NAME
Hbond - locate possible hydrogen bonds in a protein
structure
SYNOPSIS
hbond -options files
DESCRIPTION
Hbond is a program to read in Brookhaven format files and to
calculate all the possible hydrogen bonds and other
sidechain interactions of interest. The output from hbond is
a file with a .hbd extension that is primarily meant to be
used as input for joy. If you want to use the file for
other purposes it is important that you understand the
assumptions used to define the hydrogen bonds. It is your
responsibility to make sure that the input Brookhaven file
is o.k., i.e. it should not have alternate atoms, explicit
hydrogen atoms, etc; preprocessing of the Brookhaven file
with pdb2atm(1) is recommended.
The following options to hbond are available:
-h Print a help message on the standard output.
-v Be verbose in what is reported to the user, this is
mainly of use to see what is actually being read by
hbond.
-V Print the version number of hbond on the standard
output.
WHAT IS A HYDROGEN BOND ?
In the interests of simplicity, a hydrogen bond (or
analogous interaction) is flagged if the distance between
the two participants in the interaction are less than a set
distance apart (usually 3.5). For disulphide bonds the
criteria is that the two sulphur atoms are less than 3.0
apart (note, this is a generous cutoff). Due to the larger
Van der Waal's radius of sulphur atoms a 4.0 cutoff is used
for interactions involving this element type. No check is
made on the lower bound of the distance. Interactions to
HETATM records are always calculated. Some interactions
also have the relevent angles calculated, however this has
no influence as to whether an atom pair occurs in this file
(and is considered as a hydrogen bond). Note that angles
that are undefined are written as 999.99.
For mainchain to mainchain hydrogen bonds, energies are
calculated as in the paper of Kabsch and Sander.
The following points may need to be considered before you
use the program. The first is that the amide sidechains of
asparagine and glutamine are assumed to be undetermined (
i.e. you can swap the NH2 and O atoms freely), thus both
atoms are considered as donors and acceptors. The second
point is that the y atom of cysteine is considered to be a
possible donor and acceptor atom, additionally the d sulphur
of methionine is an acceptor. Finally, the protonation state
of an acidic residue (aspartic acid and glutamic acid is
considered as undefined), i.e. it can be protonated and so
again either of the oxygen atoms can be both donors and
acceptors. Finally, (again) no interactions between residues
are considered for atoms in residues closer than two
residues from the central residue.
Because of the nature of the cutoff, it is likely that not
all the interactions reported are likely to be real hydrogen
bonds. However, the definition is rigid and easily
defendable. Other methods are likely to produce lists of
hydrogen bonds that are simply subsets of the ones reported
by hbond. However, it is relatively easy either to write a
program that will select out the hydrogen bonds from the
.hbd file, or to change the cutoffs internally within the
program.
FORMAT OF THE .hbd FILE
The .hbd file contains a small header, whether HETATM
records were included in the analysis. There the follows a
list of the possible interactions, listed one per line, with
the donor listed first followed by the acceptor atom. The
format for the data in the .hbd file is (FORTRAN format):
1X, I3, 1X, A5, 1X, A1, 1X, A, 1X, A3, 1X, I3, 1X, A5, 1X,
A1, 1X, A, 1X, A3, 1X, A3, 1X, I4, 1X, F4.2, 1X, F4.2, 1X,
F6.2, 1X, F6.2
Where the fields are; the index of the first residue, the
number of the donor residue, the chain-name of the donor
residue, the residue type of the donor residue, and the atom
type of the donor atom; this is followed by a similar set of
values for the acceptor atom. This information is followed
by the class of the hydorgen bond (see table below), an
integer indicating the soacing of the donor and acceptor
atoms (can be negative), the distance between the donor and
acceptor atoms, the distance between the hydrogen atom and
the acceptor (see below for how hydrogen positions are
calculated), the angle between the donor, hydrogen and
acceptor and finally the angle between the donor, acceptor
and carbonyl carbon. An example of a .hbd file is:
# produced by hbond, version 1.0c (2.3b)
# parameterized for 600 residues, 20 atoms per residue, and 500 hetatoms
#
# coordinate data taken from file 2ptn.atm
# number of atoms = 1610
# number of residues = 221
# number of chains = 2
# chain 1 extent from 1 to 4, of length 5
# chain 2 extent from 5 to 221, of length 216
# number of hetatoms = 1
#
# criteria for hydrogen bond definition :-
# donor-acceptor distance cutoff (oxygen and nitrogen) = 3.50
# donor-acceptor distance cutoff (sulphur) = 4.00
# hydrogen-acceptor distance cutoff = none
# angular criteria applied = none
# energy criteria applied = none
# include HETATM records = T
# include WATER records = T
#
#----- Donor ----- ---- Acceptor --- ------- Geometry ------ - Energy -
#index - res - atm index - res - atm typ span Dd-a Dh-a <d-H-N <a-O=C kJ/mol
1 16 I N 120 142 G O MM 119 3.29 9.99 999.99 999.99 999.99
1 16 I N 121 143 N O MM 120 3.03 9.99 999.99 999.99 999.99
1 16 I N 174 194 D OD1 SN 173 2.72 9.99 999.99 120.26 * 999.99
1 16 I N 174 194 D OD2 SN 173 3.48 9.99 999.99 81.36 * 999.99
2 17 V N 169 189 D O MM 167 2.82 1.85 166.93 166.76 -3.01
4 19 G N 2 17 V O MM -2 3.42 3.40 83.36 71.35 .97
5 22 C N 133 155 L O MM 128 3.08 9.99 999.99 999.99 999.99
5 22 C SG 133 155 L O SO 128 3.52 9.99 999.99 89.43 999.99
5 22 C SG 134 156 K O SO 129 3.41 9.99 999.99 75.51 999.99
5 22 C SG 135 157 C SG DS 130 2.04 9.99 999.99 999.99 999.99
The hydrogen bonds are classified using the following scheme:
center;
l l.
class meaning
MM mainchain to mainchain
MH mainchain to hetero atom
MW mainchain to water
SO sidechain to mainchain CO
SN sidechain to mainchain NH
SS sidechain to sidechain
SH sidechain to hetero atom
SW sidechain to water
DS disulphide
CAVEATS
Please bear in mind what the program is meant to do, before
blindly using its output. The program is likely to be
changed in the near future. The main limit you are likely to
encounter is a maximum numer of residues of 500.
SEE ALSO
pdbseq(1), joy(1).
RELEASE LEVEL
This document describes hbond version 0.7 and later.