I've noticed that discussions of patent law really wilt my traffic something fierce, so I thought I'd go ahead and get another one out of my system now. Perhaps the effect isn't additive. (It'll serve me right if turns out to be nonlinear the other way).
One of the things that hits you when you start worrying about patenting chemical compounds is that there sure have been a lot of them patented. The number of compounds exemplified is pretty large all by itself, but it's just a roundoff error compared to the number that have been claimed. I've seen patent specifications that I swear would run well into the millions if every claimed variation were rung.
They aren't, of course. There's no way that they could be. The extra space is just breathing room, to try to keep the competition from coming too close to the good stuff. It's also to protect some of it for you if you happen to come across something else that's good after you file, but that's a two-edged sword - if you find it too much later, you've got a patent with years of its lifetime already expired, for one thing.
It's important to remember that these drug patents aren't just randomly distributed around the world of chemical structures, either. There are plenty of things that no one is going to be interested in patenting, because they're too reactive or toxic. And there are other structures that just lend themselves to drug discovery: piperidines, piperazines, biphenyls, imidazoles, and tricyclics are just a few that come immediately to mind. The patent space around those things has been trampled, plowed, clear-cut and strip-mined.
These things are all prior art for new patent applications to worry about. Exemplified, reduced-to-practice compounds are bulletproof prior art: if it's been described in the literature, you can't own it. You can claim a new way to make it, or a new use for it, but you can't claim the chemical matter. Section 103 of the patent code says that an invention has to be nonobvious, and you don't get much more obvious than something that's already there. But what about all those claims, those huge, inflated claim structures that reach well into the Kuiper Belt? Are all those off limits?
We'd be in big trouble if that were the case. After all, the problem gets worse each year, nonlinearly worse as the pace of research (and the pace of patenting) picks up. But there's a way to sneak into those putatively fenced-off reserves. If you turn to the MPEP, the patent examiner's manual, you find that "the fact that a claimed species or subgenus is encompassed by a prior art claim is not sufficient by itself to establish a prima facie case of obviousness." Good news, indeed.
For a patent examiner to reject a claim on the grounds of obviousness, the manual says that "it is essential that Office personnel find some motivation or suggestion to make the claimed invention in light of the prior art teachings. . .regardless of the type of disclosure, the prior art must provide some motivation to one of ordinary skill in the art to make the claimed invention. . ." And that's the key. Those massive edifices of claim language eventually have to get down to earth. As the claims roll on, page after page, they start narrowing down. A preferred embodiment of the invention is. . .a still more preferred embodiment. . .a particularly preferred embodiment. . .eventually you get down to what they really, really want to protect.
And legally, what they're doing is teaching you the invention. The claims teach toward the true invention, and the rest of the patent is supposed to teach you how to carry it out (including the "best mode" requirement I mentioned yesterday.) If you're claiming something that's way out in the fringes of the first forest of claims, then the rest of the patent is clearly going to teach away from it. And that means that you can get it for yourself.
There are complications. This is patent law; of course there are complications. It'll hurt your case if you're going for the same use as the original compounds. Your best chance there is to show that your compounds perform in some unexpected way, which the original patent claims clearly didn't anticipate. A less common way to break out of this situation is to show that the original patent's methods wouldn't even be able to produce some of the distant structures - if they don't teach how to make the structures, they can't have 'em.
So that's how it's done. That's not to say that patenting isn't getting harder all the time, because it is (after all, there are all those truly exemplified compounds piling up.) But at least it isn't getting impossible. Nothing a lot of time, effort, money, and legal resources can't attack, anyway.