Potent and selective structure-based dibenzofuran inhibitors of transthyretin amyloidogenesis: Kinetic stabilization of the native state

Article abstract of DOI:10.1021/ja044351f

Transthyretin (TTR) amyloidogenesis requires rate-limiting tetramer dissociation and partial monomer denaturation to produce a misassembly competent species. This process has been followed by turbidity to identify transthyretin amyloidogenesis inhibitors including dibenzofuran-4,6-dicarboxylic acid (1). An X-ray cocrystal structure of TTR-1₂ reveals that it only utilizes the outer portion of the two thyroxine binding pockets to bind to and inhibit TTR amyloidogenesis. Herein, structure-based design was employed to append aryl substituents at C1 of the dibenzofuran ring to complement the unused inner portion of the thyroxine binding pockets. Twenty-eight amyloidogenesis inhibitors of increased potency and dramatically increased plasma TTR binding selectivity resulted. These function by imposing kinetic stabilization on the native tetrameric structure of TTR, creating a barrier that is insurmountable under physiological conditions. Since kinetic stabilization of the TTR native state by interallelic trans suppression is known to ameliorate disease, there is reason to be optimistic that the dibenzofuran-based inhibitors will do the same. Preventing the onset of amyloidogenesis is the most conservative strategy to intervene clinically, as it remains unclear which of the TTR misassembly intermediates results in toxicity. The exceptional binding selectivity enables these inhibitors to occupy the thyroxine binding site(s) in a complex biological fluid such as blood plasma, required for inhibition of amyloidogenesis in humans. It is now established that the dibenzofuran-based amyloidogenesis inhibitors have high selectivity, affinity, and efficacy and are thus excellent candidates for further pharmacologic evaluation.

Full text of DOI:10.1021/ja044351f

Published on Web 04/15/2005
Potent and Selective Structure-Based Dibenzofuran Inhibitors
of Transthyretin Amyloidogenesis: Kinetic Stabilization of the
Native State
H. Michael Petrassi, Steven M. Johnson, Hans E. Purkey, Kyle P. Chiang,
Traci Walkup, Xin Jiang, Evan T. Powers, and Jeffery W. Kelly*
Contribution from the Department of Chemistry and The Skaggs Institute of
Chemical Biology, The Scripps Research Institute, 10550 North Torrey Pines Road,
BCC265, La Jolla, California 92037
Received September 16, 2004; E-mail: jkelly@scripps.edu.
Abstract: Transthyretin (TTR) amyloidogenesis requires rate-limiting tetramer dissociation and partial
monomer denaturation to produce a misassembly competent species. This process has been followed by
turbidity to identify transthyretin amyloidogenesis inhibitors including dibenzofuran-4,6-dicarboxylic acid
(1). An X-ray cocrystal structure of TTR‚12 reveals that it only utilizes the outer portion of the two thyroxine
binding pockets to bind to and inhibit TTR amyloidogenesis. Herein, structure-based design was employed
to append aryl substituents at C1 of the dibenzofuran ring to complement the unused inner portion of the
thyroxine binding pockets. Twenty-eight amyloidogenesis inhibitors of increased potency and dramatically
increased plasma TTR binding selectivity resulted. These function by imposing kinetic stabilization on the
native tetrameric structure of TTR, creating a barrier that is insurmountable under physiological conditions.
Since kinetic stabilization of the TTR native state by interallelic trans suppression is known to ameliorate
disease, there is reason to be optimistic that the dibenzofuran-based inhibitors will do the same. Preventing
the onset of amyloidogenesis is the most conservative strategy to intervene clinically, as it remains unclear
which of the TTR misassembly intermediates results in toxicity. The exceptional binding selectivity enables
these inhibitors to occupy the thyroxine binding site(s) in a complex biological fluid such as blood plasma,
required for inhibition of amyloidogenesis in humans. It is now established that the dibenzofuran-based
amyloidogenesis inhibitors have high selectivity, affinity, and efficacy and are thus excellent candidates for
further pharmacologic evaluation.
Introduction
There is no effective treatment available today for SSA or
FAC. FAP has been successfully treated by liver transplantation,
Amyloid formation from numerous, nearly identical tran-
sthyretin (TTR) sequences appears to cause the organ dysfunc-
tion and/or neurodegeneration characteristic of several amyloid
diseases.1-3 In senile systemic amyloidosis (SSA), the amyloid
and related aggregates derived from wild-type (WT) TTR lead
to cardiac dysfunction, affecting as much as 10% of the human
population over age 80.2 Familial amyloid cardiomyopathy
(FAC) is associated with V122I TTR amyloidogenesis in 3-4%
of African Americans, ultimately leading to congestive heart
failure.1,4 In familial amyloid polyneuropathy (FAP), one of over
100 TTR variants (e.g., V30M) primarily deposits in and leads
to dysfunction of the peripheral nervous system.5-7 Recently, a
few TTR variants that are primarily associated with central
nervous system degeneration have also been discovered.8-10
which is a crude form of gene therapy wherein the mutant gene
is replaced by a wild-type TTR gene in the liver, which appears
to be the main secretory organ for plasma TTR.11-14 However,
liver transplantation requires invasive surgery, an appropriate
donor, and the need for sustained immunosuppression; in
addition, this therapy is imperfect because WT-TTR continues
to deposit in the heart.15,16 Moreover, recent evidence suggests
that this procedure is more effective for V30M heterozygotes
(7) Suhr, O. B.; Svendsen, I. H.; Andersson, R.; Danielsson, A.; Holmgren,
G. J. Int. Med. 2003, 254, 225-235.
(8) Hammarstrom, P.; Sekijima, Y.; White, J. P.; Wiseman, R. L.; Lim, A.;
Costello, C. E.; Altland, K.; Garzuly, F.; Budka, H.; Kelly, J. W.
Biochemistry 2003, 42, 6656-6663.
(9) Sekijima, Y.; Hammarstrom, P.; Matsumura, M.; Shimizu, Y.; Iwata, M.;
Tokuda, T.; Ikeda, S.-I.; Kelly, J. W. Lab. InVest. 2003, 83, 409-417.
(10) Vidal, R.; Garzuly, F.; Budka, H.; Lalowski, M.; Linke, R. P.; Brittig, F.;
Frangione, B.; Wisniewski, T. Am. J. Pathol. 1996, 148, 361-366.
(11) Herlenius, G.; Wilczek, H. E.; Larsson, M.; Ericzon, B.-G. Transplantation
2004, 77, 64-71.
(12) Holmgren, G.; Ericzon, B. G.; Groth, C. G.; Steen, L.; Suhr, O. Clinical
improvement and amyloid regression after liver transplantation in hereditary
transthyretin amyloidosis; Department of Clinical Genetics, University
Hospital, Umea, Sweden, United Kingdom, 1993; pp 1113-1116.
(13) Jonsen, E.; Suhr, O. B.; Tashima, K.; Athlin, E. Amyloid 2001, 8, 52-57.
(14) Suhr, O. B.; Ericzon, B.-G.; Friman, S. LiVer transplantation: official
publication of the American Association for the Study of LiVer Diseases
and the International LiVer Transplantation Society 2002, 8, 787-794.
(1) Jacobson, D. R.; Pastore, R. D.; Yaghoubian, R.; Kane, I.; Gallo, G. Buck,
F. S.; Buxbaum, J. N. New Eng. J. Med. 1997, 336, 466-473.
(2) Westermark, P.; Sletten, K.; Johansson, B.; Cornwell, G. G., III. Proc. Natl.
Acad. Sci. U.S.A. 1990, 87, 2843-2845.
(3) Saraiva, M. J.; Sousa, M. M.; Cardoso, I.; Fernandes, R. J. Mol. Neurosci.
2004, 23, 35-40.
(4) Afolabi, I.; Hamidi, A. K.; Nakamura, M.; Jacobs, P.; Hendrie, H.; Benson,
M. D. Amyloid 2000, 7, 121-125.
(5) Coelho, T. Curr. Opin. Neurology 1996, 9, 355-359.
(6) Saraiva, M. J. M. Human Mutat. 1995, 5, 191-196.
9
6662
J. AM. CHEM. SOC. 2005, 127, 6662-6671
10.1021/ja044351f CCC: $30.25 © 2005 American Chemical Society

Dibenzofuran Inhibitors of TTR Amyloidogenesis
A R T I C L E S
than for carriers of other FAP mutations for reasons that remain
unclear but that may relate to synthesis of TTR by other organs
and tissues in the periphery.17
barrier for tetramer dissociation, making it insurmountable under
physiological conditions. This is highly desirable as it prevents
the process of amyloidogenesis from commencing.35 Greater
than 99% of the TTR thyroxine binding sites are unoccupied
because TTR is the tertiary carrier, providing the opportunity
to use ligands that selectively bind at one or both of these sites
to stabilize TTR in blood. On the basis of the mechanistic
analogies with interallelic trans suppression, which is known
to ameliorate human disease, there is good reason to believe
that a small molecule that selectively binds the native tetrameric
state of TTR and imparts kinetic stabilization will be effective
at preventing amyloidosis in humans.31
Several structurally distinct classes of small molecule TTR
amyloidogenesis inhibitors have been discovered, of which
dibenzofuran-4,6-dicarboxylic acid (1) is particularly interesting,
Figure 1.31,33,34,36-46 This inhibitor (1, 7.2 µM) decreases the
extent of WT-TTR (3.6 µM) amyloid formation (pH 4.4) by
90% over a time course of 72 h. The X-ray cocrystal structure
of TTR‚12 reveals that 1 exerts its impressive activity by binding
exclusively to the outer portion of each thyroxine binding site.33
This provides the opportunity to modify 1 with substructures
that project into the inner cavity of the thyroxine binding pocket,
which should increase their affinity and, most importantly, their
selectivity for binding to TTR in human blood.
Herein, we utilize structure-based design principles to append
aromatic substituents onto the dibenzofuran-4,6-dicarboxylic
acid core structure at the C1 position using three different types
of linkages, affording exceptional amyloidogenesis inhibitors.
These display increased affinity and greatly increased binding
selectivity to TTR over that of all the other plasma proteins,
relative to lead compound 1.47 Moreover, we have demonstrated
that these compounds function by imposing kinetic stabilization
on the TTR tetramer. Hence, there is reason to be optimistic
that they will be clinically useful, based on the analogy with
interallelic trans suppression, provided they exhibit suitable
pharmacology and toxicology profiles, the analysis of which is
beyond the scope of this contribution.31,35,43
Transthyretin serves as a transporter of both thyroid hormone
(i.e., thyroxine, T4) and holo-retinol binding protein in the blood
and cerebrospinal fluid.18-21 The extent of TTR aggregation does
not appear to be substantial enough to significantly compromise
normal function; however, the structure of this protein is
integrally associated with its pathology. TTR is a homotetramer
composed of 127-residue â-sheet-rich subunits.18 Amyloid
formation occurs by rate-limiting tetramer dissociation, followed
by partial monomer denaturation, rendering the protein misas-
sembly competent.22-28 Numerous aggregates are formed under
partial denaturing conditions including amorphous aggregates,
soluble cross-â-sheet aggregates and insoluble amyloid fibrils,
of which the nonfibrous deposits appear to mediate toxicity.28,29
Interallelic trans suppression of FAP has been demonstrated
in a Portuguese family who are known to have the highly
penetrant V30M disease-associated mutation, yet these individu-
als do not present with amyloidosis because they have a T119M
mutation on their other TTR allele.30,31 Trans suppression is
accomplished by the incorporation of T119M subunits into TTR
tetramers otherwise composed of V30M subunits, imparting
kinetic stabilization on the native tetrameric state of TTR,
preventing amyloidogenesis.31,32 Incorporation of T119M sub-
units into the tetramers of these compound heterozygotic
individuals raises the activation barrier for tetramer dissociation,
dramatically slowing the rate-limiting step of TTR amyloido-
genesis and thereby preventing the onset of FAP.31,32
The transthyretin homotetramer has 222-symmetry and
features a pair of funnel-shaped thyroid hormone (thyroxine)
binding sites bisected by the crystallographic C2 axis.18,33 Ligand
binding to one or both of the TTR thyroxine binding sites
imparts kinetic stabilization on the TTR tetramer, a feat that is
accomplished by selectively stabilizing the ground state over
the dissociative transition state under amyloidogenic condi-
tions.31,34 High-affinity native state binders raise the kinetic
Design and Synthesis
(15) Hornsten, R.; Wiklund, U.; Olofsson, B.-O.; Jensen, S. M.; Suhr, O. B.
Transplantation 2004, 78, 112-116.
Figure 1A depicts the two symmetry-equivalent binding
modes of 1 (green and yellow) within one of the TTR thyroxine
(16) Dubrey, S. W.; Davidoff, R.; Skinner, M.; Bergethon, P.; Lewis, D.; Falk,
R. H. Transplantation 1997, 64, 74-80.
(17) Pfeffer, B. A.; Bacerra, S. P.; Borst, D. E.; Wong, P. Mol. Vision 2004,
10, 23-30.
(34) Razavi, H.; Palaninathan, S. K.; Powers, E. T.; Wiseman, R. L.; Purkey,
H. E.; Mohamedmohaideen, N. N.; Deechonkit, S.; Chiang, K. P.; Dendle,
M. T. A.; Sacchettini, J. C.; Kelly, J. W. Angew. Chem. 2003, 42, 2758-
2761.
(18) Hornberg, A.; Eneqvist, T.; Olofsson, A.; Lundgren, E.; Sauer-Eriksson,
A. E. J. Mol. Biol. 2000, 302, 649-669.
(19) Blake, C. C. F.; Geisow, M. J.; Oatley, S. J.; Rerat, B.; Rerat, C. J. Mol.
Biol. 1978, 121, 339-356.
(35) Cohen, F. E.; Kelly, J. W. Nature 2003, 426, 905-909.
(36) Miroy, G. J.; Lai, Z.; Lashuel, H. A.; Peterson, S. A.; Strang, C.; Kelly, J.
W. Proc. Natl. Acad. Sci. U.S.A. 1996, 93, 15051-15056.
(37) Peterson, S. A.; Klabunde, T.; Lashuel, H.; Purkey, H.; Sacchettini, J. C.;
Kelly, J. W. Proc. Natl. Acad. Sci. U.S.A. 1998, 95, 12956-12960.
(38) Baures, P. W.; Peterson, S. A.; Kelly, J. W. Bioorg. Med. Chem. 1998, 6,
1389-1401.
(20) Monaco, H. L.; Rizzi, M.; Coda, A. Science 1995, 268, 1039-1041.
(21) Nilsson, S. F.; Rask, L.; Peterson, P. A. J. Biol. Chem. 1975, 250, 8554-
8563.
(22) Colon, W.; Kelly, J. W. Biochemistry 1992, 31, 8654-8660.
(23) Kelly, J. W. Curr. Opin. Struct. Biol. 1996, 6, 11-17.
(24) Lai, Z.; Colon, W.; Kelly, J. W. Biochemistry 1996, 35, 6470-6482.
(25) Lashuel, H. A.; Wurth, C.; Woo, L.; Kelly, J. W. Biochemistry 1999, 38,
13560-13573.
(39) Baures, P. W.; Oza, V. B.; Peterson, S. A.; Kelly, J. W. Bioorg. Med.
Chem. 1999, 7, 1339-1347.
(26) Liu, K.; Cho, H. S.; Lashuel, H. A.; Kelly, J. W.; Wemmer, D. E. Nat.
Struct. Biol. 2000, 7, 754-757.
(40) Petrassi, H. M.; Klabunde, T.; Sacchettini, J.; Kelly, J. W. J. Am. Chem.
Soc. 2000, 122, 2178-2192.
(27) McCutchen, S. L.; Lai, Z.; Miroy, G. J.; Kelly, J. W.; Colon, W.
Biochemistry 1995, 34, 13527-13536.
(41) McCammon, M. G.; Scott, D. J.; Keetch, C. A.; Greene, L. H.; Purkey, H.
E.; Petrassi, H. M.; Kelly, J. W.; Robinson, C. V. Structure 2002, 10, 851-
863.
(28) Reixach, N.; Deechongkit, S.; Jiang, X.; Kelly, J. W.; Buxbaum, J. N. Proc.
Natl. Acad. Sci. U.S.A. 2004, 101, 2817-2822.
(42) Oza, V. B.; Smith, C.; Raman, P.; Koepf, E. K.; Lashuel, H. A.; Petrassi,
H. M.; Chiang, K. P.; Powers, E. T.; Sacchettini, J.; Kelly, J. W. S J. Med.
Chem. 2002, 45, 321-332.
(29) Sousa, M. M.; Cardoso, I.; Fernandes, R.; Guimaraes, A.; Saraiva, M. J.
Am. J. Pathol. 2001, 159, 1993-2000.
(30) Coelho, T.; Carvalho, M.; Saraiva, M. J.; Alves, I. L.; Almeida, M. R.;
Costa, P. P. J. Rheumatol. 1993, 20, 179-179.
(43) Sacchettini, J. C.; Kelly, J. W. Nat. ReV. Drug DiscoV. 2002, 1, 267-275.
(44) Green, N. S.; Palaninathan, S. K.; Sacchettini, J. C.; Kelly, J. W. J. Am.
Chem. Soc. 2003, 125, 13404-13414.
(31) Hammarstrom, P.; Wiseman, R. L.; Powers, E. T.; Kelly, J. W. Science
2003, 299, 713-716.
(45) Adamski-Werner, S. L.; Palaninathan, S. K.; Sacchettini, J. C.; Kelly, J.
W. J. Med. Chem. 2004, 47, 355-374.
(32) Hammarstrom, P.; Schneider, F.; Kelly, J. W. Science 2001, 293, 2459-
2462.
(46) Miller, S. R.; Sekijima, Y.; Kelly, J. W. Lab. InVest. 2004, 84, 545-552.
(47) Purkey, H. E.; Dorrell, M. I.; Kelly, J. W. Proc. Natl. Acad. Sci. U.S.A.
2001, 98, 5566-5571.
(33) Klabunde, T.; Petrassi, H. M.; Oza, V. B.; Raman, P.; Kelly, J. W.;
Sacchettini, J. C. Nat. Struct. Biol. 2000, 7, 312-321.
9
J. AM. CHEM. SOC. VOL. 127, NO. 18, 2005 6663

A R T I C L E S
Petrassi et al.
a heteroatom (N or O) or directly via a Caryl-Caryl bond (not
shown). Aromatic substituents (Figure 2) were chosen to interact
with either the halogen binding pockets or hydrogen bonding
substructures within the inner cavity based on the envisioned
orientation of the phenyl ring in the inner binding cavity and
previous SAR data from other chemical series thought to
position their aryl rings similarly.31,33,34,36-46
The synthesis of C1-substituted dibenzofuran-based inhibitors
commenced with the radical phenolic homocoupling of com-
mercially available 2,4-di-tert-butyl-6-bromophenol (2) to afford
the dibenzofuran derivative 3 using potassium hexacyanoferrate
(III) as previously reported (Scheme 1).48 This tetra-tert-butyl
dibenzofuran derivative was subjected to transalkylation in
toluene to form 1-hydroxydibenzofuran (4) in 33% overall yield
from 2 (alternative strategies for the synthesis of this intermedi-
ate have appeared).49,50 After protection of the phenol, silyl ether
5 was selectively ortho metalated at the 4- and 6-positions with
sec-butyllithium (the triisopropylsilyl group on the 1-oxygen
precludes it from acting as a metalation director).51 The dianion
was quenched with gaseous CO2 and esterified to afford 6, which
was then deprotected with TBAF and converted to triflate 8 in
high overall yield.
Selected anilines were coupled to triflate 8 using a palladium-
mediated N-arylation reaction developed by Buchwald and
Hartwig to afford dibenzofuran-based biarylamine analogues
9-23 (Scheme 2).52,53 To append an aryl ether to the C1-position
of dibenzofuran, phenol 7 and several phenylboronic acids were
cross-coupled using the copper-mediated biaryl ether coup-
ling methodology of Chan and Evans, affording compounds
39-43.54,55 Compound 8 was also subjected to Suzuki coupling
conditions in the presence of several phenylboronic acids to
afford the dibenzofuran-based biaryl analogues 49-59.56
Saponification of the methyl esters in these precursors afforded
the desired dibenzofuran-4,6-dicarboxylic acid amines (24-38),
ethers (44-48), and biaryls (60-70) to be evaluated as potential
TTR amyloidogenesis inhibitors.
Figure 1. (A) X-ray crystallographic structure of TTR‚12 depicting one of
the ligand binding sites (figure adapted from reference 33). The residues
lining the binding site are displayed as stick models (oxygen in red, nitrogen
in blue, and carbon in gray), with the protein’s Connolly surface depicted
in gray. Compound 1 is shown in both of its C2 symmetry equivalent binding
modes (yellow and green). The binding channel has three sets of depressions
referred to as the halogen binding pockets (HBPs) because they interact
with the iodines of thyroxine. Compound 1 occupies only the outer portion
of the binding pocket and fills both HBP1 and 1′. The carboxylic acids of
Results
Two of the most important characteristics of an effective
small molecule amyloidogenesis inhibitor are the ability to bind
with high affinity and selectively to TTR in blood and to
stabilize the native tetrameric structure of TTR.31,34,47 Inhibi-
tion efficacy (compounds 24-38, 44-48, and 60-70) was
first evaluated using recombinant TTR in a partially denaturing
buffer that promotes amyloidogenesis (pH 4.4, 37 °C). As a
follow up, the ability of effective inhibitors to bind to TTR
selectively over all the other proteins in human plasma was
assessed.
+
1 are in proximity to the ꢀ-NH3 of K15 and K15′. (B) Line drawing
representation of the design of the C1-substituted-dibenzofuran-4,6-dicar-
boxylic acids placed in the thyroxine binding pocket where X represents
either an NH, O, or direct Caryl-Caryl linkage (not shown). R represents the
substituents of the aryl ring envisioned to complement the inner binding
cavity of TTR.
binding sites, the surface of which is outlined in gray.33 The
carboxylates at the 4 and 6 positions make electrostatic
+
interactions with the ꢀ-NH3 groups of Lys 15 and 15′ at the
entrance to the thyroxine binding site. Removal of one of the
carboxylates renders dibenzofuran much less active, as does
varying the carboxylate spacing from the aromatic ring in most
cases (Figure S1). Inspection of the TTR‚12 crystal structure in
Figure 1A reveals that the dibenzofuran ring nicely comple-
ments the shape and hydrophobicity of the outer portion of the
thyroxine binding cavity and that there is a large amount of
unoccupied volume in the inner portion of the thyroxine binding
site with 1 bound. Based on this structure, it is clear that a
substituent, such as an aryl ring, could be projected into the
inner portion of the binding site by attaching it to the C1 position
of the dibenzofuran ring of 1. As shown in Figure 1B, such a
substituent could be linked to a dibenzofuran scaffold through
Evaluating the Dibenzofuran-Based Compounds As Amy-
loidogenesis Inhibitors. TTR amyloid inhibition efficacy was
probed using a stagnant fibril formation assay described
(48) Tashiro, M.; Yoshiya, H.; Fukata, G. Synthesis 1980, 6, 495-496.
(49) Labiad, B.; Villemin, D. Synthesis 1989, 143-144.
(50) Lee, Y. R.; Suk, J. Y.; Kim, B. S. Org. Lett. 2000, 2, 1387-1389.
(51) Snieckus, V. Chem. ReV. 1990, 90, 879-933.
(52) Louie, J. D.; Michael S.; Hamann, B. C.; Hartwig, J. F. J. Org. Chem.
1997, 62, 1268-1273.
(53) Wolfe, J. P. B. Tetrahedron Lett. 1997, 38, 6359-6362.
(54) Chan, D. M. T.; Monaco, K. L.; Wang, R.-P.; Winters, M. P. Tetrahedron
Lett. 1998, 39, 2933-2936.
(55) Evans, D. A.; Katz, J. L.; West, T. R. Tetrahedron Lett. 1998, 39, 2937-
2940.
(56) Suzuki, A. Mod. Arene Chem. 2002, 53-106.
9
6664 J. AM. CHEM. SOC. VOL. 127, NO. 18, 2005

Dibenzofuran Inhibitors of TTR Amyloidogenesis
A R T I C L E S
Figure 2. (A) Summary of dibenzofuran-based amyloid inhibition activity (3.6 µM) against WT-TTR (3.6 µM) fibril formation (pH 4.4, 72 h, 37 °C) in
vitro and binding stoichiometry of these inhibitors to TTR in human blood plasma. % Fibril formation (f.f.) values in the middle column represent the extent
of f.f., and thus inhibitor efficacy, relative to WT-TTR f.f. in the absence of inhibitor (assigned to be 100%). Complete inhibition is equivalent to 0% f.f.
Measurement error is (5%. The right column depicts the observed stoichiometry of inhibitor (dosed at 10.8 µM, ∼2-3× the concentration of plasma TTR)
bound to TTR in blood plasma as determined using the antibody capture method as described herein and elsewhere.47 Measurement error is (0.15. (B)
Inhibitor binding stoichiometry to TTR in plasma plotted vs fibril formation inhibition efficacy in vitro (omitting data for inhibitor 34, which was not
analyzed for binding selectivity due to poor f.f. inhibition). The area shaded gray corresponds to our definition of high activity and high selectivity (<40%
fibril formation and a binding stoichiometry >1). Biaryls are identified as b, biarylamines as 2, and biaryl ethers as [; the lead compound 1 ) 0.
previously, wherein partial denaturation was triggered by
acidification (pH 4.4, 37 °C).22 Briefly, a test compound (7.2
or 3.6 µM) is incubated with TTR (3.6 µM) for 30 min in pH
7 buffer. Amyloidogenesis is then initiated by lowering the
pH to 4.4, where maximal fibril formation is observed with
WT-TTR after 72 h (37 °C). The turbidity in the presence of a
potential inhibitor (Ttest) is compared to that of a solution lacking
a test compound (Tcontrol). Exceptional inhibitors yield 0% fibril
formation, whereas compounds not functioning as an inhibitor
afford 100% fibril formation. From experience we know that
excellent inhibitors allow <10% fibril formation at a small
molecule concentration of 7.2 µM and <40% fibril formation
at a concentration equal to that of WT-TTR (3.6 µM).31,33,34,36-46
Of the 31 compounds evaluated, all but one (34) completely
inhibit fibril formation at a concentration twice that of TTR
(7.2 µM inhibitor), Figure S2 (at 7.2 µM there is enough test
compound added to occupy both of the binding sites of TTR
(TTR‚I2), provided their dissociation constants are both in the
low nanomolar range at pH 4.4). Small molecules typically bind
to TTR with negative cooperativity; hence, Kd1 and Kd2 are often
separated by 1 or 2 orders of magnitude. Therefore, when both
the ligand and TTR are at equal concentrations, a population
9
J. AM. CHEM. SOC. VOL. 127, NO. 18, 2005 6665

A R T I C L E S
Petrassi et al.
Scheme 1. Synthesis of 1-Hydroxydibenzofuran-4,6-dicarboxylate Dimethyl Ester and the Corresponding Triflatea
a Conditions: a) K3[Fe(CN)6], KOH, H2O, benzene; b) AlCl3, toluene, 33% for both steps; c) TIPSCl, DMAP, CH2Cl2, 77%; d) sec-BuLi, Et2O, -78 °C,
gaseous CO2, TMSCHN2, 43%; e) TBAF, THF, 97%; f) Tf2O, pyridine, 92%.
Scheme 2. Synthesis of 1-Phenyl-, Phenoxy-, and Phenylamine-dibenzofuran-4,6-dicarboxylate Dimethyl Esters and the Corresponding
Dicarboxylatesa
a Conditions: a) Pd2(DBA)3, (()binap, Cs2CO3, toluene 100 °C; b) LiOH‚H2O, THF/MeOH/H2O (3:1:1); c) CuII(OAc)2, pyridine, 4 Å molecular sieves,
CH2Cl2; d) Pd(PPh3)4, LiCl, aq. Na2CO3, toluene, MeOH, 80°C.
of TTR‚I, TTR‚I2, and unliganded TTR is observed, dictated
by the dissociation constants. It is now established by other
studies that inhibitor occupancy of only one of the two TTR
binding sites is sufficient to stabilize the entire tetramer against
amyloidogenesis.57 This is further supported by the observation
herein that 26 of these dibenzofurans are excellent TTR
amyloidogenesis inhibitors (<40% fibril formation) at a con-
centration equal to that of TTR (3.6 µM, Figure 2A). Repre-
sentative small molecules from all three series (25, 26, 27, 30,
45, 47, 62; 3.6 µM) were subjected to the acid-mediated
amyloidogenic conditions (pH 4.4, 37 °C, 72 h) in the absence
of TTR, revealing no measurable precipitation.
inhibitors (10.8 µM, ∼2-3× the natural concentration of TTR)
are incubated with human blood plasma for 24 h at 37 °C.
Quenched sepharose resin is then added to the plasma to remove
any small molecules that would bind to the resin as opposed to
TTR. TTR and any TTR-bound small molecule is then immu-
nocaptured using a polyclonal TTR antibody covalently attached
to sepharose resin. After washing the resin (5 × 10-min washes),
the antibody-TTR complex is dissociated at high pH and
analyzed by RP-HPLC. The relative stoichiometry between TTR
and inhibitor is then calculated from their HPLC peak areas
using standard curves. Wash-associated losses are typically
observed for inhibitors that have high dissociation rates;
therefore, this analysis can underestimate true binding stoichi-
ometry but gives faithful results for compounds exhibiting low
Evaluating the Plasma TTR Binding Selectivity of the
Dibenzofuran-Based Inhibitors. Inhibitor binding selectivity
to TTR in human blood plasma was assessed using a previously
established antibody capture method.47 In this evaluation,
(57) Wiseman, R. L.; Johnson, S. M.; Kelker, M. S.; Foss, T.; Wilson, I. A.;
Kelly, J. W. J. Am. Chem. Soc. 2005, 127, 5540-5551.
9
6666 J. AM. CHEM. SOC. VOL. 127, NO. 18, 2005

Dibenzofuran Inhibitors of TTR Amyloidogenesis
A R T I C L E S
dissociation constants and off-rates. Twenty-one inhibitors
exhibit a binding stoichiometry exceeding one (Figure 2A; two
being the maximal binding stoichiometry possible), 19 of which
exhibit <40% fibril formation at a concentration of 3.6 µM
(Figure 2B, shaded box). The high plasma TTR binding
selectivities observed are remarkable, considering that the parent
dibenzofuran-4,6-dicarboxylic acid (1) displays no binding
selectivity to TTR, demonstrating the importance of the C1 aryl
substituent in terms of endowing binding selectivity to TTR
over all the other plasma proteins.
Stabilization of the Tetrameric Quaternary Structure
under Amyloidogenic Conditions. To ensure that these C1-
arylated dibenzofurans inhibit TTR fibril formation by native
state stabilization (i.e., tetramer stabilization), we studied the
TTR quaternary structure by analytical ultracentrifugation after
a preincubation period of 72 h under amyloidogenic conditions
(pH 4.4, 37 °C). In the presence of 27 (7.2 µM), TTR (3.6 µM)
was found to have hydrodynamic molecular weights of 57.1 (
0.3 and 55.1 ( 0.4 kDa by sedimentation velocity (Figure 3A)
and equilibrium analytical ultracentrifugation (Figure 3B),
respectively, comparable to the expected molecular weight of
tetrameric TTR (55.0 kDa). In the absence of 27, TTR
aggregated into very high molecular weight oligomers that
sedimented rapidly in the ultracentrifugation experiment (data
not shown).
Do the Dibenzofuran-Based Inhibitors Impose Kinetic
Stabilization on TTR? The ability of these inhibitors to impose
kinetic stability on tetrameric TTR is best evaluated by assessing
the rate of TTR tetramer dissociation.31,58 Under acidic condi-
tions (pH 4.4, 37 °C) tetramer dissociation leads to amyloido-
genesis, whereas in the presence of chaotropes (6 M urea, 25
°C) tetramer dissociation leads to monomer unfolding. The
influence of inhibitors 25, 47, and 64, representing the three
structural classes of dibenzofuran-based inhibitors, on the rates
of tetramer dissociation under acid- and urea-mediated denatur-
ing conditions was probed. TTR amyloidogenesis mediated by
partial acidification is dramatically slowed in a dose-dependent
fashion relative to control (no inhibitor) by 25, 47, and 64
(Figure 4A). The rate of TTR tetramer dissociation in 6 M urea
is easily monitored by linking the slow quaternary structural
changes to rapid tertiary structural changes that are easily
monitored by spectroscopic methods.58 The rate of TTR tetramer
(3.6 µM) dissociation monitored by far-UV circular dichroism
(CD) is markedly slowed in a dose-dependent fashion in 6 M
urea by 25, 47, and 64 (Figure 4B). These results are con-
sistent with differential stabilization of the ground state vs
the dissociative transition state by the binding of 25, 47, and
64.
Figure 3. Sedimentation velocity (A) and equilibrium ultracentrifugation
studies (B) on TTR (3.6 µM) after being preincubated with 27 (7.2 µM, 30
min, pH 7) and after another incubation period where the pH was dropped
to 4.4 for 72 h at 37 °C, a time frame that results in near maximal amyloid
formation in the absence of inhibitor. (A) Velocity analysis: overlay of
data sets taken 15 min apart at 50,000 rpm. The data (symbols) fit to a
single ideal species model (solid line) with a MW of 57.1 ( 0.3 kDa. (B)
TTR equilibrium concentration gradient observed after a 24 h application
of centrifugal force to the sample employing at a speed of 17,000 rpm.
The data (O) fit to a single ideal species model (solid line) with a MW of
55.1 ( 0.4 kDa. The residuals, the difference between experimental and
fitted data, are shown in the inset.
activity relationships (SAR) are deducible from the 7.2 µM
inhibitor data. However, parts A and B of Figure 2 reveal a
range of inhibitor efficacy at an inhibitor concentration equal
to that of TTR (3.6 µM), allowing some SAR conclusions to
be drawn, limited by the 31 analogues prepared and the
experimental error. Notably, all the inhibitors (except 34) display
increased potency relative to that of the parent compound (1)
at 3.6 µM (Figure 2B). Most importantly, of the 30 inhibitors
exhibiting increased potency, all but two (31 and 35) exhibit
dramatically increased binding selectivity to TTR in human
blood plasma. The exceptional binding selectivity exhibited by
the C1-substituted inhibitors is ideal for inhibiting TTR amy-
loidogenesis in complex biological systems.
Discussion
All but one of the C1-aryl substituted dibenzofurans (7.2 µM)
are exceptional inhibitors of WT-TTR (3.6 µM) acid-mediated
fibril formation in vitro (pH 4.4, 37 °C), even those bearing
unsubstituted aryl rings (Figure S2). The only dibenzofuran-
based inhibitor that was not as effective against TTR amyloido-
genesis was 34, bearing potentially four negative charges.
Because all the compounds completely inhibited TTR fibril
formation (within the (5% experimental error), no structure-
Comparison of the four C1-aryl substitution patterns (H,
3-CF3, 3,5-F2, and 3,5-Cl2) found in all three inhibitor series
reveals that the inhibitors having their aryl rings directly linked
(58) Hammarstrom, P.; Jiang, X.; Hurshman, A. R.; Powers, E. T.; Kelly, J. W.
Proc. Natl. Acad. Sci. U.S.A. 2002, 99, 16427-16432.
9
J. AM. CHEM. SOC. VOL. 127, NO. 18, 2005 6667

A R T I C L E S
Petrassi et al.
results on simple biphenyl and biphenylamine inhibitors dem-
onstrate that it can be preferable to have the carboxyl-bearing
aryl ring in the inner binding cavity (enabling H-bonding to
S117 and T119).33,42,45
Appending aryl groups to the C1 position not only increases
inhibitor potency, but more importantly dramatically increases
plasma binding selectivity to TTR, presumably by increasing
binding affinity for TTR over the other plasma proteins. The
superior binding selectivity of the C1-aryl-substituted dibenzo-
furan-based inhibitors to TTR in plasma is clearly demonstrated
by the fact that about two-thirds of the compounds prepared
display a TTR binding stoichiometry >1. This is exceedingly
interesting as the screening hit 1, utilizing only the outer cavity
of TTR for binding, displays no measurable binding selectivity
to TTR in plasma. Previous experience with amyloidogenesis
inhibitors of diverse chemical structure reveals that very few
members display binding stoichiometries exceeding 1.31,33,34,36-46
The area of Figure 2B shaded in gray contains dibenzofuran-
based compounds that meet the criteria for high in vitro activity
(<40% fibril formation) and high binding selectivity (>1 equiv
bound to TTR in plasma). The most important point is that the
activity and binding selectivity of almost all of the dibenzofuran-
based inhibitors, especially those in the gray box (Figure 2B),
are sufficient to kinetically stabilize TTR in plasma should they
display desirable oral bioavailability, pharmacokinetic, and
toxicity profiles.
It is not surprising that inhibitor efficacy in vitro (3.6 µM)
and inhibitor binding selectivity (10.8 µM) to TTR in plasma
do not correlate (Figure 2B). Compounds that exhibit superior
binding selectivity to TTR over that of all of the other plasma
proteins should be excellent inhibitors of fibril formation.
However, the converse is not necessarily true: excellent
inhibitors need not display high TTR plasma binding selectivity.
Excellent inhibitors that display high TTR plasma binding
selectivity are the most useful compounds in humans because
these can selectively stabilize the TTR native state over the
dissociative transition state and impart kinetic stabilization on
TTR in a protein-rich biological fluid. Inhibitor binding con-
stants to TTR are important because the extent of kinetic stabili-
zation is proportional to the binding constants. However, focus-
ing on binding constants and potency in vitro can be misleading
because compounds can be excellent TTR amyloidogenesis in-
hibitors in vitro, but bind to other plasma proteins and therefore
be rendered useless in humans.47 Potent in vitro inhibitors not
displaying good binding selectivity to TTR likely interact with
other plasma proteins, such as albumin.47 Because the diben-
zofuran inhibitors display unprecedented binding selectivity and
inhibitor potency as a group relative to inhibitors characterized
heretofore,31,33,34,36-46 these are ideal for further pharmacological
evaluation. Because TTR is the tertiary carrier of T4, more than
99% of its binding sites are unoccupied; therefore, inhibitor
binding to TTR should not perturb T4 homeostasis.
Figure 4. (A) Timecourse analysis of WT-TTR (3.6 µM) fibril formation
mediated by partial acid denaturation (pH 4.4, 37 °C) in the absence (2)
and presence of inhibitors 25 (7.2 µM ) O; 3.6 µM ) b), 47 (7.2 µM )
0; 3.6 µM ) 9), and 64 (7.2 µM ) ]; 3.6 µM ) [), as measured by
turbidity at 500 nm. (B) Timecourse analysis of WT-TTR (3.6 µM) tetramer
dissociation (6.0 M urea, 25 °C) in the absence (2) and presence of inhibitors
25, 47, and 64 at 7.2 and 3.6 µM as defined in graph A. Slow tetramer
dissociation is not detectable by far-UV CD spectroscopy, but this process
is linked to rapid (∼500,000× faster) monomer denaturation as monitored
by loss of â-sheet content, easily followed by CD spectroscopy.
to C1 of the dibenzofuran skeleton, hereafter referred to as the
biaryls, display slightly increased inhibitor potency relative to
that of their biarylamine and biaryl ether counterparts (Figure
2A). This may be due to structural differences that enable the
rings to be oriented differently within the inner binding cavity;
however, we caution that this preference may not hold in a very
large analogue series. It could be argued with the same qualifiers
that the other two series produce the most selective TTR binders
in plasma; however, the SAR here is confounded by the fact
that as many as 100 proteins are competing for these inhibitors.
While it is not surprising that the most potent inhibitors have
2-F or 3,5-Cl2 substituents (36, 63, and 67), likely picking up
the halogen binding pockets in the thyroxine binding site, it is
somewhat surprising that the 3-CO2H-substituted aryl inhibitors
33 and 69 are among the most potent (although their plasma
binding selectivity to TTR is modest). Previous crystallographic
In summary, the C1-substituted dibenzofuran-based TTR
amyloidogenesis inhibitors are promising because of their
amyloid inhibition potency in vitro, their superb binding
selectivity to TTR in plasma, their slow TTR dissociation rates
(which must be the case to see high plasma selectivity by the
method utilized herein), their ability to impose kinetic stabiliza-
tion upon the TTR tetramer, their chemical stability in plasma,
and their chemical stability at low pH (making them excellent
9
6668 J. AM. CHEM. SOC. VOL. 127, NO. 18, 2005

Dibenzofuran Inhibitors of TTR Amyloidogenesis
A R T I C L E S
for all reverse-phase HPLC analyses. An acetonitrile/water/trifluoro-
acetic acid solvent system was used; solvent A in the proportions of
4.8%, 95%, and 0.2%, respectively, while solvent B was of 95%, 4.8%,
and 0.2%, respectively. Following 2 min of isocratic flow at 100% A,
a linear gradient of 0 to 100% B over 8 min was run at 1.5 mL/min.
All flash chromatography was accomplished using 230-400 mesh silica
gel 60 (EM Sciences). 1H- and 13C NMR spectra were recorded at 300,
400, 500, or 600 MHz on Bruker spectrometers. Chemical shifts are
reported in parts per million downfield from the internal standard (Me4-
Si, 0.0 ppm).
candidates for oral administration). There is reason to be
optimistic that these inhibitors have the potential to be generally
useful for the treatment of TTR amyloid diseases, including
SSA, FAP, and FAC, because kinetic stabilization of TTR is
known to ameliorate FAP.30,31,32
Experimental Section
The procedures used for bacterial expression of TTR,24 the stagnant
fibril formation assay,22,24 the blood plasma binding selectivity assay,47
and analytical ultracentrifugation59 have all been described in detail
previously.
(Dibenzofuran-1-yloxy)triisopropylsilane (5). To a dry 250-mL
round-bottom flask was added phenol 448 (492 mg, 2.67 mmol) and a
stir bar, and the flask was capped with a septum. CH2Cl2 (5 mL) was
added followed by DMAP (391 mg, 3.2 mmol) and triisopropylsilyl
chloride (800 µL, 3.73 mmol). The resulting colorless solution became
a white suspension overnight. The reaction was transferred to a 250-
mL separatory funnel and washed with H2O (3 × 10 mL). The aqueous
layers were combined and extracted with CH2Cl2 (3 × 30 mL). The
organic layers were combined, dried with MgSO4, and concentrated
under reduced pressure to afford a pale-yellow oil. The oil was purified
by flash chromatography over silica (100% hexanes) to afford 0.70 g
Time Course Analysis of WT-TTR Fibril Formation Inhibition
by Compounds 25, 47, and 64. Compounds 25, 47, and 64 were
dissolved in DMSO to provide 7.2 mM primary stock solutions (10×
stocks), from which 5- and 10-fold DMSO dilutions yielded 1.44 mM
(2×) and 0.72 mM (1×) secondary stock solutions, respectively. To
disposable cuvettes were added 495 µL of 0.4 mg/mL (7.2 µM)
WT-TTR solution (10 mM sodium phosphate, 100 mM KCl, and 1
mM EDTA, pH 7.2) and 5 µL of either the 1.44 or 0.72 mM inhibitor
secondary stock solutions; these were vortexed briefly and then
incubated for 30 min at 25 °C. The pH was then adjusted to 4.4 with
addition of 500 µL of acidic buffer (100 mM acetate, 100 mM KCl, 1
mM EDTA, pH 4.2), and the final 1 mL solutions were vortexed again
and incubated in the dark at 37 °C without agitation. At 0, 4, 8, 12, 25,
49, 74, 100, 122, 145, and 169 h time points after acidification, the
solutions were vortexed, and the turbidity at 500 nm was measured.
Control samples containing 5 µL of pure DMSO were prepared and
analyzed as above for comparison. Small molecule and TTR control
samples were prepared in groups of 10 to prevent disturbance of the
cuvettes during incubation. Samples were discarded after their turbidities
were measured.
Time Course Analysis of WT-TTR Tetramer Dissociation
Inhibition by Compounds 25, 47, and 64 Evaluated by Linked-
Monomer Unfolding in Urea. Compounds 25, 47, and 64 were
dissolved in DMSO to provide 10 mM primary stock solutions, from
which 10-fold EtOH dilutions yielded 1 mM secondary stock solutions.
To 2-mL Eppendorf tubes were added 200 µL of 1.0 mg/mL (18 µM)
WT-TTR solution (50 mM sodium phosphate, 100 mM KCl, and 1
mM EDTA, pH 7.2) and either 7.2 or 3.6 µL (2× and 1×, respectively)
of 1 mM inhibitor secondary stock solutions; these were vortexed briefly
and incubated for 15 min at 25 °C. The TTR‚inhibitor solutions (100
µL) were added to 900 µL of urea buffer (6.67 M urea, 50 mM sodium
phosphate, 100 mM KCl, 1 mM EDTA, pH 7.2), and the final 1 mL
solutions were vortexed again and incubated in the dark at 25 °C without
agitation. At 0, 4, 11, 24, 49, 73, 97, 122, 146, and 170 h time points
after mixing with urea, the CD spectra were measured between 218
and 215 nm, with scanning every 0.5 nm and averaging for 5 s. After
measurements were taken, samples were returned to their respective
Eppendorf tubes and incubation was continued. Control samples
containing 7.2 µL of 10% DMSO in EtOH were prepared and analyzed
as above for comparison. The CD amplitude values were averaged
between 215 and 218 nm to determine the extent of â-sheet loss
throughout the experiment. TTR tetramer dissociation is linked to the
rapid (∼500,000× faster) monomer denaturation as measured through
this â-sheet loss.58
1
(77%) of 5 as a colorless oil. H NMR (600 MHz, CDCl3) δ 1.22 (d,
J ) 7.6 Hz, 18 H), 1.48 (s, J ) 7.6 Hz, 3 H), 7.69 (t, J ) 8.6 Hz, 1
H), 7.19 (d, J ) 8.2 Hz, 1H), 7.27-7.33 (m, 1H), 7.36 (t, J ) 7.5 Hz,
1 H), 7.44 (t, J ) 8.2 Hz, 1 H), 7.56 (t, J ) 8.2 Hz, 1 H), 8.16-8.20
(s, 1H); 13C NMR (150 MHz, CDCl3) δ 13.33, 18.04, 104.31, 111.02,
111.77, 115.74, 122.59, 122.61, 123.76, 126.09, 127.47, 152.18, 155.44,
157.67; MALDI-FTMS 341.1932 m/z (M + H)+, C21H29O2Si requires
341.1931.
1-Triisopropylsilanyloxydibenzofuran-4,6-dicarboxylic Acid Di-
methyl Ester (6). Silyl ether 5 (654 mg, 1.92 mmol) was added to a
dry 50-mL round-bottom flask followed by Et2O (7.4 mL) and TMEDA
(0.87 mL, 5.77 mmol). The flask was cooled to -78 °C in an acetone/
CO2(s) bath for 10 min before adding sec-butyllithium (4.44 mL of a
1.3 M solution in cyclohexane, 5.77 mmol) over 10 min. The resulting
orange suspension was allowed to warm to room temperature and stirred
for 24 h. The flask was cooled again to -78 °C as described above,
and a 15 psi stream of CO2(g) was bubbled through the reaction
suspension (the CO2 was dried by passing it through a drying tube
containing activated silica). Following initial addition of CO2(g), the
cooling bath was removed, and the reaction was stirred for 30 min.
The reaction mixture was poured into a 1-L beaker containing ice/
H2O (50 mL). The solution was brought to pH 9 by the slow addition
of 0.05 M KOH and then cooled to 0 °C with an ice/H2O bath. The
solution was acidified to pH 2 with 0.5 M HCl causing a white solid
to precipitate. The aqueous suspension (pH 2) was transferred into a
1-L separatory funnel and extracted with EtOAc (5 × 50 mL). The
combined extracts were dried with MgSO4 and concentrated under
reduced pressure to afford the crude diacid as an oil. The 100-mL flask
containing the crude diacid was equipped with a stir bar, capped with
a septum, and evacuated. The flask was then back-filled with argon.
Anhydrous MeOH (2 mL) and ACS reagent grade benzene (8 mL)
were added via syringe. Trimethylsilyldiazomethane (TMSCHN2; 2.5
mL of a 2 M solution in hexanes, 5 mmol) was added slowly via syringe
through the septum. Upon completion of the TMSCHN2 addition the
reaction was stirred for 10 min and the solvent removed under reduced
pressure to afford a red oil. The residue was purified by flash
chromatography over silica (15% EtOAc in hexanes) to afford 0.36 g
(43%) of 6 as a white solid. 1H NMR (600 MHz, CDCl3) δ 1.18 (d, J
) 7.6 Hz, 18 H), 1.47 (s, J ) 7.6 Hz, 3 H), 4.07 (s, 3H), 4.10 (s, 3H),
7.64 (t, J ) 8.6 Hz, 1 H), 7.44 (d, J ) 7.8 Hz, 1H), 8.06 (d, J ) 8.6
Hz, 1H), 8.12 (d, J ) 1.4 Hz, 1H), 8.14 (d, J ) 1.4 Hz, 1H); 13C NMR
(150 MHz, CDCl3) δ 13.26, 17.87, 52.12, 52.38, 108.60, 112.40, 115.07,
115.60, 123.03, 124.48, 127.04, 128.91, 131.56, 147.92, 154.26, 156.47,
156.87, 165.06, 165.27; MALDI-FTMS 479.1874 m/z (M + Na)+,
C25H32O6SiNa requires 479.1860.
Inhibitor Synthesis. Reagents for chemical synthesis were purchased
from commercial suppliers and used without further purification unless
otherwise stated. Thin-layer chromatography on silica gel 60 F254 coated
aluminum plates (EM Sciences) or analytical reverse-phase high
performance liquid chromatography (HPLC) was used to monitor
reaction progress. HPLC was performed using a Waters 600E multi-
solvent delivery system employing a Waters 486 tunable absorbance
detector and a Waters 717 plus auto sampler. A C18 Western Analytical
column was used (model 033-715, 150 Å pore size, 3 µm particles)
(59) Lashuel, H. A.; Lai, Z.; Kelly, J. W. Biochemistry 1998, 37, 17851-17864.
9
J. AM. CHEM. SOC. VOL. 127, NO. 18, 2005 6669

A R T I C L E S
Petrassi et al.
1-Hydroxydibenzofuran-4, 6-dicarboxylic Acid Dimethyl Ester
(7). A dry 100-mL round-bottom flask was equipped with a stir bar,
charged with 6 (363 mg, 0.95 mmol), capped with a septum, evacuated,
and back-filled with argon. Anhydrous THF (6.3 mL) and tetrabutyl-
ammonium fluoride (1 M in THF, 1.2 mL, 1.19 mmol) were added to
the reaction by syringe. The reaction was stirred for 1 h at room
temperature and then poured into 30 mL of H2O in a 250-mL separatory
funnel. The aqueous layer was extracted with CHCl3 (4 × 20 mL).
The organic layers were combined, dried with MgSO4, and concentrated
under reduced pressure. The residue was purified by flash chromatog-
raphy over silica (30% EtOAc in hexanes) to afford 0.23 g (97%) of 7
as a white solid. 1H NMR (600 MHz, DMSO-d6) δ 4.06 (s, 3H), 4.10
(s, 3H), 6.79 (d, J ) 8.5 Hz, 1H), 7.43 (t, J ) 7.7 Hz, 1H), 8.00 (d, J
) 8.5 Hz, 1H); 13C NMR (125 MHz, CDCl3) δ 52.11, 52.48, 108.71,
112.61, 114.32, 115.55, 122.66, 125.48, 126.99, 129.55, 131.33, 147.77,
155.11, 156.45, 158.97, 165.01, 165.29; LC-MS m/z 301, C16H12O6
requires 301.
CDCl3) δ 52.05, 52.50, 107.37, 108.82, 112.23, 115.43, 121.30, 122.90,
124.07, 124.31, 125.33, 128.63, 129.73, 131.88, 140.28, 144.58, 154.40,
156.81, 165.16; MALDI-FTMS 375.1094 m/z (M•)+, C22H17NO5
requires 375.1106.
Representative Procedure for the Copper-Mediated Cross-
Coupling of Phenol 7 with Substituted Phenylboronic Acids To
Afford 1-Phenoxydibenzofurans 39-43. The biaryl ether coupling
was directly adapted from the procedures reported by Chan and
Evans. A 20-mL scintillation vial equipped with a magnetic stir bar
was charged with phenol 7 (150 mg, 0.50 mmol), copper (II)
acetate (91 mg, 0.5 mmol), freshly activated 4 Å molecular sieves (∼250
mg), and phenylboronic acid (180 mg, 1.5 mmol). Dichloromethane
(5 mL) was added followed by pyridine (201 µL, 2.5 mmol), resulting
in an aqua-colored suspension. The cap was very loosely applied such
that the reaction suspension was partly open to the atmosphere. The
reaction was monitored by TLC. After completion, the reaction mixture
was adsorbed onto ∼6 g of silica gel, adding silica gel to the reaction
mixture and then removing the solvent under reduced pressure.
Chromatography (30% EtOAc in hexanes) of the reaction mixture over
silica afforded biaryl ether 39 as a white solid (29 mg, 15%). Refer to
the Supporting Information for specific synthetic details and charac-
terization data for compounds 40-43 analogous to that reported for
39 below.
1-Trifluoromethanesulfonyloxydibenzofuran-4,6-dicarboxylic Acid
Dimethyl Ester (8). The triflation procedure previously described by
Stille was used to synthesize 8.60 Phenol 7 (120 mg, 0.4 mmol) was
added to a dry 10-mL round-bottom flask, which was then fitted with
a septum. The solvent, anhydrous pyridine (2 mL), was added by
syringe through the septum. The reaction mixture was cooled to 0 °C
with an ice/H2O bath. To initiate the reaction, trifluoromethanesulfonic
anhydride (81 µL, 12 mmol) was added by syringe through the septum.
The ice bath was removed, and the reaction was allowed to warm to
room temperature and stirred overnight. The reaction mixture was
poured into a 250-mL beaker containing 30 mL of an ice/H2O slurry
and transferred into a 125-mL separatory funnel. The aqueous layer
was extracted with Et2O (4 × 40 mL). The organic layers were
combined, washed with saturated CuSO4 (4 × 20 mL) and brine (2 ×
20 mL), dried over MgSO4; the Et2O was then removed under reduced
pressure to afford a slightly yellow solid. The solid was purified by
flash chromatography over silica (30% EtOAc in hexanes) to afford
1-Phenoxydibenzofuran-4,6-dicarboxylic Acid Dimethyl Ester
1
(39). H NMR (500 MHz, CDCl3) δ 4.09 (s, 3H), 4.13 (s, 3H), 6.73
(d, J ) 8.8 Hz, 1H), 7.25-7.31 (m, 3H), 7.42-7.50 (m, 3H), 8.07 (d,
J ) 8.8 Hz, 1H), 8.18 (dd, J ) 1.1, 7.7 Hz, 1H), 8.38 (dd, J ) 1.1, 7.7
Hz, 1H); 13C NMR (125 MHz, CDCl3) δ 52.29, 52.46, 109.42, 110.02,
114.93, 115.28, 120.52, 123.31, 123.87, 125.34, 127.79, 129.51, 130.26,
131.56, 154.56, 154.73, 156.75, 157.55, 164.88, 165.20; MALDI-FTMS
399.0825 m/z (M + Na)+, C22H16O6Na requires 399.0839.
Representative Procedure for the Palladium-Catalyzed Cross-
Coupling of Triflate 8 with Substituted Phenylboronic Acids. A
flame-dried 10 mm by 13 cm test tube, equipped with a stir bar and
capped with a septum, was charged with 8 (100 mg, 0.23 mmol),
Pd(PPh3)4 (14 mg, 0.01 mmol), LiCl (29 mg, 0.69 mmol), Na2CO3
(300 µL of a 2 M aqueous solution), and toluene (3 mL). Phenyl-
boronic acid (43 mg, 0.35 mmol) was dissolved in EtOH (0.5 mL) and
added to the reaction mixture. MeOH replaced EtOH in this procedure
for all other compounds because transesterification was observed;
therefore compound 49 was isolated as the diethyl ester and all other
compounds as dimethyl esters. After the reagents were added, the tube
was purged with argon and the reaction mixture heated to 100 °C for
12 h in an oil bath. The reaction mixture was then filtered through
Celite. The solvent was removed under reduced pressure from the
filtrate, and the resulting dark residue was purified by flash chroma-
tography over silica to afford biaryl 49 as a white solid (52 mg, 63%).
Refer to the Supporting Information for specific synthetic details and
characterization data for compounds 50-59 analogous to that reported
for 49 below.
1-Phenyldibenzofuran-4,6-dicarboxylic Acid Diethyl Ester (49).
1H NMR (400 MHz, CDCl3) δ 1.52-1.57 (m, 6H), 4.53-4.62 (m,
4H), 7.20 (t, J ) 7.9 Hz, 1H), 7.33 (d, J ) 7.9 Hz, 1H), 7.50-7.62
(m, 5H), 7.64 (dd, J ) 1.4, 7.8 Hz, 1H), 8.09 (dd, J ) 1.3, 7.7 Hz,
1H), 8.18 (d, J ) 7.9 Hz, 1H); 13C NMR (100 MHz, CDCl3) δ 14.38,
14.41, 61.38, 114.82, 115.98, 122.45, 124.29, 124.59, 126.70, 127.08,
128.62, 128.73, 128.78, 129.70, 138.71, 142.39, 164.58, 164.61;
MALDI-FTMS 411.1197 m/z (M + Na)+, C24H20O5Na requires
411.1203.
1
159 mg (92%) of 8 as a white solid. H NMR (600 MHz, CDCl3) δ
4.11 (s, 3H), 4.12 (s, 3H), 7.44 (d, J ) 8.7 Hz, 1H), 7.56 (t, J ) 7.8
Hz, 1H), 8.24 (d, J ) 8.7 Hz, 1H), 8.26 (d, J ) 7.8 Hz, 1H), 8.36 (d,
J ) 7.8 Hz, 1H); 13C NMR (150 MHz, CDCl3) δ 115.70, 116.19,
116.41, 122.09, 124.51, 127.63, 131.43, 131.71, 146.74, 155.35, 156.61,
164.22, 164.98. FAB-MS (NBA/NaI) m/z 433.0215 (M + H)+,
C17H12F3O8S requires 433.0205.
Representative Procedure for the Palladium-Catalyzed Cross-
Coupling of 8 with Substituted Anilines. The aryl coupling pro-
cedure reported by Buchwald and Hartwig was used to prepare
compounds 9-23. A flame-dried 10 mm by 13 cm borosilicate test
tube, equipped with a stir bar and capped with a septum, was charged
with 8 (140 mg, 0.324 mmol), palladium dibenzylidene acetone,
Pd2(dba)3 (15 mg, 0.016 mmol), (()-binap (15 mg, 0.024 mmol),
Cs2CO3 (147 mg, 0.456 mmol), and aniline (32 µL, 0.356 mmol). Upon
addition of all reagents the tube was purged with argon for 10 min.
Anhydrous toluene (2.4 mL) was then added through the septum, and
the reaction mixture was heated to 100 °C for 36 h in an oil bath. The
reaction mixture was filtered through Celite, and the solvent was
removed from the filtrate under reduced pressure. The resulting dark
oil was purified by flash chromatography over silica (30% EtOAc in
hexanes) to afford biarylamine 17 as a white solid (0.12 g, 68%). Refer
to the Supporting Information for specific synthetic details and
characterization data for compounds 10-23 analogous to that reported
for 9 below.
1-Phenylaminodibenzofuran-4,6-dicarboxylic Acid Dimethyl Es-
ter (9). 1H NMR (500 MHz, CDCl3) δ 4.06 (s, 3H), 4.11 (s, 3H), 6.49
(br s, 1H), 7.11 (d, J ) 8.6 Hz, 1H), 7.14-7.20 (m, 1H), 7.28-7.31
(m, 2H), 7.38-7.44 (m, 3H), 7.97 (dd, J ) 1.3, 7.7 Hz, 1H), 8.03 (d,
J ) 8.6 Hz, 1H), 8.12 (dd, J ) 1.3, 7.7 Hz, 1H); 13C NMR (125 MHz,
Representative Procedure for Ester Hydrolysis To Afford Final
Inhibitors 24-38, 44-48, and 60-70. Methyl ester 9 (25 mg, 0.067
mmol) was saponified in THF: MeOH: H2O (3:1:1, 1 mL) in a 20-
mL scintillation vial equipped with a stir bar. LiOH‚H2O (22 mg, 0.53
mmol) was added to the suspension, and the reaction was allowed to
stir until completion (typically 4 h) as determined by TLC or analytical
reverse-phase HPLC monitoring. The reaction mixture was diluted with
(60) Echavarren, A. M.; Stille, J. K. J. Am. Chem. Soc. 1987, 109, 5478-5486.
9
6670 J. AM. CHEM. SOC. VOL. 127, NO. 18, 2005

Dibenzofuran Inhibitors of TTR Amyloidogenesis
A R T I C L E S
brine (2 mL) and acidified to pH 2 with 1 M HCl (pH paper), resulting
in a biphasic solution. The upper layer (THF) was removed, and the
aqueous layer was extracted with THF (3 × 3 mL). The combined
organic layers were dried with MgSO4 and then concentrated under
reduced pressure to afford diacid 24 as a white solid (21 mg, 92%).
Refer to the Supporting Information for specific synthetic details and
characterization data for compounds 25-38, 44-48, and 60-70
analogous to the data reported for 24 below.
1-Phenylaminodibenzofuran-4,6-dicarboxylic Acid (24). 1H NMR
(600 MHz, THF-d8) δ 7.00-7.04 (m, 1H), 7.12 (d, J ) 8.6 Hz, 1H),
7.22 (d, J ) 7.7 Hz, 1H), 7.29-7.36 (m, 3H), 7.88 (s, br, 1H), 7.98 (d,
J ) 8.5 Hz, 1H), 8.03 (d, J ) 7.5 Hz, 1H), 8.15 (d, J ) 7.7 Hz, 1H);
13C NMR (150 MHz, THF-d8) δ 108.51, 110.04, 113.70, 116.30, 120.73,
122.36, 122.84, 124.78, 126.43, 128.80, 129.51, 131.95, 142.47, 144.86,
154.87, 157.36, 164.96, 165.24; MALDI-FTMS 347.0794 m/z (M•)+,
C20H13NO5 requires 347.0788.
Acknowledgment. We thank the National Institutes Health
(DK 46335), the W. M. Keck Foundation, the Skaggs Institute
for Chemical Biology, the Lita Annenberg Hazen Foundation,
and the ACS Organic Division, pre-doctoral fellowship (H.M.P.)
for financial support. The technical expertise of Dr. Hossein
Razavi and Michael Simpson is also greatly appreciated.
Supporting Information Available: Experimental data for
compounds formed from the three key reactions, fibril formation
results for the dibenzofuran inhibitors lacking C1 appended
groups, 7.2 µM fibril formation results for all inhibitors. This
material is available free of charge via the Internet at
http://pubs.acs.org.
JA044351F
9
J. AM. CHEM. SOC. VOL. 127, NO. 18, 2005 6671