Age-onset phosphorylation of a minor actin variant promotes intestinal barrier dysfunction

SUMMARY

Age-associated decay of intercellular interactions impairs the cells’ capacity to tightly associate within tissues and form a functional barrier. This barrier dysfunction compromises organ physiology and contributes to systemic failure. The actin cytoskeleton represents a key determinant in maintaining tissue architecture. Yet it is unclear how age disrupts the actin cytoskeleton, and how this, in turn, promotes mortality. Here we show that an uncharacterized phosphorylation of a low-abundant actin variant, ACT-5, compromises integrity of the C. elegans intestinal barrier and accelerates pathogenesis. Age-related loss of the heat shock transcription factor, HSF-1, disrupts the JUN kinase/Protein Phosphatase I equilibrium and increases ACT-5 phosphorylation within its troponin-binding site. Phosphorylated ACT-5 accelerates decay of the intestinal sub-apical terminal web and impairs its interactions with cell junctions. This compromises barrier integrity, promotes pathogenesis, and drives mortality. Thus, we provide the molecular mechanism by which age-associated loss of specialized actin networks impacts tissue integrity.

Graphical Abstract

eTOC Blurb

Aging is accompanied by a general loss of organ integrity. In the intestine, this “leakiness” can lead to infection, inflammation, and disease. Herein, Egge et al. uncover how dysregulation of a major structural component of cells, actin, leads to the loss of the intestine’s barrier and drives age and mortality.

INTRODUCTION

The ability of an organism to maintain long-term tissue functionality amid various environmental challenges is fundamentally rooted in its barrier integrity. Age-onset defects in intercellular interactions reduce the cells’ junctional integrity, resulting in non-selective permeability between neighboring cells. Thus, aging and a multitude of age-related diseases may likely result from the progressive disorganization and dysfunction of tissue integrity over time (Rera et al., 2012). While most organ systems in the body display this age-onset defect (Parrish, 2017), the intestine exhibits inherent physiological properties which render it particularly sensitive to age-related decay. Perhaps due to the inherent duality of its essential functions, the intestinal epithelium must facilitate absorption of nutrients, ions, and water across its luminal surface, yet maintain a protective barrier against microorganisms, dietary antigens, and environmental toxins. Recent studies implicate barrier dysfunction as a major contributor to inflammatory bowel disease (Mohanan et al., 2018), autoimmune disease (Manfredo Vieira et al., 2018), and hyperglycemia-induced systemic infection (Thaiss et al., 2018). Furthermore, intestinal barrier dysfunction is a highly accurate predictor of mortality in several animal models of aging (Dambroise et al., 2016; Rera et al., 2012). Yet, molecular events that control age-associated changes in cellular architecture and how these alterations promote the loss of tissue barrier integrity remain to be determined.

Cell-cell junctions have been implicated as important determinates underlying barrier dysfunction (Resnik-Docampo et al., 2017). Junctional integrity within the intestine is mediated by several multi-protein adhesion complexes including tight and adherens junctions, which reside at cell-cell interfaces to facilitate barrier maintenance, intercellular communication, and tissue organization. The C. elegans intestine shares numerous similarities and conserved components with their vertebrate and drosophila counterparts (Knust and Bossinger, 2002). However, the worm intestine possesses a single junction that serves both as adherens and tight junctions, providing a simplified system for analysis of cell junctions (Armenti and Nance, 2012). These adherens junctions in C. elegans assemble on the apical surface of intestinal epithelia and are composed of cadherin adhesion molecules, which are anchored to the actin-rich subapical terminal web (Bernadskaya et al., 2011; Costa et al., 1998). Actin microfilament turnover and organization are tightly controlled to produce mechanical forces that drive the assembly, maintenance and remodeling of adherens junctions in development and cell migration (Sumi et al., 2018). How actin affects long-term junctional integrity and barrier function is not understood.

Maintenance of actin microfilament stability into old age improves organismal health and prolongs the aging process (Baird et al., 2014). Yet, the particular actin species and molecular details underlying this emerging paradigm linking actin cytoskeletal dynamics and age regulation is unknown. Whereas unicellular organisms typically possess a single gene encoding actin (Shortle et al., 1984), the advent of multicellularity has given rise to the genetic duplication and diversification of multiple actin proteins. Although highly homologous, actin variants possess unique promoter elements and amino acid substitutions to specify their tissue, cellular, and molecular functions. Compared to the human genome which harbors approximately 20 actin variants (Humphries et al., 1981), C. elegans encode five actin genes (Files et al., 1983; MacQueen et al., 2005). Thus, C. elegans present the opportunity to investigate the fundamental and physiological differences underlying actin diversification. Acting primarily in the C. elegans muscle and myofilament containing cells (Waterston et al., 1984), the ACT-1, ACT-2, and ACT-3 genes cluster within the same genomic locus and appear to be functionally redundant (Landel et al., 1984). ACT-4 is also strongly expressed in muscle (Stone and Shaw, 1993), whereas ACT-5 functions predominately in the intestine (Gobel et al., 2004). While ACT-1 through ACT-4 share 99% sequence homology, ACT-5 is the most sequence divergent, sharing 93% homology to the other actin proteins. Within the intestine, ACT-5 serves a unique physiological role, which cannot be rescued by the ectopic expression of other actin proteins (MacQueen et al., 2005). At the apical intestine, ACT-5 comprises a major structural component for microvilli extending into the intestinal lumen and the sub-apical terminal web, which provides a scaffold for anchoring of cell junctions (Figure 1A) (MacQueen et al., 2005). However, it is not understood how the regulation of ACT-5 impacts intestinal barrier maintenance during aging.

Figure 1. Loss of ACT-5 in adulthood impairs intestinal morphology and luminal integrity to promote pathogenesis.

(A) Schematic of the C. elegans intestine shows the apical organization of ACT-5 within the subapical terminal web and microvilli. (B,C) TEM ultrastructural analysis of C. elegans at day 3 and 5 of adulthood on control or act-5 RNAi. Arrows with MW indicate a microvilli tip projecting into the intestinal lumen. *TW and B denote the subapical terminal web and E. coli respectively. White box highlights the panel zoom. Scale bars: 0.5 μm (B) and 20 μm (C). (D) Bacterial colonization/invasion in day 3 adult C. elegans cultured on control or act-5 RNAi conditions. Scale bar: 200 μm. (Top) Phase of anesthetized worms, (middle) fluorescence of mCherry-expressing E. coli within the worm intestine and body cavity, (bottom) profiles of a larger worm population analyzed by flow cytometry. Dopamine neurons expressing GFP (green) used for head/tail orientation. Distribution of worm body length across the population is shown on the x-axis in arbitrary units. Animal numbers for control n=467 and act-5 RNAi n=331. (E) Worm lengths, as measured by the COPAS Biosorter, were normalized using WormProfiler on a scale of 1–100%. mCherry fluorescence intensities were determined from normalized body lengths and plotted as an average profile from the worm population on the left Y-axis. Fluorescence from GFP-expressing dopaminergic neurons was plotted on the right Y-axis for animal orientation. Day 10 adult animals on control RNAi reveal the extent of gut colonization during normal aging. (F) Micrographs of day 3 adult worms show the localization of the adherens junction protein, DLG-1, tagged with GFP under the respective RNAi conditions. White box highlights the panel zoom. Scale bar: 20 μm.

HSF-1 is a potent aging modifier (Hsu et al., 2003), whose ability to stabilize actin microfilaments under times of stress and age (Baird et al., 2014), make it an attractive regulatory candidate of the intestinal ACT-5 network. In addition to its well-defined role in protein folding and cytoplasmic stress responsiveness (Akerfelt et al., 2010), HSF-1 positively regulates expression of the actin microfilament stabilizing troponin C, PAT-10, to ensure integrity of the actin cytoskeleton (Baird et al., 2014). In support of HSF-1 operating through ACT-5, genome-wide proteomics identified ACT-5 as the only actin protein whose expression correlated with HSF-1 mediated thermal protection and lifespan extension (Baird et al., 2014). Furthermore, HSF-1 expression is required in the intestine, where ACT-5 functions, for HSF-1 mediated neurosecretory signals to extend animal lifespan (Douglas et al., 2015). Thus, we hypothesize that HSF-1 modulates ACT-5 activity to impact intestinal barrier function, pathogenesis, and age progression. Herein we report that the loss of HSF-1 activity by aging or RNA interference (RNAi) relieved repression of the JUN kinase, KGB-1, resulting in the aberrant phosphorylation and mislocalization of ACT-5. Furthermore, we have identified the Protein Phosphatase 1 (PP1) catalytic subunit beta, GSP-1, as epistatic to KGB-1 and as a potential regulator of ACT-5 dephosphorylation. Loss of ACT-5 apical polarity via phosphorylation of its serine 232 residue (S232) abolished its interactions with adherens junction proteins and promoted their rapid disorganization. This junctional disruption caused dilation of the intestinal lumen and loss of barrier function, as evidenced by pathogenic invasion into the body cavity and systemic infection. These studies highlight the continuum between HSF-1 mediated age regulation and intestinal barrier dysfunction, which include a newly identified post-translational modification at a highly conserved site within a specialized, low-abundant variant of actin.

RESULTS

Minor actin variant, ACT-5, affects intestinal aging and barrier function

Variants of actin often serve specialized roles in animal physiology, and their function cannot readily be substituted by the ectopic expression of other highly homologous actin variants (Fyrberg et al., 1998; Kumar et al., 1997; MacQueen et al., 2005). Of the five variants in C. elegans, ACT-5 was the only actin whose protein levels correlated with stress-mediated lifespan extension and thermotolerance (Baird et al., 2014). To better understand the role of ACT-5 in the aging intestine, we first examined intestinal morphology by ultrastructural analysis via transmission electron microscopy (TEM). To bypass developmental defects, worms were grown until late larval stages on control RNAi before shifting to act-5 RNAi. By day 3 of adulthood, worms treated with act-5 RNAi displayed several intestinal abnormalities similar to those observed upon aging (McGee et al., 2011) including luminal dilation, microvilli shortening (0.76 μm in control conditions versus 0.46 μm upon act-5 RNAi), and bacterial colonization (Figure 1B and S1A,B). These phenotypes are consistent with a previous report which highlights defective luminal and microvilli development upon the loss of act-5 expression (MacQueen et al., 2005). Although microvilli defects arose at day 3 of adulthood, the sub-apical terminal web remained intact (Figure 1B). However, by day 5, animals treated with act-5 RNAi displayed advanced stages of intestinal degeneration including a complete loss of microvilli (0.75 μm in control worms versus 0.053 μm), bacterial colonization, and an ill-defined luminal barrier (Figure 1C and S1B). Thus, our data indicate that the ACT-5 network in microvilli is the first to become dysfunctional prior to disruption of the sub-apical terminal web.

To ensure that act-5 RNAi did not globally impact all actin networks, western blot analysis with a pan-actin antibody revealed no change in total actin levels (Figure S1C). Comparative western blot analysis with an ACT-5-specific antibody and the GFP∷ACT-5 transgene (Szumowski et al., 2016) showed that endogenous ACT-5 comprises approximately 1.6% of total worm actin (Figure S1C,D). These results indicate that intestinal epithelia require expression of this low-abundant actin variant with age to maintain tissue architecture.

Consistent with ultrastructural analysis, loss of act-5 promoted rapid intestinal colonization of mCherry-expressing E. coli detected by fluorescence microscopy and large-particle flow cytometry (Figure 1D). Whole body fluorescence profiling showed the strongest bacterial colonization in the proximal intestine (Figure 1E). Normal aging resulted in intestinal colonization by day 10 of adulthood (Figure S1E), similar to previous reports of age-onset pathogenesis (Garigan et al., 2002; McGee et al., 2011), with bacterial colonization initiating in the proximal intestine. Thus, we hypothesize that a loss of functional ACT-5 networks represents a normal consequence of age that directly compromises the intestinal barrier.

In C. elegans, the discs large homolog 1, DLG-1, localizes to actin-tethered adherens junctions (McMahon et al., 2001). We believe that ACT-5 anchors these junctional proteins to facilitate barrier integrity of the apical intestine. To test whether ACT-5 affects localization of junctional proteins, we treated worms expressing a transgenic DLG-1∷GFP fusion protein with act-5 RNAi. Upon act-5 knockdown, DLG-1 adopted an aberrant morphology in the proximal intestinal cells, indicating disorganization of adherens junctions (Figure 1F). In the pharyngeal cells, which do not express act-5, DLG-1 localization remained unperturbed, indicating that defects in junctional morphology due to act-5 knockdown are specific for intestinal cells. Moreover, normal aging impaired DLG-1∷GFP localization within the intestine by day 15 of adulthood (Figure S1F). Therefore, age-associated defects in ACT-5 and adherens junctions may explain the disruption in intestinal barrier morphology over the life of the animal.

HSF-1 regulates ACT-5 localization and function early in adulthood

Our data indicate that ACT-5 functions at the apical surface of intestinal epithelia (Figure 1A) to maintain microvilli, ensure proper adherens junction localization, and sustain the luminal barrier. Exogenous expression of ACT-5 localizes to the appropriate subcellular regions of the apical intestine and rescues growth arrest defects in loss-of-function ACT-5 mutants (MacQueen et al., 2005). To determine how aging influences this ACT-5 network, we examined GFP∷ACT-5 which progressively mislocalized with age and adopted a more cytoplasmic distribution in the epithelia by day 10 of adulthood (Figure 2A,B). Redistribution of GFP∷ACT-5 from the apical intestine coincided with aberrant microvilli and adherens junction morphologies (Figure S1A,B,F). Yet, the cause of age-associated ACT-5 mislocalization from the apical epithelium requires additional clarification.

Figure 2. ACT-5 localization and intestinal morphology require HSF-1.

(A) Distribution of GFP∷ACT-5 in day 3 adults on control or hsf-1 RNAi and day 10 adults on control RNAi. White box highlights the panel zoom. Arrows denote basolateral intestine. Scale bar: 20 μm. (B) Quantification of apical GFP∷ACT-5 fluorescence. Error bars denote SEM with ***p=0.0007 and **p=0.0045 as determined by one-way ANOVA with Tukey comparison. (C) Lifespan analysis of hsf-1(CT) worms overexpressing the C-terminal truncation of HSF-1. (D) TEM ultrastructural analysis of C. elegans at day 3 and 5 of adulthood on control or hsf-1 RNAi. Arrows with MV highlight a microvilli tip projecting into intestinal lumen. *TW and B denote the subapical terminal web and undigested E. coli, respectively. Scale bars: 0.5 μm. (E) Western blot analysis of tubulin (red) and both endogenous ACT-5 and ectopically expressed GFP∷ACT-5 (green). WT or transgenic animals ectopically expressing GFP∷ACT-5 were collected at day 3 of adulthood on respective RNAi conditions. (F) Micrographs of the fluorescent actin filament indicator, LifeAct∷mRuby, in the intestine of day 3 adults under respective RNAi conditions. White box highlights the panel zoom. Arrows denote basolateral intestine. Scale bar: 20 μm.

HSF-1 is a potent aging regulator whose activity declines precipitously with age (Ben-Zvi et al., 2009; Labbadia and Morimoto, 2015) and has a role in actin cytoskeletal regulation (Baird et al., 2014). Thus we examined whether HSF-1 acts through ACT-5 to modulate the aging process. Treatment with act-5 RNAi during adulthood abolished lifespan extension in long-lived animals overexpressing the carboxy-terminal (CT) deletion of hsf-1 (Figure 2C and Table S1). Hsf-1 (CT) was utilized to minimize contributions of molecular chaperones and focus on its cytoskeletal role (Baird et al., 2014). Conversely, we sought to understand whether age-associated loss of HSF-1 activity promotes aberrant ACT-5 function. Following reduction of hsf-1 levels by approximately 4-fold (Figure S2A) via systemic RNAi, ultrastructural analysis revealed these animals exhibited similar aberrant intestinal morphologies and aging kinetics to those observed upon act-5 RNAi (Figure 2D and S2B). These included luminal dilation, accumulation of undigested bacteria, and microvilli degeneration (0.69 μm versus 0.2 μm).

Hsf-1 knockdown caused an approximate 3-fold increase in the steady-state levels of both endogenous ACT-5 and the exogenous GFP∷ACT-5 fusion protein as seen by western blot (Figure 2E). Flow cytometry confirmed that hsf-1 RNAi increased the fluorescence intensity of GFP∷ACT-5 in day 3 transgenic animals (Figure S2C). Despite elevated protein levels, we hypothesized that ACT-5 function was compromised due to its inability to remain incorporated within the apical actin network. Similar to the effects of aging, the subcellular distribution of GFP∷ACT-5 was dramatically altered by day 3 of adulthood upon hsf-1 RNAi, favoring a cytoplasmic localization rather than an apical polarization (Figure 2A,B). In combination with ultrastructural defects observed within the intestine (Figure 2D), our data suggest that HSF-1 promotes structural integrity of the intestinal actin network through its ability to properly localize ACT-5.

C. elegans harbors several distinct actin networks within different tissues (MacQueen et al., 2005). We sought to further understand how HSF-1 influences the filamentous state of the ACT-5 network in relation to other actin networks in the worm. A fluorescent indicator of filamentous actin, LifeAct∷mRuby, was ectopically expressed in the intestine or in the body-wall muscle, which is void of ACT-5 expression (Higuchi-Sanabria et al., 2018). LifeAct∷mRuby decorated apically-localized ACT-5 microfilaments in the intestine of young day 1 adult animals, and this signal was abolished upon act-5 RNAi treatments (Figure S2D,E). Although fluorescence intensity remained constant (Figure S2E), hsf-1 RNAi altered the subcellular distribution of intestinal LifeAct∷mRuby, favoring its cytoplasmic accumulation into fluorescent foci (Figure 2F). Conversely, LifeAct∷mRuby localization within the body-wall muscle remained unaffected by either hsf-1 or act-5 RNAi at day 3 of adulthood (Figure S2F), despite overall levels being modestly elevated upon hsf-1 RNAi (Figure S2G). Taken together, loss of HSF-1 appears to selectively target the subcellular distribution of intestinal ACT-5 in early adulthood.

Loss of HSF-1 function impairs adherens junctions and intestinal barrier

Our data suggest that junctional stability and tissue morphology require proper ACT-5 expression and localization. Yet it remained to be determined whether loss of HSF-1 can disrupt junctional organization, leading to a loss of intestinal membrane integrity and sensitivity to pathogenesis. We examined the subcellular distribution of different adherens junction proteins under hsf-1 RNAi conditions. Transgenic worms expressing DLG-1∷GFP displayed a selective loss of steady-state DLG-1 protein in the epithelial cells neighboring the intestinal-pharyngeal valve by day 3 of adulthood (Figure 3A and S3B). It is unclear why these particular cells display a dramatic decrease in steady-state protein levels, but this impairment remains consistent with the proximal intestine’s increased sensitivity to bacterial colonization and potential invasion. Indirect immunofluorescence of a different adherens junction protein, AJM-1, which requires DLG-1 for proper localization (Firestein and Rongo, 2001), confirmed that hsf-1 RNAi disrupts junctional morphology in the same proximal intestinal epithelia (Figure S3A). Interestingly, hsf-1 RNAi in dead bacteria resulted in similar morphologic defects, indicating that breakdown of adherens junctions occurs independently of pathogenesis (Figure S3A). Although potentially refractory to RNAi treatments, abnormal junctional morphologies were not observed within the actin-rich pharyngeal muscle, which does not express ACT-5. Therefore, disruption of adherens junctions due to loss of HSF-1 appears to selectively occur in cells expressing ACT-5, indicating an ACT-5 dependent process.

Figure 3. Loss of HSF-1 expression impairs ACT-5 binding to adherens junction proteins and accelerates pathogenesis in the proximal intestine.

(A) Confocal micrographs of day 3 adult worms show localization of the adherens junction protein, DLG-1, tagged with GFP under control and hsf-1 RNAi conditions. White box highlights the panel zoom. Scale bars: 20 μm. (B) Fold-change in differential binding of ACT-5 to DLG-1, AJM-1, or PAT-3 upon hsf-1 RNAi as determined by MS/MS. (C) Western blot analysis of GFP∷ACT-5 Co-IPs from day 3 adults on control and hsf-1 RNAi (n=4). (D) Bacterial colonization/invasion in day 3 adult C. elegans treated with control or hsf-1 RNAi. Scale bar: 200 μm. (Top) Phase of 8 anesthetized worms, (middle) representative fluorescence of mCherry-expressing E. coli within the worm intestine and body cavity, (bottom) fluorescence profiles of worms analyzed by flow cytometry. GFP-expressing dopamine neurons (green) orient head/tail of individual animals. Worm length shown on the x-axis in arbitrary units. Animal numbers for control (n=467) and hsf-1 RNAi (n=404). (E) TEM ultrastructural analysis of C. elegans at day 5 of adulthood on hsf-1 RNAi. Disruption of terminal web is highlighted by *TW, and B denotes invading E. coli. White box highlights the panel zoom. Scale bars: 2 μm. (F) Quantification of population averages of bacterial colonization/invasion across the length of the worm in control and hsf-1 RNAi conditions. (G) Lifespan analysis of worms cultured on living or dead E. coli harboring control, hsf-1, or act-5 RNAi.

We sought to determine why knockdown of hsf-1 or act-5 dramatically altered junctional morphology within the intestinal epithelia. Adherens junctions tether cells through anchorage to terminal web microfilaments (Hartsock and Nelson, 2008). Loss of ACT-5 binding may weaken cell-cell tethering and increase the likelihood of junctional defects. Liquid chromatography-tandem mass spectrometry (LC-MS/MS) analysis of GFP∷ACT-5 co-immunoprecipitations (Co-IPs) indicated that hsf-1 RNAi resulted in reduced binding of ACT-5 to the adherens junction proteins, AJM-1 and DLG-1 (Figure 3B). ACT-5 interactions with other junctional markers, including proteins resident to tight and gap junctions from other tissues, were not detected by mass spectrometry. Western blot analysis of GFP∷ACT-5 Co-IPs confirmed an approximate 70% decrease in binding between ACT-5 and DLG-1 under hsf-1 RNAi conditions (Figure 3C and S3C).

Emerging studies implicate impairment of cell junctions as a key determinant of intestinal barrier integrity (Resnik-Docampo et al., 2017). We hypothesize that disrupted interactions between adherens junctions and ACT-5 microfilaments of the subapical terminal web weaken inter-cellular contacts between neighboring epithelia and promote pathogenesis. Under conditions of whole body and intestine-specific hsf-1 knockdown, we observed accelerated colonization of fluorescent bacteria (Figure 3D and S3D,E). In fact, loss of hsf-1 expression specifically in the intestine was sufficient to reduce lifespan (Figure S3F and Table S1). Attempts to reduce act-5 expression specifically in the intestine were confounded by the developmental requirement to knock down act-5 at late larval stages, at which point intestine-specific expression of the transgenic RNAi transporter, SID-1, was markedly reduced (data not shown). TEM micrographs confirmed that undigested E. coli within the intestinal lumen invade into the body cavity at rupture sites of the subapical terminal web in day 5 adult worms treated with hsf-1 RNAi (Figure 3E). Similar to act-5 RNAi and age, the onset of bacterial accumulation began in the proximal intestine where junctional defects first became apparent in early adulthood (Figure 3F). Consistent with other findings (Dambroise et al., 2016; Garigan et al., 2002), we believe that bacterial invasion across the intestinal epithelia and into the body cavity significantly contributes to mortality in C. elegans. Indeed, age-onset colonization and invasion of fluorescent bacteria preceded worm death as determined by staining with the membrane impermeable nucleic acid dye, Sytox (Figure S3G,H) (Gill et al., 2003).

To determine the contribution of pathogenesis to the aging process, we assessed lifespan on both living and dead bacteria. To assure death of the bacteria while preserving RNAi efficiency, E. coli were subjected to a brief formaldehyde fixation after induced expression of the respective RNAi in liquid culture. We observed a complete loss of bacterial proliferation (Figure S3I) yet still retained effective RNAi comparable to living bacteria, as seen by the effects of hsf-1 and act-5 RNAi on the steady-state levels of GFP∷ACT-5 (Figure S3J). We confirmed through indirect immunofluorescence of AJM-1 that act-5 and hsf-1 RNAi in dead bacteria disrupted junctional morphology in day 3 adults, similar to living E. coli (Figure S3A). However, worms cultured on dead bacteria were substantially longer lived under control, hsf-1, or act-5 RNAi conditions (Figure 3G and Table S1), suggesting that pathogenesis significantly contributes to lifespan determination by HSF-1 and ACT-5.

Phosphorylation within the troponin binding site of ACT-5 promotes its mislocalization

We observed that mislocalization of GFP∷ACT-5 during hsf-1 knockdown correlated with the emergence of a slower-migrating band upon immunoprecipitation and resolution by SDS-PAGE (Figure 4A). Tyrosine phosphorylation on actin has previously been reported under stress conditions (Howard et al., 1993). However, the slower migrating GFP∷ACT-5 species was resistant to treatments with the tyrosine-specific phosphatase, YopH (Guan and Dixon, 1990). Yet it was sensitive to the Lambda phosphatase, indicating phosphorylation of a serine, threonine, or histidine residue. In addition to hsf-1 RNAi, normal aging conditions resulted in increased phosphorylation of ACT-5. Long-lived worms overexpressing hsf-1(CT) showed a reduction in overall ACT-5 phosphorylation (Figure 4B).

Figure 4. Phosphorylation within the troponin binding site of ACT-5 disrupts growth and intestinal morphology.

(A) Western blot analysis of GFP∷ACT-5 IPs under control and hsf-1 RNAi conditions. GFP∷ACT-5 immunoprecipitates treated with the tyrosine-specific phosphatase, YopH (Y), or the lambda phosphatase (SYT) (n=3). * denotes phosphorylated GFP∷ACT-5. (B) Western blot analysis of GFP∷ACT-5 at day 3, 5, 10 and 15 of adulthood with and without hsf-1(CT) overexpression (n=3). * denotes phosphorylated GFP∷ACT-5. (C) Sequence alignment from yeast actin to human β actin. Highlighted are phosphorylated residues detected by MS/MS from day 3 adults on hsf-1 RNAi. (D) Structural models of F-actin and tropomyosin docked into density envelopes determined by TEM. Left, potential phosphorylation sites at S232 and S239 denoted on ribbon models. Middle, reconstruction of F-actin-tropomyosin complex decorated with the C-terminal peptide of troponin I. Right, overlay. (E) Percentage of growth-arrested transgenic animals at day 1 of adulthood determined by flow cytometry. Error bars denote SEM with ****p<0.0001 as determined by one-way ANOVA with Tukey’s multiple comparison (n≥3). (F) Body length of different transgenic GFP∷ACT-5 animals at day 1 of adulthood with control or hsf-1 RNAi. Animals retaining larvae sizes were excluded from this analysis. Error bars denote SEM with ****p≤0.0001, ***p=0.0007, and **p=0.0034 as determined by two-way ANOVA with Tukey comparison (n≥3). (G) Representative micrographs of day 1 adults ectopically expressing different GFP∷ACT-5 mutations. Scale bar: 20 μm. (H) Proteins isolated in complex with GFP∷ACT-5 were identified by MS/MS. Plot depicts the fold-change in differential binding of ACT-5 to actin-associated machinery upon hsf-1 RNAi as determined via label-free quantification.

To identify post-translationally modified ACT-5 residues, immunoprecipitations of GFP∷ACT-5 were resolved by LC-MS/MS. With over 69% peptide coverage of ACT-5 (Figure S4A), phosphates were identified on two highly conserved serine residues, S232 and S239 (Figure 4C and S4B,C). Based on the actin microfilament structure (von der Ecken et al., 2015), both candidate phosphorylation sites reside on a solvent exposed loop, which allows accessibility to phosphorylation machinery and minimizes the possibility of protein monomer instability. Since previous studies implicated the troponin C subunit, PAT-10, in HSF-1 mediated lifespan regulation (Baird et al., 2014), we examined whether either of these putative phosphorylation sites on ACT-5 might exist within its troponin binding site. Structural modeling of conjugated phosphates revealed that neither of these serine residues interfered with tropomyosin interactions (Galinska-Rakoczy et al., 2008). However, only S232 resided within the troponin I binding site of filamentous actin (Figure 4D).

We sought to determine whether phosphorylation at S232 was sufficient to disrupt GFP∷ACT-5 assembly and localization. We generated phosphomimetic mutations at S232 and S239 within the GFP∷ACT-5 transgene using glutamate to mimic the bulky negative charge of a phosphate group. Flow cytometry confirmed that S232D and S239D were expressed at levels similar to the wild type (WT) transgene in early larvae stages (Figure S4D). Yet only expression of the S232D mutation caused severe developmental and adulthood growth defects. Over 40% of the transgenic animals expressing S232D exhibited larval arrest (Figure 4E). Similar to hsf-1 RNAi treatments, the S232D mutant animals that reached adulthood were significantly smaller than WT and S239D transgenic animals (Figure 4F). Unlike the S239D transgene, the S232D mutation resulted in early age-onset aggregation and mislocalization of GFP∷ACT-5 (Figure 4G and S4E). Replacement of the bulky negatively-charged glutamate on the GFP∷ACT-5 phosphomimetic with a small, neutral alanine residue (S232A), resulted in transgenic animals with no gross developmental or morphological defects (Figure 4E,F,G and S4E). As the onset of ACT-5 phosphorylation in adulthood coincided with aberrant intestinal morphology (Figure 2A,B and 4B), transgenic expression of the deleterious phosphomimetic, S232D, during development may act in a dominant negative fashion to compromise assembly of endogenous ACT-5 networks. This may yield the larval arrest and aberrant growth phenotypes, similar to those reportedly induced by a loss of ACT-5 during development (MacQueen et al., 2005).

We hypothesized that S232 phosphorylation alters interactions of the troponin complex with the ACT-5 microfilaments. We identified several ACT-5 interacting partners through LC-MS/MS analysis of GFP∷ACT-5 Co-IPs. Several subunits of the troponin regulatory complex, including PAT-10, displayed increased affinity for GFP∷ACT-5 under hsf-1 RNAi conditions (Figure 4H). Consistent with the troponin complex competing with myosin for actin microfilament binding (Smith and Geeves, 2003), we observed a decrease in non-muscle myosins bound to GFP∷ACT-5 upon hsf-1 RNAi. The predicted troponin-actin binding site resides in actin residues 222–232 (Yang et al., 2014), and harbors several bulky negatively charged amino acids within the binding cleft, including D222, E224, and E226 (Figure 4C). Introduction of a negatively charged phosphate within the binding cleft on ACT-5 may increase actin-troponin binding affinity and hyper-stabilize microfilaments. Indeed, pharmacological stabilization of F-actin impairs assembly/disassembly dynamics and leads to the formation of large aggregated structures (Ayscough, 2000), similar to what we observe in the worm upon hsf-1 RNAi (Figure 2F).

GSP-1 promotes ACT-5 dephosphorylation and intestinal barrier integrity

Our data suggest that HSF-1 modulates aging and pathogenesis by influencing ACT-5 dynamics in the intestine via S232 phosphorylation. LC-MS/MS analysis of GFP∷ACT-5 Co-IPs identified several potential serine/threonine phosphatases as interacting partners. We hypothesized that compromising phosphatase activity would increase ACT-5 phosphorylation at S232 and promote aberrant ACT-5 phenotypes similar to those observed upon hsf-1 RNAi. To this end, we performed a visual screen to determine whether knockdown of individual candidate phosphatases via systemic RNAi would alter fluorescence intensity and subcellular localization of GFP∷ACT-5 (Figure 5A). The scoring system for the targeted phosphatase screen ranged from 0–6 (0–3 points for subcellular localization and 0–3 points for GFP∷ACT-5 fluorescence). A total score of 0 phenocopied control RNAi (all apically localized with normal expression) while a total score of 6 phenocopied hsf-1 RNAi (not apical but rather distributed throughout the cytoplasm with elevated expression) (Figure S5A). Overrepresented within this phosphatase set of ACT-5 binding partners were several subunits of the Protein Phosphatase 1 (PP1) complex. These included the beta catalytic subunits GSP-1, PPH-1, ZK938.1, and F52H3.6, alpha catalytic subunit GSP-2, and regulatory subunit SDS-22. The PP1 complex has already been shown to directly bind actin monomers and regulate actin filament assembly (Chambers et al., 2015), although the mechanism of action remains unclear. From our screen, two subunits of this PP1 complex, GSP-1 and GSP-2, were identified as top candidates whose expression was required for proper GFP∷ACT-5 expression and localization (Figure 5B and S5B).

Figure 5. GSP-1 antagonizes ACT-5 phosphorylation, pathogenesis, and age.

(A) Heat map depicting qualitative phosphatase RNAi screen of GFP∷ACT-5 localization and expression. C. elegans were scored on a scale of 0–6 (0 phenocopies control conditions and 6 phenocopies hsf-1 RNAi). (B) Confocal micrographs show the apical distribution of GFP∷ACT-5 in day 3 adult animals upon control, gsp-1, and gsp-2 RNAi. White box highlights the panel zoom. Arrows denote basolateral intestine. Scale bar: 20 μm. (C-E) Western blot analysis of steady-state GFP∷ACT-5 levels at: (C) Day 3 of adulthood under the respective RNAi conditions (n=5). (D) Day 3 for either WT or S232A GFP∷ACT-5 upon control or gsp-1 RNAi (n=3). (E) Day 3 and 5 with and without hsf-1(CT) overexpression upon control and gsp-1 RNAi (n=3). (F) Bacterial colonization/invasion in day 3 adult C. elegans treated with control or gsp-1 RNAi. Scale bar: 200 μm. (Top) Phase of anesthetized worms, (middle) representative fluorescence of mCherry-expressing E. coli within the worm intestine and body cavity, (bottom) fluorescence profiles from flow cytometry. GFP-expressing dopamine neurons (green) orient head/tail of individual animals. Worm length shown on the x-axis in arbitrary units. Control (n=467) and gsp-1 RNAi (n=383). (G) Quantification of bacterial colonization/invasion (mCherry, left y-axis) in control and gsp-1 RNAi conditions using WormProfiler. Dopaminergic neurons (GFP, right y-axis).

Utilizing transgenic animals expressing GSP-1∷GFP and GSP-2∷GFP from their endogenous promoter, GSP-1∷GFP was more prominent in the intestine compared to GSP-2∷GFP (Figure S5C). Thus, we focused on GSP-1 due to its abundance in the same tissues as ACT-5. Defects in GFP∷ACT-5 localization resulting from gsp-1 RNAi were coincident with the accumulation of phosphorylated GFP∷ACT-5 in day 3 adult worms (Figure 5C). Importantly, GFP∷ACT-5 phosphorylation under gsp-1 RNAi conditions was not observed in transgenic animals expressing the S232A mutation, indicating that GSP-1 acts to dephosphorylate this serine residue (Figure 5D). Furthermore, overexpression of hsf-1(CT) dramatically reduced GFP∷ACT-5 phosphorylation observed upon gsp-1 RNAi at day 3 and 5 of adulthood (Figure 5E).

To understand how ACT-5 phosphorylation upon gsp-1 RNAi impacted animal health, we examined several physiological parameters. Similar to effects elicited by hsf-1 and act-5 RNAi, worms treated with gsp-1 RNAi appeared unhealthy with a smaller body size (Figure S5D) and displayed early age-onset intestinal colonization/invasion of fluorescently-labeled E. coli (Figure 5F) predominately in the proximal intestine (Figure 5G). Moreover, gsp-1 RNAi caused a significant reduction in lifespan (Figure S5E and Table S1). Transcript levels of gsp-1 were significantly decreased by day 10 of adulthood and may, in part, account for age-onset phosphorylation of ACT-5 (Figure S5F). Yet, hsf-1 RNAi did not affect gsp-1 transcript levels (Figure S5G), indicating that HSF-1 mediated regulation of ACT-5 phosphorylation likely operates independent of GSP-1.

KGB-1 antagonizes GSP-1 function by catalyzing ACT-5 phosphorylation

HSF-1 may influence ACT-5 phosphorylation and barrier function through regulation of kinase activity. Leveraging the kinase substrate prediction algorithm, PhosphoNET (Safaei et al., 2011), three mammalian JUN kinases (JNKs) were identified as top candidates to catalyze S232 phosphorylation of the homologous human β-actin (Figure S6A). Our LC-MS/MS analysis identified KGB-1, the C. elegans JNK3 homolog, as the most abundant JUN kinase isolated in complex with ACT-5 exclusively upon hsf-1 RNAi (Figure S6B). The regulatory functions of JNKs have previously been linked with maintenance of intestinal epithelial integrity and junctional dynamics (Biteau et al., 2008; Resnik-Docampo et al., 2017). In particular, JNK activation remodels both the actin and junctional networks in human intestinal epithelia, which correlate with barrier dysfunction (Samak et al., 2015).

Consistent with previous reports showing that HSF-1 represses JNK (Su et al., 2016), we observed that hsf-1 RNAi significantly elevated transcript levels of kgb-1 (Figure 6A and S6C). Since KGB-1 directly activates the AP1 transcriptional complex, we also examined the well-characterized AP1 transcriptional target, LYS-3, to report on KGB-1 activity (Hattori et al., 2013). Hsf-1 RNAi increased lys-3 transcript levels, whereas HSF-1 activation by transient heat shock repressed transcription of lys-3 (Figure 6B). Importantly, lys-3 repression by heat shock was dependent on HSF-1 expression. These results support HSF-1 as a repressor of KGB-1. Furthermore, age-related decline in hsf-1 transcript abundance inversely correlated with increasing kgb-1 expression and activity as seen by kgb-1 and lys-3 transcript abundance, respectively (Figure S6D).

Figure 6. Epistasis between KGB-1 and GSP-1 activity on GFP∷ACT-5 localization, pathogenesis, and age.

(A) Transcript abundance of the JUN kinase, KGB-1, as determined by RNAseq upon control and hsf-1 RNAi. Error bars denote SEM with **p=0.0065 as determined by student’s t-test. (B) Transcript abundance of the KGB-1 target, LYS-3, in late larval stages (L4). Animals on the respective RNAi conditions were subject to a transient heat shock. Error bars denote SEM with **0.0027 and ***p=0.0003 as determined by two-way ANOVA with Tukey comparison. (C) Confocal micrographs show the apical distribution of GFP∷ACT-5 of day 3 adults with different RNAi combinations. Arrows indicate basolateral intestine. White box highlights the panel zoom. Scale bar: 20 μm. (D) Quantification of apical GFP∷ACT-5 fluorescence. Error bars denote SEM with ***p=0.0007 and ****p<0.0001 as determined by one-way ANOVA with Tukey comparison. (E) Western blot analysis of GFP∷ACT-5, GSP-1 and KGB-1 levels in day 3 adults under the respective RNAi combinations (n=3). (F) Bacterial colonization/invasion in day 3 adults treated with the respective RNAi combinations. Scale bar: 200 μm. (Top) Phase of anesthetized worms, (middle) representative fluorescence of mCherry expressing E. coli within the worm intestine and body cavity, (bottom) fluorescence profiles of worms analyzed by flow cytometry. GFP-expressing dopamine neurons (green) orient head/tail of individual animals. Worm length shown on the x-axis in arbitrary units. Animal numbers for control (n=467), gsp-1/control (n=383), and gsp-1/kgb-1 (n=454) RNAi conditions. (G) Quantification of bacterial colonization and invasion (mCherry, left y-axis) in gsp-1/control or gsp-1/kgb-1 RNAi conditions using WormProfiler. Dopaminergic neurons (GFP, right y-axis). (H) Lifespan analysis of CF512 worms cultured on the respective RNAi combinations.

We hypothesized that an imbalance in JUN kinase and PP1 phosphatase activities would result in the aberrant accumulation of phosphorylated ACT-5. To examine this interplay, we determined whether modulating KGB-1 expression and activity would impact the deleterious phenotypes observed with gsp-1 RNAi. Others have reported that loss of KGB-1 modestly shortens lifespan (Twumasi-Boateng et al., 2012). However, co-administration of kgb-1 RNAi reduced the deleterious effects of gsp-1 RNAi on GFP∷ACT-5 localization and worm body length (Figure 6C,D and S6E). Moreover, kgb-1 RNAi suppressed GFP∷ACT-5 phosphorylation associated with gsp-1 RNAi (Figure 6E). Conversely, increasing KGB-1 activity by removing its negative regulator, VHP-1, (Mizuno et al., 2004; Zhang et al., 2017) was sufficient to increase GFP∷ACT-5 phosphorylation in day 3 adults. The rescue of intestinal architecture by co-administering kgb-1 and gsp-1 RNAi coincided with reduced rates of intestinal pathogenesis and improved lifespan, as compared to gsp-1 RNAi alone (Figure 6F–H and Table S1). Moreover, intestine-specific gsp-1 RNAi was sufficient to shorten lifespan and accelerate early-onset bacterial colonization (Figure S6F–H). Similar to whole-body, intestine-specific co-administration of kgb-1 RNAi mildly rescued the deleterious effects of gsp-1 RNAi on lifespan. Likewise, co-administration of kgb-1 RNAi in animals overexpressing hsf-1(CT) extended lifespan compared to gsp-1 RNAi alone (Figure S6I and Table S1). In combination, these findings indicate that intestinal integrity becomes compromised during normal aging through a loss of this kinase/phosphatase equilibrium, initiated by loss of KGB-1 repression through age-onset decline in HSF-1 activity (Figure 7A).

Figure 7. Model for age-associated intestinal barrier dysfunction in C. elegans.

(A) Proposed model for the ACT-5 phosphorylation cycle with the respective regulatory machinery. (B) Schematic depicting a young healthy intestine (left) compared to an aged intestine (right). In young epithelial cells, HSF-1 activity maintains KGB-1 in a repressed state, favoring an unphosphorylated, apical ACT-5. Upon loss of HSF-1 activity, KGB-1 promotes ACT-5 phosphorylation and mislocalization, which results in reduced binding to adherens junction proteins, impairs cellular connectivity, and promotes paracellular bacterial invasion. The subcellular locations for KGB-1 and GSP-1 have not been experimentally validated.

DISCUSSION

Our findings detail the molecular events underlying age-related decay of the apical barrier of the intestinal epithelium (Figure 7B), which accurately predicts animal mortality in several metazoan and vertebrate models of aging (Dambroise et al., 2016; Rera et al., 2012). We identify the stress-responsive transcription factor, HSF-1, as a regulator of intestinal barrier homeostasis and pathogenesis. In addition to acting through conventional protein folding enzymes (Akerfelt et al., 2010), HSF-1 represses the JUN kinase, KGB-1, which impairs polarization of actin networks and accelerates tissue barrier dysfunction, age progression and pathogenesis. Furthermore, we have uncovered an epistatic interaction between the JNK family of stress-activated kinases and the PP1 holoenzyme. Overall, these studies begin to uncover core molecular mechanisms of age regulation, which extend the continuum from protein homeostasis and cytoskeletal dynamics into the barrier function of tissues.

Central to this aging mechanism is the identification of an uncharacterized and highly conserved residue within a specialized, tissue-specific actin, which is essential for barrier tissue health and normal age progression. Early studies hinted that actin phosphorylation mediates a more dynamic role in stress-induced alternation of membrane morphology (Howard et al., 1993). However, the role of actin phosphorylation in membrane remodeling and its biological significance remained unclear. Subsequent studies have since identified at least 35 amino acids residues within actin that can be modified by phosphorylation (Terman and Kashina, 2013). In the slime mold, Distyostelium, phosphorylation at Y53 reduces F-actin levels and likely acts to sterically impede actin subunit-subunit contacts (Liu et al., 2006). Alternatively, phosphorylation at T201–203 residues increases F-actin levels and likely acts to inhibit binding by the actin severing factor, Fragmin (Furuhashi and Hatano, 1990). The phosphorylation described herein at S232 represents a newly identified post-translational modification, which we suggest directly regulates interactions between the actin microfilament and the troponin complex. Whether ACT-5 phosphorylation at S232 represents an aberrant modification leading to an intestinal pathology or serves an important physiologic role for the intestine remains unclear.

In our search for phosphorylation machinery acting on ACT-5, we discovered an epistatic interaction between the stress-responsive JUN kinase (KGB-1) and a catalytic subunit of the PP1 complex, GSP-1. The JUN kinase has been implicated with intestinal barrier dysfunction through its ability to aberrantly regulate intestinal stem cell proliferation/differentiation (Biteau et al., 2008). Herein, we report a role for the JUN kinase in the somatic epithelium where it regulates dynamics of an intestinal-specific actin to impact tissue architecture and barrier integrity of non-dividing cells. Furthermore, our data suggests that age-associated changes in JNK activity likely arise from a decline in HSF-1 activity. The PP1 complex has been implicated as a regulator of actin microfilaments and directly interacts with actin monomers (Chambers et al., 2015). We provide an additional role for the PP1 subunit, GSP-1, within the context of intestinal barrier dysfunction, bacterial colonization, and aging. It remains to be determined what other catalytic and regulatory subunits of the PP1 holoenzyme are required to modulate actin phosphorylation. Furthermore, additional analysis is needed to demonstrate direct action of this phosphorylation machinery on ACT-5 and to determine whether this is specific for particular actin variants in the cell.

While extensive studies have defined biophysical and cellular functions of actin, little work has focused on the diversity of actins within an organism and how they may contribute to different physiological roles in tissue structure and function. Canonically, six actin genes have been reported in the literature for mammals (Vedula and Kashina, 2018), yet several actin-like proteins are not annotated as actins despite sharing high amino acid sequence homology and similar molecular weights. Southern blot analysis indicates at least 20 different forms of actin within the human genome (Humphries et al., 1981). Thus, future studies should examine in greater detail the array of actins and actin-like proteins encoded within the genome. Our studies begin to highlight the functional importance of low-abundant, tissue-specific variants of actin, which still remain poorly characterized.

STAR METHODS

Lead Contact and Materials Availability

Further information and requests for resources and reagents should be directed to and will be fulfilled by the Lead Contact, Peter Douglas (peter.douglas@utsouthwestern.edu). Worm strains, plasmids, and antibodies generated in this study are available upon request from the Lead Contact with a completed Materials Transfer Agreement.

Experimental Model and Subject Details

Microbe Strains

OP50 was purchased from the CGC. Stock OP50 E.coli bacteria were maintained on Luria Bertani (LB) plates. HT115 E. coli from our Ahringer and Vidal RNAi libraries were cultured under carbenicillin selection for all RNAi experiments. All bacteria were grown at 37°C and broth cultures were shaken at 250 rpm. Following overnight culturing of RNAi bacteria at 37°C, cultures were induced with 1 mM IPTG for 4 hrs at 37°C.

C. elegans Strains

C. elegans were maintained on nematode growth media (NGM) plates seeded with Escherichia coli OP50 as described (Brenner, 1974). Worm strains that were used for experimental purposes (lifespans, western blotting, flow cytometry, etc.) were grown on NGM plates supplemented with HT115 E. coli at 20°C or 25°C for all temperature sensitive stra ins. Listed below are all C. elegans strains used in this study. The following strains were obtained from the CGC: N2 (ancestral) as WT, CF512: rrf-3(b26) II; fem-1(hc17) IV, ML2615: dlg-1(mc103[dlg-1∷GFP]) X, BC14231: (dpy-5(e907) I; sEX14231, [RcESf29f11.6∷GFP + PcEH361], BC10077: (dpy-5(e907) I; sEX10077[RcESf56c9.1∷GFP + pCeh361]. Strain ERT60: jyIs13 [act-5p∷GFP∷act-5 + (pRF4) rol-6(su1006)] II) was obtained from the laboratory of Emily Troemel. Strain MAH728: sid-1(qt9) V; aIxIs6[vha-6p∷sid-1∷sl2∷GFP]) was obtained from the laboratory of Malene Hansen and was used for intestine-specific RNAi treatments. Strains AGD612 (uthIS225[sur-5p∷hsf-1 CT; myo-2p∷tdTomato]), AGD1657: (unc-119(ed3) III; uthSi13[gly-19∷LifeAct∷mRuby∷unc-54 3’UTR∷cb-unc-119(+)] IV, and AGD1651: unc-119(ed3) III; uthSi7[myo-3p∷LifeAct∷mRuby∷unc-54 3’UTR∷cb-unc-119(+)] IV were obtained from the laboratory of Andrew Dillin. All other strains were generated within our laboratory, including PMD13 (egIs1[dat-1p∷GFP]; (rrf-3(b26) II)), PMD19: jyIs13 [act-5p∷GFP∷act-5 + (pRF4) rol-6(su1006)] II; rrf-3(b26) II; fem-1(hc17) IV, PMD109: jyIs13 [act-5p∷GFP∷act-5 + (pRF4) rol-6(su1006)] II; uthIS225[sur-5p∷hsf-1 CT; myo-2p∷tdTomato], PMD47: unc-119(ed3) III; utsEx2[unc-119(+) + act-5p∷GFP∷act-5 WT], PMD70: unc-119(ed3) III; utsEx4[unc-119(+) + act-5p∷GFP∷act-5 S239D], PMD84: unc-119(ed3) III; utsEx5[unc-119(+) + act-5p∷GFP∷act-5 S232D], PMD97: unc-119(ed3) III; utsEx6[unc-119(+) + act-5p∷GFP∷act-5 S232A], PMD95: (dpy-5(e907) I; sEX14231, [RcESf29f11.6∷GFP + PcEH361]; jyIs17(vha-6p∷mCherry∷act-5; ttx-3p∷RFP) IV), and PMD96: (dpy-5(e907) I; sEX10077[RcESf56c9.1∷GFP + pCeh361]; jyIs17[vha-6p∷mCherry∷act-5; ttx-3p∷RFP] IV). GFP∷ACT-5 transgenic animals were generated by injecting unc-119(ed3) mutant worms with 75 ng/μl of the unc-119 rescue plasmid and 75 ng/μl of act-5p∷GFP∷act-5 plasmid. Strains PMD95 and PMD96 contain an intestinally expressed N-terminal mCherry tagged act-5 (pET187) transgene.

Method Details

Antibodies

GFP mouse (Sigma 11814460001) and rabbit (Invitrogen A6455); tubulin mouse (Sigma T6074) and rabbit (Abcam ab4074), ACTA1 C4 monoclonal (Abcam ab3280); β-actin (Cell Signaling #4970). Antisera from rabbit raised against a synthetic ACT-5 peptide fragment (VAHDFESELAAA) was used to generate anti-ACT-5 enriched antibodies (New England Peptide). Monoclonal anti-AJM-1 (MH27) and anti-DLG1 (DLG-1) antibodies developed by R. H. Waterston and M. L. Nonet, respectively, were obtained from the Developmental Studies Hybridoma Bank (University of Iowa, Iowa City, IA). Anti-KGB-1 antibody was generated in rabbits against the KGB-1 peptide corresponding to residues 352–367 (SENRYDQEIDFADKTL) by New England Peptide (Gardner, MA).

RNAi administration

All experiments were performed on HT115 E. coli, which harbored either the L4440 empty vector plasmid as a control treatment or different RNAi constructs from either the Ahringer or Vidal RNAi libraries (Rual et al., 2004). To prevent developmental arrest due to reduced expression of ACT-5, larval worms were cultivated on HT115 harboring empty vector L4440 until reaching L3–L4 developmental stages at which time they were rinsed off plates with M9 buffer, washed twice by centrifugation at 1000 × g in a room temperature (RT) clinical centrifuge before transferring to plates possessing act-5 RNAi. For large 100 mm plates, HT115 were grown in small 2–3 ml cultures before transferring 500 μl to larger 50 ml cultures in Terrific broth and grown overnight for 15 hr. In the morning, 1 mM IPTG was supplemented directly to liquid cultures and incubated while shaking for an additional 4 hr (37°C), at which time cultures were centrifuged at 4000 × g in a bucket centrifuge for 10 min and bacterial pellets were resuspended in 5–8 ml of TB before being spread on NGM plates containing a final concentration of 1 mM IPTG and 0.1 mg/ml carbenicillin. For experiments requiring multiple RNAi combinations, the optical density (OD600) of different HT115 bacterial cultures was standardized to equalize cell number and enable ease of mixing equivalent volumes. In these cases, single RNAi treatments were equally diluted with HT115 bacteria containing empty L4440 “control” vector.

Lifespan analysis

Lifespan experiments were conducted as previously described (Baird et al., 2014). For CF512 temperature-sensitive sterile worms, lifespan assays were performed at 25°C on worms fed HT115 E. coli, using the pre-fertile period of adulthood as day 0. For N2 worms, lifespan assays were performed at 20°C and worms were transf erred to fresh plates every second day until day 10 of adulthood. In cases of excessive bag-of-worm phenotypes, a final concentration of 0.1 mg/ml of 5-fluoro-2′-deoxyuridine (FUdR) was supplemented to plates containing egg-laying day 1 adult worms. All lifespan analysis was based on a minimum of two biological repeats. Prism 8 software was used for all statistical analysis.

Dead bacteria were generated by growing large 100 ml or 300 ml cultures of HT115 E. coli harboring various RNAi constructs in TB broth supplemented with carbenicillin. IPTG was added to overnight cultures at a final concentration of 1 mM and incubated with vigorous shaking at 37°C for an additional 4 hr. Cultures were pelleted at 3700 × g for 10 min in a bucket centrifuge, resuspended in 10 ml of M9 containing 3.7% formaldehyde and incubated on a nutator at RT for 30 min. Fixed cells were pelleted at 3700 × g for 10 min and subject to 7 rounds of rinses with 25 ml of M9. After the final wash, fixed cells were concentrated in M9 or LB at a final volume of 2–5 ml to generate a dense slurry of fixed bacteria, which was subsequently plated on the same NGM plates used for living E. coli RNAi cultures. Fixed cells could be stored for 1 to 2 weeks at 4°C and still retain effective RNAi capability.

Ultrastructural analysis

For all transmission electron microscopy, age-synchronized C. elegans were fixed in 2.5% glutaraldehyde, 1% paraformaldehyde in 0.05M sodium cacodylate buffer, (pH 7.4) plus 3.0% sucrose overnight at 4°C. After several rinses in 0.1M cacodylate buffer, samples were embedded in 3% agarose and sliced into small blocks (1 mm³), rinsed with the same buffer three times and post-fixed in 1% osmium tetroxide in 0.1M cacodylate buffer for 2 hr at RT. Blocks were rinsed in buffer, then in double distilled water and en bloc stained with 2% aqueous uranyl acetate for 1 hr at RT. Next, they were dehydrated with increasing concentrations of ethanol, transitioned into Spurr’s resin with propylene oxide, infiltrated with Spurr’s resin and polymerized in a 70°C oven overnight. Blocks were sectioned with a diamond knife (Diatome) on a Leica Ultracut 7 ultramicrotome (Leica Microsystems), transferred to copper grids, and post stained with 2% aqueous uranyl acetate and lead citrate. Images were acquired on a Tecnai G² spirit transmission electron microscope (Thermo Fisher) equipped with a LaB₆ source using a voltage of 120 kV.

Fluorescence microscopy

Age synchronized C. elegans expressing different fluorescently tagged transgenes were mounted on microscope slides with M9 buffer. Samples were rinsed three times with M9 buffer to remove debris and eggs. 1 mM levamisole was used to paralyze worms prior to placing coverslips (0.15 mm, #1 thickness) over samples. Confocal microscopy was performed on an A1R system configured on an automated Nikon TiE inverted microscope using a Nikon 20X air immersion CFI Plan Apochromat VC objective with 0.75 numerical aperture (NA) and 1 mm working distance (WD). An LD laser at 488 nm and DPSS laser at 561 nm were used as light source (LU-N4 laser unit, 15 mW). A hybrid A1 scan head with resonant and galvano scanners were used to adjust laser incidence angles. Images were collected using an A1-DU4G detector unit with 2 GaAsP and 2 PMTs. Acquisition was controlled by NIS Elements imaging software and 3X optical zoom was employed during acquisition. Epifluorescence imaging was performed on an Axio Observer inverted microscope by Zeiss, using a Zeiss 20X Plan-Apochromat air immersion objective (0.8 NA, 0.55 mm WD). Images were acquired using standard filter settings for excitation and emission of fluorescence probes/proteins and recorded on a CCD camera (Zeiss AxioCam MRm). Zeiss Zen software was used to control acquisition. Fiji software was utilized for all image processing and analysis.

Indirect Immunofluorescence

Approximately 0.1–0.5 ml of C. elegans worms were fixed in 1.25 ml of 1% freshly-made fixative solution (1% paraformaldehyde solution, 80 mM KCl, 20 mM NaCl, 10 mM EDTA, 5 mM spermidine, 15 mM PIPES, pH 7.4, 25% methanol). The worm pellet emulsion was mixed well by gentle inversion for 20–120 min at RT. The emulsion was then snap frozen in liquid nitrogen and stored at −80°C for later use. Next, the frozen emulsion was thawed on ice with occasional agitation for 1–2 hr. The worms were centrifuged at 376 × g (2000 rpm) for 2 min, followed by aspiration of the supernatant. Worms were washed twice with 1X TTB (100 mM HEPES, pH 7.4, 1% Triton X-100, 1 mM EDTA) followed by re-suspension in 1 ml 1X TTB (with 1% β-mercaptoethanol added). Worms were incubated at 37°C for 1.5–2 hr with gentle shaking. Centrifugation at 376 × g for 2 min followed by aspiration of the supernatant was performed. The worms were washed once in 1 ml 1X BO₃ buffer (1 M H₃BO₃, 500 mM NaOH. Final pH 9.2), incubated in 1 ml 1X BO₃ buffer (plus 10 mM DTT) for 15 min at RT with gentle shaking, washed once in 1 ml 1X BO₃ buffer, and incubated in 1 ml 1X BO₃ buffer (plus 0.3% H₂O₂) for 15 min at RT with gentle shaking. A final wash in 1 ml 1X BO₃ buffer was performed prior to introduction of the antibody. Mounting and visualization were performed on a 2% agarose pad.

Image Quantification

All fluorescence images were analyzed and processed using Fiji – Image J software. For quantification of apical-to-cytoplasmic localization of ACT-5 protein in C. elegans, z-stacks of transgenic animals were acquired and resultant series of DIC micrographs and fluorescent images were displayed as summed projections. We selected the three central image slices within the fluorescent (GFP∷ACT-5) z-stack for the clearest demarcations of apical/luminal perimeter and sum projected the series. The apical region of interest (ROI) was drawn onto the central projected images with the Fiji line tool. The apical ROI was transposed onto the projected DIC image and the Fiji line tool was used to outline cytoplasmic regions, which juxtapose either side of the luminal region. The Fiji “measure” function was used to measure cytoplasmic and apical concentrations of fluorescent ACT-5 protein inside the manually segmented boundaries. The relationship between apical and cytoplasmic concentration of fluorescent ACT-5 protein is reported as a ratio of apical-to-cytosolic localization of GFP∷ACT-5 average signal intensity. All fluorescence images were analyzed and processed using Fiji – Image J software.

For quantification of DLG-1∷GFP localized in the first two proximal epithelial cells in C. elegans, z-stacks of transgenic animals were acquired and resultant series of DIC micrographs and fluorescent images were displayed as summed projections. Quantification of DLG-1∷GFP signal intensity was measured from sum projected images. The Fiji line tool was used to outline the ROI along the perimeter of the two proximal epithelial cells. The Fiji “measure” function was used to quantify average signal intensity along the perimeter of the epithelial cells. This measurement was compared to average signal intensity of DLG-1∷GFP along the pharyngeal terminal bulb. Accordingly, the Fiji line tool was used to outline define the ROI along the terminal bulb and the Fiji “measure” function was used to quantify signal intensity emitted by DLG-1∷GFP. The relationship between terminal bulb and proximal epithelial cells’ fluorescence was reported as a ratio of bulb-to-epithelial cell localization of DLG-1∷GFP average signal intensity.

Western blot analysis

All western blots were performed as previously described (Baird et al., 2014). Age synchronized populations of worms were cultivated on nematode growth media (NGM) agarose plates supplemented with HT115 E. coli at 20°C or 25°C for all temperature sensitive stra ins. Unless otherwise noted in the text, all western blot analysis was performed on day 3 adult animals. In brief, animals were washed off the plates with M9 buffer, centrifuged at 1000 × g for 30 sec in a RT clinical centrifuge and washed twice with M9 before being transferred to 1.5 ml Eppendorf tubes and rapidly flash frozen in liquid nitrogen.

Worm extracts were generated by glass bead disruption in non-denaturing lysis buffer [50 mM Hepes pH 7.4, 150 mM NaCl, 1mM EDTA, 1% Triton, EDTA-free mini-protease inhibitor cocktail (Roche), phosSTOP phosphatase inhibitor cocktail (Roche)]. Crude lysates were subject to centrifugation at 7500 × g at 4°C for 5 min. All extracts were subject to protein determination with a BCA protein quantification kit (Thermo Scientific). For co-immunoprecipitations, precleared worm extracts were supplemented first with 1 μl of monoclonal GFP antibody and incubated for 1 hr before adding 35 μl of 50/50 slurry of protein-G conjugated magnetic beads (Bio-Rad). After an additional 1 hr incubation, magnetic G-protein beads were magnetically precipitated and gently washed 3 separate times with 1 ml of lysis buffer before being resuspended in 2X SDS sample buffer [50mM TrisCl at pH 6.8, 2 mM EDTA, 4% glycerol, 2% SDS, Coomassie Blue, EDTA-free mini-protease inhibitor cocktail (Roche)]. For analysis of total protein, supernatants were supplemented with 2X SDS sample buffer. Samples were boiled at 90°C for 10 min, resolved by SDS-PAG E, transferred to nitrocellulose membranes and subject to western blot analysis. To better visualize GFP∷ACT-5 mobility shifts, SDS-PAGE gels (8%) were prepared the night before and stored at 4°C in pre-wet paper towels. Electrophoresis was performed at a constant 80 V until the 40 kDa protein standard of the pre-stained protein ladder (EZ-RUN, Thermo-Fisher) reached the bottom of the SDS-PAGE gels. All antibodies were prepared in 5% BSA/PBST. Mouse monoclonal anti-GFP was used at 1:5000, rabbit polyclonal anti-GFP at 1:5000, mouse and rabbit anti-tubulin were used at 1:10000–1:20000. Mouse and rabbit anti-actin were used at 1:10000–1:15000. Anti-ACT-5 was used at 1:5000, anti-DLG-1 at 1:2000, anti-KGB-1 at 1:3000 and anti-GSP-1 at 1:2000.

Western blots were quantified using Image Studio software (LI-COR Biosciences, Lincoln, NE). Quantified bands of interest were standardized based on band signal intensity of either tubulin, GFP∷ACT-5, or total actin (C4).

Transcriptomic and qPCR analysis

For quantitative PCR, C. elegans strain CF512 (rrf-3(b26) II; fem-1(hc17) IV) was grown on L4440 RNAi at 25°C to prevent offspring producti on. The worms were age synchronized, harvested in TRIzol® (ThermoFisher Scientific) at various developmental stages (L4, D1, D3, D5, D10, D12), flash-frozen in liquid nitrogen, and stored at −80°C. For total RNA extraction, worms were freeze-thawed 3 times, followed by a chloroform/isopropanol extraction process. The RNA pellets were washed twice with ethanol, allowed to air-dry, and subsequently reconstituted in 20 μl ddH₂O. RNA quantification was performed using the DeNovix spectrometer (DS-11 FX+); 260/280 and 260/230 ratios were used to assess RNA quality. cDNA synthesis was conducted using the QuantiTect Reverse Transcription kit (Qiagen) as per the manufacturer’s guidelines. qPCR was performed with the iTaq™ Universal SYBR® Green Supermix kit (Bio-Rad) and reactions were run using the CFX384 Real Time System (Bio-Rad). Primer sequences were designed to target genes of interest. See also Table S2. These include:

• act-5_for 5′-ACCACCGGAATCGTTTTGGA-3′,

• act-5_rev 5′-GGCATGTGGGAGGGCATATC-3′,

• hsf-1_for 5′- ATGCGTGCGATGCGAGAAAA-3′,

• hsf-1_rev 5′-GCGACACGCTTCGACAATC-3′,

• kgb-1_for 5′- GACGATGAGGTAAACGCCCC-3′,

• kgb-1_rev 5′- GTGAAAATGTCGTGGTCGGC-3′,

• vhp-1_for 5′-AATGTCTCCGATCATCCGGC-3′,

• vhp-1_rev 5′-CGTCGAGCAAATTGACACCG-3′,

• gsp-1_for 5′-ACAAATTCGCCGTGTGATGC-3′,

• gsp-1_rev 5′-TCCCCATCCGGTAACATCCT-3′,

• gsp-2_for 5′-TGGACCGAGGGAAACAATCG-3′,

• gsp-2_rev 5′-CCGCTTGCACTCGTCATAGA-3′,

• lys-3_for 5′-TTGCACCAATGGCTGTGAGA-3′,

• lys-3_rev 5′-TCCAGCCTCCTGTGATTTCC-3′

For each gene, 3 independent biological repeats and 3 technical repeats were included in the analysis. The ΔCt method was used to calculate the relative transcriptional abundance of the target genes and 2 housekeeping genes, alpha tubulin (tba-1) and a putative iron-sulfur cluster assembly enzyme (Y45F10D.4), were utilized for normalization of the transcriptional data.

RNA-sequencing datasets (Brunquell et al., 2016) were obtained from SRA database (Accession number SRP078295) and contained two replicates per condition: empty vector RNAi +/− heat shock and hsf-1 RNAi +/− heat shock. Successful reads were mapped using Qiagen Bioinformatics CLC Genomics Workbench (https://www.qiagenbioinformatics.com/products/clcgenomics-workbench/) version 9.5 against Caenorhabditis elegans genome (WBcel235) from Ensembl. RNA-seq analysis was performed at the gene level using CLC Genomics Workbench version 9.5 with these settings: mismatch cost = 2, insertion cost = 3, deletion cost = 3, length fraction = 0.8, similarity fraction = 0.8, and max number of hits per read = 10. Expression values were set as RPKM, and expectation-maximum estimation was used. Statistical analysis of RNA-seq data was carried out in CLC Genomics Workbench at the gene level using the proportions-based Baggerly’s test on total counts, and p-values are displayed as FDR adjusted p-values. FDR p-values ≤0.05 are considered to be significant.

Mass spectrometry

In-Gel Digestion:

Immunoprecipitated protein samples were separated on SDS-PAGE and stained with GelCode™ Blue Stain (Thermo Fisher Scientific, San Jose, CA) according the manufacturer’s protocol. Protein bands from each lane were then cut into slices. Each excised gel slices was then further chopped into 1 mm cubes. In-gel digestion was performed following the protocol below. The gel pieces were destained with 50 mM triethylammonium bicarbonate (TEAB)/acetonitrile (1:1, v/v) by incubation at 37°C for 30 min. Gel pieces were dehydrated with acetonitrile at RT, followed by reduction/alkylation using DTT and iodoacetamide. Gel pieces were then dehydrated with acetonitrile and rehydrated with trypsin solution (200 ng/μl in 50 mM TEAB). Trypsin digestion was carried out at 37°C overnight. Peptides were extracted after 30 min incubation at 37°C with extraction buffer to a final concentration of 66.7% acetonitrile and 5% Trifluoroacetic Acid (TFA). All steps were carried out on a thermomixer shaker (Eppendorf, NJ) unless stated otherwise. Extracts were dried in vacuum centrifuge. Salts were removed using Oasis HLB Elution plate (Waters, MA) before LC-MS/MS analysis.

Liquid chromatography and tandem mass spectrometry:

Mass spectrometry data were collected using an Orbitrap Fusion mass spectrometer (Thermo Fisher Scientific, San Jose, CA) coupled to a coupled to an Ultimate 3000 RSLCnano HPLC system (Dionex). Peptides were loaded onto a 75 μm × 50 cm, 2m Easy-Spray column (Thermo) and eluted with a gradient from 1–28% buffer B over 120 min at 250 nl/min. Buffer A contained 2% (v/v) acetonitrile (ACN) and 0.1% formic acid in water, and buffer B contained 80% (v/v) ACN, 10% (v/v) trifluoroethanol, and 0.08% formic acid in water. The scan sequence began with an MS1 spectrum (Orbitrap analysis; resolution 120,000; mass range 400–1600 m/z; automatic gain control (AGC) target 4 × 10⁵; maximum injection time 50 ms). Precursors for MS2 analysis were selected using a Top Speed of 3 sec. MS2 analysis consisted of HCD (quadrupole ion trap analysis; 15K resolution; AGC 1× 10⁴; normalized collision energy (NCE) 33%; maximum injection time 100 ms).

Data Analysis:

Data from the Orbitrap Fusion were processed using Proteome Discoverer Software (ver. 2.1.0.62). MS2 spectra were searched using Sequest HT and MS Amanda against C.elegans UniProt database. Search parameters were specified as: trypsin enzyme, 3 missed cleavages allowed, minimum peptide length of 6, precursor mass tolerance of 10 ppm, and a fragment mass tolerance of 0.2 Da. Carbamidomethylation of cysteine residues (+57.021 Da) were set as static modifications, while oxidation of methionine residues (+15.995 Da), ubiquitination (+GG; +114.043 Da) and phosphorylation (+79.966 Da) was set as a dynamic modification. Peptide-spectral matches (PSMs) error rates were determined using the target-decoy strategy coupled to Percolator modeling of positive and false matches (Kall et al., 2007; Spivak et al., 2009).

Label-free quantification:

Quantification was performed using MaxQuant (ver. 1.5.8.3) against the C. elegans database, which was concatenated with their reversed sequences as decoys to determine FDR. Searches restricted the precursor ion tolerance to 10 ppm, and product ion tolerance window was set to 0.7 m/z. Searches allowed up to two missed cleavages, included static carbamidomethylation of cysteine residues (+57.021 Da), and variable oxidation of methionine residues (+15.995 Da). The FDR was set to 1% on peptide spectrum match (PSM), PTM site and Protein level. MaxQuant applies a target-decoy search strategy to estimate and control the extent of false-positive identifications using the concept of posterior error probability (PEP) to integrate multiple peptide properties, such as length, charge, number of modifications and Andromeda score into a single quantity reflecting the quality of a PSM. Match-between-runs (MBR) was enabled with default settings across the samples (match time window = 0.5 min, alignment time window = 15). Quantification was via MaxQuant’s LFQ algorithm which combines and adjusts peptide intensities into a protein intensity value.

Large-particle flow cytometry

Flow cytometry of C. elegans was performed on a COPAS FP-250 flow cytometer (Union Biometrica, Holliston, MA) using either the built-in sample cup or an attached LP Sampler (Union Biometrica) fitted with 96-well plates. M9 buffer was utilized as the sample solution for worm flow, and the sheath solution was composed of a proprietary recipe, COPAS GP SHEATH REAGENT (PN: 200-5070-100, Union Biometrica). Flow data was collected in FlowPilot software (Union Biometrica). Extinction signal was collected using a 488 nm laser with a 1.3ND filter and a gain of 1.0. GFP and mCherry were excited using 488 and 561 nm lasers, respectively. Fluorescence signal gains were set to 2.0 and the PMT voltage was varied to acquire signal intensity within the limits of detection of the instrument, though kept consistent for every sample within an experiment. For experiments pertaining to LifeAct fluorescence, a data collection threshold of 256 worms per sample was employed.

Worm length normalization and profiling

WormProfiler, an in-house program written for the MATLAB runtime, was used to standardize worm lengths and calculate averages for fluorescence intensity profiles from a worm population. To standardize worm lengths, time-of-flight (TOF) data from individual worms analyzed by the COPAS biosorter were transformed into vectors of length 100, where each data point represents a 1% section of the total worm length. Fluorescence values were likewise downsampled by a moving-average filter into length-100 vectors. For average values, the mean of a fluorescent section was calculated from the worm sample population. Neuronal dat-1p∷GFP fluorescence was utilized to orient each individual worm, where the head was identified by the highest peak GFP signal by flow cytometry.

Pathogenesis assay

Bacterial cultures (HT115) were grown as described above. In brief, overnight TB cultures were inoculated with an 8 hr starter culture at 1:1000 dilution. All bacteria were cultured under carbenicillin selection. Following overnight culturing at 37°C, cultures were induced with 1 mM IPTG for 4 hr at 37°C, collected, and concentrat ed 10x by centrifugation. Concentrated RNAi cultures were mixed at a ratio of 4:1 with concentrated HT115 bacteria driving mCherry expression from the pDP151 plasmid (4 parts HT115 RNAi culture, 1 part HT115 expressing mCherry) before seeding 500 μl each onto 100 mm RNAi plates and allowed to dry at RT for 24 hr. Age-synchronized PMD13 worms were obtained by hypochlorite treatments of gravid adults to obtain eggs, which were added to seeded RNAi plates at a density of 1000 eggs per plate. These worms were grown at 25°C and analyzed by larg e-particle flow cytometry and fluorescence microscopy at day 1, 3, 5, and 10 of adulthood. Prior to analysis, worms were gently removed from the RNAi/mCherry agar plates, washed three times with M9 buffer, and placed on OP50 plates to feed for 2 hr to remove residual fluorescent bacteria from the gut.

To analyze bacterial invasion or colonization by large-particle flow, the worms were rinsed off the OP50 plates, washed 2x with M9, and transferred in 100 μl of M9 per well to a 96-well plate. Each worm sample in the 96 well plate was separated by an empty well of M9 to avoid cross-contamination during sample analysis by flow. The 96 well plate was loaded onto an LP sampler and analyzed by large particle flow cytometer. Fluorescence images were obtained from animals anesthetized in 250 mM NaN₃ on an unseeded NGM plate and imaged on a Zeiss Axio Zoom.V16 (Zeiss, Oberkocken, Germany) at 50X magnification.

Sytox staining of dead worms

Adapting from (Gill et al., 2003), CF512 worms were cultured on 100 mm NGM agar plates, which contained HT115 E. coli food harboring the “control” L4440 plasmid. At the indicated ages, worms were gently rinsed off the agar plates with liquid M9 media and washed twice in the same buffer by centrifugation at 1000 × g in a RT clinical centrifuge. Worms in M9 suspension were supplemented with SYTOX green (Molecular Probes) at a final concentration of 1 μM before incubation at RT for 15 min prior to flow cytometry.

Quantification and Statistical Analysis

Statistical details of all experiments, including exact values of n (number of biological repeats and number of technical repeats), type of statistical test performed (e.g. one or two-way ANOVA, multiple comparisons test, Fisher’s exact test, student’s t-test, etc.), the use of precision measures (e.g. SEM or SD), and the definition of significance for each experiment can be found in the figure legends, figures, and results sections. Lifespan values are represented as the mean of all median lifespans from each biological repeat. Statistics on lifespan analyses can be found in Table S1. GraphPad Prism 8 software was utilized for all statistical analyses. All t-tests were performed as unpaired, two-tailed tests. For ANOVA tests, post-hoc analysis was performed using Tukey’s multiple comparisons test. Adjusted p-values ≤ 0.05 were considered significant.

Data and Code Availability

The code “WormProfiler” generated in this study for analysis of profiles by large-particle flow cytometry is available at GitHub (https://github.com/MadGradLad/WormProfiler).

KEY RESOURCES TABLE

REAGENT or RESOURCE	SOURCE	IDENTIFIER
Antibodies
Mouse monoclonal anti-GFP (mixture of clones 13.1 and 7.1)	Sigma-Aldrich	Cat#11814460001 (Roche), RRID: AB_390913
Rabbit polyclonal anti-GFP	Invitrogen	Cat#A-6455, RRID: AB_221570
Mouse monoclonal anti-α tubulin (clone B-5-1-2)	Sigma-Aldrich	Cat# T6074, RRID: AB_477582
Rabbit polyclonal anti-α tubulin	Abcam	Cat# ab4074, RRID: AB_2288001
Mouse monoclonal anti-actin [clone ACTN05 (C4)]	Abcam	Cat#ab3280, RRID: AB 303668
Rabbit monoclonal anti-β actin (clone 13E5)	Cell Signaling	Cat#4970, RRID: AB_2223172
Rabbit polyclonal anti-ACT-5	This paper	N/A
Mouse monoclonal anti-AJM-1	DSHB	Cat#MH27, RRID: AB_531819
Mouse monoclonal anti-DLG-1	DSHB	Cat#DLG1, RRID: AB_2617529
Rabbit polyclonal anti-KGB-1	This paper	N/A
Bacterial and Virus Strains
HT115 (DE3) Escherichia coli	Rual et al., 2004	CGC Cat# HT115(DE3), RRID:WB-STRAIN:HT115(DE3)
OP50 Escherichia coli	Brenner, 1974	CGC Cat# OP50, RRIDWB-STRAIN:OP50
Biological Samples
None used	N/A	N/A
Chemicals, Peptides, and Recombinant Proteins
Sodium cacodylate buffer, 0.2M, pH7.4	Electron Microscopy Sciences	Cat#11650
Osmium tetroxide	Electron Microscopy Sciences	Cat#19100
Uranyl acetate solution, 4%	Electron Microscopy Sciences	Cat#224004
Spurr’s resin	Electron Microscopy Sciences	Cat#14300
Propylene oxide	Electron Microscopy Sciences	Cat#20401
Lead citrate	Electron Microscopy Sciences	Cat#17800
Levamisole, 99%	Acros Organics	Cat#187870100
Spermidine, 99%	Acros Organics	Cat#132740010
PIPES, 98.5%	Fisher Scientific	Cat#172610250
Complete, EDTA-free mini protease inhibitor cocktail	Sigma-Aldrich	Cat#11836170001
PhosSTOP phosphatase inhibitor tablets	Sigma-Aldrich	Cat#4906837001
EZ-Run prestained rec protein ladder	Thermo Fisher Scientific	Cat#BP36031
Lambda protein phosphatase	New England Biolabs (NEB)	Cat#P0753S
SureBeads protein G magnetic beads	Bio-Rad	Cat#161–4023
GelCode™ blue stain	Thermo Scientific	Cat#24592
COPAS GP sheath reagent	Union Biometrica	Cat#300-5070-100
SYTOX green nucleic acid stain	Thermo Fisher Scientific	Cat#S7020
(+)-5-fluoro-2′-deoxyuridine (FUdR)	Fisher Scientific	Cat# AC227601000
Critical Commercial Assays
QuantiTect Reverse Transcription kit	Qiagen	Cat#205313
iTaq Universal SYBER Green Supermix kit	Bio-Rad	Cat#1725124
Deposited Data
None	N/A	N/A
Experimental Models: Cell Lines
None	N/A	N/A
Experimental Models: Organisms/Strains
C. elegans: Strain N2 (ancestral) as wild type (WT)	Caenorhabditis Genetics Center	CGC Cat# N2 (ancestral), RRID: WBSTRAIN:N2_(ancestral)
C. elegans: Strain CF512: rrf-3(b26) II; fem-1(hc17) IV	Caenorhabditis Genetics Center	CGC Cat# CF512, RRID:WB-STRAIN:CF512
C. elegans: Strain: ERT60: jyIs13 [act-5p::GFP::act-5 + (pRF4) rol-6(su1006)] II	Caenorhabditis Genetics Center	CGC Cat# ERT60, RRID:WB-STRAIN:ERT60
C. elegans: Strain PMD19: jyIs13 [act-5p::GFP::act-5 + (pRF4) rol-6(su1006)] II; rrf-3(b26) II; fem-1(hc17) IV	This paper	N/A
C. elegans: Strain PMD47: unc-119(ed3) III; utsEx2[unc-119(+) + act-5p::GFP::act-5 WT]	This paper	N/A
C. elegans: Strain PMD70: unc-119(ed3) III; utsEx4[unc-119(+) + act-5p::GFP::act-5 S239D]	This paper	N/A
C. elegans: Strain PMD84: unc-119(ed3) III; utsEx5[unc-119(+) + act-5p::GFP::act-5 S232D]	This paper	N/A
C. elegans: Strain: PMD97: unc-119(ed3) III; utsEx6[unc-119(+) + act-5p::GFP::act-5 S232A]	This paper	N/A
C. elegans: Strain: PMD95: (dpy-5(e907) I; sEX14231, [RcESf29f11.6::GFP + PcEH361]; jyIs17(vha-6p::mCherry::act-5; ttx-3p::RFP) IV)	This paper	N/A
C. elegans: Strain PMD96: (dpy-5(e907) I; sEX10077[RcESf56c9.1::GFP + pCeh361]; jyIs 17[vha-6p::mCherry::act-5; ttx-3p::RFP] IV)	This paper	N/A
C. elegans: Strain ML2615: dlg-1(mc103[dlg-1::GFP]) X	Caenorhabditis Genetics Center	CGC Cat# ML2615, RRID:WB-STRAIN:ML2615
C. elegans: Strain AGD794: uthIS225[sur-5p::hsf-1 CT; myo-2p::tdTomato]	Baird et al., 2014.	CGC Cat# AGD794
C. elegans: Strain PMD13: egIs1[dat-1p::GFP]; rrf-3(b26) II	This paper	N/A
C. elegans: Strain AGD1657: (unc-119(ed3) III; uthSi13[gly-19::LifeAct::mRuby::unc-54 3’UTR::cb-unc-119(+)] IV	Higuchi-Sanabria et al., 2018	N/A
C. elegans: Strain AGD1651: unc-119(ed3) III; uthSi7[myo-3p::LifeAct::mRuby::unc-54 3’UTR::cb-unc-119(+)] IV	Higuchi-Sanabria et al., 2018	N/A
C. elegans: Strain MAH728: sid-1(qt9) V; aIxIs6[vha-6p::sid-1::sl2::GFP)	Laboratory of Malene Hansen	Ref: MAH728
C. elegans: Strain PMD109: jyIs13 [act-5p:-GFP::act-5 + (pRF4) rol-6(su1006)] II; uthIS225[sur-5p::hsf-1 CT; myo-2p::tdTomato]	This paper	N/A
Oligonucleotides
Refer to Table S2 for qPCR primer list	This study	N/A
Refer to Table S2 for RNAi oligonucleotide list	Ahringer and Vidal Libraries	N/A
Recombinant DNA
None	N/A	N/A
Software and Algorithms
ZEN Software (ver. 2.3)	Carl Zeiss Microscopy, LLC	https://www.zeiss.com/microscopy/us/products/microscope-software/zen.html
Fiji Software	ImageJ	https://imagej.net/Fiji
GraphPad Prism (ver. 8.0.0 for Windows)	GraphPad Software	https://www.graphpad.com
CLC Genomics Workbench (ver. 9.5)	Qiagen Bioinformatics	https://www.qiagenbioinformatics.com/products/clc-genomics-workbench/
Proteome Discoverer Software (ver. 2.1.0.62)	Thermo Fisher Scientific	https://www.thermofisher.com/order/catalog/product/OPTON-30795
MaxQuant (ver. 1.5.8.3)	Computational Systems Biochemistry (Prof. Jurgen Cox)	https://www.maxquant.org/
FlowPilot Software	Union Biometrica	https://www.unionbio.com/copas/
WormProfiler (ver.1.2)	This paper	https://github.com/MadGradLad/WormProfiler
Other
None	N/A	N/A

Highlights.

1. The low-abundant, intestine-specific actin (ACT-5) affects animal aging

2. ACT-5 phosphorylation within its troponin binding site destabilizes actin networks

3. The Jun kinase (KGB-1) and PP1 phosphatase (GSP-1) regulate ACT-5 phosphorylation

4. Repression of KGB-1 by HSF-1 impacts the intestinal barrier and pathogenesis