Article pubs.acs.org/Langmuir
On the Function and Fate of Chloride Ions in Amyloidogenic SelfAssembly of Insulin in an Acidic Environment: Salt-Induced Condensation of Fibrils Viktoria Babenko,† Weronika Surmacz-Chwedoruk,‡ and Wojciech Dzwolak*,† †
Department of Chemistry, Biological and Chemical Research Centre, University of Warsaw Pasteura 1, Warsaw 02-093, Poland Institute of Biotechnology and Antibiotics, Staroscinska 5, Warsaw 02-516, Poland
‡
S Supporting Information *
ABSTRACT: Formation of amyloid fibrils is often facilitated in the presence of specific charge-compensating ions. Dissolved sodium chloride is known to accelerate insulin fibrillation at low pH that has been attributed to the shielding of electrostatic repulsion between positively charged insulin molecules by chloride ions. However, the subsequent fate of Cl− anions; that is, possible entrapment within elongating fibrils or escape into the bulk solvent, remains unclear. Here, we show that, while the presence of NaCl at the onset of insulin aggregation induces structural variants of amyloid with distinct fingerprint infrared features, a delayed addition of salt to fibrils that have been already formed in its absence and under quiescent conditions triggers a “condensation effect”: amyloid superstructures with strong chiroptical properties are formed. Chloride ions appear to stabilize these superstructures in a manner similar to stabilization of DNA condensates by polyvalent cations. The concentration of residual chloride ions trapped within bovine insulin fibrils grown in 0.1 M NaCl, at pD 1.9, and rinsed extensively with water afterward is less than 1 anion per 16 insulin monomers (as estimated using ion chromatography) implying absence of defined solvent-sequestered nesting sites for chloride counterions. Our results have been discussed in the context of mechanisms of insulin aggregation.
■
multivalent cations (e.g., Cu2+ and Zu2+) and amyloidogenic precursors such as Aβ peptide,14,15 or tau protein.16 Arguably, the role of such ion-binding is not limited to charge compensation but also involves a local conformational change on-pathway with the global transitions toward the amyloid structure. Certainly, when an ion stabilizes the native structure, its presence is expected to inhibit fibrillation, as is the case of transthyretin and chloride and iodide anions.17,18 A recent computational study suggested preference of chloride ions for sites with particular positively charged side chains (arginines).19 Hence, pH and dissolved salts affect the rate of amyloid formation and the thermodynamic stability of fibrils through both nonspecific and specific mechanisms.20−22 Another interesting problem linking amyloidogenesis and electrostatics is polymorphism of fibrils. Within the range of physicochemical conditions conducive to fibrillation, a change of pH or ionic strength may favor self-assembly of an alternative amyloid variant (e.g., see refs 23−25). Interestingly, it has been shown that a shift in pH or salt concentration may also switch between amyloid morphologies of fibrils already formed.26,27 Electrostatic interactions within and between individual fibrils are the
INTRODUCTION Amyloidogenic self-assembly of misfolded protein molecules has been attracting a lot of interest primarily because of the fact that formation of amyloid deposits in vivo is involved in pathology of many degenerative maladies including Alzheimer’s disease and diabetes mellitus type II.1,2 In spite of the growing number of examples of benign biologically functional amyloid fibrils,3,4 the medical context remains the main reason for undertaking research in this field. The ability to form amyloid fibrils is now recognized as a common property of polypeptide chains,5 as many synthetic peptides and proteins that normally do not aggregate in vivo have been shown to form fibrils in vitro under appropriate (i.e., more or less denaturing) conditions. There are several aspects of the involvement of electrostatics in protein fibrillation. Extreme, far from the isoelectric point pH may trigger aggregation-prone states of amyloidogenic precursor proteins, but the aggregation itself may be slow due to repulsive Coulombic interactions between charged monomers (e.g., see refs 6−9). In such a situation, Debye screening provided by ions from dissolved salts is expected to accelerate fibrillation.8,10 However, the picture becomes more intricate when Hofmeister series effects11,12 and ion binding13 to specific protein sites occur. In fact, there is a large body of work devoted to fibrillation-promoting docking interactions between © XXXX American Chemical Society
Received: May 22, 2014 Revised: January 10, 2015
A
DOI: 10.1021/la5048694 Langmuir XXXX, XXX, XXX−XXX
Article
Langmuir
monitored at 45 °C with gentle agitation at 300 rpm. Each kinetic trace was calculated as the average of six trajectories. Preparation of Insulin Fibrils at 60 °C in the Presence of Various NaCl Concentrations. Insulin was dissolved at 0.2 wt % concentration in D2O containing NaCl at specified concentrations with the pD subsequently lowered to 1.9 with diluted DCl. Insulin aggregation took place during quiescent 4 day long incubation of asprepared precursor solutions at 60 °C in an Eppendorf Thermomixer Comfort accessory. Subsequently, precipitated aggregates were subjected to atomic force microscopy (AFM)/ICD/FT-IR analysis. Preparation of Insulin Fibrils at 60 °C in the Absence of NaCl Followed by Addition of NaCl. Insulin fibrils for experiments depicted in Figures 4 and 5 were prepared in the absence of salt that was only added afterward, that is, to already fibrillar insulin. Because insulin fibrillation at low pD/pH and in the absence of NaCl is slow (which facilitates undesirable covalent modifications of vulnerable soluble peptide), the process was accelerated by addition of sonicated preformed insulin fibrils (0.2 wt % insulin in 0.1 M NaCl, D2O, pD 1.9 incubated for 4 days/60 °C/0 rpm) at a 100:1 native insulin/fibrils mass ratio. The 100:1 seeding ratio is typically used to induce insulin aggregation, as it provides both reasonably fast fibrillation and daughter fibrils without excessive contamination with residual mother templates. The seeded insulin solution was left to aggregate at pD 1.9 for 4 days/60 °C/0 rpm. In this way, intact daughter fibrils without large quantities of NaCl were obtained. Salt-induced association of individual fibrils was carried out by adding appropriate amounts of crystalline NaCl to thus formed suspensions of daughter fibrils followed by brief mixing and quiescent incubation at 25 °C for 96 h. AFM. Samples of aggregates were initially diluted 250 times with deionized water. A small 8 μL droplet of amyloid suspension was swiftly deposited onto freshly cleaved mica and left to dry overnight. AFM tapping-mode measurements were carried out using a Nanoscope III atomic force microscope from Veeco, USA, and TAP300-Al sensors (res. frequency 300 kHz) from BudgetSensors, Bulgaria. ICD Spectroscopy. Portions (75 μL) of insulin fibrils in acidified D2O-based solutions were further diluted with 2000 μL of 0.1 M NaCl in H2O, pH 1.9, containing 65 μM ThT. All ICD spectra were collected at 25 °C and under quiescent conditions on a J-815 Spectropolarimeter from Jasco, Japan, using 10 mm quartz cuvettes. FT-IR Spectroscopy. For FT-IR measurements, a CaF2 transmission cell and a 0.025 mm Teflon spacer were used. Spectra were collected on a NEXUS Nicolet FT-IR spectrometer (Thermo, USA) equipped with a liquid-nitrogen-cooled MCT detector. For a single spectrum, 256 interferograms of 2 cm−1 resolution were coadded. During measurements, the sample chamber was continuously purged with dry CO2-depleted air. All spectra were corrected by subtracting the proper amount of buffer and water vapor spectra prior to being baseline-corrected. Data processing was performed using GRAMS software (Thermo, USA). Quantification of Chloride Ions Trapped within Insulin Fibrils Using IC. Portions (1000 μL) of freshly prepared 1 w/v % insulin solution in 0.1 M NaCl, D2O, pD 1.9, were incubated in Eppendorfs at 60 °C/1400 rpm/24 h (5 identical probes) or at 60 °C/ 0 rpm/24 h (another 5 identical probes) using an Eppendorf Thermomixer Comfort accessory. After a 24 h long incubation, aggregated samples were sonicated (1 s long pulses repeated 60 times) using a UP200S Ultrasonic Processor manufactured by Hielscher Ultrasonic GmbH operating at 40% of its maximum power output and 24 kHz frequency. Sonicated fibrils were transferred to separate dialysis tubings (Spectra/Por Biotech Regenerated Cellulose Dialysis Membranes MWCO 3500 Da manufactured by Spectrum Laboratories Inc.) and dialyzed against deionized water for 20 h at 25 °C. Dialyzed insulin fibrils were centrifugated at 14 500 rpm and washed with 10fold volumetric excess of deionized water. This procedure was repeated 4 times after which the aggregates were dried in air at 60 °C for 16 h. Finally, 7 mg portions of dry powder of insulin fibrils were dissolved in a 500 μL volume of 0.03 M NaOH at room temperature. Samples were centrifuged on a Nanosep 3K membrane prior to IC analysis. IC measurements were carried out on a Dionex ICS-5000 ion chromatograph system equipped with an IonPac AS11-HC analytical
key factor controlling self-assembly of amyloid superstructures.28 We have shown previously that, in the presence of NaCl, a particular type of amyloid superstructures with strong chiroptical properties precipitates from agitated solutions of acidified insulin.29−31 Because of the chiral transfer between the amyloid superstructure and an orderly bound amyloid-specific achiral dye (such as thioflavin T (ThT)), the electronic transitions of dye molecules become circular dichroism (CD) active. Therefore, the chiral bias of amyloid superstructures may manifest in induced CD (ICD) spectra of bound ThT (“extrinsic Cotton effect”). Previous work has shown that insulin fibrillation under the conditions of strong agitation could lead to two different types of insulin superstructures with quasi-opposite chiroptical phenotypes as defined by the sign (positive or negative) of the extrinsic Cotton effect (induced circular dichroism) induced in amyloid-bound ThT molecules around 450 nm.29−31 Hence, these two types of amyloid superstructures were named +ICD and −ICD fibrils, respectively. Obviously, as no actual global reversal of molecular chirality is plausible when a superstructure with an opposite ICD sign is assembled from L-only mammalian insulins, the relationship between +ICD and −ICD fibrils is diastereomeric.29−31 So far, it remains unclear whether the two structural variants are formed from the same or different individual fibrils. The presence of NaCl is a strict requirement for the selfassembly of either +ICD or −ICD fibrils.29,30 In the absence of NaCl, amyloid fibrils would still form but without merging into the higher-order structures. Elution of salt from insulin amyloid superstructures formed in its presence leads to a gradual disassembly of these entities reflected through decay of their chiroptical properties.30 In this study, we are revisiting the self-assembly of insulin amyloid superstructures in the context of the role and structural destination of chloride counterions.
■
MATERIALS AND METHODS
Insulin from bovine pancreas (cat. no. I6634, lot SLBF4205 V), ThT, and deuterium chloride used for pD adjustment (35 wt % DCl solution in D2O, 99 atom % D) were from Sigma−Aldrich, USA. D2O (“99.8 atom % D” grade) was from ARMAR Chemicals, Switzerland. Deuterated compounds were used in order to enable acquisition of Fourier transform IR (FT-IR) spectra in the amide I′ band region. The batch of commercial bovine insulin used in this study was initially characterized in terms of residual content of chlorides using ion chromatography (IC), as described in the following subchapter. According to the IC assay (Supporting Information, Figure S1), the amount of chloride ions in 10 mg/mL solution of insulin dry powder in 0.03 M NaOH was below the limit of detection estimated as 5.78 μg/mL; that is, there was less than 0.1 mol of chloride ions per 1 mol of insulin. Thus, the concentration of Cl− anions in the insulin samples prepared in this study was effectively determined by the amounts of NaCl and DCl added. Kinetic Measurements of ThT Fluorescence. ThT is a stain with high and selective affinity toward amyloid fibrils and as such is commonly applied in the field of protein aggregation.32 ThT is a molecular rotor whose quantum yield of fluorescence increases when the intramolecular rotation is hindered, for example, because of intercalation on amyloid surfaces.33 Samples for kinetic measurements were prepared by dissolving insulin (with pD adjustment to 1.9 using diluted DCl, uncorrected pH meter readout) at 0.2 wt % concentration in 20 μM ThT in D2O containing a specified concentration of NaCl. Measurements were carried out using a Fluoroskan Ascent FL fluorometer equipped with a pair of λex = 440 nm/λem = 485 nm optical filters and 96-well black microplates (with a 170 μL portion of the insulin solution per well). Aggregation of insulin samples was B
DOI: 10.1021/la5048694 Langmuir XXXX, XXX, XXX−XXX
Article
Langmuir column (4 × 250 mm), an IonPac AG11-HC guard column (4 × 50 mm), and an ISC-5000 Analytical CD conductivity detector. The eluent was a freshly prepared 35 mM aqueous solution of NaOH with a flow rate of 1.5 mL/min. During analysis, the temperature of the columns was set at 30.0 °C, and the temperature of the autosampler was maintained at 5.0 °C. The analysis time was 8 min. Each sample was analyzed twice (10 μL injections). A 163 μM solution of NaCl in 30 mM NaOH was used as the reference standard (corresponding to 5.78 μg Cl−/mL). The limit of detection for chloride ions was estimated experimentally by analyzing six parallel preparations of blank sample and six parallel preparations of standard solution with a low concentration of analyte. The standard solution for which the average measured signal was about 3 times greater than the average measured signal from the blank sample was assumed as the detection limit of the method.
insulin fibrils with different types, numbers, and affinities of ThT binding sites. Figure 2 presents AFM images of fibrils grown at 60 °C in the presence of various NaCl concentrations. The increased
■
RESULTS AND DISCUSSION Trajectories of ThT-fluorescence-monitored insulin aggregation at 45 °C in D2O, pD 1.9, and in the presence of varying concentrations of NaCl are shown in Figure 1. In the absence Figure 2. Amplitude AFM images of insulin fibrils obtained through 4 day long quiescent incubation of 0.2 wt % insulin in D2O, pD 1.9, at 60 °C in the presence of various concentrations of NaCl (as indicated). The scale bar is the same for all panels.
temperature is typical for insulin fibrillation protocols.10 Moreover, at 60 °C, growth of NaCl-free fibrils becomes sufficiently fast to allow insulin to avoid extensive acid-induced degradation. According to the data in Figure 2, with increasing salt concentration, short singly dispersed fibrils begin to stick together into disordered clumps, a rather unsurprising morphological manifestation of Debye screening of positively charged insulin fibrils (pI of insulin being ∼5.3) provided by chloride anions. The agglomeration of fibrils cannot be explained by artifactual interactions between positively charged amyloid and the negatively charged mica surface as essentially the same kind of clustering is observed on neutral HOPG surfaces (Supporting Information, Figure S2). Moreover, the tendency of insulin fibrils to form larger aggregates in the presence of NaCl in bulk samples is also detectable by light scattering (Supporting Information, Figure S3). Interestingly, despite quiescent conditions of incubation, fibrils appear to break into shorter fragments with the increasing ionic strength. Substitution of H2O with D2O allowed us to acquire FT-IR spectra in the amide I′ band region of the same samples as used for AFM measurements (Figure 3). The wavenumber range of the amide I′ peaks of fibril samples grown at different salt concentrations is typical for parallel β-sheet structure, the main secondary constituent of insulin amyloid with the shoulder band at approximately 1660 cm−1 attributed to the looplike region connecting individual β strands.35,36 However, the variations between the individual spectra point to subtle conformational differences within the stacked sheets and possibly different patterns of interstrand hydrogen bonding.37,38 Without additional studies employing Raman spectroscopy, fiber X-ray diffraction, and possibly solid-state NMR, it would be exceedingly risky to attribute such fine spectral effects, for example, the NaCl-induced upshift of the amide I′ band beyond 1625 cm−1 (Figure 3), exclusively to a single underlying factor such as weakening of the hydrogen bonds within the stacked sheets. The data shown in Figures 1−3 illustrates the previously reported tendency that the increasing concentration of NaCl not only accelerates aggregation of insulin in an acidic
Figure 1. Kinetics of spontaneous insulin fibrillation at 45 °C in the presence of different NaCl concentrations monitored by ThT fluorescence (emission at 485 nm). Insulin concentration was 0.2 wt %, pD 1.9. Each kinetic trace was calculated as the average of six trajectories (standard deviations are superimposed as error bars).
of salt, no aggregation is observed within the first 24 h, while raising the NaCl concentration to 5 mM (which represents an approximately 30-fold molar excess relative to the concentration of insulin dimers) results in a tiny increase in fibrillation rate. Only at higher salt concentrations (20 mM and above), both nucleation and elongation significantly accelerate, in accordance with previous studies.10 The considerable spread of trajectories reflects the immanent stochasticity of the de novo fibrillation, in particular, bifurcation events observed for bovine insulin under similar conditions.29,30 It is likely that increasing the NaCl concentration also affects other processes that have been recognized to contribute strongly to the overall rate of insulin fibrillation.34 The fact that significant molar excess of salt is needed to affect the kinetics of insulin aggregation implicates nonspecific effects (e.g., anion shielding of electrostatic repulsion) rather than specific ion−protein interactions.22 Insulin fibrils formed at the different conditions vary in heights of corresponding ThT fluorescence plateaus. Because the quantum yield of bound ThT is predominantly affected by the angle of intramolecular twist around the C−C bond linking the benzothiazole and the dimethylaminobenzene rings (and effectively controlled by the local conformation of the ThT binding amyloid moiety),33 this is an early hint that changing the salt concentration may result in formation of variants of C
DOI: 10.1021/la5048694 Langmuir XXXX, XXX, XXX−XXX
Article
Langmuir
characteristics of +/−ICD superstructures of insulin amyloid fibrils.29,30 The characteristic feature of these superstructures is the chiral bias of surface ThT binding moieties manifesting in ICD spectra around 450 nm.39 Hence, insulin amyloid samples subjected to the delayed incubation in NaCl solutions were subsequently stained with ThT and subjected to CD measurements in the wavelength range corresponding to ThT absorption. The ICD spectra of as-prepared samples are juxtaposed with corresponding FT-IR spectra in Figure 5.
Figure 3. FT-IR spectra corresponding to the amyloid samples with morphologies shown in Figure 2. Amide I′ IR band of insulin amyloid fibrils formed through 4 day long quiescent incubation of 0.2 wt % insulin in D2O, pD 1.9, at 60 °C in the presence of various concentrations of NaCl.
environment but may also trigger polymorphic variants of fibrils (e.g., see refs 10 and 28). Because in the so-far discussed experiments NaCl was present from the earliest (i.e., partial unfolding and nucleation) until the latest stages of aggregation (formation of higher-order structures of individual fibrils), the effects of chloride ions on the polymorphism on the levels of secondary and tertiary/ quaternary structures cannot be separated. Hence, we have designed different conditions of self-assembly of insulin amyloid superstructures that have led to the key experimental results of this work. Namely, individual amyloid fibrils were grown in quiescent NaCl-free insulin samples and were subjected to the NaCl-treatment afterward. Morphologies of insulin fibrils prepared in this way are shown in Figure 4. As the
Figure 5. Spectral data corresponding to the amyloid samples with morphologies shown in Figure 4. (A) ICD spectra of ThT-stained insulin fibrils grown in the absence of NaCl but subjected to NaCltreatment afterward. (B) Amide I′ IR band of as-grown insulin fibrils formed in the absence of NaCl but subjected to incubations with various concentrations of NaCl afterward.
The striking result of this experiment is the strong positive ICD signal obtained for insulin fibrils subjected to incubations in 200 or 100 mM NaCl; individual fibrils grown in the absence of salt (and with “zero” ICD spectral trait29−31) self-assemble into higher-order structures with the characteristic chiroptical properties. This is an unexpected and very interesting observation for several reasons, namely, (i) this is the first time that +ICD superstructures are reported to form under quiescent conditions (a quiescent incubation of insulin in the presence of NaCl would normally result in very weak or negligible ICD signal (Supporting Information, S4 and refs 29−31)); (ii) the superstructures are assembled from fibrils grown at 60 °C, the temperature range strongly favoring (in the presence of NaCl and agitation29,30) the opposite chiral phenotype “−ICD”; (iii) this experimental outcome suggests that the +ICD trait is defined on the superstructural level. Examination of the corresponding IR data in Figure 5B gives support to this observation. The amide I′ band of insulin fibrils grown in the absence of salt split into two components at 1628 and 1619 cm−1 is not visibly affected by the subsequent salt treatment despite all the changes happening on the super-
Figure 4. Amplitude AFM images of insulin fibrils first grown through quiescent seeding of BI with preformed insulin fibrils and in the absence of salt (0.2 wt % insulin, D2O, pD 1.9, 0 mM NaCl, 60 °C/4 days, 1:100 seed/native insulin mass ratio) followed by addition of NaCl to the final concentrations specified in panels and 96 h incubation. The scale bar is the same for all panels.
AFM images reveal, again, the increasing ionic strength favors association of fibrils; however, at higher NaCl concentrations, it is not accompanied by excessive fragmentation. Instead of disordered “clumps”, aggregates with a high aspect ratio are formed with the appearance of thick braids, especially in the case of samples incubated in 100/200 mM NaCl solutions. In fact, these morphological traits are reminiscent of AFM D
DOI: 10.1021/la5048694 Langmuir XXXX, XXX, XXX−XXX
Article
Langmuir
structures, through quiescent incubation at 60 °C/24 h and −ICD fibrils, which form upon intensive vortexing at 1400 rpm at 60 °C for 24 h. Subsequently, the as-formed fibrils of either type were briefly sonicated and the excess of unbound NaCl was removed by extensive dialysis against water followed by a repetitive centrifugation-and-rinsing-with-water routine (Materials and Methods). Eventually, the amyloid pellet was transferred to diluted NaOH that was accompanied by complete disassembly of fibrils and release of trapped counterions. Samples were then analyzed using IC. The corresponding chromatograms are presented in Figure 7. The
structural levels that have been captured through AFM and ICD spectroscopy. While we have shown recently that the secondary structure of amyloid fibrils (controlled by seeding) does play an important role in determining the preference of superstructures for either the +ICD or −ICD phenotype,40 this result clearly resonates with findings by Kurouski et al. that led them to conclude that superstructural chirality of insulin fibrils (determined by vibrational CD) grown at different pH values and under quiescent conditions is determined on higher than secondary order levels.25 The here reported conditions of +ICD superstructural selfassembly indicate that charge-balancing Cl− ions are not trapped within amyloid fibrils but rather are placed between them. However, the situation can be quite different when all the stages of fibrillation, including unfolding and nucleation, take place in NaCl solution, as is the case of agitation-assisted selfassembly of −ICD superstructures.29−31 We have addressed this question in this work by carrying out an IC-based determination of chloride concentration in liquid samples obtained by dissolving in NaOH insulin fibrils formed earlier in the presence of sodium chloride. We have employed a dedicated analytical procedure briefly summarized in Figure 6. By using the same starting solution of 1 wt % bovine insulin in 0.1 M NaCl, D2O, pD 1.9, two types of samples were prepared: ZERO ICD fibrils, which do not form the chiral super-
Figure 7. Analysis of residual chloride ions through IC of NaOHdissolved insulin fibrils preformed under quiescent conditions (A) or agitation at 1400 rpm (B). Retention time range corresponding to the elution of chloride ions is marked with the black frame. Red chromatograms obtained for the standard solution containing 5.78 μg Cl−/mL mark the detection limit of chloride ions under the given conditions in the method. Chromatograms obtained for five independent insulin amyloid samples are marked with dotted lines. Black solid line is the chromatogram of the blank sample (without insulin) subjected to all steps of the analytical procedure shown in Figure 6. Results indicate that the amounts of chloride ions trapped within single insulin fibrils (A) or their vortex-induced superstructures (B) lie below the detection limit of the method, which corresponds to more than 16 insulin monomers per a single Cl− anion.
signal intensities corresponding to chloride ions for five independent analyses of either type of sample are clearly below the estimated limit of detection of this method, 5.78 μg Cl−/mL, which under the conditions employed in this study, corresponds to a single chloride ion per 16 insulin monomers. This substoichiometric result strongly implies lack of solventsequestered binding sites for Cl− ions capable of trapping them within the bulk amyloid. It appears that all insulin−amyloid− chloride interactions take place on the surfaces of the fibrils. This is in accordance with the recent study by Buell et al.22 employing ζ-potential measurements. Our previous work on
Figure 6. Routine sample preparation for IC measurements of amyloid-trapped chloride ions. Insulin fibrils were grown in the presence of 0.1 M NaCl under quiescent conditions or through intensive agitation at 1400 rpm favoring formation of superstructures. Separation of unbound chloride anions involved a preliminary dialysis of sonicated aggregates against water and four consecutive rounds of centrifugation and elution with an excess of water. Thus-rinsed fibrils were dissolved in aqueous NaOH and centrifuged on a membrane prior to acquisition of the ion chromatograms (step not shown). E
DOI: 10.1021/la5048694 Langmuir XXXX, XXX, XXX−XXX
Article
Langmuir the violet inclusion complex of insulin fibrils and oligoiodide ions that forms in statu nascendi insulin fibrillation demonstrated complete accessibility of all amyloid-bound I3−/ I5− ions to chemical reactants added afterward that themselves did not affect amyloid stability.41 Hence, at least for the certain types of insulin fibrils, the densely packed amyloid structure does not permit encapsulation of relatively small counterions such as Cl− or I3−/I5−. The condensation of individual fibrils as evidenced by the data in Figures 4 and 5A is therefore mediated by counterions placed between but not within them. The rapid self-assembly of the higher-order amyloid structures as in the case of the formation of +ICD fibrils (Figure 5A) upon addition of concentrated salts is reminiscent of the phase transition taking place in vitro in DNA in the presence of multivalent cations.42 This may be a surprising common physicochemical aspect of salt-induced transformations in these two different biopolymers. Certainly, further studies are needed at this point to elucidate mechanisms that lead to the swift self-assembly of individual insulin fibrils into +ICD superstructures reported in this study. In conclusion, we have shown that the presence of high salt concentration, so far thought to be condition sine qua non of self-assembly of the chiral superstructures of amyloid fibrils, is critical only at the final stages of this process: lateral association of individual fibrils. Chloride ions responsible for compensation of repulsive interactions are placed outside of fibrils and are solvent-accessible. This hold true for superstructures formed in the presence of NaCl from the earliest stages of amyloidogenesis.
■
(3) Fowler, D. M.; Koulov, A.V.; Balch, W. E.; Kelly, J. W. Functional amyloid − from bacteria to humans. Trends Biochem. Sci. 2007, 32, 217−224. (4) Romero, D.; Aguilar, C.; Losick, R.; Kolter, R. Amyloid fibers provide structural integrity to Bacillus subtilis biofilms. Proc. Natl. Acad. Sci. U.S.A. 2010, 107, 2230−2234. (5) Fändrich, M.; Dobson, C. M. The behaviour of polyamino acids reveals an inverse side chain effect in amyloid structure formation. EMBO J. 2002, 21, 5682−5690. (6) Barrow, C. J.; Yasuda, A.; Kenny, P.; Zagorski, M. G. Solution conformations and aggregational properties of synthetic amyloid βpeptides of Alzheimer’s disease: Analysis of circular dichroism spectra. J. Mol. Biol. 1992, 225, 1075−1093. (7) Jha, S.; Snell, J. M.; Sheftic, S. R.; Patil, S. M.; Daniels, S. B.; Kolling, F. W.; Alexandrescu, A. T. pH-Dependence of amylin fibrillization. Biochemistry 2013, 53, 300−310. (8) Topping, T. B.; Gloss, L. M. The impact of solubility and electrostatics on fibril formation by the H3 and H4 histones. Protein Sci. 2011, 20, 2060−2073. (9) Picotti, P.; De Franceschi, G.; Frare, E.; Spolaore, B.; Zambonin, M.; Chiti, F.; de Laureto, P. P.; Fontana, A. Amyloid fibril formation and disaggregation of fragment 1-29 of apomyoglobin: Insights into the effect of pH on protein fibrillogenesis. J. Mol. Biol. 2007, 367, 1237−1245. (10) Nielsen, L.; Khurana, R.; Coats, A.; Frokjaer, S.; Brange, J.; Vyas, S.; Uversky, V. N.; Fink, A. L. Effect of environmental factors on the kinetics of insulin fibril formation: Elucidation of the molecular mechanism. Biochemistry 2001, 40, 6036−6046. (11) Marek, P. J.; Patsalo, V.; Green, D. F.; Raleigh, D. P. Ionic strength effects on amyloid formation by amylin are a complicated interplay among Debye screening, ion selectivity, and Hofmeister effects. Biochemistry 2012, 51, 8478−8490. (12) Diaz-Espinoza, R.; Mukherjee, A.; Soto, C. Kosmotropic anions promote conversion of recombinant prion protein into a PrPSc-like misfolded form. PLoS One 2012, 7, No. e31678. (13) Raman, B.; Chatani, E.; Kihara, M.; Ban, T.; Sakai, M.; Hasegawa, K.; Naiki, H.; Rao, Ch. M.; Goto, Y. Critical balance of electrostatic and hydrophobic interactions is required for β2microglobulin amyloid fibril growth and stability. Biochemistry 2005, 44, 1288−1299. (14) Atwood, C. S.; Scarpa, R. C.; Huang, X.; Moir, R. D.; Jones, W. D.; Fairlie, D. P.; Tanzi, R. E.; Bush, A. I. Characterization of copper interactions with Alzheimer amyloid β peptides. J. Neurochem. 2000, 75, 1219−1233. (15) Gaggelli, E.; Kozlowski, H.; Valensin, D.; Valensin, G. Copper homeostasis and neurodegenerative disorders (Alzheimer’s, prion, and Parkinson’s diseases and amyotrophic lateral sclerosis). Chem. Rev. 2006, 106, 1995−2044. (16) Savelieff, M. G.; Lee, S.; Liu, Y.; Lim, M. H. Untangling amyloidβ, tau, and metals in Alzheimer’s disease. ACS Chem. Biol. 2013, 8, 856−865. (17) Hammarström, P.; Jiang, X.; Deechongkit, S.; Kelly, J. W. Anion shielding of electrostatic repulsions in transthyretin modulates stability and amyloidosis: Insight into the chaotrope unfolding dichotomy. Biochemistry 2001, 40, 11453−11459. (18) Hörnberg, A.; Hultdin, U. W.; Olofsson, A.; Sauer-Eriksson, A. E. The effect of iodide and chloride on transthyretin structure and stability. Biochemistry 2005, 44, 9290−9299. (19) Friedman, R. Ions and the protein surface revisited: Extensive molecular dynamics simulations and analysis of protein structures in alkali-chloride solutions. J. Phys. Chem. B 2011, 115, 9213−9223. (20) Malisauskas, M.; Weise, C.; Yanamandra, K.; Wolf-Watz, M.; Morozova-Roche, L. Lability landscape and protease resistance of human insulin amyloid: a new insight into its molecular properties. J. Mol. Biol. 2010, 396, 60−74. (21) Shammas, S. L.; Knowles, T. P.; Baldwin, A. J.; MacPhee, C. E.; Welland, M. E.; Dobson, C. M.; Devlin, G. L. Perturbation of the stability of amyloid fibrils through alteration of electrostatic interactions. Biophys. J. 2011, 100, 2783−2791.
ASSOCIATED CONTENT
S Supporting Information *
Additional IC, AFM, light scattering, and ICD data. This material is available free of charge via the Internet at http:// pubs.acs.org.
■
AUTHOR INFORMATION
Corresponding Author
*Phone: +48 22 8220211 ext. 528; fax: +48 22 822 5996; email:
[email protected]. Notes
The authors declare no competing financial interest.
■
ACKNOWLEDGMENTS This work was supported by the National Science Centre of Poland, grant no. 2011/03/N/ST4/00736 (to V.B.) and in part by University of Warsaw (project 501/64-BST-169709 to W.D.). IC equipment was funded by the POIG Key Research Project (contract POIG 01.01.02-00-007/08-00).
■
ABBREVIATIONS USED AFM, atomic force microscopy; FT-IR, Fourier transform infrared; IC, ion chromatography; ICD, induced circular dichroism; ThT, thioflavin T
■
REFERENCES
(1) Chiti, F.; Dobson, C. M. Protein misfolding, functional amyloid, and human disease. Annu. Rev. Biochem. 2006, 75, 333−366. (2) Uversky, V. N.; Fink, A. L. Conformational constraints for amyloid fibrillation: The importance of being unfolded. Biochim. Biophys. Acta 2004, 1698, 131−153. F
DOI: 10.1021/la5048694 Langmuir XXXX, XXX, XXX−XXX
Article
Langmuir (22) Buell, A. K.; Hung, P.; Salvatella, X.; Welland, M. E.; Dobson, C. M.; Knowles, T. P. Electrostatic effects in filamentous protein aggregation. Biophys. J. 2013, 104, 1116−1126. (23) Tao, K.; Wang, J.; Zhou, P.; Wang, C.; Xu, H.; Zhao, X.; Lu, J. R. Self-assembly of short Aβ (16− 22) peptides: Effect of terminal capping and the role of electrostatic interaction. Langmuir 2011, 27, 2723−2730. (24) Kurouski, D.; Lombardi, R. A.; Dukor, R. K.; Lednev, I. K.; Nafie, L. A. Direct observation and pH control of reversed supramolecular chirality in insulin fibrils by vibrational circular dichroism. Chem. Commun. 2010, 46, 7154−7156. (25) Kurouski, D.; Dukor, R. K.; Lu, X.; Nafie, L. A.; Lednev, I. K. Normal and reversed supramolecular chirality of insulin fibrils probed by vibrational circular dichroism at the protofilament level of fibril structure. Biophys. J. 2012, 103, 522−531. (26) Kurouski, D.; Dukor, R. K.; Lu, X.; Nafie, L. A.; Lednev, I. K. Spontaneous inter-conversion of insulin fibril chirality. Chem. Commun. 2012, 48, 2837−2839. (27) Adamcik, J.; Mezzenga, R. Adjustable twisting periodic pitch of amyloid fibrils. Soft Matter 2011, 7, 5437−5443. (28) Foderà, V.; Zaccone, A.; Lattuada, M.; Donald, A. M. Electrostatics controls the formation of amyloid superstructures in protein aggregation. Phys. Rev. Lett. 2013, 111, 108105. (29) Dzwolak, W.; Loksztejn, A.; Galinska-Rakoczy, A.; Adach, R.; Goto, Y.; Rupnicki, L. Conformational indeterminism in protein misfolding: Chiral amplification on amyloidogenic pathway of insulin. J. Am. Chem. Soc. 2007, 129, 7517−7522. (30) Loksztejn, A.; Dzwolak, W. Chiral bifurcation in aggregating insulin: An induced circular dichroism study. J. Mol. Biol. 2008, 379, 9−16. (31) Babenko, V.; Harada, T.; Yagi, H.; Goto, Y.; Kuroda, R.; Dzwolak, W. Chiral superstructures of insulin amyloid fibrils. Chirality 2011, 23, 638−646. (32) Groenning, M. Binding mode of thioflavin T and other molecular probes in the context of amyloid fibrilsCurrent status. J. Chem. Biol. 2010, 3.1, 1−18. (33) Stsiapura, V. I.; Maskevich, A. A.; Kuzmitsky, V. A.; Turoverov, K. K.; Kuznetsova, I. M. Computational study of thioflavin T torsional relaxation in the excited state. J. Phys. Chem. A 2007, 111, 4829−4835. (34) Manno, M.; Craparo, E. F.; Martorana, V.; Bulone, D.; San Biagio, P.L. Kinetics of insulin aggregation: Disentanglement of amyloid fibrillation from large-size cluster formation. Biophys. J. 2006, 90, 4585−4591. (35) Ivanova, M. I.; Sievers, S. A.; Sawaya, M. R.; Wall, J. S.; Eisenberg, D. Molecular basis for insulin fibril assembly. Proc. Natl. Acad. Sci. U.S.A. 2009, 106, 18990−18995. (36) Zandomeneghi, G.; Krebs, M. R. H.; McCammon, M. G.; Fändrich, M. FTIR reveals structural differences between native βsheet proteins and amyloid fibrils. Protein Sci. 2004, 13, 3314−3321. (37) Nielsen, L.; Frokjaer, S.; Carpenter, J. F.; Brange, J. Studies of the structure of insulin fibrils by Fourier transform infrared (FTIR) spectroscopy and electron microscopy. J. Pharm. Sci. 2001, 90, 29−37. (38) Carpenter, J. F.; Kendrick, B. S.; Chang, B. S.; Manning, M. C.; Randolph, T. W. Inhibition of stress-induced aggregation of protein therapeutics. Methods Enzymol. 1999, 309, 236−255. (39) Dzwolak, W.; Pecul, M. Chiral bias of amyloid fibrils revealed by the twisted conformation of thioflavin T: An induced circular dichroism/DFT study. FEBS Lett. 2005, 579, 6601−6603. (40) Dzwolak, W.; Surmacz-Chwedoruk, W.; Babenko, V. Conformational memory effect reverses chirality of vortex-induced insulin amyloid superstructures. Langmuir 2013, 29, 365−370. (41) Dzwolak, W. Insulin amyloid fibrils form an inclusion complex with molecular iodine: A misfolded protein as a nanoscale scaffold. Biochemistry 2007, 46, 1568−1572. (42) Teif, V. B.; Bohinc, K. Condensed DNA: condensing the concepts. Prog. Biophys. Mol. Biol. 2011, 105, 208−222.
G
DOI: 10.1021/la5048694 Langmuir XXXX, XXX, XXX−XXX
