Effect of Oxidation and Mutation on the Conformational Dynamics and

Sep 25, 2009 - Applied Physics, School of Applied Sciences, RMIT University, Melbourne, VIC, Australia; Department of Biochemistry and Molecular Biolo...
0 downloads 0 Views 3MB Size
Download PDF
14006

J. Phys. Chem. B 2009, 113, 14006–14014

Effect of Oxidation and Mutation on the Conformational Dynamics and Fibril Assembly of Amyloidogenic Peptides Derived from Apolipoprotein C-II F. S. Legge,† K. J. Binger,‡ M. D. W. Griffin,‡ G. J. Howlett,‡ D. Scanlon,‡ H. Treutlein,§ and I. Yarovsky*,† Applied Physics, School of Applied Sciences, RMIT UniVersity, Melbourne, VIC, Australia; Department of Biochemistry and Molecular Biology, Bio21 Molecular Science and Biotechnology Institute, The UniVersity of Melbourne, ParkVille, VIC 3010, Australia; and Cytopia Research Pty. Ltd., Melbourne, VIC, Australia ReceiVed: April 27, 2009; ReVised Manuscript ReceiVed: August 5, 2009

The oxidation of methionine residues in proteins can inhibit the self-assembly of proteins to form amyloid fibrils. For human apolipoprotein (apo) C-II the oxidation of methionine at position 60 inhibits fibril formation by the mature protein and by the core peptides apoC-II56-76 and apoC-II60-70. To investigate the molecular nature of these effects, we carried out fully solvated, all-atom molecular dynamics simulations of the structural changes in apoC-II56-76 associated with substitutions of oxidized methionine (Met ox) at position 60. The results with apoC-II56-76 (Met ox) showed less flexibility in structure, leading to a perturbation of the hydrophobic core. Valine substitution at position 60 showed an increased tendency to explore a wide range of conformational space, whereas the behavior of the Gln substitution mutant was similar to the wild-type peptide. These simulations are consistent with kinetic measurements which showed that a Met60Gln substitution within apoC-II56-76 had little effect on the rate of fibril formation whereas substitution of Met ox or Val at position 60 lead to significant inhibition of peptide fibril formation. The effects of amino acid modification and substitutions on the kinetics of peptide fibril formation differ from the effects observed with full-length apoC-II inferring that additional mechanisms are involved in fibril formation by mature apoC-II. Introduction The misfolding and self-assembly of proteins to form insoluble amyloid fibrils occurs in a number of debilitating diseases including Alzheimer’s disease (AD), Parkinson’s disease, variant Creutzfeldt-Jakob disease (vCJD), and type II diabetes.1-3 Interest in the mechanism for the formation of specific partially folded intermediates and their assembly into stable oligomeric intermediates arises from recent studies that show the oligomeric species are the toxic species.4 Partially folded intermediates can arise as a result of various specific structural transitions. One such characteristic transition is the conversion of R-helices into β-strands,5 which combine to form the typical cross β-sheet structure of amyloid fibrils.6 Recent NMR studies illustrate the important role of “turns” defined by specific geometric features in amyloid forming peptides in the nucleating mechanism of amyloid formation.7 Experimental and simulation studies of the β-amyloid (Aβ) peptide associated with AD identify a region with an intrinsic propensity to form turn structure (Val24-Lys28), shedding light on the role of interactions between hydrophobic regions in the stabilization of such structures.7-9 Solid-state NMR experiments of amyloid fibrils have also observed this type of turn structure.10 The selfassembly of amyloidogenic monomers arising from the formation of turn structure, are believed to expose hydrophobic patches that initiate the early stages of oligomerization.11 The oxidation of Met in proteins reduces or eliminates biological activity and protein-protein interactions for native proteins.12 The effects of Met oxidation are of particular interest in the amyloidosis mechanism because of the potential inhibitory †

RMIT University. The University of Melbourne. § Cytopia Research Pty. Ltd. ‡

role and the reversibility of the process. However, the means by which this occurs is not clear. There are several studies which show that Met oxidation affects the propensity for fibril formation by a number of amyloidogenic proteins, including the Aβ peptides,13-15 prion protein,16,17 human R-synuclein,18 transthyretin,19 immunoglobulin light chain LEN,20 and apolipoprotein C-II.21 Studies of the early stages of Aβ-assembly, suggest that the importance of hydrophobic and bendlike structures in the amyloid formation process, and that oxidation of Met35 blocks the critical hydrophobic association in the initial transformations.14 For the Aβ1-40 and Aβ1-42 peptides, oxidation of the Met-35 side chain to methionine sulfoxide (Met-35 ox), caused a significant reduction in the rate of fibril formation at physiological pH,13,15 resulting in reduced neurotoxicity. It was reported that the Met-35 ox decreased the propensity for β-strand structure at two hydrophobic regions (Leu17-Ala21 and Ile31Val36), and turn-type structure at Asp7-Glu11 and Phe20-Ser26. Our understanding of the effect of Met oxidation on fibril formation and disease is further complicated by the association of oxidized fibrils and toxicity. For example, the human prion protein which is associated with variant Creutzfeldt-Jakob disease (vCJD), contains Met residues and even though the oxidation of the methionine reduces the propensity for fibrillization, the neurotoxic effect of the oxidized peptide in vivo is slightly increased.17 Human apolipoprotein (apo) C-II is a lipid-binding protein and a member of a large apolipoprotein family involved in blood lipid transport. ApoC-II is typical of several members of the apolipoprotein family which share a tendency to misfold in the lipid-free state to form amyloid fibrils.22 Immunohistochemical studies have identified apolipoprotein aggregates, including apoC-II, in atherosclerotic plaques. These aggregates may contribute to an age-related decrease in aortic elasticity.23,24

10.1021/jp903842u CCC: $40.75  2009 American Chemical Society Published on Web 09/25/2009

Conformational Dynamics of Amyloidogenic Peptides

J. Phys. Chem. B, Vol. 113, No. 42, 2009 14007 TABLE 1: Summary of apoC-II56-76 Peptide Systems Labeling, and Chemical Formula

Figure 1. Ribbon diagram showing structural features of the NMR structure of apoc-II (ref 31) and the peptide 56-76 with amino acid sequence. Residue 60 has been shown in liquorice representation.

ApoC-II amyloid fibrils bind to the CD36 receptor of macrophages and activate the macrophage inflammatory response,25 further implicating apoC-II fibrils in cardiovascular disease. ApoC-II amyloid fibrils have a cross-β sheet structure,26 composed of two core regions (19-37 and 54-74) which remain resistant to hydrogen-deuterium exchange and proteolysis.27 This region includes a tryptic peptide apoC-II56-76, and a smaller peptide apoC-II60-70 which retain the ability to assemble into amyloid fibrils.27 Full-length apoC-II contains two Met residues; Met-9 and Met-60 which are sensitive to oxidation. Hydrogen peroxide induced oxidation of apoC-II21 inhibits fibril formation and leads to an increase in the free pool of apoC-II in equilibrium with the fibrils. Oxidation of Met-60, contained within the C-terminal core region of apoC-II play a key role in the oxidation-induced inhibition of fibril formation. Met60Gln mutants, designed to mimic the oxidized Met residue with a similar change in hydrophobicity in this region, also showed reduced rates of fibril formation. In contrast, Met60Val mutants assembled into amyloid fibrils with kinetics very similar to that of the wild type.21 As discussed above, previous studies have shown the effects of Met oxidation on a number of proteins, including the full length apoC-II. In addition, the usefulness of studies of peptides of amyloid proteins, such Aβ and prion peptides,14,16 in providing valuable insight into the fibril formation mechanism, has also been shown. The effect of Met oxidation of the apoCII56-76 peptide, which is the subject of our investigation, is not known. To examine the molecular effect of oxidation and mutation on apoC-II fibril assembly, the conformational dynamics of the apoC-II56-76 peptide was analyzed by both computational and experimental techniques (see Figure 1). We have previously used computational techniques to explore the dynamic behavior of the apoC-II56-76.28 These studies showed that apoC-II56-76 populated an ensemble of turn structures, stabilized by hydrogen bonds and hydrophobic interactions enabling the formation of a strong hydrophobic core postulated to provide the conditions required to initiate aggregation. MD analysis showed a hydrophobic core forming between two hydrophobic

regions in the peptide, residues 58-60, 66-67, and a turn region between residues 62-64. Here, fully solvated all-atom molecular dynamics simulations totalling ∼800 ns were used to compare the structure and dynamics of the wild-type apoC-II56-76 with the behavior of peptide variants containing methionine sulphoxide at position 60. To explore the effects of altered hydrophobicity at position 60, variants with Gln or Val mutations at this position were also studied. To validate and complement our computational study, the results from the MD simulations were subsequently compared with in vitro studies. Kinetic measurements of fibril formation were performed on these peptides and showed to be in good agreement with the computational analysis. Interestingly, the pattern of fibril formation by the apoC-II peptides, as predicted both theoretically and experimentally, differs from that of the corresponding modifications in mature apoC-II, suggesting that additional processes govern the rate of fibril formation by full length apoC-II. Method Simulations. The simulations were performed with the NAMD software package29 and the all-atom CHARMM27 force field30 on the tryptic peptide comprising residues 56 to 76 of apoC-II (see Figure 1). Simulations of four fully solvated peptides of the NMR experimental structure in dodecylphosphocholine (PDB code: 1SOH)31 were performed: (1) the NMR or wild-type structure (56-76), (2) the peptide with oxidized Met 60 (56-76ox), and (3,4) substituted Val (56-76Met60Val) or Gln (56-76Met60Gln). The peptides will hereafter be denoted as indicated in the closed brackets. Two sets of simulations for each system were performed. The simulations were performed twice to gain a better estimate of statistical properties from the calculation. The two sets of simulations for each system are denoted as [1] and [2]. Table 1 shows a summary of the peptide systemssincluded are the system labeling, and the chemical structures of residue 60. It should be noted that the system 56-76[1] was discussed in our previous paper with the aim of comparing the effects of different starting conformations.28 Protonation of side chains was consistent with pH 7. The peptides were soaked with ∼11800 TIP3 water molecules32 with density of 1 g/cm3 in a periodic box of size 70 Å × 70 Å × 70

14008

J. Phys. Chem. B, Vol. 113, No. 42, 2009

Legge et al.

TABLE 2: Cluster Analysis for All Systems of apoC-II56-76 56-76[1] 848 10.8 36.4

56-76[2] 431 19.6 51.7

56-76 ox[1] 644

56-76 ox[2] 534

56-76 M60 V[1] Number of States 505

56-76 M60 V[2] 2747

56-76 M60Q[1] 669

56-76 M60Q[2] 841

16.9

Percentage Time Occupied by the Most Populated State 11.7 7.7 2.6

2.5

9.2

46.5

Percentage Time Occupied by 6 Most Populated States 41.6 35.2 10.4

22.3

27.4

Å. The systems were minimized by conjugate gradient method for 1000 steps to remove steric clashes. All of the systems were equilibrated by slow heating with temperature reassignment to 300 K (over 24 ps) to a total of 50 ps with monitoring of the potential energy. During the data collection of 100 ns for each system, the velocities were coupled to a temperature bath of 300 K. Atom-based cutoff of 12 Å with switching at 10 Å for nonbonded van der Waals interactions was used. The particle mesh Ewald (PME) summation33 was applied for long-range electrostatic interactions. Bonds containing hydrogen atoms were constrained using the SHAKE algorithm to their energy minimized values, thus allowing a numerical integration time step of 2 fs to be used in the simulation. The atomic coordinates were saved every 2500 steps (5 ps) for analysis. Unless otherwise specified, analysis was performed on all frames generated from the simulation, that is, structures were collected for analysis every 5 ps over the simulation time. Met is an easily oxidized amino acid and can undergo reversible oxidation to the sulfoxide (Met(O)) or sulfone (Met(O2). The oxidation of Met to the sulfoxide has been shown to occur in a wide variety of proteins34 and is the oxidized form used in this study. A modified form of the standard Charmm topology file containing a Met sulfoxide was constructed using the parameters for dimethyl sulfoxide from Strader and Feller (2002).35 The Met sulfoxide was in the S-methionine S-sulfoxide configuration. The oxidized Met 60 residue will be referred to as Met ox 60. Peptide Synthesis and Oxidized Met 60 Preparation. Native56-76(STAAMSTYTGIFTDQVLSVLK),56-76Met60Gln (STAAQSTYTGIFTDQVLSVLK) and 56-76Met60Val (STAAVSTYTGIFTDQVLSVLK) were assembled fully manually by Bio21 Peptide Technologies (Vic, Australia) using Fmoc solid phase synthesis. The purity (>95%) and identity of the peptide were confirmed by reversed-phase high performance liquid chromatography (HPLC) and mass spectrometry using a Q-TOF LC mass spectrometer (Agilent Technologies, USA). Met ox 60 peptide was prepared by incubation of native 56-76 peptide with 0.1% H2O2 in 8 M guanidine hydrochloride, 10 mM Tris.HCl, pH 8.0 for 2 h. The success of the oxidation procedure was determined by HPLC. Concentrated stocks of apoC-II 56-76, 56-76ox, 56-76Met60Gln and 56-76Met60Val were maintained in 8 M GuHCl, 10 mM Tris.HCl pH 8.0 at concentrations of 15-25 mg/mL, and stored at -20 °C. The concentration of each stock solution was determined with an extinction coefficient at 280 nm of 1,290 M-1 cm-1. Thioflavin T Fluorescence Measurements. Fibril formation by native, mutant, and Met ox apoC-II56-76 peptides was initiated by dilution from concentrated guanidine hydrochloride stocks into 100 mM sodium phosphate pH 7.4, and incubating at room temperature at a final concentration of 0.5 mg/mL. To monitor fibril assembly, at the required time points 12 µL aliquots of sample were added to 220 µL of 8 µM ThT in refolding buffer. Fluorescence intensities were measured in duplicate using an fmax fluorescence plate reader (Molecular Devices, Sunnyvale,

CA) with excitation and emission filters of 444 and 485 nm, respectively. Control measurements were performed with a 0.1 mg/mL sample of bovine serum albumin (Sigma, MO) which was incubated at room temperature in refolding buffer with ThT fluorescence measurements as per the peptide procedure. These control readings were then subtracted from the measured ThT fluorescence of the apoC-II56-76 peptide samples to give a baseline corrected ThT fluorescence. Results and Discussion Structural Analysis. The program PEPCAT (PEPtide Conformational Analysis Tool)36 was used to carry out a cluster analysis of apoC-II56-76 structures generated during the simulations to monitor the effects of oxidation and mutation of the peptide. PEPCAT is particularly useful in monitoring changes in the specific elements of the secondary structure of proteins. The structures were processed according to a set of geometric descriptors (e.g., atom-atom distances, φ and Ψ dihedral angles), and subsequently classified into a set of conformational states, based on similarity in the descriptors. The conformational changes that occur during the MD simulation are reflected in the classified states. In this study, the geometric descriptors were based on the distances between the CR atoms: CRn to CRn+3 distance 10 times more prevalent than previously believed.41 There is sufficient evidence to suggest that the π-helix is a real event of the dynamic behavior of the protein and plays an important role in function.39,41,42 The N terminus region showed continued fluctuation, although was largely stabilized at ∼70 ns lying close to the central helix.

Snapshots of these structures are shown in Figure 3A, including the most populated conformational state at ∼30 ns as classified in the cluster analysis. The second simulation of the wild-type system 56-76[2] as shown in Figure 2B displayed greater stability. The R- and π-helix in the central helix region (between residues 62 and 71) was more clearly defined throughout the simulation, and although the turn structure was still observed in the N terminus, it was less frequent. Figure 3B shows snapshots of structures that formed early in the simulation and were conserved throughout the dynamics, with the most populated conformation at 4.8 ns. Figure 4 illustrates the conserved helical content of the central helix region and the low helical content of the N terminus region in this system. Overall, both simulations of the wild-type system consistently showed a conserved central helix region, together with a more flexible N terminus region. The secondary structure of the oxidized system showed the R- and π-helix conservation in the central region, but also helical presence in the N terminal region, containing the Met ox. In the first simulation, 56-76ox[1], the R-helix remained in the N terminal region for up to ∼53 ns (Figure 2C). At this time point, and again at ∼ 82 ns, a distinct change in the structure with the R-helix forming turn structure occurred. Figure 3C contains snapshots of these structures, with the most populated helical type structure formed at ∼2.3 ns and conserved for more than half of the simulation. The second simulation, 56-76ox[2], retained the helical structure in the N terminus until ∼32 ns, at which time turn structure was formed (Figure 2D). Typical structures illustrated by Figure 3D show the helical conformation in the N terminus at 0.1 and 24.0 ns, and the most populated structure represented by a snapshot taken at 39.4 ns, observed after the formation of turn structure. Figure 4 illustrates the notable difference in helical content in the N terminus region between the oxidized system and the previously described wild type systems. The oxidized system showed the typical central helix region; however, unlike the other systems, there was conspicuous helical conformation formed in the N terminus. This is of particular interest, as a typical conformational transition in amyloid formation is the conversion of R-helices into β-strands, which then associate to form β-sheet structure of amyloid. As discussed previously, this transition is a

Conformational Dynamics of Amyloidogenic Peptides characteristic feature in the formation of the disease-related PrPSc proteins leading to prion infectivity.5 The secondary structure plots for the mutant systems 56-76M60V and 56-76M60Q are shown in Figure 2, E,F and G,H, respectively. The Met60Val system showed the most flexibility and there was also variation in conformations sampled within each mutant system. The first simulation, 56-76M60V[1], retained R- and π-helical structure in the central region; however, in the second simulation, 56-76M60V[2], the peptide showed much more flexibility and lost most of its R-helical structure. Snapshots of these structures are shown in Figure 3E,F, and illustrate the early appearance of the most populated conformations at 8.8 and 1.5 ns, respectively. The Met60Gln system (56-76M60Q[1] and 56-76M60Q[2]) showed less flexibility, and more similarity to the wild-type system with the central helical region conserved and some distinct bridge or β-turn structure formed in both systems. Snapshots of these structures are shown in Figure 3, G and H. Figure 4 shows the pattern of helical content in the mutated peptides was similar to that of the wild-type system, with no helix in the N terminus region and high helical content in the central region. The exception was the 56-76M60V[2] simulation where the Met60Val peptide lost its helical structure in the central region. Potential Role of Residue 60 in Aggregation. For aggregation to occur, we propose that the monomeric peptide undergoes a structural transition which then leads to self-assembly. The previous secondary structure and cluster analysis showed a variation in conformation between the systems of the N-terminus region containing residue 60. While the wild-type and mutant systems showed flexibility and predominantly turn structure, the oxidized system favored helical rather than turn conformation, which also showed higher stability. The increased stability of an oxidized variant has been observed in previous MD simulation of a de novo designed ccβ peptide model system which showed that an oxidized variant of ccβ-Met prevented amyloid fibril formation. These calculations suggested that the ccβ-MetO strands are also more stable in water than in the amyloid fibril.43 If we compare the average of the amount of turn vs helical secondary structure of residue 60 in the systems (56-76, 56-76ox, 56-76M60V, 56-76M60Q), we can readily see the lower turn content and higher helical content in the oxidized system: turn (53%, 45%, 71%, 56%), helical (1%, 31%, 6%, 1%), respectively. As was discussed in our previous work, we propose that the turn formation in this region is important in the formation of the hydrophobic core as an early step in oligomerization.28 To investigate this further, the solvent accessible surface area (SASA) and hydrogen bond (H bond) analysis of the systems were performed. The analysis of the solvent-accessible surface area (see Supporting Information) supported the previous conclusions showing flexibility between the two simulations of the same system, with the exception of the oxidized system. This was particularly evident in the two largely hydrophobic regions (58-60 and 66-71), which showed consistently lower SASA values in the fibril-inhibiting oxidized system compared to the other systems. This may be a factor in limiting the exposure of the hydrophobic patches, a process believed to drive the early stages of oligomerization in the formation of amyloid fibrils.11 These residues have been identified previously as part of the group forming the hydrophobic core of the peptide.28 As a result, it appears that in the Met ox system there was restricted mobility of the hydrophobic residues affecting the exposure of the hydrophobic core necessary for aggregation.

J. Phys. Chem. B, Vol. 113, No. 42, 2009 14011

Figure 5. Fraction of simulation time (%) when H bond interactions of residue 60 were present: (A) H bonds between residue 60 and the rest of the peptide; (B) H bonds between residue 60 and water.

The average number of H bond for structures generated every 50 ps were calculated for each simulation (see Supporting Information). The H bond definition used in this study is that the distance between the donor and acceptor atoms is shorter than 3.5 Å and the angle formed by the donor-hydrogen-acceptor is larger than 150°. Both the average number of intramolecular H bonds for the peptides and the average number of H bonds between the peptides and water showed little difference between the systems. The first Val mutant simulation showed slightly less H bonding, although with a higher variation. To obtain further insight into the role of residue 60 in influencing specific peptide conformations, a detailed examination of the interactions of this residue with the other residues and water was carried out. Of particular interest is the effect of the Met ox 60 on the formation of the R-helix in the N-terminal region. The wild-type and Val mutant systems have no side chain H bonding capability, whereas the Met ox has the potential to form one side chain H bond and the Gln mutant has the potential to form two H bonds with its side chain amide group. The H bond formation of residue 60 was computed by dividing the number of snapshots showing H bonds by the total number of snapshots. This parameter was computed to give an indication of the amount of time this residue was involved in H bonding. Figure 5 shows the calculated frequency of H bonds formed by residue 60 with the rest of the peptide (Figure 5A) and with water (Figure 5B) over the simulation time. Figure 5A shows that the frequency of H bonds of residue 60 to the rest of the peptide for the two simulations of the oxidized system (34%, 18%, respectively) were higher than for the other systems, particularly in the first oxidized simulation. Most of this H bonding involved the backbone atoms of Met ox 60. As expected, Figure 5B shows greater H bonding with water in the oxidized system and Gln mutant system persistent for nearly 100% of the simulation time. The H bonds involve both

14012

J. Phys. Chem. B, Vol. 113, No. 42, 2009

backbone and side chain atoms of Met ox 60 and Gln 60. The similar values for each system pair of simulations observed in Figure 5B indicate the persistence of the interactions and statistical significance of the results. The previous results showed that, although the overall number of H bonds in the systems was similar, there was a greater propensity for H bonding of Met ox 60 in the oxidized system with the rest of the peptide. These interactions predominantly involved the backbone of Met ox 60. The number of intramolecular H bonds over time showed a strong correlation with the R-helical formation in the N terminal in the oxidized systems, which occurred over the first ∼50 ns. For example, in the first simulation of the oxidized system, 86% of the backbone H bonding occurs in the first 53 ns of the simulation corresponding to the formation of the N terminal R-helix region. In the second simulation, 89% of the backbone H bonding occurs in the first 32 ns of the simulation corresponding to the N terminal R-helix region. This also correlated with the lower, more stable SASA observed in these systems. Figure 6 illustrates the types of H bonds which occurred in the oxidized system. The most important H bond of Met ox 60 backbone was with the N terminal residue, Ser 56, with further stabilization by H bonds between Ser 61 and Thr 57 (Figure 6A). This structure was classified in the cluster analysis as the most populated conformational state. The analysis showed a very low frequency of intrapeptide H bonds with Met ox 60 side chain. However, close inspection of the generated conformations in the oxidized system showed the occurrence of water-mediated intrapeptide H bond between the side chain of Met ox 60 and Ser 56 (Figure 6B). This structure was classified in the cluster analysis of 56-76ox[1] as the second most populated conformational state of 56-76ox[1]. Figure 6C shows an example of Met ox 60 H bonding to a water molecule. The average distance between residues 60 and 56 (6.5 Å) over 100 ns for the oxidized system was within the range for potential water-mediated interactions (data not shown). Inspection of the most populated states revealed this type of interaction also occurred consistently in 56-76ox[2], but was not observed in the other systems. The specific behavior of Met ox 60 which favored helix over turn formation was observed in the oxidized system and not in the other systems. The oxidation of Met transforms the hydrophobic Met into a polar residue with H-bonding capabilities. As a result, the increased tendency for solvation of the residue side chain provides one obvious explanation for the differences observed in the structure around residue 60 between the wild-type and Val mutant. It should be noted that Gln 60 and Met ox 60 are residues with a similar chemical character, yet the Gln mutant system showed little backbone H bonding with the rest of the peptide. This may be an entropic effect caused by the slightly different geometry of the side chains of the two residues and the resulting side-chain dynamics. The amide group in the side chain of Gln has planar geometry, whereas the sulfoxide of the oxidized system has tetrahedral geometry. This may restrict the flexibility of the Met ox 60 side chain, but at the same time placing the sulfoxide in an ideal geometry to interact with water. As a result, the backbone is in a good position to interact with residue 56 of the peptide either directly, or through water-mediated interactions, thus stabilizing the helical structure in the N terminus. ThT Fluorescence Experiments. To test our predictions from the modeling results described above, the wild-type and mutant apoC-II56-76 peptides were synthesized: 56-76, 56-76M60V, and 56-76M60Q. Wild-type peptide with Met 60 oxidized (56v76ox) was prepared as described in the Method section.

Legge et al.

Figure 6. Molecular fragments showing the H-bonding patterns in 56-76ox[1]: (A) backbone-Met ox 60 to Ser 56, Ser 61 to Thr 57; (B) water-mediated side chain Met ox 60 to Ser 56; (C) Met ox 60 side chain-water H bonds (hydrogens are not displayed). Hydrogen bonds are shown in green.

The fibril formation of the four peptides was monitored in Vitro by ThT fluorescence at concentration of 0.5 mg/mL (Figure 7). Wild type 56-76 readily assembled into fibrils after an initial delay or lag phase of 24 h. In contrast, the 56-76ox peptide formed fibrils with much slower kinetics; a lag phase of ∼48 h was apparent, but was not completely inhibited. It was observed that the 56-76M60Q peptide behaved very differently to the 56-76ox peptide, and instead had almost identical fibril formation kinetics to the wild type. Interestingly, the fibril formation results follow closely the predictions of the modeling, rather than the results on the equivalent full-length apoC-II protein.21

Conformational Dynamics of Amyloidogenic Peptides

Figure 7. Thioflavin T fluorescence monitoring of mutated apoC-II56-76 peptides: 56-76 (closed circles, solid line), 56-76ox (filled triangles, dotted line), 56-76M60Q (open squares, dashed line), and 56-76M60V (open diamonds, dash-dot line).

The most intriguing result was the kinetics of the 56-76M60V peptide. In the full-length protein, the Val mutant behaves exactly the same as the wild type. However, with the peptides examined, it is the slowest to form fibrils. These results are supported by the predictions from the simulations, which show that the 56-76M60V behaves very differently to the other 3 peptides. It may be that this peptide is very flexible and occupies a large number of conformations, but does not spend enough time in a fibril-competent conformation, which may explain the reduced aggregation kinetics. Circular dichroism spectroscopy was also used to monitor the secondary structural changes of the peptides as they assembled into fibrils. No significant difference was evident for the wild-type and mutated monomeric peptides with a characteristic spectral minimum at 202 nm, indicating a predominantly disordered structure. Upon the formation of peptide fibrils, this spectral minimum at 202 nm was lost and a minimum at 220 nm was apparent, indicating the conversion from disordered structure to increased β-sheet structure (see Supporting Information). Conclusion The present study explores the effect of oxidation and mutation on behavior of the 56-76 fragment of apoC-II. Simulation of the wild-type system showed variation in the N terminus region containing Met 60 including the formation of β-turn structure. After structural rearrangement, a relatively consistent turn structure was formed, thus enabling the formation of a hydrophobic core. This transition is believed to be an essential step in self-assembly of the monomers and amyloid formation and is accompanied by dynamic fluctuation in the hydrophobic residues which make up the hydrophobic core. In contrast, the oxidized peptide showed less flexibility and a tendency to form helical conformation in the N-terminus region containing Met ox 60, stabilized by H bonds to nearby residues, in particular Ser 56, and to water. This may provide an explanation for the inhibition in fibril formation, observed in the kinetic study. This helical conformation and decreased flexibility leads to a lesser tendency for formation of turn

J. Phys. Chem. B, Vol. 113, No. 42, 2009 14013 structure in the N terminus. Such a change in secondary structure may be linked to an entropic effect due to the increased solvation of the oxidized Met which results in a lower free energy of the helical structure in the N terminus. Less fluctuation observed in the hydrophobic residues inferred that the oxidized Met also caused a restriction of the behavior of the hydrophobic residues, a factor associated with the perturbation of the hydrophobic core, a critical step in the amyloid formation mechanism. The type of structure observed in the oxidized system was not observed in the mutant systems. The study showed an interesting difference in the behavior of the 56-76 mutants when compared to the behavior of the equivalent full-length mutants. Unlike the full-length protein, the hydrophobic Val mutation appeared to destabilize the peptide and showed a greater tendency to explore a wide range of conformational space, when compared to the wild-type system. This increased flexibility explains the observed slow kinetics of fibril formation. In addition, unlike the full-length protein the Gln mutant showed closer similarity in structure to the wild-type system. The relatively high content of the turn conformation allows the peptides to combine and form fibrils. This was verified by the similar fibril formation kinetics of the Gln mutant and the wild type. These results highlight the potential difference in mechanism of fibril formation between the 56-76 apoC-II peptide and the full-length protein. The detailed analysis of the effect of substitution of a single Met residue in the amyloidogenic apoC-II provides a molecular basis for the understanding of the stability and aggregation propensity of the native protein, as well as the potential inhibition of fibrils. Abbreviations: apo, apolipoprotein; NMR, nuclear magnetic resonance; Alzheimer’s disease, AD; Aβ, β-amyloid; MD, molecular dynamics; PME, particle mesh Ewald: VMD, visual molecular dynamics; PEPCAT, peptide conformational analysis tool; SASA, solvent-accessible surface area; H bond, hydrogen bond; Met ox, methionine sulfoxide. Acknowledgment. The authors acknowledge the ARC (Australian Research Council) grant no. LP0562041 and Cytopia Pty. Ltd. for providing funding for this project, and to APAC (Australian Partnership for Advanced Computing) for providing a grant of computer time. We also acknowledge colleagues in our group, Andrew Hung and Nevena Todorova. Supporting Information Available: SASA of apoC-II(5676) peptides; average number of H bond for structures generated every 50 ps; and circular dichroism spectroscopy of apoC-II5676 peptide monomer and fibrils. This material is available free of charge via the Internet at http://pubs.acs.org. References and Notes (1) Dobson, C. M. Nature 2002, 418, 729. (2) Sipe, J. D.; Cohen, A. S. J. Struct. Biol. 2000, 130, 88. (3) Selkoe, D. J. Nature 2003, 426, 900. (4) Walsh, D. M.; Klyubin, I.; Fadeeva, J. V.; Rowan, M. J.; Selkoe, D. J. Biochem. Soc. Trans. 2002, 30, 552. (5) Pan, K. M.; Baldwin, M.; Nguyen, J.; Gasset, M.; Serban, A.; Groth, D.; Mehlhorn, I.; Huang, Z.; Fletterick, R. J.; Cohen, F. E.; et al. Proc. Natl. Acad. Sci. U.S.A. 1993, 90, 10962. (6) Sunde, M.; Blake, C. AdV. Protein Chem. 1997, 50, 123. (7) Lazo, N. D.; Grant, M. A.; Condron, M. C.; Rigby, A. C.; Teplow, D. B. Protein Sci. 2005, 14, 1581. (8) Borreguero, J. M.; Urbanc, B.; Lazo, N. D.; Buldyrev, S. V.; Teplow, D. B.; Stanley, H. E. Proc. Natl. Acad. Sci. U.S.A. 2005, 102, 6015. (9) Torok, M.; Milton, S.; Kayed, R.; Wu, P.; McIntire, T.; Glabe, C. G.; Langen, R. J. Biol. Chem. 2002, 277, 40810.

14014

J. Phys. Chem. B, Vol. 113, No. 42, 2009

(10) Petkova, A. T.; Ishii, Y.; Balbach, J. J.; Antzutkin, O. N.; Leapman, R. D.; Delaglio, F.; Tycko, R. Proc. Natl. Acad. Sci. U.S.A. 2002, 99, 16742. (11) Tcherkasskaya, O.; Sanders, W.; Chynwat, V.; Davidson, E. A.; Orser, C. S. J. Biomol. Struct. Dyn. 2003, 21, 353. (12) Kim, Y. H.; Berry, A. H.; Spencer, D. S.; Stites, W. E. Protein Eng. 2001, 14, 343. (13) Hou, L.; Kang, I.; Marchant, R. E.; Zagorski, M. G. J. Biol. Chem. 2002, 277, 40173. (14) Hou, L.; Shao, H.; Zhang, Y.; Li, H.; Menon, N. K.; Neuhaus, E. B.; Brewer, J. M.; Byeon, I. J.; Ray, D. G.; Vitek, M. P.; Iwashita, T.; Makula, R. A.; Przybyla, A. B.; Zagorski, M. G. J. Am. Chem. Soc. 2004, 126, 1992. (15) Johansson, A. S.; Bergquist, J.; Volbracht, C.; Paivio, A.; Leist, M.; Lannfelt, L.; Westlind-Danielsson, A. Neuroreport 2007, 18, 559. (16) Breydo, L.; Bocharova, O. V.; Makarava, N.; Salnikov, V. V.; Anderson, M.; Baskakov, I. V. Biochemistry 2005, 44, 15534. (17) Bergstrom, A. L.; Chabry, J.; Bastholm, L.; Heegaard, P. M. Biochim. Biophys. Acta 2007, 1774, 1118. (18) Uversky, V. N.; Yamin, G.; Souillac, P. O.; Goers, J.; Glaser, C. B.; Fink, A. L. FEBS Lett. 2002, 517, 239. (19) Maleknia, S. D.; Reixach, N.; Buxbaum, J. N. FEBS J. 2006, 273, 5400. (20) Hu, D.; Qin, Z.; Xue, B.; Fink, A. L.; Uversky, V. N. Biochemistry 2008, 47, 8665. (21) Binger, K. J.; Griffin, M. D.; Howlett, G. J. Biochem. 2008, 47, 10208. (22) Hatters, D. M.; Howlett, G. J. Eur. Biophys. J. 2002, 31, 2. (23) Mucchiano, G. I.; Jonasson, L.; Haggqvist, B.; Einarsson, E.; Westermark, P. Am. J. Clin. Pathol. 2001, 115, 298. (24) Haggqvist, B.; Naslund, J.; Sletten, K.; Westermark, G. T.; Mucchiano, G.; Tjernberg, L. O.; Nordstedt, C.; Engstrom, U.; Westermark, P. Proc. Natl. Acad. Sci. U.S.A. 1999, 96, 8669. (25) Medeiros, L. A.; Khan, T.; El Khoury, J. B.; Pham, C. L.; Hatters, D. M.; Howlett, G. J.; Lopez, R.; O’Brien, K. D.; Moore, K. J. J. Biol. Chem. 2004, 279, 10643. (26) Hatters, D. M.; MacPhee, C. E.; Lawrence, L. J.; Sawyer, W. H.; Howlett, G. J. Biochem. 2000, 39, 8276.

Legge et al. (27) Wilson, L. M.; Mok, Y. F.; Binger, K. J.; Griffin, M. D.; Mertens, H. D.; Lin, F.; Wade, J. D.; Gooley, P. R.; Howlett, G. J. J. Mol. Biol. 2006. (28) Legge, F. S.; Treutlein, H.; Howlett, G. J.; Yarovsky, I. Biophys. Chem. 2007, 130, 102. (29) Kale´, L.; Skeel, R.; Bhandarkar, M.; Brunner, R.; Gursoy, A.; Krawetz, N.; Phillips, J.; Shinozaki, A.; Varadarajan, K.; Schulten, K. J. Comput. Phys. 1999, 151, 283. (30) MacKerell, A. D., Jr.; Bashford, D.; Bellott, M.; Dunbrack, R. L., Jr.; Evanseck, J. D.; Field, M. J.; Fischer, S.; Gao, J.; Guo, H.; Ha, S.; Joseph-McCarthy, D.; Kuchnir, L.; Kuczera, K.; Lau, F. T. K.; Mattos, C.; Michnick, S.; Ngo, T.; Nguyen, D. T.; Prodhom, B.; Reiher III, W. E.; Roux, B.; Schlenkrich, M.; Smith, J. C.; Stote, R.; Straub, J.; Watanabe, M.; Wio´rkiewicz-Kuczera, J.; Yin, D.; Karplus, M. J. Phys. Chem. B 1998, 102, 3586. (31) MacRaild, C. A.; Howlett, G. J.; Gooley, P. R. Biochemistry 2004, 43, 8084. (32) Jorgensen, W. L.; Chandrasekhar, J.; Madura, J. D.; Impey, R. W.; Klein, M. L. J. Chem. Phys. 1983, 79, 926. (33) Cheatham, T. E.; Miller, J. L.; Fox, T.; Darden, T. A.; Kollman, P. A. J. Am. Chem. Soc. 1995, 117, 4193. (34) Vogt, W. Free Radical Biol. Med. 1995, 18, 93. (35) Strader, M. L.; Feller, S. E. J. Phys. Chem. A 2002, 106, 1074. (36) O’Donohue, M.; Minasian, E.; Leach, S. J.; Burgess, A. W.; Treutlein, H. R. J. Comput. Chem. 2000, 21, 446. (37) Frishman, D.; Argos, P. Proteins 1995, 23, 566. (38) Humphrey, W.; Dalke, A.; Schulten, K. J. Mol. Graphics 1996, 14, 33. (39) Armen, R.; Alonso, D. O.; Daggett, V. Protein Sci. 2003, 12, 1145. (40) Feig, M.; MacKerell, A. D.; Brooks, C. L. J. Phys. Chem. B 2003, 107, 2831. (41) Fodje, M. N.; Al-Karadaghi, S. Protein Eng. 2002, 15, 353. (42) Legge, F. S.; Budi, A.; Treutlein, H.; Yarovsky, I. Biophys. Chem. 2006, 119, 146. (43) Steinmetz, M. O.; Gattin, Z.; Verel, R.; Ciani, B.; Stromer, T.; Green, J. M.; Tittmann, P.; Schulze-Briese, C.; Gross, H.; van Gunsteren, W. F.; Meier, B. H.; Serpell, L. C.; Muller, S. A.; Kammerer, R. A. J. Mol. Biol. 2008, 376, 898.

JP903842U