Novel synthesis of [1-11C]γ-vinyl-γ-aminobutyric acid ([1-11C]GVG) for pharmacokinetic studies of addiction treatment

Article abstract of DOI:10.1002/jlcr.544

γ-Vinyl-γ-aminobutyric acid (GVG, Vigabatrin), a suicide inhibitor of GABA-transaminase (GABA-T), has been suggested as a new drug for the treatment of substance abuse. In order to better understand its pharmacokinetics and potential side effects, we have developed a radiosynthesis of carbon-11 (t_1/2=20min) labeled GVG for positron emission tomographic (PET) studies. We report here a novel synthetic strategy to prepare the precursor and to efficiently label GVG with C-11. 5-Bromo-3-(carbobenzyloxy)amino-1-pentene was synthesized in five steps from homoserine lactone, including reduction and methylenation. This was used in a one-pot, two-step radiosynthesis. Displacement of bromide with no-carrier-added [¹¹C]cyanide followed by acid hydrolysis afforded [1-¹¹C]GVG with decay corrected radiochemical yields of 27 ± 9% (n=6, not optimized) with respect to [¹¹C]cyanide in a synthesis time of 45 min. Copyright

Full text of DOI:10.1002/jlcr.544

JOURNAL OF LABELLED COMPOUNDS AND RADIOPHARMACEUTICALS
J Label Compd Radiopharm 2002; 45: 199–211.
DOI: 10.1002/jlcr.544
Research Article
Novel synthesis of [1-11C]g-vinyl-g-
aminobutyric acid ([1–11C]GVG) for
pharmacokinetic studies of addiction
treatment
Zongren Zhang, Yu-Shin Ding*, Andrei R. Studenov,
Madina R. Gerasimov and Richard A. Ferrieri
Chemistry Department, Brookhaven National Laboratory, Upton, NY 11973,
USA
Summary
g-Vinyl-g-aminobutyric acid (GVG, Vigabatrin1), a suicide inhibitor of
GABA-transaminase (GABA-T), has been suggested as a new drug for the
treatment of substance abuse. In order to better understand its pharmacoki-
netics and potential side effects, we have developed a radiosynthesis of carbon-
11 (t1/2=20 min) labeled GVG for positron emission tomographic (PET)
studies. We report here a novel synthetic strategy to prepare the precursor and
to efficiently label GVG with C-11. 5-Bromo-3-(carbobenzyloxy)amino-1-
pentene was synthesized in five steps from homoserine lactone, including
reduction and methylenation. This was used in a one-pot, two-step radio-
synthesis. Displacement of bromide with no-carrier-added [11C]cyanide
followed by acid hydrolysis afforded [1-11C]GVG with decay corrected radio-
chemical yields of 27 Æ 9% (n=6, not optimized) with respect to [11C]cyanide
in a synthesis time of 45min. Copyright # 2002 John Wiley & Sons, Ltd.
Key Words: GVG; vinyl-GABA; Vigabatrin; GABA-T; C-11 positron emis-
sion tomography (PET)
*Correspondence to: Y.-S. Ding, Brookhaven National Laboratory, Chemistry Department,
Upton, NY 11973, USA.
Contract grant sponsor: Brookhaven National Laboratory; contract grant number: DE-AC02-
98CH10886
Contract grant sponsor: US Department of Energy and Office of Biological Environmental
Research; National Institute of Health
Copyright # 2002 John Wiley & Sons, Ltd.
Received 1 June 2001
Revised 18 October 2001
Accepted 27 October 2001

200
Z. ZHANG ET AL.
Introduction
g-Vinyl-g-aminobutyric acid (g-vinyl-GABA, GVG, Vigabatrin1) is a
suicide inhibitor of GABA transaminase (GABA-T).1,2 Inhibition of
GABA-T by GVG increases the concentration of g-aminobutyric acid
(GABA) in CNS3 which accounts for the antiepileptic effect of GVG.4,5
We have shown that GVG inhibits the effects of heroin, cocaine,
methamphetamine, nicotine and alcohol both biochemically (attenua-
tion of drug-induced neostriatal synaptic dopamine increase) and
behaviorally in rodents and primates. These data suggest that the use
of GVG may represent a new strategy for the treatment of substance
abuse.6–9 GVG has been used as an antiepileptic drug in several
countries. However, side effects of GVG, such as visual field
constriction,10 place a sense of urgency on better understanding the
mechanisms underlying its efficacy in addiction treatment and those
associated with vision impairment. Therefore, we developed a radio-
synthesis of [11C]GVG (t1/2 for C-11=20 min) for PET studies in order
to measure its pharmacokinetics in the brain and in the body using PET.
Several synthetic routes for the preparation of GVG including
radiolabeling with long-lived 14C (t1/2 for C-14=5730 year) have been
published; however, they are not suitable for C-11 radiosynthesis.11–15
For example, a procedure to prepare the key 2-oxopyrrolidine-5-
carboxaldehyde precursor that would in principle allow the incorpora-
tion of C-11 by employing the Wittig condensation with [11C]methyl-
triphenylphosphonium iodide (Scheme 1) was not straightforward.14–16
We experienced this in our initial trials, and other researchers (authors
in References14,15) made similar observations. Difficulties encountered
during the isolation of the aldehyde precursor, its instability for storage,
as well as the subsequent time-consuming Wittig reaction prompted us
to pursue another synthetic pathway.
Scheme 1. Previously reported synthesis of GVG via Wittig reaction14
Copyright # 2002 John Wiley & Sons, Ltd.
J Label Compd Radiopharm 2002; 45: 199–211

NOVEL SYNTHESIS OF
201
Scheme 2. Synthesis of [11C]GVG via bromide via [11C]cyanide displacement
We report here a novel synthetic route to prepare the precursor (from
homoserine lactone) and to efficiently label GVG with C-11 (Scheme 2).
The entire synthetic procedure should be extendable to the labeling of
the pharmacologically active S- form of GVG when S-homoserine
lactone is used as the starting material.
Results and discussion
Five-Bromo-3-(carbobenzyloxy)amino-1-pentene (6) was synthesized in
five steps from homoserine lactone (Scheme 2). Homoserine lactone
was chosen as the starting material in anticipation of developing a
synthetic procedure for both racemic mixture and individual enantio-
mers of [11C]GVG. The amino group of homoserine lactone 1 was
protected as a benzyl carbamate (Cbz-) 2. Compound 2 was obtained in
low yield when aqueous NaOH was used as a base (34%); however, a
significant improvement in yield (97%) was achieved when Et3N was
used as the base in nonaqueous conditions. Reduction of 2 with
diisobutylaluminum hydride (DIBAL-H) afforded 2-(carbobenzyl-
oxy)amino-4-butyrolactol (3) in 72% yield.17,18 Efforts were made to
optimize reaction conditions by modifying the ratios of DIBAL-H to 2,
reaction temperature and reaction time. 1H NMR of the product (3) was
Copyright # 2002 John Wiley & Sons, Ltd.
J Label Compd Radiopharm 2002; 45: 199–211

202
Z. ZHANG ET AL.
rather complex, most likely due to a mixture of diastereomers arising
from lactol–aldehyde equilibrium (although an aldehyde peak was
always present, as expected). Compound 3 was found to be stable in
solid state and neutral organic solvents. It appeared to be less stable in
acidic solution. Wittig methylenation of 3 with DMSO–NaH–
Ph3PCH3Br19,20 led to unidentifiable products; however, THF–BuLi–
Ph3PCH3Br21 or THF–NaNH2-Ph3PCH3Br22 afforded a crude mixture
containing-3-amino-4-penten-1-ol (4).
In order to facilitate product isolation, this crude mixture was further
treated with benzyl chloroformate to convert 4 to 5 (15% yield, not
optimized). We attribute the low yields of the Wittig reaction to the
cleavage of the Cbz- protecting group leading to unknown side
reactions. Ideally, a protecting group should survive all the chemical
manipulations prior to its deprotection, it should bear a chromophore
for easy detection and be readily removed at the end of synthesis. We
have tested a few other protecting groups, and the choice of benzyl
chloroformate to protect amino group, though not ideal, was the best so
far.23,24
Bromination of 5 with CBr4/PPh3 led to the desired precursor 5-
bromo-3-(carbobenzyloxy)amino-1-pentene (6) in 89% yield. Kinetic
studies of the cyanation of 6 showed that the reaction went to near 50%
completion in DMF at an equimolar ratio of 6 and LiCN in 10 min at
1108C. The reaction was complete without the formation of noticeable
byproduct in THF, even with excess LiCN under extended refluxing. No
desired product was found when KCN was used. Solubility of metal
cyanides might be a critical factor influencing the cyanation.25,26
DMSO, MeOH, EtOH and CH3CN were found to be poor solvents
for rapid synthesis. Reaction in DMSO lead to formation of significant
quantities of byproducts, while the rest of the solvents gave
unacceptable product yields (55%).
The isolation of 5 as the major product, when the reaction was carried
out in KCN/DMSO at 908C for 5h, suggests that this displacement
reaction might be sensitive to the presence of H2O. 4-(Carbobenzyloxy)-
amino-5-hexenenitrile (7) was isolated by treating compound 6 with
excess LiCN for 2 h in refluxing THF. Compound 7 afforded GVG (8)
as essentially the only product after being heated in concentrated DCl
(deuterated HCl, for the purpose of monitoring reaction via NMR) for
20 min. The alkene of GVG appeared to be unreactive toward
concentrated DCl at 1008C. Acrylamide and allyl cyanide underwent
partial hydrochlorination under the similar conditions.
Copyright # 2002 John Wiley & Sons, Ltd.
J Label Compd Radiopharm 2002; 45: 199–211

NOVEL SYNTHESIS OF
203
Gas phase [11C]cyanide was produced from a [11C]CO2 based on a
procedure described previously; it was then delivered to a hot cell.27–29 It
requires 15min for the radiosynthesis of [ 11C]cyanide after EOB.
Initially, [11C]cyanide was trapped in an aqueous solution of LiOH
(0.05M, 0.2 ml). Water was removed under reduced pressure as an
azeotrope along with CH3CN. [11C]4-(Carbobenzyloxy)amino-5-hex-
enenitrile ([11C]7) was rapidly generated (10 min) in DMF at 1108C.
However, the radiochemical yields were unreliable and depended on
whether or not the reaction was carried out under anhydrous
conditions. [11C]Cyanide trapping was eventually achieved at À58C
(ethanol–ice bath) with a solid mixture of Kryptofix and K2CO3 without
using a conventional aqueous trapping procedure.29 The efficiency for
the ‘‘dry’’ trapping procedure was 84 Æ 10% (n=6, the error is a
standard deviation). At this temperature, most of the unreacted NH3
from the system that may interfere with the synthesis would not be
trapped.30 This procedure would be advantageous to any moisture
sensitive radiosynthetic steps, as it was the case for our displacement
reaction. When a conventional aqueous trapping procedure was used,
any trace amount of water left, even after prolonged heating, resulted in
either no reaction or extremely low yields for displacement. Further-
more, the use of a column filled with P2O5 on silica support (Sicapent1)
could reduce the excess amount of ammonia in the gas stream from the
cyanide production system. However, directly trapping [11C]cyanide in
the reaction solvent (DMF), instead of dry trapping, would result in low
radiochemical yields of [11C]7, probably due to inhibition of the reaction
by ammonia that transferred from the cyanide system and broke
through the Sicapent trap.31
Hydrolysis of [11C]7 in a procedure similar to that for the synthesis of
GVG afforded [1-11C]GVG ([11C]8) quantitatively in 10 min. The
resulting mixture was neutralized with aqueous (NH4)2HPO4 and
injected directly onto an HPLC column. This procedure requires less
time and is much more efficient than evaporation of HCl prior to HPLC
purification.28,32 Because of the high solubility of (NH4)2HPO4 in water,
as much as 0.5ml of the 12 M HCl can be neutralized by a small volume
of (NH4)2HPO4 (ꢀ1 ml). The peak resolution on semipreparative
HPLC used for the radiotracer purification was not affected. Compared
to alternative quenching agents such as (NH4)2CO3, the use of
(NH4)2HPO4 is superior since it does not generate gas bubbles leading
to poor HPLC separation. Though the post-column collected fraction
of [11C]GVG could be directly injected into baboon (phosphate buffer
Copyright # 2002 John Wiley & Sons, Ltd.
J Label Compd Radiopharm 2002; 45: 199–211

204
Z. ZHANG ET AL.
solution), the eluent was evaporated in order to diminish the UV-
adsorbing volatile impurities in the solution.
This two-step, one-pot radiosynthesis}displacement of bromide with
[11C]cyanide followed by acid hydrolysis}afforded [11C]GVG with
decay corrected radiochemical yield with respect to [11C]cyanide
(27 Æ 9%, n=6, the error is the standard deviation, the synthesis was
not optimized) and specific radioactivity >0.5Ci/ mmol after 45min.
(Due to poor UV adsorption of GVG, more precise measurements
could not be achieved. Typical specific radioactivity at the EOB in our
previous radiosyntheses was 2.4–3.2 Ci/mmol.34) Radiochemical purity
was >98%. This simple and straightforward radiosynthesis provides
several advantages over the possible C-11 labeling of GVG via the
Wittig reaction (Scheme 1). The starting bromide precursor 6 for our
synthesis (Scheme 2) was very stable for storage, as compared to the
extremely unstable aldehyde precursor used in the Wittig reaction. In
addition to the difficulties encountered during the isolation of the
aldehyde, it was claimed that the aldehyde was so unstable that it had to
be used immediately after its preparation.15 All of this precludes the use
of the aldehyde as a precursor for our radiosynthesis. Our current one
pot radiosynthesis involves simple cyanide displacement and hydrolysis,
avoiding the extra labor and care needed for the Wittig reaction. The
radiochemical yields were consistently high, more reliable and
reproducible than that we have experienced with the Wittig reaction.
The procedure is fully applicable to the synthesis of the routine
preparations for human PET application.
Experimental
Chemicals were purchased from Aldrich Chemical Co. Authentic GVG
was bought from Aventis Co. Anhydrous THF was prepared by
distillation from sodium under nitrogen. Anhydrous DMF was
purchased in Sure-seal1 bottles. Techniques for preparing and handling
moisture sensitive compounds were followed.33 Macherey-Nagel poly-
gram sil G/UV254 plastic-back TLC plates (4 Â 8 cm) were used. NMR
spectra were recorded with a Brucker 400 MHz NMR spectrometer.
Proton chemical shifts were reported in ppm relative to TMS (d 0 ppm)
or DSS (d 0.015ppm). Melting points were determined with a Fisher–
Johns melting point apparatus. Mass spectra were recorded with a
Finnegan-Mat GC-MS 5100 mass spectrometer using electron impact
Copyright # 2002 John Wiley & Sons, Ltd.
J Label Compd Radiopharm 2002; 45: 199–211

NOVEL SYNTHESIS OF
205
ionization at 70 eV. Microanalyses were conducted by Schwarzkopf
Microanalytical Laboratory, Inc.
[11C]Cyanide was produced from [11C]CO2 by an in-house made
automatic synthesizer following a standard synthetic route27–29 and
passed through Sicapent1 (EM Science, Gibbstown, NJ) before being
trapped in the reaction vessel. Purification and analyses of radioactive
mixtures were performed with HPLC pump, in-line UV detector (both
from Knauer) and radioactivity detector. Peak areas were measured
using two Hewlett-Packard 3390A recording integrators. A short
wavelength ultraviolet lamp, NaI well counter and automatic TLC
scanner (Berthold Automatic TLC Linear Analyzer) were used as UV-
and radioactivity detectors for TLC analysis. The radiochemical purity
was determined by radio TLC and analytical radio HPLC in the
presence of authentic unlabeled compound as a carrier.
Attempt of [6-11C]GVG synthesis (Scheme 1)
Attempt of GVG synthesis via Wittig reaction was performed according
to the procedures described in References.14,15
2-(Carbobenzyloxy)amino-4-butyrolactone (2, Scheme 2)
Method A: Solution of Benzyl chloroformate (4.25ml, 30 mmol) in
aqueous NaOH (2 N, 17.5ml) was slowly (1 h) added to a stirred
solution of racemic homoserine lactone [( Æ )-a-amino-g-butyrolactone
hydrobromide (1, 4.55 g, 25 mmol) in aqueous NaOH (2.5 M, 10 ml)
cooled in ice-water bath. The mixture was stirred for another 4 h, and
the product was extracted with CHCl3 (4 Â 60 ml). The extracts were
dried (MgSO4), and the solvent was rotary-evaporated yielding a
colorless oil that changed to a solid upon addition of diethyl ether
(20 ml). The crystals were filtered and dried to give 2.0 g (34%) of 2 as a
white solid.
Method B: Benzyl chloroformate (6.34 ml, 44 mmol) followed by
triethylamine (6.0 ml, 43 mmol) were added dropwise to a stirred
solution of 1 (6.08 g, 33 mmol ) and triethylamine (5.5 ml, 39 mmol) in
CHCl3 (60 ml) cooled in ice-water bath. The solution was left overnight
at room temperature while stirring. The mixture was acidified (12 M
HCl) and washed with water (3 Â 20 ml). After the aqueous phase was
backwashed with diethyl ether (3 Â 40 ml), organic fractions were
combined, dried (MgSO4) and rotary-evaporated leaving a crystalline
Copyright # 2002 John Wiley & Sons, Ltd.
J Label Compd Radiopharm 2002; 45: 199–211

206
Z. ZHANG ET AL.
solid. The crystals were washed with 1 : 1 diethyl ether: hexane and
vacuum-dried to give 7.65g (97%) of 2 as white solids. Mp 105–1098C.
1H NMR (CDCl3) d 7.32 (m, 5H, C 6H5-), 5.34 (bs, 1 H, -NH-), 5.12 (s,
2 H, -CH2Ph), 4.43 (m, 2 H, -CH2O-), 4.25(m, 1 H, -C H(NHZ)-), 2.76
(m. 1 H, -OCH2CH2-), 2.33 (m, 1 H, -OCH2CH2-). MS m/z 235(M +),
91 (PhCH2), 108 (PhCH2OH). Anal. Calcd. for C12H13NO4 (%):
C, 61.27; H, 5.57; N, 5.95. Found: C, 61.30; H, 5.92; N, 5.75.
2-(Carbobenzyloxy)amino-4-butyrolactol (3)
Compound 2 (2.0 g, 8.5mmol) was added to a flame dried flask under
N2 atmosphere. After the introduction of toluene (anhydrous, 20 ml) the
flask was cooled to À728C (acetone–dry ice). Diisobutylaluminum
hydride (DIBAL-H, 24 ml of 1.0 M solution in CH2Cl2) was added over
30 min while the reaction mixture was stirred. After stirring for another
6.5h at À728C, the solution was treated with ethyl acetate (50 ml),
methanol (1 ml), water (5ml) and excess of NaHCO 3 and stirred for 2 h
at room temperature. The residue was filtered and washed with ethyl
acetate. The washes were combined and concentrated to give an oil as a
crude product. Chromatography on silica gel column using ethyl acetate
followed by crystallization in diethyl ether–pentane produced 1.45g
1
(72%) 3 in a form of white crystals. Mp 80–828C. H NMR (CDCl3) d
9.58 (s, 1 H, -CHO), 7.31 (m, 5H, -C 6H5), 5.41 (m, 1 H, -NHZ), 5.28 (s,
1 H, -OH), 5.10 (m, 3 H, PhCH2-, HOCH-), 4.14-3.81 (m, 3 H,
=CHNZ, -CH2OH), 2.27-2.03 (m. 1 H, -CH2CH2OH), 1.85-1.72 (m,
1 H, -CH2CH2OH). MS m/z 219 (M+-H2O), 208 (M+ -MeCH2), 146
(M+-PhCH2), 108 (PhCH2OH), 91 (PhCH2). Anal. Calcd. for
C12H15NO4: C, 60.75; H, 6.37; N, 5.90. Found: C, 60.94; H, 6.59; N,
5.87.
3-Amino-4-penten-1-ol (4)
Synthesis of this compound was similar to reaction described in the
synthesis of 5 (see next section). After methylenation of 3 (3.43 g,
14.5mmol) using Wittig reaction, 0.124 g (4%) of 4 was isolated by
silica gel flash chromatography using 5–50% methanol in ethyl acetate
1
as eluent. H NMR (CD3OD) d 5.82 (ddd, J=5.8, 10.3, 22.7 Hz, 1 H,
-CH=), 5.24 (d, J=21.5Hz, 1 H, C H2=), 5.20 (d, J=15.9 Hz, 1 H,
CH2=), 4.21 (m, 2 H, HOCH2-), 4.03 (m, 1 H, -CH(NH)2-), 2.09 (m,
1 H, -CH2CH2OH), 1.80 (m, 1 H, -CH2CH2OH). MS m/z 101 (M+).
Copyright # 2002 John Wiley & Sons, Ltd.
J Label Compd Radiopharm 2002; 45: 199–211

NOVEL SYNTHESIS OF
207
3-(Carbobenzyloxy)amino-4-penten-1-ol (5)
Methyltriphenylphosphonium bromide (6.79 g, 19 mmol) was added to
a flame dried flask under the N2 atmosphere. After addition of THF
(25ml) the flask was cooled in ice–water bath, and butyllithium (1.6 M
in hexanes, 11.8 ml, 18.9 mmol) was added dropwise. The mixture was
stirred for 30 min, and the solution of 3 (1.50 g, 6.3 mmol) in THF (8 ml)
was added over 30 min. The mixture was stirred at room temperature
for another 6.5h. The flask was cooled in an ice-water bath, and water
(0.4 ml) was added followed by the slow addition of the solution of
benzyl chloroformate (2.7 ml, 18.9 mmol) and triethylamine (2.6 ml,
18.7 mmol) in THF (10 ml). After 16 h stirring at room temperature, the
solid was filtered and washed with THF (2 Â 20 ml). THF fractions were
combined, acidified (HCl conc.) and washed with water saturated with
NH4Cl (3 Â 10 ml). Combined aqueous phase was backwashed with
diethyl ether (3 Â 30 ml). Ether was added to THF fractions, and the
combined organic phase was dried (MgSO4) and rotary-evaporated.
Chromatography of the residue on silica gel column eluted with 1:1
1
ethyl acetate:hexanes yielded 0.23 g (15%) of 5 as a colorless oil. H
NMR (CDCl3) d 7.35(m, 5H, -C 6H5), 5.83 (m, 1 H, -CH=), 5.31-5.08
(m, 4 H, CH2=, -CH2Ph), 4.37 (s, 1 H, -CH(NHZ)-), 4.70 (m, 2 H,
HOCH2-), 2.25(bs, 1 H, -N HZ), 1.93 (m, 1 H, -CH2CH2OH), 1.55 (m,
1 H, -CH2CH2OH). MS m/z 235(M +),190 (M+ -CH2CH2OH), 91
(PhCH2).
5-Bromo-3-(carbobenzyloxy)amino-1-pentene (6)
A solution of triphenylphosphine (0.885g, 3.37 mmol) in 10 ml of
CH2Cl2 was added during 200 min to a stirred solution of 5 (0.159 g,
0.675mmol) and carbon tetrabromide (1.12 g, 3.37 mmol) in CH 2Cl2
(20 ml) cooled in ice–water bath, and the mixture was further stirred at
08C for 1 h. The solvent was rotary-evaporated, and the residue was
purified using silica gel column with 30% ethyl acetate–hexane as an
eluent. Yield was 0.179 g (89%), with 6 obtained as white crystalline
1
solid. Mp 40–428C. H NMR (CDCl3) d 7.35(m, 5H, -C 6H5), 5.76 (m,
1 H, -CH=), 5.26-5.12 (m, 4 H, CH2=, -CH2Ph), 4.75(bs, 1 H, -N H-),
4.36 (m, 1 H, -CH(NHZ)-), 3.41 (m, 2 H, HOCH2-), 2.10 (s, 2 H,
-CH2CH2OH). 1H NMR (DMSO) d 7.33 (m, 5H, -C 6H5), 5.74 (m, 1 H,
-CH=), 5.15-5.03 (m, 4 H, -CH2=, -CH2Ph), 4.15(s, 1 H, -C HNH-),
3.47 (m, 2 H, HOCH2-), 1.97 (m, 2 H). MS m/z 299 (M+, 81Br), 297
Copyright # 2002 John Wiley & Sons, Ltd.
J Label Compd Radiopharm 2002; 45: 199–211

208
Z. ZHANG ET AL.
(M+, 79Br), 91 (C7H7+). Anal. Calcd for C13H16BrNO2: C, 52.37; H,
5.41; N, 4.70. Found: C, 52.58; H, 5.73; N, 4.48.
4-(Carbobenzyloxy)amino-5-hexenenitrile (7)
Synthesis: Compound 6 (30 mg, 0.10 mmol) was added to a solution of
LiCN (66.6 mg, 2.0 mmol) in THF (5ml). The resulting mixture was
refluxed for 2 h and the solvent was rotary- evaporated. The residue was
chromatographically purified on a silica gel column using 30–40% ethyl
acetate in hexane as eluent to afford 7 as a white solid (14.4 mg, 70%
1
yield). H NMR (CDCl3, 400 MHz) d 7.35(m, 5H, -C 6H5), 5.78 (m,
1 H, -CH=), 5.36-5.21 (m, 2 H, -CH2=), 5.12 (s, 2 H, -CH2Ph), 4.74 (bs,
1 H, -NH-), 4.27 (m, 1 H, -CH(NHZ)-), 1.96 (m, 2 H, NCCH2-), 1.29 (m,
1 H, NCCH2CH2-), 0.95(m, 1 H, NCCH 2CH2-). MS m/z 244 (M+), 190
(M+-CH3CH2CN), 91 (C7H+7 ).
Finding optimal solvent for synthesis of 7
Compound 6 (2 mg, 7 mmol) and LiCN (0.33 mg, 10 mmol) were
dissolved in 0.25ml of the studied solvent and heated DMF, 105 8C;
DMSO, 908C; THF and MeOH, 808C. The samples (2 ml) were analyzed
at 0, 10 and 25min using HPLC: Phenomenex, Ultremex-10 (C18)
column, 10 mm, 250 Â 4.6 mm eluted with 10 mM KH2PO4 pH 2.6 –
55% CH3CN at 2 ml/min; tR (min): 6, 3.9; 7, 2.4. The progress of the
reaction was estimated based on changes in the areas of the peaks for
the starting material (6), product (7) and byproducts (tR of the
byproducts formed in DMSO are 2.2 and 4.1 min).
GVG (4-amino-5-hexenoic acid, 8)
A solution of 7 (5mg) in 0.8 ml of 38% DCl-D 2O was heated for 20 min
at 1008C (in 5mm diameter NMR tube) to give GVG as the only
product. No work-up was done. The NMR spectrum of product was
1
identical to that of authentic GVG. H NMR (D2O) d 5.90 (m, 1 H,
HC=), 5.54 (s, H, CH2=), 5.50 (d, J=5.8 Hz, 1 H, CH2=), 3.96 (m,
1
1 H, -CH(NH2)-), 2.56 (m, 2 H, -CH2COOH), 2.17 (m, 1 H,
-CH2CH2COOH), 2.04 (m, 1 H, -CH2CH2COOH).
Copyright # 2002 John Wiley & Sons, Ltd.
J Label Compd Radiopharm 2002; 45: 199–211

NOVEL SYNTHESIS OF
209
[1-11C]GVG ([11C]8)
[11C]Cyanide (150–450 mCi at the EOB) from home-made synthesizer
carried by nitrogen was passed through a Sicapent1 column (50 Â 5mm
column), to reduce the amount of ammonia, and trapped in an ethanol–
ice cooled vessel (À58C) containing dry K2CO3 (3.5mg, 25 mmol) and
Kryptofix 222 (4,7,13,16,21,24-Hexaoxa-1,10-diazabicyclo[8.8.8]hexa-
cosane, 12 mg, 32 mmol). [11C]Cyanide which was not trapped in the
reaction vessel was quantitatively retained on a column (70 Â 5mm)
filled with slime. The radioactivity trapping efficiency was therefore
calculated as percentage of the [11C]cyanide retained in the reaction
vessel versus the total amount of the [11C]cyanide delivered to the system.
A dry-ice cooled solution of 6 (3 mg, 10 mmol) in 0.2 ml of DMF was
added, and the vessel was heated at 1258C for 10 min. After short
cooling (cold water), aqueous HCl (0.5ml, 12 M) was introduced, and
the hydrolysis of [1-11C]7 was done at 1258C for 10 min. The solution
was rapidly cooled (cold water) and neutralized by addition of
(NH4)2HPO4 (400 mg in 0.8 ml of water). [11C]GVG was purified by
HPLC on Phenomenex, Lichrosorb RP-18, 10 mm, 250 Â 10 mm column
with eluent 10 mM KH2PO4 pH 2.6 at 2 ml/min; tR (retention time)
12.3 min. The solvent from the collected fraction was removed by rotary
evaporation, and the residue was dried by azeotroping with CH3CN
(2 Â 1 ml). The product was redissolved in saline and filtered through a
0.22 mm Millipore filter into a sterile vial. Radiochemical purity was
determined by both HPLC and TLC. HPLC was equipped with radio-
and UV (235nm) detectors. HPLC: Phenomenex, Ultremex-10 (C18),
10 mm, 250 Â 4.6 mm column with gradient system 0–7 min, 10 mM
KH2PO4 pH 2.6 at 1 ml/min and 7–15min, 10 mM KH 2PO4 pH 2.6 – 60%
CH3CN at 2 ml/min, tR 5.7 min. TLC analysis was performed
using 10 mM KH2PO4 pH 2.6 : CH3CN 1 : 1 as eluent, Rf of GVG
was 0.68.
Specific radioactivity of [11C]GVG was estimated by the injection of
0.2 ml of the final radiotracer solution on the HPLC. The same HPLC
conditions as for [11C]GVG quality control were used except an
isocratic elution was performed with 10 mM KH2PO4 pH 2.6 at 1 ml/
min. A calibration curve was obtained by the injection of 0.2 ml
solutions containing 1–10 mg (7.8–78 nmol) of authentic GVG in
saline (0.9% NaCl in water). The EOB-corrected specific radio-
activity based on the samples from two radiosyntheses exceeded
0.5Ci/ mmol.
Copyright # 2002 John Wiley & Sons, Ltd.
J Label Compd Radiopharm 2002; 45: 199–211

210
Z. ZHANG ET AL.
Conclusion
GVG, a suicide inhibitor of GABA transaminase (GABA-T), was
efficiently labeled with carbon-11 via a [11C]cyanation of 5-bromo-3-
(carbobenzyloxy)amino-1-pentene followed by acid hydrolysis. These
synthetic procedures should be applicable to the labeling of the
pharmacologically active S form of GVG when the S-enantiomer of
homoserine lactone is used. The availability of this radiotracer should
allow us to study the pharmacokinetics as well as the potential side
effects of the drug. It also sets the stage for future studies in human to
better understand its role in addiction treatment.
Acknowledgements
This research was carried out at Brookhaven National Laboratory
under contract DE-AC02-98CH10886 with the U.S. Department of
Energy and Office of Biological Environmental Research and also
supported by the National Institutes of Health (National Institute of
Neurological Diseases and Stroke NS-15380). The authors are grateful
for valuable discussions with S. Dewey and J. Fowler, and the assistance
of D. Schlyer, V. Garza, R. MacGregor, C. Shea, R. Carciello.
References
1. Jung MJ, Lippert BW, Metcalf BW, Bohlen P, Schechter PJ. J Neurochem
1977; 29: 797–802.
2. Lippert B, Metcalf BW, Jung MJ, Casara P. Eur J Biochem 1977; 74:
441–445.
3. Weber OM, Verhagen A, Duc CO, Meier D, Leenders KL, Boesiger P.
Magn Reson Imaging 1999; 17: 417–425.
4. Sabers A, Gram L. Pharm Toxicol 1992; 70: 237–243.
5. Grant SM, Heel RC. Drugs 1991; 41: 889–926.
6. Dewey SL, Morgan AE, Ashby CRJ, Horan B, Kushner SA, Logan J,
Volkow ND, Fowler JS, Gardner EL, Brodie JD. Synapse 1998; 30:
119–129.
7. Dewey SL, Brodie JD, Gerasimov M, Horan B, Gardner EL, Ashby CRJ.
Synapse 1999; 31: 76–86.
8. Gerasimov MR, Ashby CRJ, Gardner EL, Mills MJ, Brodie JD, Dewey
SL. Synapse 1999; 34: 11–19.
9. Gerasimov MR, Schiffer WK, Brodie JD, Lennon IC, Taylor SJ, Dewey
SL. Eur J Pharmacol 2000; 395: 129–135.
Copyright # 2002 John Wiley & Sons, Ltd.
J Label Compd Radiopharm 2002; 45: 199–211

NOVEL SYNTHESIS OF
211
10. Eke T, Talbot JF, Lawden MC. BMJ 1997; 314: 180–181.
11. Schuster AJ, Wagner ER. J Labelled Cpd Radiopharm 1993; 33: 213–217.
12. Gill HS. J Labelled Cpd Radiopharm 1995; 36: 425–430.
13. Wei ZY, Knaus EE. Tetrahedron 1994; 50: 5569–5578.
14. Kwon TW, Keusenkothen PF, Smith MB. J Org Chem 1992; 57:
6169–6173.
15. Wei Z-Y, Knaus EE. J Org Chem 1993; 58: 1586–1588.
16. Kihlberg T, Gullberg P, Langstron B. J Labelled Cpd Radiopharm 1990;
28: 1115–1119.
17. Barrett AGM, Broughton HB. J Org Chem 1986; 51: 495–503.
18. Yoon NM, Gyoung YS. J Org Chem 1985; 50: 2443–2450.
19. Greenwald R, Chaykovsky M, Corey E. J Org Chem 1963; 28: 1128–1129.
20. Corey EJ, Weinshenker NM, Schaaf TK, Huber W. J Am Chem Soc 1969;
91: 5675–5677.
21. Reitz AB, Nortey SO, Maryanoff BE. J Org Chem 1987; 52: 4191–4202.
22. Pike RM, Mayo DW, Butcher DJ, Hinkle RJ. J Chem Ed 1986; 63:
917–918.
23. Greene TW, Wuts PGM. Protective Groups in Organic Synthesis, Wiley:
New York, 1991.
24. Two syntheses of fully protected homoserine lactones were made at the
different stages of this research. N-(2,5,5-Tetramethyl-2,5-disila-1-azacy-
clopentyl)-homoserine lactone was found to be moisture sensitive and
decomposed during purification, while 2-[1,3-Dimethyl-1,3-diaza-2-oxo-5-
azacyclohexanyl)-4-butyrolactone was synthesized in rather low yield.
25. Grashey R. Comprehensive Organic Synthesis, Vol 6. Pergamon Press: New
York, 1991.
26. Ding Y-S, Fowler JS, Wolf AP. J Labelled Cpd Radiopharm 1993; 33:
645–654.
27. Christman DR, Finn RD, Karlstrom KI, P. WA. Int J Appl Radiat Isot
1975; 26: 435–442.
28. Iwata R, Ido T, Takahashi T, Nakanishi H, Iida S. Appl Radiat Isot 1986;
38: 97–102.
29. Ding Y-S, Antoni G, Fowler JS, Wolf AP, Langstrom B. J Labelled Cpd
Radiopharm 1989; 27: 1079–1090.
30. Ponchant M, Hinnen F, Demphel S, Crouzel C. Appl Radiat Isot 1997; 48:
755–762.
31. Westerberg G, Langstrom B. Appl Radiat Isot 1996; 48: 459–461.
32. Ponchant M, Galea H, Bottlaender M, Coulon C, Fuseau C, Ottaviani M,
Crouzel C. J Labelled Cpd Radiopharm 2000; 43: 1311–1320.
33. Brown HC. Organic syntheses via boranes, Wiley: NY, 1975.
34. Studenov AR, Wegner AM, Ding YS. J Labelled Cpd Radiopharm 2001;
44: 1–12.
Copyright # 2002 John Wiley & Sons, Ltd.
J Label Compd Radiopharm 2002; 45: 199–211