Abstract
An evolutionary perspective enhances our understanding of biological mechanisms. Comparison of sex determination and X-chromosome dosage compensation mechanisms between the closely related nematode species C. briggsae (Cbr) and C. elegans (Cel) revealed that the genetic regulatory hierarchy controlling these processes is conserved, but both the X-chromosome target specificity and mode of binding for the specialized condensin dosage compensation complex (DCC) controlling X gene expression have diverged. We identified two motifs within Cbr DCC recruitment sites that are highly enriched on X: 13-bp MEX and 30-bp MEX II. Mutating either MEX or MEX II in an endogenous recruitment site with multiple copies of one or both motifs reduced binding, but only removing all motifs eliminated binding in vivo. Hence, DCC binding to Cbr recruitment sites appears additive. In contrast, DCC binding to Cel recruitment sites is synergistic: mutating even one motif in vivo eliminated binding. Although all X-chromosome motifs share the sequence CAGGG, they have otherwise diverged so that a motif from one species cannot function in the other. This functional divergence was demonstrated in vivo and in vitro. A single nucleotide position in Cbr MEX can act as a critical determinant for whether Cel DCC binds. The rapid divergence of DCC target specificity could have contributed to nematode speciation and contrasts dramatically with the conservation of target specificity for transcription factors that control developmental processes such as body-plan specification from fruit flies to mice.
Introduction
Comparative studies have shown that different facets of metazoan development exhibit remarkably different degrees of conservation across species (Carroll, 2008). At one extreme, homeobox-containing Hox genes and Wnt pathway signaling genes play conserved roles in body plan formation (Hox) as well as cell-fate determination, neural patterning, or organogenesis (Wnt) across clades diverged by at least 600 million years (MYR) (Malicki et al., 1990; De Kumar and Darland, 2021; Rim et al., 2022). Distant orthologous genes within these ancestral pathways can often substitute for each other. For example, both the mouse Small eye (Pax-6) gene (Hill et al., 1991) and the fruit fly eyeless (ey) gene (Quiring et al., 1994; Halder et al., 1995) control eye morphogenesis and encode a transcription factor that includes a paired domain and a homeodomain. Ectopic expression of mouse Pax-6 in diverse fruit fly imaginal disc primordia induced morphologically normal ectopic compound eye structures on fruit fly wings, legs, and antennae (Halder et al., 1995). Hence, at a deep level, eye morphogenesis is under related genetic and molecular control in vertebrates and insects, despite profound differences in eye morphology and mode of development.
At the other extreme are aspects of development related to sex. For example, chromosomal strategies to determine sexual fate in mice, fruit flies, and nematodes (XY or XO males and XX females or hermaphrodites) and the mechanism needed to compensate for the consequent difference in X-chromosome dose between sexes have diverged greatly. To balance X gene expression between sexes, female mice randomly inactivate one X chromosome (Yin et al., 2021; Loda et al., 2022), male fruit flies double expression of their single X chromosome (Samata and Akhtar, 2018; Rieder et al., 2019), and hermaphrodite worms halve expression of both X chromosomes (Meyer, 2022a; Meyer, 2022b)
The divergence in these pathways is so great that comparisons among animals of the same genus can provide useful evolutionary context for understanding the developmental mechanisms that distinguish the sexes. Therefore, we determined the genetic and molecular specification of sexual fate and X-chromosome dosage compensation in the nematode Caenorhabditis briggsae (Cbr) and compared it to the wealth of knowledge amassed about these processes in Caenorhabditis elegans (Cel). These two species have diverged by 15-30 MYR (Cutter, 2008).
In C. elegans, the sex determination and dosage compensation pathways are linked by genes that coordinately control both processes. For example, in XX embryos, the switch gene sdc-2 sets the sex determination pathway to the hermaphrodite mode and triggers binding of a dosage compensation complex (DCC) onto both X chromosomes to reduce X gene expression by half and thereby match X expression with that from XO males (Meyer, 2022a). The DCC shares subunits with condensin, a protein complex that controls the structure, resolution, and segregation of mitotic and meiotic chromosomes from yeast to humans (Yatskevich et al., 2019; Meyer, 2022a).
We determined the extent to which the sex-specific gene regulatory hierarchy is conserved between C. elegans and C. briggsae and the extent to which subunits of the C. briggsae DCC correspond to those of the C. elegans DCC. We also defined the cis-acting regulatory sites that confer X-chromosome specificity and recruit the C. briggsae DCC. We found that the DCC itself and the regulatory hierarchy that determines sex and directs the DCC to X have been conserved, but remarkably, both the X-chromosome target specificity of the C. briggsae DCC and its mode of binding to X have diverged.
Results
Conservation between C. briggsae and C. elegans of the core dosage compensation machinery and genetic hierarchy that regulates dosage compensation
The pivotal hermaphrodite-specific regulatory protein that controls both sex determination and dosage compensation in C. elegans is SDC-2, a 350 kDa protein with no known homologs outside of nematodes and only a coiled-coil domain as a predicted structural feature (Meyer, 2022a). It directs the DCC to both X chromosomes of XX embryos to achieve dosage compensation and also activates the hermaphrodite program of sexual differentiation (Chuang et al., 1996; Dawes et al., 1999; Chu et al., 2002; Pferdehirt et al., 2011). Loss of Cel sdc-2 causes XX-specific lethality due to excessive X-chromosome gene expression and masculinization of escaper animals (Nusbaum and Meyer, 1989; Kruesi et al., 2013). To assess the conservation of gene function, we deployed genome-editing technology in C. briggsae to knockout sdc-2, which shares 31% amino acid identity and 44% similarity with C. elegans (Figure 1—figure supplement 1). Using a PCR-based molecular strategy to identify insertions and deletions induced by DNA repair following directed mutagenesis with zinc finger nucleases, we recovered several independent Cbr sdc-2 mutant lines (Figure 1—figure supplement 2). Homozygous Cbr sdc-2 mutations caused extensive XX-specific lethality, consistent with a defect in dosage compensation and the conservation of gene function (Figure 1A). Nearly all Cbr sdc-2 hermaphrodites died as embryos or young larvae; rare XX survivors exhibited slow growth and masculinization. Cbr sdc-2 males were viable (Figure 1A) and had wild-type body morphology.
To determine whether the hermaphrodite-specific lethality of Cbr sdc-2 mutants was caused by defects in dosage compensation, we first identified components of the C. briggsae DCC and then asked whether DCC binding to X is disrupted by a Cbr sdc-2 mutation, as it is by a Cel sdc-2 mutation. In C. elegans, five of the ten known DCC proteins are homologous to subunits of condensin, an evolutionarily conserved protein complex required to restructure and resolve chromosomes in preparation for cell divisions in mitosis and meiosis (Figure 1B) (Chuang et al., 1994; Lieb et al., 1996; Lieb et al., 1998; Chan et al., 2004; Tsai et al., 2008; Csankovszki et al., 2009; Mets and Meyer, 2009; Yatskevich et al., 2019; Meyer, 2022a). The evolutionary time scale over which condensin subunits became coopted for dosage compensation in nematodes had not been explored.
Several lines of evidence indicate that a condensin complex mediates dosage compensation in C. briggsae as well. First, BLASTP searches revealed C. briggsae orthologs of all known C. elegans DCC condensin subunits (Figure 1B). Immunofluorescence experiments using antibodies raised against Cbr DPY-27, the SMC4 ortholog of the sole Cel DCC condensin subunit (Cel DPY-27) not associated with mitotic or meiotic condensins (Chuang et al., 1994), revealed X chromosome-specific localization in hermaphrodites, but not males, indicating conservation of function (Figure 1C and Figure 2A,B). Second, mass-spectrometric analysis of proteins that co-immunoprecipitated with Cbr DPY-27 identified Cbr MIX-1 (Figure 1—table supplement 1), the SMC2 condensin subunit ortholog (Lieb et al., 1998). Immunofluorescence experiments using Cbr MIX-1 antibodies revealed co-localization of Cbr MIX-1 with Cbr DPY-27 on hermaphrodite X chromosomes (Figure 1D). Third, disruption of Cbr dpy-27 conferred hermaphrodite-specific lethality, with XX escaper animals exhibiting a dumpy phenotype, like disruption of Cel dpy-27 (Figure 1F). Cbr MIX-1 protein did not localize to X in Cbr dpy-27(y436) mutant animals, consistent with the two proteins participating in a complex (Figure 1B,E). These data demonstrate that condensin subunits play conserved roles in the dosage compensation machinery of both C. briggsae and C. elegans.
Evidence that DCC binding defects underlie the XX-specific lethality caused by Cbr sdc-2 mutations is our observation that neither Cbr DPY-27 (Figure 2C) nor Cbr MIX-1 (not shown) localizes to X chromosomes in Cbr sdc-2 mutant hermaphrodites. Thus, the role for sdc-2 in the genetic hierarchies that activate dosage compensation is also conserved.
We next explored why maternally supplied DCC subunits fail to bind to the single X chromosome of C. briggsae males. In C. elegans XO embryos, the master switch gene xol-1 (XO lethal) represses the hermaphrodite-specific sdc-2 gene required for DCC binding to X and thereby prevents other DCC subunits from functioning in males (Miller et al., 1988; Rhind et al., 1995; Dawes et al., 1999; Meyer, 2022a). Loss of Cel xol-1 activates Cel sdc-2 in XO embryos, causing DCC binding to X, reduction in X-chromosome gene expression, and consequent death. We isolated a null mutant allele of Cbr xol-1(y430) by PCR screening of a C. briggsae deletion library (Supplementary File 1). We found that the Cbr xol-1 mutation caused inappropriate binding of the DCC to the single X of XO embryos (Figure 2D) and fully penetrant male lethality (Figure 3B), as expected from the disruption of a gene that prevents the DCC machinery from functioning in C. briggsae males. Cbr xol-1 mutant XX hermaphrodites appeared wild type.
To investigate the hierarchical relationship between Cbr xol-1 and Cbr sdc-2, we asked whether a Cbr sdc-2 mutation could suppress the male lethality caused by a Cbr xol-1 mutation. Both genes are closely linked in C. briggsae, prompting us to use genome editing technology to introduce de novo mutations in cis to pre-existing lesions without relying on genetic recombination between closely linked genes. If Cbr xol-1 controls Cbr sdc-2, then mutation of Cbr sdc-2 should rescue the male lethality of Cbr xol-1 mutants (Figure 2E). This prediction proved to be correct. XO males were observed among F1 progeny from mated Cbr xol-1 hermaphrodites injected with ZFNs targeting Cbr sdc-2 (Figure 3A,B,D). Insertion and deletion mutations were found at the Cbr sdc-2 target site in more than twenty tested F1 males (examples are in Figure 1—figure supplement 2C,D). Quantification of male viability in four different xol-1 sdc-2 mutant lines revealed nearly full rescue (Figure 3B), with concomitant absence of DCC binding on the single X chromosome (Figure 2E). Therefore, Cbr xol-1 functions upstream of Cbr sdc-2 to repress it and thereby prevents DCC binding to the male X chromosome. In summary, not only is the core condensin dosage compensation machinery conserved between Caenorhabditis species, so also are the key features of the genetic hierarchy that confers sex-specificity to the dosage compensation process.
Conservation between C. briggsae and C. elegans of the genetic hierarchy that regulates sex determination
Mechanisms controlling sex determination and differentiation are dynamic over evolutionary time; major differences can exist even within individual species. For example, males within the house fly species Musca domestica can utilize one of many different male-determining factors on autosomes and sex chromosomes to determine sex depending on a factor’s linkage to other beneficial traits (Meisel et al., 2016). Within the Caenorhabditis genus, differences and similarities have been observed in the genetic pathways governing the later stages of sexual differentiation (Haag, 2005). In fact, hermaphroditism evolved independently for C. elegans and C. briggsae; sperm production differs between them (Yin et al., 2018; Cutter et al., 2019). However, the earlier stages of sex determination had not been analyzed in C. briggsae. Therefore, we asked whether xol-1 and sdc-2 control sexual fate as well as dosage compensation in C. briggsae, as they do in C. elegans.
Our analysis of Cbr sdc-2 XX mutant phenotypes revealed intersexual tail morphology in the rare animals that survived to the L3/L4 stage (Figure 3C), indicating a role for Cbr sdc-2 in sex determination. Sexual transformation to the male fate is unlikely to have resulted from a disruption in dosage compensation, since such transformation was never observed in Cbr dpy-27 XX mutants (Figure 3C). Analysis of sexual phenotypes in double mutant strains confirmed that Cbr sdc-2 controls sex determination. Specifically, Cbr xol-1 Cbr sdc-2 double mutant XO animals develop as males, whereas Cbr dpy-27; Cbr xol-1 double mutant XO animals develop as hermaphrodites (Figure 3C,D). That is, both Cbr sdc-2 and Cbr dpy-27 mutations suppress the XO lethality caused by a xol-1 mutation, but only Cbr sdc-2 mutations also suppress the sexual transformation of XO animals into hermaphrodites. These results show that both sdc-2 and dpy-27 function in C. briggsae dosage compensation, but only sdc-2 also functions in sex determination. Thus, the two master regulatory genes that control both sex determination and X-chromosome dosage compensation are conserved between C. briggsae and C. elegans.
DCC recruitment sites isolated from C. briggsae X chromosomes fail to bind the C. elegans DCC
Discovery that the dosage compensation machinery and the gene regulatory hierarchy that control sex determination and dosage compensation are functionally conserved between C. briggsae and C. elegans raised the question of whether the cis-acting regulatory sequences that recruit dosage compensation proteins to X chromosomes are also conserved. In C. elegans, the DCC binds to recruitment elements on X (rex) sites and then spreads across X to sequences lacking autonomous recruitment ability (Csankovszki et al., 2004; Jans et al., 2009; Pferdehirt et al., 2011; Albritton et al., 2017; Anderson et al., 2019). Within rex sites, combinatorial clustering of three DNA sequence motifs directs synergistic binding of the DCC (Fuda et al., 2022). To compare X-recruitment mechanisms between species, DNA binding sites for the Cbr DCC recruitment protein SDC-2 and the Cbr DCC condensin subunit DPY-27 were defined by chromatin immuno-precipitation experiments followed by sequencing of captured DNA (ChIP-seq experiments) (Figure 4A). SDC-2 sites were obtained with anti-FLAG antibodies from a genome-engineered Cbr strain encoding a FLAG-tagged version of endogenous SDC-2. DPY-27 sites were obtained from either a wild-type Cbr strain with DPY-27 antibodies or from a genome-engineered strain encoding endogenous FLAG-tagged DPY-27 with anti-FLAG antibodies.
A consistent set of twelve large, overlapping SDC-2 ChIP-seq peaks and DPY-27 ChIP-seq peaks emerged from the studies (Figure 4A), representing less than one-fourth the number of DCC peaks than on the C. elegans X chromosome, which is smaller (17.7 Mb for Cel vs. 21.5 Mb for Cbr). Confirmation that each of these twelve Cbr peaks represented an actual DCC binding site resulted from DCC recruitment assays conducted in vivo (Figure 4B). Embryos carrying extrachromosomal arrays composed of multiple copies of DNA from a single ChIP-seq peak were stained with DPY-27 antibodies and a FISH probe to the array. Typically, DPY-27 localized to 80-90% of extrachromosomal arrays carrying DNA from an individual peak (Figure 4C and 4E and Figure 4—table supplement 1A). In contrast, extrachromosomal arrays made from three regions of X lacking DCC binding in ChIP-seq experiments showed minimal recruitment (0-6% of nuclei with arrays) (Figure 4E and Figure 4—table supplement 1A). In strains with arrays comprised of Cbr DCC binding sites, the X chromosomes rarely exhibited fluorescent signal, because the arrays titrated the DCC from X (Figure 4C). The titration was so effective that brood sizes of array-bearing hermaphrodites were very low, and hermaphrodite strains carrying arrays could not be maintained. Cbr DCC binding sites that were confirmed by array assays were named recruitment elements on X (rex sites) (Table 1), like the C. elegans DCC binding sites, due to their ability to recruit the DCC when detached from X.
To determine whether rex sites from C. briggsae and C. elegans had functional overlap in DCC binding specificity, we asked whether a rex site from one species could recruit the DCC from the other. We made extrachromosomal arrays in C. elegans with DNA from C. briggsae rex sites and extrachromosomal arrays in C. briggsae with DNA from C. elegans rex sites. Arrays in C. elegans with C. briggsae rex sites failed to recruit the Cel DCC or to titrate the Cel DCC from Cel X chromosomes (Figure 4C, Cbr rex-8), indicating evolutionary divergence in rex sites between the two Caenorhabditis species. Reciprocally, extrachromosomal arrays made in C. briggsae with Cel rex sites failed to bind the Cbr DCC or titrate it from the Cbr X, confirming divergence in rex sites (Cel rex-33 in Figure 4D; Cel rex-33 and Cel rex-4 in Figure 4—table supplement 1B). In contrast, controls showed that 100% of extrachromosomal arrays made in C. elegans with DNA from either Cel rex-33 or Cel rex-4 recruited the Cel DCC (Figure 4—table supplement 1B).
Because X chromosomes and extrachromosomal arrays have different topologies, histone modifications, DNA binding proteins, and positions within nuclei, we devised a separate assay to assess the divergence of rex sites between species in a more natural chromosomal environment. We inserted six Cbr rex sites with a range of ChIP-seq scores into a location on the endogenous Cel X chromosome that lacked DCC binding (15, 574, 674 bp) (Figure 5 and Table 1). Proof of principle for the experiment came from finding that insertion of Cel rex-32, a high-affinity Cel DCC binding site, into the new location on X resulted in DCC binding that was not significantly different from binding at its endogenous location on X (p = 0.2, Figure 5). All Cbr rex sites except rex-1, which will be discussed later, failed to recruit the Cel DCC when inserted into Cel X chromosomes, confirming the divergence of rex sites between species.
Identification of motifs on Cbr X chromosomes that recruit the Cbr DCC
To understand the mechanisms underlying the selective recruitment of the Cbr DCC to X chromosomes, but not autosomes, and the basis for the divergence in X-chromosome targeting between Caenorhabditis species, we searched for DNA sequence motifs that are enriched in the twelve Cbr rex sites (Figure 6—figure supplement 1A) using the website-based program called Multiple Em fo Motif Elicitation (MEME) (Version 5.4.1) (Bailey and Elkan, 1994; Bailey et al., 2015) and compared them to motifs in C. elegans rex sites important for recruiting the Cel DCC to X (Figure 6A,B). We found two motifs enriched within Cbr rex sites that are highly enriched on Cbr X chromosomes compared to autosomes (Figure 6A; Figure 7A,B; Table 1). A 13-bp motif named MEX (Motif Enriched on X) is enriched up to 12-fold on X chromosomes versus autosomes, and a 30-bp motif named MEX II is enriched up to 30-fold on X versus autosomes (Figure 7A,B). The similarity of a motif to the consensus motif is indicated by the ln(P) score, which is the natural log of the probability that the 13-mer for MEX or the 30-mer for MEX II matches the respective consensus motif matrix as calculated by the Patser program (Hertz and Stormo, 1999). The lower the score, the better the match. For both MEX and MEX II, the lower the ln(P) score, and hence the better the match to the consensus sequence, the more highly enriched is the motif on X chromosomes compared to autosomes (Figure 7A,B).
Our analysis revealed that only the Cbr MEX (Figure 7C) or MEX II (Figure 7D) motifs on X that are located within rex sites are bound by SDC-2. Negligible SDC-2 binding was found at single, isolated MEX (Figure 7C) or MEX II (Figure 7D) motifs on X that are not in rex sites. These results implicate MEX and MEX II as important elements for Cbr DCC recruitment to rex sites.
Neither of the Cbr motifs is enriched on the X chromosomes of C. elegans, indicating motif divergence between species (Figure 7A,B). No additional enriched C. briggsae motif candidates were found when the sequences of the two motifs in the twelve rex sites were eliminated from the search by converting them to N’s and searches for potential motifs were conducted again. In addition, motif analysis of DNA from SDC-2 and DPY-27 ChIP-seq peaks with intermediate or low levels of DCC binding (i.e., lower than for rex-2) (Figure 6—figure supplement 1B) revealed no motif candidates that correlate with DCC binding.
In C. elegans, two motifs are highly enriched on X chromosomes relative to autosomes: a 12-bp motif also called MEX and a 26-bp motif called MEX II (Figure 6B) (Fuda et al., 2022). These C. elegans X-enriched motifs are not enriched on C. briggsae X chromosomes (Figure 6B and Figure 7--figure supplement 7A,B). Cbr MEX as well as Cel MEX and Cel MEX II share a common core sequence of CAGGG (Figure 6), which is necessary but not sufficient for DCC binding in C. elegans (Fuda et al., 2022). The core is likely indicative of a common evolutionary history between species. However, the Cbr and Cel motifs diverged sufficiently that the motifs from one species are not enriched on the X chromosomes of the other species. Moreover, the Cbr MEX motif has a nucleotide substitution that would render the Cel MEX motif incapable of binding the Cel DCC. Predominantly, the C. elegans consensus MEX motif has a cytosine nucleotide located two nucleotides 5’ to the core CAGGG sequence: 5’-TCGCGCAGGGAG-3’ (Figure 6B). Mutational analysis in C. elegans demonstrated that replacing that nucleotide with a guanine greatly reduced DCC binding both in vivo and in vitro (Fuda et al., 2022). The consensus Cbr MEX motif has a guanine at that critical location, and in principle, the Cbr MEX motif would not function as a Cel DCC binding motif (Figure 6), thereby offering insight into the divergence of X-chromosome binding sites between species.
In C. elegans, a 9-bp motif called Motif C also participates in Cel DCC recruitment to X but lacks enrichment on X (Figure 6B) (Fuda et al., 2022). Sequences between the clustered Motif C variants within a Cel rex site are also critical for DCC binding (Fuda et al., 2022). Evidence that C. elegans Motif C fails to participate in Cbr DCC recruitment to Cbr X chromosomes is our finding that Cbr SDC-2 binding is negligible at Cel Motif C variants on Cbr X, except in the case of rare variants (0.26% of all Cel Motif C variants on X) that are within bona fide MEX or MEX II motifs in Cbr rex sites (Figure 7— figure supplement 1C). The likely reason that Cbr rex-1 recruits the Cel DCC when inserted into Cel X chromosomes (Figure 5) is that each of the four Cbr MEX motifs includes a strong match to the consensus Cel Motif C (Figure 5 legend), and DNA sequences surrounding the Cel Motif C variants in Cbr rex-1 are highly conserved with the syntenic region of C. elegans, which includes Cel rex-34. Both Cel rex-34 and Cbr rex-1 are within coding regions of orthologous pks-1 genes. In contrast, Cbr rex-7 also contains Motif C variants but lacks necessary surrounding sequences to permit Cel DCC binding when inserted on the Cel X (Figure 5).
Mutational analysis of motifs on endogenous C. briggsae X chromosomes showed that combinatorial clustering of motifs in rex sites facilitates Cbr DCC binding but some binding can still occur with only a single motif in a rex site
To assess further the importance of the Cbr motifs and the divergence of motifs between species, we performed mutational analyses of the two Cbr X-enriched motifs. Initial demonstration that both Cbr MEX and Cbr MEX II motifs participate in DCC binding at Cbr rex sites in C. briggsae came from analysis using extrachromosomal arrays carrying wild-type and mutant forms of Cbr rex-1 (Figure 8—figure supplement 1). Eighty-nine percent of C. briggsae nuclei carrying extrachromosomal arrays composed of wild-type rex-1 sequences recruited the DCC and titrated it away from X. In contrast, only 24% of nuclei carrying arrays with mutant copies of rex-1 lacking MEX II recruited the DCC, demonstrating the importance of MEX II. Only 38% of nuclei carrying arrays with mutant copies of rex-1 lacking all four MEX motifs recruited the DCC, demonstrating the importance of MEX. DCC binding was reduced to 6% of arrays carrying mutant copies of rex-1 lacking both MEX II and the four MEX motifs. Hence, both motifs contribute to DCC binding.
This conclusion was reinforced by using genome editing to mutate the MEX II sequence or all MEX II and MEX sequences in the endogenous rex-1 site on C. briggsae X chromosomes and then assaying DCC binding (Figure 8). ChIP-seq analysis revealed significant reduction in DPY-27 binding at rex-1 lacking MEX II sequences and negligible DPY-27 binding at rex-1 lacking both MEX and MEX II sequences. Hence, clustering of motifs in the endogenous rex-1 on X is important for DCC binding (Figure 8).
To evaluate more precisely the participation of different Cbr motifs in DCC binding, we used genome editing at three endogenous rex sites to evaluate the interplay between MEX and MEX II motifs, only MEX II motifs, or only MEX motifs. Eliminating either MEX or MEX II in rex-4 reduced binding significantly, but binding was evident at the remaining motif (Figure 9 and Figure 9—figure supplement 1). Binding was dramatically reduced when both motifs were mutated. This result demonstrates that an individual MEX or MEX II motif can confer significant DCC binding at a rex site, but both motifs are needed for full DCC binding.
Equivalent results were found by mutating either of the two MEX II motifs in rex-3 or combinations of the three MEX motifs in rex-7. For rex-3, DCC binding was reduced significantly when one of the two MEX II motifs was mutated, but significant binding occurred at either of the remaining MEX II motifs (Figure 10 and Figure 10—figure supplement 1). Binding was greatly reduced when both motifs were mutated. For rex-7, DCC binding at the endogenous site lacking the MEX motif with the best match to the consensus sequence (−18.22) was not significantly different from binding at the wild-type site. In contrast, mutating different combinations of two motifs (−18.72 and −12.26 or −18.7 and −12.58) reduced binding significantly (Figure 11 and Figure 11—figure supplement 1). Mutating all three motifs reduced binding severely. Results with the four Cbr rex sites, rex-1, rex-3, rex-4, and rex-7 demonstrate that combinatorial clustering of motifs achieves maximal DCC binding at Cbr rex sites, but significant binding can occur at a single motif.
These results contrast with results in C. elegans. Mutating individual motifs, either MEX, MEX II, or Motif C, at an endogenous C. elegans rex site with multiple different motifs dramatically reduced DCC binding in vivo to nearly the same extent as mutating all motifs, demonstrating synergy in DCC binding (Fuda et al., 2022). Hence, not only have the motifs diverged between species, the mode of binding to motifs has also changed.
Functional divergence of motifs demonstrated by Cel DCC binding studies in vivo and in vitro to a Cel rex site with Cbr MEX and MEX II motifs replacing Cel motifs
To explore the divergence in motifs between species in greater detail, we replaced each of the two MEX II motifs of the endogenous Cel rex-39 site on X with a copy of MEX II from Cbr rex-3 and assayed the level of Cel SDC-3 binding in vivo by ChIP-qPCR (Figure 12A,B). SDC-3 binding in vivo was negligible at the Cel rex-39 site with the Cbr MEX II motifs and indistinguishable from binding at the Cel rex-39 site with two scrambled MEX II motifs, thus demonstrating the high degree of functional divergence between MEX II motifs of different species (Figure 12B).
We performed a similar analysis for MEX motifs. We replaced the three MEX motifs in endogenous Cel rex-33 with the three Cbr MEX motifs from endogenous Cbr rex-7 (Figure 12D). SDC-3 binding in vivo was negligible at the Cel rex-33 site with the Cbr MEX motifs and indistinguishable from binding at the Cel rex-33 site with three scrambled MEX motifs, demonstrating the functional divergence between MEX motifs of different species (Figure 12E).
As a second approach, we conducted DCC binding studies in vitro. We assayed Cel DCC binding in vitro to a Cel rex-39 site with two Cbr MEX II motifs (Figure 12C) and to the Cel rex-33 site with the three Cbr MEX motifs (Figure 12F). Our prior studies developed a robust in vitro assay using C. elegans embryo extracts to evaluate the role of motif combinations toward Cel SDC-2 binding at rex sites (Fuda et al., 2022). The advantage of this assay is that Cel SDC-2 is capable of binding to a single motif on an in vitro template, perhaps because that DNA lacks the competing binding of nucleosomes and general transcription factors that occurs in vivo (Fuda et al., 2022). If either of the Cbr MEX II motifs inserted into the Cel rex-39 site were functional or if any of the three Cbr MEX motifs inserted into the Cel rex-33 site were functional, we would detect Cel SDC-2 binding to the template in vitro.
The in vitro assay demonstrated robust binding of Cel SDC-2 to the wild-type Cel rex-39 template (Figure 12C) and to the wild-type Cel rex-33 template (Figure 12F), as shown previously (Fuda et al., 2022). However, Cel SDC-2 binding at the Cel rex-39 site with substituted Cbr MEX II motifs was indistinguishable from binding to the mutant Cel rex-39 template with two scrambled Cel MEX II motifs or to the negative control template made from Cel X DNA at a site lacking Cel DCC binding in vivo (Figure 12C). Similarly, Cel SDC-2 binding at the Cel rex-33 site with substituted Cbr MEX motifs was indistinguishable from binding to the mutant Cel rex-33 template with three scrambled Cel MEX motifs or to the negative control template (Figure 12F). Thus, the in vitro assay demonstrates that substituting Cbr MEX II or MEX motifs for Cel MEX II or MEX motifs in a Cel rex site eliminates Cel DCC binding.
A single nucleotide position in the consensus Cbr MEX motif acts as a critical determinant for whether the Cel DCC can bind to Cbr MEX
In contrast to the many nucleotide changes that mark the difference between MEX II motifs in C. briggsae versus C. elegans, the MEX motifs are strikingly similar in nucleotide composition and core CAGGG sequence between species (Figure 6). The most significant change between the consensus MEX motifs is the substitution in Cbr MEX of a guanine for the cytosine in Cel MEX located two nucleotides 5’ from the CAGGG core of both motifs (Figure 13A). That C4G transversion was not found in a functional Cel MEX motif in vivo or in vitro, and creating a C4G change in either the MEX motif of endogenous Cel rex-1 or in an in vitro Cel DNA template reduced binding (Fuda et al., 2022). In principle, that single cytosine-to-guanine transversion could be a critical evolutionary change in MEX motifs that render the motifs incapable of binding the DCC from the other species. To test this hypothesis, we made the C4G transversion in each of the three MEX motifs within the endogenous Cel rex-33 site (Figure 13B). Cel SDC-3 binding in vivo to the C4G-substituted Cel rex-33 site was reduced to the same level of binding as that at the Cel rex-33 site with all three Cel MEX motifs scrambled, confirming the functional significance of the nucleotide substitution between species (Figure 13B). Our in vitro assay comparing Cel SDC-2 binding to the C4G-substituted and the MEX-scrambled Cel rex-33 DNA templates produced the same result (Figure 13C).
If the evolutionary transversion of that C to G between Cel and Cbr MEX motifs represents an important step in the divergence of motif function, then making a G-to-C change within the Cbr MEX motifs (G7C) inserted into Cel rex-33 might enhance Cel DCC binding. Indeed, the G7C change to Cbr MEX within Cel rex-33 increased the Cel SDC-3 binding in vivo by 4.2-fold and increased the specific Cel SDC-2 binding in vitro by 4.3-fold. The G7C change increased Cel SDC-3 binding in vivo to 18% of its binding at wild-type Cel rex-33 (Figure 13B) and increased Cel SDC-2 binding in vitro to 44% of its the specific binding at the wild-type Cel rex-33 template (Figure 13C). Hence, the cytosine-to-guanine transversion between MEX motifs of C. elegans versus C. briggsae is important for the functional divergence in motifs.
Discussion
Comparison of X-chromosome dosage compensation mechanisms between the closely related Caenorhabditis species C. briggsae and C. elegans revealed that both the dosage compensation machinery and the regulatory hierarchy that directs it to hermaphrodite X chromosomes have been conserved, but remarkably, the X-chromosome target specificity of the C. briggsae machinery and its mode of binding to X have diverged, as well as the density of DCC recruitment sites. The extent of evolutionary changes in dosage compensation mechanisms between species diverged by only 15-30 MYR is in striking contrast to changes in developmental processes such as body-plan specification and eye morphogenesis from fruit flies to mice, which instead utilize highly conserved transcription factors and cis-acting DNA regulatory sequences (Malicki et al., 1990; Halder et al., 1995).
Central to the dosage compensation machinery of both species is a specialized condensin complex. Here we identified two C. briggsae dosage compensation proteins (DPY-27 and MIX-1) that are orthologs of the SMC (structural maintenance of chromosome) subunits of condensin and bind to hermaphrodite X chromosomes. As in C. elegans (Chuang et al., 1994; Lieb et al., 1998), mutation of dpy-27 causes hermaphrodite-specific lethality in C. briggsae, and MIX-1 fails to bind to X in the absence of DPY-27, consistent with both proteins acting in a complex. We also found that the hermaphrodite-specific Cbr sdc-2 gene triggers binding of the condensin subunits to X and activates the hermaphrodite mode of sexual differentiation, as in C. elegans. Mutation of Cbr sdc-2 causes XX-specific lethality, and rare XX animals that escape lethality develop as masculinized larvae. SDC-2 and condensin subunits are prevented from binding to the single X of males by the action of xol-1, the master sex-determination gene that controls both sex determination and dosage compensation and triggers the male fate by repressing sdc-2 expression. Mutation of xol-1 kills XO animals because the DCC assembles on the single male X, thereby reducing gene expression inappropriately. Mutations in sdc-2 or dpy-27 suppress the XO-specific lethality caused by xol-1 mutations, but only mutations in sdc-2 permit the rescued animals to develop as males. Just as in C. elegans, XO animals rescued by dpy-27 mutations develop as hermaphrodites, consistent with dpy-27 controlling only dosage compensation and sdc-2 controlling both sex determination and dosage compensation. Hence, the two master regulators that control sexual fate and dosage compensation are functionally conserved between the two Caenorhabditis species, as is the condensin dosage compensation machinery. In both species, SDC-2 recruits the condensin DCC subunits to X and is the likely protein to interact directly with X DNA. These 350 kDa proteins lack homology to proteins outside of Caenorhabditis, and the only predicted structural feature is a coiled-coil region.
Although the two species have conserved DCC machinery, the DCC binding sites have diverged, as has their density on X. ChIP-seq analysis of C. briggsae SDC-2 and DPY-27 revealed twelve sites of binding on X that were validated by functional analysis in vivo as being strong autonomous recruitment (rex) sites. Even though the X chromosome of C. briggsae (21.5 Mb) is larger than the X of C. elegans (17.7 Mb), it has only one-fourth the number of recruitment sites. The C. briggsae sites are sufficiently strong that extrachromosomal arrays carrying multiple copies of a single site can titrate the DCC from X and cause dosage-compensation-defective phenotypes in XX animals, including death, as in C. elegans. In contrast, extrachromosomal arrays of C. briggsae rex sites made in C. elegans fail to recruit the C. elegans DCC, and vice versa, indicating that rex sites have diverged between the two species. As a more rigorous test of divergence, we inserted individual C. briggsae rex sites in single copy into C. elegans X chromosomes and assayed binding. The C. elegans DCC failed to bind to the five C. briggsae rex sites inserted into C. elegans X chromosomes.
Not only have the rex sites diverged, the mechanism by which the Cbr DCC binds to X motifs differs from that of the Cel DCC. We identified two motifs within C. briggsae rex sites that are highly enriched on X, the 13-bp MEX motif and the 30-bp MEX II motif. Mutating one copy of either motif in endogenous rex sites with multiple motifs reduced binding, but significant binding still occurred at the sites. Binding was eliminated only when all motifs were removed. Hence, DCC binding to motifs in C. briggsae rex sites appears additive. In contrast, mutating one motif in C. elegans rex sites that have multiple different combinations of motifs reduced binding to nearly the same extent as mutating all motifs, indicating synergy in C. elegans DCC binding (Fuda et al., 2022).
Additional factors, such as yet-unidentified DNA binding proteins might alter the specificity of DCC binding between species as well as aid DCC binding at Cbr rex sites. Precedent exists in the homeodomain family of Hox DNA binding proteins that have remarkably similar DNA specificities for DNA binding in vitro but a wide range of specificities in vivo due to interactions with heterologous cofactors required for functional specificity, such as Pbx-Hox complexes (Chang et al., 1996).
The need for synergy in DCC binding to Cel rex sites is likely caused by competition between DCC binding and nucleosome formation, since nucleosomes preferentially bind to rex sites when DCC binding is precluded by mutations (Fuda et al., 2022). The status of nucleosomes on C. briggsae X chromosomes remains to be determined. Although a single MEX or MEX II motif enables some DCC binding to a Cbr rex site, equivalent motifs on X that are not in rex sites appear to lack DCC binding. Nucleosome formation may preclude DCC binding at those motifs. The X may have a paucity of DNA-binding proteins that interact with core histones and open compacted chromatin to enable DCC binding.
Although the X-chromosome motifs of both species share the core consensus sequence CAGGG, the motifs have diverged such that they function in only C. elegans or C. briggsae. This functional divergence was demonstrated through DCC binding studies in vivo and in vitro to C. elegans rex sites engineered with C. briggsae motifs substituted for C. elegans motifs. We replaced the two MEX II motifs in the endogenous C. elegans rex-39 site with C. briggsae MEX II motifs and the three MEX motifs in Cel rex-33 with Cbr MEX motifs while maintaining motif spacing appropriate for C. elegans. We found negligible C. elegans DCC binding in vivo and in vitro. A feature of the in vitro assay is that Cel SDC-2 is capable of binding to a single motif on a DNA template, likely because the DNA lacks competing binding of nucleosomes that occurs in vivo. If either Cbr MEX II or MEX motif were functional in C. elegans we would have detected binding.
While the MEX II motif has diverged sufficiently that evolutionary tracing is difficult, divergence of MEX motifs provides important insight into their evolution. A major difference in MEX motifs between the two species is the preference for a guanine instead of a cytosine two nucleotides 5’ of the conserved CAGGG sequence. We demonstrated that converting that C to G in the three Cel MEX motifs of Cel rex-33 eliminated DCC binding in vitro. Conversely, replacing the G nucleotide in each Cbr MEX motif inserted into Cel rex-33 with a C nucleotide restored Cel DCC binding in vivo and in vitro, indicating that the single nucleotide change can be important in the evolutionary divergence of this motif. The evolutionary C-to-G substitution in the Cbr MEX motif is sufficient to prevent it from functioning in the closely related C. elegans species.
Like many developmental regulatory proteins, such as Drosophila Dorsal (Schloop et al., 2020), the DCC controls hundreds of genes through its action on cis-acting target sites. However, the DCC acts in a single developmental process, the control of X-chromosome dosage compensation, while the more typical developmental regulators participate in multiple, unrelated developmental processes. For such multi-purpose proteins, target site specificity is evolutionarily constrained: protein function is changed far more by changes in the number and location of conserved cis-acting target sequences than by changes in the target sequences themselves (Carroll, 2008; Nitta et al., 2015). Hence, it was of considerable interest to know whether the nematode DCC complex with multiple targets but lacking the constraints of such pleiotropy, would exhibit a different pattern of evolution. Our results show that it does: clearly the target specificity of the DCC has changed dramatically over the 15-30 MYR that separate C. elegans and C. briggsae. This divergence of X-chromosome target sites could have been an important factor for nematode speciation.
Materials and methods
Procedures for mutant isolation
Procedures for sdc-2 mutant isolation were described previously in (Wood et al., 2011). xol-1(y430), dpy-27(y436), and mix-1(y435) were isolated from a C. briggsae deletion library provided by E. Haag using primers listed in Supplementary File 2. Resulting strains are listed in Supplementary File 1.
Protein Sequence Alignment of SDC-2
Sequence alignment of the Caenorhabditis elegans SDC-2 (UniProtKB G5EBL3) and Caenorhabditis briggsae SDC-2 (Uniprot A8XQT3) was generated using Clustal Omega (Madeira et al., 2022) and ESPript 3.0 server (https://espript.ibcp.fr) (Robert and Gouet, 2014). The coiled-coil annotations were predicted using the web server version of DeepCoil (Ludwiczak et al., 2019), part of the MPI Bioinformatics Toolkit (Zimmermann et al., 2018; Gabler et al., 2020).
Preparation of FISH probes
Chromosome FISH probes were prepared from 1 mg of total DNA, which included multiple C. briggsae BACs listed in Supplementary File 3 (BACPAC Resources Center, CHORI, Oakland, CA). BACs were purified using the QIAGEN midiprep kit (catalog number 12243). Chromosomal FISH probes were made with the Invitrogen DNA FISH-tag kit. X-chromosome probes (10 BACS covering approximately 5% of the chromosome) were labeled with AlexaFluor 594 (Molecular Probes, F32949), and chromosome III probes (3 BACS covering approximately 1% of the chromosome) were labeled with AlexaFluor 488 (Molecular Probes, F32947).
Preparation of gut nuclei for FISH and immunofluorescence
Adult worms were dissected in 4 µl egg buffer (25 mM HEPES, pH 7.4, 118 mM NaCl, 48 mM KCl, 0.2 mM CaCl2, 0.2 mM MgCl2) on a 18 mm X 18 mm coverslip. Four µl of 4% formaldehyde (in egg buffer) were added, and the solution was mixed by tapping the coverslip before it was placed onto a Superfrost/Plus glass slide (Fisherbrand, 12-550-15). Fixed samples were incubated for 5 min at room temperature in a humid chamber, then frozen in liquid nitrogen for at least 1 min. Coverslips were removed quickly with a razor blade, and slides were placed immediately into PBS-T (PBS with 1mM EDTA and 0.5% Triton X-100). Slides were subjected to three 10-min washes in PBS-T at room temperature. Slides were dehydrated in 95% ethanol for 10 min at room temperature followed by either the FISH or immunofluorescence protocol below.
FISH
Following dehydration of the slides, excess ethanol was removed, 15 µl of hybridization solution (50% formamide, 3X SSC, 10% dextran sulfate, 10 ng labeled DNA probe in water) were added, and a coverslip was placed on each slide. Slides were placed into a slide chamber, and the FISH incubation protocol was conducted in a PCR machine overnight (80 °C for 10 min, 0.5 °C/second to 50 °C, 50 °C for 1 min, 0.5 °C/second to 45 °C, 45 °C for 1 min, 0.5 °C/second to 40 °C, 40 °C for 1 min, 0.5°C/second to 38 °C, 38 °C for 1 min, 0.5 °C/second to 37 °C, 37 °C overnight). After overnight incubation at 37 °C, slides were washed at 39 °C using the following regime: three times (15 min each) in 2X SSC (0.3 M NaCl and 30 mM Na3C6H5O7) in 50% formamide, three times (10 min each) in 2X SSC in 25% formamide, three times (10 min each) in 2X SSC, and three times (1 min each) in 1X SSC. Samples were incubated in PBS-T for 10 min at room temperature, and immunofluorescence staining was performed as described below.
Immunofluorescence of gut nuclei
Following dehydration of slides subjected to immunofluorescence only or to PBS-T treatment (after FISH protocol), the excess liquid was removed (either ethanol from dehydration step, or PBS-T from FISH protocol) and 20 µl of affinity-purified primary antibodies (Cbr-DPY-27 and Cbr-MIX-1 peptide antibodies [Covance, Inc.]) in PBS-T were added at 1:200 dilution. Samples were incubated in a humid chamber for between 4h and overnight. Slides were washed three times (10 min each) in PBS-T at room temperature and then incubated in secondary antibodies for 3-6 h. Slides were washed three times (10 min each) in PBS-T at room temperature before Prolong (Molecular Probes, P36934) with DAPI (1 µg/ml) was added, and the samples were imaged using a Leica TCS SP2 AOBS. Antibodies used: anti-DPY-27 rabbit antibody raised to Cbr-DPY-27 C-terminal peptide DVQSEAPSAGRPVETDREGSYTNFD, anti-DPY-27 guinea pig antibody raised to the same Cbr-DPY-27 peptide, anti-MIX-1 rabbit antibody raised to Cbr-MIX-1 C-terminal peptide EATKKPSKKSAKKAVQNTDDEME, Alexa Flour 488 goat anti-rabbit antibody (Molecular Probes, A11034), Alexa Flour 488 goat anti-guinea pig antibody (Molecular Probes, A11073), and Alexa Flour 594 goat anti-rabbit antibody (Molecular Probes, A11037).
Immunofluorescence of embryos
Embryos were picked into 4 µl of water on poly-lysine-treated slides. After adding a coverslip, slides were frozen in liquid nitrogen for at least 1 min. Coverslips were removed rapidly with a razor blade and samples were dehydrated in 95% ethanol for 10 min. Next, 40 µl of fix solution (2% paraformaldehyde in egg buffer) were added and slides were incubated in a humid chamber for 10 min. Slides were washed three times (10 min each) in PBS-T at room temperature. Antibody staining was performed as described above for gut nuclei.
Immunoprecipitation analysis
Immunoprecipitation with DPY-27 antibodies followed by MALDI mass spectrometry of trypsinized protein bands excised from an SDS-PAGE gel was performed according to (Mets and Meyer, 2009).
Calculation of viability for C. briggsae sdc-2 mutants
XX animals: sdc-2 (X) / + hermaphrodites were crossed to JU935 males, which carry a gfp transgene integrated on the X chromosome, and the hermaphrodite cross progeny (sdc-2 + / + gfp) were moved to individual plates. Three classes of genotype are expected among the self-progeny of sdc-2 + / + gfp hermaphrodites. Two classes, (+ gfp / + gfp and sdc-2 + / + gfp) express GFP, whereas the third class, (sdc-2 + / sdc-2 +) does not. If -sdc-2 + / sdc-2 + animals are 100% viable, the expected proportion of non-green animals among the self-progeny of sdc-2 + / + gfp hermaphrodites is 25%. In each case, the expected number of viable non-green adult progeny is shown in parentheses, and the observed proportion is depicted in the chart as a percentage of the expected number. Wild-type XX viability was calculated among the self-progeny of + + / + gfp animals.
XO animals: sdc-2 + / + gfp hermaphrodites were crossed with + + / O (wild-type) males. Successfully mated hermaphrodites were identified by the presence of a copulatory plug and then moved to individual plates. Two classes of genotype were expected among the progeny of this cross. One class (+ gfp / O) expresses GFP, whereas the other (sdc-2 + / O) does not. If sdc-2 + / O animals are 100% viable, the expected proportion of non-green animals among the male progeny is 50%. In each case, the expected number of non-green animals is shown in parentheses, and the observed proportion is depicted in the chart as a percentage of the expected number. Wild-type XO viability was calculated among the male cross-progeny of + + / + gfp hermaphrodites and + + / O males.
Calculation for rescue of xol-1 XO-specific lethality in C. briggsae by an sdc-2 mutation
The percent viability of wild-type XO animals and mutant XO animals carrying combinations of xol-1 and sdc-2 mutations was calculated by formulae that follow. For wild-type XO or xol-1(y430) XO progeny from crosses of wild-type or xol-1(y430) hermaphrodites mated with wild-type males, the formula is [(number of F1 males)/(total F1 progeny/2)] x 100, a calculation that assumes successful mating and the potential for 50% male cross progeny among the F1. For xol-1 sdc-2 XO double mutants, xol-1-sdc-2 / xol-1 hermaphrodites were mated with wild-type males. Given that xol-1 XO progeny are inviable, xol-1 sdc-2 F1 males should make up 1/3 of viable F1s. Thus, % XO rescue is calculated as [(number of males)/(total progeny/3)] x 100.
Genome Editing Using CRISPR-Cas9
The Cbr rex-1 (Figure 8), Cbr rex-3 (Figure 10), Cbr rex-4 (Figure 9) and Cbr rex-7 (Figure 11) mutations, as well as Cel site 2 insertions (Figure 5) and substitutions of Cbr MEX motifs into Cel rex-33 and substitution of Cbr MEX II motifs into Cel rex-39 (Figure 12) were made with the CRISPR-Cas9 co-conversion technique using Cas9 RNP injections and species-appropriate co-injection markers (Farboud et al., 2019). C. elegans editing utilized the dpy-10 roller marker, and C. briggsae editing utilized the ben-1 marker. The tracrRNA and crRNA guides (Dharmacon) were resuspended in 600 μM of nuclease-free water (Ambion AM9937). The Cas9 RNP mixture for injections included 5 μl Cas9 protein (UC Berkeley QB3 MacroLab, 10 mg/ml), 1.15 μl 2M HEPES, pH 7.5, 0.35 μl 0.5 M KCl, 0.5 μl 600 μM dpy-10 crRNA, 1 μl target crRNA (Table S3), 5μl tracrRNA, and 7 μl nuclease-free water. The Cas9 RNP mix was incubated at 37 °C for 15 min, and 1 μl of the resulting Cas9 RNP mix was combined with 0.5 μl 10 μM dpy-10 repair oligo (IDT), 0.5 μl 10 μM rex repair oligo (IDT), and 8 μl nuclease-free water. After centrifuging at 16,100 x g for 10 min, the Cas9 RNP mix was injected into gonads of adult hermaphrodites. The target-specific sequences for Cas9 guide RNAs are listed in Supplementary File 4. The DNA sequences for the repair templates are listed in Supplementary File 5.
For C. elegans, injected adults were placed on NGM plates. After 3 days of growth at 25 °C, progeny with the roller phenotype were picked to individual plates and allowed to lay embryos. The roller parents were picked into lysis buffer, and the edited site was amplified and sequenced to identify the worms that were edited. The homozygous progeny from properly edited worms were backcrossed twice to wild-type (N2) worms before usage in experiments. For C. briggsae, mutants were isolated as published (Farboud et al., 2019). The homozygous progeny from those were backcrossed twice to AF16 worms before usage in experiments. Primers used for genotyping are listed in Supplementary File 2.
C. briggsae ChIP Extract Preparation
Mixed-stage animals were grown on MYOB agar plates with concentrated HB101 bacteria at 20 °C. Animals were cross-linked with 2% formaldehyde for 10 min and quenched with 100 mM Tris-HCl, pH 7.5. Cross-linked animals were resuspended in 1 ml of FA Buffer (150 mM NaCl, 50 mM HEPES-KOH, pH 7.6, 1 mM EDTA, 1% Triton X-100, 0.1% sodium deoxycholate, 1 mM DTT, and protease inhibitor cocktail [Calbiochem, #539134]) for every 1 gram of animals. This mixture was frozen in liquid nitrogen and then ground under liquid nitrogen by mortar and pestle for 3 min. Once thawed, the mixture was then homogenized with 50 strokes in a Dounce homogenizer. The chromatin was sheared using the Covaris S2 (20% duty factor, power level 8, 200 cycles per burst) for a total of 30 min processing time (60 sec ON, 45 sec OFF, 30 cycles). The concentration of protein in each extract was quantified using the BCA assay (Thermo Fisher, #23228).
C. elegans ChIP-seq Extract Preparation
Mixed-stage embryos were harvested from hermaphrodites grown on MYOB agar plates with concentrated HB101 bacteria at 20 °C. Embryos were cross-linked with 2% formaldehyde for 10 min and quenched with 100 mM Tris-HCl, pH 7.5. Cross-linked embryos were resuspended in 1 ml of FA Buffer (150 mM NaCl, 50 mM HEPES-KOH, pH 7.6, 1 mM EDTA, 1% Triton X-100, 0.1% sodium deoxycholate, 1 mM DTT, and protease inhibitor cocktail [Calbiochem, #539134]) for every 1 gram of embryos and homogenized with 50 strokes in a Dounce homogenizer. The chromatin was sheared using the Covaris S2 (20% duty factor, power level 8, 200 cycles per burst) for a total of 30 min processing time (60 sec ON, 45 sec OFF, 30 cycles). The concentration of protein in each extract was quantified using the BCA assay (Thermo Fisher, #23228).
C. briggsae ChIP Reactions
To perform the ChIP reactions, a 50 µl bed volume of protein A Dynabeads (Thermo Fisher, #10001D) was re-suspened in 1 ml of FA Buffer (150 mM NaCl, 50 mM HEPES-KOH, pH 7.6, 1 mM EDTA, 1% Triton X-100, 0.1% sodium deoxycholate, 1 mM DTT, and protease inhibitor cocktail [Calbiochem, #539134]). The beads were incubated in a microcentrifuge tube with 5 µg of anti-FLAG antibodies (Sigma-Aldrich, #F1804) and 5 µg of rabbit anti-mouse IgG antibodies (Jackson ImmunoResearch, #315-005-003), or 5 µg of mouse IgG (Sigma-Aldrich, #I5381) and 5 µg of rabbit anti-mouse IgG antibodies (Jackson ImmunoResearch, #315-005-003), for 90 min at room temperature. Tubes with incubated beads were placed on a magnetic rack, and the liquid was discarded.
Extracts containing 2 mg of protein ChIPs were increased in volume to 1 ml with FA buffer and then added to each tube of Dynabeads for a for 90-min incubation. The Dynabead-extract mixture was washed at room temperature twice with FA Buffer (150 mM NaCl), once with FA Buffer (1 M NaCl), once with FA Buffer (500 mM NaCl), once with TEL buffer (10 mM Tris-HCl, pH 8.0, 250 mM LiCl, 1% IGEPAL CA-630 [Sigma-Aldrich, #I3021], 1% sodium deoxycholate, 1 mM EDTA), and twice with TE Buffer (10 mM Tris, pH 8.0, 1 mM EDTA). Protein and DNA were eluted with 250 µl of buffer (1% SDS, 250 mM NaCl, 1mM EDTA) at 65 °C for 20 min.
C. elegans ChIP Reactions
To perform the ChIP reactions, a 25 µl bed volume of protein A Dynabeads (Thermo Fisher, #10001D) was re-suspened in 1 ml of FA Buffer (150 mM NaCl, 50 mM HEPES-KOH, pH 7.6, 1 mM EDTA, 1% Triton X-100, 0.1% sodium deoxycholate, 1 mM DTT, and protease inhibitor cocktail [Calbiochem, #539134]). The beads were incubated in a microcentrifuge tube with 3 µg rabbit anti-SDC-3 (lab stock), or 3 µg rabbit IgG (Jackson Immunoresearch, #301-005-003) for 90 min at room temperature. Tubes with incubated beads were placed on a magnetic rack and liquid was discarded. Protocols for the incubation of extract with beads and elution of protein and DNA from beads were the same as those described for C. briggsae ChIP reactions.
ChIP-seq, Illumina Sequencing, and Data Processing
Sequencing libraries were prepared with the eluted materials from ChIP reactions as published (Zhong et al., 2010) with minor changes: sequencing adapters were obtained from Bioo (NEXTflex), and adapters were ligated using the NEB Quick Ligation Kit (M2200). Libraries were sequenced on the Illumina HiSeq 4000 platforms. After barcode removal, reads were aligned uniquely to the C. briggsae CB4 genome using the default settings in Bowtie version 2.3.4.3. To account for read depth, ChIP signal was normalized to the total number of reads that uniquely aligned to the genome.
C. elegans qPCR
To perform qPCR reactions, protein and DNA from a C. elegans ChIP reaction or from 50% of a control extract (1 mg protein) were de-crosslinked at 65 °C for at least 4 h with 150 μg/ml Proteinase K (Sigma, #3115887001). DNA from each ChIP reaction or from control extract was isolated using the Qiagen PCR purification kit and diluted to a final volume of 200 μl with (10 mM Tris-HCl, pH 8.5). For quantitative PCR, the immunoprecipitated DNAs were quantified by comparing their threshold cycle to the standard curve from control DNA (10% and 3 serial 10-fold dilutions). For the site 2 insertions, the DCC levels at each inserted rex site were calculated for each biological replicate as a ratio of the average DCC level at 5 control rex sites (rex-8, rex-16, rex-32, rex-35, and rex-48). For all experiments involving endogenous Cel rex-39 in Figure 12B or involving endogenous Cel rex-33 in Figure 12E, the DCC levels at each inserted rex site were calculated for each biological replicate as a ratio of the average DCC level at 7 control rex sites (rex-8, rex-14, rex-16, rex-32, rex-35, rex-36, and rex-48). Primers used for qPCR are listed in Supplementary File 2.
C. briggsae qPCR
To perform the qPCR reactions, protein and DNA from a C. briggsae ChIP reaction or from 50% of a control extract (1 mg protein) were de-crosslinked at 65 °C for at least 4 h with 150 μg/ml Proteinase K (Sigma, #3115887001). DNA from each ChIP reaction or from control extract was isolated using the Qiagen PCR purification kit and diluted to a final volume of 400 μl with (10 mM Tris-HCl, pH 8.5). For quantitative PCR, the immunoprecipitated DNAs were quantified by comparing their threshold cycle to the standard curve from control DNA (10% and 3 serial 10-fold dilutions). For the endogenous rex site mutations, the DCC levels at each inserted rex site were calculated for each biological replicate as a ratio of the average DCC level at 4 control rex sites (rex-1, rex-2, rex-5, and rex-9). Primers used for qPCR are listed in Supplementary File 2.
Identification of C. briggsae DCC binding motifs
The 500 bp DNA sequence centered on each C. briggsae SDC-2 ChIP-seq peak location for the 12 Cbr rex sites was isolated from the CB4 reference genome. Motif candidates were obtained by inputting twelve 500 bp sequences onto MEME on the MEME-suite website (Bailey and Elkan, 1994; Bailey et al., 2015). The settings used to identify motif candidates were the classic mode and any number of repetitions (anr). The X:A enrichment was calculated for motif candidates. The two motif candidates enriched on the Cbr X chromosomes were named Cbr MEX for the 13-bp motif and Cbr MEX II for the 30-bp motif (Figure 6).
X:A Fold Motif Enrichment Calculation
The Patser program (version 3e) (Hertz and Stormo, 1999) was used to calculate the natural log of the probability (ln[P]) of finding a match to the Cbr MEX motif, Cbr MEX II, Cel MEX motif, and Cel MEX II motif at all positions along each chromosome, as explained in (Fuda et al., 2022). For each threshold value, the number of motifs with ln[P] values less than the value (better match) was summed for X and for autosomes. The number of autosomal motifs was divided by the total number of autosomal base pairs to find the number of motifs per base pair. The number of motifs per base pair of X was calculated similarly. The final X:A ratio was calculated by dividing the motifs per base pair for X by the motifs per base pair for the autosomes.
C. elegans DCC binding assay performed in vitro
The in vitro Cel DCC binding assays (Figure 12 and Figure 13) were performed as described previously in (Fuda et al., 2022). The sequences for the DNA templates used in the in vitro DCC binding assays are listed in Supplementary File 6.
Data
The accession number for the ChIP-seq data reported in this paper is GEO GSE214714.
Acknowledgements
We are grateful to E. Haag and his laboratory for generously providing expertise and reagents to make and screen C. briggsae deletion pools, A. Wood for initiating ZFN mutagenesis of C. briggsae sdc-2, D. Stalford for figure preparation, T. Cline and laboratory members for valuable discussions, and the QB3 Genomics Facility (RRID:SCR_022170) for DNA sequencing. This work was supported in part by NIH Grant R35 GM131845 (to B.J.M.). B.J.M. is an investigator of the Howard Hughes Medical Institute.
Footnotes
↵• Roche Diagnostics 2881 Scott Blvd. Santa Clara, CA 95050
Competing interests: The authors declare that no competing interests exist.