^{by Benjamin Drexler and Fabian Grandke}

Introduction

In this task we analyse the structure of our protein to find out what effects the point mutations have. Therefore we created mutated structures and compared them to the wild-type protein. Several tools based on different methods have been used to achieve that aim. We used the mutations that we have chosen in the previous task.

Methods

In the first step of this task we had to find available protein structures for our protein and to decide which one would be the best for our detailed analysis. We set several cut-offs to exclude improper structures. The following tools have been used to perform the energy calulations. They were used as described in the task description.

SCWRL

SCWRL was initially developed by Dunbrack et al. in 1997. We use SCWRL4<ref name=dunb>G. G. Krivov, M. V. Shapovalov, and R. L. Dunbrack, Jr. Improved prediction of protein side-chain conformations with SCWRL4. Proteins (2009)</ref> which was published in 2009. The program takes a PDB file and a sequence file as input. By usage of a rotamer library, collision detection, and a residue interaction graph the optimal side-chain conformation is calculated, based on the backbone and the mutated sequence given in the input files. The output is a PDB file containing the conformation and the total minimal energy of the graph in STDOUT.

FoldX

FoldX was developed by Serrano et al. in 2002<ref name=serr>Guerois R, Nielsen JE, Serrano L., Predicting Changes in the Stability of Proteins and Protein Complexes: A Study of More Than 1000 Mutation. Journal of Molecular Biology (2002)</ref>. We used version FoldX 3.0 beta 4. The program provides the calculation of determination of energy effects of point mutations. It provides different run modes, but basically it takes a PDB file as input calculates several single energies(e.g. Van der Waals, Electrostatics, ...) and returns the single energies together with the total energy as output.

Minimise

Before this tool from the virtual box was used we had to remove the hydrogens and waters from the PDB file with the script repairPDB. Afterwards we were able to compare the energies differences between the wildtype and the mutated protein.

GROMACS

GROMACS is mostly a package to perform molecular dynamics, but it also provides energy calculations. For the mutations we used the forcefield AMBER03 and for the wildtype AMBER03, AMBERGS and CHARMM27. Additionally to the energy calculation task we did a runtime analysis with values from nsteps=10 to nsteps=1500. The results are shown in the results section of this task. According to the task description we created an MDP file with the following content:

title = PBSA minimization in vacuum
cpp = /usr/bin/cpp
define = -DFLEXIBLE -DPOSRES
implicit_solvent = GBSA
integrator = steep
emtol = 1.0
nsteps = 500
nstenergy = 1
energygrps = System
ns_type = grid
coulombtype = cut-off
rcoulomb = 1.0
rvdw	 = 1.0
constraints = none
pbc = no

Keyword	Describtion<ref name=manual>Gromacs Manual</ref>
General
title	Name of Project
cpp	Location of c-preprocessor
Preprocessing
define	Defines to pass to the preprocessor; -DFLEXIBLE:include flexible water in stead of rigid water into your topology; -DPOSRES: include posre.itp into your topology, used for position restraints
Implicit Solvent
implicit_solvent	Simulation with implicit solvent using the Generalized Born formalism
Run Control
integrator	Steepest descent algorithm for energy minimization
nsteps	Maximum number of steps to integrate or minimize
Energy minimization
emtol	Rhe minimization is converged when the maximum force is smaller than this value
Output
nstenergy	Frequency to write energies to energy file
Tables
energygrps	Group(s) to write to energy file
Neighbor searching
ns_type	Type of neighbor searching
pbc	Remove the periodicity (make molecule whole again)
Electrostatics
coulombtype	Type of coulomb energy
rcoulomb	Distance for the Coulomb cut-off
VDW
rvdw	distance for the LJ or Buckingham cut-off
Bonds
constraints	Which constraints should be used

Within the GROMACS work step we used the script fetchpdb. It checks if the given input is a valid PDB entry. If the check was successful it downloads the PDB file, extracts it and removes the packed version.

Results

Structure Selection

There are several structure files available for our protein:

PDB ID	Resolution [Å]	ph-Value	R-Factor	Coverage [%]	Missing Residues
1R46	3.25	8.0	0.262	99.7	422-429
1R47	3.45	8.0	0.285	99.5	422-429
3GXN	3.01	NULL	0.239	88.08	422-429
3GXP	2.20	NULL	0.204	81.9	422-429
3GXT	2.70	NULL	0.245	97.29	422-429
3HG2	2.30	4.6	0.178	97.32	422-429
3HG3	1.90	6.5	0.167	98.64	427-435
3HG4	2.30	4.6	0.166	99.86	422-429
3HG5	2.30	4.6	0.192	100	422-429
3LX9	2.04	6.5	0.178	98.92	423-435
3LXA	3.04	6.5	0.216	99.52	427-435
3LXB	2.85	6.5	0.227	99.3	427-435
3LXC	2.35	6.5	0.186	98.31	423-435

We set two cutoffs to decide which structures are excluded:

ph-value: < 6.5
resolution: > 2.7

After we applied the cutoffs to our set of structures three were left (exclusion factors are colored red in the table). One of them was slightly better than the other ones so we decided to use 3HG3 (worse values are colored gray in the table). Additionally 3GH3 has the best overall resolution and R-factor (colored green). As the missing residues are very similar for all structures they are not further taken into account.

Visual Examination of the Mutations

Figure 1 shows the protein α-galactosidase A and the residues which will be mutated. In the following sections, we compare the side chain conformation of the mutated residues and discuss the influence of the mutation. Aspects will be, inter alia, loss of polar interactions and clashes with other residues.

SCWRL was used to model the side chain conformation of the mutated residue and we use the term tool-based to describe this side chain conformation. The side chain conformation which was done according to this tutorial is referred to as manual side chain conformation.

Figure 1: Representation of the protein α-galactosidase A. The residues which will be mutated are colored red. Asp170 and A231 are part of the active site and colored cyan. The ligand is colored green.

close to active site?
comparison of side chain conformation
- clashes?
polar interactions
hydrophobic/hydrophilic?
surface

M42T (Mutation 1)

Figures 2 to 4 show the side chain conformation of the residue 42 in α-galactosidase A. The only difference between the manual and the tool-based side chain conformation is a variation in the conformation of the hydroxyl group of threonine (see figure 3 and 4). The tool-based side chain conformation does not lead to any clashes with the surrounding residues.

The wildtype M42 has two hydrogen bonds with E87 and Y88. Since these hydrogen bonds are formed by the carboxylgroup of the backbone, they are also abundant in the mutation, but T42 also forms a hydrogen bond with G85.

A part of T42 is exposed to the surface (see figure 5B). This is no problem, since threonine is slightly hydrophilic. But the mutation introduces a small hole into the surface of GLA (see figure 5A and 5B).

Figure 2: Close-up of methionine (wildtype) at position 42 in the protein GLA.

Figure 3: Close-up of threonine (mutated) at position 42 in the protein GLA. The side chain conformation was done manually.

Figure 4: Close-up of threonine (mutated) at position 42 in the protein GLA. The side chain conformation was done by SCRWL.

Figure 5: Surface representation of the protein GLA. (A) The wildtype residue methionine at position 42 is colored green. (B) The mutated residue threonine at position 42 is colored red. The mutation leads to a small hole in the surface of the protein.

S65T (Mutation 2)

The side chain conformation of the wildtype and the mutated residues is shown in figure 6 to 8. The hydroxyl and methyl group of threonine point towards totally different direction in the tool-based conformation (see figure 6 and 8), but there are no clashes with other residues. The wildtype S65 forms five hydrogen bonds with surrounding residues (C63, K67, L68, F69). The mutated residue T65 forms also five hydrogen bonds, but one of them is with E66 instead of K67.

Serine is hydrophilic and the residue is exposed on the surface, but threonine is also a hydrophilic residue. There is no remarkable change of the surface of the protein.

Figure 6: Close-up of serine (wildtype) at position 65 in the protein GLA.

Figure 7: Close-up of threonine (mutated) at position 65 in the protein GLA. The side chain conformation was done manually.

Figure 8: Close-up of threonine (mutated) at position 65 in the protein GLA. The side chain conformation was done by SCRWL.

I117S (Mutation 3)

Since the tool-based side chain conformation of the mutated residue (see figure 11) is pretty similar to the side chain conformation of the wildtype (see figure 9), there are no clashes in the mutated structure. The angle of the hydroxyl group is slightly different in the manual side chain conformation (see figure 10).

The hydroxyl group in the backbone of I117 forms two hydrogen bonds with L120 and A121. These bonds are also abundant in the mutated structure. Serine is a hydrophilic amino acid, but this is no problem, because it is part of the surface. Overall there is no remarkable difference in the surface of the protein due to the mutation.

Figure 9: Close-up of isoleucine (wildtype) at position 117 in the protein GLA.

Figure 10: Close-up of serine (mutated) at position 117 in the protein GLA. The side chain conformation was done manually.

Figure 11: Close-up of serine (mutated) at position 117 in the protein GLA. The side chain conformation was done by SCRWL.

A143T (Mutation 4)

This is the only mutation which is close to the active site of the protein (see figure 12). The general orientation of the side chain is very similar between the wildtype, manual and tool confirmation (see figure 13, 14 and 15). The hydroxyl group of threonine leads to an increase of the space requirements with the tool-based side chain conformation. Hence there is a clash between the hydroxyl group of T143 and the residue D93 (see figure 16).

A143 forms two hydrogen bonds with the carboxyl group of its backbone to D92 and T141. These hydrogen bonds are also formed by T143, but a third hydrogen bond is established between the oxygen of the hydroxyl group of T143 and D92.

T143 is part of the surface, so it is no problem that this position becomes hydrophilic. The mutation changes the surface and since this mutation is nearby the active site, this will probably have an influence on the work rate of the protein (see figure 17). The entry to the active site becomes more narrow due to the mutation. Hence it will be less likely or even impossible that the ligand binds in the active site.

Figure 12: Active site of the protein GLA with the ligand (green). The residues which are involved in the active site are shown in cyan. The mutated residue T143 is shown in red.

Figure 13: Close-up of alanine (wildtype) at position 143 in the protein GLA.

Figure 14: Close-up of threonine (mutated) at position 143 in the protein GLA. The side chain conformation was done manually.

Figure 15: Close-up of threonine (mutated) at position 143 in the protein GLA. The side chain conformation was done by SCRWL.

Figure 16: Close-up of the clash between the residues D98 (green) and T143 (red) in the protein GLA. (A) The two residues are shown in sticks representation. (B) Spheres representation of the two residues.

Figure 17: Close-up of the active site of the protein GLA in surface representation. The ligand is colored in orange. (A) This is the wildtype of GLA. The residue A143 is colored in green. (B) The mutated residue T143 is colored in red. The mutation changes the surface and it less likely that the ligand will be able to bind in the active site.

H186R (Mutation 5)

close to active site?
comparison of side chain conformation
- clashes?
polar interactions
hydrophobic/hydrophilic?
surface

Figures 13 to 15 show the side chain conformations of the residue. The general orientation of the side chain is similar. Because of this, the residue points towards the exterior of the protein. There is small collision with the residue D153 which is located in alpha-helix nearby (see figure 16).

H186 forms three hydrogen bonds with its backbone to D182, L189 and A190, which are also formed by the mutated residue R186. Both amino acids, histidine and arginine, are hydrophobic even though the position itself is exposed on the surface. The change on the surface is not noteworthy.

Figure 18: Close-up of histidine (wildtype) at position 186 in the protein GLA.

Figure 19: Close-up of arginine (mutated) at position 186 in the protein GLA. The side chain conformation was done manually.

Figure 20: Close-up of arginine (mutated) at position 186 in the protein GLA. The side chain conformation was done by SCRWL.

Figure 21: Close-up of the clash between the residues D153 (green) and R186 (red) in the protein GLA. (A) The two residues are shown in sticks representation. (B) Spheres representation of the two residues.

P205T (Mutation 6)

Close-up of the wildtype residue number 205 of GLA.

Close-up of the mutated residue number 205 of GLA.

D244H (Mutation 7)

Close-up of the wildtype residue number 244 of GLA.

Close-up of the mutated residue number 244 of GLA.

Q283P (Mutation 8)

Close-up of the wildtype residue number 283 of GLA.

Close-up of the mutated residue number 283 of GLA.

Q321E (Mutation 9)

Close-up of the wildtype residue number 321 of GLA.

Close-up of the mutated residue number 321 of GLA.

R363C (Mutation 10)

Close-up of the wildtype residue number 363 of GLA.

Close-up of the mutated residue number 363 of GLA.

Energy Comparison

The results of the energy comparison are presented in the table below. Due to the fact that the result of the first run of the eighth mutation clearly differed from the other results, the run was repeated with the outcome from the first run as input. Thus, there is the number 8.2. This observation shows that minimise has a decreased tolerance for clashes in comparison to the other tools. Their results for the eighth run are not outstanding and seem not to be affected by the fact that a proline was inserted into a helix. Furthermore, their results seem to be almost equally with respect to some variance. Only the comparison of the FoldX results of the mutations with the wildtype show, that the inserted mutations have a huge influence on the energy of the protein.

Number	AA-Position	Codon change	Amino acid change	SCWRL4	FoldX	FoldX - Diff	Minimise	Minimise - Diff
WT				-	-20.93	-	-20481.23	-
1	42	ATG-ACG	Met -> Thr	343.25	157.29	-178.22	-20324.41	-156.82
2	65	AGT-ACG	Ser -> Thr	327.798	152.87	-173.8	-20339.34	-141.89
3	117	ATT-AGT	Ile -> Ser	333.027	157.97	-178.9	-20353.47	-127.76
4	143	cGCA-ACA	Ala -> Thr	333.944	154.40	-175.33	-20339.32	-141.91
5	186	CAC-CGC	His -> Arg	323.717	154.57	-175.5	-20321.32	-159.91
6	205	gCCT-ACT	Pro -> Thr	340.619	155.96	-176.89	-20345.87	-135.36
7	244	gGAC-CAC	Asp -> His	333.594	152.08	-173.01	-20393.12	-88.11
8	283	CAG-CCG	Gln -> Pro	332.631	159.91	-180.84	-8027.71	-12453.52
8.2	-	-	-	-	-	-	-19134.48	-1346,95
9	321	tCAG-TAG	Gln -> Glu	332.853	160.95	-181.88	-20246.98	-234.25
10	363	TATa-TAA	Arg -> Cys	330.56	150.50	-171.43	-20295.77	-185.46

Gromacs

Figure 11: nstep vs. Elapsed Time in Gromacs.

Wildtype

Force Field	Average	Error Estimat	RMSD	Tot-Drift (kJ/mol)
Bond
AMBERGS	1826.99	420	4409.39	-2499.37
AMBER03	1639.74	410	4358.68	-2424.42
CHARMM27	2908.14	350	4779.8	-2033.44
Angle
AMBERGS	5496.47	74	476.18	408.548
AMBER03	5324.13	72	469.75	369.24
CHARMM27	7975.2	86	798.12	432.901
Potential
AMBERGS	-114713	1200	5648.79	-7915.46
AMBER03	-91307.7	1200	5559.82	-7839.05
CHARMM27	136.699	32	64.3892	227.896

Mutations

Force Field	Average	Error Estimat	RMSD	Tot/Drift
Bond
1	1815.39	570	5166.85	-3384.48
2	1862.77	610	5331.85	-3618.04
3	1773.13	520	4937.34	-3068.93
4	1828.63	580	5229.18	-3479.09
5	1870.95	610	5361.67	-3713.22
6	1816.6	550	5091.81	-3303.34
7	1819.7	570	5173.34	-3397.07
8	2992.15	1700	-nan	-10631.8
9	2083.16	830	-nan	-4913.82
10	1867.42	620	5390.82	-3693.03
Angle
1	5183.95	85	360.959	550.303
2	5195.33	80	364.473	515.645
3	5196.5	89	353.256	586.473
4	5175.59	85	364.496	547.465
5	5113.99	81	365.511	526.244
6	5200.44	85	356.964	553.934
7	5261.77	87	365.202	565.196
8	5178.73	76	-nan	215.036
9	5201.95	76	-nan	442.141
10	5174.48	88	375.775	555.294
Potential
1	-90528.4	1600	7234.09	-10149.1
2	-90481.9	1600	7442.03	-10340
3	-90654	1500	6928.73	-9614.54
4	-90541	1600	7311.04	-10343.7
5	-91011.7	1600	7484.45	-10592.5
6	-90782.2	1600	7226.99	-10188.5
7	-90232.9	1600	7236.24	-10198
8	-87316	3600	-nan	-23670.3
9	-90090.3	1900	-nan	-12335.3
10	-89721.8	1700	7523.88	-10750.1

Mutation	Plot
1	Gromacs energy calculation for mutation 1.
2	Gromacs energy calculation for mutation 2.
3	Gromacs energy calculation for mutation 3.
4	Gromacs energy calculation for mutation 4.
5	Gromacs energy calculation for mutation 5.
6	Gromacs energy calculation for mutation 6.
7	Gromacs energy calculation for mutation 7.
8	Gromacs energy calculation for mutation 8.
9	Gromacs energy calculation for mutation 9.
10	Gromacs energy calculation for mutation 10.