WHAT would happen if you scaled the gates of Buckingham Palace and strolled
across the courtyard to have a chat with the Royal Family? You wouldn鈥檛 get far
before the guards identified you as an intruder and wrestled you to the ground.
Casual visitors aren鈥檛 welcome in those exalted precincts.
Nor are casual genes welcome in the high society of a cell鈥檚 genome. There,
too, intruders usually mean trouble. A gene that doesn鈥檛 belong might come from
a virus trying to commandeer the cell鈥檚 machinery to make more viruses. Or it
could be a transposon鈥攁 good-for-nothing bit of selfish, parasitic DNA
that hangs around the genome like an unemployed relative. Whatever their origin,
these unwanted visitors often trigger a security crackdown within the cell that
silences the offending gene every bit as swiftly and vigorously as guards would
deal with a placard-waving hooligan at the Queen鈥檚 front door.
This genomic police force operates in everything from protozoa and fungi to
plants and fruit flies. And within the past few months, as researchers have
begun to flesh out its details, they are discovering that 鈥済ene silencing鈥� may
be a defence system as old as nucleated cells themselves. Already, genetic
engineers鈥攚ho once viewed gene silencing as a nuisance, because it can
switch off the genes they have so painstakingly inserted鈥攁re using it as a
tool to switch off genes at will, potentially revealing gene functions or
generating genetically modified crops that lack natural toxins and
allergens.
Advertisement
Researchers first became aware of this security force 10 years ago when
Richard Jorgensen and his colleagues at DNA Plant Technology in Oakland,
California, tried to deepen the colour of petunia petals by inserting extra
copies of a gene for an enzyme called chalcone synthase (CHS), which plants use
to make red, blue and purple pigments. To their astonishment, they found that
some modified plants made less pigment, producing white or partially white
flowers instead of the normal pink and violet. 鈥淲e were totally surprised,鈥� says
Jorgensen, who is now at the University of Arizona in Tucson. 鈥淏asically the
conversation was `what the heck did we do wrong?鈥� It took a while before we
realised that this was something interesting.鈥�
Not only had the gene they inserted failed to work, but the plant鈥檚 own
copies of the CHS gene had fallen silent as well. That meant the plant cells
must somehow have used information about the inserted gene to seek out and
inactivate similar genes. Eventually the scientists established that the cells
had made an RNA copy, or transcript, of the inserted gene鈥攖he first step
in the production of a gene鈥檚 protein product. Ordinarily, this messenger RNA
(mRNA) carries the genetic information out of the nucleus to the sites where
proteins are made. But in Jorgensen鈥檚 petunias, the cells somehow destroyed the
mRNA before it could be used to make the CHS protein.
Other researchers later dubbed this process 鈥減ost-transcriptional gene
silencing鈥� (PTGS), and it has since been found in bread mould, nematode worms,
fruit flies and protozoa. Geneticists now suspect that PTGS has, at one time or
another, operated in all organisms with nucleated cells, though vertebrates may
have lost it during their evolution (see 鈥淎 spanner in the works鈥�).
You might think a foreign gene would be hard to spot amid all the hustle and
bustle as the cell produces mRNA from its own genes in the course of its daily
business. In fact, the interloper usually leaves a telltale footprint鈥攁
double-stranded RNA molecule, something rarely seen in a normal cell.
Double-stranded RNA is to cells what the wolf鈥檚 snout and big teeth under
Grandma鈥檚 nightcap were to Little Red Riding Hood: a sure sign that something is
amiss.
Recall that DNA, the stuff of genes, is made of two complementary strands
joined into a double helix. When a legitimate gene becomes active, the cell
always makes an mRNA copy from just one of these two strands, the one identified
by the proper starting instructions. Since the other strand is not copied, the
mRNA never finds a complementary strand to pair with, so it remains
single-stranded.
Alarm bells
Many viruses, on the other hand, use RNA as their genetic material and create
double-stranded RNA temporarily as they create new copies of their genome. Even
viruses that use DNA as their genetic material often pack genes onto both
strands to save space. If such viruses transcribe overlapping genes from
complementary strands at the same time, the two mRNA molecules that result could
pair up just as their parent DNA strands did. Bingo: double-stranded RNA that
will set the cell鈥檚 alarm bells ringing.
Something similar happens with transposons. These bits of selfish DNA carry
the gene or genes needed to cut themselves out of the host chromosome and
reinsert elsewhere. Some can also make copies of themselves. Each of a
transposon鈥檚 genes is usually transcribed from only one DNA strand鈥攍ike
the cell鈥檚 own genes鈥攖o give single-stranded, 鈥渟ense鈥� mRNA. But
transposons insert themselves into DNA more or less at random. If a transposon
inserts next to, or into, a host gene, then some or all of the transposon genes
may get transcribed when the host tries to transcribe its own gene. And if this
accidental transcription reads the 鈥渨rong鈥� strand of the
transposon鈥攂ecause it happened to insert backwards compared with the host
DNA鈥攖he result will be 鈥渁ntisense鈥� RNA. This is complementary to the sense
RNA and binds to it easily, once again yielding double-stranded RNA (see
Diagram). The more successful the transposon, the more copies of its DNA insert
into the host and the more likely this becomes. Exactly the same thing happens
to the genes inserted by genetic engineers and the genes of viruses that insert
DNA into their host鈥檚 chromosomes, for example HIV.
Of course, the gene-silencing system may respond to other alarms as well.
鈥淓verybody believes that double-stranded RNA is a trigger,鈥� says Jorgensen. 鈥淏ut
there may be triggers and triggers.鈥� Sometimes it is enough to insert a single
extra copy of the CHS gene into petunias, if that copy makes lots of mRNA. Since
genes make up only a small fraction of the genome, it is unlikely that a single
inserted copy will land close enough to an existing gene to make antisense RNA,
so perhaps cells also induce gene silencing if they are swamped by a particular
mRNA, as would happen during successful viral infections.
Whatever the other possibilities, double-stranded RNA causes a massive
gene-silencing response. Even a few double-stranded molecules per cell
injected into nematode worms can silence perhaps thousands of copies of
single-stranded mRNA from the nematode鈥檚 own versions of the gene, Andrew Fire
and his colleagues at the Carnegie Institution of Washington showed in 1998
(Nature, vol 391, p 806). Somehow鈥攊n a way that no one
understands yet鈥攖he nematode鈥檚 cells are able to take up this RNA and
trigger silencing. Geneticists now use the technique to switch off particular
genes at will to study their function. The same approach works for
Drosophila and protozoa.
In plants, too, gene silencing breaks the bounds of individual cells. It can
spread from one part of a plant to another, and even from one plant to another
through a graft. Botanists have speculated that the invaded cell might run off
many short, single-stranded, antisense RNA copies using part of the
double-stranded RNA as a template. These short RNAs, they reasoned, could act as
a combination 鈥淲anted鈥� poster and death warrant, binding to any complementary
mRNA in the cell鈥攆oreign or native鈥攁nd marking it for destruction.
By travelling through the pores that link plant cells, these warrants could
spread silencing throughout the plant.
Puzzle pieces
In the past year, researchers working on fungi, plants and nematodes have
each confirmed parts of this speculation. The first piece of the puzzle turned
up last May in a study of bread mould with a mutation that disabled the
gene-silencing process. When researchers at the University of Rome sequenced the
mutant gene, they found it closely resembled a plant gene for an enzyme called
RNA-dependent RNA polymerase (RdRP), which makes RNA copies from an RNA template
(Nature, vol 399, p 166). This enzyme is exactly what鈥檚 needed to run
off the short antisense RNA death warrants predicted by the theory.
The second piece turned up in October, when David Baulcombe and Andrew
Hamilton of the John Innes Centre in Norwich showed that tobacco plants in the
throes of PTGS really do produce short fragments of RNA, about 25 nucleotides
long, which is complementary to mRNA from the gene being silenced
(快猫短视频, 6 November 1999, p 25).
They also showed that the spread of PTGS
through the plant parallels the spread of these fragments, which implies that
the fragments are indeed the death warrants responsible for transmitting PTGS
between cells.
Also in October, Ronald Plasterk of the Netherlands Cancer Institute in
Amsterdam supplied a third piece of the puzzle. Plasterk鈥檚 team studied
nematodes with defective PTGS and found that one of the mutations, in the
MUT7 gene, blocks the production of a protein that resembles an
RNA-digesting enzyme (Cell, vol 99, p 133). Since the mutant nematodes
are otherwise healthy, they don鈥檛 appear to need this protein for the cell鈥檚
run-of-the-mill housekeeping. This suggests that it is dedicated to digesting
mRNAs during silencing.
These puzzle pieces fit together nicely enough to suggest that the picture is
more or less the same in different organisms. 鈥淭hese phenomena look similar in
plants, fungi and nematodes,鈥� says Baulcombe. 鈥淥ur working hypothesis is that
they use similar mechanisms.鈥� If so, this implies that PTGS is an ancient
defence mechanism that arose many hundreds of millions of years ago in the
common ancestor of all these groups.
Gene silencing is so effective at crushing gene expression that every
successful virus in an organism with silencing must have evolved some way to
hide from it, escape from it or fight back. Many plant viruses, says Baulcombe,
have chosen the third option. Several research groups, including Baulcombe鈥檚,
have shown that a number of plant viruses produce proteins that suppress
silencing. Exactly how the suppression works no one knows, but different viruses
clearly target different parts of the system. Some viruses bring silencing to a
halt throughout the plant, while others only block its spread to new leaves.
Not surprisingly, some plants strike back. A group at the National University
of Singapore has shown that a species of tobacco possesses an
anti-anti-silencing mechanism. The plant鈥檚 cells recognise the protein that
cucumber mosaic virus uses to inhibit PTGS. They then commit suicide, depriving
the virus of its host and stopping the infection before it spreads.
Besides its effect on viruses, gene silencing also helps to keep transposons
in check鈥攅specially in reproductive cells, where they can do the most
damage. In nematode worms, for example, transposons are usually active in normal
cells but silent in the cells that make eggs and sperm. But in many mutants with
impaired silencing, the transposons become active in these cells as well,
researchers at the University of Massachusetts reported in October (
Cell, vol 99, p 123).
Gene silencing has also proved useful in genetic engineering of crops. In
fact, one silenced product鈥攁 canned tomato pur茅e鈥攈as already
reached the market. The pur茅e is made from tomatoes genetically
engineered by Zeneca to contain a short section of the gene for
polygalacturonase, an enzyme which digests the cell walls in ripening fruit.
This induces PTGS and silences the plant鈥檚 own polygalacturonase gene, producing
tomatoes that can stay on the vine longer, develop more flavour and eventually
yield better pur茅e.
According to Wolfgang Schuch, a researcher at Zeneca鈥檚 labs in Norwich, the
company is now testing this technology in other crops. Schuch declines to say
exactly what they hope to achieve, but there are obvious targets. Switch off the
right genes, for example, and crop plants will no longer make toxins or
allergens. Allergen-free peanuts must be on the biotech industry鈥檚 wish
list.
Then, of course, there are possibilities for creating new flower colours and
patterns鈥攁lthough it turns out that this is old news. In 1838, a century
and a half before Jorgensen鈥檚 experiment, breeders making crosses between
petunia species created flowers with a star pattern of white stripes. By
accidentally duplicating a CHS gene, and probably therefore overproducing CHS
mRNA, they too had induced gene silencing.
If gene silencing really is an ancient immune system, then our own ancestors
must have had all the equipment to carry it out. But although plenty of
researchers have looked, no one so far has demonstrated a clear-cut case of
post-transcriptional gene silencing in humans or other vertebrates. It may be
that we have lost the ability in favour of a more drastic response to the
double-stranded RNA that so often gives viruses away.
In mammalian cells, double-stranded RNA activates an enzyme called PKR, which
shuts down the cell鈥檚 protein synthesis machinery, slowing viral replication.
Sometimes it also induces the cell to commit suicide, depriving the virus of its
host. In addition, double-stranded RNA triggers the production of interferon,
which among other functions increases the production of PKR. This amplifies the
effect of even a whiff of double-stranded RNA.
The importance of PKR became even clearer last October, when Allan Lau and
his colleagues at San Francisco General Hospital and the University of
California at San Francisco showed that inhibiting PKR could turn a normally
short-term infection into a chronic one (Proceedings of the National
Academy of Sciences, vol 96, p 11 860). Deprived of PKR activity, cultured human
cells infected with a virus called EMCV fail to kill themselves, instead
releasing new virus particles. Many viruses, including HIV and influenza,
produce proteins that inhibit PKR activity. Perhaps, says Randal Kaufman, an
immunologist from the Howard Hughes Medical Institute and the University of
Michigan, drugs that boost PKR function could help fight such diseases.