They can't have it both ways. If being a journal parallel to and under the same org as the journal Nature is supposed to make Nature Communications look better, then low-quality content in Nature Communications necessarily erodes the Nature brand of which they are a very deliberate part.
Again, Andrew Gelman put it succinctly:
> The paper is published at nature.com. Nature Communications is an extension of the Nature brand. Reputational inference goes both ways. By giving this journal the Nature name, they’re leveraging the Nature brand. The converse is that if the new journal published a paper with serious flaws, the Nature brand loses.