Download Role of protein methylation in chromatin remodeling and

ã
Oncogene (2001) 20, 3014 ± 3020
2001 Nature Publishing Group All rights reserved 0950 ± 9232/01 $15.00
www.nature.com/onc
Role of protein methylation in chromatin remodeling and transcriptional
regulation
Michael R Stallcup*,1,2
1
Department of Pathology, University of Southern California, Los Angeles, California, CA 90089, USA; 2Department of
Biochemistry and Molecular Biology, University of Southern California, Los Angeles, California, CA 90089, USA
Recent ®ndings suggest that lysine and arginine-speci®c
methylation of histones may cooperate with other types of
post-translational histone modi®cation to regulate chromatin structure and gene transcription. Proteins that
methylate histones on arginine residues can collaborate
with other coactivators to enhance the activity of speci®c
transcriptional activators such as nuclear receptors. Lysine
methylation of histones is associated with transcriptionally
active nuclei, regulates other types of histone modi®cations, and is necessary for proper mitotic cell divisions. The
fact that some transcription factors and proteins involved
in RNA processing can also be methylated suggests that
protein methylation may also contribute in other ways to
regulation of transcription and post-transcriptional steps
in gene regulation. In future work, it will be important to
develop methods for evaluating the precise roles of protein
methylation in the regulation of native genes in physiological settings, e.g. by using chromatin immunoprecipitation assays, di€erentiating cell culture systems, and
genetically altered cells and animals. It will also be
important to isolate additional protein methyltransferases
by molecular cloning and to characterize new methyltransferase substrates, the regulation of methyltransferase
activities, and the roles of new methyltransferases and
substrates. Oncogene (2001) 20, 3014 ± 3020.
Keywords: protein methylation; coactivators; histones;
transcriptional regulation; chromatin
DNA methylation is well known to play roles in the
regulation of chromatin structure and regulation of
transcription (Bird and Wol€e, 1999). However, a
surprising convergence of recent work in a number of
laboratories indicates that protein methylation also
contributes to the complex web of mechanisms that
govern chromatin remodeling and gene transcription.
This review will focus on recent evidence that
methylation of histones and perhaps other proteins
on lysine and arginine residues cooperates with other
types of post-translational modi®cations, such as
histone acetylation and phosphorylation, in the regulation of transcription.
*Correspondence: MR Stallcup, Department of Pathology, HMR
301, University of Southern California, 2011 Zonal Avenue, Los
Angeles, California, CA 90089-9092, USA
Types of protein methylation and their possible roles in
various signaling pathways
Protein methylation involves transfer of a methyl
group from S-adenosylmethionine to acceptor groups
on substrate proteins. Proteins can be methylated on
lysine, arginine, histidine, or carboxyl residues (Aletta
et al., 1998). In addition, when some aspartate residues
in proteins spontaneously convert to isoaspartate as a
result of protein aging, a methylation mechanism is
used in cells to reverse this process (Najbauer et al.,
1996). Speci®c roles for carboxy-methylation of
proteins such as Ras and protein phosphatase 2A have
been identi®ed in various signaling pathways, and
recent evidence has implicated arginine-speci®c protein
methyltransferases in several signaling pathways,
although in most of these cases the speci®c roles of
protein methylation and the relevant protein substrates
have not been identi®ed (Aletta et al., 1998; Gary and
Clarke, 1998; Lin et al., 1996; Abramovitch et al.,
1997). Arginine methylation is important for nuclear
export of some hnRNP proteins (Shen et al., 1998;
Nichols et al., 2000; Gary and Clarke, 1998). The
major arginine methyltransferase in mammalian cells,
PRMT1 (Tang et al., 2000a), is required for very early
stages of mouse development, suggesting a fundamental role in development (Pawlak et al., 2000).
Evidence for dynamic histone methylation
The lysine and arginine-rich N-terminal tails of
histones are the sites of many types of post-translational modi®cations such as methylation, acetylation,
and phosphorylation (Strahl and Allis, 2000; van
Holde, 1989). In nucleosomes, these N-terminal histone
tails extend beyond the DNA double helix which is
wrapped around the nucleosome core formed by the Cterminal regions of eight (two of each type) histone
molecules. In their unmodi®ed form the N-terminal
histone tails are positively charged and interact with
the negatively charged DNA backbone or core histone
regions on the same or neighboring nucleosomes; this
interaction contributes to chromatin compaction
(Luger et al., 1997; Luger and Richmond, 1998).
Neutralization of the positive charge of the N-terminal
tails by acetylation of lysines and phosphorylation of
serines weakens the binding of the N-terminal tail with
Protein methylation in transcriptional regulation
MR Stallcup
the negative regions of nucleosomes and thus contributes to chromatin remodeling.
Lysine methylation of histones in vivo is well
documented: histone H3 is methylated on lysines 4, 9,
27 and 36, although the speci®c pattern of residues
methylated may vary among species; histone H4 is
methylated on lysine 20 (van Holde, 1989; Strahl and
Allis, 2000). In contrast to lysine methylation, in vivo
methylation of arginine residues in histones is less well
documented. Arginine methylation has been dicult to
detect in native mammalian histones (Gary and Clarke,
1998), but has been reported in Drosophila (Desrosiers
and Tanguay, 1988). The latter study found changes in
lysine and arginine methylation of histone H3 and in
the N-terminal methylation of histone H2B after heat
shock of cultured Drosophila cells, suggesting that
arginine methylation of histones not only exists but is a
dynamically regulated process. Subsequent studies in
mammalian and avian cells provided further evidence
for association of dynamic histone methylation with
active (i.e. acetylated) chromatin, although it was not
determined whether the methylation was on lysine or
arginine residues (Annunziato et al., 1995; Hendzel and
Davie, 1991). The e€ect of lysine or arginine methylation on chromatin structure is currently unknown.
Families of lysine and arginine-speci®c protein
methyltransferases
Arginine methylation occurs on either or both of the
two terminal guanidino nitrogen atoms, resulting in
three possible products: monomethylarginine; NG,NGdimethylarginine, in which both methyl groups are on
the same nitrogen (asymmetric dimethylarginine); and
NG,N'G-dimethylarginine, in which each nitrogen atom
receives one methyl group (symmetric dimethylarginine) (Aletta et al., 1998; Gary and Clarke, 1998).
cDNA clones for ®ve genetically distinct but related
mammalian arginine methyltransferases have been
isolated: PRMT1 (Lin et al., 1996), PRMT2/
HRMT1L1 (Scott et al., 1998), PRMT3 (Tang et al.,
1998), CARM1 (Chen et al., 1999a), and JBP1 (Pollack
et al., 1999). A clone for one yeast protein, Hmt1 or
RMT1, belonging to this family has also been identi®ed
(Henry and Silver, 1996; Gary et al., 1996). While the
overall length of these polypeptide chains varies from
348 ± 608 amino acids, they all share a highly conserved
central domain encoding the methyltransferase activity
(Zhang et al., 2000). The methylarginine products for
PRMT1, PRMT3, RMT1 and CARM1 are monomethylarginine and asymmetric dimethylarginine (Lin
et al., 1996; Tang et al., 1998; Gary et al., 1996; BT
Schurter et al., 2001, submitted); the type of methylarginine produced by the other two family members
has not been reported. Studies conducted in vitro
indicated that each of the ®ve mammalian enzymes
recognizes a di€erent set of protein substrates. PRMT1
prefers to methylate arginine residues in glycine rich
regions, which are found in many proteins that bind
RNA and are involved in various aspects of RNA
metabolism; PRMT1 substrates include ®brillarin,
nucleolin, several hnRNPs, and histone H4 (Najbauer
et al., 1993; Gary and Clarke, 1998; Chen et al., 1999a).
Yeast RMT1 also eciently methylates arginines in
glycine rich regions of proteins (Gary and Clarke,
1998). However, CARM1 has little or no activity with
the substrates preferred by PRMT1 and RMT1. To
date the best substrate reported for CARM1 is histone
H3 (Chen et al., 1999a), and the methylation is not in
glycine rich regions (BT Schurter et al., 2001,
submitted). JBP1 can methylate histones H2A and H4
and myelin basic protein (Pollack et al., 1999). Both
PRMT1 and PRMT3 methylated poly(A)-binding
protein II, but the two enzymes produced di€erent
patterns of methylated proteins in a cell extract (Smith
et al., 1999; Tang et al., 1998). Substrates for PRMT2
have not yet been reported.
Symmetric dimethylarginine has been found in
myelin basic protein (Gary and Clarke, 1998) and
human spliceosomal Sm proteins D1 and D3, which
are components of some of the small nuclear
ribonucleoprotein complexes (Brahms et al., 2000),
but the enzymes responsible for methylating these
proteins have not yet been isolated.
Lysines may accept one, two, or three methyl groups
on the terminal amine group of the lysine side chain.
Histone H3 lysine methyltransferase activity has been
observed in several proteins containing SET domains
(Rea et al., 2000; O'Carroll et al., 2000). The
mammalian SET domain protein SUV39H1 methylates
lysine 9 of histone H3, and the SET domain was
important for this activity. The ability to methylate
histones was observed in some but not all SET domain
proteins. It remains to be determined whether the
inactive SET domain proteins lack catalytic activity or
require undetermined non-histone proteins as substrates, and whether SET domains are common
features of lysine methyltransferases.
3015
Recent evidence implicating histone methylation in
transcriptional regulation
Arginine methylation
Recent studies on the mechanism of transcriptional
regulation by the nuclear hormone receptors led to the
identi®cation of many proteins that may serve as
transcriptional coactivators. These coactivators, often
working as multi-subunit complexes, bind only to the
hormone-activated form of the nuclear receptors and
enhance the ability of these hormone-activated transcription factors to activate transcription of target
genes (Xu et al., 1999; Westin et al., 2000; McKenna et
al., 1999; Freedman, 1999; Glass and Rosenfeld, 2000).
Some of the coactivators are involved in chromatin
remodeling, by ATP-dependent mechanisms (e.g. SWI/
SNF complex) or histone acetylation (e.g. CBP, p300,
p/CAF), while others apparently help to recruit or
activate RNA polymerase II and its associated basal
transcription factors (e.g. TRAP/DRIP complex).
Oncogene
Protein methylation in transcriptional regulation
MR Stallcup
3016
Oncogene
These coactivator complexes constitute signal transduction pathways that transmit the activating signal from
the enhancer element-bound nuclear receptors to the
transcription machinery, thus enhancing transcription
from the associated promoter.
One of the best characterized coactivator complexes
contains one or more 160 kDa subunits called p160
coactivators; three genetically distinct but related
proteins make up the p160 coactivator family (Torchia
et al., 1998; Xu et al., 1999). The p160 coactivators
have multiple signal input domains that bind directly
to hormone activated nuclear receptors and multiple
signal output domains which recruit additional coactivators (Ma et al., 1999). The p160 coactivators have an
intrinsic protein acetyltransferase activity and recruit
additional coactivators such as CBP/p300 and p/CAF,
which also possess protein acetyltransferase activities
(Chen et al., 1997; Spencer et al., 1997). CBP, p300,
and p/CAF can acetylate histones, some DNA-binding
transcriptional activator proteins, and some components of the basal transcription machinery (Ogryzko et
al., 1996; Gu and Roeder, 1997; Imhof et al., 1997;
Korzus et al., 1998; Chen et al., 1999b); p160 proteins
can acetylate histones weakly (Chen et al., 1997;
Spencer et al., 1997), but eciently acetylated protein
substrates for the p160 acetyltransferase activity have
not been reported. Histone acetylation plays a major
role in nucleosome remodeling (Wol€e and Guschin,
2000; Cheung et al., 2000b); the physiological role for
acetylation of most other protein components of the
basal transcription machinery is still under investigation.
The search for additional proteins that bind to the
p160 coactivators led to the discovery of CARM1, a
protein with an arginine-speci®c histone H3 methyltransferase activity which can cooperate with p160
coactivators to enhance the ability of nuclear receptors
to activate transcription of transiently transfected
reporter genes containing the appropriate nuclear
receptor-controlled enhancer elements (Chen et al.,
1999a). In vitro CARM1 methylated histone H3 at
arginine 2 (minor site) and arginines 17 and 26 (major
sites) in the basic N-terminal region, and one or more
of four clustered arginines near the C-terminus (BT
Schurter et al., 2001, submitted). CARM1 also
methylated histone H2A but not the other core
histones (Chen et al., 1999a). The N-terminal methylation sites on histone H3 are located among the sites for
lysine methylation, lysine acetylation, and serine
phosphorylation (Strahl and Allis, 2000). The coactivator and methyltransferase activities of CARM1 led
to the proposal that methylation of histones and/or
other proteins in the transcription complex cooperates
with acetylation of histones and other components of
the transcription complex to achieve chromatin
remodeling, recruitment of RNA polymerase II, and
initiation of transcription.
Subsequent transient transfection studies demonstrated a robust synergy between CARM1 and p300
in their functions as coactivators for nuclear receptors
(Chen et al., 2000a; Lee et al., 2001). The synergy was
only observed with very low levels of transfected
nuclear receptor expression vectors and was essentially
entirely dependent on the presence of the nuclear
receptors, the appropriate hormone, and a p160
coactivator. These studies reinforce the model that
p160 coactivators, histone acetyltransferases (e.g. p300
or CBP), and histone methyltransferases function
together to make multiple cooperative covalent histone
modi®cations that lead to chromatin remodeling.
In addition to CARM1, another member of the
arginine-speci®c protein methyltransferase family,
PRMT1, was also found to function as a coactivator
for nuclear receptors in collaboration with p160
coactivators. Furthermore, at low levels of transfected
nuclear receptor expression vectors CARM1 and
PRMT1 acted synergistically to enhance nuclear
receptor function (Koh et al., 2001). Their synergy
and activities as coactivators depended upon the cotransfection of a p160 expression vector, consistent
with the model that both CARM1 and PRMT1 are
recruited to the promoter through contact with the
p160 coactivator. Since CARM1 preferentially methylates histone H3 in vitro and PRMT1 methylates
histone H4, cooperative methylation of these two
histones may be responsible for the observed synergy.
Lysine methylation
Two recent studies have also implicated lysine
methylation of histones in regulation of transcription
and chromatin structure. A methyltransferase activity
which speci®cally modi®ed lysine 4 of histone H3 in
vitro was found in transcriptionally active macronuclei
of Tetrahymena but not in transcriptionally inactive
micronuclei (Strahl et al., 1999). In the same study
methylation of yeast histones was found preferentially
on H3 molecules that were also acetylated, thus
providing another indication that histone methylation
occurs on active chromatin. Results of a separate study
suggested a link between methylation of lysine 9 in
mammalian histone H3 and the regulation of chromatin structure (Rea et al., 2000). The Su(var) group of
genes in Drosophila and S. pombe is a functionally
diverse group of genes that were discovered in genetic
screens for suppressors of position e€ect variegation.
The Drosophila SU(VAR)3 ± 9 protein is associated
with heterochromatin and defects in this protein
disrupt the organization of normally heterochromatic
regions. Mammalian homologues of this protein
(SUV39H) were found to methylate lysine 9 of histone
H3. In vitro, methylation of lysine 9 and phosphorylation of serine 10 were mutually antagonistic. Disruption of the two Suv39h genes in mice resulted in
decreased viability and developmental abnormalities,
and ®broblast cultures from the double null mice had
increased levels of serine 10 phosphorylation and
aberrant mitotic cell divisions. The above studies
suggest that methylation of histones at various sites
may play multiple roles in modulating chromatin
structure and can be associated with both activation
and repression of gene expression. This diversity of
Protein methylation in transcriptional regulation
MR Stallcup
roles for methylation is reminiscent of the ®ndings that
histone phosphorylation is apparently involved in
regulation of both transcription and mitosis (Strahl
and Allis, 2000).
Proposed roles for methylation of histones and other
proteins in regulation of gene expression; and future
directions for investigation
Cooperative roles for multiple histone modifications in
regulating chromatin structure
The speci®c roles for histone methylation, both on
lysine and arginine residues, in regulation of chromatin
structure and transcription remain to be determined.
The accumulated evidence suggests some speci®c
models. Acetylation of lysine residues neutralizes the
positive charge of the basic N-terminal tails of
histones, and thus reduces the binding of the histone
tails to DNA or to acidic regions of other proteins.
However, since methylation on lysine and arginine
residues should not alter their charge, the e€ect of
histone methylation on chromatin structure is dicult
to predict. The extra bulk of the methyl group could
inhibit protein binding or provide a new epitope for
binding of a protein in the same way that acetylation
of lysine creates a binding site for proteins with
bromodomains (Jacobson et al., 2000).
As suggested previously (Strahl and Allis, 2000), the
location of speci®c lysine and arginine methylation sites
among sites for acetylation and phosphorylation on the
N-terminal tails of histones suggests potential cooperation among the di€erent types of post-translational
histone modi®cations in modulating chromatin structure; speci®c combinations of covalent modi®cations
may disrupt and/or promote interactions of the histone
tails with DNA and other proteins. The functional
synergy between p300 and CARM1 and between
PRMT1 and CARM1 as coactivators for nuclear
receptors also suggests a possible cooperation between
acetylation and methylation of various histones (Chen
et al., 2000a; Koh et al., 2001; Lee et al., 2001).
Methylation could also in¯uence the eciency with
which other types of covalent histone modi®cations
occur, as illustrated already by the negative e€ect that
methylation of lysine 9 of histone H3 has on
phosphorylation of serine 10 (Rea et al., 2000) and
the cooperative e€ect on transcriptional activation by
acetylation and phosphorylation of lysine 14 and serine
10, respectively, on histone H3 (Cheung et al., 2000a;
Lo et al., 2000). Thus far, the activities of most
methyltransferases have been examined only with free
histone substrates. Intact nucleosomes must also be
tested as substrates.
Role of protein methylation in the coactivator function of
protein methyltransferases
The mechanism for recruitment of lysine methyltransferases to the promoter of activated genes must await
the cloning of cDNAs encoding these methyltransferases. Since there is solid evidence for methyllysine in
histones and exciting recent ®ndings which suggest a
role for lysine-speci®c methylation in regulation of
chromatin structure and transcription, identi®cation of
the responsible enzymes must obviously be an important goal for future research. The activities of these
enzymes on histones will almost certainly be regulated
as part of signaling pathways that regulate chromatin
structure and transcription. Identi®cation of the
methyltransferases will provide an entry point for
studying the signaling pathways and how they ®t into
the increasingly complex picture of chromatin structure
and transcriptional regulation.
Possible mechanisms for preferential arginine methylation of active chromatin are suggested by the physical
and functional relationships among the various nuclear
receptor coactivators in transient transfection experiments. Methyltransferases CARM1 and PRMT1 bind
to a C-terminal activation domain (AD2) of p160
coactivators and can only function as nuclear receptor
coactivators when co-expressed with a p160 coactivator
with an intact AD2 region (Chen et al., 1999a, 2000a).
The function of p300 as a nuclear receptor coactivator
is similarly dependent on co-expression of a p160
coactivator with an intact AD1 region, which is the
binding site for p300 (Chen et al., 2000a; Li et al.,
2000). These results suggest that both methyltransferases and acetyltransferases are recruited to the
promoter through their physical interaction with
speci®c domains of p160 coactivators, which are
recruited to the promoter by direct interactions with
the DNA-bound nuclear receptors.
Thus far, the activities of CARM1 and PRMT1 as
coactivators for nuclear receptors have only been
demonstrated with transiently transfected reporter
genes. It will be important to test whether these
proteins are required for or can enhance nuclear
receptor function on target genes in more physiologically relevant settings. Transient transfection experiments in cells containing stably integrated reporter
genes o€er one such setting. Chromatin immunoprecipitation, using antibodies against methyltransferase
proteins, can be used to test whether these proteins
are recruited to the promoter of native target genes
when they are activated by nuclear receptors and their
hormones. Such assays have already been used to show
recruitment of other types of nuclear receptor
coactivators to target genes (Chen et al., 1999b; Shang
et al., 2000). Similarly, it will be important to test
directly whether methylated histones are associated
with speci®c genes in the active versus the quiescent
state. Strategies that were used to show de®nitively the
involvement of histone acetylation in transcriptional
regulation should be applicable to the study of histone
methylation. Antibodies that recognize acetylated but
not unmodi®ed histones have been used in chromatin
immunoprecipitation assays to demonstrate that histone acetylation occurs in conjunction with transcriptional activation of speci®c promoters (Chen et al.,
1999b; Cheung et al., 2000a). Since some speci®c in
3017
Oncogene
Protein methylation in transcriptional regulation
MR Stallcup
3018
Oncogene
vitro sites of lysine and arginine methylation on
histones have been determined already, similar strategies should be applied to the study of histone
methylation. In conjunction with these studies, it will
be important to use recently improved mass spectrometry technology to look for more concrete evidence to
support the existence of arginine methylation of
histones in vivo, and to test whether the sites of lysine
and arginine methylation in vivo correlate with those
observed with speci®c methyltransferases in vitro.
Various procedures to accomplish a functional
knock-out of the methyltransferase proteins or their
enzymatic activities should prove to be useful for
assessing their biological functions. Of course, the yeast
system is the most convenient for this type of
experiment, and such experiments have already shown
that Hmt1/RMT1 and its methyltransferase activity are
important in RNA metabolism and for the nuclearcytoplasmic shuttling of speci®c RNA binding proteins
(Shen et al., 1998; McBride et al., 2000). However, the
larger number of arginine-speci®c methyltransferase
proteins in mammalian cells than in yeast indicates that
there may be more complex roles for these proteins in
the higher eukaryotes. In mammalian cells antisense
technology and dominant negative mutants of CARM1
and PRMT1 (if they can be developed) should prove
helpful in testing whether these proteins are required
for transcriptional activation by nuclear receptors.
Studies in di€erentiating cell culture models, such as
one which demonstrated a role for p160 coactivator
GRIP1 in a developing myocyte system (Chen et al.,
2000b), may also prove useful. Genetic knockout and
over-expression in transgenic mice will also help to
determine the roles of these proteins in hormone action
and speci®c physiological processes; studies already
performed with the p160 coactivators can serve as a
model for such studies (Xu et al., 1998, 2000). A mouse
lacking functional PRMT1 has already been established, and the phenotype of the homozygous null
mouse is embryonic lethal (Pawlak et al., 2000); but the
e€ect of this mutation on nuclear receptor function and
chromatin structure has not been examined.
The fact that CARM1 and PRMT1 have argininespeci®c protein methyltransferase activities suggests
that protein methylation may be involved in their
coactivator activities. In an initial genetic test of this
hypothesis, substitution of alanine for three highly
conserved residues (VLD) in the region of CARM1
responsible for S-adenosylmethionine binding resulted
in loss of methyltransferase and coactivator activities,
thus supporting a role for protein methylation in the
coactivator function of CARM1 (Chen et al., 1999a).
However, a recent X-ray crystal structure of the
catalytic domain of PRMT3, which is closely related
in sequence to that of CARM1, indicates that the VLD
residues are probably involved in intramolecular
folding and thus may a€ect more than just Sadenosylmethionine binding (Zhang et al., 2000). Ideal
mutations for testing the role of methylation in the
coactivator function of CARM1 should selectively
eliminate binding of S-adenosylmethionine, protein
substrate, or p160 coactivator by CARM1 without
a€ecting overall three-dimensional structure. The X-ray
crystal structure for PRMT3 and another of yeast
Hmt1/RMT1 (Weiss et al., 2000) provide new insights
into the binding of S-adenosylmethionine and the
target arginine residue of the protein substrate and
thus should help to guide design of more suitable
mutations for testing the role of methyltransferase
activity in coactivator function.
Methylation of non-histone substrates
The possibility that the lysine and arginine-speci®c
methyltransferases contribute to gene activation by
methylation of protein substrates in addition to
histones must also be considered. The histone acetyltransferases p300, CBP, and p/CAF all have the ability
to acetylate speci®c non-histone proteins found in the
transcription machinery, such as other coactivators,
transcriptional activators, and basal transcription
factors (Ogryzko et al., 1996; Gu and Roeder, 1997;
Imhof et al., 1997; Korzus et al., 1998; Chen et al.,
1999b). By analogy, methyltransferases recruited to
active promoters may have non-histone targets as well.
In fact, PRMT1 and the yeast Hmt1/RMT1 can
methylate many RNA binding proteins which are known
to be involved in various aspects of RNA processing,
transport, translation, and metabolism (Najbauer et al.,
1993; Lin et al., 1996; Gary and Clarke, 1998; Aletta et
al., 1998; Shen et al., 1998; Smith et al., 1999; Tang et al.,
2000b; Nichols et al., 2000). The fact that transcriptional
initiation and many post-transcriptional steps are known
to be coupled and coordinated (Monsalve et al., 2000)
suggests that some proteins currently classi®ed as
transcriptional coactivators could actually function by
facilitating post-transcriptional steps. For example,
recruitment of PRMT1 to the promoter through its
contact with p160 coactivators could enhance reporter
gene expression by methylation of (e.g.) histone H4, an
RNA processing factor that is associated with the
transcription initiation complex, or both. In that regard,
it is interesting that another coactivator, PGC-1 contains
RNA binding motifs and has been shown to help
stimulate and couple transcription initiation and subsequent splicing of the transcript (Monsalve et al., 2000).
In addition the yeast Hmt1/RMT1 has been shown to
interact genetically with proteins involved in RNA
processing and transport (Henry and Silver, 1996; Shen
et al., 1998), and the methyltransferase activity of Hmt1/
RMT1 is important for this function in yeast (McBride et
al., 2000). A number of unidenti®ed proteins in cell
extracts have been shown to be substrates for methylation by speci®c or unspeci®ed methyltransferases (Lin et
al., 1996; Tang et al., 1998; Frankel and Clarke, 1999;
Pollack et al., 1999). Thus, it is likely that additional
methylation targets involved in regulation of transcription or processes coupled to transcription remain to be
identi®ed.
As new methyltransferase substrates are identi®ed, it
will be desirable to address the physiological roles of
their methylation. In order to do this, it will be
Protein methylation in transcriptional regulation
MR Stallcup
necessary to know speci®c functions for the substrate
proteins and to have speci®c assays to detect these
functions (e.g. cell free functional assays, transfection
assays in cultured cells, or gene replacement strategies
in living organisms). In that case, the importance of
methylation could be addressed by ®rst determining the
methylation site(s) and modifying them genetically (e.g.
change lysine to arginine or arginine to lysine). The
activities of the mutants which cannot be methylated
could then be compared with the wild type protein in
the available assay.
Reversibility of protein methylation
Most steps in signal transduction pathways are readily
reversible so that the biological response to the stimulus
can be shut o€ after removal of the stimulus. Thus, if
protein methylation is involved in signal transduction, it
seems necessary to assume that the methylation is
reversible or that the methylated protein can be
eliminated or replaced by an unmethylated protein.
However, little is known about this process. The ®ndings
that at least some methylation of histones is dynamic
rather than static implies that methylation must be
reversible. Methylated histones can thus be used as
substrates to search for enzymes that remove the methyl
groups. An activity that removes epsilon-methyl groups
from methyllysine of histones or from free methyllysine
has been reported (Paik and Kim, 1973, 1974), but there
is no evidence for an activity that can remove methyl
groups from arginine residues of proteins. For other
types of covalent modi®cation, such as acetylation and
phosphorylation, both the addition and the removal of
the modifying moiety are highly regulated. Therefore, it
is likely that the discovery of protein demethylases will
also lead to additional signaling pathways which
contribute to the complex and cooperative mechanisms
that regulate transcription and post-transcriptional
nuclear events in gene expression.
3019
Acknowledgments
I thank Dr DW Aswad (University of California, CA,
USA), Dr BD Strahl (University of Virginia, VA, USA),
and Dr H Ma (University of Southern California, CA,
USA) for critical comments on the manuscript. This work
was supported by US Public Health Service grant
DK55274 from the National Institutes of Health.
References
Abramovich C, Yakobson B, Chebath J and Revel M.
(1997). EMBO J., 16, 260 ± 266.
Aletta JM, Cimato TR and Ettinger MJ. (1998). Trends
Biochem. Sci., 23, 89 ± 91.
Annunziato AT, Eason MB and Perry CA. (1995). Biochem.,
34, 2916 ± 2924.
Bird AP and Wol€e AP. (1999). Cell, 99, 451 ± 454.
Brahms H, Raymackers J, Union A, de Keyser F, Meheus L
and LuÈhrmann R. (2000). J. Biol. Chem., 275, 17122 ±
17129.
Chen D, Huang S-M and Stallcup MR. (2000a). J. Biol.
Chem., 275, 40810 ± 40816.
Chen D, Ma H, Hong H, Koh SS, Huang S-M, Schurter BT,
Aswad DW and Stallcup MR. (1999a). Science, 284,
2174 ± 2177.
Chen H, Lin RJ, Schiltz RL, Chakravarti D, Nash A, Nagy
L, Privalsky ML, Nakatani Y and Evans RM. (1997). Cell,
90, 569 ± 580.
Chen H, Lin RJ, Xie W, Wilpitz D and Evans RM. (1999b).
Cell, 98, 675 ± 686.
Chen SL, Dowhan DH, Hosking BM and Muscat GEO.
(2000b). Genes Dev., 14, 1209 ± 1228.
Cheung P, Tanner KG, Cheung WL, Sassone-Corsi P, Denu
JM and Allis CD. (2000a). Mol. Cell, 5, 905 ± 915.
Cheung WL, Briggs SD and Allis CD. (2000b). Curr. Opin.
Cell Biol., 12, 326 ± 333.
Desrosiers R and Tanguay RM. (1988). J. Biol. Chem., 263,
4686 ± 4692.
Frankel A and Clarke S. (1999). Biochem. Biophys. Res.
Commun., 259, 391 ± 400.
Freedman LP. (1999). Cell, 97, 5 ± 8.
Gary JD and Clarke S. (1998). Prog. Nucleic Acids Res. Mol.
Biol., 61, 65 ± 131.
Gary JD, Lin W-J, Yang MC, Herschman HR and Clarke S.
(1996). J. Biol. Chem., 271, 12585 ± 12594.
Glass CK and Rosenfeld MG. (2000). Genes Dev., 14, 121 ±
141.
Gu W and Roeder RG. (1997). Cell, 90, 595 ± 606.
Hendzel MJ and Davie JR. (1991). Biochem. J., 273, 753 ±
758.
Henry MF and Silver PA. (1996). Mol. Cell Biol., 16, 3668 ±
3678.
Imhof A, Yang X-J, Ogryzko VV, Nakatani Y, Wol€e AP
and Ge H. (1997). Curr. Biol., 7, 689 ± 692.
Jacobson RH, Ladurner AG, King DS and Tjian R. (2000).
Science, 288, 1422 ± 1425.
Koh SS, Chen D, Lee Y-H and Stallcup MR. (2001). J. Biol.
Chem., 276, 1089 ± 1091.
Korzus E, Torchia J, Rose DW, Xu L, Kurokawa R,
McInerney EM, Mullen T-M, Glass CK and Rosenfeld
MG. (1998). Science, 279, 703 ± 707.
Lee Y-H, Koh SS and Stallcup MR. (2001). in preparation.
Li J, O'Malley BW and Wong J. (2000). Mol. Cell Biol., 20,
2031 ± 2042.
Lin W-J, Gary JD, Yang MC, Clarke S and Herschman HR.
(1996). J. Biol. Chem., 271, 15034 ± 15044.
Lo W-S, Trievel RC, Rojas JR, Duggan L, Hsu J-Y, Allis
CD, Marmorstein R and Berger SL. (2000). Mol. Cell, 5,
917 ± 926.
Luger K, MaÈder AW, Richmond RK, Sargent DF and
Richmond TJ. (1997). Nature, 389, 251 ± 260.
Luger K and Richmond TJ. (1998). Curr. Opin. Genet. Dev.,
8, 140 ± 146.
Ma H, Hong H, Huang S-M, Irvine RA, Webb P, Kushner
PJ, Coetzee GA and Stallcup MR. (1999). Mol. Cell. Biol.,
19, 6164 ± 6173.
McBride AE, Weiss VH, Kim HK, Hogle JM and Silver PA.
(2000). J. Biol. Chem., 275, 3128 ± 3136.
McKenna NJ, Xu J, Nawaz Z, Tsai SY, Tsai M-J and
O'Malley BW. (1999). J. Steroid Biochem. Mol. Biol., 69,
3 ± 12.
Monsalve M, Wu Z, Adelmant G, Puigserver P, Fan M and
Spiegelman BM. (2000). Mol. Cell, 6, 307 ± 316.
Oncogene
Protein methylation in transcriptional regulation
MR Stallcup
3020
Oncogene
Najbauer J, Johnson BA, Young AL and Aswad DW. (1993).
J. Biol. Chem., 268, 10501 ± 10509.
Najbauer J, Orpiszewski J and Aswad DW. (1996). Biochem.,
35, 5183 ± 5190.
Nichols RC, Wang XW, Tang J, Hamilton BJ, High FA,
Herschman HR and Rigby WF. (2000). Exp. Cell Res.,
256, 522 ± 532.
O'Carroll D, Scherthan H, Peters AH, Opravil S, Haynes
AR, Laible G, Rea S, Schmid M, Lebersorger A, Jerratsch
M, Sattler L, Mattei MG, Denny P, Brown SD, Schweizer
D and Jenuwein T. (2000). Mol. Cell Biol., 20, 9423 ± 9433.
Ogryzko VV, Schiltz RL, Russanova V, Howard BH and
Nakatani Y. (1996). Cell, 87, 953 ± 959.
Paik WK and Kim S. (1973). Biochem. Biophys. Res.
Commun., 51, 781 ± 788.
Paik WK and Kim S. (1974). Arch. Biochem. Biophys., 165,
369 ± 378.
Pawlak MR, Scherer CA, Chen J, Roshon MJ and Ruley HE.
(2000). Mol. Cell Biol., 20, 4859 ± 4869.
Pollack BP, Kotenko SV, He W, Izotova LS, Barnoski BL
and Pestka S. (1999). J. Biol. Chem., 274, 31531 ± 31542.
Rea S, Eisenhaber F, O'Carroll D, Strahl BD, Sun Z-W,
Schmid M, Opravil S, Mechtler K, Ponting CP, Allis CD
and Jenuwein T. (2000). Nature, 406, 593 ± 599.
Scott HS, Antonarakis SE, Lalioti MD, Rossier C, Silver PA
and Henry MF. (1998). Genomics, 48, 330 ± 340.
Shang Y, Hu X, DiRenzo J, Lazar MA and Brown M.
(2000). Cell, 103, 843 ± 852.
Shen EC, Henry MF, Weiss VH, Valentini SR, Silver PA and
Lee MS. (1998). Genes Dev., 12, 679 ± 691.
Smith JJ, RuÈcknagel KP, Schierhorn A, Tang J, Nemeth A,
Linder M, Herschman HR and Wahle E. (1999). J. Biol.
Chem., 274, 13229 ± 13234.
Spencer TE, Jenster G, Burcin MM, Allis CD, Zhou J,
Mizzen CA, McKenna NJ, OnÄate SA, Tsai SY, Tsai M-J
and O'Malley BW. (1997). Nature, 389, 194 ± 198.
Strahl BD and Allis CD. (2000). Nature, 403, 41 ± 45.
Strahl BD, Ohba R, Cook RG and Allis CD. (1999). Proc.
Natl. Acad. Sci. USA, 96, 14967 ± 14972.
Tang J, Frankel A, Cook RJ, Kim S, Paik WK, Williams
KR, Clarke S and Herschman HR. (2000a). J. Biol. Chem.,
275, 7723 ± 7730.
Tang J, Gary JD, Clarke S and Herschman HR. (1998). J.
Biol. Chem., 273, 16935 ± 16945.
Tang J, Kao PN and Herschman HR. (2000b). J. Biol.
Chem., 275, 19866 ± 19876.
Torchia J, Glass C and Rosenfeld MG. (1998). Curr. Opin.
Cell Biol., 10, 373 ± 383.
van Holde KE. (1989). Chromatin, Springer-Verlag: New
York, pp. 111 ± 148.
Weiss VH, McBride AE, Soriano MA, Filman DJ, Silver PA
and Hogle JM. (2000). Nat. Struct. Biol., 7, 1165 ± 1171.
Westin S, Rosenfeld MG and Glass CK. (2000). Adv.
Pharmacol., 47, 89 ± 112.
Wol€e AP and Guschin D. (2000). J. Struct. Biol., 129, 102 ±
122.
Xu J, Liao L, Ning G, Yoshida-Komiya H, Deng C and
O'Malley BW. (2000). Proc. Natl. Acad. Sci. USA, 97,
6379 ± 6384.
Xu J, Qiu Y, DeMayo FJ, Tsai SY, Tsai M-J and O'Malley
BW. (1998). Science, 279, 1922 ± 1925.
Xu L, Glass CK and Rosenfeld MG. (1999). Curr. Opin.
Genet. Dev., 9, 140 ± 147.
Zhang X, Zhou L and Cheng X. (2000). EMBO J., 19, 3509 ±
3519.