A
prebleach postbleach 2 min 60 min 300 min
GFP-H2A
mH2A2-GFP
mH2A1.2-GFP
mH2A1.1-GFP
0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1.0
0 50 100 150 200 250 300
mH2A1.1‐GFPn=8
mH2A1.2‐GFPn=14
mH2A2‐GFPn=12
GFP‐H2An=24
GFPn=17
Rel
ativ
e In
tens
ity
Time (min)
B
Supplemental Figure 1
Supplemental Figure 1. FRAP analysis of canonical H2A and mH2A isoforms. (A) Representative images from long-term FRAP time series of HeLa cells transiently transfected with GFP-H2A or mH2A-GFP isoforms. Squares of 5 µm x 5 µm within the nucleus were photobleached and fluorescence recovery followed for at least 5 hours. Prebleach, postbleach, and 2min, 60min, 300min postbleach are shown. (B) Quantitative FRAP evaluation. mH2A1 isoforms (orange and yellow) display a decreased recovery compared to canonical H2A (blue). mH2A2 (red) displays decreased recovery compared with canonical H2A and mH2A1 isoforms. In contrast, GFP alone shows full recovery within less than a minute (green).
Protein H2A peptide count
macroH2A peptide count
Importin 4 4 19
Importin 7 25 50
Importin 8 3 14
Importin 9 120 42
KapA1 48 141
KapA2 66 140
KapA3 33 62
KapA4 20 54
KapA5 44 133
KapA6 45 127
KapB1 141 104
Exportin 1 62 97
Exportin 5 3 11
Exportin 7 0 6
Transportin 4 0
NuclearImportandExportFactorsC
B
130
9572
56
36
28
17
Imp9
ATRX
Ncln
m1.2-GFP
Npm
H2A-GFP
Imp9
NAP1 SET
A
D
Supplemental Figure 2
Supplemental Figure 2. H2A and mH2A1.2 associate with nuclear-cytoplasmic shuttling factors and chromatin-associated factors in chromatin-free extracts. (A) Proteins interacting with either GFP-H2A or mH2A1.2-GFP in chromatin-free extracts were resolved on 4-12% NuPAGE and silver stained. Each lane was cut into ten slices and MS analyzed. Chromatin-associated factors identified by MS are presented alongside the gel (see Fig. 1D for peptide counts). Blue characters = H2A or mH2A1.2-specific factors; black characters = chromatin-associated factors which interact with both H2A and mH2A1.2; green = GFP histones. (B) Table of nuclear-cytoplasmic shuttling factors that interact with H2A or mH2A1.2 in chromatin free extracts. These data were collected from MS analysis of the gel presented in (A) and Supplemental Table 1. (C) Chromatin-free IP and silver staining of proteins interacting with GFP-H2A or mH2A1.2-GFP resolved on 4-12% NuPAGE. Unique bands were excised and MS analyzed. Arrow indicates Imp9; molecular weight marker shown on right. (D) GFP, GFP-H2A or mH2A1.2-GFP IPs were resolved on 12% PAGE and IB’d for Imp9 (upper panel) or GFP (lower panel).
!Imp9
mH2A
1.2-
GFP
H2A
-GFP
GFP
mH2A
1.2-
GFP
H2A
-GFP
GFP
Input IP
mH2A1.2-GFP
GFP-H2A
GFP
!GFP
2492
1200
600
1800
300
900
1500
2100
(161-292) ADD (ATRX-DNMT3-DNMT3L)
(1189-1326) DID Daxx-interacting domain
(1441-1470) Acidic Patch
(1574-2136) ATPase Domain
163
‐71EEGTSSSEK
360‐73
LIETTANMNSSYV
K39
8‐43
1SV
LADIKKA
HLALEED
LNSEFRAMDAV
NKEKN
TK
809‐27
RQTQ
SESSNYD
SELEKEIK
871‐98
TSQ
EGSSDDAER
KQER
ETFSSA
EGTV
DK
1085
‐92NGAYGRE
K
1145
‐67RN
TKEIQSG
SSSSDAEESSED
NK
1179
‐86AV
IVKEKK
1346
‐60LTVS
DGESGEEKK
TK
1451
‐82EEEEEEEEEEEEEEED
ENDDSKSPGKG
RKKIR
1662
‐97RP
QER
SYMLQ
RWQED
GGVM
IIGYEMYR
NLAQGRN
VK
1727
‐44NEA
SAVS
KAMNSIRSRR
R
2166
‐79FLAQ
GTM
EDKIYD
R
2227
‐34RD
TPMLPK
2361
‐88EN
MNLSEA
QVQ
ALALSRQ
ASQ
ELDVK
RR
(561-595) HP1α- interacting domain
C A
Protein H2A
Peptide counts
mH2A1.2 Peptide counts
NAP1 32 72
SET 2 19
Nucleolin 79 81
Nucleophosmin 7 42
ATRX 0 19
Importin 9 120 42
B
αGFP
αATRX
αH3
Input IP
Supplemental Figure 3
Supplemental Figure 3. ATRX interacts with mH2A in chromatin-free extracts. (A) Peptide counts of chaperones identified by MS for H2A and mH2A1.2 (See corresponding gel in Supplemental Fig. 2A). (B) ATRX peptides identified by MS analysis in mH2A1.2 IPs from chromatin-free extracts. Peptides presented next to a schematic of ATRX protein with its various domains shown. Nineteen peptides in total were identified (15 are presented here as 4 peptides were overlapping). (C) IBs of GFP and ATRX from chromatin-free extracts by independent chromatin fractionation method (Mendez and Stillman 2000).
Full length ATRX N-term (1-841) Middle (800-1670) C-term (1670-2942)
Supplemental Figure 4
ATRX fragments
Supplemental Figure 4. Expression of ATRX deletion constructs. Fluorescent images of GFP-tagged ATRX constructs transiently expressed in HEK293 cells. While the N-terminal construct expressed strongly, enrichment of mH2A1 signal observed in the N-terminal IP (Fig. 2C, D) suggests that this region is sufficient for binding to mH2A1.
αATRX
Amido Black
A B
HEK 293
HeLa 1.2.11
αATRX
Amido Black
DAPI αATRX Merge
sh92
shluc
C
Supplemental Figure 5
Supplemental Figure 5. shRNA-mediated depletion of ATRX in 293 and HeLa1.2.11 cells. (A) αATRX IB of HEK 293 cells expressing five lentivirally introduced ATRX shRNAs, and control shluc. sh90 and sh92, used for subsequent studies, are highlighted in red; Amido black of histones shown for loading. (B, C) ATRX depletion in HeLa1.2.11 cells shown by IB (B) and immunofluorescence (C). Note presence of ATRX at telomeres.
657.0 657.5 658.0 658.5 659.0 659.5 660.0 660.5 661.0
m/z
0
5
10
15
20
25
30
35
40
45
50
55
60
65
70
75
80
85
90
95
100
Rel
ativ
e A
bund
ance
657.382
657.884
659.898
658.391 660.399
661.066658.900 659.392660.900
659.723656.719 658.718 661.401658.326 659.052 660.730657.771 660.065657.047 661.321659.463657.311
659.312658.636657.436 658.061 660.129
660.352
ATRX sh_ 92
sh_luc
Supplemental Figure 6. ATRX knockdown results in mH2A1 accumulation in chromatin. Comparison of an averaged full mass spectrum for the +2 charged mH2A1 peptide (SAKAGVIFPVGR) from sh92 ATRX knockdown (d0-labeled) and shluc control (d5-labeled) histones extracted from chromatin.
Supplemental Figure 6
0 1 2 3 4 5 6
luc 90
mH
2A1/
IgG
shluc
sh90
Probe: telomere repeat
shRNA:
A B
shluc
sh90
Probe: telomere repeat
sh92
0 1 2 3 4 5
luc 90 92 m
H2A
1/In
put
shRNA:
HEK 293 K562
Supplemental Figure 7
Supplemental Figure 7. Loss of ATRX results in telomeric accumulation of mH2A1. (A) One of two representative ChIP-telomere Southern blots for HEK293 cells shows the increased association of mH2A1 with telomeric chromatin in the absence of ATRX. Only sh90 was used as it induced the most efficient knockdown in this cell line (see Supplemental Fig. 4). (B) Biological replicate of telomere Southern blot in K562 cell line. mH2A1’s presence at telomeres of K562 cells is increased in the absence of ATRX in both sh90 and sh92 lines. Densitometry quantitation presented below each graph.
Supplemental Figure 8
0
5
10
15
20
luc 90 92
ATRX
0 2 4 6 8
10 12 14
luc 90 92
mH2A1
shRNA:
Normalized
toGAPD
H
A
B
Supplemental Figure 8. ATRX knockdown results in reduced RNA levels of sub-telomeric chromosome 16 genes. (A) shRNA-mediated knockdown of ATRX (sh90 and sh92) in K562 cells results in the loss of ATRX mRNA, compared to shluc, without affecting mH2A1 mRNA. (B) Knock down of ATRX results in decreased mRNA levels of genes found in the α globin cluster of sub-telomeric chromosome 16. Starting with the most telomere proximal, genes assayed include POLR3K, MPG, C16orf35, ITFG3, TMEM8A, NME4 and DECR2. Genes assayed are circled in blue in the UCSC browser above. CDK8 transcription (chromosome 13) is unaffected by ATRX knockdown, similar to mH2A1 (chromosome 5) in (A). Expression was measured relative to GAPDH and to the control shluc, whose expression values were arbitrarily set as 1.
POLR3K MPG C16orf35 ITFG3 TMEM8A NME4 DECR2 CDK8
No
rm
alized
to
GA
PD
H
0
0.2
0.4
0.6
0.8
1
1.2
1.4 sh_lu
c
sh_90
sh_92
0
0.2
0.4
0.6
0.8
1
1.2
1.4
sh_lu
c
sh_90
sh_92
0
0.2
0.4
0.6
0.8
1
1.2
1.4
sh_lu
c
sh_90
sh_92
0
0.2
0.4
0.6
0.8
1
1.2
1.4
sh_lu
c
sh_90
sh_92
0
0.2
0.4
0.6
0.8
1
1.2
1.4
sh_lu
c
sh_90
sh_92
0
0.2
0.4
0.6
0.8
1
1.2
1.4
sh_lu
c
sh_90
sh_92
0
0.2
0.4
0.6
0.8
1
1.2
1.4
sh_lu
c
sh_90
sh_92
0
0.2
0.4
0.6
0.8
1
1.2
1.4
sh_lu
c
sh_90
sh_92
Supplemental Figure 9
150
100
50
0
[FU]
mH2A‐LucmH2A‐92Input‐Luc
35 100 200 300 400 6001000 10380 [bp]
A
sh92 shluc Input (shluc) ATRX Input (ATRX)
Raw reads 67219237 56540184 148165330 9139129 18232783
Bowtie alignments
(wiggle) 69827565 59143870 145176252 5563975 18468175
Alignments analyzed (MACS)
51200836 45308710 63296750 4654137 17770258
Total peaks count
(MACS) 133763 127000 - 16848 -
Total peak length (bp)
(MACS) 113477382 94296714 - 12914783 -
B
Supplemental Figure 9. ChIP-sequencing of mH2A in shluc and sh92 K562 cells. (A) Bioanalyzer traces of ChIP (mH2A1) and Input DNA from MNase digested chromatin (High sensitivity DNA chip; Agilent Technologies). Isolated mononucleosomal DNA (arrow) was size selected for ChIP-Seq library preparation. Input (red), sh92 (blue) and shluc (green). Peaks at 35bp and 10380bp are internal size markers. (B) Raw number of reads obtained by Illumina Hi-Seq, total number of alignments (Bowtie: -m 20 -k 20 -n 2 -l 50), and alignments used for peak calling (MACS) for sh92, shluc and Input shown. Raw reads for the ATRX ChIP-Seq were downloaded from GEO (GSE22162). Also shown are the total number of peaks for sh92 and shluc (p value cutoff = 1.00e-4) and total number of base pairs covered by significant peaks (MACS). (C) Overlap of significant mH2A1 peaks (MACS) from shluc and sh92 ChIP-seq data sets (HOMER software, Heinz et al. 2010).
shluc
65413
~54%
sh92
74145
~57%
55397
C
B
Supplemental Figure 10 A
!"
!#$"
!#%"
!#&"
!#'"
!#("
!#)"
!#*"
+(!!!" +'!!!" +&!!!" +%!!!" +$!!!" !" $!!!" %!!!" &!!!" '!!!" (!!!"
!"#$%&''&()$#"*+,$-(.)/!
0&,/*)-"$/($122$3456!
7.''$8")"$,"/!
,-./%"
,-.012"
34516"
789:"
Full
High
Low
TSS TES
-4000 4000 -2000 2000 0 -4000 4000 -2000 2000 0
position
0.00
0.02
0.04
0.06
0.08
0.10
0.00
0.02
0.04
0.06
0.08
0.10
0.00
0.02
0.04
0.06
0.08
0.10
Norm
aliz
ed R
ead C
ount (p
er
Mill
ion r
eads)
!"#$%&%'()%!"#$%&%*#%
+,-(.%&%'()%
!"
!#$"
!#%"
!#&"
!#'"
!#("
!#)"
!#*"
+(!!!" +'!!!" +&!!!" +%!!!" +$!!!" !" $!!!" %!!!" &!!!" '!!!" (!!!"
!"#$%&''&()$#"*+,$-(.)/$
0&,/*)-"$/($122$3456!
7&89$":5#",,"+$8")",!
,-./%"
,-.012"
34516"
789:"
!"
!#$"
!#%"
!#&"
!#'"
!#("
!#)"
!#*"
+(!!!" +'!!!" +&!!!" +%!!!" +$!!!" !" $!!!" %!!!" &!!!" '!!!" (!!!"
!"#$%&''&()$#"*+,$-(.)/$
0&,/*)-"$/($122$3456!
7(8$"95#",,"+$:")",!
,-./%"
,-.012"
34516"
789:"
Supplemental Figure 10. TSS/TES analysis and peak overlap of mH2A1 and ATRX. (A) Read counts (200bp window) normalized to total number of reads (counts per million reads), plotted against the distance (-5Kb, +5Kb), from the nearest annotated Transcription Start Site (TSS, left), and Transcription End Site (TES, right). Genes were grouped by expression levels to Full (all annotated genes), High and Low categories based on K562 RNA-Seq data from the ENCODE project. sh92 (black), shluc (red), and Input (green). (B) Same TSS analysis as in (A) also containing ATRX (blue). Scale is different than in (A). (C) Overlap of significant mH2A1 and ATRX peaks (MACS) in shluc and sh92 samples (Law et al. 20120; Heinz et al. 2010).
C
mH2A1_shluc
125773
~99%
ATRX
15643
~93%
1141
mH2A1_sh92
132423
~99%
ATRX
15531
~92%
1245
Supplemental Figure 11
B C
Supplemental Figure 11. mH2A1 is enriched at the α globin cluster. (A) Capture of UCSC genome browser showing ~50kb region around α globin locus. ChIP –seq analysis of mH2A1 in K562 cells. Loss of ATRX (sh92) results in redistribution of mH2A1 compared to control cells (shluc) as shown in Fig. 5D (shown on the X-axis genomic position in Kb; Y-axis alignment counts in 500bp window sliding 250bp. Window counts are normalized to total number of alignments and scaled by 10^7 for mH2A1 and 10^6 for ATRX). ChIP-seq data for mH2A1 was compared to published ATRX ChIP-seq data (Law et al. 2010). Shluc Input enrichment is shown for reference. Threshold line set at 35 to facilitate visualization. Regions of significant enrichment are indicated by black bars below the respective enrichment plots. UCSC custom tracks are shown, RefSeq gene annotation, K562 Chromatin State Segmentation by HMM from ENCODE/Broad, Repeat elements by RepeatMasker, and Duplications of >1000 Bases of Non-RepeatMasked Sequence are shown, respectively. Primer pairs used in B (native ChIP) and C (fixed ChIP) shown on top. As the HBA genes are duplicated, primer pair 6 at the TSS is presented twice. (B) Validation by qPCR of the ChIP-sequencing library indicates enrichment of mH2A1 at regions upstream of the HBA1/2 genes. (C) ChIP via standard formaldehyde cross-linking and sonication demonstrates increased mH2A1 levels 1kb upstream and at TSS of the HBA1/2 genes when ATRX is depleted. Mock = no antibody. One of three replicates (two biological and one technical) shown. The stronger enrichment of mH2A1 signal in native ChIP (B) compared with formaldehyde-fixed ChIP (C) is likely due to greater antibody affinity for mH2A1 in the native ChIP protocol.
A
0
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
0.45
0.5
-1kb TSS
mH2A1 mock mH2A1 mock
Pe
rce
nta
ge
In
pu
t
!"#$%&
!"'(&
!"')&
!"
#!"
$!"
%!"
&!"
'!"
(!"
-27kb -5kb -4kb -1kb"
"Primer position: 1 2 3 4 5 6 "
Pe
rcen
tag
e In
pu
t" shluc"
sh92"mH2A1 ChIP"
190000 200000 210000 220000 230000
Nor
mal
ized
alig
nmen
ts c
ount
(5
00bp
win
dow
, slid
ing
250b
p)
CHR16:
HBZ HBA2 HBQ1 HBM HBA1 C16orf35
mH2A1 (sh92)
mH2A1 (shluc)
ATRX
Input (shluc)
RefSeq 0 -
0 -
0 -
0 -
70 -
70 -
60 -
70 -
SINE LINE LTR
DNA SIMPLE
L. COMPLEXITY
20Kb qPCR amp:
Sign
ifica
nt p
eaks
Transcription
Promoter Enhancer
Dup >1000bp
35 -
35 -
1 2 3 45 6 6
<
<
<