It's actually a pretty complex problem. The simple answer is that since PBC designed the v-bearing and VB2 bushing together, the system should be capable of 615 lbs for each bearing, so assuming a small preload, 1000 lb for the y-car (not accounting for cutting and dynamic loads).
That's the load for reliability. There may be substantial flex before that. There are 3 stages of flex (assuming the bolt is tightened to a moderate tension):
1. Removal of slop in the system. A preload should do this.
2. A ~linear portion of flex, due to compression in the bearing, bending of the bushing, and bending of the plate. The bolt actually doesn't change much: its purpose is just to keep the bearing and bushing firmly planted against plate. All the bending load is taken through the bushing--not the bolt. Normal operation should stay in this stage.
3. An increased rate of flex, when the load is high enough that the bushing starts to separate from the plate on the bushing's bottom edge. (Now the bolt would increase in tension.) If PBC engineered and tested the system thoroughly, 615 lbs won't be enough to get to this stage, but we need the bolt tension required for that.
From your system pic, the bushing OD is a little larger than the lateral distance from the plate to the center of the V (the overhanging moment arm, I'll call it w). This indicates PBC thought about the problem. The actual solution is really complex (contact mechanics). But we can easily approximate complete separation along the bushing's bottom face:
Load * w = BoltTension * OD/2.
BoltTension = Load * 2 * w / OD = ~ 615 * 2 * 0.85 = ~1000 lb
I'd guess 2000 lb would give minimal separation.
Bolt torque-tension is another complex relationship, but this pdf gives a rough guide:
http://www.fastenal.com/content/feds...0Gr8%20Gr9.pdf
So, a 1/4" grade 8 bolt is good for up to 2864 lbs tension, if it can be torqued appropriately. 2000 should be reasonable. All this indicates PBC (or BWC) did engineer the system. If currently using stainless bolts (similar to A307-A in the pdf, 859 lb max tension), upgrading to Grade 8 would be very smart.
Note that even though the system should stay in stage 2, the elastic flex could still be pretty large. To calc that stiffness (lb/in), we'd either need a complex FEA model or a simple experiment, like 6061 suggested.
Switching to 3/8" bolts has pros and cons. The pro of the PBC design is the bushing OD is > 3/8", so the bearing has a larger 'foundation'. Also the bushings have a precision fit to the bearing ID. But the bushing also pushes the system out for a larger "w"--more clearance but more bending. Using a 3/8" bolt with a small washer to space out the bearing just enough for clearance, w could be reduced by about half, for more stiffness. As Gerry suggested, a shoulder bolt with reamed holes would be best.