Mapping Post-translational Modifications of the Histone Variant MacroH2A1 Using Tandem Mass Spectrometry*S

Post-translational histone modifications modulate chromatin-templated processes and therefore affect cellular proliferation, growth, and development. Although post-translational modifications on the core histones have been under intense investigation for several years, the modifications on variant histones are poorly understood. We used tandem mass spectrometry to identify covalent modifications on a histone H2A variant, macroH2A1.2. MacroH2A1.2 can be monoubiquitinated; however, the site of monoubiquitination has not been documented. In this study we used green fluorescent protein-tagged macroH2A1.2 to determine that Lys115 is a site of ubiquitination. In addition, we found that this variant H2A is methylated on the ε amino group of lysine residues Lys17, Lys122, and Lys238 and phosphorylated on Thr128. Three of these modifications were also found to be present in the endogenous protein by mass spectrometric analysis. These results provide the first direct evidence that multiple post-translational modifications are imposed on macroH2A1.2, suggesting that, like canonical H2A, this variant H2A is subject to regulation by combinatorial use of covalent modifications.

Post-translational histone modifications modulate chromatin-templated processes and therefore affect cellular proliferation, growth, and development. Although posttranslational modifications on the core histones have been under intense investigation for several years, the modifications on variant histones are poorly understood. We used tandem mass spectrometry to identify covalent modifications on a histone H2A variant, macroH2A1.2. MacroH2A1.2 can be monoubiquitinated; however, the site of monoubiquitination has not been documented. In this study we used green fluorescent protein-tagged macroH2A1.2 to determine that Lys 115 is a site of ubiquitination. In addition, we found that this variant H2A is methylated on the ⑀ amino group of lysine residues Lys 17 , Lys 122 , and Lys 238 and phosphorylated on Thr 128 . Three of these modifications were also found to be present in the endogenous protein by mass spectrometric analysis. These results provide the first direct evidence that multiple post-translational modifications are imposed on macroH2A1.2, suggesting that, like canonical H2A, this variant H2A is subject to regulation by combinatorial use of covalent modifications.

Molecular & Cellular Proteomics 5:194 -203, 2006.
All DNA-based processes in mammalian cells must contend with chromatin. DNA is wrapped around an octamer of four core histones, H2A, H2B, H3, and H4, to form nucleosomes, which in turn are subject to higher order packaging to form chromatin. Histones are subject to covalent modifications such as acetylation, methylation, phosphorylation, and ubiquitination. Different combinations of histone modifications characterize transcriptionally active euchromatin and transcriptionally silent heterochromatin (1,2). These epigenetic marks provide binding sites for proteins that regulate gene expression and chromosome structure. Replacement of core histones with variant histones provides further structural diversity that may be used to modulate chromatin structure. The combination of regulated alterations in post-translational modifications of core histones and incorporation of histone variants affects chromatin architecture to modulate transcription, facilitate DNA repair and chromosome segregation, and define epigenetic states. Recently tandem mass spectrometry has proven advantageous in dealing effectively with the detection and assignment of this "combinatorial" post-translational complexity for individual sites as well as multisite occupancies at the protein level (3)(4)(5).
H2A is the core histone that has the most variants, and these variants are implicated in a number of processes. H2A.X is involved in DNA damage repair (6). H2A.Z is required for proper chromosome segregation and HP1␣ localization (7). H2A.BBD is associated with active chromatin (8), and macroH2A1.1, macroH2A1.2, and macroH2A2 are enriched on the facultative heterochromatin of the inactive X chromosome (9 -11). MacroH2A1.1 and macroH2A1.2, which are produced by alternative splicing of the macroH2A1 gene (H2AFY, according to Human Genome Organization nomenclature), were first identified as histone variants that replace canonical H2A in an estimated 3% of nucleosomes (12). These are vertebrate-specific variants that contain a histone domain that is highly similar to H2A and a large, globular non-histone domain. The two isotypes of macroH2A1 are identical except for a 30-amino acid region in the non-histone domain. This non-histone domain is related to a family of proteins that includes polynucleotide hydrolases and nicotinamide dinucleotide metabolite-binding proteins (13,14). MacroH2A1.1 and macroH2A1.2 differ in their expression patterns during development (15). MacroH2A1.1, but not macroH2A1.2, binds Oacetyl-ADP-ribose, a product of the nicotinamide dinucleotidedependent SirT1 deacetylation reaction (14). In addition to enrichment on the inactive X chromosome, macroH2A1.2 is also enriched in regions of heterochromatin in senescent cells, suggesting a role in repression of transcription (9,16). Depletion of macroH2A1.1 and macroH2A1.2 in female cells causes reactivation of a transgene on the inactive X chromosome, demonstrating a role for these variants in the maintenance of silenced chromatin (17). When macroH2A1.2 is incorporated into nucleosomes, transcription factor binding is inhibited, and the activity of chromatin remodeling factors is blocked, suggesting mechanisms by which this variant histone may facilitate transcriptional repression (18).
Canonical H2A is subject to several post-translational modifications (19 -23). This suggests that the macroH2A family, which shows significant conservation with H2A, would also exhibit a similar range of covalent modifications. MacroH2A1.2 is ubiquitinated (17). In this work, we used capillary HPLC electrospray tandem mass spectrometry to identify a site of ubiquitination at Lys115 on immunopurified epitope-tagged macroH2A1.2. In addition, we detected methylation of Lys 17 , Lys 122 , and Lys 238 and phosphorylation of Thr 128 . Finally we showed the presence of three of these modifications on endogenous macroH2A1.2 by mass spectrometry. Thus, like the core histone H2A it replaces, macroH2A1.2 is subject to a variety of post-translational modifications. We speculate that the covalent modifications of macroH2A1.2 are a further elaboration of the histone code, providing additional mechanisms to regulate chromatin structure.

EXPERIMENTAL PROCEDURES
Cell Lines-Full-length human macroH2A1.2 cDNA was generated using the polymerase chain reaction and cloned into a backbone originating from pBOS-H2B-GFP (Clontech), generating a construct that expresses macroH2A1.2 tagged with green fluorescent protein (GFP) 1 on the C terminus. This plasmid was transfected into human embryonic kidney cells (293 cells) using FuGENE 6 (Roche Applied Science), and stable lines were selected using blasticidin S (Invitrogen). All 293 and 293-derived cell lines were maintained in Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum and antibiotics.

SDS-PAGE Separation and
In-gel Tryptic Digestion-Affinity-purified macroH2A1-containing samples were eluted from beads with sample buffer (0.24 M Tris, 8% SDS, 2.88 M ␤-mercaptoethanol, 40% glycerol, and 0.4% bromphenol blue) and loaded onto 4 -20% SDS-PAGE Criterion Ready Gels (Bio-Rad) for separation. Proteins were visualized with Coomassie Brilliant Blue G-250. In-gel digestions on histone bands were performed utilizing a procedure described at ms-facility.ucsf.edu/ingel.html. Typically 50 ng of trypsin (porcine, side chain-protected; Promega, Madison, WI) was used for each gel band, and digestions were carried out at 37°C for 2 h. Peptides were extracted from gel pieces with 50 l of 50% acetonitrile, 2% formic acid thrice, and the extraction solution was dried down to ϳ10 l.
On-line Capillary LC-MS and LC-MS-MS Analysis-A 1-l aliquot of the digestion mixture was injected into an Ultimate capillary LC system via a FAMOS autosampler (LC Packings, Sunnyvale, CA) and separated by a 75-m ϫ 15-cm reverse-phase capillary column at a flow rate of ϳ330 nl/min. The HPLC eluent was connected directly to the micro-ion electrospray source of a QSTAR XL mass spectrometer (Applied Biosystems/MDS Sciex, Foster City, CA). Typical performance characteristics were Ͼ8000 resolution with 30-ppm mass measurement accuracy in both MS and CID spectra. LC-MS data were acquired in an information-dependent acquisition mode, cycling between 1-s MS acquisition followed by 3-s low energy CID data acquisition. The centroided peak lists of the CID spectra were searched against the National Center for Biotechnology Information (NCBI) protein database using Batch-Tag, a program in the in-house version of the University of California San Francisco ProteinProspector package (24), considering protein N terminus and lysine acetylation; lysine mono-, di-, and trimethylation; arginine methylation; phosphorylation; and lysine ubiquitination as variable modifications. The CID spectra with putative post-translational modifications were further inspected manually.
Immunofluorescence and Fluorescence in Situ Hybridization-Cells were cultured on glass coverslips, fixed in 4% paraformaldehyde for 10 min, extracted in PBS containing 0.5% Triton X-100 for 10 min, washed in PBS containing 0.2% Tween 20 (PBST), and then placed in a block solution (PBS, 0.2% cold water fish skin gelatin, 5% normal goat serum, and 0.2% Tween 20) for 20 min. The cells were then incubated with anti-macroH2A1 antibody diluted into block solution (1:500) for 1 h, washed three times in PBST, and then incubated with Texas Red-conjugated anti-rabbit antibody (Vector Laboratories) for 30 min. Coverslips were then washed in PBST, stained with 4,6diamidino-2-phenylindole dihydrochloride in PBST, washed in PBST, and mounted in Vectashield (Vector Laboratories). Fluorescence in situ hybridization for XIST RNA using directly labeled DNA probes was carried out as described previously (25).

Expression and Purification of GFP-tagged MacroH2A1.2-
To generate a source of macroH2A1.2 for affinity purification and mass spectrometry, we generated 293:macroH2A1.2-GFP, a transformed human embryonic kidney cell line stably expressing macroH2A1.2 tagged with GFP on the C terminus.
To determine whether the addition of GFP to macroH2A1.2 adversely affected expression or localization, GFP fluorescence was used to analyze the intracellular localization of the fusion protein. For comparison, the distribution of endogenous macroH2A1 in the parental 293 cell line was determined by immunofluorescence using a rabbit polyclonal antibody that recognizes macroH2A1.1 and macroH2A1.2. Like endogenous macroH2A1, macroH2A1.2-GFP was distributed uniformly throughout most of the nucleus and was enriched in one to two large contiguous domains, which coincided with the two XIST RNA-coated inactive X chromosomes in this aneuploid cell line (Fig. 1A). Therefore addition of GFP to macroH2A1.2 did not noticeably alter the localization properties of this variant histone.
To determine whether macroH2A1.2-GFP was correctly incorporated into chromatin, macroH2A1.2-GFP was immuno-purified using an anti-GFP antibody (Fig. 1B). SDS-PAGE separation of the proteins that bound to anti-GFP antibody yielded two bands of ϳ70 kDa, the expected size for the macroH2A1.2-GFP fusion protein, that were both identified as macroH2A1.2 by Western blotting and LC-MS-MS analysis (data not shown). In addition to macroH2A1.2-GFP, there was the quartet of low molecular weight bands that is characteristic of the core histones and that were identified as H3, H2A, H2B, and H4 via tandem mass spectrometric analysis (data not shown). Co-purification of the core histone proteins with macroH2A1.2-GFP indicates that tagged macroH2A1.2-GFP was appropriately incorporated into chromatin. Because addition of a C-terminal GFP tag to macroH2A1.2 did not alter the characteristics of this variant histone as assayed by its incorporation into chromatin and localization, we used affinity-purified macroH2A1.2-GFP to identify post-translational modifications by tandem mass spectrometry.
Identification of Post-translational Modifications on GFPtagged MacroH2A1.2-LC-MS-MS analysis of the tryptic digestion mixture eluted from the upper macroH2A1.2-GFP band identified macroH2A1.2 peptides and ubiquitin, suggesting ubiquitination of macroH2A1.2. Modified peptides can be identified by their characteristic mass changes due to modification, and sites of modification can be determined through fragmentation analysis. ProteinProspector reported one spectrum was potentially of a ubiquitinated peptide. The CID spectrum of a peptide with m/z 743.42 3ϩ (Table I and Fig. 2) displayed partial C-and N-terminal sequence ion ladders (y 2 to y 17 and b 2 to b 4 ) that matched a tryptic peptide of macroH2A1.2 spanning residues Gly 96 to Lys 116 . However, the molecular weight of this peptide was 114 amu higher than the unmodified peptide Gly 96 -Lys 116 , which equaled the mass of two glycine residues that remain after tryptic digestion of the ubiquitin appended to the ⑀ amino group of a lysine residue (26). The presence of the y 1 ion at m/z 147.11 showed the C-terminal lysine (Lys 116 ) was unmodified. Meanwhile the y 2 ion and all subsequent y ions had a mass 114 amu higher than predicted for the unmodified peptide. The 242-amu mass difference between y 2 and y 1 matches the residue mass of a ⑀-(Gly-Gly)-amino lysine, thus identifying Lys 115 as a site of ubiquitination. The difference in size and amount between the more abundant, faster migrating form of macroH2A1.2-GFP and the slower migrating, ubiquitin-containing form of macroH2A1.2-GFP is consistent with monoubiquitination of a subset of the total macroH2A1.2-GFP.
LC-MS-MS analysis of the tryptic digest of the major macroH2A1.2-GFP band of slightly lower molecular weight showed 89 CID spectra matching the macroH2A1.2 sequence with no ubiquitin peptides. ProteinProspector reported several potentially modified peptides, and manual inspection of the CID spectra of these peptides revealed monomethylation of Lys 17 , dimethylation of Lys 122 , phosphorylation of Thr 128 , and mono-and dimethylation of Lys 238 on macroH2A1.2-GFP (Table I).
In the CID spectrum of a peptide with a m/z value of 405.91 3ϩ (Fig. 3A), the mass values of the C-terminal sequence ions up to y 9 (y 1 to y 9   displayed a mass change of ϩ14 amu, suggesting a monomethylation within the N-terminal 15 SAKG 18 moiety. Combining information derived from y and b ions, Lys 17 can be assigned as the monomethylation site. The identity and structure of the peptide with an m/z value of 497.63 3ϩ was determined in a similar fashion (Fig. 3B). Essentially the complimentary and continuous C-and N-terminal fragment ion ladders (y 1 to y 8 and b 2 to b 8 , respectively) provided sufficient information for the conclusive assignment of the peptide sequence. However, the molecular weight of the precursor ion was 28 amu higher than that of the unmodified peptide Gly 121 -Lys 134 . The masses of the C-terminal sequence ions (y 1 to y 8 ) were the anticipated values, showing that all residues C-terminal of Ile 126 were unmodified. On the other hand, the addition of 28 amu on b ions (b 2 to b 8 ), relative to the unmodified peptide, revealed that the first two N-terminal residues of this peptide, Gly 121 and Lys 122 , were either dimethylated or doubly monomethylated. The absence of a side chain in glycine residues precludes the possibility for modifications on Gly 121 , indicating that Lys 122 is dimethylated.
The precursor ion at m/z 514.97 3ϩ had a mass increment of ϩ80 amu compared with that of the unmodified peptide Gly 121 -Lys 134 , indicating phosphorylation or sulfation of this peptide (Fig. 3C). The presence of a 98-amu neutral loss from the precursor ion at m/z 482.32 3ϩ suggested phosphorylation rather than sulfation (27). The mass values of the first six C-terminal sequence ions (y 1 to y 6 ) and the first six N-terminal sequence ions (b 2 to b 7 ) were identical to those of the un-modified peptide; thus residues from Pro 129 to Lys 134 and from Gly 121 to Ile 127 were not modified. Therefore, Thr 128 remained the only site for phosphorylation. As expected, Cand N-terminal sequence ions containing Thr 128 all had the characteristic H 3 PO 4 neutral loss (y 7 Ϫ98 to y 12 Ϫ98 and b 8 Ϫ98). The absence of any Thr 128 -containing fragment ions that matched the masses of the unmodified peptide further confirmed this assignment. This peptide also existed in the tryptic digest of monoubiquitinated macroH2A1.2-GFP, suggesting these two post-translational modifications are not mutually exclusive.
The CID spectra of the peptides with m/z of 530.30 3ϩ and 534.97 3ϩ (Fig. 4) correspond in mass to singly and doubly methylated versions of the peptide Lys 235 -Arg 248 . Both spectra exhibited the first seven C-terminal sequence ions (y 1 to y 7 ) with mass values identical to those of the unmodified peptide, suggesting that the Glu 242 -Arg 248 peptide moiety is not modified. Similarly, although weak, the mass of b 3 ion at m/z 243.2 demonstrated that the first three N-terminal residues, Lys 235 -Gly 237 , remained unmodified. The mass differences between b 4 and b 3 ions, 142 amu for the peptide with an m/z of 530.30 3ϩ (Fig. 4A) and 156 amu for the peptide with an m/z of 534.97 3ϩ (Fig. 4B), correspond to the residue mass of monoand dimethylated lysines, respectively. Furthermore, when compared with those of the unmodified peptide, b ions containing Lys 238 (b 4 to b 9 ) all displayed a mass increment of 14 amu for the peptide with an m/z of 530.30 3ϩ or 28 amu for the peptide with an m/z of 534.97 3ϩ (Fig. 4, A and B). Therefore, we conclude that Lys 238 occurs in mono-and dimethylated forms.
Post-translational Modifications on Endogenous Macro-H2A1-The phosphorylation and methylation of macro-H2A1.2-GFP may reflect biologically interesting modifications that occur on the endogenous protein or may be modifications that are consequences of overexpression or tagging of macroH2A1.2. To determine whether endogenous material is subject to these post-translational modifications, we used tandem mass spectrometry to analyze the tryptic peptides produced from macroH2A1 immunopurified from 293 cells with a rabbit polyclonal anti-macroH2A1 antibody, which recognizes macroH2A1.1 and macroH2A1.2 (data not shown). Although the macroH2A1 antiserum was not as efficient for immunoprecipitation as the GFP antiserum and less total protein was purified, LC-MS-MS analysis revealed macroH2A1 as the principal component in the sample.
Evidence for three of the modifications detected in the tagged protein was found in the analysis of the endogenous protein. The peptide containing monomethylated Lys 17 was detected and identified by MS-MS fragmentation. In addition, peptides with the masses of modified peptides containing Lys 122 dimethylation and Thr 128 phosphorylation were also detected, although they were not intense enough for fragmentation analysis (Table I). For example, Fig. 5A shows an extracted ion chromatogram (XIC) for ions of m/z 405.9. An XIC plot shows when during the LC run ions of a particular m/z were observed. The m/z 405.9 corresponds to the m/z of the triply charged peptide containing monomethylated Lys 17 that was observed in the macroH2A1.2-GFP digest, and a peak is clearly identified at Ϸ23.8 min (Fig. 5A). The mass spectrum of the peptides eluting at this time point shows that m/z 405.9 3ϩ co-eluted with peptides giving signals at m/z 445.6 3ϩ , 519.0 4ϩ , and 601.3 2ϩ (Fig. 5A, inset). Similarly the XIC for m/z 405.9 from the endogenous macroH2A1 digest also exhibited a peak at Ϸ23.8 min (Fig. 5B) that co-eluted with peptides giving signals at m/z 445.6 3ϩ , 519.0 4ϩ , and 601.3 2ϩ (Fig. 5B,  inset). This indistinguishable retention time observed in the chromatographic elution profiles of the digestion mixtures of macroH2A1.2-GFP and macroH2A1 suggests the identical chemical composition for the species with m/z 405.9 3ϩ . In addition, the tandem mass spectrum of the peptide at m/z 405.9 3ϩ in the macroH2A1 digest (Fig. 5D) contained the same major fragment ions as the MS-MS spectrum of this peptide from the macroH2A1.2-GFP digest (Fig. 5C). Therefore, monomethylation at Lys 17 was confidently identified in endogenous macroH2A1. Peptides with the same masses as the peptides containing dimethylated Lys 122 and phosphorylated Thr 128 observed in the tagged protein were also detected in macroH2A1 with the same retention times. Although not definitive proof, this is strong evidence that these two modifications are also present in endogenous protein. The peptides corresponding to mono-and dimethylation of Lys 238 were not observed in the macroH2A1 digest. DISCUSSION Post-translational modifications of histones affect most processes that involve chromatin, including transcription, replication, DNA repair, and DNA segregation. This is also true of histone variants, which recruit DNA repair machinery to lesions, differentiate chromatin at centromeres, and define transcriptionally active or silent domains. However, it remains unclear whether variant histones, which show considerable sequence conservation with the core histones they replace (28), use the same set of modifications as their canonical counterparts. In this study, we investigated covalent modifications on a histone H2A variant, macroH2A1.2, using liquid chromatographic and mass spectrometric analysis of affinitytagged material. MacroH2A1.2-GFP was ubiquitinated on Lys 115 . The difference in electrophoretic mobility of the ubiquitinated and unubiquitinated forms was consistent with monoubiquitination of this protein. Monoubiquitination has been reported for macroH2A1 (17), suggesting that Lys 115 is the site of this modification on the endogenous protein. In addition, Lys 17 monomethylation (Lys 17 me1), Lys 122 dimethylation (Lys 122 me2), and Thr 128 phosphorylation (Thr 128 P), and Lys 238 mono-and dimethylation (Lys 238 me1 and Lys 238 me2) were identified in macroH2A1.2-GFP.
We confirmed that Lys 17 me1, Lys 122 me2, and Thr 128 P occurred on endogenous material using tandem mass spectrometry to analyze protein immunopurified with antiserum that recognizes both isoforms of macroH2A1. Separation of macroH2A1.1 and macroH2A1.2 prior to mass spectrometric analysis will be necessary to determine whether these modifications occur on one or both of these isoforms. The peptides containing Lys 238 me1 and Lys 238 me2 were not detected on endogenous macroH2A1. The peptides containing these modifications were less abundant in the macroH2A1.2-GFP preparation, and limiting quantities of immunopurified macroH2A1 may have precluded their detection. MacroH2A1.2-GFP was correctly incorporated into chromatin and displayed its hallmark enrichment on the inactive X chromosome, suggesting that Lys 238 me1 and Lys 238 me2 are also likely to be of biological relevance.
Comparison of the post-translational modifications that occur on the histone domain of macroH2A1 and canonical histone H2A reveals a divergent modification map without any identical modifications between these histones (Fig. 6). This lack of overlap in covalent modifications is in part due to the discrepancy in their primary amino acid sequences. Six of the 10 documented modification sites in canonical H2A, Lys 5 , Lys 13 , Lys 15 , Lys 95 , Arg 99 , and Thr 120 , are not conserved in macroH2A1, and one of the three modification sites in the histone domain of macroH2A1, Lys 17 , is not conserved in H2A. Lys 36 and Lys 119 (H2ALys 36 and H2ALys 119 ) are conserved, and only the peptides with unmodified residues were observed in the LC-MS analysis of macroH2A1. Lys 36 resides in the region where the major structural difference between macroH2A1-and H2A-containing nucleosomes occurs (29), suggesting that this region is important for regulation of chromatin structure. The sequences surrounding H2ALys 36 are not conserved in macroH2A1, suggesting that the histone acetyltransferase that recognizes H2ALys 36 may not recognize the corresponding residue in macroH2A1. In addition, the non-histone domain of macroH2A1.2 interacts with histone deacetylases 1 and 2 (29), which may inhibit accumulation of acetylation on macroH2A1. Consistent with this idea, acetylation was not detected on macroH2A1.2-GFP or macroH2A1 and was detected on H2A peptides (data not shown). The lack of acetylation on macroH2A1 may also be due to its enrichment in silent chromatin, which is generally hypoacetylated (29).
H2ALys 119 can be ubiquitinated and acetylated, and ubiquitination was detected on Lys 115 of macroH2A1.2-GFP, which is adjacent to the residue of macroH2A1 that corresponds to H2ALys 119 . Ubiquitination of H2ALys 119 is mediated by Ring1 and Ring1b, two mammalian Polycomb proteins that are enriched on the inactive X chromosome (21, 30 -32). H2A ubiquitinated at Lys 119 is enriched on the inactive X chromosome, and this enrichment first occurs when transcriptional silencing is initiated (21,31,32). This suggests the possibility that macroH2A1 Lys 115 ubiquitination may also be mediated by Polycomb proteins and may contribute to silencing of the inactive X chromosome. In addition to monoubiquitination, macroH2A1.2 is also subject to poly-or multiubiquitination mediated by a CULLIN3/SPOP ubiquitin E3 ligase complex (17). With the identification of the monoubiquitination site, it will be possible to determine whether the same residue in macroH2A1.2 is used as the site for both types of ubiquitination. Lys 122 me2 and Thr 128 P reside in the region of macroH2A1 that binds SPOP, which plays a role in the correct localization of this variant histone (17,33). These modifications are correctly positioned to regulate the macroH2A1.2-SPOP interaction and modulate localization of macroH2A1.2.
Histone variants are under intense investigation and are recognized for their roles in transcription, DNA segregation, DNA repair, and chromatin structure. As modifications of canonical histones are also involved in many aspects of chromatin metabolism, the post-translational modifications on histone variants will likely prove to be equally important. Already H2A.X phosphorylation is recognized as a critical step during DNA damage repair (34), and acetyl-lysine in H3.3 is associated with domains of active transcription in Drosophila (35). Vertebrate-specific macroH2A1 is unique among histone variants due to its large C-terminal non-histone domain. It is associated with facultative heterochromatin and plays a role in transcriptional silencing. Determining how these newly discovered modifications affect localization and function of macroH2A1 will provide new insights into this very interesting family of H2A variants.