Chris, your CLS won't be affected, it just cuts some light pollution out, so in theory you should need to take longer exposures to get sky limited subs.
It's important to understand that you don't gain (pun intended) anything shifting the signal further right because the first "object" adjacent to the left edge is the bias curve. If you are exposing your subs sufficiently, your background sky values need to be clear of the bias curve i.e. further to the right in the histogram...this is how you get the real signal (even if that is black background sky) to swamp the read noise, and thus your background sky shouldn't be plagued with any pattern noise or similar. With the Sony sensors, the read noise is the largest source of noise at reasonably exposure lengths, as the thermal noise is minuscule.
Incidentally, how I learned to do this was guided by docos by folk like Craig Stark on how to measure read noise. The BasicCCDParameters script in PI is really informative in this case, as you take an ensemble of bias, flats and a couple of darks (of differing lengths) to calculate the gain, read noise, dynamic range, etc. This is invaluable IMO for those of us with variable gain cameras, and it helped me understand how my CMOS cameras respond.
Step 1 in your process is key to understanding why linearly your CCD responds to photons. This is also important as during the course of imaging you might want to avoid straying into the non-linear response zone as it will clip your highlights. Fortunately for me, the 1600 chip is almost perfectly linear for exposures greater than 0.5s
