Chains and Chain IDs

From Proteopedia

Chain Biochemistry

The term chain, in biochemistry, usually denotes either a polypeptide chain or a polynucleotide chain.

• Protein Chains: A polypeptide chain is a sequence of amino acids covalently linked by peptide bonds. When longer than 50 amino acids, it is called a Protein, whereas a short polypeptide consisting of 50 or fewer amino acids is termed a peptide. The chain structures of proteins are most easily visualized with backbone representations.
• Nucleic Acid Chains: A polynucleotide chain is a sequence of nucleotides covalently linked by ribose (or deoxyribose)-phosphodiester bonds, e.g. either DNA or RNA.

Polypeptide (protein) chains are linear, with rare exceptions where side-chains form protein crosslinks between two linear chains, such as disulfide bonds, or less commonly other types protein crosslinks of such as isopeptide bonds.

Each protein chain has two ends, an amino terminus (positively charged) and a carboxy terminus (negatively charged). The first residue in a protein chain becomes the amino terminus, with new amino acids being added at the carboxy terminus. The sequence of amino acids is specified by messenger RNA, which is a copy of the sequence of codons in the template strand of the DNA gene. The first residue in a nucleic acid chain becomes the 5' (phosphate) terminus, with new nucleotides being added at the 3' (hydroxy) terminus.

Protein molecules may consist of one or more polypeptide chains (see Protein primary, secondary, tertiary and quaternary structure). Those with more than one chain may be termed homo-oligomers or hetero-oligomers, homo-multimers or hetero-multimers. Functional forms of the molecule, termed biological units, often contain a different number of chains than does the crystallographic asymmetric unit. Examples are given in the article on biological units.

In a protein molecule consisting of multiple chains, the chains are usually held together by non-covalent bonds, but sometimes by covalent bonds, usually disulfide bonds. See quaternary structure.

Chain IDs

In the atomic coordinate files maintained by the wwPDB (PDB files), each polymer chain is given an ID, or chain "name". In the legacy PDB data format, chain IDs are a single letter or numeral (A-Z, a-z, 0-9), which limits the number of chains to 62. In the newer mmCIF data format (also called PDBx), chain IDs can be up to 4 letters or numerals, so the number of chains in a single structure is essentially unlimited (>10 million chains/structure could be accommodated by 4-character chain IDs).

Carbohydrate and Non-Polymer Chain IDs

An idiosyncracy of PDB files is that not only polymer chains, but all components in the structure model are assigned chain IDs, including carbohydrates (regardless of whether covalently-linked), ligands, metal ions, and water. The assignment of unique names to disaccharides and oligosaccharides began with the 2020 wwPDB Remediation of Carbohydrates. Notably, monomeric nucleotides and amino acids (not part of a polymeric chain) and monosaccharides are assigned the chain ID of the nearest protein or nucleic acid, while multimeric di- or oligo-nucleotides, and di- or oligosaccharides are given unique chain IDs. See item 2 below for the special case of dipeptides vs. tri- / oligo-peptides.

The procedure for assigning chain IDs is specified in the wwPDB Procedures section 6. Chain ID assignment. In February, 2025 that document needs two corrections in order to agree with actual wwPDB practice:

1. When protein or nucleic acid is present, ligands and water bound to carbohydrate are never assigned the chain ID of that carbohydrate, but are given the chain ID of the nearest protein/nucleic acid, even when it is >5 Å away (examples: 7LKC, 7dc4, 8g82). When the structure is carbohydrate without any protein or nucleic acid, only then are ligands and water given the chain ID of the nearest carbohydrate (examples: 1c58, 2kqo).
2. Although dinucleotides and disaccharides are assigned unique chain IDs, dipeptides are not. Traditionally deemed ligands, dipeptides are assigned the ID of the polymer chain to which they are bound. Dipeptide examples are 2cyh and 1dpp; tripeptide: 4q1L. This is correctly documented at the wwPDB in 3. Polymer sequences and sequence database reference assignment. Tripeptides and longer oligopeptides are assigned unique chain IDs.

Author vs. wwPDB Chain IDs

Regardless of the chain IDs assigned by the authors of a structure model entry deposited in the wwPDB, the wwPDB assigns its own (usually distinct) chain IDs. These PDBx chain IDs are present only in the mmCIF files, not in the PDB format files. In mmCIF files, there is no single place that lists all author-assigned or all wwPDB-assigned chain IDs.

• Author-assigned chain IDs:
  • Protein & Nucleic Acids: _pdbx_poly_seq_scheme.pdb_strand_id
  • Carbohydrates: _pdbx_branch_scheme.pdb_asym_id (supercedes .auth_asym_id after 2020 remediation)
  • All else: _pdbx_nonpoly_scheme.pdb_strand_id
• wwPDB-assigned chain IDs:
  • Protein & Nucleic Acids: _pdbx_poly_seq_scheme.asym_id
  • Carbohydrates: _pdbx_branch_scheme.asym_id
  • All else: _pdbx_nonpoly_scheme.asym_id

Monomeric ligands and other non-polymers that are assigned the chain name of the associated polymer are always assigned the author-assigned polymer IDs, when the model is compatible with the PDB format -- and hence always get single-character names.

When the entry is not compatible with PDB format because it has >62 chains, some author-assigned chain names will have at least two characters.

wwPDB-assigned chain IDs are all upper case letters, and are assigned systematically in this order: A-Z, AA-ZA, BA-ZA, ... AZ-ZZ, AAA-ZAA, ABA-ZBA, etc. Single character IDs cover 26 chains. Double letter IDs cover 26²=676. Triple-letter IDs cover 26³=17,576. Four-letter IDs are also allowed, covering an additional 456,976 IDs.

When a biological unit has more chains than the asymmetric unit, the additional wwPDB-assigned chain IDs are added following the above order. In the examples below, the ratio of biological unit polymer chains to asymmetric unit polymer chains are:

• 1hho 2x: 2 -> 4 polymer chains ("polymer chains" means protein or nucleic acid only).
• 1igt 1x: 4 -> 4 polymer chains.
• 4nia 4x: 60 -> 240 polymer chains.
• 8qqj 1x: 31 -> 31 polymer chains.
• 8qhu 1x: 86 -> 86 polymer chains.
• 7o0u 1x: 86 -> 86 polymer chains.

The numbers of polymer chains in biological units are given in the token _pdbx_struct_assembly.oligomeric_count. Symmetry operations for generating biological units are given in the category _pdbx_struct_oper_list.

Examples

These examples show that it is much easier to work with author-assigned chain IDs than with wwPDB-assigned chain IDs, despite the latter being systematically assigned. Luckily, Jmol & JSmol, iCn3D, and ChimeraX use author-assigned chain IDs. PyMOL selects single-character author-assigned chain IDs, but reports both types of chain IDs. Molstar reports both chain IDs.

* Not compatible with PDB format because of its 62-chain limit (A-Z, a-z, 0-9). Some author-assigned chain IDs must have at least 2 characters.

See Also

• Backbone representations
• Introduction to molecular visualization
• Jmol/Visualizing large molecules