Selenium and sulfur in exchange reactions: A comparative study

Article abstract of DOI:10.1021/jo1011569

Cysteamine reduces selenocystamine to form hemiselenocystamine and then cystamine. The rate constants are k₁ = 1.3 × 10⁵ M^-1 s^-1; k_-1 = 2.6 × 10⁷ M^-1 s^-1; k₂ = 11 M^-1 s^-1; and k_-2 = 1.4 × 10³ M^-1 s^-1, respectively. Rate constants for reactions of cysteine/selenocystine are similar. Reaction rates of selenium as a nucle'phile and as an electrophile are 2-3 and 4 orders of magnitude higher, respectively, than those of sulfur. Sulfides and selenides are comparable as leaving groups.

Full text of DOI:10.1021/jo1011569

pubs.acs.org/joc
available oxidation states. An important difference, though,
4
,5
Selenium and Sulfur in Exchange Reactions:
A Comparative Study
4
is that the pK of Sec is lower (5.3) than that of Cys (8.3), which
6
a
makes Sec a considerably more potent nucleophile under neutral
4
and acidic conditions. Furthermore, selenium is softer than
†
Daniel Steinmann, Thomas Nauser, and
Willem H. Koppenol*
˚
sulfur, with a polarizability volume of 3.8 A (Se) compared to
5
.9 A (S). The two elements have similar functions, and, for most
˚
2
selenoenzymes, there exist Cys-containing homologues. During
7
Laboratory of Inorganic Chemistry, Department of
Chemistry and Applied Biosciences, ETH Zurich, 8093
Z u€ rich, Switzerland. Present address: Department of
the catalytic cycles of selenoenzymes, formation and cleavage of
8-11
†
selenylsulfides via exchange reactions are common.
How-
Pharmaceutical Chemistry, University of Kansas,
095 Constant Avenue, Lawrence, KS 66047.
ever, in contrast to thiol-disulfide exchange reactions, only a few
studies of analogous reactions involving selenium have been
published. It has been shown by NMR that the selenol-
diselenide exchange reaction of selenocysteamine (SeCya) with
2
12-17
koppenol@inorg.chem.ethz.ch
Received June 14, 2010
selenocystamine (SeCya ) is 7 orders of magnitude faster at
ox
neutral pH than the analogous exchange reaction of cysteamine
18
(Cya) with cystamine (Cya ) and that selenols catalyze the
latter reaction.
ox
19
Recently, the authors of a computational study predicted
that nucleophilic attack by thiols at selenium is both kineti-
cally faster and thermodynamically more favorable than at
2
0
sulfur.
Herein, we compare sulfur and selenium as nucleophiles,
electrophiles, and leaving groups in thiol-disulfide-like ex-
change reactions to better understand differences in reactiv-
ity between selenium and sulfur in biological systems. We use
a published value for the electrode potential of the thiol/
Cysteamine reduces selenocystamine to form hemiseleno-
cystamine and then cystamine. The rate constants are
0
21
disulfide couple of dithiothreitol (DTT), E° (DTT /DTT),
ox
5
-1 -1
7
-1 -1
as the linchpin to calculate critical equilibrium and rate
constants.
k = 1.3 ꢀ 10 M
s
; k = 2.6 ꢀ 10 M
s ; k = 11
2
s , respectively. Rate
1
-1
-
1
-1
3
-1 -1
M
s
; and k = 1.4 ꢀ 10 M
-
2
We compare the kinetics of reduction of Secox by Cys and
of reduction of SeCyaox by Cya (Scheme 1), measured by
stopped-flow spectrophotometry, from which we conclude
that selenium promotes both nucleophilic and electrophilic
exchange reactions, relative to sulfur.
The reduction of the diselenides by the thiols proceeds via
a two-step process (Scheme 1) at pH 7 (Figure 1); the first
phase of the reaction, the formation of the hemiselenocystine
constants for reactions of cysteine/selenocystine are simi-
lar. Reaction rates of selenium as a nucleophile and as an
electrophile are 2-3 and 4 orders of magnitude higher,
respectively, than those of sulfur. Sulfides and selenides
are comparable as leaving groups.
According to recent proteomics reports, selenium is pre-
sent in at least 25 human protein families in the form of the
(
6) Perrin, D. D.; Sayce, I. G. J. Chem. Soc. A 1968, 53.
1
2
(7) Fomenko, D. E.; Xing, W.; Adair, B. M.; Thomas, D. J.; Gladyshev,
V. N. Science 2007, 315, 387–389.
(8) Aumann, K. D.; Bedorf, N.; Brigelius-Floh ꢀe , R.; Schomburg, D.;
Floh ꢀe , L. Biomed. Environ. Sci. 1997, 10, 136–155.
amino acid selenocysteine (Sec) . Although the functions of
only a few selenoproteins have been well characterized to
date, many can be classified as redox enzymes in which Sec is
(
(
9) Kim, H. Y.; Gladyshev, V. N. PLoS Biol. 2005, 3, 2080–2089.
10) Zhong, L.; Arn ꢀe r, E. S. J.; Holmgren, A. Proc. Natl. Acad. Sci. U.S.
3
part of the catalytic center. Sec and Cys share many proper-
ties, as there are only minor differences between sulfur and
selenium in terms of electronegativity, ionic radius, and
A. 2000, 97, 5854–5859.
11) Cheng, Q.; Sandalova, T.; Lindqvist, Y.; Arn ꢀe r, E. S. J. J. Biol. Chem.
009, 284, 3998–4008.
12) Dickson, R. C.; Tappel, A. L. Arch. Biochem. Biophys. 1969, 130,
547–550.
(13) Kice, J. L.; Slebocka-Tilk, H. J. Am. Chem. Soc. 1982, 104, 7123–
7130.
(14) Singh, R.; Kats, L. Anal. Biochem. 1995, 232, 86–91.
(
2
(
(
1) Abbreviations, systematic nomenclature (trivial names): DTT, dithio-
threitol; DTTox, trans-1,2-dithiane-4,5-diol (oxidized dithiothreitol); Sec,
selenocysteine; Secox, selenocystine; Cysox, cystine; SeCya, selenocysteamine;
SeCyaox, selenocystamine; Cya, cysteamine; Cyaox, cystamine; hSeCysox
,
hemiselenocystine; hSeCyaox, hemiselenocystamine; GPx, glutathione per-
oxidases; bis-tris, 2-[bis(2-hydroxyethyl)amino]-2-(hydroxymethyl)propane-
(15) Metanis, N.; Keinan, E.; Dawson, P. E. J. Am. Chem. Soc. 2006, 128,
16684–16691.
1
3
,3-diol; hepes, 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid; mops,
-(N-morpholino)propanesulfonic acid.
(16) Beld, J.; Woycechowsky, K. J.; Hilvert, D. Biochemistry 2007, 46,
5382–5390.
(
2) Kryukov, G. V.; Castellano, S.; Novoselov, S. V.; Lobanov, A. V.;
Zehtab, O.; Guig oꢀ , R.; Gladyshev, V. N. Science 2003, 300, 1439–1443.
3) Gromer, S.; Eubel, J. K.; Lee, B. L.; Jacob, J. Cell. Mol. Life Sci. 2005,
2, 2414–2437.
4) Huber, R. E.; Criddle, R. S. Arch. Biochem. Biophys. 1967, 122, 164–
73.
5) CRC Handbook of Chemistry and Physics; 87th ed.; CRC Press: Boca
Raton, 2006.
(17) Figueroa, J. S.; Yurkerwich, K.; Melnick, J.; Buccella, D.; Parkin, G.
Inorg. Chem. 2007, 46, 9234–9244.
(18) Pleasants, J. C.; Guo, W; Rabenstein, D. L. J. Am. Chem. Soc. 1989,
111, 6553–6558.
(19) Singh, R.; Whitesides, G. M. J. Org. Chem. 1991, 56, 6931–6933.
(20) Bachrach, S. M.; Demoin, D. W.; Luk, M.; Miller, J. V. J. Phys.
Chem. A 2004, 108, 4040–4046.
(
6
(
1
(
(21) Lees, W. J.; Whitesides, G. M. J. Org. Chem. 1993, 58, 642–647.
6
696 J. Org. Chem. 2010, 75, 6696–6699
Published on Web 08/31/2010
DOI: 10.1021/jo1011569
r 2010 American Chemical Society

Steinmann et al.
JOCNote
SCHEME 1
(hSeCys )/hemiselenocystamine (hSeCya ), is complete in
ox ox
ca. 60 ms, while the latter phase continues for ca. 20 s
Figure 1, inset) in our experiments. Rate constants for the
FIGURE 2. Reversed-phase HPLC analysis of Cyaox (RS-SR),
SeCyaox (RSe-SeR), and hSeCyaox (RSe-SR) obtained from
(
reactions can be derived from the rates of absorbance
change, and from the absorbance of the selenolate produced
after 60 ms and 20 s, the equilibrium constants K and K ,
reaction of 10 mM Cya and 2 mM SeCya with O
5
2
; Uptisphere
CN mobile
μm HDOC18 (250 mm ꢀ10 mm) column; H
2
O/CH
3
phase with 0.1% TFA; 30 °C; detection at 300 nm; molar absorp-
tivities for quatification of products obtained from UV-vis spectra
1
2
respectively, were calculated.
(
insets).
0
(
2
(
DTT ), and 0.05 mM SeCya before mixing. E° (SeCya /
ox
ox
ox
2
SeCya) = -368 ( 4 mV, from Keq ([DTT ][SeCya] )/
ox
[DTT][SeCya ]) = 0.04 ( 0.01 M and the published value
ox
0
21 19
for E° (DTT /DTT) = -327 mV. Singh et al. used a
ox
2
similar approach with K ([DTT ][SeCya] )/([DTT]-
eq
ox
[
SeCya ]) = 0.14 M to obtain a value of -352 mV.
ox
Upon rapid mixing of Cya and SeCya , we find K =
1
ox
-
3
0
(
5 ( 1) ꢀ 10 , and we calculated ΔE ° = -69 ( 3 mV.
0 0
ox 1
1
From E° (SeCya /2SeCya) = -368 ( 4 mV and ΔE ° =
-
5
0
69 ( 3 mV, we obtain E° (hSeCya /Cya,SeCya) = -299 (
ox
0
mV. We used this E° (hSeCya /Cya,SeCya) and subtracted
ox
0
16
0
2
E° (Cya /2Cya) to calculate ΔE ° = -63 ( 5 mV, from
ox
FIGURE 1. Formation of Sec, monitored at 250 nm, upon rapid
-3
-5
which we calculated K = (8 ( 4) ꢀ 10 . From the direct
2
mixing of Cys and Secox. At low concentrations of Cys (2 ꢀ 10 M)
-5
observation of equilibrium 2 upon mixing Cya with SeCya ,
ox
and Secox (5 ꢀ 10 M), the first equilibrium, Cys þ Secox =
hSeCysox þ Sec, is attained within 60 ms. Inset: At 10 mM Cys and
the determination of K and, thus, of ΔE ° is difficult because,
2
2
-5
5
Sec, is reached within 20 s.
ꢀ 10 Secox, the second equilibrium, hSeCysox þ Cys = Cysox
þ
during the 20 - 30 s that elapse before equilibrium is reached,
oxidation of the selenide by O could corrupt the absorbance
2
measurement at the end of the experiment. Further, the
presence of even very small amounts of Cyaox in the Cya shifts
the equilibrium and leads to incorrect results, and we indeed
To determine the equilibrium constants for the reactions
in Scheme 1, we needed to know the molar absorptivities of
Sec and SeCya. The selenolates are the only species that
absorb significantly at 250 nm at pH 7, and absorbance
changes caused by other species were neglected. The extinc-
tion coefficients of the selenolates were measured after
anaerobic reduction of the diselenides with sodium boro-
hydride in the stopped-flow spectrophotometer. We per-
formed the reaction at pH 10 to prevent evolution of H2
during the spectroscopic measurements. We multiplied the
obtained irreproducible values for ΔE ° from measurements of
2
K upon mixing Cya with SeCya .
0
2
ox
0
2
From the values of ΔE ° and ΔE ° , we can predict the
1
equilibrium distribution of diselenides, selenylsulfides and
disulfides in solution mixtures, e.g.:
Cyaox þ SeCyaox ¼ 2hSeCyaox
ð3Þ
-
1
molar absorptivity obtained at pH 10 (ε = 6450 ( 50 M
We mixed 10 mM Cya with 2 mM SeCya, oxidized them in
air at pH 7, and analyzed the products by HPLC (Figure 2) to
measure the equilibrium concentrations of Cya , SeCya ,
2
50
-
1
cm for both Sec and SeCya) by the ratio of selenolates
2
present at neutral pH (96-99%) to calculate the ε250 at
2
ox
ox
-
1
-1
0
pH 7 (6300 ( 200 M cm ).
and hSeCya ; from K = 0.30 ( 0.09, we calculate ΔE ° =
ox
3
3
0
To determine the two-electron electrode potential E° -
SeCya /2SeCya), we mixed SeCya with DTT at pH 7 in
-8 ( 2 mV. This value agrees well with the K = K /K
3
1
2
0
(
measured by stopped-flow spectrophotometry (ΔE ° =
ox
ox
3
0
0
2
a stopped-flow spectrophotometer. Initial concentrations
were typically 0.1 mM DTT, 5 mM oxidized dithiothreitol
(ΔE ° - ΔE ° ) = -6 ( 6 mV).
1
The consecutive equilibrium constants K and K for the
1
2
reduction of Sec by Cys are equivalent, K = K = (4 ( 1) ꢀ
ox
1
2
-
3
0 , which corresponds to electrode potential differences of
1
(22) Arnold, A. P.; Tan, K. S.; Rabenstein, D. L. Inorg. Chem. 1986, 25,
433–2437.
0
0
2
ΔE ° = ΔE ° = -72 ( 4 mV. From this difference and
1
2
J. Org. Chem. Vol. 75, No. 19, 2010 6697

Steinmann et al.
JOCNote
TABLE 1. Rate Constants for Exchange Reactions of Selenols and
Thiols with Disulfides, Selenylsulfides, Diselenides and Hydrogen
Peroxide at pH 7
in an active site must be deprotonated under physiological
2
8,29
conditions to act as a nucleophile,
and thus, only the
smaller difference in nucleophilicity of ca. 1-2 orders of
magnitude is expected for Sec versus Cys in enzymes with
similar functions. In GPx, nucleophilic attack of the thiolate/
selenolate on the peroxide substrate is the rate-limiting step:
indeed, for the Sec-containing pig heart enzyme GPx-4, the
rate constant for the attack at phosphatidylcholine hydroper-
-
1 -1
reaction
rate constant (M
s )
a
Nu
b
c
El
LG
Cya/SeCyaox
Cys/Secox
5
7
5
7
RSH RSe-SeR RShe
RShe RSe-SR RSh
k
k
k
k
k
k
k
k
1
(1.3 ( 0.2) ꢀ 10
-1 (2.6 ( 0.2) ꢀ 10
(1.0 ( 0.2) ꢀ 10
(2.6 ( 0.2) ꢀ 10
7 ( 2
RSH RS-SeR RRShe
2
11 ( 3
3
3
7
-1 -1
6
-1 -1 30,31
RShe RS-SR
RSH RS-SR
RSh
RSh
-2 (1.4 ( 0.3) ꢀ 10
(1.8 ( 0.4) ꢀ 10
oxide is 1 ꢀ 10 M
s
, compared to 1.3 ꢀ 10 M
s
d
4
5
6
7
3.6 ( 0.1
for the corresponding Cys-containing enzyme of Drosophila
melanogaster. Similarly, for the Sec-containing bovine cGPx,
the rate constant for attack at H O is 1 order of magnitude
7
e
RShe RSe-SeR RShe
RSH HO-OH HOh
RShe HO-OH HOh
a
(1.7 ( 0.2) ꢀ 10
f
1.1 ; 2.9
g
c
1.0 ; 2.9
g
2h
9.7 ꢀ 10
2 2
higher than that of the Cys-containing human peroxiredoxin
Although comparisons of enzymes from different org-
Nucleophile, RSH stands for the equilibrium mixture RSH and RSh.
32,33
2.
b
c
Electrophile. Leaving group. pD = 7.43. pD < 7.6. pH = 7.4,
d
18 e
18 f
2
4 g
25 h
26
2
0 °C. pH 7.4, 37 °C. pH = 6.8, 20 °C.
anisms or that have different functions may have limited
validity, these literature values agree reasonably well with
our findings of a 10-fold rate enhancement when selenium acts
as a nucleophile.
0
23
E° (Sec /2Sec) = -383 ( 8 mV, we calculate E° (hSeCys /
0
ox
ox
Cys,Sec) = -311 ( 9 mV. Solutions containing Cys are less
sensitive to oxidation, and we were able to measure K directly
via the absorbance changes in kinetics traces attributed to
Additionally, leaving group protonation by proximate
amino acid residues may influence the rate of reaction. It is
argued that, in a hydrophobic active-site environment, de-
protonated Cys is a worse leaving group; in an environment
where protonation of the leaving group is assisted, differ-
2
0
equilibrium 2, from which we calculate E° (Cys /2Cys) -
ox
0
E° (Sec /2Sec) = -145 mV, in excellent agreement with the
ox
0
0
2
sum ΔE ° þ ΔE ° = -144 ( 9 mV. We were also able to
1
2
-1
34
determine K ([DTT][Sec ])/([DTT ][Sec] ) = 90 M by
ox
ences in reactivity between Sec and Cys may be small.
Thus, from Table 1 it is clear that the rate enhancement of
eq
ox
stopped-flow spectrophotometry, which fits with the electrode
23
0
4
selenium over sulfur as an electrophile (ca. 10 ) is 2 orders of
potential of E° (Sec /2Sec) = -383 ( 8 mV. We can derive
ox
mechanistic information regarding the relative effectiveness of
selenium versus sulfur as an electrophile, nucleophile, or
leaving group by comparing the reactions and corresponding
rate constants in Table 1.
First, we compare reactions of each nucleophile with a
given electrophile and leaving group (e.g., reactions 1 vs 5 or
magnitude greater than that of the nucleophilic enhance-
ment; selenium as a leaving group is approximately equiva-
lent to sulfur.
The thermodynamics as well as the kinetics of an enzyme
reaction are altered as a function of whether selenium
replaces sulfur in an active site. The electrode potentials of
selenylsulfides are ca. 70 mV lower than the potentials of
corresponding disulfides according to our study and that of
-
2 vs 4): the rate constants for nucleophilic reactions of
SeCya are 2-3 orders of magnitude higher than those of
1
4
15
Cya, as reported by Singh and Kats. Similar enhancement
is noted for nucleophilic attack on H O (reactions 6 vs 7),
Metanis et al.; thus, the selenylsulfides are considerably
9
more difficult to reduce. This difference in electrode poten-
2
2
reactions that are catalyzed by Sec- or Cys-containing gluta-
thione peroxidases (GPx) in vivo. The thiolate, not the thiol,
is the reactive nucleophile, and, thus, the difference in reac-
tivity reduces to 1-2 orders of magnitude in a basic environ-
ment where both selenols and thiols are deprotonated.
Next, we compare reactions of each electrophile with a
given nucleophile (reactions 1 vs 2 and -1 vs -2): the
reactions of selenium as an electrophile are 4 orders of
magnitude faster than those of sulfur. Similarly, nucleophilic
tials may be used advantageously to tune enzyme equilibria
in different compartments.
Experimental Section
All chemicals were of highest commercially available quality
and were, aside from Cya, used as received. Millipore Milli-Q
water (18.2 MΩ resistance) was used in all experiments.
Because small amounts of disulfide impurities can strongly
influence the redox equilibria under study, thiols were analyzed
before experiments for the presence of disulfides by HPLC. Cya
attack of cyanide on PhSeSO Ar is 5 orders of magnitude
2
2
7
faster than on PhSSO Ar.
2
contained 1.5% Cyaox and was, therefore, reduced with NaBH
4
-
5
Finally, we note that rate constants for selenolate and
thiolate as the leaving group at neutral pH are comparable
(0.01-0.1 M) and trace amounts of Secox (< 10 M) as a
catalyst. DTT contained 0.5% DTTox, which was taken into
account in the equilibrium calculations, and Cys contained only
negligible levels of impurities.
(reactions 5 vs -1 and 2 vs 4).
In enzymes, e.g., GPx, the microenvironment of the active
site plays a critical role and has to be taken into account when
comparing Sec- and Cys-containing variants. A Cys residue
(29) Huang, H. H.; Arscott, L. D.; Ballou, D. P.; Williams, C. H.
Biochemistry 2008, 47, 1721–1731.
(
23) Nauser, T.; Dockheer, S.; Kissner, R.; Koppenol, W. H. Biochem-
istry 2006, 45, 6038–6043.
24) Barton, J. P.; Packer, J. E.; Sims, R. J. J. Chem. Soc., Perkin Trans. 2
973, 1547–1549.
25) Winterbourn, C. C.; Metodiewa, D. Free Radical Biol. Med. 1999, 27,
22–328.
(30) Maiorino, M.; Ursini, F.; Bosello, V.; Toppo, S.; Tosatto, S. C. E.;
Mauri, P.; Becker, K.; Roveri, A.; Bulato, C.; Benazzi, L.; De Palma, A.;
Floh ꢀe , L. J. Mol. Biol. 2007, 365, 1033–1046.
(31) Ursini, F.; Maiorino, M.; Gregolin, C. Biochim. Biophys. Acta 1985,
839, 62–70.
(32) Peskin, A. V.; Low, F. M.; Paton, L. N.; Maghzal, G. J.; Hampton,
M. B.; Winterbourn, C. C. J. Biol. Chem. 2007, 282, 11885–11892.
(33) Floh ꢀe , L.; Loschen, G.; Eichele, E.; G u€ nzler, W. A. Hoppe-Seylers
Z. Physiol. Chem. 1972, 353, 987–999.
(
1
(
3
(
(
(
26) Pr u€ tz, W. A. Z. Naturforsch., C: Biosci. 1995, 50, 209–219.
27) Gancarz, R. A.; Kice, J. L. J. Org. Chem. 1981, 46, 4899–4906.
28) Tosatto, S. C. E.; Bosello, V.; Fogolari, F.; Mauri, P.; Roveri, A.;
Toppo, S.; Floh ꢀe , L.; Ursini, F.; Maiorino, M. Antioxid. Redox Signaling
008, 10, 1515–1525.
(34) Eckenroth, B. E.; Rould, M. A.; Hondal, R. J.; Everse, S. J.
Biochemistry 2007, 46, 4694–4705.
2
6
698 J. Org. Chem. Vol. 75, No. 19, 2010

Steinmann et al.
JOCNote
Stopped-flow experiments were carried out with an Applied
Photophysics SX 17 MV stopped flow spectrophotometer at
25 °C. The instrument was flushed with nitrogen; solutions were
degassed with argon and transferred in airtight syringes (Hellma
pump, a HP-1100 diode-array detector and Agilent-1100 degasser.
The stationary phase consisted of a C-18 250 mm ꢀ10 mm column
(Uptisphere 5 μm HDOC18) from Interchim Inc., operated at
30 °C. The mobile phase was composed of mixtures of H O/
2
GmbH, M u€ llheim, Germany) to minimize dioxygen contamina-
tion. O concentrations during experiments are estimated to be
CH CN containing 0.1% trifluoroacetic acid. Detection was
at 300 nm; extinction coefficients used to quatify products
3
2
below 2 μM. Solutions were buffered with 5-10 mM bis-tris,
mops, hepes, or phosphate buffer. Kinetics traces were analyzed
with singular value decomposition and nonlinear regression model-
were obtained by UV-vis spectroscopy.
UV-vis spectroscopy was carried out with a SPECORD 200
or SPECORD 250 spectrophotometer from Analytik Jena in
0.1-2 cm quartz cells from Hellma GmbH.
35
ing by means of the Levenberg-Marquard method.
HPLC analysis was performed with a Hewlett-Packard
Agilent) HPLC, with a HP-1050 autosampler and quaternary
(
Acknowledgment. We thank Dr. Patricia L. Bounds for
discussions and editing of the manuscript.
(
35) Maeder, M.; Zuberb u€ hler, A. D. Anal. Chem. 1990, 62, 2220–2224.
J. Org. Chem. Vol. 75, No. 19, 2010 6699