[PyMOL] how to get a correct pdb output for atom names ?

[Pascal A. <p.a...@ib...>]

Dear pymol users,

I came across a problem with pymol writing out pdbfiles 
containing MG atoms like in 1EHZ.

I used the following options since a long time:

		set retain_order, on
		set pdb_retain_ids, on
		set pdb_use_ter_records, off
		set pdb_no_end_record, on
		set pdb_conect_all,off
		set pdb_unbond_cations, on
		set pdb_literal_names, on
		set pdb_truncate_residue_name, on
		set pdb_reformat_names_mode, 3

Yet, these options transform the following lines
extracted from 1EHZ

HETATM 1654 MG    MG A 590      70.566  35.530   1.665  1.00 65.44          MG  
HETATM 1655 MN    MN A 530      80.714  57.068  31.271  1.00 34.23          MN  
HETATM 1656 MG    MG A 510      57.346  47.575  47.279  1.00 62.77          MG  
HETATM 1657 MN    MN A 520      49.923  44.427  50.131  1.00 89.20          MN  
HETATM 1658 MG    MG A 540      77.110  64.307  25.357  1.00 40.08          MG  
HETATM 1659 MG    MG A 560      62.649  46.629  27.595  1.00 47.02          MG  
HETATM 1660 MG    MG A 570      73.331  43.321  11.207  1.00 50.39          MG  
HETATM 1661 MG    MG A 580      69.222  44.815  33.339  1.00 61.74          MG  
HETATM 1662 MN    MN A 550      72.301  48.513  33.894  1.00 56.51          MN  

into

HETATM 1654  MN   MN A 530      80.714  57.068  31.271  1.00 34.23          MN  
HETATM 1655  MG   MG A 510      57.346  47.575  47.279  1.00 62.77          MG  
HETATM 1656  MN   MN A 520      49.923  44.427  50.131  1.00 89.20          MN  
HETATM 1657  MG   MG A 540      77.110  64.307  25.357  1.00 40.08          MG  
HETATM 1658  MG   MG A 560      62.649  46.629  27.595  1.00 47.02          MG  
HETATM 1659  MG   MG A 570      73.331  43.321  11.207  1.00 50.39          MG  
HETATM 1660  MG   MG A 580      69.222  44.815  33.339  1.00 61.74          MG  
HETATM 1661  MN   MN A 550      72.301  48.513  33.894  1.00 56.51          MN  

Thus, MG^^is transformed into ^MG^ which is a problem and would for
example prevent to distinguish 
CA^^ as a calcium atom
from ^CA^ which would be a C-alpha based solely on atom names.

Are my settings OK ?
Did I do something wrong ?
Should I change them to get an output respecting pdb standards ?

Thanks a lot for helping,

Pascal

-----
Pascal Auffinger
IBMC/CNRS
15 rue René Descartes
67084 Strasbourg Cedex

p.a...@ib...
-----