Is there not any precedent as to what happens in cases like these? I feel like I've been hearing about patents on gene mapping, gene editing, even genes themselves, for a while now.
I'm not sure. The sequence of the original Cas9 gene found in S. pyogenes cannot be patented, as it is a naturally occurring DNA sequence. However, in this case, I believe the many modifications people have made to Cas9, as well as methods that use it to edit genes, constitute as patentable inventions.
As far as "who gets the patent", by law it's whoever was the "first to invent" this use of Cas9, but I'm not sure the exact details of how that will be argued in this case. Commenting to your friend that X could be used to do Y might be considered the first novel description of a process, but I don't think that's enough to hold water. But IANAPL.
And an aside: how ground breaking is CRISPR? I've read a little bit about it, and listened to a Radiolab episode where a science writer answers questions that the hosts put to him, but I don't quite understand its scope. CRISPR is a gene editor. Had there been no editor before?
At its heart, Cas9 it a DNA nuclease. It cuts DNA. Plenty of enzymes cut DNA, but what makes it unique is that where it cuts is controllable. Again, other the cut sites of other enzymes are controllable, but what makes Cas9 particularly unique is that that control is specified by an RNA sequence.
TALENs and ZFNs are two other controllable-DNA-cutting-enzymes, but they both are targeted by the sequence of the protein. It's hard to predict in advance how different protein sequences will interact with DNA, so for ever new site you want to target, you must mutant, screen, and evolve a new protein. That can take months to years of work and is a barrier to use by most researchers.
In contrast, if you want to target a new sequence of DNA with Cas9/CRISPR, you simply find a sequence that is complementary to that sequence. In practice, there are still some caveats, but as a starting point, that gives you a lot more speed and power to new sequences or large sets of sequences.
The "editing" part of the equation comes from the fact that when DNA is cut, the cell has a number of repair pathways to reseal that break.
The one of the major ones is Non-Homologous End Joining (NHEJ), which just glues either end back together (sometimes inserting an extra bit of DNA in between). The other is Homology Directed Repair (HDR), which looks for DNA that is similar to the sequences around the cut site and copies that sequence over while sealing the break.
So in that latter case of HDR, if you also provide a short, similar sequence with the modification you want, the cell can repair that site, copying the new DNA, achieving that final goal of an "edit". It's not perfect, often the efficiency is rather low, but it's still proven to be a robust enough tool that people have rolled with it.