Improving the representation of peptide-like inhibitor and antibiotic molecules in the Protein Data Bank

Abstract

With the accumulation of a large number and variety of molecules in the Protein Data Bank (PDB) comes the need on occasion to review and improve their representation. The Worldwide PDB (wwPDB) partners have periodically updated various aspects of structural data representation to improve the integrity and consistency of the archive. The remediation effort described here was focused on improving the representation of peptide-like inhibitor and antibiotic molecules so that they can be easily identified and analyzed. Peptide-like inhibitors or antibiotics were identified in over 1000 PDB entries, systematically reviewed and represented either as peptides with polymer sequence or as single components. For the majority of the single-component molecules, their peptide-like composition was captured in a new representation, called the subcomponent sequence. A novel concept called "group" was developed for representing complex peptide-like antibiotics and inhibitors that are composed of multiple polymer and nonpolymer components. In addition, a reference dictionary was developed with detailed information about these peptide-like molecules to aid in their annotation, identification and analysis. Based on the experience gained in this remediation, guidelines, procedures, and tools were developed to annotate new depositions containing peptide-like inhibitors and antibiotics accurately and consistently.

Keywords: Protein Data Bank; peptide-like antibiotic; peptide-like inhibitor.

Figure 1

The chemical structures and sequences of (A) thiostrepton (PDB entry 1e9w) and (B) gramicidin S (PDB entry 1tk2). The chemical diagrams show cyclizations and modifications that lead to formation of the final antibiotic molecule. Red lines indicate the boundaries of the chemical components in the polymer, while the numbers indicate the correspondence with the gene or nonribosomal product.

Figure 2

Chemical structure of three trifluoroacetyl-dipeptide-anilide inhibitors of elastase. In each case, the inhibitor's chemical component name is followed by its subcomponents (following the colon). (A) Chemical structure of inhibitor 0Z1 from PDB entry 1ela; (B) inhibitor 0Z4 from PDB entry 1elb; and (C) inhibitor 0Z3 from PDB entry 1elc.

Figure 3

Representation of grouped peptide-like molecules. (A) Example of a peptide-like antibiotic that is a derivative of teicoplanin (from PDB entry 3vfj). The molecule shown here has lost a single chlorine atom during the experiment and is chemically different from naturally occurring teicoplanin. Both chemical structure and components list show that teicoplanin has a peptide core (shown in black), decorated with three saccharides (circled in blue, green, and orange) and a fatty acid (shown as R in red). The chemical components in the peptide core are numbered from 1 to 7. Bonds highlighted in green denote the peptide linkages between residues in the peptide core, while the purple bonds mark the covalent linkages between side-chains of the peptide residues. (B) Schematic representation of residues in the 22-mer “minigramicidin” (PDB entry 1kqe) showing two copies of the terminal 11-mer domains of gramicidin A, covalently linked in a head-to-head fashion. The linker between the two molecules is succinic acid (SIN).

Figure 4

Covalent binding of the peptide-like inhibitor ACE-ASP-GLU-VAL-ASJ to its target, caspase 3, in PDB entry 4dcp. While the carbonyl carbon (marked with an arrow) of the C-terminal residue in the unbound inhibitor (of type ASA) is sp², its hybridization state in the bound hemithioacetal product is sp³ and the residue is of type ASJ.

Figure 5

Chemical structures of four members of the vancomycin family of glycopeptide antibiotics. All these molecules have the same peptide core, and the different decorations in each of the molecules are circled using different colors. (A) Vancomycin aglycon has no sugars linked to it (PDB entry 1ghg). (B) Desvancosaminyl vancomycin is an intermediate in the vancomycin-biosynthesis pathway. It has only one saccharide linked to the peptide core (PDB entry 1rrv). (C) Vancomycin has a disaccharide decorating the peptide core (PDB entry 1aa5). (D) Chloroorienticin A has a disaccharide and a monosaccharide decorating the core (PDB entry 1gac).

Figure 6

A flow-chart showing the logic used for deciding the representation of peptide-like inhibitor and antibiotic molecules in the PDB.