Supplementary Information

Transcription

1 Supplementary Information Molecular determinants of nucleosome retention at CpG-rich sequences in mouse spermatozoa Serap Erkek, Mizue Hisano, Ching-Yeu Liang, Mark Gill, Rabih Murr, Jürgen Dieker, Dirk Schübeler, Johan van der Vlag, Michael B. Stadler and Antoine H.F.M. Peters Supplementary Figures Supplementary Figure 1: Supplementary Figure 2: Supplementary Figure 3: Supplementary Figure 4: Supplementary Figure 5: Supplementary Figure 6: Supplementary Figure 7: Genome-wide distribution of nucleosome peaks and motif enrichment analysis for nucleosome peaks. Nucleosome occupancy in Saccharomyces cerevisiae and human sperm. Chromatin states of imprinting control regions (ICRs) in sperm and round spermatids. Specificity of the H3.3 antibody and average profiles of canonical and variant histone occupancy along genes in relation to expression status in embryonic stem cells. Histone proteins and modification patterns in round spermatids, sperm and ESCs. Chromatin states of genes representative of the gene clusters described in Fig. 6c and of pluripotency associated genes in sperm and round spermatids. Reproducibility of ChIP experiments in sperm. Supplementary Tables Supplementary Table 1: GO-term analysis for the gene clusters described in Fig. 6c. Supplementary Note 1

2 a b c Top 20 motifs enriched CGCCGCG CGCGGCG GGCCGCG CGCGGCC CGCGCCG GCCCGCG CGGCGCG GCGGCGG CCGCCGC GCCGCGG CCGCGGC CGCGGGC GCGCGGG CCCGCGC CGCGACG CGGCGGC CGTCGCG CCGCGGG GCCGCCG Supplementary Figure 1: Genome-wide distribution of nucleosome peaks and motif enrichment analysis for nucleosome peaks. The mouse genome is classified into promoter (± 1 kb around TSS), exon, intron, repeat and intergenic (non-repeat) regions. (a) Observed and expected fractions of nucleosome peaks by genomic region. Observed refers to experimentally identified nucleosome peaks and expected is calculated assuming uniform distribution of the same number of peaks in the genome. (b) Barplot showing the data from (a) as log2 enrichments (observed/expected) of nucleosome peaks in different genomic regions. (c) Motif abundance versus enrichment plot showing the enrichment of 7-mer motifs in nucleosome peaks compared to background abundance (see Supplementary Methods for the description of background). Top 20 enriched motifs are shown next to the plot. 2

3 a b c Pearson correlation coefficient Pearson correlation coefficient Pearson correlation coefficient A C G T AA AC AG AT CA CC CG CT GA GC GG GT TA TC TG TT A C G T AA AC AG AT CA CC CG CT GA GC GG GT TA TC TG TT e Average DNA methylation (%) Nucleosome enrichment (log2) d Pearson correlation coefficient f Average DNA methylation (%) Nucleosome enrichment (log2) Supplementary Figure 2: Nucleosome occupancy in Saccharomyces cerevisiae and human sperm. (a) and (b) Correlation of single nucleotide frequencies (left) and dinucleotide frequencies normalized for single nucleotide composition (right) with nucleosome occupancy in Saccharomyces cerevisiae 1 in 1kb windows tiling the yeast genome based on (a) in vitro reconstituted nucleosomes and (b) in vivo determined nucleosomal occupancies (YPD medium). (c) and (d) Correlation of single nucleotide frequencies (left) and dinucleotide frequencies normalized for single nucleotide composition (right) with nucleosome occupancy in human sperm in 1kb windows tiling the human genome. (e) and (f) Correlation of human sperm DNA methylation 2 with nucleosome occupancy in human sperm in 1kb windows tiling the human genome. (c) and (e) display human sperm nucleosome data from Brykczynska et al. 3 while (d) and (f) display data from Hammoud et al. 4. 3

4 a b Round spermatids H3K4me3 H3K4me3 H3.1/H3.2 H3.3 Nucleosome CGI CpG Blastocyst Oocyte DNA methylation Supplementary Figure 3: Chromatin states of imprinting control regions (ICRs) in sperm and round spermatids. From top to bottom, images show the DNA methylation status for spermatozoa, oocytes and blastocyst embryos 5, CpG density, CGI localization, nucleosome, histone variant and histone modification states in sperm and histone modification states in round spermatids. Imprinting control regions 5 for paternally imprinted genes are shown with light blue boxes and imprinting control regions for maternally imprinted genes are shown with light pink boxes. (a) H19. (b) Rasgrf1. (c) Dlk1-Meg3. (d) Kcnq1ot1. (e) Nespas-Gnas. (f) Snrpn. (g) Peg10. DNA methylation states refer to percentage of methylation at individual CpGs as determined by shot gun bisulfite sequencing 5. Chromatin images represent read count per base averaged over a moving 300 bps interval. Averaged reads counts were normalized for total read counts across samples. 4

5 c Round spermatids H3K4me3 H3K4me3 H3.1/H3.2 H3.3 Nucleosome CGI CpG Blastocyst Oocyte DNA methylation 5

6 d e Round spermatids H3K4me3 H3K4me3 H3.1/H3.2 H3.3 Nucleosome CGI CpG Blastocyst Oocyte DNA methylation 6

7 f g Blastocyst Oocyte DNA methylation H3K4me3 H3.1/H3.2 H3.3 Nucleosome CGI CpG Round spermatids H3K4me3 Sgce Peg chr6 (Mbp) 7

8 H3.3-HA over H3.2-HA ratio (log2) 2.0 H3.3-HA enrichment (log2) 1.5 H3.2-HA enrichment (log2) b 3 a TSS TES Position around transcript (kb) Expression in embryonic stem cells: Not detected Low Medium High Supplementary Figure 4: Specificity of the H3.3 antibody and average profiles of canonical and variant histone occupancy along genes in relation to expression status in embryonic stem cells. (a) Western blots showing endogenous and exogenously expressed histone variant H3.3 as detected by H3.3 and tag antibodies. 293 cells were transfected with constructs encoding tagged canonical H3.1, H3.2 and variant H3.3 histones. (b) Average profiles of histone variants H3.2-HA and H3.3-HA in embryonic stem cells (ESC)6. Genes were classified according to expression status in ESC7. Transcripts without any aligned reads were classified as not detected. Remaining transcripts were classified on the basis of increasing expression values into three equally sized groups. In the bottom panel, the ratio between H3.3HA over H3.2-HA is shown. 8

9 a H3K4me3 enrichment in sperm (log2) H3K4me3 enrichment in sperm (log2) H3K4me3 enrichment in sperm (log2) enrichment in sperm (log2) b c H3K4me3 enrichment in RS (log2) H3K4me3 enrichment in RS (log2) H3K4me3 enrichment in RS (log2) H3K4me3 relative enrichment in ESC enrichment in sperm (log2) enrichment in sperm (log2) enrichment in RS (log2) enrichment in RS (log2) enrichment in RS (log2) relative enrichment in ESC Clusters of Fig. 6c Supplementary Figure 5: Histone proteins and modification patterns in round spermatids, sperm and ESCs. (a) Scatter plots showing the comparison of the enrichments of H3K4me3 (left) and (right) in round spermatids (RS) and sperm at TSS (± 1kb); (b) in relation to the five gene clusters as described and color coded in Fig. 6c; (c) in comparison to H3K4me3 8 and 7 levels in mouse ESCs. (d) Scatter plots showing the correlation between percentage of CpGs at TSS (± 1kb) and H3.3 over H3.1/H3.2 ratio in RS in relation to relative enrichment of H3K4me3 in RS (top left), in RS (top right), H3K4me3 in sperm (bottom left) and in sperm (bottom right). (e) H3.3 enrichments (left panel) and H3.1/H3.2 enrichments (right panel) were calculated genome-wide for 1kb windows which do not intersect TSS regions (± 3kb) and CGIs and were modeled as a linear combination of explanatory variables (CpG content and histone measurements in RS). The unique contribution of each variable to observed sperm variation is shown in percentage. Combinatorial effects refer to variation which is explained by combinations of variables used in the analysis. 9

10 d H3K4me3 (RS) (RS) H3.3 over H3.1/H3.2 ratio in RS (log2) H3K4me3 () () Relative enrichment Percentage of CpG (±1kb around TSS) e Explained variation (%) CpG% H3.3 H3.1/H3.2 H3K4me3 Round spermatids Combinatorial effects CpG% H3.3 H3.1/H3.2 H3K4me3 Round spermatids Combinatorial effects 10

11 a b Round spermatids H3K4me3 H3K4me3 H3.1/H3.2 H3.3 Nucleosome CGI CpG Blastocyst Oocyte DNA methylation Supplementary Figure 6: Chromatin states of genes representative of the gene clusters described in Fig. 6c and of pluripotency associated genes in sperm and round spermatids. From top to bottom, images show at various loci the DNA methylation status for spermatozoa, oocytes and blastocyst embryos 5, CpG density, CGI localization, nucleosome, histone variant and histone modification states in sperm and histone modifications in round spermatids. (a, b) examples for cluster 1, Sfrs6 and H3f3a. (c, d) examples for cluster 2, Rps14 and Dnajb1. (e, f) examples for cluster 3, Tsen2 and Hint3. (g, h) examples for cluster 4, T (Brachyury) and Gata2. (i, j) examples for cluster 5, Olfr family and Cts6. (k) Pou5f1 (l) Nanog (m) Sox2 (n) Esrrb (o) Klf4 (p) Klf5. DNA methylation states refer to percentage of methylation at individual CpGs as determined by shot gun bisulfite sequencing 5. Chromatin images represent read count per base averaged over a moving 300 bps interval. Averaged reads counts were normalized for total read counts across samples. 11

12 c d Round spermatids H3K4me3 H3K4me3 H3.1/H3.2 H3.3 Nucleosome CGI CpG Blastocyst Oocyte DNA methylation 12

13 e f Round spermatids H3K4me3 H3K4me3 H3.1/H3.2 H3.3 Nucleosome CGI CpG Blastocyst Oocyte DNA methylation 13

14 g h Round spermatids H3K4me3 H3K4me3 H3.1/H3.2 H3.3 Nucleosome CGI CpG Blastocyst Oocyte DNA methylation 14

15 i j Round spermatids H3K4me3 H3K4me3 H3.1/H3.2 H3.3 Nucleosome CGI CpG Blastocyst Oocyte DNA methylation 15

16 k l Round spermatids H3K4me3 H3K4me3 H3.1/H3.2 H3.3 Nucleosome CGI CpG Blastocyst Oocyte DNA methylation 16

17 m n Round spermatids H3K4me3 H3K4me3 H3.1/H3.2 H3.3 Nucleosome CGI CpG Blastocyst Oocyte DNA methylation 17

18 o p Round spermatids H3K4me3 H3K4me3 H3.1/H3.2 H3.3 Nucleosome CGI CpG Blastocyst Oocyte DNA methylation 18

19 a b ± 1kb TSS region Intersecting a peak Not intersecting a peak c d e f Nucleosome H3.3 H3.1/H3.2 H3K4me3 CpG rep1 rep2 rep3 rep1 rep2 rep1 rep2 rep1 rep2 rep1 rep2 Feature density Supplementary Figure 7: Reproducibility of ChIP experiments in sperm. (a - e) Scatter plots show the correlation of the replicates for nucleosome (a), H3.3 (b), H3.1/H3.2 (c) H3K4me3 (d) and (e) ± 1kb around TSS. TSS regions which intersect a peak are shown in red. For each histone variant / modification sample, peaks were identified in a similar way as described for nucleosomes. (f) Heatmap of genes illustrating CpG density, nucleosome (3 replicates), H3.3 (2 replicates), H3.1/H3.2 (2 replicates), H3K4me3 (2 replicates) and (2 replicates) coverage around TSS (± 3kb) in sperm. Feature density shows the scaled read densities from ChIP-seq experiments genes were randomly selected for visualization. 19

20 Supplementary Table 1: GO-term analysis for the gene clusters described in Fig. 6c. Top 50 terms associated with each cluster are shown. Cluster 1 GO.ID Term Annotated Significant Expected classicfisher Over-representation GO: protein modification by small protein conjugation or removal GO: spermatogenesis GO: male gamete generation GO: protein ubiquitination GO: cellular protein catabolic process GO: ubiquitin-dependent protein catabolic process GO: proteolysis involved in cellular protein catabolic process GO: protein modification by small protein conjugation GO: sexual reproduction GO: cellular macromolecule metabolic process GO: modification-dependent protein catabolic process GO: cellular protein metabolic process GO: modification-dependent macromolecule catabolic process GO: protein catabolic process GO: cellular macromolecule catabolic process GO: proteasomal ubiquitin-dependent protein catabolic process GO: gamete generation GO: proteasomal protein catabolic process GO: cellular metabolic process GO: protein polyubiquitination GO: spermatid development GO: protein folding GO: macromolecule catabolic process GO: spermatid differentiation GO: multicellular organism reproduction GO: multicellular organismal reproductive process GO: primary metabolic process GO: macromolecule metabolic process GO: catabolic process GO: ER-associated protein catabolic process GO: protein K63-linked ubiquitination GO: germ cell development GO: mrna transport GO: cellular catabolic process GO: protein metabolic process GO: protein K48-linked ubiquitination

21 GO: nucleic acid transport GO: RNA transport GO: establishment of RNA localization GO: RNA splicing GO: translational initiation GO: macromolecule modification GO: protein polyglycylation GO: cellular protein modification process GO: protein modification process GO: nucleobase-containing compound transport GO: inactivation of MAPK activity GO: reproduction GO: RNA localization GO: protein monoubiquitination Cluster 2 GO.ID Term Annotated Significant Expected classicfisher Over-representation GO: RNA splicing < 1e GO: mrna processing < 1e GO: mrna metabolic process < 1e GO: RNA processing < 1e GO: nuclear division < 1e GO: mitosis < 1e GO: organelle fission < 1e GO: M phase of mitotic cell cycle < 1e GO: DNA repair < 1e GO: M phase < 1e GO: chromosome organization < 1e GO: cell cycle phase < 1e GO: cell cycle process < 1e GO: mitotic cell cycle < 1e GO: DNA metabolic process < 1e GO: cell cycle < 1e GO: organelle organization < 1e GO: cellular protein metabolic process < 1e cellular component organization or biogenesis at cellular GO: level < 1e GO: nucleic acid metabolic process < 1e GO: cellular macromolecule metabolic process < 1e GO: cellular component organization at cellular level < 1e GO: nucleobase-containing compound metabolic process < 1e GO: gene expression < 1e

22 GO: RNA metabolic process < 1e GO: cellular nitrogen compound metabolic process < 1e GO: nitrogen compound metabolic process < 1e GO: macromolecule metabolic process < 1e GO: cellular metabolic process < 1e GO: primary metabolic process < 1e GO: metabolic process < 1e GO: protein metabolic process E GO: cell division E GO: protein modification by small protein conjugation or removal E GO: response to DNA damage stimulus E GO: ribonucleoprotein complex biogenesis E GO: cellular macromolecule biosynthetic process E GO: macromolecule localization E GO: cellular component organization or biogenesis E GO: chromosome segregation E GO: chromatin modification E GO: intracellular transport E GO: chromatin organization E GO: cellular component biogenesis at cellular level E GO: translation E GO: protein localization E GO: ncrna metabolic process E GO: macromolecule modification E GO: microtubule-based process E GO: macromolecule biosynthetic process E Cluster 3 GO.ID Term Annotated Significant Expected classicfisher Over-representation GO: metabolic process E GO: cellular metabolic process E GO: cofactor metabolic process E GO: small molecule metabolic process E GO: primary metabolic process E GO: cofactor biosynthetic process E GO: coenzyme metabolic process E GO: cellular lipid metabolic process E GO: type I interferon production E GO: cofactor transport E GO: cellular ketone metabolic process E GO: B cell homeostasis E

23 GO: interferon-beta production E GO: B cell differentiation GO: oxidation-reduction process GO: biosynthetic process GO: lipid metabolic process GO: regulation of interferon-beta production GO: alcohol metabolic process GO: glycerolipid biosynthetic process GO: sterol metabolic process GO: coenzyme biosynthetic process GO: carboxylic acid metabolic process GO: oxoacid metabolic process GO: lipid biosynthetic process GO: iron ion homeostasis GO: lymphocyte homeostasis GO: cellular nitrogen compound metabolic process GO: magnesium ion transport GO: cellular catabolic process GO: nitrogen compound metabolic process GO: nucleoside metabolic process GO: sterol biosynthetic process GO: tetrapyrrole biosynthetic process GO: intracellular signal transduction GO: positive regulation of innate immune response GO: organic acid metabolic process GO: GPI anchor biosynthetic process GO: protein lipidation GO: B cell activation GO: carbohydrate derivative metabolic process GO: tetrapyrrole metabolic process GO: heterocycle metabolic process GO: type I interferon biosynthetic process GO: protein homotetramerization GO: glycerophospholipid biosynthetic process GO: regulation of type I interferon production GO: response to oxidative stress GO: nucleoside monophosphate biosynthetic process GO: porphyrin-containing compound metabolic process

24 Cluster 4 GO.ID Term Annotated Significant Expected classicfisher Over-representation GO: cell fate commitment <1e GO: embryonic organ morphogenesis <1e GO: axonogenesis <1e GO: regionalization <1e GO: pattern specification process <1e cell morphogenesis involved in neuron GO: differentiation <1e GO: cell morphogenesis involved in differentiation <1e GO: neuron projection morphogenesis <1e GO: brain development <1e GO: central nervous system development <1e GO: embryonic morphogenesis <1e GO: regulation of nervous system development <1e GO: regulation of cell development <1e GO: organ morphogenesis <1e GO: tissue morphogenesis <1e GO: neuron projection development <1e GO: neuron differentiation <1e GO: neuron development <1e GO: epithelium development <1e GO: generation of neurons <1e GO: neurogenesis <1e GO: cell-cell signaling <1e GO: nervous system development <1e GO: positive regulation of developmental process <1e GO: cell morphogenesis <1e GO: regulation of multicellular organismal development <1e GO: cell development <1e GO: cell migration <1e GO: tissue development <1e GO: embryo development <1e GO: regulation of cell differentiation <1e GO: locomotion <1e GO: anatomical structure morphogenesis <1e GO: regulation of developmental process <1e GO: regulation of multicellular organismal process <1e GO: organ development <1e GO: system development <1e GO: cell differentiation <1e GO: regulation of cell communication <1e GO: regulation of localization <1e

25 GO: cellular developmental process <1e GO: multicellular organismal development <1e GO: anatomical structure development <1e GO: regulation of signaling <1e GO: developmental process <1e GO: positive regulation of cellular process <1e GO: positive regulation of biological process <1e GO: negative regulation of cellular process <1e GO: negative regulation of biological process <1e GO: multicellular organismal process <1e Cluster 5 GO.ID Term Annotated Significant Expected classicfisher Over-representation GO: detection of chemical stimulus involved in sensory perception of smell < 1e GO: detection of chemical stimulus involved in sensory perception < 1e GO: sensory perception of smell < 1e GO: detection of chemical stimulus < 1e GO: sensory perception of chemical stimulus < 1e GO: detection of stimulus involved in sensory perception < 1e GO: detection of stimulus < 1e GO: sensory perception < 1e GO: defense response to bacterium < 1e GO: G-protein coupled receptor signaling pathway < 1e GO: neurological system process < 1e GO: system process < 1e GO: cell surface receptor signaling pathway < 1e GO: response to chemical stimulus < 1e GO: defense response < 1e GO: signal transduction < 1e GO: signaling < 1e GO: cell communication < 1e GO: cellular response to stimulus < 1e GO: response to stimulus < 1e GO: multicellular organismal process < 1e GO: biological regulation < 1e GO: regulation of biological process < 1e GO: regulation of cellular process E GO: response to other organism E GO: response to biotic stimulus E GO: response to bacterium E GO: immune response E

26 GO: negative regulation of peptidase activity E GO: multi-organism process E GO: inflammatory response E GO: negative regulation of endopeptidase activity E GO: sensory perception of taste E GO: detection of chemical stimulus involved in sensory perception of bitter taste E GO: response to wounding E GO: detection of chemical stimulus involved in sensory perception of taste E GO: humoral immune response E GO: sensory perception of bitter taste E GO: defense response to Gram-positive bacterium E GO: innate immune response E GO: acute-phase response E GO: interferon-gamma production E GO: negative regulation of hydrolase activity E GO: regulation of interferon-gamma production E GO: acute inflammatory response E GO: cytokine production E GO: defense response to Gram-negative bacterium E GO: regulation of peptidase activity E GO: immune effector process E GO: regulation of immune response E

27 Supplementary Note Mononucleosome-BisSeq (MN-BisSeq) library preparation The protocol was adapted from Illumina Genomic DNA Sample Preparation Guide. Briefly, 2 μg of mononucleosomal fraction DNA were end repaired by incubation at 20 C for 30 minutes with 200µM dntp, 7.5 units of T4 DNA polymerase (NEB #M0203S), 5 units of DNA Polymerase I Large Fragment (Klenow) (NEB #M0210S), 25 units of T4 PNK (NEB #M0201S), 1x T4 DNA ligase buffer containing 10mM ATP (NEB). 3 ends of DNA fragments were adenylated by incubation at 37 C for 30 minutes with 100µM datp, 1xNEB Buffer 2, 10 units Klenow Fragment (3 5 exo ) (NEB # M0212L). Single End adapter sequences were produced based on Illumina adapter sequences (Oligonucleotide sequences Illumina, Inc. All rights reserved). 5 P- GATXGGAAGAGXTXGTATGXXGTXTTXTGXTTG and 5 AXAXTXTTTXXXTAXAXGAXGXTXTTXXGATXT, where X is a methylated cytosine. Adapters were ordered as single stranded oligos (Microsynth AG), resuspended in annealing buffer (10mM Tris ph7.5, 50mM NaCl, 1mM EDTA), annealed by heating at 95 C for 10 minutes and cooling down slowly. Annealed adapters were ligated to the DNA fragments as per manufacturer s instructions for genomic DNA library construction. Adapter-ligated DNA of 250 bp was isolated on 2% agarose gel electrophoresis. Gel purified DNA was then converted with sodium bisulfite using the Imprint DNA Modification Kit (Sigma-Aldrich) following the manufacturer s instructions. Half of the bisulfite-converted, adapter-ligated DNA molecules wasenriched by 7 cycles of PCR with the following reaction composition: 2.5 U of uracilinsensitive PfuTurboCx Hotstart DNA polymerase (Stratagene), 5 μl 10X PfuTurbo reaction buffer, 25 μm dntps, 0.5µM of Illumina single end PCR primers. The thermocycling parameters were: 95 C 2 min, 98 C 30 sec, then 7 cycles of 98 C 15 sec, 65 C 30 sec and 72 C 3 min, ending with one 72 C 5 min elongation step. The reaction products were purified using the MinElute PCR purification kit (Qiagen, Valencia, CA), resolved by 2% agarose gel electrophoresis to separate the library from adapter-adapter ligation products, and purified from the gel using the MinElute gel purification kit (Qiagen, Valencia, CA). Quality of the libraries and template size distribution were assessed by running an aliquot of the library on an Agilent 2100 Bioanalyzer (Agilent Technologies). 27

28 Analysis of bisulfite converted sequencing (BisSeq) data Read filtering and alignment of the BisSeq data from this study (bisulfite converted mononucleosome associated sperm DNA), published sperm whole methylomes (mouse sperm 5 and human sperm 2 ) and published oocyte and blastocyst methylomes 5 were performed as described 7. Peak identification Peak identification for nucleosome data was performed by training a two state hidden semi- Markov model, R-mhsmm package 9 on nucleosome enrichments calculated for 1 kb windows. The two-state model (non-peak and peak states) was initialized using Gaussian emissions (means of 0, 1, and variances of 0.5, 0.5), a gamma sojourn distribution with shape=2 and scale=10, and initial state probabilities of 0.5, 0.5. Parameter estimation of the model was performed by using 1kb window enrichments on chr1 as a training data set and by selecting a maximum of 200 windows in one state. Model parameters were estimated using EM algorithm, and the fitted model with emission distribution means of -0.21, 0.57, variances of 0.071, 0.48, and sojourn distribution parameters shape of 1.32, 0.38, and scale of 52.84, 7.86 was used to predict the maximum likelihood state path for the sequence of all 1kb windows in the genome. Adjacent 1kb windows with identical state labels were fused. Classification of nucleosome peaks Nucleosome enrichments were quantified on the peaks identified by a hidden semi- Markov model. Peaks were classified into three equal sized groups according to enrichment levels, termed weak, intermediate and strong peak groups (Fig. 5a). Genome-wide modeling of nucleosome occupancy Modeling of genome-wide mouse sperm nucleosome occupancy (Fig. 2d) was performed by using enrichment values in 1kb windows. Nucleosome occupancy data was modeled by a linear model with CpG dinucleotide frequency and the average % DNA methylation in 1kb windows as regressors, including only windows with detectable levels of nucleosomes, defined as log2 nucleosome occupancy greater than 0.2. This filtering excludes the majority of 1kb windows 28

29 without nucleosomes from the analysis which is required since the majority of the genome does not contain any nucleosomes in mouse sperm. In the model, DNA methylation data which was obtained by bisulfite converted sequencing of nucleosome associated DNA was used. Average DNA methylation was calculated by taking the ratio of total number of reads for methylated C over total number of reads for all C (methylated or umethylated) per window. Windows with less than 5 total reads for all Cs were excluded from the analysis. Finally, n= (9% of all windows) were used in modeling of nucleosome occupancy in mouse sperm. The same windows were used to analyze the relationship between mouse sperm nucleosome occupancy and DNA methylation / sequence composition (Fig. 2a, 2b). To analyze the relationship between human sperm nucleosome occupancy 3,4 and DNA methylation / sequence composition (Supplementary Fig. 2c - 2f), 1kb windows were processed in a similar way as for mouse sperm. Finally, n= (29% of all windows) of 3 and n= (17% of all windows) of 4 were used in the analysis. Quantifying expression in round spermatids Round spermatid expression data was quantified by summing the total number of reads mapping to Refseq transcripts. Concerning the classification of the expression status, transcripts without any aligned reads were classified as not detected, and the remaining transcripts were classified on the basis of increasing expression values into three equally sized groups termed low, medium and high. Motif finding for histone peaks 7-mer motif frequencies in the foreground and background sets were calculated by using Bioconductor package Biostrings. Foreground refers to histone peaks identified via the hidden semi-markov model approach. As a background, we used CGI that do not intersect any of the peak regions in the foreground set. Motif enrichment (M) and abundance (A) values were calculated as follows: M=log2((fg Z /fg Total min(fg Total, bg Total ))+pscnt)-log2((bg Z /bg Total min(fg Total,bg Total ))+pscnt) A=(log2((fg Z /fg Total min(fg Total,bg Total ))+pscnt)+log2((bg Z /bg Total min(fg Total,bg Total) )+pscnt))0.5 29

30 where fg Z is the number of counts of motif Z in the foreground, fg Total is the total number of all motifs in foreground, bg Z is the number of counts of motif Z in background, bg Total is the total number of all motifs in the background, and pscnt is a constant number (8). Results were visualized in a motif enrichment abundance plot (MA plot) and the top 20 motifs enriched in the foreground set were displayed. Defining expression for oogenesis or early embryogenesis Expression data from 10 was processed as described in 3. Expression states are referred as not expressed, oocyte, 2-8 cell and blastocyst. Embryonic expression was classified according to the first expression stage during development. The distinction between maternal and embryonic expression in 2-cell embryos was made by comparing the expression levels in early embryos treated or untreated with α-amanitin. References 1. Kaplan, N. et al. The DNA-encoded nucleosome organization of a eukaryotic genome. Nature 458, (2009). 2. Molaro, A. et al. methylation profiles reveal features of epigenetic inheritance and evolution in primates. Cell 146, (2011). 3. Brykczynska, U. et al. Repressive and active histone methylation mark distinct promoters in human and mouse spermatozoa. Nat Struct Mol Biol 17, (2010). 4. Hammoud, S.S. et al. Distinctive chromatin in human sperm packages genes for embryo development. Nature 460, (2009). 5. Kobayashi, H. et al. Contribution of intragenic DNA methylation in mouse gametic DNA methylomes to establish oocyte-specific heritable marks. PLoS Genet 8, e (2012). 6. Goldberg, A.D. et al. Distinct factors control histone variant H3.3 localization at specific genomic regions. Cell 140, (2010). 7. Stadler, M.B. et al. DNA-binding factors shape the mouse methylome at distal regulatory regions. Nature 480, (2011). 8. Mikkelsen, T.S. et al. Genome-wide maps of chromatin state in pluripotent and lineagecommitted cells. Nature 448, (2007). 9. J., O.C. & S., H. Hidden Semi Markov Models for Multiple Observation Sequences: The mhsmm Package for R. Journal of Statistical Software 39, 1-22 (2011). 10. Zeng, F. & Schultz, R.M. RNA transcript profiling during zygotic gene activation in the preimplantation mouse embryo. Developmental biology 283, (2005). 30