The replicative helicase CMG is required for the divergence of cell fates during asymmetric cell divisionin vivo

The mechanisms that enable differential gene expression in daughter cells produced by asymmetric cell divisions are not well understood. We discovered that the eukaryotic replicative helicase CMG (Cdc45-MCM-GINS) is required for this process in C. elegans. During C. elegans development, some dividing cells give rise to a daughter that survives and a daughter that dies. We found that PSF-2 GINS2, a component of C. elegans CMG, is necessary for the transcriptional burst of the pro-apoptotic gene egl-1 BH3-only, which occurs in the daughter that dies immediately following mother cell division. We present evidence that this requirement is independent of the function of CMG in DNA unwinding. We propose that the recently described histone chaperone activity of CMG causes epigenetic changes at the egl-1 locus during replication in mother cells, and that these changes are required for the increase in egl- 1 transcription in the daughter that dies. We also find that PSF-2 is required for the divergence of other cell fates during C. elegans development, suggesting that this function is not restricted to the regulation of egl-1 expression. Our work uncovers a new and unexpected role of CMG in cell fate and a novel intrinsic mechanism for gene expression plasticity in the context of asymmetric cell division. .CC-BY-NC-ND 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 13, 2024. ; https://doi.org/10.1101/2024.05.09.593335doi: bioRxiv preprint 3

The ability of cells to adopt different fates is fundamentally important to life. One process through which cells of different fates are generated is through asymmetric cell division. The mechanisms through which gene expression patterns and, hence, cell fates diverge during asymmetric cell division are incompletely understood 1-4. A large body of work has demonstrated that different daughter cell fates can be established through the asymmetric inheritance of cell fate determinants. Whether the process of cell division itself plays a role is not clear. The quantal cell cycle theory proposes that changes at the chromosomal level during DNA replication make available - for transcription in daughter cells - regions of the genome that were not available for transcription in the mother cell 5,6. Whether and how this may occur in vivo is unknown. The development of the nematode C. elegans is essentially invariant and provides a unique opportunity to study cell fate and how a cell acquires a specific fate during its lineage history, i.e. the successive rounds of cell division starting from the first division of the one-cell embryo to the terminal division that generates that cell 7. Most terminal divisions during C. elegans development are asymmetric and result in two daughter cells that adopt different fates8,9. Lineage-resolved single cell transcriptome profiling of C. elegans embryos has revealed that in the case of such terminal asymmetric divisions, the genes that determine the two daughter cell fates are often co-expressed in the mother cell 10, a phenomenon observed in other organisms and referred to as multilineage priming 11-17. The intrinsic and extrinsic mechanisms through which the expression of genes subject to multilineage priming is modified during terminal asymmetric divisions to retain expression in one but not the other daughter cell, are not fully understood. Out of the 1090 somatic cells formed during the development of a C. elegans hermaphrodite, 131 reproducibly die 8,9. Cell death during C. elegans development can therefore .CC-BY-NC-ND 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 13, 2024. ; https://doi.org/10.1101/2024.05.09.593335doi: bioRxiv preprint 4 be considered a genetically programmed cell fate. Most cells that adopt the cell death fate are generated through an asymmetric cell division and after the completion of mother cell division, very rapidly (within 20-30 min) undergo apoptotic cell death 18,19. The gene that determines the fate of these cells is egl-1, which encodes a BH3-only protein, a pro-apoptotic member of the Bcl-2 superfamily of cell death regulators20-22. egl-1 is necessary and sufficient for apoptotic cell death, and its expression is essentially restricted to lineages in which an apoptotic cell death occurs. Using single molecule RNA Fluorescence In Situ Hybridization (smRNA FISH), the dynamics of egl-1 mRNA levels has been analyzed in specific cell death lineages in vivo 23. This revealed that a low level of egl-1 mRNA is already present in the mother cell. Immediately after mother cell division, the egl-1 mRNA level increases substantially in the daughter that dies but decreases to zero in the daughter that survives. This suggests that mothers of cells that die are ‘poised’ for egl-1 expression and that the regulation of egl-1 expression is binarized during the terminal asymmetric division to result in increased egl-1 expression specifically in the daughter that dies. Indeed, lineage-resolved single cell transcriptome profiling of C. elegans embryos identified egl-1 as a gene that is subject to multilineage priming 10, confirming the smRNA FISH results23. It has been proposed that the non-random segregation of direct repressors of egl-1 transcription into the daughters that survive contributes to repression of egl-1 expression in these cells24,25. How egl-1 expression is increased in their dying sister cells has so far been unclear. Here we demonstrate that the increase in egl-1 BH3-only expression in the daughters that die is dependent on the eukaryotic replicative helicase CMG (Cdc45-MCM2-7-GINS), and we provide evidence that this requirement of CMG is independent of its role in DNA unwinding. In addition, we demonstrate that the role of CMG in the divergence of cell fates during asymmetric cell .CC-BY-NC-ND 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 13, 2024. ; https://doi.org/10.1101/2024.05.09.593335doi: bioRxiv preprint 5 division is not restricted to the cell death fate. Our results uncover an intrinsic mechanism through which the expression of a gene that is subject to multilineage priming can be altered during terminal asymmetric cell divisions. Importantly, they also provide the first in vivo evidence for a role of components of the replisome in the control of gene expression during asymmetric cell division, providing experimental support for the quantal cell cycle theory. .CC-BY-NC-ND 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 13, 2024. ; https://doi.org/10.1101/2024.05.09.593335doi: bioRxiv preprint 6

Reducing psf-2 GINS2 function causes embryonic lethality and increased cell cycle lengths We identified the t3443 ts mutation by screening a collection of temperature-sensitive (ts) embryonic lethal mutants for abnormalities in the invariant pattern of cell death. At the non- permissive temperature (25°C), the morphology of early t3443ts embryos is essentially indistinguishable from that of wild-type embryos ( Fig. 1A , 4-cell stage, Pre-morphogenetic stage) (see Materials and Methods for the exact time of the shift from permissive to non- permissive temperature). The different tissues can be recognized, and a normal pre- morphogenetic stage is reached, which suggests that tissue differentiation overall is not impaired. However, t3443ts mutants undergo arrest shortly after the initiation of morphogenesis ( Fig. 1A, see ‘Final recording’). This embryonic lethal (Emb) phenotype is fully penetrant (100% embryonic lethal) ( Fig. 1B). At the permissive temperature (15°C), t3443ts animals are viable (Fig. 1B ). Furthermore, there is no significant difference in brood size (number of eggs laid) between t3443ts and wild type at the permissive or non-permissive temperature (Fig. 1B) Using snip-SNP mapping 26, we mapped the Emb phenotype of t3443ts animals to Linkage Group I (LGI) close to the variation pkP1071 at position 23.40 cM and performed whole genome sequencing. In the region identified, we found one gene that carries a missense mutation in its coding region in t3443ts animals, the gene psf-2 (yeast Partner of Sld Five). psf-2 encodes the C. elegans ortholog of Psf2, one of four subunits (Psf1, Psf2, Psf3, Sld5) of the GINS (G o- Ichi-Ni-San) complex 27. (Of note, Psf1, Psf2, Psf3 and Sld5 are also referred to as GINS1, GINS2, GINS3 and GINS4, respectively.) GINS is a subcomplex of the conserved replicative helicase CMG, which unwinds double-strand DNA prior to DNA synthesis and is therefore essential for DNA replication 28,29. t3443ts is a missense mutation that causes a cytosine to .CC-BY-NC-ND 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 13, 2024. ; https://doi.org/10.1101/2024.05.09.593335doi: bioRxiv preprint 7 thymine change at position 190 (C190T) of the coding region of the psf-2 gene (Suppl Fig. 1A). This results in a predicted proline to serine change at position 64 of the amino acid sequence of the PSF-2 protein (P64S), a residue that is conserved from S. cerevisiae to human 27 (Suppl Fig. 1B, C). To confirm that t3443 ts is an allele of psf-2, we amplified a 3.8 kb genomic fragment, which spans the psf-2 transcription unit and upstream and downstream regions (Suppl Fig. 1A, B bcEx1302 3.8kb) from wild type and injected it into t3443ts animals to generate a psf-2(+) transgene. We found that this transgene rescues the Emb phenotype of t3443ts animals. t3443ts embryos carrying the psf-2 (+) transgene complete embryogenesis and hatch ( Fig. 1A-C , psf- 2(t3443ts); (psf-2(+)). As an additional form of verification, we knocked-down the psf-2 gene by RNA interference (RNAi) in wild-type animals for 24h (see Fig. 1A and Suppl Figure 2 for phenotype at final recording). We found that this causes an Emb phenotype in the F1 progeny that resembles the phenotype observed in psf-2(t3443ts) animals at the non-permissive temperature (Fig. 1A, psf-2(RNAi)). These results demonstrate that t3443ts is a loss-of-function mutation of the gene psf-2, which is verified further by the experiments described below. To analyze the development of psf-2(t3443ts) animals in more detail, we performed 4D microscopy using Differential Interference Contrast (DIC) combined with cell lineage analyses (‘4D lineaging’) 30,31. We identified the ABarp blastomere at the 12-cell stage and measured cell cycle length during the five consecutive rounds of cell division that give rise to the ABarpppppp blastomere (also referred to as ‘V6R’) (Fig. 1C, D ). In wild-type animals, cell cycle lengths increase from an average of 22 min to an average of 40 min during these five rounds of cell divisions. In psf-2(t3443ts) animals, during the same five rounds of cell division, cell cycle .CC-BY-NC-ND 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 13, 2024. ; https://doi.org/10.1101/2024.05.09.593335doi: bioRxiv preprint 8 lengths increase from an average of 39 min to an average of 144 min ( Fig 1.C, D ). Therefore, cell cycle length in psf-2(t3443ts) animals is increased almost 2-fold at the beginning of the recordings (22 min versus 39 min) and almost 4-fold at the end of the recording (40 min versus 144 min). This ‘increased cell cycle length’ phenotype of psf-2(t3443ts) animals is rescued by the psf-2(+) transgene ( Fig. 1C, D ). Furthermore, the knock-down by RNAi of psf-2 (for 24h) increases cell cycle length similarly to what we observed in psf-2(t3443ts) animals at the non- permissive temperature (Fig. 1C, D). psf-2(RNAi) also leads to the block of some cell divisions. For example, in the cell lineages shown in Figure 1D, the divisions of the cells ABarpppaa and ABarpppap were blocked in psf-2(RNAi) animals (indicated in red). Finally, the knock-down of psf-2 by RNAi (for 48h) leads to an arrest at about the 50-cell stage (see Suppl Figure 2 for phenotype at final recording). To confirm that the increased cell cycle length phenotype observed in psf-2(t3443ts) and psf-2 (RNAi) animals is not specific to the ABarpppppp lineage, we also analyzed the MSpppppp lineage. We identified the MS blastomere at the 16-cell stage and measured cell cycle length during the six consecutive rounds of cell division that give rise to the cell MSpppppp. As in the ABarpppppp lineage, we found an increased cell cycle length phenotype in psf-2(t3443ts) animals at the non-permissive temperature, and this phenotype was rescued by psf-2(+) (Suppl Fig. 3A, B ). In summary, in line with the essential role of CMG in DNA replication, reducing psf-2 GINS2 function causes increased cell cycle lengths, ultimately resulting in a block in cell division and embryonic lethality. Reducing psf-2 GINS2 function blocks the cell death fate During C. elegans development, 131 cells adopt the cell death fate and reproducibly die 8,9. Using the same 4D recordings as above, we analyzed the number of cell deaths in psf-2(t3443ts) .CC-BY-NC-ND 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 13, 2024. ; https://doi.org/10.1101/2024.05.09.593335doi: bioRxiv preprint 9 embryos at the non-permissive temperature to determine whether the cell death fate was affected in these mutants. After the 8 th round of cell division, one cell derived from the MS blastomere dies (MSpaapp) and after the 9th round of division, 13 cells derived from the AB blastomere die, together referred to as the 1st wave of cell death. We identified the 13 cell death lineages derived from AB and determined the fate of the 13 cells that normally die. As shown in Figure 2A, we found that in wild type, all 13 cells die (+/+, 0% cell death blocked, n=4). In contrast, in ced- 3(n717) animals, which exhibit a general block in cell death or Ced phenotype 32, 100% of cell death is blocked (a cell death was considered blocked, when the cell had not turned into a cell corpse based on DIC after twice the cell cycle length of its mother cell; see Materials and Methods) ( Fig. 2B ). We analyzed four psf-2(t3443ts) embryos and found that at the non- permissive temperature, the death of many of the 13 cells is blocked (Fig. 2A, psf-2(t3443ts)). In total, we found that at the non-permissive temperature, 62% of the 1 st wave cell deaths derived from AB are blocked in psf-2(t3443ts) animals (Fig. 2B). In addition, we found that the death of MSpaapp is blocked in four out of six psf-2(t3443ts) embryos analyzed (67% cell death blocked) (Fig. 2C). Of note, psf-2(t3443ts) affects cell deaths in all cell death lineages ( Fig. 2A), which indicates that it causes a general rather than lineage-specific block in cell death. Importantly, the psf-2(+) transgene fully rescues the Ced phenotype observed in psf-2(t3443ts) animals (Fig. 2A- C). In addition, the knock-down of psf-2 by RNAi (for 24h) blocks 64% of the 1 st wave cell deaths derived from AB (Fig. 2A, B). This confirms that the Ced phenotype observed in t3443ts is caused by a reduction in psf-2 GINS2 function. For the 1 st wave cell deaths derived from AB that were not blocked in psf-2(t3443ts) animals (i.e., 38% of the cell deaths), we measured the time it took the cells to die. Specifically, we measured the time between the completion of the mother cell division that gives rise to a particular cell and that cell’s adoption of a cell corpse .CC-BY-NC-ND 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 13, 2024. ; https://doi.org/10.1101/2024.05.09.593335doi: bioRxiv preprint 10 appearance by DIC. We found that in wild type, it takes a cell on average 23.6 min to die ( Fig. 2D). In contrast, in psf-2(t3443ts) animals, it takes a cell on average 49.7 min to die ( Fig. 2D). This indicates that in the 38% of 1 st wave cell deaths that still occur in psf-2(t3443ts) animals, cell death is delayed. Of the 131 cell deaths that occur during C. elegans development, 18 occur post- embryonically8. To determine whether psf-2(t3443ts) also blocks post-embryonic cell deaths, we analyzed the death of the cell QL.pp, which dies in larvae of the 1 st larval stage (L1 larvae). Using a Q lineage specific reporter (Ptoe-2gfp)33, we found that at the non-permissive temperature, the death of QL.pp was blocked in 29% of psf-2(t3443ts) animals (Fig. 2E). Based on these results we conclude that reducing psf-2 GINS2 function not only causes a fully penetrant Emb phenotype, but also an incompletely penetrant Ced phenotype. In addition, the Ced phenotype exhibits variable expressivity, ranging from a block in cell death to an increase in the time it takes a cell to die. Hence, C. elegans psf-2 GINS2 is required for the cell death fate. The loss of other components of the replicative helicase CMG blocks the cell death fate PSF-2 is a component of the GINS complex, which comprises four proteins, PSF-1, PSF-2, PSF- 3 and SLD-5 27. To determine whether the Ced phenotype detected in psf-2(t3443ts) animals can be attributed to the loss of GINS subcomplex function rather than the specific reduction of psf-2 GINS2, we used RNAi to knock-down expression of the gene psf-3 and found that 62% of the 1st wave cell deaths derived from AB are blocked ( Fig. 2B , psf-3(RNAi)) (see Materials and

for the exact timing and duration of the RNAi knock downs). .CC-BY-NC-ND 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 13, 2024. ; https://doi.org/10.1101/2024.05.09.593335doi: bioRxiv preprint 11 The assembly of the replicative helicase CMG at replication origins occurs in a stepwise process28,29. After mitosis, a hexameric ring of the ATPases MCM2-7 is assembled around double-strand DNA at replication origins where it forms part of the pre-replication complex. CMG assembly is completed in S phase when CDC45 and the preformed GINS complex are recruited to MCM2-7 rings 28,29. The loss of psf-2 or psf-3 is predicted to disrupt the formation of the GINS complex and the assembly of CMG during S phase but not the assembly of MCM2-7 rings at replication origins after mitosis. To determine whether the Ced phenotype detected in psf-2(t3443ts) and psf-3 (RNAi) animals can be attributed to a more general loss of CMG function, we knocked-down the genes mcm-2 and mcm-7 34 and found that 55% and 22% of the 1st wave cell deaths derived from AB are blocked, respectively ( Fig. 2B , mcm-7(RNAi), mcm- 2(RNAi)). These results demonstrate that several members of the CMG complex are required for the cell death fate and suggest that the Ced phenotype is the result of the inability to assemble CMG in S phase. In psf-2(t3443ts) animals the increase in cell cycle lengths in mother cells does not correlate with the block of daughter cell death Reducing psf-2 GINS2 function causes two phenotypes, an Emb phenotype, which we have shown above is the result of increased cell cycle lengths (and ultimately a block in cell division) likely caused by replication defects, and a Ced phenotype. To address whether there is a causal relationship between these two phenotypes, we measured the cell cycle lengths of the mothers of the 1 st wave cell deaths derived from the AB lineage in the wild-type and psf-2(t3443ts) embryos that we had analyzed for a block in cell death ( Fig. 3A; and see data in Fig. 2A, B ). We found that in wild type, in which 0% of the cell deaths are blocked, the average cell cycle length of .CC-BY-NC-ND 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 13, 2024. ; https://doi.org/10.1101/2024.05.09.593335doi: bioRxiv preprint 12 mothers is 42 min (Fig. 3B ). In psf-2(t3443 ts) animals, in which 62% of the cell deaths are blocked, there are two groups of mother cells: (1) mothers, whose daughters die (‘cell death’) (38% of the mothers) and (2) mothers, whose daughters fail to die (‘cell death blocked’) (62% of the mothers). As seen in Figure 3B (psf-2(t3443ts)), in both groups of mothers, there is a broad range of cell cycle lengths, reflecting the increased cell cycle length caused by psf-2(t3443ts). The average cell cycle lengths of the two groups are 108 min and 101 min, respectively, which is significantly different from the average cell cycle length observed in wild type (+/+); however, there is no significant difference between the average cell cycle lengths of these two groups (Fig. 3B). Furthermore, in both groups (i.e. regardless of whether their daughters died or not), there are mothers with cell cycle lengths close to the average cell cycle length of wild-type mothers (42 min) and there are mothers with cell cycle lengths almost four times the average cell cycle length of wild-type mothers (168 min). The independence of the daughter cell fate from the cell cycle length of the mother cell is furthermore exemplified in Figure 3C, which depicts 4D lineaging data of two of the 1 st wave cell deaths, ABalaapapa (referred to as ‘CD1’) and ABalaappaa (referred to as ‘CD2’). In the wild-type embryo (+/+), both cells died within 20-30 min after the completion of their mothers’ divisions, and the cell cycle length of their mothers was 42 min and 44 min, respectively. In the psf-2(t3443ts) embryo, the cell cycle length of the two mother cells was increased to 97 min (mother of CD1) and 107 min (mother of CD2), respectively. CD1 failed to die (indicated in red) whereas CD2 died 20-30 min after the completion of its mother’s division. (Of note, the sister of CD1 divided like in wild type, but the sister of CD2, failed to divide (Fig. 3C)). In summary, in psf-2(t3443ts) animals, there is no correlation between the cell cycle length of a mother cell and the ability of its da ughter to die. This suggests that the roles of psf-2 .CC-BY-NC-ND 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 13, 2024. ; https://doi.org/10.1101/2024.05.09.593335doi: bioRxiv preprint 13 in DNA replication (the likely cause of the increased cell cycle length) and in the acquisition of the cell death fate represent two independent activities of psf-2 GINS2. Our finding that the Emb phenotype and the Ced phenotype observed in psf-2(t3443 ts) animals differ in penetrance and expressivity (see Fig. 1 and 2) furthermore supports this notion. Reducing tyms-1 TYMS function causes increased cell cycle lengths and embryonic lethality, but no block in the cell death fate The lack of correlation between the increased cell cycle length phenotype and the Ced phenotype observed in psf-2(t3443ts) mutants suggests that the role of PSF-2 GINS2 in the acquisition of the cell death fate is independent from its role in DNA unwinding i.e. DNA replication per se. To explore this notion further we sought to abrogate DNA replication and increase cell cycle length more directly. Like psf-2(t3443ts) embryos, at the non-permissive temperature (25°C), embryos homozygous for the temperature-sensitive mutation e2300ts initiate morphogenesis but then arrest 35 (Suppl Fig. 4A). At 25°C, e2300ts embryos exhibit a fully penetrant Emb phenotype and a reduced brood size; however, at the permissive temperature (15°C), e2300ts animals are essentially indistinguishable from wild type ( Suppl Fig. 4C, D ). We found that e2300ts is a partial loss-of-function mutation of the gene tyms-1, which encodes an ortholog of human thymidylate synthase TYMS 36 (see Material and Methods). Specifically, e2300 ts is a missense mutation that causes a guanine to thymine change at position 240 (G240T) of the coding region of the tyms-1 gene. This results in a predicted tryptophan to cysteine change at position 80 of the amino acid sequence of the TYMS-1 protein (W80C), which is a residue that is conserved between C. elegans , mouse and human ( Suppl Fig. 4E ). Thymidylate synthase is the sole enzyme capable of de novo synthesis of thymidine nucleotide precursors, and its inactivation for .CC-BY-NC-ND 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 13, 2024. ; https://doi.org/10.1101/2024.05.09.593335doi: bioRxiv preprint 14 instance through inhibitors such as 5-fluorouracil (5-FU) leads to a block in DNA replication and a block in cell division 37,38. As shown in Figure 1C, D , 4D lineaging analyses revealed that tyms-1(e2300ts) embryos exhibit increased cell cycle lengths in the five consecutive rounds of cell divisions starting from ABarp that give rise to ABarpppppp, very similar to what we observed for psf-2(t3443ts) embryos. Likewise, tyms-1(e2300ts ) animals exhibit increased cell cycle lengths in the six consecutive rounds of cell divisions starting from MS that give rise to MSpppppp (Suppl. Fig. 3A, B). Next, we analyzed 1 st wave cell deaths derived from AB in three tyms-1(e2300ts) animals and found that in contrast to psf-2(t3443ts ) animals, 0% of the cell deaths are blocked ( Fig. 2B, tyms-1 (e2300ts)). We also determined the cell cycle lengths of the mother cells of these 1 st wave cell deaths and found that the average cell cycle length was increased to 125 min, which is more than three times longer than the average cell cycle length of these mothers in wild type ( Fig. 3B). To give an example, Figure 3C depicts 4D lineaging data for CD1 and CD2 in one tyms-1(e2300ts) embryo. The cell cycle length of the CD1 and CD2 mother cells in this embryo was 170 min and 175 min, respectively, but CD1 and CD2 both died. In summary, in line with the importance of thymidylate synthase for DNA synthesis and, hence, DNA replication, reducing tyms-1 TYMS function causes increased cell cycle lengths, ultimately resulting in embryonic lethality. However, in contrast to the reduction of psf-2 GINS2 function, reducing tyms-1 TYMS function does not cause a block in the cell death fate. Together with the results presented in Figure 3C , these results demonstrate that increasing cell cycle lengths in mother cells (presumably through either compromising DNA unwinding or DNA synthesis) is not sufficient to block the deaths of daughters that die. Therefore, the role of psf-2 GINS2 in the acquisition of the cell death fate is independent of its role in DNA unwinding i.e. DNA replication per se. .CC-BY-NC-ND 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 13, 2024. ; https://doi.org/10.1101/2024.05.09.593335doi: bioRxiv preprint 15 Reducing psf-2 GINS2 function abolishes the increase in egl-1 BH3-only mRNA observed in daughters that die Most of the cell deaths that occur during C. elegans development (including all cells that die during the 1 st wave of cell death) are apoptotic cell deaths and dependent on the gene egl-1, which encodes a BH3-only protein 18,19. In contrast to the genes that act downstream of egl-1 in the apoptosis pathway (i.e. ced-9 Bcl-2, ced-4 Apaf-1, ced-3 caspase) and that are ubiquitously expressed at least during embryogenesis 39-41, egl-1 expression is mainly restricted to cell death lineages18,23. Using smRNA FISH, we previously showed that within a cell death lineage, a low concentration of egl-1 mRNA is detected in the ‘mother’ cell. The mother cell divides asymmetrically by both fate and size. The resulting smaller daughter cell dies, whereas the larger one survives. (Daughter cell size ratios differ between cell death lineages and range from about 5.0 to 1.5.) Immediately after mother cell division, egl-1 mRNA concentrations in the two daughters are similar to egl-1 mRNA concentration in the mother cell. Within a few minutes, however, egl-1 mRNA concentration decreases to undetectable levels in the daughter that survives but increases several-fold in the daughter that dies 23. Based on egl-1 transcriptional reporters, the increase in egl-1 mRNA concentration observed by smRNA FISH in the daughter that dies is most probably the result of an increase in egl-1 transcription21,25. To determine the impact of reducing psf-2 GINS2 function on egl-1 expression, we analyzed egl- 1 mRNA levels in the MSpaap lineage using smRNA FISH. The MSpaap mother cell divides to give rise to MSpaapa, which survives, and MSpaapp, which dies (referred to as MSpaapp(X)). As described above, we found that in psf-2(t3443ts) animals, 67% of MSpaapp(X) deaths are .CC-BY-NC-ND 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 13, 2024. ; https://doi.org/10.1101/2024.05.09.593335doi: bioRxiv preprint 16 blocked (Fig. 2C). In wild type animals, we found on average 14.1 and 10.7 egl-1 mRNAs in MSpaap and MSpaapp(X), respectively ( Fig. 4A, B, +/+, egl-1). (O note, the estimated cell volumes of MSpaap and MSpaapp(X) are 113 μ m3 and 22 μ m3, respectively (Sherrard et al, 2017). MSpaap is therefore about 5-times the volume of MSpaapp(X). To assess the temporal dynamics of egl-1 mRNA concentration in MSpaap and MSpaapp(X), we determined the developmental stage of the embryos analyzed based on the number of nuclei in the embryo (see

and Methods). As seen in Figure 4C (+/+), we found that in MSpaap, egl-1 mRNA concentration between embryonic nuclei stages 170 and 177 is 0.1 to 0.2 mRNAs/ μ m3. In MSpaapp(X), egl-1 mRNA concentration increases from about 0.2 mRNAs/ μ m3 at embryonic nuclei stage 182 to about 0.6 mRNAs/ μ m3 at embryonic nuclei stage 197. In psf-2(t3443ts ) animals, we found on average 11.6 egl-1 mRNAs in MSpaap, which is similar to what we detected in MSpaap in wild type (Fig. 4A, psf-2(t3443ts), egl-1 mRNA). However, in contrast to wild type, we only found on average 3.0 egl-1 mRNAs in MSpaapp(X). Indeed, in eight out of 11 embryos (about 73%), essentially no egl-1 mRNA was detectable in MSpaapp(X) ( Fig. 4B, psf-2(t3443ts), egl-1). (In the case of these eight embryos, we measured egl-1 mRNA in four cells at the position where MSpaapp(X) is usually located, determined the average egl-1 mRNA copy number of these four cells and used this as a value for MSpaapp(X).) In addition, we found that compared to wild type, egl-1 mRNA concentration and dynamics in MSpaap are largely unchanged in psf-2(t3443ts) animals (0.1 mRNAs/ μ m3) (Fig. 4C, psf-2 (t3443ts)); however, the burst of egl-1 mRNA concentration observed in MSpaapp(X) between embryonic nuclei stages 182 and 197 is essentially absent in psf-2(t3443ts) animals. Of note, we did not detect changes indicative of ectopic expression in the pattern of egl-1 mRNA in psf-2 (t3443ts) embryos, suggesting that the loss of psf-2 does not cause general mis-regulation of egl-1 expression (see .CC-BY-NC-ND 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 13, 2024. ; https://doi.org/10.1101/2024.05.09.593335doi: bioRxiv preprint 17 Fig. 4A, B, psf-2(t3443ts), egl-1). As a control, and to determine whether reducing psf-2 function affects the expression of genes other than egl-1, we analyzed ced-3 mRNA numbers and concentrations in the same cells through double-labelling. As seen in Figure 4A and B ( ced-3 mRNA), there is no significant difference between wild-type and psf-2(t3443ts) animals in the average ced-3 mRNA numbers in MSpaap (8.4 and 6.8) or MSpaapp(X) (2.5 and 3.8). In addition, between embryonic nuclei stages 170 to 200, ced-3 mRNA concentrations in MSpaap and MSpaapp(X) are relatively stable in both wild-type and psf-2(t3443ts) animals (Fig. 4D). In summary, reducing psf-2 GINS2 function abolishes the increase in egl-1 BH3-only mRNA observed immediately after MSpaap division in MSpaapp(X), but does not affect ced-3 caspase mRNA levels in either MSpaap or MSpaapp(X). These results suggest that in cell death lineages, psf-2 is specifically required for the increase in egl-1 mRNA levels in daughters that die. Thus, our results suggest that psf-2 GINS2 is required for cell fate divergence in cell death lineages. The role of psf-2 GINS2 in cell fate divergence is not restricted to cell death lineages To determine whether the role of psf-2 GINS2 in cell fate divergence is restricted to asymmetric cell divisions that generate a daughter that dies, we analyzed the ABaraappaa lineage. ABaraappaa divides during the 10 th round of cell division and generates an anterior daughter, ABaraappaaa, which differentiates into the pharyngeal motor neuron/interneuron MI, and a posterior daughter cell, ABaraappaap, which differentiates into the pharyngeal marginal cell m1DR (MI/m1DR decision) 9. A cold-sensitive (cs) gain-of-function (gf) mutation of the gene his-9, n5357 gf, one of 14 genes in the C. elegans genome that encode replication-dependent histone H3, has been proposed to interfere with H3-H3 interactions and, hence the formation of .CC-BY-NC-ND 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 13, 2024. ; https://doi.org/10.1101/2024.05.09.593335doi: bioRxiv preprint 18 H3-H4 tetramers and nucleosome assembly 42. Importantly, it was previously shown that in his- 9(n5357gf) animals, instead of differentiating into the MI neuron, ABaraappaaa adopts the fate of the pharyngeal epithelial cell e3D42. Using a MI-specific reporter (P sams-5gfp)43, we found that 100% of wild-type animals have one MI neuron, but only 51% of his-9(n5357gf) animals raised at non-permissive temperature (15°C) have an MI neuron 42 ( Fig. 5A, his-9(n5357gf)). To determine whether reducing psf-2 function affects the MI fate, we analyzed psf-2(t3443ts ) animals at the non-permissive temperature. We found that 89% of psf-2(t3443ts) animals have one MI neuron and 5% have no MI neuron ( Fig. 5A, B, psf-2(t3443ts)). Interestingly, based on reporter expression, 6% of the animals had an additional MI neuron (‘ectopic MI’). Therefore, we conclude that reducing psf-2 GINS2 function impacts the acquisition of the MI fate in the ABaraappaa lineage. Next, we analyzed the AMso cells, are a pair of bilaterally symmetrical glial cells that are born in the embryo. In males, each of the AMso cells divides asymmetrically during post-embryonic development to generate a glial cell (ABpl/rpaapapaa) and a MCM neuron (ABpl/rpaapapap) 44. We analyzed psf-2(t3443ts ) animals using the glia-subtype reporter P lin-48gfp45 and the panneuronal reporter P rab-3NLS::rfp46. As expected, given the role of psf-2 in DNA replication, the AMso division is often blocked in psf-2(t3443ts) animals raised at non-permissive temperature (Fig. 5E and F). Importantly, we found that in the cases where the AMso did divide, there was no rab-3 expression in 82% of the cases (167/204 AMso divisions). This lack of neuronal differentiation accounts for 68% of the total population of cells ( Fig. 5C right panel, psf-2(t3443ts) 25°C). These rab-3- lacking cells retained lin-48 expression into adulthood ( Fig. 5E and F ‘no neuron’), which was never observed in wild type (+/+) controls at the non- permissive temperature. In 3% of the cases, we also observed rab-3 expression in both the .CC-BY-NC-ND 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 13, 2024. ; https://doi.org/10.1101/2024.05.09.593335doi: bioRxiv preprint 19 anterior and posterior cells ( Fig. 5D ‘2 neurons’). Interestingly, at the non-permissive temperature, 4% of the cells divide ectopically to produce an extra cell (7/8 ectopic cells express the neuronal marker rab-3). Finally, all defects observed were rescued by psf-2(+). These data suggest that psf-2 GINS2 is required for the acquisition of the MCM fate in the AMso lineage in males. Based on these analyses we conclude that the role of psf-2 GINS2 in cell fate divergence is not restricted to cell death lineages. .CC-BY-NC-ND 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 13, 2024. ; https://doi.org/10.1101/2024.05.09.593335doi: bioRxiv preprint 20

Genetic studies of programmed cell death in C. elegans have been instrumental in the elucidation of the conserved apoptosis pathway 18,19. Continuing in this tradition, we have discovered a new factor required for cell death during C. elegans development, CMG (C dc45-MCM2-7-GINS). Since the discovery of the BH3-only protein EGL-1, this is the first new C. elegans factor identified that when inactivated causes a general block in cell death. CMG is also the first essential factor identified that is required for cell death during C. elegans development. Importantly, we demonstrate that CMG’s role in the acquisition of cell fate is not specific to the cell death fate. As outlined below, our genetic studies of programmed cell death have uncovered a CMG-mediated mechanism that is fundamentally important for C. elegans embryonic and post- embryonic development and that is likely conserved in higher organisms. CMG has a conserved role in the divergence of cell fates in the context of asymmetric cell division How daughter cells acquire different fates during asymmetric cell division is incompletely understood 1-4. Here we demonstrate that CMG, a core complex of the eukaryotic replisome, is required for the asymmetric acquisition of at least three different cell fates in C. elegans (cell death fate, MI fate, MCM fate), suggesting a widespread role for CMG in this process. The groups of Zhiguo Zhang, Anja Groth and Haiyun Gan have recently independently reported that Mcm2, a component of CMG, or POLE-3/POLE-4, a subcomplex of the leading strand polymerase Pol ε , are required for the ability of mouse embryonic stem cells (ESCs) to transition from a naïve cell state to a differentiated cell state 47-50. Importantly, our results not only confirm their in vitro data but provide the first in vivo evidence for a role of CMG in the divergence of .CC-BY-NC-ND 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 13, 2024. ; https://doi.org/10.1101/2024.05.09.593335doi: bioRxiv preprint 21 cell fates in the context of asymmetric cell division. Whether the depletion of POLE-3/POLE-4 also affects cell fate divergence in C. elegans remains to be determined. Our results also provide support for the quantal cell cycle theory, which proposes that genome alterations that occur during cell division enable the expression of distinct sets of genes in the daughter cells 5,6. Although coupled to replication, the function of Mcm2 in the differentiation of mouse ESCs in vitro is independent of its helicase activity and hence, independent of its well described role in DNA unwinding48-50. Similarly, we present in vivo evidence in support of the notion that the role of C. elegans CMG in cell fate divergence is independent of its role in DNA unwinding. Specifically, (1) increases in cell cycle length in psf-2(t3443ts) mutants (likely reflecting replication stress) do not correlate with the loss of the cell death fate and (2) increases in cell cycle length in tyms-1(e2300ts) mutants (also likely reflecting replication stress) do not lead to the loss of the cell death fate. Together, these findings suggest that in the context of asymmetric cell division, CMG has a conserved role in the divergence of cell fates that is distinct from its conserved role in DNA unwinding. Finally, the idea that CMG has a conserved role in cell fate is supported by previous observations implicating Psf2 GINS2 in nervous system development in Xenopus and zebrafish 51,52, Sld5 GINS4 in early embryonic development and nervous system development in mouse53,54 and MCM5 in neuronal differentiation in Drosophila55. Interestingly, genes encoding components of the pre-replication complex, such as orc-1-5 or mcm-2-7, but not genes encoding components required for the assembly of the functional replicative helicase, such as psf-1, psf-2, psf-3 or sld-5, are required for the expression of pro- invasive genes during anchor cell invasion in C. elegans larvae56. This provides additional evidence that the replisome is involved in regulating gene expression. However, it is important to note that the anchor cell is a non-dividing post-mitotic cell. The role of the pre-replication .CC-BY-NC-ND 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 13, 2024. ; https://doi.org/10.1101/2024.05.09.593335doi: bioRxiv preprint 22 complex in the expression of pro-invasive genes is therefore independent of DNA replication and, hence, independent of cell division. Working model for how C. elegans CMG causes divergence of the cell death fate Apart from their known roles as part of the replicative helicase or the leading strand polymerase Pol ε , Mcm2 and POLE3/POLE4 bind histones and act as histone chaperones during replication57- 63. In this capacity, they transfer parental nucleosomes carrying epigenetic marks on their histone H3 and H4 moieties from the parental chromosome to the lagging or leading strand, respectively63,64. The roughly equal inheritance of parental nucleosomes to leading and lagging strand is thought to be critical for the maintenance of gene expression patterns, and thus cell fate stability, during symmetric or ‘proliferative’ cell divisions 65-67. Indeed, the loss of Mcm2’s histone chaperone activity or the depletion of POLE-3 or POLE-4 in mouse ESCs in vitro has recently been reported by the Growth, Gan and Zhang groups to disrupt the equal inheritance of parental nucleosomes to leading and lagging strands and to result in sister chromatids with alternating, complementary patches of parental or new nucleosomes along the chromosomes 48-50. Unexpectedly, the ability of these mutant ESCs to undergo self-renewing proliferative cell divisions was essentially unaffected; however, as mentioned above, defects were observed in the ability of these cells to acquire a differentiated state. Bivalent genes are considered key regulators during differentiation. They are decorated by nucleosomes that carry both the repressive mark H3K27me3 and the activating mark H3K4me3 and this ‘bivalent’ condition is thought to enable rapid transcriptional activation in a cell- or lineage-specific manner 68,69. Based on transcriptome profiling combined with genome-wide analyses of epigenetic marks, it was proposed that the defects in differentiation observed in the Mcm2 and POLE-3/POLE-4 mutant .CC-BY-NC-ND 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 13, 2024. ; https://doi.org/10.1101/2024.05.09.593335doi: bioRxiv preprint 23 ESCs are in part the result of the over-enrichment of the repressive epigenetic mark H3K27me3 at bivalent genes, which impairs activation of these genes during differentiation48-50. We demonstrate that reducing C. elegans CMG function results in the inability to increase transcription of the cell death fate determinant egl-1 in daughters that die immediately following mother cell division. This establishes the BH3-only gene egl-1 as a target of CMG. Based on the following observations, we propose that CMG impacts egl-1 BH3-only expression at the epigenetic level. First, our results suggest that the cell death defect observed in response to reducing CMG is independent of CMG’s helicase activity. This makes it highly likely that the role of CMG in cell death is dependent on CMG’s recently described histone chaperone activity. Second, the mutation his-9(n5357gf), which interferes with the formation of H3-H4 tetramers,

in the loss of the MI fate 42 (see Figure 5). We show that reducing CMG also results in the loss of the MI fate. This suggests that his-9 and CMG are involved in the same process and that this process is likely to depend on nucleosome assembly. Third, the egl-1 BH3-only locus on LGV spans ~15kb and includes the egl-1 transcription unit as well as extensive cis -acting elements (i.e., enhancers), some of which are located beyond transcription units upstream and downstream of the egl-1 transcription unit18,21 (Suppl. Fig. 5). We have mined publicly available data on genome-wide distributions of nucleosomal histone H3 posttranslational modifications (PTMs) in C. elegans embryos and found that the egl-1 locus is decorated by both nucleosomes carrying repressive H3K27me3 marks and nucleosomes carrying activating H3K4me3 marks (bulk embryo data representative of non-cell death lineages) 70,71 (Suppl. Fig. 5). It remains to be ascertained whether H3K27me3 and H3K4me3 are present on the same nucleosomes within a single cell; however, considering that egl-1 BH3-only is a key developmental regulator that is expressed in a highly dynamic and spatiotemporally restricted manner, this makes it likely that .CC-BY-NC-ND 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 13, 2024. ; https://doi.org/10.1101/2024.05.09.593335doi: bioRxiv preprint 24 egl-1 BH3-only represents a bivalent gene whose expression during differentiation is affected when the equal inheritance of parental nucleosomes is disrupted during replication. (In addition, in bulk embryo data representative of non-cell death lineages, the egl-1 locus is devoid of nucleosomes carrying the activating H3K36me3 mark but is decorated with nucleosomes carrying the mark H3K4me1 70. The egl-1 locus can therefore be considered ‘poised’ for transcriptional activation (Suppl. Fig. 5).) Furthermore, based on the following observations, we propose that CMG acts during mother cell replication to enable epigenetic changes at the egl-1 BH3-only locus. First, Nakano et al demonstrated that the mutation his-9(n5357gf) acts in the MI mother cell to cause loss of the MI fate 42. Second, we demonstrate that a shift - in psf-2(t3443ts) animals - of the AMso glial cell to the non-permissive temperature (25°C) is sufficient to cause AMso to divide symmetrically and generate two AMso-like cells rather than divide asymmetrically and generate one AMso glial cell and one MCM neuron (Fig. 5). Based on these observations, we propose that CMG enables epigenetic changes at the egl- 1 BH3-only locus during mother cell replication. These changes are mediated by CMG’s histone chaperone activity and, hence, its role in replication-coupled nucleosome assembly, and they are required for the increase in egl-1 expression in daughters that die. The nature of the relevant epigenetic changes remains to be determined but may include the removal of the repressive mark H3K27me3. According to this working model, reducing CMG function would prevent these epigenetic changes and thus eliminate the increase in egl-1 expression in daughters that die. The development of methods for in vivo visualization of the chromatin state at the egl-1 BH3-only locus in specific cell death lineages in the developing embryo will be necessary to test our model 72, and this is one of our prime goals in the future. .CC-BY-NC-ND 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 13, 2024. ; https://doi.org/10.1101/2024.05.09.593335doi: bioRxiv preprint 25 How is symmetry of egl-1 BH3-only expression broken during asymmetric cell divisions? The working model described above explains how egl-1 BH3-only transcription can be activated in daughter cells. What remains unclear is how asymmetry of egl-1 transcriptional activation is created to cause egl-1 expression in one, but not the other daughter cell. As mentioned above, the non-random segregation of direct repressors of egl-1 transcription into the daughter that survives may contribute to the repression of egl-1 expression in these cells 24,25. Therefore, one possible model is that CMG-dependent epigenetic changes that enable transcriptional activation at the egl-1 locus occur on both sister chromatids. In this scenario, both daughter cells would inherit two egl-1 loci that are competent for transcriptional activation, but transcriptional activation would be blocked in the daughter that survives, because of the presence of a direct repressor of egl-1 transcription. In this scenario, symmetry breakage would be caused by the non-random segregation of trans -acting factors, e.g., transcriptional repressors. Alternatively, symmetry breakage could occur at the level of the egl-1 locus itself. Specifically, the CMG-dependent epigenetic changes at the egl-1 locus could occur on only one of the two emerging sister chromatids and the two egl-1 loci competent for transcriptional activation could be inherited specifically by the daughter that dies. How CMG-dependent epigenetic changes at the egl-1 locus could occur on only one of the two emerging sister chromatids is currently unknown. Interestingly, there is some evidence that epigenetically distinct sister chromatids can be generated and non-randomly segregated during asymmetric or ‘informative’ cell divisions in stem cell lineages. Specifically, the group of Xin Chen has reported that in asymmetrically dividing D. melanogaster germline and intestinal stem cells and mouse ESCs, distinct sister chromatids with either parental nucleosomes or new nucleosomes are non-randomly segregated into the self-renewing stem cell daughter or differentiating daughter, respectively 73-75. .CC-BY-NC-ND 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 13, 2024. ; https://doi.org/10.1101/2024.05.09.593335doi: bioRxiv preprint 26 Furthermore, Zhiguo Zhang’s group has recently shown that parental nucleosomes carrying the repressive mark H3K9me3 are preferentially segregated to the leading strand during the replication of LINE1 transposable elements and that this non-random segregation of nucleosomes is dependent on the interaction of the Human Silencing Hub (HUSH) complex with POLE-3/POLE-4 76. Additionally, scenarios can be envisioned in which both trans-acting factors and epigenetically distinct egl-1 loci could be non-randomly segregated. Finally, regardless of the specific scenario, they all require that the mother cell becomes polarized, which is expected to be dependent on extrinsic factors such as intercellular signals. Control of gene expression in multilineage priming In multilineage priming, genes that determine the fates of the two daughter cells are co-expressed in the mother cell, but after mother cell division, their expression becomes restricted to one or the other daughter cell 10. Our results uncover an intrinsic mechanism through which the expression of a C. elegans cell fate determinant that is subject to multi-lineage priming is maintained and increased in one of the daughter cells. As outlined above, we propose that CMG-dependent epigenetic changes at the egl-1 BH3-only locus during replication in the mother cell enable increased expression of egl-1 in the daughter that dies. Whether CMG is required for changes in the expression of other C. elegans genes that exhibit multilineage priming remains to be determined10. It also remains to be determined whether CMG is required for the expression of genes that are subject to multilineage priming in other organisms 11-14,16. Our working model proposes that CMG controls egl-1 expression at the epigenetic level in most if not all cell death lineages i.e. globally. egl-1 expression has previously been shown to be controlled at both the transcriptional and post-transcriptional level. Briefly, lineage-specific transcription factors act .CC-BY-NC-ND 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 13, 2024. ; https://doi.org/10.1101/2024.05.09.593335doi: bioRxiv preprint 27 through specific cis-acting elements in the egl-1 locus to control egl-1 transcription in specific cell death lineages (see Suppl. Fig. 5) and the loss of a lineage-specific transcription factor or the loss of the cis -acting element through which it acts results in the block of one or a few specific cell deaths (in the case of transcriptional activators) or the ectopic death of specific cells that normally survive (in the case of transcriptional repressors)18,77-79. At the post-transcriptional level, members of the miR-35 and miR-58 families of microRNAs act through the 3´UTR of egl-1 mRNAs to globally ensure that the level of egl-1 expression in mothers of cells that die does not reach the level necessary to trigger cell death, and the loss of miR-35 and miR-58 microRNAs causes mother cells to die precociously 23,80. How control of egl-1 expression at the epigenetic, transcriptional and post-transcriptional levels is coordinated within a particular cell death lineage to reproducibly results in the highly dynamic and spatiotemporally restricted pattern of expression observed remains to be investigated. It also remains to be investigated whether such complex control of gene expression at multiple levels underpins the expression of other C. elegans genes that are subject to multilineage priming. Multilineage priming is a phenomenon also observed during mammalian development and is well characterized during lineage commitment in the hematopoietic system 13,14,17. Inborn errors of immunity (IEI) are rare genetic conditions that are characterized by the absence or dysfunction of specific types of immune cells 81, such as for example Natural Killer (NK) cells 82- 84. Importantly, of the six genes that have so far been identified to mutate and cause NK cell deficiency (NKD), three encode components of CMG, Psf1 GINS1, Sld5 GINS4 and MCM4, and one encodes an auxiliary component of CMG, MCM10 85. Interestingly, in the case of the Sld5 GINS4 mutations, compared to wild type, no significant changes in cell cycle profiles or DNA damage were detected, suggesting that NKD is caused by the loss of a role of CMG that is .CC-BY-NC-ND 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 13, 2024. ; https://doi.org/10.1101/2024.05.09.593335doi: bioRxiv preprint 28 independent of its role in DNA replication per se 86. Based on our findings in C. elegans, we speculate that a reduction in Sld5 GINS4 results in the deregulation of the expression of determinants of the NK fate that are subject to multilineage priming, resulting in a block in cell fate divergence and, hence, the absence of NK cells. .CC-BY-NC-ND 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 13, 2024. ; https://doi.org/10.1101/2024.05.09.593335doi: bioRxiv preprint 29

General C. elegans maintenance and strains C. elegans strains were cultured and maintained as described previously 87. The Bristol N2 strain was used as wild-type strain, and the following transgenes and alleles were used in this study: LGI: tyms-1(e2300ts), psf-2(t3443ts) (this study); LGII: his-9(n5357gf) 42; LGIII: saIs14 (Plin- 48gfp)45; LGIV: ced-3 (n717)32; LGV: otIs356 ( Prab-3NLS::rfp)46; him-5(e1490)88; nIs396 (Psams- 5gfp)43. In addition, the following multicopy transgenes and extrachromosomal arrays were used: bcEx1302 (psf-2(+)) (this study), bcEx1306 (Ppsf-2psf-2::gfp::psf-2 3´UTR) (this study), bcIs133 (Ptoe-2gfp)33. Throughout our studies, we used information and tools available on WormBase (https://wormbase.org/#012-34-5)89,90. Extrachromosomal arrays generated bcEx1302 was generated by microinjection of a 3.6 kb genomic PCR fragment amplified from N2 Bristol using the primer psf-2_for 5’-ataaaagcgacaacgattgc-3’ and psf-2_rev 5’- aattcctttacgacttgcga-3’. 10ng/ul of the PCR product was injected along with 100ng/ul pRF4 into psf-2(t3443ts) mutants. Animals were incubated at 15°C until L4 rollers were visible. Rollers were selected and shifted to 25°C. Lines that grew at 25°C were considered as rescue. bcEx1306 was generated by microinjection of the plasmid pBC1695 (P psf-2psf-2::gfp::psf-2 3´UTR). pBC1695 was injected (10ng/ul) along with pRF4 (100ng/ul) into psf-2(t3443ts) mutants. Animals were incubated at 15°C until L4 rollers were visible. Rollers were selected and shifted to 25°C. Lines growing at 25°C were considered as rescue. EMS mutagenesis .CC-BY-NC-ND 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 13, 2024. ; https://doi.org/10.1101/2024.05.09.593335doi: bioRxiv preprint 30 A screen for temperature-sensitive embryonic lethal mutants using Ethyl methane sulfonate (EMS) was conducted in the laboratory of Ralf Schnabel (TU Braunschweig, Germany) using the following protocol 87. P0s were mutagenized with 50mM EMS at room temperature for 4h, distributed among large (90 mm) plates and incubated overnight at 15°C for recovery. After 24h, mutagenized P0s were picked onto large plates (25-30 worms/plate) and incubated for 7 days at 15°C. The F1 generation was picked onto large plates (25–30 worms each) and incubated for 7 days at 15°C. We estimated to get ~2,500–3,000 F2 worms per plate without running out of food. L4 stage animals were selected from these F2 populations and singled into 96-well plates using a worm sorter (COPAS, Union Biometrica). We performed four independent screens (NC, ND, NE and NF) and singled a total of ~460.000 L4 larvae. 96-well plates were incubated at the permissive temperature (15°C) for 7 days and then replicated with a Biomek FX (Beckman Coulter). These replica plates were incubated at the non-permissive temperature (25°C) for 7-10 days, after which they were analyzed for lethality by eye. Compared to wells with viable animals, wells with non-viable animals still contained food. In addition, a lot of small larvae were present in these wells. Positive clones were retested manually for embryonic lethality using an 8-channel pipette for replica plating. Before phenotypic analyses were performed, positive candidates were re-tested for temperature sensitivity a third time. psf-2 cloning and temperature shift experiments with psf-2(t3443ts) The mutant psf-2(t3443ts) was a mutant isolated in the above-described screen (NE). Using snip- SNP mapping 26, we mapped the Emb phenotype of t3443ts to LGI close to the variation pkP1071 at position 23.40 cM. The whole genome sequencing was performed in Don Moermann’s laboratory. The rescue experiments shown in Figures 1 and 2 were performed using a 3778bp .CC-BY-NC-ND 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 13, 2024. ; https://doi.org/10.1101/2024.05.09.593335doi: bioRxiv preprint 31 PCR fragment amplified from C. elegans N2 strain using the primers 5’-ataaaagcgacaacgattgc-3’ and 5’-aattcctttacgacttgcga-3’ ( bcEx1302 array; Suppl Fig. 1A ). The rescue experiment shown in Figure 5 was performed using a smaller PCR fragment of 1563 bp (Suppl Fig. 1A) amplified using the primers 5’- GAACAGAACGATGAGCAATAC -3’ and 5’- TGGAACGTTCAACAAGTCAT-3’ (bcEx1306 array). This smaller fragment rescues both the Emb and Ced phenotype. The smaller fragment was used to generate the plasmid pBC1695 (see General C. elegans maintenance and strains). Unless stated otherwise, for the analysis of the phenotype, L4 larvae were shifted to 25°C for ~16h before the then adult animals were dissected and embryos extracted for analysis. tyms-1 cloning and temperature shift experiments with tyms-1(e2300ts) For our analyses, we used animals homozygous for the mutation e2300ts, which the Schnabel lab had previously described, and which had defined the gene ‘ cib-1’ 35. We (HS and RS) have since found that e2300ts is a mutation in the thymidylate synthetase gene tyms-1. For this reason, cib-1 was renamed ‘ tyms-1’. At the non-permissive temperature, embryos homozygous for tyms- 1(e2300ts) presumably run out of thymidine 37,38. This is expected to lead to a block in DNA synthesis, replication stress and general replication failure, resulting in a defect in cell division and embryonic arrest. For example, shifting tyms-1(e2300ts) L4 larvae to 25°C results in the F1 embryos to arrest at the ~50-cell stage ( Suppl. Fig. 4B )35. To overcome this problem, for our analyses, we shifted embryos at the one- or two-cell stage to 25°C (rather than L4 larvae as we did for psf-2(t3443ts)). Using this temperature shift regime, compared to wild type, the cell cycle lengths of tyms-1(e2300ts) mutants progressively increase with every round of cell division similar to what we found in the case of psf-2(t3443ts) (see Fig 1 and Suppl Fig. 3). .CC-BY-NC-ND 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 13, 2024. ; https://doi.org/10.1101/2024.05.09.593335doi: bioRxiv preprint 32 Plasmid construction To generate pBC1695 three fragments were amplified using N2 wild-type lysate as DNA template. The first 1331bp long fragment was amplified using the primers 5’- gaacagaacgatgagcaatac-3’ and 5’-ctcctttactcattaaaggtgttgat-3’ and included the upstream regulatory regions and psf-2 Exon 1 and 2 with overhangs for egfp. The second fragment was 894bp long and included egfp with overhangs to the upstream and downstream fragments. It was amplified using the primers 5’ atcaacacctttaatgagtaaaggag-3’ and 5’-ggaataaaacactatttgtatagttc- 3’. The last fragment included the psf-2 3’UTR and the downstream regulatory region and was amplified using the primers 5’ gaactatacaaatagtgttttattcc-3’ and 5’-tggaacgttcaacaagtcat-3’ and had overhangs to egfp. All fragments were stitched via PCR stitching and cloned blunt end into the EcoRV site of pBluescript II KS. The resulting plasmid pBC1695 was sequenced for verification and injected (10ng/ul) along with pRF4 (100ng/ul) into psf-2(t3443ts) mutants. Animals were incubated at 15°C until L4 rollers were visible. Rollers were selected and shifted to 25°C. Lines growing at 25°C were considered as rescue. The psf-2 RNAi clone was generated from the psf-2(RNAi) plasmid from the Vidal library. The psf-2 cDNA fragment from the Vidal plasmid was subcloned into pBluescriptII KS(+) using EcoRV and SpeI site to generate pBC1720. All other plasmids used for RNAi were generated by amplification of cDNA fragments for each gene and were subcloned blunt end into pBluescriptII KS(+) using EcoRV site. For the psf-3 RNAi clone a 582 bp fragment was amplified from cDNA using psf- 3_cDNA_f 5’-atggctggatttgaaattc-3’ and psf-3_cDNA_r 5’-ttataacgaaagtctcttac-3’. For the mcm- 2 RNAi clone a 573 bp fragment was amplified from cDNA using mcm-2_RNAi_for 5’- gggagtagaatggatacatgttc-3’ and mcm-2_RNAi_rev 5’-cggagctcgatcagtactc-3’. For the mcm-7 .CC-BY-NC-ND 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 13, 2024. ; https://doi.org/10.1101/2024.05.09.593335doi: bioRxiv preprint 33 RNAi clone a 634 fragment was amplified from exon 2 using mcm-7_Exon-2_for 5’- gacaagcaggcaatcgttg-3’ and mcm-7_Exon-2_rev 5’-ctactgggacttgctcgc-3’. RNA Interference For RNAi experiments by microinjections 91, the following plasmids were used as PCR templates: pBC1720 ( psf-2(RNAi)), pBC1721 ( psf-3(RNAi)), pBC1719 ( mcm-7(RNAi)), pBC1962 ( mcm-2(RNAi)) and the following primers were used for amplification CMo24 (5’- ttgtaaaacgacggccag-3’) and CMo25 (5’- catgattacgccaagcgc-3’) to generate PCR products, which include at the ends of the PCR product the T7 and T3 promoter. In vitro transcription was performed with Ambion Megascript Kit T3 and T7. RNAi was performed via injection into young adult worms, which were incubated at 25°C 3-24h prior to recordings depending on the RNAi. For injections, Bristol N2 was used as the wild-type strain. RNAi injections were performed 24h prior to recording in the case of psf-2(RNAi) and psf-3(RNAi), 3h prior to recording in the case of mcm-7(RNAi) and 12h prior to recording in the case of mcm-2(RNAi). Analysis of embryonic lethality and brood size For the analysis of brood size and embryonic lethality, L4 larvae were picked and either maintained at the permissive temperature or shifted to the non-permissive temperature. The worms were transferred to fresh small (35 mm) plates with food twice a day until they no longer laid eggs. Shortly after transferring the worms to a fresh plate, the number of eggs laid was counted. After 24-36h, the number of dead eggs was counted and after 24-48h, the number of animals hatched was counted. .CC-BY-NC-ND 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 13, 2024. ; https://doi.org/10.1101/2024.05.09.593335doi: bioRxiv preprint 34 4D lineaging analysis C. elegans L4 animals were grown to the adult stage overnight at 25°C. Two- or four-cell stage embryos were harvested from the young adults and mounted on 4.5% agarose pads for differential interference contrast (DIC) and fluorescence microscopy. 4D recordings were made throughout development as described previously 30,31, using a Zeiss Axio Imager.M2 and Time to Live software (Caenotec), capturing 25 DIC and/or fluorescence z-stacks every 35 sec at 25°C. Lineage analysis of the 4D recordings was performed using SIMI©BioCell software (SIMI Reality Motion Systems, http://www.simi.com )30,31. Cells are followed by the observer and the coordinates are recorded approximately every 2 min. The cell cleavages are assessed by marking the mother cell before the cleavage furrow ingresses. The centers of the daughter cells are marked three frames later (105 /i3 s). Lineage analysis of the ABarpppppp (V6R) lineage and the MSpppppp lineage For the analysis of a specific cell (ABarp or MS) and their descendants the Software Database SIMI©BioCell was used as described above in the section ‘4D lineage analysis’. The cells ABarp and MS are followed until ABarpppppp and MSpppppp are born. These cells are the last cells in the lineage tree and differentiate into the hypodermal cell V6R (ABarpppppp) and a muscle cell (MSpppppp) respectively. For the analyses, the timepoint (min) of each cell division was taken and the difference (min) (referred to as ‘cell cycle length’) between each cell and the cell division of its corresponding mother cell was calculated. Lineage analysis of cells programmed to die .CC-BY-NC-ND 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 13, 2024. ; https://doi.org/10.1101/2024.05.09.593335doi: bioRxiv preprint 35 The 4D lineage analysis was performed as mentioned above. For the analysis of programmed cell death events, we tracked the 13 cell death events derived of the AB lineage and the MSpaapp cell death event. Cells ‘programmed to die’ and their actual fates were followed from their births until – if possible – the engulfment of their cell corpses. If a cell died appropriately, it is indicated as ‘cell death’ (see Fig. 2 ). If a cell had not died before the next round of cell division started in other cells or within a length of time corresponding to twice the cell cycle length of their mothers, it was considered inappropriately surviving and is indicated as ‘cell death blocked’. Cells are indicated as ‘lost cells’, if they did not fulfill either one of the two requirements described above. During our analysis we also encountered two more phenotypes in mothers of cells programmed to die. Cells are marked as ‘cell division of mother blocked’, if a mother failed to divide and cells are marked as ‘mother dies’, if a mother died (precocious cell death). For all cells and their descendants, cell cycles lengths were measured in minutes from birth until the next division. The cell cycle analysis of mothers of programmed cell deaths was performed from the birth of the mother cell until the mother cell divided again. Time until cell corpse formation was measured in minutes from the birth of a cell programmed to die until the formation of the ‘erythrocyte’ stage (stage, at which the distinction between nucleus and cytoplasm is lost by DIC) 92,93. smRNA FISH and image analysis smRNA FISH was performed in C. elegans embryos, as described previously 94. Embryos were harvested by bleaching healthy adults and then permitted to develop in M9 buffer at 25°C until the desired stage was reached (1h15m for MSpaap cell lineage). Stellaris FISH probes (Biosearch Technologies) were designed against the mature mRNAs of egl-1 and ced-3. The egl- .CC-BY-NC-ND 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 13, 2024. ; https://doi.org/10.1101/2024.05.09.593335doi: bioRxiv preprint 36 1 probe set contained 23 TAMRA-labeled oligonucleotides and was used at a working concentration of 250 nM in Hybridization Buffer94. The ced-3 probe set contained 48 Quasar670- labeled oligonucleotides and was used at 500 nM. Image stacks were obtained using Leica LAS AF software and a Leica TCS SP5 II confocal microscope with a 63x oil immersion lens and a z- spacing of 500 nm to capture diffraction-limited mRNA spots over several z-slices. Laser intensity was set to 10% to minimize bleaching. Image analysis was performed using Fiji software 95. The pipeline used to quantify the mRNA copy number in a cell of interest was outlined previously23. Briefly, a three-dimensional region of interest (ROI) was defined for the cell of interest as a subset of cropped z-slices that fully contained the cell. The total smRNA FISH signal intensity (SI Total) contained within this ROI was determined by summing all z-slices and measuring the total signal in the resulting z-projection. Next, background signal was subtracted by determining the total smRNA FISH signal intensity for three neighboring regions of the same size without visible mRNA signal, then subtracting their average signal intensity (SI Bkgd) from that of the ROI. Finally, the mRNA copy number was calculated by dividing the total signal intensity of the ROI by the average intensity of a single diffraction-limited mRNA spot (SISpot), or generally, the mRNA copy number in a cell was calculated as (SI Total - SI Bkgd) / SISpot. For presentation, maximum intensity z-projection images were smoothed (Gaussian blur; radius, 1.5). To calculate cellular concentrations of mRNA in the MSpaap and MSpaapp cells, first the average volumes of these two cells were determined by assuming sphericality and measuring their average diameters from confocal image stacks. Then, the mRNA copy number in each cell of interest was divided by the average volume of that cell. Finally, the total number of embryonic nuclei was counted so that the cell of interest could be mapped to a specific developmental timepoint. This nuclei count was performed using Multiview Reconstruction .CC-BY-NC-ND 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 13, 2024. ; https://doi.org/10.1101/2024.05.09.593335doi: bioRxiv preprint 37 software96 to detect interest points in the DAPI channel, with manual corrections as required. To generate a developmental time-course plot of mRNA concentration, mRNA copy numbers were first ordered chronologically by increasing embryonic nuclei count. Next, a centered running average of order 5 was applied to both embryonic nuclei count and mRNA copy number data, effectively smoothing the resulting plot. Finally, shaded areas representing the SEM of averaged data points were added to the plot. Quantification of QL.pp death QL.pp death was analyzed using the P toe-2gfp (bcIs133) transgene33, which labels cells of the Q lineage. Within the QL.p lineage, L2 larvae of wild-type animals contain two GFP-positive PVM and SDQL neurons, which are the daughters of the surviving sister QL.pa. In cell death mutants, up to two extra GFP positive cells can be seen, which are undead QL.pp cells 33. For quantification, gravid wild-type adults maintained at 25°C were allowed to lay eggs at 25°C for 1h. The adults were removed, and the eggs laid were incubated at 25°C for 24h, until they reached the L2 stage. The animals on the plate were washed off with 2mM levamisole solution in MPEG and collected by centrifugation at 400g for 1min. 5µL of the pelleted worms was then mounted on a 2% agarose pad on a glass slide and an 18x18mm coverslip (#1.5 thickness) was added on top. The number of GFP positive cells was counted using the Zeiss Axio Imager M2 with a 100X/1.3 NA oil-immersion objective lens as previously described 33. Only those worms were assessed where the PVM and SDQL (QL.pa daughter cells) had formed obvious neurite extensions. For quantification of psf-2(t3443ts ) animals at the non-permissive temperature, the strain was maintained at 15°C, and gravid adults were allowed to lay eggs at 25°C for 2h. The adults were removed, and eggs laid were further incubated at 25°C for 28-30h until they reached .CC-BY-NC-ND 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 13, 2024. ; https://doi.org/10.1101/2024.05.09.593335doi: bioRxiv preprint 38 L2 stage. The larvae were mounted and scored as described above for wild-type animals. In the psf-2(t3443ts) background, only one GFP positive cell with neurite extensions was observed in some cases, indicating a block in cell division in QL.p or QL.pa. Analysis of MI fate MI fate was analyzed using the P sams-5gfp (nIs396) transgene, which is expressed exclusively in the MI neuron in wild-type animals 42,43. Wild-type animals (15°C and 25°C), his-9(n5357gf) animals (15°C and 20°C) and psf-2(t3443ts) animals (15°C) were grown at the respective temperatures for at least two generations before assaying the MI fate. L4 larvae were mounted on 2% agarose pads containing 25 mM sodium azide in M9 buffer and an 18x18mm coverslip (#1.5 thickness) was added on top. The number of GFP positive cells was counted in the anterior pharynx using the Zeiss Axio Imager M2 with a 100X/1.3 NA oil-immersion objective lens. For scoring psf-2(t3443ts) animals at the non-permissive temperature of 25°C, the strain was maintained at 15°C and then gravid adults were allowed to lay eggs at 25°C for 6 hours. The adults were removed, and eggs laid were shifted back to permissive temperature of 15°C for 5 days. On the 5th day, L4 larvae were mounted and GFP positive cells were counted in the anterior pharynx as described above. Analysis of AMso division and MCM fate psf-2(t3443ts) and wild-type animals were grown at the permissive temperature (15ºC). Synchronised populations of L1 larvae, obtained by hypochlorite treatment and hatching in M9 buffer, were shifted to the non-permissive temperature (25ºC). The AMso is born 310 min post- fertilization, thus the L1 shift ensures that only the asymmetric division to produce the MCM is .CC-BY-NC-ND 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 13, 2024. ; https://doi.org/10.1101/2024.05.09.593335doi: bioRxiv preprint 39 affected. 1-day adult male animals were scored for AMso division and the presence of the MCM neuron. To monitor the division, the saIs14 (Plin-48gfp) transgene was used 45. In wild-type animals, the saIs14 transgene is expressed in the AMso mother (ABpl/rpaapapa) starting with its birth during embryogenesis. After the postembryonic division (approx. 32-36 hours post-L1 arrest), it is continuously expressed in the anterior AMso daughter (ABpl/rpaapapaa), temporarily retained in the posterior daughter cell (ABpl/rpaapapap) and gradually lost as it differentiates into the MCM neuron. To assess neuronal identity, the otIs356 ( P rab-3NLS::rfp) panneuronal reporter transgene was used 46. Cells per side were quantified, with wild-type animals having two lin-48 and one rab-3 expressing cell per side. psf-2(t3443ts ) animals maintained at the permissive temperature (15ºC) were scored as controls. Animals carrying the rescuing transgene were maintained and scored at the non-permissive temperature (25ºC). Animals considered wild-type carry him-5(e1490). Fluorescence Imaging For the MI fate analysis, images of anesthetised worms were acquired on the Zeiss Axio Imager M2 with a 100X/1.3 NA oil-immersion objective lens. A stack that included the entire anterior pharynx of the larvae was acquired in the GFP and DIC channel with step-size of 0.5 µm. Using Fiji (ImageJ), a maximum intensity projection of the GFP channel was generated (to ensure all GFP positive cells in the anterior pharynx were included in the image). A central slice for the DIC channel was chosen that accurately represented the GFP positive cell. The GFP and DIC images were merged and used for Figure 5B. For AMso and MCM imaging, worms were anesthetised using 50 mM sodium azide and mounted on 5% agarose pads on glass slides. Images were acquired on a Zeiss AxioImager with a 40X/1.3 NA oil-immersion objective lens .CC-BY-NC-ND 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 13, 2024. ; https://doi.org/10.1101/2024.05.09.593335doi: bioRxiv preprint 40 coupled to a 2.5 zoom, using a Zeiss Colibri LED fluorescent light source and custom TimeToLive multichannel recording software (Caenotec). Representative images are shown following maximum intensity projections of 2–10 1 mm z-stack slices edited using Fiji. .CC-BY-NC-ND 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 13, 2024. ; https://doi.org/10.1101/2024.05.09.593335doi: bioRxiv preprint 41

We thank members of the Conradt, Schnabel, Lambie and Poole labs for discussions and Stéphane Roland, for comments on the manuscript. We thank C. Struck and L. McGuinness for excellent technical support. Some strains were provided by the Caenorhabditis Genetics Center (CGC), which is funded by NIH Office of Research Infrastructure Programs (P40 OD010440). We thank Don Moerman and the C. elegans Knockout Facility in Vancouver B.C. Canada (https://www.zoology.ubc.ca/~dgmweb/ ) for providing the annotated single nucleotide variants and indels in the psf-2(t3443) strain. RS was supported by a Post-graduate Scholarship - Doctoral (PGS D) from the Natural Sciences and Engineering Research Council of Canada (NSERC). This work was supported by UCL (Division of Biosciences, UCL LSM Capital Equipment Fund to BC), a Korean Institute for Basic Science Grant to NM (IBS-R022-D1), a Wolfson Fellowship from the Royal Society (https://royalsociety.org/ ) to BC (RSWF\R1\180008), and the Biotechnology and Biological Sciences Research Council (https://bbsrc.ukri.org/) (BB/V007572/1 and BB/V015648/1 to BC). .CC-BY-NC-ND 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 13, 2024. ; https://doi.org/10.1101/2024.05.09.593335doi: bioRxiv preprint 42 Figure legends Figure 1. Reducing psf-2 GINS2 function causes increased cell cycle lengths, a block in cell division and embryonic lethality. (A) Differential Interference Contrast (DIC) images of representative wild-type (+/+), psf-2(t3443ts), psf-2(t3443ts); psf-2(+) (transgene bcEx1302) and psf-2(RNAi) embryos at the 4-cell stage, the pre-morphogenetic stage and at the final recording (terminal phenotype). The +/+ and psf-2 (t3443ts); psf-2(+) embryos completed embryogenesis and hatched. Images were taken from long-term recordings performed at 25°C. Scale bars 10 µM. ( B) Embryonic lethality [%] and Brood size at permissive (15°C) and non-permissive temperature (25°C) in wild-type (+/+), psf-2(t3443ts) and psf-2(t3443ts); psf-2(+) (transgene bcEx1302) animals. For each genotype, embryonic lethality and brood size of the entire progeny of six adults was analyzed (n=6). Mean ±SD are indicated. (C) Cell cycle length [min] of the ABarppppp cell and its ancestors in wild-type (+/+) (n=5), psf-2(t3443ts) (n=5), psf-2(t3443ts); psf-2(+) (transgene bcEx1302 ) (n=5), psf-2(RNAi) (n=3) and tyms-1 (e2300ts) (n=6) animals at 25°C. Mean ±SD are indicated. (D) ABarpppppp (V6R) lineage of representative animals of the genotypes indicated. The wild-type and psf-2 (t3443ts); psf-2(+) embryos completed embryogenesis and hatched. The blocks in cell division of ABarpppaa and ABarpppap in the psf- 2(RNAi) embryo and the corresponding lineages in the other genotypes are indicated in red. Figure 2. Reducing psf-2 GINS2 function causes a general block in cell death. (A) Cell fate analysis of the first 13 cell deaths of the AB lineage in wild-type (+/+) (n=4; data from one representative embryo is shown), psf-2(t3443ts) (n=4), psf-2(t 3443ts); psf-2(+) (transgene bcEx1302) (n=4) and psf-2(RNAi) (n=3). Cell fate was determined based on 4D lineaging .CC-BY-NC-ND 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 13, 2024. ; https://doi.org/10.1101/2024.05.09.593335doi: bioRxiv preprint 43 analyses performed on long-term recordings done at 25°C as described in Materials and Methods. In the case of psf-2(t3443ts); psf-2(+), data shown is from embryos that hatched. (B) Percentage [%] cell death blocked during first wave of cell death (13 AB-derived cell deaths). Summary of data shown in (A) and data for the following additional genotypes: ced-3(n717), psf-3(RNAi), mcm-7(RNAi), mcm-2(RNAi) and tyms-1 (e2300ts). Unless noted otherwise, recordings were performed at 25°C. ( C) Percentage [%] MSpaapp cell death blocked in wild- type (+), psf-2(t3443ts ), and psf-2(t3443ts); psf-2(+) (transgene bcEx1302) animals. In the case of psf-2(t3443ts); psf-2(+), data shown is from embryos that hatched. ( D) Time until cell corpse formation [min] in wild-type (+/+) and psf-2(t3443ts) animals. Time measured was from the birth of the cell until the cell formed a cell corpse (button-like appearance by DIC) (n=14 for both genotypes). Mean ±SD are indicated, and mean value is stated above data. P value=0.0009, unpaired t-test with Welch correction. (E) Percentage [%] QL.pp cell death blocked in wild-type (+/+) animals at 25°C and psf-2(t3443ts) animals at 15°C and 25°C using the bcIs133 (P toe-2gfp) transgene. Figure 3. Lack of correlation between the increased cell cycle length phenotype and the cell death phenotype of psf-2(t3443ts) animals. (A) Schematic of cell cycle length [min] measurements of mothers of cells that die. ‘Cell cycle length’ is defined as the time in minutes from their births after the 8 th round of cell division until their own divisions (9 th round of cell division). Data was generated from 4D lineaging analyses done on long-term recordings performed at 25°C. (B) Cell cycle length [min] of mothers of cells that die in wild type (+/+) (n=56), psf-2(t3443ts) (mothers whose daughters died (‘cell death’) (n=15) and mothers whose daughters failed to die (‘cell death blocked’) (n=25)) and tyms-1(e2300ts) (n=31). The average .CC-BY-NC-ND 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 13, 2024. ; https://doi.org/10.1101/2024.05.09.593335doi: bioRxiv preprint 44 cell cycle lengths are stated above the data. Mean ±SD are indicated, and value of mean is stated above data. P value **** < 0.0001, unpaired t-test with Welch correction was used; ns: not significant. (C) ABalaap lineage of representative animals of the genotypes indicated. Figure 4. Reducing psf-2 GINS2 function abolishes the increase in egl-1 BH3-only mRNA observed in daughters that die. smRNA FISH analysis in (A) MSpaap (mother) and (B) MSpaapp(X) (daughter that dies) in wild-type (+/+) and psf-2(t3443ts) embryos. Top. Representative fluorescent images of embryos and MSpaapp or MSpaapp(X) (insets and enlargements). Cells are indicated by white circles and are 6.0 µm (MSpaap) and 3.5 µm (MSpaapp(X)) in diameter. Nuclei are labeled with DAPI and are shown in dark blue. Labelled egl-1 mRNAs or ced-3 mRNAs are shown in orange or light blue, respectively. Scale bars 10 µM. Bottom. mRNA copy numbers in MSpaap and MSpaapp(X). Mean ±SEM are indicated, and mean value is shown above data. P value=0.004, ns: not significant. Mann -Whitney test was performed. Time course of mRNA concentration [copy number/µm 3] of (C) egl-1 mRNA and (D) ced-3 mRNA in MSpaap and MSpaapp(X) in wild-type (+/+) and psf-2(t3443ts) embryos. X axis indicates number of nuclei in embryo. As indicated by the vertical dotted line, MSpaap divides when ~180 embryonic nuclei are present in the embryo. Graphs were generated from raw data as a centered moving average of order 5 as described in materials and methods. Shaded areas represent SEM. Figure 5 – Reducing psf-2 GINS2 induces cell fate defects in the MI and AMso lineages. (A) Embryonic cell lineage of the MI neuron. (B) Heat map showing the percentage of L4 hermaphrodites that express the MI-specific reporter transgene P sams-5gfp ( nIs396) in psf- .CC-BY-NC-ND 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 13, 2024. ; https://doi.org/10.1101/2024.05.09.593335doi: bioRxiv preprint 45 2(t3443ts) and his-9(n5357gf) mutants. Rows represent different genotypes and temperatures. Columns represent different phenotypes depicted as coloured circles. (C) Differential interference contrast (DIC) and fluorescence micrographs of the anterior pharynx of three hermaphrodites showing expression of the MI-specific transgene P sams-5gfp in psf-2(t3443ts ) mutants grown at the non-permissive temperature of 25ºC. (D) Postembryonic cell lineage of the AMso glial cell and the MCM neuron. (E) Heat map showing the percentage of cells per side in adult males that express the glial reporter transgene P lin-48gfp ( saIs14; left panel) and the panneuronal marker Prab-3::NLS::rfp (otIs356; right panel) in various phenotypes. bcEx1306 was used for psf-2(t3443ts) rescue. (F) Fluorescence micrographs showing expression of the glial marker P lin-48gfp and the panneuronal marker P rab-3::NLS::rfp in the AMso and MCM cells of wild type and psf-2(t3443ts) adult animals grown at the non-permissive temperature of 25ºC. .CC-BY-NC-ND 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 13, 2024. ; https://doi.org/10.1101/2024.05.09.593335doi: bioRxiv preprint 46

1 Li, R. The ar t o f cho reographing asy m metric c ell divis ion. Dev Cell 25, 439-45 0 (2013). https://doi.org/10.101 6/j.devcel.201 3.05.00 3 2 Rose, L. & Gonczy, P. Polarity estab lis hment, asym metric divis ion and segr egation of fa te determinants in ear ly C. el egans embryos. WormBook , 1-43 (2 014). https://doi.org/10.189 5/wor mbook.1.30.2 3 Sunchu, B. & Cabernard, C. Principles and mechanisms of asymmetric cell d ivis ion. Development 147 (2020). https://doi.org/10.1 242/dev.1 67650 4 Venkei, Z. G. & Yamashita, Y. M. Eme rging mechanisms of asym metric ste m cell divis ion. J Cell Biol 217 , 3785 -379 5 (2018). https://doi.org/10. 1083 /jcb.2018 07037 5 Holtzer, H. et al. Lineages, quantal ce ll cycl es, and the generati on of ce ll di versity. Q Rev Biophys 8 , 523-557 (19 75). https://doi.org/10.10 17/s00 3358 35000 01980 6 Holtzer, H., Weintraub, H., Mayne, R. & Mochan, B. The cell cycle, cell lin e ages, and cell differ entiat ion. Curr Top Dev Biol 7 , 229- 256 (1972). https://doi.org/1 0 .1016/s0 070 - 2153(08)60 073 -3 7 Liu, J. & Murray, J. I. M echanisms of lin eage specif icat ion in Caenorha bditis elegans. Genetics 225 (2023). https://doi.org/ 10.1093 /genetics/iyad174 8 Sulston, J. E. & Horv itz, H. R. Post-embryonic cell l ineages of t he nematode, Caenorhabditis e legans. Dev Biol 56, 110-1 56 (1977). 9 Sulston, J. E., Schierenberg, E., White , J. G. & Thomson, J. N. The e mbryonic cel l lin eage of the nematode Caenorhabd itis elegans. Dev Biol 100 , 64- 119 (198 3). 10 Packer, J. S. et al. A lineage -resolved molecular atlas of C. elegans emb ryogenesis at single-cell reso lution. Science 365 (20 19). https://doi.org/10.11 26/science.aax1971 11 Briggs, J. A. et al. The dynamics of gene expression in verteb rate embryogenesis a t single-cell reso lution. Science 360 (20 18). https://doi.org/10.11 26/science.aar5780 12 Brunski ll, E. W. et al. Single cell dissection of ea rly kid ney devel opment: multilineage priming. Development 141 , 3 093 -310 1 (2014). https://doi.org/10. 1242 /dev.110601 13 Hu, M. et al. Multil in eage gene exp ression p recedes commitment in th e hemopoieti c system. Genes Dev 11, 7 74- 785 (199 7 ). https://doi.org/10.1101 /gad .11.6 .7 74 14 Laslo, P. et al. Mult ilineage transcr iptional priming and de termination of a lterna t e hematopoietic cell fates. Cell 126 , 755-766 (2006). https://doi.org/10.101 6/j.cell.2006. 0 6.052 15 Thomson, M . et al. Pluripotency fac t ors in embryonic st em cells regulate differ entiat ion into germ layers. Cell 145 , 875-8 89 (2 011). https://doi.org/10.1 016/j .cell.2011.05. 017 16 Wang , W. et al. A single-cell t ranscripti onal r oadmap for card ioph aryngeal fat e diversi ficat ion. Nat Cell Biol 21, 674 -686 (2019). https://doi.org/10. 1038 /s41556 -01 9- 0336-z 17 Nimmo, R. A. , May, G. E. & Env er, T. Primed and ready: understanding lineage commitment through single ce ll analysis. Trends Cell Biol 25, 45 9-467 (201 5). https://doi.org/10.101 6/j.tcb.201 5.0 4.004 18 Conradt, B., Wu , Y. C. & Xue, D. Pro g rammed Cell Death During Caenorhabditis el egans Development. Genetics 203 , 1 533 -1562 (2016). https://doi.org/10.153 4/genetics.115 .186247 .CC-BY-NC-ND 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 13, 2024. ; https://doi.org/10.1101/2024.05.09.593335doi: bioRxiv preprint 47 19 Horvitz, H. R. Nobel le cture. Worms, life and death. Biosci Rep 23, 239-303 (2003). 20 Lomonosova, E. & Chinnadurai, G. BH3-only proteins in ap optosi s an d beyond: an overview. Oncogene 27 Suppl 1 , S2 -1 9 (2008). https://doi.org/10. 1038 /onc.2009.39 21 Conradt, B. & Horvitz, H. R. The C. ele gans protein EGL-1 is requ ired for p ro gram med cell death and inte racts with the B cl-2-like protein CED-9. Cell 93, 519 -52 9 (19 98). 22 Puthalakath, H. & Strasser, A. Keepin g killers on a tight leash: trans crip tio nal and post- translati onal c ontro l of the pro-apop totic activ ity of BH3-only prote ins. Cell Death Differ 9 , 505 -512 . (2002). 23 Sherrard, R. et al. miRNAs cooperate in apoptos is regulatio n durin g C. eleg ans development. Genes Dev 31, 209 -222 (2017). https://doi.org/10.1 101/ gad. 288555. 116 24 Hatzold, J. & Conradt, B. Contro l of apoptosis by asymmetric cel l div isio n. Plos Biol 6 , e84 (2008). https://doi.org/07 -PLBI -R A-26 04 [pii] 10.1371 /journal.pbio.0060084 25 Wei, H. , La mbie, E. J. , Osorio, D. S., C arvalho, A. X. & Conradt, B. PIG -1 ME LK-dependent phosphorylat ion of n onmuscle myosin II promotes apoptosis throug h CES -1 Sn ail partiti oning. PLoS Genet 16, e10089 12 (2020). https://doi.org/10.137 1/journal.pgen .100891 2 26 Wicks, S. R., Yeh, R. T. , Gish, W. R ., Waterston, R. H. & Plasterk, R. H. Rapid gene mapping in Caenorhabditis elegans using a high density polymorphism m ap. Nat Genet 28, 160 -16 4 (2001). https://doi.org/1 0.1038 /8887 8 27 Takayama, Y . et al. GINS, a n ovel multipro tein complex re quired for chr o mosomal DNA replicat ion in budding yeast . Genes Dev 17, 115 3- 11 65 (2003). https://doi.org/10.110 1/g ad.106 590 3 28 Li, H. & O'Donnell, M. E. The Eukaryo tic CMG Heli case at the Repl icatio n Fork: Emerging Architectu re Re veals an Unexpected Me chanism. Bioessays 40 (2018). https://doi.org/10.100 2/bies.2017 00 208 29 Pellegrini, L. The CMG DNA helicase and the co re rep lisome. Curr Opin Struct Biol 81, 102612 (20 23). https://doi.org/10. 10 16/j.sbi.202 3.102 612 30 Bischof f, M. & Schnabel, R. Global cell sort ing is mediated by local c ell-cel l interac tion s in the C. elegans embryo. Dev Biol 294 , 432-44 4 (2006). https://d oi.org/S0012 - 1606(06)00 184 -9 [pii] 10.1016 /j.ydbio.2006 .03.0 05 31 Schnabel, R., Hutter, H., Moerman, D. & Schnabel, H . Assessing norm al embryogenesis in Caenorhabditis e legans using a 4D microscope: variabili ty of d ev elopment and regional speci ficat ion. Dev Biol 184 , 234-26 5 (1997). 32 Ellis, H. M. & Horvitz, H. R. Geneti c co ntrol o f programmed cell death in the nematode C. elegans. Cell 44, 817 -829 (1 986). 33 Mishra, N., Wei, H. & Conradt, B. Ca enorhabditi s elegans ced-3 Caspase Is Required for Asym metric Div isions That Gen erate Cells Programmed To Die. Genetics 210 , 983-9 98 (2018). https://doi.org/10.1 534/ gene tics.118.301500 34 Sonneville, R., Querenet, M., Craig, A., Gartner, A. & Blow, J. J. The dynamics of replicat ion licens ing in l ive Caenorh abditis elegans embryos. J Cell Biol 196 , 233- 246 (2012). https://doi.org/10.1 083/jcb. 2 011100 80 .CC-BY-NC-ND 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 13, 2024. ; https://doi.org/10.1101/2024.05.09.593335doi: bioRxiv preprint 48 35 Schnabel, R. & Schnabel, H. Early determination in the C. elegans embryo: a gene, cib-1, required to sp ecif y a set of stem-cell-like b lastomeres. Development 108 , 1 07- 119 (1990). https://doi.org/10.1 242/dev. 108.Supplement.1 07 36 Kim, S. , Park, D. H. & Shim, J. Thymidylate synthase and dihy dropyrimidin e dehydrogenase levels are associa ted with response to 5-fluorouracil in Caenorhabditis elegans. Mol Cells 26, 344-3 49 (2008) . 37 Danenberg, P. V., Malli, H. & S wenson, S. Thymidylate synthase inhibi tors . Semin Oncol 26, 621 -63 1 (1999). 38 Marsh, S. Thy midylate synthase p harmacogenetics. Invest New Drugs 23, 533 -53 7 (2005). https://doi.org/10.1 007/s 106 37-00 5-4 021 -7 39 Chakraborty, S., La mbie, E. J. , Bindu, S., Mikeladze-Dvali, T. & Conradt, B. Engulfment pathways promote programmed cel l death by enhancing the unequal s egregation of apoptotic poten tial. Nat Commun 6 , 1012 6 (2015). https://doi.org/10.103 8/ncom ms10 1 26 40 Chen, F. et al. Translo cation of C. elegans CED-4 to nuclear membranes during program med cell death. Science 287 , 1485- 1489 (20 00). 41 Maurer, C. W., Chiorazzi, M. & Shaham, S. Timing of the onset of a deve lopmental cel l death is cont rol led by transc ript ional induction of the C. el egans ced-3 cas pase-encoding gene. Development 134 , 13 57- 1368 ( 2007). 42 Nakano, S., Stillman, B. & Horvit z , H. R. Repli cation-coupl ed ch romatin assembly generates a neuronal bilateral asym metry in C. elegans . Cell 147 , 1525 -1536 (201 1). https://doi.org/10.101 6/j.cell.2011. 1 1.053 43 Nakano, S., Ellis, R. E. & Horvitz, H. R. Otx-dependent expression of pr oneural bHLH genes establish es a ne uronal bilat eral asymmetry in C. elegans. Development 137 , 4017- 4027 (2010). https://doi.org/ 10.12 42 /dev.058834 44 Sam mut, M. et al. Glia-deri ved ne urons are requi red f or sex-speci fic learning in C. elegans. Nature 526 , 385 -390 (20 15). https://doi.org/10.103 8/nature15700 45 Sewell, S. T. , Zhang , G. , Uttam , A. & Chamberlin, H . M. Developmental patterning in the Caenorhabditis elegans hindgut. Dev Biol 262 , 88 - 93 (2003). https://doi.org/10.101 6/s001 2-1 606( 03)00352 -x 46 Stefanakis, N., Carrera, I. & Hobert, O. Regulatory Logic of Pan-Ne uronal Gene Expression in C. elega ns. Neuron 87, 7 33- 75 0 (2015). https://doi.org/10.101 6/j.neuron.201 5.07.03 1 47 Altamirano-Pacheco, L. & Navarro, P. A for k in the road to dif ferent iatio n. Nat Genet 55, 1422- 1423 (20 23). https://doi.org/10 .1038/s4 1588 -02 3-0 1489 -6 48 Wen, Q. et al. Symmetric inhe ritanc e of parental his tones contr ibutes t o safeguarding the fate of mouse embryonic stem cells during differ entia tion. Nat Genet 55 , 15 55- 1566 (2023). https://doi.org/10.1 038/s 415 88-02 3-0 1477 -w 49 Wenger, A. et al. Symmetric inher itance of par ental hist ones govern s epigenome maintenance and embryonic stem cell ident ity. Nat Genet 55, 1567-1578 (20 23). https://doi.org/10.103 8/s415 88- 023 - 01476 -x 50 Xu, X. , Hua, X ., Brown, K. , Ren, X. & Zhang, Z. Mcm2 promotes stem cell differ entiat ion via its abi lit y to bind H3-H4 . Elife 11 (2022). https://doi.org/10. 7554 /eLife.80917 .CC-BY-NC-ND 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 13, 2024. ; https://doi.org/10.1101/2024.05.09.593335doi: bioRxiv preprint 49 51 Varga, M. et al. Tissue-Specific Req uirement fo r th e GINS Complex Du ring Zebraf ish Development. Front Cell Dev Biol 8 , 37 3 (2020). https://doi.org/10.338 9/fcell.2020.00 373 52 Walter, B. E. e t a l . Psf2 plays impor tant r oles in n ormal eye deve lopment in Xenopus laevis. Mol Vis 14, 906-921 (20 08). 53 Jia, W., Hsieh, H. Y ., Kidoya, H. & Taka kura, N. Embryonic express ion of GIN S members in the development of the mam malian nervous system. Neurochem Int 129 , 1 04465 (2019). https://doi.org/10.1 016/j .neuint.2019.1044 65 54 Mohri, T. et al. Requirement of SLD5 for early embryogenesis. PLoS One 8 , e78 961 (2013). https://doi.org/10.1 371/jour nal.pone.0078961 55 Rubio-Ferrera, I. et al. Selective rol e of the DNA heli case Mcm5 in BMP retrograde signaling during Drosophila neuronal differen tiat ion. PLoS Genet 18, e10 10255 (20 22). https://doi.org/10.137 1/journal.pgen .101025 5 56 Lattmann, E. et al. A DNA replicati o n-independent func tion of pr e-replic ation complex genes during cell invasion in C. elegans. Plos Biol 20, e3 00 1317 (20 22). https://doi.org/10.137 1/journal.pbio.300131 7 57 Bellel li, R. et al. POLE3- POLE4 Is a Histone H3-H4 Chaperone that Main tai ns Chromatin Integrity dur ing DNA Replicati on. Mol Cell 72, 112-126 e115 (2018). https://doi.org/10.101 6/j. molcel.201 8.08.04 3 58 Foltman, M. et al. Eukaryot ic r eplis ome components coopera te to pro cess his tone s during chromosome replicati on. Cell Rep 3 , 892 -9 04 (2013). https://doi.org/10.101 6/j.celrep.201 3.02.02 8 59 Groth, A. et al. Regulation of repli cation for k progressi on through hist on e supply and demand. Science 318 , 192 8-1 931 (20 07). https://doi.org/10.11 26/science.114899 2 60 Huang, H. et al. A unique binding mo de enables MCM2 to chaperone hist ones H3-H 4 at replicat ion fo rks. Nat Struct Mol Biol 22, 618-6 26 (2015). https://doi.org/10.103 8/nsmb .305 5 61 Ishimi, Y. , Ichinose, S., Omori, A ., Sato, K. & Kimura, H. Binding of human minichromosome maintenance prote ins with histone H3. J Biol Chem 271 , 24115 -24 122 (1996). https://doi.org/10.1 074/jbc. 2 71.39.2 4115 62 Jasencakova, Z. et al. R eplica tion stress in terf eres with h istone r ecycling and predeposit ion marking of new histones. Mol Cell 37, 73 6 -743 (2010). https://doi.org/10.101 6/j. molcel.201 0.01.03 3 63 Yu, C. et al. A mechanism for p reventing asymmetric h iston e segr egation ont o replicat ing DNA strands. Science 361 , 1386- 1 389 (2018). https://doi.org/10.112 6/science.aat8849 64 Petryk, N. et al. MCM2 promotes s ymmetric inheritan ce o f modified hi stones dur ing DNA replicati on. Science 361 , 13 89- 1392 (2018). https://doi.org/10.112 6/science.aau0294 65 Escobar, T. M., Loyola, A. & Reinb erg, D. Parental nucl eosome segregation and th e inheritan ce of cel lular id enti ty . Nat Rev Genet 22, 379 - 392 (2021). https://doi.org/10.103 8/s415 76- 020 - 00312 -w .CC-BY-NC-ND 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 13, 2024. ; https://doi.org/10.1101/2024.05.09.593335doi: bioRxiv preprint 50 66 Serra-Cardona, A. & Zhang , Z. Replica tion-Coupled Nucl eosome Assembly i n the Passage of Epigenetic Information and Cell Identity. Trends Biochem Sci 43, 136-148 (201 8). https://doi.org/10.101 6/j.tibs.2017 .1 2.003 67 Stewart-Morgan, K . R., Petryk, N. & Groth, A. Chromatin replication and epigenetic ce ll memory. Nat Cell Biol 22, 361-371 (2 020). https://doi.org/10.1 038/s 41556 -020 -048 7-y 68 Blanco, E., Gonzalez-Ramirez, M., Alcaine-Colet, A., Aranda, S. & Di Croce, L. The Bivalen t Genome: Characterizat ion, Structu re, and Regulation. Trends Genet 36, 118-131 (20 20). https://doi.org/10.101 6/j.tig.2 019.1 1 .004 69 Macrae, T. A., Fothergill-Robinson, J. & Ramalho-Santos, M. Regulation, f unctions and transmission of bival ent chromatin during ma m malian development. Nat Rev Mol Cell Biol 24, 6-26 (2023). https://doi.org/ 1 0.1038 /s4158 0-0 22- 00518 -2 70 Janes, J. et al. Chromatin accessibil ity dynamics across C-elegans development and ageing. Elife 7 (2018). https://doi.org /ARTN e373 44 10.7554 /eLife.37344 71 Ahringer, J. & Gasser, S. M. Repressive Chromatin in Caenorhab ditis e legans: Establishment, Composition, and Function. Genetics 208 , 491 - 511 (2018). https://doi.org/10.153 4/genetics.117 .300386 72 Memar, N., Sethi, A., Luehr, S., Lambie, E. J. & Conradt, B. In vivo label ing o f endogenous genomic loci in C. elegans us in g CRIS PR/dCas9 . MicroPubl Biol 2022 (2022). https://doi.org/10.179 12/ micropub.biology.000701 73 Liu, Y., Hen gartner, M. O. & Herr, W. Selected elements o f herp es simplex virus accessory facto r HCF are highly cons erved in Caeno rhabditi s elegans. Mol Cell Biol 19, 909-9 15 (1999). 74 Ma, B. et al. Dif ferent ial Histo ne Distribu tion Patte rns in Induce d Asymmetrically Dividing Mouse Embryonic St em Cells. Cell Rep 32, 10 8003 (202 0). https://doi.org/10.101 6/j.celrep.202 0.10800 3 75 Tran, V., Lim , C., Xie, J. & Chen, X. Asy m metric div ision of D rosophi la male germline stem cell sh ows asymmetric hist one distr ibution. Science 338 , 679-682 (20 12). https://doi.org/10.112 6/science.122 6028 76 Li, Z. et al. Asymmetric distr ibutio n of parenta l H3K9me3 in S phas e silence s L1 elements. Nature 623 , 643-6 51 (2023 ). https://doi.org/10.1038 /s415 86- 02 3-067 11- 3 77 Conradt, B. & Horvitz, H. R. The TRA- 1A sex determination protein of C. elegans regulates sexually dimorphic ce ll de aths by repr essing the egl-1 cell de ath activat o r gene. Cell 98, 317 -32 7 (1999). 78 Nehme, R. & Conradt, B. egl-1: a k ey activat or of apop tot ic cel l death in C. elegans. Oncogene 27 Suppl 1 , S30 -40 (2 008). https://doi.org/onc200941 [pii] 10.1038 /onc.2009 .41 79 Thellmann, M. , Hatzold, J. & Conrad t, B. The Snail-like CE S-1 protein of C. elegans can block the expr ession of the BH3-only cell-death activa tor gene egl-1 by antagonizing the function of bHLH proteins. Development 130 , 4057 -40 71 (2003). 80 Tran, A. T. et al. MiR-35 buffers apoptosis th reshold s in the C. elegans germline by antagonizing both MAPK and cor e a poptosis pathways. Cell Death Differ 26, 2637 -26 51 (2019). https://doi.org/10.1 038/s 414 18-01 9-0 325 -6 .CC-BY-NC-ND 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 13, 2024. ; https://doi.org/10.1101/2024.05.09.593335doi: bioRxiv preprint 51 81 Tangye, S. G. et al. Hu man Inborn Err ors of Immunity: 2022 Update on the Classificat ion from the Internati onal Union of I m munological Societies Exper t Com mittee. J Clin Immunol 42, 1473 -150 7 (2022). https://doi.org/10. 1007 /s1087 5-0 22- 0128 9-3 82 Mace, E. M. Human natural killer ce l ls: Form, function, and development. J Allergy Clin Immunol 151 , 371 -385 (2 023). https:/ /doi.org/10.10 16/j.j aci.2022.0 9.02 2 83 Seo, S. & Mace, E. M. Diversity of h uman NK cell developmental pathw ays defined b y single-cell analyses. Curr Opin Immunol 74, 106 -1 11 (2022). https://doi.org/10.101 6/j.coi.2021 .11 .001 84 Mace, E. M. & Orange , J. S. E mergi ng insights into human health and NK cell bio logy from the stud y of NK cell def icienci es. Immunol Rev 287 , 202-225 (201 9). https://doi.org/10.111 1/imr.1 2725 85 Guilz, N. C., Ahn, Y. O., Seo, S. & Mace, E. M. Unwinding the Rol e o f the CMG Helicase in Inborn Errors of Immunity. J Clin Immunol 43, 84 7-861 (2023). https://doi.org/10.100 7/s108 75- 023 - 01437 -3 86 Conte, M. I. et al. Part ial loss-of-func tion mutation s in GINS4 lead to NK c ell de fici ency with neutropen ia. JCI Insight 7 (2022). https://doi.org/10.117 2/jci.insight.1 54948 87 Brenner, S. The genetics of Caenorha bditis e legans. Genetics 77, 71- 94 (19 74). 88 SEX1 3. The Nematode Caenorhabditis elegans . Cold Spring Harbor Laborat ory Press edn, (1988). 89 Davis, P. et al. WormBase in 2022-d ata, processes, and to ols for analy zing Caenorhabditis e legans. Genetics 220 (2022). https://doi.org/10.1 093/ gen etics/iyac003 90 Sternberg, P. W. et al. WormBa se 2024: status and transitioning to Alliance infrastru cture. Genetics (2024). https: //doi.org/10. 1093/ genetics/iyae050 91 Fire, A. et al. Poten t and speci fic genetic in terfe rence by doub le-stra nded RNA in Caenorhabditis e legans [see commen ts]. Nature 391 , 806 -81 1 (1998). 92 Hoeppner, D. J., Hengartner, M. O . & Schnabel, R. En gulfment genes co operate wit h ced-3 to promote cell death in Caeno rhabditis e legans. Nature 412 , 202- 20 6. (2001). 93 Reddien, P. W., Ca meron, S. & Horv itz, H. R. Phagocytosis promotes pro gram med cell death in C. elegans. Nature 412 , 198- 202. (2001). 94 Raj, A., van den Bogaard, P., Rifkin, S. A., van Oudenaarden, A. & Tya gi, S. I ma ging individual mRNA molecules using multiple singly labe led pr obes. Nat Methods 5 , 877- 879 (2008). https://doi.org/1 0.103 8/ nmeth.1253 95 Schindelin, J. et al. Fiji: an open-source p latfo rm for biol ogical-image analysis. Nat

9 , 676-682 (2 012). https://d oi.org/10.103 8/n meth.2019 96 Preibisch, S., Saalfeld, S., Schindelin, J. & To mancak, P. Software for bead-based registration o f selec tive plane illumi nation microscopy data. Nat Methods 7 , 418 -4 19 (2010). https://doi.org/10.1 038/n me th0610-4 18 97 Madeira, F. et al. Search and seque nce analysis too ls ser vices from EMBL-EBI in 2022. Nucleic Acids Res 50, W 276 -W2 79 (20 22). https://doi.org/10.10 93/nar/ gka c240 98 Jumper, J. et al. Highly accurate pr otein st ructur e predi ction wi th Alph aFold. Nature 596 , 58 3-5 89 (2021). https://doi.org/ 10.1038 /s415 86- 021 -0381 9-2 99 Varadi, M. & Velankar, S. The impact of AlphaFold Prot ein St ructur e Dat abase on th e fields of lif e sc iences. Proteomics 23, e2200 128 (2023). https://doi.org/10.100 2/p mic.20220 0128 .CC-BY-NC-ND 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 13, 2024. ; https://doi.org/10.1101/2024.05.09.593335doi: bioRxiv preprint 52 100 Robinson, J. T. et al. Integrative genomics viewer. Nat Biotechnol 29, 24-26 (201 1). https://doi.org/10.103 8/nbt.175 4 .CC-BY-NC-ND 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 13, 2024. ; https://doi.org/10.1101/2024.05.09.593335doi: bioRxiv preprint 4-cell stage Pre- morphogenetic stage Final recording A Memar_Fig1 psf-2(t3443ts) 0 20 40 60 80 100Embryonic lethality [%] 15 °C 25 °C B C 15 °C 25 °C 0 100 200 300 400Brood size psf-2 (t3443ts); psf-2(+) psf-2(RNAi) psf-2(RNAi) 24h psf-2(t3443ts);psf-2(+) psf-2(t3443ts) +/+ tyms-1(e2300ts) +/+ psf-2(t3443ts);psf-2(+)psf-2(t3443ts)+/+ 0 50 100 150 200 250 Cell cycle length [min] ABarpp ABarpppABarppppABarppppp psf-2(t3443 ts) 50 0 100 150 200 400 250 450 500 300 350 +/+ psf-2(RNAi)tyms-1(e2300 ts) psf-2(t3443 ts); psf-2(+) ABarp ABarp ABarpABarp ABarp [min] D ABarpppppp ABarpppppp ABarpppppp ABarpppppp ABarpppap ABarpppaa ABarpppppp .CC-BY-NC-ND 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 13, 2024. ; https://doi.org/10.1101/2024.05.09.593335doi: bioRxiv preprint B D EC Genotype % cell deaths blocked during first wave n +/+ 0 52 ced-3(n717) 100 32 psf-2(t3443ts) 15°C 0 26 mcm-7(RNAi) 64 28psf-2(RNAi) psf-3(RNAi) 62 psf-2(t3443ts) 62 40 24 psf-2(t3443ts);psf-2(+) 0 50 mcm-2(RNAi) 26 tyms-1(e2300ts) 22 55 18 0 31 Genotype % MSpaapp cell death blocked n +/+ 6 psf-2(t3443ts) 67 6 6 psf-2(t3443ts);psf-2(+) 0 60 23.6 49.7 +/+ psf-2(t3443ts) 0 50 100 150 p=0.0009 time until cell corpse formation [min] ABplpappap ABarpaaapp ABaraaaapp ABalppaapa ABalppaaaa ABalappaaa ABalaapapa cell death blockedcell death ABprpppapp ABprppaaap ABplpppapp ABplppaaap cell division of mother blocked ?cell lost ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ?? ? ? #1 #2 #3 #4 #1 #2 #3 #1 #2 #3#4 +/+ A Memar_Fig2 psf-2(t3443ts)Cell psf-2(t3443ts);psf-2(+) psf-2(RNAi) mother dies Genotype % QL.pp cell death blocked n +/+ 6 psf-2(t3443ts)15°C 0 77 60 psf-2(t3443ts) 0 7529 ABalaappaa ABalapapaa .CC-BY-NC-ND 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 13, 2024. ; https://doi.org/10.1101/2024.05.09.593335doi: bioRxiv preprint B cell death cell death blocked ✱✱✱✱ ✱✱✱✱ ns ns 42 108 101 125 Cell cycle length [min] N/A +/+ N/A psf-2(t3443 ts) tyms-1(e2300 ts) 0 100 200 300 Memar_Fig3 A 9th 8th X Cell cycle length [min] tyms-1 (e2300ts) C x x ABalaap +/+ psf-2 (t3443ts) x ABalaap 50 100 150 250 200 300 350 400 450 500 550 ABalaapapa CD1 xABalaapapa CD1 42 min 44 min 97 min 107 min ABalaappaa CD2 ABalaappaa CD2 x ABalaap 170 min 175 minABalaappaa CD2 [min] .CC-BY-NC-ND 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 13, 2024. ; https://doi.org/10.1101/2024.05.09.593335doi: bioRxiv preprint MSpaap(mother) egl-1 14.1 11.6 8.4 6.8 ns ns ced-3 A +/+ psf-2(t3443 ts) psf-2(t3443 ts)+/+ 0 10 20 30 40mRNA copy number MSpaapp(X) B egl-1 mRNA ced-3 mRNA +/+ psf-2 (t3443ts) psf-2 (t3443ts) egl-1 mRNA ced-3 mRNA +/+ psf-2 (t3443ts) 10.7 3 2.5 ns 0.0004 3.83.8 egl-1 ced-3 +/+ psf-2(t3443 ts) psf-2(t3443 ts)+/+ 0 5 10 15 20 25mRNA copy number C 0.0 0.2 0.4 0.6 0.8 +/+ psf-2(t3443ts) # of nuclei in embryo # of nuclei in embryo egl-1 [mRNA/µm3] 160 170 MSpaap MSpaapp(X) 180 190 200 210 160 170 180 190 200 210 160 170 180 190 200 210160 170 180 190 200 210 0.2 0.0 0.4 0.6 0.8 0.2 0.0 0.4 0.6 0.8 0.2 0.0 0.4 0.6 0.8MSpaap MSpaapp(X) +/+ # of nuclei in embryo # of nuclei in embryo psf-2(t3443ts) ced-3 [mRNA/µm3] MSpaap MSpaapp(X) D MSpaap MSpaapp(X) Memar_Fig4 .CC-BY-NC-ND 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 13, 2024. ; https://doi.org/10.1101/2024.05.09.593335doi: bioRxiv preprint +/+ 25 ºC psf-2(t3443ts) 15 ºC psf-2(t3443ts) 25 ºC psf-2(t3443ts); psf-2(+) 25 ºC rab-3 - pan-neuronal 21 17 2 0 28 68 4 0 0 3 0 0 wt 51 8 93 100 0 0 0 0 0 1 0 0 0 3 1 0 0 0 0 0154 61 246 n 106 0 0 0 0 0 21 17 2 wt 100 79 79 97 lin-48 - glial 0 0 4 1 D A F C MI neuron DIC Psams-5gfp no MI +/+ psf-2(t3443ts) ectopic MI Prab-3rfp Plin-48gfp AMso-to-MCM phenotypes +/+ psf-2(t3443ts) no division AMso AMso MCM 2 neurons AMso.a AMso.p no neuron AMso.a AMso.p ectopic division AMso.a AMso.p1 AMso.p2 ABaraappaa Memar_Fig5 m1DR MIAMso MCM ABpl/rpaapapa AMso +/+ 25 ºC sams-5 - MI 98 n +/+ 15 ºC84 0 0 100 100 wt psf-2(t3443ts) 15 ºC111 psf-2(t3443ts) 25 ºC111 5 6 100 89 his-9(n5357gf) 20 ºC95 29 071 his-9(n5357gf) 15 ºC101 49 051 0 0 0 0 B E 10 µm 10 µm 10 µm .CC-BY-NC-ND 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 13, 2024. ; https://doi.org/10.1101/2024.05.09.593335doi: bioRxiv preprint