Abstract
Cytohesins are Arf guanine nucleotide exchange factors (GEFs) that regulate membrane trafficking and actin cytoskeletal dynamics. We report here that GRP-1, the sole Caenorhabditis elegans cytohesin, controls the asymmetric divisions of certain neuroblasts that divide to produce a larger neuronal precursor or neuron and a smaller cell fated to die. In the Q neuroblast lineage, loss of GRP-1 led to the production of daughter cells that are more similar in size and to the transformation of the normally apoptotic daughter into its sister, resulting in the production of extra neurons. Genetic interactions suggest that GRP-1 functions with the previously described Arf GAP CNT-2 and two other Arf GEFs, EFA-6 and BRIS-1, to regulate the activity of Arf GTPases. In agreement with this model, we show that GRP-1’s GEF activity, mediated by its SEC7 domain, is necessary for the posterior Q cell (Q.p) neuroblast division and that both GRP-1 and CNT-2 function in the Q.posterior Q daughter cell (Q.p) to promote its asymmetry. Although functional GFP-tagged GRP-1 proteins localized to the nucleus, the extra cell defects were rescued by targeting the Arf GEF activity of GRP-1 to the plasma membrane, suggesting that GRP-1 acts at the plasma membrane. The detection of endogenous GRP-1 protein at cytokinesis remnants, or midbodies, is consistent with GRP-1 functioning at the plasma membrane and perhaps at the cytokinetic furrow to promote the asymmetry of the divisions that require its function.
Keywords: cytohesin, arf, asymmetric cell division, cell fate, apoptosis
METAZOAN development relies on the coordinated behavior of individual cells that either adopt morphogenetic behaviors such as proliferation, migration, adhesion, and differentiation, or die at precise times and places. Despite extensive characterization of the molecular mechanisms associated with the execution of programmed cell death (PCD) during development, less is known about how a given cell chooses to live or die. PCD execution often relies on the activation of versatile proteases, the caspases, by an evolutionarily conserved molecular cascade. In the nematode Caenorhabditis elegans, somatic PCD is largely determined by asymmetric cell divisions that produce a surviving daughter cell and a daughter cell fated to die (Frank et al. 2005; Cordes et al. 2006; Hatzold and Conradt 2008; Ou et al. 2010; Singhvi et al. 2011). The invariant lineage that produces these dying cells makes C. elegans a powerful system to explore the mechanisms involved in PCD specification.
Although several studies point to the cell-specific transcriptional control of EGL-1, a BH3-only protein that can activate the caspase cascade, as a mechanism of PCD specification (Potts and Cameron 2011), other data suggest that daughter cell-size asymmetry regulates PCD (Frank et al. 2005; Cordes et al. 2006; Hatzold and Conradt 2008; Ou et al. 2010; Singhvi et al. 2011). Indeed, C. elegans divisions that generate dying cells are generally asymmetric, producing a larger surviving daughter and a smaller daughter fated to die. Several mutants affecting this size difference also perturb PCD specification, leading to the survival of both daughter cells.
The ADP-ribosylation factor (Arf) GTPase-activating protein (GAP) CNT-2 and two C. elegans Arf GTPases that function with CNT-2 were previously shown to control cell size and cell death in asymmetric neuroblast divisions by an unknown mechanism (Singhvi et al. 2011). Arfs are small GTPases that regulate secretory and endocytic pathways, as well as the actin cytoskeleton (Donaldson and Jackson 2011). Arfs fall into three classes based on sequence homology: class I (Arf1-3), class II (Arf4-5), and the more divergent class III (Arf6) (Kahn et al. 2006). Class I and II Arfs localize to Golgi and endosomal compartments and are required for protein trafficking in the secretory and endocytic pathways. Arf6, by contrast, localizes to the plasma membrane and to endosomes and has been shown to regulate events near the cell surface, including endocytosis, exocytosis, and cortical actin structure (Donaldson and Jackson 2011). Arfs exist in active (GTP bound) and inactive (GDP bound) states that are controlled by accessory proteins. Guanine nucleotide exchange factors (GEFs) facilitate GDP release and GTP binding, and GAPs like CNT-2 stimulate hydrolysis of GTP to GDP. Arf-GTP can recruit coatomer proteins and initiate the formation of membrane vesicles. The cycling between GDP- and GTP-bound states is necessary for Arfs to regulate vesicle budding (Kreis et al. 1995).
In this report, we describe the involvement of General Receptor for Phosphoinositides-1 (GRP-1), an Arf GEF of the cytohesin family, in asymmetric neuroblast divisions and PCD specification. Cytohesins contain an N-terminal coiled-coil (CC) domain, a central SEC7 domain that contains ARF GEF activity, and a C-terminal pleckstrin-homology (PH) domain (reviewed in Jackson et al. 2000; Moss and Vaughan 2002). Cytohesins have been implicated in regulating signal transduction, actin cytoskeletal dynamics, protein trafficking in the exocytic and endocytic pathways, and cell adhesion (Jackson et al. 2000; Moss and Vaughan 2002; Kolanus 2007).
Since most of the previous studies of cytohesins focused exclusively on assays conducted in cultured cell lines, the functions of these molecules during animal development are still poorly understood. Here we report that GRP-1 possibly regulates multiple C. elegans Arfs together with the previously described Arf GAP CNT-2. We show that both GRP-1 and CNT-2 act autonomously in dividing neuroblasts that produce a dying daughter. In the absence of GRP-1 function, the apoptotic daughters of these neuroblasts are transformed into their sisters, resulting in the production of extra neurons. Human cytohesins can functionally substitute for GRP-1 to regulate neuroblast divisions. We also provide evidence that other Arf GEFs, acting in parallel to GRP-1, function in these divisions. Surprisingly, we find GFP-tagged GRP-1 localized to the nucleus due to the presence of a nuclear localization signal in the GRP-1 CC domain. Targeting the Arf GEF domain to different cellular compartments, however, suggests that it acts at the plasma membrane. Supporting this model, we found that endogenous GRP-1 is associated with midbodies, cytokinesis remnants of dividing cells.
Materials and Methods
C. elegans genetics
General handling and culture of nematodes were performed as previously described (Brenner 1974). The N2 Bristol line was used as wild type, and experiments were performed at 20° unless otherwise noted. The following mutations and integrated arrays were used:
LG I. ynIs45 [Pflp-15::gfp] (Kim and Li 2004), zdIs5 [mec-4::gfp] (Clark and Chiu 2003).
LG II. rrf-3(pk1426) (Simmer et al. 2002), MosSCI transgenes integrated at the ttTi5605 site (see Table 1) (Frøkjær-Jensen et al. 2008), gmIs20 [hlh-14::gfp] (Frank et al. 2003), ynIs25 [Pflp-12::gfp] (Li et al. 1999; Kim and Li 2004).
LG III. ced-4(n1182) (Ellis and Horvitz 1986), arf-1.2(ok796) (Singhvi et al. 2011), grp-1(gm350, gk402275) (this study), grp-1(tm1956) (Denning et al. 2012), lon-1(e185) (Brenner 1974), unc-32(e189) (Brenner 1974), cnt-2(gm377) (Singhvi et al. 2011), unc-119(ed3) (Maduro and Pilgrim 1995), gmIs12 [Psrb-6::gfp] (Hawkins et al. 2005), rdvIs1 [Pegl-17::mCherry:his-24 + Pegl-17::myristoylated mCherry + pRF4] (Ou et al. 2010).
LG IV. eri-1(mg366) (Kennedy et al. 2004), pig-1(gm301) (Cordes et al. 2006), bris-1(gk592726) (this study), arf-3(tm1877) (this study), efa-6(tm3124) (O’Rourke et al. 2010), dpy-20(e1282ts) (Brenner 1974), ced-3(n717, n2436) (Ellis and Horvitz 1986; Shaham et al. 1999), unc-30(e191) (Brenner 1974), arf-6(tm1447) (Singhvi et al. 2011), mgIs71 [tph-1::gfp] (Bülow and Hobert 2004).
LG V. egl-1(n1084n3082) (Conradt and Horvitz 1998), ayIs9 [Pegl-17::gfp] (Branda and Stern 2000), gmIs65 [Psra-6::mCherry, Ptph-1::gfp] (a kind gift from Richard Ikegami).
LG X. gmIs81 [Pmec-4::mCherry, Pflp-12::EBFP2, Pgcy-32::gfp, Pegl-17::gfp] (Gurling et al. 2014).
Table 1. A/PVM phenotypes of RNAi-treated animals and mutants.
Gene | % extra A/PVMsa | n | |
---|---|---|---|
Control/L4440 | 4 | 457 | |
grp-1/ K06H7.4b | 72c | 463 | |
cnt-2/Y39A1A.15d | 49c | 218 | |
arfs and arf-related genes (RNAi) | arf-1.2/ B0336.2b | 41c | 51 |
arf-1.2/ B0336.2e | 49c | 76 | |
arf-3/ F57H12.1b | 28f | 80 | |
arf-6/ Y116A8C.12e | 30c | 98 | |
arf-1.1/ F45E4.1e | 10f | 171 | |
arl-1/ F54C9.10e | 8g | 135 | |
evl-20/arl-2/ F22B5.1b | 5h | 130 | |
arl-3/ 19FH8.3b | 6h | 142 | |
arc-1/arl-4/ ZK1320.6e | 10f | 122 | |
arl-5/ ZK632.8b | 6h | 130 | |
arl-6/ C38D4.8d | 7h | 150 | |
arl-7/ F20D1.5 (pseudogene)b | 6h | 132 | |
arl-8/ Y57G11C.13e | 5h | 131 | |
arl-13/ Y37E3.5d | 5c | 159 | |
Y54E10BR.2e | 4h | 131 | |
sar-1/ ZK180.4b | 10g | 70 | |
arfs (mutants) | |||
arf-3(tm1877, m+) | 0 | 107 | |
ced-4(n1162) | 12 | 240 | |
ced-4(n1162); arf-3(tm1877, m+) | 1 | 168 | |
arls (mutants) | ced-3(n717) | 12 | 445 |
ced-4(n1162) | 12 | 240 | |
evl-20(ar103); ced-3(n717) | 10i | 209 | |
arl-5(ok2407); ced-3(n717) | 11i | 180 | |
ced-4(n1162); arl-8(wy271) | 15i | 218 |
Extra A/PVMs in the RNAi experiments were scored in a zdIs5; rrf-3; ced-3(n2436) strain.
RNAi clones were from Kamath et al. (2001).
Significantly different from L4440 control P < 0.001.
RNAi clones were constructed for this study.
RNAi clones were from Rual et al. (2004).
Significantly different from L4440 control P < 0.02.
Significantly different from L4440 control P < 0.05.
Not significantly different from L4440 control.
Not significantly different from ced-3 or ced-4 control.
Detection of specific neurons
All neurons were detected with chromosomally integrated transcriptional reporters that express GFP under control of the indicated C. elegans promoter. The ALM, AVM, PVM, and PLM neurons were detected using the reporter zdIs5 [mec-4::gfp]. The SDQ neurons were detected using the reporter ynIs25 [flp-12::gfp]. The PHB neurons were visualized in L1 larvae using gmIs12 [Psrb-6::gfp], which is expressed both in PHB and the adjacent PHA neurons. The PHA and I2 neurons were detected using the reporter ynIs45 [flp-15::gfp]. Reducing grp-1 activity does not alter the number of PHA neurons in ced-3(+) or ced-3(n717) backgrounds (data not shown), indicating that the extra neurons observed using gmIs12 are PHBs. The HSN neurons were detected using mgIs71 [tph-1::gfp] (Sze et al. 2002; Bülow and Hobert 2004), and by immunostaining adult hermaphrodites with rabbit antiserotonin antibodies as previously described (Garriga et al. 1993). The A/PQR neurons were detected using the gmIs81 [Pgcy-32::gfp] reporter. The NSM neurons were scored using the gmIs65 [Ptph-1::gfp] reporter.
Isolation and cloning of grp-1(gm350)
We mutagenized zdIs5 [mec-4::gfp]; ced-3(n2436) hermaphrodites with 50 mM ethylmethylsulfonate and screened F2’s for mutations that strongly enhanced the penetrance of extra AVMs, PVMs, and PLMs above the level observed in ced-3(n2436) (data not shown). grp-1(gm350); ced-3(n2436) had a high penetrance of extra AVMs, PVMs, PLMs, and HSNs.
To map gm350, we first separated the mutation from ced-3(n2436) and then used single-nucleotide polymorphism (SNP) mapping to position gm350 between snp_F27B3[1] and pkP3049 on LG III (Wicks et al. 2001), a 180-kb region containing ∼45 open reading frames (ORFs).
To determine which of the 45 ORFs was likely to be mutated in gm350 mutants, we systematically inactivated all but two of the ORFs by RNA-mediated interference (RNAi) in zdIs5; ced-3(n2436) hermaphrodites (Timmons and Fire 1998; Kamath et al. 2001). The grp-1 ORF, K06H7.4, was the only gene in this region that phenocopied the original zdIs5; gm350; ced-3(n2436) mutant when inactivated by RNAi (Supporting Information, Figure S1). To determine whether the gm350 lesion was located in grp-1, we sequenced the ORF in gm350 mutants and identified a single nonsense mutation (Figure 1A).
Figure 1.
The gene grp-1 regulates apoptosis in multiple neuroblast lineages. (A) Schematic diagram of the GRP-1 domain structure and the grp-1 genomic structure and mutations. grp-1(gm350) is a nonsense mutation at the beginning of the SEC7(Arf GEF) domain (W76TaGStop). The grp-1(tm1956) mutation is a deletion that removes part of the SEC7 and PH domains and shifts the reading frame. The grp-1(gk402275) mutation is a 3′ splice site mutation that changes the last nucleotide G of the last intron to A, which is predicted to disrupt RNA splicing and result in an altered pleckstrin homology (PH) domain. (B) The wild-type Q.p neuroblast lineage produces two neuron types (A/PVMs and SDQs) and a dying cell (Q.pp). (C) The Q.p divisions of pig-1, cnt-2, and grp-1 mutants produce daughters that are more equivalent in size, and the posterior daughter cell can behave like its mitotic sister cell, surviving, dividing, and producing extra neurons with A/PVM and SDQ fates. (E and G) The wild-type PLM/ALN (E) and HSN/PHB (F) neuroblast lineages each produce two neuron types and dying cells. (D and F) The grp-1 mutations interact synergistically with ced-3 and egl-1 mutations to produce extra A/PVM and SDQ neurons in the Q.p lineage (D) and extra PLM neurons (F). (H) grp-1(gm350) and ced-3(n717) mutations interact synergistically to produce extra HSN and PHB neurons. In all lineage diagrams, the anterior cell is to the left, and apoptotic cells are indicated by an “X.” All neurons were scored using cell-specific GFP reporters as described in Materials and Methods. The number of lineages scored per genotype is shown under each bar. ND, not determined. *P < 0.0001 in comparison to either single mutant.
Protein sequence analysis
Sequence alignments and dendrograms were generated using ClustalW2 (https://www.ebi.ac.uk/Tools/msa/clustalw2/). The GenBank accession numbers for sequences used in phylogenetic analyses are: C. elegans GRP-1, NP_498764; Drosophila melanogaster steppke, NP_610120; Homo sapiens CYTH1, NP_004753.1, CYTH2, NP_059431.1, CYTH3, NP_004218.1, CYTH4, NP_037517.1; and Mus musculus Cyth1, NP_035310.2, Cyth2, NP_035311.1, Cyth3, NP_035312.2, Cyth4, NP_082471.2.
RNAi
RNAi was performed by feeding worms individual bacterial clones as described (Timmons and Fire 1998). Clones were from the libraries constructed in the Ahringer lab (Timmons and Fire 1998; Kamath et al. 2001) or the Vidal lab (Rual et al. 2004), or were generated for this study as indicated. The negative control used in these experiments was a clone containing empty vector (L4440) in the bacterial host HT115 (Timmons and Fire 1998), the same host used for the RNAi feeding experiments. A clone was constructed to target the SEC7 protein-encoding gene Y106G6G.2 (Supporting Information) and introduced into HT115 bacteria, the same approach used to generate the clones from the Arhinger and Vidal labs’ libraries (Kamath et al. 2001; Rual et al. 2004).
Analysis of neuroblast daughters size and the number of cells produced by Q.p
For Q.p and the anterior Q (Q.a) daughter cell-size measurements, anterior and posterior daughter cell areas were measured in triplicate using ImageJ, on the left side of L1 larvae carrying the Q lineage markers rdvIs1 [Pegl-17::mCherry:his-24 + Pegl-17::myristoylated mCherry] (Q.p daughters) or gmIs81 [Pegl-17::gfp] (Q.a daughters). The size ratio was calculated using the average area values from the three measurements. We only measured the relative sizes of the daughter cells when the posterior daughter cell of Q.p (Q.pp) or the anterior daughter cell of Q.a (Q.aa) did not appear apoptotic: they were not rounded and were still attached to their respective sister.
The persistence of the rdvIs1 mCherry signal in Q.p descendants allowed us to determine the final number of surviving cells in the Q.p lineage progeny before expression was lost. We only scored cells that were not rounded and were located at the position where Q.p had been. The rdvIs1 marker also labels Q.a and its daughters, which could make the identification of Q.p progeny ambiguous. Since the Q.aa cell dies before the Q.pp cell, we only analyzed lineages where we could identify the Q.aa dying cell and its surviving sister unambiguously by their positions, and thus, exclude them from the scoring.
Plasmid construction and transgenics
The detail of DNA manipulations is presented in Supporting Information. Transgenic lines were generated by injecting plasmid DNA into the syncytial hermaphrodite gonad. MosSCI lines were generated as described by injecting the EG6699 ttTi5605, unc-119(ed3) mutants (Frøkjær-Jensen et al. 2014). All transgenes used in this study are described in Table S1.
Analysis of grp-1 cell autonomy
For the mosaic analysis, zdIs5; grp-1(gm350) hermaphrodites were injected with plasmid bGG836 [grp-1(+)] and the cell-autonomous marker Pdpy-30::NLS::DsRed2. Animals carrying the resulting extrachromosomal arrays (gmEx353-5; Table S1) were rescued for the Grp-1 A/PVM phenotype (data not shown). Extrachromosomal arrays are mitotically unstable, and thus animals carrying these arrays are mosaic, frequently having lost the array from one or more lineages. We screened zdIs5; grp-1; gmEx353 mosaic animals for those with extra A/PVMs and then determined which lineages had lost the array by scoring nuclear DsRed2 expression.
For grp-1 rescue using the mab-5 promoter assay, zdIs5 [mec-4::gfp]; grp-1(gm350) mutants were injected with Pmab-5::grp-1::gfp and the cell-autonomous marker Pdpy-30::NLS::DsRed2. The Pdpy-30::NLS::DsRed2 plasmid drives ubiquitous expression of nuclear-localized DsRed2 and allowed us to unambiguously determine which cells in mosaic animals contained the extrachromosomal array. As a negative control in all experiments, animals carrying the extrachromosomal array were compared to their siblings on the same plate that had lost the array. Rescue of Grp-1 phenotypes was assessed by determining whether grp-1 mutants that did not express Pmab-5::grp-1::gfp had a significantly higher percentage of sides with an extra cell compared to transgenic animals where the array was present in the A/PVM neurons.
The arf-6 rescue experiment was performed with the integrated MosSCI transgene gmSi33 (Table S1) introduced by crosses into the ced-4; arf-6 double mutant.
For the unc-86 promoter assay, integrated MosSCI trangenes [Punc-86::grp-1::gfp] and [Punc-86::cnt-2b::gfp] (Table S1) were crossed into zdIs5 [mec-4::gfp]; grp-1(gm350) or zdIs5 [mec-4::gfp]; cnt-2(gm377) mutants, respectively.
Statistical analysis
Statistical analysis was performed using the two-sample Z-test for proportions for extra cell scoring and the Mann–Whitney U-test for cell size ratios.
Immunofluorescence
Adults and larvae were removed from starved plates by washing, leaving embryos that were collected, fixed, and permeabilized as described (Guenther and Garriga 1996). Rabbit anti-GRP-1 antibodies (Covance, raised against a 29-amino-acid peptide corresponding to GRP-1 C-terminal sequence DEDMRSWINAISRMMAPQQHLLARPKSTH) were diluted 1:1000 in PBST-A (1× PBS, 1% BSA, 0.5% Triton X-100, 0.05% NaN3, 1 mM EDTA), and the dilution was incubated overnight at room temperature with fixed embryos. To detect NMY-2::GFP and ZEN-4::GFP fusions, a chicken anti-GFP antibody diluted 1/500 was added (Molecular Probes). After washing, the embryos were rocked 4 hr at room temperature in Alexa 568 goat anti-rabbit antisera and, when needed, Alexa 488 anti-chicken antisera diluted 1:100 in PBST-A (Molecular Probes). After washing three times and DAPI counterstaining, embryos were mounted in Vectashield Mounting Medium (Vector Laboratories). Pictures were obtained from a Hamamatsu ORCA-ER digital camera attached to a Zeiss Axioskop2 microscope using Openlab software.
Confocal microscopy
For detection of GRP-1 fusion proteins, live animals were mounted in 3 μl of a solution containing 33 μM sodium azide in M9 minimal medium on an agarose pad. Confocal microscopy of live and fixed immunostained animals was performed on a Leica TCS NT confocal microscope. Projections of confocal images were generated with ImageJ software (http://imagej.nih.gov/ij/) using the standard deviation method. Images were prepared for publication using Adobe Photoshop.
Time-lapse images of the Q.p division were acquired as described (Singhvi et al. 2011). rdvIs1 and Pmab-5::gfp::efa-6 images were acquired with the 561- and 488-nm rays of a Marianas spinning disc confocal microscope (Intelligent Imaging Innovations) using a ×100 1.4-NA objective and the SlideBook software.
Results
Isolation of the grp-1(gm350) mutant
A subset of C. elegans neurons are generated by neuroblast asymmetric divisions that produce a daughter cell that dies and a neuron or neuronal precursor. The kinase PIG-1 and the Arf GAP CNT-2 have been shown to regulate the cell death decision in many of these divisions (Cordes et al. 2006; Ou et al. 2010; Singhvi et al. 2011; Hirose and Horvitz 2013). To identify additional genes that might function with pig-1 or cnt-2, we screened for mutants with extra neurons that expressed the Pmec-4::gfp transgene zdIs5 and are located in the positions normally occupied by the AVM and PVM neurons. The posterior daughter of the Q neuroblast, Q.p, divides to produce a posterior cell that dies and an anterior precursor that divides to produce the A/PVM and SDQ neurons, resulting in an AVM and SDQ on the right side and a PVM and SDQ on the left (Figure 1B). Loss of PIG-1 or CNT-2 often results in a posterior Q.p daughter cell that survives and adopts the fate of its sister, resulting in the production of extra neurons (Figure 1C) (Cordes et al. 2006; Singhvi et al. 2011). Extra neurons that express zdIs5 usually have axons that project ventrally, similar to the AVM and PVM axons. When there is more than one neuron on a side that expresses zdIs5, we refer to the additional neurons as extra A/PVMs. Since mutations in the caspase gene ced-3 strongly enhance the phenotypes of pig-1 mutants, we used ced-3(n2436), a partial loss-of-function mutation, to sensitize the background of our screen. We isolated the gm350 mutant on the basis of this extra A/PVM defect and found that it also displayed extra PLMs and HSNs, similar to pig-1 and cnt-2 mutants (Figure 1, B–H and Figure S1A). We mapped the gm350 lesion to the gene grp-1. A second allele of grp-1, tm1956, was obtained from the National Biosource Project (http://www.shigen.nig.ac.jp/c.elegans/index.jsp) and it caused a similar extra A/PVM phenotype (Figure 1, A and D). A third, hypomorphic allele, grp-1(gk402275), was identified by the Million Mutation Project (http://genome.sfu.ca/mmp/) and it caused milder defects (Figure 1, A and D).
Several observations suggest that the phenotypes observed in these mutants were caused by a loss of GRP-1 activity. First, meiotic mapping placed gm350 close to the grp-1 locus, and DNA sequencing confirmed the presence of a nonsense mutation in the grp-1 ORF (Materials and Methods and Figure 1A). Second, RNAi of the grp-1 locus generated phenotypes similar to those produced by the gm350 and tm1956 alleles (Figure 1, D, F, and H and Figure S1B). Third, the two strong alleles are predicted to impair GRP-1 function (see below). Fourth, complementation tests showed tm1956 and gk402275 failed to complement gm350 (data not shown). Finally, expressing a grp-1 cDNA from various C. elegans promoters rescued the grp-1(gm350) mutant phenotypes (see below).
GRP-1 is a member of the cytohesin family of ARF GEFs:
GRP-1 is the sole C. elegans member of the cytohesin family of Arf GEFs. Like vertebrate cytohesins, nematode GRP-1 contains an N-terminal CC domain, a central SEC7 domain that contains Arf GEF activity, and a C-terminal PH domain (Figure 1A). C. elegans GRP-1 and all four human cytohesins are 38.5% (CYTH4) to 41.7% (CYTH3) identical over the full length of the protein, and >60% similar with a high level of conservation in all three domains. The grp-1(gm350) mutation isolated in our genetic screen is a nonsense mutation (W76ochre) located at the beginning of the SEC7 domain. The grp-1(tm1956) mutation, a 983-bp deletion and single-bp insertion, removes the second half of the SEC7 domain and the first part of the PH domain and is predicted to introduce a shift in the reading frame. The grp-1(gk402275) mutation is a 3′ splice site mutation in the last intron of grp-1 (Figure 1A). Because the proteins encoded by both gm350 and tm1956 mutants are predicted to lack intact SEC7 and PH domains, and because expression of the CC domain alone cannot rescue the mutant phenotypes of grp-1 mutants (see below), we infer that gm350 and tm1956 severely reduce or eliminate grp-1 function.
grp-1 functions in parallel to the programmed cell death pathway in multiple neuroblast lineages:
The grp-1(gm350); ced-3(n2436) mutant isolated from our sensitized screen had extra AVM, PVM, PLM, and HSN neurons (Figure S1A). To characterize the grp-1 phenotype in the absence of other mutations, we scored grp-1 mutants in a wild-type ced-3 background. These mutants had extra neurons derived from the Q.p neuroblast, but had few extra neurons derived from the HSN/PHB lineages (Figure 1, D and H) in comparison to the original grp-1; ced-3(n2436) mutant strain isolated from our screen. An examination of animals with mutations in grp-1 and strong mutations in the proapoptotic genes ced-3 or egl-1 supports the hypothesis that grp-1 and ced-3 mutations interact synergistically and demonstrate that grp-1 also regulates the number of PHB neurons (Figure 1, D, F, and H). Finally, we observed that the grp-1(gm350) mutant had weak extra neuron defects in other neuroblast lineages such as the Q.a lineage (1.8% extra A/PQR neurons, n = 168), the I2 lineage (3% extra I2 neurons, n = 100), and the NSM lineage (3.6% extra NSM neurons, n = 302). Together, these observations indicate that grp-1, like pig-1, functions partly in parallel to the PCD execution pathway to regulate the asymmetric divisions of certain neuroblasts. All subsequent experiments that involve a grp-1 mutant were carried out with the gm350 allele.
GFP::GRP-1 fusions localized to nuclei and rescued the grp-1 mutant:
To characterize the GRP-1 expression pattern and subcellular localization, we introduced a translational GFP::GRP-1 cDNA fusion expressed under the control of the grp-1 promoter into grp-1 mutant worms (Figure 2). We observed that GFP::GRP-1 transgene expression was widespread in somatic cells of embryos, larvae, and adults, and was concentrated in the nuclei of interphase cells (Figure 2, A–C and data not shown). Since vertebrate cytohesins localize to the cytoplasm and the plasma membrane, the GFP::GRP-1 enrichment in nuclei was unexpected. Two independently generated transgenes generated the same expression pattern, were able to rescue the extra A/PVM phenotype caused by the grp-1 mutation (Figure 2D), and did not induce any extra cell phenotype in wild-type grp-1 background (data not shown).
Figure 2.
Functional GFP::GRP-1 localizes to nuclei. (A) Diagram of the GFP::GRP-1 construct. Expression of this GFP fusion was driven off the grp-1 promoter. (B and C) Confocal projections of zdIs5 [mec-4::gfp]; grp-1(gm350); Pgrp-1::gfp::grp-1 in an embryo (B) and a L1 larva (C). Arrows mark the ALM and PLM processes labeled by the zdIs5 transgene. Bar, 10 μm (B) and 20 μm (C). (D) Expression of Pgrp-1::gfp::grp-1 from two independently isolated extrachromosomal arrays rescued the A/PVM phenotype of grp-1 mutants. The number of lineages scored per genotype is shown under each bar. *P < 0.0001.
Human cytohesins can functionally substitute for nematode GRP-1:
To test whether human cytohesins can provide GRP-1 function, we expressed GFP fusions to cDNAs of three human family members [cytohesin-1 (CYTH1), cytohesin-2 (CYTH2), and cytohesin-4 (CYTH4)] in C. elegans, examined the localization of these fusions in embryos, and determined whether they could rescue the grp-1 mutant A/PVM phenotype (Figure S2). All of the cytohesin transgenes were expressed from the dpy-30 promoter, which drives widespread expression (Hsu et al. 1995). We found that the CYTH1 and CYTH2 fusions localized to the plasma membrane and that CYTH1 also localized weakly to nuclei, but we did not observe any specific localization for the CYTH4 fusion (Figure S2, C–E′). All three fusions rescued the extra-A/PVM phenotype of grp-1 mutants (Figure S2F), supporting the hypothesis that GRP-1 is a cytohesin ortholog. The ability of both membrane-localized CYTH2 and nuclear GRP-1 to rescue the grp-1 mutant phenotype raises the issue of where GRP-1 normally functions in the cell.
GRP-1 may function in an Arf GTPase cycle with CNT-2:
The CNT-2 GAP domain was previously shown to be required for its function in asymmetric neuroblast divisions (Singhvi et al. 2011). As both CNT-2 and GRP-1 are predicted to regulate Arf GTPases, they may be components of the same Arf cycle. In agreement with this model, the grp-1cnt-2 double mutant had the same penetrance of extra-A/PVM neurons as the cnt-2 single mutant, suggesting that grp-1 and cnt-2 act in the same genetic pathway (Figure 3). Since the completion of the Arf GTP-bound to GDP-bound cycling is thought to be essential for vesicular budding (Kreis et al. 1995), we expected to observe similar defects when the cycle is compromised by disrupting the Arf GAP. The lower penetrance of the grp-1 mutants, however, suggests that CNT-2 has other functions or acts with Arf GEFs other than GRP-1 (Figure 3).
Figure 3.
A/PVM phenotypes of grp-1 and Arf pathway mutants. The grp-1 extra-A/PVM phenotype was not enhanced by an arf-6 mutation, but was enhanced by mutations in arf-1.2 or the Arf GEF genes, efa-6 and bris-1. The cnt-2 Arf GAP mutant phenotype was not enhanced by grp-1 or arf-6 mutations. The grp-1 cnt-2 double mutant also contained a lon-1 mutation used in strain construction; control experiments showed no effect of lon-1 on the penetrance of A/PVM phenotype of either grp-1 or cnt-2 mutants (data not shown). bris-1(gk592726) is a 5′ splice site hypomorphic mutant in the third intron of the predicted long a isoform, in the first intron of predicted b and c isoforms, and not affecting the predicted d isoform (TTGgtga > TTGatga). The number of lineages scored per genotype is shown under each bar. NS, not significant; *P < 0.0001.
Additional Arf GEF proteins regulate the Q.p division:
To test whether other Arf GEF proteins may compensate for grp-1 loss, we performed an RNAi grp-1 enhancer screen using RNAi clones targeting other SEC7 domain-containing proteins of C. elegans. We found that knocking down the function of two Arf GEFs, EFA-6 and M02B7.5, which we name BRIS-1 for its homologs BRAG/IQSEC/Schizo, were able to enhance the grp-1 A/PVM phenotype, but did not induce a phenotype in animals that were wild type for grp-1 (Figure S3). Using efa-6 and bris-1 mutations, we confirmed this enhancer effect in the grp-1 mutant (Figure 3). Taken together, our genetic data are consistent with a model where GRP-1 and these other Arf GEFs function with CNT-2 to regulate the Q.p asymmetric division. Our findings do not distinguish between models where the Arf GEFs act together with CNT-2 to regulate a single membrane trafficking event or where they function with CNT-2 in separate events.
ARF-6 acts with GRP-1 to regulate the Q.p division:
To determine which Arfs are regulated by GRP-1, we analyzed the arf-1.2, arf-3, and arf-6 genes, which encode the sole C. elegans orthologs of classes I, II, and III Arfs, respectively (Li et al. 2004). Two Arf GTPases, arf-1.2 and arf-6, have been previously implicated in the regulation of the Q.p neuroblast asymmetric division (Singhvi et al. 2011). Whereas single Arf mutants have subtle phenotypes and only generate extra cells in a sensitized ced mutant background (Singhvi et al. 2011), the arf-1.2; arf-6 double mutant displayed a synthetic A/PVM phenotype, suggesting that the two Arfs can compensate for each other’s loss (Figure 3). However, the penetrance of the defects in the double Arf mutant was significantly lower than the defects observed in cnt-2 (P < 0.0001) and grp-1 (P < 0.005) mutants, suggesting that additional Arfs or Arf-like (Arl) GTPases regulate the Q.p division (Figure 3).
We used RNAi to test the role of each of the C. elegans Arf and Arl genes in the Q.p division by scoring the number of A/PVMs in a rrf-3; ced-3(n2436) mutant background (Table 1). The rrf-3 mutation sensitizes the animals to the effects of RNAi (Simmer et al. 2002), and the ced-3(n2436) mutation sensitizes the Q lineage to perturbations that alter the Q.p division (see above). RNAi of all three arf genes, arf-1.2, arf-3, and arf-6, and mutations of arf-1.2 and arf-6 but not arf-3, produced extra-A/PVM phenotypes (Table 1). In contrast to the arf genes, RNAi or mutation of arf-related genes did not produce significant extra-A/PVM phenotypes (Table 1). Although arf-3 is the best candidate for being the missing GTPase, its involvement in neuroblast asymmetric divisions that produce apoptotic cells remains uncertain (Singhvi et al. 2011). The RNAi phenotype of arf-3 could result from off-target effects on arf-1.2, or the lack of a phenotype in the arf-3 mutant could result from maternal rescue, since the mutation leads to larval arrest of homozygous mutants from heterozygous mothers.
We previously showed that cnt-2 single and cnt-2; arf-6 double mutants have a similar A/PVM phenotype, suggesting that cnt-2 and arf-6 function in the same genetic pathway (Singhvi et al. 2011). We also observed a suppression of the extra-A/PVM phenotype of a cnt-2 mutant by an arf-1.2 mutation, although this effect was explained by a failure of extra cells to express the A/PVM marker in double mutants rather than a true suppression of the extra cell production. Although the daughter cell size-symmetry defect of a cnt-2 mutant was partially suppressed by the arf-1.2 mutation, the number of extra cells and hence the apoptosis defect was not significantly altered. These results are consistent with cnt-2 functioning together with arf-1.2 and arf-6 in the same cycle or cycles to regulate the Q.pp apoptotic fate and arf-1.2 providing a separate function that regulates the differentiation of the extra cells (Singhvi et al. 2011).
To address whether arf-1.2 and arf-6 function in the same genetic pathway as grp-1, we examined the phenotypes of arf, grp-1 double mutants and found that the arf-1.2 but not the arf-6 mutation enhanced the extra-A/PVM phenotype of the grp-1 mutant (Figure 3). These data suggest either that GRP-1 could act in the same pathway as ARF-6 but in parallel to ARF-1.2 or that all three proteins could function together, with additional Arf GEFs able to compensate for grp-1 loss and to activate ARF-1.2.
Loss of GRP-1 affects the sizes of the Q.p daughters:
In wild-type animals, Q.p divides asymmetrically in size to generate a larger anterior cell that survives and a smaller posterior cell that dies (Figure 1B and Figure 5A). In pig-1, arf-1.2, and cnt-2 mutants, a reduced cell-size asymmetry of the Q.p daughters was associated with the failure to specify PCD in the Q.pp cell (Cordes et al. 2006; Singhvi et al. 2011; Chien et al. 2012). The larger Q.pp in these mutants was often able to escape the death fate and adopt the mitotic fate of its sister, the anterior daughter cell of Q.p (Q.pa), to produce additional A/PVM and SDQ neurons. We observed that the Q.p division is more symmetric in the grp-1 and arf-6 mutants as well (Figure 4, B and C). Whereas arf-6 did not enhance the cell-size defects of grp-1 or cnt-2 mutants, arf-1.2 and bris-1 were grp-1 enhancers, consistent with the genetic interactions described for the A/PVM phenotype. Ou et al. (2010) showed that the Q.p daughter cell-size defect of pig-1 mutants was not restricted to Q.p, but that the Q.a division was also more symmetric in the mutants. We observed similar Q.a division defects for grp-1 and cnt-2 (Figure S4). These results indicate that like pig-1, grp-1 and cnt-2 are general regulators of cell size asymmetries in the Q lineage.
Figure 5.
GRP-1 functions autonomously. (A and C) Schematic diagrams of the Q neuroblast lineage indicating the specific cells where the promoters are expressed. (A) The mab-5 promoter is active (shaded cells) in the left Q lineage. (B) Expressing grp-1 or arf-6 from the mab-5 promoter rescued the QL but not the QR lineage defects of the grp-1(gm350) and ced-4; arf-6 mutants, respectively. (C) The unc-86 promoter is active (shaded cell) in the left and right Q.p cells before they divide. (D) Expressing grp-1 or cnt-2 from the unc-86 promoter rescued the extra-A/PVM defects of the grp-1(gm350) and cnt-2(gm377) mutants, respectively. For each rescue experiment in D, two independent transgenic lines are shown. The number of lineages scored per genotype and side is shown under each bar. NS, not significant; *P < 0.0001 for rescue of the A/PVM defects.
Figure 4.
Analysis of QL.p cell division and progeny in grp-1 and arf mutants. (A) Schematic representation of the wild-type QL.p neuroblast development (dashed outline represents an L1 larva). QL.p divides asymmetrically (1) and gives rise to a larger Q.pa and a smaller Q.pp, which is fated to die. After the division of the surviving Q.pa (2), the two cells produced differentiate as the PVM (Q.paa) and the SDQL (Q.pap). (B) Fluorescence photomicrographs from time-lapse recordings of wild-type and grp-1(gm350) QL.p divisions. Animals contained the Pegl-17::gfp transgene ayIs9. Arrowheads indicate the position of the cleavage furrow. Numbers at the bottom of each frame represent the time in seconds. The Q.p cell exhibits a range of sizes, and the one from the grp-1 mutant shown here is at the larger end of the size range. Although this cell is larger than the wild-type Q.p cell, the size range of wild-type and mutant Q.p cells was similar. Bar, 2 μm. (C) Cell-size ratios of the QL.p daughters. The cell sizes of the Q.p daughters were determined by measuring the area of cells expressing the rdvIs1[Pegl-17::myristoylated mCherry] transgene. Each dot corresponds to the size ratio of one pair of Q.p daughters, and the horizontal bars indicate the median of the ratio distribution. The dashed horizontal line indicates a 1:1 ratio, where the Q.p daughter cells are of equivalent size. (D) Total number of surviving Q.p descendants. The number of daughter pairs (C) and lineages (D) scored per genotype is shown under each bar. NS, not significant; *P < 0.0001.
The size asymmetry of the Q.p daughter cells correlates with extra cell production:
We also scored the number of cells produced by the Q.p lineage using the Pegl-17::mCherry transgene (Figure 4D). This transgene expresses plasma membrane-bound and nuclear mCherry in the lineage and the mCherry markers persist after the A/PVM and SDQ neurons are born. To determine the number of cells produced by the Q.p lineage, we counted the number of mCherry-expressing cells in the positions of the Q.p descendants. In the grp-1 mutant, 57% of Q.p lineages produced one or two extra neurons (Figure 4D), which is higher than the frequency of lineages that produce extra-A/PVM neurons. This discrepancy may be due to the inability of some of the extra cells to choose a fate, as has been suggested for the “undead” cells produced in cell death mutants (Ellis and Horvitz 1986; Guenther and Garriga 1996; Cordes et al. 2006). The genetic interactions seen when scoring the sizes of daughter cells and the extra-Q.p-derived neurons were similar: the arf-6 mutation did not significantly alter the frequency of extra cells produced by grp-1 or cnt-2 mutants, the arf-1.2arf-6 double mutant had an increased numbers of extra cells compared to the single mutants, and the arf-1.2 mutation enhanced the grp-1 phenotype (Figure 4D).
Since these different mutant combinations displayed a range of cell size and extra-cell phenotypes, we tested whether there was a relationship between the two datasets. We indeed observed that decreased size asymmetry of the Q.p daughters and the production of extra-Q.p progeny are strongly correlated (Figure S5).
GRP-1 and ARF-6 function autonomously in the Q lineage
To determine the site of grp-1 function, we first analyzed animals that were functionally mosaic for grp-1. Since the A/PVM phenotype is incompletely penetrant, we screened for mosaic animals with extra AVMs or PVMs and determined which cells had lost the rescuing array. All of the mutant animals had lost the array in the lineage that produced the extra AVM or PVM (Figure S6). The most informative loss was in AB.plapapa, four divisions before the production of the Q.p. We conclude from the mosaic experiment that grp-1 functions in the progeny of AB.plapapa.
To ask whether grp-1 acts in the Q lineage, we generated transgenes driving grp-1::gfp expression from the mab-5 promoter. The gene mab-5 encodes a homeobox transcription factor that is expressed in many posterior cells on both sides of the animal (Cowing and Kenyon 1992). The one exception to this bilateral expression in young L1 larvae is the Q lineage: mab-5 is expressed in the left Q.p lineage, which produces PVM, but not in the right, which produces AVM (Figure 5A) (Salser and Kenyon 1992). We reasoned that if grp-1 functions autonomously in the Q lineage, then the Pmab-5::grp-1::gfp transgene would rescue the extra-PVM phenotype, but not the extra-AVM phenotype. If grp-1 functions nonautonomously then the transgene might rescue both the AVM and PVM phenotypes or neither phenotype. We used this approach to show that pig-1, cnt-2, and arf-1.2 function cell autonomously in the Q lineage (Cordes et al. 2006; Singhvi et al. 2011). In two of the three transgenic lines that we tested, we found that the Pmab-5::grp-1::gfp construct rescued the PVM but not the AVM phenotype (Figure 5B and data not shown). We performed a similar experiment to examine arf-6 autonomy. Although arf-6 mutants do not produce extra A/PVMs, they can display an extra-A/PVM phenotype in sensitized genetic backgrounds where the caspase activation cascade is impaired, such as the ced-4/APAF mutant (Singhvi et al. 2011). We observed that a Pmab-5::arf-6::GFP integrated transgene suppressed the PVM, but not the AVM defects of ced-4; arf-6 double mutants (Figure 5B). These results suggest that, like cnt-2, arf-1.2 and pig-1, grp-1 and arf-6 act autonomously in the Q lineage.
GRP-1 and CNT-2 function autonomously in Q.p:
Although our results indicate that grp-1 and cnt-2 function in the Q lineage to regulate Q.p asymmetric division, they are compatible with a function in the Q neuroblast, the Q.p cell itself, its sister Q.a, or any combination of these cells. To pinpoint where grp-1 and cnt-2 function, we drove the expression of functional GRP-1::GFP and CNT-2::GFP fusions in Q.p using the unc-86 promoter. The gene unc-86 encodes a POU-domain transcription factor required for cell-fate specification in many C. elegans neuroblast lineages (Finney et al. 1988). In the Q neuroblast lineage, the unc-86 promoter is active in Q.p before its division (Figure 5C) (Baumeister et al. 1996). We observed that expressing GRP-1::GFP and CNT-2::GFP in Q.p, rescued, respectively, the grp-1 and cnt-2 A/PVM mutant phenotypes (Figure 5D).
The ability of the unc-86 promoter-driven expression of grp-1 and cnt-2 to rescue the A/PVM phenotypes of the mutants suggests that GRP-1 and CNT-2 act in Q.p to regulate the asymmetry of its division. Both proteins are normally expressed in Q.p and its daughter cells (data not shown). This expression raises the question of whether these proteins act solely in Q.p to regulate the asymmetry of the division, which contributes to Q.pp apoptosis, or whether they act both in Q.p to regulate the asymmetry of the division and in Q.pp to promote apoptosis. We currently do not know whether GRP-1 and CNT-2 function in Q.pp to promote its apoptotic fate.
Structure/function analysis of GRP-1
We used deletion and mutational analysis to better characterize the three GRP-1 domains. All of these experiments were conducted by expressing a grp-1 cDNA or cDNA fragments fused to GFP from the dpy-30 promoter, which drives widespread expression, in zdIs5; grp-1(gm350) mutants. This broad expression allowed us to observe GFP localization in embryonic cells and to address whether the constructs rescued the A/PVM phenotype in larvae. In these experiments, GFP was fused to the C terminus of GRP-1, and like GFP::GRP-1 (Figure 2), GRP-1::GFP localized to nuclei and rescued the grp-1 A/PVM phenotype (Figure 6, A, B, B′, and G).
Figure 6.
GRP-1 structure–function analysis. (A) Diagrams of the GRP-1–GFP proteins used. All constructs were expressed from the dpy-30 promoter, which drives broad expression. (B–F) Confocal projections and (B′–F′) individual slices from the same confocals series. zdIs5 [mec-4::gfp]; grp-1(gm350) embryos expressing (B and B′) GRP-1::GFP, (C and C′) the GRP-1K270A PH domain mutant, (D and D′) the GRP-1 coiled-coil (CC) domain, (E and E′) the GRP-1E155K GEF-dead mutant, and (F and F′) the SEC7 domain. Bar, 10 μm. (G) Ability of the different transgenes to rescue the A/PVM phenotype of grp-1(gm350) mutants. The number of lineages scored per genotype is shown under each bar. NS, not significant; *P < 0.001 in comparison to nontrangenic grp-1 mutants.
In light of the importance of the CC and PH domains in regulating the activity of vertebrate cytohesins, we performed experiments to determine whether these domains were necessary or sufficient for GRP-1 activity. First, we mutated a critical lysine residue in the GRP-1 PH domain that vertebrate cytohesins require to bind PI(4,5)P2 and PI(3,4,5)P3 (Klarlund et al. 2000; Cronin et al. 2004). The GRP-1K270A mutant did not affect GRP-1 rescuing activity nor did it affect the localization of GRP-1 to nuclei (Figure 6, A, C, C′, and G). Second, we determined whether the GRP-1 PH domain could function as a dominant negative, as has been described previously for vertebrate cytohesins (Kolanus et al. 1996; Várnai et al. 2005). We targeted the GRP-1 PH domain to either the nucleus (NLS-PH) or cell cortex (MYR-PH), but did not observe an extra-A/PVM phenotype in wild-type animals or ced-3(n717) mutants sensitized to the loss of GRP-1 activity (data not shown). Third, although predominantly localized in nuclei like the full-length GRP-1, a deletion mutant containing only the GRP-1 CC domain neither rescued the grp-1 mutant (Figure 6, A, D, D′, and G) nor produced extra neurons in a wild-type grp-1 background (data not shown). Finally, the hypomorphic grp-1(gk402275) mutant, which is predicted to produce a GRP-1 protein with a PH domain truncation, displayed only a weak extra-A/PVM defect (Figure 1D).
Experiments conducted in cultured cell lines indicate that the Arf GEF activity of the SEC7 domain is required for cytohesin function. The Arf GEF activity of vertebrate cytohesins is disrupted by mutation of a conserved glutamate codon to lysine in the SEC7 domain (Cherfils et al. 1998; Mossessova et al. 1998). We found that GRP-1 with the analogous E155K mutation failed to rescue the extra-A/PVM phenotype of the grp-1(gm350) mutant but localized similarly to wild-type GRP-1 (GRP-1E155K; Figure 6, A, E, E′, and G). Expressing a mutant containing only the GRP-1 SEC7 domain rescued the extra-A/PVM phenotype (Figure 6, A and G), and this fusion localized broadly throughout both the cytoplasm and nucleus (SEC7; Figure 6, F and F′). Taken together, these findings suggest that the CC and PH domains are not essential and that Arf GEF activity is necessary and sufficient for GRP-1-dependent regulation of asymmetric neuroblast divisions, at least under conditions where the fusions were expressed from multicopy transgenes.
Nuclear localization signals in GRP-1:
GFP::GRP-1 localization to the nucleus (Figure 2) compelled us to ask which GRP-1 domains were necessary for GRP-1 function and nuclear localization. To identify regions necessary for GRP-1 subcellular localization and function, we examined deletion mutants lacking certain GRP-1 domains. Subcellular localization of two fusions proteins suggested that the GRP-1 CC domain contained a nuclear-localization sequence (NLS): a fusion containing only the CC domain was strongly enriched in nuclei, while a fusion containing only the SEC7 domain was not (Figure 6, D, D′, F, and F′). These data led us to identify a putative NLS (PKVRKRK) in the CC domain using the PPSORT II algorithm (Cherfils et al. 1998; Mossessova et al. 1998; Nakai and Horton 1999). To determine whether this sequence functions as a NLS, we mutated it to PKVAARK in a full-length GRP-1 fusion construct (Figure 7A, GRP-1mutNLS), designing these mutations so that they would not interfere with the repeated structure of the coiled-coil domain. We found that the resulting GRP-1mutNLS fusion was evenly distributed between the nucleus and cytoplasm, suggesting that this sequence contributes to the nuclear localization of GRP-1 (Figure 7, B and B′). The continued presence of GRP-1mutNLS in nuclei may be due to an additional cryptic NLS in the SEC7 domain, as the SEC7 fusion and a similar construct with an additional mRFP moiety (mRFP-SEC7) were still present in nuclei (Figure 7, F and F′).
Figure 7.
Subcellular localization and function of GRP-1::GFP proteins. (A) Diagram of mutant and modified GRP-1–GFP proteins used. All constructs were expressed from the dpy-30 promoter, which drives broad expression. (B–H) Confocal projections, and (B′–H′) individual slices from the related confocal series. zdIs5 [mec-4::gfp]; grp-1(gm350) embryos expressing (B and B′) GRP-1 carrying a mutation in the NLS found in the CC domain, (C and C′) GRP-1 containing an NES, (D and D′) myristoylated GRP-1, (E and E′) myristoylated GRP-1 containing a mutated NLS, (F and F′) an mRFP::SEC7:: GFP protein containing an NLS, (G and G′) an mRFP::SEC7:: GFP protein containing a PLCδ1 PH domain, and (H and H′) an mRFP::SEC7:: GFP protein. The signal from the mRFP::SEC7::GFP trangenes, made with constructs injected at low concentration, was fainter than the other transgenes, and contrast adjustment was used to reveal subcellular localization of the fusion proteins. Bar, 10 μm. (I) Ability of the different transgenes to rescue the A/PVM phenotype of grp-1(gm350) mutants. The number of lineages scored per genotype is shown under each bar. NS, not significant; *P < 0.02, **P < 0.0005, ***P < 0.0001 in comparison to nontransgenic grp-1 mutants.
GRP-1 may function at the plasma membrane
The novel nuclear localization of GRP-1 led us to ask whether GRP-1 functioned in nuclei to regulate asymmetric neuroblast divisions. The first series of experiments we performed was designed to prevent GRP-1’s nuclear localization. We mutated the GRP-1 NLS (GRP-1mutNLS), added a nuclear export signal (NES-GRP-1), targeted the protein to the plasma membrane using a myristoylation signal (MYR-GRP-1), and made a fusion with both a mutant NLS and a myristoylation signal (MYR-GRP-1mutNLS) (Figure 7, A–E′ and I). None of these modifications eliminated GRP-1 entirely from nuclei, and all of these constructs rescued the grp-1(gm350) mutant phenotype, providing little insight as to whether GRP-1 functions in nuclei. When placed in a wild-type background, the GRP-1mutNLS and the NES-GRP-1 did not perturb the asymmetric divisions of either the Q.p or the Q.a neuroblasts, indicating that delocalizing GRP-1 in interphase does not generate defects in the Q lineage (data not shown).
Because the SEC7 domain alone was partially active in the rescue assays (Figure 6G), we next asked whether targeting the SEC7 domain to different subcellular compartments would affect the ability of the SEC7 domain to rescue the mutant A/PVM phenotypes. We created three constructs, one targeted to the nucleus (NLS-mRFP-SEC7), one to the cell cortex (PHPLCdelta1-mRFP-SEC7), and one untargeted control (mRFP-SEC7) (Figure 7A). To prevent passive diffusion into the nucleus, all three constructs contained both mRFP and GFP (GFP is not indicated in construct name) to increase the size of the fusion protein, and we reduced the concentration of the injected DNA to minimize effects caused by excess expression (Table S1). As expected, the NLS-mRFP-SEC7 fusion was found concentrated in nuclei, and the PHPLCdelta1-mRFP-SEC7 to the plasma membrane (Figure 7, F–G′). The untargeted mRFP-SEC7 fusion was present uniformly throughout the cytoplasm and nucleus, even though it should be too large to passively diffuse into nuclei (Figure 7, H and H′). This latter result suggested that an additional cryptic NLS might exist in the SEC7 domain. When we examined these constructs in our phenotypic rescue assay, we found that the NLS-mRFP-SEC7 fusion did not rescue grp-1, that the mRFP-SEC7 fusion rescued partially, and that the PHPLCdelta1-mRFP-SEC7 fusion rescued completely (Figure 7I). There is a discrepancy between the partial rescue by the mRFP-SEC7 fusion and the complete rescue observed with the SEC7 fusion (Figure 5G). This could be explained by the addition of the mRFP at the N terminus; however, this possibility seems unlikely since this tag did not perturb the activity of the rescuing PHPLCdelta1-mRFP-SEC7 fusion. The other possibility is that the SEC7 domain cannot efficiently rescue the grp-1 phenotype at low expression levels, suggesting auxiliary roles played by the CC and PH domain in the full-length GRP-1 protein. These observations strongly suggest that GRP-1 functions at the cell periphery to regulate asymmetric neuroblast divisions.
Targeting Arf GEF activity to the cell periphery is sufficient to restore grp-1 function
Although the experiments described above suggest that the GRP-1’s Arf GEF activity is required at the cell surface, it is possible that, due to the cryptic NLS sequence in the SEC7 domain, the PHPLCdelta1-mRFP-SEC7 fusion is present in the nucleus at undetectable levels. We thus tried to restore Arf GEF activity in Q.p using an alternative GEF, EFA-6, which exclusively localized to the plasma membrane. The EFA-6 protein contains a SEC7 domain and a PH domain like GRP-1, but its structure differs in the N- and C-terminal domains. EFA-6 has been shown to regulate cortical microtubules dynamics in the C. elegans embryo, an activity dependent on a conserved short N-terminal motif (microtubule regulating region, MTR) but independent of the SEC7 domain (O’Rourke et al. 2010).
GFP fusions to a functional EFA-6, a GEF dead mutant (EFA-6E447K), and a MTR-deleted mutant (EFA-6ΔMTR) were expressed in the QL lineage using the mab-5 promoter (Figure 8A). We confirmed that EFA-6 localized at the cell periphery in the QL lineage and other cells where the mab-5 promoter is active (Figure 8, B–D′). The analysis of the A/PVM phenotype of grp-1 mutants carrying these transgenes revealed that full-length EFA-6 and EFA-6ΔMTR rescued the grp-1 mutant, but the SEC7 EFA-6E447K did not (Figure 8E). These results support the hypothesis that GRP-1 GEF activity is required at the plasma membrane and not in the nucleus.
Figure 8.
The Arf GEF EFA-6 can replace GRP-1’s requirement in a SEC7-dependent manner. (A) Diagrams of the EFA-6–GFP proteins used to rescue the grp-1 A/PVM phenotype. The mab-5 promoter drove expression in the left Q lineage. MTR, microtubule-regulating motif; PH, pleckstrin-homology domain. (B–D′) Confocal micrographs showing the localization of GFP::EFA-6 in grp-1 rdvIs1 larvae. (B–D) Labeled plasma membranes and nuclei of cells expressing the rdvIs1 mCherry markers. (B′–D′) Labeled plasma membranes (arrows) of cells expressing the Pmab-5::gfp::efa-6 transgene. The GFP::EFA-6 cells indicated by arrows are QL and V5L (B′), QL.a and QL.p (C′), and posterior P lineage-derived cells (D′). Bar, 10 μm. (E) Ability of different Pmab-5::gfp::efa-6 transgenes to rescue the extra-PVM phenotype of grp-1(gm350) mutants. The number of lineages scored per genotype and per side is shown under each bar. NS, not significant; *P < 0.0001.
Endogenous GRP-1 protein localizes to midbodies
We observed that GRP-1–GFP fusions accumulate in interphase nuclei, but our rescue experiments indicate that the ARF GEF functions at the plasma membrane. We reasoned that GRP-1 may accumulate at the membrane at the time of its cytoplasmic release by the nuclear membrane breakdown during mitosis (Figure 6, B and B′). However, the cytoplasmic GFP signal was too bright to allow us to detect a specific accumulation of GRP-1 at the surface, and lowering the expression of GRP-1::GFP fusions by the use of rescuing low-copy number transgenes (Figure 5D) rendered the GFP undetectable in Q.p cells (data not shown). To examine the endogenous localization of GRP-1 during mitosis, we raised antibodies against a GRP-1 peptide and immunolabeled C. elegans embryos.
We were able to detect punctate, ring-like signals between cells in wild-type embryos (Figure 9, A and B). These punctae were not detected in the grp-1 mutants, suggesting that the immunolabeling was specific for the GRP-1 protein (Figure 9C). The punctae appeared similar to the cytokinetic furrow remnants, or midbodies. The GRP-1 punctate signal colocalized with both NMY-2/nonmuscle myosin II heavy chain and ZEN-4/MKLP/Pavarotti, two markers of the midbody (Figure 9, D and E). Localization of endogenous GRP-1 at midbodies could have two explanations. First, GRP-1 could be recruited to the midbody after cytokinesis (Figure 9F). Second, GRP-1 could be recruited to the cleavage furrow before or during its ingression and accumulate at the midbody upon completion of cytokinesis (Figure 9G). In this latter model, we would only be able to detect GRP-1 with our antibodies when the protein is concentrated at midbodies. We were unable to detect endogenous GRP-1 in cell nuclei by immunolabeling. Together, these data support a model where the GRP-1 Arf GEF functions at the cell periphery in mitotic cells.
Figure 9.
Immunolabeling of endogenous GRP-1 protein at midbodies. DAPI (blue) was used to counterstain nuclei in all merge photomicrographs. (A and B) Immunofluorescence staining of GRP-1 (red) in an early (A) and a late (B) stage wild-type (N2) embryo. GRP-1 localized to ring-like structures between cells (arrows). (C) Absence of GRP-1 staining in a grp-1(gm350) mutant embryo. (D and E) Immunofluorescence staining of wild-type embryos that expressed the myosin II fusion NMY-2::GFP (D, green, left), the ZEN-4::GFP fusion (E, green, left), and endogenous GRP-1 (D and E, red, center). The merged panels (right) show colocalization of GRP-1 at midbodies marked with NMY-2::GFP and ZEN-4::GFP (arrows). Bars, 10 μm. (F and G) Models for the timing of GRP-1 recruitment to the cell surface. (F) GRP-1 is recruited to the midbody at the end of cytokinesis. (G) GRP-1 is recruited to the furrow during cytokinesis and accumulates at the midbody.
Discussion
We show that GRP-1, the C. elegans homolog of vertebrate cytohesin Arf GEFs, regulates the asymmetry of certain neuroblast divisions and their ability to generate a daughter cell that adopts the apoptotic fate. To our knowledge, this is the first report of a role for a cytohesin in the asymmetry of cell divisions. We demonstrate that the ARF GEF activity of GRP-1 at the plasma membrane is both necessary and sufficient for GRP-1 to regulate asymmetric neuroblast divisions and identify ARF-6 and the Arf GAP CNT-2 as potential components of the same Arf cycle that regulates these divisions.
Daughter cell-size asymmetry and cell death specification
By analyzing the sizes of Q.p daughter cells in various mutants, we observed that an increase in the cell size of the posterior daughter correlated with the loss of its apoptotic fate. Although there is no direct evidence of a causal relationship between a small cell size and cell death in C. elegans, two additional studies support a model where increased cell size interferes with the ability of a cell to adopt the apoptotic fate. First, the asymmetric division of the NSM precursor to generate the NSM neuron and its sister cell, which is fated to die, is more symmetric in dnj-11 mutants (Hatzold and Conradt 2008). Following the NSM precursor division in dnj-11 mutants, the authors observed a range of altered daughter cell sizes and noted a correlation between a larger daughter cell size and the inability of the cell to adopt an apoptotic fate. Second, Ou et al. (2010) showed that the Q.p neuroblast sister, Q.a, produces a smaller anterior cell by a mechanism dependent on anterior myosin II accumulation. CALI experiments that inactivated the asymmetrically distributed NMY-2::GFP in Q.a also resulted in a more symmetric division and differentiation of Q.aa, the cell normally fated to die.
Recent studies in HeLa cells dividing on patterned substrates showed that a failure to divide symmetrically was occasionally associated with the death of the smaller cell (Kiyomitsu and Cheeseman 2013). One possibility is that the activation of the caspase pathway and a reduction in cell size both can contribute to the induction of cell death. How the cell size might tip the balance toward cell death instead of survival remains to be determined.
C. elegans Arf GTPases can compensate for each other’s inactivation
We observed that C. elegans can survive in the absence of both class I and III Arfs: arf-1.2; arf-6 double mutants are sick but viable. The C. elegans class II Arf, ARF-3, is sufficient for viability in the absence of class I and III Arfs. This observation indicates that ARF-3 or the Arls can perform all Arf-dependent essential functions. This view is supported by studies showing that specific members of all three Arf classes can support recruitment of coatomer to Golgi membranes in vitro and that knocking down the functions of more than one Arf is necessary to disrupt specific membrane trafficking events (Liang and Kornfeld 1997; Volpicelli-Daley et al. 2005). A more recent study showed that the in vitro production of COPI vesicles can be promoted by several Arfs including Arf3, but that in the presence of other Arfs, Arf3 was excluded from the vesicles, highlighting the promiscuity of Arf function when certain Arfs are lost (Popoff et al. 2011). In summary, our studies of the arf mutants suggest that in vivo the Arfs can provide overlapping functions in the regulation of essential cellular activities such as secretion and endocytosis.
Arf pathways regulate asymmetric neuroblast divisions:
Our data indicate that ARF-6, GRP-1, and CNT-2 could act in the same Arf cycle to regulate Q.p asymmetric division. Furthermore, other Arfs, such as ARF-1.2 and ARF-3, and GEFs, such as EFA-6 and BRIS-1, may act in the same Arf cycle or define one or more additional CNT-2-dependent cycles. How could class I and III Arfs provide overlapping functions for the Q.p division? ARF-1.2 and ARF-6 might function in the same process. For example, all three Arf classes have been shown to activate PLD and PI5K, suggesting that GRP-1 could regulate asymmetric neuroblast divisions by controlling the production of a lipid intracellular signal (Godi et al. 1999; Honda et al. 1999; Jones et al. 2000). Alternatively, ARF-1.2 and ARF-6 could have different outputs; for example, they could regulate separate trafficking steps. One possibility is that GRP-1 regulates protein trafficking in the endocytic pathway, as there is evidence that Arf1 and Arf6 can regulate different aspects of trafficking to or from endosomes (Randazzo et al. 2000; Donaldson 2003; Nie et al. 2003).
Cytohesins can act as exchange factors for both Arf1 and Arf6 in vitro (Casanova 2007). Arf6-GTP and to a lesser extent Arf1-GTP can also activate cytohesins, relieving an autoinhibition mediated by interactions between the Sec7 and PH domains and converting them from their inactive cytoplasmic form to their active membrane-bound form (Cohen et al. 2007; DiNitto et al. 2007; Hofmann et al. 2007; Li et al. 2007). The findings that Arfs can function as cytohesin activators and effectors led Stalder et al. (2011) to propose that Arfs and the cytohesin ARNO act in a positive feedback loop in vitro. In C. elegans, genetic interactions are consistent with ARF-6 acting as a GRP-1 activator, effector, or both.
GRP-1 functions at the cell surface despite a predominantly nuclear localization
The localization of GFP-tagged GRP-1 to nuclei was unexpected. The ability of the PHPLCdelta1-tagged SEC7 domain but not the NLS-tagged SEC7 domain to rescue the extra-A/PVM phenotype indicates that membrane-associated and not nuclear GRP-1 functions to regulate asymmetric neuroblast divisions. The rescue of grp-1 defects by the Arf GEF EFA-6 and the human cytohesin CYTH2, which localized to the cell periphery and not the nucleus, and the detection of endogenous GRP-1 in midbodies, further support this model.
While casual inspection of grp-1 mutant nuclei and nucleoli did not reveal any obvious phenotypes, it remains possible that GRP-1 has nuclear functions unrelated to its role in asymmetric divisions.
An alternative explanation for nuclear localization is to sequester GRP-1, either to remove excess GRP-1 from its site of function at the plasma membrane or to time its access to the plasma membrane at nuclear breakdown during mitosis. This model predicts that disruption of this nuclear sequestration should generate a phenotype. When we expressed some of the mutant GRP-1 proteins with an increased cytoplasmic localization in a wild-type background, however, we observed apparently normal Q.a and Q.p divisions (data not shown), suggesting that the cytoplasmic GRP-1 localization is not detrimental for the development of the Q lineage.
Mechanism of asymmetry
Our results suggest that Arf-dependent trafficking and the PIG-1 kinase function in a general mechanism that establish daughter cell-size asymmetry in C. elegans. The genes pig-1, grp-1, arf-1.2, and arf-3 also regulate cell shedding, a process that occurs in caspase-deficient mutants and results in the death of specific cells generated early in embryogenesis (Denning et al. 2012). In addition, pig-1 functions with the LKB-1/STRAD/MO25 polarity complex in both cell shedding and the asymmetric cell divisions described in this article (Chien et al. 2012; Denning et al. 2012). Denning et al. (2012) observed that mutations in pig-1, grp-1, arf-1.2, or arf-3 suppressed the shedding and death caused by loss of caspase activity and proposed that these genes negatively regulate adhesion of the shed cells to their neighbors. In their model, mutations in these genes increased adhesion, interfering with shedding and allowing the cells to survive. Our analysis of grp-1 and the findings of other studies demonstrate that mutations in the genes pig-1, grp-1, arf-1.2, or arf-3 can cause the transformation of cells that normally die into their sister cells (Cordes et al. 2006; Ou et al. 2010; Singhvi et al. 2011; Chien et al. 2012). We propose that the cell shedding defects of pig-1, grp-1, arf-1.2, and arf-3 mutants observed by Denning et al. (2012) result from defects in the asymmetric divisions that produce the shed cells. In this model, the daughter cell that would normally be shed in the caspase-deficient mutant instead adopts the fate of its sister in the pig-1, grp-1, arf-1.2, or arf-3 mutant background. We do not disagree with the hypothesis that increased adhesion in these mutants prevents the cells from being shed, but propose that the increased adhesion results indirectly from a transformation of a cell fated to die into its surviving sister cell.
Our findings may provide insights into the mechanism of action of the Arf pathway in asymmetrically dividing cells. GRP-1 localization in midbodies is compatible with a function at the cleavage furrow during anaphase and cytokinesis, at a time when cell size asymmetry is established by spindle displacement in Q.p or myosin polarity in Q.a (Ou et al. 2010). We also previously proposed that cnt-2 is required for endocytosis, in agreement with a requirement for GRP-1 at the plasma membrane (Singhvi et al. 2011). However, recent data indicate that CNT-2 and ARF-1.2 also function during the first division of the C. elegans embryo, and that CNT-2 is required to regulate cortical actomyosin dynamics (Fievet et al. 2012). Moreover, steppke, the Drosophila homolog of GRP-1, has been shown to antagonize myosin contractility at cleavage furrows in the cellularizing blastoderm (Lee and Harris 2013). Although the mechanism by which the Arf pathway controls daughter cell-size asymmetry remains unknown, it could involve the interplay of membrane trafficking and myosin contractility at the cleavage furrow or the asymmetric trafficking of polarity regulators or membrane to one pole of the cell.
Supplementary Material
Acknowledgments
We thank Yuji Kohara for providing grp-1 cDNAs; Jon Audhya and Karen Oegema for the PLCδ1 cDNA; Sean O’Rourke and Bruce Bowerman for sharing GFP::EFA-6 reagents; Richard Ikegami for sharing the gmIs65 transgene; Ofer Rog, Abby Dernburg, Guangshuo Ou, and Ron Vale for their help with time-lapse microscopy; and Shohei Mitani and the National Biosource Project and the Million Mutations Project for providing several of the mutants used in this study. Some nematode strains used in this work were provided by the Caenorhabditis Genetics Center, which is funded by the National Institutes of Health (NIH) Office of Research Infrastructure Programs (P40 OD010440). This work was supported by NIH grant NS32057 (G.G.). J.T. was supported by a fellowship from the Association pour la Recherche sur le Cancer.
Footnotes
Supporting information is available online at http://www.genetics.org/lookup/suppl/doi:10.1534/genetics.114.167189/-/DC1.
Communicating editor: P. Sengupta
Literature Cited
- Baumeister R., Liu Y., Ruvkun G., 1996. Lineage-specific regulators couple cell lineage asymmetry to the transcription of the Caenorhabditis elegans POU gene unc-86 during neurogenesis. Genes Dev. 10: 1395–1410 [DOI] [PubMed] [Google Scholar]
- Branda C. S., Stern M. J., 2000. Mechanisms controlling sex myoblast migration in Caenorhabditis elegans hermaphrodites. Dev. Biol. 226: 137–151 [DOI] [PubMed] [Google Scholar]
- Brenner S., 1974. The genetics of Caenorhabditis elegans. Genetics 77: 71–94 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Bülow H. E., Hobert O., 2004. Differential sulfations and epimerization define heparan sulfate specificity in nervous system development. Neuron 41: 723–736 [DOI] [PubMed] [Google Scholar]
- Casanova J. E., 2007. Regulation of Arf activation: the Sec7 family of guanine nucleotide exchange factors. Traffic 8: 1476–1485 [DOI] [PubMed] [Google Scholar]
- Cherfils J., Ménétrey J., Mathieu M., Le Bras G., Robineau S., et al. , 1998. Structure of the Sec7 domain of the Arf exchange factor ARNO. Nature 392: 101–105 [DOI] [PubMed] [Google Scholar]
- Chien S. C., Brinkmann E. M., Teulière J., Garriga G., 2013. Caenorhabditis elegans PIG-1/MELK acts in a conserved PAR-4/LKB1 polarity to promote asymmetric neuroblast divisions. Genetics 193: 897–909. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Clark S. G., Chiu C., 2003. C. elegans ZAG-1, a Zn-finger-homeodomain protein, regulates axonal development and neuronal differentiation. Development 130: 3781–3794 [DOI] [PubMed] [Google Scholar]
- Cohen L. A., Honda A., Várnai P., Brown F. D., Balla T., et al. , 2007. Active Arf6 recruits ARNO/cytohesin GEFs to the PM by binding their PH domains. Mol. Biol. Cell 18: 2244–2253 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Conradt B., Horvitz H. R., 1998. The C. elegans protein EGL-1 is required for programmed cell death and interacts with the Bcl-2-like protein CED-9. Cell 93: 519–529 [DOI] [PubMed] [Google Scholar]
- Cordes S., Frank C. A., Garriga G., 2006. The C. elegans MELK ortholog PIG-1 regulates cell size asymmetry and daughter cell fate in asymmetric neuroblast divisions. Development 133: 2747–2756 [DOI] [PubMed] [Google Scholar]
- Cowing D. W., Kenyon C., 1992. Expression of the homeotic gene mab-5 during Caenorhabditis elegans embryogenesis. Development 116: 481–490 [DOI] [PubMed] [Google Scholar]
- Cronin T. C., DiNitto J. P., Czech M. P., Lambright D. G., 2004. Structural determinants of phosphoinositide selectivity in splice variants of Grp1 family PH domains. EMBO J. 23: 3711–3720 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Denning D. P., Hatch V., Horvitz H. R., 2012. Programmed elimination of cells by caspase-independent cell extrusion in C. elegans. Nature 488: 226–230 [DOI] [PMC free article] [PubMed] [Google Scholar]
- DiNitto J. P., Delprato A., Gabe Lee M.-T., Cronin T. C., Huang S., et al. , 2007. Structural basis and mechanism of autoregulation in 3-phosphoinositide-dependent Grp1 family Arf GTPase exchange factors. Mol. Cell 28: 569–583 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Donaldson J. G., 2003. Multiple roles for Arf6: sorting, structuring, and signaling at the plasma membrane. J. Biol. Chem. 278: 41573–41576 [DOI] [PubMed] [Google Scholar]
- Donaldson J. G., Jackson C. L., 2011. ARF family G proteins and their regulators: roles in membrane transport, development and disease. Nat. Rev. Mol. Cell Biol. 12: 362–375 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Ellis H. M., Horvitz H. R., 1986. Genetic control of programmed cell death in the nematode C. elegans. Cell 44: 817–829 [DOI] [PubMed] [Google Scholar]
- Fievet B. T., Rodriguez J., Naganathan S., Lee C., Zeiser E., et al. , 2012. Systematic genetic interaction screens uncover cell polarity regulators and functional redundancy. Nat. Cell Biol. 15: 103–112. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Finney M., Ruvkun G., Horvitz H. R., 1988. The C. elegans cell lineage and differentiation gene unc-86 encodes a protein with a homeodomain and extended similarity to transcription factors. Cell 55: 757–769 [DOI] [PubMed] [Google Scholar]
- Frank C. A., Baum P. D., Garriga G., 2003. HLH-14 is a C. elegans achaete-scute protein that promotes neurogenesis through asymmetric cell division. Development 130: 6507–6518 [DOI] [PubMed] [Google Scholar]
- Frank C. A., Hawkins N. C., Guenther C., Horvitz H. R., Garriga G., 2005. C. elegans HAM-1 positions the cleavage plane and regulates apoptosis in asymmetric neuroblast divisions. Dev. Biol. 284: 301–310 [DOI] [PubMed] [Google Scholar]
- Frøkjær-Jensen C., Davis M. W., Hopkins C. E., Newman B. J., Thummel J. M., et al. , 2008. Single-copy insertion of transgenes in Caenorhabditis elegans. Nat. Genet. 40: 1375–1383 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Frøkjær-Jensen C., Davis M. W., Sarov M., Taylor J., Flibotte S., et al. , 2014. Random and targeted transgene insertion in Caenorhabditis elegans using a modified Mos1 transposon. Nat. Methods 11: 529–534 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Garriga G., Desai C., Horvitz H. R., 1993. Cell interactions control the direction of outgrowth, branching and fasciculation of the HSN axons of Caenorhabditis elegans. Development 117: 1071–1087 [DOI] [PubMed] [Google Scholar]
- Godi A., Pertile P., Meyers R., Marra P., Di Tullio G., et al. , 1999. ARF mediates recruitment of PtdIns-4-OH kinase-beta and stimulates synthesis of PtdIns(4,5)P2 on the Golgi complex. Nat. Cell Biol. 1: 280–287 [DOI] [PubMed] [Google Scholar]
- Guenther C., Garriga G., 1996. Asymmetric distribution of the C. elegans HAM-1 protein in neuroblasts enables daughter cells to adopt distinct fates. Development 122: 3509–3518 [DOI] [PubMed] [Google Scholar]
- Gurling M., Talavera K., Garriga G., 2014. The DEP domain-containing protein TOE-2 promotes apoptosis in the Q lineage of C. elegans through two distinct mechanisms. Development 141: 2724–2734 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Hatzold J., Conradt B., 2008. Control of apoptosis by asymmetric cell division. PLoS Biol. 6: e84. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Hawkins N. C., Ellis G. C., Bowerman B., Garriga G., 2005. MOM-5 frizzled regulates the distribution of DSH-2 to control C. elegans asymmetric neuroblast divisions. Dev. Biol. 284: 246–259 [DOI] [PubMed] [Google Scholar]
- Hirose T., Horvitz H. R., 2013. An Sp1 transcription factor coordinates caspase-dependent and -independent apoptotic pathways. Nature 500: 354–358 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Hofmann I., Thompson A., Sanderson C. M., Munro S., 2007. The Arl4 family of small G proteins can recruit the cytohesin Arf6 exchange factors to the plasma membrane. Curr. Biol. 17: 711–716 [DOI] [PubMed] [Google Scholar]
- Honda A., Nogami M., Yokozeki T., Yamazaki M., Nakamura H., et al. , 1999. Phosphatidylinositol 4-phosphate 5-kinase alpha is a downstream effector of the small G protein ARF6 in membrane ruffle formation. Cell 99: 521–532 [DOI] [PubMed] [Google Scholar]
- Hsu D. R., Chuang P. T., Meyer B. J., 1995. DPY-30, a nuclear protein essential early in embryogenesis for Caenorhabditis elegans dosage compensation. Development 121: 3323–3334 [DOI] [PubMed] [Google Scholar]
- Jackson T. R., Kearns B. G., Theibert A. B., 2000. Cytohesins and centaurins: mediators of PI 3-kinase-regulated Arf signaling. Trends Biochem. Sci. 25: 489–495 [DOI] [PubMed] [Google Scholar]
- Jones D. H., Morris J. B., Morgan C. P., Kondo H., Irvine R. F., et al. , 2000. Type I phosphatidylinositol 4-phosphate 5-kinase directly interacts with ADP-ribosylation factor 1 and is responsible for phosphatidylinositol 4,5-bisphosphate synthesis in the golgi compartment. J. Biol. Chem. 275: 13962–13966 [DOI] [PubMed] [Google Scholar]
- Kahn R. A., Cherfils J., Elias M., Lovering R. C., Munro S., et al. , 2006. Nomenclature for the human Arf family of GTP-binding proteins: ARF, ARL, and SAR proteins. J. Cell Biol. 172: 645–650 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Kamath R. S., M. Martinez-Campos, P. Zipperlen, A. G. Fraser, J. Ahringer, 2001 Effectiveness of specific RNA-mediated interference through ingested double-stranded RNA in Caenorhabditis elegans. Genome Biol. 2: RESEARCH0002. [DOI] [PMC free article] [PubMed]
- Kennedy S., Wang D., Ruvkun G., 2004. A conserved siRNA-degrading RNase negatively regulates RNA interference in C. elegans. Nature 427: 645–649 [DOI] [PubMed] [Google Scholar]
- Kim K., Li C., 2004. Expression and regulation of an FMRFamide-related neuropeptide gene family in Caenorhabditis elegans. J. Comp. Neurol. 475: 540–550 [DOI] [PubMed] [Google Scholar]
- Kiyomitsu T., Cheeseman I. M., 2013. Cortical dynein and asymmetric membrane elongation coordinately position the spindle in anaphase. Cell 154: 391–402 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Klarlund J. K., Tsiaras W., Holik J. J., Chawla A., Czech M. P., 2000. Distinct polyphosphoinositide binding selectivities for pleckstrin homology domains of GRP1-like proteins based on diglycine vs. triglycine motifs. J. Biol. Chem. 275: 32816–32821 [DOI] [PubMed] [Google Scholar]
- Kolanus W., 2007. Guanine nucleotide exchange factors of the cytohesin family and their roles in signal transduction. Immunol. Rev. 218: 102–113 [DOI] [PubMed] [Google Scholar]
- Kolanus W., Nagel W., Schiller B., Zeitlmann L., Godar S., et al. , 1996. Alpha L beta 2 integrin/LFA-1 binding to ICAM-1 induced by cytohesin-1, a cytoplasmic regulatory molecule. Cell 86: 233–242 [DOI] [PubMed] [Google Scholar]
- Kreis T. E., Lowe M., Pepperkok R., 1995. COPs regulating membrane traffic. Annu. Rev. Cell Dev. Biol. 11: 677–706 [DOI] [PubMed] [Google Scholar]
- Lee D. M., Harris T. J. C., 2013. An Arf-GEF regulates antagonism between endocytosis and the cytoskeleton for Drosophila blastoderm development. Curr. Biol. 23: 2110–2120 [DOI] [PubMed] [Google Scholar]
- Li C., Kim K., Nelson L. S., 1999. FMRFamide-related neuropeptide gene family in Caenorhabditis elegans. Brain Res. 848: 26–34 [DOI] [PubMed] [Google Scholar]
- Li C.-C., Chiang T.-C., Wu T.-S., Pacheco-Rodriguez G., Moss J., et al. , 2007. ARL4D recruits cytohesin-2/ARNO to modulate actin remodeling. Mol. Biol. Cell 18: 4420–4437 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Li Y., Kelly W. G., Logsdon J. M., Schurko A. M., Harfe B. D., et al. , 2004. Functional genomic analysis of the ADP-ribosylation factor family of GTPases: phylogeny among diverse eukaryotes and function in C. elegans. FASEB J. 18: 1834–1850 [DOI] [PubMed] [Google Scholar]
- Liang J. O., Kornfeld S., 1997. Comparative activity of ADP-ribosylation factor family members in the early steps of coated vesicle formation on rat liver Golgi membranes. J. Biol. Chem. 272: 4141–4148 [DOI] [PubMed] [Google Scholar]
- Maduro M., Pilgrim D., 1995. Identification and cloning of unc-119, a gene expressed in the Caenorhabditis elegans nervous system. Genetics 141: 977–988 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Moss J., Vaughan M., 2002. Cytohesin-1 in 2001. Arch. Biochem. Biophys. 397: 156–161 [DOI] [PubMed] [Google Scholar]
- Mossessova E., Gulbis J. M., Goldberg J., 1998. Structure of the guanine nucleotide exchange factor Sec7 domain of human arno and analysis of the interaction with ARF GTPase. Cell 92: 415–423 [DOI] [PubMed] [Google Scholar]
- Nakai K., Horton P., 1999. PSORT: a program for detecting sorting signals in proteins and predicting their subcellular localization. Trends Biochem. Sci. 24: 34–36 [DOI] [PubMed] [Google Scholar]
- Nie Z., Hirsch D. S., Randazzo P. A., 2003. Arf and its many interactors. Curr. Opin. Cell Biol. 15: 396–404 [DOI] [PubMed] [Google Scholar]
- O’Rourke S. M., Christensen S. N., Bowerman B., 2010. Caenorhabditis elegans EFA-6 limits microtubule growth at the cell cortex. Nat. Cell Biol. 12: 1235–1241 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Ou G., Stuurman N., D’Ambrosio M., Vale R. D., 2010. Polarized myosin produces unequal-size daughters during asymmetric cell division. Science 330: 677–680 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Popoff V., Langer J. D., Reckmann I., Hellwig A., Kahn R. A., et al. , 2011. Several ADP-ribosylation factor (Arf) isoforms support COPI vesicle formation. J. Biol. Chem. 286: 35634–35642 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Potts M. B., Cameron S., 2011. Cell lineage and cell death: Caenorhabditis elegans and cancer research. Nat. Rev. Cancer 11: 50–58 [DOI] [PubMed] [Google Scholar]
- Randazzo P. A., Nie Z., Miura K., Hsu V. W., 2000. Molecular aspects of the cellular activities of ADP-ribosylation factors. Sci. STKE 2000: re1. [DOI] [PubMed] [Google Scholar]
- Rual J.-F., Ceron J., Koreth J., Hao T., Nicot A.-S., et al. , 2004. Toward improving Caenorhabditis elegans phenome mapping with an ORFeome-based RNAi library. Genome Res. 14: 2162–2168 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Salser S. J., Kenyon C., 1992. Activation of a C. elegans Antennapedia homologue in migrating cells controls their direction of migration. Nature 355: 255–258 [DOI] [PubMed] [Google Scholar]
- Shaham S., Reddien P. W., Davies B., Horvitz H. R., 1999. Mutational analysis of the Caenorhabditis elegans cell-death gene ced-3. Genetics 153: 1655–1671 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Simmer F., Tijsterman M., Parrish S., Koushika S. P., Nonet M. L., et al. , 2002. Loss of the putative RNA-directed RNA polymerase RRF-3 makes C. elegans hypersensitive to RNAi. Curr. Biol. 12: 1317–1319 [DOI] [PubMed] [Google Scholar]
- Singhvi A., Teulière J., Talavera K., Cordes S., Ou G., et al. , 2011. The Arf GAP CNT-2 regulates the apoptotic fate in C. elegans asymmetric neuroblast divisions. Curr. Biol. 21: 948–954 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Stalder D., Barelli H., Gautier R., Macia E., Jackson C. L., et al. , 2011. Kinetic studies of the Arf activator Arno on model membranes in the presence of Arf effectors suggest control by a positive feedback loop. J. Biol. Chem. 286: 3873–3883 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Sze J. Y., Zhang S., Li J., Ruvkun G., 2002. The C. elegans POU-domain transcription factor UNC-86 regulates the tph-1 tryptophan hydroxylase gene and neurite outgrowth in specific serotonergic neurons. Development 129: 3901–3911 [DOI] [PubMed] [Google Scholar]
- Timmons L., Fire A., 1998. Specific interference by ingested dsRNA. Nature 395: 854. [DOI] [PubMed] [Google Scholar]
- Várnai P., Bondeva T., Tamás P., Tóth B., Buday L., et al. , 2005. Selective cellular effects of overexpressed pleckstrin-homology domains that recognize PtdIns(3,4,5)P3 suggest their interaction with protein binding partners. J. Cell Sci. 118: 4879–4888 [DOI] [PubMed] [Google Scholar]
- Volpicelli-Daley L. A., Li Y., Zhang C.-J., Kahn R. A., 2005. Isoform-selective effects of the depletion of ADP-ribosylation factors 1–5 on membrane traffic. Mol. Biol. Cell 16: 4495–4508 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Wicks S. R., Yeh R. T., Gish W. R., Waterston R. H., Plasterk R. H., 2001. Rapid gene mapping in Caenorhabditis elegans using a high density polymorphism map. Nat. Genet. 28: 160–164 [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.