Functionally, it ties together the selbri (the "verb", if you will, of the sentence) with the sumti (argument) that follows it:
lo gerku [ku] cu klama lo zarci [ku] -- The dog goes to the store.
lo gerku cu klama be lo zarci -- The dog is a goer-to-the-store
These two sentences basically mean the same thing (there is a slight difference in meaning which we needn't go into right now unless you are interested). So why the be? Well, it becomes necessary if you want to talk about a compound subject. What if I wanted to say "The goer to the store is fast?" I can't say *"lo klama [ku] lo zarci [ku] cu sutra" because that means "the goer is fast in performing the store (whatever THAT means!?=the x2 of sutra). So in order for the location, the store, to be linked together with the action of going, I need "be":
lo klama be lo zarci [ku] [be'o] [ku] cu sutra (the goer-to-the-store is fast)
--gejyspa