The CUPS web interface would allow you to direct the print job to another printer.
However it's possible to capture a PDF file from the CUPS queue:
/private/var/spool/cups
total 472
drwx--x--- 24 root _lp 816 Mar 1 14:30 ./
-rw------- 1 root _lp 4403 Mar 1 14:30 c00025
drwxrwxr-x 13 root _lp 442 Mar 1 14:30 cache/
-rw-r----- 1 root _lp 86110 Mar 1 14:28 d00025-001
drwxrwx--T 5 root _lp 170 Mar 1 14:28 tmp/
-rw------- 1 root _lp 4391 Feb 25 12:13 c00024
root@Macbook: file d00025-001
d00025-001: PDF document, version 1.3
You can copy (in this example) d00025-001 to somewhere and rename it with a pdf extension (n this case), after fixing permissions, ownership and group you can open it.
The matching file starting with c is the cover sheet. You should be able to recover the job in the the printers native language from the tmp directory:
/var/spool/cups/tmp
total 288
drwxrwx--T 5 root _lp 170 Mar 1 14:34 ./
drwx--x--- 28 root _lp 952 Mar 1 14:50 ../
-rw-r----- 1 root _lp 1234 Mar 1 14:28 17ed254faebef
-rw-r----- 1 root _lp 1193 Mar 1 14:34 17ed254fece74
-rw------- 1 root _lp 136512 Mar 1 14:28 17f025500d2e2
david_koontz@Macbook: file *2
17f025500d2e2: PostScript document text conforming DSC level 3.0, Level 2
This all done with an aging HPLJ-5000N LaserJet specified as a PostScript printer, the print job one page out of an ODT file (OpenOffice).
The original (in this case ODT) file isn't kept in the print system, which would force you to format it again.