Calcium Biochemistry
stimulation leads to long-term potentia-
tion (LTP) and low-frequency stimulation
to long-term depression (LTD). These pro-
cesses have been implied as important
models for studying spatial learning and
memory storage. It was also reported that
both processes are induced by an increase
in intracellular Ca
in the postsynap-
tic neuron concommitant with a rise in
CaMKII activity and phosphorylation of
synapsin I.
An important link has been made be-
tween LTP and some forms of learning,
and its dependence on CaMKII by gene-
targeted disruption of CaMKII
group of Tonegawa. These studies pro-
vided evidence that those mice that de-
veloped normally were not only impaired
in spatial learning but also that induction
of LTP was blocked in hippocampal slices
from those mice.
Recently, De Koninck and Schulman
provided direct evidence that CaMKII in-
deed can decode the frequency of Ca
pulses. Owing to the complex activation
pattern of CaMKII by CaM, autophospho-
rylation, and subsequent CaM trapping,
the enzyme can become autonomous.
Autophosphorylation is an intersubunit
reaction between proximate subunits in
which CaM not only activates the ‘‘ki-
nase’’ subunit but also presents the ‘‘sub-
strate’’ subunit for phosphorylation. Sub-
sequently, the kinase is transformed into
the ‘‘trapped’’ state, that is, a cooperative,
positive feedback loop resulting in a short
molecular ‘‘memory’’ that could enable the
enzyme to respond to digital and cyclic ac-
tivation associated with Ca
In simulation calculations, it was predicted
that repetitive Ca
pulses lead to recruit-
ment of CaM, autophosphorylation and
trapping of CaM, establishing a threshold
frequency at which the activity of the en-
zyme is sustained. These predictions were
exactly conFrmed by the experiments of De
Koninck and Schulman who could demon-
strate that, independent of the Ca
pulse duration, the autonomous activation
of CaMKII increased steeply as a function
of frequency. On the other hand, once a
threshold value was achieved, it was pos-
sible to maintain the response level with
signals of substantially lower frequency.
CaMKI and IV
In contrast to CaMKII,
which is a multimeric protein, CaMKI
and CaMKIV are monomeric enzymes.
expressed, whereas the tissue distribution
of CaMKIV is restricted to nervous tissues,
the thymus, particularly T lymphocytes,
the bone marrow, keratinocytes, testis,
and ovary. Both enzymes are activated not
only by autophosphorylation but also by
another CaM-dependent protein kinase,
the CaM kinase kinase (CaMKK).
The Crystal Structure of CaMKI
In 1996,
the group of Kuriyan reported the crystal
structure of CaMKI in the autoinhibited
state, that is, in the absence of CaM. This
is the Frst three-dimensional structure
of a calmodulin-dependent enzyme deter-
mined at high resolution, as mentioned
before. The crystal structure of CaMKI
provided evidence of signiFcant contacts
between the autoinhibitory sequence and
the catalytic core of the kinase, thereby
supporting the pseudosubstrate model for
activation of calmodulin-dependent pro-
tein kinases. Another important result was
the Fnding that Trp303 corresponding to
the Frst anchoring residue of the CaM
binding domain (see Table 1) lies on the
surface of the protein. This arrangement
makes it very likely that CaM could Frst
bind to the exposed Trp303 and subse-
quently release the
-strands of the ATP
previous page 824 Encyclopedia of Molecular Cell Biology and Molecular Medicine read online next page 826 Encyclopedia of Molecular Cell Biology and Molecular Medicine read online Home Toggle text on/off