A kinetic approach to characterize the electrostatic environments of thiol groups in proteins

Article abstract of DOI:10.1006/bioo.1998.1112

In this study, we synthesized a zwitterionic DTNB derivative, 5-(2- aminoethyl)-dithio-2-nitrobenzoate (ADNB), and characterized its reactions with several cationic, anionic, and neutral thiols. Reactions with ADNB, unlike those with DTNB, are relatively insensitive to electrostatic environments and ionic strengths. At relatively low ionic strength, rate ratios, k(ADNB)/k(DTNB), varied from 0.22 for reactions with low-molecular- weight cationic thiols to 3.0 for those with low-molecular-weight anionic thiols. A k(ADNB)/k(DTNB) ratio of ~200 for Cys-34 of BSA appears to reflect a very anionic environment. k(ADNB)/k(DTNB) ratios of ~6 and ~1, respectively, for canine and equine serum albumins, which have Glu-82 → Asp and Glu-82 → Ala substitutions suggest Glu-82 is the most important anionic residues affecting the reactivity of Cys-34 in BSA. k(ADNB)/k(DTNB) ratios appear to be useful for characterizing electrostatic environments of thiol groups in proteins.

Full text of DOI:10.1006/bioo.1998.1112

BIOORGANIC CHEMISTRY 26, 356–364 (1998)
ARTICLE NO. BH981112
A Kinetic Approach to Characterize the Electrostatic
Environments of Thiol Groups in Proteins
Hao Zhang, Min Le, and Gary E. Means1
Department of Biochemistry, The Ohio State University, 484 West 12th Avenue,
Columbus, Ohio 43210
Received June 29, 1998
In this study, we synthesized a zwitterionic DTNB derivative, 5-(2-aminoethyl)-dithio-2-
nitrobenzoate (ADNB), and characterized its reactions with several cationic, anionic, and
neutral thiols. Reactions with ADNB, unlike those with DTNB, are relatively insensitive to
electrostatic environments and ionic strengths. At relatively low ionic strength, rate ratios,
kADNB/kDTNB, varied from 0.22 for reactions with low-molecular-weight cationic thiols to 3.0
for those with low-molecular-weight anionic thiols. A kADNB/kDTNB ratio of ȁ200 for Cys-34
of BSA appears to reflect a very anionic environment. kADNB/kDTNB ratios of ȁ6 and ȁ1,
respectively, for canine and equine serum albumins, which have Glu-82 Ǟ Asp and
Glu-82 Ǟ Ala substitutions suggest Glu-82 is the most important anionic residues affecting
the reactivity of Cys-34 in BSA. kADNB/kDTNB ratios appear to be useful for characterizing
electrostatic environments of thiol groups in proteins.
1998 Academic Press
INTRODUCTION
Thiol groups are required for the catalytic activities of many enzymes, are found
in ligand and receptor binding sites, and are required for the normal structure and
functional properties of many proteins. They vary in reactivity due to differences
in their accessibility, their proximity to charged groups and other features of their
environments.
5,5Ј-Dithiobis(2,2Ј-nitrobenzoate) (DTNB)2 is one of the most widely used re-
agents to determine thiol groups and to determine the effects of thiol group modifi-
cation on the properties of proteins (1). It reacts rapidly with most simple thiols
and ‘‘exposed’’ thiol groups of proteins to produce an easily detected equivalent
of 5-thio-2-nitrobenzoate (TNB) (Scheme 1). Slow reaction or failure to react is
usually attributed to the inaccessibility of a thiol group or to its being in an anionic
environment (2, 3).
Mixed disulfides produced upon the reaction of DTNB with various thiols have
also been used, in a few cases, to detect and to modify thiol groups of proteins
1 To whom correspondence should be addressed. E-mail: means.1@osu.edu.
2 Abbreviations used: DTNB, 5,5Ј-dithiobis(2,2Ј-nitrobenzoate); TNB, 5-thio-2-nitrobenzoate; ADNB,
5-(2-aminoethyl)dithio-2-nitrobenzoate; BSA, bovine serum albumin; ESA, equine serum albumin; CSA,
canine serum albumin; MEA, 2-mercaptoethylamine; GSH, glutathione; CYS, cysteine; HOMO, homo-
cysteine; TGA, thioglycolic acid; MPA, 3-mercaptopropionic acid; ͱME, ͱ-mercaptoethanol; DTT, dithi-
othreitol.
356
0045-2068/98 $25.00
Copyright 1998 by Academic Press
All rights of reproduction in any form reserved.

THIOL GROUP ENVIRONMENTS IN PROTEINS
357
SCHEME 1
(Scheme 2). 5-(n-Octyl) dithio-2-nitrobenzoate, for example, reacts more rapidly
than DTNB with some protein thiol groups and is thought to reflect their being in
hydrophobic environments (4–6).
To determine whether other mixed disulfide derivatives of DTNB might be used
to detect or characterize thiol groups in other unique environments, we synthesized
and studied the reactions of 5-(2-aminoethyl)dithio-2-nitrobenzoate (ADNB) with
a series of low-molecular-weight thiols and three closely related protein thiols.
Because ADNB is a zwitterion, with little or no net charge under the conditions
normally used, we anticipated that its reactions would be less sensitive to the ionic
environments of the thiol groups with which it reacts. A comparison of its reactivity
with a series of differently charged thiols to that of DTNB suggests that kADNB
/
kDTNB ratios may be useful for characterizing the ionic environments of thiol groups
in proteins.
MATERIALS AND METHODS
Materials
Thioglycolic acid (TGA), 3-mercaptopropoinic acid (MPA), and ͱ-mercaptoetha-
nol (ͱME) were obtained from Aldrich Chemical Co. 2-Mercaptoethylamine/HCl
(MEA), glutathione (GSH), cysteine/HCl (Cys), homocysteine (HOMO), DTNB,
bovine serum albumin (BSA, Fraction V), equine serum albumin (ESA, Fraction
V), and canine serum albumin (CSA, Fraction V) were obtained from Sigma Chemi-
cal Co. To increase their thiol contents, serum albumins were treated at pH 7.4 for
30 min with a five-fold molar excess of dithiothreitol (DTT) and passed through a
Sephadex G-50 column (7, 8).
Preparation and Characterization of ADNB
DTNB was dissolved in H2O to a final concentration of 0.126 M and adjusted to
pH 7 with NaOH. One equivalent of MEA, also neutralized with NaOH, was added
and mixed immediately (Scheme 3). After 1 h, the precipitate was collected by
SCHEME 2

358
ZHANG, LE, AND MEANS
SCHEME 3
vacuum filtration, washed three times with cold water, and dried. The resulting
solid was redissolved in hot water and slowly cooled to room temperature. The
light yellow crystals were collected, washed with cold water, and dried. The resulting
product (60–70% yield) had a melting point of 215 to 221ЊC (decomp.), gave a
single peak upon C-18 RP-HPLC and a single M ϩ 1 peak (m/z 275) upon FAB
mass spectroscopy. Its uv–visible spectrum (Fig. 1), determined with a HP 8452A
diode array spectrophotometer, showed a peak at 328 nm with an extinction coeffi-
FIG. 1. The uv–vis spectrum of 45 ȐM ADNB in 50 mM phosphate buffer containing 1 mM EDTA
at pH 7.4 (———) and after reaction with 200 ȐM MEA (—ии—).

THIOL GROUP ENVIRONMENTS IN PROTEINS
359
TABLE 1
Second Order Rate Constants for the Reactions of ADNB and DTNB with Small Thiolsa
1
1
1
1
RSH
kADNB (MϪ sϪ
)
kDTNB (MϪ sϪ
)
kADNB/kDTNB
Net charge
MEA
HOMO
ͱME
CYS
GSH
MPA
TGA
1020
220
110
550
300
38
4600
450
220
940
270
20
0.22
0.50
0.50
0.58
1.11
1.90
3.03
ϩ1
0
0
0
Ϫ1
Ϫ1
Ϫ1
94
31
a Reactions were carried out using 5 ȐM of each thiol and a large excess of ADNB or DTNB in 10
mM MES, 1 mM EDTA, at pH 6.65. Reactions were followed by monitoring absorbance changes at
412 nm.
1
1
cient of ȁ8730 MϪ cmϪ . Upon addition of excess MEA, a new peak was formed
with a maximum at 412 nm, corresponding to exactly one equivalent of TNB based
1
1
on its absorbance at 412 nm (␧ ϭ 13,600 MϪ cmϪ ) (1). ADNB stock solutions up
to 5 mM were prepared by dissolving ADNB in a small amount of 1 M HCl and
diluting it with 20 mM phosphate buffer at pH 6.65.
Kinetic Studies
Rate constants for the reactions of ADNB and DTNB with various thiols were
determined as follows: 5 ȐM of thiol or thiol protein was added and rapidly mixed
with excess (48 to 240 ȐM) of ADNB or DTNB in 10 mM MES buffer containing
1 mM EDTA at pH 6.65. The reactions were followed at 412 nm for at least eight
half-lives. Pseudo-first-order rate constants, kobs , were calculated according to the
relationship ln(At/A0) ϭ Ϫkobst. Values of kobs at different concentrations of either
ADNB or DTNB were plotted against reagent concentration to get second-order
rate constants. Ionic strength dependencies were determined in 10 mM Mes,
1 mM EDTA, pH 6.65, with increasing amounts of NaCl. pH dependencies were
determined using different buffers (sodium acetate at pH 5.3, sodium phosphate
at pH 6 to 7, and Tris/HCl at pH 7 to 8) in the presence of 1 mM EDTA and an
ionic strength of 0.02.
RESULTS AND DISCUSSION
ADNB, a zwitterionic mixed disulfide derivative of DTNB (Scheme 3), can be
synthesized and purified in good yield due to its relatively low solubility at neutral
pH. With little or no net charge at physiological pH, its reactivity with charged
thiols and thiol groups located in the vicinity of charged groups of proteins are
expected to differ from those of DTNB.
To evaluate the influence of charge on its reactivity, we compared its reactions
with seven low-molecular-weight thiols with different charge characteristics to those
of DTNB with the same thiols (Table 1). As with DTNB, reactions of ADNB

360
ZHANG, LE, AND MEANS
FIG. 2. Plots of pseudo-first-order rate constants for the reactions of various thiols versus ADNB
concentration. MEA (᭹), CYS (᭿), MPA (᭡), and BSA (᭺), 5 ȐM of each, were allowed to react with
a large excess of ADNB in 10 mM MES buffer, 1 mM EDTA, pH 6.65, followed by monitoring
absorbance changes at 412 nm.
produced one equivalent of easily detected TNB (Fig. 1). Reaction rates were first
order with respect to both reactants (Fig. 2) and increased from pH 5.3 to 8.5 in
direct proportion to concentrations of thiolate ions (data not shown), in accord
with a simple nucleophilic mechanism (9, 10). As expected, its reactions with neutral
(HOMO, ͱME, and CYS) and cationic thiols (MEA) were somewhat slower than
those of DTNB. Those with anionic thiols (GSH, MPA, and TGA), however, were
slightly faster (Table 1). The effects of charge are revealed more obviously by the
rate ratios, kADNB/kDTNB , which varied from 0.22, for reactions with 2-mercapto-
ethylamine, a cationic thiol, to about 0.5 for neutral thiols and to about 3 for the
thioglycolate anion, as shown in Table 1.
The effects of added salt on the reactions of ADNB and DTNB with an anionic,
a neutral, and one cationic thiol were quite different, as shown in Fig. 3. Reactions
of DTNB, which is a dianion under the conditions employed, with MPA anion, for
example, were strongly accelerated by the addition of salt. Those with ͱME, a
neutral thiol, also increased slightly, whereas those with the MEA cation were
slightly decreased. Reactions of ADNB, which is approximately neutral under the
same conditions, however, were all only slightly increased by the addition of salt.
With little or no net charge, its reactions are, thus, much less sensitive to nearby

THIOL GROUP ENVIRONMENTS IN PROTEINS
361
FIG. 3. Ionic strength dependencies for the reactions of thiols with DTNB and ADNB. Thiols, 5 ȐM,
were allowed to react with 120 mM ADNB (filled symbols) or DTNB (open symbols) in 10 mM MES,
1 mM EDTA, pH 6.65. Ionic strengths of solutions were adjusted by adding NaCl. MEA (᭹, ᭺), ͱME
(᭿, ᮀ), MPA (᭡ ᭝).
charged groups and to ionic strength than those of DTNB. It should, therefore, be
more useful for modifying thiol groups in anionic environments. Ratios of its reactiv-
ity as compared to those of DTNB may be a useful measure of the ionic environments
of thiol groups in proteins.
Serum albumin, the most abundant protein and thiol in plasma, has at least three
major organic ligand binding sites and is known to interact strongly with, and affect
the bioavailability of, many drugs and physiological substances (11, 12). Its single
thiol group, Cys-34, is thought to react with heavy metal cations, various oxidants,
and free radicals and to serve as a vascular carrier and/or reservoir for nitric oxide
(13, 14). X-ray structures of human serum albumin show that the thiol group is
located in a small crevice on the surface of the protein in close proximity to
His-39 and the carboxyl moiety of Glu-82 (15). Although partially exposed to
solvent, it reacts relatively slowly with anionic reagents, like DTNB (3), consistent
with its being in an anionic environment. The adjacent side chain of Glu-82 is
presumably the most important anionic group in its environment.
Under the conditions used to compare reactions of DTNB and ADNB with low-
molecular-weight thiols (Table 1), Cys-34 of BSA reacts very slowly with DTNB,
but approximately 200 times faster with ADNB (Table 2). Reaction rates were first

362
ZHANG, LE, AND MEANS
TABLE 2
Second Order Rate Constants for the Reactions of ADNB and DTNB with Serum Albumins
from Different Speciesa
1
1
1
1
Albumins
kADNB (MϪ sϪ
)
kDTNB (MϪ sϪ
)
kADNB/kDTNB
Residue-82
BSA
CSA
ESA
208
139
69.7
1.1
22.9
70.1
190
6.1
1.0
E
D
A
a Reactions were carried out using 5 ȐM of albumin (basded on thiol content) and a large excess of
ADNB or DTNB. Other conditions were the same as described in Table 1.
order with respect to either DTNB (3) or ADNB and with respect to BSA (Fig.
2). As in the case of DTNB, rates increased rapidly with pH, reflecting an apparent
dependence on the thiolate anions. The rate ratio, kADNB/kDTNB , of ȁ200, however,
was much larger than observed for any low-molecular-weight thiol (Table 1) and
appeared to reflect a very intense anionic charge in the vicinity of Cys-34.
Cys-34, His-39, and the surrounding crevice appear to be highly conserved in all
mammalian serum albumins. The adjacent Glu-82 carboxyl group in BSA, however,
is replaced by Ala-82 in ESA and by Asp-82 in CSA (11). The presence, absence,
or the small displacement of that anionic group, observed in these three different
albumins, should have quite different effects on the reactivity of Cys-34 with anionic
DTNB as compared to zwitterionic ADNB. Although slightly farther away,
Glu-38, which is replaced by Asp-38 in both ESA and CSA, may also have some
effect on the reactivity of Cys-34.
As shown in Table 2, the reaction of DTNB with ESA is approximately 70 times
faster than that with BSA and the kADNB/kDTNB ratio of ȁ1 is comparable to that
observed for glutathione and other low-molecular-weight thiols (Table 1). The
reactivity of DTNB with CSA, which has the same number of anionic groups in
the immediate vicinity of Cys-34 but arranged differently, was intermediate between
those of BSA and ESA. Its kADNB/kDTNB ratio of ȁ6 was, likewise, between those
of BSA and ESA but greater than observed for any low-molecular-weight thiol
(Table 1). The very high kADNB/kDTNB ratio (ȁ200) in the case of BSA thus appears
to be due largely to Glu-82 and to reflect both a strong anionic environment and,
probably, less conformational freedom in its transition state as compared to any of
the low-molecular-weight thiols. While disulfide exchange reactions are generally
quite sensitive to electrostatic environments (16), those involving proteins, wherein
the participating sulfur atoms may be somewhat more constrained and perhaps in
less polar environments, are likely to be even more sensitive than those between
low-molecular-weight thiols and disulfides. Relatively small differences in the ionic
environments of Cys-34 residues in different serum albumins, for example, have
relatively dramatic effects on their kADNB/kDTNB ratios.
Reactions of DTNB with BSA are known to be strongly accelerated by the
addition of salt due, presumably, to the strong anionic environment of Cys-34 (3).
Reactions of ADNB with BSA are much faster than those with DTNB. They are

THIOL GROUP ENVIRONMENTS IN PROTEINS
363
FIG. 4. Ionic strength dependencies for the reactions of BSA with DTNB and ADNB. BSA, 5 ȐM
of thiol content, was reacted with 120 ȐM ADNB (᭹) or DTNB (᭿) in 10 mM MES, 1 mM EDTA,
pH 6.65. Ionic strengths of solutions were adjusted by adding NaCl.
less sensitive to salt but still affected somewhat more than those of low-molecular-
weight thiols (Fig. 4). Specific interactions of ADNB, which still possesses two
oppositely charged groups, with the anionic environment of Cys-34 presumably
account for that slightly greater effect of salt.
Due to its high reactivity with simple thiols, with most thiol groups in proteins,
and the ease with which those reactions can be followed, DTNB is one of the most
widely used reagents for thiol group modification. ADNB also reacts readily with
both low-molecular-weight and protein thiol groups, about half as fast in the case
of simple neutral thiols (Table 1), and produces an equivalent of the same easily
detectable TNB anion (Scheme 2). Due to their different electrostatic charges,
however, ADNB differs from DTNB in its reactivity with thiol groups in ionic
environments. Whereas DTNB does not react readily with thiol groups in anionic
environment, ADNB does and can be used to distinguish such thiol groups from
those that fail to react for other reasons, so-called ‘‘buried thiol groups.’’ ADNB
may be useful for selectively modifying thiol groups in anionic environments, for
detecting them, and for comparing anionic charge intensities around thiol groups
on the basis of their kADNB/kDTNB ratios. The latter should be particularly useful
for thiol groups in closely related proteins, like the serum albumins of different
species, families of proteins produced by site directed mutagenesis, etc.

364
ZHANG, LE, AND MEANS
ADNB is very easy to prepare, reacts rapidly and selectively with thiol groups,
and is useful for characterizing thiol group environments in proteins due to its
significantly different selectivity as compared to DTNB.
ACKNOWLEDGMENT
This work was supported by the American Heart Association, Ohio–West Virginia Affiliate,
Columbus, Ohio.
REFERENCES
1. Ellman, G. L. (1959) Arch. Biochem. Biophys. 82, 70–77.
2. Jocelyn, P. C. (1987) Methods Enzymol. 143, 44–67.
3. Wilson, J. M., Wu, D., Motiu-DeGrood, R., and Hupe, D. J. (1980) J. Am. Chem. Soc. 102, 359–363.
4. Faulstich, H., and Heintz, D. (1995) Methods Enzymol. 251, 357–366.
5. Gargouri, Y., Moreau, H., Pieroni, G., and Verger, R. (1988) J. Biol. Chem. 263, 2159–2162.
6. Faulstich, H., Tews, P., and Heintz, D. (1993) Anal. Biochem. 208, 357–362.
7. Sogami, M., Nagoka, S., Era, S., and Noguchi, K. (1984) Int. J. Pept. Protein Res. 24, 96–103.
8. Le, M., and Means, G. E. (1995) Anal. Biochem. 229, 264–271.
9. Wilson, J. M., Bayer, R. J., and Hupe, D. J. (1977) J. Am. Chem. Soc. 99, 7922–7926.
10. Whitesides, G. M., Lilburn, J. E., and Szajewski, R. P. (1977) J. Org. Chem. 42, 332–338.
11. Carter, D. C., and Ho, J. X. (1994) Adv. Protein Chem. 45, 385–391.
12. Peters, T. (1996) All About Albumin, Academic Press, New York.
13. Stamler, J. S., Jaraki, O., Osbourne, J., Simon, D. I., Keaney, J., Vita, J., Singel, D., Valeri, C. R.,
and Loscalzo, J. (1992) Proc. Natl. Acad. Sci. USA 89, 7674–7677.
14. Zhang, H., and Means, G. E. (1996) Anal. Biochem. 237, 141–144.
15. He, X.-M., and Carter, D. C. (1992) Nature (London) 358, 209–215.
16. Bulaj, G., Kortemme, T., and Goldenberg, D. P. (1998) Biochemistry 37, 8965–8972.