APLYSIA. Sea snails. They are not quite what you鈥檇 expect. They are
big鈥攁bout three times the size of a garden snail鈥攆at, and ugly, and
looking at them you wonder what on earth they would possibly need to remember.
Yet Eric Kandel at Columbia University in New York has made a career of
learning
how sea snails learn and how they remember what they have learnt.
For decades, neuroscience has wrestled with the problem of how the brain
converts a short-term memory鈥攖hat is one that lasts a few
minutes鈥攊nto one that lasts a lifetime. Now, Kandel, one of the most
celebrated and controversial figures in that arena, thinks that his team鈥檚
work,
in combination with some key experiments from other laboratories, is within a
hair鈥檚 breadth of finding the answer. Those findings, Kandel says, mean that a
new type of memory pill is just around the corner.
At the heart of his vision are two proteins鈥� CREB1 and CREB2.
According
to Kandel, our ability to store information over the long-term depends on the
relative amounts of the two proteins that are produced in the brain. Taking
their cue from some of Kandel鈥檚 early findings in sea snails, other researchers
have shown that versions of CREB1 and CREB2 are vital for memory in flies and,
even more significantly, mice. That suggests that if Kandel鈥檚 hypothesis that
the CREB proteins make up a sort of master memory switch is correct, it is also
likely holds true for other mammals, including humans.
Advertisement
Kandel, who was born in Vienna, but has lived in New York long enough
to have
a bit of a Brooklyn accent, likes to speculate that it is CREB proteins that
give Talmudic scholars their edge. Some scholars have an amazing ability to
recall verbatim obscure portions of the Talmud, the gigantic body of
Jewish law.
Kandel says that the Talmudic scholars may have an excess of CREB1, or a
particularly potent type of CREB1, while absent-minded professor types
will have
more of the CREB2 protein. 鈥淚t consoles me that [the Talmudic scholars] are not
very happy people,鈥� says Kandel. 鈥淭hey have fantastic memories, but are not
profound. They have no original insights.鈥�
But Kandel does. He is often described as the father of learning and memory,
and, on occasion, as arrogant. 鈥淗e has a global vision of what memory could be
at a cellular level,鈥� says neurogeneticist Tim Tully of Cold Spring Harbor
Laboratory on Long Island, New York. 鈥淏ut because of that vision he has
sometimes ignored discrepancies in the actual data in favour of the ideal
model.
Hard core scientists bristle at that.鈥�
Back in his laboratory, Kandel, like his sea snails, is not quite what you鈥檇
expect, either. He is very charming. His labs, which are squeezed into a sixth
floor annex of Columbia鈥檚 College of Physicians and Surgeons in Harlem, are
crowded with snail tanks, refrigerators, centrifuges, electrophoresis
apparatus,
not to mention a small, ever-changing army of graduate students and postdocs.
Sitting in the middle of it all, Kandel describes with great
animation鈥攑assion even鈥攈ow his recent discoveries have meshed
perfectly with those from other laboratories. He knows that finding the key to
long-term memory storage would be one of the discoveries of the century in
memory research.
Kandel鈥檚 sea snail doesn鈥檛 have a brain to speak off, but it does have a
memory. It鈥檚 stored in a simple nervous system of about 20 000 nerve
cells. That
makes it a far easier creature to study than the human with its many
billions of
neurons. Kandel鈥檚 research鈥攁nd reputation鈥攊s built on the
assumption
that the neurological basis of a sea snail鈥檚 memory will also explain why
we may
remember a popular song forever, but soon forget who ruled France from 1270 to
1285*.
Imagine for a moment that you are a sea snail. If someone pinches your tail
you will vigorously withdraw your gill, or respiratory organ, much like a
person
would blink if something flew at their face. If the stimulus is repeated, you
will become sensitised so that you will also withdraw your gill in response to
the 鈥渨eak鈥� stimuli of having someone stroke your siphon, the small fleshy spout
you use to get rid of sea water.
The number of pinches or 鈥渘oxious鈥� stimuli determine how long a sea snail
remembers to withdraw its gills in response to siphon stroking. 鈥淚f we give one
noxious stimulus, the Aplysia will withdraw its gills in response to a weak
stimulus and that memory will last for a few minutes,鈥� says Kandel. 鈥淏ut if we
give it five noxious stimuli, it will remember to withdraw its gills to a weak
stimulus for several days.鈥�
In the mid-1980s, Kandel鈥檚 team discovered that for memories to be
stored for
any time at all, the sea snail鈥檚 neurons must first release a neurotransmitter
called serotonin. The serotonin binds to a protein on the surface of the
sensory
neurons, starting a relay race that calls into play a sequence of different
chemicals and proteins. The duration of the final memory depends on which path
the relay race takes within the sensory neuron, which depends on the number of
pulses of serotonin released, which, in turn, depends on the type of
stimuli and
training the animal receives.
Indelible ties
With short-term memories, serotonin turns on the production of a chemical
called cAMP which binds to a protein called cAMP-dependent protein kinase A or
PKA. PKA, in turn, modifies proteins in the cytoplasm of the neurons thereby
strengthening the synapses to create a short-lived memory.
But in order for that memory to become long-term and be indelibly
etched into
the mind, PKA must enter the nucleus of the neurons and trigger a sequence of
events that turns on genes, and鈥攗ltimately鈥攍ead to the synthesis of
new proteins. Most researchers in the field think that these proteins probably
have their effect by strengthening synapses in part of a process called
long-term potentiation.
In the sea snails that have had their tails pinched and their siphons
stroked, the synapses are strengthened between the sensory neurons of the
siphon
and tail and the motor neurons of the gill. According to Kandel, it is at the
point in the relay race when passing on the baton depends on turning on genes
and synthesising new proteins, that the CREB proteins come into play.
CREB proteins were discovered almost ten years ago. They turn on genes in
cells throughout the body, and are themselves turned on by PKA and by cAMP. To
Kandel, they fitted exactly the sort of profile that a protein that forges
long-lasting memories should have.
In 1990, Pramod Dash, one of Kandel鈥檚 postdocs at the time, made a
breakthrough. He measured the electrical activity in sensory and motor neurons
cut out of sea snails to get a handle on the strength of the synapses that
joined them. Next he added serotonin to the neurons every 20 minutes to
recreate
the chemical conditions that lead to different length memories in living
snails.
A single dose of serotonin, designed to mimic the triggering of short-term
memories, strengthened the synapse for a few minutes, whereas five or more
pulses, designed to mimic the triggering of long-term memories,
strengthened the
synapse for several days.
鈥淒ash reasoned that if CREB1 is involved in long-term memory storage,
then if
you block CREB1, you should also block long-term memory,鈥� says Kandel. And sure
enough, when Dash injected a chemical that prevents CREB1 from binding to its
DNA target in the sensory neurons, the serotonin pulses failed to trigger
synapse strengthening, the test-tube stand-in for memory, for longer than one
hour.
Kandel doesn鈥檛 yet known which genes CREB1 turns on, although he is certain
that the protein is capable of activating genes in the snail鈥檚 nervous system.
In 1993, he and his graduate student Bong Kiung Kaang showed that in cultured
neurons, CREB1 turned on 鈥渞eporter鈥� genes that had been artificially
attached to
pieces of DNA that the protein recognises. What is more, CREB1 only did this in
the presence of PKA, the enzyme that is essential for the formation of long and
short-term memories.
By themselves, Kandel鈥檚 experiments would probably not have convinced the
sceptics who like to see biochemical events directly influence memory in living
animals. But by identifying CREB1 as a key player in long-lasting memory,
Kandel, Dash, and Kaang set the stage for some now famous experiments in fruit
flies and mice.
Over at Cold Spring Harbor, Tully had been doggedly teaching fruit flies a
few tricks. By giving the flies electric shocks at the same time as they were
exposed to the smell of octanol it smells like aniseed) and methylcyclohexanol
(it smells like sweaty feet) in 10 training sessions 15 minutes apart,
Tully had
persuaded fruit flies to commit the odours to their long-term memories, and to
avoid them.
Then, Tully and his colleague Jerry Yin, together with Chip Quinn and other
researchers at the Massachusetts Institute of Technology in Cambridge, tracked
down the equivalent of the sea snail CREB1 gene in the fruit flies, and called
it the dCREB2 gene. The fruit-fly gene, it turned out, codes for two different
versions of the protein. When Tully and Yin genetically-engineered fruit
flies to overproduce one version of the dCREB2 protein, the creatures developed
long-term memories after only one training session, rather than the 10 training
sessions that were usually required. 鈥淭he equivalent of producing photographic
memory,鈥� says Tully.
Memory testing
When they engineered the flies to overproduce the other version of the
protein it blocked long-term memory. 鈥淏ut it has no effect on early or
short-term memory,鈥� he says. Tully鈥檚 experiments showed that in live flies the
CREB proteins are superfluous to short-term memory, but vital for converting it
to a long-term memory. And that the protein comes in one form that inhibits
long-term memory and one that enhances it.
While Tully and Yin were busy with their supersmart and superdumb flies,
other researchers at Cold Spring Harbor including Alcino Silva and Roussoudan
Bourtchuladze were creating mutant mice that produce far less of the
memory-enhancing form of CREB. They tested their memories by seeing whether they
could still locate a submerged platform after the water had been made opaque
with white paint.
鈥淭he mutant mice remember just as well as [normal mice] if we test them
within one hour of their training session,鈥� says Silva. 鈥淚f we test them later,
they have completely forgotten.鈥�
Now, the research has come full circle. At the end of last year,
Kandel, this
time with his postdocs Dusan Bartsch and Mirella Ghirardi, discovered an
inhibitor version of CREB in the sea snail, and called it CREB2. In the
dissected sea snail neurons, CREB2 blocks CREB1鈥檚 ability to turn on genes, and
to strengthen synapses after five or more serotonin pulses.
All these experiments make the bedrock for Kandel鈥檚 new theory of how
memories are etched into the brain. In Kandel鈥檚 vision, which Tully shares,
CREB1 and CREB2 in the human brain will bind to each other or to
themselves; the
relative level of CREB1 to CREB2 decides whether or not a person has a good
long-term memory. If the person has more CREB1 around, the new proteins needed
for long-term storage will be manufactured. But if a person has an excess of
CREB2, that protein will mop up all the CREB1 and stop it from doing its job,
leaving the person with a sieve-like memory.
With the Talmud scholars in mind, Kandel even has an idea why the brain
would
need a forgetfulness molecule like CREB2 in the first place: storing too much
trivia could clutter the mind and perhaps cramp creativity. Kandel admits,
however, that the CREB theory will undoubtedly get more complicated. 鈥淲e are
likely to find more and more CREB activators and CREB inhibitors that interact
to control different aspects of memory,鈥� he says. Indeed, the Kandel team has
already found two new proteins, C/EBP and ATF4, with CREB1-like activity.
And many questions remain. For instance, 鈥淐REB1 and CREB2 may control
whether
the bricks are made that are needed for memory storage,鈥� says Silva. 鈥淏ut they
are either there and you can store memories or they are missing and you can鈥檛
store memory. What controls which memories get stored remains a big
尘测蝉迟别谤测.鈥�
Meanwhile, although it is unlikely that drugs will be able to turn a
forgetful person into a Talmud scholar, Kandel envisions a time when they could
help cure more severe memory defects. 鈥淲e are trying to develop drugs based on
the cAMP signalling system,鈥� says Kandel. 鈥淐REB2 seems like a good target.鈥�
Columbia University may form a company to market whatever his lab develops, he
says.
Kandel refuses to say how far his team has got, but he does caution that a
CREB-based memory drug won鈥檛 help in retrieving long-term memories stored and
lost. Rather it will increase the number of memories that can be indelibly
etched on the brain, a treatment likely to be most helpful to people with
age-related memory loss. 鈥淲hen you get older, you forget things,鈥� says
66-year-old Kandel. 鈥淚t could help people like that.鈥�
Meanwhile, you can be sure that Kandel鈥檚 critics and admirers are out
in full
force to test the validity of Kandel鈥檚 current model of CREB1 and CREB2.
鈥淜andel is controversial because he is pushy and has very structured
ideas of
how things ought to work,鈥� says Quinn, 鈥淏ut the field of behaviour is full of
funny little people with funny little animals doing funny little things.
Without
any sort of structural framework, or model, of how things work, these funny
observations don鈥檛 mean much. Kandel is not always right, but he drives the
field by providing an intellectual structure that can then be tested.鈥� To
Tully,
it鈥檚 all par for the course. 鈥淪omeone presents a model until there is enough
evidence to either support or refute it. If you present your model first,
测辞耻鈥檙别
king of the mountain until someone else comes along to knock you off.鈥�
And for several decades, Kandel has been king.
*Phillip III, The Bold