Supplemental Material
for “Free energy of helical transmembrane peptide dimerization in OPLS-AA/Berger
force field simulations: inaccuracy and implications
for partner-specific Lennard-Jones parameters between peptides and lipids”
by Manami Nishizawa and Kazuhisa Nishizawa
Contents:
Text S1
Text S2
Table S1
Table S2
Table S3
Table S4
Figure S1 and legend
Figure S2 and legend
References
p1
p2
p3
p4
p5
p6
p7
p8
p9
Text S1
System for the FEP analysis
For the FEP analysis, a simulation system similar to that constructed by Cordomi et al.
was used1. Briefly, the isotropic coupling method was used with 1 bar as the target pressure.
A time constant of 0.5 ps and a compressibility of 4.5 × 10-5 bar -1 were used. The bond
lengths of lipids (and proteins) and water molecules were restrained using LINCS and
SETTLE, respectively. LJ potentials were switched to 0 between 0.8 and 0.9 nm. For the
temperature setting, 298 K was chosen and a time constant T = 1.0 ps was used. For
soft-core potentials,  = 0.6 and  = 0.26 nm were used to avoid singularities. The PME
method was used in the same manner as the PMF analysis. Typically, a cyclohexane system
consisted of 500 cyclohexane molecules. For hydration runs, ~3000 water molecules were
used. For both, the box size was about 4.5 × 4.5 × 4.5 nm. For each intermediate state (i.e.,
each value), a 50-ps equilibration simulation was performed before a 500-ps production
run.
1
Text S2
Results of FEP calculation for hydration, cyclohexane solvation, and
water-to-cyclohexane transfer free energies
We examined the effects of the LJ-rescaling on free energies of solvation in cyclohexane
(and also of water-to-cyclohexane transfer) for the neutral forms of 14 SCAs involving
residues commonly occurring in TM segments. Cyclohexane parameters were taken from
the Berger FFs. Although not relevant to this study, we also calculated hydration energies
and the results replicated the values reported by Cordomi et al. (Table S1).[1] Thus,
OPLS-AA systematically overestimates the hydration free energies as reported in previous
studies,[1,2] with average errors of 3.1 kJ/mol and 3.4 kJ/mol for TIP3P and SPC water,
respectively, compared with the experiment.[3]
For cyclohexane solvation (Table S2), LJ-rescaling schemes ranging from OB105 (i.e.,
1.05-fold upscaling) to OB085 (0.85-fold downscaling) were examined. Unfavorable
interactions (destabilization) were observed as more intensive LJ-downscaling was used.
While for some SCAs the original OB was better than the rescaled versions, some
LJ-rescaling schemes were more accurate for some other SCAs. For Ala, OB led to
solvation that was too favorable whereas OB09 improved this. The remedial effect of OB09
on the TM PMF analysis (Figure 2A,B) can partly be explained by the abundance of Ala in
(AALALAA)3 that caused strong solvation under the original OB. For Leu, OB095 was
the most accurate among the tested OBs (Table S2). For Trp, OB105 was the most accurate.
OB combination unsurprisingly replicated the values of the free energies of transferring
the SCAs from cyclohexane to water as reported by Cordomi et al..[1] Table S3 and S4
show the results of calculation using TIP3P and SPC as the water model, respectively.
When both cyclohexane solvation energy and cyclohexane-to-water transfer energy were
regarded as targets, overall merits/demerits of each rescaling scheme differed among SCAs.
For Ala, OB09 improved both targets (Table S2,S3,S4). For Leu and Phe, OB095 improved
both targets. However, for hydroxyl group-carrying SCAs (Ser, Thr, and Tyr) and Trp,
OB105 (that made the solvation more favorable) improved cyclohexane solvation (Table S2),
but worsened the cyclohexane-to-water transfer energy (Table S3,S4). This is because the
original OB led to Trp hydration less favorable than in the experiment (Table S1), and
OB105 that stabilized cyclohexane solvation worsened the transfer energy. For Asn and Gln,
cyclohexane solvation with OB was fairly good (Table S2), but hydration (and
consequently the transfer energy) was inaccurate (Table S1). Thus, a uniform LJ-rescaling
over SCAs has limited usefulness, whereas a 'residue-specific scheme' may be more
relevant. The results also imply a challenge in achieving high accuracy both in cyclohexane
solvation and water-to-cyclohexane transfer in parameterization.
2
Table S1. Experimental and calculated hydration free energies (kJ/mol) for neutral
analogs of the 14 amino acids side chains.
OPLS-AA amino acid analog/water
experimental
reference
(ref.3)
(ref.1)
free energy
free energy
Acetamide (Asn)
-40.5
-35.7
Propionamide (Gln)
-39.2
-36.4
Methylimidazole (His)
-42.1
-29.0
Methanol (Ser)
-21.2
-19.6
Ethanol (Thr)
-20.4
-18.4
p-Cresol (Tyr)
-25.6
-22.2
Methanethiol (Cys)
-5.2
-1.1
Methane (Ala)
8.1
9.1
Methylindole (Trp)
-24.6
-19.2
-6.2
-1.2
Toluene (Phe)
-3.2
-3.6
Propane (Val)
8.4
9.9
Butane (Ile)
9.0
10.0
Isobutane (Leu)
9.5
9.7
OPLS-AA (TIP3P)
OPLS-AA (SPC)
compound
free energy (kJ/mol)
Methyl-ethylsulfide
(Met)
average error
energy relative to experiment (kJ/mol), ±SE (kJ/mol)
-36.0
4.5, ±0.233
-34.1
5.1, ±1.073
-29.3
12.8, ±0.176
-18.4
2.8, ±0.244
-18.4
2.0, ±0.308
-23.0
2.6, ±0.431
-2.0
3.2, ±0.141
9.1
1.0, ±0.071
-20.1
4.5, ±0.308
-35.8
4.7, ±0.153
-32.3
6.9, ±0.246
-29.0
13.1, ±0.086
-19.6
1.6, ±0.182
-18.9
1.5, ±0.171
-22.9
2.7, ±0.292
-1.9
3.3, ±0.149
9.0
0.9, ±0.124
-19.7
4.9, ±0.459
-2.2
4.0, ±0.137
-1.8
4.4, ±0.105
-4.1
-0.9, ±0.181
9.6
1.2, ±0.075
10.0
1.0, ±0.233
9.6
0.1, ±0.132
-3.6
-0.4, ±0.449
10.0
1.6, ±0.275
10.2
1.2, ±0.287
10.3
0.8, ±0.240
+3.1
+3.4
3
Table S2. Experimental and calculated cyclohexane solvation free energies (kJ/mol)
for neutral analogs of 14 of the 20 naturally occurring amino acids.
cyclohexane solvation free energy (kJ/mol)
experimental
OPLS-AA amino acid analog/Berger cyclohexane
(ref.4)
original parameters
compound
free energy
(kJ/mol)
OB
(ref.1)
free energy
(kJ/mol)
Acetamide (Asn)
-12.6
-12.8
Propionamide (Gln)
-15.8
-16.4
Methylimidazole (His)
-23.4
-19.4
Methanol (Ser)
-6.9
-4.3
Ethanol (Thr)
-9.5
-7.6
p-Cresol (Tyr)
-24.6
-22.4
Methanethiol (Cys)
-10.3
-8.3
0.6
-0.9
-33.8
-29.6
-15.8
-15.8
Toluene (Phe)
-17.5
-20.7
Propane (Val)
-8.5
-8.2
Butane (Ile)
-11.4
-12.5
Isobutane (Leu)
-10.9
-11.6
Methane (Ala)
Methylindole (Trp)
Methyl-ethylsulfide
(Met)
average error
OB
LJ-rescaling
OB105
OB095
OB09
OB085
free energy (kJ/mol)
energy relative to experiment (kJ/mol), SE (kJ/mol)
-13.3
-0.7, 0.021
-16.9
-1.1, 0.146
-19.2
4.2, 0.128
-4.0
2.9, 0.132
-7.7
1.8, 0.165
-15.0
-2.4, 0.138
-19.5
-3.7, 0.323
-21.8
1.6, 0.230
-4.9
2.0, 0.087
-9.1
0.4, 0.073
-11.4
1.2, 0.074
-14.7
1.1, 0.205
-16.9
6.5, 0.046
-2.8
4.1, 0.132
-6.2
3.3, 0.111
-9.7
2.9,0.079
-12.4
3.4, 0.174
-14.4
9.0, 0.087
-1.9
5.0, 0.039
-4.7
4.8, 0.068
-7.76
4.8, 0.237
-9.7
6.1, 0.278
-12.1
11.3, 0.272
-1.0
5.9, 0.086
-3.4
6.1, 0.189
-22.9
1.7, 0.157
-8.7
1.6, 0.120
-0.7
-1.3, 0.112
-29.9
3.9, 0.244
-26.0
-1.4, 0.045
-10.0
0.3, 0.129
-1.3
-1.9, 0.105
-33.8
0.0, 0.174
-19.9
4.7, 0.127
-7.1
3.2, 0.114
0.2
-0.4, 0.053
-26.3
7.5, 0.281
-16.9
7.7, 0.160
-5.8
4.5, 0.076
0.9
0.3, 0.047
-23.1
10.7, 0.175
-14.3
10.3, 0.119
-4.5
5.8, 0.144
1.6
1.0, 0.120
-19.6
14.3, 0.243
-15.8
0.0, 0.258
-18.3
-2.5, 0.220
-13.4
2.4, 0.231
-11.4
4.4, 0.024
-8.8
7.1, 0.114
-20.0
-2.5, 0.259
-8.5
0.0, 0.209
-12.2
-0.8, 0.203
-11.9
-1.0, 0.246
-23.1
-5.6, 0.087
-9.9
-1.4, 0.033
-14.4
-3.0, 0.164
-14.3
-3.4, 0.096
-17.4
0.1, 0.034
-6.6
1.9, 0.064
-10.4
1.1, 0.159
-10.2
0.7, 0.124
-14.5
3.0, 0.135
-5.3
3.2, 0.171
-8.0
3.4, 0.060
-8.0
2.9, 0.118
-12.1
5.4, 0.124
-3.6
4.9, 0.155
-6.0
5.4, 0.195
-6.2
4.8, 0.210
+0.6
+1.5
+2.7
+4.7
+6.7
4
Table S3. Cyclohexane to water (TIP3P) transfer energy of SCAs under OPLS-AA
protein/Berger cyclohexane/SPC water parameters with and without LJ-rescaling between
protein and lipid.
cyclohexane to water (TIP3P) transfer energy (kJ/mol)
original parameters
experimental
(ref.3, 4)
compound
OB
(ref.1)
free energy
(kJ/mol)
free energy
(kJ/mol)
Acetamide (Asn)
-27.7
-22.9
Propionamide (Gln)
-22.9
-19.9
Methylimidazole (His)
-18.7
-9.6
Methanol (Ser)
-14.2
-15.3
Ethanol (Thr)
-11.1
-10.8
p-Cresol (Tyr)
-0.8
0.2
Methanethiol (Cys)
5.2
7.2
Methane (Ala)
7.7
10
Methylindole (Trp)
9.5
10.4
9.7
14.6
Toluene (Phe)
14.1
17.1
Propane (Val)
16.7
18.1
Butane (Ile)
20.2
22.5
Isobutane (Leu)
20.5
21.3
Methyl-ethylsulfide
(Met)
average error
OB
LJ-rescaling
OB105
OB095
OB09
OB085
free energy (kJ/mol)
energy relative to experiment (kJ/mol)
-22.7
5.0
-17.2
5.7
-10.2
8.5
-14.4
-0.2
-10.7
0.4
-0.1
0.7
6.6
1.4
9.8
2.1
9.8
0.3
-21.0
6.7
-14.6
8.3
-7.5
11.2
-13.5
0.7
-9.3
1.8
3.0
3.8
8.0
2.8
10.4
2.7
13.7
4.2
-24.6
3.1
-19.4
3.5
-12.4
6.3
-15.6
-1.4
-12.2
-1.1
-3.2
-2.4
5.1
-0.1
8.9
1.2
6.2
-3.3
-26.3
1.4
-21.7
1.2
-14.9
3.8
-16.5
-2.3
-13.7
-2.6
-6.2
-5.4
3.8
-1.4
8.2
0.5
3.1
-6.4
-28.2
-0.5
-24.4
-1.5
-17.2
1.5
-17.4
-3.2
-15.0
-3.9
-8.8
-8.0
2.5
-2.7
7.5
-0.2
-0.5
-10.0
13.7
4.0
16.1
6.4
11.2
1.5
9.2
0.5
6.6
-3.1
15.9
1.8
18.0
1.3
22.2
2.0
21.6
1.1
19.0
4.9
19.5
2.8
24.4
4.2
23.9
3.4
13.3
-0.8
16.2
-0.5
20.4
0.2
19.8
-0.7
10.4
-3.7
14.9
-1.8
18.0
-2.2
17.7
-2.8
8.0
-6.1
13.1
-3.6
16.0
-4.2
15.8
-4.7
+2.4
+4.6
+0.4
-1.5
-3.6
5
Table S4. Cyclohexane-to-water (SPC) transfer energy of amino acid analogs under
OPLS-AA analog/Berger cyclohexane parameters with varied LJ-rescaling schemes.
cyclohexane to water (SPC) transfer free energy (kJ/mol)
experimental
(ref.3,4)
compound
original parameters
OB (ref.1)
free energy
(kJ/mol)
free energy
(kJ/mol)
Acetamide (Asn)
-27.7
-22.9
Propionamide (Gln)
-22.9
-19.9
Methylimidazole (His)
-18.7
-9.6
Methanol (Ser)
-14.2
-15.3
Ethanol (Thr)
-11.1
-10.8
p-Cresol (Tyr)
-0.8
0.2
Methanethiol (Cys)
5.2
7.2
Methane (Ala)
7.7
10
Methylindole (Trp)
9.5
10.4
9.7
14.6
Toluene (Phe)
14.1
17.1
Propane (Val)
16.7
18.1
Butane (Ile)
20.2
22.5
Isobutane (Leu)
20.5
21.3
Methyl-ethylsulfide
(Met)
average error
OB
LJ-rescaling
OB105
OB095
OB09
OB085
free energy (kJ/mol)
energy relative to experiment (kJ/mol)
-22.6
5.1
-15.4
7.5
-9.8
8.9
-15.6
-1.4
-20.9
6.8
-12.8
10.1
-7.2
11.5
-14.7
-0.5
-24.5
3.3
-17.6
5.3
-12.1
6.7
-16.8
-2.6
-26.2
1.5
-19.9
3.0
-14.6
4.1
-17.7
-3.5
-28.1
-0.4
-22.6
0.3
-16.9
1.8
-18.6
-4.4
-11.2
-0.1
0.0
0.8
6.8
1.6
9.7
2.0
10.1
0.6
-9.8
1.3
3.1
3.9
8.1
2.9
10.3
2.6
14.0
4.5
-12.7
-1.6
-3.0
-2.2
5.2
0.0
8.8
1.1
6.5
-3.0
-14.2
-3.1
-6.0
-5.2
3.9
-1.3
8.1
0.4
3.4
-6.1
-15.5
-4.4
-8.6
-7.8
2.6
-2.6
7.4
-0.3
-0.2
-9.7
14.1
4.4
16.5
6.8
11.6
1.9
9.6
-0.1
7.0
-2.7
16.5
2.4
18.5
1.8
22.3
2.1
22.2
1.7
19.6
5.5
20.0
3.3
24.5
4.3
24.5
4.0
13.9
-0.2
16.6
-0.1
20.5
0.3
20.4
-0.1
10.9
-3.2
15.3
-1.4
18.2
-2.0
18.3
-2.2
8.6
-5.6
13.6
-3.1
16.2
-4.0
16.4
-4.1
+2.7
+4.8
+0.6
-1.4
-3.4
6
Figure S1
Legend for Figure S1
Decomposition
of
enthalpic
energies
of
the
dimerization
PMF
analysis
with
the
(AALALAA)3/DOPC system under OB and LJ-rescaled versions. Decomposed LJ potential
energies (kJ/mol) are shown as a function of interhelical separation distance (r) along with error
bars representing S.E.. The values relative to the average of the values at r = 2.1 and 2.2 nm are
plotted. (A) Specific peptide-peptide LJ energy. (B) Specific peptide-lipid LJ energy. (C) Specific
lipid-lipid LJ energy. (D) Specific peptide-peptide Coulombic energy. (E) Specific peptide-lipid
Coulombic energy. (F) Specific lipid-lipid Coulombic energy.
7
Figure S2
Legend for Figure S2
Decomposition
of
enthalpic
energies
of
the
dimerization
PMF
analysis
with
the
(AALALAA)3/DOPC system under OB, OB09 and OBA09L095. Decomposed LJ potential energies
(kJ/mol) are shown as in Figure S1.
(A) Specific peptide-peptide LJ energy.
peptide-lipid LJ energy. (C) Specific lipid-lipid LJ energy.
Coulombic energy.
(B) Specific
(D) Specific peptide-peptide
(E) Specific peptide-lipid Coulombic energy.
(F) Specific lipid-lipid
Coulombic energy.
8
References for Supplementary Material
(1) Cordomí, A.; Caltabiano, G.; Pardo, L. Membrane Protein Simulations Using
AMBER Force Field and Berger Lipid Parameters. J. Chem. Theory Comput. 2012, 8,
948–958.
(2) MacCallum, J. L.; Tieleman, D. P. Calculation of the water-cyclohexane transfer free
energies of neutral amino acid side-chain analogs using the OPLS all-atom force field. J.
Comput. Chem. 2003, 24, 1930–1935.
(3) Wolfenden, R.; Andersson, L.; Cullis, P. M.; Southgate, C. C. Affinities of amino acid
side chains for solvent water. Biochemistry 1981, 20, 849–855.
(4)
Radzicka, A.; Wolfenden, R.
Comparing the polarities of the amino acids:
side-chain distribution coefficients between the vapor phase, cyclohexane, 1-octanol, and
neutral aqueous solution.
Biochemistry 1988, 27, 1664–1670.
9