This tutorial is based on a crystal structure of sucrose-phosphatase
(spp) from Synechocystis sp. pcc6803 in a closed
conformation, PDB code 1tj3. The crystal space group is P6522
and data resolution 2.8 Å.
There are several homologous structures available in both the open
and closed conformations. For this exercise we will choose an
open-conformation model with 100% sequence identity, PDB code 1s2o.
The file 1s2oA.pdb contains chain A of the crystal structure. Check
with Coot that there are two obvious domains, which are also
provided in separate files, 1s2oA_dom1.pdb and 1s2oA_dom2.pdb.
This example demonstrates the MR search using electron density maps
from partial structure. Three different protocols for this type of
search are implemented in Molrep (see comments in §4). These methods
can be really useful for large structures containing several
This particular example is an easy case. In particular, for Phaser
it is a very straightforward run with two ensembles corresponding to
two different domains.
2. Checking the data
Matthews suggests one molecule in the asymmetric unit.
Sfcheck shows no particular problems with data.
There is a space group ambiguity. The space group cannot be selected
from the absences alone; P6522 and P6122
would both require that the only reflections observed along the c*
axis are 0 0 6n. In real experiment, you should check both space
groups and see which one gives the best result for the translation
search. If you are a very cautious person, you may wish to check all
space groups consistent with the point group - in this case P622,
P6122, P6522, P6222,
P6422 and P6322.
In Molrep interface, the space group can be selected in "Search
Options > Used SG" menu. Note that in the "MR using Phases" mode
this menu is not active.
3. Structure solution
Try to solve the structure using Molrep and the complete
Refine the MR solution
Automatic MR using the whole 1so2 model finds the solution, but
the subsequent refinement does not improve the R-factors
much, and the maps show no reasonable density for the small
domain. However a clear density for the first domain is a good
Therefore in the second attempt we use individual domains as
search models. In this process, the phases from the partial
model containing the larger domain are used in the search for
the smaller domain.
Position the first domain, 1s2oA_dom1.pdb.
From solvent content analysis, Molrep will assume that there are
two copies of the model in the asymmetric unit. Set the correct
number of copies explicitly to save some time: "Search Options
> Number of copies to find 1"
Refine the partial model and compare refinement statistics
Position the second domain using phases from partial structure
Select "MR Using Phases"
Select output mtz-file from (4) as input Data; mtz-columns
will be set automatically to FWT and PHWT
Select the second domain 1s2oA_dom2.pdb as the Model
Select output pdb-file from (4) as Fixed model
Refine the structure and check with Coot that the whole
structure matches the maps
4. Notes on Spherically Averaged Phased Translation Function
The standard Molrep protocol uses phase information from the fixed
model in the Translation Function only. This is a preferred
procedure until 1/3 of the structure is positioned. (You can try it:
select "Molecular replacement", use the model from step (3) or (4)
as Fixed model and search for the second domain.)
In (5), the phase information was used in both RF and TF. The 2Fo-Fc
and Fo-Fc map coefficients were generated by Refmac after refinement
of the structure containing the first domain. Molrep calculated one
of these maps and set electron density to zero in the area occupied
by the partial structure. Then Molrep uses the structure amplitudes
from the masked map in conventional Patterson based RF, and both
amplitudes and phases in Phased TF.
In the Molecular Replacement with SAPTF, the phase information is
used directly in both RF and TF steps but now the order of steps is
changed and an additional Phased TF step is added to confirm and
adjust the initial positions:
SAPTF: overlap function is calculated between spherically
averaged density around this point and the model density. Some
of SAPTF peaks correspond to a correct position of the model
regardless of the model orientation.
Phased Rotation Functions with the centres of rotation at
Phased Translation function to refine the position of the
5. Finding solution for the second domain using SAPTF