Heterobifunctional linker allows for selective catalyst-free ligation of two different azide-tagged substrates via strained-promoted azide-alkyne cycloaddition (SPAAC). The linker contains an azadibenzocyclooctyne (ADIBO) moiety on one end and a cyclopropenone-masked dibenzocyclooctyne (photo-DIBO) group on the other. The first azide-derivatized substrate reacts only at the ADIBO end of the linker as the photo-DIBO moiety is azide-inert. After the completion of the first SPAAC step, photo-DIBO is activated by brief exposure to 350 nm light from a fluorescent UV lamp. The unmasked DIBO group then reacts with the second azide-tagged substrate. Both click reactions are fast (k = 0.4 and 0.07 M(-1) s(-1), respectively) and produce quantitative yield of ligation in organic solvents or aqueous solutions. The utility of the new cross-linker has been demonstrated by conjugation of azide functionalized bovine serum albumin (azido-BSA) with azido-fluorescein and by the immobilization of the latter protein on azide-derivatized silica beads. The BSA-bead linker was designed to incorporate hydrolytically labile fragment, which permits release of protein under the action of dilute acid. UV activation of the second click reaction permits spatiotemporal control of the ligation process.
Heterobifunctional linker allows for selective catalyst-free ligation of two different azide-tagged substrates via strained-promoted azide-alkyne cycloaddition (SPAAC). The linker contains an azadibenzocyclooctyne (ADIBO) moiety on one end and a cyclopropenone-masked dibenzocyclooctyne (photo-DIBO) group on the other. The first azide-derivatized substrate reacts only at the ADIBO end of the linker as the photo-DIBO moiety is azide-inert. After the completion of the first SPAAC step, photo-DIBO is activated by brief exposure to 350 nm light from a fluorescent UV lamp. The unmasked DIBO group then reacts with the second azide-tagged substrate. Both click reactions are fast (k = 0.4 and 0.07 M(-1) s(-1), respectively) and produce quantitative yield of ligation in organic solvents or aqueous solutions. The utility of the new cross-linker has been demonstrated by conjugation of azide functionalized bovineserum albumin (azido-BSA) with azido-fluorescein and by the immobilization of the latter protein on azide-derivatized silica beads. The BSA-bead linker was designed to incorporate hydrolytically labile fragment, which permits release of protein under the action of dilute acid. UV activation of the second click reaction permits spatiotemporal control of the ligation process.
Structural
modifications of biomolecules and polymers, as well
as derivatization of particles and surfaces, are commonly achieved
using ″click chemistry″.[1] This term describes a set of bimolecular reactions that permits
the efficient formation of a covalent link between two substrates
(a.k.a. ligation in biochemistry) or between a substrate and a surface.
The majority of ″click″ strategies are based on 1,3-dipolar
or Diels–Alder cycloadditions, imine formation, and addition
to carbon–carbon multiple bonds.[2−4] Since azide and alkyne
moieties are very uncommon in natural products, i.e., ″bioorthogonal″,
the copper-catalyzed (CuAAC) or strain-promoted azide–alkyne
cycloadditions (SPAAC) have found many applications in labeling biomolecules,[5] protein modification,[6] synthesis of bioconjugates,[7] and developing
biotech tools.[8] The chemical stability
of the triazole linker made azide–alkyneclick reactions a
popular tool in material chemistry.[1b,1e,9]Chemoselective sequential click ligation brings
the ability to
introduce multiple functionalities to biological molecules,[10] drug delivery vehicles,[11] polymer,[12] or surfaces.[13] Cross-linking of biological molecules using sequential
click chemistry allows for the preparation of complex macromolecules.[14] There are two different approaches to selective
sequential click ligations. One method relies on two (or more) mutually
orthogonal click reactions, e.g., CuAAC and Diels–Alder,[11a,12a−12d] SPAAC and SPANC,[15,13a,13b] SPAAC and hetero-Diels–Alder,[16] as well as other combinations.[11b,17] These strategies
provide excellent selectivity but require derivatization of substrates
with different functionalities. To avoid this complication, sequential
click ligation may employ reactions of the same type, most commonly
azide–alkyne cycloaddition. This approach relies on the differences
in the reactivity of alkyne moieties[10c,13c,14c,18] or on the deprotection
of terminal acetylenes.[10a,11c,19] The use of cytotoxic copper(I) catalyst and/or deprotecting reagents,
however, somewhat reduces the utility of this method. In addition,
terminal acetylene were found to inhibit cysteine proteases by forming
thioether with the catalytically active thiol.[20]Here, we report the development of the SPAAC-based
selective sequential
click strategy, which permits cross-linking of two different azide-tagged
substrates without the use of catalysts or activating reagents (Scheme 1). Dibenzocyclooctynes are apparently too large
to fit into the active site cavity of a protease. Selection of the
reactive moieties for the construction of the sequential SPAAC cross-linking
agent was based on the orthogonality requirement. The first click
reaction should occur only at one end of the linker, while the second
ligation should not degrade connection to the first substrate. We
have chosen azadibenzocyclooctyne (ADIBO, Scheme 1)[21] as the first SPAAC moiety,
as this cyclooctyne combines high reactivity toward azides with excellent
aqueous stability and long shelf life.[22] PhotoDIBO (Scheme 1), on the other hand,
does not react with azides in the dark and possesses excellent thermal
stability.[23] Exposure of photoDIBO moiety
to a low intensity 350 nm light results in the efficient decarbonylation
and the formation of azide-reactive dibenzocyclooctyne (DIBO).[23,24] ADIBO, DIBO, and corresponding triazole adducts have virtually no
absorbance at 350 nm, and, therefore, are stable under photodecarbonylation
conditions.[23]
Scheme 1
Sequential Click–Photoclick
Conjugation of Two Azides
The heterobifunctional ADIBO-photoDIBO cross-linker
(1) was prepared by the DCC/DMAP-assisted coupling of
ADIBO-acid (5)[25] with of pehtamethylenehydroxy-derivatized
photoDIBO (6, Scheme 2).
Scheme 2
Preparation
of ADIBO–PhotoDIBO Cross-Linker (1)
The efficiency and selectivity
of cross-linker 1 were
evaluated using its reaction with benzyl azide. The accurate rate
measurements of the first and second click reaction were conducted
by UV spectroscopy at 25 ± 0.1 °C in methanol under pseudo-first-order
conditions. The UV spectrum of ADIBO-photoDIBO (1) contains
characteristic features of both ADIBO (a band at 292 nm) and photoDIBO
(intense bands at 331 and 347 nm) chromophores (Figure 1a, black trace). All stages of the conversion of 1 can be, therefore, conveniently monitored by UV spectroscopy. The
addition of the excess of benzyl azide (0.1–20 mM) to the methanol
solution of 1 leads to the rapid disappearance of the
band at 292 nm, while photoDIBO bands remain unchanged (Figure 1a, red trace). The decay ADIBO band followed the
single exponential equation well (Figure 1b).
The dependence of the observed pseudo-first-order rate constants on
azide concentration was linear and produced second-order rate constant k= 0.406 ± 0.001 M–1 s–1. This value is consistent with literature data for ADIBO.[21,22] After the completion of the first click reaction, the cyclopropenone
protection of the triple bond in 2 was removed by 2 min
irradiation of the reaction mixture with 350 nm fluorescent lamps
(Scheme 1). The conversion of photo-DIBO was
followed by the disappearance of the characteristic cyclopropenone
bands at 331 and 347 nm and the formation of the DIBO band at 319
nm (Figure 1a). The disappearance of the latter
band due to the addition of the second molecule of benzyl azide to 3 also showed clean first-order kinetics (Figure 1b). The second-order rate of this reaction (k = 0.072 ± 0.004 M–1 s–1) is similar to the values reported for DIBO.[22,23a]
Figure 1
(a)
UV spectra of 50 μM methanol solutions of ADIBO-photoDIBO
(1, black line); product of the first click reaction 2b (red line); product of photoactivation 3b (blue
line); product of the second click reaction 4b (purple
line). (b) Kinetics traces of the first (black circles) and the second
(blue hexagons) click reactions in the presence of 10 mM of benzyl
azide.
(a)
UV spectra of 50 μM methanol solutions of ADIBO-photoDIBO
(1, black line); product of the first click reaction 2b (red line); product of photoactivation 3b (blue
line); product of the second click reaction 4b (purple
line). (b) Kinetics traces of the first (black circles) and the second
(blue hexagons) click reactions in the presence of 10 mM of benzyl
azide.HPLC analysis of the “click”
– photoactivation
– second “click” sequence indicates clean and
quantitative conversion at every step (Figure 2). Thus, 50 μM methanol solution of 1 (Figure 2a) was treated with an equimolar amount of butyl
acid and incubated for 48 h under ambient conditions to ensure complete
conversion. The HPLC trace (Figure 2b) and
ESI-HRMS of the resulting product (MH+, Calc. for C50H54N5O7 836.4018, found
836.4016) confirmed the quantitative formation of 2a
(Note: Apparently, head-to-tail and head-to-head isomers have very
similar retention times on C-18 column). Photodecarbonylation step
produces 3a (MH+, Calc. for C49H54N5O6 808.4069, found 808.4067)
with no detectable amounts of side products (Figure 2c). The addition of the second equivalent of butyl azide cleanly
gives the final adduct 4a after 48 h incubation (Figure 2d; MH+, Calc. for C53H63N8O6 907.4865, found 907.4867).
Figure 2
HPLC traces of the sequential click ligation
(Scheme 1): (a) starting linker 1; (b) reaction mixture
after the reaction of 1 with equimolar amount of butyl
azide; (c) product of photodecarbonylation step (3a);
(d) product (4a) of the second click reaction with equimolar
amount of azide.
HPLC traces of the sequential click ligation
(Scheme 1): (a) starting linker 1; (b) reaction mixture
after the reaction of 1 with equimolar amount of butyl
azide; (c) product of photodecarbonylation step (3a);
(d) product (4a) of the second click reaction with equimolar
amount of azide.As an illustrative example,
we have employed ADIBO-photoDIBO cross-linker
(1) for the selective conjugation of azido-derivatized
BSA (7) with (3-azidopropylcarbamoyl)fluorescein (azido-Fl,
8, mixture of 5- and 6-isomers, Scheme 3).
After each click reaction the functionalized protein was isolated
and characterized by MALDI. The azido-derivatized BSA (7) was prepared by treating the native BSA with 1-azido-3-iodopropane.
Protein 7 was incubated with an excess of the cross-linker 1 in aqueous solution overnight and resulting BSA derivative 2c (Scheme 3) was isolated by gel filtration.
An aqueous solution of 2c was irradiated for 2 min using
350 nm fluorescent lamps and treated with 50-fold excess of azidoFL 8 overnight and purified by gel filtration (Scheme 3). The resulting Fluorescein-derivatized BSA (4c) clearly shows characteristic absorbance and emission of
the Fluorescein chromophore (Figure 3).
Scheme 3
Conjugation of Azido-Derivatized Bovine Serum Albumin
(BSA) with
Azido-Fluorescein (FL) Using ADIBO-PhotoDIBO Cross-Linker
Figure 3
Absorption (dash–dotted line) and emission
spectra (solid
line) of 11 μM PBS solution of BSA–fluorescein conjugate
(10) and emission spectrum of a 11 μM PBS solution
of fluorescein (dotted line).
The protein content of the
solution of 4c was determined
using Coomassie brilliant blue dye assay, while the emission intensity
of 4c and the absorbance in the fluorescein region were
compared to the fluorescence and absorbance of an aqueous solution
of an authentic fluorescein sample of the same concentration (Figure 3). These experiments confirm high cross-linking
efficiency (84–92%) of the ADIBO-photoDIBO (1).
The high yield of azido-BSA (7) labeling was somewhat
surprising result since it is commonly accepted that commercial BSA
samples contain 0.6–0.8 equiv of free thiols. The rest of Cys34
moieties are believed to form a disulfide with cysteine or glutathione.[27] We believe
that free thiol content of the BSA is actually higher than reported
previously, since Ellman’s test, which uniformly used in these
measurements, often underestimates SH contents in proteins,[28] especially in proteins containing acidic Cys
residues.[29,30]Absorption (dash–dotted line) and emission
spectra (solid
line) of 11 μM PBS solution of BSA–fluorescein conjugate
(10) and emission spectrum of a 11 μM PBS solution
of fluorescein (dotted line).To test the suitability of ADIBO-photoDIBO (1) for
protein immobilization, we have employed this cross-linking agent
for the attachment of azido-BSA (7) to azide-functionalized
silica microbeads. First, a photo-Diels–Alder click reaction[31] between 8-(2-(2-(2-aminoethoxy)ethoxy)ethoxy)-3-(hydroxymethyl)naphthalen-2-ol
(9, NH2-TEG-NQMP)[17d] and (2-(2-azidoethoxy)ethoxy)ethylene produced heterobifunctional
linker 10 (Scheme 4). It is important
to note that 2-alkoxybenzochroman 10 is stable under
neutral conditions, but becomes hydrolytically labile at pH < 3.[31] Linker 10 was then EDC-coupled
to commercial carboxylate-functionalized silica microbeads to yield
target azide-derivatized microbeads bearing acid-labile linker (11, Scheme 4).
Scheme 4
Photo-Diels-Alder
Click Derivatization of Silica Microbeads with
Azide Moiety
Silica beads bearing
azido functionality (11) were
first derivatized with linker 1 (click 1) by overnight
incubation with ADIBO-photoDIBO (1) in DMF (Scheme 5). The resulting microbeads were added to 10 μM
PBS solution of azido-BSA (7) and irradiated with 350
nm light for 2 min (Scheme 5). The reaction
mixture was incubated under ambient conditions for 16 h and the beads
were separated and washed. The Bradford assay of the combined supernatants
from the reaction mixture after click reaction 2 showed
that there was virtually no protein present in the solution. This
observation illustrates the high efficiency of azido-BSA (7) immobilization using linker 1.
Scheme 5
Immobilization of
Azido-BSA on Azide-Functionalized Silica Microbeads
As BSA was immobilized to silica beads via an
acid-sensitive 2-alkoxybenzochroman
fragment,[31] the protein can be cleaved
from the solid support at low pH (Scheme 6).
To validate the release of the protein, silica microbeads bearing
immobilized BSA were incubated in 0.1 M perchloric acid overnight.
The protein released from the beads was isolated from the supernatant
via spin filtration and reconstituted to the original volume in PBS.
Total protein concentration in the resulting solution was determined
to be 9.12 μM by Bradford assay, which corresponds to the ca.
91% recovery. The BSA protein is stable in an aqueous 0.1 M perchloric
acid solution, at least for the time frame (16 h) used for the release
of protein from silica microbeads.
Scheme 6
Release of BSA from Silica Microbeads
Conclusions
The
heterobifunctional ADIBO-photoDIBO linker (1)
allows for the efficient ligation of azide-tagged substrates, or for
the immobilization of azide-functionalized molecules on an azide-coated
surfaces. Since the covalent conjugation is achieved via strained-promoted
azide–alkyne cycloaddition (SPAAC) and photochemical activation,
this method does not require any catalysts or other reagents. This
sequential click strategy is applicable to both homogeneous and heterogeneous
cross-linking. The use of phototrigger in the sequential click system
enables the spatiotemporal control to the cross-linking chemistry.
Additionally, the incorporation of a hydrolytically labile fragment
in protein–surface linker, allows for the release of a protein
under the action of dilute acid.
Experimental Section
General Information
All organic solvents were dried
and freshly distilled before use. Flash chromatography was performed
using 40–63 μm silica gel. All NMR spectra were recorded
on 400 MHz instruments in CDCl3 and referenced to TMS unless
otherwise noted. Solutions for HPLC and UV–vis analysis were
prepared using HPLC grade solvents. HPLC analysis was conducted using
analytical C-18 column and a diode array detector. Photochemical decarbonylation
of photoDIBO moiety was conducted by the irradiation of the reaction
mixture for 1–2 min in a photochemical reactor equipped with
four fluorescent UV lamps (4 W, 350 nm).
Materials
Bovineserum albumin (BSA) was purchased
from Boehringer Mannheim. ADIBO-carboxylic acid (5),[25] cyclopropenone-masked DIBO-OH (6),[23a] 8-(2-(2-(2-aminoethoxy)ethoxy)ethoxy)-3-(hydroxymethyl)naphthalen-2-ol
(9),[17d] (2-azidoethoxy)ethyl
vinyl ether,[17d] 5- and 6-(3-azidopropylcarbamoyl)fluorescein
(azido-Fl, 8, mixture of isomers)[23b] were prepared following previously reported procedures.
Carboxy-functionalized silica microspheres (0.01 mmol of acid/g, 1
μm) was purchased from Polysciences Inc.
ADIBO-PhotoDIBO Cross-Linker (1)
N,N′-Dicyclohexylcarbodiimide (62
mg, 0.3 mmol) and a catalytic amount of DMAP were added to a solution
of ADIBO-carboxylic acid (5) (150 mg, 0.38 mmol) in 8
mL of dry DMF, followed by a dropwise addition of a solution of alcohol 6 (151 mg, 0.42 mmol) in 2 mL of DMF. The mixture was stirred
for 12 h at r.t., the solvent was removed in a vacuum, the residue
was dissolved in DCM, washed with NaHCO3 solution, brine,
dried over anhydrous magnesium sulfate, and concentrated under reduced
pressure. The residue was purified by column chromatography (10% MeOH
in DCM) to yield 225 mg (80%) of ADIBO-photoDIBO (1). 1H NMR: 7.93 (dd, J = 8.9, 6.6 Hz, 2H), 7.66
(dd, J = 7.4, 1.5 Hz, 1H), 7.43–7.21 (m, 7H),
6.83–6.93 (m, 4H), 6.06 (t, J = 6.6 Hz, 1H),
5.12 (d, J = 13.9 Hz, 1H), 4.07 (t, J = 6.4 Hz, 2H), 4.02 (t, J = 6.4 Hz, 2H), 3.87 (s,
3H), 3.59 (d, J = 13.9 Hz, 1H), 3.29–3.39
(m, 3H), 3.16–3.27 (m, 1H), 2.55–2.65 (m, 2H), 2.30–2.50
(m, 1H), 2.28 (t, J = 7.5 Hz, 2H), 2.04 (t, J = 7.5 13 Hz, 2H), 1.76–1.99 (m, 3H), 1.72–1.38
(m, 6H). 13C NMR: 173.6, 172.7, 172.3, 162.9, 162.5, 154.2,
151.5, 148.5, 148.3, 148.3, 142.8, 142.6, 136.3, 136.3, 132.6, 129.5,
129.1, 128.9, 128.8, 128.3, 127.7, 126.1, 123.4, 123.0, 116.8, 116.7,
116.3, 115.3, 112.8, 112.3, 108.2, 68.6, 64.8, 63.3, 55.99, 55.98,
35.9, 35.7, 35.2, 34.0, 29.5, 29.1, 26.2, 21.3. FW calc [(C46H44N2O7)H+]: 737.3221;
ESI-HRMS: 737.3217.
Sequential Labeling
of BSA with Azido-Fl
3-Azidoprop-1-yl
iodide (2.1 mg, 10 μmol) was added to a solution of BSA (66
mg, ca. 1 μmol) in 3 mL of 0.1 N phosphate buffer (pH = 8) containing
0.5 mL of acetonitrile and gently shaken for 12 h at r.t. The aqueous
layer was washed with ethyl acetate and was freeze-dried to produce
azido-derivatized BSA (80 mg, contains some phosphate salt). Ellman’s
test performed after completion of the reaction gave negative results.
Azido-derivatized BSA (7) was further purified by spin
filtration. MALDI TOF: 66546.0.3 mg of linker 1 (400 μmol) was added to the solution of azido-BSA 7 (13.2 mg, ca. 0.2 μmol) in 2 mL of PBS and the mixture was
incubated for 16 h at r.t. on a mechanical shaker. The excess of 1 was removed by filtration through a Dextran desalting column.
The product 2c was characterized by MALDI-TOF: 67283.A solution of BSA 2c (13.2 mg, ca. 0.2 μmol)
in 2 mL of PBS was irradiated for 2 min and a solution of azido-Fl 8 (0.5 mg ca 1.1 μM) in 2 μL of DMF was added.
The reaction mixture was incubated 16 h at r.t. on a mechanical shaker.
The unreacted azido-fluorescein was removed by filtration through
a Dextran desalting column. Fluorescein–BSA conjugate 4c was characterized by MALDI-TOF: 67713. Total protein content
in the solution of 4c was determined by Bradford assay
using Brilliant Blue G-250 dye. The Bradford assay standard curve
was obtained from a series of standard BSA solution in the concentration
range 1–15 μM and the unknown protein concentration was
determined from the quadratic fit of the standard data. The concentration
of protein in fluorescein–BSA conjugate 4c was
found to be 11.13 μM. The emission (519 nm) and absorbance (494
nm) of this solution were compared to corresponding spectral features
of 11 μM solution of fluorescein, confirming 84–92% of
BSA derivatization.
Preparation of
Azide-Derivatized Silica Beads.
(2-Azidoethoxy)ethyl
vinyl ether (96 mg, 0.6 mmol) was added to a solution of 8-(2-(2-(2-aminoethoxy)ethoxy)ethoxy)-3-(hydroxymethyl)
naphthalen-2-ol (9, NH2-TEG-NQMP, 70 mg, 0.22
mmol) in 20 mL of aqueous acetonitrile (1:1). The reaction mixture
was irradiated for 1 h. The solvent and excess of (2-azidoethoxy)ethyl
vinyl ether was removed under reduced pressure. The crude amine 11 was taken into the next step without further purification.
EDC.HCl (66 mg, 0.3 mmol) and a catalytic amount of DMAP were added
to a dispersion of carboxy-functionalized silica microspheres (2 g)
in 20 mL of dry DMF, followed by addition of amine 11. The mixture was stirred for 48 h at r.t. and the solvent was removed
by centrifugation. The azido derivatized silica beads (12) were thoroughly washed with DMF trice and dried under nitrogen.
Immobilization of Azido-BSA 7 on Azide-Coated Silica
Microbeads (12)
A solution of ADIBO-photoDIBO 1 (1.5 mg, 2 μM) in 10 μL of methanol was added
to a suspension of the azide-derivatized silica beads 11 (100 mg) in DMF (10 mL) and incubated overnight. The beads were
filtered, washed, and allowed to dry. Excess of silica beads (50 mg,
0.5 μmol in ADIBO-photo DIBO) were added to 3.5 mL of 10 μM
PBS solution azido-BSA 7 (2.33 mg, 0.035 μmol)
and irradiated under vigorous stirring with 350 nm light for 2 min
(Scheme 5). The reaction mixture was incubated
under ambient conditions for 16 h and the beads were separated by
centrifugation and washed.
Release
of BSA from Solid Support
50 mg of silica microbeads
with immobilized BSA were incubated in 1 mL of 0.1 M perchloric acid
overnight under mild shaking. The beads were separated and washed
with 1 mL PBS buffer 5 times. The protein was isolated from combined
supernatants (6 mL) by spin-filtration and reconstituted in 3.5 mL
of PBS. The resulting solution was analyzed by the Bradford Assay.
Authors: Ibai E Valverde; Fabien Lecaille; Gilles Lalmanach; Vincent Aucagne; Agnès F Delmas Journal: Angew Chem Int Ed Engl Date: 2011-12-05 Impact factor: 15.336
Authors: Brian C Sanders; Frédéric Friscourt; Petr A Ledin; Ngalle Eric Mbua; Selvanathan Arumugam; Jun Guo; Thomas J Boltje; Vladimir V Popik; Geert-Jan Boons Journal: J Am Chem Soc Date: 2010-12-23 Impact factor: 15.419
Authors: Merel A Nessen; Gertjan Kramer; JaapWillem Back; Jeremy M Baskin; Linde E J Smeenk; Leo J de Koning; Jan H van Maarseveen; Luitzen de Jong; Carolyn R Bertozzi; Henk Hiemstra; Chris G de Koster Journal: J Proteome Res Date: 2009-07 Impact factor: 4.466