No we don’t. Not a single case. Just flat out false.
No, just function. Functions are what matters here, not what structure implements it.
Completely made up. Please show your work. How do you go from “some function exists at a rate of 1 in every 10^15 sequences of some length L” to “therefore we probably need 10^30 mutations to evolve 2 protein domains”. It just doesn’t follow in any way that I can see.
Many more what? What proteins are those, and what is it they require, and what do you know about how this had to evolve?
Yes, if you’re going to claim evolution couldn’t produce X, you need to show the couldn’t. If you can’t show the couldn’t, you haven’t ruled out the could. If you haven’t ruled out the could, you can’t reject evolution. Phylogenies imply the could, all by themselves. And of course we have now over half a century of biochemstry that implies that, in general, yes, new proteins with novel functions can evolve by numerous different methods.
We have examples of overlapping functions in sequence space. Entire superfamilies of proteins are known to be capable of performing literally tens of thousands of distinct functions.
We have examples of novel functional proteins being generated from random sequences, of non-coding DNA evolving into de novo proteins, and of the emergence of genes with novel functions from random combinations of pieces of existing genes.
I don’t think that. What I do think is that you should try to relax your absolutist, dichotomous thinking. This thing with “all” vs “none” is a big part of your problem. Another is your insistence on some particular transition you can think of.
Maybe it can’t, but I’m not aware of those transitions being required to have occurred in the inferred history of life. Is cytochrome c required to have evolved from a globin, or keratin from histones? No. So whether those particular transitions are possible is simply flat out irrelevant.