Scaffolded multiple cyclic peptide libraries for protein mimics by native chemical ligation

Article abstract of DOI:10.1039/c4ob00190g

The accessibility to collections, libraries and arrays of cyclic peptides is increasingly important since cyclic peptides may provide better mimics of the loop-like structures ubiquitously present in and-especially-on the surface of proteins. The next important step is the preparation of libraries of ensembles of scaffolded cyclic peptides, which upon screening may lead to promising protein mimics. Here we describe the synthesis of a tri-cysteine containing scaffold as well as the simultaneous native chemical ligation of three cyclic peptides thereby affording a clean library of multiple cyclic peptides on this scaffold, representing potential mimics of gp120. Members of this collection of protein mimics showed a decent inhibition of the gp120-CD4 interaction. the Partner Organisations 2014.

Full text of DOI:10.1039/c4ob00190g

Organic &
Biomolecular Chemistry
View Article Online
View Journal | View Issue
PAPER
Scaffolded multiple cyclic peptide libraries for
protein mimics by native chemical ligation†
Cite this: Org. Biomol. Chem., 2014,
12, 4471
H. van de Langemheen,a M. van Hoeke,a H. C. Quarles van Ufford,a J. A. W. Kruijtzera
and R. M. J. Liskamp*a,b
The accessibility to collections, libraries and arrays of cyclic peptides is increasingly important since cyclic
peptides may provide better mimics of the loop-like structures ubiquitously present in and – especially –
on the surface of proteins. The next important step is the preparation of libraries of ensembles of
scaffolded cyclic peptides, which upon screening may lead to promising protein mimics. Here we
describe the synthesis of a tri-cysteine containing scaffold as well as the simultaneous native chemical
ligation of three cyclic peptides thereby affording a clean library of multiple cyclic peptides on this
scaffold, representing potential mimics of gp120. Members of this collection of protein mimics showed a
decent inhibition of the gp120-CD4 interaction.
Received 24th January 2014,
Accepted 24th April 2014
DOI: 10.1039/c4ob00190g
www.rsc.org/obc
phage display libraries of cyclic disulfide bridged2 and,
more recently, the bicyclic phage display libraries obtained by
Introduction
In the quest for protein mimics possessing desirable pro- alkylation of cysteine containing linear peptide phage display
perties, the generation of collections or even large libraries is libraries3 represented tremendous progress. However, the
instrumental in screening approaches to uncover “hits”, which number of (bio)organic chemical approaches for the prepa-
represent first steps towards these protein mimics. It remains ration of cyclic peptide libraries, let alone combined in ensem-
important to realize that “generation of collections or bles, for mimicry of discontinuous epitopes is still pretty
libraries” and “screening or evaluation of many possibilities” limited. So far most examples in the literature, which have
are very reminiscent of processes in nature where many possi- used chemoselective ligation methods to conjugate peptides to
bilities are available or generated and scanned to select the a scaffold molecule, mainly have focussed on attachment of
compound members with optimal properties.
multiple identical peptides.4–6
Although the availability of arrays or libraries of linear pep-
We have been increasingly interested in the preparation of
tides has provided excellent starting points for development of cyclic peptides especially as ensembles of different combi-
new (peptide) ligands and identification of peptide segments nations. Recently, we reported the first strides towards modest
in peptides and proteins which are crucial determinants of libraries of ensembles of scaffolded cyclic peptides using
their biological activities, it is expected that these starting the bio-orthogonal Cu(I)-catalyzed azide–alkyne-cycloaddition
points might be considerably improved by using as a starting (CuAAC).7,8 However, probably the best biocompatible/bio-
point arrays of cyclic peptides, which in principle provide orthogonal reaction to date is native chemical ligation
better mimics of the loop-like structures ubiquitously present (NCL),9,10 so next to CuAAC, we have great interest in applying
in, and especially, on the surface of proteins.1 Here the (large) native chemical ligation for the construction of protein
mimics. The recently described convenient synthesis of cyclic
peptides containing a thioester handle for native chemical lig-
ation11 enables now their attachment onto a suitably functio-
aMedicinal Chemistry & Chemical Biology, Utrecht Institute for Pharmaceutical
nalized scaffold for the preparation of libraries, which will be
Sciences, Department of Pharmaceutical Sciences, Faculty of Sciences,
described in this manuscript. The clean preparation of a
Utrecht University, P.O. Box 80082, 3508 TB Utrecht, The Netherlands.
library allowed us to carry out a facile screening and evaluation
E-mail: r.m.j.liskamp@uu.nl
bSchool of Chemistry, Joseph Black Building, University of Glasgow,
of the resulting cyclic peptide ensembles as potential protein
mimics of gp120 in a gp120-CD4 receptor ELISA binding assay.
The cyclic peptides in the gp120 mimics correspond to the
sequences of the peptide segments of loops in gp120 together
forming the discontinuous epitope interacting with CD4
(Fig. 1).12–14
University Avenue, Glasgow G12 8QQ, United Kingdom.
E-mail: robert.liskamp@glasgow.ac.uk
†Electronic supplementary information (ESI) available: NMR data of compound
4, HPLC and MALDI-TOF MS data of loop 1 (11a), loop 2 (11b), loop 3 (11c) and
compound 12, HPLC data of 5, mass data of mixture 14, and HPLC and MS data
of 13 and all purified protein mimics. See DOI: 10.1039/c4ob00190g
This journal is © The Royal Society of Chemistry 2014
Org. Biomol. Chem., 2014, 12, 4471–4478 | 4471

View Article Online
Paper
Organic & Biomolecular Chemistry
Fig. 1 Conserved discontinuous epitope of the CD4 binding site on
gp120.12–14
Results and discussion
Synthesis of the tri-cysteine TAC-scaffold and cyclic peptide
Fig. 2 Structures of the LTRDGGN (11a, loop 1), INMWQEVGKA (11b,
loop 2) and SGGDPEIVT (11c, loop 3) gp120 peptide thioester loops.11
loops containing a thioester handle
The required triazacyclophane (TAC) scaffold, capable of
accepting cyclic peptide thioesters in a NCL reaction, was pre-
pared by attachment of (three) cysteine residues to the depro-
tected TAC-scaffold 3. This was easily accessible starting from
the tri-oNBS protected TAC-scaffold (1).15 Amidation of the car-
boxylic acid moiety and removal of the oNBS groups from
scaffold 2, using in situ prepared sodium thiophenolate, was
followed by the simultaneous coupling of three protected
cysteine (Boc-Cys(Trt)-OH) residues to afford scaffold 4. Finally,
the acid-labile protecting groups were removed and the tri-
cysteine TAC-scaffold (5) was purified by preparative HPLC and
obtained in a yield of 74% (Scheme 1).
Cyclic peptide thioesters for native chemical ligation to
scaffold 4 were prepared by a solid phase peptide synthesis
Scheme 2 General SPPS procedure for obtaining cyclic peptides, con-
taining a thioester handle.11
method involving the use of a N-acyl sulfonamide linker,
which is conveniently accessible by a sulfo-click reaction
(Fig. 2).11,16–20 Briefly, our general procedure (Scheme 2) for
the preparation of cyclic peptide thioesters encompassed first
attachment of an orthogonally protected glutamic acid residue
required for peptide cyclization followed by assembly of the
linear peptide. Cyclization was performed on the solid phase
after removal of the (2-phenyl-2-trimethylsilyl)ethyl (PTMSE)
protecting group with fluoride, using (benzotriazol-1-yloxy)tris-
(dimethylamine)phosphonium hexafluorophosphate (BOP) as
a coupling reagent. After alkylation the N-acyl sulfonamide
linker can be cleaved using nucleophilic reagents, such as
Scheme 1 Synthesis of the tri-cysteine TAC-scaffold 5.
4472 | Org. Biomol. Chem., 2014, 12, 4471–4478
This journal is © The Royal Society of Chemistry 2014

View Article Online
Organic & Biomolecular Chemistry
Paper
β-mercaptopropionic acid ethyl ester, leading to the desired 200 mM MPAA, 40 mM tris(2-carboxyethyl)-phosphine hydro-
thioester. Finally, the acid-labile side-chain protecting groups chloride (TCEP), pH 7.4) with 1 mM scaffold and 3 mM cyclic
were cleaved and the resulting unprotected cyclic peptide, con- peptide thioester followed by TFA-acidification and washing
taining a thioester handle, was purified by preparative HPLC with ether to remove MPAA. Within two hours almost complete
(Scheme 2). In this way, the gp120 cyclic peptide thioesters conversion to the ligated product 12 was observed. Only small
with amino acid sequences LTRDGGN (11a), INMWQEVGKA amounts of the double- and single-ligated products were
(11b) and SGGDPEIVT (11c) were obtained (Fig. 2) in 5.6–7.0% found upon LC-MS of the reaction mixture. After preparative
overall yield, corresponding to 87–90% average yield per reac- HPLC, the triple ligated product 12 was obtained in an excel-
tion step.11
lent yield of 88% (Scheme 3). To avoid interference with bio-
logical assays, the sulfhydryl group of the scaffolded cysteine
residues was capped. To achieve complete capping and to
Library synthesis of the gp120 discontinuous epitope mimics
A library, but especially a library of protein mimics, should not avoid formation of any by-products, 12 was first treated for
contain impurities, which for example can give rise to false 20 minutes with a TCEP buffer to reduce any possible disul-
positives or negatives in the screening assay. We have pre- fides, followed by addition of a sodium phosphate buffer con-
viously described the preparation of smart libraries of discon- taining iodoacetamide. After one hour, complete and clean
tinuous epitopes.7,8 Crucial characteristics of these libraries, conversion to the triple-capped product 13 was observed. Pre-
which have been obtained through CuAAC, are that the pre- parative HPLC purification afforded 13 in a quantitative yield
pared collections of library members were very clean, virtually (Scheme 3).
free of (unknown) impurities and obtained in a very reproduci-
These initial results suggested that it would be possible to
ble manner. The alternative method using NCL might be even obtain a clean collection of cyclic peptides ligated to the three
more attractive since it is devoid of any traces of Cu(I) or Cu(II). cysteine residue-containing scaffold 5. Thus, an equimolar
However, in a NCL strategy, normally a buffer system is mixture of three different cyclic peptide thioesters 11a–c,
used with a large excess of a thiol to increase the reactivity of corresponding to the discontinuous epitope of gp120, was sub-
the peptide thioester. Nowadays, 4-mercaptophenylacetic acid jected to a NCL-reaction with TAC-scaffold 5 as was described
(MPAA) is the most popular thiol additive in NCL.21 A draw- above. After acidification and washing with ether, LC-MS ana-
back of MPAA is that the HPLC retention time of this thiol is lysis showed that almost all combinations of peptide loops
in the same range as that of most peptidic compounds. There- ligated to the TAC-scaffold had been obtained (Scheme 4).
fore, when performing HPLC analysis or purification, it is After alkylation with iodoacetamide a relatively clean reaction
possible that MPAA co-elutes with the ligated product. This co- mixture of the capped protein mimics was formed, with an
elution is an even bigger risk when a library of compounds is improved separation of the peaks of the individual mimics as
prepared. Fortunately, we found that it was possible to remove compared to the starting mixture (Scheme 4).
MPAA quickly and completely by washing the TFA-acidified
Although the LC-MS chromatogram of the iodoacetamide
NCL-reaction mixture with ether.11 First a test ligation reaction capped reaction mixture showed only one peak in the chroma-
of cyclic peptide thioester 11a and tri-cysteine scaffold 5 was togram for the mimic with the three different ligated loops,
carried out in ligation buffer (200 mM sodium phosphate, it was expected that our NCL approach has yielded three
Scheme 3 Synthesis of the alkylated triple ligated mimic 13. Analytical HPLC profiles (λ = 214 nm) of the crude NCL (bottom-left) and the crude
iodoacetamide alkylation reaction (bottom-right).
This journal is © The Royal Society of Chemistry 2014
Org. Biomol. Chem., 2014, 12, 4471–4478 | 4473

View Article Online
Paper
Organic & Biomolecular Chemistry
Scheme 4 Synthesis of the TAC-scaffolded library of gp120 mimics. Analytical HPLC profiles (λ = 214 nm) of the crude NCL (middle-bottom) and
crude iodoacetamide capping reaction (right-bottom).
Fig. 3 Analytical HPLC profiles (λ = 214 nm) of the purified [1-1-1],
[1-1-3], [1-3-3], [1-1-2], [1-2-3], [2-3-3], [1-2-2] and [2-2-3] mimics.
different isomers of the [1-2-3] construct: one with peptide
loop 1 in the middle, one with peptide loop 2 in the middle
and one with peptide loop 3 in the middle. Due to the ring-
flipping ability of the TAC-scaffold for example isomer [1-2-3]
is identical to isomer [3-2-1]. It is expected that the different
Fig. 4 Results of the competitive gp120(IIIB)-CD4 ELISA for all purified
protein mimics, different cyclic peptide thioesters loops, TAC-scaffold,
equimolar mixture of the cyclic peptide thioesters and equimolar
mixture of the cyclic peptide thioesters as well as the TAC-scaffold. The
concentration of the mimics was 125 µg mL−1, which corresponds to
31 µM for the [1-2-3] mimic. The concentration of the thioester loops
(L1, L2 and L3) was 31 µM. In the mixtures each individual component
(L1, L2, L3 and TAC-scaffold) was present in a concentration of 31 µM.
isomers have similar physical properties and therefore it is
likely that the three [1-2-3] mimics also have similar HPLC
retention times. Analogously, for example the [1-1-2] purified
construct probably consists of two different compounds
(Fig. 3).
clearly showed a synergistic effect of ligating the gp120 discon-
tinuous epitope loops to the TAC-scaffold. When comparing
Activity evaluation of the gp120 discontinuous epitope mimics
Binding of the resulting gp120 discontinuous epitope mimics the different gp120 mimics, it becomes evident that the
to CD4 was studied in a gp120-capture ELISA experiment. protein mimics that contain peptide loop 3 are stronger
Most of the gp120 mimics were able to compete with the binders of CD4 than the other protein mimics. The binding
recombinant monomeric gp120(IIIB) for binding to CD4, affinity of these molecular constructs was in the same range
whereas the individual non-ligated loop thioesters, the TAC- (low micromolar) as the TAC-scaffolded gp120 mimics, which
scaffold, an equimolar mixture of loop thioesters or an equi- have been obtained using the CuAAC reaction.7 Using the
molar mixture of TAC-scaffolds together with cyclic peptide CuAAC method, protein mimics containing loop 2 and loop 3
loops showed hardly any activity (Fig. 4). This experiment were the best inhibitors of binding of recombinant monomeric
4474 | Org. Biomol. Chem., 2014, 12, 4471–4478
This journal is © The Royal Society of Chemistry 2014

View Article Online
Organic & Biomolecular Chemistry
Paper
gp120(IIIB) to CD4. In contrast, the [1-2-2] and [2-2-3] NMR spectra were recorded on a Varian VNMRS400 400 MHz
gp120 mimics obtained in this work showed hardly any spectrometer in DMSO-d6 as the solvent at 373 K because at
binding to the CD4 receptor. This may indicate that the cyclic room temperature very broad signals were observed, most
peptide 11b representing loop 2 in this work is a less ideal likely due to the cis/trans isomerization of the tertiary amides.
mimic of the natural 424INMWQEVGKA433 loop, as compared Chemical shifts (δ) are reported in parts per million (ppm)
to the cyclic peptide used for representing loop 2 in the CuAAC relative to DMSO-d6 (2.50 ppm). Splitting patterns are desig-
method. The observation that the [1-2-3] protein mimic con- nated as singlets (s) and multiplets (m). 13C NMR spectra were
taining all three loops of the discontinuous epitope of gp120 recorded on a Varian VNMRS400 400 MHz spectrometer at
did not show the highest inhibition in this ELISA assay might 100 MHz in DMSO-d6 as the solvent. Chemical shifts (δ) are
be indicative of a non-optimal positioning of the different reported in parts per million (ppm) relative to the solvent
peptide loops on the scaffold. Moreover, the molecular con- residual signal, DMSO-d6 (39.52 ppm). 2-D NMR data (HSQC
struct [1-2-3] likely consists of three isomers. An approach for and COSY) were acquired on a Varian VNMRS400 400 MHz
synthesising the individual isomers thereby enabling the syn- spectrometer. Analytical HPLC was accomplished on a Shi-
thesis of the different isomers of the most active molecular madzu-10Avp (Class VP) with a UV-detector operating at 214
construct that is the [1-3-3] mimic, and of course of the [1-2-3] and 254 nm using a Alltech Prosphere C4 column (300 Å,
mimic will be reported soon and should enable evaluation of 5 μm, 250 × 4.60 mm, column A) or a Phenomenex Gemini
the importance of the relative positioning of the peptide loops C18 column (110 Å, 5 μm, 250 × 4.60 mm, column B) at a flow
on the scaffold.
rate of 1 mL min−1 using a standard protocol: 100% buffer A
for 5 min, followed by a linear gradient of buffer B (0–100% in
20 min, method A) or 100% buffer A for 2 min, next a linear
gradient of buffer B (0–100% in 48 min, method B). The
mobile phase was H2O–CH3CN–TFA (95 : 5 : 0.1, v/v/v, buffer A)
Conclusions
In conclusion, we have developed a native ligation method for and H2O–CH3CN–TFA (5 : 95 : 0.1, v/v/v, buffer B). Purification
obtaining molecular constructs containing multiple cyclic pep- of the peptidic compounds was performed on Prep
a
tides on a scaffold; these may act as protein mimics capable of LCMS-QP8000α HPLC system (Shimadzu) using a Alltech Pro-
mimicry of discontinuous epitopes. The required tri-cysteine sphere C4 column (300 Å, 10 μm, 250 × 22 mm, column A) or a
containing the TAC-scaffold was conveniently synthesized in a Phenomenex Gemini C18 column (110 Å, 10 μm, 250 × 20 mm,
high yield. Native chemical ligation using cyclic peptide thio- column B) at a flow rate of 12.5 mL min−1 using a standard
esters followed by capping was both a highly efficient and very protocol: 100% buffer A for 5 min, followed by a linear gradi-
clean reaction, by which to obtain these relatively complex ent of buffer B (0–60% in 100 min, method A) or 100% buffer
molecular constructs of intermediate size in a both straightfor- A for 5 min, followed by a linear gradient of buffer B (0–100%
ward and reproducible manner. Several of the obtained in 20 min, method B) using the same buffers as described for
protein gp120 discontinuous epitope mimics showed an inhi- analytical HPLC. Analytical LC-MS was performed on
a
bition of the recombinant monomeric gp120(IIIB) – CD4 Thermo-Finnigan LCQ Deca XP Max ion trap mass spectro-
binding. These library members may be used as promising meter using the same buffers and protocols as described for
starting points for optimization studies of these protein analytical HPLC. Routine electrospray ionization mass spec-
mimics. Finally, the library approach developed for the con- trometry (ESI-MS) was performed on
a
Shimadzu
struction of protein mimics might represent a promising tool LCMS-QP8000 single quadrupole bench-top mass spectro-
for constructing mimics of other discontinuous epitopes.
meter operating in a positive ionization mode or on a Thermo-
Finnigan LCQ Deca XP Max ion trap mass spectrometer. The
MALDI-TOF-MS spectrum was recorded on a Kratos Analytical
(Shimadzu) AXIMA CFR mass spectrometer using sinapic acid
as a matrix and bovine insulin oxidized B chain (monoisotopic
[M + H]+ 3494.6513) as a reference. High-resolution electro-
Experimental
General information
All reagents were obtained from commercial sources and used spray ionization (ESI) mass spectra were measured on a Bruker
without further purification. Peptide grade DiPEA and TFA micrOTOF-Q II in positive mode and calibrated with ESI
were purchased from Biosolve B.V. (Valkenswaard, The Nether- tuning mix from Agilent Technologies. The pH was measured
lands) and CH2Cl2, NMP and HPLC grade solvents were pur- using a PHM210 standard pH meter from Radiometer Analyti-
chased from Actu-All (Oss, The Netherlands). Lyophilizations cal equipped with a micro combined pH electrode pHC3359-8
were performed on Christ Alpha 1–2 apparatus. Reactions were from Radiometer Analytical. The microtiter plate reader used
carried out at ambient temperature unless stated otherwise. in the ELISA experiments was a BioTEK µQuant (Beun de
Solvents were evaporated under reduced pressure at 40 °C. Ronde, Abcoude, The Netherlands). Software used for data
Reactions in solution were monitored by TLC analysis and Rf analysis was the Full Mode-KC4 version 3.4.
values were determined on Merck pre-coated silica gel 60
(oNBS)3-TAC(N(H)Me) 2. Tri-oNBS protected TAC-scaffold
F-254 (0.25 mm) plates. Spots were visualized by UV-light and 115 (3.11 g, 3.74 mmol, 1 eq.) and BOP (1.82 g, 4.11 mmol, 1.1
by heating plates after dipping in a ninhydrin solution. 1H eq.) were dissolved in DCM (25 mL) at room temperature. Next,
This journal is © The Royal Society of Chemistry 2014
Org. Biomol. Chem., 2014, 12, 4471–4478 | 4475

View Article Online
Paper
Organic & Biomolecular Chemistry
DiPEA (2.61 mL, 15.0 mmol, 4 eq.) and methylamine (2 M 127.4, 127.4, 128.0, 128.2, 128.7, 128.8, 129.5, 130.1, 135.6,
in THF, 9.35 mL, 18.7 mmol, 5 eq.) were added and the reac- 137.8, 144.1 (Ar–C), 154.1, 154.5 (3 × CvO Boc), 165.9 (CvO
tion mixture was stirred overnight. After the evaporation of the TAC), 169.2, 169.9 (3 × CvO Cys).
solvents in vacuo, the crude product was dissolved in EtOAc
(H-Cys)3-TAC(N(H)Me) 5. The TAC-scaffold 4 (150 mg, 92.2
(100 mL) and washed twice with 1 N KHSO4 (both 100 mL) and µmol) was dissolved in a solution of TFA–H2O–TIPS–EDT
twice with 5% NaHCO3 (both 100 mL) and brine (100 mL). (10 mL, 90 : 5 : 2.5 : 2.5, v/v/v/v) for 2 h at room temperature.
Finally, the organic layer was dried over Na2SO4 and evapor- The volume of the reaction mixture was reduced to approxi-
ated in vacuo, yielding the relatively clean amidated scaffold 2, mately 2 mL by evaporation. Next, this mixture was added drop
which was used as such for the preparation of 3.
wise to a cold (4 °C) solution of MTBE–hexanes (1 : 1, v/v).
TAC(N(H)Me) 3. NaH (60% dispersion in mineral oil, 1.35 g, After centrifugation (3500 rpm, 5 min) the supernatant was
33.7 mmol, 9 eq. (3 eq. per oNBS group)) was added slowly to a decanted and the pellet was resuspended in MTBE–hexanes
solution of thiophenol (3.82 mL, 37.4 mmol, 10 eq. (3.33 eq. (1 : 1, v/v) and centrifuged again. This was repeated twice after
per oNBS group)) in DMF (45 mL) at room temperature. After which the pellet was dissolved in water containing H2O–TFA
ca. 30 min stirring at room temperature, bubbling of H2 gas (100 : 0.01, v/v) and purified by preparative HPLC (column A,
has subsided and (oNBS)3-TAC(N(H)Me) 2 (3.74 mmol, 1 eq.), method A, using H2O–TFA (100 : 0.01, v/v) as buffer A). Frac-
dissolved in DMF (5 mL), was added to the above freshly pre- tions corresponding to 5 were pooled and lyophilized to yield
pared sodium thiophenolate solution. Upon this addition, the the TFA salt of the tri-cysteine TAC-scaffold 5 as a white fluffy
homogeneous, yellow reaction mixture became a brown-red solid (64.6 mg, 74%, 3 TFA molecules were added to the mole-
suspension. After stirring overnight at room temperature, the cular weight of the scaffold). tR = 14.1 min (column A, method
solvents were evaporated in vacuo and DCM (90 mL) and Et2O A, using H2O–TFA (100 : 0.01, v/v) as buffer A)); HRMS m/z
saturated with HCl (75 mL) were added to the crude product, calcd for C25H42N7O4S3 [M + H]+ 600.2455, found 600.2493.
yielding a yellowish suspension, which was filtered over a glass
Triple-ligated product 12
filter. The residue was washed with DCM and dried in vacuo,
yielding the crude unprotected TAC-scaffold amide 3 as a N2 was bubbled through a freshly prepared solution of
white solid, which was directly used in the next reaction. NaH2PO4·2H2O (200 mM, 62.4 mg), MPAA (200 mM, 67.3 mg),
Although the crude product contained approximately 9 eq. of and TCEP (40 mM, 22.9 mg), pH 7.4 in 2 mL of H2O, for
NaCl, it was not expected that this would give rise to problems 5 minutes. Next, this solution (2 mL) was added to solid cyclic
in the next reaction step.
peptide thioester 111 (compound 11a, 7.55 mg, 6 µmol, final
(Boc-Cys(Trt))3-TAC(N(H)Me) 4. To a solution of Boc-Cys- conc.: 3 mM, the molecular weight of 11a was assumed to
(Trt)-OH (5.20 g, 11.2 mmol, 3 eq. (1 eq. per amine of the TAC- include 2 molecules of TFA) and triple-cysteine TAC-scaffold 5
scaffold)) and BOP (4.96 g, 11.2 mmol, 3 eq. (1 eq. per amine (1.88 mg, 2 µmol, final conc.:1 mM, the molecular weight of 5
of the TAC-scaffold)) in DCM (40 mL), DiPEA (7.8 mL, was assumed to include 3 molecules of TFA) and the reaction
45 mmol, 12 eq. (4 eq. per amine of the TAC-scaffold)) was mixture was stirred at room temperature. After 2 hours the
added. Next, crude TAC(N(H)Me) 3 (3.74 mmol, 1 eq.) was reaction mixture was diluted with H2O–TFA (9 : 1, v/v, 2 mL) to
added to the reaction mixture, yielding a white suspension. precipitate MPAA. After two washing steps with Et2O (4 mL) to
The reaction mixture was stirred overnight at room tempera- remove MPAA, TCEP (ca. 10 mg) was added and the solution
ture, followed by the evaporation of the solvents in vacuo. Next, was stirred for 5 min. After two additional washing steps with
the crude product was dissolved in EtOAc (100 mL) and Et2O (4 mL), purification, using preparative HPLC (column B,
washed twice with 1 N KHSO4 (both 100 mL) and twice with method A), was performed. Fractions corresponding to the
5% NaHCO3 (both 100 mL) and brine (100 mL). Finally, the ligated product 12 were pooled and lyophilized to yield the
organic layer was dried over Na2SO4, evaporated in vacuo and triple-ligated product 12 as a white fluffy solid (7.00 mg, 88%,
the product was purified by column chromatography its molecular weight was assumed to include 6 TFA molecules).
(hexanes–EtOAc, 1 : 1, v/v) to afford the TAC-scaffold (Boc-Cys- tR = 17.9 min (column B, method B); MALDI-TOF MS m/z calcd
(Trt))3-TAC(N(H)Me) 4 as a white foam (4.85 g, 80% over 3 for C133H213N46O46S3 [M + H]+ 3286.49, found 3286.49.
steps). Rf = 0.26 (Hexanes–EtOAc, 1 : 1, v/v); HRMS m/z calcd
Triple-capped product 13
for C97H108N7O10S3 [M + H]+ 1626.7314, found 1626.7358;
1H-NMR (400 MHz, DMSO-d6) δ: 0.86, 1.22 (2 m, 4H, 2 × The triple-ligated product 12 (1.00 mg, 0.252 µmol, MW.6TFA)
CH2CH2CH2), 1.33, 1.34 (2 s, 27H, 3 × C(CH3)3), 2.34, 2.55 was dissolved in 252 µL of a freshly prepared solution of
(2 m, 6H, 3 × CHCH2S), 2.72, 3.02, 3.24 (3 m, 8H, 4 × NaH2PO4·2H2O (200 mM, 62.4 mg) and TCEP (40 mM,
NCH2CH2), 2.78 (s, 3H, CvONHCH3), 4.11–4.67 (2 m, 7H, 3 × 22.9 mg), pH 7.6 in 2 mL of H2O and the mixture was stirred
CHCH2S and 2 × PhCH2), 6.37, 6.73 (2 m, 3H, 3 × NHBoc), for 20 minutes at room temperature. Next, 252 µL of a freshly
7.20–7.33, 7.69 (2 m, 48H, Ar–CH), 7.87 (m, 1H, CvONHCH3). prepared solution of NaH2PO4·2H2O (200 mM, 62.4 mg) and
13C-NMR (100 MHz, DMSO-d6): 25.7 (CvONHCH3), 26.9 (3 × iodoacetamide (60 mM, 22.2 mg), pH 7.6 in 2 mL of H2O was
CH2CH2CH2), 27.7 (3 × C(CH3)3), 33.7, 33.8 (3 × CHCH2S), 44.5 added to the reaction mixture. After stirring for 1 hour in the
(4 × NCH2CH2), 49.7, 50.5 (3 × CHCH2S), 50.7 (2 × PhCH2), dark, the mixture was directly transferred and purified by pre-
66.0, 66.3 (3 × C(Ph)3), 78.3, 78.4 (C(CH3)3), 126.2, 126.3, 126.9, parative HPLC (column B, method A). Fractions corresponding
4476 | Org. Biomol. Chem., 2014, 12, 4471–4478
This journal is © The Royal Society of Chemistry 2014

View Article Online
Organic & Biomolecular Chemistry
Paper
to 13 were pooled and lyophilized to yield the alkylated
[1-2-3] mimic: tR = 20.8 min (column B, method B); HRMS
triple-loop 1 mimic 13 as a white fluffy solid (1.15 mg, quant, m/z calcd for C171H267N50O56S4 [M + 3H]3+ 1348.2822, found
MW.6TFA). tR = 16.5 min (column B, method B); HRMS m/z calcd 1348.2861.
for C139H224N49O49S3 [M + 3H]3+ 1153.1902, found 1153.1889.
[2-3-3] mimic: tR = 21.7 min (column B, method B); HRMS
m/z calcd for C179H276N48O60S4 [M + 3H]3+ 1395.2968, found
1395.6229.
General native chemical ligation procedure of (H-Cys)3-
TAC(N(H)Me) 5 and cyclic peptide thioesters 11a–c
N2 was bubbled through a freshly prepared solution of m/z calcd for C187H290N55O55S5 [M + 3H]3+ 1448.6730, found
[1-2-2] mimic: tR = 22.1 min (column B, method B); HRMS
NaH2PO4·2H2O (200 mM, 124.8 mg), MPAA (200 mM, 1448.6644.
134.6 mg) and TCEP (40 mM, 45.9 mg), pH 7.4 in 4 mL of
[2-2-3] mimic: tR = 22.9 min (column B, method B); HRMS
H2O, for 5 minutes. Next, this solution (4 mL) was added to a m/z calcd for C195H300N53O59S5 [M + 3H]3+ 1496.0236, found
falcon tube, containing cyclic peptide thioester 111 (compound 1496.0155.
11a, 5.03 mg, 4 µmol, 1 mM, MW.2TFA), cyclic peptide thio-
ester 211 (compound 11b, 6.35 mg, 4 µmol, 1 mM, MW.1TFA),
HIV-1 gp120 capture ELISA
cyclic peptide thioester 311 (compound 11c, 4.69 mg, 4 µmol,
The HIV-1 gp120 capture ELISA experiments were performed
1 mM) and tri-cysteine TAC-scaffold 5 (3.76 mg, 4 µmol, 1 mM,
according to literature procedures.7,8 Before binding of the
MW.3TFA) and the resulting reaction mixture was stirred at
TAC-scaffold was studied in this ELISA experiment, the tem-
room temperature. After 3 hours TFA (500 µL) was added to
plate was first alkylated with iodoacetamide according to the
the reaction mixture to precipitate MPAA. After two wash steps
above described method. tR = 13.3 min (column A, method A,
with Et2O (each 5 mL), TCEP (ca. 10 mg) was added and the
using H2O–TFA (100 : 0.01, v/v) as buffer A). MS (ESI) m/z calcd
solution was stirred for 5 min. After two additional wash steps
for C31H51N10O7S3 [M + H]+ 771.31, found 771.25; calcd for
with Et2O (each 5 mL), a sample was taken from the reaction
C31H50N10NaO7S3 [M + Na]+ 793.30, found 793.40.
mixture for LC-MS analysis (column B, method B) to verify the
absence of MPAA. Finally, the reaction mixture was directly
purified by a fast preparative HPLC run (column B, method B)
and all peptide fractions were pooled and lyophilized to yield
Acknowledgements
collection 14 as a white fluffy solid, which was directly used in
the next reaction.
We thank Ing. Javier Sastre Toraño for the HRMS measure-
ments and Dr Johann Jastrzebski for the help with the variable
temperature NMR measurement. This research was (partly)
financed by Chemical Sciences of The Netherlands Organi-
zation for Scientific Research (NOW).
General procedure for the alkylation of the sulfhydryl group of
the cysteine residues in protein mimics with iodoacetamide
Collection 14 (4 µmol, 3 cysteine residues per molecule) was
dissolved in a freshly prepared solution of NaH2PO4·2H2O
(200 mM, 124.8 mg) and TCEP (40 mM, 45.9 mg), pH 7.6 in
4 mL of H2O and was stirred for 20 minutes at room tempera-
Notes and references
ture. Next, a freshly prepared solution of NaH2PO4·2H2O
(200 mM, 124.8 mg) and iodoacetamide (60 mM, 44.4 mg), pH
7.6 in 4 mL of H2O was added to the reaction mixture. After
stirring for 1 hour in the dark, a sample was taken from the
reaction mixture for LC-MS analysis (column B, method B).
Finally, the reaction mixture was lyophilized, dissolved in
H2O–CH3CN–TFA (75 : 25 : 0.01, v/v/v) and the individual
components were purified by preparative HPLC (column B,
method A).
1 P. R. Werkhoven and R. M. J. Liskamp, in Biotherapeutics,
RSC Drug Discovery Series No. 36, ed. L. H. Jones and
A. J. McKnight, The Royal Society of Chemistry, 2013, vol. 263.
2 M. A. McLafferty, R. B. Kent, R. C. Ladner and
W. Markland, Gene, 1993, 128, 29.
3 C. Heinis, T. Rutherford, S. Freund and G. Winter, Nat.
Chem. Biol., 2009, 284, 502.
4 M. Galibert, O. Renaudet, P. Dumy and D. Boturyn, Angew.
Chem., Int. Ed., 2001, 50, 1901.
[1-1-1] mimic: tR = 16.5 min (column B, method B); HRMS
m/z calcd for C139H224N49O49S3 [M + 3H]3+ 1153.1902, found
1153.1889.
[1-1-3] mimic: tR = 17.9 min (column B, method B); HRMS
m/z calcd for C147H234N47O53S3 [M + 3H]3+ 1200.5407, found
1200.5335.
[1-3-3] mimic: tR = 18.9 min (column B, method B); HRMS
m/z calcd for C155H244N45O57S3 [M + 3H]3+ 1247.8913, found
1247.8859.
[1-1-2] mimic: tR = 19.9 min (column B, method B); HRMS
m/z calcd for C163H257N52O52S4 [M + 3H]3+ 1300.9316, found
1300.9133.
5 A. Berthelmann, J. Lach, M. A. Gräwert, M. Groll and
J. Eichler, Org. Biomol. Chem., 2014, 12, 2606.
6 H. Li, Y. Guan, A. Szczepanska, A. J. Moreno-Vargas,
A. T. Carmona, I. Robina, G. K. Lewis and L.-X. Wang,
Bioorg. Med. Chem., 2007, 15, 4220.
7 G. E. Mulder, J. A. W. Kruijtzer and R. M. J. Liskamp, Chem.
Commun., 2012, 48, 10007.
8 G. E. Mulder, H. C. Quarles van Ufford, J. van Ameijde,
A. J. Brouwer, J. A. W. Kruijtzer and R. M. J. Liskamp, Org.
Biomol. Chem., 2013, 11, 2676.
9 P. E. Dawson, T. W. Muir, I. Clark-Lewis and S. B. H. Kent,
Science, 1994, 266, 776.
This journal is © The Royal Society of Chemistry 2014
Org. Biomol. Chem., 2014, 12, 4471–4478 | 4477

View Article Online
Paper
Organic & Biomolecular Chemistry
10 T. M. Hackeng, J. H. Griffin and P. E. Dawson, Proc. Natl. 16 N. Shangguan, S. Katukojvala, R. Greenberg and
Acad. Sci. U. S. A., 1999, 96, 10068. L. J. Williams, J. Am. Chem. Soc., 2003, 125, 7754.
11 H. van de Langemheen, A. J. Brouwer, J. Kemmink, 17 R. Merkx, A. J. Brouwer, D. T. S. Rijkers and R. M. J. Liskamp,
J. A. W. Kruijtzer and R. M. J. Liskamp, J. Org. Chem., 2012,
77, 10058.
12 P. D. Kwong, R. Wyatt, J. Robinson, R. W. Sweet, J. Sodroski
and W. Hendrickson, Nature, 1998, 393, 648.
13 R. Franke, T. Hirsch, H. Overwin and J. Eichler, Angew.
Chem., Int. Ed., 2007, 46, 1253.
Org. Lett., 2005, 7, 1125.
18 D. T. S. Rijkers, R. Merkx, C. B. Yim, A. J. Brouwer and
R. M. J. Liskamp, J. Pept. Sci., 2010, 16, 1.
19 C. B. Yim, I. Dijkgraaf, R. Merkx, C. Versluis, A. Eek,
G. E. Mulder, D. T. S. Rijkers, O. C. Boerman and
R. M. J. Liskamp, J. Med. Chem., 2010, 53, 3944.
14 C. Chamorro, J. A. W. Kruijtzer, M. Farsaraki, J. Balzarini 20 R. Merkx, M. J. van Haren, D. T. S. Rijkers and
and R. M. J. Liskamp, Chem. Commun., 2009, 821. R. M. J. Liskamp, J. Org. Chem., 2007, 72, 4574.
15 M. C. F. Monnee, A. J. Brouwer and R. M. J. Liskamp, QSAR 21 E. C. B. Johnson and S. B. H. Kent, J. Am. Chem. Soc., 2006,
Comb. Sci., 2004, 23, 546.
128, 6640.
4478 | Org. Biomol. Chem., 2014, 12, 4471–4478
This journal is © The Royal Society of Chemistry 2014