This is a long and drawn out post on intellectual property, possibly not suitable for those with attention deficit disorder.

<<<< Warning! This post may cause somnolence or ED. >>>>

In this post, I have attempted to make a case that current practice in granting US patents contains a flaw that may be counter to the public interest. See what you think.

In my view, there is a curious discontinuity between the practice of determining obviousness and the allowance of Markush claims in US Patent law.  A Markush claim in the context of chemistry refers to a claim of a generic chemical structure defined by symbols that represent sets of functionally related moieties or structures.  Very often a core moiety is defined and one or more substructure symbols or other symbols representing various chemical elements are attached. 

As an example of Markush claims, consider US 4,237,133, an expired Pfizer patent dated December 2, 1980.  I “randomly” found this patent by searching under the key words “bromination” and “aromatics” at the USPTO website. This patent is illustrative of the point I want to make and my use here is not meant to defame or otherwise irritate Pfizer. I have no connection to this art in any sense.

The ‘133 patent is a fairly ordinary chemical patent. It contains 10 claims- two independent claims (claims 1 and 10) and 8 dependent claims that are ultimately based on claim 1.  Claim 1 is a Markush claim that defines a set of chemical compositions that the PTO has allowed the assignee, Pfizer, to have a legal monopoly on.   Basically, Pfizer was allowed two varieties of claims: a) a composition of matter,  and b) the process of producing analgesia in mammals, based on the compounds in claim 1.

Claims 2 through 8 are a series of “necking down” refinements to more preferred embodiments that are especially meaningful to the assignee.  Preferred embodiments are specific features that the assignee apparently wished to have clear definition to avoid ambiguity.

A patent must be “enabling”. That is, the patent must teach enough of the art to allow a Phosita the chance to see and avoid the patented art.  This is the whole purpose of publishing a patent.  If the state is to grant exclusive rights to a composition of matter or a process, then the public needs to have a fair chance to avoid infringement. The content ahead of the claim section is called the specification and it must contain information that, when combined with the claims, enable a reader to understand exactly what is being claimed and under what constraints.  In the case of composition of matter, it is common to disclose the procedures used to make the composition so there is no doubt by Phosita as to what conditions lead to the claimed material.

The patent claims a tricyclic ring systen festooned with functionalities, some of which are variable.  Variable groups are R1, R2, R3, R4, Z, and W.  R1 is further subdivided into other moieties, some bearing variable groups R’ and R” and appended to a chain bearing p methylene units -(CH2)-, where p may range from 0 to 4 .  Z and W are also comprised of features subject to variability.

The point is that the set of all claimed species is quite large.  Not surprisingly, one could easily wander into claimed composition space because, ordinarily, CAS does not capture all of the compositions from the Markush claim.

It is not required that the applicant prove that they have prepared each permutation in the set of claims, nor is it required that the enabling procedures specifically address each claimed species. The ‘133 patent has 43 procedures, many of which are for intermediate compounds, at best a number that falls far short of the entire set of claimed compositions.  Usually, it is sufficient for illustrative examples or preferred embodiments to be set forth in procedure.

If you think of each group as a spatial dimension, a generic core species with n variable groups essentially maps out a set of structures occupying a kind of n-dimensional space, subject to specific exclusions. When the variable groups are defined as alkyl, aryl, alkoxy, alkenyl, etc., the number of claimed species can be quite large due to the vast number of possible combinations of groups.  Even limited ranges, i.e., R = C20 alkyl, etc., can result in huge collections of claimed species owing to structural isomerism. 

The concept of obviousness in patent prosecution is one of the most vexing and mercurial ideas I can think of.  The code is set forth in 35 USC 103.  A patent attorney will caution that there is no hard and fast universal definition in advance of litigation because what really matters is how a judge decides the matter.  In a practical sense, though, obviousness depends on how the examiners interpret the code.

On to the point of this posting.  While it is possible for an applicant to claim compositions never made or compositions that should exist by reasonable extrapolation, claims in the reverse sense are more problematic. But what do I mean?

Consider US 7,235,700, a process patent claiming the preparation of a cyclohexenone functionalized on the beta carbon with an enol ether group.  [Disclaimer: again, this patent was “randomly” chosen. I have no specific axe to grind with the assignees or the inventors.  I do, however, have an axe to grind with US patent law.]

This process is a good piece of journeyman organic synthesis featuring the preparation of an alpha/omega functionalized fragment with a Grignard functional group on one end and a silyl-protected oxygen on the other.  From Example 2 of the ‘700 patent, to the Grignard reagent, made in the customary fashion in diethyl ether with dibromoethane as an entrainment additive, was added a THF solution of the cyclohexenone enol ether.  The Grignard added to the enone in 1,2 fashion to afford a tertiary alcohol which upon acid hydrolysis, the resulting alcohol eliminates and the 3-alkoxy enol ether hydrolyzes to afford the product cyclohexenone on workup. 

The patent teaches that the inventors had a poor process before this patented process (column 1, line 37).  So, this must be an improvement, right? It seems to be. But, should it receive a patent?

From my copy of Kharasch and Reinmuth, I see that 1,2-additions of Grignard reagents to cyclohexenone were reported as early as 1941 (Whitmore, Pedlow JACS, 1941, 63, 758-760).  So the knowledge of 1,2- vs 1,4-additions by RMgX nucleophilic additions to cyclohexenones resulting in primarily 1,2-addition is not new. 

The use of nucleophiles with protected incompatible functional groups is not new.  The hydrolysis of enol ethers is not new.  Indeed, nowhere in the description do the inventors state that the disclosed transformations were “surprising” or “unexpected” in their outcome.  As a phosita myself, I look at this patent and see good solid organic synthesis.  I see the results of workers who have undergone training in the usual graduate level chemistry curriculum. Advanced organic synthesis with attention to donors and acceptors, functional group transformations, and protection/deprotection schemes.  They took known transformations and assembled the pieces into the desired molecule.

My objection is this.  Under the convention that Markush claims are allowed under current practices, many compositions of matter can be claimed by virtue of simple declaration despite the fact that homologous series or the usual genus groups of radicals (alkyl, alkenyl, alkynyl, aryl, heteroaryl, etc.) may be rather obvious additions to the list.  A Phosita would reasonably state that if methyl is feasible, then so is ethyl, propyl, butyl, …, alkyl.  Markush claims invoke a kind of obviousness that is allowed. 

However, the same principle may not apply in reverse. That aspect of the body of scientific work teaching that certain generalizations are possible does not seem to be allowed in determinations of obviousness. 

In the instant example, the generalization is that Grignards as a class might be expected to add in the fashion claimed in the ‘700 patent.  Or that enol ethers as a class would be expected to undergo acid catalyzed hydrolysis to ketones.  In the ‘700 patent, elements of the claim are novel only by virtue of being obscure members of a very large set of possibilities.

So, on the novelty and obviousness side of examination, the fact that a claim uses known transformations or schemes on heretofore unreported substrates bearing known features seems to be sufficient to cause an examiner to allow the claim.  The allowance of Markush claims then allows broad generalization into large sets of claimed structures.

But generalization from a broad area of knowledge may not necessarily bar an unreported claim element when acted upon by known influences resulting in transformations that are consistent with the broad knowledge, as in the case of the ‘700 patent.

A patent lawyer reading this might object that the novelty of the substrate and the lack of specific precedence confers novelty and non-obviousness under current precedent.  That lawyers opinion might be internally consistent with precedent and most would leave it at that. 

But the overarching concern that I want to draw attention to is that the current practice in relation to novelty and obviousness may not serve the public interest. I’m seeing far too many patents being allowed for the application of known transformations to substrates that are merely obscure.  What passes for inventorship is often just good craftsmanship. Reacting a Grignard reagent with a ketone followed by elimination is a general process that we might teach to students in a classroom. 

Indeed, the current practice of teaching chemists is to expose the student to systematic generalizations of reaction-types so that they can go out and put generalizations into practice rather that have to memorize countless specifics. 

When these chemists apply their training by reducing generalizations to practice on specific substrates, however, it seems they can claim to have made an invention under US patent law. 

The upshot is that a good deal of technology resulting from ordinary problem solving skills is barred from the public domain for 20 years.  Not that I believe that privately developed inventions should be in the public domain. But I will point out that it is quite easy for a company to get clobbered by an infringement suit for stumbling into claimed art by practicing what their chemists learned to do in graduate school. Reducing general reactions to practice.

I suspect that it is common practice for companies to believe that if something is patentable, then a patent is manditory.  Unfortunately, the current system seems over-generous in granting 20 year monopolies for dubious inventions. When the threshold for obtaining a patent is too low, when practices are too easily removed from the commons, others trying to practice the art are unreasonably restrained.

Reform of matters as basic as the definition of obviousness and novelty cannot come from the USPTO, the courts, or from patent attorneys. Applicants and their attorneys will continue to game the system to the extent allowed by the courts. Fundamental change must come from legislation.