Detection of the sulfhydryl groups in proteins with slow hydrogen exchange rates and determination of their proton/deuteron fractionation factors using the deuterium-induced effects on the 13Cβ NMR signals

Article abstract of DOI:10.1021/ja101205j

A method for identifying cysteine (Cys) residues with sulfhydryl (SH) groups exhibiting slow hydrogen exchange rates has been developed for proteins in aqueous media. The method utilizes the isotope shifts of the C _β chemical shifts induced by

Full text of DOI:10.1021/ja101205j

Published on Web 04/12/2010
Detection of the Sulfhydryl Groups in Proteins with Slow
Hydrogen Exchange Rates and Determination of Their Proton/
Deuteron Fractionation Factors Using the Deuterium-Induced
Effects on the 13Cꢀ NMR Signals
Mitsuhiro Takeda,† JunGoo Jee,‡ Tsutomu Terauchi,‡ and Masatsune Kainosho*,†,‡
Structural Biology Research Center, Graduate School of Science, Nagoya UniVersity, Furo-cho,
Chikusa-ku, Nagoya, 464-8602, Japan, and Center for Priority Areas, Graduate School of
Science and Technology, Tokyo Metropolitan UniVersity, 1-1 Minami-ohsawa,
Hachioji, 192-0397, Japan
Received February 17, 2010; E-mail: kainosho@nmr.chem.metro-u.ac.jp
Abstract: A method for identifying cysteine (Cys) residues with sulfhydryl (SH) groups exhibiting slow
hydrogen exchange rates has been developed for proteins in aqueous media. The method utilizes the
isotope shifts of the Cꢀ chemical shifts induced by the deuteration of the SH groups. The 18.2 kDa E. coli
peptidyl prolyl cis-trans isomerase b (EPPIb), which was selectively labeled with [3-13C;3,3-2H2]Cys, showed
much narrower line widths for the 13Cꢀ NMR signals, as compared to those of the proteins labeled with
either [3-13C]Cys or (3R)-[3-13C;3-2H]Cys. The 13Cꢀ signals of the two Cys residues of EPPIb, i.e. Cys-31
and Cys-121, labeled with [3-13C;3,3-2H2]Cys, split into four signals in H2O/D2O (1:1) at 40 °C and pH 7.5,
indicating that the exchange rates of the side-chain SH’s and the backbone amides are too slow to average
the chemical shift differences of the 13Cꢀ signals, due to the two- and three-bond isotope shifts. By virtue
of the well-separated signals, the proton/deuteron fractional factors for both the SH and amide groups of
the two Cys residues in EPPIb could be directly determined, as approximately 0.4-0.5 for [SD]/[SH] and
0.9-1.0 for [ND]/[NH], by the relative intensities of the NMR signals for the isotopomers. The proton NOE’s
of the two slowly exchanging SH’s were clearly identified in the NOESY spectra and were useful for the
determining the local structure of EPPIb around the Cys residues.
Introduction
To investigate protein structures in solution at atomic
resolution, NMR is undoubtedly the most powerful method.
However, the acquisition of structural information for various
polar side-chain groups such as hydroxyl or amino groups,
which are biologically and structurally important, has been
hampered due to their intrinsically fast exchange rates with the
hydrogens of the surrounding solvent molecules.8-11 This is
also the case for the SH groups of Cys residues, which are rarely
observed, except for those deeply embedded in the interior of
proteins.5,6 Therefore, this important polar side-chain functional
group has drawn little attention from the biological NMR
community, and the number of assigned SH proton NMR signals
of Cys residues deposited in the Biological Magnetic Resonance
Data Bank (BMRB)12 (http://www.bmrb.wisc.edu/) is currently
less than 60. Moreover, there is presently no direct method to
determine the fractionation factors, even for Cys residues in
proteins with very slow hydrogen exchange rates.
It is well-known that the SH groups of Cys residues in proteins
often play crucial roles in their biological functions.1,2 The SH
groups form hydrogen bonds and bind to various metal ions,
thereby stabilizing the secondary and tertiary structures of proteins.3,4
Another unique feature of SH groups, which distinguishes them
from the other polar side-chain functional groups, is their low
proton/deuteron fractionation factors ([SD]/[SH] ≈ 0.5).5,6 There-
fore, the SH groups of Cys residues preferentially bind to protons,
rather than deuterons, in a mixture of H2O/D2O.5,6 Since detailed
enzymatic mechanisms are often studied by measuring the accurate
kinetic isotope effects in D2O-containing solutions, the fractionation
factors of the catalytic SH groups in proteins are crucial parameters
to be considered.7
† Nagoya University.
‡ Tokyo Metropolitan University.
(1) Giles, N. M.; Giles, G. I.; Jacob, C. Biochem. Biophys. Res. Commun.
2003, 300, 1–4.
(7) Balasco, J. G.; Bruice, T. W.; Albery, W. J.; Knowles, J. R.
Biochemistry 1986, 25, 2558–2564.
(2) Johnson, D. C.; Dean, D. R.; Smith, A. D.; Johnson, M. K. Annu.
ReV. Biochem. 2005, 74, 247–81.
(8) Otting, G.; Wu¨thrich, K. J. Am. Chem. Soc. 1989, 111, 1871–1875.
(9) Otting, G.; Liepinsh, E.; Wu¨thrich, K. J. Am. Chem. Soc. 1991, 113,
4363–4364.
(3) Gregoret, L. M.; Rader, S. D.; Fletterick, R. J.; Cohen, F. E. Proteins
1991, 9, 99–107.
(10) Liepinsh, E.; Otting, G.; Wu¨thrich, K. J. Biomol. NMR 1992, 2, 447–
465.
(4) Zhou, P.; Tian, F.; Lv, F.; Shang, Z. Proteins 2009, 76, 151–63.
(5) Jarret, R. M.; Saunders, M. J. Am. Chem. Soc. 1985, 107, 2648–2654.
(6) (a) Gold, V. AdV. Phys. Org. Chem. 1969, 7, 259–331. (b) Anslyn,
E. V.; Dougherty, D. A. Modern Physical Organic Chemistry,
University Science Books: Sausalito, CA, 2006; pp 437-441.
(11) Liepinsh, E.; Otting, G. Magn. Reson. Med. 1996, 35, 30–42.
(12) Doreleijers, J. F.; Mading, S.; Maziuk, D.; Sojourner, K.; Yin, L.;
Zhu, J.; Markley, J. L.; Ulrich, E. L. J. Biomol. NMR 2003, 26, 139–
146.
9
6254 J. AM. CHEM. SOC. 2010, 132, 6254–6260
10.1021/ja101205j  2010 American Chemical Society

Detection of the Sulfhydryl Groups in Proteins
A R T I C L E S
Scheme 1. Synthesis of L-[3-13C;3,3-2H2]Cysteinea
a Reagents and conditions: (a) BnOH, HCl gas, 97%; (b) (i) N-(diphenylmethylene)glycine tert-butyl ester, chiral-PTC, 50% aqueous KOH, toluene; (ii)
0.5 M citric acid, dioxane, 76%; (c) (i) Boc2O, DMAP, DMF, (ii) Pd/C, H2 (4 atm), 50 °C, 76%; (d) TsCl, pyridine, 75%; (e) AcSK, DMF, 87%; (f) TFA,
room temperature, then HCl, ∆, 73%.
Preparation of E. coli Peptidyl-Prolyl cis-trans Isomerase b,
EPPIb, Selectively Labeled with Isotope-Labeled Amino Acids. The
EPPIb’s selectively labeled with each of the isotope-labeled
cysteines or [1-13C]tyrosine were prepared by the E. coli cell-free
protein expression system, as described previously.18-20 A 5 mL
reaction mixture, consisting of the cell extract from E. coli BL21
Star (DE3) (Invitrogen) and all other components, was placed in a
dialysis tube and dialyzed against 20 mL of outer medium for 8 h
at 37 °C. A total of 75 mg of the amino acid mixture, consisting of
about 2 mg each of the isotope-labeled cysteines or tyrosine and
19 other unlabeled amino acids, was used for each reaction, which
yielded 3-4 mg each of selectively labeled EPPIb’s. The obtained
labeled EPPIb’s were dissolved in 50 mM sodium phosphate buffer,
prepared from 100% H2O, 100% D2O, or H2O/D2O (1:1), respec-
tively, containing 100 mM NaCl and 0.1 mM NaN3, pH 7.5 (meter
reading).
Recently, we developed a unique method for identifying and
assigning the hydroxyl groups of tyrosine (Tyr) residues that
slowly exchange with the surrounding solvent hydrogens.13 The
most crucial prerequisite for observing and assigning the 13Cꢁ
peak Tyr in a high-resolution manner was to use the stereoarray
isotope labeled (SAIL) Tyr, (2S,3R)-[R,ꢀ,ꢁ-13C;ꢀ2,ꢀ1,2-
∼2H;15N]Tyr, which has an optimized labeling pattern for such
experiments.13,14 A similar strategy could also be applied to
observe the deuterium isotope shifts for the Cꢀ signals of Cys
residues in a protein: for example, the 18.2 kDa E. coli peptidyl
cis-trans isomerase b (EPPIb). One might expect to observe
the isotope shifts of the Cꢀ signals for the Cys residues in EPPIb
labeled with either uniformly 13C,15N-labeled Cys or SAIL Cys,
i.e. (2R,3R)-[1,2,3-13C;3-2H;15N]Cys.15 However, this was actu-
ally not the case, since the NMR line widths of the side-chain
Cꢀ atoms with directly bonded proton(s) were too broad to
observe small isotope shifts. Therefore, we optimized the isotope
labeling pattern of Cys in order to obtain narrower line widths
for the Cꢀ signals, which allowed us to identify the SH groups
of Cys residues in proteins with slow hydrogen exchange rates
and accurately determine their fractionation factors.
The sample concentrations varied from 0.2 to 0.7 mM, depending
on the available labeled proteins, but there was no concentration
dependence for the NMR spectra in this concentration range. In
the case of the 100% H2O solution, a 4.1 mm o.d. Shigemi tube
containing the protein solution was inserted into a 5 mm o.d. outer
2
tube containing pure D2O for the H lock signal.
NMR Measurements. All 1D-13C NMR spectra were measured
at 40 °C on a DRX600 spectrometer (Bruker Biospin; 150.9 MHz
for 13C), equipped with a 5 mm TCI cryogenic triple-resonance
probe. The WALTZ16 decoupling scheme was used for 1H and/or
2H during acquisition.21 The carrier frequency was set to 30 ppm
with a sweep width of 4500 Hz, and free induction decays were
acquired for 226 ms at a repetition time of 3 s. 1D-13C NMR spectra
with adequate signal-to-noise ratios were obtained within 6-24 h,
depending on the concentrations of the samples. In order to ensure
complete hydrogen-deuterium equilibration, we measured the time
course of the deuterium exchange of the sulfhydryl and amide
groups of the two Cys residues in D2O, as shown in the Supporting
Information. On the basis of this experiment, all of the protein
samples dissolved in H2O/D2O (1:1) were incubated for at least 1
day at 40 °C to ensure that the hydrogen-deuterium equilibrium
was completed before the NMR measurements.
Materials and Methods
Synthesis of Isotope-Labeled Cysteines. L-[3-13C]Cys and L-[3-
13C;3,3-2H2]Cys were synthesized from [13C]paraformaldehyde and
[13C;2H]paraformaldehyde, respectively, via the stereoselective
alkylation of the glycine-benzophenone Schiff base, using a chiral
phase-transfer catalyst ((11bR)-(-)-4,4-dibutyl-4,5-dihydro-2,6-
bis(3,4,5-trifluorophenyl)-3H-dinaphth[2,1-c:1′,2′-e]azepinium bro-
mide), as shown in Scheme 1 for L-[3-13C;3,3-2H2]Cys.16,17 After
protection of the amino group, the obtained serine derivative 3 was
converted to the cysteine derivative 6 by transforming the side-
chain moiety.15 The deprotection of compound 6 was achieved by
refluxing with 2 M HCl to give L-[3-13C;3,3-2H2]cysteine 7 as the
hydrochloride salt. The optical purity for the R-position was
determined to be 88% ee, by an HPLC analysis on a chiral stationary
column (DAICEL CROWNPAK CR+).
The 13C NMR exchange spectroscopy (EXSY) experiment was
performed for a 0.7 mM H2O/D2O (1:1) solution of the EPPIb
selectively labeled with [3-13C;3,3-2H2]Cys, with the same pulse
scheme used previously.13 During the chemical shift encoding (t1)
L-(2R,3R)-[3-13C;3-2H]Cys was synthesized by the previously
described method for preparing SAIL Cys, i.e. L-(2R,3R)-[1,2,3-
13C;3-2H;15N]Cys, using the unlabeled ethyl hippurate in lieu of
ethyl [1,2-13C;3-2H;3-15N]hippurate.15
1
and the acquisition of free induction decay (t2), decoupling on H
21
2
and H by the WALTZ16 scheme was applied. The data points
and the spectral widths were 512 (t1) × 4096 (t2) points and 1200
(13) Takeda, M.; Jee, J.; Ono, M. A.; Terauchi, T.; Kainosho, M. J. Am.
Chem. Soc. 2009, 131, 18556–18562.
(14) Kainosho, M.; Torizawa, T.; Iwashita, Y.; Terauchi, T.; Ono, A. M.;
Gu¨ntert, P. Nature 2006, 440, 52–57.
(18) Torizawa, T.; Shimizu, M.; Taoka, M.; Miyano, H.; Kainosho, M.
J. Biomol. NMR 2004, 30, 311–325.
(15) Terauchi, T.; Kobayashi, K.; Okuma, K.; Oba, M.; Nishiyama, K.;
Kainosho, M. Org. Lett. 2008, 10, 2785–2787.
(19) Takeda, M.; Ikeya, T.; Gu¨ntert, P.; Kainosho, M. Nat. Protoc. 2007,
2, 2896–2902.
(16) Connor, D. S.; Klein, G. W., Jr.; Medwid, J. B. Organic Syntheses;
House, H. O., Ed.; Wiley: New York, 1972; Vol. 52, pp 16.
(17) Kitamura, M.; Shirakawa, S.; Maruoka, K. Angew. Chem., Int. Ed.
2005, 44, 1549–1551.
(20) Kariya, E.; Ohki, S.; Hayano, T.; Kainosho, M. J. Biomol. NMR 2000,
18, 75–76.
(21) Shaka, A. J.; Keeler, J.; Frenkiel, T.; Freeman, R. J. Magn. Reson.
1969, 52, 335–338.
9
J. AM. CHEM. SOC. VOL. 132, NO. 17, 2010 6255

A R T I C L E S
Takeda et al.
Figure 1. Comparison of the Cys 13Cꢀ signals of EPPIb’s selectively labeled with [3-13C]cysteine (A), (3R)-[3-13C;3-2H2]cysteine (B), and [3-13C;3,3-
2H2]cysteine (C). The 150.9 MHz [2H,1H]-decoupled 1D-13C NMR spectra of EPPIb’s were obtained at 40 °C, pH 7.5, for 0.2-0.4 mM solutions. The
spectra in D-F were obtained in H2O, and those in G-H were in a H2O/D2O (1:1) solution.
Hz (ω1, 13C) × 9100 Hz (ω2, 13C), respectively, and the number of
scans/FID was 240. The carrier frequency of carbon was set to
25.5 ppm. The mixing time was 600 ms, and the repetition time
was 3.1 s.
frequency of carbon was set to 125.5 ppm. The NOE mixing time
for both of the 13C-edited NOESY-HSQC spectra was 145 ms.
Results and Discussion
13C NMR Line-Narrowing Effect Caused by the
Deuterium Substitution of the Methylene Protons on the Cys
13Cꢀ NMR Signals of EPPIb. Since the line widths of the 13Cꢀ
Cys residues in EPPIb, even for the SAIL EPPIb, were too broad
to resolve the isotope-shifted 13Cꢀ signals, it was necessary to
optimize the isotope-labeling pattern of Cys in order to make
the 13Cꢀ line width narrower than the signal separation due to
the isotope shift. In doing so, the effect of the proton(s) attached
to the Cꢀ on the line width was examined by synthesizing
[3-13C]Cys, (3R)-[3-13C;3-2H]Cys, and [3-13C;3,3-2H2]Cys, which
were then incorporated within EPPIb by E. coli cell-free
expression (Figure 1A-C).18–20 In these labeled cysteines, only
Cꢀ was selectively labeled with 13C, to eliminate the scalar and
dipolar effects caused by adjacent 13C nuclei, and thus the line-
broadening effect due to the hydrogen atoms directly bonded
to the 13Cꢀ atom could be precisely examined. The 1D-13C NMR
spectra of these three samples in 100% H2O showed significantly
different line widths for the 13C signals by the replacement of
either or both of the two methylene protons with a deuteron
We searched for NOE peaks involving the SH protons of Cys-
31 and Cys-121 in 15N- and 13C-edited 3D-NOESY-HSQC spectra
obtained from 0.3-0.5 mM EPPIb samples composed of 20 SAIL
amino acids with different isotope labeling patterns of the pheny-
lalanine and tyrosine residues.22 The 15N-edited 3D-NOESY-HSQC
spectra were measured with a Bruker DRX600 spectrometer
equipped with a TXI cryogenic probe. The data points and the
spectral width of the 15N-edited NOESY-HSQC were 256 (t1) ×
1
48 (t2) × 1024 (t3) points and 8400 Hz (ω1, H) × 1800 Hz (ω2,
1
15N) × 8400 Hz (ω3, H), respectively. The number of scans/FID
was 32, and the NOE mixing time was 200 ms. The repetition time
was 1 s. The 13C-edited 3D-NOESY-HSQC experiments were
performed with a DRX800 spectrometer equipped with a normal
TXI probe, and the experiments were acquired for two different
regions: the aliphatic and aromatic regions. In the first set for the
aliphatic region, the data points and the spectral width were 210
1
(t1) × 58 (t2) × 1024 (t3) points and 12 000 Hz (ω1, H) × 6700
1
Hz (ω2, 13C) × 10 000 Hz (ω3, H), respectively. The number of
scans/FID was 32. The repetition time was 1 s. The carrier
frequency of carbon was set to 40 ppm. In the second set for the
aromatic region, the data points and the spectral width were 200
1
1
(Figure 1D-F): namely, ∼18 Hz (13Cꢀ H2), ∼12 Hz (13Cꢀ H2H),
1
and e5 Hz (13Cꢀ H2) (Table 1). Note that the 13Cꢀ showed
2
(t1) × 40 (t2) × 1024 (t3) points and 12 000 Hz (ω1, H) × 5800
1
Hz (ω2, 13C) × 10 000 Hz (ω3, H), respectively. The number of
upfield chemical shifts due to the deuterium substitutions, with
an increment of 0.27 ( 0.02 ppm per deuterium. The longitu-
dinal relaxation times (T1) of the 13Cꢀ signals may become
slightly longer by the replacement of one or two of the attached
protons with a deuteron, thus resulting in the sensitivity loss
scans/FID was 48. The repetition time was 1 s. The carrier
(22) Takeda, M.; Terauchi, T.; Ono, A. M.; Kainosho, M. J. Biomol. NMR
2010, 46, 45–49.
9
6256 J. AM. CHEM. SOC. VOL. 132, NO. 17, 2010

Detection of the Sulfhydryl Groups in Proteins
A R T I C L E S
Table 1. Chemical Shifts (ppm) and Line Widths (∆ν, Hz) of the Cys 13Cꢀ Peaks in EPPIb Samplesa
100% H2O
H2O/D2O (1:1)
Cys-31 13Cꢀ
Cys-12113Cꢀ
Cys-31 13Cꢀ
Cys-12113Cꢀ
[3-13C]cysteine
28.41 (∆ν ) 18 ( 2)
28.14 (∆ν ) 12 ( 2)
27.85 (∆ν e 5)
27.65 (∆ν ) 18 ( 2)
27.40 (∆ν ) 12 ( 2)
27.12 (∆ν e 5)
n.d.
n.d.
n.d.
n.d.
(3R)-[3-13C;3-2H2]cysteine
[3-13C;3,3-2H2]cysteine
27.84 (NH/SH)
27.77 (ND/SH)
27.71 (NH/SD)
27.65 (ND/SD)
27.13 (NH/SH)
27.07 (ND/SH)
27.02 (NH/SD)
26.97 (ND/SD)
a The values in the table were obtained from the NMR spectra shown in Figure 1. n.d. indicates that the Cꢀ signals due to various isotopomers were
not resolved in H2O/D2O (1:1).
within a fixed duration. However, this was not problematic in
practice, since the significant line-narrowing effect by the
deuteration compensated for such adverse effects. The 13C NMR
spectra of these three samples were also measured in H2O/D2O
(1:1), which revealed additional line broadening due to the
isotope shifts caused by the deuteration of the exchangeable
hydrogens of the Cys residues, as discussed below. Since well-
resolved 13Cꢀ signals for Cys-31 and Cys-121 were only
observed with the EPPIb labeled with [3-13C; 3,3-2H2]Cys in
H2O/D2O (1:1) (Figure 1G-I), it was obvious that both protons
on the 13Cꢀ carbon should be deuterated to accurately observe
the isotope shifts induced by the deuterium substitution23,24 of
the Cys SH groups.
Identification of the Slowly Exchanging SH Protons and
the Fractionation Factors for the SH’s and Backbone
Amides of Cys Residues in EPPIb. The 13C NMR spectra of
the EPPIb labeled with [3-13C;3,3-2H2]Cys were also obtained
in 100% D2O, and were compared to those obtained in 100%
H2O and H2O/D2O (1:1) (Figure 2A). A comparison between
these spectra revealed that the lowest and highest field peaks
of the four well-resolved peaks in H2O/D2O (1:1), for each of
the 13Cꢀ signals for Cys-31 and Cys-121, are identical with those
observed in 100% H2O and 100% D2O, respectively (Figure
2A). Since there are two exchangeable protons for Cys, the
sulfhydryl and backbone amide groups, which may affect the
13Cꢀ chemical shifts by the deuterium substitution, the observed
Figure 2. (A) 150.9 MHz [2H,1H]-decoupled 1D 13C NMR spectra of
EPPIb, selectively labeled with [3-13C;3,3-2H2]Cys. The labeled EPPIb was
dissolved in H2O, D2O, and H2O/D2O (1:1) buffers at concentrations ranging
from 0.4 to 0.7 mM for each solution at pH 7.5 and was subjected to NMR
experiments at 40 °C. The 13Cꢀ resonances for both Cys-31 and Cys-121
were observed as the four-line signals in H2O/D2O (1:1), due to the dual
isotope effect from the backbone amide and side-chain sulfhydryl groups.
(B) Structure of the [3-13C;3,3-2H2]cysteine residue. The deuterium isotope
effects from the side-chain sulfhydryl and backbone amide groups on the
13Cꢀ nuclei are shown by red arrows. (C) 13C-EXSY spectra for the 13Cꢀ
signals of Cys-31 and Cys-121 observed for the EPPIb selectively labeled
with [3-13C;3,3-2H2]Cys, at a mixing time of 600 ms. A 0.7 mM solution
in H2O/D2O (1:1), pH 7.5, was used for the measurement at 40 °C. The
two pairs of cross peaks for Cys-121, due to the SH/SD chemical exchange,
are correlated with the diagonal peaks by red lines.
four-line signals in the H2O/D2O (1:1) buffer were caused by
the isotope shifts induced by the deuterium substitutions for
both the SH’s and backbone amides of the respective Cys
residue (Figure 2B). The slow hydrogen exchange rates observed
for both the SH and amide groups of Cys-31 and Cys-121 are
compatible with the relative solvent accessibilities of these
residues estimated from the crystal structure:25 Cys-31 HN, 0%;
Cys-31 S, 0%; Cys-121 HN, 9.1%; Cys-121 S, 0%. The two
pairwise peaks, separated by smaller isotope shifts of 0.05-0.06
ppm, were intuitively attributed to the three-bond deuterium
isotope effects on the backbone amides of the Cys residues.
To further confirm the origins of the deuterium substitution
sites that induced larger isotope shifts (0.12 ( 0.02 ppm) with
smaller fractionation factors ([D]/[H] ≈ 0.4), and smaller isotope
shifts (0.06 ( 0.01 ppm) with larger fractionation factors ([D]/
[H] ≈ 1.0), the fractionation factors of the backbone amides
were independently determined by the DEALS method. The
DEALS (deuterium-hydrogen exchange rates of amide on line
NMR signal for the (i - 1)th residue to monitor the
deuterium-hydrogen exchange rate of the amide of the ith
residue. The H/D exchange at the amide group causes a change
in the chemical shift of the carbonyl carbon.27,28 Given that the
shapes) method26-29 utilizes the line shapes of the carbonyl 13
C
(27) Kainosho, M.; Nagao, H.; Tsuji, T. Biochemistry 1987, 26, 1068–
1075.
(23) Hansen, P. E. Annu. Rep. NMR Spectrosc. 1983, 15, 105–234.
(28) Markley, J. L.; Kainosho, M. In Stable Isotope Labeling and Resonance
Assignments in Larger Proteins: NMR of Macromolecules, Roberts,
G. C. K., Ed.; Oxford University Press: New York, 1993; pp 101-
152.
(24) Dziembowska, T.; Hansen, P. E.; Rozwadowski, Z. Prog. Nucl. Magn.
Reson. Spectrosc. 2004, 45, 1–29.
(25) Edwards, K. J.; Ollis, D. L.; Dixon, N. E. J. Mol. Biol. 1997, 271,
258–265.
(29) Uchida, K.; Markley, J. L.; Kainosho, M. Biochemistry 2005, 44,
11811–11820.
(26) Kainosho, M.; Tsuji, T. Biochemistry 1982, 21, 6273–6279.
9
J. AM. CHEM. SOC. VOL. 132, NO. 17, 2010 6257

A R T I C L E S
Takeda et al.
Figure 3. DEALS experiment for the carbonyl carbon 13C NMR signals of the Tyr residues of EPPIb to monitor the amide deuterium-hydrogen exchange
rates of the Cys residues. In order to monitor the amide exchange rates for Cys-31 and Cys-121, the carbonyl carbon NMR signals of the N-terminal
neighbors of the two residues, namely Tyr-30 and Tyr-120, should be observed. For the DEALS experiment, EPPIb was dissolved in H2O/D2O (1:1) to
observe the two-bond deuterium isotope shift effect. The fractionation factors ([D]/[H]) for the amides of Cys-31 and Cys-121 were estimated by the relative
peak heights of the two isotope-shifted signals to be 0.9 and 1.0, respectively, which were identical with those obtained by the Cꢀ signals (Table 2).
Table 2. Summary of the Deuterium-Induced Isotope Shifts for the
Cꢀ Signals of Cys-31 and Cys-121 and for the C′ signals of Tyr-30
and Tyr-120 and the Fractionation Factors for the Sulfhydryl and
Amide Groups of Cys-31 and Cys-121
exchangeable group
affected C signal
isotope shift (Hz)
fractionation factor
Cys-31 SH
Cys-31 NH
Cys-31 NH
Cys-121 SH
Cys-121 NH
Cys-121 NH
Cys-31 13Cꢀ
Cys-31 13Cꢀ
Tyr -30 13C′
Cys-121 13Cꢀ
Cys-121 13Cꢀ
Tyr -120 13C′
19.2 ( 3
9.9 ( 2
0.5 ( 0.1
0.9 ( 0.1
0.9 ( 0.1
0.4 ( 0.1
1.0 ( 0.1
1.0 ( 0.1
13.5 ( 1
16.0 ( 3
8.0 ( 2
13.8 ( 1
Figure 4. Fractionation factors for the sulfhydryl groups ([SD]/[SH]) and
the backbone amides ([ND]/[NH]) of Cys-31 and Cys-121 in EPPIb
dissolved in the H2O/D2O (1:1) mixture. Line-shape analyses were
performed with the DNMR software (Bruker Biospin). For each analysis,
the line width of each isotopomer peak was set to 5 Hz. The observed 13Cꢀ
signals (black) are shown, along with the four components resolved into
each of the four isotopomer peaks: SH/NH (blue), SH/ND (brown), SD/
NH (green), and SD/ND (magenta), respectively. The relative intensity of
each signal component was adjusted to minimize the difference between
the observed (black) and simulated spectra (red).
H/D exchange is slow, as compared to the size of the isotope
shift effect, the carbonyl signal appears as a doubled peak. In
contrast, when the H/D exchange is rapid, as compared to the
size of the isotope shift effect, the carbonyl signal gives one
averaged signal. In the case of the backbone amides of the two
Cys residues in EPPIb, i.e. Cys-31 and Cys-121, their amide
exchange rates can be monitored by the carbonyl 13C NMR
signals of their N-terminal neighbors, i.e. Tyr-30 and Tyr-120,
respectively, in an H2O/D2O (1:1) solution (Figure 3). Obvi-
ously, both of the amides of the two Cys residues exhibited
slow exchange rates, and the fractionation factors were almost
unity ([D]/[H] ≈ 1.0), as shown by the relative intensities of
the isotope-shifted carbonyl carbon signals of Tyr-30 and Tyr-
120, respectively. Therefore, the larger splitting, namely 0.11
and 0.13 ppm, is due to the two-bond isotope effect on the
slowly exchanging SH groups for Cys-31 and Cys-121, respec-
tively, which can be rationalized by their burial in the EPPIb
structure. Note that the two-bond SH/SD isotope shift effect
on the 13Cꢀ chemical shifts is almost identical with that for the
13Cꢁ of the Tyr residue induced by the deuteration of OH, which
was ∼0.13 ppm.13 We also performed a preliminary EXSY
experiment for the sample dissolved in an H2O/D2O (1:1)
solution. Since the 13Cꢀ signals for both Cys-31 and Cys-121
split into four peaks due to the dual isotope effect, the
experiments were very time-consuming, and thus we only
measured the EXSY spectrum at a mixing time of 600 ms, in
order to identify the exchangeable NH/ND and SH/SD. On the
basis of the data shown in Figure 2C, we could estimate the
approximate hydrogen exchange rate for the SH group of Cys-
121 to be kex ≈ 0.9 ( 0.3 s-1. In addition, the fastest limit for
the hydrogen exchange rates for the SH of Cys-31 was estimated
9
6258 J. AM. CHEM. SOC. VOL. 132, NO. 17, 2010

Detection of the Sulfhydryl Groups in Proteins
A R T I C L E S
the auxiliary hydrogen bond between the side-chain sulfhydryl
group and the backbone carbonyl oxygen seems to contribute
toward stabilizing the R-helices.4 This bifurcated hydrogen bond
might be responsible for the slow hydrogen exchange rates for
the sulfhydryl group of Cys-31, as shown by the lack of cross
peaks in the EXSY experiment, even at a 600 ms mixing time
(Figure 2C).
The proton/deuteron fractionation factors for the backbone
amides and side chains of the sulfhydryl groups were quanti-
tatively determined for the Cys residues, by measuring the peak
areas by a line-shape analysis of the four-line 13C signals
observed for the EPPIb labeled with [3-13C;3,3-2H2]Cys dis-
solved in the 1:1 H2O/D2O (1:1) solution (Figure 4). The isotope
shift values and the fractionation factors obtained by the line-
shape analyses are summarized in Table 2. The SH groups of
the Cys residues in EPPIb exhibited fractionation factors around
∼0.4-0.5, indicating that their SH groups have stronger affinity
toward a proton over a deuteron, relative to its content in a
solution.5,6 Further accumulation of such data for SH groups
will be essential to correlate the fractionation factors to their
microenvironments.
Figure 5. Hydrogen bonds involving the sulfhydryl groups of Cys-31 (A)
and Cys-121 (B). Hydrogen, oxygen, nitrogen, and sulfur atoms are shown
by white, red, blue, and yellow balls, respectively. The deduced position
of the sulfhydryl hydrogen for each of the two Cys residues is shown by a
gray circle, and the experimentally observed NOE’s between the sulfhydryl
groups and the nearby protons are indicated by arrows. This figure was
produced by using MolMol software.32
to be kex < 0.3 s-1, and that for the amide groups of Cys-31 and
13
Cys-121 was kex < 0.2 s-1
.
The absence of SH/SD exchange
cross peaks for Cys-31 implies that the hydrogen bond was
formed between the sulfhydryl group of Cys-31 (ith residue)
and the backbone carbonyl oxygen of Phe-27 ((i - 4)th residue).
Since the amide hydrogen of Cys-31 also forms a hydrogen
bond to the carbonyl oxygen of Phe-27 in the R-helix, this extra
hydrogen bond results in the bifurcated hydrogen bond between
Cys-31 and Phe-27. Similar bifurcated hydrogen bonds are
frequently found in the Cys residues involved in R-helices, and
NOE’s Involving the SH Protons of Cysteine Residues in
EPPIb. Intuitively, a discrete SH proton signal could be
observable in H2O for the Cys residues with SH groups that
have a slow hydrogen exchange rate, thereby yielding two
separate 13Cꢀ signals in the H2O/D2O (1:1) mixture due to the
Figure 6. Strips of the 15N- and 13C-edited 3D-NOESY-HSQC spectra containing NOE’s associated with the side-chain sulfhydryl protons of Cys-31 and
Cys-121.
9
J. AM. CHEM. SOC. VOL. 132, NO. 17, 2010 6259

A R T I C L E S
Takeda et al.
SH and SD species.9 This was actually the case for the two
Cys residues in EPPIb. The two signals at 2.17 and 1.24 ppm
were identified by NOESY data obtained for the EPPIb fully
labeled with SAIL amino acids, including (3R)-[ul-13C;3-
2H;15N]Cys, SAIL Cys,15 which were assigned to the SH groups
of Cys-31 and Cys-121, respectively (Figure 5). These two peaks
could also be identified by the HSQC-TOCSY experiment for
the EPPIb selectively labeled with (3R)-[3-3C;3-2H]Cys, which
and amide groups are involved in hydrogen bonds. As a
consequence, the lifetimes of the H/D atoms on the SH/SD and
NH/ND groups were long enough to give discrete 13Cꢀ signals,
and thus the fractionation factors of both the amide and
sulfhydryl hydrogen atoms could be directly determined by the
relative peak intensities of the discrete 13Cꢀ signals. However,
the present method can also be applied to the surface SH groups,
which generally have much faster hydrogen exchange rates and
thus the signals split by the isotope shifts tend to be coalesced.
In such cases, the fractionation factors could be estimated by
measuring the averaged chemical shifts in solvents with different
H/D ratios, as often used for small organic molecules.5 Since
the hydrogen exchange rate and the fractionation factor are
presumably correlated with the solvent accessibility and the
hydrogen bond strength,6,30,31 respectively, this approach will
provide various new possibilities to study the microenvironments
of Cys residues in proteins. Note that the proteins labeled with
isotope-labeled cysteines33 can be prepared by either cell-free
protein synthesis or the standard E. coli cellular expression
system.34
1
utilizes the vicinal H-1H spin couplings between SH and Hꢀ
(data not shown). The atomic coordinates of the EPPIb crystal
structure (PDB #: 2NUL)15 revealed that the carbonyl oxygen
atoms of Phe-27 and Glu-71 are in close proximity to the sulfur
atoms of Cys-31 and Cys-121, respectively, and seem to form
hydrogen bonds with the SH groups. In solution, the locations
of the SH protons were unambiguously identified from a number
of observed NOEs and supported the existence of these hydrogen
bonds (Figure 6). It should be noted that a hydrogen bond
between the SH group of Cys and the backbone carbonyl oxygen
of the residue four residues behind in the sequence thus forms
the bifurcated hydrogen bonds which stabilize the R-helix as
described above. It should also be noted that, in the crystal
structure of EPPIb, the hydrogen bond lengths involving Cys-
31 and Cys-121 are 3.3 and 3.7 Å, respectively, which are
considerably longer than the typical hydrogen bond lengths
observed for amide and hydroxyl hydrogen bonds, which are
usually less than 3 Å. Due to the longer hydrogen bond lengths
involving sulfur, which are attributed to the larger atomic size
of sulfur as compared to those of nitrogen and oxygen atoms,4
it is expected that relatively larger distance constraints should
be imposed for the hydrogen bonds involving SH groups.
Therefore, such hydrogen bonds are useful for refining the local
structures around the SH’s (Figure 6).
Acknowledgment. This work was supported by the Targeted
Protein Research Program (MEXT) to M.K. It was also supported
in part by a Grant-in-Aid for Young Scientists (B) (21770110) to
M.T. and a Grant-in-Aid on Innovative Areas (4104) to J.J.
Supporting Information Available: A figure detailing the time
course of the deuterium exchange for the backbone amide and
sulfhydryl groups of Cys-31 and Cys-121 in EPPIb, as observed
by the [2H,1H]-decoupled 13Cꢀ NMR signals at 150.9 MHz. This
material is available free of charge via the Internet at http://
pubs.acs.org.
Conclusion
JA101205J
We have developed a method to detect the slowly exchanging
SH groups of Cys residues in proteins. The method utilizes a
protein labeled with [3-13C;3,3-2H2]Cys, which gives very
narrow line widths for the 13Cꢀ signals, thus allowing us to
observe the isotope shifts due to the deuterium substitution on
the side-chain sulfhydryl and backbone amide groups simulta-
neously. The two Cys residues in EPPIb, Cys-31 and Cys-121,
are both located in the interior of the protein, and their sulfhydryl
(30) Loh, S. W.; Markley, J. L. Biochemistry 1994, 33, 1029–1036.
(31) Harris, T. K.; Mildvan, A. S. Proteins 1999, 35, 275–282.
(32) Koradi, R.; Billeter, M.; Wu¨thrich, K. J. Mol. Graph. 1996, 14, 51–
55.
(33) All of the isotopically labeled amino acids used in this study are
available from SAIL Technologies: http://www.sail-technologies.com/
index.html.
(34) LeMaster, D. M. Biochemistry 1996, 35, 14876–14881.
9
6260 J. AM. CHEM. SOC. VOL. 132, NO. 17, 2010