New β-strand templates constrained by Huisgen cycloaddition

Article abstract of DOI:10.1021/ol3002199

New peptidic templates constrained into a β-strand geometry by linking acetylene and azide containing P₁ and P₃ residues of a tripeptide by Huisgen cycloaddition are presented. The conformations of the macrocycles are defined by NMR studies and those that best define a β-strand are shown to be potent inhibitors of the protease calpain. The β-strand templates presented and defined here are prepared under optimized conditions that should be suitable for targeting a range of proteases and other applications requiring such a geometry.

Full text of DOI:10.1021/ol3002199

ORGANIC
LETTERS
2012
Vol. 14, No. 5
1330–1333
New β-Strand Templates Constrained by
Huisgen Cycloaddition
Ashok D. Pehere and Andrew D. Abell*
School of Chemistry & Physics, The University of Adelaide, North Terrace, Adelaide SA
5005, Australia
andrew.abell@adelaide.edu.au
Received January 27, 2012
ABSTRACT
New peptidic templates constrained into a β-strand geometry by linking acetylene and azide containing P1 and P3 residues of a tripeptide by
Huisgen cycloaddition are presented. The conformations of the macrocycles are defined by NMR studies and those that best define a β-strand are
shown to be potent inhibitors of the protease calpain. The β-strand templates presented and defined here are prepared under optimized conditions
that should be suitable for targeting a range of proteases and other applications requiring such a geometry.
The biological function of a peptide is inadvertently
linked to its conformation or shape. With this in mind,
much work has been reported on covalently constraining
a linear peptide into a well-defined conformation, such as
an R-helix or a β-strand, to provide important bio-
logical probes, scaffolds, and enzymes inhibitors.1ꢀ13 Of
particular significance is the use of RCM,4 intramolecular
alkyltion,14 and other methods of macrocyclization15 to
constrain a protease inhibitor into a biologically active β
strand conformation.
The Huisgen cycloaddition reaction of an acetylene
with an azide offers much promise in this context, with
reports on its use to prepare macrocyclic peptidomimetics
to improve biostability and restrict conformational mo-
bility, but often in an ill-defined way.5 What we lack are
rationally designed templates that have a well-defined
and predictable conformation and hence wide and gen-
eral applicability. Here we report an azide alkyne Huisgen
(1) Loughlin, W. A.; Tyndall, J. D. A.; Glenn, M. P.; Hill, T. A.;
Fairlie, D. P. Chem. Rev. 2010, 110, Pr32–Pr69.
(2) Smith, A. B.; Charnley, A. K.; Hirschmann, R. Acc. Chem. Res.
2011, 44, 180–193.
(3) Vasudev, P. G.; Chatterjee, S.; Shamala, N.; Balaram, P. Chem.
Rev. 2011, 111, 657–687.
(4) Abell, A. D.; Jones, M. A.; Coxon, J. M.; Morton, J. D.; Aitken,
S. G.; McNabb, S. B.; Lee, H. Y. Y.; Mehrtens, J. M.; Alexander, N. A.;
Stuart, B. G.; Neffe, A. T.; Bickerstaff, R. Angew. Chem., Int. Ed. 2009,
48, 1455–1458.
(5) Pedersen, D. S.; Abell, A. Eur. J. Org. Chem. 2011, 2399–2411.
(6) Wyrembak, P. N.; Hamilton, A. D. J. Am. Chem. Soc. 2009, 131,
4566–4567.
(7) Nowick, J. S. Acc. Chem. Res. 2008, 41, 1319–1330.
(8) Zheng, J.; Liu, C.; Sawaya, M. R.; Vadla, B.; Khan, S.; Woods,
R. J.; Eisenberg, D.; Goux, W. J.; Nowick, J. S. J. Am. Chem. Soc. 2011,
133, 3144–3157.
(9) Liu, C.; Sawaya, M. R.; Cheng, P. N.; Zheng, J.; Nowick, J. S.;
Eisenberg, D. J. Am. Chem. Soc. 2011, 133, 6736–6744.
(10) Phillips, S. T.; Rezac, M.; Abel, U.; Kossenjans, M.; Bartlett,
P. A. J. Am. Chem. Soc. 2002, 124, 58–66.
(11) Freire, F.; Almeida, A. M.; Fisk, J. D.; Steinkruger, J. D.;
Gellman, S. H. Angew. Chem., Int. Ed. 2011, 50, 8735–8738.
(12) Janetka, J. W.; Raman, P.; Satyshur, K.; Flentke, G. R.; Rich,
D. H. J. Am. Chem. Soc. 1997, 119, 441–442.
Figure 1. Target macrocycles.
(13) White, C. J.; Yudin, A. K. Nat. Chem. 2011, 3, 509–524.
r
10.1021/ol3002199
Published on Web 02/16/2012
2012 American Chemical Society

cycloaddition-based synthesis of P1 and P3-linked pep-
tides16 and a conformational study to identify those that
best mimic a β-strand geometry. As proof of concept,
these macrocycles contain a C-terminal aldehyde to pro-
vide inhibitors of the protease calpain, which like all
proteases, is known to favor the binding of substrates
and inhibitors in this conformation.17 The inhibition of
calpain is of particular therapeutic significance since it is
involved in critical diseases,18ꢀ21 with few examples of
such cysteine proteases having reached the clinic despite
this need.22,23 The macrocycles targeted in this study
contain a component aryl group, in addition to the
triazole, to further constrain the geometry (Figure 1).
One series has this group at P3 (see 1 and 2) and the other
at P1 (see 3ꢀ5). In addition, a hydrophobic amino acid
was inserted at P2 to accommodate the known binding
preferences of calpain.4 To the best of our knowledge, this
is the first report on using a click reaction to define a
specific β-strand geometry.
give 8. Removal of the Boc group and reduction with
NaBH4 then gave 9. Reaction of N-Cbz-Tyr-OH 624 with
SOCl2 and methanol, followed by reaction with propar-
gyl bromide in presence of K2CO3, gave 10. The methyl
ester was then hydrolyzed to give 11. The azides 13aꢀd
were prepared from amines 12aꢀd with the key step being
reaction with Tf2O and NaN3.25
The macrocyclic aldeydes were then prepared as shown
in Schemes 2 and 3. Reaction of 11 with Leu-OMe in the
presence of EDCI, HOBt gave dipeptide 14, which was
hydrolyzed to give the acid 15. Separate coupling of 15
with 13a and 13b, in the presence of EDCI and HOBt,
gave tripeptides 16 and 17, respectively. These were then
cyclized, in the presence of Cu(I)Br in CH2Cl2, and the
resulting macrocyclic esters 18 and 19 reduced with
lithium borohydride to give the macrocyclic alcohols 20
and 21 that were purified by HPLC. Oxidation of these
alcohols, with Dess-Martin periodinane (DMP),23 gave
the required aldehydes 1 and 2 (Scheme 2).
The synthesis of macrocycles 1 to 5 required the pre-
paration of the key acetylenes 9 and 11 and the azides
13aꢀd as shown in Scheme 1. The acetylene 9 was
prepared from N-Boc-Tyr-OH 7 by methylation with
MeI, followed by alkylation with propargyl bromide to
Scheme 2. Synthesis of Macrocylic Aldehydes 1 and 2
Scheme 1. Synthesis of Key Building Blocks
(18) Vinciguerra, M.; Musaro, A.; Rosenthal, N. Adv. Exp. Med.
Biol. 2010, 694, 211–233.
(19) Storr, S. J.; Carragher, N. O.; Frame, M. C.; Parr, T.; Martin,
S. G. Nat. Rev. Cancer 2011, 11, 364–374.
(20) Smith, I. J.; Lecker, S. H.; Hasselgren, P. O. Am. J. Physiol.
Endocrinol. Metab. 2008, 295, E762–771.
(21) Wadosky, K. M.; Li, L.; Rodriguez, J. E.; Min, J. N.; Bogan, D.;
Gonzalez, J.; Patterson, C.; Kornegay, J. N.; Willis, M. Muscle. Nerve
2011, 44, 553–562.
(14) Reid, R. C.; Kelso, M. J.; Scanlon, M. J.; Fairlie, D. P. J. Am.
Chem. Soc. 2002, 124, 5673–5683.
(22) Pietsch, M.; Chua, K. C. H.; Abell, A. D. Curr. Top. Med. Chem.
2010, 10, 270–293.
(15) Marsault, E.; Peterson, M. L. J. Med. Chem. 2011, 54, 1961–
2004.
(23) Loser, R.; Abbenante, G.; Madala, P. K.; Halili, M.; Le, G. T.;
Fairlie, D. P. J. Med. Chem. 2010, 53, 2651–2655.
(24) Pehere, A. D.; Abell, A. D. Tetrahedron Lett. 2011, 52, 1493–
1494.
(25) Stanley, N. J.; Pedersen, D. S.; Nielsen, B.; Kvist, T.; Mathiesen,
J. M.; Brauner-Osborne, H.; Taylor, D. K.; Abell, A. D. Bioorg. Med.
Chem. Lett. 2010, 20, 7512–7515.
(16) Note the use of Schechter-Berger nomenclatureSchechter, I.;
Berger, A. Biochem. Biophys. Res. Commun. 1967, 27, 157. The residues
on the N-terminal side of the peptide bond that is cleaved are denoted (in
0
0
order) P1ꢀPn, and those on the C-terminus are denoted P1 ꢀPn . In turn,
0
the corresponding subsites on the enzyme are denoted SnꢀSn .
(17) Madala, P. K.; Tyndall, J. D. A.; Nall, T.; Fairlie, D. P. Chem.
Rev. 2010, 110, Pr1–Pr31.
Org. Lett., Vol. 14, No. 5, 2012
1331

The remaining macrocyclic aldehydes 3ꢀ5 and the
acyclic analogues 34 and 35 were prepared using a
modified protocol as shown in Scheme 3. Unlike the
sequence shown in Scheme 2 (see step v) the ester func-
tionality was reduced to the corresponding alcohol (9)
prior to coupling and cyclization. This has the advantage
of a reduced reaction time for the reduction and also
removal of the need for purification of the macrocycle by
HPLC, as was the case for 20 and 21 in Scheme 2. As such
separate reaction of 13c and 13d with Leu-OtBu and Phe-
OMe, in the presence of EDCI and HOBt, gave dipeptides
22ꢀ24 that were separately hydrolyzed to give 25ꢀ27.
Coupling of each of these with the alcohol 9, in the
presence of EDCI and HOBt, gave tripeptides 28ꢀ30
that were cyclized on treatment with Cu(I)Br in CH2Cl2
to give 31ꢀ33 in good yields. The macrocyclic alcohols
31ꢀ33 and the corresponding acyclic alcohols 28 and 29,
were oxidized with Dess-Martin periodinane (DMP) to
give the required aldehydes 3ꢀ5, 34, and 35, respectively.
The preparation of macrocycle 32 via an analogous route
to that used in Scheme 2 resulted in significantly reduced
yield.
Scheme 3. Synthesis of Aldehydes 3ꢀ5, 34, and 35
The details of the CuAAC catalyzed cyclizations shown
in Schemes 2 and 3 are worthy of comment. The cycliza-
tion of related peptidic structures (using Cu(I)/Cu-
(CH3CN)4PF6) is reported to give competing dimer and
trimer formation,26,27 necessitating purification by
HPLC. In addition, elevated temperatures are often
reported for such reactions.28 Our improved methodol-
ogy is mild, high yielding, and occurs without competing
dimerization to give the macrocycle that is purified by
simple silica-based chromatography.
Conformation of the Macrocycle. Solution structures for
the macrocycle aldehydes 1ꢀ5 and the alcohols 20, 21 and
3
observations. Characteristic29ꢀ31 NOE interactions were
observed between CRHi and (iþ1NH), βHi and (iþ1NH),
NHi and (iþ1NH) for 3, 5, 31, 32, 33, see Supporting
Information Figure S1 for an example. Thus, the macro-
cycles containing a triazole at P3 and a tyrosine analog at
P1 (see 3, 5, 31, 32, 33) appear to adopt β-strand con-
formation known to favor binding to a protease.
31ꢀ33 were determined based on JNHCRH coupling
constants, see Table 1. The magnitude of this coupling
constant is dependent on the angle Φ, as defined by the
local conformation of the polypeptide backbone.1,10,14
For a β-sheet conformation these values are typically in
the range 8 to 10 Hz, while for an unstructured random
coil a value of 5.8 to 7.3 Hz is typical.10 Macrocycles
3
3, 5, 31, 32, 33 all displayed JNHCRH coupling constant
Variable temperature NMR studies on the macrocycle
5 in DMF-D7 showed its β-starnd conformation to
be stable within the temperature range 223ꢀ353 K. The
temperature dependence coefficients (Δδ/ΔT) for the two
NH’s within the macrocycle of 5, over this temperature
range, were calculated to be 6.5 and 5.0 ppb Kꢀ1. This
large temperature dependence is characteristic of a lack of
intramolecular hydrogen bond between the P3 CdO and
P1 NH and hence a β-starnd conformation.31,32 This
conformational stability is likely associated with a com-
bination of the constituent phenyl and triazole rings,
which help to lock the macrocycle into the β-strand
conformation. Unlike other macrocyclic protease inhibi-
tors the constituent phenyl group of 5 appears to be
rotationally rigid, as evidenced by distinct resonances
>8 Hz, which is consistent with a β-strand conformation.
Only one coupling constant could be determined for 4
because of overlapping resonances. By comparison, com-
pounds 1, 2, 20 and 21 appear not to adopt a β-strand
conformation based on the JNHCRH coupling con-
stants. The determination of NOE data confirmed these
3
(26) Liu, Y. Q.; Zhang, L. H.; Wan, J. P.; Li, Y. S.; Xu, Y. H.; Pan,
Y. J. Tetrahedron 2008, 64, 10728–10734.
(27) Zhang, J. Q.; Kemmink, J.; Rijkers, D. T. S.; Liskamp, R. M. J.
Org. Lett. 2011, 13, 3438–3441.
(28) Bock, V. D.; Perciaccante, R.; Jansen, T. P.; Hiemstra, H.; van
Maarseveen, J. H. Org. Lett. 2006, 8, 919–922.
(29) Wuthrich, K.; Billeter, M.; Braun, W. J. Mol. Biol. 1983, 169,
949–961.
(30) Wuthrich, K.; Billeter, M.; Braun, W. J. Mol. Biol. 1984, 180,
715–740.
(31) Weide, T.; Modlinger, A.; Kessler, H. Top. Curr. Chem. 2007,
272, 1–50.
(32) Kessler, H. Angew. Chem., Int. Ed. Engl. 1982, 21, 512–523.
Org. Lett., Vol. 14, No. 5, 2012
1332

1
for each of the four aromatic protons in its H NMR
spectrum.14 These remain sharp over the temperature
conformation constraint imposed by the macrocycle im-
proves potency.
range 223ꢀ353 K, see Supporting Information Figure S3.
Table 2. In Vitro Inhibition Data
Table 1. 3JNHCRH Coupling Constantsa
no.
R
n ring size β-strand conformation IC50b (μM) calpain II
no.
Leu NHꢀCHR
Tyr NHꢀCHR
Phe NHꢀCHR
1
2
3
4
5
Leu 1
Leu 4
Leu 3
Leu 4
Phe 4
18
21
20
21
21
No
No
1.020
0.940
0.137
0.097
0.089
0.780
1.030
1
7.4
8.1
8.4
7.6
7.9
7.5
Yes
a
2
3
8.4
b
Yes
4
34 Leu 3
35 Leu 4
5
8.4
7.1
8.1
9.0
8.5
8.8
8.4
20
21
31
32
33
8.5
7.9
8.2
8.7
a Not determined because of overlapping resonances. b Values are the
mean of three experiments and variation between experiments is < ( 10%.
8.4
In summary, we present an optimized azide alkyne
Huisgen cycloaddition-based synthesis of macrocyclic
peptidomimetics constrained into a well-defined β-strand
geometry designed to facilitate binding to the proteases
calpain. Introduction of the macrocycle improves po-
tency, with examples adopting a β-strand geometry pro-
viding the optimum template for binding. The β-strand
templates presented and defined here are prepared under
chemically benign conditions and should be suitable for
targeting a range of proteases and other applications
requiring such a geometry. The conditions used for the
macrocyclisation avoid competing dimer and trimer for-
mation that have been reported for related systems.
a Coupling constants determined in DMSO-d6. b Not determined
because of overlapping resonances.
Biological Data. Compounds 1ꢀ5, and the acyclic analo-
gues 34 and 35 were assayed against calpain-II using a
fluorescence-based assay33 to determine in vitro potency,
with the results shown in Table 2. The 21-membered
macrocyclic aldehydes 4 and 5 were particularly potent,
with IC50 values against calpain II of 97 nM and 89 nM,
respectively. The nature of the hydrophobic P2 substitu-
ent has little effect on activity, with both Leu (4) and Phe
(5) being well tolerated at this position. The related 20-
membered macrocycle 3 had a slightly reduced potency
(IC50 of 137 nM). This data is consistent with the con-
formational analysis that showed that macrocycles 3 and
5 adopt a β-strand conformation that is known to favor
protease binding. The conformation of 4 could not be
defined but is presumed to also be a β-strand because of
its close similarity to 5. The macrocyclic aldehydes with
the triazole at P1 (see 1 and 2) were significantly less
potent against calpain II, presumably since they do not
appear to adopt a β-strand geometry. Importantly, the
acyclic aldehydes 34 and 35 were significantly less potent
than their cyclic analogues 3 and 4 respectively. Thus, the
Acknowledgment. We acknowledge Drs. M. Pietsch
(University of Adelaide) and D. Sejer Pedersen
(University of Adelaide) for preliminary discussions, As-
soc. Prof. J. Morton (Lincoln University) for the supply of
calpain, Dr. O. Zvarec (University of Adelaide) for assis-
tance with the calpain assays, and financial support from
Australian Research Council.
Supporting Information Available. Experimental pro-
cedures, characterization data, and spectra of new com-
pounds. This material is available free of charge via the
Internet at http://pubs.acs.org.
(33) Thompson, V. F.; Saldana, S.; Cong, J. Y.; Goll, D. E. Anal.
Biochem. 2000, 279, 170–178.
The authors declare no competing financial interest.
Org. Lett., Vol. 14, No. 5, 2012
1333