If you start with an image pdf (scan) there's nothing to do unless you ocr it.
If you start with a text pdf on which Adobe search works, it all depends how you convert. You can try and extract text with as much formatting as possible using this or that program - for me the best overall was just text extract and then clean up with E-book Tidy - and then convert the text/rtf/html whatever file to lrf with Calibre and then search works or you can "print to lrf" using this or that program and then you get image lrf's on which search does not work.
I used to do lots of pdf extract for the Nokia 770 and PRS 500 since both cannot handle pdf's well, but for the 700 I just dump the pdf's, maybe changing the metadata first if not correct and it's so much easier.
I used to run from pdf's as much as I could, but now I started liking them even when other formats are available.