He said it can be shown that genes demonstrably do arise(which it can), not that we can travel back in time to see how some particular extant gene evolved in the ancient past.
Nevertheless, there is actually good evidence for how the alpha and beta chains of ATP synthase evolved. They evolved and diverged by gene-duplication of a common ancestral protein functionally and structurally similar to both. In some related structures such as hexameric helicases, it is the same protein oligomerized six times, instead of two similar duplicates(alpha and beta) three times.
There’s a nice phylogenetic tree of the related proteins in both F-type and V-type ATP-synthases, and other P-loop NTPase proteins, in this paper:
Mulkidjanian AY, Makarova KS, Galperin MY, Koonin EV. Inventing the dynamo machine: the evolution of the F-type and V-type ATPases. Nat Rev Microbiol. 2007 Nov;5(11):892-9. Review. DOI:10.1038/nrmicro1767
Click to enlarge.
Key question: Why should this evidence exist? If these proteins did not evolve from common ancestors with different related functions, why can we detect high levels of nesting hierarchical structure in their shared similar sequences? Why do they all form catalytic hexamers that implement their functions by ATP binding and hydrolysis?
And hey, I have another post written 21 days ago in direct response to you on the inferred origin of P-loop NTPases. On a related subject, one of those papers on de novo evolved proteins I’ve shown here before many times describes a randomly generated peptide that gives antibiotic resistance to bacteria. Turns out the peptide forms a membrane proton channel.