The WHAT IF 'HBONDS' module contains two separate sets of options:
older 'HBO' routines, and newer 'HB2' routines.
The big advantage of the first set is that these are extremely
fast. Disadvantage is that they use a Yes/No criterium to decide
whether something is a hydrogen bond or not.
The big advantage of the second set is that the hydrogen bonds are
scored using a special Force Field that was developed for WHAT IF
using a database of accurately determined
small molecule structures. Disadvantage is that it
can be very slow. Actually, it is very fast, but it has to do so many
things that it takes a long time anyway.
When WHAT IF determines if a hydrogen bond can be formed between the hydrogen
of the donor atom and the lone pair of the acceptor atom, it uses four
parameters. These are:
1) Distance between the donor and acceptor atom.
2) Distance between the (calculated) hydrogen position, and the acceptor atom.
3) Angle from donor atom over the hydrogen to the acceptor atom.
4) Angle from the hydrogen over the acceptor to a 'virtual' atom. If the
acceptor is only covalently bound to one atom, this atom is the so-called
virtual atom. If the acceptor is covalently bound to two atoms, the
virtual atom is on the bisector of those two.
Hydrogen bonds are placed according to the following algorithm:
If the geometry fixes the hydrogen position, this position is used, wherby the
donor hydrogen distance is set to 1.0 Angstrom. If the hydrogen has a degree
of rotational freedom, then the cone on which the hydrogen can
potentially be found is calculated. This cone has a top angle of two times
sixty is one hundred twenty degrees. The hydrogen is now placed on the two
points that this cone has in common with the plane through the donor, a point
on the rotation axis of the cone and the acceptor. Figure 15.1 in the figure
book shows for all 20 amino acids where the hydrogens are placed. WHAT IF
only uses hydrogens that can be involved in hydrogen bonds. The cysteine
side chain is not considered for hydrogen bond calculations.
Any constellation that creates a donor/hydrogen/acceptor triplet that falls
within the four values described above will be accepted as a hydrogen bond.
There are several special cases to be thought about:
Water can be used either as a donor, or as an acceptor. Since it is
not possible to determine the hydrogen positions, the angular error
over the hydrogen will be zero in case water functions as a donor,
whereas the angular error over the acceptor will be zero in case water
is the acceptor in the hydrogen bond.
Hydrogen bond statistics and geometry are very well described by
Baker and Hubbard.
The best source to find out about the algorithms and parameters used is
our paper. It will only be very briefly described here.
Hydrogen bonds are scored with values between 0...1. The score for a
particular hydrogen bond is determined from the donor/acceptor types,
the H-acceptor distance, the donor-H-acceptor angle, and the position
of the H with respect to the acceptor.
A special feature of the HB2 options is that they consider the H atom
explicitly. This makes the scoring function much better, but is the
source of the long computation times for the HB2 options: the H atoms
are not just put somewhere, but the total hydrogen bond energy is
optimized by searching the best possible positions for all hydrogen
atoms SIMULTANEOUSLY. During this optimization process HIS, ASN, and
GLN side chains are allowed to flip 180 degrees around their last Chi
angle (normally crystallography can not distinguish these two
situations for these residues).
The command SHOHBO will cause WHAT IF to prompt you for two ranges. It
will then list all hydrogen bonds between these two ranges. The two
ranges may overlap, or be identical. For each hydrogen bond the four
geometric parameters will be shown, together with the used
corresponding hydrogen coordinates. Sometimes both hydrogens from one
donor can form a hydrogen bond with the same acceptor. In these cases
the residue and atom name related information will only be listed
once, followed by two lines with the geometric parameters and hydrogen
coordinates. This same is true if two different hydrogen bonds are
possible between the same two atoms (eg. two serine O-gamma's, or
serine O-gamma with water). At the end WHAT IF will display the
hydrogen bonds only if you have already used any graphics option in
this session. If you do not want to display the hydrogen bonds, you
can, as usual, give zero as mol-object number. At the end WHAT IF will
ask if you want to see the statistics for the listed hydrogen
bonds. Here you get a large amount of counting statistics, and some
averages. Be aware that the averages, when waters are included, are
biased towards perfectness.
All possible hydrogen bonds will be listed. See the HB2NET, HB2LIS,
and HB2XST options for getting a similar list for the result of a
network H-Bond analysis.
The option GRAHYD will cause WHAT IF to prompt you for a residue range.
All hydrogens within this range that potentially could be involved in a
hydrogen bond will be displayed. On machines with more
tahn 8 bitplanes the hydrogens get a colour that
is slightly different from the atom they are attached to. Hydrogens for
which a fixed position is not dictated by its donor heavy atom,
will be represented by a small circle in space that shows all places
where this hydrogen could potentially be found.
See the HB2NET and HB2GRA options for an alternative, showing all
hydrogens and hydrogen bonds after a H-bond network analysis.
The command SLFHBO will cause WHAT IF to prompt you for a residue
range. It will than make two passes over all residues in this range,
In the first pass all intra residue hydrogen bonds will be listed.
In the second pass all hydrogen bonds between covalent neighbour residues
will be listed. At the end you will be prompted for a mol-object number
and for statistics output as described for the SHOHBO option.
The command SHOPAR will cause WHAT IF to show the hydrogen bond acceptance
parameters at the terminal. See figure 15.2 in the figure book of the
writeup for the exact meaning of these parameters.
The command INIPAR will cause WHAT IF to set the hydrogen bond acceptance
parameters back to their default parameters. These are: donor-acceptor
distance = 3.5 A; hydroge-acceptor distance = 2.5A; angular error over
the hydrogen = 60 degrees; angular error over the acceptor = 90 degrees.
The command PARAMS brings you, as usual, in the parmeter menu. In this menu
you can change the four hydrogen bond acceptance parameters, and the way
WHAT IF shows hydrogen bonds at the display. Here you can choose between
a dotted line between the donor and the acceptor atom without showing
the hydrogen, and showing the hydrogen and drawing a dotted line between
the hydrogen and the acceptor (this is the default).
The parameter DONACD is the upper limit for the distance between the donor
atom (the heavy atom, not the hydrogen) and the acceptor atom. The default
value for DONACD is 3.5 A.
The parameter HYDACD is the upper limit for the hydrogen to acceptor
distance. In case the hydrogen position is not explicitly known, a position
will be calculated for it such that the hydrogen to acceptor distance becomes
minimal. The default value for HYDACD is 2.5 A.
The parameter ANGERA determines the maximal allowed deviation from perfect
for the angle over the acceptor free orbitals. This is the angle
acceptor-orbital-hydrogen. The default value is 90 degrees.
The parameter ANGERH determines the maximal allowed deviation from perfect
for the angle acceptor-hydrogen-donor. The default is 60 degrees.
LINTYP determines how the hydrogen bonds are bein displayed. By default
(LINTYP=1) WHAT IF draws a solid line from the donor to the (putative) hydrogen
position, and than a dashed line from the hydrogen to the acceptor. If LINTYP
is set at 0 (zero) only a dashed line from the donor to the acceptor will
be drawn.
If you have used the display already in this WHAT IF session, ONEHBO will
cause WHAT IF to ask you to pick two atoms, otherwise you will be prompted
for two atoms. In case you give a histidine side chain nitrogen, you will
be asked if this nitrogen should be used as a donor (protonated) or as an
acceptor (not protonated). WHAT IF will then calculate the four
hydrogen acceptance parameters for this atom pair, and display them. This
is a nice option to figure out why something is NOT a hydrogen bond.
I helps putting the molecule up at the screen first. (Picking WAIT helps
solving your problems when you forgot to do that first).
The option 1ATHBO can be used to determine all hydrogen bonds made by one
atom. You will be asked to pick an atom. If there is no graphics device
present you will be prompted to type residue and atom name.
Thereafter you will be prompted
for a range. All hydrogen bonds between the picked atom and the given
range will be listed and displayed similar as by the SHOHBO option.
I helps putting the molecule up at the screen first. (Picking WAIT helps
solving your problems when you forgot to do that first).
The command SPHHBO can be used to list and display all hydrogen bonds
in the environment of a certain residue. You will be prompted for a residue
and for a sphere radius. All hydrogen bonds between residues that fall
aproximately within this sphere radius will be listed and displayed.
WARNING: The hydrogen bond menu interpretation of a sphere is
not exactly the same as the graphics menu interpretation of a sphere.
This option should therefore only be used to evaluate the direct
environment of one (central) residue.
To show hydrogen bonds for only residues that have a certain secondary
structure, you can use the command HSTHBO. This option will prompt you
for the secondary structure type. Acceptable input is H, S, T, C and the
asterisk for helix, strand and sheet, turn, coil or everything
respectively.
Thereafter WHAT IF will execute the SHOHBO option. This means you will
be prompted for two ranges, all hydrogen bonds are shown and displayed,
and if wanted, statistics will be shown.
If the secondary structure was not yet calculated, it will be done
automatically. Both residues should have the requested secondary structure,
but do not necessarily have to sit in the same secondary structure element.
All (backbone and sidechain) atoms are used.
If you want to evaluate backbone-backbone (BB), backbone-sidechain(BS),
or sidechain-sidechain (SS) hydrogen bonds, you can use the
command BSSHBO. This option will prompt you for the kind of
hydrogen bonds you want to see. It will then proceed as for the SHOHBO
option. This means you will
be prompted for two ranges, all hydrogen bonds are shown and displayed,
and if wanted, statistics will be shown.
The command 1AAHBO will cause WHAT IF to prompt you for one amino acid.
If the graphics screen has already been used in this session, you will
be asked to pick an amino acid.
Thereafter you will be prompted for a residue range. All hydrogen bonds
between the given amino acid and the range will be listed and displayed.
I helps putting the molecule up at the screen first. (Picking WAIT helps
solving your problems when you forgot to do that first).
The command FILHBO will cause WHAT IF to prompt you for a file
name. This file should hold one line per requested hydrogen bond. The
format of these lines is 4A4 (that means four groups of four characters)
for the first residue, atom in first residue, second residue, and atom
in second residue respectively. The original residue number as read from
the PDB file should be given, and not WHAT IF's sequential numbers
in the file. The original numbers are always shown in brackets by WHAT IF.
The four hydrogen bond determining
parameters for every requested hydrogen bond will be calculated.
All requested hydrogen bonds will be accepted by WHAT IF, regardless
how long, or how bad the angles are.
All lines in the input file starting with an asterisk will be traeted
by WHAT IF as commentary, listed in the output, and skipped.
If the line in the input file starts with END, execution of this option
will stop.
The command HB2INI will (re-) initialize the HB2 module based on the
current soup. If you have used HB2 options before, but have loaded
another molecule into the soup now, or have changed torsion angles,
or the symmetry flags (with USESYM in the symmetry menu),
you need to execute this option before any of the other
HB2 options, otherwise WHAT IF might express its disgrunt in a
undesirable fashion. In principle you don't need this option for every day usage.
The command HB2RES will reset all coordinates to what they were at the
time HB2INI was executed. After HB2RES all H-atom
positions are lost. This option does NOT unlock the symmetry matrices if they
are locked!
The command HB2NET will search for the best possible hydrogen bond
network. If HB2INI has not been executed yet, that will be done
first. After that a lot of trace output will be shown on the screen,
and in the end the total amount of hydrogen bonds will be shown on the
screen. The numbers shown here are not very interesting, but this
option needs to be executed before any of the HB2L?? options and the
HB2XST option. Execution with symmetry enabled (see USESYM in the
symmetry menu) is highly
advisable.
Warning 1: It is assumed that the soup contains one protein, not a set
of overlapping NMR models. Also, results are basically meaningless if
there are other gross problems in the structure (bumps,...).
Warning 2: This option leaves the soup in an 'ideal' state. This means
that after executing HB2NET the HIS, GLN, and ASN residues that
have a better hydrogen bond scheme if they are turned around WILL be
turned around in the soup. See HB2RES.
Warning 3: After execution of this option with symmetry enabled, the
symmetry matrices will be locked (see SYMLCK), such that symmetry
transformations in any listings you generate later are
meaningful. You'll have to unlock the transformations manually if you
need to change the symmetry.
Warning 4: When there are a lot of water molecules in the soup,
executing this option can take a considerable amount of time.
Warning 5: After %USESYM, you HAVE to run HB2INI again.
Warning 6: If you run HB2NET twice in a row, then the total number of
hydrogen bonds is higher at the beginning of the second run than at the end
of the first run in case residues were flipped in the first run. This is because
in the second run the flip-penalties are no longer used. If you get flipped
residues in the second run, that either means that the flip is less probable,
or that the molecule is too complex for HB2NET's default parameters.
This option does exactly the same as MAKMOL, but it writes out all
polar hydrogen coordinates as well. Hydrogen atoms will not have a
name, but they will be numbered. All hydrogen atoms are given directly
after the atom they are connected to. The resulting file could e.g. be
used as input for further modeling with Molecular Dynamics programs.
It probably makes no sense to run this option without optimizing the
H-bond network first using HB2NET.
HB2GRA asks you for two ranges of residues, and creates a graphics
object containing all the polar hydrogen atoms in the first and all
hydrogen bonds between the first and the second range.
It probably makes no sense to run this option without optimizing the
H-bond network first using HB2NET.
HB2LIS asks you for two ranges of residues, and lists all hydrogen bonds
from any residue in the first range to any residue in the second.
It probably makes no sense to run this option without optimizing the
H-bond network first using HB2NET.
HB2LFR shows the residues that have been flipped by 'HB2NET', and all
goups of donors that have an 'unusual' amount of hydrogens attached.
An example of the latter: histidines with H atoms attached to ND1 as
well as NE2.
It probably makes no sense to run this option without optimizing the
H-bond network first using HB2NET.
HB2LUH shows all donors that have an accessibility of 0.0, but are
nevertheless not participating in a hydrogen bond. This should be very
rare. For all donors listed a search is performed for possible
acceptors; if there are any, they will be listed too.
It probably makes no sense to run this option without optimizing the
H-bond network first using HB2NET.
HB2LUA shows all acceptors that have an accessibility of 0.0, but are
nevertheless not participating in a hydrogen bond. This should be
relatively rare (not as rare as for donors). For all acceptors listed
a search is performed for possible donors; if there are any, they will
be listed too.
It probably makes no sense to run this option without optimizing the
H-bond network first using HB2NET.
HB2LUA shows some sums and averages of different types of hydrogen bonds
in the structure.
It probably makes no sense to run this option without optimizing the
H-bond network first using HB2NET.
If you know (e.g. from experimental results) that a certain hydrogen
bond is present in the protein, or you just want to experiment, the
option HB2AWB can be used to give any possible hydrogen bond a weight
of 10 x normal. This will not guarantee that it will be present after
optimization with HB2NET, but can make it much more likely.
You need to run HB2INI before this option.
This option is only available if MORE has been activated in the
HBONDS menu.
This option clears the list of wanted hydrogen bonds created with
HB2AWB.
This option is only available if MORE has been activated in the
HBONDS menu.
HB2LBF lists all bifurcated hydrogen bonds. That is: all hydrogens that
are donated to more than one acceptor at the same time.
It probably makes no sense to run this option without optimizing the
H-bond network first using HB2NET.
This option is only available if MORE has been activated in the
HBONDS menu.
Normally WHAT IF assumes that crystallographers have looked at the
hydrogen bonds surrounding HIS, GLN, and ASN residues. Whenever there
is reasonable 'energetic doubt' whether a flip of a HIS, GLN, ASN
residue is necessary, WHAT IF will choose for the original. This is
decided by the FLIP penalty: a flipped residue is penalized by a
certain fraction of a hydrogen bond value (default 0.2 is
approximately 1.2 kCal/mol). Using HB2SPN you can change this, either
for all residues, or for one single residue.
You need to run HB2INI before this option.
This option is only available if MORE has been activated in the
HBONDS menu.