Evolutionists Don’t Know What They’re Doing
And yet it is difficult to get two proteins to interact in a meaningful way. Such interactions must not be too strong or too weak. Imagine that you had two proteins that you needed to bind meaningfully to each other. If you randomly selected the amino acids at the binding patch on the surface of one of the two proteins, then meaningful binding would be unlikely. In fact, you would have to repeat the experiment millions of times before you could expect to get a good result.
But evolution does not have such resources. It cannot conduct millions of evolutionary experiments in order to luckily find amino acid sequences on protein surfaces that are required for important biological functions. And even if it could, that would only be the first step, because molecular machines are often comprised of multiple proteins, interacting with each other at multiple sites. So evolution would have to luckily find several sequences, in multiple proteins, and get them to arise in similar time frames, so the molecular machine would function.
But that is not all, for molecular machines often work in conjunction with other molecular machines. Having a molecular machine without its neighbors would often not help much.
And yet even with all this there remain more problems. For instance, most proteins are not highly modifiable. You can’t just randomly go about swapping in different amino acids. Protein function typically degrades rapidly with amino acid substitutions. So it is challenging for very much interaction site experimentation to take place in the first place. And of course another problem is that it is astronomically difficult for evolution to evolve a single protein to begin with, let alone meaningful interaction sites.
Simply put, from a scientific perspective protein-protein interaction is another problem for evolution.