Carbohydrate Remediation

As the PDB archive grows, and the related science and techniques evolve, the 3D structures represented in the Core Archive require ongoing improvement ("remediation") to ensure consistency, accuracy, and overall quality. This contributes to the goals of the wwPDB to make PDB data Findable, Accessible, Interoperable and Reusable (FAIR). While numerous tools exist to curate protein 3D structural data, few have been developed for consistent standard representation and for checking the veracity of carbohydrate 3D structural data within wwPDB deposition-validation-biocuration processes.

Therefore a new data representation for carbohydrates has been developed, as a result of consultation with the glycoscience community, the wwPDB PDBx/mmCIF Working Group and other key stakeholders. Working with the glycoscience community, carbohydrate-appropriate annotation tools were developed and implemented within the wwPDB OneDep unified system for deposition, validation, and biocuration of structures coming into the PDB Core Archive. These software tools provide standard nomenclature and consistent oligosaccharide representation that can be easily translated to other representations commonly used by glycobiologists.

This effort involved the following steps: (1) standardizing sugar nomenclature following IUPAC/IUBMB; (2) providing a uniform representation for oligosaccharides with appropriate descriptor(s); (3) adopting glycoscience community software for linear descriptors of oligosaccharides; (4) providing intra- and inter-molecular connectivity at atom level explicitly; and (5) providing glycosylation annotation and its linkage information. Due to the nature of carbohydrates, which are usually branched and/or have different glycosidic linkages, additional data descriptors have been introduced into the PDBx/mmCIF dictionary to describe new representation of oligosaccharides as branched entities. The branch components (monosaccharides) and their glycosidic bonds have been explicitly listed. The PDBx/mmCIF dictionary extensions, substantial collection of example PDBx/mmCIF files of remediated data representing various cases, and the corresponding proposed Chemical Component Dictionary definitions were provided in the early 2020 at github for testing and adoption by key stakeholders during the development stage, including the glycoscience community, refinement software developers, cheminformaticians, and 3D visualization software.

Oligosaccharide molecules are classified as a new entity type, branched, assigned a unique chain ID (_atom_site.auth_asym_id) and a new mmCIF category introduced to define the type of branching (_pdbx_entity_branch.type) . The monosaccharides constituting an oligosaccharide entity are listed in the _pdbx_entity_branch_list category. Unlike polypeptides, where the peptidic bond is implicitly assumed, the connectivity of the glycosidic bond for an oligosaccharide entity is explicitly described at the entity level in the mmCIF category _pdbx_entity_branch_link. Two or more monosaccharides connected with standard glycosidic bonds are considered as branched entity of oligosaccharides which could be linear or branched. The numbering of monosaccharides within a branched entity starts at 1 sequentially from reducing to non-reducing end following IUPAC ordering scheme (2-Carb-37.2 for linear oligosaccharides and 2-Carb-37.3 for branched oligosaccharides).

Any changes made to the data are recorded in the PDBX_AUDIT_REVISION data category. These remediated files are accessible at FTP and versioned FTP with the following audit record for this release.

_pdbx_audit_revision_details.provider repository

_pdbx_audit_revision_details.type Remediation

_pdbx_audit_revision_details.description 'Carbohydrate remediation'

Lists of PDB entries, CCDs, PRDs, and non-PDB compatible entries are created for this remediation.

Figure 1. Oligosaccharide representation compared to protein/nucleic acid representation, branched vs polymer.

For example, the oligosaccharide, Lewis Y blood group antigen with components (FUC)-NAG-GAL-FUC in PDB entry 2WMG will be represented as below branched entity.

#   2WMG
1 polymer     man 'FUCOLECTIN-RELATED PROTEIN'             66769.758 1   ? YES 'CATALYTIC MODULE, RESIDUES 31-589' ? 
2 branched    nat "alpha-L-fucopyranose-(1-2)-beta-D-galactopyranose-(1-4)-[alpha-L-fucopyranose-(1-3)]2-acetamido-2-deoxy-beta-D-glucopyranose" 681.7     1   ? ?   ?                                   ? 
3 water       nat water                                    18.015    334 ? ?   ?                                   ? 
_pdbx_entity_branch.entity_id   2
_pdbx_entity_branch.type         oligosaccharide
2 NAG 1   n 
2 GAL 2   n 
2 FUC 3   n 
2 FUC 4   n 
1 2 2 GAL C1 O1 1 NAG O4 HO4 sing ?
2 2 3 FUC C1 O1 2 GAL O2 HO2 sing ?
3 2 4 FUC C1 O1 1 NAG O3 HO3 sing ?

To minimize impact to PDB Core Archive users, the current struct_conn that describes connectivity at structure level will remain as is, with extension to explicitly describe the leaving atom and its stereochemical configuration. (N.B.: Oligosaccharides are not represented in entity_poly_seq, hence sequential sequence numbering in _atom_site.label_seq_id will be null). Based on community feedback, the glycosylation site and its binding type (N-linked glycan or O-linked glycan) needs to be identified in a machine readable format in order to be easily parsed by users. The item struct_conn.pdbx_role is used for binding site and is a controlled vocabulary.

For example, the glycosylation sites are identified in _struct_conn.pdbx_role for PDB entry 4OF3.

covale1  covale Y ASN A  93  ND2 ? ? ? OXT N 1_555  NAG C 801 C1 ? ? ? O1 R 1_555  N-Glycosylation 1.438 sing ?
covale2  covale Y ASN A 206  ND2 ? ? ? OXT N 1_555  NAG D 804 C1 ? ? ? O1 R 1_555  N-Glycosylation 1.415 sing ?
covale3  covale Y ASN B  93  ND2 ? ? ? OXT N 1_555  NAG E 801 C1 ? ? ? O1 R 1_555  N-Glycosylation 1.434 sing ?

The linear sequence descriptor of an oligosaccharide based on observed sequence will be described in _pdbx_entity_descriptor category so that it is interoperable with various representations that are used in the glycoscience community.

For example, the linear sequence descriptor for the oligosaccharide, (FUC)-NAG-GAL-FUC, is described in _pdbx_entity_descriptor.descriptor for PDB entry 2WMG.

1 2 "LFucpa1-2DGalpb1-4[LFucpa1-3]DGlcpNAcb1-ROH" 
      "Glycam Condensed Sequence" GMML   	1.0      
2 2 
"WURCS=2.0/3,4,3/[a2122h-1b_1-5_2*NCC/3=O][a1221m-1a_1-5][a2112h-1b_1-5]/1-2-3-2/a3-b1_a4-c1_c2-d1" WURCS                   	PDB2Glycan 1.1
3 2 
"[][b-D-GlcpNAc]{[(3+1)][a-L-Fucp]{}[(4+1)][b-D-Galp]{[(2+1)][a-L-Fucp]{}}}"                    	LINUCS                  	PDB-CARE   ?

In addition, we will standardize wwPDB Chemical Component Dictionary (CCD) so that all sugars are represented as monosaccharides with IUPAC style residue naming and standard atom nomenclature following IUPAC-IUBMB. For pyranoses, the leaving atom is O1, while for furanoses and sialic acid derivatives, the leaving atom is O2. These standardized atom nomenclatures will facilitate systematic numbering of components in oligosaccharides and also in validation of carbohydrate structures.

Figure 2. Standard atom nomenclature of pyranose and furanose ring sugars.

To support findability, wwPDB will preserve the commonly used names for oligosaccharides through extending an existing reference dictionary of biologically interesting molecules. To minimize the impact on the coordinate section (atom_site) of the PDB Core Archive files and on other software used in the field such as refinement programs or 3D visualization tools, the three letter code residue names will be retained in _atom_site.label_comp_id. We will introduce IUPAC style naming using existing data item, pdbx_chem_comp_identifier in the CCD for consistent annotation in the PDB Core Archive atomic coordinate files. This information will be copied to the PDB Core Archive atomic coordinate files as metadata.

For example, the IUPAC style naming b-D-GlcpNAc or GlcpNAcb or N-acetyl-b-D-glucopyranose-osamine for NAG are provided with source provenance identified.

NAG "SYSTEMATIC NAME"             ACDLabs              12.01
NAG "SYSTEMATIC NAME"             "OpenEye OEToolkits" 1.7.6
NAG "COMMON NAME"                 GMML                 1.0  
NAG "IUPAC CARB SYMBOL"           PDB-CARE             1.0  
NAG "SNFG CARB SYMBOL"            GMML                 1.0  

We will take this opportunity to improve management of synonyms in the wwPDB CCD during this remediation. We will move synonym list out of _chem_comp.pdbx_synonyms to a better organized table structure, pdbx_chem_comp_synonyms with source provenance.

For example, each synonym is listed at each row with data source identified for ligands ROC and for 58A.

ROC  Fortovase       DRUGBANK ?
ROC  "RO 31-8959"    ? ?
58A cytidinediphosphate-dioleoylglycerol   PUBCHEM ?
58A CDP-1,2-dioleoyl-sn-glycerol   CHEBI ?

Based on the feedback from glycoscience community, the relationship between a modified monosaccharide and its core structure (standard monosaccharide) is described in _pdbx_chem_comp_related category and the atom-to-atom mapping is provided in _pdbx_chem_comp_atom_related category.

For example, the core structure for modified monosaccharide, SGN(N-sulfo-6-O-sulfo-alpha-D-glucocosamine), is PA1(alpha-D-glucosamine) with "carbohydrate core" identified as relationship type.

_pdbx_chem_comp_related.comp_id		      SGN
_pdbx_chem_comp_related.related_comp_id	      PA1
_pdbx_chem_comp_related.relationship_type     'Carbohydrate core'
_pdbx_chem_comp_related.details		      ?
 1  SGN C1   PA1 C1   "Carbohydrate core"
 2  SGN C2   PA1 C2   "Carbohydrate core"
 3  SGN C3   PA1 C3   "Carbohydrate core"
 4  SGN C4   PA1 C4   "Carbohydrate core"
 5  SGN C5   PA1 C5   "Carbohydrate core"
 6  SGN C6   PA1 C6   "Carbohydrate core"
 7  SGN N2   PA1 N2   "Carbohydrate core"
 8  SGN O1   PA1 O1   "Carbohydrate core"
 9  SGN O3   PA1 O3   "Carbohydrate core"
10 SGN O4   PA1 O4   "Carbohydrate core"
11 SGN O5   PA1 O5   "Carbohydrate core"
12 SGN O6   PA1 O6   "Carbohydrate core"
13 SGN H1   PA1 H1   "Carbohydrate core"
14 SGN H2   PA1 H2   "Carbohydrate core"
15 SGN H3   PA1 H3   "Carbohydrate core"
16 SGN H4   PA1 H4   "Carbohydrate core"
17 SGN H5   PA1 H5   "Carbohydrate core"
18 SGN H61  PA1 H61  "Carbohydrate core"
19 SGN H62  PA1 H62  "Carbohydrate core"
20 SGN HN21 PA1 HN21 "Carbohydrate core"
21 SGN HO1  PA1 HO1  "Carbohydrate core"
22 SGN HO3  PA1 HO3  "Carbohydrate core"
23 SGN HO4  PA1 HO4  "Carbohydrate core"




LINUCS:, Bohne-Lang A, Lang E, Forster T, von der Lieth CW. LINUCS: linear notation for unique description of carbohydrate sequences. Carbohydr Res 2001, 336: 1-11.

GlycanBuilder2: Tsuchiya S, Aoki NP, Shinmachi D, et al. Implementation of GlycanBuilder to draw a wide variety of ambiguous glycans. Carbohydr Res. 2017;445:104‐116. doi:10.1016/j.carres.2017.04.015


The carbohydrate remediation project is a wwPDB collaborative project that is carried out principally by RCSB PDB at Rutgers, The State University of New Jersey and is funded by NIH Common Fund Glycoscience Program through the National Cancer Institute cooperative agreement U01 CA221216 to Dr. Robert Woods at Complex Carbohydrate Research Center at the University of Georgia in collaboration with Dr. Jasmine Young as sub-awardee at RCSB PDB at Rutgers, The State University of New Jersey.