Subscriber access provided by Kaohsiung Medical University
Article
Effect of Mutations on the Global and Site-Specific Stability and Folding of an Elementary Protein Structural Motif Jason K Lai, Ginka S. Kubelka, and Jan Kubelka J. Phys. Chem. B, Just Accepted Manuscript • DOI: 10.1021/acs.jpcb.8b05280 • Publication Date (Web): 09 Jul 2018 Downloaded from http://pubs.acs.org on July 10, 2018
Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.
is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.
Page 1 of 35 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
The Journal of Physical Chemistry
Effect of Mutations on the Global and Site-Specific Stability and Folding of an Elementary Protein Structural Motif Jason K. Lai, Ginka S. Kubelka, Jan Kubelka* Department of Chemistry, University of Wyoming, Laramie, WY, 82071
Author E-mail addresses: Jason K. Lai:
[email protected] Ginka S. Kubelka:
[email protected] Jan Kubelka:
[email protected] * Corresponding author ACS Paragon Plus Environment
The Journal of Physical Chemistry 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Page 2 of 35
Abstract Understanding the folding mechanism of proteins requires detailed knowledge of the roles of individual amino acid residues in stabilization of specific elements and local segments of the native structure. Recently, we have utilized the combination of circular dichroism (CD) and site-specific 13C isotopically-edited infrared spectroscopy (IR) coupled with the Ising-like model for protein folding to map the thermal unfolding at the residue level of a de novo designed helix-turn-helix motif αtα. Here we use the same methodology to study how the sequence of local thermal unfolding is affected by selected mutations introduced into the most and least stable parts of the motif. Seven different mutants of αtα are screened to find substitutions with the most pronounced effects of the overall stability. Subsequently, thermal unfolding of two mutated αtα sequences is studied with site-specific resolution, using four distinct
13
C isotopologues of each. The data are analyzed with the Ising-like model, which
builds on a previous parameterization for the original αtα sequence and tests different ways of incorporating the amino acid substitution. We show that for both more and less stable mutants only the adjustment of all interaction parameters of the model can yield a satisfactory fit to the experimental data. The stabilizing and destabilizing mutation result, respectively, in a similar increase and decrease of the stability of all probed local segments, irrespective of their position with respect to the mutation site. Consequently, the relative order of their unfolding remains essentially unchanged. These results underline the importance of the interconnectivity of the stabilizing interaction network and cooperativity of the protein structure, which is evident even in a small motif with apparently noncooperative, heterogeneous unfolding. Overall, our findings are consistent with the native structure being a dominant factor in determining the folding mechanism, regardless of the details of its overall or local thermodynamic stabilization.
ACS Paragon Plus Environment
Page 3 of 35 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
The Journal of Physical Chemistry
Introduction How specific amino acid residues and their mutations affect the stability of the native protein structure is an important problem of significant interest to a range of scientific fields, from evolutionary biology to medicine.1–5 In biophysics, engineering mutations into proteins has become one of the classic and widely used approaches for studying the mechanism of protein folding.6 These studies rely on the assumption that the amino acid substitution does not cause any significant structural perturbation, but only a change in the thermodynamic stabilities of the underlying states. Native protein structures are generally robust with respect to point mutations, despite their often significant impact on thermodynamic stability.2,7 However, given the close correspondence between the stability of the various partially folded intermediates and the folding mechanism,8,9 it can be intuitively expected that the folding pathways should be quite sensitive to mutations and examples of this behavior exist.10 Moreover, even the unfolded states of proteins often contain residual structure11 and can therefore be expected to be sensitive to sequence perturbations, further adding to the net effects of the mutations on the protein stability and folding. Unfortunately, the details of how mutations perturb the various metastable protein states at intermediate stages of folding are difficult to characterize, particularly since the mutations themselves are often the only available experimental means for their study. An alternative approach to gaining microscopic information about the partly folded states of proteins is through experiments that combine multiple, site-specific structural probes.12–19 In our laboratory, infrared (IR) spectroscopy with
13
C isotopic editing (IE)19 has been used as such site-
specific technique to investigate the unfolding thermodynamics in two similar helix-turn-helix motifs.21-23 IE IR allows probing of local unfolding via the analysis of isotopically shifted amide I’ IR bands that arise from interacting pairs or triplets of
13
C labeled amides. The isotopic labels can be
systematically placed in different locations throughout the studied protein with no perturbation to the structure or stability. Although the interpretation of the amide I’ spectral changes is not completely straightforward,24 with the aid of the microscopic Ising-like model25-27 the residue-level detail of the thermal unfolding can be reconstructed.21 Furthermore, the folding free energy surfaces derived from the optimized Ising-like model give a complete picture of folding, including the kinetic mechanism. Our results have shown that despite the nearly identical structure of the two model helix-turn-helix motifs, the details of their thermal unfolding are very different.21 One, the naturally occurring P22 subdomain unfolds from the helical termini toward the turn, while the de novo designed αtα motif does nearly the opposite. At the same time, both motifs rely predominantly on the tertiary interactions between the hydrophobic faces of the helices for stabilization, as the peptide fragments corresponding
ACS Paragon Plus Environment
The Journal of Physical Chemistry 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Page 4 of 35
to each helix are mostly unstructured in isolation. Perhaps the most interesting, and quite unexpected result was that the Ising-like models were capable to quantitatively explain all global and site-specific experimental data for the both motifs, and capture the pronounced differences in their thermal unfolding behavior, without any residue-specific interaction parameters: a single energy value for all contacts, irrespective of the amino acid type, was sufficient. Here, we use the same methodology to investigate the effects of specific mutations on the overall as well as site-specific stability of one of the model proteins, the αtα. Combination of 13C IE IR spectroscopy with selective amino acid substitutions is potentially very powerful, because it has the ability to selectively probe and compare the local unfolding at multiple regions, particularly those in close proximity to the mutation site and those far away from it. This can highlight the degree of interconnectivity between the individual regions of the protein molecule, which is closely related to the folding cooperativity. As Muñoz and coworkers pointed out, certain measurable degree of cooperativity is expected even for proteins exhibiting a highly non-cooperative behavior.13 A natural question therefore is whether in an apparently non-cooperative folder such as αtα the local mutation impacts the stability only locally, or whether and how it projects onto the whole protein. The former would suggest that folding pathways may be altered by changing the stability of a particular region, while the latter that they should not. The advantage of using microscopic Ising-like model as an integral part of the experimental data analysis is that its natural output is the complete underlying folding free energy landscape from which the folding mechanism can be inferred. On the other hand, treating mutations with Ising-like models is not entirely straightforward, particularly when only one single energy parameter encompasses all inter-residue interactions. We explore several possibilities, from complete re-optimization, to adjusting only the contact energy or, equivalently, effective contact weight of the perturbed residue.26,27 In addition to suggesting the optimal practical means for handling mutations, the comparison of different parameterizations further builds on our recent results in providing additional tests for the Ising-like models and, through their fundamental assumptions, the significance of the native structure and residue-specific interactions in protein folding.20
Experimental Methods Protein Synthesis The protein αtα, its mutants,
13
C isotopically labeled mutants, and two peptide fragments,
corresponding to the two individual α-helices in the original sequence (Table 1), were synthesized by standard FMOC-based solid-phase peptide synthesis techniques on a Tribute automated peptide ACS Paragon Plus Environment
Page 5 of 35 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
The Journal of Physical Chemistry
synthesizer (Protein Technologies, Inc., Tucson, AZ). The selection of suitable mutations, expected to either stabilize or destabilize the protein, was aided by the Rosetta Design software.28 Isotopic labels were introduced through
13
C=O labeled amino acids (Cambridge Isotopes Laboratories, Tewksbury,
MA). Purification of the crude product was performed by reversed-phase HPLC, and the purity of the final product was confirmed by MALDI-TOF mass spectrometry.
CD Spectroscopy To determine the effects of the mutations on the overall stability, the CD of the αtα proteins and its mutants were measured as a function of temperature in H2O 20 mM phosphate buffer (pH 2.3) at a protein concentration of approximately 20 µM. For the selected mutants (E15L and L32V) whose sitespecific unfolding was investigated by isotopically-edited IR spectroscopy (Table 2), along with the two peptide fragments, the CD experiments were also carried out in a D2O 100 mM phosphate buffer to ensure compatibility with the IR experiments. The exact sample concentration was determined spectrophotometrically. The sample was placed in a 1 mm path length quartz cell and the CD spectra were collected every 5oC between 0°C and 85°C (nominally) on a JASCO J-815 CD spectropolarimeter equipped with Peltier device for precise temperature control. The actual sample temperature was determined with a thermocouple inserted into the sample cell. Reversible refolding was confirmed by an additional CD scan at room temperature. For baseline subtraction the CD spectra of just the buffer were collected under identical conditions. IR Spectroscopy IR experiments were conducted in a D2O 100 mM phosphate buffer at pH 2.3 (uncorrected). The complete H/D exchange of labile protons was achieved by repeated dissolution in D2O and lyophilization. Residual trifluoroacetic acid (TFA) from peptide synthesis and purification was removed by adding 0.1 M HCl prior to lyophilizing. The samples were placed in a custom-built transmission IR cell with CaF2 windows separated by a 50 µm Teflon spacers. The IR spectra were collected on a Bruker Tensor 27 FTIR equipped with a RT-DLaTGS detector every 3°C from (nominally) 0°C to 87°C. Temperature was controlled by a liquid bath connected to the spectra acquisition software and the exact sample temperature was calibrated using a thermocouple. At each temperature 512 scans were averaged at a resolution of 4 cm-1. Three samples were independently prepared and measured for each peptide variant to ensure reproducibility. Buffer and dilute TFA spectra were collected under identical conditions and for the subsequent subtraction from peptide spectra. All spectra were also corrected for the temperature variation of the optical path length.29 ACS Paragon Plus Environment
The Journal of Physical Chemistry 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Page 6 of 35
Data Analysis and Modeling Two-state analysis of the overall folding stability All data processing and modeling was performed using Matlab (MathWorks, Inc., Natick, MA) with programs written in-house. For the determination of the overall stability, the thermal unfolding CD data were analyzed using a two-state thermodynamic model with temperature dependent intensity baselines. We model the CD mean residue molar ellipticity at 222 nm, whose sensitivity to the αhelical content is best established, but as we have shown, global analysis of the CD spectra yields essentially identical results.22 The observed mean residue molar ellipticity at 222 nm as a function of temperature [θ]222(T) is expressed as:
= +
(1)
where [θ]F222(T) and [θ]U222(T) are the temperature dependent ellipticities of the folded and unfolded protein, respectively, and XF and XU are the corresponding molar fractions:
= 1 − =
(2)
The unfolding equilibrium constant K relates the molar fractions to the fundamental thermodynamic parameters Gibbs free energy (∆G), enthalpy (∆H) and midpoint transition temperature (Tm): = −
= −
−
(3)
where T is the absolute temperature and R is the universal gas constant. The temperature dependence of ∆H and ∆S has been neglected in this data analysis, which is equivalent to assuming the unfolding heat capacity (∆Cp) to be zero. As has been shown previously,22 the unfolding heat capacity ∆Cp is very small, as expected for such a small protein, and including it in the thermodynamic model does not significantly alter the results. The ellipticities of the folded and unfolded states are assumed to be linearly temperature dependent:
= + ! − = + ! −
(4)
where T0 is the reference temperature most commonly set to 0°C. Limits for the parameters can be estimated from the well-established dependence of the ellipticity on helix length and temperature:30 = "# +
$%& $
− '
()*+ (
(5)
where Θ∞ denotes the ellipticity for an infinitely long helix, n is the number of helical amides and kn is a constant. For αtα, the lengths of the α-helices are 14 - 15 for helix 1 and 9 - 10 for helix 2.31 Based on the values for # , , # ⁄, , and kn previously reported by several groups,30 the expected ranges are ACS Paragon Plus Environment
Page 7 of 35 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
The Journal of Physical Chemistry
(22,000 ≤ 0≤ 15,000) deg.cm2.dmol-1 and (60 ≤ αF ≤ 120) deg.cm2.dmol-1.K-1. As before, the ellipticity of the unfolded state was left unconstrained, as any residual helical structure in the unfolded state is likely to lead to deviations from the standard “random coil” ellipticity. The parameters ∆H, Tm,
0, 0, αF and αU were optimized to fit the two-state thermodynamic model to the experimental data, using standard non-linear least squares procedures.
Analysis of site-specific thermal unfolding with Ising-like statistical mechanical model For detailed modeling of microscopic states of the polypeptide chain during thermal unfolding, all experimental data including CD, 13C IE amide I’ IR and CD of peptide fragments corresponding to the individual helices, were combined together within an Ising-like statistical mechanical model. The processing of the IR data and model parameterization of the Ising-like model closely followed procedures developed and tested previously for the analysis of thermal unfolding of αtα and P22 subdomain proteins.21 Ising-like models assume each individual peptide bond to adopt only two possible states: either native or unfolded. The model considers interactions only between native residues and all non-native interactions are neglected. Specific implementations of Ising-like models differ by additional simplifications and approximations invoked to enumerate the partition function.32 We have found the double-sequence approximation with loops (DSA/L) variant, which was developed and rigorously tested by Eaton and coworkers 26,27,33–35 to be both tractable and sufficiently powerful to describe extensive sets of global and site-specific unfolding data for two small proteins, including the original αtα sequence.21 The DSA/L partition function is written as: :)6
7 /012⁄3 = 1 + ∑: 67 ; ∑97 ;
567 ,97 +
:)67 :)97 67 < ∑: ∑9< ; ∑:)6 67 ; ∑97 ; 9< ;97 67 567 ,97 56< ,9< =1 + >67 ,97 ;6< ,9< @
(6)
where wj,i is the statistical weight matrix for a stretch starting from position i of length j amide bonds. The statistical weights are described by: 56,9 = =−A6,9 ⁄B@
(7)
with the total energy contribution: 96)
A6,9 = ∑*;9
∑96) E;* C∆*,E − ∆F ∗ H
(8)
where k and l are amino acid residues and ∆k,l is the contact map matrix,ɛ is the inter-residue contact (free) energy and ∆s the average enropy cost of fixing a peptide bond in the native conformation. The vji;lk term, which accounts for the contacts between native segments ji and lk separated by an unfolded loop is given by:
ACS Paragon Plus Environment
The Journal of Physical Chemistry 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
)
6 9 )
7 7 >67 ,97 ;6< ,9< = I ∑J;9 7
Page 8 of 35
6< 9< ) 97 67 ;9< ) ∑(;9 =C∆J,( @ − =∆KELLM @N
