As usual the N-version of the mosfet is a die shrink of the non-N version.
This is confirmed by looking at the gate charge numbers and the SOA-graph.
Depending on your application this could be a dealbreaker to go with the N-version.
Smaller die means lower pulse power handling capability (and pulse power dissipation is what you get a lot of in a normal class AB amp).
The A version looks like a die shrink as well (lower gate charge numbers) but the SOA-graph indicates it is a very good match to the original 140.
I would say this is a near spot on candidate.
As a sidenote Fairchild even characterized the SOA of this device at DC. Something IR never bothered to do with the original (one could suspect IR have something to hide there - pisspour performance @ DC maybe).