Synthesis and conformational preferences of a potential β-sheet nucleator based on the 9,9-dimethylxanthene skeleton

Article abstract of DOI:10.1021/jo9605758

A 4,5-disubstituted-9,9-dimethylxanthene-based amino acid (10) has been synthesized for incorporation into peptide sequences which have a propensity to adopt β-sheet structure. Molecular dynamics studies support the FT-IR and NMR results which demonstrate that amides based on this residue utilize the NH and the C=O from the xanthene residue to form an intramolecular hydrogen bond (13-membered ring), unlike the previously studied dibenzofuran-based amino acid residues in which the NH and the C=O of the attached amide groups participate in intramolecular hydrogen bonding (15-membered ring). Interestingly, residue 10 derivatized as a simple amide prefers to adopt a trans conformation where the aliphatic side chains are placed on opposite sides of the plane of the 9,9-dimethylxanthene ring system. This is different than the conformational preferences of the dibenzofuran-based amino acids which adopt a cis conformation that is preorganized to nucleate β-sheet formation. It will be interesting to see how these conformational differences effect nucleation in aqueous solution.

Full text of DOI:10.1021/jo9605758

7408
J . Org. Chem. 1996, 61, 7408 7414
Syn th esis a n d Con for m a tion a l P r efer en ces of a P oten tia l â-Sh eet
Nu clea tor Ba sed on th e 9,9-Dim eth ylxa n th en e Sk eleton
Kurt McWilliams and J effery W. Kelly*
Department of Chemistry, Texas A&M University, College Station, Texas, 77843-3255
Received March 27, 1996X
A 4,5-disubstituted-9,9-dimethylxanthene-based amino acid (10) has been synthesized for incorpora-
tion into peptide sequences which have a propensity to adopt â-sheet structure. Molecular dynamics
studies support the FT-IR and NMR results which demonstrate that amides based on this residue
utilize the NH and the CdO from the xanthene residue to form an intramolecular hydrogen bond
(13-membered ring), unlike the previously studied dibenzofuran-based amino acid residues in which
the NH and the CdO of the attached amide groups participate in intramolecular hydrogen bonding
(15-membered ring). Interestingly, residue 10 derivatized as a simple amide prefers to adopt a
trans conformation where the aliphatic side chains are placed on opposite sides of the plane of the
9,9-dimethylxanthene ring system. This is different than the conformational preferences of the
dibenzofuran-based amino acids which adopt a cis conformation that is preorganized to nucleate
â-sheet formation. It will be interesting to see how these conformational differences effect nucleation
in aqueous solution.
In tr od u ction
tors function by forming a hydrogen-bonded hydrophobic
cluster which preorganizes pendant peptide chains for
hydrogen bonding and subsequent â-sheet formation.
While these templates eliminated most of the entropic
penalty associated with chain reversal due to their U
shape, the divergent projection of the peptide strands in
the dibenzofuran-based peptides and the rotational isomer-
ization in the biphenyl- and bipryidyl-based templates
could limit nucleating efficacy. In designing a potentially
improved nucleator it was desirable to enhance the
characteristics associated with nucleation (hydrophobic
cluster formation, chain reversal, and appropriate in-
tramolecular hydrogen bonding) and modify the proper-
ties which possibly hinder â-sheet nucleation.
Unnatural amino acids have been widely used as
templates to induce protein-like secondary structure in
small peptides.1 For â-sheet nucleators, these templates
are generally conformationally restricted molecules, which
in addition to reversing the chain direction provide
stabilizing contacts with the attached -amino acid
15
sequence.2
These contacts include hydrophobic and/
or electrostatic interactions and have been used to
nucleate both parallel and antiparallel â-sheet structures.
Recently, this laboratory reported the synthesis of a
series of unnatural amino acids and diacids which are
capable of nucleating â-sheet structure in peptide se-
quences of less than 20 -amino acids.16
These nuclea-
19
An unnatural amino acid based on 9,9-dimethylxan-
thene (1) was chosen as the target because it offered
several potential advantages relative to the amino acids
previously prepared by this labortory. Unlike the planar
dibenzofuran skeleton, the incorporation of an isopropy-
* To whom correspondence should be addressed: phone 409-845-
4242; fax 409-845-9452; e-mail kelly@chemvx.tamu.ed.
X Abstract published in Advance ACS Abstracts, August 15, 1996.
(1) Schneider, J .; Kelly, J . Chem. Rev. 1995, 95, 2169 2187.
(2) Brandmeier, V.; Feigel, M.; Bremer, M. Angew. Chem., Int. Ed.
Engl. 1989, 28, 486 488.
(3) Brandmeier, V.; Feigel, M. Tetrahedron 1989, 45, 1365 1376.
(4) Brandmeier, V.; Sauer, W. H. B.; Feigel, M. Helv. Chim. Acta
1994, 77, 70 85.
4.7 Å
O
6
7
3
2
4
1
5
8
(5) Gardner, R. R.; Liang, G. B.; Gellman, S. H. J . Am. Chem. Soc.
1995, 117, 3280 3281.
9
(6) Kemp, D. S.; Bowen, B. R. Tetrahedron Lett. 1988, 29, 5081
5082.
1
(7) Kemp, D. S.; Bowen, B. R.; Muendel, C. C. J . Org. Chem. 1990,
55, 4650 4657.
lidene spacer at the nine position results in a nonplanar
xanthene ring system. This leads to the realization of a
range of accessible xanthene “puckering angles” which
could prove useful in optimizing packing within the
desired hydrophobic cluster.20 In addition to enhancing
the hydrophobic cluster through packing interactions,
hydrophobicity is also improved as a result of the
projection of one of the methyl groups of the isopropy-
lidene spacer into a void existing between the side chains
of the flanking -amino acid residues. Our expectation
is that the combination of these two effects should
compensate for any increase in the entropic penalty
associated with hydrophobic cluster formation as a result
of the increased flexibility of the xanthene skeleton.
Previous studies involving the rigid aromatic templates
(8) Kemp, D. S. Trends Biotechnol. 1990, 8, 249 255.
(9) Nowick, J . S.; Powell, N. A.; Martinez, E. J .; Smith, E. M.;
Noronha, G. J . Org. Chem. 1992, 57, 3763 3765.
(10) Nowick, J . S.; Abdi, M.; Bellamo, K. A.; Love, J . A.; Martinez,
E. J .; Smith, E. M.; Noronha, G.; Ziller, J . W. J . Am. Chem. Soc. 1995,
117, 88 99.
(11) Smith, A. B., III; Guzman, M. C.; Sprengeler, P. A.; Keenan, T.
P.; Holcomb, R.; Wood, J . L.; Carroll, P. J .; Hirschmann, R. J . J . Am.
Chem. Soc. 1994, 116, 9947 9962.
(12) Wagner, G.; Feigel, M. Tetrahedron 1993, 49, 10831 10842.
(13) Winningham, M. J .; Sogah, D. Y. J . Am. Chem. Soc. 1994, 116,
11173 11174.
(14) Nesloney, C.; Kelly, J . Bioorg. Med. Chem. 1996. In Press.
(15) Mueller, K.; Obrecht, D.; al., e. In Perspectives in Medicinal
Chemistry; Testa, B., Ed.; VCH: New York, 1994.
(16) Choo, D.; Schneider, J .; Graciani, N.; Kelly, J . Macromolecules
1996, 29, 355 366.
(17) Diaz, H.; Kelly, J . Tetrahedron Lett. 1991, 32, 5725 5728.
(18) Schneider, J .; Kelly, J . J . Am. Chem. Soc. 1995, 117, 2533
2546.
(19) Tsang, K.; Diaz, H.; Graciani, N.; Kelly, J . J . Am. Chem. Soc.
1994, 116, 3988 4005.
(20) Schaefer, T.; Sebastian, R. Can. J . Chem. 1990, 68, 1548.
S0022-3263(96)00575-0 CCC: $12.00 © 1996 American Chemical Society

Second Generation â-Sheet Nucleator
J . Org. Chem., Vol. 61, No. 21, 1996 7409
chains of the flanking hydrophobic -amino acids to
interact with the face of the hydrophobic xanthene ring
system. The placement of these spacers at the 4 and 5
positions of the 9,9-dimethylxanthene skeleton appears
optimal for sheet formation with attachment points 4.7
Å apart, which is close to the strand strand separation
observed in naturally occurring â-sheets (4.85 Å).
One anticipated difference between the previously
studied dibenzofuran-based system and this xanthene-
based system is in the preferred intramolecularly hydro-
gen-bonded conformations (vide infra). In dibenzofuran-
based simple amides, the CdO and the NH of attached
amide groups intramolecularly hydrogen bond affording
a 15-membered ring exclusively.19 This arises, in part,
from divergent projection of substituents from the fused
ring skeleton. For xanthene, substituents radiating from
the 4 and 5 positions are parallel with respect to each
other creating the possibility of a 13-membered intramo-
lecular hydrogen-bonded ring (eq 1). It will be interesting
to discern what effect these anticipated conformational
differences have on â-sheet nucleation in aqueous solu-
tion.
F igu r e 1. Ball-and-stick rendition of a representative low-
energy conformer of benzyl diamide 12, determined by molec-
ular dynamics/molecular mechanics. Arrows indicate hydro-
gens used for trans/cis conformer determination. Note that the
ethylene spacers adopt a trans conformation.
A
B
Owing to the potential of a 4,5-disubstituted-9,9-
dimethylxanthene-based amino acid, a convenient route
for its regioselective synthesis was therefore sought. The
synthesis of xanthene derivatives has previously been
accomplished utilizing substituted phenols in the Ull-
mann condensation to form the xanthone skeleton. In
some cases, subsequent electrophilic aromatic substi-
tution was necessary to further functionalize the skel-
eton. Unfortunately, the Ullmann condensation requires
strongly acidic conditions and is sensitive to the substitu-
F igu r e 2. Newman projections of trans and cis conformations
of the propionic acid and aminoethyl substructures within the
xanthene-based unnatural amino acid residue.
28
tion pattern of the phenol.23
The commercial avail-
ability of both xanthene and xanthone have made the
electrophilic aromatic substitution route popular. How-
ever, due to the enhanced reactivity of the 2 and 7
positions to classical electrophilic aromatic substitution
conditions blocking groups must be used at the 2 and 7
positions to achieve 4,5 electrophilic substitution.29 Since
have demonstrated the need for flexible ethylene linkers
between the aromatic ring system and the peptide
strands to achieve the desired “perpendicular” conforma-
tion associated with the hydrophobic cluster conforma-
tion. This perpendicular conformation refers to the
preferred perpendicular orientation of the aliphatic car-
bon carbon bond relative to the plane of the aromatic
ring (Figures 1 and 2).21,22 In a peptide, these CH2CH2
linkers should orient the xanthene ring system perpen-
dicular to the plane of the â-sheet, allowing the side
(23) Goldberg, A.; Walker, H. J . Chem. Soc. 1953, 1348 1357.
(24) Kimura, M.; Okabayashi, I. Chem. Pharm. Bull. 1987, 35, 136
141.
(25) Shi, J .; Zhang, X.; Neckers, D. J . Org. Chem. 1992, 57, 4418
4421.
(26) Rewcastle, G.; Atwell, G.; Baguley, B.; Calveley, S.; Denny, W.
J . Med. Chem. 1989, 32, 793 799.
(27) Rewcastle, G.; Atwell, G.; Zhuang, L.; Baguley, B.; Denny, W.
J . Med. Chem. 1991, 34, 217 222.
(28) Rewcastle, G.; Atwell, G.; Palmer, B.; Boyd, P.; Baguley, B.;
Denny, W. J . Med. Chem. 1991, 34, 491 496.
(21) Kriz, J .; J akes, J . J . Mol. Struct. 1972, 12, 367.
(22) Scharfenberg, P.; Rozsondai, B.; Hargittai, I. Naturforscher
1980, 35a, 431.

7410 J . Org. Chem., Vol. 61, No. 21, 1996
McWilliams and Kelly
Sch em e 1a
a
(a) 2 mol % Pd (OAc)2, 4 mol % P (Ph)3, TEA, ethyl acrylate, acetonitrile, 75 °C, (78%); (b) NaOH, ethanol, reflux, (95%); (c) 10% Pd/C,
H2, H2O, (93%); (d) 2.5 equiv of DIC, 2.5 equiv pentafluorophenol, ethyl acetate/DMF 8:3, (75%); (e) 4 mol % Pd (OAc)2, 2 mol % P (Ph)3,
ethyl acrylate, TEA, 75 °C, (25%); (f) 2 mol % Pd (OAc)2, 4 mol % P (Ph)3, TEA, acrylic acid, 75 °C, (67%); (g) 1.1 equiv diphenylphosphoryl
azide, TEA, toluene, reflux, 2.0 equiv of tert-butyl alcohol, (54%); (h) 10% Pd/C, 35 psi H2, ethanol, acetic acid, (97%); (i) LiOH H2O,
ethanol, 0 °C, (92%).
a tetrasubstituted product (2, 4, 5, and 7 positions) is
inconsistent with the design target, an alternative route
was pursued. Metalation studies on xanthone using
lithium 1,3-diaminopropane-N,N,N′,N′-d4 as the base and
deuterium oxide as the quencher have demonstrated that
deprotonation occurs very rapidly at the 4 and 5 posi-
tions.30 Therefore, this anionic strategy was adopted for
the regioselective synthesis of 4,5-disubstituted xan-
thenes and the subsequent synthesis of the unnatural
amino acid 10 and the potential parallel â-sheet nuclea-
tor, active diester 6 (Scheme 1).
A number of other iodine-quenching agents were
tested; however, these sources of iodine proved to be
problematic with regard to purification and/or product
yield.31 Carbon tetrabromide and tetrachloride quench-
ing agents gave better yields, but the resulting products
were for the most part unreactive to the cross-coupling
conditions employed in subsequent steps.32
The activated diester 6 was prepared by conversion of
the common intermediate 2 to the symmetrical diester 3
through a cross-coupling with ethyl acrylate (2.5 equiv)
in the presence of palladium acetate (2 mol %), triph-
enylphosphine (4 mol %), and triethyl amine at 75 °C in
acetonitrile.31,33 Diester 3 was saponified with NaOH in
ethanol followed by hydrogenation with 10% Pd/C and
H2 to produce the saturated diacid 5. Diacid 5 was
converted to the activated potential parallel â-sheet
nucleator, diester 6, by treatment with pentafluorophenol
in the presence of diisopropylcarbodiimide. Diester 6
could be directly incorporated into peptides by solution
fragment condensation with the amino terminus of
purified, side chain protected peptides.
Unnatural amino acid 10 was prepared from the
common intermediate 2 through sequential palladium-
catalyzed cross-coupling reactions to produce the unsym-
metrical unsaturated monoacid monoester 8. The first
cross-coupling employed palladium acetate (4 mol %),
triphenylphosphine (2 mol %), ethyl acrylate (1.5 equiv),
and triethyl amine at 75 °C to produce the monoiodide
monoester 7 in a 25% yield.31,33 The remaining starting
material (75%) did not react and was recovered and
recycled. A second cross-coupling employing 7 in the
Resu lts a n d Discu ssion
Xanthone was converted to 9,9-dimethylxanthene 1
using trimethylaluminum following a previously reported
procedure.29 The preparation of 6 and 10 from 1 employs
the common synthetic intermediate 4,5-diiodo-9,9-dim-
ethylxanthene (2) (Scheme 1). Diiodide 2 was formed by
the treatment of 1 with n-BuLi in the presence of
N,N,N′,N′-tetramethylethylenediamine (TMEDA) at 78
°C producing the 4,5-dianion, which was subsequently
quenched by rapid addition of diiodomethane (eq 2).
Less than 5% of the undesirable alkylation product
resulting from nucleophilic displacement of iodine from
diiodomethane was observed. However, slow addition of
diiodomethane to the dianion resulted in a greater yield
of the undesirable nucleophilic addition product.
(31) Arnold, R.; Kulenovic, S. J . Org. Chem 1978, 43, 3687 3689.
(32) Quenching the reaction mixture of 1 with carbon tetraiodide
produced 2 in a 40% yield; however, purification proved difficult due
to iodinated xanthene side products. The aryl halide products resulting
from quenching the reaction mixture with either carbon tetrabromide
(88% yield) or carbon tetrachloride (74% yield) were easily purified
but were unreactive in subsequent palladium cross-coupling reactions.
(33) Plevyak, J .; Dickerson, J .; Heck, R. J . Org. Chem. 1979, 44,
4078 4080.
(29) Nowick, J .; Ballester, P.; Ebmeyer, F.; Rebek, J . J . Am. Chem.
Soc. 1990, 112, 8902 8906.
(30) Abrams, S. J . Labelled Compd. Radiopharm. 1987, 29, 941
948.

Second Generation â-Sheet Nucleator
J . Org. Chem., Vol. 61, No. 21, 1996 7411
F igu r e 4. Temperature dependency of amide proton chemical
shift for diamide 12 in 1.5 mM CDCl3 solution. Data points at
268, 278, and 298 K for amide proton A are not represented
due to overlap with the aromatic region of xanthene: ( ) NH
of 12, amide A; (9) NH of 12, amide B.
dichloromethane at room temperature for 50 min. The
free amine was treated with phenylacetic acid in the
presence of BOP to afford diamide 12.
Previous hydrogen-bonding studies on acyclic amides,
in noncompetitive solvents such as CH2Cl2, have shown
that clear differences exist between the temperature
F igu r e 3. Conversion of compound 10 to compound 12
through 11: (a) 1.5 equiv of BOP, 1.5 equiv of benzylamine,
3.0 equiv of DIEA, CH2Cl2; (b) 25% TFA/CH2Cl2, 50 min; (c)
1.5 equiv of BOP, 2.0 equiv phenylacetic acid, 3.5 equiv of
DIEA, CH2Cl2, (25%).
dependence of hydrogen-bonded vs non-hydrogen-bonded
43
amide proton chemical shifts.41
A temperature coef-
ficient less than 10 ppb/K is indicative of a hydrogen-
bonded amide proton, while a dependence greater than
5 ppb/K is characteristic of a non-hydrogen-bonded
amide proton. Generally, hydrogen-bonded amides are
downfield of a free amide NH. FT-IR spectroscopy has
also been used to detect the presence of hydrogen-bonded
amide hydrogens due to an altered force constant which
presence of palladium acetate (2 mol %), triphenylphos-
phine (4 mol %), and acrylic acid gave the monoacid
monoester 8 in a 67% yield.31,33 The initial strategy called
for hydrogenation of the monoacid monoester and sub-
sequent Curtius rearrangement of the propionic acid side
chain. This strategy yielded a myriad of products via
cyclic intramolecular intermediates, presumably due to
the proximity of the propionic acid derived side chains
1
shifts the NH stretching band from 3400 to 3500 cm
for non-hydrogen-bonded amide hydrogens to a broad-
1
37
to one another throughout the course of the reaction.34
ened 3200 3400 cm band for hydrogen-bonded amide
NHs.
Therefore, the unsymmetrical unsaturated monoacid
monoester 8 was treated with diphenylphosphoryl azide
and triethylamine in refluxing toluene to produce the
unsaturated monoester monoisocyanate intermediate,
which was trapped with anhydrous tert-butyl alcohol to
Variable temperature NMR experiments verified the
presence of both a hydrogen-bonded and non-hydrogen-
bonded amide pair in amide 12 (Figure 4). Corroboration
was provided by the IR spectra of 12 in chloroform which
showed a sharp band at 3443 cm 1, indicative of a non-
hydrogen-bonded amide hydrogen, and a broad band at
3327 cm 1, corresponding to a hydrogen-bonded amide
hydrogen (see supporting information, concentration
independent from 0.5 to 6 mM). In order to assign the
amide proton involved in an intramolecular hydrogen
bond, a DQCOSY experiment was carried out at 258 K
to unambiguously assign both protons. It was deter-
mined that the amide signal at 7.46 ppm (258 K)
corresponded to amide A and that the signal at 6.7 ppm
(258 K) was that of amide B, (eq 1). This result
demonstrates that a 13-membered hydrogen-bonded ring
is present in CDCl3 solution (eq 1) due to the fact that
amide A exhibited a temperature dependence of 12 ppb/
K, characteristic of hydrogen bonding, while amide B
showed a temperature dependence of only 4 ppb/K.44
The curvature in the temperature dependent amide plots
afford the unsaturated monoester monocarbamate 9 in
40
a 67% yield.38
The purified material was hydrogenated
using 10% Pd/C and H2 in ethanol/acetic acid (4:1)
followed by hydrolysis with LiOH in ethanol at 25 °C to
give the amino acid 10 directly, in a 40% overall yield
from 7.
Diamide 12 was prepared using the route shown in
Figure 3, to evaluate the preferred hydrogen-bonding
pattern in amides derived from 10. The Boc-protected
xanthene-based amino acid 10 was treated with benzy-
lamine in the presence of (benzotriazol-1-yloxy)tris(dim-
ethylamino)phosphonium hexafluorophosphate (BOP) to
afford monobenzamide monocarbamate 11. The BOC
protecting group was then removed using 25% TFA in
(34) Chin, J .; Breslow, R. Tetrahedron Lett. 1982, 23, 4221 4224.
(35) Kluger, R.; Hunt, J . J . Am. Chem. Soc. 1989, 111, 5921 5925.
(36) Sauers, C.; Gould, C.; Ioannou, E. J . Am. Chem. Soc. 1972, 94,
8156 8163.
(37) Ishikawa, K.; Endo, T. J . Am. Chem. Soc. 1988, 110, 2016
2017.
(38) Hocking, M. Can. J . Chem. 1968, 46, 2275 2282.
(39) Pfister, J .; Wymann, W. Synthesis 1983, 38 40.
(40) Shioiri, T.; Ninomiya, S.; Yamada, S. J . Am. Chem. Soc. 1972,
94, 6203 6205.
(41) Stevens, E.; Sugawara, N.; Bonora, G.; Toniolo, C. J . Am. Chem.
Soc. 1980, 102, 7048 7050.
(42) Liang, G.; Rito, C.; Gellman, S. J . Am. Chem. Soc. 1992, 114,
4440 4442.
(43) Gellman, S.; Adams, B.; Dado, G. J . Am. Chem. Soc. 1990, 112,
460 461.

7412 J . Org. Chem., Vol. 61, No. 21, 1996
McWilliams and Kelly
the amide substituent decreased from benzyl to methyl,
the population of trans conformers sampled increased.
This result seems to indicate that as the steric bulk of
the amide substituents is decreased, the molecule is
capable of adopting a more extended conformation, which
favors the trans arrangement of the ethylene spacers
(Figure 2).
It is important to point out that for the 4-(2-aminoet-
hyl)-6-dibenzofuranpropionic acid derived amides, the 15-
membered ring hydrogen-bonded cis conformation was
favored. This conformation is thought to be important
for â-sheet nucleation since it allows intramolecular
hydrogen bonding between the flanking -amino acids,
analogous to the hydrogen bond pair that forms between
the i and i 3 residues in a naturally occurring â-turn.
In addition, this conformation facilitates hydrophobic
cluster formation involving the dibenzofuran skeleton
and the hydrophobic side chains of the flanking -amino
acids. It will be interesting to see what effect, if any,
this altered hydrogen bond preference has on the â-sheet
nucleating ability of 10 in aqueous solution. It is possible
that the 15-membered ring cis conformation would be
favored in peptides composed of 10 dissolved in aqueous
buffers as a result of the hydrophobic effect, which is not
operational in hydrocarbon solvents. In any case, residue
10 will prove useful in developing a further understand-
ing of the requirements for â-sheet nucleation.
Con clu sion
The potentially useful parallel â-sheet nucleator 6 and
antiparallel â-sheet nucleator 10, based on 9,9-dimeth-
ylxanthene, were regioselectively synthesized for future
incorporation into peptides composed of less than 20
-amino acid residues. Amides based on residue 10
prefer to adopt a 13-membered ring hydrogen-bonded
trans conformation, unlike the 15-membered ring hydro-
gen-bonded cis conformation adopted by the dibenzofu-
ran-based nucleators. Peptides composed of 10 will serve
as useful probes to evaluate the mechanistic requirement
of an effective â-sheet nucleator and to determine whether
hydrophobic cluster formation could change the preferred
conformation of 10 in the context of a water soluble
peptide.
F igu r e 5. Molecular dynamics results for benzyl diamide 12.
(A) Total energy in kilocalories for each minimized structure.
(B) Distance between indicated interior hydrogens (Figure 1)
of the -carbons of the ethylene spacer. Distances less than
4.0 Å are for trans arrangement of the side chains, while those
greater than 4.0 Å are characteristic of cis arrangement
(Figure 2). (C) Hydrogen bond distance of the 13-membered
intramolecular ring measured from the nitrogen of amide A
to the carbonyl oxygen of amide B.
at low temperature indicates some intermolecular as-
sociation; however, this is not a problem at ambient
temperature since the FT-IR peaks are concentration
independent from 0.5 to 6.5 mM.
In an effort to understand the hydrogen-bonding
results, molecular modeling studies on alkyl amide 12
and related amides (where the benzyl group is replaced
by an ethyl or methyl group) were performed. A molec-
ular dynamics study was carried out for 100 ps at 700 K
starting from the extended trans conformation. Con-
formers were sampled every picosecond and minimized
to a low-energy conformation, resulting in a trajectory
of 100 possible conformations. Figure 1 is a ball-and-
stick representation of the lowest energy conformer of
12, as determined from these studies.
The ethylene spacers adopt a “perpendicular conforma-
tion”, not the cis conformer observed in dibenzofuran
systems, but rather a trans conformation, as shown in
Figure 2.
Approximately 50% of the conformers of 12 sampled
were in this trans conformation (see Figures 1 and 5, and
all of the low-energy conformers had trans, 13-membered
intramolecularly hydrogen-bonded rings. As the size of
Exp er im en ta l Section
Gen er a l Meth od s a n d Ma ter ia l for Syn th esis of 6, 10,
a n d 12. The xanthone used in these studies was purchased
from Aldrich. The n-BuLi used for the metalation reactions
was purchased from Aldrich and titrated with 9-phenylan-
thracene and isopropyl alcohol prior to use. Diethyl ether was
distilled from sodium/benzophenone ketyl. Toluene was dis-
tilled from calcium hydride. Acetonitrile was distilled spectral
grade purchased from EM Science. N,N,N′,N′- tetramethyl-
ethylenediamine (TMEDA), diisopropylethylamine (DIEA),
and triethylamine (TEA) were purchased from Aldrich and
refluxed over ninhydrin, distilled, and distilled again from
KOH. tert-Butyl alcohol was dried with magnesium sulfate
and fractionally distilled onto activated 4 Å molecular sieves.
Triphenylphosphine was purchased from Aldrich and recrys-
tallized from hexanes. All other reagents were used without
further purification. Thin layer chromatography was per-
formed using aluminum sheets precoated with a 200 µm
coating of 60 F254 silica gel as purchased from Alltech. Flash
chromatography was performed according to Still.45 Melting
points were determined using a Fisher melting point ap-
paratus and are uncorrected. Routine NMR spectra were
(44) Variable temperature NMR studies were conducted on 12 at
6.5, 1, and 0.5 mM concentrations and indicated no concentration
dependence. There was downfield deviation observed for amide B at
6.5 mM, below 243 K. We ascribe this deviation to the beginning of
intermolecular aggregation.
(45) Still, W. C.; Kahn, M.; Maitra, A. J . Org. Chem. 1978, 43, 2923.

Second Generation â-Sheet Nucleator
J . Org. Chem., Vol. 61, No. 21, 1996 7413
recorded on a Varian XL-200E, unless otherwise specified.
Variable temperature studies and DQCOSY spectra were
recorded on a Varian 500 Unity Plus spectrometer. FT-IR data
were collected on a Galaxy 6021 spectrometer using a CaF2
solution cell having a 3 mm path length. Analytical HPLC
was carried out using a Perkin Elmer series 410 Bio pump
equipped with a variable wavelength UV detector and integra-
tor. HPLC was reverse phase employing C18 column packing
unless otherwise specified. Solvent A was 95% distilled water,
5 % spectral grade acetonitrile, and 0.2% triflouroacetic acid.
Solvent B was 95 % spectral grade acetonitrile, 5% distilled
water, and 0.2% triflouroacetic acid. Mass determinations
were carried out by the Texas A&M University Mass Spec-
trometry Applications Laboratory using a VG-70S double-
focusing high-resolution mass spectrometer. Molecular dy-
namics runs of 100 ps were carried out at 700 K without
constraints using the CVFF force field in discover (Biosym
Technologies). Conformations were sampled at 1 ps intervals.
Each structure was then minimized using conjugate gradients
until the first derivative reached a minimum value of 0.001
kcal mol 1 Å 1. No constraints were used in the dynamics run
or the minimization. The bulk dielectric used in the trajec-
tories was 1, to mimic the solvent properties of CH2Cl2 in which
the experimental studies were carried out. Modeling graphics
were created using MolScript.46
Syn th esis of 4,5-Diiod o-9,9-d im eth ylxa n th en e (2). An
oven-dried 250 mL round bottom flask cooled under argon was
charged with 20.0 g (95 mmol) of 9,9-dimethylxanthene (1) and
125 mL of freshly distilled diethyl ether. The flask was flushed
with argon, and the mixture was stirred for 10 min at room
temperature. N,N,N′,N′-Tetramethylethylenediamine [TME-
DA, 36 mL (236 mmol)] was added via syringe, and the
mixture was cooled to 78 °C using a dry ice/acetone bath.
After 20 min of stirring at 78 °C, 140 mL (280 mmol) of
n-BuLi (2.0 M in hexanes) was added via syringe at a rate of
4 mL/min. The resulting dark red solution was slowly warmed
to room temperature and stirred for 24 h. To an oven-dried
500 mL round bottom flask cooled under argon were added
150 mL of freshly distilled diethyl ether and 20 mL (248 mmol)
of diiodomethane. The diiodomethane/ether solution was
cooled to 0 °C using an ice bath and stirred for 20 min. The
solution of dimethylxanthene dianion prepared above was
cooled to 78 °C and transferred to the diiodomethane/ether
(0 °C) solution via cannula. Upon complete addition of the
dimethylxanthene dianion, the cream-colored suspension was
warmed to room temperature and stirred for 12 h. Distilled
water (200 mL) was added, and the mixture was stirred until
two clear layers were observed. The layers were filtered
through a Whatman (no.1) filter paper and separated, with
the water layer being extracted with two 100 mL portions of
diethyl ether. The ether was removed from the combined
ethereal layers under vacuum to give a viscous brown oil. The
oil was dissolved in acetone, filtered to remove inorganic salts,
and evaporated to dryness by rotary evaporator to give 30 g
(68% crude yield) of a dark brown oil. The oil was used without
further purification. The reaction was monitored by TLC (95:5
petroleum ether/ethyl acetate): diiodoxanthene (Rf 0.56) and
monoiodoxanthene (Rf 0.64). 1H NMR (CDCl3, 200 MHz) δ
The reaction was held at 75 °C for 6 h or until precipitated
palladium could be seen. The flask was cooled to room
temperature, and the DMF, ethyl acrylate, and triethylamine
were removed under high vacuum. The resulting dark oil was
partitioned between diethyl ether and water and filtered using
a 45 µm nylon membrane. The layers were separated, with
the water layer being extracted twice with 50 mL portions of
diethyl ether. The ether layers were combined, and the ether
was removed under vacuum to give a dark thick oil, further
purified by flash chromatography (90:10 hexanes/ethyl ac-
etate). The purification resulted in 2.34 g (67% yield) of 3:
1H (CDCl3, 200 MHz) δ 1.35 (t, J
7.2 Hz, 6H), 4.36 (q, J
16.0 Hz, 2H), 7.12 (t, J 7.8 Hz,
16.0 Hz, 1H). 13C (CDCl3, 50
7.2 Hz, 4H), 6.50 (d, J ,
2H), 7.48 (m, 4H), 8.32 (d, J
MHz) δ 14.56, 31.56, 35.25, 60.04, 84.91, 119.36, 122.88,
123.58, 125.23, 125.89, 127.74, 131.38, 137.41, 138.94, 167.04
C25H26O5 MS (FAB-NBA) calcd [M
[M H] 407.1846.
H]
407.1859 found
Hyd r olysis a n d Hyd r ogen a tion To Affor d Dia cid 5. An
oven-dried 100 mL round bottom flask cooled in a desiccator
over calcium sulfate (Drierite) was charged with 1.0 g (2.5
mmol) of 3, 1.0 g (25 mmol) of NaOH, and 50 mL of absolute
ethanol. The flask was equipped with an oven-dried con-
denser, and the mixture was refluxed for 12 h. The flask was
cooled to room temperature, and the diacid product was
collected by filtration through a Whatman (no. 1) filter paper.
The solid material was transferred to a 500 mL Parr hydro-
genation vessel containing 0.100 g of 10% Pd/C and 200 mL
of distilled water. The vessel was degassed by attaching an
aspirator to the exit valve of the hydrogenation apparatus,
evacuating the vessel, and pressurizing with H2. These actions
were performed three times, to insure that a H2 atmosphere
was present. The vessel was then pressurized to 55 psi and
agitated for 12 h. The palladium was removed by filtration
using a 45 µm nylon filter. The product was collected by
neutralization with 12 M HCl and filtration through a 45 µm
nylon membrane. The material was then redissolved into 100
mL of boiling ethanol, which caused aggregation of any
remaining Pd/C, and was subsequently removed by filtration
through a 45 µm nylon membrane. The ethanol was removed
under aspirator vacuum to give 0.83 g (94% yield) of a white
powder 5. 1H (acetone, 200 MHz) δ 1.76 (s, 6H), 2.88 (t, J
7.8 Hz, 4H), 3.29 (t, J
7.8 Hz, 4H), 7.19 (t, J
7.6 Hz, 2H),
7.34 (d, J 7.4 Hz, 2H), 7.56 (d, J 7.8 Hz, 2H). 13C (acetone,
50 MHz) δ 41.53, 47.72, 49.74, 50.07, 138.92, 140.41, 144.03,
144.1, 146.00, 164.15, 189.25. C21H22NO5 MS (FAB-thioglyc-
erol/1% TFA) calcd [M Na]
354.1492.
354.1467 found [M Na]
P en ta flu or op h en ol Active Ester 6. An oven-dried 25 mL
round bottom flask cooled in a desiccator over calcium sulfate
(Drierite) was charged with 50 mg (0.1 mmol) of 5. The flask
was held under high vacuum for 2 h, followed by addition of
50 mg (0.25 mmol) of pentafluorophenol, 0.05 mL of diisopro-
pylcarbodiimide, and 10 mL of an ethyl acetate/DMF solution
(70/30 v/v). The flask was capped with a rubber septum,
purged with argon for 10 min, and then stirred under argon
for 12 h. The precipitated DIC urea was removed by filtration
through a 45 µm nylon membrane. The ethyl acetate/DMF
solution was removed under aspirator vacuum, and the
resulting solid was purified by flash chromatography (85:15
hexanes:ethyl acetate) to give 74 mg of 6 (76% yield). 1H
(CDCl3, 200 MHz) δ 1.63 (s, 6H), 3.08 (m, 4H), 3.26 (m, 4H),
1.59 (s, 6H), 6.85 (t, J
2H), 7.71 (dd, J
7.8 Hz, 2H), 7.36 (dd, J
1.4, 7.8 Hz,
1.4, 7.8 Hz, 2H). 13C CDCl3, 50 MHz) δ
31.99, 84.48, 125.21, 125.95, 131.45, 137.73, 150.51 C15H12OI2
MS (EI) calcd [M ] 461.8978 found [M ] 461.8956.
Syn th esis of Dieth yl 9,9-Dim eth ylxa n th en e-4,5-d ip r o-
p en oa te (3). An oven-dried 250 mL round bottom flask cooled
in a desiccator over calcium sulfate (Drierite) was charged with
4.0 g (8.6 mmol) of 2. The flask was held under high vacuum
for 2 h, followed by addition of 0.100 g (5 mol %) of palladium
acetate and 0.46 g (20 mol %) of triphenylphosphine. The flask
was capped with a rubber septum and purged with argon for
10 min. Distilled anhydrous DMF (75 mL) was added via
syringe, followed by 10 mL of ethyl acrylate (92 mmol) and
2.5 mL of distilled triethylamine. The septum was replaced
with an oven-dried condenser cooled under argon, and the flask
was heated to 75 °C using a thermostat-regulated oil bath.
7.05 (t, J
(dd, J
7.4 Hz, 2H), 7.13 (dd, J
2.1, 7.8 Hz, 2H), 7.33
2.1, 7.6 Hz, 2H). 13C (CDCl3, 50 MHz) δ 25.69, 32.39,
33.49, 34.22, 123.11, 124.98, 126.34, 127.85, 130.15, 147.96,
168.78. C33H20F10O5 MS (FAB-NBA/MeOH) calcd [M F]
705.1135; found [M
F]
705.1136, [M]
686.1152.
Syn th esis of Eth yl 4-Iod o-9,9-d im eth ylxa n th en e-5-p r o-
p en oa te (7). An oven-dried 250 mL round bottom flask cooled
in a desiccator over calcium sulfate (Drierite) was charged with
10 g (216 mmol) of 2. The flask was held under high vacuum
for 2 h, followed by addition of 0.100 g (4 mol %) of palladium
acetate and 0.05 g (2 mol %) of triphenylphosphine. The flask
was capped with a rubber septum and purged with argon for
10 min. Distilled spectral grade acetonitrile (50 mL) was
added via syringe, followed by 2.5 mL of ethyl acrylate (240
(46) Kraulis, J . J . Appl. Crystallogr. 1991, 24, 946 950.

7414 J . Org. Chem., Vol. 61, No. 21, 1996
McWilliams and Kelly
mmol) and 7.6 mL of distilled triethylamine. The septum was
replaced with an oven-dried condenser cooled under argon, and
the flask was heated to 75 °C using a thermostat-regulated
oil bath. The reaction was held at 75 °C for 6 h or until
precipitated palladium could be seen. The flask was cooled
to room temperature, and the acetonitrile was removed under
aspirator vacuum. The resulting dark oil was partitioned
between diethyl ether and water and filtered using a 45 µm
nylon membrane. The layers were separated, with the water
layer being extracted twice with 50 mL portions of diethyl
ether. The ether layers were combined, and the ether was
removed under vacuum to give a dark thick oil which was
further purified by flash chromatography (95:5 hexanes/ethyl
acetate). the purification resulted in 2.23 g (24% yield) of 7
75 mL Parr hydrogenation vessel containing 0.400 g of 10%
Pd/C. The vessel was degassed by attaching an aspirator to
the outlet valve of the hydrogenation apparatus, evacuating
the vessel, and pressurizing with H2. These actions were
performed three times to insure that a H2 atmosphere was
present. The vessel was then pressurized to 35 psi and
agitated for 12 h. The palladium was removed by filtration
through a 45 µm nylon membrane. The product was isolated
by removing the ethanol and acetic acid under high vacuum
to give 0.69 (52% yield) of 9 and approximately 10% of the
tert-butyl ester ethyl carbamate product from exchange of the
tert-butyl and ethyl protecting groups. This product was
difficult to remove by flash chromatography; however, failure
of the tert-butyl ester to be hydrolyzed in the subsequent step
provided a convenient means to separate the two compounds.
and 7.5 g of 2. 1H (CDCl3, 200 MHz) δ 1.36 (t, J
3H), 1.59 (s, 6H), 4.30 (q, J 7.2 Hz, 2H), 6.56 (d, J
7.2 Hz,
16.2
For 9: 1H (CDCl3, 200 MHz) δ 1.26 (t, J
(s, 9H), 1.60 (s, 6H), 2.67 (t, J
Hz, 2H), 3.10 (t, J
7.0 Hz, 3H), 1.43
8.0 Hz, 2H), 2.97 (t, J 7.0
8.0 Hz, 2H), 3.45 (m, 2H), 4.19 (q, J
Hz, 1H), 6.85 (t, J
7.8 Hz, 1H), 7.11 (t, J
7.2 Hz, 1H), 7.39
16.2 Hz, 1H).
(m, 2H), 7.50 (m, 1H), 7.70 (m, 1H), 8.62 (d, J
13C (CDCl3, 50 MHz) δ 15.84, 32.96, 36.29, 61.78, 86.20, 120.74,
124.15, 124.96, 126.37, 126.50, 127.27, 129.03, 132.50, 132.70,
150.43, 150.77, 168.38. C20H20O3I MS (FAB-NBA/CH2Cl2)
7.0 Hz, 2H), 7.04 (m, 4H), 7.30 (d, J 7.6 Hz, 2H). 13C (CDCl3,
50 MHz) δ 14.21, 26.45, 28.36, 31.29, 34.15, 34.83, 40.69, 60.57,
78.86, 122.84, 124.41, 124.45, 126.12, 127.64, 127.82, 128.59,
129.99, 130.0, 148.23, 156.02, 173.36. C27H35NO2 MS (FAB-
calcd [M
H] 435.0454 found [M
H] 435.0466.
Syn th esis of 4-((Eth oxyca r bon yl)eth en yl)-9,9-d im eth -
ylxa n th en e-5-p r op en oic Acid (8). An oven-dried 100 mL
round bottom flask cooled in a desiccator over calcium sulfate
(Drierite) was charged with 3.13 g (72 mmol) of 7. The flask
was held under high vacuum for 2 h, followed by addition of
0.032 g (2 mol %) of palladium acetate and 0.075 g (4 mol %)
of triphenylphosphine. The flask was capped with a rubber
septum and purged with argon for 10 min. Distilled spectral
grade acetonitrile (50 mL) was added via syringe, followed by
0.54 mL of acrylic acid (7.9 mmol) and 2.5 mL of triethylamine.
The septum was replaced with an oven-dried condensor cooled
under argon. The flask was heated to 75 °C using a thermostat-
regulated oil bath and held for 6 h or until precipitated
palladium could be seen. The flask was cooled to room
temperature, and the acetonitrile was removed under aspirator
vacuum. The resulting dark thick oil was purified by flash
chromatography (70:29:1 hexanes/ethyl acetate/acetic acid) to
give 1.73 g (63% yield) of a light brown oil. 1H (CDCl3, 200
NBA/NaI/acetone) calcd [M H]
454.2598.
454.2593; found [M H]
Hyd r olysis od 9 To Affor d Un n a tu r a l Xa n th en e Am in o
Acid 10. An oven-dried 100 mL round bottom flask cooled in
a desiccator over calcium sulfate (Drierite) was charged with
0.18 g (0.4 mmol) of 9 and 50 mL of absolute ethanol. This
mixture was cooled to 0 °C under argon using an ice bath and
stirred for 10 min. LiOH H2O (0.100 g, 2.3 mmol) was added,
and the mixture was slowly warmed to room temperature. The
saponification reaction was monitored by HPLC at 254 nm,
and upon completion, the reaction was quenched with 10 mL
of 1 M citric acid. The ethanol was removed under vacuum,
resulting in the precipitation of the product, which was
collected by extraction with dichloromethane. The methylene
chloride was removed under aspirator vacuum, and the
material was held under high vacuum overnight to give 0.155
g (92%) of 10. 1H (CDCl3, 200 MHz) δ 1.40 (s, 9H), 1.61 (s,
6H), 2.71 (t, J
7.8 Hz, 2H), 3.40 (m, 2H), 5.06 (bs, 1H), 7.05 (m, 4H),
7.29 (d, J 7.8 Hz, 2H). C25H31NO5 MS (FAB-NBA/NaI/
polyethylene glycol 400 and 600) calcd [M Na] 448.2099;
found [M Na] 448.2103.
7.8 Hz, 2H), 2.95 (t, J
7.6 Hz, 2H), 3.12 (t,
MHz) δ 1.34 (t, J
Hz, 2H), 6.51 (d, J
7.1 Hz, 3H), 1.62 (s, 6H), 4.34 (q, J
7.1
J
16.2 Hz, 1H), 6.56 (d, J
16.0 Hz, 1H),
7.12 (m, 2H), 7.47 (m, 4H), 8.30 (d, J
J
16.2 Hz, 1H), 8.38 (d,
16.0 Hz, 1H), 9.25 (broad s, 1H) 13C (CDCl3, 50 MHz) δ
14.38, 31.77, 34.32, 60.75, 119.10, 120.36, 122.56, 122.99,
123.52, 125.70, 126.14, 127.70, 128.21, 130.84, 131.01, 138.09,
140.72, 148.38, 148.70, 166.93, 171.57. C23H22O5 MS (FAB-
BOP Cou p lin g To Affor d Dia m id e 12. Following a
previously reported procedure, diamide 12 was prepared in a
25% yield) of 10:19 1H (CDCl3, 500 MHz) δ 1.6 (s, 6H), 2.6 (m,
2H), 2.9 (m, 2H), 3.1 (m, 2H), 3.4 (m, 4H), 4.4 (d, J
2H), 6.4 (t, J 8.2 Hz, 1H), 7.0 (m, 4H), 7.2 (m, 13H)
C35H36N2O4 MS (FAB-NBA/CH2Cl) calcd [M H] 533.2804
found [M H] 533.2833.
NBA/TFA/CHCl3) calcd [M H]
379.1523.
379.1546; found [M H]
6.0 Hz,
Cu r tiu s Rea r r a n gem en t a n d Su bsequ en t Hyd r ogen a -
tion of 8 To Affor d 9. An oven-dried 100 mL round bottom
flask cooled in a desiccator over calcium sulfate (Drierite) was
charged with 1.25 g (3.3 mmol) of 8. The flask was held under
vacuum for 48 h, backfilled with argon, capped with a rubber
septum, and purged with argon for 30 min. Freshly distilled
toluene (50 mL) was added via syringe, followed by 0.80 mL
of diphenylphosphoryl azide (3.7 mmol, 1.1 equiv) and 1.2 mL
of distilled triethylamine (2.6 equiv). The septum was replaced
with an oven-dried condenser cooled under argon, and the flask
was heated to reflux. After 90 min, the flask was cooled to
room temperature and 0.24 mL (1 equiv) of dried tert-butyl
alcohol was added through the condenser. The flask was again
heated to reflux. After 8 h, another equivalent of dried tert-
butyl alcohol was added and the flask was heated to 45 °C for
an additional 8 h. The flask was then cooled to room
temperature, and the toluene was removed under aspirator
vacuum to give a thick pale green oil which was purified by
flash chromatography (85:15 hexanes/ethyl acetate) to give
0.81 g (54% yield) of the unsaturated amino acid precursor.
The majority (0.71 g) of this material was dissolved in 30 mL
of ethanol and 3 mL of glacial acetic acid and transferred to a
Ack n ow led gm en t. We thank Dr. Wayne R. Fiori
for substantial help with molecular modeling and the
Laboratories for Biological Mass Spectrometry and
Macromolecular Design for infrastructure support. We
gratefully acknowledge primary support from the Na-
tional Institutes of Health (R01GM 51105) and second-
ary support for this project from the Robert A. Welch
Foundation, the Camille and Henry Dreyfus Foundation
Teacher Scholars Program, and the Searle Scholars
Program.
Su p p or tin g In for m a tion Ava ila ble: IR data, concentra-
tion dependent VT-NMR for 12, and NMR spectra for all
compounds (16 pages). This material is contained in libraries
on microfiche, immediately follows this article in the microfilm
version of the journal, and can be ordered from the ACS; see
any current masthead page for ordering information.
J O9605758