Unaided trifluoroacetic acid pretreatment solubilizes polyglutamine peptides and retains their biophysical properties of aggregation

Analytical Biochemistry 494 (2016) 23e30

Contents lists available at ScienceDirect

Analytical Biochemistry journal homepage: www.elsevier.com/locate/yabio

Unaided trifluoroacetic acid pretreatment solubilizes polyglutamine peptides and retains their biophysical properties of aggregation Gunasekhar Burra, Ashwani Kumar Thakur* Department of Biological Sciences and Bioengineering, Indian Institute of Technology, Kanpur 208 016, India

a r t i c l e i n f o

a b s t r a c t

Article history: Received 18 July 2015 Received in revised form 14 October 2015 Accepted 15 October 2015 Available online 26 October 2015

Understanding the biophysical mechanism of polyglutamine (polyGln) aggregation is important to unravel the role of aggregates in the pathology of polyGln repeat disorders. To achieve this, synthetic polyGln peptides are widely used. Their disaggregation and solubilization is essential because it plays a crucial role in reproducing biophysical experimental data under in vitro conditions. Pretreatment with trifluoroacetic acid (TFA) and hexafluoroisopropanol (HFIP) at a 1:1 ratio is currently the method of choice to achieve solubility of polyGln peptides. Here we report that the disaggregation and solubilization of polyGln peptides can be achieved by TFA alone. We tested TFA due to the close similarity of it with HFIP in the nature of H-bond breakage and formation, higher cost, and the problems faced by us in the availability of HFIP. Our results demonstrate that the TFA disaggregated polyGln sequences give similar solubilization yield, aggregation kinetics, thioflavin T (ThT) binding, and structural features in comparison with the TFA/HFIP method. Furthermore, we show by limited validation studies that the proposed TFA method can replace the existing TFA/HFIP disaggregation method of polyGln sequences. © 2015 Elsevier Inc. All rights reserved.

Keywords: Polyglutamine Huntington's disease Disaggregation Solubilization Trifluoroacetic acid Hexafluoroisopropanol

Protein aggregation is associated with at least 20 diseases of brain, systemic, and localized origin. Some of the proteins/polypeptides involved are beta-amyloid peptide, a-synuclein, prions, polyglutamines, insulin amyloid polypeptide, and transthyretin [1]. Nine different polyglutamine-containing proteins are involved in Kennedy disease, different forms of ataxia (types 1, 2, 3, 6, 7, and 17), and Huntington's disease [1,2]. The expansion of polyglutamine (polyGln) length in mutant huntingtin (mHtt) above the threshold limit of approximately 35 residues due to corresponding CAG expansion mutation in the huntingtin (IT15) gene is responsible for Huntington's disease [3]. In vitro approaches have been used extensively to examine protein aggregation [4e19]. A major difficulty faced in these studies is the preparation of monomeric solutions of aggregationprone proteins and peptides. This step is required to ensure that

Abbreviations used: polyGln, polyglutamine; DMSO, dimethyl sulfoxide; TFA, trifluoroacetic acid; HFIP, hexafluoroisopropanol; ThT, thioflavin T; PBS, phosphatebuffered saline; RP-HPLC, reversed-phase high-performance liquid chromatography; TMeAFM, tapping modeeatomic force microscopy; TEM, transmission electron microscopy; FTIR, Fourier transform infrared; RSD, relative standard deviation. * Corresponding author. E-mail address: [email protected] (A.K. Thakur). http://dx.doi.org/10.1016/j.ab.2015.10.006 0003-2697/© 2015 Elsevier Inc. All rights reserved.

aggregation is initiated from the same starting conditions. Inadequate solubilization can lead to misinterpretation of kinetics and thermodynamics of aggregating reactions. To resolve the solubility problem, nonvolatile denaturing solvent such as dimethyl sulfoxide (DMSO) and volatile denaturing solvents such as trifluoroacetic acid (TFA), hexafluoroisopropanol (HFIP), and a combination of both (TFA/HFIP) have been tested for their ability to disaggregate and solubilize insoluble synthetic peptides [20]. The use of DMSO is imperfect due to its inability to solubilize the synthetic peptides completely. Besides, its continuous presence in the reaction mixture may lead to biased results. Hence, for more than a decade, the TFA and HFIP combination had been the preferred choice for disaggregating polyGln-containing peptides. The biophysical understanding of polyGln aggregation was derived from synthetic polyGln peptides of different lengths [21]. These studies were possible because of the development of a robust procedure where synthetic polyGln peptides are pretreated for 3 h to overnight in a 1:1 ratio of TFA/HFIP at room temperature. After incubation, volatile solvents are evaporated, followed by a vacuum drying step. The obtained disaggregated peptide residue is solubilized in watereTFA (pH 3.0) [22,23]. This procedure improved the solubilization of otherwise insoluble polyGln peptides. It further

24

G. Burra, A.K. Thakur / Analytical Biochemistry 494 (2016) 23e30

resolved the problem of non-reproducibility of experimental data reported for similar peptides [7,24,25]. We noticed that the cost of HFIP per milliliter is higher than that of TFA per milliliter. In addition, we observed that HFIP availability at times is low. Moreover, both TFA and HFIP may act in a similar fashion for making or breaking H-bonds during disaggregation of insoluble peptides (see Table 1 in Ref. [26]). Therefore, we tested anhydrous TFA alone and compared it with the TFA/HFIP method of polyGln disaggregation. Based on the experimental evidence, we demonstrate here that the recovery of peptides after disaggregation, overall kinetics, and structure of aggregates formed remain the same after disaggregation with TFA when compared with the TFA/HFIP method. Materials and methods Materials All of the peptides (Table 1) were procured in crude from the Keck Biotechnology Center at Yale University. TFA, thioflavin T (ThT), and HFIP were purchased from SigmaeAldrich, and sodium azide was purchased from Sd Fine Chemicals. Phosphate-buffered saline (PBS) was prepared as per the standard procedure described in Cold Spring Harbor protocols. Peptide design Table 1 shows the peptides chosen for this study. Peptides are free at the N and C termini. Q35 peptide was chosen as a representative of simple polyGln peptides. Because the lag time of polyGln aggregation is length dependent [12], PGQ9Q, a 46-residue polyGln peptide representing a model peptide of b-sheet formation, was chosen [27]. The role of flanking sequences, NT17 and P10, in polyGln aggregation was evaluated with the help of NT17Q35P10 and Q35P10 peptides, respectively. NT17 is expected to enhance [13], and P10 delays [9,15], the rate of polyGln aggregation. In this study, NT17Q35P10 was chosen based on its ability to aggregate as a fulllength exon1 encoded huntingtin sequence [28]. We also chose PGQ9I, a variant of PGQ9Q, to understand the effect of isoleucine (I) substitution mutation on polyGln aggregation (R. Mishra and A. K. Thakur, unpublished data).

Disaggregating polyGln-containing peptide sequences Q35, Q35P10, PGQ9Q, and PGQ9I peptides were disaggregated under two different conditions: (i) overnight incubation in a 1:1 ratio of TFA/HFIP at 0.5 mg/ml peptide concentration [23,29] and (ii) overnight incubation in only TFA at 0.5 mg/ml concentration. In both conditions, the final concentration maintained for NT17Q35P10 peptide was 0.1 mg/ml. This is because its disaggregation at a lower concentration improves recovery of this peptide [22,29]. After overnight incubation, the volatile solvents were evaporated in a chemical fume hood using a gentle stream of nitrogen gas, producing a thin peptide film on the glass walls. To further ensure the complete removal of any residual TFA or TFA/HFIP, they were subjected to drying under vacuum in a desiccator for approximately 1 h. The peptide film was then solubilized in watereTFA (pH 3.0) and subjected to gentle swirling to ensure complete solubility [23,29]. Visual inspection showed no cloudiness or insoluble aggregates, suggesting a clear and transparent solution in both conditions. Comparison of yield after disaggregation under TFA and TFA/HFIP conditions Peptide solubilized in watereTFA was ultracentrifuged at 305,611 g and 4  C for 4 h to remove insoluble peptide or microaggregates. Two-thirds of the supernatant was carefully removed, and the peptide concentration of the supernatant was determined by injecting in an Agilent Eclipse Plus C18 column (4.6
100 mm) connected to a reversed phase (RP)eHPLC device. This was achieved by comparing the observed peak area with the standard curve generated for each peptide. The standard curve was generated from the stock concentration that was determined based on the molar extinction coefficient of peptides at 214 nm in an ultraviolet (UV) spectrophotometer [10,30]. The determined concentration was then converted to percentage yield by dividing the supernatant peptide concentration (mg/ml) by the concentration of the initial sample prior to ultracentrifugation. To determine the effect of time of disaggregation, PGQ9I peptide was dissolved in TFA at 0.5 mg/ml for 1e12 h, dried, and resuspended in pH 3.0 water. Solubility was determined as described. To determine the effect of concentration during disaggregation, PGQ9I was dissolved in TFA at various concentrations (0.3e2 mg/ml) for 12 h and solubility was determined as described.

Methods Peptide purification and storage Approximately 1 mg (analytical balance, Sartorius BSA 224SCW) of crude peptide was dissolved in 200 ml of formic acid. It was then adjusted to 20% formic acid with high-performance liquid chromatography (HPLC)-grade water before injecting into a Zorbax SB C3 semi-preparative column (9.4
250 mm) attached to a BioRad (Biologic Duoflow) fast protein liquid chromatography system for purification. The pure fraction of peptide was collected in one glass vial (Borosil, 15 ml) based on the retention time of the main peak that was earlier characterized by mass spectrometry. Subsequently, it was dried by lyophilization and stored at 80  C for future use [22,29]. Table 1 PolyGln containing peptide sequences chosen for the disaggregation and solubilization study. Peptide

Amino acid sequence

Sequence length

Q35 PGQ9Q NT17Q35P10 Q35P10 PGQ9I

K2-Q35-K2 K2-Q9-PG-Q9-PG-Q9-PG-Q9-K2 MATLEKLMKAFESLKSF-Q35-P10-K2 K2-Q35-P10-K2 K2-Q9-PG-Q4-I-Q4-PG-Q9-PG-Q9-K2

39 46 64 49 46

Spontaneous aggregation kinetics analysis The aggregation reaction setup and sedimentation assay were carried out as described previously [22]. All of the aggregation reactions were monitored at pH 7.2 in PBS at 37  C. Sodium azide (0.05%) was added to the reaction mixture to avoid microbial contamination. The aggregation kinetics was monitored by taking supernatant from an aliquot of ongoing aggregation reaction after subjecting it to ultracentrifugation at 305,611 g and 4  C for 30 min. The leftover monomer concentration at regular time intervals was determined by the RPeHPLC method [29,30]. ThT and light scattering assays were performed to corroborate the aggregation kinetics monitored by sedimentation assay [5]. In a 1-ml cuvette, an aliquot of 600 ml sample from ongoing reaction was taken at different time points. The intensity of light scattering was measured by setting both the excitation and emission wavelengths at 450 nm (both of the slit widths were adjusted to 2.5 nm at a voltage of 650). To this, 20 ml of 2.5 mM ThT was added and the spectra were recorded by resetting the excitation and emission wavelengths to 450 nm (slit width of 5 nm) and 489 nm (slit width of 10 nm), respectively, on a PerkinElmer model LS 51 spectrofluorimeter. The final spectra were obtained after averaging over three scans for each spectrum. The light scattering and ThT intensity was normalized against blank and then converted to

G. Burra, A.K. Thakur / Analytical Biochemistry 494 (2016) 23e30

percentage aggregation as described below in the “Data analysis” section. Nucleation kinetic analysis This assay was performed as described previously [22]. Seven different concentrations (90, 70, 50, 40, 30, 20, and 10 mM) of freshly disaggregated Q35P10 were prepared in PBS at pH 7.2 and 37  C. The monomer concentration of the ongoing reactions was determined at different time points based on sedimentation assay. First, 20% monomer drop was used to construct plots of monomer concentration (mM) versus time2 (s2) for each concentration. Using the slopes from t2 plots, logelog plots were constructed between log [initial concentration (M)] and log [slope] of the t2 plots. The slope of the linear fit of logelog plots was used to determine critical nucleus size (n*) based on an expression of slope ¼ n* þ 2. Seeding (and cross-seeding) competency analysis Seeded and cross-seeded reactions were carried out using Q35P10 monomers (30 mM) in PBS at pH 7.2 and 37  C after disaggregation under TFA and TFA/HFIP conditions. Appropriate volumes of preformed seeds (8%, w/w) of Q35P10 generated by using TFA and TFA/HFIP disaggregation protocols were used. The TFA condition generated seeds were incubated to TFA generated and TFA/HFIP generated Q35P10 monomer solutions. Similarly, TFA/HFIP generated seeds were used to seed TFA/HFIP and TFA generated Q35P10 solutions [22]. Similar reactions to which preformed seeds were not added were used as controls for data comparison. The ongoing aggregation reactions were monitored by determining monomer concentration at different time points by using sedimentation assay [22]. The data plotted between monomer concentration (mM) and time (h) was used for comparing elongation rates at different conditions. Fibril imaging using TMeAFM and TEM For tapping modeeatomic force microscopy (TMeAFM), 20 ml of PGQ9I sample containing final aggregates was incubated on a freshly cleaved mica surface and dried using a gentle stream of N2 gas. Furthermore, the surface containing aggregates was washed with 500 ml of HPLC-grade water and dried again [31]. Images of approximately 2
2 mm were collected by using a MultiMode8 modeled TMeAFM instrument (Bruker). For this, tapping mode in air was employed by using the TAP 150A tip with resonance frequency of 150 KHz and spring constant of 5 N/m. The obtained images were processed by using NanoScope image analysis software provided along with the instrument. Samples for transmission electron microscopy (TEM) analysis were prepared by inverting the carbon-coated side of the grid on a 5-ml droplet of aggregate sample and incubating for 30 s. Excess sample was removed with the help of tissue paper. After rinsing the grid with 5 ml of water, it was stained by inverting on a 5-ml droplet of 2% uranyl acetate for 5e10 s [32]. Samples were analyzed on a Tecnai 200-kV transmission electron microscope. FTIR spectroscopy analysis Fourier transform infrared (FTIR) spectroscopy analysis of PGQ9I aggregates generated under TFA and TFA/HFIP disaggregation conditions were carried out using a Tensor 27 FTIR instrument (Bruker). The aggregates were harvested by centrifuging 300 ml of reaction sample (~30 mM) at 25,000 g for 30 min and washed thoroughly eight times using HPLC-grade water. After washing, aggregates were resuspended in 50 ml of water and loaded into a Bio-ATR II cell. Total scans collected for the spectra were 120 at 4 cm1 resolution under constant purging with nitrogen and room temperature. Second derivatives for the amide I region of the

25

primary spectrum were calculated using nine-point smoothing with the Opus 7.2 spectroscopy software package (Bruker Optics) [13,33]. Validation of the method The developed method was subjected to validation as described previously [34e36] to determine its accuracy and precision. Intraassay, inter-assay, and inter-analyst variability study was performed. Because we use the TFA/HFIP method for disaggregating polyGln peptides on a regular basis, we collected spontaneous aggregation kinetics data by sedimentation assay for Q35 (n ¼ 8), NT17Q35P10 (n ¼ 14), and PGQ9I (n ¼ 10) peptides generated by lab members independently on different days. For these peptides, the average aggregation kinetics and standard deviations were determined. The obtained data were compared with those of both the TFA alone and TFA/HFIP disaggregation conditions used in the current study. The inter-analyst variation was analyzed by providing the protocol to a non-specialist in the peptide aggregation area. The analyst was given training to follow the disaggregation protocol step by step. In parallel, the percentage yield obtained after disaggregation and solubilization under both conditions was used to determine the deviation (S) and percentage relative standard deviation (RSD), a measure of method precision [34,35]. Data analysis Sedimentation assay data sets were obtained in triplicates, and the standard deviations were calculated based on all three data sets. The intensities observed at different time points through ThT and light scattering assays were converted to percentage aggregation by dividing the respective value by the final value. This calculation was done by assuming the final intensity value as 100% aggregation, which is also evident from sedimentation assay data where monomer concentration has gone down beyond the limit of detection. The entire data analysis was done using OriginPro 8.5 version data analysis and graphing software [12]. B-spline curve and linear regression fits were chosen to fit all of the time and time2 and logelog plots, respectively. Results and discussion Proteins play a crucial role in the functioning of living cells. However, misfolding and aggregation of certain proteins result in pathological conditions. The limitations associated with in vivo studies require understanding the in vitro mechanism of protein aggregation to correlate to their in vivo significance. Even so, working with such proteins is difficult due to their insoluble nature under normal buffer conditions. Hence, it is important to solubilize them by giving solvent treatments before bringing into aqueous buffer conditions. Among many solvents available, TFA and HFIP are the preferred solvents for disaggregating polyGln-containing peptides and some other amyloid-forming peptides. TFA plays an important role in peptide synthesis, separation, and purification by acting as an ion-pairing agent [37e40]. It has the ability to solubilize a wide variety of proteins and peptides, thereby aiding in improved recovery of their monomeric states without any contamination due to its high volatility [20,41]. Apart from this, spectroscopic studies strongly suggest that the presence of fluorinated alcohols, trifluoroethanol (TFE), and HFIP, even at a moderate concentration, induces secondary structure creating an unnatural state of peptides in aqueous buffers [42e46]. Owing to these facts, we intended to understand the ability of TFA alone to disaggregate and solubilize the polyGln peptides. The extent of solubility was quantified by measuring the monomer concentration of peptides using RPeHPLC assay. This is because,

26

G. Burra, A.K. Thakur / Analytical Biochemistry 494 (2016) 23e30

Table 2 Percentage yield of peptides after disaggregation under the experimental TFA and controlled TFA/HFIP conditions. Peptide

Q35 PGQ9Q NT17Q35P10 Q35P10 PGQ9Ia

Yield (%) TFA/HFIP

TFA

84.75 73.28 85.61 86.93 72.20

89.86 73.02 92.18 84.36 74.51

Standard deviation (S)

Precision (% RSD)

3.6 0.2 4.7 1.8 1.6

4.1 0.3 5.2 2.1 2.2

Note. The table also shows the deviation and precision observed across the two methods. a To enhance the reliability of the data, PGQ9I disaggregation was performed in triplicates under both conditions. The reported yield of PGQ9I is the average yield calculated based on all three data sets. The deviation (S) and percentage relative standard deviation (RSD) observed within TFA/HFIP (n ¼ 3) and TFA (n ¼ 3) disaggregation conditions are 4.3 and 5.9% and 1.8 and 2.4%, respectively.

unlike other spectroscopic techniques, the RPeHPLC method is based on the detector response that correlates directly between the protein/peptide peak area and its concentration [47,48]. In addition, the concentration determination is based on the theoretical molar extinction coefficient obtained by summing the molar extinction coefficients of individual amino acid residues [10,30,47]. Hence, the determined peptide/protein concentration is highly accurate (within 2% error) and reproducible [47]. Peptides were disaggregated and solubilized with either TFA or TFA/HFIP, and the yield was determined. For all five peptides, yields for the two methods were statistically indistinguishable (Table 2). This confirms the efficiency of disaggregation by the TFA method. In addition, the size exclusion chromatography analysis suggests that the soluble fraction of peptide is monomeric and similar when compared with the TFA/HFIP disaggregated condition (see Fig. 1 in Ref. [26]). The comparative yield obtained might be due to prolonged incubation of peptides and replacement of the same volumes of TFA in place of HFIP. The latter was presumed because of the similar ability of TFA to act as both hydrogen bond donor and acceptor as compared with HFIP. This was confirmed by carrying out time- and concentration-dependent disaggregation studies to optimize the disaggregation conditions. This suggests that the incubation of peptides for 2e3 h is sufficient for achieving maximum solubility (Fig. 1A). Incubating for more than 2 h and even up to 12 h resulted in approximately similar solubility (Fig. 1A). From Fig. 1B, it can be observed that maximum solubility was achieved at a 0.5-mg/ml concentration and it remained

constant even on reducing the concentration further to 0.3 mg/ml by adding more TFA. It is also likely that incubation of peptides in TFA can lead to trifluoroacetylation [49]. This was analyzed by comparing the RPeHPLC chromatogram and mass spectra for both of the disaggregation conditions. The presence of a small shoulder at the end of the chromatogram hinted at the possible modification (see Fig. 2 in Ref. [26]). This was confirmed by mass spectrometry, where the presence of a modest amount of ions indicating trifluoroacetylation (m/z 1423, 1448, and 1471) was observed when compared with the expected ions (m/z 1398) (see Fig. 3 in Ref. [26]). It is also possible that these modifications happened even at the time of peptide synthesis because TFA is used as a deprotecting agent to remove main chain and side chain protecting groups [50e52]. The solubilized peptides were then brought to physiological buffer conditions for aggregation analysis. The spontaneous aggregation reaction kinetics was determined by RPeHPLC-based sedimentation assay. RPeHPLC determines the quantity of monomers left in the aggregation reaction at different time intervals [22]. Similarly, the change in the light scattered by growing aggregates was measured to see the progress of aggregation reaction [5,13]. ThT assay was employed to monitor the amyloid-like characteristics of the growing aggregates over time [53]. The spontaneous aggregation kinetics experiment with TFA disaggregated Q35 peptide showed nucleation-dependent aggregation (Fig. 2). The addition of flanking sequences to polyGln peptide altered its overall kinetics of aggregation [9,13,15]. As expected, the addition of a P10 sequence at the C terminus of Q35 peptide (Q35P10) delayed overall aggregation with a distinct and larger lag phase (Fig. 2) compared with simple Q35 peptide, whereas in the case of PGQ9Q the rate of aggregation was faster with a smaller lag phase compared with Q35, which in turn is faster than Q35P10 peptide (Fig. 2). This clearly supports the earlier findings that showed a polyGln length and sequence dependence on rate of aggregation [9,12,13,15]. Overall, the data obtained through sedimentation assay (Fig. 2A and D) was in agreement with that of light scattering (Fig. 2B and E) and ThT (Fig. 2C and F) assays. In addition, the kinetics observed (Fig. 2DeF) was similar to that of the TFA/ HFIP disaggregated condition (Fig. 2AeC). The comparison between TFA and TFA/HFIP disaggregation was further tested on NT17Q35P10 peptide. The addition of N-terminal 17 residues (NT17) to simple polyGln peptides (e.g., Q35) completely alters the aggregation pathway from nucleation-dependent to nonnucleated, downhill aggregation pathway (Fig. 2) [13,15]. As observed earlier, the rate of decrease in monomer concentration

Fig.1. Optimization of disaggregation conditions: Time-dependent (A) and concentration-dependent (B) disaggregation of PGQ9I peptide in TFA.

G. Burra, A.K. Thakur / Analytical Biochemistry 494 (2016) 23e30

27

Fig.2. Spontaneous aggregation kinetics of polyGln-containing peptides. Each peptide (~30 mM) incubated at pH 7.4 and 37  C was monitored by sedimentation (A,D), light scattering (B,E), and ThT (C,F) assays. (AeC) Kinetics under TFA/HFIP disaggregation condition. (DeF) Kinetics under TFA disaggregation condition.

observed through sedimentation assay (Fig. 2A and D) was in agreement with the rate of formation of aggregates observed using light scattering (Fig. 2B and E) and ThT binding (Fig. 2C and F) assays [6,10,13]. Although the aggregation kinetics in TFA and TFA/HFIP disaggregated conditions are in compliance with each other, we further confirmed the similarity with respect to the critical size of nucleus formed during nucleation by carrying out nucleation kinetics analysis for Q35P10 peptide. As expected, based on time plots (Fig. 3A and D), the size of the nucleus obtained through the experimental TFA condition (~0.8) was in compliance with that of the controlled TFA/HFIP condition (~1.1). Both of the above values suggest a critical nucleus of size 1 (n* ¼ 1 after rounding off), indicating that a highly unfavorable folding reaction occurs within the polyGln monomer [14]. We then checked the seeding competency of aggregates obtained through TFA and TFA/HFIP disaggregated conditions to determine structural similarity of monomers and aggregates obtained in both cases [22,27]. Abrogation of nucleation phase (inset in Fig. 4) and overlapping of elongation phases of seeded and crossseeded reactions under TFA and TFA/HFIP conditions (Fig. 4) suggest a similar kind of fiber having the same growth sites for monomer addition. To further confirm the structural similarity between the aggregates, FTIR spectroscopy was employed on PGQ9I aggregates generated after disaggregation in both the TFA and TFA/HFIP conditions. Similar infrared spectra with an absorbance at approximately 1624 cm1 (Fig. 5A) strongly suggest a b-sheet-rich secondary structure for final aggregates generated under both conditions. The morphological features observed with TMeAFM

for PGQ9I aggregates (Fig. 5B and C) and TEM for Q35 aggregates (Fig. 5D and E) are in agreement with the data obtained by using FTIR spectroscopy. These images clearly show long fibrillar structures of similar morphology irrespective of TFA/HFIP (Fig. 5B and D) and TFA (Fig. 5C and E) disaggregated conditions. Moreover, the observed morphologies are in agreement with aggregate structures shown earlier for similar polyGln peptides [12,15]. To enhance the credibility, the proposed methodology was validated for its accuracy and precision. This is because synthetic polyGln peptides are prone to exhibit some batch-to-batch variation without compromising their actual aggregation tendency. In addition, slight changes in pH [54] and salt concentration [22] due to handling errors, lab-to-lab variations, and inter-analyst variations are expected in any method. These possible variations are captured for RPeHPLC-based sedimentation assay for the existing TFA/HFIP method. This was done by collecting and averaging the data generated by all lab members individually on separate experiments for Q35 (n ¼ 8), PGQ9I (n ¼ 10), and NT17Q35P10 (n ¼ 14) peptides (Fig. 6). These data were compared for similar peptides disaggregated in the TFA condition. Interestingly, the data fit very well in the observed range of deviation for all of the peptides (Fig. 6). We also analyzed inter-analyst variations to ensure handling errors by choosing PGQ9I peptide. This was performed by a nonspecialist in this field of study. Interestingly, the amount of peptide recovered (Table 2) and the aggregation kinetics (Fig. 2) followed under the experimental TFA condition were comparable to those of the TFA/HFIP study. In addition, the percentage aggregation observed using light scattering (Fig. 2B and E) and ThT assays (Fig. 2C and F) were in agreement with the percentage decrease in

28

G. Burra, A.K. Thakur / Analytical Biochemistry 494 (2016) 23e30

Fig.3. Nucleation kinetics analysis for Q35P10 peptide at pH 7.4 and 37  C. The top row (AeC) represents the TFA/HFIP disaggregation condition, and the bottom row (DeF) represents the TFA disaggregation condition. Panels A and D show spontaneous aggregation kinetics at different concentrations, and panels B and E show initial 20% drop in monomer concentration obtained from panels A and D, plotted against time2 (s2) to obtain time2 plots. Logelog plots (C,F) were generated by plotting the slopes obtained from panels B and E against respective initial concentrations (M).

soluble monomer observed using the sedimentation assay (Fig. 2A and D). This suggests the reproducibility and accuracy of the proposed method. To gain more confidence in disaggregation and solubilization, the solubilization of longer peptide, PGQ9I, was performed in triplicates under both conditions. The observed deviation (S) and percentage RSD within the TFA/HFIP (n ¼ 3) and TFA (n ¼ 3) disaggregation conditions were 4.5 and 5.9% and 1.8 and 2.4%, respectively. Thus, the RSD value, a measure of precision less than 10% under both conditions, further suggests that the proposed method of disaggregation is precise. Interestingly, the range of RSD values obtained between the TFA and TFA/HFIP methods across the five different peptides was also found to be less than 10%, indicating the reproducibility of the proposed method. Even this suggests that one may generally expect an RSD value of 10% while working with polyGln peptides. Finally, it seems reasonable to state that the proposed method is reproducible, robust, and precise and does not differ from the current method of disaggregation.

Conclusion

Fig.4. Seeding and cross-seeding analysis of Q35P10 peptide (30 mM) in PBS at 37  C. Upward-pointing triangles ( ) and diamonds ( ) represent Q35P10 (TFA disaggregated) monomer seeded with Q35P10 aggregates generated after TFA and TFA/HFIP disaggregation, respectively. Similarly, downward-pointing triangles ( ) and left-pointing triangles ( ) represent Q35P10 (TFA/HFIP disaggregated) monomer seeded with Q35P10 aggregates generated after TFA/HFIP and TFA disaggregation, respectively. Squares ( ) and circles ( ) represent unseeded reaction controls of Q35P10 under TFA and TFA/HFIP disaggregation conditions. The inset shows the comparison between spontaneous aggregation of seeded reactions and nucleation-dependent aggregation of unseeded reactions.

The purpose of the current study was to check the ability of TFA to disaggregate and solubilize polyGln peptides in comparison with the existing protocol that uses TFA/HFIP solvents. On comparison, the data were found to agree not only with the TFA/HFIP method but also with the data shown for similar peptides in previous studies [5,6,9,10,12e15,27]. All of these observations suggest the accuracy, repeatability, and reproducibility of the proposed method of disaggregating and solubilizing polyGln or polyGln-containing peptide sequences of up to 46 Gln residues. Furthermore, the study revealed that the disaggregation of polyGln peptides in TFA at

G. Burra, A.K. Thakur / Analytical Biochemistry 494 (2016) 23e30

29

0.5 mg/ml for approximately 2e3 h is sufficient for achieving maximum solubility. In addition, validation studies strongly suggest that the method has a wide scope. Hence, we believe that it will work well in other laboratories where different analysts are involved. It can also stand as a better alternative for carrying out experiments involving structural determination because it is completely devoid of HFIP, which would otherwise induce some abnormal secondary structural features to peptides. Because the cost of TFA per milliliter is cheaper than the cost of HFIP per milliliter, the proposed method is economical compared with the existing method [26]. Considering all of this, we presume that the proposed method can replace the existing method of polyGln pretreatment for its disaggregation and solubilization. This method can be further used to solubilize various types of polyGln sequences as well as other homo-amino acid stretches prone to aggregation. Conflicts of interest This study was funded by a grant (BT/PR3041/NNT/28/545/ 2011) from Department of Biotechnology, Government of India, to Ashwani Kumar Thakur. The authors declare conflicts of interest according to the patent bearing application number 757/DEL/ 2015. Acknowledgments We greatly acknowledge the assistance provided by Chinmay Kumar Das while validating our results. We are thankful to Rahul Mishra, Mangesh Bawankar, and Itika Saha for their generosity in sharing data for comparison and for useful discussions. Fig.5. Characterization of polyGln aggregates. (A) FTIR spectra of final aggregates of PGQ9I peptide. (BeE) AFM images (B,C) and TEM images (D,E) of PGQ9I and Q35 peptide aggregates, respectively. Shown are images of final aggregates generated through TFA/ HFIP (B,D) and TFA (C,E) methods of disaggregation and solubilization.

Fig.6. Validation of the proposed methodology. Aggregation kinetics of NT17Q35P10 (n ¼ 14), PGQ9I (n ¼ 10), and Q35 (n ¼ 8) peptides disaggregated under the existing TFA/HFIP method by lab members (their unpublished data) were compared with the data produced in triplicates using the TFA method. Squares, upward-pointing triangles, and diamonds ( , , and ) represent the aggregation kinetics (of PGQ9I, Q35, and NT17Q35P10) under the TFA/HFIP condition. Circles, downward-pointing triangles, and left-pointing triangles ( , , and ) represent the aggregation kinetics (of PGQ9I, Q35, and NT17Q35P10) under the TFA condition.

References [1] F. Chiti, C.M. Dobson, Protein misfolding, functional amyloid, and human disease, Annu. Rev. Biochem. 75 (2006) 333e366. [2] C.J. Cummings, H.Y. Zoghbi, Fourteen and counting: Unraveling trinucleotide repeat diseases, Hum. Mol. Genet. 9 (2000) 909e916. [3] Huntington's Disease Collaborative Research Group, A novel gene containing a trinucleotide repeat that is expanded and unstable on Huntington's disease chromosomes, Cell 72 (1993) 971e983. [4] R. Wetzel, W. Yang, V. Berthelier, S. Chen, Polyglutamine repeat sequences, Google Patents, US 20060234385 A1 (2006). [5] S. Chen, V. Berthelier, J.B. Hamilton, B. O'Nuallain, R. Wetzel, Amyloid-like features of polyglutamine aggregates and their assembly kinetics, Biochemistry 41 (2002) 7391e7399. [6] M. Jayaraman, R. Mishra, R. Kodali, A.K. Thakur, L.M.I. Koharudin, A.M. Gronenborn, R. Wetzel, Kinetically competing huntingtin aggregation pathways control amyloid polymorphism and properties, Biochemistry 51 (2012) 2706e2716. [7] D. Sharma, S. Sharma, S. Pasha, S.K. Brahmachari, Peptide models for inherited neurodegenerative disorders: conformation and aggregation properties of long polyglutamine peptides with and without interruptions, FEBS Lett. 456 (1999) 181e185. [8] W. Yang, J.R. Dunlap, R.B. Andrews, R. Wetzel, Aggregated polyglutamine peptides delivered to nuclei are toxic to mammalian cells, Hum. Mol. Genet. 11 (2002) 2905e2917. [9] A. Bhattacharyya, A.K. Thakur, V.M. Chellgren, G. Thiagarajan, A.D. Williams, B.W. Chellgren, T.P. Creamer, R. Wetzel, Oligoproline effects on polyglutamine conformation and aggregation, J. Mol. Biol. 355 (2006) 524e535. [10] M. Jayaraman, R. Kodali, B. Sahoo, A.K. Thakur, A. Mayasundari, R. Mishra, C.B. Peterson, R. Wetzel, Slow amyloid nucleation via a-helix-rich oligomeric intermediates in short polyglutamine-containing huntingtin fragments, J. Mol. Biol. 415 (2012) 881e899. [11] O. Sumner Makin, L.C. Serpell, Structural characterisation of islet amyloid polypeptide fibrils, J. Mol. Biol. 335 (2004) 1279e1288. [12] K. Kar, M. Jayaraman, B. Sahoo, R. Kodali, R. Wetzel, Critical nucleus size for disease-related polyglutamine aggregation is repeat-length dependent, Nat. Struct. Mol. Biol. 18 (2011) 328e336. [13] A.K. Thakur, M. Jayaraman, R. Mishra, M. Thakur, V.M. Chellgren, I.-J.L. Byeon, D.H. Anjum, R. Kodali, T.P. Creamer, J.F. Conway, A. M Gronenborn, R. Wetzel, Polyglutamine disruption of the huntingtin exon 1 N-terminus triggers a complex aggregation mechanism, Nat. Struct. Mol. Biol. 16 (2009) 380e389.

30

G. Burra, A.K. Thakur / Analytical Biochemistry 494 (2016) 23e30

[14] S. Chen, F.A. Ferrone, R. Wetzel, Huntington's disease age-of-onset linked to polyglutamine aggregation nucleation, Proc. Natl. Acad. Sci. U. S. A. 99 (2002) 11884e11889. [15] S.L. Crick, K.M. Ruff, K. Garai, C. Frieden, R.V. Pappu, Unmasking the roles of Nand C-terminal flanking sequences from exon 1 of huntingtin as modulators of polyglutamine aggregation, Proc. Natl. Acad. Sci. U. S. A. 110 (2013) 20075e20080. [16] C.L. Hoop, H.-K. Lin, K. Kar, Z. Hou, M.A. Poirier, R. Wetzel, P.C.A. van der Wel, Polyglutamine amyloid core boundaries and flanking domain dynamics in huntingtin fragment fibrils determined by solid-state nuclear magnetic resonance, Biochemistry 53 (2014) 6653e6666. [17] C.C. Lee, R.H. Walters, R.M. Murphy, Reconsidering the mechanism of polyglutamine peptide aggregation, Biochemistry 46 (2007) 12810e12820. [18] K.A. Burke, K.J. Kauffman, C.S. Umbaugh, S.L. Frey, J. Legleiter, The interaction of polyglutamine peptides with lipid membranes is regulated by flanking sequences associated with huntingtin, J. Biol. Chem. 288 (2013) 14993e15005. [19] R.H. Walters, R.M. Murphy, Aggregation kinetics of interrupted polyglutamine peptides, J. Mol. Biol. 412 (2011) 505e519. [20] S.-C. Jao, K. Ma, J. Talafous, R. Orlando, M.G. Zagorski, Trifluoroacetic acid pretreatment reproducibly disaggregates the amyloid b-peptide, Amyloid 4 (1997) 240e252. [21] R. Wetzel, Physical chemistry of polyglutamine: Intriguing tales of a monotonous sequence, J. Mol. Biol. 421 (2012) 466e490. [22] B. O'Nuallain, A.K. Thakur, A.D. Williams, A.M. Bhattacharyya, S. Chen, G. Thiagarajan, R. Wetzel, Kinetics and thermodynamics of amyloid assembly using a high-performance liquid chromatography-based sedimentation assay, Methods Enzymol. 413 (2006) 34e74. [23] S. Chen, R. Wetzel, Solubilization and disaggregation of polyglutamine peptides, Protein Sci. 10 (2001) 887e891. [24] M.F. Perutz, T. Johnson, M. Suzuki, J.T. Finch, Glutamine repeats as polar zippers: their possible role in inherited neurodegenerative diseases, Proc. Natl. Acad. Sci. U. S. A. 91 (1994) 5355e5358. [25] E. Scherzinger, A. Sittler, K. Schweiger, V. Heiser, R. Lurz, R. Hasenbank, G.P. Bates, H. Lehrach, E.E. Wanker, Self-assembly of polyglutaminecontaining huntingtin fragments into amyloid-like fibrils: implications for Huntington's disease pathology, Proc. Natl. Acad. Sci. U. S. A. 96 (1999) 4604e4609. [26] B. Gunasekhar, A.K. Thakur, Anhydrous trifluoroacetic acid pretreatment convert insoluble polyglutamine peptides to soluble monomers, Data Brief (2015) (submitted for publication). [27] A.K. Thakur, R. Wetzel, Mutational analysis of the structural organization of polyglutamine aggregates, Proc. Natl. Acad. Sci. U. S. A. 99 (2002) 17014e17019. [28] B. Sahoo, D. Singer, R. Kodali, T. Zuchner, R. Wetzel, Aggregation behavior of chemically synthesized, full-length huntingtin exon1, Biochemistry 53 (2014) 3897e3907. [29] M. Jayaraman, A.K. Thakur, K. Kar, R. Kodali, R. Wetzel, Assays for studying nucleated aggregation of polyglutamine proteins, Methods 53 (2011) 246e254. [30] B.J.H. Kuipers, H. Gruppen, Prediction of molar extinction coefficients of proteins and peptides using UV absorption of the constituent amino acids at 214 nm to enable quantitative reverse phase high-performance liquid chromatographyemass spectrometry analysis, J. Agric. Food Chem. 55 (2007) 5445e5451. [31] W.-F. Xue, A.L. Hellewell, W.S. Gosal, S.W. Homans, E.W. Hewitt, S.E. Radford, Fibril fragmentation enhances amyloid cytotoxicity, J. Biol. Chem. 284 (2009) 34272e34282. [32] M. Etienne, N. Edwin, J. Aucoin, P. Russo, R. McCarley, R. Hammer, b-Amyloid protein aggregation, in: G. Fields (Ed.), Peptide Characterization and Application Protocols, Humana Press, Totowa, NJ, 2007, pp. 203e225.

[33] V.G. Ostapchenko, M.R. Sawaya, N. Makarava, R. Savtchenko, K.P.R. Nilsson, D. Eisenberg, I.V. Baskakov, Two amyloid states of the prion protein display significantly different folding patterns, J. Mol. Biol. 400 (2010) 908e921. [34] J.M. Green, A practical guide to analytical method validation, Anal. Chem. 68 (1996) 305e309. [35] R. Singh, HPLC method development and validationdan overview, J. Pharm. Educ. Res. 4 (2013) 26e33. [36] V. Shah, K. Midha, J.A. Findlay, H. Hill, J. Hulse, I. McGilveray, G. McKay, K. Miller, R. Patnaik, M. Powell, A. Tonelli, C.T. Viswanathan, A. Yacobi, Bioanalytical method validationda revisit with a decade of progress, Pharm. Res. 17 (2000) 1551e1557. [37] W.C. Mahoney, M.A. Hermodson, Separation of large denatured peptides by reverse phase high performance liquid chromatography: trifluoroacetic acid as a peptide solvent, J. Biol. Chem. 255 (1980) 11199e11203. [38] P.-M. Yuan, H. Pande, B.R. Clark, J.E. Shively, Microsequence analysis of peptides and proteins: I. Preparation of samples by reverse-phase liquid chromatography, Anal. Biochem. 120 (1982) 289e301. [39] F.E. Regnier, K.M. Gooding, High-performance liquid chromatography of proteins, Anal. Biochem. 103 (1980) 1e25. [40] F.E. Regnier, HPLC of proteins, peptides, and polynucleotides, Anal. Chem. 55 (1983), 1298Ae1306A. [41] J.J. Katz, Anhydrous trifluoroacetic acid as a solvent for proteins, Nature 174 (1954) 509. [42] D.S. Dwyer, R.J. Bradley, Chemical properties of alcohols and their protein binding sites, Cell. Mol. Life Sci. 57 (2000) 265e275. [43] M.R. Nichols, M.A. Moss, D.K. Reed, S. Cratic-McDaniel, J.H. Hoh, T.L. Rosenberry, Amyloid-b protofibrils differ from amyloid-b aggregates induced in dilute hexafluoroisopropanol in stability and morphology, J. Biol. Chem. 280 (2005) 2471e2480. [44] K. Yanagi, M. Ashizaki, H. Yagi, K. Sakurai, Y.-H. Lee, Y. Goto, Hexafluoroisopropanol induces amyloid fibrils of islet amyloid polypeptide by enhancing both hydrophobic and electrostatic interactions, J. Biol. Chem. 286 (2011) 23959e23966. [45] S. Kumaran, R.P. Roy, Helix-enhancing propensity of fluoro and alkyl alcohols: influence of pH, temperature, and cosolvent concentration on the helical conformation of peptides, J. Pept. Res. 53 (1999) 284e293. [46] N. Hirota, Y. Goto, K. Mizuno, Cooperative a-helix formation of b-lactoglobulin and melittin induced by hexafluoroisopropanol, Protein Sci. 6 (1997) 416e421. [47] G.A. Eberlein, Quantitation of proteins using HPLC-detector response rather than standard curve comparison, J. Pharm. Biomed. Anal. 13 (1995) 1263e1271. [48] F. Moffatt, P. Senkans, D. Ricketts, Approaches towards the quantitative analysis of peptides and proteins by reversed-phase high-performance liquid chromatography in the absence of a pure reference sample, J. Chromatogr A 891 (2000) 235e242. [49] J.-G. Kim, D.O. Jang, Trifluoroacetylation of amines with trifluoroacetic acid in the presence of trichloroacetonitrile and triphenylphosphine, Tetrahedron Lett. 51 (2010) 683e685. [50] S.A. Kates, F. Albericio, Solid-phase Synthesis: A Practical Guide, Marcel Dekker, New York, 2000, p. 848. € lzer, P. Alewood, A. Jones, D. Alewood, S.H. Kent, In situ neutraliza[51] M. Schno tion in boc-chemistry solid phase peptide synthesis, Int. J. Pept. Res. Ther. 13 (2007) 31e44. [52] T.J. Curphey, Trifluoroacetylation of amino acids and peptides by ethyl trifluoroacetate, J. Org. Chem. 44 (1979) 2805e2807. [53] H. LeVine 3rd, Quantification of beta-sheet amyloid fibril structures with thioflavin T, Methods Enzymol. 309 (1999) 274e284. [54] S.J. Wood, B. Maleeff, T. Hart, R. Wetzel, Physical, morphological, and functional differences between pH 5.8 and 7.4 aggregates of the Alzheimer's amyloid peptide Ab, J. Mol. Biol. 256 (1996) 870e877.