Constructing hybrid protein zymogens through protective dendritic assembly

Article abstract of DOI:10.1002/anie.201308533

The modulation of protein uptake and activity in response to physiological changes forms an integral part of smart protein therapeutics. We describe herein the self-assembly of a pH-responsive dendrimer shell onto the surface of active enzymes (trypsin, papain, DNase I) as a supramolecular protecting group to form a hybrid dendrimer-enzyme complex. The attachment is based on the interaction between boronic acid and salicyl hydroxamate, thus allowing the macromolecular assembly to respond to changes in pH between 5.0 and 7.4 in a highly reversible fashion. Catalytic activity is efficiently blocked in the presence of the dendrimer shell but is quantitatively restored upon shell degradation under acidic conditions. Unlike the native proteases, the hybrid constructs are shown to be efficiently taken up by A549 cells and colocalized in the acidic compartments. The programmed intracellular release of the proteases induced cytotoxicity, thereby uncovering a new avenue for precision biotherapeutics. The programmed self-assembly of a dendritic shell onto enzymes was used to modulate enzyme activity as well as induce cellular entry and release of the active proteins. The defined dendritic construct represents a contemporary avenue for smart protein therapeutics. Copyright

Full text of DOI:10.1002/anie.201308533

.
Angewandte
Communications
DOI: 10.1002/anie.201308533
Supramolecular Enzyme Hybrids
Constructing Hybrid Protein Zymogens through Protective Dendritic
Assembly**
David Y. W. Ng, Matthias Arzt, Yuzhou Wu, Seah Ling Kuan, Markus Lamla, and Tanja Weil*
Abstract: The modulation of protein uptake and activity in
response to physiological changes forms an integral part of
smart protein therapeutics. We describe herein the self-assem-
bly of a pH-responsive dendrimer shell onto the surface of
active enzymes (trypsin, papain, DNase I) as a supramolecular
protecting group to form a hybrid dendrimer–enzyme com-
plex. The attachment is based on the interaction between
boronic acid and salicyl hydroxamate, thus allowing the
macromolecular assembly to respond to changes in pH
between 5.0 and 7.4 in a highly reversible fashion. Catalytic
activity is efficiently blocked in the presence of the dendrimer
shell but is quantitatively restored upon shell degradation
under acidic conditions. Unlike the native proteases, the hybrid
constructs are shown to be efficiently taken up by A549 cells
and colocalized in the acidic compartments. The programmed
intracellular release of the proteases induced cytotoxicity,
thereby uncovering a new avenue for precision biotherapeutics.
therapeutics. As such, the design of a synthetic strategy
incorporating a responsive “assemble and release” system
mimicking the biosynthesis of zymogens represents an ideal
model. Importantly, the innate proteolytic activity of pro-
teases precludes them from most biological delivery systems
(i.e. cell-penetrating peptides, fusion proteins).
The modification of active proteases to include stimulus-
responsive moieties is challenged by protein denaturation and
post-modification inactivation as well as by their chemo-
selectivity which often necessitates the use of mild bioor-
thogonal approaches.[4] Conventionally, site-specific modifi-
cation through single-residue mutation has been developed to
incorporate thermoresponsive steric handles near the cata-
lytic site acting as a temperature-dependent switch.[5–7] This
concept subsequently evolved into chemical strategies
towards the development of protein–polymer conjugates
assembled using in situ atom-transfer radical polymerization
(ATRP).[8,9] However, the site-specific mutation of different
classes of enzymes is often challenging owing to their highly
diverse activity and stability, whereas metal-catalyzed reac-
tions on proteins are plagued by the adsorption of metal
contaminants. As a result, the application of these enzymatic
constructs is still highly confined and underdeveloped.
In this context, macromolecular protecting groups involv-
ing metal-free reactions under physiological conditions for
the protection/deprotection of enzyme catalytic sites offer
a distinct advantage over existing systems. Inspired by recent
results on engineered aptamers[10] as macromolecular pro-
tecting groups, we report herein the use of a stimulus-
responsive dendritic core–shell system that self-assembles
into a supramolecular dendrimer–enzyme complex in a facile
and bioorthogonal manner. Small defined dendrimer seg-
ments (dendrons) provide a high volume to molecular weight
ratio due to their branched structure and thus serve as an ideal
T
he application of stimulus-responsive protective groups
offering precise control of protein activity is an appealing
approach in the preparation of protein-based drugs with
potentially high therapeutic index.[1] In this context, nature
introduces peptidic protecting groups to proteases (trypsin,
pepsin, caspases) resulting in the formation of zymogens,
which facilitate efficient intra/extracellular transportation
and prevent unintended cellular degradation.[2] We were
inspired by this transcendental efficacy and considered that
the introduction of synthetic macromolecular protecting
groups would provide a contemporary chemical perspective
by masking the function of an enzyme through steric bulk
prior to exposure to a designated stimulus.
Proteases play a central role in cancer progression and
metastasis, regulating processes such as proliferation, apop-
tosis, differentiation, and evasion of the immune system.[3]
The delivery of proteases is important in the elucidation of
their function as well as the development of new protein
candidate as
a sterically demanding, bulky protecting
group.[11] In addition, the dendrons confer additional phar-
macologically attractive properties (increased uptake,
potency)[12] depending on their surface functionalities.[13]
These dendrons are self-assembled onto the enzyme utilizing
an acid-labile boronic acid/salicyl hydroxamate based ligation
method, resulting in a pH-responsive dendrimer–enzyme
hybrid which functions as a synthetic zymogen (Scheme 1a).
We have chosen catalytically unique enzymes (trypsin,
papain, DNase I), each representing an enzyme class with
high therapeutic significance, to demonstrate the broad
applicability of this strategy.[3]
[*] D. Y. W. Ng, M. Arzt, Y. Wu, Dr. S. L. Kuan, M. Lamla, Prof. Dr. T. Weil
Institute of Organic Chemistry III, Ulm University
Albert-Einstein-Allee 11, 89081 Ulm (Germany)
E-mail: tanja.weil@uni-ulm.de
D. Y. W. Ng, Prof. Dr. T. Weil
Max Planck Institute for Polymer Research
Ackermannweg 10, 55128 Mainz (Germany)
[**] This work was supported by the Deutsche Forschungsgemeinschaft
(DFG) within the SFB 625, the BMBF project Biotechnologie
2020+, and the ERC Synergy Grant 319130-BioQ. S.L.K. is grateful
to the Alexander von Humboldt Foundation for providing a research
fellowship.
The synthesis of the dendrimer protective group is based
on the bifunctional second-generation poly(amido)amine
(PAMAM) dendron bearing a single azido group as well as
four primary amino groups (N3-G2 PAMAM, Scheme 1b)
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/anie.201308533.
324
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2014, 53, 324 –328

Angewandte
Chemie
Scheme 1. a) Synthesis and supramolecular dendritic assembly for the construction of hybrid zymogens. Catalytic residues are highlighted in
green. b) Synthesis of the salicyl hydroxamate core dendron.
which has been described previously.[14] The azido core group
of the dendron was converted into a salicyl hydroxamate
moiety by a copper-catalyzed azide–alkyne cycloaddition.
Salicyl hydroxamates have been previously reported to
“click” with aryl boronic acids to form a stable boronate
complex.[15] The kinetics and orthogonality of this fast
complexation have been studied and utilized to perform
efficient peptide ligation.[16] To construct the dendron core
group, 4-pentynoic acid was activated by N-hydroxysuccini-
mide and coupled with 4-aminosalicyclic acid to afford 3.
Condensation of the ethynyl salicyclic acid 3 with trityloxy-
amine in excess afforded 4 in moderate yield. The resulting
protected ethynyl hydroxamic acid underwent copper-cata-
lyzed cycloaddition with the Boc-protected azido G2
PAMAM dendron with high yield. Subsequent deprotection
using trifluoroacetic acid in the presence of triisopropylsilane
as a radical scavenger afforded the final product 5 as
a trifluoroacetate salt.
was found to be > 80% of that of their native counterparts
(Figure S2).
The construction of the dendritic assembly with dendron 5
was monitored using a fluorescence assay based on Alizarin
Red S. Alizarin Red S complexes with each aryl boronic acid
on its catechol moiety quantitatively to produce a fluores-
cence complex (lex = 495 nm, lem = 600 nm).[20] As observed,
the fluorescence quenching upon the addition of dendron 5
corresponds to the displacement of Alizarin Red S from the
protein surface (Figure 1a). The affinity constants of boronic
acid/catechol and boronic acid/salicyl hydroxamate at pH 7.4
were reported to be 800mÀ1 and 17800mÀ1, respectively,[16,21]
allowing the quantification of the number of boronic acid
groups and dendrons bound on each enzyme molecule
through a fluorescence-based titration. For each protein, the
endpoint of the titration with dendron 5 (10 equiv for trypsin
and papain, 4 equiv for DNase I) corresponds to the results
obtained earlier by MALDI-TOF analysis.
The native enzymes trypsin, papain, and DNase I were
functionalized by condensation of their surface lysine residues
with an excess of preactivated 4-carboxyphenylboronic acid
(Scheme 1a). As the catalytic domains of the proteases
primarily do not involve lysine residues, the surface lysines
become an attractive target for modification.[17–19] The modi-
fied proteins were purified using ultrafiltration and charac-
terized by MALDI-TOF mass spectrometry (Figure S1 in the
Supporting Information). Trypsin, papain, and DNase I were
functionalized with an average of ten, ten, and four boronic
acid residues, respectively, due to their different numbers of
solvent-accessible lysines. The degree of functionalization for
trypsin and papain is 90%, whereas only 50% of the surface
lysines were modified for DNase I. These modified proteins
were subsequently assayed for their catalytic activity, which
To evaluate the protection efficiency of the dendritic
assembly, the proteolytic activity of the boronic acid modified
proteases (trypsin and papain) were tested and quantified
using Na-benzoyl-l-arginine 4-nitroanilide hydrochloride as
the substrate. The dendrons were first conjugated onto both
proteins by incubating the two components at pH 7.4 prior to
the analysis. Incidentally, the activity of the hybrid construct
was reduced to < 10%, suggesting that the catalytic site of the
enzymes was sterically blocked by the attached dendrons
(Figure 1b,c). Upon reduction to pH 5.0, the dendrons
dissociated and were removed by ultrafiltration and the
activity of the enzymes was measured again at pH 7.4. The
recovery of activity upon the dissociation of dendrons was
calculated to be > 90%, supporting the reversibility of the
boronic acid/salicyl hydroxamate binding. To further substan-
Angew. Chem. Int. Ed. 2014, 53, 324 –328
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
325

.
Angewandte
Communications
Figure 1. a) Fluorescence titration curves displaying the stoichiometric substitution of the fluorogenic Alizarin Red S/boronic acid complex on
each protein by dendron 5. b,c) Colorimetric enzyme kinetics for modified trypsin and papain based on Na-benzoyl-l-arginine 4-nitroanilide
hydrochloride as a molecular substrate. d) Agarose gel electrophoretic analysis, using SYBR safe stain, of the activity of modified DNase I on
pUC19 DNA plasmid. Fluorescence intensity of bands in lanes iii and vi are shown in detail. e,f) SDS-PAGE analysis of the activity of modified
trypsin and papain on human serum albumin as a protein substrate.
tiate this observation, the dendronized proteases were tested
at pH 7.4 against human serum albumin (HSA) as a protein
substrate followed by SDS-PAGE analysis. The addition of
dendron 5 to both boronic acid modified proteases signifi-
cantly blocked the proteolytic activity, with observable native
albumin (67 kDa) clearly seen (Figure 1e,f). Correspondingly,
we found further support for the reversibility of the protective
dendritic shell when the dendrons were removed at pH 5.0
prior to the proteolytic digestion (Figure S3). In a similar
context, the activity of modified DNase I was tested against
a 2000 base pair plasmid (pUC19) in which the degradation of
the supercoiled plasmid into its linear form was effectively
blocked in the presence of dendron 5 at pH 7.4 and
subsequently recovered at pH 5.0 (Figure 1d).
The bioorthogonality and possible aggregation behavior
of the dendritic assembly was investigated using fluorescence
polarization. The interaction between the boronic acid
modified protease with dendron 5 demonstrated an expected
increase in size and displayed consistency in serum-fortified
DMEM (Figure S4). The absence of aggregation behavior of
these dendronized proteins[22] and the specificity of the
boronic acid/salicyl hydroxamate interaction agree with
previous reports.[16]
In addition to conferring pH-dependent “assemble and
release” properties to the modified enzymes, these positively
charged dendrons facilitate cellular uptake through electro-
static interactions with the negatively charged cell membrane.
The uptake proceeds by means of clathrin-dependent endo-
cytosis and intracellular release is driven by the proton sponge
effect as described previously.[14,23] Consequently, these modi-
fied proteases showed efficient dendron-mediated uptake
which was observed by confocal microscopy (Figure 2). As
the dissociation of the dendritic shell directly relied on low
intracellular pH, localization of these constructs in acidic
compartments was necessary for the release of the active
protease. Colocalization studies using LysoTracker, which is
fluorescent in acidic organelles, directly showed that the
dendronized proteins were indeed located within these
cellular compartments. In contrast, the nondendronized
proteins displayed minimal cellular uptake.
Following cellular entry facilitated by the dendritic shell,
these proteases could become active when the attached
dendrons dissociate, directly causing rampant proteolytic
degradation and cell death. Hence, upon addition of these
dendronized proteases into A549 cells, the impact on cell
viability and their potency was immediately observed after 3 h
of incubation at 378C and 5% CO2. As controls, the
326
www.angewandte.org
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2014, 53, 324 –328

Angewandte
Chemie
in the intracellular trans-
portation but also as
a benign steric moiety.
In conclusion, we
have presented the syn-
thesis of
demanding
a
sterically
dendritic
supramolecular protect-
ing group, which is
responsive to changes
between
pH 5.0, by utilizing a bio-
orthogonal ligation
pH 7.4
and
method based on boronic
acid/salicyl hydroxamate
complexation. The den-
dritic protecting group
shields the large binding
sites of macromolecules
in a highly reversible fash-
ion. In addition, the for-
mation of
a dendritic
PAMAM shell imparts
the capability for efficient
membrane translocation
and localization into the
acidic intracellular com-
partments. These attri-
butes are of high signifi-
cance since the assembly
and disassembly of the
Figure 2. Confocal laser scanning microscopy using A549 cell lines. Colocalization of dendronized proteins
(using Dylight647-conjugated proteins) performed with LysoTracker. The control observation is representative of
nondendronized HSA, trypsin, and papain. Scale bar: 20 mm.
respective native proteases, covalently dendronized G2-HSA
and dendron 5 were used for comparison (Figure 3). The non-
cytotoxic behavior of both covalently and noncovalently
dendronized HSA at much higher concentrations implied that
the observed toxicity of the dendronized proteases cannot be
attributed to the dendrons themselves or the cationic
dendron–protein hybrid structure. These observations sug-
gested that the dendritic shell not only plays an important role
dendritic shell occurs at physiological conditions, conferring
stability at near-neutral pH and release inside acidic cellular
lysosomes. The resulting active proteases in the cell severely
decreased the cell viability. In perspective, this integrated
bioorthogonal dendritic assembly provides a contemporary
dimension in creating highly efficient hybrid zymogens as
smart protein therapeutics.
Received: September 30, 2013
Keywords: dendrimers · enzymes · supramolecular chemistry
.
[1] P. Plainkum, S. M. Fuchs, S. Wiyakrutta, R. T. Raines, Nat. Struct.
Mol. Biol. 2003, 10, 115.
[2] A. V. Karginov, F. Ding, P. Kota, N. V. Dokholyan, K. M. Hahn,
Nat. Biotechnol. 2010, 28, 743.
[3] C. Lꢀpez-Otꢁn, L. M. Matrisian, Nat. Rev. Cancer 2007, 7, 800.
[4] E. M. Sletten, C. R. Bertozzi, Angew. Chem. 2009, 121, 7108;
Angew. Chem. Int. Ed. 2009, 48, 6974.
[5] Z. Ding, C. J. Long, Y. Hayashi, E. V. Bulmus, A. S. Hoffman,
P. S. Stayton, Bioconjugate Chem. 1999, 10, 395.
[6] P. De, M. Li, S. R. Gondi, B. S. Sumerlin, J. Am. Chem. Soc. 2008,
130, 11288.
[7] P. S. Stayton, T. Shimoboji, C. Long, A. Chilkoti, G. Ghen, J. M.
Harris, A. S. Hoffman, Nature 1995, 378, 472.
Figure 3. Comparison of the cytotoxicities of native proteases and their
dendronized counterparts. Cell viability was referenced against blank
controls. Data expressed as mean ÆSEM, n=3. [a] G2-HSA was
constructed by covalently conjugating 32 G2 PAMAM dendrons onto
human serum albumin as previously reported.[22]
[8] W. Gao, W. Liu, T. Christensen, M. R. Zalutsky, A. Chilkoti,
Proc. Natl. Acad. Sci. USA 2010, 107, 16432.
Angew. Chem. Int. Ed. 2014, 53, 324 –328
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
327

.
Angewandte
Communications
[9] M. Yan, J. Du, Z. Gu, M. Liang, Y. Hu, W. Zhang, S. Priceman, L.
Wu, Z. H. Zhou, Z. Liu, T. Segura, Y. Tang, Y. Lu, Nat.
Nanotechnol. 2010, 5, 48.
[10] A. A. Bastian, A. Marcozzi, A. Herrmann, Nat. Chem. 2012, 4,
789.
[11] D. Astruc, E. Boisselier, C. Ornelas, Chem. Rev. 2010, 110, 1857.
[12] L. Albertazzi, B. Storti, L. Marchetti, F. Beltram, J. Am. Chem.
Soc. 2010, 132, 18158.
[13] S. K. Hamilton, E. Harth, ACS Nano 2009, 3, 402.
[14] D. Y. W. Ng, J. Fahrer, Y. Wu, K. Eisele, S. L. Kuan, H. Barth, T.
Weil, Adv. Healthcare. Mater. 2013, DOI: 10.1002/
adhm.201200419.
[16] S. B. Y. Shin, R. D. Almeida, G. Gerona-Navarro, C. Bracken,
S. R. Jaffrey, Chem. Biol. 2010, 17, 1171.
[17] C. Q. Pan, J. S. Uumer, A. Herzka, R. A. Lazarus, Protein Sci.
1998, 7, 628.
[18] P. Carter, J. A. Wells, Nature 1988, 332, 564.
[19] H. A. Chapman, R. J. Riese, G.-P. Shi, Annu. Rev. Physiol. 1997,
59, 63.
[20] G. Springsteen, B. Wang, Chem. Commun. 2001, 1608.
[21] G. Springsteen, B. Wang, Tetrahedron 2002, 58, 5291.
[22] S. L. Kuan, B. Stçckle, J. Reichenwallner, D. Y. W. Ng, Y. Wu, M.
Doroshenko, K. Koynov, D. Hinderberger, K. Mꢂllen, T. Weil,
Biomacromolecules 2013, 14, 367.
[15] M. L. Stolowitz, C. Ahlem, K. A. Hughes, R. J. Kaiser, E. A.
Kesicki, G. Li, K. P. Lund, S. M. Torkelson, J. P. Wiley, Biocon-
jugate Chem. 2001, 12, 229.
[23] L. Albertazzi, M. Serresi, A. Albanese, F. Beltram, Mol. Pharm.
2010, 7, 680.
328
www.angewandte.org
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2014, 53, 324 –328