I don't see how a shear bolt like this gets "stretched and relaxed" during the use - we're not talking thousands of RPM here and the distance from the center of rotation is very small; the bolt head may have 15g outward on it, which isn't actually that much for a 1/2" steel bolt (consider, when you hold the threads of the bolt with the head down, it's got 1g on it; imagine the head of the bolt now weighing 16 times as much - what's that, a few ounces? - it's not stretching much from this), and at a constant 540RPM that 15g isn't changing until you disengage the PTO or spin the clutch.
Note that if you torque a bolt down, it gets narrower as it stretches, so a snug bolt gets slightly less snug and is more likely to get pounded. What is the point of torquing a shear bolt? We're looking for shear, not tensile, and I think torquing it is going to reduce the shear strength.
Shear bolt should have a nut on it to keep it in, but it doesn't take much. I use a nyloc and have a few extra on hand, though not as many extra as bolts, since I usually can sometimes find the old one.
IMO the bolts are getting hammered because they're too small at this point for the hole, for whatever reason. Increase the bolt size (like go the next slightly larger metric) and ream the hole out to so that whatever goes in is tight; make an effort to have the shoulder go all the way through; if the shoulder extends beyond, don't care, just snug the nut on it. *if* the bolt moves during the spin (I strongly doubt it will if it's not way too small) it's going to either have the cap against or the nut against.