Someone will propose a more elegant coding solution, but in the meantime if you just want to do this 'quick and dirty', I'd suggest the Table Browser intersection function or Galaxy join function. I can give you a more complete step-by-step, but it might have to wait till I'm at my computer and not my iPad :D.
Edited to add:
can't promise much more than this at the moment, got busy and travel coming up :D... but here's the basic outline that I think will work. I'm sure there is a more elegant solution.
Use Table Browser
Choose UCSC Known Genes (use whatever gene position, IDs you have as position)
Create custom track
Choose "introns" or "upstream (5000kb)
create track
Create a second track
Choose conservation in comparative. There are many to choose from, look at the documentation to determine which (across vetebrates, mammals, etc)
Filter for a value you prefer for conservation (again, look at documentation)
create a second custom track.
In Table Browser choose first custom track,
intersect with second custom track
get sequence as output
That should get you a FASTA record of only those regions that have an overlap with a high conservation score you choose. I'm sure there must be a more elegant way (and to get very specific regions) than this, but that's the 'quick and dirty' method.
This is helpful, thanks again.