Neuronal CBP-1 is Required for Enhanced Body Muscle Proteostasis in Response to Reduced Translation Downstream of mTOR

2.1. Nematode Culture and Strains

C. elegans strains were cultured and maintained with standard procedures as described in [39] unless otherwise specified. The N2 Bristol strain was used as the reference wild type. Genotypes of strains acquired from the Caenorhabditis Genetics Center are as follows: MAH23 rrf-1(pk1417), VP303 rde-1(ne219); kbIs7[nhx-2p::rde-1 + rol-6(su1006)], NR222 rde-1(ne219); kbIs9[lin-26p::nls::GFP, lin-26p::rde-1 + rol-6(su1006)], WM118 rde-1(ne300); neIs9[myo-3p::HA::rde-1 + pRF4(rol-6(su1006))], TU3335 lin-15B(n744); uIs57 [unc-119p::YFP + unc-119p::sid-1 + mec-6p::mec-6], PS3551 hsf-1 (syf441), CF1038 daf-16(mu86): uthIs225 [sur5p::hsf-1(CT-Delta)::unc-54 3’ UTR + myo-2p::tdTomato::unc-54 3’ UTR], and NL5901 pkIs2386 [unc-54p::α-synuclein::YFP + unc-119(+)]. Transgenic lines ANR149 rrf-1(pk1417); pkIs2386[unc-54p:: α-synuclein::YFP + unc-119(+)], ANR153 rde-1(ne300); neIs9[mvo-3p::HA::rde-1 + pRF4(rol-6(su1006))]; pkIs2386[unc-54p:: α-synuclein::YFP + unc-119(+)], and ANR168 lin-15b(n744); pkIs2386[unc-54p::alphasynuclein::YFP + unc-119(+)]; uIs57[unc-119p::YFP + unc-119p::sid-1 + mec-6p::mec-6] were created by crossing NL5901 males with MAH23, WM118 or TU3335 hermaphrodites, respectively. Polymerase Chain Reaction (PCR) combined with targeted DNA sequencing was performed to validate genotypes.

2.2. RNAi Experiments

RNAi knockdown treatments were performed as described in [40]. RNAi bacteria strains included empty vector L4440 (Addgene, Cambridge, MA, USA), ifg-1 (M110.4), cbp-1 (R10E11.1), daf-15 (C10C5.6, gifted from Han Lab), and rict-1 (F29C12.3) (Ahringer library, Source BioScience, Nottingham, UK). RNAi empty vector L4440 is referred to in text as control RNAi. RNAi was carried out from day 0 of adulthood, defined by the period immediately following L4 stage, but before animals have developed visible germline. RNAi empty vector L4440 is referred to in text (Figs. 1,2,3,4, Supplementary Figs. 1–4) as control RNAi. In Fig. 4 and RNAi screens shown in Supplementary Fig. 4, a dual RNAi strategy was used in which day 0 adults were placed on RNAi for ifg-1 or control for the first two days before being transferred to plates containing RNAi for cbp-1 or other screen target genes for 5 days prior to analysis. For a full list of RNAis used in screens shown in Supplementary Fig. 4, refer to Supplementary Table 1. For all assays except RNAi screen, animals were syncronized via 2–5 hours timed egg lays, duirng which gravid adults were allowed to lay eggs, before all animals were removed from plate, leaving behind only eggs. For the RNAi screen, animals were synchronized via bleaching gravid animals to obtain similarly staged embryos.

2.3. Thermotolerance

Approximately 120 synchronized animals were maintained for each condition at 20 °C until exposure to heat stress (35 °C for 4 hours or 37 °C for 5 hours). Animals were allowed to recover for 1 hour before being scored for survival. Thereafter, survival was scored at daily intervals.

2.4. RNA Processing and Qualitative Polymerase Chain Reaction (qPCR)

RNA was isolated using Trizol Reagent (Fisher Scientific, Waltham, MA, USA) following the manufacturer’s instructions using 1–200 worms per sample. RNA samples were then processed with either Sureprep RNA Cleanup and Concentration Kit (Fisher BioReagents, Fair Lawn, NJ, USA) or RNA Clean and Concentrator (Zymo Research, Irvine, CA, USA). 200 ng RNA was reverse transcribed using QuantiTect Reverse Transcription Kit (Qiagen, Valencia, CA, USA). qPCR was performed in technical duplicate or triplicate using SYBR FAST qPCR Master Mix (Kapa Biosystems, Cape Town, South Africa) on a LightCycler 480 (Roche Applied Science, Indianapolis, IN, USA). Target gene mRNA expression was normalized to the housekeeping gene cdc-42 expression. Relative expression was determined by normalizing to control samples. Primer sequences are provided in Supplementary Table 2.

2.5. Paralysis Analysis

Synchronized NL5901, ANR149, ANR153, and ANR168 strains were maintained on control or ifg-1 RNAi plates starting at adulthood and transferred fresh RNAi plates daily. On days of paralysis measurement, 100–150 adults were transferred to a new RNAi plate with their bodies aligned in a straight line. After 10 minutes, worms that had not moved from their original location were gently tapped with a sterile platinum wire 2–3 times on the head. Worms able to move only their head (from the pharynx bulb to the tip of the head) were scored as paralyzed. Worms able to move between the pharynx and tail were considered not paralyzed. Worms unable to respond to touch were scored as dead. Paralyzed or dead worms were removed from the plate on the day of paralysis measurement.

2.6. Western Blotting

Synchronized NL5901 and ANR168 strains were maintained as they were in paralysis assays until day 7 when total protein extraction occurred. To determine the levels of alpha-synuclein proteins, western blotting was performed in triplicate. Total protein extraction and preparation was performed as previously described in [ 11]. Alpha-synuclein was detected using an anti-alpha-synuclein mouse monoclonal antibody (1:500 dilution) (Santa Cruz biotechnology, Santa Cruz, CA, USA). Beta-tubulin was detected using an anti-beta-tubulin mouse monoclonal antibody (E7, 1:500 dilution) (DSHB, Iowa City, IA, USA). Peroxidase-conjugated goat anti-mouse IgG secondary antibody (1:5000 dilution) was from Pierce (Rockford, IL, USA). The density of the bands was determined using ImageJ software (1.54i, NIH LOCI, University of Wisconsin (Madison, WI, USA)) and normalized according to beta-tubulin or total protein.

2.7. Imaging

Synchronized NL5901 animals were maintained were maintained on control or ifg-1 RNAi plates starting at adulthood and transferred fresh RNAi plates daily. Visualization and quantification of alpha-synuclein:: YFP expression was conducted after 7 days of control or ifg-1 RNAi exposure. Individual worms were mounted on a 2% agarose pad and immobilized in a drop of 25 mM levamisol. Worms were imaged on a Leica M 165 FC microscope with the YFP filter (excitation 510/20 nm, emission 560/40 nm). The intensity of YFP fluorescence was measured in ImageJ by closely tracing around individual worms. In total, approximately 30–40 worms were used per condition.

2.8. Statistical Analysis

All statistics were performed using GraphPad Prism (6, GraphPad Software Inc. (Lo Jolla, CA, USA)). Kaplan-Meier survival curves were plotted for paralysis assays and compared using the Mantel-Cox log rank test. Western blots, survival assays, and fluorescent expression analyses were compared using unpaired two-tailed t-tests with Welch’s correction. Data from quantitative PCR were assessed by performing two-way ANOVA or unpaired two-tailed t-tests with Welch’s correction.

3.1. Reduced Translation in Neurons or Germline Primes the HSR and Increases Thermotolerance

Since HSR expression is correlated with somatic protection from perturbations in proteostasis, we sought to first resolve the kinetics of HSR priming upon translation attenuation. For this, we employed RNA interference (RNAi) targeting eIF4G, which is known in C. elegans as IFG-1 and encoded by the ifg-1 gene. This treatment started on the first day of adulthood after all tissues are fully developed and continued for one week. Four HSP genes measured as a proxy for the HSR showed a decrease in the first two days after exposure to ifg-1 RNAi followed by a steady increase to peak levels by the end of the week when normalized to control RNAi for each timepoint (Fig. 1A). The delay in HSR induction corresponds with a lack of thermoprotection at two days compared to seven days of low translation observed in a previous study [11]. Thus, the time in between the rapid drop in translation and peak in HSP gene expression is considered the period of adaptation to low translation conditions.

Having established a time-course for HSR priming under ifg-1 RNAi (hereafter also referred to as a low translation condition), we sought to investigate tissue-specific effects of low ifg-1 expression. Several C. elegans strains were chosen that allow ifg-1 to be attenuated in select tissues via RNAi as previously characterized [33,41–44|. Tissues selectively targeted by RNAi include the germline, neurons, body muscle, intestine, and hypodermis (for strain details, see methods). Using the same conditions as above, we looked to see whether ifg-1 RNAi showed signs of HSR priming in these strains. Only low translation in neurons or germline elicited significantly increased HSP expression, though to varying levels, whereas other tissues exhibited a reduction (Fig. 1B, Supplementary Fig. 1). Based on these outcomes, we hypothesized that low translation in neurons or germline tissue would confer protection to thermal stress as observed for whole-body RNAi in a previous study [11].

To determine whether low translation in individual tissues could protect the entire worm from heat-derived unfolded protein stress, wild-type and tissue-specific RNAi strains were treated with ifg-1 RNAi for one week as in Fig. 1B, subjected to thermal stress for four hours, and tracked for survival. Low ifg-1 expression in N2 wild-type as well as germline or neuronal tissue increased thermotolerance, but did not increase thermotolerance through other tissues (Fig. 1C). Analysis of HSP gene expression showed that only germline- and neuron-specific RNAi strains exhibited more robust HSR gene expression following heat treatment compared to heat-treated controls (Fig. 1D). Conversely, muscle- or intestine-specific low translation significantly lowered many HSR proteins compared to heat-treated controls (Fig. 1D). Findings indicate that either low germline or neuronal translation confers enhanced protection from thermal stress.

Since eIF4G/IFG-1 is downstream of mTORC1, but not mTORC2, we tested both the pathway and tissue-specificity of the response using RNAi targeting Raptor/daf-15, a part of the mTORC1 complex and Rictor/rict-1, a part of the mTORC2 complex (Supplementary Fig. 2A). daf-15 RNAi elicited HSR priming in wild-type and germline-specific RNAi animals, but not in the neuronal RNAi strain (Supplementary Fig. 2B). However, enhanced resistance to heat was observed for the neuronal RNAi strain despite the lack of HSR priming (Supplementary Fig. 2B, lower panel). No HSR priming nor protection were conferred by rict-1 knockdown. Collectively, the data indicates the importance of mTORC1, but not mTORC2 with HSR priming and enhanced thermotolerance.

The transcription factor hsf and the C. elegans FOXO transcription factor DAF-16 regulate proteostasis and control stress resistance via the insulin-like signaling pathway [45]. To determine whether the molecular alterations and enhanced thermotolerance observed under low translation are dependent on these factors, mutant strains for daf-16 and hsf-1 were tested. HSR priming was evident in a daf-16(mu86) null mutant in response to ifg-1 RNAi. Although thermal stress survival tended to be slightly improved, results did not reach statistical significance (Supplementary Fig. 3A). Null mutants for hsf-1 are not viable and cannot be tested, however, an hsf-1(sy441) mutant lacking the carboxy-terminus DNA binding domain was available for testing and failed to induce the same level of HSR priming phenotype, demonstrating the importance of this domain for HSR induction in general (Supplementary Fig. 3B). However, despite greatly reduced HSR priming before heat challenge, following heat treatment, it resulted in a more robust HSR response and improved survival. Interestingly, a previous study showed that overexpression of hsf-1 lacking the carboxy-terminus (strain AGD794) exhibits enhanced survival [46]. Results indicate that the carboxy-terminus of HSF-1 is not necessary for enhanced thermotolerance in general. Collectively, HSR priming and thermotolerance with low translation are partially dependent on the insulin-like signaling pathway.

3.2. Reducing Translation in Neurons or the Germline Improves Motility and Restores Youthful Transcription of Muscle Maintenance Genes

The fact that selectively lowering translation in neurons or the germline improve survival from challenge with heat indicates that they are likely to improve maintenance of protein folding in other tissues. Testing this possibility requires a tissue-specific model of protein unfolding stress. For this, we employed motility assays using strain NL5901, which expresses alpha-synuclein fused with YFP in body muscle. This tissue-specific expression is due to the fact that the transgene is controlled by the major heavy myosin chain unc-54 gene promoter. Aging animals lose the ability to maintain proteostasis, which is highly exacerbated in body muscle in this strain due to expression of alpha-synuclein, resulting in paralysis in worm middle-age [47]. Here, we used paralysis to monitor the onset of proteostatic collapse in body muscle.

Prior to looking at tissue-specific effects, we tested the effect of whole body attenuation of ifg-1 during adulthood, which increased the motile period in NL5901 by an average of 44% (Fig. 2A). Unexpectedly, we observed that low translation led to increased alpha-synuclein::YFP fluorescence intensity by the end of the first week of adulthood (Fig. 2B,C). Increased protein expression was observed also in a Western blot probed with a monoclonal antibody specific for alpha-synuclein (Fig. 2D). Thus, despite increased motility under this condition, lowering systemic translation resulted in increased total protein expression for alpha-synuclein::YFP.

We wondered whether this result could be explained by increased transcriptional activity of the major heavy myosin chain gene (unc-54) promoter used to drive body muscle expression of alpha-synuclein. After seven days of ifg-1 RNAi started at the onset of adulthood, the transcript level of endogenous unc-54 was increased by more than 10-fold (Fig. 2E). In addition to unc-54, we tested expression of several other muscle structural and regulatory genes including muscle actin (act-4), minor myosin heavy chain (myo-3), troponin C (pat-10), tropomyosin (lev-11), myogenic transcription factors (hlh-1, unc-120), and the unc-23 gene encoding a negative regulator of proteosomal degradation. Expression of these genes increased in NL5901 subjected to ifg-1 RNAi. To rule out the possibility of this phenomenon being specific to this strain, we tested wild-type N2 animals with ifg-1 RNAi. A previous study showed that unc-54 decreased in the first week of adulthood in C. elegans [48]. We also observed decreased expression at the end of the first week of adulthood in wild-type animals for unc-54 and several other muscle structural and regulatory genes (Fig. 2F). Transcript expression of these muscle specific genes showed that lowering translation reversed the age-related loss of muscle specific gene expression in wild-type animals (Fig. 2G). Tissue-specific RNAi strains demonstrated that increased muscle gene expression resulted from lowering translation in the germline or neurons (Fig. 2H). Interestingly, lowering translation selectively in body muscle tissue did not induce body muscle expression significantly (Fig. 2H).

Because selectively lowering translation in neurons and germline resulted in HSR priming (Fig. 1) and an increase in muscle specific transcripts (Fig. 2), we hypothesized that selectively lowering translation in neurons or germline in the alpha-synuclein proteotoxicity model could extend their motile period and delay disease onset. We also wondered what effect lowering translation selectively in body muscle, where proteotoxicity occurs, could improve conditions in this model. Thus, we crossed the corresponding tissue-specific RNAi strains with NL5901. We confirmed that muscle structural and regulatory genes increased when translation was reduced selectively in the germline or neurons (Fig. 3A). Interestingly, slightly increased expression of muscle structural and regulatory genes was also observed when translation was selectively lowered in muscle (Fig. 3A). This result differs slightly from the results with wild-type N2 animals (Fig. 2H), which may be due to the constitutively perturbed muscle proteostasis resulting from expression of alpha-synuclein. The paralysis assay demonstrated that lowering translation in the germline or neuronal tissue resulted in the greatest increase in motility with age, whereas lowering translation in muscle resulted in a small protective effect (Fig. 3B). In summary, low translation in the neurons or germline improved body muscle proteostasis.

3.3. Neuronal CBP-1 is Required for Improved Proteostasis Due to Low Translation

Next, we sought to determine a potential mechanism behind the benefits of low translation in neurons. We chose to further examine neurons, as we previously examined other trade-offs between low translation and reproduction [33]. We performed a small screen of genes using RNAi to determine whether certain transcription factors, chromatin remodeling factors, or muscle-enriched genes were required for the increase in muscle gene expression observed under neuronal low translation conditions. Wild-type N2 animals or neuronal-specific RNAi strain TU3335 were subjected to two days of ifg-1 RNAi to lower translation before being transferred to another RNAi of interest for an additional 5 days. Translation remained low as no progeny were detected in the days following removal from ifg-1 RNAi. Because unc-54 had a robust increase in mRNA transcript abundance under low translation conditions, its expression change was used as a proxy for other muscle-specific gene expression changes. Results for a portion of the RNAi tested are shown in Supplementary Fig. 4 and Supplementary Table 1. From the RNAi screen, the transcriptional regulator gene cbp-1 was the only gene found to completely suppress the increase in unc-54 mRNA in the neuron-specific RNAi strain, but not in N2 (Supplementary Fig. 4).

Previous studies have shown that cbp-1 plays an important role in dietary restriction (DR)-mediated lifespan extension, specifically in neurons [49,50]. Based on this, and its role as a transcriptional regulator, we used the neuron-specific RNAi strain TU3335 and the dual RNAi technique used in the RNAi screen to determine whether changes in expression of other body muscle-specific genes were influenced when both ifg-1 and cbp-1 expression were reduced in neurons. Results showed that lowering neuronal cbp-1 expression significantly reduced muscle-specific gene expression changes compared with lowering ifg-1 by itself (Fig. 4A).

To determine whether CBP-1 played a role in the maintenance of muscular proteostasis with low neuronal translation, we crossed the alpha-synuclein proteotoxicity model with neuronal-specific RNAi strain to create ANR168. The combination of ifg-1 and cbp-1 RNAi prevented a significant increase alpha-synuclein protein level (Fig. 4B). In addition, knocking down expression of neuronal cbp-1 prevented a robust increase in proteostasis associated with low translation in the muscle paralysis model (Fig. 4C). Similarly, cbp-1 RNAi prevented a robust increase in proteostasis from low translation in the neuronal-specific RNAi strain TU3335 under conditions of heat stress (Fig. 4D). Together, results indicate that neuronal CBP-1 is required for the full beneficial effects of low neuronal translation on proteostasis in tissue outside the nervous system, particularly in muscle.

4.1. Responses to Low Translation are Partitioned Among Tissues and Capable of Cellular Cross-talk

The basis for loss of proteostatic maintenance with age is centered on inability to properly regulate transcriptional activation of stress response pathways and maintain the protein turnover apparatus [25,26]. Results of this study address the roles of several major tissues in mediating effects of low translation on survival under perturbed proteostasis. Low translation results from downregulation of the mTOR pathway, as occurs when nutrients are scarce, leading to slowed growth but also increased lifespan and resilience to stress. Data support a model in which neurons and the germline are tissues that control physiological responses to low translation, including enhanced somatic maintenance and function of body muscle. Although we do not know how muscle maintenance factors are invoked downstream of translation, we know some of their expression changes are dependent on neuronal expression of cbp-1.

CREB-binding factor protein (CBP) acts as a histone acetyltransferase (HAT) to acetylate key transcription factors and histones, and as a recruiter for additional transcriptional machinery [49,51]. Though expressed ubiquitously due to its localization in the nuclei of most somatic cells, in mice hypothalamic expression of CBP is correlated to increases in lifespan [47]. At the same time, decreases in its expression are associated with age and diabetes [49]. In C. elegans, CBP-1 has been shown to act almost exclusively in GABAergic neurons to double lifespan in DR worms [50]. Under DR conditions, its knockdown results in loss of lifespan extension associated with this intervention [49]. The current study shows that this factor plays a critical role in adaptive changes downstream of DR and mTOR at the level of translation in neurons.

Based on results from another study, body muscle stands out because it is the only major tissue in which selectively lowering translation reverses, at least in part, the trade-offs usually observed between longevity and development [33]. The mTOR pathway is upstream of translation. When low mTOR is driven by nutrient scarcity, it increases mitochondrial respiration in skeletal muscle and mitigates normal respiratory decline observed with age in mice [52]. In addition, stem cell availability increases under dietary restriction and low mTOR signaling in injured muscles [53]. Thus, it may be that, while translation-inhibiting conditions are detrimental during growth, in adult animals, at least certain aspects of basal muscle maintenance and function are preserved better with age when translation downstream of mTOR is reduced. Our study of the effects of low translation shows an indirect association with enhanced muscle maintenance, requiring low translation in non-muscle tissue. This could help explain the otherwise paradoxical connection between catabolism-inducing dietary restriction or low mTOR signaling conditions and improved muscle maintenance with age.

4.2. The Connection between Translational Regulation and Proteostasis

At the heart of improved muscle function and enhanced resistance to unfolded cellular protein resulting from low translation is reversal of age-related diminution of proteostatic maintenance. Many factors are involved in maintaining proteostasis, which involves synthesis, folding, and turnover of cellular protein. Nascent peptide chains emerging from the ribosome are managed by folding chaperones and are subject to ubiquitination, making translation a hub for all these processes [54]. The fact that everything from protein synthesis to degradation is regulated at the same location makes translation a critical process for maintaining this balance. We show that limiting protein synthesis controlled by an essential nutrient-responsive translation factor, eIF4G/IFG-1, acts as a lever to restore robustness of stress responsiveness through the HSR. Previously, we showed that at least one other proteostasis mechanism, the ER unfolded protein response, is enhanced when translation is reduced in a manner that depends on the gene encoding the HSR transcription factor, hsf-1 [11]. Experiments with single-celled organisms and mammalian tissue culture showed that intracellular stress signaling pathways can cross-activate one another to maintain cellular proteostasis [30,55–62]. Thus, HSR priming may improve function of other proteostasis mechanisms, a possibility that future experiments will need to address.

4.3. Limitations

Like all research studies, our study does have limitations. Firstly, though C. elegans are a highly informative, simplistic model organism, further studies are needed to determine whether cell non-autonomous sigaling under conditions of reduced translation is conserved in more complex model organisms. Secondly, we conducted a limited RNAi screen that led to the importance of neuronal CBP-1 under conditions of low translation. However, CBP-1 is most likely not the only factor at play. A larger RNAi screen is needed to determine other pathways and key regulators that provide protection from proteostatic stress under low translation. Third, our studies commenced with the discovery of the importance of neuronal CBP-1 downstream of low translation, however more research is needed into the genes regulated by CBP-1 acetylation activity. Such genes would be great candidates for RNAi screens. Lastly, though CBP-1 is ubiquitously expressed, we focused our research on the importance of neuronal expression. Future studies can examine a broader role of CBP-1 in other tissues which may explain why we observed differing tissue-specific effects of low translation.