Well, the lux operon seems to be very difficult to PCR. We tried it several times with brand new phusion (proof-reading) polymerase. No amplification.
Then we tried with phire polymerase, it works after some modification.
Phire is not proof-reading, yet it makes only half the errors as Taq. So, to see if there are no point mutations (that destroy the function of a lux gene) we'd need a promoter that reads off the lux genes in E coli. Designing new pirmers again... I'm glad these finally worked, so a plasmid with T5 or someting (not induced or induced) would seem awesome..