Latchman. Eukariotic Transciption Factors

Подождите немного. Документ загружается.

system of the embryo. However, in moving from the 5’ to the 3’ end of the

cluster (i.e. from Hoxb-9 (2.5) to Hoxb-1 (2.9) in Fig. 7.11) each successive

gene is expressed earlier in development and also displays a more anterior

boundary of expression within the central nervous system (Figs 7.12 and 7.13).

Similar expression patterns have also been observed in Drosophila where each

successive gene in the Bithorax and Antennapedia clusters is expressed more

REGULATION OF TRANSCRIPTION FACTOR SYNTHESIS 221

Figure 7.12

Comparison of the

expression pattern of the

Hoxb-9 (2.5), b-8 (2.4)

and b-6 (2.2) genes in

the 12.5-day mouse

embryo. The top panel

shows in situ

hybridization with the

appropriate gene probe to

a section of the entire

embryo while the middle

row shows a high power

view of the region in

which the anterior limit of

gene expression occurs.

In these panels, which

show the sections in

bright field, hybridization

of the probe and

therefore gene expression

is indicated by the dark

areas. In the lower panel,

which shows the same

area in dark field,

hybridization is indicated

by the bright areas. Note

the progressively more

anterior boundary of

expression of Hoxb-6

(2.2) compared to Hoxb-8

(2.4) and to Hoxb-9 (2.5)

and compare with their

positions in the Hoxb

(Hox 2) complex in Figure

7.11.

anteriorly and affects progressively more anterior segments when it is

mutated. Indeed, studies in which regulatory elements from the invertebrate

Amphioxus were tested in mouse and chick embryos have indicated that the

elements regulating homeobox gene expression have been highly conserved

in evolution with the Amphioxus elements functioning in these very different

species (Manzanares et al., 2000) .

In the case of the mouse genes a possible molecular mechanis m for the

differential expression pattern across a cluster is provided by the finding that

genes in the 3’ half of the Hoxb cluster are activated in cultured cells by

treatment with low levels of retinoic acid, whereas genes in the 5’ half of

the cluster require much higher levels of retinoic acid for their activa tion

(for review see Conlon, 1995; Tabin, 1995). Considerable eviden ce exists

that retinoic acid can act as a morphogen in vertebrate development and it

has been suggested that a gradient of retinoic acid concentration may exist

across the developing embryo (reviewed by Conlon, 1995; Tabin, 1995).

Hence, the observed difference in expression of the Hoxb genes could be

controlled by a retinoic acid gradient (Fig. 7.14). In turn, the Hoxb genes like

their Drosophila counterparts would switch on other genes required in cells at

222 EUKARYOTIC TRANSCRIPTION FACTORS

Figure 7.13

Summary of the anterior

boundary of expression of

the genes in the Hoxb (2)

complex indicated on a

section of a 12.5-day

mouse embryo and

compared to the position

of the gene in the Hoxb

(2) cluster. Note the

progressively more

anterior boundary of

expression from the 5’ to

the 3’ end of the Hoxb

(2) cluster.

particular positions in the embryo accounting for the morphogenetic effects

of retinoic acid.

Retinoic acid functions by binding to and activating specific recept ors

which are members of the steroid-thyroid hormone receptor super family

and which in turn bind to specific sequences within retinoic acid responsive

genes activating their expression (see Chapter 4, section 4.4, and Chapter 8,

section 8.2.2). Hence, the activation of regulatory genes and the initiation of a

regulatory cascade can be achieved by the activation of specific receptors-

transcription factors by an inducing stimulus.

This illustrates therefore how the synthesis of one set of transcription

factors (the homeobox proteins) can be regulated by the activation of another

set of transcription factors (the retinoic acid receptors). In agreement with

this idea, the treatment of mouse embryos with retinoic acid results, for

example, in changes in the expression pattern of the Hoxb-1 gene which

contains a retinoic acid response element in its 3’ regulatory region.

Moreover, the inactivation of this element so that it no longer binds the

retinoic acid receptors, abolishes expression of Hoxb-1 in the neuroectoderm

of the early embryo, providing direct evidence that the retinoic acid response

element is necessary to produce the expression pattern of this gene observed

in the developing embryo (Marshall et al., 1994; for review see Stern and

Foley, 1998).

REGULATION OF TRANSCRIPTION FACTOR SYNTHESIS 223

Figure 7.14

Model for the

progressively more

anterior expression of the

genes in the Hoxb (2)

cluster in which

expression is controlled

by a posterior to anterior

gradient in retinoic acid

concentration and the

increasing sensitivity to

induction by retinoic acid

which occurs from the 5’

to the 3’ end of the

cluster. Thus because

genes at the 3’ end of

the cluster are inducible

by very low levels of

retinoic acid they will be

expressed in anterior

points of the embryo

where the retinoic acid

level will be too low to

induce the genes at the

5’ end of the cluster

which require a much

higher level of retinoic

acid to be activated.

Interestingly, the regulation of Hox gene expression by such DNA response

elements located adjacent to the individual genes appears to interact with

other regulatory processes which operate over the whole gene cluster. Thus,

in experiments where individual Hox genes (with their adjacent control ele-

ments) were moved to a different position withi n the gene cluster, their

pattern of expression was altered so that they behaved similarly to genes

normally located at that position in the cluster for example, in terms of the

time at which they were switched on during development (van der Hoeven et

al., 1996) (Fig. 7.15).

In the case of the HoxD cluster, this effect appears to involve the order of

the genes relative to a distant enhancer element located at least 100 000 bases

away. Thus, in this cluster the first gene, HoxD13 is expressed most anteriorly

and at the highest level with each successive gene being exposed at lower

levels and more posteriorly. If Hox D13 is deleted, the next gene in the cluster

HoxD12 is now expressed in the manner typical of Hox D13 even though it

remains in its normal position (Kmita et al., 2002; for review see Zeller and

Deschamps, 2002). In this case therefore, the genes appear to compete to

interact with the distant enhancer element so that the closest gene is

expressed in a particular pattern and so on (Fig. 7.16). This effect is evidently

reminiscent of the locus control region (LCR) in the -globin gene cluster (see

Chapter 1, section 1.3.4) where the globin genes were expressed in a specific

order in development which is determined by their position relative to the

LCR.

The specific pattern of expression of individual homeobox genes, which is

determined by their position in the cluster is absolutely critical to their func-

tion. Indeed, it appears that it is the different patterns of regulation rather

224 EUKARYOTIC TRANSCRIPTION FACTORS

Figure 7.15

Each gene in a Hox

cluster has its own

specific expression

pattern (panel i). Moving

a particular gene to a

new position in the

cluster results in it having

the expression pattern of

the gene which is

normally located at that

position (panel ii).

than the different proteins that they encode, that determine the different

roles of specific genes in a cluster. Thus, if an individual gene in a cluster is

deleted, the other genes in the cluster cannot substitute for it and an abnor-

mal anim al results (Fig. 7.17). However, if the deleted gene is replaced by a

further copy of another gene in the cluster, then a normal animal results (for

review see Dubou le, 2000) (Fig. 7.17). This occurs because the expression of

the inserted gene is now determined by its position in the cluster and it is

therefore express ed in the manner characteristic of the deleted gene. Hence,

the products of different genes in a cluster can functio nally substitute for one

another but only if they are expressed in the appropriate pattern, as deter-

mined by their position in the cluster. This illustrates the critical role of the

regulated synthesis of transcription factors in allowing them to produce their

functional effects.

The manner in which the regulated synthesis of multiple homeobox factors

can regulate the production of several different cell types has been analysed in

detail in the ven tral neural tube. In this case, the system is regulated by a

gradient in the concentration of a protein signalling molecule known as sonic

hedgehog (Shh) rather than via a retinoic acid gradient. The expression of

several homeobox factors (Dbx1, Dbx2, Irx3 and Pax6) is repressed by Shh,

but their sensitivity to such repression differs so that Pax6, for example, is

expressed at higher Shh concentrations than Irx3 and so on (Fig. 7.18). In

contrast, two oth er homeobox genes are activated by Shh but their sensitivity

differs so that Nkx6.1 is expressed at lower levels of Shh than Nkx2.2 (Fig.

7.18).

REGULATION OF TRANSCRIPTION FACTOR SYNTHESIS 225

Figure 7.16

In the HoxD cluster, the

expression of the genes

is affected by their order

relative to a distant

enhancer element (E)

(panel i). Deletion of

gene A results in gene B

being the closest to the

enhancer. It is therefore

expressed in the normal

pattern for gene A even

though its physical

location is unchanged

(panel ii).

226 EUKARYOTIC TRANSCRIPTION FACTORS

Figure 7.18

In the ventral neural tube, a

gradient of sonic hedgehog

regulates the expression of

several homeobox genes. Dbx1,

Dbx2, Irx3 and Pax6 are

repressed by sonic hedgehog

but differ in their sensitivity to

repression. Thus, Pax6, which

is the least sensitive to

repression, is expressed at

higher sonic hedgehog

concentrations than Irx3 and so

on. Conversely, Nkx6.1 and

Nkx2.2 are activated by sonic

hedgehog with Nkx6.1 being

activated at lower

concentrations than Nkx2.2.

Together these effects create a

homeodomain code in which

each region has a different

pattern of expression of the six

genes and hence different

neuronal types (1–5) form at

each point.

Figure 7.17

The normal expression pattern

of Hox genes results in the

production of a normal animal

(panel i). Inactivation of a

specific gene results in a

mutant animal being produced

(panel ii). However, if the

mutant gene is replaced with a

further copy of another gene in

the cluster, this gene is

expressed in the same way as

the deleted gene and a normal

animal results (panel iii).

The different expression patterns of these genes, therefore, convert the

gradient of Shh expression into a homeobox code in which each region has

a unique pattern of expression of the different homeobox genes. In turn, this

results in five different neuronal types forming at different positions in the

ventral neural tube (Fig. 7.18) (Briscoe et al., 2000). Hence, in this case, the

precise combination of specific homeobox genes expressed in each position

controls the precise cell type that is formed (for review see Marquardt and

Pfaff, 2001).

In our discussion so far, it has been assumed that a homeobox factor is

either present in a particular cell or is entirely absen t. In fact, however, a

further level of complexity exists since many homeo box factors are not

expressed in a simple on/off manner but rather show a concentration gradi-

ent ranging from high levels in one part of the embryo via intermediate levels

to low levels in another part. For example, in Drosophila, the bicoid protein

(bcd) whose absence leads to the development of a fly without head and

thoracic structures is found at high levels in the anterior part of the embryo

and declines progressively posteriorly, being absent in the posterior one third

of the embryo (Fig. 7.19).

Most interestingly, genes that are activated in response to bicoid contain

binding sites in their promoters which have either high affinity or low affinity

for the bicoid protein. If these sites are linked to a marker gene, it can be

demonstrated that genes with low affinity binding sites are only activated at

high concentrations of bicoid and are therefore expressed only at the extreme

anterior end of the embryo. In contrast, genes that have higher affinity binding

REGULATION OF TRANSCRIPTION FACTOR SYNTHESIS 227

Figure 7.19

The gradient in Bicoid

concentration from the

anterior to the posterior

point of the embryo

results in bicoid-

dependent genes with

only low affinity binding

sites for the protein being

active only at the extreme

anterior part of the

embryo, whereas genes

with high affinity binding

sites are active more

posteriorly. Note that, in

addition to the different

posterior boundaries in

the expression of genes

with high and low affinity

binding sites, genes with

high affinity binding sites

will be expressed at a

higher level than genes

with low affinity binding

sites at any point in the

embryo.

sites are active at much lower protein concentrations and will be active both at

the anterior end and more posteriorly. Moreover, the greater the number of

higher affinity sites the greater the level of gene expression that will occur at

any particular point in the gradient (Dr iever et al., 1989; Fig. 7.19).

The gradient in bicoid expression can be translated therefore into the

differential expression of various bicoid-dependent genes along the anterior

part of the embryo. Each cell in the anterior region will be able to ‘sense’ its

position within the embryo and respond by activating specific genes. One of

the genes activated by bicoi d is the homeobox-containing segmentation gene

hunchback. In turn, this protein regulates the expression of the gap gen es,

Kruppel and giant (Struhl et al., 1992). All four of these proteins then act on

the eve gene with bicoid and hunchback activating its expression while

Kruppel and giant repress it. The concentration gradients of these four fac-

tors thus result in the spatial localization of eve gene expression in a defined

region of the embryo where it exerts its inhibitory effects on gene express ion

(Small et al., 1991; Fig. 7.20). Hence, the gradient in bicoid gene expression

results in changes in the express ion of other genes encoding regulat ory pro-

teins leading to the activation of regulatory networks involving the controlled

synthesis of multiple transcription factors.

The bicoid factor therefore has all the properties of a morphogen whose

concentration gradient determines position in the anterior part of the

embryo. This idea is strongly supported by the results of genetic experiments

in which the bicoid gradient was artificially manipulated, cells containing

artificially increased levels of bicoid assuming a phenotype characteristic of

more anterior cells which normally contain the new level of bicoid and vice

versa (Driever and Nusslein-Volhard, 1988).

The anterior to posterior gradient in bicoid leve ls is required to produce

the opposite posterior to anterior gradient in the level of another protein,

caudal. However, the caudal mRNA is equally distributed throughout the

embryo indicating that the bicoid gradient does not regulate transcription

of the caudal gene. Rather, the bicoid protein binds to the caudal mRNA

and represses its translation into protein so that caudal protein is not pro-

duced when bicoid levels are high (reviewed by Carr, 1996; Chan and Struhl,

1997). As well as providing further evidence for the key role of the bicoid

factor, this finding also show s that homeodomain proteins can bind to RNA as

well as to DNA and that they may therefore act at the post-transcriptional level

as well as at transcription.

The bicoid case clearly illustrates therefore how the regulated synthesis of

an individual factor, resulting in a gradient in its concentration, can alter the

expression of a regulatory network of other genes and ultimately control the

differentiation of specific cells during development.

228 EUKARYOTIC TRANSCRIPTION FACTORS

7.3 MECHANISMS REGULATING THE SYNTHESIS OF

TRANSCRIPTION FACTORS

The cases discussed in the previous section illustrate therefore that where a

factor must be active in a particular cell type or at a spec ific point in devel-

opment, this is frequently achieved by the factor being present only in the

particular cells where it is required. Clearly, such regulated synthesis of a

specific transcription factor could be achieved by any of the methods that

are normally used to regulate the production of individual proteins, such as

the regulation of gene transcription, RNA splicing or translation of the mRNA

(Fig. 7.21, for review of the levels at which gene regulation can occur see

REGULATION OF TRANSCRIPTION FACTOR SYNTHESIS 229

Figure 7.20

Model illustrating how the

concentration gradients

of the activators bicoid

and hunchback and the

repressors Kruppel and

giant produce a stripe of

eve gene expression.

Note that the bicoid

gradient (a) affects

hunchback expression (b)

which in turn affects

giant and kruppel

expression (c). Eve gene

activation (+) by

hunchback and bicoid

and its repression (-) by

giant and Kruppel then

produces a specific stripe

or region of the embryo

in which eve is expressed

(d).

Latchman, 2002). Several of these mechanisms of gene regulation are utilized

in the case of individual transcription factors and these will be discussed in

turn.

7.3.1 REGULATION OF TRANSCRIPTION

As discussed above, a number of cases where the cell type-specific expression

of a transcription factor is paral leled by the presence of its corresponding

mRNA in the same cell type have now been described. In turn this cell type-

specific expression of the transcription factor mRNA is likely to result from

the regulated transcription of the gene encoding the transcriptio n factor.

230 EUKARYOTIC TRANSCRIPTION FACTORS

Figure 7.21

Potential regulatory

stages in the expression

of a gene encoding a

transcription factor.