Inhibition of elongin C promotes longevity and protein homeostasis via HIF‐1 in C. elegans

Summary The transcription factor hypoxia‐inducible factor 1 (HIF‐1) is crucial for responses to low oxygen and promotes longevity in Caenorhabditis elegans. We previously performed a genomewide RNA interference screen and identified many genes that act as potential negative regulators of HIF‐1. Here, we functionally characterized these genes and found several novel genes that affected lifespan. The worm ortholog of elongin C, elc‐1, encodes a subunit of E3 ligase and transcription elongation factor. We found that knockdown of elc‐1 prolonged lifespan and delayed paralysis caused by impaired protein homeostasis. We further showed that elc‐1 RNA interference increased lifespan and protein homeostasis by upregulating HIF‐1. The roles of elongin C and HIF‐1 are well conserved in eukaryotes. Thus, our study may provide insights into the aging regulatory pathway consisting of elongin C and HIF‐1 in complex metazoans.


Introduction
Proper levels of oxygen are essential for the survival of aerobic organisms. Hypoxia-inducible factor 1 (HIF-1) is a key transcription factor that governs cellular responses to low oxygen (reviewed in Semenza, 2012). In normal oxygen conditions, HIF-1 is hydroxylated by the proline hydroxylase EGL-9. This leads to the ubiquitination and degradation of HIF-1 by an E3 ligase containing the von Hippel-Lindau-1 (VHL-1) tumor suppressor. VHL-1 determines the substrate specificity of the E3 ligase. In contrast, hypoxia or inhibition of EGL-9 or VHL-1 promotes the stabilization of HIF-1. Stabilized HIF-1 translocates to the nucleus and regulates the transcription of genes that control hypoxic responses. The crucial functions of human HIF-1 are highlighted by the findings that HIF-1 is associated with various diseases and pathological conditions, including cancer, arterial diseases, and organ transplant rejection (reviewed in Semenza, 2012).
In our previous report, we performed a genomewide RNA interference (RNAi) screen using an HIF-1 reporter, nhr-57p::gfp transgenic C. elegans. We found 245 putative HIF-1 regulators (Lee et al., 2010;Fig. S1, Supporting information). Here, we characterized the functions of these potential HIF-1 regulatory genes in lifespan regulation. We found six genes whose knockdown increased the lifespan of worms. Among those, knockdown of elc-1, which encodes a worm homolog of elongin C, lengthened lifespan by stabilizing HIF-1. In addition, genetic inhibition of ELC-1 increased protein homeostasis in a HIF-1-dependent manner. Elongin C is evolutionarily conserved and therefore may affect aging in complex animals, such as mammals as well as C. elegans.

Results
Our previous genomewide screen was performed in a liquid culture system (Lee et al., 2010); however, conventional lifespan assays are performed in solid culture systems. Therefore, we re-examined 53 RNAi clones that were strong nhr-57p::gfp inducers and found 16 RNAi clones that robustly increased nhr-57p::gfp levels in the solid culture system [ Fig. S1, Table S1, Supporting information, and Fig. 1A. Note: Commercially available RNAi clones that were designed to target elc-1 and Y82E9BR.16 have another common target, Y82E9BR.3. We therefore designated these two RNAi clones as elc-1/Y82E9BR.3 RNAi and Y82E9BR.16/Y82E9BR.3 RNAi (Fig. S2, Supporting information)]. Surprisingly, we found that 12 of the 16 RNAi clones induced nhr-57 independently of HIF-1 (Fig. 1A). These data suggest that factors other than HIF-1 also regulate the induction of nhr-57.
Next, we performed lifespan assays with the 16 strong nhr-57 inducer RNAi clones and found that six RNAi clones significantly increased lifespan ( Fig. 1B-H, Fig. S3, and Table S2, Supporting information). RNAi targeting elc-1, a worm homolog of mammalian elongin C, and Y82E9BR.3, a worm homolog of ATP synthase subunit C, significantly promoted longevity (Fig. 1B,H). Likewise, RNAi targeting Y82E9BR.16, a worm homolog of solute carrier family 22 member 21, and Y82E9BR.3 promoted longevity (Fig. 1C,H). In addition, knockdown of the nematode-specific gene ril-2 increased lifespan (Fig. 1D,H); this result is consistent with those presented in a previous report (Hansen et al., 2005). We also found that knockdown of the mitochondrial genes F29C4.2 (a worm homolog of cytochrome C oxidase subunit 6C), C16A3.5 (a worm homolog of NADH dehydrogenase [ubiquinone] 1 b subcomplex subunit 9), and C34C12.8 (a worm homolog of mitochondrial GrpE) extended lifespan ( Fig. 1E-H). These results are consistent with many reports showing that mild inhibition of mitochondrial components confers longevity (reviewed in Van Raamsdonk & Hekimi, 2010;Hwang et al., 2012). We then examined whether the longevity caused by these six RNAi clones was dependent on HIF-1. The RNAi clone that targeted elc-1/Y82E9BR.3 was the only one that increased lifespan in a slightly hif-1-dependent manner (Figs 1H and S4,Supporting information Fig. 1 Semiquantification of nhr-57p::gfp levels and the effects of nhr-57 inducer RNAi clones on lifespan. (A) Among 53 candidate RNAi clones selected from our previous screen in a liquid culture system (Lee et al., 2010), 16 RNAi clones consistently increased the level of nhr-57p:: gfp in a solid culture system, mostly in a hif-1-independent manner. The arbitrary cutoff value was 0.5 as indicated by a dotted line. egl-9 RNAi was used as a positive control. Error bars indicate standard error of the mean (SEM) (n > 12). (B-G) Lifespan curves of wild-type (WT) animals treated with commercially available RNAi clones targeting elc-1/Y82E9BR.3 (B), Y82E9BR.16/Y82E9BR.3 (C), ril-2 (D), F29C4.2 (E), C16A3.5 (F), or C34C12.8 (G). Lifespan assays were performed at least twice independently. See Fig. S3 for the results of lifespan assays upon treating with other nhr-57p::gfp inducer RNAi clones that did not increase lifespan. (H) Percent changes in the lifespan of WT and hif-1 mutant worms after treatment with RNAi clones shown in Figs 1B-G and S4. The mean lifespan was compared with those of control RNAi-treated worms in at least two trials, and error bars indicate SEM. See Table S2 for additional trials and statistical analysis for lifespan data shown in this figure.
targeting individual genes (see Figs S2 and S5, Supporting information for detailed description). We found that elc-1-specific knockdown increased lifespan in a largely hif-1-dependent manner (in five of nine trials) ( Fig. 2A,B Table 1 and Table S2). Based on these results, we focused on the regulation of HIF-1 by ELC-1.
ELC-1 is a worm homolog of mammalian elongin C. The amino acid sequences and structures of these proteins are well conserved from yeast to humans ( Fig. 3A-C). Elongin C has two distinct roles. First, elongin C acts as a component of a transcription elongation factor in association with elongin A and elongin B (Bradsher et al., 1993a,b;Shilatifard et al., 2003). Second, elongin C functions as a component of an E3 ubiquitin ligase by binding to other components, including elongin B and pVHL; this complex determines the specificity for HIF-1a degradation (Duan et al., 1995;Kim & Kaelin, 2003). Although mammalian elongin C has been functionally characterized, it is unknown whether C. elegans elc-1 regulates HIF-1 and modulates HIF-1-dependent phenotypes.
We generated GFP-fused elc-1-expressing transgenic animals to determine the expression patterns of elc-1. We detected bright expression of ELC-1::GFP in the vulval muscle and dim expression in the pharynx, hypodermis, and intestine ( Fig. 3D-F). We found that ELC-1 localized to both the cytoplasm and the nucleus (Fig. 3F). This result is consistent with the dual roles of mammalian elongin C as a transcription elongation factor and component of an E3 ligase.
We examined whether knockdown of elc-1 affected other phenotypes, including impaired reproduction and improved protein homeostasis, which are caused by upregulation of HIF-1 (Mehta et al., 2009). We found that elc-1 RNAi conferred a severe sterile phenotype, which was mostly independent of hif-1 (Fig. 5A). Age-dependent paralysis in a transgenic worm model of Huntington's disease caused by expression of aggregation-prone Q35 was reduced by elc-1 knockdown (Fig. 5B). The effect of elc-1 RNAi was similar to that of vhl-1 RNAi (Fig. 5C and Mehta et al., 2009). Knockdown of elc-1 or vhl-1 did not improve the motility of Q35 transgenic animals in a hif-1-mutant background ( Fig. 5B,C). In addition, similar to vhl-1 RNAi (Mehta et al., 2009), elc-1 RNAi also delayed the paralysis caused by overexpression of aggregation-prone Ab, a worm model of Alzheimer's disease ( Fig. 5D; two of three trials). Together, these data suggest that inhibition of ELC-1 reduces proteotoxicity via HIF-1 but affects reproduction independently of HIF-1.

Discussion
In this study, we analyzed the lifespan-regulatory roles of putative HIF-1 regulators in C. elegans. We showed that the inhibition of C. elegans elongin C promoted longevity by upregulating HIF-1. Inhibition of elongin C also delayed paralysis in a C. elegans model of Huntington's disease that expresses a polyglutamine protein, in a HIF-1-dependent manner. In contrast, elongin C affected reproduction independently of HIF-1. Thus, we propose that ELC-1 regulates different aspects of animal physiology through both HIF-1-dependent and HIF-1-independent mechanisms.
Longevity caused by knockdown of the putative HIF-1 regulators in this study was independent of hif-1 with the exception of elc-1 knockdown. In addition, a majority of the strong nhr-57 inducer RNAi clones increased the nhr-57p::gfp levels in a largely hif-1-independent fashion. These findings are consistent with those of our previous report; This table contains summary of the lifespan results in this study except lifespan screen data. Percent changes in mean lifespan of RNAi-treated wild-type and mutant animals were calculated against control RNAi-treated wild-type and mutant worms, respectively. See Table S2 for the results of each trial and statistical analysis for lifespan data.  Table 1 and Table S2 for additional information for lifespan data shown in this figure. The lifespan-extending effect of elc-1 RNAi was modest (+12% on average) compared to that of vhl-1 mutation (+12% to +62%, Mehta et al., 2009). However, the lifespan-extending effect of elc-1 RNAi is actually comparable to that of vhl-1 RNAi on wild-type (+11%, Mehta et al., 2009). Please note that we were unable to determine the lifespan of elc-1 deletion mutants because they display a lethal phenotype. , Drosophila elongin C (79% identities), mouse elongin C (75% identities), and human elongin C (75% identities). The sequence identity values were calculated by comparing each sequence to that of C. elegans ELC-1. BLAST was conducted using Clustal W and revisualized using Clustal X2 (Larkin et al., 2007). '*' indicates an identical residue. ':' indicates a residue that has conserved amino acids with strong similarities. '.' indicates a residue that has conserved amino acids with weak similarities. Gray bars indicate scores for evolutionary conservation (Thompson et al., 1997). (B) A phylogenetic tree of elongin C in various species. This tree was generated using Clustal W2 and revisualized by Phylowidget (Larkin et al., 2007;Jordan & Piel, 2008). (C) Alignment of predicted human elongin C and C. elegans ELC-1 protein structures displays strong similarities. (D) Bright-field image of an elc-1p::elc-1::gfp animal shown in panel E. (E, F) elc-1p::elc-1::gfp was strongly expressed in the vulval muscle (E) and detected in the intestine (arrow), pharynx (arrowhead), and hypodermis (asterisk) after a longer exposure (F). Scale bar = 100 lm. Young adult worms were used for these images.
Genetic inhibition of VHL-1 and ELC-1 exerted similar and distinct effects on C. elegans physiology. Both elc-1 RNAi and vhl-1 mutation reduced fertility (shown in this study and in Mehta et al., 2009). However, elc-1 RNAi compromised reproduction independently of HIF-1 (this study), whereas vhl-1 mutations do so in a HIF-1-dependent manner (Mehta et al., 2009). These differences may be due to the different roles of ELC-1 and VHL-1 in the E3 ligase complex. ELC-1 is a core factor of E3 ligase and binds to multiple substrate-specific subunits, including VHL-1, which has limited specificity for the degradation of substrate proteins. Thus, it seems highly likely that ELC-1 affects the stability of a broader range of substrates than VHL-1 does. Indeed, an ELC-1-containing E3 ligase that contains ZIF-1 as a substrate-specific subunit affects reproduction by destabilizing PIE-1, an essential factor for germ line development (DeRenzo et al., 2003). Thus, the reduced fertility caused by ELC-1 may act through both the VHL-1/HIF-1 and ZIF-1/PIE-1 axes.
Studies have shown that elongin C has dual functions as an E3 ubiquitin ligase component and a transcription elongation factor (reviewed in Kim & Kaelin, 2003;Shilatifard et al., 2003). Nevertheless, the postembryonic functions of metazoan elongin C remain unclear. This is mainly due to the lethal phenotype of Drosophila (Mummery-Widmer et al., 2009) and C. elegans (http://www.wormbase.org/species/c_elegans/gene/elc-1) mutants and the lack of a knockout mouse model. Here, we showed that elongin C modulated several physiologic aspects, including aging and reproduction. Thus, our work will serve as a guide for future research by improving our understanding of how elongin C regulates specific physiologic processes in vivo.
Examination of nhr-57p::gfp expression upon RNAi treatment RNAi bacteria were seeded onto the wells of 24-well NGM plates in triplicate, and dsRNA was induced with 1 mM isopropyl-b-D-thiogalactopyranoside (IPTG, Gold biotechnology, St. Louis, MO, USA) at room temperature for 24 h. nhr-57p::gfp and hif-1(ia4); nhr-57p::gfp transgenic worms were synchronized on the RNAi bacteria lawns, and GFP expression was scored by three researchers in three independent experimental sets. GFP expression was scored as zero to three based on the intensity of GFP fluorescence. Zero indicated no induction, while three indicated the highest induction of GFP upon RNAi treatment. We set the criteria for the GFP scores using control RNAi (score: zero)-and egl-9 RNAi (score: three)-treated nhr-57p::gfp. Among 53 strong candidates, RNAi targeting 16 genes increased the nhr-57p::gfp levels (arbitrary cutoff value = 0.5) in the solid culture system (Fig. 1A).

Lifespan assays
Lifespan assays were performed as described previously with some modifications (Seo et al., 2013). Briefly, synchronized young adult worms were transferred onto 5 lM 5-fluoro-2 0 -deoxyuridine (FUdR, Sigma, St. Louis, MO, USA)-treated NGM plates with E. coli food. In the case of lifespan assays without FUdR treatment, worms were transferred to a new plate every one or 2 days until they stopped laying eggs. Approximately 100 worms for each condition were examined for death every 2 or 3 days until all the animals were dead. Animals that ruptured, bagged, burrowed, or crawled off the plates were censored but used as censored subjects for the statistical analysis. Lifespan assays were performed at 20°C. OASIS (http://sbi.postech.ac.kr/oasis) was used for statistical analysis (Yang et al., 2011).

Quantification of paralysis
The quantification of paralyzed worms was performed as described previously (Mehta et al., 2009) with some modifications. Briefly, the paralysis of Q35 (Q35::YFP)-and Ab-expressing animals was determined by visual analysis. Worms were classified as paralyzed if they did not show any forward movement in response to tapping. Approximately 100 worms for each condition were examined for paralysis every 3 or 4 days until day 14 to day 16. Animals that died, ruptured, bagged, burrowed, or crawled off were censored but used for statistical analysis as censored subjects. All the paralysis assays were performed at 20°C. OASIS (http:// sbi.postech.ac.kr/oasis) was used for statistical analysis (Yang et al., 2011). The format of Table S3 (Supporting information) that shows the statistical analysis of the paralysis experiments was based on a previous report (Zhang et al., 2013).

Measurement of brood size
RNAi-treated single L4 stage worm was placed on each RNAi bacteriaseeded plate. Worms were transferred to new plates every day until they stopped laying eggs. The number of hatched progeny was counted. Six to nine P0 hermaphrodites were used for measuring average brood sizes.

Western blot analysis
Synchronized young adult worms were harvested and washed using M9 buffer and then centrifuged at 2000 g for 5-10 seconds. More than 1000 worms (approximately 50 lL of worm pellets) for each condition were used for one set of sample. Worms were then immediately frozen at À80°C and mixed with 29 SDS sample buffer. The samples were boiled at 100°C for 10 min and were vortexed until the samples were broken. After 30-min centrifugation at 15 000 g, supernatant was used for the assay. The worm lysates were electrophoresed using 8% SDS-PAGE and transferred to PVDF membrane. The membrane was treated with 5% skim milk for blocking and subsequently incubated with primary antibodies against c-Myc (Santa Cruz, Paso Robles, CA, USA; 1:1000) or a-tubulin (Santa Cruz, 1:1000). The membrane was then incubated with goat anti-mouse secondary antibody conjugated with horseradish peroxidase (Thermo, Waltham, MA, USA, 1:10 000). The PVDF membrane was then treated with the chemiluminescent horseradish peroxidase substrate (Thermo) for 1 min, and the signal was detected using X-ray film (Agfa, Mortsel, Belgium). The band intensity was quantified using IMAGEJ (http:// imagej.nih.gov/ij/).

Quantitative RT-PCR
Approximately 500-1000 RNAi-treated young adult worms were used for the quantitative RT-PCR analysis. Preparation of cDNA samples was performed as previously described (Lee et al., 2010). Quantitative PCR from the cDNA was executed in a StepOne Real Time PCR System (Applied Biosystems, Foster City, CA, USA) and analyzed using comparative C T method. mRNA levels of ama-1 (the large subunit of RNA polymerase II) were used for normalization.

Supporting Information
Additional Supporting Information may be found in the online version of this article at the publisher's web-site.     Table S1 The list of RNAi clones that highly increased the level of nhr-57p:: gfp in a liquid culture system.