C. elegans yolk is comprised of lipid associated with six apoB-like vitellogenins. Yolk is synthesized in the hermaphrodite intestine and secreted into the pseudocoelom (body cavity) before being endocytosed into maturing oocytes (Kimble and Sharrock, 1983; Grant and Hirsh, 1999; Hall et al. , 1999; Perez and Lehner, 2019). During embryogenesis yolk granules becomes distributed amongst the dividing blastomeres. Once the primordial intestine is formed, yolk accumulates in the intestinal cells (Bossinger and Schierenberg, 1996). Little is known about how yolk is distributed during embryogenesis.

Arf GTPases are regulators of membrane trafficking that cycle between a GTP-bound “on” state and a GDP-bound “off” state regulated by guanine nucleotide exchange factors (GEFs) and GTPase activating proteins (GAPs), respectively (Jackson and Bouvet, 2014). When bound to GTP, Arfs can interact with effector proteins to regulate membrane trafficking. C. elegans AGEF-1 is a putative GEF orthologous to yeast Sec7p and human ARFGEF1 and ARFGEF2 that activate class I and II Arf GTPases (Togawa et al. , 1999; Sato et al. , 2006; Ishizaki et al. , 2008; Skorobogata et al. , 2014). We previously reported that a missense allele, agef-1 ( vh4 ) , that results in a Glutamic acid to Lysine change in the HDS2 domain ( Figure 1A ) caused mislocalization of the LET-23 /Epidermal Growth Factor Receptor (EGFR) in the vulva precursor cells and caused enlargement of endosomes and lysosomes in coelomocytes (Skorobogata et al. , 2014). Both phenotypes were phenocopied by agef-1 (RNAi) , and a deletion allele was zygotic lethal, suggesting that agef-1 ( vh4 ) was a hypomorphic allele (Tang et al. , 2012; Skorobogata et al. , 2014).

Here we report that agef-1 ( vh4 ) mutant embryos had a yolk trafficking phenotype. Unlike wild type, agef-1 ( vh4 ) embryos accumulated extraembryonic yolk in between the eggshell and the developing embryo as determined by differential interference contrast microscopy and confirmed with a yolk protein fusion VIT-2::GFP ( bIs1 ) ( Figure 1B ). In nearly all agef-1 ( vh4 ) embryos VIT-2::GFP was found in pools or blobs rather than in the intestine ( Figure 1B, C). While many yolk blobs were clearly floating between the embryo and the eggshell, many appeared to accumulate internally ( Figure 1D ). Analysis of newly hatched larvae confirmed that 53% (n=41) had yolk blobs that failed to extrude from the embryo ( Figure 1C, upper right). However, it was not clear if this phenotype was caused by loss of agef-1 or a background mutation as we did not observe a yolk blob phenotype by agef-1 (RNAi) (0% yolk blobs across 3 replicates, n= 72-84 embryos/replicate).

Recent structural analysis of the yeast ortholog of AGEF-1 , Sec7p, revealed that the HDS2 domain interacts with the SEC7 GEF domain, blocking the interaction with Arf1 (Brownfield et al. , 2024). Consistent with this being an autoinhibitory interaction, an engineered L1376D mutation in the HDS2 domain greatly increased Sec7p GEF activity toward Arf1 in vitro . This Leucine is conserved in AGEF-1 and is adjacent to the Glutamic Acid that is mutated in agef-1 ( vh4 ) suggesting that this allele could also be an activating allele ( Figure 1A ). To test if agef-1 ( vh4 ) is indeed an activating allele, we performed agef-1 RNAi on the agef-1 ( vh4 ) mutant and found that it potently suppressed the yolk blob phenotype while control empty vector (EV) RNAi had no effect ( Figure 1C, E). Thus, agef-1 ( vh4 ) is likely a hypermorphic allele and the yolk blobs may be a result of increased Arf GTPase activity.

AGEF-1 functions with ARF-1 and ARF-5 to antagonize LET-23 /EGFR localization and signaling during vulva development (Skorobogata et al. , 2014). We found that RNAi of either arf-1 or arf-5 strongly suppressed the agef-1 ( vh4 ) yolk blob phenotype suggesting that both are required ( Figure 1C, E). To ensure that this suppression is not caused by decreased yolk uptake into maturing oocytes we measured fluorescence intensity of VIT-2::GFP in bean-stage embryos treated with RNAi targeting agef-1 , arf-1 or arf-5 . We found no decrease in VIT-2::GFP fluorescence intensity from these RNAi knockdowns and in fact arf-1 (RNAi) significantly increased the levels of VIT-2::GFP ( Figure 1F ). Thus, ARF-1 and ARF-5 activity are required for the yolk blob phenotype of agef-1 ( vh4 ) without blocking yolk uptake into embryos.

The agef-1 ( vh4 ) mutant behaved as an activating allele in the context of yolk trafficking but acted as a loss of function mutation during LET-23 /EGFR-mediated vulva development and in regulation of lysosome size in coelomocytes (Skorobogata et al. , 2014). In mammalian cells Arf1 activation is required for AP-1 recruitment (Stamnes and Rothman, 1993), but GTP hydrolysis appears to be required for AP-1 uncoating and subsequent vesicle trafficking (Tanigawa et al. , 1993; Zhu et al. , 1998; Meyer et al. , 2005). We hypothesized that the AP-1 complex would not be required for the agef-1 ( vh4 ) yolk blob phenotype. We found that RNAi targeting AP-1 complex components apg-1 , apm-1 or aps-1 did not suppress the agef-1 ( vh4 ) yolk blob phenotype despite causing a potent dead egg phenotype ( Figure 1C, G). Therefore, the aberrant yolk trafficking seen in agef-1 ( vh4 ) happens independently of the AP-1 clathrin adaptor complex.

Here we show that agef-1 ( vh4 ) is likely a hypermorphic allele and that aberrant AGEF-1 activity results in mislocalization of yolk during embryogenesis. Unlike other agef-1 ( vh4 ) phenotypes that are phenocopied by RNAi the yolk trafficking phenotype is suppressed by agef-1 (RNAi) as well as with RNAi targeting arf-1 and arf-5 . However, the agef-1 ( vh4 ) yolk trafficking phenotype is not suppressed by RNAi targeting components of the AP-1 complex. These data are consistent with a model whereby increased AGEF-1 activity activates ARF-1 /5, which in case of yolk trafficking, engages an unidentified effector that does not require cycling like AP-1 ( Figure 1H ). The strong phenotypes caused by RNAi targeting either arf-1 or arf-5 could suggest a requirement for both GTPases. However, we noted that arf-1 and arf-5 coding sequences and hence their RNAi clones share stretches of identical nucleotide sequences that could cause some cross reactivity. Therefore, further experiments with mutants will be required to determine whether one or both GTPases regulate yolk trafficking. Furthermore, we used VIT-2::GFP as a proxy for yolk, thus we cannot exclude the possibility that yolk comprised of other vitellogenins are trafficked normally in agef-1 ( vh4 ) mutants.

A similar yolk trafficking phenotype was reported for loss alfa-1 /C9orf72 and smcr-8 /SMCR8 , where ALFA-1 and SMCR-8 were found to regulate endolysosomal trafficking (Corrionero and Horvitz, 2018). Intriguingly, structural and biochemical data suggests that C9orf72 and SMCR8 function together as an Arf GAP (Su et al. , 2020; Su et al. , 2021). If this is the case, then Arf GTPase activity would be increased in alfa-1 and smcr-8 mutants. Further analysis will be required to determine if ALFA-1 and SMCR-8 function in a common pathway with AGEF-1 to regulate yolk trafficking.

Human ARFGEF1 variants are associated with developmental delay with and without epilepsy (Takata et al. , 2019; Thomas et al. , 2021; Xu et al. , 2022). While most ARFGEF1 mutants introduce premature stop codons and frameshifts some are missense mutations. Notably the I1180R mutation in the HDS2 domain of ARFGEF1 could be an activating allele as the analogous residue in yeast makes contact with the Sec7 GEF domain in the autoinhibited state (Brownfield et al. , 2024). If so, we would expect the corresponding mutant in agef-1 would cause a yolk trafficking phenotype. Thus, C. elegans embryonic yolk trafficking could be used to assay the effects of conserved ARFGEF1 disease alleles.

Wormbase ( https://wormbase.org ) was an invaluable resource to the planning and execution of this work (Sternberg et al. , 2024). All strains were maintained at 20°C on Nematode Growth Medium (NGM) and fed HB101 E. coli as a food source, as previously described (Brenner, 1974; Stiernagle, 2006). All strains were derived from the wild type N2 strain. RNAi by feeding was performed as previously reported (Kamath et al. , 2003). All RNAi experiments were performed a minimum of 3 times. All images were collected on an Axio Imager A1 using the Axiocam 305 with Axio Vision software (Zeiss). Embryos were mounted on 2% agarose pads in water and imaged using glass coverslips between 0.16 to 0.19 mm (Fisher). For yolk blob scoring, a minimum of one VIT-2::GFP positive blob was the threshold for the presence of a yolk blob. A percent of blobs present was taken after each round of RNAi. For total VIT-2::GFP amounts, images were captured as previously described (Chotard et al. , 2010b) and mean fluorescence intensity was measured using the selection tool brush in Fiji, and subtracting background ( https://imagej.net/software/fiji/ ). Statistics and graphical analysis were performed using Graph Pad Prism 10. Sequence alignment was performed using EMBOSS Water Pairwise Sequence Alignment (PSA) ( https://www.ebi.ac.uk/jdispatcher/psa/emboss_water?format=clustal ). Visualization and final alignment was performed using Jalview version 2.11.4.1 ( https://www.jalview.org/ ).