I would investigate fitting a small pump into a pressure vessel and connecting the fluid lines through the enclosure cap with NPT or NPTF tube fittings. You might want to try using soft tubing that gets pressed onto “hose barbs”. You would have to see if that type of connection will be suitable for your pressure. Re-engineering the T-200 into a pump would take some considerable knowledge of prop/pump design, which might not be worth your time.
I wouldn’t worry about the other things getting crushed from the atmospheric pressure. However, you will need a limit switch for the balloon to press against once it is full. The switch will prevent the balloon from overinflating. You could always house your electronics inside another small enclosure inside that enclosure as well.
You should easily be able to fit a suitable pump into one of the BlueRobotics enclosures. At 40 meters, you’re only going to be working with about 60 psi of pressure, so it shouldn’t be too difficult to source a 12 or 24 VDC pump. In the setup that you described, if you use anything other than a peristaltic pump, you will also need to run some electronic valve between the flexible ballast and the pump so that water doesn’t get immediately pushed back out once you shut off the pump. You should try to have the pump below surface level so it doesn’t suck in air, which will decrease efficiency.
It would be best if you considered implementing small zinc screens into the fluid path before the inlet to serve as a sacrificial anode to help mitigate corrosion if it’s going to be in seawater. Furthermore, you will need a filter to prevent aggregate from getting sucked into your ballast and damaging it.
You could consider using a hard-plastic ballast tank (like a plastic bottle) housed inside a larger pressure vessel. If you do this, then you will need a higher-pressure pump. Air will get compressed inside the ballast as it fills and has nowhere to escape and little space to compress, which means you can only flood the tank to about 3/4 full. You can circumvent this if you implement a vent valve at the top, but that’s another valve (another failure point + incurred cost). I think your setup is the best value for your operating depth.
So, to reiterate, try to find a peristaltic pump that will fit inside an enclosure, and make sure to implement a shut-off switch, so the pump knows when to stop. Best of luck!