Mechanism of sulfoxide formation through reaction of sulfur radical cation complexes with superoxide or hydroxide ion in oxygenated aqueous solution

Article abstract of DOI:10.1021/ja962032l

We have characterized and quantified several pathways which transform aliphatic sulfur radical cations into sulfoxides in aqueous solution. Sulfur radical cations were produced photochemically via one-electron photooxidation through triplet 4-carboxybenzophenone. Sulfur radical cations and superoxide yield sulfoxide, confirmed by oxygen product isotope effects and an inhibitory role of superoxide dismutase. On the basis of competition experiments with superoxide dismutase the rate constant for the reaction between dimethylsulfide radical cations and superoxide was derived as (2.3 ± 1.2) x 10¹¹ M^-1 s^-1. A demetalated variant of superoxide dismutase did not inhibit superoxide mediated sulfoxide formation, confirming the importance of an active site of the enzyme for inhibition. The stoichiometry of 2 equiv of sulfoxide per reaction of superoxide with a sulfur radical cation suggests a pathway like the singlet oxygen mediated sulfoxide formation, i.e., via a persulfoxide intermediate formed via (i) direct coupling of superoxide with the sulfur radical cation or (ii) electron transfer followed by addition of the product singlet oxygen to a nonoxidized sulfide. In aqueous solution the persulfoxide may add water to yield a hydroperoxy sulfurane prior to its reaction with a second nonoxidized sulfide. At pH values larger than 9, hydroxide ion starts to compete with superoxide for sulfur radical cations and reacts with the persulfoxide or hydroperoxy sulfurane intermediates, initiating less effective pathways of sulfoxide formation. One pathway involves the formation of hydroxysulfuranyl radicals and their reaction with oxygen, supported by product and solvent isotope effects. Besides superoxide and hydroxide-mediated sulfoxide formation there is an additional route involving methylthiomethylperoxyl radicals. Based on oxygen product isotope effects, the latter appear to transfer oxygen onto the sulfide rather than reacting via electron transfer.

Full text of DOI:10.1021/ja962032l

11014
J. Am. Chem. Soc. 1996, 118, 11014-11025
Mechanism of Sulfoxide Formation through Reaction of Sulfur
Radical Cation Complexes with Superoxide or Hydroxide Ion in
Oxygenated Aqueous Solution
Brian L. Miller,† Todd D. Williams,§ and Christian Scho1neich*,†
Contribution from the Department of Pharmaceutical Chemistry, 2095 Constant AVenue,
UniVersity of Kansas, Lawrence, Kansas 66047, and Mass Spectrometry Laboratory,
UniVersity of Kansas, Lawrence, Kansas 66045
ReceiVed June 17, 1996X
Abstract: We have characterized and quantified several pathways which transform aliphatic sulfur radical cations
into sulfoxides in aqueous solution. Sulfur radical cations were produced photochemically via one-electron
photooxidation through triplet 4-carboxybenzophenone. Sulfur radical cations and superoxide yield sulfoxide,
confirmed by oxygen product isotope effects and an inhibitory role of superoxide dismutase. On the basis of
competition experiments with superoxide dismutase the rate constant for the reaction between dimethylsulfide radical
cations and superoxide was derived as (2.3 ( 1.2) × 1011 M-1 s-1. A demetalated variant of superoxide dismutase
did not inhibit superoxide mediated sulfoxide formation, confirming the importance of an active site of the enzyme
for inhibition. The stoichiometry of 2 equiv of sulfoxide per reaction of superoxide with a sulfur radical cation
suggests a pathway like the singlet oxygen mediated sulfoxide formation, i.e., via a persulfoxide intermediate formed
via (i) direct coupling of superoxide with the sulfur radical cation or (ii) electron transfer followed by addition of the
product singlet oxygen to a nonoxidized sulfide. In aqueous solution the persulfoxide may add water to yield a
hydroperoxy sulfurane prior to its reaction with a second nonoxidized sulfide. At pH values larger than 9, hydroxide
ion starts to compete with superoxide for sulfur radical cations and reacts with the persulfoxide or hydroperoxy
sulfurane intermediates, initiating less effective pathways of sulfoxide formation. One pathway involves the formation
of hydroxysulfuranyl radicals and their reaction with oxygen, supported by product and solvent isotope effects.
Besides superoxide and hydroxide-mediated sulfoxide formation there is an additional route involving methylthio-
methylperoxyl radicals. Based on oxygen product isotope effects, the latter appear to transfer oxygen onto the sulfide
rather than reacting via electron transfer.
Introduction
These sulfuranyl radicals react with molecular oxygen to yield
sulfoxide and superoxide anion. Mere´nyi et al. obtained the
overall rate constant for the reaction of 4 with oxygen as k )
2.0 × 108 M-1 s-1 4 and a relatively stable intermediate 5 can
be postulated by analogy to a similar species from the model
Organic sulfides serve an important function in many
materials such as organic polymers and biological macromol-
ecules. They are highly susceptible to oxidation, and different
pathways have been established depending on the nature of the
oxidizing species. Potential intermediates have been character-
ized such as zwitterionic structures, >(+)S-O-O(-), available
through the addition of singlet oxygen to a thioether.1,2
Recently, we have identified a mechanism by which aliphatic
sulfide radical cations (here from dimethylsulfide, DMS), >S•+
(1a), and in particular the radical cation complexes [>S S<]+
(2a), can form sulfoxide in oxygen-containing aqueous solution.3
sulfide 2-(methylthio)methyl acetate (k6,2-(methythio)methyl acetate
)
1.1 × 108 M-1 s-1).5
CH3 CH3
2a + HO–
HO
S
S
(3)
•
CH3
3
CH3
>S + Ox f Ox•- + >S•+ (1a)
(1)
(2)
2a + HO- h >S + HO-S•< (4)
(4)
(5)
>S
>S + Ox 8 Ox•- + [>S S<]+(2a)
•-
3 + O2 9
8 >S + >SdO + H+ + O2
This mechanism is summarized in reactions 1-7, initiated
by one-electron oxidation of an aliphatic sulfide (reactions 1
and 2), and completed through hydroxide/water attack on 2a
yielding sulfuranyl radicals of structure 3 or 4, respectively.
CH3
S
•
O
4 + O2
HO
O
(6)
CH3
5
* Correspondence: FAX: (913) 864-5736; email: schoneich@
smissman.hbc.ukans.edu.
† Department of Pharmaceutical Chemistry
§ Mass Spectrometry Laboratory.
•-
5
9
8 >SdO + H+ + O2
(7)
X Abstract published in AdVance ACS Abstracts, November 1, 1996.
(1) Jensen, F.; Foote, C. S. J. Am. Chem. Soc. 1988, 110, 2368-2375.
(2) Akasaka, T.; Yabe, A.; Ando, W. J. Am. Chem. Soc. 1987, 109,
8085-8087.
Apart from reactions 3-7, sulfoxide formation could alter-
natively proceed via the reaction of superoxide anion with
2a (reaction 8), in particular under conditions where signifi-
(3) Scho¨neich, Ch.; Aced, A.; Asmus, K.-D. J. Am. Chem. Soc. 1993,
115, 11376-11383.
S0002-7863(96)02032-X CCC: $12.00 © 1996 American Chemical Society

Transformation of Sulfur Radical Cations into Sulfoxides
J. Am. Chem. Soc., Vol. 118, No. 45, 1996 11015
cant concentrations of 2a and superoxide are generated
will protonate, depending on pH (pKa,16 ) 8.211). Depending
on the actual concentration of DMS, the monomeric 1a will
associate with DMS to yield the dimeric species 2a (reaction
17; K17 ) 2.0 × 105 M-1).12
simultaneously.6-8
2a + O2•- 98 [intermediate(s)]
(98sulfoxide + products)
(8)
3CBP + CH3SCH3
3CBP + CH3SCH3
3CBP + CH3SCH3 h [δ-CBP‚‚‚S<δ+] (6)
8 CBP + CH3SCH3
9
8 products
(9)
This process may involve the formation of a zwitterionic
intermediate such as formed during the reaction of singlet
oxygen with sulfides and, thus, show product stoichiometries
comparable to those of singlet oxygen pathways.
9
8 CBP + CH3SCH3
(10)
(11)
(12)
Considering these reactions it appears that sulfur radical
cations or their dimeric forms are potential key intermediates
en route to sulfoxide formation in many chemical and biological
systems. This necessitates a detailed understanding of the
individual mechanisms underlying sulfoxide formation from
sulfur radical cations. In the present study we have quantified
and characterized several of these pathways by subjecting
aliphatic sulfides to one electron-transfer photooxygenation by
triplet 4-carboxybenzophenone (3CBP) in aqueous solution. An
emphasis on aliphatic sulfides is particularly important with
respect to their model character for biological molecules such
as methionine. The existence of reaction 8 was unambigously
proven by performing the experiments in the presence of native
and des-Cu superoxide dismutase, and corresponding rate
constants were derived. Further mechanistic details were
obtained on the basis of solvent isotope and product isotope
effects.
6
9
-
6 + H+ 98 (PhCO2 )(Ph)C•-OH (7) + (CH3)2S•+ (1a) (13)
-
6
9
8 (PhCO2 )(Ph)C•-O- (7-) + 1a
(14)
(15)
(16)
(17)
6
9
8 7 + •CH2SCH3
7- + H+ h 7
1a + >S yk z [(CH3)2S S(CH3)2]+ (2a)
17
k-17
By measurement of the photolytic yield of reaction 15 in our
steady-state system we will be able to obtain the photolytic
yields of the electron transfer pathways (reactions 13 and 14)
based on the ratio of Φel/ΦH established by laser photolysis10
3
for the CBP/DMS system.
Results
At neutral pH, there are two possible pathways to irreversibly
consume CBP in the absence of oxygen, namely reactions 18
(formation of pinacol 8) and 19. On the other hand, reverse
electron transfer from 7 (or 7-) to 1a or 2a, respectively, will
reconstitute ground state CBP (reaction 20).
A. Photolytic Yields of 3CBP and Dimethyl Sulfide
Radical Cations in H2O and D2O. The photolytic yields of
3CBP and sulfur radical cations in H2O and D2O in our steady-
state photolysis systems were quantified by a chemical dosimeter
3
based on reaction 15 between CBP and DMS in N2-saturated
-
-
aqueous (H2O or D2O) solutions, pH 7.0, containing 2.5 × 10-4
M CBP and various concentrations of DMS between 2 × 10-3
M and 3.4 × 10-2 M.
27
9
8 (PhCO2 )(Ph)(HO)C-C(OH)(Ph)(PhCO2 ) (8)
(18)
-
7 + •CH2SCH3
9
8 (PhCO2 )(Ph)(HO)C-CH2SCH3 (9) (19)
The quenching of 3CBP by DMS in aqueous solution occurs
with k9 ) 1.5 × 109 M-1 s-1 9
.
By laser-photolysis the
7 + 1a/2a
9
8 H+ + DMS/2DMS + CBP
(20)
respective quantum yields Φ for processes 10-15 in H2O and
D2O were determined at both pL 6.7 and pL 10 (L ) H, D).10
These quantum yields are essentially independent of pL in this
region. A non-productive reaction of 3CBP with DMS to yield
ground-state CBP occurs with Φn,H O ) 0.72 and Φn,D O ) 0.77.
In order to obtain a quantitative relationship between the
consumption of CBP and the initial yield of CBP we have to
ensure that any deprotonation of 1a (reaction 21) will not
compete against the reverse electron transfer reaction 20.
3
2
2
These nonproductive pathways represent either physical quench-
ing (reaction 10) or back-electron transfer (reaction 12) within
the initially formed charge-transfer complex 6 (reaction 11).
Competitively, complex 6 decomposes via electron transfer
(reactions 13 and 14) (Φel,H O ) 0.17 and Φel,D O ) 0.10, related
1a
9
8 •CH2SCH3 + H+
(21)
This deprotonation reaction proceeds significantly faster from
the monomeric species 1a than from the dimer 2a. Given that
the irreversible deprotonation from 1a could remove 1a/2a from
the solution, then the consumption of CBP should be dependent
on the concentration of DMS. With increasing DMS concentra-
tions the final consumption of CBP should become representa-
tive for reactions 15, 18 and 19 since high DMS concentrations
stabilize 2a, available for reverse electron transfer. In fact,
during pulse radiolysis experiments at DMS concentrations
>10-2 M, species 2a was observed to entirely decay via second-
order radical-radical reactions, indicating that deprotonation
via reactions -17 and 21 is negligible at such high DMS
concentrations.13,14
2
2
to initial [3CBP]), and hydrogen transfer (reaction 15) (ΦH,H O
2
) 0.12 and ΦH,D O ) 0.13). Thus, only the electron transfer
2
pathway shows a significant normal product isotope effect of
Φ
el,H O/Φel,D O ) 1.7. This mechanism initially yields 7- which
2
2
(4) Mere´nyi, G.; Lind, J.; Engman, L. J. Phys. Chem. 1996, 100, 8875-
8881.
(5) (a) Bobrowski, K.; Scho¨neich, Ch. J. Chem. Soc., Chem. Commun.
1993, 795-797. (b) Scho¨neich, Ch.; Bobrowski, K. J. Phys. Chem. 1994,
98, 12613-12620.
(6) Eriksen, J.; Foote, C. S.; Parker, T. L. J. Am. Chem. Soc. 1977, 99,
6455-6456.
(7) Ando, W.; Kabe, Y.; Kobayashi, S.; Takyu, C.; Yamagishi, A.; Inaba,
H. J. Am. Chem. Soc. 1980, 102, 4526-4528.
(8) Inoue, K.; Matsuura, T.; Saito, I. Tetrahedron 1985, 41, 2177-2181.
(9) Bobrowski, K.; Marciniak, B.; Hug, G. L. J. Photochem. Photobiol.,
A 1994, 81, 159-168.
(11) Inbar, S.; Linschitz, H.; Cohen, S. G. J. Am. Chem. Soc. 1981, 103,
7323-7328.
(12) Mo¨nig, J.; Goslich, R.; Asmus, K.-D. Ber. Bunsenges. Phys. Chem.
1986, 90, 115-121.
(10) Bobrowski, K.; Hug, G. L.; Marciniak, B.; Miller, B. L.; Scho¨neich,
Ch., manuscript in preparation
(13) Bonifaˇcic´, M.; Mo¨ckel, H.; Bahnemann, D.; Asmus, K.-D. J. Chem.
Soc. Perkin Trans. 2 1975, 675-685.

11016 J. Am. Chem. Soc., Vol. 118, No. 45, 1996
Table 1. Rate Constants Used or Determined in this Paper
Miller et al.
process
rate constant
ref
k3
k6
k8
k9
k22
k23
k26
k27
k28
k31
k33
k37
k56
2.6 × 109 M-1 s-1
1.1 × 108 M-1 s-1
(2.3 ( 1.2) × 1011 M-1 s-1
1.5 × 109 M-1 s-1
4.2 × 106 M-1 s-1
1.5 × 105 s-1
35
4,5
this work
9
11
11
this work
this work
(3.5 ( 0.3) × 107 M-1 s-1
e3.0 × 104 M-1 s-1
4.0 × 109 M-1 s-1
e5.0 × 109 M-1 s-1
e1.0 × 108 M-1 s-1
≈102 M-1 s-1
20
estimated from 23
estimated from 28
estimated from 26
34
2.0 × 109 M-1 s-1
The photolysis of our N2-saturated aqueous solutions yields
a final consumption of CBP of (5.0 ( 0.5) × 10-7 M s-1
regardless of DMS concentrations between 2 × 10-3 M and
3.4 × 10-2 M, and regardless of whether the reaction was
performed in H2O or D2O (because of the broader spectrum of
incident light in the steady-state photolysis experiments the
yields are presented in units of M s-1 rather than quantum yields
in order to avoid confusion with laser photolysis experiments
employing a defined wavelength of incident light; see Experi-
mental). For a quantitative assessment we have to take into
account that 3CBP can competitively suffer physical quenching
according to reactions 22 (k22 ) 4.2 × 106 M-1 s-1) and 23
(k23 ) 1.5 × 105 s-1).11
3CBP + CBP
3CBP
8 CBP
9
8 2CBP
(22)
(23)
9
Table 1 summarizes all rate constants used or determined in
this paper. With k9 ) 1.5 × 109 M-1 s-1, we derive that at
[DMS] g 2.0 × 10-3 M more than 95% of 3CBP will directly
attack DMS. Thus, the yield of CBP consumption at [DMS]
g 2.0 × 10-3 M should be representative for reactions 15, 18,
and 19. On the basis that reaction 15 leads to the consumption
of 1 equiv CBP per hydrogen transfer event, and the quantum
yields10 ΦH,H O ) 0.12 and ΦH,D O ) 0.13, we calculate that
Figure 1. Photolytic yields of DMSO as a function of pH. Air-saturated
aqueous solutions, containing 2.5 × 10-4 M CBP, 1.0 × 10-2
M
phosphate buffer, and (A) 6.8 × 10-3 M DMS (system I) or (B) 3.4 ×
10-2 M DMS (system II). (C) Oxygen-saturated solution containing
2.5 × 10-4 M CBP, 1.0 × 10-2 M phosphate buffer, and 3.4 × 10-2
M DMS (system III). Curves a (9): total DMSO yields in absence of
SOD; curves b (b): SOD-dependent DMSO yields obtained by
subtraction of curves c from curves a; curves c ([): SOD-independent
DMSO yields obtained in the presence of 1,000 U/mL SOD.
2
2
3
the initial yield of CBP in our steady-state system is on the
order of (4.2 ( 0.5) × 10-6 M s-1 in both solvents H2O and
D2O, respectively.
From the ratios10 [Φel/ΦH]H O ) 1.42 and [Φel/ΦH]D O ) 0.77
U/mg and a molecular weight of 34 000 Da for the SOD dimer).
These remaining yields will be referred to as SOD-independent
yields. Identical results as with 1000 U/mL SOD were obtained
with 500 U/mL and 2000 U/mL SOD, respectively (data not
shown), indicating that the SOD effect was independent of the
SOD concentration at these SOD levels, and therefore at
maximum. A tendency of increasing SOD-independent DMSO
yields is observed with increasing pH in the presence of 1000
U/mL Cu,Zn superoxide dismutase (SOD) (curves c). Subtrac-
tion of the SOD-independent yields from the total DMSO yields
gives the SOD-dependent DMSO yields (curves b). The latter
show rather constant values between pH 6.2 and 9.0 and drop
significantly only for pH > 9.0 at high concentrations of DMS
(curves b). By assessing the SOD activity before and after each
experiment we confirmed that there was no measurable inac-
tivation of the enzyme during the photolysis.
Figure 1C displays the DMSO yields obtained after photolysis
of oxygen-saturated aqueous (H2O) solutions containing 1.0 ×
10-2 M sodium phosphate, 2.5 × 10-4 M CBP, and 3.4 × 10-2
M DMS (system III) at various pH. It is important to note that
there is no significant difference between the total DMSO yields
obtained with 3.4 × 10-2 M DMS under conditions of air-
saturation and oxygen-saturation, respectively (curves a in
2
2
we derive that the electron transfer processes (reactions 13 and
14) should generate 1a/2a at levels of Φel,H O × [3CBP] ) 7.1
2
× 10-7 M s-1 in H2O and Φel,D O × [3CBP] ) 3.85 × 10-7
M
2
s-1 in D2O. As we will see in the discussion, these yields are
well in accord with the observed material balances.
B. The Aerobic Oxidation of DMS to Dimethyl Sulfoxide
3
by CBP. 1. Influence of pH, Superoxide Dismutase, and
Oxygen Concentration. Figure 1A,B display the yields of
dimethyl sulfoxide (DMSO) obtained as a function of pH during
the photolysis of air-saturated aqueous (H2O) solutions, contain-
ing 2.5 × 10-4 M CBP, 1.0 × 10-2 M sodium phosphate, and
6.8 × 10-3 M DMS (system I) or 3.4 × 10-2 M DMS (system
II), respectively. At pH values between 6.2 and 10, the total
DMSO yields are rather constant and drop only slightly at pH
11 (curves a). A significant reduction of the DMSO yields (e.g.
by 90% at pH 6.2 for 3.4 × 10-2 M DMS) is observed when
the experiments are performed in the presence of 1000 U/mL
Cu,Zn superoxide dismutase (SOD) (curves c), corresponding
to 6.7 × 10-6 M SOD dimer (taking a specific activity of 4400
(14) Asmus, K.-D. In Sulfur-Centered ReactiVe Intermediates in Chem-
istry and Biology; Chatgilialoglu, C., Asmus, K.-D., Eds; NATO ASI
Series: Plenum Press: New York, 1990; Vol. 197, 155-172.

Transformation of Sulfur Radical Cations into Sulfoxides
J. Am. Chem. Soc., Vol. 118, No. 45, 1996 11017
Table 2. Yields of DMSO as a Function of DMS Concentration
and SODa
DMSO, 10-7 M s-1
SOD,
U/mL
0.68 × 10-2
2.04 × 10-2
3.4 × 10-2
M DMS
M DMS
M DMS
0
500
1000
8.20 ( 0.2
1.71 ( 0.13
1.78 ( 0.02
9.61 ( 0.41
2.59 ( 0.03
2.27 ( 0.06
11.7 ( 0.7
3.48 ( 0.05
3.67 ( 0.17
a Air-saturated solutions, pH 8.0, containing 2.5 × 10-4 M CBP and
1.0 × 10-2 M sodium phosphate.
Figure 1B,C). However, in particular at pH 6-7 there is a larger
fraction of SOD-independent DMSO formation (curve c) in the
oxygen-saturated system.
In contrast to oxygen-containing solutions there was no
significant sulfoxide formation during the photolysis of N2-
saturated solutions. This result demonstrates the importance
of oxygen in the mechanistic scheme. Under our experimental
conditions this can be rationalized by the competition of reverse
electron transfer (reaction 20) with disproportionation. How-
ever, in earlier radiation chemical experiments we had excluded
that disproportionation of 1a or 2a contributes to sulfoxide
formation even under conditions where reverse electron transfer
reactions are absent, i.e., in the absence of species such as 7
and 7-.3
Table 2 displays the SOD-independent DMSO yields obtained
as a function of DMS concentration at pH 8.0 (air-saturated
aqueous solution containing 2.5 × 10-4 M CBP and 1.0 × 10-2
M sodium phosphate). An important feature is that the yields
of DMSO gradually increase with increasing DMS concentration
independent of the absence or presence of 500 or 1000 U/mL
SOD.
2. Native SOD and Demetalated Apoenzyme. Sulfide
radical cations such as 1a or 2a are very strong one-electron
oxidants.15 Therefore, we had to design an experiment to
confirm that the inhibition of sulfoxide formation by SOD was
not merely caused by the competitive reaction of sulfide radical
cations with easily oxidizable amino acids X of the SOD protein
skeleton (reaction 24) where X in bovine erythrocyte SOD could
be Tyr or Met.16,17
Figure 2. (A) Activity of native and des-Cu SOD at inhibiting the
xanthine/xanthine oxidase induced reduction of cytochrome c. (B)
Efficiency of various concentrations of native and des-Cu SOD at
inhibiting the photolytic formation of DMSO; conditions: air-saturated
aqueous solutions, pH 8.0, containing 2.5 × 10-4 M CBP, 3.4 × 10-2
M DMS, and 1.0 × 10-2 M phosphate buffer.
2a + SOD-(X)
9
8 2>S + SOD-(X•+)
(24)
For this purpose a demetalated des-Cu variant (apo-SOD) of
SOD was prepared by partial unfolding of the enzyme, extrac-
tion of Cu2+ with EDTA, and refolding (see Experimental).
Figure 2A displays the activities of native enzyme and apoen-
zyme in terms of inhibition of the reduction of ferricytochrome
c by superoxide, generated by the xanthine/xanthine oxidase
system. The apo-SOD shows approximately 9% of the activity
as compared to the native enzyme, essentially comparable to
the preparation of apoenzyme by McCord and Fridovich.18 This
residual activity is most probably caused by an incomplete
removal of Cu2+. When air-saturated solutions containing 1.0
× 10-2 M sodium phosphate, pH 8.0, 2.5 × 10-4 M CBP, and
3.4 × 10-2 M DMS were photolyzed in the presence of either
native SOD or apoenzyme, the apoenzyme was much less
effective in inhibiting the formation of DMSO than was native
SOD, as shown in Figure 2B. These data suggest that it is the
dismutation of superoxide by SOD which causes the observed
Figure 3. Competition plot displaying the effect of the active fraction
of SOD on the photolytic formation of DMSO (conditions as for Figure
2B). The active fractions of SOD were obtained by multiplication of
the enzyme concentration with fa ) 1.0 for native and fa ) 0.09 for
des-Cu SOD. Native SOD: b, des-Cu SOD: 9. Data were corrected
for SOD-independent DMSO through processes not initiated by sulfur
radical cation formation, i.e., from the measured DMSO yields at
various (low) SOD concentrations between 5 × 10-10 and 3.7 × 10-8
M were subtracted the DMSO yields at high SOD concentrations of
6.7 × 10-6 M.
inhibition of sulfoxide formation and not an unspecific oxidation
of amino acid residues of SOD by 2a according to reaction 24.
The fact that the apoenzyme still inhibits sulfoxide formation,
though to lesser extent, can be related to its residual 9% SOD
activity. This becomes evident from Figure 3 which shows a
plot of [DMSO]max/[DMSO] vs fa[SOD] where [DMSO]max and
[DMSO] represent the yields of DMSO in the absence and
presence of SOD, respectively, fa represents the fraction of active
(15) Bonifaˇcic´, M.; Weiss, J.; Chaudhri, S. A.; Asmus, K.-D. J. Phys.
Chem. 1985, 89, 3910-3914.
(16) Keele, Jr., B. B.; McCord, J. M.; Fridovich, I. J. Biol. Chem. 1971,
246, 2875-2880.
(17) Forman, H. J.; Evans, H. J.; Hill, R. L.; Fridovich, I. Biochemistry
1973, 12, 823-827.
(18) McCord, J. M.; Fridovich, I. J. Biol. Chem. 1969, 244, 6049-6055.

11018 J. Am. Chem. Soc., Vol. 118, No. 45, 1996
Miller et al.
SOD, and [SOD] represents the concentration of SOD dimer.
In this plot all the points fall on a line regardless whether
obtained with native SOD or apoenzyme.
3. Exclusion of a Direct Oxidation of DMS to DMSO by
Superoxide. The fact that SOD inhibits the formation of DMSO
necessitates experimental evidence as to what role superoxide
plays during the oxidation reaction. One potential pathway
would be the direct oxidation of DMS to DMSO by superoxide
or its conjugate acid, HOO• (reaction 25).
(CH3)2S + O2•- 98 (CH3)2SdO + products
(25)
A series of experiments was carried out in oxygen-saturated
aqueous (D2O) solution, pD 10, containing 1.0 × 10-2
M
Figure 4. Solvent isotope effect on the photolytic formation of DMSO.
Conditions: Aqueous (L2O, L ) H, D) solutions, pL 8.0, containing
1.0 × 10-2 M phosphate buffer, 2.5 × 10-4 M CBP, various
concentrations DMS, and no (curves a and b) or 1,000 U/mL superoxide
dismutase (curves c and d).
sodium phosphate, 2.5 × 10-4 M CBP, 6.8 × 10-3 M DMS,
and (i) no further additive, (ii) 1.33 M sodium formate, or (iii)
1.33 M sodium acetate. Conditions of pD 10 were selected in
order to accelerate superoxide formation from 7 (via 7-; see
below) and to slow down the bimolecular dismutation of
superoxide.19 Analysis of the photolyzed reaction mixtures by
1H-NMR revealed that identical yields of DMSO were formed
in the absence of any further additive and in the presence of
1.33 M sodium acetate, respectively, whereas the presence of
1.33 M sodium formate caused a 65% reduction of the DMSO
yields (1H-NMR analysis was preferred to HPLC for these
experiments since the high salt concentrations compromised the
retention behavior of DMS, DMSO, and CBP during reversed
phase chromatography).
then we would not expect any protection by SOD which
dismutates superoxide to H2O2).
4. The Rate Constants k26 and k27. The rate constant k26
was determined by standard competition kinetics measuring the
decomposition of CBP, -d[CBP], in N2-saturated aqueous
solutions, pH 7.0, containing 2 × 10-2 M sodium phosphate,
2.5 × 10-4 M CBP, and various concentrations of sodium
formate.
A plot of d[-CBP]max/[-CBP] vs formate concentration yielded
a straight line with a slope equal to (k22 + k23 [CBP])/k26 from
which k26 ) (3.5 ( 0.3) × 107 M-1 s-1 was derived. There
was no measurable decomposition of CBP in the presence of
up to 5.0 × 10-2 M sodium acetate instead of formate, and we
can obtain an upper limit of [-CBP]max/[CBP] g 100 for [acetate]
-
3CBP + HCO2- 98 7 + •CO2
(26)
(27)
3CBP + CH3CO2- 98 7 + •CH2CO2
-
e 5.0 × 10-2 M which leads to k27 e 3.0 × 104 M-1 s-1
.
•-
•CO2- + O2
98 CO2 + O2
(28)
5. Solvent Isotope Effects. Figure 4 displays the yields of
DMSO obtained as a function of L2O (L ) H, D) for the
photolysis of various concentrations of DMS in air-saturated
aqueous solutions containing 1.0 × 10-2 M sodium phosphate,
pL ) 8.0, and 2.5 × 10-4 M CBP. There are several important
features. (i) The total yields of DMSO are significantly higher
in H2O (curve a) as compared to D2O (curve b), i.e., by a factor
of 1.94 for [DMS] ) 6.8 × 10-3 M and 2.04 at [DMS] ) 3.4
× 10-2 M. (ii) The SOD-independent pathway in the presence
of 1,000 U/mL SOD (curves c and d) produced DMSO to
comparable extents in H2O and D2O for [DMS] e 2.0 × 10-2
M. However, for 3.4 × 10-2 M DMS the SOD-independent
pathway gave 1.58 times higher yields of DMSO in H2O (3.67
× 10-7 M s-1) than in D2O (2.32 × 10-7 M s-1). There were
no differences between experiments employing 500, 1000, or
2000 U/mL SOD in both H2O and D2O, indicating that an SOD
activity of 1,000 U/mL was sufficient for suppressing the SOD-
dependent pathway in both solvents. These results mean that
in D2O a larger fraction of DMSO is generated via the SOD-
independent pathway though the total yields of sulfoxide are
significantly smaller than in H2O. This is particularly obvious
for the smaller DMS concentrations. The variation of oxygen
concentration between 2.5 × 10-4 M (air-saturated) and 1.25
× 10-3 M (oxygen-saturated) had little influence on the total
DMSO formation in H2O, as displayed in Figure 5. Only in
D2O an effect is noticable for 6.8 × 10-3 M DMS where the
yield of DMSO increases from 3.7 × 10-7 M s-1 in air-saturated
solution to 4.7 × 10-7 M s-1 in oxygen-saturated solution.
•CO2- + CBP
7- + O2
9
8 CO2 + 7-
(29)
(30)
•-
9
8 CBP + O2
From competition experiments (see below) we obtained k26
) (3.5 ( 0.3) × 107 M-1 s-1 and k27 e 3 × 104 M-1 s-1. In
the acetate-containing system, superoxide should be formed via
a sequence of reactions 13-16 and 30, and the extent of reaction
27 should be negligible (k9[DMS] . k27[CH3CO2-]). Reaction
•-
30 will be rapid at pD 10. Thus, we expect 1 equiv of O2
3
per equivalent of CBP in the acetate-containing system. The
-
•CO2 radical anion rapidly reduces molecular oxygen either
directly (reaction 28; k28 ) 4 × 109 M-1 s-1 20) or via initial
reduction of ground state CBP (analogous to the reaction
between (CH3)2C•-OH and CBP) followed by reaction 30. Thus,
the reaction sequence 26 and 28-30 yields 2 equiv of
superoxide per equivalent of 3CBP reacting with formate.
Competition kinetics predict that, in the presence of 1.33 M
sodium formate and 6.8 × 10-3 M DMS, a fraction of 82% of
3CBP should directly react with formate, overall producing a
1.8 fold higher yield of O2•- in the formate system than in the
acetate system. However, in the formate system we observe a
65% reduction of the formation of DMSO which (i) suggest
that superoxide does not directly oxidize DMS under our
experimental conditions, and (ii) is reasonably close to the
expected drop of the DMSO yields (82%) based on competition
kinetics (any significant direct oxidation of either DMS or 2a
by the dismutation product H2O2 can also be excluded since
6. The Source of Oxygen: Experiments with 16O2 in
18OH2. The source of the oxygen in DMSO was determined
by GC-MS experiments subsequent to the photolysis of air-
saturated (16O2/N2, 1:4, v/v) solutions containing 1.25 × 10-4
(19) Bielski, B. H. J.; Cabelli, D. E. Int. J. Radiat. Biol. 1991, 59, 291-
319.

Transformation of Sulfur Radical Cations into Sulfoxides
J. Am. Chem. Soc., Vol. 118, No. 45, 1996 11019
Table 5 compares the yields of diethylsulfoxide (DESO) and
DMSO obtained after photolysis of air-saturated aqueous
solutions containing 2.5 × 10-4 M CBP and 1.0 × 10-2
M
sodium phosphate, pH 7.0, in the absence and the presence of
SOD.
The most notable feature is that the ratio of total sulfoxide
yields to the SOD-independent sulfoxide yields are much lower
for DES than for DMS, in particular at the lower sulfide
concentration. The SOD-independent DESO yields are higher
than the SOD-independent DMSO yields at both sulfide
concentrations.
Discussion
1. Sulfur Radical Cations as Intermediates for Sulfoxide
Formation in the SOD-Dependent But Not the SOD-
Independent Pathway. During the photooxidation of DMS by
3CBP in oxygen-containing solutions, sulfur radical cations can
theoretically form via two distinct pathways. A type I mech-
anism is based on the direct one-electron oxidation of DMS by
3CBP (reactions 11, 13, 14, and 17). Alternatively, the
interaction of 3CBP with triplet molecular oxygen can generate
singlet oxygen (reaction 31).21-23
Figure 5. Influence of oxygen concentration of the photolytic formation
of DMSO in H2O and D2O at various concentrations of DMS.
Conditions: Aqueous solutions (L2O, L ) H, D), pL ) 8.0, containing
2.5 × 10-4 M CBP and 1.0 × 10-2 M phosphate buffer.
M CBP and 2.0 × 10-2 M sodium phosphate, pL 10 (L ) H or
D), in either H218O, H218O/H216O (1:1, v/v), or H218O/D216O
(1:1, v/v), respectively. The results are summarized in Table
3. In control experiments we analyzed synthetic DMS16O
dissolved in H218O (entry 2), and DMS18O in H216O (entry 3).
There was no oxygen exchange between DMSO and water under
our experimental conditions. Further control experiments
(entries 9-11) were done particularly with regard to the possible
incorporation of deuterium into DMSO during photolytic
reactions carried out in D2O/H218O mixtures (entry 7). There
was no incorporation of deuterium into DMSO during photolytic
oxidation of DMS in D2O (entries 10 and 11), showing that the
significant yield of the isotopic cluster of m/z ) 80 in entry 7
is caused by the incorporation of 18O from the solvent.
In general, the incorporation of 18O into DMSO is small under
all experimental conditions, indicating that the oxygen is derived
mainly from molecular oxygen (16O2). However, we find that
a significantly higher fraction of DMS18O is present whenever
the experiments are carried out in the presence of SOD. Even
higher levels of 18O incorporation into DMSO are detectable
when the experiments are carried out in mixtures of H2O and
D2O. Mechanistically we can conclude that sulfoxide formation
via the SOD-independent pathway involves at least one inter-
mediate different from the intermediate formed in the absence
of SOD (hence different fractions of 18O incorporation under
both experimental conditions). Furthermore, the incorporation
of 18O is promoted by the presence of D2O.
3CBP + 3O2
9
8 CBP + 1O2(1∆g, 1Σ+ )
(31)
g
Singlet oxygen could theoretically produce sulfur radical
cations via one-electron oxidation of DMS (reaction 32)
although this process appears rather unlikely in light of the fact
that such electron transfer mechanisms gave small yields only
with electron rich sulfides.8
•-
1O2 + >S
98 1a + O2
(32)
Singlet oxygen rather adds to DMS to produce the zwitter-
ionic intermediate 10 (reaction 33).1
1O2 + >S h >S(+)-O-O(-) (10)
(33)
We note that in water the lifetime of 10 may be rather short
not only because of its reaction with a second DMS to yield 2
equiv of DMSO per equivalent of 10 (reaction 34)1 but also
due to a possible addition of HO-/H+ to yield a hydroperoxy
sulfurane, HO-S(CH3)2-OOH (11) (see below).
10 + >S
98 2>SdO
(34)
7. The Formation of Formaldehyde. There is little
variation of formaldehyde yields over a range of pH 6.2-11.0,
representatively measured for system II, with an average
formation of (7.3 ( 0.6) × 10-7 M s-1. The additional presence
of 1000 U/mL SOD had no influence on the formaldehyde
yields.
8. Other Reaction Products. In order to obtain a complete
material balance we monitored the formation of several other
possible reaction products which could originate from the
oxidation and subsequent reactions of DMS. The yields of
methanethiol (CH3SH), dimethyldisulfide (CH3SSCH3), meth-
anesulfonic acid (CH3SO3H), and dimethylsulfone ((CH3)2SO2)
were quantified by 1H-NMR by comparison to known quantities
of authentic standards, as displayed in Table 4.
For our mechanistic considerations of the SOD-dependent
pathway of sulfoxide formation the mere formation of sulfur
radical cations 1a, regardless whether via reactions 11, 13, and
14 or reaction 32, would be sufficient. However, for a
quantitative assessment of sulfoxide formation via the SOD-
independent pathway, it is important to determine the potential
relative contributions of singlet oxygen (via the initial inter-
mediate 10; reactions 33 and 34) and the hydroxide pathway
(reactions 3-7).
The individual processes of reaction 31 are currently under
extensive investigation with regard to the yields of singlet
oxygen and intermediates such as exciplexes.21-23 In general,
it appears that the efficiency of singlet oxygen formation, S∆,
decreases with decreasing oxidation potential of the sensitizer
and increasing rate constant of the quenching process.23 By
comparison with values for benzophenone in solvents of
It is evident that all the products listed in Table 4 are formed
with only minor yields when compared to the yields of dimethyl
sulfoxide and formaldehyde.
3
C. The Aerobic Oxidation of Diethyl Sulfide by CBP.
(20) Simic, M. G. Methods Enzymol. 1990, 186, 89-100.
(21) McLean, A. J.; Rodgers, M. A. J. J. Am. Chem. Soc. 1992, 114,
3145-3147.
Conclusive information with regard to the involvement of
dimeric sulfur radical cations in the sulfoxide formation can be
obtained from a comparison of the sulfoxide yields from the
photolysis of DMS and diethyl sulfide (DES), respectively.
(22) McLean, A. J.; Rodgers, M. A. J. J. Am. Chem. Soc. 1993, 115,
9874-9875.
(23) Grewer, C.; Brauer, H.-D. J. Phys. Chem. 1994, 98, 4230-4235.

11020 J. Am. Chem. Soc., Vol. 118, No. 45, 1996
Miller et al.
Table 3. Incorporation of 18O into DMSO in the Absence and Presence of SOD and D2Oa
DMSO molecular ion (MW 78)
incorporation
system
(SOD
isotopic cluster m/z 80 as % m/z 80a
of 18O (%)c
1
2
3
4
DMS16O/H216O
4.60 ( 0.37
4.60 ( 0.40
94.0
DMS16O/H218O
0.0
94.0
DMS18O/H216O
DMS16O/H216O
4.51 ( 0.04
control, entries 5-7
3.4 × 10-2 M DMS,
H218O, pH 10
5
hν
6
-SOD
+SOD
7.14 ( 0.01b
10.4 ( 0.28b
2.63
5.90
3.4 × 10-2 M DMS,
H218O/H216O (1:1),
pH 10
hν
-SOD
+SOD
5.03 ( 0.5b
5.94 ( 0.4b
0.52
1.43
7
hν
3.4 × 10-2 M DMS,
H218O/D216O (1:1),
pH 10
-SOD
+SOD
7.61 ( 0.26b
8.09 ( 0.12b
3.10
3.58
8
9
DMSO/H216O
control, entries 9-11
3.0 × 10-2 M DMSO/
D216O, pD 10, no hν
3.0 × 10-2 M DMS/
D216O, pD10, +catalase
3.0 × 10-2 M DMS,
D216O, pD 10
4.63 ( 0.06
4.75 ( 0.09
+SOD
+SOD
+SOD
(0.12)d
10
hν
11
hν
4.50 ( 0.42
4.66 ( 0.09
0
0
a Standard deviation from n ) 3. b n ) 2. c m/z 80 from isotopic experiment corrected for experimental control value for natural abundance. The
predicted natural abundance of the DMSO molecular ion isotopomers is m/z 78 (100%) and m/z 80 (4.67%). d Incorporation of trace amount of
deuterium. e Photolytic experiments (entries 5-7, 10, and 11, labeled hν), are air-saturated and contain 1.25 × 10-4 M CBP.
DES ([(CH3CH2)2S S(CH2CH3)2]+, 2b) is less stable than the
Table 4. Reaction Products from the Photolytic Oxidation of 1.0
× 10-2 M DMS in N2- and O2-Saturated Solution Exposed to a
analogous species from DMS, 2a, and suffers much faster
Photon Flux of 1.48 × 10-5 M photons/s at 350 nm
deprotonation (via the monomer 1b) (reactions 35 and 36).13
yield, M s-1
2b h (CH3CH2)2S•+ (1b) + (CH3CH2)2S
8 H+ + CH3C•H-S-CH2CH3
(35)
(36)
conditions pH
N2-sat 6.5
CH3SH CH3SSCH3 CH3SO3H CH3SO2CH3
1.23 × 10-8
N2-sat 12.0 6.5 × 10-8 1.23 × 10-8 5.5 × 10-9
1b
9
O2-sat
O2-sat 12.0
6.5 5.0 × 10-8 1.4 × 10-7 6.6 × 10-8 2.4 × 10-8
2.72 × 10-8
Any potential reaction of superoxide with 2b has to compete
against deprotonation, and, therefore, sulfoxide formation via
sulfur radical cations will be less efficient for DES as compared
to DMS. The higher SOD-independent sulfoxide yields from
DES (compared to DMS) then suggest that these do not directly
originate from intermediary sulfur radical cations.
Table 5. Photolytic Yields of Sulfoxide in Air-Saturated Solutions,
pH 7, Containing 2.5 × 10-4 M CBP, 1.0 × 10-2 M Sodium
Phosphate, and Various Concentrations of Sulfide and SOD
(1000 U/mL
sulfoxide,
sulfide
DMS
concentration
SOD
10-7 M s-1
ratioa
4.1
0.68 × 10-2
0.68 × 10-2
3.4 × 10-2
3.4 × 10-2
M
-SOD
+SOD
-SOD
+SOD
-SOD
+SOD
-SOD
+SOD
7.4 ( 0.2
1.8 ( 0.1
4.6 ( 0.3
3.5 ( 0.1
12.2 ( 0.3
1.8 ( 0.1
8.4 ( 0.5
5.0 ( 0.3
A second line of evidence for the involvement of 1a/2a in
the SOD-dependent formation of DMSO derives from the
experiments in H2O and D2O. The experimentally observed
product isotope effect of [DMSO]H O/[DMSO]D O ) 2.0 cor-
DES
DMS
DES
M
1.3
6.8
1.7
2
2
M
relates with the product isotope effect of [2a]H O/[2a]D O
)
2
2
Φ
el,H O/Φel,D O ) 1.7, derived by laser photolysis.10 The slight
2
2
M
difference between 2.0 and 1.7 can be rationalized by additonal
secondary (solvent) isotope effects on the formation of DMSO,
i.e., a more efficient (competitive) reaction of DO- as compared
to HO- 25 with intermediates 2a and 10 (see also below).
2. The Potential Role of Hydrogen Peroxide for the SOD-
Independent Pathway. A major contribution of hydrogen
peroxide to the SOD-independent pathway can be excluded. For
all our photolysis experiments any potential contribution of
hydrogen peroxide was minimized by the addition of catalase
immediately after irradiation (catalase cannot be present during
the photoexperiment due to photolytic inactivation of the
enzyme). Thus, catalase was present in the reaction mixtures
at times e 180 s after start of the photolysis (depending on the
duration of photolysis which was generally between 0 and 100
s), limiting the reaction time for hydrogen peroxide and DMS
to at most 180 s. An upper limit for the concentration of
photolytically generated hydrogen peroxide concentrations can
be obtained (representatively in H2O) from the quantum
efficiency of the CBP reduction. The combined yields of CBP
a [Total sulfoxide]/[sulfoxide in presence of SOD].
different polarity23 we estimate that for CBP in water S∆ ≈ 0.30
T
T
and k31()kq ) e 5.0 × 109 M-1 s-1. On the basis of k31()kq )
e 5.0 × 109 M-1 s-1, competition kinetics predict that in our
experimental systems 3CBP will directly react with DMS to the
extent of g98% in system II (3.4 × 10-2 M DMS, 2.5 × 10-4
M O2; air-saturated), g89% in system III (3.4 × 10-2 M DMS,
1.25 × 10-3 M O2; oxygen-saturated), and g88% in system I
(6.8 × 10-3 M DMS, 2.5 × 10-4 M O2; air-saturated). From
these calculations we conclude that direct oxidation of DMS
3
by CBP should produce the major fraction of sulfur radical
cations, responsible for the SOD-dependent DMSO formation
via reaction sequences 11, 13, 14 and 8 in systems I-III. Before
a detailed consideration of the reaction mechanisms we shall
summarize the evidence for the involvement of sulfur radical
cations in the SOD-dependent sulfoxide formation. A com-
parison of DMS and DES reveals significantly smaller SOD-
dependent but significantly higher SOD-independent sulfoxide
yields for DES (see Table 4). The dimeric radical cation of
(24) Clennan, E. L.; Yang, K. J. Org. Chem. 1992, 57, 4477-4487.
(25) Schowen, R. L. Prog. Phys. Org. Chem. 1972, 9, 275-332.

Transformation of Sulfur Radical Cations into Sulfoxides
J. Am. Chem. Soc., Vol. 118, No. 45, 1996 11021
reduction by DMS (via electron and hydrogen transfer) would
amount to -d[CBP]/dt ) 1.22 × 10-6 M s-1. One-electron
reduced CBP (species 7 or 7-) then produces superoxide
(reactions 16 and 30) with d[O2•-]/dt ) -d[CBP]/dt ) 1.22 ×
10-6 M s-1, which dismutates to hydrogen peroxide with
d[H2O2]/dt ) 0.5 (d[O2•-]/dt) ) 6.1 × 10-7 M s-1. The rate
constant for the oxidation of DMS by hydrogen peroxide can
singlet oxygen (reaction 33 and 34; see below) and hydrogen
peroxide (reaction 37). However, at pH > 9.0, the following
intermediates have to be considered as major sources for a
hydroxide-mediated sulfoxide formation: the monomeric hy-
droxy sulfuranyl radical 4, its adduct with a second nonoxidized
thioether 3, and the methylthiomethylperoxyl radical 12 which
forms via hydroxide-mediated deprotonation of 1a (or 2a) to
yield the methylthiomethyl radical, followed by addition of
molecular oxygen (reactions 21, 41, and 42).
be approximated as k37 ≈ 10-2 M-1 s-1 26
.
H2O2 + >S
98 H2O + >SdO
(37)
1a + HO- 98 •CH2SCH3 + H2O
(21)
(41)
(42)
Equation (I), where [H2O2]0 ) 6.1 × 10-7 M s-1, yields an
upper limit for hydrogen peroxide-derived DMSO formation
within 180 s of 7.4 × 10-9 M s-1 for 6.8 × 10-3 M DMS
(system I) and 3.6 × 10-8 M s-1 for 3.4 × 10-2 M DMS
(systems II and III).
2a + HO- 98 (CH3)2S + •CH2SCH3 + H2O
•CH2SCH3 + O2 8 CH3SCH2OO• (12)
9
A defined concentration of 12 will always be present indepen-
dent of reactions 21 and 41 since the hydrogen abstraction
reaction 15 directly yields methylthiomethyl radicals. The
following quantification shall primarily be based on results
derived from the photolysis of system II, although crossreference
will be made to the other systems I and III for specific
information and control of the calculations.
37[DMS]t
[DMSO] ) ([H2O2]0 - [H2O2]) ) [H2O2]0(1 - e-k
)
(I)
Both values are significantly lower than the experimentally
obtained sulfoxide yields, even for the less efficient SOD-
independent pathway. Thus, we can exclude hydrogen peroxide
as a major source of DMSO formation under our experimental
conditions.
4. The Quantitative Contribution of the Individual
Processes to the SOD-Independent Sulfoxide Formation. In
systems II and III, both containing 3.4 × 10-2 M DMS, maximal
3.6 × 10-8 M s-1 of the SOD-independent sulfoxide formation
will result from the direct oxidation of DMS by hydrogen
peroxide (reaction 37) (see above). The contribution of singlet
oxygen to the SOD-independent sulfoxide formation can now
be estimated from the difference of the SOD-independent
sulfoxide yields at pH 6.2 of systems II and III (∆DMSO-
(III-II)SOD-ind,pH 6.2). This difference ∆DMSO(III-II)SOD-ind,pH 6.2
should be due to different initial yields of 1O2 in both systems
since the yields of hydrogen peroxide and 12 should be essen-
tially similar, only depending on the DMS concentration (deter-
mining photolytic conversion of CBP) and pH (determining hy-
droxide-catalyzed deprotonation of 1a and 2a). Moreover, the
hydroxide-based processes such as deprotonation of 1a and 2a
or formation of the adducts 3 and 4 are relatively inefficient at
pH 6.2.3 From equation II we calculate that the ratio of singlet
oxygen formation in systems III and II, RIII/II, amounts to RIII/II
) 4.56, where k31 e 5 × 109 M-1 s-1 (see above), k9 ) 1.5 ×
109 M-1 s-1,9 [DMS] ) 3.4 × 10-2 M, and the oxygen concen-
trations in air-saturated and oxygen-saturated solution are [O2]air
) 2.5 × 10-4 M and [O2]oxy ) 1.25 × 10-3 M, respectively.
Scheme 1
3. The Mechanism. A mechanism which rationalizes all
our experimental findings includes the reactions displayed
representatively for DMS in Scheme 1. On the SOD-dependent
side, initially present dimeric sulfur radical cations 2a react with
superoxide anion to yield the zwitterionic structure 10 either
via direct coupling of 2a and 10 (reaction 8a) or via electron
transfer (reaction 8b) followed by addition of singlet oxygen to
DMS (reaction 33). Intermediate 10 either reacts directly with
a second nonoxidized molecule of DMS (reaction 34)1 or
converts into the hydroperoxy sulfurane 11 (reaction 38) which
may also oxidize DMS to generate 2 equiv of DMSO per
equivalent of 11 (reaction 39). This mechanism shows little
variation over the pH range 6.2-9.0. At pH > 9.0, hydroxide
ions competitively react with 2a and/or 10, lowering the DMSO
yields, where the reaction of hydroxide ion with 2a yields
intermediate(s) representing precursor(s) for superoxide-inde-
pendent sulfoxide formation (see below). We note that besides
the reaction of HO- with a “naked” radical cation 2a (reaction
3) we have to consider the more general possibility of proton
abstraction by HO- from a hydrated dimeric sulfur radical cation
leading to 3 (reaction 40). Theoretical calculations have
suggested that sulfur radical cations are present as hydrated
forms where the bond dissociation energy of the R2S•+‚‚‚OH2
bond amounts to approximately 16.8 kcal/mol.27
[1O2]III [O2]oxy k31[O2]air + k9[DMS]
RIII/II
)
)
×
(II)
[1O2]II
[O2]air k31[O2]oxy + k9[DMS]
On the basis of literature values for the reaction of singlet
oxygen with various thioethers28 we approximate that k33 e 1.0
× 108 M-1 s-1, and, therefore, equation III predicts that singlet
oxygen will nearly completely react with DMS when [DMS]
) 3.4 × 10-2 M rather than suffer quenching by the solvent
(reaction 43; k43,H O ) 3.5 × 105 s-1 32).
2
k33[DMS]
[DMSO] ) [1O2]total
1O2 + H2O
×
(III)
(43)
k33[DMS] + k43
8 3O2 + H2O
9
(26) Sysak, P. K.; Foote, C. S.; Ching, T.-Y. Photochem. Photobiol. 1977,
HO- + [(H2O)m‚‚‚>S S<]+ 98 mH2O + 3
On the SOD-independent side some DMSO originates from
(40)
26, 19-27.
(27) Clark, T. in Sulfur-Centered ReactiVe Intermediates in Chemistry
and Biology; Chatgilialoglu, C., Asmus, K.-D., Eds.; NATO ASI Series:
Plenum Press: New York, 1990; Vol. 197, 13-18.

11022 J. Am. Chem. Soc., Vol. 118, No. 45, 1996
Miller et al.
With ∆DMSO(III-II)SOD-ind,pH 6.2 ) 7 × 10-8 M s-1, a
stoichiometry of 2 equiv of sulfoxide per equivalent of singlet
oxygen, and RIII/II ) 4.56 we calculate that the initial yields of
singlet oxygen in system II should be on the order of [1O2]II ≈
0.77 × 10-8 M s-1. This initial yield corresponds well to a
theoretical value of [1O2]II e 1.5 × 10-8 M s-1, predicted
according to equation IV on the basis of a competitive reaction
Considering that reactions 44 and 45 lead to 1 equiv of DMS18O
per equivalent of 10, an 11% reduction of the SOD-dependent
yields is consistent with a fraction of 22% of 10 decomposing
via reactions 44 and 45, and 78% of 10 decomposing via
reaction 34. With the SOD-dependent DMSO yields accounting
for ca. 2/3 of the total DMSO yields, and the SOD-independent
pathway contributing a fraction of 5.9% of DMS18O (entry 5,
Table 3), we derive that such reaction should yield total DMSO
yields of which a fraction of 10.2% contains 18O. Experimen-
tally we observed a fraction of 2.63% (Table 3, entry 5),
demonstrating that reaction sequence 44 and 45 alone is unlikely
to cause the hydroxide mediated reduction of the DMSO yields.
(ii) Alternatively, hydroxide ion could attack species 10 on the
terminal oxygen which, according to a suggestion of Foote and
co-workers for protic solvents,33 should accept protons from
water (reaction 46). Such a mechanism could account for a
hydroxide-mediated reduction of the DMSO yields without any
significant incorporation of 18O into DMSO.
3
3
of CBP with DMS and O2.
k31[O2]
[1O2] e [3CBP]S∆ ×
(IV)
k31[O2] + k9[DMS]
With [1O2]II ≈ 0.77 × 10-8 M s-1, the singlet oxygen-derived
DMSO yields in system II are on the order of 1.54 × 10-8
M
s-1. Subtraction of the hydrogen peroxide- and singlet oxygen-
derived yields from the total SOD-independent sulfoxide yield
at pH 6.2, 1.30 × 10-7 M s-1, leaves 7.8 × 10-8 M s-1 for
DMSO formation through pathways other than involving
hydrogen peroxide and singlet oxygen. These reactions likely
involve 12, the adducts 3 and 4, and, though speculative at
present, eventually a reaction of 2a with hydrogen peroxide.
We note an increase of the SOD-independent sulfoxide yields
with increasing pH between pH 6.2 and 9.0. Since hydrogen
peroxide and singlet oxygen are not expected to show any
significant pH-dependent variation of reactivity toward DMS
in this pH region,26 the pH-dependent increasing yields of
DMSO may be rationalized by a pH-dependent reactivity of 3,
4, 12, and/or 2a (besides the addition of HO-).
(iii) If reaction 44 were not followed by the elimination of
hydrogen peroxide, a direct oxidation of of a second molecule
DMS could occur (reactions 47 and 48).
H18O-S(CH3)2-16O-16O- + >S
9
8 16OdS< +
H18O-S(CH3)2-16O- (47)
5. Sulfoxide Formation at pH > 9.0. There is a significant
drop of the SOD-dependent sulfoxide yields at pH > 9.0
whereas the SOD-independent sulfoxide yields still increase.
In system II, this decrease of 1.2 × 10-7 M s-1 of the SOD-
dependent DMSO yields between pH 9.0 and 10.0 is paralleled
only by a 1.4 × 10-8 M s-1 increase of the SOD-independent
DMSO yields, indicating that a potential reaction of HO- with
2a leads to intermediates which yield sulfoxide less efficiently.
For mechanistic considerations we shall first discuss the
formation of DMS18O in 16O2/H218O systems, pH 10 (Table
3, entries 4 and 5), which yields information as to which
species contribute to the decreasing DMSO yields between pH
9 and 10.
H18O-S(CH3)2-16O- 98 16/18OH- + 16/18OdS< (48)
However, since reaction sequence 47 and 48, like species
10, yields 2 equiv of DMSO, such reaction of hydroperoxy
sulfurane intermediates would not lead to a pH-dependent
reduction of the DMSO yields unless the hydroperoxy sulfurane
reacts via additional competitive pathways which yield less or
no sulfoxide. The possibility of sulfone formation (reaction
49)24 can be excluded since sulfone yields in our systems were
generally low (see Table 4).
Several pathways are possible: (i) An addition of hydroxide
ion to the electrophilic sulfur of 10 initially yields a hydro-
peroxysulfurane anion (reaction 44). According to a proposal
by Sysak et al.,26 the latter may eliminate hydrogen peroxide
to yield sulfoxide (reaction 45).
HO-S(CH3)2-O-O- 98 HO- + (CH3)2SO2
(49)
(iv) An alternative would be the reaction of hydroxide ion
with the hydroperoxysulfurane (reaction 50), followed by
reaction 48.
H18O- + >S(+)-16O-16O(-) 98 H18O-S(CH3)2-16O-16O- (44)
H18O-S(CH3)2-16O-16OH + 18OH- 9
8
H18O-S(CH3)2-16O-16O- + H+ 98 18OdS< + H216O2
(45)
H216/18O2 + H18O-S(CH3)2-16O- (50)
Based on the effect of pH on the singlet oxygen-mediated
sulfoxide yields of N-formylmethionine amide,26 we expect the
reaction of hydroxide with 10 to be responsible for an 11%
reduction of the DMSO yields on going from pH 9 to pH 10.
Analogous to the calculations above, this pathway would lead
to an overall incorporation of ca. 5.1% 18O into the total DMSO
yields, a value, though closer, still higher than experimentally
determined for DMSO formation in the absence of SOD
(2.63%). Thus, at present reaction 46 represents the most
probable mechanism based on the incorporation of 18O and the
overall yield of DMSO. However, we cannot exclude that at
(28) There are several examples showing that the rate constants for the
reaction of 1O2 with sulfides are quite similar in solvents of different polarity,
e.g. for (i-C3H7)2S, k ) 2.5 × 106 M-1 s-1 in CH3OH29 and k ) 2.2 × 106
M-1 s-1 in CHCl3,30 for (t-C4H9)2S, k ) 1.5 × 105 M-1 s-1 in CH3OH29
and k ) 1.3 × 105 M-1 s-1 in CHCl3,30 and for methionine, k ) 8.6 × 106
M-1 s-1 in H2O31 and k ) 1.4 × 107 M-1 s-1 (for the CBZ-L-methionine
methyl ester) in CHCl3.30 Furthermore, it was shown that, although important
for sulfides with branched substituents (see above), steric effects appear to
be less significant for n-alkyl substituted sulfides, e.g. compare the rate
constants for (n-C4H9)2S and n-C4H9SCH3 in CHCl3 which are 2.3 × 107
(29) Kacher, M. L.; Foote, C. S. Photochem. Photobiol. 1979, 29, 765-
769.
(30) Monroe, B. M. Photochem. Photobiol. 1979, 29, 761-764.
(31) Kraljic, I.; Sharpatyi, V. A. Photochem. Photobiol. 1978, 28, 583-
586.
1
and 2.9 × 107 M-1 s-1, respectively.30 The oxidation of (C2H5)2S by O2
(32) Merkel, P. B.; Kearns, D. R. J. Am. Chem. Soc. 1972, 94, 1029-
1030.
(33) Gu, C.; Foote, C. S.; Kacher, M. L. J. Am. Chem. Soc. 1981, 103,
5949-5951.
in methanol occurs with k ) 1.7 × 107 M-1 s-1 29
.
Thus, an upper limit for
1
the oxidation of DMS by O2 in water may be set at k e 1.0 × 108 M-1
s-1
.

Transformation of Sulfur Radical Cations into Sulfoxides
J. Am. Chem. Soc., Vol. 118, No. 45, 1996 11023
least part of the reaction of hydroxide with 10 proceeds through
reactions 47, 48, and 50.
d[2a]
dt
•-
) Φel[3CBP] - k3[HO-][2a] - k8[2a][O2
]
(V)
In this respect, we note that the SOD-independent DMSO
formation shows higher levels of 18O incorporation (5.9%) than
the total DMSO formation (2.63%). Thus, very likely the
mechanism leading to the incorporation of 18O into DMSO does
not at all involve intermediate 10. Given that the hydroxide
pathway via species 5, or the analogous species HO-S(CH3)2-
S(CH3)2-OO• (13), a potential intermediate during reaction 5,3
leads to incorporation of the water oxygen into DMSO,3 we
derive that at most 1.8 × 10-8 M s-1 of the SOD-independent
DMSO yields at pH 10 can originate from reactions 3-7. This
fraction of 1.8 × 10-8 M s-1 is close to the observed increase
of the SOD-independent DMSO yields on going from pH 9.0
to 10.0 (1.4 × 10-8 M s-1) and may reflect the hydroxide
pathway (intermediates 5 and/or 13) as an additional competitive
pathway reducing the DMSO yields at pH 10. From these
considerations it becomes clear that 5 and 13 (as well as
hydrogen peroxide and singlet oxygen) are not major contribut-
ing species to the SOD-independent sulfoxide formation in this
pH region. However, it appears that the hydroxide pathway
(reactions 3-7) is responsible for the experimentally observed
incorporation of 18O into DMSO, a conclusion supported by
solvent isotope effects (see below).
For d[2a]/dt ) 0 equation (VI) follows.
Φel[3CBP]
[2a]s )
(VI)
•-
k3[HO-] + k8[O2
]
s
•-
The calculation of [O2
] is based on the fact that each
s
dismutation of superoxide by SOD consumes 2 equiv of
superoxide and that k56 is rate-determining,34 with k56 ) 2.0 ×
109 M-1 s-1 for superoxide concentrations between 10-13 and
5.0 × 10-5 M.34 We note that in the presence of SOD we do
not have to include a term of uncatalyzed dismutation of
superoxide (reaction 58) since its rate is negligible compared
to the reaction of superoxide with SOD.19
2O2•- + 2H+ 98 H2O2
(58)
Equations VII and, for d[O2•-]/dt ) 0, VIII follow where
•-
the initial yield of O2 is determined by the initial yield of 7
and 7-, (Φel + ΦH)[3CBP], on the basis of reactions 13-16
and 30.
The majority of the SOD-independent DMSO must, therefore,
originate from other species, e.g., the peroxyl radical 12.
Peroxyl radicals are known to oxidize sulfides, and such
reactions will become particularly important at the high DMS
concentration of 3.4 × 10-2 M (i.e., at low [DMS] ) 1.0 ×
10-3 M, 12 did not significantly oxidize DMS3). A potential
mechanism would involve an oxygen transfer reaction between
12 and a second sulfide (reaction 51) since the source of oxygen
has to be molecular oxygen, present through addition of O2 at
•CH2SCH3. The resulting oxyl radical will react via fragmenta-
tion (reaction 52), rearrangement (reaction 54), or hydrogen
abstraction from a second sulfide (reaction 55). Reactions 54
and 55 appear, however, less probable based on the expected
poor stability of CH3SCH2O• with regard to fragmentation
reaction 52.
•-
d[O2
]
) (Φel + ΦH)[3CBP] - k8[2a]s[O2•-] -
2k56[SOD][O2•-] (VII)
dt
(Φel + ΦH)[3CBP]
k8[2a]s[O2•-] + 2k56[SOD]
[O2•-]s )
(VIII)
Figure 3 shows that [SdO]max/[SdO] ) 2.0 at [SOD] ) 2.3 ×
10-9 M (for native SOD, fa ) 1.0). At this ratio we theoretically
expect that k8[2a] ) 2k56[SOD] ) 9.2 where k56 ) 2.0 × 109
M-1 s-1 34
With k8[2a] ) 9.2, Φel ) 0.17, ΦH ) 0.12, and
.
[3CBP] ) 4.2 × 10-6 M s-1 (see above), equation VIII yields
•-
[O2
]
) 6.6 × 10-8 M for [SOD] ) 2.3 × 10-9 M.
s
Rearrangement of equation VIII leads to equation IX which
yields k8 ) 2.3 × 1011 M-1 s-1 where k3 ) 2.6 × 109 M-1
s-1 35 and [HO-] ) 1 × 10-6 M.
12 + S<
9
8 CH3SCH2O• + >SdO
(51)
(52)
CH3SCH2O• 98 CH3S• + CH2O
2CH3S• 98 CH3SSCH3
k3[HO-]
k8 )
×
[O2•-]sΦel[3CBP]
(53)
(54)
-1
1
1
3
-
(Φel + ΦH)[ CBP] - 2k56[O2•-]s[SOD] Φel[ CBP]
3
[
]
CH3SCH2O• 98 CH3-S-•CH-OH
(IX)
CH3SCH2O• + S<
9
8 CH3SCH2OH + •CH2SCH3 (55)
This value for k8 should be given with at least (50% error,
i.e., k8 ) (2.3 ( 1.2) × 1011 M-1 s-1, in the light of possible
errors of 10-20% each in the determination of 3[CBP], k3, and
[O2•-]s. Nevertheless, the reaction of superoxide with 2a occurs
with a very high rate constant, comparable to the neutralization
of hydroxide ions by protons.36 This high rate constant may
actually reflect that the electron transfer process (reaction 8b)
predominates over the radical-radical coupling of superoxide
with 2a (reaction 8a).
6. The Rate Constant k8. The rate constant k8 ()k8a
+
k8b) can be determined from the plot displayed in Figure 3,
reflecting the competitive reaction of superoxide with the SOD
dimer (reactions 56 and 57), or with radical cation 2a (reaction
8), respectively.
2a + O2•- 98 10 + DMS
O2•- + SOD
O2•- + SODred + 2H+ 98 H2O2
98 (DMSO + products) (8)
(34) (a) Klug, D.; Rabani, J.; Fridovich, I. J. Biol. Chem. 1972, 247,
4839-4842. (b) Rotilio, G.; Bray, R. C.; Fielden, M. Biochim. Biophys.
Acta 1972, 268, 605-609. (c) Forman, H. J.; Fridovich, I. Arch. Biochem.
Biophys. 1973, 158, 396-400.
9
8 SODred
(56)
(57)
(35) By analogy to the reaction of hydroxide ion with the intramolecular
three-electron bonded sulfur radical cation of the peptide cyclo-Met-Met.
Holcman, J.; Bobrowski, K.; Scho¨neich, Ch.; Asmus, K.-D. Radiat. Phys.
Chem. 1991, 37, 473-478.
The steady-state concentrations of radical cation 2a, [2a]s,
and superoxide, [O2•-]s, are derived from equations (V)-(VIII).
(36) Eigen, M.; De Maeyer, L. Z. Elektrochem. 1955, 59, 986.

11024 J. Am. Chem. Soc., Vol. 118, No. 45, 1996
Miller et al.
7. The Competition of Hydroxide Ion and Superoxide
for 2a. We can now derive an estimate for the competition of
hydroxide ion and superoxide for 2a at pH 10. For this
treatment we need the steady-state concentration of superoxide
ion in the absence of SOD. In a first approximation the initial
yield of superoxide, equal to (Φel + ΦH) [3CBP], will be reduced
by the fast reaction of superoxide with radical cation 2a of an
initial yield Φel [3CBP] (reaction 8). A residual amount of
superoxide, equal to ΦH [3CBP], remains available for reactions
58 and 59.
fractionation factors Φ < 1.0 in the transition state (structure
14).25,38 Since reaction 51 does not contain any (rate determin-
ing) proton transfer step, this situation is consistent with an
overall normal solvent isotope effect based on relation XIII (R
) reactant sites; TS ) transition state sites).
CH3SCH2OO• (12) + O2•- 98 products
(59)
We cannot derive the steady state concentration of 12 and
k59 is not known. Neglecting reaction 59 in a first approximation
enables us to use equations X and XI to obtain [O2•-]s e 2.9 ×
10-5 M in the absence of SOD where the observed rate constant
ΦR
∏
i
kH
for reaction 58 at pH 10 is k58,pH10 ) 6 × 102 M-1 s-1 19
.
i
)
(XIII)
kD
•-
ΦTS
d[O2
]
∏
j
•- 2
e ΦH[3CBP] - k58[O2
]
(X)
j
dt
The peroxyl radical 12 alone may exist in a weakly dipolar
structure (associated with some hydrogen bonding to the
surrounding water). But this hydrogen bonding should become
even stronger in the transition structure 14, consistent with an
overall normal solvent isotope effect.
ΦH[3CBP]
[O2•-]s e
(XI)
k58
x
The application of equation XII then predicts that an increase
from pH 9 to pH 10 in system II should reduce the efficiency
of reaction 8, f8, to an extent of ca. 5% based on a competition
of superoxide and HO- for 2a.
The comparison of entries 6 and 7 in Table 3 reveals that
there is a significantly higher incorporation of 18O into DMSO
when the reaction was carried out in H218O/D216O (1:1, v/v)
instead of H218O/H216O (1:1, v/v) with an inverse product isotope
effect of 2.50 for the SOD-independent pathway. Such an
inverse isotope effect is consistent with the contribution of an
equilibrium isotope effect involving several protons with
fractionation factors of the reactant state of ΦR < 1.0 and
product state fractionation factors ΦP ≈ 1.0.38 We can neglect
any contribution of a reaction of lyoxide (LO-; L ) H, D) with
10 to this observed isotope effect since 10 does not form in
significant amounts in the presence of SOD. Thus, this isotope
effect can be rationalized by a reaction of lyoxide with 2a and
subsequent reactions. In equilibrium 60 lyoxide reacts with a
hydrated species 2a to yield L2O and 3 (which is in equilibrium
with 4).
•-
k8[O2
]
s
f8 ≈
(XII)
k8[O2•-]s + k3[HO-]
Together with the additional 11% reduction of the sul-
foxide yields through reaction of hydroxide ion with 1026 (see
above), the combined action of hydroxide on 10 and, competi-
tively with superoxide, on 2a is expected to reduce the SOD-
dependent DMSO yields by ca. 16% on going from pH
9 to 10.
Experimentally, we observe that the SOD-dependent DMSO
yields in system II decrease by 16.5% when the pH is increased
from 9.0 to 10.0. Thus, the experimental and theoretically
predicted yields are correlating well, supporting the quantitative
analysis of our proposed reaction mechanism.
8. Solvent Isotope Effects. The effect of H2O/D2O on
the SOD-dependent DMSO yields has been rationalized by the
different yields of 2a in both solvents. For the SOD-independent
DMSO yields the product isotope effects clearly support the
conclusion that singlet oxygen is not a major source of DMSO.
In particular at [DMS] ) 6.8 × 10-3 M we would have expected
higher DMSO yields in D2O due to the increased lifetime of
1O2 in D2O.37 Instead, we observe nearly similar SOD-
independent DMSO yields in H2O and D2O for [DMS] e 2.0
× 10-2 M. Moreover, a normal solvent isotope effect of 1.58
is derived for [DMS] ) 3.4 × 10-2 M. At high DMS
concentrations the major process leading to the SOD-indepen-
dent DMSO formation is reaction 51 (see above). On the basis
of the highly dipolar structure of DMSO we believe that the
oxygen transfer process is associated with solvation and
hydrogen bonding of the developing DMSO molecule as
displayed in the transition structure 14.
LO- + [(L2O)m‚‚‚>S S<]+ h
mL2O + LO-S(CH3)2-(H3C)2S• (3) (60)
The fractionation factors25 of the protons of the hydrating water
will depend on the partial positive charge x on the hydrating
water molecule. We expect that with more than one water
molecule involved in the hydration of 2a, a higher average
positive partial charge will reside on the hydrating water
molecules, i.e., the value x will approach x ) 1/m. The
application of the Gross-Butler equation results in equation
XIV where Φ1R ) 0.5, Φ2R ) 0.69, Φ1P ) 1.0, Φ2P ) 1.0, and
n ) 0.5 (atom fraction deuterium; [H218O]/[D216O] ) 0.5).
P
P
(1 - n + nΦ1 )(1 - n + nΦ2 )
KD,60 ) KH,60
(XIV)
R 1/m 2m
R
(1 - n + nΦ1 )(1 - n + n[Φ2 ]
)
For a reasonable average of three hydrating water molecules
(m ) 3) we can assume that x ≈ 0.33, resulting in KD,60/KH,60
) 1.90. In order to predict the complete solvent isotope effect
Thus, in the course of reaction 51 protons of solvent water
molecules will change their fractionation factors of Φ ) 1.0 to
(37) (a) Merkel, P. B.; Nilsson, R.; Kearns, D. R. J. Am. Chem. Soc.
1972, 94, 1030-1031. (b)Merkel, P. B.; Kearns, D. R. J. Am. Chem. Soc.
1972, 94, 7244-7253.
(38) Schowen, K. B. J. In Transition States of Biochemical Processes;
Gandour, R. D., Schowen, R. L., Eds.; Plenum Press: New York, 1978,
225-283.

Transformation of Sulfur Radical Cations into Sulfoxides
J. Am. Chem. Soc., Vol. 118, No. 45, 1996 11025
for lyoxide-mediated sulfoxide formation, we have now to
consider subsequent reactions of the intermediate 5. This
intermediate 5 should be relatively stable (as experimentally
shown for the analogous intermediate from 2-methylthiomethyl
acetate5). Equilibrium 61 will then be important for sulfoxide
formation with the deprotonated species 5- being more reactive
with respect to superoxide extrusion than the protonated species
5 which likely enters several competitive reaction pathways such
as, e.g., reactions 62 and 63.
6.2-9.0 less than 6%) whereas the majority of ∆2a will suffer
•
deprotonation to yield CH2SCH3 (reaction 21), available for
formaldehyde production via 12 and subsequent reactions
(including sulfoxide formation via reaction 51). In this respect
∆
2a ) 2.04 × 10-7 M s-1 correlates well with ∆H CO ) 2.26 ×
2
10-7 M s-1. Reactions 51-53 demonstrate that CH3S• and
CH3SH are characteristic byproducts en route to formaldehyde.
Table 4 displays that, at pH 6.5, the combined yield of CH3SH,
CH3SSCH3 (representing the initial formation of two CH3S•)
and CH3SO3H (4.0 × 10-7 M s-1) is well on the order of the
observed formaldehyde yields (among the missing products
possibly being CH3SO2H). The latter were obtained for system
III but should be quite similar to the ones obtained in system
II, based on the general similarity of both systems with respect
to the product yields (see Figure 1). Thus, our product quantities
correlate well with the initial photolytic yields of 3CBP,
permitting a full quantitative analysis of our experimental
system.
LO- + LO-S(CH3)2OO• (5) h
L2O + -O-S(CH3)2OO• (5-) (61)
2LO(CH3)2S-OO• 98 O2 + 2LO(CH3)2S-O•
LO-S(CH3)2O• 98 CH3SO2- + L+ + •CH3
(62)
(63)
Because of the relative long lifetime of 5 (and 5-),5
equilibrium 61 shall be rapid compared to reaction 7. Thus,
no primary isotope effect will result from the proton transfer
but rather an equilibrium isotope effect. For equilibrium 61
10. Conclusion. In the present work we have quantified
several pathways which lead to the conversion of aliphatic sulfur
radical cations to sulfoxide in aqueous solution. Clearly,
superoxide can react with sulfur radical cation complexes to
produce 2 equiv of sulfoxide per equivalent of one-electron
oxidized sulfur. An important parameter is the stability of such
sulfur radical cation complexes which slows down competing
reaction pathways such as deprotonation reactions. In this
respect we shall note that [R2S SR2]+ radical cation complexes
are, for example, observable when the dipeptide L-Met-L-Met
is subjected to one-electron oxidation by hydroxyl radicals39 or
3CBP.40 Preliminary experiments with L-Met-L-Met demonstrate
that significant yields of the disulfoxide L-Met(O)-L-Met(O) are
formed by reaction of superoxide with one-electron oxidized
L-Met-L-Met.41 Such pathway may be highly biologically
relevant for the intracellular protein calmodulin. When isolated
from aged brain, calmodulin contains significant amounts of
methionine sulfoxide on the subsequence Met71-Met7242 which
can be rationalized by the involvement of reactive oxygen
species (including superoxide) in age-related protein modifica-
tions.
R
the solvent isotope effect is given by equation XV where Φ1
-
) 0.5 and n ) 0.5 (assuming φ5 ≈ 1.0).
1
KD,61 ) KH,61
) 1.33
(XV)
R
(1 - n + nΦ1 )
The combination of both isotope effects yields (KD/KH)total
)
(KD,60/KH,60) (KD,61/KH,61) ) 2.53, corresponding well to our
experimentally observed isotope effect of 2.50.
9. Material Balance. In order to avoid secondary reactions
of primary oxidants (3CBP) with oxidation products, our
experimental system was based on very low conversion of
excess substrate, i.e., a photolytic conversion of <2.0 × 10-4
M DMS of an initial concentration of g 6.8 × 10-3 M. Thus,
within experimental error we cannot directly measure the
consumption of DMS and compare this value with the product
yields. However, based on the known quantum yields of DMS
3
oxidation by CBP, the known (see above) initial photolytic
Experimental Section
3
production of CBP, and the quantitative analysis of reaction
See Supporting Information.
products, a material balance can be obtained. Between pH 6.2
3
Acknowledgment. This work was supported by a Allen and
Lila Self Fellowship (to B.L.M.) and NIH (PO1AG12993-01).
We wish to thank Mr. Robert Drake for his efforts in acquiring
the GC-MS data, and we would like to express our thanks to
one reviewer for valuable suggestions on the chemistry of
intermediate 10 in aqueous solution.
and 9.0 (system II), the oxidation of DMS by CBP generates
12 with ΦH[3CBP] ) 5.04 × 10-7 M s-1 (reactions 15 and
51). Reactions 51-53 imply a stoichiometry of one molecule
formaldehyde per molecule 12 so that about 5.04 × 10-7
s-1 of the total observed formaldehyde yields (7.3 × 10-7
M
M
s-1) can be accounted for by the initial reaction 15. This leaves
a residual amount of ∆H CO ) 2.26 × 10-7 M s-1 of
2
Supporting Information Available: Experimental section
(4 pages). See any current masthead page for ordering and
Internet access instructions.
formaldehyde representative for other pathways. Between pH
6.2 and 9.0 an average of 1.02 × 10-6 M s-1 of DMSO is
formed via the SOD-dependent pathway which has a stoichi-
ometry of 2 molecules of DMSO per intermediate 10, i.e., per
2a. Thus, a fraction of 0.51 × 10-6 M s-1 2a forms SOD-
dependent DMSO. However, 2a is generated with Φel[3CBP]
) 7.14 × 10-7 M s-1, so that the difference between
Φel[3CBP] and 0.51 × 10-6 M s-1 2a, a residual yield of ∆2a
) 2.04 × 10-7 M s-1, reacts through other pathways. As
discussed above a small fraction of ∆2a will form SOD-
independent DMSO yields via the hydroxide pathway (at pH
JA962032L
(39) Bobrowski, K.; Holcman, J. J. Phys. Chem. 1989, 93, 6381-6387.
(40) Hug, G. L.; Marciniak, B.; Bobrowski, K. J. Photochem. Photobiol.
A 1996, 95, 81-88.
(41) Scho¨neich, Ch.; Zhao, F.; Yang, J.; Miller, B. L. In Stability,
Formulation and DeliVery of Peptide and Proteins; American Chemical
Society Symposium Series 1996, in press.
(42) Michaelis, M. L.; Bigelow, D. J.; Scho¨neich, Ch.; Williams, T. D.;
Ramonda, L.; Yin, D.; Hu¨hmer, A. F. R.; Yao, Y.; Gao, J.; Squier, T. C.
Life Sci. 1996, 59, 405-412.