Kinetics and equilibria of the thiol/disulfide exchange reactions of somatostatin with glutathione

Article abstract of DOI:10.1021/jo960917+

Rate and equilibrium constants are reported for the thiol/disulfide exchange reactions of the peptide hormone somatostatin with glutathione (GSH). GSH reacts with the disulfide bond of somatostatin to form somatostatin-glutathione mixed disulfides (Cys³-SH, Cys¹⁴-SSG and Cys³-SSG, Cys¹⁴-SH), each of which can react with another molecule of GSH to give the reduced dithiol form of somatostatin and GSSG. The mixed disulfides also can undergo intramolecular thiol/disulfide exchange reactions to re-form the disulfide bond of somatostatin or to interconvert to the other mixed disulfide. Analysis of the forward and reverse rate constants indicates that, at physiological concentrations of GSH, the intramolecular thiol/disulfide exchange reactions that re-form the disulfide bond of somatostatin are much faster than reaction of the mixed disulfides with another molecule of GSH, even though the intramolecular reaction involves closure of a 38-membered ring. Thus, even though the disulfide bond of somatostatin is readily cleaved by thiol/disulfide exchange, it is rapidly reformed by intramolecular thiol/disulfide exchange reactions of the somatostatin-glutathione mixed disulfides. By comparison with rate constants reported for analogous reactions of model peptides measured under random coil conditions, it is concluded that disulfide bond formation by intramolecular thiol/disulfide exchange in the somatostatin-glutathione mixed disulfides is not completely random, but rather it is directed to some extent by conformational properties of the mixed disulfides that place the thiol and mixed disulfide groups in close proximity. A reduction potential of -0.221 V was calculated for the disulfide bond of somatostatin from the thiol/disulfide exchange equilibrium constant.

Full text of DOI:10.1021/jo960917+

J . Org. Chem. 1996, 61, 7391-7397
7391
Kin etics a n d Equ ilibr ia of th e Th iol/Disu lfid e Exch a n ge Rea ction s
of Som a tosta tin w ith Glu ta th ion e
Dallas L. Rabenstein* and Kim H. Weaver
Department of Chemistry, University of California, Riverside, California 92521
Received May 17, 1996X
Rate and equilibrium constants are reported for the thiol/disulfide exchange reactions of the peptide
hormone somatostatin with glutathione (GSH). GSH reacts with the disulfide bond of somatostatin
to form somatostatin-glutathione mixed disulfides (Cys3-SH, Cys14-SSG and Cys3-SSG, Cys14-SH),
each of which can react with another molecule of GSH to give the reduced dithiol form of
somatostatin and GSSG. The mixed disulfides also can undergo intramolecular thiol/disulfide
exchange reactions to re-form the disulfide bond of somatostatin or to interconvert to the other
mixed disulfide. Analysis of the forward and reverse rate constants indicates that, at physiological
concentrations of GSH, the intramolecular thiol/disulfide exchange reactions that re-form the
disulfide bond of somatostatin are much faster than reaction of the mixed disulfides with another
molecule of GSH, even though the intramolecular reaction involves closure of a 38-membered ring.
Thus, even though the disulfide bond of somatostatin is readily cleaved by thiol/disulfide exchange,
it is rapidly reformed by intramolecular thiol/disulfide exchange reactions of the somatostatin-
glutathione mixed disulfides. By comparison with rate constants reported for analogous reactions
of model peptides measured under random coil conditions, it is concluded that disulfide bond
formation by intramolecular thiol/disulfide exchange in the somatostatin-glutathione mixed
disulfides is not completely random, but rather it is directed to some extent by conformational
properties of the mixed disulfides that place the thiol and mixed disulfide groups in close proximity.
A reduction potential of -0.221 V was calculated for the disulfide bond of somatostatin from the
thiol/disulfide exchange equilibrium constant.
In tr od u ction
disulfide exchange with glutathione (γ-L-glutamyl-L-
cysteinylglycine, GSH) and oxidized glutathione (GSSG).7,8
The reactions take place in two steps:
Disulfide bonds are structural elements in many
biologically active peptides, including some peptide hor-
mones and peptide toxins. For example, the nonapeptide
hormones oxytocin (OT) and arginine vasopressin (AVP)
both have a disulfide bond between cysteine residues at
positions one and six, and the tetradecapeptide hormone
somatostatin (SS) has a disulfide bond between cysteine
residues at positions 3 and 14. In early reports, the
biological activity of OT and AVP was accounted for in
terms of a mechanism in which they interacted with their
receptors by covalent bond formation via thiol/disulfide
exchange reactions.1 However, structure-activity stud-
ies have shown that the disulfide bond is not involved in
the mechanism of action, but rather it is the cyclic
arrangement of amino acids 1-6 that is essential for high
biological activity.2 The disulfide bonds in OT, AVP, and
somatostatin serve to form and keep molecular confor-
mations suitable for noncovalent interaction with their
receptors.
Rate constant k1 was found to be 1-2 orders of
magnitude larger than rate constants for thiol/disulfide
exchange reactions of disulfide bonds between acyclic
peptides. However, it was also found that, at physiologi-
cal concentrations of GSH,9 the intermediate mixed
disulfides formed in this first step of the two-step reaction
sequence undergo intramolecular thiol/disulfide exchange
to re-form the native disulfide bonds much faster than
they react with another molecule of GSH via the second
reaction, even though the intramolecular thiol/disulfide
exchange reactions involve closure of 20-membered rings.
Furthermore, the rate constants for intramolecular thiol/
disulfide exchange are significantly larger than reported
for analogous reactions of model peptides under random
coil conditions, which suggests that the AVP-GSH and
OT-GSH mixed disulfides exist, at least part of the time,
in precyclic conformations that place the thiol and mixed
disulfide groups in close proximity.7
Disulfide bonds are kinetically unstable structural
elements, being readily cleaved by thiols via thiol/
disulfide exchange.3-7 In a previous study, we character-
ized the kinetics and equilibria for reduction and forma-
tion of the disulfide bonds of OT and AVP by thiol/
* To whom correspondence should be addressed. Phone: (909) 787-
3585. Fax: (909) 787-4713. E-mail: dlrab@mail.ucr.edu.
X Abstract published in Advance ACS Abstracts, October 1, 1996.
(1) (a) Fong, C. T. O.; Schwartz, I. L.; Popenoe, E. A.; Silver, L.;
Schoesler, M. A. J . Am. Chem. Soc. 1959, 81, 2592-2593. (b) Fong, C.
T. O.; Silver, L.; Christman, D. R.; Schwartz, I. L. Proc. Natl. Acad.
Sci. U.S.A. 1960, 46, 1273-1277. (c) Rasmussen, H.; Schwartz, I. L.;
Schoesler, M. A.; Hochster, G. Proc. Natl. Acad. Sci. U.S.A. 1960, 46,
1278-1287. (d) Schwartz, I. L.; Fong, C. T. O.; Popenoe, E. A.; Silver,
L.; Schoesler, M. A. J . Clin. Invest. 1959, 38, 1041. (e) Schwartz, I. L.;
Rasmussen, H.; Schoesler, M. A.; Silver, L.; Fong, C. T.O. Proc. Natl.
Acad. Sci. U.S.A. 1960, 46, 1288-1298.
(3) Szajewski, R. P.; Whitesides, G. M. J . Am. Chem. Soc. 1980, 102,
2011-2026.
(4) Ziegler, D. M. Ann. Rev. Biochem. 1985, 54, 305-329.
(5) Gilbert, H. F. Adv. Enzymol. 1990, 63, 69-172.
(6) Keire, D. A.; Strauss, E.; Guo, W.; Nosza´l, B.; Rabenstein, D. L.
J . Org. Chem. 1992, 57, 123-127.
(7) Rabenstein, D. L.; Yeo, P. L. J . Org. Chem. 1994, 59, 4223-4229.
(8) Rabenstein, D. L.; Yeo, P. L. Bioorg. Chem. 1995, 23, 109-118.
(9) The concentration of GSH in human plasma is in the range of
2-5 µM: (a) Henning, S. M.; Zhang, J . Z.; McKee, R. W.; Swendseid,
M. E.; J acob, R. A. J . Nutr. 1991, 121, 1969-1975. (b) Mansoor, M.
A.; Svardal, A. M.; Ueland, P. M. Anal. Biochem. 1992, 200, 218-229.
(2) J osˇt, K. In Handbook of Neurohypophyseal Peptide Hormones;
J osˇt, K., Lebl, M., Brtnik, F., Eds.; CRC Press, Inc.: Boca Raton, FL,
1987; Vol. 1, Part 2, pp 144-155.
S0022-3263(96)00917-6 CCC: $12.00 © 1996 American Chemical Society

7392 J . Org. Chem., Vol. 61, No. 21, 1996
Rabenstein and Weaver
F igu r e 1. Thiol/disulfide exchange reactions of somatostatin (SS) with glutathione (GSH) and glutathione disulfide (GSSG).
MD1 and MD2 are single somatostatin-glutathione mixed disulfides, with the disulfide bonds at Cys14 and Cys3 of somatostatin,
respectively, and SH2 is the reduced, dithiol form of somatostatin.
In view of this finding, it is of interest to determine if
the structurally important disulfide bonds in other pep-
tide hormones are also rapidly reformed from their mixed
disulfides by intramolecular thiol/disulfide exchange. In
this paper, we report the results of a study of the kinetics
and equilibria of the formation and reduction of the
disulfide bond in somatostatin, which has an even larger
(38 membered) disulfide-containing ring.
disulfides with another molecule of GSH and of SH2 with
GSSG to form the mixed disulfides were characterized
in terms of the rate constants defined by eq 2.
Th iol/Disu lfid e Exch a n ge Equ ilibr iu m Con sta n ts.
The equilibrium constants defined by eqs 3-11 were
determined at 25 °C in pH 7.0 sodium phosphate buffer
(0.075 M).
[MD2]
K1a
K1b
)
)
(3)
(4)
Ala-Gly-
[SS][GSH]
Cys-Lys-Asn-Phe-Phe-Trp-Lys-Thr-Phe-Thr-Ser-Cys
[MD1]
[SS][GSH]
Equilibrium constants and forward and reverse rate
constants for the formation and reduction of the disulfide
bond of somatostatin by thiol/disulfide exchange with the
GSH/GSSG system were determined, including specific
rate constants for the intramolecular thiol/disulfide
exchange reactions of each of the two possible mixed
disulfides. Also, the redox potential of the disulfide bond
of somatostatin was determined from the equilibrium
constant for its thiol/disulfide exchange reaction with
GSH.
[SH2][GSSG]
[MD2][GSH]
K2a
)
)
(5)
(6)
(7)
(8)
(9)
[SH2][GSSG]
[MD1][GSH]
K2b
[MD1] + [MD2]
K1 )
[SS][GSH]
Resu lts
[SH2][GSSG]
K2 )
The thiol/disulfide exchange reactions of somatostatin
(SS) with GSH and GSSG are summarized in Figure 1.
GSH reacts with the disulfide bond of SS to form a mixed
disulfide (MD1 or MD2), which in turn reacts with
another molecule of GSH to form reduced somatostatin
(SH2) and GSSG. The reactions are reversible, and
GSSG reacts with the thiol groups of SH2 to form MD1
or MD2, which in turn undergo intramolecular thiol/
disulfide exchange to form the disulfide bond of SS. In
addition, the mixed disulfides can interconvert by in-
tramolecular thiol/disulfide exchange and they can react
with another molecule of GSSG to form the double mixed
disulfide (DMD) (not shown in Figure 1). Equilibrium
constants were determined for each of the steps shown
in Figure 1 and for formation of the double mixed
disulfide. The kinetics of the reaction of SS with GSH
to form MD1 and MD2 and of the intramolecular thiol/
disulfide exchange reactions of MD1 and MD2 were
characterized in terms of the rate constants shown in
Figure 1, while the kinetics of the reaction of the mixed
([MD1]+[MD2])[GSH]
[SH2][GSSG]
[SS][GSH]2
Kov
)
[MD1]
K3 )
(10)
(11)
[MD2]
[DMD][GSH]2
[SH2][GSSG]2
K4 )
The concentrations of SS and SH2 and of the two
somatostatin-glutathione mixed disulfides were deter-
mined in equilibrium mixtures by HPLC. A typical
chromatogram is shown in Figure 2. The peaks for SS
and SH2 were assigned by comparison with chromato-
grams of each compound. The peaks labeled MD1 and
MD2 were assigned previously by 1H NMR to the specific

Thiol/Disulfide Exchange Reactions of Somatostatin
J . Org. Chem., Vol. 61, No. 21, 1996 7393
Ta ble 1. Equ ilibr iu m Con sta n ts for Th iol/Disu lfid e
Exch a n ge Rea ction s of Som a tosta tin , Ar gin in e
Va sop r essin , a n d Oxytocin w ith Glu ta th ion e
arginine
somatostatina-c
vasopressind
oxytocind
140
K1a, M-1
71 ( 9
50 ( 6
K
1b, M-1
K1, M-1
K2a
121 ( 13
0.35 ( 0.03
0.49 ( 0.06
0.20 ( 0.02
24.7 ( 3.5
0.71 ( 0.06
7 ( 2
60
K2b
K2
0.24
14.5
0.26
36
K
ov, M-1
K3
K4
E°′, V
-0.221
-0.228
-0.216
a
b
25 °C and 0.075 M pH 7.00 phosphate buffer. Reactant
concentrations covered the following ranges: 8-30 µM for the total
somatostatin concentration; 4-18 mM for the GSH concentration;
and 0.4-2 mM for the GSSG concentration. c The average relative
F igu r e 2. Chromatogram of a somatostatin/GSH reaction
mixture at equilibrium (25 °C, pH 7.0 phosphate buffer). The
peak labels are defined in the text. The concentrations
determined from the peak areas are 5.73 µM SS, 14.9 µM SH2,
3.18 µM MD1, and 5.73 µM MD2. The concentration of the
internal standard FFF was 15.0 µM. The mobile phase
contained pH 3.0 phosphate buffer (0.1 M) and 26% acetoni-
trile.
d
standard deviation of the equilibrium constants is 13%. 25 °C
and pH ) 7.00; 0.15 M KCl.7
somatostatin-glutathione mixed disulfides10
Ala–Gly–Cys–Lys–Asn–Phe–Phe–Trp–Lys–Thr–Phe–Thr–Ser–Cys
γ–Glu–Cys–Gly
(MD1)
Ala–Gly–Cys–Lys–Asn–Phe–Phe–Trp–Lys–Thr–Phe–Thr–Ser–Cys
γ–Glu–Cys–Gly
(MD2)
and the peak labeled DMD is for the double mixed
disulfide
Ala–Gly–Cys–Lys–Asn–Phe–Phe–Trp–Lys–Thr–Phe–Thr–Ser–Cys
γ–Glu–Cys–Gly
γ–Glu–Cys–Gly
F igu r e 3. Chromatograms of samples taken as a function of
time after raising the pH of an ∼4 µM solution of MD1 and
internal standard FFF to 7.0.
(DMD)
The peak labeled FFF is for (phenylalanyl)(phenylalanyl)-
phenylalanine, which was added as an internal intensity
standard.
over a range of reactant concentrations are reported in
Table 1. Also listed for comparison in Table 1 are
literature values for the analogous equilibrium constants
for the OT/GSH and AVP/GSH systems.7
To establish that the reactions are reversible, equilib-
rium was approached from both directions.11 After a
solution prepared by combining stock solutions of GSH,
GSSG, and somatostatin had reached equilibrium, as
indicated by HPLC analysis, the equilibrium was shifted
twice, first by addition of additional GSH and then by
addition of additional GSSG. After each addition, ali-
quots of the resulting solutions were analyzed as a
function of time until there was no further change in
concentration. Equilibrium constants calculated using
concentrations from the three separate equilibrium con-
ditions were equal within experimental error. For ex-
ample, the three values determined for K1 from such an
experiment were 121, 128, and 129 M-1 and the values
determined for K2 were 20.6, 20.8, and 22.4 M-1. These
results verify that the thiol/disulfide exchange reactions
are reversible under the conditions used and that equi-
librium was reached. The average values obtained for
the equilibrium constants from 20 separate experiments
Kin etics of Th iol/Disu lfid e Exch a n ge Rea ction s.
Rate constants k-1a and k3 and rate constants k-1b and
k-3 (Figure 1) were determined directly by monitoring
the conversion of MD2 and MD1, respectively, to SS and
the other mixed disulfide. The procedure involved in-
creasing the pH of solutions of MD1 or MD2 (at concen-
trations in the 1-10 µM range) from pH ∼2.5, where
thiol/disulfide exchange is very slow, to pH 7.0. Aliquots
were then removed as a function of time, quenched and
analyzed by HPLC. Representative chromatograms from
a study of the conversion of MD1 to MD2 and SS are
presented in Figure 3. At time 0, the chromatogram only
shows peaks for MD1 and the internal intensity standard
FFF. In the chromatogram for the aliquot taken at 20
s, peaks are also observed for both MD2 and SS, indicat-
ing that the cyclization reaction to form SS and the
isomerization reaction to form MD2 are taking place at
similar rates. As the reaction proceeds, the SS peak
increases in intensity, and eventually both MD1 and the
MD2 formed from MD1 are converted to SS. Since the
solution only contains MD1 at t ) 0, the cyclization and
(10) Kaerner, A.; Weaver, K. H.; Rabenstein, D. L. Magn. Reson.
Chem. 1996, 34, 587-594.
(11) Yeo, P. L.; Rabenstein, D. L. Anal. Chem. 1993, 65, 3061-3066.

7394 J . Org. Chem., Vol. 61, No. 21, 1996
Rabenstein and Weaver
Ta ble 2. Ra te Con sta n ts for Th iol/Disu lfid e Exch a n ge
Rea ction s of Som a tosta tin , Ar gin in e Va sop r essin , a n d
Oxytocin w ith Glu ta th ion ea
resulting solution of SH2/FFF at pH 7.0, and then
aliquots were removed as a function of time, quenched,
and analyzed by HPLC. Rate constant k-2 was deter-
mined from the initial rate of the reaction. The value
reported in Table 2 is the average from six experiments
in which the concentration of SH2 and GSSG were in the
range of 20-50 µM and 0.3-0.5 mM, respectively. The
value listed for rate constant k2 was calculated from k-2
and K2.
arginine
somatostatina
vasopressinb
oxytocinb
k1a, M-1 s-1
k-1a, s-1
0.39 ( 0.09
0.0055 ( 0.0011
0.49 ( 0.18
k
1b, M-1 s-1
k-1b, s-1
k1, s-1
0.0097 ( 0.0034
0.88 ( 0.20
38
110
k-1, s-1
0.0072 ( 0.0020
0.0030 ( 0.0006
0.0039 ( 0.0009
1.83 ( 0.22
0.63
0.76
k3, s-1
Discu ssion
k-3, s-1
k2, M-1 s-1
k-2, M-1 s-1
0.74
3.1
0.84
3.2
Thiol/disulfide exchange occurs by an SN2 displacement
mechanism in which a thiolate anion, the reactive form
of the thiol in thiol/disulfide exchange reactions, ap-
proaches the disulfide bond along its S-S axis.3,5,14,15 The
equilibrium and rate constants reported in Tables 1 and
2 are for pH 7.00, where only a fraction of the thiol groups
of GSH, MD1, MD2, and the dithiol form of somatostatin
are in the thiolate form, and thus they are conditional
constants for pH 7.00. pH-independent, intrinsic equi-
librium, and rate constants for the reactions in terms of
thiolate species are reported in Table 3. The intrinsic
equilibrium constants, Ki, were calculated using the
values reported in Table 1 by accounting for the fractional
thiolate concentrations at pH 7.00 with equations of the
type Ki1a ) (RMD2/RGSH)K1a where R is the fraction of the
thiol group of the compound indicated in the thiolate form
at pH 7.00. R was calculated to be 0.0116 for GSH,
0.0661 for the Cys3 thiol group of MD1, 0.001 20 for the
Cys14 thiol group of MD2 using pKa values of 8.93 for
GSH, 8.15 for the Cys3 thiol group of MD1, and 9.92 for
the Cys14 thiol group of MD2.12,16,17 The fraction of the
dithiol form of somatostatin with both thiol groups
ionized was calculated to be 7.9 × 10-5 using pKa values
of 8.15 and 9.92 for the Cys3 and Cys14 thiol groups. The
intrinsic rate constants were calculated from the condi-
tional rate constants in Table 2 using relationships of
9.5 ( 0.6
a
b
25 °C, pH 7.00, 0.15 M phosphate buffer. 25 °C, pH 7.00,
0.15 M KCl.7
Ta ble 3. In tr in sic Equ ilibr iu m a n d Ra te Con sta n ts for
Th iol/Disu lfid e Exch a n ge Rea ction s of Som a tosta tin w ith
Glu ta th ion e
equilibrium constants
rate constants
Ki1a, M-1
7.3
290
297
0.036
2.8
0.0355
ki1a, M-1 s-1
33.6
4.58
42.2
0.147
2.50
0.059
Ki1b, M-1
Ki1, M-1
Ki2a
ki-1a, s-1
ki1b, M-1 s-1
ki-1b, s-1
ki3, s-1
Ki2b
Ki2
ki-3, s-1
Kiov, M-1
Ki3
10.5
42.4
interconversion reactions are the only reactions that are
taking place initially, and thus, k-1b and k-3 were
obtained directly from the initial rates of formation of
SS and MD2, respectively. Values determined for k-1a
,
k3 and k-1b, k-3 from multiple experiments with MD2 and
MD1, respectively, are presented in Table 2. Also listed
in Table 2 are values for k1a and k1b, which were
calculated from k-1a and K1a and k-1b and K1b using the
relations k1a ) K1ak-1a and k1b ) K1bk-1b
.
The value listed in Table 2 for rate constant k1 (eq 1)
was calculated from k1a and k1b using the relation k1 )
k1a + k1b. Rate constant k1 was also estimated from the
initial rate of disappearance of SS by reaction with GSH.
The procedure involved addition of GSH to a solution of
SS and FFF, and then aliquots were removed as a
function of time, quenched, and analyzed by HPLC. SS
concentrations in the range 16-34 µM were used. To
approximate pseudo-first-order reaction conditions, GSH
concentrations in the 2-10 mM range were used. Be-
cause the rate of the reaction was found to be too fast at
pH 7.0 to follow accurately by removing and quenching
aliquots as a function of time, the reaction was run at
lower pH where a smaller fraction of the GSH is in the
reactive thiolate form. Rate constant k1 was determined
from the initial rate measured in 12 experiments in the
pH range 5.3-6.2. Each of these rate constants was then
used to estimate a value for k1 at pH 7.00 by accounting
for the increased concentration of the reactive thiolate
form at pH 7.00; a pK of 8.93 was used for the thiol group
of GSH.12 An average value of 2.00 ( 0.55 M-1 s-1 was
estimated for k1 at pH 7.00.
the type ki1a ) k1a/RMD2
.
Kin etic Sta bility of th e Disu lfid e Bon d of Som a -
tosta tin . A main objective of this research was to
characterize the kinetic stability of the disulfide bond of
somatostatin with respect to cleavage by thiol/disulfide
exchange. The results in Table 2 indicate that the
disulfide bond of somatostatin is readily cleaved by a two-
step reaction with GSH and that the rate constants for
its two possible reactions with GSH in the first step (k1a
and k1b in Figure 1) are essentially equal. Also, they are
of a magnitude that is typical of rate constants for
reaction of thiol-containing amino acids and peptides with
intermolecular peptide disulfide bonds, which suggests
that, even though it is part of a ring system, the disulfide
bond of somatostatin is not particularly unstable and it
reacts as a normal disulfide bond. For comparison, rate
constants for the second-order reaction of GSH and
cysteine with the disulfide bond of GSSG at pH 7.0 are
0.41 and 0.33 M-1 s-1, respectively.6 This is in contrast
(14) Hupe, D. M.; Wu, D. J . Org. Chem. 1980, 45, 3100-3103.
(15) Rabenstein, D. L.; Theriault, Y. Can. J . Chem. 1984, 62, 1672-
1680.
Rate constant k-2 (eq 2) was determined by reaction
of reduced somatostatin with GSSG at pH 7.0. The
reduced somatostatin was prepared by electrochemical
reduction of a SS/FFF solution.13 The procedure for
determination of k-2 involved addition of GSSG to the
(16) The pKa’s of the mixed disulfides of somatostatin cannot be
measured due to intramolecular thiol/disulfide exchange. The pKa of
Cys3 was estimated to be 8.15 using the pKa value reported for Cys6 of
arginine vasopressin in D2O.17 The pKa of Cys14 was estimated to be
9.92 using the pKa’s reported for Cys6 of tocinoic acid in D2O.17 The
pKa values for D2O solution were converted to H2O solution using the
relation pKa(H2O) ) pKa(D2O) - 0.50.6
(12) Rabenstein, D. L. J . Am. Chem. Soc. 1973, 95, 2797-2803.
(13) Saetre, R.; Rabenstein, D. L. Anal. Chem. 1978, 50, 276-280.
(17) Nosza´l, B.; Guo, W.; Rabenstein, D. L. J . Org. Chem. 1992, 57,
2327-2334.

Thiol/Disulfide Exchange Reactions of Somatostatin
J . Org. Chem., Vol. 61, No. 21, 1996 7395
to the rate constants measured for reaction of GSH with
the disulfide bonds of arginine vasopressin and oxytocin
(k1, eq 1), which are significantly larger than rate con-
stants for typical thiol/disulfide exchange reactions,7 and
they are larger than rate constant k1 for somatostatin.18
by accounting for the protonation state of the thiol group
at pH 7.0 using the relation k-1 ) Rki-1, where R is the
fraction in the thiolate form.21 The values calculated for
k-1 are 0.06, 0.12, 0.011, 0.011, and 0.0068 s-1 for n ) 1,
2, 3, 4, and 5, respectively. Comparison with the values
listed in Table 2 indicates that k-1 for arginine vaso-
pressin and oxytocin is some 60-70 times larger than
k-1 for the n ) 4 random coil homolog, which also forms
a 20-membered ring. No rate constants are available for
homologs that form a 38-membered ring for comparison
with k-1 for somatostatin. However, k-1 for somatostatin
is essentially the same as k-1 for the n ) 5 homolog,
which forms a 23-membered ring; considering the steady
decrease in k-1 for the series of random coil homologs as
the size of the ring increases, it seems likely that k-1 for
reaction of a random coil mixed disulfide to form a 38-
membered ring would be significantly less than is found
for somatostatin. That is, this analysis suggests that
disulfide bond formation by intramolecular thiol/disulfide
exchange in the somatostatin-glutathione mixed disul-
fides is not completely random, but rather it is directed
to some extent by conformational properties of the mixed
disulfides that place the thiol and the mixed disulfide
groups in close proximity.
With respect to the kinetic stability of somatostatin,
it is important to consider the fate of the mixed disulfides
formed in the first step. The mixed disulfides can react
with another molecule of GSH to give fully reduced
somatostatin in the second step or they can undergo
intramolecular thiol/disulfide exchange to re-form the
disulfide bond of somatostatin. Rate constants k-1 and
k2 (eqs 1 and 2) can be used to determine the relative
tendencies for reaction by these two pathways if k2 is
converted to the pseudo-first-order rate constant k2′ ()
k2[GSH]). Using the results in Table 2 for somatostatin,
it can be shown that k-1 > k2′ when [GSH] < 0.004 M;
i.e., intramolecular thiol/disulfide exchange to re-form the
disulfide bond of somatostatin is faster than reaction with
another molecule of GSH when [GSH] < 0.004 M. This
suggests that in human blood plasma, where the con-
centration of GSH and other nonprotein thiols is much
less than 0.004 M,9,19 mixed disulfides of somatostatin
will tend to undergo intramolecular thiol/disulfide ex-
change to re-form the native disulfide bond rather than
react with another molecule of thiol to give the reduced,
dithiol form of somatostatin. That is, even though the
disulfide bond of somatostatin is susceptible to cleavage
by thiol/disulfide exchange, biologically active somatosta-
tin is readily re-formed by intramolecular thiol/disulfide
exchange. This is an important concept to consider when
investigating the behavior of other disulfide-containing
peptide hormones in biological systems and the use of
disulfide bonds as structural elements in peptide and
peptidomimetic drugs.
That some fraction of the somatostatin-glutathione
mixed disulfides exist in precyclic conformations is sup-
ported by evidence from 1H NMR studies of the two
somatostatin-glutathione mixed disulfides.10 1H NMR
results, as well as results from semiempirical energy
calculations and time-resolved fluorescence spectroscopy
experiments, indicate that the cyclic disulfide form of
somatostatin exists as an equilibrium of several rapidly
interconverting, low-energy conformations, including con-
formations that have âII turns over Trp8-Lys9 and Thr10
-
Phe11.10,22-27 Nuclear Overhauser enhancement (NOE)
data and the temperature coefficients of selected NH
chemical shifts, together with chemical shift data for CRH
and NH protons, indicate that the mixed disulfides are
interconverting between multiple conformations, some of
which have secondary structure, and that some elements
of the secondary structure are similar to those of the
cyclic disulfide form of somatostatin.10
A similar analysis using the rate constants in Table 2
for arginine vasopressin and oxytocin indicates that the
tendency for formation of their disulfide bonds by in-
tramolecular thiol/disulfide exchange is even greater.7
Specifically, k-1 > k2′ when [GSH] < 0.85 and 0.90 M for
arginine vasopressin and oxytocin, respectively. This
even greater tendency for intramolecular thiol/disulfide
exchange is most likely because smaller rings are formed
by the disulfide bond in arginine vasopressin and oxy-
tocin (20-membered as compared to the 38-membered
ring of somatostatin). Nevertheless, the rate constant
k-1 for somatostatin is surprisingly large. This can be
seen, for example, by comparison of k-1 to rate constants
reported for the analogous reactions of a homologous
series of peptides of the type Cys-(Ala)n-Cys, where n
varies from 1 to 5, which corresponds to ring sizes of 11,
14, 17, 20, and 23 atoms for the disulfide form of the
peptides.20 The rate constants for the homologous series
were measured under conditions where the mixed dis-
ulfides are random coil, and they were reported as
intrinsic rate constants, ki-1, i.e., rate constants for
conditions where the reacting thiol groups are completely
in the deprotonated thiolate form. For purposes of
comparison with the values reported in Table 2, we have
converted them to conditional rate constants at pH 7.0
Kin etics of In tr a m olecu la r Th iol/Disu lfid e Ex-
ch a n ge Rea ction s of Som a tosta tin -Glu ta th ion e
Mixed Disu lfid es. The rate constants in Table 2 are
for pH 7.00, where only a small fraction of the various
thiol groups are in the reactive thiolate form. The pKa
values of 8.15 for the Cys3 thiol group of MD1 and 9.92
for the Cys14 thiol group of MD216 indicate that a much
smaller fraction of MD2 will be in the thiolate form at
pH 7.0, and thus, k-1a is predicted to be much less than
k-1b. However, rate constants for thiol/disulfide exchange
also depend on the nucleophilicity of the thiolate anion,
(21) A pKa of 8.9 was used to calculate R because this value was
used to convert the measured rate constants to intrinsic rate con-
stants.20
(22) Hallenga, K.; Van Binst, G.; Scarso, A.; Michel, A.; Knappen-
berg, M.; Dremier, C.; Brison, J .; Dirkx, J . FEBS Lett. 1980, 119, 47-
52.
(23) Knappenberg, M.; Michel, A.; Scarso, A.; Brison, J .; Zanen, J .;
Hallenga, K.; Deschrijver, P.; Van Binst, G. Biochim. Biophys. Acta
1982, 700, 229-246.
(24) J ans, A. W. H.; Hallenga, K.; Van Binst, G.; Michel, A.; Scarso,
A.; Zanen, J . Biochim.. Biophys. Acta 1985, 827, 447-452.
(25) Van Den Berg, E. M. M.; J ans, A. W. H.; Van Binst, G.
Biopolymers 1986, 25, 1895-1908.
(18) The comparison is made in terms of k1 because values have
not been reported for k1a and k1b for arginine vasopressin and oxytocin.
(19) The concentrations of cysteine, cysteinglglycine, and homocys-
teine in human plasma are in the region of 9, 3, and 0.25 µM,
respectively.9b
(20) Zhang, R.; Snyder, G. H. J . Biol. Chem. 1989, 264, 18472-
18479.
(26) Verheyden, P.; DeWolf, E.; J aspers, H.; Van Binst, G. Int. J .
Peptide Protein Res. 1994, 44, 401-409.
(27) Elofsson, A.; Nilsson, L.; Rigler, R. Int. J . Peptide Protein Res.
1990, 36, 297-301.

7396 J . Org. Chem., Vol. 61, No. 21, 1996
Rabenstein and Weaver
as indicated by its Bronsted basicity, and the Bronsted
basicities of the central and leaving group sulfur atoms.3
The dependence on the Bronsted basicity of the thiolate
nucleophile and the central and leaving group sulfurs is
given by eq 12
E°′SS/SH2 ) E°′GSSG/GSH + (RT/ nF) ln Kov (13)
A value of -0.221 V was calculated for E°′SS/SH2 using
the value reported in Table 1 for Kov and E°′GSSG/GSH
)
-0.262 V.29
E°′SS/SH2 is similar to E°′ for the disulfide bonds of AVP
and oxytocin (-0.228 and -0.216 V, respectively), as
expected since Kov is similar for somatostatin, AVP, and
oxytocin. It is also of interest to note that E°′SS/SH2 is also
similar to E°′ values reported for the disulfide bonds of
octapeptide models for the active sites of the thiol-
protein oxidoreductases thioredoxin, thioredoxin reduc-
tase, glutaredoxin, and protein disulfide isomerase, even
though the model peptides form smaller (14-membered)
disulfide-containing rings. The redox potentials of the
model peptides, all of which are of the general sequence
Ac-X-Cys-X-X-Cys-X-X-X-NH2, are -0.212, -0.232, -0.237,
and -0.227 V, respectively.30 It will be of interest to
determine the effect of the kinetics of the intramolecular
thiol/disulfide exchange reactions on the equilibrium
constants and redox potentials for the model peptides,
as was done above for somatostatin, AVP, and oxytocin.
Finally, E°′ values for a series of nonpeptide molecules
that form 5-11 membered disulfide-containing rings
range from -0.354 V to -0.240 V.31 For example, the
E°′ value for dithiothreitol (DTT), lipoic acid, and 6,6′-
sucrosedithiol, which form 6-, 5-, and 11-membered rings,
are -0.327, -0.288, and -0.245 V, respectively. E°′ for
somatostatin is less reducing; however, it is larger than
might be expected, considering that formation of the
disulfide bond of somatostatin involves closure of a 38-
membered ring.
lg
log ki ) C + ânucpKanuc + âcpKac + âlgpKa (12)
where ki is the intrinsic rate constant.3 The constant C
and the Bronsted coefficients ânuc, âc, and âlg have been
estimated to be 7.0, 0.50, -0.27, and -0.73, respectively,
using rate data for intermolecular thiol/disulfide ex-
change reactions.3 While these values for C and the
Bronsted coeficients are not directly applicable to rate
constants for intramolecular thiol/disulfide exchange,
they can be used to account qualitatively for some
features of the rate constants in Tables 2 and 3. For
example, ki-1a is predicted by eq 12 to be significantly
larger than ki-1b, because the Bronsted basicity of the
thiolate anion of MD2 is larger and that of the central
sulfur is smaller in the reaction described by ki-1a
.
Likewise, ki3 is predicted to be much larger than ki-3, as
observed, because the Bronsted basicity of the thiolate
nucleophile is larger and that of the leaving group is
smaller in the reaction described by ki3. Thus, the
smaller fraction of Cys14 in the thiolate form is compen-
sated for by the greater nucleophilicity of the Cys14
thiolate anion and the dependence of thiol/disulfide
exchange rate constants on the Bronsted basicity of the
central and leaving group sulfur atoms, with the result
that k-1a and k-1b, and k3 and k-3, are similar in
magnitude.
Th iol/Disu lfid e Exch a n ge Equ ilibr ia a n d th e Re-
d ox P oten tia l of th e Som a tosta tin Disu lfid e Bon d .
Even though the disulfide bond in somatostatin is part
of a much larger ring system, and the rate constants for
its thiol/disulfide exchange reactions with GSH are
significantly different from those of arginine vasopressin
and oxytocin, the stepwise and overall equilibrium con-
stants for its thiol/disulfide exchange reactions with GSH
are essentially the same as those for arginine vasopressin
and oxytocin at pH 7.00. The equilibrium constants are
ratios of rate constants, i.e., K1 ) k1/k-1, K2 ) k2/k-2, and
Kov ) (k1k2)/(k-1k-2). Rate constant k1 for reaction of GSH
with somatostatin is significantly less than for reaction
with AVP and oxytocin, but k-1 is also much less with
the apparently fortuitous result that their ratio, and thus
K1, is essentially the same for somatostatin, AVP, and
oxytocin. The rate constant for the second step in the
overall reduction reaction, k2, is the same within a factor
Exp er im en ta l Section
Ch em ica ls. Somatostatin was obtained from Bachem, Inc.,
Torrance, CA. Glutathione, glutathione disulfide, dithiothrei-
tol (DTT), and FFF were supplied by Sigma Chemical Co.
HPLC-grade acetonitrile and reagent grade phosphoric acid
were purchased from Fisher Scientific Co. The peptide content
of the GSH, GSSG, and somatostatin was determined by 500
MHz 1H NMR using an internal intensity standard. The GSH
was found to be 99.7% GSH and 0.3% GSSG. The GSSG was
found to be 95% GSSG; the remainder is assumed to be water
because no extra resonances were observed in the 1H NMR
spectrum. The somatostatin was determined to be 68.7%
somatostatin, with the balance being trifluoroacetate and
water.
Reduced somatostatin was prepared either by reduction of
somatostatin with DTT or by electrochemical reduction under
nitrogen in a glove bag.13 Somatostatin was dissolved in 0.15
M KCl to give concentrations of 20-50 µM, FFF stock solution
was added, and the solution was then reduced at -1.0 V using
a three-electrode system (a Hg pool working electrode, an Ag/
AgCl reference electrode in saturated KCl and a Pt auxiliary
electrode in 0.15 M KCl). After 25-30 min, 90-95% of the
somatostatin was reduced as determined by HPLC.
of 2-3 for the three compounds, as is rate constant k-2
,
as might be expected since both k2 and k-2 are for thiol/
disulfide exchange reactions of acyclic compounds.6 Thus,
K2 and Kov are also similar for somatostatin, AVP, and
oxytocin. It also is of interest to note that the equilibrium
constants are similar in magnitude to those for the
reaction of two one-disulfide analogs of apamin, in each
of which two of the four cysteines of apamin are replaced
by serine.28 The two cysteines in the two analogs are
separated by either 9 or 11 amino acids, as compared to
10 for somatostatin.
Syn th esis a n d Isola tion of Mixed Disu lfid es. The two
somatostatin-glutathione mixed disulfides were synthesized
by reacting somatostatin with a mixture of GSH and GSSG.
Typical concentrations were 0.5-2 mM GSSG, 1-8 mM GSH,
and 100-400 µM somatostatin. Mixtures were allowed to
react for ∼5 min at pH 7.0 and then were quenched by
lowering the pH to 2.5. The best yields of mixed disulfide were
obtained with GSH:GSSG ratios of 5:1 to 10:1.
The equilibrium constant for the overall reaction of
somatostatin with GSH is related to the half-cell poten-
tial for its disulfide bond:
(29) Millis, K. K.; Weaver, K. H.; Rabenstein, D. L. J . Org. Chem.
1993, 58, 4144-4146.
(30) Siedler, F.; Rudolph-Bo¨hner, S.; Doi, M.; Musiol, H.-J .; Moroder,
L. Biochemistry 1993, 32, 7488-7495.
(28) Huyghues-Despointes, B. M. P.; Nelson, J . W. Biochemistry
1992, 31, 1476-1483.
(31) Lees, W. J .; Whitesides, G. M. J . Org. Chem. 1993, 58, 642-
647.

Thiol/Disulfide Exchange Reactions of Somatostatin
J . Org. Chem., Vol. 61, No. 21, 1996 7397
The mixed disulfides were isolated by reversed-phase HPLC
on a 10 mm × 250 mm C18 reversed-phase column. Because
the mixed disulfides readily undergo intramolecular thiol/
disulfide exchange reactions at neutral pH, they were isolated
at pH 2.5 with a mobile phase containing 0.1% trifluoroacetic
acid in acetonitrile (29%) and water. Because less than 20%
of the somatostatin was converted to mixed disulfide in the
reaction, the native disulfide and reduced dithiol forms of
somatostatin were also collected and reacted again with a
mixture of GSH and GSSG to form more mixed disulfide. The
isolated mixed disulfides were lyophilized, rechromatographed
to increase purity, and then stored in dry ice until used.
Kin etic a n d Equ ilibr iu m Stu d ies. Reaction mixtures
were analyzed by HPLC on a 3.2 mm × 100 mm C18 reversed-
phase column (particle size 3 µm) and a 15 mm × 3.2 mm
guard column. The detector was set at 215 nm. Chromato-
graphic conditions were optimized by varying the pH, the
percent acetonitrile, and the concentration of phosphate in the
mobile phase. A pH 3.0 mobile phase containing 26% aceto-
nitrile and 0.1 M phosphate buffer was found to give good
separation of the reduced, dithiol, and mixed disulfide forms
of somatostatin and the internal standard FFF. The mobile
phase was prepared by adding phosphoric acid and acetonitrile
to water that had been purified with a Millipore water
purification system and adjusting the pH with sodium hy-
droxide. The mobile phase was filtered through a 0.5 µm filter
to remove any particulates. The mobile phase was sparged
daily with helium to remove dissolved gas to prevent postcol-
umn degassing, which caused a noisy base line.
The response of the detector was calibrated by measuring
the ratio of analyte to internal standard peak areas vs the
concentration of analyte for somatostatin and reduced soma-
tostatin.11 Somatostatin and reduced somatostatin concentra-
tions ranged from 0.1 to 100 µM and 0.1 to 50 µM, respectively.
The concentration of FFF was 19-21 µM. The reduced
somatostatin solutions used to calibrate detector response were
prepared by reducing somatostatin with DTT. Detector re-
sponse was found to be linear over the above concentration
ranges for somatostatin and reduced somatostatin. The limit
of detection for somatostatin was determined to be ∼18 nM
at a signal-to-noise ratio of 3, which corresponds to ∼0.36 pmol.
Detector calibration factors were determined for each mixed
disulfide by measuring a chromatogram for an aliquot of a
solution of the mixed disulfide. The pH of the remaining
solution was then increased to allow the mixed disulfide to
convert to the native disulfide form by intramolecular thiol/
disulfide exchange, and a chromatogram was measured for an
aliquot of the solution. The calibration factor for the mixed
disulfide was determined from the relative peak areas for the
cyclic and mixed disulfide forms and the calibration factor for
the cyclic disulfide.
All solutions used in the kinetic and equilibrium studies
were deoxygenated by bubbling with oxygen-scrubbed argon
through a glass dispersion tube. To exclude oxygen, experi-
ments were conducted in a nitrogen-filled glovebag. Samples
were then removed from the glovebag and analyzed by HPLC.
The general procedure used in the kinetic experiments has
been described previously.7,11 To summarize the procedure for
measurement of k1, aliquots of GSH, somatostatin, and FFF
standard solutions in pH 5.6-5.7 buffer (0.0375 M phosphate,
0.0375 M acetate) were combined. Samples were then re-
moved as a function of time, quenched by lowering the pH to
2.5, and analyzed by HPLC. The procedure for determining
k-2 was similar and involved mixing aliquots of standard
solutions of GSSG, reduced somatostatin, and FFF in 0.15 M
pH 7.0 buffer.
Equilibrium constants for the thiol/disulfide exchange reac-
tions were determined at pH 7.0 in 0.075 M phosphate buffer
by measuring the concentrations of the native disulfide,
reduced dithiol, and mixed disulfide forms of somatostatin in
solutions containing known, excess concentrations of GSH and
GSSG, i.e., in solutions containing a GSH/GSSG redox buffer.11
The procedure involved combining appropriate volumes of
standard solutions; the reaction mixture was then allowed to
react for 1 h, after which time an aliquot was removed,
quenched, by lowering the pH with HCl to minimize air
oxidation after removal from the glove bag, and analyzed by
HPLC. After an additional 20 min, another aliquot was
removed, quenched, and analyzed by HPLC. This procedure
was repeated two more times to ensure that equilibrium was
achieved, as indicated by no further change in concentration
with time. The equilibrium concentrations of somatostatin,
reduced somatostatin, and the two somatostatin-glutathione
mixed disulfides were determined from the HPLC analysis,
while the equilibrium concentrations of GSH and GSSG were
calculated from their initial concentrations and the equilibrium
concentrations of the various forms of somatostatin.
Kinetics experiments were conducted by combining stock
solutions of the reactants, which were equilibrated at 25 °C
before combining and were maintained at 25 °C throughout
the experiment. Aliquots were removed as a function of time
and quenched by lowering the pH to 2.5. The quenched
solutions were then analyed by HPLC.
Ack n ow led gm en t. This research was supported in
part by National Institutes of Health Grant No. GM
37000.
J O960917+