Supplementary Materials and Methods
Generation of Suv4-20h ko mice
Targeting constructs for both Suv4-20h enzymes were generated by recombination in bacteria (Liu et al. 2003). The targeting vectors were linearized and electroporated into 129/Sv embryonic stem (ES) cells. About 400 G418-resistant ES cell clones were screened for each targeting construct. Southern blot analysis was used to identify homologous recombinants. For Suv4-20h1, the PCR-derived probe hybridizes to a 16.5 kb XhoI/BamHI restriction fragment in wt cells and a 8.5 kb fragment in homologous recombinants. A PCR-derived probe for Suv4-20h2 detects a 7.6 kb EcoRI/XbaI fragment in wt cells and a 6.2 kb fragment in recombinants. For each gene targeting, three independent ES cell clones with homologous integration at the targeting site were injected into C57BL/6J blastocysts to generate chimeric mice. These chimeras were subsequently crossed with C57BL/6J females. Heterozygous mice with successful germ line transmission of the targeted allele were crossed with C57BL/6J mice expressing Flp recombinase to remove the neomycin resistance cassette resulting in “flox”-alleles. To disrupt Suv4-20h enzymes, heterozygous Suv4-20h1f/+ and Suv4-20h2f/+ mice were crossed with Mox2-Cre mice (Tallquist and Soriano 2000), resulting in heterozygous deletion of the SET domain coding regions, and intercrossed to generate Suv4-20h1-/-, Suv4-20h2-/- and Suv4-20h dn mice. To delete Suv4-20h enzymes in the hematopoietic system, Suv4-20h dnvav mice were engineered by crossing vav-Cre mice (de Boer et al. 2003) with Suv4-20h1f/f, Suv4-20h2-/- mice.
Northern blots
To analyze expression of Suv4-20h enzymes in embryonic development and in different tissues of adult mice, a 1.8kb fragment of the Suv4-20h1 cDNA and a 1.2kb fragment of the Suv4-20h2 cDNA was hybridized to multiple tissue northern blots (BLOT2 and BLOT3, Sigma) following the manufacturers recommendation.
Mass-spectrometry
Bulk histones from pMEFs and B cells were isolated by acid extraction. Mass-spec analysis was performed after propionylation and tryptic digest as described (Peters et al. 2003). Propionylation prevents tryptic cleavage at lysine residues, however, we never observed an effect on cleavage efficiency at arginine residues adjacent to propionylated lysines (e.g. H3R8-K9, H3R26-K27). For the quantification of H4K20 methylation, we first synthesized four different peptides H4 (12-31)-cys [KGGAKRHRKVLRDNIQGITK-cys], which were unmodified, mono-, di-, or tri-methylated at the H4K20 position. Equal amounts of these synthetic peptides were then subjected to the same propionylation-tryptic digest-propionylation-HPLC-MS procedure as histone samples isolated from cells. This confirmed that only fragments H4(20-23) but not longer fragments were generated, irrespective of the initial lysine methylation state. Furthermore, synthetic peptides were used to correct for possible effects of different lysine modifications on the final mass spectrometric signal, thereby taking into account different behavior of differentially methylated peptides during sample preparation, pre-column enrichment, HPLC separation, ionization and mass-spectrometric detection.
For 1 pmole of synthetic peptide, we obtained the following mass-spectrometric read-outs (from full-scan spectra, summing up expected 1+ and 2+ ion m/z values):
Starting peptide / Analyte / Mass spec response[counts/pmol starting material]
KGGAKRHRKVLRDNIQGITKC / Pr-KprVLR / 8281842
KGGAKRHRKme1VLRDNIQGITKC / Pr-KprmeVLR / 6426685
KGGAKRHRKme2VLRDNIQGITKC / Pr-Kme2VLR / 6453876
KGGAKRHRKme3VLRDNIQGITKC / Pr-Kme3VLR / 1773877
These values were then used to correct the peak areas of the differentially modified Pr-K(mod)VLR peptides obtained from histones from cellular extracts.
Cell proliferation assays
Primary mouse embryonic fibroblasts (pMEFs) were isolated from E14.5 embryos by standard procedures. For cumulative proliferation assays, Suv4-20h dn and genetically matched wt pMEFs were plated at a density of 3x105 cells in 6cm dishes and counted every three days. For cell cycle analysis, pMEFs were plated at a density of 3x105 cells in 6cm dishes. At day 2, 10mM BrdU was added for 30 min to label cells in S-phase. Cells were fixed with 80% methanol and stained with a-BrdU-FITC (BD Pharmingen) and propidium iodide. Cell cycle profiles were then determined by FACS analysis for BrdU/PI. To examine kinetics of S-phase entry, Suv4-20h dn and wt pMEFs were synchronized by serum starvation. Cells were grown until they reached confluence, followed by two days of incubation with medium containing low (0.5%) serum. Synchronized cells were trypsinized and cultured in medium containing 15% serum to allow cell cycle re-entry. S-Phase entry was determined by BrdU pulse-labeling and FACS analysis. The experiment was repeated three times.
To measure non-cycling cells, pMEFs that were cultured under normal and low oxygen conditions were incubated with BrdU for 24h. BrdU-negative cells (=non-cycling) were quantified by FACS analysis.
DNA damage assays
Primary MEFs (2 independent wt and dn cell lines) were seeded on chamber slides (LabTek), treated with 2Gy ionizing radiation and fixed at different time points post irradiation. Immunofluorescence was conducted using gH2A.X and 53BP1 antibodies. Per time point >200 cells were scored for 53BP1 foci.
To analyze chromosomal abnormalities post DNA damage, pMEFs were treated with 2Gy ionizing radiation and incubated for 1h. Then cells were incubated with 10mM nocodazole for 4-5h prior to mitotic spread preparation (Morales et al. 2003). In a second series of experiments the G2-M checkpoint was blocked with 5mM caffeine one hour post irradiation. Over 100 metaphase spreads were scored for chromatid gaps and breaks.
For colony formation assays, MEFs were plated in triplicate at 500 cells/10 cm dish. Post treatment, the cells were incubated to form colonies for 7 days and were subsequently fixed in methanol/acetic acid and stained with crystal violet. Colonies per dish were counted and the results are represented as percentage of surviving colonies relative to the untreated control.
G2-M checkpoint
Damage-induced G2-M arrest was tested by treating wt and Suv4-20h dn pMEFs with different doses of ionizing radiation. One hour post irradiation cells were fixed in 70% methanol and stained with PI and H3S10p (Hendzel et al. 1997; Difilippantonio et al. 2007; Kim et al. 2007).
Nucleosome binding assay
Generation of nucleosomal substrates
Octamers and oligonucleosomes with recombinant Xenopus histones were prepared as described (Luger et al. 1999). Biotinylation was performed in addition to the described standard protocol. Briefly, 1mg of Histone H2B was incubated with 10mM EZ-link sulfo NHS-Biotin (Pierce Biotechnology) for 1hr and mixed with equimolar amounts of histones H3, H2A and H4. Octamers were reconstituted on a FPLC sizing column. Plasmid containing 5S ribosomal DNA repeats was used for assembly of oligonucleosomes.
HMTase Assays
GST fusion proteins of Suv4-20h2 (aa 1-280) and G9a (aa 948-1244) containing the SET domain were expressed in Escherichia coli. HMTase assays were performed as described (Nishioka et al. 2002). In particular, 150mg of biotin nucleosomes were incubated with 75mg Suv4-20h2 or 1mg G9a for 1 h at 30°C in reaction buffer (50 mM Tris-HCl at pH 8.5, 5 mM MgCl2, 4 mM DTT) and 20mM S-Adenosyl Methionine (Amersham Pharmacia Biotech). The methylation was confirmed by immunoblotting for H4K20me2, H4K20me3 and H3K9me2 (Perez-Burgos et al. 2004). 100mg of methylated biotin nucleosomes were hybridized on the CADOR (chromatin associated-domain array) chip and detected as described (Kim et al. 2006).
V-DJ recombination
Pro B cells (c-kit+, CD19+, IgM-) were isolated from bone marrow of 4 weeks old wild type (n=4) and Suv4-20h dnvav (n=4) mice using FACS Aria cell sorter. For analysis for TCRb rearrangements double positive (CD4+ CD8+ CD90.2+) and double negative (CD4- CD8- CD90.2+) thymocytes were sorted from the thymus of 4 to 6 week old wt or Suv4-20h dnvav mice. Sorted cells were digested with proteinase K and DNA was extracted with phenol/chloroform and then ethanol precipitated.
V(D)J rearrangements were amplified using published primers (Fuxa et al. 2004; Cobaleda et al. 2007) and visualized using ethidium bromide. Rearrangement dilutions were normalized against thy1 (CD90) using primers fw CCATCCAGCATGAGTTCAGC and rev CTTGACCAGCTTGTCTGTGG.
Class-switch recombination
Spleens of Suv4-20h1f/f, Suv4-20h2-/-, vavCre (= Suv4-20h dnvav) and age-matched wt mice were isolated and B220+ cells were purified by magnetic cell separation (MACS) using magnetic B220 beads (Miltenyi Biotec). To analyze cell proliferation, cells were labeled with 1mM CFSE for 10 min at 37°C. B cells were cultured at a density of 5x105 cells/ml in RPMI-medium containing glutamine, non-essential amino acids, b-mercaptoethanol, 15% fetal calf serum supplemented with 25mg/ml LPS and 10ng/ml IL-4 to stimulate class-switch recombination to IgG1 or 25mg/ml LPS for CSR to IgG3. After 4 days cells were harvested and stained with a-IgG1 or a-IgG3 antibodies (BD Pharmingen). Cell proliferation and IgG1/IgG3-positive cells were determined by FACS analysis. IgH-associated breaks were analyzed in activated B cells as described (Callen et al. 2007).
RT-PCR analysis and Affymetrix expression profiling
Pro B (c-kit+, CD19+, IgM-) and pre B (CD25+, CD19+, IgM-) cells were isolated from bone marrow of 4 weeks old wild type and Suv4-20h dnvav mice by FACS Aria cell sorter. Mature splenic B220+ B cells were isolated by MACS sorting. Activated B cells were isolated after LPS/IL-4 stimulation (day 4); apoptotic cells were removed by standard Ficoll centrifugation.
Total RNA was extracted with TRIzol (Invitrogen), reverse transcribed with random hexamers and SuperScript II reverse transcriptase (Invitrogen). First-strand cDNA was used for PCR and SYBR-green fluorogenic dye real-time PCR (Roche). Primer pairs for RT-PCR analysis are listed in Table S4.
For Affymetrix expression analysis, total RNA from wild type (n=2) and Suv4-20h dn (n=2) ES cells, wild type (n=2) and Suv4-20h dnvav (n=2) B220+ resting B cells and wild type (n=3) and Suv4-20h dnvav (n=3) activated B cells (LPS/IL-4 day 4) was prepared. RNA was processed according to standard procedures and hybridized to Affymetrix Mouse Genome 2.0 arrays (RZPD, Berlin). Raw data was normalized with quantiles and background corrected with mas algorithms. P-values were calculated from a t-test analysis.
FACS analysis of B and T cell lineages
The following FITC-, phycoerythrin (PE)- or allophycocyanin (APC)-coupled antibodies were used for flow cytometry: anti-B220 (RA3-6B2), CD4 (L3T4), CD8 (53-6.7), CD19 (1D3), CD25 (PC61.5), c-Kit (2B8), sca-1 (D7), IgM (M41.42), IgD (1.19), Mac-1 (M1/70), Gr-1 (RB6-8C5), M-CSF-R (AFS98), CD45.1 (A20), CD45.2 (107), IgG1 (A85-1), IgG3 (R40-82). Unspecific antibody binding was suppressed by preincubation of cells with CD16/CD32 Fc-block solution (BD Pharmingen).
Competitive bone marrow reconstitution
Hematopoietic stem cells were analyzed in bone marrow from 2 months old wild type (n=3) and Suv4-20h dnvav (n=3) mice. First, lineage positive cells were depleted by AutoMACS (Miltenij Biotec). Then, HSCs were determined by FACS analysis for lin-, Sca-1+, c-kit+.
Lethally irradiated B6.SJL recipient mice (CD45.1) were injected with 5x105 bone marrow cells (1:1) derived from wt (CD45.1) and Suv4-20h dnvav (CD45.2) mice. Hematopoietic lineages in recipient mice were analyzed for CD45.1 and CD45.2 7 weeks and 11 months after transplantation. The following lineage markers were used: granulocytes (Mac-1+, Gr-1+), macrophages (M-CSF-R+, Mac-1+), bone marrow B cells (CD19+), splenic B cells (B220+), T cells (CD4+; CD8+; CD4+,CD8+).
Supplementary Figures
Supplementary Figure 1
Expression profiles of Suv4-20h1 and Suv4-20h2 were analyzed in adult tissues by northern blotting. Suv4-20h1 is abundantly expressed in all adult tissues. By contrast, Suv4-20h2 shows a restricted expression profile in adult tissues. Br - brain, H – heart, Li – liver, Ki – kidney, Sp – spleen, Te – testis, Lu – lung, Th – thymus, Pl – placenta, Mu – muscle.
Supplementary Figure 2
Analysis of H4K20me3 in embryonic and adult tissues.
(A) Paraffin sections from E14.5 wt and Suv4-20h2-/- embryos were stained for H4K20me3 (brown=H4K20me3, blue=cytoplasmic counter stain). In wt tissues, H4K20me3 is very abundant. Notably, in Suv4-20h2-/- embryos H4K20me3 is strongly reduced but still detectable in some cells (enlargement). (B) H4K20me3 was analyzed in different tissues of adult wt and Suv4-20h2-/- mice. H4K20me3 is readily detectable in tissues of adult Suv4-20h2-/- mice, suggesting compensation by Suv4-20h1 or by a different HMTase.
Supplementary Figure 3
Mass-spec analysis of different histone H3 and H4 methylation and acetylation states in wt and Suv4-20h dn pMEFs. For histone H4, our method does not allow to distinguish positions of acetylation (K5, K8, K12 or K16) but only the number of acetyl groups on the H4 (aa4-17) fragment. No significant differences between wt and Suv4-20h dn were observed in methylation and acetylation states.
Supplementary Figure 4
Suv4-20h dn pMEFs display proliferation defects but normal G2-M checkpoint.
(A) Logarithmically growing pMEFs under normal and low oxygen conditions were labeled with BrdU for 24h and, numbers of BrdU negative (= non-cycling) cells were determined by FACS analysis. Higher percentages of non-cycling cells were observed in Suv4-20h dn populations. (B) Wild type and Suv4-20h dn pMEFs were treated with 0.5 and 2Gy IR and incubated for one hour. Following this, cells were fixed and stained with H3S10p antibody and propidium iodide (PI). Numbers of mitotic cells were quantified by FACS analysis. IR treated pMEFs show reduced numbers of mitotic cells due to an active G2-M checkpoint. Both wt and Suv4-20h dn cells show a similar G2-M arrest.
Supplementary Figure 5
The DNA damage response protein 53BP1 associates with Suv4-20h-methylated nucleosomes.
(A) Schematic showing the generation of biotinylated recombinant nucleosomes. To generate methylated nucleosomes, H2B was biotin-tagged and assembled in nucleosomes which were then modified in vitro using the Suv4-20h2 HMTase. (B) Nucleosomes were methylated with G9a and Suv4-20h2 and tested by western blotting for H3K9 and H4K20 methylation states. G9a induces H3K9me2, whereas Suv4-20h2 mediates H4K20me2 and me3. (C) Modified nucleosomes were incubated with the CADOR protein array (Kim et al. 2006), which contains around 150 royal family domains, such as chromo, tudor, MBT, PWWP, bromo and SANT (Maurer-Stroh et al. 2003). Nucleosomal binding was detected using streptavidin-Cy5. Although some unspecific signals were detected using unmodified nucleosomes, G9a-modified nucleosomes selectively associated with chromodomains of the three HP1 isoforms. The only specific binder for Suv4-20h2-methylated nucleosomes was the tudor domains of the DNA damage protein 53BP1.
Supplementary Figure 6
Deletion of Suv4-20h enzymes results in genome-wide conversion to H4K20me1 in B cells.
(A) Expression of Suv4-20h1 and Suv4-20h2 was analyzed in wt and Suv4-20h dnvav splenic B cells (CD19+). No transcripts were detectable for Suv4-20h enzymes in Suv4-20h dnvav B cells, whereas controls (Hprt and Bach2) were unchanged. (B) Mass-spec analysis of H4K20 methylation states in wt and Suv4-20h dnvav splenic B cells (CD19+). In wt cells, H4K20me2 is the most abundant H4K20 methylation state. Upon abrogation of the Suv4-20h enzymes, similar to pMEFs, B cells show a conversion from H4K20me2 and H4K20me3 to H4K20me1.