Fluorescent inhibitors for IspF, an enzyme in the non-mevalonate pathway for isoprenoid biosynthesis and a potential target for antimalarial therapy

Article abstract of DOI:10.1002/anie.200503003

Designed inhibitors (like 1) of IspF, a key enzyme in the non-mevalonate pathway for terpene biosynthesis and a potential antimalarial target, were synthesized and evaluated. Since fluorescent probes were introduced in these ligands, their affinity toward

Full text of DOI:10.1002/anie.200503003

Angewandte
Chemie
Inhibitors
DOI: 10.1002/anie.200503003
Fluorescent Inhibitors for IspF, an Enzyme in the
Non-Mevalonate Pathway for Isoprenoid Biosyn-
thesis and a Potential Target for Antimalarial
Therapy**
Christine M. Crane, Johannes Kaiser,
Nicola L. Ramsden, Susan Lauw, Felix Rohdich,
Wolfgang Eisenreich, William N. Hunter,*
Adelbert Bacher,* and François Diederich*
Malaria remains a leading cause of illness and death in
endemic areas, infecting 300–500 million and killing 2.5–5
million people annually. As a result of rapidly emerging
[*] N. L. Ramsden, Prof. Dr. W. N. Hunter
Division of Biological Chemistry and Molecular Microbiology
School of Life Sciences
MSI/WTB Complex
University of Dundee
Dow Street, Dundee DD15EH (UK)
Fax: (+44)138-232-2558
E-mail: w.n.hunter@dundee.ac.uk
Dr. J. Kaiser, S. Lauw, Dr. F. Rohdich, Dr. W. Eisenreich,
Prof. Dr. A. Bacher
Lehrstuhl für Organische Chemie und Biochemie
Technische Universität München
Lichtenbergstrasse 4, 85748 Garching (Germany)
Fax: (+49)89-289-13363
E-mail: adelbert.bacher@ch.tum.de
C. M. Crane, Prof. Dr. F. Diederich
Laboratorium für Organische Chemie
ETH Hönggerberg
HCI, 8093 Zürich (Switzerland)
Fax: (+41)44-632-1109
E-mail: diederich@org.chem.ethz.ch
[**] We gratefully acknowledge financial support for this work from the
following agencies: the Hans-Fischer Gesellschaft e.V. (to J.K.), The
Wellcome Trust (to W.N.H.), the Biotechnology and Biological
Sciences Research Council (UK) (to W.N.H.), and InPharmatica (to
W.N.H.). The authors also thank Prof. D. Arigoni and Dr. L. Kemp for
help and advice. IspF=2C-Methyl-d-erythritol 2,4-cyclodiphosphate
synthase.
Supporting information for this article is available on the WWW
under http://www.angewandte.org or from the author.
Angew. Chem. Int. Ed. 2006, 45, 1069 –1074
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
1069

Communications
multiple-drug-resistant strains of the various Plasmodium
residues lining these pockets, see Figure 1b). The latter
pocket also contains a tetrahedrally coordinated ZnII ion.
parasites, the search for new therapies with novel modes of
action is of urgent necessity.[1] Enzymes in the non-mevalo-
nate pathway, utilized for assembling the C5 precursors to
terpenes, isopentenyl diphosphate (IPP) 1 and dimethylallyl
diphosphate (DMAPP) 2, were recently identified as targets
for antimalarial and antimicrobial drugs.[2,3] The non-meval-
onate pathway[4] is characterized by the condensation of
pyruvate 3 and glyceraldehyde 3-phosphate 4 (Scheme 1) and
Scheme 1. The non-mevalonate pathway for the biosynthesis of the C5
precursors to terpenes, IPP 1 and DMAPP 2 (see the Supporting
Information). DXS=1-deoxy-d-xylulose-5-phosphate synthase,
CMP=cytidine 5’-monophosphate.[4–8]
Figure 1. Binding mode of 7 in the ternary complex of IspF with ZnII as
determined by X-ray crystallography to 2.3-Š resolution.[17] Ligand
binding shown occurs at the interface of subunits A and C (primes
designate residues from an adjacent monomer). Five of the six
independent active sites of IspF show the depicted binding conforma-
tion. a) Schematic representation of the binding mode of 7 (in blue)
with potential hydrogen bonds depicted as red dashed lines. b) Fo-Fc
omit electron density map surrounded by residues within 4 Š, con-
toured at the 2s level. Color code: C gray, O red, N blue, S green, P
yellow, ZnII pink ball.
is the sole source for 1 and 2 in plastids of higher plants[5a–c]
and in many bacteria[4b,5c,6] including some responsible for
serious diseases such as Mycobacterium tuberculosis[6b] and
the protozoan Plasmodium parasites (Apicomplexa).[3] Since
mammals exclusively utilize the mevalonate pathway,[6] the
development of small-molecule lead compounds inhibiting
the enzymes of the non-mevalonate pathway may be a key
step towards new antimalarial drugs.[1b,2,3]
We have chosen the enzyme IspF (2C-methyl-d-erythritol
2,4-cyclodiphosphate synthase, ygbB) as a target for struc-
ture-based lead generation.[7,8] IspF is the fifth enzyme in the
non-mevalonate pathway and catalyzes the cyclization of 4-
diphosphocytidyl-2C-methyl-d-erythritol 2-phosphate (5), to
the key cyclic diphosphate intermediate, 2C-methyl-d-eryth-
ritol 2,4-cyclodiphosphate (6). Published crystal structures
(Protein Data Bank (PDB)[8,9] codes 1GX1 and 1JY8) show
IspF to be a C3-symmetric homotrimer. The topologically
equivalent active sites are located at the interfaces of adjacent
subunits. The rigid, well-conserved “Pocket III” of one
monomer binds the cytidine moiety of 5, and the larger,
more flexible “Pocket II” of an adjacent monomer binds the
2C-methyl-d-erythritol moieties of 5 and 6 (for the protein
Inhibitors 7–9, occupying both pockets, were designed with
the help of the molecular modeling software MOLOC.[10,11] In
the absence of known inhibitors, we chose in a first step to
maintain the CDP moiety of the natural substrate to occupy
Pocket III while connecting the diphosphate by means of an
appropriately sized linker to an aromatic residue, for occu-
pation of the hydrophobic cleft of Pocket II, defined by
Leu76’, Phe61’, and Ile57’ (Figure 1). Since at a first stage, we
were also interested in developing a fluorescence-based
enzyme inhibition assay, fluorescent anthranilate (2-amino-
benzoate) and dansyl (5,5-dimethylaminonaphthalenesulfa-
moyl) residues were chosen as aromatic moieties reaching
into Pocket II.[12]
The synthesis of target compound 7 is shown in Scheme 2
(for the synthesis of compounds 8 and 9, see the Supporting
Information). Reduction of nitro derivative 10 afforded
1070 www.angewandte.org
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2006, 45, 1069 –1074

Angewandte
Chemie
Figure 2. Typical fluorescence titration curves of E. coli IspF (6.25 mm)
^
*
with inhibitors 7 (&), 8 ( ), and 9 ( ), measured from 0–24 mm in
0.3-mm aliquots (0.6-mm steps shown for clarity).[15]
ation constants were obtained by nonlinear least-squares
approximation and are given in Table 1. All three fluorescent
probes show activities in the lower double-digit micromolar
Table 1: Affinities of ligands against E. coli IspF measured by fluores-
cence (7–9) and 13C NMR assays (9, CMP, CDP).
Compound
Kd [mm][a]
IC50 [mm][b]
7
8
9
36ꢀ5
23ꢀ2
15ꢀ0.3
3.0
15.0
7.3
CMP
CDP
[a] Determined at 258C; for details, see ref. [15]. [b] Determined at 378C;
Scheme 2. Synthesis of fluorescent ligand 7. a) Zn, AcOH, MeOH,
08C!208C, 5 h, quant.; b) HOPO(OBn)2, DIAD, Ph3P, THF, 208C,
63%; c) H2, 10% Pd/C, CH2Cl2/EtOH, 7 h, 208C, 98%; d) Et3N,
cytidine 5’-monophosphomorpholidate 4-morpholine-N,N’-dicyclohexyl-
carboxamidine salt, 1H-tetrazole, pyridine, 208C, 48 h, then DOWEX
ion-exchange (NH4+), 45%. THF=tetrahydrofuran, Bn=benzyl,
DIAD=diisopropyl azodicarboxylate.
for details, see ref. [16] and the Supporting Information.
range. In parallel, the IC50 values (IC50 = concentration of
inhibitor at which 50% of the maximum initial velocity is
observed) for 9, cytidine 5’-diphosphate (CDP), and cytidine
5’-monophosphate (CMP) were determined in enzymatic
assays monitored by 13C NMR spectroscopy (see the Support-
ing Information).[16] These results show inhibitor 9 has a lower
IC50 value (3.0 mm) than both CDP (7.3 mm) and CMP
(15.0 mm), suggesting that binding free enthalpy is gained
upon introduction of the aromatic fluorophore on inhibitor 9.
The participation of the fluorescent probes in the binding to
the E. coli enzyme, as suggested by the fluorescence assays,
was subsequently confirmed by X-ray crystal structure
analyses of inhibitors 7 and 9 bound to IspF.
The structures of the ternary complexes of 7 (Figure 1)
and 9 (see the Supporting Information) with tetrahedrally
coordinated ZnII in the active site of E. coli IspF were
determined to 2.3-Š and to 2.5-Š resolution, respectively.[17]
The complex of 7 consists of six subunits (two trimers) in the
asymmetric unit, and the complex of 9 consists of one subunit.
The crystal structures confirm that the cytidine moieties in
both complexes are firmly anchored in Pocket III as pre-
dicted. There are four highly conserved hydrogen bonds
between the nucleobase, p-stacked against Ala131, and the
backbone from Ala100 to Leu106 (see Figure 1a and the
Supporting Information).[8,9] The ribose moiety is further held
by hydrogen bonds between the ribose C(2’) and C(3’)
hydroxyl groups and the carboxylate of Asp56’. The catalytic
ZnII ions in both structures are tetrahedrally coordinated to
anthranilate 11, which was phosphorylated using a modified
Mitsunobu reaction[13] to furnish phosphotriester 12. Removal
of the benzyl residues by hydrogenation provided the free
phosphate 13. The desired diphosphate was obtained using a
modified Moffatt condensation of 13 with cytidine 5’-mono-
phosphomorpholidate 4-morpholine-N,N’-dicyclohexylcar-
boxamidine salt, catalyzed by 1H-tetrazole.[14] Ion-exchange
+
chromatography (DOWEX 50 WX8 (NH4 )) gave the
diammonium salt of 7, which was purified by column
chromatography on cellulose and fully characterized (see
the Supporting Information).
Fluorescence binding titrations were performed to deter-
mine the dissociation constants, Kd, for the complexes of IspF
from Escherichia coli with ligands 7–9.[15] In the presence of
the enzyme, the emission of the fluorescent probes is shifted
hypsochromically and increases in intensity (hyperchromism),
which is most likely attributed to complexation in an
environment of reduced polarity. Affinity data were found
to depend substantially on the concentration of ZnII ions
present in solution; therefore Zn(OAc)2 was added in a
concentration sufficient to ensure saturation of the enzyme.
Typical titration curves[15] are depicted in Figure 2. Dissoci-
Angew. Chem. Int. Ed. 2006, 45, 1069 –1074
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
1071

Communications
the side chains of His10’, His42’, and Asp8’, the fourth
structures of the ternary complex IspF-ZnII-7 and the
quaternary complex IspF-ZnII-MnII-CDP (PDB code
1GX1).[8b] On the other hand, His34’ assumes the “closed”
conformation in the latter complex to interact with the b-
phosphate of CDP (see the Supporting Information).
The crystals of the ternary complex of IspF bound to 9 and
ZnII contain a single monomer in the asymmetric unit.
Interactions in the cytidine pocket mirror those seen for the
complex with 7 (see the Supporting Information). His34’
adopts an “open” conformation. A superimposition of the X-
ray crystal structures of the enzyme bound to 7 and 9 (see the
Supporting Information) provides additional evidence for the
large conformational flexibility of the hydrophobic cleft of
Pocket II. Compared to the complex of 7, the lipophilic
residues Ile56’, Phe61’, Phe68’, and Leu76’ interacting with 9
have undergone a considerable reorientation to accommo-
date the larger dansyl residue and engage in hydrophobic
enzyme–ligand contacts. The Fo-Fc omit electron density map
of bound 9 shows only weak electron density about the dansyl
ring and elevated thermal parameters suggest a large degree
of flexibility (see the Supporting Information).
coordination site being occupied by the b-phosphate of the
inhibitors. The structures confirm our predicted binding mode
with the anthranilate and dansyl fluorophores bound in
Pocket II.
Two different conformations are observed for the six
independent active sites (arbitrarily designated A–F) of IspF
in the crystal of the ternary complex with ligand 7 and ZnII. In
active sites B–F, the ligand orientation and interactions with
the protein are well conserved (see the Supporting Informa-
tion). The CDP moiety of 7 takes a similar orientation in all
six active sites A–F. The imidazole of His34’ in active site A is
observed to be in a “closed” (into the active site) conforma-
tion, in contrast to the “open” one seen in active sites B–F.[18]
Pocket II in structure A shows a higher degree of flexibility,
which is mirrored in much larger thermal parameters (see the
Supporting Information).
Figure 3 shows the superimposition of the X-ray crystal
structures of the ternary complex of IspF-ZnII-7 (conformers
B–F) with the previously reported structure of the quaternary
In conclusion, we have reported the synthesis, biological
evaluation, and co-crystal structures of the first designed
inhibitors of IspF. By introducing fluorescent probes in
inhibitors 7–9, we could determine their affinity towards
IspF by fluorescence binding titrations. Efforts to validate the
utility of the fluorescent inhibitors as standards in rapid and
economic affinity assays are ongoing. The predicted binding
modes of 7 and 9 were largely confirmed by X-ray crystal
structure analysis. Although the hydrophobic region in
Pocket II apparently is very flexible, its occupation by
aromatic rings leads to a measurable gain in binding free
enthalpy. The results provide fruitful guidance for the future
development of new inhibitors of the non-mevalonate path-
way, by structure-based design, with the promise to ultimately
generate new classes of antimalarials and other antimicro-
bials.
Received: August 23, 2005
Published online: January 3, 2005
Keywords: fluorescence · inhibitors · IspF · non-mevalonate
.
pathway · protein crystal structure analysis
Figure 3. Superimposition of the X-ray crystal structures of the ternary
complex of IspF-ZnII-7 (conformers B–F; C atoms: light-green) with the
previously reported structure of the quaternary complex of IspF with
CMP, 2C-methyl-d-erythritol 2,4-cyclodiphosphate (6), and ZnII (PDB
code 1JY8;[8a] C atoms: gray). Color code: O red, N blue, S red-orange,
P purple, ZnII ion: purple ball.[17]
[1] a) U. Weiss (Ed. ), “Malaria”, in Nature 2002, 415, 669–715; b) L.
Schofield, M. C. Hewitt, K. Evans, M.-A. Siomos, P. H. See-
berger, Nature 2002, 418, 785 – 789; c) J. K. Baird, N. Engl. J.
Med. 2005, 352, 1565 – 1577.
[2] Fosmidomycin (OHC-N(OH)-(CH2)3-PO3H2) is a potent inhib-
itor of 1-deoxy-d-xylulose-5-phosphate reductoisomerase (IspC)
in the non-mevalonate pathway. For some selected publications,
see: a) J. Wiesner, R. Ortmann, H. Jomaa, M. Schlitzer, Angew.
Chem. 2003, 115, 5432 – 5451; Angew. Chem. Int. Ed. 2003, 42,
5274 – 5293; b) T. Kuzuyama, T. Shimizu, S. Takahashi, H. Seto,
Tetrahedron Lett. 1998, 39, 7913 – 7916; c) M. A. Missinou, S.
Borrmann, A. Schindler, S. Issifou, A. A. Adegnika, P.-B.
Matsiegui, R. Binder, B. Lell, J. Wiesner, T. Baranek, H.
Jomaa, P. G. Kremsner, Lancet 2002, 360, 1941 – 1942; d) B.
Lell, R. Ruangweerayut, J. Wiesner, M. A. Missinou, A.
Schindler, T. Baranek, M. Hintz, D. Hutchinson, H. Jomaa,
complex of IspF with CMP, 6, and ZnII (PDB code 1JY8).[8a]
In both structures, His34’ adopts the “open” conformation.
The anthranilate nicely superimposes with the hydrophobic
region of cyclodiphosphate 6. The loop Leu60’–Phe68’, which
shapes the hydrophobic cleft of Pocket II, shows substantial
flexibility, approaching the ligands to varying degrees in the
two complexes. Interestingly, this flexible loop adopts a nearly
identical orientation in the superimposition of the crystal
1072
www.angewandte.org
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2006, 45, 1069 –1074

Angewandte
Chemie
P. G. Kremsner, Antimicrob. Agents Chemother. 2003, 47, 735 –
738.
where Irel is the relative fluorescence intensity at the recorded
emission wavelength, Imax the maximal fluorescence enhance-
ment of bound inhibitor at saturation, and Fo is the concentration
of the fluorophore. Background emissions of 7–9 (from 0–24 mm)
as well as of IspF in 50 mm Tris hydrochloride, pH 8.2, containing
10 mm MgSO4 and 2 mm Zn(OAc)2, were subtracted to obtain
the saturation curves shown in Figure 2. The relative fluores-
[3] H. Jomaa, J. Wiesner, S. Sanderbrand, B. Altincicek, C.
Weidemeyer, M. Hintz, I. Türbachova, M. Eberl, J. Zeidler,
H. K. Lichtenthaler, D. Soldati, E. Beck, Science 1999, 285,
1573 – 1576.
[4] a) W. Eisenreich, M. Schwarz, A. Cartayrade, D. Arigoni, M. H.
Zenk, A. Bacher, Chem. Biol. 1998, 5, R221 – R233; b) M.
Rohmer, M. Knani, P. Simonin, B. Sutter, H. Sahm, Biochem. J.
1993, 295, 517 – 524.
[5] a) M. K. Schwarz, PhD Dissertation, ETH Zürich, No. 10951,
1994; b) W. Eisenreich, B. Menhard, P. J. Hylands, M. H. Zenk,
A. Bacher, Proc. Natl. Acad. Sci. USA 1996, 93, 6431 – 6436;
c) S. T. J. Broers, PhD Dissertation, ETH Zürich, No. 10978,
1994.
[6] a) F. Rohdich, S. Hecht, A. Bacher, W. Eisenreich, Pure Appl.
Chem. 2003, 75, 2 – 3, 393 – 405; b) Y. Boucher, W. F. Doolittle,
Mol. Microbiol. 2000, 37, 703 – 716; c) F. Rohdich, J. Wungsinta-
weekul, M. Fellermeier, S. Sanger, S. Herz, K. Kis, W. Eisenreich,
A. Bacher, M. H. Zenk, Proc. Natl. Acad. Sci. USA 1999, 96,
11758 – 11763.
[7] For selected publications, see: a) F. Rohdich, W. Eisenreich, J.
Wungsintaweekul, S. Hecht, C. A. Schuhr, A. Bacher, Eur. J.
Biochem. 2001, 268, 3190 – 3197; b) S. B. Richard, J.-L. Ferrer,
M. E. Bowman, A. M. Lillo, C. N. Tetzlaff, D. E. Cane, J. P. Noel,
J. Biol. Chem. 2002, 277, 8667 – 8672; c) T. L. Campbell, E. D.
Brown, J. Bacteriol. 2002, 184, 5609 – 5618.
[8] a) S. Steinbacher, J. Kaiser, J. Wungsintaweekul, S. Hecht, W.
Eisenreich, S. Gerhardt, A. Bacher, F. Rohdich, J. Mol. Biol.
2002, 316, 79 – 88; b) L. E. Kemp, C. S. Bond, W. N. Hunter,
Proc. Natl. Acad. Sci. USA 2002, 99, 6591 – 6596.
[9] Published crystal structures of IspF: a) H. Kishida, T. Wada, S.
Unzai, T. Kuzuyama, M. Takagi, T. Terada, M. Shirouzu, S.
Yokoyama, J. R. H. Tame, S.-Y. Park, Acta Crystallogr. D 2003,
59, 23 – 31; b) S. Ni, H. Robinson, G. C. Marsing, D. E. Bussiere,
M. A. Kennedy, Acta Crystallogr. D 2004, 60, 1949 – 1957; c) C.
Lehmann, K. Lim, J. Toedt, W. Krajewski, A. Howard, E.
Eisenstein, O. Herzberg, Proteins: Struct. Funct. Bioinf. 2002, 49,
135 – 138; d) L. E. Kemp, M. S. Alphey, C. S. Bond, M. A. J.
Ferguson, S. Hecht, A. Bacher, W. Eisenreich, F. Rohdich, W. N.
Hunter, Acta Crystallogr. D 2005, 61, 45 – 52; e) T. Sgraja, L. E.
Kemp, N. L. Ramsden, W. N. Hunter, Acta Crystallogr. F 2005,
61, 625 – 629.
[10] a) P. R. Gerber, K. Müller, J. Comput.-Aided Mol. Des. 1995, 9,
251 – 268; b) Gerber Molecular Design (http://www.moloc.ch).
[11] For structure-based design of inhibitors of plasmepsin II,
another potential target in antimalarial therapy, see: D. Carc-
ache, S. Hörtner, A. Bertogg, C. Bonkert, D. Bur, H. P. Märki, A.
Dorn, F. Diederich, ChemBioChem 2002, 3, 1137 – 1141.
[12] a) Y.-H. Li, L.-M. Chan, L. Tyler, R. T. Moody, C. M. Himel,
D. M. Hercules, J. Am. Chem. Soc. 1975, 97, 3118 – 3126; b) F.
Diederich, K. Dick, D. Griebel, Chem. Ber. 1985, 118, 3588 –
3619.
cence enhancements (ImaxꢁI)/I, where
I is the maximum
fluorescence of the unbound inhibitor, were 20% (7), 30% (8),
and 80% (9). Binding data were analyzed by nonlinear least-
squares approximation (GraphPad Prism 4 Software, San Diego,
CA, 2005). Uncertainties given are standard deviations, with R2
for all fits ꢂ 0.96); b) R. S. Sarfati, V. K. Kansal, H. Munier, P.
Glaser, A.-M. Gilles, E. Labruyꢀre, M. Mock, A. Danchin, O.
Bꢁrzu, J. Biol. Chem. 1990, 265, 18902 – 18906; c) W. O.
McClure, G. M. Edelman, Biochemistry 1967, 6, 559 – 566.
[16] The IC50 values for CMP, CDP, and 9 were determined in
enzymatic reactions catalyzed by the E. coli IspF enzyme using
[1,3,4-13C3]-5 (1 mm) as substrate. Product formation was moni-
tored by 13C NMR spectroscopy (see the Supporting Informa-
tion).
[17] a) Crystals of E. coli IspF complexed with ligands 7 or 9 were
grown by the vapor-diffusion method.[8b] The reservoir solution
was 0.1m ammonium sulfate and 0.1m sodium acetate, pH 5, and
10% and 8% monomethylether polyethylene glycol 200 for
compounds 7 and 9, respectively. The hanging drops consisted of
1 mL of reservoir and 3 mL of a solution of IspF (5.5 mgmLꢁ1) in
50 mm Tris (tris(hydroxymethyl)aminomethane) hydrochloride,
pH 7.7, containing 50 mm NaCl and 2 mm ligand. Crystals of the
complex with 7 are monoclinic plates in the space group P21 with
unit cell dimensions a = 54.03, b = 115.27, c = 87.61 Š, b =
90.188. Crystals of the complex with 9 are cubic blocks in the
space group I213 with a = 145.1 Š. Data were measured on
beamline ID29 at the European Synchrotron Radiation Facility
(Grenoble, France) and processed with MOSFLM (A. G. W.
Leslie, H. R. Powell, G. Winter, O. Svensson, D. Spruce, S.
McSweeney, D. Love, S. Kinder, E. Duke, C. Nave, Acta
Crystallogr. D 2002, 58, 1924 – 1928). The dataset for the complex
of 7 is 99.5% complete to 2.3-Š resolution with an Rsym value of
9.7%, 27.1% in the highest resolution bin. The dataset for the
complex of 9 is 99% complete to 2.5-Š resolution with an Rsym
value of 8%, 68.5% in the highest resolution bin. The starting
model for both analyses was the structure of E. coli IspF in
complex with CDP (PDB code 1GX1). Molecular replacement
methods (Collaborative Computational Project Number 4, Acta
Crystallogr. D 1994, 50, 760–763; J. Navaza, Acta Crystallogr.
Sect. D 2001, 57, 1367 – 1372) were used to generate the starting
models for refinement, which consisted of six subunits (two
trimers) for the complex with 7 and a single subunit for the
complex with 9. The structures were refined using a combination
of O (T. A. Jones, J. Y. Zou, S. W. Cowan, M. Kjeldgaard, Acta
Crystallogr. A 1991, 47, 110 – 119) and refmac5 (G. N. Murshu-
dov, A. A. Vagin, E. J. Dodson, Acta Crystallogr. D 1997, 53,
240 – 255) to an R factor of 24.3% and R-free of 27.9% for the
complex with 7; an R factor of 21.5% and R-free of 25.9% for
the complex of 9. The model for the complex of 7 includes 233
water molecules, six ZnII ions, one in each active site and two
molecules of geranyldiphosphate (GPP).[9d] A total of 92.1% of
the residues are in the most favored regions of the Ramachan-
dran plot with none in disallowed regions. The model for the
complex of 9 includes 56 water molecules, a ZnII ion, and one
geranyldiphosphate; 90.8% of the residues are in the most
favored regions of the Ramachandran plot with none in
disallowed regions. Further details are included in the Protein
Data Bank depositions PDB codes: 2AMT and 2AO4; b) It
should be noted that a screen of crystallization conditions were
[13] M. Saady, L. Lebeau, C. Mioskowski, Tetrahedron Lett. 1995, 36,
2239 – 2242.
[14] V. Wittmann, C.-H. Wong, J. Org. Chem. 1997, 62, 2144 – 2147.
[15] a) The fluorescent inhibitors 7–9 were titrated into a solution
(2 mL) of IspF (17 kDa per subunit) from E. coli (6.25 mm) in
50 mm Tris hydrochloride, pH 8.2, containing 10 mm MgSO4 and
2 mm ZnO(Ac)2 in a total volume of 3 mL at 208C (Tris =
tris(hydroxymethyl)aminomethane). Spectral data were
recorded at lexc = 320 nm, lem = 423 nm for compounds 7 and 8,
and lexc = 366 nm, lem = 532 nm for 9, using a Hitachi F-2000
fluorescence spectrophotometer (1-cm quartz cuvette, 3-mL
volume). Kd values were derived using Equation (1):
I
rel = (Imax Fo)(Kd + Fo)ꢁ1
(1),
Angew. Chem. Int. Ed. 2006, 45, 1069 –1074
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
1073

Communications
explored to generate useful crystals, and the best conditions did
not have MgII present for either ligands 7 and 9. Crystals with
other ligands in the presence and absence of MgII or MnII have
been grown in the past; structural differences that can be
attributed to the presence/absence of the divalent cations could
not be observed (unpublished data).
[18] a) The imidazole ring of His34’ in active sites B–F is observed to
varying degrees to be in an “open” (away from the active site)
conformation. Distances from the anthranilate NH2 to the
imidazole ring range from 3.8 to 4.5 Š, thereby precluding the
presence of any strong interaction. The “open” and “closed”
conformations of His34’ have been observed in some of the
previously reported crystal structures of IspF.[8,9] b) Deviation in
active site A cannot be attributed to crystal-packing effects.
1074
www.angewandte.org
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2006, 45, 1069 –1074