In RNA secondary structure prediction servers like Mfold or Sfold there is an option to choose whether you want to fold it as a prokaryote or eukaryote. If one chooses to use a prokaryote option, then Sfold algorithm will use a maximum distance (of 50 bases) between paired bases. But is it really true, that for example mRNA of some prokaryote in reality there is some kind of restriction relating to maximum distance between paired bases? Is it because of circular nature of DNA the folding of mRNA will be more local, i.e. paires do not form between very distance bases?
From my experience the fold is relatively local and allowing for distal base-pairing won't change the predicted structure a lot. I think this has to do more with the capabilities of the algorithm than the biology. There can be such interactions but these interactions will usually involve pseudoknots which basic RNA structure prediction algorithms won't deal with anyway.