A tRNA processing enzyme is a key regulator of the mitochondrial unfolded protein response

  1. James P Held
  2. Gaomin Feng
  3. Benjamin R Saunders
  4. Claudia V Pereira
  5. Kristopher Burkewitz
  6. Maulik R Patel  Is a corresponding author
  1. Department of Biological Sciences, Vanderbilt University, United States
  2. Department of Cell and Developmental Biology, Vanderbilt University, United States
  3. Diabetes Research and Training Center, Vanderbilt University School of Medicine, United States

Abstract

The mitochondrial unfolded protein response (UPRmt) has emerged as a predominant mechanism that preserves mitochondrial function. Consequently, multiple pathways likely exist to modulate UPRmt. We discovered that the tRNA processing enzyme, homolog of ELAC2 (HOE-1), is key to UPRmt regulation in Caenorhabditis elegans. We find that nuclear HOE-1 is necessary and sufficient to robustly activate UPRmt. We show that HOE-1 acts via transcription factors ATFS-1 and DVE-1 that are crucial for UPRmt. Mechanistically, we show that HOE-1 likely mediates its effects via tRNAs, as blocking tRNA export prevents HOE-1-induced UPRmt. Interestingly, we find that HOE-1 does not act via the integrated stress response, which can be activated by uncharged tRNAs, pointing toward its reliance on a new mechanism. Finally, we show that the subcellular localization of HOE-1 is responsive to mitochondrial stress and is subject to negative regulation via ATFS-1. Together, we have discovered a novel RNA-based cellular pathway that modulates UPRmt.

Editor's evaluation

This manuscript reports a novel RNA-based cellular pathway that modulates mitochondrial UPR (UPRmt). It advances our understanding of the mitochondrial-to-nuclear communication mediated by a tRNA processing enzyme.

https://doi.org/10.7554/eLife.71634.sa0

Introduction

Mitochondria are central to a myriad of cellular processes including energy production, cellular signaling, biogenesis of small molecules, and regulation of cell death via apoptosis (Nunnari and Suomalainen, 2012). Mitochondrial dysfunction can lead to metabolic and neurological disorders, cardiovascular disease, and cancers (Vafai and Mootha, 2012). To maintain proper mitochondrial function cellular mechanisms have evolved that respond to, and mitigate, mitochondrial stress (Baker et al., 2012; Wang and Chen, 2015; Wrobel et al., 2015; Munkácsy et al., 2016; Tjahjono and Kirienko, 2017; Weidberg and Amon, 2018; Naresh and Haynes, 2019; Fessler et al., 2020; Guo et al., 2020).

One of the predominant mitochondrial stress response mechanisms is the mitochondrial unfolded protein response (UPRmt). Although first discovered in mammals (Zhao et al., 2002), UPRmt has been best characterized in Caenorhabditis elegans (Naresh and Haynes, 2019). UPRmt is primarily characterized by transcriptional upregulation of genes whose products respond to and ameliorate mitochondrial stress (Yoneda et al., 2004; Nargund et al., 2012).

In C. elegans, activation of UPRmt relies on the transcription factor ATFS-1 that primarily localizes to mitochondria, but under mitochondrial-stress conditions is trafficked to the nucleus where it drives the expression of mitochondrial stress response genes (Haynes et al., 2010; Nargund et al., 2012; Nargund et al., 2015). However, it has become increasingly apparent that UPRmt is under multiple levels of control: Mitochondrial stress in neurons can activate intestinal UPRmt non-cell-autonomously via retromer-dependent Wnt signaling (Durieux et al., 2011; Berendzen et al., 2016; Zhang et al., 2018); overexpression of two conserved histone demethylases are independently sufficient to activate UPRmt (Merkwirth et al., 2016); and ATFS-1 is post-translationally modified to facilitate its stability and subsequent UPRmt activation (Gao et al., 2019). Given mitochondrial integration into many diverse cellular signaling and metabolic pathways, there are likely yet-to-be identified pathways regulating UPRmt.

In conducting a small-scale RNAi screen to interrogate the effects of perturbing mitochondrial RNA processing we discovered that the 3’-tRNA zinc phosphodiesterase, homolog of ELAC2 (HOE-1), is a key regulator of UPRmt in C. elegans. ELAC2 is an essential endonuclease that cleaves 3’-trailer sequences from nascent tRNAs—a necessary step of tRNA maturation—in both nuclei and mitochondria (Nashimoto et al., 1999; Mayer et al., 2000; Schiffer et al., 2002; Takaku et al., 2003; Dubrovsky et al., 2004; Brzezniak et al., 2011; Sanchez et al., 2011; Siira et al., 2018). ELAC2 has also been reported to cleave other structured RNAs yielding tRNA fragments, small nucleolar RNAs (snoRNAs) and micro RNAs (miRNAs) (Kruszka et al., 2003; Lee et al., 2009; Bogerd et al., 2010; Siira et al., 2018). In humans, mutations in ELAC2 are associated with hypertrophic cardiomyopathy (Haack et al., 2013; Shinwari et al., 2017; Saoura et al., 2019) and prostate cancer (Tavtigian et al., 2001; Korver et al., 2003; Noda et al., 2006) while in C. elegans, loss of HOE-1 has been shown to compromise fertility (Smith and Levitan, 2004).

Surprisingly, we find that it is not the mitochondrial, but rather the nuclear activity of HOE-1 that is required for activation of UPRmt. Remarkably, compromising nuclear export of HOE-1 is sufficient to specifically and robustly activate UPRmt. Blocking tRNA export from the nucleus suppresses this HOE-1-dependent UPRmt induction, suggesting that HOE-1 generates RNA species required in the cytosol to trigger UPRmt. Finally, we show that HOE-1 nuclear levels are dynamically regulated under conditions of mitochondrial stress, supporting a physiological role for HOE-1 in mitochondrial stress response. Taken together, our results provide a novel mechanism by which UPRmt is regulated as well as provide critical insight into the biological role of the conserved tRNA processing enzyme, HOE-1.

Results

hoe-1 is required for maximal UPRmt activation

We discovered that RNAi against hoe-1, a gene encoding a 3’-tRNA phosphodiesterase, attenuates hsp-6p::GFP induction—a fluorescence based transcriptional reporter of UPRmt activation (Yoneda et al., 2004). Knockdown of hoe-1 by RNAi is sufficient to attenuate UPRmt reporter activation induced by a loss-of-function mutation in the mitochondrial electron transport chain (ETC) complex I subunit NUO-6 (nuo-6(qm200)) (Figure 1A and B).

hoe-1 is required for maximal UPRmt activation.

(A) Fluorescence images of UPRmt reporter (hsp-6p::GFP) activation in L4 nuo-6(qm200) animals on control and hoe-1 RNAi. Scale bar 200 μm. (B) Fluorescence intensity quantification of hsp-6p::GFP in individual L4 nuo-6(qm200) animals on control and hoe-1 RNAi normalized to hsp-6p::GFP in a wildtype background on control RNAi (n = 8 and 15 respectively, mean and SD shown, unpaired t-test). (C) Fluorescence images of UPRmt reporter (hsp-6p::GFP) activation in L3/L4 wildtype and hoe-1 null (hoe-1(-/-)) animals on control, cco-1, and spg-7 RNAi. Scale bar 200 μm. (D) Fluorescence intensity quantification of hsp-6p::GFP in individual L3/L4 wildtype and hoe-1(-/-) animals on control and cco-1 RNAi (n = 8,12,6 and 13 respectively, mean and SD shown, ordinary two-way ANOVA with Tukey’s multiple comparisons test). (E) Fluorescence intensity quantification of hsp-6p::GFP in individual L3/L4 wildtype and hoe-1(-/-) animals on control and spg-7 RNAi (n = 7,15,6 and 18 respectively, mean and SD shown, ordinary two-way ANOVA with Tukey’s multiple comparisons test). (F) Fluorescence images of UPRmt reporter (hsp-6p::GFP) activation in L3/L4 nuo-6(qm200) animals with (hoe-1(+/+)) and without (hoe-1(-/-)) hoe-1. Scale bar 200 μm. (G) Fluorescence intensity quantification of hsp-6p::GFP in individual L3/L4 nuo-6(qm200) animals with (hoe-1(+/+)) and without (hoe-1(-/-)) hoe-1 normalized to hsp-6p::GFP in a wildtype background (n = 22 for each condition, mean and SD shown, unpaired t-test).

To further interrogate the potential role of hoe-1 in UPRmt regulation, we used CRISPR/Cas9 to generate a hoe-1 null mutant (hoe-1(-/-)) by deleting the open reading frame of hoe-1 (Dokshin et al., 2018). The hoe-1 null mutants do not develop past late larval stage 3, thus the allele is maintained over a balancer chromosome, tmC25 (Dejima et al., 2018). UPRmt induced by the knockdown of both the mitochondrial protease, spg-7, and ETC complex IV subunit, cco-1, is robustly attenuated in hoe-1 null animals (Figure 1C–E). Furthermore, UPRmt induced by nuo-6(qm200) is attenuated in hoe-1 null animals similarly to what is seen in nuo-6(qm200) animals on hoe-1 RNAi (Figure 1F and G). Taken together, these findings suggest that HOE-1 is generally required for maximal UPRmt activation.

HOE-1 is dual-targeted to nuclei and mitochondria

To better understand the role of HOE-1 in UPRmt regulation, we sought to identify where HOE-1 functions in the cell. HOE-1 is predicted to localize to both nuclei and mitochondria and this dual-localization has been shown for HOE-1 orthologs in Drosophila, mice, and human cell lines (Dubrovsky et al., 2004; Brzezniak et al., 2011; Rossmanith, 2011; Siira et al., 2018). To determine the subcellular localization of HOE-1 in C. elegans, we C-terminally tagged HOE-1 with GFP at its endogenous locus (HOE-1::GFP). Both hoe-1::GFP homozygous and hoe-1::GFP/hoe-1(-/-) trans-heterozygous animals grow and develop indistinguishably from wildtype animals suggesting that GFP-tagging HOE-1 does not compromise its essential functions (Figure 2—figure supplement 1A). We found that HOE-1 localizes to both mitochondria and nuclei (Figure 2A).

Figure 2 with 5 supplements see all
Nuclear HOE-1 is required for maximal UPRmt activation.

(A) Fluorescence images of a terminal intestinal cell in a wildtype animal expressing HOE-1::GFP (green) stained with TMRE (magenta) to visualize mitochondria. GFP and TMRE co-localization shown in white in merged image. Arrow indicates nuclei. Scale bar 20 μm. Representative line segment analysis of individual mitochondrion. (B) Schematic of HOE-1 protein showing the mitochondrial targeting sequence (MTS) and nuclear localization signals (NLS). ΔMTS allele created by replacing START codon with an alanine (M1A). Transcription begins at M74 for nuclear localized HOE-1. ΔNLS allele created by compromising the most N-terminal NLS (636KRPR > AAPA). (C) Fluorescence images of UPRmt reporter (hsp-6p::GFP) in L4 wildtype and hoe-1(ΔMTS) animals on control and spg-7 RNAi. Scale bar 200 μm. (D) Fluorescence intensity quantification of hsp-6p::GFP in individual L4 wildtype and hoe-1(ΔMTS) animals on control and spg-7 RNAi (n = 15,20,17, and 19 respectively, mean and SD shown, ordinary two-way ANOVA with Tukey’s multiple comparisons test). (E) Fluorescence images of UPRmt reporter (hsp-6p::GFP) in L4 wildtype and hoe-1(ΔNLS) animals on control and spg-7 RNAi. Scale bar 200 μm. (F) Fluorescence intensity quantification of hsp-6p::GFP in individual L4 wildtype and hoe-1(ΔNLS) animals on control and spg-7 RNAi (n = 15 for each condition, mean and SD shown, ordinary two-way ANOVA with Tukey’s multiple comparisons test). (G) Fluorescence images of UPRmt reporter in L4 nuo-6(qm200) animals in wildtype and hoe-1(ΔNLS) backgrounds. Scale bar 200 μm. (H) Fluorescence intensity of hsp-6p::GFP in individual L4 nuo-6(qm200) animals in wildtype and hoe-1(ΔNLS) backgrounds (n = 30 for each condition, mean and SD shown, unpaired t-test). (I) mRNA transcript quantification of hsp-6 in L4 wildtype and hoe-1(ΔNLS) animals on control and spg-7 RNAi normalized to ama-1 (n = 4 for each condition, mean and SD shown, ordinary two-way ANOVA with Tukey’s multiple comparisons test).

Mitochondrial HOE-1 is not required for UPRmt activation

Given the dual-localization of HOE-1, we questioned whether it is mitochondrial or nuclear HOE-1 that is required for UPRmt activation. To address this question, we created mitochondrial and nuclear compartment-specific loss-of-function mutants of HOE-1 (Figure 2B). hoe-1 contains two functional start codons. Translation beginning from the first start codon (encoding methionine 1 (M1)) produces a protein containing a mitochondrial targeting sequence (MTS). Translation beginning from the second start codon (encoding methionine 74 (M74)), which is 3’ to the MTS, produces a nuclear specific protein. This feature is conserved in human ELAC2 and it has been shown that mutating M1 to an alanine produces a mitochondrial-specific knockout (Brzezniak et al., 2011). Thus, we used the same approach to create a mitochondrial-specific knockout of C. elegans HOE-1 (hoe-1(ΔMTS)). This mutation is sufficient to strongly attenuate mitochondrial targeting without impacting nuclear localization (Figure 2—figure supplement 2A).

UPRmt reporter activation by spg-7 and cco-1 RNAi is not attenuated in hoe-1(ΔMTS) animals (Figure 2C and D and Figure 2—figure supplement 3A and B). In fact, UPRmt reporter activation is slightly elevated in hoe-1(ΔMTS) animals relative to wildtype. These data suggest that mitochondrial HOE-1 is not required for UPRmt activation.

Nuclear HOE-1 is required for UPRmt activation

HOE-1 is predicted to contain two nuclear localization signals (NLS). Given that hoe-1 null mutant animals are developmentally arrested and hoe-1(ΔMTS) animals are superficially wildtype we reasoned that completely ablating nuclear localization of HOE-1 may result in recapitulation of the null phenotype. In effort to disentangle the developmental effects from the effect on UPRmt we ablated only one of the nuclear localization signals of HOE-1. To compromise nuclear localization, we mutated the positively charged residues of the most N-terminal NLS to alanines (hoe-1(ΔNLS)). These mutations are sufficient to strongly attenuate, but not completely ablate, HOE-1 nuclear localization whilst still allowing animals to develop to adulthood (Figure 2—figure supplement 4A–C).

In contrast to loss of mitochondrial HOE-1, loss of nuclear HOE-1 robustly attenuates UPRmt reporter activation induced by spg-7 RNAi (Figure 2E and F) and attenuates UPRmt reporter activation induced by nuo-6(qm200) (Figure 2G and H). Furthermore, loss of nuclear HOE-1 attenuates the transcriptional upregulation of UPRmt target genes hsp-6 (Figure 2I) and cyp-14A1.4 (Figure 2—figure supplement 5A) under conditions of mitochondrial stress. Together these data suggest that HOE-1 is required in the nucleus to facilitate UPRmt activation.

Compromising HOE-1 nuclear export is sufficient to activate UPRmt

Like many nuclear localized proteins (la Cour et al., 2004), HOE-1 has both nuclear localization signals and a nuclear export signal (NES). Given that loss of nuclear HOE-1 results in UPRmt attenuation we questioned if compromising HOE-1 nuclear export, by ablating the NES of HOE-1, is sufficient to activate UPRmt. We created a HOE-1 NES knockout mutant (hoe-1(ΔNES)) by replacing the strong hydrophobic residues of the predicted NES with alanines (Figure 3—figure supplement 1A). hoe-1(ΔNES) animals are superficially wildtype in their development but are sterile. Thus, the allele is balanced with tmC25. Homozygous hoe-1(ΔNES) animals have elevated nuclear HOE-1 accumulation relative to wildtype (Figure 3—figure supplement 1B, Figure 2—figure supplement 4B and C).

Strikingly, the UPRmt reporter hsp-6p::GFP is robustly activated in adult hoe-1(ΔNES) animals similarly to that seen in mitochondrial stressor nuo-6(qm200) and constitutive UPRmt activation in atfs-1 gain-of-function (atfs-1(et15)) mutant animals (Figure 3A and B). hoe-1(ΔNES) also mildly induces the less sensitive UPRmt reporter hsp-60p::GFP (Figure 3—figure supplement 2A and B).

Figure 3 with 5 supplements see all
Nuclear export defective HOE-1 is sufficient to specifically activate UPRmt.

(A) Fluorescence images of UPRmt reporter (hsp-6p::GFP) activation in day 2 adult wildtype, nuo-6(qm200), atfs-1(et15), and hoe-1(ΔNES) animals. Scale bar 200 μm. (B) Fluorescence intensity quantification of hsp-6p::GFP in individual day 2 adult wildtype, nuo-6(qm200), atfs-1(et15), and hoe-1(ΔNES) animals (n = 10 for each condition, mean and SD shown, ordinary one-way ANOVA with Tukey’s multiple comparisons test). (C–E) mRNA transcript quantification of hsp-6, clec-47, and cyp-14A1.4, respectively, in day 2 adult wildtype and hoe-1(ΔNES) animals normalized to ama-1 mRNA levels (n = 4 for each condition, mean and SD shown, unpaired t-test). (F) Fluorescence images of UPRmt reporter (hsp-6p::GFP) activation in day 2 adult hoe-1(ΔNES) animals on control and atfs-1 RNAi. Scale bar 200 μm. (G) Fluorescence intensity quantification of hsp-6p::GFP in individual day 2 adult wildtype and hoe-1(ΔNES) animals on control and atfs-1 RNAi (n = 10 for each condition, mean and SD shown, ordinary two-way ANOVA with Tukey’s multiple comparisons test). (H) Fluorescence images of UPRER reporter (hsp-4p::GFP) activation in day 2 adult wildtype and hoe-1(ΔNES) animals. Scale bar 200 μm. (I) Fluorescence intensity quantification of hsp-4p::GFP in individual day 2 adult wildtype and hoe-1(ΔNES) animals (n = 10 for each condition, mean and SD shown, unpaired t-test). (J) Fluorescence images of intestinal-specific basal protein reporter (ges-1p::GFPcyto) activation in day 2 adult wildtype and hoe-1(ΔNES) animals. Scale bar 200 μm. (K) Fluorescence intensity quantification of ges-1p::GFPcyto in individual day 2 adult wildtype and hoe-1(ΔNES) animals (n = 10 for each condition, mean and SD shown, unpaired t-test).

UPRmt activation is characterized by the transcriptional upregulation of a suite of mitochondrial stress response genes that encode chaperone proteins, proteases, and detoxification enzymes that function to restore mitochondrial homeostasis (Nargund et al., 2012). To interrogate the extent of UPRmt induction in hoe-1(ΔNES) animals, we measured transcript levels of a diverse set of UPRmt associated genes. We found that the UPRmt genes encoding a chaperone protein (hsp-6), stress response involved C-type lectin (clec-47), and P450 enzyme (cyp-14A4.1) are all upregulated in hoe-1(ΔNES) animals (Figure 3C, D and E). These data support hoe-1(ΔNES) being sufficient to activate the UPRmt transcriptional response.

UPRmt activation is dependent upon the transcription factor ATFS-1 (Haynes et al., 2010; Nargund et al., 2012). Thus, we tested if UPRmt reporter activation in hoe-1(ΔNES) animals is ATFS-1 dependent. Knockdown of atfs-1 is sufficient to completely attenuate UPRmt reporter activation in hoe-1(ΔNES) animals (Figure 3F and G), showing that UPRmt induction by hoe-1(ΔNES) is ATFS-1 dependent.

Elevated nuclear HOE-1 levels in hoe-1(ΔNES) animals is likely responsible for UPRmt activation

To further interrogate how UPRmt is activated in hoe-1(ΔNES) animals, we made double localization mutants of hoe-1. If UPRmt is activated in hoe-1(ΔNES) animals due to elevated nuclear HOE-1 levels we reasoned that introducing a hoe-1(ΔNLS) mutation in the hoe-1(ΔNES) background (hoe-1(ΔNLS+ΔNES)) should be sufficient to attenuate UPRmt activation. Indeed, hoe-1(ΔNLS+ΔNES) animals have UPRmt reporter activation comparable to wildtype animals (Figure 3—figure supplement 3A and B). Furthermore, we reasoned that compromising mitochondrial localization of HOE-1 in a hoe-1(ΔNES) background (hoe-1(ΔMTS+ΔNES)) may further enhance hoe-1(ΔNES)-induced UPRmt activation as what would be the mitochondrial targeted HOE-1 pool should be diverted to the nucleus as well. Consistent with our hypothesis, hoe-1(ΔMTS+ΔNES) animals have even higher activation of UPRmt than hoe-1(ΔNES) alone (Figure 3—figure supplement 4A and B). Taken together, these data strongly suggest that hoe-1(ΔNES) triggers UPRmt activation due to elevated nuclear HOE-1 levels.

Compromising HOE-1 nuclear export activates UPRmt cell-autonomously in the intestine

Contrary to UPRmt induced by nuo-6(qm200) and atfs-1(et15), hoe-1(ΔNES) animals appear to have UPRmt activated specifically in the intestine (Figure 3A). We questioned if this UPRmt activation is occurring cell autonomously or non-cell autonomously as UPRmt has been shown to be able to be signaled across tissues, particularly from neurons to intestine (Durieux et al., 2011; Berendzen et al., 2016; Zhang et al., 2018). To determine which tissue HOE-1 is required in for UPRmt activation we took advantage of the auxin-inducible degradation (AID) system that allows for tissue-specific protein degradation (Zhang et al., 2015). Briefly, degron-tagged proteins will be degraded in the presence of the plant hormone auxin but only in tissues wherein E3 ubiquitin ligase subunit, TIR1, is expressed. We C-terminally degron-tagged hoe-1(ΔNES) (hoe-1(ΔNES)::degron) and crossed this allele into backgrounds in which TIR1 is driven under an intestinal-specific (ges-1p::TIR1) or a neuronal-specific (rgef-1p::TIR) promoter (Ashley et al., 2021). hoe-1(ΔNES)-induced UPRmt is only attenuated when HOE-1 is selectively degraded in the intestine (Figure 3—figure supplement 5A and B). This data strongly suggests that compromised nuclear export of HOE-1 activates UPRmt cell-autonomously in the intestine.

Compromising HOE-1 nuclear export specifically activates UPRmt

Changes in protein synthesis rates and associated protein folding capacity can broadly activate cellular stress response mechanisms (Wang and Kaufman, 2016; Das et al., 2017; Boos et al., 2019). Given the role of hoe-1 in tRNA maturation, we questioned if the robust upregulation of UPRmt in hoe-1(ΔNES) animals may be the result of compromised cellular proteostasis in general rather than specific activation of UPRmt. One stress response mechanism that is sensitive to global proteotoxic stress is the endoplasmic reticulum unfolded protein response (UPRER) (Preissler and Ron, 2019). We find that the UPRER reporter hsp-4p::GFP is not activated in hoe-1(ΔNES) animals (Figure 3H and I), suggesting that hoe-1(ΔNES) does not cause ER stress nor cellular proteotoxic stress. Additionally, a basal reporter of GFP that has been used to proxy general protein expression (Gitschlag et al., 2016), ges-1p::GFPcyto, is only mildly upregulated in hoe-1(ΔNES) animals relative to wildtype (Figure 3J and K). Together these findings support that impaired nuclear export of HOE-1 specifically activates UPRmt.

Compromising HOE-1 nuclear export reduces mitochondrial membrane potential

UPRmt is known to be triggered when mitochondrial membrane potential is compromised (Rolland et al., 2019; Shpilka et al., 2021). Thus, we assessed mitochondrial membrane potential, using TMRE staining, in adult hoe-1(ΔNES) animals where UPRmt is robustly activated. Consistent with UPRmt activation, we found that mitochondrial membrane potential is severely depleted in adult hoe-1(ΔNES) animals (Figure 4A and B). However, hoe-1(ΔNLS) animals also exhibit compromised mitochondrial membrane potential without UPRmt activation suggesting that decreased membrane potential does not guarantee UPRmt induction (Figure 4A and B). Compromised mitochondrial membrane potential can be both a cause and consequence of UPRmt activation (Rolland et al., 2019; Shpilka et al., 2021). Thus, we assessed whether or not compromised membrane potential in hoe-1(ΔNES) animals is atfs-1-dependent. Mitochondrial membrane potential is not rescued in hoe-1(ΔNES) animals on atfs-1 RNAi (Figure 4C and D) suggesting that reduced mitochondrial membrane potential in hoe-1(ΔNES) animals is not a result of UPRmt activation. Taken together, these data show that compromised nuclear export of HOE-1 results in depletion of mitochondrial membrane potential. Furthermore, this depletion in membrane potential correlates with UPRmt activation, consistent with the possibility that hoe-1(ΔNES) activates UPRmt via depletion of mitochondrial membrane potential.

Nuclear export defective HOE-1 activates UPRmt, correlating with reduced mitochondrial membrane potential.

(A) Fluorescence images of TMRE stained day 1 adult wildtype, hoe-1(ΔNES), and hoe-1(ΔNLS) individuals. Scale bar 20 μm. (B) Fluorescence intensity quantification of TMRE staining in individual day 1 adult wildtype, hoe-1(ΔNES), and hoe-1(ΔNLS) animals (n = 57, 60, and 63 respectively, mean and SD shown, ordinary one-way ANOVA with Tukey’s multiple comparisons test). (C) Fluorescence images of TMRE stained day 1 adult wildtype and hoe-1(ΔNES) animals on control and atfs-1 RNAi. Scale bar 20 μm. (D) Fluorescence intensity quantification of TMRE staining in individual day 1 adult wildtype and hoe-1(ΔNES) animals on control and atfs-1 RNAi (n = 65, 62, 65, and 61 respectively, mean and SD shown, ordinary two-way ANOVA with Tukey’s multiple comparisons test).

Compromising HOE-1 nuclear export elevates nuclear levels of UPRmt transcription factors ATFS-1 and DVE-1

UPRmt activation is dependent upon nuclear accumulation of the transcription factor ATFS-1 (Nargund et al., 2012; Nargund et al., 2015). Thus, we tested if ATFS-1 accumulates in nuclei of hoe-1(ΔNES) animals by assessing the fluorescence intensity of ectopically expressed mCherry-tagged ATFS-1 (atfs-1p::ATFS-1::mCherry) in wildtype, hoe-1(ΔNES), and mitochondrial-stressed nuo-6(qm200) animals. Both hoe-1(ΔNES) and nuo-6(qm200) animals have elevated nuclear accumulation of ATFS-1 relative to wildtype (Figure 5A and B). However, while nuo-6(qm200) animals exhibit elevated levels of total cellular and extranuclear levels of ATFS-1::mCherry relative to wildtype, hoe-1(ΔNES) animals do not (Figure 5C and Figure 5—figure supplement 1A). We find that atfs-1 mRNA levels are also elevated in hoe-1(ΔNES) animals relative to wildtype comparable to that seen in nuo-6(qm200) animals (Figure 5D).

Figure 5 with 1 supplement see all
Nuclear export defective HOE-1 animals have increased nuclear accumulation of UPRmt transcription factors ATFS-1 and DVE-1.

(A), Fluorescence images of ATFS-1::mCherry in the terminal intestine of day 2 adult wildtype hoe-1(ΔNES), and nuo-6(qm200) individuals (tip of the tail is in the bottom of each panel). Intestinal nuclei outlined with dashed white line. Scale bar 20 μm. (B) Fluorescence intensity quantification of nuclear ATFS-1::mCherry in wildtype, hoe-1(ΔNES), and nuo-6(qm200) individuals (n = 65, 74, and 72 respectively, mean and SD shown, ordinary one-way ANOVA with Tukey’s multiple comparisons test). (C) Fluorescence intensity quantification of total cellular ATFS-1::mCherry in wildtype, hoe-1(ΔNES), and nuo-6(qm200) individuals (n = 61, 62, and 67 respectively, mean and SD shown, ordinary one-way ANOVA with Tukey’s multiple comparisons test). (D) mRNA transcript quantification of atfs-1 in day 2 adult wildtype, nuo-6(qm200), and hoe-1(ΔNES) animals normalized to ama-1 (n = 4 for each condition, mean and SD shown, ordinary one-way ANOVA with Tukey’s multiple comparisons test). (E) Fluorescence images of dve-1p::DVE-1::GFP in day 2 adult wildtype and hoe-1(ΔNES) animals. Scale bar 200 μm. (F) Number of intestinal cell nuclei with DVE-1::GFP puncta above brightness threshold of 25 in day 2 adult wildtype and hoe-1(ΔNES) animals (n = 33 and 41 respectively, unpaired t-test). (G) Western blot for DVE-1::GFP and actin from day 1 adult wildtype and hoe-1(ΔNES) animals. (H) Quantification of DVE-1::GFP western blot band intensity from day 1 adult wildtype and hoe-1(ΔNES) animals normalized to total protein (n = 4 for each condition, mean and SD shown, unpaired t-test).

Figure 5—source data 1

Blots for wildtype and hoe-1(ΔNES) animals with DVE-1::GFP (Figure 4G and H).

All panels are the same membrane. (A) Image of stain-free blot for total protein from day 1 adult wildtype and hoe-1(ΔNES) animals. Four biological replicates of each condition: Lane #1 BR Spectra Protein Ladder – ladder bands in kDa denoted, Lane #2–5 wildtype and #6–9 hoe-1(ΔNES). (B) Chemiluminescence image of blot for DVE-1::GFP using GFP primary antibody. (C) Composite image of chemiluminescence and colorimetric images of blot for DVE-1::GFP to show bands relative to ladder. (D) Chemiluminescence image of blot for actin using β-actin primary antibody. (E) Composite image of chemiluminescence and colorimetric images of blot for actin to show bands relative to ladder.

https://cdn.elifesciences.org/articles/71634/elife-71634-fig5-data1-v3.zip

The transcription factor DVE-1 is required for full UPRmt activation (Haynes et al., 2007; Tian et al., 2016). Thus, we asked if DVE-1::GFP nuclear expression is higher in hoe-1(ΔNES) than in wildtype animals. We found that accumulation of DVE-1::GFP in intestinal cell nuclei is significantly higher in hoe-1(ΔNES) than in wildtype animals (Figure 5E and F). Qualitatively, cellular DVE-1::GFP levels appear mildly elevated in hoe-1(ΔNES) animals based on actin (Figure 5G, Figure 5—source data 1), though the difference in DVE-1::GFP levels is not significant when normalized to total protein (Figure 5H). Thus, while we cannot rule out the possibility of a slight increase in the cellular levels of DVE-1, elevation in the nuclear localization of DVE-1 in hoe-1(ΔNES) animals is the more robust phenotype. Together, these data suggest that UPRmt induction in hoe-1(ΔNES) animals is a result of increased nuclear accumulation of UPRmt transcription factors ATFS-1 and DVE-1.

UPRmt is activated by altered tRNA processing in animals with compromised HOE-1 nuclear export

The canonical role of HOE-1 is to cleave 3’-trailer sequences from nascent tRNAs (Nashimoto et al., 1999; Mayer et al., 2000; Schiffer et al., 2002; Takaku et al., 2003; Dubrovsky et al., 2004; Brzezniak et al., 2011; Sanchez et al., 2011; Siira et al., 2018). This enzymatic function is dependent upon zinc binding (Ma et al., 2017; Bienert et al., 2017). Thus, we queried if UPRmt activation by hoe-1(ΔNES) is dependent upon the catalytic activity of HOE-1. To test this, we generated a catalytically-dead HOE-1 mutant by changing an essential aspartate of the zinc-binding pocket of HOE-1 to alanine in both a wildtype (hoe-1(D624A)) and hoe-1(ΔNES) (hoe-1(D624A+ΔNES)) background. Animals homozygous for D624A recapitulate the growth arrest phenotype of the hoe-1 null mutant precluding us from assessing the impact of D624A on UPRmt induction in adult hoe-1(ΔNES) animals. To overcome this constraint, we assessed UPRmt activation in hoe-1(ΔNES) versus hoe-1(ΔNES)/hoe-1(D624A+ΔNES) trans-heterozygous animals. A single copy of catalytically dead hoe-1 is sufficient to attenuate hoe-1(ΔNES)-induced UPRmt (Figure 6—figure supplement 1A and B). These data suggest that hoe-1(ΔNES)-induced UPRmt requires the catalytic activity of HOE-1.

Given that HOE-1 catalytic activity is required for UPRmt, we further interrogated the potential role of tRNA processing as a mechanism by which HOE-1 may modulate UPRmt induction. Production of mature tRNAs begins with transcription of tRNA gene loci by RNA polymerase III followed by sequential cleavage of 5’-leader and 3’-trailer sequences from immature tRNA transcripts by RNase P and HOE-1, respectively. Following cleavage of 3’-trailer sequences, tRNAs can be transported to the cytosol by tRNA exportin (Hopper and Nostramo, 2019).

Given that HOE-1 nuclear levels are elevated in hoe-1(ΔNES) animals, we reasoned that 3’-tRNA processing should be elevated due to increased nuclear activity of HOE-1. Thus, we questioned if UPRmt induction in hoe-1(ΔNES) animals is a result of elevated 3’-tRNA processing. First, we knocked-down RNA pol III subunit rpc-1 to attenuate the production of total RNA pol III-dependent transcripts in hoe-1(ΔNES) animals. If hoe-1(ΔNES)-induced UPRmt is due to elevated processing of tRNAs we hypothesized that restriction of nascent tRNA production should attenuate UPRmt activation. Indeed, we found that rpc-1 RNAi robustly attenuates hoe-1(ΔNES)-induced UPRmt (Figure 6—figure supplement 2A and B). Interestingly, rpc-1 RNAi has little impact on mitochondrial stress-induced UPRmt (nuo-6(qm200)) (Figure 6—figure supplement 2C and D). These data show that rpc-1 is required for hoe-1(ΔNES)-induced UPRmt and support our hypothesis that increased 3’-tRNA processing by HOE-1 activates UPRmt.

For the majority of tRNAs 5’-end processing by the RNase P complex is a prerequisite for 3’-end processing by HOE-1 (Frendewey et al., 1985; Yoo and Wolin, 1997). Thus, if increased 3’-tRNA end processing is responsible for UPRmt activation, compromising 5’-end processing by RNAi against RNAse P should attenuate hoe-1(ΔNES)-induced UPRmt. RNAi against a subunit of the RNase P complex, popl-1, attenuates UPRmt induction in hoe-1(ΔNES) animals (Figure 6A and B). popl-1 RNAi also attenuates both UPRmt induced by nuo-6(qm200) (Figure 6C and D) as well as basal induction of ges-1p::GFPcyto (Figure 6E and F), albeit to a lesser extent than the attenuation seen in hoe-1(ΔNES) animals. These data suggest that popl-1 RNAi may have a broad impact on protein expression but supports that elevated 3’-tRNA processing in hoe-1(ΔNES) animals is responsible for UPRmt activation given that popl-1 RNAi strongly attenuates hoe-1(ΔNES)-induced UPRmt.

Figure 6 with 3 supplements see all
Nuclear export defective HOE-1 activates UPRmt via altered tRNA processing.

(A) Fluorescence images of UPRmt reporter (hsp-6p::GFP) activation in day 2 adult wildtype and hoe-1(ΔNES) animals on control and popl-1 RNAi. Scale bar 200 μm. (B) Fluorescence intensity quantification of hsp-6p::GFP in individual day 2 adult wildtype and hoe-1(ΔNES) animals on control and popl-1 RNAi (n = 24 for each condition, mean and SD shown, ordinary two-way ANOVA with Tukey’s multiple comparisons test). (C) Fluorescence images of UPRmt reporter (hsp-6p::GFP) activation in day 2 adult wildtype and nuo-6(qm200) animals on control and popl-1 RNAi. Scale bar 200 μm. (D) Fluorescence intensity quantification of hsp-6p::GFP in individual day 2 adult wildtype and nuo-6(qm200) animals on control and popl-1 RNAi (n = 24 for each condition, mean and SD shown, ordinary two-way ANOVA with Tukey’s multiple comparisons test). (E) Fluorescence images of intestinal-specific basal protein reporter (ges-1p::GFPcyto) activation in day 2 adult wildtype animals on control and popl-1 RNAi. Scale bar 200 μm. (F) Fluorescence intensity quantification of ges-1p::GFPcyto in individual day 2 adult wildtype animals on control and popl-1 RNAi (n = 24 for each condition, mean and SD shown, unpaired t-test). (G) Fluorescence images of UPRmt reporter (hsp-6p::GFP) activation in day 2 adult wildtype and hoe-1(ΔNES) animals on control and xpo-3 RNAi. Scale bar 200 μm. (H) Fluorescence intensity quantification of hsp-6p::GFP in individual day 2 adult wildtype and hoe-1(ΔNES) animals on control and xpo-3 RNAi (n = 24 for each condition, mean and SD shown, ordinary two-way ANOVA with Tukey’s multiple comparisons test). (I) Fluorescence images of UPRmt reporter (hsp-6p::GFP) activation in day 2 adult wildtype and nuo-6(qm200) animals on control and xpo-3 RNAi. Scale bar 200 μm. (J) Fluorescence intensity quantification of hsp-6p::GFP in individual day 2 adult wildtype and nuo-6(qm200) animals on control and xpo-3 RNAi (n = 24 for each condition, mean and SD shown, ordinary two-way ANOVA with Tukey’s multiple comparisons test). (K) Fluorescence images of intestinal-specific basal protein reporter (ges-1p::GFPcyto) activation in day 2 adult wildtype animals on control and xpo-3 RNAi. Scale bar 200 μm. (L) Fluorescence intensity quantification of ges-1p::GFPcyto in individual day 2 adult wildtype animals on control and xpo-3 RNAi (n = 24 for each condition, mean and SD shown, unpaired t-test).

Following 3’-end processing in the nuclei, tRNAs can be exported to the cytosol by tRNA exportin (Hopper and Nostramo, 2019). To test if elevated levels of 3’-processed tRNAs are required in the cytosol to activate UPRmt, we asked if restricting tRNA nuclear export via RNAi against tRNA exportin, xpo-3, attenuates hoe-1(ΔNES)-induced UPRmt. Strikingly, xpo-3 RNAi robustly attenuates hoe-1(ΔNES)-induced UPRmt(Figure 6G and H). However, xpo-3 RNAi does not attenuate nuo-6(qm200) induced UPRmt (Figure 6I and J) nor basal ges-1p::GFP levels (Figure 6K and L). These data suggest that in hoe-1(ΔNES) animals 3’-processed tRNAs are required in the cytosol to activate UPRmt.

While 5’- and 3’-tRNA processing are the only steps known to be required for tRNA export from the nucleus, there are other downstream tRNA maturation processes that occur (Hopper and Nostramo, 2019). Some nascent tRNAs include introns that need to be removed and then ligated by a tRNA ligase (Englert and Beier, 2005; Popow et al., 2012). For tRNAs to be charged with corresponding amino acids, nascent tRNAs must contain a CCA sequence as part of the 3’ acceptor stem. This can be achieved by a CCA-adding tRNA nucleotidyl transferase (Hou, 2010). Knockdown of both tRNA ligase, rtcb-1, and tRNA nucleotidyl transferase, hpo-31 mildly attenuate hoe-1(ΔNES)-induced UPRmt (Figure 6—figure supplement 3A–C). However, rtcb-1 RNAi also mildly attenuates nuo-6(qm200)-induced UPRmt (Figure 6—figure supplement 3D and E). Knockdown of hpo-31 severely compromised growth of nuo-6(qm200) animals and thus the impact on UPRmt could not accurately be assessed. These data suggest that tRNA ligation and CCA addition have limited involvement in hoe-1(ΔNES)-induced UPRmt.

Taken together, these data suggest that UPRmt induction by nuclear export deficient HOE-1 is the result of increased 3’-tRNA processing and that these tRNA species are required in the cytosol to trigger UPRmt.

Compromised HOE-1 nuclear export does not activate UPRmt via GCN2 or eIF2α

Alteration to tRNA processing can activate cellular signaling pathways (Raina and Ibba, 2014). One such pathway is the integrated stress response (ISR) in which uncharged tRNAs activate the kinase GCN2 which, in turn, phosphorylates the eukaryotic translation initiation factor, eIF2α. This inhibitory phosphorylation of eIF2α leads to upregulation of a select number of proteins including the transcription factor ATF4 (Pakos-Zebrucka et al., 2016; Costa-Mattioli and Walter, 2020). Interestingly, ATF4 and one of its targets, ATF5, are orthologs of ATFS-1 (Fiorese et al., 2016). Moreover, GCN2 and ISR in general have been shown to be responsive to mitochondrial stress (Baker et al., 2012; Fessler et al., 2020; Guo et al., 2020; Koncha et al., 2021). Thus, we questioned if UPRmt activation by hoe-1(ΔNES) is mediated via GCN2 and eIF2α phosphorylation. We found that hoe-1(ΔNES)-induced UPRmt is only slightly reduced in both a gcn-2 null (gcn-2(ok871)) and an eIF2α non-phosphorylatable mutant (eIF2α(S46A,S49A)) background (Figure 7A and B). These data suggest that a mechanism independent of ISR must largely be responsible for UPRmt activation by hoe-1(ΔNES) animals.

Nuclear export defective HOE-1 induced UPRmt is not gcn-2 or eIF2α dependent.

(A) Fluorescence images of UPRmt reporter (hsp-6p::GFP) activation in day 2 adult wildtype, gcn-2(ok871), eIF2α(S46A,S49A), hoe-1(ΔNES), hoe-1(ΔNES);gcn-2(ok871), and hoe-1(ΔNES);eIF2α(S46A,S49A) animals. Scale bar 200 μm. (B) Fluorescence intensity quantification of hsp-6p::GFP in individual day 2 adult wildtype, gcn-2(ok871), eIF2α(S46A,S49A), hoe-1(ΔNES), hoe-1(ΔNES);gcn-2(ok871), and hoe-1(ΔNES);eIF2α(S46A,S49A) animals (n = 24 for each condition, mean and SD shown, ordinary two-way ANOVA with Tukey’s multiple comparisons test).

Nuclear HOE-1 is dynamically responsive to mitochondrial stress and negatively regulated by ATFS-1

To better understand the potential physiological implications of HOE-1 in UPRmt, we assessed hoe-1 expression and subcellular dynamics of HOE-1 during mitochondrial stress. It is predicted that two major transcripts are produced from the hoe-1 gene locus: one that contains an MTS and one that does not, which are translated into mitochondrial- and nuclear-targeted HOE-1, respectively. However, it has been shown in other systems that hoe-1 orthologs produce a single transcript that encodes both a mitochondrial targeted and nuclear targeted HOE-1 isoform via alternative translation initiation (Rossmanith, 2011). Thus, we first sought to determine which mechanism is used for hoe-1 expression. To do so, we designed two sets of primers complementary to hoe-1 mRNA one of which amplifies only mRNA that includes the MTS and the other which amplifies all hoe-1 mRNA (spans a sequence that is included in all predicted HOE-1 isoforms). If there are two independent hoe-1 transcripts, we expected there to be higher levels of hoe-1 mRNA measured by the primer pair for total transcripts than for the mitochondrial specific pair. However, we found that both primer pairs measured similar levels of hoe-1 mRNA (Figure 8—figure supplement 1A) suggesting that, like in other systems, there is a single hoe-1 transcript. Next, we assessed hoe-1 mRNA levels in non-stress versus mitochondrial stress conditions. We found, using both primer pairs, that hoe-1 mRNA levels are modestly elevated in nuo-6(qm200) animals relative to wildtype (Figure 8—figure supplement 1B and C) suggesting that hoe-1 may be transcriptionally upregulated under conditions of mitochondrial stress.

Next, we assessed the subcellular dynamics of HOE-1 in response to mitochondrial stress. We found that HOE-1::GFP nuclear levels are markedly diminished under mitochondrial stress induced by nuo-6(qm200), cco-1 RNAi, and spg-7 RNAi (Figure 8A and B and Figure 8—figure supplement 2A and B). This observation was unexpected given that hoe-1 transcript levels are elevated during mitochondrial stress and it runs contrary to the fact that compromising HOE-1 nuclear export is sufficient to induce UPRmt (Figure 3A and B). A common feature of signaling pathways is negative regulation. Thus, we questioned if reduced nuclear HOE-1 is a result of negative feedback rather than a direct result of mitochondrial stress. Given that mitochondrial stress activates UPRmt, we assessed HOE-1::GFP status in a mitochondrial stress background wherein atfs-1 is knocked down by RNAi. HOE-1 levels are significantly upregulated in nuclei of nuo-6(qm200) animals on atfs-1 RNAi relative to nuo-6(qm200) animals on control RNAi, as well as both wildtype animals on control and atfs-1 RNAi (Figure 8A and B and Figure 8—figure supplement 3A–C). Moreover, total cellular HOE-1 levels are elevated under mitochondrial stress in an atfs-1 RNAi background (Figure 8C and D and Figure 8—source data 1A–E and Figure 8—source data 2A–E). Additionally, mitochondrial HOE-1 levels are elevated under mitochondrial stress conditions irrespective of RNAi treatment (Figure 8—figure supplement 3D). Together these data suggest that HOE-1 is upregulated and accumulates in nuclei upon mitochondrial stress. Then, nuclear HOE-1 is negatively regulated by ATFS-1 once UPRmt is activated.

Figure 8 with 4 supplements see all
Nuclear HOE-1 levels are elevated during mitochondrial stress in the absence of ATFS-1 but decreased in the presence of ATFS-1.

(A) Fluorescence images of HOE-1::GFP in day 1 adult wildtype and nuo-6(qm200) animals on control and atfs-1 RNAi. Scale bar 200 μm. (B) Fluorescence intensity quantification of intestinal nuclei relative to extranuclear signal in day 1 adult wildtype and nuo-6(qm200) animals on control and atfs-1 RNAi (n = 40 for each condition, mean and SD shown, ordinary two-way ANOVA with Tukey’s multiple comparisons test). (C) Western blot for HOE-1::GFP and actin from day 1 adult wildtype and nuo-6(qm200) animals on control and atfs-1 RNAi. (D) Quantification of HOE-1::GFP western blot band intensity from day 1 adult wildtype and nuo-6(qm200) animals on control and atfs-1 RNAi normalized to total protein (n = 4 for each condition, mean and SD shown, ordinary two-way ANOVA with Tukey’s multiple comparisons test). (E) Fluorescence images of HOE-1::GFP in day 1 adult wildtype and atfs-1(et15) animals. Scale bar 200 μm. (F) Fluorescence intensity quantification of intestinal nuclei relative to extranuclear signal in day 1 adult wildtype and atfs-1(et15) animals (n = 40 for each condition, mean and SD shown, unpaired t-test). (G) Western blot for HOE-1::GFP and actin from day 1 adult wildtype and atfs-1(et15) animals. (H) Quantification of HOE-1::GFP western blot band intensity from day 1 adult wildtype and atfs-1(et15) animals normalized to total protein (n = 4 for each condition, mean and SD shown, unpaired t-test). (I) Mitochondrial stress triggers activation of HOE-1 resulting in altered RNA processing that facilitates UPRmt via ATFS-1. Activation of UPRmt negatively regulates HOE-1.

Figure 8—source data 1

Blots for wildtype and nuo-6(qm200) animals on control and atfs-1 RNAi (Figure 8C).

All panels are the same membrane. (A) Image of stain-free blot for total protein from day 1 adult wildtype and nuo-6(qm200) animals on control and atfs-1 RNAi. Two biological replicates of each condition: Lane # 1&11 BR Spectra Protein Ladder – ladder bands in kDa denoted. Lane # 2&7 wildtype on control RNAi, 3&8 wildtype on atfs-1 RNAi, 4&9 nuo-6(qm200) on control RNAi, and 5&10 nuo-6(qm200) on atfs-1 RNAi. Lane # 6 empty. (B) Chemiluminescence image of blot for HOE-1::GFP using GFP primary antibody. (C) Composite image of chemiluminescence and colorimetric images of blot for HOE-1::GFP to show bands relative to ladder. (D) Chemiluminescence image of blot for actin using β-actin primary antibody. (E) Composite image of chemiluminescence and colorimetric images of blot for actin to show bands relative to ladder.

https://cdn.elifesciences.org/articles/71634/elife-71634-fig8-data1-v3.zip
Figure 8—source data 2

Blots for wildtype and nuo-6(qm200) animals on control and atfs-1 RNAi (Figure 8D).

Samples were loaded and ran on two separate membranes simultaneously (Membrane A and Membrane B). All panels in each column are the same membrane. (A) Image of stain-free blots for total protein from day 1 adult wildtype and nuo-6(qm200) animals on control and atfs-1 RNAi. Two biological replicates on each blot of each condition: Lane # 1&10 BR Spectra Protein Ladder – ladder bands in kDa denoted. Lane # 2&6 wildtype on control RNAi, 3&7 wildtype on atfs-1 RNAi, 4&8 nuo-6(qm200) on control RNAi, and 5&9 nuo-6(qm200) on atfs-1 RNAi. (B) Chemiluminescence image of blots for HOE-1::GFP using GFP primary antibody. (C) Composite images of chemiluminescence and colorimetric images of blots for HOE-1::GFP to show bands relative to ladder. (D) Chemiluminescence images of blots for actin using β-actin primary antibody. (E) Composite images of chemiluminescence and colorimetric images of blots for actin to show bands relative to ladder.

https://cdn.elifesciences.org/articles/71634/elife-71634-fig8-data2-v3.zip
Figure 8—source data 3

Blots for wildtype and atfs-1(et15) animals (Figure 8G).

All panels are the same membrane. (A) Image of stain-free blot for total protein from day 1 adult wildtype and atfs-1(et15) animals. Two biological replicates of each condition: Lane # 1&6 BR Spectra Protein Ladder – ladder bands in kDa denoted. Lane #2&4 wildtype and #3&5 atfs-1(et15). (B) Chemiluminescence image of blot for HOE-1::GFP using GFP primary antibody. (C) Composite image of chemiluminescence and colorimetric images of blot for HOE-1::GFP to show bands relative to ladder. (D) Chemiluminescence image of blot for actin using β-actin primary antibody. (E) Composite image of chemiluminescence and colorimetric images of blot for actin to show bands relative to ladder.

https://cdn.elifesciences.org/articles/71634/elife-71634-fig8-data3-v3.zip
Figure 8—source data 4

Blots for wildtype and atfs-1(et15) animals (Figure 8H).

All panels are the same membrane. (A), Image of stain-free blot for total protein from day 1 adult wildtype and atfs-1(et15) animals. Four biological replicates of each condition: Lane # 1 BR Spectra Protein Ladder – ladder bands in kDa denoted. Lanes #2,4,6,8 wildtype and #3,5,7,9 atfs-1(et15). (B) Chemiluminescence image of blot for HOE-1::GFP using GFP primary antibody. (C) Composite image of chemiluminescence and colorimetric images of blot for HOE-1::GFP to show bands relative to ladder. (D) Chemiluminescence image of blot for actin using β-actin primary antibody. (E) Composite image of chemiluminescence and colorimetric images of blot for actin to show bands relative to ladder.

https://cdn.elifesciences.org/articles/71634/elife-71634-fig8-data4-v3.zip

To further test if nuclear HOE-1 is negatively regulated by UPRmt activation rather than by mitochondrial stress, we assessed HOE-1 localization in ATFS-1 gain-of-function animals (atfs-1(et15)). atfs-1(et15) constitutively activates UPRmt in the absence of mitochondrial stress (Rauthan et al., 2013). Thus, we asked if atfs-1(et15) is sufficient to reduce nuclear HOE-1 levels. Indeed, nuclear HOE-1 levels are markedly reduced in atfs-1(et15) animals relative to wildtype (Figure 8E and F and Figure 8—figure supplement 4A–C) while total and mitochondrial HOE-1 protein levels are largely unperturbed (Figure 8G and H, Figure 8—figure supplement 4D, Figure 8—source data 3A–E and Figure 8—source data 4A–E). These data further support that UPRmt activation negatively regulates nuclear HOE-1.

Discussion

Regulation of UPRmt is not completely understood and elucidating this mechanism has broad implications for understanding cellular response to mitochondrial dysfunction. Here, we describe a novel mechanism by which mitochondrial stress is transduced to activate UPRmt and how that response is regulated through a feedback mechanism (Figure 8I).

Multiple factors have been identified that are required for maximal activation of UPRmt. This includes the mitochondrial localized proteins, CLPP-1 protease and peptide transmembrane transporter HAF-1 (Haynes et al., 2007; Haynes et al., 2010). Additionally, the transcription factors ATFS-1 and DVE-1 along with the co-transcriptional activator UBL-5 are required for UPRmt activation (Benedetti et al., 2006; Haynes et al., 2007; Haynes et al., 2010; Nargund et al., 2012; Nargund et al., 2015; Tian et al., 2016). Histone modifications, chromatin remodeling, and post-translational modifications of ATFS-1 are also involved in fully activating UPRmt (Tian et al., 2016; Merkwirth et al., 2016; Gao et al., 2019; Shao et al., 2020). We show for the first time that nuclear HOE-1 is required for maximal activation of UPRmt as its induction by various stressors is attenuated in hoe-1 RNAi, hoe-1 null, and hoe-1(ΔNLS) backgrounds.

We show that loss of hoe-1 results in varied attenuation of UPRmt depending on how UPRmt is activated. UPRmt induction by RNAi (cco-1 and spg-7) is robustly attenuated by loss of hoe-1 while nuo-6(qm200)-induced UPRmt is only modestly attenuated. RNAi by feeding works well in all tissues except neurons (Timmons et al., 2001; Kamath et al., 2003). Importantly, UPRmt can be activated non-cell autonomously in the intestine by mitochondrial stress in neurons (Durieux et al., 2011; Berendzen et al., 2016; Zhang et al., 2018). UPRmt induced cell-autonomously in the intestine by RNAi may be hoe-1 dependent while neuron-to-intestine UPRmt induction may work primarily in a hoe-1-independent manner. Consistent with this, increased nuclear accumulation of HOE-1 only activates UPRmt in the intestine. These results further exemplify the complexity of UPRmt signaling.

UPRmt is generally triggered via compromised mitochondrial membrane potential which facilitates the nuclear accumulation of ATFS-1 (Rolland et al., 2019; Shpilka et al., 2021). We find that UPRmt activation via hoe-1(ΔNES) correlates with a decrease in mitochondrial membrane potential providing a potential trigger for UPRmt induction. Furthermore, we show that the UPRmt transcription factors ATFS-1 and DVE-1 have increased nuclear localization in hoe-1(ΔNES) animals, thus likely facilitating the robust UPRmt activation.

HOE-1 functions in tRNA processing (Nashimoto et al., 1999; Mayer et al., 2000; Schiffer et al., 2002; Takaku et al., 2003; Dubrovsky et al., 2004; Brzezniak et al., 2011; Sanchez et al., 2011; Siira et al., 2018). Here, we show that increased 3’-tRNA processing by HOE-1 is likely responsible for UPRmt activation. Restricting HOE-1-dependent 3’-tRNA trailer sequence cleavage indirectly by RNAi against RNA polymerase III subunit, rpc-1, and RNase P subunit, popl-1, strongly attenuate hoe-1(ΔNES)-induced UPRmt. Moreover, these RNA species must be required in the cytosol to activate UPRmt as RNAi against tRNA exportin xpo-3 is sufficient to robustly attenuate hoe-1(ΔNES)-induced UPRmt. Our findings herein are the first reported connection between altered tRNA processing and UPRmt in C. elegans. Given the general requirement for tRNAs in protein translation on the one hand, and the mitochondria-specific nature of UPRmt on the other, our findings of a connection between the two are intriguing. However, besides performing their core housekeeping function in protein translation, tRNAs have also emerged as small RNAs with important regulatory roles inside cells (Avcilar-Kucukgoze and Kashina, 2020). Perhaps, the most well-characterized regulatory role for tRNAs is in the activation of the integrated stress response (ISR). In ISR, uncharged tRNAs activate the eIF2α kinase, GCN2, resulting in the upregulation of ATFS-1 orthologs ATF4 and ATF5 (Pakos-Zebrucka et al., 2016; Costa-Mattioli and Walter, 2020). However, we show that gcn-2 and eIF2α are not required for hoe-1(ΔNES)-induced UPRmt activation suggesting that a different mechanism is responsible. The lack of involvement of ISR in HOE-1’s role in UPRmt is not too surprising as there may be a greater pool of fully mature tRNAs in the cytosol in hoe-1(ΔNES) animals due to increased 3’-end processing of tRNAs above wildtype levels. This would result in an excess of charged tRNAs in the cytosol, the opposite of what is required to trigger GCN2-dependent ISR. Instead, we can speculate on several additional possibilities for the consequences of increased levels of charged tRNAs that can explain the role of HOE-1 in UPRmt regulation. For example, the use of amino acids to charge excess tRNAs in hoe-1(ΔNES) animals may limit the pool of free amino acids available for mitochondrial import, thus affecting translation of proteins encoded by the mitochondrial genome. This may result in stoichiometric imbalance between nuclear and mitochondrial-encoded components of the electron transport chain, which is known to compromise mitochondrial membrane potential and trigger UPRmt (Houtkooper et al., 2013). Alternatively, mito-nuclear imbalance in hoe-1(ΔNES) animals may result from excessive translation of nuclear-encoded mitochondrial proteins due to increased abundance of available charged tRNAs in the cytosol. In yet another scenario, UPRmt may not be the consequence of a global increase in the levels of all cytosolic tRNAs but rather, may be due to changes in the levels of specific tRNAs that preferentially impact translation of genes enriched for the corresponding codons. Such selective upregulation of tRNAs has been shown previously to have specific cellular consequences (Gingold et al., 2014; Goodarzi et al., 2016). Finally, it is possible that a tRNA-like RNA or other small RNA species such as tRNA fragments are responsible for UPRmt induction in hoe-1(ΔNES) animals (Kruszka et al., 2003; Lee et al., 2009; Bogerd et al., 2010; Siira et al., 2018). However, if this is the case, our data argue that such an RNA species would need to be transported to the cytosol by tRNA exportin. Non-tRNA transport by an ortholog of xpo-3 has not yet been reported (Hopper and Nostramo, 2019).

We show that nuclear HOE-1 is dynamically regulated by mitochondrial stress. In the presence of stress, nuclear HOE-1 levels are depleted. However, this is UPRmt dependent as HOE-1 nuclear levels under mitochondrial stress are elevated above wild-type levels when UPRmt is blocked by atfs-1 RNAi. These data, paired with the fact that compromising HOE-1 nuclear export triggers UPRmt, lead us to hypothesize that upon mitochondrial stress, nuclear HOE-1 levels are elevated. This upregulation of nuclear HOE-1 elevates 3’-tRNA processing thereby triggering a signaling cascade that results in elevated nuclear ATFS-1 and DVE-1 and subsequent UPRmt induction. Activated UPRmt then negatively regulates HOE-1 nuclear levels thus providing a feedback mechanism to tightly control mitochondrial stress response. UPRmt-negative regulation of HOE-1 is further supported by our data showing that constitutive activation of UPRmt by atfs-1(et15) is sufficient to reduce nuclear HOE-1 levels in the absence of mitochondrial stress. How it is that mitochondrial stress activates HOE-1 is still unknown. Multiple mitochondrial derived small molecules have been reported to communicate mitochondrial status including reactive oxygen species (ROS), NAD+, and acetyl-CoA (Baker et al., 2012; Mouchiroud et al., 2013; Ramachandran et al., 2019; Tjahjono et al., 2020; Zhu et al., 2020) We look forward to further investigating whether these, or other molecules, are involved in HOE-1 regulation.

In humans, mutations in the ortholog of HOE-1, ELAC2, are associated with both hypertrophic cardiomyopathy (Haack et al., 2013; Shinwari et al., 2017; Saoura et al., 2019) and prostate cancer (Tavtigian et al., 2001; Korver et al., 2003; Noda et al., 2006). Historically, it has been suggested that mutations in ELAC2 cause disease because of a loss of mature tRNA production. Our works suggests an intriguing alternative whereby ELAC2 mutations lead to altered tRNA processing that triggers aberrant stress response signaling resulting in disease state. Our system provides a convenient opportunity to interrogate these disease causing variants.

Taken together, our findings provide a novel mechanism—involving the tRNA processing enzyme HOE-1—by which mitochondrial stress is transduced to activate UPRmt thus providing important insight into the regulation of mitochondrial stress response.

Methods

Worm maintenance

Worms were grown on nematode growth media (NGM) seeded with OP50 E. coli bacteria and maintained at 20 °C.

Mutants and transgenic lines

A complete list of C. elegans strains used can be found in Supplementary file 1. All new mutant and transgenic strains generated via CRISPR/Cas9 for this study were confirmed by Sanger sequencing.

CRISPR/Cas9

CRISPR was conducted as previously described (Dokshin et al. Genetics 2018; Paix et al. Genetics 2015) using Alt-R S.p. Cas9 Nuclease V3 (IDT #1081058) and tracrRNA (IDT #1072532). A complete list of crRNA and repair template sequences purchased from IDT can be found in Supplementary file 2.

Genetic crosses

Strains resulting from genetic crosses were generated by crossing ~20 heterozygous males of a given strain to 5–8 L4 hermaphrodites of another strain (heterozygous males were generated by first crossing L4 hermaphrodites of that strain to N2 males). F1, L4 hermaphrodites were then cloned out and allowed to have self-progeny. F2 progeny were cloned out and once they had progeny were genotyped or screened (if fluorescent marker) for presence of alleles of interest. All genotyping primers were purchased from IDT and can be found in Supplementary file 2.

Fluorescence microscopy

All whole animal imaging was done using Zeiss Axio Zoom V16 stereo zoom microscope. For all whole animal imaging, worms were immobilized on 2% agar pads on microscope slides in ~1 μl of 100 mM levamisole (ThermoFisher #AC187870100) and then coverslip applied.

Fluorescence image analysis

For whole animal fluorescence intensity quantification, total pixels (determined by tracing individual animals and summing the total number of pixels within the bounds of the trace) and pixel fluorescence intensity (pixel fluorescence intensity on 1–255 scale) were quantified using imageJ and mean fluorescence intensity for each worm was calculated (sum total of fluorescence intensity divided by total number of pixels within bounds of the trace). For DVE-1::GFP image analysis (Figure 5E&F), brightness threshold was set to 25 in imageJ and then the number of gut cell nuclei that were saturated at this threshold were counted. For Figure 8A&B and Figure 8E&F, and Figure 8—figure supplement 2A&B, mean fluorescence intensity was calculated within the bounds of gut cell nuclei and outside of the bounds of gut cell nuclei and then graphed as the ratio fluorescence intensity of nuclear to extranuclear signal.

RNAi

RNAi by feeding was conducted as previously described (Gitschlag et al. Cell Met. 2016). Briefly, RNAi clones were grown overnight from single colony in 2 ml liquid culture of LB supplemented with 50 μg/ml ampicillin. To make 16 RNAi plates, 50 ml of LB supplemented with 50 μg/ml ampicillin was inoculated with 500 μl of overnight culture and then incubated while shaking at 37 °C for 4–5 hours (to an OD550-600 of about 0.8). Cultures were then induced by adding 50 ml additional LB supplemented with 50 μg/ml ampicillin and 4 mM IPTG and then continued incubating while shaking at 37 °C for 4 hours. Following incubation, bacteria were pelleted by centrifugation at 3900 rpm for 6 min. Supernatant was decanted and pellets were gently resuspended in 4 ml of LB supplemented with 8 mM IPTG. 250 μl of resuspension was seeded onto standard NGM plates containing 1 mM IPTG. Plates were left to dry overnight and then used within 1 week. Bacterial RNAi feeder strains were all from Ahringer RNAi Feeding Library, grown from single colony and identity confirmed by Sanger sequencing. atfs-1 (ZC376.7), cco-1 (F26E4.9), hoe-1 (E04A4.4), hpo-31 (F55B12.4), popl-1 (C05D11.9), rpc-1 (C42D4.8), rtcb-1 (F16A11.2), spg-7 (Y47G6A.10), xpo-3 (C49H3.10).

Quantification of gene expression

cDNA was synthesized using Maxima H Minus First Strand cDNA Synthesis Kit, with dsDNase (ThermoFisher #K1682) according to manufacturer’s directions. Lysates for cDNA synthesis were made by transferring 10, day 2 adult worms to 10 μl of lysis buffer supplemented with 20 mg/ml proteinase K and incubating at 65 °C for 10 min, 85 °C for 1 min and 4 °C for 2 min. Quantification of gene expression was performed using droplet digital PCR (ddPCR) with Bio-Rad QX200 ddPCR EvaGreen Supermix (Bio-Rad #1864034). Primers used for ddPCR can be found in Supplementary file 2.

TMRE staining

A total of 500 μl of 1 mM TMRE (ThermoFisher #T669) solution in M9 buffer (prepared from a stock TMRE solution of 0.5 M in DMSO) was supplemented on top of standard NGM plates pre-seeded with 200 ul lawn of OP50 and allowed to dry overnight in the dark. The following day young L4 animals were transferred to TMRE plates and incubated on TMRE for 16 hr. After 16 hr, animals were transferred from TMRE plates to seeded standard NGM plates for 1 hr to remove any non-specific TMRE signal from cuticle and intestinal lumen. Animals were then imaged via confocal microscopy as described below.

Confocal fluorescence imaging

Worms were grown at 20 °C and age-synchronized by timed egg-lays on NGM plates seeded with OP50 or HT115 bacteria for RNAi experiments. Before imaging, worms were immobilized with 3 μl 0.05 µm Polybead microsphere suspension (Polysciences) on a 10% agarose pad with a coverslip (1). Images were taken in the mid- or posterior intestine using a Nikon Ti2 with CSU-W1 spinning disk and Plan-Apochromat 100 X/1.49 NA objective. HOE-1::GFP was imaged by 488 nm laser excitation and ET525/36 m emission filter. 2 X integration was applied (Nikon Elements) to increase signal strength. TMRE and ATFS-1::mCherry were imaged with 561 nm laser excitation and ET605/52 M emission filter.

Image processing and analysis was performed with Nikon Elements software. Raw images were subjected to deconvolution and rolling ball background subtraction. Mitochondrial networks were segmented using the TMRE signal after excluding dye aggregates via Bright Spot Detection. To objectively set threshold parameters across groups with different TMRE intensity levels, the low threshold for segmentation was calculated based on a linear correlation with mean TMRE intensity within each group, y = 0.6411*x + 89.71 (x = mean TMRE intensity and constants derived from an initial manual validation). Regions of interest (ROIs) were manually drawn to encompass a single intestinal cell, and nuclei were identified and segmented manually using brightfield images. Mean intensities were measured within the resulting masks.

To detect localization of HOE-1::GFP in mitochondria, images of TMRE-stained intestinal cells of control and ΔMTS worms were collected and blinded. Mitochondria were segmented by TMRE signals as above. For each cell, one representative line scan was drawn manually across the mitochondrial short axis.

Western blot

Fifty adult worms were transferred into a tube containing 20 μl of M9 Buffer. Then, 20 μl of 2 x Laemmli Buffer (BioRad #161–0737) supplemented with 2-mercaptoethanol (i.e. βME) was added to worm suspension and gently pipetted up and down 5 times to mix. Worms were lysed at 95 °C for 10 min in thermocycler followed by ramp down to room temperature (25 °C). Lysates were then pipetted up and down 10 times to complete disrupt and homogenize suspension. Samples were briefly centrifuged to pellet any worm debris. 20 μl of lysate supernatant was loaded onto precast Mini-PROTEAN TGX Stain-Free Gel (BioRad #4568045). Gel was run for 30 min at 100 V and then an additional 40–45 min at 130 V in 1 x Tris/Glycine/SDS Running Buffer (BioRad #1610732). Following electrophoresis gel was activated and imaged for total protein. Gel was equilibrated in Trans-Blot Turbo Transfer Buffer (BioRad #10026938) and transferred to activated and equilibrated Trans-Blot Turbo LF PVDF Membrane (BioRad #10026934) for 7 min at 2.5 A/25 V on Trans-Blot Turbo Transfer System. Following transfer, stain-free membrane was imaged for total protein. Membrane was then blocked in 5% milk in TBST for 2 hr rocking at room temperature. Following blocking, membrane was incubated in primary antibody overnight rocking at 4 °C. Mouse monoclonal anti-β-actin (Santa Cruz Biotechnology #sc-47778) or mouse monoclonal anti-GFP (#sc-9996) were used at a dilution of 1:2,500 in 5% milk in TBST. The following day the membrane was washed three times for 5 min each with TBST and then incubated with HRP-conjugated goat anti-mouse antibody (sc-2005) at 1:2000 in 5% milk in TBST for 2 hours at room temperature. Membrane was again washed three times for 5 min each with TBST. Membranes were then incubated for 5 min in Clarity Western ECL Substrate (BioRad #1705060) and immediately imaged on a BioRad ChemiDoc MP imager. Band intensity was quantified using imageJ.

Statistical analysis

Experiment-specific details regarding sample size and statistical test used can be found in the corresponding Figure Legends. Significant p-values under 0.05 are denoted on all graphs and p-values above 0.05 are considered non-significant (ns). All statistical analysis was performed in GraphPad Prism 9. All data points for each experiment are included (no outlier exclusion was performed). For all whole animal fluorescence analysis, a sample size of 24 animals was generally used, each animal considered a biological replicate. Statistical analysis of high resolution fluorescence confocal imaging (HOE-1::GFP, ATFS-1::mCherry, and TMRE) was conducted on sample sizes between 60 and 80 animals of which animals were collected and imaged on three independent days, each animal considered a biological replicate. For western blot analysis, four independent samples were used for each condition, each sample (containing 50 worms each) is considered a biological replicate. For ddPCR analysis, a sample size of 4 was used for each condition, each sample (containing 10 worms each) is considered a biological replicate, each biological replicate was run in technical duplicate of which the average value was used for analysis.

Data availability

All data generated or analyzed during this study are included in the manuscript and supporting files. Source data files have been provided for Figures 5 and 8.

References

Decision letter

  1. Xiaochen Wang
    Reviewing Editor; Institute of Biophysics Chinese Academy of Sciences, China
  2. David Ron
    Senior Editor; University of Cambridge, United Kingdom

Our editorial process produces two outputs: i) public reviews designed to be posted alongside the preprint for the benefit of readers; ii) feedback on the manuscript for the authors, including requests for revisions, shown below. We also include an acceptance summary that explains what the editors found interesting or important about the work.

Decision letter after peer review:

Thank you for submitting your article "A tRNA processing enzyme is a central regulator of the mitochondrial unfolded protein response" for consideration by eLife. Your article has been reviewed by 3 peer reviewers, one of whom is a member of our Board of Reviewing Editors, and the evaluation has been overseen by David Ron as the Senior Editor. The reviewers have opted to remain anonymous.

The reviewers have discussed their reviews with one another, and the Reviewing Editor has drafted this to help you prepare a revised submission.

Essential revisions:

The following points (especially the major experiments) should be addressed to strengthen the conclusion that HOE-1 plays a specific role in the activation of mtUPR.

Major experiments:

1. The authors use the transcriptional reporter hsp-6::gfp as a mtUPR reporter.

However, a fluorescent signal requires not only transcription from the hsp-6 promoter (the parameter of interest) but also translation of the derived mRNA. As HOE-1 is a tRNA processing enzyme whose inactivation may affect protein synthesis, qRT-PCR analysis (or some alternative analytical strategy) should be performed to quantify the effects of HOE-1 inhibition on the mtUPR transcription response, independent of the translation of a reporter.

Several groups have shown that inhibition of S6 kinase inhibits mtUPR activation. As HOE-1 is presumably required for protein synthesis, perhaps the mechanism is related? Does inhibition of other genes affecting tRNA levels also impair mtUPR or is it specific to HOE-1?

2. It seems that the HOE-1 protein with a mitochondrial targeting sequence is

transcribed from the same gene as HOE-1 without the MTS. And there are separate transcriptional start sites for each mRNA/protein. Considering the number of claims related to subcellular localization of HOE-1, the authors must determine if transcription from either site is altered during mitochondrial stress. During mitochondrial stress, does the ratio of HOE-1 transcript change? For example, is the hoe-1 variant transcript lacking the MTS increased?

3. There is an important caveat regarding the interpretation of the hoe-1(∆NES) strain which causes mtUPR activation: It remains unclear if nuclear accumulation is an event driving mtUPR activation or if the activation reflects a different feature of the ∆NES mutation.

The authors suggest that mtUPR induction in hoe-1(∆NES) is a result of increased 3'-tRNA processing. Whether 3'-tRNA processing is elevated in hoe-1(∆NES) should be tested more directly. Is it possible to determine the tRNA species that are elevated in hoe-1(∆NES) strain by sequencing? Or that the authors can express hoe-1(∆NES) that lacks the enzymatic activity and see whether it can still activate mtUPR.

4. There is a concern In regards to the finding that hoe-1(ΔNES) mutant is sufficient to induce the nuclear accumulation of the ATFS-1 and the subsequent up-regulation of the mtUPR reporter gene: the authors did not rule out the possibility that mitochondrial protein homeostasis was already disrupted in hoe-1(ΔNES) mutants so that the mtUPR was induced. Does HOE-1∆NES cause mitochondrial dysfunction which increases mtUPR activation? The authors only showed that mitochondrial membrane potential was not changed in hoe-1(ΔNES) mutants. More characterization of mitochondrial function in hoe-1(ΔNES) mutants is required, such as OCR and mitochondrial morphology. It seems that hoe-1(ΔNES) mutants are smaller than wild-type animals. Alternatively, the ∆NES mutation could be combined with the ∆MTS mutation.

5. The authors generate a beautiful ATFS-1::mCherry fusion protein and

demonstrate that it accumulates within nuclei during mitochondrial stress. Why is the overall level of ATFS-1 dramatically increased in hoe-1(ΔNES) mutants (Figure 4a)? This is not consistent with only two-fold up-regulation of atfs-1 transcript levels. Does hoe-1 inhibition affect translation/synthesis of ATFS-1::mCherry or nuclear accumulation of ATFS-1::mCherry? Or, DVE-1?

The authors also need to show the ATFS-1::GFP expression pattern in the nuo-6 mutants as a control.

6. Regarding the specific involvement of HOE-1 in the regulation of mtUPR, since

tRNA processing, the tRNA exporter xpo-3, as well as the RNase P complex popl-1, are all general regulators for protein synthesis. How to explain the specific involvement of these regulators only in the regulation of the mtUPR? The authors mentioned that HOE-1 homolog ELAC2 is not only required for tRNA maturation but also essential for the formation of tRNA fragments, snoRNAs, and miRNAs, are these non-coding RNAs account for the activation of the mtUPR?

It is also confusing that HOE-1(∆NLS) mutants suppressed the mtUPR induction in nuo-6 mutants, however, xpo-3 which functions in the same pathway as HOE-1 in terms of tRNA processing and export did not suppress the mtUPR induction in nuo-6 mutants in Figure 6i and 6j.

Reviewer #2 (Recommendations for the authors):

It was reported that the epigenetic regulation of the UPRmt is in parallel with the ATFS-1 pathway (PMID: 27133168, PMID: 27133166). Whether these epigenetic factors are required for the induction of UPRmt in hoe-1(ΔNES) mutants. Similarly, whether HOE-1(NLS) suppressed the epigenetic changes or the accumulation of epigenetic factors (PMID: 27133166, PMID: 32789178, PMID: 32934238 ) in response to mitochondrial stresses.

https://doi.org/10.7554/eLife.71634.sa1

Author response

Essential revisions:

The following points (especially the major experiments) should be addressed to strengthen the conclusion that HOE-1 plays a specific role in the activation of mtUPR.

Major experiments:

1. The authors use the transcriptional reporter hsp-6::gfp as a mtUPR reporter.

However, a fluorescent signal requires not only transcription from the hsp-6 promoter (the parameter of interest) but also translation of the derived mRNA. As HOE-1 is a tRNA processing enzyme whose inactivation may affect protein synthesis, qRT-PCR analysis (or some alternative analytical strategy) should be performed to quantify the effects of HOE-1 inhibition on the mtUPR transcription response, independent of the translation of a reporter.

The reviewers raise an important point. To determine the effects of hoe-1 inhibition on the UPRmt transcriptional response independent of translation of the UPRmt reporter (hsp-6p::GFP) we performed droplet digital PCR to quantify transcripts of genes upregulated upon UPRmt activation (i.e. hsp-6 and cyp-14A4.1) in a wildtype and hoe-1(ΔNLS) background in the absence and presence of mitochondrial stress (control and spg-7 RNAi, respectively). We find that loss of nuclear HOE-1 results in attenuation of both hsp-6 and cyp-14A4.1 transcript levels in mitochondrial stress conditions (Figure 2I and Figure 2 —figure supplement 5A). This finding is consistent with the effect of loss of nuclear HOE-1 on UPRmt reporter induction and further suggests that nuclear HOE-1 is directly involved in UPRmt transcriptional response.

Several groups have shown that inhibition of S6 kinase inhibits mtUPR activation. As HOE-1 is presumably required for protein synthesis, perhaps the mechanism is related? Does inhibition of other genes affecting tRNA levels also impair mtUPR or is it specific to HOE-1?

The reviewers query whether inhibition of other genes affecting tRNA levels also impair UPRmt. To address this question we assessed UPRmt reporter activation in mitochondrial stressed animals when other tRNA processing genes are knocked-down. These include RNA polymerase III subunit, rpc-1 (Figure 6 —figure supplement 2C, 2D), RNAse P subunit, popl-1 (Figure 6C, 6D), and tRNA ligase, rtcb-1 (Figure 6 —figure supplement 3D, 3E). RNA polymerase III transcribes tRNAs, RNAse P processes 5’ ends of nascent tRNAs before they are processed at the 3’ end by HOE-1, and the tRNA ligase is involved in splicing of intron-containing tRNAs. Knockdown of rpc-1 did not significantly impact nuo-6(qm200) induced UPRmt. Knock-down of popl-1 and rtcb-1 partially attenuate UPRmt activation by nuo-6(qm200). These data suggest that the inhibition of UPRmt is not specific to hoe-1 loss-of-function and further strengthen the connection between tRNA biology and UPRmt.

Although not asked for directly, prompted by the reviewer suggestion, we also tested whether rpc1 and rtcb-1 knockdown impairs hoe-1(ΔNES)-induced UPRmt (we had already reported in the original manuscript that popl-1 RNAi suppresses hoe-1(ΔNES)-induced UPRmt). Like popl-1 RNAi, rpc-1 RNAi robustly attenuates hoe-1(ΔNES)-induced UPRmt (Figure 6 —figure supplement 2A, 2B) further suggesting that limiting the availability of tRNA substrates for HOE-1 to act on suppresses hoe-1(ΔNES)-induced UPRmt Knock-down of rtcb-1 also mildly attenuates hoe1(ΔNES)-induced UPRmt (Figure 6 —figure supplement 3A, 3B) suggesting that downstream rates of tRNA processing may also impact hoe-1(ΔNES)-induced UPRmt. These data provide further support for the role of tRNAs in inducing UPRmt.

The reviewers raise an interesting possibility that mTOR and HOE-1 mechanisms of UPRmt induction may be related or intertwined. Given the broad involvement of mTOR signaling in cellular processes we would need to fully investigate any potential connection between these pathways (i.e., direct interaction) in future work.

2. It seems that the HOE-1 protein with a mitochondrial targeting sequence is

transcribed from the same gene as HOE-1 without the MTS. And there are separate transcriptional start sites for each mRNA/protein. Considering the number of claims related to subcellular localization of HOE-1, the authors must determine if transcription from either site is altered during mitochondrial stress. During mitochondrial stress, does the ratio of HOE-1 transcript change? For example, is the hoe-1 variant transcript lacking the MTS increased?

We appreciate the reviewers’ suggestion to assess transcript dynamics of hoe-1. HOE-1 protein with and without a mitochondrial targeting sequence are indeed transcribed from the same gene locus. However, whether the two protein isoforms are independently transcribed is not clear. In fact, in human cell culture it has been shown that both mitochondrial and nuclear-targeted HOE1 are produced from the same transcript via alternative translation initiation (Rossmanith, PMID: 21559454). Thus, we first endeavored to determine the mode by which mitochondrial and nuclear HOE-1 are individually produced. We designed two sets of primers for measuring hoe-1 transcript levels. One set that amplifies only transcripts containing the sequence encoding the mitochondrial targeting sequence and one set that amplifies all HOE-1 transcripts (i.e., sequence that is found in both mitochondrial and nuclear isoforms). If the two isoforms are a result of independent transcription, we would expect the number of mitochondrial specific transcripts to be lower than total transcript levels. However, using droplet digital PCR, we find that the number of transcripts that include a mitochondrial targeting sequence were nearly identical to the number of total hoe1 transcripts (Figure 8 —figure supplement 1A). This finding suggests, that like in higher eukaryotes, HOE-1 is dual-targeted via differential translation of a single transcript.

Given the above finding, we next endeavored to determine if hoe-1 transcript levels are altered upon mitochondrial stress. We find that hoe-1 transcript levels are mildly elevated under conditions of mitochondrial stress (i.e., nuo-6(qm200) worms) relative to wildtype when measured by ddPCR using both sets of aforementioned primers (Figure 8 —figure supplement 1B, 1C). These findings are consistent with our HOE-1 protein level analysis and support our finding that nuclear HOE-1 levels are elevated upon mitochondrial stress.

3. There is an important caveat regarding the interpretation of the hoe-1(∆NES) strain which causes mtUPR activation: It remains unclear if nuclear accumulation is an event driving mtUPR activation or if the activation reflects a different feature of the ∆NES mutation.

The authors suggest that mtUPR induction in hoe-1(∆NES) is a result of increased 3'-tRNA processing. Whether 3'-tRNA processing is elevated in hoe-1(∆NES) should be tested more directly. Is it possible to determine the tRNA species that are elevated in hoe-1(∆NES) strain by sequencing? Or that the authors can express hoe-1(∆NES) that lacks the enzymatic activity and see whether it can still activate mtUPR.

The reviewers raise an important point regarding the functional nature of the hoe-1(ΔNES) mutant that we generated and used in the manuscript. To validate the function of the hoe-1(ΔNES) allele we conducted three complimentary experiments. First, as suggested, we created a catalyticallydead hoe-1(ΔNES) allele by introducing a point mutation (D624A) in hoe-1 that ablates zinc binding. The endonuclease activity of HOE-1 is dependent upon zinc binding as it is a zinc phosphodiesterase. Homozygous hoe-1(D624A+ΔNES) animals have the same arrest phenotype as hoe-1 null animals. Given that UPRmt is not activated in hoe-1(ΔNES) animals until late in development we needed to be able to assess the impact of the D624A mutation later in development. To overcome this constraint we established hoe-1(ΔNES)/hoe-1(D624A+ΔNES) trans-heterozygous animals that expressed the UPRmt reporter hsp-6p::GFP. These animals were able to grow to adulthood and thus we could assess impact on UPRmt activation. hoe1(ΔNES)/hoe-1(D624A+ΔNES) trans-heterozygous animals had markedly diminished UPRmt activation relative to homozygous hoe-1(ΔNES) animals (Figure 6 —figure supplement 1A, 1B) suggesting that the ability of hoe-1(ΔNES) to activate UPRmt requires the RNA processing function of HOE-1.

Secondly, for hoe-1(ΔNES) to facilitate increased 3’-tRNA processing this would likely require there to be elevated nuclear HOE-1 levels in hoe-1(ΔNES) animals. To assess this we generated a C-terminally GFP-tagged hoe-1(ΔNES) allele hoe-1(ΔNES::GFP) and compared it’s subcellular expression to wildtype hoe-1::GFP. Based on high resolution imaging and its quantification, there is elevated HOE-1::GFP signal in nuclei of the hoe-1(ΔNES) background relative to wildtype (Figure 3 —figure supplement 1B, Figure 2 —figure supplement 4B, 4C). This finding is consistent with our hypothesis that there is increased 3’-tRNA processing in hoe-1(ΔNES) animals.

Third, and finally, if elevated nuclear HOE-1 levels are responsible for UPRmt activation we reasoned that ablating HOE-1 nuclear localization in hoe-1(ΔNES) animals (hoe-1(ΔNLS+ΔNES)) should inactivate hoe-1(ΔNES)-induced UPRmt. Indeed we found that compromising HOE-1 nuclear localization was sufficient to completely attenuate UPRmt induced by hoe-1(ΔNES) (Figure 3 —figure supplement 3A, 3B). This finding strongly suggests that HOE-1 is required in the nucleus to activate UPRmt.

Combined, these experiments suggest that UPRmt in hoe-1(ΔNES) animals is induced by increased 3’-tRNA processing that is a result of elevated nuclear levels of HOE-1.

4. There is a concern In regards to the finding that hoe-1(ΔNES) mutant is sufficient to induce the nuclear accumulation of the ATFS-1 and the subsequent up-regulation of the mtUPR reporter gene: the authors did not rule out the possibility that mitochondrial protein homeostasis was already disrupted in hoe-1(ΔNES) mutants so that the mtUPR was induced. Does HOE-1∆NES cause mitochondrial dysfunction which increases mtUPR activation? The authors only showed that mitochondrial membrane potential was not changed in hoe-1(ΔNES) mutants. More characterization of mitochondrial function in hoe-1(ΔNES) mutants is required, such as OCR and mitochondrial morphology. It seems that hoe-1(ΔNES) mutants are smaller than wild-type animals. Alternatively, the ∆NES mutation could be combined with the ∆MTS mutation.

We thank the reviewers for making this important suggestion to more thoroughly investigate the relationship between UPRmt and mitochondrial function in hoe-1(ΔNES) animals. The experiments we conducted in response to these suggestions proved to be very informative. Compromised mitochondrial membrane potential has been shown to be the driving factor for UPRmt activation as decreased membrane potential impairs mitochondrial import of proteins with weakly charged mitochondrial targeting sequences including ATFS-1 (Rolland et al., PMID: 31412237, Shpilka et al., PMID: 33473112). In the original draft of the manuscript, we had measured membrane potential in L4 stage animals and had not observed any differences between wildtype and hoe1(ΔNES) animals. However, as the UPRmt is most robustly induced in 2-day old adult hoe1(ΔNES) animals, we reassessed membrane potential at this later stage. Furthermore, in collaboration with the Burkewitz Lab, this measurement was done using high resolution microscopy as opposed to whole animal imaging. We conducted TMRE staining on adult hoe1(ΔNES) and wildtype animals and found that mitochondrial membrane potential is significantly reduced in hoe-1(ΔNES) relative to wildtype (Figure 4A, 4B). Thus, these data are consistent with the reviewers’ surmise that there may be mitochondrial dysfunction in hoe-1(ΔNES) animals. Interestingly, hoe-1(ΔNLS) animals also show a similarly drastic decline in mitochondrial membrane potential (Figure 4A, 4B), despite the fact that UPRmt is attenuated in this background. Thus, while there is a correlation between decreased membrane potential and UPRmt induction in hoe-1(ΔNES) animals, it is difficult to infer causality between the two.

UPRmt induction has been reported to cause a decrease in mitochondrial membrane potential. Therefore, we wondered whether UPRmt causes decline in mitochondrial membrane potential in hoe-1(ΔNES) animals. To test for this possibility, we measured mitochondrial membrane potential using TMRE in hoe-1(ΔNES) on atfs-1 RNAi. Loss of atfs-1 did not rescue membrane potential in hoe-1(ΔNES) background. Based on these data, we conclude in the manuscript that hoe-1(ΔNES) directly causes a decrease in mitochondrial membrane potential independent of UPRmt.

In addition, we took the reviewers’ suggestion of creating a hoe-1(ΔMTS+ΔNES) mutant to address whether hoe-1(ΔNES) may be having a compromising effect directly in the mitochondria. If hoe-1(ΔNES) is causing UPRmt by acting in the mitochondria, then impairing its mitochondrial localization should attenuate hoe-1(ΔNES)-induced UPRmt. If instead, as we hypothesized, hoe1(ΔNES) activates UPRmt through its nuclear role, then compromising mitochondrial localization of HOE-1 should not attenuate hoe-1(ΔNES)-induced UPRmt. We found that hoe-1(ΔMTS+ΔNES) animals have higher UPRmt activation than hoe-1(ΔNES) alone (Figure 3 —figure supplement 4A, 4B). This is consistent with HOE-1 activating UPRmt via increased nuclear accumulation and rules out the possibility that mitochondrial localized HOE-1 induces UPRmt in hoe-1(ΔNES) animals.

5. The authors generate a beautiful ATFS-1::mCherry fusion protein and

demonstrate that it accumulates within nuclei during mitochondrial stress. Why is the overall level of ATFS-1 dramatically increased in hoe-1(ΔNES) mutants (Figure 4a)? This is not consistent with only two-fold up-regulation of atfs-1 transcript levels. Does hoe-1 inhibition affect translation/synthesis of ATFS-1::mCherry or nuclear accumulation of ATFS-1::mCherry? Or, DVE-1?

The authors also need to show the ATFS-1::GFP expression pattern in the nuo-6 mutants as a control.

To more thoroughly investigate ATFS-1 levels across backgrounds, with the help from the Burkewitz Lab, we conducted the ATFS-1::mCherry imaging experiments at high resolution using confocal microscopy as opposed to our original imaging which was done on a Nikon Ti-E Fluorescence Motorized DIC Polarization Microscope. In addition to wildtype and hoe-1(ΔNES) animals, we also imaged ATFS-1::mCherry in nuo-6(qm200) animals as a positive control, as suggested by the reviewers. High resolution microscopy of ATFS-1::mCherry confirmed our previous findings that nuclear ATFS-1 levels are elevated in hoe-1(ΔNES) (Figure 5A, 5B). Importantly, nuclear ATFS-1 levels were also elevated under mitochondrial stress (i.e. nuo6(qm200) animals) as expected (Nargund et al., PMID: 22700657). We also quantified total cellular and extranuclear ATFS-1::mCherry fluorescence levels to address the reviewers’ question regarding the impact of hoe-1(ΔNES) on ATFS-1 translation/synthesis. hoe-1(ΔNES) animals do not exhibit elevated total or extranuclear ATFS-1::mCherry levels (Figure 5C and Figure 5 —figure supplement 1A). These data suggest that hoe-1(ΔNES) results in elevated nuclear localization but not increased ATFS-1 protein levels.

Similarly, to address the impact of hoe-1(ΔNES) on DVE-1 translation/synthesis level we conducted a western blot for DVE-1::GFP in a wildtype vs hoe-1(ΔNES) background. DVE-1 levels are not significantly different between wildtype and hoe-1(ΔNES) (Figure 5G, 5H, Figure 5 – source data 1) suggesting that hoe-1(ΔNES) triggers nuclear accumulation of DVE-1 as opposed to upregulating total DVE-1 protein levels.

6. Regarding the specific involvement of HOE-1 in the regulation of mtUPR, since

tRNA processing, the tRNA exporter xpo-3, as well as the RNase P complex popl-1, are all general regulators for protein synthesis. How to explain the specific involvement of these regulators only in the regulation of the mtUPR? The authors mentioned that HOE-1 homolog ELAC2 is not only required for tRNA maturation but also essential for the formation of tRNA fragments, snoRNAs, and miRNAs, are these non-coding RNAs account for the activation of the mtUPR? It is also confusing that HOE-1(∆NLS) mutants suppressed the mtUPR induction in nuo-6 mutants, however, xpo-3 which functions in the same pathway as HOE-1 in terms of tRNA processing and export did not suppress the mtUPR induction in nuo-6 mutants in Figure 6i and 6j.

We appreciate these reviewer comments. While indeed xpo-3 and popl-1 should be required for protein synthesis it is clear from our results that modulating their activity can specifically impact UPRmt. While surprising, these data support the idea that in addition to their role in protein synthesis more generally, tRNAs (or other putative HOE-1-processed RNAs) play a specific signaling role in modulating UPRmt. This idea explains how RNAi against essential tRNA processing machinery, while strong enough to compromise UPRmt activation, is not strong enough to significantly impact protein synthesis. Indeed this reasoning is supported by the fact that animals on xpo-3 and popl-1 RNAi grow to adulthood.

Orthologs of HOE-1 have been reported to be capable of processing other RNA species. If those species are involved in UPRmt regulation they would need to be transported by tRNA exportin (xpo-3)—while such xpo-3 dependent transport of non-tRNAs has not been shown to date, it is plausible. We have addressed this possibility in the discussion (manuscript page 17, lines 8-10) and look forward to identifying the causal RNA in future studies.

We agree that the differential impact of xpo-3 RNAi on hoe-1(ΔNES)- and nuo-6(qm200)-induced UPRmt is interesting. One reasonable hypothesis to explain this data is that while HOE-1 processed tRNAs play a role in activating UPRmt in response to mitochondrial stress, ATFS-1 is also capable of activating UPRmt directly. In contrast, HOE-1 processed tRNAs are presumably solely responsible for UPRmt activation in hoe-1(ΔNES) animals and hence completely dependent on their exporter XPO-3. We hope to formally test this hypothesis once we identify the causal RNA species in the future.

Reviewer #2 (Recommendations for the authors):

It was reported that the epigenetic regulation of the UPRmt is in parallel with the ATFS-1 pathway (PMID: 27133168, PMID: 27133166). Whether these epigenetic factors are required for the induction of UPRmt in hoe-1(ΔNES) mutants. Similarly, whether HOE-1(NLS) suppressed the epigenetic changes or the accumulation of epigenetic factors (PMID: 27133166, PMID: 32789178, PMID: 32934238 ) in response to mitochondrial stresses.

We thank the reviewer for this suggestion. We agree that the involvement of epigenetic regulation is a plausible and intriguing possibility. To thoroughly assess such involvement, we feel, is outside of the scope of the current manuscript. We look forward to addressing this in future studies.

https://doi.org/10.7554/eLife.71634.sa2

Article and author information

Author details

  1. James P Held

    Department of Biological Sciences, Vanderbilt University, Nashville, United States
    Contribution
    Conceptualization, Formal analysis, Investigation, Methodology, Project administration, Validation, Visualization, Writing - original draft, Writing – review and editing
    Competing interests
    No competing interests declared
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0003-4322-2108
  2. Gaomin Feng

    Department of Cell and Developmental Biology, Vanderbilt University, Nashville, United States
    Contribution
    Formal analysis, Investigation, Methodology, Validation, Writing – review and editing
    Competing interests
    No competing interests declared
  3. Benjamin R Saunders

    Department of Biological Sciences, Vanderbilt University, Nashville, United States
    Contribution
    Investigation
    Competing interests
    No competing interests declared
  4. Claudia V Pereira

    Department of Biological Sciences, Vanderbilt University, Nashville, United States
    Contribution
    Investigation, Writing – review and editing
    Competing interests
    No competing interests declared
  5. Kristopher Burkewitz

    Department of Cell and Developmental Biology, Vanderbilt University, Nashville, United States
    Contribution
    Funding acquisition, Supervision, Writing – review and editing
    Competing interests
    No competing interests declared
  6. Maulik R Patel

    1. Department of Biological Sciences, Vanderbilt University, Nashville, United States
    2. Department of Cell and Developmental Biology, Vanderbilt University, Nashville, United States
    3. Diabetes Research and Training Center, Vanderbilt University School of Medicine, Nashville, United States
    Contribution
    Conceptualization, Funding acquisition, Project administration, Resources, Supervision, Writing – review and editing
    For correspondence
    maulik.r.patel@vanderbilt.edu
    Competing interests
    No competing interests declared
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0003-3749-0122

Funding

National Institute of General Medical Sciences (R01GM123260)

  • James P Held
  • Benjamin R Saunders
  • Maulik R Patel
  • Claudia V Pereira

National Institute on Aging (R00AG052666)

  • Gaomin Feng
  • Kristopher Burkewitz

National Institute of Environmental Health Sciences (T32ES007028)

  • James P Held

National Institute of General Medical Sciences (R35GM145378)

  • James P Held
  • Benjamin R Saunders
  • Maulik R Patel

The funders had no role in study design, data collection and interpretation, or the decision to submit the work for publication.

Acknowledgements

We thank Lantana K Grub and Cassidy A Johnson for their valuable feedback on the manuscript. We thank WormBase for invaluable tools and information used to plan and execute the research described. Worm strain itSi001 was graciously shared with us by Sasha de Henau. Some strains were provided by the CGC, which is funded by NIH Office of Research Infrastructure Programs (P40 OD010440). This work was generously supported by R01 GM123260 (MRP), R35 GM145378 (MRP), R00 AG052666 (KB), and by the support provided to JPH by the Training Program in Environmental Toxicology (T32ES007028). Some confocal microscopy imaging was performed through the Vanderbilt Cell Imaging Shared Resource (supported by NIH grants CA68485, DK20593, DK58404, DK59637 and EY08126). Droplet Digital PCR to quantify transcript levels was performed through the Vanderbilt University Medical Center’s Immunogenomics, Microbial Genetics and Single Cell Technologies core.

Senior Editor

  1. David Ron, University of Cambridge, United Kingdom

Reviewing Editor

  1. Xiaochen Wang, Institute of Biophysics Chinese Academy of Sciences, China

Publication history

  1. Preprint posted: June 22, 2021 (view preprint)
  2. Received: June 25, 2021
  3. Accepted: April 21, 2022
  4. Accepted Manuscript published: April 22, 2022 (version 1)
  5. Version of Record published: May 3, 2022 (version 2)
  6. Version of Record updated: May 17, 2022 (version 3)

Copyright

© 2022, Held et al.

This article is distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use and redistribution provided that the original author and source are credited.

Metrics

  • 1,493
    Page views
  • 329
    Downloads
  • 0
    Citations

Article citation count generated by polling the highest count across the following sources: Crossref, PubMed Central, Scopus.

Download links

A two-part list of links to download the article, or parts of the article, in various formats.

Downloads (link to download the article as PDF)

Open citations (links to open the citations from this article in various online reference manager services)

Cite this article (links to download the citations from this article in formats compatible with various reference manager tools)

  1. James P Held
  2. Gaomin Feng
  3. Benjamin R Saunders
  4. Claudia V Pereira
  5. Kristopher Burkewitz
  6. Maulik R Patel
(2022)
A tRNA processing enzyme is a key regulator of the mitochondrial unfolded protein response
eLife 11:e71634.
https://doi.org/10.7554/eLife.71634
  1. Further reading

Further reading

    1. Cell Biology
    2. Genetics and Genomics
    Shawn P Shortill, Mia S Frier ... Elizabeth Conibear
    Research Article Updated

    Membrane trafficking pathways perform important roles in establishing and maintaining the endosomal network. Retrograde protein sorting from the endosome is promoted by conserved SNX-BAR-containing coat complexes including retromer which enrich cargo at tubular microdomains and generate transport carriers. In metazoans, retromer cooperates with VARP, a conserved VPS9-domain GEF, to direct an endosomal recycling pathway. The function of the yeast VARP homolog Vrl1 has been overlooked due to an inactivating mutation found in commonly studied strains. Here, we demonstrate that Vrl1 has features of a SNX-BAR coat protein and forms an obligate complex with Vin1, the paralog of the retromer SNX-BAR protein Vps5. Unique features in the Vin1 N-terminus allow Vrl1 to distinguish it from Vps5, thereby forming a complex that we have named VINE. The VINE complex occupies endosomal tubules and redistributes a conserved mannose 6-phosphate receptor-like protein from endosomes. We also find that membrane recruitment by Vin1 is essential for Vrl1 GEF activity, suggesting that VINE is a multifunctional coat complex that regulates trafficking and signaling events at the endosome.

    1. Cell Biology
    2. Chromosomes and Gene Expression
    Brandt Warecki, Simon William Abraham Titen ... William Sullivan
    Research Article

    Wolbachia, a vertically transmitted endosymbiont infecting many insects, spreads rapidly through uninfected populations by a mechanism known as cytoplasmic incompatibility (CI). In CI, a paternally delivered modification of the sperm leads to chromatin defects and lethality during and after the first mitosis of embryonic development in multiple species. However, whether CI-induced defects in later stage embryos are a consequence of the first division errors or caused by independent defects remains unresolved. To address this question, we focused on ~1/3 of embryos from CI crosses in Drosophila simulans that develop apparently normally through the first and subsequent pre-blastoderm divisions before exhibiting mitotic errors during the mid-blastula transition and gastrulation. We performed single embryo PCR and whole genome sequencing to find a large percentage of these developed CI-derived embryos bypass the first division defect. Using fluorescence in situ hybridization, we find increased chromosome segregation errors in gastrulating CI-derived embryos that had avoided the first division defect. Thus, Wolbachia action in the sperm induces developmentally deferred defects that are not a consequence of the first division errors. Like the immediate defect, the delayed defect is rescued through crosses to infected females. These studies inform current models on the molecular and cellular basis of CI.