Development and Optimization of Glaser-Hay Bioconjugations

Article abstract of DOI:10.1002/anie.201502676

The prevalence of bioconjugates in the biomedical sciences necessitates the development of novel mechanisms to facilitate their preparation. Towards this end, the translation of the Glaser-Hay coupling to an aqueous environment is examined, and its potent

Full text of DOI:10.1002/anie.201502676

Angewandte
Chemie
International Edition: DOI: 10.1002/anie.201502676
German Edition: DOI: 10.1002/ange.201502676
Bioconjugation
Development and Optimization of Glaser–Hay Bioconjugations**
Jessica S. Lampkowski, Jordan K. Villa, Travis S. Young, and Douglas D. Young*
Abstract: The prevalence of bioconjugates in the biomedical
sciences necessitates the development of novel mechanisms to
facilitate their preparation. Towards this end, the translation of
the Glaser–Hay coupling to an aqueous environment is
examined, and its potential as a bioorthogonal conjugation
reaction is demonstrated. This optimized, novel, and aqueous
Glaser–Hay reaction is applied towards the development of
bioconjugates utilizing protein expressed with an alkynyl
unnatural amino acid. Unnatural amino acid technology
provides a degree of bioorthognality and specificity not feasible
with other methods. Moreover, the scope of the reaction is
demonstrated through protein–small molecule couplings,
small-molecule–solid-support couplings, and protein–solid-
support immobilizations.
bioconjugation strategies are necessary to maximize the
utility of bioconjugates.
While bioconjugations dependent on the natural chemical
functionality in proteins have found usefulness, they are often
associated with a lack of control over both the number of
conjugation sites and their location.[9] Thus, the introduction
of bioorthogonal handles to proteins proffers an additional
degree of control over the reaction. One mechanism that is
extremely effective in the introduction of novel functionality
is through site-specific incorporation of unnatural amino acids
(UAAs). Utilizing an evolved aminoacyl-tRNA synthetase
(aaRS)/tRNA cognate pair, a plethora of UAAs have been
introduced site-specifically into proteins in response to
a mutated TAG stop codon within the mRNA.[10] This
method has been employed in numerous applications; how-
ever, it is perhaps most useful in the installation of a bio-
orthogonal handle for bioconjugations.[11] A recent example
includes employing a Sonogashira reaction to couple an
alkyne UAA within a protein to an aryl halide.[5,12]
Specifically, we became interested in applying the Glaser–
Hay coupling of terminal alkynes as a novel biochemical
conjugation strategy (Figure 1A).[13] The Glaser–Hay reac-
tion affords an ideal conjugation strategy as it confers the
formation of a highly stable and rigid carbon–carbon bond,
can be conducted under mild conditions, does not utilize
potentially photosensitive azides, and confers a geometrically
well-defined linear conjugate. Moreover, the reagents/cata-
lysts are cost-efficient, numerous alkyne linkers and con-
jugation partners are commercially available, and the product
is a highly oxidized diyne capable of numerous additional
reactions.
To assess the feasibility of employing a Glaser–Hay
reaction towards bioconjugations, the compatibility of the
reaction to aqueous conditions had to be assessed. To our
knowledge a Glaser–Hay reaction has not previously been
reported in an aqueous solution. Consequently, proof-of-
concept couplings were conducted based on previously
optimized Glaser–Hay conditions in organic solvents involv-
ing the homodimerizaton of either phenylacylene or prop-
argyl alcohol. Gratifyingly, the reaction proceeds to comple-
tion in an aqueous solvent after 16 h at room temperature
with CuI/tetramethylethylenediamine (TMEDA) in greater
than 95% yields.
B
ioconjugates have found a wide degree of relevance as
both diagnostics and therapeutics, along with functional
materials.[1,2] Consequently, the development of novel meth-
ods for their preparation has far-reaching applications.
Various protein bioconjugation reactions are well-known
and regularly employed, including reactions involving mala-
mides, isocyanates, NHS esters, and iodoacetamides with
nucleophilic amino acid residues.[3] Moreover, novel chemical
functionalities have been introduced to further the technol-
ogy and afford 1,3-dipolar cycloadditions, photo-cross-link-
ing, transition-metal-mediated alkyne couplings, and oxime
formations.[4,5] These reactions all must meet several key
requirements to be useful in the generation of bioconjugates.
These include compatibility with aqueous media, a physiolog-
ically stable interaction between the two coupling partners,
biologically compatible and mild reaction conditions, and
ideally a degree of chemoselectivity.[1,6] Arguably the most
widely utilized bioconjugation reaction is the Husigen cycli-
zation reaction involving an alkyne and azide to yield a highly
stable triazole linker.[7] This reaction has been employed in
the preparation of a wide range of bioconjugates and a range
of variant reactions have been developed in the absence of
catalysts to increase its biocompatibility.[8] Despite the robust
nature of this reaction, further investigations into other
[*] J. S. Lampkowski, J. K. Villa, Dr. D. D. Young
Department of Chemistry, College of William & Mary
P.O. Box 8795, Williamsburg, VA 23187 (USA)
E-mail: dyoung01@wm.edu
To utilize this reaction within the context of a protein, an
alkyne moiety must be introduced into the protein as
a bioconjugation handle. Conveniently an aaRS that recog-
nizes propargyloxyphenylalanine (1) has already been
evolved, and a convenient synthesis of the alkynyl UAA is
known (Figure 1B).[14] Based on the fluorescent properties of
green fluorescent protein (GFP), it is an ideal protein for
optimization of the bioconjugation reaction. As such, a TAG
codon was introduced at residue 151, which has previously
Dr. T. S. Young
California Institute of Biomedical Research
11119 N. Torrey Pines Rd, La Jolla, CA 92037 (USA)
[**] We would like to acknowledge the Jeffress Memorial Trust and the
College of William & Mary for financial support. We would also like
to thank Dr. Peter Schultz at The Scripps Research Institute for the
pEVOL-pPrF plasmid.
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/anie.201502676.
Angew. Chem. Int. Ed. 2015, 54, 9343 –9346
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
9343

.
Angewandte
Communications
Figure 2. Glaser–Hay bioconjugation at both 378C and 48C. A) SDS-
PAGE analysis after Coomassie staining showing the successful
incorporation of pPrF into GFP due to presence of a GFP band in
lane 2 and lack of GFP in lane 3. B) SDS-PAGE analysis of the identical
gel imaged for fluorescence (Ex 280 nm/Em 512) prior to Coomasie
staining. The gel indicates the successful Glaser–Hay coupling as
fluorescence is present in lanes 2 and 4 indicating the presence of the
fluorophore, while the lack of fluorescence in lanes 6 indicate no
attached fluorophore in the absence of CuI/TMEDA despite the
presence of fluorophore in the reaction.
Figure 1. Glaser–Hay reaction and components. A) Standard Glaser–
Hay reaction employing catalytic copper to couple a terminal alkyne
moiety. B) Propargyloxyphenylalanine (pPrF) unnatural amino acid
incorporated into proteins for Glaser–Hay bioconjugations. C) Crystal
structure of GFP indicating residue 151 targeted for alkyne introduc-
tion on the b-barrel, adapted from PDB: 1EMA.[16]
4). Control reactions in the absence of CuI and TMEDA
demonstrated the presence of protein, but no fluorescent
label (Figure 2; Lane 6). Moreover, couplings performed with
the wild-type GFP and the fluorophore resulted in no
fluorescent signal on a gel. This confirms that the labeling is
due to the Glaser–Hay coupling and not non-covalent
interactions of the protein with the fluorophore. To validate
this comparison, the GFP was also coupled to a SRFluor-680
alkyne under identical conditions. The absorbance at 649 nm
been demonstrated to be an ideal site for UAA incorporation
as it is a surface-exposed position on the rigid b-barrel and
does not impact GFP fluorescence (Figure 1C).[15] BL-21-
(DE3) E. coli were co-transformed with the pEVOL-pPrF
and pET-GFP-Y151TAG plasmids, and GFP was expressed in
the presence and absence of 1 (Figure 2; Lanes 6 and 7). The
lack of full-length protein in the absence of 1 corresponds to
the termination of translation at the TAG codon. With a GFP-
pPrF-151 mutant in hand, Glaser–Hay optimization could
proceed.
Initial bioconjugations were attempted between the
mutant GFP and an AlexaFluor-488 modified alkyne. The
Cu/TMEDA was first prepared by mixing CuI with TMEDA
(ca. 50 equiv) in water for 10 min at room temperature. The
fluorophore was then added in approximately ten-fold excess
followed by the GFP and the reaction was incubated at either
378C or 48C. While some fluorophore dimerization does
occur, this product, the CuI/TMEDA, and excess fluorophore
were easily removed by centrifugal concentration accompa-
nied by washings in PBS, and the reactions were analyzed by
SDS-PAGE and LC/MS (Figure 2). Gel electrophoresis
demonstrated successful Glaser–Hay coupling when compar-
ing the identical gel under fluorescent imaging (Figure 2B)
and Coomassie staining (Figure 2A). In the presence of the
CuI/TMEDA and an alkyne fluorophore, fluorescent labeling
of the protein was observed at both temperatures, with 48C
affording slightly higher fluorescence (Figure 2; Lanes 2 and
that
corresponds
to
the
fluorophore
(e =
257800 LmolÀ1 cmÀ1) was then normalized to the GFP
absorbance, affording coupling yields of 71% after 4 h at
48C, and 93% after 6 h at 48C (see the Supporting Informa-
tion). These were roughly in agreement with what was
observed from the PAGE results. Moreover, the successful
coupling was confirmed by MS analysis (see the Supporting
Information).
Extended couplings led to decreased fluorescence and
overall lower amounts of GFP upon staining. We hypothe-
sized that this may be a result of CuI oxidative degradation of
the protein over time. To fully optimize the reaction, time
courses were performed, sampling the reaction at different
time points and comparing the ratios of intact GFP with
fluorescence (Figure 3). Based on these results, 6 h at 48C
appears to be the optimal conditions for the Glaser–Hay
bioconjugation. Additionally, reduction of the CuI/TMEDA
concentrations, even by a factor of five, resulted in no
observable coupling. Attempts to accelerate the reaction by
increasing oxygen concentrations, either by bubbling air
through the reaction or by leaving the Eppendorf tubes
9344 www.angewandte.org
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2015, 54, 9343 –9346

Angewandte
Chemie
orientation and coupling of the two terminal alkyne moieties
(see the Supporting Information). Even when protein con-
centrations were doubled in the reaction, dimerization levels
did not increase.
To probe the utility of this novel bioconjugation, we next
set out to examine its relevance towards solid-supported
reactions. To rapidly assess the method, a propargyl alcohol or
a 5-hexyn-1-ol derivitized Sepharose 6B resin was reacted
under similar conditions with the AlexaFluor 488 alkyne.
Control reactions were performed in the absence of both the
fluorophore and the CuI/TMEDA. After several PBS washes,
the fluorescence of the resin was assessed, and fluorescence
was only detected in the presence of both fluorophore and
copper (see the Supporting Information). Based on previous
results using UAAs for protein immobilization, attempts were
made to translate this reaction to the immobilization of
GFP.[15]
A GFP-pPrF-151 mutant was expressed and utilized to
investigate the ability to conjugate the protein to a solid
support via a Glaser–Hay reaction. Using the optimized
conditions, the GFP was immobilized on both a propargyl
alcohol and 1-hexynol resins with similar effect (Figure 4).
Control reactions in the absence of either GFP or CuI/
TMEDA yielded little to no detectable fluorescence on the
Figure 3. Glaser–Hay time course. Ratio of total protein to fluores-
cence based on densiometry measurements obtained by SDS-PAGE.
Gels were first imaged for fluorescence to ascertain the relative
amounts of fluorophore conjugate, then stained to assess the concen-
tration of total GFP protein. Reactions at each time point were
conducted in triplicate to generate error bars indicating standard
deviation. Results indicate highest coupling and optimum reaction
conditions at 48C for 6 h. Extended periods of time were found to lead
to protein degradation.
open, did not afford the desired effect, as increased protein
degradation was observed. This suggests that at these low
concentrations of substrate, the CuI/TMEDA is not catalytic,
unlike previously observed Glaser–Hay reactions on larger
scales.
The Glaser–Hay reaction time is comparable to reported
azide–alkyne click couplings, and slightly slower than the
previously reported Sonogashira couplings that ranged from
0.5–2 h.[5] However, advantages of the Glaser–Hay approach
relative to the Sonogashira is the use of substantially cheaper
and more common reagents than palladium catalysts, as well
as lower fluorophore equivalents (10 equiv, compared to
50 equiv in the Sonogashira).[12] Interestingly, the Glaser–Hay
rate was comparable between the 48C conditions and the
378C conditions. This may be due to the increased protein
degradation observed at the higher temperature, limiting
reaction rates via divergent pathways. It appears that as long
as the reaction is quenched by the 6 h time point, the
oxidative damage is minimal, as confirmed by the mass
spectrum of the protein subjected to reaction conditions in the
absence of fluorophore remaining identical to protein not
subjected to the Glaser–Hay conditions (see the Supporting
Information). Oxidative damage was observed by MS after
extended reaction times of 10 h. Additionally, based on this
result, bioorthogonality was confirmed as no cysteine cross-
linking or thiol–yne additions were detected. While some
chemoselectivity issues exist, as a consequence of the Glaser–
Hay mechanism, these reactions are minimal, or easily
addressed by purification. As previously noted, fluorophore
dimerization is easily removed by size-exclusion centrifuga-
tion, and GFP dimerization occurs in approximately 5–13%
yield as determined by analysis of PAGE protein band
intensities (Figure 2 and the Supporting Information).
Attempts to enhance and analyze dimerization by reaction
of the pPrF-GFP mutant under optimized conditions in the
absence of the fluorophore failed to produce any increased
GFP dimer, and occurred at less than 10% yield. This is most
likely due to the steric bulk of the protein hindering proper
Figure 4. Protein-resin Glaser–Hay immobilization. A) Reaction includ-
ing GFP-pPrF-151 and alkyne derivatized Sepharose 6B resin. Strong
fluorescence in the presence of both CuI/TMEDA and GFP-pPrF-151
indicate a successful coupling. Lack of fluorescence in control reac-
tions omitting either protein or copper indicate no coupling occurred.
B) Fluorescence data of completed reactions with both propargyl
alcohol and hexynol loaded Sepharose resins. Controls with no CuI/
TMEDA indicate low background fluorescence, while GFP-pPrF-151
protein reacted with the CuI/TMEDA and resin displays strong
fluorescence. Standard deviations depicted in error bars from triplicate
couplings.
Angew. Chem. Int. Ed. 2015, 54, 9343 –9346
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
9345

.
Angewandte
Communications
[2] G. Pasut, F. Veronese, R. SatchiFainaro, R. Duncan, Polym.
Therap. I 2006, 192, 95 – 134; J. Kalia, R. Raines, Curr. Org.
Chem. 2010, 14, 138 – 147; K. Boeneman, J. Deschamps, S.
Buckhout-White, D. Prasuhn, J. Blanco-Canosa, P. Dawson, M.
Stewart, K. Susumu, E. Goldman, M. Ancona, I. Medintz, ACS
Nano 2010, 4, 7253 – 7266; K. Thomas, D. Sherman, T. Amiss, S.
Andaluz, J. Pitner, Diabetes Technol. Ther. 2006, 8, 261 – 268; C.
Niemeyer, Angew. Chem. Int. Ed. 2001, 40, 4128 – 4158; Angew.
Chem. 2001, 113, 4254 – 4287; T. Allen, Nat. Rev. Cancer 2002, 2,
750 – 763.
[3] S. Ghosh, P. Kao, A. Mccue, H. Chappelle, Bioconjugate Chem.
1990, 1, 71 – 76; M. E. Annunziato, U. S. Patel, M. Ranade, P. S.
Palumbo, Bioconjugate Chem. 1993, 4, 212 – 218; I. Johnson,
Histochem. J. 1998, 30, 123 – 140; K. Lo, J. Lau, D. Ng, N. Zhu, J.
Chem. Soc. Dalton Trans. 2002, 1753 – 1756; N. Stephanopoulos,
M. Francis, Nat. Chem. Biol. 2011, 7, 876 – 884; M. A. Gauthier,
H. A. Klok, Chem. Commun. 2008, 2591 – 2611.
[4] A. Dirksen, P. Dawson, Bioconjugate Chem. 2008, 19, 2543 –
2548; H. Kolb, K. Sharpless, Drug Discovery Today 2003, 8,
1128 – 1137; J. Baskin, C. Bertozzi, QSAR Comb. Sci. 2007, 26,
1211 – 1219; C. Chou, R. Uprety, L. Davis, J. Chin, A. Deiters,
Chem. Sci. 2011, 2, 480 – 483; H. Dibowski, F. Schmidtchen,
Angew. Chem. Int. Ed. 1998, 37, 476 – 478; Angew. Chem. 1998,
110, 487 – 489; J. Chalker, C. Wood, B. Davis, J. Am. Chem. Soc.
2009, 131, 16346; B. Hutchins, S. Kazane, K. Staflin, J. Forsyth, B.
Felding-Habermann, P. Schultz, V. Smider, J. Mol. Biol. 2011,
406, 595 – 603.
resin; however, significant immobilization was detected when
the resin was reacted in the presence of both the GFP mutant
and the CuI/TMEDA for 6 h at 48C. Moreover, controls with
wild-type GFP under similar reaction conditions did not
result in resin immobilization (see the Supporting Informa-
tion). Thus, no non-covalent interactions between the protein
and resin are occurring and the fluorescence can be attributed
to the successful Glaser–Hay bioconjugation.
In conclusion, along with the transfer of the Glaser–Hay
reaction to an aqueous environment, a novel bioconjugation
has been developed utilizing the reaction and unnatural
amino acid technologies. The generation of a new linear
carbon–carbon bond affords a highly stable linkage utilizing
mild reaction conditions and photochemically inert starting
materials. This bioconjugation has been applied to small
molecule dimerizations, protein–fluorophore conjugations,
fluorophore–resin conjugations, and protein–resin immobili-
zations, demonstrating the breadth of the reaction substrates
and applicability. Future work is underway to expand the
scope of the conjugations to biologically relevant applications
and therapeutic agents. This technology represents a viable
alternative to click conjugations and has the potential to find
widespread utility.
[5] N. Li, R. Lim, S. Edwardraja, Q. Lin, J. Am. Chem. Soc. 2011,
133, 15316 – 15319.
[6] E. Sletten, C. Bertozzi, Angew. Chem. Int. Ed. 2009, 48, 6974 –
6998; Angew. Chem. 2009, 121, 7108 – 7133.
[7] J. C. Jewett, C. R. Bertozzi, Chem. Soc. Rev. 2010, 39, 1272 –
1279; K. Nwe, M. W. Brechbiel, Cancer Biother. Radiopharm.
2009, 24, 289 – 302.
[8] J. F. Lutz, Angew. Chem. Int. Ed. 2008, 47, 2182 – 2184; Angew.
Chem. 2008, 120, 2212 – 2214.
[9] L. Presta, Adv. Drug Delivery Rev. 2006, 58, 640 – 656; A. Wu, P.
Senter, Nat. Biotechnol. 2005, 23, 1137 – 1146; S. Smith, Curr.
Opin. Mol. Ther. 2004, 6, 666 – 674; L. Wang, G. Amphlett, W.
Blattler, J. Lambert, W. Zhang, Protein Sci. 2005, 14, 2436 – 2446.
[10] L. Wang, P. G. Schultz, Angew. Chem. Int. Ed. 2005, 44, 34 – 66;
Angew. Chem. 2005, 117, 34 – 68; T. S. Young, P. G. Schultz, J.
Biol. Chem. 2010, 285, 11039 – 11044; C. Liu, P. Schultz, R.
Kornberg, C. Raetz, J. Rothman, J. Thorner, Annu. Rev.
Biochem. 2010, 79, 413 – 444.
[11] K. Lang, J. Chin, Chem. Rev. 2014, 114, 4764 – 4806.
[12] N. Li, C. Ramil, R. Lim, Q. Lin, ACS Chem. Biol. 2015, 10, 379 –
384.
[13] P. Siemsen, R. Livingston, F. Diederich, Angew. Chem. Int. Ed.
2000, 39, 2632 – 2657; Angew. Chem. 2000, 112, 2740 – 2767; C.
Glaser, Ber. Dtsch. Chem. Ges. 1869, 2, 422 – 424; A. Hay, J. Org.
Chem. 1962, 27, 3320 – 3321; V. Tripp, J. Lampkowski, R. Tyler,
D. Young, ACS Comb. Sci. 2014, 16, 164 – 167; J. S. Lampkowski,
C. E. Durham, M. S. Padilla, D. D. Young, Org. Biomol. Chem.
2015, 13, 424 – 427.
Experimental Section
Glaser–Hay aqueous conditions: Aqueous conditions for the Glaser–
Hay reaction were optimized by preparing a phenylacetlyene homo-
dimer. TMEDA (10 mL, 0.06 mmol) and CuI (10 mg, 0.05 mmol) were
added to a vial containing H2O (3 mL), forming the copper complex.
Phenylacetylene (37 mg, 0.364 mmol) was then added and the
reaction was allowed to stir at room temperature for 16 h. The
reaction was extracted using EtOAc and H2O washes (4 ” 5 mL),
concentrated, and dried in vacuo. The product was obtained as
a white solid: 88 mg, 0.349 mmol, 96% yield; 1H NMR (400 MHz;
CDCl3): d = 7.38–7.56 ppm (m, J = 7.2 Hz, 10H).
Protein–fluorophore Glaser–Hay bioconjugation: The expressed
GFP-pPrF-151 was coupled to AlexaFluor 488 alkyne using Glaser–
Hay reaction conditions. In an Eppendorf tube, CuI (5 mL, 500 mm)
and TMEDA (5 mL, 500 mm) were mixed and equilibrated at 378C.
After 10 min, AlexaFluor 488 alkyne (10 mL, 1 mm) was added and
equilibrated at 378C for 10 min. Finally, GFP-pPrF-151 (20 mL,
0.5 mgmLÀ1) was added. A control reaction was also prepared with
the same concentrations of fluorophore and protein, but with the CuI/
TMEDA replaced with PBS buffer (10 mL). The reactions were
incubated for various times at 378C or 48C. Reactions were then
purified through centrifugal concentration on Spin-X UF colums
(Corning), with wash cycles of PBS buffer (5 ” 100 mL) until flow-
through was free of fluorophore. The protein was then analyzed by
SDS-PAGE gel to verify coupling of the fluorophore to the protein.
Time-course experiments were analyzed by comparing densitometry
of fluorescent bands to their Coomasie-stained bands using a BioRad
Molecular Imager Gel Doc XR + system.
[14] A. Deiters, P. G. Schultz, Bioorg. Med. Chem. Lett. 2005, 15,
1521 – 1524.
[15] B. K. Raliski, C. A. Howard, D. D. Young, Bioconjugate Chem.
2014, 25, 1916 – 1920.
[16] M. Ormç, A. B. Cubitt, K. Kallio, L. A. Gross, R. Y. Tsien, S. J.
Remington, Science 1996, 273, 1392 – 1395.
Keywords: alkynes · bioconjugation · Glaser–Hay ·
solid supports · unnatural amino acids
How to cite: Angew. Chem. Int. Ed. 2015, 54, 9343–9346
Angew. Chem. 2015, 127, 9475–9478
Received: March 23, 2015
Revised: April 24, 2015
Published online: June 18, 2015
[1] G. T. Hermanson, Bioconjugate techniques, 3rd ed., Academic
Press, London, 1996.
9346 www.angewandte.org
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2015, 54, 9343 –9346