I have not worked with Apollo in awhile but you may have to manually go into the database to insert the new reference sequences. the code lines for this are here on github
you can see here, there is a line that says "if the sequences for the organism are not loaded yet, then run the loadRefSeqs code"
ed
and here is the loadRefSeqs code
so there is a database entry for the chromosomes/refseqs that gets initialized at startup, but if you later update the on disk file, it wouldn't get mirrored by the database, and in that case, you can manually edit the database or just restart from some amount of scratch (create a new organism directory may work, or total scratch if it gets messy)
hope that helps, I don't have the exact SQL commands you can use but I hope that 'codewise reverse engineering' helps clarify the behavior you are seeing
-Colin