There's a simple solution to all the problems with reproductase: remove reproductase entirely.
I'm pretty sure that's already the path we decided on -- now that we've figured out how to model the-stuff-a-cell-is-made-of (with 'locked' compound pools), we no longer need a magical compound to represent the-stuff-we-need-to-make-a-new-cell, because those are one and the same. If an average, healthy cell has a certain amount of each compound, then two of them would have twice as much. Therefore, for a cell to be able to split into two average, healthy cells, it must have twice as much, and then a bit extra to cover the energy needed to reproduce. Note my wording -- cancer cells, for example, generally reproduce too early.
How strict we should be about having double of everything is another question entirely (for example, if we have a glut of food protein, can we split earlier even if we don't have large enough glycogen reserves?), but it provides interesting tradeoffs for gameplay.
The other question, of what we do with changing cell mass through evolution, is simple: The population still has the same amount of each compound, so if each organism is bigger, there are now fewer of them.
I'll fork off a discussion before this gets even further off-track.