diff --git a/disco/util/sanitize.py b/disco/util/sanitize.py
index e5ecf74..23a11e4 100644
--- a/disco/util/sanitize.py
+++ b/disco/util/sanitize.py
@@ -9,15 +9,15 @@ MODIFIER_GRAVE_ACCENT = u'\u02CB'
 
 # Regex which matches all possible mention combinations, this may be over-zealous
 #  but its better safe than sorry.
-MENTION_RE = re.compile('<([@|#][!|&]?[0-9]+>|@everyone|@here)')
+MENTION_RE = re.compile('<?([@|#][!|&]?[0-9]+|@everyone|@here)>?')
 
 
 def _re_sub_mention(mention):
     mention = mention.group(1)
     if '#' in mention:
-        return ZERO_WIDTH_SPACE.join(mention.split('#', 1))
+        return (u'#' + ZERO_WIDTH_SPACE).join(mention.split('#', 1))
     elif '@' in mention:
-        return ZERO_WIDTH_SPACE.join(mention.split('@', 1))
+        return (u'@' + ZERO_WIDTH_SPACE).join(mention.split('@', 1))
     else:
         return mention