Reason 2.5 does not have a sidechain on the compressor. You can't easily do what you're asking.
On the other hand, Reason 2.5 does have a vocoder with 16 CV outputs for the 16 frequency bands. Theoretically if you merge these 16 CV outputs together (using five Spider CV's), you obtain a CV which corresponds to the overall volume of the input. If you invert that CV (again using a Spider CV), then merge it with a constant high CV (say, from a Matrix curve with one step), you get a CV which is high when there's no signal, and low when there is signal. Connect that CV to the "Level CV In" of a mixer, and you get a channel that is attenuated when the input to the vocoder is loud. So if you were to connect your bass to the mixer channel, and your bass drum to the vocoder, when the drum hits, the bass will be automatically reduced in volume. This might work well for you.