Controlling intracellular macrocyclization for the imaging of protease activity

Article abstract of DOI:10.1002/anie.201006140

(Figure Presented) Reporters on the ground: An intramolecular macrocyclization reaction took place highly efficiently in live cells under the control of a specific enzyme and reduction by glutathione (GSH; see picture). Macrocyclic products (represented as blue rings) synthesized in cells self-assembled into nanoparticles aggregated and retained at the site near the enzyme location to report local proteolytic activity in live cells.

Full text of DOI:10.1002/anie.201006140

DOI: 10.1002/anie.201006140
Macrocyclization in Cells
Controlling Intracellular Macrocyclization for the Imaging of Protease
Activity**
Deju Ye, Gaolin Liang, Man Lung Ma, and Jianghong Rao*
Macrocycles are an important class
of molecules for their profound
chemical and pharmacological prop-
erties derived from the preorganized
ring structures.[1–3] They are either
isolated from microorganisms or
prepared from acyclic precursors
through chemical macrocyclization,
such as ring-closing olefin metathe-
sis.[4–8] On the other hand, the con-
trol of macrocylization to synthesize
macrocycles directly from acyclic
precursors in living mammalian
cells has been little explored. For
example, a chemically stable, bio-
logically inactive acyclic precursor
(such as 1 in Figure 1) would enter
cells and be converted into a macro-
cycle (such as 3). Specific cellular
processes may be exploited to con-
trol the synthesis through regulation
of the conversion of the chemically
Figure 1. Proposed enzyme-controlled macrocyclization reaction in cells. A cell-permeable probe 1
enters cells and is converted by reduction and/or enzymatic processing into an intermediate 2,
which quickly undergoes intramolecular cyclization to generate the macrocyclic product 3. The self-
assembly of 3 leads to the formation of nanoparticles accumulated locally at or near the enzyme
location in cells.
stable precursor into the reactive
intermediate (2 in Figure 1) for sub-
sequent macrocyclization. Because
of the unique properties of macro-
cycles, the conversion from the acy-
clic into the cyclic form may lead to
novel functions and applications for the probing of cellular
biochemistry and biology. Herein, we describe an example of
the control of macrocyclization in living cells to image local
protease activity.
The conceptual design in Figure 1 was demonstrated with
an intramolecular macrocyclization system derived from a
biocompatible condensation reaction between 2-cyanobenzo-
thiazole (CBT) and free cysteine.[9–12] We have previously
shown that this bimolecular condensation reaction can lead to
the formation of oligomers in vitro and in cells.[12] For efficient
macrocyclization in cells, intermolecular condensation of the
acyclic precursors with intracellular endogenous free cysteine
should be minimal. We hypothesized that if we used a CBT
analogue with significantly reduced reactivity towards cys-
teine, the acyclic precursor would not react intermolecularly
with free cysteine before activation by the target enzyme in
cells, but only by an intramolecular cyclization after enzyme-
mediated conversion into the reactive intermediate 2.
To discover such CBT analogues, we screened a series of
cyano-substituted aromatic compounds to determine their
reaction rates with l-cysteine in phosphate-buffered saline
(PBS) in an HPLC assay. A 13-fold decrease in the second-
order rate constant was observed for 4-methyl-2-thiazolecar-
bonitrile, and a more than 480-fold decrease for 2-cyano-6-
hydroxyquinoline (CHQ). The other three analogues gave no
detectable condensation product on the HPLC column in
5 hours (see Table S1 in the Supporting Information). Amino-
thiol substrates were also screened for improved activity
towards CBT; however, the second-order rate constant of all
[*] Dr. D. Ye, Dr. G. Liang, M. L. Ma, Prof. J. Rao
Molecular Imaging Program at Stanford
Departments of Radiology and Chemistry
Biophysics, Cancer Biology Programs, Stanford University
1201 Welch Road, Stanford, CA 94305-5484 (USA)
Fax: (+1)650-736-7925
E-mail: jrao@stanford.edu
Homepage: http://raolab.stanford.edu
[**] This research was supported by the Stanford University National
Cancer Institute (NCI) Centers of Cancer Nanotechnology Excel-
lence (1U54A151459-01), the NCI ICMIC@Stanford (1P50A114747-
06), and an IDEA award from the Department of Defense Breast
Cancer Research Program (W81XWH-09-1-0057).
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/anie.201006140.
Angew. Chem. Int. Ed. 2011, 50, 2275 –2279
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
2275

Communications
of the tested analogues was lower than that of free cysteine
than 1%, especially for 4b, which has more than 6.4-fold
faster kinetics than 4c (see Figure S4).
(see Table S1). Therefore, we selected CHQ and cysteine as
the reacting pair for the intramolecular-condensation system.
Three precursors 4a–c containing CHQ and cysteine
moieties (Scheme 1) were synthesized to evaluate this
hypothesis. A varying number of glycine residues were
A derivative of 4b containing a disulfide-protected
cysteine residue and an l-lysine residue in place of glycine
(5 in Figure 2a) was prepared to evaluate whether disulfide
reduction could trigger intramolecular macrocyclization.
Compound 5 was stable at pH 7.4, but upon treatment with
the reducing reagent tris(2-carboxyethyl)phosphane (TCEP)
at pH 3, it was converted into the reduced form 5-I (Fig-
ure 2b). Subsequent adjustment of the pH value to 7.4
resulted in intramolecular cyclization. Interestingly, in con-
trast to the reactions of precursors 4a–c, two product peaks
were observed by HPLC analysis with distinct retention
times: 5-II-1 and 5-II-2, which share an identical molecular
weight (see Figure S5a). Heteronuclear multiple-bond corre-
lation (HMBC) NMR spectroscopy confirmed that both
compounds were macrocyclization products derived from 5
(see Figure S5b). Owing to the presence of l-lysine, they are
probably diastereoisomers that arise from two different ring-
closing orientations (see Figure S6).
Scheme 1. Intramolecular cyclization of 4a, 4b, and 4c to produce 4a-
I, 4b-I, and 4c-I, respectively.
Interestingly, UV/Vis spectroscopy and dynamic light
scattering (DLS) revealed that the macrocyclic products
derived from 5 could further assemble into nanoparticles, as
observed with cyclic oligomers in our previous study.[12] Upon
the reduction of 5 with TCEP, the spectrum showed signifi-
cantly increased broad scattering around 500–700 nm because
of the reduction-triggered intramolecular cyclization and
aggregation (see Figure S7c,e). DLS further confirmed the
formation of particles with a mean diameter of 4–5 mm upon
the reduction with TCEP (see Figure S7d,f). TEM images of
the products derived from 5 revealed that the macrocycles can
self-assemble into nanofibers with an average length of
3–5 mm and an average diameter of 10–20 nm (Figure 2c;
see also Figure S7g).
When the hydrophobic dye dansyl chloride was conju-
gated to the l-lysine side chain of 5, the new derivative 5d
underwent phase transfer from a homogeneous solution to a
suspension upon reduction with TCEP (see Figure S8). TEM
images of the cyclized products of 5d showed self-assembled
nanoparticles with diameters of 40–60 nm that could further
assemble into clusters (see Figure S8e). Fluorescein isothio-
cyanate (FITC) was similarly conjugated to form derivative
5 f, which displayed the same reduction-induced macrocycli-
zation and self-assembly of nanoparticles in a buffer and in
cell lysates as observed for 5 and 5d (see Figure S9). These
results suggest that the modification of the l-lysine side chain
has little effect on the intramolecular cyclization reaction. On
the other hand, the control compound 5c with an
S-methylated cysteine residue did not form any cyclized
products or aggregates upon reduction with TCEP in the
phosphate buffer at a concentration of 100 mm (see Fig-
ure S10).
inserted to produce cyclization products with macrocyclic
rings of different sizes. All precursors were purified by HPLC
and characterized by NMR spectroscopy and MS to confirm
their structures (see the Supporting Information).
A similar HPLC assay was employed to measure the
reaction rate of the intramolecular cyclization. Upon adjust-
ment of the pH value to 7.4 in the PBS buffer, all three
compounds afforded the expected cyclization products, which
were cleanly separated by HPLC from the precursors (see
Figure S2 in the Supporting Information). The first-order
reaction rates are (3.18 Æ 0.18) ꢀ 10À4 sÀ1 for 4a, (5.79 Æ
0.49) ꢀ 10À3 sÀ1 for 4b, and (9.04 Æ 0.22) ꢀ 10À4 sÀ1 for 4c.
This result demonstrates that the intramolecular condensa-
tion can occur in aqueous solution, and that the reaction rate
is affected by the size of the macrocyclic ring formed; 4b
displayed the fastest kinetics with a half-life (t1/2) of (119.8 Æ
10.2) s.
To further validate that the intramolecular cyclization of
CHQ is more favorable than its intermolecular condensation
with cysteine, we performed a competition assay with 4c
(5 mm) and varying concentrations of free l-cysteine (0, 5 mm,
500 mm, 5 mm, and 50 mm). The reaction solutions were mixed
in the PBS buffer and stirred overnight before analysis by
HPLC (see Figure S3). As expected, the amount of product
formed by intermolecular condensation slowly increased with
the concentration of free l-cysteine; at a concentration of
50 mm of l-cysteine, the ratio of the product of intermolecular
condensation to the product of intramolecular macrocycliza-
tion was 1:0.76. This result agrees well with the effective
molarity of about 48 mm calculated from the ratio of the rate
constant of the intermolecular condensation of CHQ and
cysteine ((0.019 Æ 0.001)mÀ1 sÀ1) to that of the intramolecular
cyclization of 4c ((9.04 Æ 0.22) ꢀ 10À4 sÀ1). Since the intra-
cellular cysteine concentration is generally around 20–
100 mm,[13,14] intermolecular condensation between acyclic
precursors or with endogenous free cysteine should be less
We then investigated whether the intramolecular cycliza-
tion could occur in cells. We first incubated 5 (200 mm) in
lysates of the human breast cancer cell line MDA-MB-468 for
2 hours. HPLC analysis indicated that the same cyclization
products formed as those in the buffer. This result suggests
that the intramolecular cyclization can take place in a
complex cellular environment containing free intracellular
2276 www.angewandte.org
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2011, 50, 2275 –2279

Figure 2. Characterization of the cyclization and nanoassembly induced by chemical or cellular reduction in vitro. a) Proposed two-step reduction-
induced cyclization of probe 5. b) HPLC traces of the cyclization of 5 induced by TCEP reduction. Bottom (black): 5 alone; middle (blue): 5-I
formed by the reduction of 5 (100 mm) by TCEP (200 mm) at pH 3 at room temperature for 20 min; top (red): two cyclized diastereoisomeric
products 5-II formed by the reduction of 5 by TCEP at pH 3 followed by adjustment of the pH value to 7.4. c) TEM image of the products of the
TCEP-induced cyclization of 5 (100 mm) in water; 36000ꢀ magnification (scale bar: 500 nm). d) HPLC analysis of the intracellular condensation
reaction of 5 in cell lysate. Bottom (black): the two cyclized products 5-II-1 and 5-II-2 formed upon the treatment of 5 (200 mm) with TCEP
(400 mm) at pH 7.4 at room temperature for 30 min; middle (red): the MDA-MB-468 cell lysate without 5; top (blue): HPLC trace observed after
the incubation of 5 (200 mm) in the MDA-MB-468 cell lysate for 2 h. e) HPLC analysis of the intracellular cyclization reaction of 5 (100 mm) in live
cells. Bottom (black): culture medium alone (Dulbecco modified Eagle medium, DMEM); middle (red): the collected medium after the incubation
of 5 for 8 h in live MDA-MB-468 cells; top (blue): the lysate of MDA-MB-468 cells incubated with 5 for 8 h.
cysteine. Notably, a side product of 5 with its cysteine residue
removed was observed (Figure 2d); this compound, 5-S1, was
presumably formed by proteolytic hydrolysis. Similar results
were obtained with HeLa cell lysates (see Figure S11).
cells. When a d-cysteine residue replaced the l-cysteine
residue in 5, neither side product was observed by HPLC (see
Figure S12).
The control compound 5c, on the other hand, did not
produce any cyclization products after similar incubation with
the MDA-MB-468 cell lysates for 2 hours at 200 mm, although
the hydrolysis side products were also observed (see Fig-
ure S10b). These results together demonstrate that intra-
cellular reduction by glutathione can induce the intramolec-
ular macrocyclization of CHQ–Cys probes.
To apply this novel chemistry to image protease activity,
we designed a probe 6, which is a conjugate of 5 f with the
peptide substrate of the trans-Golgi protease furin (RVRR;
Figure 3a).[15] Furin is a convertase that plays crucial roles in
development, homeostasis, and diseases ranging from anthrax
and Ebola fever to Alzheimerꢁs disease and cancer.[16–19] As
We next incubated 5 with live MDA-MB-468 cells for
8 hours and analyzed the cell lysates and culture medium by
HPLC and MS spectrometry (Figure 2e). Both intramolecu-
lar cyclization products, 5-II-1 and 5-II-2, were detected in the
cell lysates but not in the culture medium, which indicates that
the intramolecular cyclization can proceed in live cells, and
that the formed macrocycles are retained inside cells. The side
product observed in the incubation experiment in cell lysates
was also found in the cell medium. An additional side
product, 5-S2, present in the cell medium resulted from the
further hydrolysis of 5-S1 and removal of its lysine residue
(see Figure S11). Both side products appeared to leak out of
Angew. Chem. Int. Ed. 2011, 50, 2275 –2279
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
2277

Communications
expected, the treatment of 6
(5 mm)
with
furin
(100 pmolUÀ1) in vitro gen-
erated the two cyclization
products 6-II-1 and 6-II-2 as
diastereoisomers
(Fig-
ure 3b). Only the reduced
form of 6, not 6-I, was
observed by HPLC, which
suggests that furin cleavage
is the rate-limiting step of
the multiple-step process.
Similarly, overnight incuba-
tion of 6 with MDA-MB-
468 cell lysates or live cells
produced the macrocycles
6-II, as revealed by the
HPLC assay (Figure 3c).
No cyclization products
were detected by HPLC in
the medium used for the
incubation of live cells,
which confirms that the
formed macrocycles are
trapped inside cells. This
result is consistent with the
observed nanoparticle for-
mation after the macrocyc-
lization of 5 f (see Fig-
ure S9).
MDA-MB-468
cells
were then incubated with 6
(1 mm) for 8 h, followed by
costaining with
a Golgi
marker (fluorescent BODI-
PY TR C5-ceramide–BSA
complexes;
boron
BODIPY=
dipyrromethene,
BSA = bovine serum albu-
min) and imaging under a
fluorescence microscope. A
good overlap of green (from
the assembled macrocycles)
and red fluorescence (from
the Golgi marker) was
observed. This result sup-
ports the hypothesis that the
intramolecular cyclization
occurred at or near Golgi
bodies in cells (Figure 3d).
When we used the control
probe 6c (see Figure S13),
which contains an S-methy-
lated cysteine residue and
thus cannot undergo macro-
cyclization after furin cleav-
age, or when we used a furin
inhibitor (decanoyl-RVRR-
cmk; cmk = chloromethyl
Figure 3. Furin-triggered cyclization in vitro and in live cells. a) Proposed furin-triggered cyclization of probe 6
in cells. Glutathione reduction and furin hydrolysis generate 6-I. The intracellular cyclization of 6-I gives 6-II,
which subsequently self-assembles into nanoparticles. b) HPLC traces of 6 (5 mm) in water (bottom, black),
after incubation with glutathione (100 mm) for 2 h at 308C (middle, red), and after incubation with furin
(100 pmolUÀ1) for 8 h at 308C in furin buffer (pH 7.5; top, blue). Compound 6-r is the disulfide reduced form
of 6. c) HPLC traces of the DMEM medium (green), of the MDA-MB-468 cell lysate after incubation with 6
(20 mm) along with 1 mm CaCl2 at 308C overnight (black), of the incubation medium after the incubation of
MDA-MB-468 live cells with 6 (20 mm) overnight (blue), and of the cell lysate after the removal of the
incubation medium following the incubation of MDA-MB-468 live cells with 6 (20 mm) overnight (red).
d) Fluorescence imaging of furin-triggered localized macrocyclization in live cells. The differential interference
contrast (DIC) images are overlaid with the fluorescence images; the three images on the right are an overlay
of the green and red fluorescence images on top of the DIC images. MDA-MB-468 cells were incubated with
probe 6 (1 mm, top), the control probe 6c (1 mm, middle), or 6 with the furin inhibitor (100 mm, bottom) for
8 h, followed by costaining with a Golgi marker (BODIPY TR C5-ceramide–BSA complexes). Scale bars: 10 mm.
2278
www.angewandte.org
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2011, 50, 2275 –2279

ketone)[20] together with 6, we observed no concentrated
fluorescence signals but instead weak and uniform green
fluorescence, which confirmed that furin was responsible for
the localized accumulation of 6 near and at the Golgi bodies
(Figure 3d; see also Figure S14). Similar imaging results were
observed with 6 in other cell lines (see Figure S15). These
results demonstrate that this intramolecular macrocyclization
reaction can be applied to image local furin activity in living
cells.
In comparison to previously reported CBT and cysteine
bimolecular condensation systems, an important advantage of
this system is that it is not subject to endogenous cysteine
competition owing to the low reactivity of the CHQ moiety
toward cysteine. Furthermore, the condensation takes place
intramolecularly, and thus is not concentration-dependent,
and is kinetically fast. The enzyme cleavage is therefore the
rate-limiting step in the formation of the aggregated macro-
cycles, which results in a better correlation of the aggregation
sites with the location of enzyme activity.
In summary, we have described a bioorthogonal intra-
molecular macrocyclization reaction that can occur highly
efficiently in live cells. We have shown that macrocyclic
products can be synthesized in cells under the control of a
proteolytic enzyme and glutathione reduction to image local
proteolytic activity. Other applications may be possible, such
as the in situ synthesis of macrocyclic molecules to probe
cellular function through receptor binding.[21]
[2] J. Taunton, J. L. Collins, S. L. Schreiber, J. Am. Chem. Soc. 1996,
118, 10412 – 10422.
[3] D. T. Bong, T. D. Clark, J. R. Granja, M. R. Ghadiri, Angew.
Chem. 2001, 113, 1016 – 1041; Angew. Chem. Int. Ed. 2001, 40,
988 – 1011.
[4] L. A. Wessjohann, E. Ruijter, D. Garcia-Rivera, W. Brandt, Mol.
Diversity 2005, 9, 171 – 186.
[5] L. A. Wessjohann, E. Ruijter, Top. Curr. Chem. 2005, 243, 137 –
184.
[6] A. Gradillas, J. Pꢂrez-Castells, Angew. Chem. 2006, 118, 6232 –
6247; Angew. Chem. Int. Ed. 2006, 45, 6086 – 6101.
[7] O. David, W. J. N. Meester, H. Bierꢃugel, H. E. Schoemaker, H.
Hiemstra, J. H. van Maarseveen, Angew. Chem. 2003, 115, 4509 –
4511; Angew. Chem. Int. Ed. 2003, 42, 4373 – 4375.
[8] S. Muthana, H. Yu, H. Cao, J. Cheng, X. Chen, J. Org. Chem.
2009, 74, 2928 – 2936.
[9] E. H. White, F. McCapra, G. F. Field, W. D. McElroy, J. Am.
Chem. Soc. 1961, 83, 2402 – 2403.
[10] K. Okada, H. Iio, I. Kubota, T. Goto, Tetrahedron Lett. 1974, 32,
2771 – 2774.
[11] H. Ren, F. Xiao, K. Zhan, Y. P. Young, H. Xie, Z. Xia, J. Rao,
Angew. Chem. 2009, 121, 9838 – 9842; Angew. Chem. Int. Ed.
2009, 48, 9658 – 9662.
[12] G. Liang, H. Ren, J. Rao, Nat. Chem. 2010, 2, 54 – 60.
[13] S. Park, J. A. Imlay, J. Bacteriol. 2003, 185, 1942 – 1950.
[14] M. H. Stipanuk, J. E. Dominy, J.-I. Lee, R. M. Coloso, J. Nutr.
2006, 136, 1652S – 1659S.
[15] M. Hosaka, M. Nagahama, W. S. Kim, T. Watanabe, K.
Hatsuzawa, J. Ikemizu, K. Murakami, K. Nakayama, J. Biol.
Chem. 1991, 266, 12127 – 12130.
[16] G. Thomas, Nat. Rev. Mol. Cell Biol. 2002, 3, 753 – 766.
[17] J. Shapiro, N. Sciaky, J. Lee, H. Bosshart, R. H. Angeletti, J. S.
Bonifacino, J. Histochem. Cytochem. 1997, 45, 3 – 12.
[18] R. V. Talanian, C. Quinlan, S. Trautz, M. C. Hackett, J. A.
Mankovich, D. Banach, T. Ghayur, K. D. Brady, W. W. Wong, J.
Biol. Chem. 1997, 272, 9677 – 9682.
Received: September 30, 2010
Revised: December 27, 2010
Published online: February 14, 2011
[19] A. Dragulescu-Andrasi, G. Liang, J. Rao, Bioconjugate Chem.
2009, 20, 1660 – 1666.
[20] S. Henrich, A. Cameron, G. P. Bourenkov, R. Kiefersauer, R.
Huber, I. Lindberg, W. Bode, M. E. Than, Nat. Struct. Biol. 2003,
10, 520 – 526.
Keywords: bioorthogonal reactions · macrocycles ·
nanoparticles · protease imaging · self-assembly
.
[1] E. M. Driggers, S. P. Hale, J. Lee, N. K. Terrett, Nat. Rev. Drug
Discovery 2008, 7, 608 – 624.
[21] Z. Yang, G. Liang, B. Xu, Acc. Chem. Res. 2008, 41, 315 – 326.
Angew. Chem. Int. Ed. 2011, 50, 2275 –2279
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org 2279