WITH a rough draft of the entire human DNA sequence in the bag, it might seem
time for biologists to take a well-deserved rest. In fact, the real work is only
just beginning. If the molecular drama that constitutes human biology were a
movie, the DNA sequence of the human genome would be no more than a cast
list鈥攁nd one written in a foreign language, at that.
The challenge now is to find out how cells actually use the information in
our DNA. Biologists are already busy identifying which parts of that DNA
represent genes鈥攖hose sections that can generate RNA copies, which in turn
are used to build protein molecules. And by studying the roles played by DNA,
RNA and proteins, they hope to answer key questions about life鈥檚 drama. Who
plays the leading roles in a liver cell, say, and who are the extras? Who are
the villains behind cancer and other diseases?
The role of DNA and RNA is primarily to store information, while proteins are
the major actors in the biochemistry of the cell. Failure of any of these three
players can be disastrous. A genetic disease could result from a deletion of the
protein-coding sequence of a gene, or defects in sequences of DNA that regulate
genes, resulting in low levels of RNA. Even more subtle genetic defects still
can result in a crippled protein.
Advertisement
The goal of the publicly funded Human Genome Project and commercially driven
Celera has been to record all 3 billion DNA 鈥渓etters鈥�, or base pairs, on the 23
chromosomes in our cells. Information from genome sequencing is already helping
biologists to discover how changes to DNA can lurk behind an illness. They can
do this by tracking down or 鈥渕apping鈥� exactly where on a particular chromosome
these changes occur. Such mapping projects depend on the fact that despite
obvious differences between people, more than 99 per cent of your DNA is
identical to that of anyone else. It is in the remaining fraction of a per cent
that the differences lie, from hair colour to height, from a susceptibility to
cancer to a tendency towards obesity.
Geneticists can use variations in a single DNA letter, called single
nucleotide polymorphisms (SNPs), as landmarks on the vast tracts of each
chromosome. This allows them to follow those gene variants in a population, and
assess the part they play in disease. In 1999, an academic-commercial
partnership called the SNP Consortium set out to map 150 000 SNPs spaced evenly
across the human genome by 2002. Last month, on its first anniversary, the
consortium announced that it had already mapped 102 719 SNPs, and might complete
its initial goal by the end of this year. 鈥淭he flood of information from
sequencing projects has put us far ahead of any schedule we imagined,鈥� says
Dalia Cohen of Novartis in Basle, Switzerland, one of the companies
involved.
Once particular DNA sequences are linked to a disease, the next step is to
home in on the genes in that region by looking for typical sequences genes tend
to contain. But that鈥檚 harder than it sounds. Genes account for only about 3 per
cent of the genome, and are mixed in with a large excess of 鈥渏unk鈥� DNA that does
not encode RNA or proteins. In addition, human genes are split up into pieces
along the chromosome which are spliced together only after the RNA copy is
generated. 鈥淕ene prediction is as much an art as a science,鈥� says Gerald Rubin
of the University of California at Berkeley, who led the academic team that
collaborated with Celera to deliver the DNA sequence of all the genes in the
fruit fly Drosophila melanogaster (Science, vol 287, p
2185).
Computers are playing a big part in detecting genes. A report last month in
Genome Research (vol 10, p 483) described the results when 12 labs
unleashed their gene-sniffing programs on a well-studied 2.9 million base-pair
chunk of the Drosophila genome. The programs spied an impressive 86 to
97 per cent of the protein-coding DNA in the region. But they fared less well
arranging those sequences into discrete genes and separating genes from junk.
The programs completely missed between 5 and 16 per cent of genes in the region.
And between 11 and 52 per cent of what they found didn鈥檛 appear to be genes at
all.
Finding out whether a certain gene has a counterpart in another organism will
add to our understanding of the human gene map. This type of comparative
genomics also helps with another goal: identifying regulatory DNA sequences that
lie hidden in junk DNA. Eddy Rubin of Lawrence Berkeley National Laboratory in
California says the sequences that control when and where genes are turned on
play a key role in evolution. He points out that all mammals, from mice to
humans, have very similar genes and make similar proteins. They turn out so
different because they switch genes on and off in different ways. 鈥淚t鈥檚 like
having bricks. You can build a garage out of them or a skyscraper. It isn鈥檛 the
material, it鈥檚 the way you use it.鈥�
Last month, Eddy Rubin鈥檚 team took a region of human DNA a million base pairs
long and hunted for similar non-gene sequences in the analogous region in mice,
reasoning they must have an important role in mammalian evolution. In a region
with just 23 genes, they found 90 highly conserved sequences, what Rubin calls
鈥渏ewels within the junk DNA鈥�. Tests on one of these regions showed it has
dramatic effects on the expression of genes thousands of bases away. 鈥淭his shows
genes are just the tips of icebergs of information we can abstract from the
genome,鈥� he says.
Another way to gain more information is to compare DNA and RNA. Finding an
RNA copy of a DNA sequence, for instance, is excellent proof that a DNA region
is a gene. And because RNA analysis reveals which genes are switched on in any
cell, it is an important clue to cell function. Companies such as Human Genome
Sciences in Maryland and Incyte Genomics in California have been busy for years
cataloguing all the human RNA molecules they can find, banking on this being a
quick way to discover marketable biological secrets. 鈥淲hen you look at RNA, you
are looking at what evolution has programmed a cell to do in a particular
situation,鈥� says Randall Scott, president of Incyte.
The last member of life鈥檚 molecular trinity, the proteins, is the most
daunting group to study. After proteins are made, they can undergo a startling
array of transformations such as having pieces chopped off, being decorated with
sugars, or changing their shape鈥攁ll of which can alter their function.
Biologists suspect the functional set of human proteins鈥攖he
鈥減roteome鈥濃�攃ould outnumber human genes by a factor of 10.
Aiming for a quick snapshot of every protein in a cell, biologists at Oxford
GlycoSciences have taught some old biochemistry techniques new tricks. Over the
past four years, they have automated the entire process of gel electrophoresis,
the old standby of protein analysis that sorts protein molecules by their charge
and size.
This has allowed them to quickly find out how well a drug stops production of
proteins, for instance. Proteins from cells that have or have not been treated
with the drug are run out on gels, stained with a dye, and then compared for any
significant changes in protein production. Robots cut out these regions of the
gels, purify the protein and send it to a mass spectrometer, which can determine
the mass and amino acid sequence of the protein, which in turn leads back to a
gene in the human DNA blueprint.
Eventually, an even faster way to study proteins might come from dispensing
with gels altogether. A team led by Ruedi Aebersold at the University of
Washington in Seattle has devised an ingenious way to measure relative amounts
of proteins in, say, cancerous and non-cancerous cells, which could lead to new
treatments. First, proteins in both samples are digested into fragments and
attached to molecular tags containing two different chemical isotopes. At that
point, the samples can be mixed and fed into the mass spectrometer together. The
machine can then separate out the proteins, measure the relative amounts of the
two different isotopes, and then identify the fragments (Nature
Biotechnology, vol 17, p 994). 鈥淲e can also deal with proteins that are too
big, too small, or too greasy to enter a gel,鈥� says Aebersold.
The hope is that this new wave of production-line biology will quickly give
us invaluable information, such as revealing which proteins are involved in
cancer. But further down the line, Sydney Brenner of the Molecular Sciences
Institute in Berkeley sees its limitations. Really understanding what any
protein, RNA or gene does requires 鈥済oing back to the bench鈥� and studying each
molecule鈥檚 idiosyncrasies, he argues.
Brenner reckons that each gene and its products will require about 40 years
of study. 鈥淭he human genome will inspire at least 50 000 professorships,鈥� he
says. It seems that resting is the last thing biologists will be doing.
