Evolutionary conservation of systemic and reversible amyloid aggregation

Eukaryotic cells must tightly regulate the synthesis, folding and degradation of proteins, as the disruption of these processes can dysregulate proteostasis (Cox and Walter, 1996; Harding et al., 2002; Wickner et al., 1999). A growing wealth of literature suggests that the impairment of protein dynamics is not exclusively associated with genetic abnormalities in the regulatory machinery, but can also be driven by changing cellular conditions, as cells are exposed to suboptimal growth environments (Carmeliet et al., 1998; Diamant and Goloubinoff, 1998; Escusa-Toret et al., 2013; Okamura et al., 2000). Elevated temperature is a particularly damaging stressor, as it can globally weaken hydrogen bonds and electrostatic interactions that are necessary to maintain a native protein conformation (Martin et al., 1992; Mine et al., 1990; Nakamura et al., 1978). This often results in the accumulation of misfolded polypeptides, which can coalesce into either amorphous aggregates or highly organized amyloid-based structures (Jahn and Radford, 2005). The amyloid fold is a conformation rich in β-sheets that allows proteins to stack into stable and unbranching fibrils (Jahn and Radford, 2005; Knowles et al., 2014). Historically, this state of protein organization has been viewed as toxic, and is prominently associated with debilitating neuropathies, such as Alzheimer's, Parkinson's and prion diseases (Rambaran and Serpell, 2008; Stefani and Dobson, 2003). What is less well established is the functional role of amyloid-based aggregation in physiological pathways. To date, the observation of functional amyloids in humans is rare, with most work citing a limited number of proteins in very specific cellular settings (Fowler et al., 2005; Li et al., 2012; Maji et al., 2009). Whether amyloid aggregation acts as a natural and ubiquitous component of proteostasis remains an outstanding question in cellular biology.

Recently, widespread and functional amyloid aggregates have been observed in cultured human cell lines that were subjected to environmental stressors including heat shock, extracellular acidosis, and transcriptional/proteotoxic stress (Audas et al., 2016). Upon exposure, a diverse array of cellular proteins were found to be targeted to nucleoli (Audas et al., 2016; Marijan et al., 2019), where they rapidly transition from a liquid phase to a solid amyloid-like state (Audas et al., 2016; Wang et al., 2018). This cellular process impairs ribogenesis, disrupts nucleolar organization, and results in the formation of a new subnuclear domain, termed the amyloid body (A-body) (Audas et al., 2016; Jacob et al., 2013). Within this structure, targeted proteins adopt amyloid biophysical characteristics, as sequestration is accompanied by protein immobilization, insolubilization, fibrillation and an affinity for amyloidophilic dyes (such as Congo Red and Thioflavin S) (Audas et al., 2016; Mekhail et al., 2005). Regulation of A-body formation is dependent on a family of low-complexity long noncoding RNA (lncRNA) transcripts that are expressed from the intergenic spacer regions of the rDNA cassettes (Audas et al., 2012; Wang et al., 2018). Each stressor upregulates one or more of these ribosomal intergenic spacer lncRNAs (rIGSRNA), to drive the recruitment of a large family of stimuli-specific A-body constituents (Audas et al., 2016, 2012; Jacob et al., 2012; Marijan et al., 2019). The early stages of this process appear to mimic the architectural RNA-mediated recruitment of proteins to other liquid-like biomolecular condensates, such as stress granules, P-bodies, nuclear paraspeckles and omega speckles (Chujo et al., 2016; Fox et al., 2002; Kedersha et al., 1999; Liu et al., 2005; Prasanth et al., 2000). However, A-bodies quickly solidify the initial liquid-phase structures (Wang et al., 2018), although the mechanism regulating this biological transition remains unknown. Functionally, A-body formation has been shown to induce a global state of cellular dormancy, reducing energy demand during unfavorable growth conditions (Audas et al., 2016). This is potentially through the sequestration and inactivation of key cellular regulators including cell division cycle protein 73 homolog (CDC73), cyclin-dependent kinase 1, and the delta catalytic subunit of DNA polymerase. Local nuclear protein synthesis (LNPS) has also been shown to occur within the A-bodies, as mRNAs induced by the transcription factor heat shock factor protein 1 (HSF-1) are recruited and synthesized on ribosomes bound to the amyloid fibrils (Theodoridis et al., 2021). These data demonstrate a biochemically active role for these amyloid-based structures, and further highlights the unconventional nature of these nuclear foci. Additionally, returning cells to physiological growth conditions leads to the rapid disassembly of A-bodies, where constituent proteins return to their native conformation/localization, cell proliferation is restored and LNPS is halted (Audas et al., 2016; Theodoridis et al., 2021). Biophysically, this is a remarkable event, as the formation of pathological amyloid-based structures has not been shown to be reversible (Mukherjee et al., 2015; Ross and Poirier, 2004; Stefani and Dobson, 2003). Therefore, studying this physiological example of the widespread, rapid and reversible adoption of an amyloid conformation could provide key insights into the maintenance of proteostasis and protein folding diseases.

To date, functional amyloids have been observed in several organismal settings. Bacteria biofilms (Epstein and Chapman, 2008), silk moth eggshells (Iconomidou et al., 2000) and yeast protein-based inheritance (Kochneva-Pervukhova et al., 1998; Liebman and Chernoff, 2012) have each been linked to polypeptides adopting the amyloid conformation. As A-body formation has only been examined in cultured human cell lines, we sought to determine whether this structure was evolutionarily conserved across a variety of eukaryotic species (e.g. mammals, birds, fish, flies and yeast). Our data supports the notion that A-bodies are well conserved, and suggests that their formation is an important adaptive pathway that may aid in our understanding of stress response mechanisms and physiological/pathological amyloid aggregation events.

A-bodies were discovered in cultured human cells, and form in response to stressful stimuli such as elevated temperature, extracellular acidification, and impaired transcription/proteostasis (Audas et al., 2016). As environmental perturbation is a universal obstacle encountered by virtually all species, we sought to investigate whether these functional amyloid aggregates were conserved in other eukaryotic settings. Using established treatment parameters and the biophysical characteristics of A-bodies as a guide (Audas et al., 2016), we assessed a panel of mammalian cell lines derived from human (MCF-7), African green monkey (COS-1), rat (A7R5) and mouse (Swiss 3T3) tissue. Our data demonstrated that each of these cell lines could form subnuclear protein aggregates that stained positively with the amyloidophilic dyes Congo Red and Thioflavin S, under the same conditions previously shown to induce A-bodies in human cells [heat shock of 43°C; extracellular acidosis of pH 6.0+1% O₂; transcriptional/proteotoxic stress (TPS) of 4 µM actinomycin D and 8 µM MG132] (Fig. 1A). To further validate these structures as A-bodies, and not another subnuclear domain, we assessed the localization, mobility and solubility of CDC73, a protein known to be targeted to human A-bodies in response to all established stressors (Marijan et al., 2019). As expected, endogenous CDC73 shifted from a diffuse nuclear localization in untreated cells into distinct subnuclear foci when exposed to heat-shock conditions (Fig. 1B). This change in CDC73 localization was accompanied by the hallmark losses of mobility (Fig. 1C) and solubility (Fig. 1D), which are seen when proteins aggregate within amyloid structures (Audas et al., 2016). Conversely, the control proteins GAPDH and histone H3, which are not targeted to A-bodies, retained their solubility and mobility profiles in these mammalian cell lines (Fig. 1C,D). Finally, as A-bodies disassemble in human cells upon stimuli termination, we examined whether CDC73 reverted to a diffuse nuclear localization when heat-shock-treated cells were returned to normal growth conditions (37°C). After a 24-h recovery period, CDC73 was no longer present within subnuclear foci, demonstrating the disaggregation of these subnuclear domains (Fig. 1B). Together, these data suggest that A-bodies are a conserved stress-inducible structure in a panel (human, monkey, rat and mouse) of mammalian cell lines.

Adaptive stress response pathways are designed to help cells maintain homeostasis when conditions deviate from a standard growth environment (Carmeliet et al., 1998; Morimoto et al., 1992). However, different organisms can have highly variable physiological norms, which would require cells to calibrate stress response induction thresholds. With this in mind, we looked for A-body formation in embryonic chicken cells (DF-1), as the resting body temperature of chickens (39°C) is higher than most mammals (37°C) (Mezquita et al., 1998), and chicken embryos are less-responsive to extracellular pH fluctuations (Shartau et al., 2016). Despite the established TPS-treatment parameters inducing A-bodies efficiently in DF-1 cells, the mammalian thermal (43°C) and extracellular acidity (pH 6.0) thresholds did not significantly induce the formation of Congo Red-positive subnuclear foci (Fig. 2A–C). Instead, incubation of DF-1 cells in a 46°C or a pH 5.2 environment was necessary to generate putative A-bodies of similar intensity to those seen in the mammalian cell lines (Fig. 2A,B). Using these species-specific conditions, we validated cell viability (Fig. 2D) during the severe treatments, and demonstrated the insolubilization (Fig. 2E), reversible targeting (Fig. 2F) and immobilization (Fig. 2G) of CDC73 within these amyloid structures. Additionally, fibrillarin co-staining (Fig. 2F, inset) showed that these subnuclear condensates occupied the nucleolar space, which is another hallmark of the A-body (Audas et al., 2016; Jacob et al., 2013).

Having demonstrated that A-body induction thresholds can vary depending on the adaptive needs of an organism, we further examined the activation and conservation of this amyloid aggregation pathway in a more evolutionarily distant setting, the ectothermic Bluegill sunfish cell line (BF-2). Like many fish species, the Bluegill displays remarkable adaptive ability, as they can be found in waters with temperatures ranging from 10°C–30°C and a pH of 6.5–8.5 (Jones et al., 2008; Stuber et al., 1982). We exposed these fish cells to a thermal and pH gradient and found that a temperature of 38.5°C or extracellular pH of 5.2 was required to induce subnuclear amyloid aggregation (Fig. 3A,B). It is interesting to note that the acidotic and TPS conditions necessary for BF-2 and DF-1 cells to generate amyloid aggregates were the same (Figs 2B,C, 3B,C), despite the marked divergence in the thermal thresholds (Figs 2A and 3A). Despite these severe stress parameters, propidium iodide staining revealed no significant loss in viability under these conditions (Fig. 3D). Using the hallmarks employed to validate these structures as A-bodies in the mammalian and avian settings as a guide (Figs 1 and 2), the BF-2 stress-inducible subnuclear condensates were confirmed to be amyloid-like in nature (i.e. Congo Red- and Thioflavin S-positive) (Fig. 3C) and capable of immobilizing (Fig. 3E), insolubilizing (Fig. 3F) and reversibly targeting exogenously expressed CDC73 to the nucleolar space (Fig. 3G). Here, we used human CDC73, as the sequence of the Bluegill homolog is not available, and the protein is not cross-reactive with our antibodies. Together, this data suggests that A-bodies are evolutionarily conserved in chickens and fish; however, the precise conditions that cause these amyloid foci to form are not universal among species, but rather tailored to the environmental norms of the divergent organisms.

To this point, our study of A-bodies has been limited to vertebrate species. Therefore, we sought to explore the formation of these amyloid aggregates in lower eukaryotes. We first examined Drosophila melanogaster, by incubating the embryonic cell line S2 under increasingly severe conditions. Here, Congo Red-positive structures were observed upon exposure to 38.5°C temperatures or an extracellular pH of 5.2 (Fig. 4A–C). Unlike what was seen in the mammal, avian and fish settings, concentrations of actinomycin D (1 µM) and MG132 (50 µM) needed to be adjusted to activate TPS-mediated formation of these putative A-bodies (Fig. 4A), likely owing to the different concentrations needed to elicit a similar transcriptional and proteotoxic stress response in fruit flies (Bönisch et al., 2007; Velentzas et al., 2011). As above, our data indicates that S2 cells can survive these extreme conditions (Fig. 4D), further suggesting that this form of amyloid-based aggregation represents a response pathway for severe stress exposure. To confirm that these foci are evolutionarily related to A-bodies, we co-stained with the amyloidophilic dyes Congo Red and Thioflavin S (Fig. 4A) and detected the reversible targeting of endogenous Hyrax, the D. melanogaster homolog of CDC73, within subnuclear foci that colocalize with fibrillarin (Fig. 4E). Hyrax/CDC73 was also shown to insolubilize and immobilize within these aggregates (Fig. 4F,G). Given the large evolutionary gap, we felt it was prudent to evaluate whether these amyloid-based structures colocalized with other subnuclear domains in D. melanogaster. Here, we focused our analysis on omega speckles (Prasanth et al., 2000) and nucleolar aggresomes (Latonen et al., 2011), as these biomolecular condensates are either induced under similar conditions or occupy a similar localization within the nucleolus, respectively. Omega speckle (Squid and NonA) and nucleolar aggresome (CDK2 and CDK4) marker proteins were exogenously expressed, and their localization was assessed. Squid- and NonA-containing foci failed to colocalize with Congo Red-positive aggregates (Fig. 4H) or immobilize the marker protein under these severe heat shock conditions (Fig. 4G), while the nucleolar aggresome residents CDK2 and CDK4 were not targeted to a subnuclear structure, indicating the absence of nucleolar aggresomes altogether under the conditions tested (Fig. 4H). This data increases our confidence that these amyloidogenic foci are evolutionarily related to A-bodies, and distinct from other subnuclear domains.

To date, all previous work has relied on the examination of A-bodies in a cultured cell setting (Audas et al., 2012, 2016; Jacob et al., 2013; Wang et al., 2018; Marijan et al., 2019; Theodoridis et al., 2021); thus, we wanted to see whether these structures could form within living organisms. Expanding on our results in the S2 cell line, we examined adult and embryo D. melanogaster that were incubated at 38.5°C, prior to dissection, fixation and Congo Red staining. Our data shows that a short 30-min heat shock treatment was sufficient to induce the formation of putative A-bodies within the egg chambers of adult flies (Fig. 5A), though a slightly longer 90-min exposure was necessary to generate these structures in the epidermis and amnioserosa of the embryo (Fig. 5B). Next, we chose to explore whether A-bodies were conserved in the single-celled eukaryotic organism Saccharomyces cerevisiae. Yeast cultures were exposed to increasingly severe heat-shock conditions and observed for the formation of aggregates with amyloid-like properties. Here, a severe heat-shock treatment of 45°C was required to induce the formation of Congo Red-positive structures within S. cerevisiae (Fig. 5C). These foci could also be generated by reducing the extracellular pH to 3 (Fig. 5C), another common A-body-inducing stimulus. Co-staining of treated yeast with Congo Red and Thioflavin S validated the amyloid-like nature of these aggregates (Fig. 5D), and colocalization with endogenous fibrillarin indicates these structures have formed within the nucleolar region (Fig. 5E). These data combined with the disassembly of these structures following a return to normal growth conditions (Fig. 5F) suggests that these amyloid aggregates are distant relatives of the mammalian A-body. As S. cerevisiae remained viable in response to these extreme/harsh conditions (Fig. 5G), we speculate that A-body formation is a natural adaptive response across the eukaryotic domain that is activated in response to severe, yet sublethal, environmental insults. These findings are also notable as they represent the first examples of putative A-body formation within the tissue or cells of a living organism.

A-body formation is seeded by stress-inducible low-complexity lncRNA derived from the intergenic spacer region of the rDNA cassette (Audas et al., 2012; Wang et al., 2018). While this region is poorly annotated (rat, Bluegill sunfish) or highly repetitive (D. melanogaster) in some species, the rDNA sequences of humans, mice and chickens is available (Fig. 6A). Therefore, we performed RNA-sequencing and quantitative PCR (qPCR) analysis to identify heat-shock-inducible transcripts within this region. As previously published (Audas et al., 2012; Wang et al., 2018), human cells significantly upregulated transcripts from a region ∼16 kilobases (kb) downstream of the rDNA transcriptional start site (rIGS₁₆RNA) in the early stages of thermal stress (Fig. 6B,C). In mice, a strongly expressed heat-shock-inducible transcript was also observed from a similarly positioned region, ∼14 kb downstream of the rDNA start site (Fig. 6D,E), whereas chicken cells expressed lncRNA from multiple regions, including the ∼12 kb locus of the rDNA cassette (Fig. 6F,G).

An analysis of the heat-shock-inducible transcripts derived from the region directly downstream of the rRNA 3′ external transcribed spacers showed that each species possessed similar areas of low complexity (Fig. 7A), a characteristic that has previously been shown to be necessary for inducing A-body formation in human cells (Wang et al., 2018). Using siRNA, we inhibited the human (rIGS₁₆RNA), mouse (mrIGS₁₄RNA), or chicken (crIGS₁₂RNA) transcripts and assessed for A-body formation. Each of these siRNA delayed or impaired A-body aggregation as detected by Congo Red and CDC73 staining (Fig. 7B,C). Together, these data highlight the conservation of the low complexity rIGSRNAs and their role in A-body formation.

Finally, it has recently been shown that A-bodies are a site of nuclear protein synthesis (Theodoridis et al., 2021). Therefore, we examined whether this unconventional biological function was conserved across the panel of species we tested. As previously established (Theodoridis et al., 2021), puromycin-tagged proteins present within the A-bodies are derived from localized nuclear translation events, and not cytoplasmically synthesized polypeptides that have been released by the ribosome and transported to the nucleus (Hobson et al., 2020). Here, we incubated the cells/organisms with puromycin and detected the localization of the puromycin-incorporated polypeptides. Like the human cell line MCF-7, chicken- (DF-1) and bluegill sunfish-derived (BF-2) cells had newly synthesized puromycin-positive proteins within the A-bodies (Fig. 7D), with this nuclear translation event being markedly impaired by siRNA targeting the species-specific low-complexity rIGSRNAs (Fig. 7E). Interestingly, S2 cells (D. melanogaster) and S. cerevisiae did not display active translation within the A-bodies in response to prolonged heat exposure (Fig. 7D), suggesting that A-body-mediated local nuclear protein synthesis may have evolved in upper eukaryotes.

In humans, the amyloid state has historically been associated with disease (Wang et al., 2017). However, the discovery of A-bodies (Audas et al., 2016) and other physiological amyloid aggregation events (Matiiv et al., 2020) has led to a changing perception of this protein conformation. Functional amyloids have been observed in many organisms (Maury, 2009), yet the conservation of a particular amyloid aggregate across a variety of species has not been rigorously investigated. In the present study, we assessed whether A-bodies were a conserved structure within a cross-section of evolutionarily distinct organisms. Our data shows that the species tested are capable of forming subnuclear foci, which shared the same amyloid-like biophysical characteristics ascribed to human A-bodies. The conservation of a pathway that allows cells to rapidly, inducibly and reversibly harness the amyloid state highlights the utility and significance of this adaptive mechanism.

While some elements of A-body biology remain constant from yeast to human, the conditions needed to induce amyloid aggregation were highly variable. It is well established that organisms evolve to thrive within their environmental niches, and this would necessitate species-specific changes to various stress-induction thresholds. Despite these expected differences, our threshold analyses provide some insight into the mechanisms controlling functional amyloidogenesis. Fish and chicken A-bodies were shown to form when cells were exposed to the same level of extracellular acidity (pH 5.2), while the thermal thresholds were markedly different (38.5°C and 46°C) (Figs 2 and 3). This may imply that components of the heat-shock and acidotic amyloid aggregation signaling networks evolved independently of each other and likely use an array of distinct regulators. It could also suggest that acidotic sensory molecules possess a higher level of sequence conservation, relative to the heat-shock regulators, within these species. Mechanistically, differences in thermal induction conditions can occur through mutations in biomolecules, such as RNA ‘thermometers’. Here, variations in the nucleotide sequence of an lncRNA could strengthen or weaken base pairing within the secondary structure, tailoring the activation or inactivation of this biomolecule to a specific temperature (Giuliodori et al., 2010; Johansson et al., 2002; Morita et al., 1999). The rIGSRNAs themselves seem unlikely to employ this mechanism, as a consistent theme for these transcripts is the presence of low complexity sequences (Fig. 7A), which are critical for their functionality (Wang et al., 2018). Since these regions are not predicted to adopt any secondary structure, their very nature makes them incompatible with the basic principles of being an RNA thermometer. However, other regulatory lncRNA transcripts or even thermo-sensitive proteins could be controlling the expression of the rIGSRNAs and/or other elements of the A-body induction pathway. We believe that as new regulators of A-body formation are discovered, comparing their homologs in the species assessed here will enhance our understanding of this physiological pathway.

The work presented here focuses on the evolutionary conservation of A-bodies and does not delve deeply into the biological function of these subnuclear structures. Based on human proteomic studies, >150 proteins have been detected in acidotic- (Audas et al., 2016) and heat-shock-induced (Marijan et al., 2019) A-bodies, suggesting that these biomolecular condensates play an important role in many cellular processes. Here, we assessed whether the A-body marker molecule CDC73 (D. melanogaster homolog Hyrax) was localized to A-bodies across our panel of species, but it would be fascinating to determine whether targeting of additional protein constituents was likewise conserved. Functionally, we assessed whether a recently described form of nuclear protein synthesis (Theodoridis et al., 2021) occurred in the A-bodies of the cells and organisms examined. Although the presented work does not employ all of the LNPS detection assays used in the original report identifying this process (Theodoridis et al., 2021), the detection of puromycin-incorporated polypeptides in these subnuclear foci indicates that LNPS may be a conserved feature of A-body biology in mammalian, chicken and fish-derived cells. Interestingly, A-body-associated LNPS appears to be absent in both D. melanogaster and S. cerevisiae, suggesting that this nuclear protein synthesis pathway may have evolved around the time vertebrates and invertebrate species diverged. A brief analysis of homologs for the critical cytoplasmic and nuclear translation factors, eIF4A1 and eIF4B (Theodoridis et al., 2021), respectively, demonstrated a notable difference in the level of conservation of these proteins. The human and largemouth bass (a close relative of the Bluegill, with a sequenced genome) eIF4B proteins are 611 and 635 amino acids long and share 67% sequence identity, while the closest fly and yeast homologs have 15–20% sequence identity to the human protein, and between 388–436 amino acids in length. If this is compared to eIF4A1 (392–406 amino acids long and 64–89% sequence identity across all tested species), it is tempting to speculate that the drastic reduction in conservation of this essential nuclear protein synthesis factor could play a role in the absence of LNPS in the lower eukaryotes.

The discovery of putative A-bodies in D. melanogaster tissue and yeast marks the first occasion where stress-induced subnuclear foci with these amyloid-like characteristics were shown to form within an organismal setting. By inducing the aggregation of these structures outside of a cultured cell model, it dispels the notion that this type of amyloidogenesis is simply an artifact of the tissue culture system. It also presents several exciting opportunities. While the fruit fly tissue used in this report was limited to embryos and ovaries (Fig. 5A,B), moving forward additional organ systems could be assessed. Detection of A-bodies in other D. melanogaster tissue could generate a road map to identify human sites that may also respond to stressors through reversible amyloid aggregation. Additionally, we believe the discovery of A-bodies in D. melanogaster and yeast will unlock the extremely powerful genetic tools available for these model organisms. Flies and yeast have been a staple of biological research for over a century, enhancing our understanding of basic cell biology process ranging from genetic inheritance to embryonic development (Artavanis-Tsakonas et al., 1999; Reiter et al., 2001). Adding these tools to an A-body researchers tool kit should facilitate studies on the mechanisms and biological roles of this emerging cellular structure.

Mapping the evolutionary conservation of these amyloid aggregates across the eukaryotic domain highlights the significance of A-bodies as an adaptive mechanism. The high thermal and acidity thresholds needed to induce the formation of these biomolecular condensates in a range of species demonstrates that A-bodies are a stress-response pathway for severe, but sublethal, perturbations. This suggests that A-bodies could represent a pathway of last resort, as cells and organism struggle to survive extreme environmental insults. This level of conservation also furthers the notion that amyloidogenesis is a widespread and prominent physiological process, challenging the predominantly pathological implications that have been associated with the amyloid state of protein organization. Ultimately, the discovery of these aggregates in multiple organisms could assist our understanding of protein aggregation events that have been tied to numerous devastating neurological diseases and enhance our understanding of cellular adaptation to changing environmental conditions.

MCF-7, A7R5, COS-1, Swiss 3T3, DF-1, BF-2, and S2 cell lines were purchased from ATCC and maintained with the recommended media according to manufacturers' instructions. MCF-7, A7R5, COS-1, Swiss 3T3, DF-1 and BF-2 cells were seeded 24 h before treatments and grown to >80% confluence. S2 cells were seeded onto coverslips pre-treated with concanavalin A (0.5 mg/ml) 1 h prior to treatments. All stress treatments were administered for 2 h unless otherwise stated. Heat shock was performed at the temperature indicated. Extracellular acidification occurred within a H35 HypOxystation at 37°C in a 1% O₂, 5% CO₂, and N₂-balanced environment (Don Whitley Scientific, Frederick, MD, USA) following the replacement of standard growth medium with serum-free medium at the indicated pH. TPS treatments were induced by adding actinomycin D and MG132 to the media at the indicated concentrations. Flies (w¹¹¹⁸; Bloomington Drosophila Stock Center) were maintained at 25°C under standard conditions prior to treatment and harvest. S. cerevisiae strain BJ2168 (from Christopher Beh, Simon Fraser University, Burnaby, Canada) was validated using a Bruker MALDI Biotyper and cultured at 30°C in liquid yeast extract peptone dextrose media. Yeast cells were mounted on poly-L-lysine-treated coverslips prior to staining procedures.

The plasmids expressing human CDC73 (pEGFP-CDC73) and GAPDH (pEGFP-GAPDH) were previously described (Marijan et al., 2019). Hyrax, GAPDH, squid, NonA, cdk2 and cdk4 cDNAs from D. melanogaster were generated by RT-PCR from whole flies and cloned with the GFP-encoding sequence into the NotI/BamHI sites of the pAc5.1/V5-His plasmid. These constructs were transfected into MCF-7, A7R5, COS-1, Swiss 3T3, DF-1 or BF-2 cells that were seeded 24 h prior to transfections and grown to 50–70% confluence. 1 µg of plasmid DNA was transfected using polyethylenimine (PEI). S2 cells were seeded at a density of 3,000,000 cells/ml and transfected with 1 µg of plasmid DNA using Effectene (Qiagen). To induce sufficient protein expression, S2 cells were treated with 0.7 mM copper sulfate for 30 min, 48 h post-transfection. Experiments were performed 48 h after induction. Control siRNA (Silencer Select Negative Control No. 1 siRNA, Invitrogen) and siRNA targeting human rIGS₁₆RNA (target sequence: ACGCTGCCGTGTATGAACATA) mouse mrIGS₁₅RNA (target sequence: GTCGACCAGATGTCAGAAAGT), or chicken crIGS₁₂RNA (target sequence: CGCCGTTTTTGTTCTTATTTT) transcripts were purchased from Life Technologies (Thermo Fisher Scientific) and transfected using Lipofectamine 3000 (Thermo Fisher Scientific) for 24 h, then maintained in normal culture conditions for 24 h prior to harvesting the samples.

Immunofluorescence detection was performed as previously described (Marijan et al., 2019) using anti-CDC73 (1:100; Thermo Fisher Scientific, PA5-26189), anti-Hyrax (1:100 Thermo Fisher Scientific, PA5-40082) and anti-fibrillarin (1:100; EMD Millipore, MABE1154), donkey-anti-rabbit-IgG conjugated to Alexa Fluor 488 (1:200; Invitrogen, #A21206), and donkey-anti-mouse-IgG conjugated to Alexa Fluor 555 (1:200; Invitrogen, # A31570) antibodies. Yeast immunofluorescence detection of endogenous fibrillarin was performed as previously described (Ayscough and Drubin, 1997), replacing the methanol/acetone permeabilization with 0.5% Triton X-100 for 5 min. Samples stained with amyloidophilic dyes were fixed for 20 min in 4% formaldehyde and permeabilized for 1 min with 0.5% Triton X-100. Following two washes with 1× PBS, the samples were rinsed twice with ddH₂O (2 min). Either 0.05% Congo Red (Fisher Chemical), 0.01% Thioflavin S (Sigma-Aldrich), or equal parts 0.1% Congo Red and 0.02% Thioflavin S were used to stain cells for 15 min. After two washes in ddH₂O, samples were mounted on slides in 5% glycerol and visualized immediately. All images were captured on a Zeiss LSM880 laser scanning microscope (Carl Zeiss Microscopy, Germany) with Airyscan and ZEN 2.3 software (Carl Zeiss Microscopy, Germany). Images were analyzed using ImageJ (National Institutes of Health, USA), and quantification of the relative A-body intensity (I_r) was calculated as I_r=(I_A−I_b)/(I_n−I_b), with I_A, I_n, and I_b, being the intensity of the A-bodies, nucleus and background, respectively. Data represents the mean relative intensity of 10 cells captured in at least three independent replicates.

Fluorescence recovery after photobleaching (FRAP) experiments were performed as previously described (Audas et al., 2016, 2012; Jacob et al., 2012; Marijan et al., 2019). Briefly, transfected cells were grown on glass-bottom 35 mm culture dishes (MatTek Corporation, USA) and visualized on a Zeiss LSM880 laser scanning microscope (Carl Zeiss Microscopy, Jena, Germany). Full A-bodies or equal sized portions of the nucleoplasm or cytoplasm were bleached (100% argon laser pulse at 488 nm) and imaged for the indicated times post-bleaching. The intensities of the regions were measured in ImageJ software (National Institutes of Health, USA) and normalized as previously described (Audas et al., 2016, 2012; Jacob et al., 2012; Marijan et al., 2019). All fluorescence recovery after photobleaching experiment data represent the average of 10 or more cells.

Following a previously established methodology (Audas et al., 2016; Marijan et al., 2019), treated cells were rinsed in 1× PBS, lysed in NP40 buffer (1% NP40, 50 mM Tris-HCl, 150 mM NaCl) on ice for 5 min, and subjected to two rounds of sonication (Branson Ultrasonics, USA) at 10% power for 30 s (1 s pulse on/off). Whole cell lysate (WCL) aliquots were taken, and the remaining sample was pelleted at 6010 g for 10 min. Pellets were resuspended in fresh NP40 buffer and sonicated briefly (4 s at 10% power) to resuspend the insoluble (Insol) fraction. Anti-CDC73 (1:1000; Thermo Fisher Scientific, PA5-26189), anti-Hyrax (1:1000; Thermo Fisher Scientific, PA5-40082) anti-GAPDH (1:1000; Santa Cruz Biotechnology, sc-25778), anti-β-Actin (1:1000; Abcam, ab8224), anti-histone H3 (1:10,000; Cell Signaling, #9715), donkey anti-rabbit-IgG conjugated to HRP (1:10,000; Invitrogen, A16023), and donkey anti-mouse-IgG conjugated to HRP (1:10,000; Invitrogen, A16011) antibodies were used to detect proteins by western blotting.

Cell viability was determined for DF-1, BF-2, S2, and S. cerevisiae cells grown under standard, heat-shock or acidotic conditions for 2 h. Untreated cells exposed to 0.1% Triton X-100 were used as a positive control. Each sample was stained with 2 mM propidium iodide (Thermo Fisher Scientific) and Hoechst 33342 (Thermo Fisher Scientific) for 15 min, immediately prior to imaging on an EVOS FL Auto 2 fluorescence microscope (Thermo Fisher Scientific). The percentage of dead cells was determined by counting the number of propidium iodide-positive cells per field versus Hoechst-stained nuclei. Experiments were performed in triplicate.

Total RNA was isolated from MCF7, Swiss 3T3, and DF-1 cells following the indicated treatment using TriZol reagent (Thermo Fisher Scientific) according to manufacturer's instructions, then submitted for total RNA sequencing (Michael Smith Genome Sciences Centre). Ribosomal RNA was depleted, and Illumina PET 150 sequencing was performed on pooled samples. Adapter sequences were removed from reads using Cutadapt v3.1 (http://dx.doi.org/10.14806/ej.17.1.200), then aligned to the appropriate genome (GRCh38, GRCm38 or GRCg6a) with the addition of the intergenic spacer region of the rDNA cassette (U13369, BK000964, MG967540) using STAR v2.7.6a (www.ncbi.nlm.nih.gov/pubmed/23104886). Graphical representations of RNA-seq data were produced using IGV as previously described (Wang et al., 2018). Quantitative PCR validations were performed as previously described (Marijan et al., 2019), using cDNA produced by the qScript Flex cDNA Synthesis Kit (QuantaBio) according to manufacturer's instructions with random primers.

Embryos ranging from 13.5–15.5 h post-egg laying [either untreated (25°C) or heat-shocked (1.5 h, 38.5°C)] were dechorionated and fixed using established protocols (Rothwell and Sullivan, 2007). Congo Red staining was performed as described above, following PFA fixation. Wild-type flies were collected upon eclosion and placed in vials containing wet yeast. Untreated (25°C) and heat-shocked (0.5 h, 38.5°C) adult flies were dissected in 1× PBS, then ovaries were removed using fine-point forceps and fixed in 4% paraformaldehyde for 15 min prior to Congo staining.

MCF7, DF-1, BF-2, S2, and S. cerevisiae were incubated with 2 µM puromycin (Thermo Fisher Scientific) for the last 15 min of stimulus. Following treatment, MCF7, DF-1, BF-2 and S2 cells were fixed with methanol, and S. cerevisiae were fixed with formaldehyde. Fixed cells were permeabilized with 0.5% Triton X-100 for 10 min, as previously described (Theodoridis et al., 2021), then immunofluorescence detection of puromycin-incorporating peptides was performed (DSHB, PMY-24A).

All graphs represent the mean values of three biological replicates, with individual data points (black dots) overlaid on each bar graph. All sample sizes used in the experiments were in line with those reported in the literature for similar experiments. Error bars represent the standard error of the mean (s.e.m.), and P values were calculated using the two-tailed unpaired Student's t-test, with the significance level indicated in figure legends.

The authors acknowledge Esther Verheyen and Gritta Tettweiler (Simon Fraser University) for providing reagents and insight into D. melanogaster S2 cell protocols, and Christopher Beh (Simon Fraser University) for providing the S. cerevisiae cultures used in this study.

The peer review history is available online at https://journals.biologists.com/jcs/article-lookup/doi/10.1242/jcs.258907.