LSE Logo MBoC Logo

A novel function for the MAP kinase SMA-5 in intestinal tube stability

Published Online:


Intermediate filaments are major cytoskeletal components whose assembly into complex networks and isotype-specific functions are still largely unknown. Caenorhabditis elegans provides an excellent model system to study intermediate filament organization and function in vivo. Its intestinal intermediate filaments localize exclusively to the endotube, a circumferential sheet just below the actin-based terminal web. A genetic screen for defects in the organization of intermediate filaments identified a mutation in the catalytic domain of the MAP kinase 7 orthologue sma-5(kc1). In sma-5(kc1) mutants, pockets of lumen penetrate the cytoplasm of the intestinal cells. These membrane hernias increase over time without affecting epithelial integrity and polarity. A more pronounced phenotype was observed in the deletion allele sma-5(n678) and in intestine-specific sma-5(RNAi). Besides reduced body length, an increased time of development, reduced brood size, and reduced life span were observed in the mutants, indicating compromised food uptake. Ultrastructural analyses revealed that the luminal pockets include the subapical cytoskeleton and coincide with local thinning and gaps in the endotube that are often enlarged in other regions. Increased intermediate filament phosphorylation was detected by two-dimensional immunoblotting, suggesting that loss of SMA-5 function leads to reduced intestinal tube stability due to altered intermediate filament network phosphorylation.


Intermediate filaments, actin filaments, and microtubules are the main components of the metazoan cytoskeleton and are involved in most cellular processes. The contribution and properties of the three components differ, however. The least investigated elements are the molecularly highly diverse intermediate filaments. In contrast to actin filaments and microtubules, they are not essential for either cell division or basic cell functions. Yet mutations in intermediate filament genes have been implicated in more than 80 human diseases (Omary, 2009).

Intermediate filaments have unique mechanical properties. They are highly flexible and extensible but support, at the same time, cellular stiffness (Stamenovic and Wang, 2000; Ramms et al., 2013; Seltmann et al., 2013; Mendez et al., 2014). Their contribution to mechanical stress resilience is reflected by the large number of blister-forming diseases that are caused by single-point mutations of epithelial keratin intermediate filament–encoding genes (Homberg et al., 2015). Besides these mechanical functions, intermediate filaments are involved in the regulation of cell growth, apoptosis, and vesicle trafficking to name but a few other functions (Kim and Coulombe, 2007; Magin et al., 2007; Toivola et al., 2010). To facilitate these functions, intermediate filaments form complex three-dimensional networks whose arrangement depends on the cell type, cell cycle stage, and functional state.

A particularly striking arrangement of the intermediate filament cytoskeleton is encountered in intestinal epithelial cells. The bulk of intermediate filaments is concentrated in a dense fibrous layer just below the apical terminal web of actin and spectrin in these cells. These layers are anchored to the apical junctional complex in vertebrates including fish (Markl and Franke, 1988), amphibians (Maurizii et al., 2000), and mammals (Iwatsuki and Suda, 2010). This remarkable distribution is also conserved in the nematode Caenorhabditis elegans, in which intermediate filaments localize to the prominent endotube (Bossinger et al., 2004; Carberry et al., 2009; Coch and Leube, 2016). It has been also observed in the basal hexapod Isotumurus maculatus, which—in contrast to other hexapods—expresses cytoplasmic intermediate filaments (Mencarelli et al., 2011). Taken together, these findings indicate it is likely that the highly conserved distribution pattern is advantageous to intestinal cells and suggests that it strengthens the mucosal barrier function (Geisler and Leube, 2016).

The intermediate filament–rich subapical network might form an intracellular barrier to the cytoplasmic space beneath and is closely attached to the organelle-free terminal web above, which serves as an anchor for the actin bundles that protrude into the microvilli of the overlying brush border (Hirokawa et al., 1982). The terminal web contains myosins, spectrins, actin, and an assortment of actin-binding proteins (see reviews in Mooseker, 1985; Fath and Burgess, 1995; Ku et al., 1999; Thomas, 2001). Assembly of the terminal web has been studied during gut development in vertebrates (Chambers and Grey, 1979), but its relationship to the intermediate filament network is still largely unexplored.

To identify structural determinants and regulatory components of the intestinal cytoskeleton and to study its functional contribution to intestinal physiology, we performed a mutagenesis screen using a reporter strain producing cyan fluorescent protein (CFP)-tagged intermediate filament protein IFB-2::CFP exclusively in the intestinal endotube (Hüsken et al., 2008). Using this strategy, we recently identified the novel intestinal filament organizer IFO-1, which acts as a structural component to localize the intermediate filament–rich endotube to the periluminal subapical cytoplasmic region of intestinal cells (Carberry et al., 2012). Here we characterize mutants of the MAP kinase 7 orthologue SMA-5, which has been linked to growth regulation (Watanabe et al., 2005), and show that sma-5 alleles induce the development of bubble-shaped invaginations of the apical plasma membrane together with the associated actin and modified intermediate filament cytoskeleton into the cytoplasm of intestinal cells.


SMA-5 mutations induce IFB-2::CFP–labeled cytoplasmic invaginations in the intestine

Allele kc1 was selected in a mutagenesis screen of IFB-2::CFP–expressing strain BJ52 (details in Carberry et al., 2012). Fluorescence microscopy of outcrossed homozygous strain BJ132 revealed that the circumferential localization of the IFB-2–labeled endotube was disrupted presenting multiple bubble-shaped IFB-2::CFP–positive protrusions into the apical cytoplasm along the entire length of the adluminal surface of all intestinal epithelial cells (Figure 1, A and B). These alterations were not detectable during embryogenesis (unpublished data) and larval development (Figure 1C) but emerged from early adulthood onward (Figure 1D), increasing in number throughout the intestine during aging (Figures 1B and 2, B, E, and H).


FIGURE 1: Mutations in the sma-5 gene induce bubble-shaped protrusions of IFB-2::CFP into the cytoplasm of intestinal cells. The fluorescence images (inverse presentation; composite images in B, E, and F) show IFB-2::CFP distribution. Note the smooth and even periluminal localization of IFB-2::CFP in the adult wild-type background (WT; strain BJ52; A), while multiple protrusions along the entire intestine are readily detectable in the adult kc1 mutant (B). Note also that the kc1 L1 larva in C is still devoid of invaginations and that the young adult in D presents some invaginations but not as many as the older adult in B. Enlargements correspond to boxed areas. (E, F) Similar multiple cytoplasmic invaginations of IFB-2 are also present in worms treated with sma-5(RNAi) and in mutants carrying a partial deletion of the sma-5 gene (allele n678). Scale bars: 100 μm in A, B, D, E, F, and inset in D; 50 μm in inset of B; 10 μm in C and inset in C.


FIGURE 2: Development of the sma-5 phenotype is age-dependent. IFB-2::CFP fluorescence (inverse presentation) was recorded 1 d (A–C), 2 d (D–F), and 4 d (G–I) after hatching. The percentage of animals showing a cytoplasmic invagination phenotype is shown at right. Wild-type animals were completely devoid of invaginations at all time points (A, D, G; 1 d: 0% [n = 0/16]; 2 d: 0% [n = 0/13]; 4 d: 0% [n = 0/21]), while a few sma-5(kc1) animals displayed scarce invaginations at 1 d (B; 9.10% [n = 2/22]). A substantial increase in number of affected animals and density of cytoplasmic invaginations occurred during aging (E, H; 2 d: 50.82% [n = 31/61]; 4 d: 87.04% [n = 47/54]). Most animals carrying allele n678 presented cytoplasmic invaginations already 1 d after hatching (C; 94.12% [n = 16/17]) and 2 d after hatching (F; 2 d: 94.74% [n = 18/19]), leading to full penetrance by 4 d after hatching (I; 4 d: 100% [n = 28/28]). Scale bars: 20 μm in A–I; 10 μm in inset of F; 20 μm in all other insets.

We used chromosomal and subchromosomal single-nucleotide polymorphism mapping to localize the kc1 mutation to the interval between 6.6 and 6.97 cM on the X chromosome, a region encompassing 94 protein-coding genes. The kc1-like phenotype was then mapped by bacterial feeding and microinjection of dsRNA to the sma-5 gene W06B3.2 (Figure 1E), which had been recently characterized by Watanabe and colleagues (2005). They showed that its mutation (allele n678) causes small body size despite unaltered cell numbers, slow growth, and abnormal intestinal granules. Aberrations of the intestinal cytoskeleton, however, were not investigated in this study. We therefore crossed the sma-5(n678) allele into the ifb-2::cfp reporter background. The resulting worms presented an even more severe phenotype of IFB-2::CFP distribution than those carrying the sma-5(kc1) allele, with many more invaginations in adults and already detectable invaginations in early larval stages (Figures 1F and 2, C, F, and I).

To demonstrate that a lack of functional SMA-5 is responsible for the kc1 phenotype, we performed rescue experiments with a fosmid construct that we obtained from the TransgeneOme Project ( This construct encodes the SMA-5–enhanced green fluorescent protein (EGFP) fusion protein SMA-5::EGFP. Microinjection of fosmid DNA into the gonads of sma-5(kc1) and sma-5(n678) animals efficiently rescued the bubble-shaped invaginations observed in matched control animals (Figure 3, A–B″). Similarly, intestine-specific expression of SMA-5 using the vha-6 promoter (Oka et al., 2001) also rescued the phenotype (Figure 3, C–D′). Furthermore, the SMA-5::EGFP fluorescence was exclusively detected in the cytoplasm of intestinal cells in strain BJ269 (Figure 3, B′ and B″). This was somewhat in contrast to the previous observations of Watanabe et al. (2005), who found expression of two sma-5 reporter constructs, which contain 1.6 kb of the upstream region and either almost none or all of the gene-coding region, not only in the intestine but also in the excretory cell and to some degree in the hypodermis. In addition, our observations were in disagreement with the distribution of another reporter containing 4.9 kb of the region upstream of exon 1 described by the same group, the expression of which was noted only in the pharynx and hypodermis. To exclude that the kc1 background is responsible for the observed differences, we also injected fosmid DNA into strain DP38 unc-119(ed3) leading to progeny expressing the wild-type copy of unc-119 together with the sma-5::egfp transgene. Fluorescence analyses of resulting non-unc strain BJ263 again showed intestine-restricted cytoplasmic expression of the fluorescent reporter with no detectable fluorescence in the pharynx, hypodermis, or excretory cell (Figure 4). A localized dot-like signal was seen in the somatic gonad in situ and in dissected worms when using anti-GFP antibodies after long exposure times (Figure 4, D′, E′, and F′). Of note, SMA-5::EGFP fluorescence intensity was highest at the apical pole of intestinal cells (Figure 4B′). To further assess whether SMA-5 function for intermediate filament organization is restricted to the intestine, we investigated the effect of sma-5(RNAi) on IFA-1 distribution in the marginal cells of the pharynx and of IFB-1 distribution in the hypodermis. In contrast to IFB-2, these intermediate filaments remained properly arranged upon SMA-5 depletion (Supplemental Figure S1; compare with wild-type distribution in Karabinos et al., 2003; Woo et al., 2004).


FIGURE 3: The sma-5 phenotype can be fully rescued by sma-5::egfp and intestine-specific vha-6p::sma-5 expression. (A–B″) The images show immunofluorescence micrographs of dissected adult intestines depicting the staining of anti–IFB-2 (A, B) and anti-GFP delineating the distribution of SMA-5::EGFP (A′, B′; merged images in A″ and B″) in sma-5(kc1) animals (A–A″) and in sma-5(kc1) animals producing a SMA-5::EGFP fusion protein (B–B″). Note the disappearance of the multiple cytoplasmic invaginations of the IFB-2 network in the rescued intestine. The pancytoplasmic localization of SMA-5::EGFP in the intestine can be appreciated in B′ and B″ (nuclei devoid of fluorescence marked by asterisks). The outline of the intestine is demarcated by broken lines in A′. (C–D′) Fluorescence micrographs (inverse presentation) of living worms depicting the distribution of IFB-2::CFP in sma-5(n678) animals (C) and in sma-5 animals microinjected with the intestine-specific expression construct vha-6p::sma-5 (strain BJ301; low and high magnification in D and D′, respectively). Note the efficient rescue of the intestinal cytoplasmic invagination phenotype in the microinjected worm. Scale bars: 20 μm.


FIGURE 4: SMA-5::EGFP is detected throughout the entire intestine and weakly in restricted regions of the somatic gonad. The in vivo fluorescence pictures in A′–D′ (corresponding interference contrast images in A–D) and immunofluorescence pictures of dissected intestines detecting GFP (E–F′) show expression of an extrachromosomal sma-5::egfp–containing fosmid. Note that the fluorescence is restricted to the cytoplasm of enterocytes, with prominent apical enrichment in some instances (arrowheads in B) with no detectable expression in either pharynx (P) or gonad (Go) even after overexposure (E′ and F′). Occasionally, localized expression was noted in the somatic gonad (asterisks in D′, E′, and F′) Scale bars: 50 μm in A–D′; 25 μm in E–F′.

sma-5(kc1) carries a point mutation in the region encoding a highly conserved catalytic serine–threonine kinase domain

The WormBase gene model ( of the sma-5 gene predicts four protein isoforms: a, b, c, and d. They give rise to polypeptides with a molecular weight of 56.7 kDa (a), 55 kDa (b), 57.6 kDa (c), and 51.3 kDa (d) and isoelectric points of 7.4, 6.61, 6.71, or 6.94, respectively. Further analysis based on RNAseq data suggests the appearance of alternative splice variants of isoform c (c1 and c2) and isoform d (d1 and d2) (Supplemental Figure S2). Of note, the cDNA corresponding to isoform a was used to successfully rescue the intestinal phenotype using the vha-6 promoter (Figure 3, D and D′).

A BLAST (Basic Local Alignment Search Tool) search revealed that SMA-5 orthologues are present in other nematodes including Caenorhabditis remanei, Caenorhabditis japonica, Caenorhabditis briggsae, Pristionchus pacificus, and Caenorhabditis brenneri with amino acid similarities of 99.2, 98.2, 94.9, 78.2, and 77.4%, respectively. More importantly, searches in the KEGG database ( showed that SMA-5 orthologues are conserved throughout the animal kingdom, including in the model organisms Drosophila melanogaster and Danio rerio (Supplemental Figure S3). The human SMA-5 orthologue still shows a similarity of 36% to the nematode counterpart. It encodes the 816 amino acid isoform 1 of mitogen-activated protein kinase 7 (BMK1/ERK5; ENSP00000311005).

Sequencing the entire sma-5 gene of kc1 animals showed that the gene contains a single point mutation (position 998 in CELE_W06B3.2a) resulting in a change of R125 (CGC) into H125 (CAC) (marked in yellow in Supplemental Figure S3). Protein domain analysis using the SMART-server ( identified SMA-5 as a serine–threonine kinase with a typical catalytic domain located between amino acids 81 and 388 in isoform a. Remarkably, R125 is contained in a highly conserved sequence motif, which shows similarities to D. melanogaster (38.5%), D. rerio (40.4%), and Homo sapiens (42.3%). R125 is just three amino acids away from one of the essential and also fully conserved amino acids (E129; marked in blue in Supplemental Figure S3). This can be taken as an indication for dysfunctional kinase activity of the kc1-encoded SMA-5 mutant protein. In contrast, the sma-5(n678) allele contains a 226 base pair deletion of the 3′ coding region of sma-5, leading to truncation of the last 75 amino acids in addition to a complete ∼15 kb deletion of the downstream genes C49F5.3, C49F5.5, C49F5.6, and C49F5.7 (Watanabe et al., 2005).

sma-5 mutants are small and have reduced brood size, slower larval development, and decreased life span

Next we wanted to investigate the consequences of the altered intestine for worm growth, replication, development, and survival. The histogram in Figure 5A shows that sma-5(kc1) worms were smaller (1.11 ± 0.05 mm; n = 38) than the wild type (1.27 ± 0.06 mm; n = 54; p < 0.001) and that sma-5(n678) worms were even smaller (0.72 ± 0.10 mm; n = 37; p < 0.001). Furthermore, the number of progeny was significantly (p < 0.001) less in sma-5(kc1) (168 ± 29; n = 40) and sma-5(n678) (69 ± 22; n = 44) than in the wild type (209 ± 21; n = 50; Figure 5B). Increased embryonic lethality was not observed in the mutants (unpublished data), but development was considerably (p < 0.001) prolonged in the mutants (Figure 5C). Time of development at 18°C was 5.01 ± 0.11 d in the wild type (n = 265), 5.59 ± 0.49 d in sma-5(kc1) (n = 269), and 7.36 ± 0.65 d in sma-5(n678) (n = 181). In addition, a reduced life span was observed in both sma-5 alleles (kc1: 19 d; n = 468; n678: 17 d; n = 461) versus the wild type (25 d; n = 512; p < 0.001) (Figure 5D).


FIGURE 5: sma-5 mutations lead to impaired growth, reduced brood size, slowed development, and decreased survival. (A) The histogram shows that sma-5(kc1) and sma-5(n678) worms are smaller (1.11 ± 0.05 mm [n = 38] and 0.72 ± 0.10 mm [n = 37], respectively) than the wild type (1.27 ± 0.06 mm; n = 54; p < 0.001). (B) The number of progeny is significantly (p < 0.001) smaller in sma-5(kc1) (168 ± 29; n = 40) and sma-5(n678) (69 ± 22; n = 44) than in the wild type (209 ± 21; n = 50). (C) The time of development at 18°C is considerably (p < 0.001) prolonged in sma-5(kc1) (5.59 ± 0.49 d; n = 269) and sma-5(n678) (7.36 ± 0.65 d; n = 181) compared with the wild type (5.01 ± 0.11 d; n = 265). (D) Life span is considerably reduced in both sma-5 alleles (kc1: 19 d [n = 468]; n678: 17 d [n = 461]) vs. the wild type (25 d; n = 512; p < 0.001).

To demonstrate that the phenotype was caused by intestine-specific loss of SMA-5 function, we performed sma-5(RNAi) experiments in strain OLB11, which facilitates selective RNA interference (RNAi) in the intestine (Pilipiuk et al., 2009). These animals not only replicated the characteristic cytoplasmic invaginations in the intestine but also showed reduced growth, retarded development, and reduced brood size (Supplemental Figure S4). Conversely, SMA-5::EGFP expression in n678 mutants not only rescued the intestinal phenotype but also rescued body size (935.0 ± 10.42 μm; n = 46 in rescued worms vs. 1035 ± 14.63 μm; n = 41 in wild type and 603.4 ± 15.99 μm; n = 47 in n678 mutants; p < 0.0001). Taken together, the observed phenotypes are indicative of compromised food resorption.

sma-5 mutations induce an overall reorganization of the apical cytoskeleton in the intestine but do not perturb junction localization and integrity

Differential interference contrast microscopy revealed that the bubble-shaped structures bulged out from the intestinal lumen (arrows in Figure 6, B–D). These invaginations were much more frequent and pronounced in n678 than in kc1 mutants. The adjacent apical domain, however, was still inconspicuous (arrowheads in Figure 6, B–D). Luminal width was often enlarged, especially in n678 mutants (asterisks in Figure 6D). To find out whether these invaginations are connected to the intestinal lumen and whether they are surrounded by IFB-2, we performed feeding experiments with the membrane-impermeable dye Texas Red-dextran in IFB-2::CFP reporter strains (n = 50). The dye was restricted to the intestinal lumen in wild-type and mutant worms, demonstrating that the intestinal barrier remains intact. In addition, the dye filled all the luminal pockets that herniate into the cytoplasm in the kc1 (n = 50) and n678 mutants (n = 50; Figure 6, E–G″). These cytoplasmic invaginations were surrounded by IFB-2::CFP fluorescence, which often appeared to be discontinuous.


FIGURE 6: sma-5 mutations induce cytoplasmic IFB-2–positive invaginations connected to the intestinal lumen that is often dilated. (A–D) The differential interference contrast microscopic images depict the characteristic even-shaped lumen (arrowheads) in wild-type N2 (WT; A), while multiple cytoplasmic invaginations (arrows) and local luminal expansions (asterisks) occur in sma-5(kc1) and sma-5(n678) mutants next to normal-appearing sections (arrowheads; B–D). (E–G″) Fluorescence micrographs detect IFB-2::CFP and Texas Red–labeled dextran after feeding in wild-type (E–E″; n = 50), sma-5(kc1) (F–F″; n = 50), and sma-5(n678) (G–G″; n = 50) intestines. Note that the membrane-impermeable dye is restricted to the even-shaped lumen in the wild type but enters bubble-shaped protrusions that are surrounded by cytoplasmic IFB-2::CFP in both mutants. The intestinal barrier remains intact in the mutants. Scale bars: 20 μm in A–D; 10 μm in E–G″.

To examine the structural and topological integrity of the C. elegans apical junction (CeAJ), we performed immunofluorescence microscopy on isolated adult intestines with antibodies directed against the junctional MAGUK protein DLG-1 (Figure 7, A″–C″). The localization and continuity of immunostaining was unperturbed in the mutants presenting the characteristic ladder pattern. This result further supports the dextran-feeding experiments demonstrating that the intestinal barrier function is still fully intact. Next we wanted to study the effect of sma-5 mutation on other intermediate filament polypeptides in the intestine. Figure 7, A–C′, shows that IFC-2 perfectly colocalizes with IFB-2 in both the wild-type and mutant intestines. To test, whether and how other components of the apical cytoskeleton are affected, we studied the distribution of F-actin. As expected, phalloidin staining showed apical F-actin enrichment next to IFB-2 immunoreactivity (e.g., Carberry et al., 2012). This codistribution was maintained in the mutants with F-actin staining in the IFB-2–positive cytoplasmic invaginations (Figure 7, D–F″). In the following set of experiments, we looked at the cytoskeletal linker ezrin–radixin–moesin C. elegans orthologue ERM-1, which links actin to the plasma membrane (Göbel et al., 2004; van Fürden et al., 2004; Fievet et al., 2007; Fehon et al., 2010). As exemplified in Figure 7, H–I″, ERM-1 also localizes to the cytoplasmic invaginations of both sma-5 mutants. Taking these findings together, we conclude that the entire apical membrane domain is displaced into the cytoplasm of sma-5 mutants, and these invaginations remain in direct contact with the intestinal lumen.


FIGURE 7: The subapical membrane domain invaginates coordinately into the apical cytoplasm of intestinal cells without affecting junction localization in kc1 and n678 mutants. The micrographs depict results of fluorescence microscopic detection of apical components in dissected intestines. (A–C″) Triple immunofluorescence detecting the endotube components IFB-2 (left) and IFC-2 (middle) together with the CeAJ component DLG-1 (right) in N2 (WT) and sma-5 mutants (sma-5(kc1), sma-5(n678)). (D–F″) Double fluorescence microscopy localizing IFB-2 with antibodies (left) and F-actin with Alexa Fluor 546–conjugated phalloidin (middle; superimposed images at right) in wild-type and sma-5 mutant intestines. (G–I″) Double immunofluorescence microscopy using anti-IFB2 (left) and anti-ERM-1 antibodies (middle; merged images at right) in dissected wild-type and mutant intestine. Note the colocalization of IFB-2, IFC-2, F-actin, and ERM-1 in cytoplasmic invaginations (arrows). Scale bars: 10 μm.

The intestinal filament organizer IFO-1 acts additively with SMA-5

Because the recently characterized intestinal filament organizer IFO-1 plays a crucial role in correct localization of the subapical intermediate filament and actin cytoskeleton in the intestine (Carberry et al., 2012), we examined the distribution of IFO-1 in the sma-5 mutant background. As expected, IFO-1 codistributed closely with IFB-2 in the subapical intestinal cytoplasm (Figure 8, A–A″). It also codistributed in mature invaginations of sma-5(RNAi)-treated animals (Figure 8, B–B″). Slight differences between IFO-1 and IFB-2 distribution were, however, occasionally detectable, especially in smaller invaginations, where strong IFO-1 fluorescence was detectable without significant IFB-2 (Figure 8, B–B″ and Supplemental Figure S5). The findings indicate either that IFO-1 precedes cytoplasmic translocation of IFB-2 or that cytoplasmic invaginations occur at positions of a discontinuous, that is incomplete, IFB-2 network.


FIGURE 8: Intestine-specific sma-5(RNAi) animals and sma-5 mutants develop IFB-2- and IFO-1–positive cytoplasmic invaginations. (A–C″) Antibody staining detecting IFB-2 and IFO-1 in the dissected adult intestine of intestine-specific RNAi strain OLB11 (A–A″), OLB11 treated with sma-5(RNAi) (B–B″), and sma-5(n678) mutants (C–C″). (D–F″) In vivo fluorescence detection of IFB-2::CFP and IFO-1::YFP (inverted presentation in D, D′; E, E′, F, F′; corresponding interference contrast images in D″, E″, F″, respectively) in wild-type strain BJ276 (WT; D–D″) and sma-5(kc1) animals of strain BJ277 (E–F″). The images depict the characteristic apical colocalization of IFB-2 and IFO-1 that is maintained in the cytoplasmic invaginations forming in the mutant backgrounds. Note that invaginations are usually marked by both (arrowheads), although occasionally IFO-1–positive invaginations lack IFB-2 (e.g., arrows in B–B″). Scale bars: 10 μm in A–C″; 20 μm in D–F″; 10 μm in insets of E–E′; 20 μm in insets of F–F′; 5 μm in insets of B–B″.

To further investigate the genetic relationship between sma-5 and ifo-1, we prepared double knockdowns. Supplemental Figure S6 shows that ifo-1(RNAi) in sma-5(kc1) induced both the sma-5–dependent cytoplasmic invaginations and the enrichment of IFB-2 at the CeAJ that is typically found in ifo-1 mutants. Junctional integrity remained intact, and no additional phenotype was detected. The additive nature of the phenotype indicates that the mutations act on intermediate filament network organization at least in part through separate pathways.

SMA-5 is needed for structural maintenance of the intermediate filament–rich endotube and ordered arrangement of microvilli

Electron microscopy was performed to study luminal alterations at the ultrastructural level. In the wild-type situation (n = 3), the intestinal lumen, in transverse section, has a regular ellipsoid shape and is surrounded by evenly sized and regularly arranged microvilli of the brush border (Figure 9, A–C; see also Carberry et al., 2012; The intermediate filament–rich endotube appears as an electron-dense structure surrounding the lumen below the organelle-free terminal web (green arrowheads). It is anchored to the CeAJ at cell–cell borders (white arrowheads). When viewed in longitudinal section, kc1 and n678 mutants (Figure 9, D–I; n = 4 for kc1; and n = 4 for n678) showed luminal pockets extending into the intestinal cytoplasm that were not observed in the wild type. The invaginations were still connected to the intestinal lumen, and no additional lumina were observed. These pockets are formed at transition zones where the endotube is considerably thinned (yellow arrowheads) or completely missing (red arrowheads). Overall local endotube thinning was observed in both mutant backgrounds, although it was more pronounced and combined with much larger gaps in n678. Local endotube thickening was not seen in kc1 but was quite pronounced in n678 (blue arrowheads). Taken together, these findings indicate that the continuity of the endotube is compromised in the mutant intestine, leading to cytoplasmic invaginations, and likely explains the reduced/absent decoration by IFB-2 (Figure 8, B–B″, and Supplemental Figure S5). Furthermore, the regular microvillar arrangement was considerably perturbed in several regions (e.g., Figure 9, I and L). Regular spacing and orthogonal protrusion of microvilli, which are so characteristic for a genuine brush border, were lost. Finally, the CeAJs were morphologically inconspicuous, supporting the observed functional integrity of the intestinal barrier (Figure 9, J–L).


FIGURE 9: Luminal perturbations, disordered microvilli, and endotube alterations with local gaps are present in the intestine of sma-5 mutants. The electron micrographs are taken from wild type (transverse sections, A–C, J), sma-5(kc1) (longitudinal sections, D–F, K), and sma-5(n678) mutants (longitudinal sections, G–I, L). Differently sized luminal invaginations into the cytoplasm of enterocytes are seen in both mutants (D–H). Note that the endotube layer is indistinguishable from that of the wild type in some regions (green arrowheads) but is drastically altered in other regions. In sma-5(kc1) local endotube thinning (yellow arrowheads) and endotube gaps (red arrowheads) are observed. In sma-5(n678), localized and extensive endotube thickening (blue arrowheads) and thinning (yellow arrowheads) with complete endotube loss (red arrowheads) are frequent. Note also that loss of endotube and endotube thinning coincide with the rim of cytoplasmic invaginations (E, F, and H). In addition, the brush border (B) is irregular in many regions, with variable spacing and orientation (e.g., I and L). The CeAJ appears unperturbed (white arrowheads in A–C, J–L) and is still associated with the material of the endotube. L, lumen; Y, yolk granule; F, lipid droplet. Scale bars: 2 μm in A, D, and G; 500 nm in B–C, E–F, and H–L.

SMA-5 activity affects phosphorylation of intermediate filament proteins

To test whether the observed structural alterations of the IF system are linked to altered phosphorylation of intermediate filament polypeptides, we performed two-dimensional gel electrophoresis and immunoblotting to detect changes in isoelectric points. Figure 10 shows results using anti–IFB-2 antibodies. In wild-type N2, a single isoelectric variant could be detected for the higher molecular weight isoform IFB-2a, while two isoelectric variants were visible for the lower molecular weight isoform IFB-2c with a major spot (2c and 2c+1P, respectively). Phosphatase treatment resulted in a shift of isoelectric points toward basic with the appearance of new immunoreactive signals labeled as 2a−1P and 2c−1P. A different pattern of IFB-2 immunoreactivity was observed in lysates obtained from sma-5(n678) worms with an overall reduction of isoelectric points and appearance of new signals in the more acidic part of the pH gradient (2a+1p, 2a+2P, 2c+2P). Phosphatase treatment reversed this pattern toward more basic isoelectric variants. These data support the notion that intermediate filament polypeptides are phosphorylated and indicate that SMA-5 inactivation leads to increased intermediate filament phosphorylation. To examine whether this paradoxical effect might be mediated through altered ERM-1 phosphorylation, we performed immunoblots (Supplemental Figure S7). They show that the overall ERM-1 expression is reduced, while the fraction of phosphorylated ERM-1 remains unaltered in n678 mutants.


FIGURE 10: Two-dimensional immunoblot analysis of IFB-2 detects SMA-5–dependent phosphorylation. The picture shows immunoblots of polypeptides obtained from total worm lysates that were separated by two-dimensional isoelectric focusing gel electrophoresis (pH gradient of second dimension indicated on top) and reacted with monoclonal anti–IFB-2 antibodies. The higher magnifications at right depict details of the different IFB-2 isoforms that were detected. They differ in size (larger a and smaller c variant) and isoelectric points as a consequence of phosphorylation (+P, −P). Note that, at steady state, one a variant (denoted 2a) and two c variants, that is, a major spot denoted as 2c and a minor spot denoted as 2c+1P, are visible in the wild-type N2 background. As expected, phosphatase treatment leads to a shift to more basic isoelectric points (2a−1P and 2c−1P). In contrast, isoelectric points are shifted toward acidic in the sma-5(n678) background, with new spots at 2a+1p, 2a+2P, and 2c+2P. Again, phosphatase treatment shifts the variants toward the nonphosphorylated positions. Indication of molecular weight in kilodaltons.


We provide genetic evidence that kinase activity is linked to intermediate filament network stability in the polarized simple epithelium of the C. elegans intestine by identifying the serine–threonine MAPK7 kinase orthologue SMA-5 as a key regulator of intermediate filament network organization. In sma-5 mutants, the intermediate filament protein IFB-2 is hyperphosphorylated rather than hypophosphorylated, as one might expect if IFB-2 was a direct target of SMA-5. Therefore it is more likely that SMA-5 activates another kinase or inhibits a phosphatase that could then act on intermediate filaments. We describe phenotypic changes in the two mutant sma-5 alleles, n678 and kc1. The more severe phenotype in n678 mutants is caused by a carboxy-terminal deletion of 75 amino acids, presumably yielding a nonfunctional polypeptide, while the much milder phenotype in kc1 mutants is linked to a mutation in a single conserved residue within the catalytic domain that may render the encoded enzyme less active but not completely dysfunctional.

Our study extends previous work that first identified and characterized worms carrying the n678 mutation (Watanabe et al., 2005, 2007). However, these authors did not describe the drastic changes in cytoskeletal organization and luminal defects that we characterize here. Furthermore, our data are in contrast with the reported rescue of the sma-5 phenotype by expression in the hypodermis. Instead, we show a complete rescue of the sma-5 phenotype by intestine-specific reintroduction of a sma-5 reporter. In accordance, we find a phenocopy of the sma-5 phenotype, that is, reduced body length, growth retardation, reduced brood size, and shortened life span upon intestine-specific sma-5(RNAi). We therefore conclude that the sma-5 phenotype is caused by intestine-specific SMA-5 function.

In sma-5 mutants, a morphologically normal subapical cytoskeleton forms around the lumen of the intestine initially but becomes progressively disorganized over time. We interpret the incremental appearance of the pathological phenotype as a consequence of kinase dysfunction resulting in impaired resilience of the endotube, which manifests as cytoplasmic invaginations upon prolonged and increasing wear and tear. However, the endotube in sma-5 mutants appears to be fragile and eventually ruptures, perhaps because of stress from normal wear and tear or because of growth during development. The worm intestine seems to be sensitive to stress, and invaginations and luminal dilation occur in worms subjected to challenges such as microbial infection (Estes et al., 2011) or exposure to fungal lectins (Stutz et al., 2015) or pore-forming toxins (Los et al., 2011). In fact, the intestinal filament organizer IFO-1 is up-regulated in response to pore-forming toxins, and this up-regulation requires another MAP kinase, KGB-1 (Kao et al., 2011). In support of a protective role for the SMA-5 MAP kinase, the vertebrate homologue MAPK7 is activated by oxidative stress and hypo-osmolarity (Abe et al., 1996; Chao et al., 1999).

IFO-1 is a critical component of the organization of intermediate filaments in the intestine. During embryogenesis, IFO-1 accumulates at the apical plasma membrane before localization of intermediate filaments (Carberry et al., 2012). Moreover, in ifo-1 mutants, intermediate filaments collapse onto the CeAJ. In sma-5 mutants, sites of IFO-1 delamination seem to precede the subsequent herniation of the intermediate filament protein IFB-2 and the lumen of the intestine. We stress, however, that the functions of IFO-1 and SMA-5 are separable, rather than being organized in a simple pathway. Loss of IFO-1 leads to intermediate filament accumulation at junctions, whereas loss of SMA-5 leads to a weakening of the intermediate filament–rich endotube leading to cytoplasmic invaginations. As expected, we find that sma-5 and ifo-1 phenotypes are additive in double knockdowns. Furthermore, loss of SMA-5 is coupled to intermediate filament hyperphosphorylation, whereas loss of IFO-1 does not lead to changes in intermediate filament phosphorylation (F.G., unpublished observations). Thus, even though IFO-1 is a potential target of kinases (phosphorylation site database,, there must be additional SMA-5 targets whose phosphorylation affects intermediate filament phosphorylation and overall apical domain stability. A suppressor screen of the sma-5 mutants could possibly identify the pathway that links SMA-5 to intermediate filament phosphorylation and network organization. It may also help to understand how and why SMA-5 dysfunction results in two opposing morphologies, namely endotube thinning and endotube thickening. It is possible that phosphorylation of different targets by SMA-5 may be important for network formation and network disassembly.

Our data suggest that the endotube itself becomes organized in a trilaminar structure, with an actin network closest to the plasma membrane anchoring the microvilli, followed by an IFB-2 intermediate filament layer and an IFO-1 layer below. First, we sometimes see that the IFO-1 layer has delaminated and buckled, whereas the IFB-2 layer is still intact. Second, we observe that the actin cytoskeleton is also disrupted in advanced endotube hernias. The observation that the displacement of actin, the actin-binding protein ERM-1, IFB-2, and IFC-2 occurs in concert and coincident with the formation of dextran-filled invaginations suggests that the intermediate filament network provides a major scaffolding on which the entire apical membrane domain rests. The existence of cross-bridges among the components in C. elegans is also supported by the observation of similar invaginations in act-5(RNAi), act-5(ok1397), erm-1(RNAi), and ifc-2(RNAi) intestines (Göbel et al., 2004; Hüsken et al., 2008; Saegusa et al., 2014). In accordance, ultrastructural and molecular links have been described among these components in mice (Hirokawa et al., 1982; Grimm-Günter et al., 2009).

Previous studies in cultured epithelial cells have indicated a role for kinase signaling in epithelial intermediate filament network organization. The ratio of filamentous to soluble keratin correlates with keratin phosphorylation in most, though not all, instances (Chou and Omary, 1993; Liao et al., 1996; Strnad et al., 2002). Furthermore, inhibition of either tyrosine phosphatases or serine–threonine phosphatases resulted in rapid keratin network disruption (Strnad et al., 2001, 2002). Specifically, an involvement of the p38 MAP kinase has been demonstrated in phosphatase inhibitor–induced network disassembly into granular aggregates (Ku et al., 2002; Wöll et al., 2007). In accordance, stimulation of p38 activity induced network disorganization, leading to the formation of large granules containing phosphorylated keratins (Wöll et al., 2007). JNK and p38 MAP kinase–dependent keratin phosphorylation has been described for several stress paradigms such as osmotic imbalance and heat and chemical stress (D’Alessandro et al., 2002; Ku et al., 2002; Wöll et al., 2007). It was further shown that cooperation between p38 MAP kinase and MAPKAP kinase MK2/3 is responsible for keratin phosphorylation in intestinal HT29-MTX cells (Menon et al., 2010). Remarkably, MK2-deficient mice become less sensitive to dextran sodium sulfate (DSS)-induced colitis (Li et al., 2013). This situation is analogous to that encountered in our sma-5 mutants. We suggest that the protective effect against stress is mediated in both instances by reduced intermediate filament phosphorylation resulting in a more resilient filament network. Furthermore, similar mechanisms may act in other cell systems. Thus it has been demonstrated that tension induces a mechanotransduction pathway in the C. elegans hypodermis, whose output is increased intermediate filament phosphorylation (Zhang et al., 2011). It also involves a kinase (PAK-1) whose activity is stimulated by a hemidesmosomal mechanosensor acting through hemidesmosomal adaptor proteins and the Rac GTPase CED-10. It will be interesting to find out whether similar pathways involving SMA-5 exist in the C. elegans intestine.

The evolutionary conservation of the subapical distribution of intermediate filaments anchored to the apical junctional complex, either in the form of the prominent endotube in C. elegans or as a dense network below the terminal web in Isotumurus maculatus (Mencarelli et al., 2011), to mouse (Hirokawa et al., 1982) and man (Mooseker, 1985; Carboni et al., 1987) can be taken as evidence for the functional importance of this arrangement. The small body size, reduced brood size, developmental retardation, and reduced life span observed in sma-5 mutants (this study) and ifo-1 mutants (unpublished data; Carberry et al., 2012) all strongly support this notion. In addition, mechanical functions have been ascribed to intestinal intermediate filaments (Jahnel et al., 2016) and keratin mutations have been linked to intestinal inflammation in mice lacking keratins (Baribault et al., 1994; Habtezion et al., 2005). Interestingly, keratin polymorphisms have been associated with intestinal bowel disease in human patients (Owens and Lane, 2004; Owens et al., 2004; Zupancic et al., 2014). In vitro observations provide further support for a function of intermediate filaments in intestinal barrier formation and its regulation by phosphorylation. Thus interleukin-6 (IL-6) activity evoked increasing phosphorylation of keratins 8 and 18 coincident with subapical intermediate filament accumulation in cultured colon carcinoma–derived Caco2-BBE cells (Wang et al., 2007). In accordance, IL-6 knockout mice presented with reduced keratin 8 levels and showed increased intestinal permeability after administration of dextran sulfate sodium (DSS) (Wang et al., 2007). The relationship between intestinal barrier function and keratin phosphorylation is also suggested by the work of Tao et al. (2006), who demonstrated that hypo-osmosis induced site-specific dephosphorylation of keratin 8 in HT29 cells, whereas hyper-osmosis induced keratin 8 hyperphosphorylation. These observations are in contrast to reports in the same study noting keratin hyperphosphorylation in human HRT18 and Caco2 cells and in ex vivo colon cultures upon hypo-osmosis. But they also underscore the importance, though presently poorly understood relevance, of intermediate filament phosphorylation in different stress paradigms.

Taking these findings together, we conclude that intermediate filaments are crucial ingredients for the maintenance of the C. elegans intestinal tube, which is an important prerequisite for efficient food uptake. It is likely that they fulfill similar functions in other lumen-containing epithelial structures such as the excretory cells in C. elegans (Kolotuev et al., 2013).


DNA constructs

Fosmid WRM0633B_G09(pRedFlp-Hgr)(sma-5[26688]::S0001_pR6K_Amp_2xTY1ce_EGFP_FRT_rpsl_neo_FRT_3xFlag)dFRT::unc-119-Nat (here referred to as fosmid 3170) was obtained from the TransgeneOme Project of the Max Planck Institute of Molecular Cell Biology and Genetics (Sarov et al., 2006; Transgene­Ome Project, pWD553 was made by first cloning the sma-5 cDNA into the PstI-XmaI sites of plasmid pPifb-2:cfp (1759; Hüsken et al., 2008). Next the ifb-2 promoter was replaced by the vha-6 promoter. This was done with a Gibson reaction (Gibson et al., 2009) of two fragments. The first fragment was a PstI-HindIII digest. The second was a PCR of the vha-6 promoter using primers TAACAACTTGGAAATGAAATAAGCTTaagtatactatttattactcgatacttttgttca and TGTGGAGGAGACATCTGCAGtttttatggg­ttttggtaggttttagtcgccctg. The PCR template was pML25 [vha-6p]. pWD554 was made by replacing the cfp coding sequence and the unc-54 3′-untranslated region (UTR) of pWD553 with the let-858 3′ UTR. This was done with a Gibson reaction (Gibson et al., 2009) of two fragments. The first was a SmaI-SpeI digest of pWD553. The second was a PCR of the let-858 UTR using the primers ATTGCGCAGCAACCCGGGTAAcTCGACGCCaACGTCGTTGAattttc and GGCCCGTACGGCCGAAACCAAGCGAGGACAATTCTcatcg and pADA126 (Watanabe et al., 2011) as a template. pWD555 was made by adding three native introns into the sma-5 cDNA. This was done using a Gibson reaction (Gibson et al., 2009) of two fragments. The first was a BamHI-BglII digest of pWD554, and the second was a PCR from worm genomic DNA using primers tcttcatcatatcatttatggaaacgaggatccattggaggaaca and catgctgatcaattcttgatgattgagatcttcaacttgcatc. The sma-5(RNAi) feeding vector was produced by cloning the coding sequence of W06B3.2a exon 5 into the XbaI/BamHI restriction sites of vector L4440 (Addgene, Cambridge, MA) using forward primer ACGGATCCTATCTTCATGCTGCATGCATCGCT and reverse primer ACTCTAGAATCATCTCCTCGAGCTTTGCAGAA.

C. elegans strains

Wild-type Bristol strain N2 and Hawaiian strain CB4856 were obtained from the Caenorhabditis Genetics Center (CGC; University of Minnesota, Minneapolis, MN). Strain BJ52 kcIs21[ifb-2p::ifb2::cfp]V was recently described (Hüsken et al., 2008; for chromosomal mapping, see Hüsken, 2008). sma-5 mutant strains carrying allele kc1 were BJ132 sma-5(kc1)X;kcIs21[ifb-2p::ifb-2::cfp]V and BJ141 sma-5(kc1)X. Strain FK312 sma-5(n678)X (Watanabe et al., 2007) was provided by the CGC. It was crossed with BJ49 kcIs6[ifb-2p::ifb-2::cfp]IV (Hüsken et al., 2008; for chromosomal mapping, see Hüsken, 2008) to obtain strain OLB18 sma-5(n678)X;kcIs6[ifb2-p::ifb-2::cfp]IV. BJ276 kcIs30[ifo-1p::ifo-1::yfp;myo-3p::mCherry::unc-54];kcIs21[ifb-2p::ifb2::cfp]V and BJ277 sma-5(kc1)X;kcIs30[ifo-1p::ifo-1::yfp,myo-3p::mCherry::unc-54];kcIs21[ifb-2p::ifb2::cfp]V were both generated in one step by crossing BJ132 with BJ186 kcIs30[ifo-1p::ifo-1::yfp;myo-3p::mCherry::unc-54]III animals. Microinjection of fosmid 3170 into the gonad of DP38 unc-119(ed3) animals (provided by the CGC) resulted in strain BJ263 kcEx73[pRedFlp-Hgr)(sma-5[26688]:: S0001_pR6K_Amp_2xTY1ce_EGFP_FRT_rpsl_neo_FRT_3xFlag)dFRT::unc-119-Nat] producing SMA-5::EGFP. Rescue strain BJ269 sma-5(kc1)X;kcEx73[pRedFlp-Hgr)(sma-5[26688]::S0001_pR6K_Amp_ 2xTY1ce_EGFP_FRT_rpsl_neo_FRT_ 3xFlag)dFRT::unc-119-Nat] was generated by microinjection of fosmid 3170 into BJ141 sma-5(kc1)X animals. Rescue strains BJ300/301 sma-5(n678)X; kcEx75[vha-6p::sma-5;myo-2p::gfp::H2B] were generated by microinjection of plasmid pWD555 and pCFJ770 [myo-2p::gfp::H2B] into OLB18 animals. Microinjection of plasmid 1772 [ifb-2p::ifb-2::cfp] (Hüsken et al., 2008) into the gonad of strain EC668 kcEx41[unc-119(+);ifa-1a::gfp];unc-119(ed3)III (Karabinos et al., 2003) yielded strain BJ37 kcEx4[ifb-2p::ifb-2::cfp]; kcEx41[unc-119(+);ifa-1a::gfp];unc-119(ed3)III. Strain OLB11 rde-1(ne219); kcIs39[pOLB11 (elt-2p::rde-1(+)); pRF4(rol-6(su1006))], which allows intestine-specific RNAi, was previously described (Pilipiuk et al., 2009). Maintenance and propagation of strains was done as previously described (Brenner, 1974).


RNAi by feeding was performed as previously described (Hüsken et al., 2008) with minor modifications. Nematode growth medium (NGM) plates contained 100 μg/ml ampicillin, 12.5 μg/ml tetracycline, and 2 mM isopropyl β-d-1-thiogalactopyranoside (IPTG). Plasmid clones for RNAi were either taken from the Ahringer RNAi library (Geneservice, Cambridge, UK) or obtained by cloning gene sequences of at least 500 base pairs containing a minimum of one coding exon into the feeding vector L4440 (Addgene) followed by chemical transformation into RNaseIII-deficient HT115 Escherichia coli. Liquid overnight cultures in Luria broth with 100 μg/ml ampicillin and 12.5 μg/ml tetracycline were supplemented with 7 μl/ml 1 M IPTG before plates were inoculated with a volume of 300 μl each. After overnight incubation at room temperature, 10–20 L4 larvae were added; this was followed by incubation at 15°C or 18°C. After 48 h, adult grown animals were transferred to new plates and incubated as before until F1 progeny could be analyzed. Alternatively, RNAi by microinjection was performed. To this end, dsRNA was first amplified in vitro using the MAXIscript T7 Kit (Ambion, Austin, TX) and then injected into L4 larvae using the inverse Leica DM IRB microscope (Leica, Wetzlar, Germany). Injected animals were subsequently incubated for 24 h at 20°C on normal NGM plates and transferred onto new plates for examination of F1 progeny.

Isolation and identification of sma-5 mutants

BJ52 kcIs21[ifb-2p::ifb2::cfp]V animals were subjected to chemical mutagenesis by incubating adult worms with 47 mM ethyl methane sulfonate for 4 h at room temperature. F2 progeny corresponding to 10,000 haploid genomes were screened for alterations of IFB-2::CFP fluorescence. Isolated animals were backcrossed five times with N2 to obtain strain BJ132 sma-5(kc1)X;kcIs21[ifb-2p::ifb-2::cfp]V. Further identification of the mutated gene was done by single-nucleotide polymorphism mapping (Davis et al., 2005; Carberry et al., 2012) and RNAi targeting of the remaining 94 candidate genes within the mapped region. An RNAi screen was performed by bacterial feeding (Ahringer library) starting either with L1 or L4 larva of strain BJ52 at 18°C or 20°C, resulting in down-regulation of postembryonic or embryonic gene function, respectively. In some instances, RNAi by microinjection of dsRNA was also performed, when appropriate feeding clones were not available. Animals were screened for a copy of the sma-5(kc1) phenotype using a fluorescence binocular.

Immunostaining and fluorescence labeling

For immunohistology, animals were dissected in phosphate-buffered saline on a poly-l-lysine–coated glass slide; this was followed by mounting a coverslip on top, removing excessive fluid, and freezing the sample in liquid nitrogen. The coverslip was then blasted off, resulting in permeabilization of the cuticle. Samples were then fixed by incubation at −20°C first for 10 min in methanol and then for 20 min in acetone, followed by incubation in a graded ethanol series (5 min in 90% ethanol at −20°C, 5 min in 60% ethanol at −20°C, and 5 min in 30% ethanol at room temperature). After a 10 min washing in TBST (TBS buffer [20 mM (wt/vol) Tris(hydroxymethyl)-aminomethane, pH 7.6, 0.15 M (wt/vol) NaCl] + 0.2% Tween 20), samples were incubated with primary antibodies dissolved in blocking solution (1% [wt/vol] nonfat milk powder [Roth, Karlsruhe, Germany], 1% [wt/vol] bovine serum albumin, and 0.02% [wt/vol] sodium azide) overnight at 4°C. Following a 10 min wash in TBST, samples were incubated with secondary antibodies dissolved in blocking solution for 2 h at room temperature. After a final wash for 10 min in TBST, samples were embedded in Mowiol 4–88 (Sigma, St. Louis, MO) supplemented with DABCO (Roth) and covered with a glass coverslip. The following primary antibodies were used: guinea pig anti–IFO-1 (1:100; Carberry et al., 2012), guinea pig anti–IFC-2 (1:10; Karabinos et al., 2002), rabbit anti-DLG-1 (1:200; Segbert et al., 2004), rabbit anti–ERM-1 (1:200; van Fürden et al., 2004), mouse monoclonal anti–IFB-2 (MH33, 1:200; Francis and Waterston, 1985), rabbit anti–GFP (1:1000; Invitrogen, Carlsbad, CA), and mouse monoclonal anti–AJM-1 (MH27, 1:200; Francis and Waterston, 1985). Secondary antibodies used at 1:200 dilution were Cy2-, Cy3-conjugated affinity-purified anti-mouse immunoglobulin G (IgG; Rockland, Limerick, PA), Alexa Fluor 488–conjugated affinity-purified anti-mouse IgG (Invitrogen), Alexa Fluor 555–conjugated affinity-purified anti-mouse IgG (Invitrogen), Dylight488- and Cy3-conjugated affinity-purified anti–guinea pig IgG, Cy2-, Cy3-, Cy5-conjugated anti-rabbit IgG (Jackson ImmunoResearch Laboratories, West Grove, PA), and Alexa Fluor 488–conjugated affinity-purified anti-rabbit IgG (Invitrogen). Actin filaments were stained by Alexa Fluor 546–phalloidin (Molecular Probes, Eugene, OR).

To label the intestinal lumen and to check for intestinal tightness, young adults were incubated on agar plates containing 100 μl of 0.3% (wt/vol) Texas Red–conjugated dextran (10,000 MW; Molecular Probes) in the presence of OP50 bacteria at room temperature for 24 h. They were subsequently washed three times in M9 medium before being imaged.


Microscopy was performed with confocal laser-scanning microscopes (LSM 510 META and LSM 710, Zeiss; Jena, Germany; TCS SP5; Leica), a Zeiss Apotome fitted with a Zeiss AxioCamMRm camera, and a Zeiss Axioplan 2 fitted with an ORCA-ER camera (Hamamatsu, Herrsching, Germany). For electron microscopy, worms were cryofixed by rapid high-pressure freezing with the help of a Leica EM Pact high-pressure freezer. For each genotype, at least 10 samples, each consisting of 5–15 worms, were fixed. Quick-freeze substitution was done using two alternative fixatives: 1) 1% OsO4, 0.2% uranyl acetate in acetone; and 2) 1% OsO4, 0.5% uranyl acetate, and 5% H2O in acetone. Embedding into epoxy resin was performed as previously described (McDonald and Webb, 2011). Following quick-freeze substitution, 50 nm transverse and longitudinal ultrathin sections of the worms were prepared with a Leica UC6/FC6 ultramicrotome, contrasted for 10 min each in 1% uranyl acetate in ethanol followed by Reynolds lead citrate. Sections were viewed at 100 kV with a Hitachi H-7600 transmission electron microscope (Tokyo, Japan).


For each assay, 60 adult worms of the wild type or 120 sma-5(n678) mutant animals were handpicked, transferred into 30 μl dH2O, and immediately frozen at −80°C. After quick thawing, samples were sucked up and down through a 30G hypodermic syringe (BD Medical, Heidelberg, Germany) three times. Subsequently, 7.5 μl of 5× Laemmli loading buffer was added, and samples were incubated for 5 min at 90°C. Proteins were separated in a 10% SDS polyacrylamide gel, and transferred onto polyvinylidene difluoride membrane by wet-tank blotting at 100 V for 60 min. The membranes were blocked for 1–2 h at room temperature in Roti-Block (Roth) and then incubated with primary antibodies in Roti-Block at 4°C overnight. Antibodies against total ERM-1 (murine monoclonal antibodies from the Developmental Studies Hybridoma Bank [DHSB, Iowa City, IA]; Hadwiger et al., 2010) were diluted 1:1000, against phosphorylated ERM-1 (rabbit monoclonal antibodies ab76247 from Abcam, Cambridge, UK) 1:500, and against actin (rabbit polyclonal antibodies A2066 from Sigma) 1:1000. Three 10 min washings with TBST (20 mM Tris(hydroxymethyl)-aminomethane, 0.15 M NaCl, 0.1% Tween 20 [vol/vol], pH 7.6) ensued. Incubation with secondary antibodies (goat anti-mouse IgG antibodies and goat anti-rabbit IgG antibodies coupled to horseradish peroxidase from DAKO [Santa Clara, CA] at 1:5000 in Roti-Block) was done for 1 h at room temperature. Final washes (three at 10 min each) were in TBST. Bound antibodies were detected with Fusion Solo (Vilber, Eberhardzell, Germany) using chemiluminescence substrate AceGlow (Peqlab, Darmstadt, Germany).

For two-dimensional gel electrophoresis, 60 adult worms were handpicked for each assay, transferred into 15 μl dH2O, and immediately frozen at −80°C. After quick thawing, each sample was aspirated by a syringe needle three times, resulting in homogenization, and stored on ice for at least 30 min after addition of 15 μl 2× lysis buffer (9.5 M urea, 4% [wt/wt] 3-[(3-cholamidopropyl)dimethylammonio]-1-propanesulfonate, 0.04% [wt/vol] bromophenol blue, 2% [wt/vol] dithiothreitol [DTT], 2% [vol/vol] Immobilized pH Gradients buffer [GE Healthcare, Freiburg, Germany]). Then 170 μl 2× lysis buffer was added, and all of the solution (200 μl) was transferred to an Immobiline DryStrip (pH 4-7, 11 cm; GE Healthcare) for overnight rehydration. During this time, the strip was covered by silicone oil to prevent sulfuring of urea. For isoelectric focusing, the following protocol was used: 150 V, 75 V/h; 300 V, 225 V/h; 1000 V, 500 V/h; and 3000 V, 27,000 V/h. Strips were sometimes stored at −80°C until further usage. For subsequent immunoblot analysis, strips were first washed for 15 min in 0.1 g DTT/10 ml equilibration buffer (6 M urea, 2% [wt/vol] SDS, 50 mM [wt/vol] Tris(hydroxymethyl)-aminomethane, pH 8.8, 0.02% [wt/vol] bromophenol blue, 30% [vol/vol] glycerol); this was followed by another washing step for 15 min in 0.25 g iodoacetamide/10 ml equilibration buffer. Strips were then placed on top of the solidified stacking gel and further embedded by agarose (0.5% [wt/vol] agarose, 0.02% [wt/vol] bromophenol blue in running buffer [0.25 M Tris(hydroxymethyl)-aminomethane, 1.92 M [wt/vol] glycine, 1% [wt/vol] SDS]). After gel electrophoretic separation, proteins were transferred onto polyvinylidene difluoride membrane by tank blotting at 100 V for 70 min. After blocking for 1–2 h at room temperature in Roti-Block (Roth), immunostaining against IFB-2 was done by incubating the blots with primary antibody MH33 (1:1000 dilution in Roti-Block) at 4°C overnight, washing 3 × 10 min in TBST, incubating with secondary mouse anti-mouse IgG antibodies coupled to horseradish peroxidase from DAKO at 1:5000 in Roti-Block, diluting for 1 h at room temperature, and administering final washes in TBST. Bound antibodies were detected with Fusion Solo using chemiluminescence substrate AceGlow.


This article was published online ahead of print in MBoC in Press ( on October 12, 2016.

Abbreviations used:

Basic Local Alignment Search Tool


Caenorhabditis Genetics Center


dextran sodium sulfate




immunoglobulin G




isopropyl β-d-1-thiogalactopyranoside


RNA interference


untranslated region


We thank Bärbel Bonn and Sabine Eisner for superb technical support and students Carina Keller and Carina Eherer for their help. This work was supported by grants from the German Research Council (LE566/14-1, LE566/14-3, BO 1061/11-1, and BO1061/11-3). Some strains were provided by the CGC, which is funded by National Institutes of Health (NIH) Office of Research Infrastructure Programs (P40 OD010440). The monoclonal antibodies MH27 and MH33 developed by Waterston and ERM-1 (Hadwiger et al., 2010) were obtained from the Developmental Studies Hybridoma Bank, created by the National Institute of Child Health and Human Development of the NIH and maintained at the Department of Biology, University of Iowa. We also acknowledge Anton Karabinos for generously sharing anti–IFC-2 antibody. Fosmids were obtained from the TransgeneOme Project of the MPI-CBG. Plasmid pML25 was generously shared by Matt Labella. Plasmid pCFJ770 was obtained from Christian Frøkjær-Jensen.


  • Abe J, Kusuhara M, Ulevitch RJ, Berk BC, Lee JD (1996). Big mitogen-activated protein kinase 1 (BMK1) is a redox-sensitive kinase. J Biol Chem 271, 16586–16590. Crossref, MedlineGoogle Scholar
  • Baribault H, Penner J, Iozzo RV, Wilson-Heiner M (1994). Colorectal hyperplasia and inflammation in keratin 8-deficient FVB/N mice. Genes Dev 8, 2964–2973. Crossref, MedlineGoogle Scholar
  • Bossinger O, Fukushige T, Claeys M, Borgonie G, McGhee JD (2004). The apical disposition of the Caenorhabditis elegans intestinal terminal web is maintained by LET-413. Dev Biol 268, 448–456. Crossref, MedlineGoogle Scholar
  • Brenner S (1974). The genetics of Caenorhabditis elegans. Genetics 77, 71–94. Crossref, MedlineGoogle Scholar
  • Carberry K, Wiesenfahrt T, Geisler F, Stocker S, Gerhardus H, Uberbach D, Davis W, Jorgensen E, Leube RE, Bossinger O (2012). The novel intestinal filament organizer IFO-1 contributes to epithelial integrity in concert with ERM-1 and DLG-1. Development 139, 1851–1862. Crossref, MedlineGoogle Scholar
  • Carberry K, Wiesenfahrt T, Windoffer R, Bossinger O, Leube RE (2009). Intermediate filaments in Caenorhabditis elegans. Cell Motil Cytoskeleton 66, 852–864. Crossref, MedlineGoogle Scholar
  • Carboni JM, Howe CL, West AB, Barwick KW, Mooseker MS, Morrow JS (1987). Characterization of intestinal brush border cytoskeletal proteins of normal and neoplastic human epithelial cells. A comparison with the avian brush border. Am J Pathol 129, 589–600. MedlineGoogle Scholar
  • Chambers C, Grey RD (1979). Development of the structural components of the brush border in absorptive cells of the chick intestine. Cell Tissue Res 204, 387–405. Crossref, MedlineGoogle Scholar
  • Chao TH, Hayashi M, Tapping RI, Kato Y, Lee JD (1999). MEKK3 directly regulates MEK5 activity as part of the big mitogen-activated protein kinase 1 (BMK1) signaling pathway. J Biol Chem 274, 36035–36038. Crossref, MedlineGoogle Scholar
  • Chou CF, Omary MB (1993). Mitotic arrest-associated enhancement of O-linked glycosylation and phosphorylation of human keratins 8 and 18. J Biol Chem 268, 4465–4472. MedlineGoogle Scholar
  • Coch RA, Leube RE (2016). Intermediate filaments and polarization in the intestinal epithelium. Cells 5, 32 CrossrefGoogle Scholar
  • D’Alessandro M, Russell D, Morley SM, Davies AM, Lane EB (2002). Keratin mutations of epidermolysis bullosa simplex alter the kinetics of stress response to osmotic shock. J Cell Sci 115, 4341–4351. Crossref, MedlineGoogle Scholar
  • Davis MW, Hammarlund M, Harrach T, Hullett P, Olsen S, Jorgensen EM (2005). Rapid single nucleotide polymorphism mapping in C. elegans. BMC Genomics 6, 118 Crossref, MedlineGoogle Scholar
  • Estes KA, Szumowski SC, Troemel ER (2011). Non-lytic, actin-based exit of intracellular parasites from C. elegans intestinal cells. PLoS Pathog 7, e1002227 Crossref, MedlineGoogle Scholar
  • Fath KR, Burgess DR (1995). Microvillus assembly. Not actin alone. Curr Biol 5, 591–593. Crossref, MedlineGoogle Scholar
  • Fehon RG, McClatchey AI, Bretscher A (2010). Organizing the cell cortex: the role of ERM proteins. Nat Rev Mol Cell Biol 11, 276–287. Crossref, MedlineGoogle Scholar
  • Fievet B, Louvard D, Arpin M (2007). ERM proteins in epithelial cell organization and functions. Biochim Biophys Acta 1773, 653–660. Crossref, MedlineGoogle Scholar
  • Francis GR, Waterston RH (1985). Muscle organization in Caenorhabditis elegans: localization of proteins implicated in thin filament attachment and I-band organization. J Cell Biol 101, 1532–1549. Crossref, MedlineGoogle Scholar
  • Geisler F, Leube RE (2016). Epithelial intermediate filaments: guardians against microbial infection. Cells 5, 29 CrossrefGoogle Scholar
  • Gibson DG, Young L, Chuang RY, Venter JC, Hutchison CA, Smith HO (2009). Enzymatic assembly of DNA molecules up to several hundred kilobases. Nat Methods 6, 343–345. Crossref, MedlineGoogle Scholar
  • Göbel V, Barrett PL, Hall DH, Fleming JT (2004). Lumen morphogenesis in C. elegans requires the membrane-cytoskeleton linker erm-1. Dev Cell 6, 865–873. Crossref, MedlineGoogle Scholar
  • Grimm-Günter EM, Revenu C, Ramos S, Hurbain I, Smyth N, Ferrary E, Louvard D, Robine S, Rivero F (2009). Plastin 1 binds to keratin and is required for terminal web assembly in the intestinal epithelium. Mol Biol Cell 20, 2549–2562. LinkGoogle Scholar
  • Habtezion A, Toivola DM, Butcher EC, Omary MB (2005). Keratin-8-deficient mice develop chronic spontaneous Th2 colitis amenable to antibiotic treatment. J Cell Sci 118, 1971–1980. Crossref, MedlineGoogle Scholar
  • Hadwiger G, Dour S, Arur S, Fox P, Nonet ML (2010). A monoclonal antibody toolkit for C. elegans. PLoS One 5, e10161 Crossref, MedlineGoogle Scholar
  • Hirokawa N, Tilney LG, Fujiwara K, Heuser JE (1982). Organization of actin, myosin, and intermediate filaments in the brush border of intestinal epithelial cells. J Cell Biol 94, 425–443. Crossref, MedlineGoogle Scholar
  • Homberg M, Ramms L, Schwarz N, Dreissen G, Leube RE, Merkel R, Hoffmann B, Magin TM (2015). Distinct impact of two keratin mutations causing epidermolysis bullosa simplex on keratinocyte adhesion and stiffness. J Invest Dermatol 135, 2437–2445. Crossref, MedlineGoogle Scholar
  • Hüsken K (2008). Identifizierung und Charakterisierung von Intermediärfilament-Organisatoren in Caenorhabditis elegans. PhD Thesis., Mainz, Germany: Johannes Gutenberg-Universität Mainz Google Scholar
  • Hüsken K, Wiesenfahrt T, Abraham C, Windoffer R, Bossinger O, Leube RE (2008). Maintenance of the intestinal tube in Caenorhabditis elegans: the role of the intermediate filament protein IFC-2. Differentiation 76, 881–896. Crossref, MedlineGoogle Scholar
  • Iwatsuki H, Suda M (2010). Seven kinds of intermediate filament networks in the cytoplasm of polarized cells: structure and function. Acta Histochem Cytochem 43, 19–31. Crossref, MedlineGoogle Scholar
  • Jahnel O, Hoffmann B, Merkel R, Bossinger O, Leube RE (2016). Mechanical probing of the intermediate filament-rich Caenorhabditis elegans intestine. Methods Enzymol 568, 681–706. Crossref, MedlineGoogle Scholar
  • Kao CY, Los FC, Huffman DL, Wachi S, Kloft N, Husmann M, Karabrahimi V, Schwartz JL, Bellier A, Ha C et al.(2011). Global functional analyses of cellular responses to pore-forming toxins. PLoS Pathog 7, e1001314 Crossref, MedlineGoogle Scholar
  • Karabinos A, Schulze E, Klisch T, Wang J, Weber K (2002). Expression profiles of the essential intermediate filament (IF) protein A2 and the IF protein C2 in the nematode Caenorhabditis elegans. Mech Dev 117, 311–314. Crossref, MedlineGoogle Scholar
  • Karabinos A, Schulze E, Schunemann J, Parry DA, Weber K (2003). In vivo and in vitro evidence that the four essential intermediate filament (IF) proteins A1, A2, A3 and B1 of the nematode Caenorhabditis elegans form an obligate heteropolymeric IF system. J Mol Biol 333, 307–319. Crossref, MedlineGoogle Scholar
  • Kim S, Coulombe PA (2007). Intermediate filament scaffolds fulfill mechanical, organizational, and signaling functions in the cytoplasm. Genes Dev 21, 1581–1597. Crossref, MedlineGoogle Scholar
  • Kolotuev I, Hyenne V, Schwab Y, Rodriguez D, Labouesse M (2013). A pathway for unicellular tube extension depending on the lymphatic vessel determinant Prox1 and on osmoregulation. Nat Cell Biol 15, 157–168. Crossref, MedlineGoogle Scholar
  • Ku NO, Azhar S, Omary MB (2002). Keratin 8 phosphorylation by p38 kinase regulates cellular keratin filament reorganization: modulation by a keratin 1-like disease causing mutation. J Biol Chem 277, 10775–10782. Crossref, MedlineGoogle Scholar
  • Ku NO, Zhou X, Toivola DM, Omary MB (1999). The cytoskeleton of digestive epithelia in health and disease. Am J Physiol 277, G1108–G1137. MedlineGoogle Scholar
  • Li YY, Yuece B, Cao HM, Lin HX, Lv S, Chen JC, Ochs S, Sibaev A, Deindl E, Schaefer C, Storr M (2013). Inhibition of p38/Mk2 signaling pathway improves the anti-inflammatory effect of WIN55 on mouse experimental colitis. Lab Invest 93, 322–333. Crossref, MedlineGoogle Scholar
  • Liao J, Ku NO, Omary MB (1996). Two-dimensional gel analysis of glandular keratin intermediate filament phosphorylation. Electrophoresis 17, 1671–1676. Crossref, MedlineGoogle Scholar
  • Los FC, Kao CY, Smitham J, McDonald KL, Ha C, Peixoto CA, Aroian RV (2011). RAB-5- and RAB-11-dependent vesicle-trafficking pathways are required for plasma membrane repair after attack by bacterial pore-forming toxin. Cell Host Microbe 9, 147–157. Crossref, MedlineGoogle Scholar
  • Magin TM, Vijayaraj P, Leube RE (2007). Structural and regulatory functions of keratins. Exp Cell Res 313, 2021–2032. Crossref, MedlineGoogle Scholar
  • Markl J, Franke WW (1988). Localization of cytokeratins in tissues of the rainbow trout: fundamental differences in expression pattern between fish and higher vertebrates. Differentiation 39, 97–122. Crossref, MedlineGoogle Scholar
  • Maurizii MG, Alibardi L, Taddei C (2000). Organization and characterization of the keratin cytoskeleton in the previtellogenic ovarian follicle of the lizard Podarcis sicula Raf. Mol Reprod Dev 57, 159–166. Crossref, MedlineGoogle Scholar
  • McDonald KL, Webb RI (2011). Freeze substitution in 3 hours or less. J Microsc 243, 227–233. Crossref, MedlineGoogle Scholar
  • Mencarelli C, Ciolfi S, Caroti D, Lupetti P, Dallai R (2011). Isomin: a novel cytoplasmic intermediate filament protein from an arthropod species. BMC Biol 9, 17 Crossref, MedlineGoogle Scholar
  • Mendez MG, Restle D, Janmey PA (2014). Vimentin enhances cell elastic behavior and protects against compressive stress. Biophys J 107, 314–323. Crossref, MedlineGoogle Scholar
  • Menon MB, Schwermann J, Singh AK, Franz-Wachtel M, Pabst O, Seidler U, Omary MB, Kotlyarov A, Gaestel M (2010). p38 MAP kinase and MAPKAP kinases MK2/3 cooperatively phosphorylate epithelial keratins. J Biol Chem 285, 33242–33251. Crossref, MedlineGoogle Scholar
  • Mooseker MS (1985). Organization, chemistry, and assembly of the cytoskeletal apparatus of the intestinal brush border. Annu Rev Cell Biol 1, 209–241. Crossref, MedlineGoogle Scholar
  • Oka T, Toyomura T, Honjo K, Wada Y, Futai M (2001). Four subunit a isoforms of Caenorhabditis elegans vacuolar H+-ATPase. Cell-specific expression during development. J Biol Chem 276, 33079–33085. Crossref, MedlineGoogle Scholar
  • Omary MB (2009). “IF-pathies”: a broad spectrum of intermediate filament-associated diseases. J Clin Invest 119, 1756–1762. Crossref, MedlineGoogle Scholar
  • Owens DW, Lane EB (2004). Keratin mutations and intestinal pathology. J Pathol 204, 377–385. Crossref, MedlineGoogle Scholar
  • Owens DW, Wilson NJ, Hill AJ, Rugg EL, Porter RM, Hutcheson AM, Quinlan RA, van Heel D, Parkes M, Jewell DP et al.(2004). Human keratin 8 mutations that disturb filament assembly observed in inflammatory bowel disease patients. J Cell Sci 117, 1989–1999. Crossref, MedlineGoogle Scholar
  • Pilipiuk J, Lefebvre C, Wiesenfahrt T, Legouis R, Bossinger O (2009). Increased IP3/Ca2+signaling compensates depletion of LET-413/DLG-1 in C. elegans epithelial junction assembly. Dev Biol 327, 34–47. Crossref, MedlineGoogle Scholar
  • Ramms L, Fabris G, Windoffer R, Schwarz N, Springer R, Zhou C, Lazar J, Stiefel S, Hersch N, Schnakenberg U et al.(2013). Keratins as the main component for the mechanical integrity of keratinocytes. Proc Natl Acad Sci USA 110, 18513–18518. Crossref, MedlineGoogle Scholar
  • Saegusa K, Sato M, Sato K, Nakajima-Shimada J, Harada A (2014). Caenorhabditis elegans chaperonin CCT/TRiC is required for actin and tubulin biogenesis and microvillus formation in intestinal epithelial cells. Mol Biol Cell 25, 3095–3104. LinkGoogle Scholar
  • Sarov M, Schneider S, Pozniakovski A, Roguev A, Ernst S, Zhang Y, Hyman AA, Stewart AF (2006). A recombineering pipeline for functional genomics applied to Caenorhabditis elegans. Nat Methods 3, 839–844. Crossref, MedlineGoogle Scholar
  • Segbert C, Johnson K, Theres C, van Furden D, Bossinger O (2004). Molecular and functional analysis of apical junction formation in the gut epithelium of Caenorhabditis elegans. Dev Biol 266, 17–26. Crossref, MedlineGoogle Scholar
  • Seltmann K, Fritsch AW, Kas JA, Magin TM (2013). Keratins significantly contribute to cell stiffness and impact invasive behavior. Proc Natl Acad Sci USA 110, 18507–18512. Crossref, MedlineGoogle Scholar
  • Stamenovic D, Wang N (2000). Invited review: engineering approaches to cytoskeletal mechanics. J Appl Physiol 89, 2085–2090. Crossref, MedlineGoogle Scholar
  • Strnad P, Windoffer R, Leube RE (2001). In vivo detection of cytokeratin filament network breakdown in cells treated with the phosphatase inhibitor okadaic acid. Cell Tissue Res 306, 277–293. Crossref, MedlineGoogle Scholar
  • Strnad P, Windoffer R, Leube RE (2002). Induction of rapid and reversible cytokeratin filament network remodeling by inhibition of tyrosine phosphatases. J Cell Sci 115, 4133–4148. Crossref, MedlineGoogle Scholar
  • Stutz K, Kaech A, Aebi M, Kunzler M, Hengartner MO (2015). Disruption of the C. elegans intestinal brush border by the fungal lectin CCL2 phenocopies dietary lectin toxicity in mammals. PLoS One 10, e0129381 Crossref, MedlineGoogle Scholar
  • Tao GZ, Toivola DM, Zhou Q, Strnad P, Xu B, Michie SA, Omary MB (2006). Protein phosphatase-2A associates with and dephosphorylates keratin 8 after hyposmotic stress in a site- and cell-specific manner. J Cell Sci 119, 1425–1432. Crossref, MedlineGoogle Scholar
  • Thomas GH (2001). Spectrin: the ghost in the machine. Bioessays 23, 152–160. Crossref, MedlineGoogle Scholar
  • Toivola DM, Strnad P, Habtezion A, Omary MB (2010). Intermediate filaments take the heat as stress proteins. Trends Cell Biol 20, 79–91. Crossref, MedlineGoogle Scholar
  • van Fürden D, Johnson K, Segbert C, Bossinger O (2004). The C. elegans ezrin-radixin-moesin protein ERM-1 is necessary for apical junction remodelling and tubulogenesis in the intestine. Dev Biol 272, 262–276. Crossref, MedlineGoogle Scholar
  • Wang L, Srinivasan S, Theiss AL, Merlin D, Sitaraman SV (2007). Interleukin-6 induces keratin expression in intestinal epithelial cells: potential role of keratin-8 in interleukin-6-induced barrier function alterations. J Biol Chem 282, 8219–8227. Crossref, MedlineGoogle Scholar
  • Watanabe N, Ishihara T, Ohshima Y (2007). Mutants carrying two sma mutations are super small in the nematode C. elegans. Genes Cells 12, 603–609. Crossref, MedlineGoogle Scholar
  • Watanabe N, Nagamatsu Y, Gengyo-Ando K, Mitani S, Ohshima Y (2005). Control of body size by SMA-5, a homolog of MAP kinase BMK1/ERK5, in C. elegans. Development 132, 3175–3184. Crossref, MedlineGoogle Scholar
  • Watanabe S, Punge A, Hollopeter G, Willig KI, Hobson RJ, Davis MW, Hell SW, Jorgensen EM (2011). Protein localization in electron micrographs using fluorescence nanoscopy. Nat Methods 8, 80–84. Crossref, MedlineGoogle Scholar
  • Wöll S, Windoffer R, Leube RE (2007). p38 MAPK-dependent shaping of the keratin cytoskeleton in cultured cells. J Cell Biol 177, 795–807. Crossref, MedlineGoogle Scholar
  • Woo WM, Goncharov A, Jin Y, Chisholm AD (2004). Intermediate filaments are required for C. elegans epidermal elongation. Dev Biol 267, 216–229. Crossref, MedlineGoogle Scholar
  • Zhang H, Landmann F, Zahreddine H, Rodriguez D, Koch M, Labouesse M (2011). A tension-induced mechanotransduction pathway promotes epithelial morphogenesis. Nature 471, 99–103. Crossref, MedlineGoogle Scholar
  • Zupancic T, Stojan J, Lane EB, Komel R, Bedina-Zavec A, Liovic M (2014). Intestinal cell barrier function in vitro is severely compromised by keratin 8 and 18 mutations identified in patients with inflammatory bowel disease. PLoS One 9, e99398 Crossref, MedlineGoogle Scholar