Thermal Unfolding Kinetics of Ubiquitin in the Microsecond-to-Second

9912

J. Phys. Chem. B 2010, 114, 9912–9919

Thermal Unfolding Kinetics of Ubiquitin in the Microsecond-to-Second Time Range Probed by Tyr-59 Fluorescence Melinda Noronha,*,† Hana Gerbelova´,† Tiago Q. Faria,‡ Daniel N. Lund,§,⊥ D. Alastair Smith,§,⊥ Helena Santos,‡ and Anto´nio L. Mac¸anita*,† Centro de Quı´mica Estrutural, Instituto Superior Te´cnico, Technical UniVersity of Lisbon, 1049-001 Lisboa, Portugal, Instituto de Tecnologia Quı´mica e Biolo´gica, UniVersidade NoVa de Lisboa, Rua da Quinta Grande 6, Apartado 127, 2780-156 Oeiras, Portugal, and School of Physics and Astronomy, UniVersity of Leeds, Leeds, LS2 9JT, United Kingdom ReceiVed: May 7, 2010; ReVised Manuscript ReceiVed: June 10, 2010

Thermal folding/unfolding kinetics of wild-type ubiquitin (wt-UBQ) was studied in a wide time range, from microseconds to seconds, by combining rapid-mixing T-jump and laser T-jump with fluorescence detection (MTJ-F and LTJ-F, respectively) to monitor the fluorescence changes of Tyr-59 located on the 310-helix. The kinetics occurs exclusively in the millisecond to second time range, and the decays are strictly single exponential. From global analysis of folding and unfolding decays, the kf and ku values were determined, without use of the equilibrium constant Ku. The activation enthalpy of folding is negative (∆Hf#(Tm) ) -10.8 kcal/mol), but the free energy of the transition state is substantially larger than that of the unfolded state (∆Gf#(Tm) ) 7.6 kcal/mol . RTm). Thus, wt-UBQ behaves as a two-state folder, when folding is monitored by the fluorescence of Tyr-59. The observation of kinetics on the microsecond time scale, when folding is monitored by the disruption of hydrogen bonds between β-strands, using nonlinear infrared spectroscopy of the amide I vibrations (LTJ-DVE) [Chung, H. S.; Tokmakoff, A. Proteins: Struct., Funct., Bioinf. 2008, 72, 474-487], seems to result from the fact that MTJ-F monitors the effective unfolding (backbone exposure to water) of the thermally excited protein alone, while LTJ-DVE also monitors preliminary events (hydrogen-bond breaking) and thermal re-equilibration of the thermally excited protein. Introduction Kinetics of the thermal unfolding of proteins has been studied over the past 10 years using the laser temperature jump (LTJ) technique.1 LTJ can produce temperature jumps of 10-15 °C while keeping constant the final temperature (Tf) for ca. 2-3 ms.1 Hence, only the so-called fast folders (peptides2-4 and small proteins5-11) were initially studied. Later, the upper limit of 3 ms of the LTJ technique was extended to longer time scales (10-47 ms) by taking into account the temperature time drift after the T-jump.12 The appearance of the rapid-mixing stopped-flow technique in 2008 (MTJ)13 provided access to both positive and negative T-jumps, with a time resolution of a few milliseconds (dead time ca. 4 ms) and a constant final temperature up to hours. Therefore, by combining LTJ and MTJ, thermal unfolding/ folding kinetics can be studied for any protein from a few microseconds to hours. wt-Ubiquitin (wt-UBQ), a small protein of 76 amino acid residues (8.5 kDa), has been extensively used as a model for both chemical14-20 and thermal unfolding studies10-12,21 (Figure 1). The resulting data have raised a number of questions concerning the folding/unfolding kinetics of this protein and/or of its mutants. Specifically, deviations from simple kinetics have * To whom correspondence should be addressed. E-mail: macanita@ ist.utl.pt (A.L.M); [email protected] (M.N). † Technical University of Lisbon. ‡ Universidade Nova de Lisboa. § University of Leeds. ⊥ Present address: Avacta Group Plc, York Biocentre, Innovation Way, York, YO10 5NY, U.K.

Figure 1. Structure of wt-ubiquitin showing the location of Tyr-59 at the end of the 310-helix between β-strands IV and V (βIV and βV).

been observed, depending on experimental conditions17 and detection methods.18,19 In some cases, the deviations were consistent with the presence of intermediate states, while in other cases the deviation consisted of a fast decay component, either directly observed12 or indirectly deduced (burst phase).14-16

10.1021/jp104167h  2010 American Chemical Society Published on Web 07/15/2010

Thermal Folding/Unfolding Kinetics of wt-UBQ Intermediate states have been observed essentially in experiments involving chemical denaturation of tryptophan-mutated UBQ.22 The fast decay component was observed with thermal (LTJDVE) denaturation of wt-UBQ by probing the disruption of interhydrogen bonds between β-strands with nonlinear infrared spectroscopy of the amide I vibrations in acidic conditions.11,12 The nonexponential kinetics was observed in a time range from tenths of a microsecond to 2.7 ms, although the major change in the probe signal occurred within 10-47 ms where the traces could be fitted with single-exponential functions.11,12 The fast nonexponential component accounted for a 7.7% variation of signal intensity and was dependent on the size of the T-jump (∆T) and final temperature (Tf). The authors attributed the fast component to downhill unfolding of a small population of the native protein located energetically above the transition state immediately after the T-jump.12 Previously we studied the thermal unfolding of wt-UBQ at pH 1.5 using time-resolved fluorescence spectroscopy (TRFS) by following the fluorescence of its single tyrosine, Tyr-59, located on the 310-helix.23 The purpose was to determine whether the fluorescence of this tyrosine residue could be used as an intrinsic probe to study the protein thermal folding/unfolding, thus allowing the use of wt-UBQ as a model protein in our studies on the mechanisms underlying protein stabilization by osmolytes.24,25 The results showed that indeed the fluorescence of Tyr-59 was an adequate probe, and, moreover, they were consistent with the two-state model.23 However, because there was a discrepancy in the Tm values (2.4 °C) and in the unfolding enthalpy values (5.1 kcal/mol) determined by TRFS and DSC, the apparent simplicity of the folding/unfolding mechanism was not free of doubt. In this work, we combine LTJ-F and MTJ-F to further study the thermal folding/unfolding kinetics of wtUBQ using the fluorescence of Tyr-59 in an effort to resolve this issue. Materials and Methods Samples. Bovine ubiquitin was purchased from Sigma and purified with a Mono-S column using a salt gradient as previously described.23 N-Acetyltyrosinamide (NAYA) was purchased from Sigma-Aldrich. Sodium acetate (Merck), sodium chloride (Merck), acetic acid (Riedel-deHae¨n), and hydrochloric acid (Riedel-deHae¨n) were of pro analysis grade. A final protein concentration of 40 µM was used in all rapid-mixing T-jump experiments. Double-distilled water was used in all samples except for samples in D2O (Aldrich) where the pD was adjusted using deuterium chloride 99 atom % (Aldrich). Rapid-Mixing T-jumps (MTJ-F). Rapid-mixing T-jump (MTJ-F) measurements were carried out using a new temperature jump accessory from BioLogic coupled to an existing Biologic SFM-4 with a MPS-52 microprocessor unit connected via a fiber optic cable to a xenon mercury lamp and equipped with a circulating water bath for temperature control.13,26 The accessory achieves temperature jumps as high as 40 °C by rapid mixing of two solutions initially at different temperatures T1 and T2. The final temperature of the mixture (Tf) is determined both by the initial temperatures T1 and T2 and by the volume ratio of the two solutions. Three Peltier elements (BioLogic TCU-250) are used to control the initial temperatures of the two solutions and the observation cell after mixing. The heat generated by the Peltier elements is dissipated using a water bath circulator. Prior to the T-jump experiments for each Tf, test shots were carried out for both unfolding and refolding to determine the optimal conditions of the (1) appropriate voltage

J. Phys. Chem. B, Vol. 114, No. 30, 2010 9913 (V) for a sufficiently strong signal intensity without saturation, (2) correct flow rates (Q) to ensure that the temperature is kept constant in the region between the mixer chamber and the observation cell (Qf ) 12 mL/s), (3) correct dilution factors to avoid problems of aggregation while obtaining a strong enough signal change (protein (P) to buffer (B) ratio of 1:2), and (4) volume per shot to reduce noise (VB ) 0.146 mL, VP ) 0.073 mL, Vf ) 0.219 mL). For each Tf, the following data were collected: T-jumps for unfolding (38 °C f Tf) and refolding (70 °C f Tf), and shots with only UBQ and buffer both at the same Tf, to guarantee that the three Peltier elements are intercalibrated and to ensure that upon mixing of protein and buffer no deviations are observed from the temperature (Tf) of the observation cell, which if present could lead to artifacts in kinetic curves. Sampling periods of 2.3 s were used after checking the absence of fluorescence changes at longer times (23 s sampling period). The folding/unfolding kinetics of wtUBQ was monitored by changes in the intrinsic fluorescence of Tyr-59 with excitation at λ ) 275 nm, and the emission was collected above 290 nm using a 290 nm Schott cutoff filter. Experiments were carried out in 12.5 mM sodium acetate buffer, and the pH was adjusted with HCl to final concentrations of [H+] ) 33 and 100 mM (pH ) 1.5 and 1.0). Further experiments in D2O were carried out also in 12.5 mM sodium acetate buffer and a final concentration of [D+] ) 0.1 M (pD ) 1) using DCl. Laser-Induced T-jumps. Laser-induced T-jumps were carried out using a home-built LTJ-F apparatus as described elsewhere.9 The T-jumps were generated by Raman shifting the 1064 nm output (1 J per pulse) of a Q-switched Nd:YAG Laser [Spectra-Physics (Hemel Hempstead, U.K.) Quanta Ray Lab 170] in a 1 m length stainless steel tube of methane at 30 atm to produce an infrared pulse at 1550 nm with 10-30% conversion efficiency. The protein sample is contained in a quartz cell of path length 0.5 mm [Custom (Custom LC, Hellma, U.K.)], thermostatted by a Peltier device and Peltier controller [Marlow (Dallas) SE5010] to maintain the sample at constant temperature. Excitation was carried out at 275 nm by using the frequency-tripled output of a femtosecond mode-locked Ti: sapphire laser [Coherent (Santa Clara, CA) MIRA 900/VERDI]. The mode-locked output at 37 MHz was pulse-picked to give a pulse separation of 130 ns and a pulse width of 290 nm). Fu,f(0) are the preexponential coefficients at t ) 0 for the unfolding and refolding traces; fN and fU are the fluorescence signal of native and unfolded protein, respectively.

(no T-jump) is much less than that observed after the T-jump, which indicates that practically all the unfolding kinetics of wtUBQ is on a time scale greater than 1.8 ms. Similar results were obtained upon a T-jump from 50 to 60 °C (data not shown). Discussion Kinetic Analysis. The foregoing data shows that the thermal unfolding/folding kinetics of wt-UBQ at pH 1.5 is adequately described with the two-state model represented in scheme 1 below, where ku and kf are the rate constants of unfolding and folding and N and U the native and unfolded states, respectively. ku

N {\} U

scheme 1

kf

The time evolution of the native N(t) and unfolded U(t) conformations obeys eqs 1 and 2, which express the well-known result that the decay functions, N(t) and U(t), are sums of a

Thermal Folding/Unfolding Kinetics of wt-UBQ

J. Phys. Chem. B, Vol. 114, No. 30, 2010 9915

TABLE 1: Rate Constant (kobs) λ2), Fluorescence Intensity Signals (in V) at t ) 0 of the Unfolding Fu(0) and Folding Ff(0) Fluorescence Traces, and Fluorescence Intensity Signal (in V) at t ) ∞, F(∞) T (°C)

kobs (s-1)

Fu(0)

Ff(0)

F(∞)

45 47 49 51 53 55 57 58 59 62 63 65

11.0 11.3 11.9 12.1 12.7 16.2 19.2 27.7 31.3 52.6 69.1 110.2

1.02 0.99 0.94 0.91 0.95 0.74 0.72 0.57 0.64 0.54 0.70 0.35

0.61 0.52 0.49 0.36 0.34 0.31 0.23 0.25 0.20 0.19 0.14 0.15

0.99 0.93 0.86 0.76 0.65 0.52 0.41 0.36 0.32 0.21 0.19 0.14

constant term with an exponential term with rate constant kobs )λ2 ) ku + kf (see Supporting Information for details).

[]

][ ]

[

kf ku 1 N (t) ) (1/λ2) -λ2t k -k U u u e

[]

[

][ ]

kf -kf 1 N (t) ) (1/λ2) ku kf e-λ2t U

(for unfolding)

(1)

(for folding)

(2)

The determination of ku and kf is usually carried out by coupling the value of kobs ) ku + kf, obtained from either a folding or unfolding trace, to that of the equilibrium constant Ku ) ku/kf, determined from DSC or any other suitable method.23 However, the availability of both folding and unfolding traces, from MTJ-F experiments, provides an alternative, more robust methodology to extract the values of ku and kf from the kinetic data alone, using the information contained the pre-exponential coefficients (the values of N(0), U(0), and N(∞) ) U(∞)). In order to profit from this information, the experimental signal (UV absorption, fluorescence, etc.) has to be converted into concentrations (or mole fractions) of the native and unfolded protein. The experimental signal, in our case the fluorescence signal F(t), is the summed fluorescence signals of N and U, which are in turn proportional to their respective concentrations, N(t) and U(t),23 (eq 3, where the coefficients fN and fU are the fluorescence quantum yields of N and U, respectively, apart from a common instrumental constant).

F(t) ) fNN(t) + fUU(t)

(3)

Ff(t) ) [fNkf + fUku + (fUkf - fNkf) exp(-λ2t)]/(ku + kf) (5) Therefore, the unfolding Fu(t) and folding Ff(t) traces contain all the information required for the evaluation of ku and kf, without using the equilibrium constant obtained from independent experiments, i.e, two equations (eqs 6 and 7) for the two unknowns (kf and ku), with all data, fN, fU, Fu,f(∞), and kobs, taken from the decays (see Figure 3).

kf + ku ) kobs

(6)

fNkf + fUku ) Fu,f(∞)/kobs

(7)

The global analysis of each pair of folding and unfolding traces, subjected to the foregoing constrains (eqs 6 and 7), besides being more robust, provides the possibility of comparing the equilibrium constant derived from kinetic data Kukin to that independently obtained from equilibrium measurements KuTRFS, and thus verifying the internal consistency of all (equilibrium and kinetic) data. An additional and useful advantage of this comparison is the possibility of checking the presence/absence of fast decay components (on the submillisecond time scale) from the disagreement/agreement of equilibrium and kinetic data. Table 2 shows the values of kf and ku derived from the global fit of eqs 6 and 7 to each pair of unfolding Fu(t) and folding Ff(t) experimental traces (fits in Figure 3) and compares the values of the equilibrium constant calculated from the kinetic data (Kukin ) ku/kf) to those obtained from equilibrium studies (KuTRFS ) xU/xN) using time-resolved fluorescence spectroscopy (Figure 2b).23 The values of Kukin and KuTRFS are in reasonable agreement within experimental error. Temperature Dependence of the Rate Constants. Arrhenius plots of kf and ku (Figure 5) show that ku is strongly activated (Eau) 49 ( 1 kcal mol-1, A0u ) 8.3 ( 7 × 1033 s-1), whereas kf is slightly negatively activated (Eaf ) -6 ( 1 kcal mol-1, A0f ) 4.3 ( 2 × 10-4 s-1). The non-Arrhenius behavior of kf (negative Eaf) has been previously reported for several other proteins.9,12,13,28-32 Because the negative Eaf could result from a positive ∆Hf# coupled to a sufficiently negative activation heat capacity of folding (∆CPf#), as reported for CI2 (∆CPf# ) -0.31 kcal mol-1 K-1),28 we have analyzedthetemperaturedependenceofkf withtheGibbs-Helmholtz equation (eq 8), by globally fitting both folding and unfolding rate constants with Eyring’s equation, eq 9, subject to the constrain ∆CPu# - ∆CPf# ) ∆CPTRFS ) 0.89 kcal mol-1 K-1 from equilibrium (TRFS) experiments.23

(

# # # # ∆Gu/f ) ∆Hu/f (Tm) + ∆CPu/f (T - Tm) - T ∆Su/f (Tm) +

( ))

# ln ∆CPu/f

Combination of eqs 1-3 leads to eqs 4 and 5 (see also Supporting Information), which show that the values of the unfolding Fu(t) and folding Ff(t) traces at t ) 0 are equal to fN and fU, respectively, and the traces take a common value of (fNkf + fUku)/kobs at t ) ∞ (see Figure 3).

Fu(t) ) [fNkf + fUku + (fNku - fUku) exp(-λ2t)]/(ku + kf) (4)

ku/f )

(

# ∆Gu/f k BT 〈κ〉 exp h RT

)

T Tm

(8)

(9)

Acceptable fits (Figure 6a) could be obtained only with negative ∆Hf#(Tm) values. Positive ∆Hf# values always induced large errors in the temperature range below the Tm.

9916

J. Phys. Chem. B, Vol. 114, No. 30, 2010

Noronha et al.

Figure 4. (a) Normalized fluorescence signal (F) of NAYA at 65 °C (gray) and the laser T-jump from Ti ) 55 °C to Tf ) 65 °C (black); (b) fluorescence signal of wt-UBQ, pH 1.5, at 65 °C (gray), and the laser T-jump from Ti ) 55 °C to Tf ) 65 °C (black) shows a negligible decrease in the fluorescence signal up to 1.8 ms.

TABLE 2: Rate Constants of Unfolding (ku) and Folding (kf) Calculated from Global Analysis of Folding and Unfolding Traces, and Equilibrium Constants Calculated from Kinetic (Kukin ) ku/kf) and from Equilibrium TRFS (KuTRFS) Data T (°C)

ku (s-1)

kf (s-1)

Kkin u

KTRFS u

45 47 49 51 53 55 57 58 59 62 63 65

1.1 1.4 2.5 3.7 5.1 8.2 12.3 20.3 24.0 46.5 63.3 105

9.9 9.9 9.5 8.6 7.6 8.0 6.9 7.4 7.2 6.1 5.8 5.6

0.11 0.14 0.26 0.43 0.67 1.03 1.78 2.74 3.33 7.62 10.9 18.8

0.09 0.14 0.24 0.39 0.66 1.06 1.90 2.45 3.28 7.67 10.1 17.9

As expected, inclusion of solvent friction using Kramer’s equation (eq 10)13,33 provides an even more negative value for ∆Hf# (Table 3).

( ) (

ku/f ) ν

ηTm η

# ∆Gu/f exp RT

)

(10)

Interestingly, our data for wt-UBQ folding kinetics is very close to that obtained by Torrent et al.13 for RNase A using MTJ-F (Table 3, third row). A negative ∆Hf#(Tm) value means that the enthalpy decrease in the transition from the unfolded to the transition state (resulting from the intraprotein hydrophobic forces) is larger

Figure 5. Arrhenius plots of rate constant of folding kf (black) and unfolding ku (gray) of wt-UBQ at pH 1.5 using the fluorescence of Tyr-59 showing that ku is strongly activated whereas kf is slightly negatively activated.

than the increase in enthalpy that is required for reaching the transition state through rotations of the protein-backbone bonds in the presence of solvent friction. This situation has been predicted from theory.34-38 In this case, the folding free energy barrier ∆Gf#(Tm), if existent, is exclusively entropic. The value of ∆Gf#(Tm), obtained from the data in the second row of Table 3 and eq 10, is 7.6 kcal/mol, ca. 1 order of magnitude larger than RTm. Thus, within the limitations arising from the application of eq 10, there is a sufficiently high free energy barrier separating the native and unfolded states so that transition-state theory can be applied. Summarizing, the foregoing results show that thermal unfolding of wt-UBQ at pH 1.5 probed by Tyr-59, from both equilibrium (TRFS and DSC) and kinetic experiments, follows a two-state model, occurring in the millisecond time scale, with no kinetics observed in the µs-2 ms time range. Comparison with Literature Data. The kinetics of thermal unfolding/folding of wt-UBQ has been previously studied in D2O at pD 1.0, using LTJ with infrared (IR) or differential vibrational echo spectroscopy (DVE) detection to probe the temperature-induced changes in the amide I vibrational mode of β-strands I-V.11,12,21 Comparison of the data in ref 12with ours shows qualitative agreement on the millisecond time scale; namely, “single-exponential kinetics” was observed, indicating the absence of intermediate states. However, there are a number of discrepancies, namely, (1) observation of kinetics on the microsecond time scale vs our observation of its absence, and (2) different values of kf and ku leading to (3) different Tm values (64 °C, in D2O at pD 112,21 vs 54.6 °C in H2O, at pH 1.523). Because the discrepancies in the Tm and rate constants could arise from the different experimental conditions of the two studies, we have carried out rapid-mixing T-jumps at pH 1 ([H+] ) 0.1 M) and at pD 1 ([D+] ) 0.1 M in D2O). The folding and unfolding decays were qualitatively similar to those obtained in H2O at pH 1.5; namely, they could be globally fitted with single-exponential functions, and there was no indication of kinetics in the microsecond time range.39 The resulting values of kf and ku are plotted in Figure 7, together with those obtained in ref 12. Our values measured in H2O at pH 1.5 are also shown for comparison. Decreasing the pH from 1.5 to 1 stabilizes wt-UBQ by increasing only kf (Figure 7a, red squares) with no change in ku (Figure 7b, red squares). The consequent stabilization (ca. 3 °C in the Tm, from 57.2 °C at pH 1.523 to 60.9 °C at pH 1.0) is in apparent contrast with the reported destabilization of wt-UBQ with decreasing pH.40 The larger kf (and unchanged ku) likely results from the screening of the intraprotein electrostatic repulsion (positive charges at pH 1) at higher [Cl-] (ionic

Thermal Folding/Unfolding Kinetics of wt-UBQ

J. Phys. Chem. B, Vol. 114, No. 30, 2010 9917

Figure 6. Fitting of kf values calculated using Erying transtion-state theory (open circles) to the experimental (a) kf and (b) ku (black squares) of UBQ at pH 1.5, imposing ∆CPu# - ∆CPf# ) ∆CPTRFS ) 890.0 cal mol-1 K-1 (obtained from TRFS) and at Tm ) 54.6 °C. The values of ∆S# are affected by the pre-exponential factor of Eyring’s equation (kBT/h ) 6.8 × 1012 s-1) and should be regarded just as fitting parameters.

TABLE 3: Values of the Activation Parameters Obtained from Fitting Eqs 9-10 to kf and ku for Wild-Type Ubiquitin # a (wt-UBQ) at pH 1.5, Using Eyring’s Transition-State Theory and Kramers Model Corrected for ∆Cpu/f

wt-UBQ (Eyring) wt-UBQ (Kramers) RNase A (Kramers)

∆Hf#(Tm), kcal mol-1

∆Sf#(Tm)b, kcal mol-1 K-1

∆Cpf#, kcal mol-1 K-1

∆H#u(Tm), kcal mol-1

∆S#u(Tm)b, kcal mol-1 K-1

∆Cpu#, kcal mol-1 K-1

-5.6 -10.8 -10.5

-0.072 -0.056 -0.064

-0.113 -0.093

49.6 44.9 105.2

0.097 0.114 0.284

0.757 0.797

a The third row shows data obtained by Torrent et al.13for wild-type ribonuclease A (RNase A) at pH 5.0 for comparison of the folding parameters. b The values of ∆S# are affected by the pre-exponential factors of either eq 9 or 10.

Figure 7. Comparison of the rate constants of (a) folding and (b) unfolding obtained by probing the 310-helix (black; pH 1.5), (red; pH 1), (blue; pD 1), and the β-strands III-V (gray, pD 1; Chung and Tokmakoff) of wt-UBQ.

strength) in the denatured state. Actually, it has been reported that addition of 0.5 M NaCl to wt-UBQ at pH 2 increases the Tm by 22.5 °C.41 Changing from H2O at 0.1 M HCl to D2O at 0.1 M DCl induces an additional increase in kf (Figure 7b, blue squares) and a decrease in ku (Figure 7a, blue squares), consistent with a further increase in the Tm (Tm ) 62.7 °C from DSC). In terms of activation energies, it is interesting to note that kf becomes even more negatively activated (Eaf ) -15 ( 1 kcal mol-1) whereas ku has a similar activation energy as in H2O (Eau) 50 ( 2 kcal mol-1, Table 4). Comparison of our data with that presented by Chung and Tokmakoff in the same solvent conditions (Table 4, third and fourth row, respectively) shows that the differences persist: Our values for kf are smaller (Figure 7a), although the discrepancies result from very small differences in the values of Ea,f and Af (Table 4, 3rd and fourth row), and both Eu,f and Au are substantially different, although the ku values

themselves are not much different (Figure 7b). Moreover, no indication of kinetics on the microsecond time scale was found. Probe-Dependent Kinetics. There are three possible explanations for the observed differences. The first is that the differences result from accumulated intrinsic errors of the three experimental techniques used to derive the data (DSC, MTJ-F, LTJ-F, and LTJ-DVE). However, this could account for differences in rate constants and activation energies, but not for the presence vs absence of kinetics on the microsecond time scale. The second is that the 310-helix and β-strands III-V simply unfold differently from each other. As Tyr-59 is located on the 310-helix of the C-terminal of wt-UBQ, while data from Chung and Tokmakoff probe β-strands III-V, the foregoing differences in ku could suggest that unfolding of the β-strands of wt-UBQ at pD 1 begins at a lower temperature and ends at a higher temperature than the 310-helix unfolding; i.e., at low temperatures

9918

J. Phys. Chem. B, Vol. 114, No. 30, 2010

Noronha et al.

TABLE 4: Values of Activation Energies and Pre-exponential Coefficients of Unfolding (Ea,u, Au), and Folding (Ea,f, Af) of wt-UBQ, from Arrhenius Plots of kf and ku Measured under Different Experimental Conditions and Different Probes experiment in H2O, H2O, D2O, D2O, a

pH pH pD pD

1.5 1.0 1.0 1.0a

probed by 310-helix 310-helix 310-helix β-strands

Ea,u, kcal mol-1 49 48 50 21

Ea,f, kcal mol-1 -6 -9 -14.9 -14.7

Au, s-1 8 × 10 4 × 1030 5 × 1032 1.2 × 1015 33

Af, s-1

Tm, °C -4

4.3 × 10 1.0 × 10-5 5.9 × 10-9 5.8 × 10-9

57.2b 60.9b 62.7b 64.0

Values calculated with data from ref 12. b From DSC measurements.

the 310-helix region would be more stable than the β-strands, whereas at higher temperatures the inverse would be true. The 310-helix and the R-helix of wt-UBQ have been reported as recognition sites through noncovalent protein-protein interactions,42 and their higher stability at temperatures below the Tm may have functional importance. Although interesting, this possibility is odd because the 310-helix is located between β-strands IV and V (see Figure 1) that do unfold. The third, more plausible, explanation for the difference in rate constants on the millisecond time scale and the observation of kinetics by LTJ-DVE versus the absence of kinetics by MTJ-F on the microsecond time scale results from the fact that the two kinetic techniques probe different phenomena at the molecular level. The MTJ-F probes the exposure of tyrosine to water, while the LTJ-DVE probes the breaking of interstrand hydrogen bonds. It is, therefore, intuitive that for unfolding, the incipient hydrogen-bond breaking should be immediately detected by LTJ-DVE, while observation of a fluorescence signal change (by MTJ-F) requires further displacements and backbone rotations leading to water exposure of Tyr-59. Consequently, MTJ-F monitors the unfolding of the thermally excited protein alone, while LTJ-DVE also monitors the preparation and thermal re-equilibration of the thermally excited protein. This explains the observation of microsecond kinetics by LTJ-DVE and not by MTJ-F. Furthermore, the resulting greater complexity of the LTJDVE signal and the low upper limit of its temporal range make an accurate characterization of the longest decay time difficult, due to mixing with the shorter decay times (or stretched exponential). This difficulty, possibly leading to larger kobs values, coupled to the preliminary nature of our data in D2O may explain the differences observed in the rate constants. A global analysis of the MTJ-F, LTJ-F, and LTJ-DVE experimental data would certainly provide a more accurate basis for modeling the folding/unfolding pathway of wt-UBQ. Conclusions The thermal unfolding of wt-UBQ at pH 1.5, followed by the fluorescence of its single tyrosine residue Tyr-59, from both equilibrium and kinetic studies, shows that wt-UBQ behaves as a two-state folder on the millisecond time scale. No kinetics was observed in the microsecond time region, and no evidence for intermediate states was found. Arrhenius plots for kf and ku at pH 1.5 show that kf is slightly negatively activated (Eaf ) -6 ( 1 kcal mol-1), whereas ku is highly activated, Eau ) 49 ( 1 kcal mol-1). Analysis of the temperature dependence of kf with the Gibbs-Helmholtz equation results in negative values for both ∆H#f and∆Cpf#. Thus, the negative value of ∆Cpf# alone cannot explain the observation of a decreasing kf in the entire transition temperature range (from 38 to 70 °C). Comparison of kf and ku values at two different pH values shows an increase only in kf at pH 1 as compared to pH 1.5 (due to Cl--induced acceleration of folding), whereas compari-

son of kf and ku in H2O and D2O shows that D2O has a stabilizing effect brought about by both an increase in kf and a decrease in k u. Comparison of data from LTJ-DVE (ref 12)with our results from MTJ-F in the same solvent conditions, at pD 1 in D2O, revealed differences in the nature of the kinetic traces and in the kf and ku values. We propose that the differences result from the fact that MTJ-F monitors the actual unfolding (backbone exposure to water) of the thermally excited protein alone, while LTJ-DVE also monitors the preliminary events (hydrogen-bond breaking) and thermal re-equilibration of the thermally excited protein. Acknowledgment. The work was supported by the Fundac¸a˜opara a Cieˆncia e a Tecnologia (FCT), Portugal, Projects POCI/QUI/56585/04 and POCI/BIA-PRO/57263/04, and by the European Commission, 5th Framework Programme contract QLK3-CT-2000-00640. M.N. and T.F. acknowledge the FCT for postdoc grants SRFH/BPD/27128/2006 and SRFH/BPD/ 44428/2008. Supporting Information Available: This information is available free of charge via the Internet at http://pubs.acs.org. References and Notes (1) Kubelka, J. Photochem. Photobiol. Sci. 2009, 8, 499–512. (2) Eaton, W. A.; Munoz, V.; Thompson, P. A.; Henry, E. R.; Hofrichter, J. Acc. Chem. Res. 1998, 31, 745–753. (3) Callender, R. H.; Dyer, R. B.; Gilmanshin, R.; Woodruff, W. H. Annu. ReV. Phys. Chem. 1998, 49, 173–202. (4) Eaton, W. A.; Munoz, V.; Hagen, S. J.; Jas, G. S.; Lapidus, L. J.; Henry, E. R.; Hofrichter, J. Annu. ReV. Biophys. Biomol. Struct. 2000, 29, 327–359. (5) Ballew, R. M.; Sabelko, J.; Gruebele, M. Nat. Struct. Biol. 1996, 3, 923–926. (6) Yang, W. Y.; Gruebele, M. Nature 2003, 423, 193–197. (7) Liu, F.; Du, D. G.; Fuller, A. A.; Davoren, J. E.; Wipf, P.; Kelly, J. W.; Gruebele, M. Proc. Natl. Acad. Sci. U.S.A. 2008, 105, 2369–2374. (8) Kubelka, J.; Henry, E. R.; Cellmer, T.; Hofrichter, J.; Eaton, W. E. Proc. Natl. Acad. Sci. U.S.A. 2008, 105, 18655–18662. (9) Dimitriadis, G.; Drysdale, A.; Myers, J. K.; Arora, P.; Radford, S. E.; Oas, T. G.; Smith, D. A. Proc. Natl. Acad. Sci. U.S.A. 2004, 101, 3809–3814. (10) Sabelko, J.; Ervin, J.; Gruebele, M. Proc. Natl. Acad. Sci. U.S.A. 1999, 96, 6031–6036. (11) Chung, H. S.; Ganim, Z.; Jones, K. C.; Tokmakoff, A. Proc. Natl. Acad. Sci. U.S.A. 2007, 104, 14237–14242. (12) Chung, H. S.; Tokmakoff, A. Proteins: Struct., Funct., Bioinf. 2008, 72, 474–487. (13) Torrent, J.; Marchal, S.; Ribo, M.; Vilanova, M.; Georges, C.; Dupont, Y.; Lange, R. Biophys. J. 2008, 94, 4056–4065. (14) Khorasanizadeh, S.; Peters, I. D.; Butt, T. R.; Roder, H. Biochemistry 1993, 32, 7054–7063. (15) Khorasanizadeh, S.; Peters, I. D.; Roder, H. Nat. Struct. Biol. 1996, 3, 193–205. (16) Krantz, B. A.; Sosnick, T. R. Biochemistry 2000, 39, 11696–11701. (17) Went, H. M.; Benitez-Cardoza, C. B.; Jackson, S. E. FEBS Lett. 2004, 567, 333–338. (18) Qin, Z.; Ervin, J.; Larios, E.; Gruebele, M.; Kihara, H. J. Phys. Chem. B. 2002, 106, 13040–13046. (19) Larios, E.; Li, J. S.; Schulten, K.; Kihara, H.; Gruebele, M. J. Mol. Biol. 2004, 340, 115–125.

Thermal Folding/Unfolding Kinetics of wt-UBQ (20) Jackson, S. E. Org. Biomol. Chem. 2006, 4, 1845–1853. (21) Colley, C. S.; Clark, I. P.; Griffiths-Jones, S. R.; George, M. W.; Searle, M. S. Chem. Commun. 2000, 16, 1493–1494. (22) Valle´e-Be´lisle, A.; Michnick, S. W. J. Mol. Biol. 2007, 374, 791– 805. (23) Noronha, M.; Lima, J. C.; Bastos, M.; Santos, H.; Mac¸anita, A. L. Biophys. J. 2004, 87, 2609–2620. (24) Faria, T. Q.; Lima, J. C.; Bastos, M.; Mac¸anita, A. L.; Santos, H. J. Biol. Chem. 2004, 279, 48680–48691. (25) Pais, T. M.; Lamosa, P.; Garcia-Moreno, B.; Turner, D. L.; Santos, H. J. Mol. Biol. 2009, 394, 237–250. (26) Biologic user’s manual, mT-JUMP ACCESSORY for stopped-flow apparatus, V1.4, August 2006. (27) Noronha, M.; Lima, J. C.; Lamosa, P.; Santos, H.; Maycock, C.; Ventura, R.; Mac¸anita, A. L. J. Phys. Chem. A 2004, 108, 2155–2166. (28) Oliveberg, M.; Tan, Y. J.; Fersht, A. R. Proc. Natl. Acad. Sci. U.S.A. 1995, 92, 8926–8929. (29) Scalley, M. L.; Baker, D. Proc. Natl. Acad. Sci. U.S.A. 1997, 94, 10636–10640. (30) Schindler, T.; Schmid, F. X. Biochemistry 1996, 35, 16833–16842. (31) Crane, J. C.; Koepf, E. K.; Kelly, J. W.; Gruebele, M. J. Mol. Biol. 2000, 298, 283–292.

J. Phys. Chem. B, Vol. 114, No. 30, 2010 9919 (32) Yang, W. Y.; Gruebele, M. Biochemistry 2004, 43, 13018–13025. (33) Ansari, A.; Jones, C. M.; Henry, E. R.; Hofrichter, J.; Eaton, W. A. Science 1992, 256, 1796–1798. (34) Bryngelson, J. D.; Onuchic, J. N.; Socci, N. D.; Wolynes, P. G. Protein: Struct., Funct., Genet. 1995, 21, 167–195. (35) Socci, N. D.; Onuchic, J. N.; Wolynes, P. G. J. Chem. Phys. 1996, 104, 5860–5868. (36) Karplus, M. J. Phys. Chem. B. 2000, 104, 11–27. (37) Bicout, D. J.; Szabo, A. Protein Sci. 2000, 9, 452–465. (38) Best, R. B.; Hummer, G. Phys. ReV. Lett. 2006, 96, 228104. (39) The presence of kinetics would lead to values of FU(0) lower than the fluorescence signal at the initial temperature corrected for the change in fluorescence quantum yield from the initial to the final temperature of the T-jump (and/or the reciprocal for FF(0)). (40) Wintrode, P. L.; Makhatadze, G. I.; Privalov, P. L. Protein: Struct., Funct., Genet. 1994, 18, 246–253. (41) Makhatadze, G. I.; Lopez, M. M.; Richardson, J. M., III; Thomas, S. T. Protein Sci. 1998, 7, 689–697. (42) Hurley, J. H.; Lee, S.; Prag, G. Biochem. J. 2006, 399, 361–372.

JP104167H