Synthesis and liposome encapsulation of a novel 18F-conjugate of ω-conotoxin GVIA for the potential imaging of N-type Ca2+ channels in the brain by positron emission tomography

Article abstract of DOI:10.1002/jlcr.1029

ω-Conotoxin GVIA is a potent, irreversible antagonist of N-type voltage gated Ca²⁺ channels. A radiofluorinated analogue of GVIA could be useful in assessing regional synaptic density of the brain, in vivo, using positron emission tomography. N

Full text of DOI:10.1002/jlcr.1029

JOURNAL OF LABELLED COMPOUNDS AND RADIOPHARMACEUTICALS
J Label Compd Radiopharm 2006; 49: 269–283.
Published online in Wiley InterScience (www.interscience.wiley.com). DOI: 10.1002/jlcr.1029
Synthesis and liposome encapsulation of a novel
F-conjugate of x-conotoxin GVIA for the potential
1
8
2+
imaging of N-type Ca channels in the brain by
positron emission tomography
y
y
Vahe Azarian , Anne Gangloff , Yann Seimbille, Sibylle Delaloye, Johannes
Czernin, Michael E. Phelps and Daniel H.S. Silverman*
Department of Molecular and Medical Pharmacology and Ahmanson Biological
Imaging Division, David Geffen School of Medicine, University of California,
Los Angeles, USA
Summary
2
+
o-Conotoxin GVIA is a potent, irreversible antagonist of N-type voltage gated Ca
channels. A radiofluorinated analogue of GVIA could be useful in assessing regional
synaptic density of the brain, in vivo, using positron emission tomography. N-hydroxy
1
8
succinimidyl 4-[ F]fluorobenzoate was employed to site-specifically label GVIA,
preserving native binding affinity. The tracer was characterized with MALDI-TOF
mass spectrometry and colorimetric protein assay. Radiochemical decay-corrected
1
8
yield of the lysine-24 labeled analogue of [ F]GVIA was 5%. Specific activity of this
5
species was determined to be 1.2 ꢀ 10 Ci/mmol. Encapsulation of the tracer in
sulfatide containing liposomes, a potential method for enhancing blood–brain
penetrance, was accomplished with 40% efficiency. Copyright # 2006 John Wiley
&
Sons, Ltd.
Key Words: fluorine-18; o-conotoxin GVIA; peptide labeling; PET brain imaging;
liposome
Introduction
2
N-type Ca
+
channels are key members of a family of voltage-gated
perisynaptic ion channel found throughout the nervous system. The current
mediated by voltage-gated calcium channels (VGCC) initiates a cascade of
protein–protein interactions, which ultimately result in exocytosis of vesicular
transmitter during neurotransmission. Structurally, N-type channels are
*Correspondence to: D. H. S. Silverman, Nuclear Medicine, CHS AR-144, MC694215, Los Angeles, CA,
9
0095 6942 USA. E-mail: dsilver@ucla.edu
y
Contributed equally to this project.
Contract/grant sponsors: National Institute of Mental Health and NIH; contract/grant number: MH64679
Copyright # 2006 John Wiley & Sons, Ltd.
Received 30 September 2005
Revised 26 October 2005
Accepted 28 October 2005

2
70
V. AZARIAN ET AL.
pentameric transmembrane proteins. The large, pore-forming a1 subunit is
B
1
–4
the target of o-conotoxins GVIA and MVIIA.
N-type Ca
Quantification of
2
+
channels by positron emission tomography (PET), using
radiofluorinated bioconjugates of o-conotoxins, could serve as an accurate
marker for regional synaptic density of the brain in vivo. Furthermore,
investigations regarding variations in regional synaptic density stemming from
age, memory/learning, or neurodegenerative disease may be possible using
such tracers. To this end, we describe here the radiochemical synthesis of a
1
8
biologically active conjugate of o-conotoxin GVIA ([ F]GVIA) with yields
1
8
and labeling conditions suitable for use of the F-labeled product in acquiring
PET data.
The best characterized o-conotoxin, GVIA, is an approximately 3kD,
7-residue peptide isolated from the venom of the marine cone snail, Conus
2
geographus. Containing three disulfide bonds (1–16, 8–19, and 15–26 linkage)
as well as three hydoxyproline residues, the secondary structure of GVIA is
2
constrained and resistant to denaturation under non-reducing conditions.
,4
Site-specific acetylation studies have shown that residues near the N-terminus,
such as Cys-1 and Lys-2, are critical for GVIA’s binding affinity and specificity
for N-type VGCC, while residues further along, such as Lys-24, can be
1
chemically modified with little to no effect on binding activity. Therefore, site-
specific labeling strategies are important for preserving biological activity of
these compounds.
This study compares the efficacy of two well-known approaches for site-
specific fluorination in terms of their application to small hydrophilic peptides
such as GVIA. The first pathway directly utilizes fluorobenzoic acid (FBA), in
presence of diethyl cyanophosphonate and diisopropyl ethylamine, to label
GVIA under anhydrous reaction conditions. The second strategy relies on the
compound N-succinimidyl 4-fluorobenzoate (SFB), which is more difficult and
time consuming to prepare than FBA, but has the advantage of being stable in
aqueous media, facilitating labeling. Both strategies begin from an identical
triflate precursor ð1Þ, which after fluorination and acid-mediated hydrolysis
yield FBA.
SFB is an effective labeling agent for targeting the a- and e-NH groups
2
associated with the N-termini and Lys residues, respectively, of proteins under
mildly basic conditions (pH 8.6). SFB labels free amino groups via a trans-
1
8
amidification reaction resulting in the covalent addition of a [ F]fluoroben-
zoyl moiety. Furthermore, SFB shows virtually no reactivity towards
guanidino or alcohol moieties, and is stable and reactive in aqueous media,
unlike FBA. Thus, to develop a synthetic scheme for a biologically active
radiofluorinated analogue of GVIA (denoted as K24-FGVIA) we relied on the
chemical reactivity and specificity of SFB to target lysine residues.
Subsequently, the desired Lys-24 labeled radioanologue was separated from
Copyright # 2006 John Wiley & Sons, Ltd.
J Label Compd Radiopharm 2006; 49: 269–283

1
8
SYNTHESIS AND ENCAPSULATION OF [ F] GVIA
271
byproducts and reaction components by reverse-phase HPLC (RP-HPLC) to
ensure radiochemical purity of the tracer.
Finally, in developing a peptide based tracer, the blood–brain barrier poses
a significant obstacle. To enhance the pharmacokinetic and biodistribution
18
properties of K24-[ F]GVIA, we have adapted a protocol for encapsulation
of the radiotracer into large, lipid-based vesicles known as liposomes. It has
been previously demonstrated that liposomal incorporation of brain specific
5
,6
fatty acid esters can increase penetrance of peptide ligands. In the adult
mammalian brain, sulfatide (cerebroside 3-sulfate) is mainly found in myelin
sheaths, and accounts for approximately 3.8% of total lipid weight. We have
utilized a variety of liposome, containing 15% (mole fraction) sulfatide, in an
5
attempt to maximize cerebral uptake.
Imaging regional synaptic density, through the use of an N-type
VGCC antagonist and PET, may bring important insights into the dynamic
macroscopic organization of the brain. However, to use a peptide tracer,
such as K24-FGVIA, three issues must be considered: (1) development
of a labeling strategy consistent with site-specific requirements of the protein;
(2) purification of biologically active radioanalogues from other isoforms
via RP-HPLC; (3) increasing blood–brain penetrance of the peptide tracer
to yield viable PET data. This study presents the development of
such an approach, based on the native binding properties of o-conotoxin
GVIA.
Results and discussion
This study compared two alternate routes of synthesis for a lysine-24
fluoroconjugate of GVIA, from a common triflate precursor, with yields and
preparative conditions suitable for neuro-PET studies.
Pentamethyl 4-(trimethylammonium trifluoromethanesulfonate) benzoate
7–10
ð1Þ was synthesized in two steps.
The coupling of pentamethylbenzyl
chloride to 4-(N,N-dimethylamino) benzoic acid, in the presence of pyridine,
to form pentamethylbenzyl 4-(N,N-dimethylamino) benzoate, had a yield of
6
3%. Next, methylation of the dimethylamino group to form 1 was 37%
efficient (Figure 1(a)). Similarly, the esterification of 4-fluorobenzoic acid with
disuccinimidyl carbonate to form N-succinimidyl 4-fluorobenzoate yielded
7
6% of the final purified product (Figure 1(b)).
19
Site-specific conjugation of GVIA with [ F] SFB proceeded to completion
in 30 min. The HPLC elution profile of the reaction mixture showed four
major components with native GVIA having a molecular weight of 3037, and
conjugated isoforms of GVIA exhibiting an increase in mass of 122, per added
fluorobenzoyl group. Thus, analysis of the four fractions via MALDI-TOF
MS revealed the first peak to be native GVIA (M þ m=z ¼ 3036:76), while the
next two signals were isomeric mono-conjugates (M þ m=z ¼ 3158:9).
Copyright # 2006 John Wiley & Sons, Ltd.
J Label Compd Radiopharm 2006; 49: 269–283

2
72
V. AZARIAN ET AL.
O
O
O
OH
Cl
Et N, DMF
O
3
+
N
N
MeOTf, CH Cl
2
2
-
O
OTf
+
N
1
O
O
1
8 -
F , K CO /K
2
3
222
O
TFA, Anisole
1
OH
18F
18F
3
F18
NH2
GVIA−Lys , iPr NEt
x
2
O
O
O
NH
O P
O
CN
GVIA−Lysx
F18
OH
18F
NH2
GVIA−Lysx
O
O
3
O N
O
DSC, DMAP
NH
1
8
F
O
4
GVIA−Lysx
Figure 1. Two step preparation of radiochemical precursor, pentamethyl
-(trimethylammonium trifluoromethanesulfonate) benzoate (PMBTf), and
4
subsequent radiofluorination of x-conotoxin GVIA via aqueous and anhydrous
pathways
The final peak was di-conjugated GVIA (M þ m=z ¼ 3280:9), presumably
labeled at each of the two lysine sites (Figure 3).
Similar experiments with FBA showed virtually no conjugation of
GVIA after 2 h. The basis for such complete reaction failure was the chemical
incompatibility of GVIA with the labeling reagents. FBA was only
slightly soluble in water (0.4 g/100 ml), while diethyl cyanophosphonate and
diisopropyl ethylamine were insoluble. Conversely, GVIA was unstable
1
and insoluble in aqueous buffers containing more than 40% (v/v) acetonitrile.
Therefore, the lack of a mutually compatible medium for GVIA and the
anhydrous labeling agents prevented the FBA-mediated direct conjugation
of GVIA.
1
8
Prior to radiochemical development of an [ F]GVIA conjugate, the target
site of each of the mono-labeled species must be determined to ensure that at
Copyright # 2006 John Wiley & Sons, Ltd.
J Label Compd Radiopharm 2006; 49: 269–283

1
8
SYNTHESIS AND ENCAPSULATION OF [ F] GVIA
273
least one isomer is predicted to be bio-active. To this end we employed a
trypsin-digest protocol to verify which HPLC fraction corresponded to the
1
Lys-2 and Lys-24 isomer, respectively. The procedure was based on the native
proteolytic activity of trypsin for peptides containing basic residues such as
lysine and arginine. Trypsin cleaves peptides by hydrolysis on the amino side
of a basic residue. Figure 2 is a schematic representation of trypsin’s predicted
activity on native, mono- and di-conjugated GVIA. Depending on which
lysine residue(s) is labeled, thus interrupting tryptic cleavage at that site, the
corresponding GVIA fragment will have a unique molecular weight. There-
fore, by digesting each of the labeled species of GVIA with trypsin, purifying
the major fragment via HPLC, and using MALDI-TOF MS to determine
molecular weight, it was possible to ascertain the labeling site(s) of the
conjugates. The predicted fragment mass of each GVIA species was as follows:
native GVIA (2935), K2-FGVIA (3040), K24-FGVIA (3214), K2+24-
F GVIA (3299). Experimentally determined M+ fragment masses for each
2
eluting fraction correlates well with predicted values, and are were as follows:
1
2.5 min [m=z ¼ 2934:6], 13.5 min [m=z ¼ 3039:6], 18.5 min [m=z ¼ 3213:1],
and 20.1 min [m=z ¼ 3297:6]. Thus, native GVIA elutes at 12.5 min,
Native GVIA: Fragment Mass = 2935
C–K–S–O–G–S–S–C–S–O–T–S–Y–N–C–C–R–S–C–N–O–Y–T–K–R–C–Y–NH2
K2-FGVIA: Fragment Mass = 3040
1
8
F
C–K–S–O–G–S–S–C–S–O–T–S–Y–N–C–C–R–S–C–N–O–Y–T–K–R–C–Y–NH2
K24-FGVIA: Fragment Mass = 3214
1
8
F
C–K–S–O–G–S–S–C–S–O–T–S–Y–N–C–C–R–S–C–N–O–Y–T–K–R–C–Y–NH2
K2+K24-F2GVIA: Fragment Mass = 3299
1
8
F
1
8
F
C–K–S–O–G–S–S–C–S–O–T–S–Y–N–C–C–R–S–C–N–O–Y–T–K–R–C–Y–NH2
Figure 2. Graphic representation of expected fragment masses from trypsin
mediated digestion of each isomer of GVIA. Solid vertical bars illustrate trypsin
cleavage sites, while labeling site(s) are depicted by square boxes inscribed
1
8
with ‘ F.’
Copyright # 2006 John Wiley & Sons, Ltd.
J Label Compd Radiopharm 2006; 49: 269–283

2
74
V. AZARIAN ET AL.
Table 1. Results of HPLC and MALDI-TOF MS analyses of GVIA cold labeling
reaction with SFB
Species
Retention
time (min)
Predicted
mass
Experimental
mass (M+)
Predicted
fragment
mass
Experimental
fragment
mass (M+)
GVIA
12.8
13.5
18.5
20.1
3037
3159
3159
3281
3036.8
3158.9
3158.9
3280.1
2935
3040
3214
3299
2934.6
3039.9
3213.1
3297.6
K2-FGVIA
K24-FGVIA
K2+24-F GVIA
2
K2+K24-F GVIA
2
K2-FGVIA
Mass:3159
600
Mass: 3159
tR = 20.1 min
tR = 13.5 min
K24-FGVIA
Mass:3159
tR = 18.5 min
4
2
00
00
Native GVIA
Mass = 3037
tR = 12.8 min
0
−200
0
5
10
Time (min)
15
20
Figure 3. HPLC elution profile of GVIA cold-labeling reaction with SFB. The
four major components of the reaction mixture are depicted, along with the
identity, mass and retention time of each peak
K2-FGVIA at 13.5 min, K24-FGVIA at 18.5 min, and K2+24-F GVIA at
2
2
0.1 min (Table 1).
In order to determine the specific activity of K24-[ F]GVIA, an association
1
8
between absorbance at 215 nm (as measured by an online HPLC spectro-
photometer) and molar quantity was made. A method commonly used for
Copyright # 2006 John Wiley & Sons, Ltd.
J Label Compd Radiopharm 2006; 49: 269–283

1
8
SYNTHESIS AND ENCAPSULATION OF [ F] GVIA
275
1
1
quantifying collagen was adapted for this purpose. The assay relies on a
chemical reaction between hydroxyproline, Chloramine-T and p-dimethyla-
minobenzaldehyde to produce a chromogenic compound, whose absorbance is
measured at 550 nm. The reaction mixture changes from a translucent yellow
to vibrant peach with increasing concentrations of Hyp. Application of native
GVIA to the assay yielded a standard curve. Finally, an unknown quantity of
K24-FGVIA was assayed and the result divided by its absorption area
(mAU ꢀ Sec) at 215 nm to give the following linear equation (at mM
concentrations):
K24-FGVIA ðnmolÞ ¼ ½3:14 ꢀ 10ꢁ ꢂ6ꢀ10 ꢃꢀ½Absorption area at 215 nmꢃ
4
ꢁ6
1
8
It was possible to measure the specific activity of K24-[ F]GVIA by collecting
the appropriate HPLC fraction, measuring its activity, and dividing by the
mass quantity of labeled peptide as determined by the equation above. Thus,
18
5
we determined the specific activity of K24- FGVIA to be 1.2 Ci ꢀ 10 /mmol.
This set the stage for the radiochemical preparation and liposome
18
encapsulation of [ F]GVIA. The aqueous pathway required four radio-
1
8
chemical and two HPLC steps to yield K24-[ F]GVIA with a total
preparation time of 165 min (Figure 1(b)). Radiofluorination of 1 and C18
solid phase extraction was 60% efficient. Hydrolysis of the fluorinated ester 2
1
8
to form [ F]FBA ð3Þ, and subsequent coupling to DSC/DMAP yielded
1
8
>
70% [ F]SFB ð4Þ, as determined by HPLC. The fractions corresponding to
and 4 eluted at 12 and 22 min, respectively (Figure 4). The purified
3
and reconstituted 4 was 90% consumed in the labeling reaction with
native GVIA after 30 min. The radioconjugation produced approximately
1
8
18
18
3
0% K2-[ F]GVIA, 40% K24-[ F]GVIA, and 10% K2+24-[ F]GVIA
Figure 5). The remaining 20% of activity was either hydrolyzed into 3 or
(
1
8
remained unused as 4. The fraction corresponding to K24-[ F]GVIA was
isolated and prepared for liposome encapsulation. The procedure required
2
5 min and was 45% efficient, as determined by size-exclusion chromato-
18
graphy. Non-encapsulated [ F]GVIA was retained by Sephadex-G60 gel well
after the encapsulated species eluted, offering virtually complete separation
(Figure 6).
The microscale fluoroconjugation of GVIA with SFB and FBA, respec-
tively, was predictive of the radiochemical behavior of each molecule. In
examining labeling agent efficacy, and thus choosing the optimal strategy, we
had to weigh ease of radiochemical preparation against labeling efficiency and
site-specificity. Although the anhydrous labeling method has been shown to be
an effective and amine-specific labeling procedure, the insolubility of GVIA in
organic media prevented its use in this particular application. On the other
hand, SFB is soluble, specific and reactive enough in aqueous buffer to
consistently produce two mono-conjugated isomers of GVIA. While the
Copyright # 2006 John Wiley & Sons, Ltd.
J Label Compd Radiopharm 2006; 49: 269–283

2
76
V. AZARIAN ET AL.
400
300
200
100
0
0
5
10
15
20
25
30
Time (min)
1
Figure 4. HPLC elution profile of the radiochemical synthesis of F-SFB from
8
1
8
F-FBA. Depicted are traces from the high sensitivity c detector (red), low
sensitivity c detector (blue), and UV absorption (green) at 235.6 nm
aqueous pathway will require longer radiochemical preparation than the direct
anhydrous approach, the lack of reactivity exhibited by FBA makes its use
impractical. Thus, we have chosen to develop the aqueous radiochemical
strategy to site-specifically label GVIA.
Quantification of regional synaptic density by PET may offer fresh insight to
the pathology of neurodegenerative disease, neurodynamics of aging, and the
macroscopic plasticity involved in learning/memory. This study presents the
radiochemical synthesis of a biologically active radiofluorinated analogue of
o-conotoxin GVIA, a natural antagonist of N-type channel. The potency of
GVIA to bind to channels, and inhibit neurotransmission, is already being
exploited by the pharmaceutical industry for analgesia. We describe here a
four-step aqueous labeling strategy and one-step encapsulation protocol which
provides practical means for preparing a radiofluorinated GVIA analogue
potentially useful for exploring distribution of N-type calcium channels
with PET.
Copyright # 2006 John Wiley & Sons, Ltd.
J Label Compd Radiopharm 2006; 49: 269–283

1
8
SYNTHESIS AND ENCAPSULATION OF [ F] GVIA
277
5
4
00
00
3
2
1
00
00
00
0
0
5
10
15
20
25
30
Time (min)
Figure 5. HPLC elution profile of the site-specific radiochemical labeling of GVIA
18
with F-SFB. UV (green) and c-ray (red) traces are displayed simultaneously.
Corresponding gamma peaks appear directly below their respective UV peaks
Experimental
Materials
o-Conotoxin GVIA was purchased from Peptides International, Inc., Louisville,
KY. Trypsin (Gold) mass spectrometry grade was acquired from Promega Co.,
Madison, WI. HPLC was performed using an Agilent 1100 series apparatus with
an online degasser, UV/Vis spectrophotometer, and gamma detector from
Agilent Technologies, Inc. Palo Alto, CA. HPLC eluates were monitored for
their UV absorbance and radioactive content by connecting the outlet of the UV
photometer to a NaI detector. The recorded data was processed by Agilent
ChemStation software. All other chemicals were purchased from Sigma-Aldrich
Co. or Fisher Scientific International without further purification.
Synthesis of pentamethyl 4-(trimethylammonium trifluoromethanesulfonate)
benzoate (1)
A solution of 4-(N,N-dimethylamino) benzoic acid (570 mg, 3.4 mmol),
pentamethylbenzoyl chloride (650 mg, 3.3 mmol), and triethylamine (0.46 ml,
Copyright # 2006 John Wiley & Sons, Ltd.
J Label Compd Radiopharm 2006; 49: 269–283

2
78
V. AZARIAN ET AL.
4
5000
0000
5000
0000
5000
0000
5000
0000
4
3
3
2
2
1
1
Liposome Encapuslated
18
[
F]GVIA
Free [18F]GVIA
5000
0
1
3
5
7
9
11
13
15
17
19
21
Volume (mL)
Figure 6. Gel filtration elution profile of liposome encapsulated and non-
1
encapsulated F-GVIA
8
3
.3 mmol) in N,N-dimethylformamide (15 ml) was stirred overnight at room
temperature, and subsequently poured into a solution of 10% NaHCO3
80 ml). The resulting white precipitate was filtered off and dried under
(
vacuum. The crude product was then recrystallized from ethyl acetate to
1
afford pentamethyl 4-(N,N-dimethylamino)benzoate (680 mg, 63%): H-NMR
(
(
CDCl ) d 2.2–2.4 (3s, 15H), 3.01 (s, 6H), 5.42 (s, 2H), 6.60 (d, 2H), 7.90
d, 2H). Pentamethyl 4-(N,N-dimethylamino)benzoate (255 mg, 0.75 mmol)
3
was then dissolved in anhydrous methylene chloride (5 ml). Trifluorometha-
nesulfonate (90 ml, 0.8 mmol) was added and the reaction was stirred
overnight. 10–20% hexanes by volume is added to the reaction mixture and
the resulting solution was set into a ꢁ208C freezer overnight. Compound 1 was
collected by filtration as a fine, white needle-like crystal (140 mg, 37%):
1
H-NMR (CDCl ) d 2.1–2.3 (m, 15H), 2.75 (s, 3H), 3.8 (s, 9H), 5.5 (s, 2H),
3
7
.8–8.2 (m, 3H).
Synthesis of succinimidyl 4-fluorobenzoate (SFB)
A solution of 4-fluorobenzoic acid (140 mg, 1 mmol), disuccinimidyl carbonate
(256 mg, 1 mmol), and pyridine (81 ml, 1 mmol) in anhydrous acetonitrile (1 ml)
was stirred for 90 min at room temperature. The crude product was
Copyright # 2006 John Wiley & Sons, Ltd.
J Label Compd Radiopharm 2006; 49: 269–283

1
8
SYNTHESIS AND ENCAPSULATION OF [ F] GVIA
279
evaporated and applied to a silica gel column and eluted with 50:50 ethyl
acetate:hexanes. Fractions containing SFB, as assessed by thin-layer
chromatography, were pooled and evaporated under reduced pressure to
1
yield white crystals (181 mg, 76%): H-NMR (CDCl ) d 2.91 (s, 4H), 7.19
3
(m, 2H), 8.18 (m, 2H).
Conjugation of o-conotoxin GVIA via succinimidyl 4-fluorobenzoate
To a 50 ml solution of GVIA (5 nmol) in borate buffer (50 mM, pH 8.3), 100 ml
1
2,13
of deionized water and 5 ml of SFB (5 nmol in CH CN) were added.
The
reaction was allowed to sit for 30 min, after which 800 ml of 0.1 M NaCl, pH
.4, was added. RP-HPLC was used to isolate and purify conjugated species.
The reaction mixture was injected onto an Agilent Zorbax Rx-C8 Column
4.6 mm, 150 mm ꢀ 5 mm), and eluted at 1 ml/min under a gradient ranging
3
2
(
from 99:1 to 21:79 0.1 M NaCl pH 2.4: acetonitrile in 50 min. Eluates were
visualized at 215 nm and fractions corresponding to the different conjugated
species were collected for further analysis.
Conjugation of o-conotoxin GVIA via 4-fluorobenzoic acid
To 100 ml of GVIA solution (10 nmol) in deionized water, FBA (1.4 mg,
1
0 nmol) solution in 110 ml of acetonitrile, cyanophosphonate (1 mg, 6.1 mmol)
solution in 50 ml acetonitrile, and 10 ml iPr NEt were added. The reaction was
2
allowed to sit for 30 min, after which 800 ml of 0.1 M NaCl (pH 2.4) was added.
The reaction mixture was then injected onto HPLC, using the conditions
described above.
Identification of conjugated species and determination of labeling site(s)
The HPLC fractions were lyophilized and reconstituted with 50 ml borate
buffer (50 mM, pH 8.3). Each solution was then analyzed by MALDI-TOF
mass spectrometry, using a-cyano cinnamic acid as matrix, to determine the
presence of native, mono- or poly-conjugated species of GVIA. Those
fractions corresponding to an isoform of GVIA were further assayed to
investigate site(s) of conjugation.
Labeling site determination was performed using a protocol similar to that
1
developed by Lampe et al. in their acetylation studies of GVIA. Briefly,
1
1
00 mg of trypsin was reconstituted with 100 ml of buffer (20 mM CaCl2,
00 mM Tris, pH 7.8), yielding a solution with a specific activity of 15 units/ml.
About 2 ml of trypsin solution was added to each GVIA-associated HPLC
fraction, which was previously reconstituted in 48 ml of the previous buffer.
The solution was vortexed and incubated for 20 h at 378C. After incubation,
the major digest fragment of each species was isolated via RP-HPLC, dried
and reconstituted in 50 ml of water. Thereafter, each fraction was subjected to
Copyright # 2006 John Wiley & Sons, Ltd.
J Label Compd Radiopharm 2006; 49: 269–283

2
80
V. AZARIAN ET AL.
MALDI-TOF MS, including a control sample of pure trypsin. The resulting
mass was compared to previously predicted values of fragment mass, based on
the hydrolytic reactivity of trypsin towards GVIA (Figure 2).
Calculating the specific activity (mAU/nmol) of K24-FGVIA
Specific activity was determined by a chromogenic protein assay, exclusive for
1
2
peptides containing hydroxyproline residues. First, a series of standard
solutions containing 0–5 mg of GVIA (in increments of 0.5 mg) were prepared
in water. Each solution was dried under reduced pressure, reconstituted in
6
00 ml of 6 M HCl, and transferred to individual test tubes. The tubes were
sealed and refluxed at 1108C for 16 h to ensure complete hydrolysis, and dried
under reduced pressure. Second, the assay buffer was prepared containing
anhydrous citric acid (4.57 g, 23.8 mmol), 1.2 ml glacial acetic acid, sodium
acetate trihydrate (12.0 g, 146 mmol), sodium hydroxide (3.4 g, 85.0 mmol),
and 100 ml toluene in deionized water (1 l). Each tube of dried hydrosylate was
reconstituted in 2 ml of assay buffer. The standards were then treated with 1 ml
of chloramine-T reagent (1.41 g chloramine-T, 26 ml n-propanol, and 53.3 ml
assay buffer per 100 ml of solution), and allowed to stand at room temperature
for 20 min. A 1 ml aliquot of a solution containing p-dimethylaminobenzalde-
hyde (15 g, 101 mmol), n-propanol (60 ml), and 26 ml perchloric acid (40%)
was added to each tube. The tubes were placed in a water bath at 608C for
1
5 min, and immediately transferred to a new water bath at room temperature
for 5 min. Absorption measurements were made at 550 nm for each standard
solution, and a curve was formulated. To correlate this standard curve with
absorption values at 215 nm for K24-FGVIA, the conjugate was purified by
RP-HPLC, the peak area at 215 nm (mAU ꢀ Sec) recorded, and the
hydroxyproline assay applied. The resulting molar amount of K24-FGVIA,
as determined by the Hyp-assay, was divided by the absorption area at 215 nm
to yield a conversion factor between area and mass quantity.
1
8
Preparation of K[ F]F-K complex
222
1
8
18 18
F] fluoride ion, in deionized water, was produced using the O(p,n) F
[
18
14
reaction by bombarding [ O]H O in a low volume (300 ml) silver target. The
2
activity, ranging from 5.6 to 6.7 GBq, was transferred to a solution containing
3
0 ml of 0.2 M K CO and 4,7,13,16,21,24-hexaoxa-1,10-diazabicyclo[8.8.8]-
2
3
hexacosane (4.5 mg, 6 nmol) in 50 ml of acetonitrile. Water was removed via
azeotropic evaporation (4 ꢀ 200 ml) of anhydrous acetonitrile under an argon
stream at 1508C.
18
18
Preparation of 4-[ F]fluorobenzoic acid ([ F]FBA, 3)
The preparation of 3 was modified from the protocols described by Lee and
7
Gangloff et al.
,15
18
The azeotropically dried F-ion was reconstituted with a
Copyright # 2006 John Wiley & Sons, Ltd.
J Label Compd Radiopharm 2006; 49: 269–283

1
8
SYNTHESIS AND ENCAPSULATION OF [ F] GVIA
281
1
20 ml solution of DMSO containing 1 (3 mg, 6 mmol). The mixture was
vigorously vortexed and heated to 1508C for 15 min. The reaction vial was
cooled and the contents poured into water (30 ml), and subsequently passed
through a C18 Sep Pak. Compound 2 was then eluted with 5 ml of ether, and
dried over a short Na SO column. Approximately 5 ml of anisole was added to
2
4
the ether, which was then evaporated under an argon stream. Once the ether
had been evaporated, 200 ml of trifluoroacetic acid (TFA) was added and
allowed to sit at room temperature for 5 min. The TFA was removed under an
argon stream at room temperature to yield 3. Typically, 2.59 GBq of 3 could
be obtained in 65 min starting from 6.29 GBq aliquot of a cyclotron produced
1
[
8
ꢁ
F]F batch (60% decay corrected yield).
Radiofluorniation of GVIA via direct-coupling with [ F]FBA
was reconstituted with 100 ml of anhydrous GVIA solution (1 nmol/10 ml
1
8
3
water). To this solution 50 ml of cyanophosphonate solution (1 mg/50 ml
acetonitrile) and 10 ml iPr NEt were added. The reaction vessel was sealed and
2
allowed to sit for 30 min at room temperature. The reaction mixture was
analyzed by RP-HPLC using the same chromatographic parameters as
19
described for the [ F]fluoroconjugation studies. No significant amount of
radiofluorinated GVIA was achieved.
18
18
Preparation of succinimidyl 4-[ F]fluorobenzoate ( F-SFB, 4)
was treated with N,N-succinimidyl carbonate (5 mg, 19.5 mmol) and 4-
dimethylaminopyridine (2 mg, 16.4 mmol) in acetonitrile (150 ml) at 908C for
min. Thereafter, 900 ml of water was added, and the reaction mixture was
3
5
loaded onto a Phenomonex Luna C-18 column (5 mm, 250 mm ꢀ 5 mm) and
eluted isocratically at 4 ml/min using 35:65:0.1 acetonitrile:water:acetic acid.
The eluatess were visualized at 235 nm and activity monitored via an online
gamma detector. The HPLC fraction corresponding to 4 was collected around
2
2 min and diluted to a volume of 30 ml using deionized water. The sample was
then applied to a C18 Sep-Pak. The retained activity was eluted with 5 ml of
ether and passed through a Na SO plug. The ether was evaporated under an
2
4
argon stream at room temperature. Typically, 1.30 GBq of 4 was prepared
18
from 6.29 GBq of F-ion in 120 min (44% decay corrected yield).
1
Radiofluorination of GVIA via succinimidyl 4-[ F]fluorobenzoate (4)
8
To 4 in acetonitrile (35 ml) was added 85 ml of a GVIA solution (3 nmol/10 ml
5
0 mM borate buffer). The reaction vessel was sealed and allowed to sit for
3
0 min at room temperature. The reaction mixture was analyzed by RP-HPLC
1
using the same chromatographic parameters as described for the F-
9
1
fluoroconjugation studies. Routinely, 0.11 GBq of K24-[ F]GVIA was
8
18
obtained from 6.29 GBq of F-ion in 165 min (5.0% decay corrected yield).
Copyright # 2006 John Wiley & Sons, Ltd.
J Label Compd Radiopharm 2006; 49: 269–283

2
82
V. AZARIAN ET AL.
Preparation of sulfatide containing liposome and K24-FGVIA encapsulation
Sulfatide containing liposomes were prepared essentially according to the
5
method described by Chen and Lee. Briefly, a 7:10:3 molar ratio solution of
phosphatidyl choline (PC): cholesterol (CH): sulfatide (SF) was prepared in
2
.5 ml of 1:1 methanol:chloroform. The solution was evaporated under
reduced pressure, and the residue immediately refrigerated at ꢁ208C until use.
The dry film was then reconstituted with 750 ml of ether and 1.5 ml of buffer
(5 mM Tris, 20 mM NaCl, pH 7.4). The volume of the HPLC fraction
corresponding to K24-FGVIA was adjusted to 1 ml and added to the lipid
mixture. The solution was vortexed for 5 min, followed by sonication for an
additional 5 min. The ether was removed under reduced pressure, and the
aqueous suspension applied to a short column (1 cm) of Sephadex G-60 gel for
size-exclusion chromatography. The fraction was eluted under pressure and
assayed for activity. An encapsulation efficiency of 40% was normally
observed.
Conclusion
We have accomplished the radiochemical synthesis of a biologically active
conjugate of o-conotoxin GVIA, with yields and labeling conditions suitable
1
8
for use of the F-lableled product in acquiring PET data. Preclinical studies to
assess the pharmacology and biodistribution of both the free and liposome
encapsulated form of the radiotracer are currently underway using small
18
animal-dedicated PET to examine K24-[ F]GVIA distribution in rats.
Acknowledgements
The authors would like to acknowledge the staff of the UCLA Biomedical
18
Cyclotron for the production of F-ion; Dr Lang and Dr Kiesewetter from
1
8
the National Institutes of Health (NIH) for aiding in the synthesis of [ F]SFB;
and the UCLA Pasarow Mass Spectrometry Laboratory for technical advice
and use of apparatus.
References
1
. Lampe RA, Lo MMS, Keith RA, Horn MB, McLane MW, Herman JL, Spreen
RC. Biochemistry 1993; 32: 3255–3260.
2
. Flinn JP, Pallaghy PK, Lew MJ, Murphy R, Angus JA, Norton RS. Eur J
Biochem 1999; 262: 447–455.
3
. Woessner JF Jr. Arch Biochem Biophys 1961; 93: 440–447.
. Lang L, Jagoda E, Schmall B, Vuong BK, Adams HR, Nelson DL, Carson RE,
Eckelman WC. J Med Chem 1999; 42: 1576–1586.
4
5
. Cheng D, Li QT, Lee KH. Biochem Intl 1992; 28: 297–304.
. Chen D, Lee KH. Biochim Biophys Acta 1993; 1158: 244–250.
6
Copyright # 2006 John Wiley & Sons, Ltd.
J Label Compd Radiopharm 2006; 49: 269–283

1
8
SYNTHESIS AND ENCAPSULATION OF [ F] GVIA
283
7
8
9
. Lee YCL, Kiesewetter DO, Lang L, Jagoda EM, Shirani J, Dilsizian V, Eckelman
WC. J Label Compd Radiopharm 2001; 44: S268–S270.
. Lang L, Jagoda E, Schmall B, Vuong BK, Adams HR, Nelson DL, Carson RE,
Eckelman WC. J Med Chem 1999; 42: 1576–1586.
. Olivera BM, Gray WR, Zeikus R, McIntosh JM, Varga J, Rivier J, de Santos V,
Cruz LJ. Science 1985; 230: 1338.
10. Vaidyanathan G, Zalutsky MR. Bioconjugate Chem 1994; 5: 352–356.
11. Keith RA, LaMonte D, Salama AI. J Auton Pharmacol 1990; 10: 139.
12. Attina M, Cacace G, Wolf AP. J Label Compd Radiopharm 1983; 20: 501.
13. Block D, Coenen HH, Stocklin G. J Label Compd Radiopharm 1988; 25: 185–200.
14. Wieland BW, Hendry GO, Schmidt DG, Bida G, Ruth TJ. J Label Compd
Radiopharm 1986; 23: 1202.
1
5. Gangloff A, Kesner AL, Hsueh WA, Kiesewetter DO, Eckelman W, Czernin J.
J Nucl Med 2003; 44: 297.
Copyright # 2006 John Wiley & Sons, Ltd.
J Label Compd Radiopharm 2006; 49: 269–283