open your .pdb file containing the protein structure with selenomethionine residues using your favorite text editor.

serarch for selenomethionine residues using the search function and MSE string

example:

ATOM 987 N MSE A 130 22.981 77.122 151.979 1.00 134.80 N

ATOM 988 CA MSE A 130 23.053 78.542 152.305 1.00 142.62 C

ATOM 989 C MSE A 130 22.961 78.756 153.813 1.00 144.95 C

ATOM 990 O MSE A 130 23.580 79.671 154.360 1.00 154.36 O

ATOM 991 CB MSE A 130 21.943 79.318 151.593 1.00 152.32 C

ATOM 992 CG MSE A 130 22.056 79.317 150.073 1.00 154.45 C

ATOM 993 SE MSE A 130 20.604 80.273 149.182 1.00 198.00 Se

ATOM 994 CE MSE A 130 21.081 79.894 147.328 1.00 124.46 C

replace MSE with MET and Se with S and SE with SD (indicated in bold above and below)

ATOM 987 N MET A 130 22.981 77.122 151.979 1.00 134.80 N

ATOM 988 CA MET A 130 23.053 78.542 152.305 1.00 142.62 C

ATOM 989 C MET A 130 22.961 78.756 153.813 1.00 144.95 C

ATOM 990 O MET A 130 23.580 79.671 154.360 1.00 154.36 O

ATOM 991 CB MET A 130 21.943 79.318 151.593 1.00 152.32 C

ATOM 992 CG MET A 130 22.056 79.317 150.073 1.00 154.45 C

ATOM 993 SD MET A 130 20.604 80.273 149.182 1.00 198.00 S

ATOM 994 CE MET A 130 21.081 79.894 147.328 1.00 124.46 C

repeat until you changed all MSE to MET and all Se to S.

save your modified ,pdb under a different name.

EDIT:

thanks Rogelio for pointing this out, I originally omitted the atom type, which could cause problems. fixed.

The simple way would be to open the pdb file in model building software COOT, and then use the mutate and auto-fit tool from calculate tab.

Hey Appu,

yes, that is one way to do it. However, coot will change the coordinates of your MET sidechain atoms by doing so. I guess the OP wants to refine against native data? In that case that shouldn't matter. 

If I use coot to replace the Se I usually double check with a text editor to make sure I didn't miss MSE in the file.

phenix.pdbtools can convert SeMet too....phenix does everything these days.

If you computer is UNIX compliant (i.e. linux, MacOS X , free BSD or equivalent), open a terminal, change to the folder containing your file (cd path-to-your-folder) and use sed:

 sed 's/\([AH][TE][OT][MA][ T][ M]......\)SE\(...\)MSE\(........................................................\)SE/\1SG\2MET\3 S/ ; s/\([AH][TE][OT][MA][ T][ M]...........\)MSE/\1MET/ '  inputfile.pdb > outputfile.pdb

(the above is one line command in the terminal plain text, without end-of line characters, just change inputfile.pdb  to whatever the name of your file is, and and outputfile.pdb to wherever your want to put the modified file).

If your PDB is properly qualified the above would change all MSE into MET, the SE atom type into SG and the SE element into S. The modified file would be outputfile.pdb

If your also want to change other data data (such as SEQRES, LINK, and ANISOU records), you would need to add more substitution commands. In that case probably awk would be better, but programming it may take a bit more time.

Make sure the number of points, slashs, backslashs, backets, quotes, blanks and so on are preserved exactly.

Best wishes

Rogelio

Stefan, Appu, and Rogelio, thank you very much. If using SAD solving protein structure, native protein crystal X-ray data is not necessary?

Lixia,

Well, that depends on what you would like to show with that structure. Native data can have better quality than SAD data and many people do the native data by default.

Stefan,

Thank you very much. You help me a lot.

Lixia,

I agree with Stefan. However, if you do not go on to obtain native data, then when you report the structure you need to have SeMet built in your model.

Nektaria

Hey Stefan, 

I agree with you. I was under the impression that Pan collected the high resolution native data and want to use the phase from the SAD data. But, even if he has native data collected from the selenomethionine derivatized crystal, he needs to refine methionine as SeMet because methionines would not have Se atom, and refining as Met would give difference Fourier density. 

Thanks Stefan Gajewski,

happy to know that the method I followed  is being suggested here.  But selenium having double the mass of sulfur, wont it make any variation if protein's function..??

Aditya,

I don't know of people using Selenium derivative protein for anything else but experimental phasing in crystallography. Se vs. S structures are usually highly similar.

You would expect there to be a difference though, the question you are asking is: is that difference big enough to lead to an effect that falls within the detection limits of your assays?

Stefan

It appears to me your problem is only correcting the final structure data to fit the native protein. That would be easily done by Editing the PDB file with a text editor (plain text, don't use word).

Alternatively, you can use Yasara viewer, Chimera, VMD, Swiss PDB viewer, CUT or Maestro (amongst some other programs) to "mutate" the MSE to MET, you just need to learn to use them. If you did the structure solution and refinment yourself, you probably know CUT.

 HOW TO EDIT THE PDB

replace the name of residues MSE to MET for all atoms

MSE

with

MET

replace the name of the Selenium atom in atom name columns

SE

with

SG

Replace the atom symbol at columns 77-78

SE

to

 S

Notice that this last change requires you to put space in column 77 and S in column 78

make sure to keep the column positions, i.e. do not add or remove spaces other than the one mentioned above. PDB format has strict rules and requires every label or number to be at specific positions in the line. If you have doubt on the PDB site you can find the PDB format specifications.

Best wishes

Rogelio

Thank you Stefan for the clarification. My doubt was since selenium has twice the mass of sulphur won’t it occupy more space than sulfur. If so, then is it not it has affected the surrounding atoms and the final structure.?

And thank you Rogelio. I was able to replace with methionine in a text editor. Though se-met is a naturally occurring amino acid since it is not being coded by the genetic code it can not be in a protein. Whenever we find se-met (SME) then we have to replace it with MET for functional studies right..?

Selenium has significant differences in its oxidation/reduction potentials and is a softer base. In its free form is extremely toxic to living cells, and it can be incorporated in proteins as Selenecysteine, which may be coded in the genetic code with very complicated mechanisms. Its synthesis does take place in the Cys-tRNAcys molecule, as the selenocysteine is also very toxic to cells. MSE enriched yeast cultures have been used to provide Se in Human and animal diets (G. N. Schrauzer, The journal of nutrition 130 (2000) 1653-1656). But I do not know if the toxicity of purified MSE has been duly investigated in humans, though it has been studied in Yeast and is toxic (Plateau et al., Scientific Reports 7 (March 17, 2017): 44761.)

Proteins labelled with selenomethionine are not so easy to produce and usually expensive, but methionine is usually not a very reactive residue and should have only a small effect on the structure and function of an enzyme or a protein, except in very specific cases. Because there are now a good deal of proteins  solved with and without MSE, you should be able to investigate the impact of introducing the MSE on the protein structure, if you do a careful search in the PDB and compare a number of examples. I know of some cases where the spectroscopic and binding properties are modified by MSE (Yuan, A. M. Weljie, H. J. Vogel,  Biochemistry 37 (1998) 3187-95.), but it is certainly an interesting subject to study.

Best wishes

Rogelio