(C) PLOS One
This story was originally published by PLOS One and is unaltered.
. . . . . . . . . .



Zyxin contributes to coupling between cell junctions and contractile actomyosin networks during apical constriction [1]

['Mark M. Slabodnick', 'Biology Department', 'University Of North Carolina At Chapel Hill', 'Chapel Hill', 'North Carolina', 'United States Of America', 'Department Of Biology', 'Knox College', 'Galesburg', 'Illinois']

Date: 2023-04

One of the most common cell shape changes driving morphogenesis in diverse animals is the constriction of the apical cell surface. Apical constriction depends on contraction of an actomyosin network in the apical cell cortex, but such actomyosin networks have been shown to undergo continual, conveyor belt-like contractions before the shrinking of an apical surface begins. This finding suggests that apical constriction is not necessarily triggered by the contraction of actomyosin networks, but rather can be triggered by unidentified, temporally-regulated mechanical links between actomyosin and junctions. Here, we used C. elegans gastrulation as a model to seek genes that contribute to such dynamic linkage. We found that α-catenin and β-catenin initially failed to move centripetally with contracting cortical actomyosin networks, suggesting that linkage is regulated between intact cadherin-catenin complexes and actomyosin. We used proteomic and transcriptomic approaches to identify new players, including the candidate linkers AFD-1/afadin and ZYX-1/zyxin, as contributing to C. elegans gastrulation. We found that ZYX-1/zyxin is among a family of LIM domain proteins that have transcripts that become enriched in multiple cells just before they undergo apical constriction. We developed a semi-automated image analysis tool and used it to find that ZYX-1/zyxin contributes to cell-cell junctions’ centripetal movement in concert with contracting actomyosin networks. These results identify several new genes that contribute to C. elegans gastrulation, and they identify zyxin as a key protein important for actomyosin networks to effectively pull cell-cell junctions inward during apical constriction. The transcriptional upregulation of ZYX-1/zyxin in specific cells in C. elegans points to one way that developmental patterning spatiotemporally regulates cell biological mechanisms in vivo. Because zyxin and related proteins contribute to membrane-cytoskeleton linkage in other systems, we anticipate that its roles in regulating apical constriction in this manner may be conserved.

Animals take shape during development in large part by the bending of tissues. Failures in this process are common causes of human birth defects. Such tissue bending is driven primarily by individual cells changing shape: in many examples, one side of a cell shrinks, pulling on junctions that connect the cell to neighboring cells. But the networks that drive one side of a cell to shrink are not always connected to junctions. As a result, focus has turned to understanding how connections between such networks and junctions are dynamically regulated to trigger cell shape change. We sought to identify genes that contribute to these dynamic connections. Here, we describe proteomic and transcriptomic methods that we used to identify proteins that contribute to cell shape change. We developed a new image analysis tool and used it to reveal that loss of one of these genes results in networks moving without efficiently pulling in junctions. Our results identify new molecular players, and they pinpoint a key gene whose products might contribute to dynamically connecting networks to junctions to trigger tissue shape changes in C. elegans and other organisms.

Data Availability: C. elegans strains generated in this study have been deposited to the Caenorhabditis Genetics Center, accession numbers can be found in S1 Table . An interactive online visualization of transcriptomic data is available at https://n2t.net/ark:/84478/d/2bbpmsq3 . RNA-seq reads, alignments, and RPKM files are available on NCBI GEO accession number GSE205061. A Jupyter Notebook with the code for the image analysis is available on GitHub ( https://github.com/fjug/BobSeg ). All other relevant data are in the manuscript and its Supporting information files.

Here, we sought to identify proteins that could contribute to such temporally-regulated linkage either directly or indirectly. We anticipated that identifying key proteins would require integrating diverse methodologies including generating new image analysis resources, as well as screening for defects in a complex process involving cellular and subcellular dynamics. First, we hypothesized that temporal regulation of apical constriction could feasibly result from either the transcriptional regulation or posttranscriptional regulation of key linking proteins, or both. We found that members of the C. elegans cadherin-catenin complex (CCC) remained at apical cell-cell contacts as cortical actomyosin contractions began, suggesting that disassembly of these complexes cannot explain junctions’ initial failure to move. We then used both proteomic and transcriptomic approaches to find proteins that might interact physically with the C. elegans CCC as well as genes whose expression is upregulated specifically in the EPCs prior to cell internalization. Screening through the resulting list identified several new contributors to C. elegans gastrulation, including two genes, afd-1/afadin and zyx-1/zyxin, that encode candidate linkers and that we found were required for timely internalization of the EPCs. We found that AFD-1/afadin is localized broadly to cell junctions with patterns consistent with members of the CCC. Single-cell RNA-seq on multiple internalizing cells identified zyx-1/zyxin and other LIM domain-encoding genes as upregulated in multiple internalizing C. elegans cell lineages. To determine whether zyx-1/zyxin and afd-1/afadin contribute to linking contracting actomyosin networks to junctions in vivo, we developed a new, semi-automated image analysis workflow for quantifying actomyosin network and junctional movements. The results identified zyxin as a contributor to triggering apical constriction through regulating directly or indirectly the dynamic, temporally-regulated coupling of actomyosin contractions to cell-cell junctions.

Although many proteins have been identified at sites of cadherin-based adhesion that could feasibly serve as such temporally regulated links [ 14 ], the specific, temporally-regulated links relevant to triggering apical constriction are not yet known in any system. Identifying such temporally-regulated links from among candidate linkers or unknown players is an important step toward understanding how developmental mechanisms orchestrate the cytoskeletal mechanisms of apical constriction with spatial and temporal precision.

A previous study investigated actomyosin dynamics in the EPCs during gastrulation and found, unexpectedly, that contractions of the actomyosin cortex initially occurred in a conveyer belt-like fashion without pulling junctions centripetally, i.e., with junctions apparently uncoupled to the inward movement of actomyosin components toward the center of the apical cell surface [ 13 ]. It was only after several minutes of seemingly unproductive actomyosin contraction that cell-cell junctions began to move increasingly in concert with the contracting networks. This phenomenon was also observed in Drosophila melanogaster shortly before ventral furrow formation, where myosin accumulated and coalesced periodically in weak contractions that preceded the shrinking of apical cell profiles [ 13 ]. These observations suggest that in these model systems, and potentially more generally, apical constriction is not triggered by myosin activation; rather, it is likely to be triggered by gradually connecting an already-contracting apical actomyosin cytoskeleton to cell-cell junctions, via unknown links.

The nematode Caenorhabditis elegans has been a valuable model for studying mechanisms of morphogenesis [ 9 ]. Gastrulation in C. elegans begins at the 26-cell stage when a non-muscle myosin II becomes enriched in the apical cortex of two endodermal precursor cells (EPCs) [ 10 ], which then internalize by apical constriction [ 11 , 12 ].

The force-producing mechanisms that drive apical constriction are well conserved, relying on cortical networks of actin filaments and non-muscle myosin II motors, which drive contraction of the apical cell cortex [ 5 ]. The forces that contract the apical cell cortex are transmitted to neighboring cells through apical cell junctions. As a result, the contraction of a cortical network can shrink the exposed apical surface of the cell [ 6 , 7 ]. How this mechanism is developmentally regulated, driving specific cells to constrict only their apical surfaces and at specific times, remains incompletely understood in most model systems [ 2 , 8 ].

During embryogenesis, molecular forces drive the tissue shape changes that give form to the developing organism [ 1 ]. Among the mechanisms that drive such tissue shape changes, apical constriction is one of the most commonly used [ 2 ]. For example, the neural tube of vertebrate embryos forms as some cells of the neural plate constrict apically, bending the neural plate into a tube and internalizing from the embryo’s surface [ 3 ]. Neural tube formation fails frequently in human development [ 4 ]. Understanding the mechanisms that control changes to cell shape is essential to understanding disease states as well as the fundamental mechanisms by which embryos develop.

Results

Cortical actomyosin initially contracts away from stably positioned cadherin and catenins We considered that the failure of initial contractions of actomyosin networks to pull cell-cell junctions inward in C. elegans gastrulation [13] might be explained by an initial failure of α-catenin and/or β-catenin to remain stably at cell-cell junctions with cadherin. To test this hypothesis, we filmed embryos expressing fluorescent tags inserted into the native locus of each protein without disrupting each protein’s function [15]. As expected, mKate2-tagged HMR-1/cadherin localized at apical cell-cell boundaries, as did GFP::HMP-2/β-catenin and HMP-1/α-catenin::GFP (Fig 1A and 1B). We visualized the movements of actomyosin along with the CCC components during the early stage, when there are uncoupled contractions (2–8 minutes after the initiation of cytokinetic furrow formation in MSa and MSp cells [13]) using strains co-expressing a red fluorescently tagged non-muscle myosin II heavy chain (NMY-2::mKate) along with a GFP-tagged CCC component. We mounted these embryos ventrally to visualize en face centripetal actomyosin dynamics in the apical cortices of endoderm precursor cells (EPCs). During this early stage, we observed robust centripetal movement of myosin particles, but the bulk of the α-catenin-GFP and GFP-β-catenin remained stably at apical cell-cell boundaries, and failed to move centripetally (Fig 1B). We conclude that the initial inability of contracting actomyosin networks to efficiently pull cell-cell junctions centripetally cannot be explained by α-catenin and/or β-catenin being pulled away from cell-cell junctions. Although F-actin can associate with α-catenin various organisms including C. elegans [16,17], and this connection can be strengthened under force in some systems [18], these results imply that a strong connection between the contracting apical actomyosin network and junctional α-catenin is initially missing in this system (Fig 1B). We next sought to identify proteins that might contribute to this connection. PPT PowerPoint slide

PNG larger image

TIFF original image Download: Fig 1. All three CCC components remain associated with apical membrane borders during early stage actomyosin contractions. (A) Two-color spinning disk confocal fluorescence images of HMR-1::mKate2 (cadherin) with GFP::HMP-2 (β-catenin), left, and HMP-1::GFP (α-catenin) with NMY-2::mKate2 (myosin), right. Apically constricting cells are labeled (asterisks) (B) Representative line scans of the fluorescence intensities across the cell cortex from anterior (left) to posterior (right) with corresponding images of these junctions over time. As indicated by arrowheads, β-catenin (n = 3) and α-catenin (n = 7) remained at apical membrane borders while myosin particles moved centripetally. (A) Scale Bar = 5 μm (B) inset scale bar = 1 μm, kymograph frames represent 3 second intervals. https://doi.org/10.1371/journal.pgen.1010319.g001

Identification of candidate proteins that might contribute to coupling of contracting actomyosin networks and junctions We used two screening approaches to identify proteins that could feasibly contribute to connecting actomyosin to junctions in EPCs: We screened for proteins from early-stage embryos that co-immunoprecipitate (co-IP) with α-catenin, and proteins whose mRNAs became enriched in EPCs just prior to the onset of apical constriction. To identify proteins that co-IP with α-catenin, we used a strain expressing endogenously-tagged HMP-1::GFP/α-catenin. We performed co-IP using anti-GFP antibodies to pull down the CCC and any associated proteins. Because we were interested in proteins present during gastrulation in early embryogenesis, we enriched our samples for early stage (<50 cell) embryos (see Materials and methods). We used a strain expressing soluble GFP alone as a control from which to subtract contaminating proteins that co-purify with GFP. Our initial list of co-IP’d proteins contained each of the other CCC proteins at high peptide counts as expected, confirming that we pulled down intact CCCs from early stage embryos (S3 Table). Our list also contained more than 200 other proteins with at least one detectable peptide. We expect that this list includes proteins that interact with α-catenin in one or more cells as well as false positives. We view the possibility that the list may be enriched for gene products of interest as sufficient for our purpose of further screening. We further narrowed this list to only candidates whose genes were predicted to be expressed before or during the 24-cell stage of development using published single-cell mRNA sequencing data [19], and we removed common housekeeping genes (see Materials and methods). This resulted in 11 candidates with the potential to physically interact with the CCC for further screening (Table 1). PPT PowerPoint slide

PNG larger image

TIFF original image Download: Table 1. Genes Identified from a Combined Proteomic and Transcriptomic Screen for candidates. Gene names, α-catenin co-IP peptide counts, and EPC mRNA enrichment for each candidate are presented. Enrichment data are shown for genes with EPC enrichment over the whole embryo at that stage if EPC(s) also had higher RPKM than cells at earlier stages. Primers were designed to amplify each target gene sequence from cDNA (S2 Table). Each dsRNA was injected into young adults of the wild-type N2 strain of C. elegans. Embryos laid by injected worms were scored 24 hours after injection, filmed by DIC microscopy, and examined for Gad phenotypes. gdi-1 RNAi resulted in 100% sterility, failing to produce embryos for analysis, as seen previously [28]. https://doi.org/10.1371/journal.pgen.1010319.t001 Gastrulation in C. elegans relies on embryonic transcription [20]. Therefore it is possible that some proteins that contribute to the dynamically regulated linkage between actomyosin and junctions are encoded by genes that become transcribed in EPCs near the time that these cells are born. To identify a second set of candidates from genes expressed specifically in EPCs, we examined our previously-generated single-cell RNA-seq data of the first several cell cycles of development [19]. We considered genes whose mRNAs were enriched in EPCs compared to the rest of the embryo just prior to apical constriction (i.e. in the endodermal precursor cell E at the 8-cell stage, or its daughter cells Ea and Ep at the 24-cell stage) and enriched compared to cells of earlier stages. From this list, we selected 21 genes whose mRNAs were enriched at least 8-fold in EPCs vs the rest of the embryo at the 8-cell stage (16 genes) or the 24-cell stage (5 genes). To identify genes from both lists above that contribute to gastrulation in vivo, we then screened candidates by RNA interference (RNAi). Rather than feeding bacteria expressing double-stranded RNAs (dsRNAs), we used the more laborious method of injecting dsRNAs targeting each gene in order to maximize the likelihood of strongly disrupting gene functions [21]. We filmed embryos released from injected mothers and examined them for a gastrulation-defective (Gad) phenotype, defined as the two EPCs failing to fully internalize (i.e. with part of at least one of the cells not covered by any other cells) before dividing. Our RNAi screening identified 21 candidate genes with at least a low frequency Gad phenotype. These results define a set of genes that will be of interest for future studies of gastrulation mechanisms (Table 1). Among the genes we identified were two we chose to focus on because they encode proteins that associate directly or indirectly with junctional proteins and/or actin networks in C. elegans or other organisms: afd-1/afadin and zyx-1/zyxin [14,22–25]. The Drosophila afadin homolog Canoe is required along with β-catenin for medioapical actomyosin to remain connected to adherens junctions during apical constriction [26,27], although it has not been implicated in the kind of temporally-regulated linkage during normal development that we sought here. We first pursued afd-1/afadin’s roles in C. elegans apical constriction. zyx-1/zyxin is discussed further below.

RNA-seq of internalizing cell types to search for transcripts enriched in apically constricting cells of multiple cell lineages Multiple cell lineages of the early C. elegans embryo internalize by apical constriction [10,38]. Our identification of 21 genes with enriched expression in just one of these lineages, the EPCs, made us wonder if there exist any C. elegans genes with expression enriched in multiple independently internalizing cell types. No such gene might exist, but we considered this issue worth investigating because if such a gene existed, we would view it as a candidate for orchestrating apical constriction, akin to snail family genes that orchestrate another cell shape change–epithelial-to-mesenchymal transitions–in multiple animal models [39–42]. Alternatively, finding that no such gene exists by exhaustive RNAseq analysis could also inform future models of apical constriction mechanisms by suggesting that apical constriction may be orchestrated by different regulators in different cells. We collected cells from multiple internalizing cell lineages for RNA-seq (S1 Fig). We selected four groups of internalizing cells–MS descendants, E descendants, D descendants, and descendants of Cap and Cpp, hereafter referred to together as Cxp descendants (Fig 4A, red circles). We selected two non-internalizing groups as negative controls–ABp descendants (because only 2 out of 32 great-great granddaughter cells of ABp internalize) and Cxa descendants (none of which internalize) [38]. For each of the internalizing cell types, we sought transcripts whose abundance increased during the cell cycles leading up to internalization. To do this, we collected two transcriptomes from each internalizing cell lineage–one before internalization (2–3 cell cycles before cell internalization), and one at the start of internalization. The two non-internalizing cell types were used to exclude transcripts that became broadly enriched in all cell types at the relevant stages (Fig 4A, black circles). All transcriptomes from the 1- to 15-cell stages were previously published [19], and in this study we expanded the previous resource with transcriptomes from internalizing lineages. We present the results of the cumulative transcript dataset as a resource in an interactive online form, to facilitate querying the dataset, using the Differential Expression Gene Explorer, at http://dredge.bio.unc.edu/c-elegans-transcriptional-lineage-with-late-gastrulation/ (Permalink: https://n2t.net/ark:/84478/d/2bbpmsq3) [43]. PPT PowerPoint slide

PNG larger image

TIFF original image Download: Fig 4. Transcriptome profiling of multiple divergent, gastrulating cell types reveals internalization-correlated expression pattern of LIM domain-containing gene family. (A) C. elegans lineage map indicating the four internalizing lineages. All internalizing cells are represented with red branches. The internalizing MS, E, D, and Cxp lineages were dissected for transcriptome profiling and compared against two negative control lineages that do not internalize—the ABp and Cxa lineages. (B) Heatmap showing transcript abundance for the 99 internalization-correlated genes. Open circles indicate samples from a non-internalizing cell while closed circles indicate samples from an internalizing cell. (C) Pictograms of embryos representing relative gene expression levels of LIM domain-containing genes in the different internalizing cell lineages (key in S3 Fig). A mockup of an idealized gene expression pattern for a hypothesized orchestrator of gastrulation is shown on the left. The combined gene expression of all three LIM domain-containing genes is shown on the right. (D) DIC imaging of C. elegans embryos around the start of gastrulation in control (N2) embryos and in zyx-1Δ (LP831) embryos. The surfaces of internalizing EPCs that were not covered by other cells in 3 dimensions are outlined, and the black arrowhead points to these exposed EPC cell surfaces at the time of cell division, indicating a gastrulation defect that is not seen in the control. https://doi.org/10.1371/journal.pgen.1010319.g004 To identify any transcripts that became enriched in internalizing cell types, we filtered the 7,998 transcripts that we detected (see Materials and methods) for those that became enriched at least two-fold over time in at least two of the four internalizing cell types, and that did not become enriched by more than two-fold in the two negative control, non-internalizing cell types. This analysis yielded 839 genes. Of the four internalizing cell types sampled, all but the E lineage generate some muscle cells (all D descendants, all Cxp descendants, and 17/52 MS descendants will become body muscle) [44]. To avoid genes that are exclusively associated with muscle fate, we filtered the 839 transcripts for those that became enriched over time in the E lineage and at least one other internalizing lineage, and then for only genes whose transcripts are enriched by at least two-fold in the internalizing C descendants (Cxp) compared to the non-internalizing C descendants (Cxa). This reduced our filtered list to 150 genes. From this list we removed genes whose maximum transcript abundance in any non-internalizing cell types exceeded by more than two-fold the transcript abundance in the internalizing cell types where enrichment had been found. Of 99 genes that remained (Fig 4B), 55 transcripts were enriched in two of the four internalizing cell types, 33 were enriched in three of the four, and 11 were enriched in all four (as in the complete internalization-correlated pattern shown in Fig 4C). We found that all 11 were either only weakly expressed in some of the internalizing cells or exhibited relatively high expression in some non-internalizing cells as well (S3 Fig). We conclude that no single gene can be found by this kind of RNA-seq analysis that satisfies our expectations for a C. elegans orchestrator of apical constriction in multiple cell lineages. Therefore, we considered next the possibility that multiple members of a gene family might together fulfill this expected pattern.

Transcripts encoding a group of LIM domain-containing proteins become enriched in multiple apically constricting cells To expand our analysis to include groups of genes that are similar to each other in sequence, we created groups of genes based on similarity (see Materials and methods), calculating a cumulative transcript profile for each such homology group by summing the transcript profiles of all the genes in the group (as shown in the last pictogram in Fig 4C). We evaluated each of these summed transcript profiles as we had the individual genes above. This analysis yielded 51 homology groups: 27 groups that had transcripts that became enriched in two of the four internalizing cell types, 21 groups in three of the four internalizing cell types, and three groups enriched in all four internalizing cell types. Of the three homology groups whose transcripts became enriched in all four internalizing cell types, two included genes encoding F-box domains (S1 and S2 Files). C. elegans has an unusually large number of F-box-encoding genes [45]. None of the genes in the two groups we identified has a known function or known orthologs outside of the Caenorhabditid nematodes. The third group consisted of three genes encoding proteins that have LIM domains; lim-9 (an ortholog of LIMPET in Drosophila and FHL2 in vertebrates), pxl-1 (an ortholog of paxillin in Drosophila and vertebrates), and the zyx-1/zyxin gene that we had identified in a separate set of experiments above. Within this homology group, zyx-1 transcripts were enriched in the E lineage as described above; lim-9 transcripts were enriched in Cxp descendants (i.e., the cells of the C lineage that internalize) and not their non-internalizing Cxa sister cells; and pxl-1 transcripts were enriched in MS descendants and D descendants (Fig 4C). These genes encode proteins whose homologs are involved in actin filament organization in muscle cells [46–49], focal adhesions and mechanotransduction [50,51], stretch-induced gene expression [52], planar cell polarity, and asymmetric cell division [53]. LIM domain-containing proteins have been shown to interact physically with components of the actin cytoskeleton such as vinculin and α-actinin [46]. We consider members of this homology group as interesting candidates for regulators of cell behaviors during gastrulation based on the transcript enrichments that we found, their broad conservation across animals, as well as their known involvement in cytoskeletal organization. Therefore, we attempted to test whether these LIM domain-encoding genes are required during gastrulation in their respective cell lineages. We targeted each candidate individually by dsRNA injection, filmed embryos, and assayed for gastrulation defects among the internalizing cell lineages in which each candidate was found to be upregulated. Besides zyx-1 RNAi (see Table 1), none of the candidates yielded gastrulation defects in their respective cell lineages (pxl-1 RNAi, 0% MS or D lineage internalization defects, n = 24; lim-9 RNAi, 0% C lineage internalization defects, n = 46). C. elegans also has another LIM domain-encoding gene, unc-97, that is expressed at high levels in internalizing C lineage cells and moderately high levels in other internalizing lineages (see Differential Expression Gene Explorer link above). Because complex genetic redundancy among a multi-gene family and/or failure to sufficiently knock down transcript levels by RNAi might have prevented us from observing phenotypes after pxl-1 RNAi and lim-9 RNAi, and because we had found that targeting zyx-1/zyxin did result in gastrulation defects, we decided to set aside work on the larger set of LIM domain proteins to focus on characterizing ZYX-1/zyxin’s role in EPCs during gastrulation. Overall, our differential gene expression results led us to conclude that few genes can be found with expression patterns matching that expected for transcriptionally-regulated genes that might orchestrate apical constriction in diverse cell types, and it highlighted a possible role for LIM-domain-encoding genes including zyx-1. We confirmed zyx-1/zyxin’s role in gastrulation by generating a CRISPR knockout of zyx-1 (zyx-1Δ, LP831), removing the protein coding region and replacing it with a cassette encoding a codon-optimized GFP expressed under the control of the myo-2 promoter, driving expression in the pharynx to allow for easy visual identification of the allele. Consistent with the phenotype seen by RNAi, we also found gastrulation defects in the knockout strain. This phenotype was more penetrant than we had observed in the zyx-1 dsRNA injection (41.3% vs. 25%, 19/46 Gad embryos, Fig 4D). To determine if AFD-1 and ZYX-1 might function redundantly, we performed afd-1 RNAi in the zyx-1 Δ strain and found that afd-1 RNAi did not increase penetrance (9/36 Gad, 25%). Before examining whether ZYX-1 is involved in linking junctions to contracting apical actomyosin networks, we attempted to characterize its localization. We anticipated that the low level of zyx-1/zyxin transcripts that we detected in EPCs might make it difficult to visualize protein localization; indeed ZYX-1 might function specifically at a low level, during a time when levels are only beginning to rise transiently after initial gene expression. Previous authors have reported that zyx-1/zyxin produces 2 protein isoforms: a longer, 603 amino acid isoform called ZYX-1a, and a shorter, 200 amino acid isoform called ZYX-1b [46]. ZYX-1a contains 3 polyproline-rich repeats, a predicted nuclear export signal, and 3 tandem LIM domains (S4A Fig). We created a strain with mNeonGreen (mNG) inserted at the endogenous N-terminus to tag ZYX-1a, but consistent with its low predicted expression at this stage of development, we were unable to detect mNG signal in the EPCs, and we were unable to detect ZYX-1 by immunostaining methods that employed signal amplification (S4B and S4C Fig). In young adults we could see mNG::ZYX-1a readily in differentiated body wall muscle, neurons, gonads, and spermatheca, in line with where its expression was previously described [46,54] (S4D Fig). To assess where ZYX-1 could associate when accumulating in gastrulating cells, we examined where over-expressed ZYX-1 would localize using single-copy transgenes driven by the sdz-1 promoter (Psdz-1), which is predicted to drive ~20-fold overexpression compared to zyx-1 expression levels in EMS, E, and MS cell lineages (S5A Fig). We created two mNG-tagged constructs: one expressing full length ZYX-1a, and another expressing only the LIM domain-containing region (LCR) of ZYX-1 to examine where the LCR alone could direct localization. For both constructs, cytoplasmic mNG signal could be detected in E and MS cells as predicted, and small foci could be detected at the apical surfaces of internalizing EPCs (S5B Fig). Additionally, the predicted nuclear export signal of ZYX-1a appeared to be functional in full-length mNG::ZYX-1a, because mNG::ZYX-1a was excluded from the nucleus while mNG::LCRZYX-1 was not (S5C Fig). We conclude that ZYX-1a is likely expressed normally at too low a level as EPCs internalize to detect by current methods, and that it and its LCR can be recruited to apical foci in EPCs when overexpressed. One hypothesis consistent with our expression data and our phenotype data is that zyxin may be a limiting component required for triggering apical constriction that is expressed only briefly and at a low level at the onset of cell internalization. We attempted to test this hypothesis using the Psdz-1 overexpression construct to drive expression in the EMS cell (which produces both the E and MS lineages) to see if expressing zyxin early and at higher than normal levels might result in early cell internalization, but we did not see cell internalization occurring earlier in this strain (S6 Fig). In line with zyx-1’s expression enriched in only EPCs, Psdz-1-driven overexpression of full length mNG::ZYX-1a was able to rescue most of the defects seen in the zyx-1Δ background (2/27 Gad, p = 0.0003). Overexpression of mNG::LCRZYX-1 was not able to rescue similarly (7/22 Gad, p = 0.366). We conclude that ZYX-1 contributes to gastrulation, and that domains beyond the LIM-domain-containing region are important for this function. Next, we investigated whether ZYX-1/zyxin and AFD-1/afadin contribute specifically to coupling cell junctions to contracting actomyosin networks during gastrulation.

[END]
---
[1] Url: https://journals.plos.org/plosgenetics/article?id=10.1371/journal.pgen.1010319

Published and (C) by PLOS One
Content appears here under this condition or license: Creative Commons - Attribution BY 4.0.

via Magical.Fish Gopher News Feeds:
gopher://magical.fish/1/feeds/news/plosone/

You are viewing proxied material from magical.fish. The copyright of proxied material belongs to its original authors. Any comments or complaints in relation to proxied material should be directed to the original authors of the content concerned. Please see the disclaimer for more details.