It looks like what you really need to do is parse some HTML and manipulate some of the elements of the parsed text. However it's not clear just what you want to be able to manipulate, particularly as the sample you've given isn't a well formed fragment of HTML - you seem to have mismatched quote characters. It doesn't help that it's not clear just which TXT instances in the sample ought to be replaced.

For modest sized documents HTML::TreeParser may be the most appropriate tool for the job.