Secondary structure restraints

At lower resolution (lower than ~3A) refinement of biological macromolecules may be unstable. In some cases even secondary structure (SS) elements may be distorted during coordinate refinement. SS elements include helices and sheets in protein structures, base-pairs and stacking bases in nucleic acids. This applies to refinement against both X-ray and EM data. To preserve correct geometry of SS it may be useful to use specifically designed secondary structure restraints. Using SS restraints in moderate-to-low resolution refinement may help to preserve correct geometry of SS elements in the structure and sometimes correct slightly distorted SS elements.

General description

SS restraints are available in most refinement tools in Phenix package. Their behaviour is regulated by the same scope of parameters (see the bottom of the page for the full listing). They could be turned on in GUI by checking "Use secondary structure restraints" checkbox. This is equivalent to command-line option secondary_structure.enabled=True. In this case SS elements will be found automatically and outputted to the resulting .pdb and log files. We recommend to check carefully automatic assignments of SS elements because incorrect assignments can lead to wrong refinement results.

Proteins

Secondary structure elements of proteins are helices and sheets. The main stabilizing factor is their hydrogen bonds.

The PDB format supports 10 different types of helix, but only three of these (alpha, pi, and 3_10) are common in naturally occurring proteins and have easily deciphered hydrogen-bonding rules. User-specified helices default to alpha form (hydrogen bonds between residue n and residue n+4). Sheets may be either parallel or antiparallel.

The most common way to specify secondary structure is to use HELIX and SHEET records in the beginning of .pdb file. There are several ways to obtain such records. Examples of such records:

HELIX    1  HA GLY A   86  GLY A   94  1                                   9
SHEET    1   A 5 THR A 107  ARG A 110  0
SHEET    2   A 5 ILE A  96  THR A  99 -1  N  LYS A  98   O  THR A 107
SHEET    3   A 5 ARG A  87  SER A  91 -1  N  LEU A  89   O  TYR A  97
SHEET    4   A 5 TRP A  71  ASP A  75 -1  N  ALA A  74   O  ILE A  88
SHEET    5   A 5 GLY A  52  PHE A  56 -1  N  PHE A  56   O  TRP A  71

Full reference to HELIX and SHEET records PDB format may be found here Hydrogen bond lengths for protein secondary structure are restrained at 2.9A with sigma 0.05A for N-O. Outliers (defined as N-O bonds greater than 3.5A) will be automatically removed, but you may prevent this with secondary_structure.protein.remove_outliers=False. This may be helpful for very poor low-resolution structures, e.g. to force proper structure on a manually built helix, but it will also cause completely spurious bond restraints to be overlooked. Three hydrogen bond angles are also restrained for alpha-helices. The first and the last bond angles are restrained with RMSD of 10 degrees, inner angles restrained with RMSD of 5 degrees. Target values are the following:

These angle restraints can be turned off by setting secondary_structure.protein.restrain_hbond_angles=False.

How to obtain HELIX/SHEET records

There are various programs available that can produce HELIX/SHEET records suitable for use in refinement. There are several ways to obtain HELIX/SHEET records for particular structure:

  • Check your .pdb file. They may be already there
  • Run any refinement program from Phenix suite with secondary structure restraints turned on and look in the resulting .pdb file
  • Use phenix.secondary_structure_restraints command-line utility.
  • Use any other third-party program that have an option to export secondary structure definition in PDB format.

It is highly recommended to check manually all SS annotations coming from any source. They may often contain errors. We encourage users to verify all secondary structure annotations carefully with model. The format of HELIX/SHEET records is described here. The most common errors are: - Lack of registration part of SHEET record (second part of SHEET record when sense is not 0) - Strands are not combined into sheets, i.e. all SHEET records have different - Wrong sense in SHEET record - Not necessary pi-helices containing only 3 residues

Nucleic acids

Secondary structure elements for nucleic acids are base-pairs and stacking pairs. Base-pairs are supposed to be somewhat planar and stacking nucleobases are supposed to be parallel to each other. Several types of restraints to make these features present in resulting model are available in Phenix. None of them are symmetry-aware, e.g. for PDB ID 9dna basepair hydrogen bonds could not be determined automatically. Nevertheless, they still could be specified manually via geometry_restraints.edits mechanism.

Base-pairs

Restraints to keep correct geometry of base-pairs involves hydrogen bonding, planarity and paralleltiy restraints. Hydrogen bond restraints include distance and angle restraints. Two types of restraints available to keep base-pairs planar: standard planarity restraints and parallelity restraint. We use parallelity by default. User can change it to planarity or even use both of them (not recommended) separately for each defined base-pair in parameter file.

Stacking pairs

We use parallelity restraint to keep stacked pairs parallel to each other.

How to obtain annotations

The annotation could be done automatically during refinement and it will be printed out to .def file for future runs. Another way is to use phenix.secondary_structure_restraints command-line utility which can produce annotations in suitable format. PDB does not have suitable records, therefore the only way now to supply custom annotations is to use Phenix parameter files.

Graphical editor

Documentation coming soon!

References

  • Programming new geometry restraints: parallelity of atomic groups. O.V. Sobolev, P.V. Afonine, P.D. Adams, and A. Urzhumtsev. J. Appl. Cryst. 48, 1130-1141 (2015).

List of all available keywords

  • secondary_structure
    • ss_by_chain = True Find secondary structure only within individual chains. Alternative is to allow H-bonds between chains. Can be much slower with ss_by_chain=False. If your model is complete use ss_by_chain=True. If your model is many fragments, use ss_by_chain=False.
    • max_rmsd = 1 Maximum rmsd to consider two chains with identical sequences as the same for ss identification
    • use_representative_chains = True Use a representative of all chains with the same sequence. Alternative is to examine each chain individually. Can be much slower with use_representative_of_chain=False if there are many symmetry copies. Ignored unless ss_by_chain is True.
    • max_representative_chains = 100 Maximum number of representative chains
    • enabled = False Turn on secondary structure restraints (main switch)
    • protein
      • enabled = True Turn on secondary structure restraints for protein
      • search_method = *ksdssp mmtbx_dssp from_ca cablam Particular method to search protein secondary structure.
      • distance_ideal_n_o = 2.9 Target length for N-O hydrogen bond
      • distance_cut_n_o = 3.5 Hydrogen bond with length exceeding this value will not be established
      • remove_outliers = True If true, h-bonds exceeding distance_cut_n_o length will not be established
      • restrain_hbond_angles = True
      • helix
        • serial_number = None
        • helix_identifier = None
        • enabled = True Restrain this particular helix
        • selection = None
        • helix_type = *alpha pi 3_10 unknown Type of helix, defaults to alpha. Only alpha, pi, and 3_10 helices are used for hydrogen-bond restraints.
        • sigma = 0.05
        • slack = 0
        • top_out = False
        • hbond
          • donor = None
          • acceptor = None
      • sheet
        • enabled = True Restrain this particular sheet
        • first_strand = None
        • sheet_id = None
        • sigma = 0.05
        • slack = 0
        • top_out = False
        • strand
          • selection = None
          • sense = parallel antiparallel *unknown
          • bond_start_current = None
          • bond_start_previous = None
        • hbond
          • donor = None
          • acceptor = None
    • nucleic_acid
      • enabled = True Turn on secondary structure restraints for nucleic acids
      • hbond_distance_cutoff = 3.4 Hydrogen bonds with length exceeding this limit will not be established
      • angle_between_bond_and_nucleobase_cutoff = 35.0 If angle between supposed hydrogen bond and basepair plane (defined by C4, C5, C6 atoms) is less than this value (in degrees), the bond will not be established.
      • base_pair
        • enabled = True Restraint this particular base-pair
        • base1 = None Selection string selecting at least one atom in the desired residue
        • base2 = None Selection string selecting at least one atom in the desired residue
        • saenger_class = 0 Type of base-pairing
        • restrain_planarity = False Apply planarity restraint to this base-pair
        • planarity_sigma = 0.176
        • restrain_hbonds = True Restrain hydrogen bonds length for this base-pair
        • restrain_hb_angles = True Restrain angles around hydrogen bonds for this base-pair
        • restrain_parallelity = True Apply parallelity restraint to this base-pair
        • parallelity_target = 0
        • parallelity_sigma = 0.0335
      • stacking_pair
        • enabled = True Restraint this particular base-pair
        • base1 = None Selection string selecting at least one atom in the desired residue
        • base2 = None Selection string selecting at least one atom in the desired residue
        • angle = 0
        • sigma = 0.027