Methods 34 (2004) 15–27

www.elsevier.com/locate/ymeth

Rapid mixing methods for exploring the kinetics of protein folding

Heinrich Roder,a,b,* Kosuke Maki,a,1 Hong Cheng,a and M.C. Ramachandra Shastrya,2

a Basic Science Division, Fox Chase Cancer Center, 333 Cottman Ave., Philadelphia, PA 19111, USAb Department of Biochemistry and Biophysics, University of Pennsylvania, Philadelphia, PA 19104, USA

Accepted 5 March 2004

Available online 7 June 2004

Abstract

Information on the time-dependence of molecular species is critical for elucidating reaction mechanisms in chemistry and bi-

ology. Rapid flow experiments involving turbulent mixing of two or more solutions continue to be the main source of kinetic in-

formation on protein folding and other biochemical processes, such as ligand binding and enzymatic reactions. Recent advances in

mixer design and detection methods have opened a new window for exploring conformational changes in proteins on the micro-

second time scale. These developments have been especially important for exploring early stages of protein folding.

� 2004 Elsevier Inc. All rights reserved.

Keywords: Ultrafast mixing; Stopped-flow; Continuous-flow; Fluorescence; NMR

1. Introduction

To elucidate the mechanism of any reaction, be it

chemical or biochemical, a binding process or a mac-

romolecular conformational change, we need to deter-

mine the time-dependent evolution of the molecular

species involved. Kinetic techniques fall into two broad

categories, equilibrium and relaxation methods. Equi-

librium methods extract rate-information withoutphysically or chemically perturbing a system, for ex-

ample, by measuring dynamic effects on spectral line

shape (NMR, EPR or optical), or by observing molec-

ular fluctuations. Relaxation techniques generally rely

on a rapid change of an extrinsic variable (e.g., tem-

perature, pressure or solvent composition) to perturb

the system and follow its response as it evolves toward a

new equilibrium position. The development of laser-based photochemical or thermal triggers has opened

new time windows for exploring the dynamics of bio-

logical molecules on the micro- and nano-second time

* Corresponding author. Fax: 1-215-728-3574.

E-mail address: roder@fccc.edu (H. Roder).1 Present address: Department of Physics, University of Tokyo,

Tokyo 113-0033, Japan.2 Present address: Colgate-Palmolive Co., Piscataway, NJ 08855,

USA.

1046-2023/$ - see front matter � 2004 Elsevier Inc. All rights reserved.

doi:10.1016/j.ymeth.2004.03.003

scale [1–5]. However, the most common approach fortriggering chemical and biological processes relies on

turbulent mixing to achieve a rapid change in solvent

composition. Historically, some of the earliest rapid

kinetic measurements with millisecond time resolution

used a continuous-flow arrangement combined with

absorbance measurements of the reaction progress at

different points downstream [6]. However, continuous-

flow experiments were later replaced by the more ver-satile and economic stopped-flow experiment, which can

be coupled with a wide range of spectroscopic probes to

monitor reactions with millisecond time resolution [7,8].

Continuous-flow techniques have experienced a re-

naissance in recent years due to advances in mixer de-

sign and detection methods, which made it possible to

push the time resolution into the microsecond time

range [9–12]. Other techniques make use of two or moreconsecutive mixing steps to prepare the system in a

particular initial state (double-jump stopped-flow), or to

execute multiple reaction steps in sequence (quenched-

flow). If a reaction can be quenched by manipulating

solution conditions (e.g., pH) or lowering temperature,

quenched-flow or freeze-quench protocols can be used in

combination with slower analytical techniques, such as

NMR, EPR or mass spectrometry [13]. To achieve effi-cient turbulent mixing conditions requires high flow

16 H. Roder et al. / Methods 34 (2004) 15–27

rates and relatively large channel dimensions, which canconsume substantial amounts of material. A more eco-

nomic technique that uses hydrodynamic focusing to

mix solutions under laminar flow conditions has recently

been introduced [14,15].

Rapid mixing techniques play a particularly promi-

nent role in kinetic studies of protein folding [16–21]. As

with any complex process, time-resolved data are es-

sential for elucidating the mechanism. Even in caseswhere the whole process of folding occurs in a single

step, which is the case for many small proteins, the ki-

netics of folding and unfolding provide valuable infor-

mation on the rate-limiting barrier. The effects of

temperature and denaturant concentration give insight

into activation energies and solvent-accessibility of the

transition state ensemble, and by measuring the kinetic

effects of mutations, one can gain more detailed struc-tural insight [22–24]. If the protein folding process oc-

curs in stages, i.e., if partially structured intermediate

states accumulate, kinetic studies can potentially offer

much additional insight into the structural and ther-

modynamic properties of intermediate states and inter-

vening barriers [18–20,25,26]. The combination of

quenched-flow techniques with hydrogen exchange la-

beling and NMR has proven to be particularly fruitfulfor the structural characterization of transient folding

intermediates [27–29].

Fig. 1. Stopped-flow fluorescence evidence for an unresolved rapid

process (burst phase) during folding of cytochrome c (pH 5, 10 �C). (A)

Tryptophan fluorescence changes during refolding of acid-unfolded

cytochrome c (pH 2, �15mM HCl) at a final GuHCl concentration of

0.7M. The initial signal Sð0Þ at t ¼ 0 (determined on the basis of a

separate dead-time measurement) falls short of the signal for the un-

folded state under refolding conditions, SpredðUÞ, obtained by linear

extrapolation of the unfolded-state baseline (see dashed line in (B)). (B)

Effect of the denaturant concentration on the initial (squares) and final

(circles) fluorescence signals, Sð0Þ and Sð1Þ, measured in a series

of stopped-flow refolding experiments at different final GuHCl

concentrations.

2. Burst-phase signals in stopped-flow experiments

In many cases, stopped-flow and quenched-flow

measurements of protein folding reactions show indi-cations of unresolved rapid processes occurring within

the dead time (e.g. [30–35]). This is illustrated by Fig. 1,

which shows the kinetics of refolding of cytochrome c

measured by stopped-flow fluorescence (panel A) along

with equilibrium fluorescence data vs. denaturant con-

centration (panel B). The protein was unfolded by ad-

dition of 4.5M guanidine hydrochloride (GuHCl),

which lies in the baseline region above the cooperativeunfolding transition (Fig. 1B). The refolding reaction

was triggered by 6-fold dilution with buffer (0.1M so-

dium acetate, pH 5), resulting in a final GuHCl con-

centration well within the folded baseline region. The

data points in Fig. 1A were recorded by sampling the

fluorescence emission above 325 nm (using a glass cutoff

filter) at logarithmically spaced time intervals. The first

time point corresponds to the instrumental dead time of2.5ms, which was calibrated using a standard test re-

action ([36]; see below). The observed decay fits to a

series of two exponential phases (solid line), a major one

with a time constant of about 8ms and a minor one with

a time constant in the 100ms range. Extrapolation of the

observed kinetics back to t ¼ 0 yields the initial signal,

Sð0Þ, which is compared in Fig. 1B (arrow) with the

equilibrium unfolding transition plotted on the same

fluorescence scale (relative to unfolded cytochrome c at

4.5M GuHCl). The initial signal observed in this and a

series of additional stopped-flow experiments at different

final GuHCl concentrations are consistently below therelative fluorescence of the unfolded state, SpredðUÞ,predicted by linear extrapolation from the unfolded

baseline region to lower GuHCl concentrations (dashed

line in panel B). The difference between the predicted

and observed initial amplitude, SpredðUÞ � Sð0Þ, often

called the burst phase, reflects conformational events

occurring within the dead time of the stopped-flow ex-

periment. Similar observations on many different pro-teins using various spectroscopic parameters gave clear

evidence for the existence of rapid conformational

events that cannot be resolved with conventional mixing

H. Roder et al. / Methods 34 (2004) 15–27 17

techniques, and provided a strong incentive for the de-velopment of faster methods for triggering and observ-

ing structural changes during the first millisecond of

refolding.

3. Turbulent mixing

Most rapid mixing schemes rely on turbulent mixingto achieve complete mixing of two (or more) solutions.

Mixers of various design are in use ranging from a

simple T-arrangement to more elaborate geometries,

such as the Berger ball mixer [37]. The goal is to achieve

highly turbulent flow conditions in a small volume. The

turbulent eddies thus generated can intersperse the two

components down to the micrometer scale. However,

the ultimate step in any mixing process relies on diffu-sion to achieve a homogeneous mixture at the molecular

level. Given that the diffusion time, t, varies as the

square of the distance, r, over which molecules have to

diffuse, it takes a molecule with a diffusion constant

D ¼ 10�5 cm2/s about 1ms to diffuse over a distance of

1 lm (t ¼ r2=D). Thus, the mechanical mixing step has

to intersperse the two components on a length scale of

less than 1 lm to achieve sub-millisecond mixing times.The onset of turbulence is governed by the Reynolds

number, Re, defined as

Re ¼ qvd=g; ð1Þwhere q is the density (g/cm3), v is the flow velocity (cm/

s), d describes the characteristic dimensions of the

channel (cm), and g is the viscosity of the fluid (e.g., 0.01

poise for water at 20 �C). To maintain turbulent flow

conditions in a cylindrical tube, Re has to exceed values

of about 2000.

Turbulence is important not only for achieving effi-cient mixing, but also for maintaining favorable flow

conditions during observation. In stopped-flow and

quenched-flow experiments, turbulent flow insures effi-

cient purging of the flow lines. In continuous-flow

measurements, turbulent flow conditions in the obser-

vation channel lead to an approximate ‘‘plug flow’’

profile, which greatly simplifies data analysis compared

to the parabolic profile obtained under laminar flowconditions. The time resolution of a rapid mixing ex-

periment is governed not only by the mixing time, which

in practice is difficult to quantify, but also the delay

between mixing and observation. The effective delay

between initiation of the reaction and the first reliably

measurable data point is defined as the dead time, Dtd.In both stopped- and continuous-flow experiments, any

unobservable volume (dead volume), DV , between thepoint where mixing is complete and the point of obser-

vation contributes an increment Dt ¼ DV =ðdV =dtÞ to

the dead time (dV =dt is the flow rate). Additional con-

tributions to the effective dead time include the time

delay to stop the flow and any artifacts that can obscureearly parts of the kinetic trace (see below).

Increasing the flow rate promotes more efficient

mixing by generating smaller turbulent eddies, and

yields shorter time delays Dt, and thus should lead to

shorter dead times. However, this trend does not con-

tinue indefinitely. Aside from practical problems due to

back pressure and, in the case of stopped-flow mea-

surements, various stopping artifacts, the time resolu-tion of a rapid mixing experiment is ultimately limited

by cavitation phenomena [38]. Under extreme condi-

tions, the pressure gradients across turbulent eddies can

become so large that the solvent begins to evaporate,

forming small vapor bubbles that can take a long time to

dissolve. The result is an intensely scattering plume that

makes meaningful detection of the kinetic signal virtu-

ally impossible.

4. Instrumentation

4.1. Stopped-flow

In a typical stopped-flow experiment, a few hundred

microliters of solution are delivered to the mixer via twosyringes driven by a pneumatic actuator or stepper-

motors. Total flow rates in the range of 5–10ml/s with

channel diameters of the order of 1mm insure turbulent

flow conditions (Re > 5000). After delivering a volume

sufficient to purge and fill the observation cell with

freshly mixed solution, the flow is stopped abruptly

when a third syringe hits a stopping block or by closing

a valve. Commercial instruments can routinely reachdead times of a few milliseconds. Recent improvements

in mixer and flow-cell design by several manufacturers

of stopped-flow instruments resulted in dead times well

under 1ms. The upper end of the time scale that can be

reliably measured in a stopped-flow experiment is de-

termined by the stability of the mixture in the flow cell,

which is limited by convective flow or diffusion of re-

agents in and out of the observation volume. For slowreactions with time constants longer than a few minutes,

manual mixing experiments are generally more reliable.

Stopped-flow mixing is usually coupled with real-time

optical observation using absorbance (UV through IR),

fluorescence emission or circular dichroism spectros-

copy, but other biophysical techniques, including fluo-

rescence lifetime measurements [39,40], NMR [41,42],

and small-angle X-ray scattering (SAXS [43]), have alsobeen implemented.

The interpretation of stopped-flow data requires a

careful calibration of the instrumental dead time by

measuring a pseudo-first-order reaction tuned to the

time scale of interest (i.e., a single-exponential process

with a rate-constant approaching the expected dead

time) and an optical signal matching the application.

18 H. Roder et al. / Methods 34 (2004) 15–27

Common test reactions for absorbance measurementsinclude the reduction of 2,6-dichlorophenolindophe-

nol (DCIP) or ferricyanide by ascorbic acid [44]. A

convenient test reaction for tryptophan fluorescence

measurements is the irreversible quenching of N-acet-

yltryptophanamide (NATA) by N-bromosuccinimide

(NBS). For fluorescence studies in or near the visible

range, one can follow the pH-dependent association of

the Mg2þ ion with 8-hydroxyquinoline, which results ina fluorescent chelate [45], or the binding of the hydro-

phobic dye 1-anilino-8-naphthalene-sulfonic acid (ANS)

to bovine serum albumin (BSA), which is associated

with a large increase in fluorescence yield [46].

As a practical example, Fig. 2 shows a series of DCIP

absorbance measurements at several ascorbic acid con-

centrations used to estimate the dead time of our Bio-

Logic SFM-4 stopped-flow instrument equipped with aFC-08 microcuvette accessory (Molecular Kinetics, In-

dianapolis, IN). The reaction was started by mixing

equal parts of DCIP (0.75mM in water at neutral pH,

where the dye is stable) with LL-ascrobic acid at pH 2 at

final concentrations ranging from 6 to 25mM. Absor-

bance changes measured at 525 nm, near the isosbestic

point between the acidic and basic forms of DCIP, are

plotted relative to the absorbance of the reactant inwater, measured in a separate control (Fig. 2, inset).

Fig. 2. Estimation of the dead time of stopped-flow absorbance measurement

using the reduction of DCIP by ascorbic acid as a test reaction. Equal part

centrations of 6 (circles), 14 (squares), and 25mM (triangles) were mixed at

times with dashed lines indicating the dead time (�0.5ms). Absorbance ch

(diamonds) measured by mixing DCIP with 10mM HCl.

After the flow comes to a full stop, the absorbance de-cays exponentially with rate constants of about 290, 780,

and 1400 s�1 at ascorbate concentrations of 6, 14, and

25mM, respectively, which is consistent with a pseudo-

first-order reduction process with a second-order rate

constant k00 � 5:7� 105 s�1 M�1. At shorter times, the

exponential fits intersect at an absorbance change

DA ¼ 0, which corresponds to the level of the oxidized

DCIP measured in the control. The delay between thisintercept and the first data point that joins the fitted

exponential provides an estimate of the instrumental

dead time, td ¼ 0:55� 0:1ms. Alternatively, the dead

time can be estimated from the signal level of the con-

tinuous-flow regime prior to closure of the stop valve,

Icf , using the following equation:

tcf ¼ � ln½ðIcf � I1Þ=ðI0 � I1Þ�=k; ð2Þ

where I1 is the baseline at long times, I0 is the signal of

the reactant (in this case, DCIP mixed with an equal

volume of water), and k is the first-order rate constantobtained by exponential fitting. Note that tcf does not

account for the finite stop time and other stop-related

artifacts, and is thus always shorter than td, which ex-

plains why some manufacturers of stopped-flow instru-

ments prefer to cite tcf over the more realistic

operational dead time, td. In the present example, Eq. (2)

on a Biologic SFM-4 instrument with FC-08 micro-cuvette accessory,

s of DCIP in water (pH 7) and sodium ascorbate (pH 2) at final con-

a total flow rate of 12ml/s. The inset shows an expanded plot at early

anges at 525 nm are plotted relative to the absorbance of a control

H. Roder et al. / Methods 34 (2004) 15–27 19

yields tcf in the range of 0.25–0.4ms, compared totd � 0:55ms obtained by the extrapolation (Fig. 2).

4.2. Continuous-flow

In a continuous-flow experiment, the reaction is again

triggered by turbulent mixing, but, in contrast to stop-

ped-flow, the progress of the reaction is sampled under

steady-state flow conditions as a function of the distancedownstream from the mixer [47,48]. This avoids artifacts

related to arresting the flow and makes it possible to use

relatively insensitive detection methods. Thus, continu-

ous-flow measurements can achieve shorter dead times

compared to stopped-flow, but this comes at the expense

of sample economy. Most earlier versions of this ex-

periment involved point-by-point sampling of the reac-

tion profile while maintaining constant flow at high rates(several ml/s for a conventional mixer). The prohibitive

amounts of sample consumed limited the impact of

continuous-flow techniques until advances in mixer de-

sign made it possible to achieve highly efficient mixing at

lower flow rates [9,11,12,49], and an improved detection

scheme allowed simultaneous recoding of a complete

reaction profile in a few seconds [12]. These develop-

ments lowered both the dead time and sample con-sumption by at least an order of magnitude, and made

routine measurements on precious samples with dead

times as short as 50 ls possible.In 1985, Regenfuss et al. [9] described a capillary jet

mixer consisting of two coaxial glass capillaries with a

platinum sphere placed at their junction. The reaction

progress was monitored in a free-flowing jet, using

conventional photography to measure fluorescence vs.distance from the mixer. Measurements of the binding

kinetics of ANS to bovine serum albumin indicated that

dead times less than 100 ls can be achieved with this

mixer design. More recently, several laboratories re-

Fig. 3. Continuous-flow capillary mixing apparatus in fluorescence mode. (A)

cell, and optical arrangement. (B) Expanded view of the mixer/flow cell asse

ported continuous-flow resonance Raman and fluores-cence studies of enzyme and protein folding reactions on

the sub-millisecond time scale, using machined mixers

with dead-times of about 100 ls [10,11,49]. More wide-

spread use of these methods has been hampered by a

number of technical and experimental difficulties. Con-

tinuous-flow experiments involving a free-flowing jet

[9,11,49,50] are fraught with difficulties due to instability

and scattering artifacts. The use of a conventionalcamera with high-speed monochrome film for fluores-

cence detection is inadequate due to the low sensitivity

of the film in the UV region, limited dynamic range, and

the non-linearity of the film response. Finally, prohibi-

tive sample consumption makes continuous-flow ex-

periments that record a kinetic trace one point at a time

feasibly only for highly abundant proteins [10,51–53].

We were able to overcome many of these limitationsby combining a highly efficient quartz capillary mixer,

based on the design of Regenfuss et al. [9], with a flow

cell and an improved detection system involving a digital

camera system with a UV-sensitized CCD detector [12].

A diagram of the experimental arrangement is shown in

Fig. 3. Two Hamilton syringes driven by an Update

(Madison, WI) quenched-flow apparatus deliver the re-

agents to be studied at moderate pressure (<10 atm) intoeach of the two coaxial capillaries. The outer capillary

consists of a thick-walled (6mm o.d. and 2mm i.d.)

quartz tube, which is pulled to a fine tip (�200 lm i.d. at

the end), using a glassblowing lathe or a simple gravity

method. The inner capillary (360 lm o.d., 150–180 lmi.d., purchased from Polymicro Technologies, Phoenix,

AZ) with a �250 lm platinum sphere suspended at the

end is positioned inside the tapered end of the outercapillary. The sphere is formed by melting the end of

50 lm diameter platinum wire. Thin glass rods fused to

the inner wall of the outer capillary (tapering down to

diameter of �20 lm) prevent the sphere from plugging

Schematic diagram of the solution delivery system, mixer, observation

mbly.

20 H. Roder et al. / Methods 34 (2004) 15–27

the outlet. The reagents are forced through the narrowgap between the sphere and the outer wall where mixing

occurs under highly turbulent flow conditions. The fully

homogeneous mixture emerging from the mixer is in-

jected into the 250 lm� 250 lm flow channel of a fused-

silica observation cell (Hellma, Plainview, NY) joined to

the outer capillary by means of a hemispherical ground-

glass joint. Typical flow rates are 0.6–1.0ml/s resulting

in a linear flow velocities of 10–16m/s through the0.25� 0.25mm2 channel of the observation cell.

The reaction progress in a continuous-flow mixing

experiment is measured by recording the fluorescence

profile vs. distance downstream from the mixer. A

conventional light source consisting of an arc lamp (we

currently use a 350W Hg arc lamp in a lamp housing

from Oriel, Stratford, CT), collimating optics and

monochromator, is used for fluorescence excitation.Relatively uniform illumination of the flow channel over

a length of 10–15mm is achieved by means of a cylin-

drical lens. A complete fluorescence vs. distance profile

is obtained by imaging the fluorescent light emitted at a

90� angle onto the CCD detector of a digital camera

system (Micromax, Roper Scientific, Princeton NJ)

containing a UV-coated Kodak CCD chip with an array

of 1317� 1035 pixels. The camera is equipped with afused silica magnifying lens and a high-pass glass filter

or a band-pass interference filter to suppress scattered

incident light.

Fig. 4 illustrates a typical continuous-flow experi-

ment, using the quenching of NATA by NBS as a test

Fig. 4. A typical continuous-flowmixing experiment. The upper panels show ra

trace (NATA mixed with NBS). (B) fluorescence control (NATA mixed with w

(D) Corrected kinetic trace calculated according to Eq. (3) (lower trace), and a

reaction. The upper panels show raw data obtained byaveraging the intensity across the flow channel vs. the

distance d downstream from the mixer. Panel (A) shows

the intensity profile for the quenching reaction, IeðdÞ, atfinal NATA and NBS concentrations of 40 lM and

4mM, respectively (mixing 1 part of 440 lM NATA in

water with 10 parts of 4.4mM NBS in water). Panel (B)

shows the distribution of incident light intensity, IcðdÞ,measured by mixing the same NATA solution withwater. Panel (C) shows the scattering background, IbðdÞ,measured by passing water through both capillaries.

Panel (D) shows the corrected kinetic trace, flrelðtÞ,obtained according to

flrel ¼ ðIe � IbÞ=ðIc � IbÞ: ð3Þ

Distance was converted into time on the basis of the

known flow rate (0.8ml/s in this example), the cross-

sectional area of the flow channel (0.0625mm2), and the

length of the channel being imaged (12.5mm, or 9.5 lmper pixel). The signal measured at points below the en-trance to the flow channel is well fitted by a single-ex-

ponential decay. The trace at flrel ¼ 1 in Fig. 4D

represents a control for the mixing efficiency measured

as follows. The 440 lM NATA stock solution used in

the experiment above was diluted 11-fold with water and

filled into both syringes. The continuous-flow trace re-

corded at the same flow rate was then background-

corrected and normalized according to Eq. (3), using theIc trace recorded previously in which the same NATA

solution was diluted 11-fold in the capillary mixer.

w intensity profiles vs. distance from the mixing region. (A) Raw kinetic

ater). (C) Scattering background (water delivered from both syringes).

control designed to determine mixing efficiency (upper trace; see text).

H. Roder et al. / Methods 34 (2004) 15–27 21

Immediately after entering the flow channel, the ratioapproaches unity, indicating that NATA is completely

mixed with water. Alternatively, mixing efficiency could

be assessed using a very fast reaction completed within

the dead time, such as the quenching of tryptophan (or

NATA) by sodium iodide.

To estimate the dead time of the continuous-flow

experiment, we measured the pseudo-first-order NATA-

NBS quenching reaction at a final NATA concentrationof 40 lM and several NBS concentrations in the range

2–32mM. Fig. 5A shows a semilogarithmic plot of the

relative fluorescence, flrel, calculated according to Eq.

(2), along with exponential fits (solid lines). Since

flrel ¼ 1 corresponds to the unquenched NATA signal

expected at t ¼ 0, the intercept of the fits with flrel ¼ 1

indicates the time point t ¼ 0 where the mixing reaction

begins. The delay between this point and the first datapoint that falls onto the exponential fit corresponds to

the dead time of the experiment, Dtd, which in this ex-

ample is 45� 5 ls. In Fig. 5B, the rate constants ob-

tained by exponential fitting are plotted as a function of

Fig. 5. Continuous-flow measurements of the quenching of NATA

fluorescence by NBS used to determine the experimental dead-time.

(A) Semi-logarithmic plot of NATA fluorescence (>324nm) vs. time at

several NBS concentrations. (B) NATA-NBS reaction rates from ex-

ponential fitting of the data in (A) vs. NBS concentration. Linear re-

gression (line) yields a second-order rate constant of 7.9� 105 M�1 s�1.

NBS concentration. The slope of a linear fit (solid line)yields a second-order rate constant for the NBS-induced

chemical quenching of NATA of 7.9� 105 M�1 s�1,

which is consistent with data obtained by stopped-flow

measurements at lower NBS concentration [36]. This

agreement, together with the linearity of the second-

order rate plot (Fig. 5B), demonstrates the accuracy of

the kinetic data. Continuous-flow measurements on our

instrument and others built according to our designhave provided much new insight into the kinetics of

protein folding and enzyme reactions on the submilli-

scond time scale [54–63].

5. Detection methods

5.1. Tryptophan fluorescence

Table 1 lists common detection methods used in rapid

mixing studies of the kinetics of folding and other con-

formational changes in proteins. The fluorescence

emission properties of tryptophan and tyrosine side

chains provide information on the local environment of

these intrinsic chromophores. For example, a fully sol-

vent-exposed tryptophan in the denatured state of aprotein typically shows a broad emission spectrum with

a maximum near 350 nm and quantum yield of �0.14,

similar to that of free tryptophan or its derivative,

NATA. Burial of the tryptophan side chain in an apolar

environment within the native state or a compact fold-

ing intermediate can result in a substantial blue-shift of

the emission maximum (by as much as 25 nm) and en-

hanced fluorescence yield. These changes are a conse-quence of the decrease in local dielectric constant and

shielding from quenchers, such as water and polar side

chains. In other cases, close contact with a polar side-

chain or backbone moiety gives rise to a decrease in

fluorescence yield upon folding. Most polar amino acid

side chains (as well as main chain carbonyl and the

terminal amino and carboxyl groups) are known to

quench tryptophan fluorescence, probably via excited-state electron or proton transfer [64,65]. Thus, the

straightforward measurement of fluorescence intensity

vs. folding or unfolding time can provide useful infor-

mation on solvent accessibility and proximity to

quenchers of an individual fluorescence probe. Com-

plications due to the presence of multiple fluorophores

can be avoided by using mutagenesis to replace any

additional tryptophans (e.g. [66,67]). Because Trp is arelatively rare amino acid, proteins with only one tryp-

tophan are not uncommon, or in the case of Trp-free

proteins, a unique fluorophore can be introduced by

using site-directed mutagenesis (e.g. [68,69]).

The use of tryptophan fluorescence to explore early

stages of protein folding is illustrated in Fig. 6, which

shows results on staphylococcal nuclease (SNase)

Table 1

Common detection methods used in kinetic studies of protein folding

Method Probe Properties probed Sensitivity

Fluorescence Trp, Tyr Solvent shielding tertiary contacts (quenching) +++

ANS Hydrophobic clusters, collapse +++

FRET Donor–acceptor distance ++

Absorbance Trp, Tyr, cofactors Polarity, solvent effects ++

Far-UV CD Peptide bond 2� structure )Near-UV CD Tyr, Trp, co-factor Side-chain packing, mobility )Vibrational: IR res. Raman Peptide bond cofactor 2� structure metal coordination +

SAXS Heavy atoms Size (rG), shape )

Fig. 6. Folding mechanism of SNase probed by tryptophan fluores-

cence. (A) Fluorescence emission spectra of the Trp76 variant of SNase

under native and denaturing conditions (solid) and a folding inter-

mediate populated at equilibrium (dashed). The spectrum of the in-

termediate was determined by global analysis of the fluorescence

spectra as a function of urea concentration (pH 5.2, 15 �C). (B) Time-

course of folding (triggered by a pH-jump from 2 to 5.2) for wild-type

SNase (Trp140) and a single-tryptophan variant (Trp76) measured by

continuous-flow (<10�3 s) and stopped-flow (>10�3 s) fluorescence.

22 H. Roder et al. / Methods 34 (2004) 15–27

recently obtained in our laboratory [70]. A variant with

a unique tryptophan fluorophore in the N-terminalb-barrel domain (Trp76 SNase) was obtained by re-

placing the single typtophan in wild-type SNase,

Trp140, with His in combination with Trp substitution

of Phe76. The fluorescence of Trp76 is strongly en-

hanced and blue-shifted under native conditions relative

to the denatured state in the presence of urea (Fig. 6A),

indicating that upon folding the indole ring of Trp76

moves from a solvent-exposed location to an apolar

environment within the native structure. An intermedi-

ate state with a fluorescence emission spectrum similarto, but clearly distinct from, the native state was de-

tected in equilibrium unfolding experiments (dashed line

in Fig. 6A). In contrast to WT* SNase (P47G, P117G,

and H124L background), which shows no changes in

tryptophan fluorescence prior to the rate-limiting fold-

ing step (�100ms), the F76W/W140H variant shows

additional changes (enhancement) during an early

folding phase with a time constant of about 80 ls(Fig. 6B). The fact that both variants exhibit the same

number of kinetic phases with very similar rates con-

firms that the folding mechanism is not perturbed by the

F76W/W140H mutations. However, the Trp at position

76 reports on the rapid formation of a hydrophobic

cluster in the N-terminal b-sheet region while the wild-

type Trp140 is silent during this early stage of folding.

5.2. ANS fluorescence

Valuable complementary information on the forma-

tion of hydrophobic clusters at early stages of folding

can be obtained by using ANS as extrinsic fluorescence

probe [71,72]. Fig. 7 illustrates this with recent results on

the Trp76 variant of SNase introduced above. Panel (A)

shows continuous-flow measurements of the enhance-ment in ANS fluorescence that accompanies early stages

of SNase folding under native conditions (U!N), and

during formation of the A-state, the compact acid-de-

natured state of SNase (U!A). Also shown is the ki-

netics of ANS binding to the pre-formed A-state. While

the rate of ANS binding to the A-state in the presence of

1M KCl shows the linear dependence on ANS con-

centration characteristic of a second-order bindingprocess (panel (B)), the rates observed under refolding

conditions (both to the native state and compact A-

state) level off at �120 lM ANS. The limiting ANS-in-

dependent rate at higher concentrations thus is due to an

Fig. 7. An early folding event in Trp76 SNase detected by ANS fluo-

rescence. (A) ANS fluorescence changes during ANS binding/folding

of Trp76 SNase measured by continuous-flow experiments at 15 �C in

the presence of 160lM ANS. U!A, salt concentration jump from

0M to 1M KCl at pH 2.0; U!N, refolding induced by a pH-jump

from 2.0 to 5.2; A+ANS!A�ANS, ANS binding kinetics in the

presence of 1M KCl at pH 2.0; and native control, ANS binding ki-

netics under the native condition (pH 5.2). (B) ANS concentration

dependence of the rates for the major (fast) kinetic phases observed

during the U!A (m), U!N (d), and A+ANS!A�ANS (j) re-

actions.

Fig. 8. FRET-detection of an early folding intermediate in a helix-

bundle protein, ACBP. (A) Ribbon diagram of ACBP, based on an

NMR structure. The two tryptophan residues and the mutated

C-terminal isoleucine are shown in ball and stick. The two lower panels

show refolding kinetics of unmodified ACBP (B) and AEDANS-la-

beled ACBP,I86C (C) in pH 5.3 buffer containing 0.34M GuHCl at

26 �C. In both panels, data from continuous-flow (s) and stopped-flow

(,) experiments were matched and combined.

H. Roder et al. / Methods 34 (2004) 15–27 23

intramolecular conformational event that precedes ANS

binding. The rate of this process closely matches that of

the earliest phase detected by intrinsic fluorescence of

Trp76 (Fig. 6B), confirming that both processes reflect a

common early folding step. The results are consistent

with the rapid accumulation of an ensemble of states

containing a loosely packed hydrophobic core involving

primarily the b-barrel domain. In contrast, the specificinteractions in the a-helical domain involving Trp140

are formed only during the final stages of folding.

5.3. FRET

Fluorescence resonant energy transfer (FRET) can

potentially give more specific information on the chan-

ges in average distance between fluorescence donors andacceptors. For example, cytochrome c contains an in-

trinsic fluorescence donor–acceptor pair, Trp59, and the

covalently attached heme group, which quenches tryp-

tophan fluorescence via excited-state energy transfer[73]. We have made extensive use of this property to

characterize the folding mechanism of cytochrome c

[28,74,75], including the initial collapse of the chain on

the microsecond time scale [54,55].

We recently combined ultrafast mixing experiments

with FRET to monitor large-scale structure changes

during early stages of folding of acyl-CoA binding

protein (ACBP), a small (86 residue) four-helix bundleprotein [61]. ACBP contains two tryptophan residues on

adjacent turns of helix 3, which served as fluorescence

donors, and an AEDANS fluorophore covalently at-

tached to a C-terminal cystein residue (introduced by

mutation of Ile86 to Cys) was used as an acceptor

(Fig. 8A). Earlier equilibrium and kinetic studies, using

Fig. 9. Initial stages of refolding of acid-denatured oxidized

cytochrome c at pH 5 monitored by continuous-flow absorbance

measurements at different wavelengths spanning the Soret heme ab-

sorbance band. The lines represent a global fit of a four-state folding

mechanism to the family of kinetic traces.

24 H. Roder et al. / Methods 34 (2004) 15–27

intrinsic tryptophan fluorescence, showed a cooperativeunfolding transition and single-exponential (un)folding

kinetics consistent with an apparent two-state transition.

Even when using continuous-flow mixing to measure

intrinsic tryptophan fluorescence changes on the sub-

millisecond time scale (Fig. 8B), we found only minor

deviations from two-state folding behavior. However,

when we monitored the fluorescence of the C-terminal

AEDANS group while exciting the tryptophans, weobserved a large increase in fluorescence during a fast

kinetic phase with a time constant of 80 ls, followed by

a decaying phase with a time constant ranging about 10–

500ms, depending on denaturant concentration

(Fig. 8C). The large enhancement in FRET efficiency is

attributed to a major decrease in the average distance

between helix 3 and C-terminus of ACBP. The fact that

the early changes are exponential in character suggeststhat the initial compaction of the polypeptide is limited

by an energy barrier rather than chain diffusion. The

subsequent decrease in AEDANS fluorescence during

the final stages of folding is attributed to a sharp de-

crease in the intrinsic fluorescence yield of the two try-

ptophans due to intramolecular quenching. The specific

side chain interactions responsible for quenching are

established only in the close-packed native structure andare not present during the initial folding event. These

observations indicate that the early (80 ls) folding phase

marks the formation of a collapsed, but loosely packed

and highly dynamic, ensemble of states with overall di-

mensions (in terms of fluorescence donor–acceptor dis-

tance) similar to that of the native state. Accumulation

of partially structured states with some native-like fea-

tures may facilitate the search for the native conforma-tion. Because of their short lifetime and low stability,

such intermediates can easily be missed by conventional

kinetic techniques, whereas the continuous-flow FRET

technique offers the temporal resolution and structural

sensitivity to detect even marginally stable intermediates

populated during early stages of folding.

5.4. Continuous-flow absorbance

Although fluorescence is inherently more sensitive,

our capillary mixing instrument can also be adapted for

continuous-flow absorbance measurements on the mi-

crosecond time scale. The fully transparent flow cell

used for fluorescence measurements is replaced with a

custom-made partially opaque absorbance flow cell of

the same dimensions (0.25mm pathlength). Relativelyuniform illumination with minimal changes to the op-

tical arrangement (Fig. 3) was achieved by using a 2mm

fluorescence cuvette filled with a highly turbid suspen-

sion (non-dairy creamer works well) as scattering cell.

As in fluorescence measurements, a complete reaction

profile can be recorded in a single 2–3 s continuous-flow

run by imaging the flow channel onto the CCD chip.

Using the DCIP-ascorbate reaction described above (cf.Fig. 2), we measured dead times as short as 40 ls at thehighest flow rate tested (1.1ml/s).

To validate the technique, we measured the changes

in heme absorbance in the Soret region (�360–430 nm)

associated with the folding of oxidized horse cyto-

chrome c. The reaction was initiated by a rapid jump

from pH 2, where the protein is fully unfolded, to pH

4.7, where folding occurs rapidly with minimal compli-cations due to non-native histidine–heme ligands. A

series of kinetic traces covering the time window from

40 ls to 1.2ms were measured at different wavelengths

spanning the Soret region (Fig. 9). A parallel series of

stopped-flow experiments (data not shown) was per-

formed under matching conditions to extend the data to

longer times (2ms–10 s). Global fitting of the family of

kinetic traces to sums of exponential terms yielded threemajor kinetic phases with time constants of 65 ls,500 ls, and 2ms, respectively, consistent with accumu-

lation of two intermediate species, I1 and I2, with ab-

sorbance properties distinct from both the initial (U)

and final (N) states. In previous continuous-flow fluo-

rescence measurements on cytochrome c [54,55], we also

observed three kinetic phases with very similar time

constants, indicating that a basic four-state mechanismis sufficient to describe the folding process of cyto-

chrome c in the absence of complications due to

non-native heme ligation and other slow events, such as

cis–trans isomerization of peptide bonds.

H. Roder et al. / Methods 34 (2004) 15–27 25

5.5. Other detection methods for ultrafast folding studies

Continuous-flow measurements have been coupled

with several other biophysical techniques, including

resonance Raman spectroscopy [10], CD [52], EPR

[76,77], and SAXS [53,78,79]. In their pioneering work,

Takahashi et al. [10] used resonance Raman spectros-

copy to monitor changes in heme coordination during

folding of cytochrome c on the sub-milliscond timescale. Their findings confirmed and extended prior re-

sults on the involvement of heme ligation in folding of

cytochrome c, based on stopped-flow absorbance and

fluorescence measurements [80–82].

CD spectroscopy in the far-UV (peptide) region

provides an overall measure of secondary structure

content, and is thus an especially valuable technique for

protein folding studies (e.g. [30,31]). However, the lowinherent sensitivity of the technique, together with var-

ious flow artifacts, such as strain-induced birefringence,

has limited the resolution of stopped-flow CD mea-

surements to the 10 ms time range [83]. Akiyama et al.

[52] were able to extend the time resolution down to the

400 ls range by coupling an efficient turbulent mixer (T-

design) with a commercial CD spectrometer. Their

continuous-flow measurements of CD spectral changesin the far-UV region revealed the formation of (helical)

secondary structure during the second and third (final)

stages of cytochrome c folding. The same group recently

designed a mixer/flow-cell assembly with a dead time as

short as 160 ls for continuous-flow SAXS measure-

ments on a synchrotron [53]. They were thus able to

follow the changes in size (radius of gyration, Rg) and

shape (pair distribution derived from scattering profiles)associated with refolding of acid-denatured cytochrome

c under conditions similar to those used in our absor-

bance measurements (Fig. 9). The intermediate formed

within their dead time, which corresponds to the prod-

uct of the 65 ls process in Fig. 9, is substantially more

compact (Rg � 20�A) than the acid-denatured state

(Rg ¼ 24�A). This finding clearly shows that cytochrome

c undergoes a partial chain collapse during the initialfolding phase, confirming earlier fluorescence data

[54,84].

Fig. 10. A capillary mixing device for quenched-flow measurements on

the microsecond time scale. (A) A capillary mixer similar to that in Fig.

3, but without flow cell, is used to generate a fast free-flow jet. A

second mixing event occurs upon impact of the jet with a quench buffer

solution. (B) Quenched-flow NMR measurement of the H–D exchange

reaction of backbone NH groups of a model peptide (YGLFG). Ex-

trapolation of the exponential decay in normalized NH resonance in-

tensity (solid curves) yields an estimated dead time of 60 ls.

6. A quenched-flow method for H–D exchange labeling

studies on the microsecond time scale

H–D exchange labeling experiments coupled withNMR detection [27–29,85,86] are important sources of

structural information on protein folding intermediates.

These experiments generally rely on commercial

quenched-flow equipment to carry out two or three se-

quential mixing steps, which limits the time resolution to

a few milliseconds or longer. To push the dead time of

quenched-flow measurements into the microsecond time

range, we made use of our highly efficient capillarymixers [12]. The device (illustrated in Fig. 10A) uses a

quartz capillary mixer similar to that used for optical

measurements, but without observation cell, to generate

a homogeneous mixture of solutions A and B. The

mixture emerges from the capillary as a fine (200 lmdiameter) jet with a linear velocity of up to 40m/s at the

highest flow rate used (1.25ml/s). A second mixing event

can be achieved simply by injecting the jet into a testtube containing a third solution C; the extremely high

flow velocity ensures very efficient mixing. To determine

the dead time (i.e., the shortest delay between the two

mixing events), we carried out a series of H–D exchange

experiments on a pentapeptide (YGGFL). Rapid ex-

change of the backbone amide protons with solvent

26 H. Roder et al. / Methods 34 (2004) 15–27

deuterons was achieved by mixing a H2O solution of thepeptide with 5-fold excess of D2O buffered at pH� 9.7

(uncorrected pH meter reading). The exchange reaction

was quenched by injecting the mixture into ice-cooled

acetate buffer at pH� 3. Under these quench conditions,

the rate of exchange for some of the peptide NH groups

(Gly3, Phe4, and Leu5) is sufficiently slow (10, 45, and

70min, respectively) to determine their residual NH in-

tensity by recording one-dimensional 1H NMR spectra.Fig. 10B shows the results of a series of experiments in

which the capillary was raised from direct contact with

the quench solutions to a distance of about 40mm

corresponding to a ‘‘time-of-flight’’ of �1ms. For Gly3

and Phe4, exponential fits of the decay in residual NH

intensity with the incremented time delay yield exchange

rates of 5600 and 4400 s�1, respectively, in agreement

with published intrinsic exchange rates [87] (the rate ofthe C-terminal amide group is too slow to be measurable

over a 1 ms time window, and the Gly2 NH continues to

exchange under quench conditions). Extrapolation of

the fits up to the NH intensity expected at t ¼ 0 (mea-

sured in a separate control) indicates that the first

measurement corresponds to an effective exchange time

of 60� 10 ls, thus defining the dead time of the mea-

surement. We previously used a similar setup to measurethe protection of amide protons at early times of re-

folding of b-lactoglobulin [58].

Acknowledgments

This work was supported by Grants GM56250 and

CA06927 from the National Institutes of Health, Grant

MCB-079148 from the National Science Foundation,

and an Appropriation from the Commonwealth of

Pennsylvania.

References

[1] C.M. Jones, E.R. Henry, Y. Hu, C.-K. Chan, S.D. Luck, A.

Bhuyan, H. Roder, J. Hofrichter, W.A. Eaton, Proc. Natl. Acad.

Sci. USA 90 (1993) 11860–11864.

[2] R.H. Callender, R.B. Dyer, R. Gilmanshin, W.H. Woodruff,

Annu. Rev. Phys. Chem. 49 (1998) 173–202.

[3] M. Gruebele, Annu. Rev. Phys. Chem. 50 (1999) 485–516.

[4] W.A. Eaton, V. Munoz, S.J. Hagen, G.S. Jas, L.J. Lapidus, E.R.

Henry, J. Hofrichter, Annu. Rev. Biophys. Biomol. Struct. 29

(2000) 327–359.

[5] C.Y. Huang, S. He, W.F. DeGrado, D.G. McCafferty, F. Gai, J.

Am. Chem. Soc. 124 (2002) 12674–12675.

[6] H. Hartridge, F.J.W. Roughton, Proc. R. Soc. (Lond.) B Biol. Sci.

94 (1923) 336–367.

[7] Q.H. Gibson, L. Milnes, Biochem. J. 91 (1964) 161.

[8] B. Chance, Rapid Mixing and Sampling Techniques in Biochem-

istry, Academic Press, New York, 1964.

[9] P. Regenfuss, R.M. Clegg, M.J. Fulwyler, F.J. Barrantes, T.M.

Jovin, Rev. Sci. Instrum. 56 (1985) 283–290.

[10] S. Takahashi, S.-R. Yeh, T.K. Das, C.-K. Chan, D.S. Gottfried,

D.L. Rousseau, Nat. Struct. Biol. 4 (1997) 44–50.

[11] C.-K. Chan, Y. Hu, S. Takahashi, D.L. Rousseau, W.A. Eaton, J.

Hofrichter, Proc. Natl. Acad. Sci. USA 94 (1997) 1779–1784.

[12] M.C.R. Shastry, S.D. Luck, H. Roder, Biophys. J. 74 (1998)

2714–2721.

[13] K.A. Johnson, Methods Enzymol. 249 (1995) 38–61.

[14] J.B. Knight, A. Vishwanath, J.P. Brody, R.H. Austin, Phys. Rev.

Lett. 80 (1998) 3863–3866.

[15] S.A. Pabit, S.J. Hagen, Biophys. J. 83 (2002) 2872–2878.

[16] C. Tanford, K.C. Aune, A. Ikai, J. Mol. Biol. 73 (1973) 185–197.

[17] P.S. Kim, R.L. Baldwin, Annu. Rev. Biochem. 51 (1982) 459–489.

[18] C.R. Matthews, Annu. Rev. Biochem. 62 (1993) 653–683.

[19] P.A. Evans, S.E. Radford, Curr. Opin. Struct. Biol. 4 (1994) 100–

106.

[20] H. Roder, W. Col�on, Curr. Opin. Struct. Biol. 7 (1997) 15–28.

[21] H. Roder, M.C.R. Shastry, Curr. Opin. Struct. Biol. 9 (1999) 620–

626.

[22] C.R. Matthews, Methods Enzymol. 154 (1987) 498–511.

[23] A. Matouschek, J.T. Kellis Jr., L. Serrano, A.R. Fersht, Nature

340 (1989) 122–126.

[24] S.E. Jackson, N. elMasry, A.R. Fersht, Biochemistry 32 (1993)

11270–11278.

[25] H. Roder, G.A. El€ove, R.M.C. Shastry, in: R.H. Pain (Ed.),

Mechanisms of Protein Folding, Oxford University Press, New

York, 2000, pp. 65–104.

[26] R.L. Baldwin, G.D. Rose, Trends Biochem. Sci. 24 (1999) 77–83.

[27] H. Roder, K. W€uthrich, Proteins 1 (1986) 34–42.

[28] H. Roder, G.A. El€ove, S.W. Englander, Nature 335 (1988) 700–

704.

[29] J.B. Udgaonkar, R.L. Baldwin, Nature 335 (1988) 694–699.

[30] K. Kuwajima, H. Yamaya, S. Miwa, S. Sugai, T. Nagamura,

FEBS Lett. 221 (1987) 115–118.

[31] G.A. El€ove, A.F. Chaffotte, H. Roder, M.E. Goldberg, Biochem-

istry 31 (1992) 6876–6883.

[32] P.A. Jennings, P.E. Wright, Science 262 (1993) 892–895.

[33] B.E. Jones, C.R. Matthews, Protein Sci. 4 (1995) 167–177.

[34] S. Khorasanizadeh, I.D. Peters, H. Roder, Nat. Struct. Biol. 3

(1996) 193–205.

[35] W.F. Walkenhorst, J.A. Edwards, J.L. Markley, H. Roder,

Protein Sci. 11 (2002) 82–91.

[36] B.F. Peterman, Anal. Biochem. 93 (1979) 442–444.

[37] R.L. Berger, B.B.H.F. Chapman, Rev. Sci. Instrum. 39 (1968)

493–498.

[38] B.W. Zeff, D.D. Lanterman, R. McAllister, R. Roy, E.J.

Kostelich, D.P. Lathrop, Nature 421 (2003) 146–149.

[39] J.M. Beechem, L. James, L. Brand, SPIE Proc. 1204 (1990) 686–

698.

[40] B.E. Jones, J.M. Beechem, C.R. Matthews, Biochemistry 34

(1995) 1867–1877.

[41] J. Balbach, V. Forge, N.A.J. van Nuland, S.L. Winder, P.J. Hore,

C.M. Dobson, Nat. Struct. Biol. 2 (1995) 865–870.

[42] C. Frieden, Biochemistry 42 (2003) 12439–12446.

[43] D.J. Segel, A. Bachmann, J. Hofrichter, K.O. Hodgson, S.

Doniach, T. Kiefhaber, J. Mol. Biol. 288 (1999) 489–499.

[44] B. Tonomura, H. Nakatani, M. Ohnishi, J. Yamaguchi-Ito, K.

Hiromi, Anal. Biochem. 84 (1978) 370–383.

[45] P. Brissette, D.P. Ballou, V. Massey, Anal. Biochem. 181 (1989)

234–238.

[46] Q.H. Gibson, E. Antonini, Kinetics of heme–protein intractions,

in: B. Chance, R.W. Estabrook, T. Yonetani (Eds.), Hemes and

Hemeproteins, Academic Press, New York, 1966, pp. 67–78.

[47] H. Hartridge, F.J.W. Roughton, Proc. R. Soc. (Lond.): A Math.

Phys. Sci. 104 (1923) 376–394.

[48] H. Gutfreund, Methods Enzymol. 16 (1969) 229–249.

[49] S. Takahashi, Y.-c. Ching, J. Wang, D.L. Rousseau, J. Biol.

Chem. 270 (1995) 8405–8407.

H. Roder et al. / Methods 34 (2004) 15–27 27

[50] K. Paeng, I. Paeng, J. Kincaid, Anal. Sci. 10 (1994) 157–159.

[51] S.-R. Yeh, S. Takahashi, B. Fan, D.L. Rousseau, Nat. Struct.

Biol. 4 (1997) 51–56.

[52] S. Akiyama, S. Takahashi, K. Ishimori, I. Morishima, Nat. Struct.

Biol. 7 (2000) 514–520.

[53] S. Akiyama, S. Takahashi, T. Kimura, K. Ishimori, I. Morishima,

Y. Nishikawa, T. Fujisawa, Proc. Natl. Acad. Sci. USA 99 (2002)

1329–1334.

[54] M.C.R. Shastry, H. Roder, Nat. Struct. Biol. 5 (1998) 385–392.

[55] M.C.R. Shastry, J.M. Sauder, H. Roder, Accts. Chem. Res. 31

(1998) 717–725.

[56] S.-H. Park, M.C.R. Shastry, H. Roder, Nat. Struct. Biol. 6 (1999)

943–947.

[57] A.P. Capaldi, R.M.C. Shastry, C. Kleanthous, H. Roder, S.E.

Radford, Nat. Struct. Biol. 8 (2001) 68–72.

[58] K. Kuwata, R. Shastry, H. Cheng, M. Hoshino, C.A. Batt, Y.

Goto, H. Roder, Nat. Struct. Biol. 8 (2001) 151–155.

[59] N. Ferguson, C.M. Johnson, M. Macias, H. Oschkinat, A. Fersht,

Proc. Natl. Acad. Sci. USA 98 (2001) 13002–13007.

[60] T.J. Pinheiro, J.D. Venning, J.B. Jackson, J. Biol. Chem. 276

(2001) 44757–44761.

[61] K. Teilum, K. Maki, B.B. Kragelund, F.M. Poulsen, H. Roder,

Proc. Natl. Acad. Sci. USA 99 (2002) 9807–9812.

[62] S. Gianni, C. Travaglini-Allocatelli, F. Cutruzzola, M. Brunori,

M.C. Shastry, H. Roder, J. Mol. Biol. 330 (2003) 1145–1152.

[63] F. Khan, J.I. Chuang, S. Gianni, A.R. Fersht, J. Mol. Biol. 333

(2003) 169–186.

[64] Y. Chen, M.D. Barkley, Biochemistry 37 (1998) 9976–9982.

[65] M. Van Gilst, C. Tang, A. Roth, B. Hudson, Fluorescence 4

(1994) 203–207.

[66] P.L. Clark, B.F. Weston, L.M. Gierasch, Fold. Des. 3 (1998) 401–

412.

[67] X. Shao, C.R. Matthews, Biochemistry 37 (1998) 7850–7858.

[68] S. Khorasanizadeh, I.D. Peters, T.R. Butt, H. Roder, Biochem-

istry 32 (1993) 7054–7063.

[69] M.R. Eftink, M.C.R. Shastry, Methods Enzymol. 278 (1997) 258–

286.

[70] K. Maki, H. Cheng, D.A. Dolgikh, M.C.R. Shastry, H. Roder, J.

Mol. Biol., in press.

[71] G.V. Semisotnov, N.A. Rodionova, V.P. Kutyshenko, B. Ebert, J.

Blanck, O.B. Ptitsyn, FEBS Lett. 224 (1987) 9.

[72] M. Engelhard, P.A. Evans, Protein Sci. 4 (1995) 1553–

1562.

[73] T.Y. Tsong, J. Biol. Chem. 249 (1974) 1988–1990.

[74] W. Col�on, G.A. El€ove, L.P. Wakem, F. Sherman, H. Roder,

Biochemistry 35 (1996) 5538–5549.

[75] W. Col�on, H. Roder, Nat. Struct. Biol. 3 (1996) 1019–1025.

[76] V.M. Grigoryants, A.V. Veselov, C.P. Scholes, Biophys. J. 78

(2000) 2702–2708.

[77] K. DeWeerd, V. Grigoryants, Y. Sun, J.S. Fetrow, C.P. Scholes,

Biochemistry 40 (2001) 15846–15855.

[78] L. Pollack, M.W. Tate, N.C. Darnton, J.B. Knight, S.M. Gruner,

W.A. Eaton, R.H. Austin, Proc. Natl. Acad. Sci. USA 96 (1999)

10115–10117.

[79] L. Pollack, M.W. Tate, A.C. Finnefrock, C. Kalidas, S. Trotter,

N.C. Darnton, L. Lurio, R.H. Austin, C.A. Batt, S.M. Gruner,

S.G. Mochrie, Phys. Rev. Lett. 86 (2001) 4962–4965.

[80] G.A. El€ove, A.K. Bhuyan, H. Roder, Biochemistry 33 (1994)

6925–6935.

[81] T.R. Sosnick, L. Mayne, R. Hiller, S.W. Englander, Nat. Struct.

Biol. 1 (1994) 149–156.

[82] W. Col�on, L.P. Wakem, F. Sherman, H. Roder, Biochemistry 36

(1997) 12535–12541.

[83] K. Kuwajima, in: G.D. Fasman (Ed.), Circular Dichroism and the

Conformational Analysis of Biomolecules, Plenum Press, New

York, 1996.

[84] S.J. Hagen, W.A. Eaton, J. Mol. Biol. 297 (2000) 781–789.

[85] S.T. Gladwin, P.A. Evans, Fold. Des. 1 (1996) 407–417.

[86] J.M. Sauder, H. Roder, Fold. Des. 3 (1998) 293–301.

[87] Y. Bai, J.S. Milne, S.W. Englander, Proteins 17 (1993) 75–86.