Redox Properties of Met35 in Neurotoxic Î²-Amyloid Peptide. A

408

Chem. Res. Toxicol. 2002, 15, 408-418

Redox Properties of Met35 in Neurotoxic β-Amyloid Peptide. A Molecular Modeling Study Dariusz Pogocki†,‡ and Christian Scho¨neich*,† Department of Pharmaceutical Chemistry, University of Kansas, 2095 Constant Avenue, Lawrence, Kansas 66047, and Institute of Nuclear Chemistry and Technology, Dorodna 16, 03-195 Warsaw, Poland Received September 25, 2001

The β-amyloid peptide (βAP) is the principal component of plaque associated with the pathology of Alzheimer’s disease. Part of its neurotoxicity appears to correlate with the ability of the peptide to reduce CuII and form free radicals. Both processes are dependent on the presence and oxidizability of Met35 in the C-terminus of the peptide but no mechanistic details on the reactions leading to Met oxidation are known. On the basis of previous studies with model peptides, we hypothesize that a one-electron oxidation of Met35 in βAP is facilitated through a neighboring group effect. Complexed to CuII and/or in a lipid-mimicking environment, the solution structure of βAP includes a large R-helical part. The solution NMR structure of βAP1-40 in aqueous SDS micelles reveals an R-helix between residues 27 and 36, containing Met35. In this helical C-terminus of βAP, the peptide bond CdO group C-terminal of Ile31 is located very close to the Met35 sulfur and could stabilize a Met35 sulfide radical cation through formation of an (S-O) three-electron bond. In the present paper, we have computationally validated this hypothesis using Langevin dynamics methods to determine the collision frequency of the Met35 thioether sulfur and the oxygen atoms of several peptide bonds in the βAP sequence. Nanosecond time scale computations were carried out for four distinct βAP congeners, βAP2640, βAP26-36, βAP26-40(Ile31Pro), βAP40-26, and their respective Met35-sulfur-centered cation radicals. Here, βAP26-40, βAP26-40(Ile31Pro) and βAP40-26 are representative fragments of the full length βAP1-42 or βAP42-1 sequence, respectively, whereas βAP2636 represents a unique βAP sequence for which biological data are available. Initial structures of βAP26-40, βAP26-40(Ile31Pro), and βAP26-36 were selected to be identical to that of the βAP26-40 or βAP26-36 sequence in full-length βAP1-40. As the structures of βAP40-26 and βAP42-1 are not known, various initial conformations such as R-helix and antiparallel β-sheet were selected for βAP40-26. Our computational results show that βAP26-40, representative for the same sequence in full-length βAP1-42, has the highest tendency to form (S-O) bonds between Ile31CdO and Met35S. We conclude that native βAP1-42 has a higher tendency to support Met35 oxidation through (S-O) bond formation, consistent with the experimental observation that βAP1-42 is more neurotoxic compared to the other investigated sequences.

Introduction The formation and aggregation of β-amyloid peptide, βAP, is a key to the formation of neurotoxic amyloid deposits associated with the pathogenesis of Alzheimer’s disease (AD) (1-3). The major βAP sequence circulating in the cerebrospinal fluid is βAP1-40 while the major sequence identified in senile plaques is βAP1-42 (4, 5). At present, the detailed molecular mechanisms of βAP neurotoxicity are not understood (3). One important experimental observation is that neurotoxicity may be caused by the βAP-dependent formation of free radicals and/or reactive oxygen species (ROS) (6-10). In fact, AD brain is characterized by extensive oxidative stress (1113), central to the pathogenesis of this disorder (3, 10, 14, 15). Hensley et al. reported the apparent “spontaneous” oxidation and fragmentation of βAP1-40 in aqueous buffer, paralleled by the formation of free radicals (7), although no detailed mechanisms have been characterized (2, 7, 16-18). More recently, the high propensity of * To whom correspondence should be addressed. E-mail: schoneic@ ukans.edu. Phone: (785) 864-4880. Fax: (785) 864-5736. † University of Kansas. ‡ Institute of Nuclear Chemistry and Technology.

full length βAP to reduce peptide-bound CuII has been recognized (19, 20). Importantly, neither the C-terminal truncated sequence βAP1-28 nor the N-terminal truncated sequence βAP25-35 reduces CuII. These data can be interpreted such that CuII is bound to the His residues at the N-terminus whereas the electron for reduction originates from the C-terminal Met residue (19-21). In support of this interpretation, βAP1-42 containing either methionine sulfoxide (MetO) or Nle at position 35 does 0 not reduce CuII (19). The difference between ECu I/IIβAP ) 0.5-0.55 V (20) and the peak potential of anodic Met oxidation, 1.5 V (22), is equal to ca. 1.0 V. Such difference in the potentials would normally ensure that equilibrium 1 is located far on the left-hand side. k1

MetS + CuII y\ z MetS•+ + CuI k -1

(1)

On the other hand, if both reaction products, MetS•+ and CuI, are efficiently removed from equilibrium 1, it may be shifted to the right-hand side. Evidence for thermodynamically unfavorable electron-transfer steps driven by subsequent reactions exists in the organic chemistry literature, e.g., the oxidation of p-xylene by CeIV (23). The

10.1021/tx0101550 CCC: $22.00 © 2002 American Chemical Society Published on Web 02/13/2002

Redox Properties of Met35

Chem. Res. Toxicol., Vol. 15, No. 3, 2002 409

O2-dependent formation of H2O2 during the incubation of βAP1-42 (20) suggests that CuI is removed from equilibrium 1, most likely via intermediary CuII/superoxo complexes (reaction 2) which are common intermediates in many CuI/O2 reactions (24-28). k2

CuI + O2 y\ z CuII‚‚‚O2•k -2

(2)

However, there are several pathways which may affect equilibrium 1 by promoting the formation of MetS•+, and these are the focus of this paper. In general, the redox processes of organic sulfides are affected by neighboring groups which can kinetically and thermodynamically stabilize sulfide radical cations (such as MetS•+) through the formation of radical cationnucleophile complexes (29-33). This is displayed in the general reaction 3 (L ) H or organic ligands) where X represents the heteroatoms S, Se, Te, O (m ) 2, n ) 0), N, P (m ) 3, n ) 0), or Cl, Br, and I (m ) 0, n ) 1). k3

z [R2S-XLm](1-n)+ R2S+• + (LmX)n- y\ k -3

In the present work, we have used molecular mechanics calculations to model and quantitate the interaction of the Met35 sulfur with several amide bonds in various βAP sequences. A favorable interaction between the Met35 sulfur and any amide carbonyl group would ensure a more facile one-electron oxidation of Met35 and, therefore, a higher tendency for the βAP sequence to produce reactive oxygen species. In contrast, a less favorable interaction of the Met35 sulfur with the amide oxygen would likely result in a more stable and less neurotoxic sequence. Molecular mechanics studies were performed for four βAP-peptide congeners, βAP26-40 (P1), βAP2640(Ile31Pro) (P2), βAP40-26 (P3), and βAP26-36 (P4) (shown below), and their cation radicals located on the sulfur of Met35, P1(S•+), P2(S•+), P3(S•+), and P4(S•+).

P1: CH3-CO-NH-Ser26-Asn27-Lys28-Gly29-Ala30-Ile31Ile32-Gly33-Leu34-Met35-Val36-Gly37-Gly38-Val39Val40-CONH2 P2:CH3-CO-NH-Ser26-Asn27-Lys28-Gly29-Ala30-Pro31-

(3)

For Met and its sulfide radical cation, MetS•+, such interactions may lower the reduction potential, enhance k1, and/or lower k-1. The only nucleophile in the immediate environment of Met35 in βAP is the peptide amide bond. Recently, we have reported experimental evidence for sulfide radical cation-amide association during the one-electron oxidation of Met in the model compound N-acetylmethionine amide (32), supporting that such mechanism may promote Met oxidation in βAP. The solution structure of βAP1-40 in aqueous sodium dodecyl sulfate micelles shows an R-helix between residues 27 and 36, which includes Met35 (34). This structure is representative for our calculations in a lipid-mimicking environment. NMR studies of the more water-soluble subsequence βAP10-35 show no evidence for an R-helical structure in water (35). However, recent results by Curtain et al. suggest that association between Cu2+ and βAP1-42 causes R-helix formation (up to 57%), enabling the peptide to assemble in helical multimers, which can insert into membranes (36). Hence, during the reduction of peptide-bound Cu2+, conditions involving the electron transfer from Met, we can reasonable assume that the C-termini of βAP1-40 and βAP1-42 exist in R-helical conformation even in water, important for our calculations in this medium. In the helical C-terminus of βAP140 (34) the stabilization of oxidized Met35 through association with the CdO group of the peptide bond C-terminal to Ile31 may be very favorable, as the ca. 3.6 Å average S-O distance between Met35 and Ile31-CdO in the energy optimized structures (34) is close to the sum of the van der Waals radii of the two atoms (37). Chemically, such preexisting bond formations have been shown to promote the one-electron oxidation reactions in model compounds (33), and analogous mechanisms may assist the oxidation of βAP. Unfortunately, the direct experimental detection of (S-O)-bonded structures in βAP is difficult by time-resolved spectroscopy due to the low solubility of these peptides. However, mechanistic details can be obtained through molecular modeling using parameters for sulfide radical cations derived from organic model compounds.

Ile32-Gly33-Leu34-Met35-Val36-Gly37-Gly38-Val39Val40-CONH2 P3: CH3-CO-NH-Val40-Val39-Gly38-Gly37-Val36Met35-Leu34-Gly33-Ile32-Ile31-Ala30-Gly29-Lys28-Asn27Ser26- CONH2 P4:CH3-CO-NH-Ser26-Asn27-Lys28-Gly29-Ala30Ile31-Ile32-Gly33-Leu34-Met35-Val36-CONH2 Here P1, P2, and P3 are representative parts of the full length or reverse sequences. As experimental results show different neurotoxicity for the representative full length sequences, we expect significantly different propensities of these peptides to stabilize oxidized Met35, which appears likely based on our calculations. P1 represents the R-helical C-terminus of full length βAP140 where we expect that (S-O)-bond formation is supported by the close spatial proximity of the Met35 sulfur to the CdO group C-terminal to Ile31. Full length βAP140 is neurotoxic. P2 represents an artificial peptide, in which we have arbitrarily replaced Ile31 by Pro31. Pro is a “helix breaker” (38) and, therefore, we expect that the Met35 sulfur and the CdO group of Pro31 show a lower efficiency of (S-O)-bond formation.1 Peptide P3 represents the reverse sequence of P1. As the reverse sequence of βAP40-1 is nontoxic (17, 39, 40), we expect P3 to show little efficiency to form (S-O)-bonded structures. P4 represents a βAP sequence which is nontoxic (19), and therefore, we expect little efficiency for (S-O)-bond formation.

Molecular Simulations and Computational Details Structures of Peptides. The initial structures of model peptides P1, P2, and P4 were derived from the published solution structure of βAP1-40, which was obtained by distance geometry calculations employing NMR-derived NOE restraints 1 Preliminary data reveal that this peptide is nontoxic and shows negligible tendency to aggregate and to form radicals (Kanski, J., Aksenova, M., Scho¨neich, Ch., Butterfield, D. A., unpublished results).

410

Chem. Res. Toxicol., Vol. 15, No. 3, 2002

Pogocki and Scho¨ neich

Table 1. Potential Energy Parameters Used to Simulate the Sulfur Centered Cation-Radical of Meta (a) Bond Parameters bond type S-C

force constant (kcal mol-1Å-2)

bond length (Å)

205

1.795

(b) Bond Angle Parameters angle type

force constant (kcal mol-1 radian-2)

bond angle (radian)

C-S-C C-C-S

38 50

1.796 1.918

(c) Electrostatic Charge Distribution “atom” type

charge (1 electron)

CH2 CH2 S•+ CH3

+0.100 +0.240 +0.390 +0.270

(34). Due to the lack of a specific structure of the “reverse” peptide P3, two alternative initial structures were examined: (a) a regular R-helix (P3A) and (b) an antiparallel β-sheet (P3B).2 During 20 ns dynamics calculation the starting structures have sufficient time to equilibrate with energetically more favorable conformations. In all structures, the N-terminal amine functionality was acylated and the C-terminal carboxylate was converted into an amide functionality. Computational Details. All simulations were performed in the extended atom model. As potential energy function we employed the CHARMM potential (41, 42) in its HyperChem implementation (43). To simulate the presence of water solvent we utilized a scale factor () equal to 80, which screened the charge-charge interactions. Such a simplification has been shown to give results that quantitatively agree with solvent simulations (44). The presence of a lipid environment was mimicked by equal to 3, which is comparable to experimental values for liquid fatty acids and esters (45). The modeling of non-oxidized peptides was done with the default Bio85+ set of parameters that is an equivalent of the PARAM19 parameter set (46). However, to model the one electron oxidized peptides we parametrized the sulfur-centered cation radical (>S•+) of Met. These parameters were deconvoluted from the UHF/6-31G(d) gas-phase calculation, which we have performed for the sulfur-centered cation radicals of dimethyl, ethylmethyl, and propylmethyl sulfide. Bond lengths, bond angles, and dihedral angels were taken from calculated data. The force constants of bond stretching, bond angle bending, dihedral angle deformation, improper dihedral angle deformation, and nonbonded interaction for the sulfur cation radicals were copied from a set of parameters for the thioether sulfur (46). The partial charges were averaged from the charge distributions calculated with Merz-Kollman (47, 48) and Chelp (49, 50) models, implemented into the Gaussian’98W suite of programs (51) and then normalized to yield +1 charge within the side chain of the cation radical of Met. Parameters for the cation radical of Met are summarized in Table 1. The statistical distribution of the distances separating the sulfur of Met35 from the oxygen of neighboring peptide bonds was calculated by means of Langevin dynamics (LD) (52-58) without explicit solvent molecules, with a collision frequency (γ) for all heavy atoms equal to ca. 50 ps-1 in water, and ca. 1000 ps-1 in the lipid environment; these were set proportional to the respective viscosities of water (ca. 1 cP) and a lipid bilayer (ca. 200 cP) required to fit molecular dynamics and experimental data (53, 59). The free LD simulations were done with the 2 fs 2 We have modeled a single peptide chain, in which the angles are identical to an ideal antiparallel β-sheet structure, Φι ) -139° and Ψι ) +135°.

time step and 20 ps - 20 ns propagation time, preceded by 12 ps of heating from 0 to 300 K and 18 ps of equilibration. For practical purposes (1 ns dynamics required ca. 10 CPU hours) we have limited the LD propagation time to 20 ns, a time at least 1 order of magnitude longer than the average time required for the rotation of side chains on the surface of proteins [typically 10-11-10-10 s (60)]. For the force field calculations we used either the 4.5 or the 6-trial version of the HyperChem PC molecular modeling package (61), which integrates the Langevin equation of motion using the method of Allen and Tildesley (62). The data analysis was done using self-written procedures within the SigmaPlot 3.02 (63) program using its math features (64). For the nonlinear regression, the Marquardt-Levenberg algorithm (65-69) provided by the software has been utilized. Methods for the Computational Characterization of the Oxidation Susceptibility of the Peptides. In our approach, we hypothesized that the effect of the βAP-peptide sequence on a potential stabilization of the Met35 sulfide cation radical can be related to the ability of the individual peptide to promote the interaction between the sulfur of Met35 and the oxygen atoms of neighboring peptide bonds. As a benchmark of this ability we first decided to use the LD-calculated relative equilibrium constants and rate constants for the formation of sulfur cation radical-oxygen bonds. In the native peptide, such interaction may be manifested by a “nonbonded” association of the thioether sulfur of Met and the oxygen of the peptide bond. Crystallographic and computational studies have provided many examples for “nonbonded” sulfur-oxygen interactions, predominantly for the 1,5-type but also for the 1,4- and 1,6-type (70). For the Met sulfur in a peptide, the dynamics of this process can be schematically described by the first order, reversible reaction 4. Here, any transient “nonbonded” sulfur-oxygen association would be of the 1,(6+n)-type with amide bonds C-terminal of Met and 1,(7+n)-type with amide bonds N-terminal of Met (n ) 0, 1, 2, ...). Nevertheless, such nonbonded interaction may be promoted by the secondary structure of a peptide (such as an R-helical conformation in the C-terminus of βAP).

In the one-electron oxidized peptide, the interaction between a cation radical of Met and the oxygen of the peptide bond leads to the reversible formation of a sulfuranyl radical (reaction 5), and experimental evidence for a 1,6-cyclization has been provided (32).

For the purpose of this study, we define the products of reactions 4 and 5 as the (S-O)-bonded isomers of the peptide of the conformation in which the Oi-oxygen of the ith amino acid residue and the sulfur atom of Met35 are located within a distance, rS-O, shorter than the sum of their van der Waals radii (ca. 3.32 Å) (37), which defines a sulfur-oxygen collision within the framework of collision theory in chemical kinetics. To evaluate the relative rate constants of (S-O)-bond formation, kSO, based on the results of our LD simulations, we have adopted the methodology of Chandler for reactions of isomerization (71). Note that kSO represents a frequency at which the reactive system crosses over the transition state barrier, which can be treated as a relative rate constant if we assume that the



Figure 1. Rate correlation function (solid line) from eq 13 and logarithm (doted line with scale on right): LD simulation at ) 80, γ ) 50 ps-1 of peptide P2 (calculated τrxn ) 4.7 ( 0.3 ps). transmission coefficient (fraction of transition state complex which crosses the barrier to form product) is equal for all investigated peptides. The rate of (SO)-bond formation is described by the first-order rate law shown in eq 6.

d〈NSO〉na(t) ) kSO〈NS〉ne(t) - kS〈NSO〉ne(t) dt

(6)

Figure 2. Distribution of the r(S-O)-distances in native peptides: P1(Met35-Ile31) (s), P2(Met35-Pro31) (- - -), P3A(Met35Val39) (‚‚‚‚), P3B(Met35Val39) (-‚‚-), and P4(Met35-Ile31) (-‚-‚-); obtained via 20 ns LD simulation at ) 80, γ ) 50 ps-1 (“water”), and ) 3, γ ) 1000 ps-1 (“lipid”).

Here, the average fraction of the SO-bonded 〈NSO〉 and the fraction of all other 〈NS〉 peptide conformers under equilibrium conditions is replaced by their time dependent nonequilibrium values 〈NSO〉ne(t) and 〈NS〉ne(t). Equation 6 is valid only for a closed system in which N ) NSO(t) + NS(t) ) 1. The solution of eq 6 is eq 7

〈NSO〉ne(t) ) 〈NSO〉 + [〈NSO〉ne(0) - 〈NSO〉] × e-t/τrxn

(7)

where the relaxation time τrxn is given by expression 8.

τrxn-1 ) kSO + kS

(8)

On the basis of the equilibrium conditions in eq 9,

KSO )

kSO 〈NSO〉 ) kS 〈NS〉

(9)

kSO can be calculated from eq 10

〈NSO〉 τrxn

kSO )

(10)

where 〈NSO〉 and 〈NS〉 are the average fractions of the conformers calculated from the statistical distribution evaluated over a 20 ns time LD trajectory. The relative values of kSO can then be evaluated from the trajectories by means of number correlation function (71-73), providing an accurate rate at arbitrary collision frequency (γ). The relaxation time, τrxn, is obtained by the integration of the normalized number correlation function CN(t)

τrxn )

∫C ∞

0

N(t)dt

(11)

of the form

CN(t) )

〈δNSO(t)δNSO(0)〉 〈δNSO2〉

Figure 3. Distribution of the r(S-O)-distances in radical cations of peptides: P1(Met35-Ile31) (s), P1(Met35-Ile32) (b), P2(Met35Pro31) (- - -), P3A(Met35-Met35) (O), P3A(Met35-Val39) (‚‚‚‚), P3B(Met35-Val39) (-‚‚-), P4(Met35-Ile31) (-‚-‚-), P4(Met35-Met35) (9); obtained via 20 ns LD simulation at ) 80, γ ) 50 ps-1 (“water”), and ) 3, γ ) 1000 ps-1 (“lipid”). where δNSO(t) ) NSO(t) - 〈NSO〉 and NSO(t) ) 1 for rSO e 3.32 Å and NSO(t) ) 0 for rSO > 3.32 Å. Note that 〈δNSO2〉 ) 〈NSO〉 〈NSO〉2. Thus, values for kSO can be evaluated from the LD-trajectories. However, the replacement of an equilibrium ensemble average by a time average over a finite time of the length of the simulation introduces an error which can be calculated by the formula of Zwanzig and Ailawadi (73-75),

err ) (2τrxn/T)1/2

(13)

(12) where T is the length of the dynamics simulation. The loga-

412



Table 2. Relative Equilibrium Constants, KSO ) 〈NSO〉/〈NS〉, of the (S-O)-Bonded Complex Formation, Obtained in the 20 ns LD Simulationa KSO × 103 ) 80, γ ) 50 peptide

Met35-Ile(Pro)31

Met35-Ile32

P1 P2 P3A P3B P4 P1(S•+) P2(S•+) P3A(S•+) P3B(S•+) P4(S•+)

9.7 ( 1.4 5.6 ( 1.1

1.6 ( 0.3

a

0.6 ( 0.2 2.8 ( 0.8 17.0 ( 5.0 1.0 ( 0.5

4.8 ( 1.0 3.2 ( 0.8 2.4 ( 0.7

ps-1

Met35-Met35

) 3, γ ) 1000 ps-1 Met35-Val39

3.2 ( 0.8 10.1 ( 1.4 3.0 ( 0.8 3.8 ( 0.9 2.2 ( 0.7

0.6 ( 0.35 0.65 ( 0.11 -

Met35-Ile(Pro)31

Met35-Ile32

68.3 ( 3.5 39.5 ( 2.7

0.2 ( 0.1

9.5 ( 1.4 97 ( 6.0 7.2 ( 1.2

Met35-Met35

4.8 ( 1.0

89.1 ( 3.9

Met35-Val39

10.5 ( 1.5 62.2 ( 3.5 149 ( 6.0 -

Only nonzero values are shown.

rithms of the correlation functions, ln(CN(t)), were found to be linear over at least one relaxation time, implying that they are described by a single exponential; typical results are shown in Figure 1. The relaxation times were, therefore, calculated from single-exponential fits of the correlation function, rather then by calculating the integral of eq 11.

Results and Discussion Distribution of the “(S-O)-Bonded Conformers”. Relative Equilibrium Constants. The LD simulation shows that all investigated peptides can adopt a certain number of conformations which may favor the formation of intramolecular sulfur-oxygen bonds both in aqueous and in lipid environment. (The statistical distributions of the peptide conformers calculated in 20 ns LD simulations are presented in Figures 2 and 3.) During 20 ns simulation, all investigated peptides, native and oneelectron oxidized, adopt conformations in which the sulfur of Met35 collides with the oxygen of one of the neighboring peptide functions [Ile31 (Pro31), Ile32, Val39 or Met35]. Thus, in principle the formation of an (S-O)-bonded structure may be possible for all peptides. However, a comparison of the calculated equilibrium (stability) constants of the “(S-O)-bonded” conformations, KSO, shows significant differences between the peptides, depending on both the oxidation state and the environment, water, or lipid. (The stability constants KSO, calculated by integration of the statistical distributions, are summarized in Table 2). The calculated values of KSO cover the region of 3 orders of magnitude (1.5 × 10-1-2 × 10-4); a more detailed analysis of these data shows that the observed differences between the peptides are not very meaningful. On the basis of the highest KSO observed for a particular peptide, we calculate the expected difference in the free energy of (S-O)-bond formation, ∆G0289. In the extreme cases for two native peptides, P3B and P4 in “water”, the calculated ∆G0289 is equal to 1.67 kcal mol-1, and for the cation radicals of peptides P3B(S•+) and P2(S•+) in “lipid”, ∆G0289 is equal to 1.79 kcal mol-1. The first value amounts to only ca. 20% of the energy of a “nonbonded” interaction between sulfur and oxygen for the comparable interaction in (acylamino)thiadiazolines (ca. 8 kcal mol-1), calculated by Nagao et al. (70). The second value is lower than 8% of the energy of an (S-O)-three-electron-bond (22.7 kcal mol-1), calculated by Rauk et al. (21). Considering that ∆G0289 is on the order of 2 kcal mol-1, we decided not to use KSO to quantitate the relative susceptibility of the peptides toward oxidation. An energy of such low an amplitude can easily be overcome in a

Figure 4. Rate correlation functions calculated using eq 13 for LD simulation at ) 80, γ ) 50 ps-1 (“water”), and ) 3, γ ) 1000 ps-1 (“lipid”) of peptides: P1(Met35-Ile31) (s), P2(Met35Pro31) (- - -), P3A(Met35-Val39) (‚‚‚‚), P3B(Met35-Val39) (-‚-‚-), P4(Met35-Ile31) (-‚‚-).

peptide by the energy of other interactions such as formation of hydrogen bonds of energies up to 8 kcal mol-1 (76) or other local and nonlocal interactions (7779). A second problem associated with the use of KSO to predict the susceptibility of Met35 toward oxidation is the lifetime of the sulfur radical cation. Usually, such species are short-lived and decompose via deprotonation into R-(alkylthio)alkyl radicals (reactions 14 and 15).

MetS•+ f H+ + Met(CH2SCH2•)

(14)

MetS•+ f H+ + Met(•CHSCH3)

(15)

Therefore, the peptide has not unlimited time to equilibrate between open chain and (S-O)-bonded derivative. A better way to predict whether (S-O)-bond formation can assist the oxidation of Met is, therefore, to use peptide dynamics, as outlined below. Dynamics of Peptides. Relative Rate Constants. Figure 4 shows 100 ps periods of the CN(t)-correlation functions for native peptides (the correlation functions for the radical cations look similar). The shapes of CN(t)



Table 3. Relative Rate Constants, kSO ) 〈NSO〉/τrxv, of the (S-O)-Bonded Complex Formationa kSO × 10-9 (s-1) ) 80, γ ) 50 peptide

Met35-Ile(Pro)31

P1 P2 P3A P3B P4 P1(S•+) P2(S•+) P3A(S•+) P3B(S•+) P4(S•+)

10.7 ( 2.2 1.2 ( 0.25

a

1.2 ( 0.86 7.0 ( 2.0 1.5 ( 0.5 0.04 ( 0.02

Met35-Ile32

ps-1

Met35-Met35

) 3, γ ) 1000 ps-1 Met35-Val39

0.35 ( 0.09 1.1 ( 0.2 2.6 ( 0.7 2.0 ( 0.7 0.06 ( 0.02

0.35 ( 0.09 1.3 ( 0.3 0.89 ( 0.28

6.3 ( 0.5 0.76 ( 0.14 -

Met35-Ile(Pro)31

Met35-Ile32

Met35-Met35

20.9 ( 1.96 2.0 ( 0.17 0.4 ( 0.07 17.8 ( 2 0.9 ( 0.15 1.6 ( 0.08

0.01 ( 0.005

Met35-Val39

0.5 ( 0.09 2.8 ( 0.2 5.4 ( 0.4 -

Only nonzero values are shown.

Figure 5. Time series for ∆r(S-O) and ∆r(CR-CR) distance displacements between respective atoms of residues Met35and Ile31 in P1, and Met35 and Val39 in P3B, for LD at ) 80, γ ) 50 ps-1 (“water” run), and ) 3, γ ) 1000 ps-1 (“lipid” run).

functions and, thus, the relaxation times, τrxn, show pronounced differences between the individual peptides, which finally manifest in distinct frequencies kSO, shown in Table 3. In general, (S-O)-bond formation is most efficient for the peptide P1 and most efficient for (S-O)bond formation between sulfide (radical cation) and the CdO function C-terminal to Ile31. To understand this dynamic behavior, we examined which kind of internal motions of a particular peptide may be responsible for the formation of kSO. To obtain additional insight into the nature of the motions, we illustrate the time series for certain coordinates. As an example, Figure 5 shows a time series of fluctuation of the ∆r(S-O) displacements from the mean distances for Met35 and Ile31 in P1 and Met35 and Val39 in P3B. The picture also shows such a time series of fluctuation for the displacement of the distances between the R-carbons and the respective amino acid residues [∆r(CR-CR)], which should be representative for the motions of a peptide

backbone. ∆r(S-O) and ∆r(CR-CR) show high-frequency, small-amplitude oscillations (e1 ps) which are partially suppressed in the “lipid” environment. In addition, there is a strong indication of lower frequency, large amplitude contributions to the motions. From the time series and correlation functions, it is evident that the atomic motions represent a superposition of high-frequency oscillations and lower frequency fluctuations. To separate the contributions of these two types of motion to the mean square displacements, we can determine subaveraged mean square displacement values; that is, the entire trajectory is divided into a series of time intervals of given length, the mean square displacement relative to the mean for each interval are calculated, and results are averaged for the entire trajectory (80). The root-meansquare displacement (RMSD ) x〈(∆r)2〉) from the 0.1-, 1-, 5-, 10-, and 100-“ps” averages for r(S-O) and r(CR-CR) distances in the native peptides and their radical cations

414



Figure 6. Root-mean-square displacement subaverages (0.1, 1, 5, 10, 100 “ps”) of distances r(S-O) (dark columns) and r(CR-CR) (white columns) between respective atoms in residues Met35 and Ile31 in native peptides P1 and P4, Met35 and Val39 in P3A and P3B, and Met35and Pro31 in P2; for “water” and “lipid” runs (see text).

are displayed in Figures 6 and 7. They show the higher amplitudes of motions for the larger subaveraging intervals; thus, the motions that open and close the “(S-O)bond” are prominent in the low frequency. The amplitudes of motions are substantially damped with the transition from the “water” to the “lipid” environment. This effect is more pronounced for r(S-O) than for r(CR-CR), probably due to the better exposure of the Met35-side chain to the “viscous solvent” than the CR-atoms of the peptide backbone. However, the dumping effect seems not to be directly correlated to the value of the motion amplitude. To quantify the correlation between ∆r(S-O) and ∆r(CR-CR), which, for example, is visible in the time series presented for P3B in Figure 5, we calculated the cross-correlation coefficient for these parameters from the 100 ps trajectory

for each peptide. The respective cross-correlation coefficients C∆r, calculated from eq 16 (55),

C∆r )

〈∆r(S-O) × ∆r(CR-CR)〉

x

〈∆r(S-O)2〉〈∆r(CR-CR)2〉

(16)

are collected in Table 4. The results suggest that a strong dependence of ∆r(S-O) on ∆r(CR-CR) causes the low frequency of the “(S-O)-bond” formation in the peptide. For example, the highest correlation coefficients are calculated for the peptides P3B/P3B(S•+), which show a low kSO, the lowest for P1/P1(S•+), which show a high kSO. In other words, it appears that the formation of the (SO)-bond in peptides P3 and P4 requires some reorganization of their backbone conformations, whereas the (SO)-bond formation in P1 demands only local motions of



Figure 7. Root-mean-square displacement subaverages (0.1, 1, 5, 10, 100 “ps”) of distances r(S-O) (dark columns) and r(CR-CR) (white columns) between respective atoms in residues Met35 and Ile31 in 1e-oxidized peptides P1(S•+) and P4(S•+), Met35 and Val39 in P3A(S•+) and P3B(S•+), and Met35and Pro31 in P2(S•+); for “water” and “lipid” runs (see text).

the Met35 side chain. Interestingly, peptide P2 shows moderate negative correlation coefficients indicating small phase displacement between the motions of the peptide backbone and the Met35 side chain, which can cause the significantly lower kSO.

Summary and Conclusions Our molecular modeling results confirm a “privileged” conformation of β-amyloid peptide congeners containing an R-helical C-terminal sequence of βAP(26-40) where structural and dynamic properties can promote the formation and stabilization of MetS•+. The tendency to stabilize MetS•+ in form of an (S-O)-bond may explain the tendency of the native βAP1-40 to reduce CuII, generate free radicals, and induce protein oxidation whereas mutants not containing Met35, the reverse βAP sequences and certain truncated βAP sequences are

Table 4. Cross-correlation Coeficient C∆r ) 〈∆r(S-O) ×

x

∆r(Cr-Cr)〉/ 〈∆r(S-O)2〉〈∆r(CR-CR)2〉 for the Displacement of r(S-O) and r(Cr-Cr) during 100 ps LD Simulation cross-correlation coefficient ) 80, γ ) 50 ps-1 peptide P1 P2 P3A P3B P4 P1(S•+) P2(S•+) P3A(S•+) P3B(S•+) P4(S•+)

Met35Ile(Pro)31

Met35Val39

-0.06 -0.14

) 3, γ ) 1000 ps-1 Met35Ile(Pro)31

Met35Val39

0.13 -0.33 0.18 0.83

0.28 0.51 -0.05 -0.004 -0.15

0.47 0.21 -0.19 0.44 0.58 0.61

0.19 0.81 0.30

416


generally less reactive. Of the investigated peptides, experimental data show that full length βAP1-40, represented by P1, is neurotoxic, able to reduce CuII, and able to form free radicals whereas the sequences represented by the other peptides do not show these reactivities. Our modeling results indicate that only P1 is able to kinetically efficient form an (S-O)-bonded radical cation (see Table 3). Hence, our computational data would support the conclusion that the oxidation of βAP, proceeding via electron transfer from Met35, can be assisted by (S-O)-bond formation. Experimental evidence for Met sulfide cation radical-amide (S-O)-bond formation was recently presented by us for the model compound Nacetylmethionine amide (32). In the present paper, the successful formation of an (S-O)-bond was evaluated by the r(S-O) distance being smaller than the van der Waals radii of sulfur and oxygen. A more precise picture would be obtained if the actual bond geometry were known. For example, our recent ab initio/DFT calculations for (thiocarboxylic)acids (and comparison with experimental data) suggest that (S-O)-bonded radicals most likely adopt the σ*, threeelectron bonded configuration of very precisely defined geometry, which is subject to the influence of external factors such as steric hindrance and hydrogen bond formation (81). We note that a simulation performed using Langevin dynamics includes the averaged effects of the solvent without requiring the explicit presence of solvent molecules. Hence, the solvent influence on the dynamic behavior of the solute (by random collisions imposing a frictional drag on the motion of the solute through the solvent) is simulated by the frictional force proportional to the collision frequency (54, 55, 57, 58). The distribution of structures obtained through the LD simulation could be slightly different then that from the molecular dynamics (MD) with explicit solvent molecules, due to the consideration of potential hydrogen bonding between a peptide and a solvent. Most probably, the presence of explicit solvent molecules can alter the time dependence of the motion but not their magnitude, particularly for atoms on the surface of the molecule, due to their direct interaction with the solvent atoms (80). It clearly would be desirable to supplement the present study by more realistic models of the aqueous and the lipid environment. On the other hand, the projected discrepancy between the LD and the MD simulation is probably lower than the uncertainty arising from the simplifying assumption of the “united atom” representation, which allows to perform CHARMM-force field simulations on a PC computer.

Acknowledgment. This research was supported by the NIH (PO1AG12993). We thank Ms. Kimberly Sprence for providing computational facilities.


(4)

(5)

(6)

(7)

(8)

(9)

(10)

(11)

(12)

(13)

(14) (15) (16)

(17)

(18)

(19)

(20)

References (1) Butterfield, A. D. (1997) β-Amyloid-associated free radical oxidative stress and neurotoxicity: implications for Alzheimer’s disease. Chem. Res. Toxicol. 10, 495-506. (2) Butterfield, D. A., Hensley, K., Harris, M., Mattson, M., and Carney, J. (1994) β-Amyloid peptide free radical fragments initiate synaptosomal lipoperoxidation in a sequence-specific fashion: implications to Alzheimer’s disease. Biochem. Biophys. Res. Commun. 200, 710-715. (3) Sayre, L. M., Zagorski, M. G., Surewicz, W. K., Krafft, G. A., and Perry, G. (1997) Mechanisms of neurotoxicity associated with

(21)

(22)

amyloid beta deposition and the role of free radicals in the pathogenesis of Alzheimer’s disease: a critical appraisal. Chem. Res. Toxicol. 10, 518-526. Haass, C., Koo, E. H., Mellon, A., Hung, A. Y., and Selkoe, D. J. (1992) Targeting of cell-surface β-amyloid precursor protein to lysosomes: alternative processing into amyloid-bearing fragments. Nature 357, 500-503. Seubert, P., Vigo-Pelfrey, C., Esch, F., Lee, M., Dovey, H., Davis, D., Sinha, S., Schlossmacher, M., Whaley, J., and Swindlehurst, C. (1992) Isolation and quantification of soluble Alzheimer’s β-peptide from biological fluids. Nature 359, 325-327. Atwood, C. S., Huang, X., Khatri, A., Scarpa, R. C., Kim, Y. S., Moir, R. D., Tanzi, R. E., Roher, A. E., and Bush, A. I. (2000) Copper catalyzed oxidation of Alzheimer Aβ. Cell Mol. Biol. (Noisy-le-grand) 46, 777-783. Hensley, K., Carney, J. M., Mattson, M. P., Aksenova, M., Harris, M., Wu, J. F., Floyd, R. A., and Butterfield, D. A. (1994) A model for β-amyloid aggregation and neurotoxicity based on free radical generation by the peptide: relevance to Alzheimer disease. Proc. Natl. Acad. Sci. U.S.A. 91, 3270-3274. Lynch, T., Cherny, R. A., and Bush, A. I. (2000) Oxidative processes in Alzheimer’s disease: the role of Aβ-metal interactions. Exp. Gerontol. 35, 445-451. Smith, C. D., Carney, J. M., Starke-Reed, P. E., Oliver, C. N., Stadtman, E. R., Floyd, R. A., and Markesbery, W. R. (1991) Excess brain protein oxidation and enzyme dysfunction in normal aging and in Alzheimer disease. Proc. Natl. Acad. Sci. U.S.A. 88, 10540-10543. Varadarajan, S., Yatin, S., Aksenova, M., and Butterfield, D. A. (2000) Review: Alzheimer’s amyloid β-peptide-associated free radical oxidative stress and neurotoxicity. J. Struct. Biol. 130, 184-208. Smith, M. A., Harris, P. L. R., Sayre, L. M., and Perry, G. (1997) Iron accumulation in Alzheimer disease is a source of redoxgenerated free radicals. Proc. Natl. Acad. Sci. U.S.A. 94, 98669868. Smith, M. A., Hirai, K., Hsiao, K., Papolla, M. A., Harris, P. L. R., Siedlak, S. L., Tabaton, M., and Perry, G. (1998) Amyloid-β deposition in Alzheimer transgenic mice is associated with oxidative stress. J. Neurochem. 70, 2212-2215. Sayre, L. M., Perry, G., Harris, P. L. R., Liu, Y., Schubert, K. A., and Smith, M. A. (2000) In situ oxidative catalysis by neurofibrillary tangles and senile plaques in Alzheimer’s disease: a central role for bound transition metals. J. Neurochem. 74, 270-279. Selkoe, D. J. (1996) Amyloid β-protein and the genetics of Alzheimer’s disease. J. Biol. Chem. 271, 18295-18298. Markesbery, W. R. (1997) Oxidative stress hypothesis in Alzheimer’s disease. Free Radical Biol. Med. 23, 134-147. Butterfield, D. A., Martin, L., Carney, J. M., and Hensley, K. (1996) A beta (25-35) peptide displays H2O2-like reactivity towards aqueous Fe2+, nitroxide spin probes, and synaptosomal membrane proteins. Life Sci. 58, 217-228. Harris, M. E., Hensley, K., Butterfield, D. A., Leedle, R. A., and Carney, J. M. (1995) Direct evidence of oxidative injury produced by the Alzheimer’s β-amyloid peptide (1-40) in cultured hippocampal neurons. Exp. Neurol. 131, 193-202. Hensley, K., Aksenova, M., Carney, J. M., Harris, M., and Butterfield, D. A. (1995) Amyloid β-peptide spin trapping. II: Evidence for decomposition of the PBN spin adduct. Neuroreport 6, 493-496. Varadarajan, S., Kanski, J., Aksenova, M., Lauderback, C., and Butterfield, D. A. (2001) Different mechanisms of oxidative stress and neurotoxicity for alzheimer’s Aβ(1-42) and Aβ(25-35). J. Am. Chem. Soc. 123, 5625-5631. Huang, X., Cuajungco, M. P., Atwood, C. S., Hartshorn, M. A., Tyndall, J. D., Hanson, G. R., Stokes, K. C., Leopold, M., Multhaup, G., Goldstein, L. E., Scarpa, R. C., Saunders, A. J., Lim, J., Moir, R. D., Glabe, C., Bowden, E. F., Masters, C. L., Fairlie, D. P., Tanzi, R. E., and Bush, A. I. (1999) Cu(II) potentiation of alzheimer Aβ neurotoxicity. Correlation with cellfree hydrogen peroxide production and metal reduction. J. Biol. Chem. 274, 37111-37116. Rauk, A., Armstrong, D. A., and Fairlie, D. P. (2000) Is oxidative damage by β-amyloid and prion peptides mediated by hydrogen atom transfer from glycine R-carbon to methionine sulfur within β-sheets? J. Am. Chem. Soc. 122, 9761-9767. Sanaullah, Wilson, S., and Glass, R. S. (1994) The effect of pH and complexation of amino acid functionality on the redox chemistry of methionine and X-ray structure of [Co(en)2(L-Met)](ClO4)2‚H2O. J. Inorg. Biochem. 55, 87-99.

Redox Properties of Met35 (23) Baciocchi, E., Rol, C., and Mandolini, L. (1980) Changeover from rate-determining electron transfer to rate-determining proton transfer in the oxidation of alkyl aromatic compounds by ceric ammonium nitrate. J. Am. Chem. Soc. 102, 7597-7598. (24) Fox, S., and Karlin, K. D. (1995) Dioxygen reactivity in copper proteins and complexes. In Active oxygen in biochemistry (Selverstone Valentine, J., Foote, C. S., Greenberg, A., and Liebman, J. F., Eds.) Vol. 3, Chapter 4, Blackie Academic & Professional, Glasgow. (25) Kitajima, N., and Moro-oka, Y. (1994) Copper-dioxygen complexes. Inorganic and bioinorganic perspectives. Chem. Rev. 94, 737757. (26) Ho, R. Y. N., Liebman, J. F., and Selverstone Valentine, J. (1995) Biological reactions of dioxygen: an introduction. In Active oxygen in biochemistry (Selverstone Valentine, J., Foote, C. S., Greenberg, A., and Liebman, J. F., Eds.) Vol. 3, Chapter 1, pp 1-36, Blackie Academic & Professional, Glasgow. (27) Zuberbu¨hler, A. D. (1993) Kinetics and mechanism of CuI/O2 reactions. In Bioinorganic chemistry of copper (Karlin, K. D., and Tyeklar, Z., Eds.), Chapter 21, pp 264-276, Chapman & Hall, New York, London. (28) Solomon, E. I., Hemming, B. L., and Root, D. E. (1993) Electronic structures of active sites in copper proteins: coupled binuclear and trinuclear cluster sites. In Bioinorganic chemistry of copper (Karlin, K. D., and Tyeklar, Z., Eds.) Chapter 1, pp 3-20, Chapman & Hall, New York, London. (29) Asmus, K.-D. (1979) Stabilization of oxidized sulfur centers in organic sulfides. Radical cations and odd-electron sulfur-sulfur bonds. Acc. Chem. Res. 12, 436-442. (30) Steffen, L. K., Glass, R. S., Sabahi, M., Wilson, G. S., Scho¨neich, C., Mahling, S., and Asmus, K.-D. (1991) OH radical induced decarboxylation of amino acids. Decarboxylation vs bond formation in radical intermediates. J. Am. Chem. Soc. 113, 21412145. (31) Bobrowski, K., Hug, G. L., Marciniak, B., Miller, B. L., and Scho¨neich, C. (1997) Mechanism of one-electron oxidation of β-, γ-, and δ-hydroxyalkyl sulfides. Catalysis through intramolecular proton transfer and sulfur-oxygen bond formation. J. Am. Chem. Soc. 119, 8000-8011. (32) Scho¨neich, C., Pogocki, D., Wisniowski, P., Hug, G., and Bobrowski, K. (2000) Intramolecular sulfur-oxygen bond formation in radical cations of N-acetylmethionine amide. J. Am. Chem. Soc. 122, 10224-10225. (33) Glass, R. S. (1990) Neighboring group participation: general principles and application to sulfur-centered reactive species. In Sulfur-Centered Reactive Intermediates in Chemistry and Biology (Chatgilialoglu, C., and Asmus, K.-D., Eds.) Vol. 197, Plenum Press, New York. (34) Coles, M., Bicknell, W., Watson, A. A., Fairlie, D. P., and Craik, D. J. (1998) Solution structure of amyloid β-peptide(1-40) in a water-micelle environment. Is the membrane-spanning domain where we think it is? Biochemistry 37, 11064-11077. (35) Zhang, S., Iwata, K., Lachenmann, M. J., Peng, J. W., Li, S., Stimson, E. R., Lu, Y.-a., Felix, A. M., Maggio, J. E., and Lee, J. P. (2000) The Alzheimer’s peptide Aβ adopts a collapsed coil structure in water. J. Struct. Biol. 130, 130-141. (36) Curtain, C. C., Ali, F., Volitakis, I., Cherny, R. A., Norton, R. S., Beyreuther, K., Barrow, C. J., Masters, C. L., Bush, A. I., and Barnham, K. J. (2001) Alzheimer’s disease amyloid-β binds copper and zinc to generate an allosterically ordered membranepenetrating structure containing superoxide-dismutase-like subunits. J. Biol. Chem. 276, 20466-20473. (37) Bondi, A. J. (1964) Van der Waals volumes and radii. J. Phys. Chem. 68, 441-451. (38) Creighton, T. E. (1993) Proteins. Structure and molecular properties, 2nd ed., W. H. Freeman and Company, New York. (39) Harrigan, M. R., Kunkel, D. D., Nguyen, L. B., and Malouf, A. T. (1995) Beta amyloid is neurotoxic in hippocampal slice cultures. Neurobiol. Aging 16, 779-789. (40) Kowall, N. W., Beal, M. F., Busciglio, J., Duffy, L. K., and Yankner, B. A. (1991) An in vivo model for the neurodegenerative effects of beta amyloid and protection by substance P. Proc. Natl. Acad. Sci. U.S.A. 88, 7247-7251. (41) Brooks, B. R., Bruccoleri, R. E., Olafson, B. D., States, D. J., and Swaminathan, S. K. M. (1983) CHARMM: A program for macromolecular energy, minimization, and dynamics calculation. J. Comput. Chem. 4, 187-217. (42) Reiher, W. E. (1985) Ph.D. Dissertation, Harvard University, Cambridge MA. (43) Molecular Mechanics. (1994) In HyperChem Computational Chemistry, Chapter 10, pp 139-185, Hypercube Inc., Waterloo, Ontario, Canada.

Chem. Res. Toxicol., Vol. 15, No. 3, 2002 417 (44) Sneddon, S. F., and Brooks, C. L. I. (1992) The conformation of proline-linked donor-acceptor systems. J. Am. Chem. Soc. 114, 8220-8225. (45) Gunstone, F. D., Harwood, J. L., and Padley, F. B. (1994) The Lipid handbook, Chapman and Hall, London. (46) MacKarell, A. D., Jr., PARAM19. (1999) www.pharmacy.umab.edu/∼alex/research.html. (47) Singh, U. Ch., and Kollman, P. A. J. (1984) An approach to computing electrostatic charges for molecules. J. Comput. Chem. 5, 129-145. (48) Besler, B. H., Merz, K. M., and Kollman, P. A. J. (1990) Atomic charges derived from semiempirical methods. J. Comput. Chem. 11, 431-439. (49) Chirlian, L. E., and Francl, M. M. (1987) Atomic charges derived from electrostatic potentials: A detailed study. J. Comput. Chem. 8, 894-905. (50) Breneman, C. M., and Wiberg, K. B. J. (1990) Determining atomcentered monopoles from molecular electrostatic potentials. The need for high sampling density in formamide conformational analysis. J. Comput. Chem. 11, 361-373. (51) Frisch, M. J., Trucks, G. W., Schlegel, H. B., Scuseria, G. E., Robb, M. A., Cheeseman, J. R., Zakrzewski, V. G., Montgomery, J. A., Jr., Stratmann, R. E., Burant, J. C., Dapprich, S., Millam, J. M., Daniels, A. D., Kudin, K. N., Strain, M. C., Farkas, O., Tomasi, J., Barone, V., Cossi, M., Cammi, R., Mennucci, B., Pomelli, C., Adamo, C., Clifford, S., Ochterski, J., Petersson, G. A., Ayala, P. Y., Cui, Q., Morokuma, K., Malick, D. K., Rabuck, A. D., Raghavachari, K., Foresman, J. B., Cioslowski, J., Ortiz, J. V., Stefanov, B. B., Liu, G., Liashenko, A., Piskorz, P., Komaromi, I., Gomperts, R., Martin, R. L., Fox, D. J., Keith, T., Al-Laham, M. A., Peng, C. Y., Nanayakkara, A., Gonzalez, C., Challacombe, M., Gill, P. M. W., Johnson, B. G., Chen, W., Wong, M. W., Andres, J. L., Head-Gordon, M., Replogle, E. S., Pople, J. A. (1998) Gaussian 98, revision A.7, Gaussian, Inc., Pittsburgh, PA. (52) Levy, R. M., Karplus, M., and McKee, M. L. (1979) Diffusive Langevin dynamics of model alkanes. Chem. Phys. Lett. 65, 411. (53) Widmalm, G., and Pastor, R. W. (1992) Comparison of Langevin and molecular dynamics simulations equilibrium and dynamics of ethylene glycol in water. J. Chem. Soc., Faraday Trans. 88, 1747-1754. (54) Frenkel, D., and Smit, B. (1996) Understanding Molecular Simulation. From Algorithms to Applications, Academic Press, San Diego. (55) Leach, A. R. (1996) Molecular modelling: principles and applications, Longman, Harlow, England. (56) He, S., and Scheraga, H. A. (1998) Brownian dynamics simulations of protein folding. J. Chem. Phys. 108, 287-300. (57) Field, M. J. (1999) A Practical Introduction to the Simulation of Molecular Systems, Cambridge University Press, Cambridge. (58) Hinchliffe, A. (1999) Chemical Modeling. From Atoms to Liquids, John Wiley & Sons, Ltd., Chichester, U.K. (59) Venable, R. M., Zhang, Y., Hardy, B. J., and Pastor, R. W. (1993) Molecular dynamics simulations of a lipid bilayer and of hexadecane: an investigation of membrane fluidity. Science 262, 223226. (60) McCammon, J. A., and Harvey, S. (1991) Dynamics of proteins and nucleic acids, Cambridge University Press, Cambridge. (61) HyperChem Computational Chemistry (1996) Hypercube Inc., Waterloo, Ontario, Canada. (62) Allen, M. P., Tildesley, D. J. (1987) In Computer Simulation of Liquids, Chapter 9, p 261, Clarendon Press, Oxford. (63) Kerle, K., Norby, J., Rubenstein, S., Schwallie-Farmer, C., Senchyna, T., Turke, T. M., Youtz, P., and Zhu, Q. (1995) SigmaPlot for Windows (3.02), Jandel Corporation. (64) Kuo, J. (1994) SigmaPlot. Scientific Graphic Software. Transformation & Curve Fitting, Jandel Scientific, San Rafael, CA. (65) Marquardt, D. W. (1963) An Algorithm for Least Squares Estimation of Nonlinear Parameters. J. Soc. Ind. Appl. Math. 11, 431441. (66) Shrager, R. I. (1970) Nonlinear regression with Linear Constraints: An Extension of the Magnified Diagonal Method. J. Assoc. Comput. Machin. 17, 446-452. (67) Shrager, R. I. (1972) Quadratic Programming for N. Commun. ACM 15, 41-45. (68) Nash, J. C. (1979) Compact Numerical Methods for Computers: Linear Algebra and Function Minimization, John Wiley & Sons., New York.

418


(69) Press, W. H., Flannery, B. P., Teukolsky, S. A., and Vetterling, W. T. (1986) Numerical Recipes, Cambridge University Press, Cambridge. (70) Nagao, Y., Hirata, T., Goto, S., Sano, S., Kakehi, A., Iizuka, K., and Shiro, M. (1998) Intramolecular nonbonded S‚‚‚O interaction recognized in (acylimino)thiadiazoline derivatives as angiotensin II receptor antagonists and related compounds. J. Am. Chem. Soc. 120, 3104-3110. (71) Chandler, D. (1978) Statistical mechanics of isomerization dynamics in liquids and the transition state approximation. J. Chem. Phys. 68, 2959-2970. (72) Zwanzig, R. (1965) Time-correlation functions and transport coefficient in statistical mechanics. Annu. Rev. Phys. Chem. 16, 67-102. (73) Pastor, R. W., and Karplus, M. (1989) Internal effects in butane stochastic dynamics. J. Chem. Phys. 91, 211-218. (74) Zwanzig, R., and Ailawadi, N. K. (1969) Statistical error due to finite time averaging in computer experiments. Phys. Rev. 182, 280-283. (75) Lee, S., and Karplus, M. (1984) Brownian dynamics simulations: Statistical error of correlation functions. J. Chem. Phys. 81, 61066118.

Pogocki and Scho¨ neich (76) Dziembowska, T. (1994) Intramolecular hydrogen bonding. Pol. J. Chem. 68, 1455-1489. (77) Dill, K. A., Bromberg, S., Yue, K., Fiebig, K. M., Yee, D. P., Thomas, P. D., and Chan, H. S. (1995) Principles of protein folding a perspective from simple exact models. Protein Sci. 4, 561602. (78) Dill, K. A. (1990) Dominant forces in protein folding. Biochemistry 29, 7133-7155. (79) Wu, J., Gard, E., Bregar, J., Green, M. K., and Lebrilla, C. (1995) Studies of nearest-neighbor interactions between amino acids in gas-phase protonated peptides. J. Am. Chem. Soc. 117, 99009905. (80) van Gunsteren, W. F., and Karplus, M. (1982) Protein dynamics in solution and in a crystalline environment: A molecular dynamics study. Biochemistry 21, 2259-2274. (81) Pogocki, D., and Scho¨neich, Ch. (2002) Computational characterization of sulfur-oxygen bonded sulfuranyl radicals derived from alkyl- and (carboxyalkyl)thiopropionic acids: evidence for σ*-type radicals. J. Org. Chem. (in press).

TX0101550

Redox Properties of Met35 in Neurotoxic Î²-Amyloid Peptide. A

Recommend Documents