To your first point, sure you can get a yeast to produce tons of your protein, can you control for EXACTLY how much protein your yeast produces? Probably not? With automated programmable syntehsise you can control exactly how much protein is in your solution and I imagine immune system response is a function to how much dose you give. Let's say you have a yeast cell produce your protein. Is it easy to count 10,000 proteins out of what it produces? What if you need 100,000 proteins? We're talking about personalized medicine. Everyone's immune system is different, with an automated synthesis you could theoretically fine tune the concentrations you need to produce a response correct?
Second point. Good question, I was under the impression that the peptide sequence determines structure and folding. Could you create some sort of reactor to synthesize the peptide strand and have its end product "spit out" into a soup filled with features native to a cell environment such as chaperone proetins? Couldn't we just transfect a yeast with a chaperone protein DNA to produce the chaperones we needed?
I'm sure this sounds ridiculous, but I bet the atom bomb did too in the 1890s.