Summary Background Ageing is a sex-specific process characterised by a progressive decline in physiological integrity. DNA methylation represents a primary epigenetic hallmark of ageing, yet sex-specific patterns of epigenetic ageing within and across tissues remain poorly understood. This study aims to address these gaps through an integrated analysis of sex-moderated epigenetic ageing across eight human tissues. Methods A total of 137 DNA methylation datasets comprising over 36,000 individuals aged 10–114 years were analysed using a meta-analytic workflow to identify age-associated differentially methylated positions (aDMPs) and regions (aDMRs), meta-regression to assess sex moderation, and pathway enrichment analyses to interpret functional relevance. Findings Individual tissues displayed distinct age-related methylation trajectories, but some DMP sites showed consistent hyper- or hypomethylation across tissues. Across tissues, we identified 68,630 aDMPs (10%) robustly associated with ageing. Age-associated changes at the regional level were less common, with only 80 robust age-associated aDMRs detected across tissues, representing 0.09% of analysed regions. Sex moderation was observed for only 16 aDMPs (0.002%), indicating that sex effects on age-associated DNA methylation are largely tissue-specific rather than shared across tissues. Interpretation Our findings indicate that age-associated DNA methylation changes predominantly occur at isolated CpG sites rather than extended genomic regions and are strongly dependent on tissue and genomic context. The minimal overlap of sex-moderated methylation signals across tissues suggests that age-related sex differences at the epigenetic level are more likely attributable to tissue- and cell-type–specific variation rather than to broadly conserved epigenetic mechanisms shared across tissues. Funding This study was funded by an Australian Research Council (ARC) Discovery project (DP200101830). Séverine Lamon was funded by an ARC Future Fellowship (FT210100278). Nir Eynon was funded by NHMRC Investigator Grant (APP1194159), and a Hevolution/AFAR New Investigator Award in Aging Biology and Geroscience Research. Mandhri D. Abeysooryia was supported by an Australian Government Research Training Program (RTP) Scholarship. Research in context Evidence before this study DNA methylation is widely recognised as a central epigenetic hallmark of ageing. Previous research has demonstrated that some age-related methylation changes are conserved across tissues, forming the basis of pan-tissue epigenetic clocks. Most studies to date have primarily examined age effects in isolation. Although biological sex influences ageing trajectories and susceptibility to nearly all age-related diseases, sex-moderated epigenetic ageing has received limited investigation. Specifically, pan-tissue clocks, including GrimAge and PhenoAge, are “sex-aware” but were trained and validated in mixed-sex cohorts, limiting their capacity to disentangle tissue-specific sex effects. Consequently, it remains unclear whether sex-moderated epigenetic ageing signals are shared across tissues or are tissue-specific. Added value of this study This study provides a large-scale, comprehensive multi-tissue analysis of sex-moderated epigenetic ageing, integrating 137 DNA methylation datasets across eight human tissues and more than 36,000 male and female individuals spanning the lifespan. Our findings show that while age-associated methylation changes are widespread at the CpG level, sex-moderated effects are rare and largely tissue-specific, with minimal overlap across tissues. Implications of all the available evidence Together, the available evidence indicates that epigenetic ageing is predominantly driven by shared, conserved age-related methylation changes, whereas sex differences in epigenetic ageing are modest and context dependent. These sex-related effects are more likely to reflect tissue- and cell-type–specific variation rather than widespread, shared mechanisms. This underscores the need to develop sex-specific epigenetic clocks and to conduct longitudinal cohort and intervention studies to more precisely characterise sex-specific dynamics of epigenetic ageing across tissues.

Sex steroid hormones are not exclusively localised in the circulation and can be found in numerous extragonadal tissues, in concentrations unrelated to the circulating fraction. Existing methodology to measure intramuscular steroid hormone concentrations includes both immune-based assays and liquid chromatography-mass spectrometry (LC-MS), the gold standard for hormone measurements. To date, no LC-MS based methods validation has been published on the measurement of intramuscular sex steroid hormones, despite clear biological relevance. Here, we describe the development and validation of a simple, high-throughput LC-MS Orbitrap method for the measurement of 10 intramuscular sex steroid hormones, namely pregnenolone, progesterone, dehydroepiandrosterone, androstenedione, testosterone, epitestosterone, dihydrotestosterone, oestrone, oestradiol, and oestriol. In brief, isotope labelled standards were added to 5-6 mg of lyophilised muscle tissue, homogenised and extracted with ethyl acetate. The extracts were dried down and sequentially derivatised with 1-methylimidazole-2-sulfonyl chloride and hydroxylamine hydrochloride to target both the phenolic hydroxy groups and oxo groups. The limit of detection was 1.0 ± 1.0 pg/mg (range 0.36-3.26 pg/mg), with a R2 > 0.99 for all analytes. Matrix effects were 90-110% for all analytes except for dihydrotestosterone (143.6%), and precision was <10 CV% for all analytes in the presence of a muscle matrix. Our method allows for 20-40 samples to be prepared in ∼4 h, with a sample data acquisition time of 13 min. Moreover, our method provides the opportunity for specific analysis of steroid hormone concentrations in skeletal muscle, allowing target tissue specificity instead of relying on proxy measures from the circulation.

Initially thought to localize at the cytosol and nucleus only, emerging evidence indicates that miRNAs also localize within mitochondria where they could regulate diverse pathological and physiological processes. Therefore, the aim of the current study was to profile the population of miRNAs in isolated mitochondria and whole-tissue from human skeletal muscle at rest and in response to acute endurance exercise. Twelve healthy males (age 26 ± 4 years, mean ± SD) cycled for 60 min at 70% VO2peak and muscle biopsies were collected at rest, immediately after and 3 h after exercise. The mitochondria were isolated by immunoprecipitation, enzymatically purified, then the resident RNA was sequenced to assess the mitochondrial transcriptome. Small RNA sequencing revealed that mitochondria isolated from male skeletal muscle tissue contain a distinct population of miRNAs. Of the approximately 127 mature miRNAs detected in skeletal muscle mitochondria at each time point, the canonical muscle-specific miRs (myo-miRs) miR-1, miR-133 and miR-206 families constituted on average 45% of total mitochondria miRNA reads. However, none of these canonical myo-miRs were differentially expressed in mitochondria following endurance exercise. One miRNA, hsa-miR-146b-5p, was differentially expressed 3 h after exercise when compared to pre-exercise in both mitochondria (log2 fold-change = 5.4, p = 0.003, FDR = 0.82) and whole muscle tissue (log2 fold-change = 2.3, p < 0.0001, FDR = 0.060) but not when adjusted for multiple testing. Future research is now required to investigate miRNA-mRNA interactions in the mitochondria of skeletal muscle tissue.

Objective: To investigate and establish a baseline for sex and gender considerations in policy, research and curricula across the state of Victoria. Design and setting: Victoria was selected as a case study for Australia, using a mixed-methods approach to examine health and medical university curricula, research organisation policies and research funding between 2020-2025 prior to mandated inclusion. Main outcome measures: Primary outcomes include identification of predefined sex- and gender-related terms in university curricula descriptors and funded grant descriptions; and questionnaire responses from university course coordinators and organisational leads. Results: Data mining across nine Victorian universities (318 courses/3383 units) identified ~93% of units and ~60% of healthcare courses lacked sex and gender terms in their descriptors. Among medical research organisations operating in Victoria, including peak bodies, research institutes, hospitals and universities, ~70% (18/26) of the survey responders reported having no sex and gender policy. Rates of sex- and gender-term inclusion in research grants allocated in Victoria (3388) and Australia-wide (8974) validated Victoria as a case study for Australia for National Health and Medical Research Council (9.5%/8.9% respectively), Medical Research Future Funds (11%/10.2%), and Australian Research Council (4.8%/4.1%). One in nine Victorian awards from five government initiatives and one in ten from 12 non-government agencies included sex- and gender-term related terms in their guidelines. Conclusions: These baseline metrics indicate that sex and gender are still not widely considered in the education and research ecosystems. These findings support the need to build inclusive research policy at a national and state level, and accreditation standards across university education. ### Competing Interest Statement The authors have declared no competing interest. ### Author Declarations I confirm all relevant ethical guidelines have been followed, and any necessary IRB and/or ethics committee approvals have been obtained. Yes The details of the IRB/oversight body that provided approval or exemption for the research described are given below: Ethics approval was granted by Deakin University Human Ethics Committee (2024/HE000736). I confirm that all necessary patient/participant consent has been obtained and the appropriate institutional forms have been archived, and that any patient/participant/sample identifiers included were not known to anyone (e.g., hospital staff, patients or participants themselves) outside the research group so cannot be used to identify individuals. Yes I understand that all clinical trials and any other prospective interventional studies must be registered with an ICMJE-approved registry, such as ClinicalTrials.gov. I confirm that any such study reported in the manuscript has been registered and the trial registration ID is provided (note: if posting a prospective study registered retrospectively, please provide a statement in the trial ID field explaining why the study was not registered in advance). Yes I have followed all appropriate research reporting guidelines, such as any relevant EQUATOR Network research reporting checklist(s) and other pertinent material, if applicable. Yes All data supporting the findings of this study are available within the manuscript and its supplementary materials. Raw data are available at https://figshare.com/s/b32f49abd615ec38cee5. Supporting information files (S1-S7) are available at: https://doi.org/10.6084/m9.figshare.33093290. Any additional materials, including analytic code are publicly available via the study GitHub repository: https://github.com/severinelamon-prof/Sex-and-gender-incorporation-in-research-in-Victoria/tree/main. Victorian Department of Health, Women’s health and Wellbeing Program

Abstract:Sport has traditionally been organised into male and female categories, but the increasing visibility of transgender athletes has prompted sporting organisations worldwide to develop relevant eligibility regulations. These policies range from hormone thresholds and open categories to puberty-based exclusions and blanket bans, yet they are rarely based on direct scientific evidence. This review aims to examine the historical evolution of transgender inclusion in sport, map current international regulations and evaluate the evidence underpinning these approaches. First, we outline the temporal evolution of eligibility frameworks. We then review the roles of sex hormones in performance-relevant systems and assess how these mechanisms may result in physiological changes in transgender athletes receiving testosterone- or estradiol-based gender affirming hormone therapy. Studies demonstrate that gender affirming hormone therapy induces physiological changes that may translate to measures of physical fitness, but are limited by cross-sectional designs, small cohorts, short follow-up periods, heterogeneous treatment regimens and outcome measures that are rarely sport specific. Therefore, the certainty of this evidence remains low, and existing data are insufficient to support uniform policies across all disciplines. We therefore recommend that fairness and safety considerations should not be resolved through biology alone and propose a structured, sport-specific framework for policy development and revision under uncertainty.

Epigenetic changes, in particular DNA methylation, accumulate with age across different tissues, but whether these changes follow consistent patterns across different organs remains poorly understood. Here we show, through a meta-analysis of more than 15,000 human methylation profiles spanning 17 tissues, that aging produces both conserved and tissue-specific epigenetic signatures. We identify systemic shifts in methylation levels, increases in methylation variability, and growing molecular disorder across tissues. Network analysis revealed tightly connected gene clusters that are not modified by beneficial interventions, alongside a more modifiable cluster linked to NAD+ metabolism, supporting NAD+ as a potential therapeutic target in aging. A gene encoding a cell-adhesion protein, PCDHGA1, emerged as a conserved hub across tissues, implicating cell-to-cell communication pathways in aging across multiple organs. Our methylation atlas therefore provides a resource for dissecting the molecular basis of human aging and for identifying potential biomarkers and translational therapies.

Increased life expectancy across developed countries has highlighted the personal and societal value of healthy ageing. A well-functioning neuromuscular system is fundamental to quality of life and functional independence. The systemic deterioration of tissue and organ function during ageing is reflected in the diverse cellular and molecular mechanisms implicated in the age-related loss of muscle mass and strength and its extreme form, termed sarcopenia. Proposed contributors include neurodegeneration, impaired proteostasis, deficient regeneration, systemic hormonal decline, chronic inflammation, and dysregulation of muscle-resident cell populations such as muscle stem cells, fibro-adipogenic progenitors and immune cells. Recent efforts have added granularity to our understanding of the molecular response of muscle to ageing and started to unravel the cellular origins of these signals. Advancements in cell-targeting strategies (e.g. AAV capsids and antibody-targeted therapeutics) are opening new avenues for targeted interventions. Nonetheless, this raises the salient question - what should treatments for age-related muscle wasting target? While anabolic and catabolic signalling within muscle fibres has been the primary target of strategies to counteract age-related muscle loss, these efforts may be futile if impairment in other cell types such as muscle stem cells and motor neurons drive the wasting process. Furthermore, individual differences in activity, nutrition, sex, comorbidities and genetics are likely to influence the predominant mechanisms driving age-related muscle wasting. This multifactorial condition may therefore require a multifactorial solution, with scientists focusing on diverse causal mechanisms to identify and develop effective interventions.

BackgroundHuman primary muscle cell (HPMC) lines derived from skeletal muscle biopsies are potentially powerful tools to interrogate the molecular pathways underlying fundamental muscle mechanisms. HPMCs retain their genome in culture, but many endogenous circulating factors are not present in the in vitro environment, or at concentrations that do not mirror physiological levels. To address the assumption that HPMCs are valid models of age and sex-specificity in human muscle research, we examined to what extent differentiated HPMC lines retain their source phenotype in culture.MethodsBiopsies from the vastus lateralis muscle were collected from ten males aged 18-30, ten females aged 18-30 and ten males aged 60-75 recruited from a healthy population. A portion of the muscle was used for the establishment of 30 individual HMPC lines. The remaining sample was immediately snap frozen and stored for further analysis. RNA was extracted from muscle tissue samples and their corresponding, fully differentiated HMPCs and analysed using RNA Sequencing. To compare their transcriptomic signature, principal component analysis (PCA), differential expression analysis, single-cell deconvolution and pathway enrichment analysis were conducted in R.ResultsA comparison of the transcriptomic signature of 30 human muscle biopsies and their corresponding differentiated HPMCs indicated a near-complete lack of retention of the genes and pathways differentially regulated in vivo when compared to their in vitro equivalent, with the exception of several genes encoded on the Y-chromosome.ConclusionsThe diversity of resident cell populations in muscle tissue and the lack of sex- and age-dependent circulating factors in the cellular milieu likely contribute to these observations, which call for caution when using differentiated HPMCs as an experimental model of human muscle sex or age.

Abstract Background The maintenance of skeletal muscle health plays a pivotal role in prolonging both the lifespan and healthspan. However, muscle mass and strength exhibit significant declines with age. Ageing is associated with a reduced muscle protein synthesis response to key anabolic stimuli, including the androgen hormone testosterone, termed anabolic resistance. Testosterone enacts its anabolic effects in muscle through androgen receptor (AR) mediated pathways. Emerging evidence suggests that AR availability may represent a rate-limiting factor in androgen signalling, with AR saturation occurring below physiological testosterone levels in some tissues. Prior research in rodents has reported age-related reductions in AR expression, suggesting changes in AR protein content may constitute a key component of anabolic resistance. However, reports of the effects of age on the human skeletal muscle AR are inconclusive and limited by small sample sizes. Therefore, this study aimed to characterise age-related changes in expression of the AR, its regulators and downstream target genes in human skeletal muscle. Methods We developed and used a novel R-based pipeline, MetAR, to perform reproducible meta-analyses of publicly available bulk RNA-Seq datasets from NCBI GEO and investigate associations between target gene expression and variables of interest without the need for high-performance computing. Eligible datasets included skeletal muscle samples from healthy adult males aged ≥18 years, with an age range of ≥ 10 years and sample size ≥ 6. Raw counts data were downloaded, appraised and TMM normalised. Dataset-level associations between age and target gene expression were assessed using linear and generalised additive models (GAMs). Random-effects meta-analyses were performed, and heterogeneity, publication bias and leave-one-out sensitivity assessed. Results Sixteen skeletal muscle bulk RNA-seq datasets (n = 364; age 18-92 years) were eligible for inclusion in the meta-analyses. AR expression was negatively associated with age (β = −0.006 log 2 TMM-CPM per year, p < 0.001) corresponding to a 4.4% decrease in expression per decade. Age was also associated with a significant reduction in expression of various regulators of AR stability, transcriptional activity and nuclear transport. Additionally, steroidogenic enzymes and key downstream targets of the AR, including genes encoding for key structural proteins and mitochondrial function were negatively associated with age. Conclusions Collectively, these findings suggest a multi-faceted age-associated remodelling of AR expression, signalling and nuclear transport that may contribute to the development of anabolic resistance and consequent age-associated muscle loss.

Aging involves widespread epigenetic remodeling across tissues, yet the nature and consistency of these changes remain unclear. We conducted a meta-analysis of more than 15,000 human methylomes spanning 17 tissues, identifying both conserved and tissue-specific aging signatures. We examined linear changes via differentially methylated positions, variability shifts via variably methylated positions, and Shannon-entropy to capture methylation disorder. Network analysis revealed fragile co-methylation modules largely resistant to beneficial perturbation. Key disruptors, including PCDHGA1, MEST, HDAC4, and HOX genes, exacerbated aging signals across tissues. Notably, a resilient module enriched for NAD□ salvage metabolism supports therapeutic targeting of NAD□ in aging. PCDHGA1 emerged as a conserved cross-tissue driver, suggesting protocadherin-mediated adhesion plays a broader role in maintaining structural and signaling stability in multiple organ systems. Our open-access atlas provides a foundational resource for dissecting the molecular architecture of human aging and identifying testable targets for intervention, biomarkers, and translational epigenetic therapies. ### Competing Interest Statement The Regents of the University of California are the sole owner of patents and patent applications directed at epigenetic biomarkers for which Steve Horvath is a named inventor; SH is a founder and paid consultant of the non-profit Epigenetic Clock Development Foundation that licenses these patents. SH is a Principal Investigator at the Altos Labs, Cambridge Institute of Science, a biomedical company that works on rejuvenation. The other authors declare no conflict of interests. National Institute on Aging, https://ror.org/049v75w11 Hevolution Grant National Health and Medical Research Council, https://ror.org/011kf5r70 Australian Research Council, https://ror.org/05mmh0f86

INTRODUCTION:Serum concentrations of androgens and estrogens, the main male and female sex hormones, respectively, naturally fluctuate across the lifespan. Sex hormones are mainly produced in the gonads, but evidence suggests that they can also be locally synthesized in skeletal muscle. However, little is known about the purpose of intramuscular sex hormones and their role in skeletal muscle. This systematic review aimed to investigate (1) how intramuscular sex hormone concentrations vary across the lifespan, (2) whether exercise affects intramuscular sex hormone concentrations, and (3) whether intramuscular sex hormones are associated with skeletal muscle mass and function. METHODS:Four databases were searched, and studies were included if they contained measurements of intramuscular sex hormones from rodents or healthy males and females free from any hormonal treatment. RESULTS:Thirteen studies were included. Intramuscular testosterone was reduced in older males compared to their younger counterparts, but comparison of intramuscular sex hormone concentrations between pre- and postmenopausal females yielded inconclusive findings. Chronic exercise decreased androgens and estradiol in females but increased androgens in males. Acute exercise did not change intramuscular hormone concentrations in humans but increased them in rodents. Intramuscular androgens were positively associated with muscle mass and strength in males. In females, conflicting findings were reported for both estradiol and androgens and measures of muscle mass and function. CONCLUSION:Current evidence suggests that aging and exercise differentially modulate intramuscular sex hormone concentrations, and their association with muscle mass and function, between males and females and model systems.

Skeletal muscle mitochondria adaptation to exercise training is mediated by molecular factors that are not fully understood. Mitochondria import over 1000 proteins encoded by the nuclear genome, but the RNA population resident within the organelle is generally thought to be exclusively encoded by the mitochondrial genome. However, recent in vitro evidence suggests that specific nuclear-encoded miRNAs and other noncoding RNAs (ncRNAs) can reside within the mitochondrial matrix. Whether these are present in mitochondria of skeletal muscle tissue, and whether this is affected by endurance training-a potent metabolic stimulus for mitochondrial adaptation-remains unknown. Rats underwent 4 weeks of moderate-intensity treadmill exercise training, then were humanely killed and tissues were collected for molecular profiling. Mitochondria from gastrocnemius skeletal muscle were isolated by immunoprecipitation, further purified, and then the resident RNA was sequenced to assess the mitochondrial transcriptome. Exercise training elicited typical transcriptomic responses and functional adaptations in skeletal muscle, including increased mitochondrial respiratory capacity. We identified 24 nuclear-encoded coding or noncoding RNAs in purified mitochondria, in addition to 50 nuclear-encoded miRNAs that met a specified abundance threshold. Although none were differentially expressed in the exercise vs. control group at FDR < 0.05, exploratory analyses suggested that the abundance of 3 miRNAs was altered (p < 0.05) in mitochondria isolated from trained compared with sedentary skeletal muscle. We report the presence of a specific population of nuclear-encoded RNAs in the mitochondria isolated from rat skeletal muscle tissue, which could play a role in regulating exercise adaptations and mitochondrial biology.

Neuromuscular ageing is characterized by neural and/or skeletal muscle degeneration that decreases maximal force and power. Female neuromuscular ageing occurs earlier in life compared to males, potentially due to sex hormone changes during the menopausal transition. We quantified neuromuscular function in 88 females represented equally over each decade from 18 to 80 years of age and investigated the role of decreased ovarian hormone concentrations following menopause. Neuromuscular assessment included quadriceps maximal voluntary and evoked isometric torque and surface electromyography measurements, plus one-repetition maximum leg press. Voluntary and evoked torques and one-repetition maximum decreased non-linearly with age, with accelerated reductions starting during the fourth decade. An absence of changes in volitional recruitment of existing quadriceps motor units and Ia afferent facilitation of spinal motoneurons suggests that functional decline was largely mediated by impairment in intrinsic peripheral muscle function and/or neuromuscular transmission. Maximal muscle compound action potential amplitude decreased with increasing age for rectus femoris muscle only, indicating increased vulnerability to neuromuscular degeneration compared to vastus lateralis and medialis. In postmenopausal females, some variance was explained by inter-individual differences in quadriceps tissue composition and lifestyle factors, but changes in total or free concentrations of oestradiol, progesterone and/or testosterone were included in all correlations with age-related decreases in isometric voluntary and evoked torques. We demonstrate an accelerated onset of neuromuscular degeneration of peripheral muscular origin around menopause onset associated with changes in sex hormone concentrations. Interventions aimed at mitigating declines in ovarian hormones and their subsequent effects on neuromuscular function after menopause should be further explored. KEY POINTS: Neuromuscular deterioration with age is associated with poor physical function and quality of life in older adults, but female-specific trajectories and mechanisms remain unclear. This study is the first to map neuromuscular function across each decade of the adult lifespan in 88 females from 18 to 80 years old and to examine the potential role of hormonal changes after menopause. We show an accelerated reduction in neuromuscular function, primarily of peripheral muscular origin, that occurs during the fourth decade and coincides with menopause onset. In postmenopausal females, age-related reductions in neuromuscular function can in part be explained by quadriceps lean and intramuscular fat composition, physical activity and protein intake, and sex hormone concentrations. These findings help us better understand the factors that contribute to the loss of neuromuscular function with age in females, enabling the identification of potential therapeutic interventions to prolong female health span.

Androgens act through androgen receptor (AR) to maintain muscle mass. Evidence suggests that this pathway is influenced by "the gene for speed," ACTN3 (α-actinin-3). Given that one in five people lack α-actinin-3, it is possible that they may respond to androgens differently. Here, we show that α-actinin-3 deficiency decreases AR in muscles of mice and humans (in males and females) and that AR positively correlates with α-actinin-3 expression in a dosage-dependent manner. α-Actinin-3 deficiency exacerbates gastrocnemius mass loss with androgen deprivation in male mice and stunts the muscle growth response to dihydrotestosterone in female mice at the onset of puberty. This is mediated by differential activation of pathways regulating amino acid metabolism, intracellular transport, autophagy, mitochondrial activity, MAPK, and calcineurin signaling, likely driven by seven key genes that are both androgen sensitive and α-actinin-3-dependent in expression. Our results highlight a role for ACTN3 as a regulator of muscle mass and a genetic modifier of androgen action in skeletal muscle.

We investigated the molecular mechanisms of exercise adaptations in human muscle by integrating genome, methylome, transcriptome, and proteome data from over 1,000 participants (2,340 muscle samples). We identified distinctive signatures associated with maximal oxygen consumption (VO2max), and multi-omics integration uncovered five key genes as robust exercise markers across layers, with transcription factors functioning as activators, synergizing with DNA methylation to regulate gene expression. Minimal sex differences were observed, while modality-specific analysis highlighted distinct pathways for aerobic and resistance exercise, contrasting with muscle disuse patterns. Finally, we created a webtool, OMAx, featuring our individual omics and integration analysis. These findings provide a comprehensive multi-omics framework for understanding exercise-induced molecular adaptations, offering insights into muscle health, cardiorespiratory fitness, and their roles in aging and disease prevention.

Whether and how ovarian hormone fluctuations mediate the skeletal muscle response to ageing in females remains to be elucidated. We examined a tightly controlled, cross-sectional cohort of 96 females 18-80 years of age to map the functional and molecular trajectory of muscle ageing and determine its relationship with female sex hormones. Across every decade, we quantified body composition (using dual-energy X-ray absorptiometry), muscle morphology (using peripheral quantitative computed tomography), and voluntary and evoked muscle function. Circulating sex hormone concentrations were measured with GC-MS and immunoassays. Morphology and gene expression of vastus lateralis muscle samples were assessed with immunohistochemical staining and RNA sequencing, respectively. Age was negatively associated with muscle mass, strength and muscle fibre size, and positively associated with hybrid type I/IIa fibre prevalence and fibrosis. We found 37 unique patterns of gene expression across individual decades of age. Immune signalling, cellular adhesion and extracellular matrix organization pathways were the most upregulated with age, whilst mitochondrial function pathways were the most downregulated. Independently of age, circulating oestradiol and progesterone, but not testosterone, concentrations were positively associated with lean mass and negatively associated with hybrid muscle fibres across the lifespan. Oestrogen receptor binding sites were significantly enriched in upregulated genes in pre- versus post-menopausal muscle, suggesting a reduction in the translation of oestrogen target genes after menopause. Altogether, sex hormone fluctuations across the female lifespan may contribute to age-related muscle wasting, although longitudinal and interventional studies are needed to determine the causal nature of the relationship. KEY POINTS: Females live longer than males but experience worse disability in the later decades of life, highlighting the need to study female-specific patterns of ageing. This study mapped female body composition, muscle morphology, function and gene expression across every decade from 18 to 80 years of age in tightly controlled conditions and examined the relationships with circulating sex hormones. Unique patterns of muscle gene expression across ageing showed an overall increase in immune signalling and a decrease in mitochondrial respiration pathways, but limited associations with circulating sex hormones. Independently of age, circulating oestradiol and progesterone, but not testosterone, were associated with muscle mass and morphology across the lifespan, after adjusting for influential lifestyle factors (protein intake and physical activity). Fluctuations in female sex hormones across the lifespan should be considered when developing therapies to mitigate age-related muscle wasting and improve the female health span.

BACKGROUND:To date, most research investigating the influence of circulating sex hormones on ageing female skeletal muscle has been cross-sectional and focused only on dichotomised young and old, or pre- versus post-menopausal groups. This excludes an important transitional period from high to low circulating oestrogen. Using secondary data from the Baltimore Longitudinal Study of Aging, this study aimed to investigate cross-sectional and longitudinal associations between circulating sex hormones and skeletal muscle mass and function across a continuum of ages. METHODS:Multiple and binomial linear regression was used to map cross-sectional (n = 319) and longitudinal (n = 83) associations between circulating sex hormones (oestradiol (E2), free oestradiol index (FEI), total (TT) and bioavailable (BioT), testosterone, testosterone/oestradiol ratio (TT/E2)) and skeletal muscle mass and function in healthy females. Cross-sectional models analysed females across an ageing continuum (24-89 years) and longitudinal associations were tested across 4-6 years of ageing in females over 50 years old. Models were adjusted for age, height, physical activity, comorbidities, ethnicity, and follow-up time. RESULTS:Cross-sectionally, serum E2 and FEI were positively associated with relative appendicular lean mass (ALM; β = 0.28 and 0.20, respectively, p < 0.05) and thigh muscle percentage (β = 0.19 and 0.15, respectively, p < 0.05). E2 and FEI were negatively associated with total body fat percentage (β = -0.30 and -0.21, respectively, p < 0.05). BioT was positively associated with absolute ALM (β = 0.13, p < 0.05) and total body fat percentage (β = 0.18, p < 0.05). TT was negatively associated with total body fat percentage (β = -0.14, p < 0.05). The TT/E2 ratio was negatively associated with thigh muscle CSA (β = -0.08, p < 0.05) and hamstring strength (β = -0.12, p < 0.05). Across 4-6 years, decreases in E2 and FEI were associated with a decrease in ALM (β = 0.27 and 0.41, respectively, p < 0.05), and a decrease in FEI was associated with a decrease in handgrip strength (β = 0.21, p < 0.05). Decreases in TT and BioT were associated with an increase in total body fat (β = -0.25 for both, p < 0.05) and a decrease in TT was associated with an increase in hamstring specific force (β = -0.11, p < 0.05). CONCLUSION:This study demonstrates novel associations between sex hormone levels and skeletal muscle in females across a wide continuum of ages. We also demonstrate that longitudinal fluctuations in circulating sex hormones must be considered to gain a comprehensive understanding of female muscle ageing.

Hsp70 interactions are critical for cellular viability and the response to stress. Previous attempts to characterize Hsp70 interactions have been limited by their transient nature and inability of current technologies to distinguish direct vs bridged interactions. We report the novel use of cross-linking mass spectrometry (XL-MS) to comprehensively characterize the budding yeast Hsp70 protein interactome. Using this approach, we have gained fundamental new insights into Hsp70 function, including definitive evidence of Hsp70 self-association as well as multi-point interaction with its client proteins. In addition to identifying a novel set of direct Hsp70 interactors which can be used to probe chaperone function in cells, we have also identified a suite of PTM-associated Hsp70 interactions. The majority of these PTMs have not been previously reported and appear to be critical in the regulation of client protein function. These data indicate that one of the mechanisms by which PTMs contribute to protein function is by facilitating interaction with chaperones. Taken together, we propose that XL-MS analysis of chaperone complexes may be used as a unique way to identify biologically-important PTMs on client proteins. The maintenance of a correctly folded proteome (proteostasis) is critical for cell survival. Cells maintain proteostasis under both basal and stress conditions through the expression of folding chaperones such as Hsp70 and its associated co-chaperone regulators ([Hartl et al., 2011][1]; [Rosenzweig et al., 2019a][2]). Hsp70 function is dependent on three conserved domains: an N-terminal nucleotide binding domain (NBD), a substrate (“client”)-binding domain (SBD), and a C-terminal (“lid”) domain ([Frydman et al., 1994][3]; [Radons, 2016][4]). The binding and hydrolysis of ATP to ADP in the NBD promotes large-scale structural Hsp70 rearrangements that allow the closing of the CTD over client proteins that bind in the SBD, promoting protein folding ([Chirico et al., 1998][5]; [Rosenzweig et al., 2019b][6]). The characterized roles of Hsp70 include folding of new and denatured proteins; transport of mitochondrial proteins and disaggregation of protein complexes ([Artigues et al., 2002][7]; [Bush and Meyer, 1996][8]; [Nitika and Truman, 2017][9]). The essential nature of Hsp70 function in the cell, as well as its involvement in a variety of human pathologies such as cancer, has driven researchers to set out to characterize Hsp70 interactors. While great strides have been made towards this goal, these efforts have been hampered by limitations in the technologies used. For example, these past efforts have utilized affinity purification followed by mass spectrometry (AP-MS), yeast two-hybrid (Y2H) and proximity proteomics methodologies, all of which lack the ability to discriminate between direct and bridged protein interactions ([Gentzel et al., 2019][10]; [Millson et al., 2005][11]; [Truman et al., 2015][12]; [Vidal et al., 1996][13]; [Zhao et al., 2005][14]). Chemical cross-linking with mass spectrometry (XL-MS) is a powerful interactomic technique that circumvents this issue, providing information on direct interactions in protein complexes by using chemical cross-linkers ([Leitner et al., 2016][15]; [Liu et al., 2015][16]). Indeed, XL-MS studies are often complementary to the traditional structural biology methods such as X-ray crystallography, nuclear magnetic resonance, and cryo-electron microscopy ([Liu et al., 2015][16]). Importantly, a key role for Hsp70 function is stabilization and activation of a wide range of signaling molecules including those involved in processes such as DNA damage response, cell cycle control, autophagy and nutrient sensing ([Dubrez et al., 2019][17]; [Gupta et al., 2018][18]; [Truman et al., 2012][19]; [Yang et al., 2013][20]). The Hsp70 client proteins involved in these cellular processes tend to be either highly post-translationally modified (PTMs) or regulate PTMs on other proteins. In turn, these PTMs tightly regulate a multitude of protein properties including subcellular localization, enzymatic activity and protein interactions ([Nitika et al., 2020][21]). Advances in mass spectrometry-based methods have allowed identification of more than 200 different types of PTMs on proteins including phosphorylation, acetylation, and ubiquitination ([Beltrao et al., 2012][22]; [Catherman et al., 2014][23]; [Dushukyan et al., 2017][24]). Given the numerous PTMs identified on proteins, researchers are now facing difficult choices when selecting specific PTMs for further study. Computational methods for identifying important PTMs on proteins have been partially successful but rely on pre-existing MS data ([Beltrao et al., 2012][22]; [Swaney et al., 2013][25]). In this report, we have utilized XL-MS to comprehensively understand the Hsp70 interactome. In doing so, we have uncovered not only a new set of Hsp70 client proteins, but show that these clients bind at multiple sites on Hsp70, including the N-terminal NBD. Notably, many of the Hsp70 interactions with client proteins are in close proximity to biologically-important PTMs. All in all, our data suggest that our XL-MS approach to chaperone interactome characterization can also be used as a novel way to identify biologically-important and previously unknown PTMs on client proteins. ### Analysis of cross-linked yeast Hsp70 complexes Previous studies have identified proteins in complex with yeast Hsp70 (Ssa1) using quantitative AP-MS ([Truman et al., 2015][12]). To comprehensively identify direct Ssa1 client proteins and associated surfaces of interaction, we took a novel cross-linking proteomics approach. HIS-tagged Ssa1 was expressed in ssa1-4 Δ, a yeast strain in which all four SSA (Hsp70) genes have been deleted. HIS-Ssa1 complexes, with or without cross-linking with disuccinimidyl sulfoxide (DSSO), were characterized via mass spectrometry ([Figure 1A][26]). This approach facilitated the characterization of Ssa1 complexes without competition from other native Hsp70 isoforms. Quantitative proteomics identified 1,510 interactors associated with HIS-Ssa1 in the cross-linked complexes and 1,152 in the HIS-Ssa1 complexes without DSSO-mediated cross-linking ([Figure 1B][26]). We anticipated that proteins present in the DSSO-treated complexes may consist of direct Ssa1 interactors including client proteins and co-chaperones. To distinguish direct interactors of Ssa1 from indirect interactors, we filtered our data for cross-linked peptides where at least one of the identified peptides was Ssa1. We identified a total of 363 Ssa1-containing cross-linked peptides, out of which 177 were Ssa1-client/co-chaperone crosslinks and 106 were Ssa1-Ssa1 cross-links ([Figure 1C][26]). Validating our methodology, no cross-linked peptides were observed in the control uncross-linked sample. To determine whether the cross-linking process had enriched for any particular class of protein, we performed Gene Ontology (GO) analysis of unique candidate interactors of cross-linked and control samples. This GO analysis revealed enrichment of multiple cellular functions ([Figure 1D][26]). In the cross-linked samples, proteins involved in protein folding, trafficking and cell signaling were all enriched, synergistic with the established roles of Hsp70. ![Figure 1.][27] Figure 1. Cross-linking mass spectrometry of Ssa1 complexes. (A) Experimental workflow of cross-linking mass spectrometry of Ssa1 complexes purified from yeast cells. (B) Venn diagram representing Ssa1 complexes found in conventional IP and DSSO treated IP. (C) Pie chart showing types of cross links identified from XL-MS analysis. (D) Gene ontology analysis of DSSO treated Ssa1 immunoprecipitated complexes and crosslinked Ssa1 complexes using TheCellMap.org. ### Dimerization of Ssa1 is required for a subset of chaperone functions Nearly a third of the cross-linked peptides detected in our experiment were between two Ssa1 peptides ([Figure 1C][26]). To distinguish whether the cross-linked Ssa1-Ssa1 peptides came from dimerized Ssa1 molecules as opposed to single intramolecular cross-links, we initially mapped the identified cross-links onto homology-based Ssa1 models. Given the large conformational change Hsp70 undergoes during its folding cycle, we utilized two models, an ADP-bound closed structure and an ATP-bound open structure model ([Figure 2A and B][28]). The mapping of cross-links onto these models revealed that a substantial number of Ssa1-Ssa1 peptides had cross-linking lengths well within the spacer arm limit for DSSO ([Figure 2B][28]), while many others exceeded the lengths possible by cross-linking within a single molecule, implying dimerization. Considering that both bacterial and mammalian Hsp70 dimerize, we evaluated whether these cross-links provided evidence for yeast Hsp70 (Ssa1) dimerization. After mapping Ssa1-Ssa1 cross-linked peptides onto a possible Ssa1 dimer structure model based on PDB structure 2KHO ([Figure 2C][28]), it was evident that at least a subpopulation of Ssa1 dimerizes in yeast ([Figure 2C][28] and [2D][28]). To confirm this, we expressed both FLAG- and HA-tagged Ssa1 constructs in yeast. After immunoprecipitation of FLAG-Ssa1, the interaction between both tagged forms was observed ([Figure 2E][28]). Although self-interaction of yeast Hsp70 has been previously observed in vitro ([Sarbeng et al., 2015][29]), it has never been detected in live cells. To visualize Ssa1-Ssa1 interaction in vivo , we utilized bimolecular fluorescence complementation (BiFC). Yeast expressing Ssa1 tagged with Venus amino-terminal end (VN) and Venus carboxy-terminal end (VC) were examined using high-resolution fluorescence microscopy. Imaging of these cells revealed that Ssa1 dimers were clearly visible and that they localized primarily to the nucleus ([Figure 2F][28]). To demonstrate in vivo functionality of the Ssa1 dimer, residues identified as being present on the dimer interface based on the DnaK model (PDB: 4JNE) —E540 and N537—were mutated to A and K, respectively, generating a dimer-deficient mutant ([Sarbeng et al., 2015][29]). While yeast cells expressing the E540A/N537K dimer-deficient mutant were viable and grew at approximately WT rates, they were impaired for growth at high temperature ([Figure 2G][28]). To further explain this temperature-sensitive phenotype, we assessed the Heat Shock Response Element (HSE)-luciferase activity in WT and Ssa1 dimer-deficient cells. WT cells produced a robust HSE-luciferase signal after heat exposure and dimer-deficient cells did not ([Figure 2H][28]). Because Ssa1 is a major hub for protein folding in yeast, we set out to examine the possibility that some of the observed Ssa1-Ssa1 interactions might be the result of active Ssa1 folding a newly synthesized Ssa1 polypeptide chain. We studied the interactions of FLAG-Ssa1 (WT and substrate-binding deficient mutant V435F) with a known client, Rnr2 ([Truman et al., 2015][12]), the Ydj1 co-chaperone, and HA-Ssa1. Although WT Ssa1 co-purified with Rnr2, Ydj1 and Ssa1, the V435F mutant maintained interaction with Ydj1 and Ssa1, demonstrating that Ssa1 is not a client of other Ssa1 molecules ([Figure 2I][28]). Taken together, these findings confirm that Ssa1 dimerizes in yeast and that this self-interaction is important for a subset of Ssa1 functions. ![Figure 2.][27] Figure 2. A proportion of Ssa1 exists as dimer. (A) Ssa1 cross links identified from XL-MS analysis mapped on the monomeric structure of Ssa1 in open and closed conformation. (B) Ssa1 cross links mapped on the domains of Ssa1 in open and closed conformation. (C) Internal and External Ssa1 cross links were mapped on the dimeric structure of Ssa1 (4JNE). (D) Internal and External cross links were mapped on the crystal structure of Ssa1 (4JNE). (E) Immunoblot analysis of Flag-tagged Ssa1 purified from cells expressing HA-tagged Ssa1. (F) Fluorescence images of diploid cells expressing the N-terminally VN- and VC-tagged Ssa1. DAPI was used as a nuclear marker. Scale bars are10 µM. (G) Western blot analysis of Flag-tagged Ssa1 and Flag-tagged Ssa1-V435F mutants purified from cells expressing HA-tagged Ssa1. ### Hsp70 interacts with clients throughout its domains Our Hsp70 cross-linking strategy identified 124 new direct interactors of Ssa1 (Table S1). Unique direct binding proteins identified using XL-MS were mapped on the domain structure of Hsp70 ([Figure 3A][30]). We detected interactions on 58% of the DSSO-accessible lysines ([Figure 3A][30]). Interestingly, in contrast to the established paradigm that Hsp70 client proteins bind and interact solely at the SBD, the majority (79%) of the identified direct interactions mapped to the NBD ([Figure 3A][30]). Given that DSSO cross-links lysines, we considered that an explanation for such a high number of interactions with the Ssa1 NBD may be explained by the number and distribution of lysines present in each domain. Even when accounting for the relatively large number of cross-linkable lysines, the NBD bound over six times the number clients per cross-linkable lysine compared to the SBD ([Figure S1B][31] and [S1C][31]). While many of the NBD direct interactors relate to expected cell processes such as translation, chromatin organization and protein folding, there are several with unknown biological functions ([Figure 3B][30]). To validate our XL-MS screen, we confirmed several of our hits using co-immunoprecipitation and immunoblotting. Consistent with our MS data, Cct8, Pcl7, Ura8 and Sse1 all co-purified with Ssa1 and associated chaperones/co-chaperones Sse1, Hsp82 and Ydj1 ([Figure 3C][30]). ![Supplementary Figure S1][27] Supplementary Figure S1 (A) Venn diagram representing previously known physical interactors of Ssa1 versus direct interactors of Ssa1 identified in this study. (B) Scatter plot of number of interactors identified versus surface lysine on the domains of Ssa1. (C) Bar graph representing interactors per cross linkable lysine on domains of Ssa1. ![Figure 3.][27] Figure 3. Novel clients and post translationally modified clients identified on yeast Hsp70 based on XL-MS. (A) Schematic representation of 177 inter-protein crosslinks and identified post-translational modifications on domains of Hsp70. (B) Functional classification of direct Hsp70-client peptides. (C) Western blot analysis of HA-tag immunoprecipitated Cct8, Pcl7, Ura8 and Sse1 from yeast cells. ### Exploring the biological importance of novel XL-MS-identified PTMs 31% (55/177) of our cross-linked peptides contained a PTM such as acetylation, methylation or phosphorylation (see Table S1). A search for these PTMs using GPMDB () revealed that 95% of these PTMs had not been previously observed. After considering the possibility that these PTMs might be biologically important, we selected 3 different PTM-modified cross-links for further study from proteins in diverse cellular pathways; Pim1 (mitochondrial proteostasis), Mtw1 (chromosome segregation) and Ste11 (pheromone and osmotic stress response). ### HIR complex is novel client of Hsp70 The histone regulator (HIR) protein complex regulates histone gene transcription, nucleosome formation and heterochromatic gene silencing ([Mazzoni et al., 2005][32]; [Sharp et al., 2005][33]). Our XL-MS analysis revealed a novel direct interaction between Ssa1 and HIR complex components Hir1 and Hir2. We observed cross-linking between the SBD of Ssa1 and residues K435 of Hir1 and K452 of Hir2, adjacent to their respective nuclear localization signals ([Figure 4A][34] and [4B][34]). To validate our XL-MS finding, we carried out Co-IP and immunoblotting of Hir1 and Hir2 with key chaperone components Ssa1, Sse1, Hsp82 and Ydj1. These experiments confirmed a strong association between HIR and the chaperones tested ([Figure 4C][34]). Furthermore, we confirmed that Hir1 and Hir2 are bona fide Ssa1 client proteins by demonstrating that loss of Ssa1 function resulted in Hir1 and Hir2 destabilization ([Figure 4D][34]). ![Figure 4.][27] Figure 4. HIR complex is a novel client of Hsp70 in yeast and humans. (A) Schematic representation of Ssa1-Hir1/Hir2 inter protein cross-links detected on SBD of Ssa1 and NLS of Hir1 and NTD of Hir2. (B) Ssa1-Hir1/2 cross links mapped on the crystal structure of Ssa1, Hir1 and Hir2. (C) Hir complex interacts with the chaperone complex. (D) Hir1 and Hir2 are destabilized in Ssa1-45 mutant strain. (E) HIRA complex interacts with chaperone complexes in mammalian cells. IP analysis of the HIRA complex in mammalian cells. (F) Western blot analysis of HIRA upon addition of Hsp70 inhibitor JG-98 and proteasomal inhibitor Bortezomib. To demonstrate evolutionary conservation of the identified chaperone-HIR interaction, we examined interaction between human Hsc70 and HIRA, the major HIR complex protein in human cells. Consistent with our results in yeast, HA-HIRA co-immunoprecipitated with Hsc70, Hsp110, Hsp90, DNAJA1 ([Figure 4E][34]). To examine dependence of HIRA on Hsc70 chaperone activity, we treated HEK293 cells with the Hsp70 inhibitor JG-98 and monitored HIRA abundance over time. HIRA levels rapidly decreased after JG-98 addition, with HIRA becoming undetectable after 2 hours ([Figure 4F][34]). Given that in our system HA-HIRA was expressed under the constitutive human cytomegalovirus (CMV) promoter, we hypothesized that the effect we observed on HIRA abundance could be explained by protein degradation. Supporting this hypothesis, addition of the proteasomal inhibitor bortezomib prevented JG-98 dependent HIRA loss ([Figure 4F][34]). Taken together, our results suggest that HIR complex proteins are client proteins of the Hsp70 chaperone system in yeast and mammalian cells. ### Hsp70 plays a dual role in mitochondrial Pim1 protease activity Pim1 is an ATP-dependent yeast Lon protease that is involved in degradation of misfolded mitochondrial proteins, required for mitochondrial maintenance and biogenesis ([Van Dyck et al., 1994][35]). Our XL-MS data revealed an interaction between the Pim1 protease domain and the N-terminal domain of Ssa1 ([Figure 5A][36], [5B][36]). We first validated the Pim1 interaction with Ssa1 and associated co-chaperones using co-immunoprecipitation and immunoblotting ([Figure 5C][36]). The clearance of mitochondrial aggregates is important for cell homeostasis and, as such, many organisms express a Pim1 homologue. To examine whether the Ssa1-Pim1 interaction is conserved in mammalian cells, we performed an equivalent experiment to that shown in 5C, using mammalian Lonp-1 as the bait. Similar to our observations in yeast, mammalian Lonp-1 interacted with chaperone proteins including Hsc70, Hsp110, Hsp90 and DNAJA1 ([Figure 5D][36]). In order to determine if Lonp-1 is a client protein of Hsp70, we treated HEK293 cells with Hsp70 inhibitor JG-98 and observed Lonp-1 degradation after 2 hours of treatment. Treatment of HEK293 cells with bortezomib before addition of JG-98 prevented loss of Lonp-1, confirming that Lonp-1 is a client protein of Hsc70 ([Figure 5E][36]). ![Figure 5.][27] Figure 5. Pim1 phosphorylation regulates mitochondrial clearance in an Ssa1-dependent manner. (A) Schematic representation of Ssa1-Pim1 inter protein cross-links detected on NBD of Ssa1 and proteolytic domain of Pim1. (B) Ssa1-Pim1 cross links mapped on the crystal structure. (C) Pim1 interacts with the chaperone complex in yeast cells. (D) IP analysis of Lonp-1 interacts with chaperone complexes in mammalian cells. (E) Western blot analysis of Lonp-1 upon addition of Hsp70 inhibitor JG-98 and proteasomal inhibitor Bortezomib. (F) Growth assay of Pim1 phospho mutants in yeast. (G) Fluorescence images of cells expressing FlucSM–RFP and Tom70-GFP. Scale bars are 10 µM. (H) Western blot analysis of Pim1 wildtype and phosphomutants upon addition of Bortezomib. (I) IP analysis of Pim1 wildtype and phospho mutants with chaperone complex. (J) Western Blot analysis of Pim1 (wildtype and phospho-mutants) expressed in the presence or absence of Ssa1 in E.coli . (K) Schematic of Ssa1 regulation of Pim1. The Pim1 portion of the Ssa1-Pim1 cross-linked peptide contained a previously undiscovered Pim1 phosphorylation site (S974). Given its proximity to the interaction surface between Ssa1 and Pim1, we hypothesized that this phosphorylation site might be important for Pim1 function. To examine the impact of S974 phosphorylation on Pim1 function, we expressed Pim1 mutants lacking this phosphorylation site (S974A) or mimicking constitutive phosphorylation (S974D) from the regulatable CUP1 promoter in cells lacking Pim1. Although S974A cells grew at a similar rate to WT in standard growth media, S974D cells were substantially inhibited for growth ([Figure 5F][36]). To determine whether the growth defect of the S974D mutant was due to irregular mitochondrial protein aggregation, we examined the aggregation behavior of a previously established mitoFluc reporter ([Ruan et al., 2020][37]; [Ruan et al., 2017][38]). S974D mutants were unable to clear mitochondrial protein aggregates ([Figure 5F][36], [Figure 5G][36]). To query whether loss of Pim1 function in the S974D mutant was due to its mislocalization, we examined localization of GFP-tagged WT, S974A and S974D Pim1 proteins. Intriguingly, Pim1 localization was unaffected by the phosphorylation state of S974 ([Figure S2A][39]). ![Supplementary Figure S2][27] Supplementary Figure S2 (A). Localization of Pim1 wildtype and the mutants in yeast cells. (B) Quantification of percentage of cell with mitoFluc labeled DUMP structures in ([Figure 5G][36]). Paired t-test was used for statistical analysis. (C) Western blot showing the levels of Pim1 wildtype and the mutants in yeast cells. ![Supplementary Figure 3][27] Supplementary Figure 3 (A) Growth curves of wild-type (BY4741) and Mtw1-Y86F yeast strains. Cells were diluted from an overnight culture to OD600=0.03 and the OD600 was measured every 5 minutes for 16 hours using a microplate reader. Growth analysis was performed in 10 replicates per strain. (B) Doubling times for wild-type and mutant strains were calculated for each replicate and compared using Student’s t-test (mean=92.97 min, p-value=0.04). Like many proteases, Pim1 undergoes self-cleavage to achieve full maturation and protease activity ([Ondrovicova et al., 2005][40]). We examined Pim1 processing in WT, S974A and S974D cells. In contrast to both WT and S974A, S974D resolved as a single band on SDS-PAGE, suggesting that Pim1 self-cleavage and maturation was compromised in S974D cells ([Figure 5H][36], S2C). The proximity of S974 to the identified Ssa1-Pim1 cross-linked peptides suggested that this residue was important to the Ssa1-Pim1 interaction. Immunoprecipitation of Pim1 variants demonstrated that, although not critical for Ssa1-Pim1 interaction, S974 phosphorylation significantly enhanced the interaction between the two proteins ([Figure 5I][36]). Based on our results shown in [Figures 5A][36]-[5I][36], we hypothesized that Ssa1 may bind to phosphorylated Pim1 to inhibit its function. To examine Pim1 activity in the absence of endogenous Ssa1 activity, we expressed recombinant yeast Pim1 (WT, S974A and S974D) in E. coli ([Figure 5J][36]). In contrast to our findings in 5H, Pim1 self-cleavage was independent of S974 phosphorylation status in the absence of Ssa1 ([Figure 5J][36]). Upon co-expression of Pim1 and Ssa1 in bacteria, self-cleavage of Pim1 was restored only in the S974D mutant ([Figure 5J][36]). Taken together, our findings demonstrate that fascinatingly, Pim1 is not only a novel client of Ssa1, but that Ssa1 can prevent phosphorylation-mediated Pim1 self-cleavage ([Figure 5K][36]). ### Ssa1 regulates kinetochore function via Mtw1 Mtw1 is an essential component of the MIND kinetochore complex and it connects kinetochore subunits binding DNA to those associated with microtubules, rendering it critical to kinetochore assembly. Direct interaction between the NBD of Ssa1 and the head domain of Mtw1 was observed by XL-MS ([Figure 6A][41], [6B][41]). To confirm direct interaction between Mtw1 and the Hsp70 chaperone system proteins (Ssa1, Sse1, Hsp82 and Ydj1) we used a Co-IP and immunoblotting approach, similar to the one used above for other identified direct interactors ([Figure 6C][41]). Perturbation of Ssa1 function destabilized Mtw1 confirming its status as a new Hsp70 client protein in yeast ([Figure 6D][41]). ![Figure 6.][27] Figure 6. Mtw1 is a client of Hsp70 and is regulated by phosphorylation. (A) Schematic representation of Ssa1-Mtw1 inter protein cross-links detected on NBD of Ssa1 and head domain of Mtw1. (B) Ssa1-Mtw1 cross links mapped on the crystal structure of Ssa1 and Mtw1. (C) Mtw1 interacts with the chaperone complex. (D) Mtw1 is destabilized in Ssa1-45 mutant strain. (E) MIS12 interacts with chaperone complexes in mammalian cells. (F) Western blot analysis of MIS12 upon addition of Hsp70 inhibitor JG-98 and proteasomal inhibitor Bortezomib. (G) Growth assay analyzing the phenotype of the Mtw1 and its Y86 mutant. (H) Mtw1 was tagged with YFP in wild-type and Mtw1-Y86F mutant strains to compare Mtw1 localization at the kinetochore using Fluorescence microscopy. (I) Fluorescence intensities were quantified in wildtype and mutant Mtw1 using the semi-automated FociQuant ImageJ script ([Ledesma-Fernandez and Thorpe, 2015][42]). Intensities were compared using the Student’s t-test (p-value=1.8E-12). (J) Analysis of the impact of Mtw1 phosphorylation on interaction with chaperones. (K) Model of Mtw1 activity regulation via its phosphorylation. The mammalian equivalent of Mtw1, Mis12, is critical for correct kinetochore attachment ([Petrovic et al., 2010][43]). The Ssa1-Mtw1 interaction is conserved in mammalian cells as evidenced by the successful co-purification of Mis12 with Hsc70, Hsp110, Hsp90 and DNAJA1 ([Figure 6E][41]). Furthermore, treatment of HEK293 cells with Hsp70 inhibitor JG-98 resulted in loss of Mis12, and addition of the proteasomal inhibitor bortezomib prevented JG-98-mediated Mis12 destruction, showing that inhibition of Hsp70 leads to proteasomal degradation of Mis12 ([Figure 6F][41]). Taken together, our results suggest that Mis12 is a novel client protein of Hsp70. The identified site of interaction between Mtw1 and Ssa1 contained a previously undiscovered phosphorylation site, Y86 on Mtw1 ([Figure 6A][41]). Previous studies have shown that Y86 is at the interface between Mtw1 and a second essential kinetochore subunit, Mif2 ([Dimitrova et al., 2016][44]; [Hornung et al., 2011][45]; [Killinger et al., 2020][46]). Haploid yeast cells expressing the non-phosphorylatable Y86F Mtw1 mutant protein were viable but compromised for growth on the microtubule-perturbing agent benomyl ([Figure 6G][41], S3A and B). Notably, we were not able to generate cells expressing a phospho-mimetic Y86E Mtw1 mutant protein. We compared Mtw1 localization at the kinetochore in both WT and Y86F mutant strains. Y86F cells showed a significantly increased accumulation of Mtw1 at the kinetochore compared to wild-type cells ([Figure 6H][41] and [6I][41]). Although the Y86F mutation significantly impacted interaction between Mtw1 and either Hsp82 or Ydj1, Ssa1 was unaffected by the Y86F mutation ([Figure 6J][41]). Taken together our data suggests phosphorylation of Mtw1 and its interaction with Ssa1 affects its incorporation into kinetochores ([Figure 6K][41]). ### Ssa1 is involved in the selective activation of Ste11-mediated signaling Ste11 is a MEK kinase involved in the cellular responses to both pheromone and hypo-osmolarity ([Harris et al., 2001][47]; [Nishimura et al., 2016][48]; [Tatebayashi et al., 2006][49]; [Tatebayashi et al., 2020][50]) ([Figure 7A][51]). It was an encouraging validation of our XL-MS methodology to observe direct interaction between the N-terminus of Ssa1 and the unstructured regulatory domain of Ste11, one of the first identified client proteins of Hsp90 ([Figure 7B][51]). As with Mtw1 and Pim1, the Ssa1-Ste11 peptide contained a previous undiscovered PTM, dimethylation on Ste11 R305. To determine the functional importance of dimethylation of Ste11 R305, we created the non-methylatable mutant R305A and dimethylation-mimic R305F and expressed these in cells lacking Ste11. To examine the impact of R305 on the pheromone response, we assessed the ability of Ste11 R305A and R305F to form halos in response to alpha factor, activate a FUS1-LacZ reporter and promote Fus3 phosphorylation. In all 3 experiments, both R305A and R305F behaved in a similar manner to WT ([Figures 7C, D, E][51]). ![Figure 7.][27] Figure 7. Ste11 dimethylation impacts the osmotic stress response. (A) Depiction of Ste11 pathway under hyperosmotic stress. (B) Schematic representation of Ssa1-Ste11 inter protein cross-links detected on NBD of Ssa1 and regulatory domain of Ste11. (C) Halo assay analyzing the phenotype of the Ste11 wildtype and methylation mutants in response to alpha factor. (D) FUS1-lacZ activity of Ste11 mutants in response to pheromone. (E) Western blot analysis of the effect of Ste11 wildtype and mutants in response to pheromone signaling. (F) Growth assay analyzing the phenotype of the Ste11 wildtype and the methylation mutants in hyperosmotic stress. (G) 8xCRE lacZ activity of Ste11 mutants in response to osmotic stress. (H) Halo assay showing that none of the Ste11 variants permit pheromone response in the absence of Ste20. (I) IP analysis of Ste11 with chaperone complex. (J) Model of Ste11 activity regulation via its methylation. Although the R305 dimethylation status had minimal impact on the pheromone response, previous studies have demonstrated that some hyper-active mutations within this region (e.g. STE11-Q301P or STE11-DDD ) rescue the osmo-adaptation defect caused by deletion of upstream Ste20 kinase upon high osmolarity ([Tatebayashi et al., 2006][49]; [Tatebayashi et al., 2020][50]). We examined the ability of Ste11 R305 mutants to complement the loss of Ste11 in response to osmotic shock. Because Ste11 is essential for Hog1 MAPK activation by osmostress only when the other upstream pathway (the SLN1 branch) is inactivated, we deleted the SSK2 and SSK22 genes encoding the MEKKs for the SLN1 branch from the host cells ([Figure 7A][51]). As with the pheromone response, the R305 dimethylation status appeared to be dispensable for Ste11 function in this regard ([Figure 7F][51], upper panel). It is known that deletion of the upstream components of the Ste11-osmotic response pathway such as Ste20, renders cells sensitive to media containing NaCl ([Nishimura et al., 2016][48]; [Tatebayashi et al., 2006][49]; [Tatebayashi et al., 2020][50]). Interestingly, we found that while expression of WT and R305A Ste11 in cells lacking native Ste11, Ssk2, Ssk22 and Ste20 had no discernible effect on osmotic resistance, R305F Ste11 rendered cells resistant to NaCl ([Figure 7F][51], lower panel). To determine the impact of these R305 mutations on Hog1 activation, we performed reporter assays using the Hog1 reporter 8xCRE-lacZ on cells examined in 7F. The R305F mutant was able to induce CRE -mediated transcription at a level several fold greater than WT Ste11, albeit at a lower level than that observed for previously characterized hyperactive Ste11 mutants DDD and Q301P ([Figure 7G][51]). To determine whether this R305F phenotype in was specific to the osmotic stress response in ste11/ssk2/ssk22/ste20Δ cells, we performed a halo assay on the cells from 7G. In contrast to the results in 7F and 7G, Ste11 R305F was indistinguishable from WT or R305A ([Figure 7H][51]). We wondered whether R305 dimethylation might impact interaction with either Ssa1 or Ydj1, particularly given that the Ssa1-Ste11 cross-link was formed adjacent to the R305 site. Interaction studies using immunoprecipitated Ste11 suggest that R305 dimethylation was not essential for interaction with Ssa1 or Ydj1 ([Figure 7I][51]). Taken together, our data indicates a novel interaction of Ssa1 with the unstructured regulatory domain of Ste11 adjacent to a dimethylation site that impacts only osmotic signaling ([Figure 7J][51]). ### Towards a comprehensive Hsp70 interactome The identification and characterization of new chaperone interactions is important to understand the fundamental process of protein folding ([Bohen et al., 1995][52]). This knowledge can ultimately lead to the design of novel therapies that rely on the manipulation of chaperone function. While large-scale interactome studies of chaperones have been attempted previously, the methods used in these attempts have associated drawbacks that prevent a comprehensive analysis of the system ([Koegl and Uetz, 2007][53]; [Yugandhar et al., 2019][54]). For example, several of these technologies such as LUMIER are performed on purified proteins and thus might not have the dynamic range to detect the impacts of PTMs and scaffold proteins ([Taipale, 2018][55]; [Taipale et al., 2014][56]). Other cell-based assays such as AP-MS, Y2H and proximity labelling lack the ability to distinguish direct from bridged interactions. All of these drawbacks can be addressed by using the XL-MS technology we describe in this report. XL-MS methodologies provide the ability to both stabilize transient interactions and allow characterization of the interaction surface between two proteins ([Leitner et al., 2016][15]; [Liu et al., 2015][16]). Recently, this innovative technology was used to obtain a more complete interactor list for Hsp90, providing a more accurate picture of its molecular dynamics ([Chavez et al., 2016][57]). For these reasons, we have used XL-MS to characterize the Hsp70 interactome in yeast. Our study identified a total of 1510 different proteins complexed with Hsp70, 238 of which were confirmed to be direct interactors. Importantly, 121 of these direct interactions had never been previously observed, validating the use of XL-MS technology to comprehensively study the Hsp70 interactome in the future. It is interesting to speculate what the remaining 1274 interactors represent. They may be bridged interactors present in association with Hsp70 complexes. If that is the case, it would suggest that large-scale datasets claiming to identify direct chaperone interactions might need to be revisited for accuracy. On the other hand, these interactions may be direct bona fide interactors that for technical reasons were unable to be cross-linked to Ssa1. Previous in vitro studies suggested that the majority of interactions would be localized to the CTD of Hsp70, the domain recognized as being responsible for binding and processing of Hsp70 client proteins ([Radons, 2016][4]). Using our XL-MS method, we identified new Ssa1-interactions that map to other Hsp70 domains ([Fig. 3A][30]). Unexpectedly, 79% of the total interactions identified mapped to the NBD of Hsp70 ([Fig. S1 B][31] and [C][31]). Even after accounting for the number of cross-linkable lysines present on each Hsp70 domain, the NBD had over sixfold the number of interactions compared to the CTD. Biologically, there may be several explanations for this result. Firstly, it is possible that during the client protein-binding process, there are multiple interactions between chaperone and client-protein that engage the entirety of Hsp70. Although this phenomenon has been observed in vitro between recombinant Hsp70 and single client proteins, our work is the first to make a similar observation at the interactome level ([Mashaghi et al., 2016][58]). Secondly, we may be detecting the interaction of Hsp70 in fully-formed protein complexes. Finally, several of the N-terminal interactions may represent novel co-chaperones/regulators of Hsp70. Our goal was to identify a comprehensive clientome of Hsp70. To achieve this, we did not replenish ATP during the cross-linking and purification process, skewing the complexes towards co-chaperone free, client-bound Hsp70 complexes. In agreement with this, although several co-chaperones (Sse1, Cct8, Ydj1) were detected in complexes with Hsp70, very few were identified in our cross-linked samples. Although beyond the scope of this study, future experiments may entail purification of Hsp70 complexes in different stages of the folding cycle to trap co-chaperones rather than clients. While proteomics methods can undoubtedly produce non-native interactions, all the hits selected for follow up in this report were confirmed to be genuine Hsp70 interactors in both yeast and mammalian cells. Given the stress-dependent nature of the Hsp70 interactome, it will be important to perform variations of this XL-MS experiment under different stress conditions such as heat, cell cycle stage, DNA damage response and nutrient deprivation. ### Understanding novel Ssa1-Ssa1 interactions An advantage of XL-MS methodologies is the ability to detect a wider range of information about protein folding and structure. For example, in the case of Hsp90, XL-MS has been used to understand protomer conformational changes upon ATP binding ([Chavez et al., 2016][57]). In this study, we identified 177 internal Ssa1-Ssa1 cross-links. Although the structure of full length Ssa1 protein has yet to be obtained, sequence similarity to bacterial and mammalian Hsp70 strongly suggests that Ssa1 forms similar ATP and ADP-bound conformations ([Bertelsen et al., 2009][59]; [Rosenzweig et al., 2019b][6]; [Zhu et al., 1996][60]). Serving as an internal control to our experiment, the majority of obtained Ssa1-Ssa1 peptides could be matched to these structures. We were surprised by the number of remaining peptides that could not be matched to any known monomeric Hsp70 structure. Deeper analysis of these peptides revealed that these were likely the result of cross-linking of two different Ssa1 molecules. The evidence supporting this is twofold. Firstly, the distance between these cross-linked Ssa1-Ssa1 peptides exceeded the known DSSO cross-linker length. Secondly, several of the cross-linked Ssa1-Ssa1 peptides were symmetrical; the peptides on each side of the cross-link were the same. While clearly not at the same stoichiometry as Hsp90, several studies on bacterial and human Hsp70 have demonstrated a capacity for the purified chaperone to form higher-order structures ([Bertelsen et al., 2009][59]; [Liu et al., 2017][61]; [Morgner et al., 2015][62]; [Sarbeng et al., 2015][29]; [Takakuwa et al., 2019][63]; [Trcka et al., 2019][64]) Expression of a dimerization-deficient DnaK in bacteria produces viable cells that are sensitive to thermal stress, suggesting that dimerization is needed for a subset of DnaK functions ([Liu et al., 2017][61]). Through both co-immunoprecipitation and BiFC, we demonstrate for the first time that yeast Ssa1 can also form dimers in cells. These dimers are almost exclusively localized to the nucleus, suggesting an nuclear-specific function. While the role of the Ssa1 dimer remains to be explored, given that regulation of the heat shock response by HSF occurs in the nucleus, we hypothesize that dimerization may be a novel way to regulate HSF activation in cells. ### Using sites of Hsp70 interaction to reveal novel PTMs of functional relevance on client proteins Interactions with molecular chaperones are critical for supporting the function of proteins involved in signal transduction, particularly those involved in PTMs such as kinases or acetylases ([Chen et al., 2014][65]; [O’Regan et al., 2015][66]; [Tao et al., 2016][67]). Some of these interactions are quite stable, with client proteins requiring continuous chaperone interaction for activity ([Kim et al., 2013][68]). Others are transient, where full client protein maturation and activity requires rapid chaperone dissociation ([Rosenzweig et al., 2019b][6]). However, there is also a growing body of work that shows that PTMs play an important and novel role in chaperone interactions. For example, the client-phosphorylation status of the Hsp90-Mpk1 and Hsp90-ERK5 complexes have been shown to be important for their function ([Piper et al., 2006][69]). The chaperone code can also regulate interactions with many chaperones/co-chaperones including Hsp90, Hsp70, HSF, Cdc37 ([Cloutier and Coulombe, 2013][70]; [Nitika et al., 2020][21]; [Nitika and Truman, 2017][9]). While these PTMs have been identified on a one-at-a-time basis, ours is the first study to identify important chaperone-PTM interactions on a much larger scale. In this work, we have followed up on three Hsp70 clients that contain novel PTMs on the site of interaction with Ssa1-Mtw1, Pim1 and Ste11. In each of these cases, while the client protein PTMs had been previously undiscovered, we have shown them to regulate novel and distinct (and very different) client protein functionalities. For Mtw1, PTM appears to regulate its localization, required for full functionality. For Pim1, S974 phosphorylation doesn’t impact stability or localization, but rather self-processing and proteolytic activity. For Ste11, one the canonical Hsp90 clients, our working model is that dimethylation regulates the transition between inactive and active Ste11 conformations, either independent of or in conjunction with its regulation by phosphorylation. While beyond the scope of this study, we hypothesize that this dimethylation is selectively impacting interactions of Ste11 with components of the osmotic stress response pathway such as Ste50. Future studies will aim to decipher the stresses and enzymes that regulate these PTMs in addition to teasing apart their hierarchy of interaction with Ssa1. We currently have two working models; in the first the presence of the PTM recruits Hsp70 to alter client protein interactions and therefore function. In the second, Ssa1 is acting as a “protective cover” over the novel PTM, trapping the PTM in its on/off state, ultimately altering the kinetics of client activation. These two models, while reasonable, are hard to test given that Ssa1 is essential for cell growth and acts at multiple points in the signal transduction pathways involved. This study has unveiled new regulatory mechanisms for Hsp70 and its client proteins. In doing so, we have established novel tools, methods, and workflows that will allow the field to achieve a more complete understanding of chaperones and the ways in which cells use them to integrate signal transduction pathways. N, BZ, JET, MJW and KT conducted experiments and analyzed data with assistance from AWT and PMP. JTK, JK, RH and LF assisted with XL-MS experiments and data analysis. LR, YW and RL, contributed to the Pim1 mutation and microscopy analysis. MTT, CK and PT performed the Mtw1 microscopy. N and AWT designed the study, interpreted the data, and wrote the first draft of the manuscript. VAS edited and revised the manuscript. AWT supervised the project. All authors reviewed and approved the final manuscript. This work was supported by the NIH (R15GM139059 and R01GM139885 to AWT, R01GM057769 to PMP), the Queen Mary University of London and the Francis Crick Institute (Cancer Research UK—FC001183; UK Medical Research Council—FC001183 and the Wellcome Trust—FC001183 to PHT), a grant from Re-Stem Biotech to R.L., the JSPS Grants-in-Aid for Scientific Research (KAKENHI) (21H02422 to KT), the Institute for Fermentation, Osaka (G-2021-2-082 to KT). We thank Dr. Paola Lopez-Duarte for the use of her confocal microscope. AP-MS : Affinity purification mass spectrometry BiFC : Bimolecular fluorescence complementation CMV : human cytomegalovirus promoter CTD : C-terminal Domain DSSO : disuccinimidyl sulfoxide GO : Gene ontology HSE : Heat Shock Response Element HSP : Heat Shock Protein MS : mass spectrometry NBD : Nucleotide-binding domain PTM : Post-translational modifications SBD : Substrate-binding domain VN : Venus amino-terminal end VC : Venus carboxy-terminal end XL-MS : Cross-linking mass spectrometry Y2H : Yeast two-hybrid 1. [↵][71]Artigues, A., Iriarte, A., and Martinez-Carrion, M. (2002). Binding to chaperones allows import of a purified mitochondrial precursor into mitochondria. J Biol Chem 277, 25047–25055. 2. [↵][72]Beltrao, P., Albanese, V., Kenner, L.R., Swaney, D.L., Burlingame, A., Villen, J., Lim, W.A., Fraser, J.S., Frydman, J., and Krogan, N.J. (2012). Systematic functional prioritization of protein posttranslational modifications. Cell 150, 413–425. 3. [↵][73]Bertelsen, E.B., Chang, L., Gestwicki, J.E., and Zuiderweg, E.R. (2009). Solution conformation of wild-type E. coli Hsp70 (DnaK) chaperone complexed with ADP and substrate. Proc Natl Acad Sci U S A 106, 8471–8476. 4. [↵][74]Bohen, S.P., Kralli, A., and Yamamoto, K.R. (1995). Hold ‘em and fold ‘em: chaperones and signal transduction. Science 268, 1303–1304. 5. [↵][75]Bush, G.L., and Meyer, D.I. (1996). The refolding activity of the yeast heat shock proteins Ssa1 and Ssa2 defines their role in protein translocation. J Cell Biol 135, 1229–1237. 6. [↵][76]Catherman, A.D., Skinner, O.S., and Kelleher, N.L. (2014). Top Down proteomics: facts and perspectives. Biochem Biophys Res Commun 445, 683–693. 7. [↵][77]Chavez, J.D., Schweppe, D.K., Eng, J.K., and Bruce, J.E. (2016). In Vivo Conformational Dynamics of Hsp90 and Its Interactors. Cell Chem Biol 23, 716–726. 8. [↵][78]Chen, Y.J., Lai, K.C., Kuo, H.H., Chow, L.P., Yih, L.H., and Lee, T.C. (2014). HSP70 colocalizes with PLK1 at the centrosome and disturbs spindle dynamics in cells arrested in mitosis by arsenic trioxide. Arch Toxicol 88, 1711–1723. 9. [↵][79]Chirico, W.J., Markey, M.L., and Fink, A.L. (1998). Conformational changes of an Hsp70 molecular chaperone induced by nucleotides, polypeptides, and N-ethylmaleimide. Biochemistry 37, 13862–13870. 10. [↵][80]Cloutier, P., and Coulombe, B. (2013). Regulation of molecular chaperones through post-translational modifications: decrypting the chaperone code. Biochim Biophys Acta 1829, 443–454. 11. [↵][81]Dimitrova, Y.N., Jenni, S., Valverde, R., Khin, Y., and Harrison, S.C. (2016). Structure of the MIND Complex Defines a Regulatory Focus for Yeast Kinetochore Assembly. Cell 167, 1014–1027 e1012. 12. [↵][82]Dubrez, L., Causse, S., Borges Bonan, N., Dumetier, B., and Garrido, C. (2019). Heat-shock proteins: chaperoning DNA repair. Oncogene. 13. [↵][83]Dushukyan, N., Dunn, D.M., Sager, R.A., Woodford, M.R., Loiselle, D.R., Daneshvar, M., Baker-Williams, A.J., Chisholm, J.D., Truman, A.W., Vaughan, C.K., et al. (2017). Phosphorylation and Ubiquitination Regulate Protein Phosphatase 5 Activity and Its Prosurvival Role in Kidney Cancer. Cell Rep 21, 1883–1895. 14. [↵][84]Frydman, J., Nimmesgern, E., Ohtsuka, K., and Hartl, F.U. (1994). Folding of nascent polypeptide chains in a high molecular mass assembly with molecular chaperones. Nature 370, 111–117. 15. [↵][85]Gentzel, M., Pardo, M., Subramaniam, S., Stewart, A.F., and Choudhary, J.S. (2019). Proteomic navigation using proximity-labeling. Methods 164-165, 67–72. 16. [↵][86]Gupta, A., Puri, A., Singh, P., Sonam, S., Pandey, R., and Sharma, D. (2018). The yeast stress inducible Ssa Hsp70 reduces alpha-synuclein toxicity by promoting its degradation through autophagy. PLoS Genet 14, e1007751. 17. [↵][87]Harris, K., Lamson, R.E., Nelson, B., Hughes, T.R., Marton, M.J., Roberts, C.J., Boone, C., and Pryciak, P.M. (2001). Role of scaffolds in MAP kinase pathway specificity revealed by custom design of pathway-dedicated signaling proteins. Curr Biol 11, 1815–1824. 18. [↵][88]Hartl, F.U., Bracher, A., and Hayer-Hartl, M. (2011). Molecular chaperones in protein folding and proteostasis. Nature 475, 324–332. 19. [↵][89]Hornung, P., Maier, M., Alushin, G.M., Lander, G.C., Nogales, E., and Westermann, S. (2011). Molecular architecture and connectivity of the budding yeast Mtw1 kinetochore complex. J Mol Biol 405, 548–559. 20. [↵][90]Killinger, K., Bohm, M., Steinbach, P., Hagemann, G., Bluggel, M., Janen, K., Hohoff, S., Bayer, P., Herzog, F., and Westermann, S. (2020). Auto-inhibition of Mif2/CENP-C ensures centromere-dependent kinetochore assembly in budding yeast. EMBO J 39, e102938. 21. [↵][91]Kim, Y.E., Hipp, M.S., Bracher, A., Hayer-Hartl, M., and Hartl, F.U. (2013). Molecular chaperone functions in protein folding and proteostasis. Annual review of biochemistry 82, 323–355. 22. [↵][92]Koegl, M., and Uetz, P. (2007). Improving yeast two-hybrid screening systems. Brief Funct Genomic Proteomic 6, 302–312. 23. [↵][93]Ledesma-Fernandez, E., and Thorpe, P.H. (2015). Fluorescent foci quantitation for high-throughput analysis. J Biol Methods 2. 24. [↵][94]Leitner, A., Faini, M., Stengel, F., and Aebersold, R. (2016). Crosslinking and Mass Spectrometry: An Integrated Technology to Understand the Structure and Function of Molecular Machines. Trends Biochem Sci 41, 20–32. 25. [↵][95]Liu, F., Rijkers, D.T., Post, H., and Heck, A.J. (2015). Proteome-wide profiling of protein assemblies by cross-linking mass spectrometry. Nat Methods 12, 1179–1184. 26. [↵][96]Liu, Q., Li, H., Yang, Y., Tian, X., Su, J., Zhou, L., and Liu, Q. (2017). A disulfide-bonded DnaK dimer is maintained in an ATP-bound state. Cell Stress Chaperones 22, 201–212. 27. [↵][97]Mashaghi, A., Bezrukavnikov, S., Minde, D.P., Wentink, A.S., Kityk, R., Zachmann-Brand, B., Mayer, M.P., Kramer, G., Bukau, B., and Tans, S.J. (2016). Alternative modes of client binding enable functional plasticity of Hsp70. Nature 539, 448-451. 28. [↵][98]Mazzoni, C., Palermo, V., Torella, M., and Falcone, C. (2005). HIR1, the co-repressor of histone gene transcription of Saccharomyces cerevisiae, acts as a multicopy suppressor of the apoptotic phenotypes of the LSM4 mRNA degradation mutant. FEMS Yeast Res 5, 1229–1235. 29. [↵][99]Millson, S.H., Truman, A.W., King, V., Prodromou, C., Pearl, L.H., and Piper, P.W. (2005). A two-hybrid screen of the yeast proteome for Hsp90 interactors uncovers a novel Hsp90 chaperone requirement in the activity of a stress-activated mitogen-activated protein kinase, Slt2p (Mpk1p). Eukaryot Cell 4, 849–860. 30. [↵][100]Morgner, N., Schmidt, C., Beilsten-Edmands, V., Ebong, I.O., Patel, N.A., Clerico, E.M., Kirschke, E., Daturpalli, S., Jackson, S.E., Agard, D., et al. (2015). Hsp70 forms antiparallel dimers stabilized by post-translational modifications to position clients for transfer to Hsp90. Cell Rep 11, 759–769. 31. [↵][101]Nishimura, A., Yamamoto, K., Oyama, M., Kozuka-Hata, H., Saito, H., and Tatebayashi, K. (2016). Scaffold Protein Ahk1, Which Associates with Hkr1, Sho1, Ste11, and Pbs2, Inhibits Cross Talk Signaling from the Hkr1 Osmosensor to the Kss1 Mitogen-Activated Protein Kinase. Mol Cell Biol 36, 1109–1123. 32. [↵][102]Nitika, Porter, C.M., Truman, A.W., and Truttmann, M.C. (2020). Post-translational modifications of Hsp70 family proteins: Expanding the chaperone code. J Biol Chem 295, 10689–10708. 33. [↵][103]Nitika, and Truman, A.W. (2017). Cracking the Chaperone Code: Cellular Roles for Hsp70 Phosphorylation. Trends Biochem Sci 42, 932–935. 34. [↵][104]O’Regan, L., Sampson, J., Richards, M.W., Knebel, A., Roth, D., Hood, F.E., Straube, A., Royle, S.J., Bayliss, R., and Fry, A.M. (2015). Hsp72 is targeted to the mitotic spindle by Nek6 to promote K-fiber assembly and mitotic progression. J Cell Biol 209, 349–358. 35. [↵][105]Ondrovicova, G., Liu, T., Singh, K., Tian, B., Li, H., Gakh, O., Perecko, D., Janata, J., Granot, Z., Orly, J., et al. (2005). Cleavage site selection within a folded substrate by the ATP-dependent lon protease. J Biol Chem 280, 25103–25110. 36. [↵][106]Petrovic, A., Pasqualato, S., Dube, P., Krenn, V., Santaguida, S., Cittaro, D., Monzani, S., Massimiliano, L., Keller, J., Tarricone, A., et al. (2010). The MIS12 complex is a protein interaction hub for outer kinetochore assembly. J Cell Biol 190, 835–852. 37. [↵][107]Piper, P.W., Truman, A.W., Millson, S.H., and Nuttall, J. (2006). Hsp90 chaperone control over transcriptional regulation by the yeast Slt2(Mpk1)p and human ERK5 mitogen-activated protein kinases (MAPKs). Biochem Soc Trans 34, 783–785. 38. [↵][108]Radons, J. (2016). The human HSP70 family of chaperones: where do we stand? Cell Stress Chaperones 21, 379–404. 39. [↵][109]Rosenzweig, R., Nillegoda, N.B., Mayer, M.P., and Bukau, B. (2019a). The Hsp70 chaperone network. Nat Rev Mol Cell Biol. 40. [↵][110]Rosenzweig, R., Nillegoda, N.B., Mayer, M.P., and Bukau, B. (2019b). The Hsp70 chaperone network. Nat Rev Mol Cell Biol 20, 665–680. 41. [↵][111]Ruan, L., McNamara, J.T., Zhang, X., Chang, A.C., Zhu, J., Dong, Y., Sun, G., Peterson, A., Na, C.H., and Li, R. (2020). Solid-phase inclusion as a mechanism for regulating unfolded proteins in the mitochondrial matrix. Sci Adv 6, eabc7288. 42. [↵][112]Ruan, L., Zhou, C., Jin, E., Kucharavy, A., Zhang, Y., Wen, Z., Florens, L., and Li, R. (2017). Cytosolic proteostasis through importing of misfolded proteins into mitochondria. Nature 543, 443–446. 43. [↵][113]Sarbeng, E.B., Liu, Q., Tian, X., Yang, J., Li, H., Wong, J.L., Zhou, L., and Liu, Q. (2015). A functional DnaK dimer is essential for the efficient interaction with Hsp40 heat shock protein. J Biol Chem 290, 8849–8862. 44. [↵][114]Sharp, J.A., Rizki, G., and Kaufman, P.D. (2005). Regulation of histone deposition proteins Asf1/Hir1 by multiple DNA damage checkpoint kinases in Saccharomyces cerevisiae. Genetics 171, 885–899. 45. [↵][115]Swaney, D.L., Beltrao, P., Starita, L., Guo, A., Rush, J., Fields, S., Krogan, N.J., and Villen, J. (2013). Global analysis of phosphorylation and ubiquitylation cross-talk in protein degradation. Nature methods 10, 676–682. 46. [↵][116]Taipale, M. (2018). Quantitative Profiling of Chaperone/Client Interactions with LUMIER Assay. Methods Mol Biol 1709, 47–58. 47. [↵][117]Taipale, M., Tucker, G., Peng, J., Krykbaeva, I., Lin, Z.Y., Larsen, B., Choi, H., Berger, B., Gingras, A.C., and Lindquist, S. (2014). A quantitative chaperone interaction network reveals the architecture of cellular protein homeostasis pathways. Cell 158, 434–448. 48. [↵][118]Takakuwa, J.E., Nitika, Knighton, L.E., and Truman, A.W. (2019). Oligomerization of Hsp70: Current Perspectives on Regulation and Function. Front Mol Biosci 6, 81. 49. [↵][119]Tao, Y., Messer, J.S., Goss, K.H., Hart, J., Bissonnette, M., and Chang, E.B. (2016). Hsp70 exerts oncogenic activity in the Apc mutant Min mouse model. Carcinogenesis 37, 731–739. 50. [↵][120]Tatebayashi, K., Yamamoto, K., Tanaka, K., Tomida, T., Maruoka, T., Kasukawa, E., and Saito, H. (2006). Adaptor functions of Cdc42, Ste50, and Sho1 in the yeast osmoregulatory HOG MAPK pathway. EMBO J 25, 3033–3044. 51. [↵][121]Tatebayashi, K., Yamamoto, K., Tomida, T., Nishimura, A., Takayama, T., Oyama, M., Kozuka-Hata, H., Adachi-Akahane, S., Tokunaga, Y., and Saito, H. (2020). Osmostress enhances activating phosphorylation of Hog1 MAP kinase by mono-phosphorylated Pbs2 MAP2K. EMBO J 39, e103444. 52. [↵][122]Trcka, F., Durech, M., Vankova, P., Chmelik, J., Martinkova, V., Hausner, J., Kadek, A., Marcoux, J., Klumpler, T., Vojtesek, B., et al. (2019). Human Stress-inducible Hsp70 Has a High Propensity to Form ATP-dependent Antiparallel Dimers That Are Differentially Regulated by Cochaperone Binding. Molecular & cellular proteomics : MCP 18, 320–337. 53. [↵][123]Truman, A.W., Kristjansdottir, K., Wolfgeher, D., Hasin, N., Polier, S., Zhang, H., Perrett, S., Prodromou, C., Jones, G.W., and Kron, S.J. (2012). CDK-dependent Hsp70 Phosphorylation controls G1 cyclin abundance and cell-cycle progression. Cell 151, 1308-1318. 54. [↵][124]Truman, A.W., Kristjansdottir, K., Wolfgeher, D., Ricco, N., Mayampurath, A., Volchenboum, S.L., Clotet, J., and Kron, S.J. (2015). Quantitative proteomics of the yeast Hsp70/Hsp90 interactomes during DNA damage reveal chaperone-dependent regulation of ribonucleotide reductase. Journal of proteomics 112, 285–300. 55. [↵][125]Van Dyck, L., Pearce, D.A., and Sherman, F. (1994). PIM1 encodes a mitochondrial ATP-dependent protease that is required for mitochondrial function in the yeast Saccharomyces cerevisiae. J Biol Chem 269, 238–242. 56. [↵][126]Vidal, M., Brachmann, R.K., Fattaey, A., Harlow, E., and Boeke, J.D. (1996). Reverse two-hybrid and one-hybrid systems to detect dissociation of protein-protein and DNA-protein interactions. Proc Natl Acad Sci U S A 93, 10315–10320. 57. [↵][127]Yang, Y., Fiskus, W., Yong, B., Atadja, P., Takahashi, Y., Pandita, T.K., Wang, H.G., and Bhalla, K.N. (2013). Acetylated hsp70 and KAP1-mediated Vps34 SUMOylation is required for autophagosome creation in autophagy. Proc Natl Acad Sci U S A 110, 6841–6846. 58. [↵][128]Yugandhar, K., Gupta, S., and Yu, H. (2019). Inferring Protein-Protein Interaction Networks From Mass Spectrometry-Based Proteomic Approaches: A Mini-Review. Comput Struct Biotechnol J 17, 805–811. 59. [↵][129]Zhao, R., Davey, M., Hsu, Y.C., Kaplanek, P., Tong, A., Parsons, A.B., Krogan, N., Cagney, G., Mai, D., Greenblatt, J., et al. (2005). Navigating the chaperone network: an integrative map of physical and genetic interactions mediated by the hsp90 chaperone. Cell 120, 715–727. 60. [↵][130]Zhu, X., Zhao, X., Burkholder, W.F., Gragerov, A., Ogata, C.M., Gottesman, M.E., and Hendrickson, W.A. (1996). Structural analysis of substrate binding by the molecular chaperone DnaK. Science 272, 1606–1614. [1]: #ref-18 [2]: #ref-39 [3]: #ref-14 [4]: #ref-38 [5]: #ref-9 [6]: #ref-40 [7]: #ref-1 [8]: #ref-5 [9]: #ref-33 [10]: #ref-15 [11]: #ref-29 [12]: #ref-54 [13]: #ref-56 [14]: #ref-59 [15]: #ref-24 [16]: #ref-25 [17]: #ref-12 [18]: #ref-16 [19]: #ref-53 [20]: #ref-57 [21]: #ref-32 [22]: #ref-2 [23]: #ref-6 [24]: #ref-13 [25]: #ref-45 [26]: #F1 [27]: pending:yes [28]: #F2 [29]: #ref-43 [30]: #F3 [31]: #F8 [32]: #ref-28 [33]: #ref-44 [34]: #F4 [35]: #ref-55 [36]: #F5 [37]: #ref-41 [38]: #ref-42 [39]: #F9 [40]: #ref-35 [41]: #F6 [42]: #ref-23 [43]: #ref-36 [44]: #ref-11 [45]: #ref-19 [46]: #ref-20 [47]: #ref-17 [48]: #ref-31 [49]: #ref-50 [50]: #ref-51 [51]: #F7 [52]: #ref-4 [53]: #ref-22 [54]: #ref-58 [55]: #ref-46 [56]: #ref-47 [57]: #ref-7 [58]: #ref-27 [59]: #ref-3 [60]: #ref-60 [61]: #ref-26 [62]: #ref-30 [63]: #ref-48 [64]: #ref-52 [65]: #ref-8 [66]: #ref-34 [67]: #ref-49 [68]: #ref-21 [69]: #ref-37 [70]: #ref-10 [71]: #xref-ref-1-1 "View reference in text" [72]: #xref-ref-2-1 "View reference in text" [73]: #xref-ref-3-1 "View reference in text" [74]: #xref-ref-4-1 "View reference in text" [75]: #xref-ref-5-1 "View reference in text" [76]: #xref-ref-6-1 "View reference in text" [77]: #xref-ref-7-1 "View reference in text" [78]: #xref-ref-8-1 "View reference in text" [79]: #xref-ref-9-1 "View reference in text" [80]: #xref-ref-10-1 "View reference in text" [81]: #xref-ref-11-1 "View reference in text" [82]: #xref-ref-12-1 "View reference in text" [83]: #xref-ref-13-1 "View reference in text" [84]: #xref-ref-14-1 "View reference in text" [85]: #xref-ref-15-1 "View reference in text" [86]: #xref-ref-16-1 "View reference in text" [87]: #xref-ref-17-1 "View reference in text" [88]: #xref-ref-18-1 "View reference in text" [89]: #xref-ref-19-1 "View reference in text" [90]: #xref-ref-20-1 "View reference in text" [91]: #xref-ref-21-1 "View reference in text" [92]: #xref-ref-22-1 "View reference in text" [93]: #xref-ref-23-1 "View reference in text" [94]: #xref-ref-24-1 "View reference in text" [95]: #xref-ref-25-1 "View reference in text" [96]: #xref-ref-26-1 "View reference in text" [97]: #xref-ref-27-1 "View reference in text" [98]: #xref-ref-28-1 "View reference in text" [99]: #xref-ref-29-1 "View reference in text" [100]: #xref-ref-30-1 "View reference in text" [101]: #xref-ref-31-1 "View reference in text" [102]: #xref-ref-32-1 "View reference in text" [103]: #xref-ref-33-1 "View reference in text" [104]: #xref-ref-34-1 "View reference in text" [105]: #xref-ref-35-1 "View reference in text" [106]: #xref-ref-36-1 "View reference in text" [107]: #xref-ref-37-1 "View reference in text" [108]: #xref-ref-38-1 "View reference in text" [109]: #xref-ref-39-1 "View reference in text" [110]: #xref-ref-40-1 "View reference in text" [111]: #xref-ref-41-1 "View reference in text" [112]: #xref-ref-42-1 "View reference in text" [113]: #xref-ref-43-1 "View reference in text" [114]: #xref-ref-44-1 "View reference in text" [115]: #xref-ref-45-1 "View reference in text" [116]: #xref-ref-46-1 "View reference in text" [117]: #xref-ref-47-1 "View reference in text" [118]: #xref-ref-48-1 "View reference in text" [119]: #xref-ref-49-1 "View reference in text" [120]: #xref-ref-50-1 "View reference in text" [121]: #xref-ref-51-1 "View reference in text" [122]: #xref-ref-52-1 "View reference in text" [123]: #xref-ref-53-1 "View reference in text" [124]: #xref-ref-54-1 "View reference in text" [125]: #xref-ref-55-1 "View reference in text" [126]: #xref-ref-56-1 "View reference in text" [127]: #xref-ref-57-1 "View reference in text" [128]: #xref-ref-58-1 "View reference in text" [129]: #xref-ref-59-1 "View reference in text" [130]: #xref-ref-60-1 "View reference in text"

Neuromuscular ageing is characterized by neural and/or skeletal muscle degeneration that decreases maximal force and power. Female neuromuscular ageing occurs earlier in life compared to males, potentially due to sex hormone changes during the menopausal transition. We quantified neuromuscular function in 88 healthy females represented equally over each decade from 18-80 years of age and investigated the potential role of decreased ovarian hormone concentrations following menopause. Neuromuscular assessment included quadriceps maximal voluntary and evoked isometric torque and surface electromyography measurements, plus one-repetition maximum leg press. Voluntary and evoked torques and one-repetition maximum decreased non-linearly with age, with accelerated reductions starting during the fourth decade. An absence of changes in volitional recruitment of existing quadriceps motor units and Ia afferent facilitation of spinal motoneurons suggests that functional decline was largely mediated by impairment in intrinsic muscle function and/or neuromuscular transmission. Maximal muscle compound action potential amplitude decreased with increasing age for rectus femoris only, indicating increased vulnerability to neuromuscular degeneration compared to vastus lateralis and medialis. In postmenopausal females, some variance in data was explained by inter-individual differences in body composition and physical activity level, however, changes in total or free concentrations of oestrogen, progesterone and/or testosterone were correlated with all age-related decreases in neuromuscular variables. In conclusion, we demonstrate an accelerated onset of neuromuscular degeneration of muscular origin around menopause onset, which is associated with changes in sex hormone concentrations. Interventions aimed at mitigating declines in ovarian hormones and their subsequent effects on neuromuscular function postmenopause should be further explored. ### Competing Interest Statement The authors have declared no competing interest.

Androgens are vital for the maintenance of muscle mass and their anabolic effects are primarily exerted through the androgen receptor (AR). Accumulating evidence in humans and mice suggests that circulating androgens, AR and androgen response are influenced by ACTN3 ( α- actinin-3), also known as “the gene for speed”. One in 5 people worldwide are α-actinin-3 deficient due to homozygous inheritance of a common null polymorphism (577X) in ACTN3 . In this study, we show that α-actinin-3 deficiency decreases baseline AR in skeletal muscles of mice and humans, in both males and females, and that AR expression directly correlates with ACTN3 in a dosage dependent manner. We further demonstrate in Actn3 knockout mice that α- actinin-3 deficiency increases muscle wasting induced by androgen deprivation and reduces the muscle hypertrophic response to dihydrotestosterone and this is mediated by differential activation of pathways regulating amino acid metabolism, intracellular transport, MAPK signalling, autophagy, mitochondrial activity and calcineurin signalling. Gene set enrichment and protein analyses indicate that the absence of α-actinin-3 results in a failure to coactivate many of these pathways in response to changes in androgens, and relies on leveraging mitochondrial remodelling and calcineurin signalling to restore muscle homeostasis. We further identified 7 genes that are androgen sensitive and α-actinin-3-dependent in expression, and whose functions correspond to these processes. Our results highlight the pivotal role of α- actinin-3 in various processes associated with the regulation of protein turnover and muscle mass, and suggest that ACTN3 genotype is a genetic modifier of androgen action in skeletal muscle.### Competing Interest StatementF.J.R. receives institutional support as a coinvestigator and subcontracted by the Peter MacCallum Cancer Centre for an investigator-initiated trial which receives funding support from Sanofi/Regeneron Pharmaceuticals.