Pharmaceutical Excipients Enhance Iron-Dependent Photo-Degradation in Pharmaceutical Buffers by near UV and Visible Light: Tyrosine Modification by Reactions of the Antioxidant Methionine in Citrate Buffer

Article abstract of DOI:10.1007/s11095-021-03042-8

Purpose: To evaluate the effect of excipients, including sugars and amino acids, on photo-degradation reactions in pharmaceutical buffers induced by near UV and visible light. Methods: Solutions of citrate or acetate buffers, containing 1 or 50?μM Fe³⁺, the model peptides methionine enkephalin (MEn), leucine enkephalin (LEn) or proctolin peptide (ProP), in the presence of commonly used amino acids or sugars, were photo-irradiated with near UV or visible light. The oxidation products were analyzed by reverse-phase HPLC and HPLC-MS/MS. Results: The sugars mannitol, sucrose and trehalose, and the amino acids Arg, Lys, and His significantly promote the oxidation of peptide Met to peptide Met sulfoxide. These excipients do not increase the yields of hydrogen peroxide, suggesting that other oxidants such as peroxyl radicals are responsible for the oxidation of peptide Met. The addition of free Met reduces the oxidation of peptide Met, but, in citrate buffer, causes the addition of Met oxidation products to Tyr residues of the target peptides. Conclusions: Commonly used excipients enhance the light-induced oxidation of amino acids in model peptides.

Full text of DOI:10.1007/s11095-021-03042-8

Pharm Res
https://doi.org/10.1007/s11095-021-03042-8
RESEARCH PAPER
Pharmaceutical Excipients Enhance Iron-Dependent
Photo-Degradation in Pharmaceutical Buffers by near UV
and Visible Light: Tyrosine Modification by Reactions
of the Antioxidant Methionine in Citrate Buffer
Natalia Subelzu1,2 & Christian Schöneich1
Received: 4 January 2021 /Accepted: 5 April 2021
#
The Author(s), under exclusive licence to Springer Science+Business Media, LLC, part of Springer Nature 2021
ABSTRACT
ABBREVIATIONS
Purpose To evaluate the effect of excipients, including sugars
and amino acids, on photo-degradation reactions in pharma-
ceutical buffers induced by near UV and visible light.
Methods Solutions of citrate or acetate buffers, containing
1 or 50 μM Fe3+, the model peptides methionine enkeph-
alin (MEn), leucine enkephalin (LEn) or proctolin peptide
(ProP), in the presence of commonly used amino acids or
sugars, were photo-irradiated with near UV or visible light.
The oxidation products were analyzed by reverse-phase
HPLC and HPLC-MS/MS.
Results The sugars mannitol, sucrose and trehalose, and the
amino acids Arg, Lys, and His significantly promote the oxi-
dation of peptide Met to peptide Met sulfoxide. These exci-
pients do not increase the yields of hydrogen peroxide, sug-
gesting that other oxidants such as peroxyl radicals are respon-
sible for the oxidation of peptide Met. The addition of free
Met reduces the oxidation of peptide Met, but, in citrate buff-
er, causes the addition of Met oxidation products to Tyr res-
idues of the target peptides.
ABTS
2,2′-azino-bis(3-ethylebenzothiazole-6-sulfonic
acid) diammonium salt
Arginine
Arg
DTPA
EDTA
FA
Diethylenetriaminepentaacetic acid
Ethylenediaminetetraacetic acid
Formic acid
Gly
Glycine
His
Histidine
HPLC
HRP
LC
LEn
Leu
High-performance liquid chromatography
Horseradish peroxidase
Liquid chromatography
Leucine enkephalin
Leucine
Lys
Lysine
mAbs
MEn
Met
MS
Phe
ProP
Monoclonal antibodies
Methionine enkephalin
Methionine
Mass spectrometry
Phenylalanine
Conclusions Commonly used excipients enhance the light-
induced oxidation of amino acids in model peptides.
Proctolin peptide
Q-TOF Quadrupole time of flight
TFA
Thr
tR
Ty r
UV
Trifluoroacetic acid
Threonine
Retention time
Tyrosine
.
.
KEY WORDS excipients near UVand visible light
.
.
pharmaceutical peptides and proteins photo-degradation
photo-Fenton reaction
Ultraviolet
INTRODUCTION
Pharmaceutical proteins and peptides are susceptible to light-
induced degradation (1–3). The wavelengths of light that
pharmaceutical products are exposed to vary between visible
and near UV light, which some authors refer to as ambient
light (4, 5). Amino acid residues do not significantly absorb
near UV or visible light, except for some oxidation products of
* Christian Schöneich
schoneic@ku.edu
1
Department of Pharmaceutical Chemistry, University of Kansas, 2095
Constant Avenue, Lawrence, Kansas 66047, USA
2
Present address: Merck & Co., Inc, Rahway, NJ, USA

Pharm Res
Trp residues, e.g. N-formylkynurenine, kynurenine, and 3-
hydroxykynurenine (6–9). The exposure of proteins to ambi-
ent light can occur during manufacturing, packaging, visual
inspection of products, storage, and patient administration (5,
10–13). Such light exposure can induce protein degradation,
including the formation of post-translational modifications
and aggregates (4, 5). As a potential consequence of such type
of modifications, several side effects have been reported, in-
cluding the reduction of clinical efficacy or adverse immuno-
genic responses (14–17). There are also reports of aesthetic
changes in drug products, such as cloudiness, change of color,
or the formation of visible particles (18–22).
Mechanistically, the processes leading to protein photo-
degradation during near UV and visible light exposure are
not well established. Recently, we evaluated pharmaceutical
buffers for their propensity to promote oxidation mechanisms
via the photo-Fenton reaction under near UV and visible light
(23). Specifically citrate buffer generated a variety of reactive
species (Scheme 1), leading to multiple oxidation products of
Met- and Tyr-containing model peptides, while acetate and
succinate promoted only the formation of Met sulfoxide (23).
Here, we have extended our studies to the role of additional
excipients, specifically sugars and amino acids, in oxidation
reactions induced by near UV and visible light in pharmaceu-
tical buffers.
In order to design a stable formulation with an adequate
shelf life, it is frequently necessary to add multiple excipients to
formulations (24). These excipients function to improve the
physical and chemical characteristics of the drug and consti-
tute a vehicle for drug delivery (25). Studies have demonstrat-
ed that a combination of excipients can have a synergetic
effect on the stabilization of formulations (26–28). However,
some excipients are not inert and can afford changes in the
stability of a formulation (23).
Excipients are classified with respect to their effects on for-
mulations. For parenteral formulations, these categories are
buffers, tonicifiers or osmolites, surfactants, solubilizers, pres-
ervatives, and stabilizing agents. Chelating agents (e.g., EDTA
or DTPA) are frequently used to prevent oxidation promoted
by transition metals. Moreover, amino acids and sugars or
polyols are generally used to prevent aggregation, and/or ox-
idation (24).
Parenteral protein formulations are usually formulated in a
pH range between 4 to 7, making the selection of buffer a critical
factor. The most commonly used buffers are citrate, succinate,
acetate, phosphate, tris(hydroxymethyl)aminomethane (Tris),
Gly, and His (25, 29). The selection of buffer ultimately depends
on the stability profiles of the respective proteins (26).
Sugars and amino acids, especially Arg, His, Gly, Lys, and
Met, are commonly used stabilizers in parenteral formulations
Scheme 1 Photo-Fenton reaction in citrate buffer.

Pharm Res
(26), where positively charged amino acids can stabilize proteins
by electrostatic interactions with acidic amino acid side chains
(26, 30–35). However, Arg, Gly, and Lys can both stabilize or
denaturate proteins, depending on their concentration and the
nature of the proteins (30). Arg can increase protein solubility,
and prevent protein unfolding and aggregation induced by heat
or light (36). Lys tends to be less of a stabilizer compared to the
other amino acids (30, 37). Met has been used as an antioxidant
due to its ability to react with hydrogen peroxide (H2O2) and
other reactive oxygen species (11, 28, 38).
Sugars stabilize proteins by increasing the protein-water
potential, preferred hydration of proteins, and by increasing
the activation energy for protein unfolding (39–41). Sugars
also diminish protein aggregation during freeze-thaw and
freeze-drying cycles (42, 43). However, depending on their
structure, sugars can also promote protein degradation, for
example via glycation (44).
Azino-bis (3-ethylebenzothiazole-6-sulfonic acid) diammo-
nium salt (ABTS) was from Fluka (Saint Louis, MO). All
chemicals used were analytical grade.
Peptide Photo-Irradiation
Solutions were prepared at a final volume of 400 μL, contain-
ing 500 μM peptide in water or 10 mM buffer (unless other
specified concentration), pH 6.0, 1.0 or 50 μM Fe3+, and
either 50 mM amino acids or 120 mM sugars. Photo-
irradiation was performed in a Rayonet© reactor (The
Southern New England Ultraviolet Company, Branfort,
CT), using near UV (RPR-3500 Å, λmax = 350 nm; band-
width 305–416 nm) or visible light lamps (RPR-4190 Å,
λmax = 419 nm; bandwidth 380–480 nm).
HPLC Analysis
Our data will show that amino acids and sugars can enhance
the oxidative degradation of various model peptides (methio-
nine- and leucine-enkephalin, and proctolin peptide) in iron-
containing citrate and acetate buffer during the exposure of
formulations to small doses of near UV or visible light. We
selected iron concentrations of 1 and 50 μM, where 1 μM is
in the range of iron levels reported for some formulations of
monoclonal antibodies (1–9 μM) (45, 46), and 50 μM was
selected for mechanistic purposes. The guideline for elemental
impurities of the International Council for Harmonisation
(ICH), ICHQ3D, sets no specific limits for the permitted daily
exposure (PDE) to iron (47). We applied light doses between 9
and 35 Whm−2, where 35 Whm−2 correpond to ca. 2 h expo-
sure to UV A light of combined indoor lighting and window-
filtered daylight in a home or clinical environment (48).
The analysis of the oxidation products was achieved by
reverse-phase HPLC. The photo-irradiated solutions (50 μL)
were injected onto a C18 Vydac column (250 mm, I.D.
4.6 mm, 5 μm; Grace, Columbia, MD), maintained at 35°C
in a column heater (CTO-20A; Shimadzu, Columbia, MD).
Mobile phase A was 0.1% TFA in water, and mobile phase B
consisted of acetonitrile with 0.1% TFA. The separation of the
oxidation products was performed with a linear gradient
changing mobile phase B from 7% to 35% within 30 min,
delivered by two pumps (LC-20AT; Shimadzu, Columbia,
MD), and the elution was monitored with a UV-detector
(SPD-10Av; Shimadzu, Columbia, MD) set at λ = 214 nm.
HPLC-MS/MS Analysis
The HPLC-MS instrument consisted of a nanoAcquity
UPLC coupled to a Xevo Q-TOF mass spectrometer
(Waters Corp, Milford, MA). Mobile phase A was 0.1% for-
mic acid in water, and mobile phase B was acetonitrile con-
taining 0.1% formic acid (FA). Solutions of the photo-
irradiated peptides were diluted in water containing 0.1%
FA, and 2 μL were injected onto a CSH C18 nanocolumn
(150 mm, I.D. 75 μm, Waters Corp, Milford, MA).
Separation of the products was performed with a linear gra-
dient increasing mobile phase B from 10 to 40% within
30 min. The MSe data were acquired within a mass range
between 50 to 2000 Da, with a ramp collision energy from
20 to 30 V and the cone voltage was set at 40 V. Data analysis
was achieved with the MassLynx™ Software (Waters Corp;
Milford, MA).
MATERIALS AND METHODS
Materials
Methionine enkephalin (MEn, YGGFM) and leucine enkeph-
alin (LEn, YGGFL) were from Bachem (Bubendorf,
Switzerland). Proctolin peptide (ProP, RYLPT) was from
Genescript (Piscataway, NJ). Sodium citrate tribasic dihy-
drate, ferric chloride hexahydrate (FeCl3 × 6 H2O), sodium
acetate anhydrous, 13C5-L-methionine (13C5-L-Met), 13C2-
citric acid (C2,C4), ethylenediaminetetraacetic acid (EDTA),
18O2 (stable isotope-labeled oxygen, 97%), sodium borohy-
dride (NaBH4), catalase from bovine liver, horseradish perox-
idase (HRP), sucrose, D-mannitol, D(+)-trehalose, L-histidine
(L-His) monochloride monohydrate, L-arginine (L-Arg) hy-
drochloride, and L-glycine (L-Gly) were purchased from
Sigma (Saint Louis, MO). Diethylene-triaminepentaacetic ac-
id (DTPA), L-lysine (L-Lys) hydrate, and L-methionine (L-
Met) were purchased from Aldrich (Milwaukee, WI). 2,2′-
Detection of H2O2
Solutions of 600 μL, containing water or 10 mM buffer,
pH 6.0, 1 or 50 μM Fe3+, and either 50 mM amino acids or

Pharm Res
120 mM sugars, were photo-irradiated with near UV lamps
(RPR-3500 Å, λmax = 350 nm, 9.7 Whm−2) in a Rayonet©
reactor. Aliquots (50 μL) were drawn before and after photo-
irradiation. The aliquots were placed in a 96 well plate and
were incubated in the dark for 10 min, after the addition of
5 μL of water or an aqueous solution containing
0.75 mg mL−1 of catalase. The H2O2 concentration was mea-
sured by the addition of 300 μL of an aqueous solution of
ABTS (final concentration 0.5 mM) and HRP (final concen-
tration 10 μg mL−1) (49). The absorbance of the solutions was
measured at λ = 415 nm in a plate reader (Infinite M200 Pro
Tecan; Mannedorf, Switzerland) maintained at 25°C. The
absorbance readings were compared with a calibration curve
of standard solutions of H2O2 (0–100 μM) prepared in the
respective buffers, for which the concentrations were separate-
B, C, and D, but a new peak with tR = 28.5 min was observed
(Fig. 1a, yellow trace). This product is referred to as product H
(see below).
The photo-irradiation of citrate solutions containing 1 μM
Fe3+ (Fig. 1b, black trace), led to the formation of products A,
B, C, and D. While the yield of product A was higher, the
yields of products B, C, and D were lower compared to those
obtained with 50 μM Fe3+ (Fig. 1b, black trace). At 1 μM
Fe3+, the addition of Gly showed no increase in product yields
over citrate alone (Fig. 1b, green trace). Arg and His signifi-
cantly decreased the yields of products B, C, and D, while the
yield of product A significantly increased (Fig. 1b, red and
blue traces). Interestingly, 1 μM Fe3+ solutions containing
Lys generated the highest yield of product A (Fig. 1b, light
blue trace), and significant yields of two new products eluting
with tR 12.9 min and 14.2 min, referred to as product I (see
below). These new products were also detected upon the ad-
dition of Gly and Arg but at lower yields. Met containing
solutions showed no oxidation peaks, except for a small yield
of product H (see insert in Fig. 1b).
ly measured by absorbance at λ = 240 nm (εH2O2
,
=
240nm
43.6 M−1 cm−1) (8).
RESULTS
In acetate buffer, containing 500 μM MEn and 50 μM
Fe3+, we observed low yields of product A as the only oxida-
tion product (Fig. 1c, black trace). The addition of Gly and
Arg resulted in a small increase in the yields of product A (Fig.
1c, green and red traces), while Lys and His caused a more
significant increase in the yield of product A, where His
revealed the highest yield (Fig. 1c, light blue, and blue
traces). The addition of Met abolished all peptide oxidation
products (Fig. 1c, yellow trace).
In acetate solutions containing 500 μM MEn and 1 μM
Fe3+ we detected low yields of product A (Fig. 1d, black trace).
Upon the addition of Gly, Arg, or His, the yields of product A
increased (Fig. 1d, green, red, and blue traces, respectively).
The addition of 50 mM Lys promoted the highest yields of
product A, as well as the formation of two additional peaks
with tR 12.9 min and 14.2 min, corresponding to product I
(Fig. 1 d, light blue trace). The addition of Met abolished all
peptide oxidation products (Fig. 1d, yellow trace).
Similarly to acetate solutions, the photo-irradiation of MEn
in water containing 50 μM Fe3+, did not produce significant
yields of oxidation products (Fig. 1e, black trace). However,
the addition of Gly or Arg promoted the formation of product
A (Fig. 1e, green and red traces), while Lys and His further
increased the yields of product A. The addition of Met abol-
ished all peptide oxidation products (Fig. 1e, yellow trace).
The photo-irradiation of MEn in water containing 1 μM
Fe3+ resulted in no oxidation products (Fig. 1f, black trace).
The addition of Gly or His generated slightly higher yields of
product A (Fig. 1f, green and blue traces). However, the ad-
dition of Arg and especially Lys resulted in significantly higher
yields of product A (Fig. 1f, light blue, and blue traces).
Photo-Irradiation with near UV Light
Photo-Irradiation of MEn in the Presence of Amino Acids
Solutions of 500 μM MEn in either water, 10 mM acetate, or
10 mM citrate, pH 6.0, containing 1 or 50 μM Fe3+ and
50 mM amino acids, were photo-irradiated with near UV light
(25.2 Whm−2) in a Rayonet© reactor. Representative HPLC
traces are displayed in Fig. 1. Control samples were prepared as
described above and placed in the dark for 2 h (Fig. S1).
The photo-irradiation of MEn in citrate buffer containing
50 μM Fe3+ generated all the products which we described
previously (23). Briefly, the first peak with a retention time (tR)
of 15.6 min corresponded to Met sulfoxide (product A) (Fig.
1a). The other products were three isomers of hydroxypheny-
lalanine (product B, peaks 2, 3 and 6; tR = 16.9, 18.2 and
20.9 min, respectively), a modified Tyr residue with a mass
increase of 28 Da, referred to as product C (peak 4, tR =
20.1 min), and 3,4-dihydroxyphenylalanine (product D, peak
5, tR = 20.5 min). All these oxidation products from MEn are
listed in Table I. The unmodified MEn eluted with tR =
22 min and the MS/MS spectrum is shown in Fig. S2.
When 50 mM Gly were added to a solution containing
500 μM MEn in 10 mM citrate, pH 6.0, and 50 μM Fe3+,
photo-irradiation produced similar yields of products B, C,
and D but significantly higher yields of product A (Fig. 1a,
green trace). The addition of Arg, Lys, and His reduced the
yields of products B, C, and D, and further increased the yields
of product A (Fig. 1a, red, light blue, and blue traces). The
addition of Met significantly lowered the yields of products A,

Pharm Res
Fig. 1 Chromatograms of MEn photo-irradiated in solutions containing amino acids. Solutions of 500 μM of MEn were prepared in either water,
10 mM citrate, or 10 mM acetate, pH 6.0, containing 50 (a, c, and e) or 1 (b, d, and f) μM Fe3+ and no (black trace) or 50 mM amino acids: Gly (green), Arg (red),
Lys (light blue), His (blue), or Met (yellow). Samples were photo-irradiated with near UV light (25.2 W h m−2). HPLC chromatograms were monitored by UV
detection at λ = 214 nm. Each trace represents the average of three independent experiments.
Moreover, the addition of Lys generated product I (Fig. 1d,
light blue). In the presence of Met no peptide oxidation prod-
ucts were detected (Fig. 1f, yellow trace).
Characterization of Product H and Effect of Experimental
Variables on the Formation of Product H
We photo-irradiated solutions containing 500 μM MEn in
10 mM citrate, pH 6.0, 50 mM Met and 50 μM Fe3+, with
near UV light (25.2 Whm−2). HPLC-MS/MS analysis of
product H revealed a molecular ion with m/z 660.29
(Fig. 2a), corresponding to a Δmass = 86 a.m.u. compared to
the molecular ion of unmodified MEn (m/z = 574.3). An
immonium ion of the N-terminal Tyr residue modified by
the addition of 38 a.m.u. (Y + 38) was detected, and the ad-
dition of 38 a.m.u. was further corroborated by the ions b2,
a2, b3, and b4. No ions were detected, which would indicate a
product formed by the addition of 86 a.m.u. to the N-terminal
Tyr residue. The Phe or Gly residues were not modified as
indicated by the presence of the immonium ion of unmodified
Phe, the b2, b3, and b4 ions, and the internal fragment GF. A
fragment with m/z 612.28 was detected (Fig. 2a), which could
be generated by the loss of 48 a.m.u. (consistent with the
cleavage of CH3SH) from the molecular ion (m/z = 660.29).
Such a loss of 48 a.m.u. was not detected for unmodified MEn
Tab le I Photoproducts of MEn
Product ID tR (min)
Sequence
m/z [M+H]+
A
B
C
D
E
15.6
YGGFM(+16)
590.22
590.22
602.22
590.22
630.27
688.33
674.31
16.9, 18.2, 20.9 YGGF(+16)M
20.1
20.5
22.5
23
Y(+28)GGFM
Y(+16)GGFM
Y(+56)GGFM
Y(+114)GGFM
Y(+100)GGFM
F
G
H
I
23
28.5
12.9, 14.2
Y(+38)GGFM(+48) a 660.30
Y(+16)GGFM(+16) 606.18
a The initial modification of Tyr may be by 86 a.m.u., decomposing during
mass spectrometry analysis leaving 38 a.m.u. at the Tyr residue and adding 48
a.m.u. to the C-terminal residue (see Discussion)

Pharm Res
([M+H]+-48)-H2O
594.28
c
y2-48
297.14
[M+H]+
([M+H]+-48)-H2O
594.28
[M+H]+
660.29
F
a
a4
435.23
y2-48
297.14
120.09
a4
435.23
660.29
F
a2
231.13
b3
1
0.5
0
[M+H]+-48
612.28
[M+H]+-48
612.28
120.09
316.15
b4
a
b3
GF
1
0.5
0
b4
231.13
463.22
316.15
205.11
GF
463.22
205.11
b2
259.12
Y(+38)
174.10
y3-48
b2
259.12
Y(+38)
174.10
354.17
y3-48
354.17
100
200
300
400
m/z
500
600
700
100
200
300
400
m/z
500
600
700
b
([M+H]+-49)-H2O
597.28
([M+H]+-49)-H2O
597.28
[M+H]+
664.31
d
y2-49
297.14
[M+H]+
664.31
GF
205.11
a4
438.24
y2-49
297.14
a4
438.24
a2
234.13
[M+H]+-49
615.29
[M+H]+-49
GF
1
0.5
0
b3
b3
615.29
b4
1
0.5
0
b4
205.11
319.16
319.16
a2
234.14
466.23
466.23
F
120.09
Y(+41)
b2
262.13
b2
262.13
177.11
y3-49
y3-49
354.17
Y(+41)
177.11
354.17
F
120.09
100
200
300
400
m/z
500
600
700
100
200
300
400
m/z
500
600
700
Fig. 2 MS/MS analysis of product H obtained in solutions containing combinations of 12C- and 13C-labeled citrate and Met. Solutions
contained (a) 12C-citrate and 12C-Met, (b) 13C2-citrate (C2,C4) and 13C5-Met, (c) 13C2-citrate (C2,C4) and 12C-Met, and (d) 12C-citrate and 13C5-Met. Samples
were photo-irradiated with near UV light (25.2 Whm−2) and analyzed by HPLC-MS/MS.
(Fig. S2), indicating that modifications leading to product H
were responsible for this fragmentation. Interestingly, small
yields of fragment ions y2–48 and y3–48 were detected, sug-
gesting an apparent addition of 48 a.m.u. to the C-terminal
Met residue of MEn.
To evaluate which molecule was the source of the modifi-
cations, we photo-irradiated solutions of 500 μM MEn, con-
taining 50 μM Fe3+ and a combination of either 10 mM 13C2-
citrate (C2, C4) and 50 mM 12C-Met or 10 mM 12C-citrate
and 50 mM 13C5-Met. We observed that photo-irradiation of
solutions in 13C2- citrate and 12C Met resulted in a molecular
ion with an m/z = 660.29 a.m.u, identical to the molecular
ion obtained after of photo-irradiation in the presence of 12C-
citrate and 12C-Met (Fig. 2c). When we changed the combi-
nation to 12C2-citrate and 13C5-Met we observed the molec-
ular ion with m/z 664.31, revealing an increase by Δmass = 4
a.m.u. compared to 12C-citrate and 12C-Met (Fig. 2d). Once
again, the ions containing the modified Tyr showed an in-
crease of 3 a.m.u. while the fragment ions y3–49 and y2–49
are consistent with an increase of 1 a.m.u. at the C-terminus.
In a previous study (23), we demonstrated that ●OH radi-
cals, or their metal-bound equivalents/ hypervalent iron-oxo
species (50), were produced during the photo-Fenton reaction
and that these were responsible for the formation of oxidation
products, including products B and D (23). To evaluate if
these oxidants were involved in the formation of product H,
we photo-irradiated solutions of 500 μM MEn, in 10 mM
citrate, pH 6.0, 50 mM Met, and 50 μM Fe3+ containing
To evaluate whether the increase of 86 a.m.u. on MEn
originated from the photo-degradation of citrate or Met, we
employed 13C2-citrate (C2, C4) and 13C5-Met. Solutions were
photo-irradiated with near UV light (25.2 Whm−2) and ana-
lyzed by HPLC-MS/MS (Fig. 2b). We observed a molecular
ion with m/z 664.31, indicating a mass increase of Δmass = 4
a.m.u. as compared to experiments with 12C-citrate and 12C-
Met. The immonium ion of the modified N-terminal Tyr
residue was detected with an m/z = 177.11 a.m.u, (Fig. 2b),
which corresponded to an increase of 3 a.m.u. compared to
the same immonium ion generated in experiments with 12C-
citrate and 12C-Met. The ions b2, b3, and b4 ions showed an
increase of 3 a.m.u. each, indicating that three carbons were
incorporated into the modified Tyr residue. Further, we
detected a fragment ion [M + H]+-49, as compared to
[M + H]+-48 in experiments with 12C-citrate and 12C-Met,
as well as the fragment ions y3–49 and y2–49, consistent with
an increase of 1 a.m.u.

Pharm Res
●
0.5, 1.0 and 2.0 M methanol, a scavenger of OH radicals.
The presence of 0.5 M methanol (Fig. S3, red trace), reduced
the yield of product H by 50% compared to the solution
without added methanol (Fig. S3, yellow trace). A further
increase in the concentration of methanol to 1.0 or 2.0 M
(Fig. S3, blue and black traces, respectively), decreased the
yield of product H to 33% of the original yield (Fig. S3),
suggesting that ●OH radicals are necessary for the formation
of product H.
ion of the modified Tyr residue with m/z 174.09 a.m.u. and
and the fragment ions a4 and b4, and fragments [M + H]+-48
and ([M + H]+-48)-H2O (Fig. S7a). However, in contrast to
MEn we did not detect fragments y2–48 and y3–48 for LEn.
No ion was observed indicating the addition of 86 a.m.u. to the
Tyr residue. The photo-irradiation of LEn in the presence of
12C-citrate and 13C5-Met revealed the same pattern as for MEn
(Figs. S7a and b).
The formation of product H showed a slight dependence
on citrate concentrations between 1.0 and 75 mM, with a
maximal yield around 50 mM citrate (Fig. S4a). No significant
differences in the yields of product H were observed for Met
concentrations between 25 mM and 100 mM, but slightly
lower yields for 10 mM Met (Fig. S4b).
Photo-Irradiation of ProP in the Presence of Met
An important question was whether photo-irradiation would
lead to the modification of a Tyr residue which is neither
located N- or C-terminal. For this, we prepared solutions of
500 μM ProP in 10 mM citrate buffer, pH 6.0, containing
50 mM Met and 50 μM Fe3+. Solutions were photo-
irradiated with near UV light (25.2 Whm−2) and analyzed
by HPLC-MS/MS. Photo-irradiation resulted in the forma-
tion of a new product (H*) with a molecular ion with m/z
735.39, consistent with the addition of 86 a.m.u. to ProP
(Fig. S8a). The ions b2, a2, and b3, as well as the internal
fragment Y(+38)L-CO (m/z 335.19) indicated that the addi-
tion of 86 a.m.u. occurred on the Tyr residue.
To evaluate whether atmospheric O2 was involved in the
formation of product H, solutions containing 500 μM MEn,
in 10 mM citrate, pH 6.0, 50 mM Met and 50 μM Fe3+, were
first saturated with Ar for 30 min in the dark, and subsequent-
ly saturated with either 18O2 or 16O2 for 10 min. After photo-
irradiation with near UV light (25.2 Whm−2), HPLC-MS/
MS analysis revealed the formation of the ion with m/z =
660.29 a.m.u. in solutions saturated with either 16O2 or
18O2 (Fig. S5a and b, respectively), indicating that atmospher-
ic O2 was not incorporated into product H. Moreover, in both
solutions the immonium ions of modified Tyr showed an in-
crease of 38 a.m.u., and identical fragment ions [M + H]+-48,
([M + H]+-48)-H2O, y3–48 and y2–48 were observed. .
To evaluate whether the amino group was converted into a
Schiff base, photo-irradiated solutions containing 500 μM
MEn, in 10 mM citrate, pH 6.0, 50 mM Met and 50 μM
Fe3+ were subjected to reduction with 10 mM NaBH4 for
20 min at room temperature in the dark. We observed no
changes in the MS/MS spectrum (Fig. S6) compared to the
non-reduced samples (Fig. 2a), indicating that the modified N-
terminal Tyr residue did not contain a Schiff base.
Experiments with 13C5-Met resulted in the addition of 90
a.m.u. to the Tyr residue (Fig. S8b), consistent with the incor-
poration of four carbon atoms from Met, corroborated by the
ions b2, b3, b4, y2, and y4.
Importantly, the molecular ions of modified ProP showed a
neutral loss of 48 or 49, respectively, depending on whether
photo-irradiation had been carried out in the presence of 12C-
Met or 13C5-Met. However, no fragment was detected con-
taining Tyr(+38) or Tyr(+41), as was the case for MEn and
LEn. The differences between ProP on one hand and MEn
and LEn on the other hand are likely caused by the location of
Tyr, within the sequence vs. N-terminal, but potentially also
by the Pro residue in ProP. ProP contains the Pro residue in
the fourth position, restricting the flexibility of the backbone,
as was previously demonstrated in peptides containing Pro in
aqueous environments (51).
We attempted to fractionate and enrich product H for
analysis by NMR. However, MS analysis of the collected frac-
tion revealed that product H decomposed during purification.
Photo-Irradiation of LEn in the Presence of Met
Photo-Irradiation of MEn in the Presence of Carbohydrates
To evaluate if the C-terminal Met residue in MEn plays a role
in product formation, we tested the formation of product H′ in
LEn (here, product H′ differs from product H only with respect
to the substitution of Met by Leu). Solutions of 500 μM LEn in
10 mM citrate, pH 6.0, containing 50 mM Met and 50 μM
Fe3+ were photo-irradiated with near UV light (25.2 Whm−2)
and analyzed by HPLC-MS/MS. As for MEn, we observed a
product characterized by a mass increase of Δmass = 86 a.m.u.
compared to the unmodified LEn (Fig. S7). The MS/MS data
showed some similarity to the pattern of MEn, i.e. the addition
of 38 a.m.u. to the Tyr residue, corroborated by the immonium
We photo-irradiated solutions of 500 μM MEn, in 10 mM
citrate, pH 6.0, containing 120 mM mannitol, sucrose or tre-
halose, and 50 μM Fe3+ with near UV light (25.2 Whm−2).
The oxidation products were analyzed by HPLC. The addi-
tion of mannitol increased the yields of product A, but signif-
icantly decreased the yields of products B, C, and D, com-
pared to solutions without mannitol (Fig. 3a, green trace).
The addition of sucrose led to higher yields of product A, as
well as products E, F, and G (Fig. 3a, red trace). Product E,
with tR = 22.5 min, was characterized by the addition of 56
a.m.u. to Tyr. Products F and G coeluted at tR = 23 min.

Pharm Res
b 1.5
1.5
1
Citrate
Citrate
a
1
E
A
C D
A
B
F G
0.5
0.5
0
+ trehalose
+ sucrose
+ trehalose
+ sucrose
+ mannitol
+ mannitol
Citrate
citrate
0
0 2
0 2
0 2
15
20
20
20
25
25
25
30
30
30
0 2 10
15
20
25
30
d1.5
1.5
Acetate
Acetate
c
1
1
A
A
0.5
0.5
+ trehalose
+ sucrose
+ trehlose
+ sucrose
+ mannitol
Acetate
+ mannitol
Acetate
0
0
15
0 2
15
20
25
30
1.5
1.5
Water
Water
e
f
1
0.5
0
1
0.5
0
A
A
+ trehalose
+ sucrose
+ mannitol
+ trehlose
+ sucrose
+ mannitol
Water
Water
15
0 2
15
20
25
30
Elution time (min)
Elution time (min)
Fig. 3 Chromatograms of MEn photo-irradiated in solutions containing sugars. Solutions of 500 μM MEn were prepared in either water or 10 mM
citrate or 10 mM acetate, pH 6.0, containing 50 (a, c, and e) or 1 (b, d, and f) μM Fe3+ without (black trace) or with added 120 mM mannitol (green), sucrose
(red), or trehalose (blue). Samples were photoirradiated with near UV light (25.2 W h m-2). HPLC chromatograms were monitored by UV detection at λ =
214 nm. Each trace represents the average of three independent experiments.
Product F was characterized by the addition of 114 a.m.u.
addition to the Tyr residue, while product G was character-
ized by the addition of 100 a.m.u. to the Tyr residue. Products
G and F also form in the absence of sucrose when the concen-
tration ratio of citrate:iron is >500 (23).
without carbohydrates (Fig. 3b, black trace). However, prod-
ucts B, C, and D were not observed.
In acetate buffer, containing 50 μM Fe3+, mannitol, su-
crose, and trehalose promoted a slight increase in the yields
of product A, compared to a solution without carbohydrates
(Fig. 3c). No other oxidation products were detected. When
the Fe3+ concentration was lowered to 1 μM Fe3+, only su-
crose promoted an increase in the yield of product A (Fig. 3d,
red trace).
The addition of trehalose caused similar yields of prod-
uct A as sucrose but increased the yields of products E, F,
and G (Fig. 3, blue trace). Both sucrose- and trehalose-
containing solutions decreased the yields of products B, C,
and D with respect to the solution without sugars. As
products B, C and D are generated via hydroxylation of
The photo-irradiation of solutions containing 500 μM MEn,
120 mM carbohydrates, 50 μM Fe3+, and no buffer, demon-
strated the formation of small yields of product A (Fig. 3e,
green, red, and blue traces), while at 1 μM Fe3+, product A
was only generated in the presence of sucrose (Fig. 3f, red trace).
•
MEn by OH radicals (23), reduced yields of these prod-
ucts in the presence of large concentrations of carbohy-
drates are expected due to the competitive reaction of
•OH radicals with the carbohydrates.
To evaluate the formation of oxidation products at lower
concentrations of Fe3+, we prepared solutions of 500 μM
MEn in 10 mM citrate, pH 6.0, containing 1 μM Fe3+. We
observed a significant increase in the yields of product A in the
presence of mannitol, sucrose, or trehalose (Fig. 3b, green,
red, and blue traces, respectively), compared to citrate buffer
Formation of H2O2
We evaluated the photo-induced formation of H2O2 in water,
10 mM acetate or 10 mM citrate, pH 6.0, containing 1 or
50 μM Fe3+, and 50 mM amino acids or 120 mM sugars.

Pharm Res
Solutions of 10 mM citrate buffer, pH 6.0, containing
transient levels of photo-generated Fe2+, which react
with H2O2 (23).
Importantly, especially in citrate buffer containing 1 μM
Fe3+, none of the additional excipients increased the levels of
H2O2 (Fig. 4a), suggesting that H2O2 formation is not key to
the formation of increased yields of product A in the presence
of these excipients.
1 μM Fe3+ produced 98 μM H2O2 after photo-irradiation
with a dose of 9.7 Whm−2 of near UV light (Fig. 4a, white
column). A quantitative comparison of H2O2 yields in citrate
buffer containing amino acids and sugars shows that, except
for mannitol, all other excipients lowered the detectable yields
of H2O2 (Fig. 4a) The specificity of the assay for H2O2 was
evaluated by the addition of catalase (Fig. S9).
In 10 mM acetate or unbuffered water containing 1 μM
Fe3+, only low levels of H2O2 were detected (Figs. 4b and c).
Notably the addition of Lys or His to water promoted the
formation of small yields of H2O2 (Fig. 4c). When His was
replaced by imidazole in water, slightly higher yields of
H2O2 were generated (Fig. S10).
Formation of Oxidation Products by Peroxyl Radicals
Since the excipients did not promote an increase of H2O2 in
citrate buffer, we decided to evaluate whether other oxidants
could be responsible for the general increase in the formation
of product A. A viable candidate would be the peroxyl radical,
which forms by oxygen addition to carbon-centered radicals.
For model studies, peroxyl radicals can be generated through
the thermal decomposition of azo-initiators such as AAPH
(52, 53). Therefore, 500 μM MEn in 10 mM citrate,
pH 6.0, containing 50 μM Fe3+, and different concentrations
of AAPH (0 to 100 mM) were incubated at 37°C for 2 h and
When the concentration of Fe3+was increased to
50 μM, no significant levels of H2O2 were detected after
photo-irradiation with 9.7 Whm−2 (Fig. S11). This obser-
vation is likely the result of reactions 1 and 3 (Scheme 1),
where Fe3+ is involved in the formation and decomposi-
tion of H2O2. Higher concentrations of Fe3+ increase the
Fig. 4 Detection of H2O2 in
solutions containing amino
acids or sugars. Solutions of
water, 10 mM citrate, or 10 mM
acetate, pH 6.0, containing 1 μM
Fe3+ were photo-irradiated with
near UV light (9.7 Whm-2).
a
100
Citrate
80
60
40
20
6
Aliquots of 50 μL were taken after
photo-irradiation, placed in a 96
well plate, and incubated with 5 μL
of water or an aqueous solution
containing 0.75 mg mL-1 catalase.
After 10 min of incubation in the
dark, 300 μL ABTS and HRP solu-
tion were added to each well and
the absorbance was measured at
λ = 415 nm in in plate reader
(Infinite M200 Pro; Tecan,
Mannedorf, Switzerland): (a) citrate
buffer, (b) acetate buffer, and (c)
water. Each column represents the
average of three independent
experiments.
4
2
0
Citrate
+ Gly
+ Met
+ Arg
+ Lys
+ His
+ Mannitol + Sucrose + Trehalose
b
c
100
80
60
40
20
6
4
2
0
Acetate
Acetate
+ Gly
+ Met
+ Arg
+ Lys
+ His
+ Mannitol + Sucrose + Trehalose
100
80
60
40
20
Water
6
4
2
0
water
+ Gly
+ Met
+ Arg
+ Lys
+ His
+ Mannitol + Sucrose + Trehalose

Pharm Res
1.2
0.8
the oxidation products were analyzed by HPLC with UV
detection at λ = 214 nm (Fig. S12). Peroxyl radicals generated
by the thermal decomposition of AAPH generated mainly
product A, consistent with earlier studies on peroxyl radical
reactions with organic sulfides (54), with yields depending on
the initial AAPH concentration. The mechanism for the for-
mation of peroxyl radicals from AAPH is shown in Fig. S12.
Hence, peroxyl radical generation from excipients via the
photo-Fenton reaction may play an important role in the for-
mation of increased yields of product A.
H
1.5
1
Citrate
A
28
29
30
Elution time (min)
+ Met
+ His
+ Lys
+ Arg
+ Gly
0.5
0
Citrate
Characterization of Product I
0 210
15
20
25
30
35
1.5
We observed the formation of two peaks with tR = 12.9 min
and 14.2 min, corresponding to product I (Table S1), during
the photo-irradiation of solutions of 10 mM citrate ( 50 mM
Gly, Arg, or Lys) (Fig. 1b) or acetate (+ 50 mM Lys) (Fig. 1d),
pH 6.0, containing 500 μM MEn, and 1 μM Fe3+. Product I
was also generated in water containing 1 μM Fe3+ and 50 mM
Lys (Fig. 1f).
The HPLC-MS/MS analysis detected molecular ions
with m/z = 606.18 a.m.u. for both peaks with tR =
12.9 min and 14.2 min (Figs. S13a and b, respectively),
corresponding to the addition of 32 a.m.u. to MEn. The
fragment ions b2, b3, b4, y1, y2, and y3 were consistent
with the addition of 16 a.m.u. to each Tyr and Met, i.e.
the formation of 3,4-dihydroxyphenylalanine and Met
sulfoxide. The observation of two peaks with tR = 12.9
and 14.2 min can be rationalized by the formation of
Met sulfoxide diastereomers.
Acetate
A
1
0.5
0
+ Met
+ His
+ Lys
+ Arg
+ Gly
Acetate
0 2
15
20
25
30
35
1.5
1
Water
A
+ Met
+ His
+ Lys
+ Arg
+ Gly
0.5
0
Photo-Irradiation with Visible Light
Water
0 2
15
20
25
30
35
Photo-Irradiation of MEn in the Presence of Amino Acids
and Sugars
Elution time (min)
Fig. 5 Visible light photo-irradiation of MEn in the presence of
amino acids. Solutions of 500 μM MEn in either water, 10 mM citrate, or
10 mM acetate, pH 6.0, containing 1 μM Fe3+ without (black trace) or with
50 mM Gly (green), Arg (red), Lys (light blue), His (blue), or Met (yellow)
were photo-irradiated with visible light (34.6 Whm−2), and analyzed by
HPLC with UV detection at λ = 214 nm: (a) citrate, (b) acetate, and (c) water.
Each trace represents the average of three independent experiments.
Solutions of 500 μM MEn in 10 mM buffer, pH 6.0,
containing 50 mM amino acids, and 1 μM Fe3+, were
photo-irradiated with 34.6 Whm−2 visible light and ana-
lyzed by HPLC.
Similar yields of product A were generated in citrate buffer
alone and in citrate buffer containing 50 mM Gly or His.
However, the addition of 50 mM Arg or Lys promoted the
generation of higher yields of product A (Fig. 5a, red and light-
blue traces, respectively). The photo-irradiation of solutions
containing Met generated low yields of product H, as shown
in the inset of Fig. 5a, yellow trace.
Also in acetate solutions we did not observe the formation
of oxidation products in the absence of amino acids. However,
the addition of Gly or Arg generated low yields of product A
(Fig. 5b, green and red traces), while the addition of Lys sig-
nificantly increased the yields of product A (Fig. 5b, light blue
trace). No oxidation products were detected in the presence of
Met. Similar results were observed in water without added
buffer (Fig. 5c), suggesting that the excipients are responsible
for the generation of product A.
In addition, we monitored the generation of oxidation
products in solutions containing sugars that were exposed to
visible light (34.6 Whm−2). Solutions of 500 μM MEn in
10 mM citrate, pH 6.0, containing 120 mM mannitol, su-
crose, or trehalose displayed an increase in the yields of prod-
uct A for all sugars with higher yields for sucrose and trehalose
(Fig. 6a).

Pharm Res
containing 1 μM Fe3+ and either mannitol, sucrose, or treha-
lose (Figs. 6b and c).
Combination of Excipients with EDTA or DTPA
Solutions containing 500 μM MEn, 10 mM citrate, pH 6.0,
120 mM sucrose, 50 mM amino acids, 50 μM Fe3+, and
250 μM EDTA or DTPA, were photo-irradiated with near
UV light (25.2 Whm−2). After photo-irradiation, samples
were analyzed by HPLC with UV detection (λ = 214 nm).
When the added amino acid was Gly, we detected the forma-
tion of products A and E (Fig. 7, green traces), where the yield
of product A was slightly higher in the presence of EDTA
compared to DTPA. Similar yields of products A and E were
also detected when the added amino acid was Lys (Fig. 6,
orange traces). However, the addition of Met reduced the
formation of product A but promoted the formation of prod-
uct H. Moreover, products E, F, and G were detected (Fig. 7,
grey and black traces).
DISCUSSION
The present work shows that commonly used excipients can
promote or inhibit photo-degradation pathways in pharma-
ceutical buffers. Photo-degradation reactions were elucidated
with model peptides as oxidation targets such as MEn, LEn,
and ProP. In general, citrate promoted the oxidative degra-
dation of these peptides via the photo-Fenton oxidation. The
addition of selected amino acids (except Met) or carbohy-
drates (in citrate buffer) enhanced the oxidation yields specif-
ically of Met sulfoxide (product A). The addition of Met large-
ly prevented the photochemical formation of Met sulfoxide,
but, in citrate buffer, resulted in the generation of a Tyr mod-
ification (product H). Mechanistic features of the photo-
Fenton reaction in the presence of amino acids and carbohy-
drates will be discussed in the following.
Fig. 6 Visible light photo-irradiation of MEn solutions in the pres-
ence of sugars. Solutions of 500 μM MEn were prepared in water or
10 mM citrate, or acetate pH 6.0, containing 1 μM Fe3+ without (black trace)
or with added 120 mM sugars mannitol (green), sucrose (red), or trehalose
(blue). Samples were photo-irradiated with visible light (34.6 Whm−2), and
analyzed by HPLC with UV detection at λ = 214 nm. (a) citrate, (b) acetate,
and (c) water. Each treace represents the average of three independent
experiments.
In contrast to citrate buffer, no significant yields of oxida-
tion products were detected in either 10 mM acetate or water
Fig. 7 MEn photo-irradiated
in solutions containing a com-
bination of excipients. Solutions
of 500 μM of MEn in 10 mM citrate,
1.5
1
E
F
A
pH 6.0, containing 1 μM Fe3+
,
H
G
120 mM sucrose, 50 mM Gly, Lys
or Met, and 250 μM EDTA or
DTPA, were photo-irradiated with
near UV light (25.2 Whm−2) and
analyzed by HPLC with UV detec-
tion at λ = 214 nm.
+ Met + EDTA
+ Met + DTPA
+ Lys + EDTA
+ Lys + DTPA
+ Gly + EDTA
+ Gly + DTPA
0.5
0
0 2
15
20
25
30
Elution time (min)

Pharm Res
peroxyl radicals cause the increased yields of product A.
Peroxyl radicals can be generated from the amino acids used
●
in this study via (i) reaction with OH radicals (55) or (ii) via
direct photo-induced ligand-to-metal charge transfer (LMCT)
(56) within amino acid-iron complexes, followed by decarboxyl-
ation and the addition of oxygen to an intermediary α-amino
radical (57). In citrate buffer, hydroxyl radicals are generated via
reactions 1–3 (Scheme 1) (23). Specifically, Arg, Lys, and His
●
provide multiple sites, for OH radical attack (58), where the
ensuing carbon-centered radicals will subsequently add O2 to
•
yield peroxyl radicals. However, OH radical attack on amino
acids cannot be the only mechanism to yield peroxyl radicals.
This is evident from photo-Fenton oxidation reactions in acetate
buffer or water, reaction media which do not generate products
typical for •OH radical reactions (23). Here, direct LMCT within
amino acid-Fe3+ complexes (57) will lead to peroxyl radicals
according to reactions 8–10, displayed in Scheme 2.
Importantly, amino acids bind Fe3+ with high formation con-
stants (log K > 8.0) at acidic pH values (pH ≤ 3.35) (59). These
complexes represent chelates involving both the amino and the
carboxylate groups (59), where binding to the metal promotes
deprotonation of the amino group even at acidic pH (59, 60).
In general, the addition of Met prevented the genera-
tion of oxidation products from MEn, LEn and ProP.
This is due to the efficient reaction of thioethers with
reactive oxygen species such as hydrogen peroxide (61),
peroxyl radicals (54) and hydroxyl radicals (62) which ren-
der Met a suitable antioxidant. In acetate buffer and wa-
ter, no reactions of Met-derived products with our target
peptides were observed. However, in citrate buffer an
oxidation product of Met reacted with peptide Tyr resi-
dues, as discussed in more detail below.
Scheme 2 Photo-induced peroxyl radical formation via LMCT in amino
acid-Fe3+ complexes
Effect of Amino Acids
In citrate buffer, none of the added amino acids increased H2O2
formation (Fig. 4); however, except for Met, these excipients
caused an increase of the yields of Met sulfoxide (product A).
These data suggest that alternative oxidants are responsible for
product A formation. Based on our studies with model peroxyl
radicals from AAPH, we propose that amino acid-derived
Scheme 3 Reaction of trehalose
peroxyl radicals with Met

Pharm Res
Scheme 4 Condensation of
methional to Tyr
Effect of Sugars
Specific Reactions of the Antioxidant Met in Citrate
Buffer
The presence of sugars enhanced the formation of product A
in citrate buffer, suggesting peroxyl radical formation via the
reaction of hydroxyl radicals generated from citrate.
Scheme 3 shows the reaction of a representative peroxyl rad-
ical from trehalose with Met. An important feature of the
reaction of peroxyl radicals with organic sulfides is the gener-
ation of alkoxyl radicals (54), which are powerful oxidants and
can carry out additional oxidation reactions.
Studies by Morelli et al demonstrated that the reactivity of
the different sugars towards oxidation by ●OH radicals
depended on the number of hydroxyl groups present in the
sugar molecule, rendering disaccharides more reactive than
monosaccharides (63). However, at the employed carbohy-
drate concentrations we expect that •OH radicals react quan-
titatively with the carbohydrates. The enhanced yields of
product A in the presence of sucrose and trehalose vs. manni-
tol (Fig. 3) are likely caused by different reactivities of the
individual peroxyl radicals and/or secondary reactions of
product radicals.
Of specific interest is the modification of peptide Tyr residues
when Met was present in citrate buffer. The photo-Fenton
reaction generates various oxidants including H2O2 and
●OH radicals (Scheme 1). Obviously, H2O2 can react with
free Met to generate Met sulfoxide, a product which we
detected by HPLC-MS/MS analysis (data not shown). The
●OH radical represents a one-electron oxidant, which will
initially add to the Met sulfur atom and trigger decarboxyl-
ation via an intermediary sulfur-oxygen three-electron bond-
ed radical cation (62). The resulting α-aminoalkyl radical will
react with oxygen and convert to 3-methylthiopropanal
(methional) (64). Methional is also generated via riboflavin-
photosensitized oxidation of Met (65). Methional can add to
Tyr via phenol-aldehyde condensation (66, 67) (Scheme 4;
reactions 12–15), which would lead to the net addition of 86
a.m.u. to the Tyr residue, as observed for ProP. It is possible
that the phenolic hydroxyl group subsequently adds to the
double bond, yielding a bicyclic product (reaction 15).

Pharm Res
Methional contains four carbon atoms from the original Met
residue (one carbon atom is lost during decarboxylation of the
original Met). Hence, when 12C-Met was replaced by 13C5-
Met during the photo-Fenton reaction, the phenol-aldehyde
condensation with ProP caused a net addition of 90 a.m.u.
Interestingly, we did not detect any evidence for the addi-
tion of 86 (or 90) a.m.u. to the N-terminal Tyr residues in
MEn and LEn. However, for the addition products from all
peptides HPLC-MS/MS analysis detected a neutral loss of 48
(after photo-Fenton reaction in the presence of 12C-Met) or 49
(13C5-Met) a.m.u., consistent with the elimination of CH3SH.
Specifically, for MEn and LEn we detected the products of
such neutral losses, i.e. peptides containing Tyr(+38 a.m.u.).
These data suggest that the aldehyde-phenol condensation
products of especially N-terminal Tyr residues are unstable
during positive electrospray ionization, and eliminate
CH3SH, potentially via reaction with the N-terminal amino
groups. Interestingly, the fragment ions y2–48 and y3–48 of
product H, detected when MEn was photo-irradiated in the
presence of 12C-Met, suggest that during MS analysis CH3SH
may somehow associate with the C-terminus of MEn. One
could speculate about a gas phase addition to the C-terminal
carboxylic acid of Met in MEn, generating a 2-(alkylamino)-
1,4-bis(methylthio)butane-1,1-diol derivative, but such struc-
ture would certainly not be stable in aqueous solution.
Because of the fragments y2–48 and y3–48 (or y2–49 and
y3–49 when product H was generated from MEn in the pres-
ence of 13C5-Met) the sequences of product H from MEn in
Fig. 2 are given as Y(+38)GGFM(+48) and Y(+41)GGFM(+
49), respectively. We have kept a similar notation for product
H′ from LEn (Fig. S7), since we detected fragment ions con-
taining Y(+38) and Y(+41), respectively.
even EDTA or DTPA chelates of iron are not necessarily
redox-inert and can react with various substrates, including
hydrogen peroxide (71–76). Moreover, in iron chelates of
EDTA and DTPA the amino and carboxylate groups are
amenable to LMCT analogous to the reactions shown for
citrate in Scheme 1 and for amino acids in Scheme 2.
Earlier we had shown that EDTA and DTPA were able to
promote the photo-Fenton oxidation of MEn in acetate buffer
(23). Hence, under conditions of light exposure EDTA and
DTPA may enhance rather than prevent the oxidation of
formulation constituents.
CONCLUSIONS
In citrate buffer, photo-Fenton processes lead to iron-
dependent oxidative modifications of target peptides during
the exposure to near UV and visible light. Here, the addition
of selected excipients such as Arg, Lys, mannitol, sucrose or
trehalose changes the product patterns, generating higher
yields of Met sulfoxide. In acetate buffer or water, specifically
the addition of Lys or His promote Met oxidation depending
on the concentration of ferric iron. Not surprisingly, the addi-
tion of free Met protects the target peptides against photo-
Fenton oxidation. However, in citrate buffer an oxidation
product of free Met, methional, can modify peptide Tyr res-
idues via a condensation reaction.
SUPPLEMENTARY INFORMATION
The online version contains supplementary material available
at https://doi.org/10.1007/s11095-021-03042-8.
We note that during the photo-sensitized oxidation of Met
by riboflavin, methional was observed to undergo further deg-
radation to CH3SH and propenal (68). Technically, propenal
could add to the Tyr residue, but a subsequent elimination of
water to generate an addition product Tyr(+38) would be
difficult. Moreover, no propenal adduct of ProP was detected.
REFERENCES
1. Kerwin BA, Remmele RL Jr. Protect from light: photodegradation
and protein biologics. J Pharm Sci. 2007;96(6):1468–79.
2. Hawe A, Wiggenhorn M, van de Weert M, Garbe JH, Mahler H-
C, Jiskoot W. Forced degradation of therapeutic proteins. J Pharm
Sci. 2012;101(3):895–913.
Robustness of the Formulation
The literature offers some criteria regarding the acceptable
yields of aggregation, oxidation, and charge variants present
in a protein formulation (69). The excipients used should sta-
bilize or ensure minimal modifications in pharmaceutics pro-
duced during manufacturing and storage, which is considered
formulation robustness (70). Under our experimental condi-
tions, an increase in the number of excipients (i.e., by addition
of amino acids or carbohydrates, and EDTA or DTPA) used
for protein formulation was not necessarily helpful to the sta-
bility of the model peptides, as demonstrated in Fig. 7.
Especially the addition of EDTA or DTPA to protein formu-
lations aims at reducing metal-catalyzed oxidation. However,
3. Schöneich C. Photo-degradation of therapeutic proteins: mecha-
nistic aspects. Pharm Res. 2020;37(3):45.
4. Luis LM, Hu Y, Zamiri C, Sreedhara A. Determination of the
acceptable ambient light exposure during drug product manufac-
turing for long-term stability of monoclonal antibodies. PDA J
Pharm Sci Technol. 2018;72(4):393–403.
5. Sreedhara A, Yin J, Joyce M, Lau K, Wecksler AT, Deperalta G,
et al. Effect of ambient light on IgG1 monoclonal antibodies during
drug product processing and development. Eur J Pharm Biopharm.
2016;100:38–46.
6. Walrant P, Santus R, Grossweiner L. Photosensitizing properties of
N-formylkynurenine. Photochem Photobiol. 1975;22(1–2):63–5.
7. Parker NR, Korlimbinis A, Jamie JF, Davies MJ, Truscott RJ.
Reversible binding of kynurenine to lens proteins: potential

Pharm Res
protection by glutathione in young lenses. Invest Ophthalmol Vis
adenoviral vector through spray drying. Int J Pharm. 2016;506(1–
2):289–301.
Sci. 2007;48(8):3705–13.
8. Razygraev A, Arutiunian A. Determination of human serum glu-
tathione peroxidase activity, by using hydrogen peroxide and 5, 5′-
dithio-bis (2-nitrobenzoic acid). Klin Lab Diagn. 2006;6:13–6.
9. Ávila F, Ravello N, Zanocco AL, Gamon LF, Davies MJ, Silva E. 3-
Hydroxykynurenine bound to eye lens proteins induces oxidative
modifications in crystalline proteins through a type I photosensitiz-
ing mechanism. Free Radic Biol Med. 2019;141:103–14.
10. Rathore A, Bhambure R, Ghare V. Process analytical technology
(PAT) for biopharmaceutical products. Anal Bioanal Chem.
2010;398(1):137–54.
11. Shah DD, Zhang J, Maity H, Mallela KMG. Effect of photo-
degradation on the structure, stability, aggregation, and function
of an IgG1 monoclonal antibody. Int J Pharm. 2018;547(1–2):438–
49.
12. Patro SY, Freund E, Chang BS. Protein formulation and fill-finish
operations. 2002.
28. Schöneich C. Redox processes of methionine relevant to β-amyloid
oxidation and Alzheimer's disease. Arch Biochem Biophys.
2002;397(2):370–6.
29. Brange J, Havelund S, Hougaard P. Chemical stability of insulin. 2.
Formation of higher molecular weight transformation products
during storage of pharmaceutical preparations. Pharm Res.
1992;9(6):727–34.
30. Platts L, Falconer RJ. Controlling protein stability: mechanisms
revealed using formulations of arginine, glycine and guanidinium
HCl with three globular proteins. Int J Pharm. 2015;486(1–2):131–
5.
31. Platts L, Darby SJ, Falconer RJ. Control of globular protein ther-
mal stability in aqueous formulations by the positively charged ami-
no acid excipients. J Pharm Sci. 2016;105(12):3532–6.
32. Jorgensen L, Hostrup S, Moeller EH, Grohganz H. Recent trends
in stabilising peptides and proteins in pharmaceutical formulation–
considerations in the choice of excipients. Expert Opinion on Drug
Delivery. 2009;6(11):1219–30.
13. Cromwell ME, Hilario E, Jacobson F. Protein aggregation and
bioprocessing. AAPS J. 2006;8(3):E572–9.
14. Bessa J, Boeckle S, Beck H, Buckel T, Schlicht S, Ebeling M, et al.
The immunogenicity of antibody aggregates in a novel transgenic
mouse model. Pharm Res. 2015;32(7):2344–59.
15. Freedman MS, Pachner AR. Neutralizing antibodies to biological
therapies: a “touch of gray” vs a “black and white” story. In: AAN
Enterprises. 2007.
16. Calabresi P, Giovannoni G, Confavreux C, Galetta S, Havrdova E,
Hutchinson M, et al. The incidence and significance of anti-
natalizumab antibodies: results from AFFIRM and SENTINEL.
Neurology. 2007;69(14):1391–403.
17. Br B, Bessa J, Folzer E, Ríos Quiroz A, Schmidt R, Bulau P, et al.
Extensive chemical modifications in the primary protein structure
of IgG1 subvisible particles are necessary for breaking immune
tolerance. Mol Pharm. 2017;14(4):1292–9.
18. Doessegger L, Mahler H-C, Szczesny P, Rockstroh H, Kallmeyer
G, Langenkamp A, et al. The potential clinical relevance of visible
particles in parenteral drugs. J Pharm Sci. 2012;101(8):2635–44.
19. Carpenter JF, Randolph TW, Jiskoot W, Crommelin DJ,
Middaugh CR, Winter G, et al. Overlooking subvisible particles
in therapeutic protein products: gaps that may compromise prod-
uct quality. J Pharm Sci. 2009;98(4):1201–5.
20. Du C, Barnett G, Borwankar A, Lewandowski A, Singh N, Ghose
S, et al. Protection of therapeutic antibodies from visible light in-
duced degradation: use safe light in manufacturing and storage. Eur
J Pharm Biopharm. 2018;127:37–43.
21. Tønnesen HH. Formulation and stability testing of photolabile
drugs. Int J Pharm. 2001;225(1–2):1–14.
22. Li Y, Polozova A, Gruia F, Feng J. Characterization of the degra-
dation products of a color-changed monoclonal antibody:
tryptophan-derived chromophores. Anal Chem. 2014;86(14):
6850–7.
33. Chi EY, Krishnan S, Randolph TW, Carpenter JF. Physical stabil-
ity of proteins in aqueous solution: mechanism and driving forces in
nonnative protein aggregation. Pharm Res. 2003;20(9):1325–36.
34. Kamerzell TJ, Esfandiary R, Joshi SB, Middaugh CR, Volkin DB.
Protein–excipient interactions: mechanisms and biophysical char-
acterization applied to protein formulation development. Adv Drug
Deliv Rev. 2011;63(13):1118–59.
35. Ohtake S, Kita Y, Arakawa T. Interactions of formulation exci-
pients with proteins in solution and in the dried state. Adv Drug
Deliv Rev. 2011;63(13):1053–73.
36. Maity H, O'Dell C, Srivastava A, Goldstein J. Effects of arginine on
photostability and thermal stability of IgG1 monoclonal antibodies.
Curr Pharm Biotechnol. 2009;10(8):761–6.
37. Arakawa T, Timasheff S. The mechanism of action of Na gluta-
mate, lysine HCl, and piperazine-N, N'-bis (2-ethanesulfonic acid)
in the stabilization of tubulin and microtubule formation. J Biol
Chem. 1984;259(8):4979–86.
38. Dion MZ, Leiske D, Sharma VK, de Zafra CLZ, Salisbury CM.
Mitigation of oxidation in therapeutic antibody formulations: a
biochemical efficacy and safety evaluation of N-acetyl-tryptophan
and L-methionine. Pharm Res. 2018;35(11):222.
39. Arakawa T, Timasheff SN. Stabilization of protein structure by
sugars. Biochemistry. 1982;21(25):6536–44.
40. Sudrik CM, Cloutier T, Mody N, Sathish HA, Trout BL.
Understanding the role of preferential exclusion of sugars and pol-
yols from native state IgG1 monoclonal antibodies and its effect on
aggregation and reversible self-association. Pharm Res. 2019;36(8):
109–9.
41. Amani M, Barzegar A, Mazani M. Osmolytic effect of sucrose on
thermal denaturation of pea seedling copper amine oxidase.
Protein J. 2017;36(2):147–53.
42. Oshima H, Kinoshita M. Effects of sugars on the thermal stability of
a protein. The Journal of Chemical Physics. 2013;138(24):
06B612_611.
43. Flood A, Estrada M, McAdams D, Ji Y, Chen D. Development of a
freeze-dried, heat-stable influenza subunit vaccine formulation.
PloS One. 2016;11(11).
44. Fischer S, Hoernschemeyer J, Mahler H-C. Glycation during stor-
age and administration of monoclonal antibody formulations. Eur J
Pharm Biopharm. 2008;70(1):42–50.
45. Ouellette D, Alessandri L, Piparia R, Aikhoje A, Chin A,
Radziejewski C, et al. Elevated cleavage of human immunoglobu-
lin gamma molecules containing a lambda light chain mediated by
iron and histidine. Anal Biochem. 2009;389(2):107–17.
23. Subelzu N, Schöneich C. Near UV and visible light induce iron-
dependent photo-degradation reactions in pharmaceutical buffers:
mechanistic and product studies. Mol Pharm. 2020;17(11):4163–
79.
24. Falconer RJ. Advances in liquid formulations of parenteral thera-
peutic proteins. Biotechnol Adv. 2019;37(7):107412.
25. Pramanick S, Singodia D, Chandel V. Excipient selection in par-
enteral formulation development. Pharma Times. 2013;45(3):65–
77.
26. Wlodarczyk SR, Custódio D, Pessoa A Jr, Monteiro G. Influence
and effect of osmolytes in biopharmaceutical formulations. Eur J
Pharm Biopharm. 2018;131:92–8.
27. LeClair DA, Cranston ED, Xing Z, Thompson MR. Evaluation of
excipients for enhanced thermal stabilization of a human type 5

Pharm Res
46. Adem YT, Molina P, Liu H, Patapoff TW, Sreedhara A, Esue O.
Hexyl glucoside and hexyl maltoside inhibit light-induced oxidation
of tryptophan. J Pharm Sci. 2014;103(2):409–16.
62. Hiller K, Masloch B, Goebl M, Asmus K. Mechanism of the hy-
droxyl radical induced oxidation of methionine in aqueous solution.
J Am Chem Soc. 1981;103(10):2734–43.
63. Morelli R, Russo-Volpe S, Bruno N, Lo SR. Fenton-dependent
damage to carbohydrates: free radical scavenging activity of some
simple sugars. J Agric Food Chem. 2003;51(25):7418–25.
64. Barata-Vallejo S, Ferreri C, Postigo A, Chatgilialoglu C. Radiation
chemical studies of methionine in aqueous solution: understanding
the role of molecular oxygen. Chem Res Toxicol. 2010;23(1):258–
63.
65. Enns K, Burgess WH. The photochemical oxidation of ethylene-
diaminetetraacetic acid and methionine by Riboflavin1. J Am
Chem Soc. 1965;87(24):5766–70.
66. Udayakumar S, Ajaikumar S, Pandurangan A. Electrophilic sub-
stitution reaction of phenols with aldehydes: enhance the yield of
bisphenols by HPA and supported HPA. Catal Commun.
2007;8(3):366–74.
67. Faith HE. Aldehyde-phenol reaction products and derivatives. J
Am Chem Soc. 1950;72(2):837–9.
68. Asaduzzaman M, Scampicchio M, Biasioli F, Bremer PJ, Silcock P.
Methanethiol formation during the photochemical oxidation of
methionine-riboflavin system. Flavour and Fragrance Journal.
2020;35:34–41.
69. Sreedhara A, Wong RL, Lentz Y, Schoenhammer K, Stark C.
Application of QbD principles to late-stage formulation develop-
ment for biological liquid products. In. Quality by Design for
Biopharmaceutical Drug Product Development: Springer; 2015.
p. 115–135, Application of QbD Principles to Late-Stage
Formulation Development for Biological Liquid Products.
70. Wurth C, Demeule B, Mahler H-C, Adler M. Quality by design
approaches to formulation robustness—an antibody case study. J
Pharm Sci. 2016;105(5):1667–75.
47. International Council for Harmonisation of Technical
Requirements for Pharmaceuticals for Human Use. Guideline
for elemental impurities Q3D (R1). 2019.
48. Baertschi SW, Clapham D, Foti C, Jansen PJ, Kristensen S, Reed
RA, et al. Implications of in-use photostability: proposed guidance
for photostability testing and labeling to support the administration
of photosensitive pharmaceutical products, part 1: drug products
administered by injection. J Pharm Sci. 2013;102(11):3888–99.
49. Cai H, Liu X, Zou J, Xiao J, Yuan B, Li F, et al. Multi-wavelength
spectrophotometric determination of hydrogen peroxide in water
with peroxidase-catalyzed oxidation of ABTS. Chemosphere.
2018;193:833–9.
50. Illes E, Patra SG, Marks V, Mizrahi A, Meyerstein D. The
Fe(II)(citrate) Fenton reaction under physiological conditions. J
Inorg Biochem. 2020;206:111018.
51. Li S-C, Goto NK, Williams KA, Deber CM. Alpha-helical, but not
beta-sheet, propensity of proline is determined by peptide environ-
ment. Proc Natl Acad Sci. 1996;93(13):6676–81.
52. Bommana R, Mozziconacci O, Wang YJ, Schöneich C. An effi-
cient and rapid method to monitor the oxidative degradation of
protein pharmaceuticals: probing tyrosine oxidation with fluoro-
genic derivatization. Pharm Res. 2017;34(7):1428–43.
53. Hanlon MC, Seybert DW. The pH dependence of lipid peroxida-
tion using water-soluble Azo initiators. Free Radic Biol Med.
1997;23(5):712–9.
54. Schoneich C, Aced A, Asmus KD. Halogenated peroxyl radicals as
two-electron-transfer agents. Oxidation of organic sulfides to sulf-
oxides. J Am Chem Soc. 1991;113(1):375–6.
55. Davies MJ. The oxidative environment and protein damage.
Biochim Biophys Acta. 2005;1703(2):93–109.
71. Walling C, Kurz M, Schugar HJ. Iron(Iii)-Ethylenediaminetetraacetic
Acid-Peroxide System. Inorg Chem. 1970;9(4):931-&.
56. Sharma CP, Karim AV, Shriwastav A. Decolorization of methy-
lene blue using Fe(III)-citrate complex in a solar photo-Fenton pro-
cess: impact of solar variability on process optimization. Water Sci
Technol. 2019;80(11):2047–57.
57. Marion A, Brigante M, Mailhot G. A new source of ammonia and
carboxylic acids in cloud water: the first evidence of photochemical
process involving an iron-amino acid complex. Atmos Environ.
2018;195:179–86.
72. Rahhal S, Richter HW. Reduction of Hydrogen-Peroxide by the
Ferrous Iron Chelate of Diethylenetriamine-N,N,N',N”,N”-
Pentaacetate. J Am Chem Soc 1988;110(10):3126–3133.
73. Li LX, Abe Y, Kanagawa K, Shoji T, Mashino T, Mochizuki M,
et al. Iron-chelating agents never suppress Fenton reaction but par-
ticipate in quenching spin-trapped radicals. Anal Chim Acta.
2007;599(2):315–9.
74. Brausam A, van Eldik R. Further mechanistic information on the
reaction between Fe-III(edta) and hydrogen peroxide: observation
of a second reaction step and importance of pH. Inorg Chem.
2004;43(17):5351–9.
58. Hawkins C, Davies M. EPR studies on the selectivity of hydroxyl
radical attack on amino acids and peptides. Journal of the Chemical
Society, Perkin Transactions 2. 1998(12):2617–2622.
75. Rachmilovich-Calis S, Masarwa A, Meyerstein N, Meyerstein D,
van Eldik R. New mechanistic aspects of the Fenton reaction.
Chem-Eur J. 2009;15(33):8303–9.
76. Graf E, Mahoney JR, Bryant RG, Eaton JW. Iron-catalyzed hy-
droxyl radical formation. Stringent requirement for free iron coor-
dination site. J Biol Chem. 1984;259(6):3620–4.
59. Perrin DD. The Stability of Complexes of Ferric Ion and Amino-
Acids. J Chem Soc. 1958(Sep):3125–3128.
60. Neubecker TA, Kirksey ST, Chellappa KL, Margerum DW.
Amine Deprotonation in copper(iii)-peptide complexes. Inorg
Chem. 1979;18(2):444–8.
61. Sysak PK, Foote CS, Ching TY. Chemistry of Singlet Oxygen .25.
Photooxygenation of Methionine. Photochem Photobiol
1977;26(1):19–27, Chemistry of Singlet Oxygen—xxv.
Photooxygenation of Methionine.
Publisher’s Note Springer Nature remains neutral with regard to jurisdic-
tional claims in published maps and institutional affiliations.