You could consider replacing it with... nothing. The people who own the patent on WPRE published some "positive data". Most of the people who use it have never tested if it makes a difference. I've heard rumors that those few people who did a side-by-side comparison, with all else being equal, do not necessarily see a large, or any improvement of the expression level. It may depend on the sequence context, or the cell type, or something else.
There does not seem to be any systematic study of this published. So before you replace it with something having the "same function", you should make sure that there is in fact any desirable function at all present in your context. I'm afraid, if you want to base your decision on anything other than hearsay, you will have to do the experiment and see what happens in your particular context -- WPRE vs. not X expression level vs. virus titer.
I cannot replace either WORE in the construct with poly A signal sequence, right? Coz polyA in transfer plasmid must be placed in downstream of 3’ LTR?
Not sure what vector you are working with, but normally the 3'-LTR would contain a polyadenylation signal. So you would not need to place another one. WPRE is not a polyadenylation signal itself, but is supposed to stabilize mRNA by a different (unknown?) mechanism.
But, now, for gene therapy, almost nobody use complete/natural 3' LTR. People use SIN-LTR, isn't it? I saw lots of vector put polyA downstream of 3 'LTR.
John Schloendorn I agree. After comparing my expression plasmid with and without WPRE, I found that WPRE did not enhance either viral titers or expression levels.