Probably little reason to shoot in 4K. What are you doing for audio? That's really important.
Regarding recorder, one factor is how many channels do you need? A two channel one will always be useful, but might you ever need more, as for example when recording boom plus several wireless mics? But in general, I would spend a lot more for mics than for the recorder.
Or you could go a little higher on the Canon DSLR range.
I actually think the FS7 would be fine for live events, maybe not quite as good as an eng camera, but quite adequate. Go to the forum at DVXUSER.com, find Sony, then the FS7 section. There is a lot of discussion of zoom lenses there. (It's an endless topic of discussion).
This question only makes sense if you specify the sensor size. If you are talking super 35 sensor, I like the Rokinon kit consisting of 24, 35, 50, 85mm. I also use a metabones speedbooster, which gives you some more options (These are all full frame lenses). I would consider the 50 optional, partially because of the speedbooster. The only time I could think that you would need longer lenses is if you are forced to shoot at a distance and need to get tighter on your subjects. You should be able to get shots like extreme closeups with the 85.
My only additional thought is: attach a mirror to the inside of the windshield. Frame the shot so you don't see the edges of the mirror. On different shots, move the mirror from side to side to get your two actors. But as others have said, if actors really get into a part, they are not doing good driving. It's like texting while driving, but times 10.