Inhibitors of the herbicidal target IspD: Allosteric site binding

Article abstract of DOI:10.1002/anie.201102281

The pick of the pockets: The first inhibitors for IspD, an enzyme from the non-mevalonate pathway of isoprenoid biosynthesis, are described. High-throughput-screening revealed a hit with an IC₅₀ value of 140 nM. Co-crystal structure analyses of the binding mode in the newly formed allosteric pocket (see structure, red ball: water O atom), lead to the synthesis of a set of 17 derivatives which were tested to optimize the herbicidal activity.

Full text of DOI:10.1002/anie.201102281

DOI: 10.1002/anie.201102281
Herbicides
Inhibitors of the Herbicidal Target IspD: Allosteric Site Binding**
Matthias C. Witschel,* H. Wolfgang Hçffken, Michael Seet, Liliana Parra, Thomas Mietzner,
Frank Thater, Ricarda Niggeweg, Franz Rçhl, Boris Illarionov, Felix Rohdich, Johannes Kaiser,
Markus Fischer,* Adelbert Bacher,* and FranÅois Diederich*
In memoriam Viktoriya Illarionova
After the identification of the non-mevalonate pathway for
isoprenoid biosynthesis (Scheme 1) in the mid 1990s,[1] major
efforts in academia and industry were started to use this
essential pathway for the identification of new antibacterials,
antimalarials, and also herbicides.[2] The emerging resistance
problems of many weeds and pathogens reflect the urgent
need for new actives with novel modes of action (MOA). The
enzymes of the non-mevalonate pathway are especially
interesting targets for anti-parasitic drugs or agrochemicals,
as there exist no homologs in mammals and therefore no
target related toxicological side-effects of inhibitors of this
pathway are expected.
For commercial herbicides, the last novel MOA was
introduced in 1980 with p-hydroxyphenyl pyruvate dioxyge-
nase (HPPD).[3] Due to increasing resistance of weeds to
almost all available herbicides, especially also to the com-
mercially most important herbicide glyphosate,[4] new MOAs
in the herbicide market are of high importance for the future
world food production. Therefore, a program has been
initiated to identify inhibitors of the non-mevalonate pathway
in plants.
Besides antisense-based validation of the herbicidal
potential of several of the enzymes in the pathway,[5] two
Scheme 1. The non-mevalonate pathway for the biosynthesis of isopre-
noids.[1] DXS: 1-deoxy-d-xylulose-5-phosphate synthase (EC 2.2.1.7),
herbicidal inhibitors were already known: Ketoclomazone for
IspC/DXR: 1-deoxy-d-xylulose-5-phosphate reductoisomerase
(EC 1.1.1.267), MEP: 2C-methyl-d-erythritol 4-phosphate, IspD: 4-
diphosphocytidyl-2C-methyl-d-erythritol synthase (EC 2.7.7.60), CTP:
cytidine triphosphate, CDP-ME: 4-diphosphocytidyl-2C-methyl-d-erythri-
tol, IspE: 4-diphosphocytidyl-2C-methyl-d-erythritol kinase
(EC 2.7.1.148), IspF: 2C-methyl-d-erythritol-2,4-cyclodiphosphate syn-
thase (EC 4.6.1.12), IspG: 2C-methyl-d-erythritol-2,4-cyclodiphosphate
reductase (EC 1.17.4.3), IspH: 1-hydroxy-2(E)-methylbutenyl-4-diphos-
[*] Dr. M. C. Witschel, Dr. H. W. Hçffken, Dr. L. Parra, T. Mietzner,
F. Thater, Dr. R. Niggeweg, Dr. F. Rçhl
BASF SE
Boschstrasse 14, 67056 Ludwigshafen (Germany)
E-mail: matthias.witschel@basf.com
M. Seet, Prof. Dr. F. Diederich
Laboratorium fꢀr Organische Chemie, ETH Zꢀrich
Wolfgang-Pauli-Strasse 10, HCI
phate reductase (EC 5.3.3.2), IPP: isopentenyl diphosphate, DMAPP:
dimethylallyl diphosphate.
8093 Zꢀrich (Switzerland)
E-mail: diederich@org.chem.ethz.ch
Dr. B. Illarionov, Prof. Dr. M. Fischer, Prof. Dr. A. Bacher
Institut fꢀr Lebensmittelchemie, Universitꢁt Hamburg
Grindelallee 117, 20146 Hamburg (Germany)
the enzyme DXS (1-deoxy-d-xylulose 5-phosphate synthase,
EC 2.2.1.7)[6] and Fosmidomycin for the enzyme DXR (1-
deoxy-d-xylulose 5-phosphate reductoisomerase, also called
IspC, EC 1.1.1.267).[7] For the subsequent enzymes of the
pathway, IspD-IspH, inhibitors for IspE[8] and IspF[9] have
been reported recently.
As IspD has been validated by antisense experiments to
be essential in plants, inhibitors of this enzyme were
anticipated to be potentially herbicidal.[10] Therefore, at
BASF a proprietary library of about 100000 diversity selected
compounds was tested for the inhibition of the enzyme IspD
of the plant Arabidopsis thaliana in a high throughput
E-mail: markus.fischer@chemie.uni-hamburg.de
Dr. F. Rohdich, Dr. J. Kaiser, Prof. Dr. A. Bacher
Lehrstuhl fꢀr Biochemie, Technische Universitꢁt Mꢀnchen
Lichtenbergstrasse 4, 85748 Garching (Germany)
E-mail: adelbertbacher@aol.com
[**] Work at ETH was supported by the Swiss National Science
Foundation. IspD=4-Diphosphocytidyl-2C-methyl-d-erythritol syn-
thase, EC 2.7.7.60.
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/anie.201102281.
Angew. Chem. Int. Ed. 2011, 50, 7931 –7935
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
7931

Communications
screening (HTS, see the Supporting Information for the
Surprisingly, the inhibitor does not bind to the presumed
active site, but rather in a newly formed allosteric pocket in
close proximity, resulting in a reduction in size of the active
site.
The hydroxytriazolopyrimidine has a measured pKa value
of 3.9, and hence binds in the deprotonated form, bridging
between the side chain of Arg157 and the backbone NH of
Val266. Moreover, one nitrogen atom of the triazole sub-
structure interacts with the backbone NH of Ile265. The
second nitrogen is bound to Arg157 by a highly localized
water molecule, so that the inhibitor is fixed by a network of
four H-bonds. The heterobicycle is sandwiched between a H-
bonding array involving the side chain NH of Gln238 and the
enzyme assay).[11]
Only few hits in the sub-micromolar and in the micro-
molar range could be identified, with the 7-hydroxy-
[1,2,4]triazolo[1,5-a]pyrimidine 1 being the most active inhib-
itor with an IC50 value of (140 Æ 10) nm. Based on the
published crystal structure of the A. thaliana IspD enzyme
(protein data bank (PDB) code: 1W77),[12] no reasonable
docking mode of hit 1 in the active site could be proposed.
Thus, it was decided to attempt a co-crystal structure of 1 with
the enzyme, and a structure with 1.4 ꢀ resolution (PDB code:
2YC3) could be obtained (Figure 1, and Supporting Informa-
tion, Figures 1SI and 2SI).
=
backbone C O of Asp262 on top, and an array of lipophilic
side chains of Leu245, Ile265, and Val259 on the bottom
(Supporting Information, Figure 2SI).
The benzyl moiety of 1 fits tightly into the lipophilic rear
part of the pocket, composed of Val266, Val263, Ala202, and
Ile240. This benzyl pocket provides almost no space for
further derivatization. The ortho positions of the aromatic
core point towards the lipophilic side chains of Ile240 and
Val263, with C···C distances of 3.8 ꢀ and 3.9 ꢀ, respectively.
Potential substituents in meta position would run into
repulsion with the backbone carbonyl group of Val239
(distance CBn···O: 3.5 ꢀ) and the side chain of Val263
(distance CBn···C: 3.7 ꢀ). The para position is blocked by
the side chain of Val204 with a CBn···C distance of 3.9 ꢀ. This
explicitly apolar environment is completed by the side chain
of Ile177, interacting closely with the chlorine substituent of
the ligand (distance C···Cl: 2.9 ꢀ, see Supporting Informa-
tion, Figure 2SI).
We became intrigued by the possibility to further enhance
binding affinity through proper substitution of the crystallo-
graphically defined water molecule at the entrance channel to
the binding pocket, which forms H-bonds to N(3) of 1 (for
atom labeling, see Scheme 2) and the side chain of Arg157
(Figure 1). To explore this opportunity and to test the
hypotheses derived from the examination of the X-ray co-
crystal structure of 1, a set of 17 derivatives (2–18, Table 1)
was synthesized.
The synthesis of all 7-hydroxy[1,2,4]triazolo[1,5-a]pyrimi-
dines and 7-hydroxypyrazolo[1,5-a]pyrimidines follows the
same route illustrated in Scheme 2 for ligand 1 (experimental
details are shown in the Supporting Information). Benzyl
malonate 19 was condensed with aminotriazole 20, resulting
in the dihydroxytriazolopyrimidine 5. Subsequent chlorina-
tion with POCl3 gave the dichloropyrimidine 21, which was
regioselectively hydrolyzed in the 7-position with NaOH to
yield the targeted 5-chloro-7-hydroxypyrimidine 1.
The benzyl pocket is really tight. Introduction of an
additional methyl group in the benzylic position ((Æ )-14)
reduces the activity by a factor greater than 100 compared to
2. Similarly, substituents on the aromatic ring of the benzyl
moiety lead to a decrease in activity (comparison of 2 with 15–
17).
All changes on the heterobicyclic core besides the 3-
position resulted in significant reduction of target activity (see
4–7, 13). Replacing the N(3) of 1 by C-Me (8) resulted in the
expected drastic reduction in activity due to disruption of the
Figure 1. X-ray crystal structure of A. thaliana IspD co-crystallized with
1 at 1.4 ꢂ resolution (PDB code: 2YC3). a) Chemical formula of 1 and
schematic overview of its binding mode. b) Surface of the allosteric
pocket. Distances are given in ꢂ. C gray (enzyme) and cyan (inhibitor),
N blue, O red, Cl green.
7932
www.angewandte.org
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2011, 50, 7931 –7935

Table 1: Inhibition data of A. thaliana IspD and herbicidal activity on Abutilon theophrasti and Setaria
faberi at 3 kg/ha by various azolopyrimidines.
was present in the co-crystal struc-
ture with 1. As a result, nitrile 2
exhibits an IC50 value of (35 Æ 7)
nm, a fourfold increase in activity
compared to 1, which can be
explained by the energetically
favorable water replacement.[13]
The activity of carboxylate 3 (IC50
value: (274 Æ 15) mm), however, is
lowered by a factor of almost 2000.
This result is rather surprising, con-
sidering that the pKa values of the
carboxylic acid and the aromatic
OH group were determined to be
2.32, and 5.22, respectively, and that
3 would therefore potentially bind
as a di-anion. We explain the low
activity of 3 with the high energetic
costs to be paid for the desolvation
of the carboxylate moiety upon
binding. Carboxylic esters and
amides (9–12) are more potent
than the acid, but much less than
the lead ligands 1 and 2.
Ligand
R1
R2
R3
R4
X
IC50 [mm]
Herbicidal activity[a]
1
4
5
6
7
8
3
9
Ph
Ph
Ph
Ph
Ph
Ph
Ph
Ph
Ph
Ph
Cl
H
H
H
CO2H
H
H
H
H
H
H
H
H
H
H
N
N
N
N
0.14Æ0.01
20Æ3
+++
+++
n.t.
+
0
n.t.
0
0
0
n.t.
Me
OH
Me
Me
Cl
Cl
Cl
Cl
Cl
>500
>500
>500
CH2OH
N
H
H
H
H
H
C-Me
C-CO2H
C-CO2tBu
C-CO2Et
C-CONH
(n-pentyl)
C-CONH2
C-CN
C-CN
C-CN
C-CN
C-CN
C-CN
N
43Æ2
274Æ15
38Æ3
10
11
9.5Æ0.6
18Æ1
12
2
13
Ph
Ph
Ph
Ph
2-F-Ph
3-F-Ph
3-Cl-Ph
iPr
Cl
Cl
OH
Cl
Cl
Cl
Cl
Cl
H
H
H
H
H
H
H
H
H
H
H
Me
H
H
H
H
6Æ0.3
n.t.
+
0
0
0
0
0
n.t.
0.035Æ0.007
>500
(Æ)-14
3.1Æ0.3
0.25Æ0.01
0.27Æ0.02
0.45Æ0.05
91Æ12
15
16
17
18
[a] n.t.=not tested.
It is noteworthy to discuss that
in the crystal structure of A. thali-
ana IspD with cytidine monophos-
phate (CMP) (PDB code: 1W77), the allosteric pocket filled
by ligands, such as 1–3, is not occupied and actually not fully
developed. This is a result of the geometry and flexibility of
the loop between Glu255 and Tyr268 with partly unresolved
side chains. The loop fully opens the pocket and adopts an
ordered conformation only in the presence of a suitable
ligand, such as 2. Figure 3 shows an overlay of the two co-
crystal structures of A. thaliana IspD with CMP and 2,
respectively.
The lower part of the Figure shows an enlargement of the
loop in both structures. Whereas Glu255 of the 1W77
structure (green) and Glu255* of the structure co-crystallized
with 2 (gray) still show a comparable alignment, it can be seen
that the loop adopts a different conformation starting with the
following amino acid residues. Especially the side chain of
Glu258*, pointing outside of the protein in the complex of 2,
points inside in the 1W77 structure, into the space of the
allosteric binding pocket. Moreover, the side chain of
Asp262* turns from the outside of the protein (in 1W77)
into the direction of the MEP-binding site (in 2YC5) and thus
contributes to the allosteric inhibition effect of 2 by a
reduction in size of the active site on the one hand,
presumably by electrostatic repulsion with the anionic
substrate on the other.
The original lead 1 showed very good herbicidal activity
(+ + + ) in the greenhouse at a dose of 3 kg/ha on two key
weeds in corn, Abutilon theophrasti and Setaria faberi
(Table 1). The analogues with lower target activity showed
weaker or no observable herbicidal activity. The highly target-
active nitrile derivative 2 displayed in the same assay also
some activity in the greenhouse (+ ), but on a lower level than
lead structure 1. Changing the adjuvant to the more acidic
Scheme 2. Synthesis of triazolopyrimidine 1. a) NBu3, 1508C, 8 h,
25%; b) POCl3, HNMe3Cl, 1508C, 15 h, 48%; c) NaOH, THF, 258C,
14 h, 57%.
hydrogen-bonding network. However, introducing a H-bond
acceptor that can interact with the positively charged Arg157
side chain and thus replace the water molecule, should be
favorable.[13] Figure 2 shows the structures of two such
derivates, nitrile 2 and carboxylic acid 3, as well as the
respective co-crystal structures with the A. thaliana IspD
enzyme (2: 1.6 ꢀ resolution, PDB code: 2YC5; 3: 1.8 ꢀ
resolution, PDB code: 2YMC). Both ligands adopt a very
similar binding geometry in the allosteric site to the one seen
for 1.
The nitrile and the carboxylate participate in the H-
bonding network and thus replace the water molecule that
Angew. Chem. Int. Ed. 2011, 50, 7931 –7935
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
7933

Communications
Figure 3. Schematic overlay representation of the two co-crystal struc-
tures of A. thaliana IspD. Green: with CMP (PDB code: 1W77), gray:
with 2 (PDB code: 2YC5). a) Overview of the complete protein.
b) Flexible loop between Glu255 and Tyr268 (residues marked with an
asterix belong to the co-crystal structure with 2). Color code for the
protein side chains and ligands: C gray or green (enzyme) and
cyan (CMP and inhibitor 2), N blue, O red, Cl green, P orange.
subsequent structure optimization has resulted in an
enhanced activity down to IC50 values of (35 Æ 7) nm. For
several target active analogues, we could also show herbicidal
activity. A potential feedback regulatory function of the
allosteric pocket by binding of later products of the non-
mevalonate-pathway is currently under investigation.
Figure 2. X-ray crystal structures of a) 2 (1.6 ꢂ resolution, PDB code:
2YC5) and b) 3 (1.8 ꢂ resolution, PDB code: 2YMC) co-crystallized
with A. thaliana IspD. Distances are given in ꢂ. C gray (enzyme) and
cyan (inhibitor), N blue, O red, Cl green.
Received: April 1, 2011
Published online: July 18, 2011
DASH,[14] the one-week rating could be improved, for
example, for 2 and 15 to good activity (+ + ), indicating
limitations in the uptake of these compounds. The activity of
these compounds is at an interesting level for a lead structure.
An improvement in target- and greenhouse activity by a
factor of at least ten will still be required to approach
commercially acceptable level.
To our knowledge, the azolopyrimidines are the first
described inhibitors of the enzyme IspD. Analysis of the
binding mode in the newly formed allosteric pocket and
Keywords: allosteric binding · herbicides · medicinal chemistry ·
protein–ligand interactions · structure–activity relationship
.
[1] a) M. Rohmer, M. Knani, P. Simoni, B. Sutter, H. Sahm,
Biochem. J. 1993, 295, 517 – 524; b) M. K. Schwarz, Dissertation
No. 10951, ETH Zꢁrich, 1994; c) S. T. J. Broers, Dissertation No.
10978, ETH Zꢁrich, 1994; d) W. Eisenreich, B. Menhard, P. J.
Hylands, M. H. Zenk, A. Bacher, Proc. Natl. Acad. Sci. USA
1996, 93, 6431 – 6436; e) W. Eisenreich, M. Schwarz, A. Cartayr-
ade, D. Arigoni, M. H. Zenk, A. Bacher, Chem. Biol. 1998, 5,
7934
www.angewandte.org
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2011, 50, 7931 –7935

R221 – R233; f) W. Eisenreich, F. Rohdich, A. Bacher, Trends
Plant Sci. 2001, 6, 78 – 84; g) W. Eisenreich, A. Bacher, D.
Arigoni, F. Rohdich, Cell. Mol. Life Sci. 2004, 61, 1401 – 1426.
[2] H. Jomaa, J. Wiesner, S. Sanderbrand, B. Altincicek, C.
Weidemeyer, M. Hintz, I. Turbachova, M. Eberl, J. Zeidler,
H. K. Lichtenthaler, D. Soldati, E. Beck, Science 1999, 285,
1573 – 1576.
[3] H. Matsumoto, M. Mitzutani, T. Yamaguchi, J. Kadotani, Weed
Biol. Manage. 2002, 2, 39 – 45.
[4] S. C. Weller, M. D. K. Owen, W. G. Johnson in Glyphosate
Resistance in Crops and Weeds (Eds.: V. K. Nandula), Wiley,
Hoboken, 2010, pp. 213 – 232.
[5] K. Okada, H. Kawaide, T. Kuzuyama, H. Seto, I. S. Curtis, Y.
Kamiya, Planta 2002, 215, 339 – 344.
[6] C. Mꢁller, J. Schwender, J. Zeidler, H. K. Lichtenthaler,
Biochem. Soc. Trans. 2000, 28, 792 – 793.
[7] a) F. Rohdich, S. Lauw, J. Kaiser, R. Feicht, P. Kçhler, A. Bacher,
W. Eisenreich, FESB J. 2006, 273, 4446 – 4458; b) T. Kuzuyama,
T. Shimizu, S. Takahashi, H. Seto, Tetrahedron Lett. 1998, 39,
7913 – 7916; c) J. Zeidler, J. Schwender, C. Mꢁller, J. Wiesner, C.
Weidemeyer, E. Beck, H. Jomaa, H. K. Lichtenthaler, Z.
Naturforsch. C 1998, 53, 980 – 986.
[8] A. K. H. Hirsch, M. S. Alphey, S. Lauw, M. Seet, L. Barandun,
W. Eisenreich, F. Rohdich, W. N. Hunter, A. Bacher, F. Die-
derich, Org. Biomol. Chem. 2008, 6, 2719 – 2730.
1069 – 1074; b) J. G. Geist, S. Lauw, V. Illarionova, B. Illarionov,
M. Fischer, T. Grꢂwert, F. Rohdich, W. Eisenreich, J. Kaiser, M.
Groll, C. Scheurer, S. Wittlin, J. L. Alonso-Gꢃmez, W. B.
Schweizer, A. Bacher, F. Diederich, ChemMedChem 2010, 5,
1092 – 1101.
[10] a) F. Rohdich, J. Wungsintaweekul, W. Eisenreich, G. Richter,
C. A. Schuhr, S. Hecht, M. H. Zenk, A. Bacher, Proc. Natl. Acad.
Sci. USA 2000, 97, 6451 – 6456; b) M. Fellermeier, M. Raschke, S.
Sagner, J. Wungsintaweekul, C. A. Schuhr, S. Hecht, K. Kis, T.
Radykewicz, P. Adam, F. Rohdich, W. Eisenreich, A. Bacher, D.
Arigoni, M. H. Zenk, Eur. J. Biochem. 2001, 268, 6302 – 6310.
[11] a) V. Illarionova, J. Kaiser, E. Ostrozhenkova, A. Bacher, M.
Fischer, W. Eisenreich, F. Rohdich, J. Org. Chem. 2006, 71, 8824 –
8834; b) A. Bacher, M. H. Zenk, W. Eisenreich, M. Fellermeier,
M. Fischer, S. Hecht, S. Herz, K. Kis, H. Lꢁttgen, F. Rohdich, S.
Sagner, C. A. Schuhr, J. Wungsintaweekul, WO 2001011055,
2001.
[12] M. Gabrielsen, J. Kaiser, F. Rohdich, W. Eisenreich, R. Laupitz,
A. Bacher, C. S. Bond, W. N. Hunter, FESB J. 2006, 273, 1065 –
1073.
[13] a) M. Ellermann, R. Jakob-Roetne, C. Lerner, E. Borroni, D.
Schlatter, D. Roth, A. Ehler, M. G. Rudolph, F. Diederich,
Angew. Chem. 2009, 121, 9256 – 9260; Angew. Chem. Int. Ed.
2009, 48, 9092 – 9096; b) H. Yu, S. W. Rick, J. Am. Chem. Soc.
2009, 131, 6608 – 6613.
[9] a) C. M. Crane, J. Kaiser, N. L. Ramsden, S. Lauw, F. Rohdich,
W. Eisenreich, W. N. Hunter, A. Bacher, F. Diederich, Angew.
Chem. 2006, 118, 1082 – 1087; Angew. Chem. Int. Ed. 2006, 45,
[14] T. Whitwell, K. E. Kalmowitz, G. S. Stapleton, J. Environ.
Hortic. 1992, 10, 55 – 58.
Angew. Chem. Int. Ed. 2011, 50, 7931 –7935
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
7935