Radiofluorinated derivatives of 2-(phosphonomethyl)pentanedioic acid as inhibitors of prostate specific membrane antigen (PSMA) for the imaging of prostate cancer

Article abstract of DOI:10.1021/jm300710j

For prostate cancer, prostate specific membrane antigen (PSMA) has been identified as a diagnostic and therapeutic target. Fluorinated derivatives of 2-(phosphonomethyl)pentanedioic acid were designed and synthesized to explore whether this fluorine-substituent is tolerated in the pentanedioic acid moiety that is common to almost all PSMA targeting small molecule inhibitors. The binding affinities of the racemic and individual stereoisomers of 2-fluoro-4-(phosphonomethyl)pentanedioic acid were determined and showed that the introduction of fluorine was well tolerated. The radiosynthesis of the analogous 2-[¹⁸F]fluoro-4-(phosphonomethyl)pentanedioic acid was developed and evaluated in vivo with the PSMA positive LNCaP human prostate cancer cell. The biological results demonstrated specific binding of the tracer to PSMA positive tumors in mice. These results warrant the further evaluation of this class of compounds as radiolabeled tracers for the detection and staging of prostate cancer.

Full text of DOI:10.1021/jm300710j

Article
pubs.acs.org/jmc
Radiofluorinated Derivatives of 2‑(Phosphonomethyl)pentanedioic
Acid as Inhibitors of Prostate Specific Membrane Antigen (PSMA) for
the Imaging of Prostate Cancer
Keith Graham,* Ralf Lesche, Alexey V. Gromov, Niels Bohnke, Martina Schafer, Jorma Hassfeld,
̈
̈
Ludger Dinkelborg, and Georg Kettschau
Global Drug Discovery, Bayer HealthCare, Mullerstrasse 178, 13353, Berlin, Germany
̈
ABSTRACT: For prostate cancer, prostate specific membrane antigen
(PSMA) has been identified as a diagnostic and therapeutic target.
Fluorinated derivatives of 2-(phosphonomethyl)pentanedioic acid were
designed and synthesized to explore whether this fluorine-substituent is
tolerated in the pentanedioic acid moiety that is common to almost all
PSMA targeting small molecule inhibitors. The binding affinities of the racemic and individual stereoisomers of 2-fluoro-4-
(phosphonomethyl)pentanedioic acid were determined and showed that the introduction of fluorine was well tolerated. The
radiosynthesis of the analogous 2-[18F]fluoro-4-(phosphonomethyl)pentanedioic acid was developed and evaluated in vivo with
the PSMA positive LNCaP human prostate cancer cell. The biological results demonstrated specific binding of the tracer to
PSMA positive tumors in mice. These results warrant the further evaluation of this class of compounds as radiolabeled tracers for
the detection and staging of prostate cancer.
and its metastatic disease.5 PSMA is also known as glutamate
INTRODUCTION
■
carbopeptidase II (GCPII) and folate hydrolase I (FOLH1)
and exerts N-acetylate α-linked acidic dipeptidase (NAALA-
Dase) activity. Numerous different scaffolds are available for
inhibiting PSMA and have been the topic of several recent
reviews.14,15 A number of these different PSMA scaffolds have
been radiolabeled with SPECT and PET radioisotopes for the
imaging of prostate cancer in the preclinical setting with
promising results.16−23 These scaffolds can be approximately
split into three categories: (1) glutamate−urea heterodimers
1;16−20,23 (2) glutamate containing phosphoramidates 2;21 (3)
2-(phosphinylmethyl)pentanedioic acid 322 (Figure 1).
Prostate cancer is the second leading cancer with more than
217 000 diagnosed cases in 2010 and is responsible for >30 000
prostate cancer-related deaths per year in the United States
alone; only lung cancer has a higher mortality rate.1 Prostate
specific antigen (PSA), an in vitro diagnostic test, is currently
used to screen for prostate cancer. However, this test has
limited specificity and does not allow differentiation of localized
from metastatic disease. Therefore, imaging procedures
allowing specific detection of prostate cancer with high
anatomic resolution would be highly desirable.
Currently, the only U.S. Food and Drug Administration
approved radiopharmaceutical for imaging prostate cancer is
111In-labeled capromab pendetide (ProstaScint, Cytogen).
However, this antibody-based agent has had limited clinical
use due to its slow distribution and clearance, which
contributed to challenging image interpretation. Low molecular
weight imaging agents for prostate cancer being pursued
clinically include radiolabeled choline analogues,6−10 anti-1-
amino-3-[18F]fluorocyclobutyl-1-carboxylic acid (anti-[18F]F-
FACBC),11 [18F]fluorodihydrotestosterone ([18F]FDHT),12
and 1-(2-deoxy-2-[18F]fluoro-L-arabinofuranosyl)-5-methylura-
cil ([18F]FMAU).13 All of these agents have different uptake
mechanisms, which can be advantageous like low urinary
excretion for [11C]choline and (anti[-18F]F-FACBC), but most
have various disadvantages, i.e., short physical half-life of the
PET isotope (20 min for carbon-11 for [11C]choline), high
abdominal background for anti-[18F]F-FACBC, or low-yielding
multistep radiosynthesis for [18F]FMAU.
Figure 1. Structures of the PSMA scaffolds 1 (glutamate−urea
heterodimer), 2 (glutamate phosphoramidate), and 3 (2-
(phosphinylmethyl)pentanedioic acid), all containing a similar binding
pentanedioic acid motif (dashed box).
The majority of published work focuses on the glutamate−
urea heterodimer scaffold after the first promising results
obtained with N-[N-[(S)-1,3-dicarboxypropyl]carbamoyl]-S-
[11C]methyl-L-cysteine ([11C]DCMC) and N-[N-[(S)-1,3-
dicarboxypropyl]carbamoyl]-S-3-[125I]iodo-L-tyrosine ([125I]-
DCIT).16 Presently the glutamate−urea−lysine heterodimer
One attractive target for primary and metastatic prostate
cancer is the prostate specific membrane antigen (PSMA),2−4
a
type II integral membrane protein expressed in normal human
prostate endothelium that is up-regulated in prostate cancer
Received: May 21, 2012
© XXXX American Chemical Society
A
dx.doi.org/10.1021/jm300710j | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
seems to be the heterodimer of choice, with subnanomolar
binding affinity and impressive preclinical in vivo images being
reported with various radioisotopes, i.e., fluorine-18,19 gallium-
68,20 iodine-123,18 iodine-131,18 rhenium-188,23 and techne-
tium-99m.23 Fluorine-18 is currently the isotope of choice, and
selected examples that have shown promising preclinical results
are illustrated in Figure 2, the glutamate−urea heterodimer
scaffold (417 and 519) and the phosphoramidate scaffold (621).
Surprisingly, 2-PMPA (7), the first PSMA inhibitor
discovered with subnanomolar activity and routinely used in
vitro as the gold standard to determine the binding affinity of
PSMA inhibitors, has not been investigated to see whether a
radiolabeled derivative is useful for the imaging of prostate
cancer. As fluorine-18 is the PET radioisotope of choice, with
its 110 min half-life allowing the radiolabeled imaging agent to
be produce centrally and distributed, we decided to investigate
whether fluorinated derivatives of 2-PMPA would maintain
their potency and when radiolabeled with fluorine-18, be
suitable imaging agents for prostate cancer.
Herein, we report the synthesis, structure elucidation, and in
vitro evaluation of the different stereoisomers of 2-fluoro-4-
(phosphonomethyl)pentanedioic acid 21. A suitable precursor
21 for direct labeling was synthesized to allow the radio-
fluorinated derivative [18F]21 to be synthesized in good yield
and high radiochemical purity. This tracer was evaluated in vivo
in mice bearing LNCaP tumor xenografts.
RESULTS AND DISCUSSION
■
Chemistry: Synthesis of the Fluorinated Reference
Compounds. As outlined in Scheme 1, the key alcohol
intermediate 22 was synthesized from methyl acrylate and
methyl glyoxylate via an indium-mediated Barbier-type reaction
in good yield (74%). This alcohol 22 was then subjected to a
Figure 2. Known literature examples of the glutamate−urea−lysine
heterodimer PSMA inhibitor scaffolds labeled with fluorine-18.
Vorbruggen-type fluorination28 employing nonafluorobutylsul-
̈
fonyl fluoride and triethylamine tris-hydrofluoride to give the
fluorinated derivative 23. Subsequently, a phospho-Michael
addition of dibenzylphosphite to the α,β-unsaturated ester 23
using potassium carbonate as a base led to the desired
intermediate 24 in good yield. Diastereoselectivity in said
Michael addition was low; typically, a ratio of approximately
1.5:1 was found in favor of the (2R,4S) and (2S,4R) isomers
over (2R,4R) and (2S,4S), respectively. Acidic hydrolysis of all
protecting groups was accomplished by heating in aqueous 6 M
HCl to afford the mixture of different stereoisomers of 21.
Since we found 21 to be prone to spontaneous (re)-
esterification in the presence of methanol, we performed the
deprotection in two cycles, with intermediate evaporation of all
volatiles and renewed exposure to 6 M HCl, in order to
accomplish truly complete deprotection.
As a result of the pronounced polarity of 21, we found
ourselves unsuccessful in separating its stereoisomers by means
of chiral HPLC even when screening a significant variety of
stationary phases (data not shown). Therefore, the individual
stereoisomers of 24 were prepared via two subsequent isomer
separations using preparative chiral HPLC methods after the
Structure−activity relationship studies on the glutamate or
pentanedioic acid moiety show that any modification can be
considerably detrimental to the binding affinity. The effect of
stereochemistry at the 2-position shows that the S-isomer (2S)-
7 is 300-fold more affine than its R-isomer (2R)-7 (Figure 3).24
In addition, shortening or elongating the carbon chain between
the two carboxylic acids results in a >500-fold loss in activity of
7 (0.3 nM) over 8 (700 nM) and 9 (185 nM).25
Efforts have been made by Kozikowski et al. to explore which
substituents can be tolerated on the glutamate backbone.
Unfortunately, simple substitutions with either a methyl or
benzyl group in the 4-position of glutamate gave drastic losses
in the affinity (Figure 4), surprisingly with methyl groups
having a more pronounced effect in binding affinity (∼150-fold
decrease) than the bulkier benzyl groups (∼50-fold decrease).26
Until recently it was assumed that the pentanedioic acid
moiety within these binders could not be modified without
considerable loss of affinity. However, Pomper et al. succeeded
in identifying highly potent PSMA inhibitors featuring a
modified glutamate portion featuring Ki values in the lower
nanomolar range and thus being only moderately inferior with
regard to 2-PMPA (0.3 nM). However, this was achieved
starting from a picomolar inhibitor with intact glutamate
portion (20, Ki = 0.01 nM, Figure 5), so the drop in activity was
still considerable.27
Vorbruggen fluorination and after the phospho-Michael
̈
addition, respectively, to give the desired stereoisomers
(2R,4R)-24, (2R,4S)-24, (2S,4R)-24, and (2S,4S)-24 (Scheme
2). Most gratifyingly, the basic conditions of the phospho-
Michael addition did not at all compromise the stereochemical
integrity at the stereocenter just established in (S)-23 and (R)-
Figure 3. Effect of stereochemistry and chain elongation of the PMPA scaffold.
B
dx.doi.org/10.1021/jm300710j | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
Figure 4. Effects of substitution on the glutamate backbone.
Figure 5. Substitution successfully tolerated in the P1′ site.
by phospho-Michael addition to give intermediate 27. Silyl
ether cleavage with excess HF·pyridine complex proceeded
smoothly to give the intermediate γ-hydroxy ester which was
immediately reacted with tosyl anhydride in pyridine to give
tosylate 28 in good yield without substantial formation of the
corresponding lactone. Not surprisingly, desilylation of 27 with
TBAF instead of HF·pyridine resulted in fast and complete
formation of the corresponding lactone (data not shown).
However, the need to employ large excesses of highly toxic
HF·pyridine complex and the intermediacy of a highly
lactonization prone γ-hydroxy ester warranted scouting for a
more practical route, in particular for scaling purposes. Thus,
key intermediate 22 was reacted with p-toluenesulfonic
anhydride to give the intermediate 29. When investigating
the phospho-Michael addition to 29, we were fairly concerned
that the leaving group qualities of the tosylate could result in
elimination to give a fully conjugated unsaturated diester. When
investigating the reaction with a broad range of bases, we
surprisingly found DBU, actually well-known as a reagent to
induce elimination, to be by far the most suitable base for this
reaction to afford the desired product 28 in good yield as
outlined in Scheme 4.
Radiochemistry. To establish a suitable radiosynthesis
method, precursor 28 was used as a mixture of stereoisomers.
The first step was the nucleophilic displacement of the tosylate
group in the 4-position with [18F]fluoride. The noncarrier
added (nca) nucleophilic fluorination of precursor 28 was
accomplished in DMSO at 120 °C in the presence of cesium
carbonate and Kryptofix 2.2.2. as a phase transfer catalyst
(PTC), and the intermediate [18F]24 was purified using
semipreparative HPLC methods, as we envisioned difficulties in
purifying the final radiolabeling product away from similar polar
byproducts that could interfere with any subsequent biological
testing. The purified [18F]-labeled intermediate [18F]24 was
subjected to a strongly acidic deprotection step using 6 M
hydrochloric acid at 120 °C to remove the two methylcarbox-
ylic esters and the two benzylic phosphonate esters. The
resulting strongly acidic reaction mixture was then passed
Scheme 1. Synthesis of Racemic Fluorinated PMPA
Derivative 21
a
a
(a) In, MeOH, 0.3 M HCl aq, 4 h, rt, 74%; (b) C4F9SO2F, Et3N·3HF,
Et3N, THF, 20 h, rt, 72%; (c) (BnO)2P(O), K2CO3, DMF, 4 h, 50 °C,
84%; (d) 6 M HCl aq, 60 °C, 30 h, 50%.
23. Again, deprotection of the stereoisomers of 24 was
accomplished by acidic hydrolysis in 6 M aqueous hydrochloric
acid. With regard to chemical naming it is noteworthy that the
atom numbering of the pentadioic acid backbone reverts from
24 to 21 as a result of protecting group removal.
The respective absolute stereochemistry of the different
isomers was assigned by X-ray crystallography of (2R,4R)-21
(Figure 8), in combination with assignments based on their
enantiomeric and diastereomeric relationship with each other,
which is evident from the NMR spectra (enantiomers give
identical NMR spectra, while different NMR spectra indicate
diastereomeric relationship), but also from their respective
chiral pHPLC separation history. Said assignments of absolute
stereochemistry were further corroborated by additional X-ray
analyses data of (2S,4S)-25 (Figure 9) which was formed from
(2S,4S)-24 by catalytic hydrogenolysis (Scheme 3).29
Chemistry: Synthesis of the Radiolabeling Precursor.
For the radiolabeling and in vivo experiments described below,
precursor 28 was prepared by silylation of alcohol 22, followed
C
dx.doi.org/10.1021/jm300710j | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
a
Scheme 2. Synthesis of the Different Stereoisomers of 21
a
(a) Chiral preparative HPLC (Chiralpak IC); (b) (BnO)2P(O)H, K2CO3, DMF, 50 °C, 2 h, 96/98% yield; (c) chiral preparative HPLC (Chiralcel
OD-H); (d) 6 M HCl aq, 50 °C, 30 h, 80% range.
a
Scheme 3. Synthesis of the Derivative (2S,4S)-25 for X-ray
Crystallography
Scheme 5. Synthesis of the Radiolabeled Racemic [18F]21
a
a
(a) K[18F]F, K222, DMSO, 120 °C, 15 min; (b) 6 M HCl aq, 120 °C,
15 min.
a
(a) Pd/C, H2, THF, 2 h, rt, 70%.
biological testing. The identity of [18F]21 was confirmed by co-
injection of the corresponding cold reference, using a Corona
aerosol detector (CAD) connected to the HPLC instrument
(Figure 6), as this detector is known to detect non-UV polar
through an AG-11A8 resin and a C18 solid phase extraction
cartridge (SPE) and washed with saline to give the desired
product [18F]21 (Scheme 5) as a formulated solution ready for
a
Scheme 4. Syntheses of the Precursor 28 (Optimized Route, as Mixture of Stereoisomers)
a
Initial route: (a) TBSCl, N-Me-imidazole, CH2Cl2, 0 °C then 16 h rt, 84%; (b) (BnO)2P(O)H, K2CO3, DMF, 2 h 90 °C then 60 h rt, 98%; (c) (i)
HF·py 50 equiv, THF, 4 h rt, (ii) Tos2O, pyridine, 16 h rt, 70% overall. Optimized route: (d) Tos2O, pyridine, 2 h at 0 °C, then 16 h rt, 99%; (e)
(BnO)2P(O)H, DBU, THF, 1.5 h rt, 61%.
D
dx.doi.org/10.1021/jm300710j | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
this scaffold showed promise for imaging PSMA expression in
PSMA positive tumor bearing xenographs. Positive visualization
of the tumor would then warrant further investigations into the
different radiolabeled stereoisomers of [18F]21, as new
analytical methods would need to be developed to analyze
the final product for stereochemistry integrity.
Micro-PET/CT Imaging. Racemic [18F]21 was injected into
nude mice bearing LNCaP tumors. Images of different animals
were recorded at 45−75 min and 110−130 min after injection,
respectively (shown in Figure 7). Racemic [18F]21 allowed
clear tumor visualization at both time points, whereas uptake
into other organs was observed mainly in bladder, kidney, and
to a lesser degree, bone. No visualization of liver and gut and
continuous accumulation of the tracer into the bladder pointed
to renal elimination as the predominant excretion pathway.
Since bone uptake did not increase over time, we found no
evidence for compound disintegration caused by accumulation
of free fluoride. Instead, residual free fluoride from the
radiopreparation (3−4% estimated by TLC) contributed to
the observed bone accumulation. Preinjection of unlabeled 21
resulted in significant decrease of signal from the tumor,
confirming specific binding of [18F]21 to LNCaP tumors.
Interestingly, tracer accumulation in kidney was also reduced
upon blocking with 21, pointing to specific binding of the
tracer. This observation is in line with previous reports of high
PSMA expression in proximal tubules of rodent kidneys34,35
and comparable to high and specific binding to kidney reported
for other radiotracers targeting PSMA.20,21
Figure 6. HPLC analysis of the final product [18F]21 with co-injection
of the reference standard 21 (top chromatogram with Corona
detector, bottom chromatogram with γ detector).
CONCLUSIONS
■
compounds with a high degree of sensitivity (detection limit
using this method was determined to be 1 μg/mL, and the
specific activity of the tracer was always below the limit of
detection and calculated to be >25GBq/μmol). The final
product was analyzed by TLC, and the radiochemical was
found to be >95% with about 3−4% free [18F]fluoride30,31
Biological Characterization. In Vitro Binding. In vitro
potency of PSMA inhibitors was determined by inhibiting
NAALADase, one of the enzymatic activities reported for
PSMA.32 NAALADase activity was determined in membrane
lysates from LNCaP cells, reported to express high amounts of
PSMA.33 NAALADase inhibition was determined by measuring
the release of [3H]glutamate from the PSMA substrate
[3H]NAAG in the presence of increasing concentrations of
inhibitor and resulted in Ki values of 0.4 0.2 nM and 2.6
0.4 nM for 2-PMPA and racemic 21, respectively. Among the
four stereoisomers compounds, (2S,4S)-21 and (2R,4S)-21
showed significantly higher potency to inhibit NAALADase
activity compared to (2S,4R)-21 and (2R,4R)-21, respectively
(see Table 1 for details). The next natural step was to carry out
in vivo experiments; it was decided to investigate the
radiolabeled racemic derivative [18F]21 to determine whether
The syntheses of different fluorinated stereoisomeric derivatives
of PMPA (the in vitro gold standard) showed that fluorine was
surprisingly well tolerated in the pentanedioic acid binding
portion, despite previous reports showing that small sub-
stitutions in this portion resulted in >100-fold loss in binding
affinity.25,26 Two of the four stereoisomers, (2R,4S)-21 (Ki =
3.1 nM) and (2S,4S)-21 (Ki = 1.4 nM), were identified to be
highly potent inhibitors of PSMA. [18F]-Labeled 21 was
synthesized as a mixture of stereoisomers and showed
promising images in LNCaP tumor bearing mice. These results
warrant the further evaluation of the different stereoisomers of
[18F]-21 in vivo.
EXPERIMENTAL SECTION
■
General. Chemicals and solvents were used as received from
commercial vendors Sigma Aldrich (Taufkirchen, Germany), Tocris
Bioscience (Bristol, U.K.), Alexis Biochemical (now Enzo Life
Sciences, Loerrach, Germany), Bachem AG (Bubendorf Switzerland),
Toronto Research Chemicals Inc. (North York, Ontario, Canada), and
Acros Organics (part of Thermo Fisher-Scientific, Nidderau,
Germany). Monomeric methyl glyoxylate was purchased from
Waterstone, Inc., Carmel, IN, U.S. [18F]Fluoride was produced via
an 18O(p,n)18F nuclear reaction and was supplied by EuroPET (Berlin,
Germany). Chemical reactions were monitored by TLC and by
UPLC/MS analysis of reaction mixtures (Waters Aqcuity with single
quadrupole ESI-MS detector). Crude products were, unless stated
otherwise, purified on silica gel (KP-SIL cartridges, Biotage) on
automated flash cromatography devices (SP-4 or Isolera Four by
Biotage). Stereoisomer separations were performed by chiral HPLC
according to the methods specified below. 1H and 19F nuclear
magnetic resonance (NMR) spectra were recorded with Bruker
AVANCE 300 and 400 MHz and the Bruker DRX-600 NMR
spectrometers. Chemical shifts (δ) are expressed in parts per million
relative to tetramethylsilane. Coupling constants (J) are expressed in
hertz (Hz), and the following abbreviations are used to describe the
Table 1. NAALADase Inhibition of Fluorinated PMPA
Compounds
compd
21
IC50 (nM)
Ki (nM)
6.8 0.7
3.1 0.8
8.1 0.7
55.9 3.1
89.4 13.2
1.1 0.6
2.6 0.4
1.4 0.6
3.1 0.3
22.0 1.1
36.4 6.3
0.4 0.2
(2S,4S)-21
(2R,4S)-21
(2S,4R)-21
(2R,4R)-21
7
E
dx.doi.org/10.1021/jm300710j | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
Figure 7. PET/CT imaging of [18F]21. Human LNCaP prostate tumor-bearing mice were used for tumor visualization by PET. Whole body PET
data (in rainbow colors) were acquired and fused with the CT image (in gray). Strong tumor signals (yellow arrow) were found 45−75 min pi (a)
and 110−130 min (b). Tumor uptake was significantly reduced by preinjection of excess cold reference 21 (110−130 min pi) (c), confirming
specific binding of [18F]21 to LNCaP tumors. The high background signals observed in bladder (b) were due to excretion. Besides renal excretion
the strong signals in kidneys (filled arrows) were caused by specific binding of the tracer to PSMA which is expressed in rodent kidney.34,35 Bone
uptake (arrow heads point to skull and joints) did not increase over time and was primarily due to residual [18F]fluoride in the radiopreparation.
Open arrows point to testosterone pellets that were implanted to stimulate growth of the androgen dependent LNCaP tumor cells.
appearance of peak patterns: s (singlet), d (doublet), t (triplet), q
(quartet), quint (quintet), combinations thereof (e.g., dd, doublet of
doublet), m (multiplet), mc (centered multiplet), and br (broad
signal). Purity all compounds including 21 and its stereoisomers was
judged to be in the 95% range according to H NMR, with no chemical
impurity detectable. For the isomeric mixture and the active
stereoisomers, this was confirmed by F NMR vs fulvestrant as internal
standard; yields were adjusted accordingly. Chiral HPLC methods
were carried out using the following analytical and preparative
methods. Analytical method 1 consisted of the following: Chiralpak IC
5 μm column (150 mm × 4.6 mm) with a flow rate of 1.0 mL/min at
25 °C using UV detection at 210 nm. The solvent system used was
isocratic (hexane/2-propanol, 80:20). Analytical method 2 consisted of
the following: Chiralcel OD-H 5 μm column (150 mm × 4.6 mm)
with a flow rate of 1.0 mL/min at 25 °C using UV detection at 210
nm. The solvent system used was isocratic (hexane/ethanol, 85:15).
Preparative method 1 consisted of the following: Chiralpak IC 5 μm
column (250 mm × 30 mm) with a flow rate of 40.0 mL/min at
ambient temperature using UV detection at 210 nm. The solvent
system used was isocratic (hexane/2-propanol, 80:20). Preparative
method 2 consisted of the following: Chiralcel OD-H 5 μm column
(250 mm × 20 mm) with a flow rate of 20.0 mL/min at ambient
temperature using UV detection at 210 nm. The solvent system used
was isocratic (hexane/ethanol, 85:15). X-ray crystallography analyses
were performed at 100 K on either a Bruker X8 Proteum Microstar
diffractometer with Helios Cu X-ray optics ((2S,4S)-25) or a Bruker
SMART platform diffractometer with Osmic Cu X-ray optics (Cu Kα,
λ = 1.541 78 Å) ((2R,4R)-21), respectively. The single crystals were
mounted on a cryoloop using a protective oil. X-ray data collection and
processing of data were performed using the PROTEUM2 program.36
Absorption correction was performed using SADABS.37 SHELXS was
used for structure solution, and SHELXL was used for full matrix least-
squares refinement on F2.37 All non-hydrogen atoms were refined
anisotropically. All H atoms were added in calculated positions and
refined riding on their resident atoms. The stereochemistry could be
assigned unambiguously. The program XP in the SHELXTL package
was used for molecular representations.37 Radioanalytical HPLC runs
were performed on an Agilent 1100, and preparative HPLC runs were
performed on a Knauer Smartline 1000. Purification was done on a
semipreparative ACE RP C18 column (10 mm × 250 mm, 5 μm;
Advanced Chromatographic Technologie). The elution solution was
60% acetonitrile/40% water with 0.1% trifluoroacetic acid at a flow of
4 mL/min. Radiochemical purity was analyzed on a ZIC HILIC
column (4.6 mm × 100 mm, 5 μm; SeQuant), and radioactivity
detection was performed on a GABI Star from Raytest. The elution
solution was 0.1 M ammonium formate (pH 3.2)/actonitrile (3:7),
and the flow rate was 0.5 mL/min. The non-UV active cold reference
compound was detected on a Corona charged aerosol detector (CAD)
from ESA.The tumor cell lines were obtained from ATTC (U.S.) or
DSZM (Germany) if not otherwise indicated and maintained
according to protocols provided by the supplier.
Dimethyl 2-Hydroxy-4-methylenepentanedioate (22). In a
three-necked flask equipped with a powerful mechanical stirrer, to a
cooled (+5 °C, ice−water bath) mixture of methyl glyoxylate (10.0 g,
114 mmol), methyl (2-bromomethyl)acrylate (22.5 g, 1.10 equiv),
methanol (80 mL), and 0.3 N aqueous hydrochloric acid (80 mL) was
added powdered indium (100 mesh, 13.0 g, 1.00 equiv) in several
portions, maintaining the internal temperature below 35 °C (which
did not require full cooling during the addition of the later portions).
The cooling bath was then completely removed. Several freshly broken
glass shards (from a Pasteur pipet, to scratch clotting pieces of the
indium) were added, and the mixture was stirred vigorously for 4 h,
during which the mixture cooled to room temperature. The mixture
was decanted from all solids, and the supernatant was concentrated in
vacuo. The residue was saturated with solid sodium chloride and then
extracted with MTBE. The residue remaining hereafter was loaded on
Celite and was washed with MTBE (4×) and EtOAc (1×). The
combined organic layers were dried over sodium sulfate, passed over a
plug of Celite, and evaporated. The residue was purified by column
chromatography over silica to give 16.8 g of the target compound in
1
94% purity (74% yield): H NMR (300 MHz, CDCl3) δ ppm 2.65
(dd, 1 H), 2.87 (dd, 1 H), 3.07 (s br, 1 H), 3.77 (s, 3 H), 3.78 (s, 3 H),
4.34−4.44 (m, 1 H), 5.74 (m, 1 H), 6.30 (m, 1 H). MS (CI): [M +
H]+ = 189. MS (CI): [M + NH4]+ = 206.
Dimethyl 2-Fluoro-4-methylenepentanedioate (23). To a
solution of 22 (6.00 g, 31.9 mmol) in THF (60 mL) were added at
room temperature perfluorobutanesulfonic acid fluoride (11.7 mL,
2.00 equiv), triethylamine trihydrofluoride (10.4 mL, 2.00 equiv), and
triethylamine (26.7 mL, 6.00 equiv). After being stirred for 20 h at
room temperature, the reaction mixture was partitioned between water
F
dx.doi.org/10.1021/jm300710j | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
and dichloromethane. The organic layer was separated and washed
with a brine/water mixture (1:1), dried over sodium sulfate, and
concentrated under reduced pressure. The crude product was purified
by column chromatography (silica, hexane/EtOAc) to give the desired
2 h in a microwave oven. After cooling to room temperature, the
mixture was partitioned between EtOAc and 5% aqueous citric acid,
and the organic layer was washed with brine, dried over sodium sulfate,
and evaporated. Column chromatography over silica gel (hexane/
EtOAc) gave 1.06 g (98% yield) of the target compound as a mixture
of the two (4R) epimers, which were separated by means of chiral
HPLC (preparative method 2) to give (2S,4R)-24 (480 mg, 44%
yield) and (2R,4R)-24 (340 mg, 31% yield).
1
fluoride (4.37 g, 72% yield): H NMR (400 MHz, CDCl3) δ ppm
2.76−2.88 (m, 1 H), 2.94−3.06 (m, 1 H), 3.79 (s, 3 H), 3.80 (s, 3 H),
5.18 (ddd, 1 H), 5.78 (s, 1 H), 6.35 (s, 1 H). MS (ESI): [M + H]+ =
191.
Dimethyl (S)-2-Fluoro-4-methylenepentanedioate ((S)-23).
(S)-23 was obtained from racemic 23 by means of preparative chiral
HPLC (preparative method 1): tR = 4.8 min (chiral HPLC, analytical
method 1).
Dimethyl (R)-2-Fluoro-4-methylenepentanedioate ((R)-23).
(R)-23 was obtained from racemic 23 by means of preparative chiral
HPLC (preparative method 1): tR = 6.3 min (chiral HPLC, analytical
method 1).
Dimethyl rac-2-{[Bis(benzyloxy)phosphoryl]methyl}-4-fluo-
ropentanedioate (24). To a solution of 23 (50 mg, 0.26 mmol)
and dibenzyl phosphite (86 mg, 1.25 equiv) in DMF (3 mL) was
added finely powdered potassium carbonate (mesh 325, 55 mg, 1.50
equiv), and the resulting mixture was heated to 50 °C for 4 h in a
microwave oven. After cooling to room temperature, the mixture was
partitioned between EtOAc and 5% aqueous citric acid. The organic
layer was washed with half-concentrated brine, dried over sodium
sulfate, and evaporated. The residue was purified by column
chromatography over silica gel (hexane/EtOAc) to give the target
compound as a mixture of stereoisomers (99 mg, 84% yield): 1H NMR
(400 MHz, CDCl3) δ ppm 1.88−2.04 (m, 1 H), 2.10−2.39 (m, 3 H),
2.91−3.08 (m, 1 H), 3.55 (s, 3 H, minor diastereomer), 3.58 (s, 3 H,
major diastereomer), 3.76 (s, 3 H, minor diastereomer), 3.77 (s, 3 H,
major diastereomer), 4.82−5.08 (m, 5 H), 7.29−7.41 (m, 10 H). MS
(ESI): [M + H]+ = 453. The reaction was also performed on an 80 g
scale under somewhat modified conditions (DBU, 1.25 equiv (instead
of K2CO3), dibenzyl phosphite (2.0 equiv) in THF, 3 h room
temperature, 46%).
Dimethyl (2S,4R)-2-{[Bis(benzyloxy)phosphoryl]methyl}-4-
fluoropentanedioate ((2S,4R)-24). tR (chiral HPLC, analytical
1
method 2) = 8.8 min; H NMR (300 MHz, CDCl3) δ ppm 1.86−
2.02 (m, 1 H), 2.09−2.42 (m, 3 H), 2.93−3.10 (m, 1 H), 3.58 (s, 3 H),
3.77 (s, 3 H), 4.79−5.11 (m, 5 H), 7.27−7.44 (m, 10 H).
Dimethyl (2R,4R)-2-{[Bis(benzyloxy)phosphoryl]methyl}-4-
fluoropentanedioate ((2R,4R)-24). tR (chiral HPLC, analytical
method 2) = 9.4 min; 1H NMR (300 MHz, CDCl3) δ ppm 1.90−2.06
(m, 1 H), 2.11−2.39 (m, 3 H), 2.89−3.06 (m, 1 H), 3.55 (s, 3 H), 3.76
(s, 3 H), 4.82−5.11 (m, 5 H), 7.27−7.43 (m, 10 H). The preparation
was adapted to larger scale starting from (R)-23 (2.58 g) being reacted
with 1.40 equiv of dibenzyl phosphite and 1.60 equiv of potassium
carbonate under otherwise unchanged conditions to give (2S,4R)-24
(2.28 g) and (2R,4R)-24 (1.49 g).
rac-2-Fluoro-4-(phosphonomethyl)pentanedioic Acid (21).
A mixture of 24 (34 g, 75 mmol) in aqueous 6 N hydrochloric acid
(340 mL) was stirred for 6 h at 60 °C. The mixture was then
evaporated (to remove methanol formed during hydrolysis). Again, 6
M hydrochloric acid (340 mL) was added, and the mixture was stirred
at 60 °C overnight. After evaporation, the residue was mixed with
acetic acid (340 mL) and then evaporated at 45 °C in vacuo to remove
traces of hydrochloric acid, followed by the same procedure using a 1:1
mixture of acetonitrile and toluene (780 mL) which was repeated until
no HCl traces could be detected after evaporation (control with wet
pH paper above the residue). The residue was recrystallized from
acetonitrile to give the desired product as an off-white solid (9.8 g,
1
50% yield): H NMR (500 MHz, DMSO-d6) δ ppm 1.64−1.79 (m, 1
Dimethyl (2S,4S)-2-{[Bis(benzyloxy)phosphoryl]methyl}-4-
fluoropentanedioate ((2S,4S)-24) and Dimethyl (2R,4S)-2-{[Bis-
(benzyloxy)phosphoryl]methyl}-4-fluoropentanedioate
((2R,4S)-24). To a solution of (S)-23 (500 mg, 2.39 mmol, 91%
purity) in DMF (4.0 mL) was added finely powdered potassium
carbonate (325 mesh, 595 mg, 1.80 equiv) and dibenzyl phosphite
(1.00 g, 1.60 equiv), and the resulting mixture was stirred at 50 °C for
2 h in a microwave oven. After cooling to room temperature, the
mixture was partitioned between EtOAc and 5% aqueous citric acid,
and the organic layer was washed with brine, dried over sodium sulfate,
and evaporated. Column chromatography over silica gel (hexane/
EtOAc) gave 1.04 g (96% yield) of the target compound as a mixture
of the two (4S) epimers, which was separated by means of chiral
HPLC (preparative method 2) to give (2S,4S)-24 (340 mg, 31% yield)
and (2R,4S)-24 (470 mg, 43% yield).
Dimethyl (2S,4S)-2-{[Bis(benzyloxy)phosphoryl]methyl}-4-
fluoropentanedioate ((2S,4S)-24). tR = 10.7 min (chiral HPLC,
analytical method 2); 1H NMR (300 MHz, CDCl3) δ ppm 1.90−2.06
(m, 1 H), 2.11−2.39 (m, 3 H), 2.89−3.06 (m, 1 H), 3.55 (s, 3 H), 3.76
(s, 3 H), 4.82−5.11 (m, 5 H), 7.27−7.43 (m, 10 H).
Dimethyl (2R,4S)-2-{[Bis(benzyloxy)phosphoryl]methyl}-4-
fluoropentanedioate ((2R,4S)-24). tR = 7.1 min (chiral HPLC,
analytical method 2); 1H NMR (300 MHz, CDCl3) δ ppm 1.86−2.02
(m, 1 H), 2.09−2.42 (m, 3 H), 2.93−3.10 (m, 1 H), 3.58 (s, 3 H), 3.77
(s, 3 H), 4.79−5.11 (m, 5 H), 7.27−7.44 (m, 10 H). The preparation
was adapted to larger scale starting from (S)-23 (2.75 g) being reacted
with 1.40 equiv of dibenzyl phosphite and 1.60 equiv of potassium
carbonate under otherwise unchanged conditions to give (2S,4S)-24
(2.02 g) and (2R,4S)-24 (3.03 g).
H), 1.87−2.00 (m, 1 H), 2.02−2.28 (m, 2 H), 2.66−2.79 (m, 1 H),
4.81−5.11 (m, 1 H). CO2H and PO3H2 protons form a very broad
peak between approximately 9 and 13 ppm. 19F NMR (376 MHz,
DMSO-d6) δ ppm −191.5 (mc, 1 F, diastereomer 1), −189.2 (mc, 1 F,
diastereomer 2). MS (ESI): [M − H]− = 243.
(2S,4S)-2-Fluoro-4-(phosphonomethyl)pentanedioic Acid
((2S,4S)-21). A mixture of (2S,4S)-24 (3.83 g, 8.47 mmol) and 6 M
hydrochloric acid (42 mL) was stirred at 50 °C for 12 h and then
evaporated (to remove methanol formed by hydrolysis). Another
portion of 6 M hydrochloric acid (42 mL) was added, and the mixture
was stirred at 50 °C overnight. The mixture was again evaporated and
then further processed, analogous to the procedure described for 21,
by addition of acetic acid followed by evaporation and subsequent
repetition of the addition/evaporation procedure with an acetonitrile/
toluene mixture until no HCl traces could be detected after
evaporation (control with wet pH paper above the residue).
Recrystallization from acetonitrile gave the crystalline product (1.77
1
g, 82% yield): H NMR (400 MHz, DMSO-d6) δ ppm 1.73 (ddd, J =
17.5, 15.2, 6.8 Hz, 1 H), 1.92 (ddd, J = 18.3, 15.2, 6.8 Hz, 1 H), 2.00−
2.29 (m, 2 H), 2.63−2.77 (m, 1 H), 5.02 (ddd, J = 49.3, 9.0, 3.8 Hz, 1
H); 19F NMR (376 MHz, DMSO-d6) δ ppm −188.9. MS (ESI): [M +
H]+ = 245.
(2S,4R)-2-Fluoro-4-(phosphonomethyl)pentanedioic Acid
(2S,4R)-21). A mixture of (2R,4S)-24 (5.28 g, 11.7 mmol) and 6 M
hydrochloric acid (58 mL) was stirred at 50 °C for 12 h and then
evaporated (to remove methanol formed by hydrolysis). Another
portion of 6 M hydrochloric acid (55 mL) was added, and the mixture
was stirred at 50 °C overnight. The mixture was again evaporated and
then further processed, analogous to the procedure described for 21,
by addition of acetic acid followed by evaporation and subsequent
repetition of the addition/evaporation procedure with an acetonitrile/
toluene mixture until no HCl traces could be detected after
evaporation (control with wet pH paper above the residue).
Recrystallization from acetonitrile gave the crystalline product (2.52
Dimethyl (2S,4R)-2-{[Bis(benzyloxy)phosphoryl]methyl}-4-
fluoropentanedioate ((2S,4R)-24) and Dimethyl (2R,4R)-2-
{[Bis(benzyloxy)phosphoryl]methyl}-4-fluoropentanedioate
((2R,4R)-24). To a solution of (R)-23 (500 mg, 2.39 mmol, 91%
purity) in DMF (4.0 mL) were added finely powdered potassium
carbonate (325 mesh, 595 mg, 1.80 equiv) and dibenzyl phosphite
(1.00 g, 1.60 equiv), and the resulting mixture was stirred at 50 °C for
1
g, 84% yield): H NMR (400 MHz, DMSO-d6) δ ppm 1.69 (ddd, J =
G
dx.doi.org/10.1021/jm300710j | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
17.2, 15.4, 8.1 Hz, 1 H), 1.87−2.28 (m, 3 H), 2.65−2.80 (m, 1 H),
4.90 (ddd, J = 50.0, 10.4, 2.8 Hz, 1 H).
686 μmol) in THF (5 mL) was added a 10% palladium on charcoal
catalyst (50 mg), and the resulting mixture was stirred at room
temperature under an atmosphere of hydrogen for 2 h. The catalyst
was filtered off over a plug of Celite, washed with acetonitrile, and all
(2R,4S)-2-Fluoro-4-(phosphonomethyl)pentanedioic Acid
(2R,4S)-21). A mixture of (2S,4R)-24 (5.05 g, 11.2 mmol) and 6 M
hydrochloric acid (55 mL) was stirred at 50 °C for 15 h and then
evaporated (to remove methanol formed by hydrolysis). Another
portion of 6 M hydrochloric acid (55 mL) was added, and the mixture
was stirred at 50 °C overnight. The mixture was again evaporated and
then further processed, analogous to the procedure described for 21,
by addition of acetic acid followed by evaporation and subsequent
repetition of the addition/evaporation procedure with an acetonitrile/
toluene mixture until no HCl traces could be detected after
evaporation (control with wet pH paper above the residue).
Recrystallization from acetonitrile gave the crystalline product (2.43
g, 85% yield): 1H NMR (400 MHz, DMSO-d6) δ ppm 1.69 (ddd, J =
17.2, 15.4, 8.1 Hz, 1 H), 1.87−2.28 (m, 3 H), 2.65−2.80 (m, 1 H),
4.90 (ddd, J = 50.0, 10.4, 2.8 Hz, 1 H); 19F NMR (376 MHz, DMSO-
d6) δ ppm −191.1. MS (ESI): [M + H]+ = 245.
1
volatiles were removed in vacuo. The residue was analyzed by H
NMR, dissolved in dichloromethane (10 mL), and the mixture was
cooled to −20 °C for 4 days, whereupon 130 mg (70% yield) of the
crystalline product could be obtained. A fraction of the crystalline
product was redissolved in dichloromethane and was allowed to
concentrate slowly at −20 °C to give crystals suitable for X-ray
analysis: 1H NMR (300 MHz, DMSO-d6) δ ppm 1.67−1.96 (m, 2 H),
2.01−2.36 (m, 2 H), 2.70−2.88 (m, 1 H), 3.57 (s, 3 H), 3.69 (s, 3 H),
5.21 (ddd, 1 H); no isomeric contamination could be detected as
judged by comparison with the respective isomeric mixture prepared
from 24 as described above. MS (ESI): [M + H]+ = 273. The crystals
were subjected to X-ray analysis revealing their (2S,4S) absolute
configuration (Figure 9).
(2R,4R)-2-Fluoro-4-(phosphonomethyl)pentanedioic Acid
(2R,4R)-21). To a solution of (2R,4R)-24 (340 mg, 752 μmol) in
THF (6.5 mL) was added 6 M aqueous hydrochloric acid (6.5 mL),
and the mixture was stirred at 100 °C for 1.5 h in a microwave oven.
1
The mixture was evaporated and analyzed by H NMR. To drive the
reaction to completion, the residue was treated with 6 M aqueous
hydrochloric acid (6.5 mL) and the mixture stirred again at 100 °C for
1.5 h. All volatiles were evaporated and the residue was purified by
preparative reversed-phase HPLC to give the target compound (158
mg) which solidified upon standing. An aliquot (20 mg) was subjected
to crystallization from MTBE/THF/hexane/toluene to give crystals
1
suitable for X-ray analysis: H NMR (600 MHz, DMSO-d6) δ ppm
1.69−1.79 (m, 1 H), 1.87−1.96 (m, 1 H), 2.02−2.12 (m, 1 H), 2.14−
2.25 (m, 1 H), 2.66−2.75 (m, 1 H), 5.01 (ddd, 1 H). MS (ESI): [M +
H]+ = 245. The crystals were subjected to X-ray analysis, revealing
their (2R,4R) absolute configuration (Figure 8).
Figure 9. Empirical formula: C8H14FO7P. Formula weight: 272.16 g/
mol. Temperature: 100(2) K. Wavelength: 1.541 78 Å. Crystal system:
monoclinic. Space group: P21. Unit cell dimensions: a = 6.135(4) Å, b
= 6.066(5) Å, c = 15.731(4) Å, β = 94.04(3)°. Z = 2. Goodness-of-fit
on F2: 1.086. Final R indices [I > 2σ(I)]: R1 = 0.0232, wR2 = 0.0614.
R indices (all data): R1 = 0.0234, wR2 = 0.0615. Absolute structure
parameter: 0.04(2). The crystal structure has been deposited into the
Cambridge Structural Database (CCDC 895588).
Dimethyl (RS)-2-{[tert-Butyl(dimethyl)silyl]oxy}-4-methyle-
nepentanedioate (26). To a cooled (0 °C) solution of the starting
alcohol 22 (2.00 g, 10.6 mmol) in dichloromethane (50 mL) was
added N-methylimidazole (1.40 equiv), followed by tert-butyldimethyl-
chlorosilane (TBS chloride, 1.30 equiv). The cooling bath was
removed, and the mixture was allowed to stir overnight at room
temperature. Subsequently, the double volume of dichloromethane
was added and the solution was washed with brine twice. The organic
layer was dried over sodium sulfate and evaporated. The residue was
purified by column chromatography on silica to give 2.69 g of the
Figure 8. Empirical formula: C6H10F1O7P1. Formula weight: 244.11g/
mol. Temperature: 100(2) K. Wavelength: 1.541 78 Å. Crystal system:
monoclinic. Space group: P21. Unit cell dimensions: a = 8.9000(12) Å,
b = 5.3584(8) Å, c = 9.8181(14) Å, β = 91.273(7)°. Z = 2. Goodness-
of-fit on F2: 1.064. Final R indices [I > 2σ(I)]: R1 = 0.0227, wR2 =
0.0571. R indices (all data): R1 = 0.0231, wR2 = 0.0574. Absolute
structure parameter: 0.043(19). The crystal structure has been
deposited into the Cambridge Structural Database (CCDC 895587).
1
target compound (84% yield): H NMR (300 MHz, CDCl3) δ ppm
0.00 (s, 3 H), 0.04 (s, 3 H), 0.87 (s, 9 H), 2.59 (dd, 1 H), 2.83 (dd, 1
H), 3.72 (s, 3 H), 3.76 (s, 3 H), 4.43 (dd, 1 H), 5.67 (m, 1 H), 6.26
(m, 1 H). MS (EI): [M + H]+ = 303.
Dimethyl rac-2-{[Bis(benzyloxy)phosphoryl]methyl}-4-{[tert-
butyl(dimethyl)silyl]oxy}pentanedioate (27). A mixture of 26
(1.88 g, 6.22 mmol), dibenzyl phosphite (1.72 mL, 1.25 equiv), and
finely powdered potassium carbonate (1.29 g, 1.50 equiv) in DMF (40
mL) was heated to 90 °C for 2 h and was then allowed to stir at room
temperature for 60 h. The mixture was then evaporated and
partitioned between EtOAc and 5% aqueous citric acid. The organic
layer was washed with half-concentrated brine, dried over sodium
sulfate, and evaporated. The residue was purified by column
The preparation was scaled in analogy to the protocol described for
the three preceding stereoisomers of 21 to give 0.93 g (78% yield) of
1
(2R,4R)-21 from (2R,4R)-24 (2.10 g, 4.64 mmol): H NMR (400
MHz, DMSO-d6) δ ppm 1.73 (ddd, J = 17.5, 15.2, 6.8 Hz, 1 H), 1.92
(ddd, J = 18.3, 15.2, 6.8 Hz, 1 H), 2.00−2.29 (m, 2 H), 2.63−2.77 (m,
1 H), 5.02 (ddd, J = 49.3, 9.0, 3.8 Hz, 1 H).
Dimethyl (2S,4S)-2-Fluoro-4-(phosphonomethyl)-
pentanedioate ((2S,4S)-25). To a solution of (2S,4S)-24 (310 mg,
H
dx.doi.org/10.1021/jm300710j | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
chromatography over silica to give the desired product (3.43 g, 98%
yield) as an oil: H NMR (300 MHz, CDCl3) δ ppm −0.02 to 0.08
(m, 6 H), 0.88 (s, 9 H, major diastereomer), 0.89 (s, 9 H, minor
diastereomer), 1.84−2.35 (m, 4 H), 2.86−3.12 (m, 1 H), 3.53 (s, 3 H,
major diastereomer), 3.55 (m, 3 H, minor diastereomer), 3.67 (m, 3
H, major diastereomer), 3.68 (m, 3 H, minor diastereomer), 4.14−
4.24 (m, 1 H), 4.90−5.07 (m, 4 H), 7.28−7.40 (m, 10 H).
passed through a C18 light SPE (preconditioned with 5 mL of ethanol
and with 10 mL of water). The C18 light SPE was washed with 5 mL
of water and eluted with 1 mL of MeCN. The eluted solution was
diluted with 3 mL of water and purified over preparative HPLC (ACE
5 μm C18, 250 mm × 10 mm, isocratic 60% MeCN in 40% water +
0.1% TFA, flow of 4 mL/min). The product peak was collected and
diluted with 20 mL of water and passed through a C18 light SPE
(preconditioned with 5 mL of ethanol and with 10 mL of water). The
SPE was washed with 5 mL of water and was eluted with 1 mL of
acetonitrile. The acetonitrile solution was dried under gentle N2
stream for 10 min at 100 °C. An amount of 500 μL of 6 M HCl
was added, and the mixture was incubated for 15 min at 120 °C. After
cooling, the reaction mixture was diluted with 1 mL of water and
passed through a cartridge containing AG11A8 resin (ion retardation
resin, ∼11g) connected in series with a C18 light SPE (preconditioned
with 5 mL of ethanol and with 10 mL of water). The cartridges were
then washed with saline (5 mL) and eluents were collected to give the
desired product [18F]21 (509 MBq, 29% d.c.) with a specific activity of
>25GBq/μmol with a radiochemical purity of >95% as determined by
HPLC with the ZIC-HILIC column and by TLC (silica, 1-butanol/
acetic acid/water/ethanol (12:3:5:1.5)). The final preparation was
always found to contain 3−4% free [18F]fluoride.
NAALADase Assay. NAALAdase activity was assayed essentially as
described previously.38 Briefly, crude membrane extracts from LNCaP
tissue culture cells (obtained from ATCC) were prepared by
homogenization of the cells in lysis buffer (50 mM Tris-HCl, 150
mM NaCl, pH 7.4, glass Teflon potter, 300 rpm), separation of the
nuclear fraction (1000g, 10 min), precipitation of the membrane
fraction (100000g, 20 min), and resuspension in solubilization buffer
(lysis buffer plus 1% Triton X-100). After a final centrifugation step
(100000g, 20 min) the supernatant containing solubilized membrane
proteins was frozen in liquid nitrogen and stored at −80 °C. Release of
[3H]glutamate from [3H]-N-acetylaspartylglutamate (NAAG, assay
concentration of 0.5 μM) in 50 mM Tris-HCl buffer, pH 7.4, after 30
min at 37 °C was measured using 2 μg/mL solubilized membrane
protein. The enzymatic reaction was stopped by adding cold methanol.
Substrate and product were resolved by cation-exchange liquid
chromatography. Prefilled AG 50W-X8 columns (200−400 mesh,
(BIO RAD, Hercules, CA) were equilibrated in 0.2 M HCl before
loading with the enzymatic mixture. After a wash with 9 mL of 0.2 M
HCl, fractions containing [3H]glutamate were eluted with 6 mL of 2
M HCl, mixed with liquid scintillation fluid, and analyzed in a liquid
scintillation analyzer. Under these assay conditions less than 20% of
substrate was cleaved. Inhibition curves were generated by semi-
logarithmic plotting of enzyme activity versus inhibitor concentration
and IC50 values determined by curve fitting using the software GraFit5
(version 5.0.6). Reported values represent the mean and standard
deviation of two to four replicates of the entire inhibition study. Ki
values were calculated according to the method of Cheng and
Prusoff.39
1
Dimethyl rac-2-{[Bis(benzyloxy)phosphoryl]methyl}-4-
(tosyloxy)pentanedioate (28). To an ice-cooled solution of 27
(2.50 g, 4.43 mmol) in THF (100 mL) was added hydrogen fluoride−
pyridine complex (5.75 mL, 50 equiv) under ice cooling, and the
mixture was then stirred at room temperature for 4 h. Subsequently, a
solution prepared from potassium hydroxide (18 g), boric acid (39.5
g), and water (180 mL) was carefully added until the pH reached 7.
The mixture was carefully concentrated to largely remove THF,
partitioned between water and MTBE, and the organic layer was
washed with brine, dried over sodium sulfate, and evaporated. The
residual crude γ-hydroxy ester (2.21 g) was immediately dissolved in
pyridine (50 mL). p-Toluenesulfonic anhydride (3.19 g, 2.00 equiv
based on crude hydroxy ester) was added, and the mixture was stirred
for 16 h at room temperature. The mixture was then partitioned
between water and MTBE, and the organic layer was washed with 5%
aqueous citric acid and with brine, dried over sodium sulfate, and
evaporated. The residue was purified by column chromatography to
give the target compound as an oily mixture of stereoisomers (1.89 g,
1
70% overall yield): H NMR (400 MHz, CDCl3) δ ppm 1.83−2.07
(m, 1 H), 2.10−2.33 (m, 3 H), 2.43 (s, 3 H), 2.81−2.98 (m, 1 H), 3.51
(s, 3 H, minor diastereomer), 3.53 (s, 3 H, major diastereomer), 3.57
(s, 3 H, major diastereomer), 3.62 (s, 3 H, minor diastereomer), 4.88−
5.05 (m 5 H), 7.28−7.41 (m, 12 H), 7.75−7.81 (m, 2 H). MS (ESI):
[M + H]+ = 605.
Dimethyl (RS)-2-Methylene-4-(tosyloxy)pentanedionate
(29). To a cooled (0 °C) solution of the starting alcohol 22 (1.00
g, 5.31 mmol) in pyridine (20 mL) was added p-toluenesulfonic
anhydride (3.47 g, 2.00 equiv), and the mixture was stirred for a period
ranging from 2 h at 0 °C to 16 h at room temperature. The reaction
mixture was then partitioned between water and dichloromethane.
The organic layer was then washed with water and brine, dried over
sodium sulfate, and evaporated. The crude product (1.80 g, 99% yield)
was used without further purification: 1H NMR (400 MHz, CDCl3): δ
ppm 2.45 (s, 3 H), 2.69 (dd, 1 H), 2.90 (dd, 1 H), 3.67 (s, 3 H), 3.68
(s, 3 H), 5.06 (dd, 1 H), 5.67 (s, 1 H), 6.21 (s, 1 H), 7.32 (d, 2 H),
7.75 (d, 2 H). MS (ESI): [M + H]+ = 343.
Dimethyl rac-2-{[Bis(benzyloxy)phosphoryl]methyl}-4-
(tosyloxy)pentanedioate (28, Optimized Preparation). To a
solution of 26 (1.32 g, 3.81 mmol) in THF (40 mL) were added DBU
(1,8-diazabicyclo(5.4.0)undec-7-ene, 717 μL, 1.25 equiv) and dibenzyl
phosphite (1.70 mL, 2.00 equiv), and the mixture was stirred at room
temperature for 1.5 h. The mixture was concentrated in vacuo and
then partitioned between EtOAc and 5% aqueous citric acid. The
organic layer was washed with brine, dried over sodium sulfate, and
evaporated. The residue was purified by column chromatography on
silica (hexane/EtOAc) to give the desired product (1.42 g, 61% yield)
1
as a mixture of stereoisomers: H NMR (400 MHz, CDCl3) δ ppm
In Vivo Animal Studies. All animal experiments were performed
in compliance with the current version of the German law concerning
animal protection and welfare. PET imaging studies with tumor
bearing nude mice were performed using the Inveon small animal
PET/CT scanner (Siemens, Knoxville, TN). Male Balb/c nude mice
(n = 5; Charles-River, Sulzfeld, Germany) were used for subcutaneous
inoculation of human xenografts. At 3−4 days before injection of
tumors cells, a testosterone pellet (Innovative Research of America,
catalog no. NA-151) was implanted. LNCaP (human prostate cancer)
tumor cells (1 × 107 cells per animal in 100 μL of Matrigel) were
injected subcutaneously into the right flank and allowed to grow for
approximately 4 weeks, reaching sizes of 5−8 mm in diameter. Animals
(n = 4) were injected intravenously with 10 MBq compound [18F]21.
Immediately prior to imaging, the animals were anesthetized with
isoflurane, and the PET data were acquired over defined time periods
after injection.
1.83−2.07 (m, 1 H), 2.10−2.33 (m, 3 H), 2.43 (s, 3 H), 2.81−2.98
(m, 1 H), 3.51 (s, 3 H, minor diastereomer), 3.53 (s, 3 H, major
diastereomer), 3.57 (s, 3 H, major diastereomer), 3.62 (s, 3 H, minor
diastereomer), 4.88−5.05 (m 5 H), 7.28−7.41 (m, 12 H), 7.75−7.81
(m, 2 H). MS (ESI): [M + H]+ = 605.
rac-2-[18F]Fluoro-4-(phosphonomethyl)pentanedioic Acid
([18F]21). [18F]Fluoride (4.2 GBq) was immobilized on a precondi-
tioned QMA (Waters) cartridge (preconditioned by washing the
cartridge with 5 mL of 0.5 M potassium carbonate and 10 mL of
water), The [18F]fluoride was eluted using a solution of cesium
carbonate (2.3 mg) in 500 μL of water and Kryptofix 2.2.2. (5 mg) in
1500 μL of acetonitrile. This solution was dried at 120 °C with stirring
under vacuum, with a stream of nitrogen. Additional acetonitrile (1
mL) was added, and the drying step was repeated. A solution of
precursor 27 (4 mg) in DMSO (500 μL) was added and heated at 120
°C for 15 min. The mixture was diluted with 15 mL of water and
I
dx.doi.org/10.1021/jm300710j | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
(10) Vees, H.; Buchegger, F.; Albrecht, S.; Khan, H.; Husarik, D.;
Zaidi, H.; Soloviev, D.; Hany, T. F.; Miralbell, R. 18F-choline and/or
11C-acetate positron emission tomography: detection of residual or
progressive subclinical disease at very low prostate-specific antigen
values (<1 ng/mL) after radical prostatectomy. BJU Int. 2007, 99,
1415−1420.
(11) Schuster, D. M.; Votaw, J. R.; Nieh, P. T.; Yu, W.; Nye, J. A.;
Master, V.; Bowman, F. D.; Issa, M. M.; Goodman, M. M. Initial
experience with the radiotracer anti-1-amino-3-18F-fluorocyclobutane-
1-carboxylic acid with PET/CT in prostate carcinoma. J. Nucl. Med.
2007, 48, 56−63.
(12) Larson, S. M.; Morris, M.; Gunther, I.; Beattie, B.; Humm, J. L.;
Akhurst, T. A.; Finn, R. D.; Erdi, Y.; Pentlow, K.; Dyke, J.; Squire, O.;
Bornmann, W.; McCarthy, T.; Welch, M.; Scher, H. Tumor
localization of 16beta-18F-fluoro-5alpha-dihydrotestosterone versus
18F-FDG in patients with progressive, metastatic prostate cancer. J.
Nucl. Med. 2004, 45, 366−373.
(13) Tehrani, O. S.; Muzik, O.; Heilbrun, L. K.; Douglas, K. A.;
Lawhorn-Crews, J. M.; Sun, H.; Mangner, T. J.; Shields, A. F. Tumor
imaging using 1-(2′-deoxy-2′-18F-fluoro-beta-D-arabinofuranosyl)-
thymine and PET. J. Nucl. Med. 2007, 48, 1436−1441.
(14) Byun, Y; Mease, R. C.; Lupold, S; Pomper, M. G. Recent
Development of Diagnostic and Therapeutic Agents Targeting
Glutamate Carboxypeptidase II (GCPII). In Drug Design of Zinc-
Enzyme Inhibitors; Superan, C. T., Winum, J.-Y., Eds.; Wiley: Hoboken,
NJ, 2009.
(15) Tsukamoto, T; Wozniak, K. M.; Slusher, B. S. Progress in the
discovery and development of glutamate carboxypeptidase II
inhibitors. Drug Discovery Today 2007, 12, 767−776.
(16) Foss, C. A.; Mease, R. C.; Fan, H.; Wang, Y.; Ravert, H. T.;
Dannals, R. F.; Olszewski, R.; Heston, W. D.; Kozikowski, A. P.;
Pomper, M. G. Radiolabeled small molecule ligands for prostate-
specific membrane antigen: in vivo imaging in experimental models of
prostate cancer. Clin. Cancer Res. 2005, 11, 4022−402.
(17) Mease, R. C.; Dusich, C. L; Foss, C. A; Ravert, H. T.; Dannals,
R. F.; Seidel, J.; Prideaux, A.; Fox, J. J.; Sgouros, G.; Kozikowski, A. P.;
Pomper, M. G. N-[N-[(S)-1,3-Dicarboxypropyl]carbamoyl]-4-[18F]-
fluorobenzyl-L-cysteine, [18F]DCFBC: a new imaging probe for
prostate cancer. Clin. Cancer Res. 2008, 14 (10), 3036−3043.
(18) Maresca, K. P.; Hillier, S. M.; Fernia, F. J.; Keith, D.; Barone, C.;
Joyal, J. L.; Zimmerman, C. N.; Kozikowski, A. P.; Barrett, J. A.;
Eckelman, W. C.; Babich, J. W. A series of halogenated heterodimeric
inhibitors of prostate specific membrane antigen (PSMA) as
radiolabeled probes for targeting prostate cancer. J. Med. Chem.
2009, 52 (2), 347−357.
AUTHOR INFORMATION
■
Corresponding Author
*Phone: +49 (0) 468 192114. Fax: (+) 49 (0) 468 992114. E-
mail: keith.graham@bayer.com.
Notes
The authors declare the following competing financial
interest(s): All the authors are employees of Bayer Healthcare.
ACKNOWLEDGMENTS
■
The authors acknowledge the excellent technical assistance of
Selahattin Ede, Mario Mandau, Marion Zerna, Rene Zernicke,
Jorg Jannsen, Eva-Maria Bickel, Uwe Rettig, Yvonne Duchstein,
̈
Oliver Schenk, and Anika Hans.
ABBREVIATIONS USED
■
CAD, Corona aerosol detection; DBU, 1,8-diazabicycloundec-
7-ene; DMF, dimethylformamide; DMSO, dimethyl sulfoxide;
EtOAc, ethyl acetate; HPLC, high pressure liquid chromatog-
raphy; MTBE, methyl tert-butyl ether; NAAG, N-acetylaspar-
tylglutamate; NAALADase, N-acetylated α linked acidic
dipeptidase; nca, no-carrier-added; NMR, nuclear magnetic
resonance; PET, positron emission tomography; PTC, phase
transfer catalyst; 2-PMPA, 2-(phosphonomethyl)pentanedioic
acid; PSA, prostate specific antigen; PSMA, prostate specific
membrane antigen; SPECT, single photon emission computed
tomography; SPE, solid phase extraction; TBAF, tetrabutylam-
monium fluoride; THF, tetrahydrofuran
REFERENCES
■
(1) Jemal, A; Siegel, R; Xu, J; Ward, E. Cancer Statistics, 2010. Ca
Cancer J. Clin. 2010, 60 (Sep−Oct), 277−300.
(2) Tasch, J.; Gong, M.; Sadelain, M.; Heston, W. D. A unique folate
hydrolase, prostate-specific membrane antigen (PSMA): a target for
immunotherapy. Crit. Rev. Immunol. 2001, 21, 249−261.
(3) Schulke, N.; Varlamova, O. A.; Donovan, G. P.; Ma, D.; Gardner,
J. P.; Morrissey, D. M.; Arrigale, R. R.; Zhan, C.; Chodera, A. J.;
Surowitz, K. G.; Maddon, P. J.; Heston, W. D.; Olson, W. C. The
homodimer of prostate-specific membrane antigen is a functional
target for cancer therapy. Proc. Natl. Acad. Sci. U.S.A. 2003, 100,
12590−12595.
(4) Chang, S. S.; Reuter, V. E.; Heston, W. D.; Bander, N. H.; Grauer,
L. S.; Gaudin, P. B. Five different anti-prostate-specific membrane
antigen (PSMA) antibodies confirm PSMA expression in tumor
associated neovasculature. Cancer Res. 1999, 59, 3192−3198.
(5) Lapidus, R. G.; Tiffany, C. W.; Isaacs, J. T.; Slusher, B. S. Prostate
specific membrane antigen (PSMA) enzyme activity is elevated in
prostate cancer cells. Prostate 2000, 45, 350−354.
(6) Scher, B.; Seitz, M.; Albinger, W.; Tiling, R.; Scherr, M.; Becker,
H. C.; Souvatzogluou, M.; Gildehaus, F. J.; Wester, H. J.; Dresel, S.
Value of 11C-choline PET and PET/CT in patients with suspected
prostate cancer. Eur. J. Nucl. Med. Mol. Imaging 2007, 34, 45−53.
(7) Rinnab, L.; Mottaghy, F. M.; Blumstein, N. M.; Reske, S. N.;
Hautmann, R. E.; Hohl, K.; Moller, P.; Wiegel, T.; Kuefer, R.;
Gschwend, J. E. Evaluation of [11C]-choline positron-emission/
computed tomography in patients with increasing prostate-specific
antigen levels after primary treatment for prostate cancer. BJU Int.
2007, 100, 786−793.
(19) Chen, Y.; Foss, C. A.; Byun, Y.; Nimmagadda, S.; Pullambhatla,
M.; Fox, J. J.; Castanares, M.; Lupold, S. E.; Babich, J. W.; Mease, R.
C.; Pomper, M. G. Radiohalogenated prostate-specific membrane
antigen (PSMA)-based ureas as imaging agents for prostate cancer. J.
Med. Chem. 2008, 51 (24), 7933−7943.
(20) Banerjee, S. R.; Pullambhatla, M.; Byun, Y.; Nimmagadda, S.;
Green, G.; Fox, J. J.; Horti, A.; Mease, R. C.; Pomper, M. G. 68Ga-
Labeled inhibitors of prostate-specific membrane antigen (PSMA) for
imaging prostate cancer. J. Med. Chem. 2010, 53 (14), 5333−5341.
(21) Lapi, S. E.; Wahnishe, H.; Pham, D.; Wu, L. Y.; Nedrow-Byers, J.
R.; Liu, T.; Vejdani, K.; VanBrocklin, H. F.; Berkman, C. E.; Jones, E.
F. Assessment of an 18F-labeled phosphoramidate peptidomimetic as a
new prostate-specific membrane antigen-targeted imaging agent for
prostate cancer. J. Nucl. Med. 2009, 50 (12), 2042−2048.
(22) Misra, P.; Humblet, V.; Pannier, N.; Maison, W.; Frangioni, J. V.
Production of multimeric prostate-specific membrane antigen small-
molecule radiotracers using a solid-phase 99mTc preloading strategy. J.
Nucl. Med. 2007, 48 (8), 1379−1389.
(8) Reske, S. N.; Blumstein, N. M.; Neumaier, B.; Gottfried, H. W.;
Finsterbusch, F.; Kocot, D.; Moller, P.; Glatting, G.; Perner, S. Imaging
prostate cancer with 11C-choline PET/CT. J. Nucl. Med. 2006, 47,
1249−1254.
(23) Banerjee, S. R.; Foss, C. A.; Castanares, M.; Mease, R. C.; Byun,
Y.; Fox, J. J.; Hilton, J.; Lupold, S. E.; Kozikowski, A. P.; Pomper, M.
G. Synthesis and evaluation of technetium-99m- and rhenium-labeled
inhibitors of the prostate-specific membrane antigen (PSMA). J. Med.
Chem. 2008, 51 (15), 4504−4517.
(9) Zophel, K.; Kotzerke, J. Is 11C-choline the most appropriate
tracer for prostate cancer? Against. Eur. J. Nucl. Med. Mol. Imaging
2004, 31, 756−759.
J
dx.doi.org/10.1021/jm300710j | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
(24) Vitharana, D.; France, J. E.; Scarpetti, D.; Bonneville, G. W.;
Majer, P.; Tsukamoto, T. Synthesis and biological evaluation of (R)-
and (S)-2-(phosphonomethyl)pentanedioic acids as inhibitors of
glutamate carboxypeptidase II. Tetrahedron: Asymmetry 2002, 13,
1609−1614.
(25) Jackson, P. F.; Cole, D. C.; Slusher, B. S.; Stetz, S. L.; Ross, L. E.;
Donzanti, B. A.; Trainor, D. A. Design, synthesis, and biological
activity of a potent inhibitor of the neuropeptidase N-acetylated α-
linked acidic dipeptidase. J. Med. Chem. 1996, 39 (15), 619−622.
(26) Kozikowski, A. P.; Zhang, J.; Nan, F.; Petukhov, P. A.;
Grajkowska, E.; Wroblewski, J. T.; Yamamoto, T.; Bzdega, T.;
Wroblewska, B.; Neale, J. H. Synthesis of urea-based inhibitors as
active site probes of glutamate carboxypeptidase II: efficacy as
analgesic agents. J. Med. Chem. 2004, 47 (7), 1729−1738.
(27) Wang, H.; Byun, Y.; Barinka, C.; Pullambhatla, M.; Bhang, H.
C.; Fox, J. J.; Lubkowski, J.; Mease, R. C.; Pomper, M. G. Bioisosterism
of urea-based GCPII inhibitors: synthesis and structure−activity
relationship studies. Bioorg. Med. Chem. Lett. 2010, 20, 392−397.
(28) For reviews see the following: (a) Vorbruggen, H. Helv. Chim.
̈
Acta 2011, 94, 947−965. (b) Takamatsu, S.; Katayama, S.; Hirose, N.;
De Cock, E.; Schelkens, G.; Demillequand, M.; Brepoels, J.; Isawa, K.
Nucleotsides, Nucleotides Nucleic Acids 2002, 21, 849−861.
(29) For a more extended discussion of stereochemical assignments
see the following: Graham, K.; Kettschau, G.; Lesche, R.; Gromov, A.;
Bohnke, N.; Hassfeld, J. WO 2011/073286, 2011.
̈
(30) Paschlau, J. Schwerpunkt chromatographieCorona CAD
eine neue detektionsmethode fur die HPLC. GIT Labor-Fachz. 2005,
49, 32−33.
(31) Gamache, P. H.; McCarthy, R. S.; Freeto, S. M.; Asa, D. J.;
Woodcock, M. J.; Laws, K.; Cole, R. O. HPLC analysis of non-volatile
analytes using charged aerosol detection. LC-GC Eur. 2005, 18, 345−
354.
(32) Robinson, M. B.; Blakely, R. D.; Couto, R.; Coyle, J. T.
Hydrolysis of the brain dipeptide N-acetyl-L-aspartyl-L-glutamate. J.
Biol. Chem. 1987, 262, 14498−14506.
(33) Troyer, J. K.; Beckett, M. L.; Wright, G. L., Jr. Detection and
characterization of the prostate specific membrane antigen (PSMA) in
tissue extracts and body fluids. Int. J. Cancer. 1995, 62, 552−558.
(34) Silver, D. A.; Pellicer, I.; Fair, W. R.; Heston, W. D.; Cordon-
Cardo, C. Prostate-specific membrane antigen expression in normal
and malignant human tissues. Clin. Cancer Res. 1997, 3, 81−85.
(35) Slusher, B. S.; Tsai, G.; Yoo, G.; Coyle, J. T. Immunocy-
tochemical localization of the N-acetyl-aspartyl-glutamate (NAAG)
hydrolyzing enzyme N-acetylated alpha-linked acidic dipeptidase
(NAALADase). J. Comp. Neurol. 1992, 315, 217−229.
(36) PROTEUM2, version 2009; Bruker, AXS Inc.: Madison, WI,
U.S., 2009.
(37) Sheldrick, G. M. A short history of SHELX. Acta Crystallogr.
2008, A64, 112−122.
(38) Fuhrman, S.; Palkovits, M.; Cassidy, M.; Neale, J. H. The
regional distribution of N-acetylaspartylglutamate (NAAG) and
peptidase activity against NAAG in the rat nervous system. J.
Neurochem. 1994, 62, 275−281.
(39) Cheng, H. C. The power issue: determination of KB or Ki from
IC50. A closer look at the Cheng−Prusoff equation, the Schild plot and
related power equations. J. Pharmacol. Toxicol. Methods 2001, 46, 61−
71.
K
dx.doi.org/10.1021/jm300710j | J. Med. Chem. XXXX, XXX, XXX−XXX