Universal peptidomimetics

Article abstract of DOI:10.1021/ja1071916

This paper concerns peptidomimetic scaffolds that can present side chains in conformations resembling those of amino acids in secondary structures without incurring excessive entropic or enthalpic penalties. Compounds of this type are referred to here as minimalist mimics. The core hypothesis of this paper is that small sets of such scaffolds can be designed to analogue local pairs of amino acids (including noncontiguous ones) in any secondary structure; i.e., they are universal peptidomimetics. To illustrate this concept, we designed a set of four peptidomimetic scaffolds. Libraries based on them were made bearing side chains corresponding to many of the protein-derived amino acids. Modeling experiments were performed to give an indication of kinetic and thermodynamic accessibilities of conformations that can mimic secondary structures. Together, peptidomimetics based on these four scaffolds can adopt conformations that resemble almost any combination of local amino acid side chains in any secondary structure. Universal peptidomimetics of this kind are likely to be most useful in the design of libraries for high-throughput screening against diverse targets. Consequently, data arising from submission of these molecules to the NIH Molecular Libraries Small Molecule Repository (MLSMR) are outlined.

Full text of DOI:10.1021/ja1071916

Published on Web 12/23/2010
Universal Peptidomimetics
Eunhwa Ko,† Jing Liu,† Lisa M. Perez,‡ Genliang Lu,† Amber Schaefer,† and
Kevin Burgess*,†
Department of Chemistry and Laboratory for Molecular Simulation, Texas A&M UniVersity,
Box 30012, College Station, Texas 77842, United States
Received August 19, 2010; E-mail: burgess@tamu.edu
Abstract: This paper concerns peptidomimetic scaffolds that can present side chains in conformations
resembling those of amino acids in secondary structures without incurring excessive entropic or enthalpic
penalties. Compounds of this type are referred to here as minimalist mimics. The core hypothesis of this
paper is that small sets of such scaffolds can be designed to analogue local pairs of amino acids (including
noncontiguous ones) in any secondary structure; i.e., they are universal peptidomimetics. To illustrate this
concept, we designed a set of four peptidomimetic scaffolds. Libraries based on them were made bearing
side chains corresponding to many of the protein-derived amino acids. Modeling experiments were performed
to give an indication of kinetic and thermodynamic accessibilities of conformations that can mimic secondary
structures. Together, peptidomimetics based on these four scaffolds can adopt conformations that resemble
almost any combination of local amino acid side chains in any secondary structure. Universal peptidomi-
metics of this kind are likely to be most useful in the design of libraries for high-throughput screening against
diverse targets. Consequently, data arising from submission of these molecules to the NIH Molecular
Libraries Small Molecule Repository (MLSMR) are outlined.
are also widespread. Many helical14-17 and turn18-23 mimics
have been prepared like this, frequently by joining two amino
Introduction
Peptides consist of polyamide main chains bearing substit-
uents. The term “peptidomimetics”1,2 can encompass compounds
designed to resemble peptide main chains, side chains, or both.
Peptidomimetics that can present side chains on main-chain
scaffolds containing amide bonds are ubiquitous because
medicinal chemists frequently design analogues of bioactive
peptides using this approach. For instance, peptidomimetics
involving substitution of amide bonds with surrogates3-7 or
transition-state analogues8-13 are common. Peptide modifica-
tions to produce molecules that resemble secondary structures
acid substituents to constrain peptide fragments in relevant
conformations.24-32
(14) Patgiri, A.; Jochim, A. L.; Arora, P. S. Acc. Chem. Res. 2008, 41,
1289–1300.
(15) Wang, D.; Chen, K.; Kulp, J. L., III; Arora, P. S. J. Am. Chem. Soc.
2006, 128, 9248–9256.
(16) Wang, D.; Liao, W.; Arora, P. S. Angew. Chem., Int. Ed. 2005, 44,
6525–6529.
(17) Chapman, R. N.; Dimartino, G.; Arora, P. S. J. Am. Chem. Soc. 2004,
126, 12252–12253.
(18) Veber, D. F.; Freidinger, R. M.; Perlow, D. S.; Paleveda, W. J.; Holly,
F. W.; Strachan, R. G.; Nutt, R. F.; Arison, B. H.; Homnick, C.;
Randall, W. C.; Glitzer, M. S.; Saperstein, R.; Hirschmann, R. Nature
1981, 292, 55–58.
† Department of Chemistry.
‡ Laboratory for Molecular Simulation.
(1) Farmer, P. S. In Drug Design; Ariens, E. J., Ed.; Academic Press:
San Diego, 1980; Vol. 10, pp 119-143.
(2) Farmer, P. S.; Ariens, E. J. Trends Pharmacol. Sci. 1982, 3, 362–
365.
(19) Spanevello, R. A.; Hirschmann, R.; Raynor, K.; Reisine, T.; Nutt,
R. F. Tetrahedron Lett. 1991, 32, 4675–4678.
(20) Gierasch, L. M.; Deber, C. M.; Madison, V.; Niu, C.-H.; Blout, E. R.
Biochemistry 1981, 20, 4730–4738.
(3) Spatola, A. F. In Chemistry and Biochemistry of Amino Acids,
Peptides, and Proteins: A SurVey of Recent DeVelopments; Weinstein,
B., Ed.; Marcel Dekker: New York, 1983; Vol. 7, pp 267-357.
(4) Hruby, V. J.; Al-Obeidi, F.; Kazmierski, W. Biochem. J. 1990, 268,
249–262.
(21) Bean, J. W.; Kopple, K. D.; Peishoff, C. E. J. Am. Chem. Soc. 1992,
114, 5328–5334.
(22) MacDonald, M.; Vander Velde, D.; Aube, J. Org. Lett. 2000, 2, 1653–
1655.
(5) Lam, K. S.; Lebl, M.; Krchnak, V. Chem. ReV. 1997, 97, 411–448.
(6) Ahn, J.-M.; Boyle, N. A.; MacDonald, M. T.; Janda, K. D. Mini
ReV. Med. Chem. 2002, 2, 463–473.
(23) MacDonald, M.; Vander Velde, D.; Aube, J. J. Org. Chem. 2001,
66, 2636–2642.
(24) Jackson, D. Y.; King, D. S.; Chmielewski, J.; Singh, S.; Schultz,
P. G. J. Am. Chem. Soc. 1991, 113, 9391–9392.
(25) Bracken, C.; Gulyas, J.; Taylor, J. W.; Baum, J. J. Am. Chem. Soc.
1994, 116, 6431–6432.
(7) Hitotsuyanagi, Y.; Motegi, S.; Fukaya, H.; Takeya, K. J. Org. Chem.
2002, 67, 3266–3271.
(8) Greenlee, W. J. Med. Res. ReV. 1990, 10, 173–236.
(9) Ripka, A. S.; Rich, D. H. Curr. Opin. Chem. Biol. 1998, 2, 441–
452.
(26) Phelan, J. C.; Skelton, N. J.; Braisted, A. C.; McDowell, R. S. J. Am.
Chem. Soc. 1997, 119, 455–460.
(10) Lebon, F.; Ledecq, M. Curr. Med. Chem. 2000, 7, 455–477.
(11) Bursavich, M. G.; Rich, D. H. J. Med. Chem. 2002, 45, 541–558.
(12) Oishi, S.; Niida, A.; Kamano, T.; Odagaki, Y.; Tamamura, H.; Otaka,
A.; Hamanaka, N.; Fujii, N. Org. Lett. 2002, 4, 1051–1054.
(13) Rich, D. H.; Sun, C.-Q.; Prasad, J. V. N. V.; Pathiasseril, A.; Toth,
M. V.; Marshall, G. R.; Clare, M.; Mueller, R. A.; Houseman, K.
J. Med. Chem. 1991, 34, 1222–1225.
(27) Cabezas, E.; Satterthwait, A. C. J. Am. Chem. Soc. 1999, 121, 3862–
3875.
(28) Schafmeister, C. E.; Po, J.; Verdine, G. L. J. Am. Chem. Soc. 2000,
122, 5891–5892.
(29) Blackwell, H. E.; Sadowsky, J. D.; Howard, R. J.; Sampson, J. N.;
Chao, J. A.; Steinmetz, W. E.; O’Leary, D. J.; Grubbs, R. H. J. Org.
Chem. 2001, 66, 5291–5302.
9
462 J. AM. CHEM. SOC. 2011, 133, 462–477
10.1021/ja1071916  2011 American Chemical Society

Peptidomimetics of All Secondary Structures
A R T I C L E S
An important development in peptide mimicry has been the
emergence of analogues of peptide secondary structures that
mainly present selected side chains; i.e., the main-chain polya-
mide backbone is abbreviated or totally absent. This approach
is appealing because small molecules without polyamide
backbones are more likely to be orally bioavailable and
proteolytically stable. Early examples of this type of mimic were
from Hirschmann and Smith, who designed ꢀ-turn analogues
based on sugar,33,34 steroid,35 or even catechol36 backbones.
Similarly, Hamilton37-51 and others52-58 used biphenyl,52
terphenyl,38,39,41,47,48,59-61 and related37,39,40,44,47,62,63 scaffolds
to mimic helices. In our laboratory, we refer to compounds that
present only selected side chains to resemble peptide secondary
structures as minimalist mimics.64,65
Literature on minimalist mimics may give the impression that
these types of molecules have one preferred conformation in
solution that corresponds to the target secondary structure, but
that is not the case. Scaffolds like those in Figure 1a do not
exist as a single conformation in solution. They equilibrate
between forms representing local minima in Boltzman distribu-
tions of energy states, and their global minimum does not
necessarily correspond to the target secondary structure. Instead,
it is sufficient that the pertinent conformations for mimicry have
energies similar to the global minima so that they are populated
and that the transition-state energy barriers to arrive at them
can be overcome at ambient temperaturessin other words, that
there are no insurmountable thermodynamic or kinetic obstacles
to attaining the target conformations. On the other hand,
entropic considerations dictate that useful minimalist mimics
cannot be totally flexible. Their scaffolds must have limited
degrees of freedom to avoid significant entropic penalties on
adopting the target secondary structure conformations.
One of the original helical mimics reported by Hamilton is
discussed here to illustrate the validity of the assertions above.
Figure 1b illustrates facets of the conformational equilibria for
this helical mimic. It has only two significant degrees of freedom
that affect the orientation of the side chains. Conformers (i)-(iv)
approximate to minimum energy forms that alleviate steric
interactions between the subsituted phenyl groups.66 There can
be no conformer that is substantially lower in energy than these,
so they are thermodynamically accessible with respect to the
global minimum; i.e., they will all be significantly populated.
In fact, they are approximately equal in energy, so the
populations of each of these conformers are similar. Similarly,
the transition-state energy barriers that need to be surmounted
to equilibrate conformers of such substituted terphenyls can be
overcome at room temperature (those energies for terphenyls67,68
depend on the ortho substitutents involved),69,70 so states
(i)-(iv) are also kinetically accessible.
(30) Kelso, M. J.; Hoang, H. N.; Oliver, W.; Sokolenko, N.; March, D. R.;
Appleton, T. G.; Fairlie, D. P. Angew. Chem., Int. Ed. 2003, 42, 421–
424.
(31) Shepherd, N. E.; Abbenante, G.; Fairlie, D. P. Angew. Chem., Int.
Ed. 2004, 43, 2687–2690.
(32) Fujimoto, K.; Oimoto, N.; Katsuno, K.; Inouye, M. Chem. Commun.
2004, 11, 1280–1281.
(33) Hirschmann, R.; Nicolaou, K. C.; Pietranico, S.; Salvino, J.; Leahy,
E. M.; Sprengeler, P. A.; Furst, G.; Smith, A. B., III J. Am. Chem.
Soc. 1992, 114, 9217–9218.
(34) Hirschmann, R.; et al. J. Am. Chem. Soc. 1993, 115, 12550–12568.
(35) Hirschmann, R.; Sprengeler, P. A.; Kawasaki, T.; Leahy, J. W.;
Shakespeare, W. C.; Smith, A. B., III J. Am. Chem. Soc. 1992, 114,
9699–9701.
(36) Mowery, B. P.; Prasad, V.; Kenesky, C. S.; Angeles, A. R.; Taylor,
L. L.; Feng, J.-J.; Chen, W.-L.; Lin, A.; Cheng, F.-C.; Smith, A. B.,
III; Hirschmann, R. Org. Lett. 2006, 8, 4397–4400.
(37) Kim, I. C.; Hamilton, A. D. Org. Lett. 2006, 8, 1751–1754.
(38) Yin, H.; Lee, G.; Sedey, K. A.; Kutzki, O.; Park, H. S.; Orner, B. P.;
Ernst, J. T.; Wang, H.-G.; Sebti, S. M.; Hamilton, A. D. J. Am. Chem.
Soc. 2005, 127, 10191–10196.
(39) Davis, J. M.; Truong, A.; Hamilton, A. D. Org. Lett. 2005, 7, 5405–
5408.
(40) Ernst, J. T.; Becerril, J.; Park, H.; Yin, H.; Hamilton, A. D. Angew.
Chem., Int. Ed. 2003, 42, 535–539.
(41) Kutzki, O.; Park, H. S.; Ernst, J. T.; Orner, B. P.; Yin, H.; Hamilton,
A. D. J. Am. Chem. Soc. 2002, 124, 11838–11839.
(42) Peczuh, M. W.; Hamilton, A. D. Chem. ReV. 2000, 100, 2479–2494.
(43) Wilson, A. J.; Hong, J.; Fletcher, S.; Hamilton, A. D. Org. Biomol.
Chem. 2007, 5, 276–285.
(44) Davis, J. M.; Tsou, L. K.; Hamilton, A. D. Chem. Soc. ReV. 2007,
36, 326–334.
Conformers (i) and (iv) in Figure 1b correspond to helical
orientations of the side chains with opposite handedness, while
(ii) and (iii) do not display side chains in helical conformations.
If this mimic binds to a substrate in the helical conformation
(i), for instance, then it would do so by induced fit. This is
possible because the target conformation of the mimic is
kinetically and thermodynamically accessible, and not because
it is thermodynamically preferred.
(45) Becerril, J.; Hamilton, A. D. Angew. Chem., Int. Ed. 2007, 46, 4471–
4473.
(46) Fletcher, S.; Hamilton, A. D. Curr. Top. Med. Chem. 2007, 7, 922–
927.
(47) Yin, H.; Lee, G.-I.; Hamilton, A. D. Drug DiscoVery Res. 2007, 281–
299.
(48) Yin, H.; Hamilton, A. D. In Chemical Biology: From Small Molecules to
System Biology and Drug Design; Schreiber, S. L., Kapoor, T. M., Wess,
G., Eds.; Wiley-VCH: Weinheim, 2007; Vol. 1, pp 250-269.
(49) Becerril, J.; Rodriguez, J. M.; Saraogi, I.; Hamilton, A. D. Foldamers
2007, 195–228.
Identification of conformations in equilibrating ensembles that
are both kinetically and thermodynamically accessible can be
done only by comparing with similar systems that have been
(50) Rodriguez, J. M.; Hamilton, A. D. Angew. Chem., Int. Ed. 2007, 46,
8614–8617.
(51) Fletcher, S.; Hamilton, A. D. J. R. Soc., Interface 2006, 3, 215–233.
(52) Jacoby, E. Bioorg. Med. Chem. Lett. 2002, 12, 891–893.
(53) Jiang, H.; Leger, J.-M.; Huc, I. J. Am. Chem. Soc. 2003, 125, 3448–
3449.
(61) Yin, H.; Lee, G.-i.; Park, H. S.; Payne, G. A.; Rodriguez, J. M.;
Sebti, S. M.; Hamilton, A. D. Angew. Chem., Int. Ed. 2005, 44, 2704–
2707.
(54) Che, Y.; Brooks, B. R.; Marshall, G. R. J. Comput.-Aided Mol. Des.
2006, 20, 109–130.
(62) Rodriguez, J. M.; Hamilton, A. D. Tetrahedron Lett. 2006, 47, 7443–
7446.
(55) Okuyama, M.; Laman, H.; Kingsbury, S. R.; Visintin, C.; Leo, E.;
Eward, K. L.; Stoeber, K.; Boshoff, C.; Williams, G. H.; Selwood,
D. L. Nat. Methods 2007, 4, 153–159.
(63) Marimganti, S.; Cheemala, M. N.; Ahn, J.-M. Org. Lett. 2009, 11,
4418–4421.
(64) Chen, D.; Brahimi, F.; Angell, Y.; Li, Y.-C.; Moscowicz, J.; Saragovi,
H. U.; Burgess, K. ACS Chem. Biol. 2009, 4, 769–781.
(65) Angell, Y.; Chen, D.; Brahimi, F.; Saragovi, H. U.; Burgess, K. J. Am.
Chem. Soc. 2008, 130, 556–565.
(56) Ahn, J.-M.; Han, S.-Y. Tetrahedron Lett. 2007, 48, 3543–3547.
(57) Volonterio, A.; Moisan, L.; Rebek, J., Jr. Org. Lett. 2007, 9, 3733–
3736.
(58) Shaginian, A.; Whitby, L. R.; Hong, S.; Hwang, I.; Farooqi, B.;
Searcey, M.; Chen, J.; Vogt, P. K.; Boger, D. L. J. Am. Chem. Soc.
2009, 131, 5564–5572.
(66) Tsuzuki, S.; Uchimaru, T.; Matsumura, K.; Mikami, M.; Tanabe, K.
J. Chem. Phys. 1999, 110, 2858–2861.
(67) Johansson, M. P.; Olsen, J. J. Chem. Theory Comput. 2008, 4, 1460–
1471.
(59) Orner, B. P.; Ernst, J. T.; Hamilton, A. D. J. Am. Chem. Soc. 2001,
123, 5382–5383.
(68) Baraldi, I.; Ponterini, G. J. Mol. Struct. 1985, 23, 287–298.
(69) Leroux, F. ChemBioChem 2004, 5, 644–649.
(70) Grein, F. J. Phys. Chem. A 2002, 106, 3823–3827.
(60) Ellard, J. M.; Zollitsch, T.; Cummins, W. J.; Hamilton, A. L.; Bradley,
M. Angew. Chem., Int. Ed. 2002, 41, 3233–3236.
9
J. AM. CHEM. SOC. VOL. 133, NO. 3, 2011 463

A R T I C L E S
Ko et al.
Figure 1. (a) Illustrative examples of minimalist mimics. (b) Favorable conformations of Hamilton’s helical mimic shown in (a).
studied experimentally or via computational methods. Spectro-
scopic techniques like NMR cannot detect a preferred solution-
state conformer of a mimic like that shown in Figure 1b at room
temperature, because there is none. Coupling constants and NOE
measurements for the scaffolds would reflect conformational
aVeraging and give no useful data. Even if the mimic were
crystallized, the conformational state(s) in the crystal would be
determined by lattice packing forces. Consequently, computa-
tional methods are highly desirable to assess kinetically and
thermodynamically accessible solution conformations of mini-
malist mimics.
This paper outlines structural design criteria for ideal mini-
malist mimics. In particular, it describes how Cꢀ-Cꢀ distances
can be used to estimate the correspondence of a minimalist
mimic to a particular secondary structure, and how closer
consideration of CR-Cꢀ vectors can be used as a more stringent
test. It then describes how a small number of complementary
scaffolds can be used to target any local pair of amino acids in
any secondary structure, i.e., uniVersal mimics.71 Several
scaffold types were designed to give a set of putative universal
mimics, and libraries of each were generated to prove that a
variety of amino acid side-chain substituents could be incor-
porated. Computational methods were identified to assess kinetic
and thermodynamic accessibilities of relevant mimic conforma-
tions. Finally, an indication of the types of biophysical data that
can be obtained with these compounds is presented.
Results
Criteria for the Design of Minimalist Mimics. To the best of
our knowledge, criteria for design of minimalist mimics have
never been specifically delineated before. Here we propose the
following four structural design criteria:
• facile syntheses with most amino acid side chains (Arg, Trp,
His, etc.);
• kinetically and thermodynamically accessible conformations
for induced fit (i.e., not too rigid); but also
• only moderate loss of entropy on docking (i.e., only a few
significant degrees of freedom that influence the side-chain
orientations); and
• appropriate CR-Cꢀ coordinates of an accessible conforma-
(71) Smith, A. B., III; Xiong, H.; Charnley, A. K.; Brenner, M.; Mesaros,
E. F.; Kenesky, C. S.; Costanzo, L. D.; Christianson, D. W.;
Hirschmann, R. Org. Lett. 2010, 12, 2994-2997 used the term
“universal peptidomimetics” to indicate one scaffold (in this case
their pyrrolinone derivatives) might be used to mimic all secondary
structures, but their design criterial were very different to those
described here.
tion of the mimic matching those of the secondary structure.
Other considerations can be important, too (e.g., water
solubilities, toxicities, cell permeability, shelf lives). However,
these four key parameters define the structural basis for
9
464 J. AM. CHEM. SOC. VOL. 133, NO. 3, 2011

Peptidomimetics of All Secondary Structures
A R T I C L E S
designing minimalist mimics; if one of these is not satisfied,
then concerns about the physical and pharmacological charac-
teristics of the compounds may be inconsequential.
The first of the criteria listed above is self-evident from a
practical perspective: a scaffold that cannot be easily made with
most of the amino acid side chains found in proteins is generally
not a good mimic. For instance, designs that can be conveniently
prepared only with simple alkyl side chains would be of limited
value if it was necessary to present Asn, Asp, Arg, Gln, His,
Lys, Ser, Thr, Trp, and/or Tyr side chains.
The second and third criteria listed above, relating to the
kinetic and thermodynamic accessiblities and to the significant
degrees of freedom, have been addressed already. However, the
fourth parameter, relating to correspondence of CR-Cꢀ vectors,
is unique to this work and warrants further explanation.
Others before us have tried to reduce conformations of
secondary structures to the simplest possible terms. For instance,
ꢀ-turn mimics72 have been classified according to CR-atom
separations73 or CR-Cꢀ vectors,74 but, as far as we are aware,
there have been no papers that suggest that mimics of all
secondary structures could be broadly categorized using similar
factors.
of sophistication. This comes by considering coordinates of the
CR-Cꢀ bond vectors of side chains, because that parameter
reveals how the side chain projects into space as well as Cꢀ-Cꢀ
distances. Matching CR-Cꢀ bond vectors to a particular second-
ary structure, therefore, gives a more realistic sense of the
validity of the mimic. The only disadvantage is that consider-
ations of CR-Cꢀ bond vectors require computation; hence, it is
less convenient than the intuitive approach based on Cꢀ-Cꢀ
distances alone.
Analyzing the properties of minimalist mimics led us to
conclude that there should be special subsets of these compounds
that we now call uniVersal peptidomimetics. The next section
describes how Cꢀ-Cꢀ distances (abbreviated to “ꢀs”) can be
used as an intuitive guide to the suitability of possible mimics
as secondary structure mimics, and then how consideration of
CR-Cꢀ bond vectors provides a more stringent test.
Candidate Universal Peptidomimetics Selected on the
Basis of Cꢀ-Cꢀ Distances. Minimalist peptidomimetics do not
have fixed Cꢀ-Cꢀ separations (ꢀs values). For instance, pep-
tidomimetics based on the triazoles B could have several
different minimal energy conformations that correspond to
preferred orientations of the two side chains. Indeed, energy
barriers for interconversion between the conformers that place
the two Cꢀ atoms as close and as far apart as possible are likely
to be relatively insignificant, so this particular type of peptido-
mimetic could access a range of ꢀs values between these two
extremes. In general, extremely contracted and expanded Cꢀ-Cꢀ
separations can be denoted as ꢀsc and ꢀse, respectively.
Figure 2. (a) A ꢀ-turn and a minimalist mimic consisting of a scaffold
affording appropriate Cꢀ atom separations. (b) Minimalist mimics A and B
prepared by our group.
Figure 3. (a) Conformations of the triazole B can display side chains
aligned with Cꢀ atoms as close or as far apart as possible, and these
conformations are easily interconverted at room temperature. (b) Definition
of the term “extension factor”.
We recently proposed that Cꢀ-atom separations are critical
in the design of minimalist ꢀ-turn mimics.64,65 This is because
the Cꢀ positions represent the last atoms along the side chains
that are held with some rigidity in these secondary structures;
Cꢀ-Cγ side-chain vectors, and all those further removed from
the backbone, are flexible (Figure 2a). Scaffolds A and B were
prepared on the basis of this hypothesis (Figure 1b). Here we
suggest that, in general, accessibility of suitable distances
between Cꢀ atoms is a necessary condition for the design of
minimalist peptidomimetics. This concept is shown for a ꢀ-turn
in Figure 2, and similar diagrams could be drawn for any
secondary structure.
To calibrate the capacity of a minimalist peptidomimetic
to span different Cꢀ-Cꢀ distances, we have defined a
parameter called the extension factor, ef ) ꢀse/ꢀsc, where
ꢀse and ꢀsc represent maximally extended and contracted
Cꢀ-Cꢀ separations among the conformers that are both
kinetically and thermodynamically accessible (Figure 3).
Some peptidomimetics could have kinetically or thermody-
namically inaccessible Cꢀ-Cꢀ separations between ꢀse and
ꢀsc, but for most, all conformations between ꢀse and ꢀsc will
be accessible. Peptidomimetics with small extension factors
can mimic a limited range of Cꢀ-Cꢀ separations, while those
with large extension factors correspond to side-chain separa-
tions in more secondary structures. However, large extension
factors are not always ideal in designs of universal peptido-
mimetics because they might also represent excessive flex-
ibility in the scaffold, so ef values are predictors of scope
rather than quality. On the other hand, scaffolds with large
extension factors that can freeze in desired conformations
with minimal loss of entropic stabilization are promising for
the design of universal peptidomimetics.
Intuitively, Cꢀ-Cꢀ distances are useful as a “rough cut” to
gauge the fit of a proposed minimalist mimic to a secondary
structure. However, more accurate fitting requires a higher level
(72) Ball, J. B.; Hughes, R. A.; Alewood, P. F.; Andrews, P. R.
Tetrahedron 1993, 49, 3467–3478.
(73) Garland, S. L.; Dean, P. M. J. Comput.-Aided Mol. Des. 1999, 13,
469–483.
(74) Garland, S. L.; Dean, P. M. J. Comput.-Aided Mol. Des. 1999, 13,
485–498.
9
J. AM. CHEM. SOC. VOL. 133, NO. 3, 2011 465

A R T I C L E S
Ko et al.
Table 1. Extension Factors Calculated for the Featured
Four templates, 1-4 (listed in order of increasing average
Cꢀ-Cꢀ separations), were conceived using the structural criteria
listed for minimalist mimics. Piperidine or piperazine residues
in scaffolds 1, 3, and 4 were also included to facilitate assembly
of these molecules into heterobivalent forms,64,65,75 but this is
of no immediate consequence here.
Peptidomimetics
monomer
ꢀsc (Å)
ꢀse (Å)
ef
A
B
1
5.5
5.5
5.2
6.2
7.1
7.2
1.2
1.3
1.4
1.1
1.9
1.2
2
7.5
8.1
3
4
7.4
12.1
14.1
15.0
Table 1 shows the extension factors deduced for mimics A,
B, and 1-4. The data shown demonstrate that scaffold 1 allows
the R1 and R2 side chains to approach more closely than any
other in the series of compounds examined (i.e., ꢀsc is the
smallest located). However, the value for ꢀse is intermediate in
the series; hence, the extension factor for this mimic is not the
smallest of these compounds.
Table 2 shows data from modeling experiments we performed
to assess Cꢀ-Cꢀ separations (ꢀs) for the common elements of
secondary structure. It reveals considerable overlap between ꢀs
values for different residues in different secondary structures
with the range of ꢀs values accessible by each mimic. For
instance, the i to i + 3 ꢀs in a type I ꢀ-turn is about equal to
that for the i to i + 2 in a inverse γ-turn (ca. 5.4 Å), so a scaffold
that can access this ꢀs can mimic those side chains in both
secondary structures. Indeed, the extent of Cꢀ-Cꢀ separation
overlap in Table 2 indicates that almost all the side-chain Cꢀ-Cꢀ
separations could be mimicked with relatively few peptidomi-
metics. However, as noted above, analyses of Cꢀ-Cꢀ separations
provide rough approximations for the suitability of minimalist
mimics. This parameter does not indicate the orientation of side
chains; more stringent tests (based on CR-Cꢀ bond vectors) are
required for this.
Each of the four mimic designs 1-4 are now considered
separately. These scaffolds were very carefully designed to
conform to the four parameters delineated above for minimalist
peptidomimetics. Syntheses of small libraries based on these
scaffolds are described to illustrate a range of amino acid side
chains can be incorporated in each case. Density functional
theory (DFT) calculations to explore the kinetic accessibility
of the conformers are outlined. Thermodynamic accessibilities
of the conformers pertinent to secondary structures are then
described on the basis of a molecular dynamics and clustering
procedure based on CR-Cꢀ bond vectors.
The following procedure was used to calculate extension
factors. First, the significant degrees of freedom in the minimized
energy scaffold were identified. Typically these are bonds that
can be rotated to adjust the Cꢀ-Cꢀ separations. Rotations around
these bonds were explored by setting this as the only variable
and then monitoring energy as a function of rotation. This led
to identification of the conformations that minimize and
maximize the ꢀs values; hence, the extension factors can be
determined. Figure 4 shows these conformations for peptido-
mimetics 1.
Universal Peptidomimetics 1: 1,3,4-Oxadiazole-Based. 1,3,4-
Oxadiazoles80-82 have been used extensively in the design of
pharmaceuticals83,84 and have been shown to have a range of
bioactivities and applications, including bactericidal, fungicidal,
analgesic, anti-inflammatory, antiproteolytic, anticonvulsant,
nervous system depressant, sedative, and local anesthetic.85-89
A 1,3,4-oxadiazole-containing pharmaceutical of current im-
portance is Merck’s anti-retroviral HIV drug, Raltegravir. 1,3,4-
Oxadiazoles are also a compact five-membered heterocycle that
might be ideal for the formation of universal peptidomimetics.
Scheme 1a describes the procedure used to form the
oxadiazole80,83,84,90 mimics 1 via formation of amino acid
(75) Reyes, S. J.; Burgess, K. Chem. Soc. ReV. 2006, 35, 416–423.
(76) Kee, K. S.; Jois Seetharama, D. S. Curr. Pharm. Des. 2003, 9, 1209–
1224.
(77) Vass, E.; Hollosi, M.; Besson, F.; Buchet, R. Chem. ReV. 2003, 103,
1917–1954.
Figure 4. (a) “Significant degrees of freedom” for mimic 1 (R1 ) R2 )
Me); rotation about only these bonds in the scaffold alters the ꢀs. (b)
Conformations corresponding to the ꢀsc and ꢀse in an abbreviated structure
of compound 1 (piperidine ring omitted).
(78) DiscoVery Studio 2.5; Accelrys Inc., 2010.
(79) Parisien, M.; Major, F. 2004, Beta-Sheet’s World website http://www-
lbit.iro.umontreal.ca/bBuilder/index.html.
9
466 J. AM. CHEM. SOC. VOL. 133, NO. 3, 2011

Peptidomimetics of All Secondary Structures
A R T I C L E S
Scheme 1. Two Methods for Preparing Monovalent Mimics 1
Table 2. Correspondence of Cꢀ-Cꢀ Distances for Peptidomimetics
A, B, and 1-4 with Common Secondary Structuresa
way to store these materials. Many of these steps can be
performed in parallel to facilitate rapid formation of a chemical
library.
A series of oxadiazole compounds that complement the ones
shown in Scheme 1a was also made (Scheme 1b); these contain
N,N-dimethyl serine fragments.97 Three compounds containing
this component were prepared via the slightly modified method
shown in Scheme 1b. This is very similar to that in 1a but
involves an initial reductive amination step.
Table 3 lists the side chains incorporated into the oxadiazole-
based mimics 1. They are all derived from protein-based amino
acids, including ones that have diverse and relatively reactive
functionalities (e.g., Tyr, Lys, Glu, Ser, Asp, Ile, Thr). The patent
literature contains a synthetic route similar to that shown in
Scheme 1a98 but does not encompass functionalized amino acids.
The following DFT method was used to investigate kinetic
accessibility for compounds 1 and all the other scaffolds in the
series. Briefly, Gaussian 03 was used at the B3LYP level of
a Templates for ideal type I ꢀ-turns54,76 and for γ-turns77 were
obtained from standard torsion angles. A standard template for overlays
with an R-helix was obtained from Discovery Studio 2.5,78 and a
ꢀ-sheet template for overlays was obtained by ꢀ-sheet builder.79
hydrazides 5,91 coupling these to another amino acid,92 and then
dehydrative cyclization to give 6. Several sets of conditions were
tested for the cyclization step, but the triphenylphosphine iodine
system shown appeared to give the best results.93-95 Thereafter,
a piperidine group was added (so that these monovalent
compounds might later be assembled into bivalent ones),65,96
and the protecting groups were removed to give the target
molecules 1. This synthetic strategy involved six linear steps
and only two chromatographic separations to the protected forms
7. Those intermediates 7 probably have longer shelf-lives than
the unprotected compounds 1 and are therefore the preferred
9
J. AM. CHEM. SOC. VOL. 133, NO. 3, 2011 467

A R T I C L E S
Ko et al.
Table 3. Oxadiazole-Based Peptidomimetics 1
Figure 5. (a) Transposition of the global minima of 1 into a conformation
that mimics the i and i + 3 residues in an R-helix by rotation around one
of the significant bonds and then the other (∆G° values shown in kcal/
mol). (b) Overlay of the latter conformation on an ideal R-helix (shown
color-coded on the left and in pink on the right).
1. The global minimum energy structure that was identified was
set to 0 kcal/mol and is shown on the left side of Figure 5a.
Consideration of the likely energy barriers corresponding to
rotations about the significant degrees of freedom for mimics 1
indicated that many conformations other than the global
minimum would be accessible. Further, Table 2 indicates that
scaffolds 1 can have Cꢀ-Cꢀ separations corresponding to the i
to i + 3 residues of an R-helix. The conformation of 1 that best
overlays with the Cꢀ atoms of the i and i + 3 residues in an
R-helix is shown on the right-hand side of Figure 5a, and an
overlay of this mimic with an R-helix is illustrated in Figure
5b. That conformation is then a target for further computational
studies to monitor accessibility. Mapping the variation of energy
with rotation about the two significant degrees of freedom
reveals that it takes 1.31 kcal/mol to overcome one rotational
energy barrier and 2.43 kcal/mol to surmount the other to arrive
at the targeted R-helical conformations. Energy barriers of this
magnitude are easily surmounted at room temperature, so
induced fit is possible. Further, the targeted conformation is only
0.41 kcal/mol above the global minima, so these molecules will
populate shapes required for docking the i and i + 3 residues
in an R-helical form, even in the absence of other factors to
induce that conformation.
theory with a 6-31+G(d′) basis set and a polarized continuum
solvation model with a dielectric of 80 (see Supporting
Information). Figure 5 illustrates the data obtained for mimic
The DFT method described above facilitates calculation of
transition-state energies and relative energies of resting conforma-
tions, but it does not show the all the conformations that can be
formed and rank their relative energies, i.e., thermodynamic
accessibility. To do that, we used a complementary computational
(80) Weaver, G. W. Sci. Synth. 2004, 13, 219–251.
(81) Stolle, R. Ber. Dtsch. Chem. Ges. 1899, 32, 797–798.
(82) Pellizzari, G. Real. Accad. Lincei 1899, 8, 327–332.
(83) Hetzheim, A.; Moeckel, K. AdV. Heterocycl. Chem. 1966, 7, 183–
224.
(84) Hill, J. Compr. Heterocycl. Chem. II 1996, 4 (267-287), 905–1006.
9
468 J. AM. CHEM. SOC. VOL. 133, NO. 3, 2011

Peptidomimetics of All Secondary Structures
A R T I C L E S
Table 4. QMD Analysis of 1
a Representative structures in the families highlighted overlay with
the secondary structures as represented in the figures accompanying this
table.
technique: quenched molecular dynamics (QMD).99-102 In this
technique, the molecule is minimized and then subjected to a
molecular dynamics run at high temperature (1000 K) for a short
time (600 ps); 600 conformational states are recorded during
this run (i.e., every 1 ps) and minimized via molecular
mechanics. The lowest energy structures below a user-defined
cutoff are selected and then clustered into families on the basis
of root-mean-square (rms) deviations from user-defined atoms
(see Supporting Information).
The QMD technique applied to compound 1 (R1 ) R2 ) Me)
illustrates the concept of universal peptidomimetics well.
Figure 6. QMD data for compound 1 (R1 ) R2 ) Me). (a) Data from
family 3, illustrating overlay with an inverse γ-turn (overlay with an R-helix
is shown in Figure 5b). (b) Data from family 5, illustrating overlay with a
type 1 ꢀ-turn.
(85) Roda, K. P.; Vansdadia, R. N.; Parekh, H. J. Indian Chem. Soc. 1988,
65, 807–809.
(86) Raman, K.; Parmar, S. S.; Salzman, S. K. J. Pharm. Sci. 1989, 78,
999–1002.
(87) Ergenc, N.; Buyuktimkin, S.; Capan, G.; Baktir, G.; Rollas, S.
Pharmazie 1991, 46, 290–291.
Application of a 0.45 kcal/mol energy cutoff gave 180 (out of
600, see above) minimized conformations. These 180 structures
were clustered by oVerlay of the CR and Cꢀ atoms (to within
0.3 Å rms deviation); this method was chosen because it matches
conformations in terms of both Cꢀ-Cꢀ distances and CR-Cꢀ
bond vectors. Sorting in terms of CR-Cꢀ bond vectors as well
as Cꢀ-Cꢀ distances is a refinement that facilitates selection of
the subsets of conformations with similar side-chain orientations
from within the larger category of conformations with similar
ꢀs values. This process gave seven families of conformations
(Table 4). Family 3 has the most structures that overlaid (i.e.,
is the most “populated”), and the minimum energy structure in
this family was only 0.15 kcal/mol above the overall minimum
energy conformation of those sampled. Conformations within
this family overlay well with the i-i + 3 side chains of an ideal
R-helix conformation; this is the same conclusion that was
reached using the DFT method (Figure 6a; see also Figure 5b
above). However, structures in this family also overlay the
i-i + 2 side chains of an ideal inverse γ-turn (Figure 6a).
Furthermore, structures in family 5 overlay with an ideal type
1 ꢀ-turn at the i + 1-i + 2 side chains (Figure 6b). Thus, the
QMD technique illustrates how conformations of peptidomi-
metic 1, all of which are less than 0.45 kcal/mol from the overall
minimum energy conformation identified, can mimic three
(88) Saxena, V. K.; Singh, A. R.; Agarwai, R. K.; Mehra, S. C. J. Indian
Chem. Soc. 1983, 60, 575–577.
(89) Mazzone, G.; Bonina, F.; Puglisi, G.; Panico, A. M.; Arrigo Reina,
R. Farmaco, Ed. Sci. 1984, 39, 414–420.
(90) Jakopin, Z.; Dolenc, M. S. Curr. Org. Chem. 2008, 12, 850–898.
(91) Moutevelis-Minakakis, P.; Photaki, I. J. Chem. Soc., Perkin Trans.
1 1985, 2277–2281.
(92) Zhang, X.; Breslav, M.; Grimm, J.; Guan, K.; Huang, A.; Liu, F.;
Maryanoff, C. A.; Palmer, D.; Patel, M.; Qian, Y.; Shaw, C.; Sorgi,
K.; Stefanick, S.; Xu, D. J. Org. Chem. 2002, 67, 9471–9474.
(93) Mann, E.; Kessler, H. Org. Lett. 2003, 5, 4567–4570.
(94) Davies, J. R.; Kane, P. D.; Moody, C. J. Tetrahedron 2004, 60, 3967–
3977.
(95) Wipf, P.; Miller, C. P. J. Org. Chem. 1993, 58, 3604–3606.
(96) Pattarawarapan, M.; Reyes, S.; Xia, Z.; Zaccaro, M. C.; Saragovi,
H. U.; Burgess, K. J. Med. Chem. 2003, 46, 3565–3567.
(97) Paquette, L. A.; Mitzel, T. M.; Isaac, M. B.; Crasto, C. F.; Schomer,
W. W. J. Org. Chem. 1997, 62, 8960.
(98) Coburn, C. A.; Nantermet, P. G.; Rajapakse, H. A.; Selnick, H. G.;
Stauffer, S. R. (Merck & Co., Inc.). Patent WO 2006/078576 A2,
2006, 47 pp.
(99) Pettitt, B. M.; Matsunaga, T.; Al-Obeidi, F.; Gehrig, C.; Hruby, V. J.;
Karplus, M. Biophys. J. 1991, 60, 1540–1544.
(100) O’Connor, S. D.; Smith, P. E.; Al-Obeidi, F.; Pettitt, B. M. J. Med.
Chem. 1992, 35, 2870–2881.
(101) Phillips, J. C.; Braun, R.; Wang, W.; Gumbart, J.; Tajkhorshid, E.;
Villa, E.; Chipot, C.; Skeel, R. D.; Kale, L.; Schulten, K. J. Comput.
Chem. 2005, 26, 1781–1802.
(102) Humphrey, W.; Dalke, A.; Schulten, K. J. Mol. Graph. 1996, 14
(33-38), 27–38.
9
J. AM. CHEM. SOC. VOL. 133, NO. 3, 2011 469

A R T I C L E S
Ko et al.
Table 5. 1,3-Butadiyne-Based Peptidomimetics 2
secondary structures. It obviates the requirement for exploring
conformational space in the DFT method described above, i.e.,
that the user should have target conformations in mind. QMD
facilitates sampling of conformation space, and the DFT
approaches can be used to reveal if the desired conformations
are kinetically accessible.
Table 1 was formulated by considering all the ꢀs values that
are open to a peptidomimetic and comparing them to common
secondary structures. The QMD procedure described above is
a more stringent test of whether a compound can be a minimalist
mimic of a particular secondary structure. This is because CR
and Cꢀ atoms are used to cluster conformations to within a
specified rms deviation, and only low-energy conformations are
considered (because of the application of low cutoff energies
in the selection process). Thus, the QMD analyses find structures
that resemble all the conformations suggested in Table 1, but
suggest the most favored of those on the basis of Cꢀ and CR
atom orientations. Several secondary structures were represented
in the most favorable conformations found for compound 1;
these include helical and turn conformers (both ꢀ and γ).
Scheme 2. Method for Preparing Diyne-Based Mimics 2
Scheme 2 illustrates how this concept was reduced to practice
in a preparation of six peptidomimetics. Conversion of N-Boc
amino acids into the corresponding alkynes 9 was achieved
via routine chemistry with the Ohira-Bestmann modifi-
cation104-106 of the Gilbert Seyferth reaction as the pivotal
step; this proceeded without racemization. Coupling of two
different alkynes to give unsymmetrical diynes rarely pro-
ceeds with good selectivity for the heterocoupling. After some
experimentation, we found it best to convert the alkynes
bearing the least functional side chain to the corresponding
1-bromo derivatives 10 and then couple these to the parent
alkynes 9 via Cadiot-Chodkiewicz conditions.107,108 This
procedure gave the unsymmetrical products 11 in good yields
after chromatographic purification. Table 5 shows the side
chains involved in the mimics 2 that were prepared.
There is only one significant degree of freedom involved
in moving the two Cꢀ atoms closer to or farther away from
each other and simultaneously adjusting the orientation of
the CR-Cꢀ vectors. Consequently, these peptidomimetics are
Universal Peptidomimetics 2: 1,3-Butadiyne-Based. 1,3-
Butadiynes are atypical scaffolds for pharmaceuticals but are
found in natural products.103 They provide a scaffold suitable
for syntheses of minimalist peptidomimetics. One reason for
this is that both halves of the molecules 2 can be derived from
amino acids, so almost any side chain can be incorporated.
(103) Shi Shun, A. L. K.; Tykwinski, R. R. Angew. Chem., Int. Ed. 2006,
45, 1034–1057.
Figure 7. (a) Peptidomimetics 2 have only one significant degree of
freedom (R1 ) R2 ) Me shown). (b) Conformations corresponding to the
ꢀsc and ꢀse. (c) Overlay of one conformation of 2 with the i and i + 2 side
chains of a classical γ-turn.
(104) Ohira, S. Synth. Commun. 1989, 19, 561–564.
(105) Eymery, F.; Iorga, B.; Savignac, P. Synthesis 2000, 185–213.
(106) Roth, G. J.; Liepold, B.; Muller, S. G.; Bestmann, H. J. Synthesis
2004, 59–62.
9
470 J. AM. CHEM. SOC. VOL. 133, NO. 3, 2011

Peptidomimetics of All Secondary Structures
A R T I C L E S
Scheme 3. Method for Preparing Extended Mimics 3
Table 6. Kinked Bistriazole-Based Peptidomimetics 3
conformations were essentially equal in energy. Similarly, QMD
calculations indicate that every possible conformation that holds
the Cꢀ atoms at distances between 7.5 and 8.1 Å has almost the
same energy and is equally populated. These data show that this
peptidomimetic would surrender very little entropy on induced fit
docking that sets its Cꢀ atoms between these ranges.
Being so constrained, it would be expected that compounds
2 have relatively low extension factors, and Table 1 shows that
this parameter is lower than those for every other mimic in the
series A, B, 1-4. It follows that this compound should overlay
highly constrained to give Cꢀ separations between 7.5 and
8.1 Å (Figure 7), corresponding to an extension factor of
1.1 (Table 1).
CR-alkyne and alkyne-alkyne bonds in compounds 2 are
essentially free to rotate. Nevertheless, for completeness, DFT
calculations were performed on this system. They revealed that
the calculated energy maximum between the two conformations
shown in Figure 7b was less than 0.01 kcal/mol, and the
(107) Siemsen, P.; Livingston, R. C.; Diederich, F. Angew. Chem., Int. Ed.
2000, 39, 2632–2657.
(108) Yun, H.; Danishefsky, S. J. J. Org. Chem. 2003, 68, 4519–4522.
9
J. AM. CHEM. SOC. VOL. 133, NO. 3, 2011 471

A R T I C L E S
Ko et al.
Table 7. QMD Analysis of 3a
a A total of 15 families were identified, but F9-15 were omitted for
clarity. F9 and F12-15 have only one structure, and F10 and F11 have
four and five structures, respectively. All the conformers in F9-15 were
more than 0.17 kcal/mol less stable than the global minimum.
b Representative structures in the families highlighted overlay with the
secondary structures as represented in the figures accompanying this
table. c Antiparallel form. d Parallel form.
Universal Peptidomimetics 3: “Kinked” Bistriazole-Based.
Compounds that present side chains at relatively large separa-
tions are necessary in a set of universal peptidomimetics if
together they are to be comprehensive mimics of secondary
structures. Minimalist mimetics 3 and 4 are included in this set
to provide such extended Cꢀ separations.
A monoprotected diyne scaffold similar to compound 14
has been prepared before,109 so preparation of that particular
intermediate was relatively straightforward (Scheme 3).
Sequential click-deprotect-click reactions using amino acid-
derived azides afforded the mimics 3 with a diverse set of
side-chain functionalities corresponding to Glu, Ile, Leu, Lys,
Ser, Trp, and Tyr in the examples we elected to make (Table
6).
Figure 8a shows the four significant degrees of freedom for
the scaffold 3. An extension factor of 1.9, the maximum in the
series, was calculated for this framework, indicative of a
concertina-like ability to expand and contract the distance
between the Cꢀ atoms (Figure 8b). Table 2 indicates that this
design is capable of presenting Cꢀ atoms at separations that
correspond to distal side chains in a variety of secondary
structures.
Calculations for the viability of interconversion of favored
conformers (density functional theory method) show such
processes are relatively facile for compounds 3. For instance,
changing the global minimum conformation into one that
overlays with “cross strand” i-i′ + 3 residues of an anti-
parallel ꢀ-sheet involves surmounting an energy barrier of
only 4.04 kcal/mol and costs only 0.07 kcal/mol on arrival
at that conformation (Figure 8c and 8d). Thus, the conformers
are comparable in energy and will interconvert rapidly at
room temperature.
Figure 8. (a) Peptidomimetics 3 have four significant degrees of freedom
(R1 ) R2 ) Me shown). (b) Conformations corresponding to the ꢀsc and
ꢀse in an abbreviated model of structure 3 (piperazine ring omitted). (c)
Transposition of the global minima of 3 into conformation that mimics the
i and i′ + 3 residues in an anti-parallel ꢀ-sheet (∆G° values shown in kcal/
mol). (d) Overlay of the latter conformation on an anti-parallel ꢀ-sheet
(shown color coded on the left and in pink on the right).
side chains only with a limited number of secondary structures,
and again, Table 1 indicates this is so. However, it still can
overlay side chains with certain elements of ꢀ- and γ-turns and
ꢀ-sheets; i.e., it is a useful member of the universal peptido-
mimetic set. For instance, Figure 7 shows the overlay of this
with the i and i + 2 side chains of a classical γ-turn.
Table 7 collects the QMD data for an abbreviated structure
of peptidomimetic 3 to illustrate the range of conformers that
(109) Tobe, Y.; Utsumi, N.; Kawabata, K.; Naemura, K. Tetrahedron Lett.
1996, 37, 9325–9328.
9
472 J. AM. CHEM. SOC. VOL. 133, NO. 3, 2011

Peptidomimetics of All Secondary Structures
A R T I C L E S
Figure 9. Conformers from QMD studies on compounds 3 (see Table 7). (a) Family 3 overlaid with an R-helix; (b) family 4 overlaid with an antiparallel
ꢀ-sheet; and (c) family 7 overlaid onto a parallel ꢀ-sheet.
9
J. AM. CHEM. SOC. VOL. 133, NO. 3, 2011 473

A R T I C L E S
Ko et al.
Scheme 4. Synthesis of the “Linear” Bistriazole-Based
Peptidomimetics 4
Table 8. Synthesis of the “Linear” Bistriazole-Based
Peptidomimetics 4
antiparallel ꢀ-sheet (Figure 9), just as shown in the density
functional theory calculations for this molecule. The lowest
energy conformers in families 3 and 4 were only 0.03 and 0.07
kcal/mol less stable than the observed global minimum,
respectively. Members of family 7 overlaid well with the i-i + 4
residues of a parallel ꢀ-sheet (lowest energy conformer 0.13
kcal/mol above observed global minima).
Universal Peptidomimetics 4: “Linear” Bistriazole-Based.
Peptidomimetics 3 feature a 1,3-disubstituted benzene as the
central ring. An analogous 1,4-disubstituted benzene system
gives a greater possible span between the two pertinent side
chains; the resulting design is peptidomimetics 4. These
compounds can have Cꢀ-Cꢀ separations of up to 15.0 Å (Figure
10), but their extension factors are toward the smaller end of
the range given in Table 1. Scheme 4 shows the strategy that
was performed to obtain these materials; it is similar to that
used in Scheme 3, except that the dialkyne framework was built
during the synthesis rather than at the beginning. Table 8
summarizes the side chains involved and the compounds formed.
Table 9. QMD Analysis of 4a
a A total 10 families were identified, but F8-10 were omitted. F8
and F10 have only one structure, and F9 has three. All the conformers
in F8-10 were at least 0.92 kcal/mol less stable than the global
minimum. b Representative structures in the families highlighted overlay
with the secondary structures as represented in the figures accompanying
this table. c Parallel form.
are accessible. Application of an energy cutoff value of 0.3 kcal/
mol and clustering to within an rms deviation of 0.3 Å gave
eight major families. Conformers in family 3 overlaid with
R-helix i-i + 8 residues, while members of family 4 contained
conformers that overlaid well with the i-i′ + 3 residues of an
9
474 J. AM. CHEM. SOC. VOL. 133, NO. 3, 2011

Peptidomimetics of All Secondary Structures
A R T I C L E S
Figure 10. (a) Peptidomimetics 4 have four significant degrees of
freedom (R1 ) R2 ) Me shown). (b) Conformations corresponding to
the ꢀsc and ꢀse in an abbreviated model of structure 4 (piperazine ring
omitted). (c) Global minimum conformation of 4 from the density
functional theory method must surmount energy barriers of just over 3
kcal/mol to reach conformations that mimic the i and i + 4 residues in
a parallel ꢀ-sheet (∆G° values shown in kcal/mol). (d) Overlay of the
latter conformation on a parallel ꢀ-sheet (shown color coded on the left
and in pink on the right).
Figure 11. Conformers generated in QMD (see Table 7): (a) family 2
overlays with a parallel ꢀ-sheet, as in Figure 10, and (b) family 5 overlays
with an R-helix.
Table 2 indicates that peptidomimetics 4 conformers might
overlay with distal side chains of several secondary structures,
9
J. AM. CHEM. SOC. VOL. 133, NO. 3, 2011 475

A R T I C L E S
Ko et al.
including the i-i + 4 Cꢀ atoms from the same strand of a
parallel ꢀ-sheet. Figure 10 shows that relatively small energy
barriers must be surmounted to reach this conformation from
the observed global minimum, and only 0.46 kcal/mol must be
expended to do so.
is given in the Supporting Information, and an abbreviated
outline is given here.
An assay for the protein-protein interactions involving
the Bcl-2 family proteins Mcl-1 and Bid (PubChem Assay
ID (AID) 1021) showed some activities for compounds 1f
(PubChem Structure ID (SID) 24708217) and 3e (SID
24707927). Another active compound, the peptidomimetic
with scaffold A, where R1 ) n-butyl, R2 ) Lys (SID
24708065), was detected in a related study, also dealing with
interactions between Bcl-2 proteins, but this time Bcl-XL
with Bim (AID 2129). This is interesting because scaffolds
A, 1, and 3 are all pertinent to helical conformations from
the studies presented above, and interactions with helical
protein segments in a binding cleft is characteristic of
protein-protein interactions involving the Bcl-2 family
proteins.47,113
Two assays identified small molecules that inhibited the
PB1-domain interaction of MEK5 with either native or a Lys-
Ala mutant of MEKK2. Peptidomimetics 3a (SID 24708115)
and 3b (SID 24708166) and two mimics derived from
scaffold B (R1 ) Leu, R2 ) Trp, SID 24707989, and also R1
) Ile, R2 ) Trp, SID 24708162) were active in the assay
with the native MEKK2 protein (AID 1531). Compound 3b
was also active in the assay using a Lys-Ala mutant of
MEKK2 (AID 1530). In confirmation assays, compound 3b
showed inhibition of MEK5 with MEK Kinase 2 (WT, AID
1897), with an EC50 value of 5.51 µM. For the inhibition of
MEK5 with MEK Kinase 2 (Lys-Ala mutant, AID 1895),
the compound was active, with an EC50 value of 4.44 µM.
QMD analyses of peptidomimetics 4 gave seven major
families; of these, family 2 conformers overlaid well with a
parallel ꢀ-sheet, placing the i and i + 4 Cꢀ atoms at
appropriate distances (Table 9). Similarly, family 5 overlaid
with an R-helix. Clustered conformations in these two
families, and the minimum energy conformation in each, are
shown in Figure 11.
Biophysical Data. Scaffolds A and B were originally
designed to be mimics of ꢀ-turns in neurotrophins. There is
considerable evidence that the turn regions in the neurotro-
phins constitute hot-spots for their interactions with the Trk
receptors.110,111 This assertion is supported by two studies
from these laboratories that are now published. In the first,
fluorescently labeled derivatives of compounds in series A
were found to bind cells stably transfected with the TrkA
receptor.65 In the second study, compounds in series B were
subjected to direct binding, survival and neuritogenic, and
biochemical signal transduction assays; these revealed eight
agonistic ligands binding to the ectodomain of TrkC. These
peptidomimetics afford discrete signals leading to either cell
survival or neuritogenic differentiation, effects that are both
attributed to the parent neurotrophin, NT-3.64
Universal peptidomimetics are likely to be most useful in
the design of libraries for high-throughput screening against
diverse targets. This is because they can mimic a range of
secondary structures and present any protein amino acid side
chain; this impacts the diversity space that corresponds to
hot-spots112 in protein-protein interactions. Universal pep-
tidomimetics are also likely to be useful for known targets
where exact binding conformations are unknown. This
represents a large number of situations, including all those
for which only one protein in a protein-protein interaction
has been structurally characterized and those for which solid-
state structure analyses show a possible mode of association
that is actually different in solution or in a cellular environ-
ment. Conversely, rigid peptidomimetics that are not universal
tend to be more useful when the binding conformation is
known with a high degree of confidence and can be accurately
matched by the mimic.
Conclusion
1,3,4-Oxazoline-based mimics 1 were predicted to be useful
mimics of many pairs in amino acids in conformations that
resemble all common secondary structures (Table 2). Modeling
data for this scaffold indicated that conformations with a diverse
set of ꢀs values would be accessible; a typical energy maximum
for interconversion of the conformers was only 2.43 kcal/mol
(Figure 5). This assertion was born out by the range of
conformations that emerged from the QMD studies, which also
highlighted low-energy conformers that resemble elements of
R-helix, γ-turn, and ꢀ-turn structures.
There are insignificant barriers to interconversion of the 1,3-
butadiyne peptidomimetics 2 between conformational states that
resemble amino acid pairs in γ-turns, ꢀ-turns, and ꢀ-sheets.
These are relatively tight minimalist mimics, having only one
significant degree of freedom for separation of the Cꢀ atoms,
hence their small extension factors, but it takes very little energy
to interconvert between them. Conversely, the kinked bistriazole
scaffolds 3 allow high ꢀs variability. Favored conformations
of this scaffold can mimic, for instance, relatively close residues
of an antiparallel ꢀ-sheet or closer side chains in R-helices. All
possible Cꢀ atom separations are calculated to be accessible, as
they are for all the mimics 1-4. Mimics 4 are similar to 3 except
that they orient the triazole units para rather than meta. This
scaffold is a tool for mimicking relatively distal side chains,
even in relatively extended structures like the i and i + 4 Cꢀ
atoms from one strand of a parallel ꢀ-sheet.
Motivated by the reasoning presented above, we submitted
compounds A, B, 1, and 3 (and protected precursors to these)
to the NIH Molecular Libraries Small Molecule Repository
(MLSMR) for screening. This is ideal for testing these
compounds, many of which were not designed for any
particular target but simply as secondary structure mimics
in general. Screening centers test compounds in the MLSMR
and publish the data on PubChem; thus, samples submitted
from this work can be exposed to a wide variety of
screenssmany more than it would be practical to establish
in one laboratory. These screening centers have already
identified cases where our molecules promote or inhibit
protein-protein interactions of particular interest. A summary
Over the past decade, there has been considerable interest
in generic libraries prepared to supply high-throughput
(110) Pattarawarapan, M.; Burgess, K. J. Med. Chem. 2003, 46, 5277–
5291.
(111) Saragovi, H. U.; Burgess, K. Exp. Opin. Ther. Patents 1999, 9, 737–
(113) Youle, R. J.; Strasser, A. Nat. ReV. Mol. Cell Biol. 2008, 9, 47–59.
(114) Burgess, K. (Texas A&M University System, College Station, TX).
U.S. Patent 2009/0264315 A1, 2009, pp 1-103.
751.
(112) Leigh, D. A. Chem. Biol. 2003, 10, 1143–1144.
9
476 J. AM. CHEM. SOC. VOL. 133, NO. 3, 2011

Peptidomimetics of All Secondary Structures
A R T I C L E S
LBMS-Applications Laboratory, headed by Dr. Shane Tichy and
Dr. Yohannes Rezenom, provided mass spectrometric support,
and Dr. Gillian Lynch at University of Houston supported our
QMD work. We thank Shuhei Shimizu for some additional
experiments.
screening projects, particularly in the pharmaceutical industry,
with “drug-like” compounds to test. The small heterocyclic
structures 1-4 and B can be decorated with the most pertinent
pharmacophores for peptide and protein targets: amino acid
side chains. These specific non-peptidic structures may have
some commercial potential, and a patent application on them
has been published.114 However, it should not be difficult to
conceive other scaffolds for universal peptidomimetics, and
it is our hope that the concepts described here will facilitate
this process.
Supporting Information Available: Procedures for modeling
experiments, synthesis, and characterization data for all pepti-
domimetics, and complete ref 34. This material is available free
of charge via the Internet at http://pubs.acs.org.
Acknowledgment. Financial support for this project was
provided by the National Institutes of Health (MH070040,
GM076261) and the Robert A. Welch Foundation. The TAMU/
JA1071916
9
J. AM. CHEM. SOC. VOL. 133, NO. 3, 2011 477