Acid-labile traceless click linker for protein transduction

Article abstract of DOI:10.1021/ja302705v

Intracellular delivery of active proteins presents an interesting approach in research and therapy. We created a protein transduction shuttle based on a new traceless click linker that combines the advantages of click reactions with implementation of reversible pH-sensitive bonds. The azidomethyl-methylmaleic anhydride (AzMMMan) linker was found compatible with different click chemistries, demonstrated in bioreversible protein modification with dyes, polyethylene glycol, or a transduction carrier. Linkages were stable at physiological pH but reversible at the mild acidic pH of endosomes or lysosomes. We show that pH-reversible attachment of a defined endosome-destabilizing three-arm oligo(ethane amino)amide carrier generates an effective shuttle for protein delivery. The cargo protein nlsEGFP, when coupled via the traceless AzMMMan linker, experiences efficient cellular uptake and endosomal escape into the cytosol, followed by import into the nucleus. In contrast, irreversible linkage to the same shuttle hampers nuclear delivery of nlsEGFP which after uptake remains trapped in the cytosol. Successful intracellular delivery of bioactive -galactosidase as a model enzyme was also demonstrated using the pH-controlled shuttle system.

Full text of DOI:10.1021/ja302705v

Article
pubs.acs.org/JACS
Acid-Labile Traceless Click Linker for Protein Transduction
Kevin Maier and Ernst Wagner*
Pharmaceutical Biotechnology, Center for System-based Drug Research, and Center for Nanoscience Ludwig-Maximilians-University
Munich, Butenandtstrasse 5-13, 81377 Munich, Germany
S
* Supporting Information
ABSTRACT: Intracellular delivery of active proteins presents
an interesting approach in research and therapy. We created a
protein transduction shuttle based on a new traceless click
linker that combines the advantages of click reactions with
implementation of reversible pH-sensitive bonds. The
azidomethyl-methylmaleic anhydride (AzMMMan) linker
was found compatible with different click chemistries,
demonstrated in bioreversible protein modification with
dyes, polyethylene glycol, or a transduction carrier. Linkages were stable at physiological pH but reversible at the mild acidic
pH of endosomes or lysosomes. We show that pH-reversible attachment of a defined endosome-destabilizing three-arm
oligo(ethane amino)amide carrier generates an effective shuttle for protein delivery. The cargo protein nlsEGFP, when coupled
via the traceless AzMMMan linker, experiences efficient cellular uptake and endosomal escape into the cytosol, followed by
import into the nucleus. In contrast, irreversible linkage to the same shuttle hampers nuclear delivery of nlsEGFP which after
uptake remains trapped in the cytosol. Successful intracellular delivery of bioactive ß-galactosidase as a model enzyme was also
demonstrated using the pH-controlled shuttle system.
a
Scheme 1. Synthesis of Linker 3 (AzMMMan)
INTRODUCTION
■
In recent years click chemistry reactions, especially the copper-
catalyzed 1,3-dipolar cycloaddition (CuAAC) and the Stau-
dinger ligation, became useful tools for conjugating biomole-
cules.1−4 Both reactions have great advantages compared to
other linking strategies, like high efficiency and bioorthogon-
ality. Nevertheless, for some applications, such as drug delivery,
a completely bioreversible bond between the two conjugated
molecules would be favorable. Dimethylmaleic anhydride is
known to form amide bonds with amines that are cleaved under
mild acidic conditions.5−10 Herein we describe the synthesis
and applications of a new heterobifunctional linker based on
substituted dimethylmaleic anhydride that combines the
advantages of click chemistry with implementation of a pH-
sensitive bond between conjugated biomolecules. This linker is
labile under mild acidic conditions, which are typical for tumor
tissue11 or early endosomes.12 Moreover, this linker is cleaved
off traceless, resulting in an unmodified molecule of interest.
The utility in protein modification is demonstrated in examples
of reversible dye labeling of proteins, pH-sensitive modification
with polyethylene glycol (PEG), and intracellular transduction
of proteins in bioactive form.
a
(a) N-Bromosuccinimide, benzoyl peroxide, 56%; (b) sodium azide,
88%.
or ß-galactosidase (ß-Gal) (Scheme 2, step a) by reaction of the
maleic anhydride moiety with amino groups of the proteins.
The AzMMMan−HSA conjugate 4 was coupled with an
alkyne-bearing hemicyanine dye 12 (synthesized as shown in
Scheme S1, Supporting Information) in a copper-catalyzed
fashion (Scheme 2, step b), resulting in conjugate 7. Alkyne dye
12 was used, on the one hand, to determine the degree of
protein modification with the new linker 3 (Figure S2,
Supporting Information) and, on the other hand, to investigate
cleavage kinetics under acidic conditions (Figure 1).
As expected, the number of azides introduced in HSA was
linear to the initial concentration of 3 (Figure S2, Supporting
Information). Figure 1 displays the kinetics of pH-dependent
cleavage of dye−conjugate 7. Dye-labeled HSA 7 was incubated
in buffers of different pH. Cleavage of dye 12 from the protein
RESULTS AND DISCUSSION
■
The azidomethyl-methylmaleic anhydride linker 3 (AzMM-
Man) was synthesized from dimethylmaleic anhydride 1 by two
simple reaction steps (radical substitution with N-bromosucci-
nimide resulting in 2 followed by reaction with sodium azide,
see Scheme 1).
The heterobifunctional linker was used to introduce acid-
labile azido groups into human serum albumin (HSA), EGFP,
Received: March 20, 2012
© XXXX American Chemical Society
A
dx.doi.org/10.1021/ja302705v | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
Scheme 2. pH-Reversible Modification of Proteins with
a
Dyes, PEG, or Transduction Oligomer
Figure 2. Serum stability of conjugate 7 at 37 °C. Samples of 7 were
incubated in PBS buffer pH 7.4 containing 30% FCS. Dye remaining
conjugated with the protein was determined by measuring the
absorbance of the conjugate after purification from released dye.
dependent cleavage of the two HSA conjugates independent of
being coupled by Staudinger ligation or CuAAC.
a
(a) Modification of nlsEGFP, HSA, or ß-Gal with AzMMMan linker
Another application of linker 3 is reversible PEGylation of
proteins (Scheme 2, step d). Figure 3 gives evidence that HSA
3. CuAAC conjugation of AzMMMan−HSA 4 with (b) alkyne−
hemicyanin dye 12 or (d) alkyne−PEG, resulting in conjugate 7 or 8,
respectively. (c) Staudinger ligation of Dylight 488 (R4) to
AzMMMan−HSA resulting in conjugate 11. (e) Conjugation of
AzMMMan−nlsEGFP 5 or AzMMMan−ß-Gal 6 with carrier polymer
386 (R1) via copper-free cycloaddition resulting in conjugate 9 or 10,
respectively.
Figure 3. Acid-catalyzed release of PEG from HSA conjugate 8. SDS−
polyacrylamide gel electrophoresis after 16 h incubation: lane 1
marker, lane 2 unmodified HSA, lane 3 pH 8.5, lane 4 pH 7.3, lane 5
pH 6, lane 6 pH 5, lane 7 pH 4.
Figure 1. Acid-catalyzed release of dye 12 from conjugate 7 at 37 °C.
Samples of 7 were incubated in buffers of different pH. At each time
point, the fraction (%) of dye remaining conjugated with HSA was
determined after protein purification. Percentage of released dye is
calculated as (100 − % HSA conjugated dye).
can be effectively coupled with alkyne-bearing PEG (lane 3).
The PEG conjugate 8 shows extended stability at physiological
pH (lane 4) but is cleaved under very mild acidic conditions of
pH 6 (lane 5). At pH 5 or 4 almost all PEG is released (lanes 6
and 7).
A more detailed cleavage kinetic of PEG−AzMMMan−HSA
at the endosomal pH of 5 is displayed in Figure 4. For this
purpose PEG was distally end labeled with tetramethylrhod-
amine (TMR) dye. Release of TMR−PEG from HSA conjugate
was determined at various times by measuring the absorbance
of TMR dye remaining incorporated in purified PEGylated
protein conjugate. In contrast to the type of coupling reaction
which has a minor influence on cleavage kinetics (Figure S3,
Supporting Information) the coupled molecule has more
impact on acidic bond cleavage. PEG release from the protein
was considerably slower compared with dye released from
conjugate 7 (Figure 1).
was determined at various times by measuring residual dye
absorbance of purified protein fractions. At basic pH of 8.5 the
conjugate was quite stable with only about 25% cleavage after
24 h. In contrast, at pH 6 the same amount was released already
after 30 min. The half-life at pH 5 is about 0.5 h, whereas at pH
4 already 90% of the linker was cleaved in that time. Cleavage
of the linker appears to follow pseudo-first-order kinetics.
Figure 2 reveals that conjugate 7 is quite stable under
physiological serum conditions at 37 °C. After 2 h incubation in
buffer containing 30% FCS still 94% of the dye remained
coupled to HSA. Even after an incubation time of 12 h 60% of
the bonds still are intact. The cleavage rate does not differ much
from incubation in pure PBS buffer at physiological pH (Figure
1).
Also, the so-called Staudinger ligation can be performed with
azido-proteins.3 A phosphine-containing dye was coupled to
AzMMMan−HSA 4 (Scheme 2, step c), resulting in conjugate
11. Figure S3, Supporting Information, shows similar pH-
Protein transduction, i.e., the intracellular delivery of
proteins, can take advantage of the endosomal acidification
process.13−15 nlsEGFP bearing a nuclear localization sequence
(which mediates natural active nuclear import once a protein
B
dx.doi.org/10.1021/ja302705v | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
Figure 4. Kinetics of PEG release from HSA conjugate at an
endosomal relevant pH of 5, and stability of the conjugate at
physiological pH of 7.3. TMR-labeled PEG released from HSA protein
was determined by comparing the absorbance of the intact PEG−HSA
conjugate before and after incubation.
Figure 5. Acid-catalyzed release of TMR−386 from conjugate 9 at 37
°C. Samples of TMR-labeled 9 were incubated in PBS buffers of pH
7.3 and 5.0. Dye-labeled 386 released from the protein was determined
indirectly by measuring the absorbance of the conjugate after
purification.
resides in the cytosol) was bound to the three-arm cationic
oligo(aminoethane) amide 386 (Scheme 2, step e). Oligomer
386 (Scheme 3) was previously developed as a precise nontoxic
the tertiary structure of nlsEGFP is not significantly affected by
coupling and consecutive cleavage of the carrier 386.
The cell culture experiment shown in Figure 6 demonstrates
how important intracellular cleavage can be for protein delivery
in the case of covalent binding to a carrier molecule using the
a
Scheme 3. Structure of Transduction Oligomer 386
a
Structure of 386 (x = H; Stp, succinoyl tetraethylene pentamine; C,
cysteine; K, α,ε-modified branching lysine).16
carrier for cytosolic siRNA delivery.16,17 Cationic carriers like
386 containing diaminoethane motifs bind cells and act as
proton sponges which upon endosomal acidification become
increasingly positively charged, triggering endosome disruption
and release of its content.18,19
For intracellular delivery of nlsEGFP protein, two different
linking strategies, conventional irreversible conjugation and pH-
reversible conjugation, were compared. For the synthesis of an
irreversible conjugate, the commercial linker SMCC was used,
resulting in thioether linkage. In the new strategy, oligomer 386
was coupled with AzMMMan−nlsEGFP 5 by copper-free
cycloaddition via dibencylcyclooctyne (see Scheme 2e). Thus, 9
contains an acid-labile bond between carrier 386 and nlsEGFP,
which is expected to be reversible in the endosomal
microenvironment before transfer into the cytosol.
Successful modification of the nlsEGFP protein with
transduction oligomer 386 and acidic lability of the resulting
conjugate 9 is shown in Figure S5, Supporting Information.
The cleavage kinetics was determined in more detail (Figure 5)
by labeling the 386 carrier with TMR dye and further
processing analogously as in Figure 4. As already observed
for the PEG−HSA conjugate, coupling of carrier 386 to
nlsEGFP is retarding the cleavage kinetics. Nevertheless, after 1
h incubation at endosomal pH 5.0, >20%, and after 2 h almost
50% of the transduction oligomer was released from the
nlsEGFP protein.
Figure 6. Transduction of HeLa cells with 386-modified nlsEGFP.
Row 1: carrier 386 was coupled with nlsEGFP via the cleavable
AzMMMan linker 9. Row 2: carrier 386 was coupled by irreversible
SMCC thioether linkage. Row 3: transfection was done with
unmodified nlsEGFP. Row 4: transfection with conjugate 9 at 4 °C
in the presence of endocytosis inhibitors. Row 5: transfection with
conjugate 9 in the presence of endosome acidification inhibitors. (A)
Brightfield picture of the transfected cells, (B) EGFP fluorescence of
the transfected cells, and (C) Hoechst 33342 DNA stain of the cell
nucleus.
The native conformation of nlsEGFP after release of 386
from conjugate 9 was investigated by circular dichroism
spectroscopy. Figure S6, Supporting Information, reveals that
C
dx.doi.org/10.1021/ja302705v | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
novel linker 3. Both constructs, based on the pH-reversible
conjugation or on conventional irreversible conjugation by
SMCC thioether linkage, were able to deliver nlsEGFP into the
cytosol of HeLa cells (Figure 6, rows 1 and 2). Flow cytometry
experiments suggest that the stable thioether construct seems
to be slightly more efficient (data not shown). Cells in row 3
were transfected with unmodified nlsEGFP as a control. The
fluorescence image shows that no protein was internalized.
However, only nlsEGFP which had been coupled to the
carrier by acid-labile bonds was further translocated into the
nucleus to a large extent (Figure 6, row 1). The lack of nuclear
import of the irreversibly modified nlsEGFP is consistent with
direct (by chemical modification of the nls lysines) or indirect
(by sterical shielding) inactivation of the nuclear localization
signal. Alternatively, active nuclear transport may be hampered
by unspecific cytosolic retention of the polycationic trans-
duction shuttle. In contrast, the nuclear import of the pH-labile
click conjugate 9 can be explained by a release of unmodified
nlsEGFP in the acidifying endosome before escape into the
cytosol, followed by nls-mediated nuclear import.
Transfection of cells with conjugate 9 at 4 °C in the presence
of three different endocytosis inhibitors (amiloride (inhibitor
for macropinocytosis), chloroproamzine (inhibitor for clathrin-
mediated endocytosis), β-cyclodextrin (inhibitor for caveolae-
mediated endocytosis)) almost completely eliminates internal-
ization (Figure 6, row 4). This observation suggests an active,
energy-dependent internalization mechanism.
The experiment shown in Figure 6, row 5, highlights
acidification in the endosome as a critical requirement for
intracellular release of the carrier. Cells were preincubated and
transfected in the presence of chloroquine and ammonium
chloride. Both substances prevent acidification of the cellular
endosomes. Transfection under these conditions results in a
similar observation that was made for the noncleavable
construct. The conjugate is taken up by the cell, and the
balanced distribution suggests at least some endosomal escape,
but only minor translocation of the protein into the nucleus
was observed.
untreated cells was observed. Even after transfection with high
amounts of 80 μg of 386 nlsEGFP conjugate cell viability
decreased by only 33%. The toxicity profile of stable and
reversible bound carrier does not differ much. However, at the
highest tested concentration the irreversible SMCC construct
seems to be slightly more toxic.
The fluorescence properties and CD spectra of construct 9
indicate that the protein conformation is not significantly
affected by modification with the linker and transduction
oligomer. Therefore, we examined whether the activity of an
enzyme would be maintained and survive modification,
delivery, and intracellular release steps. As a model enzyme
we chose ß-galactosidase (ß-Gal). It is of rather big size (119
kDa), and the activity of the enzyme in the cytosol can easily be
investigated by use of the fluorescent substrates.
Carrier 386 was covalently bound to the ß-Gal protein over
the AzMMMan linker (Scheme 2, step e), resulting in
conjugate 10. This construct was able to transport biologically
active ß-galactosidase into the cytosol of HeLa cells. ß-Gal
activity was detected using the fluorescent substrate C12-FDG
(C12-fluorescein-dibeta-D-galactopyranoside), which exhibits
FITC fluorescence after cleavage. This fluorescence was
determined by flow cytometric analysis. Already with a
transfection concentration of 1 μM a significant raise in
fluorescence was observed. Transfection with 2.5 μM ß-Gal
further raised the fluorescent properties. In contrast, cells
treated with 2.5 μM unmodified ß-Gal (without shuttle) did
not show a significant shift in fluorescence. This experiment
evidences that the transduction shuttle is able to transduce
biologically active proteins into the cytosol of cells.
Transduction oligomer 386 has been shown to exhibit low
cytotoxicity when used as a carrier for siRNA or DNA.16,17 To
verify the nontoxic properties also in the case of being
covalently bound to a protein, a cell viability test was performed
that confirmed this assumption. Figure 7 shows the low toxicity
of the transduction shuttle, even at high transduction
concentrations. It exhibited only marginally higher toxicity
than the unmodified nlsEGFP protein. With the standard
concentration used for transfections above (10 μg/mL) a
decline in metabolic activity of only 10% in comparison to
Figure 8. HeLa cells were transfected with conjugate 10 (386−
AzMMMan−ßGal), washed and incubated with C12-FDG substrate,
and evaluated by flow cytometry. I: untreated cells. II: cells transfected
with 2.5 μM unmodified ß-Gal. III: cells transfected with 1 μM
conjugate 10. IV: cells treated with 2.5 μM of conjugate 10.
CONCLUSION
■
In summary, we present a new pH-sensitive linker 3 which
combines the advantages of click chemistry with implementa-
tion of a traceless cleavable bond between two conjugated
molecules. Three different click chemistries were performed
which all are compatible with the acid-labile properties.
Traceless cleavage may be a particularly important feature in
protein transduction strategies to maintain full bioactivity of
enzymes and other proteins. The current example of 386
Figure 7. Cell viability assay. HeLa cells were transfected with different
amounts of 386−nlsEGFP conjugates. Unmodified nlsEGFP was used
for comparison.
D
dx.doi.org/10.1021/ja302705v | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
(15) Murthy, N.; Campbell, J.; Fausto, N.; Hoffman, A. S.; Stayton, P.
S. Bioconjugate Chem. 2003, 14, 412.
(16) Schaffert, D.; Troiber, C.; Salcher, E. E.; Frohlich, T.; Martin, I.;
Badgujar, N.; Dohmen, C.; Edinger, D.; Klager, R.; Maiwald, G.;
Farkasova, K.; Seeber, S.; Jahn-Hofmann, K.; Hadwiger, P.; Wagner, E.
Angew. Chem., Int. Ed. Engl. 2011, 50, 8986.
(17) Frohlich, T.; Edinger, D.; Klager, R.; Troiber, C.; Salcher, E.;
Badgujar, N.; Martin, I.; Schaffert, D.; Cengizeroglu, A.; Hadwiger, P.;
Vornlocher, H. P.; Wagner, E. J. Controlled Release 2012, DOI: http://
dx.doi.org/10.1016/j.jconrel.2012.03.018.
(18) Miyata, K.; Nishiyama, N.; Kataoka, K. Chem. Soc. Rev. 2012, 41,
2562.
(19) Wagner, E. Acc. Chem. Res. 2011, DOI: http://dx.doi.org/
10.1021/ar2002232.
(20) Wagner, E. Expert Opin. Biol. Ther. 2007, 7, 587.
carrier-mediated cytosolic delivery and subsequent nuclear
import of released nls-EGFP demonstrates the advantage of the
traceless linker. To demonstrate that the modification does not
irreversibly affect the structure and biological activity of
proteins, we transported 386-modified ß-galactosidase as a
model enzyme which exhibited cytosolic activity in transduced
cells far higher than without shuttle. Aside from these
encouraging options for protein delivery and modification,
the linker might have broader use in the design of novel
programmed, acid-labile, and biodegradable drug delivery
systems.19,20 Targeted therapeutics could, after delivery into
acidic tumor areas or upon cellular uptake into endosomes, be
dismantled from their outer shell including targeting ligands.
Besides drug delivery, the linker may also be of interest for
other applications such as reversible labeling of various
biological and also chemical molecules.
ASSOCIATED CONTENT
■
S
* Supporting Information
Figures, analytical data, and experimental protocols. This
material is available free of charge via the Internet at http://
pubs.acs.org.
AUTHOR INFORMATION
■
Corresponding Author
Ernst.Wagner@cup.uni-muenchen.de
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
This work was supported by the DFG excellence cluster
Nanosystems Initiative Munich (NIM). We acknowledge the
great help by Wolfgang Rodl and Dr. Martina Ruffer.
̈
̈
REFERENCES
■
(1) Saxon, E.; Bertozzi, C. R. Science (New York, N.Y.) 2000, 287,
2007.
(2) Yu, H.; Nie, Y.; Dohmen, C.; Li, Y.; Wagner, E. Biomacromolecules
2011, 12, 2039.
(3) Vallee, M. R.; Majkut, P.; Wilkening, I.; Weise, C.; Muller, G.;
Hackenberger, C. P. Org. Lett. 2011, 13, 5440.
(4) Dohmen, C.; Frohlich, T.; Lachelt, U.; Rohl, I.; Vornlocher, H.-
P.; Hadwiger, P.; Wagner, E. Mol. Ther. Nucleic Acids 2012, 1, e7.
(5) Meyer, M.; Philipp, A.; Oskuee, R.; Schmidt, C.; Wagner, E. J.
Am. Chem. Soc. 2008, 130, 3272.
(6) Meyer, M.; Zintchenko, A.; Ogris, M.; Wagner, E. J. Gene Med.
2007, 9, 797.
(7) Nieto, M. A.; Palacian, E. Biochim. Biophys. Acta 1983, 749, 204.
(8) Meyer, M.; Dohmen, C.; Philipp, A.; Kiener, D.; Maiwald, G.;
Scheu, C.; Ogris, M.; Wagner, E. Mol. Pharmaceutics 2009, 6, 752.
(9) Rozema, D. B.; Lewis, D. L.; Wakefield, D. H.; Wong, S. C.;
Klein, J. J.; Roesch, P. L.; Bertin, S. L.; Reppen, T. W.; Chu, Q.;
Blokhin, A. V.; Hagstrom, J. E.; Wolff, J. A. Proc. Natl. Acad. Sci. U.S.A.
2007, 104, 12982.
(10) Rozema, D. B.; Ekena, K.; Lewis, D. L.; Loomis, A. G.; Wolff, J.
A. Bioconjugate Chem. 2003, 14, 51.
(11) Engin, K.; Leeper, D. B.; Cater, J. R.; Thistlethwaite, A. J.;
Tupchong, L.; McFarlane, J. D. Int. J. Hyperthermia 1995, 11, 211.
(12) Murphy, R. F.; Powers, S.; Cantor, C. R. J. Cell Biol. 1984, 98,
1757.
(13) Lee, Y.; Ishii, T.; Kim, H. J.; Nishiyama, N.; Hayakawa, Y.; Itaka,
K.; Kataoka, K. Angew. Chem., Int. Ed. Engl. 2010, 49, 2552.
(14) Bachelder, E. M.; Beaudette, T. T.; Broaders, K. E.; Dashe, J.;
Frechet, J. M. J. Am. Chem. Soc. 2008, 130, 10494.
E
dx.doi.org/10.1021/ja302705v | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX