02 - Protein preparation from RCSB¶
Crystallographic structures from the Protein Data Bank require extensive preprocessing before use in molecular dynamics simulations. Raw PDB files contain artifacts, missing atoms, and structural inconsistencies that must be addressed systematically. Our pipeline consists of five distinct stages, each handling specific aspects of structure preparation and simulation setup.
The complete workflow proceeds through: (1) protein structure cleaning and optimization, (2) force field parameterization and solvation, (3) energy minimization preparation, (4) equilibration protocol setup, and (5) production simulation configuration. Each stage builds upon the previous one, ensuring reproducibility and maintaining careful control over simulation conditions.
Initial Structure Acquisition and Processing¶
The preparation begins with downloading the target structure (1JC0) from the RCSB Protein Data Bank. This particular structure represents the reduced form of roGFP2, which serves as our reference state for Cu(I) binding investigations. The crystallographic structure contains multiple chains, solvent molecules, and crystallization artifacts that require systematic removal or modification.
export PDB_ID="1JC0"
wget https://files.rcsb.org/download/$PDB_ID.pdb -O structures/protein/0-$PDB_ID.pdb
The initial download yields the complete crystallographic asymmetric unit, including multiple protein chains and extensive crystallographic waters.
For our simulation purposes, we focus exclusively on chain A, which contains the complete roGFP2 structure with the characteristic chromophore and cysteine residues critical for Cu(I) coordination.
Chain Selection and Water Retention¶
Chain selection removes extraneous protein chains while preserving crystallographic waters that may play important structural roles. The metalflare-select-atoms utility employs MDAnalysis selection syntax to isolate chain A while maintaining all solvent molecules:
We will only be working with one protein and need to select chain A.
We use the metalflare-select-atoms script which just drives pdb.select.run_select_atoms().
metalflare-select-atoms $SAVE_DIR/0-$PDB_ID.pdb $SAVE_DIR/1-$PDB_ID-chain-A.pdb --select_str "chainID A"
We keep all crystallographic waters so the CRO-coordinating water's position is maintained.
This selective approach is crucial because crystallographic waters often occupy functionally important positions, particularly those coordinating with the chromophore or stabilizing key structural elements. Removing these waters indiscriminately can create artificial cavities or destabilize important hydrogen bonding networks during subsequent simulation.
Structure Filtering and Coordinate Optimization¶
The filtering step removes non-essential PDB records, retaining only ATOM and HETATM entries necessary for molecular dynamics. This cleanup eliminates crystallographic metadata, experimental annotations, and other records that can interfere with downstream force field assignment:
Following filtration, the structure undergoes geometric centering to position the protein at the coordinate system origin. This centering simplifies subsequent box construction and ensures consistent placement across different simulation systems:
Rotational Optimization for Solvation Efficiency¶
The rotational optimization step represents a critical but often overlooked aspect of simulation preparation. The metalflare-minimize-box utility systematically tests different protein orientations to identify the configuration that minimizes the required solvation volume:
This reduction significantly decreases computational requirements while maintaining proper solvation around all protein surfaces. The algorithm evaluates multiple rotational configurations and selects the orientation that achieves the most compact solvation shell without creating buried surfaces or artificial contacts.
Residue Numbering and Chemical State Assignment¶
Crystallographic structures often contain inconsistent residue numbering or non-standard residue designations that must be standardized for force field compatibility. The unification process ensures sequential residue numbering and consistent chain identifiers:
Chemical Modifications and Protonation State Management¶
Selenomethionine Conversion¶
Crystallographic structures frequently contain selenomethionine (MSE) residues introduced during crystallization for phasing purposes. These artificial modifications must be reverted to native methionine residues for accurate simulation:
Cysteine State Assignment¶
The roGFP2 structure contains two critical cysteine residues (Cys147 and Cys204) that coordinate Cu(I) binding. These residues must be properly configured in their reduced, thiol-bearing state to enable metal coordination. The deprotonated cysteine state (CYM) is assigned to these specific residues:
metalflare-rename-resname 5-1JC0-residues.pdb CYS CYM --include 147 204 --output 5-1JC0-residues.pdb
This assignment is crucial because standard cysteine residues (CYS) model the protonated thiol form, which cannot coordinate metals effectively. The deprotonated cysteine state (CYM) provides the anionic sulfur atoms necessary for Cu(I) coordination chemistry.
Glutamic Acid Protonation¶
Specific glutamic acid residues may require protonation state adjustment based on local electrostatic environments and pH considerations. Residue 220 is converted to the protonated form (GLH) to maintain proper charge balance and local structural stability:
Automated Protonation State Prediction¶
The PDB2PQR program provides automated protonation state assignment for ionizable residues based on local electrostatic environments and pH conditions. This tool addresses the fundamental challenge that crystallographic structures lack hydrogen atoms, which are essential for accurate force field modeling:
pdb2pqr --log-level INFO --ff=AMBER --keep-chain --ffout=AMBER 5-1JC0-residues.pdb 6-1JC0-pdb2pqr.pdb
PDB2PQR employs sophisticated algorithms to predict optimal protonation states for histidine, aspartic acid, glutamic acid, and lysine residues. The program considers local electrostatic interactions, hydrogen bonding opportunities, and bulk pH conditions to determine the most probable protonation pattern. However, PDB2PQR cannot process non-standard residues like the GFP chromophore, requiring manual intervention for these special cases.
The merge operation combines PDB2PQR output with the original structure to recover any atoms that PDB2PQR could not process:
Water Molecule Standardization¶
Crystallographic structures may contain water molecules with various naming conventions (HOH, TIP, TIP3). Standardization to WAT ensures consistent recognition by Amber force field tools:
metalflare-rename-resname 6-1JC0-pdb2pqr.pdb HOH WAT --output 7-1JC0-resnames.pdb
metalflare-rename-resname 7-1JC0-resnames.pdb TIP WAT --output 7-1JC0-resnames.pdb
metalflare-rename-resname 7-1JC0-resnames.pdb TIP3 WAT --output 7-1JC0-resnames.pdb
metalflare-unify-waters 7-1JC0-resnames.pdb --output 7-1JC0-resnames.pdb
The unification process also ensures that water molecules have consistent atom naming and connectivity, preventing force field assignment errors during subsequent parameterization steps.
Final Amber Compatibility Processing¶
The pdb4amber utility performs final compatibility checks and modifications to ensure full compatibility with Amber force field tools:
This tool addresses Amber-specific requirements such as terminal residue modifications, special residue recognition, and atom naming conventions that differ from standard PDB formatting. The processed structure becomes the final input for force field parameterization and system construction.