Summary
Biochemical studies of chromatin have typically used either artificial DNA templates with unnaturally high affinity for histones, or small genomic DNA fragments deprived of their cognate physical environment. It has thus been difficult to dissect chromatin structure and function within fully native DNA substrates. Here, we circumvent these limitations by exploiting the minimalist genome of the eukaryote Oxytricha trifallax, whose notably small ~3kb chromosomes mainly encode single genes. Guided by high-resolution epigenomic maps of nucleosome organization, transcription, and DNA N6-methyladenine (m6dA) locations, we reconstruct full-length Oxytricha chromosomes in vitro and use these synthetic facsimiles to dissect the influence of m6dA and histone post-translational modifications on nucleosome organization. We show that m6dA directly disfavors nucleosomes in a quantitative manner, leading to local decreases in nucleosome occupancy that are synergistic with histone acetylation. The effect of m6dA can be partially reversed by the action of an ATP-dependent chromatin remodeler. Furthermore, erasing m6dA marks from Oxytricha chromosomes leads to proportional increases in nucleosome occupancy across the genome. This work showcases Oxytricha chromosomes as powerful yet practical models for studying eukaryotic chromatin and transcription in the context of biologically relevant DNA substrates.
Highlights
De novo synthesis of complete, epigenetically defined Oxytricha chromosomes
Epigenomic profiles of chromatin organization in Oxytricha’s miniature chromosomes
m6dA directly disfavors nucleosome occupancy in natural and synthetic chromosomes
Histone acetylation and chromatin remodelers temper the impact of m6dA on chromatin
Introduction
Nucleosomes are the fundamental repeating unit of eukaryotic chromatin, consisting of ~146bp DNA wrapped around an octameric core of histones. Nucleosomes limit the physical accessibility of DNA to trans-acting factors, and thus directly impact DNA-based transactions, such as transcription, DNA repair, and replication. The in vitro reconstitution of chromatin is a method central to understanding these processes at the molecular level. Currently, the most widely used DNA template for such experiments is the 147bp “601” nucleosome positioning sequence. It exhibits a ~374 fold higher affinity for histone octamers than native genomic sequences (Lowary and Widom, 1998; Thåström et al., 2004), allowing the preparation of consistent, defined nucleosome arrays in vitro. Yet the biological relevance of 601 DNA remains unclear, due to its unnaturally high affinity for histones. On the other hand, reconstituting chromatin from native genomic DNA is challenging because the long length of chromosomes increases the propensity for non-specific aggregation during chromatin assembly (Kaplan et al., 2009). As a compromise, small genomic fragments are usually prepared via PCR amplification or mechanical shearing. However, such substrates are separated from their cognate physical environment, as defined by the totality of the chromosome.
Eukaryotic genomes exhibit enormous natural variation in form and function. This is exemplified by unicellular ciliates, which possess two structurally and functionally distinct nuclei within each cell (Prescott, 1994; Yerlici and Landweber, 2014). In Oxytricha trifallax, the germline micronucleus is transcriptionally silent and contains ~100 megabase-sized chromosomes, similar to widely studied model organisms (Chen et al., 2014). In contrast, the somatic macronucleus is transcriptionally active, being the sole locus of Pol II-dependent RNA production in non-developing cells (Khurana et al., 2014). The Oxytricha macronuclear genome is extraordinarily fragmented, consisting of ~16,000 unique chromosomes with a mean length of ~3.2kb, most encoding a single gene. Chromosome ends are capped with a compact 36nt telomere consisting of 5’-(T4G4)-3’ repeats, bound cooperatively by a heterodimeric protein complex (Gottschling and Zakian, 1986; Horvath et al., 1998). Macronuclear chromatin yields a characteristic ~200bp ladder upon digestion with micrococcal nuclease, indicative of regularly spaced nucleosomes (Gottschling and Cech, 1984; Lawn et al., 1978; Wada and Spear, 1980). Yet it remains unknown how and where nucleosomes are organized within these miniature chromosomes, and how transcriptional control is orchestrated in their context. Chromatin organization in Oxytricha offers a model system to shed light on these fundamental questions, and also opens exciting possibilities for constructing complete chromosomes with defined molecular composition. Such ‘designer’ chromosomes can allow investigation of histone and DNA modifications in the context of fully native DNA.
Ciliates have long served as powerful models for the study of chromatin modifications (Brownell et al., 1996; Liu et al., 2007; Strahl et al., 1999; Taverna et al., 2002; Wei et al., 1998). They also hold promise for the study of DNA methylation - in particular, N6-methyladenine (m6dA). This modification has recently been implicated in diverse biological processes in eukaryotes, including retrotransposon regulation, transgenerational epigenetic inheritance, and gene activation. m6dA is abundant in the macronuclear genomes of ciliates (0.18 - 2.5% m6dA / dA) (Ammermann et al., 1981; Cummings et al., 1974; Gorovsky et al., 1973; Rae and Spear, 1978), similar to the green algae Chlamydomonas (0.3 - 0.5%) (Fu et al., 2015; Hattman et al., 1978). High levels of m6dA (up to 2.8%) were also recently reported in basal fungi (Mondo et al., 2017). m6dA is present at very low levels in metazoa, such as C. elegans (0.01-0.4%), Drosophila (0.001-0.07%), Xenopus laevis (0.00009%), and mouse (0.0006-0.007%) (Greer et al., 2015; Koziol et al., 2015; Wu et al., 2016; Zhang et al., 2015), although it accumulates to a high level (0.1-0.2%) during vertebrate embryogenesis (Liu et al., 2016). Ciliates offer ideal systems for probing m6dA function, given the high abundance of m6dA and the availability of genetic and biochemical tools.
Intriguingly, in green algae and the ciliate Tetrahymena, m6dA is enriched in nucleosome linker regions (Fu et al., 2015; Hattman et al., 1978; Karrer and VanNuland, 1999; Pratt and Hattman, 1981, 1983; Wang et al., 2017), suggesting a role for m6dA in chromatin organization, or vice versa. Yet the functional relationship between m6dA and nucleosomes - if any - has remained unclear. Does m6dA directly disfavor nucleosomes, and if so, is it a graded or binary effect? Does the presence of histone post-translational modifications and ATP-dependent chromatin remodelers - both integral components of chromatin in vivo - modulate this interaction?
Here we address these questions by building synthetic, epigenetically defined chromosomes. Specifically, we generate chromosomes that either lack m6dA, or contain the modification at positions identical to their in vivo configuration. We also prepare chromosomes with bona fide telomeres that either lack or contain telomeric protein complexes. Using this library of synthetic chromosomes, we show that m6dA directly disfavors nucleosome occupancy in a quantitative, site-specific manner. Furthermore, this effect is modulated by histone post-translational modifications and chromatin remodelers, and is similar in magnitude to that imposed by telomeric protein complexes. Together, we demonstrate the utility of Oxytricha chromosomes as a versatile platform for functionally dissecting epigenetic modifications in native DNA.
Results
Epigenomic profiles of chromatin and transcription in Oxytricha
We generated genome-wide in vivo maps of nucleosome positioning, transcription, and m6dA in the Oxytricha macronuclear genome using MNase-seq, poly(A)+ RNA-seq, 5’-complete cDNA-seq, and single molecule real time (SMRT) sequencing (Figure 1). The presence of m6dA in Oxytricha DNA was independently validated by mass spectrometry (Figure S1). Oxytricha transcription start sites (TSSs) localize within 100bp of chromosome ends, upstream of a phased nucleosome array (Figure 1A). Strikingly, m6dA is enriched in three consecutive nucleosome depleted regions directly downstream of TSSs. Each cluster contains varying densities of m6dA (Figure 1B), with a maximum of 22 sites in the second cluster (Table S1A). Highly transcribed chromosomes tend to bear more m6dA, suggesting a positive role of this DNA modification in gene regulation (Figure 1C). Moreover, the majority of methylation was found on both DNA strands within an ApT motif (Figures 1D and 1E). m6dA occurs on sense and antisense strands with approximately equal frequency, indicating that the methylation machinery does not function strand-specifically.
m6dA localization in nucleosome linker regions of highly transcribed genes is deeply conserved across microbial eukaryotes
To better understand the conservation and evolutionary significance of m6dA, we also performed high coverage SMRT sequencing and mass spectrometry validation (Figures S1 and S2) in the ciliate Tetrahymena, which diverged from Oxytricha over 1 billion years ago (Bracht et al., 2013). The genome architecture of Tetrahymena is drastically different from Oxytricha, with chromosomes several orders of magnitude larger, spanning tens of kilobases to megabases in length (Coyne et al., 2008; Eisen et al., 2006). We find that m6dA is similarly enriched in nucleosome linker regions in Tetrahymena, consistent with earlier reports (Gorovsky et al., 1973; Hattman et al., 1978; Karrer and VanNuland, 1999; Pratt and Hattman, 1981). m6dA occurs in an ApT dinucleotide motif in both Tetrahymena and Oxytricha (Figures 1 and S2; also Bromberg et al., 1982), suggesting that the underlying enzymatic machinery responsible for m6dA deposition is conserved. Actively transcribed genes in Tetrahymena possess higher levels of m6dA, despite transcription start sites being distant from chromosome ends (Figure S2). m6dA is thus associated with transcribed DNA templates rather than proximity to telomeres per se. While m6dA patterns are broadly similar between Oxytricha and Tetrahymena, we note that its peak density is quantitatively different in Tetrahymena genes, and is considerably further downstream of TSSs than in Oxytricha (Figure 1A and Table S1A). Given that green algae possess a generally similar m6dA distribution and methylation motif as Oxytricha and Tetrahymena (Fu et al., 2015), we conclude that conserved mechanisms underlie m6dA establishment and function in ciliates and unicellular plants.
m6dA directly disfavors nucleosome occupancy across the genome
Most strikingly conserved across the m6dA patterns of Oxytricha, Tetrahymena, and green algae is the inverse correlation between m6dA and nucleosome positioning in vivo. However, SMRT-seq data alone do not indicate causality. To test this directly, we exploited the naturally fragmented architecture of the Oxytricha macronuclear genome to amplify complete chromosomes using PCR. This erases all cognate m6dA, while fully preserving DNA sequence and physical linkage within each chromosome. We selected 98 unique chromosomes that collectively reflect overall genome properties, including AT content, chromosome length and transcriptional activity (Figure 2A; Table S1B). Only high copy number chromosomes were selected to ensure high-confidence identification of m6dA marks. Full-length chromosomes were individually PCR-amplified from genomic DNA, resulting in the collective erasure of 2,344 m6dA marks. Each chromosome was purified and subsequently mixed together in stoichiometric ratios to obtain a “minigenome” (Figure 2B). Native genomic DNA (containing m6dA) and amplified minigenome DNA (lacking m6dA) were each assembled into chromatin in vitro using Xenopus or Oxytricha histone octamers (Figures S3 and S4) and analyzed using MNase-seq. We computed nucleosome occupancy from the native genome and minigenome samples across 199,795 overlapping DNA windows, spanning all basepairs in the 98 chromosomes. This allowed the direct comparison of nucleosome occupancy in each window of identical DNA sequence, with and without m6dA (Figures 2C and 2D). Windows indeed exhibit lower nucleosome occupancy with increasing m6dA, confirming the quantitative nature of this effect. Furthermore, similar trends were observed for both native Oxytricha and recombinant Xenopus histones, suggesting that the effects of m6dA on nucleosome organization arise mainly from intrinsic features of the histone octamer rather than from species-specific variants (Figure 2C and 2D).
Modular synthesis of an epigenetically defined chromosome
In principle, we reasoned that Oxytricha chromosomes could be constructed de novo via ligation of individual DNA building blocks, themselves generated in large quantities through PCR. The introduction of epigenetic modifications onto oligonucleotides before ligation would localize them to desired sites in the chromosome. We developed a streamlined chromosome synthesis scheme involving consecutive restriction enzyme digestion, ligation, and size selection steps (see Methods). Using this approach, we built a completely synthetic chromosome in vitro with a fully native DNA sequence, containing m6dA at all sites detected by SMRT-seq in vivo. We used this construct to dissect the effect of m6dA on nucleosome occupancy. The representative chromosome, Contig1781.0, is 1.3kb and contains a single highly transcribed gene with a clearly defined TSS (Figure 3A) - features characteristic of typical Oxytricha chromosomes. We independently validated the location of m6dA in vivo by sequencing chromosomal DNA immunoprecipitated with an anti-m6dA antibody (Figure 3A).
Four chromosome variants were synthesized, with cognate m6dA sites on neither, one, or both DNA strands (chromosomes 1 - 4 in Figures 3B, 3C, and S5A). Each chromosome was cloned and sequenced to verify the accuracy of construction (Figure S5C). With these chromosomal DNA templates in hand, we investigated the impact of m6dA on nucleosome occupancy. Chromatin was assembled by salt dialysis and subsequently digested with MNase to obtain mononucleosomal DNA (Figure 4A and S3). Tiling qPCR was used to quantify nucleosome occupancy at ~50bp increments along the entire length of the synthetic chromosome (Figure 4B). The fully methylated locus exhibits a ~46% reduction in nucleosome occupancy relative to the unmethylated variant, while hemimethylated chromosomes containing half the number of m6dA marks showed intermediate nucleosome occupancy at the corresponding region (Figures 4B and 4C). The reduction in nucleosome occupancy was confined to the methylated region, and not observed across the rest of the chromosome. We therefore conclude that m6dA directly disfavors nucleosome occupancy in a local, quantitative manner.
Chromatin remodelers partially restore nucleosome occupancy over m6dA sites
Nucleosome occupancy in vivo is influenced not only by DNA sequences, but also by trans-acting factors. ATP-dependent chromatin remodeling factors modulate nucleosome organization and help establish canonical nucleosome patterns near TSSs (Struhl and Segal, 2013). We used our synthetic methylated chromosomes to test how the well-studied chromatin remodeler ACF responds to m6dA in native DNA. ACF generates regularly spaced nucleosome arrays in vitro and in vivo (Clapier and Cairns, 2009; Ito et al., 1997). Its catalytic subunit ISWI is conserved across eukaryotes, including the ciliates Oxytricha and Tetrahymena (Figure S6). Synthetic chromosomes were assembled into chromatin by salt dialysis as before, then incubated with ACF in the presence of ATP (Figure S3D). We find that ACF partially - but not completely - restores nucleosome occupancy over the methylated locus in an ATP-dependent manner (Figure 4D and 4E). Chromatin remodelers may thus modulate nucleosome occupancy in concert with m6dA, each imposing distinct but interrelated effects.
Histone acetylation and m6dA exert synergistic effects on nucleosome occupancy
Nucleosomes are decorated with post-translational modifications (PTMs) in vivo, which collectively modulate chromatin structure and function. A well-documented example is lysine acetylation, particularly at histone H3 lysine 56 (H3K56ac) and at multiple residues in the histone H4 N-terminal tail (poly-acH4). H3K56 lies at the entry and exit sites of a nucleosome, and its acetylation lowers the affinity of H3-H4 tetramers for DNA (Andrews et al., 2010) and increases DNA unwrapping (Simon et al., 2011), together leading to nucleosome destabilization. H4 polyacetylation reduces the net positive charge of histone octamers, weakening histone-DNA contacts (Hong et al., 1993). It also hinders chromatin compaction and nucleosome aggregation, indicating a role in regulating higher order chromatin structure (Allahverdi et al., 2011).
Since m6dA is embedded within chromatin in vivo and likely occurs in the context of histone PTMs, we asked whether the effects of m6dA on nucleosome positioning are themselves modulated by PTMs such as H3K56ac and poly-acH4, which influence chromatin structure. Using quantitative mass spectrometry, we verified that Oxytricha histones contain H3K56ac and poly-acH4 in vivo, along with numerous other sites of acetylation on H3 and H4 (Figure S7). We prepared recombinant H3K56ac and semisynthetic poly-AcH4 using amber codon suppression and native chemical ligation, respectively (Figure S4). The modified histones were refolded into octamers in parallel with unmodified controls and subsequently assembled into chromatin on the methylated chromosomes. Different histone octamer types exhibited qualitatively similar patterns of nucleosome organization across each chromosome (Figure 4B). Curiously, poly-acH4-containing octamers gave rise to higher nucleosome occupancy near the center of the chromosome relative to flanking regions, despite a weaker affinity of these octamers for DNA per se. Since the central region is intrinsically favorable for nucleosome formation (occupancy is highest in this region, even for unmodified octamers), it may be less sensitive to decreases in octamer affinity compared to flanking regions. We then computed the fold-change in nucleosome occupancy in methylated chromosomes relative to the unmethylated control of identical DNA sequence. This calculation was performed separately for each octamer type. Chromatin assembled with H3K56ac or poly-acH4 exhibited a significantly greater reduction in nucleosome occupancy than unmodified octamers, for fully methylated DNA (Figure 4C). Therefore, H3K56ac and poly-acH4 can act synergistically with m6dA to disfavor nucleosome occupancy. These data broadly reflect the complex interplay between histone PTMs and m6dA in modulating chromatin structure. In our model, H3K56ac and poly-acH4 may not actually localize near m6dA within this specific chromosome per se, but we propose that histone PTMs that alter chromatin structure can work in concert with m6dA to modulate nucleosome organization.
Telomere proteins and m6dA disfavor nucleosome occupancy to similar extents
The synthetic chromosomes described thus far contain blunt telomeric ends, but Oxytricha chromosome termini in vivo possess 16nt single-stranded 3’ DNA tails, necessary for associating with the telomere end-binding protein complex, TeBPα and TeBPβ, together similar in mass to a histone octamer. TeBPα/β bind cooperatively to the single-stranded telomeric tail to form a ternary complex (Kd = 2nM2), stable even in 2M NaCl (Gottschling and Zakian, 1986; Horvath et al., 1998). To determine whether telomere protein binding at chromosome termini influences nucleosome occupancy, and to compare its effects to m6dA, we used our modular synthesis scheme to build synthetic chromosomes with bona fide 3’ tails (chromosomes 6 - 8 in Figure 3B, 3C and S5B). Recombinantly expressed and purified Oxytricha TeBPα and TeBPβ (Figure 5A) were both shown, in methylation protection assays, to bind cooperatively to Oxytricha gDNA termini, yielding guanine residue protection patterns consistent with previous studies (Gray et al., 1991) (Figure 5C). TeBPα and TeBPβ were then pre-bound to synthetic chromosomes and subsequently used for chromatin assembly via salt dialysis (Figure 5D). Gel shift assays confirmed that both subunits remain associated with synthetic chromosome ends in 2M NaCl, the highest salt concentration in the chromatin assembly procedure (Figures 5E and 5F). We also verified that telomere protein binding occurs independently at 5’ and 3’ chromosome termini (Figures 5E and 5F). The synthetic chromosome pre-bound with TeBP proteins at both termini exhibits a 40-50% decrease in nucleosome occupancy, within 50bp of each chromosome end (Figure 5F). This region directly overlaps with 32.1% of transcription start sites in the genome (Figure 1A), and may thus influence promoter chromatin accessibility. Indeed, TeBP binding has been reported to modulate transcription initiation in Euplotes crassus (Bender and Klein, 1997), a ciliate with similar chromosome architecture to Oxytricha. The observed reduction in nucleosome occupancy upon TeBP binding is quantitatively similar to that imposed by fully methylated chromosomal loci (Figure 4F). Therefore, both telomere protein binding and m6dA deposition sculpt nucleosome organization, though m6dA exerts a graded rather than all-or-none effect, depending on the number of m6dA marks present.
Discussion
We report that m6dA directly disfavors nucleosome occupancy and that this effect can be modulated by histone post-translational modifications and ATP-dependent chromatin remodelers. We expect the biochemical impact of m6dA to be directly pertinent across a wide diversity of eukaryotic genomes, including vertebrates, C. elegans, Drosophila and fungi, where this epigenetic modification has recently been documented (Greer et al., 2015; Koziol et al., 2015; Liu et al., 2016; Mondo et al., 2017; Wu et al., 2016; Zhang et al., 2015). The current experiments do not reveal exactly how m6dA disfavors nucleosome occupancy. Early studies suggest that N6-methylation destabilizes dA:dT base pairing, leading to a decrease in the melting temperature of DNA (Engel and von Hippel, 1978). Whether this or some other physico-chemical property of m6dA contributes to lowered nucleosome stability awaits further investigation.
While m6dA directly disfavors nucleosome occupancy, it may also be possible that nucleosomes in turn inhibit m6dA deposition by the putative methylase, establishing a positive feedback loop that reinforces the inverse relationship between nucleosome occupancy and DNA methylation. Aside from simple physical accessibility, the activity of the m6dA methylase may differ upon binding to nucleosomes, compared to naked DNA. Histone variants and PTMs commonly enriched near transcription start sites may also modulate the enzyme’s activity. Future identification of the ciliate m6dA methylase would shed light on these questions and advance our understanding of how nucleosomes and DNA methylation interact to establish chromatin architecture near TSSs.
What could be the identity of the m6dA methylase in ciliates? While typical Dam and DMNT-like DNA methylases are absent from the Oxytricha and Tetrahymena genomes, there are 5-6 candidate genes with predicted MT-A70 domains, homologous to the METTL gene family (Iyer et al., 2016; Luo et al., 2015). Although some mammalian MT-A70 proteins are known to catalyze m6A methylation on RNA, they may have been co-opted to deposit methyl groups on DNA substrates in ciliates. Functional perturbations of these candidates in vivo would test such predictions. Equally intriguing is the observation that actively transcribed genes possess high levels of m6dA. This trend is deeply conserved, being present in the distantly related ciliates Tetrahymena and Oxytricha, as well as green algae and basal fungi. It is possible that the m6dA methylase is associated with RNA polymerases, resulting in m6dA deposition during transcriptional elongation. Alternatively, transcription factors may contain “reader” domains that specifically recognize m6dA, thus increasing transcriptional output at methylated loci. Importantly, these two scenarios are not mutually exclusive. We envision that the use of synthetic Oxytricha chromosomes, in conjunction with transcriptionally competent nuclear extracts, would constitute an especially useful biochemical tool for dissecting such effects.
More broadly, our study showcases the utility of Oxytricha chromosomes in advancing chromatin biology. Each chromosome essentially comprises a nucleosome array, capped with telomeric protein complexes at both ends. Here we show how these features can be reconstructed in their entirety using synthetic chromosomes. By extending current technologies (Müller et al., 2016), it should be feasible to introduce both modified nucleosomes and DNA methylation in a site-specific manner on full-length chromosomes. Such ‘designer’ constructs will serve as powerful tools for studying DNA-templated processes such as transcription within the context of a fully native DNA environment.
Author contributions
L.Y.B. conceived the project, designed research, synthesized chromosomes, performed computational and experimental analysis for all Figures and Tables, and wrote the manuscript. G.T.D. designed research, synthesized chromosomes, and prepared all Xenopus histones. K.A.L. processed raw SMRT-seq data. K.K. and B.A.G performed mass spectrometry analysis of Oxytricha histones. E.R.H. performed m6dA IP-seq. J.R.B. prepared Oxytricha DNA for SMRT-seq. R.P.S performed SMRT-seq. T.W.M. and L.F.L. conceived the project, designed research and analyzed data. G.T.D, T.W.M, and L.F.L. edited the manuscript.
Table S1. Descriptive statistics
(A) Properties of m6dA distribution near TSSs. An m6dA site is classified as lying within a particular methyl cluster if it is within 50 bp of the peak derived from the aggregate m6dA distribution. Aggregate m6dA peak positions in Oxytricha are +56 bp, +235 bp, and +436 bp downstream of the TSS, while those in Tetrahymena are +205 bp, +400 bp, and +597 bp respectively.
(B) Properties of Oxytricha chromosomes in native genomic DNA and mini-genome DNA.
Table S2. Primer sequences
All primers are in the 5’ to 3’ direction.
Figure S1. Mass spectrometry analysis confirms the presence of m6dA in ciliate DNA
Oxytricha and Tetrahymena genomic DNA were digested into nucleosides and used for reverse-phase HPLC and subsequent mass spectrometry. Chemically synthesized dA and m6dA were used as standards, with 12 pmol and 0.3 pmol respectively loaded. Eluted peaks with expected masses of m6dA are detected in both Tetrahymena and Oxytricha nucleosides.
Figure S2. Genomic distribution of m6dA in Tetrahymena thermophila
MNase-seq data and RNA-seq data were obtained from previously published datasets (Beh et al., 2015; Xiong et al., 2012).
(A) Meta-chromosome plots overlaying MNase-seq (nucleosome positioning in vivo) and SMRT-seq (m6dA), relative to transcription start sites. m6dA lies mainly within nucleosome linker regions, between the +1, +2, and +3 nucleosomes.
(B) Frequency of m6dA modifications downstream of TSSs.
(C) Transcriptional activity is positively correlated with the total number of m6dA within the corresponding gene. RPKM denotes the number of reads per kilobase per million mapped reads.
(D) Composite analysis of 441,618 methylation sites reveals that m6dA occurs within an 5’-ApT-3’ dinucleotide motif in Tetrahymena, consistent with isotopic labeling experiments (Bromberg et al., 1982; Wang et al., 2017) and similar to Oxytricha.
(E) Distribution of various m6dA dinucleotide motifs across the genome.
(F) Organization of transcription, nucleosome organization, and m6dA in a single Tetrahymena gene.
Figure S3. Gel analysis of assembled chromatin
Xenopus or Oxytricha histone octamers were assembled on DNA through salt dialysis and subsequently digested with MNase to obtain monunucleosome-sized fragments. The resulting products were analyzed by agarose gel electrophoresis.
(A) Chromatin assembled with synthetic chromosome DNA. All assemblies were performed in the presence of an approximately 100-fold mass excess of buffer DNA relative to synthetic chromosome (see Methods). Representative assemblies with the unmethylated chromosome are shown. Methylated chromosome assemblies were separately performed in place of the unmethylated variant.
(B) Chromatin assembled on PCR-amplified mini-genome DNA.
(C) Chromatin assembled on native genomic DNA.
(D) Chromatin assembled on unmethylated synthetic chromosomes and incubated with ACF and ATP. Regularly spaced nucleosomes are observed only when ACF and ATP are present.
Figure S4. Preparation of histone octamers for chromatin assembly
Xenopus unmodified core histones were expressed recombinantly, while H3K56ac and poly-acH4 were synthesized through amber codon suppression and native chemical ligation, respectively. Oxytricha histones were acid-extracted from vegetative nuclei. Oxytricha and Xenopus histones were subsequently refolded into octamers and purified through size exclusion chromatography.
(A) Reverse-phase HPLC analysis of purified Xenopus poly-acH4.
(B) ESI mass spectrometry analysis of purified Xenopus poly-acH4.
(C) Reverse-phase HPLC analysis of purified Xenopus H3K56ac.
(D) ESI mass spectrometry analysis of purified Xenopus H3K56ac.
(E) Reverse-phase HPLC purification of acid-extracted Oxytricha histones. Fractions 15 were individually collected and analyzed by Coomassie staining and western blotting.
(F) SDS-PAGE analysis of purified Oxytricha histone fractions.
(G) Western blot analysis confirms identity of each Oxytricha histone fraction.
(H) SDS-PAGE analysis of purified histone octamers.
Figure S5. Schematic of chromosome synthesis strategy
Staggered dotted lines represent BsaI cleavage sites.
(A) Assembly of methylated chromosome variants.
(B) Assembly of chromosome variants with 3’ single-stranded telomeric tails.
(C) Sanger sequencing of cloned synthetic chromosomes. Continuous horizontal green bar represents full sequence identity between the reference chromosome and individual sequencing reads.
Figure S6. Putative ciliate ISWI orthologs
ISWI is a member of the SW12/SNF2 ATPase family that acts as chromatin remodelers. The Oxytricha and Tetrahymena genomes were queried by BLASTP using Drosophila melanogaster ISWI (UniProt ID: Q24368), and the reciprocal best hit was retrieved from each genome. BLASTP e-values were 0.0 in each case. Putative Tetrahymena and Oxytricha orthologs were queried for protein domains and associated GO terms using InterPro (Finn et al., 2017). ISWI contains an N-terminal catalytic ATPase domain, and a C-terminal HAND-SANT-SLIDE module necessary for nucleosome binding and mobilization.
Figure S7. Detection of poly-acH4 and H3K56ac in Oxytricha cells using quantitative mass spectrometry
(A) Middle-down mass spectrometry (MS) quantification of histone H3 variants. Histone H3 variants are listed along the y-axis, and are henceforth abbreviated as g60, g122, g137, g54, g33, and g10 respectively. x-axis represents total ion count (arbitrary units).
(B) Middle-down MS quantification of histone H3 (Contig4701.0.g33) acetylation. Data from four biological replicates are respectively shown. Positions of PTMs are listed along the x-axis. y-axis represents the cumulative abundance of acetylation on the modified Lys-residues as relative to the total histone H3. Each bar represents the averaged relative abundance (%) of 3 technical replicates (with exception of nt_Rep4=2); error bars represent ± standard deviation (stdev) of technical (nt_Rep1-3 = 3; nt_Rep4 = 2;) replicates.
(C) Bottom-up MS quantification of histone H3 acetylation. Positions of PTMs are listed along the x-axis. y-axis represents the cumulative abundance of acetylation on each residue. Histone peptides containing H3K56ac are KYQKSTELLIR (g122); KFQKSTELLIR (g10); KYQKSTDLLIR (g60); and RFQKSTELLIR (g33, g54, g137). Each bar represents the averaged relative abundance (%) of 4 biological replicates; nb = 4). Error bars represent ± standard deviation (stdev) of biological replicates (nb = 4).
(E) Bottom-up MS quantification of histone H4 acetylation. Positions of PTMs are listed along the x-axis. y-axis represents the cumulative abundance of acetylation on the four modified residues of the H4 peptide GKVGKGYGKVGAKR. The Oxytricha genome contains two annotated histone H4 genes with identical amino acid sequence.
(F) modified peptides from (d) as relative to total histone H4. Each bar represents the averaged relative abundance (%) of 4 biological replicates; nb = 4). Error bars represent ± standard deviation (stdev) of biological (nb = 4) replicates.
Figure S8. Tiling qPCR analysis of nucleosome occupancy in spike-in and homogeneous synthetic chromosome preparations
The blunt, unmethylated synthetic chromosome (construct #1 in Figure 3B) was used for chromatin assembly with (“Spike-in”) or without (“Homogeneous”) a hundred-fold excess of carrier DNA. In the latter case, an equivalent mass of synthetic chromosome was added in place of carrier DNA to maintain the same DNA concentration for chromatin assembly. Red and purple regions depict the corresponding regions where m6dA and TeBPα/β respectively modulate nucleosome occupancy in separate synthetic chromosomes studied in Figure 4. Black arrowheads indicate no decrease in nucleosome occupancy in these regions when carrier DNA is used.
Acknowledgments
We thank Geoffrey Dann for discussions regarding the ACF remodeler and for performing preliminary tests of its activity; Tharan Srikumar for assistance with nucleoside mass spectrometry; Wei Wang, Jessica Wiggins, and Jennifer Miller for assistance with Illumina sequencing; C. David Allis and Virginia Zakian for suggestions and comments on the project; Ana Mostafavi and Glen Liszczak for advice regarding amber codon suppression, and Jingmei Wang, Barbara Dul, and Fei Song for general laboratory support. This work was funded by NIH grants R01-GM59708 to L.F.L and R01-GM107047 to T.W.M.
Footnotes
↵7 Lead contact