Structure-activity relationships of the ultrapotent vanilloid resiniferatoxin (RTX): The side chain benzylic methylene

Article abstract of DOI:10.1016/j.bmcl.2009.11.035

The side chain benzylic methylene is a critical element for the vanilloid activity of resiniferatoxin (2a, RTX), and introduction of branching, oxygen functions, or isosteric substitution at this center proved detrimental, with a decrease of potency of 2-3 orders of magnitude compared to the natural product. Conversely, only a modest erosion of activity was observed upon α-methylation and α-methylenation of the side chain. Surprisingly, introduction of an iodine atom in the guaiacyl moiety of the oxygen isoster 2h led to an unexpected and remarkable (>1000-fold) increase of potency, affording 2i, a compound that outperforms RTX in terms of vanilloid agonism and represents the first one-digit picomolar ligand of a TRP channel discovered to date.

Full text of DOI:10.1016/j.bmcl.2009.11.035

Bioorganic & Medicinal Chemistry Letters 20 (2010) 97–99
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Structure–activity relationships of the ultrapotent vanilloid resiniferatoxin
(RTX): The side chain benzylic methylene
a
a
b,
Giovanni Appendino a, , Abdellah Ech-Chahad , Alberto Minassi , Luciano De Petrocellis
,
*
*
Vincenzo Di Marzo c
a Dipartimento di Scienze Chimiche, Alimentari, Farmaceutiche e Farmacologiche, Università del Piemonte Orientale, Via Bovio 6, 28100 Novara, Italy
b Endocannabinoid Research Group, Institute of Cibernetica ‘Eduardo Caianiello’, CNR, Via Campi Flegrei 34, Pozzuoli (NA), Italy
c Endocannabinoid Research Group, Institute of Biomolecular Chemistry, CNR, Via Campi Flegrei 34, 80078 Pozzuoli (NA), Italy
a r t i c l e i n f o
a b s t r a c t
Article history:
The side chain benzylic methylene is a critical element for the vanilloid activity of resiniferatoxin (2a,
RTX), and introduction of branching, oxygen functions, or isosteric substitution at this center proved det-
rimental, with a decrease of potency of 2–3 orders of magnitude compared to the natural product. Con-
Received 24 September 2009
Revised 6 November 2009
Accepted 9 November 2009
Available online 14 November 2009
versely, only a modest erosion of activity was observed upon a-methylation and a-methylenation of the
side chain. Surprisingly, introduction of an iodine atom in the guaiacyl moiety of the oxygen isoster 2h
led to an unexpected and remarkable (>1000-fold) increase of potency, affording 2i, a compound that
outperforms RTX in terms of vanilloid agonism and represents the first one-digit picomolar ligand of a
TRP channel discovered to date.
Dedicated to Professor Aldo Martelli on the
occasion of his retirement as a Dean of the
Faculty of Pharmacy in Novara
Ó 2009 Elsevier Ltd. All rights reserved.
Keywords:
Resiniferatoxin (RTX)
TRPV1
Structure–activity relationships
Capsaicin
Vanilloids
Ever since its isolation in 1975,1 and its identification as a cap-
saicin (1a) analogue in 1989,2 the daphnane diterpenoid resinifera-
toxin (2a) has amazed life scientists for its extraordinary and still
unmatched potency in animal models of inflammation and in cel-
lular assays of vanilloid activity.1,2 While the clinical development
of RTX seems to have come to a standstill,3 this compound remains
an indispensable tool to investigate the capsaicin receptor
(TRPV1),4 where its role as a privileged structure was further vali-
dated by the serendipitous discovery that the powerful agonistic
activity of the natural product (EC50 ca. 19 pM on hTRPV1) could
be reversed by aromatic iodination of the acyl moiety, generating
the potent (IC50 = 4.0 nM) vanilloid antagonist iodoresiniferatoxin
(I-RTX, 2b).5
product. The homovanillyl moiety has long been known to be
essential for binding of RTX to TRPV1.7 Within this element, the
benzylic methylene is critical,8 since ultrapotency was only ob-
served in analogues of the natural products where only one carbon
links the ester carbonyl and the aryl moiety, with homologation
and deletion of this methylene leading to a dramatic (ca. three or-
ders of magnitude) decrease of potency.8 By comparison, manipu-
lation of the substitution pattern of the aryl moiety was better
tolerated.8,9
With the aim of obtaining information on the molecular bases
of this observation, we have investigated the effect of inserting a
branching or an oxygen function on the critical side chain methy-
lene, and its isosteric replacement with an oxygen atom. These
changes are expected to perturb the spatial relationship between
the carbonyl oxygen and the aromatic ring, providing information
useful to map the vanilloid recognition site of TRPV1.10 Further-
more, we reasoned that by identifying alternative side chain tem-
plates for iodination, data useful to rationalize in molecular terms
the amazing reversal of activity observed upon iodination of the
natural product might also emerge.
Capitalizing on the development of an expeditious protocol to
obtain the terpenoid core of RTX (resiniferonol orthophenylacetate,
ROPA, 3), from the household ornamental Euphorbia resinifera
Berg,6 we have adventured in the still substantially uncharted field
of the structure–activity relationships of this remarkable natural
* Corresponding authors. Tel.: +39 0321375744; fax: +39 0321375621 (G.A.); tel.:
+39 0818675173 (L.D.P.).
E-mail addresses: appendino@pharm.unipmn.it (G. Appendino), l.depetrocel-
lis@cib.na.cnr.it (L. De Petrocellis).
The esters 2c–g were prepared by Mitsunobu esterification of
ROPA with the corresponding carboxylic acids.11 These were in
turn prepared from vanillin (4a) by straightforward chemistry, as
0960-894X/$ - see front matter Ó 2009 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2009.11.035

98
G. Appendino et al. / Bioorg. Med. Chem. Lett. 20 (2010) 97–99
MeO
COOH
OMe
outlined in Scheme 2. Thus, after protection of the phenolic hydro-
xyl as a pivalate, addition of trimethylsilyl cyanide (TMSCN) affor-
ded the cyanidrine 5, next converted by selective hydrolysis of its
corresponding diacetate to the vanillomandelate 6, and, by hydro-
lysis of the corresponding (2-(O)methyl) ester, to the vanillo-
O
COOH
OMe
AcO
COOH
OMe
OH
OH
OH
7
9
6
mandelate 7. Alternatively,
5 was esterified to 8, and next
oxidized, affording, after hydrolysis, the phenylglyoxylate 9.1 Wit-
tig methylenation of 8 could only be achieved after protection of
the phenolic hydroxyl as a TIPS (triisopropylsilyl) ether, providing,
after global deprotection, the phenylacrylate 10b. Alternatively,
iii
iv
vi
HO
COOH
O
COOMe
CHO
v
ii
OMe
10a was hydrogenated and next hydrolyzed to the
onate 11b.
a-phenylpropi-
OMe
OMe
After esterification with ROPA and deprotection, all crude final
products were further purified to >95% purity by preparative HPLC
on silica gel. Little diastereomeric resolution was observed by
esterification of the racemic acids 5–7, 11b with chiral and enan-
tiopure ROPA (3), a primary alcohol, and the HPLC profile of 2c–f
showed a narrowly and almost symmetrically split peak in all
cases. A similar splitting was observed in the signal of a few side
chain signal of the 1H and the 13C NMR signals. No attempt was
done to resolve the diastereomeric mixtures 2c–f and assess the
configuration of their individual components.
OH
OR
R
OH
8
5
4a H
i
4b Piv
COOR1
COOR1
ix
vii
OMe
OMe
OR2
R1 R2
OR2
R1 R2
All final products were assayed in human embryonic kidney
(HEK)-293 cells over-expressing the human TRPV1 (Scheme 1).12
Despite the limitation that 2c–f were diastereomeric pairs, the bio-
logical investigation of this set of 200-modified analogues afforded
interesting results. Thus, introduction of a hydroxyl (2c) or a meth-
oxyl (2d) on the side chain benzylic methylene caused a decrease
of activity of three orders of magnitude, while the decrease of
potency was more modest (ca. 30-fold) for an acetoxyl. On the other
hand, a certain tolerance was observed with the introduction of an
alkyl group, and only a modest (ca. sevenfold) erosion of potency
was observed in the methylene analogue 2g. These data, while con-
firming the critical role of the side chain benzylic position, also
showed that polar groups are poorly accommodated at this site, pre-
sumably because of the lipophilicity of the complementary surface
of the ligand-binding pocket of TRPV1. The better tolerance to the
introduction of a 200-methylene or a 200-methyl supports this view.
The presence of diastereomeric mixtures precluded any more re-
fined speculation on nature of the active conformation of RTX, that is
critically dependant on the spatial orientation of the substituents
around the side chain C100–C200 bond. To overcome this limitation,
isosteric substitution of the benzylic methylene was investigated,
10a Me TIPS
10b H
11a Me TIPS
11b H
viii
x
H
H
Scheme 2. Synthesis of the modified homovanillic acids 5–7, 9, 10b, and 11b.
Reagents and conditions: (i) pivaloyl chloride, pyridine (86%); (ii) (a) TMSCN, cat.
CuOTf, THF; (b) KOH, MeOH (overall 61%); (iii) (a) Ac2O (b) pyrrolidine (3 equiv),
THF (overall 31%); (iv) (a) SOCl2, MeOH, 72 h (26%); (b) LiOH (5 equiv), THF–MeOH;
(v) Jones reagent (76%); (vi) LiOH (3 equiv), THF–MeOH (72%); (vii) (a) TIPS-Cl
(4 equiv), imidazole, DMF; (b) CH2@PPh3, THF, ꢀ78% (overall 71%); (viii) LiOH
(6 equiv), THF–MeOH (71%); (ix) H2, Pd(C); (b) LiOH (6 equiv), THF–MeOH (overall
70%).
and 2h, the carbonate analogue of RTX, was prepared by esterifica-
tionof ROPAwith theprotectedand complementary guaiacylchloro-
formate. After deprotection, the carbonate 2h was assayed for
vanilloid activity, showing a decrease of activity similar to that ob-
served for the polar analogues 2c and 2d. Overall, 2h was still a good
TRPV1 agonist, ca. threefold more potent than capsaicin
(EC50 = 13.7 nM vs 40 nM), and was therefore chosen as a platform
for iodination because of its substantial structural difference with
the side chain of RTX. We had previously observed that 50-iodination
potentiates (ca. fivefold), rather than reverting, the activity of ROPA
vanillate, the lower, and considerably less potent (EC50 9400 pM vs
19 pM), homologue of RTX.8 Since carbon–oxygen bonds are consid-
erable shorter than carbon–carbon bonds (1.34 Å for C(sp2)–O bonds
vs 1.51 Å for C(sp3)–C(sp2) bonds),13 we expected a somewhat sim-
ilar situation, an important hint to suggest that the reversal effect
critically depends on the carbonyl-to-aryl distance in the side chain.
The iodinated chloroformate (16) requested for the synthesis of 2i
was prepared from commercial 5-iodovanillin (12a) as outlined in
Scheme 3. After protection of the phenolic hydroxyl as a Mem
(methyoxyethoxymethyl)ether, Dakin oxidation introduced the
missing aryl hydroxyl, then reacted with oxalyl chloride to install
the activated carbonyl function. After coupling with ROPA, removal
of the Mem ether could only be achieved by using the SnCl4–THF ad-
duct,14 sincemoreconventional method(Brønstedacidsand SnCl4 in
CH2Cl2) led to complete degradation. Surprisingly, evaluation of the
vanilloid activity of 2i showed that the introduction of the iodine
atom had caused a >1000-fold increase of activity, affording an ago-
nist that outperformed RTX in a comparative assay of vanilloid
potency (EC50 = 5.4 vs 19 pM).12 Indeed, with a one-digit picomolar
EC50 value, 2i is the most potent ligand ever reported for a TRP chan-
nel1.10,15 Given the intracellularlocalizationof the vanilloid-binding
site on TRPV1,10 it is unclear if the dramatic increase of activity
16
15 17
12
13
O 1'
O
11
18
1
O
2'
4'
7'
6
3'
O14
5'
6'
H10
N
H
9
2
8
H
8'
19
R
7
4
HO
5
O
3
20
1''
OMe
HO
6
O
2''
O
A
R
3''
8''
1a Me Δ6,7
1b
4''
H
-
7''
R
5'' OMe
OH
6''
R
A
EC50 (pM)
19
2a H CH2
2b I CH2
4,000*
O
2c H CHOH 19,000
2d H CHOMe 17,400
O
H
O
H
2e H CHOAc
2f CH-CH3
650
980
150
H
2g H CH=CH2
HO
O
OH
2h H
2i
O
O
13,700
20
I
4.7
3
*IC50 value
Scheme 1.

G. Appendino et al. / Bioorg. Med. Chem. Lett. 20 (2010) 97–99
2. Szallasi, A.; Blumberg, P. M. Neuroscience 1989, 30, 515.
99
H
3. Recent clinical studies: (a) Refractory detrusor overactivity: Kuo, H. C.; Liu, H.
T.; Yang, W. C. J. Urol. 2006, 176, 641. (b) Cancer pain: Hartel, M.; di Mola, F. F.;
Selvaggi, F.; Mascetta, G.; Wente, M. N.; Felix, K.; Giese, N. A.; Hinz, U.; De
Sebastiano, P.; Buchler, M. W.; Friess, H. Gut 2006, 55, 519, and Ghilardi, J. R.;
Robrich, H.; Lindsay, T. H.; Sevick, M. A.; Kubota, K.; Halvorson, K. G.; Poblete, J.;
Chaplan, S. R.; Dubin, A. E.; Carruthers, N. I.; Swanson, D.; Kuskowski, M.;
Flores, C. M.; Julius, D.; Manthyh, P. W. J. Neurosci. 2005, 25, 3126. RTX was
found ineffective for the treatment of interstitial cystitis (Payne, C. K.;
Mosbaugh, P. G.; Forrest, J. B.; Evans, R. J.; Withmore, K. E.; Antoci, J. P.;
Perez-Marrero, R.; Jacoby, K.; Diokno, A. C.; O’Reilly, K. J.; Greibling, T. L.;
Vasavada, S. P.; Yu, A. S.; Frumkin, L. R. J. Urol., 2005, 173, 1590).
4. Caterina, M. J.; Schumacher, M. A.; Tominaga, M.; Rosen, T. A.; Levine, J. D.;
Julius, D. Nature 1997, 389, 816.
5. Wahl, P.; Foged, C.; Tullin, S.; Thomsen, C. Mol. Pharmacol. 2001, 59, 9.
6. Fattorusso, E.; Lanzotti, V.; Taglialatela-Scafati, O.; Tron, G. C.; Appendino, G.
Eur. J. Org. Chem. 2002, 71.
7. Walpole, C. S. J.; Bevan, S.; Bloomfield, G.; Breckenridge, R.; James, I. F.; Ritchie,
T.; Szallasi, A.; Winter, J.; Wrigglesworth, R. J. Med. Chem. 1996, 39, 2939.
8. Appendino, G.; Ech-Chahad, A.; Minassi, A.; Bacchiega, S.; De Petrocellis, L.; Di
Marzo, V. Bioorg. Med. Chem. Lett. 2007, 17, 132.
H
O
O
I
O
OH
ii
iii
I
OMe
OMe
OMem
13
I
OMe
OMem
14
OR
R
12a H
i
12b Mem
Cl
O-ROPA
O
O
I
O
O
I
vi
iv
v
2i
OMe
OMe
OMem
16
OMem
15
9. Choi, H. K.; Choi, S.; Lee, Y.; Kang, D. W.; Ryu, H.; Maeng, H. J.; Chung, S. J.;
Pavlyukovets, V. A.; Pearce, L. V.; Toth, A.; Tran, R.; Wang, Y.; Morgan, M. A.;
Blumberg, P. M.; Lee, J. Bioorg. Med. Chem. 2009, 19, 690.
10. Vriens, J.; Appendino, G.; Nilius, B. Mol. Pharmacol. 2009, 75, 1262.
11. Appendino, G.; Minassi, A.; Daddario, N.; Bianchi, F.; Tron, G. C. Org. Lett. 2002,
4, 3839.
Scheme 3. Synthesis of the modified homovanillic acids 4–9. Reagents and
conditions: (i) MemCl, DIPEA, toluene (91%); (ii) MCPBA, DCM, 90%; (iii) K2CO3
(60%); (iv) COCl2, 79%; (v) (a) 3, DMAP, pyridine (41%); (b) SnCl4, THF (56%)
ROPA = 3; MCPBA = m-chloroperoxybenzoic acid; Mem = methoxyethoxymethyl;
DIPEA = diisopropyletylamine; DMAP = 4-dimethylaminopyridine.
12. Cells were grown as monolayers in minimum essential medium supplemented
with non-essential amino acids, 10% fetal calf serum and 2 mM glutamine, and
maintained under 95%/5% O2/CO2 at 37 °C. The effect of the substances on
associated to iodination of the carbonate 2h is due to higher binding,
better membrane penetration, or to a combination of both.
[Ca2+
] was determined by using Fluo-4, a selective intracellular fluorescent
i
probe for Ca2+. On the day of the experiment the cells (50–60,000 per well)
were loaded for 1 h at 25 °C with 4 M Fluo-4 methyl ester (Molecular Probes)
Our limited knowledge on the vanilloid-binding site of TRPV110
makes it difficult to translate these observations into new molecu-
lar insights. Nevertheless, it seems reasonable to assume that io-
dine plays a directing role in the interaction between TRPV1 and
halogenated vanilloids, as suggested by a study on the introduction
of various substituents on the structurally simple capsaicinoid ago-
nist nonivamide (1b).16 Thus, reversal of activity, though observed
with all halogens, followed the strength order of halogen bonding
(I > Br > Cl),17 while the replacement of iodine with alkyl, alkenyl or
alkynyl groups led to a dramatic decrease of affinity. Iodine is eas-
ily engaged in halogen bonding with tyrosine residues,17 and the
vanilloid-binding site of TRPV1 contains tyrosine residues that
are essential for ligand binding.10 It is therefore tempting to as-
sume that a switch from hydrogen-bonding to iodine-bonding
underlies both the reversal of activity of RTX (2a) and the amplifi-
cation of activity of its oxo-isoster 2h, caused by the introduction
of a iodine atom ortho to the phenolic hydroxyl. Though specula-
tive, these considerations have a chemical precedent, since compe-
tition between H- and I-bonding is well documented in
supramolecular crystal assemblies,18 thus qualifying the RTX–
TRPV1 interaction as a critical probe to investigate the biological
relevance of halogen bonding.
l
in DMSO containing 0.04% Pluoronic. After the loading, cells were washed with
Tyrode pH 7.4, trypsinized, resuspended in Tyrode and transferred to the
quartz cuvette of the fluorescence detector (Perkin–Elmer LS50B) under
continuous stirring. Experiments were carried out by measuring cell
fluorescence at 25 °C (kEX = 488 nm, kEM = 516 nm) before and after the
addition of the test compounds at various concentrations. Kinetic effects
(Szallasi, A.; Blumberg, P. M.; Annicelli, L. LK.; Krause, J. E.; Cortright, D. N. Mol.
Pharmacol. 1999, 56, 581) were not observed in the calcium fluorescence assays
as all compounds behaved exactly like RTX in terms of onset and duration of
response.
13. Smith, M. B.; March, J. March’s Advanced Organic Chemistry, 6th ed.; Wiley:
Hoboken, US, 2007.
14. Appendino, G.; Cravotto, G.; Palmisano, G.; Annunziata, R.; Szallasi, A. J. Med.
Chem. 1996, 39, 3123.
15. Foam. IR (KBr), cmꢀ1: 3450, 1758, 1705, 1622, 1509, 1375, 1226, 1028, 812; 1
H
NMR (CDCl3, 300 MHz): d 0.98 (d, J = 7.3 Hz, 18-Me), 1.53 (s, 17-Me), 1.56 (m,
H-12a), 1.81 (br d, J = 1.1 Hz, H-19), 2.10 (m, H-12b), 2.23 (d, J = 19.0 Hz, H-5a),
2.56 (d, J = 19.0 Hz, H-5b), 2.57 (m, H-11), 3.17 (br s, H-8 and H-10), 3.22 (s, H-
20), 3.95 (s, OMe), 4.27 (d, J = 2.5 Hz, H-14), 4.72 (br s, H-16a,b), 4.74 (d,
J = 12.5 Hz, H-20a), 4.80 (d, J = 12.5 Hz, H-20b), 6.02 (br s, H-7), 6.49 (s, OH), ca.
7.25 (m, 20-Ar), ca. 7.35 (m, 20-Ar), 7.47 (s, H-1), 7.51 (d, J = 2 Hz, H-400), 8.02 (d,
J = 2.1 Hz, H-800); 13C NMR (CDCl3, 75.5 MHz): d 10.4 (q, C-19), 19.0 (q, C-17),
20.0 (q, C-18), 33.3 (d, C-11), 36.0 (t, C-12), 39.5 (d, C-8), 41.2 (t, C-5), 41.1 (d,
C-20), 56.3 (d, C-10), 56.2 (q, OMe), 73.4 (s, C-4), 74.9 (t, C-20), 80.8 (d, C-14),
81.3 (s, C-9), 84.7 (s, C-13), 104.9 (d, C-400), 111.0 (t, C-16), 118.0 (s, C-10), 127.9
(d, C-7), 128.7 (d, C-50/70), 130.4 (d, C-60), 131.0 (d, C-40/80), 135.2 (s, C-6), 136.8
(s, C-2), 144.4 (s, C-600), 146.6 (s, C-300), 146.8 (s, C-15), 154.1 (s, C-100), 158.5 (d,
C-1), 208.4 (s, C-3); CI-MS m/z 757 (M+H+) (C36H37IO10+H+).
16. Appendino, G.; Daddario, N.; Minassi, A.; Schiano Morello, A.; De Petrocellis, L.;
Di Marzo, V. J. Med. Chem. 2005, 48, 4663.
17. Lu, Y.; Shi, T.; Wang, Y.; Yang, H.; Xiuhua Yan, X.; Xiaoming Luo, X.; Hualiang
Jiang, H.; Zhu, W. J. Med. Chem. 2009, 52, 2854.
18. Aakeröy, C. B.; Fasulo, M.; Schultheiss, N.; Desper, J.; Moore, C. J. Am. Chem. Soc.
2007, 129, 13772.
References and notes
1. Isolation from Euphorbia resinifera Berg: (a) Hergenhahn, M.; Adolf, W.; Hecker,
E. Tetrahedron Lett. 1975, 19, 1595; Isolation from E. poissonii Pax: (b) Schmidt,
R. J.; Evans, F. J. J. Pharm. Pharmacol. 1975, 27, 50P.