Trapping novel promoters and genes for stress and disease tolerance, or quality enhancement is a routine in modern era. Which one of these, genes or promoters is better option for genetic engineering?
But are important, but gene is IMHO superior, because you need to choose a particular gene which is responsible for requested activity. You cannot simply chance one gene for something completely different. On the other hand, the promoter doesn't have to be so perfect, often you can use constitutive expression or the promoter can be leaky and it doesn't really matter.
And you're forgeting other important parts as targeting peptides/signals etc.
But of course you need as precise as possible cooperation of all parts.