Ex parte Chuang (B.P.A.I. 2008)

    By Kevin E. Noonan --

The correspondence between a nucleic acid sequence and its encoded amino acid sequence is the basis for a great deal of how the U.S. Patent and Trademark Office examines biotechnology claims.  In many instances the Office treats these sequences differently.  For example, it is common practice for the Office to impose a restriction requirement between them, requiring applicants to elect to pursue examination of either nucleic acid claims or protein (amino acid sequence) claims in an application disclosing both.  Nucleic acid claims encompassing all of the nucleic acids encoding an amino acid sequence are typically granted, while amino acid sequence variants have become increasingly more difficult to obtain, based in part on the unpredictability of the effects of variants on biological activity.

In the recent non-precedential decision by the Board, Ex parte Chuang, the Office affirmed rejection under 35 U.S.C. § 102(b) of claims to an isolated protein over prior art disclosing a nucleic acid encoding the protein.  The rejected claims were directed to the isolated polypeptide:

Claim 4.  An isolated polypeptide, comprising the amino acid sequence of SEQ ID NO: 1.

Claim 20.  An isolated polypeptide, comprising a sequence that has at least 95% sequence identity to SEQ ID NO: 1, wherein the polypeptide reduces 15-keto prostaglandin but does not reduce leukotriene B4.

The Office cited three prior art references, directed towards large collections of isolated and sequenced mouse cDNA sequences.  These references disclosed the amino acid sequence encoded by the cDNA and identified it as a member of a zinc-containing alcohol dehydrogenase superfamily.  The references did not produce the encoded protein nor did they identify the biochemical activity of the protein encoded by the sequence.  Nevertheless, the Board affirmed the Examiner's rejection that disclosure in the prior art of a nucleic acid that encodes a protein isolated thereafter anticipates the protein, even if the cited art did not disclose the isolated protein.

The Board distinguished over an earlier Board decision, Lee v. Dryja, 79 U.S.P.Q.2d 1614 (B.P.A.I. 2005), holding that a "cDNA is at most an equivalent of its encoded protein but does not contain every element of the isolated protein encoded thereby."  Moreover, the Lee case further held:

[A]ssuming arguendo that it was well within ordinary skill in the art to prepare, isolate and purify the protein product of a given cDNA clone at the time the '163 application [the patent application at issue] was filed, the 4.7 kb cDNA described in the '163 specification would have made its encoded protein obvious at best.

The Board based its reasoning on In re Donohue, 766 F.2d 531 (Fed. Cir. 1985), which held that "[i]t is not, however, necessary that an invention disclosed in a publication shall have actually been made in order to satisfy the enablement requirement."  Accordingly, the Board reasoned that the references cited by the examiner provided an enabling disclosure of the protein encoded by the disclosed nucleic acids, and were anticipating.  The Board noted that the burden is on applicants to establish that the cited art is not enabling.

Specifically with regard to the Lee case, the Board first noted that Lee was an interference case, and the issue decided therein was whether Dryja had produced sufficient evidence that it had possession of a single enabled embodiment within the scope of the interference count.  This standard for interferences differs from the standard for anticipation, said the Board, although here that statement seems in error:  the Board would have been in the same position with regard to the cited references if each had only enabled the embodiment identified as SEQ ID NO: 1.  Second, the Board noted that Dryja did not disclose a complete coding sequence for the protein at issue in the interference count.  In that case, the presence of a translation stop codon in the cDNA prevented Dryja's cDNA from expressing full length protein.  None of the references cited against applicants in this case suffered from this infirmity, and it was undisputed that the cDNA could have (but had not) been used to produce the claimed protein.

While non-precedential, this case illustrates that the Office will now consider disclosed nucleic acids to be anticipating references to claims for isolated polypeptides.  This attitude has far-reaching implications for biotechnology, in view of the extensive public disclosure and annotation of open reading frames in sequences determined by the Human Genome Project.  It also contrasts starkly with the more stringent utility requirement for claiming such sequences, wherein applicants must show possession of not only the isolated nucleic acid and expressed protein thereof but also a specific, substantial, and credible utility for nucleic acids and proteins.  Thus, the Office imposes a standard of knowing the function of a protein encoded by a nucleic acid, but considers the mere disclosure of the nucleic acid to anticipate not only the nucleic acid but the protein encoded thereby.  Under these circumstances, it is likely that no later-filed claims to proteins encoded by open reading frames identified from the Human Genome Project will be patented, even if the protein has a unique, unappreciated biological activity.  This stance by the Office will put to the test the belief by many in the biotechnology community that patent protection is necessary to obtain investment in technologies derived from the explosion of genetic information that even now has not been completely elucidated, and that the absence of patent protection will retard if not preclude such investment.  It may be an expensive lesson to learn.

Ex parte Chuang (B.P.A.I. 2008)
Panel: Administrative Patent Judges Grimes, Lebovitz, and Fredman
Opinion by Administrative Patent Judge Fredman

Genetic code image: Madprime, Wikipedia Commons, GNU Free Documentation License

Comments

[Dan Feigelson]

Kevin, you wrote, "Thus, the Office imposes a standard of knowing the function of a protein encoded by a nucleic acid, but considers the mere disclosure of the nucleic acid to anticipate not only the nucleic acid but the protein encoded thereby. Under these circumstances, it is likely that no later-filed claims to proteins encoded by open reading frames identified from the Human Genome Project will be patented, even if the protein has a unique, unappreciated biological activity."

The observation of the first sentence, while often vexing, is unsurprising in view of the CAFC's adoption of the CCPA's reasoning that "enablement" in the 102 context is a lower standard than "enablement" under 112 (see e.g. Rasmusson v SKB, citing In re Hafner).

But the second sentence, and the implication thereof that you spell out, if it comes to pass, makes me shudder, because it leads to the wrong result from a policy perspective by being inconsistent with analogous case law in the small molecule area. With small molecules, it's been the case for years that you can claim in a later application a specific molecule that falls within an earlier-disclosed genus, without worrying about 102, as long as the later-claimed molecule wasn't specifically disclosed in the earlier disclosure. The CAFC recently reiterated this point (and how!) in the Sanofi v Apotex decision. Identifying a protein having a particular biological activity, when the DNA sequence encoding that protein is already known, is no different. Like broad molecular genus claims, there's a huge amount of information available now about the human genome, i.e. the DNA seqeuences, but that's the not the same as having made every protein that could be encoded by those sequences, just as the molecular genus claim doesn't anticipate the later species claim on the specific compound.

Let's hope that the CAFC knocks some sense into the BPAI when this one comes up on appeal.

[Kevin E. Noonan]

Dear Dan:

Yes, it is consistent with Hafner, which I recall took me a while to wrap my head around when I was first starting in this business.

I think the distinction (which does not work in our favor) over small molecule IP is that here, disclosure of an open reading frame with disclosure of the predicted amino acid sequence of the encoded protein, is enough. Taken from a PTO perspective, this makes sense, because technically the amino acid sequence IS in the prior art (albeit as merely a recitation of the sequence of an unknown protein). The first leap the BPAI took is in deciding that a claim to an isolated polypeptide did not require the art to disclose an isolated polypeptide, but rather an isolated cDNA, because that constituted an enabling disclosure (1) having the cDNA and 2) being able to use it to express the protein (good luck figuring out that the protein does).

The leap the BPAI did not take, but it is just around the corner, is saying that disclosure of a cDNA, without explicit disclosure of an encoded amino acid sequence, is anticipating. The basis for this would be that the skilled worker could "find" an open reading frame and then express the protein. This is more analogous to your small molecule scenario, and here I think there would be some occasion to push back.

Thanks for the comment.

[EG]

"The leap the BPAI did not take, but it is just around the corner, is saying that disclosure of a cDNA, without explicit disclosure of an encoded amino acid sequence, is anticipating."

Kevin,

Unless I'm missing something here, disclosure of the cDNA doesn't (or shouldn't) anticipate the isolated polypeptide from the encoded amino acid sequence, unless the polypeptide is actually expressed. But you may be staring at a prima facie case of obviousness instead.

[Kevin E. Noonan]

Dear EG:

You would be correct, outside the world of the Patent Office. In Chuang, the Board affirmatively stated, relying on the Donohue case, that the prior art did not have to practice the invention to anticipate, just show that the skilled worker would be enabled for practicing the invention.

So I think the logical extension would be that the existence of a cDNA in a database (it would have to be a cDNA, which would show that something was expressed), with or without annotation of an open reading frame (which might otherwise raise an In re Hall problem) might be enough to anticipate, based on the Chuang reasoning that this would be enough to enable the isolated protein.

I suspect one countervailing argument would be that under those circumstances, the skilled worker would not know how to isolate the protein once she had produced it, but I'm sure there is some answer to that argument that the Board could conjure (this point did not come up, it appears, in Chuang).

Thanks for the comment.

[Fan]

Another fact that the BPAI did not consider is that recombinantly expressed polypeptides invariably have different post-translational modifications, which a simple amino acid sequence cannot dictate and thus anticipate. So, "an isolated polypeptide" requires more than just a recited aa sequence.

[EG]

Kevin,

As I thought, the BPAI mischaracterized Donohue, or at least cited it out of context. Here's the key quote from Donohue:

"Appellant argues that the Fields affidavit, which states that the authors of Nomura did not make the disclosed dicarboxylic acid TMBP and dimethyl ester TMBP compounds, overcomes the PTO's rejection. It is urged that Donohue I and In re Samour, 571 F.2d 559, 197 USPQ 1 (CCPA 1978), require, inter alia, that a 35 U.S.C. Sec. 102(b) rejection based on a primary reference disclosing a claimed compound in conjunction with one or more references which teach how to make that compound, should be sustained only if the claimed compound was actually made. We disagree."

What the above quote says it that a rejection under 102(b) based on a disclosure of the compound doesn't require that the disclosed compound actually be made (actual reduction to practice) if other references would teach how to make the compound. Conversely, if the compound isn't disclosed by the reference, there is no anticipation, even under Donohue. In fact, the Federal Circuit didn't just rely on 102(b) in Donohue (if I read it correctly), but also 103.

[Evan]

This result does not seem to be a departure from longstanding CCPA and CAFC precedent. Anticipation is not judged against what the reference strictly recites. Rather, we judge anticipation against that which the reference places into the possession of the skilled artisan. After all, it is the skilled biologist who reads the reference, not the skilled janitor.

[Kevin E. Noonan]

Dear Evan:

And that's the point. I think there is some logic to the Board's decision in Chuang; I think the extension I propose might be an over-extension of the principle. And I do think we need to consider the policy implications.

Thanks for the comment.

[Grimace]

What is the argument that a disclosed amino acid sequence does not anticipate a peptide having that amino acid sequence? Put the other way around, I dont think that anyone would require a patentee to make every species explicitly disclosed in his application in order to claim those individual species. If the sequence is disclosed and the peptide enabled, that is good enough to support a claim, and should be good enough to anticipate.

[Kevin E. Noonan]

Dear Grimace:

I think your point explains why Chuang lost. But I think an important issue is that there is a disparity between what is required to enable for patenting purposes (i.e., you need a utility, stemming from the "make and use" language of 112) and the anticipation standard here that the mere recitation of an amino acid sequence in the prior art is enough. Now, in the past the existence of an isolated protein in the art was enough since someone had actually isolated the protein. Here, you have the (anomalous) result that the isolation of a cDNA (a different molecule) anticipates a claim for an isolated protein, which the Board admits was never made in the prior art. The link is the genetic code, wherein knowing the nucleotide sequence of the cDNA lets you "know" the amino acid sequence of the protein, and that's fine as far as it goes. But since there is no "isolated protein" in the art, the purpose of 102, to prevent anyone from making private what is already in the public domain, doesn't seem to be properly implicated here.

I think the result in Chuang is not particularly wrong or significant (it is non-precedential), but it is the top of the slippery slope towards having all the sequences in all the databases in all the world considered as "prior art" no matter how "known" the sequences were. As I said upthread, unrecognized open reading frames will probably have an In re Hall problem, and that may have to be the limits to the application of the Chuang doctrine. But precluding patent protection on isolated proteins based on cDNA sequences in databases may inhibit investment in those proteins as therapeutics, and that could be a problem. Or not - we'll just have to see what happens.

Thanks for the comment.