doi:10.1016/j.jmb.2005.06.076
J. Mol. Biol. (2005) 352, 105–116
Structures of Leishmania major Pteridine Reductase
Complexes Reveal the Active Site Features Important
for Ligand Binding and to Guide Inhibitor Design
Alexander W. Schüttelkopf1, Larry W. Hardy2, Stephen M. Beverley3 and
William N. Hunter1*
1
Division of Biological
Chemistry and Molecular
Microbiology, School of Life
Sciences, University of Dundee
Dundee DD1 5EH, UK
2
Sepracor Inc., 84 Waterford
Drive, Marlborough, MA 01752
USA
3
Department of Molecular
Microbiology, Washington
University School of Medicine
Campus Box 8230, 660 S. Euclid
Avenue, St. Louis, MO
63110-1093, USA
Pteridine reductase (PTR1) is an NADPH-dependent short-chain reductase
found in parasitic trypanosomatid protozoans. The enzyme participates in
the salvage of pterins and represents a target for the development of
improved therapies for infections caused by these parasites. A series of
crystallographic analyses of Leishmania major PTR1 are reported. Structures
of the enzyme in a binary complex with the cofactor NADPH, and ternary
complexes with cofactor and biopterin, 5,6-dihydrobiopterin, and 5,6,7,8tetrahydrobiopterin reveal that PTR1 does not undergo any major
conformational changes to accomplish binding and processing of
substrates, and confirm that these molecules bind in a single orientation
at the catalytic center suitable for two distinct reductions. Ternary
complexes with cofactor and CB3717 and trimethoprim (TOP), potent
inhibitors of thymidylate synthase and dihydrofolate reductase, respectively, have been characterized. The structure with CB3717 reveals that the
quinazoline moiety binds in similar fashion to the pterin substrates/
products and dominates interactions with the enzyme. In the complex with
TOP, steric restrictions enforced on the trimethoxyphenyl substituent
prevent the 2,4-diaminopyrimidine moiety from adopting the pterin mode
of binding observed in dihydrofolate reductase, and explain the inhibition
properties of a range of pyrimidine derivates. The molecular detail
provided by these complex structures identifies the important interactions
necessary to assist the structure-based development of novel enzyme
inhibitors of potential therapeutic value.
q 2005 Elsevier Ltd. All rights reserved.
*Corresponding author
Keywords: antifolates; dihydrofolate reductase; Leishmania; pterin; shortchain reductase
Abbreviations used: ASA, accessible surface area; DPI,
diffraction-component precision index; DHB, 2-amino-7,
8-dihydro-6-(1,2-dihydroxypropyl)pteridin-4(3H)-one or
7,8-dihydrobiopterin; DHFR, dihydrofolate reductase;
E.S.R.F., European Synchrotron Radiation Facility; MTX,
4-amino-N10-methyl-pteroylglutamic acid or methotrexate; NCS, non-crystallographic symmetry; pABA,
para-aminobenzoic acid; PDB, Protein Data Bank; PTR1,
pteridine reductase; SDR, short-chain dehydrogenase/
reductase; S.R.S., Synchrotron Radiation Source; TAQ, 2,4,
6-triaminoquinazoline; THB, 2-amino-5,6,7,8-tetrahydro6-(1,2-dihydroxypropyl)pteridin-4(3H)-one or 5,6,7,8tetrahydrobiopterin; TOP, 5-((3,4,5-trimethoxyphenyl)
methyl)-2,4-diaminopyrimidine or trimethoprim; TS,
thymidylate synthase.
E-mail address of the corresponding author:
[email protected]
Introduction
Infection by trypanosomatid protozoans, e.g.
Trypanosoma and Leishmania species, causes a
range of serious human diseases.1 Current treatments of these infections involve drugs such as
sodium stibogluconate and nifurtimox, which
induce serious side-effects and this, together with
an increase in drug-resistant parasites, has created
an urgent requirement for new and more effective
treatments.2 The ideal enzyme targets for the
development of such new antimicrobial drugs are
those that are essential for the survival of the
parasite, and either absent from the human host or
display markedly differing substrate specificities.
Our improved knowledge of trypanosomatid
0022-2836/$ - see front matter q 2005 Elsevier Ltd. All rights reserved.
106
biology and biochemistry, greatly facilitated by
developments in molecular genetics and genome
sequencing efforts, promotes understanding of how
existing drugs function and is helping to identify
potential targets for chemotherapeutic attack.3,4
Whilst there are increased opportunities to target
recently discovered enzymes there are also wellstudied systems that still have much to offer. Our
present study concerns the Leishmania pteridine
reductase (PTR1; EC 1.5.1.33), which is involved
in folate/pterin metabolism and, as will be
explained, presents a link to the well-known
therapeutic targets dihydrofolate reductase
(DHFR; EC 1.5.1.3) and thymidylate synthase (TS;
EC 2.1.1.45).
Folate metabolism, in particular DHFR and TS,
have been targeted successfully for the treatment of
cancer and microbial infection.5–10 TS catalyses the
conversion of dUMP to dTMP using the cofactor
N5, N10-methylene tetrahydrofolate (THF) as both
the C-donor and reductant, whilst DHFR maintains
the THF pool by the NADPH-dependent reduction
of dihydrofolate (DHF). Inhibition of either enzyme
limits the supply of dTMP required for DNA
synthesis, thus curtailing replication and leading
to cell death. In most organisms, DHFR and TS are
separate entities. Trypanosomatid protozoans
possess a bifunctional DHFR-TS, which has been
structurally characterised.11
Since folates and pterins are essential for trypanosomatid growth, antifolates targeting DHFR, in
principle, should provide an ideal treatment.
However, DHFR inhibitors are largely ineffective
for the control of trypanosomatid infections, partly
due to the presence of PTR1.12 This short-chain
dehydrogenase/reductase (SDR) family member
exhibits a broad NADPH-dependent pteridine
reductase activity, capable of reducing unconjugated (biopterin) and conjugated (folate) pterins
from either the oxidized or dihydro-state (Figure 1).
This activity is essential for parasite growth
in vitro.12 The biochemical activities of PTR1 overlap
those of DHFR but PTR1 is less susceptible to
inhibition by classical antifolates such as methotrexate (MTX); thus, it can function as a metabolic
by-pass to alleviate DHFR inhibition.13 However, an
inhibitor of PTR1 has the potential to act in
concert with known antifolates to provide a novel
approach to the treatment of trypanosomatid
infection.14
The kinetics and stereochemical course of the
reductions catalyzed by PTR1 have been studied
together with analysis of a library of inhibitors.14,15
Crystal structures are available of the enzyme from
Leishmania major (LmPTR1) in ternary complexes
with cofactor and 7,8-dihydrobiopterin (DHB),16
the archetypal antifolate drug MTX (4-amino-N10methyl-pteroylglutamic acid; Figure 2)16 and 2,4,6triaminoquinazoline (TAQ).17 Other laboratories
have reported the 2.7 Å resolution crystal structure of PTR1 from Leishmania tarentolae18 and the
crystallization of the enzyme from Trypanosoma
cruzi.19
Structures of PTR1 Ligand Complexes
Figure 1. The two-stage reduction of biopterin (2amino-6-(1,2-dihydroxypropyl)pteridin-4(3H)-one) to 7,8dihydrobiopterin (2-amino-7,8-dihydro-6-(1,2-dihydroxypropyl)pteridin-4(3H)-one), and then to 5,6,7,8-tetrahydrobiopterin (2-amino-5,6,7,8-tetrahydro-6-(1,2-dihydroxypropyl)pteridin-4(3H)-one) catalyzed by PTR1. Each stage
requires one reducing equivalent provided by the cofactor
NADPH. The carbon atoms (C6 and C7) that accept hydride
from the cofactor are marked with an asterisk (*).
We set out to characterize PTR1–ligand
complexes to derive important molecular details
necessary to understand the enzyme and to underpin a structure-based approach to inhibitor
development. Structural analyses of discrete steps
along the two-stage PTR1 catalytic process
(Figure 1) are reported. The binary complex with
cofactor (PTR1:NADPH) corresponds to the state
prior to binding substrate. The ternary complexes
with biopterin, and DHB represent the initial
substrate complexes for stages I and II, respectively.
The DHB complex also represents the product
complex at the end of stage I, and finally the
5,6,7,8-tetrahydrobiopterin
(THB)
complex
represents the product complex at the end of
stage II.
The potent TS inhibitor CB3717 (N10-propargyl5,8-dideazafolic acid, Figure 2), resembles DHF,
though with the addition of a propargyl group, and
we reasoned that it might bind to PTR1, thereby
107
Structures of PTR1 Ligand Complexes
Results and Discussion
Crystallographic details
Five medium-resolution crystal structures of
LmPTR1 ligand complexes have been determined
with diffraction data recorded using synchrotron
radiation. The structures include the binary
complex with cofactor NADPH (2.65 Å resolution),
ternary complexes of oxidized cofactor and
biopterin (2.40 Å) and THB (2.55 Å) and two ternary
complexes with the inhibitors CB3717 (2.70 Å) and
TOP (2.60 Å). The structures all display space group
P212121 and are isomorphous. A homotetramer, of
molecular mass approximately 120 kDa, constitutes
the asymmetric unit and gives a Matthews coefficient, VM, of 2.8 Å3 DaK1 and 55% (v/v) solvent
volume. The PTR1 tetramer displays 222 point
group symmetry. Most of the protein residues are
well defined by the electron density, with only two
segments disordered in all subunits; these are the
surface loops comprising residues 75–80 and
120–130. In addition, residues 230–240 are disordered in two subunits. This is discussed later. In
the structures, the cofactor, biopterin, THB and
pteridine component of CB3717 are well ordered in
each of the four active sites in the asymmetric unit.
The remainder of CB3717 is poorly ordered. The
inhibitor TOP is observed in two active sites only.
Further information is given in Table 1.
Overall structure and cofactor binding
Figure 2. The chemical structures of three PTR1
inhibitors: methotrexate, CB3717 and trimethoprim. The
molecules are depicted in a similar orientation in all
Figures. The methotrexate molecule is depicted with the
pterin-ring flipped about the N2–N5 axis relative to other
pterins as actually observed in the PTR1 active site.
raising the potential to design a molecule capable
of inhibiting both PTR1 and TS. Structures of
CB3717 in complex with TS from Escherichia coli,20
Lactobacillus casei,21 Pneumocystis carinii TS22 and the
bifunctional DHFR-TS of L. major11 are available for
comparative purposes.
One of the most successful antimicrobial antifolates is trimethoprim (5-((3,4,5-trimethoxyphenyl)
methyl)-2,4-diaminopyrimidine, TOP; Figure 2).9
TOP is a potent inhibitor of DHFR23 and structures
with the enzyme from Mycobacterium tuberculosis,24
Staphylococcus aureus,25 E. coli and chicken26 are
known. A structure of PTR1 in complex with TOP
was sought to provide a different molecular framework in the enzyme active site, one that is necessary
to understand PTR1 inhibition by a series of 2,4diaminopyrimidine derivatives.
PTR1 is a member of the SDR super family of
enzymes, of which there are over 3000 members
with sequence identities at the 15–30% level.27,28
The subunit displays the extended doubleRossmann fold typical of the SDR family, which is
based on a central seven-stranded parallel b-sheet
with three a-helices on either side (Figure 3(a)). In
addition, PTR1 has a short helix a6 following b6,
and two short strands between b3 and a3 framing a
disordered loop (residues 75–80). Another,
frequently disordered section of the PTR1 structure
occurs between b4 and a4 (residues 120–130).
Residues 230–240 form the so-called substratebinding loop29 between b6 and a6, a common
structural feature of the SDR family. Generally, this
loop is disordered or in an open conformation in
crystal structures of the apo form or binary
complexes of family members. Once substrate or
inhibitors bind, the loop adopts an ordered and
closed conformation. In the present structures, this
substrate-binding loop is ordered in two of the four
subunits that constitute the asymmetric unit, due to
interactions within the crystal lattice, in particular
the associations of Met233 and Pro234 with Tyr114
from a symmetry mate (not shown).
A comparison of seven LmPTR1 structures, the
five reported here plus the MTX and DHB
complexes,16 by a least-squares fit of all superimposable Ca atoms in the functional tetramers
108
Structures of PTR1 Ligand Complexes
Table 1. Data collection and refinement statistics
Structure
Wavelength (Å)
Unit cell:
a (Å)
b (Å)
c (Å)
Resolution range (Å)
Unique reflections
Redundancy
Completeness (%)
I/s(I)
Rsym (%)
Wilson B (Å2)
Protein residues
Molecules of solvent,
cofactor, ligand and
ethylene glycol
R/Rfree (%)
Average B (Å2):
Overall
Protein
Cofactor
Ligand
Solvent
Ethlyene glycol
RMSD bond lengths (Å)
RMSD bond angles (8)
Cruickshank’s DPI (Å)
Ramachandran plot:
percentage of residues in
Most favoured region
Add. allowed regions
Gen. allowed regions
PDB accession codes
PTR1:NADPH
PTR1:NADPC:
biopterin
PTR1:NADPC:
THB
PTR1:NADPH:
CB3717
PTR1:NADPH:
TOP
0.933
0.890
0.933
1.033
0.890
94.31
103.49
137.98
25.00–2.65
40607
3.7
96.8 (92.6)
13.6 (5.5)
8.8 (22.3)
41.8
95.27
104.55
138.07
25.00–2.40
47504
4.3
86.9 (76.5)
12.9 (3.7)
9.9 (29.3)
37.8
93.10
103.57
137.43
25.00–2.55
43624
4.8
98.2 (98.0)
17.2 (5.6)
7.6 (25.9)
51.1
94.54
103.83
137.05
30.00–2.70
35456
3.9
92.7 (85.0)
9.8 (2.9)
13.7 (40.8)
37.4
94.70
104.44
138.06
25.00–2.60
41645
4.2
98.2 (96.4)
14.4 (5.5)
9.4 (26.2)
38.0
1044
378/4/0/6
1036
268/4/4/5
1046
152/4/4/6
1039
211/4/4/6
1042
185/4/2/0
22.6/26.9
21.9/24.2
19.8/23.8
20.3/23.6
25.3/28.8
36.3
36.3
35.0
–
35.5
31.5
0.012
1.3
0.32
33.3
33.3
28.9
42.2
31.4
44.6
0.008
1.4
0.25
46.6
46.7
42.3
48.6
44.6
52.2
0.008
1.5
0.27
28.9
28.3
29.4
69.4
23.2
41.8
0.007
1.0
0.31
29.3
29.0
41.2
51.9
21.3
0.010
1.4
0.34
88.6
10.8
0.6
2BFO
88.6
10.5
0.9
2BF7
89.1
10.1
0.8
2BFP
88.2
10.6
1.2
2BFA
87.4
11.6
1.0
2BFM
Values in parentheses pertain to the highest resolution shell (widthz0.1 Å). DPIZdiffraction-component precision index.52
(approximately 1035 atoms for each structure),
gives a root-mean-square deviation (r.m.s.d.)
range of 0.17–0.58 Å with a mean of 0.34 Å. The
smallest deviation is observed for the overlay of the
ternary DHB and THB complexes, and the largest is
observed between the binary cofactor complex
paired with the ternary MTX complex. Since the
structures were refined independently, this indicates a high level of structural conservation of the
protein, irrespective of whether PTR1 is in a binary
or a ternary complex, or whether substrate, product
or an inhibitor is bound. Inspection of the superimposed molecules (not shown) confirms this high
degree of structural conservation, which extends to
the positions of side-chains, the conformation of the
cofactor, and even the positions of numerous water
molecules. We also note structural conservation
within the asymmetric unit of each complex. Such
consistency indicates that PTR1 does not undergo
large conformational changes upon binding and
processing substrate, and that it is necessary to
describe only one enzyme active site, in this case
chosen arbitrarily as that formed mostly by
subunit A.
The PTR1 active site is an elongated L-shaped
cleft (Figure 3(b)) approximately 22 Å!15 Å,
created mainly by C-terminal sections of strands
b1 to b6, and sections contributed from a1, a4 and
a5 together with the residues on the loop between
b6 and a6. The C terminus of a partner subunit
blocks one end of the active site, with Arg287 0 (the
prime ( 0 ) identifies a contribution from another
subunit) directed towards the catalytic center and
near Asp181. The cofactor binds in the active site in
an extended conformation, with the nicotinamide
creating the floor of the catalytic center; Phe113
forms an overhang under which the pterin-binding
pocket is formed. The adenine residue adopts an
anti conformation and the nicotinamide moiety
adopts a syn conformation with respect to their
ribose groups. In the binary complex, the pterinbinding site is occupied by water molecules with a
molecule of ethylene glycol binding to Asp181 and
Arg287 0 (Figure 4). The MTX complex also shows
ethylene glycol binding in this region of the active
site.
A complex network of hydrogen bonds organize
the enzyme active site and position the cofactor (not
shown).16,17 Notably, Asn147 forms interactions
with Lys198 and Ser111, positioning the latter two
residues to bind the nicotinamide ribose. The amide
group of Ser111 is also able to donate a hydrogen
bond to O4 0 of the adenine ribose. Arg17 interacts
with the main-chain carbonyl group of Val228 and
plays a critical role in binding the cofactor
pyrophosphate. The main-chain amide and
109
Structures of PTR1 Ligand Complexes
Figure 4. Electron density associated with an active site
of (a) the binary cofactor complex, (b) the ternary
biopterin complex, (c) the ternary DHB complex16 and
(d) the ternary THB complex. The maps are drawn as blue
chicken-wire and have been calculated with coefficients
(jFoKFcj), ac and contoured at 2.6s. Fo represents the
observed structure factors, Fc represents the calculated
structure factors, ac the calculated phases for which the
ligand contributions have been omitted. The atoms are
depicted in ball-and-stick mode and colored according to
type. Spheres represent: N, blue; P, pink; O, red; and cyan,
water. Green and yellow mark the PTR1 residues and
cofactor, respectively. (a) An ethylene glycol molecule that
replaces some ordered water molecules is colored purple
and the position of DHB is shown as a semitransparent
object.
the cofactor, and the carbonyl oxygen atoms of
Gly225 and Val180, respectively (not shown).
Though weak,27,30 these interactions assist the
association of protein with cofactor and help to
align the nicotinamide to facilitate hydride transfer
from C4. These C–H/O associations are commonly
observed in SDR structures.27
Figure 3. (a) A ribbon diagram of the LmPTR1 subunit.
Helices are colored cyan, b-strands are colored purple.
Cofactor bonds are drawn as sticks, colored according to
atom type: N, blue; P, purple; O, red; C, yellow. (b) Surface
representation of the cofactor and active site. Subunit A is
gray, blue identifies Arg287 0 contributed from subunit D,
and the cofactor is depicted in stick mode with the
associated surface colored as above. Selected residues are
shown, and the position of Val230 is labeled.
carbonyl groups of Ser227 form hydrogen bonds
with the carboxyamide group of the nicotinamide,
which in turn interacts with the nearby cofactor
phosphate group. The C2, C4, and C6 groups of the
nicotinamide participate in three C–H/O hydrogen bonds, interacting with a phosphodiester O of
Substrate and product complexes
The ternary complex structures with biopterin,
DHB and THB all have the pterin ligands bound in
the same orientation and participating in virtually
identical interactions with the enzyme and cofactor
(Figures 4 and 5) as described for the DHB
complex.16 An interesting observation from early
studies on DHFR is that the pteridine moiety of
MTX binds in the active site in an orientation
different from that adopted by the substrate DHF.31
Likewise, in PTR1, the substrates bind with the
pteridine rotated about the N2–N5 axis by 1808
relative to MTX.
Average surface-accessible areas have been
determined using a solvent probe of radius 1.4 Å.
110
Structures of PTR1 Ligand Complexes
Figure 5. A representation of
biopterin binding to the LmPTR1
active site. Broken lines represent
potential hydrogen bonds and the
blocks of color are the same as in
Figure 3.
Figure 6. CB3717 binding to LmPTR1. (a) Stereoview showing the omit difference density map (contoured at 2.4s) as
described for Figure 3. The same color scheme as Figure 3 is used with the addition that the glutamate tail of the
inhibitor is red and broken lines represent potential hydrogen bonds. Phe113 is depicted as a semitransparent object.
(b) A representation of the active site with the ligand in similar fashion to Figure 4.
Structures of PTR1 Ligand Complexes
The cofactor, when removed from the protein
model has an accessible surface area (ASA) of
915 Å2. In the binary complex the ASA drops to
about 110 Å2, and a further decrease to 65 Å2
accompanies binding of substrate/product. The
relatively small pterin substrates or products have
ASA values of about 410 Å2 in isolation and around
50 Å2 when bound in the active site formed by
PTR1:NADPC. Almost 90% of the molecular surface is occluded by interactions head-on with the
cofactor, sideways with Arg17 and Tyr181, and
above and below with the nicotinamide and Phe113,
respectively (Figures 4 and 5).
These crystal structures confirm that the substrates bind in a single orientation and, in conjunction with biochemical data, clearly define a
sequential two-step reduction mechanism.14,16 The
first catalytic step resembles that of other SDR
family members and exploits three residues to (a)
position the nicotinamide moiety of the cofactor
NADPH for hydride transfer (Lys198), (b) acquire a
proton from the solvent (Asp181), and (c) pass this
on to the substrate (Tyr194). The second reduction
step, which occurs on the opposite side of the
pterin, is similar to that postulated for DHFR.
Nicotinamide again provides a hydride ion and an
activated water molecule supplies the proton.
The PTR1 catalytic center appears to be relatively
rigid, with the correct alignment of functional
groups to support catalysis. Seven residues
(Arg17, Ser111, Phe113, Asp181, Tyr194, Lys198
and Arg287 0 ) are important for creating the active
site, substrate binding, or implicated in catalysis
(Figure 5). These residues and the cofactor are
amongst the most ordered parts of the structure, as
indicated by the lower than average temperature
2
factors (or B-factors, where BZ 8p2 U , and U is the
mean displacement of atoms along the normal to
the reflecting planes, data not shown).
CB3717 binding to PTR1
CB3717 is an N10-substituted conjugated pterinlike molecule similar to MTX, but with two
significant differences, 2-amino-4-oxoquinazoline
and prop-2-inyl (propargyl) replacing the 2,4diaminopteridine and N10-methyl substituents of
MTX, respectively (Figure 2). The presence of the
4-oxo group on CB3717 suggested that the quinazoline head group should adopt the pterin-like
binding mode as opposed to the MTX binding
mode and this has indeed proven to be the case
(Figure 6).
The quinazoline ring is sandwiched between the
cofactor nicotinamide and Phe113 and positioned
such that all functional groups participate in
hydrogen bonding interactions. These interactions
are with the O2 0 hydroxyl group and one of the
phosphate oxygen atoms of the cofactor together
with side-chains of Arg17 and Ser111 (Figure 6). The
N1 atom of CB3717 is 3.6 Å distant from Tyr194 OH,
and this may represent a weak interaction. The
propargyl moiety forms only a few van der Waals
111
interactions with the side-chains of Phe113 and
Leu188, in contrast to the extensive van der Waals
interactions observed for this moiety when CB3717
is bound in the active site of TS. The benzyl group,
again using van der Waals interactions, associates
with Leu188, Leu226, Leu228 and His241. At the
glutamate tail there is a single weak potential
hydrogen bond accepted from Tyr283. Ethylene
glycol is observed binding between the inhibitor,
with which it makes van der Waals contacts, and a
polar section of the active site formed by Asp181
and Arg287 0 at the same site noted in the binary
cofactor complex and the MTX complex.16 This
ethylene glycol replaces several of the highly
conserved and ordered water molecules observed
in the other structures (Figure 4). Replacement of
N5 and N8 of the ligand, as present on the
substrates, with carbon atoms ablates polar interactions with Tyr194 and a water molecule adjacent
to the Arg17 guanidine and the backbone amide
nitrogen atom at Val230.
The para-aminobenzoate (pABA)-glutamate tail
of the inhibitor is poorly ordered, as reflected by the
thermal parameters, which exceed 80 Å2, and less
well-defined electron density associated with this
group. This feature is shared with MTX and is likely
a consequence of the shape of the PTR1 ligandbinding cavity, which is relatively wide just above
the catalytic center with a concomitant lack of
specific interactions formed between the ligand and
the enzyme. A biological consequence of this
spacious entry into the active site is that PTR1 can
process a wide range of pteridine compounds,
including conjugated pteridines such as DHF.
MTX is a more potent inhibitor of DHFR (IC50
5 nM) than of PTR1 (IC50 1.1 mM),16 mainly because
the pABA group makes extensive interactions,
including two direct salt-bridge associations with
basic residues in the DHFR active site.11 The lack of
affinity of PTR1 for the pABA-Glu tail of MTX is
demonstrated indirectly by the inhibitor TAQ,
which is essentially just the pterin-component of
MTX. TAQ adopts an identical orientation in the
active site of PTR1 and binds with an IC 50
comparable to that of MTX.17
In a similar fashion, CB3717 is a highly potent
inhibitor of TS with IC50 values in the range of
30–60 nM.32 When bound to E. coli TS, for example,
the glutamate moiety binds directly to His51 and
Ser54, and exploits numerous solvent-mediated
hydrogen bonding links to the enzyme to form a
stable complex (Protein Data Bank (PDB) code
1AN5).33 In addition, the propargyl and pABA
groups participate in extensive van der Waals
interactions with hydrophobic side-chains in TS.
The molecular conformation of CB3717 in the
PTR1 ternary complex has an ASA of 200 Å2,
increasing to 765 Å2 when the inhibitor is considered in isolation. When bound in the E. coli TS
ternary complex the inhibitor in isolation, with a
slightly different conformation, has an ASA of
710 Å2 , which decreases to only 90 Å 2 upon
binding. The inhibitor is bound more tightly by
112
Structures of PTR1 Ligand Complexes
Figure 7. TOP binding to LmPTR1. Stereoview showing the omit difference density map (contoured at 2.6s). DHB,
Phe113 and Leu188 are depicted as semitransparent objects. The orientation and conventions used in previous Figures
are reproduced here.
E. coli TS, assisted by large-scale adjustments of
protein secondary structure to form a closed
conformation about the drug. The increased ASA
observed for the drug in the PTR1 complex is
principally due to the tail of the molecule stretching
out onto the surface of the ligand-binding site. The
new structure and comparisons with the MTX and
TAQ complexes suggests that the pABA-Glu moiety
contributes little to binding PTR1 and could be
dispensed with in the design of novel and potent
inhibitors.
4-Oxo-2-amino quinazolines like CB3717 have
poor affinity for PTR1, with IC50 values O10 mM in
all cases that have been examined.15 This affinity is
similar to that of the pteridine substrates for PTR1.
The loss of the polar interactions with N5 and N8 in
the quinazolines cannot fully account for the poor
affinity of PTR1 for CB3717, because several
6-substituted 2,4-diaminoquinazolines bind to the
PTR1-NADPH complex with values of K d!
20 nM.15 Replacement of the 4-oxo with 2-amino
functionality seems to be a necessary but not
sufficient feature for potent inhibition of PTR1 by
quinazolines.
Trimethoprim binding to PTR1
TOP is a potent DHFR inhibitor and structural
studies have revealed that it binds to the enzyme
with the diaminopyrimidine moiety mimicking a
pterin ring system with the TOP N4 replacing the
pterin N8.24,34 The association of TOP with PTR1,
with an IC50 of 12 mM,15 is very different. Here, the
diaminopyrimidine is displaced by approximately
2.5 Å from the pterin-equivalent binding position
(Figure 7). This can be explained by steric restrictions imposed on the trimethoxyphenyl tail of the
inhibitor, in particular by Phe113 and Leu188. This
prevents the diaminopyrimidine group from inserting deeply enough into the binding pocket to
exploit the hydrogen bonding opportunities with
the protein and cofactor, and to maximize van der
Waals interactions with Phe113 and the nicotina-
mide. Only one well-defined direct hydrogen
bond is formed between TOP and PTR1 via
donation from N2 to the hydroxyl group of
Tyr194. A water molecule is trapped between the
diaminopyrimidine and cofactor participating in
hydrogen bonds with TOP N2, a cofactor phosphate
and Ser111. This water molecule occupies the
position of N2 observed in the pterin complexes.
TOP N1 is approximately equidistant (3.6 Å)
from Asp181 OD1 and Tyr194 OH, and this
may reflect the presence of a weak three-center
hydrogen bond.
The 2,4-diaminopyrimidine ring of TOP carries
four functional groups, two hydrogen bond donor
(N2 and N4) and two acceptor groups (N1 and N3).
When TOP binds to DHFR it makes use of all
functional groups to form hydrogen bonds directly
with the protein or mediated by a well-ordered
water molecule.24 The pyridine N3 atom is worth
further comment about an interaction, which so far
has been overlooked. This group is 3.4 Å distant
from Trp6 Ca placed to participate in a C–H/N
interaction similar to the C–H/O type of interactions discussed previously. Although this is not a
strong attractive force, it would reduce the
significant destabilization likely to occur in the
presence of an unfulfilled hydrogen bond
acceptor.30
There is rotational freedom about the C5–C7 and
C7–C8 bonds in TOP (Figure 2), and conformational
variability contributes to this molecule’s enhanced
inhibition of bacterial DHFR, compared to the
eukaryotic enzymes, as the drug adapts to binding
in the active site.5,35 Also, in common with TS,
conformational changes of DHFR (in particular of a
loop at the edge of the catalytic center and the two
sub-domains of the enzyme) accompany the binding of ligands in the active site.36,37 In contrast, the
residues that form the substrate-binding site of
PTR1 are well ordered and when cofactor is present,
the amount of space available where a ligand can
bind is highly restricted. This suggests that the
limited conformational flexibility displayed by TOP
Structures of PTR1 Ligand Complexes
would not improve binding to PTR1, due in
particular to the side-chain positions of Leu188,
Leu226, Leu229 and His241.
The molecular conformation of TOP in the PTR1
ternary complex has an average ASA of 80 Å2
and 500 Å2 when considered in isolation. In the
M. tuberculosis DHFR ternary complex (PDB code
1DG5), the inhibitor in isolation adopts a slightly
different conformation (not shown) but with a
similar ASA of 500 Å2, which decreases to 70 Å2
upon binding. Although the ASA for TOP bound to
PTR1 or DHFR is comparable, the loss of hydrogen
bonding interactions together with reduced van der
Waals associations contribute to the differences in
inhibition against the two enzymes.
The inhibitory properties of 15 2,4-diaminopyrimidine derivatives against PTR1 have been
assayed15 and the TOP complex now provides a
clear explanation of the trends observed. Most 2,4diaminopyrimidines with phenyl substituents at C5
and bulky substituents at C6 (Figure 2) are very
poor inhibitors, with IC50 values ranging from
50 mM to in excess of 300 mM, most likely because
of steric restrictions in the active site imposed by
Phe113 and Leu188.
Two 2,4-diaminopyrimidines with IC50 values of
w3 mM have chlorophenyl substituents at C5 and
ethyl groups at C6. Some of the molecules
previously tested would, in addition to steric
repulsion, suffer by virtue of positioning a hydrophobic group in a polar region of the active site near
Asp181 and Arg287 0 . However, 2,4-diaminopyrimidines with 5-propylphenyl, or 5-butylphenyl substituents were improved inhibitors with IC50 values
!1.5 mM, with the most potent of this series having
a Kd value of w30 nM for the PTR1-NADPH
complex. Flexible C3 or C4 tethers attached to either
C5 or C6 (Figure 2) might circumvent the steric
restrictions imposed by the PTR1 active site and
allow the 2,4-diaminopyrimidine to bind further in
toward the cofactor, perhaps adopting the pterinlike binding mode. It would be of interest to
investigate if adjustment of the substituent conformations to optimize van der Waals interactions with
the enzyme could also occur. Such a mode of
inhibition offers scope for future design of more
potent inhibitors.
Concluding Remarks and Implications
for Inhibitor Design
The use of combinations of drugs, each with
independent modes of action, has been shown to
improve efficacy in some treatments without
increasing toxicity, and with the additional and
significant benefit of providing mutual protection
against drug resistance. Of particular note with
respect to parasitic infections are combinations of
dapsone with chloroproguanil or pyrimethamine to
combat malaria.38 For the treatment of bacterial
infections, the combination of TOP with sulfonamide drugs is often used.39 With respect to the
113
development of antifolate drugs to treat trypanosomatid infections, we note three valuable enzyme
activities, DHFR-TS and PTR1. A determination of
whether it is reasonable to expect that a single
molecule can be obtained that is a potent inhibitor
of two or more of these enzymes, or whether
different compounds will be needed for the
different enzymes, would indicate how to most
productively pursue inhibitor discovery.
Gangjee et al. have shown that dual inhibition of
DHFR and TS is feasible40 and, as discussed, MTX is
an inhibitor of both DHFR and PTR1. Furthermore,
a 2,4-diaminoquinazoline identified previously
inhibited both PTR1 and DHFR with potency of
w30 nM, although that compound, like MTX, also
inhibited human DHFR.15 On the basis of steric
considerations and arrangement of functional
groups, the TS, DHFR, and PTR1 active sites all
have some similarities, yet each has distinctive
features. In structures of ternary complexes of TS
and PTR1, the substrates or inhibitors display
extensive interactions with the enzyme and with
the other ligand. For example, in PTR1 complexes
there are often hydrogen bonds and strong
p-stacking associations involving Phe113 and the
nicotinamide cofactor. These structural features
mandate the ordered binding seen both with TS
and PTR1. DHFR has different structural features
that allow it to bind its ligands in either order, albeit
with a kinetic preference. DHFR and TS undergo
extensive conformational changes upon ternary
complex formation, whereas PTR1 is more rigid.
DHFR and TS structures show strong associations
with the pABA group of various ligands, whilst in
PTR1 this group is bound only loosely. PTR1 offers
steric restrictions preventing access, for some
compounds, to the optimum pterin-like binding
position. Both DHFR and PTR1 can each bind
tightly the 2,4-diaminopteridine, 2,4-diaminoquinazolines, or 2,4-diaminopyrimidine scaffolds, though
in a reverse orientation from that observed for the
corresponding 4-oxo-2-amino molecules. TS, on the
other hand, has no or very poor affinity for 2,4diaminopteridine and 2,4-diaminoquinazolines.
There are already a significant number of
inhibitors of DHFR-TS with well-characterized
pharmacokinetics and impressive affinity for their
targets. Since TS is one of the most highly conserved
of enzymes, it might seem daunting to consider
trying to inhibit a pathogen enzyme, given the
deleterious affects likely to occur on human TS.
Nevertheless, recent studies have shown conclusively that specificity towards microbial TS
inhibition can be achieved,32,33 thereby providing
encouragement for such an approach. Certainly,
less toxic inhibitors of human TS might prove useful
for antimicrobial drug development. However, at
this stage we suggest that the priority should be
development of PTR1 inhibitors likely to complement existing drugs that target DHFR, or that are
capable of inhibiting both PTR1 and DHFR, with
requisite selectivity against human DHFR.
Future work developing such PTR1 inhibitors can
114
Structures of PTR1 Ligand Complexes
be based on our enzyme–ligand complexes by
extending from those sections of the molecules
shown to be important for binding. The presence of
a solvent-filled cavity, occupied by conserved water
molecules (or ethylene glycol), and lined by
hydrophilic residues (Asp181 and Arg287 0 )
suggests a suitable region to target by modification
of the pterin, quinazoline or 2,4-diaminopteridine
framework.
while partially ordered side-chains were built in a
chemically sensible conformation and assigned low
occupancies. Some density features have been interpreted
as molecules of the cryo-protectant ethylene glycol. For
the TOP complex, convincing electron density for the
inhibitor was associated with two out of the four subunits
in the asymmetric unit. Model geometries were analyzed
with PROCHECK48 and WHAT_CHECK,49 and statistics
are given in Table 1. Figures were produced using
MOLSCRIPT,50 PyMOL†, and Raster3D.51
Experimental
Protein Data Bank accession codes
Materials
Biopterin and THB were obtained from Schirks
Laboratories (Jona, Switzerland) and other reagents
were purchased from Sigma-Aldrich. CB3717 was provided by Professor A. Jackman.
Crystallization and data collection
Recombinant L. major PTR1 was purified and crystallized following published methods.16,17,41 Briefly, ligand
solutions were prepared fresh (1 mM cofactor, 1 mM
ligand in 20 mM dithiothreitol buffered with 20 mM
sodium acetate (pH 5.3)) and mixed in tenfold excess
with the enzyme solution in the same buffer. The enzyme–
ligand mixtures were incubated on ice for an hour, and
concentrated to 15 mg mlK1 of protein. Crystals were
obtained using the hanging-drop method for vapor
diffusion at 293 K from drops containing 2 ml of the
enzyme–ligand complex plus 2 ml of a reservoir, comprising 6–20% (w/v) monomethoxy-polyethylene glycol
5000, 100 mM sodium acetate buffer (pH 5.5) and
40–300 mM calcium acetate. Crystals formed mainly as
clusters of fragile orthorhombic rods and grew up to
0.5 mm in length over several days. All crystals in this
present study display space group P212121 and are
isomorphous with the PTR1:NADPC:DHB complex.16
Crystals were cryo-protected with ethylene glycol and
cooled to 100 K in a stream of nitrogen gas and diffraction
data were recorded at the Synchrotron Radiation Source
(S.R.S.) Daresbury Laboratory, UK (biopterin and TOP
complexes) and the European Synchrotron Radiation
Facility, (E.S.R.F.), Grenoble, France (CB3717, binary and
THB complexes). Data were processed and scaled using
HKL,42 and then truncated and converted with CCP4
programs.43 Details are given in Table 1.
Refinement
The asymmetric unit contains a tetramer (subunits are
labeled A to D). The refinement protocol started with the
PTR1:NADPC:DHB complex (PDB code 1E92), and
applied rigid-body refinement (REFMAC) 44 and
simulated annealing (CNS)45 to that model. Cycles of
model-map inspection (O),46 identification of solvent
molecules and restrained least-squares refinement
(initially with CNS then REFMAC) followed. Coordinates
and topologies for ligands were obtained from HIC-Up.47
Throughout the refinement, non-crystallographic symmetry (NCS) restraints were applied to the four subunits
in the asymmetric unit, excluding residues around
several flexible loops and those that occupied clearly
different conformations in some chains. Completely
disordered residues have been omitted from the models,
Coordinates and structure factors have been deposited
with the Protein Data Bank (with accession codes 2BF0,
2BF7, 2BFP, 2BFA, and 2BFM) and are available for
immediate release upon acceptance.
Acknowledgements
This work was funded by grants from the
Wellcome Trust (WNH) and the NIH (AI 21903,
SMB). We thank S.R.S. Daresbury laboratory and
E.S.R.F. for synchrotron beam time and their staff for
excellent support, Charlie Bond, Alice Dawson,
David Gourley, and James Luba for advice and Ann
Jackman for a sample of CB3717.
References
1. Webster, L. T., Jr (1990). Leishmaniasis, trypanosomaisis and other protozoal infections. In The
Pharmaceutical Basis of Therapeutics (Gilman, A. G.,
Rall, T. W., Nies, S. & Taylor, P., eds) 8th edit., pp.
1008–1017, Pergammon Press, New York chapt. 43.
2. Fairlamb, A. H. (2003). Chemotherapy of human
African trypanosomiasis: current and future
prospects. Trends Parasitol. 11, 488–494.
3. Berriman, M. & Harris, M. (2004). Annotation of
parasite genomes. Methods Mol. Biol. 270, 17–44.
4. Beverley, S. M. (2003). Protozomics: trypanosomatid
parasite genetics comes of age. Nature Rev. Genet. 4,
11–19.
5. Blakley, R. L. (1995). Eukaryotic dihydrofolate
reductase. Advan. Enzymol. Relat. Areas Mol. Biol. 70,
23–102.
6. Bermingham, A. & Derrick, J. P. (2002). The folic acid
biosynthesis pathway in bacteria: evaluation of
potential for antibacterial drug discovery. BioEssays,
24, 637–648.
7. Carreras, C. W. & Santi, D. V. (1995). The catalytic
mechanism and structure of thymidylate synthase.
Annu. Rev. Biochem. 64, 721–762.
8. Gmeiner, W. H. (2005). Novel chemical strategies for
thymidylate synthase inhibition. Curr. Med. Chem. 12,
191–202.
9. Kompis, I. M., Islam, K. & Then, R. L. (2005). DNA
and RNA synthesis: antifolates. Chem. Rev. 105,
593–620.
† http://pymol.sourceforge.net/
115
Structures of PTR1 Ligand Complexes
10. Walsh, C. (2003). Other targets of antibacterial drugs.
In Antibiotics, Actions, Orgins, Resistance, pp. 79–88,
ASM Press, Washington, DC.
11. Knighton, D. R., Kan, C. C., Howland, E., Janson, C. A.,
Hostomska, Z., Welsh, K. M. & Matthews, D. A. (1994).
Structure of and kinetic channelling in bifunctional
dihydrofolate reductase-thymidylate synthase. Nature
Struct. Biol. 1, 186–194.
12. Bello, A. R., Nare, B., Freedman, D., Hardy, L. W. &
Beverley, S. M. (1994). PTR1: a reductase mediating
salvage of oxidized pteridines and methotrexate
resistance in the protozoan parasite Leishmania major.
Proc. Natl Acad. Sci. USA, 91, 11442–11446.
13. Nare, B., Hardy, L. W. & Beverley, S. M. (1997). The
roles of pteridine reductase 1 and dihydrofolate
reductase-thymidylate
synthase
in
pteridine
metabolism in the protozoan parasite Leishmania
major. J. Biol. Chem. 272, 13883–13891.
14. Hardy, L. W., Matthews, W., Nare, B. & Beverley, S. M.
(1997). Biochemical and genetic tests for inhibitors of
Leishmania pteridine pathways. Expt. Parasitol. 87,
157–169.
15. Luba, J., Nare, B., Liang, P. H., Anderson, K. S.,
Beverley, S. M. & Hardy, L. W. (1998). Leishmania major
pteridine reductase 1 belongs to the short chain
dehydrogenase family: stereochemical and kinetic
evidence. Biochemistry, 37, 4093–4104.
16. Gourley, D. G., Schüttelkopf, A. W., Leonard, G. A.,
Luba, J., Hardy, L. W., Beverley, S. M. & Hunter, W. N.
(2001). Pteridine reductase mechanism correlates
pterin metabolism with drug resistance in trypanosomatid parasites. Nature Struct. Biol. 8, 521–525.
17. McLuskey, K., Gibellini, F., Carvalho, P., Avery, M. A.
& Hunter, W. N. (2004). Inhibition of Leishmania major
pteridine reductase by 2,4,6-triaminoquinazoline:
structure of the NADPH ternary complex. Acta
Crystallog. sect. D, 60, 1780–1785.
18. Zhao, H., Bray, T., Ouellette, M., Zhao, M., Ferre, R. A.,
Matthews, D. et al. (2003). Structure of pteridine
reductase (PTR1) from Leishmania tarentolae. Acta
Crystallog. sect. D, 59, 1539–1544.
19. Schormann, N., Pal, B. & Chattopadhyay, D. (2001).
Expression, purification, crystallization and preliminary crystallographic analysis of recombinant
pteridine reductase of Trypanosoma cruzi. Acta
Crystallog. sect. D, 57, 1671–1673.
20. Stout, T. J., Sage, C. R. & Stroud, R. M. (1998). The
additivity of substrate fragments in enzyme-ligand
binding. Structure, 6, 839–848.
21. Finer-Moore, J., Fauman, E. B., Foster, P. G., Perry,
K. M., Santi, D. V. & Stroud, R. M. (1993). Refined
structures of substrate-bound and phosphatebound thymidylate synthase from Lactobacillus
casei. J. Mol. Biol. 232, 1101–1116.
22. Anderson, A. C., O’Neil, R. H., DeLano, W. L. &
Stroud, R. M. (1999). The structural mechanism for
half-the-sites reactivity in an enzyme, thymidylate
synthase, involves a relay of changes between
subunits. Biochemistry, 38, 13829–13836.
23. Baccanari, D. P. & Kuyper, L. F. (1993). Basis of
selectivity of antibacterial diaminopyrimidines.
J. Chemother. 5, 393–399.
24. Li, R., Sirawaraporn, R., Chitnumsub, P.,
Sirawaraporn, W., Wooden, J., Athappilly, F. et al.
(2000). Three-dimensional structure of M. tuberculosis
dihydrofolate reductase reveals opportunities for the
design of novel tuberculosis drugs. J. Mol. Biol. 295,
307–323.
25. Dale, G. E., Broger, C., D’Arcy, A., Hartman, P. G.,
26.
27.
28.
29.
30.
31.
32.
33.
34.
35.
36.
37.
38.
39.
40.
DeHoogt, R., Jolidon, S. et al. (1997). A single amino
acid substitution in Staphylococcus aureus dihydrofolate reductase determines trimethoprim resistance.
J. Mol. Biol. 266, 23–30.
Matthews, D. A., Bolin, J. T., Burridge, J. M., Filman,
D. J., Volz, K. W., Kaufman, B. T. et al. (1985). Refined
crystal structures of Escherichia coli and chicken liver
dihydrofolate reductase containing bound trimethoprim. J. Biol. Chem. 260, 381–391.
Duax, W. L., Pletnev, V., Addlagatta, A., Bruenn, J. &
Weeks, C. M. (2003). Rational proteomics I. Fingerprint identification and cofactor specificity in the
short-chain oxidoreductase (SCOR) enzyme family.
Proteins: Struct. Funct. Genet. 53, 931–943.
Oppermann, U., Filling, C., Hult, M., Shafqat, N., Wu,
X., Lindh, M. et al. (2003). Short-chain dehydrogenases/reductases (SDR): the 2002 update. Chem. Biol.
Interact. 143–144, 247–253.
Tanaka, N., Nonaka, T., Tanabe, T., Yoshimoto, T.,
Tsuru, D. & Mitsui, Y. (1996). Crystal structures of the
binary and ternary complexes of 7 alpha-hydroxysteroid dehydrogenase from Escherichia coli.
Biochemistry, 35, 7715–7730.
Leonard, G. A., McAuley-Hecht, K., Brown, T. &
Hunter, W. N. (1995). Do C–HO hydrogen bonds
contribute to the stability of nucleic acid base pairs?
Acta Crystallog. sect. D, 51, 136–139.
Charlton, P. A., Young, D. W., Birdsall, B. W., Feeney, J.
& Roberts, G. C. K. (1985). Stereochemistry of
reduction of the vitamin folic-acid by dihydrofolatereductase. J. Chem. Soc., Perkin Trans. 191, 1349–1353.
Ferrari, S., Costi, P. M. & Wade, R. C. (2003). Inhibitor
specificity via protein dynamics: insights from the
design of antibacterial agents targeted against
thymidylate synthase. Chem. Biol. 10, 1183–1193.
Stout, T. J., Tondi, D., Rinaldi, M., Barlocco, D.,
Pecorari, P., Santi, D. V. et al. (1999). Structure-based
design of inhibitors specific for bacterial thymidylate
synthase. Biochemistry, 38, 1607–1617.
Champness, J. N., Achari, A., Ballantine, S. P., Bryant,
P. K., Delves, C. J. & Stammers, D. K. (1994). The
structure of Pneumocystis carinii dihydrofolate
reductase to 1.9 Å resolution. Structure, 2, 915–924.
Matthews, D. A., Bolin, J. T., Burridge, J. M., Filman,
D. J., Volz, K. W. & Kraut, J. (1985). Dihydrofolate
reductase. The stereochemistry of inhibitor selectivity.
J. Biol. Chem. 260, 392–399.
Sawaya, M. R. & Kraut, J. (1997). Loop and
subdomain movements in the mechanism of
Escherichia coli dihydrofolate reductase: crystallographic evidence. Biochemistry, 36, 586–603.
Schnell, J. R., Dyson, H. J. & Wright, P. E. (2004).
Structure, dynamics, and catalytic function of
dihydrofolate reductase. Annu. Rev. Biophys. Biomol.
Struct. 33, 119–140.
Olliaro, P. L. & Taylor, W. R. (2003). Antimalarial
compounds: from bench to bedside. J. Expt. Biol. 206,
3753–3759.
Rosowsky, A., Forsch, R. A. & Queener, S. F. (2002).
Inhibition of Pneumocystis carinii, Toxoplasma gondii,
and Mycobacterium avium dihydrofolate reductases
by 2,4-diamino-5-[2-methoxy-5-(omega-carboxyalkyloxy) benzyl]pyrimidines: marked improvement in
potency relative to trimethoprim and species
selectivity relative to piritrexim. J. Med. Chem. 45,
233–241.
Gangjee, A., Yu, J., McGuire, J. J., Cody, V., Galitsky,
N., Kisliuk, R. L. & Queener, S. F. (2000). Design,
synthesis, and X-ray crystal structure of a potent dual
116
41.
42.
43.
44.
45.
46.
Structures of PTR1 Ligand Complexes
inhibitor of thymidylate synthase and dihydrofolate
reductase as an antitumor agent. J. Med. Chem. 43,
3837–3851.
Gourley, D. G., Luba, J., Hardy, L. W., Beverley, S. M.
& Hunter, W. N. (1999). Crystallization of recombinant Leishmania major pteridine reductase 1 (PTR1).
Acta Crystallog. sect. D, 55, 1608–1610.
Otwinowski, Z. & Minor, W. (1997). Processing of
X-ray diffraction data collected in oscillation mode.
Methods Enzymol. 276, 307–326.
Collaborative Computational Project Number 4
(1994). The CCP4 suite: programs for protein
crystallography. Acta Crystallog. sect. D, 50, 760–763.
Murshudov, G. N., Vagin, A. A. & Dodson, E. J. (1997).
Refinement of macromolecular structures by the
maximum-likelihood method. Acta Crystallog. sect.
D, 53, 240–255.
Brünger, A. T., Adams, P. D., Clore, G. M., DeLano, W.
L., Gros, P. W., Grosse-Kunstleve, R. W. et al. (1998).
Crystallography & NMR system: a new software suite
for macromolecular structure determination. Acta
Crystallog. sect. D, 54, 905–921.
Jones, T. A., Zou, J. Y., Cowan, S. W. & Kjeldgaard, M.
47.
48.
49.
50.
51.
52.
(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.
Kleywegt, G. J. & Jones, T. A. (1998). Databases in
protein crystallography. Acta Crystallog. sect. D, 54,
1119–1131.
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.
Hooft, R. W. W., Vriend, G., Sander, C. & Abola, E. E.
(1996). Errors in protein structures. Nature, 381, 272.
Kraulis, P. J. (1991). MOLSCRIPT—a program to
produce both detailed and schematic plots of protein
structures. J. Appl. Crystallog. 24, 946–950.
Merritt, E. A. & Bacon, D. J. (1997). Raster3D:
photorealistic molecular graphics. Methods Enzymol.
277, 505–524.
Cruickshank, D. W. J. (1999). Remarks about protein
structure precision. Acta Crystallog. sect. D, 55,
583–601.
Edited by R. Huber
(Received 16 May 2005; received in revised form 29 June 2005; accepted 30 June 2005)
Available online 14 July 2005