Sequence sensitivity and pH dependence of maleimide conjugated N-terminal cysteine peptides to thiazine rearrangement

Article abstract of DOI:10.1002/psc.3323

Thiazine formation during the conjugation of N-terminal cysteine peptides to maleimides is an underreported side reaction in the peptide literature. When the conjugation was performed at neutral and basic pH, we observed the thiazine isomer as a significant by-product. Nuclear magnetic resonance (NMR) spectroscopy confirmed the structure of the six-membered thiazine and ultra-high performance liquid chromatography (UHPLC) combined with tandem mass spectrometry (MS/MS) allowed for facile, unambiguous detection due to a unique thiazine mass fragment. Furthermore, substitution of various amino acids adjacent to the N-terminal cysteine in a tripeptide model system resulted in different rates of thiazine formation, albeit within the same order of magnitude. We also determined that varying the N-substitution of the maleimide affects the thiazine conversion rate. Altogether, our findings suggest that thiazine rearrangement for N-terminal cysteine-maleimide adducts is a general side reaction that is applicable to other peptide or protein systems. Performing the conjugation reaction under acidic conditions or avoiding the use of an N-terminal cysteine with a free amino group prevents the formation of the thiazine impurity.

Full text of DOI:10.1002/psc.3323

Received: 12 February 2021
DOI: 10.1002/psc.3323
Revised: 10 March 2021
Accepted: 11 March 2021
R E S E A R C H A R T I C L E
Sequence sensitivity and pH dependence of maleimide
conjugated N-terminal cysteine peptides to thiazine
rearrangement
Isaiah N. Gober1
| Alexander J. Riemen1
| Matteo Villain2
1Research and Development Department,
Bachem Americas, Inc., Torrance, California,
USA
Thiazine formation during the conjugation of N-terminal cysteine peptides to
maleimides is an underreported side reaction in the peptide literature. When the
conjugation was performed at neutral and basic pH, we observed the thiazine isomer
as a significant by-product. Nuclear magnetic resonance (NMR) spectroscopy
confirmed the structure of the six-membered thiazine and ultra-high performance
liquid chromatography (UHPLC) combined with tandem mass spectrometry (MS/MS)
allowed for facile, unambiguous detection due to a unique thiazine mass fragment.
Furthermore, substitution of various amino acids adjacent to the N-terminal cysteine
in a tripeptide model system resulted in different rates of thiazine formation, albeit
within the same order of magnitude. We also determined that varying the
N-substitution of the maleimide affects the thiazine conversion rate. Altogether, our
findings suggest that thiazine rearrangement for N-terminal cysteine-maleimide
adducts is a general side reaction that is applicable to other peptide or protein
systems. Performing the conjugation reaction under acidic conditions or avoiding the
use of an N-terminal cysteine with a free amino group prevents the formation of the
thiazine impurity.
2CMC Development Group, Bachem Americas,
Inc., Torrance, California, USA
Correspondence
Matteo Villain, CMC Development Group,
Bachem Americas, Inc., 3132 Kashiwa Street,
Torrance, CA 90505 USA.
Email: matteo.villain@bachem.com
1
|
INTRODUCTION
neutral pH.10,11 Despite its advantages and widespread use, thiol–
maleimide conjugation is prone to various side reactions including
The thiol–maleimide reaction is one of the most widely used
strategies for the covalent modification of peptides and proteins.1–3
The reaction has been used to install a variety of chemical labels onto
biomolecules via thiol conjugation including fluorescent dyes,4–6 PEG
moieties,7 radiolabels,8 and small molecules.9 A key aspect of thiol–
maleimide chemistry's success is the chemoselectivity and rapid kinet-
ics of the reaction between maleimides and thiols at or slightly below
hydrolysis of the thiosuccinimide linkage, retro-Michael reaction and
subsequent cross-reactivity with other thiols, and thiazine formation
when conjugation takes place using an N-terminal cysteine.12–14 In
fact, the instability of maleimido–cysteine conjugates has prompted
the development of other strategies for thiol-selective bio-
conjugation.15 Although the hydrolysis side reaction has been shown
to be advantageous in improving the pharmacokinetics and pharmaco-
dynamics of antibody drug conjugates,16 the aforementioned side
reactions are typically detrimental to the production and efficacy of
maleimido bioconjugates. Interestingly, the rearrangement of
succinimidyl thioether linkages to a thiazine structure during the
conjugation of N-terminal cysteine containing peptides appears to be
Abbreviations: MHH, 2,5-dihydro-N-(6-hydroxyhexyl)-2,5-dioxo-1H-pyrrole-1-hexanamide
or maleimido-N-6-hydroxyhexyl-1-hexanamide; MPA, 3-maleimidopropionic acid;
PEGMA10K, methoxy PEG maleimide MW: 10 kDa.
Isaiah N. Gober and Alexander J. Riemen contributed equally to this work.
This is an open access article under the terms of the Creative Commons Attribution-NonCommercial License, which permits use, distribution and reproduction in any
medium, provided the original work is properly cited and is not used for commercial purposes.
© 2021 Bachem Americas,Inc. Journal of Peptide Science published by European Peptide Society and John Wiley & Sons Ltd.
J Pep Sci. 2021;e3323.
wileyonlinelibrary.com/journal/psc
1 of 11
https://doi.org/10.1002/psc.3323

2 of 11
GOBER ET AL.
an underreported side reaction in the peptide literature. Formation of
the thiazine impurity can complicate the purification, characterization,
and storage of peptide conjugates and may lead to significant product
loss if the succinimidyl thioether linkage is intended in the final product.
Furthermore, because the succinimidyl thioether and thiazine isomers
have identical molecular weights, it is not trivial to distinguish the two
by routine liquid chromatography–mass spectrometry (LC–MS).
N-terminal amine of the cysteine. This nucleophilic attack can occur at
the carbonyl at Position 1 or the carbonyl at Position 3 (Scheme 2).
Transcyclization occurs through a fused bicyclic tetrahedral intermedi-
ate, which allows for the formation of a six-membered thiazine prod-
uct. Although nucleophilic attack at Position 3 in Scheme 2 to form
the seven-membered ring product is possible, this isomer was not
detected during this study.
We investigated both the pH dependence and sequence depen-
dence of the succinimidyl thioether-to-thiazine side reaction using a
tripeptide H-Cys-Xxx-Phe-OH (CXF) model system (where X can be
substituted for different amino acids). We observed a general occur-
rence of thiazine rearrangement in maleimide conjugated N-terminal
cysteine peptides when conjugation was performed at or above physi-
ological pH. Consistent with previous literature,14 we found that the
conversion of succinimidyl thioethers to thiazine occurs more
rapidly at elevated pH. Although nuclear magnetic resonance (NMR)
characterization was necessary for structure determination, we dem-
onstrated that UHPLC–MS/MS analysis allows for rapid detection
and differentiation of the succinimidyl thioether from the thiazine
isomer. Additionally, we demonstrated that the rate of thiazine
formation was influenced by the identity of the neighboring amino
acid, although measured rates of rearrangement were within the same
order of magnitude for the various peptides studied. Finally, we
examined the effect of different maleimide linkers on thiazine forma-
tion and observed similar propensities for the side reaction (Figure 1).
2.2
|
Initial conjugation reaction time course
A linear model peptide, H-Cys-Xxx-Phe-OH (CXF), was chosen to
investigate the thiazine side reaction that occurs during the conjuga-
tion of
a maleimide with an unprotected N-terminal cysteine
peptide. This model system was chosen to ensure facile peptide syn-
thesis and analysis and to allow for investigation of the effect of neigh-
boring amino acids by changing the amino acid derivative at the X
position. Initial experiments were conducted using H-Cys-Gly-Phe-OH
(CGF) and 3-maleimidopropionic acid (MPA) in 0.1-M potassium phos-
phate buffer (with 10% acetonitrile) at pH 7.3. Because the aqueous
solutions contain organic cosolvent, it is important to note that the pH
values reported were the apparent measured pH based on readings
from a benchtop pH meter. Additionally, all peptides used were not
corrected for trifluoroacetic acid (TFA) and water content prior to
weighing. The formation of the thiazine impurity was studied at vari-
ous time points (0, 0.5, 1, 2, and 24 h). The percent conversion to the
thiazine structure was determined by integrating the UHPLC peak area
with respect to the total conjugate-related peak area (the response
factor for the absorbance at 220 nm of the succinimidyl thioether and
thiazine isomers was assumed to be the same). To collect each time
point, the reactions were quenched with 1% TFA in water. For the 0-h
time point, peptide dissolved in phosphate buffer solution was pre-
mixed with 1% TFA in water before adding the MPA solution to pre-
vent the formation of the thiazine impurity.
2
|
RESULTS AND DISCUSSION
2.1
|
Rearrangement of succinimidyl thioether to
thiazine via transcyclization
Peptides containing cysteine react readily with maleimides to form
succinimidyl thioether conjugates (Scheme 1). For peptides that are
conjugated to maleimides through an N-terminal cysteine, the
resulting succinimide is susceptible to nucleophilic attack from the
Conjugation of the H-CGF-OH peptide to MPA was rapid, as the
starting material was almost completely consumed even under the
acidic conditions of the “time 0 h” measurement (Figure 2, black
trace). Two peaks of roughly equal abundance were formed,
corresponding to the two diastereomers of the succinimidyl thioether
conjugation product. As the reaction was allowed to proceed to longer
time points, a third peak became visible in the UHPLC chromatograph
and increased in abundance while the two succinimidyl thioether
peaks decreased in abundance.
UHPLC–MS analysis confirmed that all three peaks were
CGF-MPA conjugate isomers with the same molecular weight. To
FIGURE 1 Chemical structure of the succinimidyl thioether
versus the thiazine isomer
SCHE ME 1 Reaction of maleimide with
N-terminal Cys containing peptide via Michael
addition

GOBER ET AL.
3 of 11
SCHEME 2 Thiazine side reaction
formed after nucleophilic attack of
N-terminal amine at the carbonyl of the
succinimide and trancyclization to give a
six-membered ring product
FIGURE 2 Ultra-high performance
liquid chromatography (UHPLC) analysis
of CGF conjugation to
3-maleimidopropionic acid (MPA) at pH
7.3. The reaction mixture was analyzed at
0 h (black), 0.5 h (blue), 1 h (cyan), 2 h
(orange), and 24 h (green)
SCHEME 3 Differentiation of the thiazine isomer
from the succinimidyl thioether isomer based on the
unique tandem mass spectrometry (MS/MS)
fragmentation of the thiazine isomer
distinguish between the succinimidyl thioether and thiazine
isomers, tandem mass spectrometry (MS/MS) was used. Differentia-
tion of the isomers was possible due to the unique fragmentation
pattern of the thiazine isomer (Scheme 3). Fragmentation of the amide
bond at the aminopropionic acid moiety of the thiazine isomer should
yield a fragment with a mass of 90.05 Da and another fragment with
a mass of 406.11 Da. In the MS/MS spectrum, a species with a mass
of 406.11 Da was detected for the peak corresponding to the
supposed thiazine isomer (supporting information). In the case of the
succinimidyl thioether isomer, cleavage of either amide bond of
the aminopropionic acid moiety should not produce two fragments
due to the cyclic nature of the succinimide. Consequently, for the
two peaks corresponding to the supposed succinimidyl thioether
diastereomers, a species with a mass of 406.11 Da was not detected
during MS/MS analysis (supporting information).
2.3
|
Structure elucidation of purified CGF-MPA
isomer
To isolate the succinimidyl thioether and thiazine isomers for analysis
by NMR, the conjugation was performed on a larger scale using

4 of 11
GOBER ET AL.
500 mg of peptide. To prepare the succinimidyl thioether isomers, the
conjugation of CGF to MPA was performed in unbuffered water with
10% acetonitrile. Dissolving the peptide and maleimide in unbuffered
aqueous solution allowed the reaction conditions to remain acidic
because MPA is a weak acid and because the peptide was used as a
TFA salt. This prevented the thiazine side reaction from occurring dur-
ing the conjugation. The reaction was monitored by analytical high
performance liquid chromatography (HPLC), and after reaching com-
pletion, the two succinimidyl thioether peaks were purified by
reversed-phase HPLC (supporting information). The earlier eluting
peak was obtained in 95.9% purity (with 1.8% of the second peak),
and the later eluting peak was obtained in 97.4% purity (with 2.6% of
the first peak present in the sample).
For preparation of the thiazine isomer, the larger scale conjuga-
tion reaction was base stressed by performing the reaction at pH 8.5
(0.1-M potassium phosphate solution with 10% acetonitrile). The
reaction was monitored by UHPLC, and after 48 h, the conjugated
species were isolated by reversed-phase HPLC. The earlier eluting
peak corresponding to the supposed thiazine isomer was obtained in
94.8% purity and contained 2.5% of the putative succinimidyl
thioether isomer. Analysis of the purified thiazine isomer by MS/MS
also produced a species with a mass of 406.11 Da, indicative of the
unique fragmentation of the thiazine isomer.
The purified isomers were then characterized by NMR spectros-
copy. All the 1H NMR chemical shifts were assigned using a combina-
tion of 1-D and 2-D NMR experiments (supporting information). For
protons with overlapping chemical shift signals, DQF–COSY experi-
ments were used to make the assignments. Position assignments are
indicated in Figure 3. Correlations in the 1H–13C Heteronuclear Single
Quantum Coherence (HSQC) NMR spectrum were used for number-
ing the protons in the thiazine isomer of CGF-MPA.
FIGURE 3 Top: Labeling of 1H and 13C peaks for CGF-
3-maleimidopropionic acid (MPA) in the nuclear magnetic resonance
(NMR) spectra (13C signals are in order of increasing chemical shift).
Bottom: Confirmation of the formation of the thiazine isomer using
1H–13C Heteronuclear Multiple Bond Correlation (HMBC) NMR. The
key correlation for excluding the possibility of the succinimidyl
thioether structure (H9 to C18) is indicated in red in the 2-D NMR
spectrum
The thiazine structure could be elucidated based on correlations
between H9 and C18, H1a and C20, H1a and C5, H5 and C1, H5 and C18
,
incubated with MPA in potassium phosphate solution for 24 h before
analyzing by UHPLC. Reactions were quenched using a solution of 1%
TFA in water.
H5 and C19, and H9 and C20 in the 1H–13C Heteronuclear Multiple
Bond Correlation (HMBC) NMR spectrum (Figure 3). The key correla-
tion for differentiating the succinimidyl thioether and thiazine isomers
was the H9 and C18 correlation. For the six-membered thiazine struc-
ture, C18 is separated from H9 by three bonds, which allowed for
detection by 1H–13C HMBC NMR.17 Because the H9–C18 correlation
was observed, the succinimidyl thioether structure (where H9 and C18
are five bonds away) could be excluded.
At pH 5.0, the conjugation of MPA to CGF proceeded rapidly,
producing two peaks that corresponded to the succinimidyl thioether
diastereomers of the CGF-MPA conjugate. However, at 24 h, only
0.1% of the thiazine isomer was observed, and hydrolysis of MPA and
the succinimidyl thioether isomers was negligible. This result is consis-
tent with the supposed mechanism of thiazine rearrangement,
because protonation of the N-terminal amino group under acidic con-
ditions should prevent intramolecular nucleophilic attack of the
succinimide carbonyl and preserve the succinimidyl thioether
structure.
2.4
|
Investigation of the influence of pH on
thiazine rearrangement
Analysis of CGF and MPA conjugation at pH 7.3 revealed the thi-
azine isomer as the major product at the 24 h mark in 70% abun-
dance. Conversely, the abundance of the succinimidyl thioether
isomers decreased, which was indicative of the rearrangement.
Expectedly, the increase in the rate of thiazine formation at pH 7.3
compared to pH 5.0 was substantial. Significant hydrolysis of excess
MPA was also observed, but minimal hydrolysis products of the CGF-
MPA conjugate were detected.
To investigate the role of pH on the formation of the thiazine impu-
rity, the conjugation reaction was monitored under acidic conditions
at pH 5.0, near neutral conditions at pH 7.3, and under basic condi-
tions at pH 8.4 (Figure 4). The reactions were all carried out in aque-
ous 0.1-M potassium phosphate solution to ensure that any
differences observed in rate and extent of thiazine formation were
not due to differences in buffer. The purified H-CGF-OH peptide was

GOBER ET AL.
5 of 11
FIGURE 4 Ultra-high performance liquid
chromatography (UHPLC) analysis of CGF
conjugation to 3-maleimidopropionic acid (MPA)
after 24 h at pH 5.0 (black), pH 7.3 (green), and
pH 8.4 (orange)
FIGURE 5 Left: Increase in thiazine formation over time for different neighboring amino acids of CXF-3-maleimidopropionic acid (MPA) at
pH 7.3 in 0.1-M potassium phosphate buffer with 10% acetonitrile. For CLF, the conjugation reaction was carried out in 0.1-M potassium
phosphate buffer with 26% acetonitrile added to maintain solubility. Right: Increase in thiazine formation over time for different neighboring
amino acids of CXF-MPA at pH 8.4 in aqueous 0.1-M potassium phosphate solution with 10% acetonitrile. For CLF, the conjugation reaction was
carried out in 0.1-M potassium phosphate solution with 26% acetonitrile added to maintain solubility
Thiazine formation was rapid at pH 8.4, with nearly 90% of the
succinimidyl thioether isomers converted to the thiazine isomer after
24 h. The considerable increase in the rate of thiazine formation at
higher pH again supports a base-dependent mechanism involving
nucleophilic attack of the succinimide by the N-terminal amine. Due
to the higher pH, most of the excess MPA was hydrolyzed over this
period of time. On the other hand, the thiazine isomer appeared to be
stable under the base stressed conditions.
H-CSF-OH, and H-CLF-OH. Side chains were selected to provide a
sampling of functional groups and potential stereoelectronic interac-
tions. Because it is the smallest amino acid, glycine was selected with
the expectation of it having the least conformational restriction due
to steric hindrance. Glutamic acid and lysine were chosen to investi-
gate the effect of acidic and basic ionizable side chains, respectively,
while serine was chosen to test the role of a neighboring hydroxyl
functional group. Finally, leucine was chosen for a representative
hydrophobic sequence.
Conjugation reactions with MPA were carried out at pH 5.0, pH
7.3, and pH 8.4 for various time points. The percent conversion of
succinimidyl thioether to thiazine was measured at 0-, 0.5-, 1-, and
2-h time points by UHPLC (Figure 5). At pH 5.0, thiazine
rearrangement was suppressed for all the CXF peptides. Even after
336 h, only 1.1% conversion occurred in the fastest instance, which
made it impractical to calculate a rate of thiazine conversion at pH 5.0
2.5
|
Investigation of sequence dependence on
thiazine rearrangement kinetics
To assess the influence of neighboring amino acids (relative to
N-terminal cysteine) on the thiazine side reaction, a selection of
five peptides was prepared: H-CGF-OH, H-CEF-OH, H-CKF-OH,

6 of 11
GOBER ET AL.
(Table 1). In the case of the pH 7.3 experiments, all five peptides
showed extensive conversion to the thiazine isomer. CKF and CGF
gave the fastest rates of conversion with 8.7% and 6.1% thiazine
formed per hour, respectively (Table 2). For CLF-MPA, the rate of
thiazine conversion was the second slowest with a rate of 4.3%
thiazine formed per hour. The slowest rate of rearrangement
corresponded to the CEF peptide, which was over 2 times slower than
the fastest result.
thiosuccinimide hydrolysis when using beta amino maleimides,20 the
amine of Lys does not appear to be well suited for base catalysis as it
has a pKa of 10.5 and would therefore be protonated at pH 7.3 and
pH 8.4.21
In addition, because the conjugation and subsequent
rearrangement was performed in water, it was not certain whether
the Lys amine, the Glu carboxylate, or the Ser hydroxyl in CKF,
CEF, and CSF (respectively) would participate in electrostatic or
hydrogen bonding interactions that directly affected thiazine
formation via interaction with the N-terminal amine. In the absence
Conducting the studies at pH 8.4 resulted in a considerable
increase in the rate of thiazine formation for each of the peptides
(Figure 5 and Table 2). However, a different order of reactivity was
observed at elevated pH, with CGF giving the most rapid rate of con-
version at 17.6% thiazine formed per hour while the CLF peptide was
the slowest at 6.6% thiazine formed per hour. The trend of the CLF
and CEF peptides being considerably slower than the other peptides
was also observed for the pH 8.4 experiments.
of
a hydrophobic binding pocket, these types of noncovalent
interactions re often negligible in water due to solvation and
hydrogen bonding between water and the polar functional
groups of amino acid side chains.22–24 Furthermore, the presence
of salt (0.1-M potassium phosphate) was expected to weaken elec-
trostatic or hydrogen-bonding interactions that may have existed
between side chain functional groups and the amine of the
N-terminus.25
Although the thiazine rearrangement kinetics data appear
suggestive of side-chain specific interactions that may influence the
rearrangement reaction, a more nuanced analysis via modeling of the
stereochemical constraints and probabilities of side-chain dihedral
angles is necessary in order to fully elucidate the observed differences
in rates of rearrangement for the CXF peptides.18,19 However, we
postulated that the rate of thiazine conversion for CGF-MPA adducts
was among the most rapid due to the small size of glycine and the
presumed freedom of rotation around the Cys-Gly bond. In the case
of CKF, the rationale for the faster rate of rearrangement was less
clear. Although intramolecular base catalysis has been reported for
It is also important to note that for the CLF peptide, a higher
percentage by volume of acetonitrile (26%) was present in the con-
jugation solution compared to the other experiments. To evaluate
the effect of the greater organic solvent content, the conjugation
studies were also performed using the CGF peptide and MPA in
pH 7.4 potassium phosphate buffer solution containing 26% aceto-
nitrile (supporting information). In comparison with the CGF-MPA
studies with 10% acetonitrile cosolvent, the rate of thiazine
rearrangement was 1.3 times slower. These results support the
supposition that the slower rate of thiazine conversion for the CLF
peptide was also due to the physiochemical properties of leucine
rather than just the presence of more acetonitrile in the conjuga-
tion solution.
TABLE 1 Thiazine conversion for CXF-MPA adducts at pH 5.0
after 336 h
Thiazine conversion after 336 h
Peptide conjugate
CGF-MPA
pH 5.0 (%)
1.1
2.6
|
Investigation of maleimide N-substitution on
thiazine rearrangement
CEF-MPA
1.0
CKF-MPA
0.5
Previous reports have shown that maleimides with more electron-
withdrawing substituents at the nitrogen of the maleimide have
greater electrophilic character.16,20,26 This leads to more rapid hydro-
lysis of both the maleimide and the cysteine-conjugated thioether
adduct. Two linkers, maleimido (hydroxyhexyl)hexanamide (MHH) and
a methoxy PEG maleimide with an average molecular weight of
10,000 Da (PEGMA10K) were chosen to compare with MPA
(Figure 6). Because its longer alkyl chain is less electron withdrawing
than the propionic acid linker of MPA, we anticipated that the MHH
linker may provide higher stability toward conversion of peptide con-
jugates to the thiazine isomer. Additionally, PEGMA10K was selected
in order to assess the analytical challenge of detecting the thiazine
isomer in the context of a polydisperse, large molecular weight
conjugate. PEGMA10K was suitable for comparison to MPA because
CSF-MPA
0.9
CLF-MPA
0.3
Abbreviation: MPA, 3-maleimidopropionic acid.
TABLE 2 Rate of thiazine formation for CXF-MPA adducts at pH
7.3 and pH 8.4
Thiazine conversion after 1 h
Peptide conjugate
CGF-MPA
pH 7.3 (%)
6.1
pH 8.4 (%)
17.6
CEF-MPA
3.4
8.4
CKF-MPA
8.7
16.8
CSF-MPA
4.8
11.2
it maintains
a similar linker length (maleimido propionic amide
CLF-MPA
4.3
6.6
vs. maleimido propionic acid) but has a significantly higher overall
molecular weight.
Abbreviation: MPA, 3-maleimidopropionic acid.

GOBER ET AL.
7 of 11
The conjugation of CXF peptides to MHH was performed at
conversion than the MPA counterparts were. The greater stability of
the MHH linker was likely due to its longer hexanamide alkyl chain
that is inductively electron donating with respect to the maleimido
nitrogen. Despite the greater stability of the MHH linker, these data
support the conclusion that the susceptibility of succinimidyl
thioethers to thiazine rearrangement is a general property of all
maleimido linkers.
pH 7.3 in 0.1-M potassium phosphate buffer with 26% acetonitrile
added to ensure solubility during the course of the reaction (for
conjugation of CLF to MHH, 28% acetonitrile was present). The
thiazine concentration was measured at 0, 0.5, 1, and 2 h by
UHPLC. As seen in Figure 7, CGF reacted rapidly with MHH to
give complete conjugation within 30 min. The emergence of
a
new peak corresponding to the thiazine isomer was observed over
time. After 24 h, about 33% of the succinimidyl thioether adduct
was converted to the thiazine isomer. When compared to the
generation of the CGF-MPA thiazine adduct at pH 7.3 with 26%
acetonitrile cosolvent, the rate of thiazine rearrangement for the
CGF-MHH adduct was over two times slower. In addition, the
extent of hydrolysis of MHH over time was significantly lower
compared to MPA as the majority of the maleimide remained
intact after 24 h. Overall, these results demonstrated that linker
chemistry has a notable effect on the degree of thiazine side
reaction encountered for maleimide conjugated N-terminal cysteine
containing peptides.
For PEGMA10K, conjugation was performed using CGF and CKF,
the two peptides to give the fastest rates of thiazine rearrangement in
the MPA and MHH studies. Reactions were run in two steps to facili-
tate monitoring the progress of the reaction.
First, the CXF peptide and PEGMA10K were dissolved in
unbuffered water with 10% acetonitrile at a ratio of 2:1 peptide to
maleimide. After 1 h, a second equivalent of PEGMA10K was added,
and the conjugation reaction was allowed to proceed for another
hour. The pH of the reaction was adjusted to pH 7.7 using 0.1-M
potassium phosphate solution with 10% acetonitrile, and the reaction
was quenched with dilute TFA in water after 48 h. As shown in
Figure 9, CGF reacted completely with PEGMA10K after the addition
of the second portion of the maleimide. After increasing the pH, a
new peak with a slightly shifted retention time was observed, and the
peak corresponding to the CGF-PEGAMA10K succinimidyl thioether
was absent. When CKF was used during the conjugation with
PEGMA10K, conversion of the CKF-PEGMA10K adduct to the thia-
zine isomer could be detected even in unbuffered water with 10%
acetonitrile upon the addition of the second equivalent of
PEGMA10K as evidenced by the emergence of a second peak in the
UHPLC chromatogram (Figure 10, green trace). After base stressing
the conjugation reaction mixture at pH 7.7 for 48 h, nearly complete
conversion to the thiazine isomer was observed (Figure 10, orange
trace).
Conjugation of CKF, CEF, CSF, and CLF peptides to MHH pro-
duced succinimidyl thioether adducts that also suffered from thiazine
rearrangement over time but at slower rates. Compared to the MPA
studies, a similar order of reactivity was also observed with CGF and
CKF giving the fastest rates of rearrangement and CLF and CEF
producing the slowest rates (Figure 8 and Table S6.1). Interestingly, all
MHH conjugates were between 3 and 5 times slower toward thiazine
Direct infusion MS/MS analysis was used to confirm the presence
of the thiazine isomer for both the CGF and CKF conjugation reac-
tions after incubation with PEGMA10K at pH 7.7 for 48 h (supporting
information). Fragmentation of the CGF-PEGMA10K thiazine isomer
produced a unique mass fragment with a mass of 406.11 Da, which
was not possible for the succinimidyl thioether isomer. Similarly, frag-
mentation of the CKF-PEGA10K adduct gave a 477.18-Da mass frag-
ment, which was unique for the thiazine isomer. These results
demonstrated the power of our MS/MS strategy for unambiguous
FIGURE 6 Chemical structures of the maleimido (hydroxyhexyl)
hexanamide (MHH) and methoxy PEG maleimide MW: 10 kDa
(PEGMA10K) maleimide reagents
FIGURE 7 Ultra-high performance liquid
chromatography (UHPLC) analysis of H-CGF-OH
conjugation to maleimido (hydroxyhexyl)
hexanamide (MHH) at pH 7.3. The reaction
mixture was analyzed at 0 h (black), 0.5 h (blue),
1 h (cyan), 2 h (orange), and 24 h (green)

8 of 11
GOBER ET AL.
FIGURE 10 Ultra-high performance liquid chromatography
(UHPLC) analysis of H-CKF-OH conjugation to methoxy PEG
maleimide MW: 10 kDa (PEGMA10K). Black trace: conjugation of
CKF to PEGMA10K (2:1 ratio of peptide to maleimide) in unbuffered
water with 10% acetonitrile. Green trace: conjugation of CKF to
PEGMA10K (1:1 ratio of peptide to maleimide) in unbuffered water
with 10% acetonitrile. Orange trace: thiazine rearrangement of CKF-
PEGMA10K conjugate after 48 h (1:1 ratio of peptide to maleimide)
at pH 7.7 in 0.1-M potassium phosphate solution with 10%
acetonitrile
FIGURE 8 Increase in thiazine formation over time for different
neighboring amino acids of CXF-maleimido (hydroxyhexyl)
hexanamide (MHH) adducts at pH 7.3 in 0.1-M potassium phosphate
buffer with 26% acetonitrile. For CLF, the conjugation reaction was
carried out in 0.1-M potassium phosphate buffer with 28%
acetonitrile added to maintain solubility
3
|
CONCLUSION
In this work, we explored the pH and sequence dependence of the
rearrangement of succinimidyl thioethers to thiazine isomers in the
context of maleimide conjugation to N-terminal cysteine containing
peptides. A prominent increase in the rate of thiazine formation was
observed at basic pH, which was consistent with previous work indi-
cating that the rearrangement is base promoted. When the amino acid
adjacent to the N-terminal cysteine was substituted for various amino
acids, we observed generation of the thiazine impurity in all instances,
though the rates of thiazine formation differed for the CXF peptides
that were studied. Using a maleimide linker that we anticipated to be
more stable also gave substantial thiazine formation, which provides
additional support that the side reaction is general.
FIGURE 9 Ultra-high performance liquid chromatography
(UHPLC) analysis of H-CGF-OH conjugation to methoxy PEG
maleimide MW: 10 kDa (PEGMA10K). Black trace: conjugation of
CGF to PEGMA10K (2:1 ratio of peptide to maleimide) in unbuffered
water with 10% acetonitrile. Green trace: thiazine rearrangement of
CGF-PEGMA10K conjugate after 48 h (1:1 ratio of peptide to
maleimide) at pH 7.7 in 0.1-M potassium phosphate solution with
10% acetonitrile
NMR analysis conclusively confirmed the presence of the thiazine
impurity, which we observed as a significant side reaction during the
conjugation of CXF peptides to MPA, MHH, and PEGMA10K mal-
eimides. We also developed a powerful MS/MS strategy for identify-
ing the thiazine isomer that has general applicability and is
independent of the N-substitution of the maleimide. This provides
excellent analytical utility because MS/MS capabilities are common to
many research labs, while instrumentation for advanced 2-D NMR
experiments and the requisite expertise to perform such experiments
are less common.
detection of the thiazine isomer. Because of the nonuniformity of
pegylated maleimide reagents, analysis of thiazine rearrangement of
succinimidyl thioethers by routine LC–MS presents an analytical chal-
lenge. Typically, confirmation of the rearrangement would require
purification and isolation of the desired isomer followed by advanced
2-D NMR experiments. However, an MS/MS detection approach
greatly simplified the detection of the thiazine isomer and even
allowed confirmation of the rearrangement without preparative
chromatographic separation or specialized sample preparation prior to
MS/MS analysis.
Due to the general nature of the thiazine side reaction, it is advis-
able to avoid the use of N-terminal cysteine in peptide conjugates
where the succinimidyl thioether linkage is desired. Although per-
forming the conjugation under acidic conditions near pH 5 prevents
thiazine formation, the subsequent purification, storage, and applica-
tion of the peptide conjugates must also be carried out under acidic
conditions to avoid loss of the succinimidyl thioether. Alternatively,
acetylation of the N-terminal cysteine can be performed to prevent
formation of the thiazine impurity (supporting information).12,13

GOBER ET AL.
9 of 11
4
|
MATERIALS AND METHODS
| Materials
15.4 mmol/2.3 eq, 5.7 mmol). All reactions were performed at room
temperature. For Fmoc removal, 20% (v/v) piperidine in DMF was
used. Cleavage of the peptidyl-resin to obtain crude peptides was
performed using TFA/TIS/water/EDT (92.5:2.5:2.5:2.5 v/v). Crude
peptides were analyzed by reversed-phase UHPLC (Waters Acquity
HSS T3 100 Å, 1.8 μm, 2.1 × 150 mm) using a linear gradient of 0.1%
TFA in acetonitrile. Peptides were purified by preparative reversed-
phase HPLC using Luna ® C18(3) 10-μm particle size media packed in
an axial compression column (ModCol ®, 25 × 5 iD). After purification,
the peptides were lyophilized to give white powders. No correction
for TFA and water content was performed prior to weighing out
peptides for the maleimide conjugation experiments.
4.1
MHH was purchased from AldLab (Woburn, MA), and PEGMA10K
was purchased from JenKem Technology USA (Plano, TX).
Dimethylformamide was purchased from Univar Solutions Inc.
(Downers Grove, IL). N,N-diisopropylethylamine was purchased from
Spectrum Laboratory Products Inc. (Gardena, CA). Piperidine was pur-
chased from Tedia Company Inc. (Fairfield, OH). Isopropyl alcohol was
purchased from Brenntag Pacific Inc. (Santa Fe Springs, CA).
Diisopropyl ether was purchased from Neuchem (Sparks, NV).
N,N0-diisopropylcarbodiimide, triisopropylsilane, 1,2-ethanedithiol,
potassium phosphate monobasic, and potassium phosphate dibasic
were purchased from Sigma-Aldrich (St. Louis, MO). HPLC grade
trifluoroacetic acid was purchased from Alfa Aesar (Haverhill, MA).
HPLC grade water and acetonitrile were obtained from
MilliporeSigma (Burlington, MA). All other amino acid derivatives and
reagents were obtained from Bachem AG, Bubendorf, Switzerland.
Preparative HPLC purification was performed using a Waters
Prep LC system. HPLC and UHPLC analyses were performed on a
Thermo Scientific UltiMate 3000 UHPLC system using an Atlantis T3
100 Å, 3 μm, 4.6 mm × 150-mm column and a Waters Acquity HSS
T3 100 Å, 1.8 μm, 2.1 × 150 mm column, respectively. A Thermo Sci-
entific Vanquish UHPLC system with a Waters Acquity CSH C18
130 Å, 1.7 μm, 2.1-mm × 150-mm column was used for UHPLC–MS
separations. MS analysis was performed in positive ion mode on a
Thermo Scientific Q Exactive Focus Hybrid Quadrupole-Orbtitrap™
mass spectrometer. For MS/MS analyses, high-energy C-trap dissocia-
tion (HCD) was used for fragmentation.
4.3
|
Maleimide conjugation time course
experiments
4.3.1
|
MPA
A solution of MPA (2 mg/ml) and a solution of CXF peptide (2 mg/ml)
were prepared separately in aqueous 0.1-M potassium phosphate
solution in 10:90 acetonitrile: water (for CLF-MPA conjugation, the
peptide and MPA were dissolved in 0.1-M potassium phosphate
solution containing 26:74 acetonitrile: water to help solubilize the
peptide). Reactions were initiated by mixing the solution of MPA with
the solution of CXF. The reaction was performed in a sealed microce-
ntrifuge tube at ambient temperature. An aliquot of the reaction
solution was removed at the appropriate time point and was
quenched using an equal volume of a solution of 1% TFA in water.
The apparent pH of the solutions used during the CXF and
maleimide conjugations was measured using a Mettler Toledo™
SevenCompact™ S220-Basic pH/ion benchtop meter equipped with a
Mettler Toledo™ InLab Solids Go-ISM® electrode (pH 1–11; 0–80ꢀC).
The electrode was calibrated before each series of measurements
using standard buffer solutions purchased from Millipore Sigma
(Burlington, MA). All measurements were taken at ambient
temperature.
4.3.2
|
MHH
A solution of MHH (2 mg/ml) was dissolved in 50% acetonitrile in
water, and a solution of CXF peptide (2 mg/ml) was prepared in aque-
ous 0.1-M potassium phosphate solution in 10:90 acetonitrile: water
(for CLF-MPA conjugation, the peptide and MHH were dissolved in
0.1-M potassium phosphate solution containing 28:72 acetonitrile in
water). Reactions were initiated by mixing the solution of MPA with
the solution of CXF. The reaction was performed in a sealed microce-
ntrifuge tube at ambient temperature. An aliquot of the reaction
solution was removed at the appropriate time point and was
quenched using an equal volume of a solution of 1% TFA in water.
4.2
|
Manual Fmoc solid-phase peptide synthesis
The peptides H-Cys-Gly-Phe-OH (CGF), H-Cys-Leu-Phe-OH (CLF),
H-Cys-Glu-Phe-OH (CEF), H-Cys-Lys-Phe-OH (CKF), and H-Cys-Ser-
Phe-OH (CSF) were synthesized on a 2.5-mmol scale by first loading
Fmoc-Phe-OH onto 2-chlorotrityl chloride resin (10-mmol scale,
substitution = 1.3 mmol/g) and then splitting the resin into four lots
for the subsequent divergent coupling reactions. Loading of the first
amino acid derivative was carried out using Fmoc-Phe-OH (1.0 eq.,
10 mmol) and DIPEA (1.0 eq., 10 mmol) in dimethylformamide (DMF).
Coupling of the second amino acid derivative was performed using
TCTU/DIPEA (1.9 eq., 4.8 mmol/2.5 eq., 6.3 mmol) in DMF. Coupling
of Fmoc-Cys (Trt)-OH was performed using DIC/HOBt (6.2 eq,
4.3.3
|
PEGMA10K
A solution of PEGMA10K (25 mg/ml) and CXF peptide (1.55 mg/ml)
was dissolved in unbuffered water with 10% acetonitrile. The solu-
tions of peptide and maleimide were mixed and allowed to react for
1 h. After 1 h, another portion of PEGMA10K (22 mg/ml) in
unbuffered water with 10% acetonitrile was added. The reaction was
allowed to proceed for another hour before the pH was adjusted to

10 of 11
GOBER ET AL.
pH 7.7 using 0.1-M potassium phosphate solution in 10:90 acetoni-
trile:water. Reactions were performed in sealed amber vials at ambi-
ent temperature. Reaction solutions at pH 7.7 were quenched with an
equal volume of a solution of 1% TFA in water.
ACKNOWLEDGEMENTS
We thank Prof. Dr. Oliver Zerbe and Nadja Bross for the acquisition
and interpretation of the NMR data. We thank Dr. Sumukh Ray and
Dr. Khanh Nguyen for their contributions in acquiring the UHPLC–MS
and MS/MS data.
4.4
|
Preparation and purification of CGF-MPA
ORCID
thioether and CGF-MPA thiazine adducts
Isaiah N. Gober
https://orcid.org/0000-0002-5896-1520
https://orcid.org/0000-0001-9944-3386
Alexander J. Riemen
4.4.1
|
Preparation and purification of CGF-MPA
Matteo Villain
https://orcid.org/0000-0002-6116-8742
thioether
REFERENCES
To a solution of 500 mg of H-CGF-OH in 25 ml of 10:90 acetonitrile:
water was added a solution of 500-mg MPA in 25 ml of 10:90 aceto-
nitrile:water. The two solutions were mixed together to initiate the
conjugation reaction. The reaction progress was monitored by taking
500-μl aliquots and adding 500 μl of 0.2% TFA in water to the solu-
tions, which were subsequently analyzed by analytical HPLC. Once
the reaction was complete, the two isomer peaks of the CGF-MPA
conjugate were purified by preparative reversed-phase HPLC using
Luna ® C18(3) 10-μm particle size media packed in an axial compres-
sion column (ModCol ®, 25 × 5 iD). The peptide conjugate isomers
were lyophilized to give white powders.
1. Koniev O, Wagner A. Developments and recent advancements in the
field of endogenous amino acid selective bond forming reactions for
bioconjugation. Chem Soc Rev. 2015;44(15):5495-5551. https://doi.
org/10.1039/C5CS00048C
2. Gunnoo SB, Madder A. Chemical protein modification through cyste-
ine. ChemBioChem. 2016;17(7):529-553. https://doi.org/10.1002/
cbic.201500667
3. Ravasco JMJM, Faustino H, Trindade A, Gois PMP. Bioconjugation
with maleimides: a useful tool for chemical biology. Chem A Eur J.
2019;25(1):43-59. https://doi.org/10.1002/chem.201803174
4. Kipper K, Eremina N, Marklund E, et al. Structure-guided approach
to site-specific fluorophore labeling of the lac repressor LacI. PLoS
One. 2018;13(6):e0198416. https://doi.org/10.1371/journal.pone.
0198416
5. Anderson CR, Rychak JJ, Backer M, Backer J, Ley K, Klibanov AL.
ScVEGF microbubble ultrasound contrast agents: a novel probe for
ultrasound molecular imaging of tumor angiogenesis. Invest Radiol. 2010;
45(10):579-585. https://doi.org/10.1097/RLI.0b013e3181efd581
6. Ramachandiran V, Grigoriev V, Lan L, Ravkov E, Mertens SA,
Altman JDA. Robust method for production of mhc tetramers with
4.4.2
thiazine
|
Preparation and purification of CGF-MPA
To a solution of 500 mg of H-CGF-OH in 25 ml of pH 8.5, 20-mM
potassium phosphate buffer solution (10:90 acetonitrile:water) was
added a solution of 500-mg MPA in 25 ml of pH 8.5, 20-mM
potassium phosphate buffer solution (10:90 acetonitrile:water). The
two solutions were mixed together to initiate the conjugation
reaction. The reaction progress was monitored by taking 500-μl
aliquots and adding 500 μl of 1% TFA in water to the solutions,
which were subsequently analyzed by UHPLC. Once the reaction
was complete, the two isomer peaks of the CGF-MPA conjugate
were purified by preparative reversed-phase HPLC using Luna ®
C18(3) 10-μm particle size media packed in an axial compression
column (ModCol ®, 25 × 5 iD). The peptide conjugate was
lyophilized to give a white powder.
small molecule fluorophores.
J Immunol Methods. 2007;319(1-2):
13-20. https://doi.org/10.1016/j.jim.2006.08.014
ꢀ
7. Nawrocka D, Krzyscik MA, Opalinski Ł, Zakrzewska M, Otlewski J.
Stable fibroblast growth factor 2 dimers with high pro-survival and
mitogenic potential. Int J Mol Sci. 2020;21(11):4108. https://doi.org/
10.3390/ijms21114108
8. Adumeau P, Davydova M, Zeglis BM. Thiol-reactive bifunctional
chelators for the creation of site-selectively modified radio-
immunoconjugates with improved stability. Bioconjug Chem. 2018;
29(4):1364-1372. https://doi.org/10.1021/acs.bioconjchem.8b00081
9. Tian F, Lu Y, Manibusan A, et al. General approach to site-specific
antibody drug conjugates. Proc Natl Acad Sci U S a. 2014;111(5):
1766-1771. https://doi.org/10.1073/pnas.1321237111
10. Nair DP, Podgórski M, Chatani S, et al. The thiol-michael addition
click reaction: a powerful and widely used tool in materials chemistry.
Chem
Mater.
2014;26(1):721-744.
https://doi.org/10.1021/
cm402180t
11. Northrop BH, Frayne SH, Choudhary U. Thiol-maleimide “click”
chemistry: evaluating the influence of solvent, initiator, and thiol on
the reaction mechanism, kinetics, and selectivity. Polym Chem. 2015;
6(18):3415-3430. https://doi.org/10.1039/c5py00168d
4.5
|
NMR experiments
For characterization, 1-D 1H and 13C spectra were collected, and
12. Tong H, Zheng Y, Zhou L, et al. Enzymatic Cleavage and Subsequent
facile intramolecular transcyclization for in situ fluorescence detec-
tion of γ-glutamyltranspetidase activities. Anal Chem. 2016;88(22):
10816-10820. https://doi.org/10.1021/acs.analchem.6b03448
13. Li X, Zheng Y, Tong H, et al. Rational design of an ultrasensitive and
highly selective chemodosimeter by a dual quenching mechanism for
cysteine based on a facile michael-transcyclization cascade reaction.
various 2-D NMR experiments were performed including 1H–1H
C C HMBC, and ROESY. All
DQF–COSY, 1H–13 HSQC, 1H–13
spectra were recorded at 600 MHz on a Bruker AVNeo-600 NMR
spectrometer at 277 K. Spectra were referenced to internal
trimethylsilylpropanoic acid (TSP). Samples were prepared by
dissolving 10 mg of the lyophilized peptide in 600 ml of D2O with
0.1% TFA-d.
Chem
A Eur J. 2016;22(27):9247-9256. https://doi.org/10.1002/
chem.201601126

GOBER ET AL.
11 of 11
14. Mayr J, Hager S, Koblmüller B, et al. EGFR-targeting peptide-coupled
platinum (IV) complexes. J Biol Inorg Chem. 2017;22(4):591-603.
https://doi.org/10.1007/s00775-017-1450-7
23. Pace CN. Energetics of protein hydrogen bonds. Nat Struct Mol Biol.
2009;16(7):681-682. https://doi.org/10.1038/nsmb0709-681
24. Sheu S-Y, Yang D-Y, Selzle HL, Schlag EW. Energetics of hydrogen
15. Ochtrop P, Hackenberger CPR. Recent advances of thiol-selective
bioconjugation reactions. Curr Opin Chem Biol. 2020;58:28-36.
https://doi.org/10.1016/j.cbpa.2020.04.017
bonds in peptides. Proc Natl Acad Sci U S A. 2003;100(22):
12683-12687. https://doi.org/10.1073/pnas.2133366100
25. Lindman S, Xue WF, Szczepankiewicz O, Bauer MC, Nilsson H,
Linse S. Salting the charged surface: pH and salt dependence of pro-
tein G B1 stability. Biophys J. 2006;90(8):2911-2921. https://doi.org/
10.1529/biophysj.105.071050
16. Fontaine SD, Reid R, Robinson L, Ashley GW, Santi DV. Long-term
stabilization of maleimide-thiol conjugates. Bioconjug Chem. 2015;
26(1):145-152. https://doi.org/10.1021/bc5005262
17. Schoefberger W, Schlagnitweit J, Müller N. Recent Developments in
Heteronuclear Multiple-Bond Correlation Experiments. Vol 72. 2011.
https://doi.org/10.1016/B978-0-12-385857-3.00001-3.
18. Zhou AQ, Caballero D, O'Hern CS, Regan L. New insights into the
interdependence between amino acid stereochemistry and protein
structure. Biophys J. 2013;105(10):2403-2411. https://doi.org/10.
1016/j.bpj.2013.09.018
26. Christie RJ, Fleming R, Bezabeh B, Woods R, Mao S, Harper J,
Joseph A, Wang Q, Xu ZQ, Wu H, Gao C, Dimasi N. Stabilization of
cysteine-linked antibody drug conjugates with N-aryl maleimides.
J Controlled Release. 2015;220(Pt B):660–670. https://doi.org/10.
1016/j.jconrel.2015.09.032.
19. Zhou AQ, O'Hern CS, Regan L. Predicting the side-chain dihedral
angle distributions of nonpolar, aromatic, and polar amino acids using
hard sphere models. Proteins Struct Funct Bioinf. 2014;82(10):
2574-2584. https://doi.org/10.1002/prot.24621
SUPPORTING INFORMATION
Additional supporting information may be found online in the
Supporting Information section at the end of this article.
20. Lyon RP, Setter JR, Bovee TD, et al. Self-hydrolyzing maleimides
improve the stability and pharmacological properties of antibody-drug
conjugates. Nat Biotechnol. 2014;32(10):1059-1062. https://doi.org/
10.1038/nbt.2968
How to cite this article: Gober IN, Riemen AJ, Villain M.
Sequence sensitivity and pH dependence of maleimide
conjugated N-terminal cysteine peptides to thiazine
rearrangement. J Pep Sci. 2021;e3323. https://doi.org/10.
1002/psc.3323
21. Grimsley GR, Scholtz JM, Pace CNA. Summary of the measured PK
values of the ionizable groups in folded proteins. Protein Sci. 2009;
18(1):247-251. https://doi.org/10.1002/pro.19
22. Hendsch ZS, Tidor B. Electrostatic interactions in the GCN4 leucine
zipper: substantial contributions arise from intramolecular interac-
tions enhanced on binding. Protein Sci. 1999;8(7):1381-1392. https://
doi.org/10.1110/ps.8.7.1381