Efficient formation of heterodimers from peptides and proteins using unsymmetrical polyfluorophenyl esters of dicarboxylic acids

Article abstract of DOI:10.1002/psc.2394

An efficient method for the heteroconjugation of biomolecules carrying free amino groups was reported previously, where mixed polyfluorophenyl diesters of dicarboxylic acids with varied aliphatic chain length were shown to be efficient reagents for the conjugation of a variety of model biomolecules. The concept was based on the differential reactivity of the esters towards amines. The concept has now been further optimized, and a 2,6-difluorophenyl-pentafluorophenyl diester combination has been demonstrated to be the most efficient, both with respect to selectivity and to reaction rate. A pentafluorophenyl ester reacts faster with an amino group and requires a weaker base than a 2,6-difluorophenyl ester that requires a stronger base and longer reaction time. With the use of this combination of esters, we obtained considerably shortened reaction times compared with those reported previously, yet still retaining the desired selectivity in heteroconjugation. The increased reactivity of the bifunctional reagent allowed the construction of sophisticated peptide heteroconjugates from peptides, carbohydrates and proteins, showing a wide scope of applicability in the field of assembling functional bioconjugates.

Full text of DOI:10.1002/psc.2394

Research Article
Received: 9 November 2011
Revised: 12 December 2011
Accepted: 14 December 2011
Published online in Wiley Online Library: 5 March 2012
(wileyonlinelibrary.com) DOI 10.1002/psc.2394
Efficient formation of heterodimers
from peptides and proteins using
unsymmetrical polyfluorophenyl
esters of dicarboxylic acids
Adam T. Ślósarczyk and Lars Baltzer*
An efficient method for the heteroconjugation of biomolecules carrying free amino groups was reported previously, where
mixed polyfluorophenyl diesters of dicarboxylic acids with varied aliphatic chain length were shown to be efficient reagents
for the conjugation of a variety of model biomolecules. The concept was based on the differential reactivity of the esters
towards amines. The concept has now been further optimized, and a 2,6-difluorophenyl-pentafluorophenyl diester combination
has been demonstrated to be the most efficient, both with respect to selectivity and to reaction rate. A pentafluorophenyl ester
reacts faster with an amino group and requires a weaker base than a 2,6-difluorophenyl ester that requires a stronger base and
longer reaction time. With the use of this combination of esters, we obtained considerably shortened reaction times compared
with those reported previously, yet still retaining the desired selectivity in heteroconjugation. The increased reactivity of the
bifunctional reagent allowed the construction of sophisticated peptide heteroconjugates from peptides, carbohydrates and
proteins, showing a wide scope of applicability in the field of assembling functional bioconjugates. Copyright © 2012 European
Peptide Society and John Wiley & Sons, Ltd.
Supporting information can be found in the online version of this article.
Keywords: conjugation; bioconjugation; linker; heteroconjugation; polyfluorophenyl esters
imidodiesters [5] or N-hydroxysuccinimide diesters [6]. Although
widely used due to the availability of a variety of linker sizes,
Introduction
In recent years, bioconjugation has developed to become an
independent and rapidly developing field of research [1]. The
reagents of these types are not practical for heteroconjugate
formation because the reactivity is the same for both esters and
a statistical distribution of homoconjugates and heteroconju-
gates is obtained. Diethylsquarate [7] or thiophosgene [8], on
the other hand, can be used to link two different molecules
carrying amino groups to form heteroconjugates in high yields
because the reactions of the first and the second amino group
occur at different rates and provide selectivity in heteroconjuga-
tion. They react within minutes with the first amino group that
allows for successful formation and separation of the monoconju-
gate, which can subsequently be reacted with the second amino
group at a considerable slower rate. These reagents are very
efficient but provide no variation in linker size, which limits their
usefulness when the possibility to vary the distance between the
two heteroconjugate subunits is of interest, especially when the
heteroconjugate is designed to interact simultaneously with
two different biomolecular targets.
combination of two functional biomolecules, to form
a
heteroconjugate with the properties of both of its individual
components, has many applications in biotechnology and
biomedicine [2]. Immobilizations on solid support are ubiquitous
in chip-based in vitro diagnostics and other bioanalytical
applications. The incorporation of fluorophores, radiolabels and
other reporter groups in proteins and peptides selectively and in
high yields is
a prerequisite for molecular imaging. The
applicability of efficient conjugation techniques is vast and ranges
from fundamental biological and biochemical research to
biomedicinal applications, including medical imaging and therapy
[3]. It is fair to say that bioconjugation constitutes an indispensable
interface between synthetic organic chemistry and biology.
Bifunctional linkers are the fundaments of bioconjugate
chemistry and are essential molecular tools in biotechnology [4].
Most of the commercially available bifunctional reagents
are designed to form linkages using two different and
orthogonal functional groups that are abundant in biomolecules.
Orthogonality between the reactive ends of the reagent ensures
the formation of heteroconjugate products in two separate
steps under mild conditions.
Recently, we reported on a very simple, robust and reliable
technology to form heteroconjugates from biomolecules carrying
*
Correspondence to: Lars Baltzer, Department of Biochemistry and
Organic Chemistry, Uppsala University, PO Box 576, SE-751 23 Uppsala, Sweden.
E-mail: lars.baltzer@biorg.uu.se
It is often desirable to form a linkage between two molecules
using the same functional group, such as the primary amino
group of the side chain of lysine. The best known reagents for
amino group conjugations are based on the reactivity of
Department of Biochemistry and Organic Chemistry, Uppsala University,
PO Box 576, SE-751 23 Uppsala, Sweden
J. Pept. Sci. 2012; 18: 261–269
Copyright © 2012 European Peptide Society and John Wiley & Sons, Ltd.

ŚLÓSARCZYK AND BALTZER
free primary amino groups [9]. The strategy was based on
aliphatic dicarboxylic acid diesters where two esters were formed
from phenols with different pKa values and therefore with
different reactivities toward amino groups. A diester where the
reactivity of each of the two esters is different towards primary
amino groups provides an opportunity to react the diester with
two biomolecules on different time scales. One can separate
the initially formed monoconjugate and react it under controlled
conditions with a second biomolecule to form a heteroconjugate
in high yield. We have shown experimentally that an o-fluorophenyl
ester and a pentafluorophenyl ester separated by a linear aliphatic
chain is a convenient bifunctional reagent that allows for the
construction of conjugates from amino group containing species
such as peptides and carbohydrates. The concept provides an
opportunity to conveniently vary the distance between the
biomolecular components by selecting different diacids that are
abundantly available from commercial suppliers. To enhance the
difference in reaction rate between the esters, we performed each
reaction step under conditions that differed in basicity. As the
reaction is typically performed in dry DMSO, no problems with
solubility of the reagents or hydrolysis of the esters were observed.
Although this technology turned out to give pure conjugates, it
suffered from a very significant drawback. The time required to
complete the entire conjugation procedure was long. Although
the pentafluorophenyl ester reacted momentarily, the reaction
of the o-fluorophenyl ester with the amine from the second
biomolecule sometimes required between 4 and 6 days to reach
completion. This led on occasion to decomposition of fragile
biomolecules under the reaction conditions.
cartridges (0.7 g) purchased from Waters Corporation, Milford,
MA, USA. NMR spectra were recorded on a Varian Unity INOVA
(1H at 499.9 MHz) (Varian Inc., Palo Alto, CA, USA) or a Varian Unity
(13C at 100.6 MHz and 19F at 376.3 MHz) spectrometer. Chemical
shifts were referenced directly via the internal standards: 1H
NMR spectra were referenced to the TMS signal (1H d 0.0), 13C
NMR spectra were referenced to the middle deuterochloroform
signal (13C d 77.0), 19F NMR spectra were referenced to CCl3F
indirectly via the lock signal. All the peptides and peptide
conjugates were characterized by MALDI-TOF mass spectrometry,
on an Applied Biosystems Voyager-DE PRO instrument (Applied
Biosystems, Foster City, CA, USA), which was calibrated to Perkin
Elmer Seqazyme Peptide Mass Standards Kit Calmix 2 and
Calmix 3 (PerkinElmer, San Jose, CA, USA). Electrospray mass
spectra (ESI-MS) for direct-infused dilute methanol solutions
were recorded in both positive and negative ion modes using
a Perkin-Elmer SCIEX API 150-EX mass spectrometer. High-
resolution mass spectrometry (HRMS) characterization of the
diesters was carried out on an Apex-Qe Ultra 7T instrument
(Bruker Daltonics Inc., Billerica, MA, USA) using ESI in the positive
ion mode. The instrument was calibrated with the Tunemix
mixture (Bruker Daltonics). The mass accuracy was approximately
2–3 ppm. The instrumental parameters were as follows: scan range,
200–1200 m/z; drying gas, nitrogen; temperature of drying gas,
200 ꢀC; potential between spray needle and orifice, set at 4.5 kV;
source accumulation time, 0.5s; and ion accumulation time, 0.5s.
The samples for MS experiments (~0.05 mg) were dissolved in the
methanol containing 10À4 M NaCl. Melting points were measured
on Stuart SMP10 Melting Point Apparatus (Beacon Road, Stone,
Staffordshire, UK) and were uncorrected. IR experiments performed
on Perkin Elmer Spectrum-100 FT-IR spectrometer equipped with
an attenuated total reflectance (ATR) accessory. Analytical HPLC
was performed using a Hypersil C-18 Gold column, 170-Å pore size,
5-Å particle size, 250 Â 4.6mm diameter, eluting with a 30–70%
gradient of acetonitrile in water containing 0.1% of TFA over
40 min at a flow rate of 1.2ml/min. HPLC purifications of the
peptides were performed by HPLC semi-preparative Hypersil C-18
column (Runcom, UK) 250 Â 20 mm, 5-mm particle size eluting
with a 35–75% gradient of acetonitrile in water containing 0.1%
of TFA over 40 min at a flow rate of 10 ml/min.
Here, we wish to report on an improved method for
bioconjugation, where the difference between the pKa values
of the phenols used for diester formation is smaller. The
polyfluorophenyl esters were systematically varied and investi-
gated with regards to reactivity under the reaction conditions,
because a variety of pKa values is accessible simply by varying
the number and position of fluorine atoms in the aromatic ring.
A pair of polyfluorophenyl esters was identified, which provides
both specificity and short conjugation times. Demonstrated
reaction examples include the formation of peptide-peptide
conjugates as well as peptide-protein conjugates.
All reagents and gels used for gel electrophoresis and gel
staining were from Invitrogen (Life Technologies, Grand Island,
NY, USA). SDS-PAGE electrophoresis was conducted using
NuPAGEW Novex 4–12% Bis-Tris Gel 1.5 mm, which was
developed in 3-(N-morpholino)propanesulfonic acid running
buffer. All gels were stained using Colloidal Blue Staining Kit.
Materials and Methods
Abbreviations used in the Paper
AcOEt, ethyl acetate; DCM, dichloromethane; DIC, N,N′-
diisopropylcarbodiimide; DIPEA, N,N′-diisopropylethylamine;
DMF, N,N-dimetyloformamid; DMSO, dimethyl sulfoxide; equiv,
equivalents; HCTU, (2-(6-Chloro-1H-benzotriazole-1-yl)-1,1,3,
3-tetramethylaminium hexafluorophosphate); HEPES, hydroxyethyl
piperazineethanesulfonic acid; PyBOP, (Benzotriazol-1-yloxy)
tripyrrolidinophosphonium hexafluorophosphate; HSA, human
serum albumin; py, pyridine; NMP, N-Methyl-2-pyrrolidone; TFA,
trifluoroacetic acid; Tfa, trifluoracetyl; TIS, triisopropylsilane.
Peptide Synthesis
Peptides were synthesized using automated solid-phase
methodology on an Applied Biosystems 433A peptide synthe-
sizer, employing standard Fmoc/tBu strategy using the FastMoc
synthesis program. The syntheses were performed on 0.1-mmol
scale, and H2N-RinkAmide-ChemMatrix (PCAS BioMatrix Inc.,
Saint-Jean-sur-Richelieu, QC, Canada) resin with a loading of
0.46 mmol gÀ1 was used as the solid support. All the coupling
steps were conducted with a HCTU/DIPEA (Iris Biotech GmbH,
Marktredwitz, Germany and Peptides International Inc. Louisville,
Kentucky, USA) activation cocktail. Fmoc deprotection was
achieved by piperidine/NMP (2/8 v/v) treatment. All reagents
used in the peptide synthesis were prepared according to the
manufacturer’s protocols. The side chains of the amino acids
General Methods
All reagents and solvents were purchased from commercial
sources and were used without further purification. TLC was
performed on 60-F254 silica and 60-F254 aluminum oxide plates
(Merck, Darmstadt, Germany), and spots were visualized with
UV light (l = 254 nm) unless otherwise stated. Sep-Pak C-18
wileyonlinelibrary.com/journal/jpepsci Copyright © 2012 European Peptide Society and John Wiley & Sons, Ltd. J. Pept. Sci. 2012; 18: 261–269

EFFICIENT TOOL FOR HETEROCONJUGATION OF PROTEINS AND PEPTIDES
(Iris Biotech GmbH and Pepnet Inc.) were protected by base-
stable groups: tert-butyl ester (Asp, Glu), trityl (His, Asn, Gln)
and 2,2,4,6,7-pentamethyldihydrobenzofuran-5-sulfonyl (Arg).
Peptide cleavage from the solid support was achieved by
treatment with the cleaving cocktail (TFA/TIS/H2O : 95/2.5/2.5v/v)
Purifications of the peptides were performed by HPLC.
2,4,5-trifluorophenyl Pentafluorophenyl Adipate (2c)
Prepared analogously to the compound 2a; white solid, yield:
424 mg, 96%; m.p. 43 ꢀC, IR (ATR) 3085, 2926, 1773, 1517, 1409,
1369, 1288, 1209, 1119, 988, 854; 1H NMR (500 MHz, CDCl3) d
ppm 7.02 (m, 1H), 2.75 (m, 2H), 2.68 (m, 2H), 1.90 (m, 4H); 13C
NMR (100 MHz, CDCl3) d ppm 169.9, 168.9, 112.7, 112.5, 106.4,
106.2, 105.9, 33.2, 32.9, 24.0, 23.9; 19F NMR (376 MHz, CDCl3) d
ppm À129.8 (m), À136.6 (m), À140.7 (m), À153.3 (m), À158.6
(m), À162.8 (m); HRMS: calculated for C18H10F8O4Na: 465.0349,
observed: 465.0345.
Synthesis of the Peptide 3
A resin containing peptide (0.1mmol) with a free N-terminus,
obtained directly from the peptide synthesizer, was immersed in
the NMP solution (5ml) of 7-methoxycoumarin-3-carboxylic acid
(88 mg, 0.4mmol, 4 equiv.), PyBOP (208 mg, 0.4 mmol, 4 equiv.)
and DIPEA (104 mg, 140 ml, 0.8mmol, 8 equiv.). The reaction
mixture was gently agitated for 2 h, and then the resin was filtered
off, washed five times with NMP and twice with DCM. The peptide
was cleaved from the resin and purified by HPLC.
2,4,6-trifluorophenyl Pentafluorophenyl Adipate (2d)
Prepared analogously to the compound 2a; white solid, yield:
427 mg, 96%; m.p 43–44 ꢀC; IR (ATR) 2956, 2886, 1791, 1763,
1516, 1469, 1456, 1381, 1250, 1212, 1097, 985, 916, 839, 737; H
NMR (500 MHz, CDCl3) d ppm 6.76 (m, 2H), 2.74 (m, 4H), 1.92
(m, 4H); 13C NMR (100 MHz, CDCl3) d ppm 169.6, 168.9, 160–153
(m), 142.4–136 (m), 101.1, 100.8, 100.6, 32.9, 23.9; 19F NMR
(376 MHz, CDCl3) d ppm À110.8 (m), À123.5 (m), À153.2 (m),
À158.4 (m), À162.7 (m); HRMS: calculated for C18H10F8O4Na:
465.0349, observed: 465.0342.
1
2,4-difluorophenyl Pentafluorophenyl Adipate (2a)
The compound was prepared according to the procedure
described previously [9]. In short, pyridine (5 mmol) in CH3CN
(10 ml) was added drop wise to a solution of the adipoyl chloride
(1.83 g,10 mmol, 2 equiv) in CH3CN (50 ml) at 0 ꢀC with vigorous
stirring. The reaction mixture was allowed to attain rt (25 ꢀC)
and a solution of 2,4-difluorophenol (650 mg,5 mmol, 1 equiv)
in CH3CN (5 ml) was added and stirring was continued for 3 h.
CH3CN was removed under reduced pressure, and the residue
was dissolved in AcOEt (100 ml) and washed with 1 N HCl
(100 ml). The crude solution of the ester was dried over anhy-
drous MgSO4 and was concentrated. The obtained monoesters
were purified by flash chromatography (silica gel), eluting firstly
with DCM to remove diester byproduct, and later, the mobile
phase was changed to 30% AcOEt in DCM. Obtained monoester
1a (yield: 920 mg, 72%) was taken directly to the next stage
without further characterization. The monoester 1a (258 mg,
1 mmol) were treated with pentafluorophenol (370 mg, 2 mmol,
2 equiv) and DIC (250 mg, 2 mmol, 2 equiv) in a mixture of CH3CN
and pyridine (9 : 1, 10 ml) for 2 h at rt. The reaction mixture was
concentrated, and the residue was subjected to column
chromatography (silica gel) using CH2Cl2 as eluent to obtain the
active diesters 2a as white crystalline solid. Yield: 438 mg, 98%;
m.p. 42–44 ꢀC, IR (ATR): 2965, 2890, 1788, 1763, 1618, 1518,
2,6-difluorophenyl Pentafluorophenyl Suberate (2e)
Prepared analogously to the compound 2a; white solid, yield:
466 mg, 98%; m.p. 40–42 ꢀC; IR (ATR) 2939, 2863, 1772, 1517,
1449, 1416, 1365, 1289, 1243, 1202, 1096, 1037, 984, 856, 769;
1H NMR (500 MHz, CDCl3) d ppm 7.17 (m, 1H), 6.98 (m, 2H), 2.68
(m, 4H), 1.83 (m, 4H), 1.51 (m, 4H); 13C NMR (100 MHz, CDCl3) d
ppm 170.1, 169.4, 156.5, 154.0, 144–139 (m), 127.2, 126.3, 126.2,
126.1, 112.1, 112.0, 111.9, 111.8, 33.4, 33.1, 28.4, 24.6, 24.5; 19F
NMR (376 MHz, CDCl3) d ppm À126.8 (m), À153.2 (m), À158.5
(m), À162.8 (m); HRMS: calculated for C20H15F7NaO4: 475.0756,
observed: 475.0750.
General Procedure for the Synthesis of Peptide-linker
Conjugates 4a–e
Synthesis of the peptide derivative 4a: The dodecapeptide, 3
(7.5 mg, 5 mmol) was dissolved in anhydrous DMSO (1 ml), and
to the peptide solution thus obtained, a solution of the active
diester 2a (4 mg, 10 mmol) pyridine (0.12 ml) was added. The
reaction was followed by reverse phase analytical HPLC and
was stirred until the peptide 3 reacted completely. The reaction
times varied from 45 to 70 min. DMF (200 ml) was added to the
reaction mixture, and the peptide products were precipitated with
diethyl ether and were centrifuged. The residue was redissolved in
DMF (1ml) and precipitated again using diethyl ether. The obtained
crude product 4a was then analyzed by analytical HPLC, indicating
sufficient purity. Therefore, the obtained peptide was dissolved in
CH3CN/H2O (1: 9 v/v), lyophilized and used for further purposes,
without any extra purification; 7.4mg (84%). MALDI-MS: expected
for [M+ H]+: 1778.22, observed: 1778.43.
1
1505, 1381, 1243, 1094, 986, 856, 737; H NMR (500 MHz, CDCl3)
d ppm 7.11 (m, 1H), 6.91 (m, 2H), 2.75 (m, 2H), 2.68 (m, 4H);
13C NMR (100 MHz, CDCl3) d ppm 170.4, 168.9, 124.2, 124.1,
111.3, 111.1, 105.3, 105.1, 104.8, 33.3, 32.9, 24.0; 19F NMR
(376 MHz, CDCl3) d ppm À113.1 (m), 124.0 (m), À153.3 (m),
À158.6 (m), À162.8 (m); HRMS: calculated for C18H11F7O4Na:
447.0443, observed: 447.0433.
2,6-difluorophenyl Pentafluorophenyl Adipate (2b)
Prepared analogously to the compound 2a; white solid, yield:
424 mg, 95%; m.p. 41–43 ꢀC, IR (ATR): 2943, 2882, 1770, 1516,
Likewise, peptide conjugates 4b, 4c, 4d, 4e were synthesized
from peptide 3 and diesters 2b, 2c, 2d, 2e, respectively.
1
1603, 1516, 1384, 1292, 1246, 1100, 986, 782; H NMR (500 MHz,
CDCl3) d ppm 7.17 (m, 1H), 6.97 (m, 2H), 2.74 (m, 2H), 1.93
(m, 4H); 13C NMR (100 MHz, CDCl3) d ppm 169.6, 168.9, 156.9,
153.6, 142.8, 141.1, 139.5, 137.8, 136.2, 126.9, 125.93, 112.9–111.1
(m), 32.9, 23.9; 19F NMR (376 MHz, CDCl3) d ppm À126.8 (m),
À153.3 (m), À158.6 (m), À162.9 (m); HRMS: calculated for
C18H11F7O4Na: 447.0443, observed: 447.0432.
4b: Yield: 6.6 mg (74%), MALDI-MS: expected for [M + H]+:
1778.22, observed: 1778.69.
4c: Yield: 7.1 mg (80%) including impurities, MALDI-MS after
analytical HPLC: expected for [M + H]+: 1796.43, observed:
1796.15.
J. Pept. Sci. 2012; 18: 261–269 Copyright © 2012 European Peptide Society and John Wiley & Sons, Ltd. wileyonlinelibrary.com/journal/jpepsci

ŚLÓSARCZYK AND BALTZER
4d: Yield: 7.7 mg (85%) including impurities, MALDI-MS after
analytical HPLC: expected for [M + H]+: 1796.43, observed:
1796.55.
materials had disappeared and the mixture was diluted with H20/
AcOH (98 : 2, 3 ml), washed with AcOEt-ether 1 : 1, the aq. phase
was evaporated to approximately ½ volume, and was applied to
a Sep-Pack Plus (C-18 cartridges, Waters), which was initially
washed with CH3CN and later equilibrated with water. The
cartridge was washed first with water, then the mobile phase with
the increasing concentration of CH3CN (10% steps, up to 50%)
gave TLC-pure fractions containing 10, which were pooled,
concentrated to half volume and then lyophilized. Yield:
59 mg, 70%. 1H NMR (500 MHz, CDCl3) d ppm 7.41 (d, J = 8.79, 2H),
7.28 (m, J = 8.79, 3H), 7.07 (m, 2H), 5.24 (m, 1H), 4.62 (d, J = 7.64,
1H), 4.28 (m, 3H), 4.21 (m, 1H), 4.07 (m, 1H), 3.98 (m, 2H), 3.77
(m, 10H), 3.38 (m, 1H), 3.16 (m, 1H), 3.00 (m, 1H), 2.88 (m, 1H), 2.75
(m, 2H), 2.46 (s, 2H), 1.74 (s, 4H), 0.96 (s, 3H); 13C NMR (100 MHz,
CDCl3) d ppm 173.7, 171.8, 156.3, 154.5, 153.8, 144.1, 136.1, 135.3,
129.2, 127.2, 126.9, 121.4, 112.4, 112.2, 101.5, 98.9, 93.8, 77.3, 74.8,
72.9, 72.3, 71.6, 70.3, 69.9, 69.8, 69.5, 68.6, 68.3, 66.8, 63.9, 61.8,
61.2, 39.3, 36.4, 34.7, 33.0, 24.9, 24.3, 15.8; 112.0, 111.9, 111.8, 33.4,
33.1, 28.4, 24.6, 24.5; 19F NMR (376MHz, CDCl3) d ppm À127.4.
ESI-MS: expected for [M+ H]+: 847.3, observed: 847.3. Expected for
[M+ Na]+ 870.3, observed: 870.2.
4e: Yield: 8 mg (90%), MALDI-MS: expected for [M + H]+:
1806.54, observed: 1806.62.
General Procedure for the Synthesis of the Peptide-peptide
Conjugate 6a and 6e
Synthesis of peptide conjugate 6a: The peptide 5 (3 mg, 0.6 mmol)
was dissolved in anhydrous DMSO (1 ml), and a solution of the
peptide derivative 4a (or 4b) (1.5 mg, 1.2 mmol, 2 equiv.), pyridine
(0.2 ml) and DIPEA (20 ml) in DMSO (0.5 ml) was added. The stirred
reaction was followed by reverse phase analytical HPLC, and
stirring was continued until 5 reacted completely. The reaction
mixture was diluted with anhydrous diethyl ether and centri-
fuged. The residue was dissolved in TFA (1 ml) and precipitated
again using diethyl ether. The crude product 6a was then purified
by semi-preparative reversed-phase HPLC. The purified fractions
were analyzed by MALDI-TOF mass spectrometry and were
pooled together, concentrated and lyophilized. The isolated yield
of the functionalized peptide 6a was 3.3 mg (84%). MALDI-MS:
expected for [M + H]+: 6635.05, observed: 6635.13.
Synthesis of the Sugar-peptide Conjugate 11
Likewise, peptide conjugate 6e was obtained from peptide 5
and peptide derivative 6e using analogous procedure. Yield:
3.5 mg (88%). MALDI-MS: expected for [M + H]+: 6661.11,
observed: 6661.15.
This compound was prepared in 90% yield from peptides 10 and
3
using essentially the same coupling procedure as for
compound 6a in 24 h. MALDI-MS: expected for [M + H]+:
2255.79, observed: 2255.67.
Synthesis of the Peptide-peptide Conjugate 8a and 8b
Synthesis of the Peptide Derivative 13
These compounds were prepared in 80% and 70% yield from
peptides 7a of 7b, respectively, and 4b using essentially the
same procedure as for compound 6a, but with twice the molar
amount of peptide 4b. The reaction time for 8a was approxi-
mately 36 h and for 8b 60 h. MALDI-MS: expected for 8a
[M + H]+: 8520.79, observed: 8520.76; expected for 8b [M + H]+:
8215.23, observed: 8215.15.
This compound was prepared in 85% yield from 12 and linker 2b
using essentially the same coupling procedure as for compound
4b in 3 h. MALDI-MS: Expected for [M + H]+: 5330.48, Observed:
5330.36.
Conjugation of 13 to HSA
Lyophilized HSA (1mg) was suspended in the HEPES buffer pH= 8
(1.5ml, 50 mM), and the suspension was stirred until a clear solution
was obtained. To the solution, derivative 13 (1mg, 0.2 mmol, 10
equiv.) was added and the reaction mixture was stirred at rt. After
24 h, another portion of the peptide 13 was added (1mg, 0.2mmol,
10 equiv.). After 48 h of total reaction time, the reaction mixture was
diluted with water to a total volume of 10 ml and processed by
means of ultrafiltration (cut-off= 30 K Da, 4400rpm) until the
volume of the upper liquid reached 1 ml. The protein-containing
Synthesis of the Trisaccharide Derivative 10
A solution of 2-(p-aminophenyl)ethyl 2-O-(a-L-fucopyranosyl)-3-O-
(a-D-galactopyranosyl)-b-D-galactopyranoside [9] (60 mg, 0.1 mmol)
was dissolved in dry DMSO (0.5ml) containing pyridine (0.05 ml),
and then 2b (63 mg, 0.15 mmol) was added. After 30 min, TLC
(AcOEt/AcOH/MeOH/H20, 12: 3 :3: 2; detection with 5% sulfuric
acid + heating) showed appearance of a faster Rf 0.5 spot and
disappearance of the starting sugar Rf 0.1 spot. After 5 h, all starting
Scheme 1. The synthesis of mixed diesters (linker candidates).
wileyonlinelibrary.com/journal/jpepsci Copyright © 2012 European Peptide Society and John Wiley & Sons, Ltd. J. Pept. Sci. 2012; 18: 261–269

EFFICIENT TOOL FOR HETEROCONJUGATION OF PROTEINS AND PEPTIDES
solution was diluted up to 10 ml and ultrafiltrated again down to
the volume of 1 ml. This procedure was repeated ten times; after
which, the filtrate did not show any fluorescent properties. The
obtained protein solution was analyzed by SDS-PAGE. To deter-
mine the relative level of HSA functionalization, we have compared
the total protein concentration contained in the retentate with the
concentration of the functionalized protein determined by the UV
absorption of the coumarin (e = 22000MÀ1 cmÀ1). The concentra-
tion of the modified species obtained from the Lambert–Beer
formula gave the concentration of 16.34 mM. The sample used for
UV analysis was diluted four times and subjected to the analysis
of total protein concentration by a certified Amino Acid Analysis
Center, Uppsala, basing on mean concentration of tyrosine,
threonine and serine, as these amino acids do not occur in the
peptide used for conjugation. The concentration value thus
obtained was 3.3 mM.
ester. Four polyfluorophenyl esters formed from phenols with pKa
values in the range from 8.7 to 7.5 were prepared to identify
candidates for incorporation in the diester for improved reactivity
in the second amidation reaction. Mixed adipic acid diesters
(1a–d) were synthesized according to the procedure published
previously [9] (Scheme 1). The adipic acid dichloride was reacted
with a polyfluorophenyl ester (1 equiv.) in the presence of pyridine
(1 equiv.) to give the monoester after aqueous workup. Monoesters
obtained according to this procedure were further reacted with
pentafluorophenol in a carbodiimide mediated reaction to form
the mixed diesters 2a–d.
All four diesters were reacted with the a-helical dodecapeptide
3 having one lysine at position 4 (Scheme 2). At this stage, it was
desired that only the pentafluorophenyl ester would react with
the amino group, whereas the other ester should remain
intact. The reaction was performed in 10% pyridine in DMSO.
Table 1. Reaction times and yields for the reactions between pep-
tide 5 and peptide derivatives functionalized with polyfluorophenyl
esters
Results and Discussion
We previously identified pentafluorophenyl esters as very fast
and reliable amidation reagents for the first step of the
heteroconjugation reaction and concluded based on those
investigations that no improvement was necessary because a
pentafluorophenyl ester provides an essentially instantaneous
reaction with a lysine side chain in the presence of base. Attention
was instead focused on the slow reaction of the o-fluorophenyl
Peptide derivative
(CH2)n Reaction timea/ yieldb
4a
4b
4
4
60 h/80%
36 h/85%
O
n
O
10% py in DMSO
+
2a-d
O
-LARKFEAFARAG-NH2
NH2
HN
*
Fx
-LARKFEAFARAG-NH2
*
reaction times: 2-3 h
4e
4f
6
4
<48 h/80%
3
4a-e
for a,b,c,d: n = 4
for : n = 6
e
F
F
F
F
Approx. 50%
O
F
a
c
O
Conversion after 72 h
F
=
F
F
O
Fx
b,e
d
O
O
F
F
aThe reaction mixture was sampled every 12 h, and the obtained
aliquots were analyzed by HPLC and MALDI-MS.
bWeight obtained after lyophilization.
Scheme 2. The reaction between the linker candidates 1a–d and
dodecapeptide 3; * = 7-methoxycoumarin-3-carbonyl.
Scheme 3. Conjugation of 42-residue polypeptide 5 to peptide derivatives formed from reacting dodecapeptide 3 with bifunctional reagents.
J. Pept. Sci. 2012; 18: 261–269 Copyright © 2012 European Peptide Society and John Wiley & Sons, Ltd. wileyonlinelibrary.com/journal/jpepsci

ŚLÓSARCZYK AND BALTZER
To facilitate identification of the reaction products, the
dodecapeptide was N-terminated with the fluorophore
7-methoxycoumarin-3-carbonyl.
In all cases, the reactions between diesters 2a–d and the peptide
3 were completed within 2–3 h. However, only experiments where
compounds 2a and 2b were used gave pure functionalized peptide
3
Scheme 4. Reactions of 42-residue polypeptides 7a and 7b to form triple conjugates 8a and 8b. The peptide 7a holds four lysine residues where
Lys-15 is modified by incorporation of a dansyl fluorophore (dansyl = 5-(dimethylamino)naphthalene-1-sulfonyl), and the side chain of Lys-41 is
protected by a Tfa (trifluoracetyl) group, whereas Lys-8 and Lys-35 remain free for conjugation. The side chain of the cysteine residue Cys-24 is protected
by an Acm (acetylaminomethyl) group. The peptide 7b holds three lysine residues where Lys-39 is functionalized with a coumarin fluorophore
(coumarin = 7-methoxycoumarin-3-carbonyl), whereas Lys-25 and Lys-28 remain free for conjugation.
Scheme 5. The synthesis of the carbohydrate-peptide conjugate 11.
wileyonlinelibrary.com/journal/jpepsci Copyright © 2012 European Peptide Society and John Wiley & Sons, Ltd. J. Pept. Sci. 2012; 18: 261–269

EFFICIENT TOOL FOR HETEROCONJUGATION OF PROTEINS AND PEPTIDES
derivatives, 4a and 4b, respectively. The reactions with the
trifluorophenyl esters 2c and 2d gave rise to significant amounts
of the peptide homoconjugate formed from two copies of 3 and
the linker reagent. The results showed that it was possible to
selectively form monoamides from pentafluorophenyl esters in
the presence of 2,4-difluorophenyl and 2,6-difluorophenyl
esters. Both difluorophenyl esters could therefore be used in a bi-
functional reagent where the functionalized peptide derivative
formed in the initial reaction with the pentafluorophenyl ester
was to be reacted with a second biomolecule in the presence of
1% DIPEA to form a heteroconjugate in high yield. A bifunctional
linker thus prepared would provide not only the required selectivity
but also an improved reaction rate compared with that of the
previously published 2-fluorophenyl pentafluorophenyl diester
reagent [9]. Because the linker candidates 2c and 2d led to the
formation of homoconjugates, they were excluded from
further investigations.
To evaluate the performance of the 2,6-difluorophenyl and 2,4-
difluorophenyl esters with respect to the difference in rate of
amidation in comparison to that of the pentafluorphenyl ester,
we investigated the conjugation of the peptide derivatives 4a,
4b and 4e to a 42-residue peptide 5 [10,11] (Scheme 3). The
polypeptide was previously designed to fold into a helix-loop-
helix motif and shown to form a helix-loop-helix dimer in
aqueous solution [12]. The 42-residue N-acetylated polypeptide
5 has a single lysine residue with a free amino group at position
17. The experiment was performed in DMSO in the presence of
10% pyridine and 1% DIPEA, and the results are summarized in
Table 1. For comparison, we also conjugated the peptide
derivative 4f [9] carrying the o-fluorophenyl ester to the 42-residue
polypeptide 5.
The functionalized peptide derivatives formed in the reaction
between the dodecapeptide 3 and the bifunctional reagents
equipped with 2,4-difluorophenyl or 2,6-difluorophenyl esters
(4a and 4b) reacted with 5 at a significantly increased reaction
rate in comparison to that observed in the reaction between 5
and the dodecapeptide carrying an o-fluorophenyl ester (4f).
Under the reaction conditions, 4f did not reach more than 50%
conversion into the heteroconjucate even after 72 h. This
suggests that the use of the 2,6-difluorophenyl ester increases
the reaction rate by at least a factor of four without loss of
selectivity. We also successfully synthesized the peptide deriva-
tive 4e from peptide 3 and 2e and the 2,6-difluorophenyl ester
equipped with a longer aliphatic chain to prove that an extension
of the linker does not affect the reaction time significantly.
The improved methodology was next examined in the assem-
bly of more sophisticated peptide constructs, and two units of
peptide 4b were conjugated to the two 42-residue N-acetylated
helical peptides 7a and 7b, each possessing two lysines with free
side-chain amino groups (Scheme 4). In polypeptide 7a, each
lysine was situated on a separate helix well separated in the
sequence. The conjugation went smoothly and reached comple-
tion in approximately 36 h, giving peptide conjugate 8a. In
polypeptide 7b, two lysine residues were positioned in the same
helical segment and in close spatial proximity. In this case, it was
suspected that after successful reaction with the first 4b unit,
access to the second free lysine might become difficult.
Despite the complexity of this model system, we were able to
successfully obtain the doubly ligated peptide conjugate 8b in
70% yield in 60 h.
The optimized methodology turned out to function well in the
case of constructing neoglycoconjugates. A derivative of the
blood group B antigenic determinant, 9, equipped with a free
amino group was reacted under mildly basic conditions with
the linker 2b to give the 2,6-difluorophenyl ester 10 in 70% yield.
The activated trisaccharide 10 was then conjugated to the
peptide 3 in the presence of 1% DIPEA, resulting in the formation
of neoglycoconjugate 11 in very good (90%) yield (Scheme 5).
The astonishingly short reaction time was probably due to the
absence of sterical interferences between the reagents.
From the perspective of utilizing the current linker methodol-
ogy for protein functionalization, HSA was selected as a model
protein. The choice of this particular protein was not casual, as
it illustrates numerous advantages of the present technology. It
is one of the most abundant proteins in human blood; it is
used as a vehicle for drug delivery and has pharmacokinetic
Scheme 6. The synthesis of the peptide derivative 13. The N-acetylated
peptide 12 carries two lysine residues, where Lys-10 was derivatized with
the coumarin fluorophore (coumarin = 7-methoxycoumarin-3-carbonyl),
whereas Lys-17 is free and available for functionalization.
O
O
N
4
H
13, HEPES buffer pH 8
*
H2O
n
Scheme 7. Conjugation of HSA with the peptide derivative 13. A mixture of HSA conjugates was obtained because of the fact that the functionalizing
reagent does discriminate between the lysine residues on the surface of the HSA protein.
J. Pept. Sci. 2012; 18: 261–269 Copyright © 2012 European Peptide Society and John Wiley & Sons, Ltd. wileyonlinelibrary.com/journal/jpepsci

ŚLÓSARCZYK AND BALTZER
implications [13]. Thus, if the linker technology was to prove
successful in the functionalization of HSA, new opportunities in
drug delivery could be opened. HSA is a well-characterized and
investigated protein frequently used in conjugation reactions. It
is also cheap and readily available from commercial sources. For
conjugation reactions of proteins, DMSO cannot be used, and this
investigation was performed using HEPES buffer at pH 8 as the
solvent. At this pH, a compromise had to be found between
protein conjugation that requires unprotonated amino groups
and undesired active ester hydrolysis. For conjugation to HSA,
peptide 12 was selected and reacted with 2b in 10% pyridine
in DMSO to give compound 13 (Scheme 6). After isolation, the
peptide derivative was added to the HSA solution in HEPES buffer
at pH 8 (Scheme 7).
The reaction was discontinued after 48 h, and the protein-
peptide conjugate purified by ultrafiltration, where unreacted
13 equipped with coumarin was removed by repeated filtrations
until no fluorescence, was observed in the filtrate (Figure 1). The
retentate showed intense fluorescence apparently due to the
presence of 13 equipped with coumarin and conjugated to
HSA, which was retained by the filter because of the high
molecular weight of the peptide-protein conjugate.
Figure 2. SDS-PAGE analysis of HSA functionalized with the peptide
derivative 13. Lane 1, HSA reference. Lane 2, HSA functionalized with 13
(500 fold diluted reaction mixture). Lane 3, HSA functionalized with 13
(100 fold diluted reaction mixture). Lane 4, HSA functionalized with 13 (10
fold diluted reaction mixture).
Conclusion
In summary, a bifunctional reagent with improved properties
over those previously reported has been developed for the
covalent linkage of biomolecules with free amino groups
including peptides, proteins and carbohydrates. The optimal
bifunctional reagent for the conjugation of biomolecules where
variation of distance between subunits is required is a diester
formed from aliphatic dicarboxylic acids. An evaluation of poly-
fluorophenyl ester pairs showed that a 2,6-difluorophenyl ester
‘in concerto’ with a pentafluorophenyl ester, provides both selec-
tivity and optimal reaction times for the assembly of heteroconju-
gates. The time needed to prepare the complete conjugate
proved to be approximately four times shorter compared with
the time needed using the reagents reported previously [9],
which is of considerable importance in the conjugation of sensi-
tive biomolecules. The new linker methodology proved to be
efficient in constructing a variety of conjugates between peptides
and peptides, between peptides and carbohydrates as well as
between peptides and proteins, demonstrating that the scope
of the applicability is wide. The results suggest that this linker
technology has the capacity to significantly influence biotechno-
logical and biochemical research whenever construction of
bifunctional conjugates is required. It is envisioned that the
bifunctional linkers reported here will find their optimal use in
the heteroconjugation of proteins using lysine side chains, a
reaction for which currently few alternatives exist, if access to
spacers of variable size is required.
The supernatant was analyzed by SDS-PAGE followed by
coomassie staining (Figure 2), and several bands were observed.
Apparently, several copies of 13 were incorporated to form an
array of conjugates with a range of molecular weights. The
bifunctional reagent does not show selectivity towards the 58
lysine residues, most of which are equally distributed on the
surface of the protein, making them available to react with
other molecules in the system [14]. Therefore, it is not possible
to obtain labeled conjugates with a well-defined number of
polypeptides. Nevertheless, the separated protein conjugates
showed fluorescent properties due to the coumarin fluorophore,
thus showing the efficiency of the reagent also in aqueous
environment. A quantitative comparison between the total
protein concentration as measured by quantitative amino acid
analysis and the concentration of the fluorescent label as
determined from the absorbance measured by UV spectroscopy
resulted in an approximate 1 : 2 ratio, which indicated that on
the average, every HSA molecule was functionalized with five
peptide 13 units.
Acknowledgement
We thank the Swedish Research Council for financial support and
Prof. Thomas Norberg for a generous gift of the carbohydrate
derivative.
References
Figure 1. Illustration of functionalization of human serum albumin,
showing filtrates at different stage of ultrafiltration and supernatant after
filtration, irradiated with UV light at 366 nm. Vial 1, filtrate after first
ultrafiltration cycle. Vial 2, filtrate after sixth ultrafiltration cycle. Vial 3,
retentate after sixth ultrafiltration cycle showing strong fluroescence
emission due to peptide-protein conjugate.
1 Kalia J, Raines RT. Advances in bioconjugation. Curr. Org. Chem. 2010;
14: 138–147.
2 Zhaoyang YE, Houssam S, Houssein H, Mahato RI. Bioconjugation of
oligonucleotides for treating liver fibrosis. Oligonucleotides 2007; 17:
349–404.
wileyonlinelibrary.com/journal/jpepsci Copyright © 2012 European Peptide Society and John Wiley & Sons, Ltd. J. Pept. Sci. 2012; 18: 261–269

EFFICIENT TOOL FOR HETEROCONJUGATION OF PROTEINS AND PEPTIDES
3 Farkaš P, Bystrický S. Chemical conjugation of biomacromolecules: a
mini-review. Chem. Pap. 2010; 6: 683–695.
10 Baltzer L. Crossing borders to bind proteins – a new concept in
protein recognition based on the conjugation of small organic
molecules or short peptides to polypeptides from a designed set.
Anal. Bioanal. Chem. 2011; 400: 1653–166.
4 Hermanson G. Bioconjugate Techniques. Academic Press: San Diego, 2008.
5 Hartman FC, Wold F. Bifunctional reagents. Cross-linking of pancreatic
ribonuclease with a diimido ester. J. Am. Chem. Soc. 1966; 88: 3890–3891.
6 Bragg PD, Hou C. Subunit composition, function, and spatial arrange-
ment in the Ca2+ and Mg2+-activated adenosine triphosphatases of
Escherichia coli and Salmonella typhimurium. Arch. Biochem. Biophys.
1975; 167: 311–321.
7 Storer R, Aciro C, Jones L. Squaramides: physical properties, synthesis
and applications. Chem. Soc. Rev. 2011; 40: 2330–2346.
8 Jobbagy A, Kiraly K. Chemical characterization of fluorescein
isothiocyanate-protein conjugates. Biochim. Biophys. Acta 1966;
124: 166.
11 Ślósarczyk AT, Baltzer L. The molecular recognition of phosphorylated
proteins by de-signed polypeptides conjugated to a small molecule
that binds phosphate. Org. Biomol. Chem. 2011; 9: 7697–7704.
12 Olofsson S, Johansson G, Baltzer L. Design, synthesis and solution
structure of a helix-loop-helix dimer – a template for the rational
design of catalytically active polypeptides. Chem. Soc. Perkin Trans.
1995; 2: 2047–2056.
13 Bertucci C, Domenici E. Reversible and covalent binding of drugs
to human serum albumin: methodological approaches and
physiological relevance. Curr. Med. Chem. 2002; 9: 1463–1481.
14 Ishii T, Ito S, Kumazawa S, Sakurai T, Yamaguchi S, Mori T, Nakayama T,
Uchida K. Site-specific modification of positively-charged surfaces on
human serum albumin by malondialdehyde. Biochem. Biophys. Res.
Commun. 2008; 371: 28–32.
9 Ślósarczyk AT, Ramapanicker R, Norberg T, Baltzer L. Mixed
pentafluorophenyl and o-fluorophenyl esters of aliphatic dicarboxylic
acids: efficient tools for peptide and protein conjugation. RSC Adv.
2011. doi: 10.1039/C1RA00530H
J. Pept. Sci. 2012; 18: 261–269 Copyright © 2012 European Peptide Society and John Wiley & Sons, Ltd. wileyonlinelibrary.com/journal/jpepsci