Joe, Certainly, mitochondria play a part, but they do so through their own DNA, which mutates in the same ways as the DNA of the nucleus, or for that matter, the DNA of the chloroplast (which is also passed through the female parent). But neither of those organelles moves from rootstock to scion, or back again. And I can see how a more- or less-active mitochondrion might affect some behavior of the plant, but it would be strictly due to availability of energy; I've never heard of mitochondria producing or modifying other substances than energy-source materials. That seems to be their unique function.
And in a very real sense, introduction of a viroid does change the "genetics" of a plant -- the RNA contains nucleotide-based code which is read and the message converted to protein, in the same manner as the m-RNA coming out of the nucleus. Certainly it is extra-nuclear, but it is still genetic code. And as long as neither scion nor stock of a grafted plant is infected with any viroids or viruses, there should be no movement of any coding molecules across the graft union.
I suppose it's theoretically possible that some hormone from a rootstock would alter the rate (but not the direction) of mutation in a scion, but I think the evidence for that is absent, and circumstantial evidence against it is overwhelming. E.g., Valencia is notorious for mutating frequently. I've not seen numbers, but I have no reason to think that those mutations occur at different rates, or in different directions, based on the rootstock, and certainly Val is grown on lots of different rootstocks. The common mutations (mohawk fruit, variegated leaves, a sectorial chimera for rust mite susceptibility) seem to be pretty evenly distributed in frequency, throughout the industry, regardless of rootstock.
Of course this whole question is much larger than just in citrus. Apples, peaches, pears, plums, cherries, magnolias, roses, olives, grapes, camellias, gardenias, peonies, and countlessly many other crops are also grafted to various rootstocks. In each case, the stock has some effects on the scion. But scions taken from those plants and placed on another rootstock immediately pick up the acquired traits of the new rootstock, and immediately lose the acquired traits of the old rootstock. I've never heard of a single exception to that in all of horticulture. It might be very useful and certainly very interesting if it did happen, but as far as I can tell, it doesn't happen, ever.
I would agree that you can never guarantee 100.00 percent that something does not happen. But if there is no observable evidence at all that it does happen, I tend to think it will not.