10

Future perspectives

Nick Hastie1
MRC Human Genetics Unit, Western General Hospital, Crewe Road,
Edinburgh EH4 2XU, U.K.

Transcription just one level of regulation

In this volume you will have enjoyed an excellent series of essays on gene
regulation by leaders in their respective fields. The focus is on regulation at the
transcriptional level. Although this is justified from the volume of work done
to date, it is important to acknowledge that, in the flow of information from
gene to protein, there are several intermediate stages at which regulation can
and does take place. These include RNA splicing, mRNA transport, mRNA
stability and translation. The excellent essay on translational control by Chris
Proud (see Chapter 8 in this volume) demonstrates how this can offer many
advantages over transcription as a control point. However, this is not the
whole story, because post-translational modification of proteins is taking on
greater and greater regulatory significance, be it phosphorylation, acetylation
or methylation all strategies for rapidly modifying protein function, often
in response to extrinsic or intrinsic signals.
Although we are starting to get a glimpse of how these different processes
work, we are only scratching the surface in many respects. In reality, it is clear
that there is co-ordination between these different steps and, in the future, we
will not be able to discuss control at transcription without reference to the up-
stream and downstream regulatory events. For example, the level of a transcrip-
tion factor itself can be regulated at the translational level and as described in the
essay by Melanie Lee and Stephen Goodbourn (see Chapter 6 in this volume);
transcription factors may often be tethered in an inactive form in the cytoplasm
and only become active and targeted to the nucleus after phosphorylation.

1e-mail: [email protected]

121
122 Essays in Biochemistry volume 37 2001

One exciting new development not covered in this book is the direct link
between transcription and RNA processing. Over the past few years it has
become evident that splicing is carried out co-transcriptionally. Furthermore,
the C-terminal domain of RNA polymerase II binds directly to splice factors,
thus recruiting the splicing apparatus to nascent transcripts [1]. The link goes
beyond this, because this domain of the polymerase also interacts with capping
and termination enzymes that allow the transcripts to be completed.
Remarkably, the strength and type of promoter can also influence the pattern
of alternative splicing [2].

Transcription factors and genetics

One way to assess the importance of transcription in regulation is to peruse
the classical genetics literature. In their historic screen for developmental
mutants in Drosophila, Nusslein-Volhard et al. (for review see [3]) identified a
large number of mutants deficient in early patterning of the embryo and later
segmentation. Previously, Lewis had identified the classic series of homeotic
mutants associated with transformation of body structures from one type to
another. We now know that a very significant proportion of the genes
defective in these mutants encode transcription factors, the majority of these
being proteins containing homeobox or zinc-finger motifs [3]. Furthermore,
Drosophila developmental mutants led to the identification of the major
intercellular signalling pathways involving the Wnt, Hedgehog and BMP
(bone morphogenetic protein) families of regulatory proteins. These pathways,
which are conserved in vertebrates, all work by activating transcription factors.
To my knowledge, very few of the classical mutants involve proteins that
regulate gene expression at the post-transcriptional level; however, there might
be some bias here as DNA-binding proteins may have more recognizable
signatures. Notable exceptions include the Transformer and Transformer 2
gene products, RNA-binding proteins that control sexual differentiation in
Drosophila by regulating the alternative splicing of the Doublesex transcript
[4]. Also, the first vital step in development is to set up asymmetries in the egg.
This is achieved in part by asymmetric localization of the mRNAs encoding
the very transcription factors that regulate early pattern formation. In turn, it
is specific mRNA-binding proteins that take the messengers for these
transcription factors to their specific subcellular locations in the egg. So
mRNA-binding proteins play essential roles in the very earliest stages of
development.
Possibly the most profound demonstration of the importance of transcrip-
tion factors comes from an assessment of their involvement in human genetic
disease. Genes encoding transcription factors of many varieties have been
found to be mutated in a large number of human genetic disorders [5]. These
include many tissue-specific or developmental-stage-specific regulators that
bind directly to DNA. For example, the genes encoding proteins with the
N. Hastie 123

paired DNA-binding motif, PAX6, PAX3 and PAX2, are mutated in aniridia,
Waardenburgs syndrome and renal coloboma syndrome, respectively.
Remarkably, these are all cases of haplo-insufficiency, where one copy of the
gene is mutated and the remaining 50% dose is insufficient for normal devel-
opment. Transcription factors of the steroid receptor class are mutated in a
variety of diseases, including male breast cancer and vitamin D-resistant rickets
[5]. Maturity-onset diabetes can arise through mutations in any one of several
transcription factors that work together in pathways to control pancreas dif-
ferentiation and the expression of insulin [5].
Arguably the most dramatic insights into the physiological and clinical
significance of transcription factors have come through investigation into the
genetic mechanisms underlying cancer. One of the biggest conceptual break-
throughs in cancer has been Knudsons two-step hypothesis and the conse-
quent discovery of the tumour-suppressor genes. The most famous and com-
monly mutated tumour-suppressor genes, p53 and Rb1, both work by regulat-
ing transcription p53 as a direct DNA-binding factor and Rb1 as a co-
repressor (see the essay by H.M. Chan et al.; Chapter 7 in this volume).
Inherited mutations in these genes lead to very rare cancer syndromes.
However, somatic mutations in these two genes occur in 50%, or more, of all
of the common (sporadic) cancers. This new knowledge not only offers novel
therapeutic avenues, but has also provided powerful insights into the regula-
tion of the cell cycle and the response to DNA damage.
A vast range of transcription factors have been found to be activated by
chromosome translocation in a variety of leukaemias [6]. Perhaps the most
exciting recent development has been the finding that many chromatin modi-
fiers, rather than DNA-binding regulators, are mutated in disease. As dis-
cussed in the essay by Alan Wolffe (see Chapter 4 in this volume), the biggest
paradigm shift in the field of transcription has been the discovery that chro-
matin modification plays a key role in transcriptional regulation. Various mod-
ifying complexes, both those working as activators and those working as
repressors, have been identified and a number of components of these com-
plexes have been found to be mutated or rearranged in human disease, particu-
larly in leukaemias and lymphomas.
Clearly it is exciting that these studies are revealing insights into the mech-
anisms of human disease and raising the prospect of new therapies. However,
from a fundamental perspective it is pleasing that the human genetic findings
validate the physiological importance of these factors and help both to support
the biochemical studies and to place genes in cellular pathways.

What does the genome sequence tell us?

We are sitting at a major crossroads in biomedical science as the sequence of
the human genome is nearing completion and the sequences of yeast,
Caenorhabditis elegans and Drosophila genomes are complete. It is
124 Essays in Biochemistry volume 37 2001

illuminating to dip into the sequence databases to assess the abundance of

different classes of transcription factor gene within each genome and to see
how this has changed with evolutionary complexity. In flies, genes encoding
putative transcription factors with C2H2 zinc-finger domains are the most
highly represented gene class in the genome, with 352 members [7]. There are
also 130 genes encoding helixloophelix DNA-binding proteins, and 113
genes encoding homeobox proteins. It is interesting to note that, whereas there
are more genes in C. elegans (18,424) than Drosophila (13,601), there are fewer
representatives of each of these categories of putative transcription factor.
Hence there are predicted to be only 138 zinc-finger proteins, 88 homeo-
domain proteins and 46 helixloophelix proteins in C. elegans. So the increa-
sed complexity of putative transcriptional regulators has co-evolved with
increased biological complexity. However, another provocative observation is
that the number of genes encoding proteins with RNA-recognition motifs has
also increased from 55 in yeast, to 92 in worms, to 160 in flies. It will come as
no surprise that we have no idea about the functions of the majority of these
putative regulators, either transcriptional or post-transcriptional. We expect
that the numbers of DNA- and RNA-binding factors in humans will have
increased dramatically over the numbers found in invertebrates. Again this
shows the scale of the challenge that awaits us.
As various genome sequences are reaching completion, the part played by
alternative processing is gaining increased attention. It was a surprise to many
that a fly has fewer genes than a worm, in spite of its more complex make-up
and biology. One potential explanation for this is that there is more alternative
splicing in flies, leading to greater complexity at the protein level. This possi-
bility has been highlighted by the recent, remarkable finding that a gene
encoding a fly axon-guidance receptor can produce potentially 38000 different
proteins through a combination of differential splicing patterns [8]. Schmucker
and co-workers speculate that this molecular diversity may contribute to the
specificity of neuronal activity. This whole area of regulated alternative pro-
cessing has received much less attention than transcriptional control, but this is
likely to change in the near future.

Multifunctional regulators
It is also becoming more and more obvious that some proteins, which have
been predicted on structural grounds to be transcription factors, may be
multifunctional, acting at the post-transcriptional as well as the transcriptional
level. For example, the famous bicoid DNA-binding transactivator also binds
to the RNA transcripts of target genes, regulating translational efficiency [9].
The Tra-1 transcription factor, necessary for normal sexual development in C.
elegans, not only regulates the transcription of the Tra-2 gene but also binds to
the Tra-2 transcript, regulating its transport into the cytoplasm [10]. Another
remarkable example is the Wilms tumour-suppressor gene, WT1, which
N. Hastie 125

encodes multiple protein isoforms with four DNA-binding zinc fingers.

Whereas one of the major isoforms has high affinity for DNA and seems to act
as a transcription factor, the other major isoform binds with higher affinity to
RNA and interacts with splicing factors in the cell [11]. Such proteins may co-
ordinate the processes of transcription and post-transcriptional events in
particular developmental pathways. We must be careful in attributing a
transcriptional function to a protein just because it has predicted DNA-
binding motifs. Many DNA-binding proteins can also bind RNA with some
sequence specificity and it may turn out that this is their primary role [12].

The future massively parallel and computational

approaches
To understand fully development, physiology and the mechanisms underlying
disease, we will have to work out the functions of all 60 000100 000 human
genes how their products interact and talk to each other in complex
networks. Scientists are just starting to use computational approaches to model
complex interactions between developmental genes in Drosophila [13]. To deal
with all this complexity, new massively parallel approaches are being
developed to study gene function. These include DNA chips or microarrays
which make it possible to study the expression of tens of thousands of genes
simultaneously. This approach can be used to identify the putative targets of a
transcription factor. For example, in an elegant study, Soukas and co-workers
[14] used microarrays to identify many of the mRNAs that were induced by
the weight-controlling hormone, leptin, in adipose tissue. Among these were
the mRNAs encoding some transcription factors already known to be induced
in adipose tissue, along with the mRNAs for their suspected target genes.
However, other transcription factors not known to be involved in the
responses to leptin were also induced, with the same kinetics as several other
genes which may turn out to be their targets. Thus, in one fell swoop, we are
learning about the biological response to hormones and identifying
transcriptional circuits.
All this has led to a massive rise in a new field of science: bioinformatics.
Analysis of the change in expression of many thousands of genes simultane-
ously has necessitated the development of sophisticated computing tools and
massive databases. One laboratory may observe, for example, that ten specific
genes are frequently expressed at higher levels in breast cancer cells compared
with levels in normal mammary tissue. The scientists then access the databases
and find that others have shown that the same set of genes is repressed by
oestrogen, or perhaps down-regulated during normal mammary development.
The more the databases expand, the more that can be learned from an experi-
ment. Particularly exciting will be the information that is obtained on the
many thousands of genes that encode so-called pioneer proteins, i.e. those
with no obvious biochemical function revealed by their gene sequences.
126 Essays in Biochemistry volume 37 2001

Ultimately, we will want to have a record of the expression patterns of all

genes in time and space during development a four-dimensional pattern of
gene expression and the only way to do this is electronically. Scientists in
Edinburgh, at the MRC Human Genetics Unit and the University, have now
started to develop a computerized, three-dimensional representation of a
mouse at different stages of development (http://genex.hgu.mrc.ac.uk/). This is
already a very powerful anatomy teaching tool. Into this atlas will be placed all
the gene-expression patterns. It will then become possible to use computation-
al approaches to identify clusters of co-regulated genes during development
and to relate this to microarray data. This atlas will contribute enormously to
our understanding of developmental processes, gene networks and disease
mechanisms.

Summary

Transcription is coupled to splicing and other post-transcriptional

processes.
The importance of transcription factors in developmental biology and
disease is underlined by genetic analysis in flies and humans.
The genome project is identifying large numbers of novel transcription
factors and RNA-binding proteins.
Proteins may have multiple functions, acting at the transcriptional and
post-transcriptional levels.
The vast amount of novel biological information requires new, high-
throughput approaches and bioinformatics.

References
1. Bentley, D. (1999) Coupling RNA polymerase II transcription with pre-mRNA processing. Curr.
Opin. Cell Biol. 11, 347351
2. Cramer, P., Cceres, J.F., Cazalla, D., Kadener, S., Muro, A.F., Baralle, F.E. & Kornblihtt, A.R.
(1999) Coupling of transcription with alternative splicing: RNA pol II promoters modulate
SF2/ASF and 9G8 effects on an exonic splicing enhancer. Mol. Cell 4, 251258
3. Lawrence, P.A. (1992) The Making of a Fly: the Genetics of Animal Design, Blackwell Sciences Ltd,
Oxford
4. Nagoshi, R.N. & Baker, B.S. (1990) Regulation of sex-specific RNA splicing at the Drosophila dou-
blesex gene: cis-acting mutations in exon sequences alter sex-specific RNA splicing patterns. Genes
Dev. 4, 8997
5. Engelkamp, D. & van Heyningen, V. (1996) Transcription factors in disease. Curr. Opin. Genet. Dev.
6, 334342
6. Rabbitts, T.H. (1994) Chromosomal translocations in human cancer. Nature (London) 372,
143149
7. Rubin, G.M., Yandell, M.D., Wortman, J.R., Gabor Miklos, G.L., Nelson, C.R., Hariharan, I.K.,
Fortini, M.E., Li, P.W., Apweiler, R., Fleischmann, W., et al. (2000) Comparative genomics of the
eukaryotes. Science 287, 22042215
N. Hastie 127

8. Schmucker, D., Clemens, J.C., Shu, H., Worby, C.A., Xiao, J., Muda, M., Dixon, J.E. & Zipursky,
S.L. (2000) Drosophila Dscam is an axon guidance receptor exhibiting extraordinary molecular
diversity. Cell 101, 671684
9. Rivera-Pomar, R., Niessing, D., Schmidt-Ott, U., Gehring, W.J. & Jackle, H. (1996) RNA binding
and translational suppression by bicoid. Nature (London) 379, 746749
10. Graves, L., Segal, S. & Goodwin, E.B. (1999) TRA-1 regulates the cellular distribution of the tra-2
mRNA in C. elegans. Nature (London) 399, 802805
11. Larsson, S.H., Charlieu, J.-P., Miyagawa, K., Engelkamp, D., Ross, A., van Heyningen, V. & Hastie,
N.D. (1995) Subnuclear localization of WT1 in splicing or transcription factor domains is regulat-
ed by alternative splicing. Cell 81, 391401
12. Ladomery, M. (1997) Multifunctional proteins suggest connections between transcriptional and
post-transcriptional processes. Bioessays 19, 903909
13. Von Dassow, G., Meir, E., Munro, E.M. & Odell, G.M. (2000) The segment polarity network is a
robust developmental module. Nature (London) 406, 188192
14. Soukas, A., Cohen, P., Socci, N.D. & Friedman, J.M. (2000) Leptin-specific patterns of gene
expression in white adipose tissue. Genes Dev. 14, 963980