Structural Properties of Polyglutamine Aggregates Investigated via

Nov 18, 2008 - International School for AdVanced Studies, Via Beirut 2-4, Trieste, Italy, CNR-INFM-Democritos National. Simulation Center, Beirut 2-4,...
0 downloads 7 Views 2MB Size
J. Phys. Chem. B 2008, 112, 16843–16850

16843

Structural Properties of Polyglutamine Aggregates Investigated via Molecular Dynamics Simulations Giulia Rossetti,†,‡,§ Alessandra Magistrato,*,†,‡ Annalisa Pastore,| Francesca Persichetti,† and Paolo Carloni†,‡,§ International School for AdVanced Studies, Via Beirut 2-4, Trieste, Italy, CNR-INFM-Democritos National Simulation Center, Beirut 2-4, Trieste, Italy, Italian Institute of Technology - SISSA Unit, Via Beirut 2-4, Trieste, Italy, and National Institute for Medical Research, The Ridgeway London, NW71AA, U.K. ReceiVed: July 24, 2008; ReVised Manuscript ReceiVed: October 16, 2008

Polyglutamine (polyQ) β-stranded aggregates constitute the hallmark of Huntington disease. The disease is fully penetrant when Q residues are more than 36-40 (“disease threshold”). Here, based on a molecular dynamics study on polyQ helical structures of different shapes and oligomeric states, we suggest that the stability of the aggregates increases with the number of monomers, while it is rather insensitive to the number of Qs in each monomer. However, the stability of the single monomer does depend on the number of sidechain intramolecular H-bonds, and therefore on the number of Qs. If such number is lower than that of the disease threshold, the β-stranded monomers are unstable and hence may aggregate with lower probability, consistently with experimental findings. Our results provide a possible interpretation of the apparent polyQ length dependent-toxicity, and they do not support the so-called “structural threshold hypothesis”, which supposes a transition from random coil to a β-sheet structure only above the disease threshold. Introduction Huntington disease (HD) is an autosomal-dominant polyglutamine (polyQ) disorder.1 It is caused by expanded CAG trinucleotide repeats in the gene that encodes the protein Huntingtin (Htt). The resulting mutant (mut-Htt) with an extended polyQ tract interacts abnormally with other proteins and causes brain damage by leading to a generalized neuronal dysfunction,1,3,4 including oxidative stress, excitotossic processes, and metabolism deregulation (Figure 1).5 (Early neuropathology involves alteration in the expression of neurotransmitter receptors2,4,6-9 and deregulated mitochondrial homeostasis10,11 that consequently disrupts calcium handling.12 This in turn activates proteases such as calpain and caspase.13,14) The proteolytic processing of mut-Htt15 exposes the toxicity of the mutant protein by releasing short N-terminal polyQcontaining fragments of 100-150 residues (∼exon-1). These fragments form in vivo and in vitro insoluble, β-sheet-containing aggregates,16-22 that constitute the hallmark of the disease (Figure 1).18,19 PolyQ peptides, as well as the exon-1 with a pathogenic Q tract, are sufficient to lead to the apoptosis of the cell,23-28 (the Htt fragments have been observed in nuclear inclusion not only in Huntington but also in other polyQ diseases.3,15,29,30) to cause the disease31 and to feature the same aggregation properties of the full-length mut-Htt,28 both in vivo and in vitro. The polyQs are toxic per se and the aggregation can be therefore studied by considering only polyQ peptides.22 The polyQ length correlates with the severity of the HD and with the age of its onset.32,33 The fully penetrant form of the disease involves polyQ tracts longer than 36-40 Qs (“disease threshold”),34 as observed in other polyQ disorders.32 This threshold varies among polyQ diseases.35 Consistent with these * To whom correspondence should be addressed. † International School for Advanced Studies. ‡ CNR-INFM-Democritos National Simulation Center. § Italian Institute of Technology. | National Institute for Medical Research.

observations are the following considerations: (i) polyQ tracts longer than the toxicity threshold can form in vitro and in vivo SDS insoluble aggregates.18,21,22,36,37 (ii) Specific recognition of expanded toxic polyQ tracts by an anti-polyQ monoclonal antibody has been observed, suggesting the existence of a generic conformational epitope formed only above a certain polyQ length.38-40 These facts have led to the so-called “structural threshold hypothesis”, stating that only above the critical length of 36-40 Qs the polyQs do undergo a structural transition from random coil to a well defined structure based on β-sheet. This hypothesis has however been challenged by various experimental evidences. First, polyQ fragments shorter than the disease threshold also aggregate adopting similar structures to those peptides longer than threshold17,26,27,41 and exhibit toxicity in an eukaryote organism (Caenorhabditis elegans).42 Second, the polyQ length turned out to affect more the kinetics of the aggregates formation rather than their stability.41 In fact, the kinetics of polyQ aggregation is that of nucleated-grow polymerization;28 aggregation is initiated by a monomer that functions as the critical nucleus with the nucleation event consisting of a random coil to β-sheet transition within an individual monomer (lag phase), then fibril formation proceeds via linear addition of single polyQ chains (elongation phase).22,28,43 It has been shown that polyQ length influences the stability of the initial aggregation seed and that may affect the kinetics of its formation.43 In contrast, the kinetics of the elongation phase is independent of the polyQ length.43 Experimental structural information on polyQ aggregates could shed light on the role of polyQ length for the aggregation process. Unfortunately, due to the insolubility of the aggregates, NMR and X-ray structures are so far lacking. Molecular modeling has therefore been the method of choice to have such insights.44-55 Several theoretical models of the aggregated polyQ units, consistent with the available experimental data (electron microscopy and X-ray data)49,56-60 have been proposed. These

10.1021/jp806548p CCC: $40.75  2008 American Chemical Society Published on Web 11/18/2008

16844 J. Phys. Chem. B, Vol. 112, No. 51, 2008

Rossetti et al.

Figure 1. (a) Schematic representation of mutated hungtinin (Mut-Htt). (b) Scheme of the events leading to cell death. The figure is based on studies by Borrell et al. 1

models are based on Perutz’s suggestion that Q side chains, being similar to the backbone units, are able to establish an H-bond net.61,62 Thus, the proposed models are β-sheetcontaining structures stabilized by main and side chains Hbonds. Examples include Atkins’s model in which the H-bond network of the Q side chains allows for high-density packing49 as well as Perutz’s circular β-helix63 and triangular β-helix models,50,64 which are parallel β-sheets held together by main and side chain hydrogen bonds (Figure 2a,c and Figure 3). Molecular simulations of the proposed models have provided valuable insights into their stability,44-53,55 and the dependence of the structural stability on Q number has been partially addressed for the circular β-helix.47,48,50,54 In particular, an interesting report based on the circular β-helix model suggested that the stability of the structure increases with a repeated number of Q, and that above a critical Q number (∼30) the structure of the β-helix is kept stable.48 Unfortunately such conclusions, being based on only one structural model, might depend on the initial structure. Furthermore, as discussed by the authors, the relatively short time-investigated (1 ns) might not allow equilibrating manually built models (as opposed to X-ray or NMR structures). In addition, the key issues on if and how the presence of a number of Qs larger than the disease threshold affects the stability of the polyQ structures in different shapes still need to be considered. Here we address these issues by performing 20 ns long molecular dynamics (MD) simulations on models in aqueous solution featuring a number of Qs well beyond the disease threshold, considering the circular β-helix64 (Figure 2a) and triangular β-helix50 (Figure 2c) models. These models are the only consistent models with the structural threshold hypothesis (the outbreak of the disease above the 36-40 residues was explained on the basis of a structural change of the polyQ chain above this number). Therefore these models have been chosen to validate or discard such hypothesis. Moreover by varying systematically the number as well as the size of the polyQ units in these models, and considering both the monomeric and oligomeric states, our calculations shed light on the dependence of the stability of the β-helix structures upon the number of monomers, independently of the structural model chosen. Computational Details Model Systems. Large Monomeric Models. The coordinates of the circular β-helix nanotube63 (named as P from Perutz, who

Figure 2. PolyQ models undergoing MD simulations in water (not shown here for clarity). (a,c) side and front views of the large monomeric models. These are the circular (P) and triangular (T) β-helix monomers composed by a Q number well above the disease threshold.2 (b,d) The oligomeric models are two series of oligomers (PAD, PAC, PAB, PA, and TAD, TAC, TAB, TA, as well as the oligomer PAH25. (e) The small monomeric models are four monomers in circular β-helix conformation composed by 25, 30, 35, and 40 Qs.

introduced this model in 2002,63 Figure 2a) were kindly provided by Dr. A. Lesk. The structure is characterized by 266 Q residues with φ and Ψ angles of -162 and 159° and each turn contains 20 residues. The triangular β-helix model (named as T, Figure 2c) was constructed starting from the regularly shaped coils from UDPN-acetyl glucosamine acyltransferase64 (Protein Data Bank entry: 1LXA),65 replacing each residue with a glutamine. This model contains 179 residues and each turn is composed by 18 Qs. Both

Polyglutamine Aggregates Investigated via MD Simulations

J. Phys. Chem. B, Vol. 112, No. 51, 2008 16845 The systems were initially relaxed by imposing harmonic position restraints of 1000 Kj/(mol · nm) on solute atoms, allowing the equilibration of the solvent without distorting the solute structure. After an energy minimization of the solvent and the solute without harmonic restraints, the temperature was gradually increased from 0 to 300 K using simulated annealing (SA).75,76 The SA was performed by increasing the temperature from 0 to 300 K in 12 steps in which the temperature was increased by 25 K in 100 ps of MD. Calculated Properties. The root-mean-square deviation (RMSD), the gyration radius (Rg), hydrogen bonds and secondary structure content were calculated using the formulas reported in the Supporting Information. Results and Discussion

Figure 3. Molecular view of circular β-helix structure with particular of (a) external HB net and (b) internal H-bond net. A similar H-bond scheme is observed also for the triangular β-helix structure.

models are composed by a single polyQ chain. Both models have a number of Qs well above the number of polyQs observed at physiological conditions. Oligomeric Models. Starting from the single chain systems we build the following series of oligomers: I. Four oligomers in circular β-helix conformation composed by 4, 3, 2, 1 monomers, respectively. These models are named PAD, PAC, PAB, PA, respectively, and each monomer is composed by 40 Q residues (Figure 2b). II. Four oligomers in triangular β-helix conformation with respectively 4, 3, 2, 1 monomers. These models are symbolized by TAD, TAC, TAB, TA, respectively, and each monomer is composed by 36 Q residues (Figure 2d). III. One oligomer in circular β-helix conformation composed by 8 monomers each containing 25 Q residues (Figure 2b). The model is named PAH25. Small Monomeric Models. We considered 4 monomers in circular β-helix conformation composed by 25, 30, 35, 40 Qs and symbolized by P25, P30, P35, P40, respectively (Figure 2e). The total charge state of all the systems studied is neutral. The oligomeric and the monomeric model are constructed starting from the coordinates of the continuous systems, selecting a number of Q and considering each residue in the zwitterionic form. MD Simulation Protocol. Classical MD simulations of all systems were performed with GROMACS66,67 software in the canonical (NVT) ensemble with a time step of 2 fs for numerical integration. The LINCS algorithm was used to constrain all bond lengths68 and the temperature was kept close to 300 K by a weak coupling to an external bath69,70 with a coupling constant equal to the time step. GROMOS9671 force field was employed, which uses an explicit representation of acidic hydrogen atoms. Long-range electrostatic interactions were treated with the particle mesh Ewald (PME) method,72,73 using a grid with a spacing of 0.12 nm combined with a fourth-order B-spline interpolation to compute the potential and forces in between grid points. The cutoff radius for the Lennard-Jones interactions as well as for the real part of PME calculations was set to 0.9 nm. All systems were solvated with SPC waters74 in a periodic rectangular box large enough to contain the protein and at least 0.9 nm of solvent molecules on each side of the solute.

Predicting the thermodynamics stability of the models considered here (Figure 2) in aqueous solution is currently not feasible with molecular simulations methods. Structural features, instead, can be evaluated by MD simulations. In fact, previous theoretical works have introduced “structural stability” concepts using a variety of criteria, including the conservation of a specific conformation,50 the acquisition of β-strandlike values of backbone dihedral angles54 and the overall number of hydrogen bonds.48 To simplify our discussion we qualitatively introduce the structural stability (SS) as a quantity which increases with (i) the compactness of the structure, as measured by the plots of the RMSD of backbone atoms, as well as the Rg versus time and (ii) the hydrogen bond content (HBC), defined as the total number of H-bonds formed within the structural models, divided by the total number of H-bond donor functionalities. Each quantity on which the SS depends is calculated based on 20 ns MD simulations for each model considered. We used SS to compare systematically the polyQ stability as a function of shape, number of monomers and number of Qs in a given monomer. The dependence of SS on the shape is investigated in the two large monomeric models of circular (P) and triangular (T) β-helix reported in Figure 2a,c. These feature a number of Qs (266 Qs for P and 179 Qs for T) well above the disease threshold.2 This number of Qs has never been observed in patients suffering from HD. Therefore, these models are also used as reference for the other systems considered here. The dependence of the SS on the different number of monomers is studied in polymeric structures within the same shape (either circular or triangular shape), but with a different number of monomers, from 4 (PAD and TAD) to 1 (PA and TA) (Figure 2b,d) and composed by short monomers of 40 and 36 Qs, respectively. Finally, the dependence of the SS on the number of Qs per monomer is investigated in small monomeric models in circular β-helix shape, composed by a number of Qs below and above the disease threshold (Figure 2e). In most cases, the quantities on which SS depends turn out to fluctuate around an average value after few nanoseconds (see Supporting Information). Few models, specified in the following text, turn out not to equilibrate in the time scale investigated. Obviously, for those systems, averages cannot be taken. However, qualitative comparisons can still be made (see, for instance, ref 77). Large Monomeric Models. In our 20 ns MD runs, the RMSD of these models (P and T) fluctuate around average values of 0.18 and 0.25 nm after 0.25 and 1.0 ns, respectively (Table 1 and Supporting Information, Figure SI.1). The larger

16846 J. Phys. Chem. B, Vol. 112, No. 51, 2008

Rossetti et al.

TABLE 1: SS of All the Systems Studied Reported in Terms of Average Values of RMSDs, Rg, and HBC; βSC is also reporteda

system

RMS (nm)

∆σ (nm)

Rg (nm)

∆σ (nm)

P T PAD PAC PAB PA TAD TAC TAB a TA PAH25 P40 P35 P30 b P25

0.18 0.25 0.27 0.15 0.22 0.36 0.19 0.20 0.17 0.30 0.23 0.46 0.61 0.58 0.65

0.02 0.02 0.03 0.03 0.03 0.04 0.03 0.02 0.03

2.11 1.72 1.52 1.39 1.25 1.05 1.51 1.28 1.11 0.95 1.47 1.05 0.98 0.99 0.95

0.02 0.01 0.02 0.01 0.01 0.02 0.01 0.01 0.01

0.02 0.04 0.06 0.06

0.01 0.01 0.02 0.03

HBC system (% of donors)

∆σ (% of donors)

HBC main chain (% of donors)

∆σ (% of donors)

HBC side chain (% of donors)

∆σ (% of donors)

β-sheet content (% of res)

70.56 69.30 65.47 56.08 48.25 48.81 68.22 67.03 62.47 52.01 67.75 48.81 48.26 44.41 40.60

2.93 2.81 3.55 3.08 3.49 5.66 2.83 3.49 4.07

65.34 50.57 50.91 48.99 40.69 24.07 44.35 44.81 40.33 23.87 52.59 24.07 16.51 21.63 16.49

3.10 3.09 3.67 3.61 4.34 4.94 3.24 3.81 4.08

60.49 48.26 54.57 48.07 44.57 40.12 47.64 44.70 42.38 35.45 52.71 40.12 34.45 27.18 24.94

4.87 3.99 5.43 4.78 5.16 8.19 4.13 4.70 5.91

81.54 41.30 60.51 60.15 20.43 23.26 40.50 36.89 20.14 14.50 59.01 34.8 17.8 17.9 11.90

3.35 5.66 5.67 6.87

3.66 4.94 4.48 5.78

5.12 8.19 7.89 8.42

∆σ (% of res) 3.82 5.42 6.01 5.11 7.83 7.74 5.00 5.30 6.77 6.02 7.60 9.83 9.94

a The selected quantities are also reported in the following figures: the HBC and βSC of T and P (Figure 4) and the HBC of P, T series and monomers (Figure 5). The properties on which SS depends and βSC of PAH25 are displayed in Supporting Information (Fig. SI.8). b Systems which do not equilibrate in the time scale investigated. Values are however taken from the last snapshot of the dynamics, as the systems are not equilibrated in the 20 ns of dynamics (see Supporting Information).

Figure 4. Percentage of residues in β-sheet (green line) and in random coil (red line) conformations in the large monomeric models considered here (See Figure 1): the circular and the triangular β-helices in (a) and (b), respectively. Percentage of donor atoms involved in H-bonds in the whole protein (black line), in the main (blue line), and in the side chains (violet line) for the circular and the triangular β-helices in (c) and (d), respectively.

the RMSD and the longer equilibration time of T, relative to that of P, might be caused at least in part by the fact that T is a purely theoretical model, while P is a model based on X-ray data interpretation.63 Rg decreases by few percent values during the dynamics, indicating that the MD models are more compact than the initial structures (Supporting Information, Figure SI.2). The HBC of P and T is ∼70% (Table 1, Figure 4). The contributions of backbone and side-chains are similar (Table 1, Supporting Information, Figure SI.3). However, the β-sheet content (βSC) of P is large, while that of T is just above 50% (Table 1, Figure 4; see also Supporting Information, Figure SI.4). (The βSC is

calculated here using the definition of W. Kabsch and C. Sander78.) Thus, the stability of the two helices, which has been related to their HBC, seems not to be necessarily associated with a large content of secondary structure elements. Oligomeric Models. For both the P and T series, the RMSD increases by decreasing the monomers from four to one (Table 1, Supporting Information, Figure SI.6). The Rg values of the oligomers which feature four and three monomers (Table 1, Supporting Information, Figure SI.7) (PAD, PAC, TAD, TAC), are similar to those of P and T (Supporting Information, Figure SI.7). Those with two or one monomers

Polyglutamine Aggregates Investigated via MD Simulations

J. Phys. Chem. B, Vol. 112, No. 51, 2008 16847

Figure 5. Percentage of H-bonds in P series (first column), T series (second column), and monomeric series (third column). Blue histograms represent the total HBC, green and red ones represent main chain and side chain HBC, respectively. Asterisk (*) designates systems that do not equilibrate in the time scale investigated. Values are taken from the last snapshot of the dynamics, as the systems are not equilibrated in the 20 ns of dynamics (see Supporting Information).

(PAB, PA and TAB, TA) do not preserve compact folds (see Supporting Information). The HBC of PAD, TAD is not much smaller than those of P and T. Those featuring three monomers are still significantly large. However, the values decrease rapidly by decreasing the number of monomers (Table 1, Figure 5). In all circumstances, the largest contribution to HBC is due mainly to the backbone rather than to the side-chains. The secondary structure content follows the same trend of the HBC (Figure 6). Thus, the SS of the assemblies of four and three monomers are similar to those of the single chain systems P and T. This means that in the short monomers forming the oligomeric systems the Qs interact with each other almost as they do in the large monomeric models. This result does not depend on the shape, as PAD, PAC, TAD, TAC behave similarly. However, the SS of PAB, TAB dimers are much smaller than that of T, P, and this feature is even more pronounced in the case of the monomers PA and TA. Also in these cases the SS does not depend on the shape. Therefore, the SS decreases with a decreasing number of monomers, independently of the shapes. Because SS does not depend on shape and T features a different number of glutamines with respect to P, we suggest that the stability of oligomeric structures may not depend on the number of Qs in each monomer. To test this hypothesis, we construct an oligomer equivalent to PAD (i.e., PAH25), except that each monomer features a smaller number of Qs. This is a circular β-helix oligomer composed by 8 monomers, each 25 Q long (model PAH25 in Figure 2b). The values of the quantities on which SS depends turn out to be similar to those of PAD

Figure 6. βSC (green) and random coil conformation (red) of oligomers in circular (left panel) and triangular (right panel) β-helix conformations.

(Supporting Information, Figure SI.8), corroborating the hypothesis that SS is not influenced by the number of Qs composing each monomer.

16848 J. Phys. Chem. B, Vol. 112, No. 51, 2008

Figure 7. βSC (green) and random coil (red) conformations of P40, P35, P30, P25 monomers (see text). On the left and right side of the graph the initial and final (after 20 ns MD) geometries of each monomer are shown. Water is not shown for the sake of clarity.

We conclude that SS depends on the number of monomers independently of their shape and of their length (number of Qs in each monomer). This is consistent with NMR data of Klein et al.41 showing that aggregates formed by short and long polyQ tracts share similar structural properties. Small Monomeric Models. We finally investigate the dependence of the SS on the length, focusing only on one shape, because, as seen in the previous section, SS turns out not to depend significantly on the latter. We focus here on the circular β-helix shape and we consider monomers of lengths below and above the disease threshold by choosing four systems containing 25, 30, 35, and 40 Qs (Figure 2e). During the MD, P25 undergoes the largest structural rearrangements among the models investigated here. Its initial shape is lost after about 10 ns, as shown by a large increase of the RMSD and Rg values (Supporting Information, Figures SI.9, SI.10). The final MD structure is shown in Figure 7. In the other small monomeric models, the RMSD decreases by increasing the number of Qs (Table 1). In fact a similar trend, but with smaller RMSD and Rg values, is found for P30 (Supporting Information, Figures SI.9, SI.10). Remarkably, the RMSD and Rg of P40 and P35 are similar to those of P (Supporting Information, Figures SI.9, SI.10), preserving the original β-helix fold. The overall HBC increases with the number of Qs (Figure 4). Such an increase is caused by the side chains H-bonds, while the backbone contribution is similar in the four models (Figure 4). This result is opposite to that we found for the oligomers in which HBC mainly depends on the H-bonds between backbone atoms (Figure 5). Finally, the βSC of P40 is much larger than that of the other three models (Figure 7). On the basis of these results, we conclude that (i) the SS increases progressively with the number of Qs and (ii) SS is associated to an increase of side-chain H-bonds. This suggests that the SS of single monomers depends on the possibility of forming intraside-chains H-bonds.

Rossetti et al. Relevance of Our Results for the Toxicity Threshold Proposals. In a widely accredited hypothesis, the polyQ length affects the nucleated-grow polymerization kinetics of polyQ aggregation, rather than the stability of the aggregates.41 At the molecular level, the process would start from a random coil to β-sheet structural transition of an individual monomer, which constitutes the aggregation seed. 20,28,43 Our findings are consistent with this hypothesis, in fact, HBC and βSC of single isolated monomers increase with the number of Qs. Assuming that of HBC and βSC in the transition state leading to the formation of the initial aggregation seed have a similar relevance to that found here for the final products (the aggregates), we formulate the hypothesis that the overall aggregation kinetics may increase upon incrementing the number of Q residues. Although reasonable, this proposal, at present, can be discussed only at a speculative level. In addition, the HBC and the βSC of the oligomers depend on the number of monomers and not on the number of Qs in each monomer. This is consistent with the hypothesis that once the initial polyQ seed is formed, the kinetics of the elongation phase does not depend on the number of Q.20,43 Limitations of the Calculations. As any modeling investigation, this study has several limitations. First, the time scale, although longer than that reported in previous studies on folded polyQ peptides,48 is too short to follow the time evolution of models TA and P25, (while the other models appear to be well equilibrated (Table and Supporting Information, Figure SI.1, SI.6, SI.9)). Thus, any conclusion based on comparisons among the models is necessarily qualitative. Second, the aggregates have been experimentally characterized at several different ionic strengths, ranging from 20 to 200 mM.79 Although some of the properties may be affected by varying ion concentrations, we expect that our conclusions on the folded peptides (which have been reported at zero ionic strength) will not change at the qualitative level. In addition, we stress here that no attempt has been made to study the condensation process of the peptides, which are likely to depend dramatically on the simulation conditions. Finally, our conclusions have been drawn for a subset of proposed models (not real structures). Again, this fact prevents to make any quantitative statement, which is beyond the scope of this paper. Concluding Remarks On the basis of our 20 ns MD simulations with GROMOS96 force field for each systems studied (Figure 2), we suggest the following qualitative conclusions: (i) The two different β-helix shapes, originally invoked to explain the disease threshold, affect only the β-sheet content; the T helix displays a larger number of residues in random coil conformation than the P helix. However, the H-bond content as well as the SS of the two shapes is comparable. (ii) The structural stability of P and T oligomeric systems is not affected by the shape. This has never been shown so far. In addition, SS does not depend on the number of Qs in the monomers. Although a recent theoretical study on polyQs suggests that the stability of an aggregate should be regulated by the structural stability of its component monomers, as stated by the authors, the short time-scale investigated (1 ns of MD simulations with AMBER force field80,81) appears not to be sufficient to draw general conclusions.48 Indeed, we demonstrate that for longer simulations, the oligomer with 4 monomers of 40 Qs and the oligomer with 8 monomer of 25 Qs have similar SS. Consistently with our results a recent NMR analysis reveals

Polyglutamine Aggregates Investigated via MD Simulations no structural difference between aggregates formed by short and long polyQ peptides.41 In summary, our results do not support the structural threshold hypothesis, which suggests a specific conformation for polyQs longer than the threshold. (iii) Only the number of monomers (thus the concentration in an (in vivo or in vitro) experiment) contributes to the overall stability of the oligomers because of the additive contribution of single monomer in the H-bond net formed between backbone atoms. (iv) The H-bonds formed between Q side-chains influence mainly the stability of the single isolated monomer and to a lesser extent the stability of oligomers. (v) Interpreting our findings on the basis of the whole landscape of available experimental data,41 we suggest that the observed length-dependent toxicity threshold may be explained by a faster aggregation kinetics that occurs for longer polyQ tracts. Acknowledgment. The authors thank Dr. A. Lesk for providing the structural model of the circular β-helix. Supporting Information Available: Graphs representing the time evolution of RMSD, Rg, secondary structure content and hydrogen bond content for all the systems. The mean values of these properties are also reported in Table 1. A study of the flexibility based on root mean square fluctuation (RMSF) of the large monomeric models is also reported. This material is available free of charge via the Internet at http://pubs.acs.org. References and Notes (1) Borrell-Pagès, M.; Zala, D.; Humbert, S.; Saudou, F. Cell. Mol. Life Sci. 2006, 63, 2642. (2) Trettel, F.; Rigamonti, D.; Hilditch-Maguire, P.; Wheeler, V. C.; Sharp, A. H.; Persichetti, F.; Cattaneo, E.; MacDonald, M. E. Hum. Mol. Genet. 2000, 9, 2799. (3) DiFiglia, M.; Sapp, E.; Chase, K. O.; Davies, S. W.; Bates, G. P.; Vonsattel, J. P.; Aronin, N. Science 1997, 277, 1990. (4) Sapp, E.; Penney, J.; Young, A.; Aronin, N.; Vonsattel, J. P.; DiFiglia, M. J. Neuropathol. Exp. Neurol. 1999, 58, 165. (5) Grunewald, T.; Beal, M. F. Ann. N.Y. Acad. Sci. 1999, 893, 203. (6) Gutekunst, C. A.; Li, S. H.; Yi, H.; Mulroy, J. S.; Kuemmerle, S.; Jones, R.; Rye, D.; Ferrante, R. J.; Hersch, S. M.; Li, X. J. J. Neurosci. 1999, 19, 2522. (7) Li, H.; Li, S. H.; Johnston, H.; Shelbourne, P. F.; Li, X. J. Nat. Genet. 2000, 25, 385. (8) Li, X. J. Mol Neurobiol. 1999, 20, 111. (9) Velier, J.; Kim, M.; Schwarz, C.; Kim, T. W.; Sapp, E.; Chase, K.; Aronin, N.; DiFiglia, M. Exp. Neurol. 1998, 152, 34. (10) Benchoua, A.; Trioulier, Y.; Zala, D.; Gaillard, M. C.; Lefort, N.; Dufour, N.; Saudou, F.; Elalouf, J. M.; Hirsch, E.; Hantraye, P.; Deglon, N.; Brouillet, E. Mol. Biol. Cell 2006, 17, 1652. (11) Panov, A. V.; Gutekunst, C. A.; Leavitt, B. R.; Hayden, M. R.; Burke, J. R.; Strittmatter, W. J.; Greenamyre, J. T. Nat. Neurosci. 2002, 5, 731. (12) Bezprozvanny, I.; Hayden, M. R. Biochem. Biophys. Res. Commun. 2004, 322, 1310. (13) Bizat, N.; Hermel, J. M.; Humbert, S.; Jacquard, C.; Creminon, C.; Escartin, C.; Saudou, F.; Krajewski, S.; Hantraye, P.; Brouillet, E. J. Biol. Chem. 2003, 278, 43245. (14) Bizat, N.; Hermel, J. M.; Boyer, F.; Jacquard, C.; Creminon, C.; Ouary, S.; Escartin, C.; Hantraye, P.; Krajewski, S.; Brouillet, E. J. Neurosci. 2003, 23, 5020. (15) Lunkes, A.; Lindenberg, K. S.; Ben-Haiem, L.; Weber, C.; Devys, D.; Landwehrmeyer, G. B.; Mandel, J. L.; Trottier, Y. Mol. Cell 2002, 10, 259. (16) Bevivino, A. E.; Loll, P. J. Proc. Nat. Acad. Sci. U.S.A. 2001, 98, 11955. (17) Tanaka, M.; Morishima, I.; Akagi, T.; Hashikawa, T.; Nukina, N. J. Biol. Chem. 2001, 276, 45470. (18) Scherzinger, E.; Lurz, R.; Turmaine, M.; Mangiarini, L.; Hollenbach, B.; Hasenbank, R.; Bates, G. P.; Davies, S. W.; Lehrach, H.; Wanker, E. E. Cell 1997, 90, 549.

J. Phys. Chem. B, Vol. 112, No. 51, 2008 16849 (19) Cooper, J. K.; Schilling, G.; Peters, M. F.; Herring, W. J.; Sharp, A. H.; Kaminsky, Z.; Masone, J.; Khan, F. A.; Delanoy, M.; Borchelt, D. R.; Dawson, V. L.; Dawson, T. M.; Ross, C. A. Hum. Mol. Genet. 1998, 7, 783. (20) Chen, S. M.; Berthelier, V.; Hamilton, J. B.; O’Nuallain, B.; Wetzel, R. Biochem 2002, 41, 7391. (21) Davies, S. W.; Turmaine, M.; Cozens, B. A.; DiFiglia, M.; Sharp, A. H.; Ross, C. A.; Scherzinger, E.; Wanker, E. E.; Mangiarini, L.; Bates, G. P. Cell 1997, 90, 537. (22) Scherzinger, E.; Sittler, A.; Schweiger, K.; Heiser, V.; Lurz, R.; Hasenbank, R.; Bates, G. P.; Lehrach, H.; Wanker, E. E. Proc. Nat. Acad. Sci. U.S.A 1999, 96, 4604. (23) Temussi, P. A.; Masino, L.; Pastore, A. Embo J. 2003, 22, 355. (24) Ross, C. A.; Poirier, M. A. Nature Med. 2004, , S10. (25) Kayed, R.; Head, E.; Thompson, J. L.; McIntire, T. M.; Milton, S. C.; Cotman, C. W.; Glabe, C. G. Science 2003, 300, 486. (26) Masino, L.; Kelly, G.; Leonard, K.; Trottier, Y.; Pastore, A. FEBS Lett. 2002, 513, 267. (27) Bennett, M. J.; Huey-Tubman, K. E.; Herr, A. B.; West, A. P.; Ross, S. A.; Bjorkman, P. J. Proc. Nat. Acad. Sci. U.S.A. 2002, 99, 11634. (28) Chen, S.; Berthelier, V.; Yang, W.; Wetzel, R. J. Mol. Biol. 2001, 311, 173. (29) Wellington, C. L.; Ellerby, L. M.; Hackam, A. S.; Margolis, R. L.; Trifiro, M. A.; Singaraja, R.; McCutcheon, K.; Salvesen, G. S.; Propp, S. S.; Bromm, M.; Rowland, K. J.; Zhang, T. Q.; Rasper, D.; Roy, S.; Thornberry, N.; Pinsky, L.; Kakizuka, A.; Ross, C. A.; Nicholson, D. W.; Bredesen, D. E.; Hayden, M. R. J. Biol. Chem. 1998, 273, 9158. (30) Welch, W. J.; Diamond, M. I. Hum. Mol. Genet. 2001, 10, 3063. (31) Mangiarini, L.; Sathasivam, K.; Seller, M.; Cozens, B.; Harper, A.; Hetherington, C.; Lawton, M.; Trottier, Y.; Lehrach, H.; Davies, S. W.; Bates, G. P. Cell 1996, 87, 493. (32) Zoghbi, H. Y.; Orr, H. T. Annu. ReV. Neurosci. 2000, 23, 217. (33) Stine, O. C.; Pleasant, N.; Franz, M. L.; Abbott, M. H.; Folstein, S. E.; Ross, C. A. Hum. Mol. Genet. 1993, 2, 1547. (34) Rubinsztein, D. C.; Leggo, J.; Coles, R.; Almqvist, E.; Biancalana, V.; Cassiman, J. J.; Chotai, K.; Connarty, M.; Craufurd, D.; Curtis, A.; Curtis, D.; Davidson, M. J.; Differ, A. M.; Dode, C.; Dodge, A.; Frontali, M.; Ranen, N. G.; Stine, O. C.; Sherr, M.; Abbott, M. H.; Franz, M. L.; Graham, C. A.; Harper, P. S.; Hedreen, J. C.; Jackson, A.; Kaplan, J. C.; Losekoot, M.; MacMillan, J. C.; Morrison, P.; Trottier, Y.; Novelletto, A.; Simpson, S. A.; Theilmann, J.; Whittaker, J. L.; Folstein, S. E.; Ross, C. A.; Hayden, M. R. Am. J. Hum. Genet. 1996, 59, 16. (35) Masino, L.; Pastore, A. Biochem. Soc. Trans. 2002, 30, 548. (36) Georgalis, Y.; Starikov, E. B.; Hollenbach, B.; Lurz, R.; Scherzinger, E.; Saenger, W.; Lehrach, H.; Wanker, E. E. Proc. Nat. Acad. Sci. U.S.A. 1998, 95, 6118. (37) Hollenbach, B.; Scherzinger, E.; Schweiger, K.; Lurz, R.; Lehrach, H.; Wanker, E. E. Philos. Trans. R. Soc. London, Ser. B 1999, 354, 991–994. (38) Trottier, Y.; Lutz, Y.; Stevanin, G.; Imbert, G.; Devys, D.; Cancel, G.; Saudou, F.; Weber, C.; David, G.; Tora, L.; Agid, Y.; Brice, A.; Mandel, J. L. Nature 1995, 378, 403. (39) Li, P. W.; Huey-Tubman, K. E.; Gao, T. Y.; Li, X. J.; West, A. P.; Bennett, M. J.; Bjorkman, P. J. Nat. Struct. Mol. Biol. 2007, 14, 568. (40) Sugaya, K.; Matsubara, S.; Kagamihara, Y.; Kawata, A.; Hayashi, H. PLoS One 2007, 2, e635. (41) Klein, F. A. C.; Pastore, A.; Masino, L.; Zeder-Lutz, G.; Nierengarten, H.; Oulad-Abdeighani, M.; Altschuh, D.; Mandel, J. L.; Trottier, Y. J. Mol. Biol. 2007, 371, 235. (42) Morley, J. F.; Brignull, H. R.; Weyers, J. J.; Morimoto, R. I. Proc. Nat. Acad. Sci. U.S.A. 2002, 99, 10417. (43) Chen, S. M.; Ferrone, F. A.; Wetzel, R. Proc. Nat. Acad. Sci. U.S.A. 2002, 99, 11884. (44) Marchut, A. J.; Hall, C. K. Comput. Biol. Chem. 2006, 30, 215. (45) Marchut, A. J.; Hall, C. K. Biophys. J. 2006, 90, 4574. (46) Marchut, A. J.; Hall, C. K. Proteins: Struct., Funct., Bioinf. 2007, 66, 96. (47) Merlino, A.; Esposito, L.; Vitagliano, L. Proteins: Struct., Funct., Bioinf. 2006, 63, 918. (48) Ogawa, H.; Nakano, M.; Watanabe, H.; Starikov, E. B.; Rothstein, S. M.; Tanaka, S. Comput. Biol. Chem. 2008, 32, 102. (49) Sikorski, P.; Atkins, E. Biomacromolecules 2005, 6, 425. (50) Stork, M.; Giese, A.; Kretzschmar, H. A.; Tavan, P. Biophys. J. 2005, 88, 2442. (51) Zanuy, D.; Gunasekaran, K.; Lesk, A. M.; Nussinov, R. J. Mol. Biol. 2006, 358, 330. (52) Armen, R. S.; Bernard, B. M.; Day, R.; Alonso, D. O. V.; Daggett, V. Proc. Nat. Acad. Sci. U.S.A. 2005, 102, 13433. (53) Finke, J. M.; Onuchic, J. N. Biophys. J. 2004, 86, 340A. (54) Khare, S. D.; Ding, F.; Gwanmesia, K. N.; Dokholyan, N. V. PLoS Comput. Biol. 2005, 1, 230.

16850 J. Phys. Chem. B, Vol. 112, No. 51, 2008 (55) Esposito, L.; Paladino, A.; Pedone, C.; Vitagliano, L. Biophys. J. 2008, 94, 4031. (56) Perutz, M. F.; Pope, B. J.; Owen, D.; Wanker, E. E.; Scherzinger, E. Proc. Nat. Acad. Sci. U.S.A. 2002, 99, 5596. (57) Perutz, M. F.; Johnson, T.; Suzuki, M.; Finch, J. T. Proc. Nat. Acad. Sci. U.S.A 1994, 91, 5355. (58) Sunde, M.; Blake, C. AdV. Protein Chem. 1997, 50, 123. (59) Sunde, M.; Serpell, L. C.; Bartlam, M.; Fraser, P. E.; Pepys, M. B.; Blake, C. C. F. J. Mol. Biol. 1997, 273, 729. (60) Sharma, D.; Shinchuk, L. M.; Inouye, H.; Wetzel, R.; Kirschner, D. A. Proteins: Struct., Funct., Bioinf. 2005, 61, 398. (61) Perutz, M. Protein Sci. 1994, 3, 1629. (62) Perutz, M. F. Trends Biochem. Sci. 1999, 24, 58. (63) Perutz, M. F.; Finch, J. T.; Berriman, J.; Lesk, A. Proc. Nat. Acad. Sci. U.S.A 2002, 99, 5591. (64) Raetz, C. R. H.; Roderick, S. L. Science 1995, 270, 997. (65) Berman, H. Biophys. J. 2000, 78, 267A. (66) van der Spoel, D.; van Drunen, R.; Berendsen, H. J. C. GROningen MAchine for Chemical Simulation; BIOSON Research Institute, Nijenborgh: Groningen, 1994. (67) Berendsen, H. J. C.; Vanderspoel, D.; Vandrunen, R. Comput. Phys. Commun. 1995, 91, 43. (68) Hess, B.; Bekker, H.; Berendsen, H. J. C.; Fraaije, J. G. E. M. J. Comput. Chem. 1997, 18, 1463. (69) Berendsen, H. J. C.; Postma, J. P. M.; Vangunsteren, W. F.; Dinola, A.; Haak, J. R. J. Chem. Phys. 1984, 81, 3684.

Rossetti et al. (70) Berensen, H. J. C. Transport Properties computed by Linear Response through weak coupling to a bath. In Computer Simulation in Materials Science; Kluwer Academic Publishers: Norwell, MA, 1991. (71) van Gunsteren, W. F.; Billeter, S. R.; Eising, A. A.; Hunenberger, P. H.; Kruger, P.; Mark, A. E.; Scott, W. R. P.; Tironi, I. G. Biomolecular simulation: the GROMOS96 manual and user guide; Hochschulverlag AG an der ETH: Zurich, Switzerland, 1996. (72) Darden, T.; York, D.; Pedersen, L. J. Chem. Phys. 1993, 98, 10089. (73) Essmann, U.; Perera, L.; Berkowitz, M. L.; Darden, T.; Lee, H.; Pedersen, L. G. J. Chem. Phys. 1995, 103, 8577. (74) Berendsen, H. J. C.; Grigera, J. R.; Straatsma, T. P. J. Phys. Chem. 1987, 91, 6269. (75) Moskowitz, J. W.; Schmidt, K. E.; Wilson, S. R.; Cui, W. Int. J. Quantum Chem. 1988, 611. (76) Nilges, M.; Clore, G. M.; Gronenborn, A. M. FEBS Lett. 1988, 239, 129. (77) Berrera, M.; Pantano, S.; Carloni, P. Biophys. J. 2006, 90, 3428. (78) Kabsch, W.; Sander, C. Biopolymers 1983, 22, 2577. (79) Masino, L.; Musi, V.; Menon, R. P.; Fusi, P.; Kelly, G.; Frenkiel, T. A.; Trottier, Y.; Pastore, A. FEBS Lett. 2003, 549, 21. (80) Wang, J. M.; Cieplak, P.; Kollman, P. A. J. Comput. Chem. 2000, 21, 1049. (81) Case, D. A.; Cheatham, T. E.; Darden, T.; Gohlke, H.; Luo, R.; Merz, K. M.; Onufriev, A.; Simmerling, C.; Wang, B.; Woods, R. J. J. Comput. Chem. 2005, 26, 1668.

JP806548P