A growing body of evidence suggests that cell division and basement membrane invasion are mutually exclusive cellular behaviors. How cells switch between proliferative and invasive states is not well understood. Here, we investigated this dichotomy in vivo by examining two cell types that derive from equipotent progenitors, but exhibit distinct cell behaviors, in the developing Caenorhabditis elegans somatic gonad: the post-mitotic, invasive anchor cell and the neighboring proliferative, non-invasive ventral uterine (VU) cells. We report that the default invasive cellular state is suppressed in the VU cells through two distinct modes of regulation of the pro-invasive transcription factor NHR-67 (NR2E1/TLX). Levels of NHR-67 are important for discriminating between invasive and proliferative behavior, and nhr-67 transcription is downregulated following post-translational degradation of its direct upstream regulator, HLH-2 (E/Daughterless) in VU cells. Residual NHR-67 protein is organized into discrete punctae in the nuclei of VU cells that are dynamic over the cell cycle and exhibit liquid-like properties. Strikingly, these NHR-67 punctae are not spatiotemporally associated with active transcription, but instead associate with homologs of the transcriptional co-repressor Groucho (UNC-37 and LSY-22), as well as the TCF/LEF homolog POP-1, likely mediated by a direct interaction between UNC-37 and the intrinsically disordered region of NHR-67. Further, perturbing UNC-37, LSY-22, or POP-1 results in ectopic invasive cells. We propose a model in which these proteins together form repressive condensates to suppress a default invasive state in non-invasive cells, which complements transcriptional regulation to add robustness to the proliferative-invasive cellular switch in vivo.
This paper reports valuable data on a classic model for cell fate specification in the C. elegans gonad. The data, though extensive, provide so far incomplete support for the major claims of the paper, especially regarding the functional significance of their conclusions. The paper will be of interest to developmental biologists studying transcriptional control of cell fate specification in animals, especially once issues around the functional significance of the condensates are resolved.
Cellular proliferation and invasion are key aspects of development (reviewed in (Medwig & Matus, 2017)), and are also two of the defining hallmarks of cancer (reviewed in (Hanahan and Weinberg, 2000)). A growing body of evidence suggests that cell cycle progression and invasion through a basement membrane are mutually exclusive cellular behaviors in both development and disease states (reviewed in (Kohrman and Matus, 2017)). Switching between invasive and proliferative phenotypes has been observed in melanoma and recently in breast cancer (Hoek et al., 2008Mondal et al., 2021), but how these cell states are regulated in the context of development is not well understood. To investigate how this dichotomy in cellular behavior is controlled in vivo, we used C. elegans, leveraging its highly stereotypical development (Sulston and Horvitz, 1977), as well as its genetic and optical tractability. During development of the hermaphroditic reproductive system, the proximal granddaughters of the Z1 and Z4 somatic gonad progenitors, Z1.pp and Z4.aa, give rise to four cells that will adopt one of two cellular fates: a proliferative ventral uterine (VU) cell or the terminally differentiated, invasive anchor cell (AC) (Figure 1A) (Kimble and Hirsh, 1979). The distal cells of this competency group, Z1.ppa and Z4.aap, quickly lose their bipotentiality and become VU cells (Seydoux et al., 1990). In contrast, the proximal cells, Z1.ppp and Z4.aaa, undergo a stochastic Notch-mediated cell fate decision, giving rise to another VU cell and the post-mitotic AC (Figure 1A,B) (Greenwald et al., 1983Seydoux and Greenwald, 1989). Following fate specification, the AC undergoes invasive differentiation and breaches the underlying basement membrane, connecting the uterus to the vulval epithelium to facilitate egg-laying (Figure 1B) (Sherwood and Sternberg, 2003).
Our previous work has shown that AC invasion is dependent on G0 cell cycle arrest, which is coordinated by the pro-invasive transcription factor NHR-67 (NR2E1/TLX) (Figure 1–figure supplement 1A) (Matus et al., 2015). NHR-67 functions within a gene regulatory network comprised of four conserved transcription factors whose homologs have been implicated in several types of metastatic cancer (Liang and Wang, 2020Milde-Langosch, 2005Nelson et al., 2021Wang and Baker, 2015). We previously reported that NHR-67 is regulated by a feed-forward loop formed by EGL-43 (Evi1) and HLH-2 (E/Daughterless), which functions largely in parallel to a cell cycle-independent subcircuit controlled by FOS-1 (Fos) (Figure 1–figure supplement 1A) (Medwig-Kinney et al., 2020). EGL-43, HLH-2, and NHR-67 are reiteratively used within the Z lineage of the somatic gonad, in that they also function to independently regulate LIN-12 (Notch) signaling during the initial AC/VU cell fate decision (Medwig-Kinney et al., 2020). Despite its role in lateral inhibition between Z1.ppp and Z4.aaa, expression of LIN-12 is not absolutely required for VU cell fate (Sallee et al., 2015). Cell cycle state also cannot explain the difference between AC and VU cell fates, as arresting VU cells in G0 through ectopic expression of CKI-1 (p21/p27) does not make them invasive (Smith et al., 2022). Thus, the mechanisms responsible for maintaining AC and VU cellular identities following initial cell fate specification remain unclear.
Maintenance of differentiated cell identity is essential for ensuring tissue integrity during development and homeostasis, and the inability to restrict phenotypic plasticity is now being recognized as an integral part of cancer pathogenesis (Hanahan, 2022). In vitro studies have identified several factors that safeguard differentiated cell identity (reviewed in (Brumbaugh et al., 2019)). Despite its largely autonomous modality of development, C. elegans has emerged as an ideal model system to study cell fate maintenance in vivo. There have been several reports of cell fate transformations that occur naturally, including two epithelial-to-neural transdifferentiation events (Jarriault et al., 2008Riva et al., 2022), or following fate challenges (reviewed in (Rothman and Jarriault, 2019)). In such contexts, several epigenetic factors, including chromatin remodelers and histone chaperones, have been identified for their roles in restricting cell fate reprogramming (Hajduskova et al., 2019Kagias et al., 2012Kolundzic et al., 2018Patel et al., 2012Rahe and Hobert, 2019Zuryn et al., 2014). However, in some cases, ectopic expression of a specific transcription factor is sufficient to overcome these barriers, as was first shown through pioneering work in mouse embryonic fibroblasts (Davis et al., 1987). Indeed, there are several examples in C. elegans where ectopic expression of single lineage-specific transcription factors induces cell fate transformations (Fukushige and Krause, 2005Gilleard and McGhee, 2001Horner et al., 1998Jin et al., 1994Kiefer et al., 2007Quintin et al., 2001Richard et al., 2011Riddle et al., 2013Tursun et al., 2011Zhu et al., 1998). Moreover, C. elegans uterine tissue may be particularly amenable to fate transformations, as ectopic expression of a single GATA transcription factor, ELT-7, is sufficient to induce transorganogenesis of the somatic gonad into gut by reprogramming the mesodermally-derived tissue into endoderm (Riddle et al., 2016). Valuable insights have been made into how the function of fate-specifying transcription factors can be tuned through means such as autoregulation and dynamic heterodimerization (Leyva-Díaz and Hobert, 2019Sallee et al., 2017). We are just beginning to understand how an additional layer of control over transcriptional regulators can be achieved through the formation of higher order associations (Boija et al., 2018Lim and Levine, 2021).
Here, in our endeavor to understand how AC and VU cellular fates are maintained, we identified two mechanisms of NHR-67 regulation that together modulate the invasive-proliferative switch in C. elegans. We found that high levels of NHR-67 expression are sufficient to drive invasive differentiation and, accordingly, nhr-67 is transcriptionally downregulated in the non-invasive VU cells following the post-translational degradation of its direct upstream regulator, HLH-2. Additionally, we observed that remaining NHR-67 protein forms discrete punctae in the nuclei of non-invasive cells that exhibit liquid-like properties including dynamic assembly, fusion, and dissolution over cell cycle as well as rapid recovery kinetics after photobleaching. These NHR-67 punctae colocalize in vivo with UNC-37 and LSY-22, homologs of the transcriptional co-repressor Groucho, as well as with POP-1 (TCF/LEF), which are mediated through a direct interaction between UNC-37 and the intrinsically disordered C-terminal region of NHR-67. Through functional perturbations, we demonstrate that UNC-37, LSY-22, and POP-1 normally function in repressing the default invasive state in VU cells. We propose a model in which the interaction between NHR-67 and Groucho coordinates formation of repressive condensates that, combined with transcriptional downregulation of nhr-67, suppress invasive differentiation. This work provides new insights into how repressive nuclear condensates may coordinate cellular behaviors in vivo and highlights how transcription factors can exhibit duality in functions depending on cellular context.
Levels of NHR-67 expression are important for distinguishing AC and VU cell identity
Despite arising from initially equipotent cells, the differentiated AC and VU cells exhibit very distinct cellular behaviors. The AC terminally differentiates to invade the underlying basement membrane while the VU cells remain proliferative, undergoing several rounds of division before terminally differentiating into the pi (π) and rho (ρ) cells that function in uterine-vulval attachment. One potential explanation for this difference in cell behavior is asymmetric expression of pro-invasive transcription factors. To investigate this possibility, we examined endogenous expression levels of four transcription factors that function in the gene regulatory network coordinating AC invasion (EGL-43, FOS-1, HLH-2, and NHR-67) using previously generated GFP-tagged alleles (Medwig-Kinney et al., 2020). While FOS-1 levels of expression in the AC are nearly twice that of the VU cells (Figure 1–figure supplement 1B,C), FOS-1 has no identified role in regulating cell cycle in the AC so we did not pursue this protein further (Medwig-Kinney et al., 2021). EGL-43 also did not appear to be a promising candidate, as it is expressed in both cell types at comparable levels, with VU cells exhibiting approximately 89% of AC expression (Figure 1–figure supplement 1B,C). In contrast, HLH-2 exhibits significant asymmetry in expression, as VU cells express merely 17% of HLH-2 levels observed in the AC on average (Figure 1C,D). Previous studies have shown that post-translational, dimerization-driven degradation of HLH-2 is responsible for its downregulation in the VU cells (Benavidez et al., 2022Karp and Greenwald, 2003Sallee and Greenwald, 2015). Endogenously tagged NHR-67::GFP exhibits a similar pattern of expression with over three-fold enrichment in the AC, consistent with prior observations of transgenic reporters (Figure 1C,D) (Verghese et al., 2011). Given the known role of NHR-67 in regulating cell cycle arrest and invasion, we hypothesized that its differential expression between the AC and VU cells could potentially be contributing to their distinct cellular behaviors.
To assess if NHR-67 plays a role in regulating uterine cell identities, we manipulated its expression levels. We found that strong depletion of NHR-67 through RNA interference (RNAi) treatment results in ACs adopting VU-like characteristics. During AC/VU cell fate specification, LIN-12/Notch normally becomes restricted to the VU cells while the Delta-like ligand LAG-2 (visualized by LAG-2::P2A::H2B::mTurquoise2 (Medwig-Kinney et al., 2022)) accumulates in the AC (Wilkinson et al., 1994). Here, we observe that NHR-67 deficient ACs not only proliferated and failed to invade, as reported previously (Matus et al., 2015), but also ectopically expressed membrane-localized Notch (visualized by LIN-12::mNeonGreen (Pani et al., 2022)) (Figure 1E). Notably, NHR-67-deficient ACs expressed both LIN-12 and LAG-2, potentially indicating an intermediate state between AC and VU cell fate (Figure 1E). Next, we ectopically expressed NHR-67 ubiquitously using a heat shock inducible transgene (hsp::NHR-67::2x-BFP) (Medwig-Kinney et al., 2020). Intriguingly, we observed that ectopic expression of NHR-67 following initial AC/VU specification resulted in the presence of multiple invasive ACs at a low penetrance (approximately 5%, n > 50), denoted by ectopic expression of an AC marker (cdh-3p::mCherry::moeABD) and expansion of the basement membrane gap (Figure 1F). Since it has been previously demonstrated that proliferative ACs cannot invade (Matus et al., 2015), we concluded that these invasive ectopic ACs most likely arose from fate conversion of neighboring VU cells. Taken together, these pieces of evidence suggest that high and low levels of NHR-67 expression correlate to properties of AC and VU cell identities, respectively (Figure 1G).
NHR-67 is enriched in the AC through direct transcriptional regulation by HLH-2
Next, we investigated how NHR-67 expression levels become asymmetric between the AC and VU cells. We and others have previously shown that HLH-2 positively regulates NHR-67 expression in the context of the AC (Figure 1–figure supplement 1A) (Bodofsky et al., 2018Medwig-Kinney et al., 2020). If this regulatory interaction exists in the context of the VU cells as well, it could explain why the relative expression pattern of NHR-67 in the AC and VU cells mirrors that of HLH-2. In support of this hypothesis, we found that initial onset of HLH-2, which has shown to be asymmetric in Z1.pp and Z4.aa (Attner et al., 2019), correlates to that of NHR-67 onset (Figure 2–figure supplement 1A). To test whether HLH-2 degradation is responsible for NHR-67 downregulation in the VU, we drove ectopic expression of HLH-2 using a transgene under the control of a heat shock inducible promoter (hsp::HLH-2::2x-BFP) (Medwig-Kinney et al., 2020). We observed that ectopic expression of HLH-2 resulted in elevated NHR-67 expression in VU cells (43% increase; n > 30) (Figure 2A,B). To control against potential dimerization-driven degradation of HLH-2 in the VU cells, which the heat shock inducible transgene would still be susceptible to, we disrupted UBA-1, an E1 ubiquitin-activating enzyme that has recently been shown to be necessary for HLH-2 degradation in VU cells (Benavidez et al., 2022). Following perturbation of UBA-1 through RNAi treatment, HLH-2 expression in the VU cells increased more than four-fold and NHR-67 expression increased by nearly 60% compared to the empty vector control (Figure 2–figure supplement 1B-D). Both experiments suggest that nhr-67 expression in the VU cells is at least partially regulated by levels of HLH-2.
It has previously been proposed that the interaction between HLH-2 and nhr-67 is direct. This is based on the identification of E binding motifs within a 276 bp region of the nhr-67 promoter that is required for NHR-67 expression in the uterine tissue and encompasses the location of several hypomorphic mutations (pf2, pf88, pf159) (Figure 2C) (Bodofsky et al., 2018Verghese et al., 2011). We confirmed this interaction through a yeast one-hybrid assay after generating a bait strain containing this nhr-67 promoter region and pairing it with an HLH-2 Gal4-AD prey plasmid from an existing yeast one-hybrid library (Reece-Hoyes et al., 2005). Yeast growth on the selective SC-HIS-TRP plates containing the competitive inhibitor 3-aminotriazole (3-AT) demonstrated that HLH-2 does indeed bind directly to this 276 bp region of the nhr-67 promoter (Figure 2D). Together, these results suggest that direct transcriptional regulation of nhr-67 by HLH-2 contributes to the asymmetry in NHR-67 expression between the AC and VU cells.
NHR-67 forms dynamic punctae in VU cell nuclei that exhibit liquid properties
Upon closer examination of GFP-tagged NHR-67, it became evident that the AC and VU cells not only exhibit differences in overall NHR-67 levels, but also in localization of the protein. While NHR-67 localization is fairly uniform throughout the AC nucleus (excluding the nucleolus), we often observed discrete punctae throughout the nuclei of VU cells (Figure 3A,B). These punctae were observed with NHR-67 endogenously tagged with several different fluorescent proteins, including GFP, mNeonGreen, mScarlet-I, and TagRFP-T (Figure 3–figure supplement 1A,B). Furthermore, by utilizing a live-cell imaging approach, we would not expect to encounter artificial puncta formation that can result from tissue fixation methods (Irgen-Gioro et al., 2022). Thus, NHR-67 puncta formation in the VU cells does not appear to be an artifact of the fluorophore or sample preparation.
To characterize dynamics of these punctae during interphase states of the cell cycle, we paired GFP-tagged NHR-67 with a CDK activity sensor. The CDK activity sensor is comprised of a fragment of DNA Helicase B (DHB) fused to a fluorophore (2x-mKate2), expressed under a ubiquitous promoter (Figure 3C) (Adikes et al., 2020). DHB contains a strong nuclear localization signal (NLS), flanked by four serine sites, as well as a weaker nuclear export signal (NES). As CDK activity increases over the cell cycle, the CDK sensor is translocated from the nucleus to the cytoplasm, allowing for correlation of its relative subcellular localization to cell cycle state (Figure 3C) (Adikes et al., 2020Spencer et al., 2013). Time-lapse microscopy revealed that the number of NHR-67 punctae was dynamic over the course of the cell cycle, with punctae first appearing shortly after mitotic exit in the G1 phase, and then reducing in number to two large punctae prior to nuclear envelope breakdown before disappearing (Figure 3D,E). We collected additional recordings with finer time resolution and captured fusion, or condensation, of punctae prior to their dissolution (representative of 6 biological replicates) (Figure 3F). These punctae also exhibit relatively rapid diffusion kinetics, as observed by fluorescence recovery following photobleaching (t1/2 = 46 seconds; n = 8) at a rate within the same order of magnitude as P granule proteins PGL-1 and PGL-3 (Figure 3G,H) (Putnam et al., 2019). These properties of NHR-67 punctae are consistent with those observed with proteins that form nuclear condensates.
Groucho homologs UNC-37 and LSY-22 associate with NHR-67 punctae and contribute to VU cell fate
In a first step towards defining the role of putative NHR-67 condensates, we tested the extent to which NHR-67 punctae colocalized with homologs of other proteins known to form nuclear condensates by pairing GFP- and mScarlet-I-tagged NHR-67 with other endogenously tagged alleles. As NHR-67 is a transcription factor, it is reasonable to speculate that its punctae may represent clustering around sites of active transcription, which would be consistent with data showing RNA Polymerase II and the Mediator complex can associate with transcription factors through phase separation (Cho et al., 2018). To test this hypothesis, we co-visualized NHR-67 with a GFP-tagged allele of ama-1, the amanitin-binding subunit of RNA polymerase II (Hills-Muckey et al., 2021) and failed to observe significant colocalization between NHR-67 and AMA-1 punctae (Manders’ overlap coefficient, M = 0.066) compared to negative controls where a single channel was compared to its 90-degree rotation (M = 0.108) (Figure 4A,B). Another possibility considered is that NHR-67 localization is indicative of chromatin organization, as heterochromatin has been shown to be compartmentalized in the nucleus through phase separation (Larson et al., 2017Strom et al., 2017). However, we did not observe significant colocalization of NHR-67 with the endogenously tagged HP1 heterochromatin proteins (Patel and Hobert, 2017) HPL-1 (M = 0.076) or HPL-2 (M = 0.083) (Figure 4A,B). Recent work in Ciona embryos has shown that the transcriptional co-repressor Groucho forms repressive condensates in nuclei through phase separation (Treen et al., 2021). The C. elegans genome encodes one Groucho homolog, UNC-37, as well as a Groucho-like protein, LSY-22. To examine their localization compared to NHR-67, we tagged LSY-22 with TagRFP-T (Figure 4–figure supplement 1) and acquired a mNeonGreen-tagged allele of unc-37 (Ma et al., 2021). Strikingly, we observed significant colocalization of NHR-67 punctae with both LSY-22 (M = 0.686) and UNC-37 (M = 0.741), comparable to colocalization measures in heterozygous NHR-67::mScarlet-I/NHR-67::GFP animals (M = 0.651), which were used as positive controls (Figure 4A,B). This evidence suggests that NHR-67 punctae do not localize to sites of active transcription or chromatin compaction, but instead associate with transcriptional co-repressors.
Since the AC is the default state of the AC/VU cell fate decision (Seydoux and Greenwald, 1989), we hypothesized that the punctae including NHR-67, UNC-37, and LSY-22 may function in repressing invasive differentiation. To test this hypothesis, we depleted UNC-37 and LSY-22 utilizing the auxin inducible degron (AID) protein degradation system, in which a protein of interest is tagged with an AID that is recognized by TIR1 in the presence of auxin and ubiquitinated by the SCF E3 ubiquitin ligase complex (Figure 4C) (Martinez et al., 2020Zhang et al., 2015). We re-tagged LSY-22 with mNeonGreen::AID (Figure 4–figure supplement 1) and acquired a BFP::AID-tagged allele of unc-37 (Kurashina et al., 2021). Each AID-tagged allele was paired with a transgene encoding Arabidopsis thaliana TIR1 (AtTIR1) that was co-expressed with a nuclear-localized mCherry::HIS-11. Following auxin treatment, we observed ectopic expression of an AC marker (cdh-3p::mCherry::moeABD) in 28% of LSY-22::AID animals and 59% of UNC-37::AID animals (n = 64 for both) (Figure 4D). These results are consistent with phenotypes we observed in genetic backgrounds with unc-37 hypomorphic (unc-37(e262wd26)) and null (unc-37(wd17wd22)) mutant alleles (Figure 4–figure supplement 2). It is likely that dual depletion of UNC-37 and LSY-22 would result in a higher penetrance of ectopic ACs given their partial redundancy in function (Flowers et al., 2010), but animals possessing both AID-tagged alleles were not viable when paired with the AtTIR1 transgene.
TCF/LEF homolog POP-1 associates with NHR-67/Groucho punctae and is necessary for VU cell maintenance post-specification
While UNC-37/LSY-22 colocalization with NHR-67 punctae and ectopic AC phenotypes are consistent with roles in coordinating AC and VU cell fates, both genes are broadly expressed and exhibit comparable levels (<10% difference) between the two cell types (Figure 5A,C; Figure 5–figure supplement 1A,B). Therefore, we hypothesized that another factor must be involved that confers VU cell specificity to Groucho-mediated repression of invasiveness. It had previously been reported that the sole TCF/LEF homolog in C. elegans, POP-1, forms a repressive complex with UNC-37 in the early embryo to restrict expression of the endoderm-determining gene, END-1 (Calvo et al., 2001). Additionally, POP-1 has a known role in development of the somatic gonad, as perturbing its function results in ectopic ACs (Siegfried and Kimble, 2002). Examination of an eGFP-tagged pop-1 allele (van der Horst et al., 2019), showed significant enrichment in the VU cells (>20%) compared to the AC (Figure 5B,C; Figure 5–figure supplement 1A,B). We also observed that POP-1 forms punctae in the nuclei of VU cells, which had previously been observed during interphase in non-Wnt signaled embryonic cells (Maduro et al., 2002). We found that these POP-1 punctae colocalize with NHR-67 (M = 0.547), although to a lesser degree than UNC-37 and LSY-22, likely because the strong POP-1 fluorescence outside of punctae made them more difficult to segment (Figure 5D,E). Additionally, nhr-67(RNAi) treatment resulted in a significant increase in AC expression of eGFP::POP-1 compared to empty vector controls (225%, n > 30), a pattern we observed following depletion of other transcription factors (Medwig-Kinney et al., 2020) and chromatin modifiers (Smith et al., 2022) required for AC arrest and invasion (Figure 5F,G; Figure 5–figure supplement 2A,B). This negative regulation of POP-1 by NHR-67 may explain why the proteins have opposite patterns of enrichment.
It has previously been suggested that POP-1 may be functioning as an activator in the VU precursors Z1.ppa and Z4.aap based on the relative expression of a POP-1 transgene (Sallee et al., 2015). This view is largely dependent on the notion that high levels of POP-1 correlate to repressive function and low levels are conducive for activator roles (Shetty et al., 2005). In contrast, we did not find evidence of transcriptional activation by POP-1 in the AC/VU precursors nor their differentiated descendants using an established POPTOP (POP-1 and TCF optimal promoter) reporter, which contains seven copies of POP-1/TCF binding sites and the pes-10 minimal promoter (Figure 5F; Figure 5–figure supplement 3A,B) (Green et al., 2008). The growing consensus regarding the Wnt/β-catenin asymmetry pathway is that relative levels of POP-1 and β-catenin are more important than absolute protein levels of POP-1 (Phillips and Kimble, 2009). Our proposed model of POP-1 acting as a repressor in the proximal gonad is consistent with the finding that SYS-1 (β-catenin) expression is restricted to the distal gonad early in somatic gonad development and is not detectable in the AC or VU cells (Figure 5–figure supplement 3C) (Phillips et al., 2007Sallee et al., 2015). It is also supported by recent evidence suggesting that UNC-37/LSY-22 mutant alleles phenocopy pop-1 knockdown, which produces ectopic distal tip cells (Bekas and Phillips, 2022).
One aspect that makes studying the repressive role of POP-1 in cell fate maintenance challenging is that its activator function is required for distal cell fate specification in the somatic gonad earlier in development. Loss of either POP-1 and SYS-1 results in a Sys (symmetrical sister cell) phenotype, where all somatic gonad cells adopt the default proximal fate and thereby give rise to ectopic ACs (Siegfried and Kimble, 2002Siegfried et al., 2004). This likely occluded previous identification of the repressive role of POP-1 in maintaining VU cell fates. To achieve temporal control over POP-1 expression to tease apart its two opposing roles, we inserted an AID tag into the N-terminus of the pop-1 locus; however, this resulted in gonadal defects even in the absence of both TIR1 and auxin. Instead, using tools at hand, we paired eGFP-tagged POP-1 with a uterine-specific anti-GFP nanobody (Smith et al., 2022Wang et al., 2017). The anti-GFP nanobody is fused to ZIF-1 and serves as an adapter, recognizing GFP-tagged proteins and promoting their ubiquitination by the Cullin2-based E3 ubiquitin ligase, which ultimately targets them for degradation via the proteasome (Figure 6–figure supplement 1A) (Wang et al., 2017). This anti-GFP nanobody, visualized by nuclear expression of mCherry, was not detectable prior to or even shortly after the AC/VU cell fate decision, which allowed us to bypass disruption of initial cell specification (Figure 6–figure supplement 1B). While this method only produced a mild knockdown of POP-1 in the VU cells, we still observed the ectopic AC phenotype at low penetrance (7%, n = 60) (Figure 6–figure supplement 1C). To achieve stronger depletion, we used RNAi for further POP-1 perturbations.
To interrogate the phenotypic consequences of POP-1 perturbation, we utilized a strain expressing two markers of AC fate (cdh-3p::mCherry::moeABD and LAG-2::P2A::H2B::mTurquoise2). Following treatment with pop-1(RNAi), we observed several animals with two or more bright cdh-3/lag-2+ ACs, consistent with known phenotypes caused by cell fate misspecification in the somatic gonad (17%, n = 30) (Figure 6A). We also observed animals with invasive cells that express AC markers at inconsistent levels (53%, n = 30), suggesting that the cells did not adopt AC fate at the same time (Figure 6A). To test whether the subset of dim cdh-3/lag-2+ ACs are the result of VU-to-AC cell fate conversion, we visualized AC and VU fates simultaneously using the AC markers previously described along with an mNeonGreen-tagged allele of lag-1 (CSL), a protein downstream of Notch signaling whose expression becomes restricted to the VU cells following AC/VU cell fate specification. Following treatment with pop-1(RNAi), we found that a subset of ectopic ACs co-express AC markers and LAG-1, likely indicating an intermediate state between the two cell types (Figure 6– figure supplement 2). To visualize this process live, we used time-lapse microscopy and were able to capture ectopic ACs gradually upregulating LAG-2 (+51%, n = 3) and downregulating LAG-1 (−16%, n = 3) over time (Figure 6B,C), consistent with VU-to-AC cell fate conversion.
IDR of NHR-67 facilitates protein-protein interaction with UNC-37
Given that UNC-37, LSY-22, and POP-1 phenocopy each other with respect to AC/VU fates and all three colocalize with NHR-67 punctae, we next sought to further characterize the interactions among these proteins. Previous work has either directly identified or predicted protein-protein interactions among POP-1, UNC-37, and LSY-22 (Boxem et al., 2008Calvo et al., 2001Flowers et al., 2010Reece-Hoyes et al., 2005Simonis et al., 2009Zhong and Sternberg, 2006). Using a yeast two-hybrid assay with UNC-37 Gal4-AD prey, we confirmed that UNC-37 directly interacts with both POP-1 and LSY-22 after observing yeast growth on the selective SC-HIS-TRP-LEU plates containing 3-AT (Figure 7–figure supplement 1). Using the same technique, we found that NHR-67 binds directly to UNC-37, as previously predicted (Li et al., 2004Simonis et al., 2009), but found no evidence of it directly interacting with LSY-22 or POP-1 (Figure 7–figure supplement 1).
To further characterize the protein-protein interaction between NHR-67 and UNC-37, we assessed the protein structure of NHR-67 using AlphaFold, an artificial intelligence-based protein structure prediction tool (Jumper et al., 2021Varadi et al., 2022), and PONDR, a predictor of intrinsic disorder (Peng and Zhang, 2006). Both identify an intrinsically disordered region (IDR) at the C-terminus of NHR-67 (Figure 7A,B). IDRs are low complexity domains that lack fixed three-dimensional structure and have been shown to support dynamic protein-protein interactions (Chong et al., 2018). To determine if the IDR of NHR-67 is important for facilitating its interaction with UNC-37, we repeated the yeast two-hybrid experiment using UNC-37 Gal4-AD prey, pairing it with different fragments of the NHR-67 protein: full-length, without its IDR (ΔIDR), and its IDR alone (Figure 7C,D). Yeast growth on the selective SC-HIS-TRP-LEU plates containing the competitive inhibitor 3-aminotriazole (3-AT) demonstrates that the 108 amino acid IDR sequence of NHR-67 is necessary and sufficient to bind with UNC-37 (Figure 7C,D).
Thus, our current model supported by the data shown here proposes that NHR-67 levels are controlled through two distinct mechanisms. First, transcription of nhr-67 is directly regulated by HLH-2, resulting in enrichment in the AC compared to the VU. Second, in the AC, where NHR-67 levels are high and POP-1 is repressed, NHR-67 is free to activate genes promoting invasive differentiation. In the VU cells, where NHR-67 levels are low and POP-1 levels are high, POP-1 assembles with LSY-22, UNC-37, and NHR-67 to repress NHR-67 targets (Figure 7E). It is possible that POP-1 negatively regulates NHR-67 at the transcriptional level as well, as the nhr-67 promoter contains 7 putative TCF binding sites (Zacharias et al., 2015). Taken together, our findings reveal a dual mechanism for repressing NHR-67 activity in the proliferative VU cells, maintaining their proliferative fates while suppressing the acquisition of an invasive phenotype.
In summary, here we provide evidence that activity of the pro-invasive transcription factor, NHR-67, is simultaneously regulated by two distinct processes, which together modulate the proliferative-invasive switch in C. elegans. We show that NHR-67 is a potent fate-specifying transcription factor, in that its expression is sufficient for invasive differentiation of ACs in the somatic gonad. This could explain why NHR-67 needs to be post-translationally sequestered in addition to being transcriptionally downregulated in non-invasive cells to fully suppress its function in activating the pro-invasive program. We also discovered that NHR-67 forms nuclear foci in non-invasive cells, which exhibit liquid-like properties, indicated by observations of their condensation, dissolution, and relatively rapid recovery from photobleaching, similar to what has been described with P granules (Brangwynne et al., 2009). These NHR-67 punctae associate with Groucho homologs, UNC-37 and LSY-22, through a direct protein-protein interaction with UNC-37 mediated by the C-terminal IDR of NHR-67. We postulate that this association leads to formation of repressive condensates, as has recently been described in Ciona embryos (Treen et al., 2021). Furthermore, Groucho-mediated repression of the default invasive state appears to be dependent on expression of the TCF/LEF homologs POP-1, which clarifies our understanding of the dual roles this protein plays during development of the somatic gonad. It is also interesting to note that the dynamic punctae formed by POP-1 in non-Wnt signaled cells was first described 20 years ago (Maduro et al., 2002), but their function are only now being appreciated in light of recent advances in our understanding of the formation of higher order associations in the nucleus.
With regard to phase separation in the nucleus, most research has been through the lens of transcriptional activation through RNA Polymerase II and the mediator complex (Boija et al., 2018Cho et al., 2018Sabari et al., 2018) or repression through condensate formation with HP1 heterochromatin proteins (Larson et al., 2017Strom et al., 2017). Rather than functioning through either of these established mechanisms, we have identified here the second observed case of Groucho-mediated repressive condensates (Treen et al., 2021), which suggests that Groucho proteins may have evolutionarily conserved roles in repressing transcription through the formation of nuclear condensates within the Metazoa. Further, we demonstrate how this repressive mechanism can control cell biology in vivo using endogenously tagged alleles and a clear cell biological read-out of fate and function: invasion versus proliferation.
Still, as this is one of the first studies into the role of repressive condensates in vivo, there is much left to learn. For example, it is unknown whether DNA binding is necessary for nuclear puncta formation. The interaction between UNC-37 and NHR-67 does not appear to depend on DNA binding, as the C-terminal IDR region of NHR-67 (excluding its zinc finger domains) was sufficient for binding with UNC-37 in vitro, but it is possible that DNA binding is needed for oligomerization in vivo. Furthermore, it remains unclear if suppression of invasive differentiation is achieved by simply sequestering the pro-invasive transcription factor NHR-67 away from its transcriptional targets or through direct repression of transcription. If the latter, another question that arises is how the repressive complex gets recruited to specific genomic sites, since POP-1 and NHR-67 are both capable of binding to DNA, and whether repression is achieved through competition with transcriptional activators or recruitment of histone deacetylases. Direct targets of NHR-67 have not yet been discovered, which makes it difficult to investigate this specific aspect of the repressive mechanism at present. We see this as a promising avenue of future study as technologies advance, allowing for transcriptional profiling and target identification in specific tissues or cells (Gómez-Saldivar et al., 2020Katsanos and Barkoulas, 2022).
In this work, we have also identified several perturbations (i.e., increasing levels of NHR-67, decreasing levels of UNC-37/LSY-22) that result in incompletely penetrant transdifferentiation phenotypes and/or intermediate cell fates. We foresee these being ideal cell fate challenge backgrounds in which to perform screens to identify regulators of cellular plasticity, as has been done in other contexts (Rahe and Hobert, 2019).
Additionally, these induced fate transformations can be paired with tools to visualize and manipulate cell cycle (Adikes et al., 2020) to determine if any cell cycle state is particularly permissive for cell fate plasticity. While G1 arrest has been shown to enhance conversion of human fibroblasts to dopaminergic neurons (Jiang et al., 2015), mitosis is required for the natural K-to-DVB transdifferentiation event in C. elegans (Riva et al., 2022). As control of proliferation and invasion, as well as maintenance of differentiated cellular identities, are important for both homeostatic and disease states, it is our hope that this work will shed light on how cells switch between these states in the context of cancer growth and metastasis.
Materials and methods
Key resources table
C. elegans strains, culture, and nomenclature
Methods for C. elegans culture and genetics were followed as previously described (Brenner, 1974). Developmental synchronization for experiments was achieved through alkaline hypochlorite treatment of gravid adults to isolate eggs (Porta-de-la-Riva et al., 2012). L1 stage animals were plated on nematode growth media plates and subsequently cultured at 20°C or 25°C. Heat shock-inducible transgenes were activated by incubating animals on plates sealed with Parafilm in a 33°C water bath for 2-3 hours. In the text and in figures, promoter sequences are designated with a “p” following the gene name and gene fusions are represented by a double-colon (::) symbol.
New alleles and single-copy transgenes were generated by homology directed repair using CRISPR-based genome engineering. mScarlet::AID and mNeonGreen::AID were inserted into the C-terminus of the nhr-67 locus by injecting adult germlines with Cas9 guide-RNA ribonucleoprotein complexes and short single-stranded oligodeoxynucleotide donors, as previously described (Ghanta and Mello, 2020). Successful integrants were identified through screening for fluorescence and by PCR. The lsy-22 locus was edited by injecting a Cas9 guide RNA plasmid and repair template plasmid containing a self-excising cassette with selectable markers to facilitate screening (Dickinson et al., 2015Dickinson and Goldstein, 2016Huang et al., 2021). Repair templates used to tag lsy-22 with TagRFP-T::AID and mNeonGreen::AID were generated by cloning ∼750-850 bp homology arms into pTNM063 and pDD312, respectively (Ashley et al., 2021Dickinson et al., 2015). All guide and repair sequences used can be found in Supplemental Table 1.
The GFP-tagged alleles of the pro-invasive transcription factors (egl-43, fos-1, hlh-2, and nhr-67) and the TagRFP-T::AID-tagged nhr-67 allele were generated in preceding work (Medwig-Kinney et al., 20212020). Recent micropublications describe the P2A::H2B::mTurquoise2-tagged lag-2 and mNeonGreen-tagged lin-12 alleles used in this study (Medwig-Kinney et al., 2022Pani et al., 2022). The eGFP-tagged pop-1 allele and POPTOP reporter were previously published (Green et al., 2008van der Horst et al., 2019), as were the AID::BFP and mNeonGreen tagged alleles of unc-37 (Kurashina et al., 2021Ma et al., 2021). GFP-tagged ama-1 (Hills-Muckey et al., 2021) as well as mKate2-tagged hpl-1 and hpl-2 (Patel and Hobert, 2017) were also disseminated in prior publications. The single-copy transgenes expressing the CDK sensor and TIR1 variants under ubiquitously expressed ribosomal promoters (rps-27 and rpl-28, respectively) as well as the tissue-specific GFP-targeting nanobody are described in previous work (Adikes et al., 2020Hills-Muckey et al., 2021Smith et al., 2022Wang et al., 2017) and are located at neutral genomic sites, ttTi4348 or ttTi5605 (Frøkjær-Jensen et al., 2013). The same is true for the heat shock inducible constructs for HLH-2 and NHR-67 (Medwig-Kinney et al., 2020). The cadherin (cdh-3) anchor cell reporter and basement membrane (laminin) markers have already been characterized (Keeley et al., 2020Matus et al., 2010). The following mutant alleles were obtained from the Caenorhabditis Genetics Center: unc-37(e262wd26) and unc-37(wd17wd22) (Pflugrad et al., 1997), the latter of which was maintained using the chromosome I/III balancer hT2 (McKim et al., 1993). The genotypes of all strains used in this study can be found within the Key Resources Table.
Auxin inducible protein degradation
The auxin inducible degron (AID) system was utilized to strongly deplete proteins of interest (Zhang et al., 2015). AID-tagged alleles were paired with the Arabidopsis thaliana F-box protein, transport inhibitor response 1 (AtTIR1), and treated with the water-soluble auxin 1-Naphthaleneacetic acid (K-NAA) at 1 mM concentration (Martinez et al., 2020). Auxin was added to nematode growth media plates according to previously published protocols (Martinez and Matus, 2020), which were then seeded with OP50 E. coli. To achieve robust depletion, synchronized L1 stage animals were directly plated on auxin plates.
The RNAi clones targeting pop-1 and uba-1 as well as the corresponding empty vector control (L4440) were obtained from the Vidal library (Rual et al., 2004). The RNAi constructs targeting the pro-invasive transcription factors (egl-43, fos-1, hlh-2, and nhr-67) and chromatin modifiers (pbrm-1, swsn-4, and swsn-8) are derived from the highly efficient RNAi vector T444T (Sturm et al., 2018) and were generated in preceding work (Medwig-Kinney et al., 2020Smith et al., 2022). To avoid known AC/VU cell fate specification defects caused by hlh-2 perturbations, synchronized animals were grown on OP50 until the L2 stage when they were shifted to hlh-2 RNAi plates.
Live cell imaging
With the exception of the FRAP experiments shown in Figure 3, all micrographs were collected on a Hamamatsu Orca EM-CCD camera mounted on an upright Zeiss AxioImager A2 with a Borealis-modified CSU10 Yokagawa spinning disk scan head (Nobska Imaging) using 405 nm, 440 nm, 488 nm, 514 nm, and 561 nm Vortran lasers in a VersaLase merge and a Plan-Apochromat 100×/1.4 (NA) Oil DIC objective. MetaMorph software (Molecular Devices) was used for microscopy automation. Several experiments were scored using epifluorescence visualized on a Zeiss Axiocam MRM camera, also mounted on an upright Zeiss AxioImager A2 and a Plan-Apochromat 100×/1.4 (NA) Oil DIC objective. For static imaging, animals were mounted into a drop of M9 on a 5% Noble agar pad containing approximately 10 mM sodium azide anesthetic and topped with a coverslip. For long-term time-lapse imaging, animals were first anesthetized in 5 mM levamisole diluted in M9 for approximately 20 minutes, then transferred to a 5% Noble agar pad and topped with a coverslip sealed with VALAP (Kelley et al., 2017). For short-term time-lapse imaging, the pre-anesthetization step was omitted, and animals were transferred directly into a drop of 5 mM levamisole solution on the slide.
Fluorescence recovery after photobleaching
FRAP experiments were performed using an Acal BFi UV Optimicroscan photostimulation device mounted on a spinning disk confocal system consisting of a Nikon Ti2 inverted microscope with Yokogawa CSU-W1 SoRa spinning disk. Data were acquired using a Hamamatsu ORCA Fusion camera, 60x 1.27 NA water immersion objection, SoRa disk, and 2.8x SoRa magnifier. Single plane images were collected every 1 second.
The 276 bp fragment of the nhr-67 promoter (Bodofsky et al., 2018) was cloned into the pMW2 vector, linearized by BamHI digestion. Linearized plasmid was transformed into the Y1H yeast strain (as described in (Reece-Hoyes and Walhout, 2018)). Transformed yeast was plated on SC-HIS plates for three days before being transformed with the HLH-2 Gal4-AD plasmid. Three colonies from each transformation plate were streaked onto SC-HIS-TRP +3-aminotriazole (3-AT) plates. Protein-DNA interactions were determined by visible growth on 3-AT conditions with negative growth in empty vector controls after three days. Plates were imaged on a Fotodyne FOTO/Analyst Investigator/FX darkroom imaging station.
Plasmids containing target proteins fused to GAL-4 DNA-binding-domain + LEU and GAL-4 Activation Domain + TRP were co-transformed into the pJ69-4a Y2H yeast strain as previously described (Reece-Hoyes and Walhout, 2018). Transformed yeast was plated on SC-TRP-LEU plates for three days. Three colonies from each transformation plate were streaked onto SC-HIS-TRP-LEU 3-AT plates. Protein interactions were determined by visible growth on 3-AT conditions with negative growth in empty vector controls after three days. Plates were imaged as described in the previous section.
Quantification of protein expression and cell cycle state
Image quantification was performed in Fiji/ImageJ (Schindelin et al., 2012). Protein expression was quantified by drawing a region of interest around the nucleus of the cell of interest and measuring the mean gray value, then manually subtracting the mean gray value of a background region of similar area to account for camera noise. The CDK sensor was quantified as previously described (Adikes et al., 2020). Following rolling ball subtraction (50 pixels), mean gray value is measured in a region of interest drawn within the cytoplasm and one around the nucleus excluding the nucleolus. The cytoplasmic-to-nuclear ratio correlates to CDK activity and is used to assess cell cycle state (Adikes et al., 2020Spencer et al., 2013). Movies were collected by acquiring z-stacks at 5-minute intervals. Samples were time-aligned relative to anaphase. Cells that did not undergo anaphase during the acquisition period were aligned based on their DHB ratios. Animals that arrested in development (i.e., did not show evidence of progressing through the cell cycle) were excluded from analysis.
For colocalization analyses, single plane images were collected to avoid z drift during acquisition and prevent photobleaching, which was often non-uniform between red and green fluorophores. Micrographs were subject to background subtraction (rolling ball radius = 50) followed by thresholding to segment punctae. Manders’ overlap coefficients (M) were calculated by measuring the extent that segmented punctae of NHR-67 overlapped with that of other proteins using Just Another Colocalization Plugin (JACoP) in Fiji/ImageJ (Bolte and Cordelières, 2006Schindelin et al., 2012). Heterozygous animals for nhr-67::mScarlet and nhr-67::GFP were used as positive controls. These images were then re-analyzed following 90-degree rotation of one of the two channels being compared, resulting in random colocalization that served as a negative control.
Data visualization and statistical analyses
Representative images were processed using Fiji/ImageJ (Schindelin et al., 2012). Heat maps were generated using the Fire lookup table. Tests to determine statistical significance of data were conducted in RStudio and plots were generated using the R package ggplot2 (Wickham, 2016). Error bars represent the mean ± standard deviation. Schematics of gene loci were generated using sequences from WormBase (Harris et al., 2020) and the Exon-Intron Graphic Maker (http://wormweb.org/exonintron). Figures were assembled in Adobe Illustrator.
We are grateful to Dr. Derek Applewhite and Aidan Teran for advice on quantification of protein colocalization. Additionally, we thank Chris Zhao for constructive comments on the manuscript. Some strains were provided by the Caenorhabditis Genetics Center, which is funded by NIH Office of Research Infrastructure Programs (P40 OD010440).
Taylor N. Medwig-Kinney, Conceptualization, Formal Analysis, Investigation, Writing— original draft, Visualization, Supervision, Funding acquisition; Brian A. Kinney, Investigation, Writing—review and editing; Michael A. Q. Martinez, Investigation, Resources, Writing—review and editing; Callista Yee, Resources, Writing—review and editing; Sydney S. Sirota, Formal Analysis, Investigation, Resources; Angelina A. Mullarkey, Formal Analysis, Writing—review and editing; Neha Somineni, Formal Analysis, Resources; Justin Hippler, Formal Analysis, Resources; Wan Zhang, Resources; Kang Shen, Supervision; Christopher M. Hammell, Resources, Supervision, Writing—review and editing; Ariel M. Pani, Investigation, Writing—review and editing; David Q. Matus, Conceptualization, Supervision, Funding acquisition, Writing—review and editing
David Q. Matus and Neha Somineni are paid employees of Arcadia Science and Integra LifeSciences, respectively.
- 1.Visualizing the metazoan proliferation-quiescence decision in vivoElife 9:1–74https://doi.org/10.7554/eLife.63265
- 2.An expanded auxin-inducible degron toolkit for Caenorhabditis elegansGenetics https://doi.org/10.1093/genetics/iyab006
- 3.HLH-2/E2A Expression Links Stochastic and Deterministic Elements of a Cell Fate Decision during C. elegans GonadogenesisCurr Biol 29:1–7https://doi.org/10.1016/j.cub.2019.07.062
- 4.unc-37/Groucho and lsy-22/AES repress Wnt target genes in C. elegans asymmetric cell divisions
- 5.Influences of HLH-2 stability on anchor cell fate specification during Caenorhabditis elegans gonadogenesisG3: Genes|Genomes|Genetics
- 6.A tissue-specific enhancer of the C. elegans nhr-67/tailless gene drives coordinated expression in uterine stem cells and the differentiated anchor cellGene Expr Patterns 30:71–81https://doi.org/10.1016/j.gep.2018.10.003
- 7.Transcription Factors Activate Genes through the Phase-Separation Capacity of Their Activation DomainsCell 175:1842–1855https://doi.org/10.1016/j.cell.2018.10.042
- 8.A guided tour into subcellular colocalization analysis in light microscopyJ Microsc 224:213–232https://doi.org/10.1111/j.1365-2818.2006.01706.x
- 9.A Protein Domain-Based Interactome Network for C. elegans Early EmbryogenesisCell 134:534–545https://doi.org/10.1016/j.cell.2008.07.009
- 10.Germline P Granules Are Liquid Droplets That Localize by Controlled Dissolution/CondensationScience (80-) 324:1729–1732
- 11.The genetics of Caenorabditis elegansGenetics 77:71–94https://doi.org/10.1016/S0047-2484(78)80101-8
- 12.Reprogramming : identifying the mechanisms that safeguard cell identityDevelopment 146:1–17https://doi.org/10.1242/dev.182170
- 13.A POP-1 repressor complex restricts inappropriate cell type-specific gene transcription during Caenorhabditis elegans embryogenesisEMBO J 20:7197–7208https://doi.org/10.1093/emboj/20.24.7197
- 14.Mediator and RNA polymerase II clusters associate in transcription-dependent condensatesScience (80-) 361:412–415https://doi.org/10.1126/science.aar4199
- 15.Imaging dynamic and selective low-complexity domain interactions that control gene transcriptionScience (80-) 361:https://doi.org/10.1126/science.aar2555
- 16.Expression of a single transfected cDNA converts fibroblasts to myoblastsCell 51:987–1000https://doi.org/10.1016/0092-8674(87)90585-X
- 17.CRISPR-based methods for caenorhabditis elegans genome engineeringGenetics 202:885–901https://doi.org/10.1534/genetics.115.182162
- 18.Streamlined genome engineering with a self-excising drug selection cassetteGenetics 200:1035–1049https://doi.org/10.1534/genetics.115.178335
- 19.The Groucho ortholog UNC-37 interacts with the short Groucho-like protein LSY-22 to control developmental decisions in C. elegansDevelopment 137:1799–1805https://doi.org/10.1242/dev.046219
- 20.Improved Mos1-mediated transgenesis in C. elegansNat Meth 9:117–118https://doi.org/10.1038/mp.2011.182.doi
- 21.The myogenic potency of HLH-1 reveals wide-spread developmental plasticity in early C. elegans embryosDevelopment 132:1795–1805https://doi.org/10.1242/dev.01774
- 22.Melting dsDNA donor molecules greatly improves precision genome editing in caenorhabditis elegansGenetics 216:643–650https://doi.org/10.1534/genetics.120.303564
- 23.Activation of Hypodermal Differentiation in the Caenorhabditis elegans Embryo by GATA Transcription Factors ELT-1 and ELT-3Mol Cell Biol 21:2533–2544https://doi.org/10.1128/mcb.21.7.2533-2544.2001
- 24.Tissue-Specific Transcription Footprinting Using RNAGenetics 216:931–945
- 25.Opposing Wnt pathways orient cell polarity during organogenesisCell 134:646–656https://doi.org/10.1016/j.cell.2008.06.026.Opposing
- 26.The lin-12 locus specifies cell fates in Caenorhabditis elegansCell 34:435–444https://doi.org/10.1016/0092-8674(83)90377-X
- 27.MRG-1 / MRG15 Is a Barrier for Germ Cell to Neuron reprogrammingGenetics 211:121–139
- 28.Hallmarks of Cancer: New DimensionsCancer Discov 31–47https://doi.org/10.1158/2159-8290.CD-21-1059
- 29.The hallmarks of cancerCell 100:57–70https://doi.org/10.1007/s00262-010-0968-0
- 30.WormBase: A Modern Model Organism Information ResourceNucleic Acids Res 48:https://doi.org/10.1093/nar/gkz920
- 31.An engineered, orthogonal auxin analog/AtTIR1(F79G) pairing improves both specificity and efficacy of the auxin degradation system in Caenorhabditis elegansGenetics iyab 174:https://doi.org/10.1093/genetics/iyab174
- 32.In vivo switching of human melanoma cells between proliferative and invasive statesCancer Res 68:650–656https://doi.org/10.1158/0008-5472.CAN-07-2491
- 33.. pha-4, an HNF-3 homolog, specifies pharyngeal organ identity in Caenorhabditis elegansGenes Dev 12:1947–1952https://doi.org/10.1101/gad.12.13.1947
- 34.Improved CRISPR / Cas9 knock-in efficiency via the self-excising cassette (SEC) selection method in C. elegansMicropublication Biol
- 35.ixation can change the appearance of phase separation in living cells
- 36.A Caenorhabditis elegans model for epithelial-neuronal transdifferentiationProc Natl Acad Sci 105:3790–3795https://doi.org/10.1073/pnas.0712159105
- 37.Cell cycle and p53 gate the direct conversion of human fibroblasts to dopaminergic neuronsNat Commun 6:https://doi.org/10.1038/ncomms10100
- 38.Control of type-D GABAergic neuron differentiation by C. elegans UNC-30 homeodomain proteinNature 372:780–783https://doi.org/10.1038/372780a0
- 39.Highly accurate protein structure prediction with AlphaFoldNature 596:583–589https://doi.org/10.1038/s41586-021-03819-2
- 40.Members of the NODE (Nanog and Oct4-associated deacetylase) complex and SOX-2 promote the initiation of a natural cellular reprogramming event in vivoProc Natl Acad Sci U S A 109:6596–6601https://doi.org/10.1073/pnas.1117031109
- 41.Post-transcriptional regulation of the E/Daughterless ortholog HLH-2, negative feedback, and birth order bias during the AC/VU decision in C. elegansGenes Dev 17:3100–3111https://doi.org/10.1101/gad.1160803
- 42.Targeted DamID in C. elegans reveals a direct role for LIN-22 and NHR-25 in antagonizing the epidermal stem cell fateSci Adv 8:
- 43.Comprehensive Endogenous Tagging of Basement Membrane Components Reveals Dynamic Movement within the Matrix ScaffoldingDev Cell 54:60–74https://doi.org/10.1016/j.devcel.2020.05.022
- 44.Live-cell confocal microscopy and quantitative 4D image analysis of anchor-cell invasion through the basement membrane in Caenorhabditis elegansNat Protoc 12:2081–2096https://doi.org/10.1038/nprot.2017.093
- 45.PHA-4/FoxA cooperates with TAM-1/TRIM to regulate cell fate restriction in the C. elegans foregutDev Biol 303:611–624https://doi.org/10.1016/j.ydbio.2006.11.042
- 46.The postembryonic cell lineages of the hermaphrodite and male gonads in Caenorhabditis elegansDev Biol 70:396–417https://doi.org/10.1016/0012-1606(79)90035-6
- 47.Divide or Conquer: Cell Cycle Regulation of Invasive BehaviorTrends Cell Biol https://doi.org/10.1016/j.tcb.2016.08.003
- 48.FACT Sets a Barrier for Cell Fate Reprogramming in Caenorhabditis elegans and Human CellsDev Cell 46:611–626https://doi.org/10.1016/j.devcel.2018.07.006
- 49.Sustained expression of unc-4 homeobox gene and unc-37/groucho in postmitotic neurons specifies the spatial organization of the cholinergic synapses in c. ElegansElife 10:1–26https://doi.org/10.7554/eLife.66011
- 50.Liquid droplet formation by HP1α suggests a role for phase separation in heterochromatinNature 547:236–240https://doi.org/10.1038/nature22822
- 51.Transcription factor autoregulation is required for acquisition and maintenance of neuronal identityDevelopment 146:https://doi.org/10.1242/dev.177378
- 52.A Map of the Interactome Network of the Metazoan C. elegansScience (80-) 303:540–543https://doi.org/10.1126/science.1091403
- 53.EVI1 in leukemia and solid tumorsCancers (Basel 12:1–17https://doi.org/10.3390/cancers12092667
- 54.Enhancer-promoter communication: hubs or loops?Curr Opin Genet Dev 67:5–9https://doi.org/10.1016/j.gde.2020.10.001
- 55.A 4D single-cell protein atlas of transcription factors delineates spatiotemporal patterning during embryogenesisNat Methods 18:893–902https://doi.org/10.1038/s41592-021-01216-1
- 56.Dynamics of a developmental switch: Recursive intracellular and intranuclear redistribution of Caenorhabditis elegans POP-1 parallels Wnt-inhibited transcriptional repressionDev Biol 248:128–142https://doi.org/10.1006/dbio.2002.0721
- 57.Auxin-mediated Protein Degradation in Caenorhabditis elegansBio-Protocol 10:1–13https://doi.org/10.21769/bioprotoc.3589
- 58.Rapid Degradation of Caenorhabditis elegans Proteins at Single-Cell Resolution with a Synthetic AuxinG3 Genes, Genomes, Genet 10:267–280https://doi.org/10.1534/g3.119.400781
- 59.In vivo identification of regulators of cell invasion across basement membranesSci Signal 3:https://doi.org/10.1126/scisignal.2000654
- 60.Invasive Cell Fate Requires G1 Cell-Cycle Arrest and Histone Deacetylase-Mediated Changes in Gene ExpressionDev Cell 35:162–174https://doi.org/10.1016/j.devcel.2015.10.002
- 61.Two Types of Sites Required for Meiotic Chromosome Pairing in Caenorhabditis elegansGenetics 134:749–768
- 62.Deletion of a putative HDA-1 binding site in the hlh-2 promoter eliminates expression in C. elegans dorsal uterine cellsMicropublication Biol 1:
- 63.An in vivo toolkit to visualize endogenous LAG-2/Delta and LIN-12/Notch signaling in C. elegansmicroPublication Biol
- 64.A developmental gene regulatory network for C. elegans anchor cell invasionDevelopment 147:https://doi.org/10.1101/691337
- 65.Breaking down barriers: the evolution of cell invasionCurr Opin Genet Dev 47:33–40https://doi.org/10.1016/j.gde.2017.08.003
- 66.The Fos family of transcription factors and their role in tumourigenesisEur J Cancer 41:2449–2461https://doi.org/10.1016/j.ejca.2005.08.008
- 67.A Proliferative to Invasive Switch is Mediated by srGAP1 Downregulation Through the Activation of TGFβ2 SignalingSSRN Electron J 40:111358https://doi.org/10.1016/j.celrep.2022.111358
- 68.TLX, an Orphan Nuclear Receptor with Emerging Roles in Physiology and DiseaseEndocrinol (United States 162:1–13https://doi.org/10.1210/endocr/bqab184
- 69.A new toolkit to visualize and perturb endogenous LIN-12/Notch signalingMicropublication Biol
- 70.Coordinated control of terminal differentiation and restriction of cellular plasticityElife 6:1–26https://doi.org/10.7554/elife.24100
- 71.Removal of Polycomb Repressive Complex 2 Makes C. elegans Germ Cells Susceptible to Direct Conversion into Specific Somatic Cell TypesCell Rep 2:1178–1186https://doi.org/10.1016/j.celrep.2012.09.020
- 72.Precise positions of Phoebe determined with CCD image-overlapping calibrationMon Not R Astron Soc 366:208–212https://doi.org/10.1186/1471-2105-7-208
- 73.The Groucho-like transcription factor UNC-37 functions with the neural specificity gene unc-4 to govern motor neuron identity in C. elegansDevelopment 124:1699–1709https://doi.org/10.1242/dev.124.9.1699
- 74.Reciprocal asymmetry of SYS-1/β-catenin and POP-1/TCF controls asymmetric divisions in Caenorhabditis elegansProc Natl Acad Sci U S A 104:3231–3236https://doi.org/10.1073/pnas.0611507104
- 75.A New Look at TCF and β-Catenin through the Lens of a Divergent C. elegans Wnt PathwayDev Cell 17:27–34https://doi.org/10.1016/j.devcel.2009.07.002
- 76.Basic Caenorhabditis elegans Methods: Synchronization and ObservationJ Vis Exp 1–9https://doi.org/10.3791/4019
- 77.A gel phase promotes condensation of liquid P granules in Caenorhabditis elegans embryosNat Struct Mol Biol 26:
- 78.The Caenorhabditis elegans gene lin-26 can trigger epithelial differentiation without conferring tissue specificityDev Biol 235:410–421https://doi.org/10.1006/dbio.2001.0294
- 79.Restriction of Cellular Plasticity of Differentiated Cells Mediated by Chromatin Modifiers, Transcription Factors and Protein Kinases. G3: Genes|Genomes|Genetics 9:2287–2302https://doi.org/10.1534/g3.119.400328
- 80.A compendium of Caenorhabditis elegans regulatory transcription factors: A resource for mapping transcription regulatory networksGenome Biol 6:https://doi.org/10.1186/gb-2005-6-13-r110
- 81.High-efficiency yeast transformationCold Spring Harb Protoc 2018:563–568https://doi.org/10.1101/pdb.prot094995
- 82.Direct in vivo cellular reprogramming involves transition through discrete, non-pluripotent stepsJ Cell Sci 124:https://doi.org/10.1242/jcs.090043
- 83.Transorganogenesis and transdifferentiation in C. elegans are dependent on differentiated cell identityDev Biol 420:136–147https://doi.org/10.1016/j.ydbio.2016.09.020
- 84.Transdifferentiation and remodeling of post-embryonic C. elegans cells by a single transcription factorDevelopment 140:4844–4849https://doi.org/10.1242/dev.103010
- 85.A natural transdifferentiation event involving mitosis is empowered by integrating signaling inputs with conserved plasticity factorsCell Rep 40:111365https://doi.org/10.1016/j.celrep.2022.111365
- 86.Developmental Plasticity and Cellular ReprogrammingGenetics 213:723–757
- 87.Toward improving Caenorhabditis elegans phenome mapping with an ORFeome-based RNAi libraryGenome Res 14:2162–2168https://doi.org/10.1101/gr.2505604
- 88.Coactivator condensation at super-enhancers links phase separation and gene controlScience (80-) 361:https://doi.org/10.1126/science.aar3958
- 89.Influences of LIN-12/Notch and POP-1/TCF on the Robustness of Ventral Uterine Cell Fate Specification in Caenorhabditis elegans GonadogenesisG3: Genes|Genomes|Genetics 5:2775–2782https://doi.org/10.1534/g3.115.022608
- 90.Dimerization-driven degradation of C. elegans and human E proteinsGenes Dev 29:1356–1361https://doi.org/10.1101/gad.261917.115
- 91.A bHLH Code for Sexually Dimorphic Form and Function of the C. elegans Somatic GonadCurr Biol 27:1853–1860https://doi.org/10.1016/j.cub.2017.05.059
- 92.Fiji: an open-source platform for biological-image analysisNat Methods 9:676–82https://doi.org/10.1038/nmeth.2019
- 93.Cell autonomy of lin-12 function in a cell fate decision in C. elegansCell 57:1237–1245https://doi.org/10.1016/0092-8674(89)90060-3
- 94.Cell-cell interactions prevent a potential inductive interaction between soma and germline in C. elegansCell 61:939–951https://doi.org/10.1016/0092-8674(90)90060-R
- 95.Anchor cell invasion into the vulval epithelium in C. elegansDev Cell 5:21–31https://doi.org/10.1016/S1534-5807(03)00168-0
- 96.C. elegans TCF protein, POP-1, converts from repressor to activator as a result of Wnt-induced lowering of nuclear levelsDev Biol 285:584–592https://doi.org/10.1016/j.ydbio.2005.07.008
- 97.POP-1 controls axis formation during early gonadogenesis in C. elegansDevelopment 453:443–453
- 98.The sys-1 and sys-3 Genes Cooperate with Wnt Signaling to Establish the Proximal-Distal Axis of the Caenorhabditis elegans GonadGenetics 166:171–186https://doi.org/10.1534/genetics.166.1.171
- 99.Empirically controlled mapping of the Caenorhabditis elegans protein-protein interactome networkNat Methods 6:47–54https://doi.org/10.1038/nmeth.1279
- 100.The SWI/SNF chromatin remodeling assemblies BAF and PBAF differentially regulate cell cycle exit and cellular invasion in vivoPLOS Genetics https://doi.org/10.1371/journal.pgen.1009981
- 101.The proliferation-quiescence decision is controlled by a bifurcation in CDK2 activity at mitotic exitCell 155:369–383https://doi.org/10.1016/j.cell.2013.08.062
- 102.Phase separation drives heterochromatin domain formationNature 547:241–245https://doi.org/10.1038/nature22989
- 103.Highly efficient RNAi and Cas9-based auto-cloning systems for C. elegans researchNucleic Acids Res 46:https://doi.org/10.1093/nar/gky516
- 104.Post-embryonic cell lineages of the nematode, Caenorhabditis elegansDev Biol 56:110–156https://doi.org/10.1016/0012-1606(77)90158-0
- 105.Properties of repression condensates in living Ciona embryosNat Commun 12:1–9https://doi.org/10.1038/s41467-021-21606-5
- 106.Direct Conversion of C. elegans Germ Cells into Specific Neuron TypesScience (80-) 331:304–309
- 107.C. elegans Runx / CBFβ suppresses POP-1 (TCF) to convert asymmetric to proliferative division of stem cell-like seam cellsDevelopment 146:1–34https://doi.org/10.1242/dev.180034
- 108.AlphaFold Protein Structure Database: massively expanding the structural coverage of protein-sequence space with high-accuracy modelsNucleic Acids Res 50:https://doi.org/10.1093/nar/gkab1061
- 109.The tailless ortholog nhr-67 functions in the development of the C. elegans ventral uterusDev Biol 356:516–528https://doi.org/10.1016/j.ydbio.2011.06.007
- 110.E Proteins and ID Proteins: Helix-Loop-Helix Partners in Development and DiseaseDev Cell 35:269–280https://doi.org/10.1016/j.devcel.2015.10.019
- 111.A toolkit for GFP-mediated tissue-specific protein degradation in C. elegansDevelopment 144:2694–2701https://doi.org/10.1242/dev.150094
- 112.ggplot2: Elegant Graphics for Data Analysis
- 113.Reciprocal changes in expression of the receptor lin-12 and its ligand lag-2 prior to commitment in a C. elegans cell fate decisionCell 79:1187–1198https://doi.org/10.1016/0092-8674(94)90010-8
- 114.Quantitative Differences in Nuclear β-catenin and TCF Pattern Embryonic Cells in C. elegansPLoS Genet 11:https://doi.org/10.1371/journal.pgen.1005585
- 115.The auxin-inducible degradation (AID) system enables versatile conditional protein depletion in C. elegansDevelopment 142:4374–4384https://doi.org/10.1242/dev.129635
- 116.Genome-Wide Prediction of C. elegans Genetic InteractionsScience (80-) 311:1481–1484
- 117.Reprogramming of early embryonic blastomeres into endodermal progenitors by a Caenorhabditis elegans GATA factorGenes Dev 12:3809–3814https://doi.org/10.1101/gad.12.24.3809
- 118.Sequential histone-modifying activities determine the robustness of transdifferentiationScience (80-) 345:826–829https://doi.org/10.1126/science.1255885