Full opinion text
OPINION AND ORDER WILLIAM C. CONNER, Senior District Judge. This is an action for alleged infringement of three United States patents on acellular antigens and vaccines for pertussis, or whooping cough. The patents were issued in the name of Pavel Novotny on applications filed by his employer, Burroughs Wellcome Foundation Ltd. (“Burroughs Wellcome”). Plaintiff Medeva PLC (“Medeva”) purchased the vaccine business of Burroughs. Wellcome and the patents were assigned to Medeva’s subsidiary, plaintiff ■ Evans Medical Ltd. (“Evans”), which granted an exclusive license under the patents to plaintiff SmithKline Beecham Biologicals S.A. (“SmithKline”). Plaintiffs charge that the patents are infringed by a composite “DTaP” vaccine for diphtheria, tetanus and pertussis sold in the. United States by defendants American Cyanamid Co. (“American Cyanamid”) and American Home Products Corp. (“American Home Products”) under the brand name ACEL-IMUNE®, and incorporating acellular pertussis antigens produced in Japan by defendant Takeda Chemical Industries Ltd. (“Takeda”). The action is before the Court on a welter of pretrial motions. Defendants have moved:, (1) for a ruling by the Court under Markman v. Westview Instruments, Inc., 517 U.S. 370, 116 S.Ct. 1384, 134 L.Ed.2d 577 (1996), construing the patent claims and for summary judgment that the claims, as so construed, have not been infringed by defendants’ accused vaccine; (2) for summary judgment that the patents in suit are invalid because their specifications (all of which are identical) fail to disclose the best mode of practicing the invention known to the patentee at the time the applications were filed, as required by 35 U.S.C. § 112; (3) for summary judgment that the patents are invalid under 35 U.S.C. § 102 because the claimed inventions were anticipated by a prior patent of defendant Takeda and by the prior sale and use of Takeda’s antigen in this country; (4) for summary judgment that the patents are invalid under 35 U.S.C. §§ 102 and 103 because the claimed inventions were anticipated by or obvious in view, of a March 1985 article co-authored by Novotny, Juan Montaraz (a doctoral student under Novotny’s supervision), and Juraj Ivanyi; (5) for an order in limine excluding certain testimony of plaintiffs’ patent law expert Gerald Bjorge; and (6) for an order in limine excluding certain testimony of Tom Bozzo or any other witness regarding FDA approval of defendants’ vaccine. Plaintiffs have moved: (1) for partial summary judgment that the patents in suit are not anticipated by the prior patents and publications cited by defendants; (2) for partial summary judgment that, for purposes of determining whether the patent specification satisfies the best mode requirement of 35 U.S.C. § 112, the January 1990 deposit in the European Collection of Animal Cell Cultures of a hybridoma which secretes a monoclonal antibody used in purification of the patented antigen relates back to the May 1,1985 filing of the parent U.S. application; (3) for an order pursuant to 35 U.S.C. § 256 directing correction of the patents in suit to add Mon-taraz as an inventor; (4) for an order compelling defendants to return certain inadvertently produced documents; (5) for an order in limine excluding the expert testimony of Dr. Alison Weiss; and (6) for an order in limine excluding evidence at trial of related litigation in the United Kingdom. Following a background discussion, the motions will be discussed serially. Unless otherwise indicated, all of the background facts recited herein are undisputed. BACKGROUND Development of pertussis vaccines Pertussis, or whooping cough, is one of the most common serious diseases of infants, with an estimated 50 million cases annually causing 600,000 deaths. A vaccine introduced in the late 1940’s strikingly reduced the incidence of the disease in developed countries where large scale immunization programs were undertaken. But that vaccine was prepared from chemically killed whole cells of Bordetella pertussis (“B. pertussis”), the bacterium which causes whooping cough, and caused serious side effects in a significant number of those inoculated, including fever and persistent screaming. This resulted in curtailment of the immunization programs and a concomitant increase in the incidence of the disease. The vaccines involved in the present case are acellular, incorporating not whole bacterial cells but selected proteinaceous material extracted from the outer membrane of B. pertussis bacteria by, for example, treatment with an amino acid solution. This breaks down the proteins of the outer membrane into fragments comprising long strands of amino acids connected in a sequence which determines the properties and behavior of the material. The extracted material is then subjected to a series of purification steps designed to increase the concentration of a particular protein, now known as pertaetin, which has been found to have excellent im-munogenic effect against B. pertussis infection. The pertaetin strands have a number of active sites or epitopes which are recognized by the defensive white blood cells, or B lymphocytes, of the inoculated patient as characteristic of B, pertussis bacteria, thereby stimulating the lymphocytes to produce antibodies which will attack B. pertussis bacteria in the event of a later infection and provide effective immunity without the serious side effects of whole cell vaccines. Acel-lular antigens containing pertaetin are currently in widespread use in vaccines for B. pertussis, frequently in composite DTaP vaccines designed to immunize against diphtheria and tetanus as well. The patents in suit Plaintiffs have sued on three U.S. patents: No. 5,237,052 (the “ ’052 patent”), No. 5,438,-120 (the “120 patent”) and No. 5,648,080 (the “’080 patent”). (Exhs. 1-3 to Defs.’ Motions for Summ. J.) All three patents have identical specifications and are based upon the same parent U.S. application, Serial No. 929,257, filed May 1, 1985, which claims priority based upon the United Kingdom specification Serial No. 8,412,207, filed May 12,1984.. The patents were issued in the name of Pavel Novotny and were based in part on work done in the laboratory of Burroughs Wellcome by Novotny’s student assistant, Juan Montaraz, who used the research as the basis' for his doctoral thesis and as the subject of an article entitled “Identification of a 68-Kilodalton Protective Protein Antigen from Bordetella bronchiseptica,” published in the March 1985 issue of Infection And Immunity in the names of Montaraz, Novotny and Juraj Ivanyi (the “Montaraz et al. article”). The common specification of the patents in suit identifies the desired antigen as “protei-naceous material associated with adenylate cyclase activity” (referred to by the acronym ACAP) and as having an isoelectric point (“pi”) of “about 7,” “a relative molecular weight of about 67,000 to 73,000, particularly 69,000” and a “proline:glutamic acid ratio of about 1:1.” The specification discloses a preferred process of preparing the patented antigen. The first step of this process, denominated “Example 1,” involves incubating a culture of B. pertussis cells with an aqueous amino acid buffer with a pH of about 3 for 10-20 hours at 30-45° C. The mixture is centrifuged at 100,000g to separate the cells and particulate matter. The supernatant extract is then subjected to a multi-step purification process, the several steps being respectively denominated “Examples 2(a), 2(b) and 3.”- “Example 2(a)” consists of separating the fractions of the extracted outer membrane material by chromatography in a DEAE-Trisaeryl column. The retained protein molecules are eluted from the column by flushing with an 0.2M NaCl buffer. “Example 2(b)” consists of preparative flat-bed isoelectric focusing (“IEF”) in a granulated gel. The eluate from Example 2(a) is embedded in the gel. After electrophoresis, the material which collects in the band at a pi of 7 is scraped from the field and dissolved in distilled water. The gel suspensions are placed in columns and eluted with an 0.2M ammonium bicarbonate buffer at pH 7.0 to produce a gel-free eluate. “Example 3” consists of further purification of the separated protein material in an immunosorbent chromatographic column charged with “a monoclonal immunoglobulin specific for ACAP.” In the parent U.S. application filed May 1, 1985, the monoclonal antibody employed in this final purification step was not otherwise described. It was not until September 21, 1992, over seven years after the filing of the application, and over eight years after the claimed U.K. priority date, that a fifth-generation continuation of the parent U.S. application was amended to add the following reference for identification of the key monoclonal antibody: The hybridoma which secretes the monoclonal immunoglobulin was deposited under the Budapest Treaty at the European Collection of Animal Cell Cultures, Porton Down, United Kingdom on Jan. 5, 1990 under accession number 90010501. (Exh. 64.) As the final step in the process of preparation described in the patent specification, denominated “Example 4,” the purified ACAP is cultured in a liquid medium containing 2% agar and 5% horse blood for 48 hours at 36-37° C while agitating to provide a gas exchange rate of 20-40 p, M oxygen/hr. The liquid cultures are then used to inoculate the medium in a glass fermentor, with the pH maintained at 7.6 by the controlled addition of 2M HC1 and the oxygen saturation maintained at 5-10% by impeller agitation. The cultures are harvested before the end of the exponential growth phase, or after approximately 36 hours of incubation. “Example 5” consists of verifying the protective potency of the antigen by employing it in a standard Kendrick test on mice. “Example 6” involves analysis of the ACAP to ascertain its relative content of the respective amino acids of which it is composed. Sixteen different amino acids had measurable residues after hydrolysis and dessieation. Proline and glutamic acid had relative amino acid values of 60 and 62, respectively. “Example 7” involves use of the ACAP antigen in aqueous solutions to foivn pertussis vaccine or, in combination with diphtheria and tetanus antitoxins, to form composite DTaP vaccines. The ’052 patent has only two claims. Claim 1 reads as follows: A purified ' Bordetella pertussis antigen characterized by the following features: a relative molecular weight of between 67,000 to 73,000 as determined by 12% (w/w) polyacrylamide gel electrophoresis; a ratio of proline to glutamic acid of substantially 1:1 as determined by amino acid analysis. Claim 2 is identical to Claim 1 except that it specifies “a relative molecular weight of 69,-000.” The ’120 patent contains only one claim, which is identical to Claim 1 of the ’052 patent except that it specifies “a relative molecular weight of between 67,000 and 69,-000.” The ’080 patent contains fourteen composition claims (Nos.1-14) directed to a vaccine incorporating the antigen of the ’052 and ’120 patents and fourteen method claims (Nos.15-28) directed to inducing an immune response in a patient by administering the vaccine of Claims 1-14. Claim 1, the only independent composition claim of the ’080 patent, reads as follows: A vaccine which comprises a proteina-ceous material, which is derived from the outer membrane of Bordetella pertussis and is characterized by a relative molecular weight of about 67,000 to about 73,000 as determined by 12% (w/w) polyacrylam-ide gel electrophoresis and a proline:glu-tamie acid ratio of about 1:1 as determined by amino acid analysis in a pharmaceutically acceptable carrier or adjuvant. Claim 15, the only independent method claim, reads as follows: • A method of inducing an immune response in a patient, which method comprises administering to said patient a vaccine which comprises a proteinaceous material, which is derived from the outer membrane of Bordetella pertussis and is characterized by a relative molecular weight of about 67,000 to about 73,000 as determined by 12% (w/w) polyacrylamide gel electrophoresis and a proline:glutamic acid ratio of about 1:1, as determined by amino acid analysis in a pharmaceutically acceptable carrier or adjuvant. Prosecution history of the patents in suit The U.K. application from which priority is claimed for the patents in suit was filed May 12, 1984. Its disclosure differed somewhat from that of the parent U.S. application. It began by acknowledging that acellular' pertussis, vaccines containing antigens extracted from the outer membrane of B. pertussis bacteria were known in the prior art, but asserted that “[w]e have now discovered that adenylate cyclase, an enzyme found in the cultures of B. pertussis_is a major protective antigen against B. pertussis,” a discovery which “permits the preparation of vaccine formulations comprising antigenic preparations which are free from, or contain reduced amounts of, other B. pertussis components which are responsible for the toxic side-effects demonstrated by whole-cell vaccines.” The U.K. specification thus proceeded to describe a procedure for preparing “an antigenic preparation comprising adenylate cyclase .... [which] generally has a molecular weight of about 69,000 and an isoelectric point of 7.6-7Ü under preparative conditions .... [and which] may, if desired, contain minor quantities of other antigenic compounds .... [but is] preferably substantially free from other antigenic components.” It is therefore clear that at the time the . U.K. priority application was filed, Novotny and his associates believed that the antigenic enzyme critical to effective protection against B. pertussis infection was adenylate cyclase, and they accordingly described a method of extracting it, along with other proteinaceous material, from the outer membrane of B. pertussis bacteria and purifying it to substantially eliminate the other proteinaceous material. Indeed, Novotny admitted this in an Addendum to his 1995 paper “Identification of 68, 69 and 71 kDa Proteins of Bordetella Species,” and added that it was not until after a visit by Professor Erik Hewlett and Dr. Alison Weiss from the United States in October 1984 that he “started to have serious doubt” that the protective enzyme was ade-nylate cyclase. (Exh. 22, Addendum.) He ultimately concluded that the effective 69 kDa antigen was not adenylate cyclase, but a different enzyme with “adenylate cyclase activity.” (See Novotny Dep., June 4, 1997, Exh. 50 at 7, 43-44.) Novotny’s original U.S. application, filed May 1, 1985, accordingly described the key antigen less specifically as “proteinaceous material associated with adenylate cyclase activity (abbreviated to ‘ACAP’ hereinafter)” and added that it “may comprise the adeny-late cyclase enzyme per se or a binding protein for the enzyme.” Claim 1 of the U.S. application as filed described the antigen merely as “proteina-ceous material associated with adenylate cyc-lase activity (ACAP).” A series of dependent claims respectively added other characteristics of the material, such as “a relative molecular weight (MW) of 67,000-73,000,” “an isoe-lectric point (pi) of 7.0-7.4 under preparative isoelectric focusing (IEF) conditions” and a “ratio of proline to glutamic acid residues ... [of] substantially 1:1.” All of the claims were rejected by the PTO Examiner as anticipated by or obvious in view of a paper by Erik Hewlett and J. Wolff entitled “Soluble Adenylate Cyclase from the Culture Medium of Bordetella pertussis: Purification and Characterization,” published in the August 1976 edition of BACTERIOLOGY (the “Hewlett & Wolff article,” Gram Aff., Exh. F). That article disclosed the extraction of adenylate cyclase from B. pertussis and its purification first by DEAE-cellulose chromatography and then by SDS electrophoresis. The purified product had a molecular weight of 69,000. Through some seven year's of prosecution, and a series of continuation and divisional applications, the broadest claims were amended to import limitations from the original dependent claims, including those specifying molecular weight and the pro-line:glutamic acid ratio. Novotny then sought to overcome the rejection based on the Hewlett & Wolff article by filing a declaration under 37 CFR, § 1.132 of Erik Hewlett, one of the authors of the article, stating that the adenylate cyclase produced by the process described therein did not have a proline:glutamic acid ratio of substantially 1:1, as called for in the amended claims. The PTO Examiner also repeatedly rejected all of the claims as anticipated by or obvious in view of the prior U.S. patent No. 3,141,824 to Robert Dahlstrom which was issued July 21, 1964 on an application filed May 29, 1961 (the “Dahlstrom patent,” Exh. 6). Dahlstrom disclosed the extraction of a B. pertussis antigen from B. pertussis bacteria by a process which the Examiner concluded would produce some ACAP even if Dahlstrom had not specifically been seeking it. After the rejection of the claims had been made final, the Examiner suggested that the rejection could be overcome if the claims were further amended to limit them to “A purified ... antigen.” Such an amendment was filed August 22, 1992 (see Exh. 43) and ultimately resulted in the allowance of the applications. Defendants’ accused antigens and vaccine Takeda has manufactured and sold an acel-lular pertussis vaccine in Japan since 1981. On September 12,1980, it filed an application for a Japanese patent on the method of producing toxoid (antigen) for such a vaccine, and filed a corresponding U.S. application on January 30, 1981. U.S. patent No. 4,455,297 (the “Takeda patent”) was issued June 19, 1984 on a continuation of the U.S. application. {See Exh. 5.) Like Novotny, Takeda extracts a mixture of antigens from B. pertussis bacteria and treats the extract with formalin (an aqueous solution of formaldehyde) to remove toxins. Unlike Novotny, however, Takeda does not attempt to purify the mixture to eliminate other antigens and achieve a concentrated pertactin. Thus, pertaetin constitutes only about 4% of the Takeda mixture of antigens. The relative molecular weight of the Takeda pertaetin is approximately 69 kDa, but its proline:glutamic acid ratio is about 0.86:1, as measured by amino acid analysis. Further, Takeda’s tests show that Takeda’s pertactin has no adenylate cyclase activity, a finding which plaintiffs have not disputed. • DISCUSSION I. DEFENDANTS’ MOTION FOR A MARKMAN RULING Markman v. Westview Instruments, 517 U.S. 370, 116 S.Ct. 1384, 134 L.Ed.2d 577 (1996), holds that patent claim construction is an issue of law to be resolved by the Court. In jury eases, once the Court has construed the claims, the jury determines as an issue of fact whether the claims, as thus construed, are infringed — either literally or under the doctrine of equivalents — by the accused product or method. Wright Med. Tech., Inc. v. Osteonics Corp., 122 F.3d 1440, 1443 (Fed.Cir.1997). In determining the meaning of the claims, the Court must first examine the “intrinsic” evidence — i.e., the claims themselves, the specification and, if in evidence, the prosecution history, Vitronics Corp. v. Conceptronic, Inc., 90 F.3d 1576, 1582 (Fed.Cir.1996). If this intrinsic evidence permits an unambiguous construction of the claim language, the Court need not, and indeed should not, consider extrinsic evidence, such as testimony from expert witnesses as to special meanings which the terms of the claims have for those skilled in the art. See id. at 1585. In this ease, defendants raise three questions concerning the applicability of the language of the broadest patent claims to their accused antigens. They contend (1) that their accused ACEL-IMUNE® vaccine does not include “purified” 69k, as called for in the claims of the ’052 and T20 patents, but a mixture of antigens of which 69k constitutes only 4%; (2) that their 69k antigen does not have a proline:glutamic acid ratio of “substantially 1:1” as called for in the claims of the ’052 and T20 patents, or “about 1:1” as called for in the claims of the ’080 patent, but a ratio of 0.86:1; and (3) that the claims of all three patents in suit must be limited to cover only antigens having adenylate cyclase activity, which their antigens do not have. Thus the Markmcm issues to be resolved by the Court concern the meaning of the terms “purified” and “substantially 1:1” or “about 1:1,” and whether a requirement of adenylate cyclase activity should be read into the claims. These issues will be discussed separately in that order. A. What does “purified” mean in the context of the patents in suit? There can be no question as to the critical importance of this limitation which appears in each of the claims of the ’052 and ’120 patents. All the application claims were repeatedly rejected as unpatentable over the prior art references (especially the Dahlstrom patent) over some seven years of prosecution and through many continuation and divisional applications until, in response to the Examiner’s repeated suggestion, the claims were finally amended to limit them to a “purified” 69k antigen. However, the parties hotly dispute the meaning of this limitation. Plaintiffs contend that it means only that the desired material has been at least partially separated or isolated from an unwanted material — in other words, that any degree of purification suffices. Defendants urge a construction at the opposite end of the spectrum, one requiring that the extracted proteina-ceous material be put through all of the purification steps described in the specification, including the final immunopurification step of “Example 3.” On the basis of the intrinsic evidence alone, it is difficult to conclude that either of these extreme positions is the correct one, although that of defendants seems closer to the mark. Our analysis begins of course with the patent specification, which discloses a vaccine wherein the only active ingredient mentioned is “an antigenic preparation derived from B. pertussis comprising ACAP, optionally tox-oided e.g. using formalin, glutaraldehyde or B-propiolactone, together with a pharmaceu-tically acceptable carrier therefor.” (’052, 2/54-59.) All the specific examples of vaccine formulations given include “antigen according to the invention” as the only active ingredient. (’052, 9/36-10/26.) The specification, after discussing the prior research in “isolating and purifying the 20 or more surface antigens of the B. pertussis organism and characterising, their ability to induce immune reactions,” states that among the antigens suggested for such investigation is adenylate cyclase. (’052, 1/54-64.) The specification then discusses several prior methods of reducing the toxicity of the extracted .proteinaceous material, including that of defendant Takeda’s U.K. Patent Specification 2 047 358 A, which discloses production of a B. pertussis extract vaccine “involving removal of endotoxin from culture superna-tants.” (’052, 2/6-9.) The specification initiates its discussion of the patented invention by stating: It has now been discovered that certain proteinaceous material, associated with adenylate cyclase activity, as hereinafter described, found in the cultures of B. pertussis, is capable of providing protection against challenge by B. pertussis when administered to experimental animals. This discovery that the proteinaceous material usually associated with adenylate cyclase activity is a major protective antigen against B. pertussis permits the preparation of vaccine formulations comprising antigenic preparations which are free from, or contain reduced amounts of, other known B. pertussis components which may be responsible for the toxic side-effects demonstrated by whole cell vaccines. (’052 2/28-40; emphasis added.) The specification later adds that although the patented antigenic preparations may have “minor quantities of other antigenic compounds, in addition to the ACAP,” they are “preferably substantially free from other antigenic compounds.” (’052, 3/34-42; emphasis added.) Then, after describing the disadvantages of prior methods of extracting ACAP from the outer membrane of B. pertussis organisms, the specification states: We have now discovered that ... extraction of B. pertussis organisms using regulated, mildly acidic conditions results in the extraction of substantially increased yields (about 40x better than previously reported techniques) of adenylate cyclase from the outer membrane in a form which is water-soluble. (’052, 4/6-12.) The specification then more specifically describes the extraction process, including incubation of the B. pertussis cells with an amino acid buffer and centrifuging to separate the cells from the ACAP-eontaining supernatant. (’052, 4/13-37.) It continues to describe use of the supernatant extract in the Kendrick test on mice, from which it was learned that ... [cjontrol vaccines containing no adeny-late cyclase activity were found to provide little or no protection against challenge with B. pertussis, suggesting that ACAP may, in fact, be the most important factor in immunity. Analysis of batches of non-protective whole-cell vaccine has also shown that non-protection tends to be associated with a lack of adenylate cyclase activity, further suggesting that ACAP may be the key antigen necessary for eliciting an immune response against B. pertussis. (’052,4/41-50; emphasis added.) The specification then teaches a multi-step purification process for producing the “purified” ACAP antigen, commencing with the “Crude Outer Membrane Proteins” extracted in the first step, which is described in “Example 1.” (’052, 5/54-55.) Plaintiffs have admitted that this “crude” mixture is not a “purified” material. (Exh. 30, Interrog.# 9.) In support of their contention that the ACAP is “purified” only after it has been put through all the steps of the purification process described in the specification, including the immuno-purification step of “Example 3,” defendants point out that “Example 3” is entitled “Purification of ACAP Using a Monoclonal Immunosorbent Column,” (’052, 7/26-28), and that none of the products of the earlier steps in the process is described as a “purified” material. On the other hand, plaintiffs lay principal stress on the fact that in its initial, summary description of the purification process, the specification states: The supernatant extract [from “Example 1”] ... may ... contain the ACAP in small quantities complexed with other proteins including fragments of LPS [lipopolysac-eharide, which is suspected as the cause of toxic side effects], in which case it may be desirable to purify further the material for use in the vaccine formulations according to the invention. Thus, for example, further purification may be effected by ion exchange chromatography and/or by preparative isoelectro-focussing to eliminate complexed material... .After the above-described purification steps the ACAP may, if desired, be further purified, for example by passing the material through an immunosorbent column containing an appropriate monoclonal antibody against the ACAP. . . (’052, 4/51-68; emphasis added.) Since the process steps denominated “Example 2(a)” and “Example 2(b)” are referred to as “purification” of the extract,, plaintiffs argue-that the product of those steps must be deemec[ to have been “purified” in the lexicography of the patent. However, this argument proves too much: because these steps are described as “further purification,” carrying plaintiffs’ argument to its logical conclusion would mean that the crude extract resulting from “Example 1” would likewise have to be deemed “purified.” But, as already noted, plaintiffs have admitted that that product is not purified, as indeed they must in view of the prior art and the prosecution'history of the applications for the ’052 and ’120 patents. The limitation of the broadest claims of those applications to a “purified” ACAP material was added during prosecution to overcome the rejection of the claims as unpatentable over the cited references, particularly the Dahlstrom patent. Dahlstrom discloses the preparation of a B. pertussis antigen by extraction of proteinaceous material from B. pertussis cells in a saline solution (pH 8.5-10.5), which is later neutralized by the addition of sterile acid and centrifuged to separate the cells. The extracted antigen was tested for immunogenic potency by injection into mice and challenge by B. pertussis organisms introduced intracerebrally. Because Dahlstrom’s antigen provided excellent immunity, with a 100% survival rate at a dosage of 0.015ml, it must have included a significant amount of ACAP, if we are to believe the teaching of the patents in suit that “vaccines containing no adenylate cyclase activity were found to provide little or no protection against challenge with B. pertussis ” and that ACAP “may be the key antigen necessary for eliciting an immune response against B. pertussis.” (’052, 4/41-50.) Thus there was ample support for the PTO Examiner’s conclusion that the Dahlstrom process “[ijnherently ... would result in the extraction of Applicant’s claimed protein. Note that Applicant’s claims 28-32 do not require the antigen to be purified.” (Exh. 41 ¶ 5.) Novotny acquiesced in that conclusion by amending all of the claims of the application for the ’052 patent so that they cover only a “purified” 69k antigen. Nevertheless, in his co-pending application for the ’120 patent, Novotny made another effort to obtain the coverage which he had thus relinquished, asserting claims similar to those of the ’052 patent except that the word “purified” was replaced with the word “acel-lular” (See Exh. 31 at 3.) These broader claims were rejected as unpatentable over an article by Novotny and K. Cownley — “Effect of Growth Conditions on the Composition and Stability of the Outer Membrane of Bor-detella pertussis ” — published in 1978 as part of the Proceedings of the Third International Symposium on Pertussis (the “Novotny & Cownley article”). In rejecting the claims, the Examiner stated that “[t]he rejected claims do not contain any limitations which would distinguish the claimed products from the isolated outer membrane of the [Novotny & Cownley] article on the basis of purity.” (Exh. 32 ¶ 6.) Novotny thereafter amended all the claims to limit them to “purified” 69k. (See Exh. 34.) Therefore, there can be no dispute that the claim limitation to a “purified” material requires, at the least, sufficient purification to distinguish it from the extracted ACAP-con-taining material of the prior art, including the Dahlstrom patent and the Novotny & Cownley article, which was cited in rejecting the claims. - It is clear that a “purified” ACAP is one that results from subjecting the mixture of proteins extracted from the outer membrane of B. pertussis bacteria to one or more purification steps which, to a substantial degree, isolate ACAP and eliminate the other' antigens in the mixture; The only remaining question is how much purification is required for the material to be deemed “purified” within the meaning of the claims? Defendants’ argument that the third step (immunopurifieation), “Example 3,” is necessary to achieve a “purified” material runs contrary to the specification’s statement that the final step is optional: After the above-described purification steps [“Example 2(a)” and “Example 2(b)”] the ACAP may, if desired, be further puri- fied by passing the material through an immunosorbent column containing an appropriated monoclonal antibody against the ACAP [“Example 3”]. (’052, 4/64-68; emphasis added.) In opposition, plaintiffs argue that, because the limitation to a “purified” 69k was added to the broadest claims of the ’052 and 120 patents to overcome their rejection as anticipated by or obvious in view of the prior Dahlstrom patent, the limitation should be construed as narrowing the claims only to the minimum extent necessary to overcome the rejection based on Dahlstrom. At the oral argument of the motions, plaintiffs’ counsel conceded that the extraction and cen-trifugation steps of Dahlstrom constitute “purification.” (Transcript of May 26, 1998 Oral Argument at 27.) Yet he argued that the claims of the ’052 and 120 patents must be construed so as to cover antigens which have undergone any degree of purification beyond that performed by Dahlstrom. However, that argument considers only one type of intrinsic evidence — the prosecution history — and ignores the equally important intrinsic evidence of the specification itself, which contains strong indications as to the meaning of the term “purified.” The specification teaches that ACAP may be the “key antigen” for immunization against B. pertussis, that “the vaccine formulations according to the invention may, if desired, contain minor quantities of other antigenic compounds, in addition to the ACAP .... [but are] ... preferably substantially free from other antigenic components.” (’052, 3/33-42; emphasis added.) Moreover, the only antigen that is described in the specification as “purified” is the product of the immunopurification step of “Example 3.” At the oral argument, plaintiffs’ counsel contended, with impressive ingenuity, that this step is unnecessary to produce a safe and effective antigen and was disclosed only so that Novotny “could characterize and identify [69k] as part of his teaching.” (Tr. at 26-27.) But Novotny’s specification taught immunopurification as the final step in producing an “antigen according to the invention.” There is not the slightest suggestion that it was disclosed only as a means of identifying the new antigen which Novotny claims to have discovered. After thorough consideration of all the intrinsic evidence, the Court concludes that the term “purified” in the claims of the ’052 and 120 patents means that the mixture of proteins extracted from the outer membrane of B.' pertussis has been treated to reduce the concentration of antigens' other than 69k at least to the point where they are “minor” ingredients and 69k is the major remaining antigen. In other words, 69k must constitute, at the minimum, more than half of the extracted proteinaceous material present in the mixture. If there were any doubt about this conclusion, resort to the most persuasive extrinsic evidence — r-the testimony of the patentee No-votny against his own interest — would only narrow the definition further in favor of the defendants. In pretrial depositions, Novotny stated unequivocally that the extracted material is not “purified” even after it has been subjected to the preliminary purification steps of “Example 2(a)” and “Example 2(b)” and not until it has gone through the final immunopurification step of “Example 3”: Q: At the end of Example 2(a), after you have performed that procedure, do you have something that is purified, as you understand the meaning of that term? A: No, it was crude separation of some components which we weren’t interested in, and some components which we wanted to proceed further, which was made, at that time, electrofocus-ing. Q: What about the procedure of Example 2(b), which is: “Preparative flat bed isoeleetrofocusing [in] granulated gel” .... At the end of that procedure, do you have something that was purified, as you understand that term? A: No, it was still complex. Q: Turning to Example 3 in column 7, that example is entitled: “Purification of ACAP using a monoclonal immuno-sorbent column.” A: That’s correct. Q: What were you trying to achieve in that procedure? A: No, that’s the final step of the purification which started in column 5, because [the] immunosorbent column was specific for the protein. So, it stuck to. the column and was then eluted from the immunosorbent and you obtained a variety of a very pure preparation. Q: So, according to your understanding, it is only after someone completes the procedure of Example 3, with monoclonal antibodies, that you have a purified substance, as you understand that term? A: Yes. Q: Turning back then to column 10, where you see the claim, am I right that you understand the term, “a purified Bordetella pertussis antigen,” that refers to the product that you get after you complete Example 3? A: Yes, sir. You have to finish all of it. (Novotny Dep., June 2, 1997, Exh. 49 at 65-67.) It would be difficult to envision a clearer 'and more positive statement from the patentee himself that the position advanced by defendants as to the meaning of th'e term-“purified” in the claims of his patents is absolutely correct, and that the construction for which plaintiffs contend is plainly wrong. In light of this testimony, it is not surprising that plaintiffs urge the Court of follow the rule that where the intrinsic evidence permits an unambiguous construction of the meaning of the claim terms, reliance on extrinsic evidence is improper. However, plaintiffs themselves invite the Court to violate the rule by offering the declaration of Dr. Carine Capiau, an employee of plaintiff SmithKline, as to the meaning of the claims. She states: “In my opinion, the term ‘purified’ as used in the claims of the ’052 patent and the ’120 patent means that the desired material has been at least partially separated or isolated from an unwanted material.” However, that opinion is pure ipse dixit; Capiau’s declaration cites absolutely no evidence, intrinsic or extrinsic, in support of her opinion. She merely adds: “It is my further opinion that those skilled in the art would define ‘purified’ in such a manner.” Such conclusory statements by an employee of one of the parties carry little if any- probative weight, especially when they contradict clear and positive statements against interest by the patentee himself. Thus, the best and most reliable extrinsic evidence strongly supports the Court’s conclusion that the word “purified” in the claims of the ’052 and ’120 patents means that the proteinaceous material extracted from the outer membrane of B. pertussis bacteria must be purified at least to the extent that 69k is its major antigenic component. Indeed, if the intrinsic evidence left any doubt as to the meaning of the term “purified,” resort to the patentee’s own testimony would lead only to a further narrowing of the claim coverage to require completion of the entire purification process disclosed, including im-munopurification in a chromatographic column charged with a monoclonal antibody which selectively binds 69k. B. Should the claims of the ’080 patent be construed to require “purified” 69k? Defendants contend that, although the limitation to a “purified” material does not appear in any of the claims of the ’080 patent, those claims should be construed as if that limitation were present. That contention is based upon several undisputed facts. First, defendants point out that, in response to an interrogatory reading, “Do plaintiffs contend that the ‘proteinaceous material’ of Claim 1 of the ’080 patent contains something other than the claimed antigen of Claim 1 of the ’052 patent,” plaintiffs answered, “No.” (Exh. 30, Interrog.# 18.) Thus plaintiffs have admitted that the claims of the ’080 patent cover only the “purified” material called for in Claim 1 of the ’052 patent. Second, defendants point out that all of the original claims of the application for the ’080 patent specifically required a “purified” antigen and, during the prosecution of that application, Novotny repeatedly described the claimed antigen as “purified.” For example, in an amendment filed April 7, 1995, he stated: “The present invention is concerned with use of a purified B. pertussis antigen for inducing an immune response.” (Exh. 34 at 3; emphasis added.) However, two months later, on June 19, 1995, when Novot-ny substituted new claims which did not contain the “purified” limitation, he made no mention of that omission. (See Exh. 35 at 7.) The Examiner likewise never expressly noted the omission, perhaps because the applicant continued to emphasize that the application covered “one purified protein, 69 kD antigen.” (Deck of Peter J. Cozens, Exh. 36 ¶ 5; emphasis added.) One fact on which defendants do not rely in arguing for construction of the ’080 patent to require a “purified” antigen, but which strongly argues for such a construction, is that the specification of the ’080 patent is identical to that of the ’052 and ’120 patents, and contains all the teachings previously discussed. These include the statements that ACAP may be “the most important factor in immunity” and “the key antigen necessary for eliciting an immune response,” and that although “vaccine formulations according to the invention may, if desired, contain minor quantities of other antigenic compounds, in addition to the ACAP,” they are “preferably substantially free from other antigenic components.” Moreover, as previously noted, in all.of the illustrative formulations disclosed in the specification, ACAP is the only antigenic component listed. An even more persuasive fact, likewise not relied on by defendants in support of this' aspect of their motion, is that if the claims of the ’080 patent are not construed as being limited to “purified” proteinaceous material, their patentability over the prior art cited against the application for that patent, including the Dahlstrom patent and the Novot-ny & Cownley article, would be in serious question. Before the “purified” limitation was added to the claims of the applications for the ’052 and ’120 patents, those claims differed from those of the ’080 patent only in that the latter claims are directed to a vaccine formed by incorporating the antigen of the ’052 and 120 patents in “a pharmaceuti-cally acceptable carrier or adjuvant.” Such use of the antigen is clearly not patentable over the disclosure of the ’052 and 120 patents, which taught that precise use as the end purpose of the antigen. Thus the claims of the application for the ’080 patent were rejected by the PTO Examiner for obviousness-type double patenting over the ’052 and 120 patents, until Novotny overcame that rejection by pointing out that the application for the ’080 patent was a division of the same parent application on which the ’052 and 120 patents were based. (See Exh. 37 at 16.) As previously noted, the claims of the applications for the ’052 and 120 patents were rejected as anticipated by or obvious in view of the cited art, a rejection in which Novotny acquiesced by amending the claims to limit them to a “purified” antigen. It therefore appears clear that if the claims of the ’080 patent were not similarly limited, they would be equally vulnerable to invalidation as unpatentable over that same art. It is a well-established rule of claim construction that claims should be interpreted, if possible, so as to preserve them validity. See Amhil Enters. Ltd. v. Wawa, Inc., 81 F.3d 1554, 1561 (Fed.Cir.1996). In this case, that rule impels the Court strongly toward a limiting interpretation of the claims of the ’080 patent. Notably, in their memorandum in opposition to defendants’ motion for a Marhman ruling, plaintiffs make no response whatsoever to defendants’ argument for construction of the claims of the ’080 patent to require a “purified” proteinaceous material. Perhaps this is because they recognize that, for the reasons discussed above, such a construction is necessary to preserve the validity of the claims. Considering all of these factors, the Court concludes that the claims of the ’080 patent must be construed to limit them, like those of the other two patents in suit, to a vaccine in which the extracted proteinaceous material is purified at least to the extent that 69k constitutes its major antigenic component. C. What is the meaning of “substantially 1:1” and “about 1:1” in the context of the patents in suit? Each of the claims of the ’052 and ’120 patents calls for an antigen in which the ratio of proline to glutamic acid is “substantially 1:1,” while each of the claims of the ’080 patent describes that ratio as “about 1:1.” There can be no dispute as to the importance of this limitation in the claims, for it was added to the broadest claims during prosecution to overcome their rejection as unpatentable over the Hewlett & Wolff article, and the amended claims were allowed only after the filing of Hewlett’s declaration attesting that the adenylate cyclase disclosed in the article did not have a prolinerglutamic acid ratio of “substantially 1:1.” Both parties apparently agree that the terms “substantially” and “about” have the same meaning in the context of the patents in suit but, as might be expected, they differ considerably as to what that meaning is. Plaintiffs argue that “[tjhese are not terms which would have a unique meaning in the art to which the patent relates” and that they merely “indicate that the ratio of proline to glutamic acid explicitly does not have to be exactly 1:1.” (Pls.’ Mem. in Opp. to Non-infringement/Markman Mot. at 17.) Plaintiffs accordingly argue that it should be left for the jury to decide whether the 0.86:1 ratio of defendants’ products is “substantially” or “about” 1:1. Defendants urge the Court to interpret the terms in light of both intrinsic and extrinsic evidence. First, defendants point out that the only guide in the specification to the meaning of the terms is found in “Example 6,” which involves amino acid analysis of ACAP. This analysis found proline and glu-tamic acid (including glutamine)- residues of 60 and 62 respectively — a ratio of 0.97:1. Defendants also rely upon the deposition of Dr. Erik Hewlett, whose declaration was submitted to the PTO to overcome a rejection of the claims as unpatentable over two articles co-authored by him. In that declaration, Dr. Hewlett stated that his tests of a 70kD portion of the adenylate cyclase toxin disclosed in the aforementioned Hewlett & Wolff article (curiously, not one of the two articles cited in the rejection) was subjected to amino acid analysis and found to have a proline:glutamic acid ( +glutamine) ratio of 0.32:1. Thus plaintiffs have presented Dr. Hewlett as an expert in this branch of protein analysis. In his deposition, Dr. Hewlett expressed his understanding, as one skilled in the art, of the meaning of the term “substantially 1:1”: Q: ... At the time you were executing this declaration, I want to find out if you had some understanding in your mind what you as a person skilled in the art thought that claim meant? A: Yes. Q: Okay. And in particular what “substantially 1 to 1” meant to you. A: I can tell you what it meant to me. Q: Okay. A: Because I was— Q: Why don’t you— A. —making the interpretation. Substantially I interpreted as being the equivalent of statistically — not statistically significantly different than 1, which is essentially plus or minus 5 percent. (Hewlett Dep., Aug. 13,1997, Exh. 54 at 208-09.) In their briefs, plaintiffs attempted to stretch this range to twice its width by interpreting it as contemplating “5 percent plus or minus per amino acid ” (emphasis added), so that it would encompass the entire span from 0.9:1 to 1.1:1. That is a creative approach, but is inappropriate in this context. It would be logically applicable if there were a potential error of plus or minus 5% in measuring the value of each of the two amino acids, so that one might be 5% too high and the other 5% too, low, for a possible spread of 10%. However, the question put to Dr. Hewlett did not contemplate measurement tolerances, but merely asked what the expression “substantially 1:1” in the patent claims means to him as a person skilled in the art.' His testimony is therefore subject to only one reasonable interpretation: , that “substantially 1:1” means within plus or minus 5% of equality— that is, between 0.95:1 and 1.05:1. Moreover, even if the Court were to accept plaintiffs’ imaginative approach and construe “substantially 1:1” as covering any ratio between 0.90:1 and 1.1:1, it still would not encompass the proline:glutamic acid ratio of 0.86:1 in defendants’ antigen. Realizing this, plaintiffs’ counsel strained his ingenuity still further- and, at oral argument, contended that the plus or minus 10% range should be measured from the “starting point” of the 0.97:1 ratio calculated from “Example 6” of the specification. This would extend the range all the way down to 0.87:1, which plaintiffs argue is so close to defendants’ 0.86:1 ratio that a jury could find the difference “insubstantial.” (Tr. at 42-45.) In his desperate advocacy, plaintiffs’ able counsel ignored the fact that the patent claims specify a ratio of “substantially 1:1,” not “substantially 0.97:1.” Thus the “starting point” from which to measure whether there is “substantial” equality is clearly 1:1, not 0.97:1. The only intrinsic evidence as -the meaning of the terms “substantially 1:1” and “about 1:1” is the one example given in the specification, in which the ratio of the residue values of the two amino acids is 60:62 (or 0.97:1), which is to say that they are within 3% of one another. And the only extrinsic evidence as to such meaning which has been presented to the Court is the testimony of plaintiffs’ own expert that the terms mean to him, as a person skilled in the art, that the residue values of the two amino acids are within 5% of each other. Faced with such evidence, plaintiffs understandably urge the Court to make no Mark-man ruling as to the meaning of the terms in question and leave this determination for the jury. However, the Court believes that course of action would be an abdication of the responsibility which Markman places on the Court. The jury is not likely to have any better evidence as to the meaning of the terms than is now available to the Court. At the trial, each side would surely present its own partisan expert to testify as to what the claims mean to those skilled in the art, and surely the plaintiffs’ expert would assign to the terms a range of ratios broad enough to include the 0.86:1 proline:glutamic acid ratio in defendants’ antigen, while defendants’ equally eminent expert would narrow the range to exclude defendants’ antigen. With those contradictory opinions offsetting one another, the issue would have to be determined on the basis of the evidence now available: the intrinsic evidence of the example given in the specification (±3%) and the extrinsic evidence of the testimony against interest of defendants’ expert (±5%). In light of such evidence, the Court concludes that the terms “substantially 1:1” and “about 1:1” mean that the proline and glu-tamic acid values are within 5% of each other — that is, the ratio between these values must fall within the range of 0.95:1 to 1.05:1. There is no evidence in the record, intrinsic or extrinsic, which would support any broader reading of the language. D. Should the claims be construed to require “adenylate cyclase activity”? All of the claims of the ’052 and 120 patents are directed to “A purified Bordetella pertussis antigen,” while all of the claims of the ’080 patent call for “a vaccine which comprises a proteinaceous material.” Defendants contend that even though none of the claims specifically requires adenylate cyc-lase activity, the terms “antigen” and “protei-naceous material” should be construed, in light of the specification, as limited to a material having adenylate cyclase activity. Defendants base this contention principally upon the specification’s statement that: ... Control vaccines containing no adeny-late cyclase activity were found to provide little or no protection against challenge with B. pertussis, suggesting that ACAP may, in fact, be the most important factor in immunity. (’080, 4/18-21; emphasis added.) Defendants further point out that in all of the illustrative vaccines disclosed in the specification, “antigen according to the invention” is the only immunogenic material listed. Thus, they reason, the patents teach that the claimed “antigen” or “proteinaceous material” must have adenylate cyclase activity. In opposition, plaintiffs point out that the specification refers to: ... This discovery that proteinaceous material usually associated with adenylate cyclase activity is a major protective antigen against B. pertussis_ (’052, 2/33-35; emphasis added.) Plaintiffs argue that the use of the word “usually” means that the claimed antigen need not have adenylate cyclase activity. However, the force of this argument is somewhat weakened by the fact that the sentence immediately preceding the one partially quoted by plaintiffs, which identifies what is referred to as “[t]his discovery,” does not contain the word “usually”: It has now been discovered that certain proteinaceous material, associated with adenylate cyclase activity, as hereinafter described, found in the cultures of B. pertussis, is capable of providing protection against challenge by B. pertussis when administered to experimental animals. (’052, 2/28-33.) Moreover, immediately after stating that ACAP may be “the most important factor in immunity,” the specification adds: ... Analysis of batches of non-protective whole-cell vaccine has also shown that non-protection tends to be associated with á lack of adenylate cyclase activity, further suggesting that ACAP may be the key antigen necessary for eliciting an immune response against B. pertussis. (’052, 4/45-50; emphasis added.) The use of the words “key” and “necessary” strongly suggests that adenylate cyclase activity is not a merely optional characteristic of the antigen. Finally, as previously discussed, in the U.K. application on which plaintiffs’ claim of priority is based, the patented antigen was not described as merely having adenylate cyclase activity but as being adenylate cyc-lase per se. Before the parent U.S. application was filed, Novotny had talked to others, including Dr. Hewlett, and concluded that the critical antigen might not be adenylate cyclase but could be either “the adenylate cyclase enzyme per se or a binding protein for the enzyme.” (’052, 2/48^49.) Now plaintiffs seek to broaden the scope of their protection even further beyond the pat-entee’s original concept and the teachings of his specification to cover antigens and vaccines that do not even have adenylate cyclase activity. The broader coverage they now seek not only is unsupported by the teaching of the specification, as required by 35 U.S.C. § 112, but is flatly contradicted by that teaching. In Genentech, Inc. v. Wellcome Found. Ltd., 29 F.3d 1555 (Fed.Cir.1994), the Court ruled that the terms of the patent claims, even though worded broadly enough to cover many permutations, should be construed to cover only those permutations which are enabled by the specification: ... As Dr. Goeddel testified, an infinite number or permutations of natural t-PA are covered by these other definitions. Many of these permutations would not be capable of binding to fibrin and would thus be inoperative. There is no basis provided in the specification for determining which of these permutations are operative and which are not.... Thus we are unwilling to say that the specification satisfies the enablement requirement of 35 U.S.C. § 112 ¶ 1 (1988) with respect to these broader definitions, or that the PTO could have relied on these definitions in issuing the patent.... We conclude therefore that the phrase “human tissue plasminogen activator” appearing in the ’075 and ’330 claims means natural t-PA. Id. at 1564-65 (internal citations omitted). Here, the circumstances much more strongly favor a narrow reading of the claims, because the specification not only fails to teach the preparation of antigens having no adenylate cyclase activity, but affirmatively teaches that such antigens have little or no immunogenic efficacy. The very essence of the specification’s message is that a protein having adenylate cyclase activity, or ACAP, may be “the key antigen necessary for eliciting an immune response against B. pertussis,” and that vaccines without it “provide little or no protection against challenge with B. pertussis.” The claims surely cannot validly be interpreted to cover materials which the specification expressly declares to be inoperative. There would be no quid pro quo for granting a patentee the right to exclude others from using something that his specification clearly instructs them not to use. "At the oral argument, plaintiffs’ counsel attempted to downplay the importance of adenylate cyclase activity by making the startling assertion that none of the vaccines actually being produced and used has adeny-late cyclase activity because all of it is destroyed in the purification process. (Tr. at 98.) He argued that Novotny’s specification discusses adenylate cyclase activity not because it is needed in the vaccine, but only because it “is present at some stages so you know you’re on the right path.” (Tr. at 99.) He added that Novotny did not say “a single word ... that tells the worker to test for adenylate cyclase activity.” (Tr. at 99-100.) Defendants’ counsel immediately corrected the latter assertion by pointing out that the specification states: “Adenylate cyclase activity was assayed by the method of Hewlett, E., and Wolff, J. (J. BacteRIol.(1976) 127, 890-898).” (See Tr. at 102-03 (citing ’052, 2/50-52).) Plaintiffs’ argument is nothing more than a desperate attempt to disown the core teaching of Novotny’s specification. Novotny taught not only that “[c]ontrol vaccines corn taining no adenylate cyclase activity were found to provide little or no protection against challenge with B. pertussis ” and that “ACAP may be the key antigen necessary for eliciting an immune response against B. pertussis,” but also that “ACAP may, in fact, be the most important factor in immunity” and that “non-protection tends to be associated with a lack of adenylate cyclase activity.” (’052, 4/41-50.) Moreover, nowhere in Novotny’s specification is there the slightest suggestion that adenylate cyclase activity is destroyed in the purification process, or that it is unnecessary in the final vaccine. On the contrary, in its discussion of each successive step of purification, the specification expressly states that both the objective and the result of the step were to purify the ACAP (ie., adenylate cyclase activity protein) which had been extracted from the outer membrane of B. pertussis cells. For example, in describing the result of the first purification step, “Example 2(a),” the specification states: “The ACAP was present in the material unretarded by the column, as shown by SDS-PAGE, but was also present in the retarded material eluted by 0.2M NaCl.” (’052, 6/37-40.) In describing the next purification step, “Example 2(b),” the specification states: “The ACAP was detectable as two bands, one of pi 7.0, and the other (diffuse) band of pi 7.2-7.4. Adenylate cyclase activity was associated almost entirely with the central band (pi 7.0) but monoclonal antibodies to ACAP bound both bands strongly.” (’052, 7/3-7.) The final purification step, “Example 3,” is entitled “Purification of ACAP Using a Monoclonal Immunosorbent Column,” and the objective of this procedure is described as “[t]o separate the ACAP on the immunosorbent column ....” (’052, 7/57-58.) In view of the undisputed fact that defendants’ ACEL-IMMUNE® vaccine has no adenylate cyclase activity, it is understandable that plaintiffs would strain to rewrite Novotny’s specification, but they cannot discard it altogether because its teaching was the sole consideration for the patent grant. If what plaintiffs now say about the irrelevance of adenylate cyclase activity is correct (as it apparently is), there was a total failure of consideration. The Court therefore concludes that the terms “antigen” and “proteinaceous material” in t