Xiang-Jun,
Thats good news.
Basically, pymol uses the segment identifier code defined here:
http://deposit.rcsb.org/adit/docs/pdb_atom_format.htmlat the "Record: ATOM" line:
73 - 76        LString(4)      Segment identifier, left-justified.
See this example (A1  )       
ATOM    145  N   VAL A  25      32.433  16.336  57.540  1.00 11.92      A1   N
or (AZ00)
HETATM  489  O   HOH B  79      -9.310  -1.446  19.391  1.00 71.79           O  
HETATM  411  N1  SPM A  21     -13.297  -8.783  22.839  1.00 40.13      AZ00 N 
(sorry for the fact that I am unable to write these lines with an non-proportional font)
pymol uses these specifications.
Yet, this is an old pdb definition as you can judge from the most recent one
http://www.wwpdb.org/documentation/format33/sect9.html#ATOMthat does not mention a segment identifier.
Here is what I found in pymol documentation
---
“segment” selects atoms based on their assigned chain, which is a four-letter code.
A PyMOL Segment is a set of atoms within a single molecular object that share all property identifiers except atom names, residue and chain property identifiers.
Segments are most useful in cases where multiple copies of an identical set of chains must exist inside of a single molecular object. They also come in handy for complex proteins which have more chains than can represented with a single-character alphanumeric identifier. 
---
Would be great if you could make this work,
Thanks for considering it,
Pascal