Article pubs.acs.org/biochemistry
Combined Isothermal Titration and Differential Scanning Calorimetry Define Three-State Thermodynamics of fALS-Associated Mutant Apo SOD1 Dimers and an Increased Population of Folded Monomer Helen R. Broom, Kenrick A. Vassall,† Jessica A. O. Rumfeldt, Colleen M. Doyle, Ming Sze Tong, Julia M. Bonner,‡ and Elizabeth M. Meiering* Department of Chemistry, University of Waterloo, Waterloo, Ontario N2L 3G1, Canada S Supporting Information *
ABSTRACT: Many proteins are naturally homooligomers, homodimers most frequently. The overall stability of oligomeric proteins may be described in terms of the stability of the constituent monomers and the stability of their association; together, these stabilities determine the populations of different monomer and associated species, which generally have different roles in the function or dysfunction of the protein. Here we show how a new combined calorimetry approach, using isothermal titration calorimetry to define monomer association energetics together with differential scanning calorimetry to measure total energetics of oligomer unfolding, can be used to analyze homodimeric unmetalated (apo) superoxide dismutase (SOD1) and determine the effects on the stability of structurally diverse mutations associated with amyotrophic lateral sclerosis (ALS). Despite being located throughout the protein, all mutations studied weaken the dimer interface, while concomitantly either decreasing or increasing the marginal stability of the monomer. Analysis of the populations of dimer, monomer, and unfolded monomer under physiological conditions of temperature, pH, and protein concentration shows that all mutations promote the formation of folded monomers. These findings may help rationalize the key roles proposed for monomer forms of SOD1 in neurotoxic aggregation in ALS, as well as roles for other forms of SOD1. Thus, the results obtained here provide a valuable approach for the quantitative analysis of homooligomeric protein stabilities, which can be used to elucidate the natural and aberrant roles of different forms of these proteins and to improve methods for predicting protein stabilities.
T
of monomer stability. Using this methodology, we determine the populations of native dimer, folded monomer intermediate, and unfolded monomer, which are defined for any temperature and protein concentration, for the unmetalated (apo) form of the homodimeric Cu,Zn-superoxide dismutase (SOD1), including multiple mutant forms associated with amyotrophic lateral sclerosis (ALS) (Figure 1). Extensive data support the prominent hypothesis that the misfolding and aggregation of mutant SOD1 are central to ALS disease processes, analogous to the misfolding of other proteins associated with neurodegeneration in Huntington’s, Alzheimer’s, and prion disorders.11−14 A hallmark of ALS is the formation of protein aggregates in motor neurons,15 and SOD1 has been identified as a component in these aggregates.16−18 To date, more than 170 predominantly missense mutations distributed throughout the protein have been linked to ALS (http://alsod. iop.kcl.ac.uk/home.aspx). The mutations confer a toxic gain of function to SOD1; however, the mechanisms of toxicity remain
he structures of natural proteins are most commonly homooligomers, homodimers in particular.1 Thus, the principles governing oligomeric protein folding and misfolding are of widespread importance, in both natural and disease processes, as well as in protein engineering and design.2 A general strategy for elucidating the folding and misfolding pathways of proteins is to measure the relative stabilities and, hence, the populations of different states along the folding pathway. Calorimetry is a powerful tool for analyzing folding by directly measuring the energetics of different states, and recent advances in instrumentation have improved the sensitivity of calorimeters such that increasingly small changes in heat can now be measured accurately.3 Isothermal titration calorimetry (ITC) is widely used for measuring the binding by proteins of ligands such as small molecules, peptides, or nucleic acids, as well as other proteins; to date, ITC has been used very little, however, for measuring homooligomeric protein−protein interactions.4,5 Differential scanning calorimetry (DSC) is a complementary method commonly used to determine the global stability of proteins in monomer, ligand-bound, or oligomeric states.6−10 Here, we demonstrate a new approach, combining ITC and DSC data, to define the constituent components of total homooligomeric protein stability, i.e., the energetics of monomer association and © 2015 American Chemical Society
Received: November 2, 2015 Revised: December 18, 2015 Published: December 29, 2015 519
DOI: 10.1021/acs.biochem.5b01187 Biochemistry 2016, 55, 519−533
Article
Biochemistry
by absorbance at 280 nm using an extinction coefficient of 10800 M−1 cm−1 for the SOD1 dimer.34 Differential Scanning Calorimetry. DSC experiments were performed as described previously,32 using a LLC VP DSC instrument (MicroCal Inc., Malvern Instruments Ltd.). Samples contained ∼0.05−3 mg mL−1 SOD1 in 20 mM HEPES (pH 7.8). Buffer versus buffer thermograms were subtracted from protein versus buffer thermograms, and data were normalized to units of calories per gram per degree Celsius before fitting. As in previous studies,26,32 the reversibility of unfolding was assessed by heating the sample to the end of the unfolding endotherm, followed immediately by cooling and rescanning. Reversibility was quantified by calculating the area under the unfolding endotherm for the first scan and second scan. Samples were scanned at a rate of 1 °C min−1. Isothermal Titration Calorimetry. ITC measurements of the heat associated with dissociation of apo SOD1 dimer (N2) to folded monomers (M), N2 ↔ 2M, according to the dissociation constant, Kd,N2↔2M, were performed and analyzed as described previously, at 37 °C where the monomer is stable (i.e., predominantly folded) using a Microcal Isothermal Titration Calorimetry 200 instrument (Microcal Inc., Northhampton, MA).27,35,36 Briefly, a concentrated apo SOD1 solution [0.35−1 mM dimer in 20 mM HEPES (pH 7.8)] was diluted into an identical buffer (flow through obtained from ultrafiltration concentration of the protein solution) in the ITC reaction cell in successive small volume (0.3−0.5 μL) injections. The heat associated with each injection (Figure S1) was determined by integrating the power versus time trace; the heats were fit to a dimer dissociation model (see the legend of Figure S1). Analysis of Calorimetry Data. Calorimetry data were analyzed using reversible two-state or three-state unfolding models with various combinations of parameters fit or set to defined values (see Table S1). The two-state model describes the transition between native folded dimer (N2) and unfolded monomer (U), N2 ↔ 2U, while the three-state model also includes formation of a folded monomer intermediate (M), N2 ↔ 2M ↔ 2U.2,9,37,38 For the two-state model, individual thermograms were fit using Origin 5.0 (Microcal Inc.) as described previously for apo SOD1 to eq 1:26,32
Figure 1. Structural features of apo SOD1 dimer. Each monomer forms a Greek key β-barrel (strands are numbered) and contains a single disulfide bond (C57−C146, orange). In the absence of bound Cu and Zn, the zinc binding loops (loop IV, green) and electrostatic loops (loop VII, blue) are disordered, and the protein is more flexible than the metalbound form.83,99 The sites of ALS-associated mutations characterized herein are shown as red sticks (Protein Data Bank entry 1HL4).
unclear.12,13,19 Attempts to identify relationships between the effects of the mutations on SOD1 stability and ALS characteristics have shown that these effects are complex and not correlated with disease characteristics in a simple way.13,20,21 Previous studies have suggested that an increased population of apo monomer forms of SOD1 may promote toxic aggregation, implicating metal loss and dimer dissociation as being critical for SOD1 aggregation.22−25 Notably, we found that diverse mutations throughout the protein weaken the dimer interface of apo SOD1.26,27 Here we extend these studies and show how global fitting of DSC data as a function of protein concentration, fixing parameters defining dimer dissociation to those measured by ITC, reveals that mutations have varying effects on the stability and population of folded monomer. An advantage of this approach is that it can define populations at physiological temperatures and protein concentrations, which are difficult or impossible to measure directly using other methods. In most cases, mutations are moderately to highly destabilizing with respect to the monomer, but some actually increase monomer stability. The global analysis also shows how multiple modes of stability changes in mutant SOD1 can have the common effect, under physiologically relevant solution conditions, of increasing the population of monomers that may be involved in ALS. The method is broadly applicable to homooligomers and can provide quantitative data, for different solution conditions or mutations, to determine how the populations of different states may impact natural, regulated, and disease processes.2,28 In summary, the combined use of ITC and DSC provides valuable thermodynamic data, both for improving methods to model protein− protein binding, which still have limited accuracy,29 and for understanding the thermodynamics underlying protein function and misfolding in disease.
Cp = (A + Bt )(1 − α) + (E + Ft )α + ×
α(1 − α) 2−α
β Δhcal 2(t ) RT 2 (1)
where Cp is the total specific heat capacity at temperature t (in degrees Celsius), A (E) and B (F) are the intercept and slope of the folded (unfolded) baseline, respectively, R is the universal gas constant, β is the ratio of the van’t Hoff enthalpy (ΔHvH) to calorimetric enthalpy (ΔHcal) multiplied by the molecular weight (MW) of the SOD1 dimer, Δhcal(t) is the specific calorimetric enthalpy (ΔHcal divided by the molecular weight) of unfolding at t, α is the extent of the unfolding reaction, and T0.5 (or t0.5) is the temperature in kelvin (or °C) at which half of the protein is unfolded (α = 0.5; described further in Supplementary Methods). The number of subunits in the cooperative unfolding transition, also known as the molecularity, n, was analyzed using eq 2:32,39
■
MATERIALS AND METHODS Recombinant SOD1 Expression and Purification. Human pWT SOD1 (pseudo wild-type with nonconserved Cys 6 and Cys 111 substituted with Ala and Ser, respectively)20,30,31 and mutant SOD1 were expressed using plasmid pHSOD1ASlacI142 in Escherichia coli (strain QC779) and then purified and demetalated by dialysis against EDTA as described previously.32,33 Protein concentration was determined 520
DOI: 10.1021/acs.biochem.5b01187 Biochemistry 2016, 55, 519−533
Article
Biochemistry Table 1. Thermodynamic Parameters for Dimer Two-State Unfolding of Apo SOD1a SOD1 variant b
pWT pWTb pWTb pWTb pWTb pWTb pWTb pWTb pWTb pWTb avgd A4S A4S A4S A4S A4S avgd A4T A4T A4T A4T A4T avgd A4V A4V A4V A4V A4V A4V avgd E100G E100G E100G avgd G37R G37R G37R G37R G37R avgd G93Ab G93Ab G93Ab G93Ab avgd G93Rb G93Rb avgd G93Sb G93Sb G93Sb avgd H43R H43R H43R avgd H46R H46R H46R H46R
[SOD1] (mg mL−1) 0.05 0.20 0.21 0.27 0.44 0.73 1.42 1.50 2.99 3.00 − 0.06 0.13 0.19 0.57 0.63 − 0.08 0.18 0.20 0.50 1.00 − 0.20 0.30 0.40 0.50 1.00 1.95 − 0.20 0.50 1.20 − 0.11 0.24 0.40 0.98 1.92 − 0.10 0.12 0.50 1.00 − 0.21 0.40 − 0.25 0.29 0.40 − 0.21 0.23 0.39 − 0.08 0.17 0.32 0.39
e
t0.5 (°C) 57.8 58.3 58.4 57.8 59.1 58.8 60.0 60.3 61.3 61.9 − 44.7 45.8 46.0 47.8 48.3 − 40.6 42.9 43.0 44.2 45.9 − 48.8 48.5 49.3 48.4 48.9 50.1 − 49.5 51.0 53.0 − 48.5 49.5 51.2 51.1 52.2 − 45.6 46.6 49.0 49.1 − 47.3 48.7 − 49.6 50.0 50.8 − 46.9 47.3 48.1 − 60.4 60.9 61.2 61.0
± ± ± ± ± ± ± ± ± ±
0.6 0.1 0.1 0.0 0.2 0.0 0.0 0.0 0.0 0.0
± ± ± ± ±
0.1 0.4 0.3 0.0 0.0
± ± ± ± ±
0.2 0.1 0.4 0.3 0.1
± ± ± ± ± ±
1.6 0.1 0.1 0.1 0.2 0.0
± 0.0 ± 0.0 ± 0.0 ± ± ± ± ±
0.1 0.0 0.0 0.0 0.0
± ± ± ±
0.2 0.1 0.0 0.1
± 0.0 ± 0.0 ± 0.1 ± 0.0 ± 0.0 ± 0.1 ± 0.0 ± 0.0 ± ± ± ±
0.3 0.1 0.0 0.0
ΔCp,N2↔2Uf [kcal (mol of dimer)−1 °C−1] 3.71 2.95 3.74 3.85 4.25 4.30 4.38 5.11 5.10 3.93 4.13 2.81 1.67 2.85 2.53 4.45 2.86 6.77 3.98 1.20 2.34 2.94 3.45 −0.22 1.73 0.54 2.50 2.21 1.88 1.44 1.82 2.18 2.40 2.13 3.01 2.64 3.81 3.59 3.16 3.24 2.99 3.60 3.82 3.25 3.42 5.73 4.36 5.04 2.60 3.48 2.13 2.74 3.60 2.80 3.07 3.16 4.81 2.56 3.31 0.74
± 0.65
± 1.00
± 2.11
± 0.75
± 0.29
± 0.47
± 0.37
± 0.97
± 0.69
± 0.41
ΔHvH(t0.5)e [kcal (mol of dimer)−1] 106.5 112.8 116.8 115.4 137.6 144.1 163.6 164.4 178.5 183.3 − 132.2 156.3 140.8 155.4 159.0 − 76.0 116.3 127.7 133.9 145.7 − 81.5 82.4 86.7 77.5 93.9 120.6 − 184.7 163.6 156.9 − 168.6 171.3 173.5 180.8 182.9 − 119.8 123.1 147.2 149.7 − 123.7 144.0 − 146.4 149.8 158.6 − 123.1 120.6 128.8 − 90.0 107.8 120.2 128.6 521
± ± ± ± ± ± ± ± ± ±
18.9 6.7 9.1 2.7 9.4 2.5 2.4 4.0 1.4 1.9
± ± ± ± ±
5.8 11.6 8.1 0.9 1.0
± ± ± ± ±
6.0 5.7 18.9 9.4 2.1
± ± ± ± ± ±
17.2 1.9 3.5 1.6 3.5 2.2
± 10.2 ± 0.9 ± 0.8 ± ± ± ± ±
2.5 1.5 0.4 1.0 0.5
± ± ± ±
8.3 8.3 0.9 2.1
± 2.7 ± 2.4 ± 7.8 ± 4.1 ± 2.5 ± 2.5 ± 2.0 ± 1.7 ± ± ± ±
11.4 2.0 2.1 2.1
ΔGN2↔2Uf,g [kcal (mol of dimer)−1] tavg 10.1 9.4 9.5 9.1 9.7 9.4 9.9 10.0 10.4 10.8 9.8 5.1 4.7 4.9 4.9 5.0 4.9 4.7 3.9 3.6 3.5 3.7 3.9 6.6 6.3 6.3 6.0 5.5 5.4 6.0 6.4 6.6 7.2 6.7 6.2 6.2 6.8 6.2 6.3 6.3 5.4 5.6 5.6 5.1 5.4 5.6 5.6 5.6 6.3 6.4 6.5 6.4 5.5 5.6 5.5 5.5 9.9 10.0 10.0 10.1
± 0.5
± 0.1
± 0.5
± 0.4
± 0.4
± 0.3
± 0.2
± 0.0
± 0.1
± 0.1
ΔGN2↔2Uf [kcal (mol of dimer)−1] 37 °C 12.3 11.9 12.1 11.8 13.1 13.2 14.3 14.4 15.3 15.8 13.4 10.6 11.1 10.5 10.9 11.2 10.9 8.3 9.0 9.2 9.1 9.6 9.0 9.2 9.0 9.0 8.5 8.7 9.5 9.0 12.1 12.4 13.6 12.7 12.7 12.7 13.1 12.8 12.9 12.8 10.2 10.4 11.1 10.8 10.6 10.3 11.0 10.7 11.6 11.8 12.2 11.9 10.2 10.2 10.3 10.2 11.0 11.8 12.4 12.9
± 1.4
± 0.3
± 0.5
± 0.4
± 0.8
± 0.2
± 0.4
± 0.5
± 0.3
± 0.1
nh 1.60 1.64 1.66 1.65 1.78 1.82 1.93 1.93 2.01 2.04 − 1.98 2.16 2.04 2.15 2.18 − 1.77 2.18 2.29 2.36 2.48 − 1.43 1.43 1.46 1.41 1.50 1.64 − 2.19 2.24 2.40 − 2.12 2.14 2.16 2.21 2.22 − 1.96 1.99 2.18 2.19 − 2.50 2.74 − 2.80 2.84 2.94 − 2.12 2.10 2.18 − 1.17 1.21 1.23 1.25
DOI: 10.1021/acs.biochem.5b01187 Biochemistry 2016, 55, 519−533
Article
Biochemistry Table 1. continued SOD1 variant
[SOD1] (mg mL−1)
H46R avgd I113T I113T I113T I113Tc I113Tc I113Tc I113T avgd V148G V148G V148G V148G V148G V148G avgd V148I V148I V148I V148I V148I V148I avgd
0.76 − 0.08 0.10 0.15 0.40 0.50 0.75 1.20 − 0.12 0.23 0.29 0.53 0.92 1.56 − 0.11 0.26 0.49 0.53 0.83 2.33 −
e
t0.5 (°C) 61.1 − 43.4 45.3 44.9 46.0 47.1 46.4 47.8 − 46.1 48.3 49.8 48.2 46.8 47.8 − 57.7 59.2 59.6 59.6 60.3 61.2 −
± 0.0 ± ± ± ± ± ± ±
0.5 1.0 0.4 0.1 0.1 0.1 0.0
± ± ± ± ± ±
0.2 0.1 0.0 0.0 0.0 0.3
± ± ± ± ± ±
0.1 0.1 0.1 0.0 0.0 0.0
ΔCp,N2↔2Uf [kcal (mol of dimer)−1 °C−1] 5.18 3.32 4.91 1.86 3.03 2.44 2.79 4.75 3.35 3.30 5.55 1.26 −2.00 0.07 2.47 −2.44 0.82 6.67 3.01 6.37 3.75 3.38 3.99 4.53
ΔHvH(t0.5)e [kcal (mol of dimer)−1] 139.2 − 68.7 105.3 94.4 106.2 129.0 118.0 138.5 − 76.6 89.1 104.6 91.5 81.7 99.0 − 108.8 137.7 141.4 152.1 166.9 182.6 −
± 1.80
± 1.14
± 3.0
± 1.58
± 1.8 ± ± ± ± ± ± ±
16.8 18.1 9.0 4.0 4.1 3.9 2.0
± ± ± ± ± ±
5.6 0.7 2.0 2.1 2.9 1.0
± ± ± ± ± ±
6.7 3.0 5.9 1.7 1.8 1.1
ΔGN2↔2Uf,g [kcal (mol of dimer)−1] tavg 9.8 10.0 5.8 5.6 5.4 4.9 4.9 4.5 4.6 5.1 6.2 6.3 6.6 5.8 5.1 4.8 5.8 9.6 10.1 9.9 10.1 10.5 10.6 10.1
ΔGN2↔2Uf [kcal (mol of dimer)−1] 37 °C
± 0.1
± 0.5
± 0.7
± 0.4
12.5 12.1 8.7 9.8 9.1 9.1 9.9 9.1 9.9 9.4 9.0 9.3 10.0 8.9 8.0 8.4 8.8 12.0 13.5 13.4 14.1 15.0 15.7 13.9
± 0.7
± 0.5
± 0.5
± 1.3
nh 1.27 − 1.53 1.82 1.73 1.82 2.00 1.91 2.07 − 0.71 0.59 0.64 0.66 0.73 0.70 − 1.56 1.71 1.73 1.78 1.86 1.94 −
Data for individual scans were fit to eq 1 as described in Materials and Methods. bData reported previously.32 cData reported previously.26 dValues are averages and standard deviations. eError estimates for individual thermograms are from the fitting program. fUncertainties were estimated from the standard deviations in the values obtained from the individual thermograms. gtavg is 51.2 °C, the average value of t0.5 for apo SOD1 variants. h Molecularity values determined using eq 2 and Figure S2, as described in Materials and Methods. a
n=
−ΔH vH +1 slope × R
as α1 and α2, respectively (described further in Supplementary Methods); C and D are the intercept and slope, respectively, of the monomer intermediate baseline. Because of the additional parameters in the three-state model compared to the two-state model, fitting individual thermograms to eq 3 resulted in high uncertainties in the fitted values. Accordingly, multiple data sets were fit globally (Matlab R2013b, The MathWorks Inc.) using shared parameters and using experimentally determined values of ΔCp and the common assumption that these are temperatureindependent,40 which has been shown to be reasonable over limited temperature ranges as used here.32,41,42 Parameters describing the first transition were fixed to those determined by ITC (described in detail in Supplementary Methods). In addition, β1 and β2 were set equal to each other and allowed to vary together, as done previously for two-state fits;9,26,32 this allows for experimental uncertainties in measured protein concentrations.32 Calculation of Thermodynamic Parameters from TwoState and Three-State Fitted Parameters. Values for the Gibbs free energy [ΔG(T)], enthalpy [ΔH(T)], and entropy [ΔS(T)] at a given temperature were calculated from fitted DSC parameters according to the following equations:
(2)
where ΔHvH was obtained from the two-state fit for a given thermogram and slope is the slope of a plot of ln Pdimer versus 1/ T0.5, where Pdimer is the protein concentration in M dimer.32,39 For each SOD1 variant, slope was determined using the data for all protein concentrations. Representative plots of ln Pdimer versus 1/T0.5 are shown in Figure S2. For the dimer three-state model with a monomer intermediate, the data were fit to eq 3:37 ⎡ β Δhcal,N ↔ 2M(t ) + α2β Δhcal,M ↔ U(t ) ⎤ 1 2 2 ⎥ Cp = ⎢ ⎢⎣ ⎥⎦ RT 2 Δhcal,N2 ↔ 2M(t )
α1(1 − α1) 2 − α1
⎧⎡ β Δh ⎤ ⎪ cal,N2 ↔ 2M (t ) + α2β2 Δhcal,M ↔ U (t ) ⎥ ⎨⎢ 1 +⎪ 2 ⎥⎦ RT ⎩⎢⎣ ⎫ β Δhcal,M ↔ U(t ) ⎪ α1α2(1 − α1) ⎬ + 2 α1α2(1 − α2)⎪ 2 2 − α1 2RT ⎭ Δhcal,M ↔ U(t ) + (1 − α1)(A + Bt ) + α1(1 − α2) (C + Dt ) + α1α2(E + Ft )
(3)
where subscripts 1 and 2 refer to the first transition, N2 ↔ 2M, and second transition, M ↔ U, respectively, both defined per M dimer. The extents of the first and second transitions are shown 522
ΔG(T ) = ΔH(T ) − T ΔS(T )
(4)
ΔH(T ) = ΔH(Tref ) + ΔCp(T − Tref )
(5)
⎛T ⎞ ΔS(T ) = ΔS(Tref ) + ΔCp ln⎜ ⎟ ⎝ Tref ⎠
(6) DOI: 10.1021/acs.biochem.5b01187 Biochemistry 2016, 55, 519−533
Article
Biochemistry ΔS(Tref ) =
ΔH(Tref ) − ΔG(Tref ) Tref
ΔG(Tref ) = −RT ln K (Tref )
(7) (8)
where Tref is the reference temperature, which for DSC two-state fitting was T0.5 (t0.5) in kelvin (degrees Celsius) and Kref = 2Pdimer; for DSC three-state fitting, the first dimer dissociation step, Tref (tref) was fixed to 310.1 K (37 °C) and Kref = Kd,N2↔2M determined from ITC, and for monomer unfolding, Tref was T0.5 (t0.5) and Kref = 1. ΔH(T), ΔS(T), and ΔG(T) are the change in enthalpy, entropy, and free energy, respectively, as a function of temperature calculated knowing the Kref, the fitted or fixed Tref, the fitted ΔH(Tref), and the experimentally determined values of ΔCp (outlined in detail in Supplementary Methods). Thermodynamic values were calculated at 51.2 °C (the average of the t0.5 values in Table 1, tavg) and at physiological temperature, 37 °C. Uncertainties due to extrapolation are decreased when considering differences in stability, ΔΔGN2↔2U = (ΔGN2↔2U,pWT − ΔGN2↔2U,mut), for values obtained with the same ΔCp.32,43 The values of ΔG were used to calculate the fractions of different species (N2, M, and U) as described in Supplementary Methods. Equilibrium Urea Denaturation Curves at 37 °C. Denaturation curve experiments were performed at 37 °C in 1 mM EDTA and 20 mM HEPES (pH 7.8). Stock solutions of apo pWT were diluted 10-fold into different concentrations of urea and equilibrated at 37 °C for 10 h. Fluorescence was measured using a Fluorolog3-22 spectrofluorometer (Horiba-Jobin-Yvon Spex Inc.) equipped with a thermostated cuvette holder with excitation and emission wavelengths of 282 and 360 nm, respectively. Thermodynamic parameters were obtained by globally fitting denaturation curves at 0.20 μM (0.0063 mg/mL), 0.80 μM (0.025 mg/mL), 3 μM (0.095 mg/mL), 10 μM (0.31 mg/mL), and 25 μM (0.79 mg/mL) dimer using a three-state dimer with a monomer intermediate model using Microcal Origin 6.0 as described previously.44 The folded and unfolded baselines were determined by linear regression and were fixed during the global fitting. The fluorescence of the intermediate was determined by systematically changing the magnitude of the fluorescence to identify the value that corresponded to the lowest χ2, similar to previous denaturation curve analyses of apo pWT in guanidinium hydrochloride.26 In the global fitting, for each intermediate fluorescence value, four parameters (ΔGN2↔2M, ΔGM↔U, and the dependence of the dimer interface and monomer stability on urea concentration, mN2↔2M and mM↔U, respectively) were shared across all protein concentrations and were allowed to float. The lowest χ2 value for the fit was obtained with intermediate fluorescence and slope corresponding to 30% of the total amplitude of the transition.
Figure 2. Highly reversible DSC scans for apo SOD1 variants fit a dimer two-state unfolding model. (A) Consecutive scans for apo variants heated at a rate of 1 °C min−1 (black solid lines), cooled at a rate of 1 °C min−1, and reheated (gray dotted lines). The reversibility, determined by integrating the area under the endotherm (ΔHcal) for the first scan and the rescan, after subtraction of the baselines, is 95% or greater for all mutants except for A4V, for which it is ∼85%. (B) Representative DSC data (black solid lines) and fits (gray dashed lines) to a dimer two-state unfolding model (N2 ↔ 2U) using eq 1. Parameters for all fits are summarized in Table 1. Data sets with buffer−buffer scans subtracted and normalized for protein concentration (0.12−1.25 mg mL−1) are offset for ease of comparison and arranged from top to bottom in order of increasing t0.5.
thermograms were not significantly affected by scan rate (ranging from 0.75 to 1.5 °C min−1), which is also an indication of equilibrium unfolding behavior.32,45 Apo pWT SOD1 has a melting temperature, t0.5 (temperature at which half of the protein is unfolded), of ∼59 °C,32 while the mutants generally have t0.5 values ranging from 10 to 15 °C lower than that of pWT, indicating marked destabilization (Table 1 and Figure 2). Notable exceptions are H46R and V148I, which are slightly more stable than pWT, with t0.5 values that are up to ∼2 °C higher. Fitting of Apo SOD1 Thermal Unfolding to the Dimer Two-State Model. The DSC data for the mutants were fit initially to a two-state transition between the native folded dimer and unfolded monomers (N2 ↔ 2U), as done previously for pWT and other ALS-associated mutants.26,32 At higher protein concentrations, most of the data are well fit by the two-state model (Figure 2B and Table 1), and the data follow trends
■
RESULTS DSC Reveals the Total Stability of Apo SOD1 Dimers Is Often Decreased but Can Also Be Increased by Disease Mutations. DSC was used to assess the energetics of total unfolding of the apo SOD1 dimer to unfolded monomers. Thermodynamic analysis requires reversibility of thermal unfolding; as has been observed previously for other apo SOD1 variants,26,32 via scanning to the end of the unfolding endotherm, cooling, and rescanning (Figure 2A), the unfolding of the mutants studied here is highly reversible. The DSC 523
DOI: 10.1021/acs.biochem.5b01187 Biochemistry 2016, 55, 519−533
Article
Biochemistry
Figure 3. Predicted t0.5 values for thermal unfolding of apo SOD1 variants. The top left panel illustrates how predicted values for t0.5 were determined from the intersection of −RT ln(2Pdimer) vs t (gray lines) and ΔG vs t (black lines) for pWT (solid black line), A4S (dashed black line), and A4T (dotted black line).32,46 The line for 0.5 mg mL−1 protein (solid gray line) intersects each ΔG vs t plot at a temperature lower than the line generated for 5 mg mL−1 (dotted gray line). The ΔG curves were calculated as described in Materials and Methods using eqs 4−8, and average fitted parameters for twostate dimer unfolding. Other panels show the predicted t0.5 values (solid gray lines) from these calculations plotted as a function of protein concentration and compared to the fitted t0.5 values (■) from Table 1.
agreement with predicted t0.5 values, while the values for A4V, H46R, V148G, and to some extent G37R change less with protein concentration than predicted.26,32 Overall, the results show that the two-state model is appropriate for many apo variants at higher protein concentrations, but at low protein concentrations, and for some mutants at higher concentrations, a three-state unfolding model is required. Fitting to the Dimer Three-State with Monomer Intermediate Model. Accordingly, the DSC data for all apo SOD1 variants were also fit to a dimer three-state unfolding model with a monomer intermediate (N2 ↔ 2M ↔ 2U), by which the thermodynamics of both dimer dissociation and monomer unfolding may be quantified. To define the larger number of parameters in this model, all data for a given variant were globally fit, fixing the parameters for dimer dissociation, Kd,N2↔2M and Δhcal,N2↔2M, at 37 °C to the values measured by ITC, and the values for ΔCp to experimental values measured independently (see Materials and Methods and Supplementary Methods). The data for all mutants are generally well fit using this method (Figure 4, Table 2, and Table S2), and the results are minimally affected by potential variations in ΔCp or aggregation (see Supplementary Results). We used several fitting approaches to define the uncertainties in monomer and total stability. The
(Figure 3) similar to those reported previously for apo SOD1 variants.26,32 We further explored the extent of formation of the monomer intermediate by calculating for each thermogram the number of subunits for the observed unfolding transition, n, known as the molecularity (Materials and Methods, eq 2, and Figure S2).32,39 Values of n increase from lower values at lower protein concentrations to ∼2 at higher protein concentrations (Table 1), the value expected for two-state unfolding of a dimeric protein; thus, for most variants, there is evidence of an increased level of formation of monomer at low protein concentrations, as may be expected on the basis of mass action. For A4V, H46R, and V148G, the values of n are considerably lower than 2 for all protein concentrations and the unfolding endotherms occur over a relatively broader temperature range (Figure 4), indicating more pronounced non-two-state behavior (Figure S3). The protein concentration dependence of t0.5 can further define when monomer is being populated. The expected concentration dependence of t0.5 was determined from the intersection of plots of ΔG versus t and −RT ln 2Pdimer versus t, where R is the universal gas constant, Pdimer is the total dimer concentration in moles per liter, T is the temperature in kelvin, and t is the temperature in degrees Celsius1,3,57 (Figure 3). For pWT, A4S, A4T, and V148I, the fitted t0.5 values show close 524
DOI: 10.1021/acs.biochem.5b01187 Biochemistry 2016, 55, 519−533
Article
Biochemistry
and V148G) in which the latter method resulted in low uncertainty in the fitted values. Thus, the approach of fixing parameters defining dimer dissociation to those obtained by ITC works well. For comparison, the mutants exhibiting the most pronounced three-state behavior (A4V, H46R, and V148G) were also fit to a monomer two-state unfolding model (see Supplementary Methods). This model could in principle be applicable if the transition temperatures for each step in the three-state mechanism are well separated and if the majority of the heat of unfolding is associated with monomer unfolding.46,47 These twostate fits returned ΔHvH values that increase significantly with protein concentration (Table S3), which is inconsistent with unimolecular two-state monomer unfolding9 and indicates at least some association is occurring. Thus, generally, a three-state model is most appropriate for describing apo SOD1 thermal unfolding. Equilibrium Urea Denaturation of Apo pWT at 37 °C. To obtain an independent measure of stability, urea denaturation curves were also determined as a function of protein concentration for pWT at 37 °C (Figure 5). We used chemical denaturation previously for pWT at 25 °C to measure the total stability, ΔGN2↔2U, dissected into constituent ΔGN2↔2M and ΔGM↔U values;26 these analyses cannot be performed for mutants, however, because of their lower stability. The stability values measured for pWT at 37 °C (Table 2) are consistent with values obtained previously by chemical denaturation at lower temperatures,26,48−50 and with values determined at 37 °C by calorimetry (Tables 2 and 3 and Table S2). Notably, chemical denaturation shows the dimer interface stability is considerably weakened at increased temperatures (Table 2 and ref 26), in agreement with ITC measurements,51 and resulting in significant population of monomer at physiologically relevant temperature and pH (see below). Comparison of Two-State and Three-State Fits Shows ALS Mutations Have Diverse Effects on Apo SOD1 Stability. For total unfolding, the two-state fits with n values of ∼2 (Table 1) generally have ΔGN2↔2U values similar to the values of ΔGN2↔2M↔2U from the global three-state fits (Figure 6A and Table 3). This agreement indicates that the population of monomer in the DSC experiments is on the whole relatively low for pWT, A4S, A4T, E100G, G37R, G93S, G93A, H43R, and I113T, so two-state fitting gives a reasonably accurate measure of total stability (Figure 6A). On the other hand, the advantage of three-state fitting is that the total stability can be dissected into constituent dimer interface and monomer stabilities. The results show that most of the stability of apo SOD1 is derived from dimer formation and that the monomer has relatively low stability, which may be decreased or increased upon mutation (Table 2). The monomer stabilities for the mutants are generally consistent with limited results obtained using chemical denaturation.26,49,50,52,53 Notably, the monomer stabilities are strongly correlated with the stability of mutants in the disulfide-reduced apo form of SOD1 that is monomeric20 (Figure 6B). When the population of monomer is increased, as occurs for DSC scans at lower protein concentrations, or when mutations substantially destabilize the dimer interface (A4V, G93R, and V148G) or increase monomer stability (H46R and V148I), there are larger discrepancies in the total stability determined by the dimer two-state and three-state fits (Figure 6A). In these cases, fitting the data to a two-state model gives less accurate values.
Figure 4. Global fitting of apo SOD1 DSC data to the three-state model. Representative global fits of thermograms to the dimer three-state model with a monomer intermediate are shown. The data are shown as black solid lines, and fits are shown as gray dashed lines. In each panel, scans are offset for the sake of clarity, from the lowest concentration at the bottom to the highest at the top. The protein concentrations for thermograms are as follows: 0.20, 0.21, 0.40, 0.44, 0.73, 1.50, and 3.00 mg mL−1 for pWT; 0.08, 0.17, 0.32, 0.39, and 0.76 mg mL−1 for H46R; 0.11, 0.26, 0.49, 0.53, 0.83, and 2.33 mg mL−1 for V148I; 0.20, 0.30, 0.40, 0.50, 1.00, and 1.95 mg mL−1 for A4V; 0.21, 0.23, and 0.39 mg mL−1 for H43R; 0.12, 0.23, 0.29, 0.53, 0.92, and 1.56 mg mL−1 for V148G; 0.08, 0.10, 0.15, 0.40, 0.50, 0.75, and 1.20 mg mL−1 for I113T; 0.08, 0.18, 0.20, 0.50, and 1.00 mg mL−1 for A4T. The values for fitted parameters from the global fits are summarized in Table 2 and Table S1.
fitting results further support the idea that treating ΔCp as a constant is reasonable, and changes in ΔCp have little impact on monomer stability. We also confirmed, consistent with our previous results,32 that possible aggregation at high temperatures has little effect on fitted values by varying the amounts of fitted data beyond the peak of the unfolding endotherm. As an additional control, mutants with more than three data sets were also globally fit without constraints from the ITC data (i.e., with Δhcal,N2↔2M and Kd,N2↔2M at 37 °C also globally shared) and β1 and β2 fixed to the molecular weight of the dimer. In these fits, values of Kd,N2↔2M, and thus ΔGN2↔2M, converged to values comparable to those measured by ITC (Table S2), although the uncertainty in all fitted parameters increased. Notably, the agreement between fitted and experimentally determined Δhcal,N2↔2M and Kd,N2↔2M is good for the cases (I113T, A4V, 525
DOI: 10.1021/acs.biochem.5b01187 Biochemistry 2016, 55, 519−533
Article
Biochemistry
Table 2. Summary of Thermodynamic Parameters for Apo SOD1s Determined from Global Three-State Fits of DSC Data SOD1 varianta pWTc,h pWT V148I G93S H46R E100G G37R H43R G93A I113T A4T A4S G93R A4V V148G
ΔHN2↔2M [kcal (mol ΔGN2↔2M [kcal (mol of dimer)−1] 37 °Cb of dimer)−1] 37 °Cb nai (30.8 ± 8.8) (11.4 ± 2.2) (17.6 ± 4.6) (16.2 ± 4.4) (16.0 ± 4.8) (7.8 ± 1.8) (23.0 ± 1.4) (14.0 ± 2.0) (30.2 ± 2.5) (39.2 ± 3.8) (9.0 ± 2.6) (45.6 ± 1.8) (37.2 ± 3.8) (50.6 ± 1.4)
10.2 ± 0.7 (10.3 ± 0.5) (8.9 ± 0.2) (8.4 ± 0.3) (8.4 ± 0.4) (8.0 ± 0.4) (7.6 ± 0.2) (7.5 ± 0.0) (7.2 ± 0.3) (7.1 ± 0.2) (7.1 ± 0.2) (7.0 ± 0.0) (6.7 ± 0.1) (6.4 ± 0.3) (5.9 ± 0.3)
t0.5,M↔U (°C)c
ΔHM↔U [kcal (mol of monomer)−1] tavgc,d
ΔGM↔U [kcal (mol of monomer)−1] 37 °Ce,f
ΔGM↔U [kcal (mol of monomer)−1] tavgd,f
ΔΔGM↔U [kcal (mol of monomer)−1] tavgd,g
nai 59.5 ± 0.9 60.1 ± 0.2 49.2 ± 1.1 62.5 ± 0.1 48.0 ± 0.7 50.3 ± 0.1 47.6 ± 0.4 47.4 ± 0.3 46.7 ± 0.2 43.6 ± 0.4 46.3 ± 0.4 49.1 ± 0.2 50.9 ± 0.2 48.6 ± 0.0
nai 44.0 ± 2.0 84.0 ± 1.2 58.3 ± 3.7 71.2 ± 1.1 53.3 ± 8.4 84.1 ± 2.8 59.5 ± 5.2 58.9 ± 4.7 53.2 ± 3.7 48.9 ± 9.6 61.1 ± 10.0 87.8 ± 7.7 59.2 ± 3.4 67.8 ± 0.5
3.4 ± 0.5 2.8 ± 0.6 5.7 ± 0.8 2.0 ± 0.3 5.3 ± 0.4 1.6 ± 0.3 3.2 ± 0.8 1.7 ± 0.1 1.7 ± 0.1 1.4 ± 0.2 0.8 ± 0.3 1.5 ± 0.3 3.1 ± 0.1 2.3 ± 0.1 2.2 ± 0.1
nai 1.2 ± 0.4 2.3 ± 0.4 −0.4 ± 0.2 2.6 ± 0.2 −0.5 ± 0.1 −0.2 ± 0.0 −0.7 ± 0.0 −0.7 ± 0.1 −0.7 ± 0.1 −1.1 ± 0.1 −0.9 ± 0.2 −0.6 ± 0.0 −0.1 ± 0.1 −0.5 ± 0.0
nai nai −1.2 1.5 −1.4 1.7 1.4 1.8 1.9 1.9 2.3 2.1 1.7 1.2 1.7
a The scans at different protein concentrations used in the global fitting correspond to those shown in Figure 4 and otherwise correspond to concentrations listed in Table 1. bThermodynamics of dimer dissociation in parentheses were measured by ITC27 and fixed in the DSC three-state fits. cEstimates of uncertainties from the fitting program. dtavg is 51.2 °C, the average of all t0.5 values obtained from the two-state fits (Table 1). e ΔGM↔U values calculated at physiological temperature. fError estimates were determined from the differences in values obtained from fits allowing ΔHvH/ΔHcal to deviate from unity, compared to values obtained when ΔHvH and ΔHcal were set equal, and values obtained using a higher ΔCp,N2↔2M [2.2 kcal (mol of dimer)−1 °C−1] (Table S2). gΔΔG = ΔGpWT − ΔGmutant. A positive value indicates a lower stability of the mutant relative to that of pWT; values are calculated at tavg. hΔGN2↔2M and ΔGM↔U determined by globally fitting urea denaturation curves at 37 °C to a three-state model with a monomer intermediate. iNot applicable.
of dimer interface and monomer stabilities to changes in total stability also vary markedly. All the mutations weaken dimerization, although to different extents, while the effects on monomer stability range from being significantly stabilizing for H46R and V148I [1.4 and 1.2 kcal (mol of monomer)−1, respectively] to destabilizing for the other mutants [∼1.5−2.3 kcal (mol of monomer)−1]. Populations of N2, M, and U for Apo SOD1 Variants. Knowledge of the three-state thermodynamic parameters allows the fractions of N2, M, and U and the related contributions of the two unfolding steps to the heat measured during DSC scans to be calculated as a function of temperature for any given protein concentration (see Supplementary Methods and Figure 7).28,44 It is noteworthy that folded monomer is not significantly populated for pWT (
