Unfolding of Cytochrome c upon Interaction with Azobenzene

Sep 24, 2012 - When most of the protein is folded, that is, Trp 59 does essentially not emit fluorescence, unfolding is obviously betrayed by an incre...
0 downloads 0 Views 593KB Size
Subscriber access provided by INDIANA UNIV PURDUE UNIV AT IN

Article

Unfolding of cytochrome c upon interaction with azobenzene-modified copolymers. Jing Sun, Juliette Ruchmann, Agnes Pallier, Ludovic Jullien, Michel Desmadril, and Christophe Tribet Biomacromolecules, Just Accepted Manuscript • Publication Date (Web): 24 Sep 2012 Downloaded from http://pubs.acs.org on October 3, 2012

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 free 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 accessible to all readers and 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.

Biomacromolecules 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 32

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

Biomacromolecules

Unfolding of cytochrome c upon interaction with azobenzene-modified copolymers. Jing Sun1, Juliette Ruchmann1, A. Pallier3, L. Jullien1, M. Desmadril2, Christophe Tribet1,*. (1) Ecole Normale Supérieure, Département de chimie, UMR8640 CNRS-ENS-UPMC, 24 rue Lhomond, F-75005 Paris, France. (2) Laboratoire de Modélisation et d’Ingénierie des Protéines – UMR8619 Université de Paris-Sud, Bât 430, F-91405 ORSAY CEDEX – France (3) ESPCI, Physico-Chimie des Polymeres et Milieux Disperses, CNRS UMR7615, 10 rue Vauquelin, 75005 Paris, France Email: [email protected]

ABSTRACT. Hydrophilic or amphiphilic macromolecules are common organic matrices used to encapsulate and protect fragile drugs such as proteins. Polymer cargoes are in addition designed for remote control of protein delivery, upon imparting the macromolecules with stimuli-responsive properties, such as light-triggered polarity switches. The effect of interaction between polymers and proteins on the stability of the proteins is however rarely investigated. Here we studied the unfolding/folding equilibrium of Cytochrome c (cyt c) under its oxidized or reduced forms, in the presence of various amphiphilic copolymers (by circular dichroism and intrinsic fluorescence measurements). As models of stimuli-responsive amphiphilic chains, we considered poly(acrylic acid) derivatives, modified to contain

1 ACS Paragon Plus Environment

Biomacromolecules

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 32

hydrophobic, light-responsive azobenzene moieties. These copolymers are thus capable to develop both ionic (under their sodium forms at pH>8) and hydrophobic associations with the basic protein cyt c (isoelectric point of 10.0). In aqueous buffer upon increasing urea concentrations, cyt c underwent unfolding, at [urea] of 9-10 M, which was analyzed under the framework of the equilibrium between two states (native – unfolded). In the presence of polymers, the native folding of cyt c was preserved at low concentrations of urea (typically < 4M). However, the presence of polymers facilitated unfolding, which occurred at urea concentrations lowered by 2M to 4M as compared to unfolding in the absence of polymers (polymer/cyt c ratio of 1:1 g/g). The predominant contribution of coulombic interactions was shown by both the lack of significant impact of the amount of (neutral) azobenzene moieties in the copolymers, and the disappearance of destabilization at ionic strength higher than 150 mM. In addition, stability was similar to that of an isolated cyt c, in the presence of a neutral chain bearing acryloyl(oligoethyleneoxide) units instead of the ionized sodium acrylate moities. DSC measurements showed that in the presence of polymers, cyt c is thermally unfolded in aqueous buffer at temperatures lowered by > 20°C as compared to thermal unfolding in the absence of polymers. Upon exposure to UV light, properties of the polymers chains were perturbed in situ, upon cis:trans isomerization of the azobenzene groups. In polymers displaying a photo-responsive polarity and hydrophobicity switch (conventional azobenzene), the stability of cyt c was not affected by the exposure to light. In contrast, when photo-ionization occurred (using an hydroxyl-azobenzene whose pKa can be photo-shifted), unfolding was initiated upon exposure to light. Altogether, these results show that coulombic binding is a predominant driving force that facilitate unfolding in water:urea solutions. As regards the design of light-responsive systems for protein handling and control of folding, we conclude that remote control of coulombic interaction upon photo-ionization of chromophores can be more efficient that the more conventional photomodulation of polarity.

2 ACS Paragon Plus Environment

Page 3 of 32

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

Biomacromolecules

KEYWORDS (Word Style “BG_Keywords”). Amphiphilic polymer, polyanion/polycation complexes, artificial chaperonin, light-responsive, azobenzene.

Introduction In aqueous solutions, clusters of macromolecules and nanogels have appeared to be versatile matrices for protecting unstable proteins and addressing practical questions of encapsulation, refolding, and controlled delivery. In the context of rapid development of therapeutic and/or engineered proteins and peptides, the non-covalent binding of proteins onto synthetic macromolecules is typically proposed to prevent aggregation and denaturation upon storage 1-7. Dispersions of gels and nanogels, often comprised of amphiphilic copolymers, are developed to entrap fragile proteins which in turn enhances their shelf life in vivo, provides sustained release 8-10 and/or targets their delivery for therapeutic and vaccine applications.11-12. In addition, various macromolecular systems were designed to help in vitro renaturation upon controlled trapping/release of proteins. Some polymer assemblies mimic the effects of natural chaperons (also called heat shock proteins) by preventing thermal aggregation upon in situ immobilization of unfolded proteins into macromolecular clusters.13 Another class of “artificial chaperons” systems help the correct refolding at room temperature, in two steps: first, the formation of soluble complexes between the amphiphilic polymer and urea- or detergent-solubilized proteins. And second, stripping of the solubilizing agent upon for instance dilution (for example see the refolding of lipase from urea solutions14 or membrane proteins in the surfactant sodium dodecyl sulfate15) Recent works in this field aim at the design of “active” artificial chaperons. Activation means here the external control of the association/dissociation scheme between polymers and proteins. A common external trigger is achieved by stepwise variations of temperature, T, combined with polymers that become insoluble in water above a critical temperature threshold (e.g. polypropylene oxide,16 poly (N-isopropylacrylamide)17-18 ). Stimulation under exposure to light holds also great promises because it represents a clean and biocompatible trigger that can be rapidly switched on/off, and cycled. Experimentally, photo-controlled folding has been achieved using a modified natural chaperon, GroE, grafted with lightresponsive moieties (azobenzene) by Muramatsu et al.19 Artificial photosystems also include light-responsive surfactants, that were used to control reversible folding/unfolding cycles on bovine serum albumin20 As regards macromolecules, fully artificial systems were designed by 3 ACS Paragon Plus Environment

Biomacromolecules

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 32

K. Akiyoshi and coll.21 K. Akiyoshi’s group develops pullulan nanogels that trap unfolded proteins, and release them typically upon the competitive binding of additives (cyclodextrin or albumins). They successfully showed the protection and often renaturation of many watersoluble proteins in the presence of these pullulan nanogels.22 Trapping of proteins in the nanogels was made light-sensitive when the pullulan contains highly hydrophobic photochrome groups, such as azobenzene. In turn, photorenaturation was achieved upon lighttriggerred release of proteins.21 To our knowledge, all other examples of photocontrolled association between macromolecule and proteins report on photo-variation of affinities, but did not consider renaturation or folding equilibrium (e.g. photo-adsorption of serum albumin on particles decorated with spiropyrans23 or photo-gelation in mixtures of proteins and azobenzene-containing polymers24). In all the photo-systems cited above, exposure to light essentially switches the hydrophilicity of photochrome groups. Hydrophobic association is certainly recognized to contribute to the formation of polymer:protein complexes24-26. But in general, the multiple origins of interaction between amphiphilic polymers and proteins add complexity to the nature of association. Binding is likely to combine hydrophobic attraction, ionic bridges, and hydrogen bonds.27-29 Presumably because of this complexity, there is no widely recognized route to design an artificial (and stimuli-controlled) chaperon. It has been suggested that hydrophobicity of the polymer chain decreases the stability of the proteins.4,30 But chain length30c and stiffness31 play also a role. In contrast, the absence of significant impact of coulombic binding on the secondary structures and on bioactivity was shown on proteins encapsulated in polyanions/cations multilayers.3,32 Reversible co-precipitation of the complexes affects also significantly the properties and stability of proteins.33 Proteins appears to be protected against thermal denaturation under polymer-trapped forms.27 Upon coulombic association, and in the absence of hydrophobic groups in the polymer chains, the protection against aggregation was pointed out in the seminal works by Wittemann and Ballauff about protein encapsulation inside polyelectrolyte brushes.34 In this context, it appears important to establish whether association with photochrome groups, e.g. azobenzenes, can affect the equilibrium between folded and unfolded states of a protein,

or

whether

those

photochromes

are

mainly

involved

in

protein

attachement/immobilization into the polymer matrix. We address here the question of the impact on protein folding of the presence of azobenzene in amphiphilic copolymers, with or without exposure to light. We used a homologous set of azobenzene-modified polyacrylates, 4 ACS Paragon Plus Environment

Page 5 of 32

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

Biomacromolecules

AzoPol, soluble in water (scheme 1). Controlled amount of azobenzene were grafted in chains having varying charge densities. We could accordingly vary both hydrophobic and coulombic interactions. We studied the stability of a model protein, Cytochrome c (cyt c), in the presence of polymers. Cyt c is a small hemoprotein that is soluble under both native and unfolded forms. Its isolelectric point of 10.0 makes it bearing a net positive charge in usual buffers. Cyt c can be prepared in oxidized or reduced forms having different stability and ionization of the central heme.35 Upon unfolding, cyt c is not prone to self-aggregation, but nevertheless exposes hydrophobic residues to water. It may accordingly develop both coulombic and hydrophobic interactions with copolymers. To estimate the thermodynamic stability of the protein in the presence (or not) of AzoPol, we resorted here to measurements of circular dichroism and intrinsic fluorescence at increasing concentrations of a destabilizing agent, urea. DSC experiments were also conducted to get some insight about thermal unfolding. The relative importance of coulombic vs hydrophobic effects is discussed first on the basis of the measured impact of ionic strength, second by comparison of a neutral AzoPol and polyanionic ones having various ionic charges and degrees of integration of azobenzene. Eventually, in situ photo-variations of either the polarity36 or the ionization37 of two representative AzoPols complete the study: Photo-switching offers here the cleanest mean to vary in situ the interactions with cyt c. The present work shed light on systems that are widely used to control protein stability under the form of small hydrophobic clusters and/or microgels.

Experimental Section Material Cytochrome c (cyt c) from equine heart and sodium dithionite were purchased from Aldrich and used without further purification. Guanidinium chloride (GdnHCl), and urea were obtained from Fluka. NaCl >99% was purchased from Merck.Co.

Polymer synthesis. Azobenzene-modified polyacrylic chains were derived by the coupling of various amines onto parent polyacrylic chains using the same chemistry as described in

24

and

37

. Briefly for PAA_3.5C6azo, the poly(acrylic) acid (PAA, Mn 60 000

g·mol-1, PI ~6, Polyscience Inc, Warrington) was dissolved in N-methylpyrrolidone (NMP) in the presence of dicyclohexylcarbodiimide (coupling reagent) at 40°C. 6-amino-N-(4phenylazophenyl)hexanamide (cf synthesis in ref

24

) was added in NMP and it reacted

quantitatively. The polymer was collected by precipitation of its sodium form in NMP (upon 5 ACS Paragon Plus Environment

Biomacromolecules

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 32

addition of 2 eq. 5M NaOH), filtered, and then dissolved in water (>10 g·L-1) and precipitated twice in methanol. The parent chain for polyNAS_7C6azo_35gly and polyNAS_7C6azo_35gly_15arg, was an homopolymer of N-acryloylsuccinimide (NAS, Acros) prepared in 1,4-dioxane as follows: dry azobisisobutyronitrile (AIBN) degassed under vacuum was mixed with a nitrogen-flushed dioxane solution of the monomer to reach a monomer:AIBN molar ratio of 0.5 mol%. The polymerization ran overnight at 60°C and the chains precipitate during the reaction. This parent polymer was obtained in 90% yield after filtration, washing in acetone, and overnight drying under vacuum. Gel permeation chromatography analysis of this parent chain, under its poly (sodium acrylate) form yielded Mn 42 000 g·mol-1, PI 2.9 (analysis in aqueous 0.5 M LiNO3 solution of the hydrolyzed product obtained by overnight incubation in 1M NaOH and overnight dialysis in slide-A-lyzer, MWCO 5000, GPC Viscotek Malvern instrument France equipped with A6000M PEHMA columns). To yield the copolymers, the (non-hydrolyzed) parent polyNAS was dissolved in dimethylsulfoxide (DMSO) at 40°C and reacted on various amines (C6azo moieties, glycine, arginine) each added with 2 eq. of triethylamine. Eventually, the crude was mixed with 2.2 eq. NaOH (compared to the NAS moieties, added from a 1 M aqueous solution) and let to react for 24 h at 40°C. Upon addition of an equal volume of water, and acidification (10 M HCl), the polymer precipitated and was collected by centrifugation. Dissolution in a basic (NaOH) solution and precipitation upon acification was repeated twice to purify the samples, and the pH of the final solution was adjusted to 7.0 prior to lyophilization. In the case of the non-ionic chain containing 2-hydroxy-azobenzenetrichloride (“azoHCl3” in scheme 1) water solubility was preserved upon integration of ethylene glycol pendant groups

in

the

chain,

by

copolymerization

of

NAS

monomer

and

poly(ethyleglycol)methylmethacrylate one (PEGMMA, Aldrich). Monomers at 1/1 molar ratio of PEGMMA:NAS and 3-mercaptoethanol (2 mol% compared to monomers) were dissolved in 1,4-dioxane with AIBN (0.7 mol%) under nitrogen. The reactants were stirred at 60 °C for 6 h. The resulting crude was dissolved in a small amount of THF, precipitated in diethylether twice and dried under vacuum to yied (77 wt%) the reactive copolymer parent chain. GPC analysis in aqueous LiNO3 (similar procedure of hydrolysis/dialysis as above) indicated a Mn of 42 000 g·mol-1 and PI 2.15. The final PEGMMA:NAS ratio in the chain (56:44 mol/mol) was determined by 1H NMR.37 The amino derivative of azoHCl3 reacted quantitatively with the parent chain in dry DMSO (room T, 2h) and the excess NAS was 6 ACS Paragon Plus Environment

Page 7 of 32

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

Biomacromolecules

quenched with addition of 2 eq. excess ethanolamine in DMSO (room T, 2h) prior to 50% dilution in water and dialysis overnight against water (Slide-A-Lyzer, MWCO 5000, interchim France). The degree of integration of azobenzene was determined from 1H NMR (respectively, from peaks at 3 ppm and between 7 ppm and 8 ppm, i.e. α-amido CH2 and aromatic protons) and from UV-visible spectrophotometry using the absorption coefficient 2.32×104 L·mol-1·cm-1 at 350 nm for the N-(4-phenylazophenyl)hexanamide or 2.0×104 L·mol-1·cm-1 for the hydroxyazobenzene (cf refs 24, 37)

Methods. Unless stated otherwise, all experiments were done in 0.01 M H3PO4-NaOH buffer, pH 7.0, at room temperature. Solutions of 10 µM cyt c in buffer with or without urea (0-10 M) or GdnHCl (8 M) were incubated overnight prior to measurements. Aliquots of polymer stock solution in buffer at the same concentration of urea (resp GdnHCl) were added 15 min. prior to measurements. For experiments on oxidized ferri-cyt c solutions were kept in contact with air. For reduced ferro-cytochrome c solutions of cyt c were prepared 24h before measurements , stored, and handled under argon atmosphere and the buffers contained in addition 1 mM sodium dithionite. Aliquots of the polymer solution were added 15 min. before measurements. Note that at incubation times longer than 10 h, the azobenzene group is reduced into hydrazobenzene under this redox condition. When needed, the samples were irradiated under blue light (436 ± 10 nm at a power of ca. 2 mW·cm-2) or UV light (365± 10 nm nm at a power of ca. 10 mW·cm-2) with the photodiode system CoolLED PE-2 (Roper Scientific France) to reach the photo-stationnary state that fixed trans:cis ratio of the isomers of azobenzene. Chemical unfolding. Fluorescence measurements were carried out using a Fluoromax 3 (Horiba Jobin Yvon) spectrophotometer and LPS 220 (PTI, Monmouth Junction, NJ) spectrofluorometer equipped with a thermostated cell holder at 20 oC. The fluorescence emission spectra were recorded from 320 to 500 nm using an excitation wavelength of 295 nm and a 1×1 cm cell. Circular dichroism (CD) measurements were carried out using a Jasco J/815 spectrophotometer and a 1 mm cell at 20 oC. Thermal unfolding. DSC experiments were performed on a differential scanning calorimeter VP-DSC (Microcal). Measurements were performed on 0.29 g·L-1 cyt c solution in a 10 mM phosphate buffer pH 7.0 with or without polymer at 1:1 g/g ratio. Solutions were

7 ACS Paragon Plus Environment

Biomacromolecules

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 8 of 32

dialysed against the buffer for one night and degassed under mild vacuum just before loading into the calorimeter. Scanning was performed at 1.5 K·min-1 between 20 °C and 100°C. The excess heat capacity, Cp, was calculated by subtracting the signal recorded with buffer and no proteins, using Origin® software. These corrected data were normalized by the concentration . The molar enthalpy of unfolding, ∆H, was calculated as: of protein to give the   ∆     .

eq. 1

With Ti = 25°C and Tf =90°C, and a cubic spline as the baseline below the peak in the transition window. The apparent van’t Hoff enthalpy, ∆HVH, was obtained by fitting equation 2 to Cp data using chi-square minimizing iterative procedure.   .∆.∆    .  .   

eq. 2

assuming a true 2-state model to calculate the apparent equilibrium constant between unfolded and folded protein, K(T), as a function of the fraction of heat evolved between 25°C and the temperature T. If the folding transition is a two-state process, the values of the two enthalpies, ∆H and ∆HVH, are equal and deviation from this case point to the contribution of intermediates or oligomers.

Results and Discussion Polymer structures. The synthesis of the azobenzene-modified copolymers (AzoPols) used in this work have been described in previous studies (for PAA cf ref 24, 38, polyNAS 36 and PAEAPEG 37). In the pH conditions used for the present work, the pendant acid groups in the chains are ionized and the integration of glycine, arginine, or ethylene glycol in the polymers affects the charge density in the chains. The polymers were denoted according to their monomer composition as follows: PAA_xC6azo_ygly_zArg refers to a derivative of PAA with molar fractions of side groups “C6azo”, “glycinyl” and “argininyl” of x, y, and z respectively (cf scheme 1). Accordingly, poly(acrylic acid) (PAA) and its PAA_3.5C6azo derivative (scheme 1) have the highest charge density (and in addition the same chain length and polydispersity). A similar notation is used for compounds prepared from the poly (Nacryloylsuccinimide),

cf

experimental

section:

PolyNAS_7C6azo_35gly

and

polyNAS_7C6azo_35gly_15arg have a lower charge density because of the added glycinyl 8 ACS Paragon Plus Environment

Page 9 of 32

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

Biomacromolecules

“spacer” between the anionic moieties and the backbone and the presence of 15 mol% cationic arginine groups. Finally, the PAEA_56PEGMA_11azoCl3 was derived from a copolymer of ethyleglycolmethylmethacrylate and N-acryloylsuccinimide (post reacted with 11 mol% of hydroxyazobenzene and excess ethanolamine; cf experimental section) to yield the essentially neutral chains shown in scheme 1 (100% neutral from NMR analysis of the degree of integration of ethanolamine and PEO side groups). A low density of hydrophobic azobenzene groups was introduced without significant modification of the solubility of the polymers in water. This point has been established in previous studies by viscosity measurements (PAA_3.5C6azo)24 and spectrophotometry39 (the spectrum of azobenzene is sensitive to solvatation condition). Unavoidably because of the need to copolymerize different monomers, the polymers have diverse chain lengths and polydispersities. The number average degree of polymerization of polyNAS derivatives was close 400 and typically twice as short as that of the PAA derivatives. The commercial PAA used in this study was also markedly more polydisperse (PI 6, cf experimental section) than polyNAS. Upon association of proteins with a long polymer chain, complexes containing several proteins per chain may form,26 which introduces the possibility that chain length and polydispersity affect the composition and thus properties of the complexes. In contrast, upon association of an isolated protein with one polymer chain, the environment of the protein is essentially independent on chain length, provided that the size of the chain is larger than the radius of the protein. Here, the hydrodynamic radius of the polymers (20 nm-30 nm in water)39 was significantly larger than the radius of cytochrome c. We chose in addition experimental conditions that were likely to minimize the contribution of chain lengths: i/ the study was conducted in the presence of excess polymer compared to the protein (see below), in order to determine the maximum degree of perturbation that can be reached, and ii/ we turned to relatively long chains and degree of polymerization always corresponded to a large excess of monomers compared to the amount of charges present on one cyt c. In this study, we aim i/ to establish qualitatively whether the chemical nature of the chain plays a role in stabilization/destabilization of protein folding (of similar chemical nature than side groups in the polymers, the small molecules ethyleneglycol, glycine and arginine are common additives known to affect refolding), and ii/ to analyze primarily the relative importance of the in situ photoswitches of polarity or coulombic interaction by comparison of polymers with varying charge:azobenzene ratios.

9 ACS Paragon Plus Environment

Biomacromolecules

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 10 of 32

Chemical unfolding. We first comment in the following section the “reference” measurements performed at room temperature, in the absence of polymer, on overnightequilibrated 10 µM cyt c solutions in 10 mM phosphate buffer pH 7.0 with urea. Representative measurements by circular dichroism (Figure 1), or by fluorescence (see supporting information Figure S1) indicate the loss of native fold detected by the abrupt modification of the spectra upon increasing urea concentration by about 1 mol.L-1 above a threshold. Both optical measurements were carried out on overnight incubated solutions at varying urea concentrations. Reduced cyt c (see method section) typically unfolds beyond 9.5 M urea, whereas oxidized cyt c starts to unfold beyond 7 M urea. The unfolding equilibrium of cyt c is well documented and it has been matched with a simple two state model (i.e. intermediate states do not accumulate at long incubation time).35 The unfolding upon increasing urea concentration is characterized by a cooperative transition, N↔U, between the native (N) and unfolded conformations (U).40 This equilibrium has been extensively characterized by different techniques including fluorescence41 and circular dichroism (CD).35, 42

The stability of cyt c was accordingly analyzed here under the assumption of a simple N↔U

equilibrium. To get quantitative insight on the denaturant dependence of the degree of unfolding, the intensities of signal, I, of either CD or fluorescence were normalized under the conventional form of Σ=(I-IN)/(IU-IN). IU and IN are respectively the intensities measured for 100% native (buffer with no urea) and 100% unfolded (buffer with 8 M guanidinium chloride) respectively. Figure 2 shows variation of ΣCD data from CD, using the ellipticity at 222 nm as the signal intensity, I, to probe the response in term of secondary structure (the maximum magnitude of variation of ellipticity upon unfolding occurs at this wavelength). Similarly, calculations of ΣFL (Figures S2 and S3 in SI) use the fluorescence intensity to probe the environment of the single tryptophan in cyt c (Trp59, maximum of emission at 370 nm ± 5 nm for reduced cyt c or 350 nm ± 5 nm for oxidized one; excitation at 280 nm ± 5 nm). By definition, Σ increases from zero to one with increasing degree of unfolding. In the framework of the two states equilibrium model, Σ represents the molar fraction of unfolded protein and should obey equation 3:

10 ACS Paragon Plus Environment

Page 11 of 32

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

Biomacromolecules

Σ

 . .. 





∆ !".#

%$∆ !".#

equation 3



%$With x the molar concentration of urea, ∆GH2O the free energy of unfolding in buffer with no urea, m an adjustable parameter,43 aN and aU constant parameters used to adjust the experimental baseline displayed in the absence of unfolding. In urea, m is typically close to 4 kJ·mol-1·M-1 for globular proteins undergoing a sharp cooperative unfolding.43 Fit of equation 3 to data in the absence of polymer yielded ∆GH2O = 56 ± 10 kJ·mol-1 and m=5.4 ± 0.8 kJ·mol1

·M-1 for the reduced form, ferro-cyt c, or ∆GH2O = 29 ± 4 kJ·mol-1 and m=3.8 ±0.4 kJ·mol-1·M-

1

for the oxidized form, ferri-cyt c. These data are in reasonable agreement with results found

in literature of ∆GH2O = 75 ± 6 kJ·mol-1, for the ferro-cyt c,35 and ∆GH2O = 28 ± 2 kJ·mol-1, for the ferri-cyt c.40 Stability of ferro- and ferri-cytochrome c in the presence of AzoPols. In the presence of polymers, added at the concentration of 0.125 g·L-1, i.e. at 1:1 g/g ratio to cyt c, the protein typically underwent unfolding at a lower urea concentration than in the absence of polymers. In Figure 2A, the urea concentration corresponding to half denaturation (ΣCD = 0.5), Cm, decreases from ca. 10 mol·L-1 to a concentration < 7.7 mol·L-1, irrespective of the presence of hydrophobic azobenzene or not in the polymer chain. The magnitude of shift of Cm of ferri-cyt c that was achieved in similar conditions upon addition of PAA_3.5C6azo clearly compares in Figure 2B with that achieved on ferro-cyt c. Of importance to interpret the data, raw intensities of fluorescence measurements in 8 M guanidinium chloride (i.e. references of 100% unfolding) and in buffer with no urea were not affected by the presence of AzoPols. The reference intensities used to calculate Σ (IN and IU, see above) were accordingly fixed, irrespective of the sample composition (with or with no polymer). One can also remark that the native folding was always predominant in the absence of urea, irrespective of the presence of polymers, indicating that the free energy of interaction between cyt c and polymer is not sufficient to overcome the free energy of folding in water. A shift in Cm in the presence of PAA points however that the parent chain poly(sodium acrylate) significantly destabilizes cyt c. The presence of hydrophobic side groups in the PAA chain (e.g. PAA_3.5C6azo) does not bring any significant additional shift to Cm (Table 1), suggesting that coulombic interactions are predominant. At pH 7, ferro-cyt c is positively charged (pI > 10)

44

.

11 ACS Paragon Plus Environment

Biomacromolecules

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 12 of 32

Coulombic attraction with the polyanionic PAA is presumably an important driving force to the formation of cyt c:PAA associates, likely to result in the displacement of the equilibrium toward the unfolded form of the protein. A qualitative confirmation of the dominant contribution of ionic interaction is shown in Figure 3 on the example of PAA_3.5C6azo. At urea concentration fixed to 9 M, ΣFL is equal to ca. 0.5 at low ionic strength, but abruptly drops down to about 0.2 upon increasing the NaCl concentration above 100 mM (Figure 3). Σ FL = 0.2 is equal to the value measured in the absence of polymer at 9 M urea. The global destabilization that is induced upon addition of polymer at low ionic strength appears thus to vanish upon increasing the ionic strength. The variation of Σ FL with polymer:protein ratio brings similar indication of the importance of ionic binding. In figure 3B, ΣFL increases with increasing concentration of PAA_3.5C6azo up to a plateau reached at about 0.025 g/L polymer (i.e. a polymer:protein ratio of ~ 0.2 g/g). Within uncertainty, this value is the same for all four AzoPols studied here, i.e. is not depending on the hydrophicity (not shown here). A 0.2 g/g weight ratio when it is translated in term of anionic:cationic charge ratio correspond to about 1:1 (e.g. 0.97 mol/mol for PAA) assuming that all basic residues in cyt c accounts for one positive charge, i.e. ascribing a formal charge of +23 to cyt c). It is also interesting to note that the effect of polymer on ΣFL saturated well below the polymer/protein amount of 1:1 g/g used in the experiments shown in Figures 2 and S2. Accordingly, the experiments on urea unfolding were conducted in the presence of an excess of polymer. Finally, upon addition of the neutral polymer PAEA_PEGMA_11azoCl3, variation of Cm was close to experimental uncertainty. The presence of the neutral polymer possibly shift slightly the onset of unfolding toward higher urea concentrations, i.e. in the direction of a stabilization of ferri-cyt c. Altogether, the results above suggest that coulombic interaction with the polymers is the main driving force taking place in the destabilization of cyt c, either under its oxidized or reduced forms. To extrapolate thermodynamic parameters of the N ↔ U equilibrium, Equation 3 was fitted to data and the resulting parameters are given in Table 1. However, to be valid, this analysis should converge, within experimental errors, to the same values irrespective of the use of data from fluorescence or from CD measurements. In the case of azobenzene-containing polymers, ΣFL obtained from fluorescence measurements (see Figure S2 in SI) were noticeably lower than ΣCD under the same experimental conditions. For instance, at the maximum urea concentration of 10 M (close to the solubility of urea), when ΣCD was close to 12 ACS Paragon Plus Environment

Page 13 of 32

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

Biomacromolecules

1, Σ FL reached 0.5 – 0.6 with PAA-3.5C6azo, NAS-7C6azo-35gly, or NAS-4C6azo-35gly14arg. It is not possible to loose 100% of the native secondary structure (ΣCD=1), while preserving a low ΣFL in the two state model (ΣFL