Updated restraint dictionaries for carbohydrates in the pyranose form

Abstract

Restraint dictionaries are used during macromolecular structure refinement to encapsulate intramolecular connectivity and geometric information. These dictionaries allow previously determined `ideal' values of features such as bond lengths, angles and torsions to be used as restraint targets. During refinement, restraints influence the model to adopt a conformation that agrees with prior observation. This is especially important when refining crystal structures of glycosylated proteins, as their resolutions tend to be worse than those of nonglycosylated proteins. Pyranosides, the overwhelming majority component in all forms of protein glycosylation, often display conformational errors in crystal structures. Whilst many of these flaws usually relate to model building, refinement issues may also have their root in suboptimal restraint dictionaries. In order to avoid subsequent misinterpretation and to improve the quality of all pyranose monosaccharide entries in the CCP4 Monomer Library, new dictionaries with improved ring torsion restraints, coordinates reflecting the lowest-energy ring pucker and updated geometry have been produced and evaluated. These new dictionaries are now part of the CCP4 Monomer Library and will be released with CCP4 version 8.0.

Keywords: carbohydrates; dictionaries; pyranose; restraints; ring conformation.

Figure 1

A view of the patched torsion section in a CIF restraint dictionary entry. This is an extract of the new CCP4 restraint dictionary entry for N-acetyl-β-d-glucosamine (GlcNAc), which is represented in the PDB as ‘NAG’. The new dictionaries distinguish ring torsion angles (prepended by ‘ring_’) from the rest (‘tors_’) so they can be activated separately to keep a low-energy ring pucker. Older CCP4 dictionaries had no separation between the ring torsions (unimodal) and the rest of the torsions (periodicity 2, 3 or 6), and had a uniform uncertainty of 20.0°.

Figure 2

Carbohydrate restraint dictionary entries generated with AceDRG. (a) α-d-Glucose in the ⁴ C ₁ conformation, (b) 3,4,5-trideoxy-α-d-erythro-oct-3-en-2-ulopyranosonic acid in the ^O H ₅ conformation, (c) α-l-fucose in the ¹ C ₄ conformation and (d) N-acetyl-α-neuraminic acid (sialic acid) in the ¹ C ₄ conformation. This figure was produced with CCP4mg (McNicholas et al., 2011 ▸).

Figure 3

Numbers of sugars diagnosed by Privateer as ‘check’, ‘no’ and ‘yes’ before and after refinement. A set of structures from the PDB were refined with the CCP4-ML dictionaries, the new dictionaries generated by AceDRG and the new dictionaries with unimodal torsion restraints activated. From left to right, the coloured bars represent the number of sugars before refinement (grey) and the numbers of analysed sugars after refinement with the CCP4-ML dictionaries (red), after refinement with the new updated dictionaries (blue) and after refinement with the new dictionaries with activated unimodal torsion restraints (yellow). (a) shows all analysed pyranosides and (b) only includes pyranosides that were diagnosed as ‘check’ or ‘no’ for at least one protocol.

Figure 4

Refinement with the new dictionaries and unimodal torsion restraints leads to fewer unlikely carbohydrate conformations. (a) Sugars that are part of N/O-glycosylation; (b) other sugars. θ versus φ plot for d-sugars (blue circles) and l-sugars (yellow triangles); see Section 4 for a description of θ and the Cremer–Pople parameters. d-Sugars usually adopt the ⁴ C ₁ conformation with θ ≃ 0°; l-sugars normally adopt the ¹ C ₄ conformation with θ ≃ 180°. Use of the new unimodal torsion restraints (top) shows fewer deviations from these values. The PDB codes corresponding to entries discussed in Figs. 5 ▸ and 6 ▸ are labelled. The number of sugars in high-energy conformations (according to Privateer) is shown in the bottom right corner of each plot. Resolution ranges contain equal numbers of sugars (1668 each). High resolution is 0.9–1.8 Å, medium resolution is 1.8–1.9 Å and low resolution is 1.9–2 Å.

Figure 4

Refinement with the new dictionaries and unimodal torsion restraints leads to fewer unlikely carbohydrate conformations. (a) Sugars that are part of N/O-glycosylation; (b) other sugars. θ versus φ plot for d-sugars (blue circles) and l-sugars (yellow triangles); see Section 4 for a description of θ and the Cremer–Pople parameters. d-Sugars usually adopt the ⁴ C ₁ conformation with θ ≃ 0°; l-sugars normally adopt the ¹ C ₄ conformation with θ ≃ 180°. Use of the new unimodal torsion restraints (top) shows fewer deviations from these values. The PDB codes corresponding to entries discussed in Figs. 5 ▸ and 6 ▸ are labelled. The number of sugars in high-energy conformations (according to Privateer) is shown in the bottom right corner of each plot. Resolution ranges contain equal numbers of sugars (1668 each). High resolution is 0.9–1.8 Å, medium resolution is 1.8–1.9 Å and low resolution is 1.9–2 Å.

Figure 5

Sugars in unusual conformations after refinement with the new dictionaries with unimodal torsion restraints. (a) BMA-B-3 from PDB entry 5jug (Jin et al., 2016 ▸); (b) SIA-A-522 from PDB entry 6hg0 (M. T. Salinger, J. R. Hobbs, J. W. Murray, W. G. Laver, P. Kuhn & E. F. Garman, unpublished work); (c) NAG-E-1 from PDB entry 5o7u (Tobola et al., 2018 ▸); (d) GLC-C-1 from PDB entry 5upm (Pluvinage et al., 2017 ▸). These sugars appear as outliers in Fig. 4 ▸(b). They remain in high-energy conformations after refinement, but have a high RSCC. This figure was produced with CCP4mg (McNicholas et al., 2011 ▸). Maps are displayed at the 1σ contour level with a sampling rate of 0.5.

Figure 6

Change in conformation and real-space correlation coefficient (RSCC) after refinement. (a) Sugar in a ¹ S ₅ conformation after refinement with its old CCP4-ML restraint dictionary entry (Fig. 4 ▸ a, bottom middle panel). (b) The conformation of the sugar has been changed to the minimal energy conformation after refinement with the updated restraint dictionary entry and unimodal torsion restraints and the RSCC has increased (Fig. 4 ▸ a, top middle panel). The sugar in (a) and (b) is MAN-Q-4 from PDB entry 4uo0 (Devi et al., 2015 ▸) at 1.90 Å resolution, mean B value 34 Ų. (c) Sugar in a ² S _O conformation after refinement with its old CCP4-ML restraint dictionary entry (Fig. 4 ▸ a, bottom middle panel). (d) The minimal energy conformation of the sugar after refinement with the new restraint dictionary entry and unimodal torsion restraints; the RSCC has decreased (Fig. 4 ▸ a, top middle panel). The sugar in (c) and (d) is BMA-P-3 from PDB entry 4iic (Suzuki et al., 2013 ▸) at 1.90 Å resolution, mean B value 18 Ų. This figure was produced with CCP4mg (McNicholas et al., 2011 ▸). Maps are displayed at a 1σ contour level with a sampling rate of 0.5.