. 32Interestingly, the S. cerevisiae importin a family member SRP1, which is believed to be a homolog of C. elegans ima-2 and ima-3, is required for nuclear import of proteasomes. 18,46Importin 9 (IPO9) and 5 (IPO5) have also been implicated to play a similar role in D. melanogaster during spermatogenesis and in human cancer cells. 15,19,20C. elegans has no IPO9 homolog, 46 and we observed no phenotype with the IPO5 homolog, imb-3.Our results suggest that AKIR-1 controls proteasome nuclear import in C. elegans together with importin receptors IMA-3 and IMB-1.Interestingly, a study by Bowman and co-workers showed that the function of AKIR-1 on the synaptonemal complex in meiosis is dependent on IMA-2, but not IMA-3. 32Together, these studies indicate that AKIR-1 might be involved in regulating nuclear events through interactions with different importins.
2][23][24][25] A homolog of Sts1/Cut8 has so far been found only in D. melanogaster, 24 but de Almeida and colleagues hypothesized that the human AKIRIN2 could potentially be a functional homolog of Sts1/Cut8, even though lack of sequence homology. 19Both AKIRIN2 and Sts1 interact with proteasome, contain protein regions of predicted disorder, and are short-lived proteins. 19,21Our akir-1 RNAi experiments, starting at L1 or L4 larval stage, consistently displayed a highly penetrant nuclear phenotype in the intestinal polyubiquitin-binding reporter strain, suggesting that AKIR-1 is a short-lived protein also in C. elegans intestine.Our AKIR-1 results support that Akirins might function as adaptor proteins in the nuclear transport of proteasomes, but future studies are required to uncover whether AKIR-1 functions in the import of proteasomes to the nucleus or possibly in retaining proteasomes in the nucleus.
Our study further demonstrates a broader role of AKIR-1 in regulation of proteasome function and protein homeostasis in a multicellular organism.Lysates of akir-1 RNAi-treated animals contained slightly increased in vitro proteasome activity, which was mainly due to enhanced activity of the 20S particle.Interestingly, when tissue-level degradation in living animals was monitored, no change was detected in proteasomal degradation of the photoconverted UbG76V-Dendra2 reporter protein at the cellular level in either intestinal cells or body-wall muscle cells.This suggests that the increase in in vitro proteasome activity in response to akir-1 knockdown does not stem from these two cell types, but possibly from another cell type(s) present in the whole animal lysates.Alternatively, as AKIRIN2 has been shown to bind to several gate-forming subunits of the human 20S proteasome and has been suggested to lock the proteasome in an inactive conformation, 19 we speculate that akir-1 depletion might directly affect the proteolytic activity of the 20S proteasome or the entry of the short peptide substrate, which does not require unfolding by the 19S, into the 20S proteasome in the in vitro assay.Remarkably, when we determined proteasomal degradation rate at the subcellular level, i.e., by separately measuring degradation in the nucleus and cytoplasm of intestinal cells, the akir-1 RNAi animals showed a clear increase in the nuclear to cytoplasmic fluorescence ratio of photoconverted UbG76V-Dendra2, revealing that downregulation of akir-1 slows proteasomal degradation in the nucleus.This reduced degradation capacity in the intestinal nucleus reflects the decreased nuclear localization of the proteasome.Together, our in vivo data suggest that downregulation of akir-1 causes a subcellular redistribution of proteasomes and proteasomal degradation capacity in the C. elegans intestine.
We have previously reported tissue-specific variations in proteasome activity and regulatory mechanisms in C. elegans. 34,36,37Interestingly, here we show that although oocytes display a more pronounced reduction in the nuclear proteasome levels compared to intestinal cells, acute accumulation of endogenous polyubiquitinated proteins is not induced in the nuclei of these cells upon loss of akir-1.Cell type-specific differences in proteasomal substrates could potentially contribute to the different response in the nuclear accumulation of polyubiquitinated proteins between oocytes and intestinal cells.It has also been reported that dissected gonads from young D. melanogaster flies display elevated proteasome capacity compared to somatic tissues, 51,52 which could contribute to a faster protein turnover in oocytes.Additionally, it has been reported in human cells and C. elegans that ubiquitinated proteins are exported from the nucleus to the cytoplasm through a UBIN-POST system (UBIN, a ubiquitin-associated [UBA] domain-containing protein; POST, a polyubiquitinated substrate transporter) as a response to proteasome inhibitor treatment. 12As we occasionally observed an accumulation of polyubiquitin-positive staining at the rim of the nuclear membrane in oocytes of akir-1-depleted animals, we speculated that the akir-1 depletion-induced strong reduction in nuclear proteasomes could perhaps mimic a proteasome inhibition at this subcellular compartment and thereby potentially result in nuclear export of polyubiquitinated proteins in oocytes.We tested this hypothesis by separate or combined RNAi knockdown of the C. elegans homologs of UBIN and POST, ubql-1 and F36D4.5, respectively, in the akir-1(gk528) mutant, but we did not detect nuclear accumulation of polyubiquitinated proteins in oocytes, or any increase in polyubiquitinated proteins in the intestinal cell nuclei (data not shown).Our results imply that the UBIN-POST system would not be involved in the akir-1-mediated tissue-specific effect on polyubiquitinated proteins.Due to the complexity of proteins involved in nuclear transport and their crucial roles in various physiological events, dissecting the tissue-specific molecular mechanisms by which AKIR-1 regulates the subcellular localization of polyubiquitinated proteins in C. elegans is not trivial.Thus, more in-depth future studies are required.
In addition to oocytes, body-wall muscle cells also responded differently to akir-1 knockdown compared to the intestinal cells.In the bodywall muscle cells, we observed a later onset of the phenotype, as a slight nuclear accumulation of polyubiquitin-binding reporter was detected only in the next (F 1 ) generation after continuous exposure to akir-1 RNAi treatment.It is unlikely that this later onset is caused by a variation in RNAi efficiency based on the use of an RNAi-sensitive strain and as we have previously shown efficient RNAi capacity in both body-wall muscle cells and intestinal cells. 33This could be due to tissue-specific variation in proteasomal degradation rate, as we have demonstrated slower degradation in the body-wall muscle cells compared to intestinal cells. 34,37We speculate that due to the slower proteasomal degradation rate, body-wall muscle cells maintain functional AKIR-1 proteins longer after akir-1 knockdown compared to intestinal cells.Given the variable effects of AKIR-1 depletion, future studies are required to decipher the contribution of individual molecular mechanisms to the tissue-specific phenotypes.
Overall, our AKIR-1 study demonstrates that the role of Akirins in regulating nuclear proteasome localization is conserved between C. elegans and human cells and that Akirin family members can interact with several nuclear transport proteins.Importantly, our results reveal that akir-1 depletion causes differential outcomes on accumulation of polyubiquitinated proteins in tissues.Lastly, our study suggests a broader role for Akirins in health span regulation and maintenance of cellular protein homeostasis, with a potential tissue-specific impact in multicellular organisms.

Limitations of study
This study utilizes our new and previously generated fluorescent reporter systems 34,[36][37][38]42 to uncover cell-and tissue-specific aspects of akir-1 depletion on the in vivo function of the UPS. Unfrtunately, the expression of our current fluorescent polyubiquitin-binding reporter and photoconvertible UPS activity reporter is yet restricted to a few cell types, excluding, e.g., germ cells, which limits a more comprehensive comparative in vivo analysis of proteasome function.

METHOD DETAILS
C. elegans RNA interference (RNAi) Select gene RNAi was performed using the feeding method as described earlier 55 with a few changes.The HT115(DE3) bacterial strain carrying the empty pL4440 cloning vector was used as a control.A single bacteria colony was cultured overnight at 37 C in Luria broth (LB) medium containing 100 mg/mL ampicillin (Merck KGaA Cat# A0166, Darmstadt, Germany) and 12.5 mg/mL tetracycline (Merck KGaA Cat# T7660).

Native gel analysis
Age-synchronized animals fed with RNAi bacteria were harvested in M9 buffer at first day of adulthood and animal pellets were stored at À80 C. Animal pellets were lysed using a Dounce homogenizer and native gel lysis buffer (50 mM Tris-HCL (pH 8.0), 5 mM Mg 2 Cl 2 , 0.5 mM EDTA, and 1 mM ATP (Sigma-Aldrich Cat# A3377)).Native gel electrophoresis and the in-gel proteasome activity assay were performed as earlier reported with a few exceptions. 34,57Gels were run in an ice bath for 30 min at 20 mA and then for 2 h at 40 mA.The gels were either blotted onto a nitrocellulose membrane prior protein detection with the antibody against proteasome 20S a-subunits or incubated in developing buffer containing 160 mM of fluorogenic proteasome substrate succinyl-leu-leu-val-tyr-7-amino-4-methylcoumarin, suc-LLVY AMC (Bachem Cat# l.1395, Bubendorf, Switzerland) and imaged either with MultiImage Light Cabinet using FluorChem 8900 software (Alpha Innotech Corporation, San Leandro, CA, USA) or Gel Doc XR + System with Image Lab Software (Bio-Rad).Total protein normalization was performed from native gels using Colloidal Blue staining kit (Thermo Fisher Scientific Cat# LC6025) after in-gel proteasome activity assay and from nitrocellulose membranes using Ponceau S staining (Sigma-Aldrich Cat# P3504) after blot transfer.Image analysis were made with Fiji software. 58munofluorescence with dissected C. elegans Age-synchronized animals cultured either on OP50 seeded plates or RNAi feeding plates were harvested at first day of adulthood in M9 buffer.Animals were transferred onto a glass dish and dissected using 27-gauge syringe needles.To immobilize the animals 1 mM levamisole hydrocloride (Merck KGaA Cat# T7660) was used prior making incision close to the pharynx forcing the intestine and gonad to extrude.Dissected animals were fixed with 2xRFB (160 mM KCL, 40 mM NaCl, 20 mM EGTA, 10 mM spermidine, 30 mM PIPES pH 7.4, 50% methanol, and 1% formaldehyde).Additional 100% methanol fixation for 1 min was utilized when antibody against the proteasome 20S alpha subunits was used.Fixed dissected animals were permeabilized using 0.5 or 1% Triton X-100 in PBS and mounted with SlowFade Diamond Antifade Mountant (Thermo Fisher Scientific Cat# S36967).Antibodies against polyubiquitinated proteins, FK1 (RRID:AB_2699340) and the proteasome 20S alpha subunits, MCP231 (RRID:AB_10541045) were used in 1:200 dilution.The following Alexa Fluor 594 conjugated anti-mouse IgM (Thermo Fisher Scientific Cat# A-21044, RRID:AB_2535713) and IgG (Thermo Fisher Scientific Cat# A-11005, RRID:AB_2534073 or Cat# R37121, RRID:AB_2556549) secondary antibodies in 1:200 and 1:100 dilution respectively were used for visualization.The nuclear immunostaining was visually estimated as none, when nuclear staining intensity was similar or less compared to staining intensity in the cytoplasm.A higher nuclear staining intensity compared to the cytoplasmic staining was further estimated visually as weak or strong.DNA was stained with 4 mg/ml Hoechst 33342 (Merck KGaA Cat# B2261).

Lifespan assays and progeny counts
Lifespan experiments were performed at 20 C. Age-synchronized animals were plated on RNAi feeding plates as L1 larvae (day 1).Animals were transferred to a new plate, first every second day, and then every few days after they stopped producing offspring.Animals were checked daily, and animals failing to respond to a gentle prod with a platinum worm pick were classified as dead.Animals crawling off the plate, dying of an extruded gonad, or carrying internally hatched offspring were censored at the time of their death.Basic survival analysis was performed using an online tool, OASIS 2. 59 Data from all three separate lifespan experiments were combined and used for analysis as one dataset.For counting progeny, age-synchronized animals were plated individually on RNAi feeding plates as L1 larvae.Animals were moved to fresh RNAi feeding plates every day, until they stopped producing offspring.Total viable offspring per animal was counted, censoring the offspring of animals that crawled off the plate, died of internally hatched offspring or an extruded gonad before finishing egg-laying.

Microscopy
Age-synchronized animals were imaged at first day of adulthood unless otherwise mentioned.Animals were mounted on 3% agarose pads and immobilized with 1 mM levamisole.Group of 5-15 live animals were imaged with a Zeiss Axio Imager 2 upright wide-field light microscope and a Zeiss EC Plan Neofluar numerical aperture (NA) 10 x 0.3 objective or a Plan Apochromat NA 203 0.8 objective, or with an LSM 780 inverted confocal microscope and a Plan-Neofluor NA 40 x 1.3 or a Plan-Apochromat NA 63 x 1.40 objective, or with a Zeiss LSM 880 inverted confocal microscope and a Plan-Apochromat NA 40 x 1.40 or NA 63 x 1.4 objective (Zeiss, Oberkochen, Germany).All the microscopes run a Zeiss Zen 2 software (Zeiss).For Hoechst nuclear staining, age-synchronized animals were washed with M9, fixed for 2 min using 100% methanol, and permeabilized either with 0.5% or 1% Triton X-100 in PBS.DNA was labeled with 4 mg/ml Hoechst 33342 stain (Merck KGaA Cat# B2261).For photoconversion of the tissue-specific UbG76V-Dendra2, green Dendra2 protein was converted to red using 405-nm UV light.Degradation of the red signal was followed 16-18 (intestine) or 24 (muscle) hours later.For determining the nuclear/cytoplasmic ratio of the red UbG76V-Dendra2 signal, images were taken with the Zeiss LSM 880 inverted confocal microscope and the Plan-Apochromat NA 63 x 1.4 objective, and mean fluorescent intensity of both the nuclei and cytoplasm were quantified using Microscopy Image Browser software. 60

Fluorescence signal quantification
Fluorescence signal was quantified from the original, unmodified TIFF file format images with Fiji software or from the original CZI file format with Zeiss Zen 2 software (Zeiss).In Fiji, sliding paraboloid algorithm was used to subtract background.The mean fluorescence intensity was quantified using an adequate threshold to select fluorescence signal and using the measure function.The brightness of the fluorescence signal has been adjusted in some images using Fiji or Adobe Photoshop (Adobe, San Jose, CA, USA), and all images presented together for comparison were adjusted similarly, unless otherwise stated.In Zeiss Zen 2, profile tab was used to measure fluorescence intensity profiles along a line intersecting the cytoplasm and the nucleus.Hoechst signal was used to determine the nuclear fluorescence.To be able to compare nuclear fluorescence intensity profiles in different cells and between different treatments the nuclear signal along the profiling line was set to start after the same distance.To normalize the fluorescence intensity profiles in control treatment the signal intensity at the starting point was set to 1.The mean intensity signal of the control treatment at the starting point was used to determine the relative fold change in the fluorescence profiles upon the treatment.

QUANTIFICATION AND STATISTICAL ANALYSIS
Quantifications are described in the previous sections under the method details section.The statistical significance was determined using Welch's t-test (two-tailed distribution and unequal variance) using Microsoft Excel 2016 spreadsheet (Microsoft, Redmond, WA, USA).For lifespan analysis, statistical significance was determined with a Mantel-Cox (log rank) test using RStudio software (RStudio, Boston, MA, USA) and the lifespan data used was stratified into three independent experiments.

Figure 1 .
Figure 1.Continued (B) Representative confocal micrographs of control and akir-1 RNAi-treated polyubiquitin reporter animals with Hoechst-visualized nuclei.Insets show enlargements.Scale bar, 20 mm.(C and D) Representative confocal micrographs of polyubiquitin immunostaining (polyUb Ab) in dissected intestines of control and akir-1 RNAi-treated wild-type (N2) animals (C), and in control (N2) animals and akir-1(gk528) mutants (D).Nuclei are visualized with Hoechst.Scale bars, 10 mm.The graphs (on right) show visually quantified nuclear polyubiquitin accumulation in the intestine in control and akir-1 RNAi-treated wild-type animals (C), and in control (N2) and akir-1(gk528) animals (D).n = total number of animals.Proportions of animals with strongly positive, weakly positive, or negative (i.e., less or equal to cytoplasmic immunostaining) nuclear immunostaining are indicated in percentages.See also Figure S1.

Figure 2 .
Figure2.Continued (C) In-gel proteasome activity assay with whole animal lysates of wild-type animals exposed to control or akir-1 RNAi treatment (upper gel).Coomassie staining of the same gel (lower gel).The quantifications show the average fold change in chymotrypsin activity in akir-1 RNAi-treated wild-type animals compared to control RNAi-treated animals (set to 1; n = 5 independent experiments).Proteasome activity is indicated as total (Total; CP + RP-CP + RP2-CP), as core particle (CP), and as CP activity subtracted from total activity (Total À CP).RP, Regulatory particle.(D) Immunoblot analysis of whole animal lysates separated under non-denaturing condition using the antibody against proteasomal 20S alpha subunits.Ponceau S staining was used for total protein normalization.The quantification graph shows the average fold change of core particles (CP) in akir-1 RNAi-treated wild-type animals compared to control RNAi-treated animals (set to 1; n = 3 independent experiments).(E) Representative fluorescence micrographs of control and akir-1 RNAi-treated transgenic C. elegans expressing photoconvertible UbG76V-Dendra2 reporter (vha-6p::UbG76V::Dendra2) in intestinal cells.The 0 h (left panels) and 18 h (right panels) indicate time after photoconversion.Scale bar, 500 mm.The graph shows the mean percentages of fluorescence intensity of the photoconverted UbG76V-Dendra2 18 h after the photoconversion relative to the fluorescence at the point of photoconversion (0 h, set as 100%); n = 6 independent experiments with triplicate images of 6-7 animals per image (total number of animals is 108 per treatment).(F) Representative confocal fluorescence micrographs of intestinal cells with photoconverted UbG76V-Dendra2 (18 h after conversion) in transgenic C. elegans treated with control or akir-1 RNAi.Scale bar, 10 mm.The graph shows the ratio between nuclear and cytoplasmic mean fluorescence per cell.n = 2 independent experiments (total number of nuclei is 23 in control RNAi and 27 in akir-1 RNAi treatment).Welch's t-test (two-tailed distribution and unequal variance) was used for statistical analyses.Error bars, SD; ns, not significant; *p < 0,05; **p < 0,01; ***p < 0.001.See also FigureS2.

1 Figure 3 .
Figure 3. Intestinal nuclei display reduced proteasome levels upon akir-1 depletion (A) Representative micrographs of proteasome immunostaining (20S Ab) in dissected intestines of control (N2) animals and akir-1(gk528) mutants.The graph shows the normalized mean G SD intensity profiles of 20S immunofluorescence measured along the line intersecting the cytoplasm and the nucleus as shown in the image above the graph.Orange line represents the profiling line, dashed line represents the nucleus.During the profiling, Hoechst signal was used to determine the nuclear 20S immunofluorescence.n = total number of nuclei is 26 in control (N2) and 44 in akir-1(gk528) animals.(B) Representative confocal micrographs showing GFP fluorescence ratio between nuclei and cytoplasm of control and akir-1 RNAi-treated rpt-5p::GFP::RPT-5 animals.Intestinal cells are outlined with white dashed lines and white arrows point to intestinal cell nuclei (C) Representative confocal micrographs of ubh-4p::UBH-4::GFP animals.Nuclei are visualized with Hoechst.The graphs show the normalized mean G SD intensity profiles of fluorescence measured along the line intersecting the cytoplasm and the nucleus.n = total number of nuclei is 13 in control RNAi and 17 in akir-1 RNAi treatment (B) and 8 nuclei in control RNAi and 9 nuclei in akir-1 RNAi treatment (C).Scale bars, 20 mm.Error bars, SD.See also Figure S3.

Figure 4 .
Figure 4. Nuclear proteasome expression decreases in oocytes upon akir-1 depletion (A) Representative confocal micrographs of proteasome immunostaining (20S Ab) in dissected oocytes of control (N2) animals and akir-1(gk528) mutants.The graphs show the normalized mean G SD intensity profiles of 20S immunofluorescence measured along the line intersecting the cytoplasm and the nucleus.n = total number of nuclei is 10 in control (N2) and 11 in akir-1(gk528) animals.(B) Representative confocal micrographs showing GFP fluorescence in oocytes of control and akir-1 RNAi-treated ubh-4p::UBH-4::GFP animals.The graph shows the normalized mean G SD intensity profiles of fluorescence measured along the line intersecting the cytoplasm and the nucleus.n = total number of nuclei is 14 in control RNAi and 10 in akir-1 RNAi treatment.White arrows point to oocyte nuclei.Nuclei are visualized with Hoechst.Scale bars, 10 mm.Error bars, SD.See also Figure S4.

Figure 5 .B 1 RFigure 6 .
Figure 5.The effects of akir-1 depletion on nuclear accumulation of polyubiquitinated proteins in oocytes and body-wall muscle cells (A) Representative micrographs of polyubiquitin immunostaining (polyUb Ab) in dissected oocytes of control (N2) animals and akir-1(gk528) mutants.Nuclei are visualized with Hoechst.Scale bar, 20 mm.White arrows point to representative nuclei.(B) Representative fluorescence micrographs of F1 generation of control and akir-1 RNAi-treated rrf-3(pk1426) animals expressing the polyubiquitin (polyUb) reporter (unc-54p::UIM2-ZsProSensor) in the body-wall muscle cells.Nuclei are visualized with Hoechst.Insets show enlargements of the indicated areas.White arrows point to representative nuclei.Scale bar, 10 mm.The graph (on right) shows quantified subcellular localization of polyUb reporter fluorescence in the body-wall muscle cells.Welch's t-test (two-tailed distribution and unequal variance) was used for statistical analysis.n = total number of nuclei; **p < 0,01.See also Figure S5.