In its bitter, seven-year-old lawsuit against Promega Corp., the Roche Group has just been deprived of one of its patents covering Taq DNA polymerase, the enzyme crucial to automating PCR. The federal judge hearing the case decided last month that the inventors of a method for purifying this enzyme had dealt dishonestly with the U.S. Patent and Trademark Office (PTO) in applying for the patent in 1988 and 1989. Because of the inventors' "inequitable conduct," Judge Vaughn R. Walker of the Northern District of California ruled that all claims of the patent are unenforceable.1 The Roche Group (formerly Hoffmann-LaRoche) vows to appeal.
Two U.S. subsidiaries of Roche initiated the lawsuit in 1992, charging Promega with patent infringement. Promega defended itself first by questioning the patent's validity and later by challenging the inventors' strategy for patent approval. Recent rulings in the case have largely cut against Roche, but future developments are hard to predict. Also uncertain is the case's ultimate impact on the cost of PCR reagents. A price comparison in 1998 found that researchers could pay twice as much for licensed Taq DNA polymerase, covered by patents, as for unlicensed enzyme from Promega.2
During the patent-application process, inventors David Gelfand and Susanne Stoffel were at Cetus Corp., part of which Roche acquired in 1991. Their major hurdle was convincing the PTO that they were indeed the first scientists to purify Taq DNA polymerase. In the 1970s, researchers at the University of Cincinnati and in Moscow had tried to purify this thermally stable enzyme.3,4 The PTO initially rejected Gelfand's and Stoffel's application, contending that, in light of this earlier "prior art," their invention just wasn't new. The inventors' response to the PTO, however, convinced the agency to award them Patent No. 4,889,818 nine months later. Its three claims cover the polymerase isolated from the bacteria that normally produce it and from organisms genetically engineered to express it.
It's still not clear whether the Cincinnati or Moscow researchers purified an inoperative, degraded form of the Taq polymerase or the full-length molecule. Experts have taken opposing views.5 "If you read all of those prior-art papers, there is no evidence that anyone prior to Gelfand and Stoffel ever isolated the full-length stuff," says Thomas White, senior vice president of R&D at Roche Molecular Systems, a Roche subsidiary based in Pleasanton, Calif. The position of Madison, Wis.-based Promega, according to vice president and chief technical officer Randall Dimond, has been that "the Cincinnati group gave this enzyme to the public in 1976, and in 1989 the scientists at Cetus stole that back from the public domain."
In narrowing down the issues of the case, Judge Walker so far has sidestepped the question of how novel the work at Cetus really was. His ruling on Dec. 7, based on a trial held last February, instead presents a withering appraisal of Gelfand's and Stoffel's conduct toward the PTO. He cites eight examples of inequitable conduct.
In one instance, Judge Walker noted that the Cincinnati group had used phosphocellulose columns to purify Taq DNA polymerase. Their purification protocol yielded only degraded polymerase, according to Cetus. But Stoffel, the company's scientist, had data showing that a fragment of the polymerase did not bind to phosphocellulose columns. These data suggested that the Cincinnati group had purified full-length enzyme, instead of a fragment (which would have passed through the column). Stoffel never disclosed her finding to the PTO, evidencing an "intent to deceive" the government, according to the judge.
Roche's White responds that Stoffel's phosphocellulose study used 80 mM salt. In contrast, the Cincinnati group's salt concentration was set at 10 mM, allowing many small molecules to bind to the column. "Most biochemists would understand the critical relevance of the salt, even if the judge didn't," he says. (Promega's Dimond contends that the salt elution gradient was similar in both studies.)
In another issue cited by the judge, Cetus argued that the Cincinnati and Moscow groups had accurately determined the molecular weights of their purified protein in the 60-kd range, far lower than the 94-kd weight of full-length Taq DNA polymerase. But Gelfand and Stoffel knew of studies showing that size-exclusion chromatography could underestimate the true molecular weight of a hydrophobic protein like the Taq polymerase. The PTO was never informed of these studies, again convincing the judge of an "intent to deceive."
White notes that the Cincinnati group used Sephadex G-100 for size exclusion chromatography, while the studies known to the Cetus scientists used other types of matrices. Since "you can't predict from one matrix to another what the actual running properties of a particular separation will be," he says that the studies known to Gelfand and Stoffel wouldn't have helped the PTO in assessing the Cincinnati group's results.
In response to Judge Walker's decision, Gelfand and Stoffel, now at Roche Molecular Systems, stated: "This situation, where inventors' reputations are attacked instead of focusing on the validity of their discoveries, puts all scientists at risk of having their raw data and good faith interpretations of it be misrepresented in court."
Judge Walker has scheduled a hearing on the case for later this month, and Promega is setting its sights on Roche's other PCR-related patents, often treated as a unit for licensing purposes. "We will be asking [Judge Walker] to consider declaring the patents unenforceable because of the way they were tied together with the patent that was fraudulently obtained," says Dimond. Melinda Griffith, Roche Molecular Systems' general counsel, counters that "this ruling does not and should not affect the foundational PCR patents with their different subject matter, different inventors, and different claims."
Griffith also notes that the vast majority of Taq DNA polymerase is now made using recombinant technology. One of Roche's patents (No. 5,079,352) covers recombinant DNA vectors encoding the polymerase. Thus, in this bout, Roche may be down, but it is hardly out.
Douglas Steinberg is a freelance writer in New York.
1. The judge's order and information about the lawsuit can be found at Promega's Web site, www.promega.com/taqlegal.
2. S. Beck, "Do you have a license? Products licensed for PCR in research applications," The Scientist, 12:21, June 8, 1998.
3. A. Chien et al., "Deoxyribonucleic acid polymerase from the extreme thermophile Thermus aquaticus," Journal of Bacteriology, 127:1550-7, 1976.
4. A.S. Kaledin et al., "Isolation and properties of DNA polymerase from extremely thermophilic bacterium Thermus aquaticus YT-I," Biokhimiya, 45:644-51, 1980 (English translation published in Biochemistry, 45:494-501, 1980).
5. R. Finn, "Ongoing enzyme patent dispute may have ramifications for academic researchers," The Scientist, 10:1, Oct. 14, 1996.