Giant KASH proteins and ribosomes establish distinct cytoplasmic biophysical properties in vivo

Abstract

Understanding how cells control their biophysical properties during development remains a fundamental challenge. While macromolecular crowding affects multiple cellular processes in single cells, its regulation in living animals remains poorly understood. Using genetically encoded multimeric nanoparticles for in vivo rheology, we found that Caenorhabditis elegans tissues maintain mesoscale properties that differ from those observed across diverse systems, including bacteria, yeast species, and cultured mammalian cells. We identified two conserved mechanisms controlling particle mobility: Ribosome concentration, a known regulator of cytoplasmic crowding, works in concert with a previously unknown function for the giant KASH (Klarsicht/ANC-1/SYNE homology) protein ANC-1 in providing structural constraints through associating with the endoplasmic reticulum. These findings reveal mechanisms by which tissues establish and maintain distinct mesoscale properties, with implications for understanding cellular organization across species.

Living tissues maintain unique intracellular biophysical properties under the control of cytoplasmic constraints and crowding.

INTRODUCTION

The cytoplasm is a complex biomolecular environment consisting of a protein-rich aqueous phase (the cytosol), dynamic networks of cytoskeletal filaments, and membrane-bound organelles. Macromolecules occupy up to 40% of cytoplasmic volume, creating a crowded environment that fundamentally alters diffusion, chemical reactions, and phase separations within cells (1–8). This macromolecular crowding (referred to as “crowding” from here on) influences essential cellular processes including protein folding, metabolism, and signal transduction (9, 10). Beyond crowding effects, molecular movement in the cytoplasm is further restricted by physical barriers and interactions. These include steric hinderance from the cytoskeletal network and organelles, as well as specific binding interactions with macromolecular complexes or molecular tethers. Together, crowding and these physical constraints define the biophysical properties of the cytoplasm and influence how molecules and complexes move through it. In this study, we specifically examine mesoscale motion—the behavior of macromolecular complexes measuring tens of nanometers in diameter. This intermediate-scale mobility is uniquely determined by both molecular crowding and structural hindrances, distinguishing it from the dynamics of small molecules (few nanometers) or large organelles (hundreds to thousands of nanometers).

Despite their importance, the biophysical properties of the cytoplasm in living multicellular organisms remain poorly understood. Traditional approaches for studying cytoplasmic biophysical properties rely on passive microrheology (9, 11), where the motion of nonbiological tracer particles is used to infer cytoplasmic dynamics, viscosity, elasticity, and structure. However, such studies have primarily focused on isolated cultured cells. The application of this technique to multicellular tissues has been limited by challenges in delivering probes into cells without disrupting cellular function, resulting in most studies being restricted to early embryos (12). Thus, there is a substantial gap in our knowledge of the mechanisms that determine cytoplasmic biophysical properties in complex, multicellular organisms.

Here, we investigate cytoplasmic biophysical properties in living multicellular animals using genetically encoded multimeric nanoparticles (GEMs) (2). These bright fluorescent tracers of defined shape and size serve as rheological tools across diverse organisms, including bacteria, cultured mammal cells, and yeast (2, 5, 13). We adapted GEMs for expression in the hypodermis and intestine of Caenorhabditis elegans, combining in vivo rheology with genetic manipulations to uncover mechanisms controlling cytoplasmic biophysical properties through crowding and constraints.

RESULTS

Cytoplasmic mesoscale motion is highly constrained in C. elegans tissues

To investigate mesoscale cytoplasmic biophysical properties in living tissue, we engineered C. elegans strains with single-copy cDNA insertions encoding enhanced green fluorescent protein (EGFP)–tagged 40-nm-diameter GEMs expressed in the hypodermis or intestine using the semo-1 or ges-1 promoters, respectively (Fig. 1A) (14). These cytoplasmic nanoparticles were readily visible in both tissues (Fig. 1B and movies S1 and S2), and transgenic animals exhibited normal development, fertility, and viability (fig. S1). We analyzed GEM motion in the anterior body region between the pharynx and oocytes of adult worms, capturing single-particle trajectories at 50 Hz using spinning disc confocal microscopy (Fig. 1B). From thousands of trajectories per tissue (minimum of 10 frames), we calculated ensemble mean squared displacement (MSD) and effective diffusion coefficients (D_eff) (Fig. 1, C and D). There was no correlation between particle intensity and D_eff (fig. S1, E and F), suggesting that we measured fully assembled 40-nm GEMs and not a population of partially assembled GEMs.

Fig. 1. C. elegans tissues maintain highly crowded and constrained cytoplasmic environments.

(A) GEM expression cassette and genomic integration. A tissue-specific promoter (P_ges-1 for the intestine or P_semo-1 for the hypodermis) drives the expression of a GEM-EGFP–encoding transgene integrated at the RMCE landing site IVa (jsTi1493) on chromosome IV. (B) Inverted grayscale spinning disc confocal images of GEMs in the intestine (top) and hypodermis (bottom). Left: representative fields of view (scale bar, 10 μm) with magnified insets (scale bar, 5 μm). Right panels display particle trajectories color-coded by time (scale bar, 4 μm; concentric circles indicate 2- and 4-μm radii). (C and D) MSD analysis of GEM mobility in the intestine and hypodermis shown in linear scale (C) with indicated diffusion coefficient D_eff and log-log (D) scale with the indicated anomalous exponent (α). Individual particle trajectories (orange, intestine; blue, hypodermis) and ensemble average MSD (black lines) demonstrate restricted diffusion. Red dotted lines represent normal diffusion (α = 1). (E) 2D probability density histograms correlating D_eff with α for intestinal (left) and hypodermal (right) GEMs. Marginal distributions shown as histograms on axes. Red curves indicate two populations identified by Gaussian mixture modeling (GMM). Color intensity represents population density, revealing distinct mobility populations in both tissues. D_eff values mentioned hereafter represent diffusion over the initial 100 ms of particle motion.

GEMs in C. elegans tissues exhibited remarkably slow diffusion, with an ensemble D_eff below 0.02 μm²/s. This is more than 10-fold slower than what was previously observed across diverse single-cell systems, including budding and fission yeasts, bacteria, and cultured mammalian cells (human embryonic kidney 293 cells and mouse embryonic fibroblasts) (2, 5, 13). The MSD curves obtained from both tissues showed slopes with an anomalous diffusion exponent (α) below 1 (indicated by the dashed line) (Fig. 1, C and D), revealing subdiffusive behavior and suggesting highly constrained GEM mobility. We also observed that while most GEMs exhibited subdiffusive behavior (α < 1), a small subset of particles displayed α values slightly above 1. These apparent superdiffusive values likely reflect variability or noise in individual time-averaged MSD trajectories, especially when fit on a log-log scale. In all cases, the α distribution is centered near 1 for the unconstrained population and well below 1 for the constrained population, consistent with freely diffusing particles rather than directional or active transport. To understand the basis for this slow and constrained diffusion, we analyzed the relationship between D_eff and α. This revealed heterogeneous diffusion patterns consistent with cytoplasmic nonergodicity, with most particles displaying constrained movement. Using Gaussian mixture modeling (GMM) (15), we identified two distinct GEM populations: locally constrained particles and unconstrained particles (indicated by fitted curves; Fig. 1E). Notably, even the unconstrained GEMs diffused substantially slower than in single-cell systems (2, 16). These findings reveal that C. elegans tissues maintain an exceptionally crowded and constrained cytoplasmic state.

ANC-1 establishes size-dependent constraints on cytoplasmic mesoscale diffusion

Giant Klarsicht/ANC-1/SYNE homology (KASH) proteins are massive molecular tethers that span the outer nuclear membrane and extend into the cytoplasm, where they interact with various cytoskeletal networks critical for cellular architecture (17, 18). In C. elegans, the giant KASH protein ANC-1 stabilizes cytoplasmic dynamics by greatly reducing organelle displacement and maintains cellular integrity in the hypodermis (19). Similarly, disruption of KASH proteins in mammalian cells by a dominant negative approach reduces cellular stiffness (20). On the basis of these observations, we hypothesized that ANC-1 constrains mesoscale macromolecular motion, potentially explaining the remarkably slow diffusion we observed in C. elegans tissues. Supporting this hypothesis, GEM diffusion increased significantly in both intestinal and hypodermal tissues of anc-1(e1873) null mutants (Fig. 2A, fig. S2, and movies S1 and S2). Detailed analysis of individual trajectories revealed that ANC-1 loss altered both the distribution and mesoscale dynamics of GEM populations (Fig. 2, C to F). Bootstrap analyses demonstrated two key changes: a substantial decrease in the fraction of constrained particles and significantly faster diffusion of the unconstrained population compared to wild type (Fig. 2, G to J).

Fig. 2. ANC-1 maintains cytoplasmic constraints in the hypodermis and intestine.

(A and B) Representative inverted grayscale spinning disc confocal images showing GEM mobility in intestinal (A) and hypodermal (B) cells in wild-type or anc-1(e1873) animals. Left: representative fields of view (scale bar, 10 μm) with magnified insets (scale bar, 5 μm). Right panels display particle trajectories color-coded by time (scale bar, 4 μm; concentric circles indicate 2- and 4-μm radii). (C and E) Heatmaps correlating D_eff with α for GEMs in intestinal (C) and hypodermal (E) cells in wild-type or anc-1(e1873) animals. Marginal histograms show individual parameter distributions. Color intensity represents particle density. (D and F) GMM of GEM D_eff in the intestine (D) and hypodermis (F) reveals distinct mobility populations. Orange, wild type; green, anc-1(e1873). Dotted lines indicate population means. (G to J) Bootstrap analysis of GMM parameters showing mean distributions for intestinal [(G) and (H)] and hypodermal [(I) and (J)] cells. [(G) and (I)] Percentage of constrained GEMs. [(H) and (J)] D_eff of unconstrained GEMs. Statistical significance assessed by the Kruskal-Wallis test (****P < 0.0001).

Notably, structural constraints are size-dependent, where smaller proteins may diffuse relatively unhindered, while larger macromolecules experience confinement (21). Thus, to determine whether ANC-1’s effects were specific to large macromolecules, we performed fluorescence recovery after photobleaching (FRAP) experiments using cytoplasmic EGFP (~4 nm in diameter) expressed in the hypodermis (fig. S2) (22). Recovery times and diffusion rates were identical between wild-type and anc-1(e1873) animals, demonstrating that ANC-1 specifically constrains the motion of larger macromolecules while allowing free diffusion of smaller proteins. These findings reveal an unexpected role for ANC-1 in establishing size-dependent constraints on cytoplasmic diffusion.

Ribosomes and ANC-1 independently determine mesoscale mobility through distinct mechanisms

Ribosomes are key determinants of cytoplasmic crowding in cultured yeast and mammalian cells (2, 23, 24). Our forward genetic screen revealed that knockdown of the ribosomal subunit rps-15 caused nuclear positioning defects in the syncytial hypodermis (fig. S3), prompting us to investigate how ribosome levels influence mesoscale mobility in vivo. RNA interference (RNAi) against two different ribosomal subunits (rps-15 or rps-18) significantly increased GEM diffusion while preserving the relative proportions of constrained and unconstrained populations (Fig. 3, A, B, and D, and fig. S3). This selective effect on mesoscale diffusion, without altering constraint patterns, suggests that ribosomes primarily influence cytoplasmic crowding rather than impose structural constraints.

Fig. 3. Ribosome depletion and ANC-1 loss affect cytoplasmic mobility through distinct mechanisms.

(A) Representative inverted grayscale spinning disc confocal images showing intestinal GEM mobility under control(RNAi), rps-15(RNAi), and rps-18(RNAi) conditions. Left: representative fields of view with magnified insets (scale bars, 10 and 5 μm, respectively). Right panels display particle trajectories color-coded by time (scale bar, 4 μm; concentric circles indicate 2- and 4-μm radii). (B) Heatmaps correlating D_eff with α for intestinal GEMs under ribosomal protein knockdown conditions. Marginal histograms show parameter distributions. (C) Similar analysis in the anc-1(e1873) background with ribosomal protein knockdowns. (D) GMM analysis of GEM D_eff comparing control(RNAi) (olive), rps-15(RNAi) (blue), and rps-18(RNAi) (coral) knockdowns. Dotted lines indicate mean D_eff values for control and rps-15(RNAi) populations. (E) Similar analysis in the anc-1(e1873) background comparing control(RNAi) (green), rps-15(RNAi) (purple), and rps-18(RNAi) (pink) knockdowns. Dotted lines indicate mean D_eff values for anc-1(e1873); control(RNAi) and anc-1(e1873); rps-15(RNAi) populations. (F and G) Bootstrap analysis of GMM parameters comparing ribosomal protein knockdown conditions under wild-type or anc-1(e1873) backgrounds. The percentages of constrained GEMs (F) with mean ± SD D_eff values of unconstrained GEMs (G) are shown. Statistical significance was assessed by the Kruskal-Wallis test with Dunn’s multiple comparisons (****P < 0.0001).

To examine the relationship between ribosome-mediated mesoscale crowding and ANC-1–dependent constraints, we depleted ribosomes in anc-1(e1873) tissues (Fig. 3A). The double perturbation revealed distinct roles for each regulator: The shift toward unconstrained GEMs matched that of anc-1 single mutants, while diffusion coefficients reflected patterns seen with ribosome depletion alone (Fig. 3, C and E to G, and fig. S3). However, interpretation of absolute diffusion rates requires caution, as baseline mobility differs between standard and RNAi feeding conditions (see Materials and Methods). These findings suggest that ANC-1 and ribosomes affect cytoplasmic mesoscale macromolecular mobility through separate mechanisms—ANC-1 primarily establishing structural constraints and ribosomes modulating crowding-dependent mesoscale diffusion.

The transmembrane α helix of ANC-1 is essential for mesoscale cytoplasmic constraints

ANC-1 is a large protein (~800 kDa) composed of multiple conserved domains: N-terminal tandem calponin homology (CH) domains that bind actin, a central region containing six tandem repeats (RPs) of ~900 residues each predicted to form spectrin-like structures, and a C-terminal KASH domain with its transmembrane α helix that localizes to the nuclear envelope and endoplasmic reticulum (ER) membranes (Fig. 4A) (19, 25). Having established ANC-1’s role in regulating cytoplasmic mesoscale constraints, we investigated its mechanism of action. Our findings revealed that ANC-1 maintains cytoplasmic organization through a pathway independent of its canonical role in the LINC (linker of nucleoskeleton and cytoskeleton) complex.

Fig. 4. ANC-1’s transmembrane α helix maintains cytoplasmic constraints and ER organization.

(A) AlphaFold structural prediction of ANC-1 showing the domain architecture, including the N-terminal actin-binding domain (ABD; pink), six RP domains (RP1 to RP6; alternating blue and purple), and C-terminal transmembrane (T) domain that spans both the ER and nuclear envelope membranes with the KASH (K) peptide extending into the perinuclear space. The linear domain diagram above shows the corresponding domain organization. (B) Representative inverted grayscale spinning disc confocal images showing GEM mobility in intestinal and hypodermal tissues of anc-1(∆TK) mutants. Left: representative fields of view with magnified insets (scale bars, 10 and 5 μm, respectively). Right panels display particle trajectories color-coded by time (scale bar, 4 μm; concentric circles indicate 2- and 4-μm radii). (C) Heatmaps correlating D_eff and α for intestinal and hypodermal GEMs in anc-1(∆TK) mutants. Marginal histograms show parameter distributions. (D and E) GMM of GEM D_eff values comparing anc-1(e1873) (green) with anc-1(∆TK) mutants (beige) in the intestine (D) or hypodermis (E). Dotted lines indicate population means. (F to I) Bootstrap analysis of GMM parameters across genotypes. Percentage of constrained GEMs in the intestine (F) and hypodermis (G) and data shown with means ± 95% confidence interval (CI). D_eff values of unconstrained GEMs in the intestine (H) and hypodermis (I). (J) Representative spinning disc confocal images of hypodermal ER organization. A single-copy insertion of mKate2::tram-1 was used to visualize the ER (see Materials and Methods). Left: ER signal; right: thresholded images showing the ER outline (cyan) and hypodermal boundary (red). Scale bar, 20 μm. Zoom-in insets are magnified views of the boxed regions. (K to N) ER network analysis: ER occupancy percentage [(K) and (L)] and perimeter complexity [(M) and (N)] across genotypes. Statistical significance was assessed by the Kruskal-Wallis test with Dunn’s multiple comparisons (**P ≤ 0.01; ****P ≤ 0.0001; ns, P > 0.05).

Previous work showed that the transmembrane α helix and 6RPs are essential for organelle anchorage and cytoplasmic integrity, while the C-terminal KASH peptide plays a minor role and the N-terminal actin-binding CH domains are dispensable (19). To identify which ANC-1 domains are the key determinants for mesoscale mobility, we analyzed GEM dynamics in mutants lacking specific regions: the KASH peptide (∆KASH), the transmembrane α helix plus KASH peptide (∆TK), the 6RPs (∆6RPs), or the N-terminal actin-binding CH domains (∆CH) (Fig. 4A and fig. S4).

Our analysis revealed both tissue-specific and global requirements for different ANC-1 domains. The transmembrane α helix proved essential in both tissues, with anc-1(∆TK) mutants showing disrupted constraints comparable to anc-1(e1873) animals. In contrast, anc-1(∆KASH), anc-1(∆6RPs), and anc-1(∆CH) mutants showed minimal effects on GEM constraints (Fig. 4, B to G, and fig. S4). Tissue-specific differences emerged in the regulation of unconstrained GEM diffusion: In intestinal cells, only anc-1(∆TK) mutants showed significantly faster diffusion (Fig. 4, D and H), while in hypodermal cells, both anc-1(∆TK) and anc-1(∆KASH) mutants exhibited enhanced diffusion (Fig. 4, E and I). These findings reveal that ANC-1’s transmembrane α helix is required for establishing mesoscale constraints. Furthermore, ANC-1’s role in regulating unconstrained mesoscale diffusion varies slightly by tissue. Notably, anc-1(∆6RPs), which deletes more than half the spectrin-like domains and causes organelle positioning defects (19), did not disrupt diffusion at the mesoscale, suggesting that organelle positioning requires a longer spectrin-like region than needed for constraining mesoscale particle motion.

ANC-1 and ribosomes affect ER architecture through distinct mechanisms

The ANC-1 transmembrane α helix is essential for both ER localization (19) and maintenance of mesoscale cytoplasmic constraints. Given that ribosomes are major determinants of crowding and associate extensively with the ER, we investigated how both ANC-1–dependent structural constraints and ribosome-dependent crowding influence ER organization.

anc-1(e1873) and rps-18(RNAi) mutants exhibited distinct defects in ER architecture. Ribosome depletion resulted in large, interconnected cytoplasmic vacancies in the ER network, while anc-1(e1873) animals showed increased ER fragmentation (Fig. 4J). Quantitative analysis revealed decreased ER occupancy (see Materials and Methods) under both conditions (Fig. 4, K and L), although network complexity (see Materials and Methods) was reduced only in ribosome-depleted animals (Fig. 4, M and D). Despite the altered ER morphology in rps-15(RNAi) animals, cytoplasmic organization remained stable during locomotion, contrasting with the pronounced organelle displacement observed in anc-1(e1873) mutants (movie S3) (19). In addition, ribosome levels, measured by RPS-18::GFP intensity, remained unchanged in anc-1(e1873) mutants (fig. S5).

These findings suggest complex relationships between ANC-1, ribosomes, and ER organization. While both factors affect ER morphology, they do so in distinct ways: Ribosome depletion leads to large cytoplasmic vacancies while maintaining network connectivity, whereas ANC-1 loss results in increased network fragmentation. The specific role of ribosomes in maintaining ER organization requires further investigation, as does the relationship between crowding and ER morphology. ANC-1, through its spectrin-like repeats targeted to the ER via its transmembrane α helix, appears to provide structural support characteristic of spectrin proteins (19, 26). This scaffolding function may enable ANC-1–supported ER to both maintain its organization and contribute to mesoscale constraints in the cytoplasm in vivo.

DISCUSSION

Our study reveals fundamental mechanisms controlling mesoscale cytoplasmic biophysical properties in living animal tissues. Using GEMs, we discovered that C. elegans tissues maintain an exceptionally constrained cytoplasmic state compared to other biological systems. This tight control emerges from two distinct but complementary mechanisms: ribosome-mediated crowding and ANC-1–dependent structural constraints. We show that the giant KASH protein ANC-1 plays a previously unrecognized role in establishing size-dependent constraints on mesoscale cytoplasmic diffusion through its association with the ER network. This function requires ANC-1’s transmembrane α helix but is independent of its canonical role in nuclear anchorage.

Both intestinal cells and the hypodermis syncytium form giant cell volumes that require sophisticated organizational structures and enhanced compartmentalization, which likely contributes to the highly constrained diffusion and pronounced molecular crowding we observed. For context, a 40-nm tracer diffuses at ~10.7 μm²/s in pure water [1 cP (centipoise)] at 20°C according to the Stokes-Einstein relation. In wild-type C. elegans tissues, GEMs diffuse at only ~0.002 μm²/s, corresponding to an effective cytoplasmic viscosity of ~5350 cP—representing a 5000-fold reduction compared to water. Even in anc-1 null mutant tissues, where structural constraints are disrupted, GEMs still diffuse at just ~0.08 μm²/s, which is 40-fold faster than wild type but still 130-fold slower than water. This marked hierarchy demonstrates the extraordinarily crowded biophysical environment of the C. elegans cytoplasm, with wild-type tissues being orders of magnitude more viscous than cultured mammalian cells (fig. S7).

Our findings demonstrate that cells use both passive and active mechanisms to maintain their internal architecture: Ribosomes act as crowding agents, affecting particle mobility, while ANC-1 provides structural constraints through associating with the ER. These distinct regulatory mechanisms establish a previously unknown paradigm for understanding cellular organization, where both crowding and structural elements work together to maintain proper mesoscale properties. This dual control system may represent a conserved strategy allowing tissues to establish and maintain distinct biophysical environments appropriate for their specialized functions.

While the present study focuses on a single particle size under baseline physiological conditions, our GEM platform makes it possible in future studies to interrogate how osmotic stress, mechanical load, and aging reshape mesoscale dynamics in vivo, which can be addressed with different size particles. Future experiments will extend these approaches to studies of development, stress responses, and disease models, which will further test the broader relevance of ANC-1–dependent structural constraint and ribosome-dependent crowding.

MATERIALS AND METHODS

Chemicals and molecules

The QIAprep Spin Miniprep Kit for DNA purification was obtained from Qiagen (Hilden, Germany). Restriction enzyme Sap I and T4 DNA ligase were obtained from New England Biolabs (Ipswich, MA). The gene fragment containing C. elegans codon-optimized GEM was synthesized by Integrated DNA Technologies (Coralville, IA). RNAi feeding vectors and bacterial strains were obtained from the Ahringer RNAi library (Source Bioscience, Nottingham, UK).

GEM plasmid construction and strain generation

The 40-nm GEM expression constructs were generated using the encapsulin protein sequence from Pyrococcus furiosus (Protein Data Bank ID: 2E0Z). The sequence consisted of the encapsulin open reading frame fused to EGFP, the unc-54 3′ untranslated region, and flanking Sap I restriction sites. The sequence was codon optimized for C. elegans expression and included artificial introns to enhance its expression (fig. S2). The optimized sequence was commercially synthesized as a gene block (Integrated DNA Technologies, Coralville, IA). Two tissue-specific promoters were used for expression: the semo-1 promoter (2.9 kb) for hypodermal expression and the ges-1 promoter (2 kb) for intestinal expression. Primers were designed with complementary overhands to both the GEM gene block and backbone vector pLF3FShC (27). The final constructs were assembled using Golden Gate cloning with the Sap I restriction enzyme and T4 DNA ligase, generating plasmids pSL848 (p_ges-1) and pSL850 (p_semo-1). Plasmid sequences were verified by Sanger sequencing. C. elegans strains were maintained on nematode growth medium plates seeded with the Escherichia coli strain OP50 (28). Some strains were obtained from the Caenorhabditis Genetics Center, funded by the National Institutes of Health Office of Research Infrastructure Programs (P40 OD010440). Single-copy transgenic strains were generated using FLP recombinase-mediated cassette exchange (RMCE). Plasmids pSL848 and pSL850 (50 ng/μl) were microinjected into the gonads of the RMCE-compatible strain NM5179 to generate strains UD803 and UD838. All strains used in this study are listed in table S1. The endogenous GFP11 tag was introduced into the rps-18 locus using CRISPR-Cas9 genome editing. A CRISPR RNA (caagatgtcgttgatcattc) was used to guide Cas9 to the target site. For homology-directed repair, a DNA repair template containing GFP11 flanked by homology arms was used (IDT HDR single-stranded DNA). The repair template sequence was CTAAATTTTTTATTTTTCAGGGTCATAAAATCACGACAAGATGAGAGATCACATGGTTCTTCATGAATATGTAAATGCAGCTGGAATTACAGAACTCGGCTCAGGATCTGGTTCTTCAGCTGGTTCGTTGATCATCCCAGAGAAATTCCAGCACATTCATCGTGTGATGAACACCAACATCGAT. Successful genome editing was confirmed by polymerase chain reaction and sequencing. A list of primers is shown in table S2.

RNAi-based genetics

Expression levels of the intestinal GEM strain (UD803) were suitable for single-particle tracking, but the hypodermal strain (UD838) had expression levels too high for efficient particle tracking. To achieve optimal expression levels in the hypodermal strain, animals were fed bacteria expressing double-stranded RNA targeting gfp to reduce GEM-EGFP fusion protein levels (28). RNAi experiments were conducted by feeding HT115(DE3) E. coli transformed with L4440-based vectors expressing double-stranded RNA. RNAi clones were obtained from the Ahringer RNAi library (Source Bioscience, Nottingham, UK) and verified by Sanger sequencing before use. Bacteria were seeded onto nematode growth medium plates according to established protocols. Synchronized L2-L3 stage worms were transferred to RNAi feeding plates and maintained at 23°C for 48 hours until they reached the young adult stage before imaging. A forward genetic screen for nuclear position defects similar to anc-1 mutants was performed using RNAi feeding. Single adult kuIs54[P_sur-5::sur-5::gfp] (MH1870; gift of M. Han, University of Colorado, Boulder) animals with GFP-marked nuclei (29) were placed on plates containing E. coli expressing RNAi clones from chromosome I of the Ahringer library (Source Bioscience) (28). Offspring were examined for nuclear positioning defects in the syncytial hypodermis using a Leica MLZIII fluorescent dissecting stereomicroscope with a 2× objective (Leica, Deerfield, IL). Screening all the double-stranded RNA clones on chromosome I yielded two hits: multiple clones targeting anc-1 and clone F36A2.6 against rps-15. The nuclear anchorage phenotypes of rps-15(RNAi) and rps-18(RNAi) animals were quantified as previously described (30).

Spinning disc confocal microscopy

Imaging was performed on a Nikon Ti2 microscope (Nikon Instruments, Melville, NY) equipped with a Yokogawa CSU-X1 confocal scanner unite (Yokogawa Electric Corporation, Tokyo, Japan) and a Hamamatsu ORCA-Flash4.0 LT3 Digital sCMOS camera (Hamamatsu Photonics, Shizuoka, Japan) using a Nikon Plan Apo l 100× oil immersion objective (numerical aperture, 1.45). Images were acquired at a resolution of 1060 by 1568 pixels at 0.065 μm per pixel. For the visualization of GEMs, samples were illuminated using a Nikon 488-nm laser at 100% power and imaged with an exposure time of 20 ms per frame under continuous acquisition mode (i.e., no interframe interval). The ER was visualized using a mKate2::tram-1 marker (strain UD756; see table S1) with a Nikon 561-nm laser at 50% power and 100-ms exposure time. Ribosomes were labeled by inserting GFP₁₁ into the rps-18 locus and using an extrachromosomal array expressing col-19::GFP_1–10 (strain UD863; see table S1), visualized using a Nikon 488-nm laser at 30% power with a 50-ms exposure time. For imaging GEM diffusion, ER structure, and ribosome intensity, which require the immobilization of C. elegans, we used the conventional paralysis method with 1 mM tetramisole in M9 buffer (30). For imaging ER displacement during locomotion, animals were mounted in M9 buffer without tetramisole. Image acquisition was performed using Nikon Elements software. Images for GEMs and ER are processed and analyzed in napari GUI with GEM detection and motion analysis or ER occupancy and complexity analysis (see Materials and Methods). The ribosome marker intensity and colocalization with the ER marker were measured and analyzed with FIJI/ImageJ manually.

Cytoplasmic EGFP FRAP experiments and analysis

C. elegans L4 larvae expressing cytoplasmic EGFP under the sur-5 promoter in wild-type or anc-1(e1873) backgrounds (strains UD1053 and UD1058, respectively) were grown at 25°C for 24 hours before imaging. FRAP experiments were performed on a Zeiss 980 laser scanning confocal microscope (Zeiss AG, Jena, Germany). The hypodermis between the pharynx and the germ line was visualized using a Zeiss 63× oil immersion objective (numerical aperture, 1.40), with a pinhole size of 1 AU (Airy unit) and a zoom factor of 8, yielding a 17 by 17–μm field of view at 512 by 512 pixels. The 488-nm laser intensity was set to 0.01%, and the signal was collected over a 492- to 684-nm emission window. Bidirectional scanning with a pixel dwell time of 0.34 μs resulted in a frame time of ~100 ms. Two 3.71-μm by 3.85-μm regions of interest (ROIs) were defined for photobleaching, and photobleaching was carried out at 100% 488-nm laser power and repeated 20 times. A time series containing the baseline image and photobleaching step and recovery containing 135 images were recorded over 15 s, and the intensity profile in the ROIs was analyzed (31). Briefly, fluorescence intensities were corrected for photobleaching, baseline adjusted to initial postbleach intensity, and normalized to prebleach fluorescence. Recovery curves were fitted to a single-exponential equation I(t) = I_max(1 − e^−kt) using PRISM 9 (Dotmatics, Boston, MA), where I(t) is the fluorescence intensity at time t, I_max is the maximal recovery intensity, and k is the recovery rate constant. Half-life (t_1/2) was calculated for each recovery curve.

GEM detection and motion analysis

GEM particle detection was performed by modeling diffraction-limited fluorescent spots as two-dimensional (2D) Gaussians above a spatially varying local background (32). Initial spot detection used an 11 by 11–pixel ROI, fitted to a 2D Gaussian function using maximum-likelihood estimation for Poisson-distributed data (33). The spot detection and localization algorithms were implemented in a custom Python/C program (34). Particle tracking was performed using TrackPy version 0.6.4 (35) with the “link” function (search range, 4.2). Trajectories shorter than 10 frames (0.2 s) were excluded from MSD analysis. For each trajectory, the D_eff was calculated by linear fitting of the first five points (total of 100 ms) of the MSD versus time plot using “numpy.polyfit.” The anomalous coefficient α was determined by fitting a line through the log(MSD) versus log(time) plots using “scipy.optimize.least_squares” with the “soft-L1” method minimizing the weighted residual function

$${\mathrm{ϵ}}_i=\frac{1}{t_i}(\text{logMS}{\mathrm{D}}_i-\mathrm{\alpha }·\text{log}{t}_i-\text{offset})$$

Diffusion coefficient distributions were analyzed after log transformation, which yielded normal distributions. Two-component Gaussian mixture models were fitted using “scikit-learn.mixture.GaussianMixture” to resolve subpopulations. Mean (μ) and variance (σ²) parameters were converted from their log-transformed values (denoted by the subscript) using standard log-normal distribution formulae.

$$\mathrm{\mu }=\text{exp}({\mathrm{\mu }}_{\text{log}}+{\mathrm{\sigma }}_{\text{log}}^2/2.0)$$

$${\mathrm{\sigma }}^2=\left[\text{exp}\left({\mathrm{\sigma }}_{\text{log}}^2\right)-1\right]·\text{exp}(2{\mathrm{\mu }}_{\text{log}}+{\mathrm{\sigma }}_{\text{log}}^2)$$

Parameter distributions were estimated by bootstrapping, resampling each dataset with replacement 1000 times for Gaussian mixture model analysis.

ER occupancy and complexity analysis

ER fluorescence images were initially processed using background subtraction with an iterative wavelet transform to remove the low-frequency background (36). The original discrete wavelet transform was replaced by an undecimated wavelet transform (à trous) algorithm using a cubic B-spline kernel. We computed five levels of detail coefficients and ran the algorithm for a maximum of 25 iterations. The background-subtracted images were then denoised and smoothed by minimizing a convex objective function

$$\text{arg}\underset{f}{\text{min}}{\mid D-f\mid }_2^2+{\mathrm{\lambda }}_1{\mid \mathit{Hf}\mid }_1+{\mathrm{\lambda }}_2{\mid f\mid }_1$$

where the first term enforces data fidelity, the second term (scaled by ${\mathrm{\lambda }}_1$ ) controls smoothness and continuity, and the third term (scaled by ${\mathrm{\lambda }}_2$ ) controls output sparsity. D represents the background-subtracted image, with f as the solution variable. We used the alternating direction method of multiplier algorithm (37) to efficiently obtain solutions to the sparsity-inducing L1 norm (denoted by ${\mid ·\mid }_1$ ). H represents a second-order finite difference operator. Computation was graphics processing unit accelerated using PyTorch (38). We implemented the background-subtraction and denoising workflow as a custom napari widget (39, 40). Denoising parameters were set to $\mathrm{\rho }=0.1,{\mathrm{\lambda }}_1=10 \text{to} 40,\text{and} {\mathrm{\lambda }}_2=10 \text{to} 40.$ Denoised images were converted to binary ER masks using Otsu’s method in scikit-image (41). Worm body masks were manually drawn in napari. For 2D ER morphology quantification, we calculated ER occupancy defined by the ratio of ER mask to worm body mask, and we calculated ER complexity defined by the $P^2/(4\mathrm{\pi A})$ ratio (42), where P and A represent the shape’s perimeter and area, respectively.

Supplementary Materials

The PDF file includes:

Figs. S1 to S7

Tables S1 and S2

Legends for movies S1 to S3

References

Other Supplementary Material for this manuscript includes the following:

Movies S1 to S3