Kinetic stabilization of an oligomeric protein by a single ligand binding event

Article abstract of DOI:10.1021/ja042929f

Protein native state stabilization imposed by small molecule binding is an attractive strategy to prevent the misfolding and misassembly processes associated with amyloid diseases. Transthyretin (TTR) amyloidogenesis requires rate-limiting tetramer dissociation before misassembly of a partially denatured monomer ensues. Selective stabilization of the native TTR tetramer over the dissociative transition state by small molecule binding to both thyroxine binding sites raises the kinetic barrier of tetramer dissociation, preventing amyloidogenesis. Assessing the amyloidogenicity of a TTR tetramer having only one amyloidogenesis inhibitor (I) bound is challenging because the two small molecule binding constants are generally not distinct enough to allow for the exclusive formation of TTR·I in solution to the exclusion of TTR·I₂ and unliganded TTR. Herein, we report a method to tether one fibril formation inhibitor to TTR by disulfide bond formation. Occupancy of only one of the two thyroxine binding sites is sufficient to inhibit tetramer dissociation in 6.0 M urea and amyloidogenesis under acidic conditions by imposing kinetic stabilization on the entire tetramer. The sufficiency of single occupancy for stabilizing the native state of TTR provides the incentive to search for compounds displaying striking negative binding cooperativity (e.g., K_d1 in nanomolar range and K_d2 in the micromolar to millimolar range), enabling lower doses of inhibitor to be employed in the clinic, mitigating potential side effects.

Full text of DOI:10.1021/ja042929f

Published on Web 03/23/2005
Kinetic Stabilization of an Oligomeric Protein by a Single
Ligand Binding Event
R. Luke Wiseman,† Steven M. Johnson,† Matthew S. Kelker,‡ Ted Foss,†
Ian A. Wilson,‡ and Jeffery W. Kelly*,†
Contribution from the Department of Chemistry, the Department of Molecular Biology, and the
Skaggs Institute for Chemical Biology, The Scripps Research Institute,
10550 North Torrey Pines Road, BCC 265, La Jolla, California 92037
Received November 23, 2004; E-mail: jkelly@scripps.edu
Abstract: Protein native state stabilization imposed by small molecule binding is an attractive strategy to
prevent the misfolding and misassembly processes associated with amyloid diseases. Transthyretin (TTR)
amyloidogenesis requires rate-limiting tetramer dissociation before misassembly of a partially denatured
monomer ensues. Selective stabilization of the native TTR tetramer over the dissociative transition state
by small molecule binding to both thyroxine binding sites raises the kinetic barrier of tetramer dissociation,
preventing amyloidogenesis. Assessing the amyloidogenicity of a TTR tetramer having only one amyloido-
genesis inhibitor (I) bound is challenging because the two small molecule binding constants are generally
not distinct enough to allow for the exclusive formation of TTR‚I in solution to the exclusion of TTR‚I2 and
unliganded TTR. Herein, we report a method to tether one fibril formation inhibitor to TTR by disulfide
bond formation. Occupancy of only one of the two thyroxine binding sites is sufficient to inhibit tetramer
dissociation in 6.0 M urea and amyloidogenesis under acidic conditions by imposing kinetic stabilization
on the entire tetramer. The sufficiency of single occupancy for stabilizing the native state of TTR provides
the incentive to search for compounds displaying striking negative binding cooperativity (e.g., Kd1 in
nanomolar range and Kd2 in the micromolar to millimolar range), enabling lower doses of inhibitor to be
employed in the clinic, mitigating potential side effects.
Introduction
formation of the intermediates that have been implicated as being
more neurotoxic than the fibrils themselves.13-17
Extracellular protein misfolding and misassembly is associ-
ated with a number of neurodegenerative diseases, including
Alzheimer’s Disease and the familial amyloidoses.1-3 Many of
these diseases involve the deposition of a polypeptide into
fibrous cross â-sheet aggregates, referred to as amyloid.1,4
Numerous therapeutic strategies have been developed to treat
amyloid diseases.5-7 In cases where proteolytic processing
generates the amyloidogenic fragment (e.g., Alzheimer’s Dis-
ease), small molecule protease inhibitors offer promise.8,9 In
cases where the full-length protein deposits as amyloid,
stabilization of the natively folded state by small molecule10 or
protein binding11,12 is a conservative approach because it avoids
Ligand binding to a protein’s native state can result in kinetic
and/or thermodynamic stabilization, depending on the binding
affinity to the native state relative to the misfolding transition
state. Selective stabilization of the native state over the
denaturation-associated transition state raises the kinetic barrier,
slowing the initial misfolding steps associated with amyloid
formation. If the kinetic barriers are surmountable, this mode
of stabilization also shifts the equilibrium toward the native state
under denaturing conditions. Inhibitors that equally stabilize the
native state and transition state(s) associated with denaturation
shift the equilibrium away from the amyloidogenic conforma-
tion(s), decreasing its concentration and therefore the rate of
amyloidogenesis owing to the concentration dependence of
misassembly.18
† Department of Chemistry. The Scripps Research Institute.
‡ Department of Molecular Biology, The Scripps Research Institute.
(1) Buxbaum, J. N. Curr. Opin. Rheumatol. 2004, 16, 67-75.
(2) Soto, C. Nat. ReV. Neurosci. 2003, 4, 49-60.
(3) Hardy, J.; Selkoe, D. J. Science 2002, 297, 353-356.
(4) Petkova, A. T.; Ishii, Y.; Balbach, J. J.; Antzutkin, O. N.; Leapman, R. D.;
Delaglio, F.; Tycko, R. Proc. Natl. Acad. Sci. U.S.A. 2002, 99, 16742-
16747.
(5) Cohen, F. E.; Kelly, J. W. Nature 2003, 426, 905-909.
(6) Ghosh, A. K.; Hong, L.; Tang, J. Curr. Med. Chem. 2002, 9, 1135-1144.
(7) Dominguez, D. I.; De Strooper, B. Trends Pharmacol. Sci. 2002, 23, 324-
330.
(8) Wolfe, M. S. Curr. Top. Med. Chem. 2002, 2, 371-383.
(9) Vassar, R. AdV. Drug DeliV. ReV. 2002, 54, 1589-1602.
(10) Hammarstrom, P.; Wiseman, R. L.; Powers, E. T.; Kelly, J. W. Science
2003, 299, 713-716.
(11) White, A. R.; Enever, P.; Tayebi, M.; Mushens, R.; Linehan, J.; Brandner,
S.; Anstee, D.; Collinge, J.; Hawke, S. Nature 2003, 422, 80-83.
(12) Peretz, D.; Williamson, R. A.; Kaneko, K.; Vergara, J.; Leclerc, E.; Schmitt-
Ulms, G.; Mehlhorn, I. R.; Legname, G.; Wormald, M. R.; Rudd, P. M.;
Dwek, R. A.; Burton, D. R.; Prusiner, S. B. Nature 2001, 412, 739-743.
(13) Reixach, N.; Deechongkit, S.; Jiang, X.; Kelly, J. W.; Buxbaum, J. N. Proc.
Natl. Acad. Sci. U.S.A. 2004, 101, 2817-2822.
(14) Lambert, M. P.; Barlow, A. K.; Chromy, B. A.; Edwards, C.; Freed, R.;
Liosatos, M.; Morgan, T. E.; Rozovsky, I.; Trommer, B.; Viola, K. L.;
Wals, P.; Zhang, C.; Finch, C. E.; Krafft, G. A.; Klein, W. L. Proc. Natl.
Acad. Sci. U.S.A. 1998, 95, 6448-6453.
(15) Walsh, D. M.; Klyubin, I.; Fadeeva, J. V.; Cullen, W. K.; Anwyl, R.; Wolfe,
M. S.; Rowan, M. J.; Selkoe, D. J. Nature 2002, 416, 535-539.
(16) Volles, M. J.; Lansbury, P. T., Jr. Biochemistry 2003, 42, 7871-7878.
(17) Sousa, M. M.; Cardoso, I.; Fernandes, R.; Guimaraes, A.; Saraiva, M. J.
Am. J. Pathol. 2001, 159, 1993-2000.
9
5540
J. AM. CHEM. SOC. 2005, 127, 5540-5551
10.1021/ja042929f CCC: $30.25 © 2005 American Chemical Society

Kinetic Stabilization of an Oligomeric Protein
A R T I C L E S
Table 1. Percentage of Unbound (T), Singly Bound (T‚I), and
Transthyretin (TTR) is a homotetrameric protein that trans-
ports the small molecule hormone thyroxine (T4) and holo-
retinol binding protein in the serum and cerebrospinal fluid
(CSF).19 Over 100 point mutations within the TTR sequence
destabilize its structure, predisposing individuals to familial
amyloid diseases, including familial amyloid polyneuropathy
(FAP), familial amyloid cardiomyopathy (FAC), and CNS
selective neurodegeneration.20-22 Wild-type TTR deposition also
results in the late onset disease senile systemic amyloidosis
(SSA), leading to cardiac dysfunction in 10% of individuals
over age 80.23
Doubly Bound (T‚I2) TTR Tetramers as a Function of Inhibitor
Dissociation Constantsa
Kd1
Kd2
% T
% T
‚
I
% T‚I2
100 nM
100 nM
1 µM
10 µM
100 µM
10 nM
100 nM
1 µM
33%
33%
33%
24.20%
16.20%
16.80%
33%
19.70%
8.30%
5.80%
33%
60.60%
79.70%
82.80%
33%
60.60%
83.30%
92.80%
33%
15.20%
4.10%
0.40%
33%
19.70%
8.30%
1.50%
33%
10 nM
1 nM
10 µM
1 nM
10 nM
100 nM
1 µM
19.70%
8.30%
3.05%
60.60%
83.30%
93.90%
19.70%
8.30%
3.05%
TTR amyloid fibril formation requires rate-limiting tetramer
dissociation and misassembly of partially denatured monomeric
subunits.24-28 Studies on a family in Portugal reveal that kinetic
stabilization of the TTR tetramer, comprised in part by disease-
associated subunits, is sufficient to prevent pathology.29 These
compound heterozygotes have one allele that codes for the FAP-
associated variant V30M TTR and another that codes for a so-
called interallelic trans-suppressor variant T119M. Inclusion of
T119M suppressor subunits into tetramers otherwise composed
of V30M subunits raises the kinetic barrier for tetramer
dissociation, preventing amyloidosis through transition-state
destabilization.10,30 Because T119M interallelic trans-suppres-
sion is known to prevent the onset of FAP, we are confident
that the selective stabilization of the native state of TTR relative
to its dissociative transition state(s) with small molecules will
ameliorate pathology by a similar kinetic stabilization mecha-
nism.10 Since both T4 binding sites are >98% unoccupied in
the serum and CSF,31 these sites can be utilized for kinetic
stabilization.
a Total protein concentration ) 3.6 µM; total inhibitor concentration )
3.6 µM.
of TTR amyloidogenesis can be achieved in vitro when both
T4 binding sites are occupied.10,37 However, it is difficult to
discern whether occupancy of only one T4 site is sufficient to
impose kinetic stabilization on the entire tetramer because the
inhibitor (I) dissociation constants (Kd1 and Kd2) are too similar
to populate the TTR‚I state exclusively over the TTR‚I2 and
unbound TTR states. To achieve a population of TTR‚I above
90%, the Kd’s for the T4 binding sites must be separated by 3
orders of magnitude (Table 1), a requirement not met by the
inhibitors characterized to date. Even if this requirement were
to be met, it is necessary that the TTR‚I population be constant
throughout the experiment. The latter requirement is nearly
impossible to meet because the denaturing assays used to
characterize TTR‚I can remove TTR tetramers from solution
(by either acid-mediated aggregation or irreversible tetramer
dissociation), re-distributing the populations of TTR, TTR‚I, and
TTR‚I2 in favor of the doubly bound species (TTR‚I2). This
re-distribution precludes testing the efficiency of single-site
occupancy to prevent TTR aggregation. Despite the difficulties
it is worth developing methodology to evaluate TTR‚I. Estab-
lishing the kinetic stability of TTR‚I would provide incentive
to search for highly selective inhibitors having a nanomolar Kd1
and a micromolar or better yet a millimolar Kd2.
To date, hundreds of TTR amyloidogenesis inhibitors, span-
ning a variety of structural classes, have been synthesized.32-37
Several of these small molecules bind selectively to the two T4
binding sites in complex biological fluids, imposing kinetic
stabilization on TTR. It is established that complete inhibition
(18) Hurshman, A. R.; White, J. T.; Powers, E. T.; Kelly, J. W. Biochemistry
2004, 43, 7365-7381.
(19) Hamilton, J. A.; Benson, M. D. Cell Mol. Life Sci. 2001, 58, 1491-1521.
(20) Sekijima, Y.; Wiseman, R. L.; Matteson, J.; Hammarstrom, L.; Miller, S.
R.; Sawkar, A. R.; Balch, W. E.; Kelly, J. W. Cell 2005, in press.
(21) Ikeda, S.; Nakazato, M.; Ando, Y.; Sobue, G. Neurology 2002, 58, 1001-
1007.
Covalent tethering of an amyloidogenesis inhibitor to TTR
is a convenient approach to evaluate whether occupancy of a
single T4 binding site is sufficient to prevent tetramer dissocia-
tion and subsequent amyloidogenesis.38,39 We demonstrate that
occupancy of only one T4 binding site is sufficient to impose
kinetic stabilization on the entire TTR tetramer; therefore,
formation of TTR‚I should ameliorate human amyloid disease.
(22) Saraiva, M. J. Hum. Mutat. 2001, 17, 493-503.
(23) Westermark, P.; Sletten, K.; Johansson, B.; Cornwell, G. G., III Proc. Natl.
Acad. Sci. U.S.A. 1990, 87, 2843-2845.
(24) Colon, W.; Kelly, J. W. Biochemistry 1992, 31, 8654-8660.
(25) Kelly, J. W. Curr. Opin. Struct. Biol. 1996, 6, 11-17.
(26) Kelly, J. W. Structure 1997, 5, 595-600.
(27) Lai, Z.; Colon, W.; Kelly, J. W. Biochemistry 1996, 35, 6470-6482.
(28) Liu, K.; Cho, H. S.; Lashuel, H. A.; Kelly, J. W.; Wemmer, D. E. Nat.
Struct. Biol. 2000, 7, 754-757.
(29) Coelho, T.; Carvalho, M.; Saraiva, M. J.; et al. J. Rheumatol. 1993, 20,
179.
Results
(30) Hammarstrom, P.; Schneider, F.; Kelly, J. W. Science 2001, 293, 2459-
Methodology for Tethering a Single Inhibitor to the TTR
Tetramer. A high-yield chemo- and regioselective reaction must
be employed to covalently label TTR with one amyloidogenesis
inhibitor. We have previously taken advantage of the unique
reactivity of the Cys-10 sulfhydryl moiety within each subunit,
the only Cys within TTR, to modify the TTR tetramer with
four small molecules in >90% yield, utilizing a 2-thiopyridine
disulfide-activated small molecule.40 Cys-10 is located at the
2462.
(31) Bartalena, L.; Robbins, J. Clin. Lab. Med. 1993, 13, 583-598.
(32) Adamski-Werner, S. L.; Palaninathan, S. K.; Sacchettini, J. C.; Kelly, J.
W. J. Med. Chem. 2004, 47, 355-374.
(33) Klabunde, T.; Petrassi, H. M.; Oza, V. B.; Raman, P.; Kelly, J. W.;
Sacchettini, J. C. Nat. Struct. Biol. 2000, 7, 312-321.
(34) Johnson, S. M.; Petrassi, H. M.; Palaninathan, S. K.; Mohamedmohaideen,
N. N.; Purkey, H.; Nichols, C.; Chiang, K. P.; Walkup, T.; Sacchettini, J.
C.; Sharpless, K. B.; Kelly, J. W. J. Med. Chem. 2005, 48, 1576-1587.
(35) Oza, V. B.; Smith, C.; Raman, P.; Koepf, E. K.; Lashuel, H. A.; Petrassi,
H. M.; Chiang, K. P.; Powers, E. T.; Sachettinni, J.; Kelly, J. W. J. Med.
Chem. 2002, 45, 321-332.
(36) Razavi, H.; Palaninathan, S. K.; Powers, E. T.; Wiseman, R. L.; Purkey,
H. E.; Mohamedmohaideen, N. N.; Deechongkit, S.; Chiang, K. P.; Dendle,
M. T.; Sacchettini, J. C.; Kelly, J. W. Angew. Chem., Int. Ed. Engl. 2003,
42, 2758-2761.
(37) Green, N. S.; Palaninathan, S. K.; Sacchettini, J. C.; Kelly, J. W. J. Am.
Chem. Soc. 2003, 125, 13404-13414.
(38) Hardy, J. A.; Lam, J.; Nguyen, J. T.; O’Brien, T.; Wells, J. A. Proc. Natl.
Acad. Sci. U.S.A. 2004, 101, 12461-12466.
(39) Erlanson, D. A.; Wells, J. A.; Braisted, A. C. Annu. ReV. Biophys. Biomol.
Struct. 2004, 33, 199-223.
(40) Zhang, Q.; Kelly, J. W. Biochemistry 2003, 42, 8756-8761.
9
J. AM. CHEM. SOC. VOL. 127, NO. 15, 2005 5541

A R T I C L E S
Wiseman et al.
Figure 1. Schematic of the covalent modification procedure. (A) Anion
exchange chromatogram of coexpressed C10A/flag-tag WT TTR and the
resulting TTR heterotetramers. (B) The (flag-tag WT TTR)1(C10A)3
heterotetramer was incubated with small molecules activated with a
2-thiopyridine moiety (1i-1vi) at room temperature. (a) Excess small
molecule was removed from solution by gel filtration chromatography (b),
resulting in TTR tetramers covalently tethered to a single inhibitor. The
table defines the nomenclature of the activated inhibitors and the resulting
modified heterotetramers prepared from compounds 1i-1vi with varying
linker lengths. The prime indicates the inhibitors disulfide tethered to TTR
without the 2-thiopyridine-activating moiety.
Figure 2. Structure of compound 2 bound to TTR reveals an accessible
position for covalent attachment to TTR. (A) Chemical structure of
compound 2. (B) The percentage of unbound (TTR; black), singly bound
(TTR‚2; red), and doubly bound (TTR‚22; green) TTR as a function of the
concentration of 2. The arrow indicates the position of the curve where the
concentration of the TTR tetramer is equal to the concentration of 2. (C)
Crystal structure of TTR‚22 demonstrating the suitability and accessibility
of the para-position to mediate tethering to TTR (red line). Adapted from
ref 35.
periphery of the T4 binding site, making this flexible residue
ideal for testing the sufficiency of single occupancy for
inhibiting TTR amyloidogenesis. Disulfide tethering of the
inhibitor has the added advantage that the linker can be cleaved
with DTT, allowing control experiments requiring free ligand.
To achieve monolabeling, a tetrameric species must be
prepared that contains only one Cys-bearing TTR subunit. TTR
heterotetramers of defined stoichiometry have been prepared
previously, utilizing one subunit that has a distinguishing 16-
residue N-terminal tandem flag-tag (DYKDDDDKDYKD-
DDDK).10,41 Coexpressing flag-tag WT TTR and TTR contain-
ing the point mutation C10A results in a statistical distribution
of heterotetramers from which the (flag-tag WT)1(C10A)3
tetramer can be separated from the other four resolvable
tetramers using anion-exchange chromatography (Figure 1A).
The 2,2,2 symmetry of the TTR tetramer renders all (flag-tag
WT)1(C10A)3 heterotetramers equivalent. Importantly, the (flag-
tag WT)1(C10A)3 heterotetramer appears to be kinetically stable
at 37 °C for at least 24 h, preventing reequilibration (Figure
1A). Incubation of (flag-tag WT)1(C10A)3 with a 2-thiopyridine
disulfide-activated amyloidogenesis inhibitor (e.g., 1i-vi) will
result in the formation of TTR tetramers covalently labeled at
one subunit (Figure 1B).40
The small molecule selected for tethering in this experiment
(2-(3,5-dichlorophenylamino)benzoic acid (2)) is one of the most
efficacious TTR amyloidogenesis inhibitors characterized to date
(Figure 2A; Kd1 ) Kd2 ) 5 nM).10,35 Unfortunately, these
dissociation constants prevent the population of TTR‚I to the
exclusion of TTR and TTR‚I2 at equimolar concentrations of
free ligand and TTR tetramer (Figure 2B). The crystal structure
of TTR‚22 demonstrates that the para-position of the dichlori-
nated phenyl ring points directly out of the binding pocket into
the solvent (Figure 2C),35 presenting an ideal site for tether
attachment. Incorporation of a para-phenol enables poly-
(ethylene glycol) (PEG) chains of variable length to be attached
onto 2 (Scheme 1). These para-pegylated analogues of 2 (5a-
f) inhibit TTR aggregation as effectively as 2 in an acid-
mediated amyloidogenesis assay (Figure 3), suggesting that PEG
linkers will be effective for tethering this inhibitor to the TTR
tetramer via a disulfide-forming reaction as outlined above.
Computer modeling taking into account the flexibility of the
Cys-10 residue suggests that a PEG linker of 14-21 Å
(41) Schneider, F.; Hammarstrom, P.; Kelly, J. W. Protein Sci. 2001, 10, 1606-
1613.
9
5542 J. AM. CHEM. SOC. VOL. 127, NO. 15, 2005

Kinetic Stabilization of an Oligomeric Protein
A R T I C L E S
Scheme 1. Synthesis of Compounds To Evaluate the Feasibility
Scheme 2. Synthesis of 2-Mercaptopyridine-Activated Disulfide
of Covalent Tethering at the para-Positiona
Linkersa
a Reagents and conditions: (a) Benzoyl chloride, pyridine, CH2Cl2 (65-
82%); (b) NBS, PPh3, CH2Cl2 (79-99%); (c) thioacetic acid, NaH, DMF
(75-94%); (d) 2,2′-dithiodipyridine, sodium methoxide (0.5 M in methanol),
methanol (23-45%).
a Reagents and conditions: (a) Benzyl bromide, K2CO3, acetone (66%);
(b) methyl 2-bromobenzoate, Pd2(dba)3, BINAP, Cs2CO3, toluene, reflux
(68%); (c) 10% Pd/C, H2, 1:1 MeOH:EtOAc (97%); (d) MeI, K2CO3, DMF
(73%); (e) HO(CH2CH2O)nMe (n ) 1-3), PPh3, DIAD, THF (72-89%);
(f) LiOH‚H2O, 3:1:1 THF:MeOH:H2O (74-100%).
Scheme 3. Synthesis of the Inhibitors Activated for Disulfide
Tetheringa
Figure 3. Acid-mediated (pH 4.4) fibril formation of WT TTR (3.6 µM)
in the presence of untethered inhibitors 5a-f (7.2 µM, 72 h). The turbidity
was measured at 400 nm, normalizing the signal to WT TTR in the absence
of small molecule.
comprising 4-6 PEG units would be optimal; therefore, a range
of linkers was evaluated (Figure 1B).
Although reaction of 2,2′-dithiodipyridine with 2-mercapto-
ethanol was found to efficiently produce the shortest activated
disulfide linker (9i, n ) 1, 67%),40 the lack of longer com-
mercially available monomercaptopoly(ethylene glycol)s neces-
sitated synthetic methodology to make these linkers (Scheme
2). Commercially available poly(ethylene glycol)s (n ) 1-6)
were protected as monobenzoates, brominated,42 and subjected
to nucleophilic attack by thioacetic acid affording acetylsulfa-
nylpoly(ethylene glycol) benzoates in good overall yields (8i-
vi, 51-67%, Scheme 2). Methanolysis of the ester and thioester
protecting groups followed by thiol activation with 2,2′-
dithiodipyridine afforded the linkers (9i-vi) in moderate yields
(23-45%).40,43 The activated linkers were then coupled (76-
98%) to compound 12 via Mitsunobu reactions, taking care to
a Reagents and conditions: (a) tert-BuOH, DCC, DMAP, CH2Cl2 (61%);
(b) 4-benzyloxy-3,5-dichloroaniline (3), Pd2(dba)3, BINAP, Cs2CO3, toluene,
reflux (58%); (c) 10% Pd/C, H2, 1:1 MeOH:EtOAc (89%); (d) 9i-vi, PPh3,
DIAD, THF (85-91%); (e) TFA (100%).
preactivate the triphenylphosphine with diisopropyl azodicar-
boxylate prior to linker addition to avoid reduction of the
activated disulfide bond (Scheme 3). The methyl ester utilized
in Scheme 1 was changed to a tert-butyl ester because the
2-thiopyridine-activated disulfide group proved labile under the
alkaline conditions required to saponify the methyl ester. The
activated disulfide survives under the acidic conditions required
to liberate the tert-butyl ester protecting group (Scheme 3, step
e), affording the desired covalent inhibitor precursors (1i-vi)
in quantitative yields.44
(42) Hammerschmidt, F.; Kahlig, H. J. Org. Chem. 1991, 56, 2364-2370.
(43) DeVries, V. G.; Moran, D. B.; Allen, G. R.; Riggi, S. J. J. Med. Chem.
1976, 19, 946-957.
9
J. AM. CHEM. SOC. VOL. 127, NO. 15, 2005 5543

A R T I C L E S
Wiseman et al.
Covalent Tethering of Small Molecules to TTR Is Medi-
ated by Disulfide Bond Formation. TTR-1i′-TTR-1vi′ were
prepared by incubating the (flag-tag WT)1(C10A)3 heterotet-
ramer with an excess of activated inhibitors (1i-vi) (Figure 1B,
step a, 2 h, 25 °C). The excess small molecule was removed by
gel filtration chromatography, yielding inhibitors disulfide
tethered to the TTR heterotetramer (TTR-1i′-TTR-1vi′) in
>80% yield (Figure 1B, step b). Removal of 1i-vi from
solutions of TTR-1i′-TTR-1iv′ could be verified by HPLC in
every case except for TTR-1iv′, where the modified flag-tag
WT subunit of TTR-1iv′ coeluted with the reduced untethered
molecule 1ivSH. It is unlikely that a significant amount of free
1iv is present after gel filtration, as demonstrated in Figure 4
(see below). Mixed disulfide formation was monitored by HPLC
under conditions that denature the heterotetramer into individual
subunits, revealing labeling of the flag-tag WT TTR subunit
with the inhibitor (cf. Figure 4A and B). Integration of the HPLC
peaks corresponding to C10A subunits and the modified flag-
tag WT TTR subunit of TTR-1iv′, correcting for the difference
in molar absorptivity between the C10A and the modified flag-
tag WT TTR subunit of TTR-1iv′, demonstrates a distribution
of ∼3:1 (Figure 4B) as expected. LC-MS demonstrated that only
the flag-tag WT TTR subunits were modified and not the C10A
subunits (data not shown). Upon treatment of the covalently
labeled sample with 1 mM DTT, the covalently modified protein
completely reverted to its unlabeled state (Figure 4C). Integra-
tion of the HPLC peaks corresponding to C10A TTR subunits
and flag-tag WT TTR subunits after DTT treatment remains
3:1, verifying the reversibility of the tether attachment (cf. Figure
4A and C). The demonstrated ability of gel filtration to remove
aromatic inhibitors by adsorption, the 3:1 ratio of C10A and
modified flag-tag WT TTR subunits of TTR-1iv′ after reduction,
and the substantially different amyloidogenicity and tetramer
dissociation exhibited by untreated and DTT-treated samples
make it unlikely that any untethered inhibitor is present in
solutions of TTR-1iv′. During inhibitor tethering ∼10% of flag-
tag WT TTR was converted to a byproduct, indicated by the
small right-most peak of the HPLC trace (Figure 4B). LC-MS
could not detect this product, which is likely the 2-thiopyridine
disulfide adduct, consistent with the disappearance of this peak
upon DTT reduction (Figure 4C).
Figure 4. HPLC traces of the unmodified heterotetramer ((flag-tag WT)1-
(C10A)3), a covalently modified TTR heterotetramer (TTR-1iv′), and a
covalently modified TTR heterotetramer (TTR-1iv′) treated with DTT under
denaturing conditions. (A) Representative HPLC trace of the monomeric
subunits from the unmodified heterotetramer (flag-tag WT TTR)1(C10A)3.
Integration of C10A subunits relative to flag-tag WT TTR subunits
(corrected for the difference in molar absorptivity between flag-tag WT
and C10A subunits) is 2.8:1. (B) Representative HPLC trace of monomeric
subunits from the modified heterotetramer TTR-1iv′. Integration of C10A
subunits relative to the modified flag-tag WT subunits from TTR-1iv′
(corrected for the difference in molar absorptivity between the modified
flag-tag WT subunit from TTR-1iv′ and C10A subunits) reveals a ratio of
3:1. (C) Representative HPLC trace of monomeric subunits from the
modified heterotetramer TTR-1iv′ incubated with 1 mM DTT demonstrating
the reversibility of the disulfide attachment. The peaks corresponding to
free small molecule (1ivSH) and the modified flag-tag WT TTR subunit of
TTR-1iv′ are found to elute at the same retention time, making it difficult
to discern the disappearance of the modified flag-tag WT TTR subunit from
TTR-1iv′ upon treatment with DTT. However, a 3:1 ratio of the integrations
for the C10A subunits relative to the unmodified flag-tag WT TTR subunits
is observed upon DTT reduction, demonstrating that the reduction is
Inhibitor Tethering Does not Disturb TTR’s Native
Structure. Tethering small molecules to TTR could alter its
tertiary and/or quaternary structure. Hence, the far-UV circular
dichroism (CD) spectra of TTR‚I conjugates were compared to
(flag-tag WT)1(C10A)3 TTR, demonstrating that the â-sheet
content remains unchanged (Figure 4D). The slight difference
in intensity is most likely a result of a change in the molar
absorptivity associated with inhibitor tethering. TTR‚I conjugates
with linkers that are too short or too long to allow for efficient
intramolecular binding could result in oligomerization by
tethered inhibitor binding to a T4 site in another tetramer.
Therefore, all of the TTR‚I conjugates (TTR-1i′-TTR-1vi′)
were subjected to sedimentation velocity analytical ultracentri-
fuge experiments (Supporting Information, Figure 1). In all cases
the conjugates were tetrameric. No intermolecular binding was
observed based on the absence of structures larger than
tetramers.
quantitative. The elution time of other reduced small molecules (1iSH
-
1viSH) is distinct from the modified flag-tag WT TTR subunits of TTR-
1i′-TTR-1iv′, allowing this analysis to be made. (D) CD spectra of
modified and unmodified TTR heterotetramers demonstrating that the small-
molecule attachment has little effect on the secondary structure of the
tetrameric protein ((flag-tag WT TTR)1(C10A)3, black; TTR-1i′, red; TTR-
1ii′, blue; TTR-1iii′, green; TTR-1iv′, orange; TTR-1v′, purple; TTR-
1vi′, pink).
(44) Hayword, M. M.; Adrian, J. C.; Schepartz, A. J. Org. Chem. 1995, 60,
3924-3927.
9
5544 J. AM. CHEM. SOC. VOL. 127, NO. 15, 2005

Kinetic Stabilization of an Oligomeric Protein
A R T I C L E S
Table 2. Data Processing and Refinement Statistics for the
TTR-1iv′ Crystal Structure (Figure 5)
resolution range (outer shell)
unique reflections
completeness
30-1.69 Å (1.76-1.69 Å)
26 531 (2795)
96.5% (88.5%)
3.4 (2.3)
6.0% (60.7%)
33.2 (2.3)
redundancy
a
Rsym
average I/σ(I)
refinement
resolution range (outer shell)
27.63-1.69 Å (1.73-1.69 Å)
21.7% (33.7%)
23.9% (36.3%)
b
Rcryst
b
Rfree
protein atoms/water molecules/ligand
1712/104/42
molecules
coordinate errorc
0.084 Å
rmsd bonds (angles)
0.015 Å (1.55°)
B protein atoms/water molecules/ligand 16.2 Å2/32.6 Å2/41.9 Å2
molecules
Ramachandran statistics
additional allowed
most favored
7.4%
92.6%
a Rsym ) 100‚∑h∑i|li(h) - l(h) |/∑hl(h), where li(h) is the ith measure-
ment of the h reflection and l(h) is the average value of the reflection
b
intensity. Rcryst ) ∑|Fo| - |Fc|/∑|Fo|, where Fo and Fc are the structure
factor amplitudes from the data and the model, respectively. Rfree is Rcryst
with 5 % of test set structure factors. c Based on Maximum Likelihood.
Crystal Structure of the TTR-1iv′ Conjugate Verifies
Intramolecular Binding to the T4 Site. The crystal structure
of the TTR-1iv′ conjugate reveals a unimolecular binding mode
that is virtually identical to that of the untethered TTR‚22
Figure 5. (A) Crystal structure of TTR-1iv′ demonstrating that the
covalently attached small molecule can bind to the T4 binding site. One of
the two hormone binding sites at the dimer interface of TTR-1iv′ is shown
depicting the two symmetry-related binding modes. Ligand density is from
a 3Fo - 2Fc simulated annealing-omit map in both binding modes and is
contoured at 1.0 σ. (B) Stereoview of an overlay of apo-TTR (green) and
TTR-1iv′ (blue). The residues represented by the stick model are identical
to those in Figure 5A.
complex (cf. Figures 2C and 5A). The 1.69 Å model (Rcryst
)
21.7%, Rfree ) 23.9%, Table 2) reveals that 92.6% of the
residues are in the most favored region of the Ramachadron
plot while the remainder are in an allowed region. The first
nine residues, the last three residues, the 16 residues of the flag-
tag, as well as the PEG linker tethering the inhibitor to Cys-10
were all too flexible to provide interpretable electron density.
Because TTR-1iv′ crystallized in the P212121 space group, there
is a random distribution of occupied and unoccupied sites in
the crystal; therefore, each T4 binding site exhibits partial
occupancy. There are two symmetry-related binding modes
observed in all TTR cocrystal structures due to a C2 axis of
symmetry running through the center of the T4 sitessthese
modes were observed in the structure of TTR-1iv′. The
superimposition of the TTR-1iv′ structure with the apo-TTR
crystal structure (Figure 5B) reveals a high level of structural
homology, with rmsd’s for main chain and all atoms of 0.19
and 0.74 Å, respectively. The conformations adopted by residues
comprising the T4 binding site are identical to those of apo-
WT TTR. Dissolution of the TTR-1iv′ crystal followed by
HPLC analysis reveals that crystallization did not sever the
tether.
of T4 site occupancy. Incorporation of a single PEG unit into
the linker (TTR-1i′) affords a conjugate that was almost as
amyloidogenic as WT TTR at pH 4.4 (Figure 6A), consistent
with modeling indicating that this linker would be too short to
allow sustained binding of the small molecule to the T4 site.
Doubling the length of the PEG linker to two units (TTR-1ii′)
slightly decreases aggregation; however, the pH optimum of
TTR-1ii′ aggregation is similar to that of (flag-tag WT)1(C10A)3
TTR, suggesting that the occupancy of small molecule in the
T4 binding site was low. Increasing the linker length to 3-6
PEG units (TTR-1iii′-TTR-1vi′) markedly decreased amy-
loidogenicity as compared to (flag-tag WT)1(C10A)3 TTR. The
small amount of aggregation observed (<15%) from TTR-1iii′-
TTR-1vi′ is consistent with the fraction of TTR tetramer lacking
a tethered inhibitor (see below), implying that occupancy of
one T4 binding site prevents amyloidogenesis.
Occupancy of One T4 Binding Site Effectively Prevents
TTR Aggregation Including Amyloidogenesis. TTR can be
induced to form aggregates and amyloid via partial acid
denaturation.24,27 Aggregation, as measured by turbidity, is
maximal over the pH range of 4.0-4.8. The best inhibitors block
>90% of aggregation at a concentration twice that of the protein
concentration (7.2 µM). Tethering one inhibitor to the T4 site
allows us to test the amyloidogenicity of TTR‚I conjugates and
evaluate the sufficiency of one binding event to impose kinetic
stabilization on the entire tetramer.
Reducing the disulfide within the tether using DTT allows
the activity of the untethered inhibitors to be assessed. Incubating
TTR-1i′-TTR-1vi′ at pH 4.4 with 1 mM DTT resulted in
∼30% fibril formation (Figure 6B; red symbols), very similar
to the extent of aggregation allowed by 2 (Figure 6B, dashed
red line),10 indicating that the PEG linker is rather innocuous.
The reduced inhibitors are more effective than their tethered
counterparts in the case of TTR-1i′ and TTR-1ii′, likely owing
to the strain induced by the short linker. In contrast, the tethered
inhibitors containing 3-6 PEG units are much more effective
than their untethered counterparts (Figure 6B, compare red and
black symbols).
The efficacy of the tethered inhibitors is highly dependent
on the length of the linker, which likely correlates with the extent
9
J. AM. CHEM. SOC. VOL. 127, NO. 15, 2005 5545

A R T I C L E S
Wiseman et al.
Figure 6. Assessing the amyloidogenicity of TTR-1i′-TTR-1vi′. (A) Acid-
mediated aggregation of TTR-1i′-TTR-1vi′ (3.6 µM) as a function of pH
and the length of the PEG linker. The extent of fibril formation was
normalized to the unmodified heterotetramer at pH 4.4 ((flag-tag WT TTR)1-
(C10A)3, -black circles; TTR-1i′, red squares; TTR-1ii′, blue diamonds;
TTR-1iii′, green ×; TTR-1iv′, orange crosses; TTR-1v′, purple triangles;
TTR-1vi′, pink upside down triangles). (B) Fibril formation (pH 4.4) of
TTR-1i′-TTR-1vi′ in the absence (black) and presence (red) of 1 mM
DTT. Proteins with a tethered inhibitor clearly show a dependence on the
linker length, which is lost upon reduction with DTT. The dotted lines
represent the amyloidogenicity of the control heterotetramer ((flag-tag WT
TTR)1(C10A)3; 3.6 µM) in the absence of small molecule (black) and in
the presence of compound 2 (3.6 µM) lacking a tether (red).
A Single Tethered Inhibitor Dramatically Slows Tetramer
Dissociation. The rate of TTR tetramer dissociation is followed
by linking the slow tetramer dissociation process to the much
faster monomer unfolding event, monitored by circular dichro-
ism (CD) spectroscopy under denaturing conditions.45 This
approach has been used previously to determine the extent of
kinetic stabilization of the TTR tetramer as a function of
inhibitor concentration.10,32,36 The best untethered small mol-
ecules completely inhibit dissociation in 6.0 M urea when
incubated at a concentration (3.6 µM) twice that of the TTR
tetramer (1.8 µM).10 Reducing the concentration of these small
molecules does not allow us to conclude anything about the
extent to which single inhibitor occupancy imposes kinetic
stabilization on TTR for the reasons discussed in the Introduc-
tion.
The TTR tetramer dissociation rates from TTR-1i′ and TTR-
1ii′ were only slightly slower (0.032 and 0.019 h-1, respectively)
than the unmodified heterotetramer control (0.042 h-1), sug-
gesting that the short linkers lead to inefficient inhibitor binding
(Figure 7A and B). Lengthening the covalent linker to 3-6 PEG
units (TTR-1iii′-TTR-1vi′) afforded tetramers displaying very
Figure 7. Assessing TTR tetramer dissociation kinetics in 6.0 M urea.
(A) The dissociation kinetics of TTR-1i′-TTR-1vi′ in 6.0 M urea was
revealed by monitoring the CD signal at 215 nm. Only the longer linkers
allow sustained binding of the ligand to the T4 binding site and inhibition
of tetramer dissociation ((flag-tag WT TTR)1(C10A)3, black circles; TTR-
1i′, red squares; TTR-1ii′, blue diamonds; TTR-1iii′, green ×; TTR-1iv′,
orange crosses; TTR-1v′, purple triangles; TTR-1vi′, pink upside down
triangles). (B) Plot displaying the extent of tetramer dissociation after 168
h in 6.0 M urea. The shorter linkers demonstrate a large amount of tetramer
dissociation, which is drastically reduced as the linker length is increased.
The dotted line represents the extent of dissociation for the control
heterotetramer ((flag-tag WT TTR)1(C10A)3) in the absence of small
molecule. (C) Dissociation kinetics of TTR-1i′-TTR-1vi′ treated with 1
mM DTT in 6.0 M urea. The untethered molecules demonstrate an inverse
relationship, relative to the tethered molecules, between their ability to inhibit
tetramer dissociation and linker length.
little dissociation over the time course of this experiment (Figure
7B), indicating that occupancy of a single T4 site by an inhibitor
can impose kinetic stabilization on the entire TTR tetramer.
Experiments below demonstrate that the small (<20%) dis-
sociation amplitude observed for TTR-1iii′-TTR-1vi′ results
from unmodified TTR remaining in solution; hence, it can be
(45) 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
5546 J. AM. CHEM. SOC. VOL. 127, NO. 15, 2005

Kinetic Stabilization of an Oligomeric Protein
A R T I C L E S
concluded that TTR-1iii′-TTR-1vi′ do not dissociate in 6.0
M urea. DTT reduction of TTR-1i′-TTR-1vi′ reveals that the
efficacy of the untethered inhibitors is highly dependent on the
length of the PEG linker (Figure 7C). Surprisingly, the least
effective tethered inhibitors (TTR-1i′ and TTR-1ii′; Figure 7A)
proved to be the most effective untethered inhibitors of tetramer
dissociation (Figure 7C). This inverse relationship is most likely
a result of the length of the PEG linker reducing the binding
constants of the free ligands proportional to the number of PEG
units.
conditions (urea and acid). Tethering a small molecule to TTR
with an optimal linker length creates a very high local
concentration of ligand at the binding site, which results in a
high population of a stabilized singly occupied TTR tetramer.
Stabilization of One of the Quaternary Structural Inter-
faces Associated with Inhibitor Binding Is Sufficient To
Impose Kinetic Stabilization on the Entire TTR Tetramer.
Herein, we show that occupancy of one T4 binding site by an
aromatic inhibitor facilitates intersubunit hydrophobic contacts
that impose kinetic stabilization on the entire TTR tetramer by
selective stabilization of the ground state over the dissociation
transition state. We recently demonstrated that connecting the
C-terminus to the N-terminus on only one of the two pairs of
subunits composing the T4 binding sites can also impose kinetic
stabilization on the entire tetramer under denaturing conditions.46
From the work described herein, and elsewhere,10,46,47 it would
appear that the dissociation of TTR about the T4 binding site
interface is energetically most feasible; thus, stabilization or
perturbation of this interface is critical to achieve kinetic
stabilization.
Sedimentation velocity analytical ultracentrifuge data on a
sample of TTR-1vi′ that had been incubated in 6.0 M urea for
168 h revealed that ∼80% of the conjugate was tetrameric and
∼20% monomeric (data not shown), indicating that the observed
transition in the kinetic urea dissociation experiment is not an
artifact (Figure 7A). The dissociation rate constant reflecting
the 20% amplitude was determined to be ∼0.030 h-1 for TTR-
1iii′-TTR-1vi′, a rate very similar to that exhibited by (flag-
tag WT)1(C10A)3 TTR. The similarity in rates suggests that the
observed dissociation results from the incomplete labeling of
the (flag-tag WT)1(C10A)3 TTR heterotetramer. Consistent with
this interpretation, a 10-15% amplitude is also observed under
acid-mediated amyloidogenesis (Figure 6A). Aggregates from
these samples were subjected to centrifugation, isolated, and
redissolved in DMSO to determine their composition. HPLC
analysis revealed that the aggregates consisted primarily of
C10A and unmodified flag-tag WT TTR (Supporting Informa-
tion Figure 2A), demonstrating that the small amount of
aggregation (Figure 6A) and dissociation (Figure 7A) is
attributable to a small fraction (<20%) of unmodified (flag-tag
WT TTR)1(C10A)3 heterotetetramers.
Substoichiometric Concentrations of TTR Amyloidogen-
esis Inhibitors Prevent TTR Amyloidogenesis in Vitro.
Substantial inhibition of amyloidogenesis does not require that
every TTR tetramer be kinetically stabilized by small molecule
binding. Misfolded monomeric TTR aggregates into amyloid
via a downhill polymerization mechanism where all forward
steps in the pathway are favorable; hence, the rate of aggregation
is dependent on the concentration of the misfolded TTR
monomer.18 Since the concentration of monomer is dependent
on the dissociation of the tetramer, it is evident that controlling
tetramer dissociation energetics will allow significant control
over the amyloidogenic monomer concentration, which dictates
the rate of TTR amyloidogenesis. Even substoichiometric
concentrations (<1 equiv) of small molecule inhibitors may be
sufficient to adequately reduce the concentration of monomeric
TTR in the serum below the concentration where aggregation
is efficient, thereby preventing disease.
To provide more evidence to support this hypothesis, TTR-
1iv′ was subjected to acid-mediated aggregation conditions (pH
4.2, 37 °C) for 72 h, centrifuged, and filtered to remove the
fraction of unlabeled tetramer capable of aggregating. HPLC,
LCMS, and SDS-PAGE confirmed that the supernatant con-
sisted of one flag-tag WT subunit tethered to an inhibitor and
three C10A subunits (data not shown). This pretreated protein
was subjected to urea denaturation (6.0 M urea) to measure its
dissociation kinetics, revealing a dramatic decrease in the
amplitude of the dissociation phase (Supporting Information
Figure 2B). Collectively these data demonstrate that the un-
conjugated protein is responsible for the small amplitude
observed in the urea-mediated denaturation assay, strongly
supporting our conclusion that single inhibitor occupancy is
sufficient to impose kinetic stabilization on the native tetrameric
structure of TTR.
The sufficiency of a single binding event to stabilize the entire
TTR tetramer suggests that the most efficient molecules for
native state stabilization of TTR would display striking negative
cooperativity, with a Kd1 in the nanomolar or subnanomolar
range and a Kd2 in the micromolar (or millimolar) range,
maximizing the population of TTR‚I. At substoichiometric
ligand concentrations positive and noncooperative ligands would
bind a lower mole fraction of TTR tetramers monovalently
because some tetramers would be doubly liganded and others
unliganded as dictated by the two dissociation constants. An
ideal inhibitor would display strong negative cooperativity
between the two thyroxine binding sites, increasing the total
population of bound tetramer (T‚I + T‚I2) and lowering the
population of unbound tetramer capable of undergoing amy-
loidogenesis (Figure 8). Ideal molecules would also display high
binding selectivity for TTR over other proteins in the serum.
High binding selectivity, negative cooperativity, and a subna-
nomolar Kd1 would allow inhibitor concentrations equal to or
less than the TTR protein concentration to be employed,
minimizing pharmaceutical side effects.
Discussion
The binding of small molecules to proteins stabilizes the
native state proportional to the binding constant. This stabiliza-
tion not only shifts the equilibrium toward the folded state, but
it also can raise the kinetic barrier between the native state and
the transition state associated with denaturation, provided that
the small molecule stabilizes the native state selectively over
the transition state. We previously demonstrated that native state
kinetic stabilization is a viable therapeutic approach to prevent
TTR amyloidoses.10 Herein, we show that occupancy of one
T4 binding site within TTR is sufficient to dramatically slow
dissociation of the entire TTR tetramer under denaturing
(46) Foss, T.; Kelker, M. S.; Wiseman, R. L.; Wilson, I. A.; Kelly, J. W. J.
Mol. Biol. 2005, 347, 841-854.
(47) Schormann, N.; Murrell, J. R.; Benson, M. D. Amyloid 1998, 5, 175-187.
9
J. AM. CHEM. SOC. VOL. 127, NO. 15, 2005 5547

A R T I C L E S
Wiseman et al.
the reaction was concentrated to a slurry, diluted with water (200 mL),
and extracted with EtOAc (3 × 100 mL). The combined organics were
washed with saturated NaHCO3 (2 × 25 mL) and brine (25 mL) and
extracted with 30% HCl (3 × 30 mL), and the white precipitate was
filtered and rinsed with EtOAc. The aqueous layer of the filtrate was
separated and washed with dichloromethane (2 × 30 mL), combined
with the white precipitate, adjusted to pH ≈ 8 with NaOH pellets, and
extracted with DCM (2 × 50 mL). The combined DCM extracts were
washed with brine (25 mL), dried over Na2SO4, filtered, and concen-
trated to afford 4-benzyloxy-3,5-dichloroaniline (3) as a tan solid (5.01
1
g, 66%). H NMR (600 MHz, CDCl3) δ 7.53-7.57 (m, 2H), 7.37-
7.41 (m, 2H), 7.32-7.36 (m, 1H), 6.60 (s, 2H), 4.94 (s, 2H), 3.62 (s,
2H); 13C NMR (150 MHz, CDCl3) δ 143.68, 143.13, 136.82, 129.93,
128.62, 128.55, 128.39, 115.11, 75.22; ESI-MS 268 m/z [MH]+, C13H12-
Cl2NO requires 268.
Methyl 2-(4-Benzyloxy-3,5-dichlorophenylamino)benzoate (4a).
Toluene (10.0 mL) was added to a flask charged with 3 (1.17 g,
5.36 mmol), methyl 2-bromobenzoate (0.73 mL, 5.20 mmol), tris-
(dibenzylideneacetone)dipalladium (0) (183 mg, 0.200 mmol), (R)-(+)-
2,2′-bis(diphenylphosphino)-1,1′-binaphthyl (82.4 mg, 0.132 mmol), and
cesium carbonate (2.00 g, 6.14 mmol) and then stirred at reflux under
an argon atmosphere. After 24 h the reaction was filtered through Celite
and concentrated. Flash chromatographic purification over silica (9:
1-4:1 hexanes:EtOAc gradient elution) afforded methyl 2-(4-benzyl-
oxy-3,5-dichlorophenylamino)benzoate (4a) as a yellow solid (1.20 g,
Figure 8. Plot displaying the concentration of unbound protein as a function
of the second dissociation constant of a soluble inhibitor (TTR concentration
) 3.6 µM, Kd1 ) 1 nM). Increasing Kd2 decreases the amount of unbound
protein, demonstrating that negatively cooperative small molecules appear
to be ideal to prevent TTR amyloidogenesis.
Conclusions
Prior to this work it was established that the inhibition of
TTR tetramer dissociation, which is rate-limiting for fibril
formation, could be mediated by amyloidogenesis inhibitor
occupancy of both T4 binding sites. Using the tethering approach
described above it is now safe to conclude that occupancy of a
single T4 binding site is capable of imposing kinetic stabilization
on the entire tetrameric structure. Because binding of one T4
binding site is sufficient to dramatically inhibit amyloidogenesis,
it is now evident that the therapeutic concentration of small
molecule amyloidogenesis inhibitors could be lowered if a
negatively cooperative, TTR-selective small molecule could be
discovered that has a nanomolar Kd1 and a micromolar, or
preferably a millimolar, Kd2.
1
68%). H NMR (500 MHz, CDCl3) δ 9.44 (s, 1H), 7.98 (dd, J ) 1.7,
8.0 Hz, 1H), 7.56-7.59 (m, 2H), 7.34-7.43 (m, 2H), 7.33-7.38 (m,
2H), 7.23 (apparent dd, J ) 0.8, 8.5 Hz, 1H), 7.21 (s, 2H), 6.82
(apparent ddd, J ) 1.1, 7.0, 8.0 Hz, 1H), 5.03 (s, 2H), 3.90 (s, 3H);
13C NMR (125 MHz, CDCl3) δ 168.95, 146.88, 146.84, 138.32, 136.58,
134.47, 131.91, 130.22, 128.67, 128.62, 128.54, 122.17, 118.52, 114.59,
113.06, 75.27, 52.10; ESI-MS 402 m/z [MH]+, C21H18Cl2NO3 requires
402, 424 m/z [MNa]+, C21H17Cl2NaNO3 requires 424.
Methyl 2-(4-Hydroxy-3,5-dichlorophenylamino)benzoate (4b).
Ten percent palladium on carbon (64.3 mg, 0.0604 mmol) and 4a
(1.90 g, 4.72 mmol) were stirred in a 1:1 mixture of MeOH:EtOAc
(48 mL) under a hydrogen atmosphere for 6 h. The atmosphere was
then purged with argon, and the reaction was quenched with DCM,
filtered through Celite, and concentrated, affording methyl 2-(4-
hydroxy-3,5-dichlorophenylamino)benzoate (4b) as a greenish solid
(1.42 g, 97%). 1H NMR (500 MHz, d6-DMSO) δ 9.93 (s, 1H), 9.06 (s,
1H), 7.86 (dd, J ) 1.6, 8.0 Hz, 1H), 7.40 (apparent ddd, J ) 1.6, 7.1,
8.5 Hz, 1H), 7.26 (s, 2H), 7.02 (apparent dd, J ) 0.7, 8.5 Hz, 1H),
6.79 (apparent ddd, J ) 0.9, 7.1, 8.0 Hz, 1H), 3.83 (s, 3H); 13C NMR
(125 MHz, d6-DMSO) δ 167.84, 147.04, 145.55, 134.60, 133.50,
131.32, 123.23, 122.86, 117.81, 114.17, 112.08, 51.99; ESI-MS 312
m/z [MH]+, C14H12Cl2NO3 requires 312.
Methyl 2-(4-Methoxy-3,5-dichlorophenylamino)benzoate (4c).
Methyliodide (25.0 µL, 0.40 mmol), methyl 2-(4-hydroxy-3,5-
dichlorophenylamino)benzoate (4b) (103.0 mg, 0.330 mmol), and K2-
CO3 (50.9 mg, 0.368 mmol) were stirred in anhydrous DMF under an
argon atmosphere for 18 h. The reaction was then poured into water
(35 mL) and extracted with EtOAc (3 × 20 mL). The combined
organics were washed with water (3 × 30 mL), dried over Na2SO4,
filtered, and concentrated. Flash chromatographic purification over silica
(4:1 hexanes:EtOAc) afforded methyl 2-(4-methoxy-3,5-dichlorophe-
nylamino)benzoate (4c) as a white solid (79.1 mg, 73%). 1H NMR (500
MHz, CDCl3) δ 9.42 (s, 1H), 7.97 (dd, J ) 1.4, 8.0 Hz, 1H), 7.38
(apparent ddd, J ) 1.3, 7.1, 8.4 Hz, 1H), 7.21 (d, J ) 8.6 Hz, 1H),
7.19 (s, 2H), 6.81 (apparent ddd, J ) 0.8, 7.2, 8.0 Hz, 1H), 3.90 (s,
3H), 3.89 (s, 3H); 13C NMR (125 MHz, CDCl3) δ 168.93, 148.15,
146.88, 138.11, 134.46, 131.88, 129.85, 122.25, 118.44, 114.48, 112.93,
61.00, 52.10; ESI-MS 326 m/z [MH]+, C15H14Cl2NO3 requires 326.
Representative Procedure for the Mitsunobu Coupling of Mono-
methyl Poly(ethylene glycol)s to Methyl 2-(4-Hydroxy-3,5-dichlo-
rophenylamino)benzoate (4b). Synthesis of Methyl 2-(3,5-Dichloro-
4-(2-methoxyethoxy)phenylamino)benzoate (4d).
Experimental Section
General Synthetic Methods.
Unless otherwise stated, all chemicals were purchased from com-
mercial suppliers and used without further purification. Reaction
progress was monitored by thin-layer chromatography on silica gel 60
F254-coated glass plates (EM Sciences) and/or analytical RP-HPLC.
All flash chromatography was performed using 230-400 mesh silica
gel 60 (EM Sciences). NMR spectra were recorded on either a Bruker
500 or 600 MHz spectrometer. Chemical shifts are reported in parts
per million downfield from the internal standard Me4Si (0.0 ppm) for
CDCl3 solutions, or in the case where the Me4Si was not seen in the
13C NMR spectra, calibration was done on the solvent peaks (CDCl3
77.16 ppm). For samples in d6-DMSO, d6-acetone, or CD3OD,
calibration was done on the solvent peak at 2.49, 2.05, and 3.31 ppm,
1
respectively, for H NMR and 39.52, 29.84, and 49.00 ppm, respec-
tively, for 13C NMR. Reverse-phase high-performance liquid chroma-
tography (RP-HPLC) was carried out on a Waters 600 E multisolvent
delivery system employing a Waters 486 tunable absorbance detector
and a Waters 717 autosampler. A ThermoHypersil-Keystone Betabasic-
18 column was used for analytical reverse-phase HPLC analyses (model
71503-034630, 150 Å pore size, 3 µm particle size). Solvent system A
was 95:5 H2O:CH3CN with 0.25% trifluoroacetic acid, and solvent B
was 5:95 H2O:CH3CN with 0.25% trifluoroacetic acid; linear gradients
were run from either 0:100, 80:20, or 60:40 A:B to 0:100 A:B. All
mass spectrometry data were collected at the Scripps Research Institute
Center for Mass Spectrometry.
4-Benzyloxy-3,5-dichloroaniline (3).
Benzyl bromide (3.10 mL, 25.9 mmol) was added slowly to a stirring
mixture of potassium carbonate (3.95 g, 28.6 mmol) and 4-amino-2,6-
dichlorophenol (5.06 g, 28.4 mmol) in acetone. After stirring for 20 h
9
5548 J. AM. CHEM. SOC. VOL. 127, NO. 15, 2005

Kinetic Stabilization of an Oligomeric Protein
A R T I C L E S
Diisopropyl azodicarboxylate (40.0 µL, 0.202 mmol) was added
dropwise to a stirring solution of 2-(4-hydroxy-3,5-dichlorophenylami-
no)benzoate (4b) (63.0 mg, 0.202 mmol), 2-methoxyethanol (16.0 µL,
0.203 mmol), and triphenylphosphine (58.5 mg, 0.223 mmol) in
anhydrous THF (2.0 mL) under an argon atmosphere. After 24 h the
reaction was concentrated, and flash chromatographic purification over
silica (9:1-2:1 hexanes:EtOAc gradient elution) afforded 4d as a pale
yellow syrup (66.2 mg, 89%). 1H NMR (500 MHz, CDCl3) δ 9.42 (s,
1H), 7.97 (dd, J ) 1.6, 8.0 Hz, 1H), 7.38 (apparent ddd, J ) 1.6, 7.1,
8.5 Hz, 1H), 7.21 (dd, J ) 0.7, 8.5 Hz, 1H), 7.18, (s, 2H), 6.81 (apparent
ddd, J ) 1.0, 7.1, 8.0 Hz, 1H), 4.16-4.19 (m, 2H), 3.90 (s, 3H), 3.79-
3.82 (m, 2H), 3.48 (s, 3H); 13C NMR (125 MHz, CDCl3) δ 168.92,
147.21, 146.82, 138.13, 134.45, 131.88, 129.88, 122.11, 118.46, 114.54,
112.97, 72.67, 71.72, 59.32, 52.10; ESI-MS 370 m/z [MH]+, C17H18-
Cl2NO4 requires 370, 392 m/z [MNa]+, C17H17Cl2NaNO4 requires 392.
Mitsunobu procedures and characterization data for compounds 4e
and 4f are presented in the Supporting Information.
Bromination procedures and characterization data for compounds
7ii-vi are presented in the Supporting Information.
Representative Procedure for the Nucleophilic Coupling of
Thioacetic Acid to Bromopoly(ethylene glycol) Benzoates. Synthesis
of Acetylsulfanylethylene Glycol Benzoate (8i).
Thioacetic acid (75.0 µL, 1.05 mmol) was added to a stirring solution
of sodium hydride (41.8 mg, 1.05 mmol) in anhydrous DMF (5 mL)
under an argon atmosphere. After 30 min the 7i (219.2 mg, 0.957 mmol
in 5 mL of anhydrous DMF) was added, the reaction was stirred for
30 min, and then diluted with EtOAc (100 mL), washed with water (3
× 25 mL) and brine (25 mL), dried over Na2SO4, filtered, and
concentrated. Flash chromatographic purification over silica (9:1
hexanes:EtOAc) afforded acetylsulfanylethylene glycol benzoate (8i)
as a clear, colorless liquid (161.2 mg, 75%). 1H NMR (500 MHz,
CDCl3) δ 8.02-8.05 (m, 2H), 7.57 (tt, J ) 1.5, 7.5 Hz, 1H), 7.43-
7.47 (m, 2H), 4.43 (t, J ) 6.4 Hz, 2H), 3.28 (t, J ) 6.4 Hz, 2H), 2.37
(s, 3H); 13C NMR (125 MHz, CDCl3) δ 195.05, 166.35, 133.25, 129.94,
129.78, 128.53, 63.33, 30.69, 28.10; ESI-MS 247 m/z [MNa]+, C11H12-
NaO3S requires 247.
Representative Procedure for the Hydrolysis of Methyl Esters
4a-f. Synthesis of 2-(3,5-Dichloro-4-benzyloxyphenylamino)benzoic
Acid (5a).
Thioacetic acid nucleophilic substitution procedures and character-
ization data for compounds 8ii-vi are presented in the Supporting
Information.
LiOH‚H2O (1.33 g, 31.7 mmol) was added to a stirring solution of
4a (3.16 g, 7.86 mmol) in THF:MeOH:H2O (18:6:6 mL). After 18 h
the reaction was diluted with H2O (200 mL) and acidified to pH ≈ 1
with 30% HCl, whereupon the precipitate was filtered, washed with
H2O, collected, and dried to yield 5a as a pale yellow powder (3.05 g,
100%). 1H NMR (500 MHz, d6-DMSO) δ 9.52 (br s, 1H), 7.91 (dd, J
) 1.7, 8.0 Hz, 1H), 7.51-7.54 (m, 2H), 7.36-7.48 (m, 6H), 7.23 (dd,
J ) 0.7, 8.3 Hz, 1H), 6.88 (apparent ddd, J ) 0.9, 7.1, 8.0 Hz, 1H),
4.97 (s, 2H); 13C NMR (125 MHz, d6-DMSO) δ 169.53, 145.42, 145.18,
138.82, 136.24, 134.25, 131.89, 129.09, 128.45, 128.42, 128.39, 120.71,
119.01, 115.20, 114.48, 74.79; ESI-MS 388 m/z [MH]+, C20H16Cl2-
NO3 requires 388, 410 m/z [MNa]+, C20H15Cl2NaNO3 requires 410.
Hydrolysis procedures and characterization data for compounds 5b-f
are presented in the Supporting Information.
Representative Procedure for the Deprotection of Acetylsulfa-
nylpoly(ethylene glycol) Benzoates and Masking of Thiols as
2-Mercaptopyridine Disulfides. Synthesis of Masked Thiol 9i.
Sodium methoxide (1.40 mL of 0.5 M in MeOH, 0.70 mmol) was
added dropwise to a stirring solution of 8i (70.9 mg, 0.316 mmol) and
2,2′-dithiodipyridine (77.4 mg, 0.351 mmol) in anhydrous methanol
(3 mL) under an argon atmosphere. After 2 h the reaction was
concentrated with silica to a powder, and the crude product was purified
by flash chromatography over silica (1:1 hexanes:EtOAc) to afford 9i
as a clear, pale yellow liquid (26.3 mg, 44%). Alternatively, 2,2′-
dithiodipyridine (2.03 g, 9.21 mmol) was added to a stirring solution
of 2-mercaptoethanol (0.65 mL, 9.2 mmol) in anhydrous methanol (90
mL) under an argon atmosphere. After 1 h the reaction was concentrated
with silica to a powder, and the crude product was purified by flash
chromatography over silica (1:1 hexanes:EtOAc) to afford 9i as a clear,
Representative Procedure for the Monoprotection of Poly-
(ethylene glycol)s. Synthesis of Ethylene Glycol Monobenzoate (6i).
Benzoyl chloride (2.90 mL, 25.0 mmol) was added slowly to a
stirring solution of ethylene glycol (4.20 mL, 75.3 mmol) and pyridine
(2.25 mL, 27.6 mmol) in 25 mL of anhydrous dichloromethane (DCM)
at 0 °C under an argon atmosphere. After stirring for an additional 24
h at room temperature the reaction was then diluted with ethyl acetate
(200 mL), washed with water (3 × 50 mL) and brine (50 mL), dried
over Na2SO4, filtered, and concentrated. Flash chromatographic puri-
fication over silica (2:1-1:1 hexanes:EtOAc gradient elution) afforded
ethylene glycol monobenzoate (6i) as a clear, colorless liquid (3.30 g,
1
pale yellow liquid (1.15 g mg, 67%). H NMR (500 MHz, CDCl3) δ
8.51 (ddd, J ) 0.9, 1.8, 5.0 Hz, 1H), 7.59 (apparent dt, J ) 1.8, 7.7
Hz, 1H), 7.41 (dt, J ) 0.9, 8.1 Hz, 1H), 7.16 (ddd, J ) 0.9, 5.0, 7.3
Hz, 1H), 5.75 (t, J ) 7.0 Hz, 1H), 3.78-3.83 (m, 2H), 2.94-2.98 (m,
2H); 13C NMR (125 MHz, CDCl3) δ 159.21, 149.97, 136.97, 122.06,
121.64, 58.32, 42.80; ESI-MS 188 m/z [MH]+, C7H10OS2 requires 188,
210 m/z [MNa]+, C7H9NaOS2 requires 210.
2-Mercaptopyridine disulfide protection procedures and characteriza-
tion data for compounds 9ii-vi are presented in the Supporting
Information.
1
79%). H NMR (500 MHz, CDCl3) δ 8.04-8.08 (m, 2H), 7.55-7.60
(m, 1H), 7.42-7.47 (m, 2H), 4.45-4.48 (m, 2H), 3.96 (br s, 2H), 2.31
(br s, 1H); 13C NMR (125 MHz, CDCl3) δ 166.99, 133.20, 129.84,
129.69, 128.42, 66.67, 61.38.
Monoprotection procedures and characterization data for compounds
6ii-vi are presented in the Supporting Information.
Representative Procedure for the Bromination of Poly(ethylene
glycol) Monobenzoates. Synthesis of Bromoethylene Glycol Benzoate
(7i).
tert-Butyl 2-Bromobenzoate (10).
tert-Butyl alcohol (1.80 g, 24.2 mmol) in anhydrous dichloromethane
(10 mL) was added to a stirring mixture of 2-bromobenzoic acid (2.02
g, 10.0 mmol), N,N′-dicyclohexylcarbodiimide (2.49 g, 12.1 mmol),
and N,N-(dimethylamino)pyridine (123 mg, 1.01 mmol) in anhydrous
dichloromethane (40 mL) under an argon atmosphere. After stirring
for 24 h the reaction was concentrated and chromatographed over silica
(9:1-4:1 hexanes:EtOAc gradient elution), affording tert-butyl 2-bro-
Triphenylphosphine (5.99 g, 22.8 mmol in 45 mL of anhydrous
DCM) was added slowly to a stirring solution of N-bromosuccinimide
(4.05 g, 22.8 mmol in 100 mL of anhydrous DCM) at -78 °C under
an argon atmosphere. After 10 min the 6i (3.16 g, 19.0 mmol in 45
mL of anhydrous DCM) was added slowly, and then the reaction was
left stir while warming to room temperature. After 2 h the reaction
was concentrated with silica to a powder. Flash chromatographic
purification over silica (4:1 hexanes:EtOAc) afforded bromoethylene
1
mobenzoate (10) as a clear, colorless liquid (1.57 g, 61%). H NMR
(500 MHz, CDCl3) δ 7.68 (dd, J ) 1.8, 7.6 Hz, 1H), 7.61 (dd, J )
1.2, 7.9 Hz, 1H), 7.33 (dt, J ) 1.2, 7.6 Hz, 1H), 7.27 (dt, J ) 1.8, 7.9
Hz), 1.61 (s, 9H); 13C NMR (125 MHz, CDCl3) δ 165.83, 134.43,
134.14, 131.97, 130.92, 127.20, 121.10, 82.68, 28.26; GC-MS 256/
258 m/z [M]+, C11H13BrO2 requires 256/258, 183/185 m/z [M -
C4H9O]+, C7H4BrO requires 183/185.
tert-Butyl 2-(4-Benzyloxy-3,5-dichlorophenylamino)benzoate (11).
Toluene (11.0 mL) was added to a flask charged with 10 (1.76 g,
6.56 mmol), 3 (1.40 g, 5.44 mmol), tris(dibenzylideneacetone)-
dipalladium(0) (533 mg, 0.857 mmol), (R)-(+)-2,2′-bis)diphenylphos-
phino)-1,1′-binaphthyl (109 mg, 0.175 mmol), and cesium carbonate
1
glycol benzoate (7i) as a clear, pale yellow liquid (4.30 g, 99%). H
NMR (500 MHz, CDCl3) δ 8.06-8.09 (m, 2H), 7.58 (apparent tt, J )
1.4, 7.5 Hz, 1H), 7.44-7.48 (m, 2H), 4.63 (t, J ) 6.1 Hz, 2H), 3.65 (t,
J ) 6.1 Hz, 2H); 13C NMR (125 MHz, CDCl3) δ 166.08, 133.29,
129.75, 129.61, 128.46, 64.21, 28.80.
9
J. AM. CHEM. SOC. VOL. 127, NO. 15, 2005 5549

A R T I C L E S
Wiseman et al.
(2.48 g, 7.61 mmol) and then stirred at reflux under an argon
atmosphere. After 24 h the reaction was filtered through Celite and
concentrated. Flash chromatographic purification over silica (95:5
hexanes:EtOAc) afforded tert-butyl 2-(4-benzyloxy-3,5-dichlorophe-
nylamino)benzoate (11) as a pale yellow solid (1.40 g, 58%). 1H NMR
(600 MHz, CDCl3) δ 9.57 (s, 1H), 7.93 (dd, J ) 1.6, 8.0 Hz, 1H), 7.57
(apparent d, J ) 7.6 Hz, 2H), 7.39-7.43 (m, 2H), 7.33-7.38 (m, 2H),
7.23 (apparent d, J ) 8.4 Hz, 1H), 7.21 (s, 2H), 6.78-6.82 (m, 1H),
5.02 (s, 2H), 1.60 (s, 9H); 13C NMR (150 MHz, CDCl3) δ 168.04,
146.61, 146.50, 138.58, 136.58, 133.91, 132.11, 130.12, 128.67, 128.62,
128.52, 121.72, 118.38, 114.75, 114.57, 81.78, 75.24, 28.42; ESI-MS
478 m/z [M + Cl]-, C24H23Cl3NO3 requires 478.
tert-Butyl 2-(4-Hydroxy-3,5-dichlorophenylamino)benzoate (12).
Ten percent palladium on carbon (158 mg, 0.148 mmol) and 11 (1.32
g, 2.97 mmol) were stirred in a 1:1 mixture of MeOH:EtOAc (16 mL)
under a hydrogen atmosphere for 5 h. The atmosphere was then purged
with argon, and the reaction was quenched with DCM, filtered through
Celite, and concentrated. Flash chromatographic purification over silica
(9:1-4:1 hexanes:EtOAc gradient elution) afforded tert-butyl 2-(4-
hydroxy-3,5-dichlorophenylamino)benzoate (12) as a pale yellow solid
(932 mg, 89%). 1H NMR (500 MHz, CDCl3) δ 9.44 (s, 1H), 7.91 (dd,
J ) 1.6, 8.0 Hz, 1H), 7.31 (apparent ddd, J ) 1.6, 7.1, 8.5 Hz, 1H),
7.24 (s, 2H), 7.05 (dd, J ) 0.8, 8.5 Hz, 1H), 6.74 (apparent ddd, J )
0.9, 7.2, 8.0 Hz, 1H), 5.67 (s, 1H), 1.60 (s, 9H); 13C NMR (125 MHz,
CDCl3) δ 168.15, 147.85, 144.37, 134.68, 133.96, 132.06, 123.39,
121.43, 117.60, 113.84, 113.79, 81.61, 28.44; ESI-MS 354 m/z [MH]+,
C17H18Cl2NO3 requires 354, 376 m/z [MNa]+, C17H17Cl2NaNO3 requires
376.
Expression and Purification of (flag-tag WT TTR)1(C10A)3.
C10A TTR and flag-tag WT TTR were co-transformed into BL21
(DE3) Epicurian Gold cells (Stratagene) with both the C10A TTR/
pMMHa and the flag-tag WT TTR/pet29a expression vectors. Selection
was performed by overnight growth (37 °C) on agar plates containing
150 µg/mL kanamycin and 100 µg/mL ampicillin. Starter cultures were
prepared by the inoculation of LB media (150 µg/mL kanamycin, 100
µg/mL ampicillin) with a single colony from the agar plate. Starter
cultures were grown at 37 °C, with shaking, until bacteria growth was
observed. Expression cultures (LB media with 150 µg/mL kanamycin
and 100 µg/mL ampicillin) were then inoculated with a 1:100 dilution
of the starter cultures, grown at 37 °C until OD600 ≈ 0.8, and induced
by addition of 1 mM IPTG. Induction was allowed to proceed overnight
at 37 °C with shaking. Cells were then harvested by centrifugation
(14 000g, 10 min), resuspended in water, frozen at -80 °C (1 h), and
thawed at 37 °C (1 h). Lysates were prepared by sonication (three cycles
of 3 min) at 4 °C. Cell debris was pelleted by centrifugation for 30
min at 15 000g. The lysate was fractionated by ammonium sulfate
precipitation, collecting the pellet of the 50-90% precipitation. This
pellet was resuspended in a minimum amount of 25 mM Tris, pH 8.0,
1 mM EDTA and dialyzed overnight (4 °C) against 25 mM Tris, pH
8.0, 1 mM EDTA using 7000 MWCO dialysis tubing (Snakeskin from
Pierce Biomedical). The dialysis product was then purified using a
Source 15Q anion-exchange column (Amersham Biosciences) running
a linear gradient from 200 to 450 mM NaCl in 25 mM Tris, pH 8.0, 1
mM EDTA. Fractions containing the various stoichiometries of C10A
and flag-tag WT TTR were pooled and concentrated. The samples were
then purified by gel filtration chromatography with a Superdex 75
column (Amersham Biosciences), eluting the sample with 10 mM
sodium phosphate (pH 7.2), 100 mM KCl, 1 mM EDTA. The tetrameric
peak was collected, and the (flag-tag WT TTR)1(C10A)3 was isolated
from other stoichiometries by anion-exchange chromatography with a
Source 15Q column (Amersham Biosciences) using the same gradient
as above. Protein concentration of (flag-tag WT TTR)1(C10A)3 was
determined using the molar absorptivity of ꢀ ) 75 829 M-1 cm-1. The
purified heterotetramer was then stored at 37 °C until modification with
small molecule.
Representative Procedure for the Mitsunobu Coupling of Acti-
vated Thiol Linkers to tert-Butyl 2-(4-Hydroxy-3,5-dichloropheny-
lamino)benzoate (12). Synthesis of 13i.
Diisopropyl azodicarboxylate (58.0 µL, 0.293 mmol) was added
dropwise to a stirring solution of triphenylphosphine (76.6 mg, 0.292
mmol) in anhydrous THF (5.0 mL) under an argon atmosphere. After
30 min 9i (54.5 mg, 0.291 mmol) and 12 (51.7 mg, 0.146 mmol) in
anhydrous THF (5.0 mL) were added to the DIAD/PPh3 mixture, and
the reaction was stirred for 1 h and then concentrated. Flash chromato-
graphic purification over silica (9:1-4:1 hexanes:EtOAc gradient
Modification of (flag-tag WT TTR)1(C10A)3 with Activated Small
Molecules.
1
elution) afforded 13i as a clear, colorless syrup (69.9 mg, 91%). H
NMR (500 MHz, CDCl3) δ 9.55 (s, 1H), 8.48 (ddd, J ) 0.7, 1.7, 4.8
Hz, 1H), 7.92 (dd, J ) 1.6, 8.0 Hz, 1H), 7.77 (apparent dt, J ) 1.0,
8.1 Hz, 1H), 7.66 (ddd, J ) 1.8, 7.4, 8.0 Hz, 1H), 7.35 (apparent ddd,
J ) 1.6, 7.1, 8.5 Hz, 1H), 7.22 (dd, J ) 0.9, 8.5 Hz, 1H), 7.17 (s, 2H),
7.10 (ddd, J ) 1.0, 4.8, 7.4 Hz, 1H), 6.80 (apparent ddd, J ) 0.9, 7.2,
8.0 Hz, 1H), 4.29 (t, J ) 6.7 Hz, 2H), 3.25 (t, J ) 6.7 Hz, 2H), 1.60
(s, 9H); 13C NMR (125 MHz, CDCl3) δ 168.01, 160.07, 149.85, 146.51,
146.45, 138.64, 137.26, 133.91, 132.11, 129.75, 121.60, 120.91, 119.99,
118.45, 114.82, 114.60, 81.82, 71.61, 38.50, 28.42; ESI-MS 523 m/z
[MH]+, C24H25Cl2N2O3S2 requires 523.
Mitsunobu procedures and characterization data for compounds 13ii-
vi are presented in the Supporting Information.
Representative Procedure for the Deprotection of tert-Butyl
Esters 13i-vi. Synthesis of 1i.
Compound 13i (62.9 mg, 0.120 mmol) was stirred in trifluoroacetic
acid (5.0 mL) for 3 h, and then it was concentrated, diluted with water,
and lyophilized to afford 1i as a grayish solid in quantitative yield. 1H
NMR (600 MHz, d6-DMSO) δ 13.20 (br s, 1H), 9.48 (s, 1H), 8.45-
8.47 (m, 1H), 7.90 (dd, J ) 1.6, 7.9 Hz, 1H), 7.78-7.85 (m, 2H), 7.45
(apparent ddd, J ) 1.6, 7.1, 8.4 Hz, 1H), 7.33 (s, 2H), 7.23-7.26 (m,
2H), 6.86-6.90 (m, 1H), 4.21 (t, J ) 6.2 Hz, 2H), 3.26 (t, J ) 6.2 Hz,
2H); 13C NMR (150 MHz, d6-DMSO) δ 169.50, 158.86, 149.66, 145.36,
145.17, 138.83, 137.87, 134.24, 131.86, 128.73, 121.34, 120.63, 119.50,
119.03, 115.25, 114.48, 71.22, 38.00; MALDI-FTMS (DHB) 467.0049
m/z [MH]+, C20H17Cl2N2O3S2 requires 467.0052.
(flag-tag WT TTR)1(C10A)3 purified above was treated with either
10× activated small molecule (prepared as 10 mM Stock in DMSO)
or DMSO for 2 h at room temperature with gentle agitation. The
samples were then purified from free small molecules by gel filtration
using a Superdex 200 column (Amersham Biosciences) eluting the
protein with 10 mM sodium phosphate (pH 7.2), 100 mM KCl, 1 mM
EDTA. Purified samples were analyzed with a Waters HPLC using a
C8 column (Western Analytical Products, Inc.; 30 × 4.6 mm, 3 µm
particle size, and 150 Å pore size) using a linear gradient from 45% to
95% acetonitrile in the presence and absence of 1 mM dithiothreitol
(DTT) to ensure modification. LC-ESI-MS (Hewlett-Packard 1100-
MSD mass spectrometer) analysis was also performed running a
gradient of 15-75% acetonitrile over 10 min, demonstrating that only
the flag-tag WT TTR peak was modified by disulfide formation (C10A,
13 860 ( 1 kDa; flag-tag WT TTR, 15 880 ( 1 kDa; TTR-1i′, 16 237
( 1 kDa; TTR-1ii′, 16 281 ( 1 kDa; TTR-1iii′, 16 325 ( 1 kDa;
TTR-1iv′, 16 369 ( 1 kDa; TTR-1v′, 16 413 ( 1 kDa; TTR-1vi′,
16 457 ( 1 kDa). The molar absorptivity of the modified protein was
calculated by determining the concentration of the C10A peak on the
HPLC trace. The concentration of C10A was determined upon
comparison with a standard curve. Because the stoichiometry of C10A
to modified flag-tag WT TTR subunits is known, the molar absorptivity
can be calculated from the A280 of the stock solution of modified protein.
Structural Analysis of Modified Heterotetramers.
The tertiary structure of the modified heterotetramers was analyzed
using an Aviv model 202SF circular dichroism (CD) spectrometer. The
modified heterotetramers (1.8 µM tetramer; 10 mM sodium phosphate
tert-Butyl deprotection procedures and characterization data for
compounds 1ii-vi are presented in the Supporting Information.
9
5550 J. AM. CHEM. SOC. VOL. 127, NO. 15, 2005

Kinetic Stabilization of an Oligomeric Protein
A R T I C L E S
pH 7.2, 100 mM KCl, 1 mM EDTA) in a 2 mm cuvette were scanned
from 250 to 200 nm (0.5 nm intervals; 25 °C; 2 s averaging time)
averaging five sequential scans to determine the final CD plot.
The quaternary structure of the modified heterotetramers was
analyzed using analytical ultracentrifugation. Analytical ultracentrifu-
gation was performed on a Beckman XL-I analytical ultracentrifuge
using both interference and absorbance optics. AUC cells were
constructed with sapphire windows, 1.2 mm charcoal-filled Epon
centerpieces, and used with either an An60 Ti four-hole or an An50 Ti
eight-hole rotor. Samples for velocity experiments were allowed to
equilibrate to 20 °C for at least 1 h prior to being spun at 50 000 rpm
for 4 h, recording absorbance and interference spectra as quickly as
instrumentation allowed (ca. every 2-10 min). Scans were analyzed
using the c(s) method in SedFit.48
Coordinates and structure factors have been deposited in the Protein
Data Bank under the accession code 1U21.
Acid-Induced Aggregation of WT TTR Treated with Compounds
5a-f.
WT TTR (7.2 µM; 10 mM sodium phosphate, pH 7.2, 100 mM
KCl, 1 mM EDTA) was incubated with small molecules 5a-f (14.4
µM tetramer) for 30 min at room temperature. The protein samples
were diluted 1:1 with acidification buffer containing 100 mM sodium
acetate (pH 4.2), 100 mM KCl, 1 mM EDTA. Acidified samples were
incubated at 37 °C for 72 h without stirring. The amount of aggregation
was analyzed by turbidity measurements (25 °C) at 400 nm on a
Hewlett-Packard 8453 UV-visible spectrometer equipped with a Peltier
temperature-controlled cell holder.
Acid-Mediated Aggregation of Modified Heterotetramers.
Crystallization and Structure Determination of TTR-1iv′.
TTR-1iv′ (6.0 mg mL-1) was crystallized in 3 days from 20% PEG
4K and 0.2 M CaCl2‚2H2O at room temperature using the sitting-drop
vapor diffusion method. Crystals were harvested from drops with nylon
cryoloops (Hampton Research) and directly cryocooled to -180 °C in
cryoprotectant that consisted of well buffer and glycerol (25%). A 1.69
Å native data set was collected on a single crystal at beamline 9-1 at
the Stanford Synchrotron Radiation Laboratory (SSRL) in Menlo Park,
CA (Table 2). Data were processed and scaled in HKL2000.49 The
space group is orthorhombic P21212 with unit cell dimensions a ) 43.65
Å, b ) 86.64 Å, and c ) 63.93 Å and two monomers per asymmetric
unit (Vm ) 2.27 Å3/Da) with a solvent content of 45.8%.
Heterotetramers (7.2 µM; 10 mM sodium phosphate, pH 7.2, 100
mM KCl, 1 mM EDTA) were diluted 1:1 in acidification buffers (pH
3.2-3.6 100 mM sodium citrate, 100 mM KCl, 1 mM EDTA; pH 4.0-
5.6 100 mM sodium acetate, 100 mM KCl, 1 mM EDTA; pH 6.0-6.4
100 mM sodium phosphate, 100 mM KCl, 1 mM EDTA). Samples
were incubated at 37 °C for 72 h. The extent of aggregation was
measured by turbidity at 400 nm on an HP 8453 UV-visible
spectrometer equipped with a Peltier temperature-controlled cell holder.
Reduced samples were prepared by addition of 1 mM DTT to
heterotetrameric samples diluted 1:1 into 100 mM sodium acetate (pH
4.2), 100 mM KCl, 1 mM EDTA and treated as above.
Kinetic Urea Denaturation of Modified Heterotetramers.
Diffraction data collected in this experiment were isomorphous with
those of a previous structural study of transthyretin (pdb code
Modified heterotetramers (7.2 µM tetramer) were diluted 1:4 with
10 M urea (50 mM sodium phosphate, 100 mM KCl, 1 mM EDTA)
and buffer (50 mM sodium phosphate, 100 mM KCl, 1 mM EDTA) to
a final solution of 1.8 µM tetrameric protein and 6.0 M urea. Samples
were mixed and then immediately measured on an Aviv model 202SF
circular dichroism (CD) spectrometer measuring the CD signal at 215
nm (5.0 s averaging time). The samples were incubated at 25 °C, taking
CD measurements at the indicated time points. Plots were prepared
using Kaleidograph fitting the denaturation to a first-order exponential
function with the equation I ) Io + A(1 - e-kt), where I ) CD signal,
Io ) initial CD signal, A ) amplitude, k ) dissociation rate constant,
and t ) time. Reduced samples were prepared by the addition of 1
mM DTT to the initial sample preparation and treated as above.
33
1BZD.pdb). An initial model was generated by performing a rigid
body refinement and simulated annealing (30-4.00 Å) (starting Rcryst
) 55.1%, Rfree ) 46.5%; final Rcryst ) 32.7%, Rfree ) 33.8%) in CNS
50 using the previous model (1BZD.pdb) against the native data. Initial
electron density maps were of excellent quality. Refinement to 1.69 Å
was performed in CNS.50 The model was further rebuilt into σA-
weighted 3Fo - 2Fc and Fo - Fc electron density maps in O.51 Water
50
molecules were assigned automatically in CNS
at >3σ Fo - Fc
difference density peaks and verified by manual inspection in O.51 After
convergence in CNS (Rcryst ) 24.6%, Rfree ) 26.3%), TLS refinement
was continued with REFMAC552 with riding hydrogens included. Care
was taken to keep the test set intact during all refinement steps. The
energy-minimized ligand 1iv′ was built with InsightII (Molecular
Simulations, Inc.) and unambiguously placed in the electron density.
Both symmetry-related binding conformations of the ligand were in
good agreement with unbiased 3Fo - 2Fc simulated annealing-omit
maps calculated in the absence of the inhibitor. The final model (Rcryst
) 21.7%, Rfree ) 23.9%; Table 1) is composed of all residues except
the first nine and last three of the wild-type sequence and the 16-residue
purification tag attached to one monomer, which showed no interpret-
able electron density, and includes 107 water molecules. The quality
of the model was analyzed using the programs Mol Probity,53 WHAT
IF,54 and PROCHECK.55 Figure 5 was made using bobscript56 and raster
3D.57
Acknowledgment. This work was supported in part by NIH
Grants DK46335 (J.W.K.) and CA58896 and AI42266 (I.A.W.),
the Skaggs Institute for Chemical Biology, the Lita Annenberg
Hazen Foundation (J.W.K.), an NIH predoctoral training grant
T32 AI077606, a Skaggs predoctoral fellowship, David and
Ursula Fairchild Fellowship (T.F.), the Jairo H. Are´valo
Fellowship (M.S.K.), the Norton B. Gilula Fellowship, and the
Fletcher Jones Fellowship (R.L.W.). We thank Dr. L. Xhu for
data collection and the SSRL staff of beamline 9.0.1 for
guidance during data collection. We thank Drs. Evan T. Powers
and Patrick Braun for helpful discussions.
(48) Schuck, P. Biophys. J. 2000, 78, 1606-1619.
(49) Otwinowski, Z.; Minor, W. Methods Enzymol. 1997, 276, 307-326.
(50) Brunger, A. T.; Adams, P. D.; Clore, G. M.; DeLano, W. L.; Gros, P.;
Grosse-Kunstleve, R. W.; Jiang, J. S.; Kuszewski, J.; Nilges, M.; Pannu,
N. S.; Read, R. J.; Rice, L. M.; Simonson, T.; Warren, G. L. Acta
Crystallogr., D.: Biol. Crystallogr. 1998, 54 (Pt 5), 905-921.
(51) Jones, T. A.; Zou, J. Y.; Cowan, S. W.; Kjeldgaard Acta Crystallogr., A
1991, 47 (Pt 2), 110-119.
Supporting Information Available: Detailed synthesis and
characterization of compounds 1-13, representative sedimenta-
tion velocity analytical ultracentrifuge trace of modified (flag-
tag WT)1(C10A)3 (TTR-1iv′), HPLC traces of redissolved
pellets collected from treatment of TTR-1v′ in acid (pH 4.0,
72 h, 37 °C), and kinetic dissociation plot of TTR-1iv′ with
and without an acid pretreatment (pH 4.2, 72 h, 37 °C). This
material is available free of charge via the Internet at
http://pubs.acs.org.
(52) Winn, M. D.; Isupov, M. N.; Murshudov, G. N. Acta Crystallogr., D: Biol.
Crystallogr. 2001, 57, 122-133.
(53) Lovell, S. C.; Davis, I. W.; Arendall, W. B., III; de Bakker, P. I.; Word, J.
M.; Prisant, M. G.; Richardson, J. S.; Richardson, D. C. Proteins 2003,
50, 437-450.
(54) Vriend, G. J. Mol. Graph. 1990, 8, 52-56, 29.
(55) Laskowski, R.; MacArthur, M.; Moss, D.; Thornton, J. J. Appl. Crystallogr.
1993, 26, 283-291.
(56) Esnouf, R. M. J. Mol. Graph. Model. 1997, 15, 132-134.
(57) Merritt, E. A.; Bacon, D. J. Methods Enzymol. 1997, 277, 505-524.
JA042929F
9
J. AM. CHEM. SOC. VOL. 127, NO. 15, 2005 5551