Addition of HOBt improves the conversion of thioester-Amine chemical ligation

Article abstract of DOI:10.1002/bip.22745

The syntheses of large peptides and of those containing non-natural amino acids can be facilitated by the application of convergent approaches, dissecting the native sequence into segments connected through a ligation reaction. We describe an improvement of the ligation protocol used to prepare peptides and proteins without cysteine residues at the ligation junction. We have found that the addition of HOBt to the ligation, improves the conversion of the ligation reaction without affecting the epimerization rate or chemoselectivity, and it can be efficiently used with peptides containing phosphorylated amino acids.

Full text of DOI:10.1002/bip.22745

Addition of HOBt Improves the Conversion of Thioester-Amine
Chemical Ligation
1
1
Toni Todorovski, David Sunol, Antoni Riera,1,2 Maria J. Macias1,3
~
1Institute for Research in Biomedicine (IRB Barcelona), The Barcelona Institute of Science and Technology,
Baldiri Reixac, 10 08028 Barcelona, Spain
2
`
Departament de Quꢀımica Orgꢁanica, University of Barcelona, Martꢀı i Franques, 1, Barcelona 08028, Spain
3Catalan Institution for Research and Advanced Studies (ICREA), Passeig Lluis Companys, Barcelona 23 08010, Spain
Received 6 May 2015; revised 10 September 2015; accepted 18 September 2015
Published online 22 September 2015 in Wiley Online Library (wileyonlinelibrary.com). DOI 10.1002/bip.22745
This article was originally published online as an accepted
preprint. The “Published Online” date corresponds to the pre-
ABSTRACT:
The syntheses of large peptides and of those containing
non-natural amino acids can be facilitated by the appli-
cation of convergent approaches, dissecting the native
sequence into segments connected through a ligation reac-
tion. We describe an improvement of the ligation protocol
used to prepare peptides and proteins without cysteine
residues at the ligation junction. We have found that the
addition of HOBt to the ligation, improves the conversion
of the ligation reaction without affecting the epimeriza-
tion rate or chemoselectivity, and it can be efficiently used
print version. You can request a copy of any preprints from
the past two calendar years by emailing the Biopolymers edi-
torial office at biopolymers@wiley.com.
INTRODUCTION
olid-phase peptide synthesis (SPPS), developed in
the 60s by B. Merrifield,1 is nowadays the standard
methodology for chemical preparation of peptides.
Although peptides up to 50 residues can be obtained
2,3
S
using this methodology
the linear synthesis of
large or complex sequences frequently yield mixtures con-
taining undesired products and is usually not practical.
Improvement of these syntheses often require the application
of a convergent approach, dissecting the native sequence in
segments that are connected using a ligation reaction (Figure
1, Panel A).4–6 These protocols also facilitate the introduc-
tion of modified amino acids in proteins (including phos-
phorylated or acetylated residues or non-natural amino
acids) as well as to the addition of specific labels for antibody
recognition, protein detection and cell imaging.7
with peptides containing phosphorylated amino acids.
C
V
2015 Wiley Periodicals, Inc. Biopolymers (Pept Sci) 104: 693–702, 2015.
Keywords: ligation of peptide fragments through direct
aminolysis; cysteine-free native chemical ligation; native
chemical ligation with b-branched residues at ligation
junction; formation of highly reactive HOBt esters
Additional Supporting Information may be found in the online version of this
article.
The direct coupling of two peptides was reported for the
first time by Kemp et al. in 1974.8,9 This method allows a liga-
tion reaction to occur independently of the amino acids pairs
at the ligation junction. The reaction couples a peptide con-
taining an activated carboxylic derivative at its C-terminus to a
second peptide carrying an unprotected NH2-group at its N-
terminus. The success of the method depends on the aqueous
solubility of the designed segments and suffers from the epi-
merization of the product at the ligation site. Improvements to
the direct aminolysis method were described later by Blake
Correspondence to: Maria J. Macias, Institute for Research in Biomedicine (IRB
Barcelona), The Barcelona Institute of Science and Technology, Baldiri Reixac, 10,
08028 Barcelona, Spain;
e-mail: maria.macias@irbbarcelona.org
Contract grant sponsor: European Union’s Seventh Framework Programme for
research, technological development and demonstration
Contract grant number: IRBPostPro 246557
Contract grant sponsor: Spanish National Research Program
Contract grant number: SAF2011-25119, BFU2014-53787-P, CTQ2014-56361,
2009SGR00901
Contract grant sponsor: Consolider
Contract grant number: CSD2009-00080
C
V
2015 Wiley Periodicals, Inc.
PeptideScience Volume 104 / Number 6
693

694
Todorovski et al.
FIGURE 1 Schematic representation of the ligation reaction through direct aminolysis (Panel A)
Peptides designed for the ligation reactions used in this study. The m/z experimental values (which
are same as theoretically calculated ones) obtained for each peptide is shown next to the corre-
sponding sequences prior the ligation (Panel B). Ligated peptides sequences (the ligation site is rep-
resented with a circle) are listed, with yields and conversion percentages obtained for ligation
reactions carried in the presence or in the absence of HOBt (Panel C).
et al. and Aimoto et al.10–12 However, in these protocols the carrying the thioester group at the C-terminus via intein-
reaction conditions cause epimerization of the C-terminal mediated cleavage.13 A drawback in this process is that proteins
thioester residue, and therefore only non-epimerizable amino synthesized by NCL are generated mainly from two segments
acids (glycine or proline) can be introduced at this position.
and are often limited by the presence of cysteines in the
A breakthrough in the ligation strategy was the develop- sequence. If cysteines are not present in the native sequence,
ment of the native chemical ligation (NCL) approach.4 In this the ligated peptide would require a desulfurization step which
reaction, carried out in aqueous buffers, the peptide containing is not always straightforward.14,15
a C-terminal thioester is attacked by the thiol group of a cyste-
Novel approaches have recently been described by Wong
ine residue located at the N-terminus of the second peptide. and coworkers, that partially overcome the above mentioned
The thioester intermediate forms a native amide bond after an drawback.16,17 Their works report the chemical ligation of pep-
intramolecular rearrangement. This ligation reaction can also tides and proteins without cysteine residues at the ligation
be used with recombinant proteins to generate the segment junction with high chemoselectivity and low levels of
Biopolymers (Peptide Science)

Improvement of peptide ligation using HOBt
695
without the side-chain protecting groups. The peptide thioesters were
then precipitated in cold diethyl ether and centrifuged (3000g). The
pellet was washed twice with cold ether, dried, and stored at 2208C.
The crude thioesters were analyzed by LC-MS and, depending on
their purity, some of them were additionally purified by preparative
RP-HPLC prior to their use in the ligation reactions.
epimerization. The main limitation of these strategies is the
ligation rate, normally requiring long reaction times, ranging
from a few hours to several days.
We considered that adding 1-hydroxybenzotriazole (HOBt),
a common reagent used in SPPS, to the protocol described by
Wong and coworkers,16 could improve the ligation yield and
the conversion of the ligation reaction without affecting the
epimerization rate or chemoselectivity. To test this hypothesis,
we prepared various peptides with different amino acids at the
ligation junction including a WW domain sequence (of 35 aa)
containing at non-natural residue. Some of the peptides
included phosphorylated amino acids, which are often a chal-
lenge due to the steric hindrances that they introduce during
couplings, as well as their tendency to undergo alpha, beta-
elimination under strong reaction conditions. Our results indi-
cate that, the presence of HOBt increases the yield and conver-
sion of the ligation, thus reducing the possibility of side-
products and degradation. These experimental conditions are
especially suited for the preparation of intrinsically unfolded
protein segments that are particularly prone to degradation.
General Procedure for Free N-Terminus Peptide Synthesis.
Peptides containing a free N-terminus (Figure 1, Panel B) were syn-
thesized manually using a Rink amide AM resin and Fmoc/tBu chem-
istry with 5 equiv. of the amino acids derivatives activated by 4.9
equiv. of DIC in the presence of 4.9 equiv. of HOBt in DMF at the
100-lmol scale. The efficiency of the manual coupling was verified by
the Kaiser test.19
The resin-bound peptides were cleaved and deprotected with TFA
containing a scavenger mixture of water, thioanisol, and ethanedithiol
(90:5:2.5:2.5 by vol) at RT for 2 h. They were then precipitated in cold
diethyl ether and centrifuged (3000 g). The pellet was washed twice
with cold ether, dried and stored at 2208C.
Finally, the crude peptides were purified by preparative RP-HPLC.
Pure fractions were collected, lyophilized, and stored at 2208C. They
were then analyzed by MALDI-MS and LC-MS prior to their use in
ligation reactions.
General Procedure for Peptide Ligation Reactions. The free
N-terminus peptides (1.5 equiv. or 0.5 equiv.) were dissolved in
120 lL of ligation buffer (Ligation buffer5 NMP:6M GnHCL 1 1M
HEPES 5 4:1 (v/v). A buffer containing 6 M GuHCl and 1M HEPES
was prepared (50 mL) and adjusted to pH 8.5 using 25% NaOH solu-
tion and degassed. For each ligation trial 100 lL of the buffer was mixed
with 400 lL of NMP). The resulting solution was used to dissolve the
thioester peptide (between 0.3 and 1.5 lmol) (Depending of the pep-
tide thioester amount the final volume of the ligation buffer was different
(the volume of 80 lL and the described procedure is for peptide thioester
amount of 1 lmol). However, the free NH2 peptide was always 1.5 equiv.
higher (Table I, entry 1, 2, 5, 6 and 7) or 0.5 equiv. lower (Table I, entry
3, 4 and 8) from the peptide thioester equivalents). For each reaction,
the solution was separated into two equal parts, and 15 lL of HOBt
dissolved in the same ligation buffer (2 equiv. based on the amount of
peptide thioester) were added to the ligation series with HOBt, while
15 lL of the ligation buffer was added to the other part, for compari-
son, as a control. For the ligations with HOAt and DIPEA, 2 equiv of
HOAt, or, 4 equiv of DBU and DIPEA were added to the ligation mix-
ture, respectively.
MATERIALS AND METHODS
General Procedure for the Synthesis of Peptide Thioest-
ers. Six peptide thioesters (Figure 1, Panel B) were synthesized using
the SPPS approach and 4-sulfamylbutyryl resin (Supporting Informa-
tion). The coupling of the first residue was performed as recom-
mended by the manufacturer.18 In brief: Fmoc-AA-OH (4 equiv.),
DIC (4 equiv.) and 1-methylimidazole (4 equiv.) were dissolved in
DCM (12 lL/lmol). The solution was added to 1 equiv. of the resin
and the mixture was gently agitated for the next 18 h at 258C. The
resin was then washed 5 times with DMF and 5 times with DCM
before being dried. Coupling efficiency was estimated through resin
load determination by UV analysis of Fmoc-release. The remaining
amino acids were incorporated manually using 5 equiv. of the corre-
sponding amino acid derivatives activated with 4.9 equiv. of DIC in
the presence of 4.9 equiv. of HOBt in DMF at 100 lmol scale using
Fmoc/tBu chemistry. The efficiency of the couplings was verified with
the Kaiser test.19
After completion of the sequences, all peptides were either acety-
lated or Boc-protected at the N-terminus prior to final peptide cleav-
age from the resin. The procedure for the peptide activation was the
following: iodoacetonitrile (67 equiv.) and DIPEA (13 equiv.) in DMF
(72 lL/lmol) were added to the resin and gently agitated for the next
18 h. The resin was then washed six times with DMF, six times with
DCM and then dried. In the next step, the peptide was cleaved from
the activated resin using benzyl mercaptan (50 equiv.) and DIPEA (13
equiv.) in DMF (72 lL/lmol) in an 18 h reaction with gentle agita-
tion. The peptide solutions were filtered and collected in a round-
bottom flask. The resin was washed twice with DMF (4 mL each
time) and the combined filtrates were concentrated under vacuum in
order to completely remove DMF. The crude product was dissolved in
Finally, thiophenol (2% by volume, 1.5 lL) was added to the reac-
tions in the presence or absence of activators, and the resulting mix-
ture was incubated at 378C with gentle agitation until the reaction
was completed. At various time points the reaction was followed by
LC-MS, and in some cases by MALDI-MS and RP-HPLC, depending
on the peptide sequences and amino acids at the ligation junction
(Table I, Supporting Information Table I). At each time point, 8 lL
aliquots of the ligation mixture were taken and quenched by the addi-
tion of 0.1% TFA in water (32 lL) or tris(2-carboxyethyl)phosphine
(TCEP) solution (32 lL, of a 10 mg/mL solution) when the products
contained cysteine residues. The quenched mixtures were diluted up
6 mL of a TFA/triisopropylsilane/water mixture (95:2.5:2.5 by vol.) to 1.5 mL with % MeCN/0.1% FA in water (Supporting Information
and stirred at room temperature for 3 h, yielding the peptide thioester Table I) and stored at 2208C.
Biopolymers (Peptide Science)

696
Todorovski et al.
Table I Ligation Combinations Used in the Study With the Experimental Masses for the Product and Exchanged Thiophenol
Thioester Used for Monitoring the Reaction Kinetics
Experimental m/z of
Exchanged Thiophenol
Thioester SC7H7»
SC6H5
Yield (Y)/conv.(C)
With HOBt
addition (%)
Yield (Y)/conv.(C)
Without HOBt
Addition (%)
Experimental m/z
of the Product
Ligation Combinations
1
2
3
4
5
6
7
8
9
10
Ac-GASATVSPLG-SC7H7 1
NH2-GGPSPLGFLG-CONH2
Ac-GASATVSPLS-SC7H7 1
NH2-GGPSPLGFLG-CONH2
Ac-LYRAG-SC7H7 1
[M1H]151783.4
[M12H]215892.2
[M1H]151813.2
[M12H]215907.1
[M1H]151168.5
[M1H]15993.5
[M1H]151023.5
[M1H]15713.5
[M1H]15713.5
[M1H]151023.5
[M1H]151035.5
[M1H]151035.5
[M1H]15713.5
[M1H]15844.8
n.d
n.d
n.d
n.d
89 (Y)
n.d
83 (Y)
n.d
NH2-GSPGYS-CONH2
Ac-LYRAG-SC7H71
NH2-ASPGYS-CONH2
[M1H]151182.5
Ac-GASATVSPLS-SC7H71
NH2-VGPSPLGFLG-CONH2
Ac-GASATVSPLV-SC7H71
NH2-VGPSPLGFLG-CONH2
Ac-GASATVSPLV-SC7H71
NH2-LGPSPLGFLG-CONH2
Ac-LYRAG-SC7H71
[M1H]151855.4
[M12H]215928.2
[M1H]151867.2
[M12H]215934.1
[M1H]151881.4
[M12H]215941.2
[M1H]151274.6
36 (Y)
34 (Y)
36.7 (C)
57 (Y)
31.5 (C)
n.d.
10 (Y)
5.2 (Y)
16.7 (C)
33 (Y)
18.8 (C)
n.d
NH2-YSPGYS-CONH2
Ac-PSpSPGSV-SC7H71
NH2-LARPSVI-CONH2
Ac-GASATVSPLS-SC7H71
NH2-CGPSPLGFLG-CONH2
[M1H]151488.6
[M12H]215744.8
[M1H]151859
[M12H]215930
The ligation junctions are in bold and italic.
Synthesis of Modified Human Pin 1 Protein Using the MALDI-MS. Mass spectra were acquired on a 4700 Proteomic ana-
lyzer or on a 4800 Plus MALDI TOF/TOF Analyzer (AB Sciex) cali-
brated with Calmix (Calmix 4700 Proteomics Analyzer Calibrating
Mixture). The mass spectra were recorded in positive reflector TOF
mode in the m/z range 500–2000 or 1500–4500 at a fixed laser inten-
sity of 4800 using alpha-cyano-4-hydroxycinnamic acid as a matrix.
Spectra were analyzed by Data Explorer software (Version 4.6, Applied
Bisystems GmbH).
Cysteine-Free Ligation Protocol With HOBt. Human Pin 1
protein was synthesized using the ligation protocol with HOBt,
through ligation of segment 1 with segment 2 (Scheme 1, Sup-
porting Information figures). Segment 1 was synthesized as thio-
ester following the above described procedure and acetylated at
the N-terminus. Segment 2, up to the Asn, was synthesized in a
microwave-assisted peptide synthesizer (Liberty Blue, CEM Corpo-
ration). Starting from Phe, the rest of the sequence of segment 2
was coupled manually following the general procedure for peptide
synthesis. The ligation was performed following the protocol
described previously, using 3.2 lmol of segment 1 and 4.8 lmol
of segment 2. The final volume of the ligation buffer was 700
lL due to the tendency of segment 2 to form a viscous gel (the-
oretically the final volume of the ligation buffer should be 240
lL). The reaction was followed by MALDI-MS and stopped after
78 h. The crude product was purified by RP-HPLC, and the frac-
tions corresponding to the ligation product were collected, lyophi-
lized, and stored at 2208C. In order to remove the Lys side-
chain protecting group (ivdDe), the product was dissolved in
DMF containing 3% of hydrazine. The solution was left at room
temperature for 1h with a gentle mixing. Afterwards, DMF was
removed under reduced pressure and the solid residue was dis-
RP-HPLC. Crude peptides, peptide thioesters, and ligation mixtures
were purified using an Aqua C18-column (internal diameter 4.6 mm,
length 150 mm, particle size 5 lm, pore size 12.5 nm, Phenomenx)
with a linear gradient from 10% to 24% aqueous acetonitrile (0.1%
TFA) over 1.2 min, followed by a gradient of 24% to 57% for the next
€
42 min with a flow rate of 1 mL/min using an AKTA purifier 10
HPLC System (GE Healthcare Life Sciences). Fractions were analyzed
by MALDI-MS and those containing the products were collected,
lyophilized, and stored at 2208C.
Ligation mixtures were analyzed on the microbore Aqua C18-col-
umn using a linear gradient from 4.75% to 57% or from 9.5% to 57%
aqueous acetonitrile (0.1% TFA) for 25 min at a flow rate of 1 mL/
min.
solved in water solution containing 20% MeCN/0.1% FA. The LC-MS. The ligation mixtures (30 lL, at final concentration of
solution was lyophilized and afterwards stored in the freezer at approx. 60 lmol/L for each time point) were injected into the HPLC-
2208C.
MS system (Waters, model Alliance 2796 with a quaternary pump
Biopolymers (Peptide Science)

Improvement of peptide ligation using HOBt
697
FIGURE 2 Kinetics of ligated product formation in ligation reactions with HOBt and without
(solid or dashed lines respectively). The combinations shown are GASATVSPLS-VGPSPLGFLG
(panel A), GASATVSPLVVGPSPLGFLG (panel B), GASATVSPLV-LGPSPLGFLG (panel C) and
LYRAG-YSPGYS (panel D). The reaction was followed by LC-MS and through peak integration of
the product ions at m/z 928.2 (doubly charged, panel A), m/z 934.1 (doubly charged, panel B), m/z
941.2 (doubly charged, panel C), m/z 1274.6 (singly charged, panel D).
and UV/Vis dual absorbance detector Waters 2487 connected with product vs peptide thioester in the TIC spectra. In those cases, were
ESI-MS model Micromass ZQ). The separation was achieved on a
Sunfire C18-column (internal diameter 2.1 mm, particle size 3.5 lm,
length 100 mm) using a linear gradient from 10% or 20% to 100%
aqueous acetonitrile (0.1% FA) in 8 min at a flow rate of 0.3 mL/min.
The mass spectra were acquired for a mass range from m/z 500 to
2000 in positive ion mode using five different cone voltages ranging
from 5 to 70 V. The TIC spectra used for peak integration correspond
to the cone voltage of 30Vand were analyzed by Masslynx 4.0 software
(Waters).
the experiments were performed in triplicate the standard error was
calculated (Reagents provided as supplementary material).
RESULTS AND DISCUSSION
To test if the addition of HOBt to the reaction protocol
described by Wong and coworkers16 can improve the rate and
conversion of ligation we first selected a set of different native
sequences, present in TGIF1 and FoxH1 proteins (Supporting
Information Figure 1) and collected in Figure 1. The selected
peptides contain b-branched (Table I, entries 5, 6 and 7) or
aromatic residues (Table I, entry 8) at the ligation site, since
these residues -and also phosphorylated amino acids- are often
present in protein binding interfaces and are known to add
complexity to the ligation reaction.20,21 All sequences of the
acetylated C-terminal peptide thioesters and free N-terminal
amidated peptides used here are given in Figure 1. These pep-
tides were subjected to the chemical ligation with and without
HOBt as additive. We performed the ligation studies following
the protocol described by Wong et al.,16 in conjunction with a
set of experiments where 2 equivalents of HOBt (1:2 molar
Kinetics of Cysteine-Free Ligation With and Without HOBt.
As mentioned above, depending on the peptide sequence and amino
acids at the ligation junction, the kinetics of the ligation reaction was
followed using different time intervals (Supporting Information
Table I). At each time point, aliquots of 8 lL of the ligation mixture
were taken and quenched by the addition of 0.1% TFA in water (32
lL) or of a TCEP solution (32 lL, of a 10 mg/mL solution) if the
products contained cysteine residues. The samples were diluted up to
1.5 mL with solution containing H2O/MeCN/FA at different ratios by
volume (Supporting Information, Table I, column 4) and analyzed by
LC-MS, in some cases additionally by MALDI-MS or analytical RP-
HPLC. The kinetics was followed through integrating the peak areas
of the single or double charged ions of ligated product and exchanged
thiophenol thioester (Table I). Yields were calculated after the RP-
HPLC purification of the crude ligation mixture, while the given con-
versions where calculated from the integrated ions ratio of ligated ratio of peptide thioester:HOBt) were added to the ligation
Biopolymers (Peptide Science)

698
Todorovski et al.
FIGURE 3 Reversed-phase HPLC chromatograms of the ligation reaction with HOBt (Panel A
and C) and without HOBt (Panel B and D) using two peptides, LYRAG-YSPGYS (Panel A and B)
and GASATVSPLVVGPSPLGFLG (Panel C and D).
mixture. The combinations are shown in Figure1, panel C and while having very little or no influence on the rate of ligation
Table I. (except in the case of GY ligation junction, Figure 2 panel D).
In all cases, the final conversion, based on the peak integra- The addition of 10 equiv. of HOBt did not improve the final
tion of the product ions in the LC-MS experiments, and yield conversion or the reaction rate further (Supporting Informa-
was higher when HOBt was present in the ligation mixture tion Figure 3), but led to a reduction in the conversion rate
(Figures 1 panel C, Figure 2, Supporting Information Figure 2, (Supporting Information Figure 3 dashed lines). We attribute
Table I) independently of the peptide thioester/peptide molar this observation to a decrease of the pH of the ligation mixture
ratio. Moreover, in the cases when aromatic or b-branched caused by the acidic nature of HOBt.22 Furthermore, using the
amino acids were present at the ligation junction, the influence above-described conditions the product can be obtained with
of the added HOBt on the final conversion/yield was greater high purity since it elutes at a different retention time than the
(Figure 1 panel C, Figure 2 bold lines) than when the peptide starting peptide reagents (thioester and the free N-terminus
had less sterically hindered amino acids (Figure 1 panel C, peptide), as displayed in Figure 3.
Supporting Information Figure 2 bold lines). Interestingly, the
In another set of experiments performed as controls, we
HOBt addition specifically increased the final conversion/yield, ligated the same set of NH2-peptide sequences as before, in the
Table II NH2-Free C-Terminal Peptide Thioesters and N-Terminal Peptide-Free Amines Used in the Ligation Studies
Peptide Thioesters
m/z theor
m/z exper
Peptides
m/z theor
m/z Exper
Lig. Com.
1. NH2-GASATVSPLG-SC7H7
965.54
965.58
A. NH2-GGPSPLGFLG-CONH2
900.49
900.47
1-A
2-A
2-B
3-B
4-C
2. NH2-GASATVSPLS-SC7H7
3. NH2-GASATVSPLV-SC7H7
4. NH2-GASATVSPLL-SC7H7
995.49
1007.53
1021.54
995.46
1007.55
1021.57
B. NH2-VGPSPLGFLG-CONH2
C. NH2-CGPSPLGFLG-CONH2
942.54
946.48
942.55
946.46
Lig. Com.: ligation combinations.
Biopolymers (Peptide Science)

Improvement of peptide ligation using HOBt
699
philes (Figure 4, panel A dotted line and solid line, correspond-
ingly). To test the effect of the activators we chose Val-Val
residues at the ligation site because HOBt-enhances the liga-
tion reaction in cases where b-branched or aromatic amino
acids were present at the junction (Figure 1 panel C, Figure 2
and Table I). In the case of DIPEA, we observed limited
improvement in the final conversion (Figure 4, panel A solid
line), while for DBU the final conversion was lower, despite the
fact that up to 50 h the kinetics of the reaction was almost
identical to that achieved using DIPEA (Figure 4, panel A dot-
ted line). The observation is directly related to a transthioester-
ification step (which can be followed by a singly charged ion at
m/z 1035.5), as after 50 h the number of ions corresponding to
the thiophenol thioester exchange was close to zero in the case
of DBU (Figure 4, panel B dotted line). Our results indicate
that the presence of DBU do not contribute to increasing the
ligation rate under the experimental conditions we have
investigated, while in the case of DIPEA we could detect small
improvements regarding the final conversion.
FIGURE 4 Kinetics of ligated product formation (panel A) and
transthioesterification step (panel B) in the ligation reaction with
HOBt1DIPEA (solid line), HOBt (dashed lines), HOBt1DBU
(dotted lines) and HOAt (long-dashed lines). The ligation was per-
formed using ligation combinations GASATVSPLV-VGPSPLGFLG.
The reaction was followed by LC-MS and through peak integration
of the product ion at m/z 934.1 (doubly charged, panel A), and
through peak integration of the ion at m/z 1035.5 (singly charged,
panel B).
Potential Alternatives to HOBt
Regarding the choice of an alternative to HOBt, we used
HOAt, a molecule similar to HOBt but with a lower pKa (pKa
3.28 and 4.60 respectively)24,25 and therefore potentially a bet-
ter leaving group. However, we observed that the final conver-
sion and also the reaction rate were similar to those obtained
with HOBt in the experimental conditions tested (Figure 4,
panel A). This observation could be attributed to the fact that
HOAt requires double the induction time relative to all other
cases (Figure 4, panel A long-dashed line). This extended time
can be explained by the fact that the transthioesterification is
also slower for HOAt (Figure 4, panel B long-dashed line).
These observations imply that if reactions were allowed to pro-
ceed for longer (more than 96 h), we would achieve a higher
final conversion; however, this would also cause an increase in
side products, thereby reducing the yield of the product of
interest and adding complexity to the purification step. How-
ever, when the ligation reactions are more favorable, the use of
HOAt could be beneficial with regards to the final conversion,
as increases in the reaction time will not have a strong influ-
ence on the amount of side products.
presence of HOBt but using unprotected thioesters (at the
N-terminus) (Table II). In all cases the formation of the
expected ligation product was observed (Supporting Informa-
tion Figure 4 solid lines); however, in the presence of cyclic
side-products (Supporting Information Figure 4 dashed lines).
According to the observed m/z values, the cyclization products
are formed due to the intramolecular cyclization of the peptide
thioesters, a competitive side reaction favored by the unpro-
tected N-termini of the peptide thioester.
Finally, ligation reactions with or without N-terminal pro-
tected thioesters and peptides containing a N-term Cysteine
were much faster than in all the other ligation reactions tested,
(Supporting Information Figure 5).
The addition of HOBt suppresses the 5(4H)-oxazolone for-
mation, preventing plausible racemization, and thus, improv-
ing the aminolysis rate through the formation of more
effective acyl donors esters.23 For this reason, we hypothesized
that the presence of HOBt in the ligation mixture improves the
ligation reaction yield and also contribute to increasing the
conversion.
Synthesis of a WW Domain Containing a
Non-Natural Amino Acid
Effects of the Presence of DBU and DIPEA Activators
The addition of a base activator can improve the final conver-
sion and speed up the reaction. To test this hypothesis, we per-
formed ligation reactions with HOBt and DBU and DIPEA as
Finally, as an application of the ligation protocol here described
we synthesized a modified WW domain using the peptidyl-
prolyl isomerase I (Pin I, 34 AA, Supporting Information
Scheme 1) sequence as the template. Pin1 is a modular protein
activating bases, since they are strong bases but weak nucleo- that contains a WW domain, responsible for target recognition,
Biopolymers (Peptide Science)

700
Todorovski et al.
FIGURE 5 Reversed-phase chromatograms (Panel A and B) and MALDI-MS (Panel C and D) of
modified human Pin I protein. After purification of the crude Pin I (Panel A) the pure sample
(Panel B) was analyzed by MALDI-MS with ivdDe group present at the Lys residue (Panel C) and
after removal of the ivdDe group (Panel D). The masses corresponding to the Pin 1 protein are
labeled in bold, while the doubly charge ion of the Pin 1 protein is with asterisk. The peak in Panel
A, labeled with double asterisk corresponds to the Pin I protein with a 16 Da higher mass.
and the catalytic domain involved in the cis-trans isomeriza- obtained with high purity (Figure 5, panel B and C). After suc-
tion of phosphor-Ser/Thr-Pro bonds.26,27 The folding pathway
cessful removal of the ivdDe side chain group (Figure 5, panel
of the Pin1 WW domain has been the subject of many stud-
ies,28 which have shown that the domain can be unfolded and
refolded with high efficiency. In addition, many structures of
this WW domain in complex with target peptides are described
in the literature.29 Since folding and binding studies often
require the production of many WW proteins carrying muta-
tions, we explored the possibility of using peptide synthesis
and refolding as an efficient strategy to prepare these samples.
If some of these mutations are designed to include non-natural
amino acids, a ligation strategy might be the best choice to
optimize the segment’s design and purification. As an example,
we prepared a WW sequence with a non-natural residue 2,4,6-
trimethylphenyl Ala (Msa)30–32 at position 19 and several point
mutations: Lys 1 was replaced with Gly, whereas Arg 9 and Arg
12 were replaced by Ala. The replacement of both arginines
prevented possible intramolecular cyclization with the guanidi-
nium group of the Arg side-chain, while the Lys at position 8
was protected with ivdDe. The ligation junction was chosen at
a Ser-Gly (position 14-15) due to the expected faster ligation
rate when compared with the other options (Supporting Infor-
mation Figure 2, panel B). The crude product was purified by
D) the pure product was obtained at final yield of 15%.
Despite the low yield obtained in this example, the synthesis of
the modified Pin1 is a proof of principle of the applicability of
the method, and we consider that the yield could be signifi-
cantly improved with further optimizations.
Addition of HOBt to Peptides Containing
Phosphorylated Residues
We also tested the HOBt-ligation protocol using peptide
sequences that correspond to TGIF1 protein containing phos-
phorylated Ser (Table I entry 9) in the sequence of the acety-
lated C-terminal peptide thioester. The expression of proteins
containing phosphorylated residues is complicated and expen-
sive. In addition, the synthesis of phosphorylated polypeptides
using microwave (MW)-assisted SPPS is inefficient due to the
alpha,beta-elimination of the phosphate group under MW
conditions catalyzed by piperidine.33 Therefore, the cysteine-
free ligation with HOBt could also be useful in the synthesis of
phosphorylated polypeptides. The addition of HOBt signifi-
cantly improved the final conversion (Figure 1 panel C, Figure
RP-HPLC (Figure 5, panel A) and the modified Pin I was 6, Table I entry 9 and Supporting Information). Furthermore,
Biopolymers (Peptide Science)

Improvement of peptide ligation using HOBt
701
the phosphorylated residue did not participate in any side induction time of the reaction is related to the transthioesterifi-
reaction with the added HOBt, and the products that corre- cation step, i.e., when the transthioesterification reaches the
spond to truncated peptide sequences or sequences with
dephosphorylated Ser residue were not detected.
maximum and begins to drop, the formation of the product
starts immediately.
Mechanism of Reaction
CONCLUSIONS
Our experimental results suggest a possible general mechanism
for the ligation reactions (Supporting Information Scheme 2).
The initial conversion of the benzyl to the phenyl thioester is
followed by the subsequent substitution by the more reactive
HOBt ester. The lower pKa of HOBt (pKa 4.6) compared to
that of thiophenol (pKa 6.6) would make the former a better
leaving group.
We have found that the addition of HOBt to the ligation of
C-terminal peptide thioesters with various free N-terminus
peptides increases the ligation conversion, especially when
b-branched or aromatic amino acids are present at the ligation
junction. This increase is probably due to the in-situ formation
of a highly reactive HOBt-ester. Furthermore, all reactions pro-
ceeded without epimerization and the method is chemoselec-
tive. Only when Lys residues are present in the sequence it is
expected that the residue’s side chain react with the in-situ
formed HOBt-ester in a similar way as with a thioester previ-
ously reported by Wong et al.16 Despite the long reaction
times, the method shows potential for application in the
synthesis of peptide sequences that cannot be generated by
other ligation techniques, especially in the presence of b-
branched or aromatic amino acids at the junction site. The
improvement of the method is also applicable to the synthesis
of phosphorylated peptides that, despite recent developments
in protein chemistry, are not readily obtainable due to their
tendency to suffer alpha,beta-elimination of the phosphate
group.
Our results indicate that the limiting reaction step is the
transthioesterification, as the ligation reaction does not occur
in the absence of thiophenol. It is known that the thiophenol is
able to undergo rapid and almost complete thiol-thioester
exchange34 and is an excellent leaving group. Therefore, in the
absence of thiophenol, the initial benzyl thioesters won’t be
able to undergo aminolysis or to form active HOBt esters,
since alkanethiols are poor leaving groups.34 When thiophenol
is replaced by 4-(carboxymethyl)thiophenol (MPAA), which
has a pKa of 6.6, similar to that of thiophenol, the rate of
transthioesterification is lower (Supporting Information Figure
7, panel A dashed line). We believe that the ligation buffer can
affect this step since the presence of MPAA, which shows better
solubility in aqueous buffers, induces a higher and faster trans-
thioesterification than thiophenol.34 Moreover, in our experi-
ments, the concentration of the MPAA thioester formed after
the transthioesterification remained constant during the
experiment. Consequently, the ligation conversion is six times
lower, as shown by peak integration in the LC-MS experiments
(Supporting Information Figure 7, panel B solid and dashed
line, respectively). In addition, when HOBt, was added to the
ligation mixture in the absence of thiophenol, no ligation
product was detected (data not shown). Again, as we men-
tioned before, we believe that this is due to the fact that alkane-
thiols are poor leaving groups and therefore the preformed
benzyl thioesters are not able to react giving the active HOBt
esters. In summary, the addition of HOBt facilitates the reac-
tion of the exchanged phenyl thioester with the free N-termi-
nus peptide, probably through formation of highly reactive
HOBt-ester, However, it does not have a significant effect on
the extent of transthioesterification. These observations are
shown in Supporting Information Figure 8 (panel A-D), where
the final conversion of the transthioesterification step (Sup-
porting Information Figure 8, dotted and long-dashed lines)
differs from the final conversion of the product (Supporting
Information Figure 8, dashed and solid lines). In addition, the
T.T. is a recipient of a FP7 Marie Curie Actions (COFUND Grant
program), and D.S. is a recipient of “la Caixa”-IRB Barcelona
International PhD Fellowship. The authors thank IRB Barcelona
for support, Dr. M. Royo for discussions regarding some peptide
synthesis, Dr. M. Vilaseca and of Dr. I. Fernandez for MS analy-
sis. They also thank Alvaro Rol Rua for the preparation of the
Msa amino acid. M.J.M. is an ICREA Programme Investigator.
REFERENCES
1. Merrifield, R. B. J Am Chem Soc 1963, 85, 2149–2154.
2. Alewood, P.; Alewood, D.; Miranda, L.; Love, S.; Meutemans,
W.; Wilson, D. Methods Enzym 1997, 289, 14–29.
3. Chandrudu, S.; Simerska, P.; Toth, I. Molecules 2013, 18, 4373–
4388.
4. Dawson, P.; Muir, T.; Clark-Lewis, I.; Kent, S. Science 1994, 266,
776–779.
5. Dirksen, A.; Meijer, E. W.; Adriaens, W.; Hackeng, T. M. Chem
Commun (Camb) 2006, 15, 1667–1669.
6. Dawson, P. E.; Kent, S. B. H. Annu Rev Biochem 2000, 69, 923–
960.
7. Wombacher, R.; Cornish, V. W. J Biophotonics 2011, 4, 391–
402.
8. Kemp, D. S.; Bernstein, Z. W.; McNeil, G. N. J Org Chem 1974,
69, 2831–2835.
Biopolymers (Peptide Science)

702
Todorovski et al.
9. Kemp, D. S.; Choong, S. L. H.; Pekaar, J. J Org Chem 1974, 39, 25. Carpino, L. A. J Am Chem Soc 1993, 115, 4397–4398.
3841–3847. 26. Ping Lu, K.; Hanes, S. D.; Hunter, T. Nature 1996, 380,
10. Blake, J. Int J Pept Protein Res 1981, 17, 273–274. 544–547.
11. Aimoto, S.; Mizoguchi, N.; Hojo, H.; Yoshimura, S. Bull Chem 27. Ranganathan, R.; Lu, K. P.; Hunter, T.; Noel, J. P. Cell 1997, 89,
Soc Jpn 1989, 62, 524–531.
12. Aimoto, S. Biopolymers 1999, 51, 247–265.
875–886.
28. Luo, Z.; Ding, J.; Zhou, Y. Biophys J 2007, 93, 2152–2161.
29. Aragoꢀn, E.; Goerner, N.; Zaromytidou, A. I.; Xi, Q.;
Escobedo, A.; Massaguꢀe, J.; Macias, M. J. Genes Dev 2011, 25,
1275–1288.
13. Vila-Perelloꢀ, M.; Muir, T. W. Cell 2010, 143, 191–200.
14. Pentelute, B. L.; Kent, S. B. H. Org Lett 2007, 9, 687–690.
15. Crich, D.; Banerjee, A. J Am Chem Soc 2007, 129, 10064–10065.
16. Payne, R. J.; Ficht, S.; Greenberg, W. A.; Wong, C. H. Angew 30. Medina, E.; Moyano, A.; Pericas, M. A.; Riera, A. Helv Chim
Chem Int Ed 2008, 47, 4411–4415. Acta 2000, 83, 972–988.
17. Wang, P.; Danishefsky, S. J. J Am Chem Soc 2010, 132, 17045– 31. Martın-Gago, P.; Gomez-Caminals, M.; Ramon, R.; Verdaguer,
ꢀ
ꢀ
17051.
X.; Martin-Malpartida, P.; Aragoꢀn, E.; Fernꢀandez-Carneado, J.;
Ponsati, B.; Loꢀpez-Ruiz, P.; Cortes, M. A.; Colꢀas, B.; MacIas, M.
J.; Riera, A. Angew Chem Int Ed 2012, 51, 1820–1825.
18. Novabiochem, Novabiochem Innovations 4/99; 1999.
19. Kaiser, E.; Colescott, R. L.; Bossinger, C. D.; Cook, P. I. Anal
Biochem 1970, 34, 595–598.
32. Martꢀın-Gago, P.; Aragoꢀn, E.; Gomez-Caminals, M.; Fernꢀandez-
Carneado, J.; Ramoꢀn, R.; Martin-Malpartida, P.; Verdaguer, X.;
Loꢀpez-Ruiz, P.; Colꢀas, B.; Cortes, M. A.; Ponsati, B.; Macias, M.
J.; Riera, A. Molecules 2013, 18, 14564–14584.
20. Hackeng, T. M.; Griffin, J. H.; Dawson, P. E. Proc Natl Acad Sci
USA 1999, 96, 10068–10073.
21. Coltart, D. M. Tetrahedron 2000, 56, 3449–3491.
22. Saha, A.; Nadimpally, C.; Paul, A.; Kalita, S.; Mandal, B. Protein 33. Attard, T. J.; O’Brien-Simpson, N. M.; Reynolds, E. C. Int J Pept
Pept Lett 2014, 21, 188–193.
Res Ther 2009, 15, 69–79.
34. Johnson, E. C. B.; Kent, S. B. H. J Am Chem Soc 2006, 128,
6640–6646.
23. Ko€nig, W.; Geiger, R. Chem Ber 1970, 103, 788–798.
24. Novabiochem. Novabiochem Innovations: 2/09; 2009.
Biopolymers (Peptide Science)