You don't hear the term "junk DNA" much anymore, and for good
reason. Now we know that stretches of the genome that don't code
for proteins can affect how genes are read in many different ways.
A new study in Nature suggests that an RNA molecule made from such
a stretch may play an important role in learning and memory.
The RNA molecule in question is a microRNA (miRNA). MiRNAs are
small stretches of RNA, 18 to 25 nucleotides long, that regulate
gene expression by binding to regions on target messenger RNAs
(mRNAs) with complementary sequences. Once they bind, they either
bring about the mRNA's cleavage or inhibit the cell's ability to "translate" it — that
is, to read it and make protein based on its sequence.
Biochemical and genetic studies have revealed important functions
for miRNAs in many areas of cell function, including differentiation,
apoptosis and metabolism. A team of researchers at Children's Hospital
Boston, supported by grants from NINDS and NICHD among others,
set out to see whether miRNAs might play a role in regulating how
nerve cells in the brain connect and communicate.
Nerve cells communicate through points of contact called synapses.
By creating, strengthening and weakening these connections, nerve
cells lay the networks for learning and memory. Since some mRNAs
in nerve cells seem to be transported to sites near synapses, the
researchers reasoned that local regulation of these transported
mRNAs by miRNAs might play a key role in synapse development.
The researchers identified a miRNA called miR-134 that concentrates
near synapses in dendrites — the branch-like nerve cell extensions
that receive signals from other nerve cells. Using rat cells in
culture, they discovered that miR-134 can reduce the size of dendritic
spines, the specialized sites on dendrites at the receiving ends
of synapses.
To figure out how miR-134 affects dendritic spines, the team searched
mRNAs for potential miR-134 binding sites. This led them to a protein
called Limk1, which controls dendritic spine development. MiR-134
inhibits translation of the Limk1 mRNA. It can thus regulate spine
size by affecting how much Limk1 protein the dendrites can make.
The researchers then found that exposing nerve cells to brain-derived
neurotrophic factor (BDNF), which stimulates mRNA translation at
synapses, relieves miR-134's inhibition of Limk1 translation. So
to put this whole picture together, the researchers hypothesize
that, in cells, miR-134 might bind to Limk1 mRNA to keep it in
a dormant state while it is being transported out to synaptic sites.
At synapses, BDNF can prompt dendritic spine development where
it's needed by selectively stimulating mRNA translation.
The authors found several additional mRNAs that may be targets
for miR-134, so they suspect it might regulate a whole set of genes
involved in synapse formation and development. There are likely
other miRNAs that are involved as well. The researchers speculate
that miRNAs may even act locally at individual synapses to form
and refine connections. That could potentially explain at the cellular
level how organisms learn and respond to their environment.
The next challenge is to identify all the miRNAs in dendrites
and all their target mRNAs. The ultimate goal will be to find out
how they all work together to selectively create, strengthen and
weaken the connections between synapses. 
