Nonenzymatic, self-elimination degradation mechanism of glutathione

Article abstract of DOI:10.1002/cbdv.200800277

In the present in vitro studies, evidence is provided showing that glutathione (GSH) can undergo spontaneous, nonenzymatic auto-degradation. The initial cleavage of the Glu-Cys bond involves nucleophilic attack of the N-terminal amino group of GSH at the

Full text of DOI:10.1002/cbdv.200800277

CHEMISTRY & BIODIVERSITY – Vol. 6 (2009)
527
Nonenzymatic, Self-Elimination Degradation Mechanism of Glutathione
by Manjeet Deshmukha)b), Hilliard Kutschera), Stanley Steina)b) and Patrick Sinko*a)b)
a) Department of Pharmaceutics, Rutgers University, Piscataway, NJ 08854, USA
b) UMDNJ-Rutgers Counteract Research Center of Excellence, Rutgers University, Piscataway,
NJ 08854, USA (phone: þ1-732-445-3831*213; fax: þ1-732-445-3134; e-mail: sinko@rutgers.edu)
In the present in vitro studies, evidence is provided showing that glutathione (GSH) can undergo
spontaneous, nonenzymatic auto-degradation. The initial cleavage of the Glu-Cys bond involves
nucleophilic attack of the N-terminal amino group of GSH at the g-carbonyl side chain, followed by an
unusual cleavage of the Cys-Gly amide bond that is complete within 3–5 weeks in the physiological pH
range. These observations may be useful for the development of novel biodegradable polymers and
releasable prodrugs for sustained and targeted drug delivery applications.
Introduction. – Glutathione (GSH) is the most abundant thiol species in the
cytoplasm and the major reducing agent for almost all biochemical processes. GSH is
important in the synthesis of proteins and DNA, transport, enzyme activity,
metabolism, and protection of cells [1]. Cellular GSH levels must be maintained,
since they are constantly fluctuating due to degradation and consumption in various
biological processes [2]. The intracellular concentration of GSH (1–10 mm) is
substantially higher than the extracellular concentration (e.g., 2–20 mm in plasma)
[3]. Extracellular GSH concentrations are relatively low, because GSH is readily
oxidized to glutathione disulfide (GSSG) by direct interaction with free radicals or,
more often, when GSH acts as a cofactor for antioxidant enzymes such as GSH
peroxidases [4]. The efflux of GSSG from cells contributes to a net loss of intracellular
GSH. GSH/GSSG is a very important redox couple that plays a crucial role in
antioxidant defense, nutrient metabolism, and the regulation of this pathway is
essential for whole body homeostasis [5]. GSH Deficiency in the cell contributes to
oxidative stress, which plays a key role in aging and the pathogenesis of many diseases
including Alzheimerꢀs disease, Parkinsonꢀs disease, AIDS, cancer, and myocardial
infarction [6]. The chemical structure of GSH is shown in Fig. 1.
The biosynthesis and metabolism of GSH, which proceeds through the g-glutamyl
cycle, has been studied and thoroughly characterized [7][8]. GSH Degradation is
initiated by g-glutamyl transpeptidase, which involves either the cleavage and transfer
of the g-glutamyl moiety from GSH to an acceptor amino acid, or the hydrolysis and
Fig. 1. Chemical representation of GSH. Annotations de-
note chemical shift assignments in the NMR spectra: A:
Glu-a, B: Glu-b, C: Glu-g, D: Cys-a, E: Cys-b, F: Gly-a.
Arrows denote the first step in the proposed degradation
mechanism.
ꢁ 2009 Verlag Helvetica Chimica Acta AG, Zꢂrich

528
CHEMISTRY & BIODIVERSITY – Vol. 6 (2009)
release of glutamate accompanied by the release of cysteinyl-glycine. The resulting g-
glutamyl amino acids formed by g-glutamyl transpeptidase are transported into cells
where g-glutamyl cyclotransferase converts these compounds into the corresponding
amino acid and 5-oxo-l-proline. The intracellular enzyme 5-oxo-l-prolinase catalyzes
the conversion of 5-oxo-l-proline to l-glutamate. The cysteinyl-glycine formed in the
transpeptidase reaction is split by dipeptidase. It is known that g-glutamylcysteine is
formed intracellularly by the action of g-glutamylcysteine synthetase [9], and
intracellular g-glutamylcysteine is converted into GSH by the action of glutathione
synthetase [10].
Although, the g-glutamyl cycle is well-established, the possibility of an alternative
nonenzymatic GSH degradation pathway has not been explored. The oxidation of
glutathione into GSSG using oxidizing agents and formation of pyroglutamic acid from
GSH at high temperature was suggested in previous studies [11]. In addition to the
well-known enzymatic reactions of GSH, there are previously unreported nonenzy-
matic side reactions occurring, which include oxidation and peptide bond cleavage.
Such nonenzymatic reactions may be applicable for designing prodrugs that can be
converted to active drugs by an intrinsic mechanism.
Prodrugs that are activated by intramolecular reaction (i.e., prodrugs that are partly
or completely activated without the need for enzymatic contribution) are of great
interest. A few examples of this type of prodrugs have been reported [12]. One natural
example is the biosynthesis of luteinizing hormone releasing hormone (LHRH), in
which spontaneous cyclization of the N-terminal glutamine residue results in the
formation of pyroglutamyl peptide with release of NH3 [12][13].
We hypothesize that the nonenzymatic degradation of GSH (Scheme 1) occurs by
intramolecular rearrangement similar to the LHRH reaction in which nucleophilic
attack of the N-terminal amino group breaks the amide bond between Glu and Cys
with the release of the N-terminal primary amine of Cys, thereby initiating peptide
degradation. To determine the validity of the proposed mechanism, the nonenzymatic
degradation of GSH was systematically investigated at three different pH values: 6.2,
6.8 (i.e., representative of a common pH range in cancer cells), and 7.4 (i.e., pH of
plasma and the most commonly found pH in the body). In the current study, we
demonstrate that nonenzymatic reactions contribute significantly to the degradation of
GSH. In addition, the observed pH-dependent degradation mechanism may be
exploited as a prodrug strategy for the selective cleavage of GSH and the release of
chemotherapeutic agents in the tumor environment.
Results and Discussion. – GSH (150 mg) was dissolved in 100 mm sodium
phosphate buffer (10 ml) at pH 6.2, 6.8, or 7.4, and incubated at 378. Aliquots
(100 ml) were taken at various time points from each of the three reaction mixtures.
1
Sample aliquots were dried using a CentriVap and analyzed by H-NMR (for all time
points) and LC/MS (for the last time point, when degradation was complete).
The NMR spectrum of GSH before degradation showed a well-resolved pattern of
peaks (Exper. Part). The H-atoms (see Fig. 1) at A and D showed a triplet at 3.69 and
4.41 ppm, respectively, whereas the F H-atom showed a singlet at 3.82 ppm. Other
signals belonging to the H-atoms B, C, and E appeared at chemical shifts 2.02, 2.41, and
2.8 ppm, respectively. It was observed that the singlet of the F H-atom shifted in all

CHEMISTRY & BIODIVERSITY – Vol. 6 (2009)
529
Scheme 1. Proposed Degradation Mechanism of GSH
a) 378, 100 mm sodium phosphate buffer at three different pH values (6.2, 6.8, and 7.4).
three pH samples. At pH 6.2, it is shifted upfield from 3.82 ppm and appears at
3.74 ppm, whereas, at pH 6.8 and 7.4, it has shifted from 3.82 to 3.61 ppm. The intensity
of the Cys-a triplet (at 4.41 ppm, D) decreased with time. The rate of disappearance of
the Cys-a peak differs with the change in pH (from 6.2 to 7.4). From the disappearance
profile of the Cys-a triplet (D), we can unambiguously confirm that the amide bond
between Glu and Cys breaks during the incubation at 378 giving rise to degradation
products. NMR Resonance of D was, therefore, chosen for quantification studies (vide
infra). In the later time points, signals arising due to degradation products can be seen
as well. t-BuOH (0.15 ml per 750 ml of sample) was used as an internal standard for
NMR studies, because it does not interact with GSH or with any degradation product as
confirmed by comparison of NMR spectra with and without t-BuOH.
Half-lives for the amide bond cleavage between Glu and Cys in GSH at three
1
different pH values, i.e., 7.4, 6.8, and 6.2, were determined by quantitative H-NMR
recorded with 100 scans and a recycle delay of 15 s. Fig. 2,d, shows the changes in the
1H integral value for peak D derived from quantitative NMR spectra recorded with an

530
CHEMISTRY & BIODIVERSITY – Vol. 6 (2009)

CHEMISTRY & BIODIVERSITY – Vol. 6 (2009)
531
Scheme 2. The Cleavage Mechanism of the GS-SG Intermediate Involves Disulfide Bond Reduction and
Cleavage of Glutamic Acid
a) 378, 100 mm sodium phosphate buffer with three different pH values (6.2, 6.8, and 7.4).
interval of 2 d during the studies. Standard kinetics models were applied to the
degradation data. Unfortunately, the data did not show a good fit to any standard
kinetic model (i.e., zero-, first-, or second-order degradation). This suggests that the
degradation mechanism follows a complex order. It was observed that degradation is
faster at pH 7.4 (t1/2 ¼12 d) than at pH 6.8 (t1/2 ¼14 d) and pH 6.2 (t1/2 ¼17 d).
In the presence of g-glutamyl transpeptidase, the amide bond cleavage between Glu
and Cys in GSH was reported to occur in ca. 4–6 h as found in tissue extracts [14].
However, total formation of the first product in the absence of the enzyme varies with
pH and requires ca. 34 d at pH 6.2, 28 d at pH 6.8, and 24 d at pH 7.4 (Fig. 2,a–c).
NMR Data show that, during the time course, the signal intensities of the a
(4.02 ppm) and b (1.89 and 2.5 ppm) H-atoms of pyroglutamic acid (at all three pH
values; Fig. 2,a–c) increases with time. For the quantification studies of the first
cleavage product, pyroglutamic acid, NMR resonance of the b-pyroglutamic acid H-

532
CHEMISTRY & BIODIVERSITY – Vol. 6 (2009)
Scheme 3. The Cleavage Mechanism of the (Cys-Gly)2 Intermediate Involves Disulfide Bond Reduction
a) 378, 100 mm sodium phosphate buffer with three different pH values (6.2, 6.8, and 7.4).
atom (1.89 ppm) was integrated, and the concentration vs. time curve of pyroglutamic
acid is shown in Fig. 2,e.
After GSH degradation was essentially complete (99%), the identification of
products 5, 6, and 7 strongly suggests the formation of two disulfide intermediates: GS-
SG (Scheme 2) and (Cys-Gly)2 (Scheme 3). However, the disulfide intermediates were
not observed at any pH studied as confirmed by LC/MS (Fig. 3). In situ disulfide bond-
formation between GS-SG or (Cys-Gly)2 is theoretically possible. The disulfide bond
should reduce as per the Schoberl mechanism [15][16], in which HOꢀ attacks the
disulfide bond producing cysteine and the unstable cysteinesulfenic acid, which is
further converted to 2-amino-3-oxopropanoic acid with the release of H2S. The same
mechanism was observed in the degradation studies of GSH at all three pH values. The
oxidized compound (GS-SG) is reduced, when HOꢀ attacks the disulfide bond,
producing GSH and GSH-sulfenic acid. At the same time, the amide bond between Glu
and Cys is cleaved via the nucleophilic attack of the N-terminal amino group, and (Cys-
Gly)2 and Cys-Gly-sulfenic acid are formed after the release of pyroglutamic acid (2>
3). The results show that the disulfide bond in (Cys-Gly)2 reduces and gives Cys-Gly-
sulfenic acid and Cys-Gly. This possible degradation mechanism is described in
Schemes 2 and 3. Further, Cys-Gly-sulfenic acid breaks the amide bond between Cys
and Gly and forms cysteinesulfenic acid (5) and glycine (7). It was observed that
cysteinesulfenic acid (5) is not stable, and it hydrolyzes to 2-amino-3-oxopropanoic acid

CHEMISTRY & BIODIVERSITY – Vol. 6 (2009)
533
Fig. 3. LC/MS of the crude reaction mixture containing seven compounds. Signals are marked corres-
ponding to peaks for cysteinyl-glycine (1), pyroglutamic acid (2) and 5-oxoproline (3), glutamic acid (4),
cysteinesulfenic acid (5), 2-amino-3-oxopropanoic acid (6), and glycine (7). Degradation studies were
carried out at 378 and pH 7.4 (a), 6.8 (b), and 6.2 (c).
(6). Formation of cysteinesulfenic acid (5), 2-amino-3-oxopropanoic acid (6), and
glycine (7) through in situ reduction of (Cys-Gly)2 and GS-SG, is based on the
precedence of the Schoberl mechanism.
Thus, both initial amide bond cleavage between Glu and Cys and total GSH
degradation (99.9%) is fastest at pH 7.4 (24 d) and decreases with pH (pH 6.8: 28 d and
pH 6.2: 34 d). Peak intensities of other protons A, B, C, E, and F (Fig. 1) also decrease,
while new peaks for degradation products are observed (Scheme 1).
Based on the LC/MS analysis (Fig. 3), a nonenzymatic degradation pathway was
deduced. In this pathway, the GSH ion at m/z 307 decreased with time until it was
undetectable. After complete degradation of GSH, intense peaks at m/z 128 for
pyroglutamic acid and its tautomer, 5-oxoproline (2), as well as at m/z 145 for glutamic
acid (4) were observed. All three compounds were derived from the same glutamic acid
moiety. Also seen was the peak at m/z 179 corresponding to Cys-Gly (1), which
constitutes the remainder of the peptide after the initial cleavage reaction. There is a
peak at m/z 135 evidencing the presence of cysteinesulfenic acid (5), apparently
resulting from oxidation of the disulfide bond in (Cys-Gly)2, followed by intra-

534
CHEMISTRY & BIODIVERSITY – Vol. 6 (2009)

CHEMISTRY & BIODIVERSITY – Vol. 6 (2009)
535

536
CHEMISTRY & BIODIVERSITY – Vol. 6 (2009)
molecular cleavage of amide bond between Cys-Gly. The cysteinesulfenic acid was
converted into 2-amino-3-oxopropanoic acid (6), the peak of which was observed at
m/z 101. The peak for glycine (7) was observed at m/z 75, completing the assignment of
the ions derived from the GSH degradation.
Additional studies were performed with 100 mm sodium phosphate buffer and in
the absence of O2 (by flushing with Ar) at all three pH values. In the absence of
oxidative environment, no changes were observed to the initial kinetics and
degradation (Fig. 4,a–c). Mass spectra of the degradation products show that
degradation is greatly hindered after the cleavage of the first amide bond between
Glu and Cys, suggesting further degradation requires the presence of oxidative agents
(Fig. 5,a). To understand the effect of the presence of a chelating agent, we have
Fig. 5. Mass spectra of the crude reaction mixtures. Degradation studies were carried out at pH 7.4, a) in
the absence of oxygen, b) in 100 mm sodium phosphate buffer containing 1 mm EDTA, and c) 20 mm
sodium phosphate buffer. Similar results were obtained for other two pH values. Signals are marked
corresponding to peaks for cysteinyl-glycine (1), pyroglutamic acid (2) and 5-oxoproline (3), glutamic
acid (4), cysteinesulfenic acid (5), 2-amino-3-oxopropanoic acid (6), and glycine (7).

CHEMISTRY & BIODIVERSITY – Vol. 6 (2009)
537
repeated degradation studies, in the presence of 1 mm EDTA (containing 100 mm
sodium phosphate buffer). We found that the rate of hydrolysis was unaffected in all
three cases of pH (Fig. 4,d–f, and Fig. 5,b). To study the effect of the ionic strength on
degradation, similar studies were performed using 20 mm sodium phosphate buffer. It
was observed that different ionic strength at the same pH does not affect the
degradation time (Fig. 4,g–i, and Fig. 5,c). Hence, this supports the conclusion that the
auto-degradation of GSH follows the pathway proposed in this article. The composite
data indicates that the elimination reaction, initiated by the nucleophilic attack of the
N-terminal amino group in GSH (Scheme 1), is the most likely possibility.
Another possibility for the degradation of GSH is via thiolysis of the free ꢀSH
group of Cys with the g-COOH group of Glu followed by thioester bond cleavage. This
is not consistent with the data. The products obtained after the degradation,
cysteinesulfenic acid (5), followed by 2-amino-3-oxopropanoic acid (6) and Gly (7),
would not occur due to thiolysis followed by thioester bond degradation. Thiolysis
would cause the shift of D proton from 4.41 to 3.4–3.7 ppm. Instead, it was observed
that the intensity of the Cys-a triplet (D) gradually decreased with time, and the
resonance disappeared completely within 24 d at pH 7.4. Furthermore, after the
disappearance of the D signal (after complete degradation), the thiolysis product was
not observed in the LC/MS spectrum. The absence of the thiolysis product (thiolysis of
free ꢀSH group of Cys with g-COOH group of Glu) was confirmed at all three pH
values (Fig. 3). Further proof was obtained by analyzing the controls (tert-butoxy)-
carbonyl- and trityl-protected GSH, and other peptide sequences (GABA-Arg, eAhx-
Arg) containing g-aminobutyric acid (GABA) and 6-aminohexanoic acid (e-Ahx)
instead of glutamic acid, which did not undergo degradation. These studies prove that
other peptides (g-amino acid or e-amino acid) do not degrade as per the proposed
mechanism and confirm that the initial cleavage of the Glu-Cys bond involves
nucleophilic attack of the N-terminal amino group of GSH at the g-carbonyl side chain,
followed by an unusual cleavage of the Cys-Gly amide.
Conclusions. – In the present study, we demonstrated for the first time that GSH
degradation occurs in the absence of g-glutamyl transpeptidase. GSH undergoes
spontaneous nonenzymatic auto-degradation at various rates, depending on pH. This
pH range covers values typically observed in normal physiological (i.e., plasma pH 7.4)
and pathological conditions (i.e., pH 6.2 and 6.8 commonly found in the local tumor
environment). The mechanism of the nonenzymatic degradation of GSH involves the
formation of two disulfide intermediates, GS-SG and (Cys-Gly)2. The initial cleavage is
at the Glu-Cys amide bond by nucleophilic attack of the N-terminal amino group of
GSH at the g-carbonyl side chain, followed by an unusual cleavage of the Cys-Gly
amide bond. The proposed nonenzymatic degradation mechanism may have utility in
drug targeting especially for diseases where local pH is altered from typical
physiological levels. Furthermore, the time scale (t1/2 ca. 2–3 weeks) is useful for
sustained release applications.
This research is supported by the Parke-Davis Endowed Chair in Pharmaceutics and Controlled Drug
Delivery, Hikma Pharma PLC, and National Institutes of Health CounterACT Program through the
National Institute of Arthritis and Musculoskeletal and Skin Diseases (Award #U54AR055073). Its

538
CHEMISTRY & BIODIVERSITY – Vol. 6 (2009)
contents are solely the responsibility of the authors and do not necessarily represent the official views of
the federal government. We would like to thank Drs. Simi Gunaseelan, Yashveer Singh, and Dayuan Gao
for their assistance and fruitful discussions. Also we would like to thank Scott Pfeil for his help.
Experimental Part
1. General. Abbreviations: GSH: glutathione, Glu: glutamic acid, Cys: cysteine, Gly: glycine, GABA:
g-aminobutyric acid, e-Ahx: 6-aminohexanoic acid. LC/MS: Liquid chromatography/tandem mass
1
spectrometry. Solvents: sodium phosphate buffer. Glutathione was purchased from Sigma-Aldrich. H-
and 13C-NMR spectra: 400-MHz spectrometer (Varian) with the chemical shift values reported in d units
(ppm). LC/MS: API-150, Applied Biosciences.
1
3
2. Data of Glutathione (GSH). H-NMR (300 MHz, D2O; Fig. 6): 4.41 (t, J(Cys-a,Cys-b) ¼ 6.59,
1 H, Cys-a (D)); 3.82 (br. s, 2 H, Gly-a (F)); 3.69 (t, 3J(Glu-a,Glu-b) ¼6.32, 1 H, Glu-a (A)); 2.8 (dd,
3J(Cys-b,Cys-b)¼6.59, 3J(Cys-b,Cys-a)¼3.4, 2 H, Cys-b (E)); 2.41 (ddd, 2J(Glu-g,Glu-g) ¼15.24,
3J(Glu-g ,Glu-b)¼7.41, 3J(Glu-g,Glu-b)¼2.74, 2 H, Glu-g (C)); 2.02 (q, 3J(Glu-b,Glu-g) ¼7.42, 2 H,
Glu-b (B)). LC/MS: 307 ([Mþ1]).
Fig. 6. 1H-NMR Spectrum of GSH in D2O
REFERENCES
[1] A. Meister, M. E. Anderson, Annu. Rev. Biochem. 1983, 52, 711.
[2] G. G. Perrone, C. M. Grant, I. W. Dawes, Mol. Biol. Cell. 2005, 16, 218.
[3] S. S. M. Hassan, G. A. Rechnitz, Anal. Chem. 1982, 54, 1972; D. P. Jones, J. L. Carlson, P. S. Samiec,
P. Sternberg Jr., V. C. Mody Jr., R. L. Reed, L. A. S. Brown, Clin. Chim. Acta 1998, 275, 175; M. E.
Anderson, Chem.-Biol. Interact. 1998, 111–112, 1; C. Friesen, Y. Kiess, K.-M. Debatin, Cell Death
Differ. 2004, 11, S73.
[4] G. K. Balendiran, R. Dabur, D. Fraser, Cell Biochem. Funct. 2004, 22, 343.
[5] O. W. Griffith, Free Radical Biol. Med. 1999, 27, 922.
[6] G. Wu, Y.-Z. Fang, S. Yang, J. R. Lupton, N. D. Turner, J. Nutr. 2004, 134, 489.

CHEMISTRY & BIODIVERSITY – Vol. 6 (2009)
539
[7] A. Meister, S. S. Tate, Annu. Rev. Biochem. 1976, 45, 559; S. S. Tate, A. Meister, Mol. Cell. Biochem.
1981, 39, 357; O. W. Griffith, R. J. Bridges, A. Meister, Proc. Natl. Acad. Sci. U.S.A. 1979, 76, 6319.
[8] A. Meister, S. S. Tate, O. W. Griffith, Methods Enzymol. 1981, 77, 237.
[9] A. Meister, ꢃThe Enzymesꢀ, 3rd edn., Academic, New York, 1974, p. 10, p. 671; O. W. Griffith, R. J.
Bridges, A. Meister, Proc. Natl. Acad. Sci. U.S.A. 1981, 78, 2777.
[10] O. W. Griffith, A. Meister, Proc. Natl. Acad. Sci. U.S.A. 1979, 76, 268; O. W. Griffith, A. Meister,
Proc. Natl. Acad. Sci. U.S.A. 1980, 77, 3384.
[11] F. G. Hopkins, J. Biol. Chem. 1929, 84, 269; H. L. Mason, J. Biol. Chem. 1931, 90, 409; E. C. Kendall,
H. L. Mason, B. F. McKenzie, J. Biol. Chem. 1930, 88, 409; H. L. Mason, J. Biol. Chem. 1931, 90, 25.
[12] B. Testa, J. M. Mayer, ꢃ Hydrolysis in Drug and Prodrug Metabolismꢀ, Verlag Helvetica Chimica
Acta, Zꢂrich, Wiley-VCH, Weinheim, 2003, p. 235.
[13] A. L. Lehninger, ꢃBiochemistry: The Molecular Basis of Cell Structure and Functionsꢀ, Worth
Publishers, New York, 1975, p. 583, p. 807; G. Chen, J. B. Russell, J. Bacteriol. 1989, 171, 2981.
[14] A. G. Palekar, S. S. Tate, A. Meister, Proc. Natl. Acad. Sci. U.S.A. 1974, 71, 293; E. F. Schroeder,
M. P. Munro, L. Weil, J. Biol. Chem. 1935, 110, 181.
[15] A. Schçberl, Justus Liebigs Ann. Chem. 1933, 507, 111.
[16] F. M. Robbins, J. A. Fioriti, Nature (London) 1963, 200, 577.
Received September 8, 2008