Maternal H3K36 and H3K27 HMTs protect germline immortality via regulation of the transcription factor LIN-15B

Maternally synthesized products play critical roles in development of offspring. A premier example is the C. elegans H3K36 methyltransferase MES-4, which is essential for germline survival and development in offspring. How maternal MES-4 protects germline immortality is not well understood, but its role in H3K36 methylation hinted that it may regulate gene expression in Primordial Germ Cells (PGCs). We tested this hypothesis by profiling transcripts from single pairs of PGCs dissected from wild-type and mes-4 mutant (lacking maternal and zygotic MES-4) newly hatched larvae. We found that mes-4 PGCs display normal turn-on of most germline genes and normal repression of somatic genes, but dramatically up-regulate hundreds of genes on the X chromosome. We demonstrated that X mis-expression is the cause of germline death by generating and analyzing mes-4 mutants that inherited different endowments of X chromosome(s). Intriguingly, removal of the THAP transcription factor LIN-15B from mes-4 mutants reduced X mis-expression and prevented germline death. lin-15B is X-linked and mis-expressed in mes-4 PGCs, identifying it as a critical target for MES-4 repression. The above findings extend to the H3K27 methyltransferase MES-2/3/6, the C. elegans version of Polycomb Repressive Complex 2. We propose that maternal MES-4 and PRC2 cooperate to protect germline survival by preventing synthesis of germline-toxic products encoded by genes on the X chromosome, including the key transcription factor LIN-15B.

. Although PRC2 and MES-4 catalyze opposing flavors of histone marking, 89 the loss of either causes nearly identical mutant phenotypes (Capowski et al., 1991). Worms 90 that inherit a maternal load of gene product but cannot synthesize zygotic product (referred to 91 as mes M+Z-mutants) are fertile. Worms that do not inherit a maternal load or produce zygotic 92 gene product (mes M-Z-mutants) are sterile due to death of nascent germ cells in early-to mid-93 stage larvae. In mes M-Z+ mutants, zygotically synthesized product does not rescue fertility, 94 highlighting the critical importance of maternal product. PRC2's roles in transcriptional 95 repression and development have been intensively studied and are well defined across species, 96 including roles in C. elegans germline development (Bender et  Two findings challenge the model that MES-4 somehow promotes expression of germline 120 genes. First, among mes-4 M-Z-mutants, hermaphrodites (with 2 X chromosomes) are always 121 sterile, while males (with 1 X chromosome) can be fertile (Garvin et al., 1998). This suggested 122 that the dosage of X-linked genes matters for the Mes-4 mutant phenotype. Second, profiling To investigate the role of MES-4 in PGCs, the cells that critically rely on maternal MES-4 for 134 survival, and to formally test the model that MES-4 promotes expression of germline genes, we 135 performed transcript profiling in dissected single pairs of PGCs from wild-type versus mes-4 M-136 Z-mutant larvae. We asked if absence of maternal MES-4 causes PGCs to 1) fail to turn on 137 germline genes, 2) inappropriately turn on somatic genes, and/or 3) inappropriately turn on X-138 linked genes. We found that in mes-4 PGCs most germline genes were turned on normally, thus 139 disproving the model that the major role of MES-4 is to promote expression of germline genes 140 in PGCs. Most somatic genes were kept off, arguing that MES-4 does not protect the germline 141 by opposing somatic development. The most dramatic impact to the transcriptome in mes-4 142 PGCs was up-regulation of hundreds of X-linked genes, many of which are part of an oogenesis 143 program. Our genetic analysis of mes-4 mutants with different X-chromosome endowments 144 from the oocyte and sperm demonstrated that up-regulation of X-linked genes is the cause of 145 death of nascent germ cells in mes-4 M-Z-mutant larvae. We identified the transcription factor 146 LIN-15B, an X-linked gene up-regulated in mes-4 mutants, as a major cause of X mis-expression 147 and germline death in mes-4 mutants. Performing similar tests of PRC2 (mes-3) M-Z-mutant 148 larvae revealed that their PGCs up-regulate many X-linked genes in common with mes-4 PGCs, 149 and that removal of LIN-15B restores the health of their germline, as it does for mes-4 mutants. 150 This study revealed that maternal MES-4 and PRC2 cooperate to ensure germline survival and 151 health in offspring by preventing mis-expression of genes on the X chromosome, and that both 152 operate through a key X-encoded transcription factor. turn on a germline program ( Figure 1A). We developed a hand-dissection strategy that enables 166 us to isolate in <30 minutes single sets of 2 sister PGCs, marked by a specifically and highly 167 expressed germline marker (GLH-1::GFP), from wild-type or mes-4 M-Z-mutant larvae for RNA-168 seq library preparation. We performed differential expression analysis to identify genes that are 169 significantly down-regulated (DOWN) or up-regulated (UP) in mes-4 mutant PGCs compared to 170 wild-type (wt) PGCs. Our analysis identified 176 DOWN genes and 450 UP genes ( Figure 1B To determine whether the DOWN genes include germline genes that fail to turn on 173 normally in mes-4 PGCs, we analyzed transcript levels and fold changes (mes-4 vs. wt) for genes 174 that are members of 2 'germline' gene sets: 1) a 'germline-specific' set containing 168 genes 175 that are expressed in germline tissue but not in somatic tissues, and 2) a 'germline-enriched' set 176 containing 2176 genes that are expressed at higher levels in adults with a germline compared to 177 adults that lack a germline (see Methods for gene sets). We found that most germline-specific 178 genes (162 of 168 genes or 96%) and germline-enriched genes (2111 of 2176 genes or 97%) are 179 not significantly DOWN in mes-4 PGCs ( Figure 1B). The numbers of germline-specific genes and 180 germline-enriched genes that are DOWN are not more than expected by chance ( Figure 1E). 181 Since some gene expression defects may not manifest until after mes-4 PGCs have 182 started dividing, we used our hand-dissection strategy to isolate sets of 2 PGC descendants, 183 which we call Early Germ Cells (EGCs), from wt and mes-4 mutant L2 larvae and profiled their 184 transcripts. We found that mes-4 EGCs down-regulate more germline-specific and more 185 germline-enriched genes than mes-4 PGCs. However, like mes-4 PGCs, mes-4 EGCs turn on most 186 germline genes normally ( Figure 1-figure supplement 1).

187
As an independent test of differential expression in PGCs, we selected 3 genes and 188 performed smFISH to measure and compare their transcript levels between mes-4 and wt PGCs. 189 2 of the genes we tested, cpg-2 and pgl-3, are members of our germline-specific set and by 190 RNA-seq analysis were DOWN or not DOWN respectively, in mes-4 PGCs (Figure 2A-C, Figure  191 2-figure supplement 1). Corroborating our RNA-seq analysis, smFISH analysis showed that the 192 average transcript abundance of cpg-2 is significantly lower in mes-4 vs wt PGCs, while the 193 average transcript abundance of pgl-3 is not significantly different (Figure 2A-C). The other gene 194 we tested by smFISH, chs-1, is a member of our germline-enriched set and was consistently not Together, our RNA-seq and smFISH analysis showed that mes-4 PGCs turn on most germline 197 genes normally. 198     wt PGCs are members of a 'soma-specific' gene set that defines 861 genes expressed in soma 300 but not in germline. We found that only 19 UP genes are soma-specific, which is not a higher 301 number than expected by chance ( Figure 1C,F). Therefore, mes-4 PGCs do not mis-express a 302 soma-specific program. 303 304 MES-4 represses genes on the X chromosome including many oogenesis genes in PGCs 305 306 Repression of the X chromosomes in the C. elegans germline is essential for germline health 307 (reviewed in Strome et al., 2014). We found that 311 of the 2808 (11%) protein-coding X genes 308 are UP in mes-4 vs wt PGCs. Strikingly, more than half of all UP genes are on the X chromosome 309 (311 out of 450 genes), and this number is significantly higher than expected by chance ( Figure  310 1F). We found that mes-4 EGCs mis-express 564 X genes, including almost all of the X genes that 311 are mis-expressed in mes-4 PGCs and an additional 273 X genes ( Figure 1-figure supplement 312 1C,F). These data show that mes-4 PGCs mis-express many genes on the X chromosome and 313 that X mis-expression becomes more severe in their descendant EGCs.

314
As an independent test of differential expression, we selected 4 X-linked UP genes and 315 performed smFISH to compare their transcript levels between mes-4 vs wt PGCs (Figure 2-316 figure supplement 1). smFISH analysis showed that all 4 X genes have higher transcript 317 abundance in mes-4 PGCs than in wt PGCs ( Figure 2D-F), corroborating our transcriptome 318 analysis. These data reveal that the X chromosome is the primary focus of MES-4 regulation in 319 PGCs. 320 While most X-linked genes are repressed during germline proliferation and genes are those that are normally turned-on during oogenesis by comparing our set of X-linked 324 UP genes to a set of 'oogenesis' genes, defined as 470 X-linked and 1201 autosomal genes that 325 are expressed at higher levels in dissected adult oogenic germlines than in dissected 326 spermatogenic germlines (Ortiz et al., 2014). We found that 209 of the 311 (67%) X-linked UP 327 genes and 25 of the 149 (17%) autosomal UP genes are in the oogenesis set, which are both 328 higher numbers than expected by chance ( Figure 1F). The enrichment for oogenesis X-linked 329 genes was especially high. No other germline gene set that we tested (germline-specific, 330 germline-enriched, and spermatogenesis) was enriched in the set of UP genes in mes-4 PGCs 331 ( Figure 1F). Based on gene ontology (GO) analysis, the set of X-linked UP genes in mes-4 PGCs 332 and/or EGCs is enriched for biological process terms that characterize roles in oogenesis: 333 'reproduction' and 'embryo development ending in birth or hatching'. (Figure 1- figure  334 supplement 2F-H). We conclude that mes-4 PGCs mis-express an oogenesis program involving 335 many X-linked genes, which may interfere with the ability of mutant PGCs to proliferate. 336 337 Mis-expression of genes on the X chromosome(s) causes germline death in mes-4 mutants 338 339 Since X mis-expression is the largest defect to the transcriptome in mes-4 PGCs, we 340 hypothesized that mis-expression of the 2 X chromosomes in germlines of mes-4 mutant 341 hermaphrodites causes germline death. To test our hypothesis, we asked whether mes-4 342 mutant males, which inherit only a single X chromosome (typically from the oocyte), have 343 healthy germlines. We live imaged wild-type and mes-4 mutant M-Z-hermaphrodites and males 344 that express a germline-specific GFP reporter (GLH-1::GFP) and scored their germline health 345 qualitatively as either 'absent/tiny' germline, 'partial' germline, or 'full' germline. All live-346 imaged mes-4 mutant hermaphrodites lacked a germline ( Figure 3A). In contrast, some mes-4 347 mutant males that inherited their single X from an oocyte (X oo males) had either partial or full 348 germlines (21% and 4%, respectively). Since X chromosomes turn on during oogenesis ( inherited an X with a history of expression. Using a him-8 mutant, we generated wild-type and 351 mes-4 mutant males that instead inherited their X from a sperm (X sp males), which has a history 352 of repression because the X was not turned on previously during spermatogenesis. We tested 353 whether him-8; mes-4 mutant X sp males that inherited a single X with a history of repression 354 have healthier germlines than mes-4 mutant X oo males that inherited a single X with a history of 355 expression. Strikingly, 67% him-8; mes-4 mutant X sp males made full germlines, compared to 356 only 4% of mes-4 mutant X oo males.

357
A new and powerful genetic tool uses gpr-1 over-expression to generate hermaphrodite 358 worms that form a germline entirely composed of 2 genomes inherited from the sperm, or 359 rarely 2 genomes inherited from the oocyte (Besseling and Bringmann, 2016; Artiles et al., 360 2019). Using this tool, we tested whether mes-4 mutant hermaphrodites whose germline 361 inherited from a sperm 2 X chromosomes with a history of repression (which we called X sp /X sp 362 hermaphrodites) have healthier germlines than mes-4 mutant hermaphrodites whose germline 363 inherited from an oocyte 2 X chromosomes with a history of expression (called X oo /X oo 364 hermaphrodites). While all mes-4 mutant X oo /X oo hermaphrodites lacked a germline, some mes-365 4 mutant X sp /X sp hermaphrodites had partial or full germlines (32% and 18%, respectively) 366 ( Figure 3A). Our combined genetic analysis demonstrates that mis-expression of the X 367 chromosome(s) causes germline death in mes-4 mutants. It also underscores that mes-4 mutant 368 PGCs can launch a normal germline program. 369 370 MES-4 promotes germline health independently from its role in transmitting H3K36me3 371 patterns across generations 372 373 Transmission of epigenetic information across generations can impact the health of offspring.

374
We hypothesized that maternally loaded MES-4's role in transmitting H3K36me3 patterns from 375 parents to offspring is essential for offspring germline development. If so, then transmission of 376 parental chromosomes lacking H3K36me3 to offspring should cause their germline to die even 377 if they received maternal MES-4. To test our hypothesis, we used the gpr-1 genetic tool and the 378 GLH-1::GFP germline marker to generate F1 adult offspring whose PGCs inherited 2 379 H3K36me3(-) genomes from the sperm and either did or did not inherit maternal MES-4. 380 Importantly, the germline in both types of F1 offspring had the same genotype (met-1; mes-4); 381 the only difference between the F1s was the presence or absence of maternally loaded MES-4. 382 We found that over half (57%) of F1 adult offspring had a full germline if their PGCs inherited 2 383 H3K36me3(-) genomes and maternal MES-4 ( Figure 3B). In contrast, 0% of F1 adult offspring 384 had a full germline if their PGCs inherited 2 H3K36me3(-) genomes from the sperm and did not 385 inherit maternal MES-4 ( Figure 3B). This result shows that maternally loaded MES-4 is critical 386 for offspring germline development but that its critical role is not to transmit H3K36me3 387 patterns from parents to offspring. 388 Presence of maternal MES-4 allows many F1 offspring whose PGCs inherited 2 389 H3K36me3(-) genomes from the sperm to make a full germline. One possibility is that 390 maternally loaded H3K36 HMTs can re-establish sufficient levels of H3K36me3 marking to 391 H3K36me3 ( almost all F1 offspring whose PGCs inherited 2 H3K36me3(-) genomes from sperm to lack a 397 germline. These findings show that maternal loads of both H3K36 HMTs are required for F1 398 offspring whose PGCs inherited 2 H3K36me3(-) genomes from sperm to make a germline, and 399 suggest that newly established H3K36me3 marking of H3K36me3(-) chromosomes by 400 maternally loaded HMTs can enable PGCs to make a germline.  Fisher's exact tests were used to test whether the proportion of F1 adults with a full-sized 414 germline significantly differs between samples. P-value designations are * < 0.01 and **** < 1e- We hypothesized that LIN-15B causes mis-expression of X genes in germlines that lack 470 MES-4. To test our hypothesis, we used RNA-seq to determine whether gonads dissected from Removal of LIN-15B may only allow mes-4 M-Z-mutant hermaphrodites to make a full-497 sized germline if that germline inherited 2 X chromosomes with a history of repression (from 498 sperm), which by itself improves germline health in mes-4 M-Z-mutants ( Figure 3A, Figure 4D). 499 We analyzed the impact of loss of LIN-15B on germline health in X oo /X sp mutants that inherited 500 1 of their 2 X chromosomes with a history of expression (from the oocyte). We found that 29% 501 of mes-4 M-Z-; lin-15B M-Z-X oo /X sp adult mutant hermaphrodites made full-sized germlines 502 compared to 0% of mes-4 M-Z-; lin-15B M+Z+ adult mutant hermaphrodites (Figure 4-figure  503 supplement 3). This finding demonstrates that removal of LIN-15B can even allow mes-4 M-Z-504 X oo /X sp mutants to make a full-sized germline. 505 To investigate whether other factors contribute to sterility in mes-4 M-Z-mutants, we 506 identified candidate genes that met 1 or more of 3 criteria: 1) they are X-linked and UP in mes-4 507 PGCs and/or EGCs, 2) there is evidence of them binding to the promoter region of at least 25% 508 of X-linked UP genes, and 3) they target a DNA motif that is enriched in the promoter of X- germline stage and away from wild-type samples ( Figure 5A). Using differential expression 592 analysis, we identified 354 X genes UP in mes-3 vs wt PGCs and 443 X genes UP in mes-3 vs wt 593 EGCs. We found that stage-matched mes-3 and mes-4 samples up-regulate a highly similar set 594 of X genes ( Figure 5B-C). Next, we compared log2(fold change) (mutant vs wt) of mis-regulated 595 X genes between mes-4 and mes-3 PGCs and between mes-4 and mes-3 EGCs. We found a 596 positive, albeit small, correlation between PGCs (0.22 Spearman's correlation coefficient) and a 597 stronger correlation between EGCs (0.44 Spearman's correlation coefficient) ( Figure 5D-E). 598 Moreover, as in mes-4 M-Z-mutants, loss of LIN-15B caused mes-3 M-Z-mutants to make 599 healthier germlines (Figure 4-figure supplement 3). We conclude that MES-4 and PRC2 600 cooperate to ensure germline survival in M-Z-mutant larvae by repressing similar sets of X 601 genes and that both operate through LIN-15B.

603
The chromodomain protein MRG-1 is a candidate reader and effector of H3K36me3   Figure 5 (

B,C) Venn diagrams comparing X-UP genes (mutant vs wt) in mes-4 and mes-3 PGCs (B) and in 632 mes-4 and mes-3 EGCs (C). (D,E) Scatterplots comparing log2(fold change) (mutant vs wt) of 633 transcript abundance for X-UP genes (circles) in mes-4 or mes-3 PGCs (D) or in mes-4 or mes-3 634
EGCs (E). The Spearman correlation coefficient along with its p-value is indicated at the top of 635 each scatterplot. (F) Cartoon model illustrating how MES-4 protects germline survival by 636 repressing X genes (gray boxes). MES-4 may indirectly repress X genes, including lin-15B, by 637 concentrating a repressor (e.g. PRC2) on the X or by restricting an activator (e.g. histone 638 acetyltransferase or LIN-15B) from the X. Our findings identify LIN-15B as a key player in 639 activating X genes and causing germline death upon loss of MES-4. LIN-15B may activate X 640 genes directly by binding to those genes or indirectly by regulating 1 or more other 641 transcription factors.  645  646  647  648  649  650  651  652  653  654  655  656  657  658  659  660  661  662  663  664  665  666  667  668  669  670  671  672  673  674  675  DISCUSSION  676  677 In this study, we investigated how a maternally supplied chromatin regulator protects germline 678 immortality and promotes germline health. We found that nascent C. elegans germlines (PGCs 679 and EGCs) that completely lack maternal MES-4 mis-express over a thousand genes, most of 680 which are on the X chromosome. We further demonstrated that X mis-expression is the cause 681 of germline death in mes-4 M-Z-mutants. Removal of a single transcription factor, LIN-15B, 682 reduced X mis-expression in the germline of mes-4 mutant mothers (mes-4 M+Z-) and was 683 sufficient to allow most of their offspring (mes-4 M-Z-) to develop full-sized germlines. 684 Intriguingly, lin-15B is itself X-linked and mis-expressed in nascent germlines that lack MES-4, 685 highlighting lin-15B as a key target for MES-4 repression. We favor a model where maternal 686 MES-4 promotes offspring germline development by preventing LIN-15B from activating a 687 germline-toxic program of gene expression from the X chromosome ( Figure 5F). This work 688 underscores how maternally supplied factors can guide development of specific tissues in 689 offspring by protecting their transcriptome. importance of H3K36me3 marking is also highlighted by our finding that loss of the candidate 721 H3K36me3 'reader' MRG-1 (homolog of yeast Eaf-3, fly MSL3, and human MRG15) ( Since MES-4 binding and its HMT activity are very low across almost the entire X chromosome 727 , it is likely that MES-4 regulates expression of X genes indirectly in 728 PGCs. One possible mechanism for indirect regulation is that MES-4 generates H3K36me3 on 729 autosomes to repel and concentrate a transcriptional repressor on the X chromosome(s). An hypomorphic mutations in the helicase domain of DRH-3 that abolish production of most 22G 796 RNAs do not impact germline formation, suggesting that 22G RNAs are not needed to specify 797 germline fate (Gu et al., 2009). We propose the intriguing possibility that in C. elegans, germline 798 fate is the default, which must be protected in the germline (e.g. by MES proteins and P 799 granules) and opposed in somatic tissues (e.g. by DREAM and LIN-15B). 800 801

(Vasa) is a component of germ granules and is specifically and highly expressed in germline
The null alleles mes-3(bn199) and met-1(bn200) linked to glh-1::GFP were created by inserting 849 TAACTAACTAAAGATCT into the 1st exon of each locus. The resulting genomic edit introduced a 850 TAA stop codon in each reading frame, a frame shift in the coding sequence, and a BglII 851 restriction site (AGATCT) for genotyping. Alt-R crRNA oligos (IDT) were designed using CRISPOR 852 (Concordet and Haeussler, 2018) and the UCSC Genome Browser (ce10) to produce highly 853 efficient and specific Cas9 cleavage in the 1st exons of mes-3 and met-1. Ultramer ssDNA oligos 854 (IDT) containing 50 bp micro-homology arms were used as repair templates. A dpy-10 co-855 CRISPR strategy (Arribere et al., 2014) was used to isolate strains carrying our desired 856 mutations. Briefly, 2.0 uL of 100 µM mes-3 or met-1 crRNA and 0.5 uL of 100 µM dpy-10 cRNA 857 were annealed to 2.5 uL of 100 µM tracrRNA (IDT) by incubation at 95˚C for 2 minutes, then at 858 room temperature for 5 minutes, to produce sgRNAs. sgRNAs were complexed with 5 uL of 40 859 µM Cas9 protein at room temperature for 5 minutes, 1 uL of 40 µM mes-3 or met-1 repair 860 template and 1 uL of 40 µM dpy-10(cn64) repair template were added, and the mix was 861 centrifuged at 13,000g for 10 minutes. All RNA oligos were resuspended in duplex buffer (IDT, 862 #11-05-01-03). Mixes were injected into 1 or both gonad arms of ~30 DUP64 adults. 863 Transformant progeny were isolated and back-crossed 4x to DUP64. 864 865 Isolation of single sets of 2 sister PGCs or 2 EGCs 866 867 L1 larvae hatched within a 30-minute window in the absence of food were allowed to feed for 868 30 minutes to start PGC development. Larvae were partially immobilized in 15 uL drops of egg 869 buffer (25 mM HEPES, pH 7.5, 118 mM NaCl, 48 mM KCl, 2 mM MgCl2, 2 mM CaCl2, adjusted to 870 340 mOsm) on poly-lysine coated microscope slides and hand-dissected using 30-gauge needles 871 to release their gonad primordium (consisting of 2 connected sister PGCs and 2 somatic gonad 872 precursors). Germline-specific expression of GLH-1::GFP was used to identify PGCs, which were 873 separated from gonad precursor cells by mouth pipetting using pulled glass capillaries coated 874 with Sigmacote (Sigma, #SL2) and 1% BSA in egg buffer. 7.5 mg/mL pronase (Sigma, #P8811) 875 and 5 mM EDTA were added to reduce sticking of gonad primordia to the poly-lysine coated 876 slides and to weaken cell-cell interactions. Single sets of sister PGCs were transferred into 0.5 uL 877 drops of egg buffer placed inside the caps of 0.5 mL low-bind tubes (USA Scientific, #1405-878 2600). Only PGCs that maintained bright fluorescence of GFP throughout isolation and were 879 clearly separated from somatic gonad precursors were used for transcript profiling. Isolation of 880 EGCs differed in 3 ways: 1) EGCs were dissected from L2 larvae that were fed for 20 hours after 881 hatching, 2) the 2 EGCs that made up 1 sample may have come from different animals, and 3) 882 the stage of each EGC could not be determined and therefore may have differed between 883 samples. Tubes containing single sets of 2 sister PGCs or 2 EGCs were quickly centrifuged, flash 884 frozen in liquid nitrogen, and stored at -80˚C. A detailed protocol for isolating PGCs and EGCs 885 from larvae is available upon request. At least 11 samples (replicates) of PGCs or EGCs were 886 isolated for each condition. were cut open with 30-gauge needles in egg buffer (see recipe above, except not adjusted to 892 340 mOsm) containing 0.1% Tween and 1 mM levamisole to extrude their gonads. Gonads were 893 cut at the narrow 'bend' to separate the gonad region containing mitotic and early meiotic 894 germ cells from the region that contains oocytes and/or sperm; the former was used for RNA 895 profiling. 20-60 gonads were mouth pipetted into 500 uL Trizol reagent (Life Technologies, 896 #15596018), flash-frozen in liquid nitrogen, and stored at -80˚C for up to 1 month before RNA 897 extraction. 898 To release RNAs from gonads in Trizol, gonads were freeze-thawed 3x using liquid 899 nitrogen and a 37˚C water bath, while vortexing vigorously between cycles. RNAs immersed in 900 Trizol were added to phase-lock heavy gel tubes (Brinkmann Instruments, INC., #955-15-404-5) 901 and mixed with 100 uL of 1-bromo-3-chloropropane (BCP) (Sigma, #B9673), followed by room 902 temperature incubation for 10 minutes. Samples were then centrifuged at 13,000g and 4˚C for 903 15 minutes to separate phases. RNAs in the aqueous phase were precipitated by mixing well 904 with 0.7-0.8x volumes of ice-cold isopropanol and 1 uL 20 mg/mL glycogen, followed by 905 incubation at -80˚C for were removed from downstream analyses if they were obviously very dim in all channels. 1008 Numbers of analyzed PGCs per probe set per genotype are indicated in Figure 2. A Laplacian of 1009 Gaussian (LoG) filter was used to enhance the signal-to-noise contrast in smFISH images. Those 1010 filtered 3D images were thresholded by signal intensity to produce binary images. The 1011 'imregionalmax' function from the Imaging Processing Toolbox in MATLAB was used to find and 1012 count regional maxima (transcript foci) in those binary images. To isolate and count regional 1013 maxima in PGCs, a 2D binary image mask of the PGCs was generated in Fiji from a maximum-1014 intensity Z-projection of the GLH-1::GFP image channel and applied to all Z-slices in the 3D 1015 smFISH images. Segmentation of PGCs to create the 2D mask was done in Fiji by first blurring Z-1016 projections using a large Gaussian kernel and then detecting edges in the blurred image. Mann-1017 Whitney tests were performed to compare transcript abundance between mes-4 and wild-type 1018 PGCs.

1019
To choose an appropriate signal intensity threshold for a 3D smFISH image, the number 1020 of detected regional maxima across 100 increasingly stringent thresholds applied to the image 1021 were plotted (Raj et al., 2008). In that plot, a range of thresholds that produced similar numbers 1022 of detected regional maxima was identified, and a threshold within that range was selected. 1023 Since threshold values were similar for all images within an image set (the collection of images 1024 acquired on the same day and for one probe set), an averaged threshold (across 5 images) was      expressed mCherry driven by the eft-3 promoter was used to track X-chromosome inheritance 1694 patterns in F1 offspring. (C) To generate males that inherited their single X from the sperm, we 1695 used the him-8(e1489) allele to cause X-chromosome nondisjunction in the maternal germline, 1696 which causes some oocytes to lack an X. (D) To generate hermaphrodites whose germline 1697 inherited either 2 genomes from the oocyte or 2 genomes from the sperm, we used a gpr-1 1698 over-expression allele. Expression of mCherry driven by the myo-2 promoter in AB-derived 1699 pharyngeal muscle or P1-derived pharyngeal muscle was used to identify F1 hermaphrodite 1700 offspring with non-Mendelian inheritance patterns.