Site-specific integration of amino acid fragments into cyclic peptides

Article abstract of DOI:10.1021/ja412256f

The concept of site-specific integration of fragments into macrocyclic entities has not yet found application in the realm of synthetic chemistry. Here we show that the reduced amidicity of aziridine amide bonds provides an entry point for the site-specif

Full text of DOI:10.1021/ja412256f

Communication
pubs.acs.org/JACS
Site-Specific Integration of Amino Acid Fragments into Cyclic
Peptides
Christopher J. White, Jennifer L. Hickey, Conor C. G. Scully, and Andrei K. Yudin*
Davenport Research Laboratories, Department of Chemistry, University of Toronto, 80 St. George Street, Toronto, Ontario M5S
3H6, Canada
S
* Supporting Information
ABSTRACT: The concept of site-specific integration of
fragments into macrocyclic entities has not yet found
application in the realm of synthetic chemistry. Here we
show that the reduced amidicity of aziridine amide bonds
provides an entry point for the site-specific integration of
amino acids and peptide fragments into the homodetic
cyclic peptide architecture. This new synthetic operation
improves both the convergence and divergence of cyclic
peptide synthesis.
yclic peptides have been gaining traction in both drug
1
Cdiscovery and biological probe design. Significant promise
lies in the structural attributes of these molecules.2 Their large
Figure 1. General synthetic strategy for the integration of molecular
fragments into cyclic peptides. Box: concept of integration by site-
specific fragment incorporation.
surface area, capacity to “wrap” polar functionalities,3 and
resistance to proteolytic degradation4 make macrocycles
attractive tools for the interrogation of challenging biological
targets such as protein−protein interactions.5 Although a
Distorted amides exhibit fast transacylation kinetics, often
number of natural products with large ring structures are
known, macrocycles are still underrepresented in the screening
collections of both the pharmaceutical industry and academic
centers. This paucity of accessible structures has fueled interest in
developing new methods for peptide cyclization. Although great
strides in devising ways to close the ring have been made,6 there
exist few approaches that allow for selective modification of the
macrocycle core itself.7
comparable to that of activated esters or even acyl halides.13 This
is in part due to significant amide bond distortion as a result of
the nitrogen lone-pair pyramidalization. Because aziridine
amides can be chemoselectively hydrolyzed,14 we set out to
explore this property as a synthetic means to realize the concept
of site-specific integration. If site-specific linearization of an Azy-
containing macrocycle followed by coupling with the desired
integration fragment and macrolactamization could be per-
formed in the same pot, the resulting synthetic sequence would
exemplify such a process. A final aziridine ring-opening reaction
would then generate the native peptide bond (Figure 1). In
target-oriented applications, such a method could improve the
overall convergence of the synthesis, whereas in the diversity-
oriented domain of synthesis, this method could allow for split/
pool applications and enable integration of rare and unnatural
fragments that may not be amenable to activation/coupling
strategies.
Our recent studies in peptide macrocyclization using aziridine
aldehyde dimers8 prompted us to explore the stereoelectronic
features of aziridine amides in other contexts. N-Acyl aziridines
stand out as direct precursors to both natural and unnatural
amino acid residues in homodetic cyclic peptides through the use
of the nonstandard amino acid aziridine-2-carboxylic acid (Azy)
and its derivatives.9a,b Gin and van der Donk showed that ring
opening of Azy residues with thiol nucleophiles provides a direct
route to backbone-modified peptides.10 A disadvantage of N-acyl
aziridines, however, is their susceptibility toward amide
hydrolysis.11 Here we show that this susceptibility can be turned
into an enabling tool for site-specific integration of molecular
fragments into cyclic peptides (Figure 1). There exist several
examples showcasing function-driven integration of “foreign”
fragments into existing biological molecules. Thus, the retroviral
enzyme integrase binds both termini of viral DNA and inserts
them into a host-cell chromosome.12 Although site-specific
integration is well-known in nature, this concept has yet to find
application in the realm of synthetic chemistry.
We used our recently reported protocol9b to synthesize the
Azy-containing cyclic tetrapeptide template 1. With aqueous
LiOH, the aziridine amide bond of 1 was site-selectively cleaved
within 1 h to generate the linear tetrapeptide 2 (Scheme 1). N-H
Azy-containing peptides with a free carboxylic acid have been
reported to be unstable toward purification and storage,15 yet we
found the linearized material 2 to be stable toward handling and
prolonged storage as long as the C-terminal carboxylic acid exists in
Received: December 2, 2013
Published: February 17, 2014
© 2014 American Chemical Society
3728
dx.doi.org/10.1021/ja412256f | J. Am. Chem. Soc. 2014, 136, 3728−3731

Journal of the American Chemical Society
Communication
Scheme 1. Hydrolysis of cyclo[Leu-3-MeAzy-Phe-Gly]
the form of a carboxylate salt.16 From here, we were ready to
couple the integration fragment.17 Since the aziridine nitrogen
remains unprotected upon linearization, we feared that acylation
of the N-H aziridine would compete with the acylation of the
desired amine during the ligation step. The N-H aziridine is
indeed a competent nucleophile, for when we stirred 2 with a
coupling reagent in the absence of an external amine nucleophile,
we observed the formation of a mixture of oligomeric products of
2.
Figure 2. Oligomer recycling: a solution to undesired polymerization.
We chose H-Gly-OEt 3a as our integration fragment and
screened various coupling reagents in hopes of maximizing the
selectivity for the formation of pentapeptide 4a (Scheme 2). The
2-carboxylic acid (Aze) (entry 9). This four-membered ring
analogue of proline possesses significant ring strain (25.4 kcal/
mol) comparable to that of its cousin Azy (27 kcal/mol), yet it
survives our synthetic sequence of fragment integration.
Scheme 2. Chemoselective Ligation of Unprotected Azy-
tetrapeptide 2
We then sought to develop a telescopic transformation to
further diversify the N-acyl aziridines 5 in a site-specific fashion
by way of nucleophilic opening of their aziridine rings. We turned
to azide as a nucleophile.21 It is well-known that azide can
function as a leaving group or as a precursor to an amine or a
tetrazole which leads to further site-specific functionalization.22
When the cyclization to furnish 5 was complete, the reaction
mixture was treated with sodium azide. For substrates 5a−p, the
azide anion attacked the β-carbon of the aziridine in a
regioselective fashion, generating the corresponding azido-
functionalized α-amino acid-containing cyclic peptide 6′. A
second product 6″, corresponding to hydrazoic acid elimination,
was formed to a varying extent. In several cases, a third product
6‴, in which HOAt opened the aziridine ring, was observed in
varying amounts. This occurred as a result of HOAt being
present from the previous macrocyclization step with AOP.
Importantly, treatment with DBU at 60 °C was shown to effect the
full elimination of both the azide and HOAt-containing macrocyles
forming 6″. Our late-stage, one-pot aziridine ring opening/azide
elimination should find application in the design of electrophilic
biological probes inspired by microcystin.23
We also considered the integration of more exotic amino acid-
containing fragments. Starting from 1, we were able to
successfully ligate the unnatural amino acid fragments 3p and
3q (Figure 3a).24 After saponification of the C-terminal ester and
subsequent cyclization with AOP, we were able to isolate the
aziridine-containing macrocycles 5p and 5q in moderate yields
over a four-step sequence (Figure 3a). We then investigated
aziridine ring-opening reactions of these scaffolds with sodium
azide. When 5p was treated with excess sodium azide, results
similar to those for cyclic peptides 6j−o were observed (Figure
3a). A highly regioselective attack of the azide anion on the N-
acyl aziridine took place at the β-carbon to furnish 6p′, whose
formation was accompanied by minor amount of the elimination
product 6p″. In contrast, the regioselectivity of azide attack on 5q
was very different from that of 5p. Here, the regioisomer in which
azide attacks the α-carbon of the aziridine ring, 6q, was formed as
the major product in the reaction, and the opposite regioisomer
6q′ was formed as a minor product. The observed macrocycle-
dependent regioselectivity of aziridine ring-opening speaks to the
adoption of different reactive conformations of the N-acyl
aziridine depending on the neighboring residues and ring size.
a
HPLC yield
coupling reagent (7-azabenzotriazol-1-yloxy)tris(dimethyl-
amino)phosphonium hexafluorophosphate (AOP) was found
to be superior in this reaction (see the Supporting Information).
The success of this chemoselective coupling is rooted in the
difference between the rates of acylation of primary amines
(pKaH = 9−11) and aziridines (pKaH = 7.9). With this reagent, we
have coupled a variety of different amino acid- or peptide-based
fragments in high yields with no detectable epimerization (see
Table 1). Minor oligomeric byproducts formed through the
combination of multiple molecules of 2 and one molecule of the
integration fragment 3 were observed in varying amounts,
depending on the reaction conditions and coupling partner.
However, these byproducts are not a dead end in our synthetic
scheme, as we can leverage the reduced amidicity of the aziridine
amide during the subsequent saponification, which simulta-
neously cleaves the C-terminal alkyl ester and the internal
aziridine amide of the oligomeric byproducts, liberating one
molecule of the desired product 4′ and n molecules of the starting
material 2 (Figure 2).18
Saponification of 4a with LiOH yielded 4a′ as a carboxylate
salt. A subsequent lactamization was performed19 with AOP.20a,b
The desired cyclic pentapeptide was formed with no detectable
oligomerization by HPLC. Using our strategy, we were able to
site-specifically insert a variety of amino acid residues into 1
(Table 1, entries 1 and 3−8). Critically, there was no evidence for
oxazoline formation, a well-recognized challenge with linear acyl
aziridines.20c We next sought to insert larger peptide fragments
into 1. Integration of the dipeptides Gly-Gly and Phe-D-Pro
furnished the corresponding cyclic hexapeptides, whereas
integration of the tripeptides Gly-Phe-D-Pro and Ala-Phe-D-Pro
gave the corresponding cyclic heptapeptides (entries 12−15).
Unnatural amino acids can also be used as integration fragments
(entries 2, 9, and 10). Of particular interest is the use of azetidine-
3729
dx.doi.org/10.1021/ja412256f | J. Am. Chem. Soc. 2014, 136, 3728−3731

Journal of the American Chemical Society
Communication
Table 1. Integration Substrate Scope
integration fragment
R1
ligation
yield (%)
cyclization
yield (%)
Azy ring-opening
a
a
b
c
entry
R2
product
product
product ratio (6′:6″:6‴)
yield (%)
1
2
Gly
Sar
Et
4a
4b
4c
4d
4e
4f
98
77
96
94
95
85
87
83
97
95
95
99
92
92
92
5a
5b
5c
5d
5e
5f
96
84
91
94
88
88
90
86
92
95
94
86
97
96
93
74:13:13
51:49:0
78:6:16
65:4:31
69:19:12
67:26:7
67:13:20
70:11:19
30
14
17
10
17
35
36
18
10
14
48
31
19
17
40
Me
Me
Me
Me
Me
Me
Me
Me
Me
Me
Me
Me
Me
Me
d
d
d
d
d
d
3
Ala
4
Leu
Val
5
6
Phe
Tyr
Trp
Aze
β-Ala
7
4g
4h
4i
5g
5h
5i
8
e
9
66 :18:16
10
11
12
13
14
15
4j
5j
51:49:0
61:34:5
43:57:0
41:58:1
33:67:0
40:60:0
d
Nα-Ac-Lys
Gly-Gly
4k
4l
5k
5l
Phe-D-Pro
Gly-Phe-D-Pro
4m
4n
4o
5m
5n
5o
Ala-Phe-D-Pro
a
b
c
HPLC yields. Determined by peak integration in crude HPLC traces. Isolated yields of combined 6′, 6″, and 6‴ (after HPLC purification) with
d
respect to 1 after five steps. Cα-epimerization of the integration fragment was detected during the cyclization step. For the extent of epimerization,
see the Supporting Information. The two regioisomers of azide attack were formed in a 43:57 ratio (HPLC).
e
Figure 3. (a) Site-specific integration of unnatural amino acids. Notes: aHPLC yield of the cyclization step; bisolated yield over four steps; cdetermined
by peak integration of crude HPLC traces. (b) Proposed model for regioselectivity in Azy ring openings of cyclic peptides.
The so-called “perpendicular” conformation that is stereo-
electronically favored for N-acyl aziridines9c is expected to be the
dominant contributor to the transition state for ring opening. We
reason that the conformation of the aziridine amide plays a
significant role in rationalizing our observed regioselectivity, as
this factor is translated into the transition state for aziridine ring
opening (buildup of A1,3 strain) (Figure 3b). When the adaptive
aziridine amide9d adopts a cis conformation and azide attacks the
α-carbon of Azy, a trans-like amide forms in the transition state.
Alternatively, attack at the β-carbon of Azy leads to the
development of a less favorable cis-like amide bond. The
observed regiochemistry for the reaction with 1 is in agreement
with this hypothesis.9b The larger Azy-containing macrocycles
5a−p and linear Azy-containing peptides10 have additional
flexibility compared with 1 and thus would accommodate an
aziridine amide bond in its trans conformation. In this case our
transition-state model is reversed, and attack of azide at the β-
carbon leads to the development of a more favorable trans-like
amide bond in the transition state. We note that this is the major
isomer observed with macrocycles 5a−p as well as with the ring
opening of linear Azy-containing peptides with thiol-based
nucleophiles.10,25
3730
dx.doi.org/10.1021/ja412256f | J. Am. Chem. Soc. 2014, 136, 3728−3731

Journal of the American Chemical Society
Communication
(9) (a) Nakajima, K.; Takai, F.; Tanaka, T.; Okawa, K. Bull. Chem. Soc.
Jpn. 1978, 51, 1577. (b) White, C. J.; Yudin, A. K. Org. Lett. 2012, 14,
In closing, we have capitalized on the susceptibility of N-acyl
aziridines to amide hydrolysis, which is often considered to be
their “Achilles’ heel”, to develop a tool for site-specific
incorporation of molecular fragments into homodetic cyclic
peptides. The reduced amidicity of the N-acyl aziridine linkage
facilitated site-selective hydrolysis and was important in two
additional embodiments: it enabled us to avoid the well-known
problem of diketopiperazine formation,16 and it made the
problem of oligomerization a correctible mistake. By stressing
the reversibility of N-acyl aziridine formation under mild
conditions, our study paves a way to thermodynamically
controlled applications of active amides in dynamic combinato-
rial chemistry.26 In the course of our study, we also discovered
the peculiar stereoelectronic consequences of placing N-acyl
aziridine units into homodetic cyclic peptides of different sizes.
The products of this integrative chemistry are useful building
blocks that can be employed for the late-stage installation of
unnatural side chains. This technique should be readily adaptable
to solid-phase synthesis, be extended to split and pool protocols
using molecular fragments of varying diversity, and inspire non-
amide-bond-forming approaches to fragment integration.
2898. (c) Rademacher, P.; Wurthwein, E.-U. J. Mol. Struct. 1986, 139,
̈
315. (d) Falkenstein, R.; Mall, T.; Speth, D.; Stamm, H. J. Org. Chem.
1993, 58, 7377.
(10) (a) Galonic,
Soc. 2004, 126, 12712. (b) Galonic,
A.; Gin, D. Y. J. Am. Chem. Soc. 2005, 127, 7359.
(11) Lygo, B. Tetrahedron 1995, 51, 12859.
́
D. P.; van der Donk, W. A.; Gin, D. Y. J. Am. Chem.
́
D. P.; Ide, N. D.; van der Donk, W.
(12) Lewinski, M. K.; Bushman, F. D. Adv. Genet. 2005, 55, 147.
(13) (a) Blackburn, G. M.; Skaife, C. J.; Kay, I. T. J. Chem. Res.,
Miniprint 1980, 3650. (b) Somayaji, V.; Brown, R. S. J. Org. Chem. 1986,
51, 2676.
́
(14) (a) Slebacka-Tilk, H.; Brown, R. S. J. Org. Chem. 1987, 52, 805.
́
(b) Bennet, A. J.; Wang, Q. P.; Slebacka-Tilk, H.; Somayaji, V.; Brown,
R. S.; Santarsiero, B. D. J. Am. Chem. Soc. 1990, 112, 6383.
(15) Korn, A.; Rudolph-Bohner, S.; Moroder, L. Tetrahedron 1994, 50,
̈
1717.
(16) In peptide chemistry, the attack of the N-terminal amine at the (i +
1) peptide bond is a commonly observed problem that leads to the
formation of unwanted diketopiperazines. We do not observe this
undesired side process in our system.
(17) For a recent example showcasing amide couplings with
carboxylate salts, see: Goodreid, J. D.; Duspara, P. A.; Bosch, C.;
Batey, R. A. J. Org. Chem. 2014, 79, 943.
(18) The yields of the desired products 4 and thus 4′ increased upon
the addition of water to the oligomeric mixture. We performed a control
experiment with H-Sar-OtBu to quantify this effect. See the Supporting
Information for details.
ASSOCIATED CONTENT
* Supporting Information
■
S
Experimental details, optimization data, characterization data,
HPLC traces, and 1H and 13C NMR spectra of all new
compounds. This material is available free of charge via the
Internet at http://pubs.acs.org.
(19) Malesevic, M.; Strijowski, U.; Bachle, D.; Sewald, N. J. Biotechnol.
̈
2004, 112, 73.
(20) (a) Carpino, L. A. J. Am. Chem. Soc. 1993, 115, 4397. (b) Ehrlich,
A.; Heyne, H.-U.; Winter, R.; Beyermann, M.; Haber, H.; Carpino, L. A.;
Bienert, M. J. Org. Chem. 1996, 61, 8831. (c) Cardillo, G.; Gentilucci, L.;
Tolomelli, A.; Tomasini, C. Tetrahedron Lett. 1997, 38, 6953.
(21) (a) Legters, J.; Willems, J. G. H.; Thijs, L.; Zwanenburg, B. Recl.
Trav. Chim. Pays-Bas 1992, 111, 59. (b) Bucciarelli, M.; Forni, A.;
Moretti, I.; Prati, F.; Torre, G. Tetrahedron: Asymmetry 1995, 6, 2073.
(22) (a) Kolb, H. C.; Finn, M. G.; Sharpless, K. B. Angew. Chem., Int.
Ed. 2001, 40, 2004. (b) Tornøe, C. W.; Christensen, C.; Meldal, M. J. J.
Org. Chem. 2002, 67, 3057.
AUTHOR INFORMATION
Corresponding Author
ayudin@chem.utoronto.ca
■
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
We thank NSERC and CIHR for financial support.
■
(23) Rosenker, C. J.; Krenske, E. H.; Houk, K. N.; Wipf, P. Org. Lett.
2013, 15, 1076.
(24) Previous examples incorporating nonpeptidic fragments within
larger peptide macrocycles include: (a) Smith, J. M.; Vitali, F.; Archer, S.
A.; Fasan, R. Angew. Chem., Int. Ed. 2011, 50, 5075. (b) Chen, S.;
Morales-Sanfrutos, J.; Angelini, A.; Cutting, B.; Heinis, C. Chem-
BioChem 2012, 13, 1032. (c) Wu, X.; Wang, L.; Han, Y.; Regan, N.; Li,
P.-K.; Villalona, M. A.; Hu, X.; Briesewitz, R.; Pei, D. ACS Comb. Sci.
2011, 13, 486.
(25) We have also shown that thiol nucleophiles are amenable to our
methodology. PhSH was shown to react with 5o to yield two
regioisomeric products. See the Supporting Information for details.
REFERENCES
■
(1) (a) Driggers, E. M.; Hale, S. P.; Lee, J.; Terrett, N. K. Nat. Rev. Drug
Discovery 2008, 7, 608. (b) Marsault, E.; Peterson, M. L. J. Med. Chem.
2011, 54, 1961.
(2) Glenn, M. P.; Kelso, M. J.; Tyndall, J. D. A.; Fairlie, D. P. J. Am.
Chem. Soc. 2003, 125, 640.
(3) (a) Rezai, T.; Yu, B.; Millhauser, G. L.; Jacobson, M. P.; Lokey, R. S.
J. Am. Chem. Soc. 2006, 128, 2510. (b) Fernan
2004, 22, 1081.
́
dez, A. Nat. Biotechnol.
(4) Tyndall, J. D. A.; Nall, T.; Fairlie, D. P. Chem. Rev. 2005, 105, 973.
(5) (a) Robinson, J. A. ChemBioChem 2009, 10, 971. (b) Burgess, K.
Acc. Chem. Res. 2001, 34, 826. (c) Whitby, L. R.; Boger, D. L. Acc. Chem.
Res. 2012, 45, 1698. (d) Bullock, B. N.; Jochim, A. L.; Arora, P. S. J. Am.
Chem. Soc. 2011, 133, 14220. (e) Nowisk, J. S. Acc. Chem. Res. 2008, 41,
1319. (f) Robinson, J. A. Acc. Chem. Res. 2008, 41, 1278. (g) Angelini, A.;
Cendron, L.; Chen, S.; Touati, J.; Winter, G.; Zanotti, G.; Heinis, C. ACS
Chem. Biol. 2012, 7, 817.
(26) Ramstrom, O.; Lehn, J.-M. Nat. Rev. Drug Discovery 2002, 1, 26.
̈
(6) White, C. J.; Yudin, A. K. Nature Chem. 2011, 3, 509.
(7) (a) Seebach, D.; Beck, A. K.; Bossler, H. G.; Gerber, C.; Ko, S. Y.;
Murtiashaw, W.; Naef, R.; Shoda, S.-I.; Thaler, A.; Krieger, M.; Wenger,
R. Helv. Chim. Acta 1993, 76, 1564. (b) White, T. R.; Renzelman, C. M.;
Rand, A. C.; Rezai, T.; McEwen, C. M.; Gelev, V. M.; Turner, R. A.;
Linington, R. G.; Leung, S. S. F.; Kalgutkar, A. S.; Bauman, J. N.; Zhang,
Y.; Liras, S.; Price, D. A.; Mathiowetz, A. M.; Jacobson, M. P.; Lokey, R.
S. Nat. Chem. Biol. 2011, 7, 810.
(8) Hili, R.; Rai, V.; Yudin, A. K. J. Am. Chem. Soc. 2010, 132, 2889.
3731
dx.doi.org/10.1021/ja412256f | J. Am. Chem. Soc. 2014, 136, 3728−3731