Role of the Ada2 and Ada3 transcriptional coactivators in ... · PDF fileidentical to Gcn5 and Spt20 respectively (4). All are components of the SAGA complex. Functional interaction

Role of the Ada2 and Ada3 transcriptional coactivators

in histone acetylation

Ramakrishnan Balasubramanian1, 3, Marilyn G. Pray-Grant2, 3, William Selleck1,

Patrick A. Grant2, and Song Tan1

1Center for Gene Regulation, Department of Biochemistry & Molecular Biology, The Pennsylvania State University, University Park, PA 16802-1014

2Dept. of Biochemistry & Molecular Genetics, University of Virginia Health Sciences Center, Charlottesville, VA 22908-0733

3These authors contributed equally to this work.

Running Title: Functions of Ada2 and Ada3 in histone acetylation

Corresponding Author: Song TanCenter for Gene RegulationDepartment of Biochemistry & Molecular Biology108 Althouse LaboratoryPenn State UniversityUniversity Park, PA 16802-1014Tel: 814-865-3355, FAX: 814-863-7024email: [email protected]

Copyright 2001 by The American Society for Biochemistry and Molecular Biology, Inc.

JBC Papers in Press. Published on December 31, 2001 as Manuscript M110849200 by guest on M

ay 26, 2018http://w

ww

.jbc.org/D

ownloaded from

http://www.jbc.org/

Summary

Previous studies have shown that the transcriptional coactivator protein Gcn5/Ada4

functions as a catalytic histone acetyltransferase (HAT). In this work, we examine the

roles of the Ada2 and Ada3 coactivator proteins which are functionally linked to Gcn5.

We show that yeast Ada2, Ada3 and Gcn5 form a catalytic core of the ADA and SAGA

HAT complexes which is necessary and sufficient in vitro for nucleosomal HAT activity

and lysine specificity of the intact HAT complexes. We also demonstrate that Ada3 is

necessary for Gcn5-dependent nucleosomal HAT activity in yeast extracts. Our results

suggest that Ada2 potentiates Gcn5’s catalytic activity and that Ada3 facilitates

nucleosomal acetylation and an expanded lysine specificity.

Functions of Ada2 and Ada3 in histone acetylation

2

by guest on May 26, 2018

http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/

Introduction

In eukaryotes, DNA is packaged into chromatin by extensive interactions with histone

proteins (1). This repressive nature of chromatin can be relieved in the cell by chromatin

remodeling and chromatin modifying activities which produce an altered chromatin

structure accessible to and recognized by other components of the RNA polymerase II

transcriptional machinery (2). Acetylation is a post-translational modification of special

interest since the transcriptional activity of a gene strongly correlates with

hyperacetylation of the N-terminal tails of histone proteins present in the chromatin

template (3). The molecular basis for this observation is suggested by findings that the

transcriptional coactivator protein Gcn5 functions as a histone acetyltransferase (HAT)

[footnote 1] and that Gcn5 HAT activity correlates well with cellular Gcn5-mediated

transcriptional activation (4). In the yeast cell, Gcn5 is assembled into the 1.8

megadalton SAGA complex together with at least 12 other proteins from the ADA, SPT

and TAFs (TBP-associated factors) families of transcriptional coactivators as well as the

Tra1 protein (4). SAGA not only acetylates the tails of histone proteins assembled into

nucleosomes, but is also recruited to promoters by direct interactions with acidic

activators and the general transcription factor, TBP (4). These findings suggest that

multisubunit HAT complexes such as SAGA might coordinate their HAT and coactivator

functions to activate transcription in the cell.

The ADA genes were isolated in genetic screens for transcriptional adapters that

interacted functionally with the VP16 acidic activation domain (5,6). These experiments

identified the genes for Ada1, Ada2, Ada3, Ada4 and Ada5, the last two of which are


3


http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/

identical to Gcn5 and Spt20 respectively (4). All are components of the SAGA complex.

Functional interaction of Ada2, Ada3 and Gcn5 was suggested by additional genetic

studies which established that mutations in Ada2, Ada3 or Gcn5 result in similar

phenotypes, while deletions of two of the three genes did not cause a more severe

phenotype than a single deletion (7,8). Two-hybrid and cotranslation experiments

indicate that Ada2 interacts directly with Ada3 and Gcn5, while cotranslation

experiments demonstrate the Ada2, Ada3 and Gcn5 proteins form a ternary complex

which might represent a SAGA subcomplex (9-11). It has been suggested that Ada2

contributes to SAGA’s coactivator function by binding directly to VP16 and TBP

(12,13).

Gcn5 is the catalytic subunit responsible for SAGA’s HAT activity since Gcn5 on its own

will acetylate histone tails (14,15), and since mutations in the Gcn5 HAT domain affect

SAGA’s HAT activity in vitro as well as promoter-directed histone acetylation and

transcriptional activation in vivo (16-18). However, while SAGA acetylates both naked

and nucleosomal histones, Gcn5 appears to acetylate only naked histones efficiently (15).

Since Gcn5 interacts with other proteins in the multisubunit SAGA complex, it is likely

that other SAGA components interact with Gcn5 to modulate its HAT activity. Ada2 and

Ada3 are candidates for such SAGA components for three reasons. Firstly, as discussed

previously, Ada2, Ada3 and Gcn5 can form a ternary complex. Secondly, SAGA and a

second yeast HAT complex called ADA share these three common subunits as well as the

ability to acetylate nucleosomal histones (15). Thirdly, the yeast HAT A2 complex,

which has been shown to contain Ada2 and Gcn5 and is suspected to contain Ada3, was


4


http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/

recently shown to acetylate nucleosomal histones in vitro (19).

To analyze how SAGA components modulate the catalytic activity of Gcn5, we have

reconstituted the yeast Ada2 and Ada3 transcriptional coactivator proteins with yeast

Gcn5 in vivo using a newly developed polycistronic E. coli coexpression vector (20).

Our results suggest that Ada2 with Ada3 are necessary and sufficient to form a HAT

subcomplex with Gcn5 capable of acetylating histone tails in a chromatin template.


5


http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/

Experimental Procedures

Coexpression and purification of recombinant proteins

The Ada2, Ada3 and Gcn5 genes were amplified from yeast genomic DNA, cloned into

cloning vectors and the entire coding regions verified by DNA sequencing. Each of the

three genes was subcloned into the transfer cloning vector, pET3aTr (20) to create

pET3aTr-Ada2, pET3aTr-Ada3 and pET3aTr-Gcn5. The Ada3 and Gcn5 genes were

also subcloned into the pRET3a-HisTrxN fusion protein expression vector, a derivative

of pRET3a (21), so that the Ada3 and Gcn5 coding regions were positioned in frame with

a 6xHis-thioredoxin-NIa recognition site fusion tag to create pRET3a–HisTrxNAda3

and pRET3a-HisTrxNGcn5 respectively. The first 51 nucleotides of the Gcn5 gene were

recoded during this procedure to optimize codon usage and to remove the BamHI site

near the 5’ end of the Gcn5 gene. The transfer vectors were used to subclone the Ada2,

Ada3 and Gcn5 translational cassettes into the pST37 polycistronic expression vector to

create pST37-HisTrxNGcn5-yAda2, pST37-HisTrxNAda3-yAda2 and pST37-Ada3-

Ada2-Gcn5. Expression of proteins and complexes were performed in

BL21(DE3)pLysS cells using 0.2 mM IPTG (22). The following expression conditions

were used after optimizing temperature and length of induction to maximize recovery of

soluble protein or complex: 4 hours at 37°C for Ada2, 6 hours at 28°C for HisTrxNAda3,

4 hours at 28°C for HisTrxNGcn5, 3 hours at 28°C for HisTrxGcn5/Ada2, 4 hours at

28°C for HisTrxNAda3/Ada2 and 12 hours at 18°C for Ada3/Ada2/Gcn5.

Gcn5 was prepared by purifying the fusion HisTrxNGcn5 from the soluble extract of


6


http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/

cells by cobalt metal affinity chromatography (Talon resin, Clontech), cleaving the fusion

tag specifically using TEV NIa protease (23), and recovering the flow through from a

second cobalt affinity chromatography to remove the fusion tag. A similar procedure was

used to purify the Ada2/Gcn5 complex except that an additional Superdex S200-HR size

exclusion chromatography step (Amersham-Pharmacia) was used after the second cobalt

affinity column to remove excess Gcn5. Ada2 was purified from the insoluble fraction of

the cell extract by Sephacryl S-300 size exclusion chromatography (Amersham-

Pharmacia) of Triton X-100 washed inclusion bodies, followed by POROS HQ anion-

exchange and POROS HS cation-exchange chromatography (Applied Biosystems) with

all three chromatography steps performed in 8 M urea. The fusion HisTrxNyAda3 was

purified from the soluble extract of cells by cobalt metal affinity chromatography, and Q-

Sepharose anion-exchange chromatography (Amersham-Pharmacia). After NIa

cleavage to remove the fusion tag, the yAda3 polypeptide was purified under denaturing

conditions by Source Q anion-exchange (Amersham-Pharmacia) in urea followed by

Vydac C4 reverse phase chromatography (using an acetonitrile gradient in 0.1% TFA).

The Ada2/Ada3/Gcn5 complex used in this work was expressed and purified as a

nonfusion complex, but similar results are obtained when the HisTrxN tag is used for

purification and removed by NIa cleavage. For nonfusion purification, the soluble extract

of cells expressing the ternary complex was precipitated with 1.5 M ammonium sulfate,

and the pellet fraction further purified by Q-Sepharose, Source Q and Source S

chromatography (Amersham-Pharmacia).

Isolation of ADA complexes


7


http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/

Strains used for the purification of ADA complexes were BY4742 (MATα his3∆1

leu2∆0 lys2∆0 ura3∆0) and an ada3∆ derivative (MATα his3∆1 leu2∆0 lys2∆0 ura3∆0 ada3

∆::KAN), from Research Genetics. ADA complex was partially purified from 4L of wild-

type or ada3∆ mutant strains, grown to mid log phase in YPD media as described (15).

Briefly, whole-cell extracts were prepared by glass bead disruption and bound to 5 ml of

Nickel-NTA agarose (Qiagen). The resin was then washed in a column with 20 mM

imidazole, followed by elution of the bound proteins with 300 mM imidazole. The Ni2+-

NTA -agarose resin eluate was directly loaded onto a Mono Q HR 16/10 column

(Amersham-Pharmacia). Bound proteins were eluted with a 500 ml linear gradient from

100 to 500 mM NaCl. Peak fractions containing the ADA complex were concentrated

down to 0.5 ml using a Centriprep-30 concentrator (Millipore). Samples were then

loaded on a Superose 6 HR 10/30 column (Amersham-Pharmacia) equilibrated with 350

mM NaCl. Peak ADA fractions were pooled and used in HAT assays and Western

blotting.

HAT assays

HAT assays were performed as described previously (15) using core histones or

oligonucleosomes isolated from chicken erythrocytes (24). Lysine acetylation specificity

was determined using synthetic peptides blocked at specific lysine residues by

incorporating acetylated lysines at these positions (25).


8


http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/

Results

Recombinant complexes containing Ada2, Ada3 and Gcn5 by coexpression

We initially pursued overexpression in E. coli and purification of the individual Ada2 (50

kDa), Ada3 (79 kDa) and Gcn5 (51 kDa) polypeptides. Ada3 and Gcn5 were expressed

as fusions to a cleavable N-terminal combination hexahistidine-thioredoxin tag, while

Ada2 was expressed as a nonfusion polypeptide. Although we managed to express and

purify Ada2 (Fig. 1a), Ada3 (Fig. 1b) and Gcn5 (Fig. 1c) to near homogeneity, this

procedure was not optimal for several reasons. Firstly, we could not identify conditions

that permitted soluble expression of Ada2, even though we examined a variety of

conditions including lower temperatures and lower inducer concentrations (data not

shown). Secondly, proteolysis during purification hampered efficient purification of the

poorly expressed Ada3 polypeptide, necessitating denaturing, reverse phase purification.

Even then we were only able to purify about 0.1 - 0.2 mg of Ada3 per liter of E. coli.

Thirdly, since two of the polypeptides required denaturing purification conditions,

reconstitution of a Ada2/Ada3/Gcn5 complex would require in vitro refolding from

denatured polypeptides and subsequent purification of the functional complex from any

incorrectly folded side products. In previous experiments with the yeast TFIIA complex,

we reconstituted the much smaller TFIIA polypeptides (6, 9 and 13 kDa) into functional

complex with 20-40% recovery from 50-100 milligrams of purified polypeptides

[footnote 2]. Since the Ada2, Ada3 and Gcn5 polypeptides are much larger and the Ada3

polypeptide is available in much smaller quantities, it appeared unlikely that we would

obtain sufficient quantities of the Ada2/Ada3/Gcn5 complex for biophysical studies using


9


http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/

similar in vitro reconstitution procedures.

To solve this problem of generating active complexes containing Ada2, Ada3 and Gcn5,

we employed a novel polycistronic expression system developed in our laboratory to

successfully coexpress Ada2, Ada3 and Gcn5 in various combinations in vivo. This

expression system enables rapid construction of polycistronic expression plasmids

capable of coexpressing up to 4 individual genes (20). In particular, we have

overexpressed the binary Ada2/Ada3 and Ada2/Gcn5 complexes, and the ternary

Ada2/Ada3/Gcn5 complex each from single expression plasmids. Our unsuccessful

attempts to overexpress and purify the yeast Ada3/Gcn5 complex suggest that Ada3 and

Gcn5 do not form a stable complex on their own (data not shown), consistent with

previously published results (10). In contrast, the Ada2/Gcn5 complex remains intact

over affinity, size-exclusion and anion-exchange chromatography (Fig 2a and data not

shown). Furthermore, we find that coexpressed Ada2, Ada3 and Gcn5 copurify in a

complex over both anion and cation ion-exchange chromatography (Fig. 2b). The

components of the Ada2/Ada3/Gcn5 complex also coelute by size-exclusion

chromatography with an elution time consistent with a 434 ± 64 kDa complex assuming

a globular shape (data not shown). If one copy of each protein is present in the complex,

the 50 kDa Ada2, 79 kDa Ada3 and 51 kDa Gcn5 proteins would form a 180 kDa

complex, suggesting that either the complex is significantly elongated or that a more

complicated stoichiometry exists. We are able to isolate milligram quantities of nearly

homogenous Ada2/Ada3/Gcn5 complex under nondenaturing conditions in this manner.


10


http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/

HAT activity of Ada2/Gcn5 and Ada2/Ada3/Gcn5 complexes

We first compared the HAT activity of our recombinant Gcn5 to the Ada2/Ada3/Gcn5

complex normalized by the amount of Gcn5 present in a HAT fluorogram assay. We find

that the Ada2/Ada3/Gcn5 complex acetylates histone H3 in naked histones significantly

better than Gcn5 alone (Fig. 3a, lanes 2 and 3). Furthermore, the Ada2/Ada3/Gcn5

complex possesses substantial nucleosomal HAT activity under conditions where an

equivalent amount of Gcn5 does not acetylate nucleosomes to a detectable level (Fig. 3a,

lanes 5 and 6). Thus, the presence of Ada2 and Ada3 increases Gcn5’s HAT activity on

core histones, and enables acetylation of nucleosomal histones.

To dissect the individual roles of Ada2 and Ada3 in histone acetylation, we next

compared the HAT activity of Gcn5 alone, the Ada2/Ada3 and Ada2/Gcn5 binary

complexes and the Ada2/Ada3/Gcn5 triple complex in a liquid HAT assay. The

Ada2/Ada3 complex did not exhibit any HAT activity (data not shown), consistent with

Gcn5 being the catalytic subunit of SAGA. We used an equivalent amount of Gcn5 in

each of the remaining assays, as judged by Coomassie blue staining of an SDS-PAGE

gel. Gcn5 displays moderate HAT activity towards naked histones, whereas the

Ada2/Ada3/Gcn5 complex possesses four to five times higher activity (Fig. 3b).

Significantly, the Ada2/Gcn5 binary complex exhibits essentially the same level of HAT

activity as the Ada2/Ada3/Gcn5 triple complex. This suggests that Ada2 is sufficient to

potentiate Gcn5’s HAT activity with core histones.

Identical amounts of the same samples were assayed using a nucleosomal substrate again


11


http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/

normalizing against the amount of Gcn5 except in the case of the Ada2/Ada3 complex.

We used nucleosomal arrays containing approximately 10 nucleosomes isolated from

chicken erythrocytes, the same source of the naked histones. Consistent with most

published results, we find that Gcn5 fails to acetylate nucleosomes to any significant

degree above background (Fig. 3c). We did not detect any HAT activity for the

Ada2/Ada3 complex with nucleosomal substrates (data not shown), and the Ada2/Gcn5

complex was found to possess only modest nucleosomal HAT activity (Fig. 3c).

However, the Ada2/Ada3/Gcn5 ternary complex acetylated nucleosomal substrates about

15x better than Ada2/Gcn5 and approximately 100x better than Gcn5 (Fig. 3c). Similar

results are obtained with the nucleosomal arrays used in Fig. 3, and with

mononucleosomes (data not shown). We have also compared the activities of these

recombinant reagents with SAGA complex isolated from yeast cells. We detect

comparable HAT activity for the SAGA complex and the Ada2/Ada3/Gcn5 complex

using both core and nucleosomal histone substrates (data not shown).

Our HAT assays using samples which contain the same amount of Gcn5 assume that

Gcn5 was equally active in each sample. However, it is formally possible that our failure

to detect nucleosomal acetylation by Gcn5 results from a lower specific activity of our

Gcn5 as compared to the Ada2/Ada3/Gcn5 complex. Therefore, we also compared the

nucleosomal HAT activity of Gcn5 and the Ada2/Ada3/Gcn5 ternary complex when

normalized by their respective core histone HAT activities. The fluorogram in Fig. 3a,

lanes 7-12 shows that the Ada2/Ada3/Gcn5 complex still possesses significantly more

nucleosomal HAT activity than the barely detectable acetylation produced by Gcn5 alone.


12


http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/

Thus the failure to observe nucleosomal HAT activity of Gcn5 is not due to a lower

specific activity of a Gcn5 preparation, but rather that the presence of Ada2 and Ada3

confers upon Gcn5 the ability to efficiently acetylate histones in a nucleosomal substrate.


13


http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/

Lysine specificity of Ada2/Gcn5 and Ada2/Ada3/Gcn5 complexes

In addition to their different abilities to acetylate nucleosome, Gcn5 protein and the

SAGA complex also display distinct specificities towards the lysine residues in the N-

terminal tail of histone H3. Gcn5 shows a marked preference for Lys14 of H3, with only

background acetylation at the other available lysine residues at position 9, 18 and 23

(25,26). In contrast, the SAGA complex will acetylate an expanded set of lysine residues,

preferring Lys18, 14, 9 and 23 in decreasing order of preference (25). The acetylation

specificity of the SAGA complex was determined using H3 N-terminal peptides

synthesized with acetylated lysines blocking all but single positions at Lys 9, 14, 18 and

23. The use of these blocked synthetic peptides were previously validated in experiments

which produced essentially the same results as antibodies which recognize site-specific

acetylated histone tails and microsequencing of radiolabeled histones (25).

We have used the same synthetic H3 peptides to examine the lysine specificity of the

binary Ada2/Gcn5 and the ternary Ada2/Ada3/Gcn5 complexes, and compared these

results to those from parallel experiments using Gcn5 and the SAGA complex. The

results of these experiments are shown in Fig. 4. Our results for Gcn5 and SAGA are

very similar to those reported previously. However, we find that that Ada2/Gcn5

complex has acquired the ability to acetylated Lys18, although it still maintains a 4-5

fold preference for Lys14. In contrast, the Ada2/Ada3/Gcn5 complex displays the same

trend observed for the entire SAGA complex of Lys18 > Lys14 > Lys9 > Lys23. In this

assay, we observe the overall acetylation for the Ada2/Ada3/Gcn5 complex to be greater


14


http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/

than detected for the SAGA complex.

Ada3 is required for nucleosomal HAT activity in fractionated yeast extracts

Our result that the presence of Ada2 enhances the core histone HAT activity of Gcn5 is

consistent with the finding that the core histone HAT activity of SAGA complex isolated

from yeast cells lacking Ada2 is greatly reduced (27). We sought similar confirmation

for the physiological role of Ada3 protein by isolating SAGA and ADA complexes from

yeast cells deleted for Ada3. We find that the SAGA complex is disrupted by deletion of

Ada3, indicating that Ada3 may play an important role in maintaining the structural

integrity of SAGA (data not shown). In contrast, an apparent ADA complex complex

containing Ada2 and Gcn5 was isolated from yeast cells lacking Ada3. This complex

elutes from MonoQ anion-exchange resin under similar conditions as the ADA complex,

and elutes from Superose size exclusion resin with a slightly smaller apparent molecular

weight than the ADA complex (data not shown). This apparent Ada3 deficient ADA

complex is reduced in H3-specific core histone activity and nearly devoid of H3-specific

nucleosomal HAT activity compared to the intact ADA complex when normalized by

Gcn5 content (Fig 5). These results suggest that Ada3 plays an important role in the

structural integrity of the SAGA complex and further that Ada3 is required for Gcn5-

mediated nucleosomal HAT activity in yeast extracts.


15


http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/

Discussion

Our expression experiments confirm the benefits of in vivo coexpression for producing

protein complexes that might be difficult to reconstitute in vitro. Several protein

complexes have been observed to contain individual polypeptides which co-fold into the

final structure. For example, both the human TFIIF dimerization domain and yeast

TFIIA factor are heteromeric complexes composed of individual subunits which wrap

around each other (28,29). The individual subunits would be highly unlikely to fold on

their own since hydrophobic residues in the interior of the heteromeric complex would be

exposed in the individual subunit. Coexpression allows the component polypeptides to

fold together in a cellular environment in the presence of protein folding machinery such

as chaperones. We have used a modular, polycistronic expression system to coexpress

various combinations of Ada2, Ada3 and Gcn5 to generate both binary and ternary

complexes of these polypeptides under soluble conditions. It is interesting to note that

the Ada2 polypeptide, which apparently could only be expressed on its own in inclusion

bodies, was mostly soluble when coexpressed with Ada3 and Gcn5, albeit at lower

overall expression levels.

We find that Ada2/Ada3/Gcn5 is the minimal complex which acetylates nucleosomal

histones and matches the H3 lysine specificity of the SAGA HAT complex. We propose

that the Ada2/Ada3/Gcn5 complex is a catalytic subcomplex of both the ADA and SAGA

complexes for the following reasons. Firstly, both ADA and SAGA contain Ada2, Ada3

and Gcn5 (15), and our results here confirm previously published reports based on two-

hybrid and coimmunoprecipitation studies that Ada2, Ada3 and Gcn5 form a ternary


16


http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/

complex (10,11). Secondly, both ADA and SAGA complexes acetylate H3 tails

preferentially, with only weak acetylation observed for H2B (15). The same preference is

detected for the Ada2/Ada3/Gcn5 complex (Fig. 4 and data not shown). Thirdly, the

Ada2/Ada3/Gcn5 complex acetylates nucleosomal histones as do both the ADA and

SAGA complexes (15). Finally, both ADA and SAGA complexes display similar H3

lysine specificities observed for the Ada2/Ada3/Gcn5 complex (25).

The roles of Ada2 and Ada3 in histone acetylation are also supported by the behavior of

HAT complexes isolated from yeast strains deleted for either Ada2 or Ada3. In contrast

to Ada2, deletion of Ada3 results in the disruption of the SAGA complex, which prevents

us from analyzing the HAT activity of an Ada3-deficient SAGA complex. We do detect

an apparent Ada3-deficient ADA complex which possesses less core histone and greatly

reduced nucleosomal HAT activity. The reduced core histone activity of this Ada3-less

ADA complex was somewhat unexpected since the complex still contains both Ada2 and

Gcn5, whereas recombinant Ada2/Gcn5 complex had similar core histone HAT activity

to the recombinant Ada2/Ada3/Gcn5 complex. However, it is difficult to make a direct

comparison between these results since the ADA complex contains additional factors

which may modulate the HAT activity of the Ada2/Ada3/Gcn5 subcomplex.

Our results suggest novel roles for the Ada2 and Ada3 transcriptional coactivators in gene

regulation. Firstly, we find that Ada2 alone elevates Gcn5’s HAT activity using naked

histones. The involvement of Ada2 in potentiating the HAT activity of Gcn5 was

suggested by studies in which complexes containing Gcn5 mutants with decreased HAT


17


http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/

activity were isolated from yeast cells with or without the Ada2 gene (30). These studies

showed that the rate of HAT activity decreased for complexes isolated from cells without

Ada2. Our results demonstrate that Ada2 is directly responsible for potentiating the HAT

activity of Gcn5. We speculate that Ada2 may alter the Gcn5 histone tail binding pocket

to increase interactions with the histone tail, possibly interacting with the histone tail

itself. The SANT domain common to Ada2 and the Swi3 and Rsc8 subunits of the

SWI/SNF and RSC chromatin remodelling complexes may mediate such interactions

since deletion of the SANT domain reduces binding and catalysis of an H3 peptide by the

Ada2/Gcn5 complex (31). Secondly, our results indicate that Ada3 is required for

nucleosomal HAT activity and lysine specificity, which suggests that Ada3 might be

involved in recognizing histone tails in a nucleosome context. It is interesting to note that

the two histone tails targeted by SAGA, H3 and H2B, are the only two histone tails that

emerge between the gyres of the DNA superhelix in the nucleosome core particle

structure (32). It is possible that Ada3 alone or in combination with Ada2 and/or Gcn5

might recognize histone tails emerging from between the DNA gyres. A different

explanation is that Ada3 might be required to access the histone tails if the tails make

additional interactions with the DNA (33), although such interactions were not observed

in the nucleosome core particle crystal structure (32).

These roles for Ada2 and Ada3 in organizing a complex with Gcn5 which affects Gcn5’s

catalytic activity are substantially different from suggestions, based on experiments using

the individual proteins, that both Ada2 and Ada3 may directly mediate interactions with

activators (10,13). Although the results in this report do not exclude such potential


18


http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/

interactions of Ada2 and Ada3 with acidic activators, Brown et al have recently shown

that the Tra1 component is the direct target of acidic activators in the intact SAGA

complex and furthermore, that the Ada2/Ada3/Gcn5 complex does not interact with

acidic activators or Gcn4 activation domain (27). Based on these results, it appears

unlikely that Ada2 or Ada3 in the context of the SAGA complex interacts directly with

acidic activators. The results of Brown et al and our studies of the Ada2/Ada3/Gcn5

subcomplex suggest the possibility that the VP16 toxicity genetic screen identified Ada2,

Ada3 and Gcn5 as transcriptional coactivators not because of direct interactions with

activator proteins, but instead because of the contributions of these proteins to SAGA’s

nucleosomal HAT activity (27).

It has been reported that recombinant Gcn5 can acetylate nucleosomal histones over a

narrow range of ionic conditions in the absence of other protein components such as

Ada2 or Ada3 (34). Although we have been unable to reproduce these findings, we

cannot exclude the possibility that Gcn5 can acetylate nucleosomal histone tails to some

degree in the cell. However we believe that the Ada2/Ada3/Gcn5 complex, and not Gcn5

alone, constitutes the catalytic substructure of SAGA and ADA in part because the SAGA

and ADA complexes and the Ada2/Ada3/Gcn5 complex share similar trends of H3 lysine

acetylation whereas Gcn5 alone displays very different H3 lysine acetylation preference

even under the optimized ionic conditions that can apparently support acetylation of

nucleosomal histones (34).

Sendra et al have reported that the yeast HAT A2 complex can acetylate nucleosomal


19


http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/

histones in vitro. The HAT A2 complex elutes from size exclusion chromatography

consistent with a size of 480 kDa assuming a globular complex (19). Western blotting

indicated that both Ada2 and Gcn5 are components of the A2 complex, while the

presence of Ada3 in the complex was inferred but not proven by the observation that the

HAT A2 complex could not be isolated from a yeast strain lacking the Ada3 gene. Our

studies with recombinant, reconstituted Ada2/Ada3/Gcn5 establish that this is the

minimal complex capable of nucleosomal HAT activity. Given the comparable

molecular weight estimated by size exclusion chromatography, it is possible that the HAT

A2 complex is highly similar to the Ada2/Ada3/Gcn5 ternary complex.

We have therefore established the minimal SAGA and ADA subcomplex needed to

acetylate nucleosomal histone tails. Our results indicate that Ada2 and Ada3 regulate the

substrate specificity of Gcn5 histone acetyltransferase and thus enable Gcn5 to act on its

physiological substrate, chromatin. This requirement of protein cofactors by a catalytic

subunit may be a more general trait of chromatin modification complexes. For example,

the yeast NuA4 HAT coactivator complex, acetylates a nucleosomal substrate but like

SAGA, contains a catalytic subunit (Esa1) that on its own acetylates core histones but not

nucleosomes (35,36). The yeast elongator complex involved in transcriptional elongation

possesses robust HAT activity in the six-subunit holo-version, while the three-subunit

core-version containing its catalytic subunit Elp3 has little or no activity (37) [footnote

3]. Investigations into catalytic subcomplexes should provide insight into the functions

and mechanisms of these and other chromatin modification complexes.


20


http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/

Acknowledgements

We are grateful to Christine Brown, Jerry Workman and all other members of the Penn

State gene regulation community for discussion, support and encouragement. Supported

by NIH grant DK-58646-01 to P.A.G. and GM-60489 to S.T.. P.A.G. is the recipient of

a Burroughs Wellcome career development award and S.T. is a Pew Scholar in the

Biomedical Sciences.


21


http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/

References

1. van Holde, K. E. (1989) Chromatin, Springer-Verlag, New York

2. Workman, J. L., and Kingston, R. E. (1998) Annu Rev Biochem 67, 545-579

3. Turner, B. M., and O’Neill, L. P. (1995) Semin Cell Biol 6, 229-236

4. Sterner, D. E., and Berger, S. L. (2000) Microbiol Mol Biol Rev 64, 435-459

5. Berger, S. L., Cress, W. D., Cress, A., Triezenberg, S. J., and Guarente, L. (1990)

Cell 61, 1199-1208

6. Berger, S. L., Pina, B., Silverman, N., Marcus, G. A., Agapite, J., Regier, J. L.,

Triezenberg, S. J., and Guarente, L. (1992) Cell 70, 251-265

7. Pina, B., Berger, S., Marcus, G. A., Silverman, N., Agapite, J., and Guarente, L.

(1993) Mol Cell Biol 13, 5981-5989

8. Georgakopoulos, T., Gounalaki, N., and Thireos, G. (1995) Mol Gen Genet 246,

723-728

9. Marcus, G. A., Silverman, N., Berger, S. L., Horiuchi, J., and Guarente, L. (1994)

Embo J 13, 4807-4815

10. Horiuchi, J., Silverman, N., Marcus, G. A., and Guarente, L. (1995) Mol Cell Biol

15, 1203-1209

11. Candau, R., and Berger, S. L. (1996) J Biol Chem 271, 5237-5245

12. Silverman, N., Agapite, J., and Guarente, L. (1994) Proc Natl Acad Sci U S A 91,

11665-11668

13. Barlev, N. A., Candau, R., Wang, L., Darpino, P., Silverman, N., and Berger, S. L.

(1995) J Biol Chem 270, 19337-19344

14. Brownell, J. E., Zhou, J., Ranalli, T., Kobayashi, R., Edmondson, D. G., Roth, S.


22


http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/

Y., and Allis, C. D. (1996) Cell 84, 843-851

15. Grant, P. A., Duggan, L., Cote, J., Roberts, S. M., Brownell, J. E., Candau, R.,

Ohba, R., Owen-Hughes, T., Allis, C. D., Winston, F., Berger, S. L., and

Workman, J. L. (1997) Genes Dev 11, 1640-1650

16. Wang, L., Liu, L., and Berger, S. L. (1998) Genes Dev 12, 640-653

17. Kuo, M. H., Zhou, J., Jambeck, P., Churchill, M. E., and Allis, C. D. (1998) Genes

Dev 12, 627-639

18. Zhang, W., Bone, J. R., Edmondson, D. G., Turner, B. M., and Roth, S. Y. (1998)

Embo J 17, 3155-31 67

19. Sendra, R., Tse, C., and Hansen, J. C. (2000) J Biol Chem 275, 24928-24934

20. Tan, S. (2001) Protein Expr Purif 21, 224-234.

21. Tan, S., Hunziker, Y., Pellegrini, L., and Richmond, T. J. (2000) J Mol Biol 297,

947-959

22. Studier, F. W. (1991) J Mol Biol 219, 37-44

23. Parks, T. D., Leuther, K. K., Howard, E. D., Johnston, S. A., and Dougherty, W. G.

(1994) Anal Biochem 216, 413-417

24. Cote, J., Utley, R. T., and Workman, J. L. (1995) Methds in Mol Genet 6, 108-128

25. Grant, P. A., Eberharter, A., John, S., Cook, R. G., Turner, B. M., and Workman, J.

L. (1999) J Biol Chem 274, 5895-5900

26. Kuo, M. H., Brownell, J. E., Sobel, R. E., Ranalli, T. A., Cook, R. G., Edmondson,

D. G., Roth, S. Y., and Allis, C. D. (1996) Nature 383, 269-272

27. Brown, C. E., Howe, L., Sousa, K., Alley, S. C., Carrozza, M. J., Tan, S., and

Workman, J. L. (2001) Science 292, 2333-2337.


23


http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/

28. Gaiser, F., Tan, S., and Richmond, T. J. (2000) J Mol Biol 302, 1119-1127.

29. Tan, S., Hunziker, Y., Sargent, D. F., and Richmond, T. J. (1996) Nature 381, 127-

151.

30. Syntichaki, P., and Thireos, G. (1998) J Biol Chem 273, 24414-24419

31. Boyer, L. A., Langer, M. R., Tan, S., Denu, J. M., and Peterson, C. L. submitted

32. Luger, K., Mader, A. W., Richmond, R. K., Sargent, D. F., and Richmond, T. J.

(1997) Nature 389, 251-260

33. Luger, K., and Richmond, T. J. (1998) Curr Opin Genet Dev 8, 140-146

34. Tse, C., Georgieva, E. I., Ruiz-Garcia, A. B., Sendra, R., and Hansen, J. C. (1998)

J Biol Chem 273, 32388-32392

35. Allard, S., Utley, R. T., Savard, J., Clarke, A., Grant, P., Brandl, C. J., Pillus, L.,

Workman, J. L., and Cote, J. (1999) Embo J 18, 5108-5119.

36. Smith, E. R., Eisen, A., Gu, W., Sattah, M., Pannuti, A., Zhou, J., Cook, R. G.,

Lucchesi, J. C., and Allis, C. D. (1998) Proc Natl Acad Sci U S A 95, 3561-3565.

37. Winkler, G. S., Petrakis, T. G., Ethelberg, S., Tokunaga, M., Erdjument-Bromage,

H., Tempst, P., and Svejstrup, J. Q. (2001) J Biol Chem 276, 32743-32749.

38. Grant, P. A., Schieltz, D., Pray-Grant, M. G., Steger, D. J., Reese, J. C., Yates, J.

R., 3rd, and Workman, J. L. (1998) Cell 94, 45-53

39. Gill, S. C., and von Hippel, P. H. (1989) Anal. Biochem. 182, 319-326

Footnotes

1The abbreviations used are: HAT, histone acetyltransferase; TAF, TBP-associated


24


http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/

factor; TEV, tobacco etch virus; IPTG, isopropyl-β-D-thiogalactopyranoside.

2unpublished data, S. Tan and T.J. Richmond

3personal communication, J. Svejstrup


25


http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/

Figure Legends

Figure 1: Purification of recombinant yeast Ada2, Ada3 and Gcn5 polypeptides, as

depicted on a Coomassie blue stained SDS-PAGE gel. The positions of the Ada2, Ada3

and Gcn5 polypeptides and fusion proteins or tags are indicated by arrows to the left of

the gels, while the sizes of the molecular weight markers in kDa are shown to the right.

(a) Expression and purification of recombinant yeast Ada2. Uninduced and induced cells

are shown in lanes 1 and 2. The detergent washed inclusion bodies before or after size

exclusion chromatography in urea are shown in lanes 3 and 4 respectively, while the

pooled fractions after subsequent anion and cation exchange chromatography are show in

lanes 5 and 6 respectively. (b) Expression and purification of recombinant yeast Ada3.

Uninduced and induced cells are presented in lanes 1 and 2. The Talon cobalt affinity

and subsequent anion exchange purified fusion HisTrxnAda3 fractions are shown in lanes

3 and 4, followed by NIa digestion of the fusion protein in lane 5. Ada3 was purified

from this digest mix by anion exchange in urea and reverse phase chromatography in

lanes 6 and 7 respectively, while molecular weight markers are shown in lane 8. Both

native yeast and our recombinant Ada3 migrates anomalously at about 100 kDa (38)

although its expected molecular weight is 79.3 kDa, suggesting that the anomalous

mobility is due to inherent properties of the presumed unmodified polypeptide. (c)

Purification of recombinant yeast Gcn5. The fusion HisTrxNGcn5 was purified from the

crude extract in lane 1 by cobalt affinity chromatography (lane 2) before treatment with

NIa protease to remove the fusion tag in lane 3. Lane 4 shows the cation-exchanged

purified non-fusion Gcn5, and lane 5 shows molecular weight markers.


26


http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/

Figure 2: Purification of recombinant yeast Ada2/Gcn5 and Ada2/Ada3/Gcn5

complexes, as depicted on a Coomassie blue stained SDS-PAGE gel. The positions of

the Ada2, Ada3 and Gcn5 polypeptides and fusion proteins or tags are indicated by

arrows to the left of the gels, while the sizes of the molecular weight markers in kDa are

shown to the right. (a) Purification of Ada2/Gcn5 complex from the fusion

Ada2/HisTrxNGcn5. Uninduced and induced cells are shown in lanes 1 and 2, while the

extract pellet and supernatant are presented in lanes 3 and 4 respectively. The cobalt

affinity purified fractions before and after digestion with NIa protease are shown in lanes

5 and 6, while the pooled fraction following a second cobalt affinity column and size

exclusion chromatography is shown in lane 7. Molecular weight markers are shown in

lane 8. (b) Purification of the Ada2/Ada3/Gcn5 complex. Uninduced and induced cells

containing the polycistronic expression plasmid pST37-Ada3-Ada2-Gcn5 are shown in

lanes 1 and 2. The pooled fractions following Q-Sepharose, Source Q and Source S

chromatography are shown in lanes 3-5, respectively, and molecular weight markers are

shown in lane 6.

Figure 3: Histone acetyltransferase (HAT) activities of Gcn5 protein, Ada2/Gcn5 and

Ada2/Ada3/Gcn5 complexes (a) Fluorogram showing HAT activity of recombinant

Gcn5 and reconstituted Ada2/Ada3/Gcn5. Recombinant Gcn5 was used in lanes 2, 5, 8,

11 while the Ada2/Ada3/Gcn5 complex was used in lanes 3, 6, 9, 12. Control lanes using

buffer instead of sample are shown in lanes 1, 4, 7, 10. Samples on the left (lanes 2, 3, 5,

6) were normalized by the amount of Gcn5 on a Coomassie blue stained gel. Samples on

the right (8, 9, 11, 12) were normalized by core histone HAT activity. HAT assays were


27


http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/

performed as described previously (15). (b) HAT activity using core histones as substrate

and equivalent amounts of Gcn5. The HAT assays contained approximately 80 ng of

Gcn5, 160 ng of Ada2/Gcn5 and 300 ng of Ada2/Ada3/Gcn5, where Gcn5 amount was

measured by UV absorption using a calculated extinction coefficient of 0.89 (mg/ml)-

1cm-1 at 280 nm (39), and Ada2/Gcn5 and Ada2/Ada3/Gcn5 complex amounts were

estimated by comparing with the same Gcn5 sample on a Coomassie Blue stained SDS-

PAGE gel. (c) HAT activity using nucleosomal arrays as substrate and equivalent

amounts of Gcn5. The same amount of sample were used for both core histones and

nucleosome experiments. The charts shows the mean value of at least four

measurements.

Figure 4: Lysine acetylation specificity of Gcn5 protein and Ada2/Gcn5,

Ada2/Ada3/Gcn5 and SAGA complexes on H3 peptides (a) Schematic diagram of H3

amino-terminal peptide substrates used. Acetylated lysines residues in each peptide are

shown as open circles, while non-acetylated lysine residues are labelled. (b) Gcn5,

Ada2/Gcn5, Ada2/Ada3/Gcn5 and SAGA were incubated with the wild-type peptide

(WT), the peptide with no acetylation sites available (0) or with peptides with only single

acetylation sites available (9, 14, 18, 23). The histogram shows scintillation counts from

liquid HAT assays normalized to the activity obtained using the WT peptide as described

previously (25).

Figure 5: HAT activity of ADA complexes from wild-type or Ada3 deficient yeast


28


http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/

strains. Fluorogram shows H3-specific core histone and nucleosomal HAT activity of

Superose 6 fractionated ADA complex isolated from wild-type (WT) or Ada3 deficient

(ada3) strains. The same fractions used for HAT assays were analyzed on Western blots

using anti-Gcn5, anti-Ada2 and anti-Ada3 antibodies.


29


http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/


http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/


http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/


http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/


http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/


http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/

and Song TanRamakrishnan Balasubramanian, Marilyn G. Pray-Grant, William Selleck, Patrick A. Grant

Role of the Ada2 and Ada3 transcriptional coactivators in histone acetylation

published online December 31, 2001J. Biol. Chem.

10.1074/jbc.M110849200Access the most updated version of this article at doi:

Alerts:

When a correction for this article is posted•

When this article is cited•

to choose from all of JBC's e-mail alertsClick here


http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/lookup/doi/10.1074/jbc.M110849200

http://www.jbc.org/cgi/alerts?alertType=citedby&addAlert=cited_by&cited_by_criteria_resid=jbc;M110849200v1&saveAlert=no&return-type=article&return_url=http://www.jbc.org/content/early/2001/12/31/jbc.M110849200.citation

http://www.jbc.org/cgi/alerts?alertType=correction&addAlert=correction&correction_criteria_value=early/2001/12/31/jbc&saveAlert=no&return-type=article&return_url=http://www.jbc.org/content/early/2001/12/31/jbc.M110849200.citation

http://www.jbc.org/cgi/alerts/etoc

http://www.jbc.org/

Documents

Role of the Ada2 and Ada3 transcriptional coactivators in ... · PDF fileidentical to Gcn5 and Spt20 respectively (4). All are components of the SAGA complex. Functional interaction