Monitoring the kinetics of ion-dependent protein folding by time-resolved NMR spectroscopy at atomic resolution

Article abstract of DOI:10.1021/ja994212b

The kinetics of protein refolding have been monitored by time-resolved NMR spectroscopy. It is shown that refolding of metal binding proteins can be induced by photolysis of photo labile ion chelators, the subsequent release of Ca²⁺ ions can induce protein folding, and the changes in resonance positions can be monitored by time-resolved NMR spectroscopy. The feasibility of the approach is demonstrated by characterizing the refolding of α-lactalbumin, or protein containing a Ca²⁺ binding site, unfolded in 4 M urea at pH 7 in the absence of Ca²⁺. The refolding kinetics of the methyl groups of residues Leu15 and Val21 in the core of the protein have been determined to be mono-exponential with rates of 0.28 s^-1 and 0.23 s^-1, respectively at 300 K.

Full text of DOI:10.1021/ja994212b

J. Am. Chem. Soc. 2000, 122, 6169-6174
6169
Monitoring the Kinetics of Ion-Dependent Protein Folding by
Time-Resolved NMR Spectroscopy at Atomic Resolution
Till Ku1hn and Harald Schwalbe*
Contribution from the Massachusetts Institute of Technology, Department of Chemistry, Francis Bitter
Magnet Laboratory, 170 Albany Street, Building NW14, Cambridge, Massachusetts 02139
ReceiVed December 2, 1999. ReVised Manuscript ReceiVed May 3, 2000
Abstract: The kinetics of protein refolding have been monitored by time-resolved NMR spectroscopy. It is
shown that refolding of metal binding proteins can be induced by photolysis of photo labile ion chelators, the
subsequent release of Ca2+ ions can induce protein folding, and the changes in resonance positions can be
monitored by time-resolved NMR spectroscopy. The feasibility of the approach is demonstrated by characterizing
the refolding of R-lactalbumin, or protein containing a Ca2+ binding site, unfolded in 4 M urea at pH 7 in the
absence of Ca2+. The refolding kinetics of the methyl groups of residues Leu15 and Val21 in the core of the
protein have been determined to be mono-exponential with rates of 0.28 s-1 and 0.23 s-1, respectively at
300 K.
Introduction
criteria have to be met to follow protein refolding by time-
resolved NMR spectroscopy:
• The release time of the caged compound is a property of
the chelator and should be as short as possible.
A variety of different spectroscopic techniques has been
developed to study the kinetics of protein folding. Recently,
also direct NMR spectroscopic detection of the kinetics of
protein refolding has been reported. Experimental approaches
have been developed to induce folding by rapidly injecting high
concentrations of unfolded protein into folding buffer1 or by
mixing unfolded protein and folding buffer in specially designed
NMR probes.2
• The irradiation time τirradiate is a property of the sample
geometry, the concentration of the caged compound in solution
and the power output of the laser at the wavelength of the
absorption maximum of the photo labile compound (in W).
τirradiate has to be shorter or of the order of the folding time.
• Conditions have to be found in which the protein conforma-
tion can be triggered by the rapid release of the caged ion. This
includes that the dissociation constant of the chelator/ion
complex is orders of magnitude lower than the dissociation
constant of the protein/ion complex.
An alternative approach is to initiate protein folding3 by
exploiting the differences in protein stability in the presence
and absence of stabilizing cofactors4 such as Ca2+ and Mg2+ 5,6
.
Ion-dependent protein folding steps can be initiated by rapid
release of Ca2+ ions caged in photolabile calcium cage
compounds7 like the EDTA-analogue DM-nitrophen which has
a KD ) 5 × 10-9 M for Ca2+ binding.8,9 This approach involves
the combination of the application of short laser pulses10 with
the subsequent observation of an NMR signal. The following
• The released byproducts of the uncaged chelator must not
interfere with the folding to be studied.
Here we present the first experimental data in which the
kinetics of protein refolding have been characterized by coupling
rapid ion release to high-resolution NMR spectroscopy. We have
studied the refolding of R-lactalbumin as a model system.
Calcium free R-lactalbumin is unfolded in 4 M urea, while the
Ca2+ bound form is nativelike under these conditions as
evidenced by NMR and circular dichroism (CD) spectra.
The developed general approach should allow application of
time-resolved NMR spectroscopy to the characterization of
folding steps or reactions that can be triggered using photo-
chemical reactive caged compounds.
* To whom correspondence should be addressed at Massachusetts
Institute of Technology, Department of Chemistry, Francis Bitter Magnet
Laboratory, 170 Albany Street, Bldg. NW14, Cambridge, MA 02139.
Telephone: (617) 253-5840. Fax: (617) 253-5405. E-mail: schwalbe@
ccnmr.mit.edu.
(1) Frieden, C.; Hoeltzli, S. D.; Ropson, I. J. Protein Sci. 1993, 2, 2007-
2014; Balbach, J.; Forge, V.; van Nuland, N. A.; Winder, S. L.; Hore, P.
J.; Dobson, C. M. Nat. Struct. Biol. 1995, 2, 865-870; Hoeltzli, S. D.;
Frieden, C. Proc. Natl. Acad. Sci. U.S.A. 1995, 92, 9318-9322; Balbach,
J.; Forge, V.; Lau, W. S.; van Nuland, N. A.; Brew, K.; Dobson, C. M.
Science 1996, 274, 1161-1163.
(2) Spraul, M.; Hoffmann, M.; Schwalbe, H.; Deutsches Patent DE
19548977 C 1. Schwalbe, H.; Ku¨hn, T.; von Heyden, R.; Ball, R.; Hofmann,
M.; Spraul, M.; Bermel, W. Manuscript in preparation.
(3) Chan, C. K.; Hofrichter J.; Eaton, W. A. Science 1996, 274, 628-
629; Jones, C. M.; Henry, E. R.; Hu, Y.; Chan, C. K.; Luck, S. D.; Bhuyan,
A.; Roder, H.; Hofrichter, J.; Eaton, W. A. Proc. Natl. Acad. Sci. U.S.A.
1993, 90(24), 11860-11864.
Material and Methods
A number of different photolabile ion chelators11 have been
developed. We have chosen DM-nitrophen for our studies because of
its favorable photochemical properties (compare Figure 1).12 The
(4) Gorne-Tschelnokow, U.; Hucho, F.; Naumann, D.; Barth, A.; Ma¨ntele,
W. FEBS Lett. 1992, 309(2), 213-217.
(8) Kaplan, J. H.; Ellis-Davies, G. C. R. Proc. Natl. Acad. Sci. U.S.A.
1988, 85, 6571-6579.
(5) DelPrincipe, F.; Egger, M.; Niggli, E. Nat. Cell. Biol. 1999, 1(6),
323-329.
(9) McCray, J. A.; Fidler-Lim, N.; Ellis-Davies, G. C. R.; Kaplan, J. H.
Biochemistry 1992, 31(37), 8856-8861.
(6) Mines, G. A.; Pascher, T.; Lee, S. C.; Winkler, R. J.; Gray, H. B.
Chem. Biol. 1996, 6, 491-497.
(10) Ellis-Davies, G. C. R.; Kaplan, J. H.; Barsotti, R. J. Biophys. 1996,
70, 1006-1016.
(7) Morad, M.; Davies, N. W.; Kaplan, J. H.; Lux, H. D. Science 1988,
241, 842-844.
(11) Kaplan, J. H. Annu. ReV. Physiol. 1990, 32, 897.
10.1021/ja994212b CCC: $19.00 © 2000 American Chemical Society
Published on Web 06/16/2000

6170 J. Am. Chem. Soc., Vol. 122, No. 26, 2000
Ku¨hn and Schwalbe
Figure 1. Photochemical properties and dissociation constants of DM-nitrophen and its photo products after laser irradation.
Figure 2. Setup for the laser-NMR coupling. As a light source a Lambda Physik EMG 101 MSC excimer laser was used with a XeF gas mixture.
The emitted wavelength was 351 nm. The laser pulse energy 35 mJ per 20 ns pulse at 60 Hz. The laser output was coupled into the sample in the
NMR tube via a quartz fiber optic. The coupling attenuation was less than 45%. Laser pulses or an electronic shutter respectively were controlled
by a TTL signal from the spectrometer console.
dissociation constant of DM-nitrophen is sufficiently small to bind Ca2+
in the presence of R-lactalbumin (KDprior(Cage:Ca2+) ) 5 × 10-9 M)
and the KD of the photo products after cleavage is 6 orders of magnitude
higher (KDafter(Prod:Ca2+) ) 3 × 10-3 M). The ion release time is short
compared to the NMR acquisition period of approximately 100 ms.
The quantum yield for photocleavage is smaller than that for 4-(2-
nitrophenyl)-3,13-bis-[(ethoxycarbonyl)methyl]-6,9-dioxa-3,12-diaza-
tetradecanedioic acid (nitrophenyl-EGTA) (Φ ) 0.20). A prominent
absorption maximum of DM-nitrophen is near 351 nm, while the
absorption maximum of the photo products is near 290 nm and the
absorption coefficient of the photo products at 351 nm is smaller than
for DM-nitrophen.13 A modified synthesis of DM-nitrophen is described
in detail in the Supporting Information.
Laser Setup. In the laser experiment shown schematically in Figure
2, a Lambda Physik EMG 101 MSC excimer laser filled with a Xe/F2
gas mixture was used to tune the emitted wavelength to 351 nm. The
60 Hz pulsed laser beam was guided through a self-made electronically
controlled shutter and was focused on a multimode silica/silica fiber
(1500 µm core, 1800 µm cladding, CERAM OPTIC). For coupling of
the pulsed laser beam into the NMR tube we used a cone-shaped quartz
tip (CSQT) that was mounted to a fiber guiding quartz tube of 4.8 mm
diameter, which itself was attached inside a standard 5 mm H2O/D2O
(12) Grell, E.; Lewitzki, E.; Ruf, H.; Bamberg, E.; Ellis-Davies, G. R.
C.; Kaplan, J. H.; de Weer, P. Cell Mol. Biol. 1989, 35(5), 515-522.
(13) Ellis-Davies, G. C. R.; Kaplan, J. H. Proc. Natl. Acad. Sci. U.S.A.
1994, 91, 187-191.

Kinetics of Ion-Dependent Protein Folding
J. Am. Chem. Soc., Vol. 122, No. 26, 2000 6171
Stopped flow CD and fluorescence data were acquired on an Applied
Photophysics π* 180 stopped flow spectrometer at 300 K and with
final concentrations (after mixing) of 100 µM R-lactalbumin in 4 M
urea, 100 mM Tris/HCl at pH 7.0 and 200 µM CaCl2 which was
obtained by mixing of 100 µM R-Lactalbumin in 4 M urea and 100
mM Tris/HCl at pH 7.0 against 200 µM CaCl2 in 4 M urea and 100
mM Tris/HCl at pH 7.0. Urea background spectra were subtracted from
the kinetic traces. The general buffer was 100 mM Tris/HCl at pH 7 to
ensure the same conditions as in the NMR experiments.
Results and Discussion
CD Titration Experiments. R-Lactalbumin in its Ca2+ free
form is unstructured in 4 M urea at pH 7.0, whereas the presence
of calcium stabilizes the native state of the protein. As can be
seen in Figure 3a the molar ellipticity increases with the increase
of the urea concentration around 222 nm in the absence of
calcium, indicating a decrease in R-helicity. With increasing
urea concentration in the calcium free form, R-lactalbumin
becomes less structured. This is clearly not the case in the
presence of calcium (Figure 3b). Up to a concentration of about
4 M urea, the CD spectra do not differ significantly from the
Ca2+ free form of the protein in the absence of urea. The largest
structural change, according to CD data, will be expected upon
addition of equimolar amounts of calcium to an R-lactalbumin
solution in 4 M urea at pH 7.
Laser-Coupled NMR Spectroscopy. In a standard 5 mm
Shigemi H2O tube 0.88 mg of R-lactalbumin (0.5 mM) was
dissolved in 250 µl of 4 M urea solution in 90% H2O/10% D2O,
2 mM DM-nitrophen, and 1.5 mM CaCl2 in 100 mM Tris/HCl
at pH 7.0. The laser/NMR coupling was performed as described
above. The experiments were performed on a Bruker DRX 600
MHz spectrometer equipped with a z gradient TXI probe. The
signal-to-noise ratio of this probe measured by using standard
ethylbenzene (thin wall tube) is greater than 1250:1. The laser/
lightguide coupling was optimized as well as the lightguide/
CSQT joint with respect to the emitted light energy. The
following experiments were carried out at 300 K.
The NMR spectrum of the methyl group region of the protein
prior to calcium release is shown in the lower part of Figure 4.
The spectrum is characteristic of an unfolded protein: the methyl
groups (and the amide region, not shown) resonate at the random
coil shifts, and the observed peaks are sharp. The upper part of
Figure 4 shows the single scan NMR spectrum after release of
calcium ions in the spectrometer with individual resonance
positions assigned. The folding transitions between the two
forms are reversible. Addition of EDTA after photo-destruction
of the DM-nitrophen unfolds the protein as evidenced by NMR
spectra after each kinetic experiment. R-Lactalbumin could be
recovered in its native form by extensive dialysis and proved
to be unmodified according to ESI-MS spectral analysis. The
effects of the photo products of DM-nitrophen on the folding
kinetics of R-lactalbumin were not specifically tested. However,
since 1D NMR spectra of the protein, folded by calcium released
from photocleavage are indistinguishable from those of R-lac-
talbumin folded by calcium addition prior to the experiment,
we do not expect changes in refolding kinetics in the presence
or absence of submillimolar amounts of photo products.
Temperature changes during laser irradiation were not measur-
able using the standard Bruker temperature control unit in the
probe, and the lock signal dropped only marginally, if at all,
during irradiation.
Figure 3. CD titration spectra of bovine R-lactalbumin and increasing
concentrations of urea at pH 7.0 in the presence and absence of calcium
ions. The protein remains structured in the presence of equimolar
amounts of Ca2+ even in 4 M urea. In the absence of Ca2+
,
R-lactalbumin is unfolded at a concentration of 4 M urea at pH 7. Due
to the urea absorbance below about 210 nm, spectra in the presence of
urea can only be interpreted for wavelengths above 210 nm.
Shigemi NMR tube. The coupling efficiency between the CSQT and
the fiber was optimized using the fluorescence image of the scattered
light. Losses due to scattering were minimized by using methanol as a
mediator between the diffraction indices of the two quartz types (fiber
and tip). The CSQT was completely covered by the solution and was
fixed to the tube in such a way that the end of the tip was located 2
mm below the center of the rf coil. Gradient shimming was applied to
the setup after a gradient map for each specific sample was recorded.
In irradiation experiments performed outside the NMR spectrometer,
no diffusion of the red photo products was observed over several
seconds (compare Figure 2, after irradiation). Diffusion of released
calcium occurring during the experiment should therefore be negligible.
CD Study of r-Lactalbumin. Urea titration experiments followed
by CD of bovine milk R-lactalbumin (Fluka) were performed on a
JASCO J710 CD spectrometer. To perform equilibrium urea titration
experiments on Ca2+ free and Ca2+ loaded R-lactalbumin, the protein
was separated by ultrafiltration using a YM1000 Amicon membrane
and washed three times with 50 mL of a desalting solution of 50 mM
EDTA in 100 mM Tris/HCl pH 7.0. The protein was filtered five times
in 50 mL of 100 mM Tris/HCl at pH 7.0 to equilibrate in EDTA free
buffer. A 100 µM solution of this Ca2+ free protein was used for urea
titration CD experiments in 0-6 M urea and 100 mM tris/HCl at pH
7.0. To study the Ca2+-loaded form of R-lactalbumin, solutions of 100
µM Ca2+ free R-lactalbumin in 200 µM CaCl2 and 100 mM Tris/HCl
at pH 7.0 with urea concentrations varying from 0 to 6 M were prepared.
The results are shown in Figure 3. All of the scanning near UV CD
experiments were recorded using a 0.01 cm quartz cuvette.
For the kinetic NMR experiment, the following sequence has
been recorded. A single scan jump and return echo FID14 was
(14) Sklenar, V.; Bax, A. J. Magn. Reson. 1987, 74, 469-474.

6172 J. Am. Chem. Soc., Vol. 122, No. 26, 2000
Ku¨hn and Schwalbe
Figure 5. The pulse sequence of the time-resolved NMR experiment,
including the irradiation period. Herein, tlight is the duration of a light
pulse, tkinetic is the time between two acquisitions plus the acquisition
time (for best time resolution, tkinetic should be as small as possible),
and n represents the number of scans after single laser application. This
inner loop (n) therefore contains non-equilibrium kinetic information.
Each of these n experiments is repeated s times. The outer loop holds
thermodynamic (equilibrium) information. Therefore, each individual
signal of the inner loop represents the same state of folding after
initiation by the light pulse. In this case, for example the second signal
after the light always represents the state of folding after the acquisition
time of the first FID plus tkinetic
.
due to scattering at the coupling site and attenuation in the fiber
(about 50%).
Directly after the irradiation, a series of seven 1D experiments
(dark_2 to dark_8) were recorded. Two thousand complex data
points were recorded per FID, the recovery time after each
acquisition was set to 1.5 s. This experiment was repeated 20
times (dark_9 would be identical to dark_1 of the next series)
(compare Figure 5). The resulting FIDs were transformed using
the MSI program Felix98. The resulting 3D file was split up
into the individual FIDs which are described above.
This gives rise to several sets (s) of series (n ) 8) of 8 FIDs.
The trace of any given spectrum (n) over the 20 sets yields
thermodynamic information since it resembles a titration curve.
This is the case since only a fraction of the Ca2+ is released by
one laser irradiation period apparently and each of the FIDs is
a sum of the fraction of folded protein and the fraction of
unfolded protein. The trace over the whole series of spectra (n)
within each set (s), however, provides kinetic information since
it monitors the events directly following a single train of laser
pulses. The signal-to-noise-ratio in such a single trace is too
poor to yield reasonable kinetic data. Therefore, the kinetic
traces where averaged as described below.
Figure 4. (Bottom) Single scan 1D NMR spectrum of the methyl group
region of the denatured form of R-lactalbumin in 4 M urea with access
of “caged calcium” prior to irradiation (initial state). The spectrum of
the protein is identical to the 1D spectrum of R-lactalbumin in calcium-
free buffer in 4 M urea. (Top) Single scan NMR spectrum of the methyl
group region of refolded R-lactalbumin after irradiation and Ca2+
release. It can be seen that distinct peaks are arising in the Me-group
region. The spectrum is identical to the protein spectrum in 4 M urea
in the presence of calcium. A, B, C, and D indicate the peaks for which
individual refolding kinetics are discussed in text and Figure 7.
acquired (further referred to as dark_1) as a reference experi-
ment. After a relaxation delay of 1.5 s, the sample was irradiated
with a wavelength of 351 nm by 30 laser pulses with a rate of
60 Hz and a pulse duration of 20 ns each, the high voltage at
the laser electrodes was set to 24.6 kV. The total irradiation
time τirradiate was therefore 500 ms. This relatively long irradia-
tion period prooved to be unproblematic since protein folding
can still be detected for several seconds thereafter. Also no major
phase shifts between the rising single residue methyl peaks and
the decaying intensity of the methyl bulk peak could be detected,
as they would be expected if the rate constants of refolding
increase and thereby induce a frequency change during acquisi-
tion.15 Therefore we conclude that the reaction is too slow to
result in a major signal change during the acquisition time and
significant changes are measurable in difference experiments
as described here. Nevertheless, it is desirable to work with
shorter irradiation times for improved dead times; however, it
proved inapplicable here due to insufficient laser energy and
therefore unsatisfactory signal-to-noise ratio in experiments with
shorter irradiation time. Energetic measurements before and after
the NMR experiment revealed that the laser was running in a
stable mode and that the pulse energy was steady during the
kinetic NMR experiment. Pulse energies of 35 mJ where coupled
into the fiber at the same conditions and with the same gas
mixture as that used during the NMR experiment. The light
energy per laser irradiation in the sample during the experiment
was approximately 1 W, taking into account all of the losses
The individual experiments were stored as two-dimensional
data arrays given in eq 1:
dark_1t1rans
dark_1t2rans
dark_1_mat: )
, dark_2_mat: )
l
(
)
dark_1trans
dark_2t1rans
dark_2t2rans
l
dark_8t1rans
dark_8t2rans
, ... dark_8_mat: )
(1)
l
(
)
(
)
dark_2trans
dark_8trans
The subscript refers to the number of the irradiation cycle (s),
the superscript trans indicates that subsequent data manipulation
was performed on the Fourier transformed data set, and dark_1
resembles the first FID (n ) 1) of an irradiation cycle (s).
Inspection of the spectra reveal the time point for which the
deliberately added excess of DM-nitrophen has been consumed
by the observation of the starting point of decreasing methyl
group peak intensity of the unstructured form. In the experi-
mental series shown in Figure 6 this occurs after about two
laser irradiations. In the same way it is possible to determine
the time point for which all of the protein is folded or for which
all Ca2+ is released (methyl group peak intensity reaching
(15) Ku¨hne, R.; Schaffha¨user, T.; Wokaun, A.; Ernst, R. R. J. Magn.
Reson. 1979, 35, 39.

Kinetics of Ion-Dependent Protein Folding
J. Am. Chem. Soc., Vol. 122, No. 26, 2000 6173
Figure 6. Equilibrium information: stack plot of the methyl group
region. The 3D representation was obtained by constantly plotting the
eighth spectrum after the laser pulse train as a function of the number
of irradiation cycles (s) (only spectra of every 2nd laser irradiation are
plotted here). Since only a certain amount of Ca2+ is released after a
single laser application, the experiment represents a calcium titration
experiment.
Figure 7. Kinetic traces of individual peaks in the methyl group region.
Selected assigned peaks are annotated:17 (A) decay of the methyl group
bulk peak, indication of overall folding; (B) rise of the residual peak
at 0.66 ppm in the methyl group area; (C) folding kinetics of the single
Val21 residue in R-lactalbumin at 0.21 ppm; (D) kinetic trace for the
Leu15 moiety in R-lactalbumin at -0.23 ppm.
steady-state asymptotically). The part between these character-
istic points where the slope of methyl group peak intensity vs
number of irradiation cycles * 0, we call the releVant area. In
the three independent experiments with different samples
conducted here, the relevant area extended from set s ) 2-10,
s ) 3-14, and s ) 2-13, respectively.
Series of spectra were subtracted, and the result was summed
to improve the signal-to-noise ratio. Only spectra of irradiation
cycles within the relevant area were used, thereby avoiding the
unnecessary summation of noise. An explanation of this is given
by eq 2:
10
diffspectrum1: ) (dark_2trans - dark_1tsrans
)
∑
s
s)2
10
diffspectrum2: ) (dark_3trans - dark_1tsrans), etc. (2)
∑
s
s)2
Herein, s resembles the number of the irradiation, and the
relevant area is between the secone and tenth laser irradiation.
A 2D matrix of the resulting difference spectra (dark_2 -
dark_1 up to dark_8 - dark_1) was stored for subsequent
kinetic data analysis.
Comparison of Folding Kinetics Revealed by Stopped
Flow CD, Fluorescence, and NMR Spectroscopy. We were
able to follow the kinetics of several distinct methyl groups and
lowfield NH groups. Best fits to kinetic functions of these data
reveal different time constants for individual methyl groups
which indicates discriminative folding patterns in different
regions of the molecule. Kinetic data is shown in Figure 7. The
overall decay of the methyl-bulk peak follows a single-
exponential kinetic (A f B) with k ) 0.33 s-1 and a half time
of τ1/2 ) 2.10 s (Figure 7a). The decreasing integral of the
methyl-bulk peak clearly indicates the formation of secondary
structure in the protein. The chemical shifts of the methyl groups
start to disperse upon structural formation. Most of the other
traces gain intensity over time, since they correspond to the
formation of distinct structural features in the molecule. Their
kinetics differ from each other as well as from the kinetics of
the overall methyl peak decay. All of the clearly resolved peaks
reveal mono-exponential refolding rates.
Figure 8. Stopped flow data for the refolding of R-lactalbumin from
4 M urea upon the addition of an equimolar Ca2+ solution. Experimental
data and fitted data are shown. (A) CD kinetic trace recorded at 225
nm, indicating the formation of R helical structure in the protein. The
kinetics were fitted to a mono-exponential decay. (B) Fluorescence
kinetic trace of the tryptophan residues: excitation at 295 nm, cutoff
before detection: 320 nm. The kinetics were fitted to a biexponential
decay.
The decay of the unfolded protein, as evidenced by the decay
of the bulk of the methyl groups, is faster than the build up of
the native methyl signal that evidences native tertiary contacts
within the protein. The rates of formation of native tertiary

6174 J. Am. Chem. Soc., Vol. 122, No. 26, 2000
Ku¨hn and Schwalbe
contacts vary from k ) 0.16 s-1 for the peak at 0.66 ppm (Figure
7b) through k ) 0.23 s-1 for the Val21 residue at 0.21 ppm
(Figure 7c), to k ) 0.28 s-1 for the Leu15 moiety at -0.23
ppm (Figure 7d). The difference between the rate constants of
the slower single residues and the rate constant of the fast decay
of the methyl bulk peak is statistically significant.
Conclusions
Triggering of the refolding of R-lactalbumin by rapid photo
release of Ca2+ from DM-nitrophen and the subsequent NMR
spectroscopic detection of the reaction allowed the measurement
of refolding rates as fast as ∼0.3 Hz with atomic resolution.
The data presented stem from three independent experiments
carried out with 0.5 mM sample in 250 µL of solution on a
600 MHz spectrometer. The data therefore shows that the
kinetics of biopolymer folding and reactions, for example,
enzyme-catalyzed reaction, can be triggered and subsequently
monitored with high sensitivity by time-resolved NMR spec-
troscopy. A plethora of structural and mechanistic questions in
structural biology should be addressable using this approach of
rapid release of photochemically caged functionality.
The kinetic data obtained from high resolution NMR spec-
troscopy differs significantly from the data obtained by stopped
flow CD and fluorescence data at the same temperature. For
stopped-flow CD, a double exponential rate for the formation
of R helical structure (CD signal at 225 nm) k1 ) 8.9 s-1 and
k2 ) 1.0 s-1 is observed. The overall change of the fluorescence
signal (excitation at 295 nm, cutoff 320 nm) shows biphasic
behavior with k1 ) 11.4 s-1 and k2 ) 1.3 s-1. Kinetic traces
are shown in Figure 8. These two methods probe for the
formation of - or for the change in secondary structure. The
results obtained here with our new NMR experiment are
sensitive to changes in side-group interactions presumably by
the formation of tertiary contacts. The decreasing peak integral
of the methyl peak at 0.84 ppm arises from the bulk of the
methyl groups in the unfolded state of R-lactalbumin. The
signals of upfield shifted methyl groups in the folded state of
R-lactalbumin arise from native contacts of these methyl groups
presumably with aromatic amino acids in the interior of the
proteins. The slower rates for these processes monitored by
NMR, compared to the results obtained by time-resolved CD
and fluorescence, therefore indicate a faster rate for secondary
structure formation than for tertiary structure formation in the
case of R-lactalbumin. The fact that different residues show
different refolding rate constants and that those rate constants
can be fitted to a two-step kinetic process favors the idea of a
nucleation model16 in the case of the calcium-induced refolding
of R-lactalbumin from 4 M urea.
Acknowledgment. We thank the DFG (Schw701/3-1), the
Fonds der Chemischen Industrie, the Hoechst Foundation, and
the Massachusetts Institute of Technology for financial support.
Spectra were recorded at the Large Scale Facility for Biomo-
lecular NMR at the University of Frankfurt and at the Francis
Bitter Magnet Laboratory at MIT. H.S. was supported in part
by the Large Scale facility. We would like to thank M. Czisch
for help with preliminary experiments carried out at the Large
Scale Facility at the Bijvoet Center at the University of Utrecht/
NL. We thank Professor S. Glaser for enlightening discussions
in the early stages of the project. We are grateful to Professor
T. Prisner for letting us use the excimer laser. We thank
Professor C. Griesinger and Julia Wirmer for helpful discussions.
Supporting Information Available: Experimental details
(PDF). This material is available free of charge via the Internet
at http://pubs.acs.org.
(16) Fersht, A. R. Proc. Natl. Acad. Sci. U.S.A. 1995, 92, 10869-10873.
(17) Redfield, C. Personal communication.
(18) Ellis-Davies, G. C. R.; Kaplan, J. H. J. Org. Chem. 1988, 53, 1966-
1969.
JA994212B