People that know what they're doing (and that can't guarantee a very specific structure to their HTML) don't use regexen to parse HTML. Use HTML::TreeBuilder or HTML::TokeParser::Simple.

In the former module you'd use look_down to find <li> elements with <a> elements of the desired styles and then pull out what you want.

With the later you'd be interested in <li>. Once you see one you look for the next two anchors and again extract what you're interested in.