Synthesis and in vitro evaluation of a novel iodinated resiniferatoxin derivative that is an agonist at the human vanilloid VR1 receptor

Article abstract of DOI:10.1016/S0960-894X(02)00127-0

Using a 'directed' iodination procedure, novel iodo-resiniferatoxin congeners were synthesized from 4-acetoxy-3-methoxyphenylacetic acid and resiniferinol- 9,13,14-ortho-phenylacetate (ROPA). The 2-iodo-4-hydroxy-5-methoxyphenylacetic acid ester of resini

Full text of DOI:10.1016/S0960-894X(02)00127-0

Bioorganic & Medicinal Chemistry Letters 12 (2002) 1189–1192
Synthesis and In Vitro Evaluation of a Novel Iodinated
Resiniferatoxin Derivative that is an Agonist at the Human
Vanilloid VR1 Receptor
Mark E. McDonnell,a Sui-Po Zhang,a Adrienne E. Dubinb and Scott L. Daxa,*
aJohnson & Johnson Pharmaceutical Research and Development, Welsh and McKean Roads, Spring House, PA 19477, USA
bJohnson & Johnson Pharmaceutical Research and Development, 3210 Merryfield Row, San Diego, CA 92121, USA
Received 15November 2001; accepted 6 February 2002
Abstract—Using a ‘directed’ iodination procedure, novel iodo-resiniferatoxin congeners were synthesized from 4-acetoxy-3-meth-
oxyphenylacetic acid and resiniferinol- 9,13,14-ortho-phenylacetate (ROPA). The 2-iodo-4-hydroxy-5-methoxyphenylacetic acid
ester of resiniferinol 5 displayed high affinity binding (Ki=0.71 nM) for the human vanilloid VR1 receptor and functioned as a
partial agonist. # 2002 Elsevier Science Ltd. All rights reserved.
The pungency and antinociceptive properties of Capsi-
cum species have been recognized and used thera-
peutically for centuries. The active ingredient, capsaicin,
(trans-8-methyl-N-vanillyl-6-nonenamide)1 acts upon a
population of mammalian sensory C- and Ad-fibers
causing excitation of primary afferents followed by
neuronal desensitization. The cloning of the human and
rat VR1 receptors2À4 as the putative pharmacological
target of capsaicin has reinvigorated research in vanil-
loids with the promise of developing novel analgesic
agents that are not irritants.5,6 To date, the most potent
natural agonist at VR1 is resiniferatoxin (RTX) 1 (Fig. 1),
a phorbol terpenoid isolated from the cactus Euphoria
resinifera that binds to VR1 with subnanomolar affinity.
by its low specific activity. We sought to synthesize
novel RTX derivatives that might similarly exhibit high
affinity for VR1 but could incorporate an iodine radio-
label (i.e., 125I). Here we report the synthesis and in
vitro evaluation of novel iodinated resiniferatoxin con-
geners. A salient feature of our work is the use of a
‘directed’ iodination protocol that installed the iodine
substituent at the 2-position of the homovanillic terminus,
adjacent to the benzylic carbon center.
Homovanillic acid was protected as the O-acetate upon
reaction with acetic anhydride in the presence of catalytic
sulfuric acid. Iodination was accomplished by treatment
with a slight excess of elemental iodine and silver tri-
fluoroacetate in methylene chloride.7 Analysis (HPLC
and MS) of the crude reaction revealed that iodination
occurred exclusively at a single position and NMR
analysis revealed this site was meta to the aryl hydroxy
substituent and para to the methoxy group of the
homovanillic moiety. The position of iodination is
apparently due to the o,p-directing effects of the meth-
oxy group in conjunction with meta direction from the
electron-withdrawing O-acetate protecting group.
Interestingly, iodination of homovanillic acid itself as
well as the corresponding ethyl carboxy ester gave
complex mixtures of products using the same condi-
tions. Thus the acetate group not only masks the
phenolic OH and directs iodination specifically to a
lone carbon center, but also ameliorates reactivity of
the aryl ring affording only mono-iodinated product 3.
Tritiated-RTX (3H-RTX) is a suitable radioligand in
receptor binding assays; however, its usefulness is limited
Figure 1. The structure of resiniferatoxin (RTX).
*Corresponding author. Tel.: +1-215-628-5211; fax: +1-215-628-
4985; e-mail: sdax@prius.jnj.com
0960-894X/02/$ - see front matter # 2002 Elsevier Science Ltd. All rights reserved.
PII: S0960-894X(02)00127-0

1190
M. E. McDonnell et al. / Bioorg. Med. Chem. Lett. 12 (2002) 1189–1192
4-Acetoxy-2-iodo-5-methoxyphenylacetic acid 3 was
then coupled to resiniferinol (ROPA) using car-
bodiimide, base and catalytic DMAP to afford iodo-
resiniferatoxin O-acetate 4.8 Cleavage of the acetate
group by treatment with pyrrolidine gave 9,13,14-
orthophenylacetylresiniferonyl-20-(4-hydroxy-2-iodo-5-
methoxyphenylacetate) 59 (Scheme 1).
(non-transfected) HEK293 cells were not affected by the
RTX congeners. The EC50 values for 4 and 5 were 935
and 160 nM, respectively (Table 1), ꢀ640- and 285-fold
less potent than their ability to inhibit 3H-RTX binding.
We determined whether the compounds reached equili-
brium in the 5min functional assay by determining the
concentration needed to half-block a subsequent cap-
saicin response after 60 min incubation. The potency of
5 did increase 10-fold, however, the potency of 4 was
unaltered. Degradation of 4 is unlikely to account for its
weak potency at human VR1 since cleavage of the ace-
tate group leads to the more potent 5. Similar dis-
crepancies in binding and functional activity have been
noted for other ion channels (e.g., nicotinic acetylcho-
line receptors) and have been attributed to increased
affinity of the radioligand for the desensitized receptor,
a state not detected in functional assays. Interestingly,
the iodinated RTX congener 5 functioned as a partial
agonist at human VR1, eliciting only 50% of the max-
imal response to CAP (Table 1), whereas 5 was a full
agonist at rat VR1 (92Æ6%, n=7). The partial agonism
of 5 was confirmed using whole cell patch clamp
technology (personal communication).
The VR1 binding assay10 used was a modified version of
that previously reported by Szallasi and Blumberg.11
Membranes harvested from a HEK293 cell line stably
expressing human VR1 cDNA (hVR1/HEK293)12 were
incubated with 3H-RTX and the test compounds 1
and 4–6. After incubation and centrifugation, bound
3H-RTX was quantified.
The permeability of VR1 to Ca2+ enabled the func-
tional activity of the iodinated RTX analogues to be
determined using a Ca2+-sensitive fluorescent dye and
FLIPRTM technology (Molecular Devices, Inc.).13 The
magnitude of agonist-induced increases in intracellular
Ca2+ was compared to that evoked by a subsequent
challenge with capsaicin at a concentration eliciting a
maximal response.14 All compounds (1, RBI; 6,15
Tocris) were tested for their ability to block a sub-
sequent capsaicin response eliciting ꢀ80% of the max-
imal response. The whole cell configuration of the patch
clamp technique was used to compare the magnitude of
currents induced by 5 (10ÂEC50) with that of capsaicin
(3 mM). Cells were maintained in 0 Ca2+ buffer to allow
multiple challenges to the same cell.12
Wahl recently reported that reaction of RTX with
sodium iodide and chloramine-T afforded the 2-iodo-4-
hydroxy-3-methoxy phenylacetic acid ester of resinifer-
inol 6 (Fig. 2) and demonstrated that this iodinated
RTX analogue is a antagonist of rat VR1.15 We con-
firmed this finding and further demonstrated that 6 is an
antagonist at human VR1 (Table 1).
Utilization of our ‘directed’ iodination protocol enabled
the synthesis of new isomeric iodo-RTX analogues that
retained high affinity for VR1 but displayed unexpected
vanilloid receptor pharmacology. Resiniferatoxin
(RTX) 1 bound to human VR (Table 1) with high
affinity and functioned as a full agonist of the receptor.
The iodinated-RTX congener 5 exhibited sub-nanomo-
lar affinity for VR1 whereas the acetate precursor 4 was
slightly less potent (Table 1). In functional assays, both
iodinated resiniferatoxin derivatives (4 and 5) induced
increases in intracellular Ca2+ in VR1-expressing cell
lines. These effects were mediated through the vanilloid
channel since the VR1 antagonists capsazepine and
ruthenium red abolished calcium flux and since control
As reported here, a ‘directed-iodination’ protocol was
used to synthesize novel RTX analogues that retained
high affinity for the human VR1 receptor. Iodo-RTX 5
bound to human VR1 with subnanomolar affinity and
functioned as a partial agonist, compared to the full
agonism observed with RTX. It is striking that the site
Table 1. Human VR1 binding affinities and functional activity of
RTX and iodinated RTX congenersa
Compd
Binding
affinity
Ki (nM)
Functional
activity
(nM)
Efficacy
(% capsaicin
response)
1 (RTX) 0.48Æ0.07 (n=5) EC50=1.9Æ1.0 (n=9)
116 (Æ6)
75( Æ15)
50 (Æ13)
n.a.b
4
5
6
1.46Æ0.55 (n=5) EC50=935Æ280 (n=7)
0.71Æ0.27 (n=5) EC50=130Æ105( n=9)
1.4Æ (n=1)
IC50=27Æ11 (n=3)
aValues are meansÆstandard deviation of no. experiments (n).
bn.a., not applicable (compound is not an agonist).
Scheme 1. (a) Ac2O, cat. H2SO4 (97%); (b) I2 (1.1 equiv), AgO2CCF3
(1.1 equiv)/CH2Cl2 (89%); (c) DCC, cat. DMAP, pyrrolidine (30
equiv)/CH2Cl2 (58%); (d) pyrrolidine/CH2Cl2 (68%).
Figure 2. Iodinated resiniferatoxin reported by Wahl.

M. E. McDonnell et al. / Bioorg. Med. Chem. Lett. 12 (2002) 1189–1192
1191
of iodination plays such an important role in function-
ality in light of an isomeric iodo-RTX reported to be a
antagonist of rat VR1.15 The use of radioactive ligands
with partial or opposing functional activity will aid in
understanding the RTX/capsaicin binding pocket of
VR1.
157.8, 151.4, 146.3, 139.0, 136.5, 136.3, 134.9, 134.0, 133.0,
130.8, 128.6, 127.6, 126.5, 117.7, 114.2, 110.7, 87.9, 84.4, 81.0,
80.5, 73.1, 71.3, 56.0, 55.2, 46.3, 40.9, 39.9, 39.0, 35.6, 33.9,
32.9, 24.9, 20.6, 19.9, 18.8, 10.3. Accurate mass (FAB MH+):
calcd for C39H41I1O10, 797.1951; found, 797.1822.
9. 9,13,14-Orthophenylacetylresiniferonyl-20-(4-hydroxy-2-iodo-
5-methoxyphenylacetate) (5). 9,13,14-Orthophenylacetylresini-
feronyl-20-(3-acetoxy-4-methoxy-6-iodophenylacetate)
(17
mg, 0.02 mmol) was dissolved in dichloromethane (1 mL) and
stirred at room temperature under argon. Pyrrolidine (59.7
uL, 0.71 mmol) was then added and the reaction stirred for 2 h
after which time no starting material was visible by TLC. The
solvent was removed in vacuo and the residue was purified on
a TLC preparative plate (silica gel, CHCl3). The product was
extracted (chloroform) from the silica and the solvent evapo-
rated after filtration to give the desired iodo-RTX congener 5
as a white solid (yield=11 mg; 68%). Analytical reverse-phase
HPLC (gradient 40–90% MeCN, 0.1% TFA) tR=4.83 min,
ꢀ90%; IR (KBr) 2850, 1732, 1705, 1504, 1270, 1027 cmÀ1; 1H
NMR (CDCl3, 500 MHz) d 7.37 (t, 1H, J=1.9 Hz), 7.31 (d,
2H, J=1.5Hz), 7.23–7.15(m, 1H), 6.76 (d, 1H, J=8.0 Hz),
6.72–6.67 (m, 2H), 5.80 (s, 1H), 5.48 (s, 1H), 4.63 (s, 2H), 4.59
(q, 2H, J=12.1 Hz), 4.12 (d, 1H, J=2.7 Hz), 3.8 (s, 3H), 3.49
(s, 2H), 3.14 (s, 2H), 3.0 (d, 2H, J=13.4), 2.47 (t, 1H, J=7.5
Hz), 2.35(d, 1H, 18.9), 2.09–1.99 (m, 3H), 7.5(s, 3H), 1.61–
1.59 (s, 1H broad), 1.44 (s, 3H), 0.88 (d, 4H, J=7.1 Hz); 13C
NMR (CDCl3, 500 MHz): d 208.4, 171.4, 158.2, 146.5, 146.2,
136.5, 134.9, 134.0, 130.8, 128.5, 127.7, 126.5, 125.6, 122.2,
117.7, 114.3, 111.7, 110.7, 84.4, 81.0, 80.5, 73.3, 70.5, 65.8,
55.9, 55.3, 41.0, 40.9, 39.9, 39.0, 35.6, 33.0, 19.8, 18.8, 10.3.
Accurate mass (FAB MH+): calcd for C37H39I1O9, 755.1741;
found, 755.1717.
Acknowledgements
We thank Arne Huvar, Sean Brown, Ed Kaftan, Nadia
Nasser, Susan Yagel, Larry Weaner and Ellen Codd for
their assistance.
References and Notes
1. Jancso, J. Bull. Millard Fillmore Hosp. (Buffalo) 1960, 7,
53.
2. Caterina, M. J.; Schumacher, M. A.; Tominaga, M.;
Rosen, T. A.; Levine, J. D.; Julius, D. Nature (London) 1997,
389, 816.
3. Julius, D. J.; Caterina, M.; Brake, A. PCT Int. Appl.,
9909140A1, 1999; Chem. Abstr. 1999, 130, 207838.
4. Dubin, A.; Erlander, M.; Rogers, K. E.; Huvar, A.; Huvar,
R. PCT Int. Appl. WO 0063415A1, 2000; Chem. Abstr. 2000,
133, 318312.
5. Lee, J.; Lee, J.; Kim, J.; Kim, S. Y.; Chun, M. W.; Cho, H.;
Hwang, S. W.; Oh, U.; Park, Y. H.; Marquez, V. E.; Beheshti,
M.; Szabo, T.; Blumberg, P. M. Bioorg. Med. Chem. 2001, 9,
19.
6. Urban, L.; Campbell, E. A.; Panesar, M.; Patel, S.;
Chaudhry, N.; Kane, S.; Buchheit, K.-H.; Sandells, B.; James,
I. F. Pain 2000, 89, 65 .
10. [3H]RTX binding assay using hVR1/HEK293 cell mem-
branes. Preparation of membranes. hVR1/HEK293 were
homogenized with a Polytron twice and centrifuged at 3000
rpm for 10 min in HEPES buffer containing 20 mM HEPES,
pH 7.4, NaCl 5.8 mM, sucrose 320 mM, MgCl2 2 mM, CaCl2
0.75mM and KCl 5mM. The supernatant was centrifuged at
18,000 rpm for 20 min. The pellet was saved in a tube and 10
mL assay buffer was added into the tube. The pellet and buffer
were mixed with a Polytron. Incubation procedure. Incubations
for 60 min at 37 ꢁC were performed in a total volume of 0.5
mL that contained 120 mg/mL membrane protein and 0.3–0.6
nM 3H-RTX (NEN, Boston MA, USA) in the HEPES buffer.
After incubation, the samples were cooled on ice and 100 mg of
a-acid glycoprotein were added followed by centrifugation at
13,000 rpm for 15min. The supernatant was aspirated and the
tips of tubes were cut off into 6 mL vials. Nonspecific binding
was measured in the presence of 200 nM unlabeled RTX in 4
mL scintillation liquid using a Packard scintillation counter.
Data analysis. Percent (%) inhibition=(total bindingÀtotal
binding in presence of compound)Ã100/(total bin-
dingÀnonspecific binding). Ki values were calculated using
Prism (GraphPad, San Diego, CA, USA).
7. 4-Acetoxy-2-iodo-5-methoxyphenylacetic acid (3): 4-Ace-
toxy-3-methoxyphenylacetic acid (200 mg, 0.89 mmol), silver-
trifluoroacetate (217 mg, 0.98 mmol) and iodine (237 mg, 0.93
mmol) were combined and stirred in dry dichloromethane (20
mL) for 2 h under argon. The reaction was filtered. The filtrate
was washed (satd NaCl solution), dried (MgSO4), and solvents
removed to afford the title compound 3 as a white powder
(yield=279 mg; 89%). Analytical reverse-phase HPLC (gra-
1
dient 10–90% MeCN, 0.1% TFA) tR=3.56 min, ꢀ99%; H
NMR (CD3OD, 300 MHz) d 7.46 (s, 1H), 7.10 (s, 1H), 3.79 (s,
3H), 3.78 (s, 2H), 2.24 (s, 3H); 13C NMR (CD3OD, 300 MHz)
d 174.4, 170.7, 153.3, 140.9, 138.8, 134.1, 116.5, 88.8, 56.9,
46.9, 20.7. Anal. calcd for C11H11O5I1: C, 37.74; H, 3.17.
Found: C, 37.93; H, 3.31.
8. 9,13,14-Orthophenylacetylresiniferonyl-20-(4-acetoxy-2-iodo-
5-methoxyphenylacetate) (4): To a round-bottom flask (10 mL)
the following reagents were added: dichloromethane (4 mL),
resiniferonol-9,13,14-orthophenylacetate (10 mg, 0.02 mmol),
4-dimethylaminopyridine (0.28 mg, 0.0022 mmol), dicyclo-
hexylcarbodiimide (4.8 mg, 0.023 mmol) and 2-iodo-4-acet-
oxy-5-methoxyphenylacetic acid (8.2 mg, 0.023 mmol). The
reaction was stirred at room temperature, under argon, and in
the dark for 24 h. The solvent was removed in vacuo and the
residue chromatographed on silica gel (chloroform). Solvent
was removed from appropriate fractions to yield adduct 4 as a
white solid (yield=10 mg; 59%). Analytical reverse-phase
HPLC (gradient 60–90% MeCN, 0.1% TFA) tR=4.27 min,
11. Szallasi, A.; Blumberg, P. M. Naunyn-Schmiedeberg’s
Arch. Pharmacol. 1993, 347, 84.
12. Grant, E. R.; Dubin, A. E.; Zhang, S.-P.; Zivin, R. A.;
Zhong, Z. J. Pharmacol. Exp. Ther. 2002, 300, 9.
13. Dubin, A. E.; Huvar, R.; D’Andrea, M. R.; Pyati, J.; Zhu,
J. Y.; Joy, K. C.; Wilson, S. J.; Galindo, J. E.; Glass, C. A.;
Luo, L.; Jackson, M. R.; Lovenberg, T. W.; Erlander, M. G.
J. Biol. Chem. 1999, 274, 799.
14. hVR1/HEK cells4 or cells stably expressing rat VR12 were
seeded on poly-d-lysine coated 96-well, black-walled plates
(BD 354640) and 2 days later loaded with Fluo-3/AM for 1 h
and subsequently tested for agonist-induced increases in intra-
cellular Ca2+ levels using FLIPRTM technology.13 Rat VR1 was
cloned from a rat thalamus cDNA library and is identical to
Genbank accession #AF029310. Rat and hVR1/HEK cells were
1
ꢀ95%; H NMR (CDCl3, 500 MHz) d 7.44 (s, 1H), 7.35(s,
1H), 7.29 (d, 2H, J=6.8 Hz), 7.22–7.15(m, 2H), 6.84 (s, 1H),
5.80 (s, 1H), 4.63 (s, 2H), 4.51 (s, 2H), 4.15 (d, 2H, J=2.1 Hz),
3.79–3.66 (m, 6H), 3.13 (s, 2H), 2.99 (d, 2H, J=14.3 Hz), 2.51
(t, 1H, J=7.6 Hz), 2.39 (s, 1H), 2.26 (s, 1H), 2.22 (s, 3H),
2.09–1.96 (m, 3H), 1.74 (s, 1H), 1.43 (s, 1H), 0.88 (d, 4H,
J=7 Hz); 13C NMR (CDCl3, 500 MHz) d 208.2, 169.7, 168.9,

1192
M. E. McDonnell et al. / Bioorg. Med. Chem. Lett. 12 (2002) 1189–1192
occasionally seeded on different halves of the same plate for
more direct comparisons of compound effects on VR1 from
these species. Cells were challenged with single concentrations
of compound (1, 4, 5, and 6) and intracellular Ca++ was
measured for 5min prior to the addition of CAP to all wells to
achieve a final CAP concentration of 3 mM to fully activate
VR1. Cells in columns 1–11 received compounds in increasing
concentrations by half-log intervals from 0.1 nM (4) or 0.01
nM (5); column 12 received buffer. Antagonist potency was
determined using a similar protocol except that after incu-
bation with 1, 4, 5 and 6, cells were challenged with CAP at a
concentration eliciting ꢀ80% maximal response. Efficacy
determinations were made by normalizing the maximum iodo-
RTX-induced fluorescence change to the change elicited by a
supramaximal concentration (3 mM) of capsaicin. Data were
analyzed using Prism software.
15. Wahl, P.; Toged, C.; Soren, T.; Thomsen, C. Mol. Pharm.
2001, 59, 9.