Biology Reference
In-Depth Information
even though these compounds do not directly contact Phe201. Gln192 and Phe201
in PDE4D UCR2 are conserved across PDE4A-C, and indeed, in UCR2 from
Drosophila
(Table
1
). Despite the amino acid sequence conservation between
human and
Drosophila
UCR2,
Drosophila
PDE4 is not inhibited by rolipram
(Henkel-Tigges and Davis
1990
). The basis for this has yet to be resolved, although
it is unlikely to be a simple Phe-Tyr switch.
C. elegans
UCR2 has a conservative
substitution of Tyr at the hydrophobic docking residue equivalent to Phe201.
The UCR2 gating helix also affects the apparent
K
M
for cAMP. Human PDE4D
has a lower
K
M
for cAMP (1.5
m
M) as compared to PDE4B (7.7
m
M). Docking
studies suggest that closure of UCR2 over the active site may allow Tyr275 in
PDE4B to form a hydrogen bond to the 2
0
OH of cAMP or 5
0
AMP, and indeed
mutation of PDE4D Phe196 to Tyr raises the apparent
K
M
for cAMP to 7.5
m
Mas
predicted (Burgin et al.
2010
). Formation of a hydrogen bond between UCR2 and
the 2
0
OH of 5
0
AMP may affect
K
M
by stabilizing the product bound form of the
enzyme-substrate complex.
C. elegans
UCR2 also contains a Phe at the homolo-
gous position rather than a Tyr as in
Drosophila
, suggesting that there may have
been natural selection in
C. elegans
for a low
K
M
PDE4.
In addition to the UCR2 regulatory helix, we solved a cocrystal structure of
PMNPQ with a C-terminal
a
-helix closing across the active site (PDB ID: 3 G58),
Fig.
2
. Previous X-ray structures of PDE4B (1F0J and 1XM6) (Xu et al.
2000
; Card
et al.
2004
), a recently deposited structure (Kranz et al.
2009
), and a deposited but
unpublished structure (PDB ID: 3KKT), all show a C-terminal helix spanning
residues Asn422 to Phe434. C-terminal residues of PDE4 are important for protein
expression and therefore present in most N-terminally truncated constructs used to
crystallize the catalytic domains of PDE4B and PDE4D; however, these residues
are typically disordered in the crystal lattice and are therefore absent in most
models. The helices from the 1F0J and 1XM6 structures do not overlay and do
not fit into the same groove that we have observed with the UCR2 helix; however,
we speculated that this C-terminal helix could also come across and be visualized if
stabilized by an appropriate inhibitor. In our cocrystal structure with PMNPQ, the
PDE4D C-terminal helix was clearly defined in only one of the four monomers in
the asymmetric unit suggesting that the inhibitor only weakly interacts with this
C-terminal helix. The observed C-terminal helix does overlay with the C-terminal
helix modeled in 3KKT, and with the UCR2 helix in both PDE4D and PDE4B
(RMSD values of 0.90 and 0.76
˚
for c
a
-carbons) structures, and a conserved
phenylalanine protruded into the active site and stacked upon the aromatic group
of PMNPQ. The helix was also stabilized by a hydrophobic interaction between
the helix (Leu436) and the catalytic domain (Ile376/Met439); however, we do not
identify any potential hydrogen bond interactions between the helix and the cata-
lytic domain as for UCR2. The linker between the C-terminus of the catalytic
domain and the N-terminus of the helix were not visualized in the 1F0J and
1XM6 structures, and could not be modeled in our structure; however, the linker
region is visible in the 3KKT structure showing an intramolecular interaction.
Unlike LR2 between UCR2 and the catalytic domain, which is relatively long
(
40 amino acids), the linking portion between the catalytic domain and the
>
