Dimeric structure of p300/CBP associated factor
BMC Structural Biology
Dimeric structure of p300/CBP associated factor
Shasha Shi 0
Juanyu Lin 0
Yongfei Cai 0
Jiao Yu 0
Haiyan Hong 0
Kunmei Ji 3
Jennifer S Downey 2
Xiaodong Lu 0
Ruichuan Chen 0
Jiahuai Han 0 1
Aidong Han 0 1
0 State Key Laboratory of Cellular Stress Biology, School of Life Sciences, Xiamen University , Xiamen 361102 , China
1 Department of Biomedical Sciences, School of Life Sciences, Xiamen University , 3 S. Xiangan Road, Xiamen, Xiangan 361102 , China
2 Division of Biomedical Science, Herman Ostrow School of Dentistry of University of Southern California , Los Angeles, CA 90089 , USA
3 School of Medicine, Shenzhen University , Shenzhen, Guangdong 518060 , China
Background: p300/CBP associating factor (PCAF, also known as KAT2B for lysine acetyltransferase 2B) is a catalytic subunit of megadalton metazoan complex ATAC (Ada-Two-A containing complex) for acetylation of histones. However, relatively little is known about the regulation of the enzymatic activity of PCAF. Results: Here we present two dimeric structures of the PCAF acetyltransferase (HAT) domain. These dimerizations are mediated by either four-helical hydrophobic interactions or a -sheet extension. Our chemical cross-linking experiments in combined with site-directed mutagenesis demonstrated that the PCAF HAT domain mainly forms a dimer in solution through one of the observed interfaces. The results of maltose binding protein (MBP)-pulldown, co-immunoprecipitation and multiangle static light scattering experiments further indicated that PCAF dimeric state is detectable and may possibly exist in vivo. Conclusions: Taken together, our structural and biochemical studies indicate that PCAF appears to be a dimer in its functional ATAC complex.
PCAF; Histone acetyltransferase; Dimerization; ATAC
The unique posttranslational modification patterns on
histones have been conceptualized as epigenetic codes
that may finely tune transcription of specific genes [1,2].
Histone acetyltransferases (HATs), including p300/CBP
and PCAF/GCN5, are responsible for modification of
histones by acetylation on the exposed lysines [3-5].
PCAF/GCN5 are important members of histone
acetyltransferases. Homozygous GCN5 knockout mice died
during embryogenesis, while the majority of PCAF
knockout mice developed normal [6,7]. However, the PCAF
knockout mice later showed memory impairment,
psychological anxiety and defects in stress control [8,9].
Interestingly, a single nucleotide polymorphism in the PCAF gene
was found in patients with coronary heart abnormalities
that resulted in vascular morbidity and mortality .
Metazoan PCAF/GCN5 proteins have three conserved
domainsN terminal extension region, HAT domain and
bromodomain (BRD) . The HAT and BRD are also
highly conserved in yeast and plants [12,13]. The HAT
domain of PCAF/GCN5 has a globular structure that
contains an acetyl-CoA binding pocket [14,15].
AcetylCoA binds the HAT domain through the pyrophosphate
body and pantetheine arm [16,17]. Substrates of histones
and non-histones, such as p53, induce large
conformational changes in the active pocket of the PCAF HAT
domain through extensive interactions that anchor
specific lysines for acetylation [16,18]. Interestingly, PCAF/
GCN5 have slightly different specificity. GCN5 acetylates
histones H3 and H4 with favorable sites of lysines 9 and
14 on histone H3 and lysines 8 and 16 on histone H4
. In comparison, PCAF mainly acetylates lysine 14
on H3  and also specifically acetylates p53 at lysine
320 to enhance responses to DNA damage [21,22].
More importantly, the metazoan PCAF/GCN5 are
usually present in the megadalton complexes SAGA
(Spt-Ada-GCN5-acetyltransferase) and ATAC
(Ada-TwoA containing complex) [11,23]. The additional
components Ada2 and Ada3 (alteration/deficiency in activation)
are required to form a minimal core complex that can
efficiently acetylate histone octamer and nucleosome [24,25].
In addition to a PCAF binding domain, Ada2 has a
SWIRM domain that binds nucleosomal DNA . A
SANT domain at Ada2 N terminus has been proposed to
direct histone tails to specifically associate with PCAF
catalytic site [27-29]. Through a structurally unknown C
terminal domain Ada3 physically associates with Ada2 in
the core complex . However, the molecular
mechanism of this core complex in the regulation of PCAF/
GCN5 catalytic activity and specificity remains largely
In this report we present structure-based biochemical
characterization of the PCAF HAT domain. We show
that the PCAF HAT domain can form a dimer in a
concentration dependent manner. All our experimental data
suggest that PCAF may exist as a possible dimer in its
The PCAF HAT domain (amino acids 493-658) was
purified and crystallized in a unique condition that
contained 1.2 M ammonium sulfate and 0.2 M lithium sulfate.
The structure was solved by molecular replacement. To
our surprise, we found four HAT domains in one
asymmetric unit (Additional file 1: Figure 1S). The structure of
each HAT domain remains almost the same globular
folding (root mean square deviation (rmsd) of 0.032-0.035 for
C backbone). The overall structure is similar to the
PCAF structure (rmsd of 0.397 for C backbone) solved
by Marmorstein and his co-workers , except for the
loop 1 (L1), which was not defined due to lack of clear
electron density (Additional file 1: Figure S2A and S2B,
indicated by arrow). Our PCAF structure is also similar to
human GCN5 (rmsd of 0.34 for C backbone)  and
Tetrahymena GCN5 (rmsd of 0.902 for C backbone)
. In addition to the L1, loop 2 (L2) is also significantly
different from Tetrahymena GCN5, which is bound by a
histone peptide (Additional file 1: Figure S1C and S1D,
indicated by arrow).
We found that the four PCAF HAT domains in one
asymmetric unit form two different dimers (Figure 1A and
B). The residues of both interfaces were very well defined
on electron density maps (Additional file 1: Figure 3S).
The first dimer is formed by four-helical stacking of
helices 1 and 2 (Figure 1A and C) and has a canonical
interface with a buried surface area of 1530 2. The four
residues Leu512, Val516, Thr535 and Phe539 from both
monomers form a central hydrophobic core, while Gln519
and Asn520 create four pairs of hydrogen bonds. In
addition, Met513 and Thr535 or two His524 residues
located at the periphery of the interface make additional van
der Waals contacts.
Figure 1 Dimeric structures of the PCAF HAT domain. A) The PCAF dimer structure mediated by a four helical bundle; B) The PCAF dimer
structure mediated by -sheet extension; C) Close-up view of detailed interactions within the PCAF dimeric interface shown in A; D) Close-up
view of detailed interactions in the PCAF dimeric interface shown in B. The structures are presented in ribbon with each monomer colored in
cyan, magenta, green and blue. Helices, -strands and loops are labeled with , and L, respectively. The cofactor acetyl-CoA is represented in
sticks (A and B). Amino acids that are involved in direct contacts are represented as sticks and also labeled (C and D). Hydrogen bonds are
indicated with yellow broken lines.
The second dimer of the PCAF HAT domain is
mediated through the -sheet extension of two antiparallel
1 strands (Figure 1B and D). The major contacts of this
dimer are three hydrogen bonds between amino acids
493-495. This interface is further stabilized by a
hydrogen bond from two His588 residues and van der Waals
contacts from two Ile494 and Tyr585 and His588 from
each monomer. The buried surface area of this interface
is 879 2.
Taken together, the results show that the PCAF HAT
domain forms two distinct dimers. The clusters of
hydrophobic residues contribute their important roles to the
interactions. While the first interface is significant, however,
the second interface may likely be produced during crystal
packing due to the smaller size and insignificant score
calculated by PISA program (Additional file 1: Table S1).
PCAF dimers detected by pulldown experiments
We next wanted to know whether PCAF HAT domain
exists as a dimer in solution. We first used a maltose
binding protein (MBP) affinity pulldown experiment.
The PCAF HAT domain in MBP fusion was used to pull
down another one in His-tag fusion (Figure 2A).
Histagged PCAF was clearly observed in beads fraction
when incubated with MBP-PCAF but not in MBP alone
(lanes 5 and 6) compared with a loading control (lane 1).
In order to analyze PCAF oligomeric state in cells, we
performed a co-immunoprecipitation experiment using
HA and Flag tags (Figure 2B). The full-length PCAF was
tagged with either HA or Flag peptides at its N terminus.
Flag affinity gel was used to bind Flag-PCAF. The
Figure 2 PCAF HAT domain interacts with each other in
solution. A) MBP pulldown experiment using an MBP-PCAF HAT
domain fusion. Purified MBP-PCAF fusion proteins were treated
with 2 M NaCl to remove any bound maltose and mixed with
new amylose beads and His-tagged PCAF fragments (lanes 4-6).
The beads were then washed off unbound proteins and analyzed
by 15% SDS-PAGE. Ten percent inputs of His-tagged PCAF fragments,
MBP alone and MBP-PCAF fusion protein were used as loading
controls (lanes 1-3). B) The presence of PCAF dimer state examined
by co-immunoprecipitation. A plasmid that expresses Flag-PCAF was
co-transfected with an empty vector pLV-nHA (lane 1) or HA-PCAF
expression plasmid (lane 2). The expressions of Flag-or HA-PCAF were
confirmed using 5% input shown at the bottom, while the Flag gel
after wash was checked by western blotting as shown at the top.
HA-PCAF was readily detected when Flag-PCAF was
co-expressed (lane 2), suggesting that over-expressed
PCAF may self-associate to form oligomers in cells.
PCAF dimers disrupted by mutations in interfaces
To confirm the presence of the PCAF dimers in solution
we then used a chemical cross-linking method (Figure 3).
We observed a dominant dimer on SDS-PAGE after the
PCAF HAT domain was incubated with a cross-linker
disuccinimidyl suberate (DSS) (lane 2, indicated by green
arrow). A ladder of higher molecular-weight oligomers
was also appeared (lane 2, indicated by red arrow).
We continued to examine whether mutations in PCAF
dimerization interfaces would affect its dimers
formation. To disrupt the first interface, we generated a triple
Figure 3 PCAF HAT domain dimer detected by chemical
cross-linking. A) Chemical cross-linking of the full-length PCAF
HAT domain (amino acids 493-658). WT represents the full-length
wt PCAF HAT domain (lanes 1-2). LV512DD is a full-length HAT
domain with L512D and V516D double mutations (lanes 3-4).
LMV512AAA is a full-length HAT domain with L512A/M513A/
V516A triple mutations (lanes 5-6). F539A is a full-length HAT domain
with single F539 to alanine mutation (lanes 7-8). B) Cross-linking of a
shortened PCAF HAT domain (amino acids 496-658) in comparison
with a full-length PCAF HAT domain mutant. LVF512DDA is a
combined mutant of LV512DD and F539A in the full-length PCAF
HAT domain (lanes 3-4). LV2DD is a double mutation (L512D and
V516D) in the shortened PCAF HAT domain (lanes 5-6) while
LVF2DDA is a triple mutation (L512D, V516D and F539A) in the
shortened HAT domain (lanes 7-8). The arrows indicate monomer
(black), dimer (green) and higher oligomers (red). Lanes where DSS was
added are indicated above the gel and M denotes a protein molecular
mutant LMV512AAA (Leu512, Met513 and Val516 to
alanine) (Figure 1C). However, this mutant did not appear
to affect the PCAF dimerization (Figure 3A, lane 6). In
comparison, a mutant with the Leu512 and Val516 altered
to aspartic acid (LV512DD) and Phe539 to alanine
(F539A) did have small effects on the presence of PCAF
dimer and higher oligomers (Figure 3A, lanes 4 and 8). An
LV512DD and F539A combined mutant LVF512DDA was
clearly less prone to form a dimer (Figure 3B, lane 4).
To disrupt the second interface, we removed the first
three N terminal amino acids (amino acids 493-495)
that are responsible for -sheet extension (Figure 1D).
However, the shortened PCAF HAT domain itself
behaves as the wt PCAF in this experiment (Data not
shown). The shortened PCAF HAT domain (amino
acids 496-658) was then combined with mutations of
the first interface studied above (Leu512 to Asp,
Val516 to Asp and/or Phe539 to Ala) to generate two
mutants (LV2DD and LVF2DDA). As expected, both
mutations, in particular the LVF2DDA, clearly suppressed
dimer formation of the HAT domain (Figure 3B, lanes 6
These data support our crystallographic observation
that both interfaces contribute to the dimeric formation
of PCAF in our cross-linking experiment. However, the
second interface may result from a crystal packing as
suggested by our crystallographic analyses since the
deletion mutant (amino acids 496-658) did not change its
ability to form a dimer. It is important to note that our
introduction of these mutations into PCAF did not affect
enzymatic activity of PCAF HAT domain (Additional
file 1: Figure S4).
PCAF dimer detected by static light scattering
To seek more evidence for presence of dimers in
solution, we analyzed PCAF HAT domain using Multiangle
static light scattering (MALS), which measures absolute
molecular weight (MW) of a particle without any
assumptions. A protein is first separated on a HPLC that
is directly connected with detectors for measuring
differential refractive index (dRI) and light scattering (LS).
The theoretical MW of the HAT domain (amino acids
493-658) including N terminal His tag is 21.8 kDa. Here
three different protein concentrations at 2, 6 and 20 mg/
ml were used in this experiment (Figure 4). The peak 1
was analyzable only at 20 mg/ml, which was determined
to be 49.2 kDa. The major portion of this protein
remained in peak 2, which had the closest MW to
PCAF HAT domain monomer at 2 mg/ml. The MW of
peak 2 was 27.8 kDa at 6 mg/ml and 30.7 kDa at
20 mg/ml, suggesting that part of large particles from
peak 1 was not separable. All these data indicate that
PCAF HAT domain may remain in a dimer-monomer
equilibrium with dominant monomeric species in solution.
Protein acetylation is well known for its role in
epigenetic regulation of transcription and is also involved in
translation, protein turnover, localization and quality
control, thus linking acetylation to a variety of biological
processes such as cell shape, migration and autophagy
. Moreover, PCAF/GCN5 family members have been
implicated in carcinogenesis and drug targets for cancer
therapy . Marmorstein and his coworkers have
extensively studied the structures and enzymatic
mechanism of the catalytic domain and histone binding of these
histone acetyltransferases [15,34]. Here in this report,
our structural and biochemical analyses demonstrate
that PCAF can exit as a dimer.
We solved the crystal structure of the PCAF HAT
domain in two different dimeric states. One of these dimeric
interfaces is large (more than 1500 2) created by several
hydrogen bonds and hydrophobic contacts (Figure 1). The
second interface is rather small, which may likely be
generated during crystal packing (Additional file 1: Table S1).
The crystals in this study were grown in 0.2 M lithium
sulfate and 1.2 M ammoniumn sulfate at pH7.5, which were
packed in the space group P43. The same human PCAF
HAT domain was crystallized with space group P64 in a
slightly different condition and no interfaces were found
between its symmetric or asymmetric units . We also
crystallized human GCN5 HAT domain in a simple
precipitant of 35% tacsimate pH 7.0 from Index screen kit
(Hampton Research) and the crystal was packed in a space
group I422 (a = b =129.2 and c = 179 ) (Data not shown).
Consistently, we found 4 GCN5 monomers in one
asymmetric unit that form the same dimeric interfaces as those
described above for PCAF (Additional file 1: Figure S5A
for the major interface). The GCN5 dimer was well
aligned with the PCAF with rmsd of 1.7 (Additional
file 1: Figure S5B).
For comparison, we listed all crystal structures of
PCAF/GCN5 homologues from tetrahymena, yeast and
human (Additional file 1: Table S2). In one crystal form
of a tetrahymena GCN5 HAT domain, a dimeric
interface is also formed through anti-parallel helices H1 and
H2 between two HAT domains in one asymmetric unit
(Additional file 1: Figure S6A and S6B). Unfortunately,
based on our PISA analysis, this interface is likely
produced by a crystal packing (Additional file 1: Table S1).
Consistently, two critical residues Met513 and Gln519
that are responsible for dimerization in human PCAF
are replaced with two lysines in tetrahymena GCN5
(Additional file 1: Figure S6C).
In order to confirm whether the human PCAF HAT
domain appears dimers in solution, we performed a
series of experiments, including MBP pulldown,
crosslinking and MALS (Figures 2, 3, 4). All these data
supported that the PCAF HAT domain is able to form a
Figure 4 Molecular weight of PCAF HAT domain determined by MALS. A) A running profile of PCAF protein at concentration of 20 mg/ml.
PCAF HAT domain was first run on a gel filtration column detected by light scattering (LS11, shown in red) and refractive index (dRI, shown in
blue). The X axis is a running time taken from the HPLC. The LS11 (one of 18 light scattering detectors) and dRI intensities are aligned and scaled.
A relative scale is shown as the Y axis. Peak 1 is small and thus highlighted by an arrow. B) The molecular weight of the PCAF HAT domain at
different concentrations. The measurements are colored in green (2 mg/ml), blue (6 mg/ml) and red (20 mg/ml). The dotted lines represent
averaged values for molecular weight calculated by 18 laser detectors at each time point. The X axis is the same HPLC running time as (A) and
the Y axis indicates molecular weight. C) The molecular weights of two PCAF particles in solution. The number in parenthesis is an error rate,
which was calculated using the measurements of all time points for each peak in panel B. ND is Not able to Determine because of low signal.
dimer in solution. MALS experiment, however, indicated
that the dimer forms only at low ratio because it could
only be readily detected at higher protein concentration
(Figure 4). The apparent MW was close to the
theoretical one only at the lowest protein concentration,
suggesting the larger MW was likely resulted of presence of
the PCAF dimer.
Interestingly, the N terminal BRD domain of BRD2
protein has been found to be a dimer, allowing to
significantly enhance the binding to histone H4 and aides with
further recognition of the hypoacetylated H4K8 [35,36].
Indeed, the BRD domains of the PCAF/GCN5 family
members also form dimers in crystallographic conditions
(3D7C and 3GG3, Structural Genomics Consortium),
which have relatively large interfaces (Additional file 1:
Table S1). Even though their significant scores are 0, a
synergistic effect of the HAT and BRD domains in dimer
formation may possibly stabilize PCAF oligomeric state
Since PCAF/GCN5 always exist in megadalton
complexes, one possibility is that PCAF dimerization may
help to better associate with a nucleosome for efficient
histone acetylation, which is depicted in our proposed
model (Figure 5). PCAF/GCN5 attach to the
nucleosomal DNA through Ada2 SWIM domain. PCAF/GCN5
bromodomain dimer may further dock the histone H4
through the H4K5 and H4K15 sites, which are acetylated
before incorporated into the nucleosome . The HAT
domain dimer then recruits a histone H3 tail for
acetylations at H3K14 and neighbor sites, including H3K9 and
Figure 5 A hypothetical model of core PCAF/GCN5 complex
interacting with nuleosome for histone acetylation. Shown in
the bottom is a half nucleosome, which is made from the crystal
structure solved by Luger et al . An oligomer of histone H3 and
H4 is wrapped with two double-strand DNA helices. The N terminal
tails of H3 and H4 are extended manually. The three subunits of the
PCAF/GCN5 core complex, including Ada2 and Ada3, are drawn
schematically as a heterohexamer. The arrows indicate experimentally
verified interactions between histone tails, DNA double strands and all
H3K18. The SANT domain of Ada2 may directly
associate with the H3 and H4 tails and help to better position
these for the acetylation process. Therefore, all these
domains contribute to the overall acetylation level and
specificity of histones even though their interactions with
histones and DNA are extremely weak per se. Interestingly,
the 3D reconstruction of complex SAGA has positioned
GCN5 and other BRD-containing proteins in adjacent
region, which may lead to a better association of SAGA and
histone tails through multiple interactions .
In summary, our structural and biochemical studies
suggest that the PCAF/GCN5 HAT domain can form a dimer
in solution. We propose a model that this dimerization
may be important for acetylations on specific sites of
histones since multiple contacts may synergistically position
PCAF/GCN5 megadalton complexes on the nucleosome.
An important question whether PCAF/GCN5 are
dimers in their functional complexes is currently under
Protein expression and purification
Gene fragment encoding residues 493-658 of human
PCAF HAT domain was amplified by PCR and
subcloned into pET28a vector (Novagen) using NheI and
XhoI restriction sites. Site-specific mutations and
truncations were made using the modified Quikchange
mutagenesis protocol . The constructs were then used to
express PCAF in Echerichia. coli stain BL21/DE3 (gold)
using 0.25 mM isopropyl--D-thiogalactopyranoside
(IPTG) for 12 h at 25C. PCAF proteins were purified by
nickel affinity agarose (Qiagen) and Superdex 200 (GE
Healthcare) according to the manufactures protocol.
Purified proteins were concentrated down to ~20 mg/ml
using centricons (Millipore) and stored at 80C in a
buffer of 20 mM Tris pH8.0, 100 mM NaCl, 300 mM
ammonium acetate, 5 mM -mercaptoethanol (-ME).
Initial crystals of the PCAF HAT domain were obtained
by screening with JCSG plus (Qiagen) in one condition
of 0.2 M lithium sulfate, 0.1 M Tris pH8.5 and 1.25 M
ammonium sulfate. Further optimizations yielded the
crystals in 0.2 M lithium sulfate, 0.1 M HEPES pH7.4,
1.2 M ammonium sulfate and 10 mM trimethylamine
HCl. The crystals were harvested and snap-frozen in
liquid nitrogen for diffraction data collection after a quick
soaking in a buffer of 0.1 M HEPES pH7.4, 0.2 M lithium
sulfate, 1.6 M ammonium sulfate and 15% isopropanol.
Data were collected in Shanghai Synchrotron Radiation
Facility (SSRF) and processed by HKL2000 . The
solution was found by molecular replacement using Phaser
. The model was rebuilt in Coot . The final
structure was refined and the model statistics (Table 1)
were calculated using Phenix . The buried surface
area was calculated by PISA program . All graphics
for the various structures were produced using Pymol
(DeLano Scientific LLC).
Maltose binding protein (MBP) pulldown
MBP-PCAF HAT domain (amino acids 493-658) fusion
protein and MBP alone were expressed in pMBP-c, a
modified vector from pMAL-c2x (NEB) with convenient
restriction and thrombin cleavage sites. The fusion
protein and MBP were purified using amylose agarose (New
England Biolab). An extra C terminus of the MBP alone
(~30 amino acids) using this empty pMBP-c vector was
sensitive to protease cleavage and showed two bands in
our SDS-PAGE. All bound maltose was efficiently
removed by running these proteins on a gel filtration
Superdex 200 (GE Healthcare) using a high salt buffer of
30 mM Tris pH8.0, 2 M NaCl, 5 mM -ME and 3 mM
Table 1 Data collection and refinement statistics
RMSD bond length ()
RMSD bond angles ()
Overall B-factor (2)
***Number of total atoms
The data for the highest resolution shell are shown in parenthesis.
*Rsym = S|I- < I > |/SI, where I is the observed intensity, <I > is the statistically
weighted average intensity of multiple observations of symmetry-related
reflections. R = S||Fo||Fc||/S|Fo|, where Fo and Fc are observed and calculated
structure factor amplitudes, respectively. **Rfree is calculated for a randomly
chosen 10% of reflections. RMSDroot mean square deviation. I/(I)ratio of
mean intensity to a mean standard deviation of intensity. *** Number of protein
atoms and nucleic acid atomsthe ordered region.
EDTA. The MBP or MBP-PCAF proteins (60 g) were
then rebound with 10 l new amylose agarose beads and
60 g His-tagged PCAF (amino acids 493-658 and
493832) in a binding buffer of 50 mM Tris pH8.0, 150 mM
sodium chloride, 5 mM -ME and 3 mM EDTA for 2 h
at 4C. After 3 washes using the binding buffer added
with 0.03% Triton X-100, the agarose beads were
collected and the bound proteins were subjected to a
regular 15% SDS-PAGE. The gel was stained with coomassie
Plasmid pCI-Flag PCAF of Flag-tagged full-length
human PCAF was purchased from Addgene . The
construct expressing HA-PCAF was built on pLV-nHA
vector using EcoRI and XbaI sites and confirmed by
DNA sequencing. Cells 293 T were cultured in DMEM
and 10% fetal bovine serum (Gibco) with 100 g/ml
penicillin and streptomycin. Co-transfection of both
tagged PCAF expression plasmids (8 g each) was
carried out using standard calcium phosphate method. Cells
were harvested after 48 h and lysed in a lysis buffer
(50 mM TrisHCl pH7.4, 150 mM NaCl, 1 mM PMSF,
10 g/ml leupeptin, 2 g/ml pepstatin, 1 mM EDTA and
1% Triton X-100). The lysed supernatants were
incubated with 20 l anti-Flag M2 affinity gel (Sigma) for
4 h. The Flag gel was then harvested and washed using
the same lysis buffer and directly mixed with SDS
sample buffer. The mixture was resolved by 10% SDS-PAGE.
The gel was blotted to nitrocellulose membrane and
further detected using anti-Flag or anti-HA antibodies
Proteins (wt or mutated HAT domains) were exchanged
to a non-amine buffer (50 mM HEPES pH7.5, 50 mM
sodium sulfate) by dialysis. Protein cross-linking was
carried out using 50 mM disuccinimidyl suberate (DSS)
dissolved in pure dimethyl sulfoxide (DMSO). DSS at 5X
molar concentration was directly added to 10 l of
1 mg/ml PCAF proteins. The reactions were incubated
for 1 h at room temperature and quenched by adding
0.3 M Tris pH7.5 to reach a final concentration of
30 mM. The reactions were then analyzed using regular
12% SDS-PAGE and stained with coomassie blue.
Multiangle laser light scattering (MALS)
The protein prep of the PCAF HAT domain at 2-20 mg/
ml was first resolved on a size exclusion column
(WTC010S5, 5 m silica beads, 7.8 300 mm) in a buffer of
50 mM phosphate buffer at pH 7.0 and 150 mM NaCl at
35C. The HPLC was run on a LabAlliance series 1500
isocratic system at a flow rate of 0.5 ml/min. Data were
then collected on a DAWN HELEOS II laser photometer
at an emission of 658 nm (Wyatt, USA). Molecular mass
was calculated using ASTRA V (Wyatt, USA).
Additional file 1: Figure S1. Two dimeric PCAF HAT structures in one
asymmetric unit. Figure S2. Structural comparison of PCAF HAT domain
with its homologues. Figure S3. Electron density map of two dimeric
interfaces. Figure S4. The PCAF mutants are enzymatically active as wt
PCAF. Figure S5. Human GCN5 HAT domain crystallized in a dimeric
state. Figure S6. Dimeric structure of tetrahymena GCN5. Table S1. PISA
analyses of the dimeric interfaces of PCAF/GCN5 crystal structures.
Table S2. Crystal structures of PCAF/GCN5 homologues that have
currently been solved.
Ada: Alteration/deficiency in activation; ATAC: Ada Two-A containing
complex; BRD: Bromodomain; DMSO: Dimethyl sulfoxide; dRI: Differential
refractive index; DSS: Disuccinimidyl suberate; HAT: Histone
acetyltransferase; GCN5: General control nonderepressible 5;
IPTG: Isopropyl--D-thiogalactopyranoside; MBP: Maltose binding protein;
MALS: Multiangle light scattering; MW: Molecular weight; PCAF: p300/CBP
associated factor; rmsd: Root mean square deviation; wt: Wild-type.
SS, JL and JY crystallized this protein and did further biochemical
experiments. YC collected data and solved this structure. HH and XL did
co-IP experiment. AH, KJ and RC formulated the designs of these studies.
AH and JH managed these experiments. AH analyzed data and wrote the
manuscript. JSD revised this manuscript. All authors read and approved
the final manuscript.
We thank Drs. Jianhua He, Bo Sun and Feng You at BL17U of Shanghai
Synchrotron Radiation facility (SSRF) for data collection; Dr. Lin Chen
(University of Southern California, USC) for useful suggestions; Dr. Jin Lang
(Wyatt, Beijing) for MALS experiment and Dr. Tanwei Lin for initial diffraction
test of crystals. This work was supported by the National Science Foundation
of China (31170685, 90919036 and 30840027), Project 985 (0660ZK1022) and
Program 111 (B06016).
1. Strahl BD , Allis CD : The language of covalent histone modifications . Nature 2000 , 403 ( 6765 ): 41 - 45 .
2. Kouzarides T : Chromatin modifications and their function . Cell 2007 , 128 ( 4 ): 693 - 705 .
3. Yang XJ : The diverse superfamily of lysine acetyltransferases and their roles in leukemia and other diseases . Nucleic Acids Res 2004 , 32 ( 3 ): 959 - 976 .
4. Glozak MA , Sengupta N , Zhang X , Seto E : Acetylation and deacetylation of non-histone proteins . Gene 2005 , 363 : 15 - 23 .
5. Shahbazian MD , Grunstein M : Functions of site-specific histone acetylation and deacetylation . Annu Rev Biochem 2007 , 76 : 75 - 100 .
6. Xu W , Edmondson DG , Evrard YA , Wakamiya M , Behringer RR , Roth SY : Loss of Gcn5l2 leads to increased apoptosis and mesodermal defects during mouse development . Nat Genet 2000 , 26 ( 2 ): 229 - 232 .
7. Yamauchi T , Yamauchi J , Kuwata T , Tamura T , Yamashita T , Bae N , Westphal H , Ozato K , Nakatani Y : Distinct but overlapping roles of histone acetylase PCAF and of the closely related PCAF-B/GCN5 in mouse embryogenesis . Proc Natl Acad Sci U S A 2000 , 97 ( 21 ): 11303 - 11306 .
8. Maurice T , Duclot F , Meunier J , Naert G , Givalois L , Meffre J , Celerier A , Jacquet C , Copois V , Mechti N , et al: Altered memory capacities and response to stress in p300/CBP-associated factor (PCAF) histone acetylase knockout mice . Neuropsychopharmacology 2008 , 33 ( 7 ): 1584 - 1602 .
9. Duclot F , Jacquet C , Gongora C , Maurice T : Alteration of working memory but not in anxiety or stress response in p300/CBP associated factor (PCAF) histone acetylase knockout mice bred on a C57BL/6 background . Neurosci Lett 2010 , 475 ( 3 ): 179 - 183 .
10. Pons D , Trompet S , De Craen AJ , Thijssen PE , Quax PH , De Vries MR , Wierda RJ , van den Elsen PJ , Monraats PS , Ewing MM , et al: Genetic variation in PCAF, a key mediator in epigenetics, is associated with reduced vascular morbidity and mortality: evidence for a new concept from three independent prospective studies . Heart 2011 , 97 ( 2 ): 143 - 150 .
11. Nagy Z , Tora L : Distinct GCN5/PCAF-containing complexes function as co-activators and are involved in transcription factor and global histone acetylation . Oncogene 2007 , 26 ( 37 ): 5341 - 5357 .
12. Brownell JE , Zhou J , Ranalli T , Kobayashi R , Edmondson DG , Roth SY , Allis CD : Tetrahymena histone acetyltransferase A: a homolog to yeast Gcn5p linking histone acetylation to gene activation . Cell 1996 , 84 ( 6 ): 843 - 851 .
13. Yang XJ , Ogryzko VV , Nishikawa J , Howard BH , Nakatani Y : A p300/CBPassociated factor that competes with the adenoviral oncoprotein E1A . Nature 1996 , 382 ( 6589 ): 319 - 324 .
14. Marmorstein R : Structure of histone acetyltransferases . J Mol Biol 2001 , 311 ( 3 ): 433 - 444 .
15. Marmorstein R , Trievel RC : Histone modifying enzymes: structures , mechanisms, and specificities. Biochim Biophys Acta 2009 , 1789 ( 1 ): 58 - 68 .
16. Rojas JR , Trievel RC , Zhou J , Mo Y , Li X , Berger SL , Allis CD , Marmorstein R : Structure of Tetrahymena GCN5 bound to coenzyme A and a histone H3 peptide . Nature 1999 , 401 ( 6748 ): 93 - 98 .
17. Clements A , Rojas JR , Trievel RC , Wang L , Berger SL , Marmorstein R : Crystal structure of the histone acetyltransferase domain of the human PCAF transcriptional regulator bound to coenzyme A . EMBO J 1999 , 18 ( 13 ): 3521 - 3532 .
18. Poux AN , Marmorstein R : Molecular basis for Gcn5/PCAF histone acetyltransferase selectivity for histone and nonhistone substrates . Biochemistry 2003 , 42 ( 49 ): 14366 - 14374 .
19. Kuo MH , Brownell JE , Sobel RE , Ranalli TA , Cook RG , Edmondson DG , Roth SY , Allis CD : Transcription-linked acetylation by Gcn5p of histones H3 and H4 at specific lysines . Nature 1996 , 383 ( 6597 ): 269 - 272 .
20. Schiltz RL , Mizzen CA , Vassilev A , Cook RG , Allis CD , Nakatani Y : Overlapping but distinct patterns of histone acetylation by the human coactivators p300 and PCAF within nucleosomal substrates . J Biol Chem 1999 , 274 ( 3 ): 1189 - 1192 .
21. Sakaguchi K , Herrera JE , Saito S , Miki T , Bustin M , Vassilev A , Anderson CW , Appella E : DNA damage activates p53 through a phosphorylation-acetylation cascade . Genes Dev 1998 , 12 ( 18 ): 2831 - 2841 .
22. Liu L , Scolnick DM , Trievel RC , Zhang HB , Marmorstein R , Halazonetis TD , Berger SL : p53 sites acetylated in vitro by PCAF and p300 are acetylated in vivo in response to DNA damage . Mol Cell Biol 1999 , 19 ( 2 ): 1202 - 1209 .
23. Spedale G , Timmers HT , Pijnappel WW : ATAC-king the complexity of SAGA during evolution . Genes Dev 2012 , 26 ( 6 ): 527 - 541 .
24. Balasubramanian R , Pray-Grant MG , Selleck W , Grant PA , Tan S : Role of the Ada2 and Ada3 transcriptional coactivators in histone acetylation . J Biol Chem 2002 , 277 ( 10 ): 7989 - 7995 .
25. Gamper AM , Kim J , Roeder RG : The STAGA subunit ADA2b is an important regulator of human GCN5 catalysis . Mol Cell Biol 2009 , 29 ( 1 ): 266 - 280 .
26. Da G , Lenkart J , Zhao K , Shiekhattar R , Cairns BR , Marmorstein R : Structure and function of the SWIRM domain, a conserved protein module found in chromatin regulatory complexes . Proc Natl Acad Sci U S A 2006 , 103 ( 7 ): 2057 - 2062 .
27. Sterner DE , Wang X , Bloom MH , Simon GM , Berger SL : The SANT domain of Ada2 is required for normal acetylation of histones by the yeast SAGA complex . J Biol Chem 2002 , 277 ( 10 ): 8178 - 8186 .
28. Boyer LA , Langer MR , Crowley KA , Tan S , Denu JM , Peterson CL : Essential role for the SANT domain in the functioning of multiple chromatin remodeling enzymes . Mol Cell 2002 , 10 ( 4 ): 935 - 942 .
29. Boyer LA , Latek RR , Peterson CL : The SANT domain: a unique histone-tail-binding module? Nat Rev Mol Cell Biol 2004 , 5 ( 2 ): 158 - 163 .
30. Horiuchi J , Silverman N , Marcus GA , Guarente L : ADA3, a putative transcriptional adaptor, consists of two separable domains and interacts with ADA2 and GCN5 in a trimeric complex . Mol Cell Biol 1995 , 15 ( 3 ): 1203 - 1209 .
31. Schuetz A , Bernstein G , Dong A , Antoshenko T , Wu H , Loppnau P , Bochkarev A , Plotnikov AN : Crystal structure of a binary complex between human GCN5 histone acetyltransferase domain and acetyl coenzyme A. Proteins 2007 , 68 ( 1 ): 403 - 407 .
32. Sadoul K , Wang J , Diagouraga B , Khochbin S : The tale of protein lysine acetylation in the cytoplasm . J Biomed Biotechnol 2011 , 2011 : 970382 .
33. Dekker FJ , Haisma HJ : Histone acetyl transferases as emerging drug targets . Drug Discov Today 2009 , 14 ( 19 - 20 ): 942 - 948 .
34. Marmorstein R , Roth SY : Histone acetyltransferases: function , structure, and catalysis. Curr Opin Genet Dev 2001 , 11 ( 2 ): 155 - 161 .
35. Nakamura Y , Umehara T , Nakano K , Jang MK , Shirouzu M , Morita S , Uda-Tochio H , Hamana H , Terada T , Adachi N , et al: Crystal structure of the human BRD2 bromodomain: insights into dimerization and recognition of acetylated histone H4 . J Biol Chem 2007 , 282 ( 6 ): 4193 - 4201 .
36. Umehara T , Nakamura Y , Jang MK , Nakano K , Tanaka A , Ozato K , Padmanabhan B , Yokoyama S : Structural basis for acetylated histone H4 recognition by the human BRD2 bromodomain . J Biol Chem 2010 , 285 ( 10 ): 7610 - 7618 .
37. Sobel RE , Cook RG , Perry CA , Annunziato AT , Allis CD : Conservation of deposition-related acetylation sites in newly synthesized histones H3 and H4 . Proc Natl Acad Sci U S A 1995 , 92 ( 4 ): 1237 - 1241 .
38. Wu PY , Ruhlmann C , Winston F , Schultz P : Molecular architecture of the S cerevisiae SAGA complex . Mol Cell 2004 , 15 ( 2 ): 199 - 208 .
39. Luger K , Mader AW , Richmond RK , Sargent DF , Richmond TJ : Crystal structure of the nucleosome core particle at 2.8 A resolution . Nature 1997 , 389 ( 6648 ): 251 - 260 .
40. Mao Y , Lin J , Zhou A , Ji K , Downey JS , Chen R , Han A : Quikgene: a gene synthesis method integrated with ligation-free cloning . Anal Biochem 2011 , 415 ( 1 ): 21 - 26 .
41. Otwinowski Z , Minor W : Processing of X-ray diffraction data collected in oscillation mode . Methods Enzymol 1997 , 276 : 307 - 326 .
42. McCoy AJ , Grosse-Kunstleve RW , Adams PD , Winn MD , Storoni LC , Read RJ : Phaser crystallographic software . J Appl Crystallogr 2007 , 40 (Pt 4): 658 - 674 .
43. Emsley P , Cowtan K : Coot: model-building tools for molecular graphics . Acta Crystallogr D Biol Crystallogr 2004 , 60 (Pt 12 Pt 1): 2126 - 2132 .
44. Adams PD , Grosse-Kunstleve RW , Hung LW , Ioerger TR , McCoy AJ , Moriarty NW , Read RJ , Sacchettini JC , Sauter NK , Terwilliger TC : PHENIX: building new software for automated crystallographic structure determination . Acta Crystallogr D Biol Crystallogr 2002 , 58 (Pt 11): 1948 - 1954 .
45. Krissinel E , Henrick K : Inference of macromolecular assemblies from crystalline state . J Mol Biol 2007 , 372 ( 3 ): 774 - 797 .