Latchman. Eukariotic Transciption Factors
Подождите немного. Документ загружается.
Treisman, R. (1996) Regulation of transcription by MAP kinase cascades. Current Opinion
in Cell Biology, 8, 205–215.
Tsukiyama, T., Becker, P.B. and Wu, C. (1994). ATP-dependant nucleosome disruption at
a heat-shock promoter mediated by DNA binding of GAGA transcription factor.
Nature 367, 525–532.
Vandromme, M., Gauthier-Rouviere, C., Lamb, N. and Fernadez, A. (1996) Regulation of
transcription factor localisation: fine tuning of gene expression. Trends in
Biochemical Sciences 21, 59–64.
Weatherman, R. V., Fletterick, R. J. and Scanlan, T. S. (1999) Nuclear-receptor ligands and
ligand-binding domains. Annual Reviews of Biochemistry 68, 559–581.
Wilmann, T. and Beato, M. (1986) Steroid-free glucocorticoid receptor binds specifically
to mouse mammary tumour DNA. Nature 324, 688–691.
Wurtz, J-M., Bourget, W., Renaud, J-P. et al. (1996) A canonical structure for the ligand
binding domain of nuclear receptors. Nature Structural Biology 3, 87–94.
Zhu, H. and Bunn, H. F. (2001) How do cells sense oxygen? Science 292, 449–451.
Zimarino, V. and Wu, C. (1987) Induction of sequence-specific binding of Drosophila
heat-shock activator protein without protein synthesis. Nature 327, 727–730.
Zou, J., Guo, Y., Guettouche, T. et al. (1998) Repression of heat shock transcription factor
HSF1 activation by HSP90 (HSP90 complex) that forms a stress-sensitive complex
with HSF1. Cell 94, 471–480.
REGULATION OF TRANSCRIPTION FACTOR ACTIVITY 291
CHAPTER 9
TRANSCRIPTION FACTORS AND
HUMAN DISEASE
9.1 DISEASES CAUSED BY TRANSCRIPTION FACTOR
MUTATIONS
In previ ous chapters we have discussed a number of examples of the involve-
ment of transcription factors in normal cellular regulatory processes, for
example constitutive, inducible, cell type-specific or developmentally regu-
lated transcription. It is not surprising that aspects of this complex process
can go wrong and that the resulting defects in transcription factors can result
in disease (for reviews see Engelkamp and van Heynin gen 1996; Latchman,
1996).
For example, mutations in several classes of transcription factor that con-
trol gene expression during development have been shown to result in human
developmental disorders. Thus, mutations in the gene encoding the POU
family transcription factor Pit-1 (see Chap ter 4, section 4.2.6) result in a failure
of pituitary gland development leading to congenital dwarfism, while muta-
tions in the genes encoding the Pax family transcription factors Pax3 and Pax6
(see Chapter 4, section 4.2.7) result in eye defects. Similarly, mutations in
genes enc oding homeobox proteins (see Chapter 4, section 4.2) result in a
variety of congenital abnormalities (for review see Boncinelli, 1997).
Interestingly, the mutations in Pit-1 and Pax6 discussed above are both
dominant with one single copy of the mutant gene being sufficient to produce
the disease, even in the presence of a functional copy. However, this domi-
nance arises for different reasons (for review see Latchman, 1996). In the case
of Pit-1, the mutant Pit-1 can bind to its DNA binding site but cannot activate
gene expression. It therefore not only fails to stimulate transcription of its
target genes but can also act as a dominant negative factor inhibiting gene
activation by preventing the wild type protein from binding to DNA
(Fig. 9.1a). This mechanism is similar to one mode of action of transcriptional
repressors which act by preventing an activator from binding to its DNA
target site (see Chapter 6, section 6.2.1). In contrast, the dominant nature
of the Pax6 mutation does not reflect any dominant negative action of the
mutant protein since such mutations often involve complete deletion of the
gene. Rather it reflects a phenomenon known as haploid insufficiency in
which the amount of protein produced by a single functional copy of the
gene is not enough to allow it to activate its target genes effectively (Fig. 9.1b) .
As well as resulting from mutations in the genes encoding DNA binding
factors, developmental disorders can also result from mutations in genes
encoding other types of transcription factors such as components of the
basal transcriptional complex, co-activators or factors which alter chromatin
structure. Thus, for example, mutation in the gene encoding the SNF2 factor
which is part of the SWI/SNF chromatin remodelling complex (see Chapter 1,
section 1.2.3) results in a lack of -globin gene expression and a variety of
other symptoms such as mental retardation. This indicates that this factor is
necessary for opening the chromatin structure of the -globin genes and a
number of other genes so preventing their transcription when it is absent
(Gibbons et al., 1995) (Fig. 9.2).
Similarly, mutations in specific subunits of the basal transcription factor
TFIIH (see Chapter 3, section 3.5) result in the skin disease xeroder ma pig-
mentosum. Interestingly, the mutant TFIIH proteins found in these patients
show defects in their ability to respond to the transcriptional activato r FBP
294 EUKARYOTIC TRANSCRIPTION FACTORS
Figure 9.1
Different mechanisms by
which mutations in genes
encoding transcription
factors can be dominant,
producing disease in the
presence of a functional
copy of the gene. (a) In
the case of Pit-1, the
mutation produces a
dominant negative form
of the factor (square)
which binds to the
appropriate binding site
and not only fails to
activate transcription but
also prevents binding and
activation by the
functional protein (circle).
(b) In the case of Pax-6,
one functional gene
cannot produce enough
functional protein (circle)
to activate its target
genes (dotted circle).
and the transcriptional repressor FIR (Fig. 9.3) (Liu et al., 2001). Hence, in
this case disease results from an inability of a component of the basal tran-
scriptional complex to respon d appropriately to activating or repressing
signals.
As well as affecting DNA binding factors, chromatin remo delling factors
and components of the basal transcriptional complex , mutation can also affect
co-activator molecules which transmit the signal between DNA binding factors
and the basal complex. Thus, mutation in the gene encoding the CBP factor,
which acts as a co-activator for a variety of other transcription factors (see
Chapter 5, sectio n 5.4.3), results in the sever e developmental disorder known
as Rubinstein-Taybi syndrome which is characterized by mental retardation
and physical abnormalities (for review see D’Arcangelo and Curran, 1995)
indicating that CBP is an important co-activator for developmentally regu-
lated as well as inducible gene expression. Interestingly, no individuals with
mutations inactivating both copies of the CBP gene have ever been identified
TRANSCRIPTION FACTORS AND HUMAN DISEASE 295
Figure 9.2
(a) Functional SWI/SNF
alters the chromatin
structure of its target
genes to a more open
configuration (solid line).
(b) If SWI/SNF is inactive
these genes remain in a
closed chromatin
structure (wavy line) and
are thus not transcribed.
Figure 9.3
In the human disease
xeroderma pigmentosum,
the mutant TFIIH
(square) has lost the
ability of the wild type
protein (circle) to
respond appropriately to
the activating factor FBP
and the inhibitory factor
FIR.
and it is likely that a lack of functional CBP is incompatible with life.
Individuals with Rubinstein-Taybi syndrome have a single functional CBP
gene and a single mutant gene indicating that the mutation is dominant. As
with Pax6, however, this dominance apparently reflects a haploid insufficiency
in which a single copy of the CBP gene cannot produce enough functional
protein. This is not surprising since, as discussed in Chapter 6 (section 6.5),
the amount of CBP in the cell is limited and different transcription factors
compete for it.
As well as being the cause of Rubinstein-Taybi syndrome, CBP is also
involved in the neurodegenerative disease, Huntington’s chorea. However,
in this case, the CBP protein is entirely normal and the disease is caused by
mutations in a protein known as Huntingtin. Although Huntingtin is not a
transcription factor, it can bind to CBP and sequester it into protein aggrega-
tions (Nucifora et al., 2001) . Since the amounts of CBP in the cell are limiting,
this prevents it binding to transcriptional activators and hence causes disease.
Hence, inactivation of CBP can occur by mutation or by its binding of
another protein and consequent inactivation (Fig. 9.4). Interestingly, it has
recently been shown that as well as targeting CB P, the mutant Huntingtin can
also disrupt the interaction between the DNA binding factor Sp1 and the
transcriptional co-activator TAF
II
130 (for review see Freiman and Tjian,
2002) indicating that it can target DNA binding factors as well as co-activators.
Hence, developmental disorders can arise from mutations in genes encod-
ing DNA binding activator proteins such as Pit-1 and Pax-6, components of
the basal transcriptional complex such as TFIIH, co-activators such as CBP or
components of chromatin modulating complexes such as SNF2 (Fig. 9.5).
As well as suc h developmental defects, mutations in the genes encoding the
nuclear receptor transcription factor family (see Chapter 4, section 4.4) can
produce a failure to respond to the hormone whi ch normally binds to the
receptor and regulates transcription. Such mutations have been reported, for
example in the receptors for glucocorticoid, thyroid hormone and vitamin D
296 EUKARYOTIC TRANSCRIPTION FACTORS
Figure 9.4
CBP can be inactivated
by mutation producing
Rubinstein-Taybi
syndrome or by binding to
mutant Huntingtin protein
in Huntington’s disease.
(for review see Latchman, 1996). Similarly, mutations in the gene encoding
the peroxisome proliferator-activated receptor (PPAR), another member
of the nuclear receptor family, have been identified in patients who show
resistance to insulin, diabetes and high blood pressure (for review see
Schwartz and Kahn, 1999; Kersten et al., 2000).
9.2 CANCER
Despite the existence of transcription factor mutations producing develop-
mental defects or non-responsiveness to hormone, a special place in
the human diseases which can involve alterations in transcription factors is
occupied by canc er. Thus, because this disease results from growth in an
inappropriate place or at an inappropriate time, it can be caused not
only by deficiencies in particular genes but also by the enhanced expression
or activation of specific cellul ar genes involved in growth regulatory processes
which are normally only expressed at low levels or very transiently.
Interestingly, many cancer-causing genes of this type, known as oncogenes
(for general reviews see Bourne and Varmus, 1992; Broach and Levine, 1997;
Hunter, 1997), were originally identified within cancer-causing retroviruses
which had picked them up from the cellular genome. Within the virus, the
oncogene has become activate d either by over-expression or by mutation and
is therefore responsible for the ability of the virus to transform cells to a
cancerous phenotype. In contrast, the homologous gene within the cellular
genome is clearly not always cancer-causing since all cells are not cancerous. It
TRANSCRIPTION FACTORS AND HUMAN DISEASE 297
Figure 9.5
Mutations can occur in
genes encoding (a) DNA
binding activators (A), (b)
components of the basal
transcriptional complex
(BTC), (c) co-activators
(CA) or (d) factors which
alter chromatin structure
(SWI/SNF).
can be activated, however, into a cancer-causing form either by over-expres-
sion or by mutation and hence thes e genes can play an important role in the
generation of human cancer (Fig. 9.6). The form of the oncogene isolated
from the retrovirus and from the normal cellular genome is distinguished by
the prefixes v and c respectively, as in v-onc and c-onc.
Despite this potential to cause cancer, the c-on c genes are highly conserved
in evolution, being found not only in the species from which the original virus
was isolated but in a wide range of other eukaryotes. This indicates that the
products of these oncogenes play a critical role in the regulation of normal
cellular growth processes, their malregulation or mutation resulting theref ore
in abnormal growth and cancer. In agreement with this idea oncogenes iden-
tified in this way include genes encoding many different types of protein
involved in growth control such as growth factors, growth factor receptors
and G proteins. They also include, however, severa l genes encoding cellular
transcription factors which normally regulate specific sets of target genes.
Similarly, a number of other genes encoding transcription factor s have
been identified at the break points of the chromosomal translocations char-
acteristic of human leukaemias with their activation being involved in the
resulting cancer. Section 9.3 of this chapter therefore discusses severa l cases
of this type and the insights they have provided into the processes regulating
gene expression in normal cells and their malregulatio n in cancer (for reviews
see Rabbits, 1994; Latchman, 1996).
Following the discovery of cellular oncogenes, it subsequently became clear
that another class of genes existed whose protein products appeared to
restrain cellular growth. The deletion or mutational inactivation of these so-
298 EUKARYOTIC TRANSCRIPTION FACTORS
Figure 9.6
A cellular proto-oncogene
can be converted into a
cancer-causing oncogene
by increased expression
or by mutation.
called anti-oncogenes therefore results in the abnormal unregulated growth
characteristic of cancer cells (for reviews see Knudson, 1993; Weinberg, 1993).
As some of thes e anti-oncogenes also encode transcription factors, they are
discussed in section 9.4 of this chapter.
9.3 CELLULAR ONCOGENES AND CANCER
9.3.1 FOS, JUN AND AP1
The AP1 binding site is a DNA sequence that renders genes that contain it
inducible by tre atment with phorbol esters such as TPA. The activity binding
to this site is referred to as AP1 (activator protein 1). It is clear, however, that
preparations of AP1 purified by affinity chromatography on an AP1 binding
site contain several different proteins (for reviews see Kerppola and Curran,
1995; Karin et al., 1997; Shaulian and Karin, 2002).
A possible clue as to the identit y of one of these AP1 binding proteins was
provided by the finding that the yeast protein GCN4, which induces transcrip-
tion of several yeast genes involved in amino acid biosynthesis, does so by
binding to a site very similar to the AP1 site (Fig. 9.7). In turn, the DNA
binding region of GCN4 shows strong homology at the amino acid level to
v-jun, the oncogene of avian sarcoma virus ASV17 (Fig. 9.8). This suggested
therefore that the protein encoded by the cellular homologue of this gene, c-
jun, which was known to be a nuclearly located DNA binding protein, might
be one of the proteins which bind to the AP1 site.
TRANSCRIPTION FACTORS AND HUMAN DISEASE 299
Figure 9.8
Comparison of the
carboxyl-terminal amino-
acid sequences of the
chicken Jun protein and
the yeast transcription
factor GCN4. Boxes
indicate identical residues.
Figure 9.7
Relationship of the DNA-
binding sites for the yeast
transcription factor GCN4
and the mammalian
transcription factor AP1.
In agreement with this, antibodies against the Jun protein react with pur-
ified AP1-binding proteins, while Jun protein expressed in bacteria can bind
to AP1 binding sites. Hence the Jun protein is capable of binding to the AP1
binding site and constitutes one component of purified AP1 preparations
which also contain other Jun-related proteins such as Jun B (Fig. 9.9).
Moreover, co-transfection of a vector expressing the Jun protein with a target
promoter resulted in increased transcription if the target gene contained AP1
binding sites but not if it lacked them, indicating that Jun was capable of
stimulating transc ription via the AP1 site (Fig. 9.10). Hence the Jun oncogene
product is a sequence specific transcription factor capable of stimulating
transcription of genes containing its binding site.
In addition to Jun and Jun-related proteins, purified AP1 preparations also
contain the product of another oncogene c-fos, as well as several Fos-related
300 EUKARYOTIC TRANSCRIPTION FACTORS
Figure 9.9
Passage of total cellular
proteins through a
column containing an
AP1 site results in the
purification of several
cellular proteins including
Jun, Jun B, Fos and
Fos-related antigens
(Fras) which are capable
of binding to the AP1 site
either alone or in
combination.
Figure 9.10
Artificial expression of the
Jun protein in an
expression vector results
in the activation of a
target promoter containing
several AP1 binding sites
but has no effect on a
similar promoter lacking
these sites, indicating that
Jun can specifically
activate gene expression
via AP1 binding sites.
proteins known as the Fras (Fos-related antigens; see Fig. 9.9). Unlike Jun,
however, Fos cannot bind to the AP1 site alone but can do so only in the
presence of another protein p39, which is identical to Jun (see Chapter 4,
section 4.5). Hence, in addition to its ability to bind to AP1 sites alone, Jun can
also mediate binding to this site by the Fos protein. Such DNA binding by Fos
and Jun is dependent on the formation of a dimeric molecule. Although Jun
can form a DNA binding homodimer, Fos cannot do so. Hence DNA binding
by Fos is dependent upon the formation of a heterodimer between Fos and
Jun which binds to the AP1 site with approximately thirtyfold greater affinity
than the Jun homodimer (Fig. 9.11).
It is clear therefore that both Fos and Jun which were originally isolated in
oncogenic retroviruses are also cellular transc ription factors which play an
important role in inducing specific cellular genes following phorbol ester
treatment (for review see Ransone and Verma, 1990). Increased levels of
Fos and Jun occur in cells following treatment with phorbol esters indicating
that these substances act, at least in part, by increasing the levels of Fos and
Jun which, in turn, bind to the AP1 sites in phorbol ester-responsive genes and
activate their expression.
Similar increases in the levels of Fos and Jun as well as Jun-B and the Fos-
related protein Fra-1 are also observed when quiescent cells are stimulated to
grow by treatment with growth factors or serum, indicating that these sub-
stances act, at least in part, by increasing the levels of Fos and Jun which in
turn will switch on genes whose products are necessary for growth itself
(Fig. 9.12). In agreement with this idea, cells derived from mice in which
the c-jun gene has been inactivated grow very slowly in culture and the mice
themselves die early in embryonic development (Johnson et al., 1993). Hence
Fos and Jun play a critical role in normal cells, as transcription factors indu-
cing phorbol ester or growth dependent genes.
Normally levels of Fos and Jun incr ease only transiently following growth
factor treatment resulting in a period of brief controlled growth. Clearly
TRANSCRIPTION FACTORS AND HUMAN DISEASE 301
Figure 9.11
Heterodimer formation
between Fos and Jun
results in a complex
capable of binding to an
AP1 site with
approximately thirtyfold
greater affinity than a Jun
homodimer while a Fos
homodimer cannot bind to
the AP1 site.