doi:10.1006/jmbi.2001.5061 available online at http://www.idealibrary.com on
J. Mol. Biol. (2001) 314, 279±291
High Resolution Crystal Structures of the Catalytic
Domain of Human Phenylalanine Hydroxylase in its
Catalytically Active Fe(II) Form and Binary Complex
with Tetrahydrobiopterin
Ole Andreas Andersen1, Torgeir Flatmark2 and Edward Hough1*
1
Department of Chemistry
University of Tromsù
N-9037 Tromsù, Norway
2
Department of Biochemistry
and Molecular Biology
University of Bergen
AÊrstadveien 19, N-5009 Bergen
Norway
The crystal structures of the catalytic domain (N1-102/C428-452) of
human phenylalanine hydroxylase (hPheOH) in its catalytically competent Fe(II) form and binary complex with the reduced pterin cofactor
6(R)-L-erythro-5,6,7,8-tetrahydrobiopterin (BH4) have been determined to
Ê , respectively. When compared with the structures reported
1.7 and 1.5 A
for various catalytically inactive Fe(III) forms, several important differences have been observed, notably at the active site. Thus, the nonliganded hPheOH-Fe(II) structure revealed well de®ned electron density
for only one of the three water molecules reported to be coordinated to
the iron in the high-spin Fe(III) form, as well as poor electron density for
parts of the coordinating side-chain of Glu330. The reduced cofactor
(BH4), which adopts the expected half-semi chair conformation, is bound
in the second coordination sphere of the catalytic iron with a C4a-iron
Ê . BH4 binds at the same site as L-erythro-7,8-dihydrobiopdistance of 5.9 A
terin (BH2) in the binary hPheOH-Fe(III) BH2 complex forming an aromatic p-stacking interaction with Phe254 and a network of hydrogen
bonds. However, compared to that structure the pterin ring is displaced
Ê and rotated about 10 , and the torsion angle between the
about 0.5 A
hydroxyl groups of the cofactor in the dihydroxypropyl side-chain has
Ê)
changed by 120 enabling O20 to make a strong hydrogen bond (2.4 A
with the side-chain oxygen of Ser251. Carbon atoms in the dihydroxypropyl side-chain make several hydrophobic contacts with the protein. The
iron is six-coordinated in the binary complex, but the overall coordination geometry is slightly different from that of the Fe(III) form. Most
important was the ®nding that the binding of BH4 causes the Glu330
ligand to change its coordination to the iron when comparing with nonliganded hPheOH-Fe(III) and the binary hPheOH-Fe(III) BH2 complex.
# 2001 Academic Press
*Corresponding author
Keywords: phenylalanine hydroxylase; tetrahydrobiopterin; ferrous iron;
iron ligands; protein crystallography
Introduction
Abbreviations used: BH2, L-erythro-7,8dihydrobiopterin; BH4, 6(R)-L-erythro-5,6,7,8tetrahydrobiopterin; DTN, dithionite; DTT,
dithiothreitol; hPheOH, human phenylalanine
hydroxylase; L-Phe, L-phenylalanine; PheOH,
phenylalanine hydroxylase; P-rPheOH, phosphorylated
rat phenylalanine hydroxylase; PKU, phenylketonuria;
qBH2, quinonid dihydropterin; rPheOH, rat
phenylalanine hydroxylase; rTyrOH, rat tyrosine
hydroxylase; TyrOH, tyrosine hydroxylase; wt, wildtype.
E-mail address of the corresponding author:
[email protected]
0022-2836/01/020279±13 $35.00/0
The non-heme iron enzyme phenylalanine
hydroxylase (PheOH, phenylalanine 4-monooxygenase, EC. 1.14.16.1) catalyses the hydroxylation
of the essential aromatic amino acid L-phenylalanine (L-Phe) to L-tyrosine, which is the rate-limiting step in the complete catabolism of L-Phe.1 The
enzyme belongs to the family of aromatic amino
acid hydroxylases with three structurally and functionally closely related enzymes that require the
cofactor 6(R)-L-erythro-5,6,7,8-tetrahydrobiopterin
(BH4) (Figure 1) and dioxygen.3 In the catalytic
# 2001 Academic Press
280
Figure 1. Structure of natural tetrahydrobiopterin
6(R)-(6-L-erythro-5,6,7,8-10 ,20 -dihydroxypropyl)-5,6,7,8-tetrahydropterin showing the non-planarity of the pyrazine
ring. Hydrogen atoms are shown as white spheres. The
Figure was prepared using MOLSCRIPT.2
reaction the cofactor is oxidized to pterin 4a-hydroxytetrahydrobiopterin. In vivo the latter is then converted to a quinonoid dihydropterin (qBH2) by
pterin 4a-carbinolamine dehydratase4 from which
BH4 is regenerated by a NAD(P)H-dependent
dihydropteridine reductase5; quinonoid dihydropterin can alternatively form L-erythro-7,8-dihydrobiopterin (BH2) by isomerization. PheOH is a
cytosolic enzyme mainly located in the liver (hepatocytes) which plays an important role in mammalian metabolism by initiating detoxi®cation of
high levels of L-Phe. A de®ciency in human
phenylalanine hydroxylase (hPheOH) is linked to
the autosomal recessive disease phenylketonuria
(PKU)6, and today more than 400 mutations are
known at the
hPheOH
locus7 (http://
www.mcgill.ca/pahdb). Although BH4 is the physiological electron donor in the hydroxylation reaction it is also a potent negative effector of the
activation of the enzyme by L-Phe as well as of the
phosphorylation of Ser16 by cyclic AMP-dependent protein kinase.8,9 The oxidized cofactor (BH2)
also inhibits the activation by L-Phe, but at higher
concentrations than BH4. The dihydroxypropyl
side-chain of BH4/BH2 has been found to be essential for this inhibitory effect.10 ± 12 Substantial progress in elucidating the structure and function of
the mammalian aromatic amino acid hydroxylases3
has been made in the past decade. Crystallographic
studies of PheOH and tyrosine hydroxylase
(TyrOH) have revealed that they have a common
tetrameric architecture and that each of the chains
fold into three domains, i.e. a regulatory, catalytic
and tetramerization domain. Crystal structures
have been determined for dimeric hPheOH catalyÊ resolution13 (PDB 1PAH), tetratic domain at 2.0 A
meric hPheOH catalytic plus tetramerization
Ê resolution14 (PDB 2PAH),
domains at 3.1 A
X-ray Studies of Phenylalanine Hydroxylase
hPheOH catalytic domain with bound catechol
Ê resolution15 (PDB 3PAH,
inhibitors at 2.0-2.15 A
4PAH, 5PAH and 6PAH), hPheOH catalytic
Ê resolution10
domain in complex with BH2 at 2.0 A
(PDB 1DMW) as well as dimeric rat PheOH regulatory plus catalytic domains in phosphorylated (PÊ resolution) and non-phosphoryrPheOH) (2.2 A
Ê resolution) states16 (PDB
lated (rPheOH) (2.6 A
respectively 1PHZ and 2PHM). Tetrameric rat
TyrOH (rTyrOH) catalytic plus tetramerization
Ê
domains have been determined to 2.3 A
17
resolution (PDB 1TOH) as well as its complex
Ê resolution18 (PDB 2TOH).
with BH2, also to 2.3 A
Although these structures together with sitedirected mutagenesis have signi®cantly improved
our understanding of the catalytic and regulatory
properties of the enzyme, they have given only
partial information on the catalytic mechanism
since all the structures relate to the catalytically
inactive Fe(III) form of the enzyme and the binary
complex with the inactive BH2 pterin cofactor. In
the present study, therefore, we have determined
the high-resolution X-ray structures of anaerobic
crystals of the catalytically competent reduced
form of the dimeric ``catalytic domain'' and its
complex with the catalytically active form of the
pterin cofactor, BH4. Our structures differ from the
previously determined structures in the overall
iron-coordination geometry and conformation of
the cofactor bound at the active site.
Results
Overall structure of the reduced catalytic
domain of hPheOH, hPheOH-Fe(II)
The reduced dimeric truncated form (residues
103-427) of recombinant hPheOH crystallized in
Ê , b ˆ 108.4 A
Ê,
space group C2221 (a ˆ 66.3 A
Ê
c ˆ 124.0 A) with one monomer per asymmetric
unit and a solvent content of 57 %. The ®nal model
includes 307 residues, 278 water molecules and the
active site iron in the Fe(II) form. That the iron is
indeed reduced is supported by three criteria: (i)
At the selected anaerobic conditions the active site
iron (midpoint potential of ‡207 10 mV for wildtype hPheOH (wt-hPheOH))19 is ef®ciently reduced
by the high concentration of dithiothreitol (DTT)
(midpoint potential of ÿ330 mV) and sodium
dithionite (DTN) (midpoint potential ÿ1120 mV);
(ii) the active site structure is slightly different
from that of the Fe(III) structure including coordinating water molecules (Table 1); (iii) in the binary
complex the autooxidizable pterin cofactor (BH4)
bound at the active site is preserved in its structurally characteristic tetrahydro form (see below and
Discussion) and the reduction potential of qBH2
(‡174mV) is below that for Fe(III). The hPheOHFe(II) crystal was crystallized similarly, with even
higher concentration of reducing agent, as the
binary hPheOH-Fe(II) BH4 complex. The structure
has been re®ned to a working R-factor of 19.7 %
and a free R-factor 25.2 % for all data from 10 to
Table 1. Comparison of Fe-ligand distances and ligand B-factors
hPheOH-Fe(II) BH4
Ligand
His285Ne
His290Ne
Glu330Oe
Wat1
Wat2
Wat3
hPheOH-Fe(III)BH10
2
hPheOH-Fe(II)
hPheOH-Fe(III)13
P-rPheOH- Fe(III)16
rTyrOH- Fe(III)17
rTyrOH-Fe(III) BH18
2
Distance
Ê)
(A
B-factor
Ê 2)
(A
Distance
Ê)
(A
B-factor
Ê 2)
(A
Distance
Ê)
(A
B-factor
Ê 2)
(A
Distance
Ê)
(A
B-factor
Ê 2)
(A
Distance
Ê)
(A
B-factor
Ê 2)
(A
Distance
Ê)
(A
B-factor
Ê 2)
(A
Distance
Ê)
(A
B-factor
Ê 2)
(A
2.1
2.1
2.1
2.4
2.1
2.4
16.7
16.3
23.1
26.1
20.5
20.0
2.2
2.1
2.0
2.6
2.2
2.3
25.0
22.6
32.5
47.5
34.0
25.5
2.1
2.1
2.4
26.1
24.9
30.5
29.7
22.1
36.8
2.2
2.4
2.1
45.1
33.0
43.4
2.1
2.2
2.0
43.4
39.2
37.4
36.3
21.4
23.5
26.3
41.5
34.6
20.0
2.4
2.6
2.7
2.6
2.2
2.2
2.1
2.3
2.3
2.3
2.8
25.8
2.3
2.4
45.2
56.7
2.0
2.0
58.5
35.5
Ligand numbers for TyrOH are His331, His336 and Glu376, respectively.
282
Ê resolution, and there are no outliers in the
1.7 A
Ramachandran plot. The overall fold is very similar to the corresponding high-spin (S ˆ 5/2) Fe(III)
form of the enzyme.13 However, some interesting
differences were observed at the active site where
signi®cant electron density was found for only one
of the three water molecules coordinated to the
iron in the Fe(III) form (see also below for the
binary complex) as well as weak and poorly
de®ned electron density for Cg and the carboxyl
group of Glu330 (Figure 2(a)). The distinct metal
coordination environment of the Fe(II) form does
not extend into the second sphere hydrogen bonding pattern. The liganded Oe of Glu330, which in
Ê compared to
our structure is displaced 1.2 A
hPheOH-Fe(III), and Wat3 have slightly longer
bonds to the iron (Table 1) than in the Fe(III) form.
Weak omit density was observed for Wat1 and
Wat2 during re®nement (Figure 2(a)), but inclusion
of these two water molecules gave 2Fo ÿ Fc density
X-ray Studies of Phenylalanine Hydroxylase
Ê 2,
at the noise level and B-factors of 66.2 and 37.8 A
respectively. Thus, only Wat3 was included in the
®nal model. Similar results were obtained with a
Ê data set from a crystal that was produced
2.0 A
anaerobically by using 30 mM DTN as the reducing agent (data not shown). Several data sets on
crystals without reducing agent all revealed strong
electron density for all three coordinated water
molecules (data not shown) as previously
observed.13
Overall structure of the binary hPheOHFe(II) BH4 complex
The truncated form (residues 103-427) of the
reduced hPheOH co-crystallized with BH4 in space
Ê , b ˆ 108.5 A
Ê , c ˆ 124.4 A
Ê)
group C2221 (a ˆ 66.4 A
with one monomer per asymmetric unit and a solvent content of 57 %. The rod-shaped crystals
formed were slightly different from those without
Figure 2. Stereo picture of the electron density at the iron active site of hPheOH-Fe(II) (a) and hPheOH-Fe(II) BH4
(b). Blue electron density is from s-weighted 2Fo ÿ Fc maps at 1.7 s. No density appears for Wat1 and Wat2 in the
hPheOH-Fe(II) structure, while Glu330 is in a different conformation in the hPheOH-Fe(II) BH4 structure. Red omit
electron density is from s-weighted Fo ÿ Fc maps at 4.5 s omitting water ligands and the side-chain of Glu330. The
Figure was produced using BOBSCRIPT.20
283
X-ray Studies of Phenylalanine Hydroxylase
bound cofactor. The ®nal model includes 307 residues, 393 water molecules, the active site iron in
the ferrous form (see above) and the cofactor in its
reduced form. It has been re®ned to a working
R-factor of 15.7 % and a free R-factor 20.3 % for all
Ê resolution, and there are no
data from 10 to 1.5 A
outliers in the Ramachandran plot. The overall fold
is very similar to the corresponding ligand-free
form of the enzyme13 and BH4 binds in the second
coordination sphere of the iron (Figure 3). In
marked contrast to the ligand-free Fe(II) form, the
hPheOH-Fe(II) BH4 complex has well de®ned electron density for Glu330 and for all three water
molecules (Wat1-3; Figure 2(b)) previously shown
to be coordinated to the iron in the Fe(III) form.
The iron coordinating Wat1 is hydrogen bonded to
BH4 O4, the non-coordinating Glu330 Oe and
Tyr325 OZ in a tetrahedral fashion (Figure 4). Wat2
forms a water-mediated hydrogen bond to BH4
N5, while Wat3 forms hydrogen bonds with
Glu286 Oe2 and BH4 O4 (Figure 4). All the atoms
of the cofactor including the dihydroxypropyl sidechain have well de®ned electron densities
(Figure 5(a) and (b)). The pterin ring is approxiÊ
mately planar except for C6, which is about 0.6 A
out of the plane, con®rming that the bound cofactor has been successfully kept in the reduced
state22,23 since the oxidized form, BH2, is strictly
planar.24 Superposition of the structure on the
hPheOH-Fe(III) BH2 complex10 (PDB 1DMW)
shows that the reduced cofactor is displaced about
Ê in the direction away from Ser251, and that
0.5 A
the pterin ring is rotated about 10 (along the C4aC8a bond) with the pyrimidine ring rotated
towards Phe254. Furthermore, the dihydroxypropyl group has a completely different conformation.
The torsion angle between the hydroxyl groups in
the binary hPheOH-Fe(III)BH2 complex is about
53 , versus ÿ65 in the binary hPheOH-Fe(II) BH4
complex and versus ÿ60 in hPheOH-Fe(III) BH2
as determined by NMR.25 This enables the BH4 O20
Ê ) with the
to make a strong hydrogen bond (2.4 A
Figure 3. MOLSCRIPT2 Figure of the catalytic domain
showing the binding of BH4 in the second coordination
sphere of the iron. Iron ligands and BH4 are shown as
ball-and-stick models with larger spheres for BH4. Carbon atoms are black, nitrogen blue and oxygen red.
a-Helices are coloured green, b-strands are orange and
the iron is purple.
side-chain oxygen of Ser251, while O10 forms
water-mediated hydrogen bonds to residues
Ala322 and Glu330 (Table 2, Figure 4). In addition,
Glu330 adapts a completely different conformation
(Figure 6). The coordination of Glu330 is still
Ê , and the
monodentate, but Cg is displaced by 1.4 A
e
Ê
coordinating C is displaced about 1.0 A compared
to the crystal structures of non-liganded hPheOHFe(III) and the binary hPheOH-Fe(III) BH2 complex. Leu255 adapts the same conformation as in
the binary complex hPheOH-Fe(III) BH2, which is
different from the non-liganded structure, forming
two hydrophobic contacts with C30 at a distance of
Ê . C30 and C20 interacts with Ala322 Cb
3.7 A
Table 2. Potential hydrogen bonds and water-mediated hydrogen bonds between BH4 and hPheOH
Ê)
Distance (A
N1-Leu249 N
N2-Gly247 O
N3-Wat4
O4-Wat1a
O4-Wat3a
N5-Wat5
N8-Leu249 O
O10 -Wat6
O10 -Wat7
O20 -Ser251 Og
O20 -Wat8
O20 -Ala322 O
Wat4-Glu286 Oe1
3.1 (3.3)
2.9 (2.8)
2.7 (2.7)
2.7 (2.5)
2.7 (2.8)
2.7
2.9 (2.8)
3.1
2.7
2.4
3.3
(2.8)
2.5 (2.6)
Ê)
Distance (A
Wat4-His264 O
Wat1a-Glu330 Oe1
Wat1a-Tyr325 OZ
Wat3a-Glu286 Oe2
Wat5-Wat9
Wat5-Wat7
Wat5-Wat2a
Wat6-Ala322 O
Wat7-Glu330 Oe1
N8-Wat10
Wat10-Ser251 Og
Wat10-Leu249 O
Wat10-Wat11
2.7 (3.0)
2.5 (2.6)
2.5 (2.7)
2.7 (2.7)
2.9
2.9
2.9
2.9
2.6
(3.2)
(3.2)
(3.2)
(2.6)
Distances in parentheses are for the crystal structure of the binary hPheOH-Fe(III) BH2 complex.10 Glu330 Oe1 is the noncoordinating oxygen.
a
Water molecules liganded directly to iron.
284
X-ray Studies of Phenylalanine Hydroxylase
Figure 4. Schematic diagram of BH4-protein interactions. The BH4 molecule is shown with purple bonds, nitrogen
atoms are blue, oxygen atoms red, water oxygens green, carbon atoms black and the iron is yellow. The Figure was
produced using LIGPLOT,21 and edited using CorelDRAW 9.0.
Ê . Leu248 is in a
(Figure 4) at distances of 3.8 A
different conformation to that in the binary
Ê from the pterhPheOH-Fe(III) BH2, with Cg1 3.7 A
in ring. Some additional conformational changes
are observed upon binding of BH4 when compared
to the non-liganded Fe(III) structure.13 Residues
245-250 are closer to the iron with a maximum disÊ for Gly247 Ca. The
placement (main chain) of 1.5 A
380's loop13 is also slightly displaced with a maxiÊ.
mum movement at Thr378 Ca of 1.0 A
Many of the BH4 contacts with the enzyme are,
however, similar to those reported for the binary
hPheOH-Fe(III) BH2 structure, e.g. the hydrogen
bonding network of the pterin ring as described in
Figure 4 and Table 2, and the p-stacking interactions with Phe254. However, several additional
water-mediated hydrogen bonds were observed in
the hPheOH-Fe(II) BH4 structure, probably due to
Ê resolution.
the higher 1.5 A
Discussion
Changes in the coordination of the catalytic
iron on reduction
Most structural and spectroscopic studies to date
on PheOH have focused on different ferric forms
of the enzyme since they are experimentally more
accessible.3,26 However, from a functional point of
view structural information on the catalytically
competent ferrous form is even more relevant. So
far, only magnetic circular dichroism (MCD) and
X-ray absorption spectroscopic (XAS) studies on
the wild-type rPheOH-Fe(II) form have been performed and interpreted in terms of a six-coordinated distorted octahedral geometry.27 The iron
coordination in the reduced form was reported to
be not ``drastically'' different from that of PheOHFe(III). The active-site iron in PheOH-Fe(III), as isolated, has electron paramagnetic resonance (EPR)
285
X-ray Studies of Phenylalanine Hydroxylase
Figure 5. Top (a) and side (b) view of the electron
density of tetrahydrobiopterin contoured as 2Fo ÿ Fc
maps at 1.7 s showing that all atoms are well de®ned.
The non-planarity of the pyrazine ring can clearly
be identi®ed in (b). The Figure was produced using
BOBSCRIPT.20
signals characteristic of high-spin (S ˆ 5/2) ferric
iron,19,28,29 which disappear under catalytic turnover conditions. Although a prereduction of the
active-site iron is an obligate step prior to catalysis30 relatively little is known about the mechanism of this reduction.26 Reduction of rPheOHFe(III) by 6-methyl-5,6,7,8-tetrahydropterin (or
under anaerobic conditions by DTN) causes a
decrease in the high-spin g ˆ 4.3 EPR signal of the
oxidized enzyme,29 ± 31 and interestingly, DTNreduced rPheOH has been reported to be kinetically indistinguishable from the 6-methyl-5,6,7,8tetrahydropterin prereduced enzyme.26,31
In the recombinant catalytic domain of hPheOH
used in the present study the high-spin Fe(III) form
has been found by EPR spectroscopy to have a
more homogenous iron coordination geometry
than in e.g. the rat liver enzyme29 and in recombinant wt-hPheOH.19 In the present crystal structure
reduced, catalytically competent enzyme was
obtained by anaerobic reduction with DTT or
DTN. The structure revealed that the six-coordinated iron observed in the hPheOH-Fe(III) form13
is less apparent in the reduced form with well
de®ned electron density only observed for the
most tightly bound axial water molecule (Wat3).
However, weak omit density is present for Wat1
and Wat2 (Figure 2(a)) implying that, in spite of
the high concentration of DTT the structure may
Figure 6. Superposition of the structures hPheOHFe(III) BH2 (orange), hPheOH-Fe(II) (blue) and hPheOHFe(III) (red) on the structure hPheOH-Fe(II) BH4 (green),
showing the different positions of Glu330. Note that all
four structures contain Wat3. The Figure was prepared
using MOLSCRIPT.2
contain traces of Fe(III) or of a six-coordinate Fe(II)
form. Interestingly, Wat1 and Wat2 are also absent
in P-rPheOH16 (PDB entry 2PHZ) which also has
an ``orthorhombic'' iron with the ligands Wat1 and
Wat2 missing. However, it should be noted that
this structure is solved to a lower resolution
Ê ). Furthermore, the Fe(II)-hPheOH structure
(2.2 A
implies a displacement and possible disorder for
Glu330, which cannot be ®tted completely into the
observed density (Figures 2(a) and 6). The Feligand distances and geometry also slightly changed upon reduction, with longer iron distances for
Wat3 and the liganding Glu330 Oe (Table 1). However, no signi®cant change was observed in the
overall protein fold. Both the present Fe(II) forms
belong to space group C2221, with cell dimensions
similar to all previous crystal structures of double
truncated hPheOH10,13 so that structural changes
must be due to biochemical rather than crystal
packing effects. Thus, the only signi®cant change
that occurs in hPheOH on reduction of the iron
centre is a reduced af®nity for two of the coordinated water molecules, displacement of Glu330
and possible disorder of its side-chain.
Binding of the reduced pterin cofactor
Previous studies on the double truncated form
of hPheOH-Fe(III) have shown that the oxidized
cofactor BH2 binds at the active site by an induced
®t mechanism, involving a small conformational
change in the protein at the active site10 as well as
in the dihydroxypropyl side-chain of the
286
cofactor.10,25 A similar conclusion is reached in the
present study on the binding of the reduced cofactor BH4 to the Fe(II) form of the enzyme (Figure 4).
The side-chain of Leu255 in hPheOH-Fe(II) BH4 is
moved compared to the ligand-free structures of
hPheOH adapting the same conformation as in
hPheOH-Fe(III) BH2. C30 makes two hydrophobic
Ê with Cg and Cd of this residue
contacts of 3.7 A
Ê with Ala322 Cb. Leu248
plus one contact of 3.8 A
adopts a different conformation compared to the
Ê
binary hPheOH-Fe(III) BH2, packing Cg1 3.7 A
away from the pterin ring. Re-orientation of the
dihydroxypropyl side-chain enables the O20 to
Ê ) with the
make a strong hydrogen bond (2.4 A
side-chain oxygen of Ser251, whereas O10 forms
water-mediated hydrogen bonding to residues
Ala322 and Glu330. The torsion angles between
the hydroxyl groups in both the hPheOHFe(III) BH2 (53 ) and the hPheOH-Fe(II) BH4
(ÿ65 ) complexes are different from the theoretical
structure of BH4, which suggests two different
forms with angles of ÿ164 and 176 ,
respectively.23 A torsion angle of about ÿ60 was
found in PheOH-bound BH2 by NMR.25
The orientation of the pterin ring system in the
present BH4 structure and the previous BH2
structure10 are not the same as that in the NMR
and computer modeling structure.25 Thus, Glu286
does indeed form two water-mediated hydrogen
bonds to the ring system, but these are to O4 and
N3 (Figure 4 and Table 2) rather than the guanidinium N2 and N3 as suggested by NMR. The same
is true of the cofactor-Fe distances with Fe-N5 at
Ê , Fe-O4 at 3.8 A
Ê and Fe-C4a at 5.9 A
Ê , rather
5.7 A
Ê
than 4.4, 2.6 and 4.3 A, respectively, in the NMR
and computer modeling structure. Thus, our study
shows that BH4 is not directly coordinated to the
iron.
All present crystal structures of PheOH and
TyrOH contain a metal ion with up to six ligands,
namely two histidine residues, one glutamic acid
and one to three water molecules. Together with
the metal ion the two histidine residues and Wat3
are almost identically positioned in all these structures. However, there is considerable variation in
the orientation of Glu330 and the presence of Wat1
and Wat2. Coordination geometries for the
hPheOH structures are given in Table 1. Glu330
seems to adopt two different conformations in the
aromatic amino acid hydroxylases. The crystal
structures reveal that the conformation in
hPheOH-Fe(II) BH4 is, with some variation, similar
to that observed in rTyrOH-Fe(III)17, the binary
and P-rPheOH-Fe(III).16 The
rTyrOH-Fe(III) BH18
2
other conformation is observed in hPheOHand the present
Fe(III)13, hPheOH-Fe(III)BH10
2
hPheOH-Fe(II) structure, although in the latter case
the side-chain is poorly de®ned. The ligand Wat3,
which is present in all PheOH and rTyrOH structures reported so far, is hydrogen bonded to the
essential Glu286 (Glu332 in TyrOH). This glutamate plays an important role in binding of BH4
through two water-mediated hydrogen bonds (Oe1-
X-ray Studies of Phenylalanine Hydroxylase
Wat4-N3 and Oe2-Wat3-O4) (Figure 4), which has
been con®rmed by site-directed mutageneses in
wt-PheOH32 and the double truncated form used
in the present study.10 When present, the ligand
Wat1 is consistently hydrogen bonded to Tyr325,
which is also involved in BH4/BH2 binding via this
water-mediated hydrogen bond to the pterin O4
(Figure 4). However, Wat1 is absent in P-rPheOH
and in the present hPheOH-Fe(II) structure. The
®nal ligand, Wat2, is also absent in hPheOH-Fe(II)
and in rPheOH-Fe(III). In the hPheOH-Fe(II) BH4
Ê ) to a
structure Wat2 is hydrogen bonded (2.9 A
well de®ned water molecule, Wat5, which lies
Ê away from C4a of BH4 and in the correct
3.2 A
direction for electrophilic attack of the cofactor.
Although only hydrogen bonded, the Wat2 Wat5
pair (Figure 7) is immediately reminiscent of the
oxygen binding site proposed in the NMR and
computer modeling structure study.25 It is also
interesting to note that if the phenylalanine binding site proposed in the NMR and computer modeling structure25 is correct, Wat2 is conveniently
placed for attack on the C3-C4 bond in the phenyl
ring (Wat2-C3 and Wat2-C4 distances 3.0 and
Ê , respectively) and, although 3.2 A
Ê away,
2.8 A
Wat5 is in the correct direction for attack on C4A
in the pterin ring.
Fuctional implications
The physiological cofactor, BH4, has both a catalytic and a regulatory function in PheOH.30 The
enzyme also requires a non-heme iron, and stoichiometric reduction of the iron centre by the
cofactor is a prerequisite for catalysis.26 In addition,
binding of the cofactor to the active site induces a
conformational transition10 that results in an inhibited state of the holo-enzyme. This state is characterized by an inhibition of the phosphorylation of
Ser16 in the regulatory domain by cyclic AMPdependent protein kinase and an increase in the
concentration of L-Phe required for the several-fold
catalytic activation of the tetrameric wild-type
enzyme.8 A recent EPR study19 has also shown
that binding of the oxidized cofactor (BH2) at the
active site of hPheOH-Fe(III) results in the appearance of a less rhombic high-spin ferric iron. Potentiometric
titrations
with
subsequent
EPR
monitoring revealed that the midpoint potential for
the catalytic iron in wt-hPheOH was reduced from
‡207(10) mV to ‡110(20) mV on binding BH2.
This 100 mV reduction in the midpoint potential
was interpreted in terms of a change in the iron
coordination environment.19 The present structure
of hPheOH-Fe(II) BH4 shows that the octahedral
arrangement around the iron becomes almost perfect, due to displacement of Glu330 (Table 3,
Figure 6), from its position in the binary hPheOHcomplex and its ligand-free form.13
Fe(III) BH10
2
The iron geometry is consistent with the magnetic
circular dichroism (MCD) studies of the resting ``Tstate'' of the ferrous form upon addition of the
reduced pterin,33 suggesting that the binary
287
X-ray Studies of Phenylalanine Hydroxylase
Figure 7. Stereo picture of the active site of the binary hPheOH-Fe(II) BH4 structure plus L-Phe from the (superimposed) ternary hPheOH-Fe(III) BH2 L-Phe NMR structure.25 The Figure shows the Wat2 Wat5 pair reminiscent of the
oxygen binding site proposed for the ternary structure.25 The Figure was prepared using MOLSCRIPT.2
hPheOH-Fe(II) BH4 complex lacks an open coordination position which would allow the formation
of a highly reactive oxygen intermediate. In the
presence of both substrate and cofactor MCD has
indicated a square-pyramidal ferrous site33 implicating the iron in the coupled hydroxylation and
suggesting that a highly reactive oxygen intermediate can only be generated after the binding of both
substrate and pterin cofactor.
A theoretical calculation using the perturbative
Becke-Perdew model (pBP/DN**)34,35 of the PC
SPARTAN PRO programme package was performed for the hydroxylated cofactor (4a-hydroxytetrahydrobiopterin) and superpositioned on BH4
in the binary complex presented here. The distance
between the oxygen incorporated in the cofactor
Ê , which is too
and the iron was measured to 6.0 A
long for the formation of a Fe(II)-m-peroxo-BH4
intermediate that has been proposed during the
activation of dioxygen in TyrOH.36 However, only
minor adjustments of the crevice structure or cofactor binding site in the ternary complex (i.e. under
turnover conditions) could alter this distance. A
change in the binding site of the cofactor seems
unlikely due to the strong network of hydrogen
bonds and hydrophobic interactions, which also is
supported by the relatively low B-factors and
unambiguous electron density of the cofactor.
However, activation by substrate (L-Phe) binding
involves a global conformational change of the
enzyme3,37 as well as in the active site crevice
structure as observed for the pterin cofactor binding, implying a possible bridging of molecular oxygen between C4a and the iron. The conformational
Table 3. Comparison of ligand angles at the active site iron ( )
285Ne-Fe-Wat3
Wat3-Fe-Wat1
Wat1-Fe-330Oe
330Oe-Fe-285Ne
330Oe-Fe-290Ne
290Ne-Fe-Wat3
Wat3-Fe-Wat2
Wat2-Fe-330Oe
285Ne-Fe-290Ne
290Ne-Fe-Wat1
Wat1-Fe-Wat2
Wat2-Fe-285Ne
R.m.s. deviation
for angles
hPheOHFe(II)BH4
hPheOHFe(III)BH210
93
83
91
93
97
88
79
95
99
84
83
94
98
80
69
113
88
87
87
92
96
89
79
95
103
6.3
11.0
10.2
hPheOH-Fe(II)
89
97
86
103
hPheOHFe(III)13
P-rPheOHFe(III)16
rTyrOHFe(III)17
rTyrOHFe(III)BH218
98
80
72
112
85
87
90
91
94
95
73
98
89
93
92
105
89
86
85
93
85
87
96
88
83
101
87
85
88
93
99
88
5.4
6.4
11.3
83
8.7
288
transition which occurs in the enzyme on L-Phe
binding has been proposed to result in a reorganization of the pterin binding site, causing the pterin
to bind closer to the catalytically important iron
centre,27 thereby allowing the enzyme to perform a
tightly coupled hydroxylation. Thus, replacement
of Wat2 and Wat5 by an approaching O2 molecule
in the hPheOH-Fe(II) BH4 structure could yield an
Fe(II)-O2-BH4 intermediate rather similar to that
proposed for the NMR and modelling structure.25
The iron ligand Glu330 adapts a completely
different conformation in hPheOH-Fe(II) BH4
Ê and the coor(Figure 6), displacing the Cg by 1.4 A
Ê as compared to the crystal
dinating Oe by 1.0 A
structures of the binary hPheOH-Fe(III) BH2 complex10 and its ligand-free form.13 Carboxylate shifts
have been observed for other enzymes including
the di-iron enzymes ribonucleotide reductase38,39
and methane monooxygenase40, and it has been
shown that conformations of carboxylate ligands
vary with the oxidation state of the metals. Carboxylate shifts are believed to be the key features for
understanding the oxygen activation mechanism in
ribonucleotide reductase41,42 and methane monooxygenase.40 We therefore strongly believe that the
carboxylate shift of Glu330 and the disorder/mobility of this side-chain are particularly important for
the oxygen activation.
Evidence for only one pterin cofactor
binding site
In addition to its catalytic function, the pterin
cofactor BH4 also functions as a potent negative
effector that inhibits L-Phe activation by forming a
binary enzyme BH4 complex.43 BH4 binds with
high af®nity,8,43 and on the basis of pre-steadystate and steady-state catalytic behaviour of
rPheOH it has been proposed that the inhibitory
effect is mediated by its binding to a second, regulatory site different from the active site.43 Furthermore, on the basis of crystal structure analyses of a
dimeric C-terminal truncated form of rPheOH with
catalytic and regulatory properties of the wild-type
enzyme, a cofactor binding site in the regulatory
domain, distant from the active site, has also been
proposed.16,44 However, no signi®cant structural
evidence has so far been presented in support of
such a binding site.3,11 On the contrary, the crystal
structure analysis of the binary hPheOHFe(III) BH2 complex has revealed that pterin binding to the active site induces localised conformational changes in the crevice structure as well as
in the pterin molecule.10 Furthermore, superposition of this binary complex onto the crystal structure of the ligand-free dimeric rPheOH,16 which
contains both the regulatory and catalytic domains,
revealed that the C20 -OH group of BH2 is suf®Ê ) to form hydrogen bonds to
ciently close (2.7 A
Ser23.10,11 The present study has con®rmed this
observation and has given additional information
on the substantial conformational changes that
occur at the active site in response to BH4 binding,
X-ray Studies of Phenylalanine Hydroxylase
notably at the iron ligand Glu330, which adapts a
completely different conformation from that in the
non-liganded form of the enzyme.
Materials and Methods
Crystallization and data collection
The double truncated mutant (N1-102/C428-452)
of hPheOH was expressed and puri®ed,45,46 and crystals
were grown using some modi®cations to the previously
described method.47 The hanging-drop vapour-diffusion
technique was used with anaerobic conditions (N2
atmosphere) in a glove box. Reservoir solutions (20 %
(w/v) PEG 2000, 10 % (v/v) ethylene glycol and 0.12 M
Na-Hepes pH 6.8) were ¯ushed with N2 through a
microporous gas distributor, and solid DTT or sodium
DTN was subsequently dissolved in the reservoir solution to a concentration of 200 mM (DTT) and deoxygenated. 5 ml of the protein solution (5.5 mg/ml) in 20 mM
Na-Hepes (pH 6.8), 75 mM NaCl were mixed with an
equal volume of the anaerobic reservoir solution. The
crystallization tray with 625 ml reservoir solution was
sealed by high vacuum grease and stored in the anaerobic glove box. For the crystallisation of the binary
hPheOH-Fe(II) BH4 complex, the reservoir included
10 mM BH4 and 150 mM DTT was used. Crystal growth
was complete within four days. Crystals were selected
and ¯ash frozen in liquid nitrogen directly from the
drop, the whole operation being carried out in the glove
box. The crystals without cofactor were thin plates with
approximate dimensions 0.5 mm 0.2 mm 0.05 mm,
whereas the crystals of the binary cofactor complex
were rod-shaped with dimensions 0.7 mm 0.2 mm
0.1 mm.
Diffraction images were collected at 110 K on the
Swiss-Norwegian Beamline (BM01) at the European Synchrotron Radiation Facility (ESRF) in Grenoble (France),
using a MAR345 Research imaging plate system and a
Ê . One single crystal was used for
wavelength of 0.873 A
each data set. An additional separate data set comprising
Ê was collected on the
data between 10 and 2.54 A
hPheOH-Fe(II) BH4 (DTT) crystal using a longer crystal
to detector distance and shorter exposure time, with the
aim of reducing the background and saturation for the
low-resolution data. The data were integrated using the
programme DENZO,48 while the scaling and merging
were performed using SCALA from the CCP4 program
suite (Collaborative Computational Project, Number 4,
1994). The statistics for the data sets are presented in
Table 4.
Model building and refinement
The coordinates for the crystal structure of the catalytic domain of hPheOH-Fe(III)13 (PDB 1PAH) were used
as the starting model and for calculating initial phases
for both structures. The re®nement was initiated by
rigid-body re®nement and simulated annealing methods
using CNS version 1.0.54 Further re®nement was performed using SHELX-9755 and the peptide-chain was
adjusted manually with the aid of the graphical display
programme O.56 The electron density maps used for the
manual adjustments were s-weighted,51 2Fobs ÿ Fcalc and
Fobs ÿ Fcalc calculated by SHELX-97. 10 % of the data
were excluded in the re®nement for cross-validation.49
Water molecules were added gradually in positions with
Fobs ÿ Fcalc density exceeding 4 s and 2Fobs ÿ Fcalc
289
X-ray Studies of Phenylalanine Hydroxylase
Table 4. Summary of data-collection and re®nement statistics of the hPheOH-Fe(II) and hPheOH-Fe(II) BH4
structures
Ê)
Resolution (A
Ê)
Cell dimensions (A
Number of observations
Number of unique reflections
Multiplicity
Rmerge (%)
I/s(I)
Completeness (%)
Ê 2)
Wilson B (A
Ê)
Resolution range in refinement (A
Number of atoms in refinement
Number of solvent molecules
R-factor (%)
Rfree (%)49
Ê 2)
Average protein B-factor (A
Ê 2)
Average cofactor B-factor (A
Ê)
R.m.s. deviation for bond lengths (A
R.m.s. deviation for bond angles (deg.)
R.m.s. deviation for dihedral angles (deg.)
R.m.s deviation for improper angles (deg.)
Ê )50
Luzatti r.m.s. coordinate error (A
Ê )51
sA r.m.s. coordinate error (A
f and c angles in most favoured regions (%)a
f and c angles in additional allowed regions (%)a
f and c angles in generously allowed and disallowed regions (%)a
hPheOH-Fe(II)
hPheOH-Fe(II) BH4
25-1.7
a ˆ 66.33,
b ˆ 108.42,c ˆ 124.02
260,644 (33,396)
48,353 (6776)
5.4 (4.8)
7.0 (36.6)
6.4 (2.1)
98.7 (97.8)
24.7
10-1.7 (1.79-1.70)
2814
278
19.7
25.2
27.6
20-1.5
a ˆ 66.44, b ˆ 108.50,
c ˆ 124.42
495,560 (61,233)
71,940 (10,450)
6.9 (5.9)
5.1 (37.0)
8.7 (2.0)
99.9 (100.0)
18.4
10-1.5 (1.58-1.50)
3041
393
15.7
20.3
19.5
25.6
0.012
2.03
23.8
1.26
0.24
0.09
94.4
5.6
0
0.008
2.00
24.0
1.25
0.27
0.18
94.8
5.2
0
Statistics for highest resolution shell are given in parentheses.
Statistics from Ramachandran plots52 calculated by PROCHECK53
a
density exceeding 1.7 s at hydrogen-bonding distances
Ê distance from an oxygen or a nitrogen). Water
(2.3-3.4 A
Ê 2 were
oxygen atoms with B-factors higher than 63 A
removed prior to the next re®nement round. The iron
and all its ligands were omitted early on in the re®nement to ensure that their positions were unbiased. A
theoretical calculation of BH4 using the perturbative
Becke-Perdew model (pBP/DN**)34,35 of the PC SPARTAN PRO programme package was performed making
the basis for the cofactor restraints concerning bond
lengths and 1,3-distances. The theoretical structure
matched the electron density impressively even prior to
any re®nement. The higher resolution of the hPheOHFe(II) BH4 (DTT) complex made it possible to re®ne the
structure with anisotropic B-factors for all non-hydrogen
atoms in the protein and the cofactor. B-Factors for the
hPheOH-Fe(II) (DTT) structure were kept isotropic, but
the re®nement was performed using an overall anisotropic correction. Side-chain atoms outside 1.5 s
2Fobs ÿ Fcalc density were given occupancy zero, and
double conformations were included for some sidechains in the latter re®nement rounds. The ®nal models
gave R-factors of 15.7 and 19.7 % for the hPheOHFe(II) BH4 (DTT) and the hPheOH-Fe(II) (DTT) structures, respectively. The better quality of the re®ned
hPheOH-Fe(II) BH4 (DTT) structure is most probably
due the anisotropic B-factors, larger volume of the crystals and the separate low-resolution data set collected.
Protein Data Bank accession numbers
Atomic coordinates and structure factors have been
deposited at the Protein Data Bank, Research Collaboratory for Structural Bioinformatics, with accession numbers 1J8T for the hPheOH-Fe(II) structure and 1J8U for
the hPheOH-Fe(II) BH4 binary complex, respectively.
The coordinates will remain privileged until 2002-05-22.
Acknowledgements
This work has been supported by grants from the
Norwegian Research Council (NFR), the Norwegian
Council on Cardiovascular Diseases, Rebergs legat, L.
Meltzer Hùyskolefond, the Novo Nordisk Foundation
and the European Commission. We thank Ali Sepulveda
MunÄoz for expert technical assistance in preparing the
bacterial extracts and fusion protein, the staff of the
Swiss-Norwegian Beamlines in Grenoble, France under
data collection and Aurora MartõÂnez with co-workers for
access to the coordinates of the ternary hPheOHFe(III) BH2 L-Phe NMR structure.
References
1. Kaufman, S. (1993). The phenylalanine hydroxylating system. Advan. Enzymol. Relat. Areas Mol. Biol.
67, 77-264.
2. Kraulis, P. J. (1991). MOLSCRIPT: a program to produce both detailed and schematic plots of protein
structures. J. Appl. Crystallog. 24, 946-950.
3. Flatmark, T. & Stevens, R. C. (1999). Structural
insight into the aromatic amino acid hydroxylases
and their disease-related mutant forms. Chem. Rev.
99, 2137-2160.
4. Cronk, J. D., Endrizzi, J. A. & Alber, T. (1996). Highresolution structures of the bifunctional enzyme and
transcriptional coactivator DCoH and its complex
with a product analogue. Protein Sci. 5, 1963-1972.
290
5. Varughese, K. I., Skinner, M. M., Whiteley, J. M.,
Matthews, D. A. & Xuong, N. H. (1992). Crystalstructure of rat-liver dihydropteridine reductase.
Proc. Natl Acad. Sci. USA, 89, 6080-6084.
6. Kaufman, S. (1959). Studies on the mechanism of the
enzymatic conversion of phenylalanine to tyrosine.
J. Biol. Chem. 234, 2677-2682.
7. Scriver, C. R., Waters, P. J., Sarkissian, C., Ryan, S.,
Prevost, L., CoÃteÂ, D. et al. (2000). PAHdb: A locusspeci®c knowledgebase. Hum. Mutat. 15, 99-104.
8. Dùskeland, A. P., Dùskeland, S. O., égreid, D. &
Flatmark, T. (1984). The effect of ligands of phenylalanine 4-monooxygenase on the cAMP-dependent
phosphorylation of the enzyme. J. Biol. Chem. 259,
11242-11248.
9. Dùskeland, A. P., Haavik, J., Flatmark, T. &
Dùskeland, S. O. (1987). Modulation by pterins of
the phosphorylation and phenylalanine activation of
phenylalanine 4-mono-oxygenase. Biochem. J. 242,
867-874.
10. Erlandsen, H., Bjùrgo, E., Flatmark, T. & Stevens,
R. C. (2000). Crystal structure and site-speci®c mutagenesis of pterin-bound human phenylalanine
hydroxylase. Biochemistry, 39, 2208-2217.
11. Flatmark, T., Erlandsen, E., Bjùrgo, E., Solstad, T.,
Dùskeland, A. P. & Stevens, R. C. (2000). Regulatory
properties of tetrahydrobiopterin cofactor bound
at the active site of phenylalanine hydroxylase.
Pteridines, 11, 34-36.
12. Wang, G. A., Gu, P. & Kaufman, S. (2001). Mutagenesis of the regulatory domain of phenylalanine
hydroxylase. Proc. Natl Acad. Sci. USA, 98, 15371542.
13. Erlandsen, H., Fusetti, F., MartõÂnez, A., Hough, E.,
Flatmark, T. & Stevens, R. C. (1997). Crystal structure of the catalytic domain of human phenylalanine
hydroxylase reveals the structural basis for phenylketonuria. Nature Struct. Biol. 4, 995-1000.
14. Fusetti, F., Erlandsen, H., Flatmark, T. & Stevens,
R. C. (1998). Structure of tetrameric human phenylalanine hydroxylase and its implications for phenylketonuria. J. Biol. Chem. 273, 16962-16967.
15. Erlandsen, H., Flatmark, T., Stevens, R. C. & Hough,
E. (1998). Crystallographic analysis of the human
phenylalanine hydroxylase catalytic domain with
Ê resolution. Biobound catechol inhibitors at 2.0 A
chemistry, 37, 15638-15646.
16. Kobe, B., Jennings, I. G., House, C. M., Michell, B. J.,
Goodwill, K. E., Santarsiero, B. D. et al. (1999). Structural basis of autoregulation of phenylalanine
hydroxylase. Nature Struct. Biol. 6, 442-448.
17. Goodwill, K. E., Sabatier, C., Marks, C., Raag, R.,
Fitzpatrick, P. F. & Stevens, R. C. (1997). Crystal
Ê resolution
structure of tyrosine hydroxylase at 2.3 A
and its implications for inherited neurodegenerative
diseases. Nature Struct. Biol. 4, 578-585.
18. Goodwill, K. E., Sabatier, C. & Stevens, R. C. (1998).
Crystal structure of tyrosine hydroxylase with
Ê resolbound cofactor analogue and iron at 2.3 A
ution: self-hydroxylation of Phe300 and the pterinbinding site. Biochemistry, 37, 13437-13445.
19. Hagedoorn, P.-L., Schmidt, P. P., Andersson, K. K.,
Hagen, W. R., Flatmark, T. & MartõÂnez, A. (2001).
The effect of substrate, dihydrobiopterin, and dopamine on the EPR spectroscopic properties and the
midpoint potential of the catalytic iron in recombinant human phenylalanine hydroxylase. J. Biol.
Chem. 276, 22850-22856.
X-ray Studies of Phenylalanine Hydroxylase
20. Esnouf, R. M. (1997). An extensively modi®ed
version of MolScript that includes greatly enhanced
coloring capabilities. J. Mol. Graph. Model, 15, 132134.
21. Wallace, A. C., Laskowski, R. A. & Thornton, J. M.
(1995). LIGPLOT: a program to generate schematic
diagrams of protein-ligand interactions. Protein Eng.
8, 127-134.
22. Matsuura, S., Sugimoto, T., Murata, S., Sugawara, Y.
& Iwasaki, H. (1985). Stereochemistry of biopterin
cofactor and facile methods for determination of the
stereochemistry of a biologically active 5,6,7,8-tetrahydropterin. J. Biochem. 98, 1341-1348.
23. Katoh, S., Sueoka, T. & Kurihara, T. (1993). Theoretical stereostructure of the neutral form of natural
tetrahydrobiopterin. Pteridines, 4, 27-31.
24. Bieri, J. H. (1977). The crystal structure of 6-methyl7,8-dihydropterine-monohydrochloride-monohydrate. Helv. Chim. Acta, 60, 2303-2308.
25. Teigen, K., Frùystein, N. A. & MartõÂnez, A. (1999).
The structural basis of the recognition of phenylalanine and pterin cofactors by phenylalanine
hydroxylase: implications for the catalytic mechanism. J. Mol. Biol. 294, 807-823.
26. Kappock, T. J. & Caradonna, J. P. (1996). Pterindependent amino acid hydroxylases. Chem. Rev. 96,
2659-2756.
27. Loeb, K. E., Westre, T. E., Kappock, T. J., Mitic, N.,
Glasfeld, E., Caradonna, J. P.et al. (1997). Spectroscopic characterization of the catalytically competent
ferrous site of the resting, activated, and substratebound forms of phenylalanine hydroxylase. J. Amer.
Chem. Soc. 119, 1901-1915.
28. Fisher, D. B., Kirkwood, R. & Kaufman, S. (1972).
Rat liver phenylalanine hydroxylase, an iron
enzyme. J. Biol. Chem. 247, 5161-5167.
29. MartõÂnez, A., Andersson, K. K., Haavik, J. &
Flatmark, T. (1991). EPR and 1H-NMR spectroscopic
studies on the paramagnetic iron at the active site of
phenylalanine hydroxylase and its interaction with
substrates and inhibitors. Eur. J. Biochem. 198, 675682.
30. Marota, J. J. & Shiman, R. (1984). Stoichiometric
reduction of phenylalanine hydroxylase by its cofactor: a requirement for enzymatic activity. Biochemistry, 23, 1303-1311.
31. Wallick, D. E., Bloom, L. M., Gaffney, B. J. &
Benkovic, S. J. (1984). Reductive activation of
phenylalanine hydroxylase and its effect on the
redox state of the non-heme iron. Biochemistry, 23,
1295-1302.
32. Dickson, P. W., Jennings, I. G. & Cotton, R. G.
(1994). Delineation of the catalytic core of phenylalanine hydroxylase and identi®cation of glutamate
286 as a critical residue for pterin function. J. Biol.
Chem. 269, 20369-20375.
33. Kemsley, J. N., Mitic, N., Zaleski, K. L., Caradonna,
J. P. & Solomon, E. I. (1999). Circular dichroism and
magnetic circular dichroism spectroscopy of the catalytically competent ferrous active site of phenylalanine hydroxylase and its interaction with pterin
cofactor. J. Am. Chem. Soc. 121, 1528-1536.
34. Becke, A. D. (1988). Density-functional exchangeenergy approximation with correct asymptotic behaviour. Phys. Rev. A, 38, 3098-3100.
35. Perdew, J. P. (1986). Density-functional approximation for the correlation energy of the inhomogeneous electron gas. Phys. Rev. B, 33, 8822-8824.
291
X-ray Studies of Phenylalanine Hydroxylase
36. Francisco, W. A., Tian, G., Fitzpatrick, P. F. &
Klinman, J. P. (1998). Oxygen-18 kinetic isotope
effect studies of the tyrosine hydroxylase reaction:
evidence of rate limiting oxygen activation. J. Am.
Chem. Soc. 120, 4057-4062.
37. Flatmark, T., Stokka, A. J. & Berge, S. V. (2001). Use
of surface plasmon resonance for real-time measurements of the global conformational transition in
human phenylalanine hydroxylase in response to
substrate binding and catalytic activation. Anal.
Biochem. 294, 95-101.
38. Nordlund, P., SjoÈberg, B. M. & Eklund, H. (1990).
Three-dimentional structure of the free-radical
protein of ribonucleotide reductase. Nature, 345,
593-598.
39. Logan, D. T., Su, X. D., Aberg, A., RegnstroÈm, K.,
Hajdu, J., Eklund, H. & Nordlund, P. (1996). Crystal
structure of reduced protein R2 of ribonucleotide
reductase: the structural basis for oxygen activation
at a dinuclear iron site. Structure, 4, 1053-1064.
40. Rosenzweig, A. C., Nordlund, P., Takahara, P. M.,
Fredrick, C. A. & Lippard, S. J. (1995). Geometry of
the soluble methane monooxygenase catalytic diiron
center in two oxidation states. Chem. Biol. 2, 409-418.
41. Andersson, M. E., HoÈgbom, M., Rinaldo-Matthis, A.,
Andersson, K. K., SjoÈberg, B.-M. & Nordlund, P.
(1999). The crystal structure of an azide complex of
the diferrous R2 subunit of ribonucleotide reductase
displays a novel carboxylate shift with important
mechanistic implications for diiron-catalysed oxygen
activation. J. Am. Chem. Soc. 121, 2346-2352.
42. HoÈgbom, M., Andersson, M. E. & Nordlund, P.
(2001). Crystal structures of oxidized dinuclear
manganese centres in Mn-substituted class I ribonucleotide reductase from Escherichia coli: carboxylate
shifts with implications for O2 activation and radical
generation. J. Biol. Inorg. Chem. 6, 315-323.
43. Xia, T., Gray, D. W. & Shiman, R. (1994). Regulation
of rat liver phenylalanine hydroxylase III. Control of
catalysis by (6R)-tetrahydrobiopterin and phenylalanine. J. Biol. Chem. 269, 24657-24665.
44. Jennings, I. G., Teh, T. & Kobe, B. (2001). Essential
role of the N-terminal autoregulatory sequence in
the regulation of phenylalanine hydroxylase. FEBS
Letters, 488, 196-200.
45. MartõÂnez, A., Knappskog, P. M., Olafdottir, S.,
Dskeland, A. P., Eiken, H. G., Svebak, R. M. et al.
(1995). Expression of recombinant human phenyl-
46.
47.
48.
49.
50.
51.
52.
53.
54.
55.
56.
alanine-hydroxylase as fusion protein in Escherichia
coli circumvents proteolytic degradation by host cell
proteases. Isolation and characterization of the wildtype enzyme. Biochem. J. 306, 589-597.
Knappskog, P. M., Flatmark, T., Aarden, J. M.,
Haavik, J. & MartõÂnez, A. (1996). Structure/function
relationships in human phenylalanine hydroxylase.
Effect of terminal deletions on the oligomerization,
activation and cooperativity of substrate binding to
the enzyme. Eur. J. Biochem. 242, 813-821.
Erlandsen, H., MartõÂnez, A., Knappskog, P. M.,
Haavik, J., Hough, E. & Flatmark, T. (1997). Crystallization and preliminary diffraction analysis of a
truncated homodimer of human phenylalanine
hydroxylase. FEBS Letters, 406, 171-174.
Otwinowski, Z. & Minor, W. (1997). Processing of
X-ray diffraction data collected in oscillation mode.
Methods Enzymol. 276, 307-326.
Kleywegt, G. J. & BruÈnger, A. T. (1996). Checking
your imagination: applications of the free R value.
Structure, 4, 987-904.
Luzzati, V. (1952). Traitement statistique des errors
dans la determination des structures cristallines.
Acta Crystallog. 5, 802-810.
Read, R. J. (1986). Improved Fourier coef®cients for
maps using phases from partial structures with
errors. Acta Crystallog. sect. A, 42, 140-149.
Ramachandran, G. N. & Sasisekharan, V. (1968).
Conformation of polypeptides and proteins. Advan.
Protein Chem. 23, 283-438.
Laskowski, R. A., MacArthur, M. W., Moss, D. S. &
Thornton, J. M. (1993). PROCHECK - a program to
check the stereochemical quality of protein structures. J. Appl. Crystallog. 26, 283-291.
BruÈnger, A. T., Adams, P. D., Clore, G. M., DeLano,
W. L., Gros, P., Grosse-Kunstleve, R. W. et al. (1998).
Crystallography and NMR system: Aa new software
suite for macromolecular structure determination.
Acta Crystallog. sect. D, 54, 905-921.
Sheldrick, G. M. & Schneider, T. R. (1997). SHELXL:
High-resolution re®nement. Methods Enzymol. 277,
319-343.
Jones, T. A., Zou, J.-Y., Cowan, S. W. & Kjeldgaard,
M. (1991). Improved methods for building protein
models in electron density maps and the location of
errors in these models. Acta Crystallog. sect. A, 47,
110-119.
Edited by R. Huber
(Received 30 May 2001; received in revised form 28 August 2001; accepted 28 August 2001)