Logo of biophysjGuide for AuthorsAbout this journalExplore this journalBiophysical Journal
Biophys J. 2013 Feb 5; 104(3): 694–704.
PMCID: PMC3566456
PMID: 23442920

Direct Observation of Protein Unfolded State Compaction in the Presence of Macromolecular Crowding

Therese Mikaelsson, Jörgen Ådén, Lennart B.-Å. Johansson,∗ and Pernilla Wittung-Stafshede∗∗
Author information Article notes Copyright and License information Disclaimer

Associated Data

Supplementary Materials

Abstract

Proteins fold and function in cellular environments that are crowded with other macromolecules. As a consequence of excluded volume effects, compact folded states of proteins should be indirectly stabilized due to destabilization of extended unfolded conformations. Here, we assess the role of excluded volume in terms of protein stability, structural dimensions and folding dynamics using a sugar-based crowding agent, dextran 20, and the small ribosomal protein S16 as a model system. To specifically address dimensions, we labeled the protein with BODIPY at two positions and measured Trp-BODIPY distances under different conditions. As expected, we found that dextran 20 (200 mg/ml) stabilized the variants against urea-induced unfolding. At conditions where the protein is unfolded, Förster resonance energy transfer measurements reveal that in the presence of dextran, the unfolded ensemble is more compact and there is residual structure left as probed by far-ultraviolet circular dichroism. In the presence of a crowding agent, folding rates are faster in the two-state regime, and at low denaturant concentrations, a kinetic intermediate is favored. Our study provides direct evidence for protein unfolded-state compaction in the presence of macromolecular crowding along with its energetic and kinetic consequences.

Introduction

To function, proteins must fold from extended unfolded states to compact unique structures that are biologically active. During the last three decades, through pioneering in vitro and in silico work using, for example, protein engineering (1) and energy landscape framework (2), significant progress has been made to pinpoint mechanisms and driving forces important for protein folding. However, in reality, proteins fold inside cells, where the environment is very different from the dilute buffer solutions mostly used in in vitro experiments. The intracellular environment is highly crowded due to the presence of large amounts of macromolecules, including proteins, nucleic acids, ribosomes, and carbohydrates. This means that a significant fraction of the intracellular space is not available to other macromolecular species. It has been estimated that the concentration of macromolecules in the cytoplasm ranges from 80 to 400 mg/ml (3). All macromolecules in physiological fluids collectively occupy between 10% and 40% of the total aqua-based volume (4). The crowded environment results in increased viscosity, excluded volume effects, and the opportunity for specific, as well as nonspecific, intermolecular interactions. To mimic excluded-volume effects in in vitro experiments, one may use so-called macromolecular crowding agents, which are inert, noncharged polymers of defined size (i.e., dextrans, Ficoll) that occupy space but do not interact with target proteins.

Due to excluded-volume effects, any reaction resulting in a volume change will be affected by macromolecular crowding (5,6). It has been suggested that macromolecular crowding provides a stabilizing effect to the folded states of proteins indirectly due to destabilization of the more extended and malleable denatured states (7,8). Many in vitro experiments have shown that protein stability is increased in the presence of macromolecular crowding agents (8–11). However, the underlying mechanisms for this observation are not clear. Conceptually, excluded-volume effects can simply shift the equilibrium toward the folded state of a protein (F), since the chemical potential will be raised more for the unfolded state than for the folded state in a crowded environment. Alternatively, there may be a direct effect of crowding on the unfolded state (U), forcing it to become more compact at crowded conditions. In the latter scenario, the chemical potential of the unfolded state is still raised more than that for the folded state (since U is still larger than F), and again the folded-state stability will be increased. The major goal here is to test the hypothesis that macromolecular crowding results in more compacted unfolded-state ensembles by using a direct experimental methodology.

Compaction of the unfolded state as a result of macromolecular crowding has been suggested based on a number of experimental studies. For example, small-angle neutron scattering experiments on random polymers (12) as well as on intrinsically disordered proteins (13) indicate compaction in crowding. Also, coarse-grained molecular simulations have demonstrated that the radius of gyration values decreased in silico for the unfolded forms of apoflavodoxin, VlsE, and calmodulin in the presence of hard spheres that mimic crowding (9,10,14). In addition, there are studies demonstrating that proteins can convert from unfolded to folded or molten-globule states upon addition of large amounts of crowding agents. For instance, unfolded cytochrome c at pH 2 can adopt a molten globule structure in the presence of crowding agents (11), unfolded RNase A at pH 3 adopts a folded-like structure upon addition of 350 mg/ml PEG 20,000 or Ficoll 70 (15), and the reduced and carboxyamidated form of RNase T1 that is intrinsically unstructured at pH 7 was found to exhibit some catalytical activity upon the addition of 400 mg/ml dextran 70 (16). Using Förster resonance energy transfer (FRET) between two proteins covalently attached at each end of the PGK polypeptide, PGK folded-state compaction was observed in the presence of Ficoll 70 (17). However, in these studies, the unfolded-state compaction was not investigated directly. In a study performed in 2008 (18), fluorescence experiments indicated that unfolded Azotobacter vinelandii apoflavodoxin was compacted in the presence of crowders, although it was noted that this may be due to the coexistence of intermediate states.

In 2010, Hong and Gierasch provided indirect evidence of protein unfolded-state compaction in a crowded environment in vitro (19). For this, they probed changes in Cys and Trp residue accessibility in the CRAB I protein as the means to demonstrate compaction. Here, we applied intramolecular pairwise FRET between residues to directly probe two pairs of distances in the unfolded ensemble of the ribosomal protein S16, with and without macromolecular crowding agents. We find that the unfolded ensemble is more compact in the presence of dextran 20 as compared to the buffer system. We also report the consequences of this unfolded-state compaction on S16 folding dynamics.

Materials and Methods

Chemicals

Dextran 20 of technical grade was purchased from Pharmacosmos A/S (Holbaek, Denmark). Ultrapure urea was from MP Biomedicals (Santa Ana, CA), NaCl for analysis from Merck (Whitehouse Station, NJ), and NaAc anhydrous for analysis from Riedel-de Haën (Seelze, Germany). N-((4,4-difluoro-5,7-dimethyl-4-bora-3a,4a-diaza-s-indacene-3-yl)methyl)iodoacetamid (BODIPY) was purchased from Invitrogen (Carlsbad, CA). Water was run through a Milli-Q (Bedford, MA) water purification system with a resistivity ≥ 18 MΩ cm−1.

Sample preparation

The S16 mutants were prepared as previously described (20). The buffer contained 30 mM NaAc and 50 mM NaCl at pH 5.5, with or without addition of 200 mg/ml dextran 20 and with varying urea concentrations. To avoid the complication that volumes of different solutions are not additive, the urea concentration was expressed as moles/kg H2O in most experiments. All samples were left to equilibrate for at least 30 min before the measurements were performed and the stability over time was tested. The distance and viscosity measurements were repeated at least once.

Absorption and fluorescence

The absorption spectra were measured on a Cary 5000 UV-VIS-NIR spectrophotometer (Varian, Sydney, Australia). Corrected steady-state fluorescence spectra were recorded on a Spex Fluorolog 112 (Spex, Evanston, IL) using a spectral bandwidth of 2 nm. All measurements were performed at 20°C and the absorbance of BODIPY was kept below 0.08. For all types of fluorescence experiments, protein concentration was ∼4 μM.

Fluorescence lifetime and depolarization measurements

Fluorescence decays were collected by means of the time-correlated single-photon counting technique, using a PRA 3000 system (PRA, Victoria, BC, Canada). The excitation source used for tryptophan was a NanoLED-15 (IBH) with the excitation-peak maximum centered at ∼280 nm operating at 800 kHz. BODIPY was excited by a PicoQuant pulsed diode laser with the peak excitation maximum at 470 nm and operated at 2.5 MHz. Due to the high intensity of the PicoQuant pulsed diode laser, an interference filter centered at 470 nm (full width at half-maximum (FWHM) = 9.3 nm) was used in combination with neutral density filters in the excitation path. The appropriate emission wavelength was collected using interference filters centered at 334 nm (FWHM = 8.9 nm) for tryptophan and at 520 nm (FWHM = 28 nm) for BODIPY. To avoid pile-up distortion, the count rate was kept below 1.5% of the repetition rate. All lifetime measurements were performed at the magic-angle condition. The resolution was 26 ps/ch and the fluorescence decay was collected over 2048 channels with 20,000 photons in the peak maximum.

The time-resolved fluorescence depolarization data were collected with the excitation polarizer set in vertical (V) position and the emission polarizer rotated periodically between vertical and horizontal (H) positions. For each setting of the polarizers, photons were collected for a period of 90 s until at least 100,000 counts were achieved in the difference between the decays obtained with the polarizers set at VV and VH. From the experimental decay curves, a sum curve s(t) = FVV(t) + 2GFVH(t) and a difference curve d(t) = FVV(t) − GFVH(t) were constructed, and the corresponding anisotropy decay curve r(t) was calculated by dividing d(t) by s(t). The correction factor G accounts for the different experimental sensitivity to detecting vertically and horizontally polarized light (21).

Far-ultraviolet circular dichroism

Far-ultraviolet (UV) circular dichroism (CD) spectra (200–300 nm, 1-mm cell) were recorded with a J-810 Spectropolarimeter (Jasco, Tokyo, Japan) equipped with a Peltier element for temperature control. The protein concentration was ∼25 μM.

Viscosity

The absolute viscosity was measured on a Brookfield DV-II + Pro Viscometer (Brookfield, New York), connected to a Neslab Endocal RTE-110 temperature bath, using the Small Sample Adaptor and Spindle SC4-18. The shear rate was varied (30–200 rpm) to verify Newtonian behavior.

Stopped-flow mixing

Folding/unfolding dynamics data for S16 (WT and F10C/W74F) were acquired using stopped-flow mixing on an SX-17MV spectrofluorometer (Applied Photophysics, London, UK) equipped with a 320-nm long-pass filter to select for tryptophan emission. The initial protein concentration was 16 μM, and unfolded samples were left to equilibrate at least 30 min before measurements. Final protein concentration in the kinetic folding/unfolding experiments was 1.5 μM.

Theoretical Prerequisites and Data Analysis

Equilibrium unfolding

Assuming a two-state unfolding mechanism, it is found that the standard free energy of denaturation (ΔG°) varies linearly with urea concentration. The conformational stability of the protein in the absence of denaturant can then be estimated by assuming a linear dependence to 0 M denaturant (22). This model was fitted to the data by using a nonlinear least-squares method (23).

Analyses of Förster resonance electronic energy transfer

Assuming weak dipole-dipole coupling between the electronically excited donor (D) and acceptor (A) in its electronic ground state, the rate of energy transfer, ω, is given by

ω=3〈κ2〉2τD0(R0DR)6.
(1)

Here, 〈κ2〉,τD0,R0D, and R stand for the averaged square of the angular part or the dipole-dipole interaction, the radiative lifetime of the donor, the Förster radius, and the distance between the center of masses of the donor and acceptor, respectively. The value of 〈κ2〉 is taken to be the isotropic dynamic average, 2/3 (24,25), and the radiative lifetime of tryptophan is 18.9 ns (26). The Förster radius was calculated for each sample according to

R0D6=9000(ln10)(2/3)J128π5n4NA,
(2)

where the overlap integral is defined by

J = ∫FD(λ)εA(λ)λ4dλ.
(3)

In Eqs. 2 and 3, n, NA, FD(λ), and εA(λ) denote the refractive index of the medium, Avogadro's constant, the normalized donor fluorescence spectrum, and the acceptor extinction coefficient spectrum, respectively. Refractive indices were measured for each sample composition at 589 nm using a refractometer. The time-resolved fluorescence of the acceptor was analyzed according to a previously described procedure (27). This allows testing a single D-A distance, as well as a Gaussian distribution. In the Gaussian distribution,

ρ(R)=1σ2πexp[−(R−R¯)22σ2],
(4)

R¯ denotes the average distance between the centers of mass of the donor and acceptor groups, and the variance, σ, is related to the FWHM of the distribution according to

FWHM=2σ2ln2≈2.355σ.
(5)

The experimental data were analyzed by using a common deconvolution procedure of nonlinear least-squares analyses, which is based on the Levenberg-Marquardt algorithm (28).

Time-resolved fluorescence anisotropy

A BODIPY group covalently bound to a protein undergoes various intramolecular reorienting motions described by ρintra(t). These are local reorientations [ρloc(t)] with respect to the binding site and overall segmental reorientations [ρsegm(t)] of this site. The protein may also undergo slow tumbling motions on the fluorescence timescale, which are accounted for by ρtumb(t). For globular proteins whose tumbling resembles that of a spherical particle, the reorientations ρtumb(t) and ρintra(t) are independent (29). Therefore, the time-dependent fluorescence anisotropy (r(t)) can be written

r(t) = r(0)ρintra(t)ρtumb(t).
(6a)

Here, the initial anisotropy is r(0) ≤ r0, where r0 is the fundamental anisotropy (30,31). The tumbling rotation is very well approximated by ρtumb(t) = exp(−t/θtumb), where θtumb is the rotational correlation time of the protein molecule. The intramolecular reorientation is here modeled by

ρintra(t) = ρloc(t)ρsegm(t)
(6b)

and

ρloc(t)=∑iρiexp(−t/θi)+ρ∞.
(6c)

In Eq. 6c, θi are the local correlation times and ρ∞ denotes the residual anisotropy (29,32). The segmental reorientations are described by an effective rotational correlation time according to

ρsegm(t)=exp(−tθsegm).
(6d)

It is reasonable to assume that correlation times for local rotations are much shorter than the segmental ones. A combination of Eqs. 6b–6d then yields

ρintra(t)≃∑iρiexp(−tθi)+ρ∞exp(−tθsegm).
(7)

From depolarization experiments, the anisotropy decay has been analyzed. The initial and fast decay is then ascribed to the sum in Eq. 7. From the obtained best fit, an average rotational correlation time was calculated (see Table 3), which is given by

Θfast=∑iρiθi2∑iρiθi.
(8)

The slow anisotropy relaxation is described by the second term of Eq. 7, i.e.,

r(t)=r(0)ρ∞exp(−tΘslow).
(9)

In Eq. 9, the correlation time is Θslow = (1/θsegm+1/θtumb)−1. Taken together, the overall anisotropy decay is conveniently described by

r(t)=r(0)∑iρiexp(−tθi)+r0ρ∞exp(−tΘslow).
(10)

This equation was applied to analyze the depolarization data when assuming two fast correlation times (θ1 and θ2).

Table 3

Fast (Θfast) and slow (Θslow) rotational correlation times for BODIPY when attached to S16 at different solvent conditions

Solvent[Urea] (mol/kg H2O)Θfast (ns)
Θslow (ns)
Θfast (ns)
Θslow (ns)
Viscosity (cP)
W74CF10C/W74F
Buffer01.29.01.37.31.2
21.39.82.29.21.4
81.18.11.15.41.6
101.18.30.62.81.7
121.26.7——1.8
Dextran01.715.34.625.711.9
21.414.63.323.013.7
81.211.82.617.220.0
101.18.20.84.821.7
121.27.0——22.6

Correlation times were determined from the measured anisotropy decay. Θfast was calculated according to Eq. 8. The quality of fitting Eq. 10 to the depolarization data was acceptable in all cases, as judged from the statistical test parameters (χ2 and D.W. in Table 2), which were 1.00 ≤ χ2 ≤ 1.20 and 1.8 ≤ D.W. ≤ 2.1. The residual plots were examined, and are exemplified in Fig. S3.

Kinetic unfolding/folding data

Logarithms of rate constants were fitted to Eq. 11 (33). This equation includes a quadratic parameter, mf′, which accounts for curvature in the refolding limb:

ln kobs = ln(exp{lnkf + mf[D] + mf′[D]2} + exp{lnku + mu[D]}).
(11)

Here, kf and ku are the rate constants for folding and unfolding kinetics, whereas mf and mu are the denaturant dependence of the folding and unfolding limbs, respectively.

Results and Discussion

We chose dextran 20 (20 kDa) as a crowder, since its size is more similar to the model protein S16 (13 kDa) than other commonly used larger crowders, such as dextran 70 or Ficoll 70. It has been proposed that when the protein and crowder are more similar in size, the larger the excluded-volume effect becomes (34). The concentration of 200 mg/ml was chosen to be a condition with a high degree of crowding (i.e., in the biological range), but still a condition where biophysical, kinetic experiments can be performed. To study intraprotein distances in perpendicular directions (based on the crystal structure), we selected two S16 variants for FRET studies (Fig. 1): for W74C, where Cys74 was labeled with BODIPY, the Trp58-Cys74 distance was probed, and for F10C/W74F, Cys10 was labeled with BODIPY and the Trp58-Cys10 distance was measured. It was demonstrated in earlier work that these mutations and their labeling do not disturb the folded structure of the protein (20). In that study, moreover, these S16 variants were shown to unfold in apparent two-state equilibrium reactions (20).

An external file that holds a picture, illustration, etc. Object name is gr1.jpg

Intramolecular distances displayed on the crystal structure of wild-type S16 (20). The side chain of Trp58 (FRET donor) is colored in green. The dashed lines show the distances between the donor and the two possible acceptor sites (orange). To measure FRET distances, in individual proteins, residues 10 and 74 were changed to cysteines that were subsequently labeled with BODIPY.

Equilibrium unfolding

We find that unfolding curves as a function of urea for the two variants and wild-type S16, probed by changes in far-UV CD, can be approximated as two-state reactions with sloping baselines (Fig. 2). In the presence of 200 mg/ml dextran, the curves remain two-state-like but are shifted toward higher urea concentrations. This indicates that the variants are stabilized in the presence of crowding agent (ΔG and midpoint values are summarized in Table 1). It is important to note that in the presence of crowding agents, there is some volume excluded to the solvent, and thus, the effective urea concentration will be somewhat higher than the concentration derived based on total sample volume. We (34) and others (19) have earlier addressed this issue and proposed a way to correct the denaturant concentrations based on the partial specific volume of the crowding agent. Such corrections affect the slope of the unfolding curve (i.e., the m value) but not the extrapolated ΔG value. Another way to plot the unfolding data is to express the denaturant concentration in the unit of molal (Fig. 2). Using this approach, the unfolding data are presented as a function of mole urea/kg water added. Thus, contributions from the dissolved urea and dextran have been removed.

An external file that holds a picture, illustration, etc. Object name is gr2.jpg

Equilibrium unfolding curves for wild-type, W74C, and F10C/W74F in the presence (open circles) and absence (solid circles) of 200 mg/mL dextran. Data are plotted as a function of urea concentration in mol/L (top) and mol/kg H2O (bottom). The urea concentrations in molar for the dextran samples have been corrected for solvent exclusion effects (see text).

Table 1

Thermodynamic parameters for equilibrium unfolding of WT and mutant S16 proteins at 20°C

SolventWT
W74C
F10C/W74F
BufferDextranBufferDextranBufferDextran
ΔG (kJ/mol)18.3 ± 2.230.9 ± 3.721.4 ± 2.332.4 ± 8.613.7 ± 2.620.7 ± 3.7
m (kJ/mol/M)3.8 ± 0.45.3 ±0.74.2 ± 0.44.9 ± 1.33.5 ± 0.73.9 ± 0.7
[Urea]1/25.0 ± 0.15.4 ± 0.35.2 ± 0.26.4 ± 0.23.5 ± 0.25.4 ± 0.3

Equilibrium unfolding curves, derived from far-UV CD at 220 nm, were fitted to a two-state equation with sloping baselines. Before analysis, the urea concentrations in the dextran samples were corrected for solvent exclusion as described in the text.

Trp-to-BODIPY electronic energy transfer

To address distances between selected positions in S16 we turned to time-resolved fluorescence decay measurements. Upon electronic excitation of Trp58 in S16, electronic energy is transferred to the acceptor group (BODIPY). The BODIPY group is here attached either to an engineered Cys at position 10, or at 74 in the primary structure. From the x-ray structure (20), one ascertains that the distances from the Cα atom of the donor to corresponding atoms of the acceptor are 22 and 21 Å, respectively. We here choose to compare the protein in buffer and in dextran at similar molal urea conditions. Other possibilities would be to choose similar corrected molar urea conditions, or conditions at which the protein is unfolded to the same degree (such as 90% unfolded) in both cases. Since the intention is to study the unfolded ensemble, the latter option is not useful. Below, for all conditions given in molal urea, we also report the corresponding corrected molar value. Energy-transfer measurements were performed for both variants at 0, 2, 8, 10, and 12 m urea with and without 200 mg/ml dextran. (The values 0, 2, 8, 10, and 12 m urea correspond to 0, 1.9, 5.8, 6.9, and 7.7 M urea in buffer samples and 0, 1.7, 5.7, 7.2, 8.3 M urea in dextran samples.)

The Trp moiety was excited and the time-resolved fluorescence of BODIPY was monitored. Examples of fluorescence decays together with the best fits of models are displayed in Figs. S1 and S2 in the Supporting Material. In the data analyses the D-A distances are assumed either to be constant (R), or to obey a Gaussian distribution. In the latter case, one obtains average distances (R¯) and variance (σ) (Table 2 and Fig. 3). A single D-A distance very well describes the fluorescence relaxation for both mutants at zero or low urea concentrations in the absence of dextran, whereas at higher urea concentrations, an R-distribution is needed to obtain a statistically acceptable agreement with the experiments. In the presence of dextran, a single distance can still be fitted to the data at 8 m, whereas a distance distribution is required at 10 m urea. In the presence and absence of dextran, the Trp-BODIPY distances are very similar at low urea concentrations, indicating that the folded structures are the same with and without crowding agent. At denaturing conditions, the distances are shorter in the presence of dextran than in buffer for both variants.

An external file that holds a picture, illustration, etc. Object name is gr3.jpg

FRET distances fitted as a Gaussian distribution (lines) with an average distance (circles) between the centers of mass of Trp58 and BODIPY for W74C (left) and F10C/W74F (right) in the absence (black lines and circles) and presence (gray lines and circles) of 200 mg/ml dextran at varying concentrations of urea, as indicated in the figure. The areas of the Gaussian distribution have been normalized to unity in all cases.

Table 2

Distance between the centers of mass of Trp58 and BODIPY as determined by the rate of energy transfer at 20°C

S16 mutantSolvent[Urea] (mol/kg H2O)R0D (Å)R (Å)R¯ (Å)σ (Å)χ2D.W.
W74CBuffer032.721.921.32.21.071.90
232.522.021.91.31.071.96
832.4—24.53.21.141.93
1032.6—26.75.21.091.97
1232.6—31.15.91.121.87
Dextran032.621.821.51.51.102.01
232.422.021.71.91.071.95
832.423.322.83.31.091.88
1032.5—25.45.21.102.02
1232.9—30.34.81.191.83
F10C/W74FBuffer033.421.221.21.01.052.01
233.321.521.51.11.051.90
833.4—26.46.91.061.89
1033.3—29.26.01.081.88
Dextran033.221.721.71.21.132.02
233.121.821.91.51.021.95
833.422.422.41.21.111.95
1033.7—26.19.01.061.87

The Förster radius (R0D) was calculated for each sample and the distances were analyzed assuming one fixed distance (R) and as a Gaussian distribution of distances with an average distance (R¯) and a variance (σ). The quality of the fit was judged by the statistical parameters χ2 and Durbin-Watson (D.W.).

For the F10C/W74F variant at 8 m urea and in the presence of dextran, which represents an early point on the unfolded-state baseline (Fig. 2), the measured distance indicates a protein as compact as in the folded state. At 10 m urea with dextran, i.e., further along the unfolded-state baseline, the mean distance has increased but is still shorter than that obtained under the same condition in buffer. For the W74C variant, the same trend emerges. At 8 m and 10 m urea, the conditions represent the middle of the unfolding transition in the presence of crowding, but the end of the transition in buffer (Fig. 2). As expected based on the difference in unfolding transitions, the measured FRET distances at 8 m and 10 m are lower in the presence of crowding than in its absence. Nonetheless, at 12 m, the protein is unfolded at both conditions and the distance is somewhat shorter in the presence of crowding. The distance data are illustrated graphically in Fig. 3.

The Stokes shift of the tryptophan is sensitive to the polarity of its environment. With increasing polarity, the peak fluorescence shifts from ∼308 to 350 nm (35). The peak-fluorescence maximum of the Trp58 as a function of urea correlates with the compaction data (Fig. 4). At conditions where each variant is more compact in dextran than in buffer, the tryptophan emission maximum is found at a shorter wavelength in dextran than in buffer conditions.

An external file that holds a picture, illustration, etc. Object name is gr4.jpg

Tryptophan emission maxima as a function of urea concentration (molal) in the presence (open circles) and absence (solid circles) of 200 mg/mL dextran for W74F and F10C/W74F. The corresponding fraction of unfolding in the presence (gray line) and absence (black line) of 200 mg/mL dextran is displayed in the background.

Unfolded-state properties

One might expect that unfolded-state compaction correlates with lower m values in the equilibrium unfolding curves, since the m values report on changes in the surface-exposed area upon unfolding (36). If the unfolded state is more compact, it appears reasonable that a reduced surface area is exposed upon protein unfolding. However, the m values for unfolding in the presence of dextran are similar (or somewhat higher) to those found in buffer (Table 1). A possible explanation is that despite unfolded-state compaction, in terms of the two specific dimensions probed, significant surface area is still exposed due to a rugged exterior of the conformers within the unfolded-state ensemble.

Concurrent with the observed compaction based on FRET-derived distances, we observe a residual structure in the unfolded state of the F10C/W74F variant based on the far-UV CD spectra (Fig. 5). At 8 m and 10 m urea, the CD signal is more negative in the presence of crowder, indicative of residual structure. The W74C variant barely reaches the unfolded state in our experiments, and a comparison of CD spectra for the unfolded states is only possible at 12 m. For wild-type S16 also, additional CD amplitude is found at crowded conditions corresponding to the unfolded-state baseline (i.e., 10 m and 12 m).

An external file that holds a picture, illustration, etc. Object name is gr5.jpg

Far-UV CD spectra of S16 at different molal of urea (0 m (solid line), 2 m (long-dashed line), 8 m (short-dashed line), 10 m (dotted line), and 12 m (dash-dotted line)) with and without dextran for wild-type, W74C, and F10C/W74F.

Fluorescence anisotropy

The fluorescence anisotropy was determined for the two BODIPY-labeled S16 mutants in the presence and absence of dextran and at varying urea concentrations (an example of the experimental data is shown in Fig. S3). The time-resolved anisotropy data were analyzed using the model described above in the section Theoretical prerequisites. These analyses yield values of fast and slow rotational correlation times (cf. Eqs. 8 and 10). The obtained results are summarized in Table 3, and an example of a more detailed analysis can be found in Table S1. Clearly the short and long rotational correlation times are similar for both mutants in the absence of dextran and at low urea concentrations. The obtained value of Θslow ≈ 9 ns correlates with the protein tumbling-rate constant measured earlier by NMR (data not shown). However, at urea concentrations >8 m, the values of Θfast and Θslow decrease, and this is even more pronounced for the double mutant. The decreased values are compatible with polypeptide unfolding, which increases the BODIPY group dynamics. The accompanying decrease of Θslow is most likely due to shorter values of θsegm, i.e., more rapid segmental motions.

In the presence of dextran and at lower urea concentrations, Θfast and Θslow are longer compared to the corresponding dextran-free systems. This is most evident for the long relaxation times. In contrast to homogenous solutions with small solutes, in dextran there is a nonlinear dependence on the measured bulk viscosity (η) for Θslow(i.e., Θslow ≠ constant × η) such that the S16 tumbling rates detected are faster than those expected on the basis of macroscopic viscosities. This means that the microviscosity experienced by the protein during tumbling is lower than the bulk viscosity. This has been reported previously for both translational and rotational microviscosities in the presence of macromolecular crowding agents (37). That the effect of large molecules (such as crowders) on solvent viscosity is mostly macroscopic was noted many years ago, using, for example, barstar and barnase and poly-vinyl-pyrollidone (38). Upon increasing urea concentration the values of Θslow show a twofold decrease, which again is compatible with an increased influence of segmental motions.

Folding and unfolding dynamics

To reveal how the crowding-induced compaction of unfolded S16 changes its folding dynamics, we collected S16 folding and unfolding kinetic data as a function of urea by means of stopped-flow experiments (Fig. 6, Table S2). To do so, we monitored the changes in the Trp signal of the wild-type and F10C/W74F proteins as previously reported (20). It was earlier reported that the kinetic folding reaction of thermophilic S16 is not two-state. Instead, at low denaturant concentrations, a kinetic intermediate formed during the stopped-flow mixing step, and thus, the kinetic reaction is at least three-state. As a result of the change in mechanism between low and high urea concentrations, there is a detectable curvature in the folding arm at low urea concentrations (i.e., the mf value varies with the urea concentration) in the chevron plot (20).

An external file that holds a picture, illustration, etc. Object name is gr6.jpg

Chevron plots for folding/unfolding dynamics of wild-type S16 in buffer and in crowding conditions. (Left) Natural logarithms of folding- and unfolding-rate constants are shown as a function of urea for wild-type S16 and F10C/W74F. Solid circles represent buffer conditions and open circles crowded conditions (200 mg/mL dextran). The urea concentrations in the dextran samples have been corrected for solvent exclusion. (Upper left) Data are fitted to Eq. 11. (Lower left) Data points in linear regions are fitted to straight lines, and linear extrapolations of these to 0 M urea were used to determine ln kf and ln ku (Table S1). (Right) Change in fluorescence amplitude during each of the kinetic reactions shown at left, reported as the percent change of the initial signal, for wild-type (upper) and F10C/W74F variant (lower).

In our kinetic experiments with S16 (wild-type and F10C/W74F) at dilute conditions, as expected, there is minor curvature in the folding arm at low urea concentrations. In the presence of 200 mg/ml dextran, the folding rate is faster at intermediate urea concentrations (2–4 M) in both cases, but at urea concentrations approaching zero, the crowding data reach the same folding rates as found in buffer. It appears that when the unfolded state is more compact due to the presence of crowding agents, folding in the two-state regime becomes faster. This is in agreement with the equilibrium stabilization of the protein in crowding originating from a destabilized unfolded state. Unfolded-state destabilization will reduce the height of the barrier toward folding, since the unfolded state is shifted up in energy, closer to the transition-state energy level. In contrast, at conditions where the intermediate is favored, the reaction from the intermediate to folded state is not much affected by the crowding agent. Notably, there is more pronounced curvature in the folding arm in the presence of dextran than in its absence. For F10C/W74F, there is a distinct kink at ∼3.5 M urea. This indicates that the intermediate pathway is favored under crowded conditions and at low denaturant concentrations. Thus, the crowding-induced compacted unfolded state is more prone to rapid conversion to the intermediate state than is the more expanded unfolded state. The next step, intermediate to folded state, appears similar in terms of kinetics for both buffer and crowded conditions. This result is reasonable if the intermediate is roughly as compact as the folded state and, thus, excluded-volume effects will play no role. The mf values as a function of urea concentration were derived (tangent in each point in the curves in Fig. 6) along with apparent Tanford βT-values (corresponding to the folding-transition-state placement if it was a two-state reaction) for the wild-type data. The different apparent dependence of mf and βT on denaturant concentration for the protein in crowding versus buffer conditions (Fig. S4) provides further support for the idea that the burst-phase intermediate becomes more populated in the presence of crowding.

We note that there is no missing amplitude in the stopped-flow experiments (Fig. 6). This suggests that the intermediate formed during mixing when jumps are made to lower urea concentrations has the same Trp emission as the unfolded state. We also compared refolding kinetics from protein unfolded at crowded conditions in 6.9 M urea (9.6 molal urea) and in 9.2 M urea (15.9 molal urea), since we found a urea dependence of the unfolded-state FRET distances. It appears that folding from 6.9 M urea (+ crowding) is similar to folding from 9.2 M urea (+ crowding) at all urea concentrations (data not shown).

For the unfolding arm in the chevron plot, the kinetics in the presence of crowders is slower at all urea concentrations than unfolding in the absence of crowders. However, when extrapolated to 0 M urea, unfolding rates with and without dextran are similar. In other studies, it has been found that unfolding rates are the same with and without crowding (8,9), whereas in at least one case (19), unfolding is slower in the presence of crowding. In the latter study, increased viscosity due to the presence of crowding agents was suggested as an explanation. The bulk viscosity for 200 mg/ml dextran 20 used in our experiments is 10 times higher than in water (Table 3). However, it has been shown that the microviscosity, on the time- and lengthscales relevant for folding/unfolding dynamics, is not increased more than two- to threefold in solutions of macromolecular crowding agents as compared to dilute buffer solutions (39–43). This is supported by our anisotropy measurements (see above) that relate to the nanosecond timescale.

Our unfolding-kinetics data show that with crowding, there is a reduced response to urea concentration, resulting in increased kinetic stability of the folded form. The reduced denaturant response corresponds with a lower mu value for unfolding in crowding compared to in buffer (∼40% lower; Table S2). The mu value relates to the change in exposed surface area going from the folded to the transition state, and thus, our observation indicates that the transition state during unfolding is more compact in the presence of crowding.

Concluding Remarks

Proteins are produced and function in cellular environments that are highly crowded with other macromolecules. Macromolecular crowding results in excluded volume effects, increased viscosity, and increased opportunities for nonspecific and specific interactions compared to dilute solution. Here, we have focused on excluded volume effects on protein biophysical properties using dextran as the crowding agent and S16 as the model protein. For these types of experiments, it is important that the model system is well characterized in dilute conditions and unfolds reversibly under crowded conditions. These criteria are fulfilled for S16. Using FRET, we demonstrate explicitly that the unfolded ensemble of S16 is more compact in crowded conditions than in dilute solutions. Thus, in response to the increased chemical potential of extended conformations due to excluded volume, the system responds with structural compaction. Since we studied two specific dimensions in the protein, we can conclude that the unfolded-state compaction is not symmetrical but seems to depend on polypeptide sequence and residual structure preferences. We found that in the presence of dextran and at higher molal urea, the Trp58-F10C distance remained shorter than the Trp58-W74C distance. Based on this finding, we speculate that residual structure is more easily induced by crowding in the N-terminal part of S16 than in the C-terminal part. Regardless of the molecular details, unfolded-state compaction will result in decreased entropy of the unfolded state, and this stabilizes folded S16, since the entropic penalty for folding will be reduced. The stability increase for S16 at crowded conditions can be deconvoluted into kinetic effects: folding kinetics becomes faster in the two-state regime and unfolding in the presence of urea becomes slower. When folding involves a kinetic intermediate, crowding instead biases the reaction toward the intermediate pathway. The speed of the intermediate-to-folded reaction is not affected by crowding, which implies that the kinetic intermediate is almost as compact as the folded state.

It was earlier reported that the unfolded state of the thermophilic S16 used here is more compact than the unfolded state of the mesophilic homolog in dilute conditions (20). Our observation of further compaction in the presence of crowding for thermophilic S16 suggests that the same unfolded polypeptide chain can span a large range of structural dimensions depending on sequence details and solution conditions. Since the mesophilic S16 protein has a more extended unfolded ensemble, we predict that crowding will have even larger effects on its unfolded state (work in progress). We propose that unfolded-state compaction is a general phenomenon that should be investigated when considering folding reactions in vivo.

Acknowledgments

We thank Magnus Wolf-Watz for helpful discussions and plasmids of the protein constructs used.

The Swedish Natural Research Council (P.W.S., L.J.), the Knut and Alice Wallenberg Foundation (P.W.S.), Göran Gustafsson Foundation (P.W.S.), and Umeå University (P.W.S.) provided financial support.

Supporting Material

Document S1. Four figures and two tables:

References

1. Matouschek A., Kellis J.T., Jr., Fersht A.R. Mapping the transition state and pathway of protein folding by protein engineering. Nature. 1989;340:122–126. [PubMed] [Google Scholar]
2. Bryngelson J.D., Onuchic J.N., Wolynes P.G. Funnels, pathways, and the energy landscape of protein folding: a synthesis. Proteins. 1995;21:167–195. [PubMed] [Google Scholar]
3. Rivas G., Ferrone F., Herzfeld J. Life in a crowded world. EMBO Rep. 2004;5:23–27. [PMC free article] [PubMed] [Google Scholar]
4. Ellis R.J., Minton A.P. Cell biology: join the crowd. Nature. 2003;425:27–28. [PubMed] [Google Scholar]
5. Laurent T.C., Ogston A.G. The interaction between polysaccharides and other macromolecules. 4. The osmotic pressure of mixtures of serum albumin and hyaluronic acid. Biochem. J. 1963;89:249–253. [PMC free article] [PubMed] [Google Scholar]
6. Minton A.P. Models for excluded volume interaction between an unfolded protein and rigid macromolecular cosolutes: macromolecular crowding and protein stability revisited. Biophys. J. 2005;88:971–985. [PMC free article] [PubMed] [Google Scholar]
7. Zhou H.X. Loops, linkages, rings, catenanes, cages, and crowders: entropy-based strategies for stabilizing proteins. Acc. Chem. Res. 2004;37:123–130. [PubMed] [Google Scholar]
8. Cheung M.S., Klimov D., Thirumalai D. Molecular crowding enhances native state stability and refolding rates of globular proteins. Proc. Natl. Acad. Sci. USA. 2005;102:4753–4758. [PMC free article] [PubMed] [Google Scholar]
9. Homouz D., Perham M., Wittung-Stafshede P. Crowded, cell-like environment induces shape changes in aspherical protein. Proc. Natl. Acad. Sci. USA. 2008;105:11754–11759. [PMC free article] [PubMed] [Google Scholar]
10. Stagg L., Zhang S.Q., Wittung-Stafshede P. Molecular crowding enhances native structure and stability of α/β protein flavodoxin. Proc. Natl. Acad. Sci. USA. 2007;104:18976–18981. [PMC free article] [PubMed] [Google Scholar]
11. Sasahara K., McPhie P., Minton A.P. Effect of dextran on protein stability and conformation attributed to macromolecular crowding. J. Mol. Biol. 2003;326:1227–1237. [PubMed] [Google Scholar]
12. Le Coeur C., Teixeira J., Longeville S. Compression of random coils due to macromolecular crowding: scaling effects. Phys. Rev. E Stat. Nonlin. Soft Matter Phys. 2010;81:061914. [PubMed] [Google Scholar]
13. Johansen D., Jeffries C.M., Goldenberg D.P. Effects of macromolecular crowding on an intrinsically disordered protein characterized by small-angle neutron scattering with contrast matching. Biophys. J. 2011;100:1120–1128. [PMC free article] [PubMed] [Google Scholar]
14. Homouz D., Sanabria H., Cheung M.S. Modulation of calmodulin plasticity by the effect of macromolecular crowding. J. Mol. Biol. 2009;391:933–943. [PMC free article] [PubMed] [Google Scholar]
15. Tokuriki N., Kinjo M., Yomo T. Protein folding by the effects of macromolecular crowding. Protein Sci. 2004;13:125–133. [PMC free article] [PubMed] [Google Scholar]
16. Qu Y., Bolen D.W. Efficacy of macromolecular crowding in forcing proteins to fold. Biophys. Chem. 2002;101-102:155–165. [PubMed] [Google Scholar]
17. Dhar A., Samiotakis A., Cheung M.S. Structure, function, and folding of phosphoglycerate kinase are strongly perturbed by macromolecular crowding. Proc. Natl. Acad. Sci. USA. 2010;107:17586–17591. [PMC free article] [PubMed] [Google Scholar]
18. Engel R., Westphal A.H., van Mierlo C.P. Macromolecular crowding compacts unfolded apoflavodoxin and causes severe aggregation of the off-pathway intermediate during apoflavodoxin folding. J. Biol. Chem. 2008;283:27383–27394. [PubMed] [Google Scholar]
19. Hong J., Gierasch L.M. Macromolecular crowding remodels the energy landscape of a protein by favoring a more compact unfolded state. J. Am. Chem. Soc. 2010;132:10445–10452. [PMC free article] [PubMed] [Google Scholar]
20. Wallgren M., Adén J., Wolf-Watz M. Extreme temperature tolerance of a hyperthermophilic protein coupled to residual structure in the unfolded state. J. Mol. Biol. 2008;379:845–858. [PubMed] [Google Scholar]
21. Lakowicz J. 3rd ed. Springer; Singapore: 2006. Principles of Fluorescence Spectroscopy. [Google Scholar]
22. Greene R.F., Jr., Pace C.N. Urea and guanidine hydrochloride denaturation of ribonuclease, lysozyme, α-chymotrypsin, and β-lactoglobulin. J. Biol. Chem. 1974;249:5388–5393. [PubMed] [Google Scholar]
23. Santoro M.M., Bolen D.W. A test of the linear extrapolation of unfolding free energy changes over an extended denaturant concentration range. Biochemistry. 1992;31:4901–4907. [PubMed] [Google Scholar]
24. Valeur B., Weber G. Resolution of the fluorescence excitation spectrum of indole into the 1La and 1Lb excitation bands. Photochem. Photobiol. 1977;25:441–444. [PubMed] [Google Scholar]
25. Haas E., Katchalsky-Katzir E., Steinberg I.Z. Effect of the orientation on donor and acceptor on the probability of energy transfer involving electronic transitions of mixed polarisation. Biochemistry. 1978;17:5064–5070. [PubMed] [Google Scholar]
26. Privat J.P., Wahl P., Auchet J.C. Rates of deactivation processes of indole derivatives in water-organic solvent mixtures—application to tryptophyl fluorescence of proteins. Biophys. Chem. 1979;9:223–233. [PubMed] [Google Scholar]
27. Olofsson M., Kalinin S., Johansson L.B. Tryptophan-BODIPY: a versatile donor-acceptor pair for probing generic changes of intraprotein distances. Phys. Chem. Chem. Phys. 2006;8:3130–3140. [PubMed] [Google Scholar]
28. Levenberg K. A method for the solution of certain non-linear problems in least squares. Q. Appl. Math. 1944;2:164–168. [Google Scholar]
29. Szabo A. Theory of fluorescence depolarization in macromolecules and membranes. J. Chem. Phys. 1984;81:150–167. [Google Scholar]
30. Jablonski A. Influence of torosional vibrations of luminescent molecules on the fundamental polarisation of photoluminiscence of solutions. Acta Physiol. Pol. 1950;10:33–36. [Google Scholar]
31. Johansson B.-Å. Limiting fluorescence anisotropies of perylene and xanthene derivatives. J. Chem. Soc., Faraday Trans. 1990;86:2103–2107. [Google Scholar]
32. Heyn M.P. Determination of lipid order parameters and rotational correlation times from fluorescence depolarization experiments. FEBS Lett. 1979;108:359–364. [PubMed] [Google Scholar]
33. Silow M., Oliveberg M. High-energy channeling in protein folding. Biochemistry. 1997;36:7633–7637. [PubMed] [Google Scholar]
34. Christiansen A., Wang Q., Wittung-Stafshede P. Factors defining effects of macromolecular crowding on protein stability: an in vitro/in silico case study using cytochrome c. Biochemistry. 2010;49:6519–6530. [PubMed] [Google Scholar]
35. Eftink M.R. Fluorescence techniques for studying protein structure. Methods Biochem. Anal. 1991;35:127–205. [PubMed] [Google Scholar]
36. Auton M., Bolen D.W. Application of the transfer model to understand how naturally occurring osmolytes affect protein stability. Methods Enzymol. 2007;428:397–418. [PubMed] [Google Scholar]
37. Li C., Wang Y., Pielak G.J. Translational and rotational diffusion of a small globular protein under crowded conditions. J. Phys. Chem. B. 2009;113:13390–13392. [PMC free article] [PubMed] [Google Scholar]
38. Ladurner A.G., Fersht A.R. Upper limit of the time scale for diffusion and chain collapse in chymotrypsin inhibitor 2. Nat. Struct. Biol. 1999;6:28–31. [PubMed] [Google Scholar]
39. Neuweiler H., Löllmann M., Sauer M. Dynamics of unfolded polypeptide chains in crowded environment studied by fluorescence correlation spectroscopy. J. Mol. Biol. 2007;365:856–869. [PubMed] [Google Scholar]
40. Lavalette D., Hink M.A., Visser A.J. Proteins as micro viscosimeters: Brownian motion revisited. Eur. Biophys. J. 2006;35:517–522. [PubMed] [Google Scholar]
41. Lavalette D., Tétreau C., Blouquit Y. Microscopic viscosity and rotational diffusion of proteins in a macromolecular environment. Biophys. J. 1999;76:2744–2751. [PMC free article] [PubMed] [Google Scholar]
42. Goins A.B., Sanabria H., Waxham M.N. Macromolecular crowding and size effects on probe microviscosity. Biophys. J. 2008;95:5362–5373. [PMC free article] [PubMed] [Google Scholar]
43. Mukherjee S., Waegele M.M., Gai F. Effect of macromolecular crowding on protein folding dynamics at the secondary structure level. J. Mol. Biol. 2009;393:227–236. [PMC free article] [PubMed] [Google Scholar]
Articles from Biophysical Journal are provided here courtesy of The Biophysical Society