I have determined that Ulrich was actually correct... I was generating the database dump of my production server via ldapsearch and then copying that ldif file to the development server and using slapadd to install it.
Since the entryDN and subschemaSubentry are not in the ldif created from a slapcat of the production server, I suspect if I were to root around in the source code that the ldapsearch process adds them to the returned ldif.
Therefore, I had two options, I could filter those two attributes out of what ldapsearch returns or I could stop using ldapsearch and use slapcat. Since I assume there may well be other attributes that ldapsearch could add which would cause me headaches in the future, I have opted to use slapcat (via ssh to the server) which is the proper mirror of the slapadd (via ssh to the server) that I use to install the production database on the development server.
Sorry to (once again) ask for help with something I should be smart enough to figure out. I could have figured this out quickly had I just done a comparison of what ldapsearch generated against what slapcat generated from the production server.