A bioorthogonal ligation of cyclopropenones mediated by triarylphosphines

Article abstract of DOI:10.1021/jacs.5b06969

Bioorthogonal chemistries have been widely used to probe biopolymers in living systems. To date, though, only a handful of broadly useful transformations have been identified because of the stringent requirements placed on the reactants. Here we report a novel bioorthogonal ligation between cyclopropenones and functionalized phosphines. These components are stable in physiological buffers and react rapidly with one another to form covalent adducts. The cyclopropenone ligation is also distinct from other bioorthogonal chemistries in that it makes use of readily accessible, commercially available reagents and proceeds via a nucleophilic reaction pathway. On the basis of these features, the cyclopropenone ligation is poised to join the ranks of chemistries with utility in living systems.

Full text of DOI:10.1021/jacs.5b06969

Communication
pubs.acs.org/JACS
A Bioorthogonal Ligation of Cyclopropenones Mediated by
Triarylphosphines
Hui-Wen Shih† and Jennifer A. Prescher*,†,‡,§
‡
§
†Departments of Chemistry, Molecular Biology & Biochemistry, and Pharmaceutical Sciences, University of California, Irvine,
California 92697, United States
S
* Supporting Information
cyclopropenones. These scaffolds are attractive for biomolecule
ABSTRACT: Bioorthogonal chemistries have been
widely used to probe biopolymers in living systems. To
date, though, only a handful of broadly useful trans-
formations have been identified because of the stringent
requirements placed on the reactants. Here we report a
novel bioorthogonal ligation between cyclopropenones
and functionalized phosphines. These components are
stable in physiological buffers and react rapidly with one
another to form covalent adducts. The cyclopropenone
ligation is also distinct from other bioorthogonal
chemistries in that it makes use of readily accessible,
commercially available reagents and proceeds via a
nucleophilic reaction pathway. On the basis of these
features, the cyclopropenone ligation is poised to join the
ranks of chemistries with utility in living systems.
tagging because of their small size and biocompatibility.
Cyclopropenones are also present in natural products10−12
and synthetic drugs,12−17 suggesting that they possess some
degree of metabolic stability. Cyclopropenones are not
normally found in mammals but are tolerated by such
organisms, making them well-suited for bioorthogonal
applications in higher eukaryotes. Indeed, diarylcycloprope-
nones have been employed as photolabile masking groups in
live cells.18,19
While cyclopropenones are stable in physiologically relevant
environments, they react with soft nucleophiles (e.g.,
triarylphosphines) to produce activated electrophiles; these
intermediates can be easily trapped to form stable adducts
(Figure 1).20 Like cyclopropenones, phosphines are extremely
rare in mammalian systems and are also remarkably
bioinert.21,22 In fact, triarylphosphines are components of the
Staudinger ligationone of the first bioorthogonal reactions on
record.23,24 This reaction similarly exploits phosphine reactivity
to generate activated intermediates that are subject to rapid
covalent trapping. We aimed to capitalize on these features by
designing a polar, bioorthogonal transformation with electro-
philic cyclopropenones and nucleophilic phosphines.
complete understanding of organismal biology requires
Atools and technologies to examine biomolecules in their
native habitats. Among the most powerful methods for probing
biomolecules in cells and tissues is the bioorthogonal chemical
reporter strategy. This approach relies on the installation of
unique functional groups (reporters) into target biomolecules.
The reporters can be detected in a second step using selective
(bioorthogonal) reactions. The reporter and detection reagent
must both be nontoxic and unreactive toward endogenous
biological functionality, yet react rapidly with one another in
physiological environments. Popular reporters include organic
azides and terminal alkynes, although several other candidates
have been reported in recent years.1
Figure 1. Phosphine-mediated cyclopropenone ligation. The triar-
ylphosphine activates the cyclopropenone for attack by a pendant or
exogenously added nucleophile.
While the bioorthogonal toolbox is expanding at a rapid pace,
many bioorthogonal reagents remain too large or unstable for
tagging biomolecules in live cells. Moreover, the majority of
common bioorthogonal reagents are incompatible with one
another. This has stymied our ability to “see” multiple
biomolecules in real time and in living systems. To address
these limitations, new bioorthogonal reactions and combina-
tions of such reactions are being developed. Our group has
been exploring bioorthogonal transformations based on
cyclopropenes.2,3 These small, strained rings are broadly
compatible with a variety of cellular functionality. Indeed,
cyclopropenes have been installed in numerous biomolecules
and detected via cycloaddition reactions with complementary
probes.3−9
Key to the success of the proposed reaction would be the
generation and trapping of activated ketene ylide 1 (Figure 2).
This intermediate would form upon initial Michael-type
addition of triarylphosphine 2 to cyclopropenone 3 and
subsequent ring opening. Ketene ylides can be long-lived and
even isolated under ambient, moisture-free conditions.20
However, these moieties are susceptible to rapid reaction
with both oxygen and amine nucleophiles.20,25−28 We
hypothesized that intermediate 1 could be formed under
aqueous conditions and competitively trapped with strong
nucleophiles (e.g., amines) even in the presence of water. The
selectivity could be further enhanced by tethering the desired
To further expand the bioorthogonal toolkit, we recently
Received: December 18, 2014
turned our attention to a related class of microcycles:
© XXXX American Chemical Society
A
DOI: 10.1021/jacs.5b06969
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Communication
To evaluate the proposed transformation, we initially turned
to commercially available cyclopropenone, phosphine, and
amine reagents. Diphenylcyclopropenone (3a) was dissolved in
benzene and incubated with triphenylphosphine and isopropyl-
amine (Table 1, entry 1). The desired α,β-unsaturated amide
was isolated in excellent yield with no observed decomposition
or Baylis−Hillman intermediates. The reaction also proceeded
efficiently when the ketene ylide was preformed and stored for
24 h prior to the addition of isopropylamine (Figures S11−
S13).26 Consistent with the proposed mechanism, lower
catalyst loading slowed the reaction but did not significantly
impact the overall yield (entry 2).
Figure 2. Proposed catalytic cycle for the tricomponent ligation.
Encouraged by these results, we began to optimize the
reagents for bioorthogonal application. We first synthesized two
model alkyl-substituted cyclopropenones (3b and 3c) that are
smaller in size than 3a and thus more attractive for biomolecule
labeling. These cyclopropenones were synthesized via alkyla-
tion of the corresponding acetals.16,30 Both 3b and 3c reacted
with amines in the presence of triphenylphosphine (Table 1,
entries 3−5), with the less substituted scaffold (3c) providing
the fastest conversions. Very good yields of the desired amides
(79−83%) were observed in all cases. Importantly, 3b and 3c
were stable in physiological buffers and toward cellular
concentrations of cysteine (Figures S2−S5).31,32 The more
water-soluble cyclopropenone 3d was found to be similarly
robust (Figures S6 and S7).
Table 1. Scope of the Cyclopropenone Ligation
The ligation reaction was noticeably faster in aqueous
solution. When cyclopropenone 3d and water-soluble triar-
ylphosphine 2a were combined in the presence of primary or
secondary amines, the reaction was complete in 30 min, even
when catalytic amounts of phosphine were used (Table 1,
entries 6 and 7). Good yields of the ligation adducts were
observed in all cases, with amine trapping preferred over ketene
hydrolysis even under aqueous conditions.33 In the absence of
phosphine, direct addition of the amine to the cyclopropenone
was not observed (Figures S8 and S9).34 Furthermore, the α,β-
unsaturated ligation products were stable in physiological
buffers and showed no additional modification by nucleophiles
such as cysteine (Figure S10). Collectively, these results
suggested that cyclopropenones could be efficiently ligated in
biological settings using amine probes and triarylphosphine
reagents.
a
b
1.5 equiv of amine was used. The product was isolated as a 1.3:1
c
d
mixture of regioisomers. Containing 20% MeCN. 10 equiv of amine
was used.
nucleophile to the phosphine core or using an excess of an
exogenously delivered probe. The latter case also offered a
unique platform for bioorthogonal organocatalysis: following
the trapping of ketene ylide 1 to yield hemiaminal 4, an
elimination−proton transfer sequence would provide the
ligated product with concomitant regeneration of the
phosphine.29
To evaluate the ligation in a more relevant context, we
appended cyclopropenone scaffolds to a model protein
(lysozyme) using standard N-hydroxysuccinimide (NHS)
Figure 3. Cyclopropenones can be ligated on model proteins. (A) Cyclopropenone NHS ester 3e was appended to the model protein lysozyme. The
modified protein (CPO) was subsequently reacted with amine−rhodamine conjugate 5 in the presence of water-soluble triarylphosphine catalyst 2a.
(B) Gel analysis of CPO incubated with equimolar concentrations of 2a (25−200 μM) and 5 (25−200 μM) or no reagent (−) for 20 min. (C) Gel
analysis of CPO incubated with 2a (100 μM) and 5 (100 μM) or no reagent (−) for 1−45 min.
B
DOI: 10.1021/jacs.5b06969
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Communication
ester coupling conditions (Figures 3A and S14).35 The
modified protein sample (CPO) was then reacted with
triarylphosphine 2a in the presence of a widely available
rhodamine probe (5). The resulting adducts were analyzed via
gel electrophoresis and in-gel fluorescence imaging (Figures 3
and S15) along with mass spectrometry (Figure S17). As
depicted in Figure 3B−C, the ligation was both dose- and time-
dependent, and no reaction was observed in the absence of
either the cyclopropenone scaffold or triarylphosphine. When
100 μM concentrations of phosphine and amine were used, the
reaction was complete within 45 min. Sparing amounts of
phosphine were also sufficient to ligate CPO, with excess amine
providing more efficient reactions in all cases (Figure S15).
Lower concentrations of amine resulted in diminished signal
intensity, likely as a result of ketene hydrolysis (Figures 3B,
S15, and S17). Thus, the cyclopropenone ligation is best
performed with excess exogenous amine.
While suitable for some applications, the tricomponent
cyclopropenone ligation (Figure 3) would be difficult to
transition to biomolecule tagging in live cells, as three reactants
must converge at a single spot. To access a more tractable
bimolecular reaction for use in living systems, we envisioned
combining the nucleophile and phosphine into a single reagent
(P,X-probe 6; Figure 4A). Such a probe would rapidly trap
ketene 1 in an intramolecular fashion, mitigating against
undesired reactions with water or endogenous biological
nucleophiles. We synthesized model P,N-probe 6a bearing a
pendant aniline nucleophile. When 6a was incubated with
cyclopropenone 3c, amide 7 was isolated as the major product
(Figure 4B).
Given the efficiency of ketene trapping observed with 6a, we
anticipated that the bimolecular cyclopropenone ligation would
be suitable for use in more complex environments. Indeed,
when the model protein CPO was treated with 6a, ligated
adducts were readily detected, with no hydrolysis or undesired
ketene addition products observed (Figure 4C). The ligation
also proceeded readily in biological media and common buffers
(Figures 4C and S18). As a further check on the selectivity, we
assayed the ligation in a more complex environment. CPO was
incubated in lysate from HEK293 cells and then treated with
rhodamine-conjugated P,N-probe 6b. As shown in Figures 4D
and S16, ligated adducts were readily detected, with no
diminishment in signal due to competitive nucleophile addition.
These results indicated that ketene trapping by the tethered
amine is rapid and further suggested that weaker nucleophiles
(e.g., phenols) could potentially afford efficient bimolecular
ligation. Indeed, when P,O-probe 6c was combined with
cyclopropenone 3c, ester formation was observed (20 min, 69%
yield). The P,O-probe was also capable of ligating the model
protein CPO with no competitive hydrolysis (Figure S19).
These data suggest that the cyclopropenone ligation can be
used to access a range of biomolecule linkages.
In conclusion, a novel bioorthogonal ligation is reported.
This transformation exploits the electrophilic nature of
cyclopropenone reagents and their susceptibility to attack by
nucleophilic triarylphosphines. The reaction produces activated
ketene intermediates that can be efficiently trapped with
exogenous amine nucleophiles to form stable amide adducts.
This tricomponent ligation is advantageous in terms of reagent
accessibility (similar to the copper-catalyzed “click” reaction). A
plethora of functionalized amine fluorophores and affinity tags
are commercially available, making this a versatile and
immediately useful reaction. Moreover, the phosphine and
Figure 4. (A) Bimolecular P,X-probes react via intramolecular ketene
trapping. (B) P,N-probes react efficiently with cyclopropenones. (C)
Cyclopropenone-modified lysozyme CPO reacts selectively with P,N-
probe 6a. ESI-MS shows 1−4 ligations and no intermolecular ketene
addition products or cyclopropenone after 20 min. (D) Gel analysis of
CPO (2−0.25 μg) incubated with P,N-rhodamine probe 6b (250 μM)
and HEK293 lysate (20 μg) or no lysate (−) for 45 min. *denotes
CPO dimer.
trapping nucleophile can be merged into a single scaffold. Such
probes offer added utility in complex settings.
The cyclopropenone−phosphine ligation is also distinct from
other bioorthogonal ligations. First, this ligation is one of the
few bioorthogonal reactions described to date that proceed via
a nucleophilic reaction manifold. Most bioorthogonal reactions,
and nearly all of the ones reported in the last 15 years, comprise
cycloadditions. The most notable exception is the Staudinger
ligation, one of the first bioorthogonal reactions on record and
still one of the most selective.21,23,24 Second, the mechanism of
the cyclopropenone ligation suggests that the reaction will be
orthogonal to most bioorthogonal cycloadditions and thus
useful for multicomponent labeling. As mentioned earlier, many
popular bioorthogonal motifs are incompatible with one
another and cannot be used concurrently to probe multiple
C
DOI: 10.1021/jacs.5b06969
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Communication
(18) Arumugam, S.; Popik, V. V. J. Org. Chem. 2014, 79, 2702.
(19) Poloukhtine, A. A.; Mbua, N. E.; Wolfert, M. A.; Boons, G.-J.;
Popik, V. V. J. Am. Chem. Soc. 2009, 131, 15769.
(20) Hamada, A.; Takizawa, T. Tetrahedron Lett. 1972, 13, 1849.
(21) Saxon, E.; Bertozzi, C. R. Science 2000, 287, 2007.
(22) Komatsu, K.; Kitagawa, T. Chem. Rev. 2003, 103, 1371.
(23) Nilsson, B. L.; Kiessling, L. L.; Raines, R. T. Org. Lett. 2000, 2,
1939.
biomolecules or cellular processes in vivo. Third, the α,β-
unsaturated products afforded by the ligation can potentially be
further exploited for biomolecule tagging. Last, the cyclo-
propenone scaffold is small and highly tunable. Collectively,
these features will enable new biological applications and
inspire new reaction development.
ASSOCIATED CONTENT
* Supporting Information
The Supporting Information is available free of charge on the
ACS Publications website at DOI: 10.1021/jacs.5b06969.
(24) Saxon, E.; Armstrong, J. I.; Bertozzi, C. R. Org. Lett. 2000, 2,
2141.
(25) Kascheres, A.; Joussef, A. C.; Duarte, H. C. Tetrahedron Lett.
■
S
1983, 24, 1837.
(26) Cunha, S.; Kascheres, A. J. Braz. Chem. Soc. 2002, 13, 687.
Experimental details, full spectroscopic data for all new
compounds, and additional images (PDF)
́
́
(27) Alajarín, M.; Lop
Lorente, P.; Sanchez-Andrada, P.; Berna,
Jones, P. G. Chem. - Eur. J. 2010, 16, 3728.
́
ez-Leonardo, C.; Alvarez-García, A.; Llamas-
́
́
J.; Pastor, A.; Bautista, D.;
(28) Hamada, A.; Takizawa, T. Chem. Pharm. Bull. 1975, 23, 2933.
(29) It is postulated that the elimination−proton transfer is
concerted; see ref 26.
(30) Nakamura, M.; Wang, X. Q.; Isaka, M.; Yamago, S.; Nakamura,
E. Org. Synth. 2003, 80, 144.
AUTHOR INFORMATION
Corresponding Author
*jpresche@uci.edu
Notes
■
(31) Cyclopropenones 3c and 3d are stable toward cysteine around
pH 6.0, but some degradation is observed with increasing pH.
(32) Poloukhtine, A.; Popik, V. V. J. Org. Chem. 2003, 68, 7833.
(33) Even when 1.5 equiv of isopropylamine was used, the desired
amide product could be isolated in 66% yield.
(34) With 3d, ∼75% of the cyclopropenone remained after 24 h of
incubation.
(35) Hermanson, G. T. Bioconjugate Techniques; Academic Press: San
Diego, 1996.
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
This work was supported by the UC Irvine School of Physical
Sciences and the Alfred P. Sloan Foundation. We thank the
Dong, Overman, Guan, Jarvo, Nowick, Vanderwal, and
Chamberlin laboratories for providing reagents and equipment.
We are grateful to Beniam Berhane and the UCI Mass
Spectrometry Facility. Finally, we acknowledge members of the
Prescher lab for manuscript edits and helpful discussions.
REFERENCES
■
(1) Patterson, D. M.; Nazarova, L. A.; Prescher, J. A. ACS Chem. Biol.
2014, 9, 592.
(2) Kamber, D. N.; Nazarova, L. A.; Liang, Y.; Lopez, S. A.;
Patterson, D. M.; Shih, H.-W.; Houk, K. N.; Prescher, J. A. J. Am.
Chem. Soc. 2013, 135, 13680.
(3) Patterson, D. M.; Nazarova, L. A.; Xie, B.; Kamber, D. N.;
Prescher, J. A. J. Am. Chem. Soc. 2012, 134, 18638.
(4) Patterson, D. M.; Jones, K. A.; Prescher, J. A. Mol. BioSyst. 2014,
10, 1693.
̌
(5) Seck
e148.
̌
ute, J.; Yang, J.; Devaraj, N. K. Nucleic Acids Res. 2013, 41,
̇
(6) Brea, R. J.; Cole, C. M.; Devaraj, N. K. Angew. Chem., Int. Ed.
2014, 53, 14102.
(7) Elliott, T. S.; Townsley, F. M.; Bianco, A.; Ernst, R. J.; Sachdeva,
A.; Elsasser, S. J.; Davis, L.; Lang, K.; Pisa, R.; Greiss, S.; Lilley, K. S.;
̈
Chin, J. W. Nat. Biotechnol. 2014, 32, 465.
(8) Yu, Z.; Lin, Q. J. Am. Chem. Soc. 2014, 136, 4153.
(9) Yu, Z.; Pan, Y.; Wang, Z.; Wang, J.; Lin, Q. Angew. Chem., Int. Ed.
2012, 51, 10600.
(10) Kogen, H.; Kiho, T.; Tago, K.; Miyamoto, S.; Fujioka, T.;
Otsuka, N.; Suzuki-Konagai, K.; Ogita, T. J. Am. Chem. Soc. 2000, 122,
1842.
(11) Bohlmann, F.; Jakupovic, J.; Muller, L.; Schuster, A. Angew.
̈
Chem., Int. Ed. Engl. 1981, 20, 292.
(12) Okuda, T.; Yokose, K.; Furumai, T.; Maruyama, H. B. J. Antibiot.
1984, 37, 718.
(13) Ando, R.; Sakaki, T.; Morinaka, Y.; Takahashi, C.; Tamao, Y.;
Yoshii, N.; Katayama, S.; Saito, K.; Tokuyama, H.; Isaka, M.;
Nakamura, E. Bioorg. Med. Chem. 1999, 7, 571.
(14) Okuda, T.; Shimma, N.; Furumai, T. J. Antibiot. 1984, 37, 723.
(15) Sotiriadis, D.; Patsatsi, A.; Lazaridou, E.; Kastanis, A.; Vakirlis,
E.; Chrysomallis, F. Clin. Exp. Dermatol. 2007, 32, 48.
(16) Tokuyama, H.; Isaka, M.; Nakamura, E.; Ando, R.; Morinaka, Y.
J. Antibiot. 1992, 45, 1148.
(17) Upitis, J. A.; Krol, A. J. Cutaneous Med. Surg. 2002, 6, 214.
D
DOI: 10.1021/jacs.5b06969
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX