ÐÓ°ÉÔ­´´

Imagine a bustling metropolis, sustained by myriad restaurants. At
the heart of each is a busy kitchen with a table upon which lie volumes
of a huge, ancient cookbook. Every restaurant relies on the same cookbook.
Yet each uses a different selection of recipes – some churn out hamburgers
exclusively, while others cook up exotic dishes such as moo-shoo pork and
profiteroles. The various menus are created by mysterious behind-the-scenes
chefs who specialise in particular cuisines.

Bizarre as it sounds, this kitchen scenario is a good analogy for what
appears to be going on in the nuclei of living cells. The recipes are genes,
the cookbook chromosomes and the chefs transcription factors – protein molecules
that bind to DNA with the effect of switching genes on or off. Together,
they hold the secret to one of the greatest puzzles in biology: how the
virtually undifferentiated cells of the early embryo, all with the same
genetic blueprints, develop into distinct tissues each dominated by the
actions of distinct subsets of genes. How do liver cells, for example, manage
to switch on genes for making detoxifying enzymes while leaving other genes
dormant? How do skin cells ‘know’ that they need to draw on the genetic
instructions for making fibrous keratin, but not, say, retinal photoreceptor
proteins or the oxygen-carrying protein haemoglobin?

Over the past decade or so, biologists have been closing in on an answer,
guided by an obvious strategy: get to know the chefs. As a result, transcription
factors have become one of the hottest topics in molecular biology. Leading
science journals vie for papers describing the latest advances in understanding
how transcription factors latch on to DNA. Researchers compete to catalogue
their molecular sequences, shapes and biochemical properties. International
scientific conferences are convened to discuss the most recent results.

Now, with international efforts well under way to map and then sequence
the human genome, biologists are anticipating an avalanche of new information.
Much is already known about transcription factors from ‘simple’ organisms
such as bacteria and yeast, while research into the proteins that switch
genes on and off in the fruit fly and nematode worm – creatures with well-mapped
molecular genetics – is progressing at a hectic pace. But with mammals,
and in particular humans, it is still early days. ÐÓ°ÉÔ­´´s don’t yet know
enough about the molecular genetics of mammals to embark on a full-scale
assault on transcription factors. The human genome project could help to
change that.

Yet even at this stage, the importance of transcription factors is clear.
Without them, sun-drenched skin cells could not produce the protective pigment
melanin, the hormone oestrogen could not trigger ovulation nor testosterone
deepen the voices of teenage boys, and toxic chemicals such as dioxin could
not wreak havoc on cells. As Robert Tijan of the University of California
at Berkeley, a leading expert on transcription factors, puts it: ‘Everything
in biology comes down to how the genes are turned on and off.’

But there is much to learn. How many different transcription factors
are there in the human body? Which ones influence which genes, and how?
Do these proteins act alone or in teams? What part do they play in human
diseases? Could scientists exploit their growing knowledge of transcription
factors to combat diseases such as cancer and AIDS?

Cancer can result when genes involved in cell growth are turned on
inappropriately. AIDS develops as the human immunodeficiency virus (HIV)
invades cells, unleashes its own transcription factors, and commandeers
gene expression. And heart disease is at least partly caused by deleterious
genes that are expressed in the unlucky carriers through the normal action
of their transcription factors. Many scientists hope to gain control of
gene expression as a strategy for treating such diseases. Already, research
is under way to develop antiviral agents through inactivation of viral
transcription factors.

Gene control

Stephen Burley at the Howard Hughes Medical Institute and Rockefeller
University in New York sees this approach as working hand-in-hand with gene
therapy. This is the idea of inserting undamaged genes into cells to override
the effects of damaged genetic material. Burley’s vision is that one day,
scientists will be able to use transcription factors to control the expression
of these healthy replacement genes. ‘Like having brakes and a gas pedal,’
says Burley, ‘you could rev up the transcription level or turn it off if
it turns out to be harmful to the patient.’

But for the time being, most researchers are still looking at the basic
science of transcription factors. One of the biggest surprises so far is
how many transcription factors there are. ‘Regulatory factors were supposed
to be enormously scarce,’ says molecular biologist Michael Levine of the
University of California, San Diego, referring to conclusions drawn from
older studies on bacteria. But it is now clear that at least 1000, and perhaps
as many as 10 000, of the estimated 100 000 genes that make up the genome
of a typical mammal may encode transcription factors. Moreover, says Levine,
the transcription factors can occur in great concentrations, sometimes as
many as 20 000 molecules in a cell.

Another surprising discovery is that different transcription factors
can collaborate to produce a particular effect on a gene, much as groups
of chefs with various specialities can cooperate to create many different
menus. The number of possible combinations of transcription factors in mammals
could be vast. This may explain why our genes can be ‘interpreted’ by cells
in so many different ways. Molecular biologists are now going hell for
leather to find out how various transcription factors combine in cells to
turn on, or ‘express’, different sets of genes.

Since the 1960s it has been clear that genes become expressed through
a two-step mechanism. First, the DNA code of the gene is copied, or transcribed,
by the synthesis of a strand of messenger RNA which repeats the code found
on the gene. This process requires the help of an enzyme called RNA polymerase,
which must first bind to a region of the gene called the ‘promoter’ before
passing along the DNA, assembling the messenger RNA. Next, enzymes in the
cell use these messenger RNA molecules as templates for building proteins,
each of whose chemical structures is specified by the original DNA codes.
But, this mechanism doesn’t explain how each cell in our body manages to
select and express only certain genes.

At first researchers postulated that transcription was regulated by
special protein subunits of RNA polymerase, dubbed ‘sigma factors’, which
had to join up with the main part of the RNA polymerase, before transcription
could occur. It was in looking for these sigma factors that the true answer
began to emerge. In the early 1980s, Tijan discovered a protein called SP1
in human cells, and Rockefeller’s Robert Roeder found a protein called
TFIIIA in frogs’ eggs. Both these proteins helped RNA polymerase but they
could also activate specific genes. Now there was a plausible model for
gene-specific activation.

Since then, researchers have identified several hundred transcription
factors in yeast cells and mammals, including humans, that function similarly.
The defining characteristic of these proteins is that they all have structures
called ‘DNA-binding domains’. These allow proteins to ‘recognise’ a particular
stretch of DNA and dock into the groove between its two nucleotide strands
like a key in a lock. ÐÓ°ÉÔ­´´s have made great strides in uncovering the
shapes of these molecules using techniques such as X-ray crystallography
and nuclear magnetic resonance spectroscopy. They fall into families with
names that reflect the shape of the protein, such as leucine zippers, helix-loop-helix
proteins and zinc fingers. Each type of protein locks onto DNA grooves with
its own distinctive type of molecular key.

The leucine zipper proteins have generated much excitement because they
shed light on how transcription factors can collaborate to produce a range
of different effects on genes. The best studied are proteins called fos
and jun, which stimulate cells to divide and have been implicated in the
chain of events leading to cancer. Each of these molecules functions like
one half of a zip – the teeth being a string of leucine amino acids – which
can be completed in three different ways: with one molecule of fos linked
to another fos, one jun molecule linked to another jun, or a fos linked
to a jun. These three pairings produce leucine zippers that bind to DNA
in three different ways to produce a variety of responses.

As the research progresses, new families of transcription factors with
quite different DNA-binding domains are sure to emerge. But the story is
unlikely to end there. ‘For years we’ve thought that if we understood the
DNA-binding sequence, that would tell us about transcription regulation,’
says Tijan. Now he and others are coming to appreciate that this is not
enough. An equally important component of these molecules is a part that
enables them to interact with each other, a region called the ‘activation
»å´Ç³¾²¹¾±²Ô’.

Activators and repressors

The shapes and chemical properties of these domains are not yet known,
but they have the all-important job of deciding which transcription factors
can fit together, like pieces in a jigsaw, to form a molecular complex.
Evidence is fast accumulating that it is the overall shape and size of the
complex, not those of its protein pieces, that decides which gene it switches
on. In-deed, biologists now have a general outline of what must happen before
a gene can be switched on in a cell. First, the cell must produce each of
the required transcription factors. Then these proteins must slot together
to form a complex. Finally, this complex must have just the right shape
to plug into an intricate array of grooves on the DNA. Only after that can
an RNA polymerase enzyme begin the copying of DNA into RNA.

The detailed picture is a little more complicated. For a start, some
transcription factors act in a negative sense, to prevent molecular complexes
from activating their target genes. ‘Repressor’ proteins first came to light
two decades ago, when scientists at Harvard University discovered a protein
in phages that could block gene expression in infected bacteria. They act
by binding to DNA and preventing RNA polymerase from moving along it. Now
it is clear that repressor proteins can work in other ways, by blocking
construction of the DNA-binding complex or interfering with its binding
to DNA. ‘Activators and repressors probably duke it out with one another,’
says Levine, whose work seems to confirm this view.

Levine has spent nine years studying how genes are turned on and off
during the development of fruit fly embryos. He noticed that variations
in the levels of a transcription factor called ‘dorsal’ act to shape the
development of muscles, internal organs, skin and nervous tissues. In other
words, different levels of this transcription factor seem to switch on different
sets of genes, some characteristic of muscles, others of skin and so on.
Levine puzzled over this observation until further research produced the
key: dorsal’s versatility was all down to the fact that it could react with
two other transcription factors. Various combinations of the three transcription
factors were coming together at three binding sites on DNA. Whether dorsal
switched nearby genes on or off depended critically on the spacing between
the sites and the presence of the other two transcription factors. ‘In certain
promoter regions, dorsal finds itself next-door to bad company, co-repressors,
which block its ability to act as an activator and convert it into a potent
repressor,’ says Levine.

Research by other biologists has shown that this is just one of the
ways transcription factors can flip, Jekyll-and-Hyde like, between conflicting
roles. Some transcription actors change character in response to signals
from hormones. In effect, they behave as receptors for many of the body’s
hormones, such as thyroid. In test tube and cell culture studies, for instance,
the thyroid receptor keeps certain growth-related genes dormant. But when
thyroid hormone comes on the scene it binds to the receptor, binding it
into a gene activator.

Some proteins perform the duties of a transcription factor on a strictly
part-time basis. For example, the usual job of NF-k B, a protein important
in inflammation and immunity, is to sit tight in the cell fluid, tethered
to a second protein. But for reasons not yet clear, the bond between these
two proteins is occasionally broken, allowing NF-k B to enter the cell
nucleus and act as a transcription factor for certain genes.

Working together

Researchers also suspect that the way genes are packaged influences
transcription. In cell nuclei, DNA is not only double-stranded and helical,
it is also wrapped around large, globular proteins called nucleosomes. Numerous
other proteins also help stabilise the DNA and build up the chromosome.
Biologists are just beginning to look at whether interactions with these
proteins could make some genes more accessible than others to transcription.
Evidence already exists that the tagging of a gene with methyl chemical
groups can shield it against transcription and result in ‘genomic imprinting.’
(‘Why genes have a gender’, 22 May 1993).

Meanwhile, researchers continue to find new proteins involved in transcription.
In all, as many as 60 proteins may be working together to turn on a single
gene. Besides the gene-specific transcription factors, there are proteins
that appear to foster chemical communication between the complex of transcription
factors, RNA polymerase and DNA. Still other proteins may help direct the
construction of the complex. Putting the whole story together will be difficult.

Fortunately, we do not need to understand the entire mechanism before
medical applications can be explored, says Steven McKnight, discoverer of
the leucine zipper and now research director of Tularik, in San Francisco.
He is hoping that the use or misuse of one or several key transcription
factors will be at the root of most diseases. McKnight contends that each
transcription factor may use different mechanisms to turn some genes on
and others off, but that all of its target genes will code for proteins
involved in a single biological response, such as inflammation or cell
growth. That means it should be possible to treat some diseases by interfering
with a single transcription factor.

The challenge now is to figure out which transcription factors have
the greatest impact on diseases, such as cancer, AIDS and arthritis. ‘We’re
just beginning to know enough to say this factor controls this process,
and we only know in a few isolated examples. But over the next 10 or 15
years, people will sort out the circuitry that controls gene expression
in humans,’ McKnight predicts. ‘It’s a lot of work, but in the end, we’ll
have a better chance at picking our targets for controlling gene expression.’

Along the way, other intriguing intellectual questions may well be
answered. For example, have transcription factors become increasingly important
with the evolution of more complex creatures? ‘It will be interesting,’
says Levine, ‘to see if the percentage of transcription factors in higher
organisms is more than in lower organisms – that is, whether the computer
is the same, but the soft- ware gets more complicated.’

Already, developmental biologists have found that organisms as diverse
as corn, yeast and humans contain very similar stretches of DNA called homeoboxes.
Discovered in the 1980s in genes that control construction of the basic
body plan of fruit flies, homeoboxes consist of sequences of 180 base-pairs
that code for DNA-binding domains. Researchers now know that the genes
that contain them encode a very large and important class of transcription
factors which coordinate the activity of arrays of other nearby genes. Indeed,
these so-called homeodomain transcription factors regulate clusters of genes
that are arranged along the chromosome in a sequence that corresponds to
the order – from head to tail – of the body parts whose development they
control.

Insights into human biology and medicine are now beginning to emerge.
Some genetic diseases – certain types of leukaemia and syndromes that affect
nervous system development – have been linked to defects in homeodomain
transcription factors. Much more shall be learnt from studies of laboratory
mice, which share with humans a similar pattern of four rows of homeobox-containing
genes, each one on a separate chromosome. ‘In a general sense, what we find
in mice, we’ll find in humans,’ says Matthew Scott, a developmental biologist
at the Howard Hughes Medical Institute and Stanford University in California
and a discoverer of homeoboxes. ‘But at what level will we start to see
differences?’ That is not yet clear.

For those interested in discerning common patterns within nature, as
well as appreciating the beauty of individuality, the study of gene expression
provides rich rewards. ‘Since every gene is different, you’re looking at
a new animal every time you look at a new gene,’ explains Albert Baldwin,
from the University of North Carolina at Chapel Hill. ‘That’s what makes
transcription so exciting – it’s like snowflakes or human beings – every
gene will be regulated somewhat uniquely.

Karen Schmidt is a writer based in Chapel Hill, North Carolina.