Dendronylation: Residue-specific chemoselective attachment of
oligoglycerol dendrimers on proteins with noncanonical amino acids
Ying Ma
a
, Bala N. S. Thota
b
, Rainer Haag
b,
⇑
, Nediljko Budisa
a,
⇑
a
Technische Universität Berlin, Institut für Chemie, Müller-Breslau Str. 10, 10623 Berlin, Germany
b
Freie Universität Berlin, Berlin, Institut für Chemie und Biochemie, Takustrasse 3, 14195 Berlin, Germany
article info
Article history:
Received 21 August 2015
Revised 19 September 2015
Accepted 23 September 2015
Available online 25 September 2015
Keywords:
Click-chemistry
Noncanonical amino acids
Oligoglycerol dendrons
Protein–dendron conjugate
Unnatural protein expression
abstract
Polyglycerol dendrimers as an important class of polymeric materials especially attractive for covalent
attachment to therapeutic proteins as a useful alternative to traditional PEGylation procedures. Herein,
we combine in vivo noncanonical amino acid (ncAA) incorporation and chemoselective conjugation
in vitro to produce novel hybrid protein–dendrimer conjugates with the defined architectures. We incor-
porated Azidohomoalanine (Aha) as methionine substitute in vivo into various protein scaffolds to allow
non-invasive dendrimer conjugations (dendronylation). Our approach makes recombinant proteins
accessible for the design of multivalent dendrimer conjugates since it enables the preparation of many
sequences with various positions for regioselective dendronylation.
Ó2015 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY license
(http://creativecommons.org/licenses/by/4.0/).
Polyglycerol dendrimers as nontoxic, non-immunogenic and
highly soluble in water are especially attractive polymers for
covalent attachment to therapeutic proteins as useful alternative
to traditional PEGylation procedures.
1–3
Traditional approaches
for chemical posttranslational modification by derivatization of
solvent-accessible reactive side chains such as those of canonical
amino acids lysine or cysteine often lead to heterogeneously
labeled mixtures of modified proteins.
4
Thus, selective methods
that allow absolute control of the type and position of polymer
attachment within a protein are required. To achieve this goal
we use reprogrammed ribosomal synthesis to install a bioorthogo-
nal azide functional group at a pre-selected site, which then can be
chemoselectively modified in a way that homogeneous, struc-
turally defined conjugates are generated.
5
In particular, we use effi-
cient auxotrophy-based residue-specific methods to introduce
azidohomoalanine (Aha) at methionine (Met) positions into
engineered proteins
6
which are subsequently subjected to
Cu-catalyzed azide-alkyne cycloaddition (CuAAC) reactions
7
with
alkyne-functionalized oligoglycerol dendron propargyl-generation
2 (dOG) to obtain homogeneous protein conjugates.
We have chosen cysteine-free barstar
8
with only one methionine
exclusively at N-terminal (Nt) position (Met1; denoted as B
⁄
). In
addition, we have engineered double mutant with two methionine
side chains (Met1 + Glu47Met; denoted as B
⁄⁄
). Barstar is well suited
as a model protein due to its high stability and small and simple
structure of 90 amino acids. Barstar is the intercellular inhibition
of barnase and is widely used to study protein structure, stability
and folding.
9
Its overall protein structure consists of a b
a
bmotif
which is characteristic for some nucleic acid binding proteins. In
the present study, we have introduced Aha in both protein variants.
Other target protein used to attach more than two dendron mole-
cules to one scaffold is green fluorescent protein (GFP) engineered
to contain three solvent-exposed Met residues at positions 50, 134
and 143 engineered by site directed mutagenesis as described else-
where.
10
The structural and functional integrity of all resulting pro-
tein-conjugates is checked by far-UV CD spectroscopy (B
⁄
/B
⁄⁄
)
activity assay. Autofluorescence is the main feature of structural
and functional integrity of GFP as a chromophore in denatured form
is efficiently quenched by water as solvent.
11
Therefore, GFP with its
high stability and intrinsic fluorescence properties
12,13
of the b-bar-
rel structure makes it an almost ideal protein scaffold for the rational
decoration with multiple dOGs.
Protected propargylated [G2] dendron (Fig. 1A) was synthesized
as previously reported in the literature.
14
The dendron was depro-
tected by heating a methanolic solution with Dowex-H (2 eq. by
weight) at 65 °C for 5–6 h. The reaction mixture was filtered and
purified by reverse phase HPLC using 40% water/methanol, which
resulted pure compound as colorless viscous liquid (yield-90%).
For a high-expression of the target proteins, we choose methionine
auxotrophic Escherichia coli B834 (DE3) as the expression host.
http://dx.doi.org/10.1016/j.bmcl.2015.09.055
0960-894X/Ó2015 The Authors. Published by Elsevier Ltd.
This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/).
⇑
Corresponding authors. Tel.: +49 30 838 52633; fax: +49 30 838 53357 (R.H.);
tel.: +49 30 314 28821; fax: +49 30 314 28279 (N.B.).
(N. Budisa).
Bioorganic & Medicinal Chemistry Letters 25 (2015) 5247–5249
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Genes for both barstar variants (Met1; Met1 + Glu47Met) are car-
ried by vector pQE80L with the inducible T5 promoter. Similarly,
GFP is harbored by vector pET30b under the inducible T7
promoter.
The general approach of labeled B
⁄
,B
⁄⁄
, GFP expression is a fer-
mentation procedure based on the selective pressure incorporation
(SPI) method.
15
First, the E. coli B834 cells have been cultured in
full new minimum media (NMM)
16
containing 19 canonical amino
acids and a limited Met concentration (0.06 mM). Cells are grown
until Met is fully depleted from the growth medium, which usually
takes place in mid-log growth phase with optical density values at
600 nm about 0.8–1.0. In the second phase, 100 mg L
1
L
-Aha is
added followed by induction of the target protein synthesis by Iso-
propyl b-
D
-1-thiogalactopyranoside (IPTG) which results in global
Met ?Aha substitution during translation, including the target
protein. Under these conditions, azide-labeled B
⁄
(B
⁄
[Aha]) was
expressed in yields of about 50% (7.36 mg L
1
) compared to
native B
⁄
(14.1 mg L
1
). Similarly, azide-labeled double mutant
B
⁄⁄
(B
⁄⁄
[Aha]) was expressed in yields of about 40% (5.2 mg L
1
)
compared to native B
⁄⁄
(12.8 mg L
1
). Moreover, this fermentation
procedure yielded 2.58 mg L
1
GFP[Aha] which is about 30% com-
pared to wild-type protein (9.25 mg L
1
). High Met replacement
levels by Aha, containing no traces of the parent protein as contam-
inant, were further confirmed by electrospray mass-spectrometric
analysis (ESI-MS) as shown in Figure 1 and Table S1.
CuAAC reactions of azide-containing proteins with dOG were
carried out in an aqueous phosphate buffer with copper(I),
aminoguanidine and
L
-ascorbic acid.
17
Gel electrophoresis mobi-
lites and mass spectrometric analyses of the obtained dendron-
conjugates revealed full agreement between expected and found
masses as shown in Figure 1 and Table S2. In general, the respec-
tive Aha-containing species was not detectable by ESI-MS, indicat-
ing a high level of dendronylation after the CuAAC reaction.
The structural dependence of GFP fluorescence clearly indicates
that the protein is not structurally compromised after conjugation
with multiple dendrons. Moreover, t he level of structural
perturbation in B
⁄
can be measured by circular dichroism and ther-
mally induced unfolding.
In our previous studies, we found that incorporation of Aha in
proteins generally does not affect their structure and function.
18
Far-UV CD analysis of B
⁄
[Aha] and related dendroproteins revealed
almost superimposable spectra (Fig. 2). Importantly, the
unchanged spectral shapes clearly indicate that the overall sec-
ondary structure is not significantly changed (within the limits of
this spectroscopic method). Small variations in dichroic intensities
between variants are most probably due to minor differences in
protein concentration estimation. Although the measurement of
thermal unfolding profiles revealed that the dendronylation leads
to slightly less stable proteins variants in terms of T
m
values that
are lowered by 2–4 °C, it can be concluded that the overall protein
scaffold is not dramatically influenced by the dOG attachments.
In conclusion, our SPI approach for incorporation of Met ana-
logue Aha in combination with site-directed mutagenesis made it
possible to introduce or eliminate Met residues and subsequently
engineer protein scaffolds with various positions for regioselective
Figure 2. Spectral profiles and melting curves of barstar variants as determined by
circular dichroism (CD). Secondary structure signals (left) and thermal unfolding
profiles (right) of native and decorated barstar variants.
Figure 1. (A) Chemical structure of propargylated oligoglycerol [G2] dendron (dOG). (B) Three-dimensional models (ribbon plots drawn by PyMOL of barstar (PDB code:
1B27) and GFP (PDB code: 4GF6) clarify with marked conjugation products (as a result of CuAAC reaction of azide-containing proteins with dOG). From left to right: B
⁄
[dOG]
1
,
B
⁄⁄
[dOG]
2
, GFP [dOG]
3
with deconvoluted ESI mass spectra after modification of both B
⁄
[Aha] and B
⁄⁄
[Aha] with 1 and 2 dendrons, respectively (list of theoretical and
experimentally found masses in SI). The absence of B
⁄
[Aha] signals in the spectra after CuAAC reaction confirms virtually quantitative conjugation. (C, D) Comparison of
electrophoretic mobilities of B
⁄
[Aha] and GFP[Aha] with their conjugates B
⁄
[dOG]
1
and B
⁄⁄
[dOG]
2
as well as GFP [dOG]
3
in Coomassie-stained 17% SDS–PAGE. Detectable
shifts upon multiple dendronylation presents a reliable qualitative indication that the corresponding protein conjugates are formed.
5248 Y. Ma et al. / Bioorg. Med. Chem. Lett. 25 (2015) 5247–5249
dendronylation. Our approach also presents a promising
alternative to the widely used suppression-based methodology
which is generally limited by low production yields and a techni-
cally complicated experimental set-up.
9,19
In this context, we pre-
sent here efficient dendronylation of recombinantly expressed
proteins by using suitably functionalized noncanonical amino
acids for chemoselective conjugation in order to produce a novel
protein conjugates with single and multiple dendrons. Non-pro-
teinogenic amino acids with bioorthogonal handles such as Aha
and its alkyne-containing counterpart Homopropargyl-glycine
(Hpg) are attractive tags having small size and being genetically
encoded. They are especially suitable for chemoselective reactions
even in living cells since they are expected not to interfere with the
protein’s innate functions and structural integrity.
20
In this con-
text, we were able to demonstrate that via CuAAC, different num-
bers of dendrons can be ligated to various azide-containing model
proteins in almost quantitative yields giving access to active homo-
geneous dendroforms. Our systematic and modular approach for
the generation of multivalent dendronylated systems by using
engineered proteins as structured scaffolds offers an attractive
alternative to the generally applied PEGylation of proteins.
Acknowledgments
Y.M. and N.B. are grateful to Dr. Patrick Durkin for providing us
with chemically synthetized
L
-Aha. We are indebted to Dr. Tobias
Baumann for critical reading of our manuscript and Richard Ber-
nitzky for his help and co-work on this project. Y.M. acknowledge
financial support from China Scholarship Council (CSC) and
Berlin International Graduate School of Natural Sciences and
Engineering (BIG-NSE). R.H. and N.B. thank the SFB 765 for finan-
cial support.
Supplementary data
Supplementary data (experimental details and supplementary
information) associated with this article can be found, in the online
version, at http://dx.doi.org/10.1016/j.bmcl.2015.09.055.
References and notes
1. Haag, R.; Sunder, A.; Stumbé, J. F. J. Am. Chem. Soc. 2000,122, 2954.
2. Meng, F.; Manjula, B. N.; Smith, P. K.; Acharya, S. A. Bioconjugate Chem. 2008,19,
1352.
3. Gupta, S.; Pfeil, J.; Kumar, S.; Poulsen, C.; Lauer, U.; Hamann, A.; Hoffmann, U.;
Haag, R. Bioconjugate Chem. 2015,26, 669.
4. Prabakaran, S.; Lippens, G.; Steen, H.; Gunawardena, J. Wiley Interdiscip. Rev.
Syst. Biol. Med. 2012,4, 567.
5. Lang, K.; Chin, J. W. ACS Chem. Biol. 2014,9,16.
6. Budisa, N. Angew. Chem., Int. Ed. 2004,43, 6426.
7. Meldal, M.; Tornøe, C. W. Chem. Rev. 2008,108, 2952.
8. Nölting, B.; Golbik, R.; Fersht, A. R. Proc. Natl. Acad. Sci. U.S.A. 1995,92, 10668.
9. Lepthien, S.; Merkel, L.; Budisa, N. Angew. Chem., Int. Ed. 2010,49, 5446.
10. Nischan, N.; Herce, H. D.; Natale, F.; Bohlke, N.; Budisa, N.; Cardoso, M. C.;
Hackenberger, C. P. R. Angew. Chem., Int. Ed. 2015,54, 1950.
11. Bae, J. H.; Rubini, M.; Jung, G.; Wiegand, G.; Seifert, M. H.; Azim, M. K.; Kim, J. S.;
Zumbusch, A.; Holak, T. A.; Moroder, L.; Huber, R.; Budisa, N. J. Mol. Biol. 2003,
328, 1071.
12. Yang, F.; Moss, L. G.; Phillips, G. N. Nat. Biotechnol. 1996,14, 1246.
13. Phillips, G. N. Curr. Opin. Struct. Biol. 1997,7, 821.
14. Thota, B. N. S.; Berlepsch, H. V.; Boettcher, C.; Haag, R. Chem. Commun. 2015,
8648.
15. Minks, C.; Alefelder, S.; Moroder, L.; Huber, R.; Budisa, N. Tetrahedron 2000,56,
9431.
16. Budisa, N.; Steipe, B.; Demange, P.; Demange, P.; Eckerskorn, C.; Kellermann, J.;
Huber, R. Eur. J. Biochem. 1995,230, 788.
17. Ma, Y.; Biava, H.; Contestabile, R.; Budisa, N.; di Salvo, M. L. Molecules 2014,19,
1004.
18. Merkel, L.; Beckmann, H. S. G.; Wittmann, V.; Budisa, N. ChemBioChem 2008,9,
1220.
19. Hoesl, M. G.; Budisa, N. ChemBioChem 2011,12, 552.
20. Dieterich, D. C.; Link, A. J.; Graumann, J.; Tirrell, D. A.; Schuman, E. M. Proc. Natl.
Acad. Sci. U.S.A. 2006,103, 9482.
Y. Ma et al. / Bioorg. Med. Chem. Lett. 25 (2015) 5247–5249 5249
