[u'ABC']
would be a one-element list of unicode strings. Beautiful Soup always produces Unicode. So you need to convert the list to a single unicode string, and then convert that to ASCII.
I don’t know exaxtly how you got the one-element lists; the contents member would be a list of strings and tags, which is apparently not what you have. Assuming that you really always get a list with a single element, and that your test is really only ASCII you would use this:
soup[0].encode("ascii")
However, please double-check that your data is really ASCII. This is pretty rare. Much more likely it’s latin-1 or utf-8.
soup[0].encode("latin-1")
soup[0].encode("utf-8")
Or you ask Beautiful Soup what the original encoding was and get it back in this encoding:
soup[0].encode(soup.originalEncoding)