Logo of nihpaAbout Author manuscriptsSubmit a manuscriptHHS Public Access; Author Manuscript; Accepted for publication in peer reviewed journal;
J Mol Biol. Author manuscript; available in PMC 2010 Oct 16.
Published in final edited form as:
PMCID: PMC2754818
NIHMSID: NIHMS142769
PMID: 19682997

Effect of Macromolecular Crowding on Protein Folding Dynamics at the Secondary Structure Level

Smita Mukherjee,≠ Matthias M. Waegele,≠ Pramit Chowdhury, Lin Guo, and Feng Gai*
Author information Copyright and License information Disclaimer

Associated Data

Supplementary Materials

Abstract

Macromolecular crowding is one of the key characteristics of the cellular environment and therefore, is intimately coupled to the process of protein folding in vivo. While previous studies have provided invaluable insight into the effect of crowding on the stability and folding rate of protein tertiary structures, very little is known about how crowding affects protein folding dynamics at the secondary structure level. Herein, we examine the thermal stability and folding-unfolding kinetics of three small folding motifs, i.e., a 34-residue α-helix, a 34-residue cross-linked helix-turn-helix, and a 16-residue β hairpin, in the presence of two commonly used crowding agents, Dextran 70 (200 g/L) and Ficoll 70 (200 g/L). We find that these polymers do not induce any appreciable changes in the folding kinetics of the two helical peptides, which is somewhat surprising as the helix-coil transition kinetics have been shown to depend on viscosity. Also to our surprise and in contrast to what has been observed for larger proteins, we find that crowding leads to an appreciable decrease in the folding rate of the shortest β-hairpin peptide, indicating that besides the excluded volume effect, other factors also need to be considered when evaluating the net effect of crowding on protein folding kinetics. A model considering both the static and dynamic effects arising from the presence of the crowding agent is proposed to rationalize these results.

Keywords: Macromolecular crowding, confinement, protein folding, helix-coil transition, β-hairpin folding

Introduction

Understanding how proteins fold in vivo poses a formidable challenge. Thus, a majority of protein folding studies have been carried out in vitro and under conditions wherein only dilute aqueous solutions were used. While such studies have provided invaluable insights into our understanding of the protein folding problem, these however did not take into account the possible effects arising from macromolecular crowding; an important but often neglected aspect of the intracellular environment.1–4 For instance, the presence of macromolecules near a protein could alter its folding energy landscape simply through the excluded volume effect since folding results in a compaction of the polypeptide chain.5 Indeed, several recent experimental and computational studies have shown that such effects of volume exclusion can significantly alter the stability and folding rate of a protein.6–16

Besides such entropic effects, macromolecular crowding could also affect the dynamics of protein folding through other mechanisms (in the present case the effect of macromolecular crowding is referred to as the net effect of an inert macromolecule on the folding properties of the protein or peptide molecule in question, which includes both static effects, such as that arising from confinement, and dynamic effects, such as modulation of the frictional drag experienced by conformational motions along the reaction coordinate). For example, it is well known that a high mass macromolecular crowding agent (i.e., inert polymers) not only changes the dynamic viscosity of the solution (i.e., macroviscosity), but could also modulate the microviscosity of the protein environment in which the folding reaction occurs,17–20 thus perturbing the underlying chain dynamics, as shown by Sauer and coworkers.21 However, most of the experimental studies carried out thus far on the effects of macromolecular crowding on protein folding have dealt with proteins of fairly large size,3,4,6,7,10,12–15 wherein the excluded volume effect appears to dominate, thus obscuring other subtle but important effects arising from the presence of crowding agents. Herein, we study the folding thermodynamics and kinetics of three relatively small protein motifs in the presence of two commonly used crowding agents, Dextran 70 and Ficoll 70, with the aim of providing new insights into the effect of macromolecular crowding on folding events taking place over a relatively short length scale. These peptides form different types of protein secondary and/or supersecondary structures in solution, specifically, a 34-residue monomeric α-helix (L9:41-74),22 a 34-residue, cross-linked helix-turn-helix (HTH) motif (Z34C-m1, which is the D20A mutant of Z34C),23 and a 16-residue β-hairpin (trpzip4-m1).24 Given the critical importance of protein secondary structure formation in several protein folding models, such as the framework model,25 this study is expected to also have considerable implications for the applicability of those models in describing in vivo protein folding.

Dextran 70 is a flexible, linear (<5% branching) polymer of D-glucopyranose, which behaves as a quasi-random coil (Rh ~63 Å),20,26–29 whereas Ficoll 70 is a compact and highly cross-linked and branched co-polymer of sucrose and epichlorohydrin, which can be approximated as a semi-rigid sphere (Rh ~55 Å).27–29 Thus, comparative studies employing these two polymers allow one to examine how the nature and geometric shape of the respective crowding agent affects the folding dynamics of the protein system in question. Interestingly, only the thermodynamic stability of the shortest peptide studied here (i.e., trpzip4-m1) shows an appreciable change when its environment is crowded by Ficoll 70. Similar to that observed for large proteins, macromolecular crowding leads to an increase in the thermal stability of trpzip4-m1 in the presence of 200 g/L Ficoll 70, a concentration that falls within the concentration range that has been used in previous crowding studies.1,2,6,7,10,11,13 However, in contrast to the common notion that macromolecular crowding increases the rate of protein folding, our results show that the folding rate of trpzip4-m1 in fact decreases in the presence of either Ficoll 70 or Dextran 70. Taken together, these results indicate that besides the commonly encountered excluded volume effect, other factors need to be considered when assessing the effect of macromolecular crowding on protein folding.

Results

All crowding experiments were carried out in 20 mM phosphate buffer in D2O at pH* = 7 in the presence of either 200 g/L Dextran 70 or 200 g/L Ficoll 70, whereas all other experiments were carried out in 20 mM phosphate buffer in D2O at pH* = 7.

L9:41-74

The far-UV CD spectra of L9:41-74 in both Dextran 70 and Ficoll 70 solutions at 4 °C exhibit the characteristic double minima of α-helices at 208 and 222 nm, respectively, and also overlap with that of L9:41-74 in D2O buffer solution (Figure S1 in Supporting Information). In addition, as shown (Figure 1A), the presence of these crowding agents has little effect on the CD thermal denaturation profile of L9:41-74, indicating that the thermodynamics of the underlying helix-coil transition is not sensitive to the environmental changes induced by these crowding agents.

An external file that holds a picture, illustration, etc. Object name is nihms142769f1.jpg

(A) CD thermal melting curves of L9:41-74 in 200 g/L Dextran 70 (red) and in 200 g/L Ficoll 70 (blue), respectively. (B) Arrhenius plot of the T-jump induced conformational relaxation rates of L9:41-74 in 200 g/L Dextran 70 (red) and in 200 g/L Ficoll 70 (blue), respectively. Also shown for comparison are the CD thermal melting curve (black crosses in A) and relaxation rates (black dashed line in B) of the same peptide in 20 mM phosphate D2O buffer (derived from reference 22).

The relaxation kinetics of L9:41-74 in these polymer solutions were studied by the laser-induced temperature-jump (T-jump) infrared method, the details of which have been described elsewhere.30 Similar to those observed in dilute solution,22 the T-jump induced relaxation kinetics of L9:41-74 contain two distinct phases (Figure S2 in Supporting Information). The fast phase cannot be resolved by our setup and has been attributed to temperature-induced spectral shift of the amide-I′ band.31 On the other hand, the slow phase, which arises from the conformational redistribution process of the peptide in response to the T-jump, is well resolved and can be described by a single-exponential function. As shown (Figure 1B), within the temperature range studied, the T-jump induced conformational relaxation rates of L9:41-74 in both Dextran 70 and Ficoll 70 solutions are almost identical to that of the peptide in D2O buffer solution. For example, at 20 °C the relaxation time constants of L9:41-74 in Dextran 70 and Ficoll 70 solutions are 1.4 ± 0.2 μs and 1.5 ± 0.2 μs, respectively, whereas in D2O buffer solution the relaxation time constant is 1.17 ± 0.15 μs.22 Taken together, these thermodynamic and kinetic results indicate that the folding-unfolding transition of this α-helical peptide is not sensitive to macromolecular crowding, at least not to that induced by 200 g/L Dextran 70 or Ficoll 70.

Z34C-m1

The far-UV CD spectra of Z34C-m1 obtained at 4 °C in both Dextran 70 and Ficoll 70 solutions are almost completely superimposable with that obtained in D2O buffer solution (Figure S3 in Supporting Information), indicating that the helical nature of the folded conformation is not affected by the crowding agents. The thermal denaturation of Z34C-m1 in both Dextran 70 and Ficoll 70 solutions, measured by monitoring the change in its helical CD signal at 222 nm with increasing temperature (Figure 2A), also shows that these crowding agents do not change the thermal stability of the peptide to any appreciable extent. Indeed, globally fitting the CD data to a two-state model reveal that the thermal melting temperature (Tm) of Z34C-m1 in these crowded environments is only 2–3 °C higher than that23 in D2O buffer (Table 1). Consistent with these thermodynamic assessments, the T-jump induced relaxation rates of Z34C-m1 in these polymer solutions also show only a moderate decrease from those measured in dilute aqueous solution (Table 1 and Figure 2B).

An external file that holds a picture, illustration, etc. Object name is nihms142769f2.jpg

(A) CD thermal melting curves of Z34C-m1 in 200 g/L Dextran 70 (red) and in 200 g/L Ficoll 70 (blue), respectively. Lines are global fits of these data to the two-state model described in the text. (B) Arrhenius plot of the T-jump induced conformational relaxation rates (open circles) of Z34C-m1 in 200 g/L Dextran 70 (red) and in 200 g/L Ficoll 70 (blue), respectively. Open triangles and squares correspond to the two-state folding and unfolding rates of Z34C-m1 in 200 g/L Dextran 70. Also shown are the CD thermal melting curve (black crosses in A) and folding (black dashed line in B) and unfolding (black solid line in B) rate constants of Z34C-m1 in 20 mM phosphate D2O buffer (derived from reference 23).

Table 1

Summary of the thermodynamic and kinetic data for Z34C-m1 and trpzip4-m1 obtained under different conditions. The folding and unfolding times were obtained at 55 °C for Z34C-m1 and 35 °C for trpzip4-m1.

Z34C-m1Trpzip4-m1
D2O bufferDextran 70Ficoll 70D2O bufferDextran 70Ficoll 70
ΔHm (kcal mol−1)16.2 ± 0.717.1 ± 0.816.8 ± 0.713.8 ± 0.212.5 ± 0.315.6 ± 0.2
ΔSm (cal K−1 mol−1)49.4 ± 3.852.0 ± 2.650.9 ± 2.045.1 ± 2.540.6 ± 0.949.1 ± 0.7
ΔCp (cal K−1 mol−1)118 ± 34245 ± 105120 ± 54343 ± 41238 ± 19223 ± 15
Tm (°C)54.7 ± 1.055.3 ± 1.557.7 ± 0.532.1 ± 0.934.3 ± 0.544.0 ± 0.2
τf (μs)*5.3 ± 0.88.6 ± 1.37.3 ± 1.147.5 ± 2.396 ± 14122 ± 18
τu (μs)*5.1 ± 0.88.8 ± 1.39.0 ± 1.438.1 ± 2.092 ± 14240 ± 36

Trpzip4-m1

Similar to that obtained in D2O buffer solution,24 the far-UV CD spectra of trpzip4-m1 in both Dextran 70 and Ficoll 70 solutions exhibit a positive band centered at ~229 nm (Figure S4 in Supporting Information), arising from excitonic coupling between the paired tryptophan sidechains. As shown (Figure 3), the thermal unfolding transitions of trpzip4-m1 in these solutions, as monitored by the change in the CD signal at 229 nm, show characteristics of a cooperative thermal unfolding process. As indicated (Table 1), the thermal melting temperature (Tm) of trpzip4-m1 in Dextran 70 solution is increased only slightly compared to that (32 °C) in D2O buffer, whereas in Ficoll 70 solution the Tm of this peptide shows a substantial increase to ~44 °C (Table 1).

An external file that holds a picture, illustration, etc. Object name is nihms142769f3.jpg

CD thermal melting curves of trpzip4-m1 in 200 g/L Dextran 70 (red) and in 200 g/L Ficoll 70 (blue), respectively. Lines are global fits of these data to the two-state model described in the text. Also shown is the CD thermal melting curve (black) of the same peptide in 20 mM phosphate D2O buffer (derived from reference 24).

The relaxation kinetics of trpzip4-m1 in response to a T-jump were also probed at 1631 cm−1 where antiparallel β-sheets are known to absorb (Figure S5 in Supporting Information).24 As shown (Figure 4A and 4B), unlike L9:41-74, the folding and unfolding rates of trpzip4-m1 are distinctly slower in the presence of the crowding agents. For example, at 35 °C the folding and unfolding time constants of trpzip4-m1 in Dextran 70 solution are 96 ± 14 and 92 ± 14 μs, respectively, whereas in dilute aqueous solution24 this peptide folds in 47.5 ± 2.3 μs and unfolds in 38.1 ± 2.0 μs (Table 1).

An external file that holds a picture, illustration, etc. Object name is nihms142769f4.jpg

Arrhenius plots of the T-jump induced conformational relaxation rates (open circles) of trpzip4-m1 in 200 g/L Dextran 70 (A) and in 200 g/L Ficoll 70 (B), respectively. Open triangles and squares correspond to their respective two-state folding and unfolding rates. Also shown in each case are the folding (blue dashed line) and unfolding (red dashed line) rate constants of trpzip4-m1 in 20 mM phosphate D2O buffer (derived from reference 24).

Discussion

Since macromolecular crowding is an intrinsic feature of the cellular environment,32,33 there has been considerable interest in recent years in investigating its effect on protein folding, both experimentally3,4,6,7,10,12–15 and computationally.3,4,8,9,11,13,16 However, almost all of the previous experimental efforts in this area have been focused on large proteins,3,4,6,7,10,13–15 thus providing little, if any, information on the effect of macromolecular crowding on protein folding at the secondary structure level. While such a bias in focus is understandable because most protein secondary structural elements (e.g., monomeric α-helix and β-hairpin) are probably too small to be profoundly affected by the excluded volume effect,3,5 however, understanding the influence of crowding on the folding dynamics of such small structural moieties could provide new insights into the otherwise complex interplay between different crowding effects. In addition, in view of the important role of secondary structure formation in existing protein folding models such as the framework model,25 there is a strong need for investigation of folding dynamics of secondary structural elements in crowded environment in order to understand protein folding in vivo. Herein, we study the folding thermodynamics and kinetics of three distinct secondary structural elements, i.e., a monomeric α-helix (L9:41-74), a β-hairpin (trpzip4-m1) and a HTH motif (Z34C-m1) in the presence of two commonly used crowding agents, namely, Dextran 70 and Ficoll 70.

L9:41-74

L9:41-74 is the central α-helix of ribosomal protein L9 from the bacterium Bacillus stearothermophilus. Owing to a series of favorable sidechain-sidechain interactions, mostly electrostatic in nature, this peptide remains folded even in isolation in aqueous solution.34 Previously, we have shown that this peptide folds on a timescale that is significantly slower than that of alanine-based peptides and its folding time may be a more realistic representation of the time scale in which α-helices in proteins fold.22 Thus, it would be quite interesting to further examine how crowding affects its folding-unfolding dynamics. Our CD measurements (Figure 1A) show that the thermal stability of L9:41-74 in either Dextran 70 or Ficoll 70 solution remains practically unchanged from that in dilute aqueous solution, suggesting that the helix-coil transition experiences little, if any, crowding effect. This observation is however unexpected because the chain length of this α-helix is estimated to be ~50 Å (assuming a full helical structure), which is comparable to the hydrodynamic radii of Dextran 70 and Ficoll 70.

Besides the apparent excluded volume effect, addition of a polymer to aqueous solution is also known to affect the dynamic viscosity of the solution. For example, the dynamic viscosities of 200 g/L Dextran 70 and 200 g/L Ficoll 70 solutions are about 18 and 10 times greater than that of water at 20 °C,27,35 respectively. Previously, Eaton and coworkers36 have shown that the T-jump induced relaxation rate of an alanine based helical peptide is inversely proportional to η0.6 (where η represents the viscosity of the solution) when small viscogens, such as glucose and sucrose, were used to increase the viscosity of the solution. Thus, our observation that the T-jump induced relaxation kinetics of L9:41-74 in both Dextran 70 and Ficoll 70 solutions are almost the same as those obtained in dilute D2O solution (Figure 1B) is quite interesting and warrants further discussion.

The viscosity dependence of protein folding rates stems from the diffusive nature of the associated barrier-crossing events,37–40 namely, from the requirement of polypeptide chain motion to form the stabilizing native contacts via diffusion. Previous studies have shown that under the current crowding conditions the polymer molecules can form porous networks17,19,21 wherein the viscosity determined via the diffusion of a probe molecule can be significantly lower than the dynamic viscosity of the bulk solution.17,21,41 To differentiate these viscosities, we referred to the one measured via molecular diffusion as the microviscosity of the solution. Since the sizes of the three peptides used in present study are relatively small, it is reasonable to assume that their folding rates would respond more to the microviscosity rather than the dynamic viscosity of the polymer solution. Thus, to further probe the microviscosity of the current crowding solutions, we have also measured the characteristic translational diffusion times (τD, a quantity that is inversely proportional to the diffusion constant D) of three molecular systems of varying sizes using Fluorescence Correlation Spectroscopy (FCS), namely, (i) a fluorescent dye: Rhodamine 6G (R6G), (ii) a 37-residue fluorescently labeled pH (low) insertion peptide (pHLIP), and (iii) the Human Serum Albumin (HSA) protein complexed with the dye Nile Red (HSA-NR). As shown (Table 2), the translational diffusion times of all three probes indicate that the microviscosities of Ficoll 70 and Dextran 70 solutions are about 4–6 times larger than that of water, depending on the size of the probe and also the crowding agent.

Table 2

Translational diffusion times determined by FCS at room temperature for Rhodamine 6G, the pHLIP peptide, and HSA-NR complex.

D2ODextran 70 (200 g/L)Ficoll 70 (200 g/L)
Rhodamine 6G (μs)53 ± 2216 ± 5256 ± 5
pHLIP (μs)75 ± 3286 ± 5305 ± 8
HSA-NR (μs)419 ± 152000 ± 302460 ± 60

These FCS measurements indicate that not only the dynamic (or bulk) viscosity but even the microviscosity measurements fall short of providing a rationale behind the observation that the relaxation rate of L9:41-74 is insensitive to the presence of crowding agents, this being in stark contrast to the well-documented notion that the rate of α-helix formation is susceptible to viscosity.36,42 However, this apparent discrepancy may be reconciled by taking into account the notion that diffusion constant could also be distance-dependent.43,44 For example, the diffusion constant (D) of a probe molecule in a polymer solution (Figure S6 in Supporting Information) consisting of porous networks45 is:

D = D0exp(−ac1/2)
(1)

where c is the concentration of the polymer and D0 is the diffusion constant of the probe at c = 0. The constant a has been shown to characterize the length scale (ξ) of the pores embedded in the porous networks, at least for a simple pore model.44 Thus, for short diffusion modes or diffusion events during which the diffusing particles travels a distance smaller than the pore size ξ of the polymer solution, the particles rarely come into contact with the polymer chains and hence their diffusion constant is close to that found in bulk water. On the other hand, for long diffusion modes or diffusion events during which the particles travel a distance much larger than ξ, the resultant diffusion constant becomes smaller than that measured in bulk water, similar to those observed in our FCS measurements. In this context, it is clear that the viscosity effect exerted by a polymeric crowding agent on the folding rate of a peptide depends on the length scale that the polypeptide chain has to traverse from the unfolded to the transitions state along the folding coordinate.

The formation of an α-helix can be regarded as a series of local events wherein a hydrogen bond between the amino acids at ‘i’ and ‘i+4’ positions along the polypeptide chain is formed. Thus, the length-scale (a few Å´) over which the key α-helix folding events take place is far less than the pore size of the polymer solutions used in the current study and therefore the corresponding folding rate does not show any significant deviation from that in dilute aqueous solution. In contrast, in studies where small viscogens (e.g., glucose or sucrose) are used to increase the viscosity, the resulting solution viscosity is microscopically homogeneous and therefore the effect of increased solvent friction can be experienced down to a very short length-scale. Thus, in such cases, the viscosity dependence of the folding kinetics of protein secondary structural elements might become more pronounced, as has been shown in a previous study.36

Z34C-m1

34C-m1 is a mutant of Z34C (i.e., D20A), which forms a crossed-linked HTH structure, a common structural motif found in DNA-binding proteins.46 Previously we have shown that Z34C-m1 folds significantly slower than Z34C since the mutation deletes a hydrogen bond that is critical to the stability of the reverse turn.23 Z34C-m1 constitutes a good model system to further examine the effect of microviscosity on the folding-unfolding kinetics of α-helices in a protein context, as the disulfide cross-linker prevents the protein to become extended upon unfolding and hence makes it less likely to experience the excluded volume effect. As expected, the folding thermodynamics and kinetics of Z34C-m1 are only moderately affected by the crowding agents employed herein (Figure 2A and 2B and Table 1). Thus, these results are consistent with those of L9:41-74 and further corroborate the notion that macromolecular crowding does not affect to any significant extent the folding dynamics of α-helices that can fold independently. However, in a protein context the frictional drag along the folding coordinate of an α-helix may be position-dependent and thus, could exert a more complex effect on the folding dynamics. Apparently, for β-sheets whose folding requires relatively large-scale chain diffusions, the effect of microviscosity is expected to play a more important role as discussed below.

Trpzip4-m1

To further explore to what extent the nature of the folded topology determines the effect of macromolecular crowding, we have also studied the folding thermodynamics and kinetics of trpzip4-m1, the D46A mutant of the 16-residue β-hairpin trpzip424 in both Dextran 70 and Ficoll 70 solutions. While the thermal denaturation properties of trpzip4-m1 is almost unaffected by addition of Dextran 70, its thermal melting temperature in Ficoll 70 solution however shows a substantial increase from that in D2O buffer.24 This difference manifests the disparity in terms of crowding efficiency of these crowding agents and can be rationalized based on the overall difference in the molecular structures of these polymers. Previous studies have shown that the persistence length of Dextran is about 0.4 nm,47 which renders it to be a relatively flexible polymer. On the other hand, Ficoll 70 is comparatively more rigid due to its cross-linked structure.29,48 Thus, based on these considerations, it is conceivable that at the same concentration Dextran 70 can accommodate more interstitial spaces in comparison to Ficoll 70, therefore allowing the unfolded polypeptide chain of trpzip4-m1 more space to escape and hence leading to a weaker excluded volume effect.

In contrast to that observed in other studies,6,9,11 macromolecular crowding results in a decrease in the folding rate of trpzip4-m1. This result thus clearly indicates that a macromolecular crowding agent can affect not only the protein stability via the excluded volume effect, but also the folding dynamics through modulation of the effective viscosity of the protein environment. In the case of Dextran 70, the thermal stability of the β-hairpin exhibits only a very modest change (Table 1), thus implying that the observed reduction in the folding and unfolding rates is unlikely to result from the excluded volume effect, but rather arises from a viscosity effect. This interpretation is consistent with the notion that the dynamics of peptide loop closure and β-hairpin folding require relatively large-scale and non-local motions, in contrast to the folding of an α-helix wherein the chain primarily undergoes a series of local structural reorganizations. Thus, during such a folding event the peptide chain has a higher chance of colliding with the polymer network formed by the crowding molecules, resulting in a slower folding/unfolding rate. This picture is consistent with a recent FCS study by Sauer and coworkers,21 who showed that the intrachain diffusion rate of a series of fluorescently labeled short poly(GS)-peptides in Ficoll 70 depends on the net effect of two opposing forces, the excluded volume effect and the increased viscous drag of the polymer solution.

The effect of viscosity on protein folding or unfolding rate constant (kobs) can often be described by the hydrodynamic approximation of the Kramers equation in the high friction limit, i.e.,

kobs=Aη(T)αexp(−ΔG≠/RT)
(2)

where η(T) is the effective viscosity of the medium at temperature T and has been assumed here to be independent of the spatial location of the peptide, α is equal to unity in many cases, A is a constant, and ΔG≠ is the apparent free energy barrier at temperature T, and R is the gas constant. Thus, a crowding agent can modulate the folding and unfolding rate constants of a protein by changing either η or ΔG≠ or both. Apparently, the effect related to η is a dynamic one, whereas that related to ΔG≠ is static (or the excluded volume) in nature.

For a simple two-state scenario, assuming that A is the same for both folding and unfolding and also that it is independent of crowding, one can easily show that (irrespective of the position of the transition state)

ln(kfckf0)+ln(kucku0)=2αln(η0(T)ηc(T))+1RT(ΔGuc−0+ΔGfc−0)
(3)

kf and ku are the folding and unfolding rate constants, ΔGuc−0=Guc−Gu0 and ΔGfc−0=Gfc−Gf0 where Gu and Gf represent the free energies of the unfolded and folded states, respectively, and in all cases, the superscripts “0” and “c” represent the dilute and crowded solutions, respectively. In addition, η0(T) and ηc(T) are the effective viscosities of the dilute and crowded solutions at temperature T, respectively. For Dextran 70, the increased stabilization of the β-hairpin conformation of trpzip4-m1 arising from crowding is considered to be negligible (i.e., ΔGuc−0+ΔGfc−0≈0). Thus, using the measured folding and unfolding rate constants of trpzip4-m1 in the presence and absence of crowding agents, in conjunction with the assumption that α is approximately unity for β-hairpin folding,36 the value of ηc/η0 is calculated to be 3.5 ± 0.5 at 20 °C, which is comparable to the ratio of diffusion times τc/τ0 = 3.8 of the pHLIP peptide in 200 g/L Dextran 70 and water at room temperature (Table 2). Therefore, this agreement further corroborates the idea that Dextran 70 affects trpzip4-m1 folding via mostly the dynamic (or viscosity) effect.

In the case of Ficoll 70, the presence of the crowding agent leads to not only a decrease in the folding rate of trpzip4-m1, but also a significant increase in the stability of β-hairpin conformation (Figure 3 and ​and4B4B and Table 1), indicating that the folding and unfolding dynamics of trpzip4-m1 experience both the excluded volume and viscosity effects. As a result, the unfolding rate is decreased to a larger extent than the folding rate, compared to their respective values in D2O buffer solution (Table 1). This is because both the excluded volume and viscosity effects slow down the rate of unfolding, whereas they exhibit opposite effects on the folding rate.

Furthermore, it is easy to show that

kfckf0=(η0ηc)αexp(ΔΔG≠RT)
(4)

where ΔΔG≠=ΔG0≠−ΔGc≠ which is a quantitative measure of the excluded volume effect on the folding dynamics. Thus, it is possible to dissect the static and dynamic contributions to the total crowding effect if ηc/η0 is known. While ηc/η0 cannot be determined independently from the current study, the result obtained with Dextran 70 nevertheless suggests that it approaches the ratio τc/τ0 measured by the FCS method. Thus, using the ratio τc/τ0 = 4.1 (Table 2) we estimated the ΔΔG≠ value to be ~0.3 ± 0.1 kcal mol−1 at 35 °C for the folding of trpzip4-m1 in 200 g/L Ficoll 70 solution. As expected, the decrease in the folding free energy barrier of trpzip4-m1 due to the excluded volume effect is very modest, as the additional stabilization (i.e., ΔΔG0=∣ΔGc0∣−∣ΔG00∣) of trpzip4-m1 due to the presence of 200g/L Ficoll 70 at 35 °C is only about 0.5 ± 0.1 kcal mol−1, owing to the small size of the peptide relative to the size of the crowding agent, which is in agreement with the theoretical prediction by Minton.5 However, we expect that the approach outlined above will be very useful for dissecting the static and dynamic effects of a crowding agent for other protein systems.

Conclusions

Employing two commonly used crowding agents, Dextran 70 (200 g/L) and Ficoll 70 (200 g/L), we have studied how macromolecular crowding affects the folding-unfolding kinetics of three peptides, which fold into different conformations in solution, namely an α-helix (L9:41-74), a cross-linked helix-turn-helix (Z34C-m1), and a β-hairpin (trpzip4-m1). Interestingly, only the folding thermodynamics and kinetics of the shortest peptide, trpzip4-m1, were found to show significant changes in the Ficoll 70 solution, thus indicating that the effect of macromolecular crowding is sensitive to the size and shape of both the peptide and the crowding agent. In addition, the observation that macromolecular crowding does not significantly alter the folding dynamics of the two helical peptides is consistent with the notion that α-helix formation involves mostly local interactions. Furthermore, and in contrast to that observed for large proteins, macromolecular crowding results in a decrease in the folding rate of trpzip4-m1, indicating that the motion of the peptide chain experiences a higher friction arising from the presence of the crowding agent, and that this viscosity effect in this case outplays the excluded volume effect on the folding kinetics of the β-hairpin.

MATERIALS AND METHODS

Peptide synthesis and purification

All peptides used in this study were synthesized based on standard Fmoc-protocols on a PS3 automated peptide synthesizer (Protein Technologies, MA) and purified by reverse-phase chromatography. The peptide sequences are: Ac-41 PANLKALEAQ-51 KQKEQRQAAE-61ELANAKKLKE-71QLEK-NH2 (L9:41–74), 6 FNMQCQRRFY-16 EALHAPNLNE-26 EQRNAKIKSI-36 RDDC-NH2 (Z34C-m1), and GEWTWADATKTWTWTE-NH2 (trpzip4-m1). TMR-maleimide (Molecular Probes, CA), a thiol reactive dye, was used to label the cysteine variant of pHLIP peptide49 (sequence: ACEQNPIYWA-RYADWLFTTP-LLLLDLALLV-DADEGTG) following the protocol available in the Molecular Probes Handbook. Oxidation of the Z34C mutant was done as previously described.23 The identity of the peptide sample was further verified by matrix assisted laser desorption ionization mass spectroscopy. Multiple rounds of lyophilization against a 0.1 M DCl/D2O solution were used to remove the exchangeable hydrogen atoms and also the residual trifluoroacetic acid from the peptide synthesis. Peptide solutions used in both CD and IR measurements were prepared by directly dissolving lyophilized peptide solids in 20 mM phosphate buffer (pH* = 7) with or without the presence of the crowding agent. The final peptide concentration for L9:41-74 was determined by comparing the published CD data34 and the concentrations for other peptides were determined as previously described23,24 and were about 16–134 μM and 0.8–1.5 mM for CD and IR samples, respectively.

Crowding agents

Dextran 70 and Ficoll 70 were purchased from Fisher Scientific, PA and GE Healthcare, NJ, respectively, and were used as received.

CD spectroscopy

All CD data were collected on an Aviv 62A DS circular dichroism spectrometer (Aviv Associate, NJ). The folding-unfolding thermodynamics of each peptide were obtained by fitting its CD thermal melting transition obtained at either 222 nm (for the helical peptides) or 229 nm (for the β-hairpin) to a two-state model described in detail previously.50

Infrared T-jump setup

The time-resolved T-jump infrared apparatus used in the current study has been described in detail elsewhere.30 The relaxation kinetics of all peptides in response to a T-jump were probed at 1631 cm−1. The observed relaxation rate constants (kR) were further separated into folding (kf) and unfolding (ku) rate constants based on the thermodynamic results obtained from CD studies using the following equations:

kR = kf + ku
(5)

Keq = kf/ku
(6)

Fluorescence correlation spectroscopy

The detail of the FCS setup has been described elsewhere.51 For each measurement, either a 1 nM R6G or 1 nM TMR-labeled pHLIP peptide solution or a solution of 40 nM Nile Red and 20 μM HSA was loaded on a PEG-silane (Gelest Inc., Morrisville, PA) modified glass slip. Excitation of the dye was accomplished by the 514 nm line of an Ar+ ion laser (~100 μW before entering the microscope), and the resultant fluorescence was equally split by a nonpolarizing beamsplitter (Newport, CA) and detected by two avalanche photodiodes (Perkin Elmer, NJ) using an integration time of 0.1 μs. Correlating the fluorescence signals in the cross-correlation mode was accomplished by a Flex 03-LQ-01 correlator card (Correlator.com, NJ) for a duration of 120 s and the resulting autocorrelation traces were analyzed using the following equation,52

G(τ)=(∑i=1n1N(fi1+ττDi)(11+τω2τDi)1/2)×(1−T+T·e−ττT)
(7)

where τDi represents the characteristic diffusion time constant of species i, ω refers to the axial to lateral dimension ratio of the confocal volume element, N represents the number of fluorescent molecules in the confocal volume, fi represents the fraction of species i, τT is the triplet lifetime of the fluorophore and T represents the corresponding triplet amplitude.

Supplementary Material

Acknowledgments

We gratefully acknowledge financial support from the National Institutes of Health (GM-065978 and RR-01348).

Abbreviation used

CD
circular dichroism
FCS
fluorescence correlation spectroscopy
T-jump
temperature-jump

Footnotes

Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

References

1. Ellis RJ. Macromolecular crowding: obvious but underappreciated. Trends Biochem Sci. 2001;26:597–604. [PubMed] [Google Scholar]
2. Minton AP. The influence of macromolecular crowding and macromolecular confinement on biochemical reactions in physiological media. J Mol Biol. 2001;276:10577–10580. [PubMed] [Google Scholar]
3. Zhou HX, Rivas G, Minton AP. Macromolecular crowding and confinement: biochemical, biophysical, and potential physiological consequences. Annu Rev Biophys. 2008;37:375–397. [PMC free article] [PubMed] [Google Scholar]
4. Zhou HX. Protein folding in confined and crowded environments. Arch Biochem Biophys. 2008;469:76–82. [PMC free article] [PubMed] [Google Scholar]
5. Minton AP. 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]
6. van den Berg B, Wain R, Dobson CM, Ellis RJ. Macromolecular crowding perturbs protein refolding kinetics: implications for folding inside the cell. EMBO J. 2000;18:6927–6933. [PMC free article] [PubMed] [Google Scholar]
7. Qu YX, Bolen DW. Efficacy of macromolecular crowding in forcing proteins to fold. Biophys Chem. 2002;101:155–165. [PubMed] [Google Scholar]
8. Friedel M, Sheeler DJ, Shea JE. Effects of confinement and crowding on the thermodynamics and kinetics of folding of a minimalist beta-barrel protein. J Chem Phys. 2003;118:8106–8113. [Google Scholar]
9. Ping G, Yuan JM, Sun ZF, Wei Y. Studies of effects of macromolecular crowding and confinement on protein folding and protein stability. J Mol Recognit. 2004;17:433–440. [PubMed] [Google Scholar]
10. Tokuriki N, Kinjo M, Negi S, Hoshino M, Goto Y, Urabe I, Yomo T. Protein folding by the effects of macromolecular crowding. Protein Sci. 2004;13:125–133. [PMC free article] [PubMed] [Google Scholar]
11. Cheung MS, 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]
12. Ai XJ, Zhou Z, Bai YW, Choy WY. N-15 NMR spin relaxation dispersion study of the molecular crowding effects on protein folding under native condition. J Am Chem Soc. 2006;128:3916–3917. [PubMed] [Google Scholar]
13. Yuan JM, Chyan CL, Zhou HX, Chung TY, Peng HB, Ping GH, Yang GL. The effects of macromolecular crowding on the mechanical stability of protein molecules. Protein Sci. 2008;17:2156–2166. [PMC free article] [PubMed] [Google Scholar]
14. Charlton LM, Barnes CO, Li CG, Orans J, Young GB, Pielak GJ. Residue-level interrogation of macromolecular crowding effects on protein stability. J Am Chem Soc. 2008;130:6826–6830. [PMC free article] [PubMed] [Google Scholar]
15. Homouz D, Stagg L, Wittung-Stafshede P, Cheung MS. Macromolecular crowding modulates folding mechanism of alpha/beta protein apoflavodoxin. Biophys J. 2009;96:671–680. [PMC free article] [PubMed] [Google Scholar]
16. Rivera E, Straub J, Thirumalai D. Sequence and crowding effects in the aggregation of a 10-residue fragment derived from islet amyloid polypeptide. Biophys J. 2009;96:4552–4560. [PMC free article] [PubMed] [Google Scholar]
17. Lavalette D, Tetreau C, Tourbez M, 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]
18. Kozer N, Schreiber G. Effect of crowding on protein - protein association rates: fundamental differences between low and high mass crowding agents. J Mol Biol. 2004;336:763–774. [PubMed] [Google Scholar]
19. Kozer N, Kuttner YY, Haran G, Schreiber G. Protein-protein association in polymer solutions: from dilute to semidilute to concentrated. Biophys J. 2007;92:2139–2149. [PMC free article] [PubMed] [Google Scholar]
20. Goins AB, Sanabria H, Waxham MN. Macromolecular crowding and size effects on probe microviscosity. Biophys J. 2008;95:5362–5373. [PMC free article] [PubMed] [Google Scholar]
21. Neuweiler H, Löllmann M, Doose S, 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]
22. Mukherjee S, Chowdhury P, Bunagan MR, Gai F. Folding kinetics of a naturally occurring helical peptide: implication of the folding speed limit of helical proteins. J Phys Chem B. 2008;112:9146–9150. [PubMed] [Google Scholar]
23. Du D, Gai F. Understanding the folding mechanism of an alpha-helical hairpin. Biochemistry. 2006;45:13131–13139. [PubMed] [Google Scholar]
24. Du D, Tucker MJ, Gai F. Understanding the mechanism of beta-hairpin folding via phi-value analysis. Biochemistry. 2006;45:2668–2678. [PubMed] [Google Scholar]
25. Kim PS, Baldwin RL. Specific intermediates in the folding reactions of small proteins and the mechanism of protein folding. Annu Rev Biochem. 1982;51:459–489. [PubMed] [Google Scholar]
26. Ruhlmann C, Thieme M, Helmstedt M. Interaction between dextran and human low density lipoproteins (LDL) observed using laser light scattering. Chem Phys Lipids. 2001;110:173–181. [PubMed] [Google Scholar]
27. Wenner JR, Bloomfield VA. Crowding effects on EcoRV kinetics and binding. Biophys J. 1999;77:3234–3241. [PMC free article] [PubMed] [Google Scholar]
28. Luby-Phelps K, Castle PE, Taylor DL, Lanni F. Hindered diffusion of inert tracer particles in the cytoplasm of mouse 3T3 cells. Proc Natl Acad Sci USA. 1987;84:4910–4913. [PMC free article] [PubMed] [Google Scholar]
29. Venturoli D, Rippe B. Ficoll and dextran vs. globular proteins as probes for testing glomerular permselectivity: Effects of molecular size, shape, charge, and deformability. Am J Physiol. 2005;288:F605–F613. [PubMed] [Google Scholar]
30. Huang CY, Getahun Z, Zhu Y, Klemke JW, DeGrado WF, Gai F. Helix formation via conformation diffusion search. Proc Natl Acad Sci USA. 2002;99:2788–2793. [PMC free article] [PubMed] [Google Scholar]
31. Mukherjee S, Chowdhury P, Gai F. Infrared study of the effect of hydration on the amide I band and aggregation properties of helical peptides. J Phys Chem B. 2007;111:4596–4602. [PubMed] [Google Scholar]
32. Fulton AB. How crowded is the cytoplasm. Cell. 1982;30:345–347. [PubMed] [Google Scholar]
33. Zimmerman SB, Trach SO. Estimation of macromolecule concentrations and excluded volume effects for the cytoplasm of Escherichia-coli. J Mol Biol. 1991;222:599–620. [PubMed] [Google Scholar]
34. Kuhlman B, Yang HY, Boice JA, Fairman R, Raleigh DP. An exceptionally stable helix from the ribosomal protein L9: implications for protein folding and stability. J Mol Biol. 1997;270:640–647. [PubMed] [Google Scholar]
35. Phillies GDJ, Quinlan CA. Glass temperature effects on probe diffusion in dextran solutions. Macromolecules. 1992;25:3110–3116. [Google Scholar]
36. Jas GS, Eaton WA, Hofrichter J. Effect of viscosity on the kinetics of alpha-helix and beta-hairpin formation. J Phys Chem B. 2001;105:261–272. [Google Scholar]
37. Jacob M, Schmid FX. Protein folding as a diffusional process. Biochemistry. 1999;38:13773–13779. [PubMed] [Google Scholar]
38. Bhattacharyya RP, Sosnick TR. Viscosity dependence of the folding kinetics of a dimeric and monomeric coiled coil. Biochemistry. 1999;38:2601–2609. [PubMed] [Google Scholar]
39. Hagen SJ, Qiu LL, Pabit SA. Diffusional limits to the speed of protein folding: fact or friction? J Phys Condens Matter. 2005;17:S1503–S1514. [Google Scholar]
40. Frauenfelder H, Fenimore PW, Chen G, McMahon BH. Protein folding is slaved to solvent motions. Proc Natl Acad Sci USA. 2006;103:15469–15472. [PMC free article] [PubMed] [Google Scholar]
41. Barshtein G, Almagor A, Yedgar S, Gavish B. Inhomogeneity of viscous aqueous-solutions. Phys Rev E. 1995;52:555–557. [PubMed] [Google Scholar]
42. Klimov DK, Thirumalai D. Viscosity dependence of the folding rates of proteins. Phys Rev Lett. 1997;79:317–320. [Google Scholar]
43. Masuda A, Ushida K, Koshino H, Yamashita K, Kluge T. Novel distance dependence of diffusion constants in hyaluronan aqueous solution resulting from its characteristic nano-microstructure. J Am Chem Soc. 2001;123:11468–11471. [PubMed] [Google Scholar]
44. Masuda A, Ushida K, Nishimura G, Kinjo M, Tamura M, Koshino H, Yamashita K, Kluge T. Experimental evidence of distance-dependent diffusion coefficients of a globular protein observed in polymer aqueous solution forming a network structure on nanometer scale. J Chem Phys. 2004;121:10787–10793. [PubMed] [Google Scholar]
45. Ogston AG, Preston BN, Wells JD. Transport of compact particles through solutions of chain-polymers. Proc R Soc London Ser A. 1973;333:297–316. [Google Scholar]
46. Starovasnik MA, Braisted AC, Wells JA. Structural mimicry of a native protein by a minimized binding domain. Proc Natl Acad Sci USA. 1997;94:10080–10085. [PMC free article] [PubMed] [Google Scholar]
47. Rief M, Fernandez JM, Gaub HE. Elastically coupled two-level systems as a model for biopolymer extensibility. Phys Rev Lett. 1998;81:4764–4767. [Google Scholar]
48. Bohrer MP, Patterson GD, Carroll PJ. Hindered diffusion of dextran and ficoll in microporous membranes. Macromolecules. 1984;17:1170–1173. [Google Scholar]
49. Reshetnyak YK, Segala M, Andreev OA, Engelman DM. A monomeric membrane peptide that lives in three worlds: in solution, attached to, and inserted across lipid bilayers. Biophys J. 2007;93:2363–2372. [PMC free article] [PubMed] [Google Scholar]
50. Bunagan MR, Gao JM, Kelly JW, Gai F. Probing the folding transition state structure of the villin headpiece subdomain via side chain and backbone mutagenesis. J Am Chem Soc. 2009;131:7470–7476. [PMC free article] [PubMed] [Google Scholar]
51. Guo L, Chowdhury P, Fang JY, Gai F. Heterogeneous and anomalous diffusion inside lipid tubules. J Phys Chem B. 2007;111:14244–14249. [PubMed] [Google Scholar]
52. Haustein E, Schwille P. Fluorescence correlation spectroscopy: novel variations of an established technique. Annu Rev Biophys Biomol Struct. 2007;36:151–169. [PubMed] [Google Scholar]