Tutorial : fitting residues into electron map density
In this example, we will learn how to guess what peptide
fragment would best fit into the density, given a poly-Ala and the sequence
of the protein to solve.
Step by Step
open file pdb file 1LDB_ala (provided with the
open the dn6 map 1LDB.dn6, and set the following parameters into the
dialog: Show unit Cell, display around CA with 8.0 Å around x y
and z, and contour at 0.8 sigma (the fitting works better when ample space
within the density is given to accomodate residues, and a contouring value
below 1 sigma is usually recommended).
Load raw sequence of the protein to solve (1LDB.txt) from the SwissModel
menu and display the Layers Infos Window (from the display menu). Remove
the visible checkmark for the layer 1LDB.txt as we need it only to have
the sequence of the protein to solve.
Enable the slab mode (from the Display menu) and center the protein on
the Ala63 Alpha Carbon by option clicking (right mouse button for PC)
on Ala63 in the control panel.
Select Ala61 to Ala65 using the control Panel, and use the "Find best
fitting peptides item of the Build menu". Be a little patient, and a result
file will be opened:
DAMDF mismatch: 0 In: 19 Out: 1 Score: 11.20
LARAF mismatch: 0 In: 18 Out: 2 Score: 8.57
KGATY mismatch: 0 In: 15 Out: 2 Score: 8.54
RAILH mismatch: 0 In: 18 Out: 4 Score: 8.31
ERDVY mismatch: 0 In: 21 Out: 6 Score: 8.15
DILTY mismatch: 1 In: 18 Out: 5 Score: 7.89
The files gives you the list of existing pentapeptides that would fit
onto the backbone fragment you selected, the number of residues that do
not match well, the number of atoms inside and outside the density, and
a score value. Results are sorted by score, but you should also consider
the other parameters when looking for the best solution (and of course
rotate the protein and actually try the various solutions.
Click on the first line to have a look at the solution. Explore the various
best solutions (you can either click on a solution, or use the up and
down arrows while the text window is active. Be careful, because the up
and down arrows will have the efect of changing the sigma contouring value
when the Text window is not active.
Now display the Align window and compare your solution with the actual
sequence of the solved protein to look whether you guessed right (you
could have a view like this one in stereo).
Check that the sigma contouring value is still set to 0.8, else set it
to 0.8. Now center the protein on Ala35, select residues Ala32 to Ala37,
hit the return key to display only this section of the protein, and use
the "Find best fitting peptides" again.
ANESK mismatch: 0 In: 15 Out: 2 Score: 8.91
KNRFH mismatch: 1 In: 20 Out: 9 Score: 8.71
ARFRF mismatch: 0 In: 22 Out: 7 Score: 8.64
RIFVN mismatch: 0 In: 20 Out: 5 Score: 8.24
TNPVD mismatch: 1 In: 17 Out: 4 Score: 7.98
GARVV mismatch: 0 In: 11 Out: 3 Score: 7.67
MDFNH mismatch: 1 In: 19 Out: 6 Score: 7.44
ASYVF mismatch: 0 In: 16 Out: 5 Score: 7.27
AIFRS mismatch: 0 In: 17 Out: 4 Score: 7.20
PKPVD mismatch: 1 In: 17 Out: 5 Score: 7.13
CRDAD mismatch: 1 In: 15 Out: 3 Score: 7.13
Again, try to guess what is the best fragment. This illustrate that the
top ranking fragment is not always the best one, but usually, the best
one is among the top ones.
Now repeat the process for residues Ala38 to Ala42.
And finally, repeat the process for residues Ala43 to Ala47. Surprise:
this time the correct fragment is not among the top ranking ones. Let's
see why. Select residues 43 to 46 only, and find the best fragments. This
time, the correct solution is among the top ranking fragments. So what
happens with residue 47 ?
Click on the mutate tool in the main display window, and then click on
residue 47. Mutate to Glu, and browse the rotamer library. As you can
see, no rotamer fits even approximately into the density. Select rotamer
1, and accept the mutation. Now click on the torsion tool, and select
residue 47. Use the little arrows that have appeared at the right of the
torsion tool to alter the chi-1 and chi-2 angles of the rotamer until
it fits into the density.
Now, reset the orientation, load 1LDB.pdb and compare your fit with the
actual solved structure.
This ends this tutorial section.