It’s good that you conceptualized the “oxygen tolerance” qualities of bioluminescence. That plays perfectly into the fact that bioluminescence is much more common in prokaryotes as opposed to eukaryotes. Considering that most eukaryotes will naturally slide towards aerobic respiration for the higher energy yields, it is likely that bioluminescence would naturally become less viable.
The most annoying part about this anaerobic conundrum is that - in my opinion - there isn’t a very easy fix without somehow including new environmental factors. I guess the good news with sulfate and nitrate though is that they have immense relevance for life and the future stages as well, as they act as important limiting nutrients in many cases. Plants compete immensely for nitrate, relying on symbiotic relationships with nitrifying bacteria, and sulfate can influence growth rates and microecosystems as well. And as said above, implementing these metabolisms can result in interesting nutrient cycling simulations in Thrive; our ammonia-producing parts indirectly result in the creation of nitrate and v.v, while our chemosynthesizing proteins result in the creation of the sulfur needed to create sulfate.
It might be the case that a player can place down parts representing both “sides” of the nitrogen/sulfur cycle, but we can set up incentives so that players have to alter their body plan towards these different metabolisms, discouraging that. For example, one process could be anaerobic, while the other can be aerobic. Or we could have surface area effect each process differently.